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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
dd2e53e6-9232-4f51-9a42-ee9f8cb6dc81 | simulate-time-integrated-coarse-grained | 2204.10348 | null | https://arxiv.org/abs/2204.10348v2 | https://arxiv.org/pdf/2204.10348v2.pdf | Simulate Time-integrated Coarse-grained Molecular Dynamics with Geometric Machine Learning | Molecular dynamics (MD) simulation is the workhorse of various scientific domains but is limited by high computational cost. Learning-based force fields have made major progress in accelerating ab-initio MD simulation but are still not fast enough for many real-world applications that require long-time MD simulation. In this paper, we adopt a different machine learning approach where we coarse-grain a physical system using graph clustering, and model the system evolution with a very large time-integration step using graph neural networks. A novel score-based GNN refinement module resolves the long-standing challenge of long-time simulation instability. Despite only trained with short MD trajectory data, our learned simulator can generalize to unseen novel systems and simulate for much longer than the training trajectories. Properties requiring 10-100 ns level long-time dynamics can be accurately recovered at several-orders-of-magnitude higher speed than classical force fields. We demonstrate the effectiveness of our method on two realistic complex systems: (1) single-chain coarse-grained polymers in implicit solvent; (2) multi-component Li-ion polymer electrolyte systems. | ['Tommi Jaakkola', 'Bradley D. Olsen', 'Nathan J. Rebello', 'Tian Xie', 'Xiang Fu'] | 2022-04-21 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [ 1.35253906e-01 -4.65442568e-01 -8.68071392e-02 -7.16357306e-02
-8.24834824e-01 -6.31738842e-01 4.77981538e-01 5.26350260e-01
-5.08345902e-01 1.17918646e+00 -5.35090327e-01 -8.92419577e-01
-8.72169361e-02 -7.25012779e-01 -8.66519630e-01 -1.40245855e+00
-4.48786765e-01 1.12530804e+00 3.41318637e-01 -4.27726179e-01
3.50025237e-01 9.53601658e-01 -1.14738929e+00 -1.00226544e-01
1.33101809e+00 2.73472548e-01 1.02174751e-01 8.76388013e-01
2.18714342e-01 3.49683374e-01 -5.07075489e-02 3.69298942e-02
-8.88646245e-02 -5.42609453e-01 -4.70655650e-01 -3.53659600e-01
2.93238580e-01 1.69866636e-01 -6.78912640e-01 1.17296553e+00
6.45662844e-01 5.19467592e-01 7.25588679e-01 -6.82258010e-01
-2.38768756e-01 4.01526451e-01 -1.55061305e-01 -1.05560664e-02
1.57957211e-01 7.40792572e-01 4.92231578e-01 -5.13728499e-01
1.00170028e+00 1.13399100e+00 7.14923501e-01 6.70507312e-01
-1.70320094e+00 -7.88174927e-01 1.12758830e-01 3.93284298e-02
-1.18719935e+00 5.02483025e-02 6.82440758e-01 -4.23326135e-01
1.55589366e+00 -1.03835464e-01 9.78618026e-01 8.51890862e-01
1.05647802e+00 -2.53086597e-01 1.26809096e+00 -8.79193693e-02
7.41929293e-01 -5.20101130e-01 3.42372775e-01 5.56506574e-01
5.31421363e-01 4.92682487e-01 -6.98413625e-02 -7.99778104e-01
6.70763135e-01 1.48347199e-01 -3.25888455e-01 -5.60189307e-01
-1.08633292e+00 1.01085472e+00 2.66471654e-01 -1.65688947e-01
-3.53044987e-01 3.61264288e-01 4.70407844e-01 2.26682976e-01
3.52901131e-01 6.35647357e-01 -4.05875713e-01 -3.62794399e-01
-1.02172244e+00 1.11399126e+00 1.06509519e+00 3.17844570e-01
4.92894650e-01 2.18404472e-01 6.87741280e-01 -1.43665131e-02
2.57354945e-01 7.04652071e-01 2.05247059e-01 -9.00615752e-01
8.76866728e-02 3.15703660e-01 2.75321901e-01 -1.27932236e-01
-5.21128356e-01 -2.09696993e-01 -1.03469563e+00 9.03095305e-01
4.35109794e-01 -1.10463075e-01 -1.15554333e+00 1.38213718e+00
6.26326382e-01 -3.23839448e-02 5.56766689e-02 8.24411154e-01
3.47092390e-01 9.62743282e-01 2.16572434e-01 -7.19117045e-01
7.95125544e-01 -8.67722034e-01 -4.58723962e-01 2.44884029e-01
6.21907353e-01 -4.94196147e-01 7.56050467e-01 7.22515643e-01
-1.42244971e+00 -2.63544887e-01 -1.28879845e+00 1.35536879e-01
-3.22562486e-01 -7.18064964e-01 7.49120712e-01 3.68694484e-01
-7.63904274e-01 1.79316998e+00 -1.33788276e+00 -2.01466292e-01
-1.02818441e-02 9.01374757e-01 -3.67663175e-01 -1.34303108e-01
-9.61834967e-01 9.51474011e-01 3.24223697e-01 -2.56173849e-01
-1.01501083e+00 -8.26629996e-01 -5.80695271e-01 -3.30497354e-01
8.67134407e-02 -8.42246652e-01 1.05599916e+00 -4.49042976e-01
-1.62829423e+00 2.85789460e-01 -4.81867880e-01 -4.80372697e-01
4.85896945e-01 2.20038235e-01 -4.83987778e-01 -7.76813105e-02
-2.83584237e-01 1.74227551e-01 5.69032550e-01 -1.19859135e+00
2.70090491e-01 -3.41178447e-01 -4.44862932e-01 1.65013790e-01
5.30513108e-01 -2.71475315e-01 3.07995200e-01 -3.03451270e-01
2.44520009e-01 -1.55042458e+00 -9.56376791e-01 -3.24347883e-01
-1.19032793e-01 -1.44725934e-01 6.99413955e-01 -3.53719145e-01
1.01972926e+00 -1.43591428e+00 6.85173631e-01 3.57414931e-01
5.60121953e-01 6.62080884e-01 1.88196868e-01 1.02377975e+00
-3.59053075e-01 -4.42232862e-02 -5.64523816e-01 -1.36790723e-02
-1.88834280e-01 2.74430275e-01 -3.29344779e-01 5.17767608e-01
-2.08169475e-01 9.90181684e-01 -1.00200176e+00 -3.79506946e-02
-2.75034597e-03 6.08365595e-01 -4.44813848e-01 1.16427250e-01
-6.65238500e-01 9.95456934e-01 -2.55905181e-01 3.78052562e-01
9.21387792e-01 -6.47094607e-01 4.58034962e-01 -1.47690460e-01
-4.27741766e-01 3.93299222e-01 -8.75427902e-01 1.65585387e+00
2.81726241e-01 2.68517733e-02 9.87051055e-03 -9.16474700e-01
6.00442410e-01 4.21419516e-02 4.42251474e-01 -6.01985455e-01
-3.35417758e-03 5.53651989e-01 7.86379516e-01 -3.41122925e-01
3.62162679e-01 -7.28070676e-01 -6.06513862e-03 6.67295575e-01
-1.81154519e-01 -5.85565805e-01 2.17977434e-01 4.15322989e-01
1.18492043e+00 5.06952628e-02 2.05600888e-01 -5.13779998e-01
1.52315944e-01 5.10028601e-01 3.24100941e-01 7.31508672e-01
-9.84799042e-02 2.54985332e-01 3.56548607e-01 -6.23237848e-01
-1.71911919e+00 -1.08609116e+00 -1.70801148e-01 6.29276812e-01
2.81923532e-01 -6.48167849e-01 -6.84763134e-01 -1.82147607e-01
5.56277692e-01 3.52133989e-01 -3.10904473e-01 -2.39629984e-01
-1.00369680e+00 -1.27565908e+00 2.51235008e-01 1.95207164e-01
-2.15851232e-01 -1.27224886e+00 2.63194274e-02 8.34119856e-01
7.03846455e-01 -6.80622697e-01 -1.77906930e-01 3.83099616e-01
-1.23711669e+00 -9.63404715e-01 -5.46727121e-01 -4.76318657e-01
5.72900593e-01 1.23288311e-01 1.01179755e+00 1.72042698e-01
-5.71557045e-01 -2.19628498e-01 1.02659017e-01 1.56378642e-01
-9.53241765e-01 5.86416796e-02 5.95292449e-01 -8.64537060e-01
1.87867388e-01 -1.07302940e+00 -7.47581422e-01 -6.17152713e-02
-6.71329618e-01 -6.98792040e-02 3.13282281e-01 7.90619433e-01
8.92002881e-01 2.86615938e-02 5.71444213e-01 -7.02281892e-01
7.83327222e-01 -5.25864124e-01 -8.47787678e-01 -2.27331296e-02
-9.36723948e-01 3.98187280e-01 8.96117508e-01 -9.70454574e-01
-6.40801668e-01 -1.14642773e-02 -5.95100522e-01 -2.10900709e-01
-6.16322160e-02 5.45423806e-01 1.12489603e-01 -8.16805065e-01
5.29543638e-01 4.29288507e-01 3.09749901e-01 -5.14527321e-01
1.53862908e-01 4.56012264e-02 3.53887886e-01 -1.05645418e+00
8.24438453e-01 3.90244812e-01 5.55963814e-01 -7.66691625e-01
-4.53308551e-03 -8.47215112e-03 -7.19776094e-01 1.48177832e-01
8.07665586e-01 -7.66167283e-01 -1.57361245e+00 6.51977003e-01
-9.90907729e-01 -8.18829834e-01 -1.17622852e-01 5.28527498e-01
-3.04190010e-01 6.90914571e-01 -9.76231635e-01 -7.03823149e-01
-5.76195061e-01 -1.46208405e+00 6.70263171e-01 2.16556385e-01
-2.17267692e-01 -1.00154936e+00 8.21133614e-01 -1.73303872e-01
4.90327001e-01 6.18120193e-01 1.27779400e+00 -2.92762041e-01
-7.06521332e-01 -8.91323835e-02 3.33111100e-02 -1.52877301e-01
-2.41875023e-01 2.66778946e-01 -3.28435600e-01 -9.44051504e-01
-6.18844554e-02 -1.13109455e-01 9.85434413e-01 4.46575314e-01
9.24469769e-01 -4.40454260e-02 -8.20374131e-01 7.35701621e-01
1.45325792e+00 6.59009576e-01 4.67758656e-01 -1.04102641e-01
1.18786824e+00 1.46895453e-01 3.51127535e-01 1.61254629e-01
2.87425548e-01 5.21556318e-01 7.07043633e-02 3.89755704e-02
4.06661592e-02 -1.34814993e-01 5.17820537e-01 1.33511198e+00
-6.14564478e-01 -2.04065457e-01 -1.05713785e+00 -2.44100112e-02
-1.78334415e+00 -1.06287563e+00 -4.70010906e-01 2.19674778e+00
9.79801953e-01 3.88645649e-01 3.40124071e-01 1.63201429e-02
3.21956038e-01 -1.10118844e-01 -1.37318087e+00 -5.82201004e-01
-2.22901236e-02 5.37370861e-01 5.76107204e-01 8.82979274e-01
-6.99284136e-01 9.89287376e-01 6.41705179e+00 7.77835608e-01
-1.62651408e+00 1.75001532e-01 2.07581475e-01 -1.63908586e-01
-2.83439904e-01 3.21863800e-01 -1.02300823e+00 6.71372473e-01
1.43447566e+00 -1.77675486e-01 6.73051596e-01 6.26376927e-01
3.66200358e-01 -2.23112963e-02 -9.52831447e-01 6.41656995e-01
-3.92269194e-01 -1.71288788e+00 1.13546876e-02 4.27602202e-01
8.58322620e-01 4.87045616e-01 -1.09218396e-01 5.30508049e-02
5.78597009e-01 -1.25544870e+00 3.14682811e-01 3.69288355e-01
1.01762307e+00 -8.54335368e-01 2.88798153e-01 5.29189527e-01
-1.09279883e+00 4.70149100e-01 -4.88084257e-01 -3.17564428e-01
6.37188017e-01 6.56445444e-01 -3.92625630e-01 1.38716966e-01
1.89912111e-01 6.20985806e-01 -2.72593051e-02 1.01518118e+00
3.14496487e-01 8.24838340e-01 -4.49046135e-01 -3.78193825e-01
3.59625816e-01 -8.58554840e-01 6.84692562e-01 1.05455256e+00
2.90661186e-01 4.56978679e-01 9.81918648e-02 7.94135630e-01
1.24521933e-01 -5.43727815e-01 -5.84301829e-01 -4.39939767e-01
1.56622782e-01 9.99020159e-01 -8.71569693e-01 -3.58245671e-01
1.04821570e-01 7.86256254e-01 3.60252529e-01 5.41513324e-01
-8.96992445e-01 -3.96851361e-01 9.82580900e-01 4.22263771e-01
3.85320514e-01 -1.03183341e+00 1.71135277e-01 -1.28850830e+00
-3.60872000e-01 -8.95684958e-01 -3.29598218e-01 -3.59710455e-01
-1.45375013e+00 5.37539899e-01 -3.73208284e-01 -8.21123123e-01
-2.87074089e-01 -7.71376550e-01 -7.20108509e-01 9.20997858e-01
-1.38958585e+00 -7.36453116e-01 1.11624837e-01 3.50131184e-01
9.08092558e-02 1.37304991e-01 1.05989838e+00 2.42388844e-01
-4.24716681e-01 2.09026456e-01 9.25325394e-01 -5.58566570e-01
6.21176481e-01 -1.15947211e+00 9.57812905e-01 2.50054002e-01
-5.51732123e-01 9.59744215e-01 1.11156499e+00 -1.42689586e+00
-2.04105544e+00 -9.40311909e-01 5.07045448e-01 -2.85110891e-01
9.04216528e-01 -6.49463832e-01 -1.68374395e+00 4.16409135e-01
-8.42097104e-02 2.68576860e-01 2.83294380e-01 -2.61915594e-01
-1.92782491e-01 3.48778933e-01 -1.12283051e+00 3.98607045e-01
1.16953421e+00 -5.63747585e-01 -2.05607876e-01 6.06097639e-01
8.52189898e-01 -6.50430918e-01 -1.06449509e+00 3.63044202e-01
6.21036232e-01 -5.95531821e-01 1.23023760e+00 -1.04001057e+00
1.81177810e-01 -3.47691029e-01 3.10975105e-01 -1.20108497e+00
-3.72989178e-01 -1.11519241e+00 -4.47325587e-01 4.57617968e-01
2.95665145e-01 -9.52281892e-01 9.10325527e-01 6.84944928e-01
-1.61844939e-01 -9.69636917e-01 -8.28045249e-01 -9.12989497e-01
7.94334054e-01 2.07063612e-02 4.11303550e-01 8.50825191e-01
2.33670831e-01 2.36567318e-01 -1.02483913e-01 5.66731207e-02
1.06432140e+00 4.11812395e-01 4.30269450e-01 -1.25814831e+00
-4.32286710e-01 -2.89661378e-01 -1.64390862e-01 -6.79067791e-01
3.06196451e-01 -9.78503168e-01 -1.44752115e-01 -1.15848494e+00
4.71448243e-01 -5.92266321e-01 1.84044186e-02 -1.85704336e-01
-1.56812623e-01 -6.28569424e-02 -2.49926131e-02 4.40829098e-01
-6.22323275e-01 6.71424806e-01 1.25486970e+00 1.22171655e-01
-2.69813061e-01 -1.38724118e-01 -1.37790311e-02 5.24892747e-01
6.38060331e-01 -7.88539350e-01 -1.36890635e-01 2.53407866e-01
3.91098619e-01 3.43181312e-01 3.90656352e-01 -1.15657580e+00
3.98177534e-01 -3.96963060e-01 2.04077363e-01 -7.18420029e-01
4.99069512e-01 -3.12051862e-01 5.67482710e-01 9.23457980e-01
-4.95722443e-02 3.63987982e-01 3.06398183e-01 6.50182843e-01
1.77504882e-01 4.62027229e-02 9.98198271e-01 -2.28452355e-01
-1.04274765e-01 6.95264637e-01 -6.27873600e-01 -1.69312209e-01
8.66056979e-01 5.64400218e-02 -4.50426459e-01 -6.99156374e-02
-8.59155953e-01 -1.71714678e-01 1.22383881e+00 -2.79100448e-01
4.69757497e-01 -1.12311840e+00 -3.02975535e-01 1.04292750e-01
-3.98768306e-01 9.46826488e-02 4.60160255e-01 6.56483710e-01
-8.68035197e-01 3.78618896e-01 -1.83095664e-01 -4.96202826e-01
-1.17509091e+00 6.49961412e-01 2.74382830e-01 -6.55604005e-01
-8.28456461e-01 5.32847941e-01 -1.12646207e-01 -5.30761540e-01
-3.59408170e-01 -2.28527263e-01 5.56671560e-01 -4.58308637e-01
3.09777796e-01 4.31532770e-01 1.07593521e-01 -5.91717780e-01
-1.80413604e-01 9.71328557e-01 -4.12028551e-01 -8.72979034e-03
1.63965857e+00 3.48058462e-01 -2.46328831e-01 3.45609307e-01
1.14252365e+00 -1.34777844e-01 -1.62679935e+00 1.87163670e-02
-2.60842621e-01 8.19163099e-02 -1.58570349e-01 -5.57559431e-01
-3.41112524e-01 1.19712961e+00 5.53599894e-01 1.18320562e-01
2.72063613e-01 -1.12644479e-01 1.23956358e+00 6.12128615e-01
5.73973715e-01 -6.34493768e-01 -2.42113054e-01 7.64972210e-01
3.86293441e-01 -1.20874512e+00 3.76617938e-01 -1.22594215e-01
-1.65872633e-01 1.27402675e+00 3.12719941e-01 -5.59900582e-01
4.65580612e-01 5.54481149e-01 -4.56135720e-01 -4.46991354e-01
-7.57587612e-01 4.26520467e-01 1.26530215e-01 3.81189018e-01
2.40606576e-01 2.50502139e-01 -1.98888853e-01 3.74700308e-01
1.14767760e-01 -8.87586623e-02 5.45026541e-01 1.09651649e+00
-6.53020561e-01 -1.49814939e+00 -1.15155198e-01 2.88920432e-01
-8.64167139e-02 -1.16794199e-01 -3.23209986e-02 7.49147117e-01
-2.38931477e-01 2.52473235e-01 -1.50760323e-01 -1.23582266e-01
-1.80895440e-02 2.59420663e-01 8.14222097e-01 -3.52297902e-01
-6.64956987e-01 -1.18769564e-01 -1.45270154e-01 -5.76413691e-01
-1.07843131e-01 -7.58830726e-01 -1.80672693e+00 -1.20608103e+00
-2.51624823e-01 4.86980319e-01 8.64041448e-01 9.46409166e-01
6.61487758e-01 2.56552160e-01 1.51680127e-01 -1.68590069e+00
-5.96164167e-01 -5.69144368e-01 -6.17013872e-01 3.60135883e-01
6.18545532e-01 -6.19978726e-01 -5.23084044e-01 -3.27070773e-01] | [5.074577331542969, 5.25032377243042] |
171917a9-0704-487b-8899-fb16bf02435b | machine-learning-in-downlink-coordinated | 1608.08306 | null | http://arxiv.org/abs/1608.08306v6 | http://arxiv.org/pdf/1608.08306v6.pdf | Machine Learning in Downlink Coordinated Multipoint in Heterogeneous Networks | We propose a method for downlink coordinated multipoint (DL CoMP) in
heterogeneous fifth generation New Radio (NR) networks. The primary
contribution of our paper is an algorithm to enhance the trigger of DL CoMP
using online machine learning. We use support vector machine (SVM) classifiers
to enhance the user downlink throughput in a realistic frequency division
duplex network environment. Our simulation results show improvement in both the
macro and pico base station downlink throughputs due to the informed triggering
of the multiple radio streams as learned by the SVM classifier. | ['Brian L. Evans', 'Faris B. Mismar'] | 2016-08-30 | null | null | null | null | ['pico'] | ['natural-language-processing'] | [-2.01238245e-01 2.32819542e-01 -8.99144828e-01 -4.39723670e-01
-5.47073364e-01 -6.49125755e-01 1.27309710e-01 -5.72828114e-01
2.87728813e-02 2.09583473e+00 -3.27806175e-01 -8.63962948e-01
-2.81014293e-01 -7.84080565e-01 -8.57974738e-02 -7.53393412e-01
-6.46715283e-01 4.70958382e-01 5.84473461e-02 -5.70888638e-01
-2.54328400e-01 9.04462159e-01 -6.71512425e-01 3.25617880e-01
7.64837742e-01 1.31007314e+00 7.81334713e-02 1.04513538e+00
3.91445100e-01 5.37167788e-01 -1.21752322e+00 -5.00848554e-02
3.15422744e-01 -2.72886246e-01 -1.91802815e-01 -2.69713402e-01
-2.78605819e-01 -3.60614449e-01 -9.70327258e-01 2.41774291e-01
1.13917804e+00 -2.24993855e-01 6.91292882e-01 -1.70256031e+00
5.50866961e-01 9.23672497e-01 -2.77127028e-01 8.05811703e-01
2.45571822e-01 -3.74520391e-01 9.84156430e-01 -4.99161094e-01
2.34890804e-01 9.77977037e-01 7.04722226e-01 2.88112700e-01
-6.67260766e-01 -9.25418377e-01 -2.24027947e-01 2.25997597e-01
-1.33387506e+00 -7.35510707e-01 2.65969545e-01 7.41190836e-02
8.99051607e-01 5.70763469e-01 7.73818851e-01 5.85300982e-01
4.58867878e-01 8.94434333e-01 3.87051910e-01 -6.13332689e-01
5.70152223e-01 4.13795441e-01 -1.07770354e-01 7.93859422e-01
2.09575146e-01 8.24228898e-02 -5.11850864e-02 -7.13568866e-01
1.09475005e+00 -7.52862990e-01 -1.57655403e-01 4.99605983e-02
-6.80335462e-01 8.04360330e-01 -2.78007239e-02 2.13448212e-01
-3.66427988e-01 1.15200825e-01 5.55614755e-02 6.41464233e-01
3.24387252e-01 5.03784597e-01 -8.14168274e-01 -1.94034055e-01
-1.13802004e+00 7.77252065e-03 1.00525343e+00 1.69669116e+00
1.11771226e-01 5.27450621e-01 -3.07223380e-01 6.46396697e-01
4.54682916e-01 1.00572526e+00 4.78002317e-02 -9.18496609e-01
7.86012769e-01 -5.35044730e-01 4.49011058e-01 -2.92928338e-01
-8.46889615e-01 -1.50215888e+00 -8.67481351e-01 -1.84048459e-01
1.26006156e-01 -1.43384373e+00 -3.56924832e-01 1.18953419e+00
-6.04212359e-02 5.86384952e-01 5.63802838e-01 4.34015721e-01
2.31245100e-01 5.03819466e-01 -3.36692214e-01 -6.16007149e-01
4.69157457e-01 -1.25621045e+00 -1.08812857e+00 -3.33990872e-01
6.59105957e-01 -5.01999974e-01 -4.05760348e-01 4.48911369e-01
-1.24510729e+00 -4.44110364e-01 -1.69104326e+00 9.28616464e-01
1.23096064e-01 3.24188441e-01 9.66958821e-01 1.46919465e+00
-7.44171500e-01 2.74094105e-01 -2.43872583e-01 -2.78529733e-01
6.68161154e-01 9.64607596e-01 3.12598020e-01 2.07407102e-01
-1.59381342e+00 6.58771992e-01 9.00170654e-02 -3.10765296e-01
-1.29249552e-02 -8.16668868e-01 -4.30602729e-01 9.57990587e-02
-3.55372466e-02 -6.66423380e-01 1.41586792e+00 -8.55187833e-01
-1.42501009e+00 -2.17127591e-01 -8.07139054e-02 -6.69090211e-01
4.31392282e-01 1.43082514e-01 -1.17201221e+00 2.49308065e-01
-1.03791691e-01 4.96624075e-02 5.83577335e-01 -1.06244493e+00
-1.62515795e+00 9.13265795e-02 1.00833215e-01 1.57814518e-01
1.06018679e-02 -3.43395501e-01 -2.24598363e-01 -7.33017385e-01
-6.72141090e-02 -7.28094041e-01 -4.31868404e-01 -5.14513493e-01
-3.52534741e-01 8.34198967e-02 8.45636725e-01 -3.75159502e-01
1.45042610e+00 -1.85659528e+00 -7.96316564e-02 7.79222608e-01
-4.23254557e-02 3.14253747e-01 3.36166382e-01 3.51808310e-01
6.35279715e-02 1.33049652e-01 5.26786804e-01 -8.92605409e-02
-3.90749693e-01 2.08186299e-01 -4.12349522e-01 2.40729675e-01
-3.36273581e-01 4.10701573e-01 -8.17634881e-01 -1.43661395e-01
1.95710748e-01 -1.71583146e-01 -3.64873350e-01 -1.35606425e-02
8.34359378e-02 2.78181046e-01 -7.17299402e-01 7.86776602e-01
9.93688405e-01 5.60365319e-02 4.13751155e-01 -4.90551323e-01
7.22259358e-02 -3.86199236e-01 -1.19390476e+00 8.15755129e-01
-6.88418269e-01 6.64097607e-01 -1.33630544e-01 -1.03096807e+00
6.54332936e-01 8.48855615e-01 7.48155475e-01 -4.65158731e-01
2.82847136e-01 1.71736911e-01 -2.26091128e-02 -1.64225653e-01
9.34764072e-02 -2.44093716e-01 -2.53858924e-01 2.04929849e-03
4.28779334e-01 2.78975904e-01 -1.45685628e-01 6.20466061e-02
1.02726626e+00 -5.27625620e-01 7.26121783e-01 -1.40883043e-01
5.36233783e-01 -5.35870790e-01 7.01961815e-01 1.33401573e+00
-7.12573171e-01 6.17322735e-02 5.75409770e-01 5.89377917e-02
-8.97058308e-01 -1.30687368e+00 -2.11106271e-01 1.03732717e+00
1.08172424e-01 3.44696909e-01 -7.67628103e-02 -6.58608198e-01
4.28046912e-01 7.90852427e-01 1.90919682e-01 4.05108854e-02
-2.68982917e-01 -9.55036104e-01 7.90247321e-01 4.36064363e-01
4.01046157e-01 -7.83221498e-02 3.65517795e-01 7.56434202e-01
5.86404875e-02 -2.05511045e+00 4.73138923e-03 4.53756481e-01
-5.68794131e-01 -2.30205983e-01 -7.11018026e-01 -6.82819903e-01
-1.62405044e-01 3.21389049e-01 4.68100607e-01 -1.37924522e-01
5.67943752e-02 2.77028888e-01 -1.89329118e-01 -4.39974874e-01
-2.15947673e-01 5.52543402e-01 7.90717661e-01 -1.42003298e-01
-1.57687008e-01 -6.94395661e-01 -1.24403015e-01 7.45384753e-01
4.01637495e-01 -1.62705272e-01 6.71990752e-01 4.01611149e-01
-7.88780004e-02 3.48412782e-01 1.49047244e+00 -7.08812773e-01
2.21034199e-01 -1.09304655e+00 -6.50243402e-01 2.73594379e-01
-3.56112003e-01 -1.30186677e-01 7.87287533e-01 8.11023414e-02
-9.69519913e-01 -3.30814987e-01 -3.85497898e-01 2.38293022e-01
3.66525501e-01 4.73134428e-01 -4.99296725e-01 -6.39424801e-01
7.01151252e-01 -8.66253227e-02 -2.43964806e-01 7.23001957e-02
7.71710351e-02 1.45994496e+00 1.61357582e-01 -2.27451175e-01
1.05694294e+00 5.83063126e-01 3.21896106e-01 -1.61612403e+00
-9.04116690e-01 -5.43059886e-01 -5.08371413e-01 -4.52229977e-01
6.15872964e-02 -1.27303731e+00 -5.52121758e-01 -2.02758498e-02
-8.09833586e-01 -3.16573888e-01 6.59930110e-01 7.57796347e-01
-6.88732445e-01 2.39521638e-01 -3.98256093e-01 -9.80664849e-01
-1.40662745e-01 -4.94821012e-01 5.82067132e-01 6.72274590e-01
-1.04857512e-01 -1.07279253e+00 -4.87460434e-01 2.12749809e-01
6.32896185e-01 6.44913688e-02 7.73477733e-01 -3.02551568e-01
-3.81867111e-01 -3.03592384e-01 -2.69735396e-01 -9.41383690e-02
3.26080620e-01 -2.15899646e-01 -9.25767183e-01 -6.11695886e-01
-4.03605253e-01 2.75723666e-01 5.34739252e-03 8.59337628e-01
9.19837177e-01 1.04719914e-01 -1.14306831e+00 6.58472717e-01
1.47250569e+00 5.12716949e-01 8.55665863e-01 7.39699602e-02
8.25267583e-02 -3.03771168e-01 1.26445830e+00 1.02081358e+00
-1.55950226e-02 8.02050352e-01 1.08129367e-01 -2.13014051e-01
2.48734713e-01 4.66552734e-01 2.58468419e-01 -6.34122565e-02
-1.65407822e-01 -8.46641183e-01 -4.46167320e-01 -2.25553125e-01
-1.93932438e+00 -1.16744649e+00 -2.35881671e-01 1.85229027e+00
4.41277862e-01 6.86151206e-01 3.22921544e-01 4.89162654e-01
7.19131768e-01 -3.88685375e-01 -2.99097598e-01 -1.75931945e-01
-3.93391699e-01 3.64294678e-01 1.49265862e+00 9.69685435e-01
-1.29688644e+00 9.81643796e-01 6.96949291e+00 1.29687476e+00
-9.94277656e-01 1.04713321e-01 5.12778163e-01 -5.33440262e-02
1.79274485e-01 -4.26874071e-01 -1.21701777e+00 3.86453211e-01
1.17276216e+00 -2.09533557e-01 1.28277376e-01 5.17278910e-01
8.48245502e-01 -2.87885249e-01 -8.85039508e-01 9.57891107e-01
-5.18280983e-01 -1.90093672e+00 -1.93981975e-01 7.25623146e-02
7.30872810e-01 5.16315810e-02 -1.54899303e-02 6.65535927e-01
-2.21298382e-01 -7.76745737e-01 3.70941758e-01 9.50521708e-01
8.29604983e-01 -1.15481091e+00 1.19999623e+00 3.22651747e-03
-1.15105855e+00 -7.08551884e-01 3.59002128e-02 -4.89642359e-02
2.64612794e-01 8.86063218e-01 -1.26087761e+00 6.97328568e-01
-1.02675177e-01 2.84454048e-01 4.88710664e-02 1.59452343e+00
1.43186554e-01 1.14510107e+00 -4.40748155e-01 -3.00016016e-01
2.22765863e-01 4.74552929e-01 7.07052588e-01 1.19505847e+00
1.92439377e-01 2.66151279e-01 3.75639230e-01 -3.15752923e-01
3.33211660e-01 -2.14812383e-01 2.01426726e-02 5.33348620e-01
1.13597548e+00 1.13938415e+00 -3.02664191e-01 -9.34322625e-02
-6.54453754e-01 1.01536703e+00 -5.04742861e-01 7.67367065e-01
-1.34514570e+00 -9.32559848e-01 8.03709149e-01 -6.30059168e-02
4.32969958e-01 -6.41866624e-01 -8.98269951e-01 -5.89255452e-01
-3.47059786e-01 -3.46367240e-01 3.68672103e-01 -4.70721066e-01
-8.98344576e-01 -6.51022943e-04 -5.34238935e-01 -1.53890705e+00
-1.57680318e-01 -5.46469033e-01 -3.00325245e-01 7.72642612e-01
-1.65582848e+00 -8.41553032e-01 -1.04697989e-02 4.62062299e-01
4.61106807e-01 -1.11056626e+00 1.03173006e+00 6.60726905e-01
-4.12815720e-01 1.35696757e+00 5.99248409e-01 -1.55217901e-01
5.46636105e-01 -1.28873718e+00 -2.57838458e-01 2.65581787e-01
-6.82425797e-01 -1.01639926e-01 9.41726685e-01 -4.95077729e-01
-1.24536109e+00 -1.07441258e+00 6.62273288e-01 1.32947996e-01
5.63035727e-01 -2.61356652e-01 2.81100422e-01 4.21362072e-01
-1.71718374e-02 -1.27677038e-01 1.20364261e+00 1.18924677e-01
8.13578606e-01 -3.61603051e-01 -1.45538437e+00 7.56239533e-01
5.94567537e-01 7.31341029e-03 1.44604132e-01 4.35419947e-01
3.89793098e-01 -3.13431144e-01 -1.06205630e+00 9.31636244e-03
7.17681766e-01 -4.18169916e-01 1.07153642e+00 -5.67608476e-01
-5.44395149e-01 -8.43463913e-02 -6.71963334e-01 -1.25640452e+00
-4.31724101e-01 -1.07430673e+00 -6.58396780e-01 7.84806728e-01
1.05170035e+00 -5.10652900e-01 1.13427269e+00 -2.22914159e-01
-2.73261289e-03 -7.58925200e-01 -1.15784419e+00 -9.58393157e-01
-4.61987674e-01 -3.55471134e-01 2.34280929e-01 6.85831606e-01
4.25611943e-01 7.20325053e-01 -5.58324635e-01 1.15310478e+00
6.54335678e-01 -5.15140533e-01 5.53332925e-01 -1.05855310e+00
-6.74096525e-01 -1.30323544e-01 -6.34079397e-01 -1.29400289e+00
-3.57343405e-02 -8.67215574e-01 -6.00377023e-01 -1.35457361e+00
-4.51184213e-01 -8.40489507e-01 -3.69109154e-01 2.75842428e-01
2.40543053e-01 -6.23685447e-03 -3.54481399e-01 -4.99647826e-01
-4.68581259e-01 5.49281359e-01 7.54174054e-01 2.51836460e-02
-4.28626001e-01 1.26743066e+00 -5.98175168e-01 4.87691611e-01
1.11511695e+00 -5.67019768e-02 -2.81441629e-01 1.62604123e-01
-1.37074709e-01 8.72947693e-01 -2.41126254e-01 -1.63757277e+00
1.31566720e-02 -1.76341996e-01 8.69658113e-01 -6.23782873e-01
5.07906497e-01 -8.15579653e-01 -5.49303532e-01 6.47733152e-01
3.17958355e-01 -9.94313896e-01 3.92533004e-01 7.75992393e-01
4.59906578e-01 4.01749350e-02 1.04959393e+00 6.88624084e-01
-4.95393366e-01 4.54899520e-01 -1.54441822e+00 -1.62891462e-01
1.17713845e+00 -2.63241410e-01 -1.96181253e-01 -1.09784102e+00
-1.05064726e+00 6.61638439e-01 -7.34521389e-01 1.64434120e-01
4.77918565e-01 -1.34387696e+00 -6.12426162e-01 -1.51313588e-01
-2.25565225e-01 -1.37827444e+00 -1.44908398e-01 7.32164204e-01
-3.42526525e-01 1.02908087e+00 -4.09723192e-01 -3.69199038e-01
-1.38251317e+00 -3.29328328e-01 9.65553522e-01 -3.09728812e-02
-3.63242328e-02 8.83716285e-01 -9.70020115e-01 3.25488657e-01
4.28973466e-01 8.46118778e-02 -3.99095446e-01 -1.28085107e-01
1.82686269e-01 7.82671392e-01 1.03917606e-01 3.97914462e-02
-3.73610854e-01 -9.18708816e-02 -1.00358762e-01 -4.45046693e-01
8.87463570e-01 -4.53192234e-01 3.10716897e-01 1.57579035e-01
1.04862249e+00 4.35089886e-01 -5.73019028e-01 -2.57449653e-02
2.76397914e-02 -4.11737472e-01 4.24111277e-01 -9.22173619e-01
-6.44216537e-01 3.94747332e-02 1.23635375e+00 3.23995113e-01
8.07465196e-01 -3.03068668e-01 9.01431143e-01 7.71851897e-01
1.12444544e+00 -1.33555126e+00 -1.84932455e-01 7.20161378e-01
2.20051587e-01 -9.57621276e-01 5.74705862e-02 -8.24834824e-01
-2.94659436e-01 1.65009475e+00 3.68547410e-01 1.58911094e-01
1.17915535e+00 5.89984119e-01 3.97866545e-03 4.98207092e-01
-5.27436912e-01 -4.18708444e-01 -2.62527019e-01 8.03019941e-01
2.23061353e-01 3.82012188e-01 -4.59635198e-01 1.00985742e+00
-1.89813301e-01 1.33843645e-02 1.01146281e+00 8.06645691e-01
-1.11616766e+00 -1.32931828e+00 -3.07449639e-01 1.12897956e+00
-2.86318511e-01 3.14385146e-02 3.06903720e-01 5.32437086e-01
8.54311287e-02 1.41466475e+00 -1.15814455e-01 -4.77204382e-01
4.57349792e-02 -5.10685444e-01 6.16011977e-01 -3.53001624e-01
4.85262483e-01 -3.95211317e-02 1.02614784e+00 -2.32685357e-01
2.73762643e-01 -6.10366523e-01 -1.70925319e+00 -5.83483614e-02
-4.41990405e-01 2.86557227e-01 4.39279586e-01 1.59906542e+00
-1.06463499e-01 8.90948355e-01 1.56926274e+00 -5.06538153e-01
-1.75293520e-01 -4.61568415e-01 -1.08903909e+00 -1.03142226e+00
5.43543339e-01 -7.83165395e-01 -2.65045971e-01 -6.67623878e-01] | [6.123929023742676, 1.5284688472747803] |
4d8b01e7-b565-425c-abe6-bdd31aee9c4c | visual-prompt-tuning-for-few-shot-text | null | null | https://aclanthology.org/2022.coling-1.492 | https://aclanthology.org/2022.coling-1.492.pdf | Visual Prompt Tuning for Few-Shot Text Classification | Deploying large-scale pre-trained models in the prompt-tuning paradigm has demonstrated promising performance in few-shot learning. Particularly, vision-language pre-training models (VL-PTMs) have been intensively explored in various few-shot downstream tasks. However, most existing works only apply VL-PTMs to visual tasks like image classification, with few attempts being made on language tasks like text classification. In few-shot text classification, a feasible paradigm for deploying VL-PTMs is to align the input samples and their category names via the text encoders. However, it leads to the waste of visual information learned by the image encoders of VL-PTMs. To overcome this drawback, we propose a novel method named Visual Prompt Tuning (VPT). To our best knowledge, this method is the first attempt to deploy VL-PTM in few-shot text classification task. The main idea is to generate the image embeddings w.r.t. category names as visual prompt and then add them to the aligning process. Extensive experiments show that our VPT can achieve significant improvements under both zero-shot and few-shot settings. Importantly, our VPT even outperforms the most recent prompt-tuning methods on five public text classification datasets. | ['Zhao Cao', 'Jie Jiang', 'Hao Jiang', 'Zhiwu Lu', 'Guoxing Yang', 'Nanyi Fei', 'Yutian Luo', 'Jingyuan Wen'] | null | null | null | null | coling-2022-10 | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 3.11690450e-01 -1.74634486e-01 -4.04000461e-01 -3.28123540e-01
-5.95597148e-01 -5.30478954e-02 1.02114868e+00 1.71766877e-01
-6.42643988e-01 2.35603452e-01 2.14374438e-01 -3.27334762e-01
4.04279232e-01 -6.48901880e-01 -5.48629761e-01 -5.82655251e-01
7.14595973e-01 3.32164049e-01 4.09677207e-01 -2.40718424e-01
4.38712239e-01 -2.19497770e-01 -1.47474074e+00 3.97503912e-01
6.65424883e-01 7.37300038e-01 6.18586123e-01 6.73488498e-01
-7.56134331e-01 8.67979050e-01 -4.25913632e-01 -4.98683006e-01
7.32577546e-03 -6.09001100e-01 -5.94304562e-01 2.32124686e-01
4.30827141e-01 -5.66896915e-01 -3.61375391e-01 9.32081401e-01
6.58463597e-01 6.38514400e-01 8.52353036e-01 -1.45903027e+00
-1.19771338e+00 7.08854437e-01 -7.36269951e-01 3.76686454e-01
3.95422503e-02 4.38478887e-01 1.11941862e+00 -1.42900193e+00
6.57465339e-01 1.25916207e+00 3.67291570e-01 8.69098425e-01
-1.15231848e+00 -3.92247140e-01 4.14927602e-01 5.29102921e-01
-1.24609804e+00 -5.00742793e-01 6.06694818e-01 -5.84976912e-01
1.11468863e+00 -1.57131180e-01 5.27733386e-01 1.50196874e+00
-3.27734724e-02 1.08587801e+00 7.98365533e-01 -7.32507944e-01
3.36687088e-01 2.40847677e-01 4.66167152e-01 7.55653977e-01
-6.69170320e-02 -3.35088819e-01 -6.58066630e-01 1.24578208e-01
4.16131377e-01 3.69555116e-01 -1.12390317e-01 -3.25402796e-01
-1.18012035e+00 9.49644804e-01 3.72873724e-01 4.88791645e-01
-1.48220256e-01 1.72006309e-01 5.93172014e-01 2.08642274e-01
6.29447699e-01 2.11516321e-01 -1.30164742e-01 -4.98537235e-02
-9.73086417e-01 -1.99774161e-01 4.03800935e-01 1.13869762e+00
7.03947544e-01 8.49306881e-02 -1.07601213e+00 1.05306339e+00
2.83083856e-01 2.70544499e-01 7.50336826e-01 -2.53444791e-01
6.48364246e-01 4.77734029e-01 -1.24546438e-01 -6.69115663e-01
-8.60987902e-02 -5.45135736e-02 -7.29362726e-01 -7.12164640e-02
1.33022189e-01 -4.86079603e-02 -1.32543087e+00 1.36133337e+00
1.40438601e-01 3.97897542e-01 1.18932605e-01 8.48583579e-01
1.19165492e+00 9.84527707e-01 3.91904980e-01 -2.58281112e-01
1.52508128e+00 -1.55452514e+00 -8.44779134e-01 -4.96335387e-01
7.86059082e-01 -7.36970425e-01 1.66061091e+00 -2.22047865e-01
-6.72785163e-01 -7.35004067e-01 -8.96712661e-01 -2.54257619e-01
-4.51233715e-01 2.30809346e-01 3.67248118e-01 4.19126809e-01
-7.37810969e-01 3.81606549e-01 -5.00304937e-01 -8.11102331e-01
7.05509901e-01 -1.85801715e-01 -9.87192392e-02 -3.40804815e-01
-1.01773095e+00 7.40059555e-01 3.36951166e-01 -3.57919365e-01
-1.10197520e+00 -6.19311333e-01 -8.89685810e-01 3.96943659e-01
6.10029101e-01 -6.72577977e-01 1.37643433e+00 -8.54344845e-01
-1.48513806e+00 8.94052982e-01 -4.72578168e-01 -3.89283240e-01
4.26843941e-01 -7.49573186e-02 -1.31059736e-01 1.27948418e-01
2.56252408e-01 9.05186236e-01 1.29174709e+00 -1.09568596e+00
-5.16986370e-01 -1.28234521e-01 1.79016721e-02 2.54911929e-01
-9.68177915e-01 5.94929829e-02 -7.42736816e-01 -5.75073123e-01
-3.62481236e-01 -5.75053632e-01 -1.98622257e-01 2.37305850e-01
-3.12061191e-01 -6.55858815e-01 8.74255180e-01 -2.99466491e-01
1.12883210e+00 -2.12131047e+00 5.02614900e-02 -5.33969223e-01
1.29124641e-01 6.88293040e-01 -4.88344520e-01 5.79696178e-01
1.94931120e-01 1.91412978e-02 9.07942206e-02 -8.12987566e-01
1.34172738e-01 1.77278638e-01 -6.89951301e-01 2.06623957e-01
1.73805237e-01 1.39504397e+00 -9.76382613e-01 -8.14892173e-01
5.53554773e-01 3.01379144e-01 -3.45275640e-01 2.74703264e-01
-5.38278103e-01 5.14715463e-02 -3.68390054e-01 4.28615749e-01
3.64126503e-01 -6.94690049e-01 -2.95432895e-01 -1.28944516e-01
-1.24644108e-01 -6.31602779e-02 -6.45442069e-01 1.76905751e+00
-3.81162792e-01 8.12695384e-01 -5.24593711e-01 -1.12442589e+00
8.82073641e-01 4.27185714e-01 9.32837501e-02 -8.27913225e-01
3.87683064e-01 -2.05362186e-01 -2.58784205e-01 -7.19307542e-01
5.62693536e-01 -2.97482848e-01 4.15351987e-02 5.45264363e-01
4.78155553e-01 1.49709821e-01 3.68969679e-01 5.90382814e-01
7.69880414e-01 1.24402426e-01 5.78957915e-01 1.64944977e-01
4.25864816e-01 1.45421535e-01 2.10226744e-01 9.84998345e-01
-5.27103484e-01 6.80797935e-01 3.62680197e-01 -2.60378301e-01
-1.24631071e+00 -7.17549205e-01 1.58857048e-01 1.68681753e+00
2.09739342e-01 -5.51433206e-01 -5.87857723e-01 -8.11998606e-01
-2.19040543e-01 9.37221766e-01 -7.97149777e-01 -3.02839458e-01
-1.72117844e-01 -5.35047770e-01 3.52746874e-01 6.52366936e-01
3.87084126e-01 -1.12803912e+00 -5.24065375e-01 1.73782423e-01
-6.96964338e-02 -1.39084768e+00 -8.61192226e-01 1.07663319e-01
-6.91990852e-01 -7.94195294e-01 -1.13755882e+00 -1.10817254e+00
7.78494418e-01 9.66234326e-01 6.40304744e-01 1.13605812e-01
-4.71232831e-01 4.12838310e-01 -7.73795485e-01 -4.47326928e-01
-1.83386549e-01 -5.06283417e-02 -2.03442276e-01 1.73488483e-01
7.03962922e-01 -1.34221256e-01 -6.33159101e-01 1.85845211e-01
-6.84330940e-01 3.95951152e-01 5.16399503e-01 1.11220658e+00
4.97316897e-01 -4.88467783e-01 4.20049429e-01 -7.60621071e-01
5.51404357e-01 -3.50407839e-01 -3.30737650e-01 5.04285991e-01
-5.51366329e-01 2.83909757e-02 8.99110436e-01 -7.19371736e-01
-1.09612989e+00 -9.55300778e-02 -1.72656663e-02 -8.65005493e-01
-2.02994972e-01 2.47605234e-01 2.64573872e-01 2.34540418e-01
7.29501009e-01 5.25828004e-01 -2.81509161e-01 -3.25407684e-01
6.79649711e-01 9.12288725e-01 2.47589126e-01 -2.22044975e-01
7.90596545e-01 5.91400683e-01 -3.40620071e-01 -1.08065915e+00
-1.25807142e+00 -8.94307613e-01 -6.41100883e-01 -7.82815740e-02
1.09206915e+00 -9.37399745e-01 -2.55371213e-01 3.17391127e-01
-1.26704907e+00 -4.01911378e-01 -1.95853502e-01 3.28791559e-01
-4.99004006e-01 5.12988448e-01 -3.39105099e-01 -7.74407268e-01
-7.17773020e-01 -1.00247467e+00 1.19090724e+00 3.60653937e-01
3.46529633e-02 -9.80826497e-01 4.12759110e-02 2.35348821e-01
4.26387668e-01 -5.27073920e-01 9.03174400e-01 -7.38350749e-01
-2.68905520e-01 4.67610583e-02 -5.48705339e-01 -3.50712170e-03
8.92309323e-02 -1.42434999e-01 -1.17912197e+00 -4.14831936e-01
-2.81060874e-01 -6.26408041e-01 1.15803504e+00 4.94605601e-01
1.14363694e+00 -2.73995660e-02 -4.94229376e-01 6.03330612e-01
1.45067954e+00 1.90650538e-01 4.04103279e-01 2.63331443e-01
9.94627833e-01 4.59274828e-01 8.53423417e-01 5.49586058e-01
1.25654459e-01 8.16533923e-01 2.09284887e-01 -5.29199615e-02
-4.47972536e-01 -5.29483318e-01 3.38018149e-01 7.56859601e-01
1.39678061e-01 -4.24652845e-01 -8.08729827e-01 6.19263291e-01
-2.09806204e+00 -1.03409803e+00 1.04229171e-02 1.82635784e+00
6.82167828e-01 3.46191339e-02 -1.26247406e-01 -2.68355578e-01
9.41807866e-01 5.09565473e-01 -5.29756844e-01 -3.06357056e-01
8.95500109e-02 7.72229284e-02 5.98162040e-02 2.78950930e-01
-1.18294120e+00 1.38297665e+00 5.08951950e+00 9.10118937e-01
-9.41125453e-01 5.16230524e-01 4.26868111e-01 -3.07455361e-01
-1.12422824e-01 2.09803104e-01 -1.26899672e+00 4.24831867e-01
6.16890311e-01 -4.93920356e-01 1.53746307e-01 1.05377698e+00
2.45273069e-01 1.87124535e-01 -1.02248073e+00 1.14495170e+00
5.93320489e-01 -1.40715396e+00 4.21782643e-01 -2.60559678e-01
8.66283774e-01 -8.78661778e-03 5.31870574e-02 7.82035470e-01
7.36383796e-02 -8.71060252e-01 5.16487837e-01 2.82819092e-01
9.24042046e-01 -4.90002394e-01 5.42563021e-01 4.55281019e-01
-1.23492789e+00 -1.80723920e-01 -8.92650962e-01 -1.76967420e-02
3.23925138e-01 2.83372134e-01 -1.04878736e+00 1.66547179e-01
5.40603459e-01 9.61351216e-01 -6.14987433e-01 1.01070714e+00
-3.53769541e-01 5.26763737e-01 3.19302768e-01 -5.06754279e-01
5.98276973e-01 9.04843956e-02 3.99461865e-01 1.15790582e+00
3.15627009e-01 -2.57649720e-02 3.48242939e-01 8.09979975e-01
-1.76917672e-01 5.04526913e-01 -6.79982066e-01 -4.37538743e-01
3.76697779e-01 1.34287214e+00 -8.96675408e-01 -7.94107020e-01
-7.81141758e-01 1.17719531e+00 5.08507729e-01 3.76037538e-01
-8.13787699e-01 -5.15866160e-01 3.90118837e-01 -1.89828604e-01
6.77264631e-01 1.14791747e-02 -1.10564217e-01 -1.45930791e+00
-2.00646132e-01 -5.50290942e-01 2.91886598e-01 -9.73377943e-01
-1.32507491e+00 3.72069359e-01 -6.63476065e-02 -1.20860672e+00
2.57969424e-02 -7.27188885e-01 -8.83707464e-01 5.64615846e-01
-1.52902400e+00 -1.29345202e+00 -5.54566383e-01 7.12793589e-01
1.42502213e+00 -4.00537908e-01 7.68115759e-01 -1.99550018e-03
-7.50258267e-01 6.94180727e-01 1.33250058e-01 1.69966653e-01
9.91754115e-01 -1.07244968e+00 6.00871742e-01 9.56517458e-01
4.82214391e-01 5.02563417e-01 6.69535518e-01 -6.06271982e-01
-1.29215813e+00 -1.24848211e+00 8.10527921e-01 -2.52290249e-01
7.75645494e-01 -5.30771375e-01 -1.08829284e+00 7.79286504e-01
6.19410872e-01 1.70926571e-01 5.60574174e-01 -9.94261876e-02
-4.64567125e-01 1.68228477e-01 -4.99499053e-01 9.26326632e-01
9.97160316e-01 -5.39464414e-01 -9.53799367e-01 5.90052605e-01
9.05969739e-01 -4.30581793e-02 -3.52160811e-01 -7.07077906e-02
2.78222859e-01 -5.38535714e-01 8.71885300e-01 -7.28527129e-01
7.95312464e-01 -1.09234691e-01 -8.26395527e-02 -1.36025155e+00
-3.00938725e-01 -3.49833637e-01 -2.34411493e-01 1.39536250e+00
6.27966225e-02 -2.63633311e-01 5.64897358e-01 7.07413554e-02
-1.83079541e-01 -5.30761480e-01 -7.18262613e-01 -7.69933462e-01
-2.13680305e-02 -2.17899129e-01 1.02543563e-01 1.03554320e+00
1.19589437e-02 9.70567942e-01 -7.07078099e-01 -2.55419970e-01
6.41899645e-01 2.21515402e-01 9.16065693e-01 -1.11870933e+00
-4.28330541e-01 -5.04573226e-01 -1.88620061e-01 -1.21294844e+00
1.79936931e-01 -1.04271984e+00 4.17455345e-01 -1.88217330e+00
8.10573041e-01 -1.48151647e-02 -2.87746161e-01 7.17877388e-01
-6.22371674e-01 8.43188018e-02 4.40935314e-01 2.21737623e-01
-9.08356309e-01 8.98477972e-01 1.28989863e+00 -5.13563514e-01
-9.79205668e-02 -3.04900020e-01 -4.98091549e-01 5.34923911e-01
7.36329854e-01 -5.13252854e-01 -6.03842974e-01 -4.24247563e-01
-1.59107834e-01 -2.85776556e-01 3.71842235e-01 -7.15354145e-01
5.09988189e-01 -3.69573355e-01 2.68908113e-01 -7.66258538e-01
2.58580327e-01 -4.04170364e-01 -6.28439665e-01 4.69436765e-01
-4.46074158e-01 -2.35610187e-01 -2.78888084e-02 7.24005818e-01
-4.90996093e-02 -6.31483614e-01 9.30993676e-01 -1.66930407e-01
-1.38425243e+00 4.31808472e-01 -3.20719779e-01 1.87949806e-01
1.25359368e+00 -1.35341614e-01 -7.49354362e-01 -1.59452632e-01
-4.11255419e-01 4.01765168e-01 3.94180298e-01 7.52968729e-01
8.11704755e-01 -1.25844932e+00 -5.14048398e-01 6.72144443e-02
8.12368155e-01 -1.97065324e-01 4.57061291e-01 9.22739983e-01
1.69147141e-02 6.93585455e-01 -2.63688534e-01 -7.01971591e-01
-1.54048800e+00 1.03916430e+00 1.62350424e-02 -9.58439559e-02
-1.14303374e+00 9.24571037e-01 5.79491854e-01 -5.43260872e-02
3.62525553e-01 1.00314088e-01 -4.79665935e-01 2.91042328e-01
9.11601424e-01 2.81375088e-02 -2.82527983e-01 -4.67510670e-01
-8.66286084e-02 6.85406208e-01 -5.29344320e-01 -7.84061104e-02
1.15272343e+00 -2.64064133e-01 3.50655198e-01 7.73491442e-01
1.19403982e+00 -4.95646596e-01 -1.42009115e+00 -5.69029808e-01
-1.67491406e-01 -4.87750649e-01 6.71928078e-02 -2.79514492e-01
-8.40913773e-01 1.49351954e+00 4.01230931e-01 -1.08802885e-01
7.52226472e-01 1.83625609e-01 7.86198497e-01 5.54783702e-01
1.65760338e-01 -1.23645186e+00 8.15899372e-01 7.10605979e-01
4.76070166e-01 -1.54075754e+00 -1.88954726e-01 -1.00641198e-01
-1.06756020e+00 1.04561567e+00 9.60788131e-01 3.34372930e-02
2.74058044e-01 -1.96766585e-01 9.09502357e-02 1.11808917e-02
-1.12915742e+00 -6.43729627e-01 2.52338767e-01 5.46854556e-01
3.56476068e-01 -2.66836822e-01 -3.59905541e-01 4.63630527e-01
3.43766600e-01 9.44767147e-02 4.69286174e-01 9.44146216e-01
-8.83662164e-01 -9.04654145e-01 -8.50821882e-02 5.16004920e-01
-7.90593252e-02 -4.16331887e-01 -2.40776867e-01 4.92977500e-01
-1.14222497e-01 9.18315768e-01 1.78296044e-01 -2.03367382e-01
2.61850715e-01 4.16585982e-01 2.94176340e-01 -1.09923565e+00
-2.43033186e-01 2.11204682e-02 -2.82162368e-01 -1.92216754e-01
-2.84872830e-01 -3.28345776e-01 -1.19404411e+00 7.22839236e-02
-3.67575049e-01 -6.25225753e-02 4.99187052e-01 1.15949571e+00
1.52133152e-01 7.45006323e-01 5.67552984e-01 -8.47383440e-01
-6.87436521e-01 -1.21871567e+00 -3.81707907e-01 5.39382219e-01
1.88816249e-01 -8.07361484e-01 -3.50480646e-01 1.43956646e-01] | [10.127581596374512, 2.354889392852783] |
47e5883d-2b9b-49e2-8f12-0999c76fb094 | inter-annotator-agreement-in-sentiment | null | null | https://aclanthology.org/R17-1015 | https://aclanthology.org/R17-1015.pdf | Inter-Annotator Agreement in Sentiment Analysis: Machine Learning Perspective | Manual text annotation is an essential part of Big Text analytics. Although annotators work with limited parts of data sets, their results are extrapolated by automated text classification and affect the final classification results. Reliability of annotations and adequacy of assigned labels are especially important in the case of sentiment annotations. In the current study we examine inter-annotator agreement in multi-class, multi-label sentiment annotation of messages. We used several annotation agreement measures, as well as statistical analysis and Machine Learning to assess the resulting annotations. | ['Victoria Bobicev', 'Marina Sokolova'] | 2017-09-01 | null | null | null | ranlp-2017-9 | ['text-annotation'] | ['natural-language-processing'] | [ 9.09641385e-02 5.41604578e-01 -1.83329836e-01 -8.40964913e-01
-6.70940042e-01 -1.06644428e+00 4.50144380e-01 1.11609924e+00
-5.47449887e-01 7.56614149e-01 4.60837305e-01 -7.40165636e-02
1.44905403e-01 -5.45685649e-01 -1.25266865e-01 -3.29809099e-01
6.07850611e-01 5.66251755e-01 1.62826389e-01 -1.96948186e-01
4.15089071e-01 -1.02689445e-01 -1.36857891e+00 8.27701092e-01
7.70725191e-01 1.07667744e+00 -3.49541157e-01 5.87663591e-01
-7.43927240e-01 1.24948299e+00 -8.13621759e-01 -7.71786749e-01
-4.68876287e-02 -2.94319391e-01 -1.13267076e+00 7.97928199e-02
1.37719974e-01 3.55788141e-01 8.09560359e-01 9.80203569e-01
2.48030096e-01 -3.23244482e-01 7.01980948e-01 -1.14825726e+00
-2.29622066e-01 9.54787493e-01 -3.74496907e-01 -7.20671862e-02
5.80062747e-01 -4.40938234e-01 1.16085494e+00 -8.94448221e-01
7.08261311e-01 7.88271368e-01 1.03798759e+00 -3.49214710e-02
-8.44532669e-01 -5.69731474e-01 -1.65293679e-01 -2.05114797e-01
-1.13237560e+00 -2.21300736e-01 2.07168818e-01 -1.01894510e+00
7.41567791e-01 2.23871857e-01 3.31103265e-01 5.36347747e-01
1.11945458e-01 2.07228795e-01 1.58569360e+00 -7.06612289e-01
3.42002988e-01 8.20141494e-01 5.42166650e-01 4.01881546e-01
3.56013983e-01 -7.14258432e-01 -4.31793660e-01 -4.43907648e-01
-3.56134057e-01 -1.89284816e-01 2.07562774e-01 1.75376371e-01
-8.17915082e-01 9.09448743e-01 1.55805750e-02 5.43242574e-01
-3.88654083e-01 -2.58010149e-01 9.55141544e-01 5.89039862e-01
9.97259438e-01 5.44233382e-01 -7.82187462e-01 -2.46954367e-01
-9.03121531e-01 -1.07645385e-01 1.11546183e+00 6.66915059e-01
8.97586405e-01 -4.40252841e-01 -1.06941909e-01 9.68729436e-01
2.58921176e-01 2.13339776e-01 6.78623080e-01 -6.14135206e-01
3.54291677e-01 1.19786131e+00 1.33485094e-01 -1.16422522e+00
-8.77369583e-01 -2.74368197e-01 -5.79274535e-01 1.00223580e-02
4.05403107e-01 -4.08320516e-01 -5.11743307e-01 1.01723588e+00
4.10644948e-01 -9.52482522e-01 5.87104931e-02 5.93442857e-01
9.99386311e-01 2.60713935e-01 5.46029270e-01 -3.07972282e-01
1.49408245e+00 -7.78366387e-01 -1.04611409e+00 7.49458075e-02
1.17648757e+00 -1.10941267e+00 1.05238175e+00 3.85561317e-01
-5.93680978e-01 -4.63210285e-01 -5.39855242e-01 -3.77545580e-02
-8.55882764e-01 2.84417093e-01 5.07573187e-01 8.59832764e-01
-7.02396274e-01 3.23934913e-01 -4.44747865e-01 -3.66055250e-01
3.39278609e-01 3.59047592e-01 -6.46891236e-01 4.33087617e-01
-1.21344686e+00 1.02105081e+00 4.53894198e-01 -6.58576936e-02
-4.55445126e-02 -5.17992020e-01 -6.36515081e-01 -4.07707989e-02
2.19148546e-01 5.24342023e-02 1.44945943e+00 -1.44927168e+00
-9.78757381e-01 1.20253956e+00 5.24856709e-02 -1.41365305e-01
5.24721503e-01 -1.17873758e-01 -4.33234215e-01 -9.67821255e-02
2.95265943e-01 3.95666957e-01 3.83935571e-01 -1.25282574e+00
-9.53328013e-01 -4.04589236e-01 -2.20034972e-01 4.55103330e-02
-5.58117747e-01 4.96168345e-01 1.16918720e-01 -4.68012094e-01
-1.51024982e-01 -9.90926802e-01 -1.60832509e-01 -5.62803268e-01
-1.40531167e-01 -4.88363832e-01 6.46543086e-01 -6.83449924e-01
1.22226584e+00 -1.93788934e+00 -3.98284495e-01 3.75640124e-01
3.14524353e-01 -1.46130010e-01 3.45775783e-01 6.32399261e-01
8.33383128e-02 5.88616729e-01 1.76828012e-01 -5.03041625e-01
-1.76353590e-03 -2.28129383e-02 -1.77497581e-01 3.34422886e-01
-1.04384430e-01 5.81029832e-01 -7.88173795e-01 -9.25781667e-01
3.06448899e-02 2.66584218e-01 -1.71453163e-01 3.40332426e-02
-9.23575088e-02 3.56506556e-01 -4.94528443e-01 4.81750190e-01
1.82426080e-01 -2.94745386e-01 2.82117218e-01 -4.13195193e-01
-3.19598287e-01 1.59637734e-01 -8.91649842e-01 1.10036933e+00
-5.10925949e-01 7.04125643e-01 -1.47992030e-01 -7.40133882e-01
1.10032451e+00 5.40416598e-01 5.57650149e-01 -3.07801366e-01
5.16304553e-01 4.13445473e-01 -9.48256329e-02 -5.24841607e-01
9.56120908e-01 -1.76408708e-01 -3.84178609e-01 8.16943586e-01
1.06332317e-01 -2.13058084e-01 3.94875318e-01 2.45415777e-01
7.97796905e-01 -2.16301411e-01 6.47810280e-01 -4.00551945e-01
4.40443218e-01 5.08443177e-01 2.54140317e-01 4.37529474e-01
2.99083389e-04 4.91802990e-01 8.74886215e-01 -6.34150624e-01
-1.21196890e+00 1.39097974e-01 -3.43771607e-01 1.36080551e+00
-3.58034879e-01 -7.94423521e-01 -7.05116987e-01 -1.06379962e+00
-3.17580104e-01 5.96589565e-01 -8.25812340e-01 1.84122145e-01
6.17654435e-02 -8.20214272e-01 3.42804909e-01 3.28067809e-01
1.04237506e-02 -1.03261638e+00 -6.48452103e-01 1.47456333e-01
-1.52203679e-01 -1.19479513e+00 -9.35368389e-02 5.43440163e-01
-4.13242608e-01 -1.14024436e+00 -1.57107458e-01 -6.92960262e-01
7.77375162e-01 -2.33642340e-01 1.23401558e+00 4.21575010e-01
1.82697609e-01 5.38610518e-02 -9.94721949e-01 -7.79507518e-01
-7.81696379e-01 5.93991339e-01 -2.07661226e-01 -8.58483762e-02
7.31931567e-01 1.30173415e-01 -1.18727118e-01 7.20528305e-01
-1.00090992e+00 -1.69206753e-01 3.09429109e-01 6.51269853e-01
4.30099636e-01 2.70523131e-01 7.32751369e-01 -1.47614062e+00
9.41557825e-01 -3.54181230e-01 -4.57629591e-01 2.47740105e-01
-9.18125808e-01 -1.87265009e-01 6.54857159e-01 -1.41842172e-01
-8.46515417e-01 3.16492580e-02 -2.30746299e-01 3.26557457e-01
-1.73045605e-01 7.61601508e-01 3.58472347e-01 -1.22452229e-01
8.68228674e-01 -7.44611740e-01 -1.11384660e-01 -3.02098453e-01
9.47808847e-03 1.20076680e+00 -1.55242607e-01 -2.22009361e-01
2.18929216e-01 2.30352193e-01 -2.76732564e-01 -3.05507988e-01
-1.59817934e+00 -7.33257055e-01 -9.99301255e-01 -3.58920276e-01
7.64225185e-01 -7.78807461e-01 -4.48898315e-01 1.34325340e-01
-8.36160898e-01 -4.19668317e-01 -4.84981060e-01 3.53545636e-01
-2.68601477e-02 1.78429484e-01 -3.37334961e-01 -7.14141965e-01
-4.66108710e-01 -1.09351361e+00 1.11795175e+00 1.38413444e-01
-1.05199409e+00 -1.20564806e+00 2.91425258e-01 8.82652044e-01
3.15639019e-01 1.04058683e-01 4.92658883e-01 -1.36071479e+00
6.92463040e-01 -9.61271703e-01 -1.85194001e-01 3.23253751e-01
1.15225367e-01 2.85603315e-01 -1.14240706e+00 9.66327637e-02
-2.78272599e-01 -7.80533493e-01 4.45794731e-01 2.79783577e-01
1.06878710e+00 -2.74887741e-01 -3.79710078e-01 -4.39112753e-01
1.16773450e+00 -2.31500462e-01 3.36644590e-01 4.72595781e-01
6.20813847e-01 1.17505610e+00 1.03191245e+00 5.16089082e-01
4.86280799e-01 4.11849648e-01 1.27869710e-01 7.36934319e-02
3.93589795e-01 1.01168633e-01 1.14949539e-01 8.79385471e-01
4.00234982e-02 -4.02034044e-01 -1.35922730e+00 2.79295534e-01
-1.72496545e+00 -6.10477746e-01 -6.97245598e-01 1.83931899e+00
9.73907411e-01 3.51122737e-01 1.77902594e-01 5.21447539e-01
8.33851874e-01 -2.81225473e-01 2.66760260e-01 -6.47495747e-01
-1.00611292e-01 -1.25549421e-01 6.75140262e-01 4.05165881e-01
-1.01518059e+00 8.07885528e-01 7.30568457e+00 7.02293515e-01
-7.70214021e-01 6.07569635e-01 7.65925288e-01 1.61908314e-01
-1.38980165e-01 3.94342281e-02 -8.12655270e-01 5.48636198e-01
1.03233433e+00 3.80288333e-01 -3.12955111e-01 8.04416478e-01
1.39226481e-01 -4.81523901e-01 -6.76208735e-01 3.81608099e-01
8.51893574e-02 -9.96546328e-01 -4.45845515e-01 1.09277263e-01
1.03692698e+00 -9.01487023e-02 -5.03288507e-01 2.96955109e-01
3.92328948e-01 -9.00481284e-01 7.59598732e-01 2.37524047e-01
7.11283267e-01 -5.54085076e-01 1.63268173e+00 6.62114471e-02
-6.73946559e-01 -4.87723015e-02 -1.15365550e-01 -2.91493744e-01
2.36594483e-01 9.54741597e-01 -8.21337223e-01 1.61905617e-01
6.81148767e-01 3.97375017e-01 -7.98831999e-01 4.41623688e-01
-2.85169929e-01 7.00722396e-01 -9.92970690e-02 -3.09949338e-01
6.88669458e-02 -2.47097462e-02 -2.23010272e-01 1.27836621e+00
-1.28703609e-01 -5.88796958e-02 4.52038407e-01 7.51405060e-02
-1.82148352e-01 8.14206898e-01 -2.34586075e-01 -2.95907348e-01
2.86296576e-01 1.97075498e+00 -1.40941465e+00 -5.92706263e-01
-6.87723279e-01 4.13116455e-01 3.25668275e-01 -1.92768693e-01
-3.08082938e-01 -2.99390107e-01 1.99074522e-02 1.51431516e-01
-1.60049081e-01 2.14126483e-01 -6.23113692e-01 -7.88201094e-01
-2.55931675e-01 -8.48721623e-01 6.17780387e-01 -8.43066931e-01
-1.18462145e+00 7.00811863e-01 -2.73247302e-01 -1.13207984e+00
3.99371274e-02 -5.63414216e-01 -3.05614591e-01 3.05428147e-01
-9.31969404e-01 -1.20298159e+00 -6.67539895e-01 5.41733503e-02
4.38129753e-01 -1.23104893e-01 9.13982868e-01 3.96479905e-01
-2.16158673e-01 2.85041332e-01 -4.90076952e-02 2.68202007e-01
1.04680729e+00 -1.28822446e+00 -2.18869880e-01 1.56544656e-01
-1.22012816e-01 1.87505186e-01 9.19929266e-01 -9.59682226e-01
-5.08111775e-01 -8.37804675e-01 1.08575487e+00 -6.98805571e-01
1.01331663e+00 -1.34151235e-01 -7.76631057e-01 5.41495085e-01
3.09958458e-01 2.69981604e-02 1.49421799e+00 4.81488645e-01
-2.86201924e-01 1.52100638e-01 -1.11400926e+00 1.12375477e-03
3.35739195e-01 -5.34225702e-01 -2.98550278e-01 7.32510328e-01
3.92367035e-01 -4.65302855e-01 -1.35760081e+00 2.08810285e-01
7.38239646e-01 -5.56659222e-01 2.71271378e-01 -6.01116598e-01
4.53120768e-01 -1.92652181e-01 -4.56924513e-02 -1.29439771e+00
7.77916834e-02 4.31707315e-02 6.00564480e-01 1.54780066e+00
9.32389915e-01 -3.35214496e-01 4.96047705e-01 9.76020873e-01
-6.32774457e-02 -2.41396815e-01 -2.62284338e-01 -2.25076675e-01
-4.52929316e-03 -5.86680949e-01 3.80179316e-01 1.24893916e+00
6.04100585e-01 4.29136693e-01 -2.63431847e-01 -2.36468211e-01
1.38884440e-01 -1.46678329e-01 7.71834433e-01 -1.63810563e+00
2.86422849e-01 -3.40421706e-01 -4.08913523e-01 4.06211436e-01
3.89490604e-01 -7.92915642e-01 5.70554025e-02 -1.58697975e+00
2.54108399e-01 -5.77438474e-01 -2.52623595e-02 7.01019228e-01
-6.30467609e-02 7.62712061e-01 -2.30703905e-01 5.04418075e-01
-9.82153416e-01 -7.00740516e-02 7.23713875e-01 -2.79786065e-03
-1.06927149e-01 1.20784519e-02 -7.29178548e-01 8.78922403e-01
1.23057425e+00 -9.96970236e-01 -7.56177828e-02 -9.68105942e-02
1.30435741e+00 -6.40636384e-01 -4.82058562e-02 -7.94251502e-01
-2.34323014e-02 -1.00067407e-01 3.52092862e-01 -3.29156846e-01
-2.97560006e-01 -9.69411075e-01 2.59613544e-01 2.46083468e-01
-8.63320589e-01 1.37737259e-01 -2.66544130e-02 8.47326145e-02
-4.08134818e-01 -8.10580194e-01 4.95853752e-01 -1.89683378e-01
-2.37428680e-01 -1.54088631e-01 -6.37621701e-01 1.37117982e-01
9.45232213e-01 -1.57371968e-01 -4.02022243e-01 -3.63746911e-01
-8.91289890e-01 1.65132046e-01 6.64872229e-01 3.19996804e-01
-2.43221462e-01 -1.03509009e+00 -6.94600821e-01 -3.38592857e-01
4.99499381e-01 -1.21224634e-01 -1.20846324e-01 1.19060922e+00
-5.25167823e-01 3.12559217e-01 -2.30190642e-02 -4.21241939e-01
-1.55345106e+00 3.71509820e-01 -6.75085708e-02 -4.93683189e-01
-9.64151397e-02 3.88465881e-01 -1.96764201e-01 -5.31776786e-01
-1.18425481e-01 -1.41389325e-01 -8.85365188e-01 8.12727451e-01
5.10975599e-01 3.66770834e-01 4.19463217e-01 -9.26225245e-01
-2.38744199e-01 6.11015737e-01 -1.24839053e-01 -9.08653215e-02
1.10619855e+00 -3.83392692e-01 -4.72121269e-01 9.15231645e-01
9.83918428e-01 3.01136047e-01 -2.99847066e-01 5.60274124e-02
3.84655744e-01 -3.35738987e-01 1.90730140e-01 -7.94746816e-01
-6.51973963e-01 6.43950880e-01 3.50150198e-01 9.48533297e-01
7.11902678e-01 1.63559511e-01 1.21284969e-01 1.62786037e-01
9.86805633e-02 -1.73074567e+00 -1.41769592e-02 3.81203562e-01
5.74127555e-01 -1.68694055e+00 2.97743201e-01 -3.94525021e-01
-1.03934109e+00 1.11808634e+00 4.76355016e-01 5.21700025e-01
7.33408391e-01 4.28362280e-01 4.04414266e-01 -5.73010147e-01
-6.23759151e-01 3.86447124e-02 3.79263043e-01 1.95557535e-01
1.20025635e+00 -1.64465889e-01 -8.53395700e-01 6.81286395e-01
-4.12578434e-01 4.05131690e-02 6.46829426e-01 1.01324296e+00
-6.50850594e-01 -9.40571606e-01 -3.75382096e-01 7.24615395e-01
-1.10000432e+00 9.01288614e-02 -8.18120122e-01 4.23951000e-01
9.98917371e-02 1.23304951e+00 4.28041518e-02 -2.88046002e-01
-2.76332069e-02 2.51320601e-01 -1.99477464e-01 -8.41770530e-01
-1.27303398e+00 -4.44046408e-02 5.92094064e-01 -1.05224796e-01
-9.76468742e-01 -7.78028965e-01 -1.20225263e+00 -2.28465587e-01
-9.43659246e-01 6.85320675e-01 1.14586830e+00 1.14415443e+00
2.03105628e-01 4.05596972e-01 4.97293830e-01 -3.00446719e-01
-1.24068514e-01 -1.47306025e+00 -2.85723388e-01 6.11656904e-01
-1.45713046e-01 -3.45340401e-01 -5.30943036e-01 4.83767390e-01] | [10.82544994354248, 6.787208557128906] |
fe980b79-cf84-42ee-a26e-47d2cd88590a | expressive-body-capture-3d-hands-face-and | 1904.05866 | null | http://arxiv.org/abs/1904.05866v1 | http://arxiv.org/pdf/1904.05866v1.pdf | Expressive Body Capture: 3D Hands, Face, and Body from a Single Image | To facilitate the analysis of human actions, interactions and emotions, we
compute a 3D model of human body pose, hand pose, and facial expression from a
single monocular image. To achieve this, we use thousands of 3D scans to train
a new, unified, 3D model of the human body, SMPL-X, that extends SMPL with
fully articulated hands and an expressive face. Learning to regress the
parameters of SMPL-X directly from images is challenging without paired images
and 3D ground truth. Consequently, we follow the approach of SMPLify, which
estimates 2D features and then optimizes model parameters to fit the features.
We improve on SMPLify in several significant ways: (1) we detect 2D features
corresponding to the face, hands, and feet and fit the full SMPL-X model to
these; (2) we train a new neural network pose prior using a large MoCap
dataset; (3) we define a new interpenetration penalty that is both fast and
accurate; (4) we automatically detect gender and the appropriate body models
(male, female, or neutral); (5) our PyTorch implementation achieves a speedup
of more than 8x over Chumpy. We use the new method, SMPLify-X, to fit SMPL-X to
both controlled images and images in the wild. We evaluate 3D accuracy on a new
curated dataset comprising 100 images with pseudo ground-truth. This is a step
towards automatic expressive human capture from monocular RGB data. The models,
code, and data are available for research purposes at
https://smpl-x.is.tue.mpg.de. | ['Ahmed A. A. Osman', 'Vasileios Choutas', 'Georgios Pavlakos', 'Dimitrios Tzionas', 'Michael J. Black', 'Nima Ghorbani', 'Timo Bolkart'] | 2019-04-11 | expressive-body-capture-3d-hands-face-and-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Pavlakos_Expressive_Body_Capture_3D_Hands_Face_and_Body_From_a_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Pavlakos_Expressive_Body_Capture_3D_Hands_Face_and_Body_From_a_CVPR_2019_paper.pdf | cvpr-2019-6 | ['3d-human-reconstruction', '3d-multi-person-mesh-recovery'] | ['computer-vision', 'computer-vision'] | [-1.21229380e-01 2.39042155e-02 -3.66465212e-03 -4.59221452e-01
-5.72483599e-01 -4.60526139e-01 4.06716973e-01 -4.64956999e-01
-4.63668436e-01 5.52206397e-01 3.92685741e-01 3.95922989e-01
4.23079312e-01 -3.04626375e-01 -8.20135236e-01 -2.83889860e-01
-5.66339642e-02 7.65248895e-01 -4.03934605e-02 -6.35295734e-02
-2.04086632e-01 6.85455501e-01 -1.53324020e+00 1.03295401e-01
2.19000489e-01 1.12739551e+00 -2.94824183e-01 8.81025016e-01
4.40012068e-01 4.33681428e-01 -1.69029370e-01 -4.14887369e-01
6.31873310e-01 -3.02769333e-01 -7.43373275e-01 2.40576088e-01
6.55850530e-01 -7.47010529e-01 -2.37703383e-01 5.29328823e-01
7.00003982e-01 3.82065661e-02 4.69222277e-01 -1.45382607e+00
-5.40908091e-02 1.63031369e-01 -8.41903090e-01 -5.36713958e-01
6.40366197e-01 4.93176758e-01 6.49295509e-01 -8.79040420e-01
8.53270411e-01 1.49831152e+00 7.57366359e-01 9.09057379e-01
-1.30676341e+00 -8.51454735e-01 -5.14744371e-02 -4.37319905e-01
-1.54124665e+00 -5.39563835e-01 6.16678834e-01 -6.03345215e-01
6.74216807e-01 2.94808805e-01 1.24510324e+00 1.38443577e+00
-9.30788666e-02 8.52398753e-01 1.10555768e+00 -3.60280424e-01
-1.76698957e-02 -2.04581335e-01 -2.72889957e-02 9.84763801e-01
-3.28409135e-01 1.38042867e-01 -6.68416560e-01 -2.59569526e-01
1.13022947e+00 -1.34057179e-01 -1.91435665e-01 -6.43952310e-01
-1.10853624e+00 5.31712592e-01 2.56769389e-01 -1.99086517e-01
-3.74559581e-01 4.95558619e-01 2.77191520e-01 -1.32517159e-01
2.51139641e-01 3.62677693e-01 -5.80126345e-01 -2.72557437e-01
-7.13835955e-01 8.12314272e-01 7.57860839e-01 7.77826726e-01
6.40114903e-01 -2.73221374e-01 -2.35608499e-02 7.79174984e-01
1.96649849e-01 7.29775906e-01 1.56697184e-01 -1.54565787e+00
1.46577552e-01 5.70917010e-01 1.92865297e-01 -8.16481054e-01
-8.12068343e-01 1.04789883e-02 -4.66506779e-01 4.66062605e-01
5.37292182e-01 -2.56212831e-01 -9.49407697e-01 1.97821081e+00
6.67266130e-01 -2.30452165e-01 -3.88838619e-01 1.28983092e+00
8.14283967e-01 3.46893221e-01 5.04592173e-02 2.82904208e-01
1.18819869e+00 -7.78975666e-01 -2.53987879e-01 -3.32003713e-01
2.62490392e-01 -5.38160443e-01 1.18604887e+00 5.31162262e-01
-1.39052057e+00 -5.05509675e-01 -6.61817312e-01 -2.58003026e-01
-6.55858815e-02 4.20781493e-01 7.30874360e-01 3.92445117e-01
-9.05391514e-01 6.57319725e-01 -1.11287999e+00 -4.82132137e-01
2.12604851e-01 7.51027524e-01 -7.21051455e-01 2.62262076e-01
-8.56765747e-01 8.54879379e-01 2.15703294e-01 2.26635486e-01
-7.20061302e-01 -3.77604455e-01 -1.15457559e+00 -3.52648079e-01
3.89417797e-01 -9.01697218e-01 1.27329707e+00 -1.08730495e+00
-1.82171559e+00 1.48359776e+00 8.79325569e-02 -4.16892171e-02
8.95623684e-01 -6.74881697e-01 2.27177784e-01 2.63634652e-01
-1.06317870e-01 1.24170077e+00 6.94241703e-01 -1.25289702e+00
-7.08462819e-02 -5.79801679e-01 -1.10551491e-01 3.23316813e-01
4.41150665e-02 2.24231780e-01 -9.06779587e-01 -4.29783195e-01
-8.00919533e-02 -1.24275672e+00 -2.89452560e-02 6.32594347e-01
-5.25674164e-01 2.51359165e-01 4.02917057e-01 -1.00257444e+00
6.31703258e-01 -2.05026484e+00 6.29654109e-01 3.49873304e-01
3.40691119e-01 8.09997693e-02 -6.51191771e-02 2.70277970e-02
1.18299853e-02 -2.34690964e-01 -1.89353079e-01 -6.71308339e-01
2.11432815e-01 2.60185868e-01 1.27191886e-01 6.13980293e-01
1.15509741e-02 1.06354690e+00 -6.06909215e-01 -6.91240788e-01
3.96531820e-01 5.98879218e-01 -8.46708417e-01 3.69139165e-01
-2.21339583e-01 7.92261660e-01 -1.39620841e-01 8.38149965e-01
4.59060192e-01 -5.97730577e-02 1.95443958e-01 -2.99180567e-01
4.75917459e-02 -2.54261732e-01 -1.13769841e+00 1.82064939e+00
-2.60855496e-01 2.85870403e-01 4.67459708e-01 -3.05786312e-01
8.09397101e-01 2.10524768e-01 8.16826642e-01 -1.34678170e-01
5.40993571e-01 4.31178324e-03 -4.28662032e-01 -4.64975178e-01
1.37187496e-01 -3.18597585e-01 -2.93203533e-01 3.64549041e-01
2.26286590e-01 -4.72104818e-01 -9.16522443e-02 -1.77723348e-01
8.24797988e-01 8.46356690e-01 2.23511443e-01 1.53615251e-01
3.09925899e-02 -2.50232667e-01 4.79268342e-01 2.13045612e-01
-2.67068028e-01 7.65559912e-01 4.79907364e-01 -4.87752825e-01
-1.00031650e+00 -1.28470254e+00 2.17985183e-01 1.07035589e+00
-8.97458270e-02 -4.53576237e-01 -8.74069989e-01 -3.26062232e-01
2.46530652e-01 2.91783005e-01 -7.63900518e-01 1.28341476e-02
-6.53169870e-01 -3.47052962e-01 6.13515377e-01 7.11392343e-01
4.52592611e-01 -9.80581760e-01 -1.02240002e+00 -2.11617351e-01
-2.94360220e-01 -1.14425647e+00 -5.64090014e-01 -6.40793741e-02
-4.44392383e-01 -1.13579047e+00 -6.99163735e-01 -3.18217635e-01
6.61535859e-01 -5.72721541e-01 1.06182790e+00 3.92704681e-02
-5.34134567e-01 6.24090135e-01 -4.88835722e-02 -2.45035738e-01
-1.53206363e-01 -1.77439556e-01 2.96775609e-01 -7.58455396e-02
1.62452146e-01 -5.28988421e-01 -4.56120938e-01 3.16102803e-01
-3.91775817e-01 4.90896940e-01 2.93581307e-01 5.71683526e-01
5.94424963e-01 -6.63815141e-01 -1.24831647e-01 -3.71410668e-01
2.47124553e-01 8.63978453e-03 -6.52600884e-01 -4.48869802e-02
8.52038786e-02 -1.48533791e-01 1.06890045e-01 -5.62789977e-01
-8.55716407e-01 8.20551157e-01 -3.27056587e-01 -8.86426270e-01
-2.71059632e-01 1.34255663e-01 -2.44412243e-01 1.49965566e-03
6.89775527e-01 -1.26977056e-01 3.73454571e-01 -4.79411215e-01
3.19878370e-01 5.37314773e-01 9.94697869e-01 -9.07414675e-01
6.54084027e-01 5.36683023e-01 5.37307598e-02 -7.46247292e-01
-8.16178322e-01 -9.98586565e-02 -1.14654565e+00 -5.51351190e-01
1.00319183e+00 -9.10804868e-01 -1.22944856e+00 6.49736345e-01
-9.69756305e-01 -8.18668246e-01 -1.61672935e-01 5.71152508e-01
-8.64157856e-01 2.08943069e-01 -8.66219163e-01 -9.24494922e-01
-4.14727271e-01 -8.98219824e-01 1.62500405e+00 -6.14294782e-03
-9.28912044e-01 -5.20040274e-01 2.50316679e-01 5.30375242e-01
-2.36010134e-01 9.36235726e-01 3.86392236e-01 -3.68477078e-03
-1.02838658e-01 -3.40233773e-01 -1.49161387e-02 3.47739577e-01
-1.12991087e-01 2.33606681e-01 -9.15965378e-01 -2.03793317e-01
-4.58377451e-01 -8.96010756e-01 3.07432473e-01 5.06458640e-01
9.88042235e-01 -1.66135654e-01 -2.50515372e-01 9.24447656e-01
8.37625861e-01 -3.45466256e-01 2.85252333e-01 2.93766968e-02
7.73926854e-01 7.10456491e-01 6.35654449e-01 6.87183559e-01
3.40827316e-01 9.46792781e-01 3.82440984e-01 -2.90039897e-01
3.72086242e-02 -5.57004035e-01 3.82872254e-01 4.63147998e-01
-4.97692406e-01 3.47423434e-01 -1.00132382e+00 2.64341235e-01
-1.70296729e+00 -7.07617879e-01 1.83555871e-01 2.08448052e+00
9.62530613e-01 -5.68889007e-02 6.56757176e-01 -6.61648214e-02
4.87296134e-01 -2.33049288e-01 -5.47356248e-01 -1.71883196e-01
1.39704138e-01 3.23872060e-01 3.16290230e-01 5.85811794e-01
-1.15867853e+00 1.05336058e+00 6.25675535e+00 6.58408403e-02
-1.25862110e+00 -1.33045956e-01 3.88588339e-01 -7.51910627e-01
2.82610625e-01 -2.91905582e-01 -7.10064590e-01 3.06941569e-01
3.95511091e-01 2.19416693e-01 6.28861725e-01 8.60350311e-01
1.82439238e-01 -1.78645942e-02 -1.24463451e+00 1.28987694e+00
1.34068623e-01 -6.94729030e-01 -2.57042974e-01 9.59237814e-02
3.41287792e-01 -6.38526827e-02 -1.66298091e-01 2.15912193e-01
4.20928419e-01 -1.06677127e+00 1.09382617e+00 6.35729432e-01
9.53713477e-01 -6.10386252e-01 2.83711582e-01 4.26354438e-01
-9.86713052e-01 4.34872992e-02 -1.04403691e-02 -2.51745492e-01
2.73489088e-01 1.10702794e-02 -6.01345241e-01 1.49456769e-01
8.56134057e-01 5.53558707e-01 -3.59475851e-01 5.82977295e-01
-4.40527320e-01 2.17465594e-01 -7.07269073e-01 1.98872864e-01
-1.79809645e-01 -2.87547745e-02 4.40727532e-01 9.64803457e-01
1.06583931e-01 1.80769578e-01 3.08735520e-01 9.10372853e-01
-1.63091104e-02 -2.05420442e-02 -3.14152360e-01 1.81504712e-01
8.30801204e-02 1.27592373e+00 -5.06330192e-01 -7.13974163e-02
9.46626365e-02 1.16621339e+00 3.76337796e-01 8.12312812e-02
-8.78407359e-01 1.40352130e-01 7.83211350e-01 3.63938451e-01
3.40678357e-02 -3.29540968e-01 1.49366884e-02 -1.26965892e+00
-1.96627458e-03 -9.47079182e-01 2.46880457e-01 -1.28502226e+00
-1.05870128e+00 5.24666309e-01 3.45863640e-01 -8.91980648e-01
-6.00899577e-01 -8.21784377e-01 -1.24518216e-01 6.93065941e-01
-7.28144944e-01 -1.59944546e+00 -6.80774748e-01 6.85044110e-01
8.86923596e-02 3.56500953e-01 1.01717556e+00 1.80878058e-01
-4.83798116e-01 3.88190836e-01 -6.89059734e-01 4.17260677e-01
8.49221706e-01 -1.06934321e+00 3.95593643e-01 2.77488679e-01
-9.09581631e-02 4.33114409e-01 7.26297915e-01 -6.97859228e-01
-1.34668136e+00 -8.17618906e-01 4.52078283e-01 -9.38906252e-01
2.59793222e-01 -7.26466298e-01 -3.50967795e-01 1.14236021e+00
-2.35359967e-01 1.64889872e-01 6.28267229e-01 1.86466724e-01
-2.18174040e-01 1.88419744e-02 -1.16217232e+00 6.72863305e-01
1.21056068e+00 -2.75225282e-01 -4.67738420e-01 1.02522403e-01
1.30690679e-01 -1.00773633e+00 -9.11357343e-01 3.70681494e-01
1.34594572e+00 -9.31683958e-01 9.29031372e-01 -4.31560189e-01
5.15247881e-01 -1.23076200e-01 -2.28640229e-01 -1.12956548e+00
3.46639045e-02 -7.24183440e-01 -8.89397115e-02 8.95741284e-01
-3.63159068e-02 -2.41674274e-01 9.32450116e-01 9.56113279e-01
3.03341508e-01 -8.65061343e-01 -7.89456308e-01 -4.88453418e-01
6.07636059e-03 -4.54962969e-01 4.78456885e-01 6.97303891e-01
4.04457673e-02 1.75625518e-01 -7.76608348e-01 -9.60428789e-02
5.95953107e-01 1.60921112e-01 1.49882317e+00 -1.12963271e+00
-5.42390525e-01 -2.12147221e-01 -3.33748937e-01 -9.55670834e-01
3.65931779e-01 -5.93592644e-01 7.12237880e-03 -1.11252666e+00
3.49601775e-01 -1.21580489e-01 2.90343702e-01 1.07129848e+00
2.10788697e-02 6.36566222e-01 3.84501725e-01 2.34178528e-02
-2.81381875e-01 5.03730953e-01 1.19986558e+00 2.88665563e-01
-1.24673106e-01 -1.24062225e-01 -2.23531380e-01 1.12444210e+00
5.78599334e-01 -1.28101125e-01 4.72040214e-02 -2.63354868e-01
7.86402375e-02 2.72047192e-01 9.14802849e-01 -9.66629803e-01
-2.93197840e-01 -1.91078961e-01 9.32610810e-01 -5.48226058e-01
1.01801848e+00 -8.01651776e-01 4.26269174e-01 4.68502939e-01
-1.94405705e-01 -1.57881871e-01 2.64700711e-01 -1.28769517e-01
2.86528319e-01 2.44994208e-01 8.51812482e-01 -4.58078563e-01
-3.60574394e-01 2.69094110e-01 -3.07702497e-02 -6.56426027e-02
9.18121517e-01 -2.71130085e-01 2.89680094e-01 -6.24973893e-01
-1.00415730e+00 2.99597472e-01 9.35832441e-01 3.94670665e-01
4.76948649e-01 -1.36614907e+00 -6.04297101e-01 4.39509273e-01
-3.09081543e-02 6.15554154e-02 -4.53600250e-02 6.85105443e-01
-7.38877296e-01 -4.99118567e-02 -5.91158390e-01 -6.61394477e-01
-1.70399392e+00 2.64120996e-01 6.23027205e-01 4.71953414e-02
-5.75641453e-01 7.95464754e-01 2.41570726e-01 -9.48718369e-01
1.89179793e-01 -2.30751574e-01 3.14761400e-01 -3.44787478e-01
2.35036060e-01 3.30589592e-01 -5.42162180e-01 -1.01645732e+00
-3.86811972e-01 9.70300853e-01 4.70060319e-01 -4.56160039e-01
1.41034734e+00 3.51532847e-02 -1.95057429e-02 3.96082193e-01
1.11474442e+00 1.87997505e-01 -1.68437886e+00 1.63968280e-01
-5.28415918e-01 -5.04225731e-01 -2.58998632e-01 -8.61196160e-01
-1.01229322e+00 7.96559930e-01 5.03363788e-01 -6.33412421e-01
1.01645923e+00 2.19900846e-01 7.20600545e-01 4.70013544e-02
4.67249185e-01 -1.18567455e+00 1.14222400e-01 2.66250640e-01
1.22160733e+00 -9.43195224e-01 2.18030691e-01 -3.79930705e-01
-6.67766750e-01 8.90419781e-01 8.60521257e-01 -7.01858997e-02
4.58513469e-01 4.53828454e-01 2.62492627e-01 -3.16362113e-01
-4.41037714e-01 -9.10991356e-02 2.65710026e-01 3.15993160e-01
3.28611165e-01 1.31904885e-01 1.19711444e-01 6.78378224e-01
-6.60424054e-01 4.42746639e-01 -1.01350024e-01 1.03462005e+00
6.71705455e-02 -9.17253613e-01 -6.08087480e-01 1.62076503e-01
-3.27074736e-01 4.25094426e-01 -8.33310187e-01 9.70671892e-01
3.20510179e-01 4.07771468e-01 -4.93764430e-02 -5.34960985e-01
5.51306605e-01 1.23268336e-01 9.81504738e-01 -4.90403801e-01
-4.91735339e-01 1.95432365e-01 8.94335955e-02 -9.13877964e-01
-2.88464904e-01 -7.25738585e-01 -1.34766698e+00 -2.95570016e-01
-1.05789691e-01 -3.28375906e-01 5.93343914e-01 7.30145156e-01
2.65704513e-01 4.21381332e-02 2.12461725e-01 -1.62671936e+00
-4.77248251e-01 -8.99115741e-01 -6.21846259e-01 5.14204443e-01
1.11313485e-01 -9.76974547e-01 -1.53866872e-01 1.63851991e-01] | [7.09281063079834, -1.0978424549102783] |
365f4322-e442-41e3-bda2-7e614d312a5c | probabilistic-loss-and-its-online | 2105.05789 | null | https://arxiv.org/abs/2105.05789v1 | https://arxiv.org/pdf/2105.05789v1.pdf | Probabilistic Loss and its Online Characterization for Simplified Decision Making Under Uncertainty | It is a long-standing objective to ease the computation burden incurred by the decision making process. Identification of this mechanism's sensitivity to simplification has tremendous ramifications. Yet, algorithms for decision making under uncertainty usually lean on approximations or heuristics without quantifying their effect. Therefore, challenging scenarios could severely impair the performance of such methods. In this paper, we extend the decision making mechanism to the whole by removing standard approximations and considering all previously suppressed stochastic sources of variability. On top of this extension, our key contribution is a novel framework to simplify decision making while assessing and controlling online the simplification's impact. Furthermore, we present novel stochastic bounds on the return and characterize online the effect of simplification using this framework on a particular simplification technique - reducing the number of samples in belief representation for planning. Finally, we verify the advantages of our approach through extensive simulations. | ['Vadim Indelman', 'Andrey Zhitnikov'] | 2021-05-12 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 2.46166229e-01 4.11424905e-01 -1.15350991e-01 -3.29148173e-01
-6.75892532e-01 -5.97101808e-01 5.49827754e-01 5.38148105e-01
-5.37916660e-01 9.16123390e-01 -7.12182298e-02 -5.26758134e-01
-6.67994201e-01 -1.09547114e+00 -7.15597391e-01 -5.61325490e-01
-1.19725130e-01 6.00710511e-01 1.24150023e-01 -6.48547411e-02
5.15395224e-01 5.93399584e-01 -1.38738787e+00 -3.74196380e-01
1.04946661e+00 1.04779553e+00 1.52161062e-01 4.13938820e-01
9.29153264e-02 5.95170498e-01 -7.00951695e-01 -3.25329304e-01
5.42049825e-01 -8.35846215e-02 -5.08495271e-01 4.57913876e-02
-5.01119494e-01 -7.35049188e-01 1.91355333e-01 1.09680867e+00
5.97263277e-01 4.12049532e-01 6.00328863e-01 -1.14602923e+00
2.26436928e-01 9.69402909e-01 -4.83958304e-01 2.55679518e-01
2.54590660e-01 1.57486960e-01 5.82908332e-01 -2.67050266e-01
4.99459207e-01 1.43795550e+00 2.39507139e-01 -1.96744353e-02
-1.23052680e+00 -3.96204740e-01 5.62397182e-01 -3.77816968e-02
-1.56739914e+00 -5.30962765e-01 6.53851748e-01 -4.75398570e-01
7.15860188e-01 1.35829493e-01 5.41893899e-01 6.37391567e-01
3.34834248e-01 6.42425597e-01 1.28503239e+00 -6.84615791e-01
7.29689181e-01 1.37245074e-01 1.47855639e-01 4.02127653e-01
8.19918156e-01 3.43057066e-01 -2.03508779e-01 -3.13264966e-01
2.57879257e-01 -1.80919960e-01 -1.47106424e-02 -4.92460459e-01
-4.86940742e-01 7.38915980e-01 -1.17497846e-01 -1.93653792e-01
-3.88771892e-01 1.80908144e-01 2.69702524e-01 1.61443785e-01
3.15751195e-01 3.91482323e-01 -1.08547896e-01 -3.57664019e-01
-8.06874037e-01 5.37149489e-01 8.05035412e-01 9.09108698e-01
5.03364980e-01 -1.54449865e-01 -5.44954181e-01 3.82793039e-01
2.06291929e-01 3.75867456e-01 -1.71533093e-01 -9.68749940e-01
6.31576777e-01 3.24593663e-01 7.54314423e-01 -9.36084926e-01
-5.14011979e-01 -7.12341368e-01 -3.16743106e-01 3.66152555e-01
5.25218248e-01 -4.19358850e-01 -6.22535288e-01 1.60477114e+00
5.18579066e-01 -4.86063272e-01 -1.18637614e-01 7.33941793e-01
-3.61637861e-01 4.71536458e-01 -4.58196066e-02 -5.58155835e-01
8.79222691e-01 -4.02659476e-01 -8.41853321e-01 1.20283306e-01
4.45331126e-01 -4.63854581e-01 7.92944074e-01 5.77841759e-01
-1.29028535e+00 -5.89324869e-02 -1.03388309e+00 4.69595075e-01
-1.13888629e-01 -2.18670428e-01 7.25009382e-01 9.93191004e-01
-5.55868566e-01 9.42872763e-01 -9.28913474e-01 -5.05533107e-02
9.66652855e-02 3.75035495e-01 5.54606974e-01 -1.45050541e-01
-1.08351505e+00 1.28167605e+00 4.53006297e-01 3.04966807e-01
-9.56376433e-01 -5.71295261e-01 -5.52072108e-01 2.84923285e-01
1.17714965e+00 -6.99888647e-01 1.62319398e+00 -6.23915136e-01
-1.80667865e+00 -8.69459361e-02 5.65429293e-02 -8.08703244e-01
1.23187184e+00 -9.27027389e-02 -3.12104672e-02 -2.88374782e-01
1.00246765e-01 1.78012103e-01 6.53723419e-01 -1.27344251e+00
-7.21873820e-01 -3.64908457e-01 5.11071324e-01 3.60434949e-01
1.02502413e-01 -2.16580898e-01 -1.99337766e-01 -3.02694857e-01
-1.23884037e-01 -9.71211255e-01 -8.01038682e-01 -5.61168849e-01
-5.43165267e-01 2.47064039e-01 1.48722986e-02 -3.36967200e-01
1.42635834e+00 -1.94119942e+00 -1.05817892e-01 5.81537962e-01
-3.11349243e-01 -2.48605371e-01 2.50625998e-01 7.82867432e-01
5.29173195e-01 2.25322098e-01 -3.98761749e-01 -2.10191682e-01
4.31552172e-01 2.94419616e-01 -4.33400363e-01 5.66299498e-01
-9.22796223e-03 2.51962662e-01 -7.63842106e-01 -2.77982295e-01
4.56830323e-01 3.21749859e-02 -5.89019716e-01 -1.00769728e-01
-3.58260930e-01 1.17722914e-01 -6.40201390e-01 7.54343092e-01
6.78749442e-01 1.69649184e-01 4.63492364e-01 1.91013828e-01
-2.90985495e-01 2.92980641e-01 -1.50631356e+00 9.72808659e-01
-6.39321089e-01 -4.22037728e-02 2.47422308e-01 -8.25036585e-01
5.60621381e-01 -3.30373406e-01 2.60639310e-01 -4.64560091e-01
4.40515518e-01 1.41991317e-01 2.99725115e-01 -6.99300393e-02
6.62001610e-01 -1.29928336e-01 -2.27783337e-01 4.16279554e-01
-5.03697634e-01 -3.01524043e-01 4.10350502e-01 -3.48153114e-02
8.09201598e-01 5.91820776e-02 7.20636964e-01 -4.84628230e-01
1.99144825e-01 -2.60518435e-02 7.09216475e-01 9.65511262e-01
-3.50162387e-01 -7.77162537e-02 7.28817582e-01 -3.76965329e-02
-6.42078400e-01 -6.93902910e-01 -2.79844999e-01 5.90050995e-01
2.74567127e-01 -8.76718462e-02 -6.80477142e-01 -3.64422381e-01
4.66856390e-01 1.54690182e+00 -6.62214816e-01 -5.62895723e-02
-3.75180840e-02 -1.08374238e+00 2.22494751e-01 2.33085394e-01
1.57627672e-01 -1.80480853e-01 -1.10472739e+00 3.46200138e-01
-1.59234982e-02 -9.44998384e-01 2.30976641e-02 2.28415549e-01
-8.70543003e-01 -9.29730713e-01 -5.22685587e-01 4.98069555e-01
7.11110532e-01 1.67263851e-01 8.31390142e-01 -3.66251081e-01
4.42123041e-02 4.90301400e-01 -2.94149607e-01 -9.40589666e-01
-4.32151139e-01 -7.80317411e-02 1.73048362e-01 -9.04231146e-02
-2.71938622e-01 -3.91658992e-01 -3.51516753e-01 1.97525650e-01
-6.09786093e-01 -5.53213432e-02 6.25323355e-01 4.25804764e-01
6.06179178e-01 6.53979838e-01 4.00128126e-01 -9.70703363e-01
1.04578519e+00 -5.02112925e-01 -1.16193342e+00 3.90395373e-01
-9.67732608e-01 4.22237188e-01 4.41309601e-01 -1.34920463e-01
-1.41119754e+00 -1.17570288e-01 4.07665253e-01 -1.46704301e-01
1.38694257e-01 6.84030890e-01 -2.62524188e-01 -1.05178788e-01
1.82193607e-01 -4.41914089e-02 -2.43034676e-01 -1.72268271e-01
1.65542826e-01 2.12250382e-01 2.73834225e-02 -1.02434170e+00
6.69990182e-01 4.53219652e-01 4.78519708e-01 -5.80824494e-01
-8.12426627e-01 1.00188628e-02 -2.45310366e-01 -3.98217142e-01
2.21281365e-01 -5.55627942e-01 -7.74822652e-01 2.10864050e-03
-8.27808321e-01 -2.97103673e-01 -3.96941781e-01 4.79757607e-01
-7.11627662e-01 5.00258446e-01 -1.10211305e-01 -1.61416233e+00
9.16540399e-02 -1.18190336e+00 4.80753243e-01 2.21578047e-01
-8.92057940e-02 -6.91958249e-01 -4.72884811e-03 2.42800251e-01
2.89761752e-01 4.20579463e-01 7.71108031e-01 -4.69744623e-01
-8.28843951e-01 -5.85052855e-02 -8.14476386e-02 4.31753770e-02
-2.42971614e-01 8.07139054e-02 -7.16275573e-01 -3.30297351e-01
1.13677725e-01 1.06113158e-01 6.05238914e-01 6.57175303e-01
1.06685197e+00 -3.76340687e-01 -2.41390407e-01 3.27417195e-01
1.50187635e+00 5.70398629e-01 3.66601974e-01 6.45272553e-01
-7.74128065e-02 8.79774749e-01 1.19926596e+00 1.10399354e+00
3.44663322e-01 8.26961696e-01 6.29904270e-01 5.09069920e-01
4.46280777e-01 -1.33297861e-01 3.58714402e-01 3.30262065e-01
-3.96521449e-01 -4.90145564e-01 -9.48526978e-01 3.74973029e-01
-1.83197486e+00 -7.08450079e-01 3.12352777e-01 2.70378423e+00
4.36248511e-01 5.46445012e-01 1.34892121e-01 3.65448654e-01
6.54622257e-01 -1.13094710e-01 -5.80120862e-01 -7.20976412e-01
1.44696474e-01 -2.21941710e-01 1.18303251e+00 7.13406205e-01
-6.97228253e-01 4.24630284e-01 6.21176100e+00 8.25997710e-01
-7.31336176e-01 -1.21284321e-01 6.83424950e-01 -3.56349498e-01
-4.86491174e-01 2.07589433e-01 -9.74669695e-01 6.33699238e-01
1.13238966e+00 -6.64480150e-01 5.68622828e-01 8.04134429e-01
5.85098207e-01 -8.94531071e-01 -1.13167167e+00 4.32839245e-01
-3.30993742e-01 -1.14992857e+00 -8.62897411e-02 2.44581491e-01
5.82476139e-01 -5.15171409e-01 -3.23517248e-03 2.79232472e-01
8.04149449e-01 -7.67794609e-01 1.12218666e+00 6.43583417e-01
3.71490955e-01 -1.29222786e+00 9.49500978e-01 4.52909678e-01
-6.87283278e-01 -4.94736969e-01 -2.06378490e-01 -4.77724552e-01
4.26179588e-01 1.04732251e+00 -1.17133403e+00 8.41534734e-01
2.16031492e-01 -2.90463746e-01 -2.98367441e-01 1.12146974e+00
-8.93095806e-02 5.53235769e-01 -5.94854891e-01 -2.23885760e-01
3.25441569e-01 -3.14107239e-01 5.57377934e-01 1.09283257e+00
4.13272202e-01 2.38200054e-01 3.26477379e-01 8.59825194e-01
5.01204193e-01 -7.31600598e-02 -5.43621302e-01 -6.56739175e-02
8.84288728e-01 5.35363734e-01 -1.01834548e+00 -1.01835160e-02
1.10351043e-02 3.56161326e-01 1.68081179e-01 2.81206429e-01
-8.54760647e-01 -1.07113957e-01 3.64892662e-01 2.12517232e-02
1.76031515e-01 -3.40791941e-01 -5.89182377e-01 -5.82389951e-01
1.90988988e-01 -4.85547096e-01 3.84632379e-01 -2.16873020e-01
-7.21325696e-01 -1.03403166e-01 5.08407593e-01 -9.20531273e-01
-4.15859878e-01 -2.25415781e-01 -3.96541506e-01 7.57512987e-01
-1.63332915e+00 -5.55555522e-01 1.48073256e-01 1.24598823e-01
3.17833215e-01 3.22473228e-01 3.30419630e-01 6.55419081e-02
-7.77332723e-01 5.12928069e-01 4.12210435e-01 -5.63172162e-01
1.89799294e-01 -9.85589206e-01 -1.05701976e-01 1.14777458e+00
-5.70992947e-01 5.31550050e-01 1.06761467e+00 -7.59621203e-01
-1.41031694e+00 -7.59810805e-01 5.03444552e-01 -1.57937422e-01
6.13684535e-01 -1.94744006e-01 -3.13376814e-01 3.93942893e-01
-3.37134063e-01 -6.08571231e-01 3.73120010e-01 3.13407212e-01
2.40700915e-01 -1.57939672e-01 -1.40107000e+00 7.77059257e-01
8.31924498e-01 -1.55971751e-01 -3.92514706e-01 2.20324278e-01
6.06254637e-01 -3.98452312e-01 -7.20099866e-01 4.73181516e-01
5.14519811e-01 -1.10884249e+00 6.77892447e-01 -1.19956590e-01
1.63393512e-01 -1.78706244e-01 -3.95848393e-01 -1.15509498e+00
-3.91447693e-02 -9.58193600e-01 -1.01926684e-01 1.13811648e+00
2.96251506e-01 -7.71470010e-01 5.79363823e-01 9.28999960e-01
-1.21418685e-01 -7.95159459e-01 -1.04817390e+00 -1.10576820e+00
-1.06055550e-01 -5.76154649e-01 6.93082154e-01 5.32247424e-01
-2.62261450e-01 -1.86416045e-01 -1.94402561e-01 3.44809800e-01
8.25200617e-01 2.44858682e-01 7.07029700e-01 -9.76388812e-01
-4.31099385e-01 -3.49527776e-01 -1.73018146e-02 -5.38619757e-01
-1.61890030e-01 -2.99604446e-01 4.78046089e-02 -1.31094003e+00
8.18972811e-02 -8.18214476e-01 -2.84918338e-01 -1.13787405e-01
1.72301054e-01 -8.95169437e-01 5.13434291e-01 1.10490687e-01
-4.25199628e-01 6.19885385e-01 1.21729541e+00 8.95344019e-02
-2.74467498e-01 4.30896938e-01 -6.75857723e-01 9.13089693e-01
7.43999958e-01 -4.63488460e-01 -7.25720942e-01 -2.52159148e-01
4.71895337e-01 5.63934267e-01 9.96423513e-02 -8.57815862e-01
8.02897066e-02 -7.56768405e-01 -1.70527413e-01 -7.54115283e-01
2.14907110e-01 -9.48734999e-01 3.58247280e-01 6.90523267e-01
-4.29230899e-01 -8.65506381e-02 2.14321092e-01 1.03354084e+00
1.27186447e-01 -6.40677989e-01 5.63115478e-01 -8.83840322e-02
-5.09580195e-01 -1.56841174e-01 -4.90321010e-01 -1.26278818e-01
1.27491069e+00 -1.76794291e-01 6.33090138e-02 -4.50947285e-01
-6.83334947e-01 4.41513062e-01 5.75394928e-01 -1.57744944e-01
1.34242862e-01 -7.19368160e-01 -2.81770587e-01 -3.46874952e-01
-2.67243862e-01 -7.23870145e-03 3.69279295e-01 7.96319306e-01
-3.37521285e-01 7.04786181e-01 -1.86091557e-01 -1.86937779e-01
-7.83402920e-01 5.96729338e-01 3.09404343e-01 -6.66436613e-01
-2.23771423e-01 5.57015240e-01 6.43090392e-03 -2.05012918e-01
4.72431749e-01 -6.43878937e-01 -4.03655544e-02 3.06920528e-01
2.59897947e-01 8.70061278e-01 1.99082822e-01 5.82787283e-02
-4.25408840e-01 4.10424210e-02 2.46160701e-02 -4.93085772e-01
1.03060722e+00 -6.11456692e-01 2.62352169e-01 5.61859727e-01
2.55411386e-01 1.61036536e-01 -1.32217348e+00 1.17458314e-01
2.73223430e-01 -5.74071407e-01 2.72197157e-01 -9.79898512e-01
-6.32479370e-01 6.21214151e-01 4.14592296e-01 1.52401075e-01
9.72171843e-01 -6.12038314e-01 2.15848476e-01 7.55199254e-01
8.56726706e-01 -1.58175373e+00 -6.92303121e-01 2.61371255e-01
7.98718989e-01 -9.56532300e-01 4.77537125e-01 -6.70141935e-01
-4.39072847e-01 8.14172983e-01 2.18237191e-01 9.64419022e-02
3.67278099e-01 3.96621972e-01 -3.90545964e-01 1.41034186e-01
-8.05625498e-01 -1.57277390e-01 -2.71542221e-01 4.85238999e-01
-8.52039978e-02 5.20890594e-01 -6.75925732e-01 8.02227080e-01
-1.96238801e-01 1.99151769e-01 1.06286073e+00 1.21680951e+00
-4.16617453e-01 -8.98221195e-01 -6.34892821e-01 5.05890965e-01
-5.08077741e-01 1.65055096e-01 -2.99882330e-02 9.91530478e-01
-3.96118760e-02 1.00596774e+00 -9.57039148e-02 -8.07255805e-02
4.13441598e-01 -2.30160609e-01 4.85981435e-01 -4.65712607e-01
-3.49305212e-01 -3.65118124e-02 6.55232370e-01 -5.86474061e-01
-2.16300741e-01 -9.44963515e-01 -9.27636683e-01 -3.62794250e-01
-3.89229894e-01 1.76488653e-01 7.89102793e-01 1.00067174e+00
2.75491416e-01 6.62017405e-01 5.37970006e-01 -7.74585724e-01
-1.48428977e+00 -5.46949327e-01 -6.57439888e-01 -2.36925125e-01
-4.52452078e-02 -1.30248225e+00 -3.96038413e-01 -6.47350073e-01] | [4.721267223358154, 2.9532253742218018] |
91de13a7-a0d9-45f6-a3b5-22a8d847abbe | deep-semantic-role-labeling-what-works-and | null | null | https://aclanthology.org/P17-1044 | https://aclanthology.org/P17-1044.pdf | Deep Semantic Role Labeling: What Works and What's Next | We introduce a new deep learning model for semantic role labeling (SRL) that significantly improves the state of the art, along with detailed analyses to reveal its strengths and limitations. We use a deep highway BiLSTM architecture with constrained decoding, while observing a number of recent best practices for initialization and regularization. Our 8-layer ensemble model achieves 83.2 F1 on theCoNLL 2005 test set and 83.4 F1 on CoNLL 2012, roughly a 10{\%} relative error reduction over the previous state of the art. Extensive empirical analysis of these gains show that (1) deep models excel at recovering long-distance dependencies but can still make surprisingly obvious errors, and (2) that there is still room for syntactic parsers to improve these results. | ['Luke Zettlemoyer', 'Mike Lewis', 'Kenton Lee', 'Luheng He'] | 2017-07-01 | null | null | null | acl-2017-7 | ['predicate-detection'] | ['natural-language-processing'] | [ 3.94054323e-01 5.64896464e-01 -3.93804491e-01 -6.19343996e-01
-1.13460135e+00 -6.91439867e-01 4.01552945e-01 1.57890216e-01
-8.02042842e-01 8.39180768e-01 6.03194535e-01 -6.34748399e-01
2.27113329e-02 -2.62255996e-01 -7.85803974e-01 -5.14303923e-01
-9.31102186e-02 5.08634090e-01 2.30793983e-01 -4.53078151e-01
1.89697929e-02 8.19264427e-02 -1.31517923e+00 6.37267053e-01
5.57273686e-01 8.06299448e-01 5.92048280e-02 7.06147194e-01
2.02239510e-02 1.56108940e+00 -1.04882288e+00 -3.90578300e-01
-3.50605786e-01 -2.62579352e-01 -1.28597748e+00 -3.90972406e-01
5.02527893e-01 -3.46627593e-01 -3.00591946e-01 7.33512461e-01
5.82185566e-01 1.56070605e-01 2.62789607e-01 -8.08831036e-01
-9.24275339e-01 1.04873037e+00 -2.30163351e-01 3.16667676e-01
5.12953289e-03 -7.02043399e-02 1.60750008e+00 -6.85389340e-01
4.73727554e-01 1.30864966e+00 6.77190542e-01 8.49915922e-01
-1.05174983e+00 -6.42521381e-01 4.98918682e-01 1.92115247e-01
-8.61367643e-01 -5.96578658e-01 3.99403691e-01 -1.03501014e-01
1.52035177e+00 -2.52513457e-02 1.31991655e-01 1.31163669e+00
-2.31129721e-01 1.02013910e+00 9.58923757e-01 -5.72259068e-01
-1.73824996e-01 -3.96526486e-01 7.13105917e-01 7.70027876e-01
2.53375530e-01 -2.44815782e-01 -6.80740476e-01 -4.45869677e-02
4.81524408e-01 -6.95536017e-01 6.87127709e-02 3.12219709e-01
-9.17666912e-01 8.12403679e-01 3.46955210e-01 5.27038336e-01
2.96565741e-01 7.33066797e-01 4.37725693e-01 4.69044566e-01
7.11370826e-01 6.00525856e-01 -8.45432162e-01 -3.73026103e-01
-6.76610589e-01 1.03265151e-01 6.63184881e-01 6.32862568e-01
2.74547905e-01 3.38607460e-01 -4.39377762e-02 1.12863517e+00
3.57148230e-01 2.15450928e-01 3.29293966e-01 -1.32939982e+00
4.75970149e-01 4.56398904e-01 -9.60786566e-02 -4.70402241e-01
-5.15792429e-01 -5.56422412e-01 -2.66077638e-01 1.10727269e-02
6.57752335e-01 -1.21151105e-01 -1.07048011e+00 2.19229841e+00
-4.79274064e-01 2.55474169e-02 3.91341783e-02 3.43249679e-01
9.45268095e-01 3.44838470e-01 7.11483240e-01 1.29711136e-01
1.19071233e+00 -9.78732944e-01 -6.89452052e-01 -7.53193378e-01
1.14936674e+00 -4.53582078e-01 1.21046340e+00 2.08726391e-01
-1.21722984e+00 -5.11196971e-01 -1.21563351e+00 -5.15163302e-01
-8.26487169e-02 2.30660751e-01 8.80907118e-01 7.98267365e-01
-1.38523555e+00 6.50241554e-01 -9.41009939e-01 -1.65182143e-01
5.87425232e-01 4.92366791e-01 -3.43187273e-01 -1.80222064e-01
-1.41702247e+00 1.00772989e+00 3.91696990e-01 -4.33899052e-02
-9.73681927e-01 -6.90854788e-01 -1.16089332e+00 7.39260092e-02
2.88206726e-01 -3.62142563e-01 1.68423545e+00 -8.33235025e-01
-1.24634743e+00 1.28204989e+00 -3.45603317e-01 -7.27685273e-01
9.42950025e-02 -5.59090555e-01 -2.16447189e-01 -1.50929596e-02
1.40436977e-01 7.83269048e-01 2.02007607e-01 -1.17210901e+00
-7.65630841e-01 -1.90211296e-01 4.45495665e-01 3.48467171e-01
-2.78803855e-01 4.14473355e-01 -1.71972439e-01 -7.14087129e-01
-1.54194917e-04 -7.99159348e-01 -2.08058774e-01 -3.94537061e-01
-1.90337047e-01 -6.63801312e-01 4.16024625e-01 -8.52599680e-01
1.13896668e+00 -1.93625236e+00 3.53845232e-03 -3.57993454e-01
2.09579706e-01 3.66233885e-01 -3.46239865e-01 7.11745694e-02
-3.05789888e-01 6.29981637e-01 -5.23488879e-01 -7.82949150e-01
-9.59591344e-02 4.57091510e-01 -2.62811273e-01 3.15551311e-01
2.98866957e-01 9.16013777e-01 -8.92412841e-01 -1.78763326e-02
-4.68081012e-02 2.03356087e-01 -5.10038257e-01 -5.28666116e-02
-3.95459980e-01 5.41786849e-01 -9.06131566e-02 7.01547384e-01
2.85013199e-01 -3.80933493e-01 5.42026222e-01 1.92057714e-01
9.87268146e-03 1.11148751e+00 -5.33207774e-01 1.89832461e+00
-4.11853582e-01 7.68021107e-01 2.31488347e-01 -1.15016747e+00
5.21099567e-01 3.18079472e-01 1.88028648e-01 -6.55896008e-01
1.75348997e-01 3.85369092e-01 4.49510336e-01 -1.24798626e-01
2.22140759e-01 -9.36526731e-02 -1.73188508e-01 4.82791215e-01
1.99184731e-01 1.11137167e-01 7.91990981e-02 4.91588551e-04
1.43239605e+00 1.81291208e-01 6.62700310e-02 -5.18149495e-01
4.48488146e-01 -1.36974957e-02 6.66744828e-01 8.86771739e-01
-2.91957617e-01 5.39462149e-01 7.52111197e-01 -3.60153079e-01
-8.29882383e-01 -9.21517789e-01 3.71793495e-03 1.66687834e+00
-1.66235417e-01 -3.55902791e-01 -5.87858498e-01 -1.01712334e+00
-1.53732687e-01 1.15886605e+00 -5.91889799e-01 -9.91922915e-02
-1.11387062e+00 -1.05246222e+00 1.20424712e+00 1.05868196e+00
5.77570796e-01 -1.14553475e+00 -4.29937571e-01 1.49819672e-01
-3.31897289e-01 -1.33247042e+00 8.27050880e-02 5.54114044e-01
-8.13135982e-01 -1.15631807e+00 -2.60776609e-01 -1.00479364e+00
3.12227130e-01 -1.75687671e-03 1.60558975e+00 3.35940570e-01
2.27227852e-01 -3.25942747e-02 -4.06127244e-01 -4.18125212e-01
-4.49306965e-01 3.32913965e-01 -1.83276102e-01 -7.38724828e-01
3.68069172e-01 -5.37487626e-01 -2.52844900e-01 3.14793177e-02
-4.87514704e-01 -8.88597369e-02 3.90832365e-01 1.04956114e+00
4.42261964e-01 -1.76216677e-01 7.34862864e-01 -1.21784341e+00
4.97790754e-01 -4.40898806e-01 -2.77187616e-01 1.87028646e-01
-6.69502914e-01 1.89425632e-01 3.94224435e-01 2.23535731e-01
-1.25452459e+00 -3.28468561e-01 -6.16554260e-01 2.14181677e-01
-1.87931195e-01 4.02060449e-01 -2.77374178e-01 2.15436593e-01
6.91473603e-01 -2.20298812e-01 -1.58245668e-01 -8.10256004e-01
3.51266801e-01 4.68068302e-01 5.22670865e-01 -8.58303547e-01
2.74785578e-01 3.79820645e-01 -2.61282116e-01 -5.77984810e-01
-1.69587195e+00 -8.93031433e-02 -5.60589015e-01 4.61193383e-01
1.05968285e+00 -1.18454552e+00 -6.02988303e-01 4.75233525e-01
-1.31790686e+00 -9.09874082e-01 -2.18779594e-01 8.86115879e-02
-3.63651067e-01 1.39778644e-01 -1.05496848e+00 -5.15131772e-01
-2.49727979e-01 -1.03632689e+00 9.35642362e-01 -1.18113458e-01
-3.62886190e-01 -1.17272592e+00 -1.04795307e-01 7.75735259e-01
2.53247082e-01 -6.43723533e-02 1.31942439e+00 -1.06532097e+00
-2.15706334e-01 1.93160355e-01 -3.19624037e-01 7.78551459e-01
-9.31489989e-02 -5.20217657e-01 -1.29389226e+00 -2.65737832e-01
-7.28998035e-02 -5.64388454e-01 1.15872741e+00 4.69427407e-01
1.30749869e+00 7.68314600e-02 -1.86526641e-01 5.36806107e-01
8.42367887e-01 3.78578275e-01 5.96148074e-01 4.46936786e-01
9.75276649e-01 5.18454313e-01 4.81410742e-01 -2.12953180e-01
6.93372846e-01 4.13064748e-01 3.83361936e-01 -3.20237279e-02
-3.92548263e-01 -2.91938782e-01 5.22779465e-01 8.12657893e-01
-8.97863358e-02 -4.39445198e-01 -1.06588995e+00 6.35791600e-01
-2.02156997e+00 -7.39464402e-01 -1.97305560e-01 1.88854182e+00
8.86831224e-01 5.02267420e-01 -1.14446275e-01 2.48571247e-01
5.26325226e-01 5.59262991e-01 -6.28011167e-01 -5.49956322e-01
-4.15290445e-01 7.15207815e-01 7.32949853e-01 7.27282226e-01
-1.21565282e+00 1.55739450e+00 7.66469574e+00 7.42248774e-01
-6.04189932e-01 5.21028638e-01 9.60882127e-01 -1.85653102e-02
-4.51781154e-01 -7.95088243e-03 -1.00123966e+00 1.63529441e-01
1.23156333e+00 1.23139724e-01 2.10731402e-01 7.19563007e-01
-1.47143140e-01 -3.76862995e-02 -1.01958883e+00 6.98594689e-01
1.96350411e-01 -1.24849296e+00 -2.64634378e-02 -1.92907155e-01
5.30805826e-01 3.92606676e-01 3.93751077e-02 5.78313410e-01
9.78243291e-01 -1.63519442e+00 7.28796542e-01 -1.84897721e-01
9.45383132e-01 -7.21390545e-01 8.84564936e-01 2.16482952e-01
-7.75036812e-01 -1.99467763e-01 -3.16994548e-01 -5.22473633e-01
2.93176267e-02 2.71875858e-01 -6.30523026e-01 3.31609577e-01
7.38884628e-01 1.02032387e+00 -6.91952944e-01 3.38925481e-01
-1.14615870e+00 1.25876260e+00 -1.87947914e-01 1.90921366e-01
3.32304507e-01 4.60206926e-01 3.23083401e-01 1.28608346e+00
-7.58595243e-02 -2.48379284e-03 -2.90258415e-02 5.41954637e-01
-5.10720491e-01 -3.02634925e-01 -4.59509671e-01 -7.57135898e-02
6.50840878e-01 7.60382473e-01 -5.71338832e-01 -1.85643777e-01
-3.23676735e-01 8.68652344e-01 6.67101324e-01 2.45106861e-01
-8.64060819e-01 -6.63167238e-02 9.00772691e-01 -1.13075294e-01
9.30686761e-03 -4.21578735e-01 -7.35773027e-01 -1.01153040e+00
-2.03675836e-01 -9.08864439e-01 7.91875720e-01 -6.74280286e-01
-1.00833488e+00 5.75962305e-01 7.14038312e-02 -4.58504885e-01
-2.91643560e-01 -1.06145811e+00 -2.63647139e-01 7.40262330e-01
-1.67313635e+00 -1.32246554e+00 1.73827812e-01 1.78967103e-01
8.40782583e-01 -1.84473410e-01 1.01850879e+00 2.32102975e-01
-5.73604345e-01 8.92514944e-01 -1.70559436e-01 3.37714136e-01
6.87569082e-01 -1.60671198e+00 9.30091739e-01 1.06648707e+00
2.49521896e-01 6.67625904e-01 4.05357182e-01 -3.80777866e-01
-7.78873980e-01 -9.59275842e-01 1.26807714e+00 -9.14990544e-01
7.35974908e-01 -4.52580094e-01 -6.80027425e-01 1.33576941e+00
1.93114236e-01 -1.59112975e-01 8.43998432e-01 5.28782725e-01
-5.48214555e-01 2.29012713e-01 -9.08892810e-01 5.21641612e-01
1.52140009e+00 -5.51502466e-01 -7.60771453e-01 2.05998272e-01
1.10773492e+00 -3.86268765e-01 -5.95114052e-01 5.59627414e-01
4.46622163e-01 -9.21258390e-01 9.93567765e-01 -7.98801839e-01
5.24300218e-01 2.64688972e-02 -3.95232230e-01 -1.35095727e+00
-4.60711807e-01 -4.19364184e-01 -1.13909272e-02 1.04161918e+00
7.47605443e-01 -4.92436767e-01 7.90866911e-01 6.12156332e-01
-6.15445971e-01 -6.13614678e-01 -9.62632596e-01 -5.32994509e-01
5.38904071e-01 -9.05339956e-01 1.75047979e-01 9.34194207e-01
-3.57484996e-01 8.04668427e-01 -2.12722093e-01 3.10160201e-02
4.98246014e-01 -3.33522230e-01 9.00918618e-02 -1.24732578e+00
-3.48615974e-01 -3.95842195e-01 2.53827404e-02 -1.13759899e+00
9.64128375e-01 -1.17107522e+00 -1.25210002e-01 -1.84898686e+00
9.68576595e-02 -5.67708015e-01 -6.49127901e-01 1.06829214e+00
-2.46095046e-01 3.29190433e-01 2.03676283e-01 5.57246804e-02
-7.61854827e-01 3.06330949e-01 8.41629982e-01 -7.54516199e-02
1.47771776e-01 -2.13991612e-01 -1.39668798e+00 1.12085843e+00
8.94573092e-01 -5.11272252e-01 -4.58245218e-01 -1.25115538e+00
2.60578573e-01 -2.63094038e-01 1.95674390e-01 -8.77188861e-01
-1.49282232e-01 1.38985589e-01 3.13857287e-01 -2.25281075e-01
3.44403148e-01 -2.36073077e-01 -6.50342643e-01 3.29954922e-01
-7.59212017e-01 1.50628939e-01 5.48107266e-01 4.18509752e-01
-1.84343368e-01 -3.70935142e-01 7.61418045e-01 -2.76310354e-01
-1.01755226e+00 -2.10916281e-01 -4.58957970e-01 4.46605295e-01
3.45002621e-01 1.47101387e-01 -6.87959909e-01 -3.58674556e-01
-8.24047506e-01 3.94795895e-01 8.27358738e-02 4.75136101e-01
1.86475351e-01 -9.61209595e-01 -8.28066885e-01 -1.20457284e-01
-3.20738964e-02 1.77514583e-01 -3.66605595e-02 3.95471364e-01
-4.09115285e-01 3.88320982e-01 1.85723215e-01 -2.51547605e-01
-1.26937628e+00 -2.82365352e-01 4.02385890e-01 -7.00655878e-01
-3.93857628e-01 1.58807039e+00 1.44252241e-01 -5.85582852e-01
2.73772269e-01 -2.49226764e-01 -4.09070849e-01 9.55835171e-03
4.13834304e-01 2.30633765e-01 2.63208717e-01 -4.76574153e-01
-5.94443500e-01 1.46379992e-01 -2.05621839e-01 -2.03109801e-01
1.46554935e+00 1.49395779e-01 -2.16118440e-01 5.96861005e-01
9.88427103e-01 -9.04859323e-03 -1.22300434e+00 -3.28091346e-02
6.47315264e-01 9.06490907e-02 -1.51587036e-02 -1.25505126e+00
-6.43052578e-01 1.01540565e+00 2.83980310e-01 -1.28003567e-01
8.35839868e-01 2.06158489e-01 8.01032126e-01 5.12327433e-01
1.36169583e-01 -1.00339651e+00 1.87900923e-02 1.25647604e+00
5.55049896e-01 -1.07403588e+00 -1.95184678e-01 -3.54256272e-01
-5.64462185e-01 7.80900896e-01 6.91496313e-01 -2.53357142e-01
4.70665753e-01 5.72494388e-01 1.37009054e-01 -2.27460325e-01
-9.96632278e-01 -4.34602141e-01 -1.05397224e-01 5.27763724e-01
9.84881938e-01 1.00143872e-01 -3.56804430e-01 1.00171566e+00
-3.49365085e-01 -3.43739331e-01 3.04352343e-01 8.22306633e-01
-5.19860923e-01 -1.30900204e+00 1.50680780e-01 1.90693945e-01
-1.06234062e+00 -4.82219011e-01 -4.55706269e-01 9.54890132e-01
7.82093257e-02 1.22238719e+00 2.46081859e-01 -3.15527171e-01
4.23407346e-01 4.39081728e-01 5.41655958e-01 -1.07998443e+00
-7.21678495e-01 -1.08394600e-01 8.62184763e-01 -6.78952634e-01
-3.89785141e-01 -4.88629758e-01 -1.60134041e+00 4.06009667e-02
6.45955503e-02 -7.27896243e-02 4.45266128e-01 1.09273612e+00
2.61845559e-01 8.16071987e-01 7.53196329e-02 -5.69875777e-01
-6.44835830e-01 -1.13581967e+00 -1.85984775e-01 2.73211062e-01
2.91337460e-01 -6.29254520e-01 -3.68346572e-01 -7.08876997e-02] | [10.439498901367188, 9.400346755981445] |
caccb78d-cd2c-4b3f-bb7c-0f74853e6fe0 | continual-machine-reading-comprehension-via-1 | 2208.05217 | null | https://arxiv.org/abs/2208.05217v1 | https://arxiv.org/pdf/2208.05217v1.pdf | Continual Machine Reading Comprehension via Uncertainty-aware Fixed Memory and Adversarial Domain Adaptation | Continual Machine Reading Comprehension aims to incrementally learn from a continuous data stream across time without access the previous seen data, which is crucial for the development of real-world MRC systems. However, it is a great challenge to learn a new domain incrementally without catastrophically forgetting previous knowledge. In this paper, MA-MRC, a continual MRC model with uncertainty-aware fixed Memory and Adversarial domain adaptation, is proposed. In MA-MRC, a fixed size memory stores a small number of samples in previous domain data along with an uncertainty-aware updating strategy when new domain data arrives. For incremental learning, MA-MRC not only keeps a stable understanding by learning both memory and new domain data, but also makes full use of the domain adaptation relationship between them by adversarial learning strategy. The experimental results show that MA-MRC is superior to strong baselines and has a substantial incremental learning ability without catastrophically forgetting under two different continual MRC settings. | ['Kai Gao', 'Jingliang Fang', 'Hua Xu', 'Zhijing Wu'] | 2022-08-10 | continual-machine-reading-comprehension-via | https://aclanthology.org/2022.findings-naacl.179 | https://aclanthology.org/2022.findings-naacl.179.pdf | findings-naacl-2022-7 | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 2.88111448e-01 5.98342083e-02 -2.38741934e-01 -4.60055113e-01
-7.66332448e-01 -6.28235877e-01 3.67039740e-01 3.93151075e-01
-6.33658290e-01 1.16734028e+00 3.50661308e-01 -3.25506955e-01
-6.26225909e-03 -1.00597215e+00 -1.18422163e+00 -3.50975811e-01
6.23637661e-02 5.59417486e-01 6.01884723e-01 -3.47884268e-01
6.15980141e-02 3.18300053e-02 -1.57521737e+00 2.82016516e-01
1.12808621e+00 9.48868871e-01 5.18867731e-01 8.28514695e-01
-1.16937131e-01 1.16122591e+00 -5.42453825e-01 -1.39795735e-01
-9.51323099e-03 -2.48233661e-01 -8.34432960e-01 -2.91317225e-01
2.36412093e-01 -8.57689738e-01 -7.61243403e-01 5.51424980e-01
5.40758014e-01 6.56225383e-01 4.60561484e-01 -1.02284205e+00
-1.07369173e+00 8.94871294e-01 -2.60853231e-01 5.41645765e-01
6.33386016e-01 4.88561034e-01 4.03370380e-01 -6.90354407e-01
3.31029624e-01 1.24860823e+00 6.22164130e-01 7.73235321e-01
-8.96422923e-01 -6.63742244e-01 7.44225860e-01 5.99110067e-01
-1.06921804e+00 -5.02514362e-01 5.93172789e-01 -7.43132904e-02
1.04090583e+00 -3.36713493e-01 3.23753387e-01 1.52859282e+00
3.59960049e-01 1.03917563e+00 1.09523165e+00 -4.36294347e-01
6.38829291e-01 -1.29198059e-01 3.32872093e-01 2.75311500e-01
-2.43423544e-02 4.41133916e-01 -8.35679173e-01 2.48431414e-02
4.32022393e-01 7.22033083e-02 -3.40708047e-01 -2.80841172e-01
-1.13901424e+00 3.29274237e-01 3.07323307e-01 -5.76283894e-02
-2.73279577e-01 5.85522540e-02 3.84754479e-01 8.97205770e-01
1.34121090e-01 2.16466978e-01 -8.60151172e-01 -4.40603852e-01
-5.82898915e-01 4.04404923e-02 5.04104078e-01 1.28977597e+00
4.51681405e-01 2.50164449e-01 -2.77166635e-01 6.08028889e-01
-1.39496416e-01 9.35912788e-01 9.51400578e-01 -7.69587159e-01
6.90558314e-01 4.17226493e-01 2.26968870e-01 -4.10107970e-01
-3.60345036e-01 -4.27616656e-01 -9.94942009e-01 -2.46719904e-02
2.16393605e-01 -2.70441562e-01 -1.08637846e+00 2.02203131e+00
2.57295609e-01 7.51236737e-01 4.96182412e-01 2.75398344e-01
5.64126611e-01 5.52476168e-01 1.27345413e-01 -5.05266786e-01
7.51639843e-01 -7.18224168e-01 -7.58001804e-01 -4.08239067e-01
-1.25194574e-03 -6.42886311e-02 1.42641592e+00 5.98745942e-01
-1.23347485e+00 -9.03171837e-01 -1.51319695e+00 -2.33386219e-01
-2.85092920e-01 -6.79079533e-01 1.53184175e-01 2.71017224e-01
-6.69709682e-01 5.12674928e-01 -1.02940381e+00 2.49979615e-01
4.38381940e-01 1.54737890e-01 -1.40035763e-01 -5.46604097e-01
-1.75380516e+00 9.04643774e-01 8.18208992e-01 -3.90656233e-01
-1.14479268e+00 -9.60493922e-01 -9.64043140e-01 -5.82228377e-02
6.27231181e-01 -8.48904967e-01 1.79493117e+00 -8.15654516e-01
-1.76061547e+00 1.06994145e-01 -3.57796967e-01 -8.40168417e-01
7.09455013e-01 -7.69317329e-01 -6.92305982e-01 -5.32546043e-02
-2.58733243e-01 3.66884440e-01 1.27820456e+00 -1.06149292e+00
-7.76865602e-01 -2.39756376e-01 3.95317599e-02 5.03035247e-01
-2.28614375e-01 -8.40106487e-01 -1.16980642e-01 -6.69703245e-01
-4.47648801e-02 -4.44707900e-01 2.09270746e-01 -4.26850975e-01
3.57343763e-01 -1.77223518e-01 7.24555194e-01 -7.38312364e-01
1.33692837e+00 -1.97444153e+00 3.11884284e-01 -1.99507281e-01
1.23730764e-01 3.35545689e-01 -2.19093308e-01 1.37878776e-01
2.20634699e-01 -2.66873509e-01 -3.93917352e-01 -9.98160020e-02
-1.67835891e-01 3.19019794e-01 -7.32113361e-01 -6.02169521e-02
1.04371399e-01 1.10488772e+00 -1.52072859e+00 -2.22789515e-02
1.07578821e-01 5.83859086e-02 -2.28531718e-01 3.69535238e-01
-8.14573109e-01 3.73810589e-01 -7.13152960e-02 4.49410796e-01
8.15640807e-01 -1.86906666e-01 8.12803134e-02 3.88512671e-01
4.15615618e-01 3.68862264e-02 -1.19050157e+00 2.04564643e+00
-5.39584935e-01 2.67352343e-01 -4.35367614e-01 -7.69302070e-01
8.14306855e-01 2.41979420e-01 -1.87571645e-01 -1.16439271e+00
-1.30525514e-01 -4.46041711e-02 -1.73062086e-01 -2.20377401e-01
6.42159343e-01 -3.20536017e-01 -2.82899350e-01 6.22706652e-01
1.23346113e-01 -1.32568516e-02 -2.99212098e-01 3.29475641e-01
1.39752913e+00 5.49952351e-02 5.41848958e-01 4.10231143e-01
5.65319002e-01 -5.61166480e-02 7.59334147e-01 1.09842515e+00
-3.91691625e-01 5.20448744e-01 -8.80896971e-02 -2.10946098e-01
-8.31167877e-01 -1.59871638e+00 2.41022304e-01 1.39296472e+00
3.10316980e-01 1.82559416e-01 -3.26967984e-01 -8.76823723e-01
3.29972088e-01 1.27468753e+00 -5.46128929e-01 -9.76426125e-01
-6.94252610e-01 -4.16753292e-01 4.11298394e-01 1.01291156e+00
9.81353998e-01 -1.02383363e+00 -5.34337044e-01 6.13981903e-01
-3.44720334e-01 -7.13826478e-01 -6.23481333e-01 3.79652143e-01
-9.81936693e-01 -9.05361414e-01 -4.58764046e-01 -6.31424427e-01
2.81476736e-01 3.79006445e-01 1.32405996e+00 -3.14838260e-01
1.74643025e-01 7.42511332e-01 -4.96111035e-01 -7.04750121e-01
-7.12212205e-01 2.85482407e-01 5.01710713e-01 -3.67462635e-01
6.26798153e-01 -6.44487977e-01 -3.75836790e-01 3.03274810e-01
-9.65308368e-01 -1.58116162e-01 6.66405976e-01 1.11827290e+00
6.95118368e-01 3.84447366e-01 1.40538442e+00 -8.71975422e-01
6.89303160e-01 -8.82653832e-01 -3.68993968e-01 5.77305436e-01
-7.65732288e-01 2.06860989e-01 9.32993054e-01 -8.47271025e-01
-1.56844366e+00 -2.05061778e-01 5.90374582e-02 -3.91865492e-01
-9.31228697e-02 4.29848254e-01 -5.29648900e-01 6.08007133e-01
8.89094770e-01 6.98143601e-01 -3.84992324e-02 -3.20930541e-01
8.29956055e-01 7.58012891e-01 1.20942676e+00 -8.70678008e-01
7.89212108e-01 1.05465852e-01 -6.21417046e-01 -3.51422846e-01
-9.87710595e-01 -3.58661264e-01 -9.86128271e-01 -1.84616894e-02
3.42031240e-01 -1.27955234e+00 -2.35614076e-01 1.19561577e+00
-7.08894312e-01 -7.33683944e-01 -7.47224450e-01 2.04826117e-01
-7.92082310e-01 3.31327915e-01 -4.49195772e-01 -4.18788016e-01
-3.03547531e-01 -4.79947656e-01 4.55062389e-01 4.02304053e-01
-5.07479012e-02 -1.02679670e+00 5.85276447e-02 3.22262466e-01
6.19393170e-01 -1.12035638e-02 1.03898156e+00 -4.84935880e-01
-5.90874553e-01 -2.62222942e-02 1.68641686e-01 6.89740717e-01
3.01091522e-01 -7.47436702e-01 -9.30428505e-01 -6.95569873e-01
1.53832585e-01 -7.97080457e-01 9.65519667e-01 -6.37415797e-02
1.17267597e+00 -7.46592283e-02 -7.64067024e-02 2.75555462e-01
1.24469948e+00 4.74625349e-01 5.51730335e-01 3.32761019e-01
3.44675690e-01 -4.57221828e-02 1.00938201e+00 5.16779840e-01
7.70450532e-01 2.43361387e-03 3.01859170e-01 6.57613873e-01
-3.25797468e-01 -8.04707885e-01 6.42106473e-01 1.04810190e+00
3.32486331e-01 -2.74972945e-01 -9.61442649e-01 7.07333624e-01
-1.95509684e+00 -1.08609021e+00 6.53051972e-01 2.47581267e+00
1.34121513e+00 4.62403476e-01 -2.04614684e-01 1.23730309e-01
5.28800666e-01 -9.51486528e-02 -1.70018959e+00 -1.68838039e-01
-2.67222166e-01 3.28110218e-01 2.50230134e-01 5.19071162e-01
-9.64916289e-01 8.21659684e-01 6.59663439e+00 6.19010627e-01
-9.97740090e-01 1.71789736e-01 3.47425967e-01 -2.50614017e-01
-1.47908807e-01 -4.00389701e-01 -8.50534916e-01 5.85234344e-01
1.24468076e+00 -5.62206209e-01 5.21646798e-01 8.87854457e-01
-3.29311758e-01 -2.75627047e-01 -1.34700966e+00 6.29131377e-01
3.76158766e-02 -1.11254060e+00 1.66253746e-01 -5.80713749e-01
8.93076837e-01 7.06033930e-02 4.05501276e-01 9.99432027e-01
8.41599762e-01 -9.59389031e-01 6.10851705e-01 9.27904844e-01
9.65422511e-01 -9.40195680e-01 6.27976596e-01 9.71899867e-01
-9.23448503e-01 -5.99105418e-01 -5.28616250e-01 -2.43819922e-01
7.30800033e-02 2.59826034e-01 -7.37305284e-01 5.81318915e-01
5.53750277e-01 6.75550640e-01 -5.80736578e-01 7.42586851e-01
-3.97928655e-01 8.24636936e-01 -1.26409709e-01 3.39520812e-01
-2.85107583e-01 5.25042415e-01 4.24219161e-01 7.35726357e-01
2.79217809e-01 4.42185313e-01 2.27685302e-01 2.93605894e-01
-3.25804174e-01 -3.97176325e-01 -4.17244822e-01 3.85601848e-01
1.16787076e+00 2.77189344e-01 1.70428872e-01 -5.19579351e-01
-3.87027979e-01 1.12262475e+00 7.35481083e-01 2.58948177e-01
-7.63286054e-01 -3.08728456e-01 3.79431009e-01 6.56504929e-02
4.05601263e-01 -3.54757160e-01 -1.03302121e-01 -1.26818383e+00
1.44317731e-01 -1.08226025e+00 7.08810329e-01 -9.01118398e-01
-1.57899570e+00 1.80200577e-01 1.28629103e-01 -1.16463733e+00
-3.09517294e-01 -9.05040801e-02 -5.76328218e-01 7.19280183e-01
-1.84119785e+00 -9.20896709e-01 -4.96363401e-01 9.68082964e-01
8.74821723e-01 -4.38867301e-01 7.21854210e-01 -1.87098399e-01
-3.03377301e-01 1.08638251e+00 6.33583128e-01 -1.74741894e-01
9.75154936e-01 -1.52509856e+00 6.06727421e-01 8.37408483e-01
-3.33308071e-01 6.93525016e-01 4.37762976e-01 -9.12747920e-01
-1.45295346e+00 -1.38130474e+00 3.59256506e-01 -7.83918679e-01
5.77192247e-01 -2.35200480e-01 -1.62578678e+00 9.04710114e-01
1.63030311e-01 -1.88083917e-01 6.02352023e-01 2.75469068e-02
-5.99100828e-01 -3.60982865e-01 -1.27893400e+00 3.48220587e-01
1.20911598e+00 -5.24859071e-01 -1.48487759e+00 -2.84545198e-02
1.41494858e+00 -7.17734694e-01 -9.66367126e-01 5.19469440e-01
3.41569602e-01 -5.65212846e-01 9.40387309e-01 -6.76827252e-01
1.58896968e-01 -2.49773547e-01 -6.16938733e-02 -1.75814760e+00
-3.45384061e-01 -4.21846896e-01 -9.86666322e-01 1.13731217e+00
1.33485258e-01 -6.84483349e-01 5.16329288e-01 6.13277733e-01
-3.37069738e-03 -3.35854203e-01 -9.28796589e-01 -1.16176820e+00
7.03365803e-01 -5.12950480e-01 1.03418529e+00 6.52192354e-01
-4.76587564e-02 4.84674484e-01 -3.96243334e-01 3.57739031e-01
5.74127793e-01 -1.95691865e-02 8.04280877e-01 -1.09926820e+00
-3.95348698e-01 1.69246688e-01 5.16379997e-02 -1.46617532e+00
1.66117653e-01 -5.50286412e-01 8.38873535e-02 -1.19449556e+00
7.19098374e-02 -4.29109365e-01 -9.11994755e-01 4.09464091e-01
-5.75946748e-01 -7.52777278e-01 5.83213940e-03 9.82517451e-02
-1.01357031e+00 8.24182868e-01 1.18940473e+00 -2.81970829e-01
-5.04568040e-01 8.69471952e-02 -9.05008256e-01 5.83364427e-01
8.20909798e-01 -2.05052227e-01 -1.19250047e+00 -5.05225241e-01
6.51427507e-02 2.19547138e-01 5.32978512e-02 -1.32271433e+00
6.32513940e-01 -2.76825815e-01 6.85049176e-01 -8.71997118e-01
1.86128438e-01 -7.23845005e-01 -9.65968966e-02 2.57749826e-01
-4.55455691e-01 -4.55515049e-02 4.44436759e-01 1.38361681e+00
-9.73237082e-02 -5.76495342e-02 8.18401396e-01 -3.71143937e-01
-1.37197256e+00 1.83845967e-01 -2.27127194e-01 5.70234001e-01
1.18783498e+00 -1.12295346e-02 -6.32213771e-01 -4.29411441e-01
-1.07369387e+00 8.96802962e-01 4.09391105e-01 9.02947485e-01
9.73348498e-01 -1.15413606e+00 -5.83208919e-01 2.89061189e-01
2.07492575e-01 6.64947212e-01 6.85282052e-01 -1.31995574e-01
9.90811139e-02 1.09184042e-01 -2.62248367e-01 -5.04820943e-01
-9.47220087e-01 1.06120360e+00 1.24068864e-01 -3.34559292e-01
-5.09863198e-01 7.32948899e-01 -3.24971646e-01 -6.13275945e-01
5.89094579e-01 -4.08352315e-01 -1.49221271e-01 -9.17247832e-02
1.05047357e+00 5.92728078e-01 2.30888203e-02 4.56835330e-02
6.15004897e-02 1.63979128e-01 -6.21877849e-01 -9.71421227e-02
9.86428320e-01 -6.67838871e-01 3.96692008e-01 1.00365770e+00
5.50052822e-01 -4.08303440e-01 -1.73970997e+00 -9.82102036e-01
-2.05009766e-02 -3.34722757e-01 -8.50556791e-02 -1.54801428e+00
-3.62675697e-01 8.50617468e-01 5.76541662e-01 -5.51534235e-01
1.32749832e+00 -4.36588138e-01 1.18085468e+00 7.99971461e-01
8.86814296e-01 -1.39357495e+00 5.44914186e-01 8.85025501e-01
9.27461445e-01 -1.44291365e+00 -3.14316936e-02 1.64930612e-01
-7.69137323e-01 7.75389910e-01 1.15112782e+00 2.58287400e-01
6.93485796e-01 1.39620602e-01 -2.30954075e-03 5.87049186e-01
-1.27813137e+00 5.12542343e-03 -3.17647196e-02 1.14521015e+00
-3.92177850e-01 4.20245230e-02 3.25424582e-01 1.12047589e+00
-1.46829858e-01 3.83326679e-01 6.03496313e-01 1.28256321e+00
-8.15528095e-01 -9.47308779e-01 -2.29708970e-01 4.24829990e-01
1.42730445e-01 4.42786328e-02 -1.05450219e-02 6.20702267e-01
1.39491260e-01 1.00218666e+00 -3.58188748e-02 -3.97162288e-01
5.39300621e-01 5.45930266e-01 5.54751575e-01 -6.41477227e-01
-2.63441235e-01 -6.59300148e-01 -3.74411374e-01 -2.40941897e-01
4.81905639e-02 -7.93096125e-01 -1.59505153e+00 -3.88834268e-01
-1.37859449e-01 -4.51229848e-02 1.27988264e-01 8.45029891e-01
5.56041479e-01 6.87590301e-01 5.44208944e-01 5.69175892e-02
-1.19101477e+00 -1.01109660e+00 -3.27762753e-01 2.18811542e-01
6.75797701e-01 -7.16281354e-01 -2.20751628e-01 2.35356063e-01] | [9.895796775817871, 3.451050281524658] |
162dcfe7-f187-45b8-b0b0-daa0b8136ae8 | country-level-arabic-dialect-identification | null | null | https://aclanthology.org/2021.wanlp-1.30 | https://aclanthology.org/2021.wanlp-1.30.pdf | Country-level Arabic Dialect Identification Using Small Datasets with Integrated Machine Learning Techniques and Deep Learning Models | Arabic is characterised by a considerable number of varieties including spoken dialects. In this paper, we presented our models developed to participate in the NADI subtask 1.2 that requires building a system to distinguish between 21 country-level dialects. We investigated several classical machine learning approaches and deep learning models using small datasets. We examined an integration technique between two machine learning approaches. Additionally, we created dictionaries automatically based on Pointwise Mutual Information and labelled datasets, which enriched the feature space when training models. A semi-supervised learning approach was also examined and compared to other methods that exploit large unlabelled datasets, such as building pre-trained word embeddings. Our winning model was the Support Vector Machine with dictionary-based features and Pointwise Mutual Information values, achieving an 18.94% macros-average F1-score. | ['Maha J. Althobaiti'] | null | null | null | null | eacl-wanlp-2021-4 | ['dialect-identification'] | ['natural-language-processing'] | [-3.58978510e-01 3.11938912e-01 -1.90062504e-02 -7.39529550e-01
-6.62497401e-01 -8.28512669e-01 1.08174324e+00 2.92741090e-01
-7.02902377e-01 8.46564710e-01 2.73762643e-01 -2.61784226e-01
-3.01915795e-01 -9.74569678e-01 -2.42358759e-01 -7.31144965e-01
-3.04419845e-01 9.44807053e-01 -2.99888283e-01 -8.20373178e-01
3.95085961e-01 5.35337567e-01 -1.65179622e+00 1.14179067e-01
1.23173761e+00 6.40128076e-01 3.99809293e-02 2.73735762e-01
-2.08832294e-01 6.74638748e-01 -5.50166309e-01 -6.32031858e-01
1.82706267e-01 -3.83771122e-01 -8.71865392e-01 -1.72923923e-01
2.52381653e-01 -2.14412466e-01 1.71460688e-01 8.06184471e-01
6.11030757e-01 2.51278281e-03 1.03150153e+00 -9.12476122e-01
-7.91065395e-01 1.06014252e+00 -3.64885777e-02 -2.53231138e-01
5.44294655e-01 -1.30947769e-01 1.18874323e+00 -1.26340735e+00
6.19641602e-01 9.31724191e-01 8.81388187e-01 2.68355876e-01
-1.20918477e+00 -3.66928399e-01 -3.45482826e-01 4.06903595e-01
-1.50250554e+00 -3.80207419e-01 6.63790822e-01 -5.35273969e-01
1.19423580e+00 1.58035487e-01 6.36005342e-01 7.51099169e-01
-1.28501162e-01 6.91345870e-01 1.64321971e+00 -1.11414444e+00
6.55643120e-02 7.17266917e-01 2.58839816e-01 7.01945543e-01
-4.52713408e-02 -5.46886697e-02 -3.39990765e-01 -2.87372619e-02
5.80705345e-01 -6.05066717e-01 -3.69115062e-02 -9.74273682e-02
-1.18475306e+00 1.31282842e+00 4.63997684e-02 8.32873404e-01
-3.02163005e-01 -8.55173230e-01 2.92952359e-01 7.29257286e-01
3.27585399e-01 6.06967568e-01 -8.18132401e-01 -1.21393837e-01
-7.55460560e-01 3.32373381e-02 1.11982751e+00 6.78241253e-01
9.98303652e-01 1.97951585e-01 3.28517735e-01 1.40489364e+00
5.64978719e-01 3.53015840e-01 8.78039241e-01 -3.78673315e-01
9.80273634e-02 8.40377271e-01 -1.94868445e-01 -8.73063743e-01
-5.18240333e-01 -1.90623358e-01 -7.15261936e-01 1.18402369e-01
7.41624951e-01 -5.58632553e-01 -6.96472764e-01 1.49562967e+00
2.48598278e-01 -5.31404793e-01 3.02520543e-01 8.37486744e-01
7.57778943e-01 6.34240031e-01 -4.35612261e-01 -9.44692269e-02
1.09615147e+00 -8.01311135e-01 -7.53796577e-01 3.15266848e-01
1.03814673e+00 -1.12077963e+00 1.09948146e+00 8.21519732e-01
-8.24429989e-01 -5.30100524e-01 -1.08645511e+00 1.54557824e-01
-1.00696015e+00 2.54563749e-01 4.84409779e-01 1.20713639e+00
-1.32475936e+00 5.67293167e-01 -1.61271304e-01 -3.78789395e-01
-1.91212013e-01 5.33023000e-01 -7.19970226e-01 7.85632655e-02
-1.61309755e+00 1.34797084e+00 7.41873384e-01 7.61453435e-02
-5.64709008e-01 -1.97213724e-01 -1.07396972e+00 -3.65201384e-01
-2.16256067e-01 3.03568691e-01 6.90482616e-01 -1.34219265e+00
-1.96020269e+00 1.30453718e+00 3.81739944e-01 -3.49114776e-01
2.90889800e-01 -1.50960788e-01 -6.74347520e-01 -1.82955101e-01
-1.26638666e-01 3.42829138e-01 5.14791310e-01 -1.26721096e+00
-5.26163638e-01 -4.38560754e-01 8.46643299e-02 2.50612110e-01
-5.13901711e-01 1.02044709e-01 2.06053853e-01 -6.85301483e-01
1.15105081e-02 -6.90425575e-01 -8.40563178e-02 -8.61587167e-01
-6.80675060e-02 -2.62273520e-01 2.82542855e-01 -1.09128404e+00
1.17504489e+00 -1.71503890e+00 1.30981147e-01 6.67206049e-01
-6.57701343e-02 5.18719733e-01 -2.58244962e-01 7.81682849e-01
-7.29425102e-02 -1.88389480e-01 -2.79906958e-01 6.18040562e-02
3.29331547e-01 3.09812129e-01 1.92034803e-02 4.89346832e-01
5.22048473e-01 3.66878092e-01 -9.06113207e-01 -4.01618868e-01
2.20641777e-01 4.48332191e-01 -2.52234519e-01 2.31336460e-01
-3.69519778e-02 1.47973999e-01 3.55191082e-01 7.73426771e-01
7.24344552e-01 6.00825310e-01 6.39806867e-01 -3.66736874e-02
-3.96831840e-01 3.58946264e-01 -1.27063918e+00 1.52196634e+00
-6.45028234e-01 6.08464658e-01 -1.06188096e-01 -1.19790888e+00
1.53368974e+00 3.01199675e-01 2.57033110e-01 -5.94265044e-01
4.20520157e-01 6.74917340e-01 1.89665720e-01 -4.24656361e-01
5.54502368e-01 -1.76000595e-01 -1.08701743e-01 4.58938658e-01
7.78997123e-01 -8.10721964e-02 3.13647956e-01 -2.47571677e-01
3.90536994e-01 3.74075830e-01 5.65192282e-01 -8.20989132e-01
8.50285292e-01 2.02120870e-01 2.52710432e-01 4.10951555e-01
-1.09481104e-02 5.45560002e-01 4.46979702e-01 -6.38675809e-01
-8.58319342e-01 -8.42731595e-01 -7.07167625e-01 1.29424071e+00
-4.05528665e-01 -4.57519293e-01 -8.17921579e-01 -8.33306789e-01
-3.29107434e-01 7.63279080e-01 -7.19651043e-01 7.60040358e-02
-5.00857472e-01 -1.09571779e+00 9.40685272e-01 4.71832454e-02
1.80031210e-01 -8.92309070e-01 -5.95252514e-02 3.28807324e-01
2.93156356e-02 -4.44484979e-01 8.24762657e-02 4.82844770e-01
-4.52505827e-01 -1.05141103e+00 -7.46452212e-01 -1.40231907e+00
3.36156160e-01 -5.20849049e-01 1.42279828e+00 -1.68080598e-01
7.84335956e-02 5.08844741e-02 -9.15247262e-01 -5.41401625e-01
-6.49857998e-01 3.25906813e-01 2.53350735e-01 1.81737468e-01
8.51293564e-01 -2.74432003e-01 4.02719900e-02 3.19237947e-01
-8.31555486e-01 -4.63944256e-01 5.51331639e-01 1.11287749e+00
2.13545203e-01 -2.68981218e-01 8.68968248e-01 -8.69620562e-01
6.09925926e-01 -5.07430851e-01 -5.02886176e-01 3.48439485e-01
-5.29662430e-01 2.48395890e-01 4.77912366e-01 -3.59127790e-01
-9.34075296e-01 2.19290301e-01 -8.56717944e-01 5.95718026e-01
-4.53537673e-01 8.05881321e-01 -5.38462400e-01 -1.01536550e-01
8.17641854e-01 3.84453595e-01 2.44241804e-01 -4.67958093e-01
5.98299921e-01 1.35336864e+00 7.95199722e-02 -7.76071370e-01
3.64528596e-01 -2.29587466e-01 -5.29329538e-01 -1.14420342e+00
-2.42503688e-01 -3.40874702e-01 -1.40274286e+00 -2.68773139e-01
5.90193510e-01 -8.57714355e-01 -1.13236152e-01 9.02486324e-01
-9.91968632e-01 -5.24126947e-01 -1.02179617e-01 6.42647028e-01
-5.46421528e-01 2.70753771e-01 -5.06359398e-01 -6.86766684e-01
-1.88783377e-01 -9.03186977e-01 7.42326140e-01 -5.46348654e-02
-3.63895267e-01 -1.44472885e+00 6.09213233e-01 -9.37426463e-02
3.68800789e-01 2.53891081e-01 9.96387362e-01 -1.40733886e+00
3.15857887e-01 -1.21918879e-01 9.44705456e-02 6.38326943e-01
3.40359628e-01 7.29366019e-02 -1.18813992e+00 -1.04464553e-01
-6.20708875e-02 -6.03055000e-01 4.89945829e-01 -3.05905119e-02
2.47419834e-01 -2.76145041e-01 4.77510691e-01 2.63036937e-01
1.57086217e+00 3.92135352e-01 6.70579076e-01 6.65904999e-01
4.43404377e-01 8.98705602e-01 6.76864982e-01 3.97626370e-01
5.40544808e-01 4.97480661e-01 8.91375393e-02 1.43256644e-02
3.10029656e-01 2.36407727e-01 5.16733646e-01 1.44691193e+00
-2.71387011e-01 2.60391742e-01 -1.30655837e+00 8.25391769e-01
-1.36356354e+00 -6.70474708e-01 -3.44596148e-01 2.14978647e+00
1.30558121e+00 1.04151596e-03 3.06178331e-01 5.68979383e-01
4.13118511e-01 -1.05470292e-01 1.57629132e-01 -8.62690866e-01
-6.10966027e-01 6.60999954e-01 1.95249423e-01 7.95303226e-01
-1.16710591e+00 1.20821357e+00 6.34356976e+00 8.28507483e-01
-1.13218379e+00 1.11985795e-01 2.94167161e-01 4.16198015e-01
-1.61740094e-01 -3.75975817e-01 -8.22300315e-01 2.05137655e-01
1.25632501e+00 9.48410928e-02 2.16533437e-01 6.27288520e-01
-2.44181231e-01 -8.73372406e-02 -9.21336055e-01 6.21509373e-01
5.94655335e-01 -9.62180614e-01 1.63216561e-01 1.35461375e-01
7.99679697e-01 6.23770431e-02 -1.65151209e-01 4.90398496e-01
5.60519874e-01 -1.10425699e+00 6.72457039e-01 4.15182829e-01
8.12956572e-01 -9.37341809e-01 1.23501968e+00 3.62147599e-01
-6.99597120e-01 6.68772832e-02 -2.64782727e-01 -3.65984231e-01
-2.04228669e-01 5.98661423e-01 -8.86217952e-01 7.37343371e-01
4.19647038e-01 7.20182121e-01 -5.82665682e-01 7.61567414e-01
-2.34508112e-01 8.11023176e-01 -3.96620899e-01 -3.44589591e-01
5.52473545e-01 -6.25202000e-01 1.26235887e-01 1.55858719e+00
3.03147227e-01 -4.05178517e-01 -1.85387179e-01 1.98234171e-01
3.54316652e-01 7.45612621e-01 -7.61921823e-01 1.91891372e-01
3.00926477e-01 1.08118272e+00 -5.68877280e-01 -3.00859034e-01
-4.62470472e-01 8.56585205e-01 1.89143777e-01 -1.08791329e-01
-6.27935708e-01 -7.32065439e-01 5.73689342e-01 -1.16527036e-01
1.92679450e-01 -3.76144141e-01 -3.15487355e-01 -1.21622050e+00
-2.89902627e-01 -1.21985233e+00 3.29073548e-01 -1.57310233e-01
-1.48475134e+00 1.17864096e+00 -1.06423967e-01 -9.84392285e-01
-6.00559235e-01 -1.18738496e+00 -1.69206306e-01 1.00294971e+00
-1.37252700e+00 -1.40738809e+00 6.26820922e-02 6.86444938e-01
3.65988374e-01 -9.50662494e-01 1.53058839e+00 4.00470495e-01
-2.92765737e-01 6.27501249e-01 7.30422676e-01 5.29783666e-01
1.05532968e+00 -1.68247366e+00 -5.71948588e-02 5.00049829e-01
5.06476402e-01 5.47341347e-01 2.61703879e-01 -3.12792629e-01
-1.10932791e+00 -6.99874938e-01 1.53959513e+00 -4.92336631e-01
8.64324391e-01 -4.59415346e-01 -7.55068362e-01 6.17296875e-01
6.40355945e-01 -5.02416968e-01 1.28404844e+00 3.82047206e-01
-4.21518087e-01 -4.30356674e-02 -1.21984899e+00 1.88272759e-01
4.70393240e-01 -5.72220564e-01 -1.02632809e+00 3.17576081e-01
8.24715048e-02 -2.67780926e-02 -1.33111453e+00 1.44859061e-01
8.01275373e-01 -9.74157810e-01 7.13609219e-01 -6.28931344e-01
3.09350431e-01 1.25455797e-01 -5.51538944e-01 -1.82314157e+00
-1.41758636e-01 -3.17530453e-01 3.41670275e-01 1.56862390e+00
6.79327011e-01 -7.83656538e-01 3.26480418e-01 2.41384372e-01
5.36353812e-02 -4.42073554e-01 -7.04452336e-01 -7.59998739e-01
7.06641138e-01 -2.88885146e-01 7.53710449e-01 1.60652828e+00
1.98695228e-01 4.37548578e-01 -2.09495589e-01 -9.77617353e-02
1.47516504e-01 -1.34929582e-01 6.36583030e-01 -1.43341279e+00
2.07838248e-02 -6.52348816e-01 -5.95321119e-01 -3.85486752e-01
3.26994359e-01 -1.10196590e+00 -1.97650883e-02 -1.18926835e+00
-5.43009102e-01 -6.50519073e-01 -2.44163111e-01 5.22387326e-01
1.39238089e-01 4.10735995e-01 -8.58667716e-02 -1.82060704e-01
-4.48646285e-02 3.34331274e-01 5.98666310e-01 -1.58107981e-01
-4.18939948e-01 -2.22986147e-01 -5.46698093e-01 7.37243772e-01
1.21193671e+00 -2.15539053e-01 -9.48297158e-02 -4.57433283e-01
2.49384925e-01 -6.51587248e-01 -2.09983915e-01 -9.44636226e-01
-1.61589235e-01 -2.63048690e-02 1.99206576e-01 -3.17374885e-01
1.09431721e-01 -7.77010798e-01 -1.55322507e-01 4.23205614e-01
-1.07599653e-01 7.96625167e-02 1.99884892e-01 -3.45790714e-01
-5.42162418e-01 -6.11651003e-01 7.98008680e-01 -8.75676051e-02
-8.39666665e-01 -2.94571072e-01 -8.04828346e-01 -1.16886549e-01
9.72735167e-01 -1.70964822e-01 2.80512184e-01 -7.42081106e-02
-8.90698016e-01 -2.52342522e-01 3.05599213e-01 2.43435949e-01
3.31503510e-01 -1.38345504e+00 -1.14220667e+00 6.26924872e-01
2.36184344e-01 -5.07712424e-01 -2.09694088e-01 6.04110539e-01
-1.02871573e+00 3.36121172e-01 -9.35929596e-01 -2.31364340e-01
-1.13393831e+00 6.20434619e-02 2.49536201e-01 -2.76035070e-01
2.82903686e-02 6.64347172e-01 -6.48915529e-01 -1.47829401e+00
1.31232724e-01 -2.38461718e-02 -8.26903045e-01 8.48894298e-01
4.77754533e-01 2.92289913e-01 4.00647402e-01 -1.28692937e+00
-3.30865949e-01 4.87506688e-01 1.55837327e-01 -4.12291020e-01
1.68618667e+00 1.23527594e-01 -5.38848042e-01 5.75410187e-01
1.20943701e+00 3.03974509e-01 -5.82033813e-01 -3.23621690e-01
4.01492715e-01 -2.27090105e-01 -1.25646561e-01 -1.06869602e+00
-6.74383104e-01 8.84671152e-01 7.64601469e-01 4.62543845e-01
1.00402009e+00 -4.55157965e-01 3.43835652e-01 5.72816372e-01
5.25713742e-01 -1.34246290e+00 -4.41561937e-01 1.04806614e+00
7.58356035e-01 -1.50724161e+00 -3.35456073e-01 -1.37469172e-01
-7.04268873e-01 1.45435977e+00 3.47558618e-01 -2.07050875e-01
1.02046907e+00 3.30445915e-01 5.15449047e-01 5.29483892e-02
-2.12420374e-01 -5.98211765e-01 3.67052227e-01 1.11064661e+00
8.78514946e-01 3.05579096e-01 -8.23741734e-01 1.00861144e+00
-5.91826737e-01 -1.99063659e-01 4.11276132e-01 7.28474557e-01
-3.85648549e-01 -1.60909057e+00 -4.74728167e-01 3.51579875e-01
-2.39615709e-01 -3.33188176e-01 -7.56894767e-01 1.05604696e+00
5.51860929e-01 1.00806153e+00 4.39304225e-02 -7.77434409e-01
1.31152242e-01 4.13774043e-01 5.33161044e-01 -4.27938193e-01
-9.03973818e-01 -2.06732795e-01 4.08292055e-01 1.05440967e-01
-5.86938739e-01 -7.65314937e-01 -8.58655035e-01 -2.70921558e-01
-4.01829183e-01 2.51551390e-01 9.10536289e-01 1.04814076e+00
-2.20715553e-01 -1.68705821e-01 8.86349976e-01 -9.16480720e-01
-5.45161843e-01 -1.37736034e+00 -8.35271180e-01 2.40152285e-01
-4.99520935e-02 -4.76099730e-01 -1.28512770e-01 3.42196941e-01] | [10.280868530273438, 10.544598579406738] |
642a5663-bfae-4de7-bd6b-4d8036baa86a | spatial-deep-learning-for-wireless-scheduling | 1808.01486 | null | https://arxiv.org/abs/1808.01486v3 | https://arxiv.org/pdf/1808.01486v3.pdf | Spatial Deep Learning for Wireless Scheduling | The optimal scheduling of interfering links in a dense wireless network with full frequency reuse is a challenging task. The traditional method involves first estimating all the interfering channel strengths then optimizing the scheduling based on the model. This model-based method is however resource intensive and computationally hard because channel estimation is expensive in dense networks; furthermore, finding even a locally optimal solution of the resulting optimization problem may be computationally complex. This paper shows that by using a deep learning approach, it is possible to bypass the channel estimation and to schedule links efficiently based solely on the geographic locations of the transmitters and the receivers, due to the fact that in many propagation environments, the wireless channel strength is largely a function of the distance dependent path-loss. This is accomplished by unsupervised training over randomly deployed networks, and by using a novel neural network architecture that computes the geographic spatial convolutions of the interfering or interfered neighboring nodes along with subsequent multiple feedback stages to learn the optimum solution. The resulting neural network gives near-optimal performance for sum-rate maximization and is capable of generalizing to larger deployment areas and to deployments of different link densities. Moreover, to provide fairness, this paper proposes a novel scheduling approach that utilizes the sum-rate optimal scheduling algorithm over judiciously chosen subsets of links for maximizing a proportional fairness objective over the network. The proposed approach shows highly competitive and generalizable network utility maximization results. | ['Wei Yu', 'Kaiming Shen', 'Wei Cui'] | 2018-08-04 | null | null | null | null | ['few-shot-camera-adaptive-color-constancy', 'few-shot-camera-adaptive-color-constancy'] | ['computer-vision', 'methodology'] | [ 1.56975865e-01 3.19296122e-01 -3.09060335e-01 -2.51556244e-02
-2.89230943e-01 -2.66878217e-01 -3.14288512e-02 -7.89423883e-02
-4.50403601e-01 1.13099360e+00 -3.30410570e-01 -7.06781447e-01
-7.23332226e-01 -9.63675141e-01 -3.68977726e-01 -1.08898091e+00
-1.09394169e+00 3.78380418e-01 -2.25311831e-01 -2.64206260e-01
1.77841876e-02 6.13317728e-01 -9.05203700e-01 -5.53711414e-01
6.46761954e-01 1.38688612e+00 4.33292061e-01 6.82365716e-01
-2.88603920e-02 4.61529016e-01 -8.07848155e-01 1.30381547e-02
4.45778638e-01 -3.86140138e-01 -5.94949305e-01 -5.64464480e-02
-2.15009212e-01 -3.11826020e-01 -5.77641010e-01 5.02844214e-01
7.33444929e-01 1.34701684e-01 4.59420919e-01 -1.23174095e+00
-1.54855877e-01 6.59600019e-01 -4.49242622e-01 2.37450451e-01
2.18560714e-02 -4.00820583e-01 9.11838412e-01 -1.23832166e-01
8.48832801e-02 8.17858338e-01 6.79740846e-01 2.10461557e-01
-1.08095133e+00 -8.68110120e-01 1.39237925e-01 -3.94996069e-02
-1.74574530e+00 -4.60875422e-01 5.08649051e-01 8.60221088e-02
7.00101435e-01 2.90798277e-01 5.20183921e-01 3.58240455e-01
2.36428827e-01 3.43730003e-01 3.33693027e-01 -6.92698777e-01
4.63870615e-01 3.14559415e-02 -5.71475565e-01 7.09635794e-01
6.20265789e-02 9.55785364e-02 -8.48721340e-02 -1.37367055e-01
9.50104415e-01 -1.16220549e-01 -4.57847357e-01 -3.36526036e-01
-8.37851763e-01 6.59205258e-01 9.79341030e-01 2.83343166e-01
-5.84720075e-01 5.39327860e-01 -9.01293606e-02 6.26979053e-01
3.81562293e-01 6.39549971e-01 -2.25896895e-01 3.76561254e-01
-1.45962691e+00 1.92781106e-01 1.06076109e+00 8.91820669e-01
8.22734058e-01 1.42416731e-01 -7.44413510e-02 6.31499827e-01
3.62090766e-01 5.67972481e-01 -2.52624243e-01 -1.17387974e+00
4.25963432e-01 -1.78361729e-01 2.13645503e-01 -1.13329816e+00
-8.80097091e-01 -1.36662304e+00 -1.09685969e+00 2.26028815e-01
4.12504792e-01 -1.04585910e+00 -5.21553338e-01 1.86428201e+00
-3.48841138e-02 2.56284982e-01 1.25844311e-02 8.90463710e-01
-8.82657096e-02 6.91572070e-01 4.39366177e-02 -4.91152227e-01
5.56331456e-01 -7.51304686e-01 -6.17624998e-01 -2.28234336e-01
6.52977884e-01 -6.15273237e-01 -4.32744138e-02 3.95337120e-02
-1.22556210e+00 -7.75282644e-03 -1.24062371e+00 6.94631398e-01
-1.99682489e-01 5.27716465e-02 6.52592063e-01 9.66383815e-01
-1.47403383e+00 5.40447235e-01 -3.67203712e-01 -4.44320798e-01
5.90772867e-01 1.00816083e+00 1.06943168e-01 -2.12070227e-01
-1.26211715e+00 7.27595806e-01 3.93036366e-01 3.37273985e-01
-5.06979048e-01 -7.00355411e-01 -7.89064288e-01 5.65136194e-01
4.42651838e-01 -5.76481998e-01 1.00771070e+00 -1.08363497e+00
-1.32667542e+00 -1.02312095e-01 -1.55158773e-01 -6.45304859e-01
2.76043922e-01 3.29504579e-01 -3.75997663e-01 1.14657611e-01
1.80516746e-02 4.22603011e-01 3.88815373e-01 -1.12389541e+00
-7.94896960e-01 5.64754158e-02 4.89435911e-01 3.09077710e-01
-4.79889095e-01 -2.84651965e-01 -6.37673438e-01 -5.26876748e-01
-3.48916762e-02 -8.32550585e-01 -8.23572397e-01 1.32912710e-01
-4.05767232e-01 3.22178423e-01 6.01634860e-01 -4.38465446e-01
1.55687070e+00 -2.00894523e+00 -5.90300304e-04 1.08916879e+00
3.18303436e-01 4.04300019e-02 -2.68877745e-01 4.69226271e-01
2.18408823e-01 4.94698659e-02 8.12960491e-02 -2.00081274e-01
-2.34881461e-01 5.65554686e-02 3.66276622e-01 5.37956357e-01
-1.66115910e-01 3.94397020e-01 -1.00903881e+00 -1.72595069e-01
2.18459755e-01 3.35508972e-01 -5.71913540e-01 1.46205619e-01
1.44223943e-01 2.63171583e-01 -5.30246198e-01 3.51912618e-01
8.42116654e-01 -2.97927678e-01 5.81931889e-01 7.38400072e-02
-1.82470754e-01 -3.72237787e-02 -1.19214439e+00 1.42731237e+00
-1.02029991e+00 7.26623714e-01 6.06383622e-01 -1.54627657e+00
5.44374585e-01 5.25116384e-01 1.02614534e+00 -8.00902843e-01
2.61815161e-01 3.25360179e-01 8.72745290e-02 -2.17794359e-01
8.39833319e-02 -3.07073295e-01 2.04063412e-02 5.17211199e-01
1.15054742e-01 5.36325157e-01 2.37473220e-01 2.30977297e-01
1.38841391e+00 -5.74178755e-01 1.54750377e-01 -4.18355674e-01
2.79546201e-01 -4.77837890e-01 3.55679721e-01 1.09870601e+00
-2.27118954e-02 1.79041818e-01 5.78739703e-01 -1.96016178e-01
-7.34761178e-01 -1.18462729e+00 4.92833145e-02 9.52153146e-01
4.22729552e-01 2.09708899e-01 -5.26111841e-01 -1.91686913e-01
-3.72948050e-02 5.03573298e-01 -5.21320343e-01 -1.70057401e-01
-1.55725658e-01 -8.09796989e-01 3.58402640e-01 5.27483560e-02
6.02146626e-01 -4.69128013e-01 -2.08018318e-01 6.41343236e-01
-6.16215877e-02 -1.11428809e+00 -2.59794205e-01 6.23142958e-01
-2.41770059e-01 -7.75589466e-01 -1.00458062e+00 -7.08758831e-01
7.61785984e-01 4.91368085e-01 8.35530818e-01 4.56815004e-01
-9.16634798e-02 1.35948509e-01 -9.36080888e-02 -1.86101258e-01
4.97246757e-02 6.13538921e-01 -1.38604790e-01 -3.31512876e-02
-3.35923821e-01 -7.48210013e-01 -5.60596347e-01 3.18233103e-01
-4.58524555e-01 -1.81916177e-01 7.88103759e-01 5.91664374e-01
3.24253701e-02 7.99912155e-01 7.88686812e-01 -6.86289549e-01
5.67207634e-01 -1.04340363e+00 -6.26474082e-01 2.06617773e-01
-1.66646346e-01 1.06032521e-01 8.90900373e-01 1.54711783e-01
-9.63496804e-01 -7.48070627e-02 -1.95529699e-01 3.10749263e-02
2.96778172e-01 5.79179108e-01 -1.91736922e-01 -6.54003799e-01
4.40837860e-01 -2.20436618e-01 1.17207550e-01 7.89011866e-02
1.44298626e-02 5.19756377e-01 -2.15819050e-02 -2.60483325e-01
8.99533391e-01 3.56296033e-01 4.41942871e-01 -1.06358194e+00
-7.82578111e-01 -3.18711609e-01 -2.28005469e-01 -5.27961373e-01
3.08734268e-01 -8.18081200e-01 -7.75641203e-01 2.65838522e-02
-9.30715621e-01 -7.35535026e-01 1.57889917e-01 6.07405126e-01
-3.09367031e-01 6.08903356e-02 -2.41257042e-01 -9.78124619e-01
-1.09763674e-01 -8.77104402e-01 4.43419158e-01 2.63553411e-01
6.90302551e-02 -1.31999123e+00 -5.51905334e-01 -3.12478751e-01
1.04177177e+00 2.46717319e-01 7.55625010e-01 4.06801328e-02
-6.71411872e-01 -3.89025867e-01 -4.93579805e-01 -6.25558048e-02
9.16596279e-02 -3.33826721e-01 -5.19483447e-01 -3.90948355e-01
-6.62101567e-01 1.83827803e-01 6.15405142e-01 9.57865596e-01
1.32728148e+00 -3.79376411e-01 -6.45792484e-01 6.98740304e-01
1.68002129e+00 8.98418203e-02 5.88016987e-01 1.39574453e-01
3.45066249e-01 3.92378390e-01 3.01176161e-01 7.81019628e-01
4.84925091e-01 6.85652435e-01 8.51373076e-01 -5.43546021e-01
1.23001441e-01 4.41874593e-01 1.75208207e-02 -8.48529711e-02
-1.42438352e-01 -9.03382182e-01 -5.20805359e-01 3.41022968e-01
-1.76776493e+00 -1.14404881e+00 2.78434813e-01 2.35219812e+00
2.02412412e-01 2.18971699e-01 6.07104525e-02 2.28406459e-01
6.24460518e-01 -6.66146055e-02 -3.11658174e-01 -3.52958769e-01
7.84163475e-02 3.37019533e-01 1.30241644e+00 7.85424590e-01
-1.07020748e+00 4.12788153e-01 6.59432268e+00 8.76735210e-01
-1.21784103e+00 -9.51953307e-02 6.27026260e-01 -5.03044665e-01
-1.72005117e-01 -2.37656057e-01 -2.34057516e-01 4.68984723e-01
9.01628017e-01 -1.46909669e-01 9.30604577e-01 3.66249263e-01
8.08386445e-01 -3.16450387e-01 -6.01109445e-01 5.89254558e-01
-4.08122331e-01 -1.58709800e+00 -4.25162703e-01 4.17288721e-01
6.41747236e-01 7.66767338e-02 -2.84610018e-02 1.08946972e-01
2.39499971e-01 -1.04096317e+00 5.50398409e-01 5.07724047e-01
5.94599426e-01 -1.02213347e+00 9.97486115e-01 2.00073570e-01
-1.17172718e+00 -5.66343129e-01 -1.86761066e-01 -4.20689285e-01
2.47480839e-01 9.61018384e-01 -5.96867323e-01 6.11276567e-01
4.56153542e-01 3.48037332e-01 1.37668744e-01 1.53201306e+00
2.15536490e-01 4.13159817e-01 -6.72178864e-01 -1.22494034e-01
5.61446190e-01 2.19752744e-01 2.36716837e-01 1.16996968e+00
7.30226219e-01 -1.02530591e-01 3.31948727e-01 4.98739570e-01
-2.07398534e-01 -4.00789566e-02 -4.80521291e-01 5.14827788e-01
8.11043024e-01 1.31300688e+00 -7.36816049e-01 2.13662922e-01
-4.79087293e-01 6.62502825e-01 3.74528468e-01 9.71441746e-01
-7.16084003e-01 -7.07584381e-01 8.06507885e-01 2.12791994e-01
4.00819063e-01 -4.97196615e-01 -2.98235655e-01 -3.41821522e-01
-1.28720373e-01 -5.61811589e-02 1.85101911e-01 -1.77508488e-01
-6.86037898e-01 3.28696162e-01 -3.00031930e-01 -1.07151544e+00
-8.08452219e-02 -3.96133333e-01 -8.31634939e-01 1.09255195e+00
-1.92918706e+00 -5.77507019e-01 -1.83112144e-01 4.17035162e-01
-1.77061856e-02 -2.53886044e-01 8.15073490e-01 8.52903843e-01
-6.67813599e-01 8.70185733e-01 4.35285538e-01 8.04433972e-03
1.63979143e-01 -1.11010587e+00 -4.39874262e-01 7.43650019e-01
-4.92021799e-01 3.94293606e-01 7.25721538e-01 -1.10404016e-02
-1.12689149e+00 -1.01273036e+00 7.27181017e-01 7.44386613e-01
5.03539145e-01 -2.68904358e-01 -2.59399235e-01 9.65293795e-02
3.01576883e-01 1.08555868e-01 9.52517092e-01 2.60093927e-01
5.02901733e-01 -3.62306684e-01 -1.16184556e+00 6.89051986e-01
9.15392101e-01 9.91574675e-02 8.25057387e-01 4.39959913e-01
2.00163409e-01 -1.96192533e-01 -6.38177395e-01 3.25981863e-02
6.07034683e-01 -7.78499126e-01 9.18978035e-01 1.78301125e-03
-6.22624010e-02 -1.86578006e-01 -1.45820037e-01 -1.54239500e+00
-5.32330573e-01 -7.69338369e-01 1.32433008e-02 8.12885940e-01
8.23344946e-01 -3.89993727e-01 9.25707102e-01 3.95181865e-01
1.10722288e-01 -9.39180315e-01 -1.02437091e+00 -6.17515802e-01
-2.72050649e-01 -3.67117286e-01 4.35778618e-01 5.95890641e-01
3.59394215e-02 1.98915094e-01 -4.02694702e-01 7.51937807e-01
7.11773872e-01 -2.18261182e-01 4.32091296e-01 -1.00887001e+00
-2.60745972e-01 -6.28907084e-01 -3.16438407e-01 -1.31331217e+00
3.26451063e-01 -8.01825583e-01 9.77127030e-02 -1.58847332e+00
-5.29622197e-01 -1.33014369e+00 -5.09310603e-01 2.78418273e-01
1.44062668e-01 2.46670499e-01 -2.86168456e-02 -1.84410453e-01
-7.04642057e-01 2.30279803e-01 1.07256722e+00 -9.88840237e-02
-2.28408903e-01 5.68332553e-01 -8.96959424e-01 2.29110718e-01
1.10718465e+00 -1.45246148e-01 -6.21506512e-01 -6.11877561e-01
1.44418687e-01 4.26386178e-01 2.82302469e-01 -1.30568945e+00
2.30812952e-01 -1.93967432e-01 4.88857985e-01 -2.87950542e-02
2.89075613e-01 -1.27651179e+00 -8.45294744e-02 5.04288077e-01
-3.29683363e-01 -5.83412111e-01 7.76706487e-02 7.03706145e-01
1.21429034e-01 -9.53388959e-02 6.90425158e-01 7.11468160e-02
-7.09586799e-01 4.35376585e-01 -7.58305967e-01 -3.71479034e-01
1.07165980e+00 -2.47103557e-01 2.90607780e-01 -9.72630024e-01
-8.52167010e-01 7.24527955e-01 -1.19741462e-01 -2.17721656e-01
1.65905595e-01 -1.09198856e+00 -5.87249398e-01 1.02817677e-02
-3.92353415e-01 -4.74742681e-01 3.18811536e-01 8.44057381e-01
-5.55055261e-01 4.75061178e-01 -1.34123400e-01 -1.68848649e-01
-7.62078464e-01 5.07350750e-02 8.07310939e-01 -3.91300768e-01
8.23423713e-02 9.37971950e-01 -1.26129538e-01 -5.20378575e-02
4.56285477e-01 4.01617736e-01 -9.06901434e-02 -3.08673322e-01
2.45869994e-01 3.57976347e-01 1.45803288e-01 -2.95164853e-01
-2.18738332e-01 7.88003355e-02 1.22766629e-01 8.98487270e-02
1.15889549e+00 -6.25552773e-01 -1.34711504e-01 -4.33974266e-01
1.35317194e+00 -1.63892478e-01 -1.10689461e+00 -3.63345295e-01
-2.24997595e-01 -4.51151758e-01 6.74196005e-01 -5.76425135e-01
-1.62303460e+00 5.14398873e-01 5.64703107e-01 6.21284723e-01
1.41597450e+00 -2.88541287e-01 5.35574615e-01 4.33674902e-01
6.31842792e-01 -1.10898376e+00 -4.05497849e-01 5.70326209e-01
2.42507190e-01 -9.54397023e-01 -4.36425321e-02 -5.60600519e-01
2.24377319e-01 1.21740496e+00 4.27244842e-01 3.12005146e-03
1.20500410e+00 4.96824086e-01 3.20509225e-02 7.39807487e-02
-3.81244302e-01 -4.61498320e-01 -2.73111343e-01 6.54454708e-01
6.30301714e-01 1.90404266e-01 -2.49607563e-01 4.21036631e-02
1.94780380e-02 -2.77691454e-01 4.42347497e-01 7.89076388e-01
-6.54133618e-01 -1.18229353e+00 -2.44098753e-01 9.12763119e-01
-4.13959861e-01 -1.25792384e-01 3.21285695e-01 6.08782470e-01
1.37164339e-01 1.24823797e+00 1.31757036e-01 -2.20691666e-01
7.17864260e-02 -6.42515242e-01 2.96831548e-01 -4.12606031e-01
-8.21039528e-02 -2.08831895e-02 2.07848236e-01 -4.37157005e-01
-3.53671640e-01 -1.98881894e-01 -1.31453955e+00 -6.82623446e-01
-5.24333596e-01 3.72208148e-01 5.84061623e-01 1.21275997e+00
2.51948535e-01 1.01003969e+00 1.19508076e+00 -1.08165312e+00
-1.25676217e-02 -3.35006982e-01 -1.00467134e+00 -4.34940755e-01
6.17125332e-01 -7.44564652e-01 -1.47647291e-01 -5.73298395e-01] | [6.036064147949219, 1.541663408279419] |
57cd7c8e-f595-4af4-a697-fe06d3552355 | med-danet-dynamic-architecture-network-for | 2206.06575 | null | https://arxiv.org/abs/2206.06575v2 | https://arxiv.org/pdf/2206.06575v2.pdf | Med-DANet: Dynamic Architecture Network for Efficient Medical Volumetric Segmentation | For 3D medical image (e.g. CT and MRI) segmentation, the difficulty of segmenting each slice in a clinical case varies greatly. Previous research on volumetric medical image segmentation in a slice-by-slice manner conventionally use the identical 2D deep neural network to segment all the slices of the same case, ignoring the data heterogeneity among image slices. In this paper, we focus on multi-modal 3D MRI brain tumor segmentation and propose a dynamic architecture network named Med-DANet based on adaptive model selection to achieve effective accuracy and efficiency trade-off. For each slice of the input 3D MRI volume, our proposed method learns a slice-specific decision by the Decision Network to dynamically select a suitable model from the predefined Model Bank for the subsequent 2D segmentation task. Extensive experimental results on both BraTS 2019 and 2020 datasets show that our proposed method achieves comparable or better results than previous state-of-the-art methods for 3D MRI brain tumor segmentation with much less model complexity. Compared with the state-of-the-art 3D method TransBTS, the proposed framework improves the model efficiency by up to 3.5x without sacrificing the accuracy. Our code will be publicly available soon. | ['Jiangyun Li', 'Yan Zhang', 'Sen Zha', 'Jing Wang', 'Chen Chen', 'Wenxuan Wang'] | 2022-06-14 | null | null | null | null | ['volumetric-medical-image-segmentation', 'brain-tumor-segmentation'] | ['medical', 'medical'] | [ 1.12499066e-01 2.90420055e-01 -2.26691678e-01 -6.11584008e-01
-7.87857175e-01 -3.04148197e-01 9.51881856e-02 5.06099500e-02
-6.95890844e-01 3.86703849e-01 -8.48321542e-02 -7.47283459e-01
-1.32303044e-01 -6.92054987e-01 -2.29050592e-01 -7.90371776e-01
-7.60906786e-02 1.04401565e+00 7.04781055e-01 2.16724381e-01
1.22873202e-01 7.15611815e-01 -5.58356583e-01 1.34955734e-01
8.32946599e-01 1.23337746e+00 5.69867074e-01 3.74947339e-01
-5.14128089e-01 4.33144778e-01 -3.75383079e-01 1.18341230e-01
5.02711236e-01 -3.91037107e-01 -1.23462701e+00 4.13684368e-01
1.21000111e-01 -4.39116597e-01 -1.93673924e-01 1.15692008e+00
7.86258757e-01 -2.37441421e-01 6.89256132e-01 -6.83043599e-01
-6.89020231e-02 7.77290761e-01 -8.45800281e-01 7.01199293e-01
-4.75167960e-01 5.95185123e-02 2.22408831e-01 -6.01954222e-01
5.66815853e-01 7.19066918e-01 5.79383790e-01 6.45708859e-01
-9.04392600e-01 -5.73939741e-01 1.78236946e-01 -3.87552045e-02
-1.41880322e+00 -3.49716581e-02 7.81750500e-01 -6.33077025e-01
6.39843643e-01 1.87861368e-01 1.03926933e+00 4.41856891e-01
6.60636842e-01 7.95635283e-01 1.15195680e+00 -1.28470004e-01
2.43140325e-01 -3.22372556e-01 4.26312387e-01 9.41038966e-01
5.12477681e-02 -2.47799084e-01 1.86166540e-01 7.56730214e-02
9.72273052e-01 1.61715131e-03 -2.84205407e-01 -5.47806382e-01
-1.23353469e+00 7.28590190e-01 7.00891197e-01 5.66318989e-01
-4.55244452e-01 -3.68214422e-03 6.50508404e-01 -8.15233402e-03
6.32933378e-01 6.09137863e-02 -2.95430988e-01 3.93850863e-01
-1.30137110e+00 -5.31572336e-03 3.48669320e-01 5.77493072e-01
1.31293043e-01 -6.28712326e-02 -3.17463428e-01 7.27081776e-01
3.25697869e-01 2.87024844e-02 9.89608169e-01 -5.16356587e-01
2.68187761e-01 5.72035193e-01 -3.10103387e-01 -3.97537559e-01
-1.03625906e+00 -7.79913425e-01 -1.09851348e+00 1.23014629e-01
3.36778790e-01 -2.08934084e-01 -1.75841022e+00 1.24205625e+00
7.39089549e-01 4.73600393e-03 -2.83763319e-01 1.05529201e+00
1.16795313e+00 2.24649519e-01 3.69850323e-02 -2.45087251e-01
1.27492583e+00 -1.11557174e+00 -4.76168454e-01 -1.29356176e-01
7.70356297e-01 -4.77071851e-01 5.79696655e-01 1.17992900e-01
-1.28559661e+00 -9.29289460e-02 -8.78345430e-01 1.77940473e-01
-2.46266481e-02 -1.57303318e-01 6.65387452e-01 8.06654394e-01
-1.22993505e+00 2.02049732e-01 -1.30293584e+00 -1.12660766e-01
8.36790264e-01 7.87440896e-01 -7.95876682e-02 1.69118240e-01
-8.97255182e-01 8.71709585e-01 6.30532086e-01 9.30784419e-02
-1.23135233e+00 -1.01960135e+00 -5.21066844e-01 -3.07232052e-01
3.74691039e-01 -6.65024996e-01 1.52500391e+00 -7.64993787e-01
-1.55098200e+00 1.36819971e+00 2.29486916e-02 -6.50712669e-01
8.61444831e-01 3.78426582e-01 -1.24920093e-01 3.27847093e-01
1.01269178e-01 8.50098789e-01 4.48021680e-01 -1.15923464e+00
-4.90932792e-01 -6.61692142e-01 -5.97833237e-03 3.66803944e-01
1.44092157e-01 7.40809366e-02 -7.08719254e-01 -4.41449165e-01
6.34924710e-01 -8.70939672e-01 -8.09468925e-01 -5.98067902e-02
-6.26174927e-01 1.42254069e-01 7.57149756e-01 -6.24269783e-01
9.39405262e-01 -1.63606620e+00 1.09387927e-01 2.66780496e-01
6.36415720e-01 1.20243885e-01 2.84604430e-01 -5.30997574e-01
-1.09711654e-01 8.61883909e-02 -6.87250614e-01 -2.36486331e-01
-5.21818697e-01 -6.45094365e-02 4.70270187e-01 6.89360738e-01
-3.56225461e-01 8.20019901e-01 -6.11746550e-01 -1.00430989e+00
2.45973960e-01 2.69377202e-01 -4.38617945e-01 1.49817422e-01
1.49014980e-01 8.99138689e-01 -7.19775975e-01 6.90494359e-01
8.26016128e-01 -5.87281764e-01 1.07166119e-01 -2.44830683e-01
3.86812538e-02 6.28930656e-03 -7.49922872e-01 1.94527447e+00
-4.67603028e-01 2.73115844e-01 2.30280578e-01 -1.32130075e+00
7.58657992e-01 4.32656109e-01 1.19343829e+00 -6.76029325e-01
4.79237705e-01 3.45074654e-01 4.78653491e-01 -5.04237235e-01
1.33760676e-01 -3.51208508e-01 -6.02996312e-02 4.95611638e-01
2.41315346e-02 -4.30175304e-01 5.64068444e-02 -9.39487815e-02
9.68490183e-01 -3.05724055e-01 2.97136694e-01 -6.24683738e-01
5.07965684e-01 2.25859687e-01 5.73798656e-01 6.38979852e-01
-7.15066910e-01 6.13848150e-01 3.21398407e-01 -5.87448180e-01
-8.35537851e-01 -1.06826794e+00 -5.48976779e-01 4.35736090e-01
2.97387868e-01 3.15108269e-01 -1.22180176e+00 -9.72369730e-01
-3.25366318e-01 4.13418263e-01 -7.78865039e-01 2.07439050e-01
-7.92731106e-01 -1.29624808e+00 3.85784417e-01 3.25494319e-01
8.07644844e-01 -9.01258707e-01 -1.05535257e+00 2.86661476e-01
-9.43840444e-02 -1.02288198e+00 -6.27838016e-01 4.48367625e-01
-1.43342626e+00 -8.92537713e-01 -1.13235629e+00 -9.89234567e-01
1.04310322e+00 1.58914089e-01 1.11858499e+00 1.23683639e-01
-2.30071992e-01 -1.03919007e-01 -2.77058445e-02 -3.18573058e-01
-4.45227712e-01 5.55017352e-01 -2.90838003e-01 -2.65962273e-01
8.23504180e-02 -3.82187009e-01 -9.23693717e-01 2.09763288e-01
-9.76613760e-01 5.89455187e-01 5.45263529e-01 8.35056067e-01
1.00005281e+00 2.44492702e-02 3.32077622e-01 -1.10639083e+00
5.17569423e-01 -4.53849435e-01 -5.70616782e-01 2.45943844e-01
-5.09298086e-01 -1.67485088e-01 3.29446644e-01 -3.43259871e-01
-8.53928030e-01 3.67396086e-01 -2.99708873e-01 -3.39676708e-01
-1.72950193e-01 5.94459534e-01 1.55081883e-01 -4.51838195e-01
3.31142247e-01 1.93752497e-01 2.17909757e-02 -3.43131661e-01
-1.26347989e-01 3.67817819e-01 3.29654545e-01 -1.21053569e-01
4.06937391e-01 5.77910304e-01 1.34653375e-01 -2.67644823e-01
-7.69258916e-01 -3.31240296e-01 -1.02009559e+00 -3.57390076e-01
1.19816089e+00 -6.13117933e-01 -3.70838672e-01 6.57371342e-01
-9.51390028e-01 -5.82819521e-01 -1.01705082e-01 5.15957892e-01
-4.52308297e-01 1.52674690e-01 -8.46501887e-01 -2.45674253e-01
-7.98153579e-01 -2.00939608e+00 1.00398386e+00 2.76229858e-01
7.47100860e-02 -1.02677023e+00 -1.74501285e-01 4.49731261e-01
5.43468535e-01 3.27343702e-01 1.09517622e+00 -7.27164924e-01
-4.64596868e-01 -1.38108686e-01 -1.76745966e-01 -8.92535299e-02
2.83867300e-01 -4.70735222e-01 -5.75767279e-01 -1.52686462e-01
2.27673486e-01 -3.51484753e-02 8.51427138e-01 1.09016955e+00
1.80430806e+00 1.82629541e-01 -5.51617384e-01 9.20781493e-01
1.51482546e+00 6.47084951e-01 1.71762064e-01 3.07458133e-01
8.08892131e-01 1.66982234e-01 2.42513910e-01 2.06726536e-01
3.64647448e-01 4.93321210e-01 6.52089238e-01 -5.96329093e-01
-6.31447062e-02 4.02367502e-01 -3.87906462e-01 9.85110879e-01
1.21576175e-01 -1.32742673e-01 -1.41675603e+00 7.03500986e-01
-1.36947417e+00 -4.38096970e-01 8.40356946e-02 1.94105053e+00
9.12333429e-01 4.12154645e-01 3.95135954e-02 -4.26202305e-02
7.54004419e-01 1.71513289e-01 -8.48322213e-01 -1.31716132e-01
3.62391025e-01 2.84003884e-01 8.20846915e-01 5.48889518e-01
-1.26576555e+00 8.38245928e-01 6.50624228e+00 1.03474045e+00
-1.61230326e+00 5.19262671e-01 1.32469368e+00 -3.24423522e-01
-6.81860521e-02 -4.76957232e-01 -5.81930280e-01 3.39207709e-01
6.40772760e-01 -3.87669285e-03 8.53636581e-03 5.82365155e-01
4.28965598e-01 -2.84302860e-01 -7.72059202e-01 8.51102650e-01
-1.18733063e-01 -1.51983321e+00 3.89204081e-03 1.62614495e-01
7.76791394e-01 3.20372432e-01 -1.58493146e-02 -8.17675814e-02
1.18978381e-01 -1.07556593e+00 6.96813345e-01 4.13812548e-01
7.06733048e-01 -7.36004353e-01 7.97568560e-01 3.20347399e-01
-9.51745629e-01 2.51369804e-01 -1.03944197e-01 6.19326890e-01
3.94127458e-01 7.28223860e-01 -1.11293316e+00 5.27617395e-01
6.66771233e-01 2.62795508e-01 -5.23763061e-01 1.39210773e+00
3.02709430e-01 6.57834470e-01 -2.25567088e-01 1.49963737e-01
6.38246298e-01 -1.15411744e-01 4.25885975e-01 1.02950752e+00
2.85904676e-01 2.56835341e-01 4.23206925e-01 6.78331673e-01
-1.04807034e-01 2.47015655e-01 -1.82945743e-01 4.48746622e-01
1.46815673e-01 1.21940458e+00 -1.53809798e+00 -5.81836581e-01
-1.57057062e-01 7.65638173e-01 1.09362900e-01 -3.06601431e-02
-9.32219863e-01 1.14444352e-03 3.33477706e-02 2.74230033e-01
1.91397056e-01 -1.71682283e-01 -7.52513230e-01 -7.13242233e-01
-2.78044045e-01 -5.34685552e-01 4.11090165e-01 -4.51474339e-01
-9.73373413e-01 1.08762336e+00 1.86069727e-01 -1.09469604e+00
-1.03342727e-01 -4.65319425e-01 -6.42957747e-01 7.52683699e-01
-1.32665908e+00 -9.18906987e-01 -4.16911006e-01 5.34617782e-01
6.40854836e-01 -1.09684365e-02 4.89326626e-01 2.37196967e-01
-4.53640163e-01 3.48992556e-01 -4.80925515e-02 3.63489121e-01
3.35983425e-01 -1.11937094e+00 3.09105605e-01 6.55526519e-01
-4.20875251e-01 -6.03184151e-03 2.57113636e-01 -6.41040385e-01
-8.70608151e-01 -1.21094406e+00 3.75306845e-01 1.28079087e-01
4.36915308e-01 -3.50632966e-02 -7.16651857e-01 5.39376974e-01
3.89155090e-01 3.36780399e-01 5.96788704e-01 -5.52330256e-01
4.44091529e-01 -9.63755995e-02 -1.70463431e+00 5.17550409e-01
1.02206302e+00 1.83694325e-02 -3.24372798e-01 4.83611465e-01
8.82224619e-01 -1.15099609e+00 -1.15255308e+00 7.37764299e-01
2.78830141e-01 -8.76609385e-01 9.25411463e-01 -3.13885659e-01
2.63673395e-01 -1.39638439e-01 1.13415621e-01 -1.37461793e+00
-2.79213607e-01 -7.99362957e-02 2.24688560e-01 3.73734742e-01
4.63419199e-01 -4.60473806e-01 1.06962001e+00 6.97069526e-01
-5.72134018e-01 -1.34822679e+00 -1.30668366e+00 -4.39788401e-01
4.93875891e-01 -3.94198090e-01 8.47432733e-01 8.49936068e-01
-5.15555561e-01 -1.12487152e-01 2.43823305e-01 1.82105035e-01
7.88238168e-01 2.16998994e-01 1.05453275e-01 -9.95918572e-01
-7.97148719e-02 -8.64346564e-01 -4.20298159e-01 -1.01049531e+00
1.62147582e-02 -1.20357943e+00 2.84714326e-02 -1.70389974e+00
4.49841321e-01 -6.75001562e-01 -4.47296590e-01 3.25104266e-01
2.61042938e-02 2.91286737e-01 7.78501257e-02 9.50956792e-02
-4.03773427e-01 2.64051080e-01 2.02163577e+00 -5.19573927e-01
-2.83531278e-01 1.33263856e-01 -2.82387465e-01 8.24267924e-01
9.04198050e-01 -5.62927306e-01 -4.80288714e-01 -7.47982442e-01
-8.07629049e-01 6.20374143e-01 1.68017715e-01 -1.09993970e+00
3.07341933e-01 -1.70298949e-01 6.13396823e-01 -9.49899793e-01
9.91032794e-02 -8.16100478e-01 -1.87387094e-02 8.13554704e-01
-3.39479297e-01 8.80235899e-03 3.47473264e-01 -8.06883816e-03
-5.61021566e-02 -4.51911271e-01 1.16531813e+00 -5.83625376e-01
-4.69346315e-01 1.04569221e+00 -5.39227903e-01 9.22183543e-02
1.13660896e+00 -3.39797437e-01 1.70800418e-01 7.86214620e-02
-1.03574228e+00 1.57032415e-01 1.61485314e-01 5.39442338e-02
5.35149157e-01 -1.14506185e+00 -7.19175637e-01 2.64905822e-02
-2.90505856e-01 4.98071402e-01 5.74470282e-01 1.14308763e+00
-7.15243638e-01 4.86979902e-01 -3.66853088e-01 -1.06388009e+00
-1.13466096e+00 2.10633427e-01 1.17582858e+00 -8.76141369e-01
-7.13277042e-01 8.96387458e-01 4.79917109e-01 -5.21997631e-01
4.33271043e-02 -6.19960010e-01 -1.71903104e-01 -2.94131458e-01
2.46044651e-01 -1.56010106e-01 4.41935122e-01 -6.46523416e-01
-5.54926515e-01 5.13019800e-01 -3.61762643e-01 -1.10612519e-01
1.31172538e+00 2.54202913e-02 -8.88966471e-02 2.72154897e-01
1.17128646e+00 -4.99551386e-01 -1.11336863e+00 -2.57637590e-01
-1.99219763e-01 -2.18431249e-01 6.73136353e-01 -9.37155664e-01
-1.91984296e+00 7.52086699e-01 1.12948382e+00 -1.12562208e-02
1.32305562e+00 -1.11238249e-02 9.67376947e-01 -1.97034791e-01
5.67447782e-01 -8.09987366e-01 -3.17787647e-01 4.10435438e-01
6.62666619e-01 -1.41787112e+00 -7.65276328e-03 -3.78969371e-01
-5.30223131e-01 9.75773573e-01 8.06258857e-01 -1.09309234e-01
1.03140831e+00 4.19451326e-01 1.55951798e-01 -4.49744254e-01
-3.19380075e-01 1.00490950e-01 2.86011606e-01 4.60175067e-01
5.64933836e-01 3.48718703e-01 -3.07888269e-01 4.79252130e-01
-7.25301057e-02 4.74465564e-02 4.26462770e-01 7.81436026e-01
-3.72718304e-01 -8.46144915e-01 -4.14298326e-01 8.08065534e-01
-6.01228297e-01 -1.35580346e-01 -3.07161473e-02 8.64828587e-01
2.38392912e-02 4.69446480e-01 1.80446923e-01 -2.89994162e-02
1.47483293e-02 -7.44123831e-02 6.80094004e-01 -4.78765517e-01
-9.44216907e-01 2.92949945e-01 -2.70812035e-01 -3.92548859e-01
-5.23457825e-01 -6.29749894e-01 -1.69628441e+00 -2.28541940e-02
-1.22563682e-01 -4.56289835e-02 6.28450692e-01 9.97021973e-01
1.26983210e-01 8.68045330e-01 6.12322509e-01 -9.48446155e-01
-3.11347425e-01 -8.97504270e-01 -5.56059599e-01 -4.96710055e-02
2.37633973e-01 -7.48284936e-01 1.76317781e-01 -1.66828424e-01] | [14.561861991882324, -2.4257001876831055] |
ffbbbcb4-a16d-4ddb-abf7-57782c576661 | argscichat-a-dataset-for-argumentative | 2202.06690 | null | https://arxiv.org/abs/2202.06690v3 | https://arxiv.org/pdf/2202.06690v3.pdf | ArgSciChat: A Dataset for Argumentative Dialogues on Scientific Papers | The applications of conversational agents for scientific disciplines (as expert domains) are understudied due to the lack of dialogue data to train such agents. While most data collection frameworks, such as Amazon Mechanical Turk, foster data collection for generic domains by connecting crowd workers and task designers, these frameworks are not much optimized for data collection in expert domains. Scientists are rarely present in these frameworks due to their limited time budget. Therefore, we introduce a novel framework to collect dialogues between scientists as domain experts on scientific papers. Our framework lets scientists present their scientific papers as groundings for dialogues and participate in dialogue they like its paper title. We use our framework to collect a novel argumentative dialogue dataset, ArgSciChat. It consists of 498 messages collected from 41 dialogues on 20 scientific papers. Alongside extensive analysis on ArgSciChat, we evaluate a recent conversational agent on our dataset. Experimental results show that this agent poorly performs on ArgSciChat, motivating further research on argumentative scientific agents. We release our framework and the dataset. | ['Iryna Gurevych', 'Mohsen Mesgar', 'Federico Ruggeri'] | 2022-02-14 | null | null | null | null | ['fact-selection'] | ['natural-language-processing'] | [-4.01938498e-01 6.90258563e-01 -1.26305491e-01 -4.57354963e-01
-6.74785078e-01 -1.29714048e+00 1.18472111e+00 3.35719466e-01
-5.26537299e-01 1.22115505e+00 5.20279348e-01 -6.70856416e-01
3.78127508e-02 -6.52703583e-01 -5.34522355e-01 -3.43449324e-01
5.46637416e-01 1.20789707e+00 1.74679741e-01 -3.75717044e-01
5.54842591e-01 9.35224518e-02 -9.18116271e-01 4.44554240e-01
1.02052653e+00 2.19486598e-02 1.99501753e-01 9.76037860e-01
-6.51780605e-01 1.20332766e+00 -1.27212596e+00 -8.12679231e-01
-6.16336763e-02 -4.76906270e-01 -1.55782926e+00 -2.37377286e-01
1.50701374e-01 -2.80795544e-01 1.69675887e-01 5.86542189e-01
7.65169919e-01 1.17017925e-01 4.35077697e-01 -1.54245579e+00
-6.40069127e-01 1.30194116e+00 -2.50873357e-01 1.85790941e-01
7.47681260e-01 8.52983147e-02 1.25185323e+00 -7.07231402e-01
1.24893367e+00 1.65345359e+00 4.29759502e-01 9.40470278e-01
-6.93917096e-01 -4.40298319e-01 1.88184440e-01 1.93147883e-01
-4.41774726e-01 -5.38989604e-01 7.47507751e-01 -6.43956959e-01
5.41752577e-01 4.59812224e-01 5.84867239e-01 1.65240252e+00
-5.24654150e-01 1.13872457e+00 1.20351899e+00 -2.93513775e-01
5.90436697e-01 3.12275261e-01 5.15567243e-01 2.17467830e-01
1.61932200e-01 -7.29224622e-01 -9.01805341e-01 -6.45023525e-01
4.62575912e-01 -7.00591624e-01 -1.13196322e-03 1.01216264e-01
-1.50831616e+00 1.11480629e+00 -5.07856496e-02 2.01089263e-01
-3.48869324e-01 -2.44640633e-01 5.63199341e-01 4.01250273e-01
5.96174717e-01 1.00195134e+00 -2.43920043e-01 -6.47258103e-01
-1.19174547e-01 9.15574014e-01 1.98941505e+00 1.01156235e+00
5.26839554e-01 -6.08865917e-01 -3.85182023e-01 9.00385797e-01
2.83425510e-01 3.50376427e-01 -3.71813662e-02 -1.59865153e+00
5.31747878e-01 8.33973527e-01 6.46513104e-01 -6.80421650e-01
-3.68172973e-01 1.01367332e-01 -3.12468052e-01 -1.05454870e-01
9.24357653e-01 -6.91005170e-01 3.27060781e-02 1.31917191e+00
6.06726348e-01 -1.92734316e-01 5.52768826e-01 1.07747149e+00
1.75855732e+00 4.96014237e-01 2.46267244e-01 -5.15306070e-02
1.67643690e+00 -1.03299081e+00 -8.29918683e-01 -7.56710023e-02
7.46880531e-01 -1.02432597e+00 1.00619447e+00 2.40809962e-01
-1.29026330e+00 6.41114712e-02 -4.76217747e-01 -3.81105006e-01
-4.49398160e-01 -1.04632139e-01 8.45740855e-01 5.31126857e-01
-9.28166211e-01 1.95276842e-01 -3.43163073e-01 -5.61690927e-01
2.93664753e-01 -6.35765120e-02 -4.48192097e-02 3.59766632e-01
-1.23301113e+00 9.23114359e-01 -1.19007982e-01 -7.62852505e-02
-8.72281849e-01 -8.17813754e-01 -4.20287460e-01 -3.08930993e-01
4.28818345e-01 -7.31133759e-01 1.77110755e+00 -3.94460350e-01
-1.84229994e+00 1.38109314e+00 -1.29963472e-01 -4.52933669e-01
1.01180339e+00 -8.91151577e-02 2.52950728e-01 1.84086680e-01
4.99978840e-01 6.16000354e-01 -2.78991126e-02 -1.42863679e+00
-7.74204135e-01 3.24582122e-02 6.54863060e-01 2.20266193e-01
-1.22363143e-01 3.79394591e-01 -1.54161409e-01 -1.84247553e-01
-6.83617949e-01 -1.09562445e+00 -1.23564824e-01 -2.99496382e-01
-7.02852070e-01 -1.26032114e+00 8.19681287e-01 -1.70690849e-01
5.11675358e-01 -1.44131541e+00 4.95352820e-02 -1.85460851e-01
6.36166573e-01 3.30590963e-01 -1.56222895e-01 7.38865316e-01
4.54328954e-01 4.26163226e-01 3.82120386e-02 -3.84073883e-01
3.00930053e-01 5.18928945e-01 -4.07068461e-01 7.88873732e-02
7.21714273e-02 8.75243068e-01 -1.52056801e+00 -6.59376562e-01
-1.18485399e-01 -3.01372837e-02 -2.53739715e-01 6.02048278e-01
-7.77729392e-01 6.74417198e-01 -1.16197097e+00 5.01357019e-01
6.43651068e-01 -5.35276711e-01 3.46295804e-01 5.35262764e-01
-5.52210987e-01 6.05384588e-01 -4.99940157e-01 1.64983881e+00
-2.10875511e-01 9.45455551e-01 6.07824802e-01 -8.42005372e-01
1.01486492e+00 5.11378288e-01 3.74461621e-01 -2.53040075e-01
-1.75768211e-01 3.89109164e-01 1.51550114e-01 -8.45568478e-01
4.16112721e-01 3.77067953e-01 -1.01029344e-01 1.17495179e+00
-4.07371581e-01 -4.32185739e-01 7.29618371e-01 8.14433038e-01
1.29466200e+00 -1.13541055e-02 -2.70787895e-01 -5.09247601e-01
6.65004790e-01 5.93139946e-01 4.38474625e-01 1.24892497e+00
-3.79924864e-01 9.24655348e-02 1.00756335e+00 -7.40950286e-01
-9.61501360e-01 -5.66974640e-01 -6.37870058e-02 1.63923025e+00
-1.06575880e-02 -3.45122367e-01 -8.41361463e-01 -8.97971332e-01
4.49671261e-02 5.71176648e-01 -5.35056531e-01 8.23096693e-01
-6.49412334e-01 -7.18027234e-01 6.94559932e-01 1.48986857e-02
6.35189891e-01 -1.48591197e+00 -6.61516249e-01 3.51813585e-01
-6.17448866e-01 -1.34551764e+00 -2.76364833e-02 -2.09264860e-01
-1.99460551e-01 -1.43943977e+00 -9.11277294e-01 -6.70759320e-01
2.66919136e-01 2.84517944e-01 1.59382522e+00 2.02273205e-01
2.13540010e-02 6.03135943e-01 -4.67927009e-01 -1.06576717e+00
-9.96779144e-01 5.23439109e-01 -2.00776801e-01 -5.28589666e-01
7.35694766e-01 -4.18256104e-01 -6.27685785e-01 6.06266379e-01
-4.11654741e-01 1.71461821e-01 1.77903429e-01 6.06526434e-01
-5.17312407e-01 -9.15714025e-01 1.00684929e+00 -1.33196652e+00
1.49131310e+00 -7.56213367e-01 -3.66240889e-01 3.29786152e-01
-2.14268774e-01 -1.68055445e-01 4.35164630e-01 -1.69393703e-01
-1.38038909e+00 -5.84331930e-01 7.65017346e-02 3.73000175e-01
-1.90093458e-01 5.01388967e-01 3.40757281e-01 1.36462733e-01
1.02942336e+00 -3.93191010e-01 3.48531716e-02 -5.48480570e-01
2.17996180e-01 9.26339388e-01 4.10302818e-01 -1.41021276e+00
3.94379735e-01 3.35872054e-01 -4.10139203e-01 -8.98015082e-01
-1.01224053e+00 -4.97117370e-01 -1.94144741e-01 -3.18315387e-01
7.96909094e-01 -7.46296704e-01 -1.78148985e+00 8.13484043e-02
-1.94699335e+00 -4.46408331e-01 -1.08452037e-01 3.37028682e-01
-2.39679858e-01 3.19124609e-01 -8.15712631e-01 -1.03506982e+00
-4.98705417e-01 -1.01405203e+00 8.43905985e-01 5.34811318e-01
-7.32779860e-01 -1.24556375e+00 4.80957806e-01 8.46544921e-01
4.71074432e-01 1.87162429e-01 7.02750981e-01 -1.10905349e+00
-2.69392669e-01 2.09648713e-01 -3.48685175e-01 -4.25473213e-01
-1.65026858e-01 2.41555691e-01 -1.11399221e+00 2.27249399e-01
-3.54194850e-01 -8.76226902e-01 4.47663546e-01 -4.24877778e-02
8.31422091e-01 -3.77684116e-01 -3.54124784e-01 -2.01599658e-01
3.48742187e-01 2.08724767e-01 1.78248435e-01 7.93431997e-01
4.34385419e-01 1.29868567e+00 3.47036421e-01 6.98257685e-01
8.53733957e-01 2.75956511e-01 6.84760064e-02 -2.58939583e-02
1.76593632e-01 1.53398842e-01 1.34648189e-01 8.25337410e-01
-5.21489441e-01 -6.33557975e-01 -1.38262475e+00 7.53465831e-01
-2.41058350e+00 -8.74770761e-01 -5.56005597e-01 1.19415510e+00
1.37150180e+00 -2.28792161e-01 3.51962507e-01 -4.73223418e-01
7.28710532e-01 1.37884185e-01 -5.46938658e-01 -5.14309525e-01
-4.13963437e-01 -1.16928272e-01 -1.13148570e-01 6.81442320e-01
-7.99436390e-01 9.77726161e-01 6.34139538e+00 3.77943635e-01
-4.06855911e-01 8.29687268e-02 6.28694475e-01 2.19898984e-01
-5.93002081e-01 1.84535950e-01 -7.50399411e-01 4.22189623e-01
7.60037661e-01 -6.41425252e-01 1.72547311e-01 7.21600115e-01
4.27175522e-01 -2.57456422e-01 -1.51494849e+00 6.27520204e-01
-3.29285175e-01 -1.71821606e+00 -1.07688300e-01 2.41573751e-02
6.91813469e-01 -4.09248322e-02 -4.02835429e-01 2.81430423e-01
1.43505430e+00 -1.00309968e+00 3.22594076e-01 3.26116920e-01
-6.29138947e-02 -2.63515025e-01 7.45715559e-01 5.89451373e-01
-2.01244265e-01 3.90254587e-01 -3.66569698e-01 -4.06554729e-01
8.41173306e-02 5.93703508e-01 -1.47884405e+00 2.82836556e-01
6.37360692e-01 6.63940072e-01 -2.52555400e-01 5.82669377e-01
-3.45917821e-01 7.18571126e-01 -2.28253961e-01 -8.27775121e-01
4.24472839e-01 -3.67597073e-01 7.45591819e-01 1.42677557e+00
-2.94191241e-01 5.13472140e-01 3.96446109e-01 1.09442306e+00
-6.49873734e-01 -4.55368385e-02 -5.96758723e-01 -3.83071184e-01
8.41863155e-01 1.54948616e+00 -5.69251001e-01 -5.07629335e-01
-2.06646740e-01 3.25298846e-01 4.03332710e-01 4.78342801e-01
-3.80648166e-01 -2.64801145e-01 5.77766478e-01 -1.54685825e-01
-5.04268110e-01 -5.07037453e-02 -4.00765210e-01 -9.22875285e-01
-4.34547551e-02 -1.19336784e+00 3.50055546e-01 -7.10117698e-01
-1.79752982e+00 4.82805401e-01 -1.53024927e-01 -5.74945629e-01
-5.44486403e-01 -4.49766278e-01 -9.60633159e-01 8.31147313e-01
-1.35694242e+00 -1.02805471e+00 -5.24144709e-01 4.18387473e-01
6.17430151e-01 -4.46655720e-01 9.83149350e-01 -2.03500792e-01
-5.98007143e-01 8.07006732e-02 -1.79663435e-01 4.67891604e-01
1.17330289e+00 -1.39789927e+00 5.82640707e-01 8.33913013e-02
-2.72621036e-01 9.12913680e-01 1.04854667e+00 -7.81618834e-01
-1.69239676e+00 -3.84006441e-01 9.99588251e-01 -1.06704569e+00
9.65131879e-01 -5.34331203e-01 -9.03964996e-01 4.30739999e-01
9.92538452e-01 -4.26531196e-01 8.20136428e-01 4.48178500e-01
-1.49624914e-01 4.85851675e-01 -8.81872714e-01 6.96964979e-01
1.01606655e+00 -5.15106976e-01 -8.26013625e-01 1.04104900e+00
5.09349823e-01 -5.21248043e-01 -8.58634055e-01 -1.82800665e-01
4.36717391e-01 -7.35232353e-01 6.77540958e-01 -9.56658244e-01
9.03787255e-01 -3.11347935e-02 5.47531545e-01 -1.39014661e+00
2.29625821e-01 -1.20825243e+00 7.46696815e-02 1.38780594e+00
5.80725193e-01 -6.55432284e-01 9.70793605e-01 1.31665754e+00
5.34498245e-02 -1.30209267e-01 -6.48033381e-01 -3.50929856e-01
4.55467641e-01 3.95448506e-03 4.77599382e-01 1.49314153e+00
4.38364178e-01 8.10029209e-01 -4.29190509e-02 -3.98193985e-01
5.79824507e-01 3.60893130e-01 1.32213080e+00 -1.64944184e+00
1.06117604e-02 -5.76706111e-01 4.77061838e-01 -9.91973460e-01
5.11009753e-01 -7.00117946e-01 1.24107748e-01 -1.87265599e+00
2.83105731e-01 -6.61368787e-01 6.23887300e-01 5.03848553e-01
-2.50995755e-01 -2.84925640e-01 -5.85437156e-02 4.16007608e-01
-9.31108773e-01 3.11146230e-01 1.42252123e+00 -1.96947992e-01
-2.74484575e-01 -1.24829412e-01 -9.30294931e-01 7.79704511e-01
6.83278620e-01 -3.28350484e-01 -2.07948267e-01 -5.83912373e-01
5.89008689e-01 5.71511164e-02 3.85796040e-01 -2.00800925e-01
7.39493430e-01 -4.94534314e-01 3.63876224e-02 -4.09820348e-01
-7.31054321e-02 -4.57759276e-02 -3.32022876e-01 -8.16491339e-03
-8.74934435e-01 -1.00829333e-01 -9.59615186e-02 4.00750011e-01
-4.74493764e-02 -4.00862157e-01 1.09752461e-01 -4.85228002e-01
-3.51740003e-01 -2.02466637e-01 -6.96843684e-01 6.78308547e-01
9.00877893e-01 3.06519598e-01 -1.25715756e+00 -6.82052612e-01
-5.21623850e-01 8.83223891e-01 3.58769238e-01 3.80663007e-01
1.41407132e-01 -8.09260011e-01 -1.49039567e+00 -6.93555474e-01
7.90818362e-04 3.56343895e-01 -6.35849088e-02 4.80260164e-01
-6.60297751e-01 4.50597495e-01 -8.09478611e-02 -4.86772478e-01
-1.15470207e+00 4.52116951e-02 -4.23273221e-02 -2.62555718e-01
-5.26659012e-01 1.00234652e+00 2.44914487e-01 -7.49170125e-01
3.17714989e-01 1.58967167e-01 -5.27468443e-01 4.01207119e-01
6.48206830e-01 5.27824163e-01 -1.76187217e-01 3.57698393e-03
-4.35263246e-01 -1.43621536e-02 -2.84383327e-01 -4.94226158e-01
1.38218701e+00 1.69754177e-02 -4.99868184e-01 4.68057275e-01
6.24938369e-01 2.90020674e-01 -8.38683963e-01 -3.28460187e-01
4.04570967e-01 -2.27288947e-01 -4.61961091e-01 -9.49323833e-01
-3.75129253e-01 5.78571558e-01 -4.35955554e-01 8.06261241e-01
-3.38621214e-02 3.18612754e-01 4.89791095e-01 9.76359785e-01
9.91247073e-02 -1.33999658e+00 1.92545325e-01 9.35070217e-01
1.15688324e+00 -1.63517439e+00 4.57042828e-02 -4.58700210e-01
-8.03918898e-01 1.21116483e+00 7.54240394e-01 1.12969264e-01
2.19845191e-01 3.54206175e-01 5.62520802e-01 -6.24481499e-01
-1.27494359e+00 8.92926827e-02 -3.16638261e-01 5.48694253e-01
8.79786491e-01 6.82034940e-02 -6.03463292e-01 6.24727190e-01
-4.96811360e-01 -1.18378997e-01 1.05521607e+00 1.07661974e+00
-2.01771647e-01 -1.18536818e+00 -4.29520160e-01 -4.59986217e-02
-5.13507366e-01 1.95733041e-01 -1.44400692e+00 6.29051745e-01
-3.79701048e-01 1.45930290e+00 2.55215801e-02 3.62849593e-01
3.56538206e-01 -3.21732052e-02 -2.12869290e-02 -6.00926936e-01
-1.19266117e+00 -3.32981765e-01 9.35929716e-01 -4.26809005e-02
-9.32409108e-01 -5.27568996e-01 -1.10881066e+00 -8.81625652e-01
2.25747108e-01 1.03073692e+00 8.07168961e-01 9.30682957e-01
3.47417623e-01 1.97776392e-01 4.83258277e-01 -5.98186255e-01
-2.38893032e-01 -1.11838353e+00 2.14843657e-02 5.42616189e-01
7.76303113e-02 -3.70276451e-01 -2.56068587e-01 -4.55262475e-02] | [12.553019523620605, 7.994938850402832] |
296dc339-5732-4de2-a86a-bb95610548e9 | resources-for-brewing-beir-reproducible | 2306.07471 | null | https://arxiv.org/abs/2306.07471v1 | https://arxiv.org/pdf/2306.07471v1.pdf | Resources for Brewing BEIR: Reproducible Reference Models and an Official Leaderboard | BEIR is a benchmark dataset for zero-shot evaluation of information retrieval models across 18 different domain/task combinations. In recent years, we have witnessed the growing popularity of a representation learning approach to building retrieval models, typically using pretrained transformers in a supervised setting. This naturally begs the question: How effective are these models when presented with queries and documents that differ from the training data? Examples include searching in different domains (e.g., medical or legal text) and with different types of queries (e.g., keywords vs. well-formed questions). While BEIR was designed to answer these questions, our work addresses two shortcomings that prevent the benchmark from achieving its full potential: First, the sophistication of modern neural methods and the complexity of current software infrastructure create barriers to entry for newcomers. To this end, we provide reproducible reference implementations that cover the two main classes of approaches: learned dense and sparse models. Second, there does not exist a single authoritative nexus for reporting the effectiveness of different models on BEIR, which has led to difficulty in comparing different methods. To remedy this, we present an official self-service BEIR leaderboard that provides fair and consistent comparisons of retrieval models. By addressing both shortcomings, our work facilitates future explorations in a range of interesting research questions that BEIR enables. | ['Jimmy Lin', 'Jheng-Hong Yang', 'Xueguang Ma', 'Carlos Lassance', 'Nandan Thakur', 'Ehsan Kamalloo'] | 2023-06-13 | null | null | null | null | ['information-retrieval'] | ['natural-language-processing'] | [-9.62341204e-02 -3.61336619e-01 -5.28974652e-01 -1.09293342e-01
-1.23735499e+00 -7.54169405e-01 9.34932828e-01 3.66989255e-01
-8.07458580e-01 4.94965047e-01 5.72340310e-01 -5.30284822e-01
-5.04459202e-01 -6.36614919e-01 -3.89285028e-01 1.01510407e-02
1.91411674e-01 7.24512219e-01 2.07280919e-01 -5.71419537e-01
4.23001200e-01 3.60251516e-01 -1.70866764e+00 2.82436609e-01
4.46272671e-01 8.72091115e-01 1.48662999e-01 4.92074877e-01
-1.97572961e-01 7.37876296e-01 -7.76972950e-01 -4.11296397e-01
2.54036784e-01 -1.45521760e-01 -9.95012760e-01 -4.85787541e-01
4.51524287e-01 -4.00962621e-01 -6.80154622e-01 7.22665846e-01
6.82683647e-01 3.04162830e-01 8.12689841e-01 -9.05876994e-01
-9.98776078e-01 5.46841681e-01 -3.63480598e-01 5.41534126e-01
5.56138992e-01 4.91429120e-02 1.35008061e+00 -8.89613450e-01
1.07169032e+00 1.22460639e+00 4.84748930e-01 5.87020159e-01
-1.13653100e+00 -5.50175607e-01 -8.84805713e-03 2.31338650e-01
-1.35970831e+00 -6.51478410e-01 4.24220443e-01 -3.49143922e-01
1.23905110e+00 3.91967833e-01 2.17384562e-01 1.40882528e+00
-1.99846491e-01 7.14177966e-01 9.26853836e-01 -4.22188193e-01
1.76386982e-01 2.99508542e-01 5.76731861e-01 1.12235740e-01
4.48209196e-01 2.10222062e-02 -5.00004053e-01 -5.95686376e-01
4.90183413e-01 3.97483073e-02 -4.21878904e-01 -1.02347404e-01
-1.03927088e+00 9.33175802e-01 2.71519184e-01 6.09210789e-01
-1.83516502e-01 -1.44617021e-01 6.35222614e-01 6.05955958e-01
4.07455057e-01 1.02618408e+00 -4.54300314e-01 -2.94457465e-01
-1.14847553e+00 5.55089593e-01 1.12486267e+00 8.04032981e-01
6.52820289e-01 -2.92170823e-01 -1.91438004e-01 1.09626162e+00
2.97339648e-01 1.32555991e-01 6.16500139e-01 -8.27337503e-01
3.20474863e-01 3.44022363e-01 1.86079174e-01 -1.08138859e+00
-1.63372338e-01 -3.72016937e-01 -5.45888543e-01 -2.17489958e-01
5.24406850e-01 -3.43695432e-02 -8.63365650e-01 1.40224719e+00
-2.07610220e-01 -1.62890598e-01 1.61704287e-01 1.04299378e+00
1.24893713e+00 6.68320179e-01 3.79544161e-02 1.08586408e-01
1.43351567e+00 -6.62903488e-01 -4.90799665e-01 -2.92414308e-01
6.44551337e-01 -9.35555160e-01 1.18633759e+00 2.35042974e-01
-1.19201922e+00 -1.12938173e-01 -8.75993192e-01 -4.17323828e-01
-8.25613737e-01 -3.93317640e-01 6.31352186e-01 2.87368000e-01
-1.08526719e+00 5.03218234e-01 -4.68046933e-01 -7.38795757e-01
8.96969289e-02 -1.09259121e-01 -3.43403250e-01 -4.26299453e-01
-1.55764914e+00 1.19533515e+00 6.94819838e-02 -3.72317404e-01
-7.97443688e-01 -7.94762731e-01 -5.93524754e-01 2.34187141e-01
3.02821666e-01 -5.72068453e-01 1.48630691e+00 -6.32792532e-01
-7.99223423e-01 1.02411115e+00 2.67747993e-04 -5.13221860e-01
3.23821694e-01 -2.44752452e-01 -4.74050999e-01 1.23826832e-01
3.59272733e-02 4.11206186e-01 3.86543542e-01 -9.21576202e-01
-2.32897699e-01 -3.59287262e-01 3.81391466e-01 2.97317445e-01
-4.13001776e-01 5.38485110e-01 -6.88280344e-01 -4.90738243e-01
-3.93252343e-01 -5.92206359e-01 -1.55315131e-01 -1.69467822e-01
3.29730213e-02 -4.98192191e-01 4.55663592e-01 -5.69755256e-01
1.42276418e+00 -2.02365398e+00 -1.52000422e-02 6.72751740e-02
2.24587694e-01 3.99533540e-01 -2.97755867e-01 9.75529909e-01
1.00792021e-01 5.83137989e-01 2.01489255e-01 -1.80089056e-01
9.16312337e-02 2.13727076e-02 -6.11920953e-01 2.98537970e-01
2.83261016e-03 1.02744544e+00 -9.34992969e-01 -5.98016441e-01
-2.45733216e-01 6.56567574e-01 -4.32085007e-01 1.38895705e-01
-3.15475017e-01 -2.42763944e-02 -5.54107487e-01 7.15263605e-01
1.61155030e-01 -5.14262080e-01 -3.47054265e-02 -6.18951470e-02
-1.56500101e-01 8.44326198e-01 -9.78519559e-01 1.83937073e+00
-3.43598038e-01 7.13402331e-01 9.16090906e-02 -8.91112566e-01
7.68742800e-01 3.91069531e-01 3.61341327e-01 -8.70518923e-01
-3.20141986e-02 4.07238632e-01 -1.34995118e-01 -3.31220537e-01
9.05238867e-01 6.07911609e-02 -8.65524635e-02 7.25348473e-01
2.25702748e-01 -1.76901072e-02 2.45375052e-01 4.83509451e-01
1.40711284e+00 -3.32678229e-01 2.21856952e-01 -2.98152447e-01
-2.26450294e-01 1.72741234e-01 2.13812605e-01 1.24330032e+00
-1.22532323e-01 7.56805182e-01 2.53209263e-01 -3.77537549e-01
-9.71622884e-01 -8.32534850e-01 -3.55311811e-01 1.35423517e+00
-9.14642289e-02 -6.41605258e-01 -1.64739668e-01 -4.78786021e-01
1.55401722e-01 5.83559811e-01 -6.63270891e-01 -7.30899945e-02
-4.06828374e-01 -4.56310391e-01 5.98986924e-01 3.74103993e-01
3.41760814e-02 -9.73917842e-01 -7.02686489e-01 3.13459896e-02
-1.61547542e-01 -7.56800413e-01 -3.35847259e-01 2.50567317e-01
-7.02928483e-01 -1.11897767e+00 -1.16102076e+00 -6.95192456e-01
1.84148312e-01 6.69260740e-01 1.70093918e+00 3.34670067e-01
-3.44153225e-01 9.91074979e-01 -6.41745567e-01 -4.58740354e-01
-1.85888857e-01 4.84315425e-01 -3.02496403e-01 -6.53709233e-01
8.47448111e-01 -3.47998202e-01 -7.63713360e-01 2.54197747e-01
-1.21330822e+00 -4.23114210e-01 6.12839103e-01 7.54049778e-01
2.37812981e-01 -4.40140843e-01 7.58879483e-01 -9.98969316e-01
1.28284144e+00 -1.01076877e+00 -4.08058167e-01 5.54719388e-01
-8.08597803e-01 -1.69964693e-02 2.35125139e-01 -3.76564622e-01
-7.38208055e-01 -6.71723545e-01 -1.41912892e-01 -3.17388505e-01
-1.61368027e-01 1.02079189e+00 5.67411661e-01 2.75786608e-01
1.16540122e+00 2.08082646e-01 -1.62424833e-01 -8.23048711e-01
3.98860991e-01 9.94665444e-01 2.31689095e-01 -6.83216035e-01
6.23423040e-01 1.12975508e-01 -7.30707824e-01 -9.17041600e-01
-6.55984759e-01 -8.99258554e-01 1.01794556e-01 8.14457685e-02
5.56063592e-01 -9.66392875e-01 -2.23082602e-01 1.31731341e-02
-1.12152457e+00 -2.00144410e-01 -4.56344903e-01 3.17911237e-01
-1.99088916e-01 2.83649385e-01 -7.76194394e-01 -5.54090142e-01
-5.75394690e-01 -9.51543808e-01 1.01838160e+00 1.91525191e-01
-4.58103001e-01 -1.02812731e+00 4.54185635e-01 2.94800460e-01
1.02066088e+00 -7.34428465e-02 9.95922029e-01 -1.07121003e+00
-3.27507377e-01 -4.97486264e-01 -2.99158573e-01 8.29476416e-02
-1.37722224e-01 -6.37301505e-02 -8.57039034e-01 -4.82712299e-01
-1.21965222e-01 -9.26144838e-01 1.02906024e+00 9.37925801e-02
1.09081221e+00 -2.12210953e-01 -4.06698823e-01 1.67219892e-01
1.28262603e+00 1.15155622e-01 5.61776519e-01 5.31987131e-01
5.75805083e-02 6.72624767e-01 4.72722203e-01 3.76755297e-01
4.30787474e-01 6.84446335e-01 1.01262093e-01 -2.35209614e-02
-7.59157836e-02 -2.08297446e-01 4.64881584e-02 6.08526111e-01
5.00846170e-02 -3.06166321e-01 -1.15086639e+00 7.75844038e-01
-1.65288496e+00 -9.60092187e-01 4.33971316e-01 2.31415248e+00
1.01269960e+00 -6.11115620e-02 -9.62365698e-03 -2.50011444e-01
3.43262285e-01 4.70147043e-01 -5.52197993e-01 -2.26666346e-01
-9.22428519e-02 4.19487029e-01 2.06110865e-01 3.06568146e-01
-7.42794931e-01 7.00896323e-01 6.95379686e+00 7.00876713e-01
-1.11945939e+00 9.67441313e-03 4.76866663e-01 -2.42212713e-01
-6.78897381e-01 -1.31921330e-02 -8.84322941e-01 3.04605961e-01
1.01657081e+00 -6.48655474e-01 4.92964238e-01 7.65328228e-01
-2.76907295e-01 1.12728886e-01 -1.23389161e+00 1.04027379e+00
2.35948801e-01 -1.50361860e+00 -3.71856168e-02 4.13478129e-02
4.05325532e-01 5.94328821e-01 1.71973109e-01 7.15711415e-01
5.04657090e-01 -1.15931988e+00 2.94721931e-01 3.90661448e-01
7.65702605e-01 -3.06923449e-01 5.21830320e-01 3.59891385e-01
-6.90458596e-01 8.97650644e-02 -4.48165625e-01 -2.11223532e-02
-7.29495585e-02 2.22068727e-01 -7.14205205e-01 4.02179182e-01
7.39287794e-01 6.66948617e-01 -7.33894706e-01 1.23179030e+00
2.61321515e-01 5.72978437e-01 -3.08765262e-01 -1.77380428e-01
2.51651585e-01 1.74573511e-01 4.76611972e-01 1.42206836e+00
1.01348631e-01 -2.94315666e-02 2.36584209e-02 6.50506675e-01
-3.75804752e-01 2.61881202e-01 -9.58068192e-01 -4.11249280e-01
7.19814301e-01 1.25364506e+00 -3.70948881e-01 -2.87534267e-01
-6.89683616e-01 4.67358142e-01 5.61562359e-01 6.42691672e-01
-3.65677178e-01 -5.19445539e-01 5.60624778e-01 4.90771174e-01
-1.16881818e-01 -1.32429495e-01 1.81056261e-01 -1.43122661e+00
-3.40257511e-02 -1.42213011e+00 8.03739727e-01 -7.27264762e-01
-1.49832833e+00 5.93989730e-01 3.54390115e-01 -8.05914819e-01
-4.08824295e-01 -3.73972267e-01 -3.31192076e-01 9.07068789e-01
-1.62459886e+00 -6.73276603e-01 -1.56222999e-01 4.44387466e-01
5.72835088e-01 -2.57541806e-01 1.07406282e+00 4.63393331e-01
-3.12035084e-01 5.62141776e-01 1.77682355e-01 2.30803534e-01
1.07243526e+00 -1.08762264e+00 4.41837192e-01 4.03436869e-01
5.41268051e-01 1.18861675e+00 6.53165519e-01 -3.62347066e-01
-1.50551033e+00 -6.43115461e-01 1.01582754e+00 -7.20304489e-01
8.51437569e-01 -2.91098624e-01 -1.13202369e+00 7.38720775e-01
4.67745960e-01 -2.19783232e-01 8.92373979e-01 7.12629139e-01
-7.03853071e-01 6.05876856e-02 -8.85535181e-01 5.30182838e-01
7.59146154e-01 -7.59977698e-01 -9.60368931e-01 5.04552484e-01
6.75307751e-01 -2.16195479e-01 -1.00350153e+00 3.70621473e-01
5.62427640e-01 -6.65545881e-01 9.97064173e-01 -9.92215574e-01
5.08434832e-01 1.92130402e-01 -2.27103308e-01 -1.12184882e+00
-1.79546595e-01 -4.79750812e-01 -2.92092353e-01 1.13616300e+00
4.64504957e-01 -7.15891540e-01 5.28644741e-01 9.35264051e-01
1.03896677e-01 -9.47684348e-01 -7.32511282e-01 -5.91497242e-01
4.98477817e-01 -4.20770288e-01 2.78128564e-01 1.15530312e+00
1.47082776e-01 5.28714359e-01 -1.26822829e-01 -5.65965950e-01
1.64764598e-01 8.38639364e-02 6.88213110e-01 -1.47975969e+00
-4.16283935e-01 -5.60458660e-01 -2.06180409e-01 -1.17368591e+00
7.06845969e-02 -1.00040126e+00 -3.33392650e-01 -1.80103171e+00
4.60143089e-01 -6.89889312e-01 -5.20160437e-01 4.95861113e-01
-1.88691225e-02 9.18640867e-02 1.01291522e-01 5.83307087e-01
-8.55799854e-01 1.83372408e-01 7.16502964e-01 -1.83684826e-01
-8.24239850e-02 -1.90965652e-01 -1.20768499e+00 4.03240502e-01
6.42915964e-01 -3.68156075e-01 -4.66106206e-01 -7.72607923e-01
5.34777761e-01 9.85263195e-03 2.75705814e-01 -6.97508991e-01
4.00278836e-01 3.12968865e-02 1.07252173e-01 -3.61935616e-01
3.18671316e-01 -5.45403421e-01 -1.10928178e-01 9.42306221e-02
-7.96271086e-01 3.14831585e-01 1.77799344e-01 5.31532168e-01
-4.20528114e-01 -4.05002832e-01 5.22163451e-01 -4.47136194e-01
-6.39144599e-01 2.12341875e-01 -3.12055558e-01 6.45589948e-01
5.05762577e-01 7.83942826e-03 -6.79389477e-01 -7.47044981e-01
-2.69650459e-01 3.57380271e-01 6.41564429e-01 8.36163282e-01
5.77378213e-01 -9.39033389e-01 -8.21705818e-01 -2.14522511e-01
4.93452370e-01 -2.09028363e-01 -4.06734534e-02 5.83700478e-01
-2.18465537e-01 7.75219142e-01 1.05282955e-01 -1.82210505e-01
-1.09026921e+00 5.82971454e-01 1.34585679e-01 -5.92629254e-01
-5.72476387e-01 5.81707537e-01 -7.32426196e-02 -2.43502632e-01
5.54360330e-01 6.62665889e-02 -3.78984451e-01 3.63295555e-01
8.51565063e-01 2.49553859e-01 2.59857297e-01 -3.55847985e-01
-3.11679602e-01 9.84682962e-02 -6.33193493e-01 -2.83577323e-01
1.45730174e+00 2.79761821e-01 9.47073102e-02 6.35054827e-01
1.32704937e+00 -1.32085115e-01 -5.34713566e-01 -5.41629016e-01
2.14877173e-01 -3.10582906e-01 1.99051246e-01 -1.07300365e+00
-7.04981148e-01 8.46054912e-01 4.63362157e-01 4.16702539e-01
9.44454193e-01 2.51134574e-01 7.15079606e-01 7.07783997e-01
4.08525974e-01 -1.07153118e+00 8.63275677e-02 7.31164753e-01
9.17934120e-01 -1.28753304e+00 7.36410692e-02 3.17589372e-01
-4.82922345e-01 8.36012423e-01 3.14760357e-01 -2.88242269e-02
6.00450814e-01 -8.99369363e-03 1.43968239e-01 -4.98219818e-01
-1.10374117e+00 -3.29920858e-01 3.77403438e-01 4.72757518e-01
8.58927965e-01 -3.96851957e-01 -5.60688555e-01 2.61797249e-01
-4.43276996e-03 4.01178468e-03 3.40338975e-01 1.05509996e+00
-3.46287608e-01 -1.07979774e+00 -1.38875887e-01 7.64594316e-01
-8.56523871e-01 -3.63162756e-01 -6.41454816e-01 9.81551051e-01
-5.14464319e-01 9.45195138e-01 2.30838060e-02 -2.06769422e-01
3.44418585e-01 2.30835855e-01 2.63497889e-01 -9.73893285e-01
-8.29356909e-01 -1.11382119e-01 2.34609753e-01 -4.17684019e-01
-1.71080723e-01 -4.87684548e-01 -4.62944716e-01 -1.57764554e-01
-3.05487275e-01 4.95167106e-01 6.49799287e-01 7.06172943e-01
6.16250217e-01 1.31147340e-01 1.22152627e-01 -5.62013328e-01
-9.87020075e-01 -1.06223357e+00 -3.32912683e-01 5.21008253e-01
2.10962623e-01 -5.79778135e-01 -5.03013313e-01 -4.51185435e-01] | [11.372443199157715, 7.737179279327393] |
fa47c2bf-476a-4dd2-97e0-b4e10dbbc907 | todd-topological-compound-fingerprinting-in | 2211.03808 | null | https://arxiv.org/abs/2211.03808v1 | https://arxiv.org/pdf/2211.03808v1.pdf | ToDD: Topological Compound Fingerprinting in Computer-Aided Drug Discovery | In computer-aided drug discovery (CADD), virtual screening (VS) is used for identifying the drug candidates that are most likely to bind to a molecular target in a large library of compounds. Most VS methods to date have focused on using canonical compound representations (e.g., SMILES strings, Morgan fingerprints) or generating alternative fingerprints of the compounds by training progressively more complex variational autoencoders (VAEs) and graph neural networks (GNNs). Although VAEs and GNNs led to significant improvements in VS performance, these methods suffer from reduced performance when scaling to large virtual compound datasets. The performance of these methods has shown only incremental improvements in the past few years. To address this problem, we developed a novel method using multiparameter persistence (MP) homology that produces topological fingerprints of the compounds as multidimensional vectors. Our primary contribution is framing the VS process as a new topology-based graph ranking problem by partitioning a compound into chemical substructures informed by the periodic properties of its atoms and extracting their persistent homology features at multiple resolution levels. We show that the margin loss fine-tuning of pretrained Triplet networks attains highly competitive results in differentiating between compounds in the embedding space and ranking their likelihood of becoming effective drug candidates. We further establish theoretical guarantees for the stability properties of our proposed MP signatures, and demonstrate that our models, enhanced by the MP signatures, outperform state-of-the-art methods on benchmark datasets by a wide and highly statistically significant margin (e.g., 93% gain for Cleves-Jain and 54% gain for DUD-E Diverse dataset). | ['Bulent Kiziltan', 'Yulia Gel', 'Yuzhou Chen', 'Ignacio Segovia-Dominguez', 'Baris Coskunuzer', 'Andac Demir'] | 2022-11-07 | null | null | null | null | ['graph-ranking'] | ['graphs'] | [ 3.33122581e-01 -8.65771174e-02 -5.64847767e-01 1.38767451e-01
-7.71578133e-01 -9.11109507e-01 6.96566880e-01 4.51552123e-01
-1.26321942e-01 1.01187956e+00 -4.24256921e-02 -5.84465802e-01
-4.71814573e-01 -9.10247445e-01 -8.44360471e-01 -8.41503978e-01
-4.91375327e-01 6.67973042e-01 1.68304920e-01 -2.85576284e-01
4.53378975e-01 1.10205925e+00 -1.03586400e+00 1.41705975e-01
8.02818358e-01 7.59984672e-01 -1.56849101e-01 4.18758661e-01
5.71499690e-02 1.63459271e-01 -2.95335054e-01 -3.38739157e-01
2.22443119e-01 -3.48641753e-01 -9.27975237e-01 -4.18452680e-01
5.57445467e-01 1.06309436e-01 -3.97095561e-01 9.49021339e-01
6.64073408e-01 9.29567367e-02 1.02901065e+00 -8.72984171e-01
-8.13312531e-01 2.73265362e-01 -3.65596384e-01 -7.31261820e-02
4.12026495e-01 -1.42991871e-01 1.29210711e+00 -1.18628168e+00
1.03882658e+00 1.09931493e+00 7.58166730e-01 4.64139283e-01
-1.81611383e+00 -5.27837276e-01 -3.02678287e-01 3.44192028e-01
-1.51549792e+00 -2.02605382e-01 7.74569809e-01 -7.42365241e-01
1.11391044e+00 3.07479650e-01 3.65841657e-01 1.20843577e+00
3.96927357e-01 3.64378244e-01 7.50552177e-01 -2.43797347e-01
3.93672645e-01 -1.81404210e-03 2.11676359e-01 8.45479786e-01
5.55421770e-01 2.23692074e-01 -3.19391876e-01 -7.11317658e-01
7.43693233e-01 -7.43013546e-02 -4.51541781e-01 -8.24861884e-01
-1.12449241e+00 1.24643135e+00 6.38453543e-01 3.24835151e-01
-4.82448280e-01 -3.62676680e-02 2.98826873e-01 9.37091112e-02
1.86608568e-01 1.01336730e+00 -4.66050029e-01 4.29385215e-01
-7.08416522e-01 2.47602388e-01 6.70864701e-01 4.78602678e-01
6.54539406e-01 -1.66263744e-01 -1.04603179e-01 5.30947387e-01
2.28309825e-01 1.64732207e-02 2.39814758e-01 -5.86260915e-01
1.30263269e-01 5.92188478e-01 -4.02804725e-02 -1.14239538e+00
-5.10440648e-01 -5.43548703e-01 -8.91148746e-01 1.35425791e-01
2.67596155e-01 1.29198864e-01 -9.59475458e-01 1.87436473e+00
4.04610664e-01 2.22700357e-01 9.62636247e-02 4.63192791e-01
7.62039661e-01 6.30243421e-01 2.15731308e-01 -3.28706324e-01
9.95676339e-01 -4.17559206e-01 -2.00169966e-01 5.49954355e-01
7.44454205e-01 -4.50562745e-01 6.62521064e-01 2.99938262e-01
-8.29650581e-01 -3.58676463e-01 -1.33420610e+00 2.45080799e-01
-6.60640240e-01 -1.82692096e-01 7.08270192e-01 5.66322982e-01
-9.18308258e-01 1.43419862e+00 -7.38365889e-01 -1.09191097e-01
6.10628366e-01 8.98487508e-01 -5.87779045e-01 2.29636520e-01
-1.26208568e+00 7.79241383e-01 4.96406615e-01 -1.89990267e-01
-8.08974385e-01 -9.82440770e-01 -6.84802115e-01 -4.59733792e-03
1.55383945e-01 -6.70190811e-01 4.44322228e-01 -5.88908970e-01
-1.51049912e+00 3.81541103e-01 1.19652763e-01 -5.15196919e-01
2.89651871e-01 3.10586929e-01 -4.01961237e-01 2.88103074e-01
1.11160176e-02 6.35522127e-01 3.54535937e-01 -1.06047130e+00
5.04230196e-03 -5.15831053e-01 -2.45977536e-01 -1.05612211e-01
-4.07517612e-01 -3.31867933e-01 -1.95056558e-01 -4.64801162e-01
-6.76974980e-03 -1.16096258e+00 -3.75511587e-01 -1.35957524e-01
-8.49064767e-01 -3.33148360e-01 6.24267280e-01 -3.59160125e-01
1.14355433e+00 -1.67932427e+00 6.27233386e-01 6.56171262e-01
5.72788835e-01 4.65445101e-01 -4.16725814e-01 6.60687804e-01
-4.81947362e-01 4.02582496e-01 -1.40767202e-01 3.40624243e-01
-2.14691624e-01 -1.30176619e-01 -2.01261342e-01 5.92531741e-01
2.94577062e-01 8.95991802e-01 -8.83102536e-01 -3.65512878e-01
6.65690005e-02 8.63903046e-01 -6.18042290e-01 -1.44323990e-01
-3.39364290e-01 3.00655723e-01 -7.77805209e-01 6.47801876e-01
6.30970597e-01 -7.98181474e-01 5.65931439e-01 -3.87246430e-01
7.37777576e-02 2.10901946e-02 -9.02882755e-01 1.28274989e+00
2.56867092e-02 3.35544050e-01 -6.00819647e-01 -1.08458626e+00
7.08679020e-01 2.08029345e-01 7.41123080e-01 -5.49976945e-01
-9.15292948e-02 4.75864559e-01 1.60515621e-01 3.84562439e-03
1.93456516e-01 -2.63627898e-03 3.00558805e-01 -1.01925597e-01
8.68935138e-02 3.75391543e-01 2.18888402e-01 1.38695687e-01
1.08445001e+00 8.93539041e-02 2.47946471e-01 -2.90966183e-01
7.12221026e-01 1.59761623e-01 3.07108164e-01 4.74196911e-01
6.72870204e-02 3.22933167e-01 6.48984134e-01 -5.93466938e-01
-1.17504859e+00 -1.01799643e+00 -6.44920230e-01 6.57950640e-01
1.25410140e-01 -4.41719204e-01 -5.65087378e-01 -6.71420157e-01
2.81314224e-01 3.64109159e-01 -8.21112633e-01 -2.98952609e-01
-4.49500889e-01 -1.00768363e+00 5.88058591e-01 2.80629128e-01
-2.61908360e-02 -7.14877784e-01 -5.44621162e-02 5.41992486e-01
4.29569274e-01 -7.85175860e-01 -2.37540245e-01 3.22749317e-01
-9.14254606e-01 -1.39495480e+00 -9.58151996e-01 -7.13348806e-01
5.10433197e-01 -5.29323472e-03 8.86737943e-01 -1.63182870e-01
-3.69602084e-01 -2.16065302e-01 -6.68058023e-02 -9.62191001e-02
-5.03905118e-01 2.46899147e-02 3.74788016e-01 -8.51865262e-02
1.19019769e-01 -6.85149848e-01 -8.12692285e-01 3.84530842e-01
-6.94238722e-01 -3.13176781e-01 6.53285325e-01 1.21027792e+00
1.08476567e+00 -1.46975249e-01 6.02015436e-01 -9.09851611e-01
6.88963711e-01 -5.22374094e-01 -6.44597828e-01 2.89542317e-01
-9.67438221e-01 5.02508938e-01 9.80217695e-01 -4.73803133e-01
-2.37502202e-01 2.84406126e-01 -1.08664937e-01 -4.81049150e-01
1.72779366e-01 7.09016919e-01 9.65844020e-02 -7.33993113e-01
9.82179046e-01 1.17014408e-01 3.21473082e-04 -4.16097313e-01
2.32051119e-01 2.90457487e-01 2.00506106e-01 -5.65567553e-01
7.33191848e-01 1.54375896e-01 8.16696346e-01 -8.92272949e-01
-2.18251288e-01 -3.50763142e-01 -5.05480349e-01 3.59391332e-01
7.82096446e-01 -6.17089093e-01 -1.07074082e+00 -7.46887177e-02
-1.03622019e+00 -2.93859933e-02 2.18720168e-01 4.03871596e-01
-4.41638559e-01 7.67560899e-01 -4.76095378e-01 -3.16227674e-01
-4.17656749e-01 -1.45745778e+00 9.07495022e-01 -1.22171462e-01
-2.49407247e-01 -9.35430408e-01 6.19924307e-01 -4.71690893e-02
1.58027574e-01 1.06899595e+00 1.58041930e+00 -9.27657247e-01
-6.58789635e-01 -3.23033720e-01 -1.35473102e-01 -8.64193216e-02
8.36898088e-02 2.53616005e-01 -6.58454657e-01 -4.60685581e-01
-7.59063900e-01 -1.07950523e-01 9.14761901e-01 5.90941548e-01
1.15030873e+00 -2.83799350e-01 -8.56863976e-01 7.42537260e-01
1.69951630e+00 5.28775990e-01 5.71601868e-01 2.89134681e-01
8.34610522e-01 2.82889724e-01 -2.21727546e-02 1.52710259e-01
-2.34408513e-01 9.95463848e-01 4.56873715e-01 -1.75240293e-01
7.43855238e-02 -2.34080255e-01 1.59004629e-01 3.43689233e-01
-3.51451486e-01 -2.88817137e-01 -8.58857810e-01 2.66881943e-01
-1.65309155e+00 -9.86984134e-01 -1.38331115e-01 2.42906189e+00
8.94128263e-01 6.60553649e-02 3.99235874e-01 6.30476931e-03
7.08262086e-01 -5.40681854e-02 -9.49373782e-01 -3.35672528e-01
-2.59281963e-01 6.79563999e-01 5.66547811e-01 4.87806261e-01
-1.02445507e+00 7.47319639e-01 6.42321873e+00 9.14002657e-01
-1.20123816e+00 -3.98551226e-01 7.76493907e-01 2.26441503e-01
-3.24111015e-01 -9.95097458e-02 -6.28400564e-01 2.52725810e-01
1.21022379e+00 -1.83049753e-01 2.67380804e-01 8.35272133e-01
-5.67518063e-02 5.17882645e-01 -1.29583561e+00 8.82135689e-01
-3.10771346e-01 -2.14679742e+00 4.96176034e-01 4.79309648e-01
7.55920351e-01 1.59147233e-01 3.28968495e-01 -1.16504490e-01
1.10920094e-01 -1.46686387e+00 1.44446701e-01 3.11814755e-01
1.11091483e+00 -9.08293009e-01 4.35335308e-01 -2.08934903e-01
-1.04828274e+00 1.09926783e-01 -5.07818103e-01 5.54434419e-01
-3.02690655e-01 2.99397320e-01 -9.85247016e-01 8.08414578e-01
7.90584981e-02 7.36562133e-01 -5.60440361e-01 1.04924822e+00
3.73585939e-01 3.26789796e-01 -2.15557218e-01 -3.74334484e-01
4.70978171e-01 -3.96985531e-01 5.72610021e-01 9.74068403e-01
2.10530445e-01 -6.41454458e-02 1.00610487e-01 9.01076615e-01
-3.04261088e-01 3.88987958e-01 -7.19830453e-01 -5.11096239e-01
3.66993040e-01 9.78001535e-01 -5.71197867e-01 -1.66085005e-01
-5.41060343e-02 8.00805986e-01 2.11743146e-01 4.39473540e-01
-6.80848837e-01 -7.43050992e-01 7.68473685e-01 1.34766340e-01
4.07822967e-01 -9.24742892e-02 3.15603875e-02 -8.45645487e-01
-2.72841603e-01 -8.85333121e-01 3.14064950e-01 -2.09231347e-01
-1.24184346e+00 8.81621778e-01 -2.66608149e-01 -1.23553753e+00
-7.21213073e-02 -1.13417530e+00 -4.25192982e-01 9.03038561e-01
-1.36328232e+00 -7.73497403e-01 2.12052286e-01 5.09109557e-01
-6.86045503e-03 -2.96256781e-01 1.21576607e+00 2.51894474e-01
-6.18355095e-01 7.51997471e-01 7.19987512e-01 -1.41361639e-01
4.42524910e-01 -1.30699623e+00 3.22154403e-01 1.68844059e-01
3.31211060e-01 7.90587366e-01 6.98110640e-01 -6.72063649e-01
-1.60971010e+00 -9.49765444e-01 5.75281262e-01 -5.80040514e-01
7.42667079e-01 -2.01179877e-01 -8.91765296e-01 2.31367946e-01
-3.02680105e-01 -5.19047938e-02 9.47629035e-01 3.37217748e-02
-5.29579222e-01 3.31581682e-01 -1.06754875e+00 6.28722489e-01
9.88539338e-01 -5.41618884e-01 -1.42562121e-01 7.46877491e-01
6.48793578e-01 -2.22255304e-01 -1.25161707e+00 5.70735991e-01
6.22617126e-01 -7.43744254e-01 1.39365339e+00 -1.20583355e+00
4.74293709e-01 -3.83394778e-01 -8.11555535e-02 -1.08995175e+00
-6.25029743e-01 -6.61099255e-01 -1.19504951e-01 3.08536053e-01
7.96552956e-01 -6.06871009e-01 1.09010994e+00 1.79099739e-01
4.99002170e-03 -1.27629542e+00 -1.00045276e+00 -8.63711894e-01
4.29583222e-01 2.36381039e-01 5.57614028e-01 1.08143628e+00
9.97170806e-03 5.20046115e-01 -1.77738279e-01 7.60283768e-02
6.70318305e-01 1.67294428e-01 3.26603204e-01 -1.68420410e+00
-4.20750022e-01 -8.20788443e-01 -7.99948275e-01 -6.65195704e-01
1.45637542e-01 -1.23319900e+00 -6.28214717e-01 -1.30292380e+00
3.33028913e-01 -2.67358333e-01 -5.53893209e-01 3.53449821e-01
7.19113648e-02 1.37540415e-01 -2.74613053e-01 1.96908429e-01
-3.74985874e-01 4.71304387e-01 1.16915071e+00 -4.79602188e-01
-5.36391497e-01 -1.81621686e-01 -6.77284181e-01 2.85610229e-01
5.94449162e-01 -2.86941111e-01 -2.54923970e-01 2.42550388e-01
3.53398114e-01 1.66442975e-01 3.21610659e-01 -8.12290490e-01
-3.37821171e-02 -1.54836625e-01 7.88549066e-01 -3.94895881e-01
2.32291028e-01 -4.29408997e-01 6.10772789e-01 7.61853278e-01
-3.02567273e-01 -9.87002105e-02 3.29263538e-01 9.34956074e-01
-1.16868742e-01 1.40787154e-01 7.71040022e-01 1.40673414e-01
-3.12058359e-01 8.16619217e-01 -6.13521151e-02 -3.35599601e-01
8.45407426e-01 -4.97504473e-01 -3.37288886e-01 1.50233544e-02
-7.29109228e-01 -1.28692821e-01 5.20731449e-01 1.52926117e-01
6.16899669e-01 -1.42972791e+00 -5.52002847e-01 -4.51772511e-02
2.23830149e-01 -4.08410758e-01 -5.68804815e-02 6.86855078e-01
-6.48484766e-01 7.63011515e-01 -2.62207389e-01 -4.35400099e-01
-1.34212351e+00 7.21201777e-01 5.25276065e-01 -3.54365885e-01
-3.66490781e-01 8.05705249e-01 1.59331918e-01 -1.21100657e-01
3.53133716e-02 1.61146596e-02 -2.70862699e-01 6.62880763e-02
1.90036774e-01 2.75297016e-01 2.39261955e-01 -7.44439602e-01
-6.00740135e-01 9.50755179e-01 -2.77256697e-01 4.52476919e-01
1.63254631e+00 7.41432965e-01 6.94303140e-02 -2.35488772e-01
1.64976478e+00 -7.46011138e-02 -8.69399667e-01 -2.94028614e-02
9.06939805e-02 -5.97214252e-02 1.66890547e-01 -7.40252674e-01
-7.42256224e-01 6.14157140e-01 8.02464783e-01 1.56485930e-01
6.33839130e-01 1.75589900e-02 6.54482245e-01 7.47000337e-01
2.51762182e-01 -7.23617435e-01 2.44084641e-01 3.01335454e-01
8.50253820e-01 -9.94898975e-01 1.66283026e-01 -3.06791753e-01
-3.21084291e-01 1.38792324e+00 -4.25065821e-03 -1.26663983e-01
5.55073738e-01 -5.41926205e-01 -4.75454509e-01 -6.48637056e-01
-4.07929957e-01 1.14256665e-01 7.05495179e-01 4.40502137e-01
4.72928554e-01 1.66539103e-01 -3.67903918e-01 2.10120946e-01
1.19772553e-01 -3.46220940e-01 2.85540074e-01 6.14798725e-01
-3.17458510e-01 -1.55483186e+00 -1.42106488e-01 4.38598633e-01
-4.59522456e-01 -3.64593446e-01 -8.30661237e-01 8.09271514e-01
-1.92101508e-01 4.94554669e-01 -4.71612722e-01 -6.07128859e-01
3.32534432e-01 5.31391650e-02 5.82821548e-01 -2.52806634e-01
-2.91489750e-01 2.10044198e-02 7.82392919e-02 -5.04154682e-01
-1.57507524e-01 -3.13010246e-01 -1.13115740e+00 -2.55872965e-01
-5.66205919e-01 3.64006430e-01 5.31769335e-01 5.08880436e-01
8.39999139e-01 3.39750975e-01 8.50276470e-01 -8.35865021e-01
-7.05188632e-01 -3.91103804e-01 -4.95399982e-01 3.99813801e-01
2.67753124e-01 -8.95260870e-01 -2.38279045e-01 -3.22553188e-01] | [5.182130813598633, 5.799956798553467] |
2fc00516-9246-462a-b97e-e2124e20e81c | tractable-probabilistic-graph-representation | 2305.10544 | null | https://arxiv.org/abs/2305.10544v1 | https://arxiv.org/pdf/2305.10544v1.pdf | Tractable Probabilistic Graph Representation Learning with Graph-Induced Sum-Product Networks | We introduce Graph-Induced Sum-Product Networks (GSPNs), a new probabilistic framework for graph representation learning that can tractably answer probabilistic queries. Inspired by the computational trees induced by vertices in the context of message-passing neural networks, we build hierarchies of sum-product networks (SPNs) where the parameters of a parent SPN are learnable transformations of the a-posterior mixing probabilities of its children's sum units. Due to weight sharing and the tree-shaped computation graphs of GSPNs, we obtain the efficiency and efficacy of deep graph networks with the additional advantages of a purely probabilistic model. We show the model's competitiveness on scarce supervision scenarios, handling missing data, and graph classification in comparison to popular neural models. We complement the experiments with qualitative analyses on hyper-parameters and the model's ability to answer probabilistic queries. | ['Mathias Niepert', 'Federico Errica'] | 2023-05-17 | null | null | null | null | ['graph-classification', 'graph-representation-learning'] | ['graphs', 'methodology'] | [ 1.89513370e-01 9.39753890e-01 -1.87931895e-01 -3.26697916e-01
-5.78818142e-01 -4.20631588e-01 7.36946046e-01 3.96270841e-01
-1.46082953e-01 6.22915447e-01 2.36125588e-01 -5.35863400e-01
-5.84911108e-01 -1.31065667e+00 -1.11215532e+00 -4.67456698e-01
-5.24778068e-01 1.01300609e+00 4.29484397e-01 2.27783471e-02
-1.88680708e-01 5.38630247e-01 -1.53964126e+00 2.15109333e-01
4.02377337e-01 9.17022526e-01 -1.69877574e-01 8.77882838e-01
-2.61892110e-01 9.47478294e-01 -2.32505113e-01 -8.83413732e-01
4.07687500e-02 5.86464331e-02 -9.65283632e-01 -3.07997167e-01
4.27101910e-01 -1.37911126e-01 -1.02000606e+00 1.10224640e+00
6.72589689e-02 1.78614482e-01 7.60659575e-01 -1.70885849e+00
-9.46434259e-01 1.38362086e+00 -2.63687760e-01 1.15749888e-01
1.78557277e-01 -2.22447757e-02 1.79654527e+00 -2.88221866e-01
5.59375346e-01 1.61318290e+00 7.17053235e-01 2.74033904e-01
-1.66130543e+00 -2.76353538e-01 1.42703488e-01 9.10244361e-02
-1.37846982e+00 -3.36776488e-02 5.98211646e-01 -1.85244113e-01
1.12527788e+00 8.99339467e-02 5.45493305e-01 1.00038159e+00
2.74575412e-01 8.50159466e-01 4.62053448e-01 -2.60629565e-01
2.82794446e-01 -5.88527210e-02 3.50385606e-01 1.17172909e+00
4.87097293e-01 2.07821041e-01 -8.31882000e-01 -5.65137386e-01
7.11120963e-01 2.74173748e-02 -6.89270645e-02 -8.77505958e-01
-8.14478934e-01 9.49975491e-01 6.02321386e-01 1.05783165e-01
-1.53970078e-01 9.19371963e-01 2.16521114e-01 1.13839753e-01
3.11223000e-01 1.84934437e-01 -6.38250053e-01 3.54840845e-01
-5.36022604e-01 3.37772012e-01 1.28086698e+00 1.21936381e+00
9.64691520e-01 -1.81030020e-01 -4.76865888e-01 4.11035746e-01
6.10611856e-01 4.11785096e-01 -1.19054243e-01 -1.04820132e+00
2.47588784e-01 5.66479921e-01 -2.00371444e-01 -9.52574253e-01
-4.47872549e-01 -5.62511683e-01 -7.30290949e-01 -2.89503813e-01
5.15945017e-01 1.53061245e-02 -9.27104890e-01 2.01633668e+00
8.65593478e-02 5.15348613e-02 -1.31420076e-01 2.43627742e-01
6.08670533e-01 4.48876441e-01 2.22090319e-01 1.75065860e-01
1.28135335e+00 -5.64815164e-01 -1.48355678e-01 -5.98403290e-02
8.17907393e-01 5.79816774e-02 9.19217169e-01 7.61873946e-02
-1.04936409e+00 -1.05222814e-01 -8.12083423e-01 -1.88622028e-01
-4.34233874e-01 -5.21019816e-01 1.12885439e+00 8.81236613e-01
-1.70605814e+00 9.30151403e-01 -9.98875558e-01 -1.66346237e-01
7.64147937e-01 5.60047388e-01 -2.02618852e-01 -2.24451959e-01
-1.32628977e+00 6.79157495e-01 6.32028461e-01 -1.57360911e-01
-1.11574852e+00 -6.81139827e-01 -1.07559359e+00 6.59809709e-01
4.00271624e-01 -1.11496902e+00 1.26828420e+00 -2.34508276e-01
-1.17794979e+00 6.05516076e-01 3.69167812e-02 -7.81154752e-01
1.30840078e-01 3.13370049e-01 1.70439914e-01 2.11969063e-01
-2.35467494e-01 8.47008526e-01 6.14715755e-01 -9.19543743e-01
-5.57159662e-01 -3.76329303e-01 4.10794318e-01 9.00956094e-02
-2.65201956e-01 -3.20181459e-01 -3.04769635e-01 -1.11273535e-01
1.61917090e-01 -7.99784362e-01 -3.61625969e-01 -1.57392807e-02
-7.05593348e-01 -6.66853368e-01 2.85419852e-01 -1.18713059e-01
8.23710740e-01 -1.94861817e+00 1.09493047e-01 5.80044210e-01
6.96188927e-01 -3.99951935e-01 -3.85907680e-01 5.45009553e-01
-8.38263892e-03 3.12624365e-01 -1.73422247e-01 -3.99017692e-01
5.82171023e-01 6.40980244e-01 -3.84215742e-01 2.78599530e-01
2.77769506e-01 1.08246958e+00 -9.53986168e-01 -6.03903651e-01
-1.62246436e-01 5.84577501e-01 -6.53521240e-01 2.48512700e-02
-7.42667437e-01 -3.26010168e-01 -2.99409837e-01 3.30018610e-01
3.12483251e-01 -7.27939606e-01 4.76247579e-01 -1.36073858e-01
6.54677093e-01 6.28123522e-01 -1.09679127e+00 1.39817464e+00
-2.96501994e-01 3.26013118e-01 -7.59762200e-03 -7.72555828e-01
6.26451969e-01 1.95869058e-01 1.39770940e-01 -2.57710993e-01
-2.23892797e-02 -9.82847661e-02 -6.19551130e-02 -9.55019966e-02
4.49554145e-01 1.19401887e-01 1.40451808e-02 7.84939647e-01
6.37535095e-01 -3.77927840e-01 3.99708718e-01 9.25290227e-01
1.43740916e+00 1.10231258e-03 -2.82688141e-01 -3.88665646e-01
1.00431077e-01 -4.51993227e-01 2.75160730e-01 1.40492761e+00
4.83938977e-02 3.06463689e-01 1.19959927e+00 -3.92455280e-01
-8.61715674e-01 -1.39834225e+00 1.28617764e-01 1.71816337e+00
-3.80761147e-01 -6.82538450e-01 -5.75295866e-01 -6.80354297e-01
2.13404074e-01 7.29201198e-01 -6.16956592e-01 -3.45303595e-01
-1.74172878e-01 -8.22469592e-01 7.47263610e-01 6.06265604e-01
-4.27313671e-02 -9.28626418e-01 -6.01204447e-02 3.24975960e-02
6.55163378e-02 -1.05166173e+00 -3.50966364e-01 4.84380275e-01
-9.54394996e-01 -9.94427919e-01 -2.03977019e-01 -7.04901814e-01
6.43675327e-01 -1.82810187e-01 1.50047159e+00 1.22995220e-01
1.99174210e-02 6.85227931e-01 2.45858908e-01 -2.94667155e-01
-5.07401764e-01 3.89102668e-01 -3.26577164e-02 -3.00445348e-01
1.65273294e-01 -1.25785649e+00 -3.16705823e-01 -9.21010673e-02
-1.07992029e+00 -5.44491373e-02 5.01097322e-01 6.40249908e-01
4.64866191e-01 2.98826247e-01 1.48162648e-01 -1.20594382e+00
6.39840722e-01 -5.56017637e-01 -9.79788780e-01 4.49111074e-01
-6.61354005e-01 7.84671843e-01 2.21145317e-01 -3.24584991e-01
-7.73375392e-01 3.92089859e-02 2.82197505e-01 -2.95703858e-01
7.13798553e-02 7.08608150e-01 4.12997697e-03 -8.98167267e-02
7.40539908e-01 1.50503358e-02 -1.03241034e-01 -1.27830356e-01
9.44468558e-01 -4.88905199e-02 6.44859970e-01 -1.06402588e+00
7.08991587e-01 4.01959419e-01 5.37871182e-01 -4.55444157e-01
-1.06949687e+00 8.01527426e-02 -2.33574390e-01 2.15845391e-01
7.03201890e-01 -7.26986825e-01 -1.37075925e+00 1.79720864e-01
-1.41911006e+00 -2.40354478e-01 -3.81408244e-01 2.04112619e-01
-5.39533198e-01 3.57258976e-01 -1.07009637e+00 -1.01626885e+00
-2.28462487e-01 -7.79000640e-01 8.81688595e-01 2.17820883e-01
-6.26192018e-02 -1.27237976e+00 3.89317498e-02 1.14433594e-01
4.60486591e-01 -2.01309592e-01 1.66170669e+00 -1.12373757e+00
-1.01874864e+00 -4.55037579e-02 -5.85011363e-01 1.92952752e-01
-5.74349105e-01 9.46865305e-02 -1.04537022e+00 -2.94492375e-02
-2.80277550e-01 -2.26250678e-01 1.04683495e+00 3.66676569e-01
1.10658038e+00 -5.98240852e-01 -4.56648916e-01 3.47606361e-01
1.28050005e+00 -4.97549593e-01 2.83069223e-01 -2.95843571e-01
8.76492560e-01 6.23552263e-01 -5.57238460e-01 2.57335454e-01
7.56099105e-01 -1.50524369e-02 7.78436065e-01 4.81498420e-01
4.55004387e-02 -8.63796651e-01 1.02996737e-01 4.76133317e-01
1.58663735e-01 -4.71297503e-01 -1.13710439e+00 6.50226712e-01
-1.89814532e+00 -5.77489316e-01 -7.50967339e-02 2.09155512e+00
6.83940232e-01 5.28552413e-01 -2.26581804e-02 -1.06706858e-01
7.24338055e-01 3.95073593e-01 -3.14481437e-01 -2.78592557e-01
-7.87095800e-02 5.87887228e-01 7.73665309e-01 6.62303150e-01
-8.16910148e-01 5.98435163e-01 7.24931479e+00 7.65205443e-01
-1.50282145e-01 1.54000297e-01 6.42800927e-01 2.21321825e-02
-9.15035903e-01 2.05293387e-01 -7.37777948e-01 1.00880854e-01
1.31957257e+00 -1.09416492e-01 7.63161480e-01 7.94422567e-01
-6.42479181e-01 8.39678422e-02 -1.67343950e+00 6.98336065e-01
-3.40088159e-01 -1.57343614e+00 1.78668618e-01 4.80292849e-02
6.78240120e-01 5.25180995e-01 6.00617789e-02 5.60955405e-01
1.41670442e+00 -1.20646083e+00 5.01629472e-01 4.39123273e-01
3.63547742e-01 -7.38161802e-01 5.95168114e-01 3.09069216e-01
-9.30145204e-01 -1.20437138e-01 -5.47838569e-01 -6.87201247e-02
-3.16752605e-02 7.95848012e-01 -1.01096046e+00 4.05026048e-01
5.50928771e-01 2.37633213e-01 -5.09331167e-01 6.70508862e-01
-5.51594198e-01 7.81898320e-01 -8.10160935e-01 -2.40319610e-01
1.14260085e-01 -2.20419213e-01 3.73443991e-01 1.07228565e+00
-1.33109372e-03 -2.19861060e-01 3.49000990e-02 1.26271403e+00
-6.54048324e-01 -3.70326906e-01 -7.46903181e-01 -3.51256222e-01
6.23111188e-01 1.11699760e+00 -6.03902459e-01 -3.72139990e-01
-1.68603629e-01 3.33119422e-01 8.41862798e-01 4.63801503e-01
-6.02392852e-01 -2.61801600e-01 3.82037759e-01 5.90070412e-02
5.92262626e-01 -5.64404316e-02 -2.24047512e-01 -7.26804018e-01
-2.38894485e-02 -4.30340916e-01 8.15663636e-01 -7.62493014e-01
-1.58047891e+00 3.93098682e-01 1.93112999e-01 -1.11632377e-01
-2.33998030e-01 -6.50779307e-01 -5.81635237e-01 8.27513874e-01
-1.11097240e+00 -1.41372037e+00 3.82261783e-01 4.94954616e-01
-4.64578390e-01 3.12362462e-01 9.30426478e-01 -1.78958088e-01
-2.99911886e-01 5.22999406e-01 -1.56094745e-01 1.54164463e-01
-1.30175307e-01 -1.53003144e+00 8.38556945e-01 5.48635483e-01
6.35631621e-01 5.67012131e-01 8.30497921e-01 -4.74396706e-01
-1.52514541e+00 -8.10216784e-01 1.10763311e+00 -7.81687319e-01
1.20179224e+00 -6.45013332e-01 -8.65021110e-01 1.04144931e+00
-4.70527224e-02 3.06757510e-01 3.96597594e-01 7.66699076e-01
-9.99460697e-01 -6.73682094e-02 -1.04300547e+00 4.72688168e-01
1.35982645e+00 -9.25798178e-01 -4.92146730e-01 7.02644408e-01
1.29097795e+00 -2.89598942e-01 -9.81713414e-01 2.17330888e-01
4.90431607e-01 -9.29540336e-01 9.28584814e-01 -9.13863957e-01
2.57821828e-01 1.29879639e-01 -2.10748821e-01 -1.03143668e+00
-3.25935245e-01 -7.95808613e-01 -4.83370125e-01 8.86997104e-01
7.32993186e-01 -7.59014249e-01 1.14657593e+00 9.45967197e-01
1.96280420e-01 -6.87256932e-01 -1.00581574e+00 -4.51260000e-01
7.71242194e-03 -6.76298797e-01 7.97667325e-01 4.68089968e-01
1.28061362e-02 5.25631726e-01 7.65866041e-02 4.83272761e-01
9.13973391e-01 -1.21248268e-01 3.24198097e-01 -1.55702543e+00
-8.11947107e-01 -5.17236173e-01 -5.49629748e-01 -8.51618886e-01
2.61311054e-01 -1.29310596e+00 3.38227302e-02 -1.54404271e+00
3.38494182e-01 -2.16719434e-01 -4.35117960e-01 6.44772947e-01
1.27469704e-01 -5.07674925e-02 4.44734059e-02 -1.43684506e-01
-9.84100699e-01 5.01648545e-01 8.57812464e-01 -1.21369340e-01
3.87073942e-02 1.82375446e-01 -8.20528746e-01 6.06645405e-01
4.97611642e-01 -8.91993403e-01 -6.51398838e-01 -5.29061675e-01
1.01760185e+00 2.19224557e-01 3.99839371e-01 -6.73964381e-01
5.08400917e-01 1.74159065e-01 9.23267193e-03 -5.27663767e-01
2.00191319e-01 -5.09283245e-01 1.62898943e-01 3.51220012e-01
-6.93837821e-01 -3.91823165e-02 -1.19162977e-01 9.83523905e-01
2.47589245e-01 -1.38066914e-02 3.63146126e-01 -1.38857082e-01
1.48313046e-01 6.42167568e-01 -3.57540846e-01 2.63385504e-01
2.76697755e-01 2.28987098e-01 -5.99197447e-01 -5.82160056e-01
-8.88194799e-01 4.14114058e-01 1.15241356e-01 6.94368184e-02
3.45434040e-01 -1.00004065e+00 -5.63777506e-01 -2.40710955e-02
3.06240446e-03 4.13408041e-01 1.48984507e-01 6.16287589e-01
-2.76986331e-01 4.04340297e-01 1.88789785e-01 -4.26024556e-01
-5.66391289e-01 6.44420743e-01 3.48056465e-01 -7.63017654e-01
-2.94479758e-01 1.16883600e+00 2.02547312e-01 -8.08457851e-01
4.54214245e-01 -6.08265460e-01 1.20360948e-01 -4.77575630e-01
5.98195158e-02 4.21035230e-01 9.04294662e-03 -1.84271056e-02
-1.47614583e-01 -3.53187025e-01 -2.47692868e-01 -2.56786317e-01
1.32229984e+00 2.35936090e-01 -5.42003512e-01 3.61451566e-01
9.51099277e-01 -4.50549603e-01 -1.09336293e+00 -5.37348151e-01
2.62773812e-01 1.57557160e-01 -1.23752110e-01 -5.18132031e-01
-9.66395080e-01 7.31834233e-01 2.63899025e-02 5.59046388e-01
4.04892147e-01 6.36849344e-01 3.86378139e-01 9.29712236e-01
4.27803576e-01 -7.60079920e-01 -2.25914896e-01 6.62301540e-01
4.38658267e-01 -7.45124876e-01 -9.72482115e-02 -4.40843880e-01
9.17906500e-03 8.45573485e-01 2.28252366e-01 -1.64675325e-01
8.95651281e-01 2.39616528e-01 -6.50372922e-01 -5.14579833e-01
-1.24324191e+00 -5.67571484e-02 2.14167804e-01 8.62953424e-01
5.88191077e-02 2.47298777e-01 2.59631157e-01 6.31397426e-01
-3.59898776e-01 -2.26836726e-01 4.31016862e-01 6.34940207e-01
-3.84177297e-01 -9.77737963e-01 8.70805085e-02 8.27386081e-01
-3.41018677e-01 -3.95859540e-01 -2.94282854e-01 6.80790544e-01
-3.70863169e-01 8.33142161e-01 1.16182275e-01 -3.75014901e-01
-2.05359831e-02 2.97450632e-01 6.45876288e-01 -7.86286473e-01
-2.76061445e-01 -7.06618965e-01 2.28119820e-01 -5.62443614e-01
-1.09214760e-01 -3.57689559e-01 -1.15133822e+00 -5.20271003e-01
-2.42764130e-01 2.12346137e-01 6.21176600e-01 9.53589618e-01
6.00113809e-01 3.37610275e-01 2.09865257e-01 -4.95371103e-01
-1.08662450e+00 -8.76063049e-01 -8.13674390e-01 1.36751607e-01
3.69680822e-02 -3.42449069e-01 -5.43863952e-01 -4.12099957e-01] | [7.269837856292725, 6.4087934494018555] |
e9244f5d-d4e5-441b-9ed7-0868fd4b7af5 | 2211-09238 | 2211.09238 | null | https://arxiv.org/abs/2211.09238v1 | https://arxiv.org/pdf/2211.09238v1.pdf | Learning unfolded networks with a cyclic group structure | Deep neural networks lack straightforward ways to incorporate domain knowledge and are notoriously considered black boxes. Prior works attempted to inject domain knowledge into architectures implicitly through data augmentation. Building on recent advances on equivariant neural networks, we propose networks that explicitly encode domain knowledge, specifically equivariance with respect to rotations. By using unfolded architectures, a rich framework that originated from sparse coding and has theoretical guarantees, we present interpretable networks with sparse activations. The equivariant unfolded networks compete favorably with baselines, with only a fraction of their parameters, as showcased on (rotated) MNIST and CIFAR-10. | ['Demba Ba', 'Emmanouil Theodosis'] | 2022-11-16 | null | null | null | null | ['rotated-mnist'] | ['computer-vision'] | [ 2.08732352e-01 8.64444554e-01 -4.11700219e-01 -7.00816095e-01
-1.73588887e-01 -7.66978800e-01 9.70901787e-01 -5.44947684e-01
-4.78435516e-01 7.39710689e-01 7.10145473e-01 -3.62169117e-01
-9.65761915e-02 -5.90409756e-01 -1.14813113e+00 -3.38986248e-01
-5.65587766e-02 4.54327404e-01 -5.16689122e-01 -6.97854757e-01
-3.17362636e-01 4.92492855e-01 -8.93861830e-01 4.72187489e-01
4.90502238e-01 5.04379809e-01 -3.77900332e-01 4.01542187e-01
7.10433424e-02 1.15423143e+00 -1.83199540e-01 -6.14263654e-01
3.69282126e-01 -2.07181662e-01 -1.03830945e+00 4.98219915e-02
7.99459696e-01 -5.71516275e-01 -9.90907729e-01 9.27759647e-01
-1.39164910e-01 2.33396217e-01 6.97410405e-01 -1.06424749e+00
-1.59062707e+00 1.05710042e+00 -1.68358743e-01 2.23930880e-01
-1.86232731e-01 8.04191530e-02 1.44169748e+00 -1.16003537e+00
8.24285269e-01 1.46529615e+00 6.78105295e-01 6.71729147e-01
-1.54531181e+00 -3.42775136e-01 5.43617606e-01 2.86877573e-01
-1.09146893e+00 -5.16469896e-01 1.00838673e+00 -3.07140380e-01
1.05438066e+00 5.38656600e-02 3.89150649e-01 1.73000741e+00
-1.54840052e-01 8.02043259e-01 7.59469867e-01 -2.79028416e-01
2.01609537e-01 -4.93058451e-02 3.77325505e-01 5.81522703e-01
3.96496445e-01 1.64270580e-01 -3.13114822e-01 1.61399245e-01
1.19570696e+00 1.44721806e-01 -1.57325044e-01 -7.20674574e-01
-1.49816477e+00 1.14442217e+00 1.03916025e+00 1.13643534e-01
-7.58976564e-02 4.21865195e-01 3.25079709e-01 3.87187421e-01
3.36619794e-01 9.04507220e-01 -6.21251941e-01 4.13208514e-01
-5.81328928e-01 2.04939559e-01 6.79820716e-01 1.01087904e+00
7.69062936e-01 8.49267602e-01 4.53169905e-02 6.31507456e-01
1.99460313e-01 5.68325222e-01 5.64132273e-01 -1.15040493e+00
3.29241067e-01 4.65371966e-01 -1.50092930e-01 -9.57395017e-01
-4.22531545e-01 -5.55793583e-01 -1.06023228e+00 1.19732767e-01
3.49915087e-01 -2.54779041e-01 -1.11191213e+00 2.18655729e+00
-4.06955481e-01 3.35836023e-01 5.01287043e-01 1.14018083e+00
8.42659414e-01 3.86862993e-01 -5.09467646e-02 4.79867965e-01
1.19671488e+00 -1.08813787e+00 -7.15472043e-01 -4.78753418e-01
3.73991340e-01 -2.20094100e-01 1.21275985e+00 2.86222667e-01
-1.09723413e+00 -4.68363374e-01 -1.28333926e+00 -4.83308643e-01
-4.28071827e-01 -1.22162640e-01 9.22587991e-01 3.84132564e-01
-1.35397661e+00 5.24376392e-01 -1.09445739e+00 -2.37898752e-01
5.34103274e-01 2.90950298e-01 -6.63078666e-01 8.75174329e-02
-1.31165469e+00 8.95210862e-01 4.80055600e-01 7.41710067e-02
-1.12749016e+00 -7.20718086e-01 -1.34900463e+00 1.29637986e-01
3.01720768e-01 -1.05644512e+00 1.45403910e+00 -1.46140897e+00
-1.67488945e+00 6.42771065e-01 1.25248998e-01 -9.52200651e-01
3.40428382e-01 -5.18851757e-01 -4.16678518e-01 2.15736598e-01
-1.83183759e-01 1.02600288e+00 9.76754427e-01 -1.14189863e+00
8.42005536e-02 -2.89742112e-01 4.83940512e-01 1.84281886e-01
-6.11437023e-01 -3.16611648e-01 9.08138603e-02 -1.17546618e+00
1.58839479e-01 -1.06007218e+00 -5.14213204e-01 5.82925323e-03
-4.27908808e-01 1.88942030e-02 1.07309067e+00 -5.28377295e-01
3.81935000e-01 -1.98586726e+00 4.47640598e-01 9.28804055e-02
4.87213969e-01 1.93946332e-01 -5.36775470e-01 -1.24019021e-02
-6.55746520e-01 2.70929188e-01 -2.59674489e-01 -2.12133229e-01
2.47854680e-01 6.87548101e-01 -9.24589455e-01 5.15883446e-01
7.08526969e-01 1.17070925e+00 -8.45585942e-01 1.24391600e-01
2.35838205e-01 7.81974554e-01 -9.97733235e-01 1.48661405e-01
-3.97109866e-01 3.22686225e-01 -1.43292800e-01 3.64150792e-01
4.48378325e-01 -5.86392343e-01 3.54895383e-01 -2.84598529e-01
2.83764780e-01 4.29552853e-01 -9.67343211e-01 2.06558609e+00
-3.96449208e-01 8.17968190e-01 1.05715677e-01 -1.45330083e+00
7.39142835e-01 3.76311094e-01 2.85985380e-01 -4.27335799e-01
1.20019831e-01 -1.38491824e-01 -6.96234927e-02 -1.50382921e-01
5.37695527e-01 -8.03209990e-02 2.93479431e-02 3.26093942e-01
7.11416841e-01 6.32464327e-03 -1.85774267e-01 4.86219019e-01
8.56010318e-01 3.33435744e-01 2.22071305e-01 -3.98583472e-01
2.06401184e-01 -8.73139277e-02 4.54533398e-01 5.89357257e-01
4.25306931e-02 8.85961294e-01 6.67155445e-01 -9.94012654e-01
-1.42028296e+00 -1.33328247e+00 -1.29158258e-01 1.19075859e+00
-1.59000129e-01 -2.13711172e-01 -7.62311518e-01 -7.17267454e-01
-4.75844555e-02 6.78589463e-01 -9.34149742e-01 -3.75865132e-01
-6.76765323e-01 -5.29553235e-01 7.25386262e-01 1.06117392e+00
4.67407346e-01 -8.96949410e-01 -1.12653472e-01 -1.31240822e-02
1.29565476e-02 -1.44965374e+00 -1.61800802e-01 5.50687909e-01
-1.12444973e+00 -8.19403231e-01 -6.90511525e-01 -6.89390421e-01
7.53485680e-01 9.54730883e-02 1.26779711e+00 -4.56821859e-01
1.52574420e-01 4.08320040e-01 -3.07775587e-01 -3.79026324e-01
-2.01179326e-01 2.64742672e-01 3.59706283e-01 -1.05807729e-01
6.35418743e-02 -9.01846290e-01 -4.76425916e-01 1.56786621e-01
-1.03776443e+00 1.54860858e-02 7.59999931e-01 1.14669204e+00
3.65532786e-01 -8.27762127e-01 7.05433071e-01 -1.18143690e+00
2.97375798e-01 -6.10560775e-01 -4.00064677e-01 -1.38297230e-01
-3.85926574e-01 6.74762309e-01 8.67382467e-01 -2.88948894e-01
-1.11064577e+00 1.85028166e-01 -1.47977382e-01 -8.87817323e-01
-3.68341893e-01 5.32648385e-01 -2.48347327e-01 -5.59977330e-02
1.11781263e+00 -2.10235387e-01 9.78313535e-02 -5.41872382e-01
1.16112483e+00 -1.19084036e-02 1.13192117e+00 -7.53529191e-01
9.69481587e-01 8.26087415e-01 -1.22534752e-01 -7.17090666e-01
-9.08694208e-01 1.40802085e-01 -8.26729417e-01 6.52651429e-01
7.50394583e-01 -1.30746222e+00 -3.62929821e-01 4.53562737e-02
-1.24334455e+00 -2.40016848e-01 -5.87397277e-01 5.01338422e-01
-5.85665762e-01 3.56538683e-01 -6.60954595e-01 -9.39383656e-02
-9.19476375e-02 -8.86268497e-01 7.02158988e-01 -1.25934094e-01
-3.60969663e-01 -1.28260005e+00 -1.31361902e-01 1.18167654e-01
6.16053760e-01 3.28379929e-01 9.14748728e-01 -1.00195253e+00
-4.80560064e-01 1.07341722e-01 -2.79207230e-01 5.44601202e-01
-9.48847830e-02 -2.35367551e-01 -1.29652095e+00 -2.67210871e-01
-4.75203879e-02 -5.17611206e-01 1.10669577e+00 3.03433299e-01
1.48638165e+00 -8.31682742e-01 1.42605066e-01 1.26711798e+00
1.04912710e+00 -3.57821375e-01 4.89482820e-01 3.48322809e-01
1.03532934e+00 3.06296080e-01 -2.40374967e-01 2.31350690e-01
5.29586673e-01 4.91078675e-01 9.16171849e-01 -4.80804116e-01
-2.10737571e-01 -3.73033375e-01 3.32497418e-01 7.04712391e-01
-4.12984520e-01 6.07774667e-02 -7.71084368e-01 6.38797164e-01
-1.80801928e+00 -8.43711913e-01 2.18796775e-01 1.47511601e+00
7.42655814e-01 5.70659377e-02 -4.20955606e-02 -1.12221137e-01
2.94713348e-01 4.70667750e-01 -8.11405241e-01 -5.13337135e-01
-5.57443321e-01 5.17360330e-01 5.70708096e-01 4.81519789e-01
-1.28801119e+00 1.12281656e+00 7.50983334e+00 1.66586101e-01
-1.17001367e+00 6.11797124e-02 5.31735837e-01 -8.09272453e-02
-6.96727931e-01 -1.57742519e-02 -3.24162424e-01 -1.92614883e-01
9.85566974e-01 -8.74429643e-02 4.93896812e-01 1.20740831e+00
-1.35550812e-01 9.71132576e-01 -1.53034115e+00 7.31126547e-01
1.20615043e-01 -1.60801208e+00 5.25331676e-01 -1.01295456e-01
1.05144489e+00 4.70585465e-01 5.75631738e-01 5.73090136e-01
1.13961530e+00 -1.45117939e+00 7.82599568e-01 2.95942396e-01
7.96535909e-01 -6.31064475e-01 5.19968808e-01 -1.24484077e-01
-7.04361796e-01 -4.74944785e-02 -5.82342505e-01 -3.79103601e-01
-1.32930651e-01 2.36041859e-01 -9.36448932e-01 3.50184411e-01
7.52795517e-01 1.13658774e+00 -5.30590594e-01 1.88078836e-01
-6.64663911e-01 8.15612733e-01 -2.28717163e-01 5.45057893e-01
5.45446038e-01 -1.06530748e-01 5.51255107e-01 1.11453509e+00
-2.05087233e-02 1.14414170e-01 -4.17592116e-02 1.31330538e+00
-6.09176755e-01 -3.37640971e-01 -1.01816356e+00 -6.68708161e-02
5.48386127e-02 1.17123461e+00 -2.97193050e-01 -3.42202812e-01
-6.81079447e-01 8.29531550e-01 6.03731573e-01 1.01722467e+00
-7.40004778e-01 -2.81975448e-01 1.08776319e+00 -9.78725925e-02
3.18491906e-01 -5.91078877e-01 -5.93558788e-01 -1.73343647e+00
-1.98860630e-01 -1.02754915e+00 3.63462687e-01 -6.85653865e-01
-1.34008598e+00 7.04628110e-01 -4.53072824e-02 -8.93837094e-01
-5.59747875e-01 -1.20999944e+00 -3.42968106e-01 6.52378678e-01
-1.74276996e+00 -1.45306194e+00 -1.81923091e-01 9.51035976e-01
4.30363715e-01 -4.57643658e-01 9.99534309e-01 1.15745209e-01
-4.49066043e-01 7.85802782e-01 1.39467791e-01 6.60795152e-01
6.74572110e-01 -1.46716452e+00 8.41747940e-01 8.57957244e-01
7.38906920e-01 1.04116821e+00 6.40591562e-01 -7.99770430e-02
-1.54815817e+00 -1.22217286e+00 2.50236034e-01 -7.82261908e-01
8.51233840e-01 -5.09399235e-01 -9.53843057e-01 1.49453366e+00
4.78537798e-01 5.33156991e-01 5.14838696e-01 4.60275620e-01
-9.80436563e-01 -5.53498184e-03 -7.03660786e-01 7.12153256e-01
1.19349861e+00 -7.63120234e-01 -1.07744515e+00 2.95843184e-01
1.25425386e+00 -4.94662672e-01 -6.30883396e-01 5.39481223e-01
3.22276413e-01 -6.52554452e-01 1.23633182e+00 -1.37272918e+00
5.37492454e-01 -2.56086830e-02 -3.66031468e-01 -1.49290705e+00
-5.18508732e-01 -5.64641774e-01 -3.19305897e-01 6.98339164e-01
4.04432088e-01 -5.56695461e-01 8.74687672e-01 5.80822885e-01
-4.08022493e-01 -1.29217789e-01 -5.29765427e-01 -8.36724102e-01
5.74564934e-01 -3.41948867e-01 5.82847178e-01 1.52413988e+00
-5.38332462e-02 5.37255883e-01 -5.03098488e-01 1.85857981e-01
4.18881565e-01 -2.19966337e-01 5.72265148e-01 -1.22941124e+00
-2.62048662e-01 -4.75460261e-01 -4.64975357e-01 -1.03092027e+00
7.66132176e-01 -1.33643019e+00 -3.65920722e-01 -1.16343522e+00
1.15692049e-01 9.31561068e-02 -5.85400701e-01 9.60123658e-01
7.23785087e-02 3.86888117e-01 2.31748372e-01 8.57809186e-02
-5.75149417e-01 6.56031907e-01 1.01898217e+00 -2.49117315e-01
7.18850046e-02 -5.66125989e-01 -1.18659127e+00 1.08133566e+00
7.49381542e-01 -1.14748813e-01 -6.48531020e-01 -1.22152829e+00
3.38972121e-01 -2.66646802e-01 7.59340644e-01 -8.46710563e-01
-9.56684873e-02 -1.58384994e-01 5.03520608e-01 2.45573483e-02
3.65271240e-01 -8.61349702e-01 -1.12279028e-01 1.40633658e-01
-6.77730262e-01 1.48285836e-01 2.92067826e-01 6.86241984e-01
-4.19368446e-01 1.91565752e-01 6.55444682e-01 -1.92005396e-01
-1.04586327e+00 4.36668992e-01 -1.70925111e-01 1.79758802e-01
7.72823393e-01 2.03102753e-01 -6.34793878e-01 -5.75811505e-01
-1.15163398e+00 7.05719441e-02 4.20904011e-01 6.31690979e-01
6.20249391e-01 -1.40948462e+00 -7.07804620e-01 2.38462687e-01
8.28644354e-03 2.66720742e-01 1.16422949e-02 3.52435231e-01
-4.44714278e-01 7.06842184e-01 -6.81817710e-01 -5.36767125e-01
-4.45731342e-01 8.37602377e-01 3.54155272e-01 -6.85229450e-02
-7.09915519e-01 7.24997699e-01 5.32696903e-01 -7.59066999e-01
1.43414691e-01 -6.91627800e-01 -1.66448474e-01 -2.77723193e-01
1.87499389e-01 -5.16384952e-02 1.79592520e-01 -5.93039751e-01
-2.80398160e-01 1.70646459e-01 -1.21558644e-01 -2.47905761e-01
1.64471149e+00 3.25893164e-01 1.14679679e-01 4.31907512e-02
1.23567748e+00 -3.85019600e-01 -1.49431288e+00 -6.50643349e-01
-2.88791656e-02 -1.44262716e-01 3.21617462e-02 -6.71575665e-01
-1.22418571e+00 1.25329936e+00 3.00465375e-01 -1.99540451e-01
6.95381165e-01 -1.38378710e-01 2.84117758e-01 1.11272025e+00
-7.65018091e-02 -6.47691965e-01 4.71017212e-01 1.03608990e+00
1.05278409e+00 -1.38112736e+00 -1.31930396e-01 1.05482060e-03
-8.02595556e-01 1.09982431e+00 7.88643599e-01 -5.74464142e-01
5.60807884e-01 2.51072198e-01 8.98698196e-02 -5.09666353e-02
-6.80015743e-01 9.71443485e-03 4.15320694e-01 8.14535558e-01
4.93209213e-01 2.14429423e-02 9.21583101e-02 9.53877270e-01
-4.70322728e-01 -2.06726953e-01 6.03027761e-01 5.85527718e-01
-1.90272734e-01 -9.16510105e-01 -2.49977022e-01 7.19154999e-03
-5.77643275e-01 -2.82172441e-01 -4.91948515e-01 9.30838585e-01
-3.31119359e-01 4.36467707e-01 1.87500060e-01 -1.38581842e-01
2.62734622e-01 3.32396440e-02 1.79591104e-01 -6.36194289e-01
-1.31335720e-01 -2.51245797e-01 3.82119603e-02 -6.36290133e-01
-3.96518230e-01 -4.09777880e-01 -1.30309069e+00 -2.05989286e-01
3.67158800e-01 -1.33793935e-01 5.46060860e-01 8.73109519e-01
4.25912559e-01 6.99187636e-01 3.47430348e-01 -8.74272585e-01
-8.05073202e-01 -8.63588214e-01 -3.41242164e-01 6.86180353e-01
7.08158195e-01 -4.90670413e-01 -1.66646153e-01 4.77182955e-01] | [9.389510154724121, 2.547260046005249] |
8c6a9cdf-99f9-409d-927c-984c709e76ca | density-aware-reinforcement-learning-to | 2306.08785 | null | https://arxiv.org/abs/2306.08785v1 | https://arxiv.org/pdf/2306.08785v1.pdf | Density-Aware Reinforcement Learning to Optimise Energy Efficiency in UAV-Assisted Networks | Unmanned aerial vehicles (UAVs) serving as aerial base stations can be deployed to provide wireless connectivity to mobile users, such as vehicles. However, the density of vehicles on roads often varies spatially and temporally primarily due to mobility and traffic situations in a geographical area, making it difficult to provide ubiquitous service. Moreover, as energy-constrained UAVs hover in the sky while serving mobile users, they may be faced with interference from nearby UAV cells or other access points sharing the same frequency band, thereby impacting the system's energy efficiency (EE). Recent multi-agent reinforcement learning (MARL) approaches applied to optimise the users' coverage worked well in reasonably even densities but might not perform as well in uneven users' distribution, i.e., in urban road networks with uneven concentration of vehicles. In this work, we propose a density-aware communication-enabled multi-agent decentralised double deep Q-network (DACEMAD-DDQN) approach that maximises the total system's EE by jointly optimising the trajectory of each UAV, the number of connected users, and the UAVs' energy consumption while keeping track of dense and uneven users' distribution. Our result outperforms state-of-the-art MARL approaches in terms of EE by as much as 65% - 85%. | ['Ivana Dusparic', 'Boris Galkin', 'Babatunji Omoniwa'] | 2023-06-14 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [-4.26083744e-01 2.48927116e-01 -2.22270817e-01 5.22290289e-01
-8.69926140e-02 -5.92507184e-01 3.18888336e-01 8.04720372e-02
-3.41989905e-01 1.29428387e+00 -5.77735186e-01 -5.28535664e-01
-7.21458733e-01 -1.21714067e+00 -5.01591444e-01 -1.10075974e+00
-7.15346515e-01 3.83792400e-01 1.31301865e-01 -4.63776588e-01
-3.64206314e-01 6.47592902e-01 -1.47196054e+00 -9.76641476e-01
1.12879753e+00 9.75342393e-01 3.22258353e-01 7.18605936e-01
4.39107209e-01 4.36251849e-01 -8.44019890e-01 -1.20904677e-01
3.03481758e-01 -2.23491848e-01 -4.66388553e-01 2.63207495e-01
-4.13744092e-01 -4.85031575e-01 -6.40425086e-01 7.20629334e-01
5.94191313e-01 3.75903368e-01 7.70583272e-01 -1.93392575e+00
-1.95724115e-01 8.35293904e-02 -8.81247044e-01 1.55065119e-01
-7.18628839e-02 2.64078081e-01 6.38011217e-01 3.50732170e-03
1.88874021e-01 6.50992215e-01 3.16527963e-01 3.58262181e-01
-7.06220150e-01 -4.87410307e-01 3.05111229e-01 3.07332903e-01
-1.49098194e+00 -4.91800755e-01 2.92886883e-01 6.83822334e-02
8.34209323e-01 1.60635993e-01 1.21308827e+00 5.28934658e-01
4.04321611e-01 4.71046746e-01 2.73650438e-01 -1.08313613e-01
5.71683049e-01 -1.06713712e-01 -1.02114260e+00 5.60170233e-01
4.33601022e-01 6.49188608e-02 1.90033227e-01 -7.34192953e-02
5.80514431e-01 -3.70842069e-01 -3.02135527e-01 -4.43912834e-01
-9.85436857e-01 9.25514400e-01 8.60262692e-01 1.68440297e-01
-1.17517924e+00 2.79876828e-01 2.87651476e-02 4.88398850e-01
3.38959992e-01 4.17229116e-01 -4.63353693e-01 -1.79967172e-02
-7.15313315e-01 1.84994668e-01 7.22385466e-01 1.35045576e+00
5.96142292e-01 2.42868423e-01 1.06836684e-01 6.14475667e-01
4.21377271e-01 8.38713109e-01 4.12028655e-02 -1.15460908e+00
1.47395536e-01 3.60244751e-01 5.12204409e-01 -7.23376870e-01
-5.37621737e-01 -6.84566140e-01 -1.35965490e+00 2.50543326e-01
9.52672511e-02 -1.19506240e+00 -6.38269603e-01 1.69284320e+00
5.67677438e-01 1.58414125e-01 3.92265022e-01 1.05281401e+00
1.33045346e-01 1.03552926e+00 5.38216047e-02 -7.33600736e-01
1.05865860e+00 -7.78140545e-01 -5.00262141e-01 -1.78938091e-01
3.72279227e-01 -2.14818463e-01 1.66538373e-01 -6.21933341e-02
-1.24810874e+00 -3.33996955e-04 -8.94141555e-01 9.55033422e-01
-4.04046208e-01 -1.44493252e-01 3.14415812e-01 6.78876042e-01
-1.37530696e+00 2.26056293e-01 -6.29781902e-01 -4.32360113e-01
6.42618775e-01 6.53429866e-01 2.47515604e-01 -1.07344985e-01
-1.26218283e+00 7.96104014e-01 1.48018384e-02 -1.18868545e-01
-1.15694869e+00 -5.15983820e-01 -5.55439115e-01 1.35330766e-01
6.73050940e-01 -9.75318432e-01 1.14821279e+00 -7.31357038e-01
-1.37823987e+00 -9.10267010e-02 2.29323506e-01 -5.85620999e-01
4.79585797e-01 4.41348344e-01 -3.91166717e-01 2.10570708e-01
-1.02387164e-02 8.98728013e-01 4.42635953e-01 -1.08728659e+00
-1.29633331e+00 -2.53911644e-01 5.48506081e-01 8.83087754e-01
-2.00938404e-01 -6.80130184e-01 -1.47699863e-01 6.61175977e-03
-7.59944260e-01 -1.11136389e+00 -5.69677591e-01 -2.67658494e-02
-2.67437965e-01 -3.79250944e-01 1.37938416e+00 -1.50961324e-03
9.69680965e-01 -1.82034290e+00 4.40966398e-01 1.48399860e-01
1.37299627e-01 3.15353215e-01 -2.54682124e-01 5.89257538e-01
8.70361805e-01 1.48893287e-02 -1.66691065e-01 -7.75994584e-02
-1.25462040e-01 4.10187155e-01 4.17703569e-01 5.35773635e-01
2.64458312e-03 4.96104628e-01 -1.26571620e+00 -2.50007302e-01
3.95758420e-01 5.50379753e-01 -2.06492856e-01 1.22822717e-01
-2.06541434e-01 3.18111867e-01 -6.34267986e-01 7.32352793e-01
6.21369421e-01 5.98194860e-02 1.00280024e-01 2.90446818e-01
-3.29174548e-01 -6.21303916e-01 -9.48947012e-01 1.17094886e+00
-7.91237712e-01 9.30152953e-01 7.17770398e-01 -1.05435908e+00
4.20423776e-01 4.59940702e-01 9.85196054e-01 -6.78445816e-01
1.51827082e-01 1.48443848e-01 6.29861951e-02 -3.23015809e-01
5.34456789e-01 9.60049704e-02 2.09032476e-01 -7.56506398e-02
-3.53351459e-02 -2.95124501e-01 1.75908059e-01 1.75812125e-01
1.29186666e+00 -5.43164730e-01 2.66598791e-01 -3.26082766e-01
2.30589718e-01 -5.14461324e-02 2.89957911e-01 3.55705112e-01
-4.46424335e-01 -4.28293467e-01 3.46675336e-01 -1.43680116e-02
-1.03928053e+00 -9.45101321e-01 -2.32780948e-02 6.05298340e-01
7.69210398e-01 2.28445515e-01 -7.07709551e-01 -3.25721413e-01
-1.52149471e-02 7.55331755e-01 -1.07111625e-01 -1.31102204e-01
2.02473551e-01 -9.51212049e-01 3.03900033e-01 -2.72501141e-01
9.25451934e-01 -4.89196777e-01 -1.07380211e+00 6.24389946e-01
-1.09048620e-01 -1.29168355e+00 -1.97079584e-01 1.13457322e-01
-2.69648671e-01 -1.19396651e+00 -1.03699934e+00 -4.80134189e-01
5.95291138e-01 6.82604551e-01 8.39208901e-01 3.43973637e-02
-6.49249479e-02 7.70324886e-01 -4.01742250e-01 -7.69755006e-01
-1.07537962e-01 4.81384963e-01 2.33830079e-01 -1.20663993e-01
-1.78667143e-01 -5.60274422e-01 -1.04832375e+00 6.10267639e-01
-6.97441936e-01 -2.49893591e-01 4.38346744e-01 4.34013814e-01
4.57874328e-01 1.12850034e+00 1.18057680e+00 1.13605587e-02
7.46460021e-01 -1.23450208e+00 -8.38508010e-01 7.67477751e-02
-4.61365819e-01 -4.99981403e-01 7.62894273e-01 -1.98262021e-01
-3.20106983e-01 -4.57403511e-02 -2.56974995e-02 -1.76958308e-01
-1.76012471e-01 1.50551498e-01 -1.66052833e-01 -4.70482349e-01
1.47701845e-01 -5.93159609e-02 7.38134235e-02 6.70171440e-01
1.92764759e-01 8.10840666e-01 2.81795394e-02 1.13965914e-01
8.13281476e-01 1.33560717e-01 5.94585776e-01 -1.37774181e+00
-1.22464530e-01 -2.66942203e-01 -9.35949981e-02 -8.13399434e-01
6.27003551e-01 -1.15170741e+00 -9.73195016e-01 5.48808277e-01
-8.69486272e-01 -7.16835797e-01 -1.94455281e-01 4.05600756e-01
-5.48878431e-01 -1.68601915e-01 -8.65833485e-04 -1.36987388e+00
-1.51270166e-01 -9.54092503e-01 8.05414021e-01 6.90482676e-01
3.73737901e-01 -1.18466544e+00 -2.23411560e-01 -3.09112012e-01
7.42541730e-01 6.28142834e-01 6.08311117e-01 2.03599334e-01
-7.89000809e-01 -2.95275338e-02 -7.75007531e-02 -7.50635490e-02
2.71088153e-01 -7.78357014e-02 -4.20709640e-01 -6.74057126e-01
-8.24015498e-01 -1.40171081e-01 1.89224198e-01 9.92630541e-01
8.46304655e-01 -6.05834842e-01 -7.26089358e-01 3.38766724e-01
1.69173372e+00 5.12049317e-01 5.76469421e-01 7.17132837e-02
1.84876278e-01 5.50547123e-01 7.13122427e-01 8.03582251e-01
5.88214874e-01 6.89512372e-01 1.52909338e+00 -2.39698336e-01
4.62952316e-01 1.60182804e-01 1.81858733e-01 3.83698009e-02
-2.62530774e-01 -1.26883316e+00 -4.44506973e-01 8.42640936e-01
-1.81590617e+00 -7.60198116e-01 6.70804158e-02 2.09851074e+00
2.16345303e-02 -2.30179369e-01 6.18929386e-01 6.77307025e-02
6.44812882e-01 4.55895998e-02 -9.39124584e-01 -2.78172761e-01
-5.39193451e-02 -5.19264877e-01 1.27637577e+00 3.88977557e-01
-7.27341533e-01 4.70969796e-01 4.67226553e+00 9.04345214e-01
-7.81970263e-01 3.74369882e-02 6.04999423e-01 -3.18227768e-01
-1.79983288e-01 -4.48645562e-01 -3.30872327e-01 6.45456493e-01
1.35174012e+00 -3.34878027e-01 9.20906007e-01 3.95403355e-01
7.11292863e-01 -5.80157280e-01 -3.31829667e-01 7.06158280e-01
-5.12621760e-01 -1.08308101e+00 -3.89950275e-01 7.77962923e-01
9.63264763e-01 3.92172247e-01 -1.24177188e-01 4.74475063e-02
4.68325585e-01 -6.98337257e-01 2.94886291e-01 4.04663384e-01
7.83283234e-01 -1.58817649e+00 8.32087278e-01 7.61877894e-01
-1.39133906e+00 -5.16076684e-01 -4.41882610e-01 8.97161439e-02
5.71379244e-01 6.50490522e-01 -5.85834444e-01 7.84037888e-01
6.00442052e-01 3.56279641e-01 1.88886121e-01 1.34916508e+00
1.19578436e-01 2.63289779e-01 -5.27488112e-01 -5.89274764e-01
8.75803351e-01 -3.49951088e-01 1.02095509e+00 4.49080050e-01
8.50194514e-01 4.67705995e-01 9.93385985e-02 3.35107863e-01
-3.29271346e-01 -2.52543241e-01 -1.14255905e+00 8.69582072e-02
8.70606959e-01 1.38682091e+00 -7.20796704e-01 6.52210712e-02
-3.02135080e-01 8.60218823e-01 -2.29696244e-01 5.13464153e-01
-8.51334214e-01 -6.66338861e-01 1.11726165e+00 1.46185607e-01
1.92781702e-01 -3.33397925e-01 4.74476516e-01 -3.86968553e-01
-3.59485894e-01 -2.49649778e-01 -1.38323128e-01 -6.77854419e-01
-9.20454085e-01 6.02141857e-01 -1.81911632e-01 -1.28219759e+00
-1.30008712e-01 -2.44356349e-01 -5.67382455e-01 4.67273295e-01
-1.81734884e+00 -7.80366182e-01 -3.68617684e-01 5.36443174e-01
5.72113156e-01 -5.36739886e-01 6.80905819e-01 3.92229229e-01
-6.59068167e-01 1.02874428e-01 5.55489361e-01 -4.21750456e-01
2.23318227e-02 -1.09600103e+00 6.65197242e-03 5.91253877e-01
-2.87719607e-01 -5.22944152e-01 6.98720276e-01 -2.59662002e-01
-1.68495107e+00 -1.41552210e+00 1.72453597e-01 2.72341698e-01
4.40923125e-01 1.45376444e-01 -7.09117949e-02 -1.04454502e-01
7.96194077e-01 -2.95944735e-02 4.42934394e-01 -3.46093535e-01
1.05973673e+00 -1.81421861e-01 -1.41662621e+00 8.40111256e-01
9.22294557e-01 2.35813439e-01 6.13418639e-01 4.82290626e-01
6.20493293e-01 -2.51968324e-01 -6.93856478e-01 3.34638834e-01
3.20371777e-01 -6.01937652e-01 6.38934255e-01 -1.44413516e-01
-1.11923069e-01 -2.92706788e-01 -1.77860662e-01 -2.32014132e+00
-3.48447829e-01 -7.62012601e-01 -3.02799225e-01 9.54038024e-01
4.31541026e-01 -6.51826084e-01 4.95977521e-01 1.35716185e-01
-1.53061688e-01 -9.50020730e-01 -1.40756631e+00 -7.88742721e-01
-1.92198217e-01 3.85507643e-02 8.24477971e-01 4.57259506e-01
-2.05114812e-01 4.30060416e-01 -3.98745298e-01 7.95542598e-01
6.81390822e-01 -6.59391701e-01 5.61881781e-01 -1.22981024e+00
2.69816160e-01 -4.09764469e-01 -3.51475149e-01 -9.66982007e-01
1.69431612e-01 -3.30689341e-01 9.59437573e-04 -2.01113915e+00
-5.33583999e-01 -6.69607699e-01 1.44494101e-01 6.78486153e-02
4.08183187e-01 1.21136039e-01 -3.78664583e-02 -2.29557335e-01
-7.69447863e-01 8.38244557e-01 1.33794332e+00 -2.97833383e-01
-3.14386994e-01 6.16296053e-01 -6.20421708e-01 3.89051974e-01
9.29707766e-01 -1.37442857e-01 -1.03670537e+00 -2.77058303e-01
1.60787776e-01 6.76392615e-01 3.52232188e-01 -1.43541574e+00
3.83558273e-01 -4.54490125e-01 1.70418143e-01 -3.77443641e-01
3.40805054e-01 -1.44794536e+00 1.33951262e-01 5.34689963e-01
3.84387165e-01 8.68013054e-02 2.27359906e-01 1.01151371e+00
4.61695530e-03 7.28889406e-02 9.81204093e-01 1.37474567e-01
-6.11362755e-01 7.79361486e-01 -1.24451649e+00 -1.85265273e-01
1.84882402e+00 -2.14376435e-01 -2.95122921e-01 -8.66895318e-01
-1.78776637e-01 1.27108526e+00 4.35636103e-01 1.16335720e-01
4.33708966e-01 -1.09587443e+00 -6.07802689e-01 -1.08502768e-01
-2.30191857e-01 9.07448679e-02 4.88821030e-01 7.84483850e-01
-3.15475583e-01 4.00191635e-01 -3.18590969e-01 -4.13852155e-01
-8.55414271e-01 1.13135062e-01 7.83222139e-01 1.04842395e-01
-4.69272025e-02 5.20168841e-01 -4.53117728e-01 -5.00712059e-02
1.61841348e-01 9.96984318e-02 -1.85869440e-01 2.84548908e-01
8.32686573e-02 7.69544542e-01 -1.50166586e-01 -5.61556518e-01
-2.36542284e-01 4.69982028e-01 5.86508870e-01 5.31270951e-02
1.19975936e+00 -7.79850602e-01 2.42613211e-01 -1.66000426e-01
7.42493689e-01 -3.89391750e-01 -1.45577896e+00 2.14914992e-01
-5.71729898e-01 -8.81030262e-02 6.50296569e-01 -6.42054141e-01
-1.38175380e+00 4.13195074e-01 5.87498665e-01 1.02396023e+00
1.46770048e+00 -9.83508900e-02 7.92757154e-01 3.41018558e-01
9.09949899e-01 -1.22470725e+00 -2.22268239e-01 2.84366637e-01
3.72249842e-01 -1.00815606e+00 -2.67719477e-01 -3.98797654e-02
-6.77249372e-01 9.28946614e-01 7.03118682e-01 7.99546838e-02
7.65359998e-01 1.27745584e-01 -4.14297342e-01 -1.10248052e-01
-9.17170644e-01 -6.31075203e-01 -3.61568004e-01 1.12944973e+00
-1.82411730e-01 4.63419110e-01 2.72347722e-02 -1.12955101e-01
7.57891983e-02 -2.90667057e-01 7.46531963e-01 7.16024876e-01
-7.71681428e-01 -8.19295228e-01 5.21345735e-02 7.83898413e-01
-1.39224038e-01 4.07615364e-01 -1.38550799e-03 8.67372215e-01
3.47806513e-01 1.44746161e+00 3.85895938e-01 -1.51064560e-01
5.69396436e-01 -7.74252832e-01 2.38592446e-01 -1.46861434e-01
1.75399378e-01 -8.07689503e-02 2.26536348e-01 -1.34979039e-01
-6.24793887e-01 -5.18709123e-01 -1.04219520e+00 -7.05743372e-01
-5.24181008e-01 5.26675224e-01 6.86758220e-01 1.00910151e+00
3.94930601e-01 9.38305736e-01 1.34008360e+00 -8.22234273e-01
4.37612273e-02 -6.55370295e-01 -9.95115578e-01 -9.84630287e-01
7.74336874e-01 -7.60000467e-01 -3.35242152e-01 -6.50933444e-01] | [5.828906536102295, 1.6099443435668945] |
81af64d0-6d5f-490e-a119-14eb11f14221 | short-term-solar-irradiance-forecasting-using | 2010.04715 | null | https://arxiv.org/abs/2010.04715v2 | https://arxiv.org/pdf/2010.04715v2.pdf | Short-Term Solar Irradiance Forecasting Using Calibrated Probabilistic Models | Advancing probabilistic solar forecasting methods is essential to supporting the integration of solar energy into the electricity grid. In this work, we develop a variety of state-of-the-art probabilistic models for forecasting solar irradiance. We investigate the use of post-hoc calibration techniques for ensuring well-calibrated probabilistic predictions. We train and evaluate the models using public data from seven stations in the SURFRAD network, and demonstrate that the best model, NGBoost, achieves higher performance at an intra-hourly resolution than the best benchmark solar irradiance forecasting model across all stations. Further, we show that NGBoost with CRUDE post-hoc calibration achieves comparable performance to a numerical weather prediction model on hourly-resolution forecasting. | ['Ram Rajagopal', 'Anand Avati', 'David Gagne', 'Andrew Y. Ng', 'Jack Kelly', 'Hao Sheng', 'Cooper Raterink', 'Jeremy Irvin', 'Sharon Zhou', 'Eric Zelikman'] | 2020-10-09 | null | null | null | null | ['solar-irradiance-forecasting'] | ['time-series'] | [-4.69414026e-01 -3.81589532e-01 -8.58993009e-02 -6.69489622e-01
-1.22957444e+00 -1.05649233e+00 9.08343136e-01 -5.05565666e-03
2.03376010e-01 1.10642385e+00 5.52844346e-01 -5.13746440e-01
-4.30591851e-01 -1.17815542e+00 -6.76997840e-01 -9.59553599e-01
1.18632726e-01 5.15848577e-01 -2.75147501e-02 -8.14978257e-02
7.45004043e-02 4.07578647e-01 -1.38499701e+00 -1.26011521e-01
1.13263476e+00 8.31567764e-01 -1.79193079e-01 9.85938728e-01
8.91399011e-02 2.47107267e-01 -6.34847462e-01 4.09869142e-02
3.44035596e-01 -2.19923586e-01 -4.15382385e-02 -7.61735201e-01
5.75362802e-01 -1.11461386e-01 1.29751517e-02 7.52220213e-01
9.15876329e-01 6.51478320e-02 9.03860331e-01 -1.32017291e+00
-2.23826617e-01 7.65684962e-01 -3.53336751e-01 4.97271597e-01
1.81650296e-01 -6.69887066e-02 9.47063386e-01 -4.81664449e-01
-3.54094446e-01 7.08354592e-01 1.29762995e+00 -2.77316064e-01
-1.37194264e+00 -7.74754107e-01 -2.43690416e-01 -3.81822288e-02
-1.52271426e+00 -2.94916958e-01 4.06880647e-01 -7.18412340e-01
1.18452621e+00 3.38098437e-01 5.17334104e-01 8.82264018e-01
4.83406216e-01 -7.10675074e-03 1.68528259e+00 -4.08317983e-01
5.42176187e-01 -1.68188751e-01 3.22688580e-01 5.70302978e-02
4.46886599e-01 8.83991599e-01 -7.96973705e-01 -6.61180973e-01
-2.74360310e-02 -4.21748668e-01 -3.69470149e-01 3.84040713e-01
-6.35234296e-01 7.75934756e-01 4.98055309e-01 6.22545741e-02
-5.78049958e-01 2.71499515e-01 -4.56517220e-01 -4.16080028e-01
9.19590533e-01 -5.47071099e-02 -8.71096432e-01 -6.00230172e-02
-1.57244909e+00 2.97071844e-01 1.18458903e+00 7.05455482e-01
7.82504559e-01 3.82757038e-01 -3.43758166e-01 6.38861060e-01
7.91348100e-01 1.50616312e+00 -2.12512258e-02 -9.08655703e-01
5.20645827e-02 -9.95847657e-02 6.03845358e-01 -4.59351569e-01
-5.86971939e-01 -6.09757006e-01 -7.28525460e-01 4.25271273e-01
-1.50124906e-02 -6.22542799e-01 -9.29135799e-01 1.24995553e+00
6.05335832e-02 4.67694640e-01 2.22330913e-01 4.32669282e-01
4.77953583e-01 1.20458961e+00 4.50152934e-01 -1.74745813e-01
9.81647670e-01 -7.22781658e-01 -5.03050268e-01 -1.50272578e-01
2.29129381e-02 -8.52105975e-01 2.84029841e-01 2.13454276e-01
-6.69226944e-01 -3.51422310e-01 -1.10997891e+00 3.65158409e-01
-6.52733207e-01 1.95878446e-02 4.23986763e-01 1.16440034e+00
-1.21166945e+00 6.34249330e-01 -9.07736659e-01 -3.03823173e-01
3.88336889e-02 -2.83607483e-01 3.00635278e-01 2.69307345e-01
-1.18054605e+00 1.12882900e+00 4.53321844e-01 2.07850248e-01
-7.74438441e-01 -1.25329220e+00 -6.36420488e-01 4.80223864e-01
-4.66926634e-01 -8.89419138e-01 1.56320214e+00 -3.31654578e-01
-1.68959200e+00 -2.00369924e-01 -5.88527262e-01 -6.14349008e-01
3.44261557e-01 1.41634662e-02 -6.98902845e-01 -4.68429953e-01
-1.18563198e-01 3.29522491e-01 3.90239507e-01 -1.58107352e+00
-9.75214720e-01 -3.53508815e-02 -3.52505386e-01 2.49934196e-01
5.64517528e-02 -2.05802709e-01 1.52182862e-01 -5.02184629e-01
-1.38694286e-01 -8.65138054e-01 -1.10950522e-01 -5.79235077e-01
-1.82476357e-01 -5.81952274e-01 5.48274696e-01 -8.93537283e-01
1.05319417e+00 -1.58955908e+00 -6.10098541e-01 7.30768919e-01
-6.76841736e-01 -1.49309710e-01 3.31329107e-01 6.31725848e-01
-1.49503499e-02 3.13907593e-01 -4.80430126e-01 -5.65600634e-01
3.89691561e-01 6.51719034e-01 -8.57605875e-01 3.23210031e-01
-1.75625816e-01 9.41833735e-01 -7.07010150e-01 -2.47256849e-02
6.64048195e-01 7.16121316e-01 8.04422647e-02 3.15996587e-01
-3.68305892e-01 2.10477263e-01 3.87315229e-02 5.99326193e-01
1.12543762e+00 -9.35648084e-02 5.65243699e-02 -7.77051374e-02
-5.41794717e-01 4.37500060e-01 -9.56941307e-01 1.18549168e+00
-4.80830103e-01 5.11215866e-01 -4.08599883e-01 -1.40545115e-01
6.73608899e-01 1.08685151e-01 1.44224674e-01 -4.85407084e-01
-3.21510792e-01 -3.39579917e-02 -5.66207349e-01 8.39467049e-02
3.86269599e-01 -1.70677155e-01 2.82440037e-01 5.81076026e-01
-3.17771137e-02 -6.88664734e-01 1.78600058e-01 -1.71717644e-01
3.56108606e-01 3.11427981e-01 3.12347002e-02 -7.92535245e-01
6.52493536e-02 -2.75090919e-03 5.81747591e-01 1.10371971e+00
2.28865772e-01 7.92541206e-01 -8.33045468e-02 -1.14740774e-01
-9.15586412e-01 -1.22115874e+00 -8.25047851e-01 9.70065475e-01
-4.75450069e-01 -2.51163691e-01 -4.14531350e-01 -3.54278594e-01
5.56099534e-01 1.69649148e+00 -5.36603987e-01 6.83263242e-01
2.73798466e-01 -1.71407020e+00 5.29553354e-01 7.00951993e-01
2.51773000e-01 -4.44823354e-01 -2.49268383e-01 2.17595175e-01
-1.27764612e-01 -6.70687973e-01 1.03963509e-01 3.75359833e-01
-6.14099562e-01 -7.35005438e-01 -5.66851676e-01 5.11436611e-02
1.64188936e-01 2.03949079e-01 1.52689850e+00 -3.90542269e-01
3.37652653e-01 5.41228592e-01 -1.75980777e-02 -1.07991672e+00
-2.79867589e-01 1.56381950e-01 -7.27696419e-02 -6.82526827e-01
2.86068290e-01 -7.26769984e-01 -5.65599680e-01 3.73755604e-01
-5.72650552e-01 -2.16250181e-01 7.75007680e-02 3.70521605e-01
5.22758305e-01 5.01287043e-01 3.22422802e-01 -4.81090456e-01
2.55107284e-01 -8.69579256e-01 -1.45161629e+00 5.12059212e-01
-1.33297169e+00 -1.36096746e-01 5.37258647e-02 4.30721492e-01
-1.50369418e+00 -1.40048852e-02 -1.64226234e-01 2.48252749e-02
-1.30736724e-01 9.51532245e-01 4.20140266e-01 -2.39202589e-01
7.30263650e-01 1.87214375e-01 -7.95979857e-01 -4.31724608e-01
4.01591569e-01 5.14725626e-01 7.16656864e-01 -6.18473947e-01
1.10918212e+00 3.38798463e-01 1.00570560e-01 -7.73801327e-01
-1.19324529e+00 -4.42461938e-01 -3.28146875e-01 -2.38270536e-01
6.30820990e-01 -1.38476086e+00 -4.06673253e-01 7.07763195e-01
-8.88924658e-01 -7.07217515e-01 -3.62451486e-02 3.94889861e-01
-1.75390095e-01 2.14945316e-01 1.93144772e-02 -1.34086013e+00
-6.60782576e-01 -4.41693306e-01 1.12948895e+00 8.14231515e-01
1.63898975e-01 -1.33259118e+00 1.09061670e+00 -1.40675027e-02
8.02795649e-01 1.88639730e-01 5.38732529e-01 -1.82571411e-01
-4.74165261e-01 -7.33729079e-02 -3.65623027e-01 3.70394826e-01
7.05832094e-02 7.96141088e-01 -1.70983064e+00 -1.44335076e-01
-2.85275221e-01 1.94282402e-04 9.25503135e-01 8.77566874e-01
1.05815423e+00 -2.24931642e-01 -2.87721455e-01 8.02826822e-01
1.68462574e+00 -1.59183487e-01 5.19873321e-01 2.25467816e-01
1.76189065e-01 1.47988334e-01 5.14533333e-02 5.63987017e-01
9.08602655e-01 -7.39725083e-02 3.88971537e-01 -1.11718737e-02
2.07168996e-01 -2.94220984e-01 4.97353263e-02 7.24575341e-01
-7.53693655e-02 -5.69313526e-01 -1.25176978e+00 4.83042836e-01
-1.85605049e+00 -1.02762485e+00 -3.78238827e-01 2.24440503e+00
7.35308766e-01 2.45317891e-02 -3.58163893e-01 -2.86344051e-01
2.90033638e-01 5.06812751e-01 -2.95026958e-01 -9.73215997e-02
-5.83523810e-01 4.70047861e-01 1.37634516e+00 8.06149125e-01
-1.13519394e+00 3.69151384e-01 8.34302235e+00 4.55398947e-01
-7.37505853e-01 3.12186033e-01 7.64329433e-01 2.55367756e-01
-6.45464063e-01 9.23596770e-02 -1.26457858e+00 7.44663715e-01
1.69135284e+00 -3.61121595e-01 7.48361349e-01 7.86893487e-01
6.85675085e-01 -7.91780591e-01 -6.01676822e-01 5.71160257e-01
1.10965364e-01 -1.63332260e+00 -4.77056146e-01 -1.09760158e-01
1.72053576e+00 9.34290886e-01 -4.81981754e-01 1.90113157e-01
1.17623377e+00 -1.18395519e+00 5.74002445e-01 1.32429242e+00
3.10978889e-01 -7.83193469e-01 1.09093237e+00 3.13292593e-01
-1.10379636e+00 -3.37502137e-02 -5.84500015e-01 -4.77794483e-02
3.58146131e-01 1.40666687e+00 -6.23617172e-01 1.18899941e+00
1.44920540e+00 5.84455550e-01 -3.12139481e-01 1.55877948e+00
-6.34860694e-01 1.31448388e+00 -1.27806902e+00 1.75462678e-01
-5.49438149e-02 -1.65785551e-01 1.08408667e-01 1.27274430e+00
9.05416906e-01 4.80623007e-01 1.34863388e-02 4.72111166e-01
9.97074507e-03 -3.96000534e-01 -4.17460650e-01 4.07128155e-01
8.33328485e-01 1.33806396e+00 2.54385304e-02 -4.10830587e-01
-1.67084098e-01 1.42464191e-01 -1.76563933e-01 4.28108752e-01
-7.12600231e-01 1.62236050e-01 7.03379333e-01 -4.94714290e-01
4.05277997e-01 -4.54651535e-01 -9.06682134e-01 -7.61763096e-01
-3.64222735e-01 -2.37007469e-01 5.63970923e-01 -1.71127820e+00
-1.88032043e+00 4.58992571e-02 2.78545737e-01 -7.47328222e-01
-3.04850698e-01 -5.40124834e-01 -1.41114926e+00 1.81678832e+00
-2.27923012e+00 -1.11415184e+00 -8.66306722e-01 1.69017762e-01
-2.48336210e-03 1.78197727e-01 1.21227944e+00 -2.07971498e-01
-3.93145472e-01 1.97388809e-02 8.84599745e-01 -3.94166768e-01
6.96276665e-01 -1.62434828e+00 6.44669175e-01 9.13862407e-01
7.21211657e-02 1.74597129e-01 8.53733599e-01 -6.56108558e-01
-1.08369350e+00 -1.15392768e+00 9.62632418e-01 -8.38582575e-01
8.78486991e-01 1.43793061e-01 -7.69201219e-01 7.01675653e-01
9.00506735e-01 -3.09666753e-01 9.31787670e-01 3.60426784e-01
-2.97783434e-01 -4.32748526e-01 -1.06129658e+00 -2.80187540e-02
9.79801863e-02 -5.81232786e-01 -8.35430443e-01 7.98753321e-01
1.81860983e-01 -4.23602611e-01 -1.41746378e+00 8.67492914e-01
8.35215271e-01 -1.00633490e+00 7.17083752e-01 -7.23634511e-02
-8.48989338e-02 -9.18356359e-01 -3.33369881e-01 -1.87498415e+00
-4.47863042e-01 -6.59768581e-01 -1.20782435e-01 1.46301436e+00
6.73017561e-01 -9.65096533e-01 6.16424263e-01 7.50357151e-01
-8.84220079e-02 7.60709569e-02 -1.16223443e+00 -5.99453092e-01
4.98153090e-01 -6.73300683e-01 8.75462174e-01 8.79940391e-01
-5.54921150e-01 -3.00427198e-01 4.44855243e-02 1.21062434e+00
9.48594868e-01 2.23306909e-01 6.61408901e-01 -1.29987323e+00
-5.72350807e-02 -4.13024634e-01 3.45913082e-01 -5.85365295e-01
6.58460408e-02 -3.17789704e-01 4.87899065e-01 -1.90519536e+00
1.10665508e-01 -5.60144007e-01 -7.83629477e-01 8.94898117e-01
-2.93758214e-01 3.52610737e-01 -2.50239879e-01 7.93471262e-02
2.98287690e-01 8.07991803e-01 7.04924539e-02 -1.59100965e-01
8.92712176e-02 2.55875736e-01 -2.74081677e-01 4.73168492e-01
1.13615298e+00 -5.22209108e-01 -1.57952055e-01 -5.03562748e-01
5.14982700e-01 -1.74664736e-01 2.88769335e-01 -1.23594940e+00
4.81473118e-01 -3.68092537e-01 8.72651935e-01 -1.51429379e+00
6.83599859e-02 -7.08747149e-01 6.02292776e-01 2.26651169e-02
2.73580462e-01 2.84685731e-01 5.96477330e-01 4.19376433e-01
1.70132026e-01 -1.76633328e-01 6.51530623e-01 3.34591240e-01
-3.11542124e-01 -3.25027965e-02 -4.08974767e-01 -3.68604690e-01
6.28896952e-01 3.98423165e-01 -1.05669701e+00 -4.15150702e-01
-3.68826509e-01 7.22676456e-01 4.20829475e-01 3.40174362e-02
-2.02203378e-01 -9.81019795e-01 -1.05110955e+00 7.97758400e-02
3.09676137e-02 -1.09595984e-01 1.94735378e-01 4.36956227e-01
-5.34630537e-01 5.66132486e-01 3.56253237e-01 -7.43101299e-01
-6.00783288e-01 -1.09861009e-01 8.76978338e-01 -4.07514572e-02
2.65753977e-02 6.76172733e-01 -3.68185818e-01 -8.11653197e-01
-8.63198861e-02 -5.74806333e-01 4.82246205e-02 1.53234765e-01
4.56014752e-01 6.60157025e-01 4.93476391e-01 -2.80511558e-01
-1.36752278e-01 5.64334095e-01 9.09210086e-01 -1.86384708e-01
1.20171034e+00 -1.27938241e-01 -2.47756690e-01 7.58222461e-01
3.15459281e-01 1.03253238e-01 -1.59305573e+00 1.27873629e-01
-1.48670956e-01 -2.02148721e-01 7.46473134e-01 -1.68649936e+00
-7.15709805e-01 3.93381268e-01 8.34888458e-01 4.47519660e-01
7.63807714e-01 -5.11803985e-01 2.42815927e-01 5.63409209e-01
3.69994372e-01 -9.98041570e-01 -1.20699382e+00 5.90874374e-01
6.89506710e-01 -1.37313664e+00 5.74303687e-01 3.85345854e-02
-2.92635649e-01 1.07382846e+00 -2.06136573e-02 -3.82748954e-02
1.46793962e+00 4.33367074e-01 3.47954035e-01 2.68326610e-01
-6.95396543e-01 -4.30328175e-02 6.61913931e-01 4.08431292e-01
1.14858694e-01 5.03986001e-01 1.24537311e-01 3.54425907e-01
-5.19963682e-01 1.20612763e-01 2.33907819e-01 7.35018909e-01
-5.06269872e-01 -7.45594442e-01 -7.11074531e-01 5.17771959e-01
-2.04271302e-01 -5.68643987e-01 9.01036188e-02 3.56753081e-01
-2.75266618e-01 1.30671835e+00 2.36018673e-01 2.19764203e-01
1.68640941e-01 4.86904174e-01 4.01479639e-02 -1.43126994e-01
-5.05398154e-01 6.91967160e-02 3.36135060e-01 -2.72467256e-01
-6.45098031e-01 -9.13985550e-01 -5.70331156e-01 -7.48982549e-01
-2.26825491e-01 4.58066583e-01 1.31416166e+00 9.48072195e-01
2.71087289e-01 4.13385481e-01 1.05935574e+00 -1.50719821e+00
-6.62886381e-01 -1.30835092e+00 -4.52344209e-01 -6.16322100e-01
1.32804811e-01 -5.18777430e-01 -7.37577140e-01 -1.10199921e-01] | [6.496255397796631, 2.986654281616211] |
c7d028d3-85e3-4c48-9d41-8365ddd76b2d | transforming-clip-to-an-open-vocabulary-video | 2302.00624 | null | https://arxiv.org/abs/2302.00624v3 | https://arxiv.org/pdf/2302.00624v3.pdf | Open-VCLIP: Transforming CLIP to an Open-vocabulary Video Model via Interpolated Weight Optimization | Contrastive Language-Image Pretraining (CLIP) has demonstrated impressive zero-shot learning abilities for image understanding, yet limited effort has been made to investigate CLIP for zero-shot video recognition. We introduce Open-VCLIP, a simple yet effective approach that transforms CLIP into a strong zero-shot video classifier that can recognize unseen actions and events at test time. Our framework extends CLIP with minimal modifications to model spatial-temporal relationships in videos, making it a specialized video classifier, while striving for generalization. We formally show that training an Open-VCLIP is equivalent to continual learning with zero historical data. To address this problem, we propose Interpolated Weight Optimization, which utilizes the benefit of weight interpolation in both training and test time. We evaluate our method on three popular and challenging action recognition datasets following various zero-shot evaluation protocols and we demonstrate our approach outperforms state-of-the-art methods by clear margins. In particular, we achieve 87.9%, 58.3%, 81.1% zero-shot accuracy on UCF, HMDB and Kinetics-600 respectively, outperforming state-of-the-art methods by 8.3%, 7.8% and 12.2%. Code is released at https://github.com/wengzejia1/Open-VCLIP. | ['Yu-Gang Jiang', 'Zuxuan Wu', 'Ang Li', 'Xitong Yang', 'Zejia Weng'] | 2023-02-01 | null | null | null | null | ['video-recognition'] | ['computer-vision'] | [ 2.13640079e-01 -3.95170242e-01 -5.62323034e-01 -1.80607647e-01
-9.45479929e-01 -2.59042382e-01 5.43553591e-01 -3.31727743e-01
-4.08605635e-01 5.62623978e-01 1.46964192e-01 -1.94030881e-01
1.27650246e-01 -4.63313073e-01 -9.82090175e-01 -5.45730352e-01
-3.19683045e-01 -1.25073925e-01 4.92832273e-01 1.61732342e-02
1.64199442e-01 -5.48009016e-02 -1.77033186e+00 7.37706661e-01
6.01929665e-01 1.15877724e+00 1.26333609e-01 9.14375484e-01
3.67183536e-02 1.33050585e+00 -4.08264428e-01 -3.59956503e-01
1.21219076e-01 -4.54932719e-01 -6.21677220e-01 2.94144511e-01
6.21782124e-01 -5.26261270e-01 -7.19759107e-01 7.84263134e-01
2.95670509e-01 3.50586772e-01 6.75369382e-01 -1.42410862e+00
-9.83202636e-01 3.48282993e-01 -5.58099985e-01 5.46778262e-01
5.04988790e-01 2.99857885e-01 8.75894487e-01 -1.15481329e+00
7.36808717e-01 9.97112215e-01 6.68377280e-01 7.81589985e-01
-7.89673269e-01 -7.37209141e-01 1.35091767e-01 7.24604726e-01
-1.33691323e+00 -5.97651958e-01 4.26816672e-01 -6.19809330e-01
1.21088827e+00 1.35658368e-01 5.02075851e-01 1.54474795e+00
1.64200276e-01 1.24375594e+00 7.30293095e-01 -4.19026494e-01
2.29214907e-01 -1.53683692e-01 2.67048061e-01 7.11189747e-01
-2.51209527e-01 6.64658770e-02 -7.28814542e-01 1.59652025e-01
5.08802116e-01 3.12366724e-01 -4.88718897e-01 -4.45399702e-01
-1.02477860e+00 5.54886460e-01 2.18164653e-01 2.11162001e-01
-1.97702557e-01 1.88785270e-01 6.65792286e-01 3.73246998e-01
4.03146088e-01 -7.48079419e-02 -3.93561691e-01 -6.18746698e-01
-8.25143099e-01 -1.70438871e-01 4.63168949e-01 9.27600861e-01
4.67307031e-01 4.28646892e-01 -4.53340322e-01 9.13308144e-01
-1.28498182e-01 4.08654362e-01 7.39884198e-01 -9.28762972e-01
3.32944185e-01 1.58999234e-01 -2.78026443e-02 -6.63240552e-01
9.57146212e-02 -2.67412394e-01 -6.70564592e-01 3.83552387e-02
1.43960997e-01 4.05412465e-02 -1.29450083e+00 1.60923839e+00
-3.87878832e-03 1.19270051e+00 2.91465670e-01 7.79023170e-01
8.52646589e-01 8.77171397e-01 2.13872612e-01 -2.89095610e-01
1.05636191e+00 -1.15191710e+00 -6.00521982e-01 5.01599535e-02
6.45075738e-01 -3.56031418e-01 1.39590752e+00 4.42240596e-01
-8.88078511e-01 -7.66683996e-01 -1.19037306e+00 1.64060161e-01
-3.58796626e-01 2.54821070e-02 5.75149655e-01 4.99397457e-01
-6.27745032e-01 6.63843393e-01 -9.56075490e-01 -4.02568400e-01
6.18742228e-01 5.41697294e-02 -4.10452217e-01 -3.66461217e-01
-1.15446806e+00 5.42888463e-01 3.10168117e-01 -4.47556823e-01
-1.40314925e+00 -9.84443367e-01 -9.95032310e-01 7.47244880e-02
7.86895037e-01 -1.43395588e-01 1.32152677e+00 -1.04719973e+00
-1.31565094e+00 7.34099567e-01 -1.47119597e-01 -6.88737273e-01
4.67010111e-01 -4.43709612e-01 -7.34631062e-01 4.92949873e-01
5.64555041e-02 6.88574076e-01 9.26630318e-01 -9.24685001e-01
-6.55025721e-01 -1.17640696e-01 1.39582515e-01 -1.79444365e-02
-4.97260094e-01 3.15605453e-03 -8.86146963e-01 -8.18951547e-01
-3.08449030e-01 -7.31945038e-01 2.79978782e-01 1.61199287e-01
1.02958933e-01 -2.33926848e-01 1.05450869e+00 -5.77066123e-01
1.19929743e+00 -2.32376957e+00 -7.64234364e-02 -4.22529459e-01
-1.08976848e-01 6.78010583e-01 -2.43752748e-01 2.47508198e-01
-7.16749877e-02 -2.74308566e-02 -1.44834250e-01 -2.27886230e-01
-1.01855978e-01 3.20533812e-01 -3.85125935e-01 3.98899704e-01
3.47382426e-01 9.24948037e-01 -9.81851161e-01 -4.15328532e-01
5.75945675e-01 6.23315096e-01 -6.51131868e-01 2.08897352e-01
-2.83339769e-02 -7.66077871e-03 -9.77772474e-02 1.07635689e+00
3.20056498e-01 -3.27885807e-01 -2.19226815e-02 -2.56846566e-02
3.47624160e-02 -3.38062376e-01 -1.08642352e+00 1.79816580e+00
-1.53950110e-01 7.63979852e-01 -4.86268997e-01 -1.25830519e+00
4.69825029e-01 5.39241850e-01 5.64297438e-01 -8.32481623e-01
1.57470539e-01 -1.81436777e-01 -2.58525848e-01 -7.06265926e-01
2.31820956e-01 6.07248861e-03 -4.36146781e-02 6.64518699e-02
7.12547123e-01 6.02670908e-01 3.81267339e-01 3.97754133e-01
1.17700136e+00 3.03086966e-01 2.24621594e-01 8.72072857e-03
5.50862610e-01 -2.27951258e-01 6.80238962e-01 8.54358435e-01
-6.09099090e-01 7.22481966e-01 3.72987330e-01 -3.13349932e-01
-7.52092898e-01 -1.18983912e+00 -7.30304867e-02 1.35232377e+00
1.90172687e-01 -6.03618920e-01 -7.30944097e-01 -5.95609128e-01
-4.03750129e-02 7.84780681e-01 -8.58369231e-01 -3.05181950e-01
-2.68641084e-01 -3.60270888e-01 4.33627218e-01 8.12908649e-01
5.67852139e-01 -1.02445984e+00 -7.08181798e-01 -1.90542284e-02
-1.32258818e-01 -1.22827232e+00 -5.06161690e-01 -2.15702988e-02
-5.38835168e-01 -1.22974813e+00 -8.64140391e-01 -7.30414510e-01
3.52878034e-01 6.87622845e-01 1.03783846e+00 9.70155299e-02
-6.26355350e-01 8.53037119e-01 -7.87120342e-01 -1.53178811e-01
7.77944326e-02 -3.94454807e-01 2.56669551e-01 2.54521400e-01
5.36135674e-01 -3.95726830e-01 -4.96819377e-01 4.49711323e-01
-8.75334442e-01 9.15357098e-03 3.38208675e-01 9.29220855e-01
6.12576187e-01 -3.52164000e-01 5.31145930e-01 -5.70825815e-01
1.99952334e-01 -6.76714301e-01 -2.88492948e-01 5.08179605e-01
-3.32801431e-01 -1.69511914e-01 5.22030890e-01 -7.73446321e-01
-9.21520650e-01 1.83614641e-02 -3.90482731e-02 -1.21396840e+00
-1.50543362e-01 3.06086630e-01 3.71899344e-02 -9.40350816e-02
6.35120928e-01 4.62041914e-01 -1.85037568e-01 -2.47045025e-01
3.87557417e-01 7.52855599e-01 8.82818758e-01 -4.54072982e-01
4.74240929e-01 4.60742861e-01 -5.49251556e-01 -1.02221656e+00
-9.46542680e-01 -7.92729497e-01 -6.17970467e-01 -5.45771599e-01
1.02636206e+00 -1.13665795e+00 -4.77208048e-01 4.93078530e-01
-6.37648642e-01 -5.32181263e-01 -1.70666322e-01 4.85419005e-01
-7.49074221e-01 4.97422725e-01 -6.37822628e-01 -6.92988157e-01
-2.56678194e-01 -9.67411876e-01 8.09091628e-01 5.30233644e-02
1.50831770e-02 -7.09100664e-01 -2.36310717e-02 3.59390199e-01
1.98735148e-01 2.01856732e-01 3.76141965e-01 -4.73241150e-01
-4.85652447e-01 -2.06107318e-01 -1.49906069e-01 5.47701478e-01
-1.69249490e-01 3.05001121e-02 -9.28870440e-01 -4.55431461e-01
-1.36994526e-01 -9.19236660e-01 1.12130117e+00 3.25518936e-01
1.43251646e+00 -4.24413243e-03 -1.10921092e-01 8.28473032e-01
1.54965758e+00 4.76226151e-01 9.38139737e-01 1.88351944e-01
4.89959836e-01 5.79746068e-03 7.68079042e-01 6.63815677e-01
1.20949276e-01 6.60577416e-01 2.86507100e-01 3.23244482e-01
-3.04944575e-01 -3.09577078e-01 6.81118369e-01 6.21383429e-01
-2.19978556e-01 -2.84957111e-01 -8.55713487e-01 7.35624671e-01
-1.98244703e+00 -1.43320620e+00 1.94543451e-01 2.11614847e+00
5.13919830e-01 2.78750420e-01 1.42236911e-02 1.05402969e-01
5.76482952e-01 4.14318115e-01 -6.01484239e-01 1.68147031e-02
-2.72031501e-02 2.45365143e-01 3.63123357e-01 3.19864810e-01
-1.44734204e+00 1.08078337e+00 5.66422176e+00 9.39275384e-01
-1.12179434e+00 3.05118203e-01 5.98521888e-01 -4.04786140e-01
3.21410954e-01 -2.03493431e-01 -6.83730543e-01 5.89786291e-01
9.75843668e-01 -3.03896248e-01 3.07757139e-01 1.08214283e+00
-2.61398721e-02 6.07426092e-02 -9.98739421e-01 1.35086751e+00
6.23756349e-01 -1.47359788e+00 8.11555386e-02 -3.55113775e-01
8.13653409e-01 2.17611283e-01 3.92280445e-02 9.37041581e-01
-3.82747576e-02 -9.16520596e-01 6.05550528e-01 5.46198249e-01
9.52609003e-01 -5.98675787e-01 5.07606089e-01 2.94910520e-01
-1.30870473e+00 -2.90315956e-01 -4.63301122e-01 -4.79953438e-02
1.99551061e-01 -1.27906322e-01 -1.79131627e-01 3.65010232e-01
8.25738311e-01 1.09655786e+00 -5.77529490e-01 1.02631342e+00
-1.35389939e-01 8.28944743e-01 1.40493825e-01 2.69035518e-01
4.05266255e-01 1.43346772e-01 2.84849852e-01 1.29350328e+00
4.03072149e-01 4.19312805e-01 4.44251150e-01 2.64257997e-01
-1.54105633e-01 4.78403531e-02 -5.08470297e-01 -8.15596581e-02
1.24249816e-01 8.25302005e-01 -3.61686975e-01 -7.71689773e-01
-9.02335703e-01 1.20594311e+00 3.76057744e-01 4.28817004e-01
-1.30781639e+00 -4.07903254e-01 7.01684535e-01 -8.12302008e-02
8.00255835e-01 -1.18239217e-01 3.82834047e-01 -1.62294078e+00
4.99963537e-02 -9.04463053e-01 6.72546685e-01 -9.32997763e-01
-1.00572801e+00 4.75508422e-01 -3.41077801e-03 -1.50690484e+00
-2.63681740e-01 -8.25238764e-01 -6.21006191e-01 1.47343963e-01
-1.37684655e+00 -1.03842056e+00 -5.21223545e-01 9.31707859e-01
1.13760817e+00 -5.02711177e-01 7.49589801e-01 4.49083269e-01
-6.77888513e-01 7.64100432e-01 1.97604090e-01 2.66166300e-01
6.77463412e-01 -8.97552729e-01 1.31006792e-01 9.15896952e-01
4.36162800e-01 2.28097484e-01 4.39200193e-01 -4.72740442e-01
-1.75535214e+00 -1.16589272e+00 4.16867018e-01 -3.43142241e-01
9.01154339e-01 -3.44159067e-01 -1.11875391e+00 7.58531272e-01
2.43838564e-01 5.07622480e-01 7.96095431e-01 -5.61642349e-02
-6.11854613e-01 -9.70364586e-02 -7.79234529e-01 5.17834902e-01
1.27130175e+00 -6.30989969e-01 -6.84639394e-01 3.17102969e-01
7.12996721e-01 -1.90362975e-01 -7.73545504e-01 4.91734147e-01
7.02245116e-01 -8.93909574e-01 1.03251445e+00 -9.55224693e-01
5.91192901e-01 -3.39754745e-02 -4.66667503e-01 -9.45956707e-01
-3.93850952e-01 -3.73832673e-01 -6.61479235e-01 8.81920934e-01
1.88611865e-01 -1.93638593e-01 8.07697535e-01 4.15323615e-01
-2.77288675e-01 -8.06378901e-01 -7.87734330e-01 -1.15255213e+00
-1.02336496e-01 -9.66459155e-01 -1.56614706e-01 9.65776682e-01
1.58831716e-01 1.41094074e-01 -9.75643396e-01 -1.69116836e-02
6.20467365e-01 -9.92448032e-02 7.70303249e-01 -6.81399047e-01
-4.85700965e-01 -1.81672648e-01 -6.93222165e-01 -1.08103240e+00
2.73471028e-01 -7.28628755e-01 -1.81438774e-01 -1.24285817e+00
4.57905501e-01 2.64206797e-01 -8.30032527e-01 5.42116046e-01
-8.83140564e-02 5.23584187e-01 5.04532814e-01 1.77125216e-01
-1.17350650e+00 7.09600627e-01 8.90928090e-01 -3.28435868e-01
6.64058980e-03 -1.84859365e-01 -2.36620799e-01 6.16201937e-01
8.56672049e-01 -1.72944322e-01 -7.22019851e-01 -4.63295162e-01
-5.06774366e-01 1.17771678e-01 3.95298779e-01 -1.44181299e+00
2.73894191e-01 -1.86569601e-01 4.00444210e-01 -5.35636902e-01
5.59971631e-01 -4.62930501e-01 -4.33865301e-02 4.51995075e-01
-3.39544952e-01 -2.99293309e-01 2.09055781e-01 8.59118938e-01
-3.36653769e-01 -1.83642611e-01 7.57328868e-01 -1.99401692e-01
-1.59330249e+00 5.48466444e-01 -2.79664069e-01 2.35644370e-01
1.49821281e+00 -3.07165056e-01 -3.12838256e-01 -3.92006934e-01
-9.26397562e-01 3.19105357e-01 2.60840267e-01 7.75626302e-01
9.73764122e-01 -1.43019104e+00 -5.56579053e-01 1.88782901e-01
4.41480368e-01 -8.07170331e-01 6.72782481e-01 8.22743297e-01
-3.20441097e-01 4.41848576e-01 -3.75475287e-01 -6.94333255e-01
-1.53504777e+00 8.77826154e-01 1.57250956e-01 7.08753318e-02
-7.98619866e-01 8.75366986e-01 1.81867734e-01 1.06785603e-01
6.67107403e-01 -1.46323204e-01 -1.16927158e-02 -6.90519297e-03
9.98900473e-01 3.03688973e-01 -2.46803507e-01 -5.94478607e-01
-3.33158880e-01 5.34305692e-01 -1.80104420e-01 5.45816161e-02
1.32636833e+00 2.20350161e-01 5.57217181e-01 6.79746687e-01
1.34350181e+00 -6.10017776e-01 -1.63894796e+00 -2.89426684e-01
-1.28150836e-01 -7.80179203e-01 1.88066568e-02 -6.30948007e-01
-9.73363698e-01 1.00354981e+00 8.27834845e-01 -1.23662584e-01
1.04531944e+00 -8.09596851e-02 7.69566953e-01 3.61084640e-01
4.78284746e-01 -1.09699428e+00 5.41884542e-01 5.82211673e-01
8.14307928e-01 -1.60507941e+00 -1.36044279e-01 -1.83459669e-01
-9.37506676e-01 8.66353810e-01 7.44891882e-01 -2.53817886e-01
6.89969897e-01 1.40813947e-01 -1.07356176e-01 3.87553684e-02
-1.17232430e+00 -3.21063012e-01 3.26261312e-01 4.86574948e-01
1.75811917e-01 -3.28924693e-02 -1.58522110e-02 5.37929535e-01
3.74029577e-01 4.76370543e-01 3.18388581e-01 1.13225305e+00
-5.61385691e-01 -6.68284059e-01 -6.20397180e-02 3.17454696e-01
-5.01345456e-01 -8.85216743e-02 1.48533778e-02 7.89918780e-01
1.94528643e-02 7.11256027e-01 2.54164428e-01 -5.88862419e-01
2.03285232e-01 3.16982716e-01 4.27339286e-01 -4.90295291e-01
1.41952988e-02 -4.21538241e-02 -1.26453172e-02 -7.97898293e-01
-4.61802632e-01 -6.86849892e-01 -1.02183115e+00 -1.13534957e-01
-9.72751528e-02 -1.11586161e-01 2.13737026e-01 7.59775162e-01
4.26818013e-01 5.67197025e-01 3.52587610e-01 -7.35868394e-01
-5.67443609e-01 -8.82466137e-01 -4.14777994e-01 7.41719067e-01
2.53694713e-01 -9.88675058e-01 -3.82542372e-01 3.97288918e-01] | [8.881330490112305, 0.8772662281990051] |
53a7198b-c480-448b-8083-de66cfbb25ce | online-spatiotemporal-action-detection-and | 2008.13759 | null | https://arxiv.org/abs/2008.13759v1 | https://arxiv.org/pdf/2008.13759v1.pdf | Online Spatiotemporal Action Detection and Prediction via Causal Representations | In this thesis, we focus on video action understanding problems from an online and real-time processing point of view. We start with the conversion of the traditional offline spatiotemporal action detection pipeline into an online spatiotemporal action tube detection system. An action tube is a set of bounding connected over time, which bounds an action instance in space and time. Next, we explore the future prediction capabilities of such detection methods by extending an existing action tube into the future by regression. Later, we seek to establish that online/causal representations can achieve similar performance to that of offline three dimensional (3D) convolutional neural networks (CNNs) on various tasks, including action recognition, temporal action segmentation and early prediction. | ['Gurkirt Singh'] | 2020-08-31 | null | null | null | null | ['action-understanding'] | ['computer-vision'] | [ 5.02634943e-01 1.27823316e-02 -6.20457768e-01 -2.23899916e-01
-3.81739199e-01 -4.45086986e-01 5.59385478e-01 1.34529183e-02
-5.03097177e-02 5.26351750e-01 5.74047983e-01 -3.60750675e-01
-3.74790907e-01 -5.82552314e-01 -7.55311489e-01 -3.10061723e-01
-6.09820664e-01 1.08282985e-02 6.29798889e-01 1.33932471e-01
3.42982799e-01 7.68584192e-01 -1.38711548e+00 6.89035058e-01
8.12045857e-02 1.22980547e+00 -3.38332325e-01 1.13841724e+00
1.84076890e-01 1.55898523e+00 -4.23595250e-01 1.32690862e-01
3.89505923e-01 -5.53611457e-01 -1.06421638e+00 1.93884417e-01
1.41715124e-01 -7.23565698e-01 -9.36096430e-01 5.48759580e-01
-1.33728504e-01 4.80418950e-01 3.55886668e-01 -1.61341035e+00
-4.45892066e-01 3.97633642e-01 -4.48144943e-01 7.74031639e-01
8.74232292e-01 2.93188840e-01 6.60050869e-01 -4.37401414e-01
9.63929772e-01 1.16416526e+00 7.40836024e-01 4.23410296e-01
-8.63090515e-01 -2.32760429e-01 4.36903775e-01 5.23206234e-01
-8.40287447e-01 -3.53477359e-01 5.75845003e-01 -7.97961831e-01
1.74307358e+00 1.36876181e-01 9.70396638e-01 1.19768226e+00
4.33272511e-01 1.25275755e+00 5.56551278e-01 -2.10301429e-01
1.72643200e-01 -7.26926684e-01 4.63419631e-02 7.03628063e-01
-5.41583121e-01 2.75283605e-01 -8.20133507e-01 1.74184352e-01
1.29649723e+00 4.46598195e-02 8.23608115e-02 -2.25884467e-01
-1.28677845e+00 5.88284612e-01 1.22837082e-01 3.23743105e-01
-3.29953104e-01 6.22022152e-01 6.74727201e-01 1.33764699e-01
7.84083545e-01 2.06985891e-01 -8.07018042e-01 -8.89494061e-01
-8.98926735e-01 2.09331125e-01 5.38506329e-01 7.93736577e-01
2.30525941e-01 -1.66538760e-01 -4.26761568e-01 7.53023252e-02
3.62687409e-02 -3.14311922e-01 4.31115866e-01 -1.44892478e+00
5.63044965e-01 8.75156641e-01 2.50860006e-01 -1.02102840e+00
-4.67469513e-01 3.96524340e-01 -2.98789233e-01 1.78150699e-01
6.65314615e-01 -2.85599697e-02 -9.16824579e-01 1.41273904e+00
4.69876438e-01 7.91548371e-01 -1.25661239e-01 7.43969738e-01
3.00920486e-01 7.38982201e-01 3.28104615e-01 -4.45308447e-01
7.98945963e-01 -8.76081169e-01 -8.44845891e-01 -7.04823360e-02
1.24350631e+00 -1.08475327e-01 5.87549269e-01 4.65252459e-01
-1.21363389e+00 -6.35160148e-01 -1.00269902e+00 -3.12075317e-01
-6.19334698e-01 1.80029154e-01 9.86131728e-01 2.75667518e-01
-8.46911907e-01 1.05936122e+00 -1.60115731e+00 -4.88922387e-01
8.31539571e-01 4.30924654e-01 -4.11808282e-01 4.14538682e-01
-1.11495149e+00 7.78141797e-01 3.91337365e-01 -1.32421153e-02
-1.11869109e+00 -8.01210821e-01 -8.92802238e-01 -4.89262283e-01
5.87296844e-01 -3.57302427e-01 1.51739478e+00 -1.22211373e+00
-1.50457478e+00 8.40420067e-01 -4.12540495e-01 -9.48942363e-01
6.85985148e-01 -7.51705289e-01 -3.05222750e-01 4.05184686e-01
1.43755183e-01 4.84714299e-01 6.02269888e-01 -3.02359998e-01
-8.49295855e-01 -4.85545576e-01 4.95827824e-01 1.35813087e-01
1.05378369e-03 5.17179728e-01 -4.73039359e-01 -5.32604814e-01
1.57441109e-01 -6.50389671e-01 -4.30133104e-01 2.44163081e-01
-2.25983754e-01 -5.04004002e-01 9.90490496e-01 -7.15792418e-01
1.30245101e+00 -2.03004336e+00 3.53587836e-01 -3.04473937e-01
4.43922505e-02 4.34911922e-02 1.96686864e-01 2.84278661e-01
-5.63235343e-01 1.63814366e-01 1.65048867e-01 3.06848116e-04
-1.77666426e-01 1.21209919e-01 -4.55400795e-01 7.22336352e-01
3.15091550e-01 1.16496646e+00 -9.78967845e-01 -7.18369842e-01
5.22170722e-01 3.31501663e-02 -2.60731876e-01 1.06357902e-01
-5.97560525e-01 6.39142632e-01 -7.44941950e-01 8.07044566e-01
-1.82217404e-01 -2.16793790e-01 -2.77701318e-02 2.28175312e-01
-3.48094910e-01 3.09262782e-01 -1.05130088e+00 1.94072664e+00
-4.15088348e-02 9.88717735e-01 -3.75664890e-01 -1.26245844e+00
5.68982124e-01 4.46806103e-01 1.54727316e+00 -7.77658403e-01
1.69099867e-02 -3.34784120e-01 -3.91647577e-01 -9.21062171e-01
4.22802240e-01 1.62059098e-01 -9.24281254e-02 6.22775078e-01
-1.37657076e-02 5.31744480e-01 3.67471188e-01 1.91224039e-01
1.55788445e+00 9.97433662e-01 2.59420484e-01 2.91628033e-01
3.71220112e-01 6.44125342e-01 4.97290790e-01 5.78394949e-01
-7.14677453e-01 3.46695065e-01 8.92803848e-01 -7.44105101e-01
-9.76146579e-01 -7.67740607e-01 1.58632383e-01 1.48879528e+00
2.19600871e-02 -5.13769209e-01 -7.13796914e-01 -8.11762452e-01
4.39483263e-02 3.71898264e-01 -8.99381042e-01 -2.18185708e-02
-9.22607183e-01 -2.09183320e-01 7.44891405e-01 1.44736803e+00
5.56938529e-01 -9.48928416e-01 -9.78647292e-01 2.88393021e-01
3.91530665e-03 -1.31553209e+00 -1.88267156e-01 -1.33354659e-03
-1.24061155e+00 -1.41176927e+00 -2.90144295e-01 -3.02024007e-01
1.63006842e-01 2.07544994e-02 8.41010630e-01 -2.97933728e-01
-5.55851638e-01 9.21814382e-01 -4.29110706e-01 -2.63519168e-01
-7.86533430e-02 -3.69631678e-01 1.92719802e-01 3.68987843e-02
7.36053586e-01 -5.25522292e-01 -6.30682290e-01 4.36966717e-01
-5.95564485e-01 1.75580874e-01 4.22160216e-02 1.52418137e-01
6.83182299e-01 1.57267064e-01 5.11843264e-02 -3.92297685e-01
1.76595062e-01 -6.17276192e-01 -5.43436468e-01 3.04845423e-01
-5.95731176e-02 -3.54580075e-01 2.79749900e-01 -6.56227648e-01
-1.14776099e+00 6.76828206e-01 2.87716359e-01 -8.34174991e-01
-4.51886028e-01 3.01462084e-01 4.27514426e-02 3.08219463e-01
6.91490531e-01 9.73634198e-02 -1.12774335e-01 -3.84348452e-01
5.70322990e-01 2.83208996e-01 4.88913834e-01 -2.86166966e-01
2.01814383e-01 8.35512757e-01 2.23123252e-01 -5.59262395e-01
-1.00013721e+00 -6.15098774e-01 -1.55862343e+00 -8.20771873e-01
1.37971342e+00 -8.19974720e-01 -9.05460835e-01 3.80059421e-01
-1.17629564e+00 -8.76444995e-01 -4.11515355e-01 4.67352718e-01
-1.22068262e+00 2.73277938e-01 -6.87031746e-01 -1.00668204e+00
4.52821881e-01 -6.76872551e-01 1.24139059e+00 1.83195040e-01
-4.65221405e-01 -1.12128031e+00 1.86992422e-01 5.32609105e-01
-2.60964215e-01 6.89857841e-01 4.33516532e-01 -5.44488251e-01
-6.78490460e-01 -4.70069677e-01 -1.97726432e-02 9.31693763e-02
-6.56267768e-03 2.36098647e-01 -6.57333612e-01 2.75012434e-01
-1.56865910e-01 -3.48028332e-01 6.94663942e-01 9.01772439e-01
1.57533813e+00 -1.10008545e-01 -6.44159198e-01 6.16884053e-01
1.12939990e+00 6.18577898e-01 7.56450534e-01 3.64381701e-01
5.65372705e-01 6.21097803e-01 1.21060848e+00 6.29634082e-01
-6.47791550e-02 5.43766141e-01 3.73811692e-01 1.08381964e-01
1.08965188e-02 -6.22127235e-01 5.30423343e-01 -2.24390596e-01
-5.19743800e-01 -9.59121808e-02 -9.95321214e-01 5.37184060e-01
-2.48589206e+00 -1.29963136e+00 -3.35408777e-01 1.83979225e+00
4.86993521e-01 9.33259279e-02 5.13269067e-01 1.16245389e-01
5.89488685e-01 3.07618707e-01 -8.32871020e-01 -4.03591514e-01
3.19730073e-01 5.06446920e-02 6.07396603e-01 1.94947109e-01
-1.77318072e+00 1.20296967e+00 7.77227592e+00 4.90609646e-01
-6.63307071e-01 -1.56807937e-02 8.36946070e-01 -3.36278886e-01
6.44836843e-01 -4.50924449e-02 -7.14879692e-01 -1.18689546e-02
1.25399351e+00 -1.52218893e-01 4.17876653e-02 1.09572160e+00
9.67904866e-01 -4.66866076e-01 -1.56839097e+00 8.33003104e-01
-6.19946495e-02 -1.61233211e+00 -3.29254180e-01 6.03333600e-02
4.57855880e-01 -2.69385487e-01 -3.45839351e-01 2.70385593e-01
3.03025395e-01 -1.25075853e+00 4.57462966e-01 7.93858886e-01
4.74544138e-01 -3.35230559e-01 7.26416558e-02 4.49568063e-01
-1.33306491e+00 -5.09671926e-01 9.76217445e-03 -6.64514244e-01
3.72168303e-01 7.05452487e-02 -4.78965312e-01 2.83621609e-01
7.14573920e-01 1.38245869e+00 -3.65944892e-01 8.65066230e-01
-1.35413989e-01 4.16964352e-01 -1.47889137e-01 2.50820015e-02
4.16304111e-01 -1.43556193e-01 4.68610615e-01 1.11654401e+00
1.86939761e-02 7.55005717e-01 2.81183213e-01 8.80218506e-01
3.61062855e-01 -4.24607366e-01 -1.06562757e+00 -4.05815661e-01
-2.39385501e-01 7.32912421e-01 -8.27308357e-01 -3.87004495e-01
-4.62129742e-01 1.06380534e+00 -9.19118896e-03 3.64177912e-01
-1.32565558e+00 1.39977753e-01 9.03575242e-01 2.27671996e-01
2.02764273e-01 -7.44319201e-01 -4.59180564e-01 -9.13813353e-01
-6.35767207e-02 -3.24611604e-01 6.32343173e-01 -9.84788716e-01
-7.75842965e-01 -2.83749551e-01 4.69249845e-01 -1.37395036e+00
-4.32485193e-01 -9.89247024e-01 -5.81321478e-01 2.26759464e-01
-6.68631196e-01 -9.44869459e-01 2.52316147e-02 9.38999951e-01
1.05022407e+00 1.31513044e-01 5.02110660e-01 -1.08337931e-01
-8.66992831e-01 -1.09096952e-01 -2.34269932e-01 5.06761491e-01
-6.15249760e-02 -1.19732332e+00 2.95374066e-01 9.80648160e-01
3.82954240e-01 2.21107379e-01 4.05732721e-01 -1.04234409e+00
-1.57714963e+00 -1.16323352e+00 5.77744246e-01 -9.36929762e-01
1.09698009e+00 -9.20758769e-02 -5.24031103e-01 1.29073656e+00
-2.26808608e-01 1.90861106e-01 6.56017900e-01 1.18919834e-01
-9.73155815e-03 2.56494373e-01 -7.12797344e-01 6.79507196e-01
1.53930366e+00 -4.58037734e-01 -6.80179298e-01 8.23992908e-01
6.02812827e-01 -7.47844756e-01 -1.18475056e+00 3.44606221e-01
6.18594944e-01 -9.54673529e-01 1.07933414e+00 -1.51551533e+00
7.92784035e-01 -9.64793842e-03 6.90581277e-02 -3.33860189e-01
-2.79852450e-01 -9.36347842e-01 -7.22480476e-01 6.32769227e-01
1.29596785e-01 6.77452236e-02 1.29796314e+00 9.29106116e-01
-1.72679439e-01 -9.72338080e-01 -1.09208965e+00 -9.26497638e-01
-9.27087143e-02 -1.06191862e+00 1.64670646e-01 7.04915166e-01
4.46343899e-01 -2.93397844e-01 -9.20832530e-02 -7.67801749e-03
1.21906132e-01 5.93459196e-02 6.59196019e-01 -7.67063737e-01
-2.01410592e-01 -4.46129948e-01 -8.39771152e-01 -1.52678311e+00
6.88867792e-02 -2.87560970e-01 6.39551654e-02 -1.52464378e+00
-8.86177421e-02 4.16690290e-01 -2.52545863e-01 7.19060600e-01
4.60753083e-01 5.18092066e-02 -1.19834900e-01 6.69015050e-02
-8.85473609e-01 2.83719003e-01 1.11621153e+00 1.97965726e-02
-4.71832603e-01 2.31490567e-01 9.29073095e-02 1.06872189e+00
7.61953294e-01 -1.53107256e-01 -5.23636103e-01 -2.68852502e-01
5.64188436e-02 5.75585544e-01 5.52812517e-01 -9.77915525e-01
3.80158007e-01 -7.57591307e-01 5.76017380e-01 -6.68967783e-01
4.32130218e-01 -4.25593436e-01 -2.66966999e-01 2.73016542e-01
-6.24572933e-01 -7.11951479e-02 2.57634521e-01 8.85547221e-01
-1.02057628e-01 2.19871193e-01 3.54679078e-01 -4.28494811e-01
-1.25834846e+00 4.79436398e-01 -6.88198149e-01 -1.50614038e-01
1.67132509e+00 -9.72477913e-01 -9.59514007e-02 -2.45891929e-01
-1.37197733e+00 1.46538839e-01 -6.93006590e-02 5.67070484e-01
8.11474323e-01 -1.12585163e+00 -2.07092524e-01 -1.03177108e-01
-2.24012613e-01 -3.31618160e-01 9.10935029e-02 1.25239325e+00
-6.06314659e-01 6.79675758e-01 -1.44687608e-01 -7.15096295e-01
-1.32516825e+00 6.42787337e-01 6.58750772e-01 -1.37120441e-01
-9.39463377e-01 8.72318268e-01 -1.05371051e-01 3.41662884e-01
5.92385948e-01 -4.92771477e-01 -3.66191417e-01 2.97672627e-03
4.78653669e-01 6.63492024e-01 -4.84612375e-01 -5.44400632e-01
-4.29680407e-01 3.91128749e-01 3.79405707e-01 -2.09255651e-01
1.36967468e+00 5.44316582e-02 1.04072794e-01 9.05793846e-01
1.04920852e+00 -8.93481553e-01 -1.75998425e+00 3.83320004e-01
4.80443150e-01 -5.47939837e-01 2.03420948e-02 -5.65315247e-01
-8.85322034e-01 6.87969744e-01 4.92149889e-01 5.16300440e-01
1.13978052e+00 2.02749610e-01 7.55378425e-01 2.49364793e-01
9.61245373e-02 -1.58405197e+00 4.72453624e-01 5.41279733e-01
7.62936532e-01 -1.14760065e+00 1.86209187e-01 -3.73293221e-01
-5.80762982e-01 1.40986001e+00 7.46721208e-01 -3.37747455e-01
8.30958784e-01 1.95269123e-01 -5.29344857e-01 -5.43102205e-01
-9.46286976e-01 -3.46331507e-01 1.38541773e-01 5.93746305e-01
3.07347000e-01 -2.26272836e-01 -1.00740142e-01 3.92401725e-01
3.28044772e-01 5.08161426e-01 3.41778785e-01 1.15032458e+00
-3.84581059e-01 -7.10829914e-01 -1.40533060e-01 2.90853202e-01
-4.65552300e-01 4.21008080e-01 -7.14343429e-01 1.01412618e+00
2.60289103e-01 1.05769694e+00 5.05638778e-01 -5.06918013e-01
2.29626194e-01 2.66067326e-01 7.15998054e-01 -6.33891642e-01
-2.40193233e-01 -7.28643015e-02 2.80907422e-01 -1.45310926e+00
-7.93737233e-01 -1.11650538e+00 -1.42516363e+00 1.18036065e-02
2.52029151e-02 -7.14113533e-01 4.42114681e-01 1.30933571e+00
2.82127023e-01 5.11445403e-01 5.15490413e-01 -7.38115013e-01
-1.26974031e-01 -8.38515162e-01 -4.08938259e-01 2.91603357e-01
9.48096141e-02 -7.59290218e-01 -1.64615437e-01 6.07370675e-01] | [8.234606742858887, 0.5688027143478394] |
b202975a-a188-43bc-8a29-3408e4a2e528 | domain-adapting-speech-emotion-recognition | 2207.12248 | null | https://arxiv.org/abs/2207.12248v2 | https://arxiv.org/pdf/2207.12248v2.pdf | Domain Adapting Deep Reinforcement Learning for Real-world Speech Emotion Recognition | Computers can understand and then engage with people in an emotionally intelligent way thanks to speech-emotion recognition (SER). However, the performance of SER in cross-corpus and real-world live data feed scenarios can be significantly improved. The inability to adapt an existing model to a new domain is one of the shortcomings of SER methods. To address this challenge, researchers have developed domain adaptation techniques that transfer knowledge learnt by a model across the domain. Although existing domain adaptation techniques have improved performances across domains, they can be improved to adapt to a real-world live data feed situation where a model can self-tune while deployed. In this paper, we present a deep reinforcement learning-based strategy (RL-DA) for adapting a pre-trained model to a real-world live data feed setting while interacting with the environment and collecting continual feedback. RL-DA is evaluated on SER tasks, including cross-corpus and cross-language domain adaption schema. Evaluation results show that in a live data feed setting, RL-DA outperforms a baseline strategy by 11% and 14% in cross-corpus and cross-language scenarios, respectively. | ['Bjorn W. Schuller', 'Sara Khalifa', 'Rajib Rana', 'Thejan Rajapakshe'] | 2022-07-07 | null | null | null | null | ['cross-corpus'] | ['computer-vision'] | [-1.46782532e-01 -1.05717599e-01 1.21667311e-01 -6.56604469e-01
-5.56823671e-01 -6.01012528e-01 3.93428445e-01 6.72115013e-03
-4.81192738e-01 7.75171161e-01 1.04889378e-01 3.60725164e-01
3.05015743e-01 -4.98024583e-01 -4.36356455e-01 -2.68912703e-01
1.02451421e-01 8.62962604e-01 -8.85200351e-02 -7.04959512e-01
-1.56020761e-01 3.74361783e-01 -1.41679323e+00 4.43614066e-01
7.07567632e-01 6.78608060e-01 8.41058344e-02 8.20047259e-01
-3.25290293e-01 6.26200140e-01 -1.06016731e+00 -3.61959755e-01
-1.65490448e-01 -5.18436372e-01 -1.08427072e+00 1.05823204e-01
-1.27494901e-01 -1.39149413e-01 6.83981180e-02 6.34977281e-01
9.69731987e-01 6.26353860e-01 2.89802134e-01 -1.33502805e+00
-6.18608057e-01 2.07059592e-01 1.75261814e-02 4.28735279e-02
8.21319699e-01 2.48558328e-01 2.10351810e-01 -4.86470133e-01
6.82179570e-01 1.39741385e+00 6.89042747e-01 1.37077522e+00
-1.27613807e+00 -6.56034768e-01 3.24859142e-01 2.12033525e-01
-9.73342657e-01 -6.08289301e-01 1.01718235e+00 -2.34610155e-01
1.45035696e+00 7.41206808e-03 6.61479235e-01 1.76749063e+00
-1.98227525e-01 7.84456432e-01 1.28991830e+00 -4.25664872e-01
6.29266381e-01 8.09396267e-01 -1.68944851e-01 6.55966699e-02
-5.77665806e-01 2.90709257e-01 -9.44196820e-01 -1.56462133e-01
2.15594843e-01 -4.99340236e-01 -7.81787857e-02 -9.71631333e-02
-8.05559039e-01 7.45276034e-01 -3.18228453e-02 5.85252583e-01
-4.58596140e-01 -4.50159639e-01 9.55464542e-01 8.79533648e-01
8.31652761e-01 8.35613430e-01 -8.70163977e-01 -8.60428989e-01
-5.95003486e-01 3.94257635e-01 1.21423805e+00 6.34779215e-01
5.24352372e-01 2.19713479e-01 1.09320253e-01 1.31658852e+00
-5.44503927e-02 5.07739067e-01 1.00351274e+00 -7.08993495e-01
2.32355893e-01 6.15660667e-01 1.32469580e-01 -5.97103000e-01
-5.17119408e-01 -1.98037922e-01 -6.98261082e-01 1.37624219e-01
1.03981167e-01 -8.00714374e-01 -4.56840277e-01 1.91726661e+00
6.09681368e-01 2.40618929e-01 6.78112924e-01 7.84256458e-01
8.00735295e-01 7.97840595e-01 4.81116861e-01 -2.88775086e-01
1.07860601e+00 -8.59386802e-01 -9.14128602e-01 -7.50519753e-01
4.85540569e-01 -4.10300583e-01 1.38724113e+00 6.59432530e-01
-9.58882332e-01 -6.35988414e-01 -8.43042195e-01 3.18059653e-01
-4.18605983e-01 -3.56505960e-01 3.18951815e-01 6.73098147e-01
-9.52526808e-01 1.37309670e-01 -5.75594842e-01 -9.99406040e-01
-4.07613628e-02 3.45710009e-01 -5.22605360e-01 1.64678037e-01
-1.56471789e+00 1.04442263e+00 4.23762321e-01 -4.45957989e-01
-6.58575952e-01 -6.98496342e-01 -6.44562840e-01 -1.56840071e-01
2.01222852e-01 -5.75099289e-01 1.71762168e+00 -1.87540448e+00
-2.23230267e+00 9.94773984e-01 1.15311839e-01 -4.89378422e-01
4.03747022e-01 -2.97125667e-01 -1.14222455e+00 9.24858078e-03
-3.82856220e-01 5.51799536e-01 7.53929794e-01 -1.22166502e+00
-5.59796989e-01 -1.29215717e-01 3.06658447e-02 4.51141566e-01
-4.77199286e-01 2.67978877e-01 3.75079885e-02 -4.26385671e-01
-7.62548685e-01 -8.33426952e-01 2.12951116e-02 -4.96022940e-01
4.39923495e-01 -4.48856890e-01 9.26996589e-01 -5.50826550e-01
1.13425672e+00 -2.28242230e+00 1.90655068e-01 -1.28331676e-01
-4.35092509e-01 7.14466155e-01 -3.02726835e-01 5.77635288e-01
-2.60433227e-01 -1.64214298e-01 3.83660849e-03 -3.72203290e-01
3.01212631e-02 5.10820985e-01 -8.22905172e-03 -1.41704053e-01
5.29306829e-01 6.00454330e-01 -9.00750875e-01 -1.79657280e-01
1.20958008e-01 4.62854534e-01 -6.37188375e-01 9.24958706e-01
-3.44107836e-01 6.00899994e-01 -2.63963640e-01 1.87541351e-01
5.13433695e-01 1.38133049e-01 1.83397546e-01 2.94101596e-01
2.56335676e-01 8.77084509e-02 -1.12313724e+00 1.84930313e+00
-8.68334174e-01 4.65837359e-01 1.66034564e-01 -1.12718511e+00
1.47666740e+00 6.37718856e-01 3.91092211e-01 -1.15663922e+00
2.58686431e-02 -1.44412294e-01 -1.79355592e-01 -1.00907636e+00
4.60463762e-01 -8.46209466e-01 -5.45882463e-01 4.78641987e-01
3.28406721e-01 -6.74513355e-02 -2.18846455e-01 2.24387664e-02
1.14042926e+00 1.71330236e-02 3.87841493e-01 2.26312906e-01
6.36531949e-01 -1.78677384e-02 6.32414937e-01 3.50361735e-01
-7.95688510e-01 1.70765474e-01 2.47049123e-01 -5.89675546e-01
-6.63306415e-01 -7.62075603e-01 3.61375183e-01 1.83908188e+00
-2.02187076e-01 -5.84270433e-02 -9.09474969e-01 -7.55101264e-01
-2.55385309e-01 1.16601086e+00 -4.93778735e-01 -8.26187968e-01
-3.71041059e-01 -4.45552617e-01 6.15881741e-01 4.69699472e-01
7.76181698e-01 -1.53147137e+00 -9.13757682e-01 7.04342246e-01
-5.75366020e-01 -1.25152290e+00 -1.90691680e-01 3.48519921e-01
-4.33503985e-01 -4.96759057e-01 -3.17033023e-01 -7.05926538e-01
2.46269196e-01 -3.41177404e-01 1.39756525e+00 -2.89544642e-01
9.40637141e-02 9.02214825e-01 -6.06250882e-01 -5.70291519e-01
-8.63988817e-01 1.51197508e-01 2.70527869e-01 5.37370220e-02
9.19837058e-01 -5.25796592e-01 -2.22440898e-01 3.55872273e-01
-7.88687825e-01 -1.01158850e-01 2.46972919e-01 8.65473509e-01
7.95900449e-02 2.24693734e-02 1.22673869e+00 -8.15886855e-01
1.23675346e+00 -6.49793327e-01 8.17243084e-02 3.82505268e-01
-3.62709224e-01 1.48239080e-02 6.13593161e-01 -9.76047754e-01
-1.69547164e+00 3.16171870e-02 -3.72442544e-01 -2.55854428e-01
-8.21863592e-01 3.98121059e-01 -3.51393342e-01 2.82513916e-01
1.03272426e+00 -1.89014599e-02 -1.08788498e-01 -3.44448715e-01
8.52600634e-02 1.15020120e+00 6.31254911e-01 -9.29781556e-01
3.70719075e-01 -1.20171122e-01 -8.93298149e-01 -3.89671624e-01
-7.09871233e-01 -4.19634879e-01 -4.92473245e-01 -4.60345119e-01
7.15960741e-01 -9.88011301e-01 -7.00367689e-01 6.03387058e-01
-1.00230277e+00 -8.75932574e-01 -4.30917978e-01 1.73639789e-01
-6.67162240e-01 -2.25343451e-01 -3.18888515e-01 -1.00772429e+00
-4.49614555e-01 -5.73524952e-01 8.13022077e-01 4.83486921e-01
-7.64801979e-01 -1.32335424e+00 6.27634406e-01 4.47314322e-01
7.00380445e-01 2.20357656e-01 5.83052754e-01 -1.20314407e+00
7.07362473e-01 -3.38789448e-02 4.00858343e-01 5.73905110e-01
-9.60539468e-03 -2.43686184e-01 -1.10352647e+00 -3.56736481e-01
1.12404391e-01 -1.04974282e+00 -1.22208120e-02 -3.33930403e-01
8.85274708e-01 -3.42541575e-01 -1.97691359e-02 2.61048377e-01
9.85393226e-01 4.95162845e-01 5.35385787e-01 6.25368178e-01
-1.15075067e-01 7.53574669e-01 8.48949790e-01 7.07470059e-01
6.43930435e-01 7.69248545e-01 -7.54253194e-02 -2.80456673e-02
7.74628520e-02 -3.25032532e-01 9.50312436e-01 6.52710736e-01
2.84342349e-01 -2.81502008e-01 -1.01283050e+00 5.89838743e-01
-1.82725251e+00 -1.07124639e+00 5.07477105e-01 1.83949292e+00
1.20519865e+00 -5.57751860e-03 6.21751368e-01 -1.11903675e-01
6.23234749e-01 -2.84408391e-01 -8.65880132e-01 -1.27843642e+00
-4.17460389e-02 4.59259421e-01 -2.61530042e-01 3.57960373e-01
-9.83661711e-01 1.02857912e+00 6.17940235e+00 4.94948000e-01
-1.60304403e+00 2.72064805e-01 5.62249005e-01 -2.73276955e-01
5.94859421e-02 -6.73328459e-01 -3.97511214e-01 4.23562795e-01
1.71509767e+00 -5.40733337e-01 6.95248246e-01 1.00399637e+00
1.69178143e-01 1.05455637e-01 -1.28394020e+00 1.25684118e+00
2.68319070e-01 -5.96693337e-01 -4.75535542e-01 -4.63959873e-01
4.60932106e-01 1.37846339e-02 6.72098920e-02 1.14622748e+00
4.56020176e-01 -8.77176523e-01 4.50894117e-01 3.96142155e-01
7.07659543e-01 -1.12035608e+00 7.11104214e-01 6.64083123e-01
-6.26602769e-01 -6.19664229e-02 -7.33642355e-02 -1.47888973e-01
7.57705122e-02 1.62588302e-02 -1.21058691e+00 2.62729555e-01
1.03623414e+00 4.30351883e-01 -3.78460675e-01 4.44200307e-01
-4.00032066e-02 7.64051855e-01 -1.91871747e-01 -7.35363141e-02
-7.46898428e-02 7.48882368e-02 5.26848793e-01 1.60698807e+00
1.52425587e-01 2.78973788e-01 1.17219307e-01 5.64541757e-01
1.29739689e-02 9.18507576e-02 -5.69806755e-01 -1.85120732e-01
3.83978575e-01 1.20305431e+00 7.68373832e-02 -3.47745687e-01
-1.90960482e-01 1.40097332e+00 5.54495454e-01 4.16460931e-01
-7.32832849e-01 -2.65029013e-01 8.79017234e-01 6.05103038e-02
-9.05317962e-02 3.38514596e-02 1.23823062e-01 -1.10364604e+00
-1.93290100e-01 -1.51997757e+00 4.91134942e-01 -9.16175425e-01
-1.65152335e+00 9.04000580e-01 5.73989898e-02 -9.48629022e-01
-6.92683220e-01 -2.62393087e-01 -5.45921266e-01 5.18485010e-01
-1.42268503e+00 -1.02782857e+00 -2.49235734e-01 9.97194231e-01
8.38709474e-01 -6.21990919e-01 1.33905852e+00 2.99869627e-01
-6.13664627e-01 7.80318737e-01 -3.73934060e-02 -2.26329640e-02
1.24487162e+00 -1.23175824e+00 2.42199823e-01 4.35574085e-01
-3.07315171e-01 1.64852589e-01 9.88000929e-01 -4.33989286e-01
-1.18006706e+00 -1.08788729e+00 7.34450459e-01 -3.32706481e-01
4.13633019e-01 -6.35950506e-01 -1.48109889e+00 5.77489376e-01
6.31855905e-01 -1.07270755e-01 1.10999525e+00 3.52872580e-01
-2.17498153e-01 -2.00403020e-01 -1.73803687e+00 3.31140190e-01
8.63104701e-01 -4.61538106e-01 -8.77770483e-01 1.16078332e-01
5.88359356e-01 -5.06072044e-01 -1.07082725e+00 7.81900138e-02
3.04928780e-01 -8.55764806e-01 6.70005739e-01 -1.01020241e+00
1.65207729e-01 1.59521386e-01 9.89131704e-02 -2.13497734e+00
-1.92957431e-01 -9.78771746e-01 -3.81992459e-02 1.66399395e+00
9.23628137e-02 -7.58966744e-01 3.58387262e-01 1.09658456e+00
6.87525943e-02 -1.64924279e-01 -9.73439515e-01 -7.19738066e-01
2.49678284e-01 -5.62039733e-01 8.50457847e-01 1.15028596e+00
5.46185434e-01 7.09316313e-01 -4.19520557e-01 -4.88677658e-02
-2.73285527e-02 -4.56155658e-01 1.07493377e+00 -1.23182786e+00
-4.76504028e-01 -1.01623885e-01 -1.18011400e-01 -5.65586209e-01
6.38148844e-01 -5.89323759e-01 1.61204517e-01 -1.11192179e+00
-1.90906644e-01 -2.71009654e-01 -3.04238379e-01 5.83892047e-01
-8.72276425e-02 -3.69588912e-01 2.16568023e-01 -3.07879657e-01
-8.03171575e-01 8.39440167e-01 7.55298674e-01 -6.04779087e-03
-6.02927625e-01 -8.90278667e-02 -6.66143537e-01 5.01527846e-01
1.11540449e+00 -4.83280838e-01 -5.18939078e-01 -2.44099334e-01
2.89031714e-01 1.06012717e-01 9.67841074e-02 -1.20915818e+00
1.15083739e-01 -4.22250867e-01 2.74846584e-01 1.05755091e-01
3.36741537e-01 -1.05009747e+00 1.80593997e-01 5.66896796e-02
-4.99324590e-01 -6.41933680e-02 8.38899136e-01 4.38968480e-01
-3.22392970e-01 -8.13990012e-02 9.31836784e-01 2.65694372e-02
-1.04103220e+00 -1.52127042e-01 -5.76906562e-01 3.01616490e-01
1.41115892e+00 -2.14927331e-01 -2.30491474e-01 -6.04498029e-01
-9.18456852e-01 3.50296497e-01 3.45173508e-01 8.85455072e-01
5.41019499e-01 -1.43649399e+00 -7.48240829e-01 4.36605811e-01
4.30267870e-01 -1.64715692e-01 4.13635194e-01 3.82421911e-01
1.67386769e-03 1.05326409e-02 -5.27013481e-01 -4.76414353e-01
-1.52386105e+00 6.04406893e-01 8.68904293e-01 -3.48869979e-01
-3.33564371e-01 8.05717945e-01 -5.51488101e-02 -8.87704015e-01
2.29319170e-01 2.27908641e-02 -1.11255430e-01 -2.11718231e-02
6.64737642e-01 9.61254835e-02 1.52264729e-01 -4.92346823e-01
-3.67561221e-01 1.17712542e-01 -9.67578068e-02 -3.48824382e-01
1.57325029e+00 -3.44477803e-01 3.64214361e-01 7.59408057e-01
9.89184618e-01 -4.87818480e-01 -1.34892237e+00 -5.19081414e-01
-2.89124344e-02 -1.19890399e-01 -1.51843056e-01 -1.59204781e+00
-8.55600476e-01 8.23912323e-01 9.47263002e-01 3.97148952e-02
1.57758653e+00 -1.69692691e-02 7.84791470e-01 5.34910083e-01
2.04925522e-01 -1.83614480e+00 5.79605877e-01 5.29351771e-01
1.16780782e+00 -1.27859330e+00 -5.70201218e-01 3.49906415e-01
-1.52223051e+00 1.04468536e+00 1.15903485e+00 1.02560498e-01
4.52725023e-01 4.15664941e-01 3.83361161e-01 5.16843684e-02
-1.43549824e+00 -1.69315271e-03 1.29048908e-02 1.21768117e+00
4.40548778e-01 1.38961643e-01 2.71627933e-01 9.69902277e-01
-2.26735622e-01 3.93053412e-01 2.20929474e-01 8.25805783e-01
-2.97719687e-01 -1.38493943e+00 -5.04786909e-01 -1.80180013e-01
-1.60208374e-01 5.57231545e-01 -8.34902823e-01 6.37675107e-01
2.49098912e-01 1.26752794e+00 1.73542142e-01 -3.93494934e-01
9.12089944e-01 7.17159986e-01 2.33386010e-01 -5.25475800e-01
-1.24275482e+00 -2.09047094e-01 4.43368614e-01 -4.11522597e-01
-5.78572333e-01 -8.45420718e-01 -1.40210605e+00 -2.65426368e-01
1.54590458e-01 2.47654215e-01 5.31702399e-01 9.65399265e-01
5.83188415e-01 5.05956590e-01 7.53524184e-01 -4.43057269e-01
-3.75389189e-01 -1.02750826e+00 -2.79675543e-01 8.99178624e-01
2.19926447e-01 -4.73831862e-01 -8.56311917e-02 2.78461009e-01] | [13.497001647949219, 5.9021172523498535] |
3f5ccaa0-e37b-41c6-a655-9850f65a3a32 | covidlies-detecting-covid-19-misinformation | null | null | https://aclanthology.org/2020.nlpcovid19-2.11 | https://aclanthology.org/2020.nlpcovid19-2.11.pdf | COVIDLies: Detecting COVID-19 Misinformation on Social Media | The ongoing pandemic has heightened the need for developing tools to flag COVID-19-related misinformation on the internet, specifically on social media such as Twitter. However, due to novel language and the rapid change of information, existing misinformation detection datasets are not effective for evaluating systems designed to detect misinformation on this topic. Misinformation detection can be divided into two sub-tasks: (i) retrieval of misconceptions relevant to posts being checked for veracity, and (ii) stance detection to identify whether the posts Agree, Disagree, or express No Stance towards the retrieved misconceptions. To facilitate research on this task, we release COVIDLies (https://ucinlp.github.io/covid19 ), a dataset of 6761 expert-annotated tweets to evaluate the performance of misinformation detection systems on 86 different pieces of COVID-19 related misinformation. We evaluate existing NLP systems on this dataset, providing initial benchmarks and identifying key challenges for future models to improve upon. | ['Sameer Singh', 'Sean Young', 'Yoshitomo Matsubara', 'Arjuna Ugarte', 'Robert L. Logan IV', 'Tamanna Hossain'] | null | null | null | null | emnlp-nlp-covid19-2020-12 | ['misconceptions'] | ['miscellaneous'] | [-1.10618897e-01 3.01239878e-01 -4.84400183e-01 -2.42191106e-01
-7.61344612e-01 -9.78652716e-01 1.09254265e+00 1.00442362e+00
-3.81571293e-01 8.14285159e-01 9.82539535e-01 -6.31018102e-01
3.56751770e-01 -6.66855216e-01 -4.72257823e-01 -1.23082191e-01
6.90227896e-02 5.74832082e-01 1.81234643e-01 -4.40475404e-01
6.83064759e-01 -2.75979042e-02 -1.05876613e+00 6.79059207e-01
8.10300708e-01 5.78896582e-01 -4.09919560e-01 5.13142884e-01
-9.14974064e-02 1.16993940e+00 -8.53801429e-01 -3.65315706e-01
-2.10635737e-01 -1.52256474e-01 -1.02516007e+00 -5.02176225e-01
1.95124820e-01 -2.74860501e-01 -1.30061224e-01 1.29712486e+00
2.60131866e-01 -2.59405404e-01 7.21695006e-01 -1.21774197e+00
-5.15148938e-01 1.02706146e+00 -3.48154813e-01 9.14170682e-01
7.48981774e-01 2.38010883e-02 9.03350830e-01 -7.93157876e-01
9.65012312e-01 1.31447244e+00 8.01691592e-01 1.43999606e-01
-7.98085272e-01 -7.54337013e-01 -1.36597097e-01 5.79043142e-02
-1.11845529e+00 -5.34190655e-01 1.99940175e-01 -1.04074073e+00
8.89324963e-01 5.31731486e-01 2.90678501e-01 1.42777300e+00
3.66874278e-01 5.33398032e-01 1.33465350e+00 1.49466872e-01
1.86333343e-01 3.73832554e-01 6.07539356e-01 2.82825738e-01
9.43064332e-01 -2.12109715e-01 -3.58468473e-01 -6.47475243e-01
-2.75661886e-01 3.75660509e-02 -3.73026043e-01 8.09658527e-01
-9.90613997e-01 1.16035318e+00 4.89016920e-01 5.78489900e-01
-6.47278845e-01 -4.65197086e-01 6.12662315e-01 2.59432077e-01
1.05887520e+00 9.47624564e-01 -4.03487086e-01 -2.57364810e-01
-9.09302711e-01 4.93158847e-01 1.40776002e+00 6.40802860e-01
6.82124436e-01 -4.31955636e-01 -1.56547621e-01 7.07446337e-01
2.78293252e-01 7.79957473e-01 4.67813492e-01 -7.93525219e-01
8.49024057e-01 5.71848333e-01 5.87845862e-01 -2.00203848e+00
-7.60383904e-01 -3.91881824e-01 -5.44454873e-01 -7.61997879e-01
7.34426454e-02 -6.92103088e-01 -6.57996833e-01 1.19787633e+00
1.03539288e-01 -2.20946759e-01 -9.90146548e-02 7.39366770e-01
1.18657076e+00 9.07263279e-01 1.74240261e-01 -2.33314022e-01
1.44248116e+00 -3.91402662e-01 -9.52614427e-01 -6.20320320e-01
8.76621008e-01 -1.13252199e+00 2.70768106e-01 1.02119803e-01
-7.58054435e-01 2.82382041e-01 -8.24367225e-01 1.73727751e-01
-8.50681543e-01 -6.05740607e-01 9.19156447e-02 3.24804187e-01
-8.78420830e-01 1.29741058e-01 -2.75281519e-01 -5.07200360e-01
3.21516722e-01 -6.15486920e-01 9.50429663e-02 -6.42079487e-02
-1.80575824e+00 1.02949369e+00 4.69388485e-01 -2.45345518e-01
-8.27933013e-01 -7.17077851e-01 -7.51888335e-01 -3.54147017e-01
3.31024796e-01 -3.85205597e-01 1.33825207e+00 -7.02549994e-01
-3.73497397e-01 1.17182279e+00 -2.47026652e-01 -6.54045761e-01
6.16207182e-01 -1.60712153e-01 -9.15117741e-01 2.23668873e-01
9.46342528e-01 2.64019847e-01 7.07558692e-01 -1.13018465e+00
-7.21789241e-01 -2.24768177e-01 -1.33595705e-01 -3.72571201e-04
-1.86206609e-01 6.30779266e-01 6.49080202e-02 -5.89496911e-01
-2.16223687e-01 -8.24443936e-01 5.68435341e-02 -8.23994100e-01
-8.41202617e-01 -2.35827506e-01 7.06191897e-01 -9.84464467e-01
1.65454853e+00 -1.76889956e+00 -7.26849735e-01 1.36553228e-01
5.42727470e-01 5.83666146e-01 2.09865883e-01 8.84642541e-01
2.14739099e-01 8.19317997e-01 1.40758380e-01 9.16031897e-02
-1.32292286e-01 2.55899243e-02 -7.47968376e-01 4.40295160e-01
-1.02468185e-01 6.53529882e-01 -1.46805382e+00 -3.66397977e-01
-4.43954080e-01 1.73929393e-01 -3.40918839e-01 -1.03278577e-01
-4.83423978e-01 2.05453336e-01 -7.29043365e-01 5.06157339e-01
7.48112977e-01 -6.89801335e-01 1.06920280e-01 1.84300113e-02
-5.10605693e-01 1.16433704e+00 -4.34769183e-01 3.81553233e-01
8.55033770e-02 8.80414546e-01 2.61150062e-01 -4.11887467e-01
5.77527046e-01 3.29332381e-01 3.78435820e-01 -4.95328218e-01
3.65704060e-01 2.03646019e-01 -3.97937775e-01 -7.62909055e-01
7.60923088e-01 4.40789089e-02 -3.62300992e-01 8.74495625e-01
-7.33677745e-01 1.15393713e-01 2.72960007e-01 6.49523914e-01
1.10150921e+00 -1.06722593e+00 3.98655772e-01 -4.60136533e-01
2.87803024e-01 7.02806413e-01 3.81422669e-01 8.53994668e-01
-3.80503118e-01 1.72428280e-01 4.02325422e-01 -3.97555053e-01
-7.02106893e-01 -5.28090477e-01 -2.86627173e-01 9.09662068e-01
-3.06404084e-02 -6.02242351e-01 -5.24874389e-01 -7.60115981e-01
2.61279911e-01 1.14573550e+00 -7.20328689e-01 4.32672679e-01
-4.06514913e-01 -1.06449389e+00 8.12243938e-01 -1.51277333e-01
5.04561245e-01 -7.30246782e-01 -3.36324185e-01 3.38361591e-01
-1.03514481e+00 -1.20239997e+00 -4.00040895e-01 -3.28829497e-01
-1.72594860e-01 -1.43030691e+00 -3.30408782e-01 -3.73600870e-01
5.44977009e-01 6.60846353e-01 1.09481490e+00 2.35914752e-01
2.55998790e-01 1.58846691e-01 -4.33743328e-01 -7.17618406e-01
-9.00737822e-01 -4.14161086e-02 1.18540347e-01 -2.78015465e-01
7.09636390e-01 -4.34444984e-03 -5.96975446e-01 2.80162573e-01
-1.01251662e+00 -3.00180793e-01 1.68909840e-02 2.69141763e-01
-2.94207960e-01 -6.38109893e-02 6.13168895e-01 -1.26520896e+00
1.16711545e+00 -1.49432421e+00 -1.98298648e-01 -2.50607610e-01
-7.74351060e-01 -3.94960076e-01 3.06522578e-01 2.25386858e-01
-8.51228595e-01 -1.06340241e+00 -2.56834418e-01 4.69426125e-01
-9.63455066e-02 1.29372478e+00 9.08311248e-01 4.47794259e-01
1.04088724e+00 -1.26713961e-01 -3.24418768e-02 -3.82322252e-01
1.54185876e-01 1.33760369e+00 1.46161705e-01 1.55137599e-01
6.53209031e-01 5.87121308e-01 -7.42598057e-01 -1.03463006e+00
-1.51917255e+00 -9.85447943e-01 1.54366316e-02 -2.17954978e-01
7.20891535e-01 -1.12009966e+00 -4.70345646e-01 3.29552263e-01
-1.42700422e+00 5.38209975e-02 6.74282312e-01 1.70013174e-01
2.44807988e-01 4.76503611e-01 -9.01882648e-01 -5.07280827e-01
-5.49406171e-01 -6.23699069e-01 3.83079171e-01 -2.32274115e-01
-9.13261890e-01 -1.22447801e+00 6.00858927e-01 8.06038022e-01
7.33489037e-01 3.70857537e-01 6.37001634e-01 -1.37011933e+00
6.88976794e-02 -3.59271169e-01 -5.47677100e-01 -8.93703401e-02
2.88079321e-01 3.64035852e-02 -7.32116282e-01 -1.81688383e-01
-7.39695728e-02 -5.52768171e-01 8.78002822e-01 4.27173860e-02
4.30651188e-01 -1.42600787e+00 -6.67920530e-01 -1.50198996e-01
1.04117870e+00 -1.86369926e-01 1.79432422e-01 7.72920549e-01
2.03821108e-01 5.92086077e-01 5.75756550e-01 7.15024173e-01
8.22598279e-01 3.53507429e-01 3.29017133e-01 3.84940684e-01
1.57529593e-01 -5.29669225e-01 5.20877302e-01 9.85660195e-01
5.55774331e-01 -4.66935873e-01 -1.42877948e+00 7.49901891e-01
-1.54761863e+00 -1.12245011e+00 -4.09019113e-01 1.58091128e+00
1.09863830e+00 2.45615587e-01 9.11741406e-02 -2.08986834e-01
8.64931881e-01 6.19181752e-01 -6.44286200e-02 -5.65265775e-01
-1.29700154e-01 -8.05823863e-01 4.85779822e-01 8.89255404e-01
-1.20946074e+00 7.46128440e-01 6.70599604e+00 3.84739459e-01
-1.19427085e+00 2.77359068e-01 5.92414379e-01 3.52191508e-01
-6.81943774e-01 -2.43102968e-01 -1.11027014e+00 9.38618302e-01
1.03697360e+00 -4.11756814e-01 7.29696453e-02 5.75426400e-01
6.87459767e-01 -2.54096948e-02 -5.09349763e-01 4.22592133e-01
3.92473102e-01 -1.66518271e+00 1.23984456e-01 -2.05449045e-01
9.26679313e-01 7.98785329e-01 1.16866250e-02 2.98452944e-01
6.03647530e-01 -7.35726357e-01 5.38326383e-01 2.37409413e-01
2.82334119e-01 -1.83239505e-01 9.69677091e-01 6.18660986e-01
-2.33575568e-01 1.35225892e-01 -2.26348415e-01 -2.05690667e-01
2.59257168e-01 1.07058191e+00 -1.71017599e+00 1.28892347e-01
6.28682256e-01 1.19582140e+00 -5.97168386e-01 8.40511203e-01
-4.78859842e-01 9.80818987e-01 -1.85946479e-01 -4.10051733e-01
7.03464031e-01 3.48694593e-01 1.13877130e+00 1.87269711e+00
-7.81827495e-02 5.48660085e-02 8.11153650e-02 8.71174812e-01
-2.68923789e-01 -1.10892415e-01 -8.69058907e-01 -6.10047460e-01
9.09172475e-01 9.05457139e-01 -3.20082694e-01 -5.70752025e-01
-1.18515626e-01 1.87737897e-01 2.25881249e-01 3.51589262e-01
-4.94089603e-01 -6.60270527e-02 5.01614511e-01 4.40796047e-01
-5.63166812e-02 -1.85396168e-02 -2.17901513e-01 -1.19390666e+00
-4.08423871e-01 -1.10917234e+00 8.23037624e-01 -6.77955866e-01
-1.36559045e+00 5.62405586e-01 -2.13303976e-02 -9.35159206e-01
-4.69422668e-01 -1.97875202e-01 -6.00332201e-01 4.94860947e-01
-1.62771142e+00 -6.10423386e-01 -2.19959006e-01 1.32145196e-01
3.23301941e-01 3.34275663e-01 5.16389787e-01 2.40780428e-01
-5.98387420e-01 1.43930748e-01 2.68948488e-02 3.78093719e-01
1.06715930e+00 -6.98662579e-01 4.13146824e-01 6.54943764e-01
-5.97527504e-01 7.24976838e-01 1.28127277e+00 -1.28752065e+00
-8.41330469e-01 -1.50841796e+00 1.77740276e+00 -9.24832463e-01
1.28920174e+00 1.69204902e-02 -9.00238752e-01 1.00189102e+00
4.08732653e-01 -9.16299760e-01 8.12929273e-01 9.55672711e-02
-6.91195667e-01 3.98776144e-01 -1.19268191e+00 5.66583753e-01
5.19255996e-01 -3.28601897e-01 -1.13471639e+00 8.93961608e-01
7.03990281e-01 -1.36622787e-01 -6.59520984e-01 4.51988801e-02
1.29699156e-01 -6.68223679e-01 6.68677568e-01 -8.64862025e-01
5.64141810e-01 -6.66739717e-02 7.40477890e-02 -1.51903391e+00
-4.24705327e-01 -8.19755316e-01 1.78012535e-01 1.02418995e+00
9.03657734e-01 -1.07852364e+00 1.65753081e-01 5.06751776e-01
2.61670351e-02 -2.20890552e-01 -5.87741137e-01 -4.63557303e-01
5.28920852e-02 -3.53121221e-01 2.88127691e-01 1.51091325e+00
6.16113186e-01 4.52074707e-01 -2.34054357e-01 3.14286441e-01
4.92371142e-01 5.58134839e-02 4.35019642e-01 -9.90483642e-01
3.29453856e-01 -5.12880266e-01 1.08872019e-01 -8.51286232e-01
-9.69774798e-02 -9.10527527e-01 5.29061258e-02 -1.57219255e+00
4.66561377e-01 -4.82880801e-01 3.10168769e-02 3.73272747e-01
-2.83487231e-01 2.49352038e-01 -1.13398388e-01 7.26778984e-01
-9.86018479e-01 -2.21301569e-03 7.91258633e-01 -4.04852420e-01
-1.81185510e-02 -1.70564651e-01 -1.20626330e+00 1.02468586e+00
9.59692299e-01 -8.21725368e-01 1.55399829e-01 -3.51819813e-01
1.14742529e+00 -2.32581571e-01 2.05029398e-01 -5.27564287e-01
3.37082624e-01 -2.56771803e-01 -3.48643474e-02 -7.71071315e-01
-5.13167605e-02 -1.06060922e-01 -1.93509370e-01 7.40624905e-01
-5.89986980e-01 2.96192497e-01 1.63050205e-01 4.68161672e-01
-3.18088830e-01 -9.23438668e-02 5.45096099e-01 -2.47046918e-01
-3.83567154e-01 -1.38622541e-02 -1.24279177e+00 9.05613601e-01
5.38848400e-01 4.71487015e-01 -1.35394084e+00 -8.53491008e-01
-4.16625619e-01 4.22147989e-01 3.93449754e-01 5.16392112e-01
5.14776766e-01 -9.18269336e-01 -1.23711836e+00 -2.34628707e-01
3.40517074e-01 -6.07843995e-01 2.95880232e-02 1.14969313e+00
-4.74522799e-01 1.05444324e+00 5.74888177e-02 -1.75318465e-01
-9.87702250e-01 6.52601719e-01 4.48356709e-03 -2.56419569e-01
-1.85546994e-01 4.77589220e-01 -1.95768148e-01 -3.35491538e-01
-3.54511216e-02 -1.73400730e-01 -6.31118417e-01 5.09320140e-01
1.04271293e+00 7.01651096e-01 5.49588911e-02 -9.30377960e-01
-6.22763455e-01 -2.14132100e-01 -2.67509282e-01 9.26707014e-02
1.12728977e+00 -5.02028644e-01 -5.70449173e-01 3.72095823e-01
1.32125258e+00 3.15910012e-01 -1.40812665e-01 -2.91402876e-01
2.90273100e-01 -4.58998561e-01 6.07449077e-02 -1.20369470e+00
-2.59562790e-01 3.15222025e-01 -4.02627140e-01 1.01395762e+00
3.36037904e-01 1.28186554e-01 9.50470030e-01 2.30347216e-01
5.91602512e-02 -9.96563613e-01 1.87717415e-02 1.05271471e+00
1.31511188e+00 -1.32778907e+00 5.75370118e-02 -4.02938902e-01
-6.80905759e-01 8.10655415e-01 1.24356538e-01 2.63698936e-01
1.14634907e+00 -1.08511969e-01 4.12573963e-01 -7.47520328e-01
-9.71331179e-01 1.55153275e-01 3.13182324e-01 2.03858137e-01
4.44074124e-01 4.18013960e-01 -7.49636829e-01 4.01178688e-01
-2.65196949e-01 -1.96723998e-01 9.72923577e-01 7.17086434e-01
-8.84286284e-01 -3.01652819e-01 -4.18327183e-01 6.95083737e-01
-9.77977395e-01 -3.03652972e-01 -8.96232486e-01 4.56572682e-01
-1.53564245e-01 1.56347108e+00 -5.40365651e-02 -4.77647960e-01
-5.36156744e-02 -2.45030314e-01 -5.48800468e-01 -6.97914898e-01
-7.17099309e-01 -3.05768758e-01 1.00935793e+00 -5.43344319e-01
-3.75605702e-01 -8.95030320e-01 -9.71579075e-01 -7.74865329e-01
1.50476068e-01 5.19111395e-01 5.69131911e-01 1.09836972e+00
7.97459483e-01 -2.44923607e-01 6.16764128e-01 -1.98609784e-01
-6.18769526e-01 -1.12147951e+00 1.16632122e-03 5.82567990e-01
7.86818206e-01 -2.67974436e-01 -7.61446416e-01 -3.03197175e-01] | [8.413398742675781, 9.81877326965332] |
7cb24021-6d24-44e8-8dcc-c4fc1149f3ea | painting-3d-nature-in-2d-view-synthesis-of | 2302.07224 | null | https://arxiv.org/abs/2302.07224v1 | https://arxiv.org/pdf/2302.07224v1.pdf | Painting 3D Nature in 2D: View Synthesis of Natural Scenes from a Single Semantic Mask | We introduce a novel approach that takes a single semantic mask as input to synthesize multi-view consistent color images of natural scenes, trained with a collection of single images from the Internet. Prior works on 3D-aware image synthesis either require multi-view supervision or learning category-level prior for specific classes of objects, which can hardly work for natural scenes. Our key idea to solve this challenging problem is to use a semantic field as the intermediate representation, which is easier to reconstruct from an input semantic mask and then translate to a radiance field with the assistance of off-the-shelf semantic image synthesis models. Experiments show that our method outperforms baseline methods and produces photorealistic, multi-view consistent videos of a variety of natural scenes. | ['Xiaowei Zhou', 'Yiyi Liao', 'Kaicheng Yu', 'Haotong Lin', 'Linzhan Mou', 'Tianrun Chen', 'Sida Peng', 'Shangzhan Zhang'] | 2023-02-14 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Painting_3D_Nature_in_2D_View_Synthesis_of_Natural_Scenes_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_Painting_3D_Nature_in_2D_View_Synthesis_of_Natural_Scenes_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-aware-image-synthesis'] | ['computer-vision'] | [ 5.18774211e-01 -7.28845224e-02 6.87957928e-02 -4.11800832e-01
-5.63898563e-01 -8.84082019e-01 7.44040549e-01 -5.54458737e-01
5.06047830e-02 5.18684506e-01 -4.87789735e-02 5.75126968e-02
3.52031738e-01 -7.97130525e-01 -1.09166706e+00 -4.82542992e-01
6.12717152e-01 3.71595830e-01 5.49030125e-01 -4.19272065e-01
1.94020137e-01 5.33385217e-01 -1.67261362e+00 6.35924757e-01
5.91961265e-01 1.02226079e+00 6.26410902e-01 7.50697672e-01
-4.54155654e-02 8.04012001e-01 -2.47034639e-01 -4.48379934e-01
8.17954600e-01 -6.29980087e-01 -8.95376086e-01 7.84087598e-01
9.37152326e-01 -6.85044646e-01 -8.50564986e-02 1.15278709e+00
6.29471242e-02 1.36548281e-01 4.98443753e-01 -1.19529903e+00
-6.13048732e-01 -1.34809986e-01 -6.24567032e-01 -2.15617552e-01
5.85971534e-01 3.00606370e-01 8.28547537e-01 -8.42438340e-01
9.11329567e-01 1.44209194e+00 2.55506963e-01 4.61545855e-01
-1.49715388e+00 -4.10337687e-01 4.23080444e-01 5.85503578e-02
-1.18062139e+00 -5.45911670e-01 9.23050523e-01 -2.73848534e-01
7.12602437e-01 4.73774038e-02 7.55321145e-01 1.00541317e+00
1.43347904e-01 5.86823642e-01 1.47236669e+00 -4.47141528e-01
2.01019153e-01 2.04163909e-01 -6.07067406e-01 7.62793720e-01
8.26185346e-02 1.80202618e-01 -6.84553981e-01 1.64545447e-01
1.23206139e+00 1.09258793e-01 -4.08295691e-01 -8.75927925e-01
-1.29502213e+00 5.73241889e-01 5.09277046e-01 -8.60581174e-02
-3.29463631e-01 1.60244375e-01 -1.00206517e-01 2.68205106e-01
5.45689702e-01 3.35777014e-01 -4.83773440e-01 4.06485051e-01
-8.15214634e-01 1.73680693e-01 4.53420728e-01 1.19148719e+00
1.18908226e+00 1.01667970e-01 3.26319367e-01 6.48470163e-01
1.93232551e-01 7.23617017e-01 -2.01836631e-01 -1.59570479e+00
2.95930564e-01 3.76739442e-01 4.27799881e-01 -7.29746580e-01
1.33005261e-01 -5.43127172e-02 -5.44175804e-01 6.20962143e-01
3.77604038e-01 1.86668530e-01 -1.12937951e+00 1.47573245e+00
4.92053628e-01 -3.84091698e-02 8.33323002e-02 1.07346594e+00
5.54660261e-01 7.81760991e-01 -2.64690101e-01 1.01667136e-01
1.11860394e+00 -1.30578518e+00 -3.54163855e-01 -5.57819307e-01
-9.76373106e-02 -9.17885542e-01 1.11944163e+00 6.06554925e-01
-1.47067130e+00 -7.15544999e-01 -8.26006949e-01 -3.88477981e-01
-3.13144386e-01 -1.10635683e-01 6.28259063e-01 4.32070404e-01
-1.23746276e+00 3.60816360e-01 -5.15267074e-01 -2.91312069e-01
2.37728268e-01 -1.69102579e-01 -5.79338968e-01 -6.43379331e-01
-8.34220529e-01 9.50037658e-01 3.22128922e-01 -1.25916436e-01
-1.23307025e+00 -8.58795047e-01 -1.05167437e+00 -1.97996825e-01
5.45060754e-01 -1.10926437e+00 1.03945005e+00 -1.61893523e+00
-1.64264810e+00 1.10512769e+00 -1.80836380e-01 8.05114880e-02
4.88488048e-01 -1.15177423e-01 -5.14012463e-02 7.02383935e-01
2.26007685e-01 1.04779172e+00 1.20549786e+00 -1.94875383e+00
-5.69322586e-01 -1.75309554e-01 5.94697595e-01 4.97111499e-01
1.74000606e-01 -1.31717637e-01 -6.78881228e-01 -6.36198103e-01
1.08746096e-01 -8.32229435e-01 -3.54794502e-01 5.33142507e-01
-3.31928611e-01 3.67052615e-01 8.25115681e-01 -6.62786305e-01
1.70346707e-01 -1.91675210e+00 4.11251575e-01 -1.37222689e-02
-2.18975320e-01 -1.72919840e-01 -3.78209472e-01 3.65078181e-01
-1.44856080e-01 -2.84393609e-01 -3.44057739e-01 -3.31874192e-01
-3.65148574e-01 2.14134321e-01 -4.46422309e-01 4.10824060e-01
1.41863152e-01 6.98978186e-01 -1.17780745e+00 -2.94573575e-01
7.08388448e-01 5.09664595e-01 -8.96384895e-01 4.96163219e-01
-7.12811470e-01 8.30389082e-01 -1.43338829e-01 6.63261712e-01
8.92245710e-01 -4.32048887e-01 7.95019642e-02 -4.62081909e-01
-3.67226638e-02 1.76478088e-01 -1.31382096e+00 2.47853446e+00
-7.69782603e-01 4.21220928e-01 2.42758885e-01 -9.33414698e-01
7.15548933e-01 -1.29231615e-02 4.92284507e-01 -4.88333523e-01
-5.99534288e-02 9.32266116e-02 -6.43305719e-01 -2.76274770e-01
3.58417928e-01 -3.96421343e-01 1.01227000e-01 5.09685814e-01
1.97125316e-01 -9.41588521e-01 2.05326267e-02 3.13581020e-01
6.13687634e-01 6.37113512e-01 3.74452285e-02 -8.87642875e-02
5.47283709e-01 1.13830082e-01 4.63552952e-01 5.17632127e-01
3.65245491e-01 1.26387501e+00 1.51438579e-01 -4.97903794e-01
-1.10301054e+00 -1.47101879e+00 2.21442565e-01 8.24590266e-01
5.28111160e-01 -1.65205002e-01 -6.48032725e-01 -6.56290174e-01
-2.54077762e-01 7.27673233e-01 -3.90001982e-01 1.24829993e-01
-4.48385775e-01 -1.82697535e-01 -2.40274608e-01 3.77840132e-01
7.86750436e-01 -6.91102684e-01 -6.88214302e-01 -1.70625031e-01
-4.69129473e-01 -1.60992122e+00 -5.75417757e-01 2.02578418e-02
-7.84134507e-01 -1.23033607e+00 -6.22572601e-01 -7.08585501e-01
9.41613138e-01 1.05212939e+00 1.36525917e+00 1.04466140e-01
-2.56555885e-01 6.79985821e-01 -2.24745974e-01 -1.44543365e-01
-4.29556578e-01 -5.47178566e-01 -2.51608729e-01 3.34679604e-01
-3.20559770e-01 -5.82326949e-01 -8.09208930e-01 2.80896008e-01
-1.21147740e+00 6.19546354e-01 4.27913308e-01 6.89484954e-01
6.84506357e-01 -2.02839687e-01 1.02324121e-01 -7.38609493e-01
-3.32250804e-01 -1.82197765e-01 -8.19755018e-01 4.50870454e-01
-2.06220776e-01 -4.49984111e-02 9.29243147e-01 -1.29831597e-01
-1.41029954e+00 5.46116233e-01 2.60774642e-01 -8.00242126e-01
-3.00247639e-01 -2.08702788e-01 -4.08460587e-01 -2.24352419e-01
4.30801302e-01 4.08737123e-01 1.58415008e-02 -3.59374225e-01
8.29945147e-01 3.02862942e-01 6.41151249e-01 -6.34565175e-01
9.41856384e-01 1.10195053e+00 3.55558917e-02 -7.74901450e-01
-1.05878639e+00 -3.57983798e-01 -9.10188377e-01 -2.38862887e-01
1.00294960e+00 -1.45884931e+00 -1.71750337e-01 5.81617951e-01
-1.33921826e+00 -5.42233586e-01 -2.96136439e-01 2.41707206e-01
-9.25585747e-01 4.16230619e-01 -3.43481481e-01 -2.14203998e-01
4.58347127e-02 -1.13689673e+00 1.61495173e+00 1.74278304e-01
1.43934503e-01 -9.79763806e-01 -2.30665430e-01 7.34029412e-01
2.70115077e-01 1.68968901e-01 8.01430941e-01 3.64288598e-01
-1.29414189e+00 3.19590747e-01 -3.69359940e-01 5.98547280e-01
4.17661577e-01 -1.23987496e-02 -1.06845939e+00 -2.02838331e-01
7.27280276e-03 -5.06293058e-01 7.31796384e-01 3.92137796e-01
1.16861761e+00 2.83468477e-02 -1.78147256e-01 9.23506677e-01
1.71423256e+00 -6.21257303e-03 6.32227600e-01 9.35526118e-02
8.54843974e-01 6.38251722e-01 5.25269866e-01 1.55526981e-01
5.71307659e-01 7.99992740e-01 5.36653936e-01 -3.73228282e-01
-5.47929585e-01 -4.65968728e-01 3.91153991e-01 2.96927065e-01
-4.97454964e-02 -1.85799867e-01 -4.38542366e-01 5.06594300e-01
-1.42479122e+00 -9.51845288e-01 1.67029113e-01 2.00284624e+00
7.43509531e-01 -2.69503981e-01 -5.36072142e-02 -1.43984675e-01
5.35622060e-01 2.83618510e-01 -4.92417812e-01 -2.03938186e-01
-1.00799702e-01 1.02023758e-01 4.78476048e-01 6.29182935e-01
-8.36169422e-01 1.17463326e+00 6.94916487e+00 4.53255773e-01
-1.30906534e+00 1.11179404e-01 7.31688142e-01 -4.62786406e-02
-7.14808106e-01 2.85015196e-01 -4.29773927e-01 2.16040060e-01
4.26365852e-01 1.73795298e-01 7.47766435e-01 6.79602921e-01
2.30215609e-01 -5.05488575e-01 -1.23613501e+00 1.18314612e+00
3.93236488e-01 -1.48109472e+00 3.79918933e-01 -2.10162967e-01
1.27823532e+00 -8.90416056e-02 -2.24487595e-02 -4.73626971e-01
6.62046075e-01 -7.30935395e-01 1.03976429e+00 4.91293997e-01
8.97226155e-01 -3.50948483e-01 -1.08476892e-01 1.63010448e-01
-1.06786323e+00 1.31431714e-01 -4.08919275e-01 1.19636178e-01
4.69010115e-01 4.88991588e-01 -3.52881104e-01 8.06188583e-01
9.24144983e-01 1.07855570e+00 -4.58533376e-01 6.86968505e-01
-4.07366157e-01 -5.27891144e-02 -2.61015773e-01 6.17552876e-01
8.62316787e-02 -5.13617814e-01 2.98903614e-01 6.57744706e-01
2.37185970e-01 1.93165645e-01 1.61113739e-01 1.11728275e+00
3.77469733e-02 -3.11342001e-01 -8.55109036e-01 4.32436466e-01
2.99110450e-02 1.26612437e+00 -8.61017764e-01 -4.10557985e-01
-7.10413218e-01 1.55996668e+00 8.87089223e-02 6.07546508e-01
-6.82146788e-01 3.34196016e-02 6.48131788e-01 2.09600881e-01
5.89610398e-01 -3.31442475e-01 -2.27697760e-01 -1.53818500e+00
-2.11686715e-02 -8.99275720e-01 1.24931574e-01 -1.63267744e+00
-1.15543759e+00 4.96900946e-01 2.52880752e-01 -1.49295366e+00
-3.37343842e-01 -7.13677943e-01 -4.70631808e-01 7.91303515e-01
-1.88897872e+00 -1.32517910e+00 -6.17904663e-01 9.08568323e-01
8.38981748e-01 1.51006982e-01 6.53299630e-01 5.10144867e-02
9.18406323e-02 -1.03209585e-01 -7.35296085e-02 -1.89008102e-01
7.45676756e-01 -1.10763180e+00 3.73241872e-01 1.01662290e+00
2.19309002e-01 1.72358572e-01 5.48907578e-01 -3.15125406e-01
-1.64402807e+00 -1.08945918e+00 3.84506494e-01 -6.15633011e-01
1.76658109e-01 -3.76743585e-01 -5.24888217e-01 6.22340500e-01
5.16487658e-01 4.25970346e-01 2.22077936e-01 -4.75517362e-01
-5.59063196e-01 -3.27284992e-01 -1.09283149e+00 5.51074326e-01
1.31769931e+00 -5.67715287e-01 -3.93608183e-01 3.42476100e-01
5.80118179e-01 -5.04434526e-01 -6.33392513e-01 1.61612645e-01
4.08634782e-01 -1.37531614e+00 1.44904280e+00 -3.46192688e-01
7.85233259e-01 -6.49830043e-01 -3.69721532e-01 -1.62309086e+00
-7.23368954e-03 -4.76042390e-01 4.37027574e-01 7.99956083e-01
3.69215496e-02 -3.29624593e-01 5.51894426e-01 4.68589187e-01
-2.01328576e-01 -1.90116704e-01 -4.86264914e-01 -7.52321362e-01
-5.58407865e-02 -1.81630790e-01 4.72098380e-01 8.22199523e-01
-6.13956511e-01 2.44714558e-01 -5.14413059e-01 2.08851263e-01
9.10687029e-01 7.09837317e-01 1.14645648e+00 -9.32479858e-01
-3.31158638e-01 -1.31064672e-02 -2.32498527e-01 -1.19237614e+00
1.84154212e-01 -7.65008926e-01 1.76533148e-01 -1.69473159e+00
8.90077874e-02 -3.57449621e-01 1.80153772e-01 2.86553085e-01
5.01688756e-02 5.09797454e-01 3.13569069e-01 -1.55308247e-02
-6.19793832e-01 7.07869589e-01 1.64838600e+00 -1.35904208e-01
6.35886416e-02 -3.51945788e-01 -7.28851497e-01 7.69108951e-01
4.44744408e-01 -2.60459423e-01 -8.02146375e-01 -7.75326312e-01
1.08497530e-01 1.07831590e-01 7.53699720e-01 -7.53262818e-01
-1.20285228e-01 -6.78155303e-01 6.40105009e-01 -4.16259408e-01
6.87487185e-01 -9.82844055e-01 4.38641459e-01 2.18136624e-01
-6.46755472e-02 -4.55476679e-02 6.91091234e-04 6.29487753e-01
-1.71262071e-01 2.80171335e-02 1.01113188e+00 -6.86569393e-01
-8.56910825e-01 2.60817647e-01 -6.89302087e-02 9.57994238e-02
9.98681247e-01 -2.93768227e-01 -1.71422452e-01 -4.52358902e-01
-3.74965101e-01 -3.74044627e-02 1.10957313e+00 6.96896553e-01
7.33991027e-01 -1.30869305e+00 -5.91824174e-01 4.48183656e-01
1.23583704e-01 3.49546164e-01 3.20000827e-01 2.62505710e-01
-8.31658840e-01 2.40317732e-01 -5.48428059e-01 -8.60234857e-01
-9.82528806e-01 5.76072514e-01 5.28441131e-01 2.13315338e-01
-7.54566848e-01 7.03998685e-01 7.39804387e-01 -4.02044952e-01
-1.13659859e-01 -3.71868342e-01 4.65894014e-01 -4.15710181e-01
4.90085244e-01 -9.45331082e-02 4.32730541e-02 -7.36879587e-01
-2.44125664e-01 1.02600646e+00 2.67382860e-01 -3.33750904e-01
1.45841956e+00 -4.71440881e-01 -1.44890696e-01 3.40580255e-01
1.19546831e+00 -7.95452520e-02 -1.89094460e+00 -2.96113938e-01
-6.35492384e-01 -1.11558902e+00 1.41189352e-01 -6.45113170e-01
-1.45459437e+00 7.03559637e-01 2.29888439e-01 -2.19288275e-01
1.31045616e+00 9.01888311e-02 6.71961665e-01 2.10380808e-01
7.29093492e-01 -9.71895337e-01 4.96168405e-01 1.69561177e-01
9.91097391e-01 -1.45842600e+00 4.33062091e-02 -6.90690517e-01
-7.03098178e-01 1.17452109e+00 6.87238455e-01 -1.77843511e-01
6.57964170e-01 5.28912805e-02 1.89162344e-01 -1.00725986e-01
-7.45819390e-01 -2.48939648e-01 3.06960791e-01 7.57420480e-01
1.55288622e-01 -3.36565435e-01 3.29269141e-01 -9.87900421e-02
1.25982031e-01 -1.14899814e-01 6.86287522e-01 8.06678534e-01
-2.40714669e-01 -1.16659856e+00 -3.55135500e-01 -6.32421374e-02
-1.01277314e-01 -1.05642140e-01 -2.76253670e-01 4.73056436e-01
2.10355803e-01 9.60634232e-01 3.79499123e-02 -1.72876820e-01
2.82011330e-01 -6.93687871e-02 1.06512177e+00 -7.77872026e-01
-2.27155149e-01 1.72270402e-01 -1.29941314e-01 -9.71681714e-01
-1.00105345e+00 -7.24185169e-01 -9.03758943e-01 -1.05424412e-01
6.34285510e-02 -3.43834311e-01 6.54276609e-01 8.57838035e-01
2.94574380e-01 1.33344620e-01 9.35710132e-01 -1.29408157e+00
5.85296378e-02 -3.42303902e-01 -4.99769479e-01 6.49331510e-01
4.82348025e-01 -7.12407470e-01 -3.25160474e-01 6.27698839e-01] | [9.260663032531738, -3.0997822284698486] |
b19d4898-bb46-4932-a332-f83377aaf7b8 | improving-graph-neural-networks-on-multi-node | 2304.10074 | null | https://arxiv.org/abs/2304.10074v1 | https://arxiv.org/pdf/2304.10074v1.pdf | Improving Graph Neural Networks on Multi-node Tasks with Labeling Tricks | In this paper, we provide a theory of using graph neural networks (GNNs) for \textit{multi-node representation learning}, where we are interested in learning a representation for a set of more than one node such as a link. Existing GNNs are mainly designed to learn single-node representations. When we want to learn a node-set representation involving multiple nodes, a common practice in previous works is to directly aggregate the single-node representations obtained by a GNN. In this paper, we show a fundamental limitation of such an approach, namely the inability to capture the dependence among multiple nodes in a node set, and argue that directly aggregating individual node representations fails to produce an effective joint representation for multiple nodes. A straightforward solution is to distinguish target nodes from others. Formalizing this idea, we propose \text{labeling trick}, which first labels nodes in the graph according to their relationships with the target node set before applying a GNN and then aggregates node representations obtained in the labeled graph for multi-node representations. The labeling trick also unifies a few previous successful works for multi-node representation learning, including SEAL, Distance Encoding, ID-GNN, and NBFNet. Besides node sets in graphs, we also extend labeling tricks to posets, subsets and hypergraphs. Experiments verify that the labeling trick technique can boost GNNs on various tasks, including undirected link prediction, directed link prediction, hyperedge prediction, and subgraph prediction. Our work explains the superior performance of previous node-labeling-based methods and establishes a theoretical foundation for using GNNs for multi-node representation learning. | ['Muhan Zhang', 'Pan Li', 'Xiyuan Wang'] | 2023-04-20 | null | null | null | null | ['hyperedge-prediction'] | ['graphs'] | [ 4.31011409e-01 6.48590803e-01 -6.87775314e-01 -2.84131378e-01
-3.49765658e-01 -5.64602017e-01 4.32953358e-01 4.94558603e-01
1.33134276e-01 6.43859684e-01 -5.41075952e-02 -5.94049335e-01
-5.89729846e-01 -1.44254315e+00 -7.29875624e-01 -6.55005157e-01
-4.45676208e-01 7.47106552e-01 1.06796868e-01 -4.26331580e-01
-2.13653281e-01 7.81102896e-01 -1.19353569e+00 1.71876270e-02
4.76412505e-01 6.40548348e-01 -7.42385536e-03 5.77749133e-01
-2.41367936e-01 8.90817404e-01 -6.26585364e-01 -4.08934355e-01
2.14938089e-01 -3.62654030e-01 -1.13488030e+00 -2.53859699e-01
5.39419234e-01 -6.90477807e-03 -6.32134736e-01 9.55145538e-01
2.26413026e-01 8.04194286e-02 9.95521963e-01 -1.73790896e+00
-6.91994429e-01 1.23796272e+00 -7.41031051e-01 2.55656894e-02
3.32934171e-01 -4.53432024e-01 1.69523942e+00 -3.28892171e-01
6.09423399e-01 1.41743028e+00 9.76572871e-01 5.83903849e-01
-1.29431891e+00 -4.85248506e-01 2.82821566e-01 5.19216135e-02
-1.48939681e+00 1.56982511e-01 7.82248080e-01 -2.11586580e-01
8.50511432e-01 3.38129699e-01 4.11391705e-01 9.70483541e-01
-2.42587253e-02 8.56070578e-01 5.93505740e-01 -5.84852576e-01
-3.55495661e-01 -2.04776615e-01 6.19596004e-01 1.02429879e+00
5.73228538e-01 3.94102819e-02 3.39156725e-02 -1.52179152e-01
8.30571592e-01 1.81944281e-01 -2.53736168e-01 -5.68777740e-01
-8.19095969e-01 1.14285111e+00 1.12210977e+00 5.45803189e-01
-1.12532459e-01 6.40313447e-01 5.42776942e-01 5.44645488e-01
4.52963054e-01 4.55410242e-01 -3.11624169e-01 6.31035149e-01
-6.67024016e-01 -1.55298591e-01 8.86133611e-01 9.19886470e-01
1.17432451e+00 1.27948001e-01 -2.40641907e-01 6.92330122e-01
2.64319807e-01 1.66500494e-01 2.18894333e-01 -5.67854047e-01
3.95227581e-01 8.54844570e-01 -7.31739998e-01 -1.32245898e+00
-8.58922720e-01 -5.47019720e-01 -1.19023359e+00 -7.96830375e-03
2.08432510e-01 -7.36695826e-02 -1.11360013e+00 2.19265389e+00
-3.22978608e-02 3.24274957e-01 1.21986113e-01 3.60039473e-01
1.28719378e+00 5.04321516e-01 1.15774743e-01 -9.88624021e-02
1.01523042e+00 -1.13250399e+00 -3.11541587e-01 -1.28051087e-01
1.37635338e+00 4.28400449e-02 3.70485008e-01 -1.14900582e-01
-7.98803627e-01 -4.97133940e-01 -9.19883072e-01 1.34295747e-01
-6.98279500e-01 -2.74284303e-01 1.10806084e+00 6.32034838e-01
-1.40183723e+00 9.11472142e-01 -3.92393410e-01 -5.63821137e-01
3.57970417e-01 7.03198731e-01 -5.88831484e-01 -2.06532508e-01
-1.38922179e+00 9.83571172e-01 8.48434329e-01 -1.44565389e-01
-8.44579101e-01 -4.92640525e-01 -1.29503286e+00 4.58109856e-01
4.83991355e-01 -8.01335752e-01 8.50774050e-01 -8.94587815e-01
-6.29954636e-01 7.67815053e-01 7.04655275e-02 -5.87354779e-01
-1.72248289e-01 3.85260046e-01 -5.45962453e-01 2.29707658e-01
2.71201320e-02 7.29716599e-01 5.22907436e-01 -1.39928401e+00
-3.23532611e-01 -1.87109530e-01 4.82411444e-01 1.29232302e-01
-6.30954802e-01 -3.53462517e-01 -2.91760355e-01 -5.91666818e-01
2.33565032e-01 -8.85111809e-01 -4.16941226e-01 -2.25862771e-01
-7.74729609e-01 -7.17619479e-01 8.61353338e-01 -8.73523206e-02
1.41880929e+00 -1.81902432e+00 3.92095327e-01 8.34811807e-01
8.28509033e-01 3.61966848e-01 -6.88732028e-01 6.57202899e-01
-5.60135186e-01 5.26734352e-01 -1.35556748e-02 -2.35356063e-01
-1.72778890e-02 6.33821547e-01 -1.22400327e-02 3.43531311e-01
-1.46422125e-02 1.19187582e+00 -9.75094318e-01 -5.20342708e-01
4.13679220e-02 3.82004797e-01 -2.51311064e-01 -6.93198759e-04
-8.03047940e-02 -9.38431621e-02 -3.62786233e-01 6.02296114e-01
5.84758937e-01 -7.54317343e-01 5.90934932e-01 -2.95469433e-01
7.74230123e-01 2.18525246e-01 -1.19198000e+00 1.37238586e+00
-4.41010922e-01 4.43273097e-01 -1.16124235e-01 -1.42403114e+00
1.08951426e+00 1.81591585e-01 5.08438826e-01 -3.73453587e-01
-9.38171614e-03 1.43543303e-01 9.82154459e-02 -1.36265486e-01
2.86714971e-01 -5.78008257e-02 -1.21293038e-01 6.45873249e-01
4.30358320e-01 2.17375919e-01 4.25692379e-01 7.02728570e-01
1.46530020e+00 -2.11220279e-01 5.69839597e-01 -7.75178000e-02
4.38056469e-01 -5.27608454e-01 3.22435498e-01 9.17289853e-01
2.57320821e-01 5.20697057e-01 8.55880439e-01 -4.93176669e-01
-6.36468887e-01 -9.48290229e-01 1.97615772e-01 1.43382370e+00
8.54920149e-02 -9.10397232e-01 -2.15301886e-01 -1.13613033e+00
1.66942492e-01 4.14208084e-01 -9.16991889e-01 -4.98294532e-01
-7.24503338e-01 -7.07400262e-01 8.08858395e-01 8.64747226e-01
1.30592072e-02 -9.85887349e-01 -1.84220877e-02 8.95419791e-02
3.24128084e-02 -9.02203321e-01 -1.63744688e-01 6.54926419e-01
-9.02813554e-01 -1.26476216e+00 -4.60692167e-01 -1.17118526e+00
8.82694066e-01 3.92135352e-01 1.45141470e+00 8.53087246e-01
-8.01266953e-02 4.14586693e-01 -3.84562075e-01 5.10619618e-02
-4.55795437e-01 4.96526182e-01 1.20167527e-02 -2.99105376e-01
3.37800011e-02 -8.57519269e-01 -3.15633342e-02 2.42381915e-01
-8.69719803e-01 -1.22457765e-01 7.40490019e-01 8.96767914e-01
4.04888362e-01 8.82872343e-02 7.10030496e-01 -1.50660241e+00
5.50102532e-01 -8.68084133e-01 -7.42428526e-02 7.52955496e-01
-6.43384397e-01 2.92254567e-01 7.43428946e-01 -3.87874514e-01
-3.50451291e-01 -4.91351774e-03 -1.35403901e-01 -4.45557892e-01
1.02840647e-01 9.78368700e-01 -8.51799995e-02 -4.15233582e-01
6.23829126e-01 -1.41573884e-02 -1.05798310e-02 -3.55423540e-01
5.28204560e-01 8.98733735e-02 3.28780651e-01 -6.71966434e-01
9.86528099e-01 2.51886547e-02 7.97738791e-01 -5.16382813e-01
-7.89150059e-01 -3.96270663e-01 -7.75053144e-01 5.55288307e-02
3.77242774e-01 -6.87095344e-01 -8.38438451e-01 -4.68009971e-02
-1.11137784e+00 -1.35050163e-01 -2.24449500e-01 1.95527092e-01
-3.02113086e-01 5.76976538e-01 -6.79921150e-01 -5.00056565e-01
-3.37181032e-01 -8.06728005e-01 8.40528786e-01 2.52202600e-02
-1.73770711e-01 -1.72619224e+00 -2.37624012e-02 -5.87055311e-02
2.99670428e-01 3.96874696e-01 1.51557159e+00 -1.19441187e+00
-2.61564404e-01 -2.19438225e-01 -3.58944267e-01 9.40764472e-02
1.10565484e-01 -7.80120343e-02 -6.42090440e-01 -6.17321253e-01
-7.60623157e-01 -6.18128777e-01 1.20731604e+00 3.09310019e-01
1.26696670e+00 -3.12367260e-01 -1.02474725e+00 7.84094155e-01
1.53832698e+00 -1.97718352e-01 4.84034061e-01 -7.44056981e-03
1.11617565e+00 4.77736831e-01 4.89043854e-02 5.47703095e-02
4.21327204e-01 5.57134092e-01 8.24240983e-01 -2.73848146e-01
-3.53573561e-01 -3.28741640e-01 1.38951465e-01 6.78915560e-01
-1.84018478e-01 -8.96687448e-01 -9.03858840e-01 5.16502321e-01
-1.77969849e+00 -6.68177485e-01 -2.68705696e-01 1.71891737e+00
3.68349224e-01 4.13900428e-03 2.42645681e-01 1.48989573e-01
1.00002611e+00 4.28023875e-01 -3.04225475e-01 -2.87007958e-01
-2.13635176e-01 4.18416113e-01 5.14281094e-01 3.97930026e-01
-1.09893608e+00 9.36373532e-01 6.57134914e+00 9.03033495e-01
-7.45361865e-01 -6.90426975e-02 5.26689529e-01 5.17830193e-01
-7.37378359e-01 5.61408028e-02 -8.57509434e-01 -9.17209312e-02
9.93923008e-01 -1.86540544e-01 2.25170195e-01 7.97954261e-01
-7.67796636e-01 4.15063828e-01 -1.46033764e+00 6.83886588e-01
2.32769251e-01 -1.45376623e+00 5.10124326e-01 1.43069670e-01
6.27723575e-01 -9.74035859e-02 -8.29519629e-02 6.55178189e-01
7.36964464e-01 -1.56494498e+00 -5.06349690e-02 1.81931376e-01
7.92806983e-01 -6.30912185e-01 7.52281129e-01 2.79604137e-01
-1.70226693e+00 -1.24918438e-01 -3.81168157e-01 1.29530784e-02
-6.20561801e-02 3.05456609e-01 -1.02992642e+00 1.19102001e+00
2.75587082e-01 9.06406224e-01 -7.50381887e-01 8.93605232e-01
-3.51943374e-01 5.69504619e-01 -4.18973058e-01 -5.20897508e-02
4.14278269e-01 2.85327584e-02 3.72334540e-01 1.14518273e+00
3.26681167e-01 -1.34983718e-01 5.17128825e-01 6.63057804e-01
-4.48614627e-01 -7.83779696e-02 -1.03618813e+00 -5.39650209e-02
5.88207841e-01 1.49119949e+00 -8.40911567e-01 -1.57213882e-01
-4.99168038e-01 6.11532271e-01 8.77082586e-01 3.04978907e-01
-6.64229393e-01 -4.26732540e-01 2.55065560e-01 1.87327061e-02
3.49681497e-01 5.12176454e-02 -1.98688600e-02 -7.36263454e-01
-2.75517344e-01 -5.60708165e-01 1.00701201e+00 -5.63694537e-01
-1.41465104e+00 7.41727769e-01 2.33075157e-01 -1.05667973e+00
-3.06476086e-01 -7.60242105e-01 -8.18920016e-01 6.85998499e-01
-1.40331054e+00 -1.56655574e+00 -2.03953072e-01 6.47255599e-01
-5.50375357e-02 -1.99574411e-01 1.16693330e+00 1.42207503e-01
-3.97412509e-01 9.10726428e-01 -2.04767659e-01 4.52048123e-01
1.69407114e-01 -1.42977822e+00 3.53667736e-01 5.60129642e-01
5.75137794e-01 7.19592333e-01 3.01747859e-01 -5.72433174e-01
-1.22240067e+00 -1.12990940e+00 7.75814593e-01 -1.64360046e-01
6.51619971e-01 -1.78880006e-01 -9.63253140e-01 1.24110830e+00
-1.18083516e-02 5.05036116e-01 7.99896538e-01 6.30515933e-01
-7.18900681e-01 -8.59880894e-02 -1.02006698e+00 3.89258534e-01
1.32397771e+00 -4.74337786e-01 -4.04311806e-01 5.46374798e-01
8.56165171e-01 -6.32909834e-02 -1.09083200e+00 8.14132810e-01
2.68487573e-01 -4.92356777e-01 1.30336797e+00 -1.04071403e+00
2.51443893e-01 -7.97165409e-02 -1.79446250e-01 -1.68360937e+00
-7.25721955e-01 -4.14640933e-01 -5.35646558e-01 1.17279029e+00
6.05691969e-01 -5.99832475e-01 9.71375644e-01 -5.91214821e-02
-1.65259242e-01 -9.10898805e-01 -7.74476051e-01 -8.95442069e-01
1.44041464e-01 -4.88772653e-02 6.67831481e-01 1.05931032e+00
-5.92796765e-02 7.72326410e-01 -4.55117464e-01 2.86276132e-01
6.13617837e-01 2.16327935e-01 5.55255294e-01 -1.83579886e+00
-3.86755884e-01 -5.16459167e-01 -7.81192660e-01 -1.05917180e+00
6.53244734e-01 -1.65820503e+00 -3.67570639e-01 -2.01855254e+00
3.04832727e-01 -8.46330464e-01 -5.87476254e-01 9.72787619e-01
-6.76626638e-02 2.38632619e-01 1.93875000e-01 -4.41906117e-02
-8.08376312e-01 2.37420410e-01 1.20187938e+00 -3.66912216e-01
9.67258811e-02 1.03486739e-01 -9.52559531e-01 5.98927796e-01
6.30856574e-01 -5.61886430e-01 -6.03058577e-01 -3.26180100e-01
4.21842337e-01 9.72992480e-02 3.78857553e-01 -8.94290149e-01
3.05081666e-01 1.51308611e-01 3.05314094e-01 -4.27458137e-01
3.82887274e-01 -9.17599857e-01 2.17780620e-01 6.17173851e-01
-4.76327628e-01 -1.74977798e-02 -1.44851565e-01 6.91263080e-01
-3.27963717e-02 -5.69795072e-01 4.89892095e-01 -2.47927561e-01
-8.62923861e-01 4.90043581e-01 1.23305105e-01 1.70791745e-02
1.08401000e+00 -2.59645581e-01 -6.49977028e-01 -5.16020179e-01
-9.92829382e-01 5.03612041e-01 -2.29630433e-02 2.68285543e-01
6.35058939e-01 -1.54651117e+00 -6.79650307e-01 9.24892947e-02
2.16645375e-01 -3.09537761e-02 3.04408148e-02 6.17827833e-01
-1.90891191e-01 2.14887187e-01 -1.64826557e-01 -4.23709214e-01
-1.43932045e+00 7.34157681e-01 3.69880825e-01 -6.94824278e-01
-5.50868571e-01 1.04389358e+00 1.21895373e-01 -6.68668628e-01
3.76083046e-01 -6.66624606e-02 -5.08474827e-01 1.28001824e-01
1.02851450e-01 3.13813776e-01 -1.00916706e-01 -7.72349238e-01
-3.32558453e-01 6.47549808e-01 -4.41469193e-01 6.72504663e-01
1.44130898e+00 2.20135421e-01 -3.70225519e-01 3.17224145e-01
1.52494335e+00 -4.20191526e-01 -5.00110865e-01 -5.31607628e-01
6.80906773e-02 -8.77326205e-02 -1.91827625e-01 -5.77288687e-01
-1.41729641e+00 7.00365126e-01 1.09787636e-01 6.19270265e-01
9.66235340e-01 5.52430212e-01 6.03596330e-01 8.06210220e-01
3.63409936e-01 -4.46253538e-01 2.30619833e-01 5.20178676e-01
5.96249282e-01 -1.13227606e+00 1.75355345e-01 -8.70982528e-01
-2.63437361e-01 1.46231377e+00 8.68215382e-01 -1.13999642e-01
7.62013018e-01 1.43089801e-01 -4.66383457e-01 -5.13921559e-01
-5.57896197e-01 -4.78606164e-01 4.09057766e-01 8.11253250e-01
3.03900987e-01 2.80809134e-01 1.41010396e-02 3.63573074e-01
6.04590103e-02 -5.34503400e-01 4.62527841e-01 7.17087924e-01
-4.16401684e-01 -1.39953482e+00 -6.57700747e-03 7.34393299e-01
-2.85845459e-01 -2.31718019e-01 -5.60616910e-01 1.04891849e+00
1.18814651e-02 6.52195871e-01 5.52233793e-02 -7.82000244e-01
5.22969756e-03 -9.50037912e-02 5.27671337e-01 -1.00921488e+00
-6.85301602e-01 -5.05216181e-01 3.06010306e-01 -1.78647727e-01
-5.97833216e-01 -1.35771289e-01 -1.34745789e+00 -3.71152520e-01
-4.97236341e-01 2.54943341e-01 1.20316140e-01 7.78065562e-01
9.36746672e-02 7.18514144e-01 5.47706366e-01 -6.27052546e-01
-4.17610139e-01 -9.62467253e-01 -7.79215813e-01 5.16955316e-01
3.77176166e-01 -7.83195257e-01 -4.90706891e-01 -7.11911857e-01] | [7.079738616943359, 6.302099704742432] |
13947280-be52-4d95-9bd9-75d677bfc1da | auto-fedrl-federated-hyperparameter | 2203.06338 | null | https://arxiv.org/abs/2203.06338v2 | https://arxiv.org/pdf/2203.06338v2.pdf | Auto-FedRL: Federated Hyperparameter Optimization for Multi-institutional Medical Image Segmentation | Federated learning (FL) is a distributed machine learning technique that enables collaborative model training while avoiding explicit data sharing. The inherent privacy-preserving property of FL algorithms makes them especially attractive to the medical field. However, in case of heterogeneous client data distributions, standard FL methods are unstable and require intensive hyperparameter tuning to achieve optimal performance. Conventional hyperparameter optimization algorithms are impractical in real-world FL applications as they involve numerous training trials, which are often not affordable with limited compute budgets. In this work, we propose an efficient reinforcement learning (RL)-based federated hyperparameter optimization algorithm, termed Auto-FedRL, in which an online RL agent can dynamically adjust hyperparameters of each client based on the current training progress. Extensive experiments are conducted to investigate different search strategies and RL agents. The effectiveness of the proposed method is validated on a heterogeneous data split of the CIFAR-10 dataset as well as two real-world medical image segmentation datasets for COVID-19 lesion segmentation in chest CT and pancreas segmentation in abdominal CT. | ['Holger R. Roth', 'Vishal M. Patel', 'Gianpaolo Carrafiello', 'Elvira Stellato', 'Francesca Patella', 'Bradford Wood', 'Baris Turkbey', 'Evrim Turkbey', 'Stephanie Harmon', 'Daguang Xu', 'Can Zhao', 'Wenqi Li', 'Ziyue Xu', 'An Xu', 'Ali Hatamizadeh', 'Dong Yang', 'Pengfei Guo'] | 2022-03-12 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [-9.15608481e-02 -3.91897298e-02 -5.48272371e-01 -4.11589324e-01
-9.35119867e-01 -5.24948537e-01 1.07394278e-01 4.22182262e-01
-7.80135393e-01 9.88564730e-01 -3.86949778e-02 -3.54229003e-01
-3.97465438e-01 -5.92016160e-01 -4.40452188e-01 -1.26019692e+00
-2.33598322e-01 7.74031818e-01 -1.52165428e-01 3.30082864e-01
-2.78233290e-01 3.98957312e-01 -9.91348207e-01 1.82816505e-01
9.37630236e-01 1.10052347e+00 -1.49601579e-01 6.20608151e-01
-9.85220224e-02 9.47101653e-01 -4.68798637e-01 -4.33228999e-01
5.82539141e-01 -3.21263075e-01 -7.10810363e-01 -5.77004552e-02
-8.33019912e-02 -3.61931920e-01 -1.30397245e-01 1.13560688e+00
8.22056472e-01 1.75964981e-01 1.69036314e-01 -1.35606539e+00
-1.14798032e-01 9.64210331e-01 -3.66862476e-01 1.12433478e-01
-8.86253640e-02 2.77749717e-01 8.23585331e-01 -1.06810100e-01
6.57544732e-01 6.53850377e-01 6.59958124e-01 5.84199846e-01
-9.29641783e-01 -6.09875381e-01 -3.83399776e-03 2.04853639e-01
-1.14687669e+00 -1.11288399e-01 6.48496509e-01 -4.41136211e-03
6.44351721e-01 4.76499140e-01 6.44413114e-01 8.60665202e-01
2.95839190e-01 1.07768285e+00 1.15858543e+00 -2.16063693e-01
8.17272723e-01 3.24248195e-01 -2.09250033e-01 6.24392927e-01
3.50920022e-01 7.62553811e-02 -5.48574209e-01 -1.00201011e+00
5.06123602e-01 2.80256569e-01 -3.26307386e-01 -9.24120903e-01
-1.14738619e+00 1.11527193e+00 5.25783896e-01 2.55791973e-02
-7.60769129e-01 4.46180403e-02 8.17698181e-01 2.37400353e-01
3.15488458e-01 1.56723052e-01 -4.84014183e-01 1.27806857e-01
-5.99780858e-01 2.65568107e-01 9.96008635e-01 5.53022623e-01
4.57300752e-01 -1.45165816e-01 -1.69045329e-01 7.33001828e-01
3.15854132e-01 3.04936796e-01 1.00765407e+00 -9.19294059e-01
4.19168890e-01 4.03149664e-01 1.66598678e-01 -6.72132850e-01
-4.83057529e-01 -4.37881768e-01 -1.01196921e+00 5.28684035e-02
3.33679020e-01 -5.50089121e-01 -3.17177773e-01 1.50057125e+00
1.12413967e+00 2.80412257e-01 3.80909175e-01 1.12147629e+00
6.06055200e-01 2.01442048e-01 2.91363537e-01 -3.60844821e-01
1.34105158e+00 -1.07102263e+00 -6.66323721e-01 2.59880334e-01
7.96766639e-01 -4.37708229e-01 7.05746889e-01 4.51539338e-01
-9.46703672e-01 3.59533548e-01 -8.88295472e-01 5.83343625e-01
-1.15732327e-01 -1.05399236e-01 1.01566708e+00 9.00688767e-01
-7.29986668e-01 4.24790442e-01 -1.15254521e+00 -1.87830254e-01
8.40629816e-01 6.77977979e-01 -2.29677662e-01 -3.25376205e-02
-1.09515846e+00 3.76148015e-01 3.37954789e-01 -2.11579159e-01
-9.63762224e-01 -9.06148791e-01 -4.43949372e-01 4.39317375e-02
5.81665635e-01 -1.00167274e+00 1.48255622e+00 -1.09450018e+00
-1.81425834e+00 5.32032669e-01 5.12306392e-01 -9.92893875e-01
9.59260225e-01 2.81834751e-01 -3.47112477e-01 2.68544585e-01
-3.56465667e-01 1.99878842e-01 9.77391303e-01 -9.76161659e-01
-7.36895084e-01 -4.95288342e-01 -2.24469855e-01 5.79415917e-01
-5.55035114e-01 -2.21195966e-01 -9.70729291e-02 -5.32361984e-01
-2.25513220e-01 -9.82404768e-01 -7.40168393e-01 8.71896446e-02
-2.37574309e-01 5.65115325e-02 7.16913998e-01 -1.64134428e-01
1.06950617e+00 -2.07679224e+00 -2.35097095e-01 6.33069336e-01
2.28849798e-01 3.32774222e-01 9.31642726e-02 1.69573918e-01
3.62136483e-01 -1.59900278e-01 -2.73296118e-01 -1.35642871e-01
-3.22381556e-02 3.65077376e-01 1.44594222e-01 8.47952127e-01
-7.59239912e-01 7.87499487e-01 -7.50403643e-01 -9.29959178e-01
4.32365462e-02 2.51420021e-01 -7.09947705e-01 2.58865029e-01
-5.67206442e-01 5.41723788e-01 -1.03806555e+00 6.09045148e-01
6.16907120e-01 -6.86097503e-01 5.94393909e-01 -3.42168100e-02
2.79895037e-01 -5.43209612e-01 -1.13953483e+00 1.65594888e+00
-3.55166018e-01 -1.98989779e-01 4.45987523e-01 -7.46771872e-01
4.88584429e-01 5.71158767e-01 1.11040652e+00 -4.09746587e-01
3.99192750e-01 1.48487851e-01 -7.52319917e-02 -6.45540357e-01
1.27530470e-01 6.94615021e-02 2.03217551e-01 9.32027996e-01
-2.34787539e-01 2.20231265e-01 -2.99365968e-01 5.06953113e-02
1.39720392e+00 -3.87626708e-01 5.68212986e-01 -5.61307706e-02
6.31533563e-01 1.52364299e-01 6.94330454e-01 8.62026095e-01
-5.39198399e-01 1.55221000e-01 8.16422030e-02 -8.06412041e-01
-5.90267956e-01 -6.45968080e-01 -8.43524560e-02 1.00391543e+00
1.44477621e-01 -4.73656766e-02 -7.47405410e-01 -1.14713979e+00
3.23281825e-01 4.99670744e-01 -5.31229973e-01 7.32868835e-02
-4.57314521e-01 -1.09367836e+00 5.35559595e-01 -1.04594845e-02
5.31695485e-01 -1.14938450e+00 -1.20304859e+00 5.15407026e-01
-1.77852318e-01 -8.93753350e-01 -4.91147280e-01 1.01905704e-01
-9.52298045e-01 -1.17390656e+00 -7.01396585e-01 -4.25100505e-01
7.27410555e-01 -3.76991145e-02 8.29326570e-01 1.95428327e-01
-7.00763881e-01 5.84506333e-01 -2.04366624e-01 -5.18209159e-01
-5.79676211e-01 2.39599839e-01 -1.69441000e-01 3.99157315e-01
1.53480858e-01 -1.80372819e-01 -1.11918342e+00 4.39262271e-01
-9.81107116e-01 -3.40396494e-01 4.47025985e-01 1.22284424e+00
1.02214813e+00 -1.43825263e-01 7.28045464e-01 -1.48632228e+00
7.62952030e-01 -8.07168841e-01 -8.75398278e-01 6.04391336e-01
-9.66128409e-01 -1.29311353e-01 8.32068503e-01 -4.76390600e-01
-1.27460098e+00 4.25993800e-01 2.11240813e-01 -6.78488135e-01
-2.18887646e-02 5.09958863e-01 2.01536611e-01 -3.91958833e-01
7.71382451e-01 1.33487461e-02 4.78684068e-01 -3.12808484e-01
2.67294228e-01 6.58775210e-01 3.33250850e-01 -5.57148278e-01
1.35029227e-01 6.50973499e-01 -1.12397403e-01 -3.49373847e-01
-3.15490335e-01 -4.44090068e-01 1.39642343e-01 3.46233323e-02
5.49422622e-01 -7.87045836e-01 -1.16074049e+00 2.80057669e-01
-4.43023473e-01 -3.23656797e-01 -5.18965662e-01 8.00117254e-01
-8.46332431e-01 2.03125551e-01 -5.59552252e-01 -4.94387031e-01
-1.01871705e+00 -1.18889058e+00 4.68464464e-01 4.06842947e-01
1.43503159e-01 -1.11374414e+00 3.34242314e-01 5.40537775e-01
7.99227357e-01 3.27327281e-01 9.84959483e-01 -1.04863930e+00
-5.17900944e-01 -2.03762397e-01 2.91806906e-01 -1.80895492e-01
1.29106790e-01 -3.29895705e-01 -7.54233062e-01 -6.56636238e-01
6.12952672e-02 -6.51612580e-01 9.03203487e-02 4.58710730e-01
1.52915323e+00 -7.50767052e-01 -4.28182572e-01 8.80271792e-01
1.61569369e+00 8.75952989e-02 8.31014216e-02 4.03742224e-01
2.63399661e-01 1.12073295e-01 7.35738814e-01 1.25713861e+00
2.89910644e-01 2.70720959e-01 6.38437688e-01 -8.87515675e-03
4.24386531e-01 -5.45444302e-02 -2.22064376e-01 4.19477254e-01
3.15037191e-01 -3.45217615e-01 -8.51114810e-01 3.80139142e-01
-2.06640601e+00 -5.67743540e-01 5.26888132e-01 2.26984096e+00
1.02019072e+00 -5.67774892e-01 1.48877636e-01 -4.01571453e-01
5.86746752e-01 -1.08280577e-01 -1.24272227e+00 -3.94252717e-01
-2.48311404e-02 -9.63536650e-03 8.50700140e-01 -1.20949090e-01
-1.10611165e+00 5.72442353e-01 5.47237492e+00 8.44411671e-01
-1.22816253e+00 4.71801162e-01 8.45225036e-01 -4.00161475e-01
-2.08050072e-01 -4.03168648e-01 -2.40495339e-01 2.22345874e-01
8.52553308e-01 -3.51324201e-01 6.16738141e-01 1.18786883e+00
-1.60051629e-01 -7.52849728e-02 -9.26977217e-01 1.28609574e+00
-3.15585405e-01 -1.69779980e+00 -2.75016934e-01 -6.30289689e-02
8.41172040e-01 5.10334849e-01 2.08216712e-01 -3.13376933e-02
7.65930951e-01 -8.98941398e-01 7.49832690e-02 1.92163095e-01
5.00482857e-01 -9.97702122e-01 6.97026551e-01 4.77455914e-01
-4.64676738e-01 -3.84634107e-01 -2.83407271e-01 8.47407222e-01
-8.45662728e-02 2.10101485e-01 -1.09671724e+00 6.54445648e-01
9.60675836e-01 1.41140535e-01 -2.16480643e-01 1.34585583e+00
3.19433451e-01 6.48290753e-01 -5.21965384e-01 -2.35385001e-01
3.10349941e-01 -6.94991499e-02 4.95946646e-01 6.65633082e-01
6.87966719e-02 2.41228163e-01 9.31802467e-02 2.06394210e-01
-2.52073675e-01 6.72039807e-01 -3.81594986e-01 1.14985190e-01
5.88084340e-01 1.58090365e+00 -5.74169636e-01 -8.19213539e-02
-2.73183554e-01 6.60730302e-01 2.90230274e-01 1.94668278e-01
-8.09633672e-01 -8.73811264e-03 6.01040542e-01 -3.48654181e-01
4.59193319e-01 3.78695846e-01 -6.97706779e-03 -1.02085304e+00
-3.44058245e-01 -1.38941228e+00 1.25173295e+00 2.29414869e-02
-1.68509829e+00 7.25503564e-01 -2.11212590e-01 -1.39637852e+00
-5.67164898e-01 9.91399139e-02 -3.35277498e-01 2.93747008e-01
-1.44391131e+00 -1.02318847e+00 -2.32516855e-01 1.29469645e+00
8.66565183e-02 -5.22007883e-01 1.13455534e+00 7.75464103e-02
-5.66502452e-01 1.16009307e+00 6.86066449e-01 -1.83432400e-01
6.95110679e-01 -8.33661079e-01 -4.88897651e-01 1.91287503e-01
-2.26832423e-02 2.78144509e-01 4.73830968e-01 -4.05323565e-01
-1.89405549e+00 -1.18692648e+00 6.51089996e-02 3.03767323e-01
5.68088531e-01 2.67203748e-02 -8.01784515e-01 5.16757071e-01
2.62673974e-01 7.63967574e-01 1.01128280e+00 -2.97261059e-01
3.60813923e-02 -3.52266818e-01 -1.93970573e+00 5.29036224e-01
4.02854562e-01 1.61564406e-02 1.51720330e-01 8.74026775e-01
5.25915682e-01 -6.82663918e-01 -1.36478853e+00 1.47342652e-01
3.15558463e-01 -6.17926002e-01 8.65752459e-01 -7.26426244e-01
-3.38469058e-01 1.12146204e-02 -5.29475957e-02 -1.29002023e+00
9.80971828e-02 -1.14606273e+00 -1.82660460e-01 7.95697868e-01
3.33507836e-01 -1.14241719e+00 1.27301073e+00 1.11644185e+00
4.93606925e-01 -1.02332199e+00 -1.27989924e+00 -5.11590421e-01
-1.24967404e-01 -6.04424812e-02 1.18798459e+00 1.12445390e+00
1.91409141e-02 -4.53608036e-01 -1.87707379e-01 1.32752329e-01
1.07488859e+00 2.90757507e-01 6.94507003e-01 -6.53095603e-01
-6.94924235e-01 -8.72515291e-02 -8.86471048e-02 -2.08112672e-02
1.39225453e-01 -8.66556823e-01 -1.95519403e-01 -8.54410648e-01
1.41821936e-01 -1.10461354e+00 -6.40555263e-01 7.35346138e-01
5.29091395e-02 -1.68454349e-01 1.25753656e-01 3.68434399e-01
-8.53178918e-01 5.86369455e-01 1.03143454e+00 -1.49443865e-01
-2.09663093e-01 5.30151129e-01 -3.88104349e-01 5.11102259e-01
9.27265227e-01 -7.71944046e-01 -6.60009801e-01 -3.02840352e-01
-1.81733556e-02 6.29910529e-01 5.73438965e-02 -5.57493448e-01
6.85248554e-01 -3.27542871e-01 3.27286452e-01 -3.59390050e-01
-6.79159686e-02 -1.27315032e+00 5.57883799e-01 6.94480181e-01
-5.45718849e-01 3.71359512e-02 -1.36846662e-01 8.44685376e-01
-1.25125572e-01 -2.04634011e-01 8.82907987e-01 -2.07327142e-01
-2.84693182e-01 8.08122993e-01 -2.57977486e-01 4.51640710e-02
1.56200647e+00 1.84734643e-01 -2.73631305e-01 -3.53090346e-01
-6.83331728e-01 6.58190846e-01 4.14345920e-01 -5.06516509e-02
6.29506528e-01 -1.07662642e+00 -6.06600523e-01 2.27532595e-01
2.35939603e-02 7.78892040e-02 5.79501152e-01 1.03270078e+00
-6.25207365e-01 1.16074413e-01 -1.31922722e-01 -5.80320954e-01
-1.25911391e+00 8.57888758e-01 5.78252852e-01 -6.24570847e-01
-8.32410336e-01 6.97551072e-01 -1.09639987e-01 -6.09139383e-01
5.91236770e-01 1.06300041e-01 4.56437096e-02 -2.53640145e-01
4.36402529e-01 4.21664506e-01 3.28280151e-01 -1.67079598e-01
-4.35846657e-01 -1.58529252e-01 -2.67809987e-01 -2.49987170e-02
1.37129378e+00 -6.77743927e-02 -4.38001612e-03 -2.43855953e-01
1.21431792e+00 -3.88838232e-01 -1.13499665e+00 -6.24152601e-01
-1.90845311e-01 -5.93720138e-01 1.85207456e-01 -8.99721146e-01
-1.51922989e+00 1.72966182e-01 9.33342755e-01 -1.83633029e-01
1.22578931e+00 -3.78328383e-01 9.56660390e-01 4.72519726e-01
7.19572723e-01 -1.08621860e+00 -2.64828473e-01 -1.86383039e-01
4.44849759e-01 -1.34234965e+00 2.83147424e-01 6.57957643e-02
-1.05536199e+00 9.21992302e-01 4.90433842e-01 6.76649809e-02
7.62847245e-01 3.60144645e-01 3.46979767e-01 -1.00083493e-01
-9.69980359e-01 5.54235637e-01 -3.32565904e-01 2.23212436e-01
-3.64160627e-01 1.67439073e-01 -2.94328213e-01 6.44876301e-01
1.85795069e-01 2.00415209e-01 1.87156424e-01 1.28216207e+00
1.53548330e-01 -1.14328933e+00 -4.79416192e-01 5.39411724e-01
-9.60152090e-01 3.45196962e-01 -7.39550823e-03 6.32066488e-01
-2.92194754e-01 7.13052750e-01 -3.09708446e-01 9.36878845e-02
-9.49211195e-02 -1.17853418e-01 3.20393533e-01 -1.33376522e-02
-1.30315876e+00 1.24292403e-01 -2.37060726e-01 -8.41997206e-01
-2.63513952e-01 -8.89498711e-01 -1.60887909e+00 -2.15383172e-01
-3.43709737e-01 4.44045514e-01 1.07387578e+00 5.19030273e-01
5.71365774e-01 9.00121406e-02 1.05358958e+00 -2.57789969e-01
-1.31993020e+00 -2.81642556e-01 -6.24704897e-01 3.86710912e-01
1.82314456e-01 -2.50402510e-01 -1.87099144e-01 -2.33800218e-01] | [6.072597980499268, 6.452235221862793] |
80d5623e-ed4e-4254-baa6-758b88326ad9 | cdiffmr-can-we-replace-the-gaussian-noise | 2306.14350 | null | https://arxiv.org/abs/2306.14350v1 | https://arxiv.org/pdf/2306.14350v1.pdf | CDiffMR: Can We Replace the Gaussian Noise with K-Space Undersampling for Fast MRI? | Deep learning has shown the capability to substantially accelerate MRI reconstruction while acquiring fewer measurements. Recently, diffusion models have gained burgeoning interests as a novel group of deep learning-based generative methods. These methods seek to sample data points that belong to a target distribution from a Gaussian distribution, which has been successfully extended to MRI reconstruction. In this work, we proposed a Cold Diffusion-based MRI reconstruction method called CDiffMR. Different from conventional diffusion models, the degradation operation of our CDiffMR is based on \textit{k}-space undersampling instead of adding Gaussian noise, and the restoration network is trained to harness a de-aliaseing function. We also design starting point and data consistency conditioning strategies to guide and accelerate the reverse process. More intriguingly, the pre-trained CDiffMR model can be reused for reconstruction tasks with different undersampling rates. We demonstrated, through extensive numerical and visual experiments, that the proposed CDiffMR can achieve comparable or even superior reconstruction results than state-of-the-art models. Compared to the diffusion model-based counterpart, CDiffMR reaches readily competing results using only $1.6 \sim 3.4\%$ for inference time. The code is publicly available at https://github.com/ayanglab/CDiffMR. | ['Guang Yang', 'Carola-Bibiane Schönlieb', 'Angelica Aviles-Rivero', 'Jiahao Huang'] | 2023-06-25 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 3.27136740e-02 4.87258509e-02 2.83205826e-02 -2.61967719e-01
-1.00950968e+00 -9.01368409e-02 6.70753479e-01 -1.55347630e-01
-4.95864540e-01 7.23130286e-01 2.50026703e-01 -2.00902164e-01
-2.43201420e-01 -6.93582892e-01 -6.87548161e-01 -1.18606877e+00
-8.75077918e-02 6.35588646e-01 1.81132793e-01 1.20127894e-01
3.26108038e-02 4.93551046e-01 -7.79757798e-01 -1.55356275e-02
1.05723393e+00 8.56856585e-01 4.70610142e-01 3.76888245e-01
1.53506681e-01 7.19211340e-01 -2.90785730e-01 -1.94546819e-01
1.47846356e-01 -5.36805689e-01 -6.00151777e-01 -2.42638793e-02
-5.10320105e-02 -5.29795349e-01 -7.90642202e-01 1.09955692e+00
1.00831413e+00 1.91771746e-01 7.43569195e-01 -8.39800298e-01
-8.82152259e-01 5.80897808e-01 -6.04419231e-01 5.25075376e-01
-4.59022373e-02 1.81510061e-01 3.76114190e-01 -9.88075435e-01
6.05810225e-01 8.60105932e-01 5.46316624e-01 7.33627260e-01
-1.49417436e+00 -7.10009992e-01 -1.33886635e-01 1.85269415e-01
-1.39327645e+00 -3.32222670e-01 8.43546689e-01 -5.17063081e-01
5.58189631e-01 1.05756328e-01 5.94481349e-01 1.31971872e+00
4.06600714e-01 8.81278336e-01 1.40893781e+00 -1.00030839e-01
4.40844506e-01 -2.06417158e-01 -8.88931900e-02 5.17294407e-01
8.91504362e-02 2.19331995e-01 -2.77888119e-01 -2.64056057e-01
1.16069853e+00 1.87025338e-01 -6.13817751e-01 -4.11439180e-01
-1.24620914e+00 1.02267838e+00 7.19947994e-01 3.88286829e-01
-5.38778424e-01 2.85989970e-01 2.63742954e-01 -1.09866977e-01
6.97132766e-01 1.13018796e-01 5.81225157e-02 -4.20800373e-02
-1.22101343e+00 2.54909009e-01 4.72126216e-01 6.54974043e-01
4.31263357e-01 2.17560202e-01 -3.06820899e-01 1.00703859e+00
3.00581694e-01 6.53422594e-01 6.86335921e-01 -8.57751191e-01
1.72412917e-01 -4.88566272e-02 1.31926145e-02 -9.02747035e-01
-4.49781209e-01 -8.75984967e-01 -1.44960856e+00 7.62908831e-02
4.12548423e-01 -1.32731810e-01 -1.00919449e+00 1.67508113e+00
3.63258779e-01 4.72791433e-01 -1.07283868e-01 1.14271033e+00
5.71225882e-01 5.91835856e-01 -1.38454705e-01 -2.35623732e-01
9.76546645e-01 -7.84329534e-01 -6.96720719e-01 5.26378453e-02
3.61363798e-01 -7.51161158e-01 9.79960024e-01 6.82853103e-01
-1.30631185e+00 -3.48609179e-01 -8.66454899e-01 1.29934147e-01
3.28706533e-01 2.06379797e-02 5.91982782e-01 5.08464158e-01
-1.22227633e+00 8.12441111e-01 -1.22629845e+00 8.27488303e-02
7.29518056e-01 7.70367607e-02 -1.34005189e-01 -4.85063016e-01
-1.03350949e+00 7.61462629e-01 -1.86900180e-02 2.78010309e-01
-1.25570321e+00 -8.62186134e-01 -6.24534428e-01 -2.78871149e-01
1.69301555e-01 -7.66090214e-01 1.07762849e+00 -4.90610689e-01
-1.51317000e+00 6.63093150e-01 -1.11424059e-01 -6.21047854e-01
7.54033089e-01 -1.74623847e-01 -2.94459581e-01 3.65693390e-01
7.42471516e-02 4.46056962e-01 1.15493846e+00 -1.24444687e+00
1.95776418e-01 -3.93183649e-01 -3.53417814e-01 -5.43688349e-02
6.91262484e-02 -2.14667201e-01 -3.91618192e-01 -1.04770494e+00
4.34150457e-01 -1.00386500e+00 -5.05657077e-01 1.31511986e-01
-4.33013618e-01 7.93418065e-02 3.79271775e-01 -7.55634546e-01
1.11141896e+00 -1.93507147e+00 3.35494787e-01 1.77670538e-01
6.52658701e-01 1.73165694e-01 6.39559776e-02 3.40096653e-01
-1.18440136e-01 -1.99823081e-01 -6.66601241e-01 -4.00503695e-01
-7.79978409e-02 2.08263382e-01 -1.72699526e-01 9.39149857e-01
-3.66945267e-02 8.55609059e-01 -9.46143389e-01 -2.18372211e-01
1.64386570e-01 9.09532070e-01 -7.24711001e-01 9.84260291e-02
2.01405495e-01 8.70701849e-01 -3.59547287e-01 3.19128394e-01
1.04000914e+00 -3.79741967e-01 1.84444591e-01 -2.85492510e-01
2.19632685e-01 8.31995718e-03 -1.00178194e+00 1.98446715e+00
-4.68694627e-01 3.20671856e-01 6.24698363e-02 -1.27087510e+00
9.08778667e-01 1.35046467e-01 5.74264050e-01 -7.27367997e-01
6.81717992e-02 3.67395461e-01 1.91553548e-01 -4.06696081e-01
1.07999198e-01 -5.93090951e-01 2.94348449e-01 5.36322951e-01
8.89647976e-02 -2.76944399e-01 -5.69442958e-02 2.55378246e-01
1.08757091e+00 1.02953121e-01 -1.90194789e-02 -4.87296849e-01
2.49995574e-01 -4.03201222e-01 4.16624516e-01 8.07638943e-01
-1.04449064e-01 8.91351402e-01 1.71164393e-01 -2.10475460e-01
-9.45841730e-01 -1.44579148e+00 -5.75915098e-01 4.59080338e-01
1.03932336e-01 -7.33086988e-02 -8.33634853e-01 -4.84748811e-01
-2.66449332e-01 6.78085923e-01 -6.43953025e-01 -2.39042431e-01
-6.20322287e-01 -1.24932134e+00 4.59962368e-01 5.56507111e-01
4.89009827e-01 -8.09732199e-01 -3.54198039e-01 3.87809455e-01
-2.19490990e-01 -7.59480357e-01 -5.29755533e-01 1.64164439e-01
-9.91877675e-01 -6.92034900e-01 -1.31457090e+00 -5.06727278e-01
8.17300737e-01 1.12310037e-01 9.07746434e-01 2.93789934e-02
-2.80972958e-01 3.06488454e-01 -2.94421226e-01 1.07315872e-02
-4.66538161e-01 -6.26097471e-02 3.80438641e-02 8.22135340e-03
-2.31822487e-02 -9.93956745e-01 -1.18671715e+00 2.28786349e-01
-1.18028605e+00 1.11097410e-01 5.13212800e-01 1.05566311e+00
8.07475269e-01 -1.19032681e-01 7.11713314e-01 -6.88908160e-01
6.22064471e-01 -7.72455215e-01 -3.42975020e-01 -2.34989703e-01
-6.70988500e-01 2.31120616e-01 5.44840574e-01 -6.19876266e-01
-8.14794600e-01 -3.26844871e-01 -6.76100969e-01 -5.88424504e-01
1.08925663e-01 4.15800899e-01 1.25235185e-01 7.72336051e-02
5.32035708e-01 6.27498031e-01 3.45435202e-01 -6.02000952e-01
2.77424186e-01 3.36665720e-01 4.01653558e-01 -6.96468472e-01
5.57262957e-01 8.23615432e-01 -7.39429817e-02 -6.87578738e-01
-7.13723779e-01 -2.06300646e-01 -3.73573542e-01 -2.73387343e-01
7.91074038e-01 -7.04500914e-01 -4.22820061e-01 7.42073059e-01
-8.62931907e-01 -6.39194548e-01 -4.07695681e-01 6.95685565e-01
-6.12186253e-01 5.54763854e-01 -8.97559524e-01 -5.48871577e-01
-5.37172198e-01 -1.46839094e+00 9.11983669e-01 -4.11193334e-02
1.21795917e-02 -1.08648050e+00 1.27743796e-01 2.09509045e-01
7.75239944e-01 1.36317149e-01 6.88958049e-01 -4.84720439e-01
-4.77581710e-01 -3.50369550e-02 -1.38786346e-01 6.01552308e-01
2.84277480e-02 -5.63877761e-01 -7.16864705e-01 -5.85490048e-01
4.74909574e-01 -1.54037222e-01 9.08951461e-01 7.62713790e-01
1.44367218e+00 -2.14434236e-01 -2.38965452e-01 8.43591154e-01
1.37941933e+00 1.84800699e-01 9.00572598e-01 1.33182496e-01
5.32774985e-01 9.40501038e-03 1.31499335e-01 6.10061347e-01
2.58235157e-01 7.20697582e-01 3.06532592e-01 -2.30805546e-01
-3.51234168e-01 -1.41379312e-01 -3.06997262e-02 1.03227818e+00
-6.92054033e-02 -8.63934308e-02 -9.41464782e-01 4.25331414e-01
-1.61442912e+00 -9.48127627e-01 -1.59922034e-01 2.21516871e+00
9.70779836e-01 8.16917568e-02 3.95175244e-04 4.57003564e-02
5.42672813e-01 1.68726608e-01 -7.36971378e-01 2.22097039e-01
8.24730396e-02 5.18209100e-01 4.00241286e-01 4.15308625e-01
-7.10778177e-01 4.43936318e-01 5.65067911e+00 1.10783601e+00
-1.34796107e+00 6.60284936e-01 7.39150941e-01 -1.22774698e-01
-5.23274839e-01 -2.51129210e-01 -4.68489438e-01 6.60251439e-01
6.64921522e-01 -6.98358864e-02 4.66273814e-01 6.21849477e-01
4.20048505e-01 -1.40844196e-01 -7.36960948e-01 1.11351037e+00
9.26268324e-02 -1.43850744e+00 -9.66121852e-02 1.71787053e-01
6.80245876e-01 3.76377910e-01 1.98302969e-01 1.41034856e-01
2.90648669e-01 -1.15670514e+00 6.66696906e-01 7.51476228e-01
7.94788003e-01 -6.22298896e-01 5.76610684e-01 4.07419115e-01
-6.97332203e-01 2.56743282e-01 -4.72306877e-01 3.35762501e-01
4.52202827e-01 1.12492430e+00 -5.78283846e-01 7.03119457e-01
6.56504095e-01 6.47710502e-01 -2.15696454e-01 9.91404355e-01
-1.69185087e-01 7.23487020e-01 -2.21693829e-01 2.89027840e-01
1.42500252e-01 -4.27665174e-01 5.55125952e-01 1.01818800e+00
5.84194779e-01 1.15898825e-01 2.56816968e-02 1.03484249e+00
3.34106609e-02 -1.52892068e-01 -2.74419636e-01 2.85657853e-01
3.04758847e-01 1.16864836e+00 -7.05748916e-01 -4.13643330e-01
-9.93087366e-02 1.12552726e+00 2.91649699e-01 3.93221498e-01
-1.13029945e+00 5.17487414e-02 4.12131071e-01 4.65704739e-01
3.78928423e-01 -4.52829123e-01 -6.34315088e-02 -1.28738260e+00
-1.24715574e-01 -7.17065394e-01 1.15788445e-01 -7.43291795e-01
-1.52507031e+00 8.75990927e-01 6.47109151e-02 -1.15573466e+00
-5.08478433e-02 -2.83485264e-01 -5.00008821e-01 9.74892557e-01
-1.40111876e+00 -8.12416613e-01 -3.42774332e-01 7.15712667e-01
3.91678184e-01 8.86452124e-02 5.93416214e-01 5.39686084e-01
-4.98030096e-01 3.58411938e-01 4.67981398e-01 -2.01053899e-02
4.90963042e-01 -1.16571403e+00 1.27161592e-01 8.31865311e-01
-1.11277342e-01 7.21105278e-01 7.90627718e-01 -6.81682944e-01
-1.35342729e+00 -1.04062247e+00 1.71049073e-01 -1.10631771e-01
7.48728871e-01 -1.79646000e-01 -1.13276958e+00 5.22132099e-01
2.16645882e-01 3.67559940e-01 5.83167791e-01 -4.59541678e-01
-4.18662727e-02 -1.51312754e-01 -1.27599347e+00 4.51185882e-01
1.06704092e+00 -3.19676608e-01 -3.78970653e-01 3.51265103e-01
3.60836595e-01 -5.83594084e-01 -1.09121168e+00 3.05637181e-01
2.78066814e-01 -9.50233936e-01 1.11276591e+00 -1.38245061e-01
4.82232541e-01 -2.10377634e-01 -9.90081728e-02 -1.58013761e+00
-3.56145859e-01 -5.59475064e-01 -2.91849285e-01 7.86257744e-01
1.09888144e-01 -7.93822646e-01 5.81764221e-01 3.46552581e-01
-2.90002316e-01 -1.07155073e+00 -1.14812469e+00 -8.81870449e-01
5.43321490e-01 -5.60739815e-01 2.89948642e-01 8.08201253e-01
-3.59220892e-01 4.63048890e-02 -3.72537911e-01 1.21781103e-01
9.60124314e-01 -7.44764181e-03 4.12793100e-01 -7.31159747e-01
-6.52334332e-01 -3.29930007e-01 -7.22498074e-02 -1.37607551e+00
-1.96461771e-02 -1.27786493e+00 -2.77182888e-02 -1.74726653e+00
1.93518966e-01 -8.27700257e-01 -2.87602723e-01 1.19248800e-01
-8.39785561e-02 3.69291514e-01 6.77174479e-02 3.18361521e-01
-1.15678810e-01 8.53184640e-01 1.68648207e+00 -5.30017503e-02
1.63905490e-02 6.47863448e-02 -5.87307632e-01 4.71113592e-01
9.64194059e-01 -5.45536280e-01 -5.66439152e-01 -4.63549644e-01
-2.04799641e-02 2.97389030e-01 5.86302102e-01 -1.07008517e+00
1.19855605e-01 1.68122977e-01 4.44682837e-01 -4.81431365e-01
3.55390400e-01 -5.29684424e-01 3.71691883e-01 7.21506238e-01
-1.92405820e-01 -1.60660979e-03 2.15499867e-02 5.91273069e-01
1.57560464e-02 -3.00201297e-01 1.08444929e+00 -1.83191746e-01
-1.91997513e-01 5.76594949e-01 -4.47522432e-01 1.84453025e-01
7.62347937e-01 1.77398562e-01 -1.78357899e-01 -3.60995710e-01
-1.08970392e+00 -2.78564274e-01 2.90408373e-01 4.34160754e-02
8.92149270e-01 -1.38622653e+00 -8.06959987e-01 2.23817512e-01
-3.57185483e-01 8.40471983e-02 7.48495936e-01 1.46212780e+00
-4.14211929e-01 3.85077251e-03 6.77626953e-02 -7.91367888e-01
-5.26213408e-01 6.79800868e-01 5.27614534e-01 -4.69996423e-01
-1.14326704e+00 7.46384680e-01 2.94623047e-01 -3.17846447e-01
-6.98582828e-02 -3.99456769e-01 2.93074489e-01 -2.72194207e-01
6.15649462e-01 3.82757962e-01 1.36640847e-01 -2.60315806e-01
-3.27181101e-01 2.87956774e-01 -1.58426404e-01 -2.63518989e-01
1.56249213e+00 -5.31665124e-02 1.97001118e-02 1.17769770e-01
1.15280128e+00 8.09519216e-02 -1.50849831e+00 -2.94683307e-01
-3.98881614e-01 -3.88149559e-01 3.14934671e-01 -7.58346856e-01
-1.46223617e+00 8.37303460e-01 6.81320369e-01 9.28236842e-02
1.16228926e+00 1.45576984e-01 8.83074760e-01 -2.98935264e-01
5.36218107e-01 -6.59792900e-01 1.90544412e-01 9.18879136e-02
1.14556956e+00 -1.03387463e+00 5.80480620e-02 -8.47383589e-02
-5.56620777e-01 8.21633160e-01 2.33264267e-01 -4.69628274e-01
8.47373366e-01 5.23115098e-01 -1.18766025e-01 -2.49495357e-01
-3.66629064e-01 1.70783117e-01 1.82125792e-01 6.07393146e-01
3.62594306e-01 8.11452493e-02 -4.73775744e-01 6.25959992e-01
6.25415444e-02 2.65164107e-01 4.72001851e-01 7.97855198e-01
-9.65610966e-02 -1.07186174e+00 -2.73370296e-01 5.99957108e-01
-4.09292430e-01 -3.51112515e-01 5.39481699e-01 6.59960270e-01
-2.08210036e-01 4.86629337e-01 -8.47365260e-02 -6.46746010e-02
7.06283525e-02 -2.76582241e-01 7.44374454e-01 -3.30823720e-01
-2.72331595e-01 3.63283426e-01 -3.71512145e-01 -6.19720340e-01
-4.04031307e-01 -7.52888501e-01 -1.31588852e+00 -2.96206951e-01
-1.50156021e-01 1.11041255e-01 5.35209775e-01 8.32124710e-01
3.97469491e-01 6.47114873e-01 4.69203591e-01 -9.86049712e-01
-7.29859769e-01 -1.00465095e+00 -7.27768540e-01 2.72759050e-01
2.33282104e-01 -7.85473049e-01 -3.46130699e-01 -2.11942881e-01] | [13.48302173614502, -2.3713948726654053] |
7e7c41aa-a9bf-48c3-9fd3-2bfeae445ae4 | self-attention-encoding-and-pooling-for | 2008.01077 | null | https://arxiv.org/abs/2008.01077v1 | https://arxiv.org/pdf/2008.01077v1.pdf | Self-attention encoding and pooling for speaker recognition | The computing power of mobile devices limits the end-user applications in terms of storage size, processing, memory and energy consumption. These limitations motivate researchers for the design of more efficient deep models. On the other hand, self-attention networks based on Transformer architecture have attracted remarkable interests due to their high parallelization capabilities and strong performance on a variety of Natural Language Processing (NLP) applications. Inspired by the Transformer, we propose a tandem Self-Attention Encoding and Pooling (SAEP) mechanism to obtain a discriminative speaker embedding given non-fixed length speech utterances. SAEP is a stack of identical blocks solely relied on self-attention and position-wise feed-forward networks to create vector representation of speakers. This approach encodes short-term speaker spectral features into speaker embeddings to be used in text-independent speaker verification. We have evaluated this approach on both VoxCeleb1 & 2 datasets. The proposed architecture is able to outperform the baseline x-vector, and shows competitive performance to some other benchmarks based on convolutions, with a significant reduction in model size. It employs 94%, 95%, and 73% less parameters compared to ResNet-34, ResNet-50, and x-vector, respectively. This indicates that the proposed fully attention based architecture is more efficient in extracting time-invariant features from speaker utterances. | ['Javier Hernando', 'Pooyan Safari', 'Miquel India'] | 2020-08-03 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [ 4.13923636e-02 -1.71630859e-01 1.52523875e-01 -5.84807336e-01
-6.27033830e-01 -1.13017187e-01 4.13211942e-01 -1.96365252e-01
-5.69175482e-01 2.47409940e-01 4.16789562e-01 -3.10844988e-01
1.47729889e-01 -4.65447992e-01 -4.80584949e-01 -8.18624914e-01
1.20324008e-01 -8.16242993e-02 -3.88953872e-02 -4.91131768e-02
1.57401919e-01 4.81326878e-01 -1.63056350e+00 -5.90466009e-03
5.78156412e-01 1.08649921e+00 3.86201084e-01 6.29833937e-01
-2.49087825e-01 4.76886719e-01 -5.45666754e-01 -3.61105829e-01
-1.82186559e-01 -2.10771963e-01 -5.31532526e-01 -2.33833775e-01
2.77990311e-01 -1.66704297e-01 -4.75433439e-01 9.61158633e-01
1.01744103e+00 3.68109941e-01 4.28065836e-01 -1.17201436e+00
-9.24381793e-01 8.24475586e-01 -3.78952891e-01 5.87722778e-01
9.20091048e-02 2.64621880e-02 8.76691401e-01 -1.04173946e+00
-1.87106319e-02 1.28363883e+00 7.05385566e-01 6.87531888e-01
-8.04971039e-01 -8.86365116e-01 1.39158309e-01 6.46421671e-01
-1.60441947e+00 -1.04908276e+00 9.44125831e-01 -2.15054359e-02
1.61426115e+00 5.64547062e-01 3.08224291e-01 1.26058638e+00
1.07115388e-01 6.79961145e-01 6.68023825e-01 -4.76197243e-01
2.03945652e-01 4.52127516e-01 4.42866266e-01 5.69741070e-01
-2.57650703e-01 -1.33351222e-01 -7.58023739e-01 -4.06195633e-02
3.46219242e-01 1.19921416e-01 -3.14580292e-01 2.64609903e-01
-1.01053715e+00 7.88020670e-01 5.27163625e-01 6.73619866e-01
-5.22314727e-01 1.06313817e-01 6.78389370e-01 1.51711822e-01
4.65595990e-01 1.68131843e-01 -4.15329486e-01 -2.30238125e-01
-1.04272783e+00 -1.32406741e-01 4.29494262e-01 7.62804627e-01
3.52217048e-01 7.41194606e-01 -2.79566973e-01 1.16213787e+00
4.11315620e-01 5.48264027e-01 1.35887241e+00 -9.73453969e-02
4.21903342e-01 4.32251841e-01 -4.70635414e-01 -8.77183795e-01
-3.75656039e-01 -6.59364223e-01 -1.10002959e+00 -2.09238827e-01
-3.54883730e-01 -5.45189157e-02 -1.06703055e+00 1.82103682e+00
1.39859214e-01 4.37821925e-01 2.93242127e-01 9.38215196e-01
1.17313600e+00 1.06485569e+00 1.38193771e-01 -5.65067604e-02
1.69074428e+00 -1.11230338e+00 -9.43337739e-01 -1.06298842e-01
5.54992817e-02 -7.05221951e-01 1.03005445e+00 1.26773352e-02
-1.00740576e+00 -9.16292846e-01 -1.21370208e+00 -2.38183528e-01
-5.11548936e-01 3.24670225e-01 3.28908205e-01 9.56058800e-01
-1.24464738e+00 4.07493919e-01 -8.28011155e-01 -3.52964759e-01
3.55141193e-01 5.47247231e-01 -2.86035538e-01 4.95475262e-01
-1.21478343e+00 6.79731250e-01 1.57085225e-01 4.44335938e-01
-7.24748492e-01 -6.75050139e-01 -9.80945885e-01 6.94078565e-01
-2.72031963e-01 -3.32070291e-01 1.09532166e+00 -8.82930458e-01
-2.10651469e+00 4.84960347e-01 -4.81180906e-01 -7.60654569e-01
-6.91561922e-02 -1.06399126e-01 -7.11629510e-01 -1.55432105e-01
-3.34466279e-01 5.35416126e-01 1.04386556e+00 -2.38425165e-01
-3.78499091e-01 -5.12207866e-01 -3.80633354e-01 1.39144450e-01
-1.02081013e+00 4.45766121e-01 -2.66460389e-01 -6.62728071e-01
9.54125375e-02 -7.49796033e-01 1.04840197e-01 -4.32051986e-01
-5.73841155e-01 -6.29355013e-01 1.15095615e+00 -8.94114852e-01
1.25938582e+00 -2.35001373e+00 1.15068248e-02 -7.15928748e-02
-4.08224426e-02 6.75211549e-01 -1.77488938e-01 1.44668832e-01
-2.65559971e-01 6.32686615e-02 1.56063540e-02 -6.56283140e-01
1.65570751e-01 -2.19482109e-01 -4.85589087e-01 5.03280938e-01
1.42346844e-01 8.08290482e-01 -1.70587420e-01 -4.45830196e-01
4.88323599e-01 9.88769650e-01 -4.66919810e-01 3.43433827e-01
2.91791230e-01 4.37338538e-02 -2.91889638e-01 4.97874141e-01
8.24794054e-01 1.41263470e-01 -9.79831070e-02 -1.84088632e-01
-1.66469425e-01 5.45611143e-01 -8.33066702e-01 1.75147450e+00
-8.65123332e-01 1.00184190e+00 1.77318528e-01 -1.11020112e+00
1.11442041e+00 7.69257486e-01 1.31418586e-01 -6.27847135e-01
4.65488613e-01 5.85656101e-03 -4.31198515e-02 -6.12312376e-01
3.86698991e-01 -1.20204844e-01 7.92301893e-02 2.60374218e-01
4.55564082e-01 3.77405018e-01 -2.88316131e-01 -2.29699776e-01
8.79568517e-01 -3.46684933e-01 7.00410977e-02 -3.58951718e-01
1.03811729e+00 -7.64221847e-01 5.16750157e-01 2.91913033e-01
-3.01102757e-01 4.18495595e-01 -5.75475320e-02 -4.15323168e-01
-8.91160190e-01 -7.49991000e-01 -2.88451195e-01 1.46958220e+00
-1.82419911e-01 -3.32209855e-01 -8.62188339e-01 -3.40581119e-01
-2.06982568e-01 7.31846631e-01 -4.73507673e-01 -1.52207106e-01
-6.55307591e-01 -7.01860905e-01 8.02908778e-01 7.19732344e-01
6.76011145e-01 -1.32796443e+00 -5.67840755e-01 3.44577074e-01
-6.53671473e-02 -1.13238990e+00 -6.54297829e-01 1.62018403e-01
-6.49741590e-01 -4.27124858e-01 -9.45361197e-01 -9.91811275e-01
2.82003254e-01 1.48103476e-01 6.13912582e-01 -2.28349626e-01
-1.42870173e-01 2.31893230e-02 -2.68837541e-01 -7.07200885e-01
-8.61036703e-02 4.35766369e-01 3.57701987e-01 4.10733193e-01
6.58532321e-01 -6.92203999e-01 -5.41546881e-01 1.71907827e-01
-5.34156084e-01 -2.44712710e-01 5.23428857e-01 8.32608998e-01
2.32537284e-01 -1.17530331e-01 8.74721289e-01 -1.72672361e-01
6.96128190e-01 -2.91560054e-01 -5.17970860e-01 2.17264548e-01
-2.97932446e-01 2.02206254e-01 7.91875958e-01 -4.72409546e-01
-1.14447808e+00 -2.18340214e-02 -6.81651473e-01 -6.32103622e-01
-7.33085498e-02 1.31470576e-01 -3.33783299e-01 -1.49959428e-02
3.35420430e-01 5.52990496e-01 1.89332422e-02 -7.81454861e-01
-2.84306165e-02 1.30674970e+00 3.82122666e-01 9.24929157e-02
4.41532761e-01 -9.88275185e-03 -5.82273483e-01 -1.31714439e+00
-3.15128267e-01 -3.87650490e-01 -2.38680810e-01 1.59220919e-01
9.88926649e-01 -9.77688789e-01 -9.78193700e-01 5.75411379e-01
-1.20945466e+00 1.93580776e-01 -9.40091312e-02 8.17800820e-01
-1.58111781e-01 2.02490151e-01 -5.35448432e-01 -1.07166779e+00
-1.14108551e+00 -1.30810118e+00 9.95114684e-01 3.72102708e-01
-2.60050613e-02 -6.41840935e-01 -8.68620202e-02 2.37380683e-01
1.12671518e+00 -4.60945427e-01 6.48050368e-01 -9.78794396e-01
-2.57392615e-01 -2.95189440e-01 -1.13246307e-01 6.66531265e-01
2.23679729e-02 -1.82015926e-01 -1.67151010e+00 -4.01496798e-01
3.11567038e-01 -3.19278017e-02 9.37385499e-01 5.54665267e-01
1.48028660e+00 -4.08786714e-01 -3.34344029e-01 8.13694954e-01
1.16927254e+00 2.67181456e-01 5.55506706e-01 -3.65855694e-02
5.68152726e-01 2.82066226e-01 -1.72195375e-01 1.88450605e-01
1.87612817e-01 8.64616692e-01 2.28813991e-01 -7.23235011e-02
-2.18661472e-01 -1.08450586e-02 5.53135574e-01 1.23218620e+00
2.06119986e-03 -3.15338790e-01 -7.15180337e-01 7.00203955e-01
-1.37164617e+00 -1.04161477e+00 2.90873647e-01 2.07394409e+00
5.14220595e-01 -4.78924438e-02 -3.54441889e-02 4.16650385e-01
9.42287624e-01 3.38726282e-01 -5.94406307e-01 -8.25568676e-01
1.50053483e-02 5.43020308e-01 1.81337893e-01 3.25509250e-01
-1.02474177e+00 7.38058269e-01 5.99286175e+00 9.20881927e-01
-1.76965332e+00 5.17456651e-01 5.14477193e-01 -2.78035849e-01
8.05652365e-02 -6.71513796e-01 -9.53753591e-01 5.83708882e-01
1.60110092e+00 -2.30876893e-01 3.16396564e-01 1.07687747e+00
5.37966341e-02 7.21053839e-01 -9.22360182e-01 1.25039780e+00
3.56027812e-01 -1.21880794e+00 -1.32213488e-01 -1.15008898e-01
1.84956357e-01 2.74707854e-01 2.13841423e-01 5.73315978e-01
-2.77673393e-01 -1.10668480e+00 6.04424655e-01 7.37564042e-02
8.61568868e-01 -8.60240400e-01 1.06076324e+00 1.04674779e-01
-1.41558456e+00 -2.95907944e-01 -5.10305166e-01 -6.23215514e-04
4.99769524e-02 3.31294388e-01 -9.57482874e-01 3.42199177e-01
9.02270079e-01 3.69956583e-01 -3.09693784e-01 6.59785748e-01
1.74833864e-01 8.36631894e-01 -3.56723100e-01 -4.48988110e-01
2.27162912e-01 2.87007004e-01 4.81405795e-01 1.44160092e+00
4.46891308e-01 -1.18887514e-01 -4.98779684e-01 7.07619190e-01
-3.04612458e-01 3.40728194e-01 -3.10490131e-01 -1.22072166e-02
5.31812072e-01 1.36676323e+00 -3.00382733e-01 -5.03075838e-01
-3.07780504e-01 9.26150680e-01 2.61234373e-01 2.93068975e-01
-1.07710457e+00 -8.42098117e-01 9.55111265e-01 -9.55372453e-02
5.63687205e-01 -3.42522860e-02 -1.35955393e-01 -1.05820155e+00
1.71653003e-01 -6.72994375e-01 4.88297716e-02 -5.91354251e-01
-1.06298006e+00 1.40714157e+00 -4.73204255e-01 -1.03011644e+00
-2.71871865e-01 -5.09378314e-01 -8.88817132e-01 1.23198116e+00
-1.62171936e+00 -1.14583778e+00 -4.26999062e-01 6.86351299e-01
9.00604904e-01 -5.39204717e-01 1.11780107e+00 4.14889485e-01
-9.90542710e-01 1.07909977e+00 6.70799315e-02 1.70161381e-01
4.25639927e-01 -9.11730647e-01 6.88929319e-01 8.32562685e-01
2.99578756e-01 6.36581361e-01 4.05236095e-01 7.71694034e-02
-1.45971394e+00 -1.11558878e+00 1.08182263e+00 5.82671948e-02
1.95864663e-01 -6.20373070e-01 -9.95180130e-01 5.79925537e-01
6.66791677e-01 1.88266614e-03 8.15524399e-01 1.60665020e-01
-2.74489701e-01 -5.05676031e-01 -1.15498233e+00 2.37920940e-01
7.76545942e-01 -7.47445047e-01 -6.72615528e-01 1.71939284e-01
8.66270602e-01 -1.38992593e-01 -6.22950435e-01 1.98014855e-01
5.44704437e-01 -8.62232029e-01 9.73183155e-01 -3.75364512e-01
1.58431847e-02 5.01863360e-02 -2.34480560e-01 -1.14425933e+00
-5.12029827e-01 -6.29792988e-01 4.69039232e-02 1.55817807e+00
4.19015527e-01 -8.70025516e-01 6.16202116e-01 4.29850727e-01
-3.51084352e-01 -6.40289724e-01 -1.37991917e+00 -6.71102762e-01
-2.39815190e-01 -4.87947792e-01 1.09852076e+00 7.26095200e-01
9.76577029e-02 5.36059141e-01 -4.15589631e-01 2.25652218e-01
2.57773221e-01 -7.94943701e-03 4.67315555e-01 -9.80967164e-01
-7.87556469e-02 -5.11633813e-01 -7.32213020e-01 -9.34812963e-01
4.42378432e-01 -8.64430428e-01 1.04792509e-02 -1.14046144e+00
-2.92680655e-02 -2.14008749e-01 -7.85219789e-01 4.80593771e-01
-4.69115116e-02 3.23747456e-01 1.33477394e-02 -1.60012230e-01
-2.00487837e-01 9.16115701e-01 4.89283949e-01 -3.06325257e-01
-3.95833075e-01 -1.13294050e-02 -4.93766785e-01 5.85878134e-01
8.82608533e-01 -2.03543067e-01 -2.53115296e-01 -5.68542838e-01
-5.50201595e-01 -4.94948588e-02 1.45730942e-01 -1.20884609e+00
4.31097120e-01 3.95165145e-01 3.95066082e-01 -6.07874990e-01
6.84692264e-01 -8.04530919e-01 -1.00463614e-01 3.91989350e-01
-3.73491824e-01 1.06447510e-01 4.64787483e-01 1.21264383e-01
-4.72577721e-01 -1.79062754e-01 6.43461287e-01 2.04621047e-01
-5.84307551e-01 2.85831690e-01 -3.55517805e-01 -5.39871156e-01
8.77990663e-01 -1.79400176e-01 6.92823231e-02 -1.77930698e-01
-5.38406074e-01 -2.51754075e-01 -3.18811744e-01 7.46672392e-01
7.09392667e-01 -1.36775732e+00 -8.83485615e-01 8.38710368e-01
-8.12095255e-02 -3.55819911e-01 8.32900345e-01 6.04469895e-01
-2.59726167e-01 9.01940286e-01 -2.73446381e-01 -6.26083195e-01
-1.57664943e+00 5.09335697e-01 3.53485554e-01 8.98985565e-02
-5.95796525e-01 1.19316030e+00 1.48711592e-01 -2.23727643e-01
5.21270692e-01 -5.37875950e-01 -2.98531532e-01 -8.67602751e-02
8.67114842e-01 3.90210003e-01 3.78550917e-01 -7.96309352e-01
-7.34021127e-01 5.95642030e-01 -1.67980149e-01 -2.96943840e-02
1.48948824e+00 -4.09652777e-02 1.33252352e-01 2.21403718e-01
1.46940362e+00 1.94187760e-02 -7.52305150e-01 -3.04107070e-01
-3.41547489e-01 -2.20525429e-01 4.90135252e-01 -4.69396740e-01
-1.35279489e+00 1.27052689e+00 9.50397074e-01 2.91464061e-01
1.25421762e+00 -1.55441597e-01 9.80636597e-01 2.08333001e-01
-1.25677481e-01 -8.31786335e-01 -2.43415490e-01 2.71717459e-01
1.11706758e+00 -1.04541147e+00 -3.71727675e-01 1.13938814e-02
-4.51798290e-01 1.05198264e+00 6.15303695e-01 -6.27072155e-02
8.65489841e-01 3.68991852e-01 4.71690632e-02 6.69658557e-02
-7.69659221e-01 1.01962745e-01 4.18074191e-01 3.77206147e-01
6.50185049e-01 1.08939655e-01 5.88894375e-02 8.19530785e-01
-5.21548033e-01 -2.00092256e-01 -1.34481145e-02 4.49479967e-01
-2.76739895e-01 -7.26545632e-01 -3.88905138e-01 2.80817538e-01
-5.95374167e-01 -2.32756153e-01 4.19369340e-02 2.57887840e-01
-7.80982152e-02 1.05399108e+00 3.49179775e-01 -6.44459963e-01
3.30868334e-01 4.02458698e-01 1.07301317e-01 -2.73790896e-01
-9.07627642e-01 1.94868315e-02 -2.56953835e-01 -3.59269440e-01
-1.94978029e-01 -3.36955905e-01 -9.71543968e-01 -3.61262083e-01
-6.02882624e-01 2.71087229e-01 1.30748057e+00 6.86427951e-01
7.67233729e-01 8.81487012e-01 8.14773560e-01 -9.20815766e-01
-7.25649595e-01 -1.59645355e+00 -3.80058587e-01 7.84532949e-02
5.16719162e-01 -4.38883483e-01 -3.89545768e-01 -1.23764150e-01] | [14.367122650146484, 6.065057277679443] |
8203159a-68d5-4311-bb79-f7bbaa584382 | classifier-calibration-with-implications-to | 2102.05143 | null | https://arxiv.org/abs/2102.05143v3 | https://arxiv.org/pdf/2102.05143v3.pdf | Classifier Calibration: with application to threat scores in cybersecurity | This paper explores the calibration of a classifier output score in binary classification problems. A calibrator is a function that maps the arbitrary classifier score, of a testing observation, onto $[0,1]$ to provide an estimate for the posterior probability of belonging to one of the two classes. Calibration is important for two reasons; first, it provides a meaningful score, that is the posterior probability; second, it puts the scores of different classifiers on the same scale for comparable interpretation. The paper presents three main contributions: (1) Introducing multi-score calibration, when more than one classifier provides a score for a single observation. (2) Introducing the idea that the classifier scores to a calibration process are nothing but features to a classifier, hence proposing expanding the classifier scores to higher dimensions to boost the calibrator's performance. (3) Conducting a massive simulation study, in the order of 24,000 experiments, that incorporates different configurations, in addition to experimenting on two real datasets from the cybersecurity domain. The results show that there is no overall winner among the different calibrators and different configurations. However, general advices for practitioners include the following: the Platt's calibrator~\citep{Platt1999ProbabilisticOutputsForSupport}, a version of the logistic regression that decreases bias for a small sample size, has a very stable and acceptable performance among all experiments; our suggested multi-score calibration provides better performance than single score calibration in the majority of experiments, including the two real datasets. In addition, expanding the scores can help in some experiments. | ['William Briguglio', 'Issa Traore', 'Waleed A. Yousef'] | 2021-02-09 | null | null | null | null | ['classifier-calibration', 'classifier-calibration'] | ['computer-vision', 'miscellaneous'] | [ 9.64565724e-02 -1.57092214e-01 -3.39763135e-01 -6.55730844e-01
-6.21984601e-01 -5.71843326e-01 2.70336568e-01 2.61491895e-01
-3.93786371e-01 7.10147083e-01 -2.98124164e-01 -8.21083188e-01
-5.00615239e-01 -9.31621611e-01 -3.81740749e-01 -7.75963604e-01
-2.66914256e-02 4.73251998e-01 4.20010597e-01 -1.60584152e-02
5.07429600e-01 5.31509459e-01 -1.51049829e+00 1.70736223e-01
5.66658318e-01 1.17576146e+00 -1.96666196e-01 7.26244450e-01
2.27265507e-01 4.28313553e-01 -1.09981441e+00 -5.27822971e-01
4.34331596e-01 -3.43459189e-01 -6.03781402e-01 -4.31119740e-01
1.04704373e-01 1.09505979e-02 3.54926348e-01 1.22151124e+00
-5.61532900e-02 -3.19647193e-01 9.64225650e-01 -1.83876145e+00
-3.41930449e-01 5.50096750e-01 -5.45394838e-01 1.86277732e-01
3.83420527e-01 1.71946622e-02 1.02662480e+00 -2.35458851e-01
2.02251956e-01 1.09617233e+00 7.48853505e-01 5.66811487e-02
-1.27161026e+00 -1.12569714e+00 -1.04807347e-01 5.77602983e-02
-1.53214169e+00 3.98941517e-01 3.88707817e-01 -4.84276295e-01
6.71473026e-01 5.98932564e-01 4.53835428e-01 9.98488069e-01
5.22262096e-01 2.86077708e-01 1.52596354e+00 -5.73614717e-01
4.06797618e-01 9.21293676e-01 5.38990557e-01 3.20620984e-01
6.92292273e-01 3.35596502e-01 -3.72022949e-02 -5.24313986e-01
4.80135590e-01 -1.15170255e-01 -3.82027216e-02 -2.21424237e-01
-9.26674664e-01 1.03004718e+00 2.23297358e-01 3.96149814e-01
-7.43028745e-02 -4.51106466e-02 2.10993424e-01 7.03777432e-01
1.11831486e-01 6.06029868e-01 -7.07252324e-01 -1.81406513e-01
-5.25650024e-01 1.75322443e-01 1.12581980e+00 8.58980238e-01
5.87052941e-01 -2.70116299e-01 2.16881454e-01 5.31020761e-01
1.81274012e-01 5.58037937e-01 4.02508408e-01 -7.83634961e-01
4.85073715e-01 6.73769772e-01 2.16268092e-01 -9.87009943e-01
-5.38305938e-01 -4.85931128e-01 -5.01419902e-01 6.50428116e-01
6.23980045e-01 -2.66305774e-01 -5.39335012e-01 1.82558334e+00
1.73873156e-02 -8.57890472e-02 9.36103985e-02 6.18328929e-01
3.15960318e-01 5.55599630e-01 2.10393593e-01 -1.41696364e-01
1.34926093e+00 -3.33569616e-01 -3.17481279e-01 -1.85124073e-02
6.29815161e-01 -8.05904448e-01 1.15523171e+00 8.45204771e-01
-3.35622162e-01 -6.27031088e-01 -1.43224001e+00 7.93466628e-01
-7.03381181e-01 -2.30919439e-02 6.94258988e-01 1.45846272e+00
-7.31090724e-01 5.39317012e-01 -4.56539214e-01 -4.39637274e-01
-2.10444137e-01 4.88963664e-01 -3.78712326e-01 9.13544297e-02
-1.18241429e+00 1.22647285e+00 5.93386471e-01 -6.96107626e-01
-2.77395308e-01 -2.05845773e-01 -4.33816880e-01 2.02870339e-01
7.08004013e-02 -3.04946452e-01 8.11150432e-01 -9.70299780e-01
-1.08759463e+00 6.06082857e-01 1.50316775e-01 -2.88328677e-01
4.02038187e-01 3.65307815e-02 -6.84252441e-01 -1.56164363e-01
-6.18144013e-02 3.84572059e-01 3.79307181e-01 -1.24369192e+00
-9.67690527e-01 -2.55412072e-01 1.31675467e-01 -1.74169868e-01
-2.92290270e-01 1.39564455e-01 -5.96467638e-03 -5.72529435e-01
1.76705122e-01 -9.20548499e-01 6.06001914e-02 -5.65274000e-01
-2.08044574e-01 -2.52065957e-01 7.13647604e-01 -2.85144180e-01
1.48909378e+00 -2.22835135e+00 -1.28634781e-01 7.49929905e-01
-1.92549780e-01 9.71682668e-02 1.09229639e-01 3.93325090e-01
-5.87385476e-01 3.83748621e-01 -3.01407456e-01 1.14388727e-01
-2.85477638e-02 1.84381798e-01 -2.75658429e-01 5.09831548e-01
3.46995652e-01 2.98559904e-01 -5.59039950e-01 -3.84396106e-01
1.49176732e-01 9.62722767e-03 -2.84638792e-01 8.22688341e-02
3.09099317e-01 -3.51341814e-02 -2.72471368e-01 5.11738598e-01
6.85112834e-01 -7.26438910e-02 2.11469531e-01 1.60659477e-01
1.67472079e-01 -1.56630293e-01 -1.56673765e+00 6.10372365e-01
-1.25913590e-01 4.22471941e-01 -2.75153846e-01 -1.09635544e+00
1.28711808e+00 1.72129855e-01 2.95440525e-01 -1.65142819e-01
4.64807302e-01 2.77298063e-01 2.26990506e-01 -2.10313559e-01
2.21064791e-01 -2.47417539e-01 -5.30890942e-01 6.50387943e-01
1.03562839e-01 -3.27110618e-01 1.12963520e-01 8.36902559e-02
1.15668905e+00 -2.29479477e-01 5.12250543e-01 -3.15747470e-01
4.71506834e-01 -1.62972003e-01 5.87206244e-01 6.91486478e-01
-3.28072399e-01 3.40501159e-01 9.45632577e-01 -2.73093700e-01
-9.26489353e-01 -1.23057735e+00 -5.84318757e-01 7.99496293e-01
2.06333980e-01 -2.12121233e-01 -7.50549197e-01 -9.03682768e-01
1.75918698e-01 1.00054371e+00 -6.71231747e-01 -3.16065192e-01
-1.24742664e-01 -1.22095668e+00 6.09815538e-01 5.73220134e-01
2.38654003e-01 -7.14073956e-01 -5.90937316e-01 -5.21407835e-02
-1.30420998e-01 -7.54597664e-01 9.92181227e-02 7.31523693e-01
-8.08737695e-01 -1.48225331e+00 -9.39311013e-02 -4.57459331e-01
6.06339931e-01 6.19514771e-02 8.96090209e-01 2.37166137e-01
6.31879792e-02 2.40468219e-01 -5.59951186e-01 -7.18845129e-01
-6.26195908e-01 9.84084676e-04 2.74591982e-01 -3.03371757e-01
8.13770413e-01 -3.31317544e-01 -6.38117194e-02 7.53308594e-01
-6.21726871e-01 -5.98542213e-01 6.68118298e-01 9.23363030e-01
8.95477310e-02 3.60202402e-01 8.68407071e-01 -1.02477062e+00
7.52453446e-01 -6.02684677e-01 -6.75366879e-01 4.21009779e-01
-1.15396917e+00 -2.71773022e-02 5.49651086e-01 -5.55941045e-01
-5.99981308e-01 -2.36949503e-01 4.66733202e-02 -9.56925675e-02
-2.87914008e-01 3.04591238e-01 -8.46820325e-02 2.00767349e-03
1.02887678e+00 -2.41135016e-01 8.91359299e-02 -2.68665642e-01
-1.57389209e-01 9.01030242e-01 2.76798874e-01 -6.56002164e-01
8.75685990e-01 -7.93462396e-02 1.76659778e-01 -3.18510324e-01
-5.06393194e-01 -5.23215592e-01 -6.43637598e-01 -8.30460861e-02
4.39072967e-01 -4.32524920e-01 -9.48314130e-01 2.54863858e-01
-5.95863342e-01 -6.07304275e-02 -1.77342847e-01 8.67291689e-01
-3.97014558e-01 2.89416254e-01 -4.27516311e-01 -1.01558709e+00
5.56498058e-02 -1.25852406e+00 4.87294048e-01 4.26582187e-01
-4.78853345e-01 -8.68996918e-01 -1.15139760e-01 7.33005926e-02
1.28164321e-01 3.57904315e-01 1.00690830e+00 -1.10237086e+00
-8.12327638e-02 -8.57389271e-01 -1.99906379e-01 7.24960685e-01
-5.31031527e-02 2.96887189e-01 -1.02353084e+00 -4.77011889e-01
1.41002029e-01 -2.86946058e-01 4.55512941e-01 1.85845837e-01
1.00565600e+00 -1.93640310e-02 -3.37115884e-01 3.26969236e-01
1.32647502e+00 7.51767635e-01 5.98606825e-01 5.71287096e-01
-1.79414392e-01 6.08458221e-01 9.90309060e-01 2.80511588e-01
1.57987654e-01 5.02204657e-01 4.55207914e-01 -1.81919515e-01
3.70989799e-01 -5.41021265e-02 3.47423464e-01 5.91781437e-01
3.72078903e-02 -2.16034979e-01 -7.99740553e-01 4.63959724e-02
-1.57122922e+00 -9.00819540e-01 -2.84478962e-01 2.63716197e+00
5.86046934e-01 6.60377383e-01 2.52163112e-01 6.93999231e-01
8.46019387e-01 -3.96067351e-01 -2.34378457e-01 -6.67997956e-01
6.89721778e-02 4.32369739e-01 6.67421699e-01 4.54854369e-01
-1.11536908e+00 3.81373286e-01 7.28464317e+00 7.25603342e-01
-1.02250481e+00 1.67157091e-02 8.59179854e-01 2.86928236e-01
4.60242741e-02 3.31951320e-01 -9.55208480e-01 6.78216696e-01
1.29071248e+00 -3.25330079e-01 1.05717003e-01 1.08147097e+00
-3.74370933e-01 -5.59026659e-01 -1.03511059e+00 6.74104452e-01
1.24421552e-01 -7.38678336e-01 -3.55333835e-01 2.70911455e-01
4.24718320e-01 -2.17908442e-01 1.05851434e-01 5.20926178e-01
6.15569711e-01 -8.88384581e-01 8.21708798e-01 3.24534297e-01
7.52270281e-01 -9.40652549e-01 1.47431886e+00 4.55897003e-01
-8.15630317e-01 -4.53992784e-01 -3.60329866e-01 -8.90051648e-02
-2.63624847e-01 5.11772156e-01 -1.07297134e+00 5.69562614e-01
8.01776290e-01 9.79543775e-02 -8.76128376e-01 1.06741369e+00
-2.24258617e-01 8.35720122e-01 -4.03922528e-01 -2.91182160e-01
-8.31269175e-02 -2.09428713e-01 1.61493421e-01 1.30689001e+00
3.85624439e-01 1.18807472e-01 7.58929625e-02 5.80334961e-01
5.43909073e-01 -1.04352362e-01 -4.05270100e-01 3.60662401e-01
8.46395791e-01 1.04748785e+00 -9.00902629e-01 -3.57079595e-01
-3.00768524e-01 1.51795328e-01 -1.07639290e-01 2.08324030e-01
-9.23577845e-01 -8.01770806e-01 4.20787662e-01 -8.60095024e-02
3.44712101e-02 -3.69377211e-02 -9.12424564e-01 -6.90005779e-01
-1.73638836e-01 -7.63074815e-01 7.81479895e-01 -5.69409788e-01
-1.33601630e+00 4.85483527e-01 3.57527316e-01 -1.53599834e+00
-3.65497231e-01 -9.75691795e-01 -4.19390738e-01 1.08024347e+00
-9.67285693e-01 -5.59694767e-01 -4.64935869e-01 3.83568615e-01
-1.07292682e-01 -3.45409155e-01 1.08481944e+00 1.36454301e-02
-4.50403035e-01 9.35877860e-01 1.12365028e-02 1.72756851e-01
9.37861681e-01 -1.32967019e+00 8.58751312e-02 7.26358771e-01
-1.17168039e-01 8.12075019e-01 7.51165152e-01 -5.21369576e-01
-7.13049173e-01 -5.15274465e-01 5.66561103e-01 -8.30560923e-01
6.10505879e-01 -8.18372741e-02 -7.55801082e-01 5.53795338e-01
-2.01926142e-01 -3.80691141e-01 1.02976739e+00 3.99063379e-01
-5.38990915e-01 -3.60651076e-01 -1.59433591e+00 1.55477434e-01
4.11113888e-01 -2.65502930e-01 -7.14226127e-01 5.82461879e-02
4.49835509e-01 -9.20085460e-02 -1.27099025e+00 4.48038638e-01
7.28226066e-01 -1.22338426e+00 8.02188098e-01 -4.72445488e-01
4.90104370e-02 -3.04793954e-01 -3.88409555e-01 -1.25657260e+00
-4.44804728e-01 -1.48481485e-02 3.68582457e-01 1.24649787e+00
6.65203333e-01 -1.22153330e+00 5.83088458e-01 7.94437468e-01
1.70195222e-01 -6.33192718e-01 -1.02090871e+00 -1.00822699e+00
2.32967064e-01 -5.91382742e-01 8.94302189e-01 1.05836880e+00
2.90450633e-01 1.01718962e-01 4.38363627e-02 1.80105940e-01
5.83313167e-01 -1.51992664e-01 9.45962965e-01 -1.48644328e+00
-4.97989714e-01 -5.38315833e-01 -8.38088632e-01 -3.17040771e-01
-2.93484300e-01 -6.19067132e-01 -2.63202131e-01 -8.36190701e-01
2.96289384e-01 -8.13056111e-01 -6.17877722e-01 5.83480656e-01
-1.97363064e-01 1.29208580e-01 1.91928297e-01 1.76337183e-01
1.13878153e-01 -2.81356364e-01 7.07820356e-01 -1.08966744e-02
3.59751564e-03 3.59041750e-01 -8.86888385e-01 6.84279382e-01
8.46350133e-01 -8.00385058e-01 -3.80961031e-01 3.20097357e-01
5.59473522e-02 1.80736795e-01 2.14712679e-01 -1.18523526e+00
6.52252212e-02 -4.06584322e-01 7.00856209e-01 -3.78935158e-01
1.74455360e-01 -1.04624879e+00 4.50014561e-01 8.20538580e-01
-1.38805702e-01 3.01318944e-01 2.94006951e-02 4.29974347e-01
-2.22771868e-01 -8.09240580e-01 1.10007465e+00 1.33366242e-01
-2.93762684e-01 -3.89687210e-01 -1.87781230e-01 -2.23786175e-01
1.36280763e+00 -5.66637039e-01 -4.84549999e-01 -3.44486296e-01
-5.43839931e-01 -1.03230208e-01 6.07950687e-01 4.56090391e-01
1.96519554e-01 -1.02523434e+00 -4.84865904e-01 3.90056849e-01
2.19873637e-01 -5.75106919e-01 -4.00232822e-01 5.95861197e-01
-5.26804090e-01 4.21269208e-01 -3.34978789e-01 -7.78242409e-01
-1.32682669e+00 4.81834561e-01 1.10414624e-01 -2.31051832e-01
-2.47860715e-01 7.27942765e-01 -2.03438550e-02 -4.05304909e-01
3.12429845e-01 -4.56701249e-01 -2.32889995e-01 5.71637154e-02
5.32319605e-01 4.41565067e-01 1.23861358e-01 -4.13607359e-01
-3.45559955e-01 5.56156516e-01 1.86841134e-02 -2.15777650e-01
1.15467763e+00 3.14750820e-01 7.08595812e-02 5.32267094e-01
9.74940896e-01 -7.88158476e-02 -8.00373614e-01 2.49995172e-01
2.47734517e-01 -8.06919873e-01 -3.67663264e-01 -1.13600969e+00
-7.30464280e-01 5.11045635e-01 7.99637973e-01 6.29637063e-01
1.11749887e+00 -2.35317513e-01 5.25793806e-02 3.43038112e-01
9.07782078e-01 -1.03401124e+00 -8.49160459e-03 3.07491541e-01
6.03204906e-01 -1.05928230e+00 1.20174959e-01 -3.55777353e-01
-8.59468639e-01 1.19044518e+00 5.99302292e-01 -1.06874347e-01
7.40422070e-01 3.42392921e-01 1.07631542e-01 6.24254793e-02
-6.19573832e-01 4.22532827e-01 -1.08338119e-02 6.78689957e-01
2.78516144e-01 5.03867090e-01 -6.92779779e-01 7.74548650e-01
-5.15704453e-01 -8.46715420e-02 7.16029763e-01 7.69330084e-01
-5.17431736e-01 -1.26950228e+00 -8.30513656e-01 7.11669147e-01
-2.76316553e-01 2.62282521e-01 -3.97268832e-01 1.05753469e+00
2.48979121e-01 1.22574294e+00 -2.06829589e-02 -8.97347331e-01
5.40738761e-01 1.52841315e-01 2.22311704e-03 -4.46857035e-01
-7.10672855e-01 -3.56850505e-01 2.59005409e-02 -3.54358912e-01
-7.21986592e-02 -9.02089596e-01 -9.19789612e-01 -5.52810490e-01
-7.61668742e-01 5.49681544e-01 9.06995595e-01 5.95004261e-01
-8.87925327e-02 2.74706244e-01 9.28565145e-01 -2.78367728e-01
-1.02132273e+00 -1.21180582e+00 -8.36082757e-01 3.71944010e-01
-2.33972386e-01 -9.61227953e-01 -8.87225389e-01 -2.58459955e-01] | [8.343284606933594, 4.284278869628906] |
6a0edf64-b5ca-4876-97dd-3383879b1cc2 | cova-context-aware-visual-attention-for | 2110.12320 | null | https://arxiv.org/abs/2110.12320v1 | https://arxiv.org/pdf/2110.12320v1.pdf | CoVA: Context-aware Visual Attention for Webpage Information Extraction | Webpage information extraction (WIE) is an important step to create knowledge bases. For this, classical WIE methods leverage the Document Object Model (DOM) tree of a website. However, use of the DOM tree poses significant challenges as context and appearance are encoded in an abstract manner. To address this challenge we propose to reformulate WIE as a context-aware Webpage Object Detection task. Specifically, we develop a Context-aware Visual Attention-based (CoVA) detection pipeline which combines appearance features with syntactical structure from the DOM tree. To study the approach we collect a new large-scale dataset of e-commerce websites for which we manually annotate every web element with four labels: product price, product title, product image and background. On this dataset we show that the proposed CoVA approach is a new challenging baseline which improves upon prior state-of-the-art methods. | ['Alexander Schwing', 'Kevin Chen-Chuan Chang', 'Jingjin Wang', 'Keval Morabia', 'Anurendra Kumar'] | 2021-10-24 | null | https://aclanthology.org/2022.ecnlp-1.11 | https://aclanthology.org/2022.ecnlp-1.11.pdf | ecnlp-acl-2022-5 | ['webpage-object-detection'] | ['computer-vision'] | [ 3.98822993e-01 -1.39337555e-01 -1.96388662e-01 -2.74760872e-01
-8.79707992e-01 -1.16270018e+00 7.32756555e-01 3.14201623e-01
-7.03407601e-02 -5.60852420e-03 -4.63770032e-02 -3.11021507e-01
3.30403358e-01 -5.97762823e-01 -1.01984775e+00 -3.36821526e-01
2.30764002e-01 -3.23100225e-03 3.59227598e-01 2.03992024e-01
2.32133821e-01 2.40368307e-01 -1.47723055e+00 7.30688155e-01
6.58871353e-01 1.35257518e+00 2.05392435e-01 4.49435979e-01
-5.39936960e-01 6.09456182e-01 -3.67174029e-01 -9.96347129e-01
4.42375183e-01 -3.84577215e-02 -7.94206262e-01 5.37727714e-01
1.00568652e+00 -3.87598068e-01 -1.44266961e-02 1.29782224e+00
-6.72308803e-02 -2.65166551e-01 7.12544441e-01 -1.22925687e+00
-8.70553136e-01 4.57173228e-01 -8.44660282e-01 3.70519906e-02
1.85850307e-01 1.92446366e-01 1.37105691e+00 -1.19011199e+00
9.87091243e-01 1.02658951e+00 6.43605649e-01 7.06672147e-02
-1.42350161e+00 -1.13761604e-01 7.02249527e-01 3.00574154e-01
-1.10879505e+00 -1.71381742e-01 9.77448702e-01 -5.33077180e-01
6.71785891e-01 2.63418168e-01 5.03825784e-01 1.33427680e+00
-2.37985879e-01 1.40541291e+00 1.25836349e+00 -7.76902199e-01
2.55021065e-01 3.94968122e-01 5.47040820e-01 8.35140049e-01
4.58983362e-01 -4.29709852e-01 -3.18445414e-01 5.97843155e-02
5.83574057e-01 7.06038252e-02 6.05416521e-02 -9.12872732e-01
-6.85556054e-01 5.80418825e-01 4.77894932e-01 -8.84667486e-02
-4.20580238e-01 1.41939878e-01 3.18048120e-01 -1.63227558e-01
1.65803120e-01 4.82615829e-02 -5.96783280e-01 1.94807604e-01
-5.88695765e-01 1.55239269e-01 8.69651139e-01 1.09054899e+00
6.74217820e-01 -3.22219938e-01 -2.00919107e-01 8.73950183e-01
5.65815032e-01 2.44897127e-01 2.45545302e-02 -7.55596220e-01
7.41937757e-01 9.19859290e-01 2.08610877e-01 -8.93231511e-01
-8.16094652e-02 -2.30098933e-01 -9.31720510e-02 4.68390346e-01
6.03666961e-01 2.70029992e-01 -1.08653915e+00 1.28453970e+00
7.11542249e-01 -4.27288830e-01 -3.16132724e-01 8.62129211e-01
6.88936412e-01 4.82804358e-01 4.59057629e-01 1.48132473e-01
2.08754516e+00 -1.38733065e+00 -7.40976274e-01 -3.65592659e-01
2.93982178e-01 -9.34021950e-01 1.25754273e+00 5.16451597e-01
-8.31377506e-01 -4.20952260e-01 -1.27676904e+00 -3.61406118e-01
-9.40174341e-01 2.87359208e-01 6.92625940e-01 6.33507013e-01
-4.80611414e-01 8.93727690e-02 -6.21110320e-01 -5.42032123e-01
5.54822743e-01 -2.70881295e-01 -2.69994617e-01 -2.84085751e-01
-6.32132828e-01 6.48490667e-01 2.05802366e-01 2.80454047e-02
-6.65279627e-01 -4.26985264e-01 -9.81445849e-01 1.76191628e-01
9.23751652e-01 -5.57328939e-01 1.43189442e+00 -1.21745658e+00
-1.12752342e+00 1.05053294e+00 7.38464743e-02 -2.67167717e-01
4.04641181e-01 -4.41149235e-01 -3.19026530e-01 2.71020502e-01
-5.28616793e-02 1.82780251e-01 1.13853002e+00 -1.85597479e+00
-8.43655884e-01 -6.88691258e-01 2.75542110e-01 -8.63492414e-02
-3.85580450e-01 1.40035778e-01 -1.14146793e+00 -8.52826774e-01
-1.76231250e-01 -7.46917188e-01 1.13338768e-01 2.54839838e-01
-6.28455818e-01 -1.98654026e-01 9.33589101e-01 -9.68122184e-01
1.25039339e+00 -1.99715853e+00 -3.72668147e-01 2.00829148e-01
3.56325358e-01 2.86616981e-01 -2.32849628e-01 3.36433113e-01
6.10574931e-02 2.59664148e-01 1.96979940e-02 -2.23186091e-01
4.20875072e-01 -1.27668321e-01 -1.06306985e-01 2.38153408e-03
3.09439510e-01 1.04359758e+00 -6.39047265e-01 -7.95071244e-01
9.95823741e-02 4.12157565e-01 -4.54642057e-01 9.19553190e-02
-6.14418983e-01 -4.14957464e-01 -5.46572804e-01 1.18511975e+00
9.04654264e-01 -5.46314418e-01 4.66637850e-01 -6.49237990e-01
-1.13409750e-01 1.28476053e-01 -1.11728036e+00 1.66343653e+00
-4.61801529e-01 5.68980634e-01 3.98207426e-01 -6.28928125e-01
4.80312675e-01 -1.00808062e-01 4.08736080e-01 -7.94873059e-01
1.03752367e-01 -1.72593519e-01 -4.67874020e-01 -7.15541422e-01
3.14442337e-01 5.59629202e-01 4.04172540e-02 2.14507610e-01
-1.86088684e-04 5.28783441e-01 4.05484617e-01 4.37144816e-01
1.21998012e+00 7.08268285e-01 2.15420574e-01 9.16406885e-03
2.24699289e-01 2.38870516e-01 4.48177308e-01 8.60288143e-01
-4.75189596e-01 7.10724115e-01 6.09692037e-01 -5.30781746e-01
-1.23002636e+00 -1.19449091e+00 -1.60253309e-02 1.31505287e+00
5.79752065e-02 -7.19304204e-01 -1.14221597e+00 -1.54785800e+00
1.64594904e-01 4.85955179e-01 -7.69176900e-01 5.00615239e-01
-5.59152126e-01 -4.67339635e-01 -1.45428210e-01 6.52384162e-01
5.15830874e-01 -7.68052936e-01 -3.80649775e-01 1.05747342e-01
-1.02965146e-01 -1.65718079e+00 -7.42304802e-01 1.88144773e-01
-4.10773546e-01 -1.32122278e+00 -4.71473366e-01 -9.44460869e-01
6.70114040e-01 3.55362624e-01 1.37229860e+00 -2.17022151e-01
-7.37869620e-01 6.82746530e-01 -4.61986274e-01 -4.31592554e-01
-1.52011469e-01 -1.08899988e-01 -5.69752574e-01 3.57531190e-01
6.29146874e-01 -1.86243996e-01 -7.71414340e-01 6.75412044e-02
-1.06723225e+00 1.94550768e-01 8.71413171e-01 4.41807270e-01
6.70741081e-01 -4.24844539e-03 -4.57939059e-02 -1.23122454e+00
2.85850823e-01 -2.84618586e-01 -9.42011535e-01 6.91892326e-01
-6.85152590e-01 9.06397998e-02 4.47556257e-01 -2.65122145e-01
-1.25514972e+00 6.24291837e-01 2.52867162e-01 -3.11413705e-01
-3.51296872e-01 5.47637224e-01 -4.98310298e-01 1.48836404e-01
3.19930047e-01 7.61243999e-02 -4.22549933e-01 -1.03550875e+00
6.58319533e-01 8.20811808e-01 5.92684150e-01 -5.44860303e-01
7.83018768e-01 5.36045313e-01 -1.90535530e-01 -5.54086924e-01
-1.02999878e+00 -6.99034691e-01 -8.51846337e-01 -2.30412140e-01
8.61101389e-01 -7.60551453e-01 -6.86064124e-01 1.51537225e-01
-1.23758054e+00 -4.96453196e-02 1.76665872e-01 -2.62298100e-02
-2.35866651e-01 6.72245383e-01 -5.64896643e-01 -8.80572855e-01
-3.71105105e-01 -9.35354590e-01 1.22700691e+00 9.62153152e-02
1.97080404e-01 -6.17589772e-01 -1.82862192e-01 7.74308801e-01
8.98539349e-02 4.53307420e-01 1.05732691e+00 -6.17905021e-01
-9.43387866e-01 -2.65586883e-01 -7.48128235e-01 2.06726685e-01
-1.28099760e-02 2.38444820e-01 -1.03852057e+00 1.45616129e-01
-3.56930196e-01 -1.72104724e-02 1.05796623e+00 -7.08388984e-02
1.11336625e+00 -5.93951166e-01 -4.04547036e-01 4.57351178e-01
1.91217244e+00 2.20829733e-02 3.99214596e-01 8.03077042e-01
8.78108203e-01 7.23322868e-01 5.24352491e-01 4.31572735e-01
5.63450098e-01 9.59065497e-01 7.15204000e-01 -6.29107058e-02
-6.18431747e-01 -4.07142818e-01 4.88696456e-01 4.47589844e-01
1.91730633e-01 -2.24211201e-01 -8.40831161e-01 5.45083702e-01
-1.94035411e+00 -6.77176893e-01 -3.05342883e-01 1.87761533e+00
7.80077100e-01 2.35477403e-01 2.35263869e-01 -2.29748785e-01
7.04818666e-01 1.40689481e-02 -3.48288178e-01 -2.53722250e-01
7.15856403e-02 -2.44678736e-01 5.09213448e-01 7.82385543e-02
-1.70909727e+00 8.10907662e-01 6.13865709e+00 6.26060069e-01
-6.97986960e-01 9.27521810e-02 3.58677745e-01 2.00810596e-01
-2.13745818e-01 -6.91218451e-02 -1.06278718e+00 5.64767599e-01
5.53403020e-01 5.11919677e-01 3.80998641e-01 1.48353267e+00
-1.25040784e-01 -4.88159023e-02 -1.22954106e+00 9.86098886e-01
3.59541655e-01 -1.16826737e+00 1.73882440e-01 1.38842866e-01
6.80048048e-01 -2.56125510e-01 1.72376767e-01 3.15476090e-01
5.08535862e-01 -5.62049866e-01 8.37600172e-01 2.14256242e-01
5.26550114e-01 -2.54736334e-01 4.00643170e-01 -2.28677154e-01
-1.35342848e+00 -1.34771138e-01 3.65486398e-04 4.79462773e-01
-6.43197894e-02 5.20056248e-01 -7.28018165e-01 3.45147192e-01
8.88549626e-01 5.34595668e-01 -1.12338257e+00 1.10914445e+00
-2.40331888e-01 5.52979350e-01 -2.60644913e-01 1.60501674e-02
2.75959641e-01 -3.24179456e-02 3.44327867e-01 1.50299311e+00
-1.60813048e-01 -3.30023229e-01 2.57028610e-01 1.01677763e+00
-3.92380178e-01 2.86704332e-01 -4.14124817e-01 -2.94223905e-01
7.17514232e-02 1.76518857e+00 -9.53980505e-01 -2.56932288e-01
-1.16604209e+00 1.28730297e+00 5.24084568e-01 5.77494085e-01
-8.47749412e-01 -4.49105412e-01 5.41715980e-01 2.02253595e-01
1.04800987e+00 -1.11565247e-01 -3.30433786e-01 -1.33235919e+00
4.72773671e-01 -7.61548638e-01 5.21236956e-01 -1.13932514e+00
-1.38571501e+00 2.40893602e-01 -2.54316002e-01 -1.09823608e+00
8.38722512e-02 -1.12768388e+00 -2.74883002e-01 3.95869941e-01
-1.85347593e+00 -1.68448997e+00 -5.33967018e-01 3.50503176e-01
9.78798449e-01 1.66250929e-01 5.27070761e-01 2.56939024e-01
-8.71427417e-01 4.65004385e-01 3.84951711e-01 6.10787034e-01
6.46258235e-01 -1.54816234e+00 4.08319443e-01 9.65432405e-01
4.92814720e-01 7.59924293e-01 4.64943528e-01 -6.95236981e-01
-1.85653400e+00 -1.08743107e+00 6.17492437e-01 -9.71399069e-01
7.79516280e-01 -7.34743059e-01 -7.18050838e-01 7.40853727e-01
4.05522823e-01 1.71630800e-01 5.28883696e-01 2.22635865e-01
-1.06063998e+00 -1.41745716e-01 -8.70743394e-01 6.71689749e-01
9.11111772e-01 -4.16594088e-01 -5.66947997e-01 2.58910865e-01
5.54390728e-01 -4.61947955e-02 -7.69093752e-01 1.06369339e-01
7.94285059e-01 -6.36196077e-01 1.23188066e+00 -6.25539362e-01
5.13951302e-01 -3.18509966e-01 -1.38886258e-01 -7.49604285e-01
-2.91054308e-01 -4.02457595e-01 -5.68452120e-01 1.62188268e+00
3.79946023e-01 1.09575719e-01 7.96449900e-01 8.62945378e-01
5.70735075e-02 -4.05948639e-01 -3.74003679e-01 -7.98932195e-01
-3.63656193e-01 -4.05846179e-01 3.11922550e-01 7.40383565e-01
7.63593465e-02 6.08797133e-01 -7.51598999e-02 2.11426631e-01
9.68766630e-01 4.52394813e-01 6.62095189e-01 -1.23872566e+00
-2.79348195e-01 -5.09706438e-01 -1.62920967e-01 -9.41851139e-01
-3.36981833e-01 -7.59601831e-01 -9.54536442e-03 -1.75244188e+00
7.47459769e-01 -6.18356504e-02 -4.19086128e-01 5.48009574e-01
-2.07630545e-01 3.15605432e-01 4.54710722e-01 2.69200563e-01
-1.18191838e+00 -2.05161739e-02 9.02889311e-01 -4.44242358e-01
2.03902900e-01 -3.65105957e-01 -5.55844188e-01 8.21497262e-01
4.85695720e-01 -2.94840783e-01 4.71442901e-02 -2.74808466e-01
3.57873201e-01 -4.91123885e-01 4.45358396e-01 -5.09282172e-01
9.48187262e-02 -1.31604284e-01 7.16472626e-01 -7.26309896e-01
9.90260690e-02 -1.18590117e+00 -3.77908200e-01 3.36530432e-02
-3.55028033e-01 -9.16230977e-02 8.02641138e-02 9.70552742e-01
4.58525568e-02 -4.40165699e-01 4.48911279e-01 -3.49401563e-01
-1.10658026e+00 2.54738443e-02 -3.38198423e-01 -4.51463126e-02
9.59427059e-01 -7.61500597e-02 -5.51097333e-01 2.11089607e-02
-6.63774431e-01 -1.02919815e-02 4.61502612e-01 6.88962102e-01
3.36835444e-01 -1.28310823e+00 -3.59744489e-01 3.10711879e-02
6.20823920e-01 -3.69552433e-01 1.67518705e-02 5.33505440e-01
-5.42508483e-01 4.65824157e-01 2.56087091e-02 -3.62528831e-01
-1.37028384e+00 1.28370273e+00 -2.29271322e-01 -3.16186577e-01
-6.75559521e-01 4.32770789e-01 5.00816405e-01 -4.42957729e-02
5.38168669e-01 -2.87315875e-01 -3.35316062e-01 3.97766307e-02
6.11670613e-01 6.06198274e-02 2.55210727e-01 -4.36175317e-01
-2.45926708e-01 5.73785126e-01 -4.17740136e-01 -6.92460313e-02
1.20654285e+00 -4.88946319e-01 2.57097960e-01 9.12384465e-02
1.25215244e+00 1.19582728e-01 -1.52923858e+00 -4.27509904e-01
3.74542236e-01 -6.29784405e-01 1.73616827e-01 -1.33334744e+00
-9.08368587e-01 7.03752756e-01 7.12160766e-01 5.04437029e-01
8.70600283e-01 1.22690946e-01 6.53717160e-01 2.79642999e-01
1.00690171e-01 -1.58271682e+00 3.48924428e-01 1.60451144e-01
8.54429603e-01 -1.63333201e+00 4.59051393e-02 -8.84398878e-01
-7.15861201e-01 1.22352743e+00 7.43963778e-01 2.06143692e-01
5.61394274e-01 2.78336197e-01 1.52905971e-01 -2.99408495e-01
-5.46729505e-01 -7.28660345e-01 6.24242663e-01 6.83169782e-01
3.05850089e-01 -2.65260100e-01 -1.77651823e-01 9.62615132e-01
4.94953692e-01 -2.27758199e-01 2.45940402e-01 1.14930904e+00
-2.72250146e-01 -1.14174759e+00 -4.15668011e-01 1.56188414e-01
-7.03051984e-01 -2.58067608e-01 -6.00604594e-01 8.12089026e-01
-9.51338038e-02 7.33060122e-01 -9.54553410e-02 -5.03517203e-02
2.02755138e-01 4.52193141e-01 5.12559474e-01 -4.84766543e-01
-5.30887544e-01 4.67794538e-01 2.03435630e-01 -7.20170021e-01
-1.72789261e-01 -8.27013671e-01 -9.06084716e-01 3.22000921e-01
-2.07974494e-01 -3.44634742e-01 1.23694766e+00 6.01622820e-01
5.36135435e-01 4.39958274e-01 3.59542847e-01 -5.44545591e-01
-4.72124130e-01 -6.44554675e-01 -3.96963507e-01 1.09393072e+00
2.09097758e-01 -5.73029041e-01 -5.13201766e-02 7.41502225e-01] | [11.510533332824707, 2.3810975551605225] |
36825125-ccd4-4289-8e4d-39afb9ce82ad | klmo-knowledge-graph-enhanced-pretrained | null | null | https://aclanthology.org/2021.findings-emnlp.384 | https://aclanthology.org/2021.findings-emnlp.384.pdf | KLMo: Knowledge Graph Enhanced Pretrained Language Model with Fine-Grained Relationships | Interactions between entities in knowledge graph (KG) provide rich knowledge for language representation learning. However, existing knowledge-enhanced pretrained language models (PLMs) only focus on entity information and ignore the fine-grained relationships between entities. In this work, we propose to incorporate KG (including both entities and relations) into the language learning process to obtain KG-enhanced pretrained Language Model, namely KLMo. Specifically, a novel knowledge aggregator is designed to explicitly model the interaction between entity spans in text and all entities and relations in a contextual KG. An relation prediction objective is utilized to incorporate relation information by distant supervision. An entity linking objective is further utilized to link entity spans in text to entities in KG. In this way, the structured knowledge can be effectively integrated into language representations. Experimental results demonstrate that KLMo achieves great improvements on several knowledge-driven tasks, such as entity typing and relation classification, comparing with the state-of-the-art knowledge-enhanced PLMs. | ['Feng Zhang', 'Tao Yang', 'Suncong Zheng', 'Lei He'] | null | null | null | null | findings-emnlp-2021-11 | ['relation-classification'] | ['natural-language-processing'] | [-2.51385897e-01 3.90507430e-01 -7.28083014e-01 -4.79916394e-01
-2.79653639e-01 -3.51134658e-01 4.86192286e-01 6.41649961e-01
-4.16208416e-01 9.14881289e-01 3.49257112e-01 -2.36627504e-01
-1.56734660e-01 -1.34771276e+00 -8.15831363e-01 -2.74383456e-01
-2.07716405e-01 4.49600667e-01 2.49202400e-01 -4.06179965e-01
-4.23845708e-01 1.59434706e-01 -9.14336920e-01 2.33648598e-01
1.27271032e+00 8.65186214e-01 6.07506670e-02 -1.28459083e-02
-7.32469022e-01 1.41138351e+00 -1.72643393e-01 -7.61768579e-01
-3.32946241e-01 4.68982682e-02 -1.00433004e+00 -2.67785221e-01
9.17790160e-02 7.73400143e-02 -5.80920815e-01 8.21010232e-01
2.27186128e-01 2.44444504e-01 6.36283159e-01 -1.01963055e+00
-1.13452518e+00 1.35768890e+00 -5.66786051e-01 7.18020946e-02
3.67941499e-01 -4.66465771e-01 1.47687626e+00 -9.29917932e-01
7.18471050e-01 1.33960903e+00 4.61893618e-01 7.14359209e-02
-1.02622724e+00 -6.09552324e-01 5.97087860e-01 4.59042668e-01
-1.72331357e+00 -6.37556091e-02 7.33482838e-01 -3.77731234e-01
1.39071870e+00 -1.55648321e-01 3.54491055e-01 5.27713120e-01
-2.12373003e-01 9.31119323e-01 4.44224209e-01 -6.22834563e-01
-4.01432067e-01 4.52344239e-01 5.53869009e-01 1.01208401e+00
6.64133370e-01 -3.69470954e-01 -6.13905072e-01 -3.07899881e-02
6.27853990e-01 -1.06752872e-01 -1.87798768e-01 -4.45018619e-01
-9.22683239e-01 6.14436507e-01 8.01510692e-01 4.16934818e-01
-3.70995343e-01 -2.95703653e-02 4.57206607e-01 8.75802040e-02
5.56447089e-01 2.76976973e-01 -8.40945303e-01 3.92039418e-01
-2.27979705e-01 -4.11125273e-02 1.02649105e+00 1.33816648e+00
1.28493261e+00 -3.12336311e-02 -1.59015134e-01 9.93477345e-01
5.99233449e-01 4.19895858e-01 3.55455488e-01 -9.50748622e-02
9.40004408e-01 1.36084592e+00 -1.85166135e-01 -1.13298476e+00
-4.48720127e-01 -7.91973710e-01 -6.65701449e-01 -7.51756132e-01
-2.57463723e-01 -2.58100390e-01 -7.62882710e-01 1.64789462e+00
5.22793829e-01 3.73105705e-01 5.70706069e-01 4.87412810e-01
1.46136546e+00 5.22943854e-01 6.51277900e-01 -1.52166963e-01
1.48937547e+00 -1.11553228e+00 -8.71412277e-01 -2.84687340e-01
1.28979218e+00 -3.29083264e-01 8.53772581e-01 -3.58668774e-01
-6.09163821e-01 -4.71022338e-01 -8.70859504e-01 -3.00775111e-01
-8.72983158e-01 3.66385192e-01 9.10269856e-01 3.05464715e-01
-3.99902523e-01 1.90375388e-01 -6.92158222e-01 -2.42185399e-01
3.79400492e-01 3.32078397e-01 -5.14866352e-01 -1.75398871e-01
-2.00093818e+00 1.11545098e+00 1.21317089e+00 1.21261366e-01
-2.78190643e-01 -7.71935403e-01 -1.53455472e+00 1.75901532e-01
8.59760880e-01 -9.58364546e-01 7.81695127e-01 -3.62682492e-01
-1.01179659e+00 7.64567316e-01 -3.54034126e-01 -5.24091423e-01
-1.33906722e-01 -4.27594453e-01 -9.36510026e-01 -2.89761573e-01
1.19595423e-01 2.06612468e-01 1.16008110e-02 -1.30551791e+00
-7.63349593e-01 -1.88536227e-01 5.00175774e-01 5.47083676e-01
-4.91484523e-01 -8.29320326e-02 -7.77628541e-01 -6.73069775e-01
-1.11471079e-01 -5.61439872e-01 2.82084029e-02 -7.27520049e-01
-7.50917077e-01 -6.38257682e-01 5.19076705e-01 -7.91565120e-01
1.75656796e+00 -1.64120913e+00 1.23593718e-01 3.29755515e-01
2.45418966e-01 4.03894991e-01 4.49065194e-02 6.51842475e-01
-4.31641713e-02 2.27261201e-01 3.64978127e-02 -2.63373464e-01
1.75650597e-01 4.94489521e-01 -3.29900622e-01 -2.54698336e-01
3.03017169e-01 1.47764432e+00 -9.94263709e-01 -8.10887158e-01
-9.78847519e-02 4.29964185e-01 -2.10550800e-01 1.70537889e-01
-3.73381853e-01 5.64811416e-02 -8.34283590e-01 6.92442656e-01
4.79516387e-01 -5.74048281e-01 5.07283330e-01 -5.30114770e-01
2.70800591e-01 6.26759946e-01 -1.03995967e+00 1.44981623e+00
-7.56428301e-01 2.50405595e-02 -4.73293841e-01 -8.48553896e-01
9.01392221e-01 3.06128263e-01 2.08689913e-01 -3.14473540e-01
-2.11478308e-01 9.90505889e-02 -5.66310845e-02 -5.52398205e-01
5.49513519e-01 -3.75415720e-02 -4.62356359e-02 1.49389595e-01
3.58021617e-01 4.54872847e-01 4.02492553e-01 5.72742105e-01
8.37413788e-01 2.50119448e-01 6.20880723e-01 1.25719547e-01
9.09973323e-01 -1.42805561e-01 8.02229464e-01 3.77981961e-01
3.74574125e-01 -4.15115476e-01 3.25660408e-01 -6.75862581e-02
-3.66186589e-01 -1.00576437e+00 3.98068782e-03 1.32851505e+00
2.47968897e-01 -9.09096777e-01 -1.74223855e-01 -1.14464009e+00
4.25085604e-01 6.13811672e-01 -4.87879068e-01 -3.88355523e-01
-5.71631610e-01 -6.62338257e-01 5.38517773e-01 9.88696754e-01
4.89136040e-01 -9.84695971e-01 5.56924641e-01 3.48456502e-01
-1.19314231e-01 -1.43271708e+00 -2.44240969e-01 1.48567215e-01
-5.32580137e-01 -1.06020832e+00 -1.52977526e-01 -1.07116854e+00
7.37351656e-01 -7.22768977e-02 1.20476484e+00 -1.72488950e-02
1.30825311e-01 2.39908740e-01 -5.89267433e-01 -2.60233968e-01
3.05142328e-02 5.10412276e-01 1.25741333e-01 -1.58529356e-02
7.44046271e-01 -5.12032211e-01 -1.15506306e-01 1.60139084e-01
-7.29480743e-01 9.87967178e-02 8.04500341e-01 8.18141878e-01
6.39501095e-01 4.07115698e-01 8.93208325e-01 -1.57907224e+00
3.67870480e-01 -7.54343271e-01 -1.76521018e-01 9.03581679e-01
-6.87456310e-01 3.92087489e-01 6.09633803e-01 -4.02225047e-01
-1.46698999e+00 -2.18180791e-01 1.51234791e-02 -8.02637264e-02
1.72283858e-01 1.56329930e+00 -7.41043866e-01 -1.06272154e-01
2.96662331e-01 1.99960411e-01 -9.14005458e-01 -6.22411549e-01
7.01367855e-01 3.76323104e-01 3.62539679e-01 -9.06645954e-01
9.60709333e-01 1.17702544e-01 -1.81704372e-01 -5.01249492e-01
-1.42251062e+00 -6.81552410e-01 -8.52018297e-01 3.59370500e-01
4.94923949e-01 -1.36028779e+00 -4.24909353e-01 1.75327972e-01
-9.98783290e-01 -2.71430351e-02 -2.12014049e-01 6.84683681e-01
-8.63292217e-02 2.93093920e-01 -6.18311942e-01 -6.01937056e-01
-2.94073641e-01 -3.40432048e-01 7.35056639e-01 4.34953839e-01
1.70081675e-01 -1.58850193e+00 7.81452581e-02 5.41691184e-01
4.41942327e-02 1.14581861e-01 1.30189514e+00 -1.03117383e+00
-5.74739754e-01 -2.05314219e-01 -5.09350121e-01 1.03487894e-01
6.34133399e-01 -2.57098764e-01 -5.03903747e-01 4.80762720e-02
-7.86393940e-01 -3.56745899e-01 1.06448221e+00 -1.91750169e-01
7.17586458e-01 -4.82629180e-01 -8.21969151e-01 5.49525619e-01
1.32867491e+00 -1.46096572e-01 2.21954957e-01 2.48227119e-01
1.55251849e+00 6.23785317e-01 7.66776681e-01 1.53114051e-01
1.24815226e+00 6.61373317e-01 -1.83580443e-01 -1.87486392e-02
2.16514915e-02 -8.89662147e-01 1.76683366e-01 1.19433212e+00
-9.56156030e-02 -3.63461077e-01 -1.09223318e+00 5.72825789e-01
-2.10312319e+00 -5.47300577e-01 -1.70720443e-01 1.74661314e+00
1.29176426e+00 1.30951136e-01 -2.58451730e-01 -3.38865668e-01
7.50804722e-01 1.91319615e-01 -6.28861904e-01 1.56567752e-01
-4.41351593e-01 -4.65518758e-02 4.20245677e-01 6.28678501e-01
-1.33192718e+00 1.55746794e+00 4.54013109e+00 9.86050129e-01
-6.83597505e-01 -1.85050387e-02 -4.99005839e-02 4.71945196e-01
-4.67784524e-01 1.82994917e-01 -1.41054380e+00 1.92243278e-01
5.93145788e-01 -7.57557154e-01 1.20989652e-02 8.07560265e-01
-1.77599285e-02 2.78865784e-01 -1.02585173e+00 6.35445654e-01
1.02126271e-01 -1.14245760e+00 5.31833053e-01 -7.85812736e-02
8.06135416e-01 -1.91798255e-01 -3.97777021e-01 9.51707244e-01
7.64121413e-01 -9.43980992e-01 1.42639950e-01 8.21884334e-01
6.12006664e-01 -7.89651632e-01 1.05037594e+00 4.05162424e-01
-1.99976563e+00 1.88907102e-01 -3.40469122e-01 2.62358338e-01
3.18487734e-01 7.11834848e-01 -9.27282512e-01 1.43539274e+00
2.71056622e-01 1.11109972e+00 -5.52701116e-01 6.94741309e-01
-8.16972256e-01 5.80381513e-01 -1.90239847e-01 1.54891446e-01
6.19446561e-02 -1.63954981e-02 1.88526154e-01 1.49052823e+00
-1.26460388e-01 2.61541069e-01 6.38842225e-01 6.60085320e-01
-5.68712473e-01 6.10891700e-01 -5.04552960e-01 -3.90696555e-01
7.19901562e-01 1.23441744e+00 -7.29328319e-02 -5.56897938e-01
-8.10845971e-01 6.06917202e-01 1.02802885e+00 5.23682356e-01
-6.28972828e-01 -6.09314024e-01 5.60446322e-01 -1.15494207e-01
5.03271163e-01 -2.64375925e-01 9.82101932e-02 -1.56783092e+00
1.53399631e-01 -2.14277148e-01 9.15373385e-01 -6.00359440e-01
-1.62831044e+00 3.63321275e-01 4.31141295e-02 -8.76335323e-01
-1.97547361e-01 -4.15888667e-01 -3.31890792e-01 1.08320510e+00
-2.07824969e+00 -1.86191130e+00 -1.86405987e-01 5.44079840e-01
-9.08635706e-02 -1.57078803e-01 7.84186125e-01 5.26039302e-01
-7.95396745e-01 8.60424936e-01 -1.59207493e-01 7.64825583e-01
7.21483290e-01 -1.41883302e+00 2.73662388e-01 5.53887725e-01
4.29167897e-01 1.15866160e+00 8.05999339e-02 -1.10073709e+00
-1.19902217e+00 -1.60294294e+00 1.36334729e+00 -4.05431449e-01
1.03995550e+00 -1.49570450e-01 -1.33543718e+00 1.25473607e+00
-1.59776732e-01 3.76301676e-01 1.15417206e+00 6.84284747e-01
-7.64955044e-01 -1.32781222e-01 -5.40735424e-01 3.47813725e-01
1.24504101e+00 -9.00808394e-01 -9.55744922e-01 1.94490775e-01
1.03694296e+00 -4.44525003e-01 -1.36847293e+00 7.73106396e-01
3.59515578e-01 1.01025276e-01 1.02870727e+00 -1.10165191e+00
2.14951932e-01 -5.09965062e-01 9.07667577e-02 -1.40909970e+00
-2.75956810e-01 -5.60948327e-02 -1.05096281e+00 1.80317295e+00
8.44875157e-01 -7.59356141e-01 7.14992702e-01 5.20234466e-01
2.38181762e-02 -8.78281474e-01 -3.27462435e-01 -9.30796146e-01
-1.76033244e-01 -9.26546454e-02 5.54735661e-01 1.54260850e+00
4.65329051e-01 8.14401925e-01 -2.29754776e-01 6.15847409e-01
3.86478662e-01 3.86209488e-01 5.76255977e-01 -1.29663694e+00
-3.11749548e-01 -8.89833346e-02 -5.54473519e-01 -1.21402216e+00
7.18725443e-01 -1.25406241e+00 -4.43356782e-01 -1.86166203e+00
2.95658082e-01 -8.54730248e-01 -6.11511171e-01 9.26503897e-01
-7.95240104e-01 -3.13180923e-01 -9.35334936e-02 8.79072398e-02
-8.62098694e-01 7.85548985e-01 1.06588566e+00 -2.23144814e-01
-2.80119479e-01 -2.50485182e-01 -8.82854402e-01 8.18171084e-01
5.35036266e-01 -3.18499774e-01 -6.06709957e-01 -4.82762188e-01
4.58685875e-01 -1.41985819e-01 1.16983997e-02 -5.34429073e-01
6.16812706e-01 -2.47874677e-01 1.33526385e-01 -4.79917765e-01
1.77505493e-01 -6.90668941e-01 -6.91679260e-03 -1.21019647e-01
-4.22381818e-01 -5.07540286e-01 6.47161305e-02 8.04583311e-01
-5.99282146e-01 -7.34285191e-02 4.92526107e-02 -2.02605892e-02
-1.12428582e+00 6.24612331e-01 3.93174350e-01 3.50584000e-01
9.07757282e-01 3.11652541e-01 -6.10041857e-01 -5.10812551e-02
-1.08706486e+00 8.39418590e-01 1.45566002e-01 5.73981643e-01
3.50763232e-01 -1.59602082e+00 -6.54228985e-01 -1.79904386e-01
6.18754268e-01 4.02413785e-01 2.82441944e-01 6.08901799e-01
-1.73913196e-01 6.25052869e-01 3.48798573e-01 -1.26222894e-02
-1.34705889e+00 4.71958309e-01 2.17792362e-01 -9.37064946e-01
-4.69985187e-01 1.01437795e+00 2.94167161e-01 -7.89654076e-01
2.44290009e-01 -2.93734789e-01 -7.41709232e-01 1.58811688e-01
3.29039305e-01 -1.18660547e-01 -2.68386826e-02 -6.47545934e-01
-4.88942266e-01 5.13655901e-01 -4.63356107e-01 4.03745115e-01
1.20066845e+00 -1.83150709e-01 -3.43013227e-01 5.85756660e-01
9.40839291e-01 3.56391996e-01 -6.32390618e-01 -1.09536052e+00
5.42548239e-01 -1.22149259e-01 -6.47066906e-02 -9.61038232e-01
-9.73045051e-01 6.19359076e-01 -3.12078446e-01 -5.61570637e-02
7.19178796e-01 4.19466615e-01 8.40877354e-01 8.20913374e-01
5.52117169e-01 -7.11712480e-01 -3.42278212e-01 8.50044250e-01
5.68688035e-01 -1.17844546e+00 1.63959175e-01 -1.17711723e+00
-7.66723931e-01 6.55751705e-01 1.06894910e+00 3.16819847e-01
7.74594545e-01 1.28035873e-01 -2.10590035e-01 -2.26793438e-01
-8.55549634e-01 -7.26940751e-01 7.63127983e-01 5.31085074e-01
7.61259854e-01 2.14398384e-01 -2.73293048e-01 1.12809157e+00
-8.54956061e-02 -3.63371670e-02 -4.16069031e-02 9.75433290e-01
-3.47067565e-01 -1.34097362e+00 4.28397268e-01 4.97489899e-01
-1.53988600e-01 -5.39454043e-01 -5.20642877e-01 6.02261782e-01
3.37397456e-01 6.58410966e-01 -3.53008986e-01 -3.86694551e-01
3.55018795e-01 2.48601496e-01 2.67116278e-01 -1.21710765e+00
-4.50155795e-01 -5.34346223e-01 6.60417616e-01 -9.50450897e-02
-4.18661028e-01 -2.32636571e-01 -1.60790920e+00 -7.09860995e-02
-7.73257732e-01 4.65945482e-01 1.06466673e-01 1.22337282e+00
3.85107130e-01 8.10118258e-01 3.31745267e-01 4.18773368e-02
6.93865307e-03 -9.35141444e-01 -8.18297446e-01 3.12141448e-01
1.48582924e-02 -8.62984478e-01 1.08543579e-02 1.74772918e-01] | [9.090047836303711, 8.24321460723877] |
f84ba9aa-e36f-4c1e-be7c-3cbe946b69f6 | neural-symbolic-argumentation-mining-an | 1905.09103 | null | https://arxiv.org/abs/1905.09103v3 | https://arxiv.org/pdf/1905.09103v3.pdf | Neural-Symbolic Argumentation Mining: an Argument in Favor of Deep Learning and Reasoning | Deep learning is bringing remarkable contributions to the field of argumentation mining, but the existing approaches still need to fill the gap toward performing advanced reasoning tasks. In this position paper, we posit that neural-symbolic and statistical relational learning could play a crucial role in the integration of symbolic and sub-symbolic methods to achieve this goal. | ['Andrea Galassi', 'Xiaoting Shao', 'Marco Lippi', 'Kristian Kersting', 'Paolo Torroni'] | 2019-05-22 | null | null | null | null | ['component-classification'] | ['natural-language-processing'] | [-1.75838858e-01 6.30221069e-01 -5.15951455e-01 -4.53030974e-01
-3.12610060e-01 -8.24827626e-02 7.59482920e-01 5.93727708e-01
-6.21214919e-02 8.36194038e-01 1.23249702e-01 -8.80640626e-01
-4.58399326e-01 -1.14757562e+00 -5.81358910e-01 -3.04011226e-01
-8.94984007e-02 5.29542387e-01 2.34883517e-01 -8.79831493e-01
4.14102316e-01 3.24292749e-01 -1.26050496e+00 7.31623709e-01
6.25585854e-01 1.05780149e+00 -8.12111020e-01 1.77848443e-01
-5.40089369e-01 1.56858253e+00 -4.97365087e-01 -9.36697543e-01
-1.35446340e-01 -3.31174672e-01 -1.06472135e+00 -7.89377391e-01
-1.05981380e-01 -2.42020532e-01 -9.49874073e-02 1.12880874e+00
1.96899533e-01 3.74465659e-02 6.65215850e-01 -1.08016300e+00
-4.18039441e-01 1.48408735e+00 -5.29966950e-01 1.96162924e-01
3.63796592e-01 -3.65754038e-01 1.18664122e+00 -6.19076133e-01
2.09406808e-01 1.26865804e+00 8.28409910e-01 3.45128268e-01
-7.03979075e-01 -5.99712431e-01 1.15632780e-01 5.50332487e-01
-7.15593398e-01 -1.70658693e-01 1.27157962e+00 -4.07739162e-01
8.38517547e-01 -1.08327530e-01 8.05347085e-01 1.35955954e+00
-6.73203170e-02 9.33878362e-01 1.00753808e+00 -6.70593321e-01
1.49210066e-01 -1.06675908e-01 3.43058050e-01 9.01855707e-01
4.17136371e-01 2.82404423e-01 -8.44384730e-01 -2.53262728e-01
6.32766485e-01 -2.29495600e-01 3.25145781e-01 -2.66306937e-01
-9.33925986e-01 1.34694993e+00 5.02905428e-01 6.97562337e-01
-1.58487067e-01 2.43473262e-01 8.98193479e-01 3.42471272e-01
4.33492392e-01 6.90402031e-01 -5.27733684e-01 -3.73367727e-01
-9.69793499e-01 5.77788472e-01 1.00558841e+00 2.43484899e-01
2.00468734e-01 2.47016698e-01 -9.84402746e-02 3.25904489e-01
4.30220574e-01 -1.06566129e-02 5.05933762e-02 -1.28946781e+00
5.06717205e-01 9.83119607e-01 -2.24211186e-01 -9.71297741e-01
-3.80596280e-01 -2.39436388e-01 -7.48244643e-01 2.69247979e-01
5.36849260e-01 -1.41281173e-01 -1.56173468e-01 1.19102871e+00
2.94458687e-01 1.54858846e-02 2.28294492e-01 5.25012970e-01
9.01592374e-01 2.43865132e-01 2.12878007e-02 1.91131700e-02
1.08000088e+00 -7.84901321e-01 -7.27392435e-01 9.64982361e-02
6.32031441e-01 -2.93091927e-02 7.75076210e-01 6.49399221e-01
-1.33452988e+00 -2.55205333e-01 -1.34137404e+00 9.33996439e-02
-6.06588542e-01 -1.93048954e-01 1.19696140e+00 4.02683526e-01
-4.65781599e-01 5.52361310e-01 -7.46926367e-01 3.73774260e-01
7.52901495e-01 4.25330490e-01 -1.04950882e-01 3.47402096e-01
-1.81288755e+00 1.17849886e+00 6.92955077e-01 4.49826568e-01
-2.70169377e-01 -2.21994773e-01 -9.89436150e-01 6.78267777e-02
6.69006824e-01 -5.33800602e-01 1.17327607e+00 -4.37131196e-01
-1.52870202e+00 7.16807067e-01 2.38203973e-01 -1.07646000e+00
4.61676002e-01 -3.78416419e-01 3.54309240e-03 -1.16514098e-02
-9.19199139e-02 2.88839757e-01 3.43122631e-01 -9.13655519e-01
-4.41849381e-01 -7.14261055e-01 4.44462508e-01 -2.57894337e-01
-3.19668621e-01 -5.46905398e-02 3.10262293e-01 -4.61393118e-01
6.95459768e-02 -5.79180181e-01 -1.30801126e-01 -3.47738266e-01
-1.67686656e-01 -1.03482854e+00 6.62004888e-01 -3.50318283e-01
9.21835840e-01 -1.80244720e+00 3.16107035e-01 3.28244686e-01
3.05750281e-01 6.88325703e-01 4.53944236e-01 4.44415003e-01
1.99578807e-01 1.44854516e-01 -2.23019235e-02 -1.11605832e-02
1.90917647e-03 2.81980932e-01 -6.45671725e-01 8.67141783e-03
3.22273433e-01 1.19301748e+00 -9.67603564e-01 -5.03935635e-01
1.12425491e-01 3.98310989e-01 -3.70951384e-01 -1.32375464e-01
-8.26474488e-01 3.20917606e-01 -8.61652493e-01 6.64136469e-01
1.03614507e-02 -3.97627279e-02 5.10205805e-01 2.05429897e-01
1.65228784e-01 5.52070022e-01 -8.22084546e-01 1.49737620e+00
-3.27223465e-02 7.17328727e-01 1.30644478e-02 -1.93771434e+00
1.01502943e+00 2.53348827e-01 3.21848691e-01 -7.73156822e-01
5.73525012e-01 2.30023995e-01 3.36655796e-01 -5.44351757e-01
3.85011673e-01 -3.04120183e-01 -1.33251414e-01 3.76138210e-01
-2.33222350e-01 -1.51779354e-01 1.50703922e-01 2.10835058e-02
7.02061176e-01 5.96505105e-01 3.76755446e-01 1.80511683e-01
7.59653270e-01 2.13409051e-01 3.95945579e-01 4.44439262e-01
-1.34382576e-01 -1.21006720e-01 1.16124237e+00 -8.49560738e-01
-7.42749989e-01 -3.93688560e-01 -3.12674344e-02 1.18159294e+00
-2.94584334e-01 -2.52955705e-01 -5.77185035e-01 -9.17108655e-01
1.26299605e-01 6.41573608e-01 -7.97661543e-01 -1.17454976e-01
-7.37872005e-01 -3.55138421e-01 1.05418205e+00 9.27491188e-01
4.41921711e-01 -1.44400787e+00 -5.49441695e-01 2.55831122e-01
-2.40378812e-01 -8.91270220e-01 1.06862962e+00 5.88434637e-01
-9.16294694e-01 -1.34984732e+00 -1.79437771e-01 -5.68590820e-01
-1.32060014e-02 -2.32315376e-01 8.85636389e-01 2.47468024e-01
1.35370851e-01 -2.57578373e-01 -5.44184387e-01 -8.31885099e-01
-5.30106664e-01 4.11815405e-01 -3.70149136e-01 -5.42543888e-01
6.08713806e-01 -7.37580597e-01 -1.01568654e-01 -3.25882316e-01
-5.72081387e-01 -1.06243856e-01 2.41661921e-01 8.21641505e-01
1.67729408e-02 4.00607064e-02 6.36851251e-01 -9.03576612e-01
1.14420986e+00 -5.84375799e-01 -4.77517009e-01 3.78037304e-01
-5.88813245e-01 3.42419177e-01 4.20774817e-01 -1.14726916e-01
-6.61253929e-01 -7.33990192e-01 -1.95907310e-01 -1.82111800e-01
1.29291862e-01 1.31966782e+00 1.14864841e-01 1.50558382e-01
6.74588501e-01 -1.10476494e-01 1.70803279e-01 -1.00748248e-01
5.06307781e-01 3.53724301e-01 1.93602324e-01 -1.13867676e+00
4.96712774e-01 5.18047988e-01 5.19017756e-01 -1.10312432e-01
-1.16759765e+00 2.30531514e-01 -5.70786536e-01 8.47768337e-02
6.53230131e-01 -5.47618747e-01 -1.09025717e+00 3.19434516e-02
-1.20614219e+00 -4.81669575e-01 -4.06659633e-01 2.46066958e-01
-4.04679000e-01 3.87123913e-01 -5.95684171e-01 -1.07032824e+00
-3.98921490e-01 -8.86269689e-01 7.53381014e-01 -1.68856531e-02
-4.58094746e-01 -1.04777277e+00 1.30212963e-01 6.51923656e-01
2.15807885e-01 5.72856665e-01 1.15472341e+00 -1.07894814e+00
-1.44683067e-02 -4.22679842e-01 -1.29120618e-01 3.70965034e-01
-3.91695112e-01 6.83953762e-02 -6.93814337e-01 3.89765888e-01
6.18307367e-02 -8.10954034e-01 8.25492740e-01 2.83810586e-01
1.13083327e+00 1.04223415e-02 -8.28056931e-02 1.72229901e-01
7.55818665e-01 1.13501623e-01 4.75679368e-01 7.46337414e-01
4.71127450e-01 9.36151087e-01 7.85117149e-01 3.66177708e-01
7.62863338e-01 5.44196606e-01 5.99214494e-01 1.12771504e-01
5.27913384e-02 -2.60364950e-01 1.34457693e-01 3.97670329e-01
-5.86877286e-01 2.84700543e-01 -1.43824315e+00 2.52843857e-01
-2.28868937e+00 -1.21837080e+00 -4.11743969e-01 1.23088026e+00
9.86430287e-01 6.76505327e-01 1.33662522e-01 8.97548795e-01
1.88183010e-01 1.24680117e-01 -3.23462963e-01 -7.69868433e-01
-2.99246758e-01 3.81480604e-01 -2.91081041e-01 3.96221936e-01
-9.88091528e-01 9.97955799e-01 6.67452669e+00 6.25382781e-01
-9.03936088e-01 -1.25890002e-01 4.46823329e-01 1.74370050e-01
-3.63605976e-01 -9.70613211e-02 -5.51934540e-01 2.54257862e-02
1.12616301e+00 2.92722642e-01 3.11794311e-01 8.54311407e-01
-8.87291878e-02 -1.12238422e-01 -1.09962261e+00 3.89912158e-01
2.39779195e-03 -1.70516276e+00 4.92965579e-02 1.39208347e-01
5.88251770e-01 -2.07067385e-01 -3.65849324e-02 6.70950234e-01
5.12692928e-01 -1.18779469e+00 5.46735704e-01 7.01836765e-01
1.15551628e-01 -6.77294195e-01 1.27017784e+00 4.67616856e-01
-6.85845971e-01 -3.39917421e-01 -9.51062441e-02 -6.75870955e-01
-1.96695015e-01 4.15948242e-01 -5.46449482e-01 8.25942636e-01
5.84601998e-01 6.58658206e-01 -2.50405073e-01 2.50012010e-01
-6.49146318e-01 6.60521448e-01 1.05002522e-01 -2.58805245e-01
3.63920093e-01 -1.12390414e-01 2.01086536e-01 7.10383832e-01
-2.21161738e-01 -1.07543893e-01 -1.67907588e-02 9.40688074e-01
5.71094789e-02 -8.41425881e-02 -7.10151672e-01 -2.48752356e-01
2.13982329e-01 7.75306463e-01 -6.40961587e-01 -3.76830637e-01
-2.71230400e-01 7.92190954e-02 4.75279152e-01 7.41272792e-02
-1.02692163e+00 -2.19907567e-01 1.09670281e-01 1.47839030e-03
2.21570969e-01 -4.61980343e-01 -8.61983001e-01 -1.01110709e+00
-2.04476882e-02 -1.09213090e+00 3.81902397e-01 -7.50155091e-01
-1.02469301e+00 2.90152013e-01 7.36068562e-02 -5.40919542e-01
-6.00058496e-01 -8.77797127e-01 -6.87228143e-01 3.53072762e-01
-1.49284005e+00 -1.50781739e+00 5.07794358e-02 4.98440951e-01
3.18592370e-01 -5.61274648e-01 9.60642636e-01 -1.20521463e-01
-2.83032566e-01 3.77584845e-01 -3.20917875e-01 4.58220541e-01
1.34301931e-01 -1.05720127e+00 7.68959597e-02 2.45174810e-01
5.48871934e-01 6.32790565e-01 3.84203434e-01 -3.47525567e-01
-1.12092340e+00 -3.13659310e-01 9.61065829e-01 -6.52364075e-01
9.56884027e-01 -8.39205459e-02 -9.24713731e-01 3.88657540e-01
2.17650533e-01 -1.48980007e-01 8.79741669e-01 7.38745272e-01
-5.37809908e-01 7.21460804e-02 -8.43531728e-01 4.34840411e-01
7.20420599e-01 -7.87656546e-01 -1.19942045e+00 -1.82082906e-01
6.80870891e-01 -3.81273925e-01 -9.50691402e-01 7.85396755e-01
6.02447569e-01 -1.08065498e+00 9.65809762e-01 -1.06841612e+00
9.85512257e-01 -4.51788455e-02 -6.39212579e-02 -5.42188942e-01
1.97107971e-01 -3.76133054e-01 -8.54864895e-01 1.14767826e+00
4.54469174e-01 -3.20801944e-01 1.00251925e+00 2.90859789e-01
9.09904614e-02 -1.32575810e+00 -7.22791970e-01 -3.20550829e-01
6.51822150e-01 -8.07479978e-01 6.60773218e-01 1.08435524e+00
6.16462946e-01 4.58802432e-01 -1.85791068e-02 -3.74881625e-01
2.86076009e-01 7.85615668e-02 5.59843361e-01 -1.69368255e+00
8.48699883e-02 -8.46778631e-01 -1.51469871e-01 -1.92814663e-01
7.67524719e-01 -9.27062869e-01 -4.71832335e-01 -1.82723773e+00
-9.19779688e-02 -2.40333095e-01 -5.04092336e-01 6.94213212e-01
1.71490327e-01 -6.92690089e-02 -1.00348637e-01 -5.13407495e-03
-6.66056514e-01 6.03412747e-01 1.01455438e+00 -3.15081000e-01
-2.88016419e-03 1.13914363e-01 -1.02441740e+00 9.27930176e-01
8.91222179e-01 -2.50732780e-01 -2.92273223e-01 -1.33631751e-01
1.00082552e+00 2.98339605e-01 5.24850905e-01 -6.23144448e-01
3.35651278e-01 -2.45269805e-01 7.37395659e-02 -4.81775612e-01
1.86417669e-01 -4.59815353e-01 -7.57155597e-01 4.58499134e-01
-6.47260070e-01 -1.06115878e-01 1.68435797e-01 3.32215965e-01
-5.74585021e-01 -2.28332698e-01 2.19498649e-01 -6.96658343e-02
-3.03848863e-01 -3.27797532e-01 -2.01578408e-01 2.39817306e-01
6.61908567e-01 -2.39797950e-01 -2.51579583e-01 -2.25396410e-01
-8.94359112e-01 3.48174646e-02 -2.07947895e-01 2.94278532e-01
5.65626979e-01 -8.76738071e-01 -6.79659188e-01 -6.10335469e-02
-2.92630047e-01 1.85262129e-01 -4.51120466e-01 9.30189908e-01
-3.09800684e-01 7.28742123e-01 -1.99772820e-01 -1.33090988e-01
-8.43807638e-01 3.69315684e-01 2.31685117e-01 -7.04616666e-01
-4.30359036e-01 7.47659028e-01 -7.44988859e-01 -4.52228427e-01
6.35468006e-01 -2.48270735e-01 -6.74901307e-01 2.92376995e-01
2.13809177e-01 4.34620678e-01 9.09065548e-03 7.77265709e-03
-1.88312218e-01 2.72783786e-01 1.48581848e-01 -6.64075762e-02
1.65780270e+00 6.20655417e-01 -3.85197192e-01 7.34380662e-01
5.38277507e-01 5.60753979e-02 -7.86534905e-01 -2.81084359e-01
4.70224857e-01 2.08168611e-01 1.97071638e-02 -8.27844799e-01
-7.60744035e-01 1.30352581e+00 -5.07566370e-02 5.57759106e-01
4.46639925e-01 2.72675067e-01 5.89929581e-01 6.55361414e-01
4.69476253e-01 -8.28548431e-01 3.00540298e-01 8.79376590e-01
7.25001752e-01 -1.33381152e+00 2.89551467e-01 -1.12322167e-01
-6.48267746e-01 1.40941906e+00 5.88061988e-01 -3.49910587e-01
6.52186632e-01 3.87420475e-01 -3.15823138e-01 -7.17921853e-01
-5.93048811e-01 -2.52486944e-01 3.30684662e-01 4.26830798e-01
8.62142086e-01 4.90745492e-02 -3.31198215e-01 9.04129744e-01
-3.88207197e-01 2.26751044e-01 -1.77853983e-02 8.53221059e-01
-2.97405064e-01 -1.44630599e+00 -1.87880799e-01 3.09724584e-02
-6.01223886e-01 -1.39817521e-01 -6.79230869e-01 9.41532016e-01
3.21059585e-01 7.51584709e-01 -2.60434628e-01 -3.66371661e-01
1.24705404e-01 2.49504372e-01 5.80858529e-01 -3.61085415e-01
-8.42995524e-01 -1.86590016e-01 4.19156551e-01 -4.38042372e-01
-7.57265866e-01 -5.00959754e-01 -1.46517301e+00 -5.26154280e-01
-2.59023219e-01 4.30073828e-01 6.93834722e-01 1.24545860e+00
4.36976664e-02 6.38518453e-01 -8.82871523e-02 -4.78458643e-01
-7.13280439e-01 -9.31317091e-01 -1.48613214e-01 -9.89305303e-02
6.43599033e-03 -8.34615886e-01 -1.67323127e-01 -3.84468138e-01] | [9.039240837097168, 7.126823902130127] |
d0e28359-4703-4e50-bf7e-b0a8fb124a2c | improving-multitask-retrieval-by-promoting | 2307.00342 | null | https://arxiv.org/abs/2307.00342v1 | https://arxiv.org/pdf/2307.00342v1.pdf | Improving Multitask Retrieval by Promoting Task Specialization | In multitask retrieval, a single retriever is trained to retrieve relevant contexts for multiple tasks. Despite its practical appeal, naive multitask retrieval lags behind task-specific retrieval in which a separate retriever is trained for each task. We show that it is possible to train a multitask retriever that outperforms task-specific retrievers by promoting task specialization. The main ingredients are: (1) a better choice of pretrained model (one that is explicitly optimized for multitasking) along with compatible prompting, and (2) a novel adaptive learning method that encourages each parameter to specialize in a particular task. The resulting multitask retriever is highly performant on the KILT benchmark. Upon analysis, we find that the model indeed learns parameters that are more task-specialized compared to naive multitasking without prompting or adaptive learning. | ['Arnold Overwijk', 'Karl Stratos', 'Chenyan Xiong', 'Wenzheng Zhang'] | 2023-07-01 | null | null | null | null | ['retrieval'] | ['methodology'] | [ 3.38325471e-01 -5.58101714e-01 -2.46952906e-01 -2.44394884e-01
-1.48910570e+00 -1.02051699e+00 6.89515173e-01 7.05938563e-02
-7.85711944e-01 5.40119231e-01 2.54024774e-01 -2.05014139e-01
-5.71389198e-01 -1.53116405e-01 -4.54988450e-01 -7.65382469e-01
1.46015108e-01 1.05242157e+00 3.77979010e-01 -3.94612074e-01
5.60168803e-01 1.86180130e-01 -1.58301461e+00 5.55947542e-01
8.67728114e-01 6.94503605e-01 1.07576966e+00 7.40759134e-01
3.04095410e-02 3.90112221e-01 -6.59592390e-01 3.30654457e-02
2.25036308e-01 1.66819900e-01 -9.54803467e-01 -3.91133517e-01
7.32452035e-01 8.88413861e-02 3.85851488e-02 4.49612767e-01
6.73299074e-01 5.49090743e-01 9.72306728e-01 -7.35230803e-01
-7.06213355e-01 3.22523713e-01 -3.98310363e-01 6.21043026e-01
2.77623087e-01 3.73809785e-02 1.34683156e+00 -1.11942339e+00
5.23885190e-01 1.02753985e+00 1.64447650e-01 4.44052666e-01
-1.33475196e+00 -6.29980862e-01 1.51766062e-01 -2.32907012e-01
-1.31554627e+00 -4.78412896e-01 4.70944375e-01 -3.47763449e-01
1.35474002e+00 3.52190435e-01 3.19273472e-01 1.21076250e+00
4.60127532e-01 8.27379286e-01 1.14255679e+00 -5.33129752e-01
1.59422662e-02 2.36419782e-01 5.11961520e-01 1.93456590e-01
2.39040777e-01 3.66244875e-02 -8.40802372e-01 -5.23489654e-01
3.65070194e-01 2.46804222e-01 -1.81460381e-01 -4.30328399e-01
-1.43176591e+00 5.03837168e-01 -5.19701764e-02 5.33473253e-01
-4.77280051e-01 3.71105313e-01 5.75879633e-01 8.68025661e-01
6.09552324e-01 1.11345947e+00 -8.54305089e-01 1.22229755e-01
-8.00323486e-01 4.69950736e-01 6.43966496e-01 1.08446169e+00
1.15071726e+00 -1.98228702e-01 -7.72990406e-01 1.04885256e+00
7.18057156e-04 7.83357322e-01 8.10734570e-01 -6.93071485e-01
4.26181257e-01 2.75101960e-01 2.53000468e-01 -4.22751933e-01
-3.86085153e-01 -4.72644538e-01 -2.25754976e-01 -4.38504159e-01
7.24479631e-02 -3.51425447e-02 -7.44353712e-01 1.67276645e+00
-4.79168557e-02 -1.37104973e-01 1.96180165e-01 7.30456948e-01
6.62026167e-01 6.83904111e-01 6.55517042e-01 -1.80381805e-01
1.52447510e+00 -1.07980132e+00 -6.59142494e-01 -8.45645308e-01
1.13384140e+00 -1.02325487e+00 1.50621712e+00 3.14314634e-01
-9.85990286e-01 -4.67886060e-01 -7.17978477e-01 -3.65616977e-01
-6.23078644e-01 5.17584719e-02 9.79089439e-01 3.86991233e-01
-1.31602526e+00 -1.25282675e-01 2.42917649e-02 -4.09942091e-01
-1.39603466e-01 4.70316410e-01 -3.30771148e-01 -1.42766714e-01
-1.33816016e+00 1.25676000e+00 6.78422689e-01 -3.00550014e-01
-1.23078299e+00 -7.98269331e-01 -6.06253922e-01 2.56853014e-01
7.17285395e-01 -1.01775098e+00 1.45085108e+00 -8.12352180e-01
-9.62913334e-01 1.04449606e+00 -3.29916030e-01 -6.00547567e-02
-1.81467384e-01 -5.53023756e-01 -9.76799354e-02 1.07018761e-01
3.25054914e-01 5.47341347e-01 1.34368396e+00 -1.40864789e+00
-6.36283815e-01 -2.62103409e-01 9.75011364e-02 7.51025736e-01
-5.33296585e-01 3.51057619e-01 -6.57099187e-01 -7.38197565e-01
-2.62145638e-01 -9.69976187e-01 -3.39566693e-02 -7.06277430e-01
-9.01183411e-02 -7.08258092e-01 7.39563942e-01 -1.57551527e-01
1.15001690e+00 -2.05444098e+00 4.15723443e-01 6.42964095e-02
3.30272406e-01 2.53815025e-01 -5.95503330e-01 4.29767340e-01
-4.93229702e-02 5.60073070e-02 2.36871675e-01 -4.81376290e-01
-2.36497834e-01 7.13706315e-02 -5.25799036e-01 1.15006760e-01
-1.11651331e-01 1.25724852e+00 -1.12223613e+00 -6.79810882e-01
-9.84704643e-02 1.72091827e-01 -4.44594890e-01 2.48111352e-01
-4.72634077e-01 9.93741304e-02 -9.92135704e-01 6.88129246e-01
2.80311167e-01 -6.38861895e-01 2.64854193e-01 5.81215583e-02
2.77236998e-01 5.40054917e-01 -7.93127537e-01 2.04192019e+00
-9.69229043e-01 3.97074014e-01 1.08649321e-01 -7.18989372e-01
7.83907354e-01 6.10717416e-01 2.97959089e-01 -7.98286259e-01
-2.31130362e-01 3.97308975e-01 -3.23734492e-01 -5.51879287e-01
1.09697962e+00 -1.09969541e-01 -2.56920666e-01 1.13360202e+00
2.30373189e-01 -3.95866573e-01 7.27362856e-02 5.41706085e-01
1.26884735e+00 -5.84075190e-02 5.06784581e-02 -6.60763264e-01
1.80875838e-01 -1.53820151e-02 3.60951386e-02 1.20223880e+00
9.27512869e-02 2.78326929e-01 -1.94323778e-01 -3.78339291e-01
-6.98666155e-01 -7.21880734e-01 6.41116202e-02 2.50894332e+00
1.33826479e-01 -4.96389687e-01 1.99504197e-02 -8.57506394e-01
1.78410903e-01 6.49761975e-01 -6.27155721e-01 -3.15752387e-01
-6.42544031e-01 -6.65641427e-01 1.89963281e-01 3.12052816e-01
-3.59146781e-02 -1.52213299e+00 -5.47277868e-01 1.28520250e-01
-3.82583708e-01 -8.03964674e-01 -1.08924496e+00 9.13662732e-01
-1.09497106e+00 -8.53600085e-01 -8.72226119e-01 -9.21955705e-01
4.22170162e-01 1.16371620e+00 1.73618782e+00 4.13355231e-01
1.41783133e-01 8.83872986e-01 -2.62118399e-01 -3.48802835e-01
-7.75367096e-02 9.64069128e-01 -6.54563978e-02 -4.15814906e-01
4.81754214e-01 -3.80474657e-01 -4.08476621e-01 3.88532728e-01
-1.22194529e+00 -2.03918636e-01 9.64283586e-01 8.67797494e-01
4.03829008e-01 -4.02885646e-01 1.05701172e+00 -1.19005322e+00
1.18023980e+00 -6.30171955e-01 -3.22787493e-01 9.09285426e-01
-8.86534750e-01 2.52325952e-01 2.19676971e-01 -7.55249798e-01
-1.18189526e+00 -6.18127994e-02 5.21888316e-01 -1.37061268e-01
1.15937807e-01 5.18478632e-01 4.09548193e-01 -3.97144705e-02
1.10934222e+00 3.56700271e-01 -3.09871703e-01 -5.05646169e-01
2.86002278e-01 5.09729743e-01 -9.02581960e-02 -1.25036192e+00
6.68536007e-01 -1.87052134e-02 -3.18270922e-01 -5.08176565e-01
-1.28684127e+00 -9.59509730e-01 -3.17120612e-01 -1.37090787e-01
6.49036586e-01 -1.20589626e+00 -5.91732264e-01 5.22133708e-02
-1.07747567e+00 -7.08008945e-01 -2.43201405e-01 2.14536399e-01
-4.83197898e-01 -1.21328309e-01 -3.28404427e-01 -5.66238344e-01
-6.05459332e-01 -1.17758477e+00 1.62362933e+00 5.50464019e-02
-2.14004830e-01 -1.26273167e+00 1.79557309e-01 3.74214530e-01
8.24677050e-01 -8.15609694e-01 1.09083688e+00 -1.26709867e+00
-7.39517570e-01 -1.11311622e-01 -2.67630100e-01 8.39567930e-02
1.29903361e-01 -4.87809569e-01 -1.02822888e+00 -6.06210411e-01
1.87960286e-02 -8.68435144e-01 1.07593513e+00 2.97564626e-01
9.51680005e-01 -2.27112956e-02 -6.79115891e-01 2.40666538e-01
1.29396486e+00 5.00265174e-02 2.60125041e-01 5.44133008e-01
4.88665342e-01 6.16056263e-01 7.73497283e-01 -1.34092391e-01
3.70195478e-01 7.90992498e-01 2.87587419e-02 2.28263795e-01
3.45650762e-02 1.30858064e-01 3.00398618e-01 6.53628528e-01
-2.45868806e-02 -1.13138042e-01 -1.02941418e+00 7.39353597e-01
-1.91246796e+00 -6.50040269e-01 2.08691403e-01 2.17885613e+00
1.17404020e+00 -7.30908439e-02 3.04345880e-03 -5.21267295e-01
3.46931845e-01 3.38265896e-01 -4.63571340e-01 -1.84234768e-01
-2.19586566e-01 4.22296643e-01 2.78868556e-01 6.09732032e-01
-9.76828098e-01 1.41643155e+00 7.68626451e+00 1.09366906e+00
-8.07138681e-01 2.17361853e-01 1.84026077e-01 -3.36595565e-01
-6.48056090e-01 1.56291679e-01 -1.13868868e+00 9.46697667e-02
8.18348825e-01 -3.98793697e-01 5.87433815e-01 8.11648905e-01
-4.26657021e-01 -3.47852975e-01 -1.20018816e+00 6.50223672e-01
2.09516257e-01 -1.05980444e+00 3.14199209e-01 -6.90141022e-02
7.49454141e-01 -6.44949749e-02 2.96687007e-01 8.60646725e-01
4.95510995e-01 -8.64905238e-01 3.57984632e-01 5.19098520e-01
4.65467244e-01 -3.95338684e-01 3.39089602e-01 4.66909051e-01
-8.50115418e-01 -1.45832732e-01 -5.02055228e-01 3.05319309e-01
-1.97482124e-01 4.04991060e-01 -7.78314829e-01 4.57630336e-01
7.51737714e-01 4.60308373e-01 -9.04364645e-01 8.84773195e-01
-1.59185722e-01 3.67440909e-01 -1.21338844e-01 -4.36758064e-02
6.40088767e-02 1.51273206e-01 6.29982889e-01 1.37548745e+00
1.77830145e-01 -1.43212691e-01 3.93108994e-01 2.56360918e-01
-9.73350853e-02 3.12753052e-01 -9.40397501e-01 1.25059664e-01
7.70142317e-01 1.51211298e+00 -1.93252251e-01 -4.98767674e-01
-1.78007171e-01 8.09156537e-01 6.09225869e-01 8.69701028e-01
-1.77151769e-01 -7.25593492e-02 2.87123889e-01 -6.26172572e-02
1.08032659e-01 -2.07161173e-01 -2.50236630e-01 -1.07227159e+00
-1.23312525e-01 -1.08780086e+00 7.32343435e-01 -1.09830689e+00
-1.41883850e+00 5.79786718e-01 3.70462418e-01 -5.98468482e-01
-4.23515409e-01 -3.91303658e-01 -2.34197721e-01 1.19488156e+00
-1.78277671e+00 -1.07885718e+00 -3.05037312e-02 7.39553809e-01
7.63768375e-01 -4.00464892e-01 1.04944694e+00 2.11101115e-01
-3.09318990e-01 4.10493881e-01 -1.09163441e-01 -5.89481950e-01
1.58441341e+00 -1.36808205e+00 -5.49146011e-02 2.58138776e-01
1.22847550e-01 1.30597985e+00 7.29547501e-01 -5.29243410e-01
-1.77807868e+00 -6.72693014e-01 1.28578067e+00 -9.93064880e-01
6.37395740e-01 -3.36937249e-01 -1.01325190e+00 1.17367375e+00
4.13599581e-01 -2.28368744e-01 7.82021165e-01 8.67669106e-01
-6.78226233e-01 -1.96773291e-01 -6.74691617e-01 4.79885727e-01
7.72980511e-01 -8.50777566e-01 -9.13902819e-01 8.71703565e-01
8.29876900e-01 -2.79138595e-01 -6.56197608e-01 2.02359691e-01
5.20590246e-01 -4.89446461e-01 1.20107162e+00 -8.11118305e-01
-5.39571755e-02 4.08326499e-02 -1.66316316e-01 -1.41634011e+00
-4.09005105e-01 -6.54001534e-01 -1.29741609e-01 7.12211013e-01
5.79103172e-01 -7.91585505e-01 3.57431024e-01 6.90797508e-01
-2.16775611e-01 -5.22732913e-01 -5.63373148e-01 -8.92091453e-01
2.77141660e-01 -1.08628064e-01 2.04119578e-01 9.45165515e-01
-2.79405266e-01 9.25013602e-01 -4.33501929e-01 -2.03891024e-01
2.86805928e-01 2.89243191e-01 7.73543000e-01 -1.54492474e+00
-2.95617580e-01 -2.80328095e-01 7.23594546e-01 -1.49090135e+00
3.60405326e-01 -1.10992908e+00 2.59992689e-01 -1.38593304e+00
5.90698838e-01 -9.31696892e-01 -6.89594209e-01 7.60834277e-01
-6.18938923e-01 -6.18101619e-02 -4.03493037e-03 6.82986081e-01
-1.25903130e+00 1.89591348e-01 1.48472786e+00 -1.65526047e-01
-4.75056857e-01 -9.73803699e-02 -1.13181961e+00 1.63857162e-01
6.63233280e-01 -9.31136370e-01 -6.52168810e-01 -6.85893416e-01
8.52429569e-01 1.09014802e-01 2.06670582e-01 -4.40827131e-01
5.66630363e-01 -1.71048135e-01 1.29059210e-01 -4.72352058e-01
4.74296659e-01 -7.83915162e-01 -1.10964663e-01 -1.52540579e-01
-7.68381894e-01 3.08450371e-01 3.38983715e-01 5.83948374e-01
-1.34970678e-03 -5.89673579e-01 1.96237251e-01 -2.56107122e-01
-7.35853791e-01 2.38995567e-01 -5.69826543e-01 2.90515333e-01
6.34460211e-01 1.44422084e-01 -6.68672204e-01 -2.61239350e-01
-5.04078269e-01 3.85717869e-01 2.60109633e-01 6.03187025e-01
3.92471492e-01 -1.22101295e+00 -6.99875832e-01 -1.17289655e-01
6.94548130e-01 -1.00066587e-01 2.30028346e-01 6.09019101e-01
3.33014995e-01 7.92665184e-01 2.49441430e-01 -5.79667032e-01
-1.20381427e+00 7.13726819e-01 7.03818575e-02 -8.44653726e-01
-3.44188362e-01 7.46307552e-01 4.99432057e-01 -4.39370215e-01
1.83454618e-01 1.47217125e-01 -1.93607599e-01 3.35955560e-01
3.56911272e-01 -5.60032167e-02 3.10472965e-01 -2.54078805e-01
-2.33203799e-01 5.50692558e-01 -8.35027218e-01 -3.10112655e-01
1.12749588e+00 -4.37176339e-02 -1.98433548e-01 5.27035892e-01
8.92741024e-01 -7.32823387e-02 -7.49622941e-01 -6.73215985e-01
3.76271218e-01 -4.46518838e-01 2.52004504e-01 -1.08088493e+00
-7.57942915e-01 4.94439334e-01 1.66343510e-01 4.44007218e-01
1.20036972e+00 1.17089868e-01 5.24270773e-01 1.20352328e+00
6.39538109e-01 -1.15000582e+00 6.77812159e-01 9.44234788e-01
1.01148069e+00 -1.20739639e+00 1.31742850e-01 1.46510014e-02
-7.39763439e-01 7.12598324e-01 6.09283447e-01 -1.11705333e-01
6.59720719e-01 5.11369221e-02 9.62685421e-02 -7.13900089e-01
-1.29919803e+00 -2.55878419e-01 6.19394898e-01 4.56717730e-01
6.29786193e-01 -2.04602480e-01 -1.83436975e-01 2.83250630e-01
2.71128237e-01 -2.31042549e-01 -1.60126448e-01 1.18130875e+00
-6.75749958e-01 -1.34887302e+00 -2.49510825e-01 6.72457099e-01
-4.01304424e-01 -5.68715453e-01 -2.69611955e-01 7.60247767e-01
-5.47078311e-01 8.69648218e-01 -2.51803011e-01 -5.45088537e-02
2.59865910e-01 6.11422956e-01 5.60234010e-01 -1.20848775e+00
-1.28946769e+00 2.13331684e-01 4.06415947e-02 -3.87142986e-01
-2.95864671e-01 -5.36167622e-01 -6.06974065e-01 1.13905139e-01
-5.75338840e-01 3.43178481e-01 4.96116102e-01 9.30222511e-01
5.78815103e-01 4.10988122e-01 3.89759183e-01 -7.54633725e-01
-5.40344238e-01 -1.14763033e+00 -5.94243765e-01 2.33810425e-01
3.85748714e-01 -7.73444593e-01 -4.56402212e-01 -1.69087917e-01] | [11.393807411193848, 7.702139854431152] |
23dd13e0-1533-4120-ad41-17a40ca99254 | vote-from-the-center-6-dof-pose-estimation-in | 2104.02527 | null | https://arxiv.org/abs/2104.02527v4 | https://arxiv.org/pdf/2104.02527v4.pdf | Vote from the Center: 6 DoF Pose Estimation in RGB-D Images by Radial Keypoint Voting | We propose a novel keypoint voting scheme based on intersecting spheres, that is more accurate than existing schemes and allows for fewer, more disperse keypoints. The scheme is based upon the distance between points, which as a 1D quantity can be regressed more accurately than the 2D and 3D vector and offset quantities regressed in previous work, yielding more accurate keypoint localization. The scheme forms the basis of the proposed RCVPose method for 6 DoF pose estimation of 3D objects in RGB-D data, which is particularly effective at handling occlusions. A CNN is trained to estimate the distance between the 3D point corresponding to the depth mode of each RGB pixel, and a set of 3 disperse keypoints defined in the object frame. At inference, a sphere centered at each 3D point is generated, of radius equal to this estimated distance. The surfaces of these spheres vote to increment a 3D accumulator space, the peaks of which indicate keypoint locations. The proposed radial voting scheme is more accurate than previous vector or offset schemes, and is robust to disperse keypoints. Experiments demonstrate RCVPose to be highly accurate and competitive, achieving state-of-the-art results on the LINEMOD 99.7% and YCB-Video 97.2% datasets, notably scoring +4.9% higher 71.1% than previous methods on the challenging Occlusion LINEMOD dataset, and on average outperforming all other published results from the BOP benchmark for these 3 datasets. Our code is available at http://www.github.com/aaronwool/rcvpose. | ['Michael Greenspan', 'Ali Etemad', 'Mohsen Zand', 'Yangzheng Wu'] | 2021-04-06 | null | null | null | null | ['6d-pose-estimation-using-rgbd'] | ['computer-vision'] | [-2.80984819e-01 -2.39427745e-01 -2.79610723e-01 -9.04673412e-02
-8.96527410e-01 -5.13519764e-01 4.54124689e-01 6.12474792e-02
-4.33705896e-01 5.06660998e-01 -6.19005077e-02 -3.08104511e-02
3.04082944e-03 -5.82440197e-01 -9.75429416e-01 -7.39033103e-01
-7.61568546e-02 5.91674685e-01 4.65303957e-01 -1.04032807e-01
5.21745980e-01 1.07385457e+00 -1.60642195e+00 -2.05229878e-01
5.04141629e-01 1.45529675e+00 2.32788194e-02 6.75246596e-01
1.98513359e-01 3.10595125e-01 -7.66816199e-01 -2.03812197e-01
4.49354231e-01 2.45578170e-01 -3.80911738e-01 -2.64025450e-01
1.17738533e+00 -5.10758281e-01 -4.20651495e-01 7.33256578e-01
6.14740431e-01 9.53744426e-02 4.74648982e-01 -1.40175605e+00
-4.08746153e-01 -6.85847551e-02 -8.85753751e-01 4.43380289e-02
5.70511878e-01 -2.17697829e-01 1.03174603e+00 -1.35615981e+00
6.88015938e-01 1.14335561e+00 7.98016310e-01 2.42743164e-01
-9.11788821e-01 -7.49340653e-01 -2.52383538e-02 1.55303225e-01
-1.75324631e+00 -1.53348312e-01 7.81729937e-01 -1.40551686e-01
1.05204642e+00 5.19200146e-01 9.42650020e-01 5.30033827e-01
3.16017896e-01 7.69942522e-01 7.61161566e-01 -3.24863523e-01
1.19747803e-01 -3.07459116e-01 -4.28275615e-02 6.92015588e-01
3.81350696e-01 1.69886183e-02 -7.01161146e-01 -4.28631842e-01
1.15992665e+00 1.90359399e-01 -4.85747457e-01 -9.81670618e-01
-1.31121361e+00 7.30152786e-01 9.94110644e-01 -6.09158315e-02
-2.98468858e-01 5.02109647e-01 -6.76677451e-02 -2.49164492e-01
4.67541844e-01 2.48275071e-01 -6.51299179e-01 -1.73640698e-01
-7.72792220e-01 5.69009602e-01 5.72614551e-01 1.06099772e+00
6.80527866e-01 -2.59909034e-01 -3.32047008e-02 6.32531643e-01
5.22526205e-01 9.01358008e-01 6.71929047e-02 -1.27937424e+00
4.61625993e-01 7.46928751e-01 3.47079396e-01 -1.19059610e+00
-6.36749804e-01 -2.99595147e-01 -5.99292576e-01 5.62363684e-01
4.90283281e-01 3.55324656e-01 -1.07946110e+00 1.30317080e+00
8.42830777e-01 2.45005324e-01 -2.67920017e-01 1.15088153e+00
1.00045717e+00 5.20319104e-01 -4.09888059e-01 8.19600821e-02
1.18349445e+00 -8.50440681e-01 -3.58029455e-01 1.85478702e-02
4.32601601e-01 -1.01484585e+00 8.29374313e-01 6.31765366e-01
-9.70757902e-01 -5.79490364e-01 -1.04027617e+00 -4.17521089e-01
-2.63370574e-01 1.83120668e-01 6.69702947e-01 3.58884066e-01
-1.03864789e+00 4.49038863e-01 -8.60120595e-01 -1.45546630e-01
4.76885498e-01 5.53339005e-01 -3.97144079e-01 -1.14714839e-01
-7.33240128e-01 8.03247631e-01 9.81295481e-02 3.15681040e-01
-4.17495459e-01 -9.26193595e-01 -9.29198682e-01 -2.91858107e-01
2.42127478e-01 -6.15914285e-01 9.77320135e-01 -3.05658281e-01
-1.32811368e+00 8.86991382e-01 -1.71785176e-01 -4.22705144e-01
7.18420804e-01 -6.23026609e-01 -7.74477199e-02 1.74279124e-01
1.17104612e-01 1.04548788e+00 8.76691997e-01 -1.33331537e+00
-8.08888257e-01 -5.95230520e-01 5.75309768e-02 4.43271309e-01
1.88840121e-01 -3.03863376e-01 -9.33945060e-01 -4.54635382e-01
7.11451352e-01 -1.16136241e+00 -1.04834892e-01 5.51452518e-01
-4.69807446e-01 -4.48534131e-01 9.40216243e-01 -2.69002944e-01
8.98384750e-01 -2.09079576e+00 1.45627126e-01 2.59449601e-01
4.25510466e-01 2.54708022e-01 1.73743993e-01 8.45900029e-02
-6.99577928e-02 -2.35908642e-01 -5.56186922e-02 -4.39782947e-01
-5.30376881e-02 1.48787554e-02 -3.26233357e-01 9.78114367e-01
1.14117563e-02 8.07733834e-01 -7.97387898e-01 -2.61044621e-01
6.39618397e-01 9.97485876e-01 -3.50054413e-01 -5.89053668e-02
-6.48530992e-03 1.86929584e-01 -3.09248418e-01 1.01152790e+00
1.08587718e+00 -1.62051469e-01 -4.31039453e-01 -5.27060926e-01
-1.74343437e-01 -2.02143844e-02 -1.49882603e+00 1.80648768e+00
-1.05695307e-01 7.61694551e-01 -3.55883181e-01 -3.80994409e-01
1.15143764e+00 -4.54338156e-02 7.35581815e-01 -5.57202995e-01
1.15958318e-01 3.40503454e-01 -4.01744902e-01 -1.05767652e-01
7.23196030e-01 5.08522213e-01 -5.00703268e-02 -8.04772303e-02
-1.81208968e-01 -7.25993812e-01 -1.06940947e-01 1.74199000e-01
9.50967789e-01 3.55408728e-01 3.65615666e-01 1.24218483e-02
4.39297587e-01 5.11838682e-02 5.66518903e-01 5.89201748e-01
-4.27572995e-01 1.04244471e+00 4.01379973e-01 -6.93707824e-01
-8.63333642e-01 -1.06683409e+00 -5.63139617e-01 4.29690063e-01
7.67794907e-01 -5.57547510e-01 -3.79152685e-01 -4.21199471e-01
5.06442726e-01 2.93364435e-01 -6.60714507e-01 1.01643205e-01
-8.11981738e-01 -3.40354562e-01 2.02385634e-01 5.80755830e-01
4.40135628e-01 -7.10305870e-01 -9.30704415e-01 8.36755559e-02
-1.99494168e-01 -1.07228017e+00 -2.76911557e-01 2.12112978e-01
-9.16586339e-01 -1.28947556e+00 -8.40094507e-01 -4.03426528e-01
6.98962331e-01 5.18296599e-01 1.08224154e+00 2.44624585e-01
-4.66598094e-01 4.43022758e-01 -4.09104407e-01 -3.66306365e-01
3.86795878e-01 -1.06348224e-01 9.17022005e-02 -2.20151827e-01
4.41064596e-01 -7.77946860e-02 -1.06709313e+00 5.54095268e-01
-5.49045205e-01 -7.97789097e-02 2.21330822e-01 6.37651980e-01
1.06131482e+00 -3.25322360e-01 -2.70129502e-01 -1.57185122e-01
-5.88622615e-02 -1.59119383e-01 -9.17885184e-01 -6.45953342e-02
-2.86979854e-01 -1.88573956e-01 1.04707763e-01 -4.39557910e-01
-4.67267245e-01 3.11561495e-01 -3.26520554e-03 -6.34558320e-01
-8.84582475e-02 -1.70641154e-01 1.06308624e-01 -6.50474191e-01
5.50046980e-01 -2.53258020e-01 -3.56504738e-01 -2.90525526e-01
4.67975289e-01 4.24915820e-01 5.56651354e-01 -4.63692158e-01
9.26489234e-01 9.22590792e-01 3.49666178e-01 -6.07716918e-01
-6.45780623e-01 -7.70576715e-01 -8.42488289e-01 -2.99496114e-01
7.19994426e-01 -9.05964673e-01 -9.06507194e-01 7.04000592e-01
-1.26623678e+00 -1.35989994e-01 -2.85575420e-01 5.81805825e-01
-6.18328333e-01 1.51364341e-01 -3.19069356e-01 -5.48014343e-01
-2.93060571e-01 -1.32442045e+00 1.55837834e+00 4.41240609e-01
-1.60440266e-01 -6.31026506e-01 1.02891371e-01 2.81617761e-01
9.84806642e-02 5.93565404e-01 3.83414060e-01 -1.82106853e-01
-8.49597812e-01 -4.62232649e-01 -1.93206787e-01 -2.38476861e-02
-9.85570475e-02 3.63858730e-01 -8.38806510e-01 -4.44841772e-01
-1.73511833e-01 -2.39910901e-01 7.46196508e-01 6.57843232e-01
9.84654427e-01 2.76088536e-01 -6.15211546e-01 9.89999413e-01
1.53688991e+00 -9.58791003e-03 7.04922020e-01 6.21575654e-01
8.55626106e-01 5.40545806e-02 1.03369713e+00 6.01785123e-01
6.07136250e-01 1.01155472e+00 1.02583635e+00 -2.64269024e-01
-2.14764670e-01 -5.81731088e-02 -5.86594716e-02 4.71566349e-01
-3.01048875e-01 -5.95797934e-02 -9.50865805e-01 4.12974626e-01
-1.77770770e+00 -3.77903759e-01 -5.59214473e-01 2.32678628e+00
5.56241870e-01 1.65286869e-01 -8.04148540e-02 1.75842106e-01
5.50240099e-01 2.86326855e-01 -8.04602921e-01 -5.37128858e-02
-1.97668001e-01 3.49971056e-01 8.70703936e-01 5.61134398e-01
-1.20319736e+00 9.64802563e-01 5.46533632e+00 6.30366981e-01
-1.03556550e+00 -1.49221048e-01 3.71473253e-01 -2.94756830e-01
6.26328588e-02 3.24541680e-03 -1.14500654e+00 2.10698143e-01
1.91253677e-01 3.57678443e-01 1.10410325e-01 8.83587003e-01
-1.38429537e-01 -5.69419444e-01 -9.26164806e-01 1.22460032e+00
1.98459908e-01 -1.23559904e+00 -2.08755314e-01 7.91130513e-02
7.58041203e-01 2.56340265e-01 2.43673697e-01 -1.32842988e-01
6.76435605e-03 -8.59434724e-01 9.43927050e-01 4.06622648e-01
6.61721289e-01 -7.41627634e-01 6.74110532e-01 5.50001226e-02
-1.40152156e+00 6.97159469e-02 -5.49825609e-01 1.91915333e-01
-2.46756114e-02 5.31793535e-01 -7.40143001e-01 4.35584009e-01
1.24010730e+00 6.43813431e-01 -5.51529586e-01 1.38832390e+00
-4.33288068e-01 -2.03343760e-02 -6.89859688e-01 -1.22817881e-01
1.15222126e-01 -4.43635173e-02 5.87023199e-01 7.17233360e-01
4.65756029e-01 1.28821447e-01 -1.30425557e-01 4.46340621e-01
-5.30259237e-02 4.21123058e-02 -4.01416272e-01 8.24771166e-01
8.20744455e-01 1.35642648e+00 -9.12912667e-01 -2.09080920e-01
-2.86046326e-01 1.01538384e+00 4.29723710e-01 2.48111352e-01
-1.08014190e+00 -5.43949425e-01 8.46915662e-01 2.63751927e-03
6.81545138e-01 -4.96616960e-01 -1.92532271e-01 -9.99811947e-01
3.22946489e-01 -6.54763460e-01 2.51069814e-01 -1.15147054e+00
-7.59172082e-01 5.49052417e-01 2.45900336e-03 -1.53946626e+00
2.08688900e-02 -8.40207219e-01 -1.78650841e-01 6.91240251e-01
-1.57903826e+00 -1.09871447e+00 -6.78210914e-01 7.15247929e-01
3.24049830e-01 3.29055846e-01 7.64978170e-01 1.08524784e-01
-1.91373393e-01 4.74328846e-01 9.09840018e-02 -6.38953969e-02
7.74599254e-01 -1.21929407e+00 5.52054465e-01 6.01882637e-01
3.88175219e-01 4.59052384e-01 5.23195803e-01 -5.63927233e-01
-1.63879287e+00 -6.99723721e-01 6.14276409e-01 -9.75399971e-01
4.30037886e-01 -5.08687973e-01 -6.74876869e-01 6.09580517e-01
-4.15320396e-01 5.92117667e-01 3.26502949e-01 -2.68883586e-01
-4.76134151e-01 -2.99273759e-01 -1.18601608e+00 5.20168602e-01
1.04044795e+00 -2.56202132e-01 -4.70983475e-01 3.94961804e-01
8.06058109e-01 -1.30599630e+00 -1.07609296e+00 6.61277294e-01
7.33951986e-01 -1.11370289e+00 1.57669413e+00 6.62234053e-02
3.15495841e-02 -6.37795508e-01 -3.77342671e-01 -1.03625667e+00
-2.31965542e-01 -5.44042647e-01 -4.66406584e-01 8.11972678e-01
6.59392029e-02 -5.19238770e-01 8.66300642e-01 5.05072176e-01
-4.50451067e-03 -1.00379145e+00 -1.42205751e+00 -5.70521057e-01
-2.25615606e-01 -6.33505583e-01 7.85180032e-01 6.45962358e-01
-4.68421102e-01 -2.04335019e-01 -1.03944011e-01 4.89031583e-01
8.41980875e-01 1.81386217e-01 1.14787757e+00 -1.23699284e+00
2.54077882e-01 -4.06338334e-01 -8.01552713e-01 -1.46317077e+00
-1.87433571e-01 -5.34249425e-01 -1.41391829e-01 -1.39937294e+00
-3.29421431e-01 -7.68790364e-01 -2.29482681e-01 4.26431209e-01
-8.16705301e-02 7.76501596e-01 2.53047884e-01 3.49655300e-01
-4.95720357e-01 4.41816479e-01 1.30629623e+00 -9.05905887e-02
-2.63278067e-01 -2.55410131e-02 -2.91938514e-01 8.43579829e-01
6.50353611e-01 -4.50467348e-01 9.98414755e-02 -4.12161946e-01
2.04563007e-01 -2.91285105e-02 6.69407845e-01 -1.04945004e+00
4.07686710e-01 6.74648136e-02 8.76135409e-01 -1.31793475e+00
7.80002892e-01 -1.09753394e+00 8.90660807e-02 5.24236500e-01
2.63953090e-01 8.45880508e-02 2.22384945e-01 3.42873544e-01
1.42405955e-02 1.07660694e-02 6.44846201e-01 1.14620462e-01
-8.72940600e-01 6.23709917e-01 3.77710819e-01 -1.81879714e-01
1.20713592e+00 -5.48398733e-01 -4.25326526e-01 -1.43968061e-01
-3.45822155e-01 4.77249697e-02 6.36815071e-01 5.37707388e-01
9.99311328e-01 -1.68910861e+00 -5.87269247e-01 3.21743429e-01
3.02604973e-01 4.88161385e-01 5.38047124e-03 8.52394402e-01
-1.00575888e+00 4.27381665e-01 2.06259713e-02 -1.16274083e+00
-1.25966942e+00 2.50652373e-01 3.65220338e-01 1.73321143e-01
-7.38149405e-01 1.13182890e+00 -1.52713910e-01 -4.41264689e-01
4.62638110e-01 -7.56902695e-01 1.05299298e-02 -1.41229611e-02
3.86870772e-01 6.20015979e-01 2.34067336e-01 -1.00669444e+00
-6.27840519e-01 1.34697998e+00 7.67998323e-02 1.34980053e-01
1.34434009e+00 9.01665390e-02 -1.75880507e-01 2.01817885e-01
1.35248470e+00 1.58910841e-01 -1.61567426e+00 -1.84488565e-01
-3.07270378e-01 -9.49970007e-01 8.22087601e-02 -4.86253887e-01
-1.04972577e+00 5.97773910e-01 9.04858112e-01 9.80356261e-02
8.54190528e-01 2.52948940e-01 7.09415674e-01 1.04560040e-01
6.05933428e-01 -8.33121419e-01 7.11080953e-02 3.91208410e-01
1.01726186e+00 -1.35680485e+00 3.92105013e-01 -4.67427850e-01
-1.42273828e-01 1.32666659e+00 6.68392897e-01 -3.28246504e-01
5.10722995e-01 2.88367897e-01 1.51627362e-01 -3.63866001e-01
-1.22943990e-01 1.04124233e-01 5.77111423e-01 6.25282943e-01
1.09653875e-01 1.00200817e-01 -2.56611519e-02 -6.58122376e-02
-3.07438642e-01 -2.22258553e-01 1.40434980e-01 1.02401519e+00
-5.01417994e-01 -7.82849252e-01 -8.78556430e-01 1.06093213e-01
-1.83721408e-01 1.32667013e-02 -8.52293819e-02 1.14263439e+00
1.63936809e-01 6.62896037e-01 4.56125379e-01 -1.62382305e-01
5.14352620e-01 -5.89482605e-01 6.50070369e-01 -1.70684472e-01
-1.85184300e-01 5.48549518e-02 -3.35466236e-01 -1.10216784e+00
-5.08852005e-01 -6.12293661e-01 -1.43524587e+00 -2.52233267e-01
-6.68606281e-01 -1.99697420e-01 9.42677677e-01 4.47595745e-01
2.94310153e-01 3.75827514e-02 7.00912476e-01 -1.36378789e+00
-2.17595026e-01 -5.75542510e-01 -3.75265002e-01 2.67750949e-01
4.87154305e-01 -9.00635302e-01 -5.50987184e-01 -3.69523346e-01] | [7.571232795715332, -2.6258087158203125] |
2174a87d-a585-49ba-a052-239ade9914aa | a-bayesian-approach-to-robust-reinforcement | 1905.08188 | null | https://arxiv.org/abs/1905.08188v2 | https://arxiv.org/pdf/1905.08188v2.pdf | A Bayesian Approach to Robust Reinforcement Learning | Robust Markov Decision Processes (RMDPs) intend to ensure robustness with respect to changing or adversarial system behavior. In this framework, transitions are modeled as arbitrary elements of a known and properly structured uncertainty set and a robust optimal policy can be derived under the worst-case scenario. In this study, we address the issue of learning in RMDPs using a Bayesian approach. We introduce the Uncertainty Robust Bellman Equation (URBE) which encourages safe exploration for adapting the uncertainty set to new observations while preserving robustness. We propose a URBE-based algorithm, DQN-URBE, that scales this method to higher dimensional domains. Our experiments show that the derived URBE-based strategy leads to a better trade-off between less conservative solutions and robustness in the presence of model misspecification. In addition, we show that the DQN-URBE algorithm can adapt significantly faster to changing dynamics online compared to existing robust techniques with fixed uncertainty sets. | ['Daniel Mankowitz', 'Esther Derman', 'Timothy Mann', 'Shie Mannor'] | 2019-05-20 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 2.17694789e-02 4.34876174e-01 -4.11363840e-02 1.24016376e-02
-1.11523151e+00 -6.66109204e-01 6.50183618e-01 1.35266662e-01
-5.80236018e-01 1.16964614e+00 -7.06203505e-02 -2.96465814e-01
-5.93301892e-01 -6.70022130e-01 -7.98743784e-01 -8.88151586e-01
-3.96037489e-01 5.63073456e-01 5.45352578e-01 1.36140838e-01
1.84164822e-01 7.21987367e-01 -1.07766032e+00 -5.49203575e-01
6.73183143e-01 9.42713439e-01 -1.25468299e-01 8.33353043e-01
5.30073941e-01 6.20979726e-01 -5.65087736e-01 9.72949043e-02
6.47563636e-01 2.87056807e-02 -5.45580149e-01 -5.31897768e-02
-3.16721857e-01 -4.59420681e-01 -3.26859087e-01 1.13683057e+00
5.44833958e-01 7.44983435e-01 8.01891923e-01 -1.30791867e+00
6.49189204e-02 7.36543179e-01 -4.60006148e-01 1.77966818e-01
2.49873891e-01 3.28751564e-01 4.17306483e-01 -1.26334354e-01
4.90845978e-01 1.60679257e+00 3.54479343e-01 7.19609261e-01
-1.38764477e+00 -3.92957538e-01 5.67469716e-01 -1.09627187e-01
-1.29235685e+00 -2.61556953e-01 3.28585535e-01 -2.80956775e-01
5.16786218e-01 1.57274544e-01 1.22185446e-01 1.19490981e+00
7.38860905e-01 5.91030240e-01 1.27282822e+00 -1.51750267e-01
9.72106278e-01 4.75371219e-02 -1.23836264e-01 2.87725717e-01
5.43769419e-01 6.26608312e-01 1.05278874e-02 -3.85778874e-01
7.15833902e-01 -1.02864943e-01 -2.35437021e-01 -6.15359902e-01
-1.07562208e+00 5.81867695e-01 -4.80934307e-02 -8.50958154e-02
-5.32389402e-01 3.26557010e-01 2.16873109e-01 5.64596832e-01
1.13415048e-01 4.09540713e-01 -3.97374272e-01 -1.92792252e-01
-4.49613929e-01 5.41842520e-01 1.05096138e+00 8.92537832e-01
1.82525873e-01 3.75905812e-01 -4.80397046e-01 6.71707615e-02
5.80805421e-01 8.55898678e-01 1.32330760e-01 -1.27149129e+00
4.73018020e-01 -2.10270539e-01 8.66584122e-01 -6.05668247e-01
-3.36244643e-01 -3.55510741e-01 -5.85005224e-01 5.93204856e-01
4.76250291e-01 -6.90421164e-01 -9.29432154e-01 1.98698115e+00
6.68735504e-01 2.65516281e-01 4.06473100e-01 4.30413425e-01
-4.02588099e-01 7.99785376e-01 -2.31161207e-01 -7.27859080e-01
5.91226518e-01 -1.57554656e-01 -8.92998099e-01 6.13129288e-02
4.08373736e-02 -1.81769729e-01 6.01305604e-01 5.93858421e-01
-9.52942550e-01 -4.97090518e-02 -1.06854594e+00 1.00064290e+00
-5.59900105e-02 -8.62184823e-01 -2.61166006e-01 7.73958266e-01
-8.43065202e-01 7.83591568e-01 -1.10012341e+00 -1.04135305e-01
4.32456136e-02 4.47147042e-01 -3.42016704e-02 2.26903602e-01
-1.21919596e+00 1.23076820e+00 7.54993975e-01 6.83554634e-02
-1.60482264e+00 -3.87798250e-01 -6.98388994e-01 -1.74766853e-01
1.13264287e+00 -3.75062376e-01 1.58154690e+00 -3.27101618e-01
-2.12412167e+00 -1.03082091e-01 1.81120858e-01 -9.04843092e-01
1.04591441e+00 -2.33664483e-01 -3.41182858e-01 1.96364015e-01
-2.95050383e-01 6.73055276e-03 1.27904558e+00 -1.38377130e+00
-5.34779549e-01 -1.87765330e-01 1.63287833e-01 1.39087275e-01
1.31116435e-01 -1.15939148e-01 8.15917775e-02 -5.36303401e-01
-8.40240419e-02 -1.19381702e+00 -7.49900222e-01 -3.44126284e-01
-4.62981433e-01 6.43393099e-02 8.37024570e-01 -2.45302469e-01
1.23396218e+00 -1.80257499e+00 3.99296105e-01 7.16843069e-01
-2.93826431e-01 2.19958052e-01 1.17599189e-01 5.76831937e-01
2.78644651e-01 1.72948554e-01 -4.21448559e-01 -1.82364345e-01
3.63664895e-01 5.33241689e-01 -6.94135904e-01 6.95780993e-01
4.86603640e-02 2.95557410e-01 -9.80673134e-01 -2.87775397e-01
2.00214699e-01 4.19293456e-02 -1.50646687e-01 2.72411704e-01
-3.79206121e-01 6.30789220e-01 -6.46831453e-01 6.08041465e-01
5.74257970e-01 4.49710995e-01 1.35061309e-01 4.95543957e-01
-4.79325205e-02 -5.55278897e-01 -1.90296364e+00 9.56700087e-01
-5.20171463e-01 1.50955737e-01 3.36950034e-01 -7.71507442e-01
7.71915734e-01 4.88811225e-01 3.60293031e-01 -7.76080117e-02
4.73501921e-01 2.77138986e-02 -2.73295287e-02 -5.94663387e-03
1.95279419e-01 -2.67834604e-01 -4.02273893e-01 5.46464860e-01
-1.35565132e-01 -4.84962106e-01 -1.70362461e-02 -1.13401935e-02
1.11730504e+00 2.69891024e-01 6.24749124e-01 -4.54383373e-01
5.31049728e-01 -5.76514959e-01 8.73764217e-01 1.14222455e+00
-5.86856186e-01 1.08760018e-02 6.41275883e-01 -7.94991180e-02
-6.39707863e-01 -1.10912347e+00 -2.65988857e-01 4.19942588e-01
1.81762040e-01 1.86773583e-01 -7.06979692e-01 -7.46862173e-01
1.09616689e-01 1.13301933e+00 -7.06801653e-01 -3.59167784e-01
-3.92015517e-01 -6.55754209e-01 2.47194082e-01 2.47128248e-01
3.02061498e-01 -5.78719079e-01 -9.50844705e-01 5.78926861e-01
4.25569355e-01 -8.26782703e-01 -3.15970391e-01 2.81487823e-01
-8.67569685e-01 -8.24266016e-01 -5.21697700e-01 2.34246165e-01
5.06528497e-01 -3.05290014e-01 4.30552274e-01 -9.28148031e-01
3.23544830e-01 7.85355806e-01 -4.50659543e-02 -6.95292711e-01
-8.49125624e-01 -3.03059757e-01 6.43201053e-01 2.34507978e-01
-2.86468834e-01 -4.81493771e-01 -1.02624886e-01 4.86220121e-01
-1.20419312e+00 -6.34080470e-01 2.60303944e-01 6.14996135e-01
7.71932781e-01 4.67709780e-01 6.45161211e-01 -6.72539175e-01
8.64991605e-01 -5.42063594e-01 -1.29026282e+00 3.41606975e-01
-8.66801858e-01 6.95302308e-01 6.37762368e-01 -7.43565679e-01
-1.26439822e+00 3.04187834e-02 3.42124373e-01 -5.27894735e-01
-1.11457981e-01 1.40848458e-01 -2.62926370e-01 -6.88763782e-02
4.83791381e-01 -3.09854764e-02 -6.12641312e-02 -2.79696763e-01
4.39524740e-01 5.34427762e-01 4.48295474e-01 -1.09434187e+00
1.14701390e+00 5.29741347e-01 4.98433322e-01 -4.37928796e-01
-5.86534917e-01 -8.10775533e-02 -5.57659447e-01 -5.06597579e-01
4.60498035e-01 -5.28039575e-01 -8.11679840e-01 3.44502658e-01
-6.30462468e-01 -4.82127845e-01 -5.82121313e-01 4.90890533e-01
-1.08431077e+00 4.19235826e-01 -4.31830853e-01 -1.65078342e+00
-3.50138173e-02 -1.10370302e+00 6.76057816e-01 3.42858076e-01
-1.01618990e-02 -1.00642097e+00 5.00306070e-01 -3.14116985e-01
3.04264754e-01 9.32915747e-01 4.64320362e-01 -7.70879865e-01
-5.05149961e-01 -3.41997564e-01 4.65426654e-01 2.93079048e-01
-8.14166963e-02 1.20948300e-01 -6.86523616e-01 -6.23362541e-01
2.85247684e-01 -1.08203486e-01 5.79868674e-01 3.60677242e-01
7.05699205e-01 -7.25250185e-01 -1.72941774e-01 2.19572097e-01
1.49134529e+00 6.02185428e-01 2.79309928e-01 6.16484106e-01
8.57787281e-02 5.08460224e-01 1.02306926e+00 8.98232877e-01
1.55419856e-01 4.57578719e-01 7.90736675e-01 6.21618092e-01
7.76258230e-01 -2.03129187e-01 7.90617466e-01 8.63277242e-02
6.21119700e-03 -3.20544779e-01 -9.49930787e-01 4.73662734e-01
-2.13747478e+00 -9.29181218e-01 4.43509579e-01 2.76211524e+00
7.91068256e-01 5.70116639e-01 2.00772420e-01 3.56350876e-02
7.73926437e-01 -1.49991289e-01 -1.02582943e+00 -6.42559409e-01
1.24162823e-01 -1.51594177e-01 9.72957075e-01 7.92026043e-01
-1.00502205e+00 4.56402212e-01 6.60689449e+00 6.03234231e-01
-7.10399270e-01 5.34525774e-02 3.58742356e-01 -1.13926880e-01
-2.61653401e-02 -3.18607800e-02 -7.64682055e-01 4.60516930e-01
1.42780030e+00 -5.11648357e-01 2.91904271e-01 8.27783585e-01
5.72270513e-01 -2.22529680e-01 -1.07423341e+00 4.28897351e-01
-4.39056396e-01 -7.33792484e-01 -3.87070060e-01 1.77953243e-01
9.83814120e-01 -2.08668172e-01 1.68648630e-01 4.21188265e-01
1.05610025e+00 -6.77240849e-01 8.11263323e-01 9.37366545e-01
3.26168388e-01 -1.23389494e+00 8.41415346e-01 5.45656443e-01
-8.48205566e-01 -5.24923623e-01 -1.13877855e-01 2.48718306e-01
4.13980037e-01 4.13086444e-01 -7.13420510e-01 6.57584012e-01
3.75966847e-01 2.06055045e-01 1.29412815e-01 1.01212084e+00
-3.19727719e-01 4.50925171e-01 -6.31033719e-01 6.14155196e-02
4.03059095e-01 -1.56600326e-01 1.22198617e+00 8.28323901e-01
3.61222446e-01 -8.59287530e-02 4.70005095e-01 5.36503077e-01
5.24710774e-01 -3.25767845e-01 -6.02824926e-01 2.09674910e-01
5.87125301e-01 6.89896345e-01 -6.75806761e-01 -2.48251364e-01
2.30867863e-01 6.16006196e-01 -3.04036066e-02 4.84002501e-01
-8.20036411e-01 -2.59398073e-01 6.11388266e-01 -2.88399071e-01
2.75830895e-01 -2.93974847e-01 5.91845177e-02 -1.02153897e+00
-1.97191224e-01 -8.53852332e-01 5.77574551e-01 -1.81500331e-01
-1.12496006e+00 4.41384435e-01 5.62380195e-01 -1.26508629e+00
-6.83609247e-01 -2.68009871e-01 -3.92808437e-01 5.96382320e-01
-1.38324380e+00 -6.22972727e-01 5.47803462e-01 6.04599535e-01
4.08599406e-01 1.69237498e-02 6.22176349e-01 -5.07593453e-01
-9.76631939e-01 2.19830766e-01 8.16303790e-01 -5.22935450e-01
5.90754688e-01 -1.55235326e+00 3.50312243e-04 1.27626967e+00
-4.26117510e-01 4.65837866e-01 1.26449454e+00 -7.49368310e-01
-1.28251779e+00 -1.31819272e+00 -9.90558714e-02 -5.16363382e-01
9.13006902e-01 -1.09741904e-01 -7.70845890e-01 7.36302018e-01
-1.27998516e-01 -1.04298189e-01 2.12875366e-01 -3.63190919e-01
-2.34940529e-01 -1.94920570e-01 -1.63837981e+00 7.48378158e-01
5.78400433e-01 -1.25714943e-01 -7.46531606e-01 1.38664648e-01
7.90861607e-01 -3.97843122e-01 -1.14374638e+00 4.89480168e-01
3.23786199e-01 -4.82030928e-01 7.16224730e-01 -6.44390523e-01
-4.93382871e-01 -5.85042477e-01 -3.33021134e-01 -1.51629782e+00
-2.63848230e-02 -1.54860687e+00 -8.64149511e-01 1.08444488e+00
2.73565650e-01 -8.90301347e-01 1.57815471e-01 8.66891325e-01
1.93053365e-01 -5.24623930e-01 -1.51092374e+00 -1.44314182e+00
2.25634962e-01 -3.87359530e-01 5.62898457e-01 3.16025764e-01
1.81866214e-02 -4.17668581e-01 -6.15243375e-01 6.41649365e-01
1.05465198e+00 -4.85820062e-02 5.27915776e-01 -9.71541464e-01
-5.70747793e-01 -1.71067238e-01 -2.32442111e-01 -5.38430512e-01
2.33469516e-01 -1.48689216e-02 5.55959225e-01 -1.15859640e+00
-1.54421180e-01 -2.58200735e-01 -6.41380727e-01 1.25902236e-01
-2.21796483e-01 -6.30144596e-01 3.68113160e-01 5.39779365e-02
-8.07851374e-01 6.99250937e-01 9.44525898e-01 8.86262879e-02
-4.63710338e-01 6.57524049e-01 -3.21676254e-01 6.91471815e-01
1.01725805e+00 -6.59856737e-01 -6.44001544e-01 5.24758734e-02
-1.11293495e-01 4.66394544e-01 -1.72937326e-02 -1.04460979e+00
3.84135619e-02 -8.05716395e-01 -1.68148443e-01 -4.82949227e-01
1.77588791e-01 -1.08866036e+00 3.19825023e-01 8.65555048e-01
-4.47688520e-01 -1.26355350e-01 2.08043620e-01 1.34504128e+00
1.69616073e-01 -3.41878772e-01 1.03245962e+00 -4.28071097e-02
-5.26044548e-01 3.30081075e-01 -8.72532010e-01 -6.64600544e-03
1.57605135e+00 -1.88917797e-02 5.65601885e-02 -7.47663736e-01
-8.68102074e-01 5.82876444e-01 2.81940997e-01 1.72410518e-01
3.86332154e-01 -9.32855308e-01 -4.12564993e-01 -3.03270936e-01
-2.90268272e-01 3.06258183e-02 1.51855111e-01 5.50879061e-01
-3.03812660e-02 3.25387627e-01 -5.67648895e-02 -3.24838191e-01
-8.51207733e-01 8.58716428e-01 5.35383344e-01 -5.47131658e-01
-3.27494174e-01 4.17984962e-01 -3.80253911e-01 -3.02538723e-01
4.99599069e-01 -4.40246016e-01 -5.48600405e-02 -6.26790300e-02
6.63022280e-01 7.95745552e-01 -2.13108167e-01 -3.32217872e-01
-1.31053969e-01 2.61603594e-01 -1.19599197e-02 -7.29401886e-01
1.05259621e+00 -4.97541040e-01 3.15741420e-01 6.78612769e-01
4.38103974e-01 -1.23579897e-01 -2.02559090e+00 -3.95089656e-01
4.00091469e-01 -4.37148333e-01 1.10615931e-01 -6.38411283e-01
-5.04387975e-01 3.60175103e-01 7.55099416e-01 2.42320940e-01
8.70819569e-01 -4.13171887e-01 3.09976250e-01 6.93727851e-01
7.03690290e-01 -1.34887218e+00 -2.12037116e-01 5.22206903e-01
7.88782477e-01 -9.07494724e-01 -6.31804317e-02 6.91319332e-02
-7.54782736e-01 1.05905116e+00 4.90429431e-01 -1.71782494e-01
7.26390481e-01 4.84828800e-01 -2.87551880e-01 5.48506498e-01
-9.43513215e-01 -2.59196460e-01 -9.94723216e-02 9.65050399e-01
-7.20829487e-01 2.52243280e-01 -4.73095328e-02 3.62449199e-01
3.29912841e-01 -2.38446966e-01 8.13083649e-01 1.30955935e+00
-7.37706482e-01 -1.00126445e+00 -9.06931818e-01 8.67739692e-02
-4.35025126e-01 4.82159287e-01 5.60204089e-02 7.87189364e-01
-3.62526804e-01 1.07751715e+00 -2.92214125e-01 2.81255152e-02
4.59061176e-01 -5.31832781e-03 3.41855943e-01 -3.85788590e-01
-9.97051373e-02 1.76246047e-01 4.69807945e-02 -8.02364349e-01
-1.84696004e-01 -9.40000534e-01 -1.12152576e+00 -1.76583961e-01
-2.40615487e-01 1.78628907e-01 5.01813173e-01 1.06229734e+00
1.76059037e-01 4.90135401e-01 8.82105291e-01 -6.21108830e-01
-1.67549098e+00 -7.34413922e-01 -7.35937953e-01 -2.07903311e-01
5.38168728e-01 -7.22269237e-01 -4.84664083e-01 -3.31627160e-01] | [4.620416641235352, 2.394618511199951] |
74702137-3fc4-44ab-8d63-56a58ad67d94 | melbourne-at-semeval-2016-task-11-classifying | null | null | https://aclanthology.org/S16-1150 | https://aclanthology.org/S16-1150.pdf | Melbourne at SemEval 2016 Task 11: Classifying Type-level Word Complexity using Random Forests with Corpus and Word List Features | null | ['ra', 'Timothy Baldwin', 'Julian Brooke', 'Alex Uitdenbogerd'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['complex-word-identification'] | ['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] | [-7.3356804847717285, 3.5935192108154297] |
e66bdf4d-1e85-4049-b5c2-0d90b9f2bdaa | unsupervised-syntactically-controlled | 2211.00881 | null | https://arxiv.org/abs/2211.00881v1 | https://arxiv.org/pdf/2211.00881v1.pdf | Unsupervised Syntactically Controlled Paraphrase Generation with Abstract Meaning Representations | Syntactically controlled paraphrase generation has become an emerging research direction in recent years. Most existing approaches require annotated paraphrase pairs for training and are thus costly to extend to new domains. Unsupervised approaches, on the other hand, do not need paraphrase pairs but suffer from relatively poor performance in terms of syntactic control and quality of generated paraphrases. In this paper, we demonstrate that leveraging Abstract Meaning Representations (AMR) can greatly improve the performance of unsupervised syntactically controlled paraphrase generation. Our proposed model, AMR-enhanced Paraphrase Generator (AMRPG), separately encodes the AMR graph and the constituency parse of the input sentence into two disentangled semantic and syntactic embeddings. A decoder is then learned to reconstruct the input sentence from the semantic and syntactic embeddings. Our experiments show that AMRPG generates more accurate syntactically controlled paraphrases, both quantitatively and qualitatively, compared to the existing unsupervised approaches. We also demonstrate that the paraphrases generated by AMRPG can be used for data augmentation to improve the robustness of NLP models. | ['Aram Galstyan', 'Kai-Wei Chang', 'Sriram Venkatapathy', 'Anoop Kumar', 'Varun Iyer', 'Kuan-Hao Huang'] | 2022-11-02 | null | null | null | null | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 3.64186823e-01 3.57589990e-01 -3.49805415e-01 -6.44233167e-01
-8.59923780e-01 -6.36116624e-01 5.37623167e-01 2.80767739e-01
-2.45769173e-02 6.95927322e-01 8.31270933e-01 -3.10641527e-01
2.19229326e-01 -9.95165706e-01 -9.12710309e-01 -2.83323467e-01
7.51513898e-01 3.85462314e-01 -1.97323784e-01 -2.54990637e-01
2.94140160e-01 1.34147719e-01 -1.16576099e+00 4.13685977e-01
1.24786043e+00 4.99086887e-01 2.28827208e-01 2.77317673e-01
-4.82638448e-01 8.36204946e-01 -6.69826448e-01 -5.18240154e-01
4.12689932e-02 -8.02793145e-01 -9.17295992e-01 -1.60707131e-01
1.80219442e-01 7.79088661e-02 -4.74518448e-01 9.98272240e-01
2.64410973e-01 2.44052778e-03 5.86343646e-01 -1.09926546e+00
-1.38683426e+00 9.80487049e-01 -2.94800699e-01 6.76236674e-02
7.04305470e-01 -8.12654849e-03 1.45925915e+00 -1.24043250e+00
6.05782509e-01 1.12802362e+00 4.40498382e-01 7.56066442e-01
-1.52876937e+00 -7.21677601e-01 2.12956872e-02 1.92144409e-01
-9.23624396e-01 -2.42125660e-01 1.22630739e+00 -1.30788177e-01
1.20667195e+00 5.40520959e-02 5.41786075e-01 1.40318024e+00
1.19953062e-02 7.57260621e-01 1.04368424e+00 -6.87839687e-01
4.50670063e-01 1.34554222e-01 3.44576865e-01 5.57957709e-01
4.37282026e-01 -1.12993672e-01 -5.82383275e-01 -1.14006817e-01
6.26593292e-01 8.78685992e-03 -2.67305404e-01 -4.21313584e-01
-1.05525362e+00 1.22990322e+00 5.69750905e-01 3.08290392e-01
-1.66638747e-01 3.86690982e-02 4.53787386e-01 5.94060302e-01
4.53582704e-01 1.12759256e+00 -3.30858022e-01 1.24808028e-01
-9.07814145e-01 2.61078626e-01 8.68948996e-01 1.18043888e+00
7.65767157e-01 -7.50114769e-02 -2.55837232e-01 1.17034078e+00
1.95780575e-01 5.40832937e-01 8.96188557e-01 -7.99812734e-01
9.09044862e-01 9.29912984e-01 -5.33242477e-04 -9.62025821e-01
4.71104234e-02 -1.29906729e-01 -7.00785875e-01 -4.75105703e-01
6.82060048e-02 -6.82895258e-02 -7.21082449e-01 1.99448180e+00
-1.67989537e-01 -2.40175098e-01 4.52268034e-01 6.07512534e-01
8.32731187e-01 9.06411052e-01 1.05935536e-01 1.36846313e-02
1.13658667e+00 -1.20238686e+00 -5.90390027e-01 -7.75762081e-01
6.29227579e-01 -6.63206160e-01 1.50029087e+00 -2.36776784e-01
-1.15667129e+00 -5.37137747e-01 -1.06168520e+00 -2.44070709e-01
-6.30914271e-02 1.63618810e-02 5.00903964e-01 3.59856397e-01
-9.35211778e-01 4.63505179e-01 -5.09514034e-01 -3.78460109e-01
3.67658466e-01 1.90004200e-01 -3.63447577e-01 -3.43041003e-01
-1.41517186e+00 1.02289903e+00 3.45451385e-01 -2.01155856e-01
-4.42672193e-01 -6.98647439e-01 -1.48782539e+00 3.15720081e-01
3.96025330e-02 -9.51666176e-01 1.23734069e+00 -8.59356463e-01
-1.56029618e+00 8.94454420e-01 -4.46425736e-01 -5.59781194e-01
-8.62258673e-02 -1.14481978e-01 -9.13101900e-03 1.14623837e-01
3.87903064e-01 6.33109927e-01 6.40241206e-01 -9.32346761e-01
-2.30611097e-02 -1.55069157e-01 1.37699485e-01 3.61799717e-01
-4.26642060e-01 3.18623669e-02 -1.84161663e-02 -9.88860548e-01
1.43996447e-01 -7.86232889e-01 -3.37800026e-01 -3.94093066e-01
-4.76612478e-01 -2.47784719e-01 4.55389410e-01 -7.32579410e-01
1.09051716e+00 -1.89955592e+00 6.79583073e-01 -1.78698525e-01
-1.30473450e-01 2.63641357e-01 -4.78788555e-01 8.02170098e-01
-3.42382491e-01 2.68320888e-01 -4.41058218e-01 -5.08414209e-01
7.01518208e-02 5.01294792e-01 -6.90811217e-01 -1.07497670e-01
4.72302109e-01 1.34167397e+00 -1.15325844e+00 -3.34682375e-01
1.11177392e-01 -1.61406845e-01 -7.90982366e-01 5.33922791e-01
-3.24441850e-01 3.44205797e-01 -5.09633362e-01 2.38871187e-01
4.79656518e-01 -2.95419902e-01 1.63516521e-01 4.06652391e-02
3.84471834e-01 7.42619097e-01 -2.36419693e-01 2.05066562e+00
-1.09076774e+00 5.11614919e-01 -4.58427697e-01 -1.21445727e+00
1.17807889e+00 1.85341552e-01 -1.32434264e-01 -7.33967543e-01
-9.53401104e-02 3.31910223e-01 -2.49976650e-01 -4.00862962e-01
4.93717998e-01 -6.08491898e-01 -5.67662656e-01 6.94721401e-01
1.24786705e-01 -5.92311621e-01 1.65354997e-01 4.78455633e-01
1.21984208e+00 2.34402306e-02 5.11979461e-01 -5.71015254e-02
6.04221523e-01 2.01940522e-01 4.92143929e-01 7.95756996e-01
1.89659983e-01 8.18534255e-01 5.47636926e-01 -8.56399313e-02
-1.47746730e+00 -1.30160809e+00 2.66237050e-01 6.60463095e-01
1.20169766e-01 -4.24488604e-01 -7.46325612e-01 -9.61173952e-01
5.90157742e-03 1.24803638e+00 -3.40889722e-01 -5.78237057e-01
-8.20424378e-01 -4.02872354e-01 5.53735137e-01 8.85331035e-01
3.44617426e-01 -1.43058991e+00 4.16011810e-02 1.57509819e-01
-4.39736545e-01 -1.23767877e+00 -5.12092769e-01 -9.70205590e-02
-9.72520769e-01 -8.09297502e-01 -7.62860000e-01 -1.23625362e+00
9.95670378e-01 3.07385147e-01 1.11948252e+00 -1.88471764e-01
1.05299197e-01 1.00481391e-01 -6.97159767e-01 -1.11873366e-01
-1.05312455e+00 1.31851196e-01 -1.04579993e-01 -3.02468330e-01
5.33361614e-01 -8.76213491e-01 -5.09752989e-01 6.51009455e-02
-9.34619486e-01 3.79067808e-01 6.89139903e-01 1.03040111e+00
4.28282976e-01 -5.90012968e-01 1.19555032e+00 -1.06720495e+00
1.06804419e+00 -5.36530316e-01 -1.87867999e-01 2.23109037e-01
-5.26789248e-01 4.46105480e-01 1.34212756e+00 -1.31455705e-01
-1.14605987e+00 -1.81261435e-01 -2.79989332e-01 -3.29563439e-01
-1.24604866e-01 4.96484786e-01 -2.55796254e-01 3.48596543e-01
7.60738492e-01 6.08264625e-01 8.73033926e-02 -5.02251089e-01
7.46386409e-01 8.68671238e-01 4.32108849e-01 -6.13425791e-01
1.00753093e+00 1.67335302e-01 -2.60530621e-01 -5.94472945e-01
-1.28474522e+00 -2.90701717e-01 -3.45911950e-01 4.50979501e-01
6.01555228e-01 -1.03583372e+00 3.19412023e-01 1.44128308e-01
-1.41175222e+00 -7.13611618e-02 -6.98738992e-01 1.38198778e-01
-6.39164090e-01 7.72574067e-01 -6.54840410e-01 -1.75172910e-01
-6.77878439e-01 -9.15825963e-01 1.07133102e+00 1.54737383e-01
-5.52471697e-01 -9.40623581e-01 2.91610092e-01 6.52047217e-01
3.28080803e-01 2.41699796e-02 1.42181492e+00 -8.62026811e-01
-4.09951419e-01 -3.97046417e-01 -3.73491824e-01 7.30528295e-01
3.79446208e-01 -6.02579594e-01 -6.77509367e-01 -1.06578715e-01
2.39791051e-01 -5.35156369e-01 6.68982446e-01 5.93548222e-03
9.19289827e-01 -6.44496560e-01 -1.27321199e-01 4.42308933e-01
1.18110907e+00 -2.80936688e-01 5.82706988e-01 1.43366858e-01
6.85031772e-01 6.22877657e-01 5.61325073e-01 -4.49046008e-02
2.31637210e-01 5.84875822e-01 -1.11610971e-01 1.85035050e-01
-3.88555348e-01 -9.71171916e-01 5.40761590e-01 1.27622259e+00
5.09917676e-01 -8.73580948e-02 -5.78516722e-01 7.31135130e-01
-1.81056285e+00 -8.09579790e-01 1.16385065e-01 1.98074090e+00
1.21553195e+00 -4.31228355e-02 -3.20390046e-01 9.56725106e-02
7.30206788e-01 3.49683464e-01 -4.70503598e-01 -7.69948959e-01
9.13599040e-04 7.54736483e-01 4.17938791e-02 4.52141225e-01
-5.25909364e-01 1.20292485e+00 5.49004507e+00 5.46507776e-01
-4.83221889e-01 1.80329517e-01 2.13388696e-01 1.22599788e-01
-7.94643641e-01 3.26041013e-01 -5.26466072e-01 5.61244845e-01
8.91063094e-01 -5.89251637e-01 3.51178348e-01 1.03161490e+00
3.51837099e-01 2.76576340e-01 -1.44018376e+00 8.03309381e-01
3.10389221e-01 -1.40499961e+00 6.34998739e-01 -3.78841162e-01
8.29793930e-01 -2.51089483e-01 -2.14140654e-01 5.30000389e-01
3.01256686e-01 -9.14807200e-01 4.16760206e-01 2.23659463e-02
7.24644780e-01 -6.59537435e-01 7.92645991e-01 3.81874770e-01
-9.10799742e-01 -7.05003962e-02 -7.37230062e-01 -2.59759247e-01
3.76858920e-01 5.00292957e-01 -7.54324555e-01 6.50352299e-01
2.69134827e-02 7.87147164e-01 -6.94926083e-01 6.37629330e-01
-1.08628094e+00 6.95833206e-01 3.75437476e-02 -3.58250767e-01
1.00174241e-01 -2.10442677e-01 6.39152586e-01 1.11712813e+00
2.40049958e-01 -2.33136699e-01 -1.09545976e-01 1.30737829e+00
-3.36178660e-01 2.41921812e-01 -9.33859706e-01 -3.33591968e-01
6.23374939e-01 8.99769247e-01 -1.50829986e-01 -3.86074752e-01
-4.99949008e-01 1.32469535e+00 6.28893495e-01 3.66728812e-01
-7.13660836e-01 -6.57750130e-01 5.34864128e-01 -1.31875739e-01
1.05197214e-01 -1.61650985e-01 -4.48973984e-01 -1.83808935e+00
2.93596476e-01 -8.20946038e-01 9.35568959e-02 -9.54401672e-01
-1.60226691e+00 4.78716820e-01 -1.95762254e-02 -9.74092960e-01
-4.51190233e-01 -5.02629638e-01 -9.18255925e-01 1.06352890e+00
-1.48097575e+00 -1.18494284e+00 -2.98616290e-01 4.78652157e-02
1.08540761e+00 -1.28675267e-01 1.08160412e+00 -1.83867276e-01
-5.93094468e-01 7.47163236e-01 -6.18664846e-02 2.58361161e-01
3.66910130e-01 -1.25760877e+00 9.76079047e-01 9.18466926e-01
3.91489655e-01 9.12197709e-01 6.76078200e-01 -5.78939259e-01
-1.19253600e+00 -1.29996347e+00 1.36591756e+00 -4.73664492e-01
7.43137419e-01 -6.51677668e-01 -9.72451448e-01 6.25152171e-01
3.48518312e-01 -5.05091310e-01 7.64955759e-01 -3.45359072e-02
-6.58173144e-01 1.52761474e-01 -1.01377261e+00 9.50818360e-01
1.21437824e+00 -7.38803089e-01 -1.49890125e+00 4.75807905e-01
1.26020718e+00 -6.04530014e-02 -5.30080080e-01 2.63650984e-01
1.02510042e-01 -7.58738458e-01 8.79226267e-01 -9.86797273e-01
1.08627558e+00 1.31980032e-01 -8.45465586e-02 -1.52938128e+00
-2.68250316e-01 -4.91014361e-01 -4.57463376e-02 1.30357742e+00
6.88543797e-01 -7.02859938e-01 9.94701087e-01 5.40693760e-01
-1.41364142e-01 -7.31515765e-01 -7.42023110e-01 -9.26579475e-01
3.93017411e-01 -1.14792936e-01 4.72222030e-01 7.15080976e-01
6.29215777e-01 1.02656770e+00 -8.60592723e-02 -1.00306772e-01
5.16667128e-01 6.42311454e-01 8.23360980e-01 -7.02786505e-01
-5.80681741e-01 -1.99244469e-01 -2.17149839e-01 -1.24187434e+00
6.48757160e-01 -1.52009547e+00 -6.88567683e-02 -1.91421473e+00
5.12030303e-01 -1.82458729e-01 -2.83207949e-02 4.03605133e-01
-3.08410704e-01 -4.68102004e-03 2.42537767e-01 3.20931107e-01
-1.35447428e-01 8.58262718e-01 1.16758657e+00 -7.05151409e-02
-1.77455902e-01 -7.75527805e-02 -1.06250310e+00 6.82159781e-01
8.92147183e-01 -5.14626145e-01 -6.18233085e-01 -6.38755381e-01
2.84633785e-01 3.38482335e-02 3.03650409e-01 -7.20257044e-01
-8.45560580e-02 -1.72795370e-01 -7.01585133e-03 -1.08915091e-01
1.12107851e-01 -3.51164609e-01 -2.88820297e-01 2.44097948e-01
-7.82581389e-01 2.60577530e-01 4.49933559e-02 7.87917972e-01
-3.88206959e-01 -7.35880256e-01 7.24400282e-01 -2.94550687e-01
-1.76770300e-01 -1.78595483e-01 -2.02906549e-01 4.20261890e-01
6.78310931e-01 -1.43414482e-01 -2.46176332e-01 -4.40060645e-01
-4.94850516e-01 1.33538127e-01 7.06138253e-01 5.88952780e-01
9.09618795e-01 -1.35266137e+00 -5.85315287e-01 2.84892410e-01
3.14961970e-01 7.87560567e-02 2.17048712e-02 2.16503724e-01
-3.82230729e-01 4.30942297e-01 1.16631147e-02 -1.01304829e-01
-8.67163241e-01 6.87427402e-01 -1.16491832e-01 -5.80662191e-01
-7.68522739e-01 6.44979894e-01 3.46939176e-01 -5.83537042e-01
-8.81612301e-02 -4.87872213e-01 1.20618837e-02 -2.66356617e-01
2.47490749e-01 -1.55673310e-01 -2.04099983e-01 -3.72626871e-01
-2.77191531e-02 5.41955054e-01 -3.86022836e-01 -1.48655295e-01
1.37414026e+00 -3.50926742e-02 -8.86796266e-02 1.23906061e-01
1.39349365e+00 1.05165996e-01 -5.97375810e-01 -2.93130547e-01
1.13557816e-01 -3.78003955e-01 -5.15404046e-01 -6.44776464e-01
-6.39324009e-01 9.25554693e-01 -3.34608048e-01 7.32382163e-02
9.11091685e-01 2.68924296e-01 1.28018391e+00 4.74483758e-01
4.63427335e-01 -7.18092561e-01 4.15127903e-01 4.84024644e-01
1.11587322e+00 -9.48329091e-01 -2.94686049e-01 -7.25999296e-01
-7.84769118e-01 9.03603971e-01 4.99816597e-01 -6.19588077e-01
5.70012182e-02 -6.49114177e-02 -2.06707999e-01 2.40998482e-03
-6.39838696e-01 1.89546227e-01 9.94085371e-02 5.57164431e-01
3.14912707e-01 -2.03211799e-01 -5.88602424e-01 8.84374619e-01
-5.58890104e-01 2.84712072e-02 6.58608377e-01 9.48837280e-01
-3.43466848e-01 -1.57333302e+00 1.34285986e-01 3.89679104e-01
-1.33841202e-01 -4.24702495e-01 -7.26292968e-01 4.66249347e-01
-4.89727616e-01 8.74293864e-01 -2.02862114e-01 -2.33844414e-01
5.35442770e-01 3.52022260e-01 5.84315717e-01 -1.47371364e+00
-5.02040148e-01 -6.29308343e-01 2.87710279e-01 -3.26933295e-01
-8.71593431e-02 -2.29886517e-01 -1.24406087e+00 1.05483845e-01
-2.66575485e-01 5.23057818e-01 3.81023854e-01 8.86955559e-01
6.88853025e-01 1.66484132e-01 8.78043234e-01 -5.02070129e-01
-1.04534304e+00 -1.06183386e+00 -2.08655089e-01 9.18846309e-01
-1.24196179e-01 -2.83074409e-01 -5.82898200e-01 -1.78534817e-02] | [11.672795295715332, 9.288416862487793] |
616a2590-c748-4af6-804e-3e08419c14ff | fedos-using-open-set-learning-to-stabilize | 2208.11512 | null | https://arxiv.org/abs/2208.11512v2 | https://arxiv.org/pdf/2208.11512v2.pdf | FedOS: using open-set learning to stabilize training in federated learning | Federated Learning is a recent approach to train statistical models on distributed datasets without violating privacy constraints. The data locality principle is preserved by sharing the model instead of the data between clients and the server. This brings many advantages but also poses new challenges. In this report, we explore this new research area and perform several experiments to deepen our understanding of what these challenges are and how different problem settings affect the performance of the final model. Finally, we present a novel approach to one of these challenges and compare it to other methods found in literature. | ['Juan Segundo Argayo', 'Julian Neubert', 'Mohamad Mohamad'] | 2022-08-22 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [-1.81362942e-01 -1.41316056e-01 -3.24994028e-01 -6.55605137e-01
-7.64181614e-01 -6.94547236e-01 6.42632842e-01 -4.03596014e-02
-4.82696503e-01 8.91938746e-01 1.11194827e-01 -2.97815472e-01
-3.50155741e-01 -5.65752387e-01 -7.91751206e-01 -7.93855846e-01
-2.39133924e-01 3.83212179e-01 2.22718701e-01 2.54389107e-01
1.14267416e-01 5.09194851e-01 -1.62818944e+00 4.85455424e-01
4.83889282e-01 1.05547619e+00 -3.28144580e-01 2.43765235e-01
-1.76734746e-01 1.04428113e+00 -4.92062658e-01 -7.99777746e-01
8.44506621e-01 -1.20757848e-01 -1.00407767e+00 -4.95763451e-01
6.98157907e-01 -4.97426778e-01 -2.07106203e-01 1.07741952e+00
4.36622649e-01 1.03935376e-01 -1.55835431e-02 -1.72020066e+00
-4.86278355e-01 7.05004036e-01 -2.07411230e-01 7.30375247e-03
9.58662033e-02 -1.35706246e-01 9.76298869e-01 -5.50942361e-01
8.98395598e-01 8.26058507e-01 7.16094792e-01 4.53425825e-01
-1.05808866e+00 -6.98692381e-01 1.05696611e-01 4.81170744e-01
-1.45295095e+00 -7.25281715e-01 5.45545638e-01 -1.60746098e-01
5.78634739e-01 6.32222354e-01 -7.90294334e-02 1.12214828e+00
7.73477331e-02 6.89691544e-01 1.47185314e+00 -3.94716203e-01
5.79025805e-01 6.72095239e-01 3.77315909e-01 4.69069690e-01
4.69823718e-01 1.07558519e-01 -9.30214763e-01 -9.71314907e-01
1.47198722e-01 2.94036657e-01 -2.31541380e-01 -9.82886970e-01
-7.53562808e-01 8.81471395e-01 2.03182340e-01 4.00804013e-01
-1.82839364e-01 -1.19955085e-01 5.18740416e-01 6.16720796e-01
4.41858590e-01 5.37605360e-02 -7.38345325e-01 4.62635010e-02
-9.93691266e-01 1.46062732e-01 1.37907112e+00 7.79578745e-01
9.10418153e-01 -4.01345134e-01 4.26300317e-02 4.16157275e-01
2.31664807e-01 7.72879347e-02 3.32560599e-01 -9.21336949e-01
4.34779316e-01 2.98920423e-01 7.89082795e-02 -7.89956093e-01
-3.66130248e-02 -8.97380635e-02 -5.27022839e-01 2.96374112e-02
5.82129121e-01 -3.30261379e-01 -2.28642792e-01 1.61114347e+00
6.37053490e-01 4.02728587e-01 1.96442176e-02 9.03836787e-01
3.99904758e-01 2.08122104e-01 9.64090750e-02 3.48065794e-02
1.01975858e+00 -9.78990018e-01 -7.84590065e-01 9.60720927e-02
6.11698329e-01 -5.53256452e-01 6.56003356e-01 3.58924001e-01
-7.69339204e-01 7.65607506e-02 -7.49598742e-01 -6.14934787e-02
-6.26553357e-01 -4.20172572e-01 9.93930042e-01 9.57809567e-01
-1.18187308e+00 7.42738485e-01 -8.97441268e-01 -6.18825972e-01
6.05085850e-01 4.30705994e-01 -7.09954619e-01 -3.03407341e-01
-1.09402800e+00 9.94166315e-01 1.43594474e-01 -4.28086430e-01
-6.89909995e-01 -7.06414580e-01 -4.75826323e-01 1.56154588e-01
4.88156706e-01 -5.78601182e-01 1.21877694e+00 -5.66148043e-01
-1.23471189e+00 8.49859715e-01 -2.58985579e-01 -5.71614742e-01
6.99223280e-01 -1.37473077e-01 -3.01591992e-01 -3.04226011e-01
-3.02273184e-01 -1.21887282e-01 4.51676369e-01 -1.13572991e+00
-9.32610333e-01 -8.26065779e-01 -2.27146327e-01 -2.70659924e-01
-6.62719727e-01 5.19283414e-01 -2.99802244e-01 -2.00984359e-01
-2.02760950e-01 -7.97208190e-01 -3.44517708e-01 1.96148396e-01
-1.62731543e-01 -2.56431758e-01 1.08752191e+00 -4.23840553e-01
9.68289971e-01 -2.15729833e+00 -1.65291980e-01 4.61484492e-01
2.15743706e-01 1.53469220e-01 -4.14881967e-02 7.79674709e-01
3.22370380e-01 3.41083258e-01 -1.80352181e-01 -7.28701890e-01
2.34667227e-01 3.43565106e-01 -3.85354578e-01 7.48863280e-01
-3.10036957e-01 6.54340327e-01 -4.10342723e-01 -4.65729475e-01
-2.92429384e-02 5.38463652e-01 -4.81728077e-01 3.96838874e-01
-1.98243499e-01 4.43005264e-01 -5.27341247e-01 7.24182963e-01
1.00867975e+00 -8.83420482e-02 6.58655584e-01 1.50047868e-01
-2.06437692e-01 2.91931450e-01 -1.38344586e+00 1.51125991e+00
-2.11900070e-01 1.51012570e-01 5.86690903e-01 -1.01992476e+00
7.54684746e-01 4.60244924e-01 6.42991483e-01 -5.20522237e-01
1.28117148e-02 2.29498103e-01 -4.64257449e-01 -5.51162660e-01
5.69299571e-02 -1.11692481e-01 5.99767752e-02 9.61350858e-01
2.41517350e-02 6.50698125e-01 -4.49306577e-01 -1.27830267e-01
1.16812575e+00 -1.13539688e-01 2.69387990e-01 -3.37047815e-01
3.21042925e-01 -1.22402102e-01 8.27940047e-01 1.06050181e+00
-5.15474498e-01 8.07896852e-02 3.47861737e-01 -8.32330942e-01
-6.41498387e-01 -1.00137293e+00 -1.12725213e-01 1.51820385e+00
-1.18483484e-01 -6.09226286e-01 -6.42862678e-01 -1.16772246e+00
4.99666631e-01 6.61841154e-01 -5.06191075e-01 4.91416194e-02
-4.43861127e-01 -5.18554270e-01 5.55826962e-01 3.30223292e-01
2.52622604e-01 -7.61845827e-01 -4.57784474e-01 -2.38390937e-01
-2.30490848e-01 -1.09877539e+00 -2.24719644e-01 1.80064470e-01
-9.15398240e-01 -1.07056999e+00 6.17778115e-02 -5.33899903e-01
3.52940500e-01 3.52886289e-01 1.19790900e+00 1.05835453e-01
-2.43122235e-01 1.83529228e-01 -2.83910483e-01 -7.49637663e-01
-1.76449686e-01 2.46255994e-01 7.44108409e-02 3.05191100e-01
9.55976248e-01 -5.93494296e-01 -2.57591069e-01 2.77315557e-01
-1.14897132e+00 -5.54566085e-01 2.28927702e-01 5.44321120e-01
5.46658814e-01 -8.61686654e-04 3.91093582e-01 -1.36233628e+00
5.26965380e-01 -8.96648765e-01 -7.75443733e-01 5.02077103e-01
-9.08277512e-01 1.79143265e-01 8.26296031e-01 -1.18310213e-01
-9.98349726e-01 2.50733852e-01 1.92444339e-01 -5.56855857e-01
-3.56744498e-01 2.90506393e-01 -3.48092765e-01 -3.26179266e-01
4.44916129e-01 -7.44997803e-03 2.69618183e-01 -1.12357998e+00
4.27036345e-01 8.88590991e-01 1.60514385e-01 -7.22078621e-01
6.00198030e-01 9.39706564e-01 -1.41648203e-01 -3.51904571e-01
-7.42248833e-01 -3.97374392e-01 -5.83911777e-01 4.05129433e-01
3.12800169e-01 -6.48851097e-01 -8.13863456e-01 2.81322330e-01
-9.15676296e-01 -1.14432178e-01 -4.09930110e-01 4.78102833e-01
-4.60424781e-01 2.23810986e-01 -5.32503486e-01 -8.12783778e-01
-4.59535807e-01 -8.27088356e-01 5.44796646e-01 -2.03458946e-02
-4.11774591e-02 -1.00406492e+00 3.78616601e-01 5.64902663e-01
9.55832124e-01 2.14541420e-01 5.35799921e-01 -1.25515378e+00
-6.53988659e-01 -3.80483270e-01 -3.77627206e-03 2.35571623e-01
9.54214558e-02 -1.89087927e-01 -1.26237202e+00 -6.13566518e-01
3.77153277e-01 -2.79275745e-01 5.37591994e-01 -2.03833908e-01
1.42281055e+00 -8.07297945e-01 -2.82412261e-01 7.92826056e-01
1.75405777e+00 -2.98885077e-01 4.26866472e-01 4.10916299e-01
4.10930157e-01 8.27761114e-01 3.80519837e-01 7.77179122e-01
4.62510794e-01 6.22480392e-01 4.28068548e-01 3.89903821e-02
3.09565693e-01 -2.39394426e-01 2.48861909e-01 3.28357965e-01
2.75190353e-01 4.70136404e-02 -8.76339376e-01 5.14480174e-01
-2.10919833e+00 -9.69151497e-01 8.21033716e-02 2.38354349e+00
8.14734876e-01 -6.68980300e-01 2.54107267e-01 -3.26375037e-01
6.01959229e-01 1.60386056e-01 -5.37949204e-01 -5.73147118e-01
-1.44136459e-01 3.53105813e-01 7.59269655e-01 1.81563243e-01
-1.03294969e+00 9.19320583e-01 7.48947573e+00 3.03690910e-01
-1.26260757e+00 5.37205458e-01 3.82968366e-01 -4.55732673e-01
-3.12789112e-01 2.65700489e-01 -6.49561405e-01 3.63067538e-01
1.27581275e+00 -3.31202090e-01 7.29673088e-01 1.15060568e+00
-1.09269701e-01 1.90136820e-01 -1.41982758e+00 7.33258247e-01
7.10793585e-02 -1.32452857e+00 -2.72868603e-01 2.52067924e-01
6.55018032e-01 7.73350477e-01 6.96105435e-02 3.05252597e-02
7.39052892e-01 -1.06009698e+00 4.04913425e-01 6.07408285e-01
2.89398015e-01 -7.56530046e-01 8.61281514e-01 6.07769787e-01
-6.16712093e-01 -2.42973566e-01 -4.12996173e-01 -2.06275940e-01
-3.10362935e-01 4.24676269e-01 -6.03089094e-01 7.47007847e-01
1.10938883e+00 2.29544744e-01 -5.40683866e-01 1.13937128e+00
1.34096652e-01 7.32356966e-01 -5.59970677e-01 1.46881133e-01
-4.71790247e-02 -7.86664411e-02 3.00621420e-01 1.05990458e+00
8.55509192e-02 -2.78206974e-01 1.32983014e-01 8.52892220e-01
-2.56322682e-01 4.08598274e-01 -9.27167714e-01 2.44174615e-01
7.54855752e-01 1.17544675e+00 8.36842656e-02 -4.03614677e-02
-8.11141491e-01 7.40904808e-01 7.43106961e-01 2.47459307e-01
-4.80431437e-01 -1.70132387e-02 1.10513854e+00 1.34102823e-02
4.32920575e-01 1.10877529e-02 -2.71579921e-01 -1.34472942e+00
3.26037675e-01 -1.20481837e+00 1.07302713e+00 -4.34569456e-02
-1.89429617e+00 4.01968479e-01 -1.50513276e-01 -8.12078595e-01
-9.55483913e-02 -2.01802567e-01 -6.03342295e-01 7.95784593e-01
-1.70874524e+00 -1.22910392e+00 3.51036079e-02 1.00706673e+00
-3.05027872e-01 -1.98965877e-01 1.34721839e+00 3.46911401e-01
-7.18419313e-01 9.71146941e-01 7.43212461e-01 2.65012626e-02
1.16367388e+00 -8.28751326e-01 2.04280481e-01 7.68504500e-01
3.16297233e-01 9.73090231e-01 4.90433425e-01 -3.42656672e-01
-1.75726295e+00 -1.10727775e+00 1.14896750e+00 -7.07515538e-01
5.35053790e-01 -5.92109203e-01 -9.91795003e-01 1.02740800e+00
3.43105823e-01 5.79043627e-01 1.30716276e+00 5.71116567e-01
-8.41859996e-01 -4.84868616e-01 -1.83782625e+00 1.00878261e-01
7.35008836e-01 -4.96875852e-01 -1.13141082e-01 3.48134428e-01
6.91877782e-01 -6.52446598e-02 -1.04798079e+00 1.39635891e-01
5.45892775e-01 -1.36535501e+00 6.24438047e-01 -1.22287691e+00
-3.07545871e-01 -7.57066309e-02 -5.59051156e-01 -9.09370065e-01
-9.13814157e-02 -8.98848534e-01 -4.85094249e-01 1.40852153e+00
1.59588441e-01 -1.16113353e+00 1.08269501e+00 1.27392960e+00
5.06530166e-01 -4.13538277e-01 -1.27301180e+00 -9.44586337e-01
3.89183700e-01 -3.69126707e-01 1.32727790e+00 1.40391934e+00
2.79632080e-02 -4.16304022e-01 -5.60421288e-01 2.94400334e-01
9.48919892e-01 3.91474962e-01 1.01014590e+00 -1.27666664e+00
-1.16243549e-01 9.25019383e-02 -3.21851581e-01 -2.44302601e-01
3.24931115e-01 -1.03659093e+00 -5.15030563e-01 -1.01024878e+00
2.76131302e-01 -6.32460475e-01 -7.93165624e-01 9.49117839e-01
1.97862774e-01 -9.05273855e-02 3.90366286e-01 4.80125934e-01
-8.22715402e-01 1.80810228e-01 4.05946851e-01 2.53535956e-01
1.00341663e-01 1.93503082e-01 -1.01587510e+00 3.58074784e-01
9.44688559e-01 -8.48904848e-01 -1.85727268e-01 -5.78548312e-01
-1.82460144e-01 -2.37324163e-01 3.45253468e-01 -6.34800136e-01
6.90133750e-01 -5.66960871e-01 5.19056320e-02 -2.76705146e-01
-4.19325083e-02 -1.38462102e+00 3.53912413e-01 3.23676318e-01
-3.26089591e-01 3.35075185e-02 -1.24623239e-01 5.88765204e-01
-1.62056953e-01 3.82545841e-04 7.84143031e-01 -1.42325044e-01
-4.82987761e-01 3.40805471e-01 8.65056291e-02 -7.89825395e-02
1.24425972e+00 2.46037975e-01 -4.45724726e-01 -3.38893890e-01
-5.31287968e-01 2.72983938e-01 8.04138005e-01 5.42021990e-01
1.63498640e-01 -1.11810291e+00 -8.17877412e-01 5.23799717e-01
6.14763349e-02 -3.55474353e-01 1.16832510e-01 8.08649957e-01
-8.82440060e-02 5.65308511e-01 -1.35748759e-01 -2.46573269e-01
-1.42418170e+00 7.79750764e-01 2.36567512e-01 -1.41394094e-01
-4.71841961e-01 5.72061002e-01 -4.04833823e-01 -9.37611878e-01
7.24772692e-01 2.86847711e-01 3.66313636e-01 -2.59686232e-01
6.50659502e-01 5.03998339e-01 3.15254420e-01 -4.06512618e-01
-7.28973687e-01 -9.49229971e-02 -2.43216887e-01 2.32529446e-01
1.64049375e+00 -1.35563537e-01 -6.34732962e-01 2.98670650e-01
1.47280490e+00 1.72233909e-01 -1.01694810e+00 -7.56588399e-01
3.47527385e-01 -8.23058367e-01 -3.66535061e-03 -9.03016150e-01
-1.24287784e+00 4.63892937e-01 7.12355852e-01 1.21300004e-01
9.40921485e-01 9.83360708e-02 7.50122190e-01 4.31076258e-01
7.21833944e-01 -1.06824613e+00 -6.90422237e-01 3.16403240e-01
3.93041074e-01 -1.39891517e+00 8.22098702e-02 -2.05645308e-01
-6.29590631e-01 6.85977757e-01 4.48748142e-01 -6.72127232e-02
9.36368465e-01 5.66103578e-01 3.24374348e-01 -1.17459118e-01
-1.06411350e+00 2.38933384e-01 -1.49944901e-01 7.04198122e-01
3.13852668e-01 9.69743542e-03 -3.53550315e-01 8.64082158e-01
-1.39551565e-01 3.28060329e-01 1.43762559e-01 1.29256701e+00
1.32524937e-01 -1.59357131e+00 -5.06099164e-01 3.25502664e-01
-7.78837323e-01 2.03959152e-01 -8.69217217e-01 5.36334395e-01
8.14363286e-02 9.77566838e-01 -1.42808542e-01 -4.60484356e-01
1.76123068e-01 4.96616781e-01 1.49962371e-02 -2.98413962e-01
-1.00789714e+00 -4.21019971e-01 -1.03859842e-01 -9.62484181e-01
-2.06707343e-01 -8.07539225e-01 -8.39967191e-01 -7.35999584e-01
-3.58416706e-01 5.24590254e-01 1.07309043e+00 6.21103704e-01
9.05177832e-01 -2.90358335e-01 1.04545164e+00 -8.50065127e-02
-1.37081182e+00 -5.08639932e-01 -6.05155170e-01 4.79944706e-01
2.23528713e-01 -1.72723502e-01 -5.25462806e-01 -2.07015544e-01] | [5.912691593170166, 6.535826206207275] |
75f3ffc5-52bf-434e-871b-e1e9bcab79c6 | promises-and-challenges-of-causality-for | 2201.10683 | null | https://arxiv.org/abs/2201.10683v2 | https://arxiv.org/pdf/2201.10683v2.pdf | Promises and Challenges of Causality for Ethical Machine Learning | In recent years, there has been increasing interest in causal reasoning for designing fair decision-making systems due to its compatibility with legal frameworks, interpretability for human stakeholders, and robustness to spurious correlations inherent in observational data, among other factors. The recent attention to causal fairness, however, has been accompanied with great skepticism due to practical and epistemological challenges with applying current causal fairness approaches in the literature. Motivated by the long-standing empirical work on causality in econometrics, social sciences, and biomedical sciences, in this paper we lay out the conditions for appropriate application of causal fairness under the "potential outcomes framework." We highlight key aspects of causal inference that are often ignored in the causal fairness literature. In particular, we discuss the importance of specifying the nature and timing of interventions on social categories such as race or gender. Precisely, instead of postulating an intervention on immutable attributes, we propose a shift in focus to their perceptions and discuss the implications for fairness evaluation. We argue that such conceptualization of the intervention is key in evaluating the validity of causal assumptions and conducting sound causal analysis including avoiding post-treatment bias. Subsequently, we illustrate how causality can address the limitations of existing fairness metrics, including those that depend upon statistical correlations. Specifically, we introduce causal variants of common statistical notions of fairness, and we make a novel observation that under the causal framework there is no fundamental disagreement between different notions of fairness. Finally, we conduct extensive experiments where we demonstrate our approach for evaluating and mitigating unfairness, specially when post-treatment variables are present. | ['Alice Xiang', 'Aida Rahmattalabi'] | 2022-01-26 | null | null | null | null | ['econometrics'] | ['miscellaneous'] | [ 4.39365953e-01 2.39964113e-01 -8.07208002e-01 -5.27956128e-01
-1.91723481e-01 -5.12209833e-01 6.84501529e-01 5.73484659e-01
-6.90775633e-01 8.74490440e-01 8.68768096e-01 -8.48601520e-01
-7.36188531e-01 -6.78494096e-01 -3.69480014e-01 -3.03634882e-01
1.43245012e-01 -7.67155811e-02 -5.39391935e-01 6.71245158e-02
6.39106095e-01 4.38989908e-01 -1.23532009e+00 -1.99616432e-01
1.13499117e+00 4.15222108e-01 -6.07752919e-01 2.09866494e-01
1.72457099e-01 8.55723262e-01 -3.47791284e-01 -9.12637889e-01
1.21419676e-01 -5.25387824e-01 -9.30131376e-01 -4.02968079e-01
2.85962671e-01 -6.71515286e-01 -9.54864770e-02 9.72949386e-01
4.89078015e-01 6.67178025e-03 7.58620918e-01 -1.59288526e+00
-9.60177183e-01 9.94226158e-01 -5.39501011e-01 1.27155095e-01
3.04402351e-01 2.28286237e-01 1.22527659e+00 -2.05100670e-01
5.08575618e-01 1.54589570e+00 3.91744614e-01 6.27242863e-01
-1.26743329e+00 -7.48984635e-01 3.62936825e-01 1.14101768e-02
-9.16368246e-01 -6.97053671e-01 2.43683875e-01 -7.54288316e-01
1.94042698e-01 5.94537258e-01 3.90178859e-01 1.03874588e+00
2.13963509e-01 9.73760486e-02 1.54066014e+00 -5.61020851e-01
3.04276019e-01 -2.21404538e-01 3.16037714e-01 1.35592997e-01
9.08212543e-01 6.69065714e-01 -2.97432542e-01 -5.31724989e-01
7.12768972e-01 -7.50287324e-02 -2.65668601e-01 -1.39220387e-01
-1.04473734e+00 1.04772770e+00 2.41558760e-01 1.79651365e-01
-4.10361916e-01 3.10860336e-01 5.11330605e-01 6.47272244e-02
3.88442665e-01 4.70264614e-01 -1.44308612e-01 -2.83747297e-02
-6.28994942e-01 5.53867698e-01 5.62896609e-01 2.76682466e-01
1.77347317e-01 -2.81952739e-01 -5.48494935e-01 3.41169745e-01
2.59078503e-01 4.88485068e-01 -1.70631975e-01 -1.43945611e+00
1.49062648e-01 2.58998066e-01 5.06617665e-01 -9.94099200e-01
-2.61575550e-01 -3.29795405e-02 -6.62340701e-01 3.22806001e-01
6.74832165e-01 -2.86944568e-01 -3.56863797e-01 2.15733886e+00
3.43280345e-01 -2.36997962e-01 -2.31417269e-01 9.00746226e-01
2.56147385e-01 -1.86749324e-01 9.41890180e-01 -5.33323646e-01
1.43767667e+00 4.77828197e-02 -8.24831128e-01 4.04333323e-02
4.38076049e-01 -6.45054638e-01 1.10466766e+00 -1.67593613e-01
-9.29013848e-01 4.40965258e-02 -4.13453311e-01 -1.11753576e-01
3.66270952e-02 -5.56882024e-01 8.87269795e-01 1.14515448e+00
-5.49012303e-01 7.00802863e-01 -2.36640975e-01 -5.41697025e-01
5.70259929e-01 7.05574527e-02 -1.29337505e-01 1.03291139e-01
-1.27378893e+00 1.11629438e+00 -5.66803515e-02 1.14341557e-01
-5.64808965e-01 -9.66515243e-01 -5.20902216e-01 3.10565084e-01
5.23086369e-01 -1.10566258e+00 1.01435244e+00 -1.11248624e+00
-8.95063877e-01 8.45102906e-01 -9.68362764e-02 -2.64775455e-01
8.99475098e-01 -1.01769418e-01 -4.84359950e-01 -2.28422329e-01
5.34502625e-01 2.55605072e-01 2.73492277e-01 -1.18056047e+00
-6.78184450e-01 -4.61743832e-01 4.51091975e-01 1.57489985e-01
-6.65798709e-02 7.35056698e-01 6.52299404e-01 -5.27226388e-01
-4.04495239e-01 -7.14031041e-01 -3.74734581e-01 -1.86632089e-02
-4.25346464e-01 -1.58191770e-01 2.41486832e-01 -2.23252684e-01
1.31682730e+00 -2.14743876e+00 -4.98771280e-01 1.60883620e-01
3.06967556e-01 -1.59043700e-01 1.12116568e-01 3.68001163e-01
-4.04984385e-01 7.97058284e-01 -9.68557745e-02 2.44606242e-01
3.65865827e-01 4.58441824e-02 -5.65536380e-01 7.95335293e-01
-2.10713129e-02 9.13666427e-01 -9.42698181e-01 -4.07225609e-01
2.64526248e-01 2.36127913e-01 -7.29558885e-01 -4.68245596e-02
3.69339049e-01 3.52365911e-01 -3.11001152e-01 3.88674736e-01
6.18027210e-01 -1.49709983e-02 5.72350204e-01 1.60001382e-01
-5.70832253e-01 5.76588511e-01 -9.00076032e-01 8.63892496e-01
-3.05051496e-03 2.18916222e-01 1.06728077e-01 -8.69277835e-01
2.19088078e-01 3.39758754e-01 3.31855714e-01 -6.77044153e-01
3.07118386e-01 1.24398261e-01 3.33817422e-01 -4.09115791e-01
4.44890171e-01 -8.25744212e-01 -1.67321533e-01 6.98221564e-01
-5.94241858e-01 6.53642192e-02 -5.35819791e-02 1.61282733e-01
6.62256122e-01 -9.26554129e-02 5.62392592e-01 -6.73813462e-01
9.28532928e-02 -6.28801584e-02 7.87378788e-01 8.36757898e-01
-6.58549547e-01 2.17296332e-01 8.82032931e-01 -2.50254095e-01
-8.64747524e-01 -9.67160940e-01 -4.58637267e-01 9.60369110e-01
1.93755373e-01 4.56274673e-02 -5.66573620e-01 -5.62614858e-01
4.07547891e-01 1.20177794e+00 -1.05686224e+00 -1.71462432e-01
-2.32432663e-01 -1.11774135e+00 6.52527690e-01 3.13028663e-01
3.90135460e-02 -6.52920783e-01 -1.00470924e+00 -1.29290417e-01
-2.12966800e-01 -6.54397786e-01 -3.07128966e-01 -2.92978764e-01
-7.52860665e-01 -1.40756404e+00 -3.17129523e-01 1.98776200e-01
4.68319535e-01 3.07414353e-01 9.02228951e-01 3.18574518e-01
-6.78471755e-03 5.06600142e-01 -1.26694381e-01 -6.89826965e-01
-4.59008187e-01 -5.48322082e-01 5.54078408e-02 -3.37851554e-01
3.76893044e-01 -3.84496808e-01 -8.04671764e-01 7.58267194e-02
-9.80583608e-01 -1.35155737e-01 2.14658916e-01 5.96602499e-01
-1.29613668e-01 -2.06483394e-01 9.10393953e-01 -1.28511643e+00
1.02485728e+00 -5.95140517e-01 -4.38588440e-01 2.37126783e-01
-1.15293133e+00 -2.53055841e-01 8.41726288e-02 -5.48016056e-02
-1.34998763e+00 -9.66778219e-01 3.29977065e-01 3.11330676e-01
-2.36477822e-01 5.45681417e-01 -1.35108843e-01 6.69365078e-02
7.58239985e-01 -8.57978284e-01 1.29725754e-01 -2.36639753e-02
5.72775006e-01 4.36638206e-01 1.31947100e-01 -1.14219189e+00
4.58362162e-01 7.00816751e-01 -7.23870993e-02 -2.20335409e-01
-7.97223508e-01 6.31891489e-02 -1.97948143e-01 -2.21477434e-01
7.53335238e-01 -6.56430840e-01 -1.39081240e+00 -1.92425530e-02
-9.20285642e-01 -3.46378386e-01 -4.67784464e-01 7.44109154e-01
-5.08726656e-01 4.30999368e-01 -5.58283567e-01 -1.20118546e+00
-5.39768972e-02 -9.83828008e-01 4.52472180e-01 1.47481598e-02
-7.12321460e-01 -9.78866458e-01 6.60047401e-04 5.39504170e-01
4.41202104e-01 4.32751805e-01 1.18450272e+00 -2.82863051e-01
-1.93098426e-01 4.49339077e-02 -4.64654714e-01 -1.90814480e-01
4.16948497e-01 3.95412922e-01 -9.20246363e-01 1.67087525e-01
-8.40653628e-02 -6.56089336e-02 4.84843761e-01 8.20433617e-01
8.89349699e-01 -4.93986487e-01 -1.54902652e-01 5.27543835e-02
1.42404044e+00 1.72300354e-01 5.24234474e-01 3.02394658e-01
4.54726100e-01 1.33359194e+00 5.22447646e-01 6.21041834e-01
7.69888520e-01 5.08319616e-01 4.36682016e-01 -2.84092396e-01
2.27487192e-01 -2.75313884e-01 -3.98876658e-03 -1.04141288e-01
-3.27308118e-01 4.02674228e-02 -8.17402244e-01 6.11073077e-01
-1.86290205e+00 -1.28152430e+00 -4.81694549e-01 2.72753406e+00
8.78311574e-01 -2.97211148e-02 4.38893884e-01 6.82161450e-02
9.02455330e-01 -9.82969813e-03 -8.25303644e-02 -8.61199737e-01
-8.41402933e-02 -1.03417160e-02 8.02786052e-01 8.01077604e-01
-8.28263581e-01 5.89736521e-01 7.30587387e+00 3.26127887e-01
-8.69236529e-01 -1.11771813e-02 1.11652923e+00 -8.29442516e-02
-9.85134363e-01 5.12739003e-01 -1.01360790e-01 4.11807835e-01
7.46512473e-01 -8.03636372e-01 1.61587536e-01 2.46255770e-01
9.75748420e-01 -4.55961585e-01 -1.20482683e+00 1.18475288e-01
-5.36705732e-01 -9.79349196e-01 -1.66455358e-01 2.65702873e-01
6.72752917e-01 -5.97311795e-01 1.50302216e-01 -1.21359609e-01
8.79847050e-01 -1.45354307e+00 1.02242804e+00 2.41200492e-01
8.38425636e-01 -6.25980258e-01 6.51242793e-01 -6.66731298e-02
-3.46326560e-01 -1.05184652e-01 -2.88173079e-01 -7.28916168e-01
1.38430774e-01 1.04624832e+00 -3.29698890e-01 5.29669702e-01
4.71212268e-01 2.08378106e-01 1.12906963e-01 7.68723369e-01
-3.78166944e-01 6.40242100e-01 1.50878027e-01 3.50165725e-01
2.24542525e-02 -1.32404476e-01 1.66948229e-01 9.42989290e-01
9.06229764e-03 4.20563430e-01 -4.23480093e-01 1.16335785e+00
-8.43959674e-03 2.22475842e-01 -5.53355813e-01 -7.45733827e-02
6.92686856e-01 8.36835146e-01 -6.23625457e-01 -1.32721841e-01
-5.59624910e-01 1.53476492e-01 7.02778399e-02 4.70700562e-01
-7.57151604e-01 3.25507343e-01 8.77807200e-01 2.09006876e-01
-5.94680369e-01 2.42124990e-01 -9.93980289e-01 -9.91354585e-01
-2.53177971e-01 -9.02601600e-01 5.72347164e-01 -3.64267200e-01
-1.29138458e+00 -4.42966580e-01 3.30745697e-01 -5.04606962e-01
1.65251896e-01 -2.77102709e-01 -5.15655518e-01 1.26507926e+00
-1.52908289e+00 -8.60211611e-01 1.69587091e-01 2.44916514e-01
-2.55973727e-01 6.88723862e-01 8.07559550e-01 2.89310664e-01
-5.41251659e-01 4.42057222e-01 -2.74509817e-01 -9.69739631e-02
1.12934864e+00 -1.12559485e+00 1.56981930e-01 9.28284287e-01
-5.90187430e-01 1.09069753e+00 8.95475149e-01 -7.98293710e-01
-8.07606876e-01 -7.02378035e-01 1.33274412e+00 -4.14807647e-01
7.98323214e-01 -6.32513016e-02 -6.29254043e-01 7.94844568e-01
9.70755890e-02 -3.37639302e-01 9.77976978e-01 7.37075925e-01
-5.81586838e-01 7.91765079e-02 -1.36966586e+00 9.68036413e-01
1.20676053e+00 -5.01867652e-01 -5.64021289e-01 -4.23348695e-03
5.36682785e-01 3.48618217e-02 -8.03832591e-01 3.49017948e-01
9.15253103e-01 -1.10661685e+00 8.11152577e-01 -9.31834638e-01
6.63369238e-01 -3.46663482e-02 -4.82666641e-02 -9.93522584e-01
-6.48504496e-01 -3.52525949e-01 9.36017096e-01 1.20996380e+00
1.72722489e-01 -8.54385257e-01 5.15854657e-01 1.52713013e+00
3.29975247e-01 -3.17620933e-01 -8.54320526e-01 -3.22634667e-01
5.77446878e-01 -4.05117124e-01 7.48311520e-01 1.63056195e+00
3.61143619e-01 -4.70319353e-02 -4.36531991e-01 1.05113775e-01
6.62220716e-01 3.48857403e-01 6.15652680e-01 -1.38092995e+00
-1.86642967e-02 -7.13976204e-01 -4.79565412e-02 -9.88227725e-02
1.45595744e-01 -4.27321374e-01 -3.64066154e-01 -1.34231067e+00
6.98750854e-01 -7.98990726e-01 -2.80519605e-01 3.46088916e-01
-7.13521004e-01 -2.63177067e-01 3.94189537e-01 1.04491413e-01
1.46144521e-04 3.34490597e-01 1.12528753e+00 1.58412963e-01
8.74432698e-02 -2.15481505e-01 -1.56054235e+00 5.30158281e-01
6.92197740e-01 -6.23319983e-01 -3.34700972e-01 -3.02847177e-01
4.75919455e-01 2.26781204e-01 8.94081771e-01 -9.09806117e-02
-1.17111363e-01 -1.18845296e+00 -3.83186117e-02 4.14570212e-01
-4.42862630e-01 -8.85152459e-01 5.36312759e-01 6.72524631e-01
-7.46349216e-01 -1.83542028e-01 7.78652430e-02 3.37577224e-01
1.36366248e-01 -2.23564297e-01 6.31163180e-01 -1.89304538e-02
-2.41956681e-01 3.44208255e-02 -2.19664514e-01 2.94008255e-01
9.65487719e-01 -8.00141171e-02 -8.03022325e-01 -3.90494585e-01
-3.94586354e-01 1.95520148e-01 8.11954379e-01 2.38385290e-01
-9.49877780e-03 -1.20280039e+00 -8.70532393e-01 -4.77679312e-01
1.84001699e-02 -7.17818618e-01 4.31595832e-01 1.06505167e+00
-1.05967894e-01 3.83605093e-01 -2.04642519e-01 1.55202180e-01
-1.06106257e+00 6.84403181e-01 3.28875929e-01 1.53878346e-01
-2.07872912e-01 3.18178177e-01 7.94751108e-01 -7.59189427e-02
-8.42624381e-02 -7.06527382e-02 -8.39902684e-02 1.24886461e-01
3.87597769e-01 7.96986818e-01 -3.03535372e-01 -5.39081633e-01
-5.66120267e-01 1.30027205e-01 3.66861969e-01 -3.48553628e-01
1.05369425e+00 -4.19958413e-01 -5.33022940e-01 3.53348762e-01
4.47425753e-01 5.46546161e-01 -9.07236278e-01 2.96179205e-01
1.34262979e-01 -8.27314019e-01 -2.10401699e-01 -9.36793029e-01
-6.79737389e-01 6.43386424e-01 2.42779866e-01 3.80939692e-01
9.95378435e-01 -3.93671960e-01 -1.80073604e-01 -4.46193337e-01
3.66124958e-01 -1.06790531e+00 -7.08164454e-01 -1.75131068e-01
7.19913244e-01 -1.11698532e+00 3.35005641e-01 -6.21417224e-01
-4.99976873e-01 6.36956692e-01 1.97135776e-01 1.92242980e-01
4.06421840e-01 -4.29046489e-02 8.30553621e-02 -1.39428288e-01
-5.35092413e-01 -6.26259968e-02 -1.32414242e-02 4.92953658e-01
9.55463171e-01 8.20941925e-01 -1.39594162e+00 5.27161002e-01
-2.67750770e-01 2.44466618e-01 9.22152698e-01 6.75013423e-01
-1.27664432e-01 -1.22550893e+00 -5.49142420e-01 5.35511494e-01
-1.04963541e+00 -2.59617895e-01 -4.39278662e-01 9.89373982e-01
1.12927660e-01 1.30437052e+00 -3.59465778e-02 1.84865117e-01
3.12177807e-01 -2.37409607e-01 2.94640332e-01 -4.68832552e-01
-6.65535033e-01 -8.10504928e-02 5.15995085e-01 -5.76931059e-01
-8.33314955e-01 -9.81286883e-01 -9.22369480e-01 -1.06531894e+00
-2.46506467e-01 3.17223072e-01 2.13366240e-01 1.03322017e+00
2.31910735e-01 2.74790674e-01 4.32372987e-01 -6.08385764e-02
-6.82703674e-01 -4.24104035e-01 -5.86947441e-01 5.40638328e-01
2.81395525e-01 -7.55156279e-01 -5.22003770e-01 -2.48859763e-01] | [8.782029151916504, 5.546262741088867] |
050a87a7-1fdb-4529-ac42-bf7be89e798d | evaluation-of-interpretability-for-deep | 2111.13208 | null | https://arxiv.org/abs/2111.13208v6 | https://arxiv.org/pdf/2111.13208v6.pdf | Evaluation of Interpretability for Deep Learning algorithms in EEG Emotion Recognition: A case study in Autism | Current models on Explainable Artificial Intelligence (XAI) have shown an evident and quantified lack of reliability for measuring feature-relevance when statistically entangled features are proposed for training deep classifiers. There has been an increase in the application of Deep Learning in clinical trials to predict early diagnosis of neuro-developmental disorders, such as Autism Spectrum Disorder (ASD). However, the inclusion of more reliable saliency-maps to obtain more trustworthy and interpretable metrics using neural activity features is still insufficiently mature for practical applications in diagnostics or clinical trials. Moreover, in ASD research the inclusion of deep classifiers that use neural measures to predict viewed facial emotions is relatively unexplored. Therefore, in this study we propose the evaluation of a Convolutional Neural Network (CNN) for electroencephalography (EEG)-based facial emotion recognition decoding complemented with a novel RemOve-And-Retrain (ROAR) methodology to recover highly relevant features used in the classifier. Specifically, we compare well-known relevance maps such as Layer-Wise Relevance Propagation (LRP), PatternNet, Pattern-Attribution, and Smooth-Grad Squared. This study is the first to consolidate a more transparent feature-relevance calculation for a successful EEG-based facial emotion recognition using a within-subject-trained CNN in typically-developed and ASD individuals. | ['Giuseppe Riccardi', 'Matthew D. Lerner', 'Tessa Clarkson', 'Sara Medina-DeVilliers', 'Juan Manuel Mayor-Torres'] | 2021-11-25 | null | null | null | null | ['facial-emotion-recognition', 'eeg-emotion-recognition'] | ['computer-vision', 'miscellaneous'] | [ 5.20159841e-01 4.19872731e-01 3.58705938e-01 -7.15434790e-01
-1.10446885e-01 2.29025185e-01 3.35294455e-01 3.17179292e-01
-3.32868427e-01 7.53073514e-01 8.96632597e-02 1.12681612e-01
-8.03758502e-01 -2.04844847e-01 -4.45043683e-01 -5.49546659e-01
-4.48420584e-01 2.22191602e-01 -4.08994615e-01 -2.03465223e-01
2.81423867e-01 6.90041244e-01 -2.00027943e+00 4.10465539e-01
7.86657989e-01 1.33850551e+00 5.51426634e-02 2.39512503e-01
3.62654418e-01 6.37989819e-01 -6.20019138e-01 -2.81279474e-01
-3.33646446e-01 -5.98119318e-01 -8.82348180e-01 -3.93573701e-01
-1.43564976e-04 -7.19871596e-02 1.52223140e-01 9.37796474e-01
5.58215380e-01 -1.83022007e-01 6.76155627e-01 -1.54256117e+00
-7.78813183e-01 3.52506012e-01 -2.96207387e-02 4.73666430e-01
5.27472019e-01 1.41234592e-01 8.40719402e-01 -8.86073589e-01
4.00737405e-01 1.00401270e+00 4.95199859e-01 8.08472812e-01
-8.62739742e-01 -5.72012842e-01 -4.24150042e-02 8.61192048e-01
-1.09236228e+00 -2.88077116e-01 4.73247528e-01 -4.03671741e-01
1.36572325e+00 2.44910389e-01 1.26825202e+00 1.34270942e+00
4.01723474e-01 1.91578403e-01 1.40120256e+00 -3.44543695e-01
4.95420963e-01 1.57610834e-01 3.65974195e-02 5.77046096e-01
2.71162540e-01 2.82972485e-01 -9.79012668e-01 7.19493181e-02
2.71136969e-01 -9.61747468e-02 -2.86978543e-01 3.06091100e-01
-1.04982960e+00 5.82870781e-01 6.23011768e-01 8.48453104e-01
-9.24604952e-01 1.65665343e-01 3.49201083e-01 3.93237352e-01
9.13567781e-01 7.40206182e-01 -6.71244085e-01 -4.47977424e-01
-1.06032825e+00 6.67512715e-02 3.90523583e-01 1.68530658e-01
4.49890077e-01 3.37828547e-01 -1.44515723e-01 5.54100037e-01
4.09187466e-01 2.19795749e-01 7.72985101e-01 -6.62076890e-01
-2.00319633e-01 8.97413790e-01 -4.67169344e-01 -1.16663909e+00
-1.02401936e+00 -3.96070957e-01 -8.18037271e-01 5.43591797e-01
-2.27306917e-01 -1.35192782e-01 -6.29220188e-01 1.89387238e+00
1.54592216e-01 3.26909155e-01 -3.14881951e-02 1.14163601e+00
8.21462572e-01 -1.28070325e-01 2.98226327e-01 -1.40771970e-01
1.29186893e+00 -5.01700342e-01 -5.44129848e-01 -3.53468023e-02
4.63349611e-01 -4.99638468e-02 8.10470343e-01 7.13528693e-01
-1.00543296e+00 -2.20085636e-01 -1.09168625e+00 1.46956056e-01
-6.13465965e-01 1.03983149e-01 8.95296335e-01 6.81628406e-01
-1.40417564e+00 9.92320240e-01 -6.94721818e-01 -5.09698749e-01
1.05605137e+00 8.33649516e-01 -9.25414562e-01 3.54168534e-01
-1.14083588e+00 1.50964165e+00 2.37734646e-01 3.67250919e-01
-8.35156977e-01 -8.56070280e-01 -6.39203906e-01 2.52735376e-01
-4.73993301e-01 -7.89506257e-01 7.11238444e-01 -1.66503322e+00
-1.58426821e+00 1.06611443e+00 -3.35711464e-02 -4.38870460e-01
-9.71690863e-02 -1.68131873e-01 -5.06404042e-01 4.48187411e-01
-2.25947469e-01 1.06558955e+00 1.06396341e+00 -5.00685573e-01
-3.42920512e-01 -8.03153217e-01 -3.07879627e-01 -2.05048919e-03
-5.54645538e-01 4.33014542e-01 6.87602222e-01 -4.13200647e-01
1.31828606e-01 -6.59693539e-01 1.46596283e-01 1.85849130e-01
-2.36786366e-01 -4.72188681e-01 4.96256173e-01 -7.70320117e-01
8.67529929e-01 -2.05518794e+00 3.88918340e-01 1.07064098e-01
5.14695466e-01 2.37629011e-01 -1.84182733e-01 -6.74135014e-02
-8.32453132e-01 9.40203592e-02 -1.88054174e-01 -2.38293692e-01
7.55862966e-02 -5.34089543e-02 2.18729615e-01 6.04819238e-01
9.29942906e-01 1.10133612e+00 -7.76962399e-01 1.17997222e-01
2.22020298e-01 7.57915914e-01 -5.92077434e-01 6.80343956e-02
1.16797693e-01 5.03179789e-01 -9.72924680e-02 7.49856710e-01
3.01257044e-01 -1.08561395e-02 -3.36969584e-01 -9.93908793e-02
-1.15162298e-01 3.97002131e-01 -4.23350960e-01 1.46592510e+00
-2.02360272e-01 9.96913612e-01 -2.03302398e-01 -1.19554543e+00
9.98486757e-01 6.18511677e-01 5.00512838e-01 -7.47012973e-01
6.19772851e-01 3.68173182e-01 3.78118992e-01 -6.89962685e-01
-1.41904160e-01 -2.60936588e-01 5.07458270e-01 2.52780855e-01
5.38763285e-01 1.01407342e-01 -4.02574211e-01 -4.18271333e-01
1.41507030e+00 1.79051399e-01 3.23587477e-01 -4.01420772e-01
4.22035486e-01 -3.12047392e-01 2.30320424e-01 2.58574300e-02
-3.84069204e-01 6.70367241e-01 7.02879846e-01 -4.33353066e-01
-6.30244434e-01 -4.57223773e-01 -2.85824597e-01 8.36219072e-01
-5.16246796e-01 -4.95751156e-03 -1.06812561e+00 -4.38348591e-01
-3.39766860e-01 8.26093495e-01 -9.33123708e-01 -8.37697208e-01
1.03040092e-01 -9.08344626e-01 5.32660604e-01 2.83017069e-01
1.66264236e-01 -1.65657151e+00 -1.06124604e+00 1.72035396e-01
2.70636827e-01 -5.79608202e-01 6.02682590e-01 5.86227357e-01
-7.26395547e-01 -1.08292270e+00 -5.55733740e-01 -6.42516851e-01
7.01072216e-01 -3.39395165e-01 6.18792295e-01 8.47149715e-02
-6.54499590e-01 5.81373096e-01 -5.77821672e-01 -8.88844252e-01
-6.74636960e-02 1.54471667e-02 2.91090846e-01 5.83892614e-02
7.84857392e-01 -1.00486946e+00 -5.97419620e-01 -2.45403126e-01
-7.36041367e-01 -1.33764982e-01 4.74060774e-01 7.39077270e-01
2.88465321e-01 -3.14187109e-01 9.63351250e-01 -7.17471819e-04
8.25755060e-01 -8.08734298e-01 -5.47171012e-02 1.24351531e-01
-8.90300453e-01 -1.26563624e-01 2.12501094e-01 -3.05382788e-01
-7.68640101e-01 -3.03750068e-01 -4.90627468e-01 -5.22075176e-01
-6.35171115e-01 6.19761646e-01 2.28625178e-01 -2.87397981e-01
7.52678573e-01 -1.99810073e-01 1.18488193e-01 -1.87445953e-01
-9.56485346e-02 6.08154535e-01 3.22190702e-01 -1.09793067e-01
1.31007046e-01 4.15292978e-02 2.88172364e-01 -5.53211629e-01
-5.81775188e-01 -4.00055274e-02 -4.88317370e-01 -4.37945366e-01
9.66821253e-01 -6.26605988e-01 -9.12375033e-01 5.51204026e-01
-1.27227938e+00 -1.47055879e-01 -3.71310055e-01 8.66006374e-01
-6.38547838e-01 -1.92372173e-01 -1.13772504e-01 -6.30598187e-01
-9.51234758e-01 -1.08800113e+00 1.07677329e+00 4.00277674e-01
-6.23065352e-01 -7.59985447e-01 1.48303717e-01 6.69448916e-03
6.49089754e-01 3.28447998e-01 1.01661015e+00 -9.26956356e-01
6.97376579e-02 -1.37481615e-01 -2.57900894e-01 4.29453254e-01
9.48865432e-03 6.13958798e-02 -1.37803638e+00 1.34556964e-01
1.97530970e-01 -5.05258679e-01 6.34839177e-01 3.32193315e-01
1.12284184e+00 -1.40290394e-01 7.21503748e-03 4.21772003e-01
1.04721737e+00 4.32771966e-02 8.04473937e-01 3.38508993e-01
1.95225373e-01 9.87179816e-01 3.00622672e-01 6.43408895e-01
2.89648056e-01 5.75612962e-01 7.62527406e-01 5.46534024e-02
4.59620776e-03 3.34719121e-01 3.20081860e-01 3.42814595e-01
-5.07600605e-01 1.19174749e-01 -8.01008880e-01 4.74552393e-01
-1.49851024e+00 -9.28395808e-01 -7.12086856e-02 1.98453724e+00
5.63605249e-01 -2.00957462e-01 -1.89547643e-01 4.38056648e-01
5.58331847e-01 -4.99517679e-01 -5.66165984e-01 -8.83859336e-01
-1.29166350e-01 9.02473390e-01 -3.04831684e-01 -2.10619926e-01
-7.00954854e-01 6.28202498e-01 5.68752289e+00 4.29598749e-01
-1.61908424e+00 3.67342830e-01 7.00934112e-01 -2.86842823e-01
-2.03591362e-01 -4.86422032e-01 -1.30791098e-01 3.28544110e-01
1.39726222e+00 1.99153721e-01 7.26680577e-01 8.81714702e-01
4.35793310e-01 -1.47345856e-01 -1.07606423e+00 1.24033022e+00
1.86634839e-01 -9.74324644e-01 -4.44779098e-01 -1.45514444e-01
3.18801790e-01 1.45314351e-01 3.42246383e-01 1.41859472e-01
-7.07437277e-01 -1.39765656e+00 7.76583016e-01 9.85022426e-01
8.13044190e-01 -5.59120238e-01 7.59593725e-01 -2.25127533e-01
-4.14616555e-01 -2.02182665e-01 -3.12829047e-01 -2.85907745e-01
-2.31406018e-01 6.47426546e-01 -8.66987228e-01 2.06822336e-01
1.03846848e+00 7.69538641e-01 -5.54113388e-01 1.02289784e+00
-4.20212448e-01 6.88942969e-01 -2.51486659e-01 -3.37104917e-01
1.52150065e-01 1.43427059e-01 5.73176682e-01 8.06672573e-01
7.01151371e-01 7.21509531e-02 -9.29071069e-01 1.36619198e+00
2.16091901e-01 2.94502169e-01 -4.63864565e-01 -2.40734935e-01
-3.52135785e-02 1.57128596e+00 -6.09736621e-01 1.99948102e-01
-2.34329253e-01 1.10834873e+00 4.59798247e-01 -3.07745449e-02
-6.40910089e-01 -1.79299563e-01 8.03021610e-01 -2.41923649e-02
8.60888287e-02 2.98612177e-01 -3.85133535e-01 -7.64821291e-01
7.66870454e-02 -7.29917645e-01 -7.73685500e-02 -1.24663210e+00
-1.14319921e+00 9.53256190e-01 4.07845974e-02 -7.64198065e-01
-3.78192246e-01 -8.36984396e-01 -7.94427752e-01 1.05688131e+00
-1.48846745e+00 -1.11536717e+00 -4.16457087e-01 6.51118696e-01
-2.52207648e-02 -3.23598266e-01 1.22198606e+00 1.89008102e-01
-6.97353363e-01 3.23526144e-01 -4.43130374e-01 -5.00809193e-01
2.93885171e-01 -1.11148345e+00 -1.97285488e-01 3.15649003e-01
-1.31265051e-03 3.44042748e-01 6.20487809e-01 -4.35998678e-01
-8.93154323e-01 -9.09757197e-01 9.42213833e-01 -1.91614524e-01
4.80302244e-01 7.44603854e-03 -9.12115276e-01 3.66885066e-01
3.27012211e-01 3.39990258e-02 8.33361864e-01 2.40468848e-02
-8.42292756e-02 -1.32304668e-01 -1.54446900e+00 2.84026116e-01
9.39363420e-01 -3.94364923e-01 -6.89790547e-01 2.12796733e-01
3.49486440e-01 1.70698732e-01 -9.66546297e-01 5.56172729e-01
5.12530506e-01 -1.24680197e+00 5.20805359e-01 -7.42291629e-01
5.79600513e-01 2.37592950e-01 1.04200438e-01 -1.58710134e+00
-1.78298980e-01 -4.20040458e-01 1.03713311e-01 9.74082530e-01
4.08364594e-01 -8.88007522e-01 4.88959402e-01 8.62281799e-01
-3.87412816e-01 -1.18080962e+00 -1.30249715e+00 -2.30785593e-01
-1.56946465e-01 -6.53616607e-01 6.51530445e-01 7.53699839e-01
6.37270272e-01 5.48164779e-03 2.74791010e-02 -3.54142301e-02
-1.15948439e-01 -5.33039629e-01 -1.73998952e-01 -1.56442535e+00
2.08143696e-01 -7.15623975e-01 -1.20839500e+00 5.22992015e-01
3.98327202e-01 -1.11408436e+00 -1.24682896e-01 -1.42620587e+00
2.42791492e-02 -3.94724123e-02 -7.21877933e-01 9.18114066e-01
9.60190315e-03 2.66467631e-01 -1.13424160e-01 -4.09794867e-01
-3.86612833e-01 7.64815390e-01 7.06954420e-01 8.06658417e-02
-1.02021426e-01 -1.36411235e-01 -7.51252949e-01 7.44418025e-01
1.09217691e+00 -7.84821630e-01 -4.02264833e-01 8.00518468e-02
4.97663081e-01 -2.86060899e-01 8.07824969e-01 -1.40786242e+00
-8.17040503e-02 3.82164270e-01 6.76279783e-01 3.50048346e-03
3.14032942e-01 -7.50361562e-01 1.65304124e-01 4.31527823e-01
-2.73068964e-01 -7.57247508e-02 3.70566994e-01 -1.66197158e-02
-1.60284132e-01 -4.46451247e-01 5.16083121e-01 1.46544829e-01
-5.47072709e-01 3.65503758e-01 -4.83447313e-01 -3.10658395e-01
1.03116250e+00 -3.08523238e-01 -2.23518923e-01 -1.75214052e-01
-8.77815664e-01 -5.53587317e-01 -4.95937318e-02 4.47387934e-01
9.87554431e-01 -1.15830743e+00 -6.92411244e-01 1.73501045e-01
5.86426631e-02 -5.37122667e-01 4.58326668e-01 1.46670163e+00
-2.94226050e-01 5.18805683e-01 -7.98152447e-01 -3.91471714e-01
-1.14251363e+00 3.67159307e-01 5.29621184e-01 1.24881819e-01
-4.44373816e-01 1.01478863e+00 -1.43950105e-01 -7.14014024e-02
3.58101502e-02 -4.64520574e-01 -6.30599320e-01 3.48661155e-01
6.98723614e-01 3.75203222e-01 5.46261966e-01 -6.30271554e-01
-5.22542655e-01 1.77606374e-01 2.47044116e-01 6.33407850e-04
1.85408628e+00 2.65181988e-01 -5.37520230e-01 2.52970278e-01
1.00257587e+00 -8.51815641e-01 -7.89697886e-01 4.70617235e-01
1.41902128e-02 1.48060009e-01 3.83757770e-01 -1.11446238e+00
-1.23192477e+00 1.17976642e+00 1.09477592e+00 8.62752944e-02
1.56224251e+00 -8.76148567e-02 4.72243987e-02 3.11005294e-01
5.42925358e-01 -1.02215254e+00 -1.31267973e-03 1.36720330e-01
1.58314323e+00 -9.23811853e-01 -1.01736076e-01 1.13499880e-01
-6.01857841e-01 1.43717170e+00 5.32048583e-01 2.69997511e-02
9.20651972e-01 -5.30914851e-02 -2.76556015e-01 -6.61726058e-01
-8.43884408e-01 -4.04020876e-01 6.29679024e-01 8.99347425e-01
3.63067955e-01 1.05936252e-01 -3.21931869e-01 1.22085631e+00
-3.12735111e-01 2.89588660e-01 2.21049130e-01 5.29466867e-01
-4.45405155e-01 -7.56895423e-01 -2.54341781e-01 8.78217638e-01
-7.08793640e-01 -2.55948633e-01 -4.17548448e-01 5.71129441e-01
5.15495420e-01 8.54737401e-01 -2.56882608e-02 -4.80140090e-01
1.92772359e-01 5.21925092e-01 6.88443720e-01 -7.01861441e-01
-9.98127460e-01 -6.10581517e-01 1.71852559e-02 -7.43118942e-01
-7.61693001e-01 -8.16865981e-01 -1.33315182e+00 3.13108624e-03
-3.00670773e-01 -3.94284427e-01 1.20097864e+00 1.25490689e+00
6.70555353e-01 8.03679466e-01 3.53447586e-01 -1.01922417e+00
5.49127348e-02 -1.22397816e+00 -7.24951565e-01 1.45164371e-01
1.56119898e-01 -1.02930605e+00 -3.84733856e-01 -2.82299846e-01] | [13.10433292388916, 3.4661197662353516] |
0087951a-b38c-427a-9399-b5feb009f49d | machine-learning-based-detection-of | 2303.11429 | null | https://arxiv.org/abs/2303.11429v1 | https://arxiv.org/pdf/2303.11429v1.pdf | Machine learning-based detection of cardiovascular disease using ECG signals: performance vs. complexity | Cardiovascular disease remains a significant problem in modern society. Among non-invasive techniques, the electrocardiogram (ECG) is one of the most reliable methods for detecting abnormalities in cardiac activities. However, ECG interpretation requires expert knowledge and it is time-consuming. Developing a novel method to detect the disease early could prevent death and complication. The paper presents novel various approaches for classifying cardiac diseases from ECG recordings. The first approach suggests the Poincare representation of ECG signal and deep-learning-based image classifiers (ResNet50 and DenseNet121 were learned over Poincare diagrams), which showed decent performance in predicting AF (atrial fibrillation) but not other types of arrhythmia. XGBoost, a gradient-boosting model, showed an acceptable performance in long-term data but had a long inference time due to highly-consuming calculation within the pre-processing phase. Finally, the 1D convolutional model, specifically the 1D ResNet, showed the best results in both studied CinC 2017 and CinC 2020 datasets, reaching the F1 score of 85% and 71%, respectively, and that was superior to the first-ranking solution of each challenge. The paper also investigated efficiency metrics such as power consumption and equivalent CO2 emissions, with one-dimensional models like 1D CNN and 1D ResNet being the most energy efficient. Model interpretation analysis showed that the DenseNet detected AF using heart rate variability while the 1DResNet assessed AF pattern in raw ECG signals. | ['Semen Budennyy', 'Alexey Kazakov', 'Konstantin Egorov', 'Huy Pham'] | 2023-03-10 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [-8.46508592e-02 -4.63621244e-02 -2.81479657e-02 2.65280679e-02
-3.96593034e-01 -5.68586528e-01 8.84288251e-02 2.21356899e-01
-2.52452582e-01 8.94262612e-01 -3.11154127e-02 -5.45009851e-01
-7.25601137e-01 -8.96399558e-01 -4.42694813e-01 -5.44762254e-01
-6.19716167e-01 1.78405240e-01 -5.00493646e-01 1.50991857e-01
-1.11875564e-01 5.63777626e-01 -1.10326159e+00 -6.70247599e-02
1.01482403e+00 1.37175393e+00 -2.01761782e-01 7.35895157e-01
3.96160245e-01 6.24293327e-01 -8.03292394e-01 -9.88010913e-02
2.07545295e-01 -9.07794058e-01 -5.00969410e-01 -7.65365183e-01
-6.92975596e-02 -2.25926358e-02 7.48582045e-03 5.43711126e-01
1.19480252e+00 -4.17492062e-01 5.29957473e-01 -1.03954005e+00
-1.27074063e-01 4.53978777e-01 -1.02106459e-01 6.46959543e-01
1.64824173e-01 3.71811241e-02 3.94684553e-01 -6.20151639e-01
1.92020372e-01 6.50356233e-01 1.12094390e+00 2.89399534e-01
-9.69241977e-01 -7.13775456e-01 -4.84675765e-01 3.57163757e-01
-1.56088269e+00 -9.27327666e-03 7.01712728e-01 -3.93728852e-01
9.81853962e-01 6.27451003e-01 1.30284131e+00 9.65043843e-01
7.28936851e-01 -4.88330834e-02 1.36758232e+00 -4.19063509e-01
2.73573756e-01 -4.22778837e-02 1.99793905e-01 7.04996526e-01
5.11380911e-01 4.62949902e-01 -3.90164346e-01 -4.60809588e-01
5.62443733e-01 1.69619873e-01 -3.16280752e-01 2.50073075e-01
-1.19264853e+00 4.98833090e-01 5.61490178e-01 6.00233614e-01
-6.46172583e-01 1.86039776e-01 5.84969103e-01 2.60756284e-01
5.20410955e-01 6.97221458e-01 -7.23556340e-01 -3.59492689e-01
-8.79899800e-01 -8.88199359e-02 5.93855321e-01 2.38913476e-01
5.21458723e-02 2.00648814e-01 -5.44414282e-01 5.63434660e-01
2.36124530e-01 6.40841961e-01 5.11006713e-01 -5.47714829e-01
1.35386109e-01 7.27050066e-01 -3.02596956e-01 -1.26139927e+00
-1.06227374e+00 -1.33502448e+00 -1.63034439e+00 -7.64437690e-02
2.19350964e-01 -5.63100517e-01 -9.20586646e-01 1.47919691e+00
-5.82532845e-02 2.44232595e-01 -5.20221628e-02 7.95120060e-01
8.04941058e-01 4.69752103e-01 3.07611614e-01 -5.19120157e-01
1.49463403e+00 -4.02002394e-01 -7.49780774e-01 1.40877858e-01
4.50999528e-01 -1.12937808e-01 5.21527767e-01 5.53848088e-01
-9.80400681e-01 -8.42189789e-01 -1.24739182e+00 3.91691566e-01
-2.74388433e-01 3.01934600e-01 4.63227361e-01 1.20950079e+00
-1.13170660e+00 9.68658566e-01 -6.17814898e-01 -4.89399642e-01
6.51632011e-01 2.99482405e-01 1.05622828e-01 2.31089234e-01
-1.49422693e+00 1.16897082e+00 1.76572114e-01 2.34764546e-01
-1.01098776e+00 -9.23832774e-01 -4.45454419e-01 2.65040666e-01
-2.23082095e-01 -1.06423771e+00 5.43278992e-01 -6.67048275e-01
-1.31376505e+00 5.88246942e-01 3.15083295e-01 -8.39932799e-01
6.61743760e-01 -3.32926124e-01 -7.09438205e-01 8.71024206e-02
-2.79542267e-01 2.04664692e-01 5.56162179e-01 -4.85364705e-01
-2.66459972e-01 -5.12322903e-01 -1.85853302e-01 1.05071552e-01
-2.64171392e-01 -2.59572744e-01 2.27429643e-01 -5.96916854e-01
-1.91019773e-02 -9.69881833e-01 -2.35165969e-01 -2.07598478e-01
-3.34856659e-01 -1.14089862e-01 4.85028595e-01 -9.15331721e-01
1.61166024e+00 -1.96408033e+00 -1.31496936e-01 4.40615624e-01
4.48368818e-01 3.65600646e-01 5.35916984e-01 -5.80647402e-02
-3.09388041e-01 6.21727288e-01 -1.50070965e-01 5.12781978e-01
-3.99796546e-01 -7.11896569e-02 2.35727862e-01 5.45092285e-01
9.05532986e-02 1.01996076e+00 -7.25754380e-01 -3.07425588e-01
3.78163278e-01 7.13882864e-01 -9.25645605e-02 6.77210167e-02
4.18467283e-01 4.61541086e-01 -2.67158449e-01 5.19496262e-01
5.40782750e-01 -3.47270131e-01 3.47602934e-01 -4.90126282e-01
1.29028440e-01 1.80749092e-02 -8.39075804e-01 1.81119609e+00
-4.93904859e-01 4.75649416e-01 -4.71834391e-01 -1.25762320e+00
1.06342006e+00 7.47691691e-01 7.98458815e-01 -9.57726955e-01
2.31270567e-01 2.73664206e-01 2.76651561e-01 -6.16005480e-01
-5.52014232e-01 -1.20376073e-01 -5.71627468e-02 1.98182762e-01
-1.10733844e-01 3.19622427e-01 -1.39391720e-01 -2.38462448e-01
1.24884641e+00 8.88291653e-03 6.24442816e-01 -7.76841998e-01
4.10280436e-01 -1.59927398e-01 6.85315728e-01 8.60687494e-01
-1.89172000e-01 5.78342557e-01 4.04951811e-01 -1.08398855e+00
-5.79436123e-01 -1.03958428e+00 -3.60070586e-01 2.02640563e-01
-1.10907115e-01 -4.12374407e-01 -6.45453811e-01 -7.75820732e-01
-8.36351141e-02 5.11267543e-01 -6.22308791e-01 -4.29436833e-01
-5.98401248e-01 -1.35136116e+00 9.68210220e-01 5.89226902e-01
9.61842895e-01 -8.87141526e-01 -1.25950801e+00 3.31553519e-01
-1.94491088e-01 -4.38994020e-01 1.93250090e-01 5.61052978e-01
-1.24543345e+00 -1.31343162e+00 -8.83934915e-01 -4.44448173e-01
3.92510563e-01 -5.79703629e-01 1.37484717e+00 2.27874935e-01
-6.91000342e-01 -4.80737798e-02 -1.57302380e-01 -1.01089108e+00
-1.07196636e-01 8.00548717e-02 8.19759294e-02 -2.02404484e-01
1.85768574e-01 -5.71968496e-01 -1.11449611e+00 1.54315475e-02
-1.57993898e-01 -1.26197606e-01 6.67060792e-01 6.00273013e-01
5.77377081e-01 -1.76992081e-02 1.08663929e+00 -6.67793632e-01
7.15453923e-01 -5.22624254e-01 -1.46906316e-01 1.58295602e-01
-1.30511928e+00 -3.71838182e-01 6.12557411e-01 1.06886290e-01
-4.18670923e-01 9.23624262e-02 -1.05107270e-01 -2.74275839e-01
-4.89551611e-02 4.98882741e-01 1.53185213e-02 2.33455420e-01
1.01354015e+00 3.58197000e-03 -8.61877576e-02 -4.17663962e-01
-3.48494411e-01 3.06632698e-01 2.82039165e-01 -2.18239635e-01
3.98489684e-01 6.06194325e-02 4.92992550e-01 -6.81394696e-01
-6.08663797e-01 5.29279299e-02 -4.03768033e-01 -5.16044974e-01
1.33664572e+00 -1.03155875e+00 -6.67235017e-01 4.77342099e-01
-1.08509374e+00 7.86933005e-02 -5.07383823e-01 6.63333714e-01
-9.74676460e-02 -1.02736704e-01 -2.45428443e-01 -8.01988006e-01
-1.09592736e+00 -5.52036583e-01 5.14352560e-01 1.68132648e-01
-3.86083305e-01 -8.72989476e-01 2.04764828e-01 -9.50027183e-02
7.47576475e-01 9.92817521e-01 1.19128132e+00 -5.75197518e-01
-5.12500964e-02 -2.08700255e-01 -4.41401787e-02 5.67535937e-01
2.31455296e-01 -2.42115319e-01 -1.15010297e+00 -1.39493883e-01
1.45762041e-01 2.86862433e-01 4.50508207e-01 9.17460203e-01
1.62714207e+00 -5.50005361e-02 -4.52870101e-01 6.90120041e-01
1.48136353e+00 7.18372226e-01 8.40949118e-01 1.13809362e-01
4.40853328e-01 -1.86432123e-01 9.85768065e-02 3.22505623e-01
8.38844553e-02 3.46266627e-01 4.21978176e-01 -7.20690429e-01
-1.42120034e-01 2.02392533e-01 -5.35264239e-02 8.17892194e-01
-7.85070598e-01 -2.97519356e-01 -1.14317477e+00 2.33302251e-01
-1.53098905e+00 -9.07960653e-01 -4.92516309e-01 2.21178937e+00
3.58938187e-01 1.84711009e-01 1.85499594e-01 7.74043500e-01
5.81981182e-01 -2.96779722e-01 -6.00992858e-01 -5.29130578e-01
-1.78302050e-01 7.37256467e-01 4.03181195e-01 -2.77331710e-01
-1.06563199e+00 -2.81609297e-01 6.90816593e+00 1.51297405e-01
-1.11805880e+00 3.90237421e-01 1.14348066e+00 -6.24429360e-02
2.28259012e-01 -4.39027637e-01 -4.89157774e-02 7.36914694e-01
1.49430346e+00 -2.84142494e-01 -1.73749123e-02 5.13034105e-01
4.04490322e-01 4.02929410e-02 -1.05970693e+00 1.25739002e+00
-1.15301229e-01 -1.32763970e+00 -4.67374861e-01 -2.73188390e-02
3.60433429e-01 8.30079913e-02 -4.24338877e-01 2.37980828e-01
-3.96350414e-01 -1.19960177e+00 2.79025942e-01 9.29089546e-01
1.21271372e+00 -8.70376408e-01 1.10344791e+00 1.81763232e-01
-1.13203144e+00 -3.98858339e-01 5.14205098e-02 -2.93824732e-01
-5.20838276e-02 1.16707575e+00 -6.55786633e-01 9.54715967e-01
1.17313874e+00 8.81645203e-01 -6.36504054e-01 1.26323032e+00
2.25766487e-02 1.07470691e+00 -3.27910423e-01 -5.87763116e-02
-3.38340253e-01 6.94799945e-02 4.42302108e-01 1.05899084e+00
8.58882725e-01 4.78110276e-02 5.65614812e-02 6.12223983e-01
1.53736360e-02 9.76968259e-02 -6.23927712e-01 4.57807481e-01
2.49830499e-01 1.20532763e+00 -7.50232041e-01 -4.60033089e-01
-6.11101501e-02 4.98534739e-01 -4.96585041e-01 7.77004147e-03
-1.17897332e+00 -5.42175770e-01 1.79328606e-01 3.80839497e-01
-5.11500657e-01 3.28345746e-01 -7.00066984e-01 -7.40547955e-01
1.10492304e-01 -7.61319637e-01 6.44152164e-01 -6.42638028e-01
-8.86610746e-01 7.05667436e-01 -1.42316982e-01 -1.47015667e+00
-1.69290483e-01 -3.52875739e-01 -5.81784546e-01 1.14390111e+00
-1.20660877e+00 -4.38163102e-01 -8.26625943e-01 3.97834033e-01
2.21538991e-01 -1.55953377e-01 1.35399210e+00 7.23209202e-01
-6.76833153e-01 2.82767355e-01 -2.37282723e-01 -6.07471971e-04
3.50472271e-01 -1.30025089e+00 1.01011924e-01 6.29009485e-01
-2.26614118e-01 2.67588437e-01 2.88193882e-01 -5.86644411e-01
-1.16317904e+00 -1.25765562e+00 8.17009628e-01 -2.39438742e-01
-2.08316758e-01 9.67424177e-03 -4.63100582e-01 -4.27662022e-02
2.84981668e-01 2.34098792e-01 6.37474239e-01 8.37379508e-03
3.94611269e-01 -5.15098631e-01 -1.31475377e+00 -2.58294269e-02
1.07634211e+00 -2.49655515e-01 -2.82828599e-01 3.37299824e-01
9.06218439e-02 -3.67073983e-01 -1.22901595e+00 9.47208524e-01
8.48277211e-01 -8.31753969e-01 1.00639105e+00 -4.24861521e-01
2.37303913e-01 -2.08421692e-01 3.15068811e-01 -1.37047803e+00
-5.99011004e-01 -5.08666158e-01 -4.04556841e-01 8.94962847e-01
5.32015741e-01 -6.61247551e-01 5.76679587e-01 7.19406679e-02
-1.60854816e-01 -9.52841938e-01 -8.51736486e-01 -8.55336130e-01
-1.98894262e-01 -2.29850739e-01 4.75686640e-01 9.51692462e-01
-1.54846400e-01 4.37667280e-01 -3.23365122e-01 -9.61342081e-02
4.30728406e-01 -3.35465103e-01 1.52183473e-01 -1.80514765e+00
1.38175413e-02 -2.58321196e-01 -6.90685928e-01 1.27629533e-01
-4.47260976e-01 -1.24723160e+00 -4.13765818e-01 -1.85566115e+00
2.28435457e-01 -4.33448851e-01 -1.08019996e+00 4.15026754e-01
-2.03665465e-01 3.51026714e-01 -1.29018679e-01 -6.43403158e-02
9.90060568e-02 -1.31971985e-01 9.70991731e-01 -2.46422559e-01
-3.82333577e-01 1.72382936e-01 -6.80149436e-01 6.51414394e-01
1.00455737e+00 -7.35502183e-01 -6.48370087e-01 -2.52724349e-01
4.78121996e-01 3.48180443e-01 7.32315660e-01 -1.73250449e+00
-1.75540801e-02 4.83232707e-01 1.20423067e+00 -5.70014954e-01
-6.84528127e-02 -9.04679835e-01 8.23765397e-01 1.10914600e+00
-1.21808842e-01 5.36618412e-01 3.22556943e-01 3.57787520e-01
-1.81563273e-02 2.57517457e-01 6.69349372e-01 -4.81381044e-02
-2.65404452e-02 2.70439744e-01 -4.30277258e-01 5.64425848e-02
1.15253866e+00 -2.41607815e-01 -1.94106668e-01 -2.86764186e-02
-1.00801957e+00 -1.60222217e-01 -3.87790591e-01 3.85282576e-01
4.68200266e-01 -1.28396893e+00 -8.35348010e-01 -5.94590716e-02
-2.58417010e-01 -1.14420913e-01 4.93993312e-01 1.15969241e+00
-8.16386521e-01 3.57170016e-01 -4.49223936e-01 -8.56221378e-01
-1.13411248e+00 2.18803510e-01 9.66300547e-01 -4.62138444e-01
-8.38066995e-01 5.57676673e-01 -3.28153402e-01 1.85254991e-01
3.17805767e-01 -4.41815913e-01 -3.95731449e-01 2.37462014e-01
3.07641804e-01 9.20746863e-01 7.24288821e-01 1.12662517e-01
-9.00725782e-01 8.01507831e-01 5.02108335e-01 3.25765669e-01
1.32823682e+00 2.68156677e-01 -7.22174421e-02 2.81626284e-01
1.02588367e+00 -5.40447593e-01 -5.34253836e-01 5.33299208e-01
-1.20350765e-02 1.24254569e-01 3.30042213e-01 -1.52592981e+00
-1.38497186e+00 9.64124680e-01 1.59694695e+00 4.61191356e-01
1.65829206e+00 -5.06343246e-01 5.18541217e-01 2.39387155e-01
2.21545845e-01 -9.79938686e-01 -3.18015367e-01 -8.28829855e-02
7.51019418e-01 -7.13985562e-01 2.97836155e-01 6.91471025e-02
-2.48762742e-01 1.24194849e+00 2.61538118e-01 -9.69295576e-02
1.20700300e+00 1.79180682e-01 1.25040680e-01 -4.79475945e-01
-3.93652797e-01 4.14644033e-01 3.86972755e-01 5.08955896e-01
5.96050322e-01 2.53859818e-01 -6.90923989e-01 8.12691092e-01
-1.34017438e-01 4.07998770e-01 9.47635323e-02 8.34875405e-01
1.07758716e-01 -6.39985740e-01 -1.11706153e-01 1.06883609e+00
-9.27969337e-01 2.22058855e-02 -2.31762797e-01 6.68853223e-01
6.27579808e-01 1.00130200e+00 -2.97732279e-02 -5.45676410e-01
6.48895442e-01 4.46456969e-01 3.06027532e-01 -2.50181675e-01
-1.15123808e+00 -1.54613718e-01 2.27535293e-01 -5.22924066e-01
-5.75009704e-01 -4.28906828e-01 -9.53742802e-01 1.31087378e-01
-3.33081514e-01 1.77918226e-01 1.03039134e+00 6.72388732e-01
7.38348067e-01 1.30889976e+00 7.89228737e-01 -4.05573040e-01
-1.62564486e-01 -1.06040120e+00 -5.58063745e-01 -9.99235455e-03
3.03034425e-01 -3.91338497e-01 -2.11918876e-01 7.14118853e-02] | [14.214637756347656, 3.342275381088257] |
4c786085-c55c-412a-8778-b3a6fc5ad751 | pragmatically-appropriate-diversity-for | 2304.02812 | null | https://arxiv.org/abs/2304.02812v1 | https://arxiv.org/pdf/2304.02812v1.pdf | Pragmatically Appropriate Diversity for Dialogue Evaluation | Linguistic pragmatics state that a conversation's underlying speech acts can constrain the type of response which is appropriate at each turn in the conversation. When generating dialogue responses, neural dialogue agents struggle to produce diverse responses. Currently, dialogue diversity is assessed using automatic metrics, but the underlying speech acts do not inform these metrics. To remedy this, we propose the notion of Pragmatically Appropriate Diversity, defined as the extent to which a conversation creates and constrains the creation of multiple diverse responses. Using a human-created multi-response dataset, we find significant support for the hypothesis that speech acts provide a signal for the diversity of the set of next responses. Building on this result, we propose a new human evaluation task where creative writers predict the extent to which conversations inspire the creation of multiple diverse responses. Our studies find that writers' judgments align with the Pragmatically Appropriate Diversity of conversations. Our work suggests that expectations for diversity metric scores should vary depending on the speech act. | ['Marti A. Hearst', 'Katherine Stasaski'] | 2023-04-06 | null | null | null | null | ['dialogue-evaluation'] | ['natural-language-processing'] | [ 2.03732178e-01 5.75210035e-01 -1.38588771e-01 -7.71173120e-01
-4.66581583e-01 -7.21450329e-01 1.06618369e+00 -7.77574182e-02
-1.68095008e-01 6.02706730e-01 1.39375794e+00 -1.85705289e-01
-3.11254598e-02 -5.56020260e-01 1.18424639e-01 -3.44193667e-01
8.17440689e-01 5.99401057e-01 -3.03505272e-01 -7.07352459e-01
7.31057048e-01 1.66696787e-01 -1.26052701e+00 7.95753539e-01
6.34970725e-01 3.23081613e-01 2.08074838e-01 9.79304492e-01
-3.52097809e-01 1.06755316e+00 -1.01853824e+00 -2.14774221e-01
-5.24541475e-02 -1.24184573e+00 -1.18801117e+00 1.64647222e-01
1.98217809e-01 -3.40101898e-01 1.15155438e-02 7.24018514e-01
6.08476818e-01 4.26200032e-01 6.90181017e-01 -9.38109398e-01
-6.07146561e-01 1.20824134e+00 4.30688977e-01 2.78051198e-01
8.26179385e-01 6.28647506e-01 1.50027168e+00 -4.68247205e-01
1.03085887e+00 1.59256935e+00 2.97455609e-01 1.08384514e+00
-1.22321570e+00 -8.77456367e-02 -5.12165157e-03 -4.21418607e-01
-3.92726481e-01 -1.01925135e+00 7.61644363e-01 -6.97851717e-01
8.83754790e-01 5.52407622e-01 6.74450696e-01 1.33730936e+00
1.33379130e-02 6.76412284e-01 1.28315115e+00 -3.24283421e-01
3.30304027e-01 1.91412836e-01 1.65567875e-01 2.91776862e-02
-4.56379741e-01 -3.84306945e-02 -9.10708368e-01 -3.64099205e-01
4.63820040e-01 -7.65828311e-01 -2.17039496e-01 5.68601608e-01
-1.24789298e+00 1.01542151e+00 -2.19334406e-03 6.60664380e-01
-4.60141838e-01 -3.50856215e-01 5.10710955e-01 5.14930010e-01
1.86153159e-01 1.36912012e+00 -1.71198040e-01 -7.77603209e-01
-5.63156068e-01 5.77103734e-01 1.23597276e+00 5.80482066e-01
2.89823145e-01 -1.37238249e-01 -4.86316234e-01 1.19779491e+00
2.57912070e-01 1.03874758e-01 6.53102100e-01 -1.49910450e+00
1.65217623e-01 6.68807924e-01 1.93140000e-01 -7.86287665e-01
-5.61891913e-01 1.86213508e-01 -2.21063629e-01 -9.27425474e-02
6.72817886e-01 -4.61413294e-01 -5.70986271e-02 2.07163763e+00
1.72915801e-01 -9.42931473e-01 3.27950120e-01 8.35154116e-01
6.74062848e-01 6.47263348e-01 1.52201140e-02 -4.38789099e-01
9.68372107e-01 -6.02118254e-01 -8.57804537e-01 -2.13535294e-01
7.14804530e-01 -1.10009360e+00 1.52107096e+00 2.30982378e-01
-1.15978110e+00 -3.91105890e-01 -8.45373452e-01 -1.05961002e-01
3.46604735e-01 -3.21868122e-01 6.93982005e-01 5.42972565e-01
-9.89012957e-01 3.60816360e-01 -2.10813388e-01 -3.32690597e-01
-2.27375105e-01 -9.02280733e-02 -5.06940223e-02 5.99605083e-01
-1.06119037e+00 1.22561872e+00 4.56074774e-02 -4.08227481e-02
-3.42309177e-01 -1.39637977e-01 -4.26771790e-01 -1.25284076e-01
-1.42742887e-01 -5.88708341e-01 1.65702057e+00 -1.23048735e+00
-2.04666090e+00 8.95376682e-01 2.91462545e-03 -4.32809480e-02
4.35611576e-01 2.28318900e-01 -7.12009892e-02 -1.05432831e-01
1.14724152e-01 8.19011033e-01 3.36271584e-01 -8.08708847e-01
-6.55929327e-01 -4.00659330e-02 4.26202774e-01 6.29850626e-01
-1.51376516e-01 1.94328502e-01 4.25142586e-01 -3.51857662e-01
-5.55803254e-02 -1.19818425e+00 6.66003004e-02 -5.35674274e-01
-3.99359912e-01 -8.27052295e-01 7.79054910e-02 -4.07749146e-01
1.29608440e+00 -2.03601313e+00 4.26632643e-01 5.07851131e-04
3.00558358e-01 -3.42336565e-01 -1.38025299e-01 6.13119602e-01
3.92642051e-01 4.40227687e-01 8.80575925e-02 -4.38905954e-02
4.84733850e-01 6.13688640e-02 -2.71412879e-01 1.67741060e-01
-2.97152195e-02 5.20325303e-01 -6.66674018e-01 -1.35701686e-01
-7.43347853e-02 9.69773903e-02 -8.63020182e-01 4.69105333e-01
-5.58179498e-01 7.81907439e-01 -4.13705945e-01 3.56421247e-02
-2.45911088e-02 -2.43158892e-01 4.91632402e-01 4.00769085e-01
-4.26482409e-01 1.26218462e+00 -5.20710289e-01 1.29500377e+00
-6.28886342e-01 1.00557423e+00 1.48925051e-01 -2.40216613e-01
1.04093289e+00 3.46295863e-01 1.25401421e-02 -6.14262998e-01
1.07201688e-01 3.06699902e-01 9.85162973e-01 -5.29222071e-01
6.60494268e-01 -3.46979350e-01 -4.97770995e-01 1.04676628e+00
-4.25086081e-01 -6.64857149e-01 3.24149966e-01 1.88097954e-01
1.06080711e+00 -2.28688613e-01 3.57632250e-01 -4.89077449e-01
3.72339875e-01 5.99172525e-02 6.31656349e-01 7.82427490e-01
-2.80538172e-01 3.35318863e-01 1.01598108e+00 -4.80759442e-01
-1.32640338e+00 -7.08892524e-01 -1.19159646e-01 1.48518384e+00
-2.01948941e-01 -2.90707976e-01 -9.10913467e-01 -3.50595951e-01
-4.05146360e-01 1.21018803e+00 -4.35627967e-01 1.22839786e-01
-6.65262282e-01 -3.51982713e-01 6.63672328e-01 -3.94123793e-02
1.52361849e-02 -1.57854831e+00 -7.26560950e-01 2.59913087e-01
-7.40189552e-01 -8.98287177e-01 -6.06925249e-01 -2.12381512e-01
-3.82889032e-01 -7.65087247e-01 -2.18551695e-01 -4.55447227e-01
3.37360799e-01 -1.69110656e-01 9.47889745e-01 2.17273027e-01
3.20415765e-01 2.51169652e-01 -4.18584377e-01 -3.09876613e-02
-1.38553154e+00 1.35392040e-01 5.27450740e-02 -4.01042044e-01
2.17669874e-01 -4.89625692e-01 -4.22820449e-01 2.64813513e-01
-5.12553811e-01 3.19471687e-01 1.59552410e-01 7.92925060e-01
-1.76863685e-01 -6.56394124e-01 8.26666832e-01 -8.21971416e-01
1.79916894e+00 -4.60959345e-01 1.07297055e-01 7.57944956e-02
-5.13096213e-01 2.21468881e-01 6.95911407e-01 -5.10309994e-01
-1.17878890e+00 -3.78624618e-01 -3.33570749e-01 5.88410616e-01
-1.93057209e-01 3.21049064e-01 1.05849683e-01 3.22821707e-01
1.10993481e+00 -5.06844893e-02 2.45631918e-01 2.36632992e-02
3.71147156e-01 1.07632840e+00 3.22568342e-02 -9.45871353e-01
-1.32997647e-01 -1.38161957e-01 -4.81131732e-01 -1.10129607e+00
-7.07942724e-01 -1.77692309e-01 -1.40624747e-01 -6.29970789e-01
9.15419340e-01 -6.41573727e-01 -1.00333500e+00 2.04248652e-01
-1.39727652e+00 -6.46107316e-01 -4.05185521e-02 4.83585149e-01
-1.00294173e+00 1.28462836e-01 -6.86400115e-01 -8.48923564e-01
-2.33874008e-01 -1.43706584e+00 6.72461390e-01 2.35899359e-01
-1.67193460e+00 -7.42350280e-01 1.61494255e-01 6.76981449e-01
5.21808028e-01 1.17975831e-01 1.07574177e+00 -1.00589824e+00
-1.12634555e-01 4.19199616e-01 4.43820506e-01 1.97465032e-01
2.09612682e-01 2.33957693e-01 -6.34161592e-01 4.03903872e-01
3.95758927e-01 -6.94369197e-01 1.49531275e-01 3.05552036e-01
2.26813942e-01 -7.32495427e-01 3.02749127e-01 5.23335971e-02
5.57891011e-01 3.09956133e-01 4.47110295e-01 2.31943935e-01
1.96657896e-01 1.23776901e+00 4.09295350e-01 5.49811840e-01
5.87122619e-01 7.35213161e-01 -2.64822811e-01 6.41061187e-01
-1.19244225e-01 -1.83918759e-01 5.90219915e-01 1.06064844e+00
3.76488477e-01 -5.01673400e-01 -9.28966761e-01 4.35606062e-01
-1.61121058e+00 -1.16567504e+00 -1.98251992e-01 1.87734020e+00
1.38784409e+00 2.87267059e-01 2.96480238e-01 -2.14708462e-01
6.92641437e-01 5.01061141e-01 -2.90924579e-01 -1.28860748e+00
-1.39201686e-01 -1.51229024e-01 -2.69412339e-01 1.15152311e+00
-2.72957236e-01 9.58427727e-01 6.92519140e+00 1.35298327e-01
-1.10653830e+00 -1.17751926e-01 7.24133193e-01 -2.97246963e-01
-8.86439621e-01 9.05543193e-03 -6.16711974e-01 6.17184043e-01
1.08725429e+00 -3.37868959e-01 7.18301833e-01 5.65516472e-01
5.08488476e-01 -9.14444774e-02 -1.50189078e+00 3.87679160e-01
-9.85337123e-02 -1.29288602e+00 -2.05199212e-01 2.98413895e-02
6.01863444e-01 -3.56555372e-01 -1.03457533e-01 1.89618304e-01
6.48536921e-01 -9.58351433e-01 7.51568615e-01 4.93375570e-01
3.35836798e-01 -3.26675504e-01 1.62276328e-01 6.70779347e-01
-3.09501350e-01 7.00005591e-02 -3.15055102e-02 -6.17900252e-01
2.85248518e-01 1.77885816e-01 -1.44179595e+00 -6.57090485e-01
-1.12128798e-02 -1.06345683e-01 -8.86309519e-02 6.63676336e-02
-5.57760596e-01 6.29421294e-01 -2.20810518e-01 -7.61589646e-01
2.27038920e-01 -8.89109001e-02 8.13149929e-01 1.08096147e+00
-9.15305596e-03 2.23950386e-01 1.09175049e-01 1.04006696e+00
8.12606886e-02 2.50309706e-01 -5.47728419e-01 -5.65497160e-01
9.28896129e-01 8.51829827e-01 -4.51086581e-01 -1.18906997e-01
2.80750748e-02 6.35893345e-01 3.15131038e-01 4.26890794e-03
-1.86014846e-01 1.16638102e-01 8.48791361e-01 -2.73353547e-01
-4.89951581e-01 -9.23645049e-02 -8.19807410e-01 -7.70325780e-01
-1.13819718e-01 -1.42413890e+00 -2.05834478e-01 -5.99176884e-01
-1.29553115e+00 5.91920853e-01 -4.24042374e-01 -6.04805887e-01
-9.73910034e-01 -2.84405231e-01 -6.02652133e-01 1.08500683e+00
-6.40849352e-01 -5.54461598e-01 2.68790424e-02 -8.58515352e-02
9.64568794e-01 -2.72730201e-01 9.82788146e-01 -3.98701251e-01
-2.63392746e-01 3.28410298e-01 -4.66703713e-01 5.15109673e-02
8.93846393e-01 -1.22718453e+00 6.34887636e-01 4.16558146e-01
-1.41748071e-01 9.95726824e-01 1.11670709e+00 -5.88390648e-01
-8.56826365e-01 -3.98337133e-02 1.40153742e+00 -5.22300303e-01
5.81612349e-01 2.76999664e-03 -4.87102658e-01 3.42620373e-01
5.09847164e-01 -9.05721366e-01 1.16696811e+00 4.53129977e-01
-1.15640022e-01 4.78313923e-01 -9.57086980e-01 1.01776612e+00
1.11992598e+00 -8.33157241e-01 -9.89804327e-01 4.22226906e-01
7.59426773e-01 -3.90563428e-01 -9.88397002e-01 -2.28373986e-02
6.24052525e-01 -1.25837171e+00 3.52417469e-01 -5.98270118e-01
1.00756454e+00 1.32930234e-01 -2.48172134e-01 -1.60353255e+00
-3.24003458e-01 -1.08347011e+00 3.77730668e-01 1.00913107e+00
5.70126235e-01 -4.88222867e-01 3.60384047e-01 1.20483291e+00
-3.80797237e-01 -4.91004854e-01 -8.03645492e-01 -7.26293623e-02
3.19725364e-01 -1.94869656e-02 6.39203906e-01 1.14710271e+00
6.94117904e-01 8.96726489e-01 -1.42249703e-01 -5.16479433e-01
5.98380566e-02 -9.77923349e-02 8.17256808e-01 -1.21719897e+00
-3.89251113e-01 -1.08953762e+00 7.31154829e-02 -1.02461874e+00
4.49101150e-01 -7.73902595e-01 3.94578367e-01 -1.43862498e+00
-1.19384669e-01 -5.89480758e-01 6.01371408e-01 2.10376486e-01
-1.32858351e-01 -5.74839115e-01 5.07772744e-01 4.20716554e-01
-2.11829200e-01 3.33058983e-01 1.40479028e+00 1.49130955e-01
-5.79626679e-01 6.59640133e-03 -1.15410876e+00 7.31783986e-01
1.06818187e+00 -1.56788230e-01 -5.52429378e-01 -2.77066976e-01
5.72210908e-01 5.82626939e-01 -2.35082150e-01 -5.83274364e-01
2.50113666e-01 -6.72346354e-01 -6.90072998e-02 1.66065708e-01
2.57895023e-01 -5.22979647e-02 1.76875778e-02 2.72954851e-01
-1.30744910e+00 2.25595921e-01 -1.80103794e-01 -1.44763605e-03
1.13246115e-02 -3.13186020e-01 7.86862254e-01 -2.45869502e-01
-2.64903665e-01 -5.45974731e-01 -7.59701610e-01 4.05926168e-01
6.47645235e-01 -2.44278938e-01 -5.03659129e-01 -7.13047981e-01
-7.92006195e-01 4.90136817e-02 5.46996355e-01 7.04351664e-01
3.45048577e-01 -1.07408476e+00 -9.30250883e-01 -1.37361705e-01
-1.84861571e-03 -4.22345817e-01 1.73686501e-02 5.27648032e-01
-4.30730730e-01 2.69006103e-01 -3.27777654e-01 -3.56224269e-01
-1.14920139e+00 -3.55665624e-01 3.99307579e-01 -8.32944177e-03
-4.52915043e-01 9.47303116e-01 1.55384568e-02 -5.07585168e-01
-6.92407489e-02 -1.15206651e-01 -2.96275109e-01 2.82723755e-01
5.14578640e-01 2.66105145e-01 -3.19322914e-01 -7.00048566e-01
2.54897624e-02 -2.59209603e-01 -2.02416256e-01 -8.51636827e-01
9.61932242e-01 -1.81456134e-01 -3.38211030e-01 8.45448911e-01
8.28955054e-01 4.90544170e-01 -1.17155552e+00 1.43709481e-01
8.56704917e-03 -3.91967624e-01 -4.30699497e-01 -1.10306478e+00
-7.28436336e-02 4.38304126e-01 -2.82040000e-01 6.25581503e-01
5.14582694e-01 -7.70500004e-02 6.14896655e-01 4.34029996e-01
1.65086046e-01 -1.55138779e+00 3.28623146e-01 9.44985390e-01
1.23242652e+00 -1.02000225e+00 -3.42889398e-01 -9.59234871e-03
-1.32928050e+00 1.20768750e+00 8.44818354e-01 1.86494753e-01
-1.52366664e-02 1.34638548e-01 4.10587102e-01 -1.82467222e-01
-1.34101641e+00 1.47482917e-01 -2.96577755e-02 3.58848780e-01
1.15676165e+00 4.94303316e-01 -1.03774679e+00 2.37582922e-01
-1.11996055e+00 -4.57576036e-01 8.74852479e-01 3.62236440e-01
-7.17238188e-01 -1.19293427e+00 -1.35452196e-01 3.51164550e-01
-1.08175561e-01 -1.43080786e-01 -1.26623046e+00 3.06695640e-01
-1.93731666e-01 1.51452982e+00 1.60911262e-01 -3.83190513e-01
1.45334095e-01 4.11338538e-01 3.17225009e-01 -9.95420635e-01
-1.15113544e+00 -1.19985119e-01 8.28036368e-01 -3.72993857e-01
-3.01688552e-01 -9.40083206e-01 -1.45679903e+00 -6.41295195e-01
-1.94074698e-02 2.25630149e-01 4.88944888e-01 1.11913741e+00
2.72825927e-01 3.10245275e-01 7.17784107e-01 -4.24526364e-01
-8.89042020e-01 -1.33951795e+00 2.46582441e-02 6.76979899e-01
2.56295592e-01 -1.28967494e-01 -6.34614050e-01 -2.57848054e-01] | [12.79451847076416, 8.056619644165039] |
b2c38e10-6ae2-417e-959c-5ec09805083e | program-induction-by-rationale-generation-1 | null | null | https://aclanthology.org/P17-1015 | https://aclanthology.org/P17-1015.pdf | Program Induction by Rationale Generation: Learning to Solve and Explain Algebraic Word Problems | Solving algebraic word problems requires executing a series of arithmetic operations{---}a program{---}to obtain a final answer. However, since programs can be arbitrarily complicated, inducing them directly from question-answer pairs is a formidable challenge. To make this task more feasible, we solve these problems by generating answer rationales, sequences of natural language and human-readable mathematical expressions that derive the final answer through a series of small steps. Although rationales do not explicitly specify programs, they provide a scaffolding for their structure via intermediate milestones. To evaluate our approach, we have created a new 100,000-sample dataset of questions, answers and rationales. Experimental results show that indirect supervision of program learning via answer rationales is a promising strategy for inducing arithmetic programs. | ['Wang Ling', 'Chris Dyer', 'Phil Blunsom', 'Dani Yogatama'] | 2017-07-01 | null | null | null | acl-2017-7 | ['program-induction'] | ['computer-code'] | [ 2.68398941e-01 1.93134129e-01 -2.13949651e-01 -8.82754982e-01
-9.58351433e-01 -1.02006125e+00 3.64757478e-01 4.78315592e-01
-7.46226981e-02 5.11415541e-01 -5.64666130e-02 -1.10589325e+00
1.40708029e-01 -1.38201594e+00 -1.00803709e+00 1.32263646e-01
4.12306450e-02 3.82729918e-01 3.32812041e-01 -3.86146933e-01
6.46324158e-01 1.03417836e-01 -1.45581484e+00 6.40923142e-01
1.21073973e+00 5.02649605e-01 2.00320166e-02 9.72548485e-01
-7.61872053e-01 1.43551695e+00 -5.64671993e-01 -7.40172029e-01
6.85585961e-02 -6.44742846e-01 -1.23695207e+00 -4.42227215e-01
7.80327857e-01 -4.22961712e-01 2.33974710e-01 1.16628850e+00
-1.63019538e-01 7.85363838e-03 5.75226247e-01 -1.21297753e+00
-8.32462430e-01 8.84901464e-01 -1.78014830e-01 -9.60148424e-02
1.04967308e+00 3.95049483e-01 1.61000967e+00 -8.17508221e-01
4.30782586e-01 1.22596586e+00 7.21781433e-01 5.95342517e-01
-1.62492108e+00 -5.01385570e-01 -3.03502232e-01 -2.22461186e-02
-9.61822689e-01 -2.28640437e-01 6.49901807e-01 -5.01681089e-01
9.84418452e-01 2.87630081e-01 6.52870834e-01 2.55146921e-01
-5.34940651e-03 8.23492110e-01 8.20446134e-01 -5.99401593e-01
1.82416528e-01 2.48912945e-01 9.83119309e-01 1.40996671e+00
1.72729671e-01 -2.04086393e-01 -1.32893354e-01 -3.74435782e-01
6.41917884e-01 -3.32068831e-01 -2.11918995e-01 -3.01482886e-01
-9.50709105e-01 1.22649705e+00 2.46310696e-01 1.98550336e-03
1.58061430e-01 4.38548297e-01 2.81265855e-01 9.29692328e-01
-2.86076069e-01 1.03021562e+00 -6.17628515e-01 -6.15752628e-03
-7.57965088e-01 6.86777055e-01 1.28352213e+00 9.37907994e-01
1.03628194e+00 -1.54269719e-02 -1.10450618e-01 2.74192244e-01
1.76556706e-01 4.30332661e-01 3.07708979e-02 -1.11702836e+00
5.91765642e-01 9.15320039e-01 9.12511498e-02 -1.12049484e+00
-1.73359677e-01 1.50227889e-01 -2.80480504e-01 1.53883234e-01
9.76170421e-01 -1.55327275e-01 -4.49135631e-01 1.84884572e+00
1.42115951e-01 -5.25248647e-02 -1.81207396e-02 5.09596467e-01
9.86188293e-01 9.52322066e-01 -6.50875047e-02 4.89341691e-02
1.26314938e+00 -9.72221196e-01 -4.03138518e-01 -7.37309456e-02
1.40012598e+00 -4.87920433e-01 1.86246979e+00 3.37179720e-01
-1.30136871e+00 -5.77803075e-01 -8.21689069e-01 -3.21284294e-01
-1.71660081e-01 -2.25028582e-02 1.14219809e+00 7.22925603e-01
-1.08761704e+00 4.57570970e-01 -3.17846477e-01 1.86861545e-01
1.62820369e-01 3.59729499e-01 -1.85969040e-01 -1.37812063e-01
-1.17968345e+00 8.20995629e-01 3.17201346e-01 -2.24472344e-01
-5.38863659e-01 -1.15047479e+00 -1.26085162e+00 1.58103809e-01
5.25519490e-01 -7.88410842e-01 1.53525019e+00 -7.79473364e-01
-1.45721114e+00 1.12281096e+00 -2.66879439e-01 -5.66032469e-01
-5.57716296e-04 -1.15008496e-01 3.00589688e-02 8.84761214e-02
-1.61292055e-03 5.91948330e-01 6.46133244e-01 -7.84055829e-01
-2.93914497e-01 -1.24045201e-01 8.99605930e-01 -4.39813286e-01
-1.62285522e-01 5.64411044e-01 -4.94120456e-02 -3.39169979e-01
-5.21452678e-03 -7.53231883e-01 -3.39859694e-01 5.77255338e-02
-2.99158186e-01 -6.76269829e-01 1.39744788e-01 -6.33181036e-01
1.42403305e+00 -1.86101997e+00 1.11285619e-01 2.88639247e-01
3.28648120e-01 2.02304184e-01 -3.63572091e-01 8.11498687e-02
-2.44018555e-01 4.29171294e-01 -3.54713410e-01 3.80441487e-01
4.18287605e-01 1.08770847e-01 -6.78850114e-01 -3.63254100e-02
6.04498208e-01 1.30135524e+00 -1.04114294e+00 -4.58609194e-01
4.30928394e-02 -4.00272220e-01 -1.36917472e+00 6.33839130e-01
-1.10660911e+00 7.36753419e-02 -4.10983950e-01 4.01765138e-01
5.01026273e-01 -4.16471124e-01 -1.07460171e-01 2.10820541e-01
1.88829958e-01 7.31200814e-01 -1.18973696e+00 1.58346415e+00
-7.91113496e-01 5.92372000e-01 -1.76142707e-01 -1.13862646e+00
9.97422755e-01 -5.01741655e-02 -2.75072902e-01 -3.96896631e-01
2.26230156e-02 3.89681667e-01 1.13670886e-01 -9.14760292e-01
5.45808494e-01 -3.79775405e-01 -6.54005349e-01 6.60257638e-01
1.30447093e-03 -1.08242667e+00 6.38789475e-01 4.02440846e-01
1.23771775e+00 5.55293858e-02 3.10126424e-01 -1.85074016e-01
9.66275752e-01 3.61357272e-01 3.17244917e-01 8.26120675e-01
1.90280646e-01 2.87300229e-01 1.04009306e+00 -6.94975615e-01
-9.95911181e-01 -1.27129090e+00 3.48294467e-01 1.07881725e+00
-4.32704091e-01 -7.10880876e-01 -7.56772220e-01 -7.23026156e-01
-1.28708631e-01 1.16951895e+00 -3.32366198e-01 -2.12447211e-01
-1.08156025e+00 -4.94032726e-02 7.17061758e-01 5.19724786e-01
3.33524436e-01 -1.03938365e+00 -4.48350489e-01 1.22429477e-02
-2.40697205e-01 -8.38494658e-01 -5.08507490e-01 1.78258687e-01
-7.95365393e-01 -1.29413116e+00 -2.29733005e-01 -1.01654613e+00
1.08885837e+00 9.12778545e-03 1.81831825e+00 6.59546554e-01
-1.19848646e-01 2.93425828e-01 1.09021636e-02 -2.20821247e-01
-8.00723910e-01 6.39676377e-02 -4.97538507e-01 -5.08611798e-01
6.16503894e-01 -4.82320935e-01 1.01764031e-01 -2.84150392e-01
-8.83877099e-01 -1.95988566e-02 2.24063173e-01 6.55337989e-01
2.44910449e-01 -9.27371457e-02 3.21515173e-01 -1.22760582e+00
8.25717270e-01 -2.72162229e-01 -1.17808568e+00 5.18991351e-01
-6.12106621e-02 6.70765936e-01 1.12663364e+00 -3.18572938e-01
-8.34200442e-01 1.49585277e-01 -2.31377602e-01 2.39414290e-01
-1.82842955e-01 7.39020050e-01 -7.05872625e-02 9.89997908e-02
9.24103320e-01 1.75553069e-01 -3.39094363e-02 2.78693866e-02
6.99785113e-01 1.39140964e-01 7.09477425e-01 -1.39047015e+00
1.30039656e+00 -1.71648785e-01 -8.61788914e-02 -4.76656467e-01
-1.06343651e+00 -1.36343479e-01 -4.40884978e-01 2.26085708e-01
7.60043144e-01 -6.34873629e-01 -7.83567011e-01 1.06905460e-01
-1.69026554e+00 -5.68451345e-01 -2.42505565e-01 4.67560515e-02
-5.28262198e-01 2.88236380e-01 -4.82477248e-01 -5.84291041e-01
-2.02114344e-01 -1.19877529e+00 6.61752582e-01 1.68506801e-01
-9.98784482e-01 -1.10178149e+00 6.78534731e-02 3.38212788e-01
2.60221541e-01 7.81563520e-02 1.91028464e+00 -5.87516665e-01
-9.83519673e-01 -1.69483528e-01 -2.02407598e-01 5.49730897e-01
-9.38738361e-02 3.15499008e-01 -4.03019786e-01 3.97404611e-01
1.58748075e-01 -8.09546769e-01 4.99885321e-01 -9.76254418e-02
1.43011439e+00 -5.48817337e-01 2.98291355e-01 4.29852754e-01
1.34599280e+00 -2.78860658e-01 7.10668981e-01 -1.91292930e-02
2.96496689e-01 5.94918787e-01 2.09964201e-01 1.21633969e-01
6.44783974e-01 1.56735450e-01 8.08747709e-02 3.02259505e-01
8.96656588e-02 -4.97755945e-01 5.52336752e-01 7.99017370e-01
2.66298473e-01 4.25729930e-01 -1.03833783e+00 5.70439816e-01
-1.30191255e+00 -1.08398235e+00 -7.90092170e-01 2.14786696e+00
1.50231659e+00 7.00392574e-02 4.77234051e-02 3.16941112e-01
3.76575708e-01 -1.78792235e-02 -4.47865278e-02 -6.89223826e-01
3.02108526e-01 9.73327518e-01 -1.43501148e-01 8.46675754e-01
-5.82621157e-01 9.72496033e-01 6.72452450e+00 6.07494652e-01
-9.03858304e-01 -3.28126043e-01 2.97015697e-01 2.43381754e-01
-1.13061082e+00 2.39993811e-01 -6.87833250e-01 1.62218273e-01
8.81597757e-01 -3.29723090e-01 6.75942838e-01 8.99251997e-01
-1.57121092e-01 -2.43306562e-01 -1.70996356e+00 5.80633700e-01
-1.70375541e-01 -1.53766513e+00 2.03704253e-01 -7.28156507e-01
7.09986925e-01 -7.19921947e-01 -4.18785997e-02 7.54725099e-01
8.09230149e-01 -1.28225136e+00 6.66644752e-01 3.11272860e-01
4.70560580e-01 -5.84002137e-01 3.37027073e-01 5.20522654e-01
-1.12449145e+00 -1.30472839e-01 -3.61245781e-01 -6.73631549e-01
-1.57790631e-01 7.15027571e-01 -6.28324866e-01 1.60171255e-01
3.74954604e-02 2.59464860e-01 -8.04943442e-01 7.48420238e-01
-9.90580320e-01 6.11429930e-01 -2.09623292e-01 -5.94998181e-01
2.29111508e-01 -2.40798786e-01 -8.96555334e-02 1.08463085e+00
2.29668692e-01 4.79512066e-01 8.71454701e-02 1.46801436e+00
-1.22978836e-01 6.40479475e-02 -7.78520465e-01 -3.12245160e-01
3.50168288e-01 1.04908133e+00 -4.50163960e-01 -5.76151609e-01
-5.99212229e-01 5.03443003e-01 6.96632504e-01 1.91649258e-01
-1.06665015e+00 -8.79492700e-01 4.16101605e-01 -5.91739491e-02
-1.06439546e-01 -4.86766994e-01 -7.42315412e-01 -1.36268461e+00
3.31802458e-01 -1.47364974e+00 3.32822233e-01 -9.42321301e-01
-8.65945041e-01 1.37032300e-01 -4.80683781e-02 -5.57620168e-01
-5.22320390e-01 -8.38113248e-01 -9.54137683e-01 9.13767397e-01
-1.36722195e+00 -5.66800535e-01 -2.48613596e-01 5.32130182e-01
3.36292267e-01 2.21072033e-01 1.15207469e+00 8.85961205e-02
-1.94411486e-01 5.82969606e-01 -7.18687773e-01 3.40231121e-01
3.15818787e-01 -1.40464389e+00 3.68264079e-01 7.19367623e-01
2.06682444e-01 1.18427086e+00 8.75768542e-01 -1.75726697e-01
-1.81959605e+00 -8.42723370e-01 1.30850387e+00 -7.58234560e-01
1.16097486e+00 -3.85358334e-01 -1.06310010e+00 7.11457491e-01
1.39872298e-01 -1.30331814e-01 7.37684965e-01 1.80463627e-01
-8.55480850e-01 1.41034946e-01 -8.33017290e-01 8.54039848e-01
7.89390206e-01 -9.80286479e-01 -1.28051531e+00 3.72729689e-01
9.99857903e-01 -3.06113064e-01 -5.13290405e-01 2.43584380e-01
2.17932731e-01 -7.58412063e-01 8.73985112e-01 -9.74986911e-01
1.19887638e+00 -3.91227543e-01 -1.32939473e-01 -1.11240578e+00
2.79854894e-01 -8.21673512e-01 -1.40515253e-01 1.17654765e+00
7.25962579e-01 -4.31101829e-01 8.97617817e-01 1.09592175e+00
3.73333432e-02 -5.79024315e-01 -1.65713564e-01 -6.37766123e-01
6.54102027e-01 -7.47428894e-01 7.36597836e-01 8.20018888e-01
3.99357021e-01 5.43493152e-01 1.15925193e-01 -9.43915695e-02
4.15611178e-01 6.48145258e-01 1.22272551e+00 -1.16554785e+00
-6.39400661e-01 -5.63571393e-01 1.46675900e-01 -1.31726706e+00
6.88578904e-01 -1.40450144e+00 1.39269918e-01 -1.39136159e+00
1.91590041e-02 -5.64650297e-01 4.40052450e-01 5.15825272e-01
-3.56324971e-01 -2.72702187e-01 -2.95121316e-02 -5.11942506e-01
-5.49848795e-01 1.73502207e-01 1.17769182e+00 -3.21222991e-01
1.11359321e-01 8.41625482e-02 -8.57388020e-01 1.11978674e+00
7.11312711e-01 -5.31880200e-01 -6.34793878e-01 -7.35548794e-01
1.00080931e+00 5.55896461e-01 4.13555324e-01 -5.85612059e-01
4.43767697e-01 -4.13246721e-01 -3.72452736e-01 -4.99004543e-01
-1.58794746e-01 -4.44619387e-01 -4.30326819e-01 6.28153384e-01
-7.76600242e-01 2.35774815e-01 1.31415591e-01 3.01960092e-02
-3.79681140e-01 -1.12057829e+00 5.68805337e-01 -5.59486091e-01
-5.70732236e-01 -3.00364960e-02 -3.98471475e-01 6.89317286e-01
6.36670947e-01 7.33966604e-02 -2.27922261e-01 -3.46174836e-01
-3.39464933e-01 3.92196596e-01 3.44634920e-01 1.64950714e-02
6.11453831e-01 -1.07060134e+00 -5.88431954e-01 1.96639597e-01
9.11603123e-02 1.93231836e-01 -3.88954163e-01 4.19507980e-01
-8.62375021e-01 4.77495760e-01 1.10347666e-01 -2.38965333e-01
-1.32929158e+00 5.46514213e-01 1.39210582e-01 -5.98681688e-01
3.03541943e-02 1.10508871e+00 1.55217350e-01 -1.30978370e+00
2.39928216e-01 -9.64425981e-01 8.16730596e-03 -2.39255443e-01
6.95234716e-01 -1.15863763e-01 -2.50820726e-01 2.76004881e-01
2.31791530e-02 6.05606616e-01 1.53468475e-01 2.39545494e-01
1.37314081e+00 4.60731417e-01 -8.78593981e-01 1.23153940e-01
1.29123878e+00 1.59871101e-01 -5.32810211e-01 -3.66421074e-01
5.67429245e-01 -3.96655470e-01 -7.20388830e-01 -4.31819767e-01
-6.37175143e-01 1.15769994e+00 -4.06995177e-01 2.24945813e-01
8.73251557e-01 8.74498393e-03 7.83652306e-01 1.25734246e+00
5.19508898e-01 -6.42458081e-01 6.32648468e-01 1.01292801e+00
8.12070131e-01 -1.20448542e+00 -3.95974040e-01 -7.58540273e-01
5.72604686e-02 1.46989143e+00 7.72969782e-01 -2.29460850e-01
3.35603654e-01 4.68413740e-01 -2.96356946e-01 -1.79231122e-01
-9.09686387e-01 -2.36256626e-02 -1.01813786e-02 3.34375232e-01
7.14698255e-01 -1.94105446e-01 -1.85074151e-01 9.39333558e-01
-7.08917439e-01 3.15786213e-01 7.91976631e-01 8.91477823e-01
-5.81869662e-01 -1.30263877e+00 -4.77123916e-01 4.16335970e-01
-3.17587614e-01 -5.72482049e-01 -2.72983223e-01 5.33234119e-01
-2.52950519e-01 8.62512052e-01 -2.48444721e-01 -2.24978477e-01
8.62654448e-02 3.42017770e-01 9.41013217e-01 -1.21912277e+00
-6.81536376e-01 -1.06891334e+00 4.05650288e-01 -2.45540529e-01
-1.11896992e-01 -3.03855836e-01 -1.85819340e+00 -6.54700398e-01
-2.32894614e-01 3.45843852e-01 2.51869351e-01 1.07113278e+00
-7.95602202e-02 4.45202440e-01 5.50882041e-01 -9.72662196e-02
-1.23484111e+00 -3.16594601e-01 -2.22858414e-03 4.01578754e-01
1.54183328e-01 1.29047856e-01 -3.97714555e-01 3.75722736e-01] | [9.482827186584473, 7.39417028427124] |
d5ba98b0-d89c-4730-919c-bb7e26e13709 | anticipating-human-actions-by-correlating | 2105.12414 | null | https://arxiv.org/abs/2105.12414v1 | https://arxiv.org/pdf/2105.12414v1.pdf | Anticipating human actions by correlating past with the future with Jaccard similarity measures | We propose a framework for early action recognition and anticipation by correlating past features with the future using three novel similarity measures called Jaccard vector similarity, Jaccard cross-correlation and Jaccard Frobenius inner product over covariances. Using these combinations of novel losses and using our framework, we obtain state-of-the-art results for early action recognition in UCF101 and JHMDB datasets by obtaining 91.7 % and 83.5 % accuracy respectively for an observation percentage of 20. Similarly, we obtain state-of-the-art results for Epic-Kitchen55 and Breakfast datasets for action anticipation by obtaining 20.35 and 41.8 top-1 accuracy respectively. | ['Samitha Herath', 'Basura Fernando'] | 2021-05-26 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Fernando_Anticipating_Human_Actions_by_Correlating_Past_With_the_Future_With_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Fernando_Anticipating_Human_Actions_by_Correlating_Past_With_the_Future_With_CVPR_2021_paper.pdf | cvpr-2021-1 | ['action-anticipation'] | ['computer-vision'] | [ 2.31510460e-01 -2.51880646e-01 -4.08316582e-01 -4.76716071e-01
-9.83729899e-01 -1.38168335e-01 5.78504384e-01 2.65993506e-01
-6.99770749e-01 7.62898803e-01 5.44627726e-01 4.25394595e-01
-4.55449104e-01 -4.65376288e-01 -5.50652921e-01 -4.95569646e-01
-7.28595197e-01 -6.95785508e-02 3.92388731e-01 -8.52958634e-02
3.03367555e-01 2.19487041e-01 -1.78037834e+00 9.03453648e-01
4.70624030e-01 1.26275575e+00 -6.38868332e-01 1.24779344e+00
5.89321733e-01 1.41928995e+00 -4.50508654e-01 -5.78084230e-01
3.21117967e-01 -3.51567447e-01 -1.01094973e+00 -2.02301487e-01
3.98561239e-01 -2.74045199e-01 -6.97943747e-01 5.35046995e-01
5.22264361e-01 2.41580665e-01 4.97940391e-01 -1.35010803e+00
-5.50637305e-01 5.70481658e-01 -6.62807763e-01 5.59517026e-01
8.03879976e-01 1.03689402e-01 1.25908446e+00 -6.32689536e-01
5.86247087e-01 1.06342733e+00 9.36031461e-01 5.97735703e-01
-9.97128963e-01 -5.06525755e-01 7.52871260e-02 8.35235178e-01
-1.21406543e+00 -3.48436177e-01 2.68949389e-01 -4.60231483e-01
1.41667199e+00 3.83039266e-01 9.12078559e-01 1.07643044e+00
4.49490339e-01 1.47903979e+00 8.94305944e-01 -3.58293146e-01
1.09165192e-01 -4.86865491e-01 2.25212172e-01 4.96819526e-01
-2.09211335e-01 4.87000525e-01 -8.57071936e-01 5.19274175e-02
8.87496471e-02 1.69716403e-01 2.07837254e-01 -8.17753524e-02
-1.35802901e+00 4.34668690e-01 1.53482780e-01 2.34552696e-01
-5.58661282e-01 4.10287499e-01 8.12930942e-01 2.77482927e-01
3.84691060e-01 1.74068019e-01 -4.35408890e-01 -9.82994139e-01
-7.59938478e-01 3.89156610e-01 4.60243970e-01 7.47515082e-01
3.01054120e-01 -2.31236935e-01 -6.04894519e-01 6.66061819e-01
9.48915854e-02 5.17521501e-01 6.67415023e-01 -1.07216120e+00
3.73665005e-01 5.70144117e-01 2.16161072e-01 -9.11235392e-01
-6.49097204e-01 -2.36770093e-01 -7.18552649e-01 1.64536849e-01
4.06888753e-01 -9.18804854e-02 -5.81851006e-01 1.82056808e+00
2.72631913e-01 4.83683765e-01 5.09977758e-01 4.98076528e-01
3.85194063e-01 2.37762019e-01 2.46447623e-01 -1.77378267e-01
1.06445038e+00 -9.92563546e-01 -6.53529227e-01 -1.94335103e-01
1.06265080e+00 -6.29069626e-01 6.91756785e-01 5.04539669e-01
-8.61861527e-01 -8.20694923e-01 -8.73117626e-01 2.26875246e-01
-1.81966335e-01 5.45203686e-01 9.49289739e-01 2.33311206e-01
-6.80655718e-01 1.18841827e+00 -1.18271637e+00 -5.39962471e-01
7.39619315e-01 8.37417915e-02 -4.70132262e-01 -1.13434782e-02
-1.08469367e+00 7.49621809e-01 3.29034567e-01 -1.42842624e-02
-8.78249466e-01 -8.89778316e-01 -5.45337737e-01 -2.97797143e-01
2.64978874e-02 -5.08451104e-01 1.08942759e+00 -7.26650953e-01
-1.30771065e+00 9.33285952e-01 1.95388928e-01 -1.17942357e+00
8.11503172e-01 -1.00897408e+00 -9.22309041e-01 9.26929638e-02
-6.70866966e-02 6.13529801e-01 2.63078630e-01 -2.19195068e-01
-1.04543817e+00 -5.11924446e-01 1.39146909e-01 8.87767151e-02
-1.27681106e-01 1.01144388e-01 1.55417815e-01 -6.56798244e-01
2.89754495e-02 -1.21114743e+00 -5.36878072e-02 1.56075880e-01
-2.27889419e-01 -3.29109937e-01 6.52300119e-01 -8.08775961e-01
1.45196474e+00 -2.29433084e+00 -1.67400986e-01 -3.33246410e-01
-1.36653304e-01 5.75389922e-01 -3.05279672e-01 4.51485395e-01
-5.39401531e-01 -3.41889620e-01 1.90980136e-01 -7.33781559e-03
-3.35965003e-03 6.71826899e-02 -1.06007487e-01 4.93048280e-01
3.41770053e-01 9.76382196e-01 -1.22343802e+00 -2.42452472e-01
3.22726369e-01 4.04984593e-01 -5.53929329e-01 1.05230220e-01
2.03392997e-01 1.52476013e-01 -4.74381238e-01 5.91797769e-01
2.71324217e-01 9.08459276e-02 -4.46771942e-02 -1.43215045e-01
1.62522346e-01 -1.00424796e-01 -1.01557612e+00 1.97922277e+00
-1.07972197e-01 7.93862283e-01 -8.83431137e-01 -1.18719256e+00
9.80610967e-01 2.07518354e-01 1.05788398e+00 -5.77108204e-01
-2.87488531e-02 -2.14428678e-01 -8.28876123e-02 -7.80601501e-01
2.71779150e-01 7.04873279e-02 -1.71350688e-01 2.45280921e-01
-1.02145255e-01 5.98577380e-01 5.76831222e-01 2.09935591e-01
1.70000529e+00 4.29096550e-01 3.87268275e-01 6.30106032e-02
6.90167725e-01 -1.51531592e-01 7.71248698e-01 6.93725049e-01
-6.92201734e-01 4.77427810e-01 4.28842992e-01 -7.08834767e-01
-6.82065785e-01 -1.23493385e+00 1.41778007e-01 1.02351546e+00
-6.02833092e-01 -8.13319504e-01 -4.56471235e-01 -8.77718985e-01
3.03436935e-01 7.65163004e-01 -1.03279150e+00 -7.80335903e-01
-5.54690301e-01 -7.40031838e-01 8.21581185e-01 1.09544790e+00
7.48933911e-01 -9.21273172e-01 -8.01297545e-01 2.09025070e-01
-1.91545710e-01 -1.07814121e+00 -1.46708265e-01 -2.85362393e-01
-1.06673086e+00 -1.51038599e+00 -7.23784268e-01 -1.32945672e-01
8.37421417e-02 -1.18500009e-01 1.10461986e+00 -2.90413290e-01
-7.23061144e-01 5.86322784e-01 -7.24838793e-01 -1.91659987e-01
-1.54594511e-01 -3.84065926e-01 3.16303939e-01 2.75537997e-01
6.26536071e-01 -6.42933726e-01 -8.04097295e-01 1.37173399e-01
-4.49216962e-01 -2.78860122e-01 6.79258108e-01 8.42220545e-01
5.90821683e-01 -2.92544812e-01 4.39534724e-01 -3.17450494e-01
6.54572397e-02 -1.57824054e-01 -1.67412519e-01 5.35128355e-01
-5.28631985e-01 3.82493176e-02 2.27042049e-01 -5.59676886e-01
-9.33611155e-01 3.67193520e-01 3.76107655e-02 -4.05228645e-01
-3.12233448e-01 9.08694565e-02 2.85303742e-01 2.75782108e-01
7.39759147e-01 3.93129617e-01 -2.31590807e-01 -5.46740592e-01
4.71788168e-01 4.88357663e-01 6.38366699e-01 -1.01469211e-01
1.09629884e-01 6.45813823e-01 1.07505828e-01 -5.71382940e-01
-1.24712133e+00 -6.80935442e-01 -1.05940926e+00 -3.26159894e-01
1.13031781e+00 -8.94783914e-01 -1.13137901e+00 9.96658206e-01
-7.60764778e-01 -5.35422787e-02 -6.60261214e-01 8.85788918e-01
-9.88075852e-01 4.40320015e-01 -4.24891621e-01 -1.06497753e+00
-4.80160207e-01 -5.02373636e-01 1.02433789e+00 8.95614251e-02
-5.40490866e-01 -5.99150717e-01 6.05242550e-01 5.16773820e-01
-5.32741286e-02 6.58593833e-01 1.93869039e-01 -5.01084507e-01
-1.54401839e-01 -5.35601735e-01 -9.12580639e-02 7.15791523e-01
4.19809781e-02 1.86079413e-01 -7.84430742e-01 -5.45082577e-02
-7.46015191e-01 -5.39969206e-01 1.20646036e+00 3.39107990e-01
9.91402924e-01 -6.08016644e-03 -4.05055970e-01 3.80532324e-01
1.09794319e+00 4.27899092e-01 1.03720844e+00 4.24092919e-01
4.13929135e-01 3.50767821e-01 1.32160461e+00 1.12571573e+00
2.38000661e-01 6.89995110e-01 8.46331045e-02 4.83451098e-01
-1.77029669e-01 -2.43916720e-01 6.26193643e-01 3.18981916e-01
-3.08476090e-01 5.10030501e-02 -1.03848839e+00 7.40824521e-01
-2.22463489e+00 -1.42938733e+00 -1.51511282e-01 2.10155392e+00
7.44722962e-01 2.12732255e-01 4.23597842e-01 5.19280672e-01
3.97137076e-01 3.04101497e-01 -4.53124970e-01 -2.31578022e-01
-6.34954423e-02 7.52892941e-02 4.54206973e-01 7.28292912e-02
-1.69639075e+00 8.93653274e-01 6.54680395e+00 7.79020548e-01
-5.27868509e-01 -1.59294620e-01 6.26390457e-01 -4.12604570e-01
7.65598476e-01 -4.78040799e-02 -7.38802314e-01 4.54970866e-01
1.20848370e+00 -9.37282071e-02 7.78734162e-02 9.52947438e-01
7.50390962e-02 -5.81807435e-01 -1.31641293e+00 1.31770420e+00
9.11400542e-02 -1.19884932e+00 -4.44349647e-01 -5.73213160e-01
7.72077084e-01 1.53484464e-01 5.88182285e-02 3.57182980e-01
3.59024286e-01 -6.04449749e-01 4.57958728e-01 1.31026864e+00
5.21395504e-01 -8.54115605e-01 5.22511482e-01 -1.53673152e-02
-1.19830525e+00 -4.60356921e-01 -1.82192251e-01 -2.46746421e-01
2.46701121e-01 6.39835417e-01 -4.80291158e-01 4.58351195e-01
9.08603728e-01 1.44323862e+00 -8.60953450e-01 1.24450302e+00
-1.18695006e-01 4.31710601e-01 -1.38881326e-01 -7.20514581e-02
2.73265481e-01 3.52125704e-01 4.07710254e-01 1.06528342e+00
1.86465815e-01 1.23981722e-01 -9.73159745e-02 8.88325572e-02
4.37835097e-01 -4.98634391e-02 -4.25936282e-01 -2.28550360e-01
-7.56224319e-02 6.65396750e-01 -1.68309450e-01 -5.60552239e-01
-1.91798925e-01 1.17078340e+00 1.63017899e-01 -1.62334985e-03
-1.07051051e+00 -5.24358630e-01 1.21560299e+00 -4.55125362e-01
4.02939856e-01 -6.12045862e-02 6.71750754e-02 -1.15981519e+00
1.44733936e-01 -7.81816900e-01 9.65077519e-01 -7.69622922e-01
-1.10332859e+00 2.65279233e-01 1.36947036e-01 -1.50203431e+00
-4.26908463e-01 -5.72294295e-01 -5.15094399e-01 7.63695389e-02
-1.02031565e+00 -9.09471452e-01 -2.12765872e-01 4.41278785e-01
3.57467175e-01 -3.64598483e-01 9.58073080e-01 4.58666116e-01
-6.38633132e-01 7.45292783e-01 3.99799734e-01 3.53414565e-01
7.24125087e-01 -1.04158247e+00 5.44304430e-01 5.86692750e-01
3.23581874e-01 -1.53061882e-01 5.81324816e-01 -5.73504090e-01
-1.18058014e+00 -1.10163677e+00 1.20362592e+00 -7.63619721e-01
8.48008096e-01 -1.31033361e-02 -3.74215841e-01 7.15703249e-01
-2.25383341e-01 3.15465182e-01 7.50762641e-01 1.75695390e-01
-3.88078332e-01 -5.25835454e-01 -1.06541073e+00 5.26099324e-01
1.59038675e+00 -4.47610140e-01 -4.81060058e-01 7.56277323e-01
3.01123321e-01 -2.20116571e-01 -1.31327140e+00 6.91388071e-01
1.18168604e+00 -1.26179087e+00 1.31697404e+00 -1.33641934e+00
3.63814712e-01 5.49743399e-02 -4.49199319e-01 -8.94618392e-01
-4.88233089e-01 -6.17084026e-01 -2.70599663e-01 1.07394195e+00
3.91730994e-01 -2.76881218e-01 9.46391642e-01 4.73721862e-01
6.63478225e-02 -1.06699288e+00 -9.42777157e-01 -1.20387650e+00
-2.54372209e-01 -8.83743584e-01 2.94274032e-01 5.71883798e-01
1.61618516e-01 -8.30318704e-02 -6.59072280e-01 -3.28620613e-01
4.60354626e-01 1.86559185e-02 7.99213171e-01 -9.46087122e-01
-3.37081939e-01 -2.05026165e-01 -1.14689589e+00 -7.63409853e-01
6.05694056e-02 -6.74899578e-01 -3.21460456e-01 -1.26673579e+00
3.15014839e-01 1.33205816e-01 -9.10467505e-01 5.60084701e-01
7.43469894e-02 3.69408518e-01 2.60134041e-01 -1.39899448e-01
-1.15409780e+00 7.66784668e-01 8.22767675e-01 -1.65135756e-01
8.35443884e-02 1.95179820e-01 -2.34100506e-01 8.32964420e-01
7.51847327e-01 -5.75550079e-01 -2.37088844e-01 -1.48248106e-01
1.45508170e-01 -4.20583226e-02 4.43204790e-01 -1.56096399e+00
-1.47807151e-01 -1.85625717e-01 5.67932367e-01 -8.56659830e-01
5.87461531e-01 -4.72554088e-01 -3.12141869e-02 9.38734353e-01
-7.78846204e-01 8.64550620e-02 1.28169790e-01 9.36773181e-01
-1.61318958e-01 1.54352173e-01 8.00586998e-01 2.04317316e-01
-1.09150600e+00 1.40893430e-01 -2.90239871e-01 3.30547094e-01
1.51041567e+00 -1.03671513e-01 -3.09890330e-01 -2.62267530e-01
-1.20835125e+00 1.68281719e-01 -1.50585711e-01 7.74962842e-01
7.74009347e-01 -1.77852786e+00 -9.57778275e-01 2.90217362e-02
4.28975761e-01 -1.11703527e+00 6.18109703e-01 1.36032450e+00
-2.64876693e-01 4.02848363e-01 -4.89896625e-01 -5.99778533e-01
-1.76891768e+00 4.39767331e-01 1.05256625e-01 -5.03274918e-01
-5.23153961e-01 1.11256361e+00 -3.12496692e-01 -5.69370091e-02
3.87365997e-01 -5.00770986e-01 -1.94215864e-01 2.91147649e-01
9.65538919e-01 9.62176144e-01 -2.66351968e-01 -5.85797548e-01
-1.01930273e+00 5.71378648e-01 -7.00864717e-02 1.46044627e-01
1.38999772e+00 2.50878185e-01 3.37786973e-01 5.03346205e-01
1.39034617e+00 -7.59814918e-01 -1.32191491e+00 -1.27271354e-01
3.63537848e-01 -8.04605901e-01 -2.12571815e-01 -1.06968760e+00
-8.90295923e-01 8.03789496e-01 1.27348447e+00 -2.50243515e-01
1.28087234e+00 -1.61598891e-01 8.12024295e-01 7.06038237e-01
3.42316069e-02 -1.47764015e+00 3.91345084e-01 6.25769973e-01
9.53335822e-01 -1.22599804e+00 3.40732425e-01 1.75083298e-02
-9.12350476e-01 9.00453687e-01 3.14428002e-01 -4.66334462e-01
8.08598638e-01 -8.58340189e-02 -1.43481433e-01 4.38376032e-02
-1.09792936e+00 -3.25119942e-01 4.86906558e-01 5.98167896e-01
5.39370894e-01 2.00979009e-01 -5.53958774e-01 5.21860421e-01
-2.05305107e-02 3.38555783e-01 -4.15683500e-02 9.87563670e-01
-2.90687382e-01 -9.46765423e-01 2.13796034e-01 4.24177438e-01
-6.78401887e-01 6.23079874e-02 -4.20278549e-01 6.53508961e-01
6.57972097e-02 8.36297572e-01 1.13560639e-01 -9.98706579e-01
5.20374179e-01 3.41522068e-01 7.42270172e-01 -1.92726165e-01
-4.24567342e-01 -2.68033028e-01 6.34928763e-01 -1.28452790e+00
-8.49971473e-01 -1.24134469e+00 -1.13716197e+00 -1.24037668e-01
-1.00288823e-01 -2.68302590e-01 2.78894186e-01 9.94241297e-01
7.88730383e-01 4.40449089e-01 5.80459476e-01 -1.71416640e-01
-7.34739363e-01 -1.11234236e+00 -5.53002894e-01 6.96581125e-01
-3.99080627e-02 -7.67210841e-01 -1.21250441e-02 2.11375087e-01] | [8.13033676147461, 0.6182119250297546] |
a8201886-1158-4112-8437-afc0a22f5000 | chatlaw-open-source-legal-large-language | 2306.16092 | null | https://arxiv.org/abs/2306.16092v1 | https://arxiv.org/pdf/2306.16092v1.pdf | ChatLaw: Open-Source Legal Large Language Model with Integrated External Knowledge Bases | Large Language Models (LLMs) have shown the potential to revolutionize natural language processing tasks in various domains, sparking great interest in vertical-specific large models. However, unlike proprietary models such as BloombergGPT and FinGPT, which have leveraged their unique data accumulations to make strides in the finance domain, there hasn't not many similar large language models in the Chinese legal domain to facilitate its digital transformation. In this paper, we propose an open-source legal large language model named ChatLaw. Due to the importance of data quality, we carefully designed a legal domain fine-tuning dataset. Additionally, to overcome the problem of model hallucinations in legal data screening during reference data retrieval, we introduce a method that combines vector database retrieval with keyword retrieval to effectively reduce the inaccuracy of relying solely on vector database retrieval. Furthermore, we propose a self-attention method to enhance the ability of large models to overcome errors present in reference data, further optimizing the issue of model hallucinations at the model level and improving the problem-solving capabilities of large models. We also open-sourced our model and part of the data at https://github.com/PKU-YuanGroup/ChatLaw. | ['Li Yuan', 'Bohua Chen', 'Yang Yan', 'Zongjian Li', 'Jiaxi Cui'] | 2023-06-28 | null | null | null | null | ['retrieval'] | ['methodology'] | [-5.64305365e-01 -8.21136460e-02 -3.11535001e-01 -2.36793458e-01
-1.20182478e+00 -4.14808989e-01 4.62027043e-01 -4.52510603e-02
-4.24957901e-01 4.33957487e-01 5.50114989e-01 -5.34794390e-01
-2.42739588e-01 -6.55147552e-01 -4.53039020e-01 -7.24188015e-02
3.92359436e-01 4.44740176e-01 -5.28897084e-02 -4.13234293e-01
5.48438072e-01 3.79343569e-01 -1.00003970e+00 4.07868087e-01
1.34677613e+00 4.75095361e-01 3.31038922e-01 6.63146526e-02
-3.01396489e-01 1.21613789e+00 -7.08170772e-01 -6.32304490e-01
4.64244694e-01 2.16973461e-02 -7.02972770e-01 -1.60735101e-01
3.81881654e-01 -5.63413084e-01 -6.64695680e-01 9.84976172e-01
7.25133955e-01 9.21268016e-02 3.85488480e-01 -1.06798923e+00
-1.20279634e+00 6.30457938e-01 -6.14154518e-01 2.97744840e-01
2.70219147e-01 3.51647288e-01 1.01125348e+00 -9.00148809e-01
7.97821224e-01 1.43944216e+00 5.18084228e-01 3.37914407e-01
-9.72781718e-01 -1.05134988e+00 2.49311402e-01 4.01585162e-01
-1.55194426e+00 -4.83498722e-01 7.06050813e-01 -4.54942971e-01
1.25926459e+00 1.12360548e-02 4.15876955e-01 9.37636971e-01
1.62508726e-01 9.55780089e-01 6.94351256e-01 -2.18462154e-01
2.61603016e-02 2.32397273e-01 4.13427979e-01 4.83529896e-01
1.57093570e-01 -2.27586944e-02 -3.74151975e-01 -4.35143858e-01
5.85691392e-01 2.93267611e-02 -4.60943170e-02 4.23566205e-03
-8.11587751e-01 9.95804787e-01 1.68696418e-01 4.30371314e-01
-4.98600662e-01 -1.61187425e-01 2.03636184e-01 3.67947191e-01
4.86681104e-01 7.07738400e-01 -4.85348493e-01 -3.21870834e-01
-8.15455079e-01 5.74835598e-01 6.00866556e-01 9.08195972e-01
5.70909381e-01 1.30593097e-02 -2.57987678e-01 1.23748088e+00
3.51131022e-01 4.97486591e-01 8.42427075e-01 -1.03026879e+00
8.94304693e-01 9.52467203e-01 3.18681560e-02 -1.04209304e+00
-2.23246172e-01 -3.34354669e-01 -3.76811504e-01 -2.29356378e-01
1.51169345e-01 1.03777200e-01 -5.95508397e-01 1.57786012e+00
-2.26482041e-02 -1.66439474e-01 8.44974220e-02 6.41119242e-01
5.29936314e-01 5.94130874e-01 2.72420913e-01 7.52091929e-02
9.39416170e-01 -8.85427117e-01 -6.56182170e-01 -2.79977441e-01
1.03234375e+00 -8.08061242e-01 1.37380350e+00 3.43516469e-01
-1.04168737e+00 -5.59920549e-01 -7.38996089e-01 -3.77312362e-01
-4.33972836e-01 -7.36911148e-02 7.78704226e-01 5.23723781e-01
-6.69994771e-01 3.72212201e-01 -6.33594334e-01 -3.31760764e-01
5.15893757e-01 4.51304950e-02 -2.85893202e-01 -3.56951952e-01
-1.50622725e+00 1.20677888e+00 2.58021414e-01 -5.81788272e-02
-5.22694707e-01 -8.36183846e-01 -7.78634667e-01 2.08402619e-01
3.87907892e-01 -3.22614044e-01 1.17101181e+00 -4.66338992e-01
-8.46851885e-01 4.32417750e-01 -3.60719651e-01 -3.82227719e-01
3.71653736e-01 -2.14572012e-01 -6.66197896e-01 -1.76130220e-01
4.32825834e-01 5.68894744e-01 4.43607002e-01 -9.39100742e-01
-3.68055135e-01 -3.73869866e-01 1.09720200e-01 1.24424882e-01
-8.25431228e-01 3.65620643e-01 -8.44610333e-01 -7.77085185e-01
-2.49077991e-01 -5.93342960e-01 -5.02223253e-01 -3.59806538e-01
-1.12511940e-01 -3.32559258e-01 3.88173550e-01 -9.80788112e-01
1.95547044e+00 -2.27291083e+00 -3.09232086e-01 2.82560259e-01
2.97462970e-01 6.67158425e-01 -6.57024980e-01 7.87022769e-01
7.57080987e-02 4.68913436e-01 6.20556250e-02 -2.59121537e-01
5.02671115e-02 9.67763588e-02 -4.69040245e-01 1.02157861e-01
2.11679325e-01 9.72237408e-01 -5.85114419e-01 -5.33165157e-01
-9.81156603e-02 3.10127676e-01 -8.27697277e-01 -9.99861956e-02
-1.01002812e-01 -1.09529532e-02 -6.07602179e-01 7.83412337e-01
7.05479085e-01 -3.21436316e-01 -4.88598347e-02 1.32411584e-01
-1.10586584e-01 3.94489586e-01 -1.11454761e+00 1.60930645e+00
-5.31079710e-01 3.49137217e-01 -2.17150208e-02 -6.40387774e-01
7.03490138e-01 2.51398444e-01 4.15905297e-01 -1.22044766e+00
-1.06343664e-01 1.88586965e-01 1.53621778e-01 -7.32375622e-01
7.46960759e-01 -1.13805942e-02 7.03024119e-03 5.45342505e-01
-3.57255787e-01 1.97848957e-02 6.09140038e-01 6.70556188e-01
1.08389258e+00 -3.45331639e-01 2.78172307e-02 3.22658904e-02
1.04126886e-01 2.93881059e-01 6.34448826e-01 8.89969230e-01
-1.06051862e-01 4.49748605e-01 3.89550567e-01 -7.98605084e-02
-7.25333810e-01 -6.53293550e-01 -2.43622690e-01 8.99726927e-01
-1.64648548e-01 -9.57702279e-01 -4.24886614e-01 -3.75619888e-01
3.77751648e-01 8.89507711e-01 -2.16383860e-01 -3.54780465e-01
-4.06753093e-01 -7.42174387e-01 8.06786418e-01 6.15572691e-01
4.76801336e-01 -9.49954510e-01 -6.77709654e-02 3.63963187e-01
-3.20777953e-01 -9.45745707e-01 -6.79226160e-01 -2.41525501e-01
-7.59646237e-01 -1.09027576e+00 -5.12502551e-01 -4.66256976e-01
4.52143818e-01 2.43316948e-01 8.90792966e-01 2.28832215e-01
-1.69192240e-01 1.60280779e-01 -2.20259011e-01 -5.11135817e-01
-4.50400770e-01 2.20970184e-01 7.56970420e-02 -2.93760568e-01
7.67819822e-01 -1.74328595e-01 -9.46764350e-02 1.14013672e-01
-9.64001477e-01 -7.05075115e-02 6.25730991e-01 6.38002276e-01
1.03457980e-01 -2.84419749e-02 8.16614389e-01 -9.40433860e-01
1.39935088e+00 -6.03687167e-01 -6.32887363e-01 4.23906237e-01
-9.73275006e-01 -1.08517841e-01 5.01271307e-01 -5.71121097e-01
-9.37575698e-01 -5.59298217e-01 -2.03842372e-01 -4.29270566e-01
1.47473171e-01 1.03473234e+00 -1.30208388e-01 1.24363877e-01
7.21179545e-01 3.49817611e-02 4.84466255e-02 -7.53729224e-01
3.25096130e-01 1.11541390e+00 9.10653621e-02 -6.18056118e-01
7.54962802e-01 1.48016304e-01 -7.67818451e-01 -6.74319208e-01
-6.36990368e-01 -5.96898258e-01 -3.41657728e-01 1.60886005e-01
5.50755799e-01 -1.23642409e+00 -4.19547856e-01 3.80895823e-01
-1.06438458e+00 -2.78582692e-01 -1.41219735e-01 3.79533261e-01
-3.20833139e-02 6.99246705e-01 -8.63565981e-01 -6.86693490e-01
-2.65930831e-01 -1.04517210e+00 6.79638922e-01 9.16174054e-02
-4.84733015e-01 -8.40374291e-01 1.06276972e-02 8.18543911e-01
7.51961768e-01 -4.81261969e-01 1.21927559e+00 -8.60090494e-01
-5.11734903e-01 -6.41850054e-01 -3.27056468e-01 5.01252651e-01
-7.48785585e-02 -1.72143966e-01 -7.99986780e-01 -2.06097767e-01
4.10912782e-02 -3.46640199e-01 7.41894901e-01 9.28964838e-02
1.00832736e+00 -1.27998963e-01 -1.65409476e-01 3.55618775e-01
1.12689960e+00 5.03378451e-01 6.17591679e-01 4.29524601e-01
6.77357137e-01 4.49199498e-01 5.72857797e-01 4.56491679e-01
7.36743748e-01 5.47998846e-01 -2.27020174e-01 -1.53845444e-01
-4.27835546e-02 -4.87462848e-01 3.74058992e-01 7.61322677e-01
2.91498631e-01 -1.94619764e-02 -1.20319378e+00 5.94359398e-01
-1.85424578e+00 -8.84015679e-01 9.32003185e-02 1.96775126e+00
8.20506990e-01 1.82561591e-01 -4.08796400e-01 -3.34802181e-01
3.63992661e-01 4.63250764e-02 -7.05915391e-01 -2.03569666e-01
-3.53667051e-01 -2.92790048e-02 3.78220022e-01 5.40841579e-01
-6.94165409e-01 1.28776550e+00 6.16942787e+00 1.19634366e+00
-9.37110305e-01 1.77207410e-01 5.40234506e-01 -3.51933122e-01
-5.64856291e-01 7.99393579e-02 -1.05783343e+00 4.98859525e-01
1.10713673e+00 -5.45287907e-01 4.40489739e-01 9.50038970e-01
5.45941532e-01 -6.75475374e-02 -9.20709670e-01 9.49864984e-01
1.38497353e-01 -1.37898755e+00 3.73044461e-01 3.74464542e-01
5.50043404e-01 2.71854877e-01 -4.03216183e-02 8.09450388e-01
2.40307674e-01 -9.67250705e-01 4.89415199e-01 3.21502715e-01
4.36118096e-01 -4.87089396e-01 5.27109325e-01 6.12445533e-01
-8.42842937e-01 -4.82085615e-01 -6.30410194e-01 -2.51264215e-01
2.93303311e-01 2.83112198e-01 -6.31285548e-01 3.45222294e-01
3.88570130e-01 7.98604727e-01 -9.45124209e-01 9.42396283e-01
1.50524721e-01 5.15020132e-01 -2.55886793e-01 3.18637788e-01
1.42257676e-01 -1.62140056e-01 1.88602060e-01 7.98608541e-01
1.55539006e-01 1.41644804e-03 2.46112049e-01 9.29000914e-01
-1.53914765e-01 3.49072605e-01 -6.48155212e-01 -4.19406235e-01
7.79397130e-01 8.27877462e-01 -8.21169280e-03 -5.68558216e-01
-7.25920260e-01 5.09273350e-01 4.04266000e-01 5.33585787e-01
-6.90667391e-01 -3.42675656e-01 7.63759077e-01 4.47234005e-01
-8.67002755e-02 -1.08873881e-01 -3.59484106e-01 -1.48635626e+00
1.64490268e-01 -1.30236983e+00 4.54065889e-01 -9.44357336e-01
-1.34919679e+00 4.35599267e-01 6.77449852e-02 -1.00003159e+00
-2.57430226e-01 -3.32138956e-01 -3.31675738e-01 1.05260944e+00
-1.46476209e+00 -1.06039429e+00 2.42484421e-01 8.19737256e-01
6.28915191e-01 -4.05890971e-01 6.76488340e-01 7.99128830e-01
-6.89274609e-01 6.90444469e-01 2.75647312e-01 2.96674997e-01
1.11188352e+00 -6.65965259e-01 4.95719075e-01 8.48869503e-01
9.35654994e-03 1.40423775e+00 2.48191714e-01 -9.84773159e-01
-1.25311005e+00 -7.88906515e-01 1.33942497e+00 -7.07840800e-01
8.89210522e-01 -3.09938669e-01 -1.09026062e+00 7.66103983e-01
2.81467084e-02 -5.71713805e-01 8.01306129e-01 2.17320427e-01
-4.72693652e-01 -1.59339041e-01 -9.51254785e-01 7.15789080e-01
8.31791222e-01 -8.03547442e-01 -6.30789518e-01 5.05802035e-01
5.72167754e-01 -1.49510756e-01 -8.61470044e-01 1.40317068e-01
2.87968665e-01 -6.11887038e-01 9.86646235e-01 -8.95477772e-01
3.50671172e-01 1.40405325e-02 -8.81779864e-02 -1.08555615e+00
-3.91350508e-01 -3.21406275e-01 1.22938976e-01 1.35000455e+00
6.24995530e-01 -7.81957090e-01 5.25756478e-01 1.39459395e+00
-6.72851130e-02 -5.87010324e-01 -7.07506776e-01 -7.90321112e-01
4.59473133e-01 -8.49094570e-01 6.53273702e-01 1.00746214e+00
2.78182983e-01 2.25041375e-01 -4.01206762e-01 1.81905571e-02
2.55224973e-01 1.29596621e-01 7.32882380e-01 -1.09698880e+00
-2.89595217e-01 -3.84072125e-01 2.89396103e-02 -1.00209224e+00
2.95236081e-01 -9.80143964e-01 -5.96857250e-01 -1.57747436e+00
2.81640112e-01 -4.46067482e-01 -2.86179930e-01 7.68881321e-01
-1.93904340e-01 -1.06314614e-01 5.35185099e-01 5.78142047e-01
-6.91767633e-01 3.62813205e-01 1.06586826e+00 -7.09242895e-02
-2.60315180e-01 -2.70843178e-01 -1.33858216e+00 4.97277677e-01
7.82702327e-01 -4.54542756e-01 -4.97301310e-01 -7.90630996e-01
1.52354658e-01 9.33682546e-02 5.32348789e-02 -7.78372049e-01
3.87643337e-01 -1.04189962e-01 2.22200274e-01 -3.73983651e-01
3.02117139e-01 -6.57140434e-01 -3.58507819e-02 3.06215435e-01
-5.87876558e-01 2.53559142e-01 3.23640376e-01 2.50741810e-01
-4.95630473e-01 -2.84751952e-01 4.02544647e-01 -3.36712569e-01
-7.39985526e-01 2.44046152e-01 -6.95088565e-01 2.03800201e-01
6.96772814e-01 -1.83812138e-02 -5.17407358e-01 -4.00023401e-01
-3.61772090e-01 7.24549890e-01 5.05531013e-01 8.16150248e-01
5.11479020e-01 -1.06623983e+00 -5.78426182e-01 4.97484356e-01
2.92017013e-01 -2.32037425e-01 4.55008745e-01 7.97768056e-01
-2.25388646e-01 8.55597615e-01 2.12082230e-02 7.91215748e-02
-1.13589156e+00 5.36758423e-01 4.81855497e-02 -5.82724512e-01
-4.27535564e-01 5.38493454e-01 -1.76528711e-02 -6.54802561e-01
2.91146070e-01 -3.80760819e-01 -2.50845551e-01 1.24883018e-01
7.84036160e-01 2.15165466e-01 1.14243940e-01 -4.98569697e-01
-1.35404676e-01 2.97098696e-01 -6.22246325e-01 -4.24832590e-02
1.33820772e+00 -1.14421830e-01 -1.51578575e-01 9.91625264e-02
1.03952777e+00 2.01638386e-01 -8.87917757e-01 -2.58442402e-01
2.86049694e-01 -6.12467825e-01 1.51930619e-02 -9.52732563e-01
-8.86445761e-01 7.64740169e-01 3.74443293e-01 -6.71518445e-02
8.66065621e-01 -9.72239450e-02 9.51402724e-01 8.23678672e-01
3.55437100e-01 -1.32883370e+00 -1.89474478e-01 8.13756704e-01
8.33196878e-01 -1.33426034e+00 -1.75905749e-01 5.62444776e-02
-9.92131054e-01 7.00813413e-01 7.19323397e-01 2.59613276e-01
6.35486066e-01 2.58415580e-01 5.95873237e-01 -2.15454951e-01
-9.64749396e-01 -1.73452143e-02 -2.89990101e-02 4.05520469e-01
4.98027653e-01 -3.42755944e-01 -4.87536281e-01 9.20238912e-01
-7.24143833e-02 3.65596950e-01 5.35160124e-01 8.99285674e-01
-2.49740213e-01 -1.40373361e+00 -3.82778287e-01 6.92451596e-01
-5.38792908e-01 -6.97671533e-01 -2.72953808e-01 7.02195346e-01
-2.44463280e-01 1.11623025e+00 -2.43851736e-01 -3.55102837e-01
5.24564564e-01 3.26001048e-01 -2.05868587e-01 -7.72275984e-01
-8.20045233e-01 1.75386578e-01 1.72544733e-01 -7.05625176e-01
1.65367514e-01 -4.68780875e-01 -1.11720693e+00 -3.92668903e-01
-1.93367273e-01 3.12633425e-01 4.86133337e-01 8.19470108e-01
7.73431957e-01 1.18439449e-02 1.68497518e-01 -3.26534331e-01
-1.00315213e+00 -1.22402811e+00 -6.23569548e-01 5.30363917e-01
-1.59169808e-01 -3.59487444e-01 -1.92281455e-01 -2.45832294e-01] | [10.878803253173828, 8.501314163208008] |
739e7218-83bc-4099-b9c2-25141ad7c6aa | identifiability-of-label-noise-transition | 2202.02016 | null | https://arxiv.org/abs/2202.02016v3 | https://arxiv.org/pdf/2202.02016v3.pdf | Identifiability of Label Noise Transition Matrix | The noise transition matrix plays a central role in the problem of learning with noisy labels. Among many other reasons, a large number of existing solutions rely on access to it. Identifying and estimating the transition matrix without ground truth labels is a critical and challenging task. When label noise transition depends on each instance, the problem of identifying the instance-dependent noise transition matrix becomes substantially more challenging. Despite recent works proposing solutions for learning from instance-dependent noisy labels, the field lacks a unified understanding of when such a problem remains identifiable. The goal of this paper is to characterize the identifiability of the label noise transition matrix. Building on Kruskal's identifiability results, we are able to show the necessity of multiple noisy labels in identifying the noise transition matrix for the generic case at the instance level. We further instantiate the results to explain the successes of the state-of-the-art solutions and how additional assumptions alleviated the requirement of multiple noisy labels. Our result also reveals that disentangled features are helpful in the above identification task and we provide empirical evidence. | ['Kun Zhang', 'Hao Cheng', 'Yang Liu'] | 2022-02-04 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 5.40325642e-01 3.19215357e-02 -1.42738521e-01 -3.70515198e-01
-1.19226611e+00 -8.09056044e-01 4.99538660e-01 9.74376053e-02
-1.93054840e-01 5.76012075e-01 1.41242027e-01 -1.83584362e-01
-8.30930889e-01 -2.31181875e-01 -5.16322374e-01 -9.87101555e-01
-9.73746777e-02 6.37448251e-01 -4.24273729e-01 8.23068842e-02
1.59158319e-01 9.90727916e-02 -1.57770050e+00 3.46069075e-02
3.00721675e-01 9.67476368e-01 -1.71686932e-01 5.72643042e-01
8.39839503e-02 6.81622267e-01 -2.61531711e-01 -3.61437500e-01
5.41463256e-01 -4.67583150e-01 -1.16615903e+00 3.94535273e-01
6.25978529e-01 2.82061938e-03 -3.39927286e-01 1.18911707e+00
3.66909921e-01 1.01891041e-01 6.85577512e-01 -1.69981921e+00
-4.57279086e-01 1.07320964e+00 -3.95506889e-01 2.56040722e-01
2.74606675e-01 -1.56380042e-01 1.66360545e+00 -6.43748999e-01
6.47777021e-01 1.07317841e+00 9.13181484e-01 5.88434577e-01
-1.67002380e+00 -7.17466354e-01 1.80336058e-01 2.90517986e-01
-1.37821472e+00 -5.39367259e-01 8.51127326e-01 -6.96167827e-01
3.57144594e-01 4.12506968e-01 1.70679703e-01 1.46588409e+00
-4.30224866e-01 5.58460653e-01 1.48079133e+00 -5.97261727e-01
1.04339585e-01 1.03751689e-01 7.30514169e-01 7.54964590e-01
6.66534305e-01 1.53638422e-01 -8.24293494e-01 -4.03001666e-01
5.90149343e-01 -1.91684067e-01 -2.64226347e-01 -6.25851810e-01
-1.16767073e+00 8.66136193e-01 -3.91304344e-02 2.96359211e-01
9.18383300e-02 3.64663601e-01 2.53465563e-01 6.57176077e-01
5.88986933e-01 7.39710212e-01 -5.79837382e-01 -2.24523604e-01
-8.26979518e-01 -9.89798233e-02 1.04954374e+00 8.50647390e-01
1.08137918e+00 -5.19407868e-01 -1.37671027e-02 5.69635749e-01
1.68105528e-01 3.18981618e-01 3.81419547e-02 -1.23001909e+00
5.18663943e-01 4.45006251e-01 1.66767523e-01 -9.12573099e-01
-6.95947886e-01 -7.81829834e-01 -8.83305192e-01 -2.10923240e-01
9.85962629e-01 -1.51303425e-01 -6.11288548e-01 2.18174505e+00
2.18214557e-01 4.31676924e-01 -1.76578641e-01 6.56530619e-01
3.37453127e-01 3.99117954e-02 -1.43930286e-01 -2.43766248e-01
1.28077424e+00 -5.17323196e-01 -8.78488004e-01 -2.97996551e-01
8.74429703e-01 -6.17067695e-01 8.69244456e-01 2.95119584e-01
-5.59120655e-01 -2.54566312e-01 -7.47854948e-01 -1.50452435e-01
-8.15135017e-02 2.80904293e-01 9.80694711e-01 9.29063261e-01
-8.12498271e-01 6.55030906e-01 -8.23434532e-01 -5.71947932e-01
2.37048551e-01 5.01295924e-01 -6.74527764e-01 -3.84906054e-01
-1.15527499e+00 9.02659059e-01 -9.31184813e-02 3.76972228e-01
-6.82724357e-01 -5.21606326e-01 -7.65144527e-01 3.22154649e-02
7.75035560e-01 -7.02953279e-01 1.14823520e+00 -7.15003312e-01
-1.02085757e+00 8.74431849e-01 -1.75736994e-01 -6.90128356e-02
6.95976734e-01 7.79705048e-02 -6.08754456e-02 -2.41922904e-02
4.01828974e-01 3.22997896e-03 8.12248886e-01 -1.52328098e+00
-5.21293998e-01 -5.31508207e-01 6.04554377e-02 -2.56334012e-03
-1.76757038e-01 -3.39126378e-01 -8.59621987e-02 -2.32208610e-01
7.51253903e-01 -1.25005901e+00 -1.83192760e-01 -3.91213000e-01
-8.34554970e-01 -3.68121684e-01 5.37528157e-01 -2.66521782e-01
1.20614719e+00 -2.30432367e+00 6.15464002e-02 5.23499668e-01
6.38775766e-01 -4.16732788e-01 -2.51749575e-01 6.56661153e-01
-4.46970314e-01 4.48387444e-01 -3.07644874e-01 -7.88287759e-01
4.83887851e-01 2.75981605e-01 -2.02580824e-01 9.36029315e-01
5.78051172e-02 6.26007378e-01 -1.03213716e+00 -1.67002767e-01
-1.06943615e-01 1.34615436e-01 -3.10851902e-01 -6.72023892e-02
2.80029833e-01 7.88931429e-01 -3.04592401e-01 3.52224320e-01
5.68765998e-01 -5.65555394e-01 4.61772472e-01 -2.55465895e-01
1.25338569e-01 2.73289621e-01 -1.93837357e+00 1.48706138e+00
-1.90041274e-01 5.41752160e-01 2.78210223e-01 -1.24001467e+00
4.51125324e-01 3.92541856e-01 7.46384561e-01 -1.45397618e-01
7.86843076e-02 2.16720939e-01 7.16743469e-02 -7.28273332e-01
1.84186399e-01 -3.94396305e-01 -3.53728533e-01 8.56984437e-01
1.54306993e-01 1.41215369e-01 3.03875268e-01 3.60559583e-01
1.42445099e+00 -3.09633255e-01 2.17579097e-01 -2.64743745e-01
2.24211708e-01 -2.79161066e-01 5.91658235e-01 1.30671513e+00
-1.99704975e-01 6.28656745e-01 7.89456129e-01 -1.99283138e-01
-7.37127662e-01 -8.97738695e-01 -4.07877684e-01 1.07234919e+00
2.81874407e-02 -5.60975492e-01 -5.36244452e-01 -6.57652736e-01
4.94251326e-02 3.27450395e-01 -1.03135037e+00 -1.23635553e-01
-1.18026964e-01 -8.43272567e-01 5.94280183e-01 2.47374922e-01
1.95432976e-01 -2.45267794e-01 2.15192754e-02 -6.04075305e-02
-6.87598109e-01 -1.29405499e+00 -3.60154897e-01 6.37094021e-01
-6.00204885e-01 -1.49813366e+00 4.26303372e-02 -5.10615706e-01
8.10993016e-01 2.91986763e-01 1.06432867e+00 1.02300674e-01
-2.16440246e-01 9.06691730e-01 -2.30337307e-01 -3.58349085e-02
-4.38119859e-01 3.31243187e-01 2.55973697e-01 2.68850386e-01
3.66106153e-01 -6.89409435e-01 -2.49038592e-01 3.45884234e-01
-9.28368509e-01 -2.73164541e-01 3.53795558e-01 1.01147056e+00
2.97403783e-01 2.77172983e-01 4.41879839e-01 -1.38200915e+00
6.74885333e-01 -6.08852446e-01 -5.09088933e-01 1.95296809e-01
-6.93580747e-01 5.34652472e-01 3.19373131e-01 -3.82279903e-01
-6.72055900e-01 4.44824100e-01 2.09158525e-01 1.05156342e-03
-1.98184058e-01 4.53980267e-01 -1.63596377e-01 -1.60797983e-01
7.38954067e-01 -1.88999757e-01 -2.63131142e-01 -6.69330418e-01
4.74777192e-01 4.95221972e-01 3.03381205e-01 -8.63290608e-01
8.34769309e-01 5.17625034e-01 5.10585427e-01 -4.75208879e-01
-1.39877248e+00 -8.87508929e-01 -8.67234766e-01 5.74141629e-02
5.69621861e-01 -9.15612876e-01 -1.09004653e+00 1.76267743e-01
-8.59008193e-01 -9.49199870e-02 -5.77405095e-01 4.21097726e-01
-6.43506765e-01 5.78775287e-01 -5.61770499e-01 -9.19451892e-01
4.33113277e-01 -1.16753161e+00 1.02009165e+00 -3.64575893e-01
-3.88654709e-01 -1.12327969e+00 1.83843106e-01 6.74633622e-01
1.59885198e-01 1.90679599e-02 9.66428816e-01 -8.18720639e-01
-7.76268661e-01 -3.66206884e-01 -1.35081112e-01 1.41465247e-01
1.82003468e-01 -2.49502584e-01 -1.12709546e+00 -2.10679337e-01
1.95349321e-01 -3.59694928e-01 1.08342302e+00 3.39938343e-01
8.31503212e-01 -6.78014383e-02 -9.15146992e-02 4.83211219e-01
1.34005988e+00 -5.90277374e-01 -1.53155439e-02 1.48419604e-01
8.18562329e-01 7.41022587e-01 1.02313021e-02 5.00324845e-01
5.10278761e-01 4.78991002e-01 3.74720126e-01 2.18525335e-01
-1.41241789e-01 -2.82358259e-01 1.68754328e-02 7.89532065e-01
1.72826052e-01 -1.01131395e-01 -9.35282767e-01 4.12579775e-01
-1.98505628e+00 -9.52482998e-01 -4.66667414e-01 2.15879726e+00
9.18408215e-01 -1.48910314e-01 -1.54057607e-01 5.76387823e-01
7.48913884e-01 -8.20628479e-02 -4.11510587e-01 2.24898368e-01
-1.34593025e-01 1.40730152e-02 5.71974158e-01 6.90224290e-01
-1.25855851e+00 5.44366539e-01 7.13690996e+00 5.35610199e-01
-6.64861441e-01 2.44446650e-01 3.85266036e-01 3.30839194e-02
-2.79614508e-01 2.72186279e-01 -7.99203753e-01 3.03628832e-01
7.01497078e-01 -6.77227462e-03 6.84704185e-01 5.09803891e-01
7.68759102e-02 -3.76838237e-01 -1.76085877e+00 1.26831865e+00
-4.19091322e-02 -8.46146762e-01 -3.67364913e-01 2.05865875e-01
9.77428794e-01 -2.22297564e-01 3.24061722e-01 9.89968926e-02
6.94100440e-01 -1.03737712e+00 4.52439904e-01 4.44329917e-01
7.16818333e-01 -3.33517611e-01 7.10870028e-01 5.49700439e-01
-9.37197387e-01 -3.36904347e-01 -2.09523723e-01 -3.96334052e-01
-1.95393413e-01 1.20468318e+00 -6.00695312e-01 5.68058252e-01
2.45420873e-01 7.10626006e-01 -6.84936464e-01 8.93182516e-01
-4.17883933e-01 9.35470223e-01 -4.82026577e-01 5.42548716e-01
2.21857876e-02 -2.04432860e-01 3.73258501e-01 8.90036583e-01
2.57313520e-01 -7.75402412e-02 3.67241651e-01 8.20037425e-01
-3.09432477e-01 -1.06588423e-01 -6.22917712e-01 -1.79364964e-01
3.50068927e-01 1.33083606e+00 -7.46631622e-01 7.23237693e-02
-3.51967394e-01 7.55739927e-01 4.59457606e-01 4.26314771e-01
-3.31557810e-01 2.50221431e-01 7.19651282e-01 -1.94652572e-01
4.77767773e-02 -3.21148545e-01 -6.02637410e-01 -1.24979353e+00
-2.67690071e-03 -8.59805286e-01 4.09988254e-01 -1.94637358e-01
-1.81492400e+00 -6.40649348e-03 -8.72560740e-02 -1.05190325e+00
-1.73835725e-01 -4.16133970e-01 -3.80263925e-02 7.80020833e-01
-1.16617966e+00 -9.58620429e-01 -2.11702541e-01 5.77537417e-01
3.99050228e-02 3.60435843e-01 1.00108135e+00 3.11174572e-01
-8.38525772e-01 5.51308632e-01 3.56386483e-01 1.54924333e-01
8.50666285e-01 -1.43813729e+00 8.50437582e-02 8.83084297e-01
6.43716216e-01 8.57498407e-01 8.50961924e-01 -4.27246094e-01
-1.44024563e+00 -7.76645243e-01 1.07841182e+00 -1.03130829e+00
9.95363474e-01 -7.19470799e-01 -5.19464970e-01 9.77789998e-01
-4.05194521e-01 3.15123796e-01 9.74375427e-01 6.82884991e-01
-6.66422367e-01 -7.95262828e-02 -9.83973920e-01 3.18304598e-01
1.37125921e+00 -9.37987030e-01 -2.17103332e-01 5.29666364e-01
2.01266497e-01 -2.64596611e-01 -7.23128855e-01 2.56577544e-02
3.09586644e-01 -9.63007927e-01 6.52055740e-01 -7.40321159e-01
1.37895420e-01 -2.18234256e-01 -3.21691692e-01 -1.32873178e+00
-5.91536462e-01 -7.13247478e-01 3.24853361e-02 1.27821696e+00
4.45492923e-01 -6.37513757e-01 8.33017111e-01 9.79243338e-01
2.49701276e-01 -3.21034819e-01 -1.21354103e+00 -7.53044069e-01
-2.54486799e-01 -7.00380147e-01 1.88076749e-01 1.43472505e+00
3.27966549e-02 5.97821593e-01 -5.68101168e-01 4.45717156e-01
8.85539412e-01 -8.11254457e-02 5.01676559e-01 -1.71770167e+00
-3.11390758e-01 -9.54539105e-02 -4.78598922e-01 -7.91485727e-01
4.66612101e-01 -1.16144705e+00 1.29291564e-01 -1.40640795e+00
5.11323392e-01 -7.05149651e-01 -3.68234843e-01 6.87914073e-01
-2.43055776e-01 3.22181255e-01 2.83742517e-01 2.36002117e-01
-7.47411132e-01 6.91021085e-02 1.06642210e+00 -1.98582843e-01
3.40169400e-01 1.88767195e-01 -1.10882699e+00 5.98145962e-01
6.34772301e-01 -9.72321451e-01 -5.49058139e-01 -2.68213779e-01
7.71387160e-01 -2.23988593e-01 3.52006912e-01 -7.86755085e-01
4.89373416e-01 7.22482949e-02 -9.49526802e-02 7.08962616e-05
3.48047763e-01 -1.02047956e+00 2.67316699e-01 6.44370839e-02
-7.58364201e-01 -1.06275603e-01 -3.13120395e-01 9.63595212e-01
3.73099465e-03 -5.33790708e-01 4.80833232e-01 -1.43935561e-01
-4.79162574e-01 2.70281971e-01 -3.44088614e-01 5.05611837e-01
6.52272403e-01 5.69197014e-02 -2.33014151e-01 -5.46813488e-01
-1.13561511e+00 2.23864019e-02 1.93813249e-01 3.96112278e-02
7.77912959e-02 -1.19456983e+00 -5.07227004e-01 3.25625032e-01
1.26588926e-01 -2.63847739e-01 2.91913599e-01 1.15624607e+00
2.71249563e-01 1.60415545e-01 2.95556426e-01 -5.98955393e-01
-1.24452424e+00 3.38343561e-01 2.75001556e-01 -4.39591050e-01
-2.71584064e-01 1.08372235e+00 -1.10441139e-02 -5.00017166e-01
3.45367730e-01 -3.15293282e-01 2.55156875e-01 3.34904581e-01
1.48892939e-01 6.05735183e-01 2.38053516e-01 -5.52413225e-01
-3.16487581e-01 6.81199670e-01 1.52531981e-01 -1.29274011e-01
1.23562324e+00 -5.29713988e-01 -2.83587337e-01 8.78201723e-01
1.29870427e+00 1.15857363e-01 -1.14845335e+00 -5.47976077e-01
3.15753967e-01 -2.87696600e-01 -6.31055981e-02 -7.35324681e-01
-9.77445602e-01 8.04786801e-01 4.91895586e-01 6.64692223e-01
8.09822261e-01 1.62120983e-01 2.49859169e-01 6.47056699e-01
3.73800009e-01 -9.92372751e-01 5.90326712e-02 7.33470559e-01
3.38523984e-01 -1.53095889e+00 6.78041158e-03 -8.43842685e-01
-3.75560105e-01 9.75759387e-01 5.15271612e-02 8.67226273e-02
7.27613568e-01 1.96718827e-01 1.94685325e-01 -3.26735467e-01
-6.05158806e-01 -5.08392036e-01 2.31571600e-01 6.53674483e-01
2.63169140e-01 6.03152774e-02 -1.55753091e-01 7.17386246e-01
-3.23108196e-01 -4.17481542e-01 7.52770007e-01 6.47488594e-01
-1.13064289e-01 -1.42187381e+00 -4.47155625e-01 5.89464128e-01
-4.94090617e-01 -1.56328544e-01 -4.79198307e-01 6.48013532e-01
3.46216708e-01 1.42800748e+00 -3.71552050e-01 -4.20209795e-01
1.70954019e-01 5.09223104e-01 5.98335862e-01 -8.52529168e-01
-4.49387312e-01 -2.79560626e-01 1.12372853e-01 -4.41343307e-01
-5.95588028e-01 -7.69203305e-01 -8.14940393e-01 -3.27004522e-01
-5.28153539e-01 3.12209159e-01 6.29696131e-01 1.19809091e+00
1.93437308e-01 2.21743852e-01 6.50640607e-01 -5.51131964e-01
-1.10067177e+00 -8.27149689e-01 -1.10703123e+00 9.01479483e-01
6.59727991e-01 -8.41804981e-01 -9.67467368e-01 8.68031234e-02] | [9.403112411499023, 4.142213821411133] |
2c9bc79c-5c1e-48eb-8520-852506ce5f74 | vcnet-a-self-explaining-model-for-realistic | 2212.10847 | null | https://arxiv.org/abs/2212.10847v1 | https://arxiv.org/pdf/2212.10847v1.pdf | VCNet: A self-explaining model for realistic counterfactual generation | Counterfactual explanation is a common class of methods to make local explanations of machine learning decisions. For a given instance, these methods aim to find the smallest modification of feature values that changes the predicted decision made by a machine learning model. One of the challenges of counterfactual explanation is the efficient generation of realistic counterfactuals. To address this challenge, we propose VCNet-Variational Counter Net-a model architecture that combines a predictor and a counterfactual generator that are jointly trained, for regression or classification tasks. VCNet is able to both generate predictions, and to generate counterfactual explanations without having to solve another minimisation problem. Our contribution is the generation of counterfactuals that are close to the distribution of the predicted class. This is done by learning a variational autoencoder conditionally to the output of the predictor in a join-training fashion. We present an empirical evaluation on tabular datasets and across several interpretability metrics. The results are competitive with the state-of-the-art method. | ['Alexandre Termier', 'Tassadit Bouadi', 'Thomas Guyet', 'Françoise Fessant', 'Victor Guyomard'] | 2022-12-21 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 4.03303146e-01 1.02077603e+00 -2.54806578e-01 -4.02662784e-01
-6.20894432e-01 -2.22154289e-01 1.09122562e+00 -4.20741588e-02
-4.15475629e-02 1.34464192e+00 5.31815827e-01 -6.44980311e-01
4.99805957e-02 -9.09053922e-01 -1.07888460e+00 -5.60629010e-01
2.31750891e-01 8.08974862e-01 -3.48516047e-01 -7.71028623e-02
5.00999808e-01 9.21627358e-02 -1.83657467e+00 7.89215326e-01
1.01648545e+00 8.54618549e-01 -5.99288307e-02 5.93025923e-01
-1.16959892e-01 9.24678206e-01 -6.04026020e-01 -7.82225847e-01
1.24673478e-01 -7.45759308e-01 -7.65490234e-01 -7.96613172e-02
2.91621923e-01 -2.77715057e-01 5.88541403e-02 8.93489838e-01
2.24791601e-01 1.77936718e-01 1.23964071e+00 -1.49095130e+00
-9.39338744e-01 1.10471094e+00 -9.65264365e-02 -8.52914006e-02
8.89088139e-02 2.64490813e-01 1.25508153e+00 -7.17348278e-01
7.45145321e-01 1.44538462e+00 4.50350195e-01 1.06082237e+00
-1.59078002e+00 -4.69006956e-01 1.45002261e-01 2.36371413e-01
-5.65005243e-01 -2.34568313e-01 9.11446393e-01 -3.84237200e-01
1.04827523e+00 3.43793869e-01 5.48912942e-01 1.43884945e+00
6.12178504e-01 7.12640285e-01 1.00433016e+00 -3.98193032e-01
7.03769684e-01 3.29439849e-01 -4.30874437e-01 3.25118750e-01
3.92031252e-01 6.70879424e-01 -3.43655646e-01 -2.42650583e-01
3.89283776e-01 -9.60964710e-03 -3.65184397e-01 -6.16279185e-01
-1.09844792e+00 1.43575752e+00 6.59948349e-01 1.47863984e-01
-7.24521041e-01 5.85212588e-01 2.57675380e-01 2.00804994e-01
7.34463155e-01 1.09570670e+00 -8.69654536e-01 3.75970811e-01
-9.04740214e-01 8.38105798e-01 8.25823784e-01 2.30573833e-01
5.12245417e-01 3.32383931e-01 -4.70044225e-01 3.15918595e-01
4.34255660e-01 3.43781292e-01 9.57113087e-01 -1.02497852e+00
4.27519202e-01 6.16887867e-01 2.31232688e-01 -5.50456464e-01
-1.66022480e-01 -4.77155238e-01 -6.90381348e-01 6.52139485e-01
3.72867763e-01 -5.57572961e-01 -8.66071522e-01 1.71806669e+00
4.78830814e-01 2.52599627e-01 4.95126367e-01 7.69008696e-01
4.48877931e-01 6.92380309e-01 4.10250314e-02 -4.83414948e-01
8.67481589e-01 -7.84006476e-01 -5.50365567e-01 -3.19837928e-01
6.30515158e-01 -4.24784869e-01 7.68024087e-01 1.75998077e-01
-8.60175252e-01 -4.50306803e-01 -1.07268834e+00 4.35937345e-02
-4.28063035e-01 -1.37655959e-01 7.94098914e-01 5.56535780e-01
-7.51188517e-01 1.16995668e+00 -4.59291607e-01 4.21556383e-01
5.14165819e-01 3.97872567e-01 -1.09592162e-01 4.46384639e-01
-1.20041072e+00 1.17216146e+00 6.97247803e-01 -2.21166417e-01
-9.95513976e-01 -1.08098853e+00 -1.03373170e+00 4.11741704e-01
3.10320765e-01 -1.14238095e+00 1.52663386e+00 -1.54048026e+00
-1.53456140e+00 4.33598399e-01 2.83077043e-02 -1.29081547e+00
9.68737602e-01 3.96152176e-02 -2.53766924e-01 -5.24943113e-01
3.66148621e-01 8.33835185e-01 1.04397750e+00 -1.27845848e+00
-5.55390775e-01 -3.06649536e-01 -1.16949663e-01 1.43547535e-01
4.53080565e-01 -7.33202875e-01 5.45283973e-01 -5.22020042e-01
-3.08908701e-01 -7.24942923e-01 -5.98111570e-01 -3.21930557e-01
-7.72943616e-01 -3.63768160e-01 6.78696811e-01 -3.60710174e-01
6.89761043e-01 -1.67917299e+00 3.57197784e-02 2.40379255e-02
1.42525122e-01 6.45551234e-02 8.04595500e-02 2.09677331e-02
-6.37688339e-01 3.16008776e-01 -4.02814686e-01 -3.93512785e-01
3.63930166e-01 4.07134220e-02 -1.02857637e+00 2.46461585e-01
3.16505164e-01 9.96960938e-01 -8.34877670e-01 -1.43333122e-01
3.36697131e-01 2.20996365e-01 -9.39919829e-01 3.10119271e-01
-9.41231012e-01 1.84921116e-01 -1.21665336e-01 -8.14244524e-02
5.92630148e-01 -6.25801906e-02 5.65423667e-01 2.61368215e-01
-1.79889929e-02 6.61100090e-01 -1.11733758e+00 1.09549093e+00
-5.42208552e-01 7.22840548e-01 -8.42304051e-01 -1.16321838e+00
7.76541531e-01 5.59419394e-01 -1.10049218e-01 -4.32320118e-01
2.85505444e-01 5.07294834e-01 5.68546206e-02 -1.39061913e-01
4.45261121e-01 -8.20221543e-01 -1.40092582e-01 6.96202159e-01
1.39784694e-01 -3.51068407e-01 -6.28764778e-02 -1.31284893e-01
4.33869213e-01 1.97098583e-01 9.53774631e-01 -1.34586558e-01
3.37737590e-01 6.84842393e-02 5.69402575e-01 8.89388978e-01
1.72044352e-01 7.47818828e-01 9.96623635e-01 -9.37576175e-01
-1.20900106e+00 -9.75141883e-01 1.52512237e-01 4.74153787e-01
-3.73566628e-01 2.04961583e-01 -8.12100112e-01 -1.26196492e+00
2.58695692e-01 1.77018082e+00 -1.16536105e+00 -3.49534839e-01
-2.12633312e-01 -7.11732864e-01 -1.70869809e-02 4.24991816e-01
3.59854668e-01 -1.41971326e+00 -8.54048729e-01 2.00369060e-01
-1.74699590e-01 -3.45509768e-01 -1.47325680e-01 1.20657712e-01
-1.02959621e+00 -1.22977698e+00 -5.20765245e-01 -1.64597020e-01
3.67982417e-01 -3.55106682e-01 1.34435654e+00 4.42379229e-02
1.87476814e-01 -3.26274633e-01 1.91927820e-01 -8.93929422e-01
-8.22997987e-01 -1.81413755e-01 1.14018172e-01 1.89353302e-02
2.07601726e-01 -5.15460789e-01 -5.50450623e-01 -2.70152032e-01
-6.86429143e-01 3.16277683e-01 4.87627566e-01 1.34093797e+00
6.34963989e-01 -1.31013289e-01 6.16848350e-01 -1.27997565e+00
7.91345417e-01 -6.20556235e-01 -7.57663071e-01 2.36904606e-01
-9.91379440e-01 7.53479421e-01 9.88933206e-01 -2.82296002e-01
-1.48337710e+00 -4.81907949e-02 -6.03011176e-02 -2.52143741e-01
-2.97056884e-01 3.03260118e-01 -7.13201538e-02 7.35553980e-01
8.70194733e-01 1.21581644e-01 -1.55229971e-01 -2.87786394e-01
7.71290481e-01 3.82581115e-01 5.05714834e-01 -8.04255754e-02
5.18453479e-01 4.68890727e-01 1.57170683e-01 -2.51196235e-01
-1.03941894e+00 3.18544775e-01 -3.75964314e-01 -3.29460278e-02
8.33813369e-01 -4.79978114e-01 -6.58731341e-01 -2.05669120e-01
-1.46617270e+00 -4.72238541e-01 -8.61883640e-01 4.22010630e-01
-9.57445085e-01 -3.03869784e-01 2.50805974e-01 -8.80354702e-01
-4.03506637e-01 -8.51513922e-01 7.55615652e-01 1.29027516e-01
-5.60372770e-01 -1.29194796e+00 3.80450130e-01 1.97227940e-01
1.29550084e-01 6.60094917e-01 9.39150870e-01 -9.35602844e-01
-4.65348482e-01 4.47058454e-02 3.04452032e-02 2.09351584e-01
-2.46817365e-01 -5.20722121e-02 -1.03367543e+00 1.21560335e-01
6.30782619e-02 -1.51189476e-01 1.28007233e+00 8.49418283e-01
1.08550537e+00 -9.20064270e-01 -4.35779899e-01 3.47780079e-01
1.34642053e+00 1.46154791e-01 6.39628589e-01 2.20629886e-01
2.10765362e-01 6.53498471e-01 5.06143332e-01 4.23558235e-01
2.43858978e-01 4.78103131e-01 7.99442470e-01 2.19316985e-02
6.12726398e-02 -6.46570981e-01 2.64612526e-01 -4.74435270e-01
-8.84480327e-02 -2.77533114e-01 -6.68511808e-01 7.63740718e-01
-2.01157832e+00 -1.46572793e+00 -2.11277694e-01 1.99769139e+00
5.27380109e-01 1.84996635e-01 -2.28955641e-01 1.38620794e-01
4.68416750e-01 2.12931171e-01 -6.79901123e-01 -1.02089870e+00
1.97890058e-01 1.81176737e-01 2.35635728e-01 7.17703760e-01
-9.86709535e-01 7.17218101e-01 6.45675278e+00 4.62308139e-01
-1.10319471e+00 9.84260291e-02 9.00198877e-01 -2.27778703e-01
-8.59128058e-01 8.23911652e-02 -2.80428082e-01 5.11209965e-01
1.11178803e+00 -4.16009277e-01 2.63651401e-01 1.16686738e+00
2.92787075e-01 9.56546739e-02 -1.46480191e+00 4.93328452e-01
-9.37044248e-02 -2.11581469e+00 4.13262397e-01 6.23609126e-02
1.19674718e+00 -1.21911496e-01 4.07571614e-01 4.70888674e-01
6.64779603e-01 -1.22852361e+00 1.04941070e+00 4.94262636e-01
5.35555422e-01 -9.12486792e-01 1.07671082e+00 5.15114248e-01
-4.65379089e-01 -3.50597411e-01 -4.51554954e-01 -4.32133555e-01
8.17824714e-03 5.79379737e-01 -1.30036759e+00 3.19257408e-01
1.48767203e-01 3.90740871e-01 -1.18842892e-01 6.05372131e-01
-8.14960003e-01 4.14017558e-01 1.73969015e-01 -3.58036935e-01
1.39294192e-01 2.48781636e-01 5.28620839e-01 8.32359433e-01
3.97443801e-01 -5.60801141e-02 -3.67656946e-01 1.63898802e+00
-2.59226292e-01 -1.44108444e-01 -9.23994303e-01 2.86457092e-01
9.85288769e-02 6.33383811e-01 -3.15421939e-01 -6.60191119e-01
1.24870338e-01 9.61802781e-01 4.55695331e-01 1.98073581e-01
-8.86004627e-01 6.20891154e-03 7.75358438e-01 -1.90787166e-01
4.21629965e-01 7.55287230e-01 -7.81080306e-01 -1.34951556e+00
-1.91041738e-01 -8.24760854e-01 5.53279459e-01 -8.06745589e-01
-1.09409404e+00 4.57005978e-01 -7.74492025e-02 -1.03684390e+00
-1.13208675e+00 -6.13782823e-01 -1.16018379e+00 8.30236733e-01
-1.35195720e+00 -8.68015289e-01 3.21137518e-01 1.01826675e-01
6.69305921e-01 -2.39516363e-01 9.05613244e-01 -6.14118755e-01
-2.73989350e-01 2.11373746e-01 1.66801631e-01 -3.14111829e-01
1.90515652e-01 -1.88011992e+00 6.20231450e-01 8.68425429e-01
3.83292168e-01 4.28730369e-01 1.35177863e+00 -6.07633829e-01
-5.19565821e-01 -1.19474256e+00 1.52078855e+00 -5.52053213e-01
1.55719280e-01 -7.98399150e-02 -6.02624357e-01 8.68031442e-01
2.75487065e-01 -2.32678860e-01 5.13135850e-01 2.20810667e-01
-2.17872798e-01 1.70788929e-01 -1.32620609e+00 8.16528916e-01
5.30007601e-01 -2.26327389e-01 -1.17961061e+00 2.12662190e-01
6.90533280e-01 -4.00726736e-01 -1.90841854e-01 -6.86775446e-02
5.15777946e-01 -1.39229298e+00 7.91578829e-01 -1.05457997e+00
1.31391776e+00 -8.71672481e-02 9.88979414e-02 -2.04421711e+00
-7.44488835e-03 -5.68837047e-01 -3.26065719e-01 8.15058529e-01
9.95780885e-01 -8.54876399e-01 8.43423128e-01 8.27890158e-01
1.49467647e-01 -7.31259406e-01 -1.13833368e+00 -4.72881824e-01
2.96724647e-01 -5.60817242e-01 1.24833572e+00 7.57406056e-01
1.62216216e-01 4.70713735e-01 -4.91066664e-01 -1.99919809e-02
7.11170137e-01 6.15803421e-01 8.91118467e-01 -1.08848107e+00
-4.70793992e-01 -4.94930327e-01 -1.41861528e-01 -5.63221991e-01
4.93642479e-01 -9.30172265e-01 1.00188456e-01 -1.56737113e+00
2.40397766e-01 2.39854842e-01 -5.69714643e-02 3.22188139e-01
-3.30914825e-01 -1.75594278e-02 2.59310037e-01 -2.10620165e-01
-9.40848440e-02 9.67018604e-01 1.15762913e+00 -1.24287896e-01
-2.59248644e-01 1.68432146e-01 -9.75929379e-01 9.02471423e-01
8.45712841e-01 -7.99820483e-01 -3.93476248e-01 -9.80317891e-02
8.36989358e-02 3.29759896e-01 6.56900406e-01 -6.69255495e-01
-2.63539314e-01 -3.79897654e-01 7.32893288e-01 -4.65966046e-01
1.50506541e-01 -5.04461288e-01 2.03371882e-01 9.25250232e-01
-7.32003748e-01 -1.15407445e-01 -9.07271542e-03 7.11345136e-01
-1.23995036e-01 -2.46341094e-01 5.86618543e-01 -2.61341840e-01
-4.96154606e-01 -7.97664076e-02 -4.34099346e-01 2.80148145e-02
9.91012454e-01 1.21365920e-01 -2.40919396e-01 -5.28031766e-01
-5.84072173e-01 1.55112192e-01 1.52984217e-01 4.33318883e-01
6.77810311e-01 -1.26144540e+00 -9.42613006e-01 1.64858460e-01
-1.05435088e-01 -2.76882797e-01 1.16816089e-01 3.21353555e-01
-2.29254529e-01 7.80135274e-01 -1.98341548e-01 4.10321588e-03
-8.59338999e-01 6.89197361e-01 8.02186847e-01 -7.79974818e-01
-2.84232259e-01 6.65777087e-01 3.36643457e-01 -7.62352824e-01
-4.53359216e-01 -4.25306469e-01 -3.65713120e-01 -5.64677343e-02
5.35732150e-01 2.95084476e-01 -2.14566588e-01 -1.89888284e-01
-1.67328548e-02 -3.51708889e-01 8.45501870e-02 -4.56506819e-01
1.55489945e+00 3.08774829e-01 1.31238610e-01 3.42510670e-01
9.32350278e-01 -4.04117972e-01 -1.42599225e+00 2.82194644e-01
2.07607206e-02 -4.63464350e-01 9.81607102e-03 -9.92789686e-01
-7.80731559e-01 9.06544268e-01 3.64094615e-01 3.79760534e-01
6.07714236e-01 -7.49855563e-02 3.16129297e-01 2.21735924e-01
-4.03288752e-02 -1.09471226e+00 -1.23824574e-01 2.23200515e-01
1.40181768e+00 -1.42855287e+00 4.51918598e-03 5.16893435e-03
-8.69382203e-01 1.07606483e+00 4.48134124e-01 -2.92937040e-01
3.99568111e-01 -3.01525950e-01 1.89220496e-02 -2.73775578e-01
-1.29220915e+00 8.10853168e-02 5.60979009e-01 6.76277280e-01
5.72311819e-01 5.76425970e-01 -3.17767352e-01 7.25301564e-01
-7.37479389e-01 -2.26137135e-02 6.21735036e-01 7.59190843e-02
-8.74012262e-02 -7.84519553e-01 -3.20075035e-01 6.74104273e-01
-5.25708199e-01 -1.35479346e-01 -5.68800867e-01 9.22959268e-01
2.46779144e-01 9.36246574e-01 3.01770151e-01 -4.39414009e-02
1.40086353e-01 2.39822015e-01 1.07236780e-01 -6.81264639e-01
-5.06971776e-01 -4.43031460e-01 1.71407297e-01 -6.67557895e-01
-5.52913137e-02 -9.30918813e-01 -1.13894069e+00 -3.29344273e-01
-3.56855690e-01 3.54403377e-01 5.63699186e-01 1.27540004e+00
1.83812976e-01 7.48391032e-01 5.16763926e-01 -8.97001386e-01
-1.13639915e+00 -9.19143438e-01 -3.79542649e-01 5.11512518e-01
6.59561396e-01 -5.62929928e-01 -5.47043264e-01 -5.34503385e-02] | [8.628473281860352, 5.650996685028076] |
73f0cc9d-8953-426a-a773-251dcc385152 | multiverse-causal-reasoning-using-importance | 1910.08091 | null | https://arxiv.org/abs/1910.08091v2 | https://arxiv.org/pdf/1910.08091v2.pdf | MultiVerse: Causal Reasoning using Importance Sampling in Probabilistic Programming | We elaborate on using importance sampling for causal reasoning, in particular for counterfactual inference. We show how this can be implemented natively in probabilistic programming. By considering the structure of the counterfactual query, one can significantly optimise the inference process. We also consider design choices to enable further optimisations. We introduce MultiVerse, a probabilistic programming prototype engine for approximate causal reasoning. We provide experimental results and compare with Pyro, an existing probabilistic programming framework with some of causal reasoning tools. | ['Ciarán M. Lee', 'Logan Graham', 'Kostis Gourgoulias', 'Adam Baker', 'Yura Perov', 'Saurabh Johri', 'Jonathan G. Richens'] | 2019-10-17 | null | https://openreview.net/forum?id=S1xFcknVFS | https://openreview.net/pdf?id=S1xFcknVFS | pproximateinference-aabi-symposium-2019-12 | ['counterfactual-inference'] | ['miscellaneous'] | [-1.19926453e-01 8.02415133e-01 -4.45679247e-01 -6.83050871e-01
-7.66134560e-01 -5.20189881e-01 1.13500082e+00 3.72070551e-01
-1.96033031e-01 9.56052661e-01 8.71262670e-01 -8.43862712e-01
-5.48213124e-01 -1.20212996e+00 -7.75252104e-01 -2.59513170e-01
-5.13606012e-01 9.78457153e-01 3.40911150e-01 1.21628359e-01
5.09267092e-01 2.34428078e-01 -1.73028851e+00 5.66697061e-01
6.15903020e-01 4.80945200e-01 -3.86126727e-01 7.92666912e-01
-8.07929114e-02 1.13301587e+00 -5.39730966e-01 -6.50406241e-01
-1.52316719e-01 8.10328275e-02 -1.06550360e+00 -1.22510087e+00
2.64993701e-02 -4.07741874e-01 -1.02374367e-01 8.13543916e-01
3.54738593e-01 2.11871460e-01 7.09019899e-01 -1.72215223e+00
-1.81495458e-01 1.67843056e+00 -1.98171601e-01 1.46229550e-01
9.81306374e-01 1.18359968e-01 1.10600352e+00 -4.44730729e-01
5.83435357e-01 2.03336787e+00 6.11002862e-01 1.53708503e-01
-1.57719243e+00 -5.31097531e-01 7.42299706e-02 4.42491651e-01
-1.22193635e+00 -4.84321445e-01 5.68800986e-01 -3.18127006e-01
1.18326414e+00 6.67453170e-01 5.42049766e-01 9.85873699e-01
4.94294465e-01 8.01229358e-01 1.38076329e+00 -5.49259543e-01
7.65049100e-01 -1.16028905e-01 2.49064282e-01 3.77626449e-01
1.34518787e-01 1.02355683e+00 -6.94297016e-01 -9.71780598e-01
2.38288790e-01 -4.88007993e-01 -4.08765897e-02 -3.41192752e-01
-9.79509413e-01 1.01084006e+00 8.86465758e-02 -2.95773298e-01
-4.93648440e-01 9.39806938e-01 3.94654900e-01 7.38715902e-02
4.70211580e-02 2.23595366e-01 -5.69310367e-01 -3.73073995e-01
-8.22297454e-01 1.16325831e+00 1.38237774e+00 6.58699691e-01
2.98984349e-01 -4.26025152e-01 -6.39036179e-01 3.73621434e-01
9.87316608e-01 3.99139017e-01 -1.50162756e-01 -1.56777537e+00
9.59097818e-02 3.56154554e-02 5.13056636e-01 -5.68002462e-01
-9.60493162e-02 4.63358313e-01 -1.42217316e-02 5.70210814e-01
3.57474744e-01 -1.57268688e-01 -5.21525502e-01 1.53002715e+00
6.32121861e-01 2.68860728e-01 -2.36451961e-02 5.31099379e-01
3.16474438e-01 5.34689248e-01 7.08630383e-01 -3.28044623e-01
1.46997404e+00 -3.37926224e-02 -7.61807859e-01 9.62394103e-02
3.21924418e-01 -6.78659737e-01 6.98070407e-01 5.78703403e-01
-9.47890043e-01 2.30960727e-01 -1.10588133e+00 1.31784543e-01
-1.69222996e-01 -9.30330336e-01 1.22351992e+00 1.19578409e+00
-7.91163981e-01 7.78196573e-01 -9.12801325e-01 1.74116075e-01
2.54647136e-01 1.15679145e-01 1.76348880e-01 -1.78670973e-01
-1.57128775e+00 1.08454740e+00 9.88801539e-01 -4.12467837e-01
-9.14069176e-01 -1.05391431e+00 -8.25937569e-01 1.56796962e-01
5.68926275e-01 -1.00002110e+00 1.75404918e+00 -2.01472249e-02
-1.53119862e+00 2.56062150e-01 -1.44136995e-01 -6.88199639e-01
6.53402030e-01 -1.99703589e-01 -4.78374302e-01 -2.60054857e-01
2.13942796e-01 4.20076042e-01 2.03106657e-01 -1.03410470e+00
-6.51938021e-01 -3.99590433e-01 4.48032707e-01 8.80811885e-02
8.03466976e-01 4.91448998e-01 7.00784326e-02 -4.69255954e-01
-4.39976692e-01 -6.92510009e-01 -5.40851831e-01 -1.49497494e-01
-5.91116786e-01 -5.60632527e-01 3.99309367e-01 -1.91996828e-01
1.20551097e+00 -1.40083182e+00 -3.86122376e-01 6.13674223e-01
-4.49999660e-01 -5.43954611e-01 5.06582379e-01 3.87370586e-01
-3.53473693e-01 3.88049215e-01 -4.90130931e-01 2.44080648e-01
5.88849068e-01 4.77024347e-01 -6.84506238e-01 3.27226937e-01
-8.93846080e-02 7.96284616e-01 -1.11659110e+00 -9.56988931e-01
4.98161286e-01 4.37139422e-02 -9.30339694e-01 1.88233912e-01
-8.88586998e-01 -3.69368374e-01 -4.95575488e-01 5.30100226e-01
7.13037789e-01 4.81541395e-01 7.18322754e-01 2.31213957e-01
-2.81549066e-01 9.48892415e-01 -1.70478821e+00 1.55640137e+00
-5.68365335e-01 2.79731192e-02 -2.49957353e-01 -4.17004347e-01
2.14049935e-01 6.23486161e-01 1.57451272e-01 -1.51573345e-01
-1.29693583e-01 -1.29656360e-01 -2.61153013e-01 -2.10425645e-01
4.57747757e-01 -5.55823445e-01 -4.63202566e-01 1.00381565e+00
-2.13069126e-01 -6.33189797e-01 1.29381210e-01 1.57713532e-01
9.87200320e-01 6.39859736e-01 8.75639200e-01 -6.75000548e-01
-6.23709001e-02 4.27255124e-01 4.04356092e-01 1.07779729e+00
-8.58410634e-03 1.52542070e-01 8.66726279e-01 -5.57148814e-01
-6.83493137e-01 -1.58824050e+00 -6.49484992e-01 1.08829713e+00
-2.29719818e-01 -6.57275796e-01 -4.16837513e-01 -8.15850198e-01
2.55356848e-01 1.89664638e+00 -5.85501015e-01 1.11478731e-01
-1.90907106e-01 -1.26609111e+00 6.59833193e-01 5.35267472e-01
-1.54760843e-02 -1.04731309e+00 -8.21657240e-01 3.44278455e-01
-4.38321754e-02 -2.07489729e-01 1.52288958e-01 1.68750003e-01
-7.42801309e-01 -1.23015213e+00 3.57861847e-01 4.61194932e-01
-2.22504422e-01 -4.72196817e-01 1.53213799e+00 -3.27606112e-01
1.97510142e-02 3.31543922e-01 1.96723640e-01 -8.57288182e-01
-7.92978287e-01 -7.78446138e-01 -2.11548835e-01 -1.08520889e+00
5.49219489e-01 -9.54099596e-01 -4.69729573e-01 9.57719609e-02
-7.50541925e-01 -3.67796302e-01 7.17675537e-02 6.24150276e-01
3.10506999e-01 4.83858138e-01 -1.05536720e-02 -1.20609772e+00
1.02781785e+00 -8.26308846e-01 -9.37555432e-01 2.96312124e-01
-7.94864774e-01 8.46117914e-01 -1.57809444e-02 1.09573953e-01
-1.71824408e+00 -1.10314183e-01 -3.06722283e-01 3.03299606e-01
-2.25318640e-01 7.61237860e-01 -3.75802308e-01 6.96966708e-01
9.47132826e-01 -6.06929302e-01 -3.33629370e-01 -4.99269329e-02
1.07747567e+00 4.48245764e-01 6.46554470e-01 -1.47293413e+00
2.56153941e-01 5.41066408e-01 7.80795515e-02 6.00812733e-02
-4.98458892e-01 -2.92084143e-02 -1.30741566e-01 -8.20936933e-02
3.17580909e-01 -6.82629287e-01 -1.46862376e+00 -2.37762928e-01
-1.25341082e+00 -4.45116401e-01 -4.31391269e-01 4.15023357e-01
-8.38582695e-01 4.66331616e-02 -2.29471013e-01 -1.27173650e+00
3.37893814e-02 -1.00078595e+00 8.96467626e-01 -8.82087499e-02
-8.53841066e-01 -1.07072127e+00 7.39461780e-01 -5.05050197e-02
-4.03754115e-02 4.11688924e-01 7.80449271e-01 -5.67442775e-01
-5.09272933e-01 8.04862157e-02 -7.39122033e-02 -5.22546887e-01
-7.27940917e-01 5.77391922e-01 -1.07412481e+00 6.55626357e-01
-2.40965620e-01 -1.15452245e-01 6.32563472e-01 6.71976388e-01
1.10562801e+00 -7.18014121e-01 -7.80872166e-01 -1.59396715e-02
1.39679229e+00 3.27148996e-02 9.33729947e-01 4.29443896e-01
2.04830438e-01 8.02176476e-01 9.17394638e-01 6.82651043e-01
7.38393486e-01 6.56875372e-01 4.21018392e-01 7.61748493e-01
3.29945832e-01 -4.98982161e-01 2.22981334e-01 -3.43804717e-01
-2.07367107e-01 9.81978178e-02 -1.17970717e+00 6.47853613e-01
-2.18512678e+00 -1.60856307e+00 -4.36881334e-01 2.14121079e+00
1.37380886e+00 1.90058157e-01 8.99570957e-02 6.00958057e-03
4.57215339e-01 1.13137819e-01 -9.42704231e-02 -1.07591438e+00
4.23486829e-01 6.33330703e-01 6.64840996e-01 9.40480888e-01
-9.52110052e-01 5.01237035e-01 8.31194973e+00 8.04135144e-01
-1.28232971e-01 2.34992832e-01 4.22155619e-01 -1.96185902e-01
-1.20396924e+00 7.47851968e-01 -3.75393301e-01 4.56817776e-01
1.52134430e+00 -4.72095013e-01 3.92070800e-01 1.09911203e+00
1.76047444e-01 -5.56937635e-01 -1.41570747e+00 2.34900013e-01
-7.39913762e-01 -1.59986556e+00 -9.54269767e-02 -1.88298389e-01
5.64346731e-01 -1.19206078e-01 -5.14768600e-01 3.43310267e-01
1.80576968e+00 -1.20319331e+00 1.02565074e+00 6.97553754e-01
3.81150573e-01 -1.25814688e+00 8.33693266e-01 6.74750805e-02
-7.41333008e-01 -4.08909656e-03 -7.82930776e-02 -2.06592575e-01
5.82300067e-01 1.24781251e+00 -1.01042163e+00 7.75430083e-01
8.84763718e-01 -1.49917910e-02 5.36906272e-02 9.63745594e-01
-1.01670265e+00 6.72668159e-01 -7.44241893e-01 -1.37345627e-01
-3.23356330e-01 2.57895827e-01 6.28201723e-01 1.45892096e+00
1.22815706e-01 2.11854011e-01 -9.52584669e-02 1.36703134e+00
4.19218540e-01 -4.92948830e-01 -5.91211081e-01 4.77547735e-01
9.99837756e-01 5.22877753e-01 -2.08305135e-01 -5.55342138e-01
4.73951697e-02 2.14108914e-01 9.47449952e-02 1.30636737e-01
-9.94116068e-01 -2.56405443e-01 8.20559800e-01 -1.65853515e-01
-1.20115183e-01 2.57399559e-01 -5.86230993e-01 -7.41127014e-01
-2.73153067e-01 -8.38065267e-01 1.02735519e+00 -9.73304391e-01
-1.08977365e+00 -1.53754279e-01 1.04628062e+00 -2.05849349e-01
-8.80504191e-01 -4.39693123e-01 -9.43127871e-01 1.09169078e+00
-9.81903851e-01 -7.28069067e-01 5.29002070e-01 2.18380257e-01
-1.57054424e-01 5.44492006e-01 1.04553187e+00 -3.75368565e-01
-1.60873815e-01 5.63775897e-02 -5.52470982e-01 -6.15159690e-01
5.58112621e-01 -1.92011631e+00 6.41571283e-01 8.66056979e-01
-1.23368964e-01 1.11111450e+00 1.43658638e+00 -8.34429741e-01
-1.14770567e+00 -7.81906903e-01 8.98751915e-01 -6.98430419e-01
9.37259614e-01 -6.09272309e-02 -5.48037350e-01 9.16092813e-01
4.44327086e-01 -4.81811911e-01 6.22799098e-01 9.88256931e-01
-5.93824983e-01 2.10599124e-01 -1.50763571e+00 9.84166384e-01
1.00216043e+00 -4.85537440e-01 -1.41751897e+00 2.25791931e-01
7.98583627e-01 -2.81293392e-01 -1.13440466e+00 3.58439237e-01
9.50701654e-01 -1.15033448e+00 1.35643160e+00 -5.18692672e-01
5.31911731e-01 -6.67189777e-01 -1.96267664e-01 -1.25800192e+00
-1.52284339e-01 -8.45206618e-01 -4.11100566e-01 1.18397188e+00
4.78406399e-01 -8.11661363e-01 5.95642030e-01 1.46079504e+00
2.71546662e-01 -1.31334245e-01 -1.11188233e+00 -2.84547359e-01
8.43592882e-02 -1.37982142e+00 1.22122300e+00 8.23666573e-01
7.11233675e-01 -1.67117920e-02 1.23444311e-01 5.34659505e-01
1.00578868e+00 3.74364257e-01 5.34444392e-01 -1.06723034e+00
-6.90337241e-01 -4.30658847e-01 -6.86313957e-02 -3.17142397e-01
2.25039110e-01 -6.44427896e-01 2.99703628e-01 -1.34988737e+00
2.92166501e-01 -5.93192995e-01 -5.56837907e-03 7.54162490e-01
-5.06105065e-01 -3.30294311e-01 -3.55531454e-01 -2.78816968e-01
-2.21943468e-01 3.90855551e-01 4.64644939e-01 -1.11102656e-01
-2.19889861e-02 5.14663979e-02 -8.57270181e-01 6.50264025e-01
8.74624491e-01 -1.03241360e+00 -4.51690078e-01 1.73872754e-01
6.93425536e-01 4.65119123e-01 6.93660319e-01 -6.71133518e-01
1.99156165e-01 -7.75385797e-01 1.42911106e-01 -7.71400809e-01
-1.14109471e-01 -4.65306044e-01 6.81742132e-01 4.52006280e-01
-5.11836410e-01 -1.26239613e-01 4.52192783e-01 6.85805678e-01
-4.34698500e-02 -5.12031317e-01 2.31251329e-01 -1.03441201e-01
-6.06984496e-01 -2.01811239e-01 -7.03501999e-01 -7.48891011e-02
9.62854207e-01 3.26075763e-01 -5.17164111e-01 -8.98368955e-02
-6.44504368e-01 5.11486769e-01 3.32233578e-01 4.78822887e-02
3.24989408e-01 -1.19737768e+00 -7.42457271e-01 -6.30824566e-01
1.92195028e-01 -2.41680592e-01 9.82001647e-02 3.97532463e-01
-4.94114161e-01 3.67302477e-01 -6.44303337e-02 -2.55807906e-01
-7.73842216e-01 9.33436692e-01 2.05011725e-01 -4.13010627e-01
-1.64368927e-01 5.96447349e-01 -2.61809081e-01 -8.34383190e-01
-7.78358653e-02 -4.43920404e-01 9.87410471e-02 -2.55416542e-01
8.63608956e-01 5.38831115e-01 -2.59193540e-01 4.14255768e-01
-7.42453694e-01 -3.86450917e-01 3.17533672e-01 -9.19918776e-01
1.21717966e+00 4.27358644e-03 -5.44826329e-01 4.99404073e-01
4.97716248e-01 2.72112012e-01 -1.04057539e+00 3.19686800e-01
4.24072504e-01 -4.87371415e-01 1.74681857e-01 -1.26377177e+00
-2.16659248e-01 3.27733099e-01 1.94027990e-01 1.33971974e-01
7.00465143e-01 5.15384860e-02 -2.13871375e-01 1.84551641e-01
7.28090942e-01 -9.02659953e-01 -9.91308212e-01 -7.03505874e-02
1.01951575e+00 -7.78467774e-01 6.08787477e-01 -4.54664588e-01
-3.24597120e-01 9.24093485e-01 -1.18889958e-01 -1.89099714e-01
8.55236471e-01 7.19156563e-01 -3.54356468e-01 -4.62505579e-01
-1.22483385e+00 1.70076087e-01 3.94432619e-02 8.32612991e-01
6.35464668e-01 7.75793374e-01 -4.00028199e-01 6.13606930e-01
-5.92348397e-01 2.90644854e-01 4.83445734e-01 9.78865623e-01
-1.51349768e-01 -1.53369319e+00 -9.84382868e-01 2.99593449e-01
-6.44217908e-01 -4.15549994e-01 -1.67853892e-01 1.04977763e+00
-2.41679382e-02 1.11357462e+00 8.74665231e-02 1.69970363e-01
2.42358893e-01 1.03800841e-01 6.80804431e-01 -5.19373357e-01
-4.84781086e-01 -5.20208955e-01 9.98599827e-01 -1.05540109e+00
-4.25102770e-01 -1.01047933e+00 -1.14751196e+00 -7.42192030e-01
8.33405480e-02 4.33271796e-01 7.00174630e-01 7.47402370e-01
6.68174848e-02 5.35868227e-01 -7.05492198e-02 -5.02899945e-01
-6.49051905e-01 -7.10633218e-01 -2.65947133e-01 -3.03129941e-01
-2.19226763e-01 -8.08676243e-01 -1.26431271e-01 -4.66597565e-02] | [8.32784366607666, 6.023369312286377] |
f085880a-0ddd-4204-86e0-dca12a234ea6 | deep-reinforcement-learning-for-detecting | 1905.09207 | null | https://arxiv.org/abs/1905.09207v1 | https://arxiv.org/pdf/1905.09207v1.pdf | Deep Reinforcement Learning for Detecting Malicious Websites | Phishing is the simplest form of cybercrime with the objective of baiting people into giving away delicate information such as individually recognizable data, banking and credit card details, or even credentials and passwords. This type of simple yet most effective cyber-attack is usually launched through emails, phone calls, or instant messages. The credential or private data stolen are then used to get access to critical records of the victims and can result in extensive fraud and monetary loss. Hence, sending malicious messages to victims is a stepping stone of the phishing procedure. A \textit{phisher} usually setups a deceptive website, where the victims are conned into entering credentials and sensitive information. It is therefore important to detect these types of malicious websites before causing any harmful damages to victims. Inspired by the evolving nature of the phishing websites, this paper introduces a novel approach based on deep reinforcement learning to model and detect malicious URLs. The proposed model is capable of adapting to the dynamic behavior of the phishing websites and thus learn the features associated with phishing website detection. | ['Akbar Siami Namin', 'Moitrayee Chatterjee'] | 2019-05-22 | null | null | null | null | ['phishing-website-detection'] | ['adversarial'] | [-1.87822089e-01 -2.63623744e-01 7.44039342e-02 -3.16010751e-02
-4.60381582e-02 -1.01887810e+00 4.88426894e-01 3.91837656e-01
-5.36182344e-01 8.13313663e-01 -4.38802361e-01 -4.20217216e-01
2.13717178e-01 -1.27758658e+00 -3.57567638e-01 -5.71691453e-01
-1.84523210e-01 3.37192804e-01 4.23359126e-01 -2.68442392e-01
8.32031667e-01 5.62217236e-01 -1.02219653e+00 -1.55187637e-01
8.07193816e-01 1.00502908e+00 -4.53974992e-01 8.32446992e-01
-1.90069139e-01 4.80420321e-01 -1.06363475e+00 -9.77175653e-01
3.92117232e-01 -1.37916595e-01 -5.42992413e-01 -2.78819501e-01
2.19833583e-01 -9.16150033e-01 -6.60874784e-01 1.55226719e+00
3.02988350e-01 -1.35040075e-01 3.17341715e-01 -1.27694976e+00
-8.18093598e-01 1.11655280e-01 -3.96636099e-01 4.33169544e-01
4.27067339e-01 4.61415887e-01 7.23431885e-01 -4.91140872e-01
1.48959383e-01 1.01868284e+00 5.75568020e-01 6.94918811e-01
-7.74258494e-01 -5.63652873e-01 -3.84605974e-01 4.34874892e-01
-7.59870112e-01 -2.71269172e-01 4.79415774e-01 -4.31961149e-01
3.03524345e-01 1.63651973e-01 5.73929608e-01 1.30184436e+00
3.14045906e-01 6.95762813e-01 9.55348015e-01 -1.73710808e-01
4.61316139e-01 5.96014142e-01 6.90251946e-01 3.88372481e-01
9.49121714e-01 3.96824241e-01 -1.50450710e-02 -6.26796722e-01
5.48249066e-01 3.79025012e-01 -2.17022732e-01 -8.97564217e-02
-8.52963030e-01 1.19700980e+00 5.25803745e-01 1.20376483e-01
-5.63870788e-01 5.90055175e-02 8.87112558e-01 4.54439163e-01
-1.10756978e-01 3.68858159e-01 -3.55520904e-01 -1.82707161e-01
-2.76061386e-01 1.78444937e-01 1.12965214e+00 1.91818833e-01
6.95940793e-01 -1.51471004e-01 5.62895656e-01 3.40660602e-01
5.13785839e-01 5.22408664e-01 4.60619241e-01 -5.14340580e-01
4.26534027e-01 7.69788623e-01 5.21547318e-01 -1.23160613e+00
1.03137560e-01 7.35473782e-02 -5.28767526e-01 4.05102015e-01
8.46978962e-01 -3.52611780e-01 -6.23428166e-01 1.07357109e+00
5.08846641e-01 1.02157608e-01 -2.29051203e-01 7.58941293e-01
4.12699252e-01 5.81469238e-01 2.53488779e-01 2.06701338e-01
1.56846750e+00 -4.44980055e-01 -6.00928903e-01 -1.22156598e-01
3.23856294e-01 -4.60312486e-01 1.02300513e+00 3.46433342e-01
-7.43512988e-01 -3.52670960e-02 -9.17502284e-01 4.93934602e-01
-1.06054842e+00 -6.86714470e-01 5.73938251e-01 1.33312142e+00
-7.05612183e-01 8.49169970e-01 -3.51368427e-01 -4.18100476e-01
4.86177862e-01 4.02410477e-01 -1.54412657e-01 -9.46795940e-02
-2.02147269e+00 9.82144177e-01 4.47818279e-01 8.59721154e-02
-8.35537493e-01 -3.22217375e-01 -5.25910199e-01 2.42303163e-01
4.98514295e-01 -2.85417497e-01 1.06734610e+00 -8.42037201e-01
-1.11858416e+00 7.65828311e-01 7.14523673e-01 -3.48439604e-01
7.57407427e-01 -2.21050441e-01 -6.84861243e-01 2.40642637e-01
-2.10742578e-02 -2.45156080e-01 1.37560463e+00 -8.28606308e-01
-5.20300210e-01 -7.53372014e-01 6.56286702e-02 -4.84853648e-02
-5.94628930e-01 1.86116055e-01 2.98519850e-01 -2.78830320e-01
-5.20184875e-01 -5.97740114e-01 9.19261500e-02 -4.59943324e-01
-5.11581182e-01 -2.47950584e-01 1.23219609e+00 -1.34835279e+00
1.12577438e+00 -1.83179712e+00 -6.05982125e-01 4.38600391e-01
3.97443175e-01 1.07639992e+00 3.34808737e-01 5.77450752e-01
1.34342507e-01 4.09756333e-01 2.07881644e-01 4.69234020e-01
4.35814373e-02 -5.38803264e-02 -2.55812705e-01 5.03829956e-01
-2.51987427e-01 8.32819283e-01 -1.04118574e+00 4.58302535e-02
1.70132324e-01 4.12516266e-01 -9.88161042e-02 2.43967235e-01
1.74375907e-01 2.40251899e-01 -8.88899505e-01 7.81019151e-01
7.84588873e-01 -3.02686632e-01 -3.82045209e-02 4.56891567e-01
3.88936251e-01 5.63573003e-01 -9.70013916e-01 3.46076399e-01
-1.64585695e-01 2.67024487e-01 2.80538648e-01 -6.92474008e-01
7.25083888e-01 3.29561770e-01 2.56403890e-02 -3.06031853e-01
6.14726126e-01 3.33680183e-01 -1.34018049e-01 -5.79490781e-01
3.27415884e-01 1.04358897e-01 -5.14478050e-02 9.41920698e-01
-3.50829095e-01 8.98657441e-01 6.72825705e-03 4.58333343e-01
1.31479323e+00 -3.95655185e-01 4.27200139e-01 2.22368255e-01
9.74937677e-01 -2.23862424e-01 2.82956302e-01 5.98183751e-01
-9.76110160e-01 -3.14211011e-01 5.38216531e-01 -7.11337149e-01
-1.14838791e+00 -1.16603518e+00 2.04699263e-01 9.49322402e-01
2.04604819e-01 1.55490026e-01 -9.61477280e-01 -1.53581583e+00
4.54453826e-01 6.80813551e-01 -4.03310984e-01 -5.90585947e-01
-5.62592983e-01 -5.76750457e-01 8.36913407e-01 2.22387478e-01
1.03043497e+00 -1.43267083e+00 -4.77857977e-01 3.30678284e-01
-2.35966682e-01 -5.30702293e-01 -5.15769482e-01 -2.80421019e-01
-7.13229001e-01 -1.50671518e+00 -6.43644929e-01 -3.95532310e-01
6.36923790e-01 3.64824116e-01 4.32925642e-01 4.45759892e-01
-3.89098465e-01 3.01655889e-01 -3.14933091e-01 -9.54942331e-02
-8.88330221e-01 -1.46072492e-01 1.87002674e-01 4.74821210e-01
8.57915580e-01 -2.19519436e-01 -7.27453172e-01 2.43755534e-01
-8.78493667e-01 -1.05155206e+00 4.13301885e-01 7.93743968e-01
-5.63023210e-01 5.15363574e-01 1.03212869e+00 -1.11074626e+00
8.60244572e-01 -7.43023336e-01 -6.06706977e-01 2.13248610e-01
-8.41176093e-01 -2.79964894e-01 8.38539541e-01 -5.00895023e-01
-1.08072197e+00 -3.10598671e-01 7.02846199e-02 1.53021617e-02
-6.33880973e-01 9.44817290e-02 -5.59850514e-01 -4.59706753e-01
5.85658371e-01 4.80136424e-01 1.61600634e-01 -6.54290915e-01
2.15412602e-01 9.26421344e-01 2.50952959e-01 -1.61521826e-02
1.12415659e+00 3.39946955e-01 -3.43231142e-01 -7.07518578e-01
-2.36485139e-01 -8.43237460e-01 -5.60599566e-01 -3.74989092e-01
6.08624756e-01 -1.28235608e-01 -1.56028712e+00 1.13696289e+00
-1.25891507e+00 3.07419091e-01 3.70985031e-01 -3.00675258e-02
6.64151907e-02 1.20319414e+00 -1.14011979e+00 -8.94309044e-01
-3.45638454e-01 -3.77995789e-01 2.22754106e-01 3.64676148e-01
6.53359946e-03 -1.33519876e+00 2.42932707e-01 8.07921171e-01
5.71101546e-01 6.85388371e-02 7.49304473e-01 -1.30935025e+00
-5.56785762e-01 -1.11693966e+00 -4.83813614e-01 5.90408981e-01
2.54211187e-01 -1.49584532e-01 -9.31952357e-01 -4.35359597e-01
2.64759809e-01 -3.53240877e-01 4.87122387e-01 -1.53826699e-01
5.44751704e-01 -9.06868994e-01 -1.27213493e-01 3.00156444e-01
1.43625414e+00 5.29800475e-01 9.66131151e-01 7.57888317e-01
6.67470574e-01 7.78310120e-01 4.18731570e-01 4.99223769e-01
1.52134076e-01 5.83448634e-02 8.23152065e-01 3.81817251e-01
7.03417480e-01 -4.39268678e-01 5.88701785e-01 9.37185660e-02
1.63295940e-02 -6.43364638e-02 -6.07468784e-01 2.10881382e-01
-1.49684763e+00 -1.21674180e+00 -3.75765502e-01 2.61678171e+00
4.20938671e-01 8.39325115e-02 3.48423421e-01 2.95506984e-01
1.44205284e+00 -8.16158876e-02 -9.34289575e-01 -4.06318486e-01
4.65524852e-01 -1.98757455e-01 8.59454513e-01 3.33969772e-01
-1.19589984e+00 8.58870685e-01 5.42870283e+00 3.46999794e-01
-1.00248516e+00 -2.00881809e-02 4.47152346e-01 3.55147779e-01
3.02856058e-01 2.16735434e-02 -8.72132123e-01 1.09702241e+00
1.02609634e+00 1.00431316e-01 8.03135037e-01 9.77443755e-01
2.21107304e-01 -1.56799957e-01 -7.36545563e-01 7.16094851e-01
-1.37776002e-01 -8.23096335e-01 -1.99524045e-01 4.17923182e-01
5.11722490e-02 -4.91410285e-01 6.00771308e-02 2.98830837e-01
5.79370797e-01 -7.33848274e-01 2.04252228e-01 1.23304553e-01
2.45367080e-01 -7.86223769e-01 8.66286516e-01 4.57607180e-01
-5.56406498e-01 -3.64406228e-01 -3.71921301e-01 1.68613449e-01
2.97882140e-01 3.77014816e-01 -1.28406179e+00 3.95538546e-02
6.43930018e-01 4.67345446e-01 -6.08340442e-01 1.09746933e+00
-3.71510834e-01 7.70213842e-01 5.24730375e-03 -2.26403087e-01
1.69691131e-01 -2.31959537e-01 6.94765687e-01 7.89999366e-01
-1.05388639e-02 -1.48151875e-01 -2.61660606e-01 7.87052453e-01
-4.52640615e-02 1.23871397e-02 -8.17442536e-01 -4.91264254e-01
4.10610974e-01 1.38315451e+00 -6.96688950e-01 -5.47434270e-01
-2.74246752e-01 1.19335914e+00 1.51090309e-01 3.23105127e-01
-7.21660972e-01 -8.38135540e-01 6.58099532e-01 3.86222214e-01
2.22197860e-01 4.13399842e-03 -3.11016113e-01 -1.48057246e+00
-1.26789033e-01 -9.18601036e-01 8.26462269e-01 -1.71453372e-01
-1.68295455e+00 8.93034041e-03 -7.88085580e-01 -9.14263308e-01
-6.90220948e-03 -5.40074944e-01 -8.74823749e-01 9.09856617e-01
-1.55262315e+00 -8.03757668e-01 -1.84129521e-01 9.19316411e-01
5.51229604e-02 -4.11400467e-01 4.80685562e-01 3.29364002e-01
-4.93385524e-01 3.21093768e-01 1.61201164e-01 7.19288468e-01
6.42807841e-01 -1.09740281e+00 7.27251112e-01 8.55715394e-01
-5.48162997e-01 9.20456648e-01 3.85633171e-01 -1.21594799e+00
-1.06329024e+00 -7.16855288e-01 1.00763106e+00 -7.00496256e-01
1.01330853e+00 -2.47322127e-01 -1.24385953e+00 5.68887472e-01
-9.04437751e-02 -3.89315784e-01 6.07267678e-01 -3.65043253e-01
-5.46427727e-01 4.27791327e-02 -1.71573234e+00 2.63135403e-01
2.61406213e-01 -5.32785416e-01 -6.00806177e-01 2.84762025e-01
2.94142634e-01 5.71807027e-01 -4.31455076e-01 -5.41732788e-01
4.33004886e-01 -1.10483515e+00 1.07533658e+00 -1.05590487e+00
-1.34557886e-02 1.11294247e-01 4.05182391e-01 -1.11862648e+00
-2.10173234e-01 -7.18358040e-01 -4.05010074e-01 1.21374798e+00
-2.63044178e-01 -1.18151009e+00 1.18965364e+00 9.03779030e-01
6.90533102e-01 1.58701912e-01 -6.65072858e-01 -6.45985663e-01
-1.22056916e-01 3.31078470e-01 5.28088391e-01 1.32208169e+00
3.50218207e-01 1.60933703e-01 -5.25974691e-01 2.63836682e-01
1.45944309e+00 -5.23508072e-01 5.75922847e-01 -1.34455037e+00
-1.37262180e-01 -2.90883899e-01 -3.71947706e-01 -6.50623739e-01
-3.26463938e-01 -5.94873488e-01 -3.15235674e-01 -9.14513111e-01
-1.41179070e-01 -5.36194146e-01 -3.46436292e-01 3.15544695e-01
-2.07674131e-01 6.76602349e-02 1.72515124e-01 4.71961677e-01
-4.17677939e-01 1.06888130e-01 9.33222175e-01 -7.12533742e-02
-1.91225097e-01 7.87131429e-01 -6.74561501e-01 6.39376283e-01
9.62192237e-01 -5.01911640e-01 4.41416427e-02 4.91680980e-01
3.90110493e-01 3.15636098e-01 6.59617424e-01 -4.02784020e-01
7.58900642e-02 -3.32729667e-01 3.82728845e-01 -3.37044477e-01
-7.60463029e-02 -8.71099710e-01 -6.02213502e-01 1.00059521e+00
-2.88726892e-02 -1.74599484e-01 -5.23250282e-01 1.04226243e+00
2.63791025e-01 -5.40627360e-01 1.15316236e+00 -6.18909419e-01
-7.35789359e-01 2.91241586e-01 -8.34446609e-01 -1.42170832e-01
1.36979532e+00 -2.79485792e-01 -6.46227181e-01 -6.48187518e-01
-8.05034339e-01 1.83467194e-01 4.06643152e-01 4.28105682e-01
7.22207189e-01 -1.04000270e+00 -3.59740317e-01 3.95407945e-01
-3.25996220e-01 -6.86492205e-01 1.41554505e-01 2.47943595e-01
-7.33103037e-01 3.69364858e-01 -7.10495114e-01 1.91569537e-01
-1.35664403e+00 1.01526904e+00 2.89288342e-01 -2.89867818e-01
-3.21391582e-01 3.55073035e-01 -9.82617661e-02 -5.25831878e-01
2.13809222e-01 5.08351266e-01 -5.64060092e-01 4.21020463e-02
9.19501901e-01 1.20815051e+00 -7.52857402e-02 -5.36594987e-01
-2.76342571e-01 -1.94532320e-01 -5.89331746e-01 2.09808543e-01
9.83617425e-01 -2.83136725e-01 -3.51660997e-01 -1.72726765e-01
1.10975790e+00 -2.56131411e-01 -1.01569188e+00 -6.61833882e-02
3.42695296e-01 -9.81388986e-01 -3.15521955e-01 -9.70345795e-01
-1.01647902e+00 1.06657040e+00 3.53302360e-01 9.43604469e-01
4.68476832e-01 -3.48357022e-01 1.58109593e+00 4.86006081e-01
5.72047114e-01 -1.12663472e+00 5.04354239e-01 2.39207134e-01
1.47939637e-01 -1.46748579e+00 -2.47163400e-01 -2.56057531e-01
-6.11544847e-01 1.21854579e+00 6.58297837e-01 -3.23680669e-01
6.66370749e-01 -3.27752322e-01 1.79269925e-01 -3.03217232e-01
-1.22553356e-01 4.40598696e-01 -3.23316962e-01 1.06295979e+00
-3.19070518e-01 2.98175234e-02 -2.68177360e-01 2.31787965e-01
3.21504593e-01 -2.37857148e-01 9.61808264e-01 9.01171803e-01
-9.57660258e-01 -1.22353876e+00 -8.01181793e-01 4.11201775e-01
-8.00240815e-01 -1.01990877e-02 -8.08287144e-01 4.79722470e-01
-3.21448535e-01 7.71850467e-01 -4.68136996e-01 -2.75059432e-01
-1.05567932e-01 4.01981682e-01 -2.50673741e-01 -2.58230031e-01
-8.41892242e-01 -5.51489532e-01 2.42361706e-02 -1.65275767e-01
3.01096559e-01 -6.89561963e-01 -1.03880119e+00 -1.16196167e+00
-5.04195154e-01 2.26926263e-02 7.68280566e-01 7.28891730e-01
1.88086003e-01 -2.53538758e-01 1.12373412e+00 -3.02062392e-01
-1.34529674e+00 -6.27065837e-01 -8.80902708e-01 8.27412844e-01
5.05557597e-01 -4.27432120e-01 -6.89359069e-01 -2.28704378e-01] | [7.813441753387451, 9.994608879089355] |
f148bb45-db08-468a-b5e5-d4cec1973192 | ptrail-a-python-package-for-parallel | 2108.13202 | null | https://arxiv.org/abs/2108.13202v1 | https://arxiv.org/pdf/2108.13202v1.pdf | PTRAIL -- A python package for parallel trajectory data preprocessing | Trajectory data represent a trace of an object that changes its position in space over time. This kind of data is complex to handle and analyze, since it is generally produced in huge quantities, often prone to errors generated by the geolocation device, human mishandling, or area coverage limitation. Therefore, there is a need for software specifically tailored to preprocess trajectory data. In this work we propose PTRAIL, a python package offering several trajectory preprocessing steps, including filtering, feature extraction, and interpolation. PTRAIL uses parallel computation and vectorization, being suitable for large datasets and fast compared to other python libraries. | ['Amilcar Soares', 'Vinicius Prado da Fonseca', 'Chiara Renso', 'Vania Bogorny', 'Yaksh J. Haranwala', 'Salman Haidri'] | 2021-08-26 | null | null | null | null | ['trajectory-modeling'] | ['time-series'] | [-3.21056724e-01 -7.06742644e-01 5.33823185e-02 -1.00694895e-01
-3.15918714e-01 -8.66755843e-01 6.74824357e-01 1.00825107e+00
-5.88586926e-01 8.20763111e-01 3.45075101e-01 -4.00933146e-01
-2.77690917e-01 -1.25910902e+00 -5.94935060e-01 -4.59932983e-01
-2.27410614e-01 5.02879143e-01 5.89186549e-01 -3.19035351e-02
4.53923523e-01 9.79904771e-01 -1.57899392e+00 -8.42995197e-02
9.10329163e-01 6.87742829e-01 6.42055795e-02 4.32655603e-01
-2.70150363e-01 5.76292038e-01 -6.21796370e-01 -1.57954141e-01
2.33936459e-01 -2.89993640e-02 -5.67699790e-01 -4.42271322e-01
1.30209848e-01 -2.37777412e-01 -5.65617859e-01 8.49203408e-01
2.65911400e-01 3.82358164e-01 4.58254278e-01 -1.44382501e+00
-2.32955976e-03 4.02073294e-01 -1.56569213e-01 3.56870025e-01
6.85111046e-01 7.07928687e-02 2.11518049e-01 -8.79958391e-01
7.84630001e-01 7.16843247e-01 1.08951557e+00 -2.04854950e-01
-1.16334343e+00 -3.30431730e-01 -1.39883965e-01 2.87970185e-01
-1.68512630e+00 -1.43815652e-01 3.93269598e-01 -8.60456705e-01
9.08250093e-01 7.10342824e-01 8.09835494e-01 9.53097939e-01
1.95097312e-01 4.74002749e-01 5.04410148e-01 4.09551337e-02
3.61997902e-01 7.34517425e-02 2.37737760e-01 -7.09392801e-02
4.01658386e-01 5.11470176e-02 -8.51161331e-02 -5.16218483e-01
5.27635574e-01 4.86137569e-01 -2.32699111e-01 -3.14256310e-01
-1.31869256e+00 4.11104172e-01 3.71899426e-01 2.61118054e-01
-7.78252065e-01 -2.59313062e-02 5.54894626e-01 7.32145160e-02
1.95593402e-01 1.81557328e-01 -2.15001628e-01 -7.56591022e-01
-1.16912293e+00 9.79911327e-01 5.67849696e-01 1.14020741e+00
7.63683975e-01 -2.75831670e-01 -2.82108307e-01 1.89658076e-01
-9.72012281e-02 4.90969479e-01 3.97024453e-01 -4.38700527e-01
7.86631167e-01 1.06425023e+00 3.71249467e-01 -1.43097401e+00
-8.85334313e-01 5.01185283e-02 -9.10172105e-01 -6.11315072e-02
5.44443846e-01 -1.30459204e-01 -4.26098168e-01 9.97526109e-01
4.39424396e-01 7.56673738e-02 -1.80155203e-01 7.63807356e-01
6.97526336e-01 8.73955905e-01 1.49872005e-01 -4.74068969e-02
9.23192441e-01 -3.01470190e-01 -8.44185531e-01 2.42764786e-01
8.12244952e-01 -6.70844853e-01 8.19951355e-01 1.46135285e-01
-7.87131250e-01 -4.41048235e-01 -4.93922532e-01 7.34201670e-02
-9.32450891e-01 1.99029446e-01 6.09865427e-01 6.31592512e-01
-5.70305288e-01 8.86405170e-01 -1.01175582e+00 -4.09013212e-01
3.77101719e-01 2.89141446e-01 -4.73234534e-01 1.42308325e-01
-9.57889616e-01 8.42027545e-01 5.83673120e-01 1.00713655e-01
-1.15504667e-01 -9.56611454e-01 -7.47658074e-01 1.02336444e-02
8.06364790e-02 -1.85487032e-01 7.46085942e-01 -1.61875680e-01
-8.85279834e-01 1.91309676e-01 -1.85919911e-01 -4.95917380e-01
7.83139706e-01 -1.30096138e-01 -7.95431793e-01 -4.89397377e-01
7.04807416e-02 -2.92108476e-01 4.41197008e-01 -4.98134106e-01
-7.80170083e-01 -5.38803101e-01 -4.42261308e-01 -1.78170234e-01
-8.18401203e-02 1.64321601e-01 -3.34205091e-01 -7.51527548e-01
4.79433313e-02 -7.75220871e-01 -5.24306178e-01 -3.89526188e-01
-2.82358915e-01 -2.08512917e-01 9.70038474e-01 -9.62365746e-01
1.86747324e+00 -2.11467218e+00 -2.16543764e-01 5.91841340e-01
-1.12574898e-01 5.45143664e-01 3.78798217e-01 1.03543830e+00
1.01131067e-01 -5.59775122e-02 -2.32811972e-01 1.22936212e-01
-8.77994448e-02 5.04073175e-03 -4.07486081e-01 7.72298813e-01
-2.31639832e-01 5.96168578e-01 -1.03648818e+00 -2.01027393e-01
6.06747866e-01 4.57810313e-01 -2.56254047e-01 3.42057981e-02
1.41392827e-01 6.06386065e-01 -5.65265715e-01 6.74235523e-01
9.78565395e-01 3.36311519e-01 -2.82913148e-01 1.11470215e-01
-9.27948475e-01 2.70323783e-01 -1.52162921e+00 1.48154867e+00
-6.02710061e-02 7.24496245e-01 -4.16307032e-01 -6.68599963e-01
9.18646753e-01 2.82271564e-01 8.94959688e-01 -4.24402982e-01
1.37030214e-01 2.13618755e-01 -5.74719548e-01 -7.75830388e-01
1.17275715e+00 7.16885626e-01 -2.53662527e-01 2.83060104e-01
-5.53258300e-01 7.51273707e-02 4.69878882e-01 2.26892792e-02
1.21660697e+00 -5.80495521e-02 3.41571361e-01 -1.24489360e-01
5.84117472e-01 5.29089391e-01 6.68147802e-01 4.75326449e-01
-5.83408773e-03 5.14882386e-01 2.49217570e-01 -6.80780709e-01
-1.14011168e+00 -8.69270921e-01 -3.63399357e-01 6.80874944e-01
1.18856370e-01 -6.71652675e-01 -6.13100767e-01 -2.49746636e-01
2.83858120e-01 6.35394812e-01 -2.98570544e-01 1.06685519e-01
-7.90287435e-01 -8.25530052e-01 6.05711639e-01 5.46885192e-01
4.32573467e-01 -8.57093394e-01 -8.23169351e-01 7.09253013e-01
-1.53413564e-01 -8.34299147e-01 -4.92052525e-01 -3.62468600e-01
-8.95786822e-01 -1.01430809e+00 -4.78207439e-01 -1.67806670e-01
7.29954302e-01 3.02802742e-01 7.35716462e-01 4.34320867e-02
-1.98445782e-01 8.53027329e-02 -5.57671607e-01 -4.38195467e-01
-9.68103036e-02 3.19591135e-01 6.40252754e-02 1.00742705e-01
4.34429049e-01 -3.88469070e-01 -4.31114137e-01 3.57573330e-01
-8.36207747e-01 -1.36261731e-01 -2.20783371e-02 2.47135431e-01
5.02287388e-01 5.51427782e-01 2.69074917e-01 -5.81367075e-01
7.47472882e-01 -8.56838167e-01 -1.04479432e+00 4.81051952e-02
-4.18433696e-01 -3.79061162e-01 7.14885831e-01 -1.72560096e-01
-7.44491875e-01 3.95770103e-01 -2.72343546e-01 1.12461157e-01
-5.65022409e-01 5.82269847e-01 -3.46802413e-01 1.82921489e-04
6.70887530e-01 2.74536312e-01 -1.93286419e-01 -8.15975130e-01
1.36852294e-01 8.54175925e-01 5.43949723e-01 -8.45491737e-02
8.23117733e-01 5.54534197e-01 -5.46341725e-02 -1.01458514e+00
9.54241678e-02 -7.43786037e-01 -1.07567942e+00 -3.04853141e-01
4.72314775e-01 -4.92731363e-01 -6.56832933e-01 6.95859194e-01
-1.14537346e+00 -2.30469987e-01 -1.19629040e-01 7.18878210e-01
-1.54504195e-01 1.41705111e-01 -1.76701490e-02 -6.04087114e-01
-1.63547173e-01 -8.72077405e-01 5.08524060e-01 3.24100733e-01
-3.32472175e-01 -8.62721562e-01 3.37312877e-01 -2.92014420e-01
6.04742348e-01 5.94355404e-01 2.05544174e-01 -4.18095440e-01
-5.87958336e-01 -8.79591882e-01 -1.20709352e-01 -4.01320696e-01
3.69881168e-02 3.20718974e-01 -5.65775156e-01 -3.46138835e-01
-4.28239197e-01 5.27047575e-01 3.02798212e-01 1.08970433e-01
1.48219407e+00 -3.76985759e-01 -7.58231103e-01 6.03065431e-01
1.34233284e+00 2.53990024e-01 7.71568477e-01 5.67372441e-01
6.68896914e-01 4.36859459e-01 9.12780643e-01 7.10432827e-01
5.28967261e-01 9.89806473e-01 2.32688487e-01 1.67353064e-01
3.02830666e-01 -4.05626893e-01 1.77483875e-02 4.08781886e-01
-4.86199230e-01 -1.93132117e-01 -1.13536429e+00 7.87809014e-01
-2.15381169e+00 -1.32167864e+00 -1.09546888e+00 2.60244441e+00
3.72243971e-01 -3.19294542e-01 5.19194722e-01 4.24896240e-01
5.22854090e-01 -1.56881720e-01 -1.95002332e-01 -3.00091118e-01
-2.28212234e-02 -3.67232382e-01 1.20029509e+00 1.90722883e-01
-1.21735001e+00 6.02279186e-01 6.34581518e+00 5.09501159e-01
-1.22197604e+00 -3.15316655e-02 -2.38893420e-01 4.90209460e-02
-1.54642329e-01 -1.03000596e-01 -7.54362106e-01 1.01152217e+00
9.95754719e-01 -3.88072878e-01 4.29200351e-01 8.17563772e-01
6.48557425e-01 -2.67206281e-01 -7.42839336e-01 8.96887243e-01
-3.43049765e-01 -1.40007305e+00 -3.41466874e-01 1.11179858e-01
3.90363127e-01 1.32646024e-01 -2.99611300e-01 1.33833647e-01
-1.14086889e-01 -8.37800741e-01 8.13715398e-01 7.51408458e-01
3.04259300e-01 -9.06553864e-01 7.77691245e-01 5.24197280e-01
-1.29488778e+00 -3.75424139e-02 -4.07083690e-01 -3.52143288e-01
5.78431785e-01 7.75939643e-01 -8.21147740e-01 7.63932228e-01
9.32010472e-01 4.51458961e-01 -6.46865964e-01 1.81844819e+00
1.32642522e-01 4.43160057e-01 -5.73197007e-01 -1.04451608e-02
4.14146297e-02 -3.00367832e-01 7.05622613e-01 1.37247550e+00
8.31262350e-01 -2.17235852e-02 1.51847348e-01 4.79811519e-01
5.18899620e-01 2.76351511e-01 -8.39356601e-01 6.87104985e-02
7.47489333e-01 9.97911811e-01 -6.65346801e-01 -2.15121999e-01
-3.25404972e-01 7.46789038e-01 6.98752105e-02 3.21280807e-01
-9.78605866e-01 -8.24295700e-01 7.33201265e-01 5.61626196e-01
1.78235937e-02 -6.67511821e-01 -3.55339527e-01 -9.18295026e-01
1.11446105e-01 -2.91556239e-01 3.92311394e-01 -3.11301112e-01
-7.60102332e-01 4.64619249e-01 5.58631532e-02 -1.56602001e+00
-2.86230803e-01 -3.74873668e-01 -7.94054329e-01 8.72365296e-01
-8.69451284e-01 -8.30602944e-01 -5.71328282e-01 6.89606667e-01
2.87071727e-02 -4.09933962e-02 6.87872827e-01 9.40945506e-01
-7.10330606e-01 4.39508885e-01 4.54877973e-01 1.28027350e-01
2.98475534e-01 -1.05441403e+00 7.76863277e-01 9.74792361e-01
-1.82824105e-01 6.09142900e-01 8.14399302e-01 -8.93651068e-01
-1.62645590e+00 -1.48104501e+00 1.33132744e+00 -5.60001254e-01
6.64948702e-01 -3.57458711e-01 -1.25294578e+00 8.69119227e-01
-4.15712893e-01 7.37142935e-02 5.21248996e-01 -9.28364545e-02
2.32988745e-01 -2.24324137e-01 -1.28804171e+00 4.21348989e-01
8.42948258e-01 -1.95953757e-01 -3.27713639e-01 1.38887316e-01
4.46036421e-02 -7.56873131e-01 -1.16252601e+00 2.64118612e-01
4.05310601e-01 -8.06625545e-01 1.10813749e+00 -1.52586788e-01
-4.21234846e-01 -6.45320654e-01 2.16374427e-01 -1.30648279e+00
-3.07452410e-01 -5.74671805e-01 -6.28666952e-02 1.48048842e+00
2.98065662e-01 -7.76547492e-01 5.96639097e-01 9.51976240e-01
-9.60551649e-02 -1.84111804e-01 -1.03618646e+00 -1.03234208e+00
-2.33750314e-01 -7.12596595e-01 1.52085769e+00 9.52554166e-01
2.88345218e-01 -4.97106045e-01 -4.29605335e-01 3.97806495e-01
4.93863314e-01 -2.15324871e-02 1.29845500e+00 -1.28266370e+00
5.42045355e-01 -4.39878345e-01 -6.18352890e-01 -5.75269759e-01
-3.97580594e-01 -6.72123015e-01 -1.07427970e-01 -1.60715783e+00
-5.34273505e-01 -8.69964182e-01 2.46183738e-01 2.63043165e-01
7.52676129e-02 -5.47574572e-02 -1.27013266e-01 2.61925489e-01
-2.19890952e-01 3.49989563e-01 6.22944891e-01 5.57127297e-02
-7.62317717e-01 5.48740327e-01 2.98889261e-02 7.18761146e-01
1.01210439e+00 -6.83430672e-01 -7.51442164e-02 -3.01454753e-01
3.18787128e-01 -2.40803380e-02 5.07093072e-01 -1.29494083e+00
4.71314669e-01 -4.44109380e-01 3.92618030e-01 -1.32268989e+00
1.37257040e-01 -1.18178415e+00 8.35077226e-01 4.15263295e-01
2.35857323e-01 4.82381612e-01 2.88956702e-01 3.54477525e-01
-1.57603979e-01 -2.83114165e-01 1.51930034e-01 2.07684055e-01
-8.28070581e-01 5.77451825e-01 -8.72571588e-01 -3.68181497e-01
1.26613116e+00 -3.57011855e-01 -1.97810292e-01 1.20862782e-01
-4.83831704e-01 3.29752743e-01 6.76597953e-01 5.13063550e-01
4.43042070e-01 -1.60962737e+00 -4.02921915e-01 4.17843759e-01
2.64328003e-01 2.28017624e-02 3.45740438e-01 9.88209784e-01
-9.20776904e-01 4.75501925e-01 -3.60690266e-01 -3.90845478e-01
-1.25057781e+00 6.96465552e-01 -3.95888314e-02 1.02280267e-01
-1.19050717e+00 4.93097901e-02 -8.39044452e-01 -6.15558565e-01
3.22849900e-01 -6.70727253e-01 -4.09751832e-01 3.01164806e-01
9.72707987e-01 1.26225889e+00 2.99798012e-01 -9.10488427e-01
-4.08785760e-01 4.99213576e-01 5.17122984e-01 1.32216588e-01
1.07647705e+00 -1.97193637e-01 -2.49859393e-01 3.96067172e-01
8.98571253e-01 2.27614582e-01 -1.04502177e+00 5.95338233e-02
4.47561681e-01 -9.07624900e-01 -1.02991231e-01 -2.87970155e-01
-5.95822036e-01 5.01889706e-01 5.28377175e-01 5.36058903e-01
7.82690108e-01 -4.20723498e-01 9.80548680e-01 2.70448983e-01
6.92581713e-01 -1.23199308e+00 -9.90893126e-01 6.29063725e-01
8.96178603e-01 -8.33337724e-01 -1.18602172e-03 -1.81611493e-01
-3.43689322e-01 9.86638725e-01 1.85400292e-01 5.32694831e-02
8.23813796e-01 2.26834819e-01 -1.39562115e-01 -4.88383882e-02
4.16637585e-02 -1.71762243e-01 2.26407334e-01 7.03112841e-01
1.62742898e-01 3.60772312e-01 -5.59534431e-01 5.91598570e-01
-4.54757690e-01 2.46751443e-01 5.02711415e-01 9.88622606e-01
-3.94320875e-01 -1.06862199e+00 -8.12891781e-01 5.37259579e-01
-2.73325771e-01 4.09091800e-01 -3.42657827e-02 6.72823429e-01
3.44281524e-01 8.23202074e-01 3.92029822e-01 -2.95421392e-01
7.19115376e-01 -3.75338048e-01 1.02768779e-01 -1.04443029e-01
-8.21080148e-01 -5.70016265e-01 -6.18951917e-02 -7.00482845e-01
4.97387983e-02 -1.25736666e+00 -1.52890444e+00 -7.73386180e-01
3.67482565e-02 1.27935454e-01 9.73842084e-01 6.97428644e-01
5.48383832e-01 2.62931913e-01 5.14461637e-01 -9.18946981e-01
1.37957469e-01 -8.78073990e-01 -5.05303919e-01 4.52061623e-01
2.46133685e-01 -6.91322386e-01 1.12732567e-01 -2.57470459e-01] | [6.351580619812012, 1.7081716060638428] |
263982d2-ec53-4407-a8b9-86d3bf47a4f1 | cross-modal-fusion-distillation-for-fine | 2210.10486 | null | https://arxiv.org/abs/2210.10486v1 | https://arxiv.org/pdf/2210.10486v1.pdf | Cross-Modal Fusion Distillation for Fine-Grained Sketch-Based Image Retrieval | Representation learning for sketch-based image retrieval has mostly been tackled by learning embeddings that discard modality-specific information. As instances from different modalities can often provide complementary information describing the underlying concept, we propose a cross-attention framework for Vision Transformers (XModalViT) that fuses modality-specific information instead of discarding them. Our framework first maps paired datapoints from the individual photo and sketch modalities to fused representations that unify information from both modalities. We then decouple the input space of the aforementioned modality fusion network into independent encoders of the individual modalities via contrastive and relational cross-modal knowledge distillation. Such encoders can then be applied to downstream tasks like cross-modal retrieval. We demonstrate the expressive capacity of the learned representations by performing a wide range of experiments and achieving state-of-the-art results on three fine-grained sketch-based image retrieval benchmarks: Shoe-V2, Chair-V2 and Sketchy. Implementation is available at https://github.com/abhrac/xmodal-vit. | ['Anjan Dutta', 'Zeynep Akata', 'Yanbei Chen', 'Massimiliano Mancini', 'Abhra Chaudhuri'] | 2022-10-19 | null | null | null | null | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 4.19337526e-02 -2.82781839e-01 -4.33666408e-01 -3.37424964e-01
-1.19118440e+00 -6.97446942e-01 1.16105592e+00 -3.66451144e-02
-3.00356656e-01 4.47433412e-01 4.17209774e-01 -4.20687385e-02
-2.35556126e-01 -6.69325054e-01 -8.96675885e-01 -5.75920641e-01
3.78477901e-01 4.34398085e-01 -1.76045537e-01 -1.92625418e-01
1.00418486e-01 3.10693741e-01 -1.70943403e+00 6.93497360e-01
2.42738187e-01 1.10930634e+00 1.53023764e-01 6.58313096e-01
-3.37711990e-01 8.56453478e-01 -1.69276148e-01 -5.53250194e-01
2.00481534e-01 -3.27713531e-03 -8.67497146e-01 -1.42667547e-01
9.52991664e-01 -7.04339445e-01 -9.60951447e-01 6.53045774e-01
4.42286849e-01 1.38758704e-01 8.61716688e-01 -1.36616278e+00
-1.40812254e+00 4.05381709e-01 -4.61358279e-01 -1.62516132e-01
4.05022085e-01 2.13819966e-02 1.42088723e+00 -1.19508314e+00
8.92577350e-01 1.48184693e+00 2.70308435e-01 7.52317309e-01
-1.40909612e+00 -6.08137548e-01 2.58692682e-01 3.20031077e-01
-1.53606701e+00 -6.13877952e-01 8.73118758e-01 -2.98126787e-01
7.60686874e-01 -7.87093770e-03 3.43393713e-01 1.50252914e+00
-3.86887789e-01 1.35192740e+00 9.44885790e-01 -1.76420346e-01
-1.55022576e-01 6.77693412e-02 -2.60864031e-02 7.53803611e-01
1.13006467e-02 1.06370896e-01 -8.92337382e-01 -1.42190844e-01
9.33695078e-01 6.73102379e-01 -3.74495894e-01 -6.36642516e-01
-1.59515905e+00 7.33992219e-01 7.48564601e-01 1.93053961e-01
-4.07081723e-01 6.70660436e-01 5.21655262e-01 6.18708789e-01
1.97848201e-01 1.85168877e-01 -1.96050376e-01 1.88063443e-01
-8.73214781e-01 1.92826256e-01 4.87913519e-01 1.15481091e+00
7.05249846e-01 -3.22213918e-01 -3.44234854e-01 1.07600868e+00
4.47320759e-01 7.97162712e-01 2.01984107e-01 -1.00156999e+00
3.85547042e-01 3.61833453e-01 7.26859719e-02 -5.60290515e-01
3.45347285e-01 1.85469776e-01 -7.15900004e-01 1.31893665e-01
2.01598704e-01 4.78921056e-01 -1.24361634e+00 1.88096464e+00
-9.05887708e-02 1.24312587e-01 1.93562061e-01 1.12798774e+00
1.37096834e+00 6.03455245e-01 3.60243499e-01 6.12071157e-01
1.44741154e+00 -1.19585037e+00 -5.73644221e-01 7.72403702e-02
6.43545315e-02 -6.95687711e-01 1.12232924e+00 2.54481751e-02
-1.24928296e+00 -6.57922685e-01 -9.58537638e-01 -8.61172676e-01
-7.35749662e-01 2.03468397e-01 3.65879565e-01 -2.56636161e-02
-1.04176676e+00 4.57239509e-01 -5.95659673e-01 -3.97074521e-01
7.71307707e-01 8.80084485e-02 -9.14372683e-01 -6.87744260e-01
-1.08084285e+00 8.48197281e-01 1.15437068e-01 -3.94123979e-02
-1.17153299e+00 -8.37399423e-01 -9.48935330e-01 2.72416145e-01
4.52490777e-01 -1.13822615e+00 1.00531888e+00 -8.35313678e-01
-1.18523645e+00 1.09399951e+00 -1.85311630e-01 -3.71794514e-02
3.34830850e-01 -4.37420219e-01 -2.28220165e-01 5.41054249e-01
3.11605260e-02 1.15230167e+00 1.19953620e+00 -1.55306292e+00
-3.00221562e-01 -2.77055740e-01 3.32854182e-01 1.47550017e-01
-1.98710963e-01 -2.66166687e-01 -8.78125489e-01 -6.23652220e-01
-1.28872886e-01 -7.95089662e-01 4.29567426e-01 7.07347691e-01
-2.68887371e-01 -4.04574215e-01 8.56349587e-01 -4.86484885e-01
7.65360534e-01 -2.12058663e+00 7.46666193e-01 -2.80298628e-02
3.51820588e-01 2.04698712e-01 -8.22144508e-01 9.57957745e-01
1.82299630e-03 2.76922574e-03 -2.19112366e-01 -8.20719957e-01
4.17753667e-01 4.42116022e-01 -8.04769516e-01 2.26450041e-01
4.68164772e-01 1.45655417e+00 -9.44158852e-01 -4.68192041e-01
4.87020701e-01 9.98831570e-01 -2.17545807e-01 3.79141092e-01
-2.37394303e-01 1.25702515e-01 -2.49350891e-01 1.15124369e+00
5.71763635e-01 -4.38914150e-01 8.72232020e-02 -5.66087842e-01
3.76271904e-01 9.20253769e-02 -7.03459978e-01 2.52359962e+00
-7.40910649e-01 7.83429682e-01 7.75441304e-02 -9.08321738e-01
5.98607004e-01 3.21141094e-01 2.20335901e-01 -7.31052339e-01
-1.45823210e-01 6.59138784e-02 -6.08409584e-01 -3.98073018e-01
6.68751776e-01 -1.32504210e-01 -1.88793242e-01 5.37718177e-01
8.60459447e-01 -8.01286027e-02 -1.08262442e-01 6.18608773e-01
7.70142913e-01 5.32079995e-01 -2.46419311e-02 1.30113438e-01
4.51501310e-01 -4.52766597e-01 -1.67709589e-01 8.37377429e-01
3.26382965e-02 8.53757322e-01 2.33068720e-01 -2.50033051e-01
-9.75288272e-01 -1.55109572e+00 1.46698371e-01 1.36504209e+00
3.34817588e-01 -6.08266950e-01 7.45777041e-02 -7.42115736e-01
5.40877223e-01 2.79574126e-01 -9.36394334e-01 -1.86383039e-01
-1.93445981e-01 3.08249116e-01 6.82992935e-01 6.61873758e-01
2.34533608e-01 -1.00995755e+00 -3.39770794e-01 -2.96325415e-01
-1.54316068e-01 -1.18331659e+00 -5.17976344e-01 -7.41753802e-02
-5.01761615e-01 -1.09847510e+00 -1.27657425e+00 -6.45312309e-01
4.95199680e-01 6.30507171e-01 1.31284916e+00 2.92169124e-01
-2.82686770e-01 1.30531979e+00 -4.02092725e-01 -1.21178493e-01
-1.71411305e-03 -6.67348504e-02 -2.20637769e-01 1.59219265e-01
2.34876201e-01 -5.59849143e-01 -7.29261518e-01 2.53667906e-02
-1.17388809e+00 -7.53549188e-02 7.14098513e-01 1.20740676e+00
6.74078047e-01 -8.38999987e-01 4.52616483e-01 -6.41207814e-01
4.30542827e-01 -6.20657325e-01 -3.02400768e-01 8.10087025e-01
-1.41294643e-01 3.78436536e-01 2.85148144e-01 -5.58101237e-01
-9.75752532e-01 -1.29878357e-01 2.19711408e-01 -1.36365962e+00
-1.94000736e-01 4.44343895e-01 -5.23283482e-02 4.54864418e-03
1.82209715e-01 3.76400948e-01 7.06505775e-02 -6.44582510e-01
1.12270629e+00 3.80819321e-01 8.67850065e-01 -9.05014813e-01
8.15608382e-01 6.69838667e-01 -1.37317672e-01 -7.48967767e-01
-5.57473958e-01 -5.21967649e-01 -4.55289155e-01 -5.40445633e-02
8.11250269e-01 -1.16671038e+00 -7.62449026e-01 2.17396110e-01
-1.20641530e+00 -1.53940380e-01 -3.81929010e-01 1.15487210e-01
-5.51438928e-01 4.27023560e-01 -5.78903377e-01 -6.41676307e-01
-4.66406137e-01 -1.03368986e+00 1.75791287e+00 7.73623735e-02
1.32587984e-01 -1.01348984e+00 4.24016416e-02 4.30786610e-01
3.60641152e-01 -1.85471833e-01 9.17425275e-01 -4.64861721e-01
-8.78432453e-01 -1.32134125e-01 -6.83008492e-01 2.38092959e-01
-7.12907687e-02 -8.30596760e-02 -1.21175420e+00 -3.44379783e-01
-8.39154363e-01 -9.88319755e-01 1.59260523e+00 -3.39007042e-02
1.19932413e+00 -1.14690252e-02 -3.31738174e-01 5.77026010e-01
1.59670210e+00 -4.82179165e-01 7.09871531e-01 -1.17763162e-01
9.22588050e-01 4.73195612e-01 2.52306610e-01 1.38264909e-01
6.57385588e-01 6.92945898e-01 5.12444615e-01 -3.60810384e-02
-4.96516079e-01 -6.02024734e-01 1.73301727e-01 5.79672456e-01
-4.28819731e-02 -2.81933397e-01 -5.87368965e-01 9.57978368e-01
-2.00202489e+00 -1.13059437e+00 5.34107864e-01 2.00349021e+00
7.02294886e-01 -6.08505547e-01 4.33902256e-02 -2.93633878e-01
4.93983448e-01 4.67879802e-01 -5.49668372e-01 -1.42937720e-01
-1.86622798e-01 3.56274724e-01 1.31005839e-01 4.34685200e-01
-1.11438417e+00 9.43875730e-01 5.54022789e+00 7.12353587e-01
-9.82999325e-01 1.55491590e-01 8.69225264e-02 -3.49424839e-01
-8.13644350e-01 2.86182463e-02 -1.43195644e-01 2.15525553e-01
4.13098902e-01 1.83721766e-01 7.10284233e-01 4.62195128e-01
-7.31967986e-01 6.60944581e-02 -1.55881035e+00 1.38216865e+00
1.71535820e-01 -1.61743248e+00 5.31669378e-01 -1.32476896e-01
5.96888661e-01 2.20968232e-01 2.94811428e-01 5.25024354e-01
2.47458965e-01 -1.13494527e+00 6.76367104e-01 1.10357082e+00
1.20822275e+00 -3.13773304e-01 3.26300561e-01 -1.65351078e-01
-1.22058916e+00 -1.76213421e-02 -2.94527233e-01 3.65699023e-01
3.59993316e-02 5.31392433e-02 -2.53257632e-01 8.89248252e-01
6.53092921e-01 1.05061734e+00 -3.86857271e-01 5.67219973e-01
-1.69230983e-01 7.09194168e-02 -2.79280484e-01 4.36668038e-01
2.12736160e-01 -1.87950507e-02 3.78797024e-01 1.19431520e+00
1.20065555e-01 -4.76989821e-02 -1.65640578e-01 1.16481233e+00
-5.13272762e-01 -3.82077068e-01 -8.45516741e-01 -4.84954596e-01
6.37689292e-01 1.18924665e+00 -3.57739348e-03 -6.62842751e-01
-8.19812775e-01 1.30400074e+00 5.09248614e-01 8.20503891e-01
-5.86064041e-01 -2.77666450e-01 9.58714724e-01 -3.30692887e-01
7.44955838e-01 -6.79121464e-02 1.14654839e-01 -1.63688624e+00
5.47890961e-02 -6.25789702e-01 5.46205819e-01 -1.11743641e+00
-1.77243674e+00 4.28817928e-01 1.63441762e-01 -1.13646615e+00
-2.01837450e-01 -6.48082972e-01 -4.03394878e-01 1.13006496e+00
-1.98663437e+00 -1.96603441e+00 -4.13968146e-01 9.54992473e-01
5.27557254e-01 -2.80976415e-01 1.07555962e+00 4.45965171e-01
-1.39997989e-01 7.34489977e-01 2.12732386e-02 1.66980654e-01
1.10435843e+00 -1.18746448e+00 1.92162588e-01 3.43159437e-01
4.68011051e-01 8.59673619e-01 3.82563360e-02 -2.77637571e-01
-1.85369897e+00 -8.48197877e-01 7.51684606e-01 -5.60315788e-01
7.08545566e-01 -4.59682643e-01 -9.99527216e-01 6.85014784e-01
5.36939740e-01 5.20668089e-01 6.32560432e-01 2.21339151e-01
-1.33472264e+00 -3.96688804e-02 -8.01194966e-01 5.11296928e-01
1.00066376e+00 -1.53428066e+00 -7.59892881e-01 9.40648187e-03
5.50956428e-01 -1.38250768e-01 -1.14769280e+00 2.08850354e-01
1.14939666e+00 -6.74515069e-01 1.42921126e+00 -9.89577770e-01
8.65984619e-01 -1.34578735e-01 -7.00300813e-01 -1.09209776e+00
-1.50927454e-01 -1.68981478e-01 -3.97903383e-01 1.07387829e+00
7.95450807e-02 -2.99084812e-01 2.66822070e-01 3.84585112e-01
1.46563873e-01 -7.15662301e-01 -7.88655758e-01 -5.45907378e-01
2.55798608e-01 -2.16924176e-01 6.26643777e-01 1.03430188e+00
-1.19567879e-01 3.97724628e-01 -5.73679984e-01 -9.28779133e-03
7.42597580e-01 5.91705859e-01 7.92114258e-01 -1.03755176e+00
-3.08372229e-01 -6.35786772e-01 -3.57447714e-01 -1.12707543e+00
4.53536570e-01 -1.18184042e+00 -2.44850650e-01 -1.73469079e+00
4.56427515e-01 -2.22218275e-01 -6.79405034e-01 8.44652474e-01
-1.18075334e-01 5.82382262e-01 6.83675408e-01 4.22073901e-01
-8.54872525e-01 7.66586661e-01 1.14349091e+00 -5.86107254e-01
3.23186278e-01 -7.27709234e-01 -5.16547203e-01 1.80160299e-01
3.75687957e-01 -5.78654185e-02 -6.04945660e-01 -8.91448617e-01
4.63486947e-02 2.50551462e-01 1.01431060e+00 -5.20702541e-01
2.90984273e-01 4.31239344e-02 3.44123185e-01 -5.43215692e-01
9.58769798e-01 -9.87154603e-01 -2.88973842e-02 -1.55721933e-01
-8.11817348e-01 -1.94201514e-01 2.38725752e-01 8.44612181e-01
-4.53202546e-01 1.64627686e-01 3.44632745e-01 -1.05156600e-01
-9.54501092e-01 4.74413902e-01 1.63174987e-01 -1.47498041e-01
5.97106099e-01 1.39071360e-01 -6.05599403e-01 -6.07514262e-01
-8.15720677e-01 2.81472623e-01 5.74941874e-01 7.54368901e-01
9.75820601e-01 -1.80613017e+00 -6.15241468e-01 1.65062174e-01
7.74206877e-01 -3.84946108e-01 4.87997890e-01 6.92102969e-01
-1.46484720e-02 5.02200305e-01 -5.11635959e-01 -5.80591500e-01
-1.16252041e+00 8.75387371e-01 2.06082284e-01 6.47380576e-02
-5.97054839e-01 8.13933074e-01 4.30701166e-01 -4.76613998e-01
2.56013066e-01 -1.50142506e-01 2.08486259e-01 1.11651607e-01
4.95530695e-01 1.60044387e-01 -2.95797408e-01 -6.60854936e-01
-4.94337052e-01 6.56189442e-01 -1.11424364e-01 -2.91523784e-01
1.28899682e+00 -1.41305968e-01 -7.34398663e-02 5.54363668e-01
1.72623038e+00 -3.96271914e-01 -1.26630700e+00 -6.69429362e-01
-3.98420721e-01 -5.30069351e-01 1.18518978e-01 -1.02451670e+00
-1.15404308e+00 1.28498709e+00 3.95038456e-01 -2.84140229e-01
1.07863081e+00 3.89566422e-01 7.30328500e-01 6.18744195e-01
1.40571490e-01 -7.08465576e-01 1.97584212e-01 2.17576250e-01
1.27554965e+00 -1.48926461e+00 -9.02877562e-03 7.75378495e-02
-6.62998974e-01 1.09438610e+00 2.13723734e-01 -3.72483671e-01
5.56500316e-01 -2.32693136e-01 -9.22283009e-02 -3.39271098e-01
-1.08495247e+00 -5.04882514e-01 7.55748272e-01 5.67080855e-01
2.68692851e-01 2.78859213e-02 1.34255201e-01 4.65916246e-01
6.81567907e-01 1.21867962e-01 -5.64234741e-02 9.67318654e-01
1.73184648e-01 -1.23512888e+00 -1.40971094e-01 2.67643303e-01
-6.49970537e-03 -2.84640282e-01 -6.57757878e-01 8.25526476e-01
-1.83301806e-01 5.85176229e-01 4.18154933e-02 -2.91332990e-01
1.97996646e-01 3.70860547e-01 1.01543188e+00 -2.11733520e-01
-1.88488692e-01 -3.73591259e-02 -3.78536694e-02 -7.39158452e-01
-7.11459219e-01 -3.57033104e-01 -7.21519828e-01 -2.66108692e-01
4.90563698e-02 -2.01692730e-01 5.72927356e-01 6.86628878e-01
6.41967893e-01 3.47866386e-01 1.27938718e-01 -1.13675845e+00
-4.38019842e-01 -7.65956104e-01 -3.90485078e-01 7.86153853e-01
6.86034441e-01 -9.78885472e-01 -5.35452925e-02 8.71953275e-03] | [11.334315299987793, 0.8237981796264648] |
481e7b86-0de7-4232-81c2-d221d6103ce5 | streamlining-models-with-explanations-in-the | 2302.07760 | null | https://arxiv.org/abs/2302.07760v1 | https://arxiv.org/pdf/2302.07760v1.pdf | Streamlining models with explanations in the learning loop | Several explainable AI methods allow a Machine Learning user to get insights on the classification process of a black-box model in the form of local linear explanations. With such information, the user can judge which features are locally relevant for the classification outcome, and get an understanding of how the model reasons. Standard supervised learning processes are purely driven by the original features and target labels, without any feedback loop informed by the local relevance of the features identified by the post-hoc explanations. In this paper, we exploit this newly obtained information to design a feature engineering phase, where we combine explanations with feature values. To do so, we develop two different strategies, named Iterative Dataset Weighting and Targeted Replacement Values, which generate streamlined models that better mimic the explanation process presented to the user. We show how these streamlined models compare to the original black-box classifiers, in terms of accuracy and compactness of the newly produced explanations. | ['Elvio G. Amparore', 'Alan Perotti', 'Paolo Bajardi', 'Francesco Lomuscio'] | 2023-02-15 | null | null | null | null | ['feature-engineering'] | ['methodology'] | [ 4.44174975e-01 6.04705632e-01 -4.48527753e-01 -6.73918366e-01
-1.95066452e-01 -5.58492184e-01 7.81892896e-01 6.33779109e-01
4.71068397e-02 6.45847857e-01 3.85053217e-01 -2.79777855e-01
-5.98791659e-01 -8.03909123e-01 -3.85694206e-01 -5.86110950e-01
1.66209012e-01 6.63012743e-01 7.51427710e-02 4.61381786e-02
6.41473174e-01 4.79231238e-01 -2.05275774e+00 6.77498817e-01
1.07217371e+00 7.85017610e-01 -1.25894696e-01 6.83036029e-01
-3.11100692e-01 7.60565460e-01 -3.81171197e-01 -1.67670265e-01
9.02853906e-02 -7.39407301e-01 -8.11166823e-01 3.08120906e-01
7.27272332e-02 -8.95922482e-02 2.85563648e-01 6.92278624e-01
-2.34087318e-01 -5.39278202e-02 8.88485670e-01 -1.42914820e+00
-5.45135915e-01 9.11919951e-01 3.40138264e-02 -3.37394983e-01
4.92555469e-01 -2.77593471e-02 1.31217873e+00 -9.47255194e-01
5.11738539e-01 1.20348001e+00 3.51295054e-01 6.52602434e-01
-1.35256028e+00 -3.24246705e-01 6.41950190e-01 4.66000825e-01
-8.48780990e-01 -2.12265074e-01 7.75416553e-01 -4.40434933e-01
5.47106981e-01 7.09169924e-01 8.22372854e-01 5.47635436e-01
1.22903600e-01 7.51620650e-01 1.00372458e+00 -7.88976371e-01
6.57331347e-01 7.87689388e-01 5.92675269e-01 6.93837643e-01
4.84232754e-01 2.12188989e-01 -7.39843905e-01 -2.54557163e-01
3.21493208e-01 2.22941041e-01 -3.16222280e-01 -1.01355028e+00
-9.93430197e-01 1.04715657e+00 6.12240255e-01 4.88746017e-01
-3.91828120e-01 5.86400405e-02 -6.51111156e-02 3.65218788e-01
4.62208211e-01 1.02101493e+00 -8.20270479e-01 1.38580516e-01
-6.10003173e-01 1.76380873e-01 7.96134114e-01 5.36484003e-01
1.01806986e+00 -3.49720657e-01 -1.10051967e-01 3.25670153e-01
5.64018965e-01 1.43645629e-01 5.33069730e-01 -7.85369694e-01
-7.37995654e-02 1.30712509e+00 9.94012207e-02 -8.86982262e-01
-4.29360986e-01 -5.16579688e-01 -3.16079080e-01 4.33470368e-01
4.85847771e-01 2.31509618e-02 -8.39824557e-01 1.51593053e+00
3.95330817e-01 -4.60130960e-01 2.28856862e-01 8.85315418e-01
5.69887459e-01 3.94244522e-01 -4.46508974e-02 -1.92771312e-02
1.07974839e+00 -9.89299178e-01 -5.46207488e-01 -3.46876740e-01
1.03397357e+00 -2.02995583e-01 1.15115786e+00 4.78900552e-01
-5.78616619e-01 -5.44513643e-01 -1.13476169e+00 1.84816271e-01
-4.97746706e-01 1.42140925e-01 8.37590039e-01 5.22719860e-01
-6.93122387e-01 1.06683791e+00 -4.26212341e-01 -4.12318736e-01
2.91800231e-01 5.68489790e-01 -2.83533782e-01 8.59958678e-03
-9.37885225e-01 9.85457301e-01 3.64544868e-01 -3.41049582e-02
-5.34543991e-01 -6.35356843e-01 -6.99871242e-01 4.48668420e-01
5.31169951e-01 -6.22849822e-01 1.28537059e+00 -1.30993664e+00
-1.18284249e+00 3.44711512e-01 -2.89126247e-01 -2.55142570e-01
4.69369471e-01 -2.68406630e-01 -1.25962030e-02 -1.40119612e-01
-1.30003452e-01 7.32116222e-01 8.26854825e-01 -1.49994695e+00
-8.30613852e-01 -4.43097830e-01 1.42574817e-01 1.51470527e-01
-2.66271353e-01 -3.19416165e-01 3.99769358e-02 -2.65282243e-01
2.85111129e-01 -7.28979111e-01 -4.58825141e-01 1.70422941e-01
-3.99531126e-01 -2.66412973e-01 7.49574959e-01 -3.33361663e-02
1.10020173e+00 -2.07646871e+00 1.34178266e-01 6.41969442e-01
4.32039976e-01 -2.69919876e-02 -3.50889377e-02 3.24225187e-01
-3.03460300e-01 1.92596763e-01 -1.74744651e-01 -1.08540319e-01
7.28160515e-02 3.43659669e-01 -2.69870192e-01 5.99434339e-02
3.68657917e-01 6.61402524e-01 -9.89324570e-01 -2.16305479e-01
2.51172006e-01 1.52901933e-01 -8.32496881e-01 3.88356268e-01
-3.71198505e-01 4.11162496e-01 -7.07266688e-01 3.02366823e-01
1.18003599e-01 -2.20887870e-01 1.39436409e-01 7.56742731e-02
3.47034372e-02 3.36048633e-01 -1.10711026e+00 1.05993712e+00
-3.63799721e-01 5.46929300e-01 -5.58562577e-01 -8.46520722e-01
1.25257790e+00 1.50927097e-01 1.39344230e-01 -2.92819917e-01
1.56582057e-01 3.70444179e-01 1.69570208e-01 -3.68264139e-01
1.91224083e-01 -1.20172657e-01 1.10267147e-01 7.45077372e-01
-1.54427007e-01 -1.00338228e-01 1.08274162e-01 1.18628755e-01
9.85041320e-01 1.25884950e-01 7.12133110e-01 -2.26206750e-01
5.87533474e-01 4.16754276e-01 3.39207381e-01 9.37557399e-01
3.30013692e-01 5.47348738e-01 7.61841714e-01 -9.48982835e-01
-7.69025087e-01 -5.28155804e-01 4.43347767e-02 8.54930818e-01
-3.65793845e-03 -6.52882755e-01 -7.73150682e-01 -1.11223364e+00
1.48279697e-01 1.05238175e+00 -1.09781814e+00 -5.72250545e-01
-1.19911686e-01 -1.09168373e-01 -2.63484716e-01 5.01135707e-01
3.88988853e-02 -1.22556937e+00 -1.09142590e+00 9.52844024e-02
3.21632355e-01 -3.91763747e-01 -3.24766748e-02 6.18073940e-01
-1.23640490e+00 -1.19589329e+00 -2.33663455e-01 -2.69239992e-01
1.26554072e+00 4.90200333e-02 8.81631076e-01 5.94504952e-01
-1.70697477e-02 3.38166915e-02 -6.38038456e-01 -4.70127434e-01
-5.47832191e-01 1.39150009e-01 -7.69007877e-02 1.73230737e-01
4.30014163e-01 -3.54682624e-01 -3.72719675e-01 3.30880404e-01
-7.76609302e-01 4.41346616e-01 8.69367719e-01 8.82560372e-01
2.79857516e-01 5.32759028e-03 4.91569698e-01 -1.35912180e+00
6.06517494e-01 -2.98044950e-01 -3.41634125e-01 5.50342917e-01
-1.22129440e+00 8.24628115e-01 7.67533183e-01 -5.49483836e-01
-1.00726235e+00 4.61250305e-01 3.18213791e-01 -9.11776945e-02
-3.71017128e-01 5.99004388e-01 -2.95910448e-01 3.45823057e-02
9.16386127e-01 -3.10751852e-02 -1.91267490e-01 -5.72965622e-01
6.92864478e-01 6.01764321e-01 2.47005463e-01 -4.90026563e-01
7.80545115e-01 9.23939049e-02 -1.98367044e-01 -6.22232705e-02
-9.10149336e-01 -1.55977502e-01 -1.00643384e+00 -2.27066129e-01
4.72451359e-01 -1.28029034e-01 -4.72686946e-01 -1.49179205e-01
-1.05936027e+00 -1.57908782e-01 -7.93414593e-01 4.71692979e-01
-5.59938669e-01 -2.19658390e-01 1.55013323e-01 -8.88438880e-01
-5.68797216e-02 -1.17553699e+00 8.18652928e-01 3.60281587e-01
-9.24322307e-01 -9.48458135e-01 -1.60236776e-01 6.10582344e-02
2.57400155e-01 -6.88646035e-03 1.52387881e+00 -1.33419073e+00
-4.85493809e-01 -5.52321911e-01 -9.90121067e-02 -1.16531905e-02
2.38694265e-01 5.51682785e-02 -1.17996514e+00 2.52490610e-01
-1.60015911e-01 1.01088084e-01 6.87983930e-01 1.54041469e-01
1.16491818e+00 -4.97356445e-01 -5.35908818e-01 2.30281413e-01
9.71840143e-01 2.93745339e-01 2.63301939e-01 3.87334675e-01
4.68391657e-01 9.80717301e-01 7.59916723e-01 3.47146809e-01
2.28822213e-02 5.31552374e-01 4.85553056e-01 -1.77953213e-01
9.61489081e-02 -3.56342435e-01 8.07530880e-02 1.56391904e-01
2.24633068e-01 -7.36710848e-03 -7.21991599e-01 4.80495095e-01
-2.16402054e+00 -5.82546055e-01 -1.95071191e-01 2.20287704e+00
4.89743352e-01 3.60217035e-01 -1.05475649e-01 5.39447069e-01
3.99730712e-01 -2.33027875e-01 -7.11874545e-01 -8.15493345e-01
3.74223888e-01 -4.28946912e-02 7.83921778e-03 7.10381150e-01
-6.25728130e-01 7.00993121e-01 6.42263222e+00 3.06065291e-01
-8.51712286e-01 -4.74590808e-01 5.37370145e-01 4.47629355e-02
-8.97929609e-01 5.61551809e-01 -5.47299027e-01 -5.88105172e-02
6.53341830e-01 -3.67609173e-01 3.10285866e-01 1.12621772e+00
1.86979771e-01 -2.25636557e-01 -1.85707021e+00 3.23604763e-01
-5.87732159e-02 -1.29201293e+00 4.40389007e-01 3.51376757e-02
3.61262530e-01 -7.61485934e-01 -1.45136207e-01 6.60748854e-02
3.72578233e-01 -1.00673771e+00 7.85228252e-01 5.96867621e-01
2.49718845e-01 -6.96084619e-01 8.15753400e-01 5.62512457e-01
-8.37612212e-01 -5.18952549e-01 -5.53982295e-02 -4.89224583e-01
-3.28888357e-01 5.40834963e-01 -1.19692791e+00 4.49675441e-01
2.69219309e-01 5.74663401e-01 -8.25800002e-01 9.67110157e-01
-7.04171956e-01 4.63646173e-01 -1.02519738e-02 -4.80045557e-01
3.42621431e-02 1.57279029e-01 3.70314568e-01 7.41878390e-01
3.24370772e-01 1.39393538e-01 -1.62747070e-01 1.05479288e+00
3.80244762e-01 1.78794414e-01 -5.90565443e-01 1.15710132e-01
3.31938535e-01 1.27896333e+00 -8.61752272e-01 -4.75039810e-01
-5.52139990e-02 5.34312546e-01 2.28809565e-01 7.03177899e-02
-3.90382588e-01 -3.23055804e-01 5.07938266e-01 1.33557171e-01
1.63289919e-01 2.96739221e-01 -7.61430800e-01 -1.03936112e+00
-1.64681733e-01 -9.57815349e-01 2.96091467e-01 -1.00702834e+00
-9.91982341e-01 6.98032200e-01 1.32371396e-01 -9.76738214e-01
-5.00330269e-01 -5.14255643e-01 -8.51503253e-01 5.41321278e-01
-1.12632620e+00 -7.39239872e-01 -3.58865231e-01 1.76322490e-01
5.17864406e-01 -2.04340331e-02 9.53850448e-01 -3.72597426e-01
-3.11208606e-01 2.74269342e-01 -2.24095553e-01 -3.14747632e-01
3.23976070e-01 -1.41061878e+00 2.83694565e-01 6.48479104e-01
4.26466495e-01 7.24518538e-01 1.12751031e+00 -5.37572324e-01
-9.87095356e-01 -7.16926336e-01 1.31223273e+00 -5.56097150e-01
3.41860801e-01 -2.22381219e-01 -1.06884360e+00 4.42228884e-01
-1.79848582e-01 -2.02805147e-01 6.75388277e-01 4.37568456e-01
-2.58908004e-01 -1.68082818e-01 -1.24047208e+00 7.01687634e-01
7.59067476e-01 -7.42985606e-02 -8.54713857e-01 9.95462239e-02
6.82872772e-01 1.24598362e-01 -3.59636933e-01 2.98053294e-01
7.34037936e-01 -9.04253244e-01 4.61525470e-01 -1.01575673e+00
4.69158828e-01 -3.25257093e-01 5.84742427e-02 -1.61286485e+00
-4.60468471e-01 -2.87545264e-01 1.10664539e-01 1.05103242e+00
1.00581241e+00 -6.88115954e-01 8.46010864e-01 1.13308275e+00
9.81851248e-04 -9.06504095e-01 -5.04897416e-01 -2.43059054e-01
-3.84436786e-01 -3.83393496e-01 1.08573008e+00 7.24897742e-01
5.86960018e-01 2.79070735e-01 -1.01038493e-01 5.94480895e-02
4.01435822e-01 5.45990050e-01 8.20397496e-01 -1.75699818e+00
-1.35279045e-01 -4.53670084e-01 -3.21104109e-01 -7.82759905e-01
-1.05980061e-01 -8.41252506e-01 1.72996029e-01 -1.60674918e+00
3.22252333e-01 -5.57479441e-01 -2.20611572e-01 9.41779196e-01
-3.56302828e-01 -2.08158910e-01 2.50477850e-01 1.93927154e-01
-3.41142058e-01 3.10976088e-01 1.00847173e+00 9.69858691e-02
-4.40775722e-01 1.14837423e-01 -1.17212451e+00 9.62991834e-01
6.97501659e-01 -9.31938589e-01 -5.26801527e-01 -3.09591174e-01
3.30190837e-01 -1.81598380e-01 3.23533326e-01 -8.23819041e-01
2.55377233e-01 -4.35927123e-01 6.07998192e-01 -2.36770794e-01
3.42662036e-02 -1.13153684e+00 2.14065731e-01 7.51547694e-01
-1.13073087e+00 -7.72970766e-02 -1.45019546e-01 3.67619514e-01
-6.77875504e-02 -4.93995994e-01 3.74534070e-01 7.17200935e-02
-5.31786144e-01 -1.79700956e-01 -4.56153691e-01 -6.05059326e-01
1.09386754e+00 -4.51365113e-01 -2.88806736e-01 -4.61006194e-01
-9.41264987e-01 1.93587348e-01 4.26828682e-01 4.02759343e-01
6.63305879e-01 -1.22767639e+00 -4.40062851e-01 5.27830184e-01
3.81809205e-01 -1.65294096e-01 -2.58282185e-01 5.05207777e-01
-9.91004854e-02 5.20600438e-01 -1.16722316e-01 -4.78435963e-01
-1.38839030e+00 6.02103591e-01 2.78167546e-01 -1.96290299e-01
-5.03117800e-01 6.06664836e-01 3.38779151e-01 -6.21177971e-01
2.31228158e-01 -6.79719329e-01 -4.95621502e-01 1.34815857e-01
4.58700866e-01 8.35815519e-02 7.88204223e-02 -3.72821055e-02
-3.38579834e-01 5.12029409e-01 -7.36391917e-02 -3.65000069e-02
1.52069962e+00 1.04843594e-01 1.53840706e-01 5.66897988e-01
7.65201032e-01 -7.72760883e-02 -1.26415157e+00 -2.54149407e-01
3.88308614e-01 -6.26567304e-01 -1.58825234e-01 -1.06985343e+00
-9.04285729e-01 8.58828485e-01 2.00557724e-01 4.85436708e-01
1.13321412e+00 2.55616099e-01 3.47085036e-02 5.59302092e-01
1.35242566e-01 -5.73641062e-01 1.88150182e-02 2.08340511e-01
8.97381365e-01 -9.17984009e-01 1.60528392e-01 -4.71514374e-01
-7.63392270e-01 1.53135467e+00 5.82315207e-01 5.34343757e-02
4.15237993e-01 1.28882157e-03 2.50056148e-01 -3.48553896e-01
-1.25349724e+00 -1.63136840e-01 5.75226963e-01 5.75146735e-01
3.19572568e-01 -3.34300809e-02 -2.56478637e-01 7.10084260e-01
-4.23074365e-01 3.75419930e-02 4.05139297e-01 7.82478929e-01
-8.09880495e-01 -1.28130519e+00 -3.91722560e-01 5.24373114e-01
1.89666107e-01 1.07231408e-01 -1.11920774e+00 6.98877096e-01
1.28405228e-01 1.02624214e+00 -9.88166034e-02 -6.07791066e-01
4.22117740e-01 3.12785804e-01 1.00762233e-01 -8.13459337e-01
-6.64591074e-01 -2.82379448e-01 1.37866631e-01 -4.78861898e-01
-6.16099127e-02 -6.03769302e-01 -1.44654500e+00 -2.46810373e-02
-6.60337031e-01 7.03398824e-01 8.13928604e-01 1.38831091e+00
3.40556055e-01 6.46853745e-01 8.64985466e-01 -7.15413749e-01
-6.30620539e-01 -7.04413772e-01 -2.12748051e-01 4.56555843e-01
3.98404360e-01 -5.99304259e-01 -5.76470256e-01 -8.07141047e-03] | [8.824880599975586, 5.789315700531006] |
d0706ec4-93dc-4c45-aebf-d2a644444b7d | deformable-mesh-transformer-for-3d-human-mesh | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yoshiyasu_Deformable_Mesh_Transformer_for_3D_Human_Mesh_Recovery_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yoshiyasu_Deformable_Mesh_Transformer_for_3D_Human_Mesh_Recovery_CVPR_2023_paper.pdf | Deformable Mesh Transformer for 3D Human Mesh Recovery | We present Deformable mesh transFormer (DeFormer), a novel vertex-based approach to monocular 3D human mesh recovery. DeFormer iteratively fits a body mesh model to an input image via a mesh alignment feedback loop formed within a transformer decoder that is equipped with efficient body mesh driven attention modules: 1) body sparse self-attention and 2) deformable mesh cross attention. As a result, DeFormer can effectively exploit high-resolution image feature maps and a dense mesh model which were computationally expensive to deal with in previous approaches using the standard transformer attention. Experimental results show that DeFormer achieves state-of-the-art performances on the Human3.6M and 3DPW benchmarks. Ablation study is also conducted to show the effectiveness of the DeFormer model designs for leveraging multi-scale feature maps. Code is available at https://github.com/yusukey03012/DeFormer. | ['Yusuke Yoshiyasu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['human-mesh-recovery'] | ['computer-vision'] | [ 2.47388165e-02 2.99690157e-01 -4.92819324e-02 -2.51766622e-01
-9.98123705e-01 -2.64123827e-02 3.06538790e-01 -3.91131520e-01
-3.13011035e-02 3.40765506e-01 4.38423276e-01 2.39239231e-01
3.68272930e-01 -8.23269665e-01 -1.25394142e+00 -2.52267361e-01
2.69335508e-01 8.60345721e-01 9.81280804e-02 -2.63451099e-01
-1.83565006e-01 4.49187100e-01 -1.65101480e+00 3.88584524e-01
6.46874189e-01 1.05888999e+00 1.25901520e-01 6.05808854e-01
8.57238844e-02 3.30633163e-01 -1.67164598e-02 -4.31417972e-01
5.34991443e-01 -2.15565085e-01 -8.96507323e-01 2.91886747e-01
9.80072439e-01 -6.26635313e-01 -5.70470154e-01 8.16014111e-01
5.37907839e-01 -1.16632879e-01 5.89500546e-01 -9.00940418e-01
-8.90708625e-01 2.28162572e-01 -8.30401599e-01 1.17179817e-02
5.09248376e-01 3.10373127e-01 7.36596107e-01 -1.34127164e+00
8.79450142e-01 1.49587905e+00 8.33016872e-01 6.86211169e-01
-1.14459908e+00 -7.57288992e-01 4.82821800e-02 -1.20763265e-01
-1.36288738e+00 -4.55035567e-01 8.89918804e-01 -5.74624956e-01
9.70031023e-01 2.65804708e-01 1.04898560e+00 9.94249761e-01
3.94298911e-01 7.75451183e-01 8.26338172e-01 -1.00657791e-01
-3.53635848e-01 -3.42567205e-01 -2.41073683e-01 1.15772855e+00
-4.11965000e-03 1.79539517e-01 -6.15716338e-01 -1.02692172e-01
1.39054072e+00 9.08310413e-02 -4.15741444e-01 -6.10212147e-01
-1.09199166e+00 7.39121139e-01 8.12443018e-01 -9.83082429e-02
-4.40761358e-01 6.02410793e-01 2.73273379e-01 1.40616462e-01
1.04194009e+00 1.02446094e-01 -3.08637381e-01 3.44944745e-01
-8.75180542e-01 4.60610271e-01 4.05075103e-01 1.09632480e+00
6.58834815e-01 1.93963081e-01 -1.98860630e-01 5.20215988e-01
5.68216085e-01 9.46537733e-01 1.45786002e-01 -1.07590914e+00
3.35968822e-01 7.71540642e-01 2.40651034e-02 -8.71197164e-01
-1.86762050e-01 -5.00833727e-02 -6.99313998e-01 3.21851939e-01
2.03715816e-01 2.23473728e-01 -1.33145320e+00 1.31100678e+00
7.04256713e-01 8.07493851e-02 -5.30152321e-01 1.31649280e+00
1.46275353e+00 3.32300484e-01 -3.42060365e-02 5.02088368e-01
1.37608671e+00 -1.18720090e+00 -5.28400779e-01 -1.22438826e-01
2.05966815e-01 -7.03712106e-01 1.15944660e+00 -3.04980800e-02
-1.55178607e+00 -5.36796749e-01 -7.93216288e-01 -7.67435133e-01
-3.78553644e-02 -3.65150608e-02 5.05164266e-01 1.53016061e-01
-9.58977997e-01 6.12323701e-01 -1.13038778e+00 -2.96255440e-01
6.48027360e-01 5.26228786e-01 -5.05331755e-01 -5.54329492e-02
-8.69503498e-01 5.63498616e-01 -2.37249985e-01 -1.05208100e-03
-8.83579612e-01 -9.07835364e-01 -1.31702447e+00 -2.26829261e-01
8.80772322e-02 -1.33009672e+00 1.24885380e+00 -9.22550797e-01
-1.46960866e+00 1.59711301e+00 -4.15217489e-01 -2.50450641e-01
7.49930799e-01 -6.23495102e-01 2.71294117e-01 2.22619623e-01
2.54328817e-01 7.59777188e-01 9.28110778e-01 -1.30292106e+00
-1.26019986e-02 -4.64555115e-01 -1.75914258e-01 3.38009596e-01
2.95070559e-01 -5.16716503e-02 -9.18919504e-01 -9.07967329e-01
1.60259113e-01 -9.50174749e-01 -8.56818259e-02 5.82159638e-01
-5.78407705e-01 -5.69899306e-02 6.49419069e-01 -9.74480867e-01
8.56076300e-01 -1.62230933e+00 7.03477919e-01 2.43738353e-01
4.61174369e-01 2.45637037e-02 -3.90088856e-02 1.66930497e-01
-8.36875662e-03 -1.74647551e-02 -3.75331372e-01 -8.68715167e-01
-1.51446328e-01 7.41834193e-02 -1.20886732e-02 7.30611503e-01
1.64936423e-01 1.45419776e+00 -6.43769383e-01 -5.81319511e-01
5.92411578e-01 9.47559476e-01 -7.77390599e-01 3.58415395e-01
-2.46126384e-01 7.92303979e-01 -3.99205297e-01 1.16367865e+00
5.69132447e-01 -6.34475887e-01 -1.26271278e-01 -5.61054349e-01
8.83698463e-02 -1.03760853e-01 -8.66865993e-01 2.05590343e+00
-3.61342967e-01 2.70999104e-01 4.53630120e-01 -5.72373927e-01
5.44473529e-01 2.20024854e-01 6.88841105e-01 -7.21091628e-01
5.10697722e-01 1.10628195e-01 -5.56945622e-01 -3.70472163e-01
3.30150515e-01 -5.73187470e-02 8.89796391e-02 7.16641769e-02
7.95789734e-02 -2.77204692e-01 -4.21328276e-01 3.10574844e-02
8.93621445e-01 5.75611651e-01 1.07738554e-01 -4.23837095e-01
3.90511721e-01 -1.84277594e-01 4.39245284e-01 1.60250232e-01
1.86964750e-01 1.08489132e+00 -1.74687520e-01 -7.87184536e-01
-1.19485700e+00 -1.23054349e+00 -1.77541256e-01 7.97176480e-01
2.80364096e-01 -4.45759684e-01 -7.50987887e-01 -4.38570023e-01
3.78212392e-01 2.41189487e-02 -1.01039326e+00 1.75976411e-01
-8.23416114e-01 -1.38537169e-01 1.86840922e-01 7.16743350e-01
3.65889370e-01 -1.11379969e+00 -7.35830069e-01 8.47685784e-02
-2.23993853e-01 -9.44604278e-01 -9.10554588e-01 -3.92399549e-01
-8.08583319e-01 -1.22500050e+00 -1.15248573e+00 -7.41078377e-01
8.45148504e-01 4.80118655e-02 1.40775263e+00 4.13151890e-01
-6.59761012e-01 6.36713982e-01 -2.05233902e-01 -2.86317319e-01
-1.83593079e-01 -2.80363895e-02 -4.52420376e-02 5.15816212e-02
-1.79870397e-01 -5.38241565e-01 -9.37623382e-01 3.38026106e-01
-6.43339992e-01 6.12334073e-01 4.02274966e-01 7.75330245e-01
1.16854000e+00 -7.07331896e-01 8.64451081e-02 -8.25894475e-01
-6.90601915e-02 -4.42122370e-01 -5.62034905e-01 -1.25418022e-01
-2.64324605e-01 -9.98333693e-02 1.57816522e-02 -3.38168412e-01
-7.53082454e-01 2.07121670e-01 -4.44711030e-01 -1.05695570e+00
1.88112050e-01 1.03826672e-01 -2.03825489e-01 -3.70612979e-01
3.07821482e-01 1.24592714e-01 3.08949977e-01 -8.06468368e-01
2.75434792e-01 3.60759556e-01 5.65708935e-01 -5.61921537e-01
8.77619028e-01 7.89522469e-01 -1.45908847e-01 -6.66210949e-01
-7.70521224e-01 -2.63599128e-01 -6.40671611e-01 -2.65796751e-01
1.23209643e+00 -1.22138059e+00 -5.75262249e-01 6.04177058e-01
-1.07313776e+00 -8.18935394e-01 -3.72232020e-01 1.65922359e-01
-7.89575398e-01 1.97145447e-01 -9.94009197e-01 -3.87564689e-01
-9.98107731e-01 -1.18911839e+00 1.85640991e+00 -4.24764641e-02
-3.62392783e-01 -8.46155703e-01 -3.11111268e-02 7.55430341e-01
3.48766267e-01 7.54409373e-01 5.12541175e-01 1.59512013e-01
-9.40268874e-01 9.82713234e-03 -1.79316327e-01 -2.92528560e-03
3.65482047e-02 -3.21241140e-01 -9.04496074e-01 -5.06977856e-01
-3.72328430e-01 -4.14244354e-01 7.38024056e-01 6.47401989e-01
1.34264004e+00 -2.76530713e-01 -5.33974469e-01 1.16079819e+00
1.21858299e+00 -2.75521398e-01 6.77320182e-01 -1.17478324e-02
1.43329215e+00 2.45580986e-01 3.79836679e-01 3.67852241e-01
7.21783042e-01 1.05343497e+00 4.52502370e-01 -6.94946527e-01
-7.31144011e-01 -4.17692691e-01 5.50891347e-02 7.84353018e-01
-3.49554062e-01 1.88454501e-02 -1.00264204e+00 6.50860846e-01
-1.68174386e+00 -7.38936007e-01 -1.92817137e-01 2.07832074e+00
7.62075126e-01 -1.85406476e-01 4.98062298e-02 -2.51751035e-01
4.01365012e-01 -3.15291733e-02 -7.56483912e-01 -1.95532948e-01
1.03287265e-01 6.32310271e-01 4.25590068e-01 7.29551017e-01
-1.21513975e+00 1.09843743e+00 5.57919121e+00 6.38768435e-01
-1.04130661e+00 2.71438360e-01 5.08785248e-01 -2.88500696e-01
-3.84433031e-01 -4.84092742e-01 -6.59239531e-01 5.34136117e-01
3.86073232e-01 -4.62146290e-02 3.55389148e-01 6.45726621e-01
-2.28802860e-02 1.33320615e-01 -1.14861572e+00 1.20561600e+00
1.33615792e-01 -1.51630139e+00 1.34304345e-01 3.28092158e-01
7.39296913e-01 4.49224710e-01 -1.36200383e-01 -1.85742844e-02
1.88002646e-01 -1.25693452e+00 9.52385545e-01 7.67472923e-01
1.29572451e+00 -5.95628440e-01 4.34680194e-01 5.57796657e-02
-1.60512388e+00 3.91685724e-01 -1.47291794e-01 2.24574685e-01
3.40269506e-01 5.19040465e-01 -2.96027988e-01 6.93126857e-01
1.07826757e+00 8.65995526e-01 -3.37986767e-01 8.92349005e-01
2.63340641e-02 3.02625209e-01 -4.11369920e-01 6.84008956e-01
-2.28652820e-01 3.23201604e-02 7.90428936e-01 9.02205884e-01
2.35030800e-01 1.91309169e-01 3.49200785e-01 1.10364580e+00
-3.62487078e-01 2.26111054e-01 -6.13959730e-01 3.72659147e-01
1.51483506e-01 9.93757725e-01 -5.55284500e-01 -4.07038212e-01
-4.90452439e-01 1.25924420e+00 4.90288228e-01 5.52005358e-02
-9.40240264e-01 2.77224422e-01 7.48779774e-01 8.36331964e-01
4.64251608e-01 -3.44944745e-02 -3.00267339e-01 -1.29329157e+00
8.91933963e-02 -8.52623940e-01 1.54330820e-01 -9.52140033e-01
-1.34066141e+00 6.98371828e-01 -8.44832137e-02 -1.11834705e+00
1.72642648e-01 -2.39565775e-01 -3.63316298e-01 9.84349310e-01
-1.21079314e+00 -1.69812727e+00 -6.41900420e-01 6.59535408e-01
7.95885503e-01 2.55828381e-01 8.62154365e-01 3.35579574e-01
-3.69667739e-01 7.25179911e-01 -4.65168864e-01 2.80206144e-01
5.01537502e-01 -1.09432542e+00 9.90001559e-01 4.50024694e-01
-9.81033742e-02 3.55053931e-01 3.77628386e-01 -1.15074098e+00
-1.67870128e+00 -1.36518168e+00 4.39436972e-01 -7.27736294e-01
9.24838334e-02 -4.42098796e-01 -8.60841155e-01 9.47446048e-01
1.24487981e-01 4.29944068e-01 2.95529097e-01 -3.57709169e-01
-2.08546475e-01 3.45865816e-01 -1.32234013e+00 4.92271870e-01
1.55283427e+00 -3.37989002e-01 -5.28400958e-01 3.51359934e-01
6.96003556e-01 -1.27823544e+00 -1.38129675e+00 7.59203136e-01
7.20902562e-01 -7.08513439e-01 1.29905689e+00 -3.09554130e-01
6.65455997e-01 -2.28515133e-01 -1.30840600e-01 -1.00117850e+00
-4.53179181e-01 -5.59621036e-01 -6.07621431e-01 8.05037320e-01
-8.99451599e-02 -4.36793804e-01 9.50020254e-01 6.53754592e-01
-2.46261358e-01 -1.31190097e+00 -1.01183367e+00 -4.41412389e-01
2.31190145e-01 -1.08766302e-01 6.08064711e-01 8.52968097e-01
-5.42718172e-01 5.06089106e-02 -4.06041533e-01 6.53683543e-02
9.66302395e-01 3.12035501e-01 9.54134047e-01 -1.11832798e+00
-2.56367713e-01 -1.55919969e-01 -3.94251257e-01 -1.14497745e+00
2.41531327e-01 -1.16138685e+00 -1.41649291e-01 -1.63180864e+00
2.29059309e-01 -2.86982894e-01 2.18859762e-01 6.59778774e-01
-1.39312506e-01 8.89384627e-01 1.83103204e-01 1.40498325e-01
-3.50150824e-01 8.69096518e-01 1.56995142e+00 -1.88396871e-01
-4.47597653e-02 -2.24184006e-01 -4.31046247e-01 8.89164209e-01
5.91587305e-01 -2.37702504e-01 -1.75196737e-01 -7.58408487e-01
-2.15805724e-01 1.82984576e-01 8.24681759e-01 -9.69754159e-01
-1.71210896e-02 1.19635649e-01 6.85177326e-01 -7.42205083e-01
5.70938110e-01 -6.39992476e-01 5.10758519e-01 3.41083050e-01
6.34849593e-02 1.81942657e-01 3.96951586e-01 4.87195790e-01
2.61361897e-01 5.84676206e-01 9.45251226e-01 -3.43052417e-01
-4.44940567e-01 9.09892738e-01 1.39012188e-01 3.44264030e-01
8.58174562e-01 -2.50835329e-01 2.15512335e-01 -2.07160383e-01
-9.24089074e-01 3.45090687e-01 1.06342554e+00 6.07942104e-01
1.01873517e+00 -1.64556825e+00 -8.82377505e-01 4.17336643e-01
5.34048490e-02 4.31654602e-01 4.07947898e-01 7.66443133e-01
-6.74175441e-01 1.59679018e-02 -1.81373924e-01 -8.03557813e-01
-1.63450015e+00 3.43466878e-01 7.61554360e-01 -1.86930582e-01
-1.45504081e+00 1.09774411e+00 5.67348003e-01 -5.73059797e-01
5.18306196e-02 -4.56912577e-01 3.66150469e-01 -5.29987216e-01
3.21281850e-01 1.23562321e-01 9.18530077e-02 -1.10881257e+00
-4.79584426e-01 1.34940195e+00 1.96357965e-01 1.14304580e-01
1.21727931e+00 -1.22602597e-01 -2.25103227e-03 1.50666177e-01
1.18474782e+00 3.87541801e-02 -1.36065221e+00 -3.62110645e-01
-6.72033727e-01 -6.30394936e-01 1.07181244e-01 -5.53809941e-01
-1.46079910e+00 5.54555535e-01 6.91282332e-01 -5.41132033e-01
8.28980565e-01 2.78973907e-01 1.29456770e+00 -3.13947320e-01
6.45961225e-01 -5.56714296e-01 -7.47548193e-02 1.99658930e-01
1.51149106e+00 -1.22996855e+00 1.78886235e-01 -7.61652768e-01
-2.54458636e-01 6.15968347e-01 7.88123548e-01 -4.74958479e-01
8.42099965e-01 4.27383840e-01 -3.08020059e-02 -6.96526408e-01
-4.40059960e-01 -3.19741011e-01 8.74648392e-01 3.88069510e-01
4.67960179e-01 1.44049540e-01 1.60436220e-02 3.97885889e-01
-3.30437243e-01 1.04934730e-01 -2.60363426e-03 6.82685494e-01
3.35755721e-02 -1.00531805e+00 -4.06625926e-01 5.88496923e-01
-4.71357048e-01 -1.10530302e-01 -2.48752087e-01 7.02876270e-01
2.04468459e-01 4.55812454e-01 1.97818935e-01 -3.02411288e-01
5.28337598e-01 -1.83712766e-01 7.80931532e-01 -7.45956898e-01
-6.48309827e-01 2.36557528e-01 -1.81215163e-02 -1.16654825e+00
-2.79427230e-01 -4.63674307e-01 -1.34331775e+00 -3.75713050e-01
8.17602277e-02 -4.41759825e-01 2.69182771e-01 6.75964653e-01
7.34330177e-01 6.08797014e-01 5.56729324e-02 -1.65880847e+00
8.90801474e-02 -7.68864691e-01 -1.63371280e-01 7.52347767e-01
2.88865983e-01 -9.93707716e-01 -3.67950764e-03 1.88869372e-01] | [7.082292079925537, -1.2029035091400146] |
51dfb313-3528-4508-b124-50c60d9b9915 | ecnuica-at-semeval-2021-task-11-rule-based | null | null | https://aclanthology.org/2021.semeval-1.185 | https://aclanthology.org/2021.semeval-1.185.pdf | ECNUICA at SemEval-2021 Task 11: Rule based Information Extraction Pipeline | This paper presents our endeavor for solving task11, NLPContributionGraph, of SemEval-2021. The purpose of the task was to extract triples from a paper in the Nature Language Processing field for constructing an Open Research Knowledge Graph. The task includes three sub-tasks: detecting the contribution sentences in papers, identifying scientific terms and predicate phrases from the contribution sentences; and inferring triples in the form of (subject, predicate, object) as statements for Knowledge Graph building. In this paper, we apply an ensemble of various fine-tuned pre-trained language models (PLM) for tasks one and two. In addition, self-training methods are adopted for tackling the shortage of annotated data. For the third task, rather than using classic neural open information extraction (OIE) architectures, we generate potential triples via manually designed rules and develop a binary classifier to differentiate positive ones from others. The quantitative results show that we obtain the 4th, 2nd, and 2nd rank in three evaluation phases. | ['Liang He', 'Qin Chen', 'Jiawei Liu', 'Zhiwei Wang', 'Jing Ling', 'Jiaju Lin'] | 2021-08-01 | null | null | null | semeval-2021 | ['open-information-extraction'] | ['natural-language-processing'] | [ 2.46824279e-01 6.91272795e-01 -4.80551869e-01 -2.57254690e-02
-7.24333405e-01 -6.98749721e-01 7.95784533e-01 5.93553662e-01
-2.08696067e-01 1.09636629e+00 1.24574527e-01 -6.25064671e-01
-2.37582445e-01 -8.96449625e-01 -9.77647245e-01 -8.04553181e-02
8.80491883e-02 4.37525034e-01 6.33699670e-02 2.01454997e-01
6.44621789e-01 4.60526973e-01 -1.24515784e+00 5.15674055e-01
1.02903235e+00 9.36398566e-01 1.36987448e-01 6.51623011e-01
-7.56126583e-01 1.09715903e+00 -5.25662124e-01 -7.58724093e-01
6.04451969e-02 -1.73610151e-01 -1.41172588e+00 -3.58393252e-01
2.81687468e-01 3.78516555e-01 -4.82834764e-02 1.32273781e+00
1.30678713e-01 -9.30877924e-02 9.17959273e-01 -1.29745948e+00
-9.55169976e-01 1.22131228e+00 -5.09336412e-01 3.20628613e-01
6.15013540e-01 -3.17634046e-01 1.35551763e+00 -9.42402363e-01
9.44092870e-01 1.10859203e+00 3.53296816e-01 3.51136833e-01
-6.39704883e-01 -5.80631971e-01 -1.73973337e-01 4.30538416e-01
-1.26424110e+00 -3.22207868e-01 6.54781222e-01 -6.50906086e-01
9.76443946e-01 4.04532194e-01 3.12186271e-01 1.06948507e+00
2.47463778e-01 6.58858895e-01 9.75512445e-01 -6.78503871e-01
1.90864712e-01 2.60753393e-01 7.37540305e-01 7.91497409e-01
7.20983028e-01 -3.97483498e-01 -6.90697074e-01 -1.71918333e-01
4.86543357e-01 -5.10222614e-01 -1.43500894e-01 1.81685597e-01
-1.32307136e+00 4.48051929e-01 4.27079707e-01 6.87529087e-01
-5.73833585e-01 -2.05580100e-01 1.79332197e-01 7.40610734e-02
3.74635398e-01 9.14608240e-01 -6.03493214e-01 9.01982635e-02
-8.46637845e-01 1.65167779e-01 1.15852892e+00 1.25172400e+00
7.45137751e-01 -5.61022043e-01 -8.92127395e-01 3.74634713e-01
3.06758255e-01 2.28065010e-02 2.64251113e-01 -6.92574263e-01
7.51821399e-01 1.04122698e+00 7.67821148e-02 -1.07078993e+00
-2.96411514e-01 -3.03606838e-01 -6.78778827e-01 -5.43102920e-01
2.43767008e-01 -3.09833944e-01 -9.41402078e-01 1.14891291e+00
1.66679233e-01 1.17653877e-01 3.69070947e-01 5.33949435e-01
1.68974400e+00 8.18652570e-01 4.39739048e-01 -3.23019952e-01
1.71566451e+00 -8.29400301e-01 -8.99462640e-01 -6.69325665e-02
5.87524533e-01 -5.88842154e-01 5.40112257e-01 3.48840803e-01
-9.39637125e-01 -3.05034727e-01 -8.70445013e-01 -5.24280190e-01
-9.05357182e-01 3.49039733e-01 6.96741104e-01 3.07240766e-02
-9.93468940e-01 6.03639901e-01 -2.09173098e-01 -2.33058095e-01
5.60049534e-01 8.45660344e-02 -2.01169714e-01 2.21169218e-01
-1.35199046e+00 9.47876930e-01 7.98920691e-01 6.34218231e-02
-5.81427336e-01 -9.33857441e-01 -7.16165185e-01 2.59260863e-01
4.50268090e-01 -7.14390099e-01 8.39153528e-01 -3.74180704e-01
-1.10703790e+00 1.41629362e+00 -1.13369077e-01 -5.15450895e-01
2.43985042e-01 -6.05292358e-02 -3.51300180e-01 1.30512744e-01
4.32551444e-01 5.09508789e-01 3.30458641e-01 -1.23011494e+00
-6.34227276e-01 -4.65291739e-01 2.07508877e-01 -1.46552056e-01
-4.09495085e-01 4.23286557e-01 -5.03550231e-01 -3.36509734e-01
-4.90105189e-02 -4.35602903e-01 2.54468508e-02 -5.35378039e-01
-1.02946723e+00 -9.50879812e-01 5.29819965e-01 -9.29800391e-01
1.34833431e+00 -1.52252603e+00 2.66057760e-01 2.08881930e-01
6.56866908e-01 2.50526249e-01 3.82699668e-02 2.58492291e-01
-1.09059036e-01 5.87251067e-01 -2.89728455e-02 1.36847660e-01
6.47348166e-02 -1.01803832e-01 -4.77829486e-01 -3.61557513e-01
3.45630616e-01 1.18815196e+00 -1.02739429e+00 -8.26585829e-01
-4.62563217e-01 -9.27535444e-02 1.13070875e-01 1.49478182e-01
-6.62875950e-01 1.78545982e-01 -8.13288927e-01 6.29366517e-01
4.37784284e-01 -5.39048433e-01 2.45928913e-01 -3.68460476e-01
-2.85111994e-01 4.80295599e-01 -9.32836771e-01 1.21661413e+00
-2.46610090e-01 6.02810502e-01 3.24498070e-03 -1.06443250e+00
1.01725650e+00 3.59572232e-01 2.93295860e-01 -4.46560949e-01
1.59495756e-01 1.93790421e-01 -2.64029860e-01 -9.47464526e-01
4.57107991e-01 2.71012217e-01 -1.12189934e-01 2.08143175e-01
3.68964374e-01 3.12023945e-02 7.83982038e-01 5.12904644e-01
1.27218747e+00 7.55608454e-02 4.05336529e-01 -2.91005790e-01
8.83077204e-01 -3.24598923e-02 4.62540388e-01 8.10952902e-01
5.11998460e-02 1.17837548e-01 1.07194829e+00 -4.04814690e-01
-7.75701046e-01 -7.16551960e-01 -6.38788491e-02 6.39582217e-01
-2.01187700e-01 -6.03603363e-01 -6.13911629e-01 -9.06617165e-01
1.02577601e-02 9.24301922e-01 -3.73156995e-01 1.45045653e-01
-3.93242210e-01 -5.86013079e-01 7.14195490e-01 2.23313496e-01
4.35084790e-01 -1.27008855e+00 -2.55423248e-01 -1.88651860e-01
-5.18808305e-01 -1.46753430e+00 -5.83571121e-02 2.56104022e-01
-3.89043242e-01 -1.32712758e+00 -3.88806134e-01 -1.16414630e+00
7.77763963e-01 -2.08726749e-01 1.30342972e+00 -8.56999010e-02
-1.57836080e-01 1.73477903e-02 -2.37449318e-01 -7.69411504e-01
-3.13923448e-01 2.22436935e-02 -2.70337582e-01 -2.10817456e-01
7.54157126e-01 -9.86627713e-02 1.02268666e-01 -4.39733922e-01
-5.72121620e-01 1.50556400e-01 7.64688671e-01 3.20963055e-01
5.10888636e-01 1.97407737e-01 5.98402083e-01 -1.00682700e+00
1.01538539e+00 -4.36191171e-01 -6.44834518e-01 8.32972944e-01
-5.72312832e-01 3.00578803e-01 6.30187690e-01 3.97751760e-03
-8.52479458e-01 -2.43745744e-02 2.26178989e-01 -9.10812989e-02
-1.89126208e-01 8.34138334e-01 -3.63878667e-01 -4.91014495e-02
5.41197300e-01 6.59231097e-02 -5.08260608e-01 -3.06265205e-01
4.13586348e-01 8.95128906e-01 4.87473935e-01 -7.81349719e-01
7.55374134e-01 -1.28349766e-01 1.18829153e-01 -7.14911938e-01
-1.46749401e+00 -4.37503070e-01 -5.91567099e-01 -2.58924723e-01
9.63232517e-01 -8.77599537e-01 -9.82144296e-01 1.81210414e-02
-1.72275937e+00 2.50350863e-01 -2.54289120e-01 2.82264888e-01
-1.14803590e-01 3.91474783e-01 -5.55976152e-01 -6.00572109e-01
-7.77405798e-01 -7.54034877e-01 8.96852970e-01 5.63082814e-01
-1.74399093e-01 -7.92308450e-01 -1.21502168e-01 9.07545626e-01
-1.89401358e-01 1.52373195e-01 1.17832816e+00 -9.45948303e-01
-6.52082384e-01 -1.54155150e-01 -5.78091860e-01 6.08877987e-02
-1.06156543e-01 1.52627558e-01 -8.51970315e-01 4.08164352e-01
-2.68547595e-01 -4.13295239e-01 8.87184858e-01 1.17177837e-01
1.39733768e+00 -5.81704021e-01 -5.19890189e-01 6.53282180e-02
1.12706101e+00 2.40386613e-02 5.71685791e-01 2.48452231e-01
8.98466289e-01 7.19110429e-01 1.65917143e-01 6.89068362e-02
6.40330911e-01 1.07009351e-01 6.04910180e-02 3.38787109e-01
-1.56635314e-01 -3.83391619e-01 3.07169169e-01 6.80979729e-01
-2.23613799e-01 -3.40726316e-01 -1.16653621e+00 6.30594790e-01
-1.75010109e+00 -9.07805800e-01 -6.48135185e-01 1.87189281e+00
1.29398179e+00 4.96592104e-01 -2.96366841e-01 -5.47476821e-02
6.36054754e-01 -2.26536468e-02 -1.14392675e-01 -2.77687401e-01
-1.27844736e-01 1.60029396e-01 4.25057232e-01 4.26496893e-01
-1.13948607e+00 1.08099484e+00 5.88207197e+00 7.88909912e-01
-6.93822920e-01 -8.43166932e-02 5.17969787e-01 4.40836519e-01
-6.59579635e-01 2.61751622e-01 -1.13962698e+00 2.13142827e-01
1.03542745e+00 -4.45182115e-01 3.11027735e-01 3.49280655e-01
1.46436557e-01 -1.02996022e-01 -1.24765873e+00 7.61326015e-01
2.58988112e-01 -1.74452126e+00 5.79577744e-01 -1.79167420e-01
6.74988210e-01 -2.63667464e-01 -6.30746365e-01 4.98622447e-01
4.76787060e-01 -9.04110610e-01 5.81647873e-01 8.66943061e-01
3.71317565e-01 -2.22004279e-01 6.68918431e-01 5.17075181e-01
-9.46227252e-01 4.22779582e-02 -3.07277113e-01 -6.50118515e-02
-2.89674312e-01 1.08470583e+00 -9.71994698e-01 1.05083001e+00
4.67082262e-01 9.02866006e-01 -7.13235140e-01 8.08445513e-01
-8.42662394e-01 5.00981629e-01 5.81391379e-02 -5.29730678e-01
1.32080186e-02 -1.06627978e-01 6.18799984e-01 1.25167143e+00
5.66330552e-02 1.62845507e-01 1.93738028e-01 1.36455274e+00
-8.11012387e-01 9.18367580e-02 -7.70020783e-01 -8.13564539e-01
4.17960346e-01 1.63143992e+00 -7.12962329e-01 -5.24485290e-01
-1.72312319e-01 4.63937134e-01 7.05268502e-01 2.89463311e-01
-4.52320695e-01 -6.78450584e-01 -2.03794792e-01 -3.25019471e-02
1.27371415e-01 -2.00017224e-04 -7.62249649e-01 -1.27812088e+00
1.76769450e-01 -5.30463398e-01 5.27173519e-01 -9.29130316e-01
-1.36497653e+00 3.67470056e-01 -1.23686135e-01 -4.83478427e-01
9.25706029e-02 -6.94256008e-01 -7.51101792e-01 9.23496246e-01
-1.51324141e+00 -1.17788875e+00 -1.08351327e-01 7.16216266e-02
1.95239067e-01 -1.26163229e-01 7.40892053e-01 3.00965756e-01
-8.27125788e-01 2.51518279e-01 -3.08061093e-01 5.13803244e-01
4.00543064e-01 -1.33432245e+00 1.30761996e-01 1.08112752e+00
1.78421244e-01 7.20246017e-01 6.11936152e-01 -1.06715369e+00
-1.50307190e+00 -1.03615701e+00 2.07707834e+00 -6.05624318e-01
1.01525283e+00 -3.98483455e-01 -7.51733601e-01 6.41256988e-01
2.90088952e-01 -1.62136421e-01 5.23935795e-01 2.72304207e-01
-2.24146530e-01 1.01491898e-01 -9.22938347e-01 3.85803938e-01
1.00693476e+00 -4.73809570e-01 -9.84254837e-01 7.19544232e-01
8.93849909e-01 -3.91643256e-01 -1.08935761e+00 2.49243781e-01
6.96100965e-02 -7.02946708e-02 9.40576375e-01 -1.17458951e+00
1.14415801e+00 -2.16241688e-01 2.20881015e-01 -9.22543764e-01
-3.93403232e-01 -6.31658018e-01 -4.91067171e-01 1.54332328e+00
7.92279840e-01 -4.78075556e-02 6.34229004e-01 6.50649130e-01
-1.24013342e-01 -6.86580479e-01 -6.80393338e-01 -5.75316966e-01
2.06735525e-02 -1.03264146e-01 2.60518134e-01 1.04622591e+00
4.01478708e-01 1.01891804e+00 7.49485344e-02 2.35532075e-01
7.83210874e-01 4.28129047e-01 3.54882866e-01 -1.59737420e+00
2.16549844e-01 -3.77831191e-01 -1.36060581e-01 -4.85691130e-01
5.92348337e-01 -1.44934285e+00 -8.96612108e-02 -2.15969348e+00
4.90129650e-01 -1.12290613e-01 -3.71404648e-01 8.48861992e-01
-3.01310003e-01 -3.60905975e-01 -3.50424908e-02 3.34879495e-02
-8.75917256e-01 2.41137013e-01 1.23793519e+00 -4.72144902e-01
-3.13060693e-02 -2.10296035e-01 -1.17657423e+00 5.91409683e-01
4.91338789e-01 -4.07355338e-01 -8.86498019e-02 -2.92714655e-01
5.03836989e-01 7.18293637e-02 5.49552560e-01 -6.71688557e-01
5.07422447e-01 -2.98133582e-01 2.98955858e-01 -6.84222043e-01
-3.25230181e-01 -3.76183748e-01 -1.48024425e-01 2.75650263e-01
-7.16027200e-01 -2.64844418e-01 1.35901377e-01 2.85735726e-01
-1.23379759e-01 -5.29951990e-01 1.99088424e-01 -3.70934963e-01
-7.10023820e-01 8.55451301e-02 -1.09965004e-01 4.27399129e-01
9.51174259e-01 3.22891086e-01 -7.39884675e-01 3.10226619e-01
-6.04829073e-01 5.96399128e-01 -2.59075224e-01 6.56177878e-01
6.50762796e-01 -1.04976559e+00 -8.92392635e-01 -2.26080358e-01
1.72407374e-01 6.13370091e-02 -9.98912230e-02 5.81059933e-01
-4.31265175e-01 8.47698212e-01 -6.36661649e-02 1.20384850e-01
-9.72802579e-01 6.95424616e-01 -1.25008464e-01 -8.66882384e-01
-2.90614724e-01 8.87749553e-01 -2.95654297e-01 -5.01127183e-01
4.56782371e-01 -3.38835955e-01 -9.16791916e-01 1.82396203e-01
4.60147619e-01 2.45856643e-01 2.56639242e-01 -3.31333369e-01
-4.83963668e-01 2.51049977e-02 -1.20193556e-01 4.00083840e-01
1.40412855e+00 3.71880770e-01 -8.04882348e-01 4.10700798e-01
1.00411046e+00 -4.32409607e-02 -1.42769873e-01 -2.32546434e-01
5.54274797e-01 1.77382663e-01 1.31218374e-01 -9.67660427e-01
-6.32676601e-01 5.60773194e-01 -1.36986673e-01 4.56690460e-01
6.38906300e-01 1.73347011e-01 5.71961343e-01 5.12129664e-01
1.60774529e-01 -1.05449975e+00 -2.06376478e-01 8.32732141e-01
1.02665985e+00 -1.18092656e+00 1.61948115e-01 -8.12444746e-01
-3.24433029e-01 1.29765093e+00 6.85304344e-01 3.14132541e-01
5.55487394e-01 1.69957191e-01 -4.46893781e-01 -7.20775247e-01
-7.83228099e-01 -2.54610218e-02 6.48975670e-01 2.54833102e-01
6.05595410e-01 1.69749647e-01 -8.38382721e-01 8.36465955e-01
-1.80443361e-01 3.42436343e-01 5.74144125e-01 8.36203694e-01
-4.56859976e-01 -7.95058966e-01 -1.99927166e-01 7.63958633e-01
-5.62777102e-01 -2.84145653e-01 -1.09533179e+00 1.15110628e-01
2.00739950e-01 1.00052249e+00 -2.60252744e-01 -3.85652214e-01
2.35149652e-01 1.87538043e-01 3.16877961e-01 -7.28370070e-01
-6.87059760e-01 -5.45170486e-01 5.47647953e-01 -7.57201239e-02
-4.40750569e-01 -3.60911161e-01 -1.45275319e+00 -2.85853650e-02
-2.05459490e-01 5.31972647e-01 5.37097037e-01 1.18361306e+00
5.01028776e-01 8.69200885e-01 1.44815698e-01 -2.22353816e-01
-5.05250752e-01 -9.24233913e-01 -2.30114162e-01 2.11487651e-01
-7.02427179e-02 -4.47574258e-01 1.93919912e-02 3.04080456e-01] | [9.436646461486816, 8.525623321533203] |
f8d10903-ae50-4802-b55c-e30924b9ced1 | the-search-for-stability-learning-dynamics-of | 2305.16695 | null | https://arxiv.org/abs/2305.16695v1 | https://arxiv.org/pdf/2305.16695v1.pdf | The Search for Stability: Learning Dynamics of Strategic Publishers with Initial Documents | We study a game-theoretic model of information retrieval, in which strategic publishers aim to maximize their chances of being ranked first by the search engine, while maintaining the integrity of their original documents. We show that the commonly used PRP ranking scheme results in an unstable environment where games often fail to reach pure Nash equilibrium. We propose the Relative Ranking Principle (RRP) as an alternative ranking principle, and introduce two ranking functions that are instances of the RRP. We provide both theoretical and empirical evidence that these methods lead to a stable search ecosystem, by providing positive results on the learning dynamics convergence. We also define the publishers' and users' welfare, and demonstrate a possible publisher-user trade-off, which highlights the complexity of determining which ranking function should be selected by the search engine designer. | ['Moshe Tennenholtz', 'Omer Madmon'] | 2023-05-26 | null | null | null | null | ['information-retrieval'] | ['natural-language-processing'] | [-2.40934283e-01 3.08876604e-01 -3.92610729e-01 5.57561755e-01
-5.60149074e-01 -1.07414877e+00 7.53079295e-01 -3.20458263e-02
-7.14841723e-01 4.85975564e-01 9.49877799e-02 -4.91493315e-01
-9.80780721e-01 -7.48558521e-01 -3.10269028e-01 -5.11825681e-01
-4.92039829e-01 6.21221721e-01 4.86571223e-01 -4.29629713e-01
6.75468326e-01 2.41544291e-01 -1.54674494e+00 -3.79150331e-01
9.55168724e-01 7.42609859e-01 5.38038135e-01 8.09404850e-01
-5.79174757e-02 1.14563406e+00 -8.09264302e-01 -7.06360996e-01
8.49465191e-01 -3.80682915e-01 -7.37726629e-01 -4.52188671e-01
-5.78964591e-01 -3.56185734e-01 -4.81403291e-01 1.28490007e+00
6.42303765e-01 9.87388715e-02 5.39617419e-01 -1.21524882e+00
-9.00003433e-01 7.20269263e-01 -3.67589653e-01 5.40765107e-01
3.28157037e-01 -1.70619145e-01 1.65979075e+00 -2.06540272e-01
1.06305087e+00 7.16522217e-01 3.87900732e-02 5.31769395e-01
-8.75926673e-01 -2.81898588e-01 -1.13994814e-01 -5.17520308e-03
-1.28871381e+00 -3.16577941e-01 5.97624183e-01 -2.21930191e-01
5.35957694e-01 7.44028449e-01 8.07695270e-01 3.29024434e-01
4.96315658e-01 6.81858957e-01 1.08078504e+00 -5.00020385e-01
3.28043759e-01 1.32747740e-02 4.31220196e-02 1.78623110e-01
7.16491938e-01 3.69930267e-01 -7.47665524e-01 -5.78090310e-01
9.19503093e-01 -3.87547165e-01 -3.88847351e-01 -8.28083873e-01
-5.80373108e-01 7.67703891e-01 1.15507081e-01 3.35256279e-01
-8.38672280e-01 1.41228408e-01 -1.24957085e-01 9.89526093e-01
3.03898036e-01 9.80131149e-01 -2.67674208e-01 -3.32801878e-01
-5.39232969e-01 3.54843795e-01 1.20632994e+00 8.14432502e-01
2.34188154e-01 -6.08857989e-01 -2.29161024e-01 6.45805836e-01
5.90064466e-01 2.03692362e-01 4.30430472e-01 -1.54116797e+00
-1.05127536e-01 3.43650997e-01 7.06642270e-01 -8.26848805e-01
2.50982344e-02 -8.48728538e-01 -2.78953850e-01 1.50431901e-01
2.56661773e-01 -8.82580969e-03 -1.42149016e-01 1.79658127e+00
-2.22470716e-01 -4.56140220e-01 1.44588485e-01 8.50010514e-01
2.12484866e-01 7.22429872e-01 -1.87641636e-01 -8.33478153e-01
1.08543360e+00 -5.56534767e-01 -5.89841187e-01 -8.59718472e-02
3.30390126e-01 -4.40144122e-01 9.98912454e-01 4.75712240e-01
-1.64439750e+00 5.65670490e-01 -7.72424400e-01 3.71171623e-01
-1.05520651e-01 -8.21036696e-01 3.95431072e-01 5.53430736e-01
-1.69830632e+00 5.31472921e-01 -2.25481600e-01 -3.43906879e-01
4.29804437e-02 3.89657915e-01 3.34003747e-01 3.26875269e-01
-1.39524150e+00 1.03455138e+00 -2.80827373e-01 -2.58701652e-01
-5.84879756e-01 -4.21712160e-01 1.59520358e-01 5.36064744e-01
7.23776460e-01 -8.14113975e-01 1.76166368e+00 -7.96073318e-01
-1.53470874e+00 7.28228450e-01 4.70203698e-01 -2.11893007e-01
7.36022294e-01 -1.04760448e-03 9.02262181e-02 8.52681845e-02
2.93406188e-01 -6.59870729e-02 7.45440871e-02 -1.44068694e+00
-8.51433933e-01 -2.44980857e-01 3.39508235e-01 8.42075944e-01
-3.88012171e-01 3.94094050e-01 -4.78074253e-01 -5.54837361e-02
-2.05805019e-01 -6.59542859e-01 -4.64088053e-01 -3.07345748e-01
-1.39199495e-01 -4.98889625e-01 -1.27253488e-01 -2.45498583e-01
1.53425479e+00 -1.61503375e+00 1.41431078e-01 6.42876267e-01
6.40428185e-01 -1.84310004e-01 -3.97581249e-01 6.75934613e-01
7.12218642e-01 7.29516864e-01 6.03753746e-01 1.78172469e-01
3.27765852e-01 -9.40911486e-05 -1.25446707e-01 2.41189748e-01
-6.01038098e-01 9.65456486e-01 -1.06773663e+00 -2.93545872e-01
-5.70781708e-01 -1.65169999e-01 -4.91242230e-01 1.44147366e-01
-1.20960318e-01 -2.80100614e-01 -7.50335991e-01 3.89665127e-01
1.75992638e-01 -3.84721458e-01 6.99413955e-01 7.09129989e-01
-3.53507310e-01 5.13319671e-01 -8.88680756e-01 8.06324184e-01
-1.75480470e-01 5.73366046e-01 5.98456681e-01 -6.06587172e-01
3.66180778e-01 1.16177440e-01 6.11098886e-01 -1.20529723e+00
1.28127024e-01 4.06879514e-01 2.75727183e-01 -8.28579888e-02
7.48054624e-01 2.96871550e-02 -5.18562198e-02 9.93357658e-01
-2.48549029e-01 1.53964996e-01 3.83351177e-01 7.21054316e-01
1.37110400e+00 -5.43510556e-01 3.08222055e-01 -9.23126042e-01
6.65058643e-02 4.46620444e-03 4.63485003e-01 1.59527385e+00
-3.00530493e-01 -7.33510703e-02 1.08260036e+00 1.07230023e-01
-1.03183532e+00 -9.70461071e-01 3.01020384e-01 1.34630704e+00
7.62922764e-01 -5.91236949e-01 -5.15969634e-01 -1.78747326e-01
1.84734408e-02 5.69186866e-01 -3.62509012e-01 -6.60596117e-02
4.01489809e-02 -5.13496518e-01 1.13878056e-01 -1.81939751e-01
2.57930726e-01 -9.79704857e-01 -9.37541783e-01 5.79097159e-02
-2.36248896e-01 -3.10761064e-01 -7.89222121e-01 2.91808303e-02
-6.03042722e-01 -1.15204775e+00 -6.24293983e-01 -3.87919664e-01
4.30507243e-01 6.19233787e-01 1.28540111e+00 6.06252193e-01
2.14487404e-01 9.07079101e-01 -3.36995006e-01 -2.44022220e-01
-4.86220747e-01 1.93925902e-01 1.68802530e-01 -6.46660388e-01
4.89985347e-02 -4.94323075e-01 -1.04709589e+00 4.20462132e-01
-8.08562458e-01 -2.72094399e-01 6.63927436e-01 5.13659954e-01
3.01491797e-01 3.58904779e-01 6.96074724e-01 -6.33316755e-01
1.78482175e+00 -6.47673488e-01 -1.00401723e+00 4.74826813e-01
-1.38296890e+00 7.89235011e-02 3.87934119e-01 -2.95538604e-01
-7.41770923e-01 -6.67457283e-01 6.48868680e-01 2.76542246e-01
9.37850177e-01 8.56564999e-01 -5.31551130e-02 -2.28120282e-01
6.37958765e-01 4.77883488e-01 4.45919186e-01 -3.71981144e-01
4.68260467e-01 7.22383380e-01 1.34999782e-01 -4.13430989e-01
6.61478162e-01 7.03722984e-02 -2.27321863e-01 -4.93705750e-01
-1.32472381e-01 -4.67484057e-01 1.52819887e-01 -4.38050270e-01
1.51562005e-01 -6.32041097e-01 -1.12136912e+00 1.28713652e-01
-8.69479775e-01 -2.78304994e-01 -4.04513896e-01 -3.13021354e-02
-7.34468341e-01 3.84348541e-01 -6.45710588e-01 -1.27411592e+00
-2.71316648e-01 -7.19340682e-01 1.94363758e-01 5.19310594e-01
1.65535644e-01 -5.33883929e-01 4.45929676e-01 1.06261589e-01
8.91682625e-01 -5.75183988e-01 7.38829792e-01 -1.00880623e+00
-1.05984855e+00 -1.56313390e-01 -8.10589455e-03 -4.08255160e-01
-1.93123907e-01 1.11659646e-01 -1.03016302e-01 -3.33345920e-01
1.07984103e-01 3.25964123e-01 3.94627482e-01 6.01439178e-01
2.92391360e-01 -1.01926804e+00 -3.50572914e-01 -1.38309211e-01
1.49715292e+00 5.41400433e-01 6.76766217e-01 7.33012795e-01
-2.17277437e-01 7.78149962e-01 3.84744644e-01 5.88025212e-01
3.42894495e-01 7.55282998e-01 4.69287485e-01 6.49858832e-01
4.76465642e-01 -2.26783201e-01 3.44429731e-01 8.21448505e-01
-2.52584994e-01 -5.86844087e-01 -6.92637742e-01 5.46413600e-01
-2.12925577e+00 -1.11474872e+00 3.28411907e-01 2.64024234e+00
9.02579665e-01 1.98496327e-01 5.96415579e-01 -4.21899229e-01
6.07365966e-01 -8.25073123e-02 -3.58756155e-01 -1.42602891e-01
-1.81037098e-01 -9.08277035e-02 8.99553657e-01 8.11094463e-01
-2.41186947e-01 8.41243625e-01 7.84232569e+00 6.17643893e-01
-5.57196259e-01 2.96361059e-01 3.75911862e-01 -5.68915963e-01
-8.59604478e-01 1.41598478e-01 -3.78497690e-02 4.74863827e-01
1.00553584e+00 -1.25354028e+00 1.07012701e+00 7.08743453e-01
2.39397198e-01 -1.94685102e-01 -5.24824440e-01 6.19486213e-01
-4.38935906e-01 -1.39541471e+00 -1.34349987e-01 8.13304424e-01
5.47529697e-01 7.78821707e-02 -1.56068802e-02 2.84524467e-02
1.23277402e+00 -6.88350201e-01 7.87049770e-01 5.97289860e-01
2.97886997e-01 -8.87495399e-01 2.76253611e-01 7.16272473e-01
-7.44235396e-01 -5.45619965e-01 -3.93358529e-01 -1.77594081e-01
4.85463068e-02 2.44522661e-01 -2.56761730e-01 3.47960770e-01
5.16689539e-01 -5.24913520e-03 -5.39106488e-01 1.40533614e+00
2.23141946e-02 1.32612810e-01 -2.97984421e-01 -4.73261416e-01
-1.00097872e-01 -5.00797629e-01 8.52658272e-01 3.42127830e-01
2.36273289e-01 4.04855549e-01 -2.87177801e-01 9.23969448e-01
-4.12108958e-01 2.20840499e-01 -5.93347669e-01 -2.58853525e-01
9.79195237e-01 1.07782614e+00 -8.46528411e-01 4.50616330e-02
-8.05153623e-02 7.72276759e-01 1.25772282e-01 3.61240685e-01
-1.40172794e-01 -3.76583248e-01 7.90351689e-01 2.86140919e-01
-3.49380858e-02 8.64133015e-02 -2.48754218e-01 -1.03305399e+00
1.21229365e-01 -6.08917534e-01 2.89042771e-01 -7.27425814e-01
-1.17261732e+00 2.05845475e-01 -9.63988453e-02 -7.25085735e-01
-3.42252135e-01 -2.66320556e-01 -6.32993996e-01 6.30372107e-01
-1.39555359e+00 -1.89314172e-01 3.07150304e-01 1.05637215e-01
5.05098328e-02 -1.59281015e-01 2.17211366e-01 7.44947940e-02
-1.95591182e-01 4.09540743e-01 8.25111985e-01 -4.58771050e-01
1.70148179e-01 -1.38233173e+00 8.80679861e-02 7.77296424e-01
6.91839457e-02 8.38174641e-01 8.74374390e-01 -6.77363813e-01
-1.80518937e+00 -1.50938839e-01 1.01225317e+00 -4.91110802e-01
8.35260987e-01 -2.28069037e-01 -4.03944343e-01 1.60155028e-01
3.17989409e-01 -8.50650609e-01 4.93289292e-01 1.00217797e-01
-9.27377492e-03 -1.12863168e-01 -1.05859709e+00 1.14380479e+00
1.30795705e+00 -4.94156361e-01 -2.55782068e-01 3.85003984e-01
5.96229136e-01 2.72693455e-01 -5.03121972e-01 -5.03584370e-02
8.28191400e-01 -1.00088322e+00 8.92021537e-01 -3.72805357e-01
2.88324833e-01 -5.53720370e-02 2.36470606e-02 -1.19800973e+00
-8.20245266e-01 -1.13407576e+00 -1.24766923e-01 7.91235089e-01
4.47901070e-01 -8.71912777e-01 7.62668550e-01 1.04058993e+00
4.83771384e-01 -5.61984360e-01 -1.12888896e+00 -1.00873339e+00
4.15961355e-01 2.51475096e-01 1.55384094e-01 5.49147129e-01
4.81502235e-01 2.00090125e-01 -2.03609556e-01 -2.60352463e-01
7.45077789e-01 -1.82785362e-01 3.59794408e-01 -1.76022792e+00
-4.00368601e-01 -1.25826490e+00 1.64960921e-02 -1.05788445e+00
-2.25703090e-01 -5.80036044e-01 1.62147969e-01 -1.78448868e+00
5.31662107e-01 -4.06040370e-01 -7.12292969e-01 -5.91757521e-02
1.32802591e-01 -1.54796198e-01 4.22822475e-01 9.64198291e-01
-1.19807184e+00 1.79131106e-01 9.70763683e-01 2.10181504e-01
-4.20312107e-01 2.13510469e-01 -1.73587132e+00 -1.32256551e-02
5.30547619e-01 -3.81366193e-01 -6.54983938e-01 9.41410661e-04
1.13450789e+00 4.15650427e-01 1.03007913e-01 -2.54879706e-02
8.76222610e-01 -5.90181828e-01 -5.15004337e-01 -8.29231590e-02
-4.04152572e-01 -6.49473369e-01 3.86524379e-01 7.70750821e-01
-6.53938890e-01 2.10909411e-01 -5.05182147e-01 7.33463049e-01
2.99579322e-01 -4.07030106e-01 4.17553991e-01 -2.56507874e-01
-2.64710724e-01 7.00871786e-03 -8.94455135e-01 -4.63347994e-02
8.53956342e-01 -1.71984494e-01 -5.21324992e-01 -1.14521182e+00
-2.57837951e-01 6.63135409e-01 7.63083279e-01 1.88756198e-01
2.41003230e-01 -1.04170549e+00 -6.92596495e-01 -5.44794798e-01
-7.69113153e-02 -6.70088410e-01 -1.58219010e-01 4.33594465e-01
-4.64408487e-01 3.96190345e-01 -1.34302497e-01 1.80349484e-01
-1.00430965e+00 4.67332274e-01 4.89302635e-01 -7.87219226e-01
-1.01419166e-01 5.61749578e-01 9.49651822e-02 -1.28240511e-01
1.47571027e-01 4.40745115e-01 -1.76274195e-01 7.45790154e-02
3.48081291e-01 3.31946164e-01 -2.29234263e-01 -5.87737374e-02
-4.19948280e-01 9.26625356e-03 -1.41054794e-01 -7.72994161e-01
1.23786616e+00 -5.63415110e-01 -4.46873188e-01 -1.24959990e-01
4.25713837e-01 2.92270631e-01 -8.29183221e-01 -3.00483763e-01
4.57529604e-01 -6.23741984e-01 1.29167438e-01 -1.00334227e+00
-8.99510264e-01 -6.01963326e-02 2.51180232e-01 1.15839505e+00
1.05915332e+00 1.25673845e-01 4.62573841e-02 4.61624622e-01
6.35953724e-01 -1.26975298e+00 1.73192732e-02 3.44106853e-01
6.54848993e-01 -3.98361653e-01 -9.46219712e-02 -9.04770121e-02
-5.64701378e-01 4.82654393e-01 3.58675778e-01 -1.71664767e-02
7.82947183e-01 -7.75576085e-02 -1.68989882e-01 -3.93612653e-01
-1.23733366e+00 -3.79841268e-01 -1.56582490e-01 5.89709759e-01
1.68132961e-01 7.81639442e-02 -1.25332522e+00 1.07727289e+00
-3.53051364e-01 -5.48945963e-02 1.03258061e+00 9.98625159e-01
-7.26756454e-01 -1.43344402e+00 1.33899912e-01 8.06219220e-01
-7.58407414e-01 -2.72550493e-01 -1.04771495e+00 5.03237188e-01
-7.62933791e-01 1.12795901e+00 8.90679434e-02 -3.63387346e-01
3.07170689e-01 -9.41417292e-02 2.45901868e-01 -3.27717423e-01
-7.23812461e-01 2.65313357e-01 -1.26638904e-01 -5.09602427e-01
-1.43779024e-01 -5.08796871e-01 -6.44090772e-01 -5.71601570e-01
-6.32119775e-01 8.69157910e-01 3.42836797e-01 5.83775342e-01
8.67682278e-01 -6.39789691e-03 8.54854286e-01 1.53717250e-02
-1.03588486e+00 -5.75635433e-01 -1.12796664e+00 3.29707153e-02
7.27570280e-02 -5.14680803e-01 -5.79898357e-01 -6.92050219e-01] | [4.2827277183532715, 2.9392666816711426] |
48f27819-d305-4a54-9f2b-eecdd6c2fa55 | evaluating-raw-waveforms-with-deep-learning | 2307.02820 | null | https://arxiv.org/abs/2307.02820v1 | https://arxiv.org/pdf/2307.02820v1.pdf | Evaluating raw waveforms with deep learning frameworks for speech emotion recognition | Speech emotion recognition is a challenging task in speech processing field. For this reason, feature extraction process has a crucial importance to demonstrate and process the speech signals. In this work, we represent a model, which feeds raw audio files directly into the deep neural networks without any feature extraction stage for the recognition of emotions utilizing six different data sets, EMO-DB, RAVDESS, TESS, CREMA, SAVEE, and TESS+RAVDESS. To demonstrate the contribution of proposed model, the performance of traditional feature extraction techniques namely, mel-scale spectogram, mel-frequency cepstral coefficients, are blended with machine learning algorithms, ensemble learning methods, deep and hybrid deep learning techniques. Support vector machine, decision tree, naive Bayes, random forests models are evaluated as machine learning algorithms while majority voting and stacking methods are assessed as ensemble learning techniques. Moreover, convolutional neural networks, long short-term memory networks, and hybrid CNN- LSTM model are evaluated as deep learning techniques and compared with machine learning and ensemble learning methods. To demonstrate the effectiveness of proposed model, the comparison with state-of-the-art studies are carried out. Based on the experiment results, CNN model excels existent approaches with 95.86% of accuracy for TESS+RAVDESS data set using raw audio files, thence determining the new state-of-the-art. The proposed model performs 90.34% of accuracy for EMO-DB with CNN model, 90.42% of accuracy for RAVDESS with CNN model, 99.48% of accuracy for TESS with LSTM model, 69.72% of accuracy for CREMA with CNN model, 85.76% of accuracy for SAVEE with CNN model in speaker-independent audio categorization problems. | ['Ayhan Kucukmanisa', 'Ulku Bayraktar', 'Zeynep Hilal Kilimci'] | 2023-07-06 | null | null | null | null | ['emotion-recognition', 'ensemble-learning', 'ensemble-learning', 'speech-emotion-recognition'] | ['computer-vision', 'computer-vision', 'methodology', 'speech'] | [-0.29788402 -0.3591448 0.44033724 -0.33984035 -0.29009116 -0.21239218
0.23048936 0.06058896 -0.58769214 0.8049107 0.08400458 -0.260679
-0.31282043 -0.5846425 -0.05666303 -0.49858293 -0.22427866 -0.10670739
-0.29213947 -0.52415 0.24884292 0.4865524 -1.9082059 0.5939202
0.5195187 1.6921331 -0.41555455 0.8694524 -0.17932694 0.9448071
-0.9664331 -0.22787803 -0.27580783 0.22982119 -0.62282485 -0.3799951
-0.25400805 -0.03180925 -0.35880533 0.5568543 1.0735375 0.57488954
0.9070131 -1.4981741 -0.46113098 0.31631765 -0.3192519 0.40941685
0.20722294 -0.51437783 0.46508 -1.1105419 0.03659471 0.9699583
0.9875245 0.37285262 -0.45465368 -1.1089834 -0.16245112 0.41830766
-1.3434033 -0.6821562 0.6935101 -0.38671926 1.5282149 0.2963075
0.5289301 1.0606362 0.3316682 0.5368297 1.4223628 -0.43207544
0.22969534 0.43859047 0.6103777 0.79165655 -0.5635382 0.28668082
-0.89739907 -0.37962887 0.11502415 -0.17629248 -0.16398022 0.66132945
-0.5689503 0.6485315 0.0773562 0.8465299 -0.6307215 -0.35725394
0.9079175 0.645006 0.59812886 -0.14328034 -0.864997 -0.3610091
-0.91275537 -0.12996976 0.86729205 0.46106416 0.20477733 0.75574565
0.04170724 1.3071531 0.2280602 0.20903747 1.1451113 -0.3707077
0.15378611 0.4996586 -0.21399279 -1.0786635 -0.4454253 -0.58103216
-1.2963892 0.32683516 -0.3328956 -0.54587066 -0.8803667 1.4609773
0.2098367 0.13812415 0.7424976 0.41171145 1.4159472 1.0064907
0.20726022 -0.3823251 1.3746874 -0.8541024 -0.8386857 0.22523749
0.2407866 -1.0055244 0.5520884 0.94043124 -0.69503963 -0.8458316
-1.1308793 0.47320384 -0.8976141 0.7091799 0.33482757 0.9281056
-1.0275275 0.5237537 -0.27882788 -0.22236876 0.31290382 0.6842588
-0.60617286 0.7080292 -1.5038999 0.8174869 0.56112194 0.21625865
-0.5831997 -0.41145307 -0.55618626 0.2078993 -0.29741395 -0.36913243
1.3011801 -1.1833118 -1.8947539 0.46822035 -0.07093339 -0.5572503
0.03067569 -0.01351157 -1.1165841 -0.11393822 -0.6268949 0.3130164
0.87816644 -0.77169305 -0.78569216 -0.4814109 -0.36901614 0.08299366
-0.767269 0.28911215 0.5534718 -0.5495434 0.04834692 -0.6296752
0.23363978 -0.6861622 -0.31804082 -0.61598027 1.2751714 -1.0833266
1.4003791 -2.2593493 -0.3644921 0.29953015 0.08007663 0.7317975
0.21980125 0.4772343 -0.35059777 0.09097634 0.08353533 -0.16558251
0.02890956 -0.06597064 -0.17670272 -0.05229357 0.10282601 0.38056925
-0.19042163 -0.23086298 0.5030714 0.8655455 0.0619997 0.18830857
0.62677556 0.11267018 -0.20934948 0.575548 0.5934386 0.48249507
-0.17460497 -0.4819255 -0.2893573 0.08797505 -1.1978604 1.0549845
-0.96339387 0.72669536 0.19393253 -1.245957 1.5018256 1.018355
0.11239497 -0.46328574 0.7228538 0.54268515 0.04347652 -0.8148415
0.41043797 -0.2509804 0.14959936 0.14023021 0.5552315 0.29235002
-0.44118467 -0.18280901 0.8098148 -0.3926253 0.40282652 -0.0136644
0.8996106 -0.4988096 0.2864534 0.2555822 -0.36265087 0.14238504
-0.02875755 -0.46884206 -0.6344111 -0.6353055 -0.13731635 1.2550136
-0.7151256 -0.10533454 -0.81782657 -0.5193522 -0.34113997 0.63992214
-0.4643719 -0.27648565 -0.04343791 -0.8259606 1.1247693 0.48747095
1.027917 -1.3377264 -0.6041671 0.30570856 -0.13965487 -0.90438646
0.22556753 0.5474218 -0.6559648 -0.71876156 -0.49703836 -0.9121347
-0.15248412 -0.23510808 0.970137 -0.04563451 -0.14641981 0.3169218
-0.5695631 -0.7126795 -0.19451386 -0.04039465 0.26661134 0.27888897
0.52129424 -0.9459603 -0.26424432 0.03162089 -0.41473073 -0.41401082
0.4228628 1.0049295 0.17518593 0.4784126 1.0200527 -0.21376263
0.90236706 -0.59011793 0.13330309 0.13594285 -0.32283172 -0.48643246
0.6823964 -0.479124 -1.1498414 -0.16627257 -0.6990324 -0.48472413
-0.55997914 0.6356894 0.07050848 -0.03749843 0.5739322 0.44829404
-0.2807013 -0.4667187 -0.24120076 1.7333865 0.24917828 -0.27541015
-0.12279697 -0.03390785 -0.6500766 -1.2137212 -0.30363312 -0.36142936
-0.39919314 -0.57797414 0.9014566 -0.9012286 -0.86056983 1.0232362
-1.28569 0.14515918 0.17715515 0.64844614 -0.12262337 -0.16720003
-0.5869464 -1.6343879 -1.0939628 -0.9005677 0.70756555 0.26088622
-0.37131426 -0.8026069 -0.25958344 0.3461241 0.6136554 0.24979356
0.80861044 -1.3494421 0.58564585 -0.28230643 0.09520949 0.9233337
-0.01433508 0.3557324 -1.5703372 0.02087649 0.2268814 -0.4567645
0.73187715 0.2318375 1.3674772 -0.2394963 0.15592326 0.2128022
1.0266037 1.209223 0.5520002 0.14947668 0.12739532 0.51161677
0.22852394 0.48344392 0.31713393 0.17495123 0.11031573 0.16581847
0.07187402 0.14219064 0.44600344 1.6887414 -0.03704235 -0.44560918
-0.9164112 0.29866678 -1.5243104 -0.9764089 -0.1819489 1.8786968
0.1990875 -0.15353821 0.18064994 1.2515212 0.82097554 0.01226756
-0.30108413 -1.0113962 -0.11252732 0.92805886 -0.1524855 0.16137533
-1.3656322 0.6209177 4.7957954 1.1264465 -1.5900284 0.39964274
1.011812 0.03078981 0.4119509 -0.73648524 -0.4107579 0.54079926
1.4039303 0.10025211 0.32646182 0.794938 0.11707009 0.06578789
-0.46846884 1.51731 0.06553413 -0.8552647 -0.27246574 -0.19755712
0.38667125 0.03332908 0.22130585 0.7947104 0.13737702 -1.1825192
0.62397236 0.44018558 0.5609076 -1.2332904 1.0960289 0.2377307
-1.2520516 -0.3616957 -0.13504602 -0.23483953 -0.14788347 0.52631956
-0.6643371 0.80880934 1.1710565 0.38493258 -0.07648805 0.6412015
0.3077743 0.90632755 -0.4547097 -0.3393381 0.37132522 0.16394572
0.20348538 1.4189656 0.81438047 0.44232407 -0.19612686 0.13819447
0.02579786 0.6658737 -0.55531293 -0.1814267 0.5794698 1.4478348
-0.38483864 -0.5318145 0.0958427 0.584035 0.11555997 0.2498325
-0.86286163 -0.9100627 0.44118047 -0.5839102 0.18461755 0.08347569
-0.28128955 -0.67757255 -0.13922304 -1.0549722 0.3167622 -1.1111486
-1.2708024 1.390611 -0.3490867 -1.1580182 -0.07935876 -0.753694
-1.0335383 0.80894375 -1.0314426 -0.9706902 -0.43901464 0.86315405
0.67796254 -1.0449568 1.0551193 0.59310293 -0.7526269 0.67795634
0.2757227 0.1570248 0.32111546 -0.7937411 -0.20700331 0.17671838
0.07288437 0.41419592 0.35036644 -0.07821993 -0.8027127 -0.9723458
0.858102 0.32236615 0.44674352 -0.06823669 -0.81951886 0.16497642
0.60982215 -0.09660272 1.0456663 0.14095238 -0.31404775 -0.321452
-1.4934492 -0.03017903 0.3845882 -0.76542866 -0.5594795 -0.07902963
0.26926824 -0.2088218 -1.1715597 0.5734184 0.869474 -1.3068864
0.663791 -0.72123903 0.35508752 0.03163754 -0.661203 -1.5699927
0.0695052 -0.10995479 -0.07843822 1.6219343 0.50791866 -0.88154876
0.4655595 0.19343707 -0.36829978 -1.0441515 -1.269574 -0.45621678
0.06475515 -0.76804125 0.7104235 1.1798022 -0.14031914 0.36074182
-0.27351147 0.03183291 -0.07263701 -0.53443307 0.50764453 -1.3590673
0.03583216 -0.36328754 -0.7322393 -0.12949744 0.29640046 -0.6079129
-0.3503045 -1.3570026 -0.28994185 -0.3082821 -0.8553256 0.5098567
0.36366156 0.21479739 0.09380283 -0.3446417 -0.18232992 0.68476063
0.4674674 -0.16953538 -0.246092 0.12140424 -0.31064498 0.93794066
1.2636883 -0.2593756 -0.46140411 -0.02491559 -0.1535708 0.22255968
0.3455562 -1.3908361 0.21419656 0.41465312 0.65894073 -0.70545703
0.75761646 -0.6566607 0.27101377 0.31705448 -0.15073384 0.24886742
0.44113 0.17446612 -0.78170633 -0.39806807 0.65612876 0.05365997
-0.6897938 0.14702183 -0.879356 -0.27842233 0.86271995 -0.38723668
-0.12363373 -0.42447966 -1.2752047 -0.32925352 -0.6951151 0.6487206
0.82071954 -1.3433709 -0.46190757 0.26185197 -0.19577311 -0.601026
0.7791109 0.8193196 -0.1678277 0.20738207 -0.3166554 -0.20447162
-1.8475376 0.00996336 0.58030003 0.01298859 0.24482389 1.0189803
-0.30395168 -0.88617164 0.35256195 -0.15032534 -0.7631812 0.32743832
0.35592595 0.740874 0.60963875 -0.87528545 -0.50958157 0.36664236
0.04444076 -0.19968957 1.3734741 0.15080215 -0.14609326 0.7339407
1.4094722 -0.18174754 -0.20231482 0.13042086 -0.28370267 0.14650087
0.5129 -1.4218328 -1.1963316 1.4331095 1.3514347 0.23366044
1.4553162 -0.820865 0.8256598 0.48064965 0.26422954 -1.3848869
-0.0930362 0.86624783 0.90548104 -1.1345164 -0.5775365 0.07576396
-0.7470953 1.4953661 0.7204211 -0.01997275 1.2821686 0.4219494
0.1857382 0.02483571 -1.0551667 0.0758668 0.2346132 0.5891265
0.67365956 0.10570237 -0.10034936 1.3397084 -0.69493985 0.22415085
0.09069074 0.7611766 -0.5425078 -0.5405561 -0.4675681 0.7152096
-0.7120356 -0.19332851 -0.49557903 0.680031 0.5779383 1.4615471
0.18406893 -1.2332752 0.4845606 0.7802124 -0.01051845 0.07597563
-1.2175016 -0.18568505 0.42070547 0.16768518 -0.27645048 -0.38706803
-0.98626435 -0.37545776 -0.49091214 0.6139755 1.06371 1.0988157
0.5143909 0.78655565 0.7623784 -0.66671276 -0.39805683 -1.3963044
-0.43186548 -0.07808029 0.23514867 -0.8266576 -0.39868256 -0.20537713] | [13.72793197631836, 5.523054122924805] |
a9d77c58-ffec-4a06-a03c-99846b351445 | semi-infinitely-constrained-markov-decision | 2305.00254 | null | https://arxiv.org/abs/2305.00254v1 | https://arxiv.org/pdf/2305.00254v1.pdf | Semi-Infinitely Constrained Markov Decision Processes and Efficient Reinforcement Learning | We propose a novel generalization of constrained Markov decision processes (CMDPs) that we call the \emph{semi-infinitely constrained Markov decision process} (SICMDP). Particularly, we consider a continuum of constraints instead of a finite number of constraints as in the case of ordinary CMDPs. We also devise two reinforcement learning algorithms for SICMDPs that we call SI-CRL and SI-CPO. SI-CRL is a model-based reinforcement learning algorithm. Given an estimate of the transition model, we first transform the reinforcement learning problem into a linear semi-infinitely programming (LSIP) problem and then use the dual exchange method in the LSIP literature to solve it. SI-CPO is a policy optimization algorithm. Borrowing the ideas from the cooperative stochastic approximation approach, we make alternative updates to the policy parameters to maximize the reward or minimize the cost. To the best of our knowledge, we are the first to apply tools from semi-infinitely programming (SIP) to solve constrained reinforcement learning problems. We present theoretical analysis for SI-CRL and SI-CPO, identifying their iteration complexity and sample complexity. We also conduct extensive numerical examples to illustrate the SICMDP model and demonstrate that our proposed algorithms are able to solve complex sequential decision-making tasks leveraging modern deep reinforcement learning techniques. | ['Zhihua Zhang', 'Wenhao Yang', 'Yang Peng', 'Liangyu Zhang'] | 2023-04-29 | null | null | null | null | ['model-based-reinforcement-learning'] | ['reasoning'] | [ 1.56770304e-01 3.72707248e-01 -5.45672297e-01 5.75782135e-02
-9.48331594e-01 -4.47719842e-01 2.32313424e-01 -8.92164186e-02
-4.87344474e-01 1.08330190e+00 -1.58882961e-01 -7.24042177e-01
-6.09102607e-01 -5.98501444e-01 -6.44673347e-01 -9.93757248e-01
-2.82800078e-01 6.67757630e-01 -5.00061549e-02 -2.89268848e-02
1.18335441e-01 1.55602098e-01 -7.85165608e-01 -2.76299834e-01
9.91218269e-01 1.03683877e+00 3.89374226e-01 7.39338994e-01
1.19068086e-01 9.70320880e-01 -1.43114999e-01 -1.62224874e-01
4.11855698e-01 -3.22024494e-01 -7.60850847e-01 4.11580890e-01
-5.80492139e-01 -5.45169532e-01 -2.77958959e-01 9.52385545e-01
4.12018478e-01 4.94549215e-01 6.75443649e-01 -1.42943871e+00
-3.15890223e-01 7.81729698e-01 -9.01473224e-01 1.83467865e-01
2.39564881e-01 2.41985008e-01 9.60961938e-01 -5.28108060e-01
3.68176669e-01 1.42938769e+00 1.68654442e-01 6.49160683e-01
-1.32656229e+00 -4.87983644e-01 6.99522138e-01 8.17013159e-02
-8.39816809e-01 -2.08264962e-02 5.01291037e-01 -5.02117515e-01
8.04139376e-01 -2.81783968e-01 6.39129996e-01 9.90564883e-01
1.55727103e-01 1.17939317e+00 1.49800789e+00 -6.33658409e-01
8.60393465e-01 -2.18574852e-01 1.35723487e-01 6.85740709e-01
-3.63258719e-02 5.05703330e-01 -2.45800838e-01 -2.89297909e-01
8.06861579e-01 5.59598617e-02 -2.68161092e-02 -6.07445776e-01
-1.14567757e+00 1.16635132e+00 -1.81507379e-01 -2.87630081e-01
-6.17609680e-01 4.39719290e-01 1.09689780e-01 6.22816920e-01
2.81916767e-01 -9.19323638e-02 -4.09218520e-01 -1.05113015e-01
-6.86948478e-01 5.03311455e-01 1.08781826e+00 9.97948527e-01
6.77044094e-01 2.44167611e-01 -4.02309448e-01 5.88143826e-01
4.22644883e-01 5.93562484e-01 -4.97413278e-02 -1.57297969e+00
8.66757452e-01 -3.05137575e-01 8.64104688e-01 -3.25509816e-01
-2.99311817e-01 -3.67214799e-01 -6.51436865e-01 2.87171543e-01
4.55037624e-01 -8.39288354e-01 -5.69761157e-01 1.90546489e+00
2.53192604e-01 1.93103671e-01 3.37753564e-01 6.86158955e-01
-5.58141410e-01 9.20988500e-01 7.46556148e-02 -8.15349579e-01
7.33487308e-01 -8.80094409e-01 -6.65127456e-01 -1.11343525e-01
4.61173981e-01 -2.21401498e-01 8.11068177e-01 5.15491426e-01
-1.18548191e+00 -7.63780922e-02 -9.04026031e-01 8.74293506e-01
2.37345964e-01 -7.75812864e-02 4.94376302e-01 5.77619255e-01
-8.91923249e-01 7.80614018e-01 -9.81249452e-01 5.31778820e-02
3.41427803e-01 3.80658805e-01 4.08913881e-01 -1.32520959e-01
-1.18860161e+00 6.71924651e-01 4.62135732e-01 -2.20251549e-02
-1.49087846e+00 -3.91031682e-01 -4.76304919e-01 1.59676671e-01
1.15893495e+00 -5.48746884e-01 1.83140373e+00 -9.75387454e-01
-2.07225776e+00 2.61769712e-01 1.09380437e-03 -5.72069645e-01
6.96604848e-01 -8.07699263e-02 1.75110996e-02 3.58390540e-01
1.63153663e-01 1.85401723e-01 1.04345345e+00 -1.21094680e+00
-9.20420289e-01 -3.50818411e-02 5.05047023e-01 4.48980927e-01
3.48093174e-02 2.99636479e-02 1.00769121e-02 -5.58752716e-01
-1.79520473e-01 -1.18231106e+00 -7.70330131e-01 -3.04789573e-01
-3.72486204e-01 -3.50023419e-01 5.09778500e-01 -4.60273623e-01
1.04986572e+00 -1.80697823e+00 4.97115076e-01 2.33496174e-01
-1.03671424e-01 4.99350950e-04 -5.61039522e-02 7.22251356e-01
2.15293020e-01 4.70950566e-02 -4.87075865e-01 -4.17300463e-01
3.73500675e-01 7.12485790e-01 -4.12362784e-01 3.85185629e-01
-9.64456946e-02 6.67765558e-01 -1.16807747e+00 -3.82627130e-01
9.61334407e-02 -3.71131569e-01 -7.79443800e-01 1.17589779e-01
-6.68970048e-01 4.63821828e-01 -7.97963798e-01 6.51680946e-01
4.90118414e-01 -5.62357791e-02 4.38420802e-01 6.11410737e-01
-9.57901850e-02 -2.00290129e-01 -1.38437843e+00 1.21617889e+00
-7.17471778e-01 -2.52630338e-02 2.67864317e-01 -1.54949844e+00
4.02522117e-01 4.48622346e-01 7.92254388e-01 -3.29525352e-01
-6.32674247e-02 1.12059794e-01 -2.23604634e-01 -2.52349585e-01
2.13862330e-01 -4.77618188e-01 -1.96439922e-01 6.91400349e-01
3.24585214e-02 2.64466293e-02 3.92139852e-01 2.20687717e-01
8.65445018e-01 3.66422594e-01 3.96036118e-01 -3.03119540e-01
4.60416079e-01 -1.91426545e-01 9.52211738e-01 1.19505906e+00
-4.95591670e-01 -2.92241693e-01 1.05580401e+00 9.31736901e-02
-7.18442440e-01 -1.32513297e+00 2.37302348e-01 1.23576009e+00
-8.32422748e-02 9.22366083e-02 -4.99568969e-01 -6.99675977e-01
1.46112829e-01 7.72066593e-01 -3.47836196e-01 1.33558720e-01
-3.24586540e-01 -7.51527548e-01 3.72454524e-02 3.48936200e-01
5.20554721e-01 -9.08640563e-01 -4.04477596e-01 6.51597083e-01
3.38317379e-02 -1.11827803e+00 -5.90980887e-01 4.68903661e-01
-7.16285348e-01 -8.69363070e-01 -9.54468310e-01 -6.52225971e-01
4.56549168e-01 3.79623758e-04 7.96490788e-01 -5.35079181e-01
4.87820655e-01 7.60281265e-01 -3.09509903e-01 -4.01283145e-01
-4.83490229e-01 -1.01703174e-01 3.22258651e-01 1.77850649e-01
-9.17954668e-02 -6.02385759e-01 -5.81037045e-01 3.50774467e-01
-6.08648002e-01 4.39296775e-02 5.82204342e-01 9.31051075e-01
8.70865226e-01 1.79990783e-01 8.09401929e-01 -7.79653788e-01
8.70098650e-01 -6.76630199e-01 -9.76541698e-01 1.84697628e-01
-7.70416141e-01 3.14638585e-01 7.83900797e-01 -4.43068087e-01
-1.25079691e+00 1.35843068e-01 1.89873189e-01 -6.43854320e-01
1.47660315e-01 6.24128103e-01 1.78577751e-02 1.55410290e-01
-2.00174004e-02 2.25899443e-01 -1.62794292e-01 -2.19689116e-01
2.28845850e-01 6.08043015e-01 1.78487167e-01 -1.15119493e+00
5.83460391e-01 3.26191872e-01 1.96432009e-01 -4.53010827e-01
-1.19331980e+00 -2.37210080e-01 -3.49605292e-01 -4.11459565e-01
8.23908389e-01 -8.17339838e-01 -1.17235851e+00 3.80442142e-01
-7.25135207e-01 -1.01696396e+00 -4.56805825e-01 5.53030372e-01
-1.43607914e+00 4.93481785e-01 -7.46291220e-01 -1.34285891e+00
8.80157109e-03 -1.01519680e+00 5.09798825e-01 1.68813899e-01
5.43570518e-01 -9.94554996e-01 3.40756625e-01 1.83115244e-01
-8.08730349e-02 2.75901288e-01 7.52364755e-01 -2.33800352e-01
-4.41875428e-01 2.40609229e-01 1.05725251e-01 4.17718112e-01
-1.97825387e-01 -2.28601009e-01 -3.26436192e-01 -5.53339422e-01
6.15594094e-04 -5.87152839e-01 7.97837436e-01 6.72368407e-01
1.17243588e+00 -5.30394673e-01 -2.09600136e-01 3.50520581e-01
1.66309226e+00 6.92938566e-01 1.59694612e-01 3.21145028e-01
2.03147799e-01 4.18184847e-01 9.51394558e-01 9.73661304e-01
4.46079284e-01 4.44425672e-01 6.06022596e-01 3.97912443e-01
7.88369656e-01 -4.40927297e-01 7.30781496e-01 6.35072112e-01
-2.22852170e-01 -2.28026286e-01 -7.65950263e-01 3.19801360e-01
-2.39513063e+00 -1.02004385e+00 2.75108278e-01 2.19959354e+00
7.67258465e-01 1.99406356e-01 3.00908953e-01 -5.65331541e-02
7.66488433e-01 -3.33567597e-02 -9.52392697e-01 -9.36221659e-01
2.39848331e-01 3.23645025e-01 7.67928958e-01 6.83601618e-01
-1.04307771e+00 9.33544576e-01 6.03425312e+00 8.70045424e-01
-3.65788430e-01 1.58922404e-01 6.06815457e-01 -4.53800373e-02
4.80542146e-02 1.74824163e-01 -7.66391337e-01 6.40394449e-01
1.08492410e+00 -2.48964041e-01 9.26312149e-01 9.42739427e-01
7.85249949e-01 -3.85374904e-01 -1.07116115e+00 6.75856411e-01
-7.02514648e-01 -1.03278148e+00 -2.08255470e-01 3.21365952e-01
1.08184254e+00 -1.62422299e-01 3.65058221e-02 6.85733557e-01
1.22757649e+00 -6.85162604e-01 6.95711851e-01 5.36594450e-01
5.25083005e-01 -1.25486493e+00 3.77924025e-01 6.79455876e-01
-1.05684590e+00 -7.96118498e-01 -4.01793063e-01 -2.58396387e-01
3.87809783e-01 3.86335462e-01 -2.11493582e-01 5.79680383e-01
2.51472235e-01 6.77574515e-01 4.01523203e-01 8.16239893e-01
-3.52846891e-01 8.11020017e-01 -3.06981027e-01 -5.36408164e-02
7.70830989e-01 -6.08117521e-01 5.81789255e-01 8.44699085e-01
1.89589694e-01 5.02691716e-02 7.61973202e-01 7.28820622e-01
1.33682281e-01 -1.76356450e-01 -1.12566881e-01 -1.56338111e-01
3.74867022e-01 9.03386772e-01 -6.46560073e-01 -2.17992321e-01
-4.26269084e-01 8.81627262e-01 3.65964770e-01 6.53505921e-01
-8.08539927e-01 -1.31587341e-01 7.81702578e-01 -4.79959339e-01
5.71587145e-01 -4.70617831e-01 1.75929740e-01 -1.18326712e+00
-1.51333332e-01 -7.85713375e-01 4.87151474e-01 -3.99912745e-01
-1.25599778e+00 -1.38559759e-01 1.87281877e-01 -1.12382102e+00
-5.46334088e-01 -3.50739151e-01 -5.76742887e-01 6.47120833e-01
-1.75019491e+00 -6.88975215e-01 5.64169824e-01 8.98840964e-01
7.75769055e-01 -9.92494822e-03 4.41887856e-01 -2.18577176e-01
-8.22510362e-01 1.31689087e-01 7.89924860e-01 -1.40556186e-01
1.20223418e-01 -1.53841913e+00 -1.08799405e-01 7.94465303e-01
-4.20521408e-01 6.89395219e-02 5.89130759e-01 -3.81758511e-01
-1.34916949e+00 -9.55196559e-01 1.56984285e-01 2.07172260e-01
1.02701235e+00 -2.12504745e-01 -3.03310812e-01 8.04818630e-01
7.01067224e-02 -1.93636507e-01 3.69199067e-01 -2.68831223e-01
3.31895649e-01 4.47812267e-02 -9.89608884e-01 5.84144592e-01
9.58754957e-01 -2.08076894e-01 -4.00953531e-01 5.22037625e-01
8.34730506e-01 -2.09807098e-01 -8.05170894e-01 3.07982974e-02
2.42708668e-01 -6.45165920e-01 7.20221937e-01 -6.43655360e-01
3.90259236e-01 -1.58542730e-02 -4.34491128e-01 -1.60757875e+00
-2.99407244e-01 -1.16734612e+00 -5.67166448e-01 8.43948364e-01
2.30243266e-01 -7.51948237e-01 7.10358441e-01 3.37068737e-01
-4.65073809e-02 -1.12513602e+00 -1.01455772e+00 -1.09458184e+00
5.20678818e-01 -5.07655859e-01 1.33234456e-01 6.23283923e-01
1.39747903e-01 -1.44150063e-01 -5.89512587e-01 2.51488626e-01
1.24501348e+00 4.15978849e-01 2.37738773e-01 -6.50125384e-01
-1.15277600e+00 -2.28023931e-01 5.60025573e-01 -1.32894528e+00
5.20275056e-01 -6.78007424e-01 5.24786152e-02 -1.39676440e+00
1.83722377e-01 -6.34092331e-01 -4.97409850e-01 3.62243384e-01
9.14882198e-02 -5.16945481e-01 6.36601865e-01 2.48653859e-01
-7.23327398e-01 8.86686206e-01 1.38893998e+00 6.85060695e-02
-5.24097443e-01 7.64620364e-01 -5.27513087e-01 7.05973089e-01
1.12264454e+00 -6.17702127e-01 -5.46561301e-01 -7.28214681e-02
2.47578546e-01 9.40725029e-01 7.56749064e-02 -7.94919372e-01
4.27476428e-02 -1.02000940e+00 -1.04278944e-01 -6.03179932e-01
5.74005470e-02 -5.43491662e-01 -3.63039851e-01 8.17593277e-01
-5.71583748e-01 -1.67653084e-01 -3.03790987e-01 1.17595601e+00
1.99818000e-01 -4.80033696e-01 8.35489511e-01 -3.13567489e-01
-6.10714138e-01 5.54545999e-01 -1.21854317e+00 2.25557461e-01
1.36342931e+00 2.05668986e-01 2.58949846e-01 -6.56981289e-01
-1.32452548e+00 8.83342385e-01 -1.11137569e-01 -2.05504730e-01
4.86509860e-01 -9.25291359e-01 -3.46695930e-01 -3.59435320e-01
-4.98143792e-01 -1.63443670e-01 7.57936388e-02 8.41544747e-01
-1.54778585e-01 5.41255951e-01 -6.88302666e-02 -3.62297446e-01
-5.96026719e-01 7.94029653e-01 5.06304622e-01 -9.47545290e-01
-4.62492436e-01 2.41753578e-01 -4.57230769e-02 -2.94372201e-01
2.55391240e-01 -2.66830802e-01 8.11634734e-02 3.09017394e-02
8.73640403e-02 6.32430613e-01 -5.18769979e-01 1.82720982e-02
4.17999476e-02 1.71586543e-01 1.51076335e-02 -6.59630835e-01
1.36383915e+00 -5.59588492e-01 3.35506111e-01 2.96450347e-01
6.17672801e-01 -5.17582655e-01 -1.95779741e+00 -4.60486054e-01
1.36767522e-01 -3.07596456e-02 1.54450297e-01 -5.65201640e-01
-9.41277921e-01 6.26362443e-01 4.31149632e-01 1.23287283e-01
1.01690340e+00 -2.51140386e-01 6.39472544e-01 7.32342899e-01
8.09843719e-01 -1.61845851e+00 2.93289155e-01 6.91200972e-01
5.09415388e-01 -9.11399782e-01 -2.05329046e-01 6.63136393e-02
-9.65584219e-01 1.02975106e+00 4.13738161e-01 -4.02486891e-01
8.62197101e-01 2.96884209e-01 -5.56730151e-01 4.40346777e-01
-9.96370018e-01 -6.58936858e-01 -7.03093112e-01 6.39795840e-01
-3.79416585e-01 3.15467596e-01 -4.83070433e-01 5.34975529e-01
2.78496832e-01 3.36681634e-01 7.79790461e-01 1.21096861e+00
-5.60637712e-01 -1.13711596e+00 -2.35670581e-01 3.89092028e-01
-4.75529432e-01 1.85673133e-01 9.21689942e-02 5.25595486e-01
-3.52337271e-01 1.14245737e+00 -8.79173875e-02 -1.82217360e-02
-2.11720131e-02 5.50093539e-02 5.50431013e-01 -4.50749159e-01
-4.93284017e-02 1.89225689e-01 -8.25256556e-02 -4.84575748e-01
-6.45796001e-01 -7.12132812e-01 -1.23829997e+00 -3.35655600e-01
-7.45004117e-02 9.22817811e-02 4.26637501e-01 1.17222297e+00
-3.20572592e-02 3.96075279e-01 1.07937443e+00 -8.41401041e-01
-1.33577871e+00 -6.18286788e-01 -8.82032394e-01 -2.13392079e-01
2.88225621e-01 -8.94842982e-01 -3.82654220e-01 -3.44357342e-01] | [4.241252422332764, 2.4687418937683105] |
f9f73329-2e01-4e70-8c53-a00f19e1c28b | multi-modal-representation-learning-for | 2306.07935 | null | https://arxiv.org/abs/2306.07935v1 | https://arxiv.org/pdf/2306.07935v1.pdf | Multi-modal Representation Learning for Social Post Location Inference | Inferring geographic locations via social posts is essential for many practical location-based applications such as product marketing, point-of-interest recommendation, and infector tracking for COVID-19. Unlike image-based location retrieval or social-post text embedding-based location inference, the combined effect of multi-modal information (i.e., post images, text, and hashtags) for social post positioning receives less attention. In this work, we collect real datasets of social posts with images, texts, and hashtags from Instagram and propose a novel Multi-modal Representation Learning Framework (MRLF) capable of fusing different modalities of social posts for location inference. MRLF integrates a multi-head attention mechanism to enhance location-salient information extraction while significantly improving location inference compared with single domain-based methods. To overcome the noisy user-generated textual content, we introduce a novel attention-based character-aware module that considers the relative dependencies between characters of social post texts and hashtags for flexible multi-model information fusion. The experimental results show that MRLF can make accurate location predictions and open a new door to understanding the multi-modal data of social posts for online inference tasks. | ['Fan Zhou', 'Wanlun Ma', 'Lisi Mo', 'Xucheng Luo', 'Jiayi Luo', 'Ruiting Dai'] | 2023-06-11 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 4.33354862e-02 -1.30039796e-01 -4.71037269e-01 -4.07191008e-01
-1.11736739e+00 -5.57642221e-01 1.11371553e+00 4.07123059e-01
-7.02764750e-01 5.41222513e-01 6.08216465e-01 -6.72049820e-02
-5.93387969e-02 -8.80061328e-01 -8.47841203e-01 -5.91458678e-01
2.22446118e-02 3.47160071e-01 1.85603902e-01 2.17215791e-02
4.29299414e-01 2.06555471e-01 -1.27087414e+00 4.58730817e-01
4.99872953e-01 1.00622106e+00 5.16651094e-01 7.00537384e-01
-2.38014072e-01 5.84033430e-01 -4.32123482e-01 -6.22803152e-01
-3.10946107e-01 2.83924472e-02 -3.45497817e-01 -2.87251204e-01
5.02553463e-01 -5.21442533e-01 -6.94887877e-01 9.33205068e-01
9.12271440e-01 -1.06091406e-02 6.57290280e-01 -1.20966828e+00
-1.26536405e+00 9.45042849e-01 -1.16468477e+00 2.20039219e-01
4.48183775e-01 -2.71037579e-01 8.58699977e-01 -1.09529734e+00
5.14104128e-01 1.39052820e+00 8.34920526e-01 -9.95537192e-02
-6.98519826e-01 -4.41163033e-01 1.80396050e-01 3.20867658e-01
-1.89579105e+00 -3.00616473e-01 6.89404070e-01 -2.97188073e-01
5.51002204e-01 2.22945094e-01 1.54994622e-01 1.40610099e+00
2.91335195e-01 1.12554181e+00 6.08902454e-01 -1.40200168e-01
-2.18920514e-01 3.86431247e-01 -1.34150743e-01 4.56045419e-01
9.30810943e-02 -5.34113646e-01 -6.60255253e-01 -4.55805331e-01
5.15246749e-01 6.01250589e-01 1.17641814e-01 2.56530970e-01
-1.75129187e+00 9.29159343e-01 5.49125850e-01 4.30414289e-01
-5.02043366e-01 4.28241044e-01 1.15329586e-01 -3.22110534e-01
9.99233425e-01 8.81446749e-02 -1.68181539e-01 1.13780849e-01
-1.09941506e+00 1.86846718e-01 5.07859766e-01 8.06735516e-01
1.04199266e+00 -3.52772862e-01 -2.84660727e-01 7.98932910e-01
5.81907332e-01 1.26384008e+00 5.46519756e-01 -4.56832200e-01
5.47014356e-01 3.92718762e-01 1.63543135e-01 -1.73335147e+00
-3.96154314e-01 -2.59959877e-01 -6.22866213e-01 -1.17606366e+00
1.51327640e-01 -5.64190149e-01 -5.33865571e-01 1.60600162e+00
5.52180767e-01 1.05225515e+00 -2.12582052e-01 7.37904370e-01
8.96124780e-01 9.62209463e-01 1.83844075e-01 2.24442124e-01
1.60058832e+00 -7.31698573e-01 -7.18651056e-01 -3.20277154e-01
6.35375082e-01 -9.06856000e-01 3.62176746e-01 -4.82038766e-01
-7.40273535e-01 -4.74856168e-01 -6.33916974e-01 -1.61215603e-01
-1.04130101e+00 2.05578074e-01 5.20968616e-01 1.99305117e-01
-8.73233199e-01 3.85127924e-02 -3.11202496e-01 -3.44776660e-01
3.51765126e-01 1.47642255e-01 -1.96196660e-01 -3.28241706e-01
-1.55870104e+00 5.47556996e-01 1.44659668e-01 3.43843043e-01
-3.43441784e-01 -7.15953112e-01 -1.09699976e+00 -2.47261643e-01
2.23542199e-01 -5.49933314e-01 8.29120636e-01 -4.15513724e-01
-7.03381002e-01 6.97862327e-01 -5.26570857e-01 -4.09664989e-01
2.50589788e-01 -1.84373856e-01 -7.36187756e-01 2.87892401e-01
4.97741699e-01 8.27813089e-01 1.11107242e+00 -9.88330066e-01
-8.81632924e-01 -5.37614584e-01 -4.29644771e-02 2.19044760e-01
-5.29043913e-01 -2.07080152e-02 -7.72169232e-01 -5.60364306e-01
-6.28435314e-02 -1.04613090e+00 -1.36352301e-01 -1.59204587e-01
-4.97998208e-01 -2.63347358e-01 1.24632299e+00 -7.59812832e-01
1.20351064e+00 -2.23494673e+00 -5.11970371e-02 6.59460947e-02
3.03743273e-01 2.44759712e-02 -3.69590253e-01 6.48135185e-01
6.62392437e-01 3.64982337e-02 1.72684789e-01 -7.03053713e-01
2.76658177e-01 2.93815439e-03 -3.73822808e-01 6.87016368e-01
5.16394861e-02 1.47927499e+00 -1.09622037e+00 -9.04561460e-01
4.77893263e-01 1.03776634e+00 -3.34414631e-01 -3.29764485e-01
-1.56870604e-01 3.06721181e-01 -8.68080080e-01 6.26057208e-01
9.05864775e-01 -4.58029598e-01 -1.43945292e-01 -3.56683135e-01
1.02742493e-01 -1.86924696e-01 -8.04785788e-01 1.69770932e+00
-7.17014253e-01 7.86161602e-01 -3.39333266e-02 -3.79597723e-01
5.74455798e-01 5.44116944e-02 6.76974118e-01 -5.88163912e-01
1.94770098e-01 -2.94912249e-01 -7.20482349e-01 -5.02927721e-01
1.30929971e+00 4.14712757e-01 -6.15531147e-01 6.83125854e-01
-2.01847047e-01 2.87131965e-01 -3.05919766e-01 6.39413953e-01
6.60297215e-01 -1.11906305e-01 -2.22974181e-01 1.30165949e-01
3.76315892e-01 -3.50843698e-01 7.77862146e-02 9.36866164e-01
-3.04117985e-02 6.24584079e-01 3.97209190e-02 -1.53312340e-01
-8.56516957e-01 -8.12673271e-01 -2.66798019e-01 1.42706871e+00
5.42190969e-01 -3.07869583e-01 -5.72178781e-01 -6.05708539e-01
4.66361731e-01 4.75218356e-01 -7.33723164e-01 1.10483028e-01
-3.40057224e-01 -1.04554069e+00 7.56241739e-01 2.34478891e-01
5.25831997e-01 -6.23791873e-01 3.08342963e-01 1.47349864e-01
-7.47196555e-01 -1.53831446e+00 -7.60502875e-01 -6.94255054e-01
-3.92968543e-02 -6.64678752e-01 -1.04270852e+00 -7.38896966e-01
7.11484253e-01 7.65295088e-01 6.07820570e-01 -9.57294479e-02
-1.48433283e-01 8.48322392e-01 -3.00339550e-01 -2.36704320e-01
3.24727654e-01 2.87625790e-01 4.42896708e-05 7.52774119e-01
7.41238356e-01 -2.69690871e-01 -8.45054269e-01 4.02463198e-01
-9.90482330e-01 -1.08521149e-01 6.31923854e-01 5.56005538e-01
3.05283666e-01 -2.02400833e-01 8.37690175e-01 -8.34196866e-01
5.20191073e-01 -1.35458553e+00 -1.05314538e-01 3.86741936e-01
1.01631343e-01 -3.42239261e-01 6.23395257e-02 -8.34403694e-01
-9.61714745e-01 -8.32876936e-02 -5.90966865e-02 -1.55852452e-01
1.64036490e-02 8.17406058e-01 1.04848541e-01 -1.14778064e-01
3.34451139e-01 5.09071648e-01 -2.14465097e-01 -3.61583143e-01
6.48808241e-01 1.07738948e+00 2.80637324e-01 -3.11339080e-01
6.86813116e-01 6.98886931e-01 -2.80519545e-01 -1.38691270e+00
-9.56099093e-01 -8.89495969e-01 -5.01507998e-01 -2.54518121e-01
9.19770241e-01 -1.29106641e+00 -8.43166053e-01 6.97478116e-01
-1.22561920e+00 1.34207264e-01 2.74263859e-01 6.17653131e-01
-5.83853088e-02 6.52497590e-01 -8.21886182e-01 -8.42932940e-01
-7.79031590e-02 -8.79501939e-01 1.71525252e+00 1.64497972e-01
2.66207960e-02 -1.53215253e+00 -4.90028076e-02 6.99624956e-01
5.77718437e-01 1.63359791e-01 2.10794792e-01 -6.47475958e-01
-8.29974174e-01 -7.31303692e-01 -7.65166998e-01 -3.30913991e-01
-2.76554506e-02 -3.48770827e-01 -1.09365022e+00 3.14493268e-03
-5.64538538e-01 -6.37212694e-02 9.33483064e-01 4.84956652e-01
1.14651561e+00 -4.36494410e-01 -7.25434124e-01 2.60701180e-01
1.33073401e+00 -5.68529069e-01 5.81369519e-01 -2.34787658e-01
1.26924896e+00 4.68577951e-01 7.03245819e-01 8.47527921e-01
1.11524117e+00 6.08294904e-01 3.31987381e-01 -1.76547199e-01
1.91816047e-01 -6.42953694e-01 1.46109596e-01 6.11290395e-01
4.01052266e-01 -8.40027392e-01 -5.26432991e-01 9.41481054e-01
-2.16950154e+00 -1.15992987e+00 -6.38074651e-02 1.72363937e+00
3.71747315e-01 -3.99905950e-01 -3.14444453e-02 -6.03458643e-01
1.20509887e+00 6.99594915e-01 -4.41377640e-01 2.32480034e-01
-3.13153684e-01 -5.59271872e-01 9.76900518e-01 6.14151597e-01
-1.26983213e+00 9.71494079e-01 5.02101898e+00 1.23132300e+00
-1.08106267e+00 3.92231762e-01 4.67302263e-01 2.04444043e-02
-6.06959105e-01 -4.98125046e-01 -1.36140025e+00 9.64523435e-01
9.94841516e-01 4.91777472e-02 1.73660904e-01 6.97496295e-01
2.12475315e-01 -1.74249560e-01 -5.42886555e-01 1.16091502e+00
5.65784156e-01 -1.66199803e+00 5.72176538e-02 4.43952918e-01
8.69401515e-01 4.59881574e-01 5.44185758e-01 1.27235621e-01
9.63563845e-02 -4.95952427e-01 6.58849597e-01 7.17811406e-01
5.36735296e-01 -6.49452448e-01 8.57105434e-01 3.26413125e-01
-1.31283617e+00 -2.61537661e-03 -2.40434021e-01 2.89764106e-01
6.23692572e-01 7.83069253e-01 -9.86908317e-01 3.82844567e-01
2.41328821e-01 1.14821649e+00 -6.38266325e-01 7.38614142e-01
1.03246175e-01 3.37014943e-01 -6.04104519e-01 -4.23456103e-01
6.17167115e-01 1.95526093e-01 4.07964617e-01 1.44186723e+00
4.13255304e-01 -1.80208802e-01 1.42267123e-01 5.76281786e-01
-2.50167876e-01 6.01708554e-02 -8.26775193e-01 -2.85212219e-01
7.78290212e-01 1.35201621e+00 -5.75420082e-01 -3.82535785e-01
-5.93218744e-01 1.14106131e+00 1.04402952e-01 5.40299237e-01
-1.21575832e+00 -2.72536963e-01 5.11491716e-01 3.86978358e-01
4.98043448e-01 -5.34794450e-01 6.94797710e-02 -1.21072018e+00
-1.76941797e-01 -1.35003090e-01 2.37849459e-01 -9.13831592e-01
-1.29590189e+00 1.57255679e-01 -2.29190234e-02 -9.54758048e-01
-2.25008354e-01 -1.79521963e-01 -3.90676588e-01 6.72454059e-01
-1.82577014e+00 -1.86210060e+00 -6.65994734e-02 6.33418679e-01
2.97524363e-01 7.06360042e-02 4.76090759e-01 8.10936749e-01
-4.61881429e-01 7.91900575e-01 3.66241455e-01 4.31171834e-01
8.21867585e-01 -7.75739431e-01 4.62240785e-01 6.44646168e-01
1.63687661e-01 7.44933724e-01 1.61507219e-01 -8.22275400e-01
-1.84715676e+00 -1.53024864e+00 1.29087877e+00 -7.00709403e-01
9.14244533e-01 -3.25491846e-01 -3.65872175e-01 7.59881318e-01
1.78149298e-01 -3.82159874e-02 1.03414083e+00 1.11367948e-01
-1.64026231e-01 1.18303886e-02 -1.14727557e+00 7.26843536e-01
6.96791828e-01 -1.06022239e+00 -2.84682870e-01 6.49520993e-01
8.90365958e-01 -1.60748810e-01 -8.59239936e-01 1.65813699e-01
5.05947828e-01 -2.68221945e-01 1.45238400e+00 1.15143135e-02
3.61905217e-01 -3.16309512e-01 -3.24395478e-01 -9.13925648e-01
-1.97014585e-01 -3.45861405e-01 -1.95952863e-01 1.28765976e+00
5.41879356e-01 -7.21529067e-01 6.60728991e-01 5.42375922e-01
1.65538982e-01 -2.63459146e-01 -7.44912446e-01 6.83725402e-02
-4.86905545e-01 -5.44058561e-01 6.72213972e-01 1.07306337e+00
2.98618549e-03 3.73969615e-01 -1.04295647e+00 4.62942779e-01
6.69508994e-01 1.74081683e-01 7.56121874e-01 -8.90383124e-01
-9.27713588e-02 1.62911832e-01 -5.28430641e-01 -1.53051329e+00
3.31298262e-01 -6.01326108e-01 -1.02867939e-01 -1.52772474e+00
2.17996851e-01 -5.31638324e-01 -1.91181228e-01 2.40166113e-01
6.82961755e-03 1.03877234e+00 -1.69535026e-01 1.12772770e-01
-1.38554978e+00 3.01758975e-01 1.30099511e+00 -3.94944429e-01
4.97027226e-02 -4.24107872e-02 -6.79086089e-01 5.83377302e-01
4.04897034e-01 -2.83793002e-01 -2.49806926e-01 -6.12354875e-01
6.84878886e-01 1.40769422e-01 5.13482213e-01 -4.58814085e-01
6.32982075e-01 3.26759294e-02 6.02128506e-01 -1.23043871e+00
8.57934654e-01 -7.79644489e-01 -1.76813379e-01 -4.48397323e-02
-3.88795167e-01 -2.01258242e-01 -3.00604124e-02 1.21308219e+00
-9.22896713e-02 2.19215779e-03 1.50449097e-01 -1.59701079e-01
-1.02998114e+00 4.98104066e-01 -4.57677543e-01 1.23817958e-02
8.89781177e-01 -1.27545252e-01 -6.06185853e-01 -6.79085851e-01
-6.90079033e-01 4.17838156e-01 2.16513500e-01 6.85229301e-01
6.74368262e-01 -1.37781060e+00 -8.06423485e-01 5.05496189e-02
3.15643579e-01 -2.65999228e-01 8.80103171e-01 9.12098467e-01
-1.19188629e-01 4.44647014e-01 4.50719059e-01 -7.23655343e-01
-9.81208026e-01 4.32120144e-01 -2.69597173e-01 1.35256782e-01
-1.36399090e-01 8.71399522e-01 1.36248276e-01 -5.01114011e-01
8.00121948e-02 -3.98258828e-02 -1.28726035e-01 5.67567885e-01
1.07366335e+00 1.13238446e-01 -2.53664404e-01 -1.19886279e+00
-3.28609705e-01 6.35499179e-01 -1.89590752e-01 -9.33790132e-02
1.25803483e+00 -7.69842446e-01 -4.95842025e-02 4.31540310e-01
1.59279156e+00 1.12889796e-01 -8.92679870e-01 -7.89594948e-01
-3.15113664e-01 -5.71689188e-01 2.50664741e-01 -3.78357172e-01
-8.30448091e-01 7.56521642e-01 2.40858749e-01 8.01966339e-02
4.85780895e-01 2.31692910e-01 1.38445890e+00 2.77798086e-01
4.33109105e-01 -1.17903709e+00 8.48940834e-02 4.03943121e-01
4.57976520e-01 -1.55570996e+00 -3.63500565e-02 -1.73184261e-01
-7.10980177e-01 5.29527426e-01 9.34752151e-02 2.52956569e-01
1.15311885e+00 1.13322705e-01 -2.01214895e-01 -1.29111513e-01
-3.08865935e-01 -4.03912604e-01 5.69324553e-01 5.44020712e-01
-2.02097334e-02 3.69474888e-02 2.24672839e-01 3.19446653e-01
1.97792545e-01 -3.70475143e-01 6.98438436e-02 7.71724641e-01
-4.57532704e-01 -5.39776802e-01 -3.18874389e-01 3.57273370e-01
-6.52061582e-01 -3.66571397e-01 -2.57343978e-01 2.56024092e-01
1.65889427e-01 9.41906393e-01 3.30370039e-01 -5.99416852e-01
-5.52231729e-01 -8.82276520e-02 1.57200262e-01 -4.07195926e-01
-3.23107600e-01 -1.22245196e-02 4.08659987e-02 -1.51798531e-01
-4.86009002e-01 -6.17773473e-01 -8.28298211e-01 -8.67367685e-01
-5.39165974e-01 1.03342891e-01 1.13136744e+00 9.86268818e-01
8.11594844e-01 8.31861645e-02 8.55947733e-01 -1.21160591e+00
1.42756313e-01 -9.53913331e-01 -5.58934331e-01 2.28302404e-01
4.28780854e-01 -4.95102137e-01 -3.62713784e-01 -1.39187112e-01] | [10.785290718078613, 1.469226598739624] |
db10ea50-baaa-4ce2-b3f1-33822462bb56 | membership-inference-attack-and-defense-for | 2107.12173 | null | https://arxiv.org/abs/2107.12173v1 | https://arxiv.org/pdf/2107.12173v1.pdf | Membership Inference Attack and Defense for Wireless Signal Classifiers with Deep Learning | An over-the-air membership inference attack (MIA) is presented to leak private information from a wireless signal classifier. Machine learning (ML) provides powerful means to classify wireless signals, e.g., for PHY-layer authentication. As an adversarial machine learning attack, the MIA infers whether a signal of interest has been used in the training data of a target classifier. This private information incorporates waveform, channel, and device characteristics, and if leaked, can be exploited by an adversary to identify vulnerabilities of the underlying ML model (e.g., to infiltrate the PHY-layer authentication). One challenge for the over-the-air MIA is that the received signals and consequently the RF fingerprints at the adversary and the intended receiver differ due to the discrepancy in channel conditions. Therefore, the adversary first builds a surrogate classifier by observing the spectrum and then launches the black-box MIA on this classifier. The MIA results show that the adversary can reliably infer signals (and potentially the radio and channel information) used to build the target classifier. Therefore, a proactive defense is developed against the MIA by building a shadow MIA model and fooling the adversary. This defense can successfully reduce the MIA accuracy and prevent information leakage from the wireless signal classifier. | ['Yalin E. Sagduyu', 'Yi Shi'] | 2021-07-22 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 7.95463085e-01 2.84692228e-01 -1.48636967e-01 2.38745306e-02
-1.05733633e+00 -1.10043490e+00 1.88490719e-01 -2.81591376e-04
-7.27796461e-03 5.34232497e-01 -4.83309865e-01 -9.02946472e-01
2.15303395e-02 -1.12239218e+00 -8.14326882e-01 -1.29048252e+00
-5.32698274e-01 -2.72544086e-01 -3.29057388e-02 1.20114550e-01
-1.58557147e-01 6.43981278e-01 -8.97219360e-01 4.89152461e-01
7.67640993e-02 1.50486457e+00 -4.05164123e-01 9.25596893e-01
3.63530308e-01 6.46682203e-01 -1.45427954e+00 -1.94659591e-01
2.24844173e-01 -4.49867725e-01 -3.07472885e-01 -7.32872248e-01
-2.44151652e-01 -1.04735300e-01 -3.91011119e-01 1.05269051e+00
5.17254591e-01 -6.89749241e-01 2.46629611e-01 -1.43964708e+00
5.41744590e-01 1.13141406e+00 1.91322476e-01 -1.53533459e-01
6.25403345e-01 6.00119829e-02 6.46869600e-01 -1.96316943e-01
1.02051176e-01 8.13485026e-01 7.42957652e-01 2.31301308e-01
-1.15733516e+00 -1.52689588e+00 -2.38757715e-01 -2.18953088e-01
-1.58825564e+00 -4.07911927e-01 9.52757537e-01 -1.20739795e-01
-1.21472254e-01 9.21059966e-01 3.80831420e-01 1.58877134e+00
4.54952240e-01 5.47083735e-01 1.22154140e+00 -5.74923277e-01
7.29765475e-01 5.02295673e-01 -1.52844623e-01 5.61677158e-01
2.39783421e-01 9.00633693e-01 -4.36069667e-01 -1.09560752e+00
2.23809090e-02 -2.71448463e-01 -4.38182831e-01 -4.06754650e-02
-8.99966300e-01 5.11975527e-01 1.34688631e-01 1.32390544e-01
-3.50277871e-01 2.65391111e-01 1.05760254e-01 7.15221345e-01
-7.23870248e-02 7.10291088e-01 -4.33361351e-01 8.82947966e-02
-6.70838892e-01 -6.39689639e-02 1.17355311e+00 4.87495273e-01
5.94913602e-01 2.13679314e-01 -3.33757736e-02 -1.07883312e-01
3.44833434e-01 1.14441478e+00 -3.11791860e-02 -3.34982991e-01
4.28714544e-01 -1.76738426e-01 9.58559737e-02 -1.26720834e+00
-2.95705646e-01 -7.85948813e-01 -5.01023829e-01 4.22509253e-01
5.09942591e-01 -5.76512516e-01 -4.04641509e-01 1.81294894e+00
1.96572587e-01 5.34467340e-01 5.83150864e-01 3.37779701e-01
3.81983250e-01 4.93776560e-01 -5.98610453e-02 4.13992964e-02
1.12288654e+00 3.08929421e-02 -5.21110058e-01 -1.47749662e-01
7.97872186e-01 -5.13363004e-01 2.41106749e-01 5.46985567e-01
-6.27964973e-01 -6.20393157e-01 -1.71078897e+00 1.26863754e+00
-4.52731550e-01 -1.25086606e-01 5.74437737e-01 1.55805516e+00
-2.57607460e-01 2.95578003e-01 -4.84510839e-01 2.33555317e-01
3.82091284e-01 6.74625635e-01 -1.92963257e-01 -4.57117483e-02
-1.85145164e+00 2.56367415e-01 2.69856721e-01 1.94368735e-01
-1.15211165e+00 -7.61125982e-01 -7.96773970e-01 1.91911831e-02
2.15924636e-01 -1.55939803e-01 8.24913025e-01 -7.83802688e-01
-1.40618479e+00 4.21452284e-01 3.09574127e-01 -7.48802245e-01
4.27904874e-01 3.58457506e-01 -1.38681674e+00 1.90343708e-01
-2.97313571e-01 -5.10837555e-01 1.37830544e+00 -1.44471991e+00
-7.33214200e-01 -1.59739077e-01 1.93047568e-01 -7.15517461e-01
1.81427244e-02 -2.02960134e-01 3.43702435e-01 -7.51980364e-01
3.19316447e-01 -1.17616522e+00 1.01418674e-01 -4.21202242e-01
-9.32717800e-01 7.96974540e-01 1.20476019e+00 -3.95519882e-01
1.29230988e+00 -2.43061590e+00 -5.29760420e-01 1.22879088e+00
-9.34919268e-02 2.02979326e-01 1.77360073e-01 4.47378010e-01
-2.54492342e-01 2.60087669e-01 -8.30416679e-02 4.83961813e-02
-1.98694050e-01 5.53449541e-02 -9.12195683e-01 8.62453997e-01
-1.68259621e-01 4.10092205e-01 -6.48588598e-01 2.23381296e-01
3.28970626e-02 1.02927230e-01 -1.31621376e-01 4.98164594e-01
9.29914638e-02 7.27374732e-01 -9.18773472e-01 5.67329168e-01
7.09977269e-01 2.92593330e-01 4.62483674e-01 -4.37639356e-01
1.34397164e-01 3.48077416e-01 -1.13888562e+00 1.00236249e+00
-7.77140796e-01 4.39036429e-01 3.82354915e-01 -1.07319117e+00
9.18410838e-01 7.25978434e-01 3.87880623e-01 -3.68158072e-01
5.47026336e-01 1.86207265e-01 2.69533187e-01 -3.07668388e-01
-5.10169327e-01 1.42702207e-01 -9.64973867e-01 8.80824566e-01
-1.79985046e-01 1.97980464e-01 -1.01083267e+00 -4.35441099e-02
1.58649647e+00 -2.57450402e-01 1.59502581e-01 -1.00306973e-01
9.18703020e-01 -4.70679283e-01 3.94176990e-01 1.31566107e+00
2.03264013e-01 -2.73725331e-01 2.91105926e-01 -2.31274143e-01
-2.32134283e-01 -1.22786069e+00 -1.35928392e-01 7.01312959e-01
2.58345157e-01 -1.14719465e-01 -4.86595660e-01 -8.50053310e-01
2.84052342e-01 6.34187281e-01 -6.36644423e-01 -8.41975629e-01
-2.36823857e-01 -4.18110549e-01 1.51128662e+00 4.26125340e-02
5.86721122e-01 -5.89602590e-01 -2.63552696e-01 4.43700522e-01
9.88874137e-02 -1.45790660e+00 2.08129615e-01 5.62299132e-01
-6.36989102e-02 -9.48755860e-01 3.70725304e-01 -2.72949994e-01
6.28295064e-01 -4.66037542e-02 3.67847174e-01 2.84688145e-01
-1.31086960e-01 3.02336603e-01 -3.28956425e-01 -6.84190929e-01
-1.27913702e+00 -1.05238333e-01 3.38193744e-01 6.65920973e-01
3.62119940e-03 -6.69172108e-01 -1.38198733e-01 5.27834356e-01
-5.65273404e-01 -6.26633763e-01 5.43484926e-01 4.42100346e-01
-8.10375139e-02 8.53559792e-01 6.51072979e-01 -1.07495701e+00
2.73592293e-01 -6.35659575e-01 -7.69613922e-01 3.31526011e-01
-1.52260855e-01 1.60422400e-01 7.81154096e-01 -6.97799802e-01
-6.26824081e-01 2.28139535e-02 -4.12514240e-01 1.50342315e-01
-1.78610757e-01 5.14558792e-01 -1.15673172e+00 -7.33755052e-01
5.56697965e-01 1.71389177e-01 -2.58842438e-01 -2.32449546e-01
-2.10775919e-02 1.00422287e+00 7.47546613e-01 -5.28378844e-01
1.77189076e+00 5.79844296e-01 3.61135036e-01 -6.96140289e-01
-9.17719603e-01 1.30779833e-01 -1.79790542e-01 -1.96096420e-01
4.12137449e-01 -7.05340624e-01 -9.69187081e-01 6.10306203e-01
-8.41447949e-01 -2.91655138e-02 2.22484961e-01 5.76627314e-01
-4.85720038e-02 2.37160344e-02 -2.58878887e-01 -1.27272642e+00
-1.30503654e-01 -9.77627456e-01 4.42534447e-01 -6.48847967e-02
-3.78531516e-01 -8.43573034e-01 -3.72226298e-01 3.04236919e-01
3.51374239e-01 7.17054784e-01 1.03225207e+00 -9.44011986e-01
-4.56981480e-01 -1.04447448e+00 4.95427251e-01 1.24974489e-01
1.58585712e-01 -3.66838813e-01 -1.61059821e+00 -3.87002230e-01
5.07409751e-01 9.46753323e-02 8.94858763e-02 1.29544273e-01
1.33824849e+00 -7.07946658e-01 -6.32458031e-01 9.05335307e-01
1.09432614e+00 5.50613463e-01 8.26734066e-01 4.69494574e-02
3.47317845e-01 1.80859119e-01 5.79800308e-01 2.66682088e-01
-4.14890528e-01 5.23896515e-01 4.68414128e-01 -3.63159597e-01
5.74654520e-01 -4.08171803e-01 6.50230825e-01 2.22019088e-02
5.54346561e-01 -4.36923057e-01 -3.16912800e-01 -4.01961863e-01
-1.17083991e+00 -1.10575557e+00 7.32522532e-02 2.41082430e+00
6.72677994e-01 8.65699470e-01 -1.75829262e-01 8.64742279e-01
4.74684745e-01 1.37592092e-01 -4.12073165e-01 -4.44448531e-01
-1.32797524e-01 4.80486214e-01 1.10115302e+00 5.02446949e-01
-1.27873433e+00 4.36856866e-01 5.12749243e+00 7.13069260e-01
-1.29236400e+00 -1.03416375e-03 1.53337657e-01 6.51099086e-01
-2.39827871e-01 1.58364207e-01 -6.45642757e-01 5.53856790e-01
1.07376158e+00 2.47135669e-01 4.51663405e-01 6.42379224e-01
9.36359465e-02 -3.47618312e-02 -1.24047434e+00 6.88109398e-01
3.09421076e-03 -1.15906465e+00 -4.14854556e-01 4.63450849e-01
1.55877182e-02 -5.21560490e-01 2.76636273e-01 4.52066749e-01
2.14130148e-01 -8.67426336e-01 7.46644199e-01 4.20802057e-01
8.53891790e-01 -9.36544836e-01 9.30273175e-01 5.80354214e-01
-1.07500994e+00 -3.95830095e-01 3.50450933e-01 -4.98967767e-02
-2.35726297e-01 5.37023544e-01 -9.51261818e-01 6.70671701e-01
3.13630134e-01 -2.50778884e-01 -3.20024222e-01 4.95355874e-01
-5.41968405e-01 1.38697243e+00 -4.04836804e-01 5.58017418e-02
6.97054900e-03 3.68383139e-01 7.59592354e-01 7.96947300e-01
1.17287599e-01 2.16840524e-02 2.97622144e-01 6.92395568e-01
-9.65369344e-02 -5.04227102e-01 -5.61184287e-01 -1.86406877e-02
9.88640368e-01 1.00014448e+00 -1.20802626e-01 1.95357367e-01
-5.13332058e-03 7.18827724e-01 -6.79388344e-01 5.47127724e-01
-9.30138052e-01 -5.96740186e-01 7.97919691e-01 9.00436118e-02
1.55615598e-01 -1.11794248e-02 -6.80202097e-02 -4.99941915e-01
-2.33745500e-01 -1.04943800e+00 3.26872706e-01 -2.90333927e-01
-1.02491271e+00 4.03806955e-01 -4.26277399e-01 -1.50080574e+00
-3.81269842e-01 -2.67937034e-01 -7.34354556e-01 9.79155779e-01
-1.29628432e+00 -1.21290123e+00 -5.03289402e-02 6.94391966e-01
-3.98612082e-01 -4.66592491e-01 1.44890070e+00 -2.30250210e-02
-3.27188939e-01 1.16012561e+00 1.82930510e-02 7.36363411e-01
1.07387111e-01 -6.10680521e-01 2.28985712e-01 9.92987335e-01
1.41174465e-01 4.98350531e-01 7.49728024e-01 -4.48047668e-01
-1.79132080e+00 -1.02122140e+00 1.01976201e-01 -2.07961693e-01
7.25718796e-01 -9.90104079e-01 -6.52087629e-01 2.69855976e-01
-5.53383529e-01 1.30076453e-01 1.29171538e+00 -1.65643156e-01
-4.95808125e-01 -4.68893051e-01 -1.60210860e+00 3.90849471e-01
1.17208615e-01 -8.95417929e-01 -4.61702421e-02 -1.72281172e-02
6.08841598e-01 -1.20320529e-01 -8.91975880e-01 3.74769211e-01
9.93138909e-01 -3.90512615e-01 1.06778944e+00 -6.28259301e-01
-7.24707186e-01 -4.09891486e-01 -5.37929714e-01 -9.85687017e-01
1.73836648e-01 -1.31272018e+00 -4.76604581e-01 1.34609103e+00
6.80862725e-01 -8.63038361e-01 7.19986081e-01 2.57104307e-01
6.48656130e-01 -2.67017812e-01 -1.02848697e+00 -8.96081448e-01
-2.97028005e-01 -9.75160301e-01 1.08146989e+00 9.14253235e-01
-6.61611930e-02 -1.00776017e-01 -4.77082968e-01 1.37448382e+00
9.51150477e-01 -1.03371680e-01 1.03456736e+00 -1.21077871e+00
-6.76782131e-01 7.39885196e-02 -5.22301137e-01 -7.25956440e-01
1.94799826e-01 -9.13698256e-01 -3.21345106e-02 -2.13917479e-01
-8.46338153e-01 -1.13361716e+00 -6.12216592e-01 3.78168494e-01
2.43047819e-01 6.50964528e-02 2.42607035e-02 -1.61592543e-01
1.69307381e-01 -1.38778090e-01 3.89726192e-01 -3.72823209e-01
-9.52577367e-02 1.23870683e+00 -6.13038957e-01 7.17127919e-01
1.04048741e+00 -1.00776446e+00 -1.08649537e-01 5.28788507e-01
1.08404323e-01 4.76137340e-01 6.80291593e-01 -1.52056801e+00
3.74459587e-02 1.47559851e-01 4.12355900e-01 -3.39041591e-01
1.66698515e-01 -1.51256239e+00 2.56031007e-01 9.05451000e-01
-2.85530686e-01 -8.32913935e-01 5.86216003e-02 7.07478404e-01
1.66976914e-01 -2.06410751e-01 5.92800736e-01 4.94584709e-01
-7.49609619e-02 1.57505460e-02 -6.77715003e-01 -3.07759404e-01
9.37700391e-01 -9.13274661e-02 -1.83850631e-01 -6.40571535e-01
-6.95225477e-01 -1.38961032e-01 -3.16946418e-03 2.32958153e-01
4.49005187e-01 -1.01867163e+00 -3.48816454e-01 7.89640069e-01
1.37610599e-01 -5.95891595e-01 -1.06161036e-01 2.29858220e-01
6.38726577e-02 8.07014406e-02 4.16301817e-01 -4.69866604e-01
-1.47324729e+00 4.82621521e-01 6.06448591e-01 -1.39980331e-01
-4.02568400e-01 5.32597780e-01 -2.26505727e-01 -1.11602820e-01
6.62911773e-01 -3.73797417e-02 1.93335354e-01 -3.97394210e-01
6.79388940e-01 -1.24414742e-01 2.48282135e-01 -2.76704818e-01
-5.14640570e-01 2.41426006e-01 5.48084915e-01 -1.39150560e-01
7.02966630e-01 7.83965811e-02 3.55479121e-01 2.17307553e-01
1.58819115e+00 8.39898169e-01 -7.75485158e-01 -3.88457149e-01
-8.96931905e-03 -4.34118539e-01 6.35931790e-02 -9.21143293e-01
-9.32521701e-01 5.85583508e-01 8.31334710e-01 7.09233999e-01
1.10355997e+00 -1.53660133e-01 8.42654049e-01 4.01179105e-01
9.65820849e-01 -6.12321496e-01 -3.70656550e-01 -4.99475226e-02
2.27176502e-01 -6.29114509e-01 4.04755175e-02 -3.33294570e-01
-1.53799608e-01 1.01554883e+00 -2.04228550e-01 2.07911298e-01
1.06150234e+00 1.01947451e+00 3.37147534e-01 6.51821196e-02
-1.64463520e-01 3.50273162e-01 -2.94775795e-02 1.00256240e+00
-5.22231519e-01 2.83416182e-01 4.53059763e-01 1.10528731e+00
-5.26793480e-01 -4.45790708e-01 3.82407248e-01 9.43964839e-01
-4.34958458e-01 -1.21272588e+00 -7.64595807e-01 2.29005188e-01
-6.70585454e-01 2.16218516e-01 -3.56023163e-01 5.35760641e-01
3.66080403e-01 1.35733902e+00 -4.25149411e-01 -9.78380322e-01
2.72318572e-01 -2.45160016e-04 -5.08594289e-02 -2.40243480e-01
-5.91981232e-01 -3.35816771e-01 2.69994229e-01 -4.74352241e-01
-9.82843861e-02 -4.38480496e-01 -1.20269012e+00 -1.44465774e-01
-5.37197232e-01 4.49102700e-01 8.66238177e-01 1.07637954e+00
-1.60700291e-01 5.33507943e-01 1.79682136e+00 -1.92594901e-01
-6.05603695e-01 -4.47706014e-01 -6.62263811e-01 -1.98699199e-02
6.03662491e-01 -5.95037520e-01 -9.64311123e-01 -5.33852100e-01] | [13.840845108032227, 5.80043888092041] |
67016273-5b2e-4e6c-8d63-d7338089972d | the-myth-of-culturally-agnostic-ai-models | 2211.15271 | null | https://arxiv.org/abs/2211.15271v2 | https://arxiv.org/pdf/2211.15271v2.pdf | The Myth of Culturally Agnostic AI Models | The paper discusses the potential of large vision-language models as objects of interest for empirical cultural studies. Focusing on the comparative analysis of outputs from two popular text-to-image synthesis models, DALL-E 2 and Stable Diffusion, the paper tries to tackle the pros and cons of striving towards culturally agnostic vs. culturally specific AI models. The paper discusses several examples of memorization and bias in generated outputs which showcase the trade-off between risk mitigation and cultural specificity, as well as the overall impossibility of developing culturally agnostic models. | ['Eva Cetinic'] | 2022-11-28 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [-3.08067026e-03 2.94781417e-01 4.34044078e-02 -1.40727177e-01
-5.33021927e-01 -5.81748128e-01 1.24541676e+00 1.28112882e-02
-7.11224139e-01 4.94388103e-01 8.90833020e-01 -3.00855249e-01
1.37613351e-02 -4.39765036e-01 -6.01059794e-01 -5.02108097e-01
3.53066415e-01 1.78498983e-01 -5.99191606e-01 -4.60416853e-01
6.06608808e-01 1.37478396e-01 -1.39559281e+00 1.97107658e-01
9.01151001e-01 4.29739147e-01 5.57062030e-02 6.57772422e-01
-1.19658999e-01 9.14265096e-01 -8.04362714e-01 -1.03492928e+00
1.13723926e-01 -7.81102717e-01 -3.45325261e-01 -3.53539176e-02
6.02824330e-01 -1.89142376e-01 -7.52464607e-02 1.11895239e+00
6.71715736e-01 -3.40423137e-01 1.11490881e+00 -1.24869668e+00
-1.67631161e+00 9.19826508e-01 -2.25976467e-01 1.45465776e-01
2.32174650e-01 5.37648082e-01 5.75868487e-01 -6.00290775e-01
9.37430859e-01 1.51759803e+00 5.97931206e-01 7.20246196e-01
-1.10038757e+00 -6.74483836e-01 3.25338662e-01 -1.53571203e-01
-1.43107092e+00 -6.23071671e-01 5.90808094e-01 -7.98483908e-01
9.35419500e-01 4.31691483e-02 9.85667646e-01 1.80927277e+00
5.15314698e-01 4.98204917e-01 1.55140555e+00 -4.84816760e-01
3.49837244e-01 7.44678438e-01 -2.84633011e-01 4.97348338e-01
4.40034717e-01 4.93615478e-01 -8.79418552e-01 -1.56768709e-01
9.34070706e-01 -7.59679377e-01 -1.51017979e-01 -1.61823899e-01
-1.27580369e+00 9.23545003e-01 2.11358994e-01 3.77895296e-01
-1.64293632e-01 3.22779119e-01 3.67334783e-01 4.87506181e-01
5.69531918e-01 6.57372773e-01 1.76168516e-01 -3.98449227e-02
-7.12292373e-01 3.58736813e-01 3.50299180e-01 9.00646508e-01
1.75541714e-01 6.49747968e-01 -7.69370422e-02 8.06082845e-01
3.95895332e-01 6.58455074e-01 5.72696865e-01 -1.12047398e+00
1.74184889e-01 1.29485279e-01 6.59235865e-02 -1.07027376e+00
-1.91884175e-01 -4.25504744e-01 -2.40127772e-01 3.10631961e-01
1.85539171e-01 -6.25179291e-01 -7.41110682e-01 1.89362967e+00
-3.43593150e-01 -2.93290079e-01 2.87156641e-01 9.31006432e-01
5.58696866e-01 5.74182093e-01 6.80185020e-01 -1.33536622e-01
7.98395157e-01 -6.91259682e-01 -5.28039575e-01 -5.23466110e-01
6.41148686e-01 -7.75996208e-01 1.14753115e+00 6.49470612e-02
-1.21523309e+00 -3.02461922e-01 -9.07231688e-01 6.33889809e-02
-3.66041392e-01 -7.89958984e-02 7.05567837e-01 1.28362310e+00
-1.37724423e+00 1.25967667e-01 -2.84847528e-01 -8.57278526e-01
1.78618640e-01 -1.80077031e-01 -8.19420964e-02 1.46043405e-01
-8.48335922e-01 1.21248567e+00 2.05599889e-01 -4.79642041e-02
-9.38837171e-01 -5.38491189e-01 -7.26873934e-01 -2.46289060e-01
-6.32534847e-02 -8.50990713e-01 9.62769210e-01 -2.01459455e+00
-1.38624024e+00 9.14955974e-01 1.77728996e-01 -1.51830584e-01
5.98902881e-01 -2.15403214e-01 -6.65102661e-01 -1.83862999e-01
8.24702159e-02 1.01153493e+00 7.01202989e-01 -1.55965269e+00
-5.09409368e-01 -2.20083728e-01 -1.47034720e-01 5.24317563e-01
-4.53941405e-01 4.84128624e-01 8.77889246e-02 -1.13808167e+00
-4.80959743e-01 -7.50771940e-01 -2.69130349e-01 -5.69320582e-02
3.92749831e-02 2.27644518e-01 1.08086742e-01 -7.64803708e-01
1.03448856e+00 -2.28854012e+00 2.48470813e-01 1.26909867e-01
-5.52497432e-02 -4.74707298e-02 -6.05711818e-01 5.74872434e-01
2.07495261e-02 4.11456317e-01 9.07348245e-02 -6.42623752e-02
-9.95119568e-03 -3.09970472e-02 -2.27631569e-01 3.58866215e-01
5.77890277e-01 8.36293697e-01 -9.40959334e-01 -4.21221197e-01
-1.61758453e-01 5.31235218e-01 -5.65031648e-01 -4.40892279e-01
-6.51075020e-02 8.62076040e-03 -1.68772921e-01 6.87844753e-01
2.91159630e-01 4.62749213e-01 3.64699215e-01 4.87887561e-01
-5.08817434e-01 -3.39989997e-02 -7.03658402e-01 1.58455110e+00
-3.51169437e-01 9.87635136e-01 -9.80201885e-02 -3.23682487e-01
8.12425196e-01 1.83404103e-01 -3.10921192e-01 -1.13651979e+00
3.66075307e-01 2.09098771e-01 2.37067357e-01 -2.77760416e-01
8.30508769e-01 -5.65544963e-01 -7.72483423e-02 5.65241456e-01
2.64245155e-03 -4.12513793e-01 -2.33188048e-01 1.25898212e-01
2.90986598e-01 4.46086556e-01 -4.45603020e-02 -7.89321482e-01
-1.46458801e-02 4.32864457e-01 3.23091447e-01 7.35673845e-01
-3.50447237e-01 6.38212740e-01 4.19178724e-01 -1.07227527e-01
-1.31507623e+00 -1.17942703e+00 2.05653995e-01 9.23119247e-01
-1.04405381e-01 -3.15272868e-01 -9.69764590e-01 -3.84528220e-01
-1.14288375e-01 1.43191504e+00 -9.43520963e-01 -5.67227960e-01
-3.25154662e-02 -6.18134856e-01 7.40675747e-01 4.06562507e-01
2.32827336e-01 -6.91112638e-01 -8.25445354e-01 -1.72202140e-01
1.22866668e-01 -5.76029778e-01 -2.13781491e-01 -2.72219390e-01
-8.38226438e-01 -2.44969100e-01 -1.09069800e+00 -7.51781881e-01
5.33353329e-01 3.44423085e-01 1.07558417e+00 -2.45939389e-01
-5.53545691e-02 1.03611922e+00 -1.95500165e-01 -1.08666706e+00
-6.34002388e-01 -1.68711662e-01 1.10336371e-01 -2.31873348e-01
5.26871502e-01 -2.06569701e-01 -3.75143737e-01 5.83213903e-02
-8.01562846e-01 2.35065222e-02 5.19126475e-01 4.98148203e-01
-4.89394478e-02 -3.94065708e-01 6.29122674e-01 -5.28113663e-01
1.11085331e+00 -5.75070739e-01 -2.20699832e-01 4.68503416e-01
-6.99773908e-01 -9.53778848e-02 3.49842235e-02 -5.90359151e-01
-1.50050366e+00 -2.57753283e-01 4.09971505e-01 -7.08971396e-02
2.32809093e-02 5.95813215e-01 2.10489243e-01 -1.38836890e-01
1.08587718e+00 2.57429749e-01 9.37853605e-02 4.88813315e-03
5.83991349e-01 5.12955487e-01 2.53503174e-01 -6.78063333e-01
7.32848823e-01 3.67945135e-01 -7.05465317e-01 -1.25179422e+00
-1.11771718e-01 4.39579546e-01 -2.06467971e-01 -7.84479737e-01
9.73080337e-01 -1.35975778e+00 6.66457564e-02 4.47007596e-01
-8.01328659e-01 -5.20814121e-01 -3.59557718e-01 5.13596416e-01
-7.76840210e-01 1.28177539e-01 -3.73456478e-01 -9.95047092e-01
-6.09753802e-02 -9.50759709e-01 6.89020693e-01 1.36568144e-01
-6.84822738e-01 -1.09554851e+00 2.44163990e-01 3.81776571e-01
5.42816997e-01 1.46139145e-01 1.22680497e+00 -1.17579393e-01
-3.27570230e-01 1.29580036e-01 -2.82584894e-02 1.59947440e-01
-1.55160934e-01 4.78485107e-01 -1.04351747e+00 -6.37134211e-03
-1.40845656e-01 -6.73685491e-01 6.86580241e-01 3.32323968e-01
-4.98396577e-03 -2.51029104e-01 1.36486977e-01 3.85721236e-01
1.66052330e+00 2.93655396e-01 6.62922323e-01 6.63678765e-01
4.06202376e-01 1.02633655e+00 4.35849339e-01 2.80294836e-01
4.81940985e-01 1.31154642e-01 -4.93040010e-02 1.87773019e-01
-1.30556356e-02 -4.28018034e-01 6.54258251e-01 8.30042541e-01
-2.40947202e-01 -4.53033112e-02 -1.01200306e+00 8.93276513e-01
-1.67007613e+00 -8.77443969e-01 1.00558259e-01 1.97076392e+00
5.76647103e-01 -3.32332961e-02 2.31764421e-01 -5.17296672e-01
4.78820592e-01 2.28674561e-01 -1.07046947e-01 -1.07057178e+00
-6.78514898e-01 -1.50811225e-01 5.09121776e-01 3.78894061e-01
-4.57279533e-01 1.12808001e+00 8.26533794e+00 5.89809120e-01
-1.15625536e+00 -8.36479887e-02 8.17324519e-01 -5.40487707e-01
-8.69722307e-01 4.01503965e-02 -2.00863466e-01 1.17678225e-01
1.14702702e+00 -4.41604644e-01 3.33301544e-01 6.94774330e-01
2.12232918e-01 -3.28568518e-01 -8.51054192e-01 5.84530175e-01
5.71885645e-01 -1.17717826e+00 5.11230826e-01 -9.24648792e-02
9.73627388e-01 -3.05266511e-02 6.40459239e-01 1.67699680e-01
5.37466824e-01 -1.03118742e+00 1.82031369e+00 4.17932242e-01
7.19850957e-01 -8.18164170e-01 2.89332658e-01 -2.84846853e-02
-2.68006176e-01 -3.78213435e-01 -3.82585257e-01 -4.07714844e-01
-8.07531625e-02 -2.47695461e-01 -3.59930933e-01 1.70920342e-01
7.53289342e-01 2.87665784e-01 -7.14714527e-01 6.53994679e-01
5.84170744e-02 4.91493136e-01 7.36170486e-02 -2.51920074e-01
4.76457417e-01 -3.70656282e-01 4.20739025e-01 1.36881852e+00
4.76700395e-01 -6.92968518e-02 -5.60838163e-01 8.41292560e-01
5.45451045e-01 3.97903651e-01 -1.21239281e+00 -6.23656690e-01
2.43018880e-01 5.77855527e-01 -6.02997780e-01 -2.12933809e-01
-5.55511057e-01 9.59446847e-01 3.05512100e-01 5.03432930e-01
-5.63477635e-01 3.68483141e-02 8.12608182e-01 -3.48014943e-02
-2.24964414e-02 -5.34720719e-01 -8.53209138e-01 -9.62685585e-01
-4.09744680e-01 -1.21499419e+00 1.48050711e-01 -1.15492284e+00
-1.17417467e+00 2.43650690e-01 3.86896320e-02 -4.87041056e-01
-9.99787007e-04 -4.63255554e-01 -3.73064786e-01 9.87712860e-01
-9.84807968e-01 -1.49555171e+00 7.06052631e-02 3.36485714e-01
5.26302934e-01 -4.93461460e-01 6.97328985e-01 -2.01622352e-01
-4.57102716e-01 8.19953501e-01 2.58490056e-01 -2.26185188e-01
8.55479360e-01 -6.84432089e-01 4.53981608e-01 7.63855338e-01
-4.03122269e-02 6.83512330e-01 8.14210773e-01 -9.72881615e-01
-1.23443437e+00 -7.40498543e-01 9.98467147e-01 -8.58578682e-01
6.67806506e-01 -3.41913849e-01 -2.11298943e-01 7.19753146e-01
7.83323824e-01 -1.34984684e+00 8.05913925e-01 -1.43686011e-01
-5.45907915e-01 3.47054809e-01 -9.20694828e-01 1.60373402e+00
1.11271787e+00 -7.06885338e-01 -4.72195953e-01 -6.06343821e-02
5.85176647e-01 2.63374150e-01 -5.28341234e-01 -3.16757470e-01
7.87831187e-01 -1.03170466e+00 8.05714250e-01 -6.70413136e-01
8.74173701e-01 2.26088732e-01 -3.17188948e-01 -1.39503121e+00
-7.12148130e-01 -5.89705110e-01 7.52975047e-01 1.38419080e+00
6.37054741e-01 -4.93439883e-01 3.25748056e-01 6.35202467e-01
9.26193744e-02 -1.31427811e-03 -6.96202278e-01 -6.19295537e-01
6.55763268e-01 -3.74986708e-01 3.05224270e-01 1.39258611e+00
2.80823885e-03 2.83562660e-01 -2.67030030e-01 -1.32351648e-02
4.65579599e-01 -5.54152012e-01 6.43301845e-01 -7.81333268e-01
-4.01843265e-02 -8.90312195e-01 -3.77155960e-01 1.76246658e-01
6.03504367e-02 -5.20450473e-01 -3.67034286e-01 -1.43973792e+00
2.70133823e-01 2.82091890e-02 -8.53297338e-02 -4.31992346e-03
-3.98633890e-02 2.01629758e-01 6.77140594e-01 -3.31158354e-03
-1.57442801e-02 4.32377130e-01 1.05722868e+00 -1.83402821e-01
-2.50236005e-01 -7.24946499e-01 -1.43246830e+00 6.62603378e-01
8.04314733e-01 -3.33613008e-01 -7.53041148e-01 -1.04162872e+00
6.96070790e-01 -2.68621385e-01 4.27552283e-01 -8.63826215e-01
-2.15689704e-01 -6.80462182e-01 4.47291583e-01 1.08464696e-01
2.97408253e-01 -7.07672954e-01 4.02975500e-01 5.84373593e-01
-5.83737671e-01 4.74768788e-01 3.94018382e-01 5.91505170e-01
4.16251011e-02 -2.61931807e-01 5.77757061e-01 -4.73247051e-01
-7.49609828e-01 -4.68259990e-01 -1.00435865e+00 1.07374024e-02
1.11164117e+00 -6.42676711e-01 -4.23798352e-01 -6.52069092e-01
-3.20227563e-01 6.17901199e-02 9.21276033e-01 9.64950562e-01
4.89713192e-01 -1.23319471e+00 -9.41327274e-01 -2.44777516e-01
4.64670628e-01 -1.02294159e+00 3.55592936e-01 3.79409075e-01
-6.56780243e-01 3.56601775e-01 -7.97459722e-01 -6.07144795e-02
-1.06382394e+00 5.73567033e-01 3.50440353e-01 6.72452390e-01
-2.33791754e-01 9.49421883e-01 3.84218216e-01 6.84095770e-02
7.13047162e-02 8.87844786e-02 -1.84278995e-01 1.93145350e-01
2.22553283e-01 3.57169330e-01 -5.04800558e-01 -6.18404508e-01
-1.70768261e-01 5.11592388e-01 1.91881265e-02 -8.89386177e-01
1.01174951e+00 -2.71085948e-01 -6.30106777e-02 7.16026843e-01
4.87324089e-01 2.53673773e-02 -1.03962147e+00 1.75910722e-02
-4.09771614e-02 -5.68967044e-01 5.82514480e-02 -1.09583449e+00
-7.35454500e-01 8.40623438e-01 8.27559948e-01 -4.15923655e-01
8.62110794e-01 -2.77401954e-01 8.83718878e-02 1.78131491e-01
2.40074560e-01 -1.75506043e+00 -7.65389577e-02 2.22175598e-01
1.15991759e+00 -8.48518491e-01 2.63281260e-02 1.01685159e-01
-1.45967412e+00 6.62262082e-01 7.65048563e-01 -6.94141388e-02
2.79548258e-01 9.47490856e-02 7.04168320e-01 -2.31757108e-02
-8.69233310e-01 8.06762557e-03 -3.22961621e-02 1.31582546e+00
5.27259469e-01 2.44109899e-01 -6.86876237e-01 3.38802069e-01
-4.15921032e-01 -1.32463768e-01 7.82553136e-01 1.10003507e+00
-1.90085843e-01 -7.59609103e-01 -5.47769964e-01 4.40908074e-02
-2.75722742e-01 -2.21834347e-01 -1.40911114e+00 9.82425809e-01
5.98430894e-02 8.11735988e-01 2.18178242e-01 -4.33357388e-01
2.45050807e-02 3.51184905e-01 4.88648683e-01 -2.92914718e-01
-1.25743949e+00 1.37179300e-01 5.12254357e-01 -2.15589732e-01
-2.96601564e-01 -8.11391175e-01 -4.83682096e-01 -2.98768848e-01
-1.97583791e-02 -3.74344587e-01 9.37078953e-01 5.35633564e-01
3.72736454e-01 1.29902631e-01 1.80412143e-01 -4.56734508e-01
-3.68250102e-01 -6.43486917e-01 -5.15782237e-01 1.42620444e-01
1.07608542e-01 -2.51566201e-01 -2.46541612e-02 8.17359835e-02] | [11.624587059020996, 0.8943983912467957] |
475c9003-aaeb-429e-af0a-291317fcc3cd | learning-based-real-time-detection-of | 1911.00189 | null | https://arxiv.org/abs/1911.00189v1 | https://arxiv.org/pdf/1911.00189v1.pdf | Learning-based Real-time Detection of Intrinsic Reflectional Symmetry | Reflectional symmetry is ubiquitous in nature. While extrinsic reflectional symmetry can be easily parametrized and detected, intrinsic symmetry is much harder due to the high solution space. Previous works usually solve this problem by voting or sampling, which suffer from high computational cost and randomness. In this paper, we propose \YL{a} learning-based approach to intrinsic reflectional symmetry detection. Instead of directly finding symmetric point pairs, we parametrize this self-isometry using a functional map matrix, which can be easily computed given the signs of Laplacian eigenfunctions under the symmetric mapping. Therefore, we train a novel deep neural network to predict the sign of each eigenfunction under symmetry, which in addition takes the first few eigenfunctions as intrinsic features to characterize the mesh while avoiding coping with the connectivity explicitly. Our network aims at learning the global property of functions, and consequently converts the problem defined on the manifold to the functional domain. By disentangling the prediction of the matrix into separated basis, our method generalizes well to new shapes and is invariant under perturbation of eigenfunctions. Through extensive experiments, we demonstrate the robustness of our method in challenging cases, including different topology and incomplete shapes with holes. By avoiding random sampling, our learning-based algorithm is over 100 times faster than state-of-the-art methods, and meanwhile, is more robust, achieving higher correspondence accuracy in commonly used metrics. | ['Yu-Kun Lai', 'Shu-Zhi Liu', 'Yi-Ling Qiao', 'Xilin Chen', 'Lin Gao', 'Ligang Liu'] | 2019-11-01 | null | null | null | null | ['symmetry-detection'] | ['computer-vision'] | [ 2.54156739e-01 -6.18633591e-02 -1.39833465e-01 -1.14752039e-01
-5.60069203e-01 -5.61356425e-01 4.35211629e-01 -1.71613693e-01
9.76839568e-03 5.14261127e-01 -7.67959803e-02 -8.94790590e-02
-3.62992465e-01 -1.18512702e+00 -8.26953888e-01 -8.17537904e-01
4.16874923e-02 7.14367509e-01 -1.48713917e-01 -2.27380276e-01
4.14908111e-01 9.10602748e-01 -1.40860629e+00 2.31444821e-01
8.09795201e-01 1.07341743e+00 -1.89811230e-01 1.79099917e-01
1.28275260e-01 2.20306292e-01 -2.00438276e-01 -5.02177738e-02
5.86462498e-01 -3.40189546e-01 -8.67025077e-01 1.97608903e-01
6.70471072e-01 -5.62462173e-02 -3.44014496e-01 1.14117491e+00
4.37585264e-01 6.93932176e-02 8.16983342e-01 -1.09218109e+00
-6.49867415e-01 3.86076808e-01 -5.30314386e-01 -5.21770597e-01
3.52788389e-01 -4.46342677e-02 1.34384954e+00 -1.11006856e+00
6.35738134e-01 1.00864625e+00 9.05429602e-01 2.64154822e-01
-1.60413802e+00 -6.45807028e-01 -3.24327499e-01 7.95342624e-02
-1.74431109e+00 -2.95100600e-01 1.31077468e+00 -5.30445933e-01
4.57234651e-01 3.85530740e-01 6.23572707e-01 7.06832528e-01
1.02133907e-01 3.73800188e-01 9.61988211e-01 -2.42537305e-01
1.70485899e-01 -2.91842014e-01 -3.15010667e-01 9.71986949e-01
2.08481148e-01 -1.87202618e-01 -2.63576627e-01 -1.13242380e-01
1.12692857e+00 2.90517002e-01 -5.60344875e-01 -9.88626599e-01
-1.43277967e+00 7.11995482e-01 5.86212039e-01 1.41299099e-01
-2.63112158e-01 -1.57890096e-01 1.81272075e-01 3.83263946e-01
3.04437995e-01 5.89727938e-01 -4.22342539e-01 1.64101213e-01
-8.51896048e-01 9.48407352e-02 8.43333304e-01 5.71896791e-01
1.11337829e+00 -2.09432244e-01 1.17472745e-01 8.48146856e-01
1.17850304e-01 5.95429480e-01 1.92954227e-01 -9.49036539e-01
2.85832167e-01 1.02389705e+00 -8.00710544e-02 -1.50857139e+00
-6.95896745e-01 -4.23201829e-01 -1.46426857e+00 3.58218759e-01
5.03315806e-01 2.22524479e-01 -6.60512567e-01 1.66505468e+00
1.75797969e-01 -1.32669255e-01 -3.43552142e-01 7.16444373e-01
6.54962480e-01 4.27913398e-01 -8.22450161e-01 1.42314509e-02
9.70809340e-01 -6.67349100e-01 -1.19599395e-01 3.98859501e-01
4.32457596e-01 -6.80259407e-01 1.02143145e+00 3.27824354e-01
-9.68836248e-01 -4.25356507e-01 -1.15705252e+00 6.17498858e-03
-1.03377588e-01 2.53649145e-01 2.86136657e-01 3.66447389e-01
-9.91492867e-01 1.00963748e+00 -7.47353673e-01 1.14343194e-02
5.06397963e-01 5.25310338e-01 -6.05900645e-01 -2.52510468e-03
-8.47668290e-01 5.39209604e-01 -6.26625866e-02 1.00770593e-01
-2.03206912e-01 -9.29814219e-01 -8.41082752e-01 1.76199213e-01
2.09431335e-01 -6.67086303e-01 5.46764433e-01 -7.91991293e-01
-1.62221110e+00 8.15819085e-01 -1.30209953e-01 1.16089031e-01
7.29388714e-01 3.46033722e-01 -1.86955765e-01 8.80929157e-02
9.77316964e-03 5.41433096e-01 9.40636039e-01 -1.36048186e+00
2.54801242e-03 -3.46483558e-01 7.81985670e-02 -1.05588948e-02
-3.50584149e-01 -5.28622985e-01 -1.13342203e-01 -4.66815114e-01
7.81885386e-01 -9.07284379e-01 -3.68244126e-02 3.62137109e-01
-6.16944313e-01 -1.68473378e-01 7.47232080e-01 -3.85791212e-01
1.06124890e+00 -2.10450435e+00 2.82193482e-01 8.23943853e-01
4.49614972e-01 -1.18638456e-01 -1.92790896e-01 5.05369544e-01
-3.13779086e-01 1.77913591e-01 -4.47110146e-01 5.15592545e-02
-2.78875008e-02 -1.24439923e-02 -2.15674162e-01 8.49065363e-01
2.36821428e-01 7.95525789e-01 -6.76936567e-01 -1.97234616e-01
1.18828051e-01 5.87605715e-01 -7.68598974e-01 -2.06793979e-01
5.75799569e-02 6.24129951e-01 -2.97995150e-01 5.76425135e-01
9.69023883e-01 -3.20587993e-01 2.38156497e-01 -6.85425878e-01
-9.52427164e-02 3.66176307e-01 -1.66941237e+00 1.48783910e+00
-6.77236021e-01 4.41904783e-01 3.04870810e-02 -1.56971014e+00
1.15619504e+00 3.83008383e-02 8.93235087e-01 -6.19157910e-01
8.48149359e-02 4.81431454e-01 6.25247359e-02 -1.81904316e-01
3.99550013e-02 6.47799298e-02 9.47679505e-02 6.50664806e-01
-3.23911667e-01 -1.15397997e-01 -5.65694198e-02 -3.36694479e-01
1.00819194e+00 -5.45153767e-02 1.72426134e-01 -4.99774337e-01
8.39118302e-01 -4.53289360e-01 5.95473945e-01 4.23044235e-01
4.28413630e-01 8.95296812e-01 5.07473767e-01 -8.13668966e-01
-1.06600857e+00 -1.14877725e+00 -4.93427813e-01 4.28488165e-01
3.75479579e-01 -3.78120631e-01 -6.31276190e-01 -5.07103562e-01
8.64313096e-02 1.95935667e-01 -6.23571813e-01 -2.88110256e-01
-9.79404330e-01 -5.65797508e-01 2.74800003e-01 3.09712052e-01
6.45798445e-01 -8.66401732e-01 -4.09841865e-01 -9.81659070e-02
-1.80679470e-01 -7.87343442e-01 -5.72787881e-01 -1.84783667e-01
-9.70077813e-01 -1.22493219e+00 -7.52899826e-01 -8.33911717e-01
1.07523918e+00 1.82888508e-01 1.03424382e+00 2.61183888e-01
-4.64554906e-01 1.50589690e-01 1.24695122e-01 3.64385039e-01
-3.51012796e-01 1.14709154e-01 3.25868465e-02 3.92285854e-01
-1.44697338e-01 -1.00845122e+00 -7.52688289e-01 6.45232379e-01
-8.83811414e-01 1.45656345e-02 6.57464087e-01 1.05290592e+00
8.71754587e-01 3.61975610e-01 2.38432318e-01 -6.21648252e-01
3.18481028e-01 -1.60059676e-01 -4.88207012e-01 1.95072398e-01
-3.63924950e-01 4.04367954e-01 8.15295756e-01 -2.44911492e-01
-5.61632097e-01 2.75262594e-01 1.13499694e-01 -3.09022009e-01
1.77519377e-02 4.92523313e-01 -3.15889239e-01 -6.12051725e-01
5.90304971e-01 2.93683648e-01 2.62452394e-01 -5.60650170e-01
1.47124812e-01 3.73707086e-01 2.85575122e-01 -8.26088369e-01
1.00759673e+00 8.67140055e-01 6.08126163e-01 -1.05676329e+00
-5.40557444e-01 -2.65151858e-01 -7.73628652e-01 -9.64229181e-02
4.16714847e-01 -5.24408937e-01 -1.03708160e+00 6.31795108e-01
-1.18570888e+00 2.32917140e-03 -3.80236447e-01 2.10245043e-01
-6.37761533e-01 6.47601247e-01 -3.23882312e-01 -4.00578678e-01
-2.22876936e-01 -1.22848892e+00 1.15986049e+00 -2.51433760e-01
-1.24224506e-01 -9.48024750e-01 1.48138493e-01 1.74907908e-01
3.97950351e-01 4.32594091e-01 1.28887510e+00 -2.99673051e-01
-7.06080854e-01 -3.84562105e-01 -2.97223300e-01 3.14207643e-01
2.81460851e-01 1.00451000e-01 -4.89010245e-01 -3.98120821e-01
1.09824851e-01 -2.07374275e-01 8.95037651e-01 1.29169852e-01
1.52393866e+00 -3.18533778e-01 -2.17772543e-01 8.49574864e-01
1.27898514e+00 -3.06192845e-01 6.22545660e-01 2.55173653e-01
9.11757946e-01 5.56466997e-01 -1.90218892e-02 4.70049322e-01
2.51308829e-01 8.02977443e-01 4.05470282e-01 -1.79960534e-01
-5.09006791e-02 -1.92951918e-01 2.61975169e-01 8.42592597e-01
-3.80921066e-01 1.94785938e-01 -8.96510780e-01 1.18651450e-01
-1.65325677e+00 -9.17814553e-01 7.28496239e-02 2.44222879e+00
8.01170468e-01 -5.66822290e-03 2.38159243e-02 3.57184678e-01
8.01281631e-01 8.29962641e-02 -6.40393674e-01 -9.33166295e-02
-2.03470662e-01 4.62422550e-01 4.07588929e-01 4.99350727e-01
-9.79806483e-01 4.47858810e-01 5.48925543e+00 7.62426436e-01
-1.40338409e+00 -3.03483784e-01 4.07938272e-01 2.43485644e-01
-5.66531122e-01 -4.94249016e-02 -4.48454201e-01 2.00579241e-01
7.06973346e-03 9.32452530e-02 7.03954458e-01 6.31992757e-01
-3.42427455e-02 4.31546152e-01 -1.28760564e+00 1.06731248e+00
1.48866028e-01 -1.51958609e+00 2.24719480e-01 1.81197584e-01
1.00137675e+00 -2.76289321e-02 7.96087682e-02 -2.12484255e-01
-1.82898611e-01 -1.17814517e+00 5.97102880e-01 5.53710401e-01
1.01872182e+00 -6.92576587e-01 3.61154854e-01 2.23523423e-01
-1.32892644e+00 6.76557645e-02 -3.51310819e-01 -4.10529636e-02
-2.04750031e-01 7.42490768e-01 -4.30140108e-01 5.98220706e-01
5.38775206e-01 8.99784625e-01 -2.46412277e-01 9.72206831e-01
-9.79877189e-02 1.59790695e-01 -4.94717747e-01 1.60595447e-01
-2.12670892e-01 -6.51667714e-01 6.84333205e-01 5.96422434e-01
4.76734698e-01 -1.86526075e-01 3.82292360e-01 1.24090934e+00
-3.14107925e-01 2.20577851e-01 -6.30601227e-01 2.15060636e-01
2.37089768e-01 1.35276687e+00 -9.23153579e-01 8.30998942e-02
-2.28324816e-01 8.81057441e-01 3.86773109e-01 3.91185284e-01
-5.03304064e-01 -4.92997080e-01 6.43578887e-01 3.52881044e-01
4.85652387e-01 -4.09786791e-01 -5.00265002e-01 -1.34615338e+00
4.52470541e-01 -8.17749560e-01 -4.03404655e-03 -4.44228798e-01
-1.35641372e+00 4.20884490e-01 -2.27342442e-01 -1.48737419e+00
7.12328032e-02 -1.00294948e+00 -8.13855350e-01 5.68272948e-01
-1.22703123e+00 -1.00033510e+00 -3.08352858e-01 6.45493507e-01
1.08041450e-01 -1.14555709e-01 8.29807818e-01 3.38899791e-01
-3.48397613e-01 7.15579093e-01 3.45091254e-01 3.00154239e-01
5.17233312e-01 -1.10954809e+00 2.07190797e-01 4.04433787e-01
2.66722322e-01 6.93833768e-01 2.08778709e-01 -3.86688203e-01
-1.62597048e+00 -8.64418507e-01 6.94899082e-01 -1.06607772e-01
6.96498930e-01 -4.77459282e-01 -8.03517401e-01 2.40138590e-01
-3.55647087e-01 8.28442872e-02 4.28920507e-01 -3.72487493e-02
-6.53676331e-01 -3.33654761e-01 -1.06279457e+00 7.38029778e-01
1.29409039e+00 -6.03640318e-01 -1.37071609e-01 4.86674339e-01
3.61248165e-01 -3.11912090e-01 -9.70955610e-01 7.61631966e-01
6.26380205e-01 -1.00087464e+00 1.10966456e+00 -4.69062239e-01
6.04339719e-01 -3.83024931e-01 -2.17848271e-01 -1.20816684e+00
-3.84711593e-01 -6.84169710e-01 2.11512342e-01 8.75120521e-01
4.87998396e-01 -8.50570083e-01 9.39306617e-01 2.65159905e-01
1.31336702e-02 -1.05482161e+00 -1.14238358e+00 -8.08131218e-01
2.12880805e-01 -2.03549668e-01 6.79673493e-01 1.13033390e+00
-1.74190089e-01 4.42311130e-02 -1.33625716e-01 1.25988528e-01
5.94954908e-01 7.47043252e-01 7.05987930e-01 -1.61314726e+00
-2.40145758e-01 -8.12227666e-01 -6.43727720e-01 -1.08760417e+00
4.12332296e-01 -1.23979187e+00 -1.02106288e-01 -1.20535243e+00
9.99204963e-02 -5.62183857e-01 -1.67536542e-01 4.62571561e-01
4.30970371e-01 5.65448463e-01 -2.61573493e-01 2.95166135e-01
-2.34322488e-01 7.48643458e-01 1.63044643e+00 -3.77753437e-01
-1.90661281e-01 6.66850060e-02 -3.58825654e-01 9.97679353e-01
7.33410716e-01 -2.30346739e-01 -1.29062226e-02 -2.48829767e-01
6.70597553e-01 -1.92332238e-01 5.52343667e-01 -9.48088050e-01
1.93214640e-01 -5.37384488e-02 3.72042388e-01 -4.16724563e-01
1.64518774e-01 -8.77414823e-01 1.57532990e-01 4.70918149e-01
-2.72922486e-01 7.33441487e-02 -1.25272349e-01 4.20696944e-01
-1.10800080e-01 -1.28900051e-01 7.92059839e-01 5.09362742e-02
-9.77557451e-02 6.08721912e-01 -1.66611467e-02 2.25169688e-01
5.52135885e-01 -3.10922831e-01 -4.86030206e-02 -3.10072809e-01
-5.18997550e-01 -1.50991023e-01 8.47643733e-01 2.34689251e-01
8.44575226e-01 -1.64950871e+00 -6.52601480e-01 7.42842197e-01
2.20645946e-02 2.49074101e-01 1.80556804e-01 1.00669861e+00
-6.64618254e-01 -1.83850843e-02 -3.42893541e-01 -8.15931737e-01
-9.63321686e-01 2.31641173e-01 5.01132905e-01 -9.34184641e-02
-7.72894025e-01 4.37839508e-01 1.73110276e-01 -8.78432930e-01
5.77921085e-02 -1.77578792e-01 -4.61459309e-02 -5.74296415e-02
1.06313545e-02 5.74789762e-01 1.60369948e-01 -7.96923637e-01
-3.44068438e-01 1.27422440e+00 1.59907818e-01 1.18884049e-01
1.39888549e+00 2.59786218e-01 -6.26164734e-01 1.86785802e-01
1.68807590e+00 3.52969199e-01 -9.54482555e-01 -2.68772840e-01
-3.51428062e-01 -3.46941382e-01 -1.26914427e-01 -3.38578254e-01
-1.39402521e+00 8.44955206e-01 1.66701391e-01 3.59613597e-01
9.75583553e-01 -1.01239443e-01 8.79262924e-01 7.75795937e-01
1.94217771e-01 -8.58734131e-01 8.05907547e-02 5.36330283e-01
1.18117321e+00 -1.19715440e+00 6.18342385e-02 -5.88835716e-01
5.82875907e-02 1.56485438e+00 3.62427533e-01 -5.11122882e-01
1.02589905e+00 1.82691105e-02 -1.13499068e-01 -3.40774208e-01
-4.40251604e-02 2.56968796e-01 6.93769455e-01 2.21216947e-01
1.98131353e-01 6.54803291e-02 -1.28151163e-01 1.85226962e-01
-4.61913854e-01 -3.84303600e-01 3.22924167e-01 4.53554332e-01
-2.50707388e-01 -1.29624331e+00 -2.11730257e-01 3.68080765e-01
-1.08676113e-01 6.02161922e-02 -4.90566850e-01 7.28120983e-01
2.81656738e-02 4.28791851e-01 1.17928661e-01 -4.71592814e-01
3.78587455e-01 -2.28163376e-01 4.96189594e-01 -3.42005402e-01
-6.09865151e-02 -1.94763348e-01 -3.33912522e-01 -7.27213025e-01
-4.31562275e-01 -7.07363904e-01 -1.21144760e+00 -2.60682136e-01
-1.38976157e-01 -1.04654849e-01 5.50338387e-01 8.74093473e-01
5.51464856e-01 1.68920830e-01 1.13629484e+00 -1.00194240e+00
-8.56422424e-01 -4.62200016e-01 -5.10517418e-01 7.63964117e-01
2.55416989e-01 -8.52789700e-01 -6.16097689e-01 -3.75598907e-01] | [8.348637580871582, -2.709261894226074] |
96891e56-ee91-4350-9ade-6218296dcbe4 | hyperspectral-image-denoising-based-on-multi | 2104.02304 | null | https://arxiv.org/abs/2104.02304v1 | https://arxiv.org/pdf/2104.02304v1.pdf | Hyperspectral Image Denoising Based On Multi-Stream Denoising Network | Hyperspectral images (HSIs) have been widely applied in many fields, such as military, agriculture, and environment monitoring. Nevertheless, HSIs commonly suffer from various types of noise during acquisition. Therefore, denoising is critical for HSI analysis and applications. In this paper, we propose a novel blind denoising method for HSIs based on Multi-Stream Denoising Network (MSDNet). Our network consists of the noise estimation subnetwork and denoising subnetwork. In the noise estimation subnetwork, a multiscale fusion module is designed to capture the noise from different scales. Then, the denoising subnetwork is utilized to obtain the final denoising image. The proposed MSDNet can obtain robust noise level estimation, which is capable of improving the performance of HSI denoising. Extensive experiments on HSI dataset demonstrate that the proposed method outperforms four closely related methods. | ['Junyu Dong', 'Feng Gao', 'Yan Gao'] | 2021-04-06 | null | null | null | null | ['noise-estimation'] | ['medical'] | [ 5.61888993e-01 -1.00519788e+00 5.61802030e-01 -2.41343319e-01
-6.08619273e-01 -3.11612874e-01 1.25397503e-01 -3.13175581e-02
-1.41701356e-01 3.90743673e-01 1.01513453e-01 -8.60404149e-02
-1.80998668e-01 -1.05514419e+00 -2.95129847e-02 -1.35813010e+00
4.67595071e-01 -6.73480034e-01 3.25272590e-01 -3.22140634e-01
-1.03662767e-01 4.22267079e-01 -1.26823795e+00 -1.41563192e-01
1.19768786e+00 1.39436769e+00 3.32259893e-01 2.27197140e-01
5.86947277e-02 2.18058795e-01 -4.48513031e-01 2.21808314e-01
3.89923275e-01 -6.08925760e-01 2.87259370e-01 3.50122094e-01
1.06943846e-02 -6.08862102e-01 -5.13280094e-01 1.63920331e+00
6.69046938e-01 5.14830351e-01 4.22699898e-01 -9.04928088e-01
-2.17534646e-01 3.44953656e-01 -9.58776057e-01 -1.56071493e-02
-2.23763704e-01 3.85351449e-01 4.35463637e-01 -6.51226580e-01
-1.22905821e-01 1.20621300e+00 7.11518824e-01 -2.44347118e-02
-1.11388493e+00 -8.71104896e-01 9.36643966e-03 2.39737630e-01
-1.34909356e+00 -5.67161858e-01 9.26746488e-01 -8.24438259e-02
-8.91033933e-02 2.09343940e-01 5.04746318e-01 5.95603943e-01
-5.21383509e-02 5.40526509e-01 1.26459181e+00 -1.18664727e-01
8.51487294e-02 -5.04669666e-01 4.87178341e-02 2.27841467e-01
6.17266238e-01 -1.50988335e-02 -1.97701335e-01 5.07120863e-02
8.46376717e-01 3.69666159e-01 -6.34270251e-01 2.02864692e-01
-9.25479054e-01 4.58765209e-01 7.85162091e-01 3.19455534e-01
-5.56440413e-01 -1.11285701e-01 4.09338564e-01 1.64183050e-01
5.34488380e-01 -3.26552600e-01 -2.23223343e-02 4.63175565e-01
-9.78568494e-01 -1.04471438e-01 3.64115596e-01 5.56201339e-01
7.99481153e-01 3.31937343e-01 -1.30002603e-01 9.58319783e-01
4.42530751e-01 7.96344459e-01 2.65969008e-01 -1.02229416e+00
2.68485963e-01 3.78926337e-01 -4.14508721e-03 -1.46513665e+00
-2.70144105e-01 -4.94064093e-01 -1.86536646e+00 9.88752097e-02
-9.31407958e-02 -1.44987807e-01 -7.37460256e-01 1.37754595e+00
2.89959371e-01 4.55759853e-01 1.51301250e-01 1.05744231e+00
8.79733205e-01 1.07344103e+00 1.82657242e-01 -5.21976173e-01
1.06371343e+00 -5.67616522e-01 -9.07451153e-01 -2.92208344e-01
-2.51868099e-01 -8.31396163e-01 6.35149777e-01 6.07355416e-01
-7.87421823e-01 -6.49789453e-01 -1.10629642e+00 4.39000763e-02
2.32786648e-02 4.26703185e-01 3.09669226e-01 5.07665396e-01
-7.17249632e-01 5.38595498e-01 -8.82185757e-01 -3.74636590e-01
4.68051344e-01 7.01691769e-03 -1.01587087e-01 -4.64355111e-01
-1.00210881e+00 3.70566219e-01 3.61275196e-01 9.95262504e-01
-8.26776087e-01 -2.30008662e-01 -7.61601329e-01 2.48444527e-01
5.07949114e-01 -4.94420499e-01 8.61971855e-01 -6.70372665e-01
-1.21514392e+00 1.51561290e-01 -3.66059512e-01 4.89685349e-02
3.42131883e-01 5.49956635e-02 -7.00575769e-01 3.77827168e-01
2.23128110e-01 -7.77703524e-02 1.04897714e+00 -1.27435195e+00
-4.85655338e-01 -5.56410491e-01 -4.81024057e-01 2.32593417e-01
-5.77570379e-01 -2.12850809e-01 -3.59920502e-01 -8.09333265e-01
9.45449173e-01 -2.13449389e-01 -3.05715322e-01 9.13578570e-02
-5.40801525e-01 3.13006073e-01 9.54784513e-01 -9.95207727e-01
1.05080986e+00 -2.64589095e+00 -1.59692261e-02 4.56405371e-01
1.84809864e-01 3.14440370e-01 -2.60037631e-01 1.86636910e-01
-4.52644378e-02 8.06730911e-02 -6.46935046e-01 -5.86781278e-02
-2.72680551e-01 7.82686770e-02 -9.43210348e-02 6.09618843e-01
-2.54116226e-02 2.47875676e-01 -7.21965611e-01 -3.73080134e-01
3.23582947e-01 4.73573089e-01 1.70275673e-01 3.81508619e-01
1.63945422e-01 5.62096834e-01 -5.50250709e-01 8.40390205e-01
1.27287209e+00 1.96735054e-01 -2.68971957e-02 -9.00034487e-01
-3.99963111e-01 -4.98281330e-01 -1.46646166e+00 1.33182979e+00
-9.20351446e-02 3.13466340e-01 9.31397617e-01 -1.09936380e+00
1.05720568e+00 2.16971427e-01 3.29555690e-01 -4.24382091e-01
3.78290355e-01 3.76644045e-01 -1.70710742e-01 -6.53615534e-01
1.48023576e-01 -4.28257734e-01 3.78042132e-01 -3.61752473e-02
-3.06594461e-01 -2.44919285e-01 2.39073128e-01 4.32574889e-03
8.74410927e-01 -1.73891202e-01 2.07456037e-01 -1.47102177e-01
9.92056847e-01 -1.03561625e-01 9.13154960e-01 5.50211012e-01
-4.42604631e-01 4.97518480e-01 1.77626908e-01 1.79020204e-02
-6.85411215e-01 -7.84890532e-01 6.93058316e-03 6.45038545e-01
5.04970670e-01 1.37797073e-01 -5.30672789e-01 -7.72969946e-02
-1.68913230e-01 2.77552277e-01 -1.26207739e-01 -3.27716172e-01
-2.64095366e-01 -9.80366111e-01 5.02246141e-01 2.29234159e-01
1.31198394e+00 -8.39187980e-01 1.12983331e-01 3.15481395e-01
-5.32444298e-01 -9.21179354e-01 -3.59057873e-01 -2.55632401e-02
-1.03587067e+00 -1.05645943e+00 -6.01152718e-01 -7.48353004e-01
5.54911852e-01 1.18221521e+00 4.08363193e-01 1.75149933e-01
-1.48602903e-01 -1.51274551e-03 -2.16148466e-01 -2.69937247e-01
-2.20257506e-01 -1.61592484e-01 -8.56274813e-02 3.43716145e-01
3.43545452e-02 -9.61468875e-01 -8.84428620e-01 2.96604007e-01
-1.49232340e+00 -1.10876873e-01 8.28394055e-01 8.45489025e-01
6.00735426e-01 1.05350721e+00 3.29646707e-01 -5.85194588e-01
6.35506690e-01 -4.79265779e-01 -8.30998123e-01 2.01675504e-01
-2.94385672e-01 -5.98103106e-01 8.00179064e-01 -1.47442445e-01
-1.46889424e+00 -3.83213013e-02 -1.36213303e-01 -1.55427799e-01
-4.35054302e-01 8.50099862e-01 -6.72740221e-01 -3.36572379e-01
3.31979156e-01 4.75677401e-01 1.52641714e-01 -7.07422376e-01
-1.53115407e-01 7.52797842e-01 8.74163389e-01 -1.56639516e-01
1.22899783e+00 4.50851351e-01 3.34173948e-01 -1.18540967e+00
-7.25407958e-01 -5.89366257e-01 -1.93041936e-01 -3.55799019e-01
5.54169536e-01 -1.21758509e+00 -9.04554665e-01 1.03116715e+00
-1.04730582e+00 4.36329991e-02 3.11860740e-01 4.40825939e-01
1.87390700e-01 6.10793531e-01 -7.62497783e-01 -9.39930439e-01
-4.20079678e-01 -9.98820722e-01 7.49352872e-01 7.03817368e-01
6.76923811e-01 -7.84033597e-01 -3.66676599e-01 2.82977045e-01
6.61800086e-01 3.59607130e-01 7.33707130e-01 -2.63672769e-02
-5.07220685e-01 -1.48031861e-01 -6.67712629e-01 8.28523040e-01
4.96995866e-01 -2.29081344e-02 -9.17624474e-01 -2.93957055e-01
5.34955025e-01 -4.37400155e-02 1.14755130e+00 5.64147890e-01
1.29260314e+00 3.44733596e-02 2.88774446e-02 9.03934360e-01
1.97596383e+00 2.77019024e-01 6.86496675e-01 3.71734470e-01
7.26784289e-01 5.73079765e-01 5.66717744e-01 4.42310691e-01
-1.98269263e-01 -2.09104493e-01 6.24396801e-01 -4.14498210e-01
1.47487342e-01 1.59305170e-01 3.07151109e-01 9.18870151e-01
2.42357682e-02 -3.00637186e-01 -6.49109304e-01 2.50177294e-01
-1.71703899e+00 -8.40327561e-01 -4.52818722e-01 1.80026352e+00
6.54304266e-01 -1.73462152e-01 -4.10223842e-01 3.78621072e-01
1.03261244e+00 4.51103956e-01 -8.34530175e-01 4.88017678e-01
-4.30744261e-01 1.62079468e-01 4.18752342e-01 1.57221839e-01
-1.05812347e+00 5.16875863e-01 5.27072573e+00 8.59881163e-01
-9.85504448e-01 6.83258241e-03 3.94642711e-01 4.90806639e-01
-3.56976911e-02 -9.58459005e-02 -2.56524801e-01 3.64014983e-01
1.21063463e-01 -1.02547869e-01 6.39324546e-01 3.52013946e-01
9.20727909e-01 -4.78581309e-01 -1.41695201e-01 9.44430113e-01
-1.49399832e-01 -6.77895606e-01 3.66890877e-02 -2.79316306e-01
5.90455294e-01 -3.25322360e-01 -1.28166974e-01 -1.63717300e-01
-1.40402466e-02 -5.49999118e-01 2.21196756e-01 6.42647147e-01
6.75781250e-01 -7.71978676e-01 1.04202163e+00 4.56925184e-01
-1.37027252e+00 -5.59218973e-02 -5.46760201e-01 1.12550594e-01
1.01306111e-01 1.27574098e+00 1.86690062e-01 1.07943451e+00
7.23075151e-01 1.13539374e+00 -3.96384180e-01 1.24449599e+00
-4.42570895e-01 8.29032540e-01 -2.41556674e-01 5.01228154e-01
1.01647735e-01 -9.44503784e-01 5.87898314e-01 8.44634652e-01
6.85140312e-01 7.79913902e-01 3.49055201e-01 6.41599357e-01
-8.31580535e-02 -3.97898331e-02 -2.07522690e-01 8.28307569e-02
4.24586505e-01 1.52247667e+00 -6.61003113e-01 -2.21570268e-01
-2.36978114e-01 8.72421145e-01 -4.74595785e-01 6.82657123e-01
-4.99500364e-01 -7.21382856e-01 5.93559146e-01 -2.91951478e-01
1.24379300e-01 -2.54760742e-01 -3.17313135e-01 -1.26380455e+00
-5.55590019e-02 -1.01484191e+00 2.62309432e-01 -1.01688385e+00
-1.41904676e+00 3.41437191e-01 -3.83047163e-01 -1.31499445e+00
7.90989220e-01 -5.36761940e-01 -9.41666305e-01 1.11603081e+00
-1.91293073e+00 -9.69844699e-01 -1.17606628e+00 7.07855761e-01
2.64446318e-01 8.97492766e-02 4.35324967e-01 3.49364728e-01
-1.15521049e+00 -1.22474112e-01 5.40199459e-01 9.36952829e-02
8.24430883e-01 -6.92057073e-01 -2.62880087e-01 1.41866004e+00
-7.16730654e-01 4.85062212e-01 5.70917130e-01 -7.62092948e-01
-1.17827928e+00 -1.33395147e+00 2.34263316e-02 1.01658821e+00
5.30329764e-01 2.71148175e-01 -1.13236010e+00 2.61185020e-01
1.10130146e-01 9.89264548e-02 3.69765729e-01 -4.47476357e-01
-1.68243065e-01 -8.84822547e-01 -1.26065099e+00 4.62762356e-01
5.99629164e-01 -2.82767981e-01 -6.36678189e-02 2.41707176e-01
5.16645312e-01 -2.08103791e-01 -7.51163721e-01 5.60931563e-01
3.13424587e-01 -1.16950297e+00 8.89404595e-01 1.93332642e-01
4.44988847e-01 -7.15644538e-01 -2.00415269e-01 -1.41479993e+00
-4.96887892e-01 -2.84518212e-01 3.12060326e-01 1.53697526e+00
-3.28336030e-01 -7.70234346e-01 3.64154816e-01 3.93338613e-02
-1.80503130e-01 -3.30172963e-02 -4.88538682e-01 -5.96429586e-01
-5.45242190e-01 -2.38717243e-01 5.74638188e-01 7.07195103e-01
-7.16192484e-01 2.58684814e-01 -2.01183647e-01 7.66612709e-01
1.32313681e+00 1.69898774e-02 4.34749454e-01 -1.30210304e+00
1.37335643e-01 -4.04381126e-01 -1.25341251e-01 -9.38632667e-01
-7.28988089e-03 -3.71542305e-01 5.07534027e-01 -1.86920691e+00
1.80579588e-01 -1.81235433e-01 -4.25039798e-01 3.79349023e-01
-5.69816589e-01 4.26936179e-01 -9.62328166e-02 4.35026169e-01
-2.46217683e-01 7.31090963e-01 1.20852530e+00 -3.45826626e-01
-1.04899257e-02 1.75850419e-03 -7.98218429e-01 7.92378843e-01
1.11855173e+00 -3.13483030e-01 -1.18583098e-01 -6.39709294e-01
-2.65725285e-01 1.94977716e-01 5.09917319e-01 -1.19047141e+00
4.28189874e-01 -3.02887291e-01 6.38420105e-01 -6.47414863e-01
1.39630780e-01 -1.10440767e+00 1.77636102e-01 5.05150020e-01
9.32503641e-02 -5.08710027e-01 2.15721920e-01 6.96060896e-01
-5.27253270e-01 -6.62159696e-02 9.52665031e-01 -1.93920508e-01
-7.32185245e-01 4.99505728e-01 -3.95674497e-01 -4.58475381e-01
6.33305013e-01 -5.72166825e-03 -5.20918608e-01 -4.14364874e-01
-5.77087164e-01 4.26617146e-01 2.49458313e-01 -3.16275567e-01
8.55380356e-01 -1.20895076e+00 -7.91982353e-01 3.34715754e-01
-7.09586069e-02 3.45901340e-01 5.91995716e-01 9.32831347e-01
-5.66895545e-01 -3.62386793e-01 -1.71314999e-02 -4.65572119e-01
-1.20054865e+00 6.29743189e-02 2.76509821e-01 -3.14833745e-02
-2.37161726e-01 5.01021683e-01 -1.17386654e-01 -2.42726445e-01
1.12872928e-01 -9.46135521e-02 -2.36254439e-01 6.77145869e-02
5.72539330e-01 8.82680357e-01 4.77737598e-02 -5.86276114e-01
-4.52670082e-02 5.32271743e-01 4.44743454e-01 -3.67575549e-02
1.47277093e+00 -4.94382203e-01 -9.82695043e-01 3.42134118e-01
1.02626848e+00 -2.04922333e-01 -1.02152789e+00 -4.48357940e-01
-3.75984460e-01 -4.98392999e-01 5.55289686e-01 -4.64603990e-01
-1.35280991e+00 8.98050070e-01 6.72868133e-01 3.57965797e-01
1.90974188e+00 -7.88774729e-01 9.86467302e-01 3.59622002e-01
4.81162854e-02 -8.13834608e-01 -1.39485538e-01 2.14738145e-01
5.39910018e-01 -1.30774462e+00 1.64732799e-01 -6.14960790e-01
-1.23016238e-01 1.23015273e+00 5.86896718e-01 1.21503521e-03
6.49178028e-01 9.29179788e-02 2.15381920e-01 8.21893141e-02
1.76025219e-02 -2.89883763e-01 -6.25824258e-02 4.34031755e-01
1.01801306e-01 -1.46216795e-01 -1.96908399e-01 6.46903276e-01
4.90651757e-01 -7.83522874e-02 3.90475214e-01 8.11082423e-01
-7.68554747e-01 -7.45964646e-01 -9.89168704e-01 5.55852830e-01
-1.30976915e-01 -2.31822014e-01 2.74805613e-02 2.61399955e-01
2.45237023e-01 1.58077633e+00 -3.78589302e-01 -3.82801890e-01
5.10331094e-01 -3.58912051e-01 -9.83188488e-03 -3.61000210e-01
-3.88875782e-01 5.91709554e-01 -3.98884088e-01 -3.92345071e-01
-7.48114884e-01 -5.64514041e-01 -7.73674905e-01 -4.88443881e-01
-5.03115356e-01 1.89518854e-01 5.61876535e-01 7.86280751e-01
-1.64661884e-01 6.45954967e-01 9.55663383e-01 -9.07413781e-01
-5.63144565e-01 -9.16624784e-01 -1.26905727e+00 1.66960731e-01
5.77401221e-01 -2.66838670e-01 -7.31956184e-01 8.32404718e-02] | [10.522424697875977, -2.06819748878479] |
56fd7b6f-8204-4716-80fe-1c140ccd2f6a | zero-shot-grounding-of-objects-from-natural | 1908.07129 | null | https://arxiv.org/abs/1908.07129v1 | https://arxiv.org/pdf/1908.07129v1.pdf | Zero-Shot Grounding of Objects from Natural Language Queries | A phrase grounding system localizes a particular object in an image referred to by a natural language query. In previous work, the phrases were restricted to have nouns that were encountered in training, we extend the task to Zero-Shot Grounding(ZSG) which can include novel, "unseen" nouns. Current phrase grounding systems use an explicit object detection network in a 2-stage framework where one stage generates sparse proposals and the other stage evaluates them. In the ZSG setting, generating appropriate proposals itself becomes an obstacle as the proposal generator is trained on the entities common in the detection and grounding datasets. We propose a new single-stage model called ZSGNet which combines the detector network and the grounding system and predicts classification scores and regression parameters. Evaluation of ZSG system brings additional subtleties due to the influence of the relationship between the query and learned categories; we define four distinct conditions that incorporate different levels of difficulty. We also introduce new datasets, sub-sampled from Flickr30k Entities and Visual Genome, that enable evaluations for the four conditions. Our experiments show that ZSGNet achieves state-of-the-art performance on Flickr30k and ReferIt under the usual "seen" settings and performs significantly better than baseline in the zero-shot setting. | ['Kan Chen', 'Arka Sadhu', 'Ram Nevatia'] | 2019-08-20 | zero-shot-grounding-of-objects-from-natural-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Sadhu_Zero-Shot_Grounding_of_Objects_From_Natural_Language_Queries_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Sadhu_Zero-Shot_Grounding_of_Objects_From_Natural_Language_Queries_ICCV_2019_paper.pdf | iccv-2019-10 | ['phrase-grounding'] | ['natural-language-processing'] | [ 1.07797667e-01 2.95189232e-01 -1.21842146e-01 -2.91496068e-01
-1.14995289e+00 -5.85750222e-01 8.82795990e-01 9.86797661e-02
-5.71924686e-01 4.17408586e-01 1.83417663e-01 1.09102398e-01
2.45557740e-01 -7.43819475e-01 -8.57186437e-01 -5.39692581e-01
1.22383587e-01 5.43276906e-01 6.89488292e-01 -2.80519098e-01
-7.36029120e-03 2.44815554e-02 -1.76271832e+00 5.65976083e-01
5.39490998e-01 7.87523210e-01 3.91343147e-01 8.52239370e-01
-4.41244729e-02 3.65282714e-01 -7.27200329e-01 -6.97123051e-01
4.32803452e-01 -1.74073547e-01 -7.65206397e-01 6.74309283e-02
9.41468120e-01 -2.77695149e-01 -4.73188639e-01 1.15330815e+00
6.21085584e-01 3.03076386e-01 5.48950672e-01 -1.38035655e+00
-9.91765618e-01 7.27727890e-01 -4.51163620e-01 2.98277408e-01
5.82903147e-01 2.12822109e-01 1.45258439e+00 -1.53167915e+00
9.01897728e-01 1.32624519e+00 5.43602049e-01 6.84010386e-01
-1.34445381e+00 -8.50576043e-01 4.06863540e-01 1.80201933e-01
-1.88923419e+00 -2.67912537e-01 2.45803654e-01 -4.10803050e-01
1.16095018e+00 1.53460936e-03 4.37480271e-01 1.30851138e+00
-4.19439711e-02 8.86563480e-01 6.31162703e-01 -5.27225494e-01
2.07621127e-01 2.96551704e-01 2.44590446e-01 5.47603607e-01
2.72280514e-01 8.39134231e-02 -6.93642795e-01 -2.39800066e-01
5.10964096e-01 -1.82401091e-01 -3.09363902e-01 -5.71310282e-01
-1.48793161e+00 8.79238188e-01 7.37820566e-01 7.84203485e-02
-2.99971730e-01 2.47008204e-01 1.08194642e-03 1.42299887e-02
3.94596875e-01 6.82730973e-01 -2.69604623e-01 3.94649684e-01
-1.05116057e+00 5.96564651e-01 6.85139418e-01 1.46884143e+00
9.69170094e-01 -4.36805040e-01 -7.77767539e-01 5.30496955e-01
3.66428763e-01 5.62049806e-01 3.62395108e-01 -4.31144029e-01
6.41180873e-01 5.85830688e-01 4.23514217e-01 -9.74285543e-01
-1.71369329e-01 -5.90340018e-01 -2.17032477e-01 -2.57082462e-01
2.68369131e-02 3.21863629e-02 -1.42590368e+00 1.69039893e+00
4.73409057e-01 4.87565577e-01 2.04589553e-02 1.02758038e+00
1.30114293e+00 7.48140275e-01 4.42055434e-01 1.33248448e-01
1.53330052e+00 -1.12788522e+00 -6.14289522e-01 -3.71578068e-01
6.09433830e-01 -5.37038803e-01 1.40488422e+00 5.40334135e-02
-7.82532990e-01 -5.42750895e-01 -1.02314138e+00 -1.62772030e-01
-8.13007951e-01 3.57746512e-01 3.23743790e-01 1.57777101e-01
-1.09068918e+00 2.22929731e-01 -4.07418162e-01 -8.83537233e-01
1.71575919e-01 3.69472504e-01 -2.56040633e-01 -6.01785332e-02
-1.32528365e+00 8.77675891e-01 6.79585576e-01 -8.81058052e-02
-1.18955612e+00 -4.87793028e-01 -8.79265606e-01 1.51089415e-01
8.06335092e-01 -7.91000009e-01 1.40844452e+00 -5.46245575e-01
-7.53261447e-01 1.19209230e+00 -1.78500544e-02 -4.91253793e-01
2.40817115e-01 -4.15459275e-01 -5.19552886e-01 1.74947411e-01
5.76581001e-01 1.17074251e+00 1.05958641e+00 -1.19839942e+00
-9.11373496e-01 -4.08555232e-02 4.02994722e-01 3.31068516e-01
-3.34947333e-02 1.38516262e-01 -7.70919144e-01 -7.44859636e-01
1.09494179e-01 -1.08323896e+00 -4.73164162e-03 -1.00892924e-01
-6.58309340e-01 -5.09848416e-01 6.48753226e-01 -1.94258243e-01
1.30675948e+00 -2.22274613e+00 9.68474615e-03 1.00815937e-01
2.01763019e-01 2.88704365e-01 -3.08581531e-01 5.10771692e-01
-1.12275332e-02 2.23649368e-01 1.49432778e-01 -4.00497794e-01
1.89839423e-01 1.75982058e-01 -5.68457305e-01 5.78367114e-02
5.98566771e-01 1.06618607e+00 -1.19730735e+00 -5.76156259e-01
-1.03223570e-01 1.89686999e-01 -5.49917579e-01 2.33157262e-01
-4.23475534e-01 -2.40746483e-01 -4.44800645e-01 6.53430164e-01
2.58141607e-01 -5.48585713e-01 -3.02941769e-01 -3.58770102e-01
6.24491274e-02 4.80654202e-02 -1.45496142e+00 1.76608026e+00
-1.41681209e-02 3.60118657e-01 -2.49065429e-01 -3.31091344e-01
7.04632938e-01 3.93490225e-01 -1.56794071e-01 -2.76235104e-01
-2.84077190e-02 9.78760198e-02 -2.35479489e-01 -4.36841965e-01
9.55585301e-01 4.17061150e-02 -3.56201977e-01 2.26720855e-01
6.75546348e-01 -1.78003639e-01 4.30286914e-01 7.21404254e-01
1.19530642e+00 5.68632036e-02 2.69862235e-01 -1.53738156e-01
6.25089407e-02 3.45604390e-01 3.93804282e-01 1.36819577e+00
-1.81542784e-01 9.03379500e-01 1.37833104e-01 -2.96125174e-01
-9.36688840e-01 -1.16133177e+00 8.13944638e-03 1.49782932e+00
6.00332618e-01 -7.54475892e-01 -4.56043333e-01 -1.02876949e+00
9.92847979e-02 7.88959324e-01 -8.59106660e-01 -1.57704532e-01
-2.52860129e-01 -5.99007547e-01 3.86065602e-01 5.27863741e-01
4.19842690e-01 -1.14386249e+00 -4.35531050e-01 8.83242264e-02
-2.36203462e-01 -1.32357132e+00 -5.19123852e-01 1.39670625e-01
-2.62151033e-01 -1.01795924e+00 -7.39778936e-01 -8.09159338e-01
6.49988115e-01 5.38575172e-01 1.39030981e+00 1.76359974e-02
-4.20951724e-01 5.65119743e-01 -6.15076780e-01 -5.91074467e-01
-1.51556328e-01 3.15282762e-01 -1.25547305e-01 1.50871426e-01
7.71779716e-01 7.13408515e-02 -6.98655963e-01 3.12525034e-01
-9.25843775e-01 5.10991886e-02 7.17004299e-01 7.88428247e-01
7.22198427e-01 -1.31651834e-01 4.84373063e-01 -7.67966926e-01
6.03180349e-01 -5.75909555e-01 -3.92366260e-01 4.25389707e-01
-4.03871506e-01 6.60237893e-02 4.37104180e-02 -6.17278099e-01
-7.77561426e-01 1.83201879e-01 1.33391008e-01 -5.48372686e-01
-1.26348421e-01 2.90682852e-01 -4.29575332e-02 1.25951037e-01
1.00243092e+00 2.59435922e-02 -6.59427881e-01 -2.81555712e-01
6.19435489e-01 5.52322924e-01 6.50218487e-01 -4.28417981e-01
1.06465113e+00 3.80522639e-01 -5.47938406e-01 -7.26254165e-01
-1.13928616e+00 -1.05795085e+00 -3.92426640e-01 -1.34730801e-01
9.45576668e-01 -1.33365703e+00 -6.67822920e-03 4.70310971e-02
-1.20705545e+00 -7.45873898e-02 -6.82020128e-01 4.07394260e-01
-3.13553035e-01 -1.21188730e-01 -3.88350695e-01 -6.35210097e-01
-3.85695815e-01 -9.70760107e-01 1.70602906e+00 1.59058943e-01
-1.64409697e-01 -3.56623977e-01 4.34726179e-02 -1.71021417e-01
1.67859882e-01 -2.94240676e-02 6.17547333e-01 -9.68223631e-01
-7.66529381e-01 -3.02405328e-01 -4.09193605e-01 4.46263328e-03
-2.87747264e-01 -9.09873396e-02 -1.06225622e+00 -3.57475519e-01
-5.13537109e-01 -5.92503965e-01 1.12862575e+00 3.37452851e-02
6.97054803e-01 -2.62892902e-01 -6.26379251e-01 3.12358409e-01
1.46467125e+00 -1.31771117e-01 5.44336259e-01 2.17151642e-01
6.19644880e-01 4.23645020e-01 8.17214131e-01 1.26383632e-01
3.35470498e-01 8.69242549e-01 4.87738460e-01 -2.85219967e-01
-1.55005127e-01 -5.71775019e-01 1.29069850e-01 2.00679258e-01
2.02528059e-01 -5.94532251e-01 -9.00431693e-01 8.77441764e-01
-1.97320700e+00 -9.70623672e-01 1.15864150e-01 1.88228583e+00
7.95144439e-01 7.13284165e-02 5.66368401e-02 -2.96148777e-01
9.28337634e-01 4.89890203e-02 -4.37266797e-01 6.03687502e-02
-1.40486762e-01 1.34631112e-01 2.48150423e-01 3.88372928e-01
-1.37748349e+00 1.38890564e+00 6.14673281e+00 7.58999527e-01
-8.02412093e-01 2.38070190e-01 2.29566246e-01 -9.19432193e-02
-2.36019567e-01 1.35963345e-02 -1.39789367e+00 1.25784442e-01
5.09191751e-01 -1.04995206e-01 -9.62192863e-02 9.62504029e-01
-1.14846401e-01 6.99432939e-03 -1.31579936e+00 9.35134351e-01
1.90453306e-01 -9.84677374e-01 3.83045435e-01 -1.30902827e-01
7.05950260e-01 3.67960393e-01 7.08743557e-02 6.37748957e-01
5.66657484e-01 -8.16174567e-01 8.66404116e-01 4.41171050e-01
7.98264623e-01 -2.41805092e-01 7.30711639e-01 2.78131783e-01
-1.23656094e+00 -5.87995127e-02 -5.26996553e-01 2.47312590e-01
1.39555171e-01 1.68327942e-01 -1.27112854e+00 3.51728022e-01
7.89487422e-01 4.16473448e-01 -8.11374784e-01 1.22895443e+00
-5.94027400e-01 4.56721872e-01 -3.18272352e-01 -1.95971504e-01
4.78615373e-01 5.72581530e-01 7.70910323e-01 1.41791904e+00
3.46187800e-01 2.62321025e-01 4.57777053e-01 1.01047564e+00
-8.96907672e-02 1.71694010e-01 -4.87136394e-01 -3.82425077e-02
6.08481705e-01 1.42351413e+00 -9.39297974e-01 -5.28514028e-01
-4.23271507e-01 8.98039162e-01 3.31965566e-01 6.40970349e-01
-7.42449701e-01 -4.16022032e-01 3.28840137e-01 4.14184690e-01
6.28183186e-01 -6.40509883e-03 4.60646421e-01 -1.43226254e+00
-3.80888768e-02 -6.53353691e-01 5.31504631e-01 -1.13021410e+00
-1.29068363e+00 6.68510079e-01 4.14113611e-01 -1.02021241e+00
-1.73193574e-01 -5.28609872e-01 -2.92890996e-01 7.73428440e-01
-1.41641343e+00 -1.32382476e+00 -5.79187751e-01 4.46783334e-01
6.07537746e-01 6.71553388e-02 8.75611603e-01 1.23231754e-01
-3.79982740e-01 4.73922670e-01 -4.71090764e-01 2.75944173e-01
7.85802722e-01 -1.40327787e+00 7.12362409e-01 1.02691698e+00
7.06184804e-01 7.47934699e-01 7.70508766e-01 -7.37079024e-01
-8.60207736e-01 -1.44734550e+00 1.19618797e+00 -7.43935406e-01
6.36565745e-01 -8.31352293e-01 -8.51226926e-01 8.30421269e-01
1.34318069e-01 3.06795418e-01 5.17975926e-01 1.23508394e-01
-5.12864888e-01 1.89234987e-01 -9.20927882e-01 6.06544793e-01
1.17807519e+00 -5.31074345e-01 -9.73978996e-01 4.67698365e-01
1.05177546e+00 -4.10025686e-01 -4.00475413e-01 3.94479662e-01
3.92776251e-01 -4.75651681e-01 9.89250600e-01 -7.17707992e-01
1.49417877e-01 -5.97588003e-01 -3.55496824e-01 -1.17548430e+00
-5.73221445e-01 -4.07214493e-01 -2.61513621e-01 1.10652518e+00
6.70148075e-01 -1.79827422e-01 7.72078872e-01 5.22426903e-01
-3.58817130e-02 -5.89589000e-01 -6.29446864e-01 -7.01049805e-01
-4.50112015e-01 -2.90795147e-01 6.74224854e-01 6.75449371e-01
-1.81286126e-01 8.03693354e-01 -1.02353796e-01 4.56216127e-01
3.87073547e-01 -3.45726982e-02 8.11139047e-01 -1.12951040e+00
-4.03108656e-01 1.10373087e-01 -5.82658708e-01 -1.20909619e+00
1.25709809e-02 -8.13285291e-01 5.06052852e-01 -1.60520232e+00
4.90449876e-01 -3.80062431e-01 -3.62965018e-01 6.24594510e-01
-5.15215337e-01 4.83306020e-01 3.83126795e-01 3.93025815e-01
-1.06393373e+00 4.15309370e-01 8.81424427e-01 -3.43414515e-01
-2.40608096e-01 -1.25587508e-01 -7.12399244e-01 6.53998315e-01
2.13734344e-01 -5.30096412e-01 -5.21563232e-01 -6.04388297e-01
5.06761193e-01 -3.78086805e-01 6.63271308e-01 -9.44249749e-01
4.03028816e-01 1.78367913e-01 1.37873933e-01 -8.65076900e-01
3.86651993e-01 -7.26893306e-01 -8.56429115e-02 2.69919280e-02
-4.97671694e-01 3.64794917e-02 3.23007442e-02 9.43480074e-01
-1.63720444e-01 -2.19437718e-01 4.59070563e-01 -2.77679801e-01
-1.22906363e+00 3.58629525e-01 3.74938101e-02 2.61608094e-01
1.19563389e+00 -2.51388073e-01 -3.33352596e-01 -3.09823692e-01
-1.05148029e+00 3.53163987e-01 2.35255122e-01 6.04579687e-01
7.13836193e-01 -1.27350628e+00 -7.52411664e-01 6.23811893e-02
7.55218744e-01 2.28455573e-01 -1.35639742e-01 4.45261002e-01
-8.02008212e-02 3.40524673e-01 4.68578190e-01 -8.82776618e-01
-1.01301754e+00 7.13399529e-01 2.29563534e-01 -1.81407139e-01
-5.70103705e-01 1.43205833e+00 6.16790652e-01 -1.88328803e-01
3.97815734e-01 -4.57598746e-01 -1.99555054e-01 2.23740414e-01
5.07141769e-01 -6.63928762e-02 1.28063619e-01 -8.38932633e-01
-2.57564873e-01 3.21007580e-01 -4.58755314e-01 -1.49872437e-01
8.61830294e-01 -1.14730179e-01 4.46727127e-01 4.33467418e-01
8.32869411e-01 -2.82766670e-01 -1.04338431e+00 -5.46092033e-01
1.67015538e-01 -2.63116270e-01 -2.62512881e-02 -7.14611411e-01
-5.15925586e-01 6.53293908e-01 6.39653623e-01 1.94963470e-01
7.95384169e-01 5.68474889e-01 5.23307025e-01 5.73090672e-01
6.05906129e-01 -9.96893942e-01 3.48851949e-01 4.04147148e-01
8.61076236e-01 -1.25927389e+00 -1.34402037e-01 -5.52549899e-01
-5.05099177e-01 5.86681485e-01 8.53336811e-01 -1.92026824e-01
3.56240958e-01 -7.79693425e-02 -1.29166707e-01 -4.78451580e-01
-7.98153579e-01 -6.11733437e-01 5.19834161e-01 3.61471802e-01
2.51872521e-02 -6.30755126e-02 -1.52356356e-01 6.70489073e-01
-1.83248505e-01 -1.24546893e-01 3.15412521e-01 8.96805167e-01
-6.35684192e-01 -7.48893201e-01 -2.21744254e-01 4.29304868e-01
-4.18825954e-01 -3.82916510e-01 -5.73440313e-01 1.02216601e+00
3.51818591e-01 8.21420372e-01 3.67146432e-02 -3.81633490e-01
5.72486401e-01 2.04067349e-01 1.60832331e-01 -1.48436797e+00
-6.38804674e-01 -3.45094390e-02 1.26897797e-01 -7.08925128e-01
-5.54912865e-01 -4.71605867e-01 -1.00141656e+00 4.01217282e-01
-1.07542777e+00 2.97519267e-01 3.66304547e-01 8.80189538e-01
6.15150452e-01 2.95124829e-01 4.78777066e-02 -8.31568122e-01
-5.95252275e-01 -1.07792950e+00 -5.20776093e-01 6.37495875e-01
2.92520642e-01 -8.76163065e-01 -2.83360094e-01 1.28235295e-01] | [10.332995414733887, 1.5013929605484009] |
decad77b-30f1-48c0-a4d7-ce1ec5334144 | l-sa-learning-under-explored-targets-in-multi | 2305.13741 | null | https://arxiv.org/abs/2305.13741v1 | https://arxiv.org/pdf/2305.13741v1.pdf | L-SA: Learning Under-Explored Targets in Multi-Target Reinforcement Learning | Tasks that involve interaction with various targets are called multi-target tasks. When applying general reinforcement learning approaches for such tasks, certain targets that are difficult to access or interact with may be neglected throughout the course of training - a predicament we call Under-explored Target Problem (UTP). To address this problem, we propose L-SA (Learning by adaptive Sampling and Active querying) framework that includes adaptive sampling and active querying. In the L-SA framework, adaptive sampling dynamically samples targets with the highest increase of success rates at a high proportion, resulting in curricular learning from easy to hard targets. Active querying prompts the agent to interact more frequently with under-explored targets that need more experience or exploration. Our experimental results on visual navigation tasks show that the L-SA framework improves sample efficiency as well as success rates on various multi-target tasks with UTP. Also, it is experimentally demonstrated that the cyclic relationship between adaptive sampling and active querying effectively improves the sample richness of under-explored targets and alleviates UTP. | ['Byoung-Tak Zhang', 'Minsu Lee', 'Moonheon Lee', 'Min Whoo Lee', 'Hyundo Lee', 'Kibeom Kim'] | 2023-05-23 | null | null | null | null | ['visual-navigation'] | ['robots'] | [ 1.37299478e-01 1.41511187e-01 -3.11233014e-01 -1.89414427e-01
-8.71094167e-01 -5.54805636e-01 2.62618035e-01 2.50430852e-01
-6.53370678e-01 8.83738041e-01 -1.80852786e-01 -9.13830325e-02
-5.41616738e-01 -9.04520512e-01 -5.91426790e-01 -7.01211870e-01
-1.81646541e-01 6.24164343e-01 5.70507407e-01 -4.82293516e-01
4.60373372e-01 4.05294031e-01 -1.93589377e+00 1.25478476e-01
1.29131377e+00 5.84263861e-01 7.09019423e-01 3.59265804e-01
-5.19976318e-01 7.22725749e-01 -7.11425483e-01 1.73006505e-01
2.09254906e-01 -2.27547526e-01 -7.27737308e-01 -2.81511009e-01
3.14898342e-01 -3.44236881e-01 3.56829554e-01 9.84242141e-01
4.51926887e-01 6.74746215e-01 4.09537792e-01 -1.30795717e+00
-8.37873369e-02 6.13144875e-01 -6.49873435e-01 4.65718597e-01
4.51418072e-01 2.34867603e-01 5.98423064e-01 -7.39366293e-01
2.72603929e-01 1.47479272e+00 2.90502816e-01 5.57461321e-01
-1.08831465e+00 -7.92059958e-01 6.12160087e-01 1.57509103e-01
-9.49027002e-01 -1.49067074e-01 6.92814887e-01 -3.65656048e-01
5.43553650e-01 5.22925913e-01 8.68839145e-01 9.79938686e-01
-4.66554649e-02 1.12528491e+00 1.08749568e+00 -4.44654703e-01
5.07535219e-01 3.30529839e-01 1.33567616e-01 6.76726222e-01
1.66459858e-01 3.25821579e-01 -8.20110977e-01 -4.90183495e-02
9.89461660e-01 -2.42609251e-02 -2.92599678e-01 -8.84332538e-01
-1.03634071e+00 1.02319181e+00 2.80099899e-01 -3.01311985e-02
-3.92182410e-01 -1.16982818e-01 3.01424265e-01 6.05625153e-01
4.13961671e-02 7.81799018e-01 -5.13602674e-01 -4.33395982e-01
-3.93078595e-01 3.54110658e-01 6.31577313e-01 1.11377788e+00
8.87538493e-01 1.86882436e-01 -2.32654542e-01 1.07027137e+00
1.75150424e-01 3.43869120e-01 6.70724630e-01 -9.54543173e-01
4.50644106e-01 9.45078850e-01 3.69651616e-01 -5.67399323e-01
-1.31079867e-01 -5.85312426e-01 -3.47750872e-01 7.83462167e-01
5.00390053e-01 -2.42278978e-01 -6.48373425e-01 1.85353851e+00
8.51208568e-01 -2.08392471e-01 1.31188575e-02 7.50430942e-01
5.43446720e-01 5.32621086e-01 4.13989484e-01 -5.91533780e-01
1.11890113e+00 -1.22181010e+00 -7.34402537e-01 -3.21503341e-01
6.02751613e-01 -5.43095052e-01 1.84003711e+00 6.08001053e-01
-1.00253367e+00 -7.74978280e-01 -8.13178599e-01 3.71423721e-01
-4.36412930e-01 -2.01578900e-01 5.05273461e-01 6.95254505e-01
-6.12167299e-01 5.49773872e-01 -6.05721891e-01 -9.50544327e-02
3.40215921e-01 3.84770721e-01 -9.63523090e-02 -2.31956616e-02
-8.37488770e-01 7.91027009e-01 4.17889923e-01 -3.19787413e-01
-1.36153007e+00 -8.91138852e-01 -6.61909461e-01 1.93466261e-01
1.08874631e+00 -2.76899517e-01 1.48565459e+00 -9.71706986e-01
-1.78471684e+00 1.13102578e-01 2.30942592e-01 5.12245446e-02
7.19246149e-01 -6.25412345e-01 4.52816598e-02 -5.54841012e-02
2.09270313e-01 6.47791147e-01 7.64445603e-01 -1.34867156e+00
-7.41893113e-01 -2.50970453e-01 3.79423946e-01 9.79955912e-01
-4.89294380e-01 -4.45078373e-01 -1.42837346e-01 -2.03283191e-01
-3.02356780e-02 -6.66816652e-01 -3.00598741e-01 -9.47139040e-02
-2.37022731e-02 -6.36864305e-01 1.13965893e+00 1.84405133e-01
1.14072156e+00 -2.08419776e+00 2.82732069e-01 1.26890749e-01
3.23443770e-01 2.86300778e-01 -3.41687918e-01 4.52547640e-01
3.12252734e-02 -9.85823795e-02 2.96799302e-01 4.44980972e-02
-1.70042947e-01 2.06170559e-01 -4.68430147e-02 -3.10792357e-01
-2.05729768e-01 5.59788704e-01 -1.09399951e+00 -6.26195729e-01
3.25037032e-01 -7.93306604e-02 -5.52832961e-01 6.70801938e-01
-4.23949420e-01 6.19503319e-01 -8.53433907e-01 6.88341141e-01
3.40892047e-01 -1.76122636e-01 -1.03846662e-01 3.04224551e-01
-2.85805702e-01 1.36694044e-01 -1.26202261e+00 1.53669512e+00
-4.89198297e-01 9.34088528e-02 8.56490806e-02 -9.00211811e-01
9.78758693e-01 1.76537409e-02 2.73199052e-01 -1.01739168e+00
-3.02684665e-01 2.07269955e-02 6.50026798e-02 -7.27853298e-01
4.70188856e-01 4.09350395e-01 1.94232121e-01 3.66272330e-01
-1.76526040e-01 -6.29619509e-02 2.91087002e-01 1.94914341e-01
9.14083242e-01 4.99302328e-01 4.61589694e-01 -3.94011974e-01
4.25484687e-01 1.45545736e-01 4.56589997e-01 1.24101472e+00
-3.85668069e-01 -3.19227338e-01 4.78110999e-01 -3.00817937e-01
-6.49990439e-01 -1.13323081e+00 2.62010634e-01 1.83518994e+00
2.97690243e-01 -1.04079030e-01 -5.63524365e-01 -8.90244126e-01
-9.85042974e-02 7.01048911e-01 -6.28625512e-01 -1.62350491e-01
-5.50458312e-01 -1.28080159e-01 -1.93115532e-01 9.91140306e-02
6.22371554e-01 -1.60381949e+00 -1.23912287e+00 2.22365111e-01
2.83554904e-02 -3.29556257e-01 -4.33935791e-01 4.88706678e-01
-1.12845933e+00 -1.22196317e+00 -8.18032026e-01 -8.61147642e-01
7.64464080e-01 3.91837388e-01 1.00097609e+00 -2.27524471e-02
-1.02136962e-01 5.34105599e-01 -4.64628369e-01 -5.63844025e-01
-6.72143176e-02 7.35465139e-02 -1.60502434e-01 -4.82769221e-01
2.39334568e-01 -4.76705700e-01 -5.87709904e-01 4.76878196e-01
-6.40050292e-01 8.22595656e-02 6.92316055e-01 9.36725140e-01
6.37099326e-01 -5.74203022e-02 9.27546620e-01 -1.18875039e+00
9.86350775e-01 -4.78924423e-01 -8.02566290e-01 4.01837081e-01
-6.55626714e-01 5.10828122e-02 5.23461163e-01 -1.19928956e+00
-1.05723464e+00 -1.26343817e-01 1.90214500e-01 -4.27246451e-01
-1.37034923e-01 5.48263192e-01 -1.05557993e-01 -2.29643881e-01
8.74268830e-01 2.68277973e-01 9.75210816e-02 -2.53410220e-01
-1.86630070e-01 2.58032113e-01 -8.49201456e-02 -8.76705468e-01
5.28019488e-01 -1.76877290e-01 -2.32583225e-01 -6.82050705e-01
-9.71506536e-01 -3.59333545e-01 -6.00059442e-02 -5.86202741e-01
2.92182028e-01 -6.77235126e-01 -1.02438438e+00 2.68324316e-01
-4.10006851e-01 -8.69709194e-01 -5.11698425e-01 5.25270104e-01
-5.32019675e-01 1.15997359e-01 -1.56799987e-01 -1.14139354e+00
-2.14778230e-01 -1.41283834e+00 6.58598840e-01 7.35152602e-01
-1.95677146e-01 -7.22021461e-01 -2.31467430e-02 3.02838355e-01
6.30139351e-01 2.07815669e-03 1.10662663e+00 -6.90471470e-01
-8.44592392e-01 1.81037962e-01 2.08363980e-01 -3.17194402e-01
1.36957034e-01 -4.84741688e-01 -5.89821637e-01 -5.37683308e-01
3.31821851e-03 -8.59603345e-01 3.31462413e-01 2.89717227e-01
1.31093407e+00 -2.61330456e-01 -2.58079916e-01 2.44470075e-01
1.35118437e+00 7.81257510e-01 3.09403956e-01 6.56583071e-01
2.00670227e-01 5.33393264e-01 1.32941926e+00 5.10903418e-01
-8.74044653e-03 5.32684565e-01 7.21578121e-01 3.89256328e-02
3.26142460e-01 -2.97357738e-01 2.13507205e-01 4.56082135e-01
2.01312020e-01 -4.32845205e-02 -7.11565077e-01 5.62471807e-01
-1.85476899e+00 -6.77690923e-01 1.96841791e-01 2.44630790e+00
1.16889596e+00 3.85558546e-01 2.97352046e-01 1.56291313e-02
4.57412183e-01 -1.40544549e-01 -1.02124596e+00 -1.17284216e-01
5.93928874e-01 1.75301880e-01 -6.31869212e-02 6.60840631e-01
-6.47141039e-01 9.87265408e-01 6.15761328e+00 9.89549875e-01
-8.96795750e-01 7.78311566e-02 4.84585106e-01 -8.83701146e-02
-1.64147153e-01 -2.31701151e-01 -1.04532349e+00 3.34893733e-01
6.37935221e-01 -1.71281755e-01 5.65127790e-01 1.33460748e+00
8.00861567e-02 -5.89336932e-01 -1.11650324e+00 6.50204182e-01
-3.65736604e-01 -7.89362907e-01 3.40719521e-02 -7.44066313e-02
4.78178442e-01 -4.11383897e-01 2.09530234e-01 9.44594562e-01
6.39004648e-01 -7.62136221e-01 5.06120503e-01 4.30291802e-01
5.55589616e-01 -8.93535614e-01 2.53412962e-01 8.08281422e-01
-1.06588852e+00 -3.14606428e-01 -3.89974505e-01 6.60329238e-02
-3.11611354e-01 1.59947976e-01 -7.61586308e-01 2.24328563e-01
7.96022177e-01 5.04257530e-02 -3.84918034e-01 1.29809427e+00
-1.57487050e-01 3.04785073e-01 -2.61490643e-01 -5.83650053e-01
1.88893080e-01 -2.71853328e-01 5.11865497e-01 5.01436949e-01
3.51227075e-01 1.61899269e-01 6.30932868e-01 7.54701018e-01
4.49960619e-01 2.78823704e-01 -6.96843207e-01 1.05084330e-01
9.07779813e-01 1.12250745e+00 -5.01349211e-01 -3.74823719e-01
3.92250083e-02 4.94875759e-01 7.08583057e-01 4.72918302e-01
-6.70344591e-01 -4.70365614e-01 2.60795176e-01 9.92214084e-02
6.88364282e-02 -1.90783560e-01 9.77991614e-03 -6.33698404e-01
-2.91088104e-01 -1.15495002e+00 4.64917570e-01 -8.36593688e-01
-8.98269296e-01 5.04794598e-01 3.92888814e-01 -1.14645195e+00
-3.38207581e-03 -2.48622999e-01 -5.55899441e-01 7.62138128e-01
-1.46447098e+00 -4.14532542e-01 -6.05882764e-01 7.50951171e-01
1.13267422e+00 -4.16253656e-01 8.38423729e-01 -1.33739173e-01
-5.40923655e-01 4.92828578e-01 -7.32141063e-02 -4.63076770e-01
5.34412563e-01 -1.34177089e+00 -3.70791495e-01 2.72444397e-01
-2.34550729e-01 6.90977037e-01 6.11267269e-01 -7.09065974e-01
-1.20363355e+00 -6.84204698e-01 1.39507309e-01 -1.27962104e-03
2.40309358e-01 -1.28691435e-01 -8.90129983e-01 5.31322062e-01
2.02046573e-01 -4.09512669e-01 7.02462137e-01 3.04722995e-01
-7.18439743e-02 -1.28880337e-01 -1.15644395e+00 8.15729022e-01
7.90798426e-01 -7.47409239e-02 -5.53773224e-01 5.56263089e-01
7.67538607e-01 -6.38091803e-01 -6.55430973e-01 2.22122461e-01
2.74428934e-01 -1.11918175e+00 9.79891419e-01 -4.16760325e-01
2.25213751e-01 -8.27418547e-03 3.05868834e-01 -1.50891137e+00
-3.35563987e-01 -4.81963098e-01 -2.68519610e-01 7.54414618e-01
1.22792698e-01 -5.95705807e-01 1.06951892e+00 2.38991439e-01
-1.15030482e-01 -9.02264118e-01 -7.06421793e-01 -6.40081644e-01
8.18802393e-04 1.69837371e-01 4.45767790e-01 9.89899218e-01
7.55388066e-02 2.06705883e-01 -2.58669764e-01 -1.35553079e-02
8.64241958e-01 6.52438104e-02 8.74435663e-01 -1.47045159e+00
-3.36747676e-01 -1.69599235e-01 4.60654289e-01 -1.42840302e+00
-3.03445518e-01 -2.14485824e-01 8.85972008e-02 -1.19012940e+00
8.49259496e-02 -7.29121447e-01 -1.70671165e-01 5.16301632e-01
-3.73545706e-01 -5.26211977e-01 -8.92911628e-02 2.46023521e-01
-9.14953470e-01 6.81101143e-01 1.84737575e+00 1.19053446e-01
-6.93699062e-01 2.50732392e-01 -4.75227475e-01 5.26362002e-01
9.11095023e-01 -4.79544461e-01 -1.03280914e+00 -1.38722956e-01
3.83098274e-02 5.41881800e-01 -1.27487242e-01 -1.11450899e+00
2.26752654e-01 -8.32241952e-01 5.26633739e-01 -6.89843893e-01
4.93857265e-01 -8.99402380e-01 -1.06130749e-01 7.35351443e-01
-7.63011932e-01 1.63859978e-01 2.72441030e-01 6.69803679e-01
1.10908309e-02 -5.64161479e-01 6.99618697e-01 -4.51084554e-01
-6.31323338e-01 6.49532899e-02 -5.31120539e-01 3.89507413e-02
1.41273451e+00 -4.68750983e-01 -3.10313284e-01 -3.67239565e-01
-9.61025417e-01 7.57041335e-01 -2.64024436e-02 3.69761229e-01
7.97924399e-01 -1.19210482e+00 -1.50149465e-01 5.03947675e-01
7.38503784e-02 3.47827613e-01 2.56092787e-01 5.86092770e-01
-1.51333794e-01 1.81331873e-01 -4.96028036e-01 -6.65341616e-01
-1.57597685e+00 4.65572894e-01 3.71289223e-01 -5.46429813e-01
-3.67205858e-01 9.64015603e-01 2.99217284e-01 -6.94529116e-01
8.76248419e-01 -7.23797977e-02 -7.05894649e-01 -2.55208835e-02
4.07453835e-01 3.90397072e-01 -3.45675021e-01 2.89281487e-01
2.66359419e-01 2.44492635e-01 -5.31846225e-01 1.77246071e-02
1.19137979e+00 -9.88437086e-02 4.19025123e-01 7.14130044e-01
3.78167629e-01 -6.26897663e-02 -1.56795609e+00 -3.13805163e-01
3.09972465e-02 -7.93771505e-01 -1.22775830e-01 -1.09574449e+00
-8.08112264e-01 6.22337222e-01 7.99757361e-01 5.21089673e-01
9.74016547e-01 -4.23535466e-01 9.08168554e-02 7.94157684e-01
5.74181736e-01 -1.15458608e+00 9.50522661e-01 5.59294283e-01
9.41883743e-01 -1.20501816e+00 -1.58707097e-01 -4.65193748e-01
-7.69581258e-01 9.80187356e-01 1.65962315e+00 1.07170872e-01
2.57937998e-01 1.00144438e-01 3.21144685e-02 -2.56959707e-01
-8.76187146e-01 4.54581566e-02 -7.24005327e-02 6.99238777e-01
1.85757950e-01 -2.24345237e-01 -2.49541402e-01 9.33731496e-02
-7.48794153e-02 -1.39604956e-01 5.16227126e-01 1.36075842e+00
-1.08739114e+00 -1.24348366e+00 -4.87861365e-01 4.63417321e-01
-5.38777523e-02 6.40462562e-02 -6.16502725e-02 9.78036344e-01
-1.19410373e-01 7.81696379e-01 -1.22450143e-01 -8.88026953e-02
3.13980758e-01 1.31792828e-01 5.06463885e-01 -8.19267035e-01
-7.79345810e-01 2.48545781e-01 -1.63408533e-01 -6.57421887e-01
-1.93605214e-01 -3.89291137e-01 -1.10972261e+00 -3.68837267e-02
-5.42324066e-01 5.97003937e-01 4.15657014e-01 6.23980343e-01
7.18169063e-02 8.62219572e-01 4.47882950e-01 -5.39660275e-01
-1.12653053e+00 -1.10061491e+00 -5.27611554e-01 8.24938864e-02
3.85962605e-01 -1.04828811e+00 -2.71333873e-01 -4.28433746e-01] | [4.0683488845825195, 1.7731633186340332] |
101b8a22-44f0-48c2-b7db-b0b89d724b59 | hooks-in-the-headline-learning-to-generate | 2004.01980 | null | https://arxiv.org/abs/2004.01980v3 | https://arxiv.org/pdf/2004.01980v3.pdf | Hooks in the Headline: Learning to Generate Headlines with Controlled Styles | Current summarization systems only produce plain, factual headlines, but do not meet the practical needs of creating memorable titles to increase exposure. We propose a new task, Stylistic Headline Generation (SHG), to enrich the headlines with three style options (humor, romance and clickbait), in order to attract more readers. With no style-specific article-headline pair (only a standard headline summarization dataset and mono-style corpora), our method TitleStylist generates style-specific headlines by combining the summarization and reconstruction tasks into a multitasking framework. We also introduced a novel parameter sharing scheme to further disentangle the style from the text. Through both automatic and human evaluation, we demonstrate that TitleStylist can generate relevant, fluent headlines with three target styles: humor, romance, and clickbait. The attraction score of our model generated headlines surpasses that of the state-of-the-art summarization model by 9.68%, and even outperforms human-written references. | ['Joey Tianyi Zhou', 'Zhijing Jin', 'Peter Szolovits', 'Lisa Orii', 'Di Jin'] | 2020-04-04 | hooks-in-the-headline-learning-to-generate-1 | https://aclanthology.org/2020.acl-main.456 | https://aclanthology.org/2020.acl-main.456.pdf | acl-2020-6 | ['headline-generation'] | ['natural-language-processing'] | [ 1.15946762e-01 1.87457249e-01 -1.46461114e-01 -1.12738319e-01
-1.29211164e+00 -6.05176926e-01 9.74187493e-01 5.78519888e-02
-3.98201048e-01 1.18002594e+00 9.62731600e-01 -1.74202755e-01
3.04726601e-01 -3.68257344e-01 -6.22418404e-01 -1.20063521e-01
6.61228597e-01 6.22737467e-01 2.06798643e-01 -5.10041595e-01
6.08637452e-01 -6.60260171e-02 -1.26457584e+00 5.41965306e-01
1.53852236e+00 2.59568572e-01 4.65322375e-01 9.80845153e-01
-2.25643814e-01 6.95761859e-01 -1.32256961e+00 -7.66913652e-01
-3.04440945e-01 -7.11520612e-01 -7.14864075e-01 -7.75612444e-02
1.00083292e+00 -3.65266830e-01 -2.93778777e-01 8.07650447e-01
8.10860455e-01 7.55171552e-02 7.42902994e-01 -8.23149621e-01
-9.99185264e-01 1.08196771e+00 -6.95549905e-01 1.04688987e-01
6.65956676e-01 7.85050448e-03 1.34008002e+00 -6.99965715e-01
7.33034492e-01 1.36684287e+00 4.33749348e-01 8.14366937e-01
-1.26167512e+00 -5.03000081e-01 7.34592006e-02 -2.93928217e-02
-9.03997838e-01 -4.47809160e-01 8.90552700e-01 -2.86125839e-01
6.84098959e-01 7.72592247e-01 4.23647761e-01 1.54305434e+00
4.85394429e-03 1.36259937e+00 9.74914074e-01 -2.67822415e-01
6.10921793e-02 2.78757244e-01 1.66911885e-01 2.48050004e-01
4.53012794e-01 -5.79543948e-01 -9.06571686e-01 -8.76444131e-02
5.36503792e-01 -4.35484141e-01 -4.60898846e-01 3.33663940e-01
-1.44766867e+00 7.27139711e-01 1.37442136e-02 1.12697333e-01
-3.15025985e-01 -6.73086867e-02 7.02147186e-01 -7.70734297e-03
6.35208130e-01 1.24479353e+00 -8.83391872e-02 -2.62702942e-01
-1.40958917e+00 7.21197546e-01 1.18130326e+00 1.36073947e+00
3.36414307e-01 1.08239271e-01 -1.05226958e+00 1.04142499e+00
-3.31148565e-01 8.61818075e-01 6.15662932e-01 -8.61178994e-01
7.15835035e-01 2.19748691e-01 4.20493811e-01 -8.94630790e-01
-3.10019404e-01 -8.17221880e-01 -8.37512314e-01 -5.66061854e-01
1.49459124e-01 -4.19689596e-01 -2.90612072e-01 1.52000880e+00
-2.01865435e-01 -4.60722953e-01 -1.44746557e-01 6.67073011e-01
1.27424586e+00 8.97636414e-01 -8.49177837e-02 -3.54516715e-01
1.43016481e+00 -1.39837539e+00 -1.05759561e+00 -3.36762995e-01
5.70982695e-01 -1.12369728e+00 1.71016574e+00 5.54076016e-01
-1.49865031e+00 -6.07466340e-01 -1.17884290e+00 -2.56012022e-01
6.47595152e-02 5.08963883e-01 1.66382432e-01 3.31724823e-01
-7.15342224e-01 5.11312485e-01 -2.41288796e-01 -1.92052588e-01
1.35966852e-01 -2.97656834e-01 1.01968490e-01 2.82870114e-01
-1.28775191e+00 9.71453011e-01 1.39095679e-01 -4.51576740e-01
-4.45131928e-01 -9.30348277e-01 -7.63178766e-01 4.88673570e-03
2.74933636e-01 -1.06285012e+00 1.67394638e+00 -6.95565164e-01
-1.51313329e+00 6.76395476e-01 -3.19833487e-01 -4.53983217e-01
8.43993843e-01 -8.99736643e-01 -2.39662841e-01 1.70990989e-01
3.59190464e-01 6.08737350e-01 8.46861422e-01 -1.33935106e+00
-5.42624533e-01 2.10091159e-01 -2.58570373e-01 5.02963603e-01
-5.03337502e-01 3.96476090e-02 -4.86749470e-01 -1.06157422e+00
-5.43278277e-01 -6.44362569e-01 -1.52753154e-02 -8.40220273e-01
-1.31775892e+00 -3.15663993e-01 5.45577824e-01 -1.03844893e+00
1.89581728e+00 -1.71348500e+00 3.07215035e-01 -3.08084786e-01
4.02744234e-01 2.44049981e-01 -1.86262459e-01 6.31645262e-01
5.61525345e-01 2.47895315e-01 8.93658623e-02 -7.41898537e-01
1.96289882e-01 -2.84756333e-01 -5.12203217e-01 -2.11209685e-01
-2.46590916e-02 1.11333990e+00 -1.19277692e+00 -6.45282388e-01
-1.51097357e-01 -1.37959505e-02 -6.21091187e-01 2.39720434e-01
-5.08859217e-01 2.22259596e-01 -4.58934873e-01 2.07083449e-01
4.00314122e-01 -3.06305528e-01 -1.88105613e-01 -2.11273953e-01
-2.47875512e-01 5.75101018e-01 -6.42240286e-01 1.52394474e+00
-7.16274500e-01 8.54358792e-01 -4.05671865e-01 3.32383402e-02
9.78773654e-01 -8.72945134e-03 -7.93514624e-02 -6.71405554e-01
-1.20460734e-01 3.13983619e-01 -2.62678981e-01 -4.82665122e-01
1.33965409e+00 9.39455628e-02 -5.60792804e-01 7.02914238e-01
-8.38216394e-02 -4.55692381e-01 6.21638060e-01 8.21096003e-01
9.45349634e-01 6.31719362e-03 1.06956325e-01 -2.54757494e-01
3.05756748e-01 -1.43103793e-01 1.74168676e-01 1.12322927e+00
2.35443890e-01 9.31275904e-01 6.78503633e-01 1.26741990e-01
-1.32810938e+00 -9.22846258e-01 2.88007081e-01 1.31160390e+00
1.09161831e-01 -8.50142360e-01 -1.03718400e+00 -5.47807574e-01
-1.50478408e-01 1.46224701e+00 -3.49901378e-01 -7.16881081e-02
-6.77125692e-01 -4.64186996e-01 7.42730439e-01 2.59338349e-01
6.91970110e-01 -1.12677634e+00 -5.80168962e-01 2.53772289e-01
-7.07619131e-01 -9.86816287e-01 -1.19741905e+00 -3.97824913e-01
-4.82810140e-01 -5.76348782e-01 -1.18618345e+00 -6.36691868e-01
3.14814448e-01 2.14031428e-01 1.37184155e+00 -2.19011635e-01
2.09317312e-01 9.52537805e-02 -4.71470743e-01 -5.02526402e-01
-6.54904366e-01 7.30801821e-01 -1.55646250e-01 -2.67252028e-01
2.93164956e-03 -2.77319968e-01 -6.52004004e-01 7.14225769e-02
-7.90350616e-01 7.45110154e-01 6.96808159e-01 9.77041721e-01
6.35392889e-02 -6.00320339e-01 1.06437111e+00 -1.31637383e+00
1.32389045e+00 -1.95821688e-01 6.24250025e-02 3.53515625e-01
-4.69439954e-01 1.39054164e-01 8.27080786e-01 -5.79506814e-01
-1.41418910e+00 -4.82749343e-01 -4.97061610e-02 6.71202242e-02
1.49811819e-01 5.36811769e-01 -1.58680499e-01 6.09319329e-01
9.51509178e-01 4.64987755e-01 -2.03273982e-01 -6.53210223e-01
7.27232397e-01 7.32722580e-01 8.66357267e-01 -6.12505257e-01
8.16082776e-01 6.48110174e-03 -4.51291412e-01 -9.68558490e-01
-1.26151383e+00 -4.92092550e-01 -3.45287412e-01 -2.74893314e-01
5.81428647e-01 -8.67753267e-01 -2.82266527e-01 4.97118205e-01
-1.63278389e+00 -2.42947340e-01 -3.82969290e-01 -1.78042743e-02
-5.00766933e-01 5.52706122e-01 -7.72348881e-01 -5.37844479e-01
-9.18313503e-01 -7.87036121e-01 1.22760248e+00 3.95324290e-01
-8.99785936e-01 -8.50185692e-01 6.50521442e-02 4.79777634e-01
5.78046501e-01 1.42172530e-01 8.81649196e-01 -9.72847104e-01
-2.70410210e-01 -2.93475002e-01 -2.88498461e-01 2.54483193e-01
-9.62134637e-03 1.17009141e-01 -6.32793069e-01 -1.85002144e-02
-3.13536018e-01 -5.14977455e-01 9.55727100e-01 3.18010271e-01
9.02571023e-01 -9.94433701e-01 -1.30722418e-01 2.34923452e-01
7.04937518e-01 -2.15770245e-01 7.07334995e-01 3.78178954e-01
8.48054469e-01 5.74119151e-01 5.03756106e-01 7.99693882e-01
5.56911290e-01 6.13538384e-01 -1.71516865e-01 -2.00820178e-01
-4.58693981e-01 -8.17492664e-01 3.79851818e-01 1.20817173e+00
2.52530992e-01 -5.55945277e-01 -3.72170389e-01 5.07516801e-01
-1.71965945e+00 -1.21957076e+00 -3.44074965e-01 1.95221734e+00
1.39718390e+00 4.11337703e-01 4.67363626e-01 -1.43173367e-01
7.20174432e-01 4.24684346e-01 -4.28485066e-01 -5.31268120e-01
-3.73726726e-01 -4.73485775e-02 3.41241539e-01 5.89973092e-01
-8.10954332e-01 1.12365842e+00 6.05310154e+00 1.14289784e+00
-1.00679028e+00 -1.38672322e-01 5.53965688e-01 -3.20936769e-01
-8.14374626e-01 -2.43570134e-01 -1.13931656e+00 6.22629941e-01
6.47987008e-01 -6.74670458e-01 8.34328234e-02 7.87573874e-01
5.15206456e-01 5.92864957e-03 -1.01470566e+00 8.64301085e-01
5.99627018e-01 -1.41575384e+00 5.15870392e-01 -3.22059482e-01
9.58416820e-01 -4.59694624e-01 1.94993719e-01 5.98475695e-01
2.82427669e-01 -7.38780379e-01 1.09049070e+00 5.72788119e-01
9.34428990e-01 -5.54593265e-01 5.78590572e-01 4.33858156e-01
-4.79691863e-01 3.14547896e-01 -1.33794695e-01 5.80945015e-02
6.54740870e-01 6.63062871e-01 -8.53804767e-01 3.68402362e-01
5.45603000e-02 6.90642595e-01 -1.00218463e+00 1.20247388e+00
-4.99453098e-01 8.02929282e-01 7.62445256e-02 -6.72032416e-01
2.06989914e-01 1.88067257e-01 1.04855132e+00 1.69124758e+00
2.52491295e-01 -1.42911166e-01 7.64640421e-02 9.28569198e-01
-4.06811297e-01 2.22491980e-01 -2.69469053e-01 -8.59963521e-02
8.09293449e-01 1.29043436e+00 -2.03627378e-01 -5.88484883e-01
-2.07392918e-03 1.24035192e+00 1.50893390e-01 4.25342351e-01
-6.91377342e-01 -9.18210328e-01 -1.24507118e-02 6.10125810e-02
-1.74283814e-02 -4.32070568e-02 -8.64444852e-01 -1.36606038e+00
-7.41269886e-02 -1.17384672e+00 3.20225544e-02 -8.67144287e-01
-1.29242218e+00 6.67568743e-01 5.37820533e-02 -1.03137708e+00
-1.81863636e-01 6.32463321e-02 -1.07188320e+00 8.70633483e-01
-1.20429671e+00 -1.13785100e+00 -3.61847311e-01 -2.78692544e-02
1.18771827e+00 -1.55262485e-01 3.45838666e-01 1.06545389e-01
-7.03803718e-01 9.56061125e-01 1.14090942e-01 -2.77545184e-01
1.19901478e+00 -1.55104959e+00 7.60274887e-01 7.25170076e-01
-1.51867807e-01 7.19565868e-01 1.40726733e+00 -8.92711282e-01
-9.80744421e-01 -1.02543032e+00 1.32563722e+00 -5.41058004e-01
7.08031595e-01 -4.59051639e-01 -8.24390709e-01 2.86296606e-01
7.12833822e-01 -1.18515813e+00 6.21279597e-01 2.94654816e-01
-2.96462774e-01 1.42664254e-01 -6.25285506e-01 1.19146693e+00
8.62918079e-01 -1.55235767e-01 -9.52116430e-01 4.51289058e-01
1.16920722e+00 -4.01845008e-01 -3.47637266e-01 3.64866331e-02
4.44236308e-01 -7.69035041e-01 6.79172695e-01 -5.97045004e-01
1.17329013e+00 5.99700920e-02 2.66164064e-01 -1.65536129e+00
-4.33452606e-01 -1.11720121e+00 -2.53383994e-01 1.59960186e+00
7.02129602e-01 -2.51794457e-01 4.99735028e-01 5.16820908e-01
-5.83547890e-01 -5.64414144e-01 -4.20197040e-01 -7.46659815e-01
1.12874895e-01 1.11695506e-01 5.27289629e-01 5.04369020e-01
4.46760297e-01 1.15576339e+00 -7.99528122e-01 -6.36476219e-01
4.60097224e-01 1.11564636e-01 1.14254975e+00 -9.03899550e-01
-3.98382455e-01 -9.39755261e-01 4.96671051e-01 -1.53849089e+00
1.72289342e-01 -7.26813674e-01 1.73730180e-01 -1.76534736e+00
5.06968975e-01 -4.46591713e-02 3.03870052e-01 1.71021298e-01
-6.38423085e-01 5.47235161e-02 4.26449955e-01 3.84984165e-01
-1.01976633e+00 7.48723626e-01 1.71329343e+00 -1.49595216e-01
-5.19831598e-01 1.15699857e-01 -1.29650664e+00 4.81965899e-01
6.15076542e-01 1.15139904e-02 -3.70402366e-01 -3.00708741e-01
3.72952104e-01 2.26975992e-01 5.68659157e-02 -7.84783483e-01
2.02255800e-01 -8.08334723e-02 1.36617735e-01 -8.50312948e-01
1.55988351e-01 2.64198512e-01 -2.64876068e-01 9.91356671e-02
-1.06406534e+00 9.31236520e-02 1.55602872e-01 4.35653508e-01
-3.36114131e-02 -3.89329970e-01 5.02038062e-01 -4.88785468e-02
-1.12436205e-01 -1.51346877e-01 -4.97181863e-01 5.87208092e-01
4.22531307e-01 3.33161019e-02 -8.53854060e-01 -8.66830647e-01
-1.29216522e-01 4.34712887e-01 5.15819848e-01 3.79143357e-01
6.86666250e-01 -1.21547329e+00 -1.26007926e+00 -2.85230696e-01
6.94864318e-02 -1.41428784e-01 2.57777840e-01 5.44223905e-01
-5.36481917e-01 4.44878668e-01 -7.65816197e-02 -8.98465887e-02
-1.17890203e+00 2.91176528e-01 -2.66417474e-01 -5.21970749e-01
-5.85619032e-01 8.00538898e-01 3.63894314e-01 2.54693348e-02
2.45219320e-01 -8.82798247e-03 -3.38526130e-01 3.80156070e-01
8.18313599e-01 7.55312681e-01 -6.61709905e-02 -2.68537462e-01
3.24813485e-01 5.61605487e-03 -5.95276296e-01 -4.37793523e-01
8.65403593e-01 -2.83722550e-01 6.26953468e-02 5.68359673e-01
9.13109004e-01 4.38185692e-01 -1.01016164e+00 -1.35181785e-01
1.65667281e-01 -3.13876122e-01 -2.19997868e-01 -1.04402626e+00
-3.01876426e-01 6.42948329e-01 -5.79266846e-01 3.89361084e-01
7.57442594e-01 -2.14314135e-03 1.44274020e+00 3.87725532e-01
2.01903004e-02 -1.32970953e+00 5.73672950e-01 8.06573272e-01
1.40906227e+00 -9.83527660e-01 1.30883172e-01 -1.55401632e-01
-1.28054869e+00 9.06741977e-01 8.09251547e-01 4.14763428e-02
-2.16933191e-01 -2.27423400e-01 -1.18472680e-01 -2.78148316e-02
-9.66191769e-01 1.49729595e-01 7.37996697e-01 4.58751291e-01
6.84339762e-01 9.66837257e-02 -6.96497440e-01 9.35241401e-01
-8.85001957e-01 -4.08137649e-01 1.00248384e+00 3.77255857e-01
-7.72938371e-01 -7.26650178e-01 -2.46684998e-01 8.32621515e-01
-3.39476913e-01 -3.23813647e-01 -7.23202705e-01 4.78070110e-01
-6.74891651e-01 9.09952939e-01 -2.31335089e-01 -3.24803531e-01
4.55931872e-01 -8.37035701e-02 2.84769177e-01 -8.00713599e-01
-8.07415247e-01 3.21041346e-01 6.23194814e-01 -2.10116617e-02
1.37989566e-01 -6.32130802e-01 -8.15279722e-01 -5.07099807e-01
-3.10929239e-01 3.29845458e-01 3.68022889e-01 6.98908448e-01
3.46015453e-01 6.00223601e-01 5.97464025e-01 -1.07748222e+00
-7.64146745e-01 -1.38495851e+00 -4.70557094e-01 6.74257576e-01
3.09031397e-01 -2.29211614e-01 -3.79148662e-01 4.96613085e-02] | [12.272357940673828, 9.261817932128906] |
02bd294b-02e5-4602-b59d-1492e5185b80 | split-u-net-preventing-data-leakage-in-split | 2208.10553 | null | https://arxiv.org/abs/2208.10553v2 | https://arxiv.org/pdf/2208.10553v2.pdf | Split-U-Net: Preventing Data Leakage in Split Learning for Collaborative Multi-Modal Brain Tumor Segmentation | Split learning (SL) has been proposed to train deep learning models in a decentralized manner. For decentralized healthcare applications with vertical data partitioning, SL can be beneficial as it allows institutes with complementary features or images for a shared set of patients to jointly develop more robust and generalizable models. In this work, we propose "Split-U-Net" and successfully apply SL for collaborative biomedical image segmentation. Nonetheless, SL requires the exchanging of intermediate activation maps and gradients to allow training models across different feature spaces, which might leak data and raise privacy concerns. Therefore, we also quantify the amount of data leakage in common SL scenarios for biomedical image segmentation and provide ways to counteract such leakage by applying appropriate defense strategies. | ['Daguang Xu', 'Andriy Myronenko', 'Wenqi Li', 'Can Zhao', 'Ziyue Xu', 'Ali Hatamizadeh', 'Holger R. Roth'] | 2022-08-22 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [-5.30950800e-02 4.51507032e-01 -3.47219527e-01 -5.91775775e-01
-7.68222034e-01 -7.10308611e-01 -5.40625080e-02 2.65706003e-01
-6.08204901e-01 9.37324345e-01 5.64159714e-02 -6.60919905e-01
7.35000893e-02 -7.14075983e-01 -7.34917760e-01 -7.30860233e-01
-5.96828721e-02 2.03489810e-02 -2.21853822e-01 2.99664587e-01
-1.69874892e-01 6.99886501e-01 -3.24642092e-01 5.69501996e-01
8.67016435e-01 9.62899029e-01 -2.19206676e-01 3.05895030e-01
1.56454906e-01 7.26693988e-01 -6.65069163e-01 -7.94940948e-01
8.83547068e-01 -2.53085703e-01 -1.03103435e+00 -1.54670790e-01
3.10014576e-01 -7.49131143e-01 -3.42104971e-01 1.06998491e+00
7.40210891e-01 -1.81877881e-01 2.29188696e-01 -1.48356295e+00
-5.27383327e-01 7.21509159e-01 -5.55896103e-01 -1.93724483e-02
-3.65958922e-02 1.94570258e-01 7.43739367e-01 -1.54631451e-01
8.25873733e-01 8.78587842e-01 8.25929523e-01 9.31124628e-01
-1.30838108e+00 -1.05604601e+00 1.45633325e-01 -2.89975464e-01
-1.05159247e+00 -2.62370974e-01 6.06234550e-01 -9.61670578e-02
6.16244793e-01 5.41090906e-01 6.75481737e-01 1.12968731e+00
4.16007102e-01 1.05253243e+00 1.22053719e+00 8.12590420e-02
3.27145725e-01 5.77504933e-01 -2.20440812e-02 7.31163263e-01
4.60613549e-01 -2.47056395e-01 -3.08655679e-01 -7.52042770e-01
7.35551715e-01 4.71565038e-01 -3.77053827e-01 -6.07775629e-01
-9.22590971e-01 9.49791133e-01 8.32013428e-01 1.67214960e-01
-2.48350203e-01 5.29204682e-02 7.09138215e-01 3.31339866e-01
3.02082419e-01 5.22929251e-01 -8.24180424e-01 5.53370833e-01
-1.04138064e+00 9.05721709e-02 7.38730371e-01 8.96166265e-01
5.85237622e-01 -4.32059020e-01 -3.46140653e-01 7.47263581e-02
1.81282476e-01 8.96813422e-02 3.79448593e-01 -1.20652008e+00
6.08253419e-01 5.64168036e-01 -1.16646133e-01 -9.11602020e-01
-5.14852762e-01 -4.26486015e-01 -1.24843490e+00 -9.56484377e-02
3.51100236e-01 -8.24385464e-01 -7.49245167e-01 1.72589326e+00
3.97605389e-01 1.90234095e-01 2.92649329e-01 6.22908652e-01
7.35282004e-01 1.73298746e-01 6.31846413e-02 -1.44446537e-01
1.11502075e+00 -7.50334084e-01 -7.11114943e-01 1.06370844e-01
1.12775218e+00 -8.89120400e-02 6.76976621e-01 2.35553175e-01
-9.29098725e-01 2.04820976e-01 -8.64848018e-01 -1.14875071e-01
-3.88500512e-01 -1.29228964e-01 7.36044586e-01 1.21515048e+00
-1.19562900e+00 5.32389402e-01 -1.15478349e+00 -1.71045288e-01
1.45743763e+00 6.85119689e-01 -4.24338132e-01 8.21322855e-03
-1.19153547e+00 2.85673082e-01 -6.22590482e-02 6.33826405e-02
-8.99432600e-01 -9.22151148e-01 -7.16575503e-01 1.72899812e-01
1.07427068e-01 -9.84828472e-01 7.46798813e-01 -6.59676075e-01
-1.15320504e+00 7.50326633e-01 2.93568045e-01 -9.50477660e-01
1.03255546e+00 -5.63580319e-02 1.86529756e-01 2.24664375e-01
8.69242996e-02 9.41619575e-01 5.07278144e-01 -1.19732070e+00
-3.13003719e-01 -6.90042138e-01 -5.95624559e-02 4.25004251e-02
-6.14052534e-01 -1.16652936e-01 -4.30178344e-02 -5.83542407e-01
-4.89983819e-02 -1.01119971e+00 -9.25494313e-01 2.70429760e-01
-8.13267589e-01 3.86429310e-01 8.74915600e-01 -7.29846001e-01
9.72089231e-01 -2.25844789e+00 -3.71435434e-01 5.40479302e-01
5.78937232e-01 8.25900882e-02 9.30507407e-02 2.65674125e-02
3.80110770e-01 5.62803149e-01 -2.86557049e-01 -4.02583420e-01
-3.23886573e-01 2.11771458e-01 -3.29043001e-01 7.08779693e-01
-1.45592958e-01 1.05353117e+00 -5.17549515e-01 -5.21940768e-01
-3.50372463e-01 3.20869833e-01 -1.03267479e+00 -4.35108580e-02
1.32111266e-01 7.99981058e-01 -8.04156482e-01 7.29497254e-01
9.86044526e-01 -2.79382765e-01 5.91484249e-01 -2.04356581e-01
4.55854177e-01 -7.29549006e-02 -7.52616823e-01 1.58526003e+00
-1.20425954e-01 3.37714761e-01 5.22651851e-01 -1.03724790e+00
5.56172490e-01 4.18510526e-01 8.83776844e-01 -4.50415403e-01
2.69250065e-01 -2.23387703e-02 -3.23406845e-01 -3.31539214e-01
7.73341209e-02 3.65505852e-02 -3.14493001e-01 5.40668428e-01
-1.90687492e-01 7.98165426e-02 -7.99961984e-01 3.61199319e-01
1.15032411e+00 -4.79665756e-01 7.40412548e-02 -3.75660598e-01
2.89385438e-01 -3.16544980e-01 9.45798218e-01 1.05137134e+00
-7.56552339e-01 3.31240207e-01 6.40897393e-01 -6.19151950e-01
-6.01028502e-01 -9.78789449e-01 -1.93530604e-01 8.36751938e-01
9.25321039e-03 -2.70057678e-01 -1.02215767e+00 -1.44301426e+00
2.53400415e-01 1.87469855e-01 -4.41978127e-01 -7.61937574e-02
-3.35791707e-01 -7.57275641e-01 1.13404536e+00 4.65697616e-01
6.85025752e-01 -7.76864886e-01 -9.28857982e-01 -1.46871060e-02
-9.53134894e-02 -9.62294340e-01 -6.21054053e-01 4.15406138e-01
-1.09245670e+00 -1.05042398e+00 -7.30609477e-01 -6.26967430e-01
1.04901123e+00 9.14280638e-02 6.17208600e-01 3.12846079e-02
-4.85093057e-01 3.80834252e-01 8.63222107e-02 -5.78327179e-01
-7.73567855e-02 1.37878463e-01 8.48669186e-02 -1.61582101e-02
9.13699642e-02 -3.28535020e-01 -1.06129956e+00 1.86585590e-01
-9.31357622e-01 -2.45458454e-01 3.24090898e-01 9.77943659e-01
4.71097291e-01 -2.19117746e-01 6.87483132e-01 -1.31780386e+00
8.79527628e-01 -5.06267250e-01 -5.84887326e-01 5.62810540e-01
-7.28984118e-01 -2.50778884e-01 6.02474928e-01 -2.43572801e-01
-9.62745726e-01 2.60802597e-01 8.50691497e-02 -4.49729651e-01
1.51940621e-02 2.18823045e-01 -2.31398180e-01 -5.29742002e-01
5.28909862e-01 6.75007850e-02 1.75999910e-01 -4.17775661e-01
5.71245432e-01 7.75703907e-01 1.11467510e-01 -4.56849545e-01
3.53639573e-01 8.76707435e-01 -1.00096568e-01 -2.98876733e-01
-5.52293897e-01 -1.37115911e-01 -5.30321896e-01 3.56104016e-01
9.93705809e-01 -1.13692939e+00 -8.22822630e-01 2.91791677e-01
-1.07059371e+00 -2.53781021e-01 -4.05853212e-01 3.09428185e-01
-3.08435231e-01 3.93816769e-01 -9.40480471e-01 -4.09704566e-01
-8.16511273e-01 -1.33864021e+00 6.63548589e-01 2.91653484e-01
-8.60684812e-02 -8.16958904e-01 -2.35555202e-01 7.87032306e-01
3.78878504e-01 3.39192182e-01 7.35561073e-01 -9.59428489e-01
-8.44359100e-01 -1.40841424e-01 -1.95236474e-01 4.99751121e-01
2.82233834e-01 -5.10584593e-01 -9.86875534e-01 -8.66204381e-01
3.67970407e-01 -4.62609470e-01 6.60094857e-01 5.55699944e-01
1.63634896e+00 -1.00695145e+00 -8.09739769e-01 1.03787279e+00
1.24498594e+00 1.26338497e-01 3.95769924e-01 5.80650494e-02
6.81530297e-01 6.04758084e-01 1.92482010e-01 6.88914776e-01
3.67912114e-01 -1.17141508e-01 1.56716898e-01 -6.01522624e-01
6.00320876e-01 -1.58729795e-02 -1.41750230e-02 3.25608164e-01
4.48860794e-01 6.85668737e-02 -7.32591510e-01 4.25454021e-01
-1.65618062e+00 -4.92634922e-01 3.21930498e-01 1.87119603e+00
9.66267109e-01 -2.03628615e-01 -8.74979347e-02 -5.90042591e-01
3.92630458e-01 1.76317673e-02 -8.95492852e-01 -4.40899462e-01
3.45865935e-02 1.98194042e-01 1.18363500e+00 1.70142204e-02
-1.10825884e+00 7.05267608e-01 7.19734144e+00 5.45621634e-01
-1.23581874e+00 5.06465435e-01 1.43441367e+00 -4.43993300e-01
-4.77776796e-01 -3.23582351e-01 -5.57228863e-01 2.89245546e-01
6.86348438e-01 -1.68372020e-01 3.32430489e-02 9.36709106e-01
-4.19965982e-02 2.79814541e-01 -9.11611259e-01 9.35057640e-01
-3.21001768e-01 -1.82128608e+00 -1.85078511e-03 3.24170291e-01
1.12002254e+00 2.74910510e-01 3.19708109e-01 -9.16329771e-02
6.12570941e-01 -1.04473925e+00 -8.89272953e-04 1.38170183e-01
5.57237625e-01 -8.43941391e-01 7.77663231e-01 4.33303654e-01
-6.91323876e-01 -3.52488577e-01 -3.08920324e-01 3.84239495e-01
-1.85864523e-01 3.73203844e-01 -8.65190506e-01 6.42470777e-01
9.52782691e-01 1.93047643e-01 -5.15777946e-01 8.30833793e-01
1.22866772e-01 4.41168249e-01 -3.38603467e-01 2.27162614e-01
3.36291492e-01 -6.57521263e-02 1.80681303e-01 9.95457709e-01
2.74430305e-01 4.74157464e-03 2.46048674e-01 7.97128916e-01
-5.62804163e-01 3.24210793e-01 -8.78305316e-01 1.96760833e-01
4.19535965e-01 1.08764553e+00 -6.61012828e-01 -1.97734728e-01
-4.21178997e-01 1.04783273e+00 2.78276265e-01 3.84453237e-01
-6.02864027e-01 -1.12350360e-01 8.74083102e-01 1.20156631e-03
1.37891211e-02 9.05568600e-02 -7.22297668e-01 -1.28499615e+00
-6.46166876e-02 -8.02696824e-01 7.59279668e-01 -3.71928811e-02
-1.36586308e+00 4.74241942e-01 -3.80410314e-01 -9.33063626e-01
1.90997094e-01 -1.27716094e-01 -4.90837485e-01 5.88020563e-01
-1.33620846e+00 -1.27880156e+00 1.50990546e-01 1.09080708e+00
-1.41542956e-01 -3.19609791e-01 8.05423081e-01 2.40841955e-01
-7.18184531e-01 1.23735654e+00 3.66726756e-01 4.76776987e-01
7.79713213e-01 -8.85872126e-01 1.07135236e-01 6.69695258e-01
5.06474776e-03 8.62188518e-01 2.26264820e-01 -8.08089674e-01
-1.11169684e+00 -1.33879232e+00 5.19779742e-01 -3.19157690e-01
2.14831859e-01 -5.05403638e-01 -7.59101927e-01 1.03795314e+00
2.08890960e-01 4.87641245e-01 1.23330390e+00 -1.75384909e-01
-2.34496862e-01 -4.02501434e-01 -1.98828959e+00 6.29229844e-01
7.29182541e-01 -7.12106049e-01 9.22956765e-02 5.02741694e-01
9.27180350e-01 -2.88830459e-01 -1.07495689e+00 3.32717538e-01
4.29610908e-01 -9.08324957e-01 9.04889703e-01 -9.29758251e-01
-6.54343218e-02 1.64531067e-01 -8.82923231e-02 -9.38384354e-01
2.42756661e-02 -9.74067330e-01 9.52732041e-02 1.12513864e+00
5.04024565e-01 -1.08105767e+00 1.33500361e+00 1.17457783e+00
4.60551083e-01 -9.48184669e-01 -1.26568365e+00 -4.99086976e-01
4.11848009e-01 -1.87194988e-01 8.20127726e-01 1.32607663e+00
1.68207392e-01 -4.62122768e-01 -5.30395567e-01 2.82191783e-01
8.21108937e-01 -3.96329425e-02 6.32274330e-01 -7.56525636e-01
-2.29560472e-02 -1.73624694e-01 -2.55851865e-01 -8.25117230e-01
1.69938803e-01 -1.07718349e+00 -3.56754869e-01 -1.17814672e+00
2.99318820e-01 -8.09693813e-01 -7.42946684e-01 8.57202768e-01
7.75662437e-02 2.04818845e-01 2.04086870e-01 1.75145224e-01
-5.54088712e-01 2.12187141e-01 1.30310082e+00 -3.30130935e-01
-3.76129270e-01 2.49969691e-01 -1.19072700e+00 3.58041257e-01
9.94685113e-01 -6.26909792e-01 -5.62399924e-01 -2.60798275e-01
-1.30014509e-01 1.90905407e-01 4.22617376e-01 -5.82728982e-01
4.72483456e-01 -4.42971401e-02 5.30044019e-01 -3.93019348e-01
-1.40680462e-01 -1.16279233e+00 9.66090336e-02 9.17263746e-01
-4.87108469e-01 -2.86101431e-01 1.30290166e-01 5.70937335e-01
3.50869559e-02 1.75638080e-01 6.82394087e-01 -3.33534360e-01
1.91015080e-01 6.97687626e-01 -2.34492376e-01 -1.92999691e-01
1.38679492e+00 -6.60798922e-02 -2.35455468e-01 -1.93933085e-01
-8.43714595e-01 6.38203740e-01 4.99551356e-01 2.95871217e-02
6.80275023e-01 -1.07710373e+00 -4.52596396e-01 4.79956239e-01
-3.02336901e-01 2.91790694e-01 6.14881635e-01 9.27964032e-01
-5.58923781e-01 4.57055956e-01 -2.58901417e-01 -3.40061158e-01
-1.32827306e+00 6.71733439e-01 4.53723073e-01 -4.52870041e-01
-7.42308736e-01 9.59850729e-01 4.74864066e-01 -6.98515058e-01
4.62434560e-01 -4.99553561e-01 1.50262058e-01 7.91895688e-02
3.24177831e-01 9.00907815e-02 -9.60944146e-02 -1.28755793e-01
-5.63815534e-01 -2.22479939e-01 -4.56675500e-01 2.50043601e-01
1.33837020e+00 -2.72616565e-01 -1.95048004e-01 -3.33826303e-01
1.46580172e+00 9.60488396e-04 -1.48929048e+00 -2.44462371e-01
-1.82557538e-01 -3.27652037e-01 3.58913578e-02 -7.40027189e-01
-1.61953688e+00 6.60999656e-01 8.25610816e-01 -4.18040082e-02
1.21095502e+00 -1.64947435e-02 1.16058731e+00 3.46916646e-01
5.15426874e-01 -9.67405438e-01 -2.04233795e-01 -2.52188563e-01
3.76510233e-01 -1.22757137e+00 1.26197517e-01 -2.37454280e-01
-8.49967599e-01 7.34744668e-01 5.43692887e-01 1.84426725e-01
9.91219878e-01 6.38528645e-01 3.77309859e-01 -1.45797580e-01
-4.53409791e-01 5.29330850e-01 -1.35381455e-02 6.37277842e-01
-1.17713697e-02 3.51324856e-01 -2.32999444e-01 1.01860666e+00
1.55558795e-01 9.17844176e-02 4.35715914e-01 1.10039437e+00
2.68870920e-01 -1.24247420e+00 -1.51038930e-01 6.82376862e-01
-1.01258433e+00 -1.07557820e-02 -4.39712107e-01 3.89154434e-01
1.84937403e-01 6.11520350e-01 -1.93693116e-01 -2.53794968e-01
2.70022154e-02 -1.56594217e-01 1.85479552e-01 -2.84568667e-01
-1.09455371e+00 2.51879245e-02 -2.46022955e-01 -7.70165622e-01
-1.51530027e-01 -6.85300469e-01 -1.30640125e+00 -1.60348892e-01
-1.61249101e-01 5.39747998e-02 4.70835507e-01 6.81333363e-01
7.42844760e-01 3.01361293e-01 9.39238250e-01 -2.77890295e-01
-7.90895283e-01 -3.36738706e-01 -6.68007731e-01 3.85796398e-01
5.20613134e-01 1.35067552e-01 -1.07723020e-01 -2.15289742e-01] | [6.1095805168151855, 6.469272136688232] |
9bd52fbd-7d11-44d8-85e2-e601d2108cb0 | casnet-investigating-channel-robustness-for | 2210.15370 | null | https://arxiv.org/abs/2210.15370v1 | https://arxiv.org/pdf/2210.15370v1.pdf | CasNet: Investigating Channel Robustness for Speech Separation | Recording channel mismatch between training and testing conditions has been shown to be a serious problem for speech separation. This situation greatly reduces the separation performance, and cannot meet the requirement of daily use. In this study, inheriting the use of our previously constructed TAT-2mix corpus, we address the channel mismatch problem by proposing a channel-aware audio separation network (CasNet), a deep learning framework for end-to-end time-domain speech separation. CasNet is implemented on top of TasNet. Channel embedding (characterizing channel information in a mixture of multiple utterances) generated by Channel Encoder is introduced into the separation module by the FiLM technique. Through two training strategies, we explore two roles that channel embedding may play: 1) a real-life noise disturbance, making the model more robust, or 2) a guide, instructing the separation model to retain the desired channel information. Experimental results on TAT-2mix show that CasNet trained with both training strategies outperforms the TasNet baseline, which does not use channel embeddings. | ['Hsin-Min Wang', 'Yu Tsao', 'Hung-Shin Lee', 'Yao-Fei Cheng', 'Fan-Lin Wang'] | 2022-10-27 | null | null | null | null | ['speech-separation'] | ['speech'] | [ 2.44409889e-01 -7.88849518e-02 -5.11053428e-02 4.44537401e-03
-1.11153901e+00 -4.57314491e-01 2.35442400e-01 -1.76844195e-01
-1.14945561e-01 3.70125264e-01 4.40316498e-01 -5.38961887e-01
1.98545277e-01 -1.68076202e-01 -6.05347812e-01 -9.12982643e-01
-9.67628211e-02 -3.15012620e-03 5.61013371e-02 -5.70312329e-02
-4.09033269e-01 3.18536401e-01 -1.28553534e+00 5.09198427e-01
7.72903979e-01 1.16499150e+00 2.34186903e-01 1.04611909e+00
1.14875287e-02 5.82683921e-01 -9.46091413e-01 -2.04905003e-01
4.36959863e-01 -9.49769258e-01 -2.28565067e-01 9.24308691e-03
6.89769685e-02 -4.90696579e-01 -7.38849878e-01 1.04906738e+00
1.06202638e+00 -1.29600525e-01 5.60355425e-01 -1.17833459e+00
-5.14706820e-02 1.02008772e+00 -2.64716838e-02 2.41661951e-01
8.35895464e-02 8.12679529e-02 1.04922748e+00 -6.71808362e-01
-2.12312024e-02 1.13044620e+00 7.58044600e-01 6.83082163e-01
-1.44932175e+00 -1.01936078e+00 2.83312760e-02 9.39144716e-02
-1.22307563e+00 -1.06345832e+00 8.62220883e-01 -2.48226419e-01
9.60609555e-01 1.93461046e-01 5.80970109e-01 1.38851607e+00
-1.92283377e-01 1.14205980e+00 7.77703285e-01 -5.63219666e-01
2.53255725e-01 2.51935542e-01 -2.75307298e-02 -4.76784632e-02
-4.61260587e-01 3.78986895e-01 -7.61572003e-01 -1.77997440e-01
6.13938391e-01 -3.61946076e-01 -7.29178429e-01 -5.35612851e-02
-8.23763311e-01 5.07307291e-01 3.61279845e-02 3.04111272e-01
-1.04703933e-01 1.90100018e-02 5.20917058e-01 7.59904861e-01
2.60180891e-01 4.34759438e-01 -4.55426425e-01 -4.72773850e-01
-1.01468229e+00 4.21431325e-02 8.23549867e-01 9.55852211e-01
1.73845589e-01 5.95620751e-01 -1.61237985e-01 9.26411390e-01
2.36206502e-01 5.23719370e-01 6.36851847e-01 -5.01062930e-01
6.86628938e-01 -1.92856550e-01 -2.43762195e-01 -3.74941617e-01
-1.91569477e-01 -7.07390428e-01 -6.50291562e-01 -1.00057364e-01
4.69994605e-01 -4.53900695e-01 -9.03934836e-01 1.69205165e+00
-6.31073341e-02 4.86673504e-01 5.64111173e-01 9.11566317e-01
7.04855204e-01 8.16725791e-01 -2.22331420e-01 -3.38896781e-01
9.53767419e-01 -1.25181520e+00 -1.17218792e+00 -2.16400519e-01
3.98402691e-01 -8.92393351e-01 6.84310615e-01 6.61792874e-01
-1.07137001e+00 -5.36507547e-01 -1.37814093e+00 3.57458562e-01
-5.06760255e-02 2.31015891e-01 2.35421687e-01 8.15429926e-01
-8.08927119e-01 5.62036574e-01 -6.19393885e-01 4.66169976e-03
1.08825609e-01 2.82491326e-01 -2.67509937e-01 8.97542313e-02
-1.60519397e+00 2.26027995e-01 1.28885135e-01 3.44902039e-01
-1.32214272e+00 -6.07727706e-01 -7.68890738e-01 3.64564896e-01
2.28326246e-01 -1.35659724e-01 1.47567558e+00 -1.11150730e+00
-1.97950351e+00 1.18652038e-01 -5.73845580e-02 -7.39349782e-01
4.35212106e-01 -4.47429955e-01 -9.21080709e-01 2.59656072e-01
-2.09440961e-01 2.09724158e-01 1.31642222e+00 -1.20527279e+00
-4.24650252e-01 -4.50345911e-02 -3.33443046e-01 3.05794716e-01
-4.95301455e-01 1.92291159e-02 -5.82754850e-01 -8.51962626e-01
6.64525777e-02 -8.88226092e-01 -1.65724680e-01 -3.67802024e-01
-8.51802528e-01 3.75800192e-01 6.91967010e-01 -9.82671857e-01
1.44644940e+00 -2.66722989e+00 1.90667123e-01 4.05064255e-01
6.90665394e-02 6.66094363e-01 -2.49435976e-01 6.53909326e-01
-2.38293514e-01 5.40382834e-03 -1.85613453e-01 -6.75835907e-01
2.48654261e-02 -2.05312386e-01 -4.90924776e-01 2.87264377e-01
1.88379213e-01 1.92538083e-01 -5.94376445e-01 3.95334773e-02
-6.63487092e-02 4.77346539e-01 -7.00365365e-01 7.95094192e-01
1.63886324e-01 5.17321944e-01 1.44174909e-02 3.77829701e-01
7.29054511e-01 4.25116539e-01 3.01494539e-01 -2.02431589e-01
1.13023080e-01 8.35757613e-01 -1.16161489e+00 1.64834929e+00
-5.59852540e-01 9.20116425e-01 6.62054360e-01 -7.86840439e-01
6.84518576e-01 9.63189423e-01 1.94573373e-01 -8.86256754e-01
3.75313342e-01 3.64527524e-01 4.40126479e-01 -4.81134892e-01
1.71700507e-01 -2.45942757e-01 3.67100798e-02 2.99618751e-01
2.06915602e-01 -1.14788897e-01 -2.89517522e-01 1.79030716e-01
1.17447937e+00 -5.36744952e-01 -1.56013414e-01 3.72349173e-02
2.57913142e-01 -5.72751045e-01 6.60426497e-01 5.58770835e-01
-3.66811424e-01 8.30312610e-01 7.39153743e-01 3.90769899e-01
-8.00868869e-01 -1.25292647e+00 4.00398001e-02 8.67832184e-01
9.06167179e-02 -6.72050178e-01 -8.81098092e-01 -3.52859497e-01
-5.60881495e-02 5.04992485e-01 -2.14855835e-01 -4.26393062e-01
-2.81920731e-01 -3.56321126e-01 1.10873878e+00 7.29659021e-01
3.92069042e-01 -5.47907829e-01 -1.09765306e-01 5.17934859e-01
-3.51627290e-01 -1.28743744e+00 -7.28132010e-01 8.18213105e-01
-5.78748167e-01 -5.72798610e-01 -6.11784339e-01 -6.16953909e-01
2.20417529e-01 3.47730517e-01 5.25900960e-01 -1.50208771e-01
1.01365209e-01 2.36003011e-01 -4.73654628e-01 -5.16393960e-01
-6.39150321e-01 3.06777917e-02 2.94515163e-01 3.09675336e-01
3.06609958e-01 -7.76350915e-01 -4.75711375e-01 3.28088462e-01
-7.68847704e-01 -2.24862412e-01 6.27577126e-01 9.01973426e-01
6.40648901e-02 4.81997997e-01 7.58370519e-01 -5.26265144e-01
8.69081855e-01 -3.12895358e-01 -2.48313755e-01 -2.66800135e-01
-4.15665895e-01 -2.49420777e-02 8.45391631e-01 -6.96278036e-01
-8.52471948e-01 -3.14060189e-02 -7.00828135e-01 -8.05030823e-01
-1.67280305e-02 3.55764329e-01 -9.00533676e-01 3.91509891e-01
3.85689914e-01 2.54388869e-01 1.65800780e-01 -7.11705267e-01
8.59754756e-02 1.43298805e+00 4.87361729e-01 -2.79588759e-01
7.18367219e-01 1.52802244e-01 -7.89620936e-01 -9.71992373e-01
-3.30849975e-01 -6.65008128e-01 -4.20135289e-01 1.47864699e-01
7.95582354e-01 -1.27098930e+00 -1.45639807e-01 6.52304709e-01
-1.16616488e+00 -6.51890099e-01 1.23431668e-01 7.96857893e-01
-2.34173700e-01 3.05891372e-02 -7.36928344e-01 -8.81971359e-01
-8.67857561e-02 -1.36240137e+00 8.15935194e-01 -9.19911563e-02
-1.70595333e-01 -7.14769483e-01 -1.98249910e-02 3.97708923e-01
4.79011327e-01 -3.54592770e-01 8.11189771e-01 -8.89747202e-01
-3.73046935e-01 -6.85157031e-02 1.44761398e-01 9.05632734e-01
-1.61971841e-02 -1.85460776e-01 -1.79350555e+00 -3.81164610e-01
1.80890694e-01 -1.80710237e-02 8.72891068e-01 3.24335784e-01
9.57145333e-01 -1.26005530e-01 -2.98171461e-01 6.66585922e-01
9.13000464e-01 5.64037859e-01 9.21267807e-01 -4.30766344e-02
6.83366358e-01 3.54495794e-01 2.31976658e-01 2.52007037e-01
-4.66650724e-02 6.70937419e-01 2.21157521e-01 -4.37575400e-01
-4.15802211e-01 -3.53775978e-01 7.30317175e-01 1.35539186e+00
4.70206380e-01 -3.88752580e-01 -5.59159994e-01 5.94146192e-01
-1.51457071e+00 -7.87844062e-01 6.76274449e-02 2.22036958e+00
8.42664838e-01 4.12271351e-01 2.20941886e-01 8.35091174e-01
6.21740222e-01 1.20758183e-01 -4.01445597e-01 -5.46732605e-01
1.29964566e-02 1.72925085e-01 4.40609038e-01 5.33162057e-01
-1.06362653e+00 6.19261086e-01 6.00860548e+00 1.01464009e+00
-1.62014079e+00 1.93125218e-01 2.04935327e-01 -1.64309531e-01
-1.43274650e-01 -9.68277454e-02 -5.23052096e-01 5.14994383e-01
1.31085849e+00 2.15959251e-01 6.18694723e-01 4.56145763e-01
1.77082211e-01 3.29393089e-01 -1.46032500e+00 1.26272881e+00
6.24183603e-02 -7.22460210e-01 -3.87567312e-01 1.26503902e-02
1.37001619e-01 -2.51035690e-01 2.42220506e-01 4.71132934e-01
-2.21294552e-01 -1.05668664e+00 1.15988064e+00 -1.41784698e-01
8.78545165e-01 -6.06653869e-01 5.41712880e-01 2.28078619e-01
-1.34601212e+00 -3.64418328e-01 5.80984131e-02 3.46978568e-02
1.86522111e-01 6.39184833e-01 -1.01789796e+00 4.05153781e-01
4.71942604e-01 4.87292469e-01 -3.50951776e-02 1.05429947e+00
-3.10555667e-01 1.17178798e+00 -5.61077148e-02 3.84914756e-01
9.91828293e-02 2.98531353e-01 8.30704689e-01 1.40692258e+00
1.06373459e-01 -1.18411727e-01 2.58502085e-03 5.71230590e-01
-1.27893820e-01 -1.62910163e-01 -5.42203963e-01 -3.17423016e-01
9.01664317e-01 8.25423241e-01 -3.30011100e-01 -4.40231375e-02
-4.08236772e-01 1.14638376e+00 -3.33165675e-02 6.56508565e-01
-9.72420394e-01 -7.55422294e-01 9.43901896e-01 5.44035994e-02
5.77375948e-01 -1.49880379e-01 -7.10698664e-02 -9.84329581e-01
-3.99721600e-02 -1.27501380e+00 -1.40723124e-01 -4.85951275e-01
-9.82883394e-01 9.02046919e-01 -4.23353344e-01 -1.53248274e+00
-1.59760341e-01 -4.96330708e-01 -7.70239234e-01 9.10679460e-01
-1.50761211e+00 -7.39295900e-01 2.86393344e-01 5.77138782e-01
6.32039428e-01 -2.59858876e-01 7.06226349e-01 9.29485738e-01
-8.37274551e-01 1.10304594e+00 2.16988862e-01 3.30228209e-01
8.93070221e-01 -1.17159104e+00 2.96566427e-01 1.01467967e+00
-3.54290269e-02 5.23855627e-01 8.35493505e-01 -2.30296433e-01
-1.27115846e+00 -1.04202688e+00 6.74198031e-01 -9.80270728e-02
5.38736939e-01 -9.62044060e-01 -9.27615702e-01 3.58789027e-01
3.94579887e-01 -3.40326667e-01 1.03859973e+00 -5.35144703e-03
-4.78987753e-01 -3.84387702e-01 -7.03106165e-01 5.11270285e-01
7.84431100e-01 -1.06580555e+00 -3.56885761e-01 -2.87017614e-01
9.81669188e-01 -4.00068283e-01 -5.78316987e-01 4.56015840e-02
6.58140838e-01 -9.82322097e-01 7.45003164e-01 -3.01515788e-01
1.34848401e-01 -1.58380583e-01 -3.56251508e-01 -1.82495356e+00
8.11484009e-02 -1.22289062e+00 -7.46782199e-02 1.48228133e+00
7.87213445e-01 -5.44834554e-01 3.00831437e-01 3.91812921e-01
-5.99541426e-01 -4.56151515e-01 -9.87197280e-01 -1.05058932e+00
1.32325560e-01 -7.12485969e-01 7.64873445e-01 6.79028869e-01
2.63484657e-01 4.78550971e-01 -5.31218588e-01 5.77406943e-01
2.14328736e-01 -3.72964084e-01 7.44846761e-01 -8.35659325e-01
-7.04821050e-01 -4.90021408e-01 -2.05899015e-01 -1.47896433e+00
2.02879429e-01 -5.74121058e-01 5.05112231e-01 -1.11641681e+00
-3.71329606e-01 -5.08795619e-01 -3.23749542e-01 7.85590261e-02
1.86158977e-02 -8.75657722e-02 3.20012301e-01 8.07662457e-02
-1.60340607e-01 7.54950941e-01 8.30378592e-01 -3.94230694e-01
-4.41969424e-01 2.13698059e-01 -8.25526655e-01 3.98691326e-01
6.01844192e-01 -4.55390036e-01 -5.69810748e-01 -4.60519940e-01
-2.34575108e-01 4.53363061e-01 -8.54923576e-02 -1.32264566e+00
3.14775735e-01 4.94507790e-01 -8.76326263e-02 -2.15706199e-01
6.93614244e-01 -1.16035390e+00 -6.55993670e-02 2.53519773e-01
-4.76303935e-01 -4.54108119e-01 5.62201202e-01 6.41704619e-01
-5.14827013e-01 1.83649242e-01 6.38373017e-01 4.55257833e-01
-2.00576380e-01 -1.80138685e-02 -7.71249652e-01 -1.55822828e-01
5.59133291e-01 -7.88585022e-02 6.21875785e-02 -5.89001596e-01
-8.72375488e-01 6.94354028e-02 -6.47864044e-02 2.82767504e-01
5.39871335e-01 -1.39239144e+00 -4.41262096e-01 6.13149941e-01
3.72411534e-02 -2.60875404e-01 1.95007354e-01 1.06364822e+00
-3.93525101e-02 3.60357136e-01 2.97976553e-01 -5.57397723e-01
-1.34209120e+00 3.02114218e-01 5.53821266e-01 4.60431576e-02
-5.63105881e-01 1.00508153e+00 2.91071773e-01 -1.74909994e-01
9.04590368e-01 -5.53031385e-01 -2.19722707e-02 -8.63753632e-02
6.66509509e-01 2.01341063e-01 3.78281444e-01 -4.53464448e-01
-2.12056085e-01 1.50732487e-01 -1.20437980e-01 -4.59036767e-01
9.44631815e-01 -2.91031718e-01 3.50608468e-01 5.17222524e-01
1.27936280e+00 3.04645807e-01 -1.34589469e+00 -2.01027170e-01
-5.50169349e-01 -4.45176333e-01 3.07014316e-01 -7.94259906e-01
-1.15679419e+00 1.39345694e+00 8.34895194e-01 3.97706628e-01
1.34963524e+00 -2.15056270e-01 1.32304358e+00 -3.92024443e-02
-1.32604361e-01 -1.15420353e+00 2.00096741e-01 5.13509512e-01
6.54142737e-01 -9.75703001e-01 -6.88344657e-01 -3.37801188e-01
-6.68793976e-01 9.96987104e-01 5.00046372e-01 2.81148255e-01
8.93861532e-01 6.18138850e-01 5.91888487e-01 1.57239273e-01
-7.93900549e-01 -2.48628944e-01 2.85741668e-02 8.21444273e-01
4.30233151e-01 1.80690423e-01 2.73374647e-01 1.18080831e+00
-1.92226037e-01 -5.08670151e-01 5.82336366e-01 7.74714887e-01
-2.81248480e-01 -1.15424240e+00 -6.08052611e-01 3.31009120e-01
-5.52675307e-01 -3.18385273e-01 -6.45398676e-01 3.02684963e-01
6.42290488e-02 1.28225470e+00 1.48280114e-01 -8.43683004e-01
4.48622823e-01 1.31803453e-01 1.71329811e-01 -5.08536458e-01
-5.31628072e-01 7.67939329e-01 2.80558944e-01 -2.96121091e-01
-3.07247266e-02 -5.52481771e-01 -1.18363976e+00 -8.36137831e-02
-7.22027540e-01 4.14586425e-01 6.46991193e-01 9.18079674e-01
3.81618619e-01 1.10004115e+00 7.93863833e-01 -9.15989041e-01
-5.99796295e-01 -1.00700045e+00 -1.07561064e+00 7.95456246e-02
9.79777217e-01 -4.74021167e-01 -6.26153648e-01 1.71207227e-02] | [14.81822681427002, 6.050320148468018] |
540b9445-9259-4b7b-9dfb-59405e5c3bb5 | false-negative-reduction-in-video-instance | 2106.14474 | null | https://arxiv.org/abs/2106.14474v1 | https://arxiv.org/pdf/2106.14474v1.pdf | False Negative Reduction in Video Instance Segmentation using Uncertainty Estimates | Instance segmentation of images is an important tool for automated scene understanding. Neural networks are usually trained to optimize their overall performance in terms of accuracy. Meanwhile, in applications such as automated driving, an overlooked pedestrian seems more harmful than a falsely detected one. In this work, we present a false negative detection method for image sequences based on inconsistencies in time series of tracked instances given the availability of image sequences in online applications. As the number of instances can be greatly increased by this algorithm, we apply a false positive pruning using uncertainty estimates aggregated over instances. To this end, instance-wise metrics are constructed which characterize uncertainty and geometry of a given instance or are predicated on depth estimation. The proposed method serves as a post-processing step applicable to any neural network that can also be trained on single frames only. In our tests, we obtain an improved trade-off between false negative and false positive instances by our fused detection approach in comparison to the use of an ordinary score value provided by the instance segmentation network during inference. | ['Kira Maag'] | 2021-06-28 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 5.35355031e-01 4.19788867e-01 1.58337966e-01 -8.37847948e-01
-7.03641117e-01 -3.09663355e-01 5.83860338e-01 6.89415157e-01
-9.06269789e-01 1.06360209e+00 -7.01183498e-01 -4.82978001e-02
-1.92536071e-01 -1.07735920e+00 -1.11802304e+00 -6.22128844e-01
-1.45859495e-01 5.47490954e-01 6.87968493e-01 3.21205437e-01
2.68623680e-01 6.53765142e-01 -2.05768561e+00 1.57384247e-01
8.98807943e-01 1.33919585e+00 2.76781887e-01 6.50399983e-01
-1.20398171e-01 6.63268328e-01 -7.99019635e-01 -5.82543194e-01
2.59809136e-01 -1.64956331e-01 -3.75176340e-01 2.85629719e-01
6.45343065e-01 -3.38710964e-01 2.15178102e-01 1.38560057e+00
8.57950076e-02 1.52111188e-01 7.07132697e-01 -1.40096140e+00
3.41307640e-01 5.37300110e-01 -3.64379942e-01 2.80762434e-01
1.58551689e-02 2.43940309e-01 7.87267387e-01 -5.64091742e-01
4.63105738e-01 8.09801519e-01 4.46061820e-01 1.16919436e-01
-1.10757804e+00 -4.25924391e-01 3.19974452e-01 5.31065643e-01
-1.26098442e+00 -3.18347543e-01 7.46118605e-01 -5.93204260e-01
6.64720535e-01 2.07802877e-01 6.97688699e-01 5.87359190e-01
5.66763356e-02 5.97573817e-01 7.74467647e-01 -3.12583923e-01
5.93549430e-01 4.54966486e-01 4.74185161e-02 5.76839328e-01
7.61056244e-01 6.07079528e-02 -2.79374212e-01 2.72925258e-01
4.72136676e-01 -3.59532803e-01 -6.22672699e-02 -3.64930302e-01
-7.11638689e-01 5.96271992e-01 3.88402283e-01 1.23978212e-01
-4.05504376e-01 2.01796904e-01 4.02493894e-01 4.24373802e-03
5.86928308e-01 3.40684623e-01 -4.31698322e-01 -6.01841584e-02
-1.25948966e+00 3.08807105e-01 6.40739560e-01 5.93515635e-01
8.58332038e-01 -4.52510603e-02 -1.75381362e-01 6.93229795e-01
-7.33420476e-02 1.59026667e-01 9.21570733e-02 -8.13247919e-01
2.68866718e-01 7.26421237e-01 2.32881501e-01 -1.08176041e+00
-4.35582161e-01 -7.33318269e-01 -6.50050223e-01 4.30908322e-01
7.43357897e-01 -1.21205978e-01 -7.24061012e-01 1.79553390e+00
5.29870093e-01 3.87939513e-01 -5.02342395e-02 7.35889256e-01
3.46000850e-01 3.61780643e-01 5.43792266e-03 -4.40214992e-01
1.15540349e+00 -4.98066425e-01 -6.17227852e-01 -1.27938300e-01
5.08388579e-01 -5.79900086e-01 5.09881258e-01 7.85024226e-01
-8.98022771e-01 -6.57892585e-01 -1.13574016e+00 2.97760755e-01
-4.67416048e-01 3.97095174e-01 2.05652669e-01 8.39494228e-01
-6.21829629e-01 8.04781735e-01 -8.48050892e-01 -1.56278923e-01
4.94954139e-01 3.25545371e-01 -2.09468529e-01 1.44276887e-01
-8.26635540e-01 9.08240616e-01 8.69219422e-01 3.68904084e-01
-7.19191968e-01 -4.61371660e-01 -8.03899825e-01 3.40334065e-02
7.64522076e-01 -2.93126494e-01 1.08468819e+00 -9.61025774e-01
-1.05047786e+00 7.44840324e-01 -7.24301860e-02 -1.11103153e+00
1.09968233e+00 -3.37061405e-01 -1.01653978e-01 2.37638757e-01
1.49246499e-01 6.68338478e-01 8.37435305e-01 -1.26339972e+00
-1.01272583e+00 -3.35593551e-01 3.64670724e-01 -2.39676207e-01
1.15969092e-01 -4.77478951e-01 -2.45068654e-01 -3.01463783e-01
2.29565129e-01 -7.16140807e-01 -2.41334975e-01 1.74692780e-01
-4.66905951e-01 -9.28556919e-02 7.07164645e-01 -4.71871585e-01
9.27485645e-01 -1.81346989e+00 -1.79955497e-01 3.76059175e-01
-1.54787764e-01 1.95757866e-01 2.55531937e-01 -1.70606479e-01
7.88724944e-02 -6.89329505e-02 -5.78275442e-01 -3.66836041e-01
-2.84684122e-01 3.26876283e-01 -9.31330211e-03 5.52492917e-01
6.99206471e-01 2.01363742e-01 -8.67286026e-01 -7.82010436e-01
6.97165906e-01 2.46860400e-01 -4.19940919e-01 1.37871712e-01
-4.97399598e-01 3.54496568e-01 -1.92069262e-01 4.05526817e-01
8.22743177e-01 9.12708864e-02 2.92006694e-02 -2.86865741e-01
-2.04454675e-01 -1.02939270e-01 -1.51619399e+00 1.33490252e+00
-4.93690014e-01 5.71400166e-01 -2.89754808e-01 -1.24948251e+00
9.64707255e-01 1.16395496e-01 4.03212756e-01 -5.00110209e-01
3.15846145e-01 2.93468714e-01 -9.18298289e-02 -4.86012042e-01
6.89626753e-01 1.08748011e-01 1.47788525e-01 -5.12819365e-02
8.11585337e-02 -1.25765830e-01 6.51297748e-01 -9.16308314e-02
9.60704625e-01 2.32763410e-01 3.79396021e-01 -1.59432441e-01
8.56010556e-01 5.49844429e-02 6.15817130e-01 9.65113342e-01
-1.61933750e-01 4.58919197e-01 7.64267921e-01 -3.47034931e-01
-1.05309868e+00 -8.16939473e-01 -3.78700972e-01 2.98649848e-01
1.99197188e-01 -2.61584911e-02 -9.65860128e-01 -6.70341074e-01
-2.18110636e-01 8.59008372e-01 -5.04880786e-01 -3.43045071e-02
-4.25411254e-01 -7.75690019e-01 3.89206529e-01 3.58233303e-01
6.49873376e-01 -7.32204974e-01 -1.19540060e+00 3.28291506e-01
-4.72341068e-02 -1.40974593e+00 3.30374002e-01 3.18221629e-01
-8.32935631e-01 -1.41917622e+00 -4.72022235e-01 -7.92661607e-02
7.63874352e-01 -1.75961539e-01 1.08761287e+00 1.02588892e-01
-3.84862751e-01 2.22426698e-01 -3.87240678e-01 -5.01536846e-01
-3.93994927e-01 3.15412669e-03 -1.82817318e-02 2.46691138e-01
2.04097539e-01 -4.52202618e-01 -3.99554372e-01 1.60317987e-01
-1.07377911e+00 -1.66330978e-01 6.36295795e-01 7.29625225e-01
7.34624565e-01 3.39318842e-01 4.69415367e-01 -7.88114369e-01
2.08289355e-01 -2.71292746e-01 -1.31124532e+00 3.30205977e-01
-3.87214303e-01 3.22559029e-01 6.45404518e-01 -2.62579352e-01
-1.04938006e+00 4.09805626e-01 -2.23195896e-01 -3.40201378e-01
-4.68932062e-01 1.87841132e-01 -1.25489622e-01 -8.76055658e-02
6.16166413e-01 -5.63205816e-02 -1.89368278e-01 -2.23460898e-01
8.38005096e-02 2.87374794e-01 6.20387852e-01 -5.58095694e-01
5.17775536e-01 4.67046440e-01 4.20316666e-01 -8.37579787e-01
-8.18792880e-01 -3.76777947e-01 -7.06832230e-01 -7.83165038e-01
7.49069095e-01 -5.17171144e-01 -7.00411320e-01 3.27553749e-01
-1.32885027e+00 1.51587263e-01 -3.09951216e-01 6.00958824e-01
-6.24305010e-01 3.69067669e-01 7.79280439e-02 -1.38418114e+00
2.52569109e-01 -1.23141599e+00 9.35796201e-01 2.35222965e-01
1.00537539e-01 -6.60191953e-01 -2.87457079e-01 4.11418416e-02
-5.67141473e-02 4.15310472e-01 4.99627322e-01 -7.84278095e-01
-8.52235198e-01 -5.79105675e-01 -2.89123118e-01 6.47435486e-01
-2.80797422e-01 2.15498164e-01 -1.14960742e+00 1.68014526e-01
-2.63465829e-02 -3.04095577e-02 9.80250239e-01 7.78397441e-01
1.39209604e+00 -2.41443172e-01 -5.13006262e-02 1.66536078e-01
1.71723509e+00 2.08090097e-01 4.76909816e-01 4.14700180e-01
2.14295864e-01 9.00500238e-01 1.04462039e+00 6.77217484e-01
-5.15347160e-02 6.78141475e-01 9.74662542e-01 3.48306090e-01
1.84076890e-01 1.72224894e-01 -1.96678583e-02 -1.44971117e-01
-1.38424352e-01 -2.90822029e-01 -6.80537760e-01 7.01096892e-01
-1.95263755e+00 -1.07979655e+00 -3.12588274e-01 2.84579539e+00
3.18888903e-01 7.90758610e-01 1.24737285e-01 5.70817351e-01
9.74581659e-01 -3.17893624e-02 -4.50189441e-01 -2.12843478e-01
5.60581349e-02 -2.45494563e-02 7.01658368e-01 6.64482653e-01
-1.11384785e+00 5.37588418e-01 4.99333096e+00 8.12801182e-01
-9.06534731e-01 -2.36193184e-02 8.52621078e-01 -2.57498592e-01
1.33243605e-01 -1.40651762e-01 -8.50379884e-01 5.42690456e-01
7.42170930e-01 1.07239068e-01 -5.45102619e-02 1.11091793e+00
2.39064619e-01 -7.68426657e-01 -1.22476172e+00 8.55745554e-01
-6.63790107e-02 -1.16371596e+00 -2.15653241e-01 -3.42579447e-02
4.00854260e-01 -3.02848518e-01 -8.37786570e-02 -9.32855997e-03
-2.61963338e-01 -7.59932816e-01 1.06176019e+00 6.94115639e-01
2.13031664e-01 -1.05249405e+00 1.02333570e+00 5.84592640e-01
-1.03105438e+00 -7.16276988e-02 -2.79973388e-01 -7.29577765e-02
4.13010597e-01 1.13325608e+00 -1.02339125e+00 5.95098436e-01
4.54120547e-01 2.26208135e-01 -4.73209590e-01 1.44003010e+00
-2.08125263e-01 3.30070674e-01 -7.76466250e-01 -1.04878977e-01
1.93157479e-01 -8.51781368e-02 7.91587651e-01 1.19111323e+00
4.00125831e-01 -2.48473942e-01 1.91688761e-02 9.77437258e-01
2.22862780e-01 -8.37030560e-02 -6.56249702e-01 4.04724717e-01
3.63389552e-01 1.22248030e+00 -1.11660206e+00 -5.64122617e-01
-5.63040338e-02 6.60237074e-01 1.25361934e-01 2.31398866e-02
-1.04652715e+00 -2.39343598e-01 3.20872784e-01 9.83430594e-02
3.35287690e-01 -1.45342685e-02 -5.00485480e-01 -8.71478260e-01
4.23638433e-01 -1.91774905e-01 1.77420884e-01 -5.42671859e-01
-9.82222795e-01 7.23873317e-01 2.94015497e-01 -1.37074673e+00
-6.45450473e-01 -7.29845643e-01 -3.11955541e-01 5.35027862e-01
-1.40962601e+00 -6.78853452e-01 -3.39852154e-01 1.60986185e-01
6.81477487e-01 2.24920332e-01 3.63312989e-01 3.55792582e-01
-5.93966842e-01 4.09440935e-01 -3.81206244e-01 -8.37998912e-02
2.13801071e-01 -1.11284852e+00 -1.35279819e-01 1.40753591e+00
1.59106448e-01 9.10472572e-02 1.27195871e+00 -7.11330652e-01
-4.21029896e-01 -1.14971161e+00 6.56331182e-01 -2.11071998e-01
2.98032463e-01 7.21239224e-02 -8.95953536e-01 3.23028058e-01
-3.66172075e-01 1.14559770e-01 2.07328096e-01 -5.44063933e-02
7.27803484e-02 -3.51174444e-01 -1.42918313e+00 3.54394317e-01
7.82607377e-01 -2.09586769e-01 -3.76728445e-01 2.51449078e-01
3.80031377e-01 -4.15381819e-01 -7.26759970e-01 7.04817772e-01
5.41194022e-01 -1.46574783e+00 8.62824202e-01 -1.79643840e-01
4.50476795e-01 -5.17494440e-01 -1.84417814e-01 -9.40056920e-01
3.84465814e-01 7.02382624e-02 -3.36250775e-02 1.08031285e+00
4.32649493e-01 -3.43555361e-01 9.42863286e-01 8.05124521e-01
-8.01798031e-02 -5.80028057e-01 -1.27853262e+00 -8.05303335e-01
-5.55819154e-01 -1.00295103e+00 3.71108979e-01 4.57412720e-01
-5.58106422e-01 -2.73816973e-01 -2.40812048e-01 6.38380527e-01
1.00091350e+00 -2.90713549e-01 7.69450963e-01 -1.38437319e+00
-4.64042515e-01 -3.17549735e-01 -9.12645876e-01 -6.21420860e-01
-3.63808721e-02 -2.26462245e-01 2.77872831e-01 -1.15726507e+00
-1.64256454e-01 -3.17365259e-01 -2.00709745e-01 1.07617855e-01
-2.25493379e-04 3.52524221e-01 5.94845377e-02 -1.75764099e-01
-4.12825316e-01 2.41769373e-01 7.31599689e-01 1.62901655e-02
-7.62042999e-02 4.05026883e-01 1.79532822e-02 1.03657758e+00
7.52681851e-01 -6.07050478e-01 -3.49680305e-01 -1.08733125e-01
4.44551051e-01 1.09538399e-01 7.74191380e-01 -1.52273250e+00
2.76133895e-01 1.29052559e-02 3.94131780e-01 -8.66438687e-01
5.65503895e-01 -1.12484932e+00 1.30705833e-01 5.29340804e-01
-2.94297189e-01 -2.05189407e-01 1.35863870e-01 7.59290278e-01
-3.07624966e-01 -8.30764413e-01 8.21702003e-01 -3.30863625e-01
-8.68878305e-01 -4.74932306e-02 -2.30631083e-01 -4.71065909e-01
1.19079280e+00 -5.97100198e-01 -8.99235234e-02 -1.73785210e-01
-8.23976755e-01 8.93939566e-03 2.36467466e-01 5.80241121e-02
5.47707558e-01 -8.81951511e-01 -3.66262078e-01 -8.53298530e-02
1.62745148e-01 2.50969559e-01 3.98305535e-01 8.91791284e-01
-5.86333573e-01 2.04408661e-01 -2.41891995e-01 -9.31021333e-01
-1.37971067e+00 3.38550508e-01 3.75508815e-01 -2.30688304e-01
-3.24647695e-01 6.86023474e-01 -6.16456605e-02 -2.04643998e-02
3.57308626e-01 -6.53266668e-01 -3.56900722e-01 1.23637997e-01
4.32516098e-01 4.67354119e-01 3.15360963e-01 -5.89060068e-01
-2.64617562e-01 3.21361184e-01 1.12013936e-01 -2.76471406e-01
1.17836285e+00 -9.04703736e-02 1.22035563e-01 6.00368857e-01
7.51792789e-01 -2.54580230e-01 -1.46884978e+00 1.46493182e-01
1.42060757e-01 -5.49110293e-01 -2.77702305e-02 -6.71306551e-01
-1.04354727e+00 7.91489780e-01 8.04749250e-01 4.34753209e-01
1.13213813e+00 -1.63046360e-01 2.30815202e-01 5.57548285e-01
4.23828453e-01 -1.39061522e+00 -2.56265223e-01 1.14806637e-01
5.33825219e-01 -1.45735395e+00 -1.17678801e-02 -4.20109600e-01
-3.69464248e-01 1.20720530e+00 6.80984557e-01 -2.30240241e-01
4.17409658e-01 2.02084064e-01 -3.78261328e-01 1.24703534e-01
-5.20547807e-01 -5.24925470e-01 1.54441908e-01 4.27238584e-01
1.13098599e-01 1.38590974e-03 -5.98582923e-01 2.74138480e-01
-1.83152154e-01 1.67014003e-01 6.28080964e-01 6.83086872e-01
-7.58544326e-01 -7.89206684e-01 -4.96152133e-01 5.70128441e-01
-4.77849156e-01 2.81506330e-01 -7.67254606e-02 9.08159792e-01
6.59024179e-01 9.74069118e-01 2.64806569e-01 -3.47250193e-01
1.93151116e-01 -2.89489124e-02 5.68963826e-01 -3.46670836e-01
-1.92705289e-01 -2.53186435e-01 3.99939328e-01 -5.94680190e-01
-7.89809883e-01 -6.81587756e-01 -1.00474322e+00 5.49794734e-02
-5.77738523e-01 2.89169829e-02 1.01005602e+00 1.22429800e+00
-1.27559885e-01 6.76067650e-01 2.74737805e-01 -8.72420847e-01
-2.35589415e-01 -7.33600974e-01 -4.15683120e-01 2.13020176e-01
3.16438824e-01 -9.27784383e-01 -3.46679896e-01 6.89843595e-02] | [8.025816917419434, -0.8582249283790588] |
bfefc8c6-2a3a-4792-918b-3d04e02c60f4 | data-incubation-synthesizing-missing-data-for | 2110.07040 | null | https://arxiv.org/abs/2110.07040v1 | https://arxiv.org/pdf/2110.07040v1.pdf | Data Incubation -- Synthesizing Missing Data for Handwriting Recognition | In this paper, we demonstrate how a generative model can be used to build a better recognizer through the control of content and style. We are building an online handwriting recognizer from a modest amount of training samples. By training our controllable handwriting synthesizer on the same data, we can synthesize handwriting with previously underrepresented content (e.g., URLs and email addresses) and style (e.g., cursive and slanted). Moreover, we propose a framework to analyze a recognizer that is trained with a mixture of real and synthetic training data. We use the framework to optimize data synthesis and demonstrate significant improvement on handwriting recognition over a model trained on real data only. Overall, we achieve a 66% reduction in Character Error Rate. | ['Oncel Tuzel', 'Ryan Dixon', 'Thomas Deselaers', 'Adrien Delaye', 'Youssouf Chherawala', 'Martin Bresler', 'Jen-Hao Rick Chang'] | 2021-10-13 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 3.76723140e-01 1.21332131e-01 -3.36128771e-01 -3.44673961e-01
-6.47335052e-01 -1.05612946e+00 5.51371396e-01 -4.92968291e-01
-7.57142575e-03 5.34045577e-01 3.23909372e-01 -3.11586440e-01
5.75602531e-01 -9.28248286e-01 -9.30837214e-01 -3.11544359e-01
7.27186322e-01 6.33641005e-01 -3.05543095e-01 -2.41650730e-01
3.28453004e-01 8.37690830e-01 -1.31087315e+00 5.19680500e-01
1.02825165e+00 3.21958840e-01 1.13353223e-01 1.13839245e+00
-2.95390517e-01 6.66902006e-01 -1.12308502e+00 -6.09229624e-01
4.82721776e-01 -6.16214097e-01 -4.76501763e-01 7.62908041e-01
9.49818730e-01 -7.49705076e-01 -8.16628873e-01 9.30097461e-01
4.63540703e-01 -2.84890413e-01 6.03950083e-01 -7.25366533e-01
-1.19388127e+00 8.37545693e-01 -1.68948606e-01 -5.18779039e-01
3.38808745e-01 4.60102648e-01 9.20298457e-01 -9.25807655e-01
9.05104756e-01 1.06121433e+00 3.66071641e-01 9.60659683e-01
-1.44726098e+00 -4.17882383e-01 5.76682352e-02 -4.23872471e-01
-1.07893062e+00 -4.37541068e-01 6.41009271e-01 -2.33834654e-01
6.71312511e-01 3.66349995e-01 4.28679019e-01 1.54637253e+00
-1.08595699e-01 1.11272311e+00 1.03251326e+00 -7.02831209e-01
8.60033184e-02 1.67619288e-01 2.78439522e-01 9.45333302e-01
2.34353557e-01 -3.10854167e-02 -5.19328117e-01 -4.89097554e-03
1.01131225e+00 -6.24287091e-02 -9.80372429e-02 3.80246267e-02
-1.17560387e+00 5.28289199e-01 -5.31888269e-02 -2.46728733e-02
1.86810996e-02 9.79405493e-02 5.59890307e-02 3.57349932e-01
1.44819558e-01 8.29578519e-01 -1.30617963e-02 -3.95612597e-01
-1.15807617e+00 3.44848782e-01 1.22650421e+00 1.62504399e+00
7.33576953e-01 4.78294522e-01 -3.83762747e-01 9.38537598e-01
-5.59012592e-02 1.16890323e+00 5.48814237e-01 -7.56070137e-01
8.02860498e-01 5.73598146e-01 3.32231849e-01 -5.94164848e-01
1.62836909e-01 -2.67092913e-01 -5.07791936e-01 1.23833396e-01
5.68009257e-01 -2.88273752e-01 -1.50657499e+00 1.27097726e+00
-2.73008078e-01 -2.33722761e-01 1.21071644e-01 7.50853360e-01
3.69836509e-01 6.36808693e-01 -5.25100529e-01 1.84738010e-01
1.02764428e+00 -1.27602983e+00 -7.48784363e-01 -2.57286698e-01
3.68027717e-01 -9.19481695e-01 1.61165440e+00 3.52986366e-01
-1.11460388e+00 -7.25797474e-01 -1.20829988e+00 -1.47574738e-01
-2.17071578e-01 7.71001399e-01 3.77244294e-01 1.07558393e+00
-8.91602695e-01 7.73173273e-01 -8.25657666e-01 -1.46737769e-01
4.30126458e-01 1.01766922e-01 -1.32930174e-01 -7.75446743e-02
-5.88626206e-01 4.92171943e-01 4.80143465e-02 -2.51516134e-01
-1.02384937e+00 -5.89722633e-01 -4.79060143e-01 -7.61595294e-02
1.70114815e-01 -2.41530553e-01 1.23842788e+00 -9.64995205e-01
-1.91843438e+00 7.58672118e-01 -2.78761029e-01 -1.23773374e-01
8.86814356e-01 -1.73291788e-01 -4.57959116e-01 7.13074133e-02
-2.41531521e-01 1.45152524e-01 1.05939353e+00 -1.29453135e+00
-3.67763728e-01 -1.17586963e-01 -3.07832927e-01 -8.30878466e-02
-5.78801692e-01 -5.80812730e-02 -5.13041377e-01 -1.13514221e+00
-9.34401229e-02 -1.24239063e+00 1.43394858e-01 -5.10749267e-03
-7.80580103e-01 2.85573244e-01 7.83396661e-01 -1.01902628e+00
1.07303584e+00 -1.92986596e+00 3.28981042e-01 3.69731069e-01
8.93614739e-02 5.32705545e-01 -4.02155966e-01 5.18224478e-01
4.17867780e-01 4.71730888e-01 -6.89988807e-02 -4.35329497e-01
1.07606962e-01 3.01576763e-01 -9.18325007e-01 8.60041380e-02
3.79369617e-01 9.96695340e-01 -8.33116233e-01 -1.88237488e-01
-2.22489506e-01 1.80104077e-01 -6.65845811e-01 5.05335689e-01
-5.64847708e-01 1.55616596e-01 -4.17502880e-01 1.00976694e+00
4.55964059e-01 -2.72782207e-01 4.68350381e-01 -1.67908706e-02
-5.90439793e-03 1.64928392e-01 -1.00979781e+00 1.46740997e+00
-5.34593284e-01 9.72000003e-01 -1.67314813e-01 -8.89449418e-02
1.54332232e+00 6.37474209e-02 -1.54420629e-01 -2.73236156e-01
-4.53723036e-02 4.03370738e-01 -1.62786484e-01 -3.85726005e-01
8.93983543e-01 1.36788428e-01 4.94508892e-02 1.05344903e+00
9.76658016e-02 -2.70174772e-01 2.36294895e-01 1.95724621e-01
1.16137254e+00 1.56635314e-01 -3.03202182e-01 -3.33342850e-02
1.90354325e-02 4.61898558e-02 7.71760345e-02 1.22780883e+00
7.27248266e-02 8.65865648e-01 4.81587827e-01 -3.73089254e-01
-1.92162323e+00 -1.11526072e+00 8.04484263e-02 1.02216041e+00
-1.71706453e-01 -2.99813390e-01 -9.01410937e-01 -7.93274403e-01
2.78669298e-01 8.45337391e-01 -4.57714766e-01 -1.82149224e-02
-9.62679088e-01 -4.63582128e-01 1.18614411e+00 7.53111959e-01
3.85143459e-01 -9.04619217e-01 -4.04484123e-02 1.61022276e-01
5.43761775e-02 -9.90201652e-01 -8.81876409e-01 -1.62135601e-01
-7.73834109e-01 -4.88377035e-01 -9.10961390e-01 -9.22948301e-01
9.88302171e-01 9.41593386e-03 1.06690943e+00 1.77951112e-01
-1.81406870e-01 3.89307022e-01 -4.35307115e-01 -2.45598868e-01
-1.15671444e+00 5.25087953e-01 -8.70720968e-02 1.94476247e-01
5.51341251e-02 -3.10689658e-01 -1.65517494e-01 3.71990710e-01
-9.67593968e-01 3.05446982e-01 6.18986964e-01 1.12717140e+00
2.48971254e-01 -5.46597600e-01 4.04255897e-01 -1.19636905e+00
8.42464209e-01 7.09791332e-02 -5.18276453e-01 7.08197117e-01
-7.50803173e-01 4.62356418e-01 1.01229751e+00 -8.66114795e-01
-1.22505212e+00 1.16221964e-01 2.80075856e-02 -5.55501878e-01
-1.48956999e-01 2.13419244e-01 -1.52993739e-01 -7.62583539e-02
9.29687202e-01 6.65803134e-01 2.70727158e-01 -5.37695765e-01
5.64215302e-01 1.22613549e+00 5.93754709e-01 -9.95493174e-01
7.42569685e-01 3.50744039e-01 -4.51046973e-01 -7.34314799e-01
-4.17162776e-01 1.77763447e-01 -5.78627169e-01 -9.78335142e-02
2.98389435e-01 -7.12366521e-01 -4.13412035e-01 8.00428212e-01
-1.17119539e+00 -7.91235924e-01 -3.62851560e-01 1.69897079e-02
-7.22151995e-01 2.44162649e-01 -8.27999651e-01 -6.26165509e-01
-3.09418797e-01 -1.12781620e+00 1.20093369e+00 3.03977400e-01
-2.69606501e-01 -8.02893043e-01 4.27813120e-02 2.24859685e-01
3.69807392e-01 6.83412999e-02 7.96797216e-01 -6.87650204e-01
-9.56629336e-01 -4.77034897e-01 -2.24674106e-01 4.17087764e-01
5.36994755e-01 6.12284422e-01 -9.03840482e-01 -3.20742518e-01
-4.85582739e-01 -4.13142502e-01 7.04844713e-01 -3.51587653e-01
9.87388730e-01 -6.44974589e-01 -1.04528993e-01 8.41684520e-01
1.22198462e+00 2.18788490e-01 9.26633418e-01 -4.70730364e-02
1.00734091e+00 2.86237895e-01 2.75061280e-01 5.37027597e-01
-1.04962237e-01 5.25282085e-01 -2.65933007e-01 1.07136227e-01
-6.14792466e-01 -9.65418398e-01 3.51313859e-01 6.74302757e-01
-1.98805988e-01 -4.64907944e-01 -9.30881917e-01 4.00535256e-01
-1.58630657e+00 -9.37347412e-01 3.25786024e-01 2.00747538e+00
1.48485661e+00 2.41204966e-02 2.10690767e-01 -1.60149962e-01
8.05090845e-01 1.36747569e-01 -8.72395039e-01 -4.27529544e-01
-3.95797759e-01 3.45396042e-01 7.43988454e-01 7.08484650e-01
-7.64903367e-01 1.34510875e+00 7.52856588e+00 7.04431474e-01
-1.42137218e+00 -5.80237806e-01 5.79143465e-01 -1.48657203e-01
-6.80995643e-01 -1.64827034e-01 -1.23971403e+00 6.75112724e-01
9.17458415e-01 -3.05986345e-01 1.07750082e+00 8.47602725e-01
1.04256131e-01 5.80741167e-01 -1.27238786e+00 8.33108604e-01
4.46226299e-01 -1.65940273e+00 4.66057748e-01 1.25081405e-01
1.16406810e+00 -3.83121282e-01 2.76396990e-01 1.10882141e-01
6.58541203e-01 -1.10765970e+00 9.86147404e-01 7.37405241e-01
1.19529903e+00 -3.32633048e-01 1.62168443e-01 3.08571517e-01
-4.50368255e-01 1.20687291e-01 -3.39116693e-01 1.80639476e-01
-2.92920042e-02 4.39111590e-01 -1.11682296e+00 8.05061311e-02
-1.71222165e-01 7.12419152e-01 -6.64362490e-01 4.85082746e-01
-4.16056782e-01 8.70958328e-01 4.16076668e-02 -6.13815188e-01
-1.15952067e-01 -2.61428893e-01 4.55969095e-01 1.39221156e+00
4.23210680e-01 -1.91556700e-02 9.19090882e-02 1.25764298e+00
-5.11732638e-01 -3.68022695e-02 -6.09456837e-01 -7.30128348e-01
5.55619359e-01 9.65850711e-01 -4.83648330e-01 -7.56580830e-01
-9.88951847e-02 1.26495302e+00 4.84362990e-01 5.66088617e-01
-6.99703693e-01 -6.53358579e-01 6.17207944e-01 9.10237059e-02
2.27579683e-01 -4.96017516e-01 -7.99366832e-01 -1.62950528e+00
8.25746953e-02 -1.50181854e+00 -3.88207018e-01 -7.39593863e-01
-1.07578492e+00 5.65182924e-01 -5.85654616e-01 -1.09905934e+00
-3.17299366e-01 -9.07386839e-01 -4.01686817e-01 1.12344587e+00
-8.91658545e-01 -1.22537220e+00 -3.75647575e-01 1.00405298e-01
6.30324244e-01 -7.21484303e-01 8.13650429e-01 8.88379142e-02
-5.20678163e-01 1.20761526e+00 3.36166710e-01 7.23140478e-01
7.82435477e-01 -1.38739145e+00 1.02154696e+00 8.60031128e-01
4.45054650e-01 8.55035186e-01 6.49873912e-01 -9.24361706e-01
-1.71660876e+00 -1.16017449e+00 6.34880483e-01 -6.53676271e-01
6.58649564e-01 -4.86830473e-01 -6.98252439e-01 8.04223120e-01
-1.24529167e-03 -3.08062613e-01 6.14112675e-01 -1.39137357e-01
-7.49795079e-01 1.31650239e-01 -1.09666586e+00 1.08658493e+00
1.27075541e+00 -7.45102704e-01 -3.17163229e-01 3.99223894e-01
4.95002925e-01 -8.70517552e-01 -8.58276963e-01 -2.60051906e-01
9.48195338e-01 -5.18172622e-01 7.28327453e-01 -1.06699741e+00
8.96041036e-01 -1.52348146e-01 -4.78484370e-02 -1.45942354e+00
-2.52201647e-01 -8.97203565e-01 -2.65279949e-01 1.34548187e+00
5.72546780e-01 -5.31950295e-01 1.15118468e+00 8.45282614e-01
-1.59495369e-01 -4.15797502e-01 -7.48606399e-02 -1.09985614e+00
1.51123673e-01 1.18843034e-01 7.84846008e-01 6.37534738e-01
-4.11224645e-03 1.65177301e-01 -7.50024557e-01 -3.21527272e-02
5.63892484e-01 4.74002153e-01 1.10113835e+00 -5.87581158e-01
-5.71028113e-01 -3.52226496e-01 -1.65949568e-01 -1.54426491e+00
2.14374423e-01 -1.02256918e+00 9.89643112e-02 -1.18062842e+00
6.28262907e-02 -5.86220264e-01 4.16616142e-01 5.84132910e-01
-2.67886460e-01 1.47168115e-01 5.04644990e-01 4.35404927e-01
-4.24185395e-02 1.50160432e-01 1.58456671e+00 -4.41503644e-01
-2.23840386e-01 2.69672996e-03 -6.74158931e-01 1.82813257e-01
9.84923184e-01 -2.60782629e-01 -1.12457491e-01 -6.53202713e-01
5.85809303e-03 -1.49299633e-02 8.99003353e-03 -5.96355975e-01
1.20407797e-01 -3.90252501e-01 3.59316617e-01 -3.11671019e-01
8.53907838e-02 -3.22765946e-01 -6.15801960e-02 2.96794474e-01
-8.74322832e-01 -2.00375259e-01 2.58195326e-02 4.33911681e-01
2.75808815e-02 -5.55627048e-01 5.55527806e-01 -1.37838393e-01
-2.12439284e-01 7.50337243e-02 -3.26564670e-01 5.54080717e-02
5.53619683e-01 -1.46733344e-01 -7.89931297e-01 -3.93966764e-01
-6.78438962e-01 -3.00634742e-01 9.87907946e-01 4.07204211e-01
6.40358925e-01 -1.43886459e+00 -7.69786119e-01 6.95434511e-01
2.51693614e-02 -4.01762396e-01 -3.26858699e-01 -2.28052050e-01
-1.02314591e+00 2.30409205e-01 -3.05139542e-01 -2.75687814e-01
-1.15814900e+00 3.56331229e-01 2.42034316e-01 -1.58734426e-01
-4.02105987e-01 4.54374671e-01 -4.83666897e-01 -4.55549031e-01
2.53087401e-01 -4.45443869e-01 5.12526870e-01 -3.22749168e-01
7.35607624e-01 4.29422051e-01 -1.30144274e-02 -1.98785335e-01
2.21329287e-01 3.17847610e-01 -2.69409001e-01 -4.20709968e-01
1.10561574e+00 4.69060540e-01 1.43587828e-01 2.36595303e-01
1.07407176e+00 5.82047462e-01 -1.63785040e+00 -9.90283862e-02
-1.82509005e-01 -8.09838951e-01 -4.80882019e-01 -8.78074408e-01
-9.38608110e-01 6.76442444e-01 1.56326443e-01 2.34755829e-01
6.16335213e-01 -2.47562051e-01 6.55753851e-01 7.92199552e-01
3.80950630e-01 -1.28237665e+00 4.24835294e-01 6.16766274e-01
9.69279170e-01 -9.62886393e-01 -1.91992611e-01 -3.63715947e-01
-8.69576216e-01 1.56316483e+00 3.13258439e-01 -3.61389607e-01
2.34934032e-01 7.72792995e-01 1.97459862e-01 3.06726426e-01
-6.64984524e-01 1.40303597e-01 2.34264061e-01 5.70031822e-01
3.82853299e-01 3.67977738e-01 4.35335236e-03 4.22253698e-01
-5.42194366e-01 9.25419852e-02 1.00103807e+00 8.20577860e-01
-3.94637942e-01 -1.42106009e+00 -3.80162656e-01 6.15338564e-01
-1.58967078e-01 -6.70407936e-02 -7.59026408e-01 4.32467252e-01
-3.65096807e-01 6.76172674e-01 1.34701058e-01 -7.04528451e-01
3.65047097e-01 3.02630991e-01 7.04322577e-01 -8.98050845e-01
-5.76949596e-01 -1.30759686e-01 1.73295081e-01 -1.15195751e-01
1.92589074e-01 -6.26432002e-01 -9.25355077e-01 -5.79147398e-01
-9.41599458e-02 -3.87758315e-01 6.26430035e-01 6.35525584e-01
4.29160506e-01 3.06237549e-01 9.17162716e-01 -4.50719386e-01
-1.03042626e+00 -7.35918343e-01 -6.57227516e-01 5.76255202e-01
2.24101469e-01 4.49728332e-02 -1.99763030e-01 5.87008119e-01] | [11.645824432373047, 0.05353802442550659] |
57f997a6-6cd3-459a-bc79-d636e3b55b5b | lightweight-alpha-matting-network-using | 2210.07760 | null | https://arxiv.org/abs/2210.07760v1 | https://arxiv.org/pdf/2210.07760v1.pdf | Lightweight Alpha Matting Network Using Distillation-Based Channel Pruning | Recently, alpha matting has received a lot of attention because of its usefulness in mobile applications such as selfies. Therefore, there has been a demand for a lightweight alpha matting model due to the limited computational resources of commercial portable devices. To this end, we suggest a distillation-based channel pruning method for the alpha matting networks. In the pruning step, we remove channels of a student network having fewer impacts on mimicking the knowledge of a teacher network. Then, the pruned lightweight student network is trained by the same distillation loss. A lightweight alpha matting model from the proposed method outperforms existing lightweight methods. To show superiority of our algorithm, we provide various quantitative and qualitative experiments with in-depth analyses. Furthermore, we demonstrate the versatility of the proposed distillation-based channel pruning method by applying it to semantic segmentation. | ['Donghyeon Cho', 'Jinsun Park', 'Donggeun Yoon'] | 2022-10-14 | null | null | null | null | ['image-matting'] | ['computer-vision'] | [ 4.57603186e-01 3.26872289e-01 -3.41001630e-01 -3.70075881e-01
-2.55981982e-01 -1.10578932e-01 2.61878341e-01 2.90205032e-01
-5.44998884e-01 7.82103956e-01 -3.84945027e-03 -5.75599253e-01
-3.02847885e-02 -1.07528973e+00 -9.87096131e-01 -5.92368364e-01
1.50259614e-01 9.15515870e-02 4.06144708e-01 4.57957089e-02
6.05895147e-02 1.53176785e-01 -1.52670813e+00 5.31183422e-01
1.26667011e+00 1.08257484e+00 2.35743210e-01 2.74158478e-01
-2.90565670e-01 3.72321129e-01 -5.83039999e-01 -6.60641968e-01
3.06463301e-01 -3.47535729e-01 -4.83392388e-01 -1.91416025e-01
3.64431441e-01 -3.43189210e-01 -3.78994018e-01 1.22531414e+00
3.44855130e-01 1.84036344e-01 1.73718005e-01 -1.18640590e+00
7.04808801e-04 1.23666692e+00 -9.09598947e-01 1.17091715e-01
-1.67548109e-03 -2.72641808e-01 9.97031569e-01 -5.04840195e-01
2.85122037e-01 1.37310278e+00 5.44134200e-01 3.67811888e-01
-9.50485051e-01 -1.07262051e+00 5.72474420e-01 2.92168975e-01
-1.05341542e+00 -3.98981534e-02 7.10466743e-01 -1.09304609e-02
5.92914522e-01 1.80003479e-01 9.36121464e-01 9.72707450e-01
-1.55950233e-01 1.26903319e+00 1.17278683e+00 -5.07685363e-01
2.11255670e-01 3.99580538e-01 1.40881509e-01 8.69600296e-01
5.51593125e-01 -2.70274401e-01 -6.57529652e-01 4.84713800e-02
6.01734579e-01 4.07746956e-02 -4.05830182e-02 -2.71854401e-01
-7.95782149e-01 7.84379780e-01 5.90655029e-01 1.27529604e-02
1.71949267e-02 4.33189303e-01 4.94301409e-01 1.86620981e-01
5.97353578e-01 1.30534470e-01 -3.75653148e-01 -3.66457433e-01
-1.10978985e+00 5.02818637e-02 6.24681294e-01 9.89556432e-01
7.35593438e-01 5.36512174e-02 -1.52824307e-02 7.96512485e-01
3.97041440e-01 2.90255576e-01 4.25946385e-01 -6.41993225e-01
5.84585369e-01 9.71826673e-01 -5.52613080e-01 -8.32962215e-01
-7.56793618e-02 -6.38405383e-01 -9.82554734e-01 -2.68613137e-02
2.78139830e-01 -1.88646078e-01 -1.14304829e+00 1.61545348e+00
4.99297827e-01 4.93302882e-01 -1.37486726e-01 5.60849845e-01
6.12076700e-01 6.45338178e-01 1.32288085e-02 -5.19668534e-02
1.28516173e+00 -1.08679950e+00 -4.82706189e-01 -3.19776945e-02
6.95819795e-01 -6.84572458e-01 1.06624317e+00 6.04068816e-01
-1.07223856e+00 -3.62815440e-01 -1.26106107e+00 -7.94315636e-02
-2.54243404e-01 1.27032071e-01 1.03373492e+00 1.03228414e+00
-5.92556596e-01 6.87451005e-01 -1.02757823e+00 -2.62414336e-01
8.82469118e-01 4.98288810e-01 1.25089675e-01 7.35216774e-03
-9.55878615e-01 4.67243850e-01 5.88501692e-01 5.40766381e-02
-7.13858962e-01 -6.74943149e-01 -7.09200263e-01 3.83997500e-01
5.30553937e-01 -6.66374445e-01 1.36113966e+00 -1.01167774e+00
-1.52242446e+00 4.20899093e-01 -2.36930735e-02 -8.41410339e-01
6.62997007e-01 -4.91599202e-01 -6.45918548e-02 2.45475724e-01
-5.27780019e-02 8.52992952e-01 8.19764912e-01 -1.16097462e+00
-8.95613730e-01 -2.75418669e-01 3.61396015e-01 2.37046078e-01
-7.68668532e-01 -3.69922459e-01 -5.62394142e-01 -9.04750168e-01
2.93497592e-01 -8.26757669e-01 -2.16360271e-01 -5.62366396e-02
-7.38662779e-01 7.38015100e-02 1.01049542e+00 -4.10059690e-01
1.47076929e+00 -2.28748965e+00 -1.21232852e-01 5.69096506e-01
3.88777077e-01 3.19361418e-01 1.36059582e-01 -6.07460514e-02
1.65425271e-01 2.99887925e-01 -2.64029235e-01 -3.81371021e-01
-3.54865909e-01 3.84878844e-01 -3.88863921e-01 6.19713031e-02
-2.21956760e-01 5.48849523e-01 -7.06082165e-01 -6.32370591e-01
1.24481730e-01 3.50433469e-01 -7.54491150e-01 4.55140732e-02
-3.30028802e-01 4.05215137e-02 -5.07775068e-01 8.51987839e-01
7.24648118e-01 -7.76011050e-02 7.00973272e-02 -2.08755806e-01
-1.09724559e-01 3.28840762e-01 -1.04918659e+00 1.67613482e+00
-5.74289083e-01 4.18493778e-01 2.77208090e-01 -1.08816040e+00
6.43263519e-01 1.95907094e-02 3.58587921e-01 -7.94713616e-01
3.69064510e-01 2.76623517e-01 2.06053197e-01 -8.49766359e-02
4.91963178e-01 1.45713806e-01 1.76432222e-01 5.12842238e-01
-1.39376163e-01 1.19985923e-01 3.14934164e-01 5.77979505e-01
1.02876687e+00 4.94165272e-02 1.41518181e-02 -1.02159061e-01
2.56527573e-01 -1.51483059e-01 6.68799400e-01 8.33276391e-01
-1.42648686e-02 2.52780467e-01 5.79083383e-01 -2.71324873e-01
-7.76665986e-01 -8.61181140e-01 -6.04951903e-02 1.07549918e+00
2.72438347e-01 -7.48721480e-01 -9.73530829e-01 -8.91042411e-01
1.65521562e-01 4.59881485e-01 -6.47055864e-01 -1.74868897e-01
-5.54160297e-01 -6.61561310e-01 8.56347799e-01 6.56751812e-01
1.19486761e+00 -6.78107500e-01 -7.34636903e-01 3.70432958e-02
1.20943792e-01 -9.32062984e-01 1.48019437e-02 4.09735054e-01
-1.43698871e+00 -8.01527023e-01 -5.57693303e-01 -7.24323332e-01
8.24385464e-01 3.79122585e-01 5.44944286e-01 4.32836205e-01
-6.25319555e-02 -7.26504326e-02 -3.42513829e-01 -5.19856393e-01
-2.72105515e-01 6.09853089e-01 -1.61271289e-01 -2.42447093e-01
2.37258345e-01 -7.72387624e-01 -6.53969944e-01 1.03415996e-01
-7.64968514e-01 4.32505846e-01 1.02017474e+00 6.62592053e-01
5.61586082e-01 2.28092015e-01 4.34306622e-01 -1.48727298e+00
6.62367463e-01 -3.28764260e-01 -4.77225393e-01 7.20003620e-02
-7.59944618e-01 1.50672674e-01 6.57108009e-01 -5.41741312e-01
-1.34417605e+00 -5.29654771e-02 -1.69995695e-01 -2.99272120e-01
1.37132173e-02 4.40722883e-01 -2.39317402e-01 -1.32281885e-01
2.94667631e-01 -2.01637000e-01 -1.92385361e-01 -7.56332219e-01
3.82111907e-01 5.75099885e-01 3.87472034e-01 -8.52809608e-01
8.71348262e-01 4.95880395e-01 2.85182688e-02 -7.18751609e-01
-9.03821945e-01 -1.32229924e-01 -2.75649548e-01 3.56640853e-02
4.84779269e-01 -8.08604300e-01 -7.61769712e-01 3.80821407e-01
-9.28319097e-01 -2.25955650e-01 -2.87507474e-01 4.58846331e-01
-1.24037161e-01 3.99211496e-01 -7.07671404e-01 -6.05412424e-01
-3.95853370e-01 -1.16476059e+00 5.31035304e-01 5.84448457e-01
-4.87020314e-02 -6.80792689e-01 -2.24919632e-01 4.93193269e-01
3.03202093e-01 -1.03884250e-01 1.13359177e+00 -6.11412406e-01
-7.50801682e-01 1.37466475e-01 -4.45663363e-01 3.61586869e-01
-1.18160181e-01 -9.16160271e-03 -1.14527416e+00 -1.91194862e-01
-3.92546654e-01 -1.97548151e-01 1.49506879e+00 3.24310273e-01
1.54141486e+00 -3.65727007e-01 -5.36492944e-01 8.53158176e-01
1.06870818e+00 2.67189145e-01 3.64719689e-01 6.00684941e-01
8.90780509e-01 1.73034608e-01 5.92759430e-01 3.75518084e-01
1.96540542e-02 2.14688927e-01 5.61835647e-01 -2.60847151e-01
-1.33613914e-01 -6.19321287e-01 1.11373350e-01 8.19080770e-01
-1.29029229e-01 -1.49275184e-01 -5.28254867e-01 4.42889780e-01
-1.91704059e+00 -4.96745020e-01 1.39731765e-01 2.06390953e+00
1.03985894e+00 6.30066276e-01 -4.15809080e-02 3.99619788e-01
6.30597889e-01 7.15973750e-02 -5.07580340e-01 -6.55423820e-01
6.95983469e-02 3.98533523e-01 5.10053992e-01 2.00107574e-01
-1.01174343e+00 1.07605588e+00 5.59662819e+00 1.12766314e+00
-1.01151121e+00 5.19874971e-03 8.18816185e-01 -1.05263963e-01
-3.68146002e-01 2.15522379e-01 -8.73721838e-01 4.71663207e-01
5.46192586e-01 2.57645454e-02 1.47572502e-01 1.08463120e+00
1.45745307e-01 -4.97518510e-01 -8.82671475e-01 7.78014779e-01
-1.61259070e-01 -1.23804188e+00 2.46185571e-01 4.19575311e-02
6.69144928e-01 -4.59024549e-01 1.92998692e-01 3.13291222e-01
1.65680304e-01 -8.68133366e-01 5.99988103e-01 -1.50348872e-01
6.10738575e-01 -9.59155202e-01 6.58127248e-01 1.81453049e-01
-1.13504171e+00 -1.58799902e-01 -3.38477045e-01 -3.11530381e-01
3.44402716e-02 1.08959115e+00 -9.79270160e-01 2.92752892e-01
7.21026361e-01 3.98275346e-01 -6.43833220e-01 1.31816590e+00
-5.28121650e-01 1.07329381e+00 -5.37142158e-01 7.66576035e-03
2.72225559e-01 -2.25018159e-01 3.58452052e-01 1.30309939e+00
2.70423472e-01 7.06623793e-02 8.54208469e-02 5.96979082e-01
-5.00480771e-01 1.15956485e-01 -4.84568715e-01 -1.77666411e-01
7.18106568e-01 1.25294077e+00 -1.12104321e+00 -4.28539813e-01
-4.40943658e-01 7.22768605e-01 1.18036076e-01 2.40084842e-01
-8.43087554e-01 -5.52622020e-01 5.01172543e-01 2.13010028e-01
1.22890405e-01 -9.11816210e-02 -7.65903234e-01 -9.51113880e-01
-9.84086692e-02 -9.37952518e-01 3.86160433e-01 -4.70809817e-01
-6.16439104e-01 3.46780628e-01 1.78342760e-01 -8.48000288e-01
2.12549016e-01 -3.24050546e-01 -7.37668276e-01 4.52218145e-01
-1.29574788e+00 -1.05645204e+00 -3.92179161e-01 4.26024109e-01
5.01208663e-01 -7.83242509e-02 3.17069829e-01 5.01879871e-01
-8.16975713e-01 9.09336686e-01 -1.94352999e-01 7.33817890e-02
5.10497808e-01 -1.10172546e+00 2.44471937e-01 9.78203237e-01
9.13137496e-02 8.43803465e-01 7.11656928e-01 -8.35016012e-01
-1.02351010e+00 -1.00086570e+00 4.93502289e-01 4.17544037e-01
3.84090632e-01 -4.30581331e-01 -9.18116868e-01 6.83303893e-01
1.97890401e-01 -4.38735515e-01 4.87338483e-01 2.31949374e-01
-3.16008747e-01 -3.68473411e-01 -1.11220706e+00 6.23105109e-01
1.07255375e+00 -7.86934495e-02 -3.43436688e-01 -9.79675725e-02
8.66421580e-01 -4.71377879e-01 -4.04190421e-01 5.45888722e-01
7.71129310e-01 -8.22511435e-01 7.54609823e-01 -1.72160909e-01
3.47755522e-01 -1.40060812e-01 2.08044738e-01 -1.25186074e+00
1.02880657e-01 -5.69537163e-01 -2.97688931e-01 1.12395358e+00
4.80433851e-01 -5.16071975e-01 1.30519092e+00 3.53598803e-01
-3.03957701e-01 -1.11145985e+00 -8.24656248e-01 -5.18283248e-01
-1.66758955e-01 -4.04676408e-01 5.43597579e-01 6.93728924e-01
-6.75855204e-03 2.46017188e-01 -2.00076312e-01 -2.22795308e-04
5.34360945e-01 1.54935688e-01 8.10982108e-01 -1.30003834e+00
-4.04124111e-01 -3.07958037e-01 -2.01161370e-01 -1.50457942e+00
-1.12410821e-01 -6.40373647e-01 -1.56045601e-01 -1.32434785e+00
3.73771310e-01 -7.31278241e-01 -7.27024078e-02 7.48484492e-01
-9.80232060e-02 1.99312076e-01 1.93379894e-01 -1.58998966e-02
-4.82580155e-01 6.01729035e-01 1.31200790e+00 -2.57202208e-01
-4.01263535e-01 2.00991854e-01 -7.97076583e-01 9.55389261e-01
9.51849401e-01 -6.17246628e-01 -7.77459323e-01 -4.53617841e-01
5.54215372e-01 -4.50087786e-01 -4.99404930e-02 -1.01659954e+00
2.29895622e-01 -1.74048282e-02 2.91271597e-01 -5.92181921e-01
2.80873567e-01 -6.48121238e-01 -3.50151896e-01 5.72867632e-01
-3.41419995e-01 -1.23465285e-02 3.32393914e-01 5.35576940e-01
-1.96339130e-01 -3.09852511e-01 7.45843709e-01 -1.61927596e-01
-3.99697572e-01 1.21422142e-01 -1.94275558e-01 -1.82042811e-02
9.63536918e-01 -5.89823604e-01 -3.75286400e-01 -1.68885395e-01
-4.09109175e-01 2.69571275e-01 2.44677231e-01 1.75899401e-01
5.55691957e-01 -8.02854002e-01 -4.05698903e-02 5.06818593e-01
-3.37523669e-01 5.34287393e-01 2.56215408e-02 8.76815975e-01
-5.33636093e-01 2.31835589e-01 -2.40484789e-01 -3.74957412e-01
-1.48459387e+00 1.00089006e-01 -6.63276091e-02 -2.04831436e-01
-7.77179599e-01 9.82432067e-01 3.30995888e-01 7.95165170e-03
7.33302891e-01 -8.19163382e-01 -1.05210021e-01 3.09956521e-02
2.47838125e-01 6.64623499e-01 2.26480246e-01 -5.87697029e-02
-1.28999323e-01 7.15063363e-02 -5.46622992e-01 -1.87079832e-01
1.10675395e+00 2.94382647e-02 2.01379098e-02 2.29397044e-01
7.14322567e-01 1.11032322e-01 -1.13173139e+00 1.11139556e-02
-2.46912315e-01 -5.38724899e-01 1.39871284e-01 -6.50244236e-01
-1.35250247e+00 1.15237606e+00 4.27116215e-01 4.01871763e-02
1.16567028e+00 -2.01648071e-01 1.09355390e+00 5.19222498e-01
1.09769635e-01 -1.36934388e+00 8.97001326e-02 3.55450720e-01
1.05787553e-02 -8.49822462e-01 2.03408390e-01 -7.70197749e-01
-2.59937078e-01 9.03826237e-01 9.81507063e-01 1.40356034e-01
5.72858453e-01 4.51718986e-01 -1.33299723e-01 3.12991142e-02
-6.58400536e-01 -6.96341544e-02 -1.57214105e-01 4.12647367e-01
3.98278385e-01 1.71649933e-03 -5.30370891e-01 6.44310892e-01
-3.58939916e-01 -9.03910249e-02 5.78430355e-01 1.12846088e+00
-6.85780108e-01 -1.27011991e+00 -1.69440955e-01 7.19613671e-01
-5.23400426e-01 -1.29467681e-01 -3.70858073e-01 7.66687512e-01
3.09985727e-01 7.01560020e-01 -1.54027328e-01 -5.22747815e-01
1.11176170e-01 -2.22973693e-02 5.11327267e-01 -7.11203575e-01
-6.88221097e-01 1.78001881e-01 6.35684207e-02 -4.61417109e-01
-2.25409567e-01 -2.47721270e-01 -1.51938939e+00 -4.91911888e-01
-5.47680378e-01 2.08901554e-01 7.89199650e-01 9.63829994e-01
3.84629786e-01 6.10173643e-01 1.63118541e-01 -3.52145433e-01
-4.56896424e-01 -8.15242529e-01 -4.50085163e-01 2.27214709e-01
-9.21149850e-02 -8.17971170e-01 -2.77556747e-01 -3.06775510e-01] | [8.779496192932129, 3.261981248855591] |
3fcd96f6-84b6-4d8f-b811-e8f55bf74668 | a-comparative-study-on-recent-neural-spoofing | 2103.11326 | null | https://arxiv.org/abs/2103.11326v2 | https://arxiv.org/pdf/2103.11326v2.pdf | A Comparative Study on Recent Neural Spoofing Countermeasures for Synthetic Speech Detection | A great deal of recent research effort on speech spoofing countermeasures has been invested into back-end neural networks and training criteria. We contribute to this effort with a comparative perspective in this study. Our comparison of countermeasure models on the ASVspoof 2019 logical access task takes into account recently proposed margin-based training criteria, widely used front ends, and common strategies to deal with varied-length input trials. We also measured intra-model differences through multiple training-evaluation rounds with random initialization. Our statistical analysis demonstrates that the performance of the same model may be significantly different when just changing the random initial seed. Thus, we recommend similar analysis or multiple training-evaluation rounds for further research on the database. Despite the intra-model differences, we observed a few promising techniques such as the average pooling to process varied-length inputs and a new hyper-parameter-free loss function. The two techniques led to the best single model in our experiment, which achieved an equal error rate of 1.92% and was significantly different in statistical sense from most of the other experimental models. | ['Junich Yamagishi', 'Xin Wang'] | 2021-03-21 | null | null | null | null | ['synthetic-speech-detection'] | ['audio'] | [ 2.35436246e-01 -2.00450450e-01 -3.63947332e-01 -3.41527969e-01
-9.14003491e-01 -4.71990556e-01 6.93373501e-01 2.71266587e-02
-8.12700033e-01 5.18906116e-01 1.41703948e-01 -9.90835309e-01
6.07780926e-02 -4.08163220e-01 -6.03138864e-01 -6.57046080e-01
-1.90011442e-01 4.36216183e-02 5.17835796e-01 -3.12775284e-01
2.41641670e-01 5.69026470e-01 -1.18790948e+00 4.57843155e-01
4.81421113e-01 9.58228230e-01 2.35447008e-03 8.11747909e-01
1.71042532e-01 7.53712654e-01 -1.16073406e+00 -8.35189521e-01
2.72579163e-01 -1.91828206e-01 -8.70933652e-01 -3.70006382e-01
4.78386074e-01 -2.95142084e-01 -4.71157849e-01 1.13583875e+00
1.00712395e+00 2.15558469e-01 3.55982929e-01 -1.00164759e+00
-4.60873634e-01 1.09411061e+00 -2.57294685e-01 5.81929564e-01
1.70162827e-01 2.86400676e-01 8.58372748e-01 -7.14999318e-01
3.98197353e-01 1.23821163e+00 7.68147707e-01 6.10416889e-01
-1.03604805e+00 -8.88502777e-01 -2.14283958e-01 5.06010830e-01
-1.28364897e+00 -8.77296627e-01 7.66348600e-01 -1.60827667e-01
1.21339226e+00 4.79841232e-01 7.31957033e-02 1.40809321e+00
-6.12341538e-02 5.83391964e-01 1.10292482e+00 -7.65076578e-01
7.99071193e-02 6.00227237e-01 3.07766318e-01 4.62381959e-01
3.06412160e-01 5.85580289e-01 -3.94663006e-01 -3.41079503e-01
1.19405910e-01 -5.98991811e-01 -3.96805257e-01 2.07480006e-02
-1.06929171e+00 9.90109503e-01 2.92841256e-01 5.34606397e-01
-8.19308832e-02 4.24695015e-02 7.22265065e-01 6.28111005e-01
2.46757373e-01 5.42607903e-01 -6.71489358e-01 -2.45268568e-02
-1.25471079e+00 1.56665787e-01 7.88521528e-01 5.54013729e-01
2.38780856e-01 3.34115267e-01 -3.38118941e-01 9.23260748e-01
2.87706882e-01 5.58473647e-01 4.87866700e-01 -6.68970585e-01
1.01969051e+00 -3.96631211e-01 -3.85117158e-02 -9.80990410e-01
-3.52914423e-01 -6.85707510e-01 -6.34234488e-01 4.95563336e-02
4.42030877e-01 -2.62429625e-01 -8.40705037e-01 1.81441498e+00
-1.06556877e-01 8.46200958e-02 8.35117251e-02 5.82282543e-01
5.24531186e-01 5.05803943e-01 -8.09328929e-02 -2.73595721e-01
1.36352956e+00 -1.00509882e+00 -9.42706287e-01 -1.88067243e-01
6.52939558e-01 -1.09828150e+00 1.03327179e+00 4.64176536e-01
-9.38107312e-01 -5.34069479e-01 -1.41496480e+00 3.43028754e-01
-5.42371869e-01 1.23934902e-01 3.27395707e-01 1.53362274e+00
-9.97693717e-01 7.44122803e-01 -5.01833379e-01 -1.84959546e-01
2.69514441e-01 3.01867992e-01 -2.05460623e-01 4.18005496e-01
-1.61981142e+00 9.41663980e-01 5.07857561e-01 -8.46007466e-02
-8.51973355e-01 -3.46405119e-01 -4.68368679e-01 1.02626763e-01
3.25216025e-01 -3.47484291e-01 1.28669882e+00 -4.47689265e-01
-1.81402946e+00 7.89509296e-01 -6.54441044e-02 -1.04943275e+00
5.51420927e-01 -2.36759543e-01 -8.26048970e-01 5.03293499e-02
-4.62747008e-01 2.89022237e-01 1.15979660e+00 -1.03122735e+00
-3.41222763e-01 4.75694127e-02 -2.39114866e-01 -4.02021348e-01
-5.64328253e-01 4.76786941e-01 -2.13130489e-01 -9.83317435e-01
-7.88219944e-02 -8.45969319e-01 -1.01524070e-02 -5.72080374e-01
-5.67744374e-01 1.43057883e-01 5.89128375e-01 -8.12053740e-01
1.75144434e+00 -2.15207028e+00 -2.69835889e-01 2.89104074e-01
-1.54952168e-01 1.08092928e+00 -1.50820494e-01 3.97870839e-01
-3.05813283e-01 5.91412961e-01 -5.93099482e-02 -5.57740390e-01
3.84240523e-02 -3.21915448e-01 -6.85993552e-01 6.12211287e-01
7.80124143e-02 3.95584255e-01 -6.45566225e-01 -3.17413926e-01
1.48712188e-01 4.22244132e-01 -8.00708771e-01 3.13767403e-01
1.49330065e-01 9.17186309e-03 4.57243696e-02 3.29744786e-01
9.12944853e-01 1.39011979e-01 -9.46703330e-02 -1.33996680e-01
1.85748428e-01 8.83160055e-01 -1.03560519e+00 1.14803684e+00
-4.91748035e-01 9.71997023e-01 1.55471921e-01 -8.89282286e-01
8.50278378e-01 6.70725644e-01 -3.63291800e-02 -4.91631836e-01
3.32766175e-01 4.48974103e-01 2.88536310e-01 -4.28258389e-01
5.60190856e-01 5.74894398e-02 3.22119683e-01 3.53018254e-01
9.21450630e-02 -2.02922225e-01 -1.54200733e-01 -7.46311015e-03
9.50739801e-01 -5.08586764e-01 1.39428943e-01 -9.77615267e-02
8.03630650e-01 -6.09085977e-01 2.71979600e-01 1.16111171e+00
-6.54055774e-01 6.31291866e-01 2.46795580e-01 -2.40010440e-01
-9.87777829e-01 -7.61602521e-01 -4.34314221e-01 9.69462156e-01
-8.56530294e-02 -5.47073424e-01 -9.38135207e-01 -7.82814503e-01
-1.83662981e-01 9.03313577e-01 -3.15915614e-01 -1.61851078e-01
-8.04974198e-01 -9.27885532e-01 1.27682209e+00 1.29494801e-01
6.12772107e-01 -1.23022187e+00 -3.36080700e-01 2.23885998e-01
-2.35128999e-01 -1.34708858e+00 -2.90244281e-01 6.09930634e-01
-7.23804414e-01 -5.96334636e-01 -8.23558390e-01 -5.93573391e-01
5.00636734e-02 2.50247777e-01 8.51767123e-01 1.97271064e-01
2.06738889e-01 -3.19927722e-01 -3.43103111e-01 -3.37991983e-01
-9.20483947e-01 4.06164944e-01 2.58192569e-01 -1.20619915e-01
3.99293095e-01 -2.57700175e-01 -3.64009887e-01 4.76843685e-01
-7.42031574e-01 -6.05056643e-01 2.65648544e-01 1.08358455e+00
-3.41561995e-02 -4.05233800e-02 6.52427316e-01 -7.09822178e-01
9.98603821e-01 -4.35180843e-01 -4.55167860e-01 2.23878562e-01
-8.82270098e-01 3.92612398e-01 4.88337845e-01 -5.50656736e-01
-7.56245434e-01 -6.05522871e-01 -7.48842418e-01 -4.75684911e-01
-1.36547163e-01 1.89281225e-01 -1.58619359e-01 -8.27908963e-02
8.71630490e-01 3.79311219e-02 -7.59493858e-02 -5.85579753e-01
2.58953005e-01 1.17400098e+00 2.99086630e-01 -3.61061662e-01
7.32906282e-01 1.58390284e-01 -5.42632997e-01 -1.01288998e+00
-3.21433455e-01 -3.64008456e-01 -2.56700575e-01 3.21800053e-01
5.61416984e-01 -6.72344923e-01 -6.39711022e-01 1.01421702e+00
-1.45376396e+00 -2.26265088e-01 -3.62546630e-02 6.94496512e-01
-2.30765209e-01 7.63830125e-01 -8.06743264e-01 -9.14487362e-01
-4.86414194e-01 -1.43220460e+00 4.51718152e-01 -2.80126840e-01
-2.72229522e-01 -8.74387562e-01 5.37919775e-02 4.83020782e-01
7.76785851e-01 -3.89637679e-01 7.30349481e-01 -1.09016550e+00
-6.03849590e-02 -2.69600779e-01 -7.46679977e-02 8.55808616e-01
1.69068798e-02 8.72187018e-02 -1.51792753e+00 -6.25634491e-01
2.72768199e-01 -1.49790227e-01 1.17870271e+00 3.47450227e-01
1.17306852e+00 -3.40812832e-01 -1.51661143e-01 7.51483321e-01
8.86951566e-01 3.57618719e-01 6.63004220e-01 5.69639444e-01
3.27098280e-01 6.40198112e-01 1.43738925e-01 2.17050001e-01
-2.17631340e-01 1.02087116e+00 2.61417717e-01 1.39895072e-02
-2.07833409e-01 -2.41514921e-01 6.45069897e-01 9.48132038e-01
2.72386193e-01 -7.77913570e-01 -6.63191915e-01 4.31829065e-01
-1.16672611e+00 -1.39228165e+00 1.99540630e-01 2.40848064e+00
6.84373617e-01 6.66157365e-01 2.02896118e-01 6.66942000e-01
9.42645490e-01 6.34567201e-01 -8.81658867e-02 -8.52042675e-01
-3.69674683e-01 1.76370025e-01 7.76730180e-01 8.64251912e-01
-1.33254588e+00 9.78625715e-01 7.43045855e+00 1.22556651e+00
-1.44541621e+00 2.36678362e-01 6.28430128e-01 -8.04684758e-02
-2.07812488e-01 -2.47529224e-01 -9.73948240e-01 5.89056253e-01
1.41785145e+00 2.29707792e-01 5.20441234e-01 5.91234446e-01
5.43985739e-02 4.02009994e-01 -6.00033402e-01 8.93049061e-01
7.79820010e-02 -1.21860588e+00 -1.13361634e-01 1.49644926e-01
5.28780639e-01 3.18074912e-01 3.32359880e-01 2.65651405e-01
1.30383357e-01 -8.18563938e-01 9.64832902e-01 -2.31534570e-01
7.26805210e-01 -7.28254557e-01 7.86855519e-01 2.12161079e-01
-7.85989940e-01 -2.92681247e-01 -2.73123860e-01 1.07190013e-01
1.59972429e-01 4.79781210e-01 -1.13602459e+00 4.02421325e-01
4.91342127e-01 4.18865867e-02 -3.53840262e-01 1.03383458e+00
-3.41165304e-01 1.21497905e+00 -4.21782672e-01 -2.23026767e-01
1.88766062e-01 4.73797530e-01 6.86393797e-01 1.62214220e+00
2.09417835e-01 -5.27911127e-01 -3.07670206e-01 4.65639770e-01
-6.35218024e-02 9.17132273e-02 -4.30203795e-01 4.78466824e-02
7.80117095e-01 6.13448083e-01 -4.01215464e-01 -2.78549641e-01
-2.16290802e-01 8.17339540e-01 2.88561791e-01 3.59103382e-01
-1.03765261e+00 -6.72402859e-01 7.85317779e-01 1.28942594e-01
5.80021143e-01 -1.05775177e-01 -3.19743216e-01 -1.07702422e+00
3.66591878e-04 -1.17670286e+00 2.09536657e-01 -1.03435501e-01
-9.60551679e-01 1.14386296e+00 1.00983880e-01 -9.89307880e-01
-3.59860152e-01 -7.63564169e-01 -5.99976540e-01 9.54142272e-01
-1.51768231e+00 -5.22714615e-01 3.13952833e-01 4.14677531e-01
3.18359643e-01 -5.93802631e-01 8.49554777e-01 6.05129957e-01
-8.29713583e-01 1.48418915e+00 1.70285732e-01 1.61851436e-01
7.50424266e-01 -8.35654914e-01 8.70211840e-01 1.16538918e+00
3.90220791e-01 8.89474332e-01 8.60313237e-01 -2.95688480e-01
-7.78597891e-01 -5.70483685e-01 1.14455724e+00 -2.58475631e-01
5.83759308e-01 -3.34533662e-01 -9.53090847e-01 3.11307520e-01
4.44806665e-01 -2.67772712e-02 4.85763609e-01 5.01455963e-02
-6.82590008e-01 2.14720462e-02 -1.31249559e+00 4.90997821e-01
7.85070479e-01 -8.59678566e-01 -5.36749363e-01 7.65795782e-02
1.03654218e+00 -2.19533533e-01 -3.76016110e-01 4.11321193e-01
4.65167284e-01 -1.12203872e+00 1.22164750e+00 -5.87931633e-01
-5.77372201e-02 8.76682475e-02 -3.82197499e-01 -1.34511042e+00
-2.26694629e-01 -1.16909957e+00 -1.17280424e-01 1.35284626e+00
7.69175053e-01 -8.94836366e-01 6.69990599e-01 4.44804206e-02
-7.09808916e-02 -6.54475152e-01 -1.04064691e+00 -1.06178629e+00
3.82509440e-01 -9.07631636e-01 7.46509731e-01 6.65980935e-01
-9.98410303e-03 5.48582859e-02 -7.16436088e-01 2.92540401e-01
4.50104147e-01 -5.08422375e-01 5.88761687e-01 -5.78626037e-01
-5.53582430e-01 -6.29415214e-01 -2.75219470e-01 -1.13322735e+00
2.10933760e-01 -8.06313753e-01 -2.15069264e-01 -6.03165567e-01
-2.82099128e-01 -4.62926149e-01 -5.81854105e-01 1.47518337e-01
-2.76899785e-01 4.70556766e-02 3.24057966e-01 3.16445902e-02
8.15757513e-02 2.26232022e-01 7.78811455e-01 -3.05250823e-01
-1.95335463e-01 4.10583466e-01 -4.02897447e-01 5.26553988e-01
1.07197607e+00 -7.72147000e-01 -3.08698654e-01 -4.89320040e-01
-5.56493253e-02 -9.49831679e-03 1.94983780e-01 -1.22423756e+00
4.31979336e-02 1.30135447e-01 -1.32204175e-01 -4.78330851e-01
4.33937907e-01 -7.36337304e-01 -1.86037198e-01 8.64548743e-01
-5.18937826e-01 -2.08441131e-02 2.34369248e-01 1.75329819e-01
-2.27866516e-01 -5.39738417e-01 1.04098153e+00 2.36929357e-01
-4.21173304e-01 -1.19759992e-01 -4.50496644e-01 1.32062197e-01
5.56965411e-01 -1.07287824e-01 -5.08624554e-01 -4.33016270e-01
-4.94348794e-01 -4.30881351e-01 7.33774379e-02 4.35946375e-01
4.17277694e-01 -1.03185332e+00 -8.39024484e-01 5.47017097e-01
-7.12169483e-02 -9.72249687e-01 -6.46493360e-02 5.66367626e-01
-5.65987647e-01 7.23053336e-01 5.44202887e-03 -2.48017713e-01
-1.44658566e+00 8.24354708e-01 4.12324339e-01 -5.41008532e-01
-1.31985903e-01 1.19675183e+00 -5.44552445e-01 -5.00024736e-01
7.00458407e-01 -4.14867401e-01 -3.96349318e-02 1.34366989e-01
7.15503216e-01 6.37473166e-01 4.61236060e-01 -7.30298817e-01
-4.99415100e-01 1.55270994e-01 -2.23523751e-01 -3.70324641e-01
9.21269715e-01 -3.16368304e-02 4.63785559e-01 1.08525679e-01
1.56112051e+00 4.77902859e-01 -7.40103126e-01 -3.87257427e-01
4.03094999e-02 -4.94772375e-01 1.42187089e-01 -7.86203563e-01
-1.13649011e+00 1.27874231e+00 8.95840526e-01 5.39861381e-01
8.74060750e-01 -4.20444399e-01 8.15811098e-01 3.81944507e-01
4.09209579e-01 -7.64987051e-01 -4.60707135e-02 6.84091091e-01
7.94272661e-01 -1.26191545e+00 -7.74257481e-02 -3.12730819e-01
-4.58477587e-01 8.81637871e-01 1.69159070e-01 8.11052695e-02
7.44715393e-01 4.83891934e-01 3.55977207e-01 2.17130750e-01
-5.98712385e-01 1.54482752e-01 9.66205746e-02 6.78935409e-01
4.46963161e-01 1.70097113e-01 -3.19031596e-01 3.33922863e-01
-5.09486675e-01 -2.94325769e-01 1.81106880e-01 6.82411492e-01
-3.89953852e-01 -1.30688655e+00 -6.12525225e-01 6.57310933e-02
-1.08606827e+00 -4.36737120e-01 -1.86811760e-01 5.69159031e-01
-3.23757790e-02 1.39756751e+00 -2.59198427e-01 -9.95436490e-01
1.54880390e-01 1.27106234e-01 1.64692268e-01 -2.20374733e-01
-1.13881028e+00 -1.96336001e-01 3.05568069e-01 -3.63696247e-01
-1.92173980e-02 -6.06557548e-01 -6.86948180e-01 -6.41511142e-01
-9.09407198e-01 2.09665447e-01 8.52688730e-01 7.84469426e-01
2.63055384e-01 3.77746522e-01 8.91224027e-01 -7.27665305e-01
-1.26338172e+00 -1.39301300e+00 -2.76375353e-01 4.02534187e-01
7.54772782e-01 -3.87935430e-01 -9.17436719e-01 -2.03009948e-01] | [14.023295402526855, 5.842580795288086] |
184f02c3-1585-4402-84ad-ff61d32ddefd | information-fusion-via-symbolic-regression-a | 2306.00153 | null | https://arxiv.org/abs/2306.00153v1 | https://arxiv.org/pdf/2306.00153v1.pdf | Information Fusion via Symbolic Regression: A Tutorial in the Context of Human Health | This tutorial paper provides a general overview of symbolic regression (SR) with specific focus on standards of interpretability. We posit that interpretable modeling, although its definition is still disputed in the literature, is a practical way to support the evaluation of successful information fusion. In order to convey the benefits of SR as a modeling technique, we demonstrate an application within the field of health and nutrition using publicly available National Health and Nutrition Examination Survey (NHANES) data from the Centers for Disease Control and Prevention (CDC), fusing together anthropometric markers into a simple mathematical expression to estimate body fat percentage. We discuss the advantages and challenges associated with SR modeling and provide qualitative and quantitative analyses of the learned models. | ['Nitesh V. Chawla', 'Jennifer J. Schnur'] | 2023-05-31 | null | null | null | null | ['symbolic-regression'] | ['knowledge-base'] | [ 6.92098737e-01 5.40919244e-01 -6.49669707e-01 -8.57035041e-01
-4.47429478e-01 -2.26952489e-02 9.47116315e-02 8.80783856e-01
1.09780850e-02 5.68398833e-01 3.69779974e-01 -4.09547120e-01
-4.96104330e-01 -6.24723494e-01 -5.92570424e-01 -1.61766559e-01
-4.21824813e-01 2.74156004e-01 -5.63639641e-01 -2.23783836e-01
-2.94424415e-01 3.28013211e-01 -1.30055881e+00 2.18785107e-01
1.18696141e+00 9.75011468e-01 -2.77514547e-01 3.75262618e-01
8.55177864e-02 7.69415081e-01 -5.47993302e-01 -3.86745870e-01
4.33037393e-02 -4.83246177e-01 -5.24383962e-01 -3.08938324e-01
2.24561363e-01 -9.19812396e-02 9.64785144e-02 9.23666298e-01
1.66891932e-01 6.79809451e-02 4.66532916e-01 -1.34151566e+00
-5.51678956e-01 1.15251637e+00 -2.70623505e-01 -2.20079988e-01
6.01878762e-01 -4.61195968e-02 5.91123462e-01 -5.98535180e-01
4.31098819e-01 1.45899105e+00 1.02245581e+00 3.81531298e-01
-1.61206496e+00 -6.37466133e-01 3.84981096e-01 -1.63450420e-01
-1.60773933e+00 -6.45600677e-01 5.46057463e-01 -5.59303045e-01
7.26910710e-01 7.44350612e-01 1.05190992e+00 6.94664836e-01
1.99551493e-01 5.08496761e-01 1.21796966e+00 -6.06063187e-01
1.13894291e-01 9.24197286e-02 6.99706554e-01 8.94451797e-01
8.90760720e-01 3.47103834e-01 -5.95454395e-01 -3.89000326e-01
6.25366628e-01 3.12497150e-02 1.17757209e-01 -4.70994532e-01
-9.96853888e-01 7.90304840e-01 3.16585392e-01 1.47478074e-01
-3.84713173e-01 1.67197108e-01 2.07937241e-01 1.66712299e-01
6.47877634e-01 1.98006511e-01 -3.67277920e-01 2.25924477e-01
-9.72146451e-01 3.55288804e-01 7.61478722e-01 4.97457653e-01
3.75059903e-01 2.73276687e-01 1.60702229e-01 6.20537102e-01
9.10259068e-01 7.82896042e-01 -1.65995389e-01 -1.01017070e+00
7.12449104e-02 9.35926318e-01 1.83787525e-01 -1.47715616e+00
-7.55806148e-01 -2.61827260e-01 -8.99650037e-01 2.72806197e-01
2.72041112e-01 -2.30385303e-01 -5.21233261e-01 1.76560271e+00
3.34533125e-01 -1.28837168e-01 4.13445756e-02 7.71489680e-01
1.04931533e+00 -8.62596780e-02 9.18616474e-01 -3.07944238e-01
1.29224324e+00 -3.75587493e-01 -1.00764740e+00 8.36634170e-03
5.71356475e-01 -1.92931786e-01 4.60071564e-01 4.18066025e-01
-1.17887342e+00 -6.57305002e-01 -1.00344849e+00 -1.65736541e-01
-5.89436710e-01 2.45066628e-01 1.10843945e+00 9.20007288e-01
-4.46154952e-01 6.41902387e-01 -1.38878691e+00 -2.51329958e-01
1.53832719e-01 4.67779189e-01 -2.76484489e-01 2.13248178e-01
-1.30197036e+00 1.30220616e+00 3.98233950e-01 2.50875652e-01
-1.32822901e-01 -1.01033628e+00 -1.40967512e+00 -2.29161680e-02
2.72455961e-01 -9.13260877e-01 8.60127389e-01 -8.89001787e-01
-1.20795298e+00 7.15953350e-01 -2.24519745e-01 -6.32059872e-01
2.00684384e-01 -3.43889922e-01 -8.99259686e-01 4.35578451e-02
-2.78057396e-01 4.89426792e-01 9.59688500e-02 -9.50555921e-01
-3.36032957e-01 -6.44053161e-01 -2.12185904e-01 5.01063056e-02
2.99866945e-01 3.88289839e-01 4.12780285e-01 -7.05948055e-01
2.34426290e-01 -4.64787930e-01 -4.67680812e-01 2.06388935e-01
-2.32368261e-01 -2.15861890e-02 1.52385369e-01 -1.09227896e+00
1.58193922e+00 -2.08174896e+00 1.19574778e-01 5.61184943e-01
3.33241165e-01 8.63118991e-02 4.16462690e-01 2.31831402e-01
-3.41471970e-01 3.24371189e-01 -2.85683781e-01 -1.92700196e-02
4.85144518e-02 1.81682140e-01 1.36802137e-01 5.73702514e-01
3.41880590e-01 9.73880172e-01 -1.03399384e+00 -4.61038083e-01
7.83418655e-01 7.16625094e-01 -1.44438118e-01 -2.48183876e-01
-2.88426839e-02 3.79040837e-01 -4.38264102e-01 6.75577164e-01
3.18604916e-01 -7.95339271e-02 6.66683793e-01 -5.18658817e-01
-1.21215686e-01 3.31896782e-01 -1.33956158e+00 1.51536202e+00
1.97738841e-01 2.40540385e-01 1.44022658e-01 -9.87913132e-01
1.12605882e+00 2.36986905e-01 8.22521269e-01 -4.94480401e-01
8.15992206e-02 -6.82570487e-02 5.74751571e-02 -4.84141648e-01
2.07617253e-01 -3.92605364e-01 -9.60936248e-02 1.88482150e-01
-1.52410299e-01 1.29912898e-01 1.75741777e-01 -1.36711895e-01
6.05723500e-01 5.10087132e-01 1.05445027e+00 -3.68135244e-01
5.74537575e-01 3.08871806e-01 5.88946462e-01 4.99527186e-01
1.04912324e-02 2.15702340e-01 1.95945367e-01 -4.96556550e-01
-6.77780449e-01 -1.02353299e+00 -3.45831037e-01 8.64262521e-01
-2.20399782e-01 -6.22504413e-01 -4.16153222e-01 -4.81203228e-01
3.98217857e-01 7.85954714e-01 -8.01541328e-01 -8.36969241e-02
-5.23111880e-01 -9.50328112e-01 5.81655204e-01 5.66818297e-01
1.62534475e-01 -5.00056922e-01 -8.13185036e-01 1.99609116e-01
-2.02388272e-01 -6.68879390e-01 1.80410832e-01 4.07985002e-02
-1.11542344e+00 -1.50793159e+00 -2.25730762e-01 -2.02691689e-01
6.15191340e-01 3.21204588e-02 1.22152126e+00 3.35081667e-01
-2.13965207e-01 6.16275311e-01 -1.52717054e-01 -7.63406456e-01
-8.18923116e-01 -2.49814823e-01 2.69691013e-02 -5.37268698e-01
4.68057662e-01 -1.19102992e-01 -4.64919448e-01 1.73228949e-01
-7.93031514e-01 4.23043668e-01 2.48249590e-01 4.39059526e-01
4.31786329e-01 -8.11179634e-03 5.43561995e-01 -8.30896080e-01
4.17490304e-01 -5.84739983e-01 -5.20509303e-01 5.97238481e-01
-9.96055484e-01 -2.06190962e-02 -1.16660640e-01 -2.93135732e-01
-8.41874421e-01 2.19889715e-01 2.14182399e-02 2.67534882e-01
-1.63019955e-01 1.02969420e+00 4.66941185e-02 -5.15890159e-02
6.97434008e-01 -2.30241805e-01 4.84712750e-01 -7.46420383e-01
2.73805201e-01 4.38291579e-01 3.82545590e-01 -6.51515245e-01
3.87616336e-01 2.05453336e-01 3.49840850e-01 -5.98806798e-01
-6.38492882e-01 -6.83185384e-02 -6.78427160e-01 -1.96467325e-01
7.25705922e-01 -8.96296918e-01 -9.50897515e-01 -5.68307899e-02
-5.66091537e-01 -1.72502711e-01 -3.23659807e-01 7.37679601e-01
-4.27568406e-01 3.20170403e-01 -2.65278339e-01 -1.02951920e+00
-4.50074196e-01 -8.99926007e-01 7.83946157e-01 1.89902280e-02
-7.50923812e-01 -1.22539008e+00 7.16352165e-02 3.69852513e-01
2.88263410e-01 1.05584311e+00 1.25446022e+00 -6.65920556e-01
5.17272763e-02 -4.16162796e-02 -1.41326800e-01 1.29563421e-01
3.81342143e-01 2.53545344e-01 -4.33894187e-01 1.32245108e-01
-1.73103526e-01 8.72453675e-02 2.95730352e-01 4.91237581e-01
7.14641094e-01 -3.65465850e-01 -4.23002392e-01 1.24016598e-01
1.25307226e+00 7.09437877e-02 2.27519557e-01 -8.39646459e-02
4.22446609e-01 7.30969667e-01 7.12061107e-01 2.29544550e-01
6.48076236e-01 8.90965879e-01 1.95160806e-02 -5.22401929e-01
-1.86941132e-01 -2.48665765e-01 2.25379914e-01 3.43278199e-01
-3.20581883e-01 3.85762751e-01 -8.69275510e-01 -1.37107372e-01
-1.88369584e+00 -7.69495547e-01 -2.80685008e-01 2.26576567e+00
8.88806522e-01 -2.47992024e-01 3.73941898e-01 2.48874873e-01
4.25501227e-01 -1.72811359e-01 -1.89814076e-01 -4.48834121e-01
7.40607604e-02 2.72428811e-01 3.70854497e-01 6.30745173e-01
-8.29689503e-01 3.58932287e-01 8.37608814e+00 -1.94364451e-02
-8.99064362e-01 -2.43683323e-01 6.56155467e-01 2.57647246e-01
-3.34233075e-01 -3.03498775e-01 -4.13791329e-01 2.45964661e-01
1.28985858e+00 -2.41379559e-01 3.52728337e-01 6.94063187e-01
6.60468221e-01 -2.70576268e-01 -1.12688386e+00 3.74103218e-01
-1.33524597e-01 -1.14549708e+00 -1.86984465e-01 -2.28847876e-01
1.53397575e-01 -2.79000908e-01 -2.27254540e-01 2.20919937e-01
4.95619863e-01 -1.31816983e+00 7.79823065e-01 9.69525397e-01
6.96208060e-01 -5.04726887e-01 5.37537754e-01 7.17039686e-03
-1.00375950e+00 1.17680833e-01 -2.48860922e-02 -3.33130538e-01
1.74449176e-01 5.93569279e-01 -7.59333432e-01 9.07819271e-01
4.22083467e-01 7.09445238e-01 -6.19854331e-01 6.67166829e-01
9.25014019e-02 4.93987083e-01 -4.82141644e-01 4.53578413e-01
-4.16559845e-01 -3.05281788e-01 3.02949458e-01 1.34680676e+00
1.71040595e-01 4.16201770e-01 6.34773001e-02 1.00695133e+00
5.81549764e-01 1.97477981e-01 -4.36209887e-01 -1.89851776e-01
1.51977032e-01 8.61236274e-01 -5.56611717e-01 -5.51756978e-01
-4.72640723e-01 -3.93992737e-02 3.08040157e-03 1.23203531e-01
-7.14085221e-01 2.83520579e-01 6.96092188e-01 1.34604007e-01
-2.70425409e-01 -2.78265119e-01 -6.55440450e-01 -1.04666305e+00
-5.34536958e-01 -1.29827559e+00 7.27568388e-01 -6.66507185e-01
-8.65151763e-01 -2.30554551e-01 7.84135520e-01 -6.88328326e-01
-4.19538349e-01 -5.13818741e-01 1.29351124e-01 9.48004484e-01
-1.01292908e+00 -1.26757336e+00 -4.20996249e-01 6.58619776e-02
-2.23045028e-03 2.55668253e-01 1.11166179e+00 2.45069161e-01
-6.32922530e-01 2.99841613e-01 -2.86982417e-01 -2.30630860e-03
1.70011312e-01 -1.16429162e+00 2.46718660e-01 5.14737904e-01
-2.31988177e-01 1.12169456e+00 1.02464068e+00 -1.13435066e+00
-1.41841292e+00 -8.73715997e-01 9.94978130e-01 -2.79212207e-01
3.29077482e-01 -6.23993278e-02 -7.67014563e-01 1.03337765e+00
-2.18856409e-01 -3.57564360e-01 1.27977109e+00 2.64577359e-01
-1.69240683e-01 -1.01483271e-01 -1.65912807e+00 5.76874435e-01
8.55631113e-01 9.25297365e-02 -6.86048269e-01 1.80971026e-02
4.88508523e-01 -6.67225480e-01 -1.62605810e+00 8.06307912e-01
1.00383055e+00 -8.43494833e-01 1.40044308e+00 -7.74342597e-01
-8.90480801e-02 -4.48986411e-01 -2.91352451e-01 -8.04873407e-01
-2.63826698e-01 -2.36198753e-01 -3.81946206e-01 8.71558070e-01
1.02211833e-01 -5.99617124e-01 4.23319638e-01 1.26013339e+00
3.24538872e-02 -5.98079503e-01 -5.21971166e-01 -3.63348365e-01
-2.35002548e-01 -6.04058385e-01 9.74798739e-01 9.33773339e-01
4.40158904e-01 -3.81916016e-02 -3.34827632e-01 1.13978013e-01
7.14407027e-01 -6.28597587e-02 6.17254972e-01 -1.47002089e+00
-1.98936220e-02 -3.13383341e-01 -3.94413859e-01 -2.67969102e-01
-1.39855221e-01 -7.64491498e-01 -5.02337039e-01 -1.56533277e+00
1.21156290e-01 -4.26669985e-01 -5.04212976e-01 6.92386925e-01
-1.51264099e-02 2.06746623e-01 3.51872504e-01 -2.69078702e-01
-2.86530763e-01 -2.45410874e-02 7.94452071e-01 -1.39941067e-01
-2.96219230e-01 5.14519587e-02 -1.11760581e+00 6.59794092e-01
7.01833546e-01 -5.61759651e-01 -3.33952844e-01 -3.49394046e-04
5.42053461e-01 4.24164683e-01 5.08560717e-01 -7.23558247e-01
-3.43540609e-01 -6.49147749e-01 8.53938043e-01 -6.36043012e-01
3.25471640e-01 -1.25425744e+00 1.14945722e+00 1.01018596e+00
-7.21629739e-01 -2.19559725e-02 2.87202388e-01 2.10109279e-01
1.43262967e-01 -1.65394768e-01 5.30207217e-01 -1.74282969e-03
-1.32670432e-01 -3.47938001e-01 -3.04531157e-01 -4.79520112e-01
7.63810873e-01 -1.35903224e-01 -2.51274884e-01 -2.09160894e-01
-1.19192672e+00 2.88756847e-01 3.30756724e-01 4.36607689e-01
4.60017771e-01 -1.18215120e+00 -6.22443855e-01 3.59590381e-01
3.62445600e-03 -3.37363809e-01 -1.03978459e-02 1.05206382e+00
-3.21870238e-01 5.67269325e-01 -2.43816853e-01 -4.00022745e-01
-1.48269284e+00 5.82111299e-01 4.06296253e-01 -2.37795547e-01
-5.95532656e-01 -1.96731493e-01 -3.92522961e-01 -4.58281815e-01
2.67714232e-01 -8.50126266e-01 -3.86038721e-01 -2.08515555e-01
6.30305409e-01 9.37616289e-01 -1.39732972e-01 -5.32280922e-01
-4.25329536e-01 1.65454686e-01 5.72442174e-01 1.81949764e-01
1.29812610e+00 -4.09499377e-01 -1.51942089e-01 6.71323299e-01
5.61040282e-01 -1.18683718e-01 -5.79037309e-01 -4.84205857e-02
2.30545148e-01 -1.05934381e-01 -2.01304868e-01 -1.25577223e+00
-7.48897970e-01 3.68939608e-01 7.69334078e-01 2.28617713e-01
9.48131919e-01 -2.66714692e-01 1.55362278e-01 2.19131336e-01
1.72270432e-01 -6.43964648e-01 -8.23295593e-01 -1.45994857e-01
8.39565694e-01 -1.07269669e+00 7.36005843e-01 -7.71538079e-01
-4.82091457e-01 1.23376977e+00 8.09219405e-02 1.57783240e-01
7.02340782e-01 2.41899416e-01 7.49931037e-02 -2.49536440e-01
-3.53651553e-01 2.73016896e-02 6.61165297e-01 9.59157705e-01
8.14162552e-01 4.59736526e-01 -7.30543315e-01 7.96625376e-01
-1.67559415e-01 5.69631994e-01 -6.92155063e-02 8.42244923e-01
-3.80673319e-01 -1.17165577e+00 -5.60745776e-01 6.70016229e-01
-6.04517221e-01 1.49788752e-01 -3.49438518e-01 1.14474285e+00
1.84215784e-01 9.47225630e-01 -1.28828689e-01 -4.08458970e-02
2.46331096e-01 4.10524577e-01 6.51151299e-01 -5.96594334e-01
-8.94704342e-01 2.75985748e-02 5.96923411e-01 -4.91270572e-01
-8.65991235e-01 -8.71577621e-01 -1.52583635e+00 -3.34314466e-01
1.08299844e-01 -1.48445517e-02 8.78861666e-01 1.06169069e+00
2.54111975e-01 5.69730639e-01 9.07210931e-02 -3.30298156e-01
-3.91951650e-01 -9.02873814e-01 -3.64110947e-01 3.67980152e-01
2.97763318e-01 -8.66862357e-01 2.80512590e-02 1.61317587e-01] | [8.116683006286621, 5.581523895263672] |
7c40ccad-0582-4e80-b0ca-ead0740b2c27 | toan-target-oriented-alignment-network-for | 2005.13820 | null | https://arxiv.org/abs/2005.13820v2 | https://arxiv.org/pdf/2005.13820v2.pdf | TOAN: Target-Oriented Alignment Network for Fine-Grained Image Categorization with Few Labeled Samples | The challenges of high intra-class variance yet low inter-class fluctuations in fine-grained visual categorization are more severe with few labeled samples, \textit{i.e.,} Fine-Grained categorization problems under the Few-Shot setting (FGFS). High-order features are usually developed to uncover subtle differences between sub-categories in FGFS, but they are less effective in handling the high intra-class variance. In this paper, we propose a Target-Oriented Alignment Network (TOAN) to investigate the fine-grained relation between the target query image and support classes. The feature of each support image is transformed to match the query ones in the embedding feature space, which reduces the disparity explicitly within each category. Moreover, different from existing FGFS approaches devise the high-order features over the global image with less explicit consideration of discriminative parts, we generate discriminative fine-grained features by integrating compositional concept representations to global second-order pooling. Extensive experiments are conducted on four fine-grained benchmarks to demonstrate the effectiveness of TOAN compared with the state-of-the-art models. | ['Jun-Jie Zhang', 'Chang Xu', 'Huaxi Huang', 'Qiang Wu', 'Jian Zhang'] | 2020-05-28 | null | null | null | null | ['image-categorization', 'fine-grained-visual-categorization'] | ['computer-vision', 'computer-vision'] | [ 1.16249040e-01 -4.60227787e-01 -2.89236009e-01 -6.64715469e-01
-6.29901886e-01 -3.56882691e-01 7.29886532e-01 7.47848153e-02
-4.75589991e-01 4.83076751e-01 3.68132353e-01 4.49201763e-01
-4.55462992e-01 -6.87921643e-01 -3.84879440e-01 -7.52074778e-01
2.59691298e-01 -1.69074144e-02 4.86924946e-01 -1.41193988e-02
5.09324133e-01 2.33210906e-01 -1.88398445e+00 6.52734101e-01
8.19430649e-01 1.59753549e+00 2.13749215e-01 1.51662007e-01
-4.32514846e-01 4.81183559e-01 -6.08589411e-01 -2.29145035e-01
2.89849252e-01 -3.02246720e-01 -7.00927317e-01 3.54017019e-01
8.92339826e-01 -1.96979299e-01 -2.15027481e-01 1.45529759e+00
3.38027835e-01 2.17699543e-01 1.00198257e+00 -1.45622265e+00
-1.07350910e+00 3.63005847e-01 -8.66735339e-01 5.48103154e-01
-1.84417725e-01 1.03665911e-01 1.21969295e+00 -1.33425069e+00
4.43698585e-01 1.63580513e+00 4.16178077e-01 3.81478786e-01
-1.37744141e+00 -7.47974455e-01 7.22100258e-01 5.10997295e-01
-1.78599823e+00 -4.03759688e-01 6.16478086e-01 -7.49826133e-01
6.76839113e-01 1.80869833e-01 3.09369296e-01 9.82711613e-01
2.25664690e-01 4.65409219e-01 1.32045579e+00 -3.52501243e-01
2.88173139e-01 5.45541272e-02 7.92376816e-01 7.11805224e-01
2.43673638e-01 7.12481439e-02 -6.31513834e-01 -3.95423733e-02
5.58872819e-01 4.46095258e-01 -5.46005405e-02 -5.85912466e-01
-1.27079153e+00 1.02014959e+00 7.55042374e-01 4.08528477e-01
-1.75040871e-01 9.42902814e-04 5.15316010e-01 3.69694978e-01
5.85381389e-01 1.81402683e-01 -2.61699617e-01 2.26226941e-01
-8.94706845e-01 1.02330901e-01 2.96812832e-01 9.77536261e-01
1.10835469e+00 -1.22283272e-01 -9.51820016e-01 1.08213508e+00
2.31853634e-01 -5.32474034e-02 8.61195564e-01 -6.86236143e-01
4.07452703e-01 8.13685596e-01 -1.48774996e-01 -1.12991321e+00
-6.57067522e-02 -6.80498958e-01 -1.10770214e+00 1.09445147e-01
3.34043413e-01 4.83676106e-01 -1.07371199e+00 1.72308612e+00
3.03335130e-01 1.32446021e-01 -3.37206900e-01 9.91410613e-01
9.02209163e-01 3.97629261e-01 2.66995907e-01 -1.59452677e-01
1.58188140e+00 -1.38115978e+00 -4.41249371e-01 -4.12901312e-01
3.21928203e-01 -5.86043596e-01 1.45364249e+00 -1.21176474e-01
-4.80493546e-01 -9.66082752e-01 -1.08659685e+00 -1.02554113e-01
-7.62378097e-01 9.79692712e-02 3.76047045e-01 4.11633015e-01
-7.09208548e-01 5.81165016e-01 -2.89740294e-01 -1.79043338e-01
8.36296916e-01 2.33925767e-02 -4.03182417e-01 -4.14625406e-01
-1.14670157e+00 5.49718618e-01 3.99761856e-01 -5.09158382e-03
-9.08803880e-01 -8.61858666e-01 -7.92492509e-01 2.81639189e-01
4.14803118e-01 -5.80136478e-01 8.52799058e-01 -6.96429789e-01
-9.73644853e-01 9.89220023e-01 -2.78321892e-01 -6.27731979e-02
3.37356508e-01 4.44967113e-02 -1.71748683e-01 -2.98002660e-02
6.94426179e-01 7.17845261e-01 1.20300746e+00 -9.80077803e-01
-8.27563107e-01 -6.47103369e-01 1.55307427e-01 3.59933227e-02
-5.06205678e-01 -5.56752831e-02 -3.62977147e-01 -9.70942974e-01
1.63930535e-01 -5.42480648e-01 -1.35046899e-01 3.06975543e-01
-4.53403711e-01 -7.01423466e-01 7.43237913e-01 2.84739360e-02
1.25109410e+00 -2.35498929e+00 1.00862652e-01 -1.33292660e-01
3.65462780e-01 4.29182462e-02 -3.64281207e-01 1.76372379e-01
1.54182330e-01 1.36237279e-01 1.31886438e-01 -2.07590431e-01
3.61580968e-01 3.61999236e-02 -4.71823335e-01 3.42780828e-01
4.83498544e-01 1.01385379e+00 -7.43524790e-01 -6.66951835e-01
1.00502580e-01 8.23297948e-02 -3.23087335e-01 2.07812592e-01
1.62445381e-01 1.07894763e-01 -8.32977235e-01 8.47990453e-01
6.21121585e-01 -5.00487804e-01 -3.84623855e-01 -5.24893999e-01
-3.40029821e-02 -2.18454033e-01 -1.13791788e+00 1.66538513e+00
-2.37257913e-01 1.56969726e-01 -2.52874553e-01 -1.04220784e+00
9.62453961e-01 -4.24924612e-01 -1.07389234e-01 -9.32396173e-01
6.90391585e-02 1.05942138e-01 3.89422514e-02 -2.11772665e-01
3.81081790e-01 -2.74417132e-01 -4.97316718e-01 2.88830996e-01
3.56118590e-01 1.64387688e-01 2.39286035e-01 1.56986624e-01
6.47399366e-01 -1.77692324e-01 6.37351751e-01 -5.85563004e-01
4.25008625e-01 -2.41359845e-01 6.23999536e-01 1.13720572e+00
-6.80307686e-01 6.03527844e-01 2.63315320e-01 -4.16685283e-01
-7.86593497e-01 -9.97631907e-01 -3.57299328e-01 1.58889782e+00
6.44904494e-01 -4.65634555e-01 -6.78634584e-01 -8.53222728e-01
2.56038368e-01 2.27549225e-01 -1.12418652e+00 -4.45390701e-01
-1.53897151e-01 -7.45481789e-01 3.06544155e-01 6.18754148e-01
6.96717262e-01 -9.22155797e-01 -4.81616020e-01 1.57064244e-01
-2.83560567e-02 -9.78192329e-01 -8.89089763e-01 2.98705041e-01
-4.46955919e-01 -9.35386956e-01 -8.17038596e-01 -8.34402442e-01
4.32758987e-01 7.28598654e-01 9.85721409e-01 -1.80595562e-01
-6.99380934e-01 1.09183632e-01 -4.54808712e-01 -2.22907126e-01
3.47831458e-01 -5.88881224e-02 5.82859404e-02 3.63023639e-01
7.80619681e-01 -2.41689458e-01 -8.39821339e-01 6.71096623e-01
-6.75771117e-01 -2.15128675e-01 6.06350720e-01 1.44911432e+00
9.71354544e-01 1.88684054e-02 7.52110481e-01 -6.63374245e-01
5.31697392e-01 -3.83556455e-01 -2.30646402e-01 3.93540829e-01
-5.33675969e-01 1.82890221e-02 6.79875255e-01 -6.06713355e-01
-8.00393403e-01 -4.56400663e-01 4.50020850e-01 -7.16999650e-01
-3.12834024e-01 1.96971327e-01 -1.68305770e-01 -1.29625291e-01
5.13881087e-01 2.94364423e-01 -3.82953018e-01 -6.22445345e-01
3.59018147e-01 7.70011067e-01 2.55197704e-01 -5.73785841e-01
7.27158725e-01 4.82558370e-01 -2.21998289e-01 -6.35442853e-01
-1.41605103e+00 -6.83098197e-01 -9.00968969e-01 1.57272339e-01
8.47142041e-01 -9.24762785e-01 -2.50714332e-01 5.12366354e-01
-8.27986360e-01 4.24812697e-02 -5.52657127e-01 2.00778082e-01
-5.13891339e-01 2.40034074e-01 -3.56638402e-01 -4.54051524e-01
-2.19577461e-01 -1.20543897e+00 1.47825015e+00 4.16060865e-01
3.50247929e-03 -6.14078283e-01 -2.10543081e-01 2.47454911e-01
3.22526336e-01 -1.28360733e-01 1.05305684e+00 -6.16292536e-01
-4.62146938e-01 -4.49470617e-02 -8.69348526e-01 2.95945674e-01
3.12970251e-01 -2.95493126e-01 -1.14694881e+00 -4.82803345e-01
-8.88260752e-02 -6.84169531e-01 1.23249066e+00 3.40896547e-01
1.55403757e+00 -2.43588179e-01 -4.85641241e-01 5.97570121e-01
1.31611359e+00 -1.82196140e-01 2.18805492e-01 4.25419301e-01
8.11900556e-01 7.64592648e-01 1.19628799e+00 4.51840997e-01
2.53751963e-01 9.14713502e-01 1.84656397e-01 2.54023403e-01
-4.35712010e-01 -2.61333644e-01 -3.06312647e-02 6.21202409e-01
2.71725714e-01 3.14376950e-02 -4.88765061e-01 5.49827099e-01
-1.86102068e+00 -9.03108597e-01 2.31875584e-01 1.73830295e+00
6.64486706e-01 6.01172037e-02 9.68146045e-03 2.15117633e-02
1.03286040e+00 5.01337230e-01 -7.82866895e-01 -1.44679770e-01
-2.01080829e-01 -6.17349856e-02 9.38927904e-02 7.52993748e-02
-1.39682603e+00 1.01913238e+00 5.26730299e+00 1.64340973e+00
-1.03463709e+00 2.14657113e-01 8.45169544e-01 -2.80792452e-02
-2.06112023e-02 -2.71540523e-01 -1.25212455e+00 6.02914333e-01
4.57881808e-01 -2.19122574e-01 2.24771351e-01 9.81612444e-01
-1.66830212e-01 1.06904611e-01 -1.13339961e+00 1.12049031e+00
1.99522793e-01 -1.31286192e+00 3.13905627e-01 -7.87946731e-02
7.41943777e-01 -1.81869254e-01 1.00245796e-01 5.49026012e-01
5.96035868e-02 -1.09928215e+00 9.99615610e-01 5.03605783e-01
1.15321112e+00 -5.56655586e-01 5.69473624e-01 3.08896542e-01
-1.56613886e+00 -4.03948545e-01 -9.58297551e-01 5.07559367e-02
-1.85399801e-01 4.93702263e-01 -2.63773113e-01 4.72741663e-01
1.13071895e+00 8.10382843e-01 -9.50457931e-01 9.30269182e-01
2.63967037e-01 1.18778877e-01 1.39155328e-01 -1.34014174e-01
5.89188635e-01 -5.99022210e-03 2.54606485e-01 1.18980527e+00
1.12026758e-01 -8.64150934e-03 5.26082158e-01 8.60581815e-01
3.54223587e-02 3.09551787e-02 -2.94122279e-01 -2.80009606e-03
4.24062550e-01 1.35517681e+00 -7.29914248e-01 -4.71130908e-01
-3.64194602e-01 1.03875875e+00 6.64497733e-01 2.89505124e-01
-5.29581904e-01 -5.97031176e-01 8.52997601e-01 -7.52836391e-02
6.21599913e-01 1.58968449e-01 -1.82832479e-01 -1.39723158e+00
5.07108532e-02 -7.94522107e-01 7.01922834e-01 -4.68385190e-01
-2.15736699e+00 7.34263480e-01 -7.24358112e-03 -1.40035307e+00
-1.48510579e-02 -6.30981207e-01 -5.18944204e-01 1.01179326e+00
-1.72100949e+00 -1.45879006e+00 -6.00688040e-01 8.55602443e-01
9.95155454e-01 -1.73291117e-01 9.04416203e-01 2.21468166e-01
-5.34435809e-01 9.56908286e-01 8.40179250e-02 -4.85313684e-02
8.65669191e-01 -1.06033361e+00 2.99907494e-02 6.50335610e-01
7.96597004e-02 6.53832197e-01 3.69221628e-01 -3.97076011e-01
-9.34554577e-01 -1.53466082e+00 5.45904636e-01 -2.44838625e-01
8.90991867e-01 -6.00327909e-01 -1.05895376e+00 2.86258668e-01
-1.40688196e-01 7.76121736e-01 6.75396502e-01 3.03705633e-02
-1.02398896e+00 -3.38890105e-01 -1.03927076e+00 4.09398437e-01
1.38551271e+00 -8.44209611e-01 -7.94174135e-01 1.04211353e-01
8.31376135e-01 2.01576784e-01 -7.58952260e-01 3.60033691e-01
6.08111918e-01 -9.59457874e-01 1.07548964e+00 -8.79305840e-01
3.75190198e-01 -3.98999482e-01 -6.81580782e-01 -1.38031340e+00
-9.52318013e-01 2.42301039e-02 -7.61418790e-02 1.25910485e+00
-1.99158847e-01 -5.37919283e-01 6.01472437e-01 6.14399984e-02
-1.69873267e-01 -8.34115386e-01 -1.12424076e+00 -9.30164993e-01
1.87696651e-01 -1.74719584e-03 5.32753766e-01 8.47153485e-01
-2.54745781e-01 5.32819867e-01 -3.00684899e-01 1.77228954e-02
8.85739923e-01 8.34592462e-01 4.63784575e-01 -1.35749066e+00
-2.71311283e-01 -5.66180885e-01 -8.50371242e-01 -9.32676911e-01
2.42244333e-01 -7.89402068e-01 1.53789118e-01 -1.19719017e+00
7.50191629e-01 -4.29593682e-01 -8.13589990e-01 3.99464399e-01
-4.08790946e-01 8.35223496e-01 2.90563524e-01 4.54056859e-01
-8.79650533e-01 7.83439398e-01 1.40461445e+00 -5.89096844e-01
2.99237311e-01 -4.13809150e-01 -9.47757125e-01 6.56873763e-01
4.03202921e-01 -4.42975581e-01 -3.56923193e-01 -1.87779263e-01
-4.98949617e-01 -5.21899760e-01 6.07333779e-01 -8.23414922e-01
2.78108180e-01 -4.06018198e-01 6.61076427e-01 -5.93599796e-01
2.41638944e-01 -4.93548185e-01 -2.52432287e-01 3.25170040e-01
-4.18556839e-01 -3.37033331e-01 -1.77915655e-02 9.11293745e-01
-6.00497067e-01 -2.34739318e-01 1.12018466e+00 -1.43966973e-01
-1.24603522e+00 7.44249761e-01 1.72925085e-01 2.85688370e-01
1.17888272e+00 -4.45894599e-01 -7.19095409e-01 1.92678779e-01
-6.05339348e-01 1.19173877e-01 3.18369836e-01 7.14969397e-01
6.15582883e-01 -1.69347930e+00 -5.19427299e-01 2.43903697e-01
8.62374306e-01 -2.09244221e-01 8.58805180e-01 6.24877512e-01
3.51839036e-01 5.64523757e-01 -5.52012086e-01 -8.11479807e-01
-1.14709342e+00 8.49769175e-01 2.95908719e-01 -2.65367597e-01
-4.59365785e-01 1.15470493e+00 1.20753443e+00 -2.23512411e-01
9.03001204e-02 4.61368775e-03 -4.29624140e-01 5.19959152e-01
9.09548044e-01 1.41057342e-01 1.28652854e-03 -7.82503486e-01
-5.63649058e-01 1.02587807e+00 -5.31866729e-01 4.84370410e-01
1.16122985e+00 -4.91404116e-01 -3.00894734e-02 7.26496220e-01
1.41696286e+00 -3.23423684e-01 -1.41054249e+00 -6.05476737e-01
-3.54602039e-02 -9.42443073e-01 9.54438522e-02 -6.27537251e-01
-8.24612796e-01 1.16488397e+00 8.71572554e-01 2.85367697e-01
9.61722314e-01 2.96312958e-01 2.82551169e-01 2.04705358e-01
5.22621632e-01 -1.05012119e+00 4.35245067e-01 5.08098781e-01
9.78282154e-01 -1.40804005e+00 -1.41396761e-01 -5.08244097e-01
-5.25232673e-01 9.72910106e-01 8.44870746e-01 -3.24498266e-01
7.37357736e-01 -2.25940973e-01 -1.48202449e-01 -2.31483012e-01
-7.13729739e-01 -2.69567817e-01 7.17197895e-01 6.10945523e-01
8.26355070e-02 8.24886337e-02 -9.53754187e-02 9.79417801e-01
1.47703514e-01 -9.61524248e-02 8.20776746e-02 6.39932334e-01
-6.49684191e-01 -6.59674287e-01 -9.97011960e-02 7.87580311e-01
-2.74903417e-01 -2.48836771e-01 -3.14409345e-01 5.73409617e-01
3.24367166e-01 8.49583507e-01 2.94258922e-01 -4.38062042e-01
3.81879002e-01 -9.54257324e-02 3.49904418e-01 -9.15155292e-01
-3.41610849e-01 2.84389034e-02 -1.59300998e-01 -8.06992173e-01
-2.31880561e-01 -4.01196599e-01 -6.84547544e-01 1.89998839e-02
-2.59591311e-01 -9.51864943e-03 1.04169153e-01 8.91242385e-01
5.65467000e-01 6.19069517e-01 8.38074267e-01 -9.42923009e-01
-1.02735388e+00 -9.99851286e-01 -9.77974296e-01 5.49858809e-01
3.71069700e-01 -1.19457364e+00 -5.86341679e-01 -3.68244201e-01] | [9.690065383911133, 2.0700182914733887] |
1a668e57-8bb6-4b0b-b8a4-3233e5b4f19f | a-clustering-based-method-for-automatic | 2010.04676 | null | https://arxiv.org/abs/2010.04676v1 | https://arxiv.org/pdf/2010.04676v1.pdf | A Clustering-Based Method for Automatic Educational Video Recommendation Using Deep Face-Features of Lecturers | Discovering and accessing specific content within educational video bases is a challenging task, mainly because of the abundance of video content and its diversity. Recommender systems are often used to enhance the ability to find and select content. But, recommendation mechanisms, especially those based on textual information, exhibit some limitations, such as being error-prone to manually created keywords or due to imprecise speech recognition. This paper presents a method for generating educational video recommendation using deep face-features of lecturers without identifying them. More precisely, we use an unsupervised face clustering mechanism to create relations among the videos based on the lecturer's presence. Then, for a selected educational video taken as a reference, we recommend the ones where the presence of the same lecturers is detected. Moreover, we rank these recommended videos based on the amount of time the referenced lecturers were present. For this task, we achieved a mAP value of 99.165%. | ['Sérgio Colcher', 'Antonio J. G. Busson', 'Álan L. V. Guedes', 'Eduardo S. Vieira', 'Paulo R. C. Mendes'] | 2020-10-09 | null | null | null | null | ['face-clustering'] | ['computer-vision'] | [-9.50413570e-02 -3.23035747e-01 -1.16223589e-01 -2.38766372e-01
-6.02668643e-01 -6.43662333e-01 4.47934508e-01 2.29802370e-01
-7.43095353e-02 3.75181943e-01 5.33536553e-01 1.11656025e-01
-7.39833891e-01 -7.49400973e-01 -4.76026684e-01 -8.91451836e-01
1.14487998e-01 -1.56192444e-02 1.91406101e-01 -8.62740949e-02
6.58897877e-01 4.83034760e-01 -2.35622835e+00 5.53113639e-01
8.03773105e-01 9.25962865e-01 3.59755248e-01 3.15866172e-01
-3.30476433e-01 9.57713783e-01 -8.05274785e-01 -4.36584353e-01
-2.03921199e-01 -5.62620163e-01 -4.77681309e-01 2.46920779e-01
4.31751370e-01 -2.90433139e-01 -6.77169859e-01 9.32983875e-01
3.33835036e-01 6.95537746e-01 6.52721941e-01 -1.07051969e+00
-4.25892740e-01 8.84535074e-01 -2.13249415e-01 3.36712867e-01
8.58966112e-01 -6.84076786e-01 8.11890781e-01 -7.98039317e-01
6.54570520e-01 9.06363189e-01 1.48029670e-01 5.17258108e-01
-4.93995756e-01 -6.06759489e-01 3.39577228e-01 6.47269845e-01
-1.57646000e+00 -5.00034273e-01 8.36619496e-01 -6.51896656e-01
2.11458638e-01 5.19563675e-01 5.77002347e-01 1.06040478e+00
-2.26209387e-01 4.48921949e-01 5.76453686e-01 -3.42460215e-01
1.78403750e-01 2.54714757e-01 7.03598410e-02 6.45676196e-01
-2.23486070e-02 -6.28810346e-01 -8.13458443e-01 -6.65336102e-02
3.74441981e-01 4.06274974e-01 -4.56228673e-01 -6.70040101e-02
-8.77953410e-01 7.03152835e-01 -3.25204618e-02 6.31875753e-01
-3.70239019e-01 -5.10059774e-01 1.74688280e-01 9.95931700e-02
2.74521500e-01 2.99074024e-01 -9.09563527e-03 -1.44428477e-01
-1.13591778e+00 -3.14475149e-02 7.57572293e-01 8.12003851e-01
6.52141452e-01 -2.57158577e-01 -1.88013420e-01 7.31753051e-01
4.49559450e-01 2.93032434e-02 7.67457724e-01 -8.06660235e-01
2.19676793e-01 7.35594332e-01 -1.04120769e-01 -1.42010844e+00
-6.35385066e-02 -3.26597214e-01 -5.51678538e-01 -3.99644911e-01
2.08049759e-01 -1.42142162e-01 -3.88592303e-01 1.11379278e+00
2.68303663e-01 5.05937338e-01 3.32335494e-02 9.81246889e-01
1.55754364e+00 6.38582766e-01 -1.41774222e-01 -6.27098143e-01
1.41076660e+00 -6.15440011e-01 -6.26598537e-01 5.47937155e-01
3.90484065e-01 -9.62746143e-01 6.49496198e-01 6.70884609e-01
-6.31603956e-01 -6.12031937e-01 -5.99067032e-01 3.65004927e-01
-2.02917859e-01 4.31320012e-01 2.63671964e-01 7.53559291e-01
-1.09183466e+00 4.72738355e-01 -1.72317252e-01 -3.19613516e-01
4.47663404e-02 2.66857058e-01 -3.58566195e-01 -1.22781344e-01
-8.73878896e-01 1.82034612e-01 9.11718160e-02 -1.42808706e-01
-8.39028776e-01 -5.37132204e-01 -4.72068459e-01 3.97518128e-01
5.29203415e-01 -2.54804403e-01 7.31879771e-01 -1.16263175e+00
-1.61088419e+00 5.26977956e-01 -1.28217578e-01 -1.55319169e-01
3.15863788e-01 -1.60574138e-01 -6.91336751e-01 4.59826887e-01
-1.22729056e-01 1.22391712e-02 8.51032197e-01 -1.00716853e+00
-9.24823880e-01 -2.79576540e-01 2.47385576e-01 3.57581466e-01
-9.17855442e-01 1.11798376e-01 -8.16805780e-01 -6.16373837e-01
2.40607604e-01 -6.94870174e-01 6.37453422e-02 -5.09095967e-01
-7.59107471e-02 -5.25393188e-01 7.94518590e-01 -7.16537297e-01
1.51928103e+00 -2.20449305e+00 6.22848682e-02 6.69663668e-01
2.56275862e-01 2.18257651e-01 1.34195358e-01 4.40075934e-01
2.58924365e-01 3.16370279e-02 5.34233809e-01 1.46224260e-01
-2.64793247e-01 -2.30405912e-01 -1.77175671e-01 2.03267455e-01
-4.46275890e-01 -1.40342593e-01 -8.31127763e-01 -6.47879303e-01
2.38663912e-01 6.83357179e-01 -5.86591601e-01 4.42613661e-01
-3.67265865e-02 4.60822225e-01 -6.46270037e-01 3.35242510e-01
2.28981853e-01 2.98459884e-02 3.92809987e-01 -3.14270467e-01
-2.95269042e-01 3.39011788e-01 -1.46622300e+00 1.42136574e+00
-3.23772520e-01 5.43849468e-01 -8.50381777e-02 -9.59319592e-01
1.20316935e+00 5.72967172e-01 9.30449307e-01 -3.43374640e-01
1.25478581e-01 1.91609152e-02 -1.68120280e-01 -9.52314913e-01
5.68417609e-01 5.18677175e-01 2.96542764e-01 4.23350424e-01
1.01281598e-01 5.54643989e-01 5.04303992e-01 3.59576315e-01
9.90494370e-01 1.72641203e-01 -1.60102099e-02 -1.98693331e-02
1.02693617e+00 -4.77190942e-01 3.59324992e-01 4.91333574e-01
1.22270726e-01 5.32587469e-01 3.03665400e-01 -3.45451474e-01
-3.96823198e-01 -4.98509556e-01 5.78599013e-02 1.31178343e+00
1.12929665e-01 -8.63019526e-01 -7.30166554e-01 -7.31086671e-01
-3.87042582e-01 3.67431879e-01 -3.10330540e-01 -2.01366216e-01
-2.37732708e-01 -2.96622574e-01 1.78542063e-01 -3.30439582e-02
2.17693835e-01 -9.03868079e-01 -3.57129574e-01 1.52969390e-01
-5.98601162e-01 -9.05406833e-01 -5.07047176e-01 -4.08222228e-01
-4.40292358e-01 -1.28152263e+00 -7.65633047e-01 -9.31340635e-01
8.92735124e-01 6.86947942e-01 8.81271839e-01 4.96664315e-01
3.04170623e-02 7.31499851e-01 -8.70397985e-01 1.57566935e-01
-2.60654658e-01 -1.66851636e-02 2.14866862e-01 4.51276392e-01
4.27389860e-01 -2.98072398e-01 -6.78190410e-01 5.67543328e-01
-7.04587877e-01 -3.72734398e-01 1.96550786e-01 3.15179318e-01
4.71222013e-01 4.98824239e-01 4.89546895e-01 -7.47019947e-01
5.65546513e-01 -6.62011504e-01 -3.49425942e-01 3.67811799e-01
-2.70898193e-01 -3.56399626e-01 5.86460173e-01 -4.79656905e-01
-1.18338931e+00 2.17904598e-01 -2.46896565e-01 -5.73043227e-01
-4.91474688e-01 5.30166864e-01 -3.03128541e-01 -1.62801996e-01
4.77316946e-01 2.62054145e-01 -2.27631196e-01 -5.07475734e-01
-3.60312723e-02 1.04010093e+00 1.64731532e-01 -5.27479887e-01
3.37907940e-01 4.43673581e-02 -1.78783536e-01 -1.10307741e+00
-6.20130002e-01 -7.18732417e-01 -4.87378120e-01 -9.02702749e-01
6.74989104e-01 -9.70701993e-01 -9.81164098e-01 -2.53694039e-03
-7.88011789e-01 4.25950259e-01 2.85503805e-01 9.99583364e-01
-1.67461544e-01 6.66136324e-01 -4.06994551e-01 -8.29290390e-01
-1.69017822e-01 -1.23443246e+00 4.23292845e-01 4.00088608e-01
-1.10346869e-01 -5.26809454e-01 -2.64132917e-01 7.17114687e-01
3.05311605e-02 -2.09081560e-01 5.36398709e-01 -8.67465258e-01
-4.06904906e-01 -2.23276809e-01 1.09869577e-01 -4.31071676e-04
3.15716773e-01 5.91953993e-01 -8.45540464e-01 -9.31185633e-02
-1.47389933e-01 2.24439308e-01 5.57091773e-01 3.43107492e-01
1.60754538e+00 -5.49737692e-01 -3.41770619e-01 1.62864774e-01
9.83346939e-01 4.83215034e-01 5.92360437e-01 1.84819788e-01
7.30707645e-01 8.02016795e-01 6.57988667e-01 7.41709054e-01
2.16122642e-01 7.16615915e-01 3.72148097e-01 4.47308898e-01
1.76205590e-01 -1.91448152e-01 4.79034841e-01 1.14686728e+00
-5.01764297e-01 -4.46294636e-01 -6.23764038e-01 3.45659524e-01
-1.69487906e+00 -1.32411063e+00 -1.82827994e-01 2.42623258e+00
4.99131411e-01 -1.27761021e-01 3.71556669e-01 3.66971344e-01
9.39175189e-01 -1.90073326e-01 2.65378468e-02 -1.00690164e-01
2.14639351e-01 -1.18523836e-01 8.30057487e-02 4.24854308e-02
-9.52999830e-01 6.29740715e-01 4.88377428e+00 8.69541466e-01
-9.67291772e-01 -1.93178684e-01 5.92291772e-01 -2.55301982e-01
-3.47540289e-01 -4.49610800e-01 -9.77585971e-01 8.22012126e-01
9.55636442e-01 -2.82337852e-02 4.42532539e-01 7.02998638e-01
3.24359447e-01 -3.48901153e-02 -8.23640227e-01 1.11915064e+00
7.29545534e-01 -1.24922299e+00 2.53172606e-01 -2.88929511e-02
6.17371559e-01 -6.41875863e-01 -2.50836499e-02 1.55967072e-01
-6.25862107e-02 -6.04764163e-01 5.74831247e-01 6.05603874e-01
3.18750918e-01 -1.07433796e+00 7.09677577e-01 3.51253182e-01
-1.08089149e+00 -2.08780274e-01 -3.65331173e-01 1.32130280e-01
-4.41433728e-01 6.10322475e-01 -6.00594938e-01 4.48907256e-01
1.09674442e+00 7.37537086e-01 -3.10208857e-01 1.41800582e+00
-1.01653673e-01 6.94025517e-01 -1.58534572e-01 -2.74982423e-01
-4.32164967e-02 -5.54456413e-01 3.64280373e-01 1.04800200e+00
1.00095713e+00 1.91401899e-01 1.60108358e-01 1.29029527e-01
-2.50777066e-01 7.27349043e-01 -5.76391101e-01 1.76492959e-01
7.82509327e-01 1.50901747e+00 -9.56483901e-01 -1.23681210e-01
-5.63535810e-01 5.92760503e-01 -4.11102325e-02 1.56240717e-01
-7.51927674e-01 -3.50874364e-01 5.01202226e-01 3.37195635e-01
3.17858905e-01 9.29783359e-02 5.20644784e-01 -1.01978064e+00
-8.47454518e-02 -9.77632761e-01 5.57191193e-01 -5.88381469e-01
-1.00817657e+00 6.78371429e-01 -3.93471658e-01 -1.57876456e+00
-1.11475833e-01 -2.28439584e-01 -4.66576159e-01 3.06009352e-01
-9.43041980e-01 -5.58193564e-01 -5.55608690e-01 9.31487858e-01
5.28887928e-01 -5.20636439e-01 6.90194011e-01 8.24373066e-01
-5.94013691e-01 6.28885090e-01 2.13118896e-01 2.20636427e-01
8.24043334e-01 -6.56350315e-01 -7.30542064e-01 6.61961138e-01
7.02679694e-01 6.27837420e-01 6.18112445e-01 -4.12030905e-01
-1.53694117e+00 -8.81019652e-01 8.05857360e-01 -1.90246582e-01
5.04596233e-01 6.79194331e-02 -7.97597110e-01 2.68483728e-01
7.77392164e-02 -4.28783447e-01 1.17338121e+00 1.92804843e-01
-1.40636787e-01 -3.80905688e-01 -1.00045693e+00 6.10006034e-01
1.01004505e+00 -3.95152688e-01 -3.96973997e-01 4.09786403e-01
2.37548321e-01 -1.85877159e-01 -1.08493483e+00 -2.15468947e-02
4.47005063e-01 -8.88131678e-01 8.51616383e-01 -2.06395715e-01
4.46227700e-01 -4.50502574e-01 -4.15534116e-02 -1.16381001e+00
-3.44326049e-01 -4.89038616e-01 -2.51256600e-02 1.86551332e+00
1.02833442e-01 -8.99561495e-02 8.63234818e-01 5.35692573e-01
-5.63124940e-03 -1.93806693e-01 -8.16300690e-01 -2.64678746e-01
-6.94141984e-01 2.22984701e-02 6.69725895e-01 1.10430789e+00
3.08797926e-01 1.82482481e-01 -4.37977463e-01 1.57185867e-01
2.22430974e-01 2.54109710e-01 5.72898626e-01 -1.56024647e+00
-1.48486551e-02 -4.97446865e-01 -5.19124448e-01 -6.90543234e-01
1.65866405e-01 -7.53431737e-01 4.83907126e-02 -1.36036670e+00
1.86061695e-01 -2.40386292e-01 -4.72488910e-01 7.70513490e-02
6.07874170e-02 2.77119786e-01 -4.21275049e-02 3.16011190e-01
-9.12290812e-01 1.16489306e-01 9.64421749e-01 1.22122623e-01
-3.40392500e-01 4.13059771e-01 -6.60072565e-01 6.92601264e-01
8.10438156e-01 -3.71102065e-01 -3.97466749e-01 -1.44389957e-01
3.44856590e-01 1.42712563e-01 -2.07884256e-02 -1.18747342e+00
4.08940703e-01 -1.36371672e-01 4.11323667e-01 -4.93923813e-01
1.65469378e-01 -1.08792675e+00 3.11696619e-01 7.44505823e-02
-4.61917132e-01 -1.12348124e-01 -2.90159523e-01 5.01299381e-01
-3.84888530e-01 -5.59846640e-01 2.38244787e-01 -9.59854349e-02
-7.69248903e-01 2.40978941e-01 -8.55530441e-01 -3.72792065e-01
1.13429356e+00 -1.60518140e-01 -1.29347891e-01 -6.07622802e-01
-6.98808789e-01 -1.62391126e-01 1.02918789e-01 5.81685007e-01
8.46194923e-01 -1.03185141e+00 -5.81597209e-01 8.72238129e-02
2.01245591e-01 -4.62135494e-01 5.67445397e-01 6.78287685e-01
-3.20189565e-01 1.25579163e-01 -7.39021078e-02 -3.02522272e-01
-1.90904236e+00 5.53479791e-01 -1.25722408e-01 3.38484406e-01
-5.23944974e-01 6.49303913e-01 -1.57483965e-01 5.50289266e-02
6.48270309e-01 1.01844728e-01 -1.23363709e+00 7.47743368e-01
8.32281768e-01 5.25429428e-01 3.02155614e-01 -9.61296022e-01
-2.94685751e-01 6.20892644e-01 2.49264315e-02 4.22945946e-01
1.20654249e+00 -2.40876347e-01 1.07226837e-02 4.16213311e-02
8.62515628e-01 4.01417136e-01 -7.97891259e-01 -2.76140869e-01
-5.67135550e-02 -5.80092549e-01 2.04397798e-01 -3.79796892e-01
-1.45062649e+00 6.59828842e-01 5.62491179e-01 2.72373766e-01
1.07395315e+00 -8.78242403e-02 3.60300422e-01 4.28176373e-01
3.09953451e-01 -1.08338523e+00 -1.43161975e-02 2.41581246e-01
4.65476543e-01 -9.92284119e-01 9.22604501e-02 -4.73835498e-01
-5.48829913e-01 1.37182796e+00 5.83192825e-01 2.08503842e-01
8.60677004e-01 -1.42814204e-01 2.80881952e-02 -4.84149791e-02
-5.14562488e-01 -1.87572435e-01 6.00540698e-01 3.58617067e-01
7.81116962e-01 -2.45135240e-02 -5.63257992e-01 7.29953229e-01
-1.64209276e-01 3.02870814e-02 6.13881230e-01 5.09222507e-01
-6.11252725e-01 -7.79607058e-01 -4.71294194e-01 7.21524954e-01
-7.81237841e-01 1.63791403e-01 -2.34373465e-01 1.12993680e-01
1.43805757e-01 1.49827528e+00 2.49824688e-01 -6.55085325e-01
1.83481157e-01 -9.92210880e-02 1.44459739e-01 -6.14962161e-01
-7.35555410e-01 2.94433624e-01 -4.72436771e-02 -2.80417234e-01
-6.95222735e-01 -6.97217286e-01 -1.03247142e+00 -5.25434554e-01
-2.89916337e-01 7.74352193e-01 5.94587088e-01 8.54638398e-01
3.60209227e-01 5.26881218e-01 8.90617549e-01 -5.94462872e-01
-5.43774338e-03 -6.32773638e-01 -6.05862379e-01 5.66719711e-01
-1.66203707e-01 -6.64204895e-01 -4.04606044e-01 2.53983349e-01] | [10.148720741271973, 5.754115104675293] |
d24f6e1c-0806-4718-9c55-a5d1f77e58a9 | fact-checking-or-psycholinguistics-how-to | null | null | https://aclanthology.org/D19-6602 | https://aclanthology.org/D19-6602.pdf | Fact Checking or Psycholinguistics: How to Distinguish Fake and True Claims? | The goal of our paper is to compare psycholinguistic text features with fact checking approaches to distinguish lies from true statements. We examine both methods using data from a large ongoing study on deception and deception detection covering a mixture of factual and opinionated topics that polarize public opinion. We conclude that fact checking approaches based on Wikipedia are too limited for this task, as only a few percent of sentences from our study has enough evidence to become supported or refuted. Psycholinguistic features turn out to outperform both fact checking and human baselines, but the accuracy is not high. Overall, it appears that deception detection applicable to less-than-obvious topics is a difficult task and a problem to be solved. | ["Justyna Sarzy{\\'n}ska-Wawer", 'er', 'Aleks Wawer', 'Grzegorz Wojdyga'] | 2019-11-01 | null | null | null | ws-2019-11 | ['deception-detection'] | ['miscellaneous'] | [-3.83443654e-01 4.74859893e-01 -2.45269641e-01 -7.53706574e-01
-9.04974341e-01 -9.69805181e-01 1.02275753e+00 6.64230108e-01
-4.21464384e-01 1.14800322e+00 5.10385692e-01 -7.75372982e-01
1.09493241e-01 -3.37019891e-01 -5.73269486e-01 -6.16147369e-02
3.46881032e-01 1.61230206e-01 -5.69918565e-03 -5.78178167e-01
1.01958430e+00 3.05388361e-01 -1.22042406e+00 4.57676351e-01
1.26862204e+00 6.72454000e-01 -7.99000442e-01 4.43591028e-01
2.52393186e-02 1.38257551e+00 -1.29698992e+00 -1.39814425e+00
-1.90161020e-02 -4.36916292e-01 -1.16332626e+00 -5.85506000e-02
1.43064964e+00 -5.55325091e-01 -1.94175258e-01 1.30568159e+00
3.58470351e-01 -1.17553212e-01 7.31447220e-01 -1.00249577e+00
-1.19280732e+00 6.51830256e-01 -4.31453019e-01 8.53529692e-01
8.44857514e-01 1.85964316e-01 1.06426418e+00 -7.80283093e-01
5.93609452e-01 1.39505303e+00 4.85868543e-01 3.77426445e-01
-7.90752888e-01 -7.29232907e-01 4.40376736e-02 3.71218503e-01
-1.01271176e+00 -9.86347973e-01 6.18823349e-01 -6.12056315e-01
9.26434040e-01 2.98698425e-01 5.02137601e-01 1.38116395e+00
6.13313317e-01 5.56485713e-01 1.69886529e+00 -2.94307500e-01
2.92039782e-01 6.86357200e-01 7.70417392e-01 7.76217878e-01
9.28433359e-01 -5.85314482e-02 -1.13357162e+00 -7.45238006e-01
-4.94442582e-02 -9.41153407e-01 -5.58748662e-01 1.80738658e-01
-8.68893445e-01 1.22022462e+00 1.20150059e-01 3.09823185e-01
-3.87803614e-01 -2.90837616e-01 5.38967252e-01 6.31721258e-01
9.24974084e-01 9.14396524e-01 -2.42510706e-01 -5.43062210e-01
-1.26038015e+00 7.05820799e-01 1.48135507e+00 4.12990540e-01
2.92963058e-01 9.18785259e-02 -3.77068184e-02 5.32584071e-01
-8.28953385e-02 4.17352438e-01 3.56602877e-01 -8.04572821e-01
5.98748326e-01 5.39140761e-01 6.00270212e-01 -1.75279570e+00
-1.93967879e-01 -2.31789604e-01 -1.33697242e-01 3.36261004e-01
7.57794917e-01 -3.53631228e-01 -4.97390389e-01 1.34884810e+00
2.46596143e-01 -6.43570065e-01 -2.46975068e-02 1.00342989e+00
8.73322487e-01 3.31537426e-01 -2.55651642e-02 -1.31186619e-01
1.80523813e+00 -3.04155409e-01 -8.34947884e-01 -6.28413022e-01
5.92609644e-01 -7.58894801e-01 7.81229198e-01 6.87295139e-01
-1.23650289e+00 6.26285821e-02 -1.27296448e+00 -4.07825261e-01
-5.39345741e-01 -2.16912314e-01 9.37854111e-01 1.21052825e+00
-7.07384765e-01 3.77456337e-01 -9.69318822e-02 -2.48511270e-01
6.55843794e-01 -3.53155673e-01 -4.66807902e-01 5.74246384e-02
-1.61379075e+00 1.64487529e+00 3.71079564e-01 1.78061441e-01
-5.59125245e-01 -3.51667583e-01 -1.10554671e+00 -6.52951300e-02
4.96885836e-01 -3.74114245e-01 1.07022214e+00 -1.01442587e+00
-1.07504523e+00 1.47841287e+00 -3.06709409e-01 -6.71319902e-01
6.06933653e-01 -3.46452624e-01 -7.90772915e-01 2.55976170e-01
3.88325363e-01 -1.16515653e-02 9.66012418e-01 -1.09142983e+00
-4.44013715e-01 -5.98048508e-01 2.77153224e-01 1.00303106e-01
-1.50927007e-01 5.33273757e-01 7.81051517e-01 -4.18400019e-01
-1.83524802e-01 -3.79363745e-01 5.47472715e-01 -2.01023474e-01
-6.14732802e-01 -3.55022907e-01 6.04728043e-01 -1.06179166e+00
1.05772376e+00 -1.61810637e+00 -5.76025069e-01 -2.43657287e-02
6.47415459e-01 4.18002129e-01 4.05096501e-01 5.03597081e-01
-2.63532512e-02 6.87198758e-01 1.01328008e-01 1.59413815e-01
1.10331446e-01 -2.68918633e-01 -5.15772760e-01 8.22650552e-01
1.52288750e-01 1.04956222e+00 -9.03441966e-01 -4.58826244e-01
-2.00261340e-01 -1.11724682e-01 -6.05659075e-02 -3.33985537e-01
1.06653079e-01 -2.81674474e-01 -2.28163466e-01 7.67611504e-01
8.03201795e-01 1.93913728e-02 3.69792106e-03 1.08519427e-01
-1.86318904e-01 1.13748801e+00 -5.60616374e-01 9.77785945e-01
1.34253979e-01 1.01224482e+00 2.82202154e-01 -6.82271540e-01
8.62332404e-01 6.74003586e-02 -6.23381078e-01 -5.34112096e-01
2.76304156e-01 2.86753863e-01 3.38054568e-01 -8.36793959e-01
8.08704615e-01 -7.14849651e-01 -1.94076121e-01 6.67715371e-01
-1.89030647e-01 -4.33400422e-01 9.59050208e-02 7.50070274e-01
9.28372324e-01 -4.76732969e-01 6.83336675e-01 -7.34188616e-01
4.97713685e-01 4.89613354e-01 4.14169848e-01 1.05604279e+00
-6.47936583e-01 1.92572817e-01 9.99008417e-01 -4.52564180e-01
-6.92728758e-01 -7.72175491e-01 -3.72559160e-01 6.95815802e-01
-5.31482045e-03 -3.10756266e-01 -6.52145445e-01 -1.11626172e+00
1.57056987e-01 1.56787372e+00 -9.65227127e-01 -1.90532222e-01
-3.40973842e-03 -6.17757380e-01 6.23009801e-01 -5.89635260e-02
5.36178470e-01 -7.01716185e-01 -6.46248281e-01 -6.87605888e-02
-3.26073378e-01 -1.16698647e+00 -2.15676188e-01 -2.54930317e-01
-4.85796630e-01 -1.28966427e+00 -1.53808251e-01 -2.89128512e-01
3.11713219e-01 1.94093525e-01 1.44328046e+00 1.58544019e-01
1.14511892e-01 1.72766224e-01 -3.01224798e-01 -8.34629357e-01
-7.80522466e-01 -3.16234916e-01 5.71472533e-02 -1.88225493e-01
1.01720631e+00 -3.06932211e-01 -3.50228518e-01 -1.40138015e-01
-6.48687780e-01 -5.37673771e-01 3.17327112e-01 7.65973747e-01
-6.19375765e-01 -1.77528247e-01 6.85901642e-01 -1.11443245e+00
1.34779310e+00 -6.63701117e-01 -2.16936380e-01 5.70430048e-02
-4.48793381e-01 -2.02853918e-01 5.01994610e-01 -8.45746323e-02
-1.03731835e+00 -9.03951228e-01 1.02206558e-01 3.14218611e-01
-3.64064217e-01 6.83550894e-01 1.70084640e-01 -3.17542143e-02
1.16604924e+00 2.82079168e-02 2.17704713e-01 3.36803421e-02
1.85972869e-01 8.15109372e-01 4.61356491e-01 -3.90163511e-01
7.66079426e-01 4.10214573e-01 -5.59445858e-01 -8.35314214e-01
-1.50701773e+00 -1.81545064e-01 -4.13348675e-01 -4.52597886e-02
2.78052360e-01 -8.23482752e-01 -9.92363095e-01 1.46948829e-01
-1.33701086e+00 2.05607355e-01 1.07038192e-01 3.55159134e-01
-7.68395141e-02 7.32795060e-01 -8.23096156e-01 -1.19009268e+00
-3.69361818e-01 -4.81418967e-01 6.89478457e-01 -1.79687366e-01
-7.17881083e-01 -9.64619040e-01 1.92763694e-02 9.71174657e-01
1.96184859e-01 4.44607466e-01 7.81584203e-01 -1.23687994e+00
-3.76067944e-02 -7.13375628e-01 -1.96459472e-01 3.96045476e-01
5.52133322e-02 1.03868604e-01 -9.42610323e-01 -7.04914406e-02
6.65345073e-01 -1.16364276e+00 9.16921616e-01 2.33793650e-02
4.59952891e-01 -8.10940325e-01 5.11762872e-03 -3.64655316e-01
9.70432639e-01 -2.75695562e-01 6.82721794e-01 3.18529636e-01
2.11236373e-01 9.01161075e-01 5.33353627e-01 3.05898219e-01
7.24509656e-01 3.30542475e-01 1.70522526e-01 4.69064176e-01
1.41306996e-01 -2.10723534e-01 6.01456583e-01 1.85180217e-01
3.33404750e-01 -4.17487085e-01 -1.10354829e+00 7.34342515e-01
-1.51720142e+00 -1.48367512e+00 -5.24769127e-01 1.88651788e+00
8.20232749e-01 4.40094143e-01 2.17971742e-01 1.03559367e-01
8.15772295e-01 4.23056364e-01 -1.37752250e-01 -1.10504448e+00
-2.82338530e-01 -1.81124151e-01 -6.98410869e-02 7.58158863e-01
-8.33067954e-01 8.60663235e-01 7.42053747e+00 6.44158304e-01
-7.98889458e-01 1.07723325e-01 8.32984686e-01 -6.39642924e-02
-4.82861549e-01 -5.77139780e-02 -5.50742745e-01 6.56340063e-01
1.08466828e+00 -3.32188338e-01 -7.41798058e-02 7.12746084e-01
3.91094834e-01 -8.40812922e-01 -1.02240121e+00 6.88790739e-01
8.10931861e-01 -1.02235472e+00 5.48598617e-02 -6.73647225e-02
4.40545499e-01 -2.63307512e-01 3.94816920e-02 1.76960006e-01
5.34519374e-01 -1.26008677e+00 1.16662908e+00 1.63609594e-01
3.62947613e-01 -5.96342921e-01 1.03346312e+00 6.00246370e-01
2.59702861e-01 1.49387181e-01 -4.65434045e-01 -7.45976627e-01
2.77837783e-01 1.08693826e+00 -6.23678029e-01 2.71229744e-01
5.25990009e-01 4.32241708e-01 -8.81531596e-01 7.15405405e-01
-6.40562177e-01 7.75144041e-01 -3.42434905e-02 -5.77174962e-01
3.54174435e-01 1.89091668e-01 7.22348750e-01 1.07752883e+00
-1.07225575e-01 2.27322981e-01 -3.05425048e-01 1.11211383e+00
-2.36875504e-01 -1.35544669e-02 -8.68307233e-01 -2.38659605e-01
3.20550531e-01 1.30526567e+00 -3.40108544e-01 -4.55204159e-01
-3.13165128e-01 8.85997832e-01 8.53457034e-01 1.06089786e-01
-5.63519001e-01 -3.26881289e-01 5.00127256e-01 3.12409736e-02
-4.46325429e-02 -8.64164829e-02 -5.30665755e-01 -1.58226764e+00
3.12755585e-01 -1.16454768e+00 3.92289728e-01 -8.35397720e-01
-1.59387410e+00 3.63644987e-01 -2.58629709e-01 -5.20036697e-01
-2.36205772e-01 -7.42244065e-01 -6.58685267e-01 8.51593673e-01
-1.45427048e+00 -7.85764754e-01 -9.67274606e-02 2.50570804e-01
1.27800658e-01 2.18441218e-01 4.69646513e-01 -4.07682240e-01
-4.13253158e-01 4.99128670e-01 -3.38201374e-01 1.73421487e-01
8.21697593e-01 -1.28147280e+00 4.20113862e-01 9.88507390e-01
1.98687524e-01 1.00287175e+00 1.18237972e+00 -9.04667795e-01
-9.46734488e-01 -2.52521396e-01 1.45449030e+00 -1.34357750e+00
1.07262564e+00 -5.43823659e-01 -1.10871387e+00 6.71976805e-01
7.21506417e-01 -3.80648673e-01 7.68489420e-01 6.22433901e-01
-9.73115265e-01 3.91861230e-01 -1.44320667e+00 4.88572985e-01
7.86917090e-01 -8.08821082e-01 -1.42569828e+00 4.78322893e-01
-1.18665502e-03 -3.46033156e-01 -4.85883325e-01 -2.35326633e-01
5.24550796e-01 -1.27920556e+00 3.65739077e-01 -8.97262573e-01
9.28383410e-01 1.42606616e-01 3.73352885e-01 -1.51913285e+00
-3.01098824e-01 -6.91380858e-01 -3.68015356e-02 1.36506855e+00
6.93082750e-01 -9.25420046e-01 7.08750367e-01 1.04756975e+00
1.06688276e-01 -3.08333457e-01 -9.86475289e-01 -6.25511408e-01
6.64200068e-01 -2.75760591e-01 -2.31146067e-03 1.40781748e+00
8.85378718e-01 7.17365205e-01 -4.03615087e-01 5.29902801e-02
4.96159792e-01 2.92976230e-01 5.05644441e-01 -1.10724473e+00
2.86297739e-01 -2.77479321e-01 -5.05929887e-01 -6.31987035e-01
5.91685236e-01 -5.90180635e-01 -7.26657733e-02 -1.31065333e+00
4.78252620e-01 2.43081763e-01 2.31280759e-01 3.61578017e-01
-5.25324345e-01 1.46239012e-01 -9.23557729e-02 -9.93407294e-02
-4.92246509e-01 4.87268746e-01 1.04736006e+00 -2.46575996e-01
2.94317424e-01 -2.83042163e-01 -1.57166326e+00 7.89910138e-01
8.22625101e-01 -4.63607252e-01 -1.18708163e-01 -3.21797788e-01
4.84530360e-01 -6.89450577e-02 8.73081326e-01 -5.01782060e-01
2.83846647e-01 -1.18524238e-01 4.42002386e-01 -2.45174512e-01
3.45798373e-01 -1.95462599e-01 -8.18816483e-01 1.38456061e-01
-3.57495755e-01 7.95772374e-02 3.04139644e-01 5.62749147e-01
-2.48980358e-01 -3.91751379e-01 5.41755199e-01 -3.36843193e-01
-3.57331902e-01 -3.56399179e-01 -8.48038137e-01 5.67803800e-01
7.23496139e-01 -3.25029612e-01 -1.11347771e+00 -1.06725991e+00
-3.47289145e-01 -7.71855982e-03 5.43008149e-01 2.56835163e-01
6.56350970e-01 -8.39176595e-01 -1.26517200e+00 -4.97423351e-01
1.91870525e-01 -9.66903925e-01 5.04881367e-02 9.19817448e-01
-3.57320487e-01 5.85056543e-01 -1.01575322e-01 1.96502849e-01
-1.11475563e+00 6.07897222e-01 2.96837479e-01 2.43676871e-01
-3.07908237e-01 8.63998830e-01 -2.12608159e-01 -1.73162207e-01
-2.82843143e-01 2.35933810e-03 -2.51059085e-01 3.06421727e-01
9.63137507e-01 4.29133505e-01 2.56921113e-01 -9.35534358e-01
-5.40488541e-01 -2.48749822e-01 -5.91373801e-01 -2.25659639e-01
9.34789062e-01 -2.57708758e-01 -4.45374578e-01 5.57992995e-01
7.96120107e-01 6.75472021e-01 -4.27703351e-01 1.74581900e-01
2.94376537e-02 -9.10485566e-01 9.23941657e-02 -1.41257739e+00
-4.28394586e-01 7.23319411e-01 -3.64760250e-01 8.17731023e-01
2.61113375e-01 2.92507582e-03 3.88889641e-01 3.97345394e-01
4.11821961e-01 -1.05541265e+00 -1.32300198e-01 4.09750551e-01
1.27696896e+00 -1.57576168e+00 4.44590122e-01 -5.06056368e-01
-8.80787075e-01 1.03143561e+00 7.25327611e-01 -9.23806056e-02
2.85496563e-01 -1.12911306e-01 1.75757572e-01 -8.90054166e-01
-7.83827901e-01 1.40276104e-01 3.44041407e-01 4.29664463e-01
6.73525333e-01 -2.73146629e-01 -1.03588450e+00 5.32697082e-01
-8.31410646e-01 -4.06008422e-01 1.12445462e+00 6.38497233e-01
-5.12282372e-01 -4.53181744e-01 -6.05103910e-01 8.47381294e-01
-1.00529265e+00 -4.78156209e-01 -1.32412505e+00 9.37446058e-01
-2.78383076e-01 1.56851637e+00 -2.34812587e-01 -1.49995387e-01
4.16606255e-02 2.04806805e-01 3.92812282e-01 -6.42314315e-01
-8.59286904e-01 -7.65527308e-01 9.55873251e-01 -4.82959598e-01
-2.46813923e-01 -7.70216346e-01 -3.64178538e-01 -8.84056211e-01
-6.22265458e-01 5.20290196e-01 3.77561957e-01 1.11924636e+00
4.49685514e-01 -2.42487401e-01 6.09238632e-02 -1.29241779e-01
-7.78503299e-01 -1.00955427e+00 -8.16664994e-01 4.07975763e-01
4.50459272e-01 -5.22416353e-01 -8.58277082e-01 -5.01930356e-01] | [8.199056625366211, 10.350325584411621] |
b75e4bf9-2500-4b48-9cf4-c7294c43f754 | cpnet-exploiting-clip-based-attention | 2305.13962 | null | https://arxiv.org/abs/2305.13962v1 | https://arxiv.org/pdf/2305.13962v1.pdf | CPNet: Exploiting CLIP-based Attention Condenser and Probability Map Guidance for High-fidelity Talking Face Generation | Recently, talking face generation has drawn ever-increasing attention from the research community in computer vision due to its arduous challenges and widespread application scenarios, e.g. movie animation and virtual anchor. Although persevering efforts have been undertaken to enhance the fidelity and lip-sync quality of generated talking face videos, there is still large room for further improvements of synthesis quality and efficiency. Actually, these attempts somewhat ignore the explorations of fine-granularity feature extraction/integration and the consistency between probability distributions of landmarks, thereby recurring the issues of local details blurring and degraded fidelity. To mitigate these dilemmas, in this paper, a novel CLIP-based Attention and Probability Map Guided Network (CPNet) is delicately designed for inferring high-fidelity talking face videos. Specifically, considering the demands of fine-grained feature recalibration, a clip-based attention condenser is exploited to transfer knowledge with rich semantic priors from the prevailing CLIP model. Moreover, to guarantee the consistency in probability space and suppress the landmark ambiguity, we creatively propose the density map of facial landmark as auxiliary supervisory signal to guide the landmark distribution learning of generated frame. Extensive experiments on the widely-used benchmark dataset demonstrate the superiority of our CPNet against state of the arts in terms of image and lip-sync quality. In addition, a cohort of studies are also conducted to ablate the impacts of the individual pivotal components. | ['Meirong Ma', 'Minghao Li', 'Mingjie Wang', 'Benlai Tang', 'Jingning Xu'] | 2023-05-23 | null | null | null | null | ['talking-face-generation', 'face-generation'] | ['computer-vision', 'computer-vision'] | [ 2.24510923e-01 6.04082830e-02 -1.74784198e-01 -2.71842271e-01
-6.32191181e-01 -2.16304198e-01 6.73223197e-01 -4.75012898e-01
-2.50620488e-02 7.11372554e-01 3.42611194e-01 2.09178969e-01
-1.66700244e-01 -5.43188989e-01 -6.37475133e-01 -9.54384089e-01
1.76290527e-01 -1.56139657e-01 7.46952253e-04 -8.33964199e-02
1.30981684e-01 4.95603263e-01 -1.79971170e+00 4.27477341e-03
1.03544283e+00 1.06533229e+00 3.02340746e-01 1.09416164e-01
-1.83373138e-01 3.39810818e-01 -5.64353704e-01 -7.11682677e-01
1.41538829e-01 -4.43884581e-01 -2.53479898e-01 2.97058165e-01
6.10189915e-01 -3.61006737e-01 -4.84424412e-01 1.35560203e+00
6.45245254e-01 1.10446103e-01 4.30781543e-01 -1.48618603e+00
-5.84802687e-01 3.84365916e-01 -8.34498107e-01 1.28903553e-01
4.66765195e-01 3.57206792e-01 8.34544122e-01 -1.12826121e+00
4.58833575e-01 1.51639283e+00 5.44359565e-01 4.76280987e-01
-9.96323347e-01 -1.01614332e+00 3.24571848e-01 4.78558451e-01
-1.67177713e+00 -8.81612003e-01 1.13523006e+00 -2.31343612e-01
2.65656680e-01 2.20794305e-01 5.63154936e-01 1.18825936e+00
1.01867042e-01 5.94749510e-01 6.73011482e-01 -2.29198277e-01
3.98701653e-02 3.00681084e-01 -5.50813675e-01 7.53782451e-01
8.53316858e-02 2.05710292e-01 -7.98024297e-01 8.11437964e-02
9.65594471e-01 -5.25863320e-02 -6.70165181e-01 -3.49199533e-01
-1.08479512e+00 5.46792686e-01 4.02305812e-01 2.77179211e-01
-3.62412870e-01 -2.55611558e-02 2.40066797e-01 -1.95885316e-01
4.39668953e-01 1.64313063e-01 -1.50601482e-02 -1.43411234e-01
-1.05665290e+00 -3.85409710e-03 4.15724486e-01 1.06847501e+00
5.48513472e-01 2.12831452e-01 -3.56230021e-01 7.84810483e-01
3.28919470e-01 4.62210238e-01 1.47138149e-01 -1.09698462e+00
5.06856382e-01 2.01969877e-01 7.21752420e-02 -1.61819541e+00
-2.07965411e-02 -4.12200838e-01 -1.09773874e+00 7.70351440e-02
2.42554888e-01 -6.30947873e-02 -6.05769455e-01 2.00006270e+00
4.54910487e-01 5.95339060e-01 -2.93858588e-01 1.12887192e+00
5.96897721e-01 6.80330813e-01 1.43987209e-01 -5.60806453e-01
1.34061730e+00 -7.45519161e-01 -1.14252949e+00 8.47146809e-02
-3.38723063e-01 -7.74646103e-01 1.01330471e+00 2.36687809e-01
-1.14488542e+00 -8.72426152e-01 -9.32945311e-01 1.93510037e-02
1.52980611e-02 -1.55874118e-02 5.30778944e-01 5.76624095e-01
-8.78998041e-01 4.89013344e-01 -6.30532503e-01 -8.60410854e-02
7.45760977e-01 1.84941024e-01 -3.39790761e-01 -1.21043764e-01
-1.19972467e+00 6.01530492e-01 6.36139140e-02 4.97650564e-01
-7.63961792e-01 -8.68902087e-01 -7.72683382e-01 2.39946619e-01
3.95919085e-01 -6.83777928e-01 8.68381500e-01 -9.28532898e-01
-1.76686549e+00 4.00269538e-01 -1.72843367e-01 3.58779845e-03
6.45197093e-01 -1.91744883e-02 -4.49957371e-01 2.97047317e-01
6.06252663e-02 9.52413440e-01 1.30041254e+00 -1.29443145e+00
-7.81324983e-01 -2.47809365e-01 -4.02102945e-03 3.37749004e-01
-5.70063710e-01 -1.55254528e-01 -7.32720077e-01 -9.20739532e-01
6.24368079e-02 -7.04246044e-01 1.20518349e-01 4.60141629e-01
-3.33355963e-01 -3.14862788e-01 8.57126951e-01 -5.17547846e-01
1.12669802e+00 -2.30616951e+00 2.86548436e-01 -1.52426988e-01
1.88092828e-01 3.00424516e-01 -1.67199492e-01 1.42095564e-02
3.68618034e-02 8.92056972e-02 -1.70899883e-01 -5.75472176e-01
3.09597538e-03 2.18266025e-02 -5.43963671e-01 6.07778847e-01
5.18812835e-01 7.50291765e-01 -8.92542124e-01 -6.90664113e-01
4.52734321e-01 1.09113610e+00 -6.49146795e-01 2.25518897e-01
-1.03442110e-01 6.38648212e-01 -5.03201842e-01 7.05308080e-01
9.02622402e-01 -2.10500509e-02 -5.29646464e-02 -7.19503760e-01
-1.46222161e-02 -1.50056541e-01 -1.16111171e+00 1.92813551e+00
-4.67232436e-01 5.66023707e-01 3.09510738e-01 -6.94533706e-01
8.54166269e-01 3.70371461e-01 4.50423300e-01 -5.79085946e-01
1.99926287e-01 -1.00729004e-01 -1.74885541e-01 -5.72494805e-01
4.26832527e-01 -1.32793695e-01 1.54786021e-01 -1.74370445e-02
1.26944914e-01 -2.27568269e-01 -1.96552590e-01 1.13252020e-02
4.75504428e-01 2.67241836e-01 7.52167404e-02 -1.33563519e-01
8.35807443e-01 -5.51114559e-01 7.58248925e-01 2.00075164e-01
-4.32992756e-01 8.32542717e-01 3.49584073e-01 -1.30214706e-01
-7.85075307e-01 -1.03133523e+00 -3.60464156e-01 8.76521528e-01
5.04544318e-01 -4.10294354e-01 -7.53422379e-01 -5.18213570e-01
-2.20834881e-01 5.51672876e-01 -4.93713707e-01 -2.64754146e-01
-5.43035448e-01 -4.66015518e-01 4.68831569e-01 2.66746402e-01
7.45403945e-01 -1.10533154e+00 -3.98936272e-01 1.74337208e-01
-4.22807306e-01 -1.23515368e+00 -8.13540697e-01 -5.64803421e-01
-4.00090605e-01 -9.81483698e-01 -8.49563360e-01 -6.72578156e-01
6.89034939e-01 4.28708792e-01 7.04752505e-01 -1.03850938e-01
-2.26946920e-01 2.79654354e-01 -2.45057210e-01 -1.80615857e-01
-1.76064372e-01 -2.25525007e-01 4.60353531e-02 4.64250237e-01
9.31547359e-02 -8.06508899e-01 -7.93219864e-01 3.81014377e-01
-7.69684076e-01 1.73006117e-01 5.06832421e-01 8.39810610e-01
3.83503616e-01 2.98585892e-01 7.84450889e-01 -2.51177192e-01
5.12210548e-01 -4.29899663e-01 -5.82732499e-01 2.21462905e-01
-3.18328619e-01 -2.23413527e-01 7.53137648e-01 -5.55968583e-01
-1.31786537e+00 -1.50211126e-01 -1.57277361e-01 -7.48355746e-01
-7.65068159e-02 4.29281481e-02 -6.81489170e-01 -8.32784846e-02
1.69524312e-01 2.90227443e-01 1.81494236e-01 -1.85378134e-01
4.89734888e-01 4.94620264e-01 7.44447351e-01 -5.47087133e-01
8.56120884e-01 6.08841181e-01 -9.29652601e-02 -7.18747914e-01
-6.02847934e-01 8.07624236e-02 -2.74010777e-01 -4.55399573e-01
7.03759968e-01 -9.85284626e-01 -9.85204041e-01 5.34303427e-01
-1.16178668e+00 3.27480972e-01 -1.95732757e-01 3.15909803e-01
-5.66095889e-01 5.66703439e-01 -3.18331659e-01 -6.86960578e-01
-1.88248947e-01 -1.49317288e+00 1.02462280e+00 5.37612557e-01
7.39062205e-02 -6.04304790e-01 -3.27268511e-01 3.04031223e-01
4.96369362e-01 1.58410802e-01 6.18220985e-01 4.96889688e-02
-8.73915970e-01 9.33957398e-02 -4.62511986e-01 2.52436817e-01
4.53818172e-01 6.86151981e-02 -1.20582008e+00 -2.73605555e-01
3.83822545e-02 -1.48391336e-01 5.35862803e-01 3.40063781e-01
1.28952837e+00 -3.58021021e-01 -3.03976089e-01 7.88841963e-01
1.07805920e+00 2.30787218e-01 6.09653294e-01 -1.87903330e-01
6.97845757e-01 8.50758076e-01 5.89737177e-01 6.36928380e-01
4.09171104e-01 9.29602921e-01 4.74576592e-01 2.01245546e-02
-5.48537970e-01 -5.67651331e-01 2.40086973e-01 6.32167757e-01
1.88703723e-02 -1.69191152e-01 -3.35618109e-01 3.76522601e-01
-1.60037351e+00 -9.27995563e-01 6.28497303e-01 2.06005669e+00
9.34956968e-01 -3.79734971e-02 -2.15970159e-01 7.22643062e-02
1.01940584e+00 4.50683057e-01 -5.20788848e-01 2.19985232e-01
-8.47651288e-02 -1.48323402e-01 -1.13010369e-01 5.29786527e-01
-8.36616993e-01 9.17018890e-01 4.93874073e+00 1.27246499e+00
-1.30324197e+00 -5.71907125e-03 7.66003728e-01 1.12527926e-02
-3.21529567e-01 -2.46525437e-01 -8.22402358e-01 8.57236803e-01
4.10185158e-01 -1.65280789e-01 4.90472555e-01 6.27339542e-01
3.81937563e-01 -6.93089375e-03 -8.64585519e-01 1.25404894e+00
2.43127704e-01 -1.32826138e+00 1.24272786e-01 -5.00172377e-02
5.62536955e-01 -5.54235637e-01 3.40298831e-01 1.26012504e-01
-1.68266743e-01 -9.90935028e-01 8.62310052e-01 6.85304761e-01
1.11182463e+00 -9.29203749e-01 4.89232332e-01 4.20843586e-02
-1.21152222e+00 -2.00496018e-02 -3.03690195e-01 2.85334080e-01
3.92751008e-01 4.47773844e-01 -6.58387959e-01 5.36428332e-01
7.10961938e-01 6.78189397e-01 -2.68949360e-01 8.85518670e-01
-2.30287015e-01 3.15045029e-01 -2.55222023e-01 3.30269694e-01
-7.66746104e-02 -1.54157549e-01 6.25387907e-01 9.35243726e-01
3.41856301e-01 2.48789087e-01 -7.03789517e-02 1.07998526e+00
-9.49905664e-02 -1.16809949e-01 -4.38222080e-01 2.13816181e-01
7.63625383e-01 1.28062725e+00 -4.96575296e-01 -1.59818560e-01
-2.96213448e-01 7.17353523e-01 9.62777212e-02 3.92659813e-01
-1.12199295e+00 -3.00409079e-01 9.60064888e-01 1.38664961e-01
4.19899195e-01 -7.61424378e-02 -1.19407892e-01 -1.14936697e+00
1.35745153e-01 -8.09406221e-01 -4.03680131e-02 -6.55371308e-01
-1.25361168e+00 7.03587353e-01 3.80053036e-02 -1.28515351e+00
-1.80836856e-01 -4.54863459e-02 -5.73342800e-01 8.62925768e-01
-1.77626824e+00 -9.57758427e-01 -5.37053585e-01 8.58956039e-01
8.28372717e-01 -4.77032699e-02 4.46867675e-01 4.65727240e-01
-7.86263406e-01 9.19572473e-01 -3.24808538e-01 -7.39232004e-02
8.74702215e-01 -5.62812626e-01 4.25993465e-02 7.81192303e-01
-3.09073050e-02 4.90790367e-01 8.03016782e-01 -4.01461333e-01
-1.46529520e+00 -1.14985728e+00 6.23068035e-01 -1.06231660e-01
4.30474102e-01 -4.14938092e-01 -9.16246951e-01 1.97873533e-01
2.38307327e-01 2.97459424e-01 1.70448020e-01 -4.21275198e-01
-2.21103802e-01 -5.10472953e-01 -1.14806581e+00 8.00812006e-01
1.18458593e+00 -4.50750798e-01 -3.43947142e-01 5.39755560e-02
7.87160397e-01 -1.95275366e-01 -6.80097640e-01 6.80057704e-01
4.82586354e-01 -1.20138872e+00 1.00447893e+00 1.23853618e-02
3.04875344e-01 -4.68490064e-01 -8.70254561e-02 -1.08175743e+00
-1.38162732e-01 -9.49674845e-01 -7.48219490e-02 1.76308763e+00
-8.41561109e-02 -4.39534456e-01 6.62603855e-01 5.26246965e-01
-1.26687586e-01 -7.21261501e-01 -1.14440536e+00 -2.92247474e-01
-2.51143396e-01 -2.92768240e-01 7.56854117e-01 8.57330680e-01
-7.26485774e-02 2.38280624e-01 -6.05948806e-01 2.21979812e-01
6.78059518e-01 1.70786493e-02 7.21165299e-01 -9.44710255e-01
-6.32949993e-02 -5.62115014e-01 -3.97719562e-01 -1.12249970e+00
3.11550498e-01 -5.18049777e-01 1.91109434e-01 -1.06622171e+00
-1.04513941e-02 -4.59118962e-01 -9.74937603e-02 1.27529189e-01
-2.68136948e-01 2.76119649e-01 3.59897017e-01 1.73063397e-01
-5.52362978e-01 1.11296833e+00 1.65742457e+00 -9.17795897e-02
-6.68681562e-02 -9.27330554e-02 -7.36881852e-01 7.15802610e-01
4.37230349e-01 -2.85331130e-01 -7.18207419e-01 -3.83643001e-01
-2.37796962e-01 4.14787203e-01 4.91897911e-01 -9.28718925e-01
5.22265613e-01 -1.97936893e-01 3.61225516e-01 -4.59150374e-01
6.78482354e-01 -8.78697455e-01 1.08815566e-01 -6.72323629e-02
-1.49229243e-01 -1.60076112e-01 5.02802506e-02 7.55569577e-01
-4.22915190e-01 2.85024196e-01 9.36319888e-01 1.89987525e-01
-5.86709499e-01 6.46683276e-01 1.10710800e-01 -2.70893537e-02
9.82519209e-01 -4.05495375e-01 -6.68096691e-02 -4.39025134e-01
-4.06049967e-01 3.92694548e-02 4.01423693e-01 7.57573009e-01
7.53929079e-01 -1.29105854e+00 -6.25548482e-01 5.85036337e-01
-1.62381768e-01 1.66123211e-01 5.47509193e-01 8.02511156e-01
-8.53230581e-02 2.64275849e-01 -3.05337906e-01 -6.82656705e-01
-8.07206929e-01 5.85484922e-01 1.51330113e-01 2.18795568e-01
-7.65156090e-01 8.45105290e-01 5.36916614e-01 1.12064518e-01
6.00069642e-01 -1.22050211e-01 -1.86943397e-01 1.63658306e-01
6.31246030e-01 1.99086189e-01 -1.87842503e-01 -7.36332655e-01
-2.94532239e-01 7.35174060e-01 2.36238949e-02 5.05603328e-02
1.13567269e+00 -5.50356090e-01 8.23442712e-02 4.36650515e-02
1.01858115e+00 1.76499918e-01 -1.96851027e+00 -1.83885217e-01
-3.66884649e-01 -8.30741405e-01 3.72017827e-03 -5.30472457e-01
-1.38529122e+00 8.88634384e-01 5.00980675e-01 -1.12878904e-01
1.30812192e+00 -2.12080836e-01 6.66068971e-01 -2.35688567e-01
4.59893584e-01 -7.04087079e-01 2.00849488e-01 -2.71575861e-02
1.11419654e+00 -1.13232613e+00 -1.53892532e-01 -6.53606772e-01
-5.22573829e-01 9.76336062e-01 7.16589868e-01 -2.46034954e-02
6.14498734e-01 2.83010721e-01 -1.14824019e-01 1.10247850e-01
-5.84947228e-01 1.90421328e-01 2.29193240e-01 7.00789452e-01
1.39450908e-01 -2.67219812e-01 -9.92469639e-02 7.06698477e-01
-2.40538821e-01 1.96748003e-02 2.60790884e-01 3.33599657e-01
-2.39150628e-01 -6.03907108e-01 -4.27003473e-01 -4.69804294e-02
-3.76318932e-01 -1.72327682e-01 2.84327239e-01 8.25364709e-01
3.08913916e-01 9.78853643e-01 1.07797176e-01 -2.80703485e-01
1.81042731e-01 -2.62978166e-01 4.19031829e-01 -1.57083765e-01
-7.20233992e-02 2.58448750e-01 -2.93242455e-01 -6.37501478e-01
-4.97754693e-01 -5.04847586e-01 -9.27598000e-01 -3.96704465e-01
-4.37638074e-01 1.52493149e-01 5.56024373e-01 7.39997268e-01
6.96728528e-01 5.23914814e-01 8.49327266e-01 -1.11343563e+00
-4.75804508e-01 -9.10020053e-01 -4.42923278e-01 2.35957831e-01
4.21464711e-01 -9.26385283e-01 -5.72441161e-01 8.04311708e-02] | [13.153016090393066, -0.3103271722793579] |
6886ca33-a207-41e4-8ff4-abf9491bf90f | grouped-attention-for-content-selection-and | null | null | https://aclanthology.org/2021.findings-emnlp.166 | https://aclanthology.org/2021.findings-emnlp.166.pdf | Grouped-Attention for Content-Selection and Content-Plan Generation | Content-planning is an essential part of data-to-text generation to determine the order of data mentioned in generated texts. Recent neural data-to-text generation models employ Pointer Networks to explicitly learn content-plan given a set of attributes as input. They use LSTM to encode the input, which assumes a sequential relationship in the input. This may be sub-optimal to encode a set of attributes, where the attributes have a composite structure: the attributes are disordered while each attribute value is an ordered list of tokens. We handle this problem by proposing a neural content-planner that can capture both local and global contexts of such a structure. Specifically, we propose a novel attention mechanism called GSC-attention. A key component of the GSC-attention is grouped-attention, which is token-level attention constrained within each input attribute that enables our proposed model captures both local and global context. Moreover, our content-planner explicitly learns content-selection, which is integrated into the content-planner to select the most important data to be included in the generated text via an attention masking procedure. Experimental results show that our model outperforms the competitors by 4.92%, 4.70%, and 16.56% in terms of Damerau-Levenshtein Distance scores on three real-world datasets. | ['Qingjun Cui', 'Rui Zhang', 'Jianzhong Qi', 'Xiaojie Wang', 'Bayu Distiawan Trisedya'] | null | null | null | null | findings-emnlp-2021-11 | ['data-to-text-generation'] | ['natural-language-processing'] | [ 3.03442925e-01 3.96634907e-01 -4.02554154e-01 -3.95278066e-01
-9.43673611e-01 -4.86353606e-01 8.96630347e-01 4.66733754e-01
-5.93119562e-01 8.09470236e-01 9.11114216e-01 -1.69894755e-01
2.59438716e-03 -1.08184218e+00 -8.33392143e-01 -6.34448946e-01
-1.77999586e-03 9.63698208e-01 1.17668010e-01 -2.25145489e-01
5.97754002e-01 1.17668599e-01 -1.66644967e+00 7.62972593e-01
1.02309811e+00 9.96756315e-01 6.21528566e-01 4.79285061e-01
-7.10307598e-01 8.64208221e-01 -8.18484187e-01 -3.04073811e-01
-6.47337884e-02 -4.79376704e-01 -1.19148731e+00 -8.66335556e-02
1.25322059e-01 -3.56590658e-01 -5.64261787e-02 9.29376960e-01
5.88008344e-01 4.22445297e-01 7.51210630e-01 -1.07813275e+00
-9.77695644e-01 1.30620158e+00 -2.55643487e-01 3.42632309e-02
4.97010887e-01 1.57383859e-01 1.39137256e+00 -9.32183266e-01
4.92029756e-01 1.45001686e+00 -8.84191170e-02 6.50795519e-01
-9.00847554e-01 -3.55876595e-01 4.63133514e-01 3.89404237e-01
-1.24284160e+00 -2.93433547e-01 6.55533850e-01 -2.66596347e-01
1.17929912e+00 3.00300986e-01 5.11719763e-01 9.17404056e-01
2.11211130e-01 1.04347527e+00 6.39243245e-01 -4.47570682e-01
3.61625552e-01 -1.81797832e-01 -1.31571233e-01 4.56573397e-01
-2.13142876e-02 -1.23519994e-01 -5.75290740e-01 -1.06883913e-01
6.18634641e-01 -3.87471579e-02 -5.99618554e-02 2.56219566e-01
-1.47117770e+00 8.88659716e-01 6.25984967e-01 7.86023960e-02
-5.18881917e-01 2.85666496e-01 3.66409332e-01 -1.33016884e-01
2.83514619e-01 5.83491385e-01 -3.54076862e-01 -1.10969372e-01
-5.19055963e-01 4.95966733e-01 5.31164408e-01 1.44911289e+00
7.10818648e-01 -4.26202081e-02 -8.39448750e-01 6.36528790e-01
3.12373698e-01 2.97171026e-01 8.56360257e-01 -6.51178002e-01
1.06974339e+00 7.49249101e-01 1.63174108e-01 -7.59345889e-01
-2.51310706e-01 -3.11169952e-01 -9.46861029e-01 -2.48723447e-01
-4.46082428e-02 -1.52876884e-01 -1.07021475e+00 1.82120645e+00
2.74418026e-01 8.09072852e-02 1.26788750e-01 9.32526350e-01
1.03484535e+00 1.00491881e+00 2.58723110e-01 -8.80815312e-02
1.38166499e+00 -1.06631052e+00 -7.40969896e-01 -2.78751850e-01
5.76802135e-01 -5.34357071e-01 1.34547281e+00 6.41543940e-02
-1.25422680e+00 -5.61437309e-01 -8.00634861e-01 -3.42326730e-01
-4.90116894e-01 -2.19537914e-02 2.20901757e-01 6.73969612e-02
-9.09276128e-01 5.00740826e-01 -3.44326228e-01 -8.85530338e-02
2.58001685e-01 2.87681818e-01 -5.56284972e-02 9.25066397e-02
-1.59898114e+00 4.32616055e-01 9.94574845e-01 -2.89973896e-02
-9.64741290e-01 -5.10717392e-01 -8.55300188e-01 4.30386782e-01
5.71991324e-01 -6.99051261e-01 1.29055071e+00 -5.86091816e-01
-1.27984416e+00 4.85528082e-01 -4.03259575e-01 -2.97800273e-01
2.81912893e-01 -2.46518135e-01 -3.05248052e-01 4.93217371e-02
3.50572556e-01 9.84920204e-01 5.59795856e-01 -1.25557292e+00
-9.87142563e-01 -7.14396983e-02 1.09663516e-01 6.71515107e-01
-4.49987203e-02 -4.61373255e-02 -7.53939271e-01 -9.52428281e-01
-3.13357078e-02 -5.07619619e-01 -3.41541380e-01 -5.40293097e-01
-8.85553539e-01 -5.56611300e-01 4.38614219e-01 -3.39428246e-01
1.38569903e+00 -1.78713989e+00 1.13425590e-01 2.62960792e-02
8.29776525e-02 -9.65376198e-03 -3.77295792e-01 8.52096200e-01
8.82366151e-02 5.14452279e-01 -1.78288713e-01 -3.96859407e-01
1.10773414e-01 4.98893149e-02 -6.16119206e-01 -1.37748808e-01
4.11655098e-01 9.46020007e-01 -1.04571009e+00 -6.57076955e-01
-1.69018179e-01 1.78728998e-01 -6.90897226e-01 4.56269920e-01
-7.88275540e-01 2.01036870e-01 -8.01607430e-01 4.32280928e-01
4.23989385e-01 -2.22902700e-01 -1.08407989e-01 9.70976800e-02
-1.36545166e-01 7.55311608e-01 -1.06309175e+00 1.66779792e+00
-2.58726090e-01 1.12531863e-01 -3.70501518e-01 -5.96502066e-01
9.60982442e-01 3.41611713e-01 2.03125209e-01 -7.12504447e-01
1.27452195e-01 1.32496506e-02 7.00743049e-02 -4.20018166e-01
1.06020451e+00 -8.97819996e-02 -4.21535820e-01 6.44582570e-01
-8.52525011e-02 1.42915532e-01 4.50322270e-01 5.34266174e-01
8.14814508e-01 5.90610243e-02 2.54887342e-01 -3.22031945e-01
6.86712623e-01 -1.59503162e-01 6.36369705e-01 8.70217681e-01
2.82716960e-01 6.73882484e-01 7.64352083e-01 -3.87286782e-01
-8.75737906e-01 -5.64795673e-01 3.25603217e-01 1.31708860e+00
1.51517168e-01 -5.66663146e-01 -8.09213281e-01 -8.49709928e-01
-1.99542910e-01 1.14687788e+00 -8.00257683e-01 -1.31259173e-01
-6.90743566e-01 -3.69464546e-01 2.90969312e-01 7.22285807e-01
2.63212323e-01 -1.81197143e+00 -7.34439194e-01 4.95294839e-01
-4.82771099e-01 -9.30340827e-01 -8.92980456e-01 2.15243638e-01
-5.87876081e-01 -8.59241486e-01 -4.24737930e-01 -8.52548480e-01
7.98011005e-01 -5.43593541e-02 1.11402833e+00 1.99832723e-01
3.18578362e-01 -1.22545704e-01 -6.72837317e-01 -4.46337402e-01
-3.28506589e-01 5.17503083e-01 -2.98414201e-01 2.22740859e-01
1.97045788e-01 -2.76721388e-01 -6.17257476e-01 2.23388389e-01
-1.04246986e+00 3.53371352e-01 8.74741375e-01 8.92300963e-01
7.69537210e-01 -8.84861350e-02 6.77201629e-01 -1.09796917e+00
1.01071608e+00 -8.00719738e-01 -2.74449378e-01 2.69502372e-01
-4.69903708e-01 4.21594322e-01 1.05809140e+00 -2.89383829e-01
-1.17778468e+00 3.56031731e-02 -1.75264403e-01 -1.78512484e-01
-2.38010719e-01 9.40536499e-01 -6.82767332e-01 9.05167639e-01
2.38039821e-01 6.38417184e-01 -3.29495698e-01 -4.62701142e-01
4.89062190e-01 5.91412544e-01 5.82458735e-01 -9.90829110e-01
3.68686885e-01 1.28048360e-01 -1.65846571e-01 -3.12799156e-01
-7.70124078e-01 -2.12142318e-01 -5.65517247e-01 -1.00645246e-02
8.63151729e-01 -6.94209635e-01 -7.07180619e-01 4.35727000e-01
-1.28905880e+00 -4.05556291e-01 -3.62160027e-01 1.03169806e-01
-6.66933179e-01 1.12212293e-01 -5.33292830e-01 -5.10852695e-01
-6.17611706e-01 -1.26680660e+00 1.16528463e+00 1.74310774e-01
-8.22843090e-02 -8.25156748e-01 -2.75948077e-01 -1.63858142e-02
3.27622712e-01 1.45592228e-01 1.32321012e+00 -1.01904428e+00
-7.45946884e-01 -4.22148630e-02 -1.85599759e-01 -3.95686716e-01
9.46703702e-02 -1.15855910e-01 -6.18391275e-01 -1.30767718e-01
-4.01181102e-01 -3.59031379e-01 8.71109545e-01 1.44171193e-01
1.65926278e+00 -9.32569802e-01 -3.47782433e-01 3.84015262e-01
1.24601233e+00 5.69971442e-01 7.75953770e-01 2.40759432e-01
8.26815546e-01 8.47439408e-01 7.89513171e-01 6.94087207e-01
5.74458420e-01 5.73786139e-01 8.52837086e-01 8.04409832e-02
8.39856360e-03 -6.95851326e-01 2.17180327e-01 7.63452411e-01
9.07926559e-02 -6.36299133e-01 -8.43253314e-01 7.20834494e-01
-1.93124962e+00 -1.12492788e+00 -5.79072013e-02 2.16361737e+00
1.09970629e+00 2.63241976e-01 -2.65827421e-02 6.42381934e-03
9.32836652e-01 2.01746687e-01 -5.92899323e-01 -3.56986016e-01
-9.68470331e-03 -1.05800621e-01 2.90422708e-01 5.78992784e-01
-9.95624065e-01 1.13905895e+00 5.58228445e+00 9.32487309e-01
-8.78517926e-01 -1.38188735e-01 7.05382764e-01 -1.32631540e-01
-7.85947859e-01 -1.11484505e-01 -1.10880232e+00 8.68997335e-01
7.98001289e-01 -4.43073541e-01 2.73892730e-01 6.40396476e-01
2.62893111e-01 9.52765942e-02 -1.21256959e+00 4.94812310e-01
1.22425914e-01 -1.47816384e+00 6.42738521e-01 1.01222686e-01
6.11659706e-01 -3.46937835e-01 8.33668336e-02 4.93644536e-01
5.67248940e-01 -1.09003186e+00 1.07914329e+00 4.88842279e-01
9.48587358e-01 -9.20557320e-01 6.17276549e-01 5.03635049e-01
-1.34070277e+00 -1.44683778e-01 -4.58853930e-01 7.33150244e-02
2.71013856e-01 2.91303337e-01 -9.60457742e-01 5.29396892e-01
4.82575715e-01 6.08078420e-01 -3.40953827e-01 8.82797539e-01
-4.10418183e-01 3.77198011e-01 1.58257589e-01 -4.24874574e-01
6.75141096e-01 -1.09183798e-02 4.25772756e-01 1.15780556e+00
4.55719501e-01 3.26441884e-01 3.43302101e-01 9.88060653e-01
-3.74858767e-01 2.02311039e-01 -2.52645850e-01 1.21403269e-01
9.65575933e-01 1.01241052e+00 -5.35603404e-01 -5.68785906e-01
-1.40555605e-01 7.36903489e-01 5.29506147e-01 2.72822112e-01
-7.26595104e-01 -7.05238760e-01 4.76871312e-01 9.71655026e-02
5.36290109e-01 1.95395485e-01 -1.36766016e-01 -7.34274030e-01
2.76779756e-04 -8.49299252e-01 5.20150542e-01 -7.56559014e-01
-1.13082361e+00 7.64765203e-01 1.14238665e-01 -1.17592418e+00
-4.67991620e-01 -2.23061070e-01 -8.22433829e-01 1.12626231e+00
-1.47105992e+00 -1.11907136e+00 -2.37692237e-01 6.39236271e-01
9.52819645e-01 -2.40878507e-01 7.43232667e-01 -6.35276884e-02
-4.92631614e-01 6.44830227e-01 -7.82231092e-02 1.19506270e-01
5.15345991e-01 -1.50035238e+00 8.35715115e-01 8.56948674e-01
-7.77778476e-02 7.91457415e-01 4.28618073e-01 -9.04172421e-01
-1.16259778e+00 -1.46714211e+00 1.23764062e+00 -1.71165779e-01
3.29536080e-01 -3.95238042e-01 -9.87267733e-01 7.20331073e-01
5.17060101e-01 -2.25762308e-01 6.02896452e-01 -1.50199577e-01
-1.93660170e-01 1.40415594e-01 -8.92423332e-01 6.36073411e-01
1.00486529e+00 -1.94677263e-01 -7.31982350e-01 1.24846004e-01
1.25129724e+00 -5.94071031e-01 -5.60978949e-01 -7.42953718e-02
1.38789162e-01 -6.56130314e-01 5.54167032e-01 -7.00349748e-01
9.07269776e-01 -3.99870902e-01 -9.78612602e-02 -1.42255092e+00
-6.63287520e-01 -6.54001355e-01 -2.99239010e-01 1.46137655e+00
6.59497142e-01 -3.55693668e-01 7.14178681e-01 5.51680923e-01
-6.72287464e-01 -1.05825567e+00 -6.49731219e-01 -5.40475249e-01
2.08384871e-01 -2.04483971e-01 1.36773241e+00 7.40525901e-01
3.70676517e-02 5.89364409e-01 -3.62068921e-01 -4.45321463e-02
2.95603395e-01 1.55757338e-01 3.85336787e-01 -1.09427083e+00
-1.19243503e-01 -6.36467576e-01 1.82916716e-01 -1.37638021e+00
2.19948843e-01 -1.05703390e+00 2.81955630e-01 -1.99481046e+00
2.27259263e-01 -5.94106853e-01 -3.22858036e-01 7.51682937e-01
-5.94062865e-01 -5.57768047e-01 2.81702608e-01 3.19674700e-01
-5.95118105e-01 8.92782390e-01 1.30906224e+00 -3.79111916e-01
-3.29965115e-01 -1.18098415e-01 -1.01814723e+00 2.45459184e-01
9.43085074e-01 -5.49321175e-01 -5.95565081e-01 -7.79796422e-01
2.94512004e-01 1.86439663e-01 3.51846665e-02 -7.01482892e-01
4.41661984e-01 -5.41474462e-01 3.46520722e-01 -1.03188252e+00
1.57638207e-01 -6.33027375e-01 -1.33010119e-01 3.61536145e-01
-9.90521312e-01 3.10584903e-01 -6.70057535e-02 5.79020560e-01
-1.93646133e-01 -4.26224202e-01 2.57475525e-01 -3.24139416e-01
-7.75442600e-01 5.17052293e-01 -3.23694944e-01 3.74149442e-01
8.81952524e-01 5.49294464e-02 -5.49885094e-01 -4.32697386e-01
-3.16200644e-01 5.99790037e-01 1.43554345e-01 5.09830654e-01
8.88186455e-01 -1.69657159e+00 -8.87957096e-01 1.44679591e-01
3.29003185e-01 5.25116563e-01 1.96372584e-01 2.12889478e-01
-2.12852359e-01 5.24876416e-01 -1.51502937e-01 -1.76024467e-01
-8.24638009e-01 7.11360633e-01 2.01552007e-02 -4.20255780e-01
-5.55695593e-01 8.89254808e-01 1.75883994e-01 -2.94580400e-01
3.55371475e-01 -4.97595429e-01 -4.99964774e-01 1.91772267e-01
7.27858484e-01 9.08802003e-02 -2.41976042e-04 -7.17261493e-01
-1.58306867e-01 1.88123584e-01 -4.00325626e-01 -2.80188322e-01
1.02635634e+00 -9.41726472e-03 -9.33052152e-02 2.38436937e-01
9.43916380e-01 -1.00598641e-01 -1.20158768e+00 -4.46252465e-01
-4.35461812e-02 -2.79835761e-01 -1.44527435e-01 -9.87446845e-01
-9.40024912e-01 8.92543316e-01 -1.16004169e-01 2.79448569e-01
1.07108116e+00 3.60208750e-02 8.27248752e-01 3.21983725e-01
1.12349100e-01 -1.07656467e+00 2.43616268e-01 8.40187430e-01
1.29417276e+00 -8.34425867e-01 -2.74415731e-01 -1.50336340e-01
-9.92094338e-01 7.98711121e-01 9.84494209e-01 2.31091604e-01
2.17257664e-01 2.17507742e-02 -3.35977167e-01 -4.84138653e-02
-1.16442990e+00 -3.68256658e-01 2.87873328e-01 5.67182839e-01
4.49402958e-01 6.55448586e-02 -3.07608068e-01 8.50813329e-01
-3.18182349e-01 -3.91993582e-01 6.08471751e-01 8.62171888e-01
-6.78619504e-01 -1.12664104e+00 -1.73432514e-01 5.84597588e-01
-4.63110089e-01 -4.80347008e-01 -4.46213573e-01 5.49895108e-01
1.17570661e-01 8.68718565e-01 3.31946701e-01 -4.33138847e-01
3.33261818e-01 8.49049389e-02 -2.00201914e-01 -8.66833448e-01
-8.34267080e-01 2.15095341e-01 3.17372777e-03 -3.87224972e-01
-8.19203109e-02 -5.79438269e-01 -1.73602843e+00 -1.59315154e-01
-1.20415047e-01 4.75690693e-01 3.54359448e-01 7.33378708e-01
6.56947553e-01 7.68149853e-01 6.49859071e-01 -6.90175414e-01
-4.03761119e-01 -1.11861837e+00 -3.05762857e-01 4.94320989e-01
3.23157549e-01 -4.37699884e-01 -4.62537482e-02 1.04513139e-01] | [11.703105926513672, 8.847733497619629] |
c79e9515-0f86-44f6-9c27-ee74a52349de | 3d-deep-shape-descriptor | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Fang_3D_Deep_Shape_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Fang_3D_Deep_Shape_2015_CVPR_paper.pdf | 3D Deep Shape Descriptor | Shape descriptor is a concise yet informative representation that provides a 3D object with an identification as a member of some category. This paper developed a concise deep shape descriptor for the first time to address challenging issues from ever-growing 3D datasets in areas as diverse as engineering, medicine, and biology. Specifically, the proposed approach developed novel techniques to extract concise but geometrically informative shape descriptor, new definitions of Eigen-shape descriptor and Fisher-shape descriptor to guide the training strategy for deep neural network, and deep shape descriptor with discriminative capacity of maximizing the inter-class margin while minimizing the intra-class variance. Our approach addressed the challenges for shape analysis techniques posed by the complexity of 3D model and data representation and geometric structural variations and noise present in 3D models. The experimental results on 3D shape retrieval demonstrate that our proposed deep shape descriptor is superior to other state-of-the-art approaches on handling noise, incompleteness and 3D shape structural variations. | ['Meng Wang', 'Fan Zhu', 'Yi Fang', 'Edward Wong', 'Guoxian Dai', 'Tiantian Xu', 'Jin Xie'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['3d-shape-retrieval'] | ['computer-vision'] | [-4.50146079e-01 -3.69475365e-01 1.43582121e-01 -4.26845759e-01
-8.71246338e-01 -7.70876706e-01 4.30654734e-01 1.36957303e-01
-3.50401476e-02 5.65589964e-02 1.09449141e-01 1.61865000e-02
-6.82621121e-01 -6.67077839e-01 -3.39825809e-01 -7.55842805e-01
-6.39345199e-02 7.81533599e-01 -1.59173116e-01 -1.94075946e-02
4.50750560e-01 1.72501636e+00 -1.31437302e+00 -2.91344702e-01
1.50238693e-01 1.40459275e+00 -4.06367034e-01 1.16774164e-01
-3.40576321e-01 -2.54704446e-01 -4.89601612e-01 -6.08054176e-02
5.35879254e-01 4.27423507e-01 -2.99017936e-01 9.72917676e-02
5.39809048e-01 -2.64839917e-01 -5.40641844e-01 7.26220071e-01
1.20331609e+00 2.47702807e-01 1.36003911e+00 -9.54740703e-01
-1.21285510e+00 -2.97651261e-01 -5.42929173e-01 8.12531933e-02
2.29720414e-01 -1.58106349e-02 7.43408263e-01 -1.66651785e+00
7.28253841e-01 1.46079528e+00 8.37544620e-01 3.45774740e-01
-8.91126752e-01 -4.77990389e-01 -2.11742669e-01 -1.68758377e-01
-1.64918065e+00 -4.09282476e-01 1.24030924e+00 -4.85666603e-01
1.03516352e+00 1.52498543e-01 5.85027695e-01 7.68972814e-01
2.36475304e-01 8.97154689e-01 3.40127051e-01 5.30771120e-03
1.28752410e-01 -4.70175743e-01 1.75141633e-01 6.90373361e-01
4.54656214e-01 2.41755337e-01 -3.86678278e-02 -5.13669491e-01
1.09772301e+00 1.60802081e-01 1.80033222e-01 -1.08250809e+00
-8.13385069e-01 6.23185933e-01 4.34455156e-01 4.79345471e-01
-4.87472534e-01 -2.74087135e-02 5.23784697e-01 1.93164021e-01
4.32674617e-01 2.21417189e-01 -3.76589149e-01 7.45277181e-02
-6.88285649e-01 2.79215157e-01 6.13274932e-01 1.20058072e+00
5.79110980e-01 4.66842771e-01 -2.09095716e-01 1.01409853e+00
4.30739492e-01 1.15977025e+00 1.41662821e-01 -7.23957837e-01
4.31534983e-02 1.06242919e+00 -1.53637901e-01 -1.43518722e+00
-7.41857469e-01 -4.76110786e-01 -1.20693755e+00 9.47317258e-02
2.34196354e-02 4.49475288e-01 -9.55752134e-01 1.31108713e+00
5.43565452e-01 -2.99597621e-01 -2.06878617e-01 9.94699359e-01
1.50588953e+00 2.96853006e-01 -3.55985194e-01 2.81008691e-01
9.99461591e-01 -2.03878850e-01 -2.92001873e-01 3.83320987e-01
3.03690404e-01 -1.05851555e+00 7.33545840e-01 -1.53372556e-01
-1.15767980e+00 -6.91721201e-01 -9.14961696e-01 -2.73544759e-01
-5.56401908e-01 4.18845743e-01 4.53988373e-01 5.42495310e-01
-9.30966556e-01 5.10681152e-01 -5.42578161e-01 -2.29782909e-01
7.46040404e-01 5.25438905e-01 -7.07133889e-01 6.11279197e-02
-5.63336492e-01 7.30122387e-01 5.07504232e-02 1.36606202e-01
-7.57290065e-01 -7.24188030e-01 -9.09158230e-01 1.19940698e-01
-3.54996659e-02 -5.62045455e-01 6.94128931e-01 -8.01329240e-02
-1.23112333e+00 1.33540428e+00 1.12830415e-01 3.98994595e-01
1.42476723e-01 1.34653717e-01 -2.21497208e-01 -1.81932654e-02
-1.98124319e-01 2.90543795e-01 8.21593165e-01 -1.15579855e+00
2.64633894e-01 -9.67527568e-01 -4.81851250e-01 1.41609445e-01
2.74922084e-02 -2.24584818e-01 -3.27035934e-01 -5.39932072e-01
6.89009964e-01 -7.70996749e-01 1.94799583e-02 6.20496929e-01
-7.52257034e-02 -4.88408267e-01 1.04515660e+00 -3.31064492e-01
7.87877023e-01 -2.37115049e+00 3.53611843e-03 6.02738917e-01
2.35138476e-01 7.54790246e-01 -6.69098973e-01 4.57235307e-01
9.92530677e-03 1.52341217e-01 -8.57910737e-02 -2.54139811e-01
3.17328274e-01 1.42228976e-01 -8.28878656e-02 7.14132249e-01
4.00770456e-01 1.21055984e+00 -5.62355101e-01 -3.03853035e-01
2.97182173e-01 1.01968217e+00 -2.35513031e-01 2.22663894e-01
3.70712310e-01 6.86029568e-02 -8.98711205e-01 1.26385629e+00
1.26143479e+00 -5.95079586e-02 -4.42561120e-01 -5.24543464e-01
8.58314559e-02 -2.71861643e-01 -1.30187500e+00 1.60632670e+00
-1.67346835e-01 3.18398535e-01 1.85932159e-01 -1.06782031e+00
1.90243220e+00 1.17116809e-01 8.22193503e-01 -6.21587813e-01
3.43236536e-01 5.90484619e-01 -3.72003704e-01 -3.12674522e-01
3.46375138e-01 1.54945150e-01 -9.05460343e-02 5.34600437e-01
4.46405075e-02 -5.59611976e-01 -3.97817552e-01 -4.06970471e-01
6.85481668e-01 6.07514232e-02 4.41886276e-01 -4.24956292e-01
8.64226222e-01 -6.34095728e-01 3.81592512e-01 5.84827721e-01
-5.61429858e-01 8.42997611e-01 9.79241207e-02 -1.09482753e+00
-1.29021680e+00 -1.24737644e+00 -2.85299480e-01 5.40325820e-01
9.16385129e-02 1.11528143e-01 -1.14503980e-01 -4.85388249e-01
6.08203232e-01 5.69497533e-02 -6.08337462e-01 -3.92491698e-01
-6.42572284e-01 -1.62280560e-01 7.03631818e-01 6.11184597e-01
2.89680481e-01 -8.84693325e-01 -2.82570571e-01 1.17654968e-02
5.56461453e-01 -7.97151327e-01 -8.33454251e-01 -5.91925606e-02
-9.63843584e-01 -1.24797034e+00 -9.73957181e-01 -1.10188603e+00
7.92559743e-01 3.52089703e-01 9.78101432e-01 3.03547401e-02
-6.48766160e-01 7.83938468e-01 -1.38333827e-01 -4.52843606e-01
-2.15436384e-01 -4.62408625e-02 -2.87577398e-02 -1.81926548e-01
7.25330293e-01 -6.29327178e-01 -7.54073799e-01 3.80319923e-01
-9.91886437e-01 -8.56417775e-01 6.88737214e-01 8.85560513e-01
1.08075023e+00 -4.70713347e-01 4.42851067e-01 -1.12905286e-01
8.22887301e-01 -6.04157746e-02 -5.28699040e-01 3.61684829e-01
-3.89105648e-01 2.51635909e-01 3.34908277e-01 -3.46233964e-01
-5.14752507e-01 2.24883601e-01 -2.82004535e-01 -6.59360111e-01
-3.83903116e-01 2.82762557e-01 -2.07603335e-01 -5.45760870e-01
5.92692971e-01 6.79690182e-01 3.88128251e-01 -7.69680798e-01
2.21881792e-01 7.78910041e-01 3.03444356e-01 -7.99527049e-01
8.45266581e-01 4.91220087e-01 7.32994795e-01 -1.11057866e+00
-4.04670686e-01 -8.24465752e-01 -1.12886190e+00 -3.85071672e-02
3.54620934e-01 -6.94533229e-01 -8.97368670e-01 5.19469440e-01
-1.32191515e+00 5.78010261e-01 -3.24237883e-01 1.39820233e-01
-6.54891491e-01 7.31849909e-01 -2.31221542e-01 -6.89225197e-01
-1.08129156e+00 -1.17324769e+00 1.65600777e+00 -4.85822037e-02
9.25347432e-02 -9.62534547e-01 1.30774260e-01 -4.06297222e-02
7.51578271e-01 4.52547759e-01 1.07902288e+00 -9.22548413e-01
-3.61627400e-01 -7.04341590e-01 -6.18808270e-01 2.87073374e-01
3.35338384e-01 -2.96441559e-02 -7.28854775e-01 -4.41321373e-01
-1.14614420e-01 -2.89829195e-01 4.69278634e-01 4.51630592e-01
1.21846116e+00 1.46518603e-01 -1.50718661e-02 8.72344434e-01
1.33969617e+00 2.09644556e-01 3.02940756e-01 -5.70936352e-02
5.73617339e-01 1.52330771e-01 1.94137037e-01 6.08850896e-01
1.57819092e-02 5.93129158e-01 2.48695016e-01 8.03180563e-04
-1.61602095e-01 -6.37550950e-02 -2.53360480e-01 9.29100692e-01
2.71981526e-02 1.51701331e-01 -7.92673230e-01 3.83923590e-01
-1.52040875e+00 -6.69442654e-01 3.48158836e-01 2.00844908e+00
4.11387891e-01 -3.10305268e-01 -1.70277655e-02 -5.59072383e-03
5.08459926e-01 1.67065322e-01 -8.56558681e-01 -4.64524090e-01
-4.20048267e-01 2.46846780e-01 2.25153521e-01 1.89734668e-01
-1.24179971e+00 6.85947299e-01 6.48920298e+00 9.37776446e-01
-1.17488992e+00 -5.20224810e-01 3.54230672e-01 2.41683215e-01
-4.48295712e-01 -7.24280417e-01 -7.82383919e-01 -8.65869522e-02
9.54138190e-02 -1.72278538e-01 3.58786806e-02 9.88890648e-01
-7.13352412e-02 5.93973637e-01 -1.10941780e+00 1.48460352e+00
2.89593279e-01 -1.22898245e+00 5.17976820e-01 1.20228797e-01
6.30628109e-01 6.71237707e-02 2.58898318e-01 8.79729465e-02
-2.93194205e-01 -1.03270662e+00 4.31123257e-01 7.99067080e-01
8.89624834e-01 -7.79238045e-01 7.98354685e-01 6.32876009e-02
-1.48244536e+00 2.40479082e-01 -8.67663980e-01 4.99725252e-01
-1.72907054e-01 4.49477434e-01 -6.47047460e-01 5.27325690e-01
1.02209881e-01 7.68546760e-01 -3.56805533e-01 1.35170972e+00
3.54611278e-01 -2.85666108e-01 -6.08998001e-01 -3.22892189e-01
3.98051769e-01 -2.44259149e-01 8.90013397e-01 1.18852866e+00
8.08336973e-01 8.31778646e-02 2.47477502e-01 1.08951092e+00
1.20975554e-01 3.96110624e-01 -1.20156574e+00 -1.47330478e-01
6.33308470e-01 1.07972133e+00 -5.42232454e-01 1.09764136e-01
-1.52568266e-01 4.13022459e-01 1.03302471e-01 3.38430226e-01
-1.19460806e-01 -3.20268065e-01 6.95564270e-01 1.05031125e-01
5.15070796e-01 -7.15256751e-01 -4.91450816e-01 -7.65337288e-01
1.11905441e-01 -5.47948182e-01 3.98166329e-01 -6.65805399e-01
-1.61899209e+00 4.73306566e-01 -8.27166066e-02 -1.27196753e+00
1.13823898e-01 -1.13591528e+00 -4.11399454e-01 8.29036891e-01
-1.51912570e+00 -1.53845751e+00 -1.92122757e-01 5.32842636e-01
2.55594164e-01 -8.27140868e-01 1.04212344e+00 3.48316580e-01
-1.50014222e-01 8.22369576e-01 4.23114896e-01 3.12333673e-01
4.14731055e-01 -8.23889256e-01 5.37024260e-01 1.68027207e-01
1.53243374e-02 6.86626315e-01 4.02371995e-02 -4.47916180e-01
-1.98363376e+00 -7.05767453e-01 7.10903227e-01 -2.58318543e-01
3.01335156e-01 -4.58338037e-02 -7.97345638e-01 -8.43839534e-03
-5.69411278e-01 3.05356979e-01 7.49232411e-01 -1.58613428e-01
-5.63821912e-01 -1.47773147e-01 -1.37175179e+00 2.00011581e-01
1.14173019e+00 -6.47031248e-01 -6.19338632e-01 -3.65906465e-03
2.79952765e-01 -4.27480102e-01 -1.13421535e+00 5.81494272e-01
7.97478139e-01 -6.64781451e-01 1.50601006e+00 -6.85527861e-01
-1.51689425e-01 -1.57505974e-01 -4.99093294e-01 -1.05456948e+00
-6.38688684e-01 -3.88671398e-01 -2.06412539e-01 7.89809406e-01
-1.75025836e-02 -2.55652308e-01 9.50128317e-01 6.54943585e-01
-3.72749180e-01 -9.90354717e-01 -1.23484802e+00 -9.46656525e-01
4.14560854e-01 -2.15749413e-01 9.72452700e-01 5.59099913e-01
-7.97066510e-01 -1.56068578e-01 8.78209397e-02 -8.15796182e-02
8.36101830e-01 4.74309713e-01 7.69516170e-01 -1.71931863e+00
5.43895364e-01 -8.54567528e-01 -1.06400239e+00 -1.02820492e+00
3.11625808e-01 -1.13335121e+00 -3.47720593e-01 -1.29710138e+00
1.88895360e-01 -3.68617892e-01 -3.45223457e-01 4.03927803e-01
4.28467304e-01 1.12756938e-01 6.76242113e-02 7.32552111e-02
-3.93434197e-01 9.38619971e-01 1.70913434e+00 -5.11253357e-01
-1.16325870e-01 8.19869787e-02 -4.45811003e-01 4.84503090e-01
3.44253153e-01 -1.21035114e-01 -2.78627068e-01 -5.35241961e-01
1.58041790e-01 -1.48212299e-01 4.01845247e-01 -6.35168314e-01
2.01820329e-01 6.01046346e-02 6.98883533e-01 -1.28979075e+00
4.46593046e-01 -1.24546516e+00 1.10666126e-01 2.64350623e-01
-9.09542888e-02 -5.87184355e-02 3.81704986e-01 3.30993503e-01
-2.78740078e-01 -1.80343032e-01 9.00775313e-01 -1.12208962e-01
-6.15717292e-01 9.03360188e-01 1.03746943e-01 7.07141161e-02
7.36279666e-01 -6.83155358e-01 4.80053611e-02 -2.40508437e-01
-8.13949287e-01 -1.51641443e-01 2.99304634e-01 4.90896165e-01
1.27502584e+00 -2.04622722e+00 -7.23271966e-01 8.79139543e-01
2.21604675e-01 7.90045261e-02 2.16070980e-01 2.95420587e-01
-5.66022694e-01 6.23385727e-01 -4.82110083e-01 -7.75928855e-01
-1.18065810e+00 3.91641140e-01 4.79294151e-01 -6.99985102e-02
-5.35103619e-01 7.08638370e-01 7.38383690e-03 -1.04577053e+00
4.90586549e-01 -3.25054258e-01 -6.08513132e-02 9.64695662e-02
3.09892088e-01 5.29772460e-01 3.49352717e-01 -1.01131523e+00
-6.49704039e-01 1.49264491e+00 2.28502572e-01 6.08891249e-01
1.83291972e+00 -2.61753704e-02 -9.11596268e-02 4.07497510e-02
1.85743225e+00 -4.25035775e-01 -9.82784688e-01 -5.21297812e-01
-9.47071165e-02 -4.91116762e-01 2.96574961e-02 -5.75754881e-01
-1.19300270e+00 1.07159412e+00 8.40736806e-01 -8.35344791e-02
7.90879071e-01 1.14745582e-02 9.24435258e-01 8.72379601e-01
2.86685705e-01 -1.08083808e+00 1.92049325e-01 1.02809381e+00
1.55303764e+00 -1.20934987e+00 2.89524734e-01 5.64075261e-03
-1.22460611e-01 1.47454941e+00 2.58154869e-01 -3.53342205e-01
1.29069173e+00 1.21234551e-01 5.71677610e-02 -6.85362756e-01
-1.80246428e-01 -1.44720539e-01 1.16752100e+00 9.40697134e-01
3.93612355e-01 -8.68195966e-02 2.87321676e-02 5.44934273e-01
2.09166110e-01 -2.72834420e-01 -2.53974408e-01 6.71448529e-01
-4.92794394e-01 -9.95915294e-01 -4.05701399e-01 2.40766197e-01
-8.44095573e-02 3.62929344e-01 -7.71887839e-01 7.63916194e-01
-3.33682656e-01 8.64266679e-02 -9.69229266e-04 -1.59651414e-01
6.91372037e-01 1.23471230e-01 5.89161575e-01 -2.35970140e-01
-1.90353066e-01 3.06741595e-01 -3.42837930e-01 -5.44732928e-01
-2.68456131e-01 -5.27553976e-01 -1.06028068e+00 -9.47103500e-02
-9.99154150e-02 -3.11646223e-01 8.46606672e-01 6.14813507e-01
8.75346959e-01 -2.00510591e-01 7.66619802e-01 -1.24153364e+00
-1.04967058e+00 -6.65238678e-01 -7.90992975e-01 8.26334059e-01
3.62716883e-01 -1.02146268e+00 -2.66070694e-01 -5.57627559e-01] | [8.126739501953125, -3.859771966934204] |
ef360c7d-85c5-4991-8cb6-3a896dab5fe0 | clera-a-unified-model-for-joint-cognitive | 2306.15073 | null | https://arxiv.org/abs/2306.15073v1 | https://arxiv.org/pdf/2306.15073v1.pdf | CLERA: A Unified Model for Joint Cognitive Load and Eye Region Analysis in the Wild | Non-intrusive, real-time analysis of the dynamics of the eye region allows us to monitor humans' visual attention allocation and estimate their mental state during the performance of real-world tasks, which can potentially benefit a wide range of human-computer interaction (HCI) applications. While commercial eye-tracking devices have been frequently employed, the difficulty of customizing these devices places unnecessary constraints on the exploration of more efficient, end-to-end models of eye dynamics. In this work, we propose CLERA, a unified model for Cognitive Load and Eye Region Analysis, which achieves precise keypoint detection and spatiotemporal tracking in a joint-learning framework. Our method demonstrates significant efficiency and outperforms prior work on tasks including cognitive load estimation, eye landmark detection, and blink estimation. We also introduce a large-scale dataset of 30k human faces with joint pupil, eye-openness, and landmark annotation, which aims to support future HCI research on human factors and eye-related analysis. | ['Bryan Reimer', 'Bruce Mehler', 'Lex Fridman', 'Meng Wang', 'Aishni Parab', 'Jack Terwilliger', 'Li Ding'] | 2023-06-26 | null | null | null | null | ['keypoint-detection', 'blink-estimation'] | ['computer-vision', 'computer-vision'] | [ 3.53211649e-02 -2.84105271e-01 -1.04192793e-01 -3.28127056e-01
-2.19606027e-01 -4.75904524e-01 2.61965245e-01 -5.85268345e-03
-4.82120693e-01 2.55390972e-01 1.08609855e-01 -2.80529529e-01
-1.01802617e-01 1.48829147e-01 -1.04872636e-01 -2.57756084e-01
1.36722088e-01 -1.45027533e-01 1.28881723e-01 3.81052077e-01
6.11864746e-01 6.14421546e-01 -2.19172406e+00 1.38643896e-02
5.80338120e-01 1.15516031e+00 2.72102300e-02 1.02885962e+00
5.58443487e-01 6.02520704e-01 -5.50848782e-01 -1.82983980e-01
9.50057618e-03 -2.37677693e-01 -3.37479323e-01 1.81691796e-01
1.06762230e+00 -3.04972112e-01 -1.13703437e-01 7.79421151e-01
1.02910864e+00 3.20197225e-01 3.41862381e-01 -1.63580632e+00
-4.54558492e-01 -6.28439009e-01 -8.99138451e-01 8.32024157e-01
7.54062533e-01 4.56726074e-01 5.60084283e-01 -7.41128385e-01
2.25056499e-01 1.26690114e+00 5.08468509e-01 4.74715114e-01
-1.23869789e+00 -1.06202281e+00 2.12121978e-01 3.11006904e-01
-1.51229453e+00 -1.23740852e+00 5.37780583e-01 -5.11680424e-01
1.00682735e+00 3.00719678e-01 6.23081148e-01 8.98499489e-01
1.83722869e-01 7.94792354e-01 1.20665538e+00 -5.52663088e-01
-6.06506914e-02 3.64719003e-01 1.73227116e-01 7.36697435e-01
1.96489677e-01 1.00852057e-01 -1.07800293e+00 -2.45339945e-01
7.16048062e-01 -5.19798743e-03 -2.90395528e-01 -3.32454413e-01
-8.75673175e-01 1.70428187e-01 1.62492558e-01 -1.60813227e-01
-2.87417799e-01 -1.42619789e-01 2.40312457e-01 2.05392003e-01
8.10141921e-01 3.25648695e-01 -3.49663079e-01 -4.97981608e-01
-6.90178454e-01 3.63105923e-01 4.23241884e-01 1.08575666e+00
4.38963354e-01 -4.05949444e-01 -5.73382199e-01 5.59909523e-01
5.38478255e-01 5.50959408e-01 5.97314946e-02 -1.01291013e+00
1.29480898e-01 7.39511430e-01 6.18748903e-01 -7.02723801e-01
-7.56775677e-01 4.53633592e-02 1.35794086e-02 5.69486380e-01
6.45507038e-01 -2.73373127e-01 -6.17528796e-01 1.68523228e+00
5.44177055e-01 2.26837710e-01 -7.90160358e-01 1.01003599e+00
3.33453864e-01 -1.69925898e-01 3.95143002e-01 -5.33128440e-01
1.69392812e+00 -4.44407940e-01 -9.59766209e-01 -3.88976365e-01
5.22694468e-01 -8.19609225e-01 1.33838940e+00 5.60207188e-01
-1.25441301e+00 -7.30803430e-01 -6.63263738e-01 -3.34168524e-01
-1.75789937e-01 2.92769790e-01 7.30479360e-01 1.04803371e+00
-1.12853801e+00 -3.67260613e-02 -8.54336202e-01 -5.62530160e-01
6.05418265e-01 3.88208956e-01 -1.14556305e-01 2.97748655e-01
-4.14950162e-01 1.01189470e+00 -3.23158115e-01 8.58057737e-02
-2.85938889e-01 -8.13889027e-01 -8.31363857e-01 -9.05477181e-02
5.54856300e-01 -4.35048878e-01 1.45714545e+00 -8.23557317e-01
-1.40899241e+00 1.25262511e+00 -9.83790636e-01 8.62401351e-02
9.05010402e-02 -4.98289019e-01 -6.74982846e-01 -1.76910743e-01
-1.60084754e-01 5.93927681e-01 1.23730898e+00 -6.08000755e-01
-8.80605459e-01 -7.94302106e-01 -2.30482042e-01 3.76926512e-01
-4.48493421e-01 7.53829122e-01 -5.24106443e-01 -1.26996726e-01
-3.96192729e-01 -8.18707526e-01 3.20494890e-01 4.00069535e-01
-1.45200506e-01 -5.22100747e-01 8.00366282e-01 -5.25681257e-01
1.70420361e+00 -2.19382143e+00 -7.67240599e-02 8.83202776e-02
7.50386536e-01 3.68457437e-01 2.87125781e-02 -3.95664185e-01
1.55359684e-02 -1.87015072e-01 7.67810702e-01 -7.51899004e-01
4.18477207e-02 -6.81134284e-01 -8.30922201e-02 8.17385137e-01
-8.19353163e-02 9.71602023e-01 -6.61629438e-01 -3.88419598e-01
2.66520441e-01 5.52204788e-01 -4.22524720e-01 2.08286285e-01
1.14677332e-01 2.80623645e-01 -1.14255272e-01 8.88080180e-01
2.82882452e-01 -2.27461994e-01 -2.87177831e-01 -2.34002739e-01
-3.79318476e-01 3.55118401e-02 -9.70419765e-01 1.37711656e+00
-3.30661267e-01 1.41007054e+00 5.77422380e-02 9.78788286e-02
5.49574852e-01 4.69258279e-02 1.05668113e-01 -1.12553573e+00
5.08225322e-01 -4.45528895e-01 7.89178014e-02 -6.87101781e-01
4.00170147e-01 5.35537899e-01 3.20519358e-01 7.48439431e-01
-3.98700744e-01 5.45959055e-01 5.81045933e-02 -1.87273160e-01
8.63665640e-01 1.97069034e-01 4.54748243e-01 -1.46602005e-01
3.18422288e-01 -5.54708481e-01 1.06147803e-01 2.74498969e-01
-9.61661518e-01 2.61655837e-01 3.47387761e-01 -5.44744253e-01
-6.64342344e-01 -8.99074376e-01 -7.05768913e-02 1.82915401e+00
-1.43461525e-02 -3.95697773e-01 -1.06199217e+00 -3.02623361e-01
6.60153329e-02 5.40511608e-01 -7.01884627e-01 -2.29176417e-01
-4.64277305e-02 -4.15563762e-01 3.06767017e-01 6.22077107e-01
1.60976142e-01 -1.10078418e+00 -1.30180132e+00 -3.40702713e-01
-1.19208939e-01 -1.22417891e+00 -9.48631585e-01 -7.06969202e-01
-4.00208831e-01 -1.35170352e+00 -3.59940499e-01 -3.23264599e-01
7.32299864e-01 7.06489801e-01 9.79683399e-01 -1.14459433e-01
-1.00955057e+00 9.44387555e-01 1.87201247e-01 -9.39345539e-01
3.14837158e-01 -2.22283646e-01 4.59206730e-01 2.02837631e-01
1.32087564e+00 -2.82791127e-02 -8.73891890e-01 6.22967958e-01
-7.74637535e-02 -2.91462898e-01 2.66465634e-01 4.47435565e-02
3.36645186e-01 -1.28861219e-01 1.48911327e-01 -2.68822342e-01
1.08308971e+00 -4.04069163e-02 -9.40075040e-01 3.13257664e-01
-7.70352602e-01 -3.71586353e-01 -3.04777175e-01 -8.24061394e-01
-1.14543629e+00 -1.86285958e-01 6.99026644e-01 -5.18841267e-01
-3.78893137e-01 -2.17956021e-01 -7.81120881e-02 -3.47408801e-01
1.10127556e+00 -3.01732779e-01 2.73993790e-01 -2.19567910e-01
6.95499331e-02 1.09608614e+00 5.24753511e-01 -2.38018975e-01
1.14774436e-01 4.23418283e-01 -5.13497647e-03 -1.04532814e+00
-9.83914018e-01 -8.35160613e-01 -6.60877705e-01 -4.32502836e-01
6.62029386e-01 -1.29178870e+00 -1.70874429e+00 4.94643092e-01
-9.06874657e-01 -4.62029576e-01 2.85659045e-01 3.74536157e-01
-5.46403050e-01 1.16627790e-01 -1.26783520e-01 -1.43434024e+00
-3.04681927e-01 -8.81232679e-01 1.29754758e+00 4.94684905e-01
-8.20567191e-01 -7.16239154e-01 -9.36799869e-02 4.26985919e-01
3.21348816e-01 -8.98580160e-03 1.86942965e-01 1.40462331e-02
-4.78770792e-01 -3.30333054e-01 -4.61497366e-01 -5.11468723e-02
9.89994779e-02 -1.80043541e-02 -1.36360371e+00 -5.63644469e-01
-3.07821453e-01 -2.53397673e-01 1.93281978e-01 8.83789241e-01
1.19984078e+00 2.17309102e-01 -5.05474925e-01 5.45330703e-01
7.17744291e-01 -6.63828179e-02 6.49324954e-01 9.92282107e-02
6.04063392e-01 6.92003012e-01 7.19055474e-01 6.78382933e-01
5.42012155e-01 7.69351661e-01 3.55503112e-02 -6.27650917e-02
-1.06814124e-01 1.49958268e-01 3.24310064e-01 -4.33730692e-01
-3.62094462e-01 -2.25982085e-01 -9.09534812e-01 3.19883436e-01
-1.62685072e+00 -9.21311438e-01 -3.40174101e-02 2.56909108e+00
6.10440135e-01 6.29985183e-02 6.77471578e-01 8.48486181e-03
6.03124440e-01 -2.66775489e-01 -8.69095206e-01 -1.96481049e-01
4.25998330e-01 9.84769911e-02 5.04279494e-01 4.10962582e-01
-1.01495135e+00 7.99447954e-01 7.30412292e+00 1.77223459e-01
-9.45811093e-01 1.10751033e-01 4.52122897e-01 -8.80606115e-01
4.62456167e-01 -5.30771017e-01 -1.19610012e+00 5.32756686e-01
1.09420586e+00 -5.91125190e-02 7.21525013e-01 5.91341436e-01
5.22344589e-01 -6.74483180e-01 -1.37658429e+00 1.49116039e+00
4.13630784e-01 -7.44329751e-01 -8.61837804e-01 3.91320348e-01
8.66107568e-02 -1.65261850e-01 4.45122927e-01 5.60195409e-02
-3.17137003e-01 -1.01489961e+00 3.83701712e-01 8.56183290e-01
1.28014028e+00 -7.13273764e-01 -3.19841281e-02 1.50959596e-01
-1.04484344e+00 -4.38868046e-01 -1.37192085e-01 -4.55890357e-01
-1.23807840e-01 -8.45479667e-02 -5.04624665e-01 -6.17188156e-01
9.17619169e-01 4.28879946e-01 -8.43192518e-01 1.15131152e+00
1.22675061e-01 2.05161929e-01 -1.60154328e-01 -1.91449728e-02
-6.66469634e-01 3.74197900e-01 4.31772470e-01 9.50380087e-01
-2.78649181e-02 1.92853212e-01 -3.93329859e-02 9.02718067e-01
1.27924994e-01 -1.25673041e-01 -4.84975874e-01 -1.25700068e-02
7.16473997e-01 1.36986899e+00 -4.81845766e-01 1.04467452e-01
-6.22344375e-01 8.58946681e-01 3.08336765e-01 6.37208700e-01
-6.97034895e-01 -5.80073059e-01 1.25531173e+00 5.76841414e-01
-1.66559860e-01 -3.90723884e-01 -3.45550776e-01 -7.20026016e-01
1.04141049e-01 -7.44643629e-01 2.72751659e-01 -9.98298824e-01
-8.38367522e-01 2.09652141e-01 1.03847086e-01 -9.79940832e-01
-2.20067382e-01 -8.46204281e-01 -4.53904688e-01 1.26782703e+00
-1.38113713e+00 -1.11913002e+00 -8.74792516e-01 9.12892520e-01
4.54296201e-01 -3.28096718e-01 6.42078340e-01 1.76477671e-01
-9.22560096e-01 1.00030720e+00 -6.43192887e-01 -1.61258712e-01
1.35876095e+00 -8.74542534e-01 5.27563632e-01 7.87869334e-01
-2.23903924e-01 7.84207284e-01 6.97598875e-01 -5.63401520e-01
-1.31860185e+00 -6.22884512e-01 8.21322203e-01 -1.34493387e+00
3.75947058e-01 -5.60209095e-01 -6.58440948e-01 7.51549602e-01
-3.45047824e-02 2.05831900e-01 9.05958414e-01 6.95244551e-01
-2.84345657e-01 -1.72377735e-01 -9.70937848e-01 8.12739968e-01
1.18732011e+00 -8.47660661e-01 -2.51635164e-01 2.36708447e-02
2.33809501e-01 -4.08708394e-01 -5.80983222e-01 -8.68001804e-02
1.07866848e+00 -9.62950647e-01 1.01779950e+00 -5.81728637e-01
-4.83670950e-01 -1.45602956e-01 3.91334176e-01 -7.59674788e-01
-6.43375099e-01 -1.13803244e+00 -8.41303527e-01 8.72133970e-01
-1.11198582e-01 -5.60371935e-01 5.56040704e-01 1.37014580e+00
4.22233105e-01 -3.28421444e-01 -4.30103034e-01 -3.80661964e-01
-8.23221087e-01 -4.50371891e-01 4.80757982e-01 4.98317599e-01
2.05444664e-01 2.20606565e-01 -3.19896072e-01 1.95510000e-01
8.44698489e-01 -2.82049239e-01 1.11164951e+00 -1.72974050e+00
2.32914299e-01 -5.44551492e-01 -5.73149145e-01 -7.96387970e-01
2.74829686e-01 7.19805583e-02 -1.94929525e-01 -8.00013065e-01
1.44702792e-01 3.72424424e-02 -3.90136808e-01 3.43359143e-01
-4.28029567e-01 4.08078343e-01 -1.99555680e-01 7.88707659e-02
-8.74913394e-01 -3.15835029e-02 9.57016170e-01 5.80616355e-01
-6.61806583e-01 2.09750190e-01 -9.13522840e-01 5.55141091e-01
5.44811547e-01 1.23392724e-01 -6.13691986e-01 -2.19571814e-01
4.69057590e-01 -3.77838224e-01 6.33509636e-01 -9.32169139e-01
6.93863511e-01 -2.60746419e-01 8.39700878e-01 -4.01228905e-01
3.31008822e-01 -5.89143693e-01 -4.69608724e-01 -9.67878848e-03
-2.08273768e-01 5.80169559e-02 7.00949848e-01 6.31051898e-01
3.78245294e-01 3.59829903e-01 8.01870644e-01 1.59043476e-01
-7.05100179e-01 4.31656659e-01 -1.87738970e-01 5.36320470e-02
1.24909031e+00 -6.25788987e-01 -3.47523928e-01 -2.61519074e-01
-5.50682366e-01 2.81127781e-01 6.38997197e-01 6.69265151e-01
4.58438516e-01 -1.02242160e+00 -2.66891539e-01 6.36549890e-01
4.96509701e-01 -4.25953448e-01 2.79113203e-01 1.13819838e+00
1.87020153e-01 4.59547788e-01 -4.71061736e-01 -7.24781871e-01
-1.82872653e+00 7.34269798e-01 2.74826020e-01 6.97330356e-01
-3.44567746e-01 7.37230062e-01 9.56905335e-02 5.41375279e-01
7.31109917e-01 -3.79873626e-02 -2.94148445e-01 3.10165226e-01
1.29263842e+00 9.09847677e-01 -1.37444571e-01 -6.43005610e-01
-3.15076888e-01 4.73016143e-01 -1.12613328e-01 8.52106661e-02
6.35416150e-01 -9.20363009e-01 3.99592146e-02 4.24981833e-01
6.91677511e-01 3.73155549e-02 -1.48771715e+00 -3.76034141e-01
-1.35906860e-01 -8.32960129e-01 4.70384389e-01 -6.30809665e-01
-3.46931428e-01 9.37223017e-01 1.11702704e+00 -5.04759587e-02
1.39571559e+00 -7.42248222e-02 4.04161155e-01 3.19278330e-01
1.58026204e-01 -1.36667335e+00 2.04135656e-01 9.25048068e-03
4.70876127e-01 -1.61829913e+00 1.36635631e-01 -1.92831054e-01
-6.02393150e-01 7.34694064e-01 9.99388039e-01 4.89104390e-01
7.35710502e-01 1.94373623e-01 4.93083324e-04 -4.61665541e-01
-8.42540026e-01 -6.84323728e-01 8.53050292e-01 8.55476916e-01
4.93660808e-01 -1.49426118e-01 4.39668238e-01 3.09388727e-01
1.06450111e-01 2.38609686e-01 4.07198742e-02 8.73555660e-01
-5.41603446e-01 -4.88426507e-01 -4.58748639e-01 4.05430406e-01
-4.70530957e-01 1.47122890e-01 -3.34480286e-01 5.53655744e-01
-1.63091421e-02 1.12686825e+00 5.89587331e-01 -5.66317439e-02
5.22180319e-01 3.74747753e-01 8.63293886e-01 -6.11232340e-01
-1.78989410e-01 9.01321471e-02 -1.99404821e-01 -1.11352277e+00
-4.42225069e-01 -8.30828190e-01 -5.82753718e-01 -3.82149011e-01
-2.39725113e-01 -7.34374940e-01 6.15036726e-01 7.35494137e-01
8.74233365e-01 3.38898987e-01 2.01358542e-01 -1.02864254e+00
-6.18874766e-02 -1.10596490e+00 -5.03503799e-01 3.58132303e-01
7.05147028e-01 -1.07924938e+00 -1.95266798e-01 1.03592336e-01] | [14.103177070617676, 0.1168123185634613] |
5c9e0d1b-ce8e-4c69-b91a-046c15136df3 | grade-automatic-graph-enhanced-coherence | 2010.03994 | null | https://arxiv.org/abs/2010.03994v1 | https://arxiv.org/pdf/2010.03994v1.pdf | GRADE: Automatic Graph-Enhanced Coherence Metric for Evaluating Open-Domain Dialogue Systems | Automatically evaluating dialogue coherence is a challenging but high-demand ability for developing high-quality open-domain dialogue systems. However, current evaluation metrics consider only surface features or utterance-level semantics, without explicitly considering the fine-grained topic transition dynamics of dialogue flows. Here, we first consider that the graph structure constituted with topics in a dialogue can accurately depict the underlying communication logic, which is a more natural way to produce persuasive metrics. Capitalized on the topic-level dialogue graph, we propose a new evaluation metric GRADE, which stands for Graph-enhanced Representations for Automatic Dialogue Evaluation. Specifically, GRADE incorporates both coarse-grained utterance-level contextualized representations and fine-grained topic-level graph representations to evaluate dialogue coherence. The graph representations are obtained by reasoning over topic-level dialogue graphs enhanced with the evidence from a commonsense graph, including k-hop neighboring representations and hop-attention weights. Experimental results show that our GRADE significantly outperforms other state-of-the-art metrics on measuring diverse dialogue models in terms of the Pearson and Spearman correlations with human judgements. Besides, we release a new large-scale human evaluation benchmark to facilitate future research on automatic metrics. | ['Xiaodan Liang', 'Liang Lin', 'Jinghui Qin', 'Zheng Ye', 'Lishan Huang'] | 2020-10-08 | null | https://aclanthology.org/2020.emnlp-main.742 | https://aclanthology.org/2020.emnlp-main.742.pdf | emnlp-2020-11 | ['dialogue-evaluation'] | ['natural-language-processing'] | [ 6.70525953e-02 7.36546516e-01 -1.41409621e-01 -5.60653448e-01
-6.22011781e-01 -5.05855441e-01 1.11079013e+00 6.70664966e-01
-7.12637082e-02 8.39335859e-01 1.11749065e+00 -9.65863243e-02
-1.28510281e-01 -9.28723156e-01 9.23583657e-02 -1.57377318e-01
-2.03431189e-01 5.20637691e-01 2.11039320e-01 -9.76304412e-01
5.52717745e-01 -2.95988679e-01 -1.04872346e+00 3.75614554e-01
1.36146438e+00 9.29812551e-01 -1.64103076e-01 9.16361511e-01
-4.93599623e-01 1.18231654e+00 -9.94500935e-01 -7.19027221e-01
-6.40609682e-01 -9.97387052e-01 -1.46143270e+00 1.52456984e-01
9.40901935e-02 -1.64984167e-01 -1.72396928e-01 1.04983389e+00
2.25327671e-01 3.75179619e-01 8.44631076e-01 -1.29082298e+00
-7.14523077e-01 1.03890729e+00 -1.78545956e-02 1.29896984e-01
1.16559947e+00 2.36818656e-01 1.76682520e+00 -5.13537824e-01
6.87417984e-01 1.77053452e+00 4.58430409e-01 3.51027757e-01
-1.03243411e+00 -4.84944209e-02 3.05155307e-01 2.13471428e-01
-7.46231318e-01 -1.21256143e-01 9.85981584e-01 -5.04729986e-01
9.37712014e-01 3.23452055e-01 5.65048456e-01 1.06673157e+00
1.26747385e-01 7.12925792e-01 1.22694135e+00 -2.86078453e-01
-2.01461501e-02 1.95871200e-03 9.14303422e-01 9.15823221e-01
-1.16426617e-01 -5.19634724e-01 -6.23194754e-01 -4.64964658e-01
5.11018813e-01 -5.87295651e-01 -4.34797972e-01 -5.45020550e-02
-1.32707465e+00 1.30873060e+00 5.39606690e-01 4.14569020e-01
-3.73886645e-01 -4.34922785e-01 7.63301492e-01 4.26665187e-01
6.18168771e-01 9.30650413e-01 -1.79315850e-01 -5.25602162e-01
-3.85419488e-01 2.91414440e-01 1.48926318e+00 7.94848382e-01
7.76305974e-01 -1.31238490e-01 -8.69912207e-01 1.15670085e+00
3.48030657e-01 2.18845103e-02 5.39822340e-01 -1.02598131e+00
5.12385726e-01 1.24048746e+00 -1.13907263e-01 -1.39423013e+00
-5.89468241e-01 -5.39414957e-02 -8.45101058e-01 -5.67576051e-01
4.80913728e-01 -2.05367342e-01 2.96360045e-03 1.77441955e+00
3.95121396e-01 -2.85594672e-01 3.39699417e-01 9.26747620e-01
1.48332107e+00 6.83944166e-01 8.11955929e-02 -3.64319146e-01
1.72125053e+00 -1.18655467e+00 -1.24317718e+00 1.57107651e-01
7.14396536e-01 -4.44654167e-01 1.48148227e+00 1.20966904e-01
-1.00633430e+00 -4.04698849e-01 -1.10087430e+00 -2.21249118e-01
-4.71474588e-01 -4.50782031e-01 5.56883454e-01 4.94049072e-01
-8.74347329e-01 5.07774532e-01 3.36830020e-02 -3.74570400e-01
-2.01597527e-01 -3.01152229e-01 -9.00475457e-02 2.99387276e-01
-2.06178284e+00 1.14094055e+00 3.32018971e-01 -1.05360918e-01
-6.12263262e-01 -3.53629559e-01 -1.21918082e+00 1.07627027e-01
7.75353074e-01 -5.49391866e-01 1.36588717e+00 -2.51768917e-01
-2.09589434e+00 7.98680902e-01 9.82013494e-02 -4.14361000e-01
2.78356761e-01 1.80001892e-02 -3.32124591e-01 1.86271265e-01
1.20828651e-01 2.21555814e-01 4.90446031e-01 -1.00348914e+00
-4.37213272e-01 -1.59390226e-01 6.50945902e-01 6.47934318e-01
-4.48923230e-01 -4.85229269e-02 -1.49234235e-01 -3.84516895e-01
-2.54637629e-01 -5.78216791e-01 -8.14347118e-02 -4.10325170e-01
-8.34166825e-01 -1.03722906e+00 6.07937872e-01 -5.90245187e-01
1.72785199e+00 -1.67342830e+00 3.08445781e-01 -8.20209179e-03
7.64173388e-01 4.19554375e-02 -3.55692655e-02 7.50975311e-01
5.29954970e-01 2.09180221e-01 -1.82452768e-01 -1.33318342e-02
5.50763965e-01 -1.03350013e-01 -1.92442954e-01 4.48675081e-02
1.48384437e-01 8.99438560e-01 -1.30520022e+00 -7.14226425e-01
1.71136096e-01 1.32789752e-02 -4.23356414e-01 5.65917850e-01
-5.70622504e-01 3.25406343e-01 -6.53581440e-01 6.15730621e-02
7.96582028e-02 -6.44633591e-01 3.75384301e-01 -2.55530357e-01
2.94843346e-01 9.82408404e-01 -6.45760357e-01 1.72606409e+00
-4.25128341e-01 6.63019538e-01 -9.93129387e-02 -5.98367035e-01
1.29069519e+00 1.79663435e-01 2.27186978e-02 -5.69839120e-01
1.70205966e-01 -2.43572891e-01 1.73745573e-01 -2.90089071e-01
1.16440380e+00 2.28124797e-01 -8.46445382e-01 8.14153552e-01
1.99266776e-01 -6.77824378e-01 3.43357682e-01 8.98748159e-01
9.10419047e-01 -4.41851348e-01 8.26329350e-01 -4.22916144e-01
8.66855860e-01 -8.66396576e-02 1.79572076e-01 5.57239115e-01
-4.49729741e-01 9.03905332e-02 1.24628937e+00 -2.46978067e-02
-5.81235468e-01 -8.15595388e-01 3.78131457e-02 1.48619008e+00
2.96372235e-01 -1.05992305e+00 -9.73352849e-01 -8.97440076e-01
-1.04644887e-01 9.14450228e-01 -7.38620102e-01 -1.98324189e-01
-2.99800605e-01 -3.00076813e-01 7.09731698e-01 2.75653563e-02
7.57808506e-01 -1.04643178e+00 -1.48823872e-01 2.91794837e-01
-7.51102686e-01 -1.05463290e+00 -6.10318780e-01 -2.90234834e-01
-3.76424730e-01 -1.32854569e+00 -4.02797014e-01 -4.17116195e-01
2.88880263e-02 1.64919928e-01 1.47146940e+00 1.41032472e-01
1.67981595e-01 5.05282640e-01 -6.53033078e-01 4.15838249e-02
-7.73845017e-01 8.07234645e-03 -2.44684786e-01 -1.59613848e-01
2.65348762e-01 -6.17596544e-02 -4.59407747e-01 3.84694636e-01
-4.77138132e-01 9.86865684e-02 -1.15330681e-01 1.15498781e+00
-1.73833504e-01 -2.11318776e-01 8.68375897e-01 -1.10094225e+00
2.02331161e+00 -5.48632979e-01 7.68871699e-03 3.70595813e-01
-6.07715070e-01 1.29062146e-01 6.21763468e-01 -4.72755767e-02
-1.10957730e+00 -1.14083660e+00 7.07173795e-02 3.31194758e-01
1.52243778e-01 9.33542907e-01 -4.77312468e-02 3.38253200e-01
7.63787866e-01 -1.29577396e-02 7.57723153e-02 1.04325093e-01
9.55243111e-01 8.32002878e-01 4.29480672e-01 -9.27652299e-01
1.02411300e-01 -1.71164006e-01 -4.39522058e-01 -1.10097837e+00
-1.10074186e+00 -5.33392906e-01 -4.84299451e-01 -5.78492224e-01
1.00414670e+00 -5.36482990e-01 -1.00104868e+00 2.44266570e-01
-1.42368913e+00 -3.90596807e-01 -1.39535636e-01 1.38952166e-01
-6.65331423e-01 7.17480242e-01 -8.97631943e-01 -8.17971587e-01
-3.84636402e-01 -9.40070331e-01 9.78870511e-01 2.33344778e-01
-7.54829049e-01 -1.52655876e+00 2.55595952e-01 5.25604188e-01
3.81197810e-01 4.27127510e-01 1.10161483e+00 -1.10173655e+00
-3.75197902e-02 1.58607766e-01 -3.61580998e-01 2.04393372e-01
2.10262388e-01 -8.68317857e-03 -8.16436708e-01 1.84489377e-02
1.02709951e-02 -7.26709723e-01 5.63603818e-01 -3.20504531e-02
7.89910495e-01 -7.88527906e-01 4.97427993e-02 -3.29205990e-01
6.51355624e-01 -3.00864667e-01 4.02747154e-01 1.39031382e-02
5.91063380e-01 1.05893564e+00 8.52785289e-01 8.29951465e-01
1.08111835e+00 6.73960388e-01 1.02522122e-02 1.47328421e-01
-1.19794995e-01 -4.12277222e-01 2.61836261e-01 1.42095995e+00
6.11244626e-02 -4.26205009e-01 -9.45531726e-01 3.50450963e-01
-1.92705381e+00 -8.80802691e-01 -3.29890132e-01 1.73503900e+00
1.25482631e+00 2.03957230e-01 3.58917385e-01 1.17446966e-02
9.97434974e-01 6.54841959e-01 -2.66640335e-01 -7.10630119e-01
-5.70891537e-02 -2.46337891e-01 -4.42266405e-01 9.27004278e-01
-8.46405804e-01 1.09983099e+00 6.06901646e+00 8.16034317e-01
-5.57256758e-01 3.28380130e-02 4.29631412e-01 6.85910046e-01
-5.36550760e-01 -1.98735684e-01 -5.23292601e-01 3.35531324e-01
1.03989673e+00 -7.28146970e-01 1.07682019e-01 7.82512486e-01
3.51092964e-02 2.42753811e-02 -1.25620818e+00 6.30971968e-01
2.23459810e-01 -1.28718841e+00 1.75765038e-01 -3.08449622e-02
6.45809412e-01 -5.33368468e-01 -3.55140179e-01 6.52866304e-01
9.22512591e-01 -9.08226252e-01 2.37744763e-01 4.86011058e-01
5.61146140e-01 -4.97155726e-01 8.17624748e-01 2.66990483e-01
-1.18630755e+00 3.31003875e-01 -2.32459009e-01 -1.89825252e-01
8.29365104e-02 4.05762345e-01 -1.26249480e+00 6.41778708e-01
1.50043681e-01 7.09964931e-01 -4.84813303e-01 3.97918582e-01
-5.17039061e-01 5.07178962e-01 2.09040090e-01 -8.81868482e-01
4.77078706e-01 -2.82084346e-01 6.22498095e-01 1.58585203e+00
-2.81262696e-01 3.77681047e-01 5.25618792e-01 8.88782740e-01
-2.61277884e-01 3.14655632e-01 -6.79232180e-01 -3.58217120e-01
6.67663932e-01 1.42736220e+00 -4.00532722e-01 -5.38156629e-01
-2.69535154e-01 6.70013785e-01 6.36496663e-01 4.72734161e-02
-5.97964466e-01 -6.04016423e-01 5.99738419e-01 -4.64942604e-01
-4.07793730e-01 -3.30633551e-01 -1.64170608e-01 -1.05746686e+00
-4.18362767e-01 -9.59295630e-01 6.55348361e-01 -5.00686705e-01
-1.83831537e+00 8.50545883e-01 3.60361263e-02 -8.88076663e-01
-6.93234861e-01 -3.44738185e-01 -8.97513270e-01 7.74701834e-01
-1.33438826e+00 -8.66483033e-01 -6.28835559e-01 7.58341610e-01
7.93463767e-01 -1.20009802e-01 1.17319310e+00 -5.64218521e-01
-3.83789480e-01 5.57645261e-01 -4.25843149e-01 2.91903228e-01
7.29232907e-01 -1.56371677e+00 4.44390476e-01 1.82905808e-01
-1.12691276e-01 6.85716331e-01 7.95942307e-01 -6.15593791e-01
-1.07712710e+00 -7.22113550e-01 9.92525220e-01 -4.31943089e-01
1.17297399e+00 -3.55158031e-01 -1.20491540e+00 1.52310148e-01
7.90130615e-01 -6.67090118e-01 1.01256323e+00 7.45167136e-01
-4.08190906e-01 4.35676396e-01 -1.01602888e+00 6.30751014e-01
9.60813105e-01 -9.64866877e-01 -1.35430574e+00 5.37786484e-01
1.12492120e+00 -3.95720422e-01 -1.37564826e+00 1.78648472e-01
1.51237383e-01 -9.23646629e-01 5.40833890e-01 -7.68702924e-01
6.11974359e-01 1.24225423e-01 -1.28732666e-01 -1.60266984e+00
-2.88503468e-01 -9.83125329e-01 -1.89782485e-01 1.36553264e+00
4.98733938e-01 -6.11567795e-01 3.97043049e-01 6.52464449e-01
-2.51823604e-01 -6.55237079e-01 -6.06129825e-01 -3.90665710e-01
3.81787419e-02 -9.16421786e-03 6.26619875e-01 1.29393530e+00
1.11286378e+00 1.17520010e+00 -2.25935549e-01 -3.94226521e-01
3.34286064e-01 2.74111301e-01 9.15327847e-01 -1.47060418e+00
-2.27716491e-02 -8.53807092e-01 -3.57720941e-01 -1.16278493e+00
5.28081298e-01 -7.90772557e-01 1.91249207e-01 -1.78621721e+00
1.12169966e-01 -1.29841864e-01 7.44147003e-02 -7.46892989e-02
-6.55888259e-01 -3.47821176e-01 1.57747805e-01 6.26386181e-02
-1.13446498e+00 1.09485400e+00 1.67357004e+00 -2.56158054e-01
-5.82131483e-02 -2.48302490e-01 -8.72810245e-01 7.04976857e-01
7.05529213e-01 1.22899458e-01 -5.70612013e-01 1.61986217e-01
-3.27373892e-02 4.99253184e-01 5.11351638e-02 -5.19809484e-01
4.40617651e-01 -3.59021962e-01 -3.42808247e-01 -3.70759100e-01
4.65604007e-01 -3.57475951e-02 -7.36919701e-01 1.35797098e-01
-1.09024560e+00 -9.22800675e-02 -3.35973620e-01 8.15401196e-01
-5.86622119e-01 -1.06624059e-01 5.28452456e-01 -1.73945591e-01
-7.29697764e-01 3.05086747e-02 -3.81566644e-01 7.46975243e-01
7.24796057e-01 3.47060822e-02 -9.06226277e-01 -9.26672816e-01
-6.01454139e-01 5.45441985e-01 1.63965702e-01 6.14414454e-01
6.12562478e-01 -1.33641565e+00 -9.03014243e-01 -3.44929397e-01
3.35992426e-01 -2.89398015e-01 1.88286453e-01 4.84126717e-01
-1.28080532e-01 6.94799006e-01 -2.28265792e-01 -6.18560433e-01
-1.09946775e+00 2.06755027e-01 6.76192343e-02 -7.86009967e-01
-4.59521174e-01 7.54958808e-01 2.85936922e-01 -7.07976758e-01
2.65250504e-01 -3.17730635e-01 -7.11545944e-01 2.83840507e-01
4.24034894e-01 4.26446050e-01 -2.34280512e-01 -6.41914606e-01
-6.88078254e-02 1.34668827e-01 -7.16969818e-02 -1.82454363e-01
7.56657779e-01 -3.79062593e-01 -4.03709739e-01 8.69037628e-01
9.94195700e-01 -4.85189296e-02 -8.25641513e-01 -5.63293815e-01
1.62454411e-01 -2.50945002e-01 -2.21365824e-01 -8.29446197e-01
-2.26374373e-01 8.93100798e-01 -2.98027724e-01 1.14089739e+00
4.89032209e-01 8.78062397e-02 4.80301797e-01 6.78357124e-01
3.98069531e-01 -1.30783486e+00 7.21696258e-01 1.09084141e+00
1.40202379e+00 -1.22579122e+00 5.23030050e-02 -7.72036970e-01
-1.32848048e+00 1.15042710e+00 9.19362426e-01 3.23986970e-02
4.10876304e-01 -3.83873761e-01 -8.59093741e-02 -5.67614734e-01
-9.69409108e-01 -2.62257278e-01 5.97951233e-01 4.64574933e-01
9.34949934e-01 4.33493793e-01 -5.81020892e-01 8.83777440e-01
-6.80693686e-01 -7.13748872e-01 6.73271179e-01 3.80917042e-01
-7.44637668e-01 -7.79089153e-01 9.99178737e-02 5.72526872e-01
7.11887553e-02 -7.06561059e-02 -1.04803705e+00 7.09098876e-01
-8.22541595e-01 1.54962325e+00 -1.60096988e-01 -6.13276660e-01
3.07510376e-01 6.30220324e-02 1.70543447e-01 -9.05632854e-01
-8.94623756e-01 -4.37117189e-01 9.86830413e-01 -5.71635544e-01
-4.12607163e-01 -3.28384101e-01 -1.18535531e+00 -6.32970750e-01
-5.24524927e-01 7.33802140e-01 4.05455619e-01 9.62143123e-01
2.11771652e-01 7.51317918e-01 7.63717949e-01 -4.36240762e-01
-8.11307967e-01 -1.56467545e+00 -4.87116754e-01 5.42527020e-01
1.20438017e-01 -7.86691666e-01 -3.36208761e-01 -4.53131080e-01] | [12.70942211151123, 8.127300262451172] |
9131a627-12cf-4bb1-bd26-9ee2702c5682 | joint-learning-of-set-cardinality-and-state | 1709.04093 | null | http://arxiv.org/abs/1709.04093v2 | http://arxiv.org/pdf/1709.04093v2.pdf | Joint Learning of Set Cardinality and State Distribution | We present a novel approach for learning to predict sets using deep learning.
In recent years, deep neural networks have shown remarkable results in computer
vision, natural language processing and other related problems. Despite their
success, traditional architectures suffer from a serious limitation in that
they are built to deal with structured input and output data, i.e. vectors or
matrices. Many real-world problems, however, are naturally described as sets,
rather than vectors. Existing techniques that allow for sequential data, such
as recurrent neural networks, typically heavily depend on the input and output
order and do not guarantee a valid solution. Here, we derive in a principled
way, a mathematical formulation for set prediction where the output is
permutation invariant. In particular, our approach jointly learns both the
cardinality and the state distribution of the target set. We demonstrate the
validity of our method on the task of multi-label image classification and
achieve a new state of the art on the PASCAL VOC and MS COCO datasets. | ['S. Hamid Rezatofighi', 'Anton Milan', 'Anthony Dick', 'Qinfeng Shi', 'Ian Reid'] | 2017-09-13 | null | null | null | null | ['multi-label-image-classification'] | ['computer-vision'] | [ 5.95122039e-01 -5.13078153e-01 -2.43007466e-01 -6.20165765e-01
-4.49795187e-01 -7.50130057e-01 6.45312011e-01 2.30606124e-01
-3.53931159e-01 6.22001886e-01 -1.50468409e-01 -2.89777040e-01
-4.29534525e-01 -7.66790688e-01 -9.65152264e-01 -8.11785638e-01
5.18035851e-02 6.28416955e-01 -6.82385936e-02 -1.66591734e-01
2.13600233e-01 4.62806553e-01 -2.00191498e+00 6.21097863e-01
1.59006998e-01 1.11242318e+00 -3.85267027e-02 5.39918661e-01
-1.78638443e-01 9.56665337e-01 -4.54189718e-01 -2.16375172e-01
3.74569684e-01 -3.34810346e-01 -9.27824080e-01 1.57431811e-01
6.21905446e-01 8.85828212e-02 -2.95610934e-01 1.05108428e+00
2.29319334e-01 6.95196306e-03 9.07963157e-01 -1.41344178e+00
-4.60680097e-01 7.77189493e-01 -3.57618570e-01 -1.41233608e-01
-1.28633872e-01 -4.78949510e-02 1.46100414e+00 -7.45417058e-01
4.85865265e-01 1.10178649e+00 6.14582121e-01 6.13090277e-01
-1.44541049e+00 -6.52505517e-01 9.71293002e-02 4.67139274e-01
-1.09076393e+00 -3.62114638e-01 7.21545756e-01 -6.33889496e-01
7.77128637e-01 5.06985009e-01 4.57168013e-01 1.03087080e+00
1.23640504e-02 8.96864891e-01 1.31015015e+00 -5.23182154e-01
8.51935595e-02 1.74457505e-01 4.92178321e-01 8.77074242e-01
3.85028236e-02 1.64437834e-02 -4.28307742e-01 -1.34100243e-01
4.77988422e-01 2.45298371e-01 -9.17312428e-02 -8.10302794e-01
-1.33080530e+00 1.11893189e+00 3.52505893e-01 4.12157744e-01
-6.90233335e-02 4.70182002e-01 6.72631264e-01 4.41079110e-01
3.23036492e-01 4.31484044e-01 -5.40850401e-01 3.45512748e-01
-7.68176913e-01 1.81243971e-01 8.70035648e-01 6.83048964e-01
6.75293565e-01 -2.69630909e-01 -2.96183348e-01 8.82544458e-01
2.17773557e-01 3.63012284e-01 5.82131803e-01 -6.14128172e-01
2.64131963e-01 4.37993526e-01 -7.57362545e-02 -9.84247684e-01
-6.38407409e-01 -4.19457018e-01 -1.19462359e+00 1.50206879e-01
4.34451014e-01 1.13906324e-01 -1.01233602e+00 1.88320076e+00
6.46948665e-02 1.02505386e-01 -7.41356611e-02 6.61446095e-01
6.78749859e-01 6.25179887e-01 -3.11081678e-01 -3.01835835e-01
1.04996657e+00 -8.69147241e-01 -6.82862997e-01 -1.45626247e-01
8.48104000e-01 -4.12694126e-01 8.76935065e-01 5.08960962e-01
-6.75849974e-01 -3.66130412e-01 -1.02576900e+00 1.93278268e-01
-4.20729667e-01 8.99246782e-02 7.83962071e-01 5.93450308e-01
-1.03656507e+00 8.00492346e-01 -5.70640385e-01 -1.58780605e-01
5.23938060e-01 7.46766806e-01 -3.59385401e-01 3.73792723e-02
-9.77444410e-01 7.95942664e-01 5.58087409e-01 2.43470967e-01
-7.18831837e-01 -3.25306058e-01 -8.62938762e-01 3.13747704e-01
4.41890031e-01 -5.10144770e-01 1.14640403e+00 -1.09411299e+00
-1.45808160e+00 9.88688588e-01 -2.58841347e-02 -7.99463391e-01
2.41202518e-01 -8.56792405e-02 -8.92151222e-02 -2.21849501e-01
-1.81730270e-01 5.27398825e-01 8.69590759e-01 -1.20268965e+00
-6.18787766e-01 -3.48119169e-01 -3.87131982e-02 -6.74609989e-02
-6.86597168e-01 -1.15868673e-02 -3.92538607e-02 -4.14321214e-01
-6.03780262e-02 -1.16921294e+00 -5.31543136e-01 -1.38444290e-01
-7.06513405e-01 -5.36598563e-01 5.89233220e-01 4.77645285e-02
1.11189115e+00 -2.15719891e+00 3.80182743e-01 3.94910902e-01
3.67045194e-01 4.24467832e-01 -1.39170736e-01 3.67532492e-01
-3.52541208e-01 2.71770656e-01 -2.87262797e-01 -5.24687052e-01
1.64047971e-01 4.03855950e-01 -4.92430717e-01 7.10744739e-01
1.54741541e-01 7.33864427e-01 -8.46410453e-01 -4.35960621e-01
3.05193782e-01 1.50215164e-01 -4.66457009e-01 -5.57057634e-02
-5.68842828e-01 1.86815649e-01 -9.53650251e-02 6.58427477e-02
2.68881321e-01 -7.74739683e-01 4.49077576e-01 -7.00174943e-02
-4.24404629e-02 1.62469581e-01 -1.22746706e+00 1.53659463e+00
-4.02383089e-01 7.34567344e-01 -3.29903036e-01 -1.50424898e+00
8.10841262e-01 2.95748442e-01 8.36115897e-01 -4.01129484e-01
2.37390235e-01 2.41955146e-01 1.70738459e-01 -2.42320418e-01
4.24027950e-01 -1.45953223e-01 -9.80665609e-02 6.29701078e-01
2.10346878e-01 5.13373762e-02 4.78488654e-01 -1.65737465e-01
9.49815929e-01 -1.09334618e-01 3.76325011e-01 -2.40595385e-01
5.34694195e-01 -1.29708186e-01 4.60845977e-01 8.94393146e-01
2.20891014e-01 6.16479218e-01 5.66867232e-01 -7.07399130e-01
-1.11319458e+00 -7.49102175e-01 -9.66957584e-02 1.09991419e+00
-1.96057215e-01 -4.28820074e-01 -4.65683758e-01 -6.94456339e-01
-4.04367335e-02 4.74511892e-01 -7.29455829e-01 -1.88512787e-01
-5.90900362e-01 -7.63924003e-01 3.72553945e-01 4.11825001e-01
1.33835047e-01 -1.12293601e+00 -4.48120743e-01 3.08434159e-01
3.94538790e-02 -1.12950695e+00 -3.91309559e-01 3.55636120e-01
-7.35098779e-01 -1.06723404e+00 -3.47274125e-01 -9.63700056e-01
4.79129225e-01 1.35816157e-01 1.08173764e+00 1.73123896e-01
-2.87278444e-01 3.19953486e-02 -8.71097296e-02 -5.80022216e-01
-4.52870339e-01 6.34092912e-02 1.42096594e-01 4.82577085e-01
2.84931421e-01 -4.78489667e-01 -1.95142701e-01 3.09526443e-01
-1.11044979e+00 1.69869080e-01 5.95398664e-01 1.28788269e+00
6.77637279e-01 7.56852254e-02 5.07183492e-01 -1.29023433e+00
3.81830633e-01 -2.78164774e-01 -7.50007272e-01 3.55556577e-01
-5.17038643e-01 3.52194667e-01 1.00492299e+00 -4.23081666e-01
-4.40504372e-01 5.88974178e-01 -1.22823641e-01 -5.26661873e-01
-2.84601122e-01 5.48522890e-01 8.27926546e-02 1.95815176e-01
5.50518870e-01 3.74361664e-01 1.23592496e-01 -2.95070648e-01
2.56626636e-01 6.52377605e-01 2.03540161e-01 -3.59387249e-01
5.71986675e-01 6.38700545e-01 3.44177425e-01 -6.51063144e-01
-1.29686141e+00 -6.23173892e-01 -8.79550815e-01 -1.69951975e-01
6.37510777e-01 -5.46043754e-01 -1.09161961e+00 5.06910443e-01
-1.15741277e+00 -2.01077938e-01 -2.67750740e-01 3.68155211e-01
-8.56860459e-01 1.61086947e-01 -5.49811542e-01 -7.86260188e-01
-3.13785404e-01 -1.07010639e+00 9.42959189e-01 -1.97366461e-01
-8.33052024e-02 -1.02559471e+00 2.58586858e-03 -8.32090445e-04
1.24695405e-01 3.02126825e-01 1.28008914e+00 -8.53739679e-01
-5.51971138e-01 -2.61453371e-02 -4.77168374e-02 4.21298236e-01
2.19995319e-03 -8.34312811e-02 -9.50076759e-01 -2.94035405e-01
-5.45773841e-03 -5.92233658e-01 1.15649080e+00 2.41735831e-01
1.64747357e+00 -5.04935384e-01 -3.92551243e-01 4.27038938e-01
1.56349480e+00 -3.36411707e-02 4.24197406e-01 1.38909906e-01
8.00760269e-01 6.59131587e-01 3.21212590e-01 4.54439729e-01
5.25363870e-02 6.63857639e-01 6.34664536e-01 -3.51826847e-02
1.05592757e-01 -3.21985111e-02 2.10057557e-01 9.02803540e-01
1.69612288e-01 -4.40345258e-01 -8.30293119e-01 5.23282826e-01
-2.14391851e+00 -1.18376434e+00 -1.43837005e-01 2.35690236e+00
9.22724545e-01 1.95185646e-01 1.69783384e-01 5.76103032e-01
6.20132923e-01 1.72784016e-01 -5.49008727e-01 -3.14825743e-01
-1.94502756e-01 3.97894651e-01 5.52218437e-01 1.69304550e-01
-1.31575525e+00 7.89326966e-01 6.60046625e+00 8.32477391e-01
-1.30997288e+00 -5.06597906e-02 6.29702747e-01 -1.18566029e-01
-3.17238003e-01 -3.47107559e-01 -9.23361063e-01 3.53439510e-01
9.45450306e-01 2.12793693e-01 3.86885703e-01 5.91918349e-01
-3.49004418e-01 3.16565484e-01 -1.59598684e+00 1.21338570e+00
1.66825771e-01 -1.58502460e+00 1.68067783e-01 1.41404867e-01
7.13536143e-01 5.50565347e-02 4.21711981e-01 2.81663358e-01
4.74714488e-01 -1.40327680e+00 6.83319926e-01 5.37138999e-01
6.91926301e-01 -7.51826882e-01 6.44065917e-01 5.30183256e-01
-6.08980477e-01 -2.30895266e-01 -4.72499281e-01 -2.38141343e-01
-2.24676311e-01 6.38473153e-01 -9.41886902e-01 2.97185391e-01
5.08653224e-01 1.04567420e+00 -3.89226019e-01 9.49628711e-01
5.01333661e-02 6.90545797e-01 -4.52267468e-01 -4.82042819e-01
3.25261712e-01 -2.33864680e-01 3.16557437e-01 1.00957668e+00
3.77335660e-02 -2.48727262e-01 2.88535237e-01 7.27109730e-01
-3.10656816e-01 1.75776169e-01 -9.27392066e-01 -2.15562627e-01
9.14712772e-02 1.13936424e+00 -7.77765155e-01 -1.33878797e-01
-3.14572930e-01 6.02778912e-01 7.51155198e-01 1.49823204e-01
-7.29314029e-01 -5.98619170e-02 6.57725155e-01 -2.79516131e-01
4.61468786e-01 -3.80325079e-01 -3.92942220e-01 -1.06811333e+00
-3.07526197e-02 -1.03972292e+00 3.96711797e-01 -8.94679725e-02
-1.34797001e+00 4.69047040e-01 -2.37285405e-01 -1.25846660e+00
-4.85675544e-01 -1.04827964e+00 -1.13338508e-01 4.35805082e-01
-1.24876308e+00 -1.02104354e+00 2.94047832e-01 6.06686056e-01
4.62361157e-01 -2.66764104e-01 1.17245626e+00 3.18968832e-01
-5.58494270e-01 6.45671904e-01 6.54935598e-01 3.14729393e-01
4.24176306e-01 -1.33415067e+00 4.99320365e-02 4.53603685e-01
8.74376774e-01 3.94212276e-01 7.02851355e-01 2.38623880e-02
-1.47466969e+00 -1.15042591e+00 7.52966881e-01 -4.03909236e-01
6.50112092e-01 -7.63383865e-01 -6.32656813e-01 8.20505917e-01
2.69758385e-02 2.17281535e-01 7.98209369e-01 4.62881118e-01
-4.69660759e-01 -1.90415084e-01 -7.62182832e-01 4.91090477e-01
9.66960430e-01 -4.86411363e-01 -1.88236192e-01 8.44241500e-01
6.61818385e-01 -3.50632191e-01 -8.02767515e-01 4.51089531e-01
5.77883542e-01 -9.62915897e-01 9.31496501e-01 -1.14054012e+00
6.66031599e-01 -9.01183486e-02 -2.16462106e-01 -1.38982880e+00
-4.74292636e-01 -3.10986429e-01 -1.65829957e-01 8.86786282e-01
6.05360270e-01 -4.43669111e-01 8.23415577e-01 2.89367251e-02
1.68920774e-02 -1.00718248e+00 -7.16856778e-01 -7.90856898e-01
3.77358645e-02 -4.30937558e-01 5.17605186e-01 8.94106448e-01
-3.43623936e-01 7.21680939e-01 -8.00652683e-01 7.17028454e-02
5.90218723e-01 5.78788757e-01 5.69984198e-01 -1.40606225e+00
-4.04351085e-01 -6.58428609e-01 -5.21145582e-01 -8.18990946e-01
5.51073611e-01 -1.04585493e+00 1.63207948e-01 -1.42001987e+00
2.54365474e-01 -7.28967786e-01 -6.41381443e-01 5.91261148e-01
5.33865541e-02 3.70088428e-01 2.99783856e-01 3.05034548e-01
-8.49951208e-01 3.13695192e-01 9.44243014e-01 -3.70859176e-01
4.08990458e-02 2.34095216e-01 -4.25714046e-01 6.93006575e-01
8.93512130e-01 -7.36254394e-01 -2.01539516e-01 -5.40539742e-01
6.34626806e-01 -1.48843497e-01 3.69428694e-01 -8.81284416e-01
1.54916570e-01 -2.21012697e-01 2.60675758e-01 -6.19055033e-01
6.09781861e-01 -9.63917434e-01 9.11542177e-02 7.08336413e-01
-7.24059761e-01 -1.44547224e-01 6.49762526e-03 6.15699947e-01
-1.91568136e-01 -3.79572332e-01 7.18595386e-01 -4.53809164e-02
-6.19517505e-01 5.59803843e-01 -2.01457933e-01 -1.07438669e-01
9.95154560e-01 1.57866031e-01 -1.92457125e-01 -3.37376535e-01
-6.84225917e-01 1.99646860e-01 1.41560331e-01 4.31923538e-01
5.30060172e-01 -1.38653040e+00 -8.03142965e-01 1.36962295e-01
1.62532926e-01 -2.63872147e-02 1.71533730e-02 6.41436160e-01
-1.86746836e-01 5.97759902e-01 -2.28464201e-01 -7.84213006e-01
-1.47902334e+00 8.16573083e-01 3.30526561e-01 -5.22389472e-01
-3.73357803e-01 7.34302044e-01 1.45577982e-01 -6.16362154e-01
2.47320116e-01 -1.58404261e-01 -4.72081125e-01 7.39711896e-02
4.40551907e-01 -1.00997306e-01 1.15042731e-01 -5.13608575e-01
-1.90406919e-01 3.82548004e-01 -2.51151145e-01 1.33536890e-01
1.62765515e+00 2.03187793e-01 -3.33400488e-01 1.05412769e+00
1.54976535e+00 -3.37132692e-01 -7.66394317e-01 -6.27651393e-01
2.39962265e-01 -1.65812790e-01 -2.35570267e-01 -4.49606001e-01
-1.03918266e+00 1.05580175e+00 5.60444593e-01 4.25492793e-01
8.83080363e-01 -2.09497847e-02 6.59685254e-01 8.81731570e-01
3.67028356e-01 -9.35800254e-01 1.46931008e-01 8.99833620e-01
7.00589061e-01 -1.44881999e+00 -1.25827625e-01 -3.98908913e-01
-2.71846205e-01 1.22884321e+00 2.61700362e-01 -1.73149958e-01
7.50606656e-01 2.41213962e-01 -1.50376543e-01 -1.65790319e-01
-1.01255393e+00 -2.37226307e-01 3.60548854e-01 3.45675856e-01
5.53526223e-01 2.10682184e-01 -1.88840002e-01 2.09098205e-01
-3.46410841e-01 -1.61241323e-01 5.56955814e-01 7.46552110e-01
-4.02634621e-01 -1.17204595e+00 -3.15789878e-01 9.04000998e-01
-6.05001688e-01 -8.04626271e-02 -3.34678560e-01 5.30467391e-01
8.94323513e-02 7.54432976e-01 2.12249100e-01 -4.95515764e-01
2.76280139e-02 1.70368254e-02 6.93866313e-01 -9.30118740e-01
-5.73169172e-01 -3.92236918e-01 -7.36903846e-02 -3.59250307e-01
-6.39487386e-01 -6.87875330e-01 -1.06597090e+00 -1.87970608e-01
-4.56942022e-01 5.35428971e-02 7.55292773e-01 1.13240993e+00
1.85283542e-01 5.71556807e-01 8.05813491e-01 -5.57674468e-01
-1.06324780e+00 -9.87084270e-01 -6.28561199e-01 6.12480700e-01
4.71985549e-01 -5.93546629e-01 6.42039208e-03 8.60862508e-02] | [9.269378662109375, 2.8137853145599365] |
b9aa897d-5da7-4698-a91b-ec890c98a14a | mixture-kernel-graph-attention-network-for | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Suhail_Mixture-Kernel_Graph_Attention_Network_for_Situation_Recognition_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Suhail_Mixture-Kernel_Graph_Attention_Network_for_Situation_Recognition_ICCV_2019_paper.pdf | Mixture-Kernel Graph Attention Network for Situation Recognition | Understanding images beyond salient actions involves reasoning about scene context, objects, and the roles they play in the captured event. Situation recognition has recently been introduced as the task of jointly reasoning about the verbs (actions) and a set of semantic-role and entity (noun) pairs in the form of action frames. Labeling an image with an action frame requires an assignment of values (nouns) to the roles based on the observed image content. Among the inherent challenges are the rich conditional structured dependencies between the output role assignments and the overall semantic sparsity. In this paper, we propose a novel mixture-kernel attention graph neural network (GNN) architecture designed to address these challenges. Our GNN enables dynamic graph structure during training and inference, through the use of a graph attention mechanism, and context-aware interactions between role pairs. We illustrate the efficacy of our model and design choices by conducting experiments on imSitu benchmark dataset, with accuracy improvements of up to 10% over the state-of-the-art.
| [' Leonid Sigal', 'Mohammed Suhail'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['grounded-situation-recognition', 'situation-recognition'] | ['computer-vision', 'computer-vision'] | [ 7.62206256e-01 4.31474537e-01 -2.99388200e-01 -5.34223557e-01
-2.67205060e-01 -4.13799733e-01 8.44863534e-01 3.31338078e-01
-4.07558888e-01 3.68508339e-01 8.40215087e-01 -1.52541742e-01
-6.32526055e-02 -5.70007563e-01 -9.56319094e-01 -4.87323314e-01
-8.42541829e-02 3.02298039e-01 1.69507936e-01 4.97653931e-02
5.36921769e-02 8.61812308e-02 -1.36136377e+00 6.84494555e-01
5.31929076e-01 9.11794662e-01 4.52896833e-01 4.71107036e-01
-1.22885205e-01 1.75516498e+00 -2.82744646e-01 -4.77742761e-01
2.63500866e-02 -3.39602590e-01 -9.64063287e-01 5.85135400e-01
3.08721691e-01 -2.92557061e-01 -4.53744560e-01 1.02964115e+00
-5.11388592e-02 5.37428916e-01 5.05758286e-01 -1.39043653e+00
-9.07799304e-01 8.19568276e-01 -4.85224783e-01 2.77174801e-01
2.88087904e-01 1.33310974e-01 1.25033951e+00 -7.31275916e-01
7.64661491e-01 1.30910611e+00 3.36413048e-02 4.24816936e-01
-1.11406326e+00 -2.59273261e-01 7.37373710e-01 7.82052636e-01
-9.28171754e-01 -3.84998828e-01 9.11144853e-01 -3.78264785e-01
1.19495285e+00 -1.25738606e-01 7.02526152e-01 1.12843347e+00
-1.26756653e-01 9.99652445e-01 5.38112760e-01 -3.97155017e-01
4.13245529e-01 -1.81859314e-01 2.02221259e-01 8.33157420e-01
5.91446161e-02 -5.75552821e-01 -6.72457337e-01 -5.70539124e-02
8.77463520e-01 7.88750276e-02 -2.25830585e-01 -5.62197506e-01
-1.39813542e+00 6.20117962e-01 6.71357095e-01 1.48683071e-01
-6.09443247e-01 7.41376221e-01 3.17416489e-01 -1.79366842e-01
2.90569991e-01 2.81370759e-01 -3.81963342e-01 5.55620603e-02
-2.32720882e-01 -9.84009355e-02 5.65117002e-01 7.47266293e-01
5.05245566e-01 -1.76271051e-02 -3.11319500e-01 5.15982509e-01
3.21308345e-01 6.32219575e-03 1.65392652e-01 -1.06988704e+00
6.43431842e-01 8.72192383e-01 -2.31522005e-02 -9.80292737e-01
-1.35436282e-01 -2.10797802e-01 -5.77508211e-01 -2.35032290e-01
4.17322993e-01 7.85654709e-02 -9.69043732e-01 2.01944375e+00
4.94085908e-01 5.89368284e-01 9.80393738e-02 8.65679979e-01
7.48303294e-01 6.84389591e-01 7.42917359e-01 -3.66040841e-02
1.51296997e+00 -9.81312573e-01 -7.46375501e-01 -6.13806069e-01
4.45615619e-01 -2.17251211e-01 1.11011684e+00 -9.98176038e-02
-1.03955758e+00 -4.44634795e-01 -6.23605669e-01 -4.14796799e-01
-8.67123753e-02 7.64055774e-02 8.02097857e-01 -3.45970057e-02
-9.65028763e-01 2.61139363e-01 -7.72259712e-01 -4.04225320e-01
7.21989095e-01 2.24812001e-01 -4.69117969e-01 -2.05673277e-01
-9.90329087e-01 6.12977326e-01 4.44725573e-01 -1.34947924e-02
-9.60383356e-01 -6.69958353e-01 -1.12381577e+00 4.83934611e-01
9.01516855e-01 -8.28138232e-01 9.77009654e-01 -1.23062718e+00
-1.06037569e+00 8.86210501e-01 -1.76201597e-01 -7.34957099e-01
-1.10468054e-02 -2.28427157e-01 -1.49746642e-01 5.42459428e-01
2.14950144e-01 9.02845383e-01 8.06305230e-01 -1.07037902e+00
-6.01142108e-01 -3.47714275e-01 7.07066357e-01 6.07738853e-01
-1.68413728e-01 1.73276305e-01 -6.74343765e-01 -5.06339669e-01
7.12600872e-02 -7.57680178e-01 -3.90097588e-01 1.24830931e-01
-3.18737864e-01 -3.94149154e-01 5.84523678e-01 -6.23348653e-01
8.69375765e-01 -2.37599564e+00 3.97625029e-01 -1.70117334e-01
2.79649884e-01 -2.08236128e-01 -2.35869944e-01 1.49422795e-01
-2.34442234e-01 -1.76127538e-01 -2.11679697e-01 -3.10927302e-01
-1.19775748e-02 4.39702988e-01 -3.86164099e-01 4.25790042e-01
4.03591961e-01 1.08659625e+00 -1.16111279e+00 -3.87213618e-01
2.40034059e-01 4.85597432e-01 -5.91778517e-01 2.09704146e-01
-6.59424067e-01 3.14136595e-01 -4.60799038e-01 4.83159631e-01
2.14602333e-02 -7.02773511e-01 6.42317533e-01 -5.27205646e-01
4.02597666e-01 3.90556306e-01 -1.01861072e+00 1.55126607e+00
-3.24827075e-01 4.86359358e-01 1.23549163e-01 -1.12920284e+00
3.51825953e-01 2.46206120e-01 3.99066001e-01 -5.46573937e-01
8.59438106e-02 -4.19667333e-01 1.11230008e-01 -5.56421399e-01
2.83496112e-01 2.33636871e-02 -4.79724668e-02 4.70321774e-01
2.75097221e-01 2.70060956e-01 4.05621201e-01 6.60065770e-01
1.09927177e+00 2.51668930e-01 6.03602171e-01 -1.05580524e-01
5.00866234e-01 -1.30912960e-01 4.53024268e-01 6.71528935e-01
-1.83607563e-01 2.79194593e-01 1.00676012e+00 -5.49512982e-01
-6.81504726e-01 -8.84050012e-01 5.26020765e-01 1.39854586e+00
2.28896007e-01 -3.86130512e-01 -6.13629758e-01 -8.71295571e-01
-8.58343914e-02 8.80418777e-01 -7.91368544e-01 -2.33096689e-01
-6.34970725e-01 -4.56281602e-01 1.26032367e-01 9.12155151e-01
5.25346518e-01 -1.39604914e+00 -6.51701450e-01 5.05592190e-02
-2.85909742e-01 -1.82054460e+00 -4.81726646e-01 9.77426469e-02
-5.71148753e-01 -1.36225903e+00 -5.43419644e-02 -7.83607841e-01
9.52200830e-01 2.73902655e-01 1.25550044e+00 -8.52359682e-02
-8.46137032e-02 7.61497140e-01 -2.58076698e-01 -2.47256532e-01
-1.26948819e-01 -3.62624139e-01 -3.02927941e-01 6.67457700e-01
8.10527503e-02 -5.24028778e-01 -5.42043209e-01 -6.62640929e-02
-1.08850670e+00 5.60774326e-01 5.77788770e-01 6.44750059e-01
6.90093756e-01 -2.34809425e-03 2.96576202e-01 -1.10771883e+00
2.63558179e-01 -6.63388431e-01 -4.73664016e-01 5.48707128e-01
-1.82154849e-02 2.37841934e-01 4.14203286e-01 -4.43639904e-01
-1.35085988e+00 3.56545806e-01 4.08189446e-01 -5.67406952e-01
-2.00302616e-01 5.00465989e-01 -5.28336406e-01 3.89991850e-01
3.59786004e-01 2.04944968e-01 -2.15192258e-01 -9.08708870e-02
7.82533944e-01 -4.42366861e-02 6.56574130e-01 -6.41514361e-01
3.59555095e-01 7.56668568e-01 9.79332104e-02 -7.36592889e-01
-1.24274957e+00 -5.62351704e-01 -4.68631953e-01 -3.52712989e-01
1.27502441e+00 -1.17593086e+00 -5.82945287e-01 2.60769129e-01
-1.13384891e+00 -5.97759545e-01 -4.76129919e-01 3.67888570e-01
-5.31374872e-01 3.37507188e-01 -6.07767582e-01 -7.14907646e-01
-9.47068632e-03 -1.05595934e+00 1.01471305e+00 1.05232432e-01
-8.37483928e-02 -1.10023546e+00 -5.00619471e-01 7.64786124e-01
3.67308594e-02 3.10995609e-01 1.23666883e+00 -5.11084318e-01
-1.01959491e+00 1.31844044e-01 -5.41539848e-01 1.14481583e-01
1.16317198e-01 -4.77349281e-01 -7.26515293e-01 8.90082195e-02
5.56835160e-03 -3.12422723e-01 9.26456749e-01 5.53713024e-01
1.43227649e+00 -3.65841538e-01 -2.39510491e-01 3.12987387e-01
1.26727414e+00 5.59860766e-02 5.30236363e-01 5.44660212e-03
1.04224098e+00 6.20613396e-01 3.50992054e-01 2.72704929e-01
5.99091768e-01 5.33005655e-01 7.30379999e-01 2.05739494e-02
-2.53446281e-01 -3.48519236e-01 3.08058918e-01 5.18693328e-02
-1.08459126e-02 -3.88961464e-01 -8.07306826e-01 6.81171060e-01
-2.21299267e+00 -1.21180785e+00 -8.98440704e-02 1.80917811e+00
6.90030992e-01 1.73389420e-01 -1.75393924e-01 -2.89260507e-01
6.72851264e-01 6.12589896e-01 -6.20667994e-01 3.48964818e-02
-2.44516172e-02 -2.84060594e-02 2.97848463e-01 4.71876085e-01
-1.28393495e+00 1.08678782e+00 5.45440722e+00 4.43413645e-01
-6.86363876e-01 1.05297476e-01 8.03976715e-01 -1.48831472e-01
-2.87387013e-01 2.11375087e-01 -5.04215479e-01 1.68951198e-01
6.22740328e-01 -1.04274102e-01 4.69901890e-01 8.42284322e-01
1.35988012e-01 -3.29372346e-01 -1.16124201e+00 8.50760639e-01
3.81949365e-01 -1.47411597e+00 3.05620342e-01 -1.07120574e-01
5.66447139e-01 -8.90653059e-02 -1.03985488e-01 2.20780194e-01
4.38630551e-01 -8.30410123e-01 9.03565347e-01 3.16179067e-01
5.64747989e-01 -3.81491601e-01 3.13766599e-01 1.24442339e-01
-1.35287881e+00 -3.76615167e-01 -1.01108074e-01 -3.49813163e-01
2.34501168e-01 4.08221841e-01 -8.49282086e-01 3.47854912e-01
4.41114128e-01 1.02375376e+00 -4.16707188e-01 4.19329375e-01
-7.99385190e-01 7.72516131e-01 -8.81948769e-02 1.11210994e-01
2.97577143e-01 -1.35873690e-01 3.88104856e-01 9.03198183e-01
-3.17154378e-01 4.25794929e-01 3.97740871e-01 9.16756988e-01
-4.11757976e-01 -6.08185753e-02 -4.74301934e-01 -1.78269789e-01
2.20554829e-01 1.28271842e+00 -9.32865083e-01 -5.17258823e-01
-6.49639010e-01 1.00489199e+00 6.12617850e-01 4.75320220e-01
-9.50552285e-01 3.33935916e-01 7.53255069e-01 2.57240385e-01
3.83482993e-01 -2.03524470e-01 -9.04132500e-02 -1.22245455e+00
1.11268893e-01 -6.74082637e-01 7.15561867e-01 -1.04816484e+00
-1.04347241e+00 2.85971314e-01 2.73371369e-01 -7.58945465e-01
-8.78033936e-02 -5.77673018e-01 -5.19708335e-01 4.12246525e-01
-1.38972867e+00 -1.33561885e+00 -2.89055437e-01 6.18954599e-01
7.55446017e-01 1.26097858e-01 5.26255667e-01 7.76373148e-02
-4.77151424e-01 -1.29029408e-01 -5.52656054e-01 2.67581671e-01
2.19847828e-01 -1.29024351e+00 4.20854777e-01 9.85492408e-01
5.00348866e-01 4.42097217e-01 4.47622508e-01 -6.56080723e-01
-1.22357023e+00 -1.27550161e+00 9.13793921e-01 -5.54144502e-01
7.98990428e-01 -4.79251742e-01 -7.25474477e-01 1.03938818e+00
2.81829417e-01 2.52610385e-01 7.41034925e-01 -4.20276932e-02
-4.37405527e-01 1.16260119e-01 -7.21094966e-01 7.93044448e-01
1.38767695e+00 -6.53058708e-01 -6.98575914e-01 3.66513759e-01
9.18767929e-01 -3.77598196e-01 -2.73561686e-01 1.84857339e-01
2.15701982e-01 -7.72706270e-01 1.13017118e+00 -8.66204977e-01
5.61638117e-01 -4.17044193e-01 -3.50646377e-01 -8.49554956e-01
-5.54215431e-01 -1.82662740e-01 -4.22140062e-01 9.38687205e-01
2.34101042e-01 -1.02157399e-01 7.35365927e-01 9.02876973e-01
8.76609422e-03 -7.60806620e-01 -6.57971740e-01 -1.47113964e-01
-7.38511682e-01 -4.95996773e-01 3.01480681e-01 8.86015177e-01
-1.47016570e-01 8.11409771e-01 -2.89048821e-01 4.25261587e-01
5.29576063e-01 1.38315946e-01 5.14220774e-01 -9.24028516e-01
-5.44833243e-01 -1.07011482e-01 -4.11861658e-01 -8.87274623e-01
5.84808946e-01 -8.25846434e-01 -9.31325927e-02 -1.91212332e+00
5.44964135e-01 1.87738955e-01 -3.71998817e-01 9.31636274e-01
-5.68747282e-01 -8.52272883e-02 4.13874358e-01 3.25927325e-02
-1.20840955e+00 5.16309381e-01 1.01983917e+00 -2.39011392e-01
-4.27718237e-02 -3.90692830e-01 -8.22503090e-01 8.81497383e-01
5.28641403e-01 -3.06984425e-01 -8.93412948e-01 -6.70747817e-01
2.80033439e-01 1.32997364e-01 7.35153735e-01 -8.40163231e-01
1.91087276e-01 -3.60501587e-01 3.50191355e-01 -2.00324371e-01
4.81229424e-01 -8.30181241e-01 6.70768023e-02 1.72438160e-01
-6.79356933e-01 -8.61108974e-02 1.19620916e-02 1.00987804e+00
-2.12255716e-01 1.74755752e-01 4.28203106e-01 -3.76924872e-01
-1.32814550e+00 2.83280730e-01 -2.53029764e-01 1.27488971e-01
1.11816120e+00 8.29466432e-03 -1.04926594e-01 -5.40866554e-01
-9.13138032e-01 2.65680522e-01 2.44778737e-01 4.23218817e-01
6.74736798e-01 -1.09361029e+00 -4.11034256e-01 8.43379572e-02
2.19518244e-01 9.63678956e-02 4.16746616e-01 6.99815571e-01
-7.63593838e-02 1.68329298e-01 -9.85772908e-02 -3.52702647e-01
-1.29649603e+00 5.38629651e-01 3.17959934e-01 -5.55276453e-01
-4.62173581e-01 9.80432808e-01 7.77197778e-01 -7.66236782e-02
2.69422382e-01 -5.59308529e-01 -4.66358274e-01 -2.61365045e-02
3.66868705e-01 -8.74707848e-02 -2.93234646e-01 -6.87129200e-01
-3.23083252e-01 7.04237819e-02 5.39022461e-02 3.66070606e-02
1.38583386e+00 1.28894160e-02 -2.56791860e-01 1.86752364e-01
9.06207442e-01 -3.68440688e-01 -1.56316411e+00 -3.55845571e-01
-1.50480932e-02 -3.55689287e-01 -1.31174400e-02 -6.75696611e-01
-1.00301027e+00 4.82186139e-01 4.83361669e-02 -1.83117762e-02
1.14242566e+00 5.40318787e-01 4.14316177e-01 3.21534932e-01
1.69571668e-01 -9.24043417e-01 6.24773324e-01 4.64128554e-01
8.58038843e-01 -1.10612428e+00 -1.26762003e-01 -7.86765099e-01
-8.98055196e-01 7.68709242e-01 7.00397193e-01 -1.05399147e-01
5.65819025e-01 -8.05630833e-02 -2.23706543e-01 -4.57093865e-01
-1.01472211e+00 -5.22707283e-01 3.27266008e-01 4.66006726e-01
3.08763444e-01 1.08611986e-01 -3.31093222e-02 5.49038351e-01
4.78048086e-01 -1.47263080e-01 2.72941709e-01 9.26476061e-01
-2.38004148e-01 -8.65638912e-01 -5.32003753e-02 4.40472424e-01
-4.45106804e-01 -2.02641264e-01 -3.51754874e-01 5.95776200e-01
1.65690154e-01 9.04617071e-01 3.20234507e-01 6.08579330e-02
2.58748084e-01 1.85638502e-01 4.87789392e-01 -9.22265589e-01
-2.54602492e-01 -2.19240800e-01 2.33645290e-01 -8.83288205e-01
-7.72376001e-01 -6.80608451e-01 -1.52192092e+00 3.48278403e-01
1.13042220e-01 -1.11220844e-01 2.85862625e-01 1.15593100e+00
4.88959521e-01 7.73964107e-01 1.60754532e-01 -5.64746320e-01
-2.16137454e-01 -5.69787085e-01 -2.95064718e-01 9.57966328e-01
5.79513013e-02 -5.81990421e-01 -1.10518955e-01 5.84162116e-01] | [10.355497360229492, 1.4165197610855103] |
55e73192-9a0d-4687-acf2-e9241752df25 | leveraging-human-feedback-to-scale | 2305.12894 | null | https://arxiv.org/abs/2305.12894v1 | https://arxiv.org/pdf/2305.12894v1.pdf | Leveraging Human Feedback to Scale Educational Datasets: Combining Crowdworkers and Comparative Judgement | Machine Learning models have many potentially beneficial applications in education settings, but a key barrier to their development is securing enough data to train these models. Labelling educational data has traditionally relied on highly skilled raters using complex, multi-class rubrics, making the process expensive and difficult to scale. An alternative, more scalable approach could be to use non-expert crowdworkers to evaluate student work, however, maintaining sufficiently high levels of accuracy and inter-rater reliability when using non-expert workers is challenging. This paper reports on two experiments investigating using non-expert crowdworkers and comparative judgement to evaluate complex student data. Crowdworkers were hired to evaluate student responses to open-ended reading comprehension questions. Crowdworkers were randomly assigned to one of two conditions: the control, where they were asked to decide whether answers were correct or incorrect (i.e., a categorical judgement), or the treatment, where they were shown the same question and answers, but were instead asked to decide which of two candidate answers was more correct (i.e., a comparative/preference-based judgement). We found that using comparative judgement substantially improved inter-rater reliability on both tasks. These results are in-line with well-established literature on the benefits of comparative judgement in the field of educational assessment, as well as with recent trends in artificial intelligence research, where comparative judgement is becoming the preferred method for providing human feedback on model outputs when working with non-expert crowdworkers. However, to our knowledge, these results are novel and important in demonstrating the beneficial effects of using the combination of comparative judgement and crowdworkers to evaluate educational data. | ['Owen Henkel Libby Hills'] | 2023-05-22 | null | null | null | null | ['reading-comprehension'] | ['natural-language-processing'] | [ 2.14325324e-01 2.52213687e-01 4.45259921e-02 -6.85392499e-01
-6.28719389e-01 -9.67329800e-01 3.78950864e-01 7.55850911e-01
-1.08956897e+00 5.84078610e-01 2.09508657e-01 -7.94200122e-01
-3.07563126e-01 -5.72380006e-01 -2.85247028e-01 -4.88134235e-01
7.30375528e-01 6.55321956e-01 1.03679486e-01 -9.84311476e-03
6.18295610e-01 2.38681301e-01 -1.92681956e+00 1.85904250e-01
1.19191468e+00 4.84299123e-01 1.85143843e-01 7.18555272e-01
-2.96668619e-01 9.64771569e-01 -1.10829079e+00 -8.39862943e-01
5.73760532e-02 -6.11400783e-01 -9.22938108e-01 -8.73894710e-03
7.70898163e-01 -3.40809643e-01 6.91106856e-01 8.37611914e-01
5.78926921e-01 5.60437918e-01 7.42529929e-01 -1.00663888e+00
-9.07198250e-01 2.33982444e-01 -3.89263421e-01 1.60888553e-01
9.16143477e-01 2.29002565e-01 9.14429545e-01 -4.79154110e-01
3.83731753e-01 1.14751589e+00 6.56701207e-01 5.97518802e-01
-1.40059388e+00 -8.81954312e-01 1.44330695e-01 2.39603117e-01
-7.36204147e-01 -4.42745239e-01 2.15584934e-01 -9.19974267e-01
6.06902540e-01 7.15529099e-02 7.02924371e-01 8.69061887e-01
-1.70046672e-01 1.49509549e-01 1.97233725e+00 -6.00497425e-01
5.41488767e-01 7.49737859e-01 2.50956237e-01 3.35154146e-01
4.20673251e-01 1.63414419e-01 -5.18841505e-01 -2.07891136e-01
4.80861962e-01 -1.62752986e-01 -3.02112222e-01 -2.00228751e-01
-8.97518277e-01 1.05152726e+00 3.48907918e-01 3.05162400e-01
-2.64263153e-01 -5.08343577e-01 2.19444096e-01 8.04696739e-01
4.16880310e-01 1.02156270e+00 -2.47954816e-01 -6.09843850e-01
-8.70823264e-01 4.17848349e-01 9.67652261e-01 2.07492426e-01
3.94301444e-01 -2.50774980e-01 -2.67597795e-01 1.01855302e+00
3.57476860e-01 3.40525985e-01 7.90193796e-01 -1.06055665e+00
4.66112077e-01 7.20863283e-01 4.69006598e-01 -9.97360945e-01
-1.40054420e-01 6.01839600e-03 -1.76764220e-01 9.33001578e-01
9.20588374e-01 -2.92113602e-01 -7.47010887e-01 1.58019722e+00
4.16791826e-01 -3.75858188e-01 -1.14967287e-01 1.15537441e+00
7.64055789e-01 3.03615749e-01 4.82473552e-01 -1.20553881e-01
1.51050246e+00 -6.59409642e-01 -5.28300464e-01 -6.50319681e-02
7.84865797e-01 -9.44650531e-01 1.45207095e+00 5.23873448e-01
-1.12932277e+00 -7.76683331e-01 -8.45430970e-01 -1.02812335e-01
-3.35669667e-01 -3.72486919e-01 3.42925727e-01 1.17843568e+00
-1.06221163e+00 4.15939659e-01 -5.26468754e-01 -1.63896739e-01
2.86738455e-01 5.45330793e-02 -4.37728822e-01 -2.74821371e-01
-9.96063232e-01 1.31808734e+00 -1.54310331e-01 -2.22080663e-01
-6.91795111e-01 -6.34328306e-01 -1.01359773e+00 1.75956547e-01
1.25507787e-02 -2.56972998e-01 1.76732850e+00 -1.25011098e+00
-1.53494143e+00 1.05595815e+00 -2.61283666e-01 9.56693143e-02
6.24874949e-01 -1.49027824e-01 1.18682124e-01 -7.27681592e-02
4.17711467e-01 5.05663574e-01 3.07328522e-01 -1.06445694e+00
-7.49518812e-01 -5.65302908e-01 2.80596733e-01 5.66117764e-01
-4.04336661e-01 3.47953647e-01 4.97304529e-01 -2.93906569e-01
-6.77617118e-02 -7.69356549e-01 -3.66687588e-02 -2.54182756e-01
3.58900130e-01 -7.89289117e-01 3.94814938e-01 -5.49746215e-01
1.00306129e+00 -1.82496297e+00 -2.77625859e-01 1.96259886e-01
3.44418824e-01 6.42584145e-01 -2.57911626e-02 3.46759766e-01
-1.22422628e-01 3.66656363e-01 1.41775087e-01 -8.05187747e-02
1.49856314e-01 -8.74161720e-02 2.29163393e-01 1.33208975e-01
3.70677710e-02 4.45362508e-01 -1.24537587e+00 -3.17918867e-01
1.65598214e-01 1.46690086e-01 -2.84437209e-01 3.92168850e-01
3.62915635e-01 3.86852205e-01 -1.07525505e-01 2.66691029e-01
1.95975825e-01 1.29329547e-01 -1.21918172e-01 6.26633644e-01
-2.46643960e-01 5.33385754e-01 -1.21646500e+00 9.63319361e-01
-6.87335491e-01 9.33790028e-01 2.67543703e-01 -8.44839573e-01
1.08067393e+00 6.15245879e-01 -2.98655659e-01 -8.67415965e-01
6.56410754e-02 3.16723824e-01 3.14868122e-01 -7.10311592e-01
4.84926313e-01 -3.94388437e-01 1.92408130e-01 9.67724025e-01
-2.54167616e-01 -6.24934971e-01 3.41905177e-01 -2.76405476e-02
1.06457376e+00 -1.97453145e-03 2.39307836e-01 -3.33618402e-01
2.76447713e-01 -4.08821851e-02 4.06479478e-01 9.46056724e-01
-4.29627299e-01 2.33112857e-01 4.46689755e-01 -4.53457743e-01
-7.45152295e-01 -8.64859343e-01 9.78832319e-02 1.60063577e+00
-2.69623965e-01 7.96668902e-02 -8.58930171e-01 -6.72256529e-01
-1.00866102e-01 9.27378297e-01 -3.87972862e-01 -5.70016094e-02
-1.73726737e-01 -2.16582462e-01 2.34804139e-01 4.64492708e-01
3.83269489e-01 -1.03180277e+00 -1.15486276e+00 1.08540617e-01
-2.13884294e-01 -4.80179667e-01 -2.88422078e-01 2.79356036e-02
-7.20535576e-01 -9.97559249e-01 -8.42345715e-01 -8.63036633e-01
7.60841489e-01 5.56062281e-01 1.31125796e+00 4.68373477e-01
8.22960809e-02 5.98078310e-01 -3.68741035e-01 -8.82228911e-01
-3.70382994e-01 -8.85893926e-02 -1.27135143e-01 -4.10198748e-01
9.47644651e-01 -3.49487990e-01 -4.21096951e-01 5.75372577e-01
-8.58091414e-01 -2.67764091e-01 4.61528778e-01 8.39429796e-01
-6.19557872e-02 -5.73085155e-03 8.46042335e-01 -9.47214007e-01
1.34066200e+00 -2.52714008e-01 -4.66318876e-01 2.08405688e-01
-8.79400432e-01 -2.69137919e-01 3.29999149e-01 -6.78504705e-01
-1.17842102e+00 -3.10881466e-01 8.77668243e-03 -5.15500270e-02
-6.19299829e-01 6.44036889e-01 2.93842107e-01 -1.20699108e-01
1.02250993e+00 -5.12356102e-01 1.80386249e-02 -1.68243706e-01
-1.87569857e-01 9.46421921e-01 3.53067815e-01 -1.01175582e+00
7.29772210e-01 -2.27304503e-01 -4.12076473e-01 -5.06810904e-01
-1.01563406e+00 -5.51633239e-01 -3.92037809e-01 -4.19666767e-01
9.93267536e-01 -9.02027428e-01 -9.17653084e-01 2.34759971e-01
-8.20763469e-01 -8.32001567e-01 -2.26082727e-01 7.84268856e-01
-1.84730515e-01 -1.10101245e-01 -5.09443879e-01 -1.07703984e+00
2.72133183e-02 -1.34173107e+00 6.16780460e-01 6.61516011e-01
-7.92311013e-01 -1.22851908e+00 1.37974396e-01 1.13975430e+00
4.52256411e-01 1.17982097e-01 9.55043256e-01 -9.36055601e-01
1.48836002e-01 -3.48719597e-01 6.40758500e-02 3.04008186e-01
4.17312309e-02 -1.02861620e-01 -1.07427597e+00 -1.24190062e-01
1.85315266e-01 -1.00035870e+00 4.24624801e-01 4.17058021e-02
7.52112329e-01 -2.43265018e-01 7.05180541e-02 -2.20945701e-01
8.86368096e-01 2.54627317e-01 2.73442507e-01 4.18627143e-01
3.29524785e-01 1.12838721e+00 5.62225640e-01 -2.58754253e-01
7.27059424e-01 5.99417150e-01 -1.97754920e-01 2.56534535e-02
1.81464836e-01 -2.16957122e-01 3.63692552e-01 6.85583293e-01
-3.27351332e-01 -1.16299100e-01 -1.11140668e+00 7.22115397e-01
-1.35659611e+00 -1.04404354e+00 -1.52521700e-01 2.39749622e+00
1.01732194e+00 6.22445643e-02 1.64615303e-01 5.25835812e-01
8.49113643e-01 -5.45975566e-02 -2.18691707e-01 -1.04219317e+00
4.47365016e-01 5.29896140e-01 5.16244955e-02 6.06444418e-01
-6.13539934e-01 4.25293416e-01 5.91720343e+00 4.99610096e-01
-8.85196626e-01 1.18012764e-01 6.95588470e-01 -1.52623504e-01
-3.16094816e-01 6.63861120e-03 -4.76204127e-01 3.32015216e-01
1.10625923e+00 -1.15425354e-02 3.98992568e-01 5.00973642e-01
3.27340037e-01 -6.32472813e-01 -1.07721353e+00 5.89936435e-01
-2.55219731e-02 -5.08072495e-01 -4.67569619e-01 2.38595139e-02
1.05206788e+00 -6.37922406e-01 -1.60828009e-01 5.76213479e-01
8.19921315e-01 -1.33432853e+00 7.71828592e-01 2.00892881e-01
5.00134051e-01 -6.08654499e-01 1.14548230e+00 6.65506423e-01
-6.14518523e-01 -5.16472496e-02 -2.48272434e-01 -1.07601047e+00
-7.85602108e-02 2.46586457e-01 -1.06778944e+00 3.33117023e-02
8.63033772e-01 -1.17877327e-01 -5.43383539e-01 9.76489484e-01
-6.55695319e-01 8.21795344e-01 -3.26282228e-03 -4.42697853e-01
9.63706896e-02 -1.08800724e-01 -1.41718775e-01 8.48403990e-01
2.95594871e-01 2.20314622e-01 1.21612042e-01 5.28422415e-01
2.18205098e-02 1.35028183e-01 -6.52540982e-01 1.69926628e-01
8.45021427e-01 1.23682797e+00 -8.77835929e-01 -2.50631511e-01
-5.24648368e-01 3.92959088e-01 5.64782202e-01 3.92622322e-01
-8.53052810e-02 -6.53773546e-01 1.96319625e-01 3.36126298e-01
-1.60577133e-01 1.41570598e-01 -5.95998764e-01 -4.56797063e-01
-1.26609355e-02 -1.21420634e+00 4.62830663e-01 -9.28412974e-01
-1.35767841e+00 1.27769902e-01 1.30809024e-01 -6.50915802e-01
-2.72652805e-01 -5.84712565e-01 -7.71762073e-01 1.29754984e+00
-1.04587209e+00 -4.09473628e-01 -6.96865380e-01 2.81363159e-01
4.30861264e-01 1.78878918e-01 7.23375857e-01 1.15075316e-02
-2.75076240e-01 6.02702737e-01 -2.42641062e-01 2.53095686e-01
1.03978324e+00 -1.48285079e+00 -9.59588885e-02 3.42689663e-01
-9.36153810e-04 7.43416429e-01 5.99379778e-01 -6.91969275e-01
-4.87339228e-01 -4.20957118e-01 1.15792990e+00 -6.56531036e-01
2.94063568e-01 -8.34495574e-02 -1.23182392e+00 3.96839797e-01
2.19136015e-01 -5.86920798e-01 1.15711570e+00 3.88490468e-01
-3.82555217e-01 1.76710650e-01 -1.38435423e+00 5.00136495e-01
6.60159171e-01 -6.78942740e-01 -1.19315970e+00 1.13589644e-01
3.87840420e-01 -5.36524177e-01 -9.42041695e-01 1.14298977e-01
6.74123824e-01 -1.08906126e+00 5.77986419e-01 -5.76332569e-01
7.83393860e-01 -1.92154691e-01 2.50274658e-01 -1.80756330e+00
-2.70039350e-01 -1.41340256e-01 5.97110093e-01 1.39650571e+00
4.24035192e-01 -7.19682634e-01 5.77080309e-01 1.20997918e+00
-1.00418657e-01 -7.45455801e-01 -8.07338834e-01 -4.56545472e-01
3.71888936e-01 -3.13957520e-02 6.01324081e-01 1.15338480e+00
1.24824688e-01 4.85802323e-01 3.02192032e-01 -1.87133253e-01
7.45179653e-02 -1.71558484e-01 7.20145762e-01 -1.66003764e+00
1.33308908e-03 -7.64993548e-01 -3.80289972e-01 -3.84021342e-01
2.30185419e-01 -5.02325714e-01 3.14407527e-01 -1.76502979e+00
9.88793839e-03 -4.29081768e-01 1.78149249e-02 2.35697970e-01
-8.23628306e-01 1.21754415e-01 3.37186128e-01 3.45473923e-02
-4.67441946e-01 9.73210856e-02 1.27697241e+00 3.10262710e-01
-3.12291712e-01 2.24510103e-01 -1.03796899e+00 7.92903483e-01
8.06191981e-01 -4.62486148e-01 -6.05254591e-01 -6.45225525e-01
3.33194196e-01 -1.23836279e-01 3.56119096e-01 -9.75423694e-01
5.11664927e-01 -4.43103701e-01 6.91921353e-01 2.46267170e-01
4.11821753e-02 -5.39632559e-01 -2.04446241e-01 1.70057699e-01
-7.63754606e-01 5.64266562e-01 1.98420033e-01 9.13337246e-02
-2.17869386e-01 -1.00441957e+00 5.86720705e-01 -4.01551396e-01
-3.20591986e-01 -5.40273845e-01 -6.53250933e-01 2.72361159e-01
1.07487082e+00 -3.72902364e-01 -4.38779444e-01 -8.26362550e-01
-3.04748744e-01 2.48734906e-01 6.79969251e-01 2.62526602e-01
4.29589480e-01 -1.08783948e+00 -8.06116223e-01 7.84797687e-03
1.12945259e-01 1.67540818e-01 5.31832688e-02 7.23175168e-01
-4.38579738e-01 2.75196314e-01 -9.03018191e-02 -3.12440068e-01
-1.37153184e+00 2.71904200e-01 1.00830600e-01 -2.73934633e-01
1.06966048e-01 1.00054669e+00 -8.13167021e-02 -7.66180575e-01
3.46791863e-01 -3.51443812e-02 -5.38138986e-01 3.39662164e-01
8.35040033e-01 6.30537033e-01 2.24720389e-01 -5.43863118e-01
1.62583545e-01 5.18683016e-01 2.21204311e-02 -5.05417764e-01
8.60598624e-01 4.98311482e-02 1.39415815e-01 6.95368767e-01
5.58613420e-01 2.31432170e-01 -9.04443264e-01 -8.25687423e-02
8.53268057e-02 -6.82656348e-01 5.92404604e-02 -1.10698509e+00
-4.20422554e-01 1.01273525e+00 5.08558750e-01 4.12914991e-01
7.96622574e-01 -2.23452523e-01 -7.05894157e-02 5.67279577e-01
3.62377852e-01 -1.27259469e+00 1.88799188e-01 2.12960213e-01
7.31160879e-01 -1.48266721e+00 7.02220052e-02 3.82435843e-02
-8.26014578e-01 8.80043805e-01 1.00392210e+00 2.71820754e-01
1.26515571e-02 -3.10607076e-01 6.00026727e-01 -3.04454505e-01
-7.76002169e-01 8.86122882e-02 3.81164461e-01 8.80156338e-01
1.08833063e+00 3.15255374e-01 -4.51299578e-01 4.01944757e-01
-7.29739726e-01 3.41794752e-02 6.68005466e-01 1.09940815e+00
-6.14867032e-01 -8.77686560e-01 -1.04815328e+00 7.68784702e-01
-3.88451070e-01 1.50531739e-01 -6.45592749e-01 6.03349686e-01
1.92288086e-01 1.69668293e+00 2.02050105e-01 -4.46954817e-02
4.99174654e-01 3.83470386e-01 2.38742098e-01 -9.40577090e-01
-1.14845848e+00 -6.39882624e-01 8.01220909e-02 -1.21581733e-01
-5.05645514e-01 -8.34022701e-01 -6.49885356e-01 -5.00421822e-01
-3.76934230e-01 5.73973477e-01 5.91820002e-01 1.14125299e+00
-1.38356656e-01 2.71628022e-01 3.03497165e-01 -7.17124164e-01
-8.44248056e-01 -1.29506445e+00 -1.76495492e-01 6.96420848e-01
1.80789962e-01 -7.83570468e-01 -5.12215495e-01 -1.66913867e-01] | [11.689867973327637, 8.944730758666992] |
f50df864-ce84-4d69-b015-b7196bdf9ac0 | cross-lingual-semantic-specialization-via | null | null | https://aclanthology.org/D19-1226 | https://aclanthology.org/D19-1226.pdf | Cross-lingual Semantic Specialization via Lexical Relation Induction | Semantic specialization integrates structured linguistic knowledge from external resources (such as lexical relations in WordNet) into pretrained distributional vectors in the form of constraints. However, this technique cannot be leveraged in many languages, because their structured external resources are typically incomplete or non-existent. To bridge this gap, we propose a novel method that transfers specialization from a resource-rich source language (English) to virtually any target language. Our specialization transfer comprises two crucial steps: 1) Inducing noisy constraints in the target language through automatic word translation; and 2) Filtering the noisy constraints via a state-of-the-art relation prediction model trained on the source language constraints. This allows us to specialize any set of distributional vectors in the target language with the refined constraints. We prove the effectiveness of our method through intrinsic word similarity evaluation in 8 languages, and with 3 downstream tasks in 5 languages: lexical simplification, dialog state tracking, and semantic textual similarity. The gains over the previous state-of-art specialization methods are substantial and consistent across languages. Our results also suggest that the transfer method is effective even for lexically distant source-target language pairs. Finally, as a by-product, our method produces lists of WordNet-style lexical relations in resource-poor languages. | ['Goran Glava{\\v{s}}', "Ivan Vuli{\\'c}", 'Roi Reichart', 'Edoardo Maria Ponti', 'Anna Korhonen'] | 2019-11-01 | null | null | null | ijcnlp-2019-11 | ['lexical-simplification'] | ['natural-language-processing'] | [-5.42674735e-02 1.30594462e-01 -6.61579728e-01 -3.62184048e-01
-5.90137124e-01 -9.43922043e-01 5.28088212e-01 2.08572865e-01
-6.71142936e-01 8.51173878e-01 6.29922271e-01 -4.05415863e-01
1.85135320e-01 -6.03588581e-01 -3.29604387e-01 2.69912984e-02
3.37247372e-01 8.85969043e-01 2.99028516e-01 -7.99961984e-01
-1.78100213e-01 2.12712914e-01 -9.57618177e-01 2.23227203e-01
1.35450351e+00 5.60880065e-01 3.32950741e-01 -4.73184846e-02
-5.11718512e-01 3.71614248e-01 -4.69845504e-01 -7.98679113e-01
3.47428322e-01 -3.28590453e-01 -9.28461075e-01 -2.50471324e-01
2.07885325e-01 1.89608023e-01 -2.82425806e-02 1.10018444e+00
3.73537660e-01 3.92880946e-01 3.63350451e-01 -7.32520461e-01
-8.69540811e-01 9.78547812e-01 -2.15598028e-02 7.90149420e-02
4.17765975e-01 8.32154900e-02 1.58380413e+00 -9.77309346e-01
1.02282977e+00 1.49965715e+00 6.36945009e-01 6.21773958e-01
-1.60088658e+00 -6.23698533e-01 5.34999311e-01 -1.46003254e-02
-1.32501566e+00 -5.18671215e-01 5.69786072e-01 -3.83064210e-01
1.59852695e+00 1.27814099e-01 3.27191919e-01 1.17984855e+00
-1.13745779e-01 3.11947167e-01 7.79137254e-01 -6.24653578e-01
6.56682439e-03 4.60757434e-01 2.66313136e-01 6.54300213e-01
2.79527396e-01 -1.16524264e-01 -6.52366400e-01 -2.32460514e-01
2.80630916e-01 -2.67899871e-01 -2.23076716e-01 -1.84481576e-01
-1.12073398e+00 9.60041761e-01 -7.00922124e-03 5.25292873e-01
2.69944854e-02 -3.32883507e-01 7.17850566e-01 4.97726023e-01
6.88703299e-01 1.01534092e+00 -9.86449480e-01 1.55885801e-01
-6.08528674e-01 -2.31881253e-02 1.30413330e+00 1.39201689e+00
9.05810535e-01 -5.65936491e-02 -1.31282762e-01 1.15009463e+00
1.35822952e-01 5.68962991e-01 6.91507101e-01 -8.15724075e-01
5.72859824e-01 6.30424142e-01 -1.79663315e-01 -5.25519490e-01
-1.99235991e-01 -1.18814021e-01 -3.00553501e-01 -3.34172428e-01
3.23900223e-01 -3.90791923e-01 -5.78947902e-01 2.29555941e+00
2.50093341e-01 -2.41665855e-01 2.58847088e-01 6.51360631e-01
8.06569636e-01 4.19341058e-01 3.42078328e-01 -2.81793714e-01
1.41546750e+00 -1.04696560e+00 -7.92971432e-01 -4.49470222e-01
9.52382088e-01 -9.00855005e-01 1.65587831e+00 -8.08358043e-02
-9.92542744e-01 -3.87933433e-01 -7.36187518e-01 -3.37163925e-01
-6.59363031e-01 -2.44949013e-02 7.71587789e-01 5.53577065e-01
-8.90224814e-01 5.23273766e-01 -5.38078606e-01 -7.75582552e-01
1.11898705e-02 4.27367747e-01 -4.23994750e-01 2.89104879e-02
-1.77635527e+00 1.37414885e+00 4.47201729e-01 -4.84939098e-01
-4.43124473e-01 -9.64560390e-01 -1.16134024e+00 1.86783612e-01
7.17190385e-01 -8.40124786e-01 1.19592547e+00 -9.23850536e-01
-1.68156981e+00 1.01348233e+00 -3.88279527e-01 -1.28331646e-01
5.32562099e-03 -4.80540901e-01 -3.16511661e-01 -4.06915247e-01
4.11331087e-01 2.37533808e-01 3.45341831e-01 -1.05740070e+00
-6.02006912e-01 -1.22613065e-01 2.43270591e-01 4.81514156e-01
-4.40191388e-01 3.52682978e-01 -7.92933583e-01 -7.12005734e-01
-3.75155121e-01 -8.06703091e-01 -2.41986379e-01 -4.89520371e-01
-4.46155369e-01 -5.87499738e-01 3.85073781e-01 -7.51195073e-01
1.46575308e+00 -1.98637867e+00 5.00956893e-01 3.88807356e-01
-3.41600999e-02 5.70445299e-01 -2.71314770e-01 5.98198295e-01
4.26243767e-02 3.57470512e-01 -2.75668442e-01 -3.88875425e-01
1.04972750e-01 5.02299607e-01 -4.92956549e-01 9.71749797e-02
1.83186397e-01 1.04173183e+00 -1.02129412e+00 -4.42352653e-01
4.36177552e-02 2.35817820e-01 -7.80463934e-01 1.24291047e-01
-5.18208861e-01 3.77142698e-01 -3.09286535e-01 4.89591181e-01
1.13411918e-01 8.34596083e-02 7.91936040e-01 -2.74575353e-01
5.51029965e-02 1.02295136e+00 -8.90103102e-01 1.99541354e+00
-9.44982469e-01 6.68689162e-02 -4.76942062e-02 -7.69108653e-01
7.52228141e-01 2.07898051e-01 2.39958882e-01 -6.00226343e-01
1.18386820e-01 4.58860815e-01 1.83697134e-01 -3.85093570e-01
4.26954329e-01 -3.49452525e-01 -3.92511785e-01 4.77221072e-01
5.69249570e-01 -3.73048604e-01 4.81933385e-01 2.57095993e-01
9.25324321e-01 1.61988258e-01 5.55151105e-01 -6.03901029e-01
7.56563485e-01 -2.48585567e-02 9.20297921e-01 3.64128649e-01
3.12333163e-02 4.91418727e-02 4.13617522e-01 2.81642713e-02
-8.45784903e-01 -9.83443618e-01 -9.31287371e-03 1.61856616e+00
1.59815952e-01 -9.89720285e-01 -5.97272038e-01 -1.01594639e+00
2.28681922e-01 9.87572432e-01 -2.31189832e-01 -1.97249085e-01
-9.15210843e-01 -1.45511508e-01 7.77692556e-01 4.35037136e-01
4.07832414e-02 -1.12847733e+00 3.83690372e-02 1.76246241e-01
-1.80259764e-01 -1.37943602e+00 -9.20333207e-01 2.54609942e-01
-5.01041770e-01 -9.14586186e-01 -1.41912505e-01 -9.40178514e-01
4.11968738e-01 -4.95222509e-02 1.40010452e+00 1.52410224e-01
2.52610147e-01 1.85004234e-01 -2.43551672e-01 -1.02694787e-01
-4.99978662e-01 4.96287644e-01 4.73774672e-01 -3.71544421e-01
8.01261544e-01 -5.26812494e-01 4.72806282e-02 1.39663219e-01
-6.67194963e-01 -4.16413903e-01 1.78078651e-01 9.52606320e-01
5.23452640e-01 -3.68406355e-01 5.02643824e-01 -1.53204870e+00
9.68627095e-01 -3.95441830e-01 -5.96812725e-01 5.45082867e-01
-6.44154131e-01 4.09632534e-01 8.91858280e-01 -6.02897584e-01
-1.38697481e+00 2.81289350e-02 -1.68281972e-01 -8.51459950e-02
9.49711949e-02 7.63875365e-01 -6.28594518e-01 2.78373539e-01
8.11969817e-01 -2.65488982e-01 -2.35580415e-01 -6.53118730e-01
9.61592317e-01 5.20164967e-01 4.05989856e-01 -1.06467569e+00
8.38822722e-01 1.11709498e-01 -3.56045067e-01 -8.16535115e-01
-1.12161386e+00 -6.02346897e-01 -7.91771293e-01 6.38825178e-01
8.82803977e-01 -8.46346796e-01 -3.32787007e-01 -1.45166397e-01
-1.46759915e+00 -4.52122569e-01 -6.69631243e-01 5.26358366e-01
-1.77078754e-01 4.94738728e-01 -7.49418139e-01 -2.21916199e-01
-3.32801342e-01 -1.10904491e+00 7.73509443e-01 -7.00004846e-02
-9.23838735e-01 -1.45157456e+00 2.35857293e-01 1.75763756e-01
5.45689940e-01 -4.97742921e-01 1.37213206e+00 -1.33749080e+00
-2.04963028e-01 2.29077667e-01 -1.47156835e-01 5.54457784e-01
4.11902219e-01 -2.46394962e-01 -6.27476037e-01 -2.42676884e-02
-1.05992462e-02 -3.83539200e-01 6.78716421e-01 1.00758923e-02
4.85720068e-01 -2.35797465e-01 -4.21496063e-01 7.21875846e-01
1.15932834e+00 -5.19835725e-02 1.22507580e-01 3.08396835e-02
8.87230277e-01 8.41710031e-01 4.94610429e-01 -3.25683132e-03
4.64129686e-01 8.09442401e-01 -2.80582845e-01 -1.17559478e-01
-4.25915122e-01 -3.71492445e-01 5.33448935e-01 1.32817197e+00
1.50392815e-01 -4.03542630e-02 -9.44021523e-01 6.21282220e-01
-1.82251930e+00 -6.07617736e-01 2.26136953e-01 2.13670111e+00
1.56756115e+00 3.06229088e-02 -1.65204138e-01 -6.84738636e-01
6.46699429e-01 -1.30394427e-03 -4.76498574e-01 -4.86572444e-01
-4.45447147e-01 6.33355260e-01 3.87620777e-01 1.23817551e+00
-8.43388319e-01 1.91658843e+00 5.61456871e+00 9.08433676e-01
-1.05225968e+00 4.83643144e-01 9.91985127e-02 9.51047093e-02
-8.84792328e-01 4.39386606e-01 -1.13957512e+00 2.08288684e-01
8.07210326e-01 -4.84192789e-01 5.87426662e-01 8.17947865e-01
-6.14643879e-02 1.62818775e-01 -1.42076766e+00 7.67891884e-01
2.11815044e-01 -1.14391840e+00 2.28781700e-01 -3.01499039e-01
7.13694215e-01 1.61783665e-01 -2.72262275e-01 6.87360466e-01
8.60544086e-01 -9.72567677e-01 4.33516562e-01 7.75281638e-02
1.13669825e+00 -5.89590013e-01 6.30232453e-01 2.65897036e-01
-1.20595586e+00 3.14340532e-01 -3.78895640e-01 -1.57174002e-02
4.05546963e-01 5.91361761e-01 -5.61385512e-01 4.24991965e-01
3.15176070e-01 6.69870853e-01 -2.89603084e-01 2.75812984e-01
-8.41573834e-01 6.01018429e-01 -4.10221636e-01 -2.83645019e-02
7.20501877e-03 -3.46928120e-01 7.78642416e-01 1.67764831e+00
-3.92742790e-02 6.02449626e-02 4.60078746e-01 8.80504429e-01
-4.29608732e-01 5.68236291e-01 -7.68141150e-01 -4.57596593e-02
7.41659224e-01 1.12056887e+00 -3.19027603e-01 -4.92515534e-01
-8.27230453e-01 9.93295431e-01 7.68686116e-01 5.37752151e-01
-4.37053233e-01 -5.04057050e-01 1.04091156e+00 -1.11851305e-01
1.60574675e-01 -2.91674793e-01 -4.82726008e-01 -1.69036472e+00
-5.89894783e-03 -1.12700224e+00 3.90214711e-01 -6.22774698e-02
-1.83021712e+00 7.78393626e-01 4.95791957e-02 -5.19682050e-01
-3.20693761e-01 -6.30729675e-01 -3.65138382e-01 1.33904231e+00
-1.47959208e+00 -1.24465418e+00 2.82010555e-01 7.88245440e-01
7.37964451e-01 -5.01434982e-01 1.17355418e+00 4.60333347e-01
-4.20057654e-01 8.25400054e-01 -2.25444049e-01 2.16643102e-02
1.18758988e+00 -1.15951359e+00 5.01963675e-01 9.27480996e-01
2.08964661e-01 1.21602201e+00 4.88902032e-01 -1.03030860e+00
-1.08109713e+00 -1.02046156e+00 1.61171818e+00 -7.08374381e-01
1.08529210e+00 -8.17553699e-01 -1.05974102e+00 9.62729514e-01
4.57194716e-01 -2.35445470e-01 9.05560970e-01 8.37168217e-01
-6.98824286e-01 6.67288750e-02 -8.27820957e-01 9.68927383e-01
1.44967747e+00 -8.26922178e-01 -1.02941012e+00 4.93867785e-01
1.28021502e+00 -3.74808460e-01 -6.58775568e-01 2.69285500e-01
2.04944819e-01 -2.84078419e-01 7.38032937e-01 -1.04917717e+00
1.31905943e-01 -9.09709930e-02 -4.46964651e-01 -1.63575649e+00
-1.93911374e-01 -9.61899042e-01 1.08498111e-01 1.47560155e+00
1.00272918e+00 -7.97828794e-01 3.26463819e-01 7.36066043e-01
-2.84783423e-01 -6.81142211e-01 -7.67482996e-01 -1.01011205e+00
4.41749334e-01 -4.72108215e-01 5.65070391e-01 1.25013900e+00
4.04241711e-01 1.17967200e+00 -4.22545895e-02 -2.02080876e-01
9.96791571e-02 6.99190721e-02 6.10362709e-01 -1.28899646e+00
-4.59380627e-01 -4.12804127e-01 7.21163526e-02 -1.18061304e+00
1.11933422e+00 -1.37826073e+00 -7.41200801e-03 -1.26954031e+00
1.89076513e-02 -8.28396559e-01 -1.25672176e-01 6.80026114e-01
-3.36094528e-01 -1.49819434e-01 2.04934418e-01 1.72511607e-01
-4.88973975e-01 7.33356237e-01 9.34095204e-01 -3.96840721e-02
-4.90199506e-01 -2.28198454e-01 -9.42914486e-01 9.48382378e-01
6.78245902e-01 -4.82690930e-01 -5.32310367e-01 -5.66855192e-01
3.16108823e-01 -4.77785826e-01 -2.94743955e-01 -2.78869659e-01
2.08574429e-01 -3.87335926e-01 -3.46173048e-01 6.07295856e-02
2.20181078e-01 -7.83376575e-01 -3.05636942e-01 1.87567949e-01
-3.61090899e-01 -4.19271439e-02 2.91314065e-01 1.27153456e-01
-2.91406900e-01 -3.05040598e-01 6.67299688e-01 -2.19554961e-01
-7.49036849e-01 5.05031794e-02 -1.62366614e-01 8.04769397e-01
7.43663907e-01 2.18916267e-01 -3.26608866e-01 -2.57915318e-01
-6.91545129e-01 1.69653639e-01 4.51738298e-01 5.69744706e-01
2.15789735e-01 -1.17116642e+00 -5.74830353e-01 2.87127316e-01
1.70591369e-01 -2.40905419e-01 -4.66515422e-01 6.36834502e-01
-6.74327835e-04 5.46853185e-01 6.60635531e-02 -1.66583091e-01
-1.13248658e+00 5.86398005e-01 2.14905873e-01 -5.38271606e-01
-3.69341433e-01 9.72842216e-01 3.67613971e-01 -9.62311089e-01
2.41956457e-01 -4.08598036e-01 6.44265860e-02 9.27392244e-02
-1.37743670e-02 -6.93769306e-02 1.57916605e-01 -7.93173432e-01
-5.75394869e-01 4.72771764e-01 -2.01293305e-01 -3.76091272e-01
1.18945611e+00 -3.03662181e-01 -2.85309494e-01 3.84929121e-01
9.69756126e-01 7.46829748e-01 -5.89418054e-01 -7.82359302e-01
3.99868250e-01 -2.22339734e-01 -4.94571716e-01 -7.58497596e-01
-7.21616387e-01 6.06530190e-01 -2.64313400e-01 -2.23582163e-01
8.02056015e-01 3.44690859e-01 1.17480516e+00 6.32823110e-01
2.74553418e-01 -1.43157494e+00 -1.64809659e-01 1.41427910e+00
7.19709039e-01 -1.16863668e+00 -2.47962743e-01 -9.46502686e-01
-9.54683781e-01 6.49289966e-01 6.60873532e-01 -2.43676603e-02
7.74094641e-01 3.33228379e-01 3.29975784e-01 -2.02172384e-01
-7.71814764e-01 -5.65172851e-01 4.60505575e-01 6.64510369e-01
8.58188808e-01 -8.44975002e-05 -7.41131067e-01 1.04206896e+00
-3.20601463e-01 -4.59155142e-01 1.27485394e-01 6.80770397e-01
-1.42395481e-01 -1.88362205e+00 2.19892144e-01 2.62547344e-01
-3.91618073e-01 -8.37543488e-01 -6.67424798e-01 8.02884340e-01
2.89383203e-01 8.05783272e-01 -3.51748705e-01 -2.55517483e-01
7.18114734e-01 4.71324503e-01 1.55535907e-01 -1.43029165e+00
-9.62208748e-01 -5.97249903e-03 5.54667950e-01 -8.32946301e-01
-1.33073866e-01 -4.87109274e-01 -1.35853386e+00 -2.66323030e-01
-4.21322525e-01 2.13674709e-01 2.96672136e-01 1.15886390e+00
3.68501365e-01 2.80753791e-01 1.33159995e-01 -4.32290852e-01
-6.93721414e-01 -9.02008533e-01 -4.71799731e-01 8.46279740e-01
-1.58659950e-01 -6.74884319e-01 -3.28158140e-01 8.47915709e-02] | [10.84362506866455, 9.72349739074707] |
cf8b7659-791c-43e9-a2c1-13840f054cac | multi-channel-speech-denoising-for-machine | 2202.08793 | null | https://arxiv.org/abs/2202.08793v1 | https://arxiv.org/pdf/2202.08793v1.pdf | Multi-Channel Speech Denoising for Machine Ears | This work describes a speech denoising system for machine ears that aims to improve speech intelligibility and the overall listening experience in noisy environments. We recorded approximately 100 hours of audio data with reverberation and moderate environmental noise using a pair of microphone arrays placed around each of the two ears and then mixed sound recordings to simulate adverse acoustic scenes. Then, we trained a multi-channel speech denoising network (MCSDN) on the mixture of recordings. To improve the training, we employ an unsupervised method, complex angular central Gaussian mixture model (cACGMM), to acquire cleaner speech from noisy recordings to serve as the learning target. We propose a MCSDN-Beamforming-MCSDN framework in the inference stage. The results of the subjective evaluation show that the cACGMM improves the training data, resulting in better noise reduction and user preference, and the entire system improves the intelligibility and listening experience in noisy situations. | ['Simon Carlile', 'Malcolm Slaney', 'Kyle Hoefer', 'E. Merve Kaya', 'Cong Han'] | 2022-02-17 | null | null | null | null | ['speech-denoising'] | ['speech'] | [-1.43995002e-01 -3.77541333e-01 7.29315817e-01 -3.94574195e-01
-9.08638179e-01 -1.87911570e-01 1.30410850e-01 -4.85110819e-01
-3.83039296e-01 2.52711713e-01 6.94958746e-01 -3.80356252e-01
-8.11251774e-02 -5.42495787e-01 -3.74228179e-01 -9.38624918e-01
4.48013842e-02 -3.01883340e-01 1.97619274e-02 -2.72221059e-01
-1.96298584e-01 1.88941762e-01 -1.78213072e+00 3.29501688e-01
9.14040089e-01 1.00926602e+00 7.90077209e-01 1.11639631e+00
1.64834946e-01 5.12766838e-01 -9.26526606e-01 -3.17388773e-01
2.21763119e-01 -2.18060479e-01 8.59808475e-02 8.48160461e-02
-6.40363395e-02 -4.36811775e-01 -4.02885765e-01 1.38004315e+00
1.40911531e+00 6.94508851e-01 3.13620269e-01 -6.78984463e-01
-1.89369142e-01 5.54946005e-01 -1.39050275e-01 1.10423252e-01
2.67811000e-01 -7.43243918e-02 5.37814498e-01 -9.58915710e-01
-3.01653206e-01 1.63949955e+00 6.73156261e-01 6.33824170e-01
-9.23792064e-01 -8.54105413e-01 -1.23548834e-02 2.00394958e-01
-1.34316754e+00 -1.04850137e+00 7.98171997e-01 1.45889133e-01
7.45361984e-01 7.89755285e-01 3.14030439e-01 1.04312932e+00
-2.02463642e-01 6.89839900e-01 9.28230464e-01 -4.81287658e-01
4.96216655e-01 -5.09576173e-03 -1.08366897e-02 -6.52361885e-02
-2.30043188e-01 3.58240098e-01 -5.53331673e-01 -1.80994734e-01
6.13910615e-01 -2.72841871e-01 -4.61358517e-01 5.49649477e-01
-6.14875138e-01 4.08550441e-01 1.93107322e-01 3.87404740e-01
-5.26543975e-01 -1.70820817e-01 1.78198591e-01 4.61889267e-01
6.88824058e-01 1.27779678e-01 -3.05336177e-01 -7.28954151e-02
-8.12842429e-01 -9.30989720e-03 1.00443017e+00 5.71829319e-01
1.05961159e-01 4.17068034e-01 5.13792746e-02 1.61366665e+00
8.37380767e-01 9.30184066e-01 5.07277608e-01 -1.04907763e+00
5.32409966e-01 -4.04787034e-01 1.47878528e-01 -7.78673768e-01
-1.49394855e-01 -1.04526019e+00 -1.23682582e+00 3.29758942e-01
-1.67742640e-01 -4.72650975e-01 -1.01529610e+00 1.65644073e+00
2.81674087e-01 4.56737548e-01 3.65868688e-01 1.04794407e+00
8.72032166e-01 1.05139506e+00 -4.45454605e-02 -5.80035567e-01
9.72976685e-01 -1.09582663e+00 -1.31530213e+00 -2.51877666e-01
6.66971318e-03 -1.28072679e+00 9.42813218e-01 1.05327487e+00
-1.20509410e+00 -1.07186472e+00 -9.88543212e-01 2.96539485e-01
9.99760702e-02 1.68127522e-01 1.24810889e-01 1.24962604e+00
-1.15978777e+00 2.03992426e-01 -7.77944982e-01 2.90619731e-01
3.86905223e-02 1.72734752e-01 -3.54091711e-02 -3.34660113e-01
-1.03269100e+00 4.91963804e-01 -1.58426374e-01 6.03307962e-01
-1.08289754e+00 -2.75046974e-01 -7.73968756e-01 2.00573966e-01
-8.91755074e-02 -6.67183757e-01 1.51563060e+00 -8.72862756e-01
-1.85130858e+00 1.12175509e-01 -4.62313473e-01 -2.68767267e-01
3.43424439e-01 -5.82832694e-01 -1.10551739e+00 -2.68342942e-01
-2.48195767e-01 5.66114560e-02 1.17508924e+00 -1.68522644e+00
-5.41806042e-01 -2.66044706e-01 -5.45221329e-01 5.94619632e-01
-2.91087121e-01 6.55444115e-02 -4.96129423e-01 -9.86551225e-01
5.15623510e-01 -2.87627250e-01 -5.37442148e-01 -7.03924239e-01
-4.77602452e-01 2.28080422e-01 5.70095122e-01 -1.38602519e+00
1.63848877e+00 -2.92361879e+00 -2.71682758e-02 5.45689225e-01
-6.73243776e-02 5.75259268e-01 -1.38523504e-01 1.34599522e-01
-1.54206499e-01 -2.51173079e-01 -1.50886746e-02 -1.04971898e+00
-1.94887146e-02 2.20951170e-01 -1.88133761e-01 1.43330589e-01
-5.27327538e-01 6.82786554e-02 -5.75980961e-01 -2.22461279e-02
3.20966691e-01 8.11248124e-01 -6.66167200e-01 7.71899998e-01
2.59858906e-01 5.52852154e-01 -1.99203640e-02 3.64378393e-01
1.19931173e+00 6.79066181e-01 -3.13385844e-01 -6.63756207e-02
-5.07391952e-02 2.90385753e-01 -1.79131520e+00 1.44957697e+00
-8.15583944e-01 3.70426416e-01 1.02382290e+00 -5.40773213e-01
1.09963739e+00 6.65541708e-01 -2.00252369e-01 -5.10388911e-01
1.28671765e-01 4.33824718e-01 3.89365256e-02 -9.85699356e-01
1.29926875e-01 -1.80931211e-01 5.01121879e-01 -7.26983845e-02
2.49847416e-02 -3.76587212e-01 -3.71662050e-01 -1.77209795e-01
8.44079494e-01 -5.39859533e-01 -3.03330366e-02 1.69653162e-01
6.17015004e-01 -1.03836477e+00 3.65959346e-01 8.55244517e-01
-3.10067125e-02 7.11498201e-01 -3.07662904e-01 2.92247236e-01
-7.21831620e-01 -1.49281609e+00 1.10398035e-03 1.31632793e+00
-1.64753377e-01 -1.16228841e-01 -1.00447214e+00 2.81376839e-01
-4.52406585e-01 1.00004888e+00 1.47388548e-01 -1.49691865e-01
-1.34076446e-01 -6.64498270e-01 5.65043509e-01 2.91627586e-01
7.07399726e-01 -7.12701023e-01 4.53755915e-01 2.34258741e-01
-3.56288821e-01 -6.03767574e-01 -4.50361013e-01 4.06508058e-01
-4.74916548e-01 -2.24757090e-01 -7.85573304e-01 -1.12703037e+00
3.96777034e-01 3.30138475e-01 6.27096772e-01 -2.28940874e-01
2.12181091e-01 1.98385864e-01 -2.65018374e-01 -7.12535381e-01
-5.94293356e-01 -5.52881598e-01 4.49069649e-01 2.49708429e-01
1.16158776e-01 -1.10811830e+00 -6.14521205e-01 4.47252601e-01
-9.09699380e-01 -4.35113341e-01 5.34228504e-01 7.21841574e-01
2.93442756e-01 9.73261118e-01 7.41719961e-01 -2.04558372e-01
1.22330391e+00 -3.98381978e-01 -3.48067671e-01 -1.65291160e-01
-1.63858697e-01 -5.58415771e-01 6.08325779e-01 -5.18723369e-01
-1.79399645e+00 -2.49817461e-01 -1.06281865e+00 -2.29220405e-01
-4.80133474e-01 4.61183786e-01 -1.02760744e+00 2.66729444e-01
6.83558822e-01 2.52244562e-01 -6.90809544e-03 -9.37432408e-01
2.42766261e-01 1.59857810e+00 9.36535299e-01 -8.61687735e-02
5.88992655e-01 7.76132289e-03 -6.31639063e-01 -1.23592031e+00
-3.70259672e-01 -9.31322873e-01 -1.44167049e-02 -1.42852128e-01
6.94887161e-01 -1.29700196e+00 -4.81817812e-01 9.84700501e-01
-1.14436150e+00 -2.06836075e-01 1.54322073e-01 9.73585486e-01
-1.65851504e-01 4.11110342e-01 -7.03103602e-01 -1.58765078e+00
-3.95342439e-01 -1.20177877e+00 7.25354373e-01 2.32959956e-01
4.45872033e-03 -7.47023880e-01 -1.49503216e-01 4.97932255e-01
6.51217759e-01 -7.43787706e-01 5.43685079e-01 -5.04914701e-01
-1.04200356e-02 -2.81209499e-01 3.74669611e-01 1.32899809e+00
2.97341138e-01 -4.72738177e-01 -1.65027905e+00 -1.26669407e-01
8.99261475e-01 1.66284174e-01 5.96771002e-01 9.61580694e-01
1.09006202e+00 -1.94792315e-01 5.92358224e-02 5.58145165e-01
9.61806893e-01 5.90972424e-01 1.06734753e+00 -3.26951407e-02
2.85267651e-01 4.36445743e-01 2.95757264e-01 4.86020833e-01
-2.68519502e-02 1.44410089e-01 3.23762894e-01 -4.33997542e-01
-3.27319413e-01 -1.13129616e-01 4.48348910e-01 1.44043100e+00
5.84144741e-02 -5.96456885e-01 -4.06241536e-01 3.84537876e-01
-1.18643200e+00 -7.80068934e-01 -1.12877890e-01 2.16849637e+00
9.16321278e-01 1.62322074e-01 -1.23848908e-01 6.37775779e-01
8.76226246e-01 6.06314205e-02 -1.58177659e-01 -3.37215722e-01
-1.92740336e-01 4.49408978e-01 1.62170529e-01 9.88967836e-01
-9.21839833e-01 4.19196248e-01 6.31858110e+00 1.18530142e+00
-9.26942229e-01 1.32380977e-01 7.01072276e-01 -6.19769506e-02
-2.05813393e-01 -6.55694783e-01 -4.69489127e-01 4.39517349e-01
9.28849578e-01 3.13007534e-01 9.53683257e-01 7.49702871e-01
8.74864519e-01 -1.22735783e-01 -6.15825176e-01 1.21581614e+00
-4.64661755e-02 -4.48151827e-01 -3.29257399e-01 -3.38723153e-01
4.78268206e-01 -7.28613064e-02 4.17299181e-01 2.71206439e-01
1.58859268e-01 -1.05806172e+00 6.50511563e-01 6.99404180e-01
3.71789008e-01 -8.91142726e-01 8.79731238e-01 7.13107705e-01
-9.33759451e-01 -2.69682229e-01 -3.63170534e-01 -1.76976413e-01
2.95673400e-01 9.39095020e-01 -5.38979471e-01 4.45768356e-01
9.71176505e-01 -4.70002927e-02 -4.86748740e-02 1.51888323e+00
-3.50699037e-01 1.15478623e+00 -4.37395722e-01 1.92044318e-01
-1.97544426e-01 -3.03817004e-01 8.62291098e-01 1.15248561e+00
4.70039576e-01 2.64281034e-01 -1.10435240e-01 2.94716388e-01
-1.35603072e-02 3.77154835e-02 -1.86044276e-01 4.25375164e-01
6.50915623e-01 9.49347436e-01 -4.30095233e-02 -2.03448966e-01
-1.25657618e-01 8.66620660e-01 -5.17732263e-01 8.90649796e-01
-5.68271458e-01 -5.65749645e-01 7.48548925e-01 -2.65168220e-01
2.53094047e-01 -2.53175557e-01 -6.02634549e-01 -7.10550666e-01
9.35143828e-02 -1.19652712e+00 -5.16949221e-02 -1.22164714e+00
-1.30192256e+00 8.32517862e-01 -5.34800828e-01 -1.10465086e+00
8.07443112e-02 -3.96862149e-01 -7.54821181e-01 1.49948955e+00
-1.20747256e+00 -4.80246931e-01 -5.32223135e-02 6.15965724e-01
7.70523846e-01 -2.37411290e-01 9.63282049e-01 8.59968781e-01
-5.47161341e-01 5.79662621e-01 6.27171934e-01 -1.69512287e-01
7.42433906e-01 -1.05561841e+00 2.04611614e-01 1.07642186e+00
-1.70384571e-01 7.60272086e-01 1.15988553e+00 -5.08392692e-01
-1.00593019e+00 -1.14516294e+00 5.46787858e-01 3.66653241e-02
1.97831631e-01 -5.17090440e-01 -8.26802135e-01 2.25338563e-01
3.76665771e-01 -4.32557285e-01 9.26513672e-01 2.79571880e-02
2.95025688e-02 -4.16901618e-01 -1.12387538e+00 6.48308814e-01
7.54736543e-01 -5.85787475e-01 -6.70635164e-01 3.56867127e-02
9.47863042e-01 -4.04539198e-01 -6.07813597e-01 1.86972424e-01
3.60493451e-01 -7.87298441e-01 1.09121418e+00 -4.60850261e-02
-6.27218839e-03 -4.78166044e-01 -7.83757389e-01 -2.01028943e+00
-2.82964736e-01 -8.58387768e-01 9.88868251e-02 1.55080104e+00
6.69494271e-01 -6.48794830e-01 2.66221970e-01 2.48173997e-01
-6.09098434e-01 -2.70999223e-01 -9.33078587e-01 -6.00841284e-01
-3.28591585e-01 -1.10160971e+00 4.77560073e-01 3.77566129e-01
-3.60781968e-01 4.34931368e-01 -5.51853657e-01 8.39900732e-01
6.23604953e-01 -8.50075364e-01 6.37160659e-01 -7.27143466e-01
-6.40367150e-01 -1.30389094e-01 -4.64115962e-02 -1.63925457e+00
-2.43233040e-01 -1.47228763e-01 6.02511287e-01 -1.64710772e+00
-6.11775994e-01 -1.25074461e-02 -2.95546561e-01 -2.81223983e-01
-5.53848267e-01 -4.15426753e-02 -3.05581003e-01 -2.90471673e-01
-2.97468752e-01 7.57085979e-01 1.12848806e+00 -1.38226986e-01
-3.82861763e-01 9.15451884e-01 -6.94373786e-01 8.10775757e-01
4.61014122e-01 -2.24102393e-01 -6.51502252e-01 -6.83021367e-01
-2.38969520e-01 1.71650887e-01 1.37250051e-01 -1.23643315e+00
4.27462190e-01 3.63473505e-01 4.31750119e-01 -6.21314526e-01
7.93532670e-01 -1.09800947e+00 2.40069509e-01 2.54358143e-01
-3.29195827e-01 -6.54346406e-01 2.73445785e-01 7.24494338e-01
-5.91181695e-01 -8.54138657e-02 7.53130138e-01 9.60751846e-02
-2.82672554e-01 -2.69982427e-01 -7.06345439e-01 -6.33657932e-01
4.10884827e-01 5.60296141e-02 7.31927231e-02 -9.79968548e-01
-1.25217593e+00 5.26543595e-02 -3.99009168e-01 1.89079747e-01
8.07781041e-01 -1.26315832e+00 -7.06237793e-01 6.76099122e-01
-5.07433295e-01 6.43576905e-02 6.43045247e-01 5.47381878e-01
-2.46841922e-01 -1.02830470e-01 3.92604530e-01 -5.35986960e-01
-1.40455461e+00 2.55623311e-01 6.66927218e-01 2.78964370e-01
-5.67947477e-02 1.44997799e+00 1.49070367e-01 -5.86373031e-01
9.21067774e-01 -2.60396659e-01 -3.86663347e-01 -3.67146283e-01
7.77032614e-01 6.69112027e-01 4.42654908e-01 -3.38657081e-01
1.86961740e-01 2.15198845e-02 2.25543752e-01 -5.63394606e-01
1.27830577e+00 -6.84003830e-01 -1.91686466e-01 3.06504965e-01
1.05374551e+00 6.27179205e-01 -9.64098573e-01 -1.82575226e-01
-4.78894055e-01 -5.76475322e-01 4.85452622e-01 -1.10772407e+00
-7.82304704e-01 7.79609263e-01 1.10109055e+00 3.98433179e-01
1.78699744e+00 -4.08994496e-01 8.01519334e-01 4.50241297e-01
4.28647101e-02 -1.06433177e+00 -5.34461588e-02 4.20752108e-01
8.49962652e-01 -8.40524316e-01 -7.23308980e-01 -5.38199306e-01
-4.90824252e-01 8.07627857e-01 3.76349568e-01 1.76203251e-01
1.10873115e+00 6.56377017e-01 6.62150323e-01 3.15481693e-01
-3.37550670e-01 -7.60614201e-02 1.11053862e-01 9.01063859e-01
3.36930513e-01 1.73345640e-01 6.64605796e-02 1.28586698e+00
-6.74187481e-01 -3.24095309e-01 -5.11110350e-02 5.83874762e-01
-8.46741319e-01 -8.76567841e-01 -1.07634044e+00 3.59592050e-01
-6.07408166e-01 -4.71992314e-01 3.06581371e-02 -2.70466447e-01
2.60994941e-01 2.00157142e+00 -1.66632324e-01 -6.60318017e-01
8.01894128e-01 5.08833155e-02 -6.80811182e-02 -4.31517482e-01
-3.98591667e-01 1.10850906e+00 2.17534810e-01 -1.63413435e-01
-1.05656616e-01 -3.05020362e-01 -8.94985676e-01 -1.61056355e-01
-6.25180304e-01 4.83368218e-01 9.65573430e-01 8.43804300e-01
1.13452412e-01 1.03579509e+00 1.08895028e+00 -7.88928330e-01
-8.23259115e-01 -1.54019976e+00 -1.00580907e+00 9.59918499e-02
6.53780162e-01 -7.70483240e-02 -7.95900524e-01 1.27905190e-01] | [15.023606300354004, 5.863903522491455] |
ed7b4533-7fff-4db9-816f-64cbca08fd8b | on-the-effectiveness-of-iterative-learning | 2111.09434 | null | https://arxiv.org/abs/2111.09434v3 | https://arxiv.org/pdf/2111.09434v3.pdf | On the Effectiveness of Iterative Learning Control | Iterative learning control (ILC) is a powerful technique for high performance tracking in the presence of modeling errors for optimal control applications. There is extensive prior work showing its empirical effectiveness in applications such as chemical reactors, industrial robots and quadcopters. However, there is little prior theoretical work that explains the effectiveness of ILC even in the presence of large modeling errors, where optimal control methods using the misspecified model (MM) often perform poorly. Our work presents such a theoretical study of the performance of both ILC and MM on Linear Quadratic Regulator (LQR) problems with unknown transition dynamics. We show that the suboptimality gap, as measured with respect to the optimal LQR controller, for ILC is lower than that for MM by higher order terms that become significant in the regime of high modeling errors. A key part of our analysis is the perturbation bounds for the discrete Ricatti equation in the finite horizon setting, where the solution is not a fixed point and requires tracking the error using recursive bounds. We back our theoretical findings with empirical experiments on a toy linear dynamical system with an approximate model, a nonlinear inverted pendulum system with misspecified mass, and a nonlinear planar quadrotor system in the presence of wind. Experiments show that ILC outperforms MM significantly, in terms of the cost of computed trajectories, when modeling errors are high. | ['J. Andrew Bagnell', 'Maxim Likhachev', 'Wen Sun', 'Anirudh Vemula'] | 2021-11-17 | null | null | null | null | ['industrial-robots'] | ['robots'] | [ 9.59320143e-02 4.75961208e-01 -5.09459198e-01 8.23209584e-01
-6.03077292e-01 -7.53937364e-01 3.49206358e-01 2.72001952e-01
-1.47925377e-01 9.90947664e-01 -4.64994818e-01 -6.53350055e-01
-5.33725202e-01 -4.59073186e-02 -1.17249143e+00 -8.88094366e-01
-2.11985126e-01 1.74467087e-01 4.84014377e-02 -6.20880902e-01
-1.66201498e-02 2.67743468e-01 -9.04164195e-01 -7.10897923e-01
8.78565609e-01 7.11412787e-01 1.01240829e-01 7.60742962e-01
8.98084342e-01 6.11931205e-01 -2.68953800e-01 4.99842614e-01
7.09038854e-01 -3.96854997e-01 -2.62016147e-01 2.34979734e-01
1.97159380e-01 -4.74384055e-02 2.38955468e-02 1.15264952e+00
5.67483008e-01 5.65810978e-01 4.12705302e-01 -1.33545840e+00
-1.15319066e-01 3.06694239e-01 -6.13229036e-01 -7.52624273e-02
-3.42787728e-02 4.76741910e-01 6.53536439e-01 -6.51906490e-01
3.88219714e-01 1.13246334e+00 8.99129927e-01 3.62700045e-01
-1.18169701e+00 -3.98073316e-01 3.12440097e-01 -3.22596341e-01
-1.28967953e+00 -2.26606831e-01 1.52871802e-01 -8.45964670e-01
5.13748586e-01 6.13115206e-02 4.71288383e-01 3.86648118e-01
7.27774858e-01 1.75999641e-01 7.30072916e-01 -1.82771400e-01
2.87685186e-01 2.67192692e-01 3.08141354e-02 6.81132197e-01
7.29485929e-01 5.22243857e-01 1.30058408e-01 -5.03045395e-02
8.93224657e-01 -2.17335775e-01 -6.79931581e-01 -6.41797066e-01
-1.02691185e+00 1.05154502e+00 2.75294155e-01 1.73529070e-02
-4.12038326e-01 1.82145074e-01 1.85607150e-01 6.55228138e-01
3.85217696e-01 1.10109425e+00 -5.82284272e-01 4.22690921e-02
-2.74984062e-01 5.97642303e-01 1.00278950e+00 1.23047996e+00
1.12420537e-01 6.04155362e-01 1.22371003e-01 3.84095937e-01
1.50846643e-03 6.38995349e-01 1.57187238e-01 -1.38416982e+00
3.58940512e-01 2.49049351e-01 8.61101747e-01 -7.38186538e-01
-4.16072994e-01 -7.04902828e-01 -6.55473232e-01 4.68751878e-01
6.40841544e-01 -8.47403586e-01 -3.80391419e-01 1.75672424e+00
5.12471855e-01 -8.49387795e-02 3.20545346e-01 1.05596519e+00
-3.90138954e-01 9.49779689e-01 -5.82063198e-01 -1.01450825e+00
9.65022802e-01 -8.16623926e-01 -9.48265433e-01 -2.13098869e-01
6.83029652e-01 -5.64855039e-01 8.96012425e-01 5.05295753e-01
-1.12542343e+00 -2.37316847e-01 -1.21871734e+00 5.65216720e-01
9.13995057e-02 1.17192954e-01 -2.43551344e-01 1.80490553e-01
-7.46767402e-01 8.17698061e-01 -7.46412814e-01 -2.63970941e-01
-5.18540621e-01 3.64858091e-01 2.22964329e-03 3.82213503e-01
-1.18429303e+00 1.32326555e+00 4.14814025e-01 1.63247749e-01
-6.25406206e-01 -1.19262612e+00 -6.63770437e-01 -1.72886714e-01
1.18291140e+00 -5.22066712e-01 1.56210577e+00 -8.61959815e-01
-1.82032466e+00 -3.25669318e-01 1.57160774e-01 -5.89626193e-01
5.75995326e-01 -2.75157392e-01 3.63272965e-01 -3.71649070e-03
-1.21674426e-01 -6.76950291e-02 9.52398062e-01 -9.36139345e-01
-4.75075960e-01 -1.40142024e-01 2.85256356e-01 4.15933430e-01
1.16434954e-01 -5.35231352e-01 3.46437573e-01 -2.06534505e-01
-1.55122027e-01 -1.52508414e+00 -6.03278518e-01 -5.37002720e-02
-7.87724704e-02 -1.75091386e-01 8.45138669e-01 -4.52251881e-01
1.03751290e+00 -1.81520295e+00 4.03087407e-01 -1.14308493e-02
-3.41829240e-01 3.95469457e-01 2.70492315e-01 8.15266609e-01
-9.74793658e-02 9.40106288e-02 -7.05020279e-02 3.82417500e-01
4.37314680e-04 -9.18560177e-02 -6.39822960e-01 8.75954568e-01
1.37364939e-01 3.94670933e-01 -1.06959689e+00 1.35910749e-01
3.05114478e-01 9.73772183e-02 -4.27131474e-01 -1.15484692e-01
-3.76056373e-01 5.47720969e-01 -5.19132376e-01 1.13694027e-01
1.86286137e-01 -1.49706528e-01 1.76942676e-01 -2.37044226e-03
-8.04192066e-01 -3.66085500e-01 -1.39223218e+00 8.37103724e-01
-6.55368268e-01 6.57717288e-01 1.00733972e+00 -1.29710126e+00
4.62373406e-01 4.23317045e-01 5.20390511e-01 -1.78791091e-01
2.85668045e-01 3.28258574e-01 1.25098109e-01 -4.06133562e-01
7.91998729e-02 -5.12262404e-01 1.35678828e-01 -1.41402140e-01
-2.62721151e-01 -6.32598877e-01 2.24848494e-01 -1.49544636e-02
9.32424545e-01 -2.09248513e-01 6.50330603e-01 -7.03554273e-01
6.18778646e-01 4.06340510e-01 7.23718703e-01 4.18206424e-01
-1.54180750e-01 -3.23265567e-02 7.09433854e-01 1.73862383e-01
-1.17684281e+00 -1.36126071e-01 -1.77166462e-01 4.96010363e-01
4.00252461e-01 -1.05793782e-01 -3.65509659e-01 4.55613248e-02
4.12801117e-01 7.75210559e-01 -2.94914633e-01 -2.96490431e-01
-5.69974422e-01 -6.05597496e-01 -2.42707700e-01 2.80585140e-01
7.57950917e-03 -9.27094296e-02 -6.98739707e-01 3.65947455e-01
3.50086302e-01 -1.28180873e+00 -6.24549806e-01 1.22817002e-01
-7.61119664e-01 -1.42875493e+00 -5.39460540e-01 -4.95514482e-01
6.27233863e-01 -3.86142060e-02 4.05993670e-01 -2.58353949e-01
-1.72496974e-01 6.81050777e-01 7.54149482e-02 -8.08116674e-01
-5.67989826e-01 -2.26499453e-01 5.92558384e-01 -3.33231598e-01
-9.53844249e-01 2.54025459e-01 -2.74223924e-01 6.27256870e-01
-4.95219141e-01 -7.02092797e-02 2.33193472e-01 1.08644497e+00
5.09804428e-01 3.55840474e-01 7.41071582e-01 -5.18236518e-01
7.64245391e-01 -4.56851602e-01 -1.61261058e+00 -1.82163388e-01
-7.38315344e-01 -6.29170984e-02 1.13882744e+00 -7.80532479e-01
-6.51461124e-01 2.99161315e-01 4.33411926e-01 -6.76598608e-01
7.37970829e-01 4.99578178e-01 4.24730211e-01 -3.88055474e-01
6.22762918e-01 -2.59043425e-01 7.49026775e-01 -1.80931523e-01
1.79582551e-01 4.01730925e-01 1.44823611e-01 -4.82550830e-01
9.37282026e-01 7.10620657e-02 7.34227359e-01 -1.07889032e+00
-8.39785933e-01 -5.66765308e-01 -2.93815374e-01 -2.97835678e-01
4.22504455e-01 -9.33075964e-01 -1.15298939e+00 1.86988845e-01
-7.97607958e-01 -7.54401445e-01 -4.12467360e-01 7.15760529e-01
-9.88196433e-01 2.60135215e-02 -6.06243432e-01 -1.51420367e+00
-1.31461322e-01 -1.12748563e+00 7.13771045e-01 2.08117783e-01
1.27081990e-01 -1.28855324e+00 1.05508186e-01 -1.41809896e-01
4.04493153e-01 7.79198587e-01 6.82476580e-01 -1.27574027e-01
-2.98009723e-01 -2.93018162e-01 4.54737544e-01 4.97441441e-01
-2.24243954e-01 7.57985236e-03 -1.75904587e-01 -1.05845392e+00
3.73104095e-01 -2.65757293e-01 2.02356607e-01 5.89663804e-01
1.60899103e-01 -8.23936462e-01 -4.23601151e-01 7.42722005e-02
1.83450127e+00 4.73048896e-01 -3.50384146e-01 1.79549396e-01
5.71477950e-01 5.68085194e-01 1.03206301e+00 3.09616685e-01
-5.32108359e-03 6.64310932e-01 5.54196358e-01 6.12221472e-02
5.77392280e-01 8.18867832e-02 7.46576309e-01 6.44064963e-01
1.92338720e-01 -3.81410345e-02 -7.44993210e-01 6.45179749e-01
-2.05558443e+00 -5.80240190e-01 -1.81639224e-01 2.63794255e+00
7.02457070e-01 -1.23085357e-01 1.57807261e-01 7.59451315e-02
9.20857430e-01 -4.51188266e-01 -8.59836638e-01 -6.04002655e-01
2.85259008e-01 -3.14303339e-01 1.24409056e+00 8.55013251e-01
-1.05592239e+00 2.12734655e-01 5.94806767e+00 5.18869817e-01
-1.30459785e+00 -1.01382703e-01 1.18683353e-01 -3.31646018e-02
6.21493518e-01 6.95942566e-02 -9.44845021e-01 4.34403419e-01
1.27935004e+00 -7.62812257e-01 5.70693851e-01 1.00991285e+00
8.64597976e-01 -1.95922807e-01 -8.71631801e-01 4.01294708e-01
-2.45761573e-01 -9.77854550e-01 -7.95081615e-01 -1.07596619e-02
1.19362879e+00 -2.40342811e-01 2.00277165e-01 5.27994871e-01
2.47246459e-01 -5.16456485e-01 5.45506954e-01 2.62736231e-01
3.91198277e-01 -7.32681751e-01 6.50969803e-01 9.25427735e-01
-9.79789972e-01 -7.82332182e-01 -4.01366711e-01 -2.52762079e-01
2.57987648e-01 3.78521323e-01 -8.51516604e-01 6.44732714e-01
-1.04016647e-01 6.50547564e-01 1.10546604e-01 1.08939254e+00
-4.81665879e-02 6.24394536e-01 -6.39481008e-01 -2.31711328e-01
4.93766248e-01 -4.68823522e-01 1.12073207e+00 5.62889397e-01
3.18948358e-01 2.26619139e-01 6.02314830e-01 5.88043213e-01
4.38777089e-01 9.46860015e-02 -8.48918319e-01 -5.02794003e-03
1.71094179e-01 9.62018132e-01 -2.58372724e-01 -2.07133457e-01
-7.28887320e-02 1.29056707e-01 -1.96052372e-01 4.23327655e-01
-8.17401409e-01 -5.68785191e-01 6.44735098e-01 2.49800742e-01
2.19024852e-01 -5.89528620e-01 5.56720011e-02 -6.89916730e-01
3.87462080e-02 -7.52543986e-01 1.51546717e-01 -3.30074668e-01
-7.10521877e-01 -6.08677976e-02 2.29341000e-01 -1.41765189e+00
-7.12040961e-01 -4.69386160e-01 -4.99979109e-01 6.48948431e-01
-1.22657192e+00 -2.13713318e-01 2.98667431e-01 -8.00421312e-02
7.39519060e-01 2.39660919e-01 1.50215104e-01 4.02598120e-02
-9.17939723e-01 -1.90770999e-01 8.71180713e-01 -5.07335424e-01
6.22027397e-01 -1.33115816e+00 -3.69088531e-01 8.31650734e-01
-8.85116696e-01 6.24287367e-01 1.24343848e+00 -5.36001325e-01
-2.00893426e+00 -1.52814972e+00 2.31207117e-01 -2.30857760e-01
1.00033855e+00 -6.00018054e-02 -8.17411005e-01 7.00969815e-01
4.18690816e-02 8.53132159e-02 -3.38996351e-01 -5.20818889e-01
6.77366793e-01 -2.35235393e-01 -9.85037863e-01 5.90541303e-01
2.86626786e-01 3.75204114e-03 -2.91604668e-01 6.51034951e-01
1.08529603e+00 -6.73130453e-01 -1.07436585e+00 7.40735948e-01
3.77605289e-01 -2.33952906e-02 7.41058111e-01 -4.70950961e-01
-1.66054636e-01 -5.72946250e-01 2.89427489e-01 -1.83569598e+00
7.70319253e-02 -1.34050393e+00 -3.74013275e-01 5.47795415e-01
3.30320984e-01 -7.62264013e-01 2.20948562e-01 2.75181144e-01
-9.17855874e-02 -1.05410945e+00 -9.21861827e-01 -1.45207655e+00
4.96391207e-01 6.08303919e-02 -3.48810762e-01 8.27194810e-01
3.48283231e-01 5.81892848e-01 -2.82992363e-01 7.22694278e-01
5.80585420e-01 -1.34064183e-01 6.89496398e-01 -8.03048074e-01
-2.96544909e-01 -1.89483806e-01 4.48213853e-02 -8.20556343e-01
1.84733391e-01 -2.99471021e-01 4.32879716e-01 -1.39669383e+00
-6.63142025e-01 -3.31292748e-01 1.91642329e-01 -2.40790844e-02
-2.34592006e-01 -2.39924893e-01 3.59795719e-01 1.31682903e-01
-3.40476155e-01 4.59505439e-01 1.36416078e+00 -1.69822380e-01
-6.46651387e-01 5.12162566e-01 -1.69576555e-01 7.29093909e-01
7.70595014e-01 -1.58458620e-01 -9.19535160e-01 -6.87019974e-02
5.47102280e-02 6.75008476e-01 1.17358424e-01 -1.26092124e+00
2.81477928e-01 -4.67529595e-01 -4.20037180e-01 -2.16130883e-01
2.15468585e-01 -8.12397063e-01 3.12074333e-01 1.05292106e+00
-3.89413178e-01 5.16699143e-02 2.94735759e-01 8.54197204e-01
1.37667581e-02 -1.90626487e-01 1.23571110e+00 4.29615825e-02
-1.36688799e-01 1.64737001e-01 -6.35474384e-01 1.90149635e-01
1.30211902e+00 2.92697340e-01 -1.22711182e-01 -7.31690049e-01
-6.30161405e-01 8.14066231e-01 2.52926618e-01 2.46506691e-01
-5.82699627e-02 -8.14947665e-01 -3.85075897e-01 -1.19862847e-01
-5.10480404e-01 -8.38859379e-02 -1.45848289e-01 1.31947744e+00
-1.50577471e-01 8.55710566e-01 3.48852426e-01 -5.99252582e-01
-1.01600897e+00 7.16829181e-01 6.07647538e-01 -1.97758242e-01
-3.48250449e-01 2.10573524e-01 1.55371338e-01 -1.70465663e-01
1.28804579e-01 -7.92191207e-01 1.02051273e-01 -7.42829293e-02
1.67437211e-01 6.91273987e-01 6.77482113e-02 -2.65344262e-01
7.12468699e-02 8.18030834e-01 4.63594496e-01 -4.37807441e-02
8.59930038e-01 -1.07303336e-01 2.95359731e-01 6.24520779e-01
9.67563033e-01 -3.02007973e-01 -1.58626676e+00 1.81654710e-02
-9.03107226e-03 1.98144950e-02 1.28861934e-01 -4.02691811e-01
-7.16669440e-01 4.71022397e-01 3.95417064e-01 3.65242362e-01
6.27372980e-01 -6.42286062e-01 4.30722713e-01 6.70885444e-01
4.98744756e-01 -1.23936510e+00 -9.21743736e-03 1.04918432e+00
1.14864719e+00 -8.93976390e-01 2.38749042e-01 -4.49494630e-01
-6.55828297e-01 9.91193950e-01 7.13710785e-01 -7.86291122e-01
5.73125899e-01 3.27225804e-01 -2.62050331e-01 4.06550199e-01
-1.21558928e+00 -1.01999372e-01 2.59035942e-03 3.79151143e-02
8.55970532e-02 -1.12150267e-01 -7.08132327e-01 1.84790969e-01
3.04186791e-01 -1.89354539e-01 9.33404028e-01 1.15650952e+00
-7.18313158e-01 -5.21162212e-01 -6.39125407e-01 7.30596855e-02
-4.13204998e-01 4.77351487e-01 1.22751258e-01 1.34518480e+00
-1.85853168e-01 8.04770887e-01 -1.15228142e-03 2.16728166e-01
8.87369931e-01 -1.25274047e-01 3.54408801e-01 -6.76490963e-01
-4.31198269e-01 1.92039579e-01 9.09572542e-02 -5.46736121e-01
-5.23071140e-02 -4.52639520e-01 -1.14526880e+00 -1.97001770e-02
-1.03137982e+00 5.05898893e-01 6.24123454e-01 7.09117293e-01
1.82871133e-01 5.58007598e-01 8.04700613e-01 -7.82465816e-01
-1.35450828e+00 -7.78185785e-01 -4.25993502e-01 -2.99678355e-01
9.21254992e-01 -9.55241799e-01 -9.50927734e-01 -1.00887068e-01] | [5.027190685272217, 2.429328680038452] |
6a20f133-5500-47e8-ba2a-524f056f0357 | evaluation-of-three-deep-learning-models-for | null | null | https://www.mdpi.com/2072-4292/11/22/2673 | https://www.mdpi.com/2072-4292/11/22/2673/pdf | Evaluation of Three Deep Learning Models for Early Crop Classification Using Sentinel-1A Imagery Time Series—A Case Study in Zhanjiang, China | Timely and accurate estimation of the area and distribution of crops is vital for food security.
Optical remote sensing has been a key technique for acquiring crop area and conditions on regional
to global scales, but great challenges arise due to frequent cloudy days in southern China. This
makes optical remote sensing images usually unavailable. Synthetic aperture radar (SAR) could
bridge this gap since it is less a effected by clouds. The recent availability of Sentinel-1A (S1A) SAR
imagery with a 12-day revisit period at a high spatial resolution of about 10 m makes it possible to
fully utilize phenological information to improve early crop classification. In deep learning methods,
one-dimensional convolutional neural networks (1D CNNs), long short-term memory recurrent
neural networks (LSTM RNNs), and gated recurrent unit RNNs (GRU RNNs) have been shown
to efficiently extract temporal features for classification tasks. However, due to the complexity of
training, these three deep learning methods have been less used in early crop classification. In this
work, we attempted to combine them with an incremental classification method to avoid the need for
training optimal architectures and hyper-parameters for data from each time series. First, we trained
1D CNNs, LSTM RNNs, and GRU RNNs based on the full images’ time series to attain three classifiers
with optimal architectures and hyper-parameters. Then, starting at the first time point, we performed
an incremental classification process to train each classifier using all of the previous data, and obtained
a classification network with all parameter values (including the hyper-parameters) at each time
point. Finally, test accuracies of each time point were assessed for each crop type to determine the
optimal time series length. A case study was conducted in Suixi and Leizhou counties of Zhanjiang
City, China. To verify the effectiveness of this method, we also implemented the classic random forest
(RF) approach. The results were as follows: (i) 1D CNNs achieved the highest Kappa coefficient
(0.942) of the four classifiers, and the highest value (0.934) in the GRU RNNs time series was attained
earlier than with other classifiers; (ii) all three deep learning methods and the RF achieved F measures
above 0.900 before the end of growth seasons of banana, eucalyptus, second-season paddy rice, and
sugarcane; while, the 1D CNN classifier was the only one that could obtain an F-measure above 0.900
for pineapple before harvest. All results indicated the e effectiveness of the solution combining the
deep learning models with the incremental classification approach for early crop classification. This
method is expected to provide new perspectives for early mapping of croplands in cloudy areas. | ['Min Feng 6', 'Wenlong Jing', 'Hongwei Zhao', 'Liang Sun', 'Hao Jiang', 'Zhongxin Chen'] | 2019-11-15 | null | null | null | remote-sensing-issn-2072-4292-mdpi-2019-11 | ['crop-classification'] | ['miscellaneous'] | [ 2.13121116e-01 -5.20169973e-01 -2.70209134e-01 -1.12648174e-01
-4.40296829e-01 -6.31817997e-01 2.63687104e-01 3.68100256e-02
-2.99489617e-01 8.32559824e-01 -4.29540277e-01 -8.71986866e-01
-2.69235492e-01 -1.14203608e+00 -5.01820683e-01 -1.09670067e+00
-4.71833467e-01 -2.15312809e-01 -1.20103449e-01 -3.16661179e-01
-1.25439927e-01 8.25297594e-01 -1.72970688e+00 3.06407232e-02
1.15390396e+00 1.11177468e+00 8.16901326e-01 4.85968828e-01
-9.72847641e-03 4.45089042e-01 -4.98721898e-01 4.43068713e-01
2.12607563e-01 -3.43454808e-01 -4.62202966e-01 -7.84852542e-03
-2.20889553e-01 -5.15257359e-01 -3.48121114e-03 8.70583653e-01
4.08395946e-01 2.49376409e-02 3.35096449e-01 -7.47489035e-01
-6.74135029e-01 7.00020790e-01 -8.49082232e-01 -1.22781089e-02
-3.49807262e-01 4.46604043e-02 6.75718725e-01 -6.30702853e-01
2.27177188e-01 6.04386270e-01 6.68130875e-01 5.04247472e-03
-9.38247204e-01 -7.87896752e-01 3.51076096e-01 5.96544817e-02
-1.29026949e+00 2.25803629e-03 3.93862784e-01 -4.61034238e-01
9.18738484e-01 -8.57277121e-03 1.14063668e+00 7.00977683e-01
4.76973772e-01 5.24811745e-01 1.25356615e+00 -3.20963174e-01
1.24646790e-01 -2.82068491e-01 5.90723120e-02 3.33139360e-01
3.07925671e-01 5.05997658e-01 1.57816082e-01 1.96438283e-01
8.07745934e-01 4.71229911e-01 -2.41778791e-01 1.67153850e-01
-1.05144978e+00 1.00587916e+00 1.08717799e+00 7.72111297e-01
-8.79671991e-01 -2.49833241e-01 1.96906522e-01 3.66140783e-01
5.79144955e-01 5.13806880e-01 -8.41516316e-01 3.94536972e-01
-9.32676673e-01 -7.86675513e-02 3.27299803e-01 4.77399528e-01
7.54300952e-01 5.58251619e-01 1.67374119e-01 9.08089101e-01
-8.18884149e-02 9.15655851e-01 5.08091509e-01 -3.92883599e-01
5.35951778e-02 4.62903678e-01 1.11827210e-01 -9.79586005e-01
-5.72810173e-01 -6.01283669e-01 -1.37514758e+00 4.91684377e-02
1.97564989e-01 -5.12496114e-01 -1.27966738e+00 1.52762651e+00
-1.58939332e-01 -4.75757010e-02 3.50005627e-01 8.69286776e-01
4.15678620e-01 1.17246103e+00 2.12198690e-01 -5.98254740e-01
1.32183433e+00 -5.06917298e-01 -6.54085100e-01 -3.10293794e-01
8.01510572e-01 -6.22736335e-01 7.06892014e-01 1.12233713e-01
-2.68718481e-01 -6.54541612e-01 -1.17741740e+00 6.98692441e-01
-7.96532035e-01 6.11816525e-01 9.70865130e-01 2.22936302e-01
-8.99320245e-01 6.65701747e-01 -9.06163454e-01 -4.69752222e-01
2.19121397e-01 9.50049311e-02 -2.18290001e-01 -7.24393055e-02
-1.38505530e+00 8.57517302e-01 6.26086056e-01 1.12222862e+00
-8.18901420e-01 -4.28635359e-01 -6.62393808e-01 4.69716676e-02
1.34659708e-01 2.68866569e-02 7.80021727e-01 -1.15106475e+00
-1.39226532e+00 5.17832398e-01 1.13714635e-01 -3.83333206e-01
-3.15976769e-01 -2.34927326e-01 -5.69207907e-01 -2.00528055e-01
1.70167089e-01 5.59622765e-01 5.03539383e-01 -1.01609683e+00
-6.75028682e-01 -6.27833843e-01 -1.57798856e-01 1.57660507e-02
-1.69705361e-01 -4.23561819e-02 4.39336836e-01 -5.23168325e-01
5.89547992e-01 -1.10879862e+00 -4.72250640e-01 -4.22478199e-01
-5.38214818e-02 1.92883283e-01 8.25170577e-01 -9.80182409e-01
9.98860180e-01 -2.14831734e+00 -1.87200725e-01 8.04726407e-02
-3.12193155e-01 7.17348099e-01 -3.65754068e-01 1.70218915e-01
-2.52089620e-01 1.57610357e-01 -4.88801897e-01 7.27603197e-01
-6.54014111e-01 3.07378501e-01 -3.33273590e-01 2.37648934e-01
4.51107115e-01 7.70537734e-01 -8.42088103e-01 5.48797175e-02
2.94678062e-01 4.43692535e-01 1.90526217e-01 2.08224699e-01
-2.45658606e-01 2.23921046e-01 -3.97910714e-01 9.71324325e-01
9.79984581e-01 -7.55458102e-02 3.75843734e-01 -6.95952773e-02
-7.84065127e-01 -2.17805237e-01 -7.39202738e-01 1.13603079e+00
-5.28545737e-01 6.37200832e-01 -1.06917173e-01 -1.13345563e+00
1.45336199e+00 4.44323599e-01 4.91192967e-01 -8.11699510e-01
-1.56763420e-01 5.37124932e-01 5.63290045e-02 -4.62353379e-01
3.31312597e-01 1.24976292e-01 2.01183230e-01 1.44963607e-01
-2.60887802e-01 -5.34782559e-02 -1.34920165e-01 -6.18732929e-01
4.45362389e-01 3.04601252e-01 3.58713895e-01 -2.56406546e-01
2.89729625e-01 3.05020928e-01 6.55222595e-01 5.36258698e-01
-2.68484093e-03 3.52718413e-01 2.71281362e-01 -8.18315208e-01
-8.21435273e-01 -6.24345839e-01 -3.45102251e-01 9.40652013e-01
-2.00080395e-01 2.98187345e-01 -2.02094346e-01 -2.47911483e-01
-1.46838248e-01 7.46175885e-01 -5.08895814e-01 8.87519568e-02
-4.53954697e-01 -1.40695405e+00 4.29040194e-01 5.47651052e-01
9.54782367e-01 -1.35363793e+00 -8.69894028e-01 4.13869023e-01
-1.96855873e-01 -8.65389764e-01 2.78011709e-01 6.91388726e-01
-1.13079429e+00 -8.65173697e-01 -1.05919647e+00 -7.14028716e-01
5.62017202e-01 6.19670987e-01 6.64836049e-01 -1.88355446e-01
-1.71841323e-01 -5.51087737e-01 -5.98229408e-01 -5.91696143e-01
-1.72936440e-01 4.34997797e-01 -2.55503476e-01 -2.94282407e-01
4.62518930e-01 -4.92662698e-01 -3.37124854e-01 1.58086956e-01
-7.35970259e-01 -3.77007835e-02 9.96036172e-01 1.07073617e+00
5.76390505e-01 2.15331674e-01 7.10933030e-01 -5.40811896e-01
2.12857589e-01 -6.16701245e-01 -1.13042665e+00 3.76645297e-01
-6.57262027e-01 -2.07615286e-01 5.61920822e-01 -4.46451545e-01
-8.57066751e-01 2.45189130e-01 6.15967624e-02 -2.23732233e-01
-2.52063900e-01 1.27749789e+00 1.63240135e-01 2.22413719e-01
5.03840089e-01 3.15481871e-01 1.20818183e-01 -2.40339577e-01
-4.54078317e-02 7.98868060e-01 3.47810060e-01 -3.06093693e-02
5.61853409e-01 -3.95124927e-02 -1.24859206e-01 -1.19930983e+00
-9.12444353e-01 -1.71253547e-01 -8.33760262e-01 -3.11154008e-01
8.90481174e-01 -1.13972855e+00 -5.40384948e-01 1.01444054e+00
-1.07392573e+00 -4.31321949e-01 6.07038252e-02 1.03228259e+00
-1.30411973e-02 -1.58601940e-01 -3.72188896e-01 -9.84569490e-01
-5.17115653e-01 -8.13559353e-01 5.56535780e-01 5.16639471e-01
3.33911866e-01 -7.06292450e-01 -1.88263118e-01 -2.91219592e-01
7.50809669e-01 6.37831509e-01 1.02651191e+00 -2.07681045e-01
-3.01391929e-01 -2.47951761e-01 -5.41805804e-01 5.39395034e-01
5.58676004e-01 3.68892401e-01 -1.04524410e+00 -3.19327623e-01
-1.02439765e-02 -7.33858794e-02 9.44993615e-01 1.10817945e+00
1.01454687e+00 -4.24426571e-02 -2.21061885e-01 6.60618067e-01
1.66082263e+00 8.68670404e-01 6.57709956e-01 4.15795535e-01
4.21011597e-01 5.19642949e-01 8.00332427e-01 3.24058473e-01
-4.36088890e-02 1.78847983e-01 5.77688277e-01 -5.37129283e-01
4.67935175e-01 -1.99037027e-02 5.06421089e-01 7.29475439e-01
-4.73941624e-01 -4.58298922e-02 -9.51939821e-01 7.53860295e-01
-1.61036098e+00 -9.76821303e-01 -9.46769714e-02 2.32366085e+00
4.30434406e-01 -1.31650046e-01 -1.39082879e-01 2.17784345e-01
6.84270442e-01 4.21835601e-01 -7.00770974e-01 -3.79231095e-01
-5.07493019e-01 1.80849999e-01 9.78812277e-01 1.33046344e-01
-1.26682425e+00 1.11623991e+00 5.61279631e+00 2.50244886e-01
-1.68462133e+00 -2.98862398e-01 6.74313903e-01 3.25511336e-01
1.04245558e-01 1.58470258e-01 -6.65741980e-01 9.73210335e-02
1.07503128e+00 2.60967433e-01 4.46747750e-01 6.86064780e-01
4.32804346e-01 -3.46263289e-01 -3.42639923e-01 3.94231319e-01
-4.66869920e-01 -1.05280590e+00 -1.60683736e-01 8.89741629e-02
7.97552288e-01 3.58100474e-01 -1.64213389e-01 4.35800731e-01
3.00947785e-01 -1.03774321e+00 6.89706430e-02 4.78269637e-01
9.24455941e-01 -8.06000948e-01 1.24383569e+00 3.78025413e-01
-1.34468842e+00 -4.11442757e-01 -6.17303193e-01 -2.50885606e-01
-2.43305743e-01 9.00318503e-01 -5.95919251e-01 8.06964159e-01
7.80191898e-01 9.55166399e-01 -3.33446741e-01 6.56770110e-01
-1.03602119e-01 7.39388108e-01 -3.36430728e-01 -2.39504248e-01
6.50850117e-01 -4.06710207e-01 1.13448739e-01 8.25538158e-01
8.45929742e-01 3.07592750e-01 1.00199416e-01 6.19031072e-01
4.01894897e-01 -9.57205985e-03 -8.78347754e-01 -3.88009816e-01
5.10071516e-01 1.15555525e+00 -7.61908650e-01 2.24238634e-01
-2.60031581e-01 4.30263489e-01 -1.82081580e-01 5.96538603e-01
-5.77723503e-01 -6.33059978e-01 5.13162673e-01 -1.95684820e-01
3.55389655e-01 -3.68192494e-01 -8.11065957e-02 -9.19193625e-01
-1.04728222e-01 -7.25075305e-01 2.21751437e-01 -8.69695663e-01
-8.99489880e-01 7.88380861e-01 -3.58536020e-02 -1.24475288e+00
-4.25517738e-01 -6.93431973e-01 -4.72026855e-01 1.31132841e+00
-1.88772678e+00 -1.17695642e+00 -5.92482567e-01 1.47209346e-01
3.80277246e-01 -2.18836427e-01 1.31248474e+00 -4.25950773e-02
-7.52045631e-01 3.98534089e-02 4.15302694e-01 1.74457788e-01
3.26445162e-01 -7.29799032e-01 2.21238837e-01 1.01312923e+00
-3.68646264e-01 2.83575058e-01 2.63205230e-01 -6.10589623e-01
-1.09046364e+00 -1.31923962e+00 9.23919618e-01 5.54063380e-01
5.21624804e-01 1.58901453e-01 -9.43828106e-01 7.35443354e-01
-1.42333776e-01 4.07208689e-03 4.34740275e-01 2.26514086e-01
2.24574655e-03 -5.02548635e-01 -8.09647024e-01 2.51906872e-01
3.56965780e-01 -3.61616910e-01 9.97291729e-02 2.83364594e-01
6.61497414e-01 -1.26157925e-01 -9.39903378e-01 8.65107179e-01
7.64342070e-01 -7.03039587e-01 6.04466200e-01 -2.71048069e-01
4.24737424e-01 -2.61342555e-01 -4.50955868e-01 -1.49133658e+00
-5.10120749e-01 -1.35230780e-01 4.06391978e-01 1.01154542e+00
4.70242262e-01 -8.75620425e-01 3.20277691e-01 -1.62837759e-01
-5.25713861e-02 -7.85309792e-01 -5.06191015e-01 -7.70295858e-01
8.97240117e-02 -9.79090333e-02 7.58606195e-01 9.19338167e-01
-6.86818838e-01 3.24124396e-02 -3.16170454e-01 5.74152827e-01
1.84386075e-01 4.21166718e-01 3.46821338e-01 -1.41718233e+00
3.06383520e-01 -2.55181968e-01 -1.28081679e-01 -6.32182717e-01
3.12947407e-02 -6.06865108e-01 8.00787061e-02 -1.37246394e+00
-9.22853649e-02 -6.85060859e-01 -4.38466668e-01 8.40338945e-01
-2.11239308e-01 -2.57183641e-01 9.29159019e-03 3.18935990e-01
6.68538570e-01 3.49935085e-01 1.18769145e+00 -1.42290503e-01
-5.86839557e-01 5.60756505e-01 -5.08444071e-01 3.69654030e-01
1.03325665e+00 -2.73510098e-01 -2.86637127e-01 -6.43421412e-01
4.72682640e-02 4.31600451e-01 3.60339165e-01 -9.49258208e-01
-1.55633867e-01 -4.65771377e-01 6.96378827e-01 -1.10571265e+00
-5.95522225e-02 -6.70902371e-01 3.53848815e-01 6.65091157e-01
-9.39105749e-02 1.17825300e-01 5.35218477e-01 8.21509212e-03
-2.68739998e-01 -1.34905666e-01 7.44356513e-01 -1.41555622e-01
-9.29188311e-01 3.65355819e-01 -6.72372818e-01 -7.23035634e-01
9.49833751e-01 -2.19089776e-01 -1.76832736e-01 -8.72156173e-02
-5.21141887e-01 3.46409887e-01 -1.96446516e-02 4.29863125e-01
5.66935360e-01 -1.11876595e+00 -1.07074618e+00 5.54137230e-01
2.86804768e-03 3.46492678e-02 3.84534299e-01 6.33893013e-01
-7.21112192e-01 5.62439084e-01 -6.23486102e-01 -8.12844157e-01
-8.65821004e-01 3.67154628e-01 4.46324766e-01 -3.55386138e-01
-2.83565313e-01 5.27174830e-01 -2.03564495e-01 -5.76101422e-01
-2.11433589e-01 -5.17931104e-01 -5.86558700e-01 5.34555316e-01
2.70233870e-01 -1.71216801e-02 1.82239756e-01 -4.30582941e-01
-2.29269296e-01 5.20324707e-01 7.46902525e-02 1.51954472e-01
1.74251497e+00 1.88502252e-01 -1.95969760e-01 6.69475794e-01
1.02602625e+00 -6.39351130e-01 -1.04696894e+00 -1.61642849e-01
3.80187146e-02 -1.03844449e-01 4.42479014e-01 -8.18157971e-01
-1.33258390e+00 8.72776628e-01 9.71524119e-01 3.37067425e-01
1.61873949e+00 -5.90672493e-01 6.09861612e-01 3.88149291e-01
2.12941140e-01 -7.77445436e-01 -6.66789830e-01 7.44839251e-01
7.16505647e-01 -1.30937231e+00 -2.92376000e-02 2.77611855e-02
-4.07094836e-01 1.39173985e+00 4.07123834e-01 1.03833497e-01
8.23822379e-01 1.01006575e-01 3.26837033e-01 -8.46069306e-02
-6.31753504e-01 -4.19565976e-01 -1.26641408e-01 5.29669046e-01
6.18479490e-01 5.67116022e-01 -7.12179318e-02 2.62010247e-01
-5.14270887e-02 2.90227830e-01 3.57968032e-01 9.73400891e-01
-4.37154651e-01 -5.65677047e-01 -3.44920814e-01 6.34927869e-01
-1.54875264e-01 -6.80968538e-02 8.07286203e-02 9.42017913e-01
-7.10694641e-02 8.53473306e-01 2.01074392e-01 -4.81501400e-01
1.36401519e-01 -1.43362850e-01 4.13686186e-02 -3.49156529e-01
-3.99574041e-01 1.76578179e-01 -2.37153977e-01 -2.47785836e-01
-6.05313659e-01 -6.97416782e-01 -9.36804295e-01 -3.72267425e-01
-8.78824949e-01 1.96507871e-01 1.07007456e+00 8.57457280e-01
1.40361875e-01 4.39423978e-01 1.26774037e+00 -8.86456609e-01
-4.76134479e-01 -1.23571992e+00 -9.00428653e-01 -4.84232515e-01
4.03782487e-01 -3.65901530e-01 -2.71632999e-01 -2.22700685e-01] | [9.44820499420166, -1.708410620689392] |
f6a4b00a-b648-46a4-b6f0-9d2cd257bace | sentiment-analysis-and-opinion-mining-on | 2302.04359 | null | https://arxiv.org/abs/2302.04359v1 | https://arxiv.org/pdf/2302.04359v1.pdf | Sentiment analysis and opinion mining on educational data: A survey | Sentiment analysis AKA opinion mining is one of the most widely used NLP applications to identify human intentions from their reviews. In the education sector, opinion mining is used to listen to student opinions and enhance their learning-teaching practices pedagogically. With advancements in sentiment annotation techniques and AI methodologies, student comments can be labelled with their sentiment orientation without much human intervention. In this review article, (1) we consider the role of emotional analysis in education from four levels: document level, sentence level, entity level, and aspect level, (2) sentiment annotation techniques including lexicon-based and corpus-based approaches for unsupervised annotations are explored, (3) the role of AI in sentiment analysis with methodologies like machine learning, deep learning, and transformers are discussed, (4) the impact of sentiment analysis on educational procedures to enhance pedagogy, decision-making, and evaluation are presented. Educational institutions have been widely invested to build sentiment analysis tools and process their student feedback to draw their opinions and insights. Applications built on sentiment analysis of student feedback are reviewed in this study. Challenges in sentiment analysis like multi-polarity, polysemous, negation words, and opinion spam detection are explored and their trends in the research space are discussed. The future directions of sentiment analysis in education are discussed. | ['Linda Galligan', 'Yan Li', 'Haoran Xie', 'Christopher Dann', 'Xiaohui Tao', 'Thanveer Shaik'] | 2023-02-08 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [ 2.10073832e-02 3.36457431e-01 -5.08706987e-01 -4.27509695e-01
-1.01006940e-01 -5.45237422e-01 8.39262232e-02 9.56948757e-01
-4.10100847e-01 5.65025389e-01 5.23980677e-01 -6.30800128e-01
9.60386544e-02 -7.72347331e-01 -1.50616914e-01 -4.84086871e-01
7.59444535e-01 -7.90995583e-02 -1.35800410e-02 -9.74292874e-01
9.85879719e-01 4.51303124e-01 -1.93781495e+00 4.13673222e-01
1.00151980e+00 7.50846505e-01 -3.40417832e-01 6.87121511e-01
-6.37897134e-01 1.36923683e+00 -9.00487483e-01 -8.77521574e-01
-3.94584566e-01 -4.22283798e-01 -8.26183081e-01 2.11468056e-01
1.94705099e-01 2.36932546e-01 6.39650762e-01 1.27517498e+00
6.52361572e-01 3.05506766e-01 4.17239189e-01 -9.72851694e-01
-7.64404714e-01 4.76305693e-01 -5.80855787e-01 2.16971353e-01
6.46690071e-01 -5.93290269e-01 6.09270811e-01 -8.48227262e-01
4.45653617e-01 9.64427114e-01 5.09355962e-01 4.69498485e-01
-2.32126325e-01 -7.43113399e-01 3.82169694e-01 3.83106858e-01
-4.43832070e-01 1.41904950e-01 9.57802355e-01 -7.09129035e-01
1.00535190e+00 7.64331222e-02 1.47324705e+00 7.98859954e-01
6.06759071e-01 9.89873230e-01 1.52604055e+00 -1.02781713e+00
3.37973922e-01 1.07885551e+00 7.72788346e-01 6.68493450e-01
2.05160961e-01 -5.44270396e-01 -9.21570718e-01 1.91816330e-01
-3.16115350e-01 -2.40674034e-01 2.78325737e-01 2.63001025e-02
-3.33756715e-01 1.28870165e+00 -2.50531107e-01 2.92251587e-01
-2.57340997e-01 -5.23938000e-01 8.38470221e-01 5.24492204e-01
8.26418757e-01 7.15129375e-01 -7.91416705e-01 -5.32650232e-01
-4.98792678e-01 -4.65712734e-02 1.05007970e+00 6.53771222e-01
2.88437337e-01 2.70533234e-01 4.50995594e-01 9.25510943e-01
5.27796805e-01 5.35303533e-01 1.25560915e+00 -5.05698323e-01
-1.37969377e-02 1.22730994e+00 -4.66417879e-01 -1.20901489e+00
-2.62288868e-01 -1.06610455e-01 -2.25348175e-01 2.49318480e-01
-2.75119692e-01 -6.63815618e-01 -8.77295196e-01 8.75326753e-01
5.26798904e-01 -3.66137654e-01 4.17670101e-01 4.07380491e-01
1.36213291e+00 3.72864157e-01 3.35746616e-01 -2.78095484e-01
1.97896147e+00 -1.03965700e+00 -1.22549284e+00 -4.54247624e-01
1.01245725e+00 -1.28803575e+00 7.92347312e-01 8.09686601e-01
-1.18597674e+00 -3.92277837e-01 -8.91000032e-01 -4.10440229e-02
-1.22037303e+00 -5.32198064e-02 9.11981165e-01 1.34902215e+00
-9.00254965e-01 -9.68504045e-03 -4.86419201e-01 -3.97486687e-01
3.76046062e-01 6.30637109e-01 -1.84210286e-01 4.78721976e-01
-1.28353405e+00 1.25725031e+00 -2.18515530e-01 -1.01491027e-01
1.93209022e-01 -5.55577338e-01 -1.07508469e+00 -1.43219501e-01
-6.63030893e-02 -1.52728125e-01 1.34435534e+00 -1.54018140e+00
-1.93577194e+00 1.08088434e+00 -2.90815324e-01 3.37277986e-02
-2.40046188e-01 -2.68239260e-01 -5.26398063e-01 1.00887239e-01
1.31412998e-01 2.23099291e-01 3.77932161e-01 -6.78669155e-01
-9.70091701e-01 -4.44470823e-01 -1.41829386e-01 4.44599032e-01
-7.85673738e-01 5.97602904e-01 2.39696592e-01 -4.53075618e-01
1.10094383e-01 -7.71411657e-01 -2.01157540e-01 -9.03775573e-01
2.43110046e-01 -7.86737382e-01 1.06069434e+00 -2.82984495e-01
1.14066756e+00 -1.65078604e+00 -4.14480150e-01 4.26030934e-01
-2.86967605e-02 5.31151712e-01 5.31704426e-01 5.07101774e-01
-3.64125490e-01 3.16321999e-01 5.36563039e-01 7.21720532e-02
-1.18147053e-01 2.96486132e-02 -2.94533998e-01 4.19168361e-02
9.40923914e-02 5.38514078e-01 -1.14210904e+00 -5.59402645e-01
4.82699692e-01 3.90589774e-01 -5.31192958e-01 8.88704285e-02
2.33224988e-01 1.60080522e-01 -7.40810096e-01 8.09047878e-01
5.88802770e-02 5.87668456e-02 1.34764984e-01 1.94632277e-01
-5.12371540e-01 6.42165840e-01 -8.73703837e-01 8.79252017e-01
-7.27905512e-01 1.10678470e+00 -1.61857441e-01 -1.16778827e+00
1.13551617e+00 7.02873230e-01 2.47727081e-01 -6.13999605e-01
4.34732378e-01 2.07124621e-01 3.89258042e-02 -9.59828615e-01
8.80890787e-01 -4.33635354e-01 -1.56421542e-01 6.42119467e-01
2.14754581e-01 -6.12460375e-01 5.09161174e-01 1.92559496e-01
4.94168460e-01 2.04981472e-02 5.07785141e-01 -1.95827931e-01
9.61374521e-01 8.77268836e-02 2.47524425e-01 1.05908476e-01
-4.94959891e-01 -1.49548829e-01 8.53520572e-01 -5.52763820e-01
-3.16992104e-01 -2.03238404e-03 -2.17039213e-01 1.65751266e+00
-3.09467435e-01 -4.00351495e-01 -6.56977355e-01 -9.03716326e-01
-3.98389965e-01 7.13585198e-01 -5.53744614e-01 -1.39212564e-01
-1.52935222e-01 -7.55338907e-01 -9.29494277e-02 4.15373117e-01
6.59496784e-02 -1.55172729e+00 -5.61676025e-01 2.44715996e-02
1.92086741e-01 -7.53146768e-01 1.75923005e-01 6.49324894e-01
-6.43325984e-01 -9.03113246e-01 -1.68465048e-01 -1.22974527e+00
1.04518533e+00 1.51580721e-01 1.07475138e+00 1.37344003e-01
2.31561095e-01 9.04518127e-01 -7.16460347e-01 -1.51625800e+00
-5.40408432e-01 -5.06721102e-02 1.60063192e-01 -7.02043474e-01
1.44041502e+00 -3.24789554e-01 -5.02870917e-01 -1.31803915e-01
-7.28211641e-01 -3.70257914e-01 3.74646157e-01 5.82057774e-01
2.83345938e-01 2.36406982e-01 1.12724042e+00 -1.45465767e+00
1.15696597e+00 -4.25383240e-01 -2.65210986e-01 7.78057501e-02
-1.08829129e+00 -5.05008936e-01 4.65195566e-01 -2.35783935e-01
-1.30923665e+00 -4.53576922e-01 -3.93104225e-01 3.39359581e-01
-3.90153825e-01 8.37301254e-01 2.57211983e-01 -4.56838489e-01
5.51870942e-01 -1.72366410e-01 -5.04114665e-02 3.06300104e-01
2.18982697e-02 1.00218594e+00 -2.75035679e-01 -3.66754711e-01
7.48807788e-02 3.66770662e-02 -2.40284339e-01 -1.08672774e+00
-1.55246055e+00 -7.04608560e-01 -5.62604368e-01 -8.17248046e-01
8.64243448e-01 -9.83554065e-01 -8.84866059e-01 2.79176086e-01
-7.10772872e-01 1.76978424e-01 -5.23221135e-01 7.63933539e-01
-3.60691473e-02 1.73627347e-01 -6.70587361e-01 -1.08531940e+00
-5.92067480e-01 -1.32844007e+00 4.50193554e-01 1.07517290e+00
-7.23539710e-01 -1.47007775e+00 2.24281013e-01 1.29367971e+00
1.03102639e-01 -7.25201517e-02 7.21953511e-01 -1.15282464e+00
5.50545871e-01 -5.44897497e-01 6.13670707e-01 8.76578212e-01
-1.31143294e-02 3.64210576e-01 -9.58062351e-01 2.73133337e-01
3.32433701e-01 -6.99403107e-01 2.38611251e-01 1.82971999e-01
8.11394036e-01 -3.26153547e-01 1.80234984e-01 -2.06597731e-01
1.12346959e+00 6.63996339e-01 3.75101954e-01 7.87342548e-01
5.20439327e-01 1.27950048e+00 7.17819571e-01 1.60507873e-01
5.03391922e-01 -4.41504210e-01 9.10467505e-02 1.50024176e-01
2.87232906e-01 -3.04421270e-03 8.18369746e-01 1.83174896e+00
8.26762915e-02 -9.91894305e-02 -8.30380797e-01 8.68288040e-01
-1.54497945e+00 -8.79621089e-01 -5.07807910e-01 1.48491263e+00
9.16479886e-01 3.17063600e-01 -1.64661825e-01 2.83092141e-01
4.03186113e-01 -2.03760173e-02 -1.20274395e-01 -1.66599238e+00
-1.64032169e-02 8.85864973e-01 1.78347155e-01 3.46591294e-01
-9.63463366e-01 1.20591962e+00 5.77954388e+00 5.98120511e-01
-1.25912035e+00 -1.44111723e-01 7.57499158e-01 2.46169001e-01
-3.51423740e-01 -2.24813968e-01 -1.01392627e+00 2.24550188e-01
1.04473996e+00 -1.07813187e-01 -3.82789940e-01 1.23694134e+00
2.39876643e-01 -4.56879616e-01 -1.44002452e-01 2.15777606e-01
5.41582346e-01 -1.09631181e+00 -2.26740569e-01 -2.66860038e-01
1.38314676e+00 -2.08401710e-01 1.75634935e-01 5.95518351e-01
5.81574678e-01 -7.40729451e-01 3.24414045e-01 3.38386260e-02
-1.64879605e-01 -1.16070342e+00 1.40761364e+00 -2.33808175e-01
-6.72481954e-01 9.97654796e-02 -1.61548227e-01 -6.26837075e-01
-2.86836416e-01 4.49185073e-01 -7.84810007e-01 1.01775005e-01
7.97057390e-01 8.68853569e-01 -5.78563035e-01 3.56032550e-01
-9.61896718e-01 9.22103226e-01 1.23372741e-01 -7.77750969e-01
3.32655966e-01 -6.44881308e-01 2.14948550e-01 1.25738764e+00
4.14583571e-02 3.70769531e-01 -1.77468508e-02 -1.47960499e-01
9.34092551e-02 8.12475801e-01 -5.55048645e-01 -2.10661873e-01
1.73724383e-01 1.81405866e+00 -1.27788234e+00 -2.87237465e-01
-7.82804132e-01 3.52518260e-01 -1.09377034e-01 6.29932135e-02
-1.93546534e-01 -5.28322518e-01 5.80381393e-01 -1.16021037e-01
1.89080074e-01 5.56013167e-01 -5.69438875e-01 -8.79397571e-01
-3.70884657e-01 -1.31719315e+00 3.80090892e-01 -7.99246311e-01
-1.17622828e+00 1.99252740e-01 -4.81782407e-01 -8.69566143e-01
-8.29944015e-02 -9.30885673e-01 -9.77322757e-01 5.03590822e-01
-1.39911199e+00 -7.76087224e-01 1.15961470e-01 3.27458560e-01
5.87731004e-01 -4.28564668e-01 9.76722181e-01 1.25397602e-02
-6.73069119e-01 2.15708479e-01 -2.29561344e-01 1.08153425e-01
8.22490931e-01 -1.57699215e+00 -5.48413932e-01 5.28569877e-01
-2.30142862e-01 6.50680482e-01 7.50190496e-01 -6.66277528e-01
-1.24565661e+00 -5.03879607e-01 1.44670975e+00 -4.72059101e-01
1.03837490e+00 2.31223360e-01 -5.91274738e-01 4.53980505e-01
8.18809927e-01 -6.42674029e-01 1.91857433e+00 3.74647647e-01
2.71538526e-01 7.68218189e-02 -1.10735047e+00 6.82749152e-01
-1.80218324e-01 -4.36035037e-01 -9.80736732e-01 6.06381476e-01
3.09823871e-01 -5.78397512e-01 -1.10001135e+00 2.34126430e-02
6.27077758e-01 -9.73133087e-01 4.45026636e-01 -7.95758367e-01
8.74574661e-01 2.49281731e-02 4.81380790e-01 -1.24330056e+00
4.04741317e-01 -4.29988027e-01 8.80683511e-02 1.13434076e+00
5.19277573e-01 -5.05162179e-01 1.14784253e+00 7.59035468e-01
-3.26295942e-01 -1.19587564e+00 -2.73262709e-01 4.15221751e-01
2.67505735e-01 -6.79991663e-01 1.18380696e-01 1.32409155e+00
9.38790619e-01 7.60723591e-01 2.10152790e-01 -2.33830780e-01
-9.88599733e-02 6.53779209e-02 5.46410203e-01 -1.21729577e+00
5.99504113e-01 -6.93578422e-01 -4.97863591e-01 -3.11904848e-01
4.28703815e-01 -4.82733876e-01 -2.89929032e-01 -1.66203439e+00
-2.61417300e-01 2.17158824e-01 -2.13344708e-01 4.44432318e-01
-2.88688362e-01 1.15307175e-01 -2.31340647e-01 -5.15736938e-01
-6.73824966e-01 1.69279233e-01 1.40885723e+00 2.06617936e-01
-2.92682856e-01 2.32280195e-01 -1.41005170e+00 1.51614892e+00
9.88404989e-01 -4.24127936e-01 -4.29027498e-01 3.42444450e-01
1.22184229e+00 -4.07479465e-01 -5.98024130e-01 -4.93245006e-01
4.50654238e-01 -3.99475783e-01 4.74415690e-01 -7.18532205e-01
-1.21184960e-01 -9.05053556e-01 -9.09240246e-01 2.23075047e-01
-4.34564263e-01 3.63883048e-01 3.43395203e-01 -2.44479831e-02
-6.54189646e-01 -8.70655179e-01 5.65944731e-01 -2.22113490e-01
-6.38352871e-01 -4.98589545e-01 -1.26071799e+00 1.37007209e-02
1.25882304e+00 -2.88787484e-01 -2.49806553e-01 -6.75533831e-01
-6.22427583e-01 4.48796481e-01 -4.58774157e-03 5.52184224e-01
6.67135358e-01 -7.67164052e-01 -1.16762467e-01 2.37421334e-01
-2.75534969e-02 4.70457561e-02 2.74838388e-01 8.58560801e-01
-6.16013467e-01 5.28271973e-01 -5.20144887e-02 -8.90407339e-02
-1.60530186e+00 1.65636629e-01 -1.83648672e-02 -4.61979508e-01
1.88297108e-01 1.19548464e+00 -3.16569626e-01 -6.37565374e-01
2.88246721e-01 -4.52319048e-02 -1.39071596e+00 9.37725782e-01
5.58881581e-01 3.61324698e-01 4.24619824e-01 -7.13609517e-01
-9.26851481e-02 5.59214473e-01 -2.73084670e-01 3.85898165e-02
1.29855192e+00 -6.35498688e-02 -4.04242963e-01 6.40304267e-01
7.70196378e-01 6.47804320e-01 -3.43681611e-02 3.99771363e-01
1.46663681e-01 1.23986796e-01 3.14882547e-01 -1.09262919e+00
-1.14139712e+00 7.84118891e-01 4.38370675e-01 5.79130054e-01
1.13120270e+00 -4.23750699e-01 5.72182059e-01 3.65931153e-01
-3.93156052e-01 -1.89781165e+00 1.78589255e-01 8.87654781e-01
1.86152622e-01 -1.56575751e+00 3.63278061e-01 -1.48822978e-01
-1.05993319e+00 1.46287274e+00 9.79380250e-01 6.79022670e-02
9.90311801e-01 2.04135343e-01 6.89066291e-01 -6.85406804e-01
-7.56252944e-01 -1.54681057e-01 4.91463035e-01 2.50167131e-01
1.34595692e+00 -7.10440800e-02 -7.47825384e-01 6.81658804e-01
-7.59230912e-01 -2.46641800e-01 1.01849878e+00 1.42833006e+00
-5.96650243e-01 -1.42465997e+00 -5.52179396e-01 5.70740938e-01
-1.32289016e+00 -2.40198329e-01 -9.55274165e-01 5.78981459e-01
2.62009859e-01 1.39357114e+00 -1.47716716e-01 -2.32895687e-01
4.66386139e-01 4.32715654e-01 -1.74598154e-02 -9.86039817e-01
-1.55109441e+00 -8.45348611e-02 1.88227922e-01 5.31293452e-02
-1.11220574e+00 -5.81763446e-01 -1.20099115e+00 -2.17230454e-01
-7.67654300e-01 9.09449160e-01 1.27103674e+00 1.18778181e+00
7.48250857e-02 8.07488620e-01 3.51339161e-01 -8.06120485e-02
1.30082741e-01 -8.95281553e-01 -2.68227607e-01 -5.79326116e-02
-7.83818308e-03 -2.58290172e-01 -5.94689548e-01 1.14849508e-02] | [11.190085411071777, 6.881440162658691] |
ab9a5489-7c64-4cdf-9762-4072f16e3751 | evaluating-capability-of-deep-neural-networks | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Hao_Cheng_Evaluating_Capability_of_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Hao_Cheng_Evaluating_Capability_of_ECCV_2018_paper.pdf | Evaluating Capability of Deep Neural Networks for Image Classification via Information Plane | Inspired by the pioneering work of information bottleneck principle for Deep Neural Networks (DNNs) analysis, we design an information plane based framework to evaluate the capability of DNNs for image classification tasks, which not only helps understand the capability of DNNs, but also helps us choose a neural network which leads to higher classification accuracy more efficiently. Further, with experiments, the relationship among the model accuracy, I(X;T) and I(T;Y) are analyzed, where I(X;T) and I(T;Y) are the mutual information of DNN's output T with input X and label Y. We also show the information plane is more informative than loss curve and apply mutual information to infer the model's capability for recognizing objects of each class. Our studies would facilitate a better understanding of DNNs. | ['Shenghua Gao', 'Dongze Lian', 'Hao Cheng', 'Yanlin Geng'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['information-plane'] | ['methodology'] | [-6.00874051e-02 4.52198200e-02 -3.43595952e-01 -4.40795273e-01
9.55138281e-02 -4.70064491e-01 3.94994438e-01 -5.28820269e-02
-5.75397193e-01 5.88692725e-01 5.77653088e-02 -3.65116060e-01
-6.16678238e-01 -9.48211372e-01 -5.39656341e-01 -8.61431003e-01
-2.10006446e-01 1.57074317e-01 1.51379555e-01 2.17889741e-01
2.58941054e-01 5.75275779e-01 -1.65712178e+00 1.24753147e-01
7.49895632e-01 1.36313081e+00 3.04050416e-01 5.74192107e-01
-2.55458802e-01 9.65863109e-01 -4.98345047e-01 -3.96883428e-01
2.91858673e-01 -4.53525960e-01 -9.67354298e-01 -4.15105492e-01
-1.35776550e-01 -3.81748885e-01 -6.12361789e-01 1.34069550e+00
4.42798762e-03 -1.30281141e-02 1.00089860e+00 -1.54097629e+00
-3.98458064e-01 7.24576592e-01 -3.70485485e-01 5.68658292e-01
-3.23710591e-01 2.50145886e-02 1.08487141e+00 -2.51112223e-01
3.58491570e-01 1.28039205e+00 3.22080761e-01 6.57120287e-01
-1.05388665e+00 -8.62033606e-01 2.59742558e-01 1.97319329e-01
-9.96212661e-01 -1.85708195e-01 5.74905097e-01 -4.82082993e-01
5.10493457e-01 8.18665624e-02 6.82222426e-01 5.45120299e-01
2.47217521e-01 1.41141033e+00 7.46638775e-01 -3.98217231e-01
2.68124551e-01 2.84459144e-01 5.66561759e-01 5.97451627e-01
2.60363400e-01 2.62672752e-01 -3.64046842e-01 3.71326149e-01
9.55237687e-01 3.92908931e-01 -3.10129225e-01 -1.99937984e-01
-1.03155911e+00 8.04236293e-01 6.44248843e-01 5.20394623e-01
-2.71049410e-01 2.50834972e-01 3.26677024e-01 5.91613114e-01
1.77865878e-01 2.92968065e-01 -7.27547407e-01 -3.60644259e-03
-4.88926709e-01 -1.97494086e-02 5.98859370e-01 8.58718991e-01
8.41816783e-01 -4.64109182e-02 -6.38707206e-02 6.77141786e-01
4.01639253e-01 1.78573310e-01 4.59138840e-01 -1.14181066e+00
5.33445299e-01 8.20724308e-01 -4.66467321e-01 -6.32388830e-01
-5.20345151e-01 -5.02346337e-01 -9.73151445e-01 1.54695481e-01
7.38212705e-01 -2.61810988e-01 -7.74044991e-01 2.05176234e+00
-2.85226494e-01 -2.42820516e-01 3.13479602e-01 7.92606294e-01
7.38088548e-01 5.46054184e-01 -2.79112235e-02 -2.34551132e-01
1.08541560e+00 -6.24203920e-01 -6.74391031e-01 -7.77681544e-02
8.90192211e-01 -3.53718758e-01 6.41181111e-01 1.87706649e-01
-1.00578022e+00 -4.63297904e-01 -8.65083575e-01 1.73046723e-01
-2.80737251e-01 -3.12970169e-02 8.95125508e-01 4.70328033e-01
-1.12709212e+00 7.66095161e-01 -8.82721245e-01 -2.84491777e-01
8.90101552e-01 5.56703031e-01 -2.01383382e-01 7.51750602e-04
-1.20970297e+00 9.58833039e-01 5.01734138e-01 -1.63415462e-01
-1.12420559e+00 -6.23157620e-01 -5.32988191e-01 1.57992482e-01
1.92057192e-01 -7.90351868e-01 1.09504914e+00 -1.25101137e+00
-1.09745014e+00 5.41889489e-01 -8.86541419e-03 -5.29229283e-01
4.30446476e-01 3.05792671e-02 -9.30769518e-02 8.24448764e-02
-8.39120075e-02 1.04259109e+00 3.69039416e-01 -1.06307495e+00
-6.79520130e-01 -8.66347194e-01 2.06100978e-02 2.64726281e-01
-5.76622665e-01 -7.89445564e-02 -9.67968106e-02 -3.11870813e-01
5.02574980e-01 -5.22904992e-01 -1.89885527e-01 3.63629103e-01
-5.01755476e-01 -4.27143961e-01 8.02899003e-01 -3.10684949e-01
8.65641594e-01 -2.17960262e+00 5.04730362e-03 1.96357861e-01
6.63243771e-01 2.77209371e-01 -1.02543816e-01 1.66424140e-01
-7.85701349e-02 1.71201006e-01 -1.22115731e-01 1.64275318e-01
-1.93227798e-01 3.08966070e-01 -1.06099442e-01 1.90949559e-01
1.24192886e-01 5.82276821e-01 -6.26846790e-01 -4.17190462e-01
1.16387032e-01 3.82810801e-01 -5.03374159e-01 1.94273844e-01
9.73403454e-02 2.41579130e-01 -6.51845932e-01 1.98778316e-01
5.10111153e-01 -2.73494035e-01 -3.98927815e-02 -2.50677973e-01
-2.18431547e-01 2.01057658e-01 -6.56783104e-01 1.09405971e+00
-2.56662965e-01 1.10155666e+00 -3.78809571e-01 -1.23074305e+00
1.01971555e+00 1.79771066e-01 4.42653894e-01 -8.56458068e-01
3.72286826e-01 -1.04586087e-01 5.53891599e-01 -4.38072026e-01
-3.20878208e-01 -1.40630931e-01 4.81248647e-01 5.70935369e-01
3.42628211e-01 5.35974681e-01 1.18538119e-01 2.48420313e-01
1.10647535e+00 -3.65438223e-01 3.38069983e-02 -5.16947746e-01
4.13763881e-01 -4.70582753e-01 5.99291742e-01 8.63806844e-01
-6.81885958e-01 1.31443322e-01 1.16380036e+00 -5.87387145e-01
-1.06298089e+00 -1.16143978e+00 -4.67216313e-01 7.80702412e-01
3.24333787e-01 1.60792470e-01 -7.61517882e-01 -7.59010673e-01
-2.59335656e-02 7.40006030e-01 -7.26705909e-01 -3.71705711e-01
-2.54874509e-02 -6.99407518e-01 6.16656959e-01 7.27821231e-01
9.60007191e-01 -9.58333075e-01 -5.60393274e-01 -2.67491013e-01
-1.14383362e-01 -7.03241229e-01 -1.89683020e-01 3.17973316e-01
-1.47338641e+00 -1.04611659e+00 -7.88242400e-01 -7.86297441e-01
9.31630313e-01 4.93105724e-02 9.26436305e-01 1.35047629e-01
-1.76059663e-01 1.43521875e-01 -8.01461190e-02 -4.12517399e-01
-2.28398487e-01 -5.19361608e-02 1.95363700e-01 -2.69344926e-01
3.98220748e-01 -6.24626637e-01 -7.76084900e-01 5.75336754e-01
-9.92615998e-01 2.55979925e-01 7.07541108e-01 5.45555770e-01
1.46994203e-01 5.91982424e-01 4.43420947e-01 -7.00015187e-01
4.49409664e-01 -2.65999615e-01 -7.46241212e-01 3.31449121e-01
-8.51092339e-01 4.66262370e-01 4.89039958e-01 -3.11551362e-01
-1.00584209e+00 -1.73875511e-01 1.03776949e-02 -3.44439894e-01
1.10483589e-02 2.37869889e-01 -2.80020952e-01 3.08028072e-01
2.87091523e-01 3.72283071e-01 1.06535360e-01 -5.07814109e-01
2.37734139e-01 4.28942144e-01 2.08820567e-01 -1.61327943e-01
2.34746233e-01 2.55000561e-01 8.29924420e-02 -5.97633839e-01
-7.98955500e-01 -1.23346128e-01 -7.95804799e-01 -4.21161711e-01
9.34146106e-01 -6.66471839e-01 -1.53725839e+00 5.37570894e-01
-1.15978479e+00 -1.14623472e-01 3.69882509e-02 7.95430958e-01
-4.04737890e-01 -5.26252650e-02 -6.68421507e-01 -1.23571789e+00
-1.42243147e-01 -1.17845130e+00 4.55464631e-01 3.80468398e-01
2.57621072e-02 -1.36799943e+00 -2.67148703e-01 2.46882901e-01
1.58848479e-01 -5.36747456e-01 1.28085554e+00 -8.37815046e-01
-7.14287639e-01 1.20340586e-02 -8.28029573e-01 6.69248343e-01
-1.05507813e-01 -1.05219446e-01 -1.09850299e+00 -9.66694579e-02
2.65007287e-01 -2.63287812e-01 1.15349662e+00 8.75010252e-01
1.31165731e+00 -6.71768725e-01 -5.40862501e-01 6.33621395e-01
1.58461809e+00 5.60796082e-01 8.87888670e-01 2.32920036e-01
5.33373892e-01 8.52488637e-01 1.79410726e-01 2.77815610e-01
2.30230644e-01 2.78603047e-01 6.00319207e-01 2.69895941e-02
1.04916222e-01 -2.36458927e-01 2.55062431e-01 9.25411999e-01
-1.28588527e-01 -3.76528561e-01 -8.04606915e-01 3.21483552e-01
-1.76901960e+00 -9.06980097e-01 2.05752961e-02 2.23246431e+00
5.67897439e-01 2.71933883e-01 6.02638461e-02 2.57696301e-01
7.55527973e-01 -7.41824955e-02 -8.11698437e-01 -5.46461605e-02
-6.63613155e-03 -3.34386945e-01 6.81174219e-01 3.45264971e-01
-6.73609555e-01 5.73780715e-01 7.45770168e+00 8.76543641e-01
-8.59674454e-01 -2.35682338e-01 1.16511142e+00 2.09563091e-01
-3.27822208e-01 -8.42900109e-03 -7.72876501e-01 6.14548206e-01
6.81542158e-01 -1.30752787e-01 3.29244226e-01 7.79199600e-01
-6.92678541e-02 -2.26555705e-01 -1.50242054e+00 9.55736220e-01
-4.40842718e-01 -1.17434120e+00 2.30454609e-01 3.18706423e-01
5.54713190e-01 1.89054921e-01 9.07456651e-02 1.26413047e-01
6.19264424e-01 -8.36128533e-01 3.13511729e-01 6.99625611e-01
3.72417182e-01 -8.91761959e-01 9.15328503e-01 7.37816870e-01
-6.94072425e-01 -2.50508249e-01 -5.87384999e-01 -2.94369876e-01
-4.04977947e-01 9.07586873e-01 -8.36004734e-01 5.54974005e-02
6.02064013e-01 7.61357009e-01 -2.46733770e-01 9.25585568e-01
-2.06629768e-01 3.54737282e-01 -1.29612640e-01 -3.37190717e-01
1.31440014e-01 -3.00952375e-01 3.88999939e-01 6.81623101e-01
1.40434146e-01 1.48711845e-01 -2.51468539e-01 1.29721880e+00
-2.33973160e-01 -5.78228608e-02 -7.77079165e-01 -1.97192222e-01
5.16097486e-01 9.64937270e-01 -9.93333101e-01 -2.21146032e-01
-1.22558400e-01 6.43360972e-01 2.26255029e-01 3.45832855e-01
-4.89773035e-01 -4.66063321e-01 8.42580914e-01 -2.18697488e-01
-1.31049499e-01 1.34426802e-01 -5.01166642e-01 -9.82940853e-01
-2.15541050e-01 -1.94991753e-01 3.75485063e-01 -7.49370694e-01
-1.21918368e+00 6.40050292e-01 -1.67893082e-01 -1.15764832e+00
5.37386052e-02 -8.76041353e-01 -5.32024026e-01 8.75279665e-01
-1.26485574e+00 -3.58291686e-01 2.03002766e-01 5.46565175e-01
3.60870212e-01 -4.73343968e-01 5.84541500e-01 3.36406946e-01
-8.52465391e-01 6.85026705e-01 3.19506288e-01 6.45801187e-01
-1.75788924e-01 -9.95570362e-01 1.09621361e-01 5.38283348e-01
-3.91232446e-02 5.36889255e-01 4.21285301e-01 -3.87252808e-01
-1.06505692e+00 -7.32886195e-01 7.70394385e-01 -3.55431080e-01
5.41705132e-01 7.92795047e-03 -5.77298224e-01 6.60622180e-01
-9.57936645e-02 -3.36048871e-01 5.63544035e-01 2.69749761e-01
-2.20034257e-01 -5.12978375e-01 -1.05273128e+00 6.32575333e-01
9.59359229e-01 -4.48908716e-01 -3.08025151e-01 9.50921550e-02
7.63959110e-01 3.00237685e-01 -8.05331886e-01 3.69121611e-01
7.06659138e-01 -1.26490474e+00 7.57199824e-01 -7.80080259e-01
6.68833137e-01 1.93011940e-01 -3.32105719e-02 -1.00816691e+00
-2.18707904e-01 3.28696817e-02 3.35705519e-01 1.04895270e+00
7.11538315e-01 -7.53048360e-01 9.36282516e-01 7.69687951e-01
1.16716579e-01 -7.13442028e-01 -1.03230357e+00 -8.02254021e-01
7.11356923e-02 -5.77979922e-01 3.43332410e-01 8.68647873e-01
-2.76463628e-02 3.97174329e-01 2.07976904e-02 -6.46341071e-02
6.77142441e-01 -2.91082740e-01 1.48982495e-01 -1.49359810e+00
1.61571205e-01 -9.28678453e-01 -8.41672659e-01 -1.10203648e+00
-1.02939885e-02 -9.19054091e-01 -1.93475753e-01 -1.38861442e+00
7.47124016e-01 -6.40717626e-01 -6.16528392e-01 5.10004103e-01
2.15496331e-01 -1.56416580e-01 2.10276753e-01 5.83184719e-01
-6.34226441e-01 4.39330190e-01 1.49096787e+00 -4.99379560e-02
-7.23102167e-02 3.46426755e-01 -8.16096306e-01 9.58966255e-01
6.46385550e-01 -6.34386241e-01 -6.29951000e-01 -6.25737071e-01
2.63938636e-01 2.96782285e-01 1.60932034e-01 -1.03046322e+00
2.53331751e-01 7.14349076e-02 6.95888996e-01 -5.34414470e-01
4.53043059e-02 -8.84242713e-01 -2.65065402e-01 9.56525564e-01
-9.58424091e-01 -1.02873430e-01 -1.25599220e-01 3.80601108e-01
-1.88599482e-01 -5.07419288e-01 8.35198820e-01 -2.54488945e-01
-3.15582186e-01 5.51806808e-01 -3.65141422e-01 -1.67201996e-01
8.80576015e-01 -1.60059065e-01 -3.93501908e-01 -5.41333437e-01
-6.32658839e-01 3.44275326e-01 1.28689155e-01 1.65871114e-01
6.71801209e-01 -1.41712487e+00 -3.43544692e-01 3.54319215e-01
-1.18537530e-01 -1.11134931e-01 1.55550569e-01 7.17413962e-01
-5.04164636e-01 5.07837653e-01 -4.50003594e-01 -7.58557022e-01
-9.07509327e-01 3.87692094e-01 4.56445932e-01 -2.35788390e-01
-2.89101213e-01 1.04881859e+00 8.21860969e-01 -1.84853211e-01
5.37850082e-01 -3.01549762e-01 -4.60687339e-01 9.54510197e-02
6.30014598e-01 5.29833555e-01 -2.41421059e-01 -2.37825692e-01
-2.75797546e-01 3.72694790e-01 -3.81409675e-01 -2.17731208e-01
1.21905684e+00 -1.32803947e-01 -2.16764793e-01 7.00556397e-01
1.61564076e+00 -1.07079637e+00 -1.37724495e+00 -3.63195211e-01
1.53918698e-01 -4.57726955e-01 3.42213720e-01 -7.71065474e-01
-1.49668396e+00 1.24948549e+00 8.25643718e-01 4.48921740e-01
1.23241079e+00 1.98058456e-01 4.52158302e-01 3.91555190e-01
1.83501080e-01 -1.00154901e+00 2.56664723e-01 4.75901127e-01
3.42300475e-01 -1.22533393e+00 -2.85501331e-01 -1.37301564e-01
-5.71974337e-01 1.28455555e+00 5.14947057e-01 2.28113428e-01
8.50210011e-01 1.95679724e-01 1.09829776e-01 -2.31045023e-01
-9.87393022e-01 -1.75280944e-01 1.47396594e-01 6.46789253e-01
4.38387930e-01 -2.01630164e-02 -2.56201215e-02 4.70549464e-01
-8.25569928e-02 5.33243231e-02 1.47004694e-01 4.96036947e-01
-7.23778605e-01 -8.56453419e-01 1.25562504e-01 6.95167243e-01
-3.12477648e-01 -6.86653331e-02 -2.17667252e-01 6.30632937e-01
5.37075214e-02 6.12742007e-01 4.23332304e-01 -5.83186507e-01
2.32818779e-02 -1.09832250e-01 5.13362885e-01 -2.18641445e-01
-3.93390842e-03 -2.00027287e-01 -3.10638040e-01 -2.31344655e-01
-3.69217902e-01 -6.07467175e-01 -1.23959494e+00 -4.89223331e-01
-6.74825251e-01 1.49237746e-02 9.01906729e-01 1.17001438e+00
-6.61111176e-02 6.03445351e-01 8.88736844e-01 -3.33086073e-01
-3.63277972e-01 -7.82372534e-01 -8.65994334e-01 1.90151438e-01
2.70786971e-01 -5.02724588e-01 -4.39175993e-01 -2.20612049e-01] | [8.006011009216309, 3.551642894744873] |
9e944aa7-358a-4d18-9498-bf9b7ab66b13 | unleashing-the-potential-of-unsupervised-pre | 2112.00317 | null | https://arxiv.org/abs/2112.00317v1 | https://arxiv.org/pdf/2112.00317v1.pdf | Unleashing the Potential of Unsupervised Pre-Training with Intra-Identity Regularization for Person Re-Identification | Existing person re-identification (ReID) methods typically directly load the pre-trained ImageNet weights for initialization. However, as a fine-grained classification task, ReID is more challenging and exists a large domain gap between ImageNet classification. Inspired by the great success of self-supervised representation learning with contrastive objectives, in this paper, we design an Unsupervised Pre-training framework for ReID based on the contrastive learning (CL) pipeline, dubbed UP-ReID. During the pre-training, we attempt to address two critical issues for learning fine-grained ReID features: (1) the augmentations in CL pipeline may distort the discriminative clues in person images. (2) the fine-grained local features of person images are not fully-explored. Therefore, we introduce an intra-identity (I$^2$-)regularization in the UP-ReID, which is instantiated as two constraints coming from global image aspect and local patch aspect: a global consistency is enforced between augmented and original person images to increase robustness to augmentation, while an intrinsic contrastive constraint among local patches of each image is employed to fully explore the local discriminative clues. Extensive experiments on multiple popular Re-ID datasets, including PersonX, Market1501, CUHK03, and MSMT17, demonstrate that our UP-ReID pre-trained model can significantly benefit the downstream ReID fine-tuning and achieve state-of-the-art performance. Codes and models will be released to https://github.com/Frost-Yang-99/UP-ReID. | ['Feng Zhao', 'Kecheng Zheng', 'Xin Jin', 'Zizheng Yang'] | 2021-12-01 | null | null | null | null | ['unsupervised-pre-training'] | ['methodology'] | [-8.08905363e-02 -3.37519884e-01 6.76547270e-03 -7.04946756e-01
-3.68779689e-01 -4.77923512e-01 6.66857302e-01 -1.38874084e-01
-7.07168996e-01 5.67311585e-01 3.76033306e-01 1.74126774e-01
-1.12663344e-01 -6.39158189e-01 -5.87783098e-01 -6.16460681e-01
1.11657329e-01 3.07617962e-01 -2.44541079e-01 -2.55187899e-01
9.13403332e-02 3.46461773e-01 -1.57616448e+00 -1.65091995e-02
9.35704291e-01 6.68252826e-01 1.43478047e-02 2.43224531e-01
1.07116513e-02 2.71472573e-01 -5.40832579e-01 -7.36702323e-01
6.17793381e-01 -3.61534059e-01 -8.77613842e-01 6.28266186e-02
8.19107592e-01 -3.37635577e-01 -6.03313386e-01 1.30861139e+00
6.90030038e-01 2.21972361e-01 6.76597595e-01 -1.26209080e+00
-1.09478772e+00 4.35598493e-01 -9.11659122e-01 4.55772907e-01
3.50567967e-01 3.58863026e-01 6.18227780e-01 -9.82126594e-01
4.33645278e-01 1.42074370e+00 7.74951041e-01 7.45873749e-01
-1.24274635e+00 -1.20969021e+00 4.07169133e-01 5.23184240e-01
-1.86916244e+00 -4.08112913e-01 7.14831948e-01 -4.44763184e-01
4.99011755e-01 2.32231230e-01 3.68637502e-01 1.23390639e+00
-3.30399930e-01 6.54065669e-01 1.21565032e+00 -2.18032166e-01
-3.02032650e-01 1.50862575e-01 3.40528697e-01 5.32559514e-01
3.19806308e-01 2.78190464e-01 -4.81913209e-01 -4.86164168e-02
8.92880857e-01 1.69804066e-01 -1.22289613e-01 -1.95559740e-01
-1.07141793e+00 6.08565450e-01 7.37297297e-01 1.61242455e-01
-1.78091630e-01 -9.20928270e-02 4.15266365e-01 2.60247767e-01
3.71242970e-01 2.66352326e-01 -1.94115296e-01 1.83392048e-01
-6.84624016e-01 4.63808924e-01 2.48537913e-01 9.41930294e-01
1.09427273e+00 -2.29703292e-01 -3.84610057e-01 1.12725258e+00
9.68769118e-02 3.51277351e-01 5.75321674e-01 -4.16397244e-01
4.41315740e-01 6.77120864e-01 -8.12294856e-02 -8.96445692e-01
-3.48027170e-01 -7.09793568e-01 -1.17963278e+00 -1.79529414e-01
4.83969152e-01 4.69406433e-02 -9.91793692e-01 1.87933290e+00
4.30368692e-01 3.84286135e-01 -2.18710214e-01 1.12724507e+00
1.03042555e+00 2.15705425e-01 4.46813852e-01 2.60523528e-01
1.58440530e+00 -1.00829959e+00 -3.73703420e-01 -3.30791205e-01
2.40568995e-01 -6.33720219e-01 9.44907665e-01 -1.65071696e-01
-9.22189891e-01 -1.10917234e+00 -9.08971369e-01 -1.68098882e-01
-5.68723798e-01 2.35815555e-01 4.55849618e-01 5.02501130e-01
-1.02379441e+00 2.91488469e-01 -2.94815838e-01 -5.38389027e-01
4.50471163e-01 5.21190107e-01 -6.86447263e-01 -4.14069295e-01
-1.18964803e+00 5.98811567e-01 3.62514943e-01 1.87412053e-01
-6.48604929e-01 -8.50968242e-01 -1.00816464e+00 5.73163666e-02
1.35133520e-01 -9.27006304e-01 7.98367321e-01 -8.68235290e-01
-1.07052445e+00 1.38951492e+00 -3.15050155e-01 -1.09023988e-01
7.45087802e-01 -2.83271968e-02 -4.68589574e-01 -6.79253638e-02
5.62242091e-01 9.96585846e-01 8.63425970e-01 -1.43429267e+00
-5.51414490e-01 -5.10604739e-01 -3.22500616e-02 2.07962647e-01
-4.05765533e-01 1.84692070e-01 -8.14306498e-01 -1.08138800e+00
-8.63713548e-02 -9.68275130e-01 -6.04350194e-02 -2.90504098e-01
-4.80046153e-01 -4.94295835e-01 5.12941658e-01 -6.38817430e-01
9.15402114e-01 -2.15516114e+00 -1.71761706e-01 2.52430439e-01
1.99083909e-01 2.77933151e-01 -5.20386040e-01 9.70975012e-02
-4.21199292e-01 5.90782091e-02 -1.84014644e-02 -7.18248010e-01
2.62911916e-02 4.38053198e-02 -6.07580245e-02 6.22621059e-01
2.27824464e-01 1.05459225e+00 -8.60406518e-01 -4.85025316e-01
2.25576878e-01 5.05474448e-01 -4.24972117e-01 3.11671257e-01
4.82156485e-01 8.57509315e-01 -4.25780624e-01 6.42135739e-01
1.07255721e+00 -1.76110461e-01 -1.14572302e-01 -3.47325176e-01
-1.01683788e-01 3.04109175e-02 -1.17104828e+00 1.64720297e+00
3.91706601e-02 9.59459916e-02 -1.02134921e-01 -1.00039697e+00
8.89129758e-01 -1.36768937e-01 3.26043904e-01 -9.65598822e-01
1.48935974e-01 -1.44457504e-01 -1.35640725e-01 -2.58418441e-01
5.76498449e-01 1.36475846e-01 -1.21473722e-01 2.91591287e-01
1.35471344e-01 7.00752795e-01 2.83107758e-01 1.29555017e-01
6.85855746e-01 8.78593847e-02 3.58377434e-02 -5.45700550e-01
7.80626237e-01 -1.08156368e-01 9.09835458e-01 9.04733181e-01
-4.38345075e-01 8.07053983e-01 4.33750525e-02 -5.54316819e-01
-9.94698882e-01 -9.90317643e-01 -4.11711544e-01 1.30564177e+00
7.01912165e-01 -3.85213047e-01 -7.67646074e-01 -7.57999957e-01
2.05888242e-01 1.47130013e-01 -8.94551396e-01 -1.10763289e-01
-7.53008783e-01 -9.19765413e-01 6.05450332e-01 6.85300648e-01
9.87142205e-01 -8.74681771e-01 3.65144521e-01 5.89424185e-03
-2.34137490e-01 -1.09376311e+00 -1.10582507e+00 -4.23144430e-01
-3.30181152e-01 -1.12695038e+00 -1.06145251e+00 -9.09026444e-01
1.07829475e+00 6.99203014e-01 9.48444664e-01 3.75075579e-01
-4.72101241e-01 4.07871902e-01 -3.07348728e-01 -1.58690289e-01
8.88600275e-02 8.27155113e-02 4.14786220e-01 2.31568992e-01
6.16465211e-01 -5.17764628e-01 -9.50280309e-01 7.70205557e-01
-6.27054751e-01 -3.93505767e-03 6.06040299e-01 1.02878714e+00
6.04046047e-01 -1.26889683e-02 6.34862721e-01 -6.89250171e-01
4.68416303e-01 -4.55468565e-01 -3.36645454e-01 2.45649561e-01
-6.29604578e-01 -1.72422245e-01 4.11468118e-01 -5.37250757e-01
-1.18700254e+00 -1.23547085e-01 -1.49605349e-01 -3.97991508e-01
-3.72046709e-01 1.78592980e-01 -3.79256219e-01 -2.83338219e-01
6.23484969e-01 3.54740322e-01 -1.21713728e-01 -6.42358959e-01
2.91272700e-01 5.05865097e-01 9.80570674e-01 -8.43732774e-01
1.25143635e+00 5.02562225e-01 -4.23572391e-01 -4.93915200e-01
-8.49854827e-01 -8.26770842e-01 -8.18804085e-01 -1.21667599e-02
8.77772987e-01 -1.32681012e+00 -8.84642780e-01 6.75117970e-01
-7.92539299e-01 -1.99488461e-01 -1.11875601e-01 3.67876619e-01
-2.76885759e-02 5.62246263e-01 -6.71796262e-01 -3.52319598e-01
-4.58236873e-01 -1.11153710e+00 8.74382019e-01 6.85799479e-01
-1.54578909e-01 -7.37381101e-01 -8.84967372e-02 6.77204907e-01
3.92617196e-01 -8.66293684e-02 5.42268455e-01 -5.44007838e-01
-3.67880404e-01 -1.81699932e-01 -7.54450679e-01 2.59348929e-01
2.35291958e-01 -6.13801897e-01 -1.12786329e+00 -6.59063160e-01
-5.09641707e-01 -2.80529141e-01 1.01198232e+00 4.72929366e-02
1.42826247e+00 -2.22670838e-01 -3.95234972e-01 1.02696609e+00
1.13559496e+00 -2.25880846e-01 6.17750168e-01 6.38956666e-01
1.02851880e+00 5.61393976e-01 4.00916636e-01 3.26262027e-01
9.16660011e-01 7.17993200e-01 -2.39685420e-02 -2.57280827e-01
-4.15333122e-01 -5.56271970e-01 7.96807557e-02 3.74452919e-01
-3.52783501e-01 9.20753330e-02 -7.19991922e-01 5.98310709e-01
-1.87411058e+00 -9.78317380e-01 1.38486296e-01 2.24014163e+00
8.63514483e-01 -2.42779240e-01 3.21246624e-01 -2.32153565e-01
1.12460923e+00 6.30618408e-02 -6.94292963e-01 2.15853348e-01
-2.88724899e-01 -3.43705923e-03 5.67325830e-01 2.73003966e-01
-1.37688601e+00 1.01190794e+00 5.27269459e+00 8.56492460e-01
-9.39765215e-01 1.75039798e-01 6.06421530e-01 6.27126619e-02
2.93068797e-03 -2.08708718e-01 -1.17618179e+00 7.37855554e-01
2.51862347e-01 -1.50121182e-01 3.77078712e-01 8.03417623e-01
-9.51954946e-02 2.60459185e-01 -1.02127647e+00 1.52437377e+00
2.13980272e-01 -1.06911588e+00 4.28732373e-02 1.24252453e-01
7.96168387e-01 -1.03014141e-01 1.89338118e-01 6.31703794e-01
2.28316113e-01 -1.03798068e+00 6.81747973e-01 3.17794412e-01
1.00787294e+00 -8.18104386e-01 8.18216503e-01 5.99956624e-02
-1.43644583e+00 -4.53090779e-02 -5.60381472e-01 1.03893124e-01
-5.08059338e-02 2.64113665e-01 -4.50002164e-01 6.75100744e-01
1.15922773e+00 7.79574931e-01 -8.97363842e-01 1.07995450e+00
-3.26134324e-01 2.17993617e-01 -6.62818253e-02 5.51898897e-01
-8.41467734e-03 -1.46547258e-01 3.44318509e-01 1.39518499e+00
-6.38966337e-02 1.88377827e-01 4.38348502e-01 1.02613926e+00
-2.45670691e-01 -1.94098815e-01 -7.65908062e-02 4.91849184e-01
4.65124011e-01 1.33256733e+00 -3.19172114e-01 -3.24568868e-01
-3.47734451e-01 1.23606455e+00 4.29700464e-01 6.01561427e-01
-7.38336802e-01 -9.01999772e-02 1.10085368e+00 2.33409464e-01
1.79280579e-01 -1.07818365e-01 -1.66211754e-01 -1.39466798e+00
2.06327736e-02 -8.39621067e-01 7.54871190e-01 -3.66849035e-01
-1.96672118e+00 5.37572384e-01 1.25710502e-01 -1.11145115e+00
-6.56937733e-02 -3.26776832e-01 -7.43596435e-01 1.24646878e+00
-1.82458842e+00 -1.62443686e+00 -7.77018070e-01 1.00359190e+00
3.61927748e-01 -2.57085204e-01 6.48021758e-01 6.38736427e-01
-9.49776649e-01 1.42093587e+00 -2.14135334e-01 6.83221519e-01
1.17409909e+00 -1.06985068e+00 5.70675671e-01 9.99071360e-01
-2.82735437e-01 1.06170678e+00 3.76524538e-01 -7.30341852e-01
-1.24281871e+00 -1.27589321e+00 6.23903334e-01 -3.30044776e-01
3.31729442e-01 -4.75041658e-01 -9.89968956e-01 7.96026587e-01
4.13532406e-02 3.51633549e-01 7.29761243e-01 3.34360182e-01
-8.06443274e-01 -1.47163123e-01 -1.27712607e+00 4.74878520e-01
1.44677341e+00 -5.91531277e-01 -4.95964468e-01 1.60428971e-01
2.62977898e-01 -3.61003786e-01 -8.70077372e-01 3.91852558e-01
5.01769066e-01 -7.09996045e-01 1.41079938e+00 -4.33034271e-01
-5.90784177e-02 -4.54095393e-01 2.11720839e-01 -1.20863378e+00
-8.42452288e-01 -4.33772862e-01 3.30831200e-01 1.74544537e+00
-1.48388281e-01 -9.21461821e-01 7.93954015e-01 7.65112102e-01
3.61735970e-02 -2.83600748e-01 -7.73866713e-01 -9.04875398e-01
1.28760472e-01 7.01109096e-02 8.73562813e-01 1.26321697e+00
-2.18991384e-01 1.79427490e-01 -3.78127545e-01 4.48019832e-01
1.09589112e+00 -8.34536329e-02 1.01554382e+00 -1.17020750e+00
-7.66249299e-02 -5.41055679e-01 -5.38526356e-01 -1.13332796e+00
3.73983890e-01 -1.02242601e+00 -1.34157002e-01 -1.09721231e+00
7.68437088e-01 -9.74647045e-01 -5.31161487e-01 7.16098487e-01
-7.44294465e-01 5.47603309e-01 2.37719193e-01 6.25718832e-01
-4.98212725e-01 5.65176129e-01 1.17960227e+00 -4.17209029e-01
-6.78514019e-02 -2.25628525e-01 -1.02220571e+00 4.56101567e-01
7.45337844e-01 -2.90248781e-01 -1.93658113e-01 -5.30463099e-01
-3.03102404e-01 -6.41917050e-01 7.87364662e-01 -8.11693072e-01
3.95377100e-01 1.07348897e-01 9.48220134e-01 -4.82491195e-01
2.64108062e-01 -3.15052629e-01 8.35868251e-03 1.28619760e-01
-1.43559381e-01 1.41653642e-01 1.75331250e-01 3.43987733e-01
-2.38604352e-01 7.63569698e-02 9.11431253e-01 -3.05342674e-01
-1.06266737e+00 7.74827838e-01 2.97365040e-01 -2.54846551e-02
7.61949420e-01 -3.16212535e-01 -5.36842167e-01 -9.00934264e-02
-4.43121940e-01 5.67071080e-01 6.77809596e-01 6.64336681e-01
4.51068640e-01 -1.42730725e+00 -9.55412984e-01 4.86765265e-01
3.66382152e-01 -3.39103751e-02 7.50185966e-01 6.32294655e-01
-2.69565769e-02 1.75762087e-01 -3.88710946e-01 -5.30976534e-01
-1.25314498e+00 6.74644589e-01 3.45555186e-01 -1.73021674e-01
-7.17866063e-01 1.04453397e+00 8.30506444e-01 -6.03863597e-01
2.63186872e-01 3.41951311e-01 -2.80777365e-01 -1.09534979e-01
9.81549501e-01 2.78079391e-01 -1.60252169e-01 -9.90248561e-01
-5.17542481e-01 7.88442910e-01 -5.90954840e-01 3.38676035e-01
1.17442787e+00 -3.11733603e-01 -1.28787711e-01 -4.22239393e-01
1.23955011e+00 -1.02967285e-01 -1.32338274e+00 -5.33555925e-01
-2.43382052e-01 -6.66529953e-01 -2.53201425e-01 -7.13362694e-01
-9.71066952e-01 4.66005892e-01 9.04888332e-01 -2.04525143e-01
1.01215792e+00 9.91895795e-02 7.76625574e-01 6.96637407e-02
3.75148565e-01 -9.41273987e-01 3.82489501e-03 4.12663460e-01
8.63342643e-01 -1.51027417e+00 9.33378295e-04 -5.68769336e-01
-5.05131662e-01 7.47193217e-01 9.54252183e-01 -8.63161683e-02
3.80516380e-01 -1.20883428e-01 1.39899567e-01 -3.01010329e-02
7.09177405e-02 -4.36933845e-01 5.49766004e-01 6.69001758e-01
1.74366727e-01 9.75535624e-03 -1.81167528e-01 8.06652725e-01
-3.83438349e-01 -1.58745080e-01 3.72943282e-02 6.44913793e-01
6.39344081e-02 -1.29281127e+00 -6.84983552e-01 1.93445027e-01
-3.42080176e-01 -2.41186842e-02 -2.72702396e-01 7.53791809e-01
6.25944912e-01 8.41487825e-01 1.35143608e-01 -4.75958616e-01
4.28716600e-01 -2.63117701e-01 4.86477256e-01 -4.39715087e-01
-6.72197282e-01 -1.99901789e-01 -1.28130883e-01 -3.55220437e-01
-3.39439809e-01 -7.00769067e-01 -1.01321149e+00 -6.15720391e-01
-3.76684666e-02 3.83429639e-02 3.81644756e-01 9.03778672e-01
4.75142866e-01 2.12993965e-01 5.53473473e-01 -1.04952884e+00
-3.66527051e-01 -1.08180368e+00 -4.54754233e-01 8.88311267e-01
1.76602677e-01 -8.63908708e-01 -2.39454865e-01 1.09312385e-02] | [14.708296775817871, 0.9282689094543457] |
c2f389d9-ac56-486d-abd9-e54a4dfbf2ae | ynu-hpcc-at-semeval-2022-task-5-multi-modal | null | null | https://aclanthology.org/2022.semeval-1.104 | https://aclanthology.org/2022.semeval-1.104.pdf | YNU-HPCC at SemEval-2022 Task 5: Multi-Modal and Multi-label Emotion Classification Based on LXMERT | This paper describes our system used in the SemEval-2022 Task5 Multimedia Automatic Misogyny Identification (MAMI). This task is to use the provided text-image pairs to classify emotions. In this paper, We propose a multi-label emotion classification model based on pre-trained LXMERT. We use Faster-RCNN to extract visual representation and utilize LXMERT’s cross-attention for multi-modal alignment. Then we use the Bilinear-interaction layer to fuse these features. Our experimental results surpass the F_1 score of baseline. For Sub-task A, our F_1 score is 0.662 and Sub-task B’s F_1 score is 0.633. The code of this study is available on GitHub. | ['Xuejie Zhang', 'Jin Wang', 'Chao Han'] | null | null | null | null | semeval-naacl-2022-7 | ['emotion-classification', 'emotion-classification'] | ['computer-vision', 'natural-language-processing'] | [ 1.85394548e-02 -1.11756235e-01 -6.99514300e-02 -6.09582722e-01
-1.07916152e+00 -4.92152661e-01 4.55973834e-01 -1.73493087e-01
-8.11337531e-01 4.93685275e-01 2.08190039e-01 1.48385137e-01
6.13973498e-01 -5.67348980e-05 -5.43569446e-01 -4.88463432e-01
5.46385944e-01 1.57339647e-01 -4.56289560e-01 -6.52434751e-02
2.26242676e-01 1.11790061e-01 -1.44286811e+00 7.60733545e-01
5.98433554e-01 1.43403447e+00 -7.59861842e-02 8.65709066e-01
2.26060711e-02 6.60062194e-01 -4.69935775e-01 -5.87076962e-01
-1.81288287e-01 -3.21218342e-01 -9.43610609e-01 -2.55985200e-01
7.30926812e-01 -2.20614970e-01 -8.87062699e-02 1.13571727e+00
6.67425096e-01 3.00162166e-01 1.08661318e+00 -1.43995261e+00
-5.45136213e-01 4.21498775e-01 -9.14326549e-01 -8.43296573e-02
2.27469042e-01 -1.98995218e-01 9.83759701e-01 -1.31990612e+00
5.86257458e-01 1.32791495e+00 4.27994519e-01 9.15768623e-01
-7.77391553e-01 -1.09555495e+00 3.39730941e-02 5.94506383e-01
-1.40021050e+00 -3.95281732e-01 6.23875618e-01 -1.74502671e-01
1.13237584e+00 2.17806235e-01 3.15531373e-01 1.55401468e+00
2.98034579e-01 1.16472125e+00 1.45851529e+00 -5.33862412e-01
-3.38867694e-01 2.22363025e-01 2.98932493e-01 6.84484601e-01
-5.11968732e-01 -5.40327430e-02 -6.56114042e-01 3.58592749e-01
3.12563539e-01 -3.53494108e-01 5.20188510e-02 3.53452832e-01
-1.10723007e+00 7.09427953e-01 2.69775808e-01 5.10898769e-01
-1.95706040e-01 2.79994369e-01 8.30783725e-01 3.46603334e-01
5.04989088e-01 2.25761741e-01 -3.89385879e-01 -3.04565936e-01
-9.10601258e-01 -3.23460847e-02 2.97126681e-01 9.00766253e-01
3.81652713e-01 4.05992121e-02 -3.44283700e-01 1.27833664e+00
2.63024062e-01 4.50870991e-01 7.34865963e-01 -8.26037169e-01
2.16463506e-01 3.16938937e-01 -2.35919982e-01 -8.83175015e-01
-6.35613382e-01 -1.41463771e-01 -1.01217830e+00 3.14037859e-01
1.31602108e-01 -2.48614803e-01 -9.36890960e-01 1.79128397e+00
-2.34310925e-02 5.32004200e-02 1.25166804e-01 1.01474369e+00
1.26089442e+00 6.96096718e-01 3.65862548e-01 -4.00512405e-02
1.57993114e+00 -1.51593447e+00 -7.98056185e-01 -1.67651519e-01
7.24937975e-01 -9.44989085e-01 1.18563664e+00 4.35487837e-01
-9.94574010e-01 -7.28456914e-01 -9.45480704e-01 -2.37036541e-01
-6.16071939e-01 7.55097926e-01 1.68053925e-01 3.60469878e-01
-1.19337952e+00 1.87343627e-01 -4.21463013e-01 -4.82996792e-01
1.76304847e-01 4.15333956e-01 -5.29403806e-01 1.09613352e-01
-1.18702960e+00 1.14338195e+00 5.53793013e-01 8.68232921e-03
-8.59150767e-01 -3.09268057e-01 -9.52397048e-01 -1.45481601e-01
2.50680238e-01 -3.95864815e-01 1.11482537e+00 -1.22954297e+00
-1.36868262e+00 1.39846528e+00 -3.15659195e-01 -5.08333407e-02
2.89149880e-01 -4.33602720e-01 -6.36062324e-01 2.71114081e-01
7.74422660e-02 1.23322916e+00 7.42626369e-01 -1.11751413e+00
-6.11549020e-01 -3.59887809e-01 -2.03373358e-01 5.74275494e-01
-4.42978501e-01 6.03996277e-01 -7.84012377e-01 -6.44189954e-01
-4.00705665e-01 -1.12466061e+00 1.24053638e-02 -3.50830913e-01
-5.81246018e-01 -3.21132421e-01 8.06085169e-01 -1.02173507e+00
1.14422262e+00 -2.17132163e+00 2.72537500e-01 3.11838295e-02
1.34466559e-01 6.77473396e-02 -4.17743117e-01 -4.19615135e-02
-4.83801901e-01 1.04103737e-01 1.08735174e-01 -8.37590277e-01
1.37251303e-01 -3.79703045e-01 9.78610814e-02 3.13343585e-01
1.53456599e-01 1.08830082e+00 -4.11058486e-01 -6.09606743e-01
4.06505495e-01 5.08194029e-01 -2.90440500e-01 3.09672743e-01
4.82124835e-02 4.65926111e-01 4.59137484e-02 7.48158813e-01
6.01895094e-01 -1.71179265e-01 -1.61798224e-01 -6.43045664e-01
1.05566747e-01 -3.50692928e-01 -7.95267522e-01 1.93144751e+00
-7.77721107e-01 6.96773231e-01 7.02085868e-02 -7.30369747e-01
7.19896257e-01 4.16574150e-01 4.22786057e-01 -7.14093506e-01
7.36365795e-01 2.92659476e-02 -2.90530145e-01 -5.07564247e-01
5.31395435e-01 -7.49884024e-02 -4.77950126e-01 2.39785597e-01
4.12900120e-01 1.87953547e-01 -8.22224561e-03 1.20222926e-01
7.01765299e-01 2.50749916e-01 3.11063111e-01 -1.77992269e-01
7.24596798e-01 -3.08511466e-01 3.61590862e-01 3.31557423e-01
-2.48234034e-01 6.53319955e-01 2.96435833e-01 -2.00201809e-01
-9.08186555e-01 -7.25261807e-01 -7.77093396e-02 1.29735589e+00
5.58549315e-02 -4.47440535e-01 -7.72210479e-01 -7.79410303e-01
-3.62596065e-01 8.34032655e-01 -8.58879626e-01 -1.19792089e-01
-2.32698724e-01 -6.91203594e-01 8.02293718e-01 5.89347303e-01
6.46705568e-01 -1.34845996e+00 -4.24406379e-01 -1.48876771e-01
-7.43086934e-01 -1.27891862e+00 -7.51837373e-01 1.49206696e-02
-1.10734008e-01 -6.82196975e-01 -8.00076008e-01 -8.13263595e-01
4.64776754e-01 -2.09799513e-01 1.03882575e+00 -3.80234458e-02
-3.56371462e-01 4.25348848e-01 -4.39286679e-01 -2.57284820e-01
-1.09576166e-01 1.11856550e-01 -2.05070223e-03 3.00641563e-02
4.80166167e-01 -1.67979702e-01 -4.53529209e-01 1.65099397e-01
-6.34863496e-01 3.02015632e-01 4.61320817e-01 9.64754701e-01
6.77139580e-01 -4.38914984e-01 4.49756444e-01 -4.54585671e-01
4.30461735e-01 -2.41094723e-01 -2.92442232e-01 4.98568654e-01
-4.24755335e-01 -2.73135602e-01 3.48835051e-01 -6.03572309e-01
-1.00753117e+00 2.24817932e-01 -3.55970860e-01 -9.13210452e-01
-3.83219302e-01 2.92930603e-01 -3.22106630e-01 -1.22341514e-02
7.41005167e-02 -1.89108014e-01 -3.90804142e-01 -2.28605643e-01
5.39943814e-01 9.90499079e-01 8.90445232e-01 -5.26673019e-01
7.15247020e-02 3.20749283e-02 -3.07601154e-01 -6.26247764e-01
-1.05906606e+00 -5.34339428e-01 -4.78024840e-01 -6.46562874e-01
1.46646440e+00 -1.17349136e+00 -1.09971845e+00 6.84070349e-01
-1.11058474e+00 -1.50627255e-01 3.62225056e-01 5.72555423e-01
-5.85877895e-01 1.53302759e-01 -7.56919384e-01 -8.25056672e-01
-7.22294092e-01 -1.27372384e+00 1.34695292e+00 3.77510160e-01
-5.30842483e-01 -7.76174366e-01 6.76880553e-02 7.07720399e-01
4.19201516e-02 2.00196341e-01 7.38378346e-01 -7.33987689e-01
1.13639891e-01 4.65468019e-02 -5.83016098e-01 4.47596192e-01
-3.14652234e-01 -1.10176064e-01 -1.35496533e+00 -2.72040576e-01
-7.68434256e-02 -9.79737997e-01 1.13940620e+00 4.59101975e-01
1.50779486e+00 5.41760325e-02 -6.97058141e-02 7.13455379e-01
1.41225994e+00 1.53900117e-01 5.67926764e-01 2.85348237e-01
1.04477429e+00 8.21194053e-01 8.45389187e-01 3.23266447e-01
6.46732450e-01 9.35038328e-01 3.77998412e-01 -1.95489883e-01
-2.12181639e-02 1.68748781e-01 6.21699333e-01 1.00451601e+00
2.66195796e-02 -3.44401091e-01 -9.02197480e-01 4.94415104e-01
-1.95846188e+00 -8.63835037e-01 -9.18078348e-02 1.72198582e+00
6.85638905e-01 -3.21730137e-01 6.14100434e-02 -3.85288149e-01
7.68620133e-01 6.97111525e-03 -3.49789411e-01 -8.34679306e-01
-3.50652426e-01 8.69063884e-02 2.39456266e-01 5.86763442e-01
-1.40910041e+00 1.32676375e+00 5.64217091e+00 1.15030110e+00
-1.12815452e+00 3.13650280e-01 9.68110979e-01 -4.17070448e-01
9.17555857e-03 -5.18322229e-01 -8.49376559e-01 4.82535154e-01
1.28687322e+00 3.86044234e-02 3.73999208e-01 5.80154777e-01
-2.05520213e-01 -2.17787787e-01 -8.61164927e-01 1.49342048e+00
7.39319146e-01 -8.31964433e-01 -2.67942101e-01 -4.97919694e-02
5.68959475e-01 -2.61955671e-02 2.59077817e-01 6.03568554e-01
9.77413133e-02 -1.31935549e+00 5.78984439e-01 5.83959341e-01
1.26851892e+00 -1.11638308e+00 7.88285792e-01 -1.18036360e-01
-1.19586706e+00 2.29327783e-01 -1.00876614e-02 3.54678005e-01
9.60555598e-02 2.62863547e-01 -5.00972450e-01 5.40847659e-01
8.13609779e-01 6.72701597e-01 -6.52875483e-01 6.38328373e-01
-1.69389561e-01 4.21115875e-01 -7.74367973e-02 1.21418266e-02
2.05674946e-01 -1.47680447e-01 1.89844370e-01 1.59685647e+00
3.78284454e-01 -1.03296958e-01 4.65429835e-02 4.51276273e-01
-4.76287574e-01 5.69164932e-01 -4.27108288e-01 -7.18434453e-02
-2.01584655e-03 1.79535615e+00 -6.01368248e-01 -5.81985772e-01
-4.24619347e-01 1.72984087e+00 4.20324653e-01 9.91028026e-02
-1.22563374e+00 -4.58835006e-01 4.22752470e-01 -8.87741625e-01
1.30816162e-01 3.07488918e-01 -3.34916711e-01 -1.34053791e+00
-2.40805700e-01 -9.72235501e-01 4.65861917e-01 -1.27740216e+00
-1.17985034e+00 9.51867521e-01 -1.25333995e-01 -1.02370465e+00
-2.89251864e-01 -8.08316290e-01 -4.35071051e-01 7.80681729e-01
-1.14825082e+00 -1.66975331e+00 -3.05121958e-01 6.20451927e-01
5.31132877e-01 -1.74297035e-01 1.17500675e+00 4.29927409e-01
-8.04682076e-01 1.03553295e+00 1.19979322e-01 1.95423409e-01
1.34831345e+00 -1.24419546e+00 -4.58994061e-02 6.22095704e-01
9.91331413e-02 7.13352785e-02 5.42943180e-01 -6.66084349e-01
-6.27708614e-01 -1.14155996e+00 1.02592397e+00 -4.18711394e-01
5.91300309e-01 -3.54596585e-01 -6.36716962e-01 7.21025586e-01
7.72855341e-01 -3.12949359e-01 8.25205266e-01 1.35970488e-01
-6.21862173e-01 1.61913246e-01 -1.10812044e+00 6.28422737e-01
6.11593962e-01 -7.14988291e-01 -1.25879824e-01 1.06417887e-01
5.63184440e-01 -4.57008451e-01 -1.13659692e+00 5.52130818e-01
7.46198118e-01 -5.72409272e-01 8.84013414e-01 -4.20721740e-01
9.26346958e-01 -2.60451771e-02 -4.16794002e-01 -1.41762769e+00
-1.06885783e-01 -6.70535713e-02 2.86499262e-01 1.33798802e+00
5.22754788e-01 -9.33408551e-03 3.88933510e-01 4.37378675e-01
-7.38750398e-02 -7.17617333e-01 -9.02705133e-01 -2.60319382e-01
1.84581503e-01 -4.90551144e-01 1.01859592e-01 1.19995713e+00
8.29018727e-02 7.38733292e-01 -8.72189939e-01 9.88668483e-03
4.55034316e-01 1.08224988e-01 3.44072312e-01 -8.24383795e-01
-7.29818195e-02 -3.92714709e-01 -1.95911914e-01 -4.12552804e-01
7.09099948e-01 -9.41001058e-01 3.71567421e-02 -1.11755431e+00
6.06258690e-01 2.90555447e-01 -7.35705674e-01 6.88556015e-01
-2.65569359e-01 7.36934185e-01 4.19572502e-01 -1.49917677e-01
-9.74840462e-01 6.57368660e-01 1.00654233e+00 -2.60153890e-01
1.26012966e-01 -5.79825163e-01 -5.29460549e-01 7.65944242e-01
1.10740709e+00 -2.67237991e-01 -8.38680863e-02 -2.00038999e-01
1.26931025e-02 1.80573836e-02 1.48947507e-01 -8.96919966e-01
-5.72746843e-02 3.69370431e-02 6.89876676e-01 -9.15164590e-01
8.21834087e-01 -7.97315657e-01 -2.46318486e-02 1.04428455e-01
-5.21978498e-01 4.06853646e-01 5.91904402e-01 -2.05969904e-02
-2.81894207e-01 -3.19348782e-01 9.47281003e-01 8.16608295e-02
-9.09755051e-01 1.73777819e-01 -5.18797219e-01 3.41471992e-02
1.07278144e+00 1.38295621e-01 -3.03512514e-01 -5.28641641e-01
-9.57808256e-01 4.01211470e-01 5.27639210e-01 6.08101845e-01
7.60157466e-01 -1.67216861e+00 -7.14088857e-01 -1.66185111e-01
4.52636361e-01 -7.06212461e-01 5.83725512e-01 9.18811321e-01
-4.79212515e-02 1.98276430e-01 -7.20734596e-01 -2.51405865e-01
-1.97971201e+00 2.34657839e-01 4.49713349e-01 -4.00836676e-01
-1.33133858e-01 1.06373584e+00 2.92376310e-01 -5.55956662e-01
3.47493768e-01 1.39589623e-01 -4.89546657e-01 3.14186692e-01
6.34942055e-01 2.33217031e-01 -1.06934220e-01 -1.25055265e+00
-6.87670171e-01 8.01078260e-01 -3.12166065e-01 -6.22864723e-01
1.09432030e+00 -2.59908199e-01 -2.16260821e-01 6.73821330e-01
1.69581175e+00 -3.00112724e-01 -6.23974144e-01 2.01904327e-02
-3.01334381e-01 3.41768712e-02 1.52824521e-01 -1.19566250e+00
-1.14742935e+00 1.13578749e+00 1.00973403e+00 -5.89081943e-01
1.36611068e+00 -2.77193519e-03 6.23081863e-01 1.98140100e-01
-1.23704679e-01 -1.53587306e+00 1.93786442e-01 7.05936968e-01
1.00180006e+00 -1.53379333e+00 -3.16996127e-01 -1.46095246e-01
-1.20914388e+00 1.01716375e+00 1.07657886e+00 1.90459758e-01
3.45539629e-01 1.08417869e-01 4.55194622e-01 -2.45762005e-01
-8.30506504e-01 -1.86234400e-01 6.15990341e-01 1.70570552e-01
6.35879338e-01 1.10488161e-01 -2.67632693e-01 7.35735655e-01
-1.52490065e-01 -9.54782069e-02 2.58861482e-01 3.45750391e-01
-7.83831179e-02 -9.31705594e-01 -2.98250616e-01 4.73856568e-01
-7.15998948e-01 -2.10668147e-01 -5.09332359e-01 4.39551383e-01
1.23974711e-01 9.81741250e-01 1.59034014e-01 -9.13825631e-01
7.30532408e-02 4.29093063e-01 4.80162263e-01 -1.99617997e-01
-7.92956114e-01 3.77623826e-01 3.48661631e-01 -7.91449785e-01
-5.78975141e-01 -5.30572176e-01 -1.31780243e+00 -1.74872443e-01
-2.07264319e-01 -2.43394196e-01 8.42653811e-01 9.27000642e-01
1.25813812e-01 6.81721985e-01 4.34971213e-01 -6.85510695e-01
1.72634963e-02 -1.14963245e+00 -4.24759984e-01 5.10672688e-01
6.29596338e-02 -5.79111397e-01 -1.11250088e-01 6.37379587e-02] | [13.271745681762695, 5.135117053985596] |
877d70ea-9ac4-410a-a0af-5e2f602c6d88 | an-expressive-deep-model-for-human-action | 1502.00501 | null | http://arxiv.org/abs/1502.00501v1 | http://arxiv.org/pdf/1502.00501v1.pdf | An Expressive Deep Model for Human Action Parsing from A Single Image | This paper aims at one newly raising task in vision and multimedia research:
recognizing human actions from still images. Its main challenges lie in the
large variations in human poses and appearances, as well as the lack of
temporal motion information. Addressing these problems, we propose to develop
an expressive deep model to naturally integrate human layout and surrounding
contexts for higher level action understanding from still images. In
particular, a Deep Belief Net is trained to fuse information from different
noisy sources such as body part detection and object detection. To bridge the
semantic gap, we used manually labeled data to greatly improve the
effectiveness and efficiency of the pre-training and fine-tuning stages of the
DBN training. The resulting framework is shown to be robust to sometimes
unreliable inputs (e.g., imprecise detections of human parts and objects), and
outperforms the state-of-the-art approaches. | ['Rui Huang', 'Liang Lin', 'Xiaolong Wang', 'Zhujin Liang'] | 2015-02-02 | null | null | null | null | ['action-understanding', 'action-parsing'] | ['computer-vision', 'natural-language-processing'] | [ 2.65359670e-01 2.83793025e-02 1.30157694e-01 -3.17204773e-01
-5.42113841e-01 -4.66889828e-01 5.54766059e-01 1.88241284e-02
-6.61974967e-01 5.45479894e-01 2.22936153e-01 3.23339313e-01
2.61592537e-01 -4.22998220e-01 -8.03133368e-01 -6.06262445e-01
-5.54508343e-02 3.77303392e-01 7.56183386e-01 -1.92706883e-01
-1.84767738e-01 4.57009882e-01 -1.79308248e+00 3.64503771e-01
4.83575046e-01 1.11146343e+00 1.01819165e-01 8.41326356e-01
3.64816308e-01 8.33303213e-01 -4.60271388e-01 -3.78119916e-01
2.23083586e-01 -2.31606916e-01 -6.49991453e-01 5.27666032e-01
6.52872086e-01 -7.99462736e-01 -4.97024357e-01 1.09821332e+00
5.21751463e-01 3.24475974e-01 3.66678923e-01 -1.31684732e+00
-5.32860309e-02 1.55227512e-01 -6.24504447e-01 1.77977979e-01
3.21459144e-01 4.00152385e-01 6.07994020e-01 -8.60197127e-01
6.60297871e-01 1.51584589e+00 6.60905719e-01 6.55159414e-01
-1.02932751e+00 -2.54500628e-01 4.58771586e-01 4.14761037e-01
-1.21118248e+00 -6.88697994e-01 7.92308092e-01 -6.22601986e-01
5.30987203e-01 1.37188733e-01 7.40479112e-01 1.53953886e+00
-1.51051069e-02 1.25692642e+00 8.40043008e-01 -4.51631665e-01
2.88375348e-01 8.60235393e-02 -9.17595103e-02 7.64517367e-01
1.62419170e-01 2.14581624e-01 -3.35447729e-01 2.46593487e-02
9.08207119e-01 1.13818623e-01 -1.92295164e-01 -7.55213320e-01
-1.18485379e+00 5.41975260e-01 4.77311134e-01 1.84602439e-01
-4.20514286e-01 3.84651393e-01 3.82192522e-01 -2.56870747e-01
2.89395362e-01 -3.40328831e-03 -4.88331139e-01 -7.06193894e-02
-7.82124877e-01 4.03019249e-01 5.00536203e-01 7.41640091e-01
5.23916006e-01 -1.89568743e-01 -2.59624571e-01 7.14761972e-01
2.97731280e-01 4.42123145e-01 1.13899037e-01 -1.08713305e+00
5.21093905e-01 3.25822860e-01 4.90353853e-01 -1.18405926e+00
-5.68989515e-01 -3.26873004e-01 -6.19218290e-01 2.71253884e-01
7.64632165e-01 -2.48101130e-01 -1.23712480e+00 1.82585180e+00
6.65882647e-01 8.69865045e-02 -2.99842864e-01 1.29464447e+00
7.24150598e-01 3.09094310e-01 3.00693035e-01 7.56860971e-02
1.28373826e+00 -1.01406288e+00 -6.76514506e-01 -7.88244605e-01
2.60912895e-01 -4.34529901e-01 6.50172591e-01 2.36100152e-01
-1.17274904e+00 -8.67647171e-01 -7.59443283e-01 -2.51422524e-01
-2.82399178e-01 3.15647006e-01 5.66437066e-01 3.98153692e-01
-7.24153757e-01 5.24468660e-01 -1.18982422e+00 -5.17436624e-01
5.32580495e-01 2.70736337e-01 -6.38693213e-01 -2.76256830e-01
-9.66913402e-01 8.33023012e-01 5.14937758e-01 5.67744851e-01
-9.59714949e-01 -2.69398123e-01 -1.04527521e+00 -1.10180765e-01
9.33512866e-01 -9.32268739e-01 1.39495170e+00 -1.18704605e+00
-1.34661722e+00 8.33050132e-01 -1.17222583e-02 -3.44756097e-01
8.99413526e-01 -5.15785158e-01 -1.95044383e-01 2.71081120e-01
-3.55625115e-02 8.22820485e-01 8.84251237e-01 -1.27051687e+00
-7.80779779e-01 -7.52938688e-01 1.58197507e-01 2.11459160e-01
-4.67011780e-02 -5.32026924e-02 -8.23258340e-01 -5.97361386e-01
7.00399801e-02 -1.07946789e+00 -4.99246091e-01 2.80059457e-01
-3.60750079e-01 3.75767499e-02 7.32977808e-01 -8.03221941e-01
7.93929160e-01 -2.02550793e+00 3.59331131e-01 -2.25973912e-02
1.52535394e-01 4.33993727e-01 -1.25731051e-01 7.27883130e-02
2.12716758e-01 -3.59406114e-01 3.98769453e-02 -4.64472950e-01
-4.68389541e-02 5.34198046e-01 2.73897108e-02 6.14077151e-01
3.46996009e-01 1.09263027e+00 -1.03104699e+00 -5.62280118e-01
5.16232669e-01 5.24317801e-01 -2.95987278e-01 1.63045421e-01
-3.78206372e-01 5.98571181e-01 -3.99624318e-01 8.54494035e-01
4.02174741e-01 -1.10283881e-01 2.20894460e-02 -3.56754720e-01
2.10202396e-01 -5.27202487e-02 -1.54396749e+00 1.92176282e+00
-1.35693565e-01 4.76924062e-01 3.34934592e-01 -1.06575119e+00
3.02691162e-01 2.76565164e-01 5.27553320e-01 -6.21380210e-01
3.05588245e-01 -1.14882343e-01 6.38826042e-02 -7.12432623e-01
4.63585377e-01 1.59424841e-02 3.03048138e-02 -1.20987803e-01
2.13342905e-01 3.51988524e-01 7.82359913e-02 -7.22217113e-02
9.85100687e-01 3.74855429e-01 3.15674782e-01 6.37637451e-02
4.53471541e-01 -1.06599808e-01 7.34213650e-01 7.77847648e-01
-6.12115562e-01 7.09683836e-01 3.71361226e-01 -6.69895232e-01
-8.76190364e-01 -1.00989544e+00 2.28797287e-01 1.34283686e+00
4.08527970e-01 -2.21784681e-01 -7.47848451e-01 -8.88132274e-01
-1.06906980e-01 2.07108155e-01 -8.34359586e-01 -1.35538861e-01
-7.38538921e-01 -5.42170227e-01 3.19402963e-01 9.87693250e-01
4.02085036e-01 -1.06782138e+00 -1.19111788e+00 3.11096549e-01
-3.79501939e-01 -1.46981919e+00 -3.69223773e-01 1.54133603e-01
-6.31701052e-01 -1.21370709e+00 -8.79259646e-01 -4.77444470e-01
5.06559372e-01 8.33728686e-02 1.10426247e+00 -1.93546161e-01
-6.82154059e-01 5.39702952e-01 -2.46308610e-01 -4.18188691e-01
-2.71810085e-01 -2.61279196e-01 6.23513795e-02 1.21527977e-01
1.35005862e-01 -3.72543603e-01 -9.03207541e-01 3.54795873e-01
-8.79468203e-01 3.21231335e-02 7.09453940e-01 8.59449744e-01
4.27172154e-01 4.48661223e-02 -1.19775916e-02 -4.75360602e-01
1.21165104e-02 2.62167044e-02 -5.07251799e-01 1.78586870e-01
2.59083938e-02 -5.51565774e-02 1.86029628e-01 -6.48240864e-01
-9.71005499e-01 5.85457027e-01 -1.49321496e-01 -5.12788415e-01
-5.90971947e-01 9.31259915e-02 -4.00317758e-01 9.80835594e-03
5.34332812e-01 -4.23772894e-02 2.98313517e-03 -4.72275287e-01
5.38009703e-01 3.30083311e-01 9.65125144e-01 -3.16346914e-01
5.90465307e-01 6.71725690e-01 3.24550234e-02 -7.72738516e-01
-1.02773941e+00 -6.80114210e-01 -9.56041515e-01 -4.20943528e-01
1.02844059e+00 -1.02411389e+00 -5.19035280e-01 6.55052125e-01
-1.28341341e+00 -2.74121821e-01 -1.41680002e-01 3.23485106e-01
-6.50613844e-01 4.76538688e-01 -7.91539133e-01 -9.32483673e-01
-4.23250161e-02 -1.11126816e+00 1.38844383e+00 1.97678864e-01
-8.86326581e-02 -8.47790778e-01 -1.09845802e-01 3.96435291e-01
-6.47314172e-03 5.80228209e-01 4.67474967e-01 -4.07649219e-01
-4.88463134e-01 -3.28228444e-01 -8.80274996e-02 4.40735996e-01
-5.45143448e-02 -1.86414346e-01 -1.06438017e+00 -1.77236304e-01
-2.20085949e-01 -5.88038802e-01 9.19691801e-01 6.82030618e-01
8.79784763e-01 -2.71945775e-01 -5.52418292e-01 3.62509757e-01
1.03603625e+00 -1.32439961e-03 5.65062106e-01 2.28347912e-01
8.89127731e-01 8.58770728e-01 7.58290887e-01 5.13901711e-01
2.52299905e-01 1.00900376e+00 6.05508387e-01 -3.67522985e-01
-1.63869619e-01 -1.43221810e-01 3.12293351e-01 -6.12507239e-02
-2.12051213e-01 -8.97881761e-02 -8.96634161e-01 5.33801496e-01
-2.23516369e+00 -1.11403322e+00 6.67862222e-02 2.06524110e+00
5.76960921e-01 2.06124544e-01 4.33766812e-01 1.26475230e-01
5.54037273e-01 9.76614505e-02 -5.26429892e-01 3.97400081e-01
7.62329400e-02 -2.31265411e-01 4.57418323e-01 2.79142499e-01
-1.66827464e+00 1.05263793e+00 6.17876244e+00 4.74761993e-01
-8.49106073e-01 -1.87224417e-03 4.21132296e-01 -7.61431679e-02
3.87729377e-01 -3.83130521e-01 -6.20370626e-01 3.61494511e-01
3.90568614e-01 5.37165880e-01 9.80860069e-02 8.26359510e-01
1.72433525e-01 -2.74183065e-01 -1.31548190e+00 1.11790276e+00
2.08749145e-01 -8.00188065e-01 -3.18638802e-01 -3.93566601e-02
5.95699787e-01 7.86077790e-03 -3.03385537e-02 2.41655484e-01
2.15145692e-01 -8.72440338e-01 9.61663783e-01 4.86669034e-01
1.89875692e-01 -6.69132471e-01 8.38385820e-01 6.24176502e-01
-1.06851697e+00 -3.20058644e-01 -2.07833514e-01 -2.36324698e-01
2.36519828e-01 4.80699807e-01 -5.42301834e-01 3.77034694e-01
8.06289494e-01 7.06246316e-01 -6.32316351e-01 1.04420578e+00
-2.20746532e-01 1.81135938e-01 -4.98713672e-01 2.96072364e-01
2.22199187e-01 2.06243977e-01 4.51085567e-01 1.17981577e+00
-4.51829145e-03 1.54979289e-01 6.82812274e-01 7.31137812e-01
3.23786855e-01 -4.10354108e-01 -4.80311602e-01 2.04551741e-01
3.89837883e-02 1.18872702e+00 -8.08436573e-01 -3.65601510e-01
-3.76632333e-01 1.25840461e+00 2.98903078e-01 5.52503824e-01
-9.63344812e-01 6.59577325e-02 6.23176515e-01 2.42872968e-01
5.71743429e-01 -2.79823810e-01 5.19994497e-02 -1.19282138e+00
3.04544270e-01 -8.54848504e-01 3.85307759e-01 -7.86551058e-01
-1.09387875e+00 4.16232973e-01 1.59905583e-01 -9.81084466e-01
-4.70992535e-01 -8.09984207e-01 -1.94261223e-01 4.25572217e-01
-1.18770111e+00 -1.35233593e+00 -4.68184441e-01 6.19982302e-01
5.46407104e-01 2.35653490e-01 4.50331181e-01 3.85389119e-01
-6.04801714e-01 3.74466926e-01 -3.41399699e-01 3.85933846e-01
6.57299459e-01 -1.12340426e+00 2.95647234e-01 1.04133570e+00
3.06851655e-01 3.27199817e-01 8.97802711e-01 -4.72920895e-01
-1.24817121e+00 -1.07706797e+00 4.47828293e-01 -8.10148418e-01
3.98670703e-01 -4.91768718e-01 -6.33079588e-01 6.35810852e-01
-2.86632091e-01 4.83299255e-01 2.27304712e-01 5.32818399e-02
-2.05613285e-01 3.48019339e-02 -8.04492533e-01 6.18067265e-01
1.12722540e+00 -4.18513656e-01 -6.41396284e-01 2.56869256e-01
1.77695617e-01 -6.75240755e-01 -3.58301193e-01 6.75095081e-01
7.27817893e-01 -1.15760350e+00 1.25427735e+00 -7.57152379e-01
2.54206955e-01 -3.00777853e-01 -1.56730205e-01 -8.94074678e-01
-4.01382327e-01 -3.01137209e-01 -4.61418718e-01 9.43425596e-01
-5.56218401e-02 1.32836089e-01 8.81543100e-01 8.17575276e-01
1.17581345e-01 -5.86271703e-01 -1.01018894e+00 -5.52795470e-01
-2.89696425e-01 -4.61367637e-01 -1.00424320e-01 5.16572297e-01
-1.73069417e-01 3.88527095e-01 -7.29828238e-01 4.02074635e-01
5.60308218e-01 -8.01447034e-02 9.31805849e-01 -1.27315950e+00
-5.23827374e-01 -1.72911316e-01 -7.07054853e-01 -1.16297579e+00
-5.28038777e-02 -1.69793382e-01 4.80957866e-01 -1.55134690e+00
2.84158409e-01 8.51145908e-02 -2.33092919e-01 5.42960644e-01
-4.14200693e-01 3.87451619e-01 3.80853295e-01 -7.21594021e-02
-1.05710220e+00 6.39024436e-01 1.18965995e+00 -2.05396086e-01
-1.50684550e-01 3.07818145e-01 -2.39322186e-01 1.07010925e+00
3.63873392e-01 -2.97962427e-01 -3.21567923e-01 -3.93241614e-01
-1.65525630e-01 1.72463477e-01 1.09610188e+00 -1.22739315e+00
2.58234143e-01 -9.64784920e-02 7.19394743e-01 -7.58759260e-01
5.99049270e-01 -1.02534318e+00 7.94486550e-04 4.90253717e-01
-3.06640655e-01 -2.17148915e-01 2.06415132e-01 8.68611336e-01
-1.65319979e-01 1.50803760e-01 7.27203488e-01 -4.22862798e-01
-9.50825155e-01 3.24308425e-01 -2.31661797e-01 4.12037550e-03
1.05670285e+00 -3.84353995e-01 4.56449017e-02 -3.84618670e-01
-9.31902647e-01 2.86933213e-01 3.24711531e-01 7.31948495e-01
3.73975575e-01 -1.18212855e+00 -3.97040963e-01 1.84250638e-01
1.66537270e-01 1.07526541e-01 4.93736446e-01 9.56140578e-01
-1.76075354e-01 2.00048313e-01 -2.49794677e-01 -8.22636425e-01
-1.38170981e+00 7.05239594e-01 5.23936152e-01 -1.36176482e-01
-7.97801614e-01 7.93212831e-01 3.42666894e-01 -1.65212993e-02
7.50561118e-01 -5.82766116e-01 -4.71714102e-02 -1.20797031e-01
6.83837712e-01 5.04126191e-01 -9.32432488e-02 -8.05623829e-01
-5.71688473e-01 6.07471824e-01 1.04575247e-01 -2.10304409e-01
9.13763404e-01 -2.66775966e-01 2.95103759e-01 3.73048007e-01
9.81967747e-01 -2.98852473e-01 -1.68889248e+00 -2.95330077e-01
-6.17704391e-02 -4.35186356e-01 1.30360816e-02 -9.33336675e-01
-1.02154744e+00 9.00593936e-01 7.59546995e-01 -1.43529639e-01
9.64883983e-01 5.73757254e-02 7.36620963e-01 4.08856839e-01
3.94188404e-01 -1.39409554e+00 2.55336255e-01 3.56468022e-01
7.34888315e-01 -1.67139709e+00 6.12774976e-02 -3.47472668e-01
-6.45101964e-01 1.02933109e+00 8.02911580e-01 1.30452633e-01
4.44622040e-01 2.84376889e-01 2.31331483e-01 -7.30282292e-02
-4.89224166e-01 -5.71993828e-01 6.58110499e-01 7.44051933e-01
1.10901766e-01 -2.32401893e-01 2.46706009e-02 5.16776979e-01
3.13297659e-01 1.73082530e-01 9.74775031e-02 1.02144277e+00
-3.32005769e-01 -7.58375704e-01 -4.55549061e-01 -4.57664616e-02
-5.03126383e-01 3.40644777e-01 -5.18023252e-01 1.00101650e+00
5.51140606e-01 8.04570615e-01 -9.58840773e-02 -2.88344026e-01
4.35924262e-01 -5.19275330e-02 6.79232240e-01 -3.98618132e-01
-2.81270713e-01 4.45496768e-01 2.21554656e-02 -1.00918007e+00
-6.36731684e-01 -6.11214221e-01 -9.53380525e-01 2.97319472e-01
-2.11751699e-01 -4.56260949e-01 3.29414934e-01 1.12155545e+00
1.44562468e-01 6.70092583e-01 -2.23038346e-02 -1.34135389e+00
-7.28378236e-01 -8.71524572e-01 -5.58467388e-01 7.65623450e-01
6.09042764e-01 -9.02951181e-01 -3.57498564e-02 3.15660119e-01] | [7.859068870544434, 0.10919881612062454] |
59ce6d84-3ff4-4635-baed-b147e35d8e96 | a-novel-plug-and-play-approach-for | 2208.09449 | null | https://arxiv.org/abs/2208.09449v2 | https://arxiv.org/pdf/2208.09449v2.pdf | A Novel Plug-and-Play Approach for Adversarially Robust Generalization | In this work, we propose a robust framework that employs adversarially robust training to safeguard the machine learning models against perturbed testing data. We achieve this by incorporating the worst-case additive adversarial error within a fixed budget for each sample during model estimation. Our main focus is to provide a plug-and-play solution that can be incorporated in the existing machine learning algorithms with minimal changes. To that end, we derive the ready-to-use solution for several widely used loss functions with a variety of norm constraints on adversarial perturbation for various supervised and unsupervised ML problems, including regression, classification, two-layer neural networks, graphical models, and matrix completion. The solutions are either in closed-form, 1-D optimization, semidefinite programming, difference of convex programming or a sorting-based algorithm. Finally, we validate our approach by showing significant performance improvement on real-world datasets for supervised problems such as regression and classification, as well as for unsupervised problems such as matrix completion and learning graphical models, with very little computational overhead. | ['Jean Honorio', 'Adarsh Barik', 'Deepak Maurya'] | 2022-08-19 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [ 4.96356606e-01 3.80837053e-01 -6.89208880e-02 -3.94493252e-01
-1.23904431e+00 -7.44984746e-01 2.48032898e-01 2.22102106e-01
-3.72082561e-01 9.29096699e-01 -2.33059466e-01 -4.71650094e-01
-4.19879377e-01 -4.45455939e-01 -1.22308779e+00 -7.60494888e-01
-2.86103994e-01 5.18587172e-01 -1.41242728e-01 -1.75608173e-01
7.47822076e-02 7.12006211e-01 -8.50180149e-01 -5.01246639e-02
8.54768336e-01 7.78991997e-01 -6.18178308e-01 6.89123094e-01
5.36496818e-01 7.84572542e-01 -3.48223954e-01 -7.97197044e-01
6.69123888e-01 5.75123467e-02 -5.34014165e-01 8.85132477e-02
5.86813509e-01 -1.69438735e-01 -4.66137081e-01 1.14783537e+00
5.48210740e-01 1.22951739e-01 7.81294763e-01 -1.56525004e+00
-2.18473032e-01 5.81701875e-01 -6.36695147e-01 -3.00225496e-01
2.48630062e-01 7.96647519e-02 6.82232261e-01 -7.94376731e-01
5.15186906e-01 1.26827145e+00 8.34868670e-01 5.68586171e-01
-1.56493771e+00 -5.56600153e-01 1.75506935e-01 -1.77037910e-01
-1.27110147e+00 -4.45444942e-01 7.51988232e-01 -5.74241340e-01
5.63420236e-01 6.49123192e-01 -7.67817348e-02 1.23312700e+00
-2.76043992e-02 9.20008004e-01 1.04355085e+00 -4.21235681e-01
3.89457583e-01 4.07564849e-01 7.38722682e-02 7.77318060e-01
3.28585178e-01 -9.11445022e-02 -1.23119697e-01 -6.43444717e-01
3.36033076e-01 -2.38485411e-02 -2.26647988e-01 -6.94275498e-01
-7.65718579e-01 1.01291311e+00 1.19570032e-01 -3.03325176e-01
-8.49461406e-02 1.77918866e-01 4.66541231e-01 5.89648068e-01
6.98825240e-01 3.70506287e-01 -4.23526615e-01 2.51528263e-01
-8.17651153e-01 3.85639489e-01 1.12061703e+00 6.94925964e-01
6.75047159e-01 3.15513253e-01 -7.84177184e-02 8.10739934e-01
4.37216520e-01 5.48820615e-01 1.43241927e-01 -7.77589858e-01
9.23269749e-01 2.09688798e-01 1.48702130e-01 -1.19234204e+00
-2.27876768e-01 -3.66069376e-01 -1.13482749e+00 4.81363624e-01
4.13810879e-01 -5.18460929e-01 -6.19424999e-01 1.89263761e+00
6.28430784e-01 3.06505859e-01 1.89809382e-01 5.28453529e-01
2.22220004e-01 5.43484986e-01 -2.32569098e-01 -4.37634587e-01
6.92672491e-01 -8.56386602e-01 -5.16818404e-01 -3.70605707e-01
7.55897641e-01 -5.63751638e-01 8.66572380e-01 6.83780193e-01
-1.09839880e+00 -5.03795743e-02 -1.17941713e+00 1.87245905e-01
-3.23561907e-01 -1.14129372e-01 5.48308969e-01 1.02729630e+00
-7.99592912e-01 8.98291051e-01 -8.06474745e-01 -1.03575531e-02
3.32088888e-01 6.93415761e-01 -7.67970979e-01 -5.81065081e-02
-1.05470908e+00 6.76140249e-01 3.29451114e-01 2.00739413e-01
-9.36016500e-01 -9.86828148e-01 -1.04754794e+00 -2.57297903e-01
5.40977657e-01 -5.14798164e-01 8.83317351e-01 -8.27305198e-01
-1.55910897e+00 1.02148342e+00 1.27200320e-01 -8.02529395e-01
1.09106362e+00 -3.19023371e-01 -7.18879178e-02 -2.01339021e-01
-3.57388377e-01 -1.49751641e-02 1.09479535e+00 -1.29066765e+00
-2.35568266e-02 -4.71197963e-01 1.43201619e-01 -5.85315675e-02
-5.77290297e-01 1.15909740e-01 -2.93353796e-01 -8.28440130e-01
-1.40680313e-01 -1.21476245e+00 -6.56306505e-01 1.49820343e-01
-8.71255636e-01 2.80601352e-01 8.14900458e-01 -1.00356793e+00
1.00848985e+00 -2.05508828e+00 5.79272866e-01 6.04226768e-01
-6.23969957e-02 2.37940967e-01 -4.68402728e-02 5.46713054e-01
-3.42887849e-01 1.02952048e-01 -8.07906508e-01 -7.83526659e-01
3.61060381e-01 2.24702090e-01 -4.86832529e-01 9.29315507e-01
1.67355791e-01 5.44033766e-01 -6.24461174e-01 -2.41346076e-01
5.60250133e-02 6.67383969e-02 -7.17216253e-01 2.67317802e-01
-2.32350841e-01 3.44017386e-01 -2.67044187e-01 5.26029646e-01
8.62443745e-01 2.31663465e-01 1.90678254e-01 9.36525464e-02
4.26836491e-01 -2.87155360e-01 -1.75580037e+00 1.54979920e+00
-5.88537633e-01 4.58375037e-01 6.01855695e-01 -1.30066168e+00
6.13042891e-01 1.01165347e-01 3.82238567e-01 -6.65254425e-03
5.67040220e-02 1.50723189e-01 -3.44742984e-01 -2.64454454e-01
9.57725123e-02 2.92243548e-02 -2.72889376e-01 5.77212155e-01
-3.91759165e-02 -9.93893296e-02 2.16224775e-01 2.90672779e-01
1.16783631e+00 -1.60268024e-01 1.83493584e-01 -1.47694260e-01
6.11290753e-01 -2.64970660e-01 3.78118843e-01 7.69286513e-01
1.85229927e-01 6.23965442e-01 8.06059182e-01 -1.83244478e-02
-8.64222705e-01 -9.97195244e-01 6.79457979e-03 9.62529600e-01
-4.45487440e-01 -1.65589422e-01 -8.68846536e-01 -9.75970566e-01
3.00528079e-01 6.59419298e-01 -5.65480053e-01 -4.24723059e-01
-4.88341779e-01 -9.73142803e-01 6.24557614e-01 2.39194408e-01
8.81468058e-02 -5.82157910e-01 3.43751758e-01 -7.34788179e-02
3.01212758e-01 -1.01410472e+00 -3.96595746e-01 3.54940206e-01
-8.36652517e-01 -1.16221404e+00 -4.77055013e-01 -5.69020092e-01
9.84686077e-01 -2.57785112e-01 6.93699539e-01 -3.29287112e-01
-3.40418547e-01 5.71133971e-01 3.49447988e-02 -3.25016677e-01
-5.96214890e-01 -1.01125196e-01 4.77956623e-01 4.20859188e-01
-4.21751708e-01 -7.82506645e-01 -1.74373195e-01 2.14312539e-01
-1.35677218e+00 -4.63565171e-01 3.94145429e-01 8.39945138e-01
5.91038048e-01 1.76734347e-02 2.51964778e-01 -1.42835307e+00
8.33214521e-01 -6.69773102e-01 -9.02218342e-01 1.92998961e-01
-5.21049500e-01 3.05755675e-01 9.88038838e-01 -7.00321496e-01
-5.65197945e-01 3.18527013e-01 -1.25339881e-01 -8.09363306e-01
1.03324853e-01 6.41734004e-01 -5.12001038e-01 -4.51110542e-01
9.03484464e-01 1.67601645e-01 5.30994460e-02 -5.41368127e-01
4.95305896e-01 5.02476394e-01 3.80223840e-01 -7.97111392e-01
1.49922013e+00 3.67548406e-01 3.99792194e-01 -6.24941409e-01
-7.58363068e-01 -2.35618383e-01 -6.46049201e-01 3.10625508e-02
2.81767428e-01 -7.61032939e-01 -7.10227311e-01 6.99920356e-01
-8.93069088e-01 -5.67732930e-01 -1.73554227e-01 2.48031288e-01
-5.60976923e-01 7.39439785e-01 -4.78792816e-01 -9.26479995e-01
-3.37910533e-01 -1.03601336e+00 7.53794193e-01 -2.29740426e-01
9.88108590e-02 -1.31914711e+00 2.10719436e-01 3.55034232e-01
1.04472145e-01 8.85976255e-01 8.70600402e-01 -9.63593006e-01
-2.58503169e-01 -7.38332808e-01 1.47321686e-01 8.77248228e-01
-2.44768724e-01 1.54944241e-01 -9.21158850e-01 -7.71842659e-01
-7.59025812e-02 -6.34528995e-01 6.63231492e-01 9.55362841e-02
1.44138253e+00 -8.35047185e-01 -7.71792233e-02 1.04511750e+00
1.43980491e+00 -3.49536270e-01 5.41140437e-01 9.04538333e-02
8.90851140e-01 5.66041350e-01 5.33692002e-01 4.75388408e-01
-1.31252810e-01 5.64148247e-01 5.82456529e-01 -1.38789713e-01
6.18022144e-01 -2.18893513e-01 6.04656875e-01 5.36836863e-01
4.41504925e-01 -3.35748233e-02 -8.41372073e-01 1.69564337e-01
-2.07214046e+00 -7.60512471e-01 -1.78921685e-01 2.63160682e+00
1.00400865e+00 2.53194511e-01 1.76177606e-01 3.22680175e-01
5.27359545e-01 1.86102748e-01 -6.22150064e-01 -5.53554237e-01
-2.40357034e-02 2.65647769e-01 8.81779790e-01 6.86974406e-01
-1.37804079e+00 6.53706968e-01 6.44361496e+00 9.57973361e-01
-1.08139586e+00 -1.24169863e-03 7.18500555e-01 -1.84208408e-01
-1.08973682e-01 -4.88976426e-02 -4.58332062e-01 3.29087853e-01
1.00513434e+00 -9.61878672e-02 6.69290125e-01 1.06213355e+00
2.45651931e-01 3.66705656e-01 -1.35394728e+00 8.72734070e-01
7.98479766e-02 -1.03729117e+00 -2.11535573e-01 8.39680061e-02
8.37062359e-01 -1.89671755e-01 3.52907926e-01 3.01892400e-01
4.67671186e-01 -1.05773699e+00 5.51024914e-01 4.03069615e-01
7.77956665e-01 -8.65924358e-01 3.64375412e-01 5.03539264e-01
-5.15337288e-01 -5.31176887e-02 -3.94378006e-01 7.00366646e-02
-2.39723220e-01 8.02492380e-01 -7.05351949e-01 7.15548158e-01
1.63826853e-01 5.25771320e-01 -3.82118255e-01 9.24833357e-01
-2.35073432e-01 8.09862673e-01 -5.52954257e-01 1.85729831e-01
1.85891524e-01 -2.94287503e-01 7.34809041e-01 1.21580970e+00
-1.65396333e-01 -2.36706033e-01 3.44384819e-01 5.55425644e-01
-3.04415315e-01 3.51274520e-01 -7.35671878e-01 -1.04040578e-02
1.81501925e-01 1.28454447e+00 -1.70177564e-01 1.36295065e-01
-6.14040829e-02 1.05050504e+00 4.87852842e-01 4.92630064e-01
-1.00432110e+00 -5.27970970e-01 6.58300579e-01 8.60750824e-02
4.27199006e-02 -1.99432909e-01 -1.81629255e-01 -1.27993345e+00
3.99302572e-01 -1.15928137e+00 5.17641485e-01 -2.78042644e-01
-1.46999359e+00 9.32087079e-02 1.88741520e-01 -1.06264019e+00
-5.46167493e-01 -9.57251787e-01 -7.60811806e-01 8.42253506e-01
-1.20922709e+00 -1.02882040e+00 1.70183584e-01 8.99834394e-01
-9.38687939e-03 -4.16146070e-01 6.84612215e-01 4.34258580e-01
-1.08219969e+00 1.15921307e+00 7.11342692e-01 1.94151506e-01
8.32311332e-01 -1.50367796e+00 1.20921701e-01 1.20053589e+00
-1.12601360e-02 5.62481701e-01 8.66202593e-01 -4.14611489e-01
-1.85034955e+00 -1.36800468e+00 3.16007227e-01 -2.88402975e-01
1.06656897e+00 -7.47561932e-01 -7.30944097e-01 8.54194224e-01
-4.00328666e-01 2.56285429e-01 7.23716736e-01 1.40059173e-01
-5.22236764e-01 -3.17816615e-01 -1.60132813e+00 7.79573202e-01
8.40211987e-01 -4.92923439e-01 -4.35600020e-02 9.94958401e-01
5.95434427e-01 -7.61542439e-01 -9.49084818e-01 2.48767003e-01
2.53448337e-01 -5.40508032e-01 1.10085082e+00 -1.12925899e+00
3.64090770e-01 -1.14484735e-01 -2.95248270e-01 -1.29104555e+00
1.38174862e-01 -1.32940507e+00 -3.71729702e-01 1.26757121e+00
7.04944372e-01 -8.17326963e-01 9.29363728e-01 8.95725191e-01
6.37954324e-02 -8.33113909e-01 -9.84655440e-01 -6.83960438e-01
2.09291056e-01 -6.06370151e-01 -7.23687792e-03 9.39484656e-01
-2.29980126e-01 -9.60642472e-02 -8.66633415e-01 5.71358800e-01
9.68908072e-01 -3.90216410e-01 1.21083319e+00 -9.40938056e-01
-7.11522102e-01 -7.88574219e-02 -5.75284183e-01 -5.35660386e-01
6.17905974e-01 -1.03646588e+00 -1.48619756e-01 -1.02918935e+00
-3.94989811e-02 -4.35002118e-01 -1.52510047e-01 6.26468658e-01
-2.12678567e-01 9.27169919e-02 1.80048868e-01 -1.37010694e-01
-4.98794973e-01 2.68354028e-01 6.64650977e-01 -3.71878684e-01
5.34776933e-02 4.57390219e-01 -7.46656239e-01 6.49736166e-01
6.76318645e-01 -7.29447544e-01 -4.15267706e-01 -2.29702517e-01
2.83575594e-01 -2.02913154e-02 3.56549829e-01 -8.66422355e-01
1.90443650e-01 -2.02917188e-01 2.51096398e-01 -8.61268863e-02
3.61737967e-01 -9.57121074e-01 1.00809000e-01 4.92684633e-01
-4.59131896e-01 -1.32514939e-01 3.37271363e-01 7.56689250e-01
7.11345067e-03 -4.49036807e-01 1.05449796e+00 2.38841608e-01
-7.16586709e-02 5.21165848e-01 1.20399585e-02 3.26684535e-01
1.18707097e+00 1.01174057e-01 -5.25331236e-02 -4.75119561e-01
-9.00410771e-01 2.73526162e-01 2.85731971e-01 1.33368179e-01
4.50191081e-01 -1.19401670e+00 -1.03790510e+00 1.65669933e-01
-1.00450680e-01 -9.14580524e-02 2.26718441e-01 6.74932241e-01
-5.22658050e-01 2.41303250e-01 1.10660605e-01 -3.08905989e-01
-1.32583702e+00 6.86122954e-01 3.69906485e-01 -6.20502949e-01
2.85125449e-02 8.17382097e-01 -3.21319431e-01 -7.89125264e-01
6.54735982e-01 -1.43778160e-01 3.82080644e-01 -1.77779809e-01
3.09257120e-01 5.35162449e-01 1.55251488e-01 -2.95687795e-01
-2.19117567e-01 2.10658595e-01 -7.97682852e-02 -9.18722004e-02
1.31211603e+00 2.55154729e-01 -2.58261412e-01 4.70104396e-01
1.52163029e+00 2.70483732e-01 -1.24241650e+00 -2.67452210e-01
-1.50154144e-01 -3.89821023e-01 -1.71402305e-01 -5.12176216e-01
-1.03659284e+00 6.29983604e-01 4.09651488e-01 2.33627051e-01
1.06742096e+00 -4.40612942e-01 4.34316337e-01 7.75893450e-01
1.96485355e-01 -8.42941940e-01 -1.29095882e-01 2.95171678e-01
9.95089233e-01 -1.24676812e+00 1.81077704e-01 -4.32370216e-01
-5.23669899e-01 8.69566500e-01 2.75139928e-01 -5.01258969e-01
7.31689513e-01 3.45105112e-01 -1.58284679e-01 3.93326104e-01
-5.82441807e-01 3.55204463e-01 4.35385585e-01 7.65083313e-01
-2.72253994e-03 -1.51722103e-01 -9.83470529e-02 4.77677584e-01
-4.40519154e-02 -4.97677922e-01 4.72950786e-01 8.16090763e-01
7.12805428e-03 -1.28316128e+00 -5.09018898e-01 4.08852279e-01
-7.46276140e-01 -5.54975308e-02 -5.07613182e-01 7.80644715e-01
-4.21864092e-01 8.39847505e-01 -5.68825901e-01 -2.75161952e-01
4.38908488e-01 9.32100043e-02 4.42643493e-01 -4.86995399e-01
-6.02371454e-01 -3.46429080e-01 2.06118762e-01 -4.80723351e-01
-4.04669084e-02 -6.58133864e-01 -7.38647103e-01 -3.52335215e-01
-7.21309334e-02 5.37007973e-02 7.83628643e-01 9.47588503e-01
1.29085451e-01 1.62232235e-01 1.06559503e+00 -8.16432595e-01
-1.23254228e+00 -8.93705964e-01 -6.84449732e-01 5.36370099e-01
3.77849162e-01 -3.53420407e-01 -7.21077681e-01 -6.21908382e-02] | [5.757019519805908, 7.813692092895508] |
b0c0226e-4fda-4717-b292-910d0da35ce2 | learning-object-level-point-augmentor-for | 2212.09273 | null | https://arxiv.org/abs/2212.09273v1 | https://arxiv.org/pdf/2212.09273v1.pdf | Learning Object-level Point Augmentor for Semi-supervised 3D Object Detection | Semi-supervised object detection is important for 3D scene understanding because obtaining large-scale 3D bounding box annotations on point clouds is time-consuming and labor-intensive. Existing semi-supervised methods usually employ teacher-student knowledge distillation together with an augmentation strategy to leverage unlabeled point clouds. However, these methods adopt global augmentation with scene-level transformations and hence are sub-optimal for instance-level object detection. In this work, we propose an object-level point augmentor (OPA) that performs local transformations for semi-supervised 3D object detection. In this way, the resultant augmentor is derived to emphasize object instances rather than irrelevant backgrounds, making the augmented data more useful for object detector training. Extensive experiments on the ScanNet and SUN RGB-D datasets show that the proposed OPA performs favorably against the state-of-the-art methods under various experimental settings. The source code will be available at https://github.com/nomiaro/OPA. | ['Ming-Hsuan Yang', 'Yen-Yu Lin', 'Yi-Hsuan Tsai', 'Chen-Hsuan Tai', 'Cheng-Ju Ho'] | 2022-12-19 | null | null | null | null | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 1.18220471e-01 1.06511883e-01 -2.37262219e-01 -4.37110901e-01
-8.26116323e-01 -5.50288439e-01 4.79413509e-01 3.44003767e-01
-2.49282837e-01 1.80363566e-01 -4.66148019e-01 -3.77553314e-01
3.82323444e-01 -6.44643843e-01 -9.08600748e-01 -6.72554374e-01
1.28021449e-01 6.50725782e-01 8.06808352e-01 7.59323016e-02
1.16137460e-01 8.44761014e-01 -1.57676649e+00 -3.15898061e-02
8.76724541e-01 9.86809433e-01 3.86057943e-01 3.44177842e-01
-5.06328642e-01 4.82676238e-01 -2.46004537e-01 -3.21948975e-02
6.32934153e-01 -1.09868031e-02 -4.47498918e-01 4.08523232e-01
6.37803555e-01 -4.43878770e-01 -1.97874710e-01 1.17013085e+00
1.98920235e-01 8.50481763e-02 6.38158679e-01 -1.35110879e+00
-3.41936201e-01 1.03321224e-01 -8.62689495e-01 2.57091492e-01
5.15042059e-02 1.74760953e-01 8.48665237e-01 -1.33943105e+00
1.97989121e-01 1.25164294e+00 4.68090177e-01 3.88957053e-01
-1.22861540e+00 -8.82795513e-01 4.18795109e-01 -3.70147564e-02
-1.57529473e+00 -1.45048201e-01 1.10277236e+00 -4.13694292e-01
7.82656193e-01 2.15558246e-01 7.08111048e-01 6.02629364e-01
-5.46007454e-01 1.11405194e+00 1.22066295e+00 -5.12583554e-01
2.34370887e-01 3.31052363e-01 1.35332733e-01 7.88292468e-01
5.08995175e-01 2.57683676e-02 -3.28586608e-01 -1.61652818e-01
1.07831085e+00 3.68374795e-01 -5.54233305e-02 -1.00806069e+00
-1.06312811e+00 6.42540097e-01 8.56340349e-01 -1.81871548e-01
-3.65036726e-01 -6.71917424e-02 6.55270144e-02 -2.13423565e-01
8.38712811e-01 6.23390041e-02 -4.14944142e-01 2.87085444e-01
-7.84741163e-01 1.84543446e-01 3.89782250e-01 1.21041298e+00
1.02162564e+00 -6.10036291e-02 1.52821049e-01 6.77366555e-01
5.34186423e-01 7.09074736e-01 -1.41077593e-01 -6.96180403e-01
3.71709257e-01 1.17014396e+00 1.29462585e-01 -5.68525612e-01
-2.08058774e-01 -3.93719167e-01 -3.92980993e-01 5.88914990e-01
4.00052756e-01 2.79198766e-01 -1.41296446e+00 1.21012211e+00
9.62540269e-01 3.99036467e-01 -1.66991368e-01 1.08935928e+00
1.10177135e+00 5.59312642e-01 1.28661260e-01 1.86844990e-01
1.24602830e+00 -1.03422415e+00 -2.45071337e-01 -5.13000727e-01
5.88254809e-01 -6.57847404e-01 1.24031782e+00 -8.05770047e-03
-8.96516502e-01 -6.58017933e-01 -1.00951946e+00 -1.35486618e-01
-3.24481070e-01 2.14049235e-01 6.35716856e-01 5.01487017e-01
-4.54223841e-01 1.56812131e-01 -1.20975459e+00 -4.52484697e-01
9.27741468e-01 2.44467422e-01 -2.83379287e-01 -1.38501525e-01
-4.23650563e-01 8.42463791e-01 5.81968606e-01 -8.01232085e-02
-1.03188670e+00 -8.20708930e-01 -9.02609646e-01 -1.53146416e-01
7.05097318e-01 -4.30073828e-01 1.39205849e+00 -4.78867769e-01
-1.24826658e+00 1.22202432e+00 -3.12533341e-02 -2.08973676e-01
5.90072095e-01 -4.43382055e-01 1.67258844e-01 1.66173175e-01
1.22542642e-01 7.89212286e-01 9.55355644e-01 -1.59054923e+00
-8.55910361e-01 -6.04751647e-01 1.27506018e-01 4.53193665e-01
-1.48161918e-01 -8.09634700e-02 -7.88110852e-01 -4.20799196e-01
5.99044859e-01 -9.92223084e-01 -3.86481762e-01 4.12433445e-01
-5.33136547e-01 -4.40426171e-01 1.23097491e+00 -1.42855301e-01
4.82648969e-01 -2.22175312e+00 -2.78987885e-01 8.12331513e-02
2.78825104e-01 3.43743920e-01 7.23475888e-02 1.34373521e-02
-6.18886016e-02 -1.79812565e-01 -4.30762917e-01 -5.04128158e-01
-1.95127651e-01 2.48733118e-01 -2.54630893e-01 6.42582476e-01
5.15060186e-01 8.75854373e-01 -9.83420372e-01 -7.28776038e-01
7.12397873e-01 5.00530958e-01 -3.41881692e-01 4.26580489e-01
-4.50046420e-01 4.91592914e-01 -7.16378987e-01 1.17852771e+00
9.67643797e-01 -4.53632414e-01 -4.56829041e-01 -8.63189176e-02
-2.36915872e-01 3.21359992e-01 -1.15391123e+00 1.68689287e+00
-1.51092082e-01 4.10922229e-01 4.09695692e-02 -6.65331542e-01
1.04090369e+00 -8.37538019e-02 3.12922060e-01 -1.19050905e-01
2.52430439e-01 1.75171748e-01 -3.39816570e-01 -1.43834829e-01
2.70927399e-01 -1.42463222e-02 3.52627963e-01 2.78014809e-01
-7.94296265e-02 -7.30871320e-01 -8.02748576e-02 3.31622303e-01
8.60232234e-01 5.14460802e-01 4.06777829e-01 -2.49695405e-02
4.84297514e-01 4.54385936e-01 5.33557951e-01 6.83433473e-01
-3.05715799e-01 6.41983151e-01 2.16199365e-03 -2.80580163e-01
-9.73245561e-01 -1.27197886e+00 -2.62630254e-01 8.32148731e-01
4.37496483e-01 -1.34499177e-01 -5.14143944e-01 -9.54527557e-01
2.87575632e-01 6.65740550e-01 -3.01239878e-01 6.89179972e-02
-3.90603215e-01 -3.80823016e-01 1.43420070e-01 7.88597465e-01
4.61523175e-01 -7.93994367e-01 -7.56232977e-01 -1.67947710e-01
1.83755636e-01 -1.35927832e+00 -1.56176239e-01 4.85394865e-01
-1.19719923e+00 -1.07715690e+00 -8.19985688e-01 -7.61467934e-01
1.21523559e+00 8.57521296e-01 8.44684422e-01 4.74812090e-03
-3.44620109e-01 3.35943341e-01 -5.12605309e-01 -9.60894227e-01
-1.63089037e-01 -3.02389432e-02 1.59263745e-01 -2.68640101e-01
7.53448069e-01 -4.48919088e-01 -5.87053597e-01 4.24791425e-01
-6.80782259e-01 2.41749406e-01 8.53335798e-01 5.23812175e-01
1.09287894e+00 -2.77777076e-01 6.04397133e-02 -8.92565429e-01
-1.98156998e-01 -6.24597967e-02 -1.13181114e+00 -8.80912691e-02
-4.12257433e-01 -2.51095146e-01 1.26644582e-01 -4.93548721e-01
-1.12328935e+00 6.82655573e-01 1.59835830e-01 -1.12040782e+00
-5.75834095e-01 -5.66770956e-02 -2.77548045e-01 -4.04340059e-01
6.55090690e-01 1.95294484e-01 -2.42725402e-01 -6.96180999e-01
4.37670380e-01 6.00091040e-01 5.06004572e-01 -4.73046005e-01
1.45742393e+00 9.36959267e-01 -7.16856942e-02 -8.61676276e-01
-1.19298685e+00 -9.84970391e-01 -9.07974541e-01 -1.89520419e-01
8.02451551e-01 -1.11137247e+00 -2.12275311e-01 2.80015439e-01
-1.06352007e+00 -3.80138904e-01 -5.42828739e-01 5.03037930e-01
-3.91913801e-01 2.28644103e-01 -2.63888717e-01 -1.04585207e+00
-1.94103241e-01 -9.51821625e-01 1.38641775e+00 3.12976807e-01
3.06979507e-01 -4.58293587e-01 -1.21888131e-01 4.05822903e-01
-1.29351526e-01 2.31777832e-01 6.20578527e-01 -7.05891728e-01
-9.75281477e-01 -4.72510248e-01 -6.03565156e-01 3.42278123e-01
1.20788425e-01 -7.13505074e-02 -1.16228342e+00 -1.44592449e-01
-9.81760174e-02 -4.07713205e-01 7.67479181e-01 1.39876857e-01
1.33212101e+00 2.02152416e-01 -4.75791574e-01 6.62300944e-01
1.20208943e+00 -1.43458873e-01 1.15256615e-01 2.66007185e-01
1.01562202e+00 4.03776079e-01 1.00638890e+00 3.46298844e-01
2.72990257e-01 4.45072532e-01 6.81825340e-01 -4.54433173e-01
-1.05088919e-01 -4.30230230e-01 1.97319575e-02 4.13185120e-01
-1.65380925e-01 1.51705578e-01 -1.26477003e+00 5.81515253e-01
-1.76358354e+00 -4.07791257e-01 -3.47960234e-01 2.07130647e+00
5.80968857e-01 4.54261929e-01 1.35352492e-01 1.23324230e-01
7.11765468e-01 9.06756800e-03 -8.45187724e-01 4.45938259e-01
-6.46806927e-03 1.05846666e-01 7.08359480e-01 8.09498578e-02
-1.32419360e+00 1.14267838e+00 4.82412720e+00 6.27679527e-01
-8.15842569e-01 2.32739553e-01 2.76973784e-01 -3.89660001e-02
2.34895602e-01 8.46998170e-02 -9.39085603e-01 -2.68523898e-02
1.46325275e-01 5.68688139e-02 -1.61446840e-01 1.34491146e+00
1.23766609e-01 -1.65319875e-01 -1.12750018e+00 1.15556359e+00
-1.07484393e-01 -9.69353318e-01 -2.63042897e-02 -4.81367111e-02
7.62567461e-01 4.23218489e-01 -1.07119456e-01 2.14431584e-01
2.75063753e-01 -3.52171898e-01 7.43996263e-01 -5.17053865e-02
5.77044070e-01 -5.41566074e-01 4.83578205e-01 5.14594495e-01
-1.38795662e+00 1.62227914e-01 -4.61159080e-01 3.74981090e-02
1.14322312e-01 5.12534797e-01 -1.07828081e+00 2.91559547e-01
9.10812080e-01 6.62126839e-01 -7.78635204e-01 1.17192769e+00
-6.68168545e-01 6.14451170e-01 -7.42579758e-01 3.55004915e-03
2.92167723e-01 -5.79163805e-02 8.65403831e-01 1.09176445e+00
1.11632705e-01 5.14833093e-01 5.49172580e-01 1.00663066e+00
-5.44563495e-02 1.00846395e-01 -5.58509707e-01 4.31967862e-02
5.62397420e-01 1.37085199e+00 -1.12115431e+00 -3.74676168e-01
-5.65361917e-01 7.70491362e-01 3.36430192e-01 2.12159470e-01
-6.47483587e-01 -2.73384899e-01 5.72892725e-01 3.33255053e-01
4.40959305e-01 -4.60639775e-01 -3.45911115e-01 -1.16675031e+00
1.51294293e-02 -3.56322140e-01 4.15598571e-01 -9.26037788e-01
-1.19647419e+00 3.46344799e-01 3.17217290e-01 -1.52304852e+00
3.59909266e-01 -7.14258790e-01 -4.92876738e-01 5.36896706e-01
-1.73965859e+00 -1.41265726e+00 -8.44407618e-01 5.15810490e-01
7.08034456e-01 2.05853522e-01 4.77616996e-01 7.12878779e-02
-4.24554616e-01 2.03135148e-01 -2.40907982e-01 1.96377590e-01
4.48890179e-01 -1.35099876e+00 5.37666976e-01 9.11873698e-01
3.89820218e-01 3.59832704e-01 5.23083568e-01 -6.96142197e-01
-1.33686531e+00 -1.32266068e+00 2.39456639e-01 -7.12921679e-01
5.32502770e-01 -6.15064800e-01 -1.20463181e+00 6.24149680e-01
-2.64152408e-01 5.25590003e-01 4.68134493e-01 -9.79390964e-02
-4.18015122e-01 1.44581841e-02 -1.07196236e+00 4.14471716e-01
1.17078590e+00 -4.73487407e-01 -5.73500395e-01 6.22591019e-01
8.67442727e-01 -7.75740862e-01 -4.37650293e-01 6.64665937e-01
4.89878990e-02 -7.01445639e-01 1.14876533e+00 -4.27088261e-01
1.03333555e-01 -8.22528303e-01 -9.85800549e-02 -8.68225694e-01
-5.28422706e-02 -2.15103999e-01 -3.82769316e-01 1.02759945e+00
2.25603819e-01 -4.30469990e-01 1.13041878e+00 4.83696580e-01
-2.47113645e-01 -6.22814298e-01 -7.67268300e-01 -9.76365685e-01
-3.18297118e-01 -5.52930355e-01 4.34998453e-01 9.39827800e-01
-4.92154330e-01 2.27546766e-01 1.60743311e-01 6.62150085e-01
1.12241066e+00 3.06804538e-01 1.30602551e+00 -1.33922660e+00
6.36003464e-02 -2.92569071e-01 -5.58696568e-01 -1.20984566e+00
3.95674370e-02 -8.42491984e-01 2.81026155e-01 -1.45791674e+00
1.73023805e-01 -7.67923295e-01 -2.60886520e-01 7.16660261e-01
-3.73940110e-01 5.24798095e-01 1.84462726e-01 4.33546185e-01
-6.62911952e-01 6.26425743e-01 1.13335073e+00 -8.52405280e-02
-5.09106874e-01 2.36236185e-01 -2.19591856e-01 1.07756698e+00
8.39579225e-01 -6.24303699e-01 -2.60379374e-01 -3.55354905e-01
-4.27957118e-01 -4.65413958e-01 6.43133938e-01 -9.64925587e-01
1.92090809e-01 -2.23769620e-01 4.63944077e-01 -1.28833187e+00
5.55148840e-01 -9.97619331e-01 -3.29470217e-01 3.92416507e-01
-2.39954293e-02 -4.14115787e-01 3.38275194e-01 7.01491117e-01
5.66822551e-02 -1.02711856e-01 9.13861096e-01 -2.03000084e-01
-8.50555658e-01 5.69854498e-01 2.25225210e-01 -2.29446381e-01
1.32306290e+00 -2.56636947e-01 -3.41217220e-02 -4.95789247e-03
-4.16144133e-01 3.67564112e-01 6.28921092e-01 4.52015638e-01
8.15244198e-01 -1.11203289e+00 -5.64194143e-01 2.46966615e-01
6.38148010e-01 9.20223355e-01 -8.32688212e-02 6.63033664e-01
-5.09534955e-01 2.84767509e-01 1.45181254e-01 -1.15425336e+00
-1.43770862e+00 6.11213624e-01 -2.85338541e-03 2.24653602e-01
-8.58468950e-01 1.08476019e+00 5.06517529e-01 -5.81910491e-01
4.57046211e-01 -5.69862366e-01 1.68365851e-01 -2.91963786e-01
3.41800123e-01 1.73801869e-01 -2.73311343e-02 -5.35859168e-01
-5.39288402e-01 5.63336670e-01 -2.19384745e-01 2.39341483e-01
1.31304789e+00 4.88254847e-03 1.59863740e-01 5.41523516e-01
8.55745256e-01 -2.76018858e-01 -1.46968448e+00 -7.46720970e-01
2.80743185e-02 -7.95078099e-01 3.30867767e-01 -3.64728540e-01
-8.69640172e-01 9.45489228e-01 7.24542141e-01 -1.03866562e-01
9.31798995e-01 5.19572496e-01 3.50554526e-01 6.52362943e-01
4.69905972e-01 -6.28032029e-01 1.27994716e-01 2.81912893e-01
6.09190702e-01 -1.67975843e+00 3.13384861e-01 -8.59381437e-01
-4.41031516e-01 8.51652443e-01 1.11082268e+00 -1.53115362e-01
5.58461368e-01 2.20323354e-01 2.92288363e-01 -3.67399633e-01
-3.11453342e-01 -6.71484947e-01 4.11726058e-01 6.19077682e-01
1.03129679e-02 -3.37009355e-02 2.99875200e-01 2.06590563e-01
1.75975591e-01 -2.12232575e-01 1.19606815e-01 1.34314823e+00
-7.00716913e-01 -8.46261084e-01 -7.13632822e-01 3.70811760e-01
7.77356401e-02 1.05443776e-01 -5.49784780e-01 9.64053631e-01
-4.21893932e-02 5.77245414e-01 4.79978919e-02 -5.68970516e-02
4.93940920e-01 4.00660969e-02 5.13218999e-01 -1.08247638e+00
-8.94776508e-02 3.32687527e-01 -3.33695233e-01 -4.79674727e-01
-6.60957277e-01 -7.60696054e-01 -1.54375613e+00 1.35176912e-01
-8.10734034e-01 -4.17451598e-02 8.44871938e-01 6.54363096e-01
1.25713602e-01 3.77610207e-01 5.10290861e-01 -1.34399152e+00
-3.87692571e-01 -8.80378008e-01 -2.64785945e-01 2.62715042e-01
3.52360249e-01 -9.68124688e-01 -2.82489598e-01 4.37585786e-02] | [7.883354187011719, -2.9079480171203613] |
79785774-7c88-4bf4-a880-81804c64782a | augmented-memory-capitalizing-on-experience | 2305.16160 | null | https://arxiv.org/abs/2305.16160v1 | https://arxiv.org/pdf/2305.16160v1.pdf | Augmented Memory: Capitalizing on Experience Replay to Accelerate De Novo Molecular Design | Sample efficiency is a fundamental challenge in de novo molecular design. Ideally, molecular generative models should learn to satisfy a desired objective under minimal oracle evaluations (computational prediction or wet-lab experiment). This problem becomes more apparent when using oracles that can provide increased predictive accuracy but impose a significant cost. Consequently, these oracles cannot be directly optimized under a practical budget. Molecular generative models have shown remarkable sample efficiency when coupled with reinforcement learning, as demonstrated in the Practical Molecular Optimization (PMO) benchmark. Here, we propose a novel algorithm called Augmented Memory that combines data augmentation with experience replay. We show that scores obtained from oracle calls can be reused to update the model multiple times. We compare Augmented Memory to previously proposed algorithms and show significantly enhanced sample efficiency in an exploitation task and a drug discovery case study requiring both exploration and exploitation. Our method achieves a new state-of-the-art in the PMO benchmark which enforces a computational budget, outperforming the previous best performing method on 19/23 tasks. | ['Philippe Schwaller', 'Jeff Guo'] | 2023-05-10 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 5.83538234e-01 1.09638847e-01 -4.06881660e-01 -2.15177551e-01
-1.19935799e+00 -7.17141747e-01 3.30727220e-01 3.50240737e-01
-5.35467029e-01 1.37620282e+00 -2.95034677e-01 -5.90401232e-01
-3.36960554e-01 -7.21775115e-01 -1.19215274e+00 -9.19773161e-01
-2.54985809e-01 9.25838292e-01 -1.71911851e-01 -8.52270611e-03
3.78899783e-01 5.37204564e-01 -1.23674154e+00 8.61575454e-02
1.09406602e+00 7.73108780e-01 2.88380563e-01 7.71940947e-01
3.27978998e-01 3.75118732e-01 -5.17557681e-01 -3.97944480e-01
2.67021835e-01 -4.46439415e-01 -7.48929381e-01 -1.60806656e-01
2.19367757e-01 -2.60931373e-01 7.70834386e-02 8.29980671e-01
1.02113628e+00 3.45263213e-01 4.07841742e-01 -8.28295410e-01
-3.64094108e-01 5.63084722e-01 -8.25569779e-03 -6.45005330e-02
3.05181563e-01 5.04572392e-01 1.07839119e+00 -7.20525801e-01
8.28785956e-01 8.96964610e-01 4.78977621e-01 5.70265710e-01
-1.75503159e+00 -6.00070477e-01 4.35567684e-02 2.63773769e-01
-1.07818794e+00 -5.07392526e-01 3.13737392e-01 -2.47825444e-01
1.46700907e+00 3.87488693e-01 5.13614476e-01 1.10544288e+00
2.28685811e-01 7.13265896e-01 1.10397065e+00 -1.27812490e-01
8.77030611e-01 1.88732117e-01 -1.40512183e-01 8.01541328e-01
3.43315780e-01 5.05627096e-01 -7.06534088e-01 -4.22267288e-01
2.57360816e-01 -1.61259957e-02 -2.92225122e-01 -7.18567550e-01
-9.99472976e-01 1.07593703e+00 2.93145210e-01 -2.91612536e-01
-4.79967952e-01 3.52805912e-01 1.89565271e-01 4.83177692e-01
1.66307047e-01 1.37082946e+00 -8.08738947e-01 -3.60154092e-01
-8.55880737e-01 4.01359111e-01 1.06282485e+00 6.27476454e-01
7.14418530e-01 3.57792042e-02 -1.07907824e-01 4.31903124e-01
1.10187987e-03 1.96851894e-01 3.90637964e-01 -7.78793037e-01
2.96683609e-01 3.28440905e-01 2.73111999e-01 -2.12942660e-01
-5.41800082e-01 -1.07543027e+00 -5.53564131e-01 3.90462935e-01
2.20002577e-01 -1.79297045e-01 -1.02231622e+00 1.81880593e+00
6.54616714e-01 -1.24448808e-02 3.45377356e-01 5.17794013e-01
5.43703198e-01 6.36414111e-01 3.12553085e-02 -5.88497400e-01
1.05262983e+00 -1.07599854e+00 -6.09701872e-01 -1.48509532e-01
9.35434461e-01 -6.26336396e-01 1.03638232e+00 8.45409334e-01
-1.13776839e+00 -1.56405360e-01 -1.42405260e+00 2.25745186e-01
-3.95184338e-01 -3.40583175e-02 1.21458840e+00 1.06189871e+00
-6.72680259e-01 1.22486639e+00 -1.08777511e+00 1.38934866e-01
6.33410573e-01 1.02759290e+00 -3.24049234e-01 4.99133617e-02
-8.30860078e-01 8.02995086e-01 5.74717402e-01 -1.64548486e-01
-1.50337589e+00 -1.11981952e+00 -4.85772371e-01 1.43441096e-01
8.91895652e-01 -8.88447046e-01 1.14769948e+00 -4.16444033e-01
-2.01131511e+00 3.39852720e-01 -3.18973255e-03 -8.69856060e-01
5.61139464e-01 -3.99319023e-01 -2.54048612e-02 -1.86458483e-01
-3.17953467e-01 6.54578209e-01 5.33609927e-01 -8.08172047e-01
-1.78467840e-01 -2.70930082e-01 1.15612268e-01 1.41267151e-01
-2.11478367e-01 -4.32508945e-01 -4.52393480e-02 -4.73477632e-01
-1.09255247e-01 -1.14564395e+00 -7.37545788e-01 -3.87575090e-01
-4.98720139e-01 -1.00485444e-01 2.44863227e-01 -3.69493216e-01
1.27533138e+00 -1.50316417e+00 4.84924167e-01 2.61539012e-01
1.38112962e-01 2.19359905e-01 -1.53949365e-01 5.07096648e-01
-2.11540967e-01 1.79908425e-01 -3.14150393e-01 -2.33138710e-01
-5.44139324e-03 2.32433498e-01 -2.88805664e-01 2.73446977e-01
2.39464603e-02 1.00544059e+00 -9.09943640e-01 -5.25211655e-02
-1.74884617e-01 2.19140291e-01 -1.27590585e+00 3.07896852e-01
-7.78283000e-01 5.16368747e-01 -3.47309113e-01 7.27148652e-01
3.26378793e-01 -7.71914303e-01 4.37114596e-01 3.10040638e-02
9.97559205e-02 3.35458785e-01 -1.11858594e+00 1.94965041e+00
-3.38695049e-01 6.73488304e-02 -5.26799619e-01 -9.42035258e-01
6.04159474e-01 2.80853808e-01 3.94967526e-01 -6.12088561e-01
-2.10465491e-01 3.89913678e-01 3.44259977e-01 -2.06785753e-01
1.98936611e-01 -5.68089746e-02 3.38651717e-01 4.28090364e-01
2.97824264e-01 -3.24535400e-01 4.42539275e-01 -1.59061059e-01
1.26921427e+00 4.90663111e-01 5.18562555e-01 -1.62837923e-01
1.82465419e-01 1.33541554e-01 5.31361997e-01 1.09218037e+00
2.34774172e-01 1.52879089e-01 6.18711233e-01 -6.05537474e-01
-1.14623773e+00 -7.88210869e-01 -1.86732918e-01 1.18613398e+00
-3.63719612e-01 -6.83123589e-01 -7.13208675e-01 -7.47250855e-01
-5.88856675e-02 6.88588858e-01 -6.60692036e-01 -3.03356916e-01
-3.76057744e-01 -1.43228590e+00 2.15291843e-01 1.85776725e-01
-3.88115905e-02 -8.80114198e-01 -6.26570165e-01 6.05822146e-01
3.88454467e-01 -6.89256549e-01 -3.42427731e-01 9.11182463e-01
-1.12300944e+00 -9.37020838e-01 -5.07346749e-01 -2.46083811e-01
6.42959237e-01 -2.59563923e-01 1.09384537e+00 -2.04314038e-01
-5.17220736e-01 -1.39061034e-01 -1.42084556e-02 -4.27732140e-01
-4.68648016e-01 4.73510265e-01 1.33721590e-01 -3.55309695e-01
-7.11462051e-02 -7.98192143e-01 -8.51436377e-01 2.31957316e-01
-8.32657814e-01 1.34684086e-01 9.44167376e-01 1.48856688e+00
1.00494647e+00 -2.69322336e-01 6.78130209e-01 -1.11263180e+00
3.95642966e-01 -3.05908382e-01 -1.00400567e+00 3.10330629e-01
-1.14823556e+00 8.51981282e-01 6.32478356e-01 -5.29978573e-01
-6.65267706e-01 3.75593990e-01 -2.89271146e-01 -1.79322153e-01
2.52664387e-01 6.54710054e-01 -2.58803666e-01 -4.19408113e-01
9.46267009e-01 2.41680905e-01 -3.83980423e-02 -5.95757067e-01
2.71062911e-01 1.10868916e-01 7.81866312e-02 -7.87415147e-01
4.72600460e-01 1.72175139e-01 6.17994547e-01 -4.67803299e-01
-8.38950455e-01 -1.28806025e-01 -2.05168158e-01 3.91816288e-01
3.06431025e-01 -7.84814775e-01 -1.19914544e+00 -2.16772556e-01
-8.34523201e-01 -4.77147371e-01 -4.23088789e-01 5.69033444e-01
-8.46311867e-01 3.41100603e-01 -2.25534365e-01 -6.37977898e-01
-5.26910901e-01 -1.52364635e+00 8.23371053e-01 5.77963591e-02
-2.18777627e-01 -7.50599980e-01 2.83103645e-01 3.47975463e-01
4.26531643e-01 4.17802334e-01 1.02166855e+00 -9.26325917e-01
-9.92072523e-01 -9.07354057e-02 3.89965177e-01 4.31647301e-02
-8.23404361e-03 -3.84787947e-01 -8.38362157e-01 -6.38680518e-01
-2.66133010e-01 -6.12198710e-01 1.05292439e+00 3.13903391e-01
1.30869043e+00 -5.96882284e-01 -2.93915838e-01 9.31612670e-01
1.27180612e+00 5.87656319e-01 4.75267172e-01 3.00143957e-01
2.87520021e-01 1.75868630e-01 5.45743525e-01 7.30435252e-01
-3.34545851e-01 9.41660881e-01 5.44893444e-01 2.47574365e-03
2.56398201e-01 -3.06022793e-01 2.87430257e-01 2.67115593e-01
-2.04857498e-01 -3.20083916e-01 -6.66271985e-01 1.76869437e-01
-1.85582888e+00 -7.83206880e-01 3.69561285e-01 2.58442640e+00
1.35938048e+00 2.82397598e-01 5.45065105e-02 -4.60963100e-02
-8.84409100e-02 -2.26660311e-01 -1.23550022e+00 -3.73525381e-01
-7.29680955e-02 9.09322917e-01 6.52226269e-01 6.31880224e-01
-9.39679325e-01 8.82178009e-01 6.97230911e+00 1.13279355e+00
-9.44027781e-01 1.97023928e-01 9.28711593e-01 -5.12859225e-01
-3.68277401e-01 2.25410789e-01 -9.27398443e-01 2.44580537e-01
1.26961386e+00 -1.88617334e-01 7.41881073e-01 1.18687320e+00
-6.67530969e-02 1.14561640e-01 -1.54242373e+00 8.94623101e-01
-2.59657323e-01 -1.83140278e+00 4.77912910e-02 3.83041084e-01
9.84599590e-01 -6.56476095e-02 4.31501955e-01 2.65377760e-01
4.06677753e-01 -1.48016775e+00 1.88070923e-01 2.48666018e-01
6.25386715e-01 -9.08237517e-01 4.90769625e-01 1.59216508e-01
-5.21355212e-01 4.15540189e-02 -3.88755083e-01 2.55098999e-01
1.51606649e-02 4.61067140e-01 -1.32427001e+00 4.11832720e-01
1.06428608e-01 3.02056789e-01 -3.89646620e-01 1.15009177e+00
-1.00526772e-01 5.60467124e-01 -5.14457345e-01 -4.29095358e-01
2.69703418e-01 -2.18469262e-01 5.65006495e-01 8.65107417e-01
2.86491871e-01 -1.06091455e-01 5.71915321e-02 9.48335826e-01
-3.08388770e-01 2.24174470e-01 -4.19446468e-01 -4.35292661e-01
2.14603171e-01 9.63783324e-01 -4.56454903e-01 -3.80356282e-01
2.07142249e-01 9.54108894e-01 2.63140529e-01 2.99366504e-01
-8.77107620e-01 -2.95178503e-01 6.81765556e-01 -1.48087487e-01
5.00281215e-01 -6.64453357e-02 -1.41182795e-01 -8.61279070e-01
-2.14913115e-01 -1.17509747e+00 3.69584531e-01 -2.62760639e-01
-7.43064344e-01 4.73997682e-01 -2.98665136e-01 -9.01814163e-01
-3.16169322e-01 -8.24649692e-01 -1.42863905e-02 5.73160410e-01
-1.34989262e+00 -5.92322767e-01 4.90152314e-02 4.58519943e-02
4.68196392e-01 -2.84525901e-01 1.11474526e+00 3.29355121e-01
-6.57237470e-01 8.49460185e-01 4.87128347e-01 -8.33262444e-01
5.54044068e-01 -1.44640446e+00 2.65180081e-01 5.67858279e-01
2.39448190e-01 8.29617262e-01 9.23273981e-01 -5.47393203e-01
-1.92574596e+00 -8.76809716e-01 4.44465160e-01 -4.88712043e-01
4.13014621e-01 -3.87135297e-01 -6.52883887e-01 3.95170599e-01
-1.29078850e-01 -1.65750638e-01 1.00745726e+00 2.65290558e-01
-2.13621154e-01 -1.15038902e-01 -9.90609050e-01 6.46199763e-01
1.23814964e+00 -2.04979911e-01 -4.70217764e-02 9.97851372e-01
9.95429039e-01 -7.51575887e-01 -1.13619554e+00 5.26194394e-01
6.50498331e-01 -5.97786784e-01 1.16276860e+00 -1.32117295e+00
3.78991187e-01 -2.00263545e-01 -2.08689213e-01 -1.26600671e+00
6.17868304e-02 -1.44032300e+00 -5.38824916e-01 3.85981560e-01
9.45042074e-01 -7.13856220e-01 1.13733470e+00 5.68849444e-01
-7.83599392e-02 -1.23800826e+00 -1.01676702e+00 -1.07056689e+00
-1.04790047e-01 -8.92324895e-02 7.91518509e-01 6.42195046e-01
8.10027122e-03 3.77345026e-01 -6.39708161e-01 5.11323176e-02
6.61532819e-01 1.62057161e-01 8.65679741e-01 -7.94622600e-01
-1.16210341e+00 -4.51437056e-01 -2.43335471e-01 -1.17460024e+00
-1.91031039e-01 -8.91551852e-01 -2.46782765e-01 -9.86863196e-01
3.60998124e-01 -3.39655578e-01 -4.41626608e-01 5.38529575e-01
-3.34377915e-01 1.02029262e-04 -1.61425337e-01 -1.64936129e-02
-6.83362126e-01 7.69209504e-01 1.15681291e+00 -1.53367937e-01
-5.64201474e-01 5.09049036e-02 -6.84706926e-01 1.52247608e-01
6.79960966e-01 -6.09786749e-01 -3.81507248e-01 6.53845444e-02
7.59847403e-01 2.03190178e-01 3.65628935e-02 -1.03857899e+00
1.84013113e-01 -2.18912259e-01 3.01291376e-01 -4.28974569e-01
4.58094448e-01 -4.85958070e-01 5.91210008e-01 1.02248371e+00
-4.21542615e-01 -3.29677053e-02 1.96918741e-01 7.38216162e-01
1.54747799e-01 -2.09246114e-01 6.53640509e-01 -1.58240478e-02
-1.41762957e-01 5.27469754e-01 1.17955534e-02 -2.84705222e-01
1.01377797e+00 -9.64005291e-02 -3.15476835e-01 -4.10883017e-02
-1.02398145e+00 1.73031688e-01 2.99122363e-01 -2.07691267e-02
5.71371853e-01 -1.02917576e+00 -4.90578204e-01 1.63450152e-01
1.25567243e-01 -2.39057109e-01 1.05982579e-01 8.41299653e-01
-5.57914019e-01 7.09790409e-01 3.22081633e-02 -4.59249616e-01
-1.27031016e+00 8.06577981e-01 5.18977463e-01 -7.82686532e-01
-1.09262995e-01 9.50073779e-01 9.64194983e-02 -3.62640440e-01
2.69202322e-01 -2.69174784e-01 3.26614588e-01 -1.77689239e-01
3.40892136e-01 3.12855542e-01 3.57091576e-01 4.51318532e-01
-2.51273245e-01 1.98347345e-01 -3.52841586e-01 5.10306284e-02
1.72569752e+00 5.86786926e-01 1.74508661e-01 -6.02568127e-02
1.05026615e+00 -1.87287003e-01 -1.40956688e+00 -1.88666686e-01
4.13174480e-02 -3.15265536e-01 7.33721778e-02 -1.23540246e+00
-6.06675684e-01 7.04596817e-01 9.03668225e-01 -3.10833722e-01
8.70531619e-01 -2.46069938e-01 6.15316093e-01 1.16411030e+00
3.62245202e-01 -1.08680952e+00 2.63289750e-01 1.43533185e-01
8.15586269e-01 -1.34128129e+00 3.81990403e-01 -1.30462319e-01
-4.44642961e-01 1.02194846e+00 3.72468889e-01 1.82326600e-01
3.10414314e-01 8.38589892e-02 -6.95974171e-01 -2.04088464e-01
-1.17017281e+00 -3.62416841e-02 3.10406744e-01 3.31086874e-01
3.38218093e-01 1.93181664e-01 -2.91534632e-01 4.26042736e-01
-1.72111213e-01 -1.01465903e-01 1.32920444e-01 9.66012180e-01
-5.31850576e-01 -1.67295265e+00 2.46639419e-02 4.67767626e-01
-6.06478035e-01 -4.30302143e-01 -2.94400692e-01 5.00904560e-01
-1.65674344e-01 6.11412048e-01 -5.18163979e-01 -1.21341370e-01
2.24804096e-02 1.51424050e-01 7.66628385e-01 -5.18287122e-01
-5.52571476e-01 3.63703594e-02 2.44539738e-01 -8.53413105e-01
-1.13565639e-01 -4.02899355e-01 -9.56251442e-01 -2.47425795e-01
-5.55777371e-01 5.11058450e-01 8.76829267e-01 7.64528930e-01
8.30308020e-01 6.48004055e-01 5.26867151e-01 -5.18113375e-01
-9.81157780e-01 -7.19891608e-01 -1.47674844e-01 -9.06617474e-03
2.68862486e-01 -5.82525432e-01 -9.63116363e-02 -7.27452785e-02] | [5.096254348754883, 5.386468887329102] |
9b629fcc-8e87-4543-a168-8a621e86ef02 | viskositas-viscosity-prediction-of | 2208.01440 | null | https://arxiv.org/abs/2208.01440v5 | https://arxiv.org/pdf/2208.01440v5.pdf | Viskositas: Viscosity Prediction of Multicomponent Chemical Systems | Viscosity in the metallurgical and glass industry plays a fundamental role in its production processes, also in the area of geophysics. As its experimental measurement is financially expensive, also in terms of time, several mathematical models were built to provide viscosity results as a function of several variables, such as chemical composition and temperature, in linear and nonlinear models. A database was built in order to produce a nonlinear model by artificial neural networks by variation of hyperparameters to provide reliable predictions of viscosity in relation to chemical systems and temperatures. The model produced named Viskositas demonstrated better statistical evaluations of mean absolute error, standard deviation and coefficient of determination in relation to the test database when compared to different models from literature and 1 commercial model, offering predictions with lower errors, less variability and less generation of outliers. | ['Patrick dos Anjos'] | 2022-08-02 | null | null | null | null | ['geophysics'] | ['miscellaneous'] | [-2.56961167e-01 -2.49728132e-02 1.03658102e-02 -3.71128023e-01
1.90663099e-01 -2.95976400e-01 4.44134980e-01 6.28115118e-01
-3.44241291e-01 9.56100285e-01 -6.77265450e-02 -3.74258995e-01
-5.36184728e-01 -1.06181252e+00 -4.79184121e-01 -8.00785959e-01
-8.99095908e-02 6.84332073e-01 2.73339570e-01 -2.64114082e-01
6.78742826e-01 8.35144162e-01 -1.24334288e+00 -2.07897738e-01
9.71347272e-01 8.51667762e-01 6.00645319e-02 5.78793347e-01
-2.63486266e-01 6.57123148e-01 -6.30916655e-01 -1.79355647e-02
1.01504512e-01 -3.32367033e-01 -6.04472041e-01 -6.22315966e-02
-1.46054447e-01 -1.23794928e-01 6.79932237e-02 8.06735635e-01
-1.47739742e-02 3.81172180e-01 1.19387102e+00 -7.99843669e-01
-6.62738562e-01 7.29599595e-01 -1.23001434e-01 -2.48209581e-01
1.68303564e-01 2.20098153e-01 4.53118473e-01 -3.49172920e-01
4.69279975e-01 9.58058655e-01 3.57223898e-01 6.48638792e-03
-1.42932606e+00 -3.97228807e-01 -2.94026732e-01 7.81400576e-02
-8.82223845e-01 1.39231667e-01 7.81228065e-01 -5.03391564e-01
9.85165775e-01 2.22680002e-01 8.82383168e-01 6.28619313e-01
7.10329294e-01 -1.73114628e-01 1.23406482e+00 -5.83081782e-01
4.24838811e-01 5.19581139e-01 1.96459457e-01 2.64649719e-01
5.33143103e-01 7.13106170e-02 4.89732623e-02 1.76988319e-01
8.74911129e-01 -4.14924860e-01 -2.30344068e-02 2.39485744e-02
-5.71193695e-01 1.03109229e+00 1.75008565e-01 7.75674403e-01
-2.44729623e-01 -1.53340578e-01 3.97108436e-01 5.57304502e-01
3.05593580e-01 9.82257187e-01 -7.34399438e-01 -1.94937035e-01
-7.00038493e-01 2.86510408e-01 1.28567636e+00 4.40752894e-01
4.71688181e-01 5.64789176e-01 6.14031017e-01 7.67739356e-01
5.98468184e-01 6.98026001e-01 9.58854556e-01 -4.93439972e-01
2.93390840e-01 7.79941678e-01 -1.32825270e-01 -9.56768692e-01
-8.41579437e-01 -3.73976678e-02 -5.62435031e-01 5.03149331e-01
5.85860252e-01 -5.86765334e-02 -8.87372077e-01 1.03050601e+00
2.94939816e-01 -3.69429260e-01 2.95740068e-01 7.81608403e-01
7.83044517e-01 9.32938755e-01 1.33947581e-01 -3.28049332e-01
7.15845644e-01 -5.37356794e-01 -6.60550296e-01 2.17554927e-01
3.64850640e-01 -8.19576442e-01 9.20305431e-01 7.17116714e-01
-9.37085330e-01 -5.80870390e-01 -1.25717294e+00 2.35466644e-01
-7.01180995e-01 -2.09150061e-01 5.32072544e-01 8.81363213e-01
-4.57164496e-01 1.38352942e+00 -8.98756027e-01 -2.97383040e-01
-3.74859571e-01 2.68872231e-01 -3.81432980e-01 6.99957430e-01
-1.00861633e+00 1.20468211e+00 7.81293929e-01 4.12050784e-01
-2.63069361e-01 -4.87879038e-01 -8.25159311e-01 5.75650297e-03
-6.85216337e-02 6.89526927e-03 9.10774469e-01 -4.18382645e-01
-2.02001214e+00 2.26091608e-01 2.80333430e-01 -3.62955332e-01
6.85912669e-01 -1.25556588e-01 -7.32523322e-01 8.96560997e-02
-4.38101768e-01 -2.81183906e-02 5.94033301e-01 -1.08927333e+00
-1.73242778e-01 -9.44615677e-02 -3.18153292e-01 -4.56903458e-01
-6.46115979e-03 -3.34949732e-01 3.36131632e-01 -2.14986652e-01
6.24038875e-01 -7.55571663e-01 -2.37831786e-01 -5.66599429e-01
-2.94230103e-01 -3.71844769e-02 7.13420272e-01 -9.89985049e-01
8.06146801e-01 -1.68089366e+00 4.51906361e-02 9.02535021e-01
-1.31002054e-01 1.67323351e-01 4.56312805e-01 6.34237826e-01
-3.82457078e-01 1.97529420e-01 -2.50440955e-01 2.84298062e-01
3.54690365e-02 4.97623011e-02 1.01589710e-01 4.37380612e-01
4.06904101e-01 2.00828493e-01 -7.11907446e-01 -2.80155003e-01
5.90491712e-01 3.43602717e-01 -2.51718491e-01 5.22638895e-02
-3.22153926e-01 2.92109579e-01 -3.77165407e-01 7.73555517e-01
7.13406563e-01 2.86618650e-01 1.40285581e-01 -1.57909483e-01
-3.13462645e-01 -9.95862857e-02 -1.33470821e+00 8.85417819e-01
-5.59346616e-01 4.20857847e-01 -2.97069907e-01 -1.06767797e+00
1.86200142e+00 4.01188970e-01 5.51921129e-01 -8.06673050e-01
6.54983640e-01 4.93242860e-01 1.44880012e-01 -8.52838457e-01
6.82666123e-01 -4.24332380e-01 4.58295077e-01 -2.90124286e-02
-6.56354874e-02 -6.76607728e-01 5.57664335e-01 -3.08877140e-01
3.06838661e-01 2.98052818e-01 -1.10697091e-01 -3.25242132e-01
6.98465884e-01 -2.22020879e-01 1.28660619e-01 2.80564517e-01
2.35213608e-01 3.42372298e-01 8.90259385e-01 -3.78505200e-01
-1.49033916e+00 -7.23213315e-01 -5.62291563e-01 2.08589584e-01
9.78436843e-02 2.39823014e-01 -3.81320536e-01 2.28516638e-01
2.56542474e-01 7.23464429e-01 -3.88636857e-01 -2.62896806e-01
-4.04003561e-01 -7.95577228e-01 5.10862805e-02 3.17603022e-01
1.75561026e-01 -1.24844444e+00 -4.71703589e-01 4.44773257e-01
8.05396140e-01 -8.89662325e-01 4.39607054e-01 4.75957245e-01
-1.32567954e+00 -1.29936182e+00 -2.26893157e-01 -2.60714561e-01
2.31831118e-01 -8.23094428e-01 9.82840657e-01 3.34308386e-01
-1.91315100e-01 -3.68439287e-01 -6.10877931e-01 -7.52490938e-01
-1.16183567e+00 1.17113724e-01 -1.42569542e-01 -3.12436938e-01
2.25372672e-01 -6.26427412e-01 -3.57521296e-01 2.33873621e-01
-1.09394264e+00 -6.75655305e-01 5.30385733e-01 5.81694186e-01
8.16291256e-04 3.09995592e-01 6.09074056e-01 -9.45442557e-01
8.04045677e-01 -4.28050488e-01 -1.10228014e+00 -1.63605828e-02
-1.20857263e+00 9.71468389e-02 8.83608282e-01 -2.54786372e-01
-1.01684499e+00 -3.62137705e-01 -9.90714207e-02 -6.59876782e-03
-2.06884906e-01 6.21925592e-01 1.83465868e-01 -3.75279300e-02
6.68859422e-01 4.05094400e-02 4.11343515e-01 -6.19716823e-01
-3.29974532e-01 5.83588779e-01 2.94019967e-01 -8.76191735e-01
7.13383198e-01 -2.78269321e-01 5.96623719e-01 -1.06879759e+00
-2.95041110e-02 -1.23379990e-01 -8.47029567e-01 -5.05567133e-01
6.58813357e-01 -2.41601110e-01 -8.69186878e-01 7.55632222e-01
-7.51907945e-01 -2.34979019e-01 -1.13038741e-01 9.23815072e-01
-3.25467229e-01 3.52011830e-01 -7.49151409e-01 -1.06387556e+00
-4.31040168e-01 -1.42130554e+00 2.15684548e-01 6.12371027e-01
-1.11335993e-01 -1.29401243e+00 -1.37375250e-01 7.59657696e-02
4.42114592e-01 5.74113250e-01 1.13262534e+00 -6.00113094e-01
-2.40729466e-01 -5.78711450e-01 1.86984316e-01 7.66123414e-01
4.02554031e-03 9.49305773e-01 -5.92382073e-01 -3.04062944e-02
9.18607563e-02 -9.58753303e-02 4.03974026e-01 5.30057609e-01
6.67364061e-01 -1.58748985e-03 1.67664126e-01 4.49585140e-01
1.88085210e+00 9.56596017e-01 8.31195176e-01 7.91279495e-01
2.06273928e-01 4.69874978e-01 4.13822293e-01 3.73134434e-01
-5.12109339e-01 1.17240950e-01 3.63054663e-01 2.55632401e-02
3.98674935e-01 -1.75643340e-03 1.69447679e-02 6.07651353e-01
-5.41094244e-01 6.78757578e-02 -7.06093252e-01 1.77491501e-01
-1.36233985e+00 -8.59014630e-01 -6.21124983e-01 2.34058809e+00
4.31963742e-01 6.22471690e-01 4.09506828e-01 7.69394875e-01
4.72877920e-01 -1.78118840e-01 -2.52467304e-01 -1.21879494e+00
-5.54252714e-02 5.19446552e-01 8.17276835e-01 5.63842118e-01
-6.53926373e-01 5.56116581e-01 7.08633041e+00 1.86047107e-01
-1.53800690e+00 -6.15279377e-01 5.46621859e-01 -8.56439292e-04
-5.58185875e-02 -1.38737645e-03 -4.06107008e-01 8.21430683e-01
1.19847286e+00 -5.67054264e-02 3.23931515e-01 8.14616263e-01
4.78835762e-01 -4.85505491e-01 -7.43009925e-01 2.95324534e-01
-3.13291430e-01 -1.01935375e+00 -2.22847611e-01 3.03223263e-02
7.30213523e-01 -3.15250516e-01 -3.31755549e-01 5.60749806e-02
-1.38000757e-01 -8.37198198e-01 6.96262598e-01 9.65222836e-01
6.22542873e-02 -8.29112589e-01 1.31991482e+00 -4.62838262e-02
-7.53268957e-01 -1.95266083e-01 -3.89116675e-01 -3.37068081e-01
2.11105257e-01 7.41961420e-01 -8.49997640e-01 7.10253239e-01
6.37806058e-01 9.65747386e-02 -2.50601441e-01 1.07602549e+00
-8.48884434e-02 7.59361804e-01 -5.87657511e-01 -6.25967324e-01
2.79564202e-01 -1.18557143e+00 1.08896457e-01 8.76074910e-01
4.21396226e-01 -2.42995694e-01 -1.87890619e-01 9.85458076e-01
5.07836998e-01 6.77800119e-01 -2.29162082e-01 -1.40698075e-01
6.19086921e-01 7.41820157e-01 -8.81549299e-01 -2.23309636e-01
-7.79485181e-02 7.14083165e-02 -3.02065641e-01 3.64897728e-01
-7.59323597e-01 -7.04797208e-01 1.72672033e-01 3.24741542e-01
-7.77905285e-02 -4.10845399e-01 -4.77372259e-01 -4.80129689e-01
-2.13613249e-02 -4.50353026e-01 -3.85110476e-03 -5.44637263e-01
-8.90218794e-01 4.16153669e-01 3.19852382e-01 -9.52198923e-01
-2.84294456e-01 -1.22303951e+00 -5.94556689e-01 1.23851466e+00
-8.95121038e-01 -7.42625713e-01 -2.08529681e-01 -1.25894368e-01
2.54496187e-01 -4.25320029e-01 5.60552239e-01 2.01728731e-01
-5.81261575e-01 1.28966540e-01 5.99104464e-01 -1.26085192e-01
4.89665866e-01 -1.42490685e+00 -2.24348202e-01 4.46843117e-01
-5.78996778e-01 3.37619156e-01 1.31466579e+00 -6.66240811e-01
-1.00614870e+00 -3.33080024e-01 5.49430728e-01 1.63887963e-01
1.00128198e+00 1.40148371e-01 -1.03671670e+00 3.01250696e-01
-1.35192662e-01 -4.83329445e-01 2.86576360e-01 1.37031883e-01
1.56991377e-01 -1.89343750e-01 -1.43357027e+00 1.99414253e-01
6.16965815e-02 -2.22221032e-01 -5.02121627e-01 2.48662651e-01
6.15323067e-01 -4.68162209e-01 -1.75332355e+00 1.78727895e-01
8.23451281e-01 -1.08642232e+00 6.83575869e-01 -4.38221097e-01
6.82161093e-01 -1.57316159e-02 3.35355788e-01 -1.10524487e+00
-1.76917478e-01 -2.01407135e-01 2.38484845e-01 1.14047408e+00
9.45865095e-01 -9.82185781e-01 9.42817628e-01 1.12222838e+00
-1.55759528e-01 -7.60107577e-01 -7.21442461e-01 -9.42143679e-01
2.78963983e-01 -2.13080376e-01 7.00702310e-01 7.36579478e-01
9.22095254e-02 7.47012487e-03 -1.86164275e-01 -1.83073676e-03
2.69516885e-01 -2.46596143e-01 6.41957879e-01 -1.51773882e+00
-2.13959455e-01 -6.43677354e-01 -7.35278606e-01 5.32031879e-02
3.38464640e-02 -5.01618981e-01 -2.85402209e-01 -1.37717736e+00
-5.08074105e-01 -4.40285742e-01 -7.98120052e-02 5.67574352e-02
2.55843580e-01 -1.47250295e-01 1.75314471e-01 5.06586172e-02
7.25040913e-01 3.23633254e-01 1.37093616e+00 -1.21289946e-01
-7.42716372e-01 1.09602973e-01 -5.17127197e-03 5.62880635e-01
8.52313161e-01 -3.13784868e-01 -3.84141117e-01 1.50499046e-01
1.67138904e-01 1.37860611e-01 -3.30009192e-01 -1.32436287e+00
-3.42952549e-01 -4.50674385e-01 6.24164581e-01 -4.88039404e-01
1.89121336e-01 -1.13349724e+00 5.94148457e-01 6.41459882e-01
-1.26499340e-01 1.55128509e-01 1.47982016e-01 6.48028329e-02
-2.93747991e-01 -1.13365912e+00 7.36084461e-01 -8.40909034e-02
-3.28613222e-01 -1.59530640e-01 -2.80022323e-01 -7.52073407e-01
9.68921661e-01 -7.43774772e-01 -2.83187091e-01 -1.28643960e-01
-1.03850901e+00 -4.57781591e-02 4.91290480e-01 3.12450640e-02
1.29562244e-01 -6.77615404e-01 -4.45797861e-01 1.08472995e-01
-3.91939104e-01 6.38569742e-02 2.58446455e-01 7.67134011e-01
-1.52718902e+00 3.15545231e-01 -4.41602886e-01 -2.71016747e-01
-9.74088192e-01 5.33594906e-01 5.46763003e-01 -1.75877124e-01
-4.71267372e-01 3.40468347e-01 -8.42079401e-01 -3.10339481e-01
-1.62976876e-01 -8.71913135e-01 -5.05779743e-01 -5.79524599e-02
1.19112119e-01 6.76429808e-01 3.96507084e-01 -4.70101982e-01
8.45881253e-02 6.98626220e-01 3.17287236e-01 9.42191333e-02
1.59926200e+00 3.85083526e-01 -3.37575674e-01 6.88341379e-01
1.06716001e+00 -1.34794876e-01 -9.23215449e-01 3.65489721e-01
2.41066933e-01 -3.55991870e-01 -1.25830146e-02 -8.65785062e-01
-1.03885996e+00 3.02857310e-01 5.59194028e-01 7.01001287e-01
9.23519135e-01 -3.78196627e-01 4.30150449e-01 5.48264384e-01
-1.52808145e-01 -1.64035285e+00 -2.22728893e-01 5.64216197e-01
8.81464839e-01 -1.20081282e+00 3.08665663e-01 -3.43421638e-01
-3.88776392e-01 1.80658770e+00 5.22497237e-01 -2.22810134e-01
7.74152100e-01 2.76589602e-01 2.48074278e-01 -1.12955466e-01
-2.78601974e-01 4.32912618e-01 2.96872277e-02 4.68345225e-01
7.84335256e-01 1.46388799e-01 -9.98857141e-01 1.00680239e-01
-5.06234109e-01 1.10631958e-01 6.50526583e-01 8.93252850e-01
-6.40334606e-01 -1.13626480e+00 -8.74660075e-01 6.81095839e-01
-2.63517380e-01 4.10993278e-01 -3.25889140e-02 1.32345378e+00
8.65587294e-02 8.97301733e-01 1.54666260e-01 -2.71026284e-01
6.56672716e-01 1.97638258e-01 2.55294740e-01 8.04983824e-03
-4.91647094e-01 5.98624498e-02 3.08215827e-01 -1.12239756e-01
-3.54361087e-01 -6.08305812e-01 -1.38902080e+00 -5.38374305e-01
-5.44812083e-01 5.85801125e-01 1.45381725e+00 1.02994263e+00
-5.83557963e-01 5.83039939e-01 7.99000323e-01 -7.97862053e-01
-5.91364801e-01 -1.57627046e+00 -1.34944808e+00 6.33828521e-01
-9.00657326e-02 -6.24992132e-01 -6.92138016e-01 -1.24277502e-01] | [6.262559413909912, 3.330704927444458] |
4574d277-939c-41b8-8de2-6986690bcfac | image-based-localization-using-hourglass | 1703.07971 | null | http://arxiv.org/abs/1703.07971v3 | http://arxiv.org/pdf/1703.07971v3.pdf | Image-based Localization using Hourglass Networks | In this paper, we propose an encoder-decoder convolutional neural network
(CNN) architecture for estimating camera pose (orientation and location) from a
single RGB-image. The architecture has a hourglass shape consisting of a chain
of convolution and up-convolution layers followed by a regression part. The
up-convolution layers are introduced to preserve the fine-grained information
of the input image. Following the common practice, we train our model in
end-to-end manner utilizing transfer learning from large scale classification
data. The experiments demonstrate the performance of the approach on data
exhibiting different lighting conditions, reflections, and motion blur. The
results indicate a clear improvement over the previous state-of-the-art even
when compared to methods that utilize sequence of test frames instead of a
single frame. | ['Esa Rahtu', 'Juho Kannala', 'Juha Ylioinas', 'Iaroslav Melekhov'] | 2017-03-23 | null | null | null | null | ['image-based-localization'] | ['computer-vision'] | [ 1.76133916e-01 -9.43551511e-02 3.38723063e-01 -5.96736789e-01
-4.67273176e-01 -3.79874557e-01 4.73905832e-01 -5.47447503e-01
-6.29843891e-01 5.92372775e-01 3.10951829e-01 -2.38651708e-01
4.29414421e-01 -5.00347793e-01 -1.21141040e+00 -5.89808881e-01
1.03541590e-01 -3.09217781e-01 2.67399907e-01 -4.99623567e-02
2.40326241e-01 3.86231542e-01 -1.42786443e+00 5.49022019e-01
2.00933471e-01 1.31253505e+00 2.18099862e-01 1.01652217e+00
4.56073225e-01 1.04901397e+00 -4.02641058e-01 -3.92246097e-01
5.10889828e-01 -2.06897259e-01 -7.03210950e-01 5.40388584e-01
7.73654580e-01 -7.33155847e-01 -6.98855340e-01 8.06998610e-01
5.50170302e-01 6.74840286e-02 1.53668851e-01 -8.04220855e-01
-7.61359990e-01 -3.57023656e-01 -4.23110962e-01 1.17709219e-01
4.60848361e-01 2.66126335e-01 5.25677919e-01 -9.91491675e-01
4.49133813e-01 1.05847192e+00 1.03168499e+00 2.30518192e-01
-8.45497310e-01 -4.21213478e-01 -7.32218698e-02 2.39992335e-01
-1.31806755e+00 -3.50393534e-01 6.34732187e-01 -4.28907275e-01
1.23322129e+00 -2.29638398e-01 6.05410218e-01 1.23989654e+00
3.13731462e-01 3.75210106e-01 1.09910131e+00 -4.23016369e-01
6.38521835e-02 -7.16062486e-02 -1.87086388e-01 8.92868161e-01
8.98765214e-03 3.15643787e-01 -5.51879108e-01 3.61533970e-01
8.85508955e-01 1.35860100e-01 -3.00145149e-01 -3.94768327e-01
-1.13513923e+00 5.45673251e-01 8.32716525e-01 -2.36964121e-01
-4.32330549e-01 3.50709856e-01 3.30965668e-01 2.90217608e-01
3.87242317e-01 -2.40063947e-02 -6.98225200e-01 -1.62537858e-01
-9.31686401e-01 -1.20019643e-02 6.65196359e-01 1.02558362e+00
8.13517332e-01 6.47771657e-02 1.16008274e-01 5.19594908e-01
4.43840593e-01 2.70248115e-01 5.54937840e-01 -8.05023313e-01
5.51480114e-01 3.86912793e-01 1.25485003e-01 -7.35543132e-01
-3.59383792e-01 -3.22409809e-01 -8.53930056e-01 5.52101910e-01
3.08050603e-01 -4.28514987e-01 -1.13412666e+00 1.33305585e+00
1.13366842e-01 3.58318657e-01 1.24513142e-01 1.16301405e+00
6.42954946e-01 5.67233741e-01 -2.68953353e-01 2.57756978e-01
1.25090778e+00 -1.39895380e+00 -5.29511929e-01 -2.37859413e-01
1.13046452e-01 -9.51103330e-01 8.35128009e-01 4.47602779e-01
-9.51734245e-01 -1.02971327e+00 -1.19687676e+00 -4.07188118e-01
-5.74074864e-01 5.67137837e-01 3.31108600e-01 5.33500016e-01
-1.31101787e+00 6.07665062e-01 -8.13363552e-01 -5.14348090e-01
4.27733123e-01 5.13603628e-01 -7.25710690e-01 -2.06929237e-01
-7.58488417e-01 8.64549279e-01 2.13475958e-01 5.09793580e-01
-1.06381917e+00 -2.92387903e-01 -1.05212033e+00 -5.79715371e-02
1.03157841e-01 -6.99701548e-01 1.36371255e+00 -1.21448076e+00
-1.98321974e+00 6.28731132e-01 -1.88797340e-01 -3.94598514e-01
6.34526670e-01 -6.70711398e-01 -1.20938033e-01 3.10592465e-02
-2.50549316e-01 6.35505021e-01 1.04130614e+00 -1.05542731e+00
-7.36084759e-01 -7.34135732e-02 2.32674658e-01 7.97262564e-02
-6.11896440e-02 8.64580646e-02 -6.28064573e-01 -4.81840491e-01
-8.15370083e-02 -1.05355811e+00 -1.28192559e-01 1.06216066e-01
-4.00835812e-01 3.42180282e-01 9.40064549e-01 -6.93223596e-01
8.56258750e-01 -2.16918015e+00 8.77023265e-02 -2.33423352e-01
-1.30247742e-01 3.47616345e-01 6.95673302e-02 5.33176780e-01
-3.79197776e-01 -3.28598738e-01 -6.32947460e-02 -6.23147190e-01
-2.37910300e-01 -7.21067702e-03 -1.46175027e-01 8.04850817e-01
4.58835065e-01 7.29841053e-01 -6.34522080e-01 -9.40672960e-03
6.20595813e-01 8.52525711e-01 -4.63949382e-01 5.76863825e-01
8.52149203e-02 5.13360143e-01 1.59836989e-02 7.02543259e-01
7.41878867e-01 -2.39490822e-01 -9.88051221e-02 -4.73971695e-01
-2.12249175e-01 1.45106792e-01 -1.13115060e+00 1.98104119e+00
-5.10052443e-01 1.12994361e+00 1.56962872e-03 -6.19477391e-01
7.79810369e-01 3.54935706e-01 1.03823356e-01 -5.75593412e-01
2.66543269e-01 5.70328273e-02 -5.29010333e-02 -7.38151193e-01
4.70936835e-01 1.27919957e-01 7.46301189e-02 1.04949348e-01
3.07878375e-01 1.73516080e-01 -1.10461093e-01 -3.54877383e-01
1.04537737e+00 6.37955844e-01 2.05521375e-01 7.52521306e-02
6.30634725e-01 -2.81959057e-01 2.53575534e-01 3.20210278e-01
-1.88316181e-01 1.25353777e+00 3.35770845e-01 -9.96473014e-01
-1.43997288e+00 -7.83471048e-01 5.87111227e-02 7.67834842e-01
1.01891145e-01 -2.68014938e-01 -8.19263101e-01 -4.61846977e-01
-2.27531716e-01 5.65876588e-02 -8.54819357e-01 -1.58112720e-02
-6.08745933e-01 -5.29297709e-01 5.87567031e-01 8.19487274e-01
1.05770564e+00 -1.03593314e+00 -1.07863617e+00 4.62147929e-02
1.37107372e-01 -1.52267635e+00 -3.65424067e-01 4.03251946e-01
-6.91902399e-01 -1.23218846e+00 -6.65675104e-01 -8.11874986e-01
7.60337055e-01 2.30599999e-01 1.01002133e+00 -2.22651854e-01
-3.11991543e-01 1.31674007e-01 -2.68146247e-01 -1.41509458e-01
1.42601162e-01 -6.71019703e-02 -2.19936922e-01 2.46760204e-01
2.81287760e-01 -4.30288196e-01 -9.61649358e-01 8.72773305e-02
-8.76111984e-01 2.67516524e-01 8.64729464e-01 8.61077428e-01
2.37665132e-01 -2.70056605e-01 -8.95243362e-02 -4.60233569e-01
3.66366714e-01 -2.02792540e-01 -7.99825132e-01 1.23668306e-01
-3.71824861e-01 -5.32115847e-02 6.11639261e-01 -1.50934547e-01
-1.05914938e+00 6.21517599e-01 5.21670841e-02 -6.17931366e-01
-4.91718382e-01 1.33840367e-01 1.81670412e-02 -2.00122580e-01
2.59413600e-01 1.86332002e-01 -4.76252437e-02 -4.47867811e-01
3.26757759e-01 8.39660645e-01 9.04383302e-01 -2.03831419e-01
5.27830362e-01 4.90578949e-01 -1.28727974e-02 -6.60972476e-01
-7.40386963e-01 -5.60737848e-01 -1.10669351e+00 -3.23925108e-01
1.08843935e+00 -1.28828549e+00 -7.61208177e-01 9.25453186e-01
-1.34850383e+00 -3.10830623e-01 8.19992125e-02 6.09262764e-01
-5.60023248e-01 6.24918640e-02 -7.70537674e-01 -5.13136446e-01
-1.41099960e-01 -1.41218889e+00 1.32573867e+00 3.91803712e-01
2.55849570e-01 -7.55673051e-01 1.85964867e-01 2.28171021e-01
4.47471797e-01 3.49660158e-01 1.68902770e-01 -1.36220515e-01
-7.70658374e-01 -3.72970819e-01 -3.79019141e-01 6.76654458e-01
3.95073816e-02 1.01620734e-01 -1.39708328e+00 -3.36655617e-01
-1.01508394e-01 -5.34903884e-01 7.89526999e-01 2.80242682e-01
1.11700618e+00 -1.98002949e-01 1.82678819e-01 1.14316690e+00
1.79574466e+00 -5.24561154e-03 1.04915679e+00 6.58722639e-01
8.45116556e-01 2.83757716e-01 2.41808712e-01 3.84943336e-01
3.94257784e-01 4.45718825e-01 7.05588818e-01 -4.17787910e-01
-2.49552742e-01 -9.32404846e-02 4.47913289e-01 4.52649564e-01
-3.12355220e-01 -1.61415070e-01 -7.67864883e-01 4.82361019e-01
-1.79577947e+00 -8.94380689e-01 8.87681823e-03 2.09165311e+00
5.37641346e-01 -2.24602446e-02 -1.03817275e-02 -4.12541851e-02
4.73546296e-01 2.05843002e-01 -3.18101466e-01 -5.45241654e-01
-6.23576995e-03 1.72123611e-01 9.48364377e-01 4.76536423e-01
-1.29507756e+00 8.33279014e-01 6.97995186e+00 9.69812926e-03
-1.45018101e+00 -7.81191587e-02 6.73914194e-01 -1.02636479e-01
4.76442009e-01 -1.29207090e-01 -5.39004862e-01 3.67318690e-01
1.01717353e+00 5.59879065e-01 4.03424561e-01 7.69301116e-01
2.31942207e-01 -2.64065832e-01 -1.02753973e+00 1.00809228e+00
2.54283130e-01 -1.21666694e+00 -4.66865927e-01 8.37299749e-02
7.62617946e-01 5.16491294e-01 3.07235755e-02 8.50011483e-02
2.17502818e-01 -1.17901731e+00 8.71918797e-01 5.94068527e-01
8.53966236e-01 -6.65674448e-01 1.02358079e+00 9.09561813e-02
-1.09506845e+00 -2.40238219e-01 -3.95407647e-01 -3.36423308e-01
7.13489056e-02 1.27500579e-01 -7.74807513e-01 4.49758410e-01
1.13276744e+00 9.08912778e-01 -7.19755054e-01 1.05355024e+00
-3.27166826e-01 3.13386470e-01 -4.24901098e-02 2.95026720e-01
5.63623726e-01 -1.28210202e-01 -6.81645200e-02 1.47042084e+00
3.36683959e-01 1.17974142e-02 -3.62904295e-02 5.10004580e-01
-1.88138992e-01 -4.62114990e-01 -6.97947025e-01 3.53317797e-01
-4.19157334e-02 1.30816555e+00 -4.92565066e-01 -2.29413763e-01
-9.17775214e-01 1.38896680e+00 4.32423174e-01 4.49093103e-01
-9.21680987e-01 -4.48647380e-01 7.06749916e-01 1.23678241e-03
1.02187765e+00 -3.15235943e-01 -1.50085419e-01 -1.00715864e+00
2.01626107e-01 -7.71588504e-01 -7.46542886e-02 -1.19237947e+00
-9.83557999e-01 8.58882666e-01 -1.42372996e-01 -1.46645868e+00
-3.15057337e-01 -1.08818376e+00 -5.84078729e-01 9.19001341e-01
-1.72103548e+00 -1.39701664e+00 -7.70778298e-01 6.94582283e-01
5.98946273e-01 -1.45323277e-01 7.62673199e-01 3.14171463e-01
-5.94319403e-01 4.12101030e-01 2.27914840e-01 5.09463191e-01
7.29366064e-01 -1.24469292e+00 4.47093397e-01 1.07892776e+00
-6.77962527e-02 5.20457089e-01 5.15647292e-01 -2.08362773e-01
-1.40915632e+00 -1.09478891e+00 7.28361309e-01 -3.39002281e-01
3.87164772e-01 -4.68414009e-01 -5.63152194e-01 8.30145359e-01
8.96886706e-01 6.25194907e-01 3.82032931e-01 -1.25074565e-01
-4.97022569e-01 -3.12097907e-01 -9.49307978e-01 2.24905238e-01
6.47847235e-01 -7.73355484e-01 -4.80019450e-01 4.18697186e-02
5.28630853e-01 -8.21978509e-01 -7.57895827e-01 2.08341062e-01
8.04638565e-01 -1.42931700e+00 9.83819306e-01 -3.25279802e-01
8.03758562e-01 -5.28143823e-01 -3.20687920e-01 -1.18622422e+00
-2.66194254e-01 -6.91795528e-01 -4.09881994e-02 6.80259347e-01
2.41073415e-01 -2.38433212e-01 6.62699342e-01 5.15310049e-01
-3.02559882e-01 -7.29716361e-01 -7.13721752e-01 -4.19732094e-01
-3.89800549e-01 -2.86312401e-01 3.72537941e-01 4.20946538e-01
-3.06435823e-01 3.27793926e-01 -7.05435336e-01 3.52293521e-01
3.98644567e-01 -9.01876763e-02 8.59111726e-01 -6.57048404e-01
-2.48185843e-01 4.04560417e-02 -6.81562603e-01 -8.95719290e-01
-7.52465129e-02 -5.16111888e-02 2.64560133e-01 -1.31060374e+00
9.62989032e-02 2.96792567e-01 -3.54519904e-01 4.68079001e-01
-2.70783696e-02 6.04854941e-01 1.85278326e-01 4.49059298e-03
-7.91146696e-01 3.37819457e-01 1.06818593e+00 1.78363964e-01
1.71706930e-01 -1.14966273e-01 -3.07989180e-01 7.82840610e-01
6.27660036e-01 -3.14821541e-01 -3.17288518e-01 -8.16564620e-01
-3.14651579e-02 -1.81672558e-01 7.56631076e-01 -1.26542759e+00
3.94150227e-01 2.98806936e-01 9.11229968e-01 -6.94446921e-01
4.50931937e-01 -1.06231558e+00 1.38889238e-01 4.65066850e-01
-1.77645668e-01 6.07692182e-01 2.95171738e-01 6.66213810e-01
-5.29736459e-01 1.94015428e-01 6.36544406e-01 -1.13563247e-01
-1.03688860e+00 2.16068894e-01 -1.03586778e-01 -4.25657153e-01
1.07142496e+00 -4.15217906e-01 -2.61087954e-01 -3.64057571e-01
-4.87017095e-01 -3.45901966e-01 6.33492470e-01 4.82663691e-01
8.10444415e-01 -1.23788917e+00 -5.33635080e-01 5.78310370e-01
1.10834740e-01 2.43913040e-01 -9.96650755e-03 6.54868424e-01
-1.04138720e+00 5.58067918e-01 -5.57360947e-01 -6.39083207e-01
-1.20714557e+00 4.96827513e-01 5.77572405e-01 2.61658952e-02
-6.99597538e-01 7.87750065e-01 -1.22659821e-02 -2.68573493e-01
4.27344203e-01 -6.75480425e-01 1.38233140e-01 -4.33967412e-01
4.61713463e-01 9.98262838e-02 2.83876210e-01 -7.48932779e-01
-3.76950383e-01 6.25164032e-01 1.26338363e-01 -1.56504542e-01
1.56977069e+00 -2.50351399e-01 8.04416835e-02 3.17666292e-01
1.55392885e+00 -2.54449338e-01 -2.16256022e+00 -1.17551468e-01
-2.48016506e-01 -6.62189960e-01 2.08705708e-01 -6.72090828e-01
-1.18804324e+00 1.01495814e+00 8.35098267e-01 -2.28834823e-01
1.32855296e+00 -3.84830534e-01 5.83524883e-01 4.25859869e-01
6.16302006e-02 -9.88500059e-01 1.93970144e-01 6.82169974e-01
7.40890563e-01 -1.48911142e+00 5.18304780e-02 6.84830546e-02
-6.82411849e-01 1.50170672e+00 6.38159454e-01 -5.15137792e-01
6.46877170e-01 4.21813011e-01 2.52337992e-01 -1.21274300e-01
-8.38461936e-01 -2.94077963e-01 2.82403558e-01 5.86789489e-01
7.09350407e-01 -3.43655258e-01 4.45981860e-01 9.73643139e-02
-2.22112052e-02 3.40874285e-01 5.00384271e-01 1.01276124e+00
-1.06607042e-01 -7.98778117e-01 -3.91463727e-01 -2.17769504e-01
-5.93378186e-01 -1.35740116e-01 -3.34950954e-01 1.01474571e+00
3.69962245e-01 7.82190621e-01 2.21551284e-01 -6.09730840e-01
3.72400850e-01 -1.17916413e-01 5.86018026e-01 -3.56147856e-01
-7.47303009e-01 1.27282619e-01 -5.70444204e-02 -9.06309664e-01
-7.23999321e-01 -4.44138348e-01 -7.89454222e-01 -2.73567051e-01
-9.58300680e-02 -5.39630949e-01 9.14513528e-01 8.54411244e-01
3.33884865e-01 8.39078128e-01 7.48943508e-01 -1.18463922e+00
-4.24618870e-01 -1.17980039e+00 -2.73640096e-01 4.13899690e-01
8.98222923e-01 -3.55145842e-01 -3.84981595e-02 6.62993729e-01] | [10.123032569885254, -2.384117364883423] |
a3f707b1-5b81-4413-9e12-9affc919caf1 | video-instance-segmentation-tracking-with-a | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Lin_Video_Instance_Segmentation_Tracking_With_a_Modified_VAE_Architecture_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Lin_Video_Instance_Segmentation_Tracking_With_a_Modified_VAE_Architecture_CVPR_2020_paper.pdf | Video Instance Segmentation Tracking With a Modified VAE Architecture | We propose a modified variational autoencoder (VAE) architecture built on top of Mask R-CNN for instance-level video segmentation and tracking. The method builds a shared encoder and three parallel decoders, yielding three disjoint branches for predictions of future frames, object detection boxes, and instance segmentation masks. To effectively solve multiple learning tasks, we introduce a Gaussian Process model to enhance the statistical representation of VAE by relaxing the prior strong independent and identically distributed (iid) assumption of conventional VAEs and allowing potential correlations among extracted latent variables. The network learns embedded spatial interdependence and motion continuity in video data and creates a representation that is effective to produce high-quality segmentation masks and track multiple instances in diverse and unstructured videos. Evaluation on a variety of recently introduced datasets shows that our model outperforms previous methods and achieves the new best in class performance.
| [' Linglin He', ' Rogerio Feris', ' Ying Hung', 'Chung-Ching Lin'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['video-instance-segmentation'] | ['computer-vision'] | [ 1.24092929e-01 2.53388822e-01 -3.11066151e-01 -2.05703408e-01
-9.16782081e-01 -4.78199303e-01 3.71886134e-01 -6.52044237e-01
-3.57598215e-01 4.77148682e-01 9.29544792e-02 -4.19321768e-02
1.84738457e-01 -4.79353786e-01 -1.16863525e+00 -7.22439706e-01
-9.53013375e-02 6.71797395e-01 8.39547038e-01 3.82131279e-01
-1.74083129e-01 2.96809733e-01 -1.40528345e+00 4.32771713e-01
4.38925326e-01 1.01903319e+00 5.46433926e-01 9.77680385e-01
1.21279709e-01 1.12493885e+00 -6.33175969e-01 -4.24579024e-01
3.67691129e-01 -2.17914686e-01 -4.89071369e-01 4.88080591e-01
5.12804508e-01 -5.35904467e-01 -5.78323424e-01 7.74926960e-01
1.18045159e-01 3.14953476e-01 8.96229625e-01 -1.33383083e+00
-6.54827297e-01 5.25351048e-01 -8.40936184e-01 3.31295639e-01
-2.71756202e-01 2.83632785e-01 9.47204232e-01 -8.50093067e-01
8.71426105e-01 1.18182385e+00 7.75771499e-01 8.42700064e-01
-1.39210975e+00 -4.26160544e-01 4.24924195e-01 7.49428719e-02
-1.45366347e+00 -4.11221355e-01 4.94626492e-01 -7.04525352e-01
1.12639725e+00 -7.79466406e-02 6.45038843e-01 1.59561896e+00
3.72911453e-01 1.37578738e+00 3.63147914e-01 1.56839177e-01
2.41778165e-01 3.25812921e-02 -2.10841805e-01 7.37645924e-01
-7.40225911e-02 -1.19589865e-02 -3.46872747e-01 5.59744164e-02
1.35361803e+00 9.44183033e-04 -9.05694366e-02 -6.21025503e-01
-1.14077222e+00 8.27009439e-01 2.01633591e-02 6.27887025e-02
-5.70801198e-01 5.17007351e-01 1.76438242e-01 -2.23099619e-01
3.69273514e-01 4.11737710e-02 -7.21625090e-01 -2.27944530e-03
-1.50355244e+00 3.30606282e-01 5.69993317e-01 1.15427566e+00
5.13531268e-01 4.24547881e-01 -5.69611430e-01 5.73342443e-01
5.69709182e-01 5.17098904e-01 2.48327091e-01 -1.50414193e+00
2.75310844e-01 1.84319168e-01 1.32380333e-02 -7.21508741e-01
-6.76275343e-02 -4.85468775e-01 -6.69161558e-01 2.19504893e-01
2.09351823e-01 -4.46674138e-01 -1.30864155e+00 1.69926608e+00
1.70825645e-01 9.17597711e-01 -1.14280418e-01 7.87016749e-01
7.13323593e-01 9.15578306e-01 2.41497546e-01 -4.40171808e-02
9.97805715e-01 -1.25181150e+00 -7.56864309e-01 -3.56918722e-01
-3.48832607e-02 -3.12189132e-01 3.30220610e-01 3.77842993e-01
-1.33996165e+00 -9.01506186e-01 -7.66022503e-01 -2.00910509e-01
2.51947269e-02 2.35224515e-01 5.05729079e-01 3.00788224e-01
-1.35242951e+00 5.28084338e-01 -1.28204799e+00 3.66601676e-01
8.85190427e-01 5.01420259e-01 -2.25551143e-01 2.09788069e-01
-6.90228701e-01 5.93673706e-01 3.97980541e-01 1.02234617e-01
-1.57521260e+00 -6.99811399e-01 -1.01569700e+00 1.12763055e-01
5.05322516e-01 -7.85488665e-01 1.13514698e+00 -1.16021097e+00
-1.43166280e+00 7.22231448e-01 -3.68014455e-01 -7.56695688e-01
7.54158497e-01 -3.68928730e-01 -1.87043518e-01 2.94891149e-01
9.93595123e-02 1.37406230e+00 1.32944798e+00 -1.29814386e+00
-8.44659865e-01 2.13503782e-02 -4.32166785e-01 -3.26078339e-03
6.24996908e-02 5.34959473e-02 -1.24532950e+00 -8.14013183e-01
-2.00645402e-01 -8.09486151e-01 -5.79917312e-01 2.10987125e-02
-2.78202385e-01 -9.63811800e-02 9.96841669e-01 -1.05287731e+00
1.02015519e+00 -2.18719029e+00 6.89339459e-01 2.93385554e-02
4.23535287e-01 9.15533751e-02 -1.69965386e-01 -4.51493025e-01
7.44247735e-02 3.34013104e-02 -2.70350873e-01 -7.77171493e-01
-9.76680443e-02 5.91976166e-01 -2.77717382e-01 4.24198389e-01
5.97895026e-01 1.21896207e+00 -6.43901587e-01 -7.00851798e-01
3.77860844e-01 7.18353987e-01 -6.80734217e-01 3.40560436e-01
-7.25560427e-01 6.22035325e-01 -3.25514048e-01 5.74221015e-01
4.25502449e-01 -6.71066225e-01 1.36479542e-01 4.29047197e-02
2.01580599e-01 -2.51898527e-01 -1.40277696e+00 1.64819407e+00
1.78637162e-01 1.00254798e+00 3.57324421e-01 -9.09045815e-01
6.01354778e-01 5.45768678e-01 7.63673663e-01 -1.70114160e-01
3.01309407e-01 -3.61701995e-01 -2.90756702e-01 -4.80429977e-01
4.87414777e-01 2.40012378e-01 1.55675516e-01 -5.54783568e-02
5.82097769e-01 1.62458569e-01 9.34123695e-02 9.42741111e-02
1.04289734e+00 6.45029068e-01 -8.91556293e-02 -1.10400714e-01
2.20078617e-01 -1.98946476e-01 1.16756427e+00 8.29461098e-01
-5.51215887e-01 8.49758089e-01 5.55003166e-01 -3.85102302e-01
-1.33510661e+00 -1.33996189e+00 -5.31010516e-03 9.58540916e-01
1.16749093e-01 -3.09636235e-01 -7.56310940e-01 -6.64464355e-01
-1.12576351e-01 6.32502496e-01 -7.08383620e-01 2.37512738e-01
-7.85655439e-01 -5.25799334e-01 4.19881046e-01 1.06544280e+00
2.15731487e-01 -1.13043845e+00 -5.64388394e-01 3.41177225e-01
-2.24349961e-01 -1.48644209e+00 -2.63317734e-01 2.90102094e-01
-7.66017318e-01 -8.94532561e-01 -8.07244778e-01 -9.23466325e-01
2.54641175e-01 -1.10730827e-01 1.31296802e+00 -3.10962558e-01
-2.99484581e-01 3.95651489e-01 6.34154528e-02 -2.37736955e-01
-3.81044686e-01 -2.48571411e-01 -6.69172406e-02 2.27145460e-02
2.81007081e-01 -3.30896288e-01 -5.30687869e-01 1.61746308e-01
-7.47721076e-01 8.03549811e-02 5.10728776e-01 7.29899168e-01
1.04577780e+00 -1.22593287e-02 4.79970500e-02 -4.58886743e-01
2.65728440e-02 -6.25902414e-01 -9.15360034e-01 1.88685864e-01
-1.95729852e-01 -3.39079499e-02 2.46509135e-01 -5.40115178e-01
-1.09724319e+00 2.44129643e-01 -4.87877168e-02 -1.31342423e+00
-3.81138176e-01 -1.00064337e-01 -5.80561534e-02 3.60764921e-01
2.27475315e-01 3.38729233e-01 -1.48664087e-01 -2.92326927e-01
5.46220541e-01 1.54452309e-01 7.50512958e-01 -3.80771875e-01
5.08699298e-01 5.77533782e-01 -2.03821808e-01 -7.09316373e-01
-4.75193918e-01 -3.77332032e-01 -8.60708237e-01 -2.76844054e-01
1.62801397e+00 -1.38893902e+00 -4.76644337e-01 3.93363237e-01
-1.22140753e+00 -7.51306951e-01 -3.70976120e-01 3.70156169e-01
-7.89144695e-01 1.54932648e-01 -1.00279748e+00 -6.64472580e-01
7.64103532e-02 -1.47973442e+00 1.33771920e+00 3.03717852e-01
-1.55931100e-01 -1.01450396e+00 3.62360408e-03 3.69617373e-01
1.48364119e-02 2.77330577e-01 4.33921635e-01 -5.84329903e-01
-1.21444714e+00 1.55839205e-01 6.26394525e-04 6.04139447e-01
-3.56514424e-01 4.40621346e-01 -9.16409016e-01 -1.02190681e-01
-1.30712375e-01 -1.15057796e-01 1.24677527e+00 1.22358990e+00
1.52626765e+00 -3.59601192e-02 -6.53381288e-01 8.97876203e-01
1.22718930e+00 3.39242786e-01 6.66473448e-01 -2.79997755e-02
8.93783987e-01 3.99562985e-01 1.68583825e-01 3.88945311e-01
3.37319732e-01 4.75943029e-01 4.17650133e-01 -1.60791799e-01
-8.72814730e-02 -1.13322042e-01 4.89196330e-01 4.18108642e-01
-9.68541503e-02 -5.49362600e-01 -8.29067707e-01 7.34917939e-01
-2.05488110e+00 -1.29181707e+00 -9.71900597e-02 1.63339937e+00
4.88550603e-01 1.10050388e-01 5.72358072e-02 -5.62570453e-01
7.63023436e-01 2.91707724e-01 -6.38433576e-01 2.60080546e-02
-2.48412669e-01 3.14210832e-01 6.74691260e-01 4.21356499e-01
-1.58739054e+00 1.30767810e+00 7.45126486e+00 7.87354648e-01
-6.28446400e-01 3.48816782e-01 8.67430329e-01 -2.98014015e-01
-6.90190643e-02 -2.98003942e-01 -9.99072313e-01 3.64223778e-01
9.14883733e-01 5.75170457e-01 1.47096097e-01 1.05271113e+00
1.30377218e-01 2.52222151e-01 -1.06583011e+00 9.88246500e-01
-3.58761624e-02 -1.76200235e+00 -8.71188659e-03 6.79730549e-02
1.04135954e+00 3.88184875e-01 2.43022472e-01 3.53581876e-01
7.52493560e-01 -1.11046231e+00 8.94055247e-01 6.82857096e-01
4.39153552e-01 -5.95759213e-01 4.45169210e-01 2.79954314e-01
-1.15881228e+00 -1.64741576e-01 -2.95854241e-01 2.50957459e-01
4.99854922e-01 6.30592480e-02 -6.03984654e-01 1.71882704e-01
8.00370157e-01 9.44121242e-01 -3.20873231e-01 8.38418484e-01
-2.88779885e-01 6.72777593e-01 -2.09522739e-01 4.04281527e-01
4.01891053e-01 -2.29101837e-01 7.17677712e-01 1.27091813e+00
1.00637272e-01 -9.97130349e-02 3.22164983e-01 1.16413355e+00
-7.81842768e-02 -4.74615455e-01 -2.76644260e-01 9.76317301e-02
2.83594429e-01 1.15961349e+00 -9.91537094e-01 -6.60517454e-01
-6.82474196e-01 1.27970052e+00 1.57512277e-01 6.99353874e-01
-1.35578561e+00 4.34887558e-01 9.34623957e-01 -2.88521647e-01
1.27374935e+00 -3.67879570e-01 -1.09170586e-01 -1.34567857e+00
-1.45472094e-01 -8.30643296e-01 2.89428443e-01 -6.16205573e-01
-1.12156153e+00 5.07331371e-01 -8.20508078e-02 -9.59711790e-01
-5.53181529e-01 -8.70331228e-01 -5.00281870e-01 6.99639916e-01
-1.08270407e+00 -1.14675808e+00 8.28661174e-02 7.45799065e-01
9.87451553e-01 -2.53678918e-01 5.28180838e-01 3.10687691e-01
-8.28004003e-01 2.87693352e-01 2.03579098e-01 5.33306241e-01
3.05080742e-01 -1.38963211e+00 5.46654344e-01 1.17798471e+00
4.91738826e-01 2.71324784e-01 6.31508052e-01 -8.58000934e-01
-1.00608778e+00 -1.41010833e+00 1.82389796e-01 -9.13221598e-01
5.02964675e-01 -5.14688909e-01 -9.16852117e-01 1.27602696e+00
2.09275886e-01 2.28394091e-01 6.70010090e-01 -1.75058067e-01
-1.64606109e-01 3.77152324e-01 -7.81427622e-01 4.99443352e-01
8.98122251e-01 -3.53806674e-01 -5.97138405e-01 1.49084300e-01
8.40952456e-01 -5.96471429e-01 -7.03574479e-01 3.44185859e-01
5.05678654e-01 -8.78854930e-01 1.18183947e+00 -7.16010928e-01
6.85757697e-01 -3.16410989e-01 -2.14596659e-01 -8.28764260e-01
-6.08983755e-01 -4.97804374e-01 -8.21559250e-01 9.20075178e-01
4.79032129e-01 2.04519629e-01 1.11586022e+00 9.00634229e-01
-2.16519177e-01 -7.44300961e-01 -8.24516714e-01 -5.37665725e-01
-7.42783444e-03 -7.43883491e-01 1.83292195e-01 6.47149801e-01
-8.43127012e-01 1.67399406e-01 -6.21476531e-01 5.69251895e-01
7.90161371e-01 -2.13912472e-01 7.19342947e-01 -9.45215106e-01
-8.19289744e-01 -4.88401204e-01 -4.61656988e-01 -1.50025046e+00
4.57921475e-01 -5.77881813e-01 1.89792812e-01 -1.43881297e+00
2.57668078e-01 9.50895846e-02 -3.62126261e-01 3.74125361e-01
-2.18614876e-01 1.78269789e-01 2.18151867e-01 9.96592343e-02
-9.74795282e-01 5.86242199e-01 1.12860453e+00 -2.18564957e-01
-2.96717614e-01 1.21544097e-02 -1.63018048e-01 9.49469566e-01
2.96060145e-01 -4.29387629e-01 -3.32466096e-01 -6.36602044e-01
-2.04814374e-01 2.52121419e-01 6.66505873e-01 -1.06143081e+00
1.70668662e-01 -1.25877336e-01 1.02615476e+00 -8.39052677e-01
4.64667708e-01 -7.74671674e-01 2.12068528e-01 1.01924591e-01
-2.99625456e-01 -3.62042710e-02 2.56889910e-01 9.10340250e-01
-1.88843563e-01 6.33409619e-02 7.66445100e-01 -2.04356954e-01
-1.25583029e+00 7.05351472e-01 -5.72430313e-01 4.99505848e-02
1.26669085e+00 -3.30723315e-01 1.96686804e-01 -2.32501790e-01
-1.06195498e+00 4.98788208e-01 2.52760470e-01 6.93364382e-01
7.42219687e-01 -9.47745323e-01 -6.39737606e-01 2.31459469e-01
-3.71660888e-01 3.27096790e-01 3.79270703e-01 6.30610883e-01
-4.01388347e-01 2.12790385e-01 -1.00816205e-01 -9.36245978e-01
-1.21035278e+00 5.80541432e-01 4.45980579e-01 -2.23771736e-01
-8.88056517e-01 1.34379995e+00 3.75243604e-01 1.75757427e-02
4.20279324e-01 -3.56681079e-01 -2.69842625e-01 -3.08717601e-02
5.47903001e-01 2.33493790e-01 -5.87904453e-01 -7.68091440e-01
-1.49338141e-01 2.26174951e-01 -8.90377983e-02 -2.28883028e-01
1.42131948e+00 -2.26088509e-01 1.36110693e-01 4.28927660e-01
1.04874325e+00 -2.97557533e-01 -2.33059525e+00 -6.23999871e-02
-9.80347022e-02 -1.27272904e-01 1.80369720e-01 -3.99218917e-01
-1.22081220e+00 9.48083222e-01 4.42179114e-01 -1.43720642e-01
7.87319481e-01 2.43835360e-01 7.60382533e-01 9.66479629e-03
6.46161065e-02 -1.14593065e+00 1.02578089e-01 5.28945565e-01
3.84371728e-01 -1.18457615e+00 -3.76347363e-01 -1.44641176e-01
-9.41761672e-01 9.57709014e-01 8.60296965e-01 -3.49890858e-01
6.27832472e-01 5.49414754e-01 -1.33264791e-02 -1.84320182e-01
-9.36467707e-01 -5.47201447e-02 4.71509546e-01 7.34517872e-01
1.28716946e-01 2.49001663e-02 4.39008236e-01 7.54150808e-01
2.52020240e-01 -7.71373659e-02 1.50642768e-01 5.37646174e-01
-3.57654393e-01 -8.13994408e-01 -1.92528307e-01 4.21424329e-01
-7.14779735e-01 6.55894692e-04 8.50894377e-02 8.65065396e-01
4.65280950e-01 6.08656585e-01 5.82915485e-01 -2.70977557e-01
-2.85380423e-01 1.93808302e-01 4.69597071e-01 -4.34146285e-01
-3.25776994e-01 5.69518149e-01 -3.33023816e-01 -9.15883958e-01
-4.58017021e-01 -8.08163941e-01 -1.22781765e+00 1.03605703e-01
-1.97430164e-01 -2.11953029e-01 2.57330209e-01 1.09369016e+00
4.32280689e-01 1.05269468e+00 9.14006531e-02 -1.06414676e+00
-3.47727805e-01 -6.29624367e-01 -5.42299807e-01 3.20497274e-01
5.48772454e-01 -6.76453292e-01 -4.89988225e-03 5.74586272e-01] | [9.149539947509766, -0.03411867097020149] |
926a5d3f-6fce-40ad-8f05-4685b894d5b6 | from-image-collections-to-point-clouds-with | 2005.01939 | null | https://arxiv.org/abs/2005.01939v1 | https://arxiv.org/pdf/2005.01939v1.pdf | From Image Collections to Point Clouds with Self-supervised Shape and Pose Networks | Reconstructing 3D models from 2D images is one of the fundamental problems in computer vision. In this work, we propose a deep learning technique for 3D object reconstruction from a single image. Contrary to recent works that either use 3D supervision or multi-view supervision, we use only single view images with no pose information during training as well. This makes our approach more practical requiring only an image collection of an object category and the corresponding silhouettes. We learn both 3D point cloud reconstruction and pose estimation networks in a self-supervised manner, making use of differentiable point cloud renderer to train with 2D supervision. A key novelty of the proposed technique is to impose 3D geometric reasoning into predicted 3D point clouds by rotating them with randomly sampled poses and then enforcing cycle consistency on both 3D reconstructions and poses. In addition, using single-view supervision allows us to do test-time optimization on a given test image. Experiments on the synthetic ShapeNet and real-world Pix3D datasets demonstrate that our approach, despite using less supervision, can achieve competitive performance compared to pose-supervised and multi-view supervised approaches. | ['Wei-Chih Hung', 'Shashank Kashyap', 'R. Venkatesh Babu', 'K L Navaneet', 'Ansu Mathew', 'Varun Jampani'] | 2020-05-05 | from-image-collections-to-point-clouds-with-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Navaneet_From_Image_Collections_to_Point_Clouds_With_Self-Supervised_Shape_and_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Navaneet_From_Image_Collections_to_Point_Clouds_With_Self-Supervised_Shape_and_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-point-cloud-reconstruction', 'point-cloud-reconstruction', '3d-object-reconstruction', '3d-object-reconstruction-from-a-single-image'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-9.59976092e-02 1.47826955e-01 8.92520230e-03 -4.00537342e-01
-5.70465744e-01 -7.35081017e-01 8.13159943e-01 -2.88082778e-01
-2.91521937e-01 1.98167056e-01 -4.40192044e-01 -1.40639767e-01
7.95902461e-02 -6.75724924e-01 -1.27495122e+00 -4.13174927e-01
3.38513076e-01 1.18713188e+00 3.15510809e-01 -2.49997899e-02
2.53691167e-01 9.70136225e-01 -1.59487331e+00 -1.97235420e-01
5.08899927e-01 9.41312492e-01 3.66739184e-01 4.58007753e-01
-3.60851884e-02 1.25082642e-01 -1.39230996e-01 -2.40954459e-01
8.11226070e-01 2.33981386e-02 -5.23122013e-01 8.09887528e-01
9.21880662e-01 -5.44470489e-01 8.22732002e-02 8.93177629e-01
4.12197977e-01 -1.21069565e-01 5.76692104e-01 -1.15622103e+00
-2.82929450e-01 -2.14046106e-01 -7.27547407e-01 -3.53738040e-01
4.74248916e-01 -1.52112376e-02 7.81615078e-01 -1.19167507e+00
8.55197787e-01 1.16577291e+00 7.55095482e-01 4.57521051e-01
-1.39729953e+00 -3.21594954e-01 2.35624433e-01 -6.19015582e-02
-1.25461173e+00 -3.15602154e-01 1.20298541e+00 -5.28228045e-01
9.87032831e-01 5.61559154e-03 8.70205700e-01 8.61878753e-01
-9.08827707e-02 5.88626266e-01 1.09088397e+00 -4.35757190e-01
1.81019738e-01 1.58974305e-01 -2.55357087e-01 8.32204163e-01
1.00487433e-01 1.20250784e-01 -3.01448971e-01 -8.33031163e-02
1.24586463e+00 3.01673383e-01 -1.80206895e-01 -1.33633697e+00
-1.31841516e+00 6.07161939e-01 4.25031513e-01 -1.60643578e-01
-3.76800478e-01 1.69945508e-01 1.15409330e-01 1.54363751e-01
7.92335927e-01 9.06672031e-02 -6.56100869e-01 1.06275052e-01
-6.85996294e-01 1.25827685e-01 7.36917198e-01 1.13928175e+00
1.08416653e+00 5.44156395e-02 6.68104470e-01 5.52378476e-01
4.47648495e-01 7.98830450e-01 1.30950585e-01 -1.10272145e+00
3.69524151e-01 7.45330811e-01 2.11750269e-01 -8.28146815e-01
-3.48930478e-01 -4.55542177e-01 -5.99708080e-01 6.48403943e-01
2.42195070e-01 2.95933127e-01 -1.08645105e+00 1.28320837e+00
6.74878597e-01 2.24639401e-01 -1.02182411e-01 9.47434306e-01
4.87696350e-01 2.71062702e-01 -7.45992899e-01 2.79559791e-02
1.02886081e+00 -8.46074462e-01 -9.97858271e-02 -2.34769613e-01
3.24504167e-01 -7.42015481e-01 8.89836073e-01 3.97791147e-01
-1.16068578e+00 -5.67804515e-01 -9.99762952e-01 -2.42969781e-01
-6.60075322e-02 2.95273568e-02 2.83658504e-01 3.31921160e-01
-9.15704548e-01 6.62862182e-01 -1.07244062e+00 -1.68140620e-01
3.63991350e-01 3.71682167e-01 -7.02610254e-01 -1.78968161e-01
-4.42600340e-01 1.02090752e+00 2.57898420e-01 5.78138232e-03
-9.29195523e-01 -7.49727309e-01 -1.00678635e+00 -2.69687027e-01
4.74249899e-01 -1.01292646e+00 1.08657455e+00 -7.02994764e-01
-1.51636136e+00 1.35750389e+00 -1.48260370e-01 -2.95771658e-01
7.61622190e-01 -4.11588430e-01 3.16497475e-01 3.08143914e-01
8.33976045e-02 7.70655990e-01 9.51232910e-01 -1.76796913e+00
-2.29693115e-01 -7.41846800e-01 2.26177037e-01 4.17607039e-01
1.63127795e-01 -5.42463303e-01 -6.82155550e-01 -2.07221042e-02
7.80530393e-01 -1.05319798e+00 -2.39378557e-01 4.79777664e-01
-3.99302900e-01 -1.67573839e-01 9.98772919e-01 -4.43108708e-01
-5.09896539e-02 -1.90568614e+00 4.44550723e-01 1.86064258e-01
9.74190608e-02 3.74092050e-02 -6.70125484e-02 1.94258079e-01
-4.07418720e-02 -1.42728537e-01 -1.64415762e-01 -8.88418615e-01
-6.86256215e-02 5.13574719e-01 -2.56115019e-01 8.79159689e-01
2.44166717e-01 9.02671039e-01 -7.60249972e-01 -2.37016290e-01
6.47353590e-01 6.87040627e-01 -7.04017699e-01 2.58027911e-01
-6.23282433e-01 8.25349033e-01 -4.28956270e-01 5.06147563e-01
8.72174680e-01 -5.23365915e-01 -1.46387666e-01 -2.04566821e-01
-1.39548272e-01 2.95586973e-01 -1.21265042e+00 2.26314712e+00
-7.20023572e-01 1.14966787e-01 1.15982234e-01 -1.14501762e+00
9.99928534e-01 3.13968420e-01 6.56768382e-01 -3.18720818e-01
2.96453554e-02 2.62099683e-01 -4.73153293e-01 -2.25203007e-01
2.23825499e-03 -2.90198058e-01 3.30074370e-01 4.89204168e-01
1.42190263e-01 -8.63388658e-01 -2.92775869e-01 -3.62822562e-02
5.55085838e-01 8.00036252e-01 2.10066631e-01 1.18600920e-01
5.01683533e-01 -4.90235677e-03 5.43264091e-01 1.87144369e-01
3.42052430e-01 1.03300977e+00 2.50787735e-01 -6.75559819e-01
-1.41210973e+00 -1.21932411e+00 -1.83853805e-01 2.19975635e-01
2.93295860e-01 -1.17406972e-01 -3.72506380e-01 -8.90245557e-01
1.22908086e-01 5.02342403e-01 -3.29247147e-01 2.56322354e-01
-7.89138198e-01 -1.00266755e-01 -6.63444176e-02 3.61322254e-01
4.22521472e-01 -5.99709272e-01 -9.01679873e-01 -1.78695247e-01
3.79932262e-02 -1.39718735e+00 -3.68490934e-01 1.97739467e-01
-1.25646031e+00 -1.15128732e+00 -6.81047738e-01 -8.41618121e-01
1.04868996e+00 5.72088003e-01 1.17859387e+00 -1.68659072e-02
-8.54775310e-03 8.08466136e-01 -1.72355086e-01 -3.71663243e-01
-3.12766343e-01 -1.12745740e-01 6.05669953e-02 3.00627910e-02
2.27213558e-02 -1.03290761e+00 -5.17946541e-01 3.99798393e-01
-6.47207618e-01 3.12853783e-01 4.81523633e-01 7.72550583e-01
1.11649442e+00 -3.93850982e-01 1.72011614e-01 -7.87454188e-01
-3.04978669e-01 -1.24190778e-01 -1.09346366e+00 8.08170289e-02
-5.31781673e-01 -9.42033082e-02 6.25735581e-01 -2.56745666e-01
-7.26258934e-01 6.95133030e-01 -3.55368108e-02 -1.22458649e+00
-3.32735211e-01 1.15250926e-02 -2.28720859e-01 -1.58532768e-01
5.04364550e-01 3.89213204e-01 4.89751399e-01 -7.11373329e-01
3.04613382e-01 1.24981567e-01 2.84357011e-01 -3.40368629e-01
1.28265357e+00 1.01547468e+00 2.80667186e-01 -7.14111984e-01
-8.89884591e-01 -6.42423928e-01 -1.27440846e+00 -2.15590730e-01
8.32290530e-01 -1.10527241e+00 -7.49305785e-01 3.23845088e-01
-1.39325678e+00 -1.41820192e-01 -2.38391504e-01 5.66178620e-01
-9.88548219e-01 5.08536994e-01 -8.26545656e-02 -6.29670858e-01
-1.42667532e-01 -1.28012788e+00 1.67530537e+00 -3.77016872e-01
1.85456023e-01 -1.05601048e+00 -7.76239708e-02 4.60150063e-01
-1.04019374e-01 5.20988643e-01 5.31101942e-01 -2.70801872e-01
-1.10616422e+00 -1.94683596e-01 5.62028363e-02 4.92415994e-01
2.12096810e-01 -3.17584962e-01 -9.73132491e-01 -4.18513685e-01
4.52562928e-01 -4.68217343e-01 4.61040944e-01 2.26448685e-01
1.00640702e+00 3.59993279e-02 -2.83881098e-01 9.75093842e-01
1.55040395e+00 -1.46654859e-01 1.96189061e-01 3.37057203e-01
1.01381421e+00 5.29067576e-01 5.74476898e-01 3.39687735e-01
4.80423599e-01 8.47890198e-01 1.06995332e+00 -3.77283394e-02
7.97547922e-02 -4.68135148e-01 7.62442946e-02 7.02272594e-01
-1.94199651e-01 2.18440756e-01 -1.03693318e+00 3.49676937e-01
-1.64180100e+00 -6.05461478e-01 -1.12701498e-01 2.32301712e+00
4.63809282e-01 9.06916484e-02 -1.03397891e-01 1.03577107e-01
3.62512589e-01 8.28380063e-02 -8.67814124e-01 1.75902694e-01
1.78090945e-01 1.40148461e-01 5.89924335e-01 5.84502995e-01
-9.00555015e-01 8.09933364e-01 5.28793049e+00 2.36981839e-01
-1.34225345e+00 2.64015254e-02 3.12423140e-01 -1.92413777e-01
-5.27175844e-01 1.88106745e-01 -6.89673364e-01 1.62901916e-02
2.92211294e-01 2.05374286e-01 4.29212362e-01 1.01482987e+00
5.05803153e-02 6.03032522e-02 -1.52607167e+00 1.25989330e+00
4.48919326e-01 -1.37154162e+00 1.70960888e-01 2.30938002e-01
7.57349193e-01 3.46618742e-01 -1.26730204e-01 -1.47987723e-01
8.27080384e-02 -7.41101205e-01 9.78627682e-01 4.44835186e-01
7.31029451e-01 -5.95221460e-01 2.87057310e-01 1.00516021e+00
-8.26123118e-01 3.50660503e-01 -4.13495094e-01 -8.55319649e-02
3.52058709e-01 7.00641453e-01 -1.05297434e+00 9.17814255e-01
7.11607516e-01 1.04162610e+00 -3.42425585e-01 7.41105199e-01
-3.20821285e-01 8.21118876e-02 -5.85983634e-01 3.15416664e-01
5.17724827e-02 -4.96792734e-01 7.69911706e-01 3.80900472e-01
3.09309334e-01 -1.57290623e-01 4.23367351e-01 9.53682661e-01
7.38056973e-02 -1.87455550e-01 -9.69995141e-01 4.92711663e-01
2.27496907e-01 1.14057422e+00 -7.77650237e-01 -1.46238670e-01
-3.93460423e-01 1.02530205e+00 3.43815386e-01 1.52861282e-01
-5.72490931e-01 3.00896078e-01 3.57387155e-01 3.27588707e-01
7.01091349e-01 -7.45588481e-01 -3.69695634e-01 -1.38391089e+00
4.64965075e-01 -4.96038079e-01 -1.54447794e-01 -9.28018630e-01
-1.16764939e+00 4.70233858e-01 2.70396471e-01 -1.79794240e+00
-3.78969759e-01 -6.68676078e-01 -3.54175985e-01 8.67761433e-01
-1.81895149e+00 -1.40347672e+00 -2.78446674e-01 5.54639280e-01
6.25701904e-01 5.21867089e-02 6.69080436e-01 -7.08590960e-03
1.00762010e-01 1.00530058e-01 -2.40986630e-01 -7.14338347e-02
5.05858660e-01 -1.27958751e+00 6.62661374e-01 4.48775947e-01
4.93840545e-01 4.71068442e-01 4.29097980e-01 -4.67198223e-01
-1.73595977e+00 -9.44514990e-01 5.06382644e-01 -8.06475997e-01
2.00578034e-01 -6.00804210e-01 -8.31233084e-01 8.63810360e-01
2.59745345e-02 3.71180952e-01 2.04924315e-01 -1.08278066e-01
-4.29486156e-01 -8.88476595e-02 -1.13922012e+00 2.48294324e-01
1.35555327e+00 -3.90712708e-01 -5.26178718e-01 5.70262611e-01
7.22654998e-01 -9.72293258e-01 -7.45450437e-01 4.92078274e-01
3.24781239e-01 -1.02887225e+00 1.18380332e+00 -2.93513179e-01
3.95072520e-01 -5.37328064e-01 -1.11348599e-01 -1.38801670e+00
1.76550806e-01 -3.92964661e-01 -6.73651248e-02 7.12255597e-01
1.27858087e-01 -7.00183988e-01 1.03888488e+00 1.67197853e-01
-3.47951412e-01 -8.88224959e-01 -1.00075722e+00 -8.58465135e-01
3.15577053e-02 -4.05562729e-01 4.54321086e-01 8.75905752e-01
-7.86458671e-01 3.63934338e-01 -1.99831948e-01 5.71913898e-01
1.00478232e+00 6.81543708e-01 1.30002689e+00 -1.44793069e+00
-4.32830930e-01 -8.09011534e-02 -6.28708899e-01 -1.37950420e+00
3.72642905e-01 -1.05823004e+00 -9.74596366e-02 -1.48091447e+00
-7.15652704e-02 -5.06944835e-01 3.73013884e-01 3.93992633e-01
3.70194674e-01 3.24336916e-01 1.86511099e-01 4.13207263e-01
-2.49885574e-01 6.87338829e-01 1.52948868e+00 1.28140882e-01
-5.79598583e-02 1.51453599e-01 -1.98662370e-01 9.84208405e-01
5.37071943e-01 -2.71589309e-01 -6.24381781e-01 -8.52495015e-01
1.94325536e-01 1.21160209e-01 8.47870409e-01 -6.70614898e-01
1.35652870e-01 -1.17899776e-01 4.89567816e-01 -1.20954311e+00
7.91756094e-01 -1.17357671e+00 1.39261186e-01 3.47178727e-01
1.28027231e-01 1.32291228e-01 1.23203456e-01 6.86371386e-01
6.06673770e-02 -1.58921748e-01 6.90331578e-01 -5.51036298e-01
-4.62425441e-01 6.30625546e-01 3.84077996e-01 -3.10944289e-01
1.07679963e+00 -4.84928191e-01 2.72596389e-01 -2.09112614e-01
-7.42468119e-01 2.28465304e-01 1.05827761e+00 3.95693749e-01
8.66340160e-01 -1.32230127e+00 -6.39887214e-01 5.32015324e-01
1.07141696e-01 9.65505123e-01 1.26746744e-02 6.69708371e-01
-6.82522178e-01 3.13142449e-01 -1.03590541e-01 -1.44818199e+00
-1.16362417e+00 5.08924723e-01 4.99053895e-01 1.61184259e-02
-1.02140605e+00 5.00890970e-01 3.52989316e-01 -1.06160665e+00
1.83476374e-01 -2.77198583e-01 1.89730838e-01 -4.35358286e-01
2.81654187e-02 2.13512741e-02 2.88249761e-01 -6.00496650e-01
-2.38713890e-01 1.05114663e+00 -1.13299508e-02 -2.56111860e-01
1.67604196e+00 9.92419273e-02 -1.07834555e-01 6.48178458e-01
1.36357760e+00 -3.90711837e-02 -1.61515534e+00 -2.94904560e-01
-1.72159374e-01 -6.51872575e-01 -1.04101170e-02 -3.27647895e-01
-1.06516516e+00 9.76797760e-01 3.25199068e-01 -1.42061144e-01
8.14983666e-01 2.39258841e-01 6.14928663e-01 5.47027230e-01
6.64066553e-01 -6.60013855e-01 3.00817698e-01 5.79614282e-01
1.16504383e+00 -1.37545502e+00 1.55705541e-01 -5.41225553e-01
-2.02911958e-01 1.08377230e+00 7.01375008e-01 -6.42180085e-01
6.95116043e-01 -1.88767519e-02 6.81451485e-02 -3.64909440e-01
-6.46015882e-01 1.99904386e-02 3.82992923e-01 5.22988737e-01
6.11985363e-02 -2.68447429e-01 1.86709508e-01 -3.20887536e-01
-2.04783753e-01 -1.76936194e-01 3.95746678e-01 1.01630390e+00
-1.29274353e-01 -1.26548684e+00 -4.52207267e-01 1.27920091e-01
-2.61395946e-02 3.62036854e-01 -4.08514202e-01 8.77083480e-01
2.60679945e-02 3.19400012e-01 2.56000876e-01 -2.01644793e-01
5.75516343e-01 8.45161229e-02 9.14192915e-01 -9.72358882e-01
-1.65526167e-01 2.51236349e-01 -3.33841383e-01 -5.47128081e-01
-7.55305767e-01 -7.62641132e-01 -1.26382089e+00 5.40988147e-02
-4.16271418e-01 -1.58767805e-01 1.03640580e+00 9.46969509e-01
4.58814412e-01 8.00227672e-02 9.92854416e-01 -1.50597930e+00
-7.76415050e-01 -6.05275273e-01 -4.15621012e-01 4.61071342e-01
5.79818010e-01 -8.80142689e-01 -4.30750728e-01 1.28335238e-01] | [8.3908052444458, -3.0051615238189697] |
9d045533-82ce-472c-a403-6c1e3eb7aa2a | transfer-learning-in-biomedical-natural | 1906.05474 | null | https://arxiv.org/abs/1906.05474v2 | https://arxiv.org/pdf/1906.05474v2.pdf | Transfer Learning in Biomedical Natural Language Processing: An Evaluation of BERT and ELMo on Ten Benchmarking Datasets | null | ['Yifan Peng', 'Zhiyong Lu', 'Shankai Yan'] | 2019-06-13 | transfer-learning-in-biomedical-natural-1 | https://aclanthology.org/W19-5006 | https://aclanthology.org/W19-5006.pdf | ws-2019-8 | ['medical-relation-extraction', 'drug-drug-interaction-extraction', 'medical-named-entity-recognition'] | ['medical', 'natural-language-processing', '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] | [-7.399508953094482, 3.747267007827759] |
67595424-a8e4-4f56-9f16-7dac84ec0966 | objectron-a-large-scale-dataset-of-object | 2012.09988 | null | https://arxiv.org/abs/2012.09988v1 | https://arxiv.org/pdf/2012.09988v1.pdf | Objectron: A Large Scale Dataset of Object-Centric Videos in the Wild with Pose Annotations | 3D object detection has recently become popular due to many applications in robotics, augmented reality, autonomy, and image retrieval. We introduce the Objectron dataset to advance the state of the art in 3D object detection and foster new research and applications, such as 3D object tracking, view synthesis, and improved 3D shape representation. The dataset contains object-centric short videos with pose annotations for nine categories and includes 4 million annotated images in 14,819 annotated videos. We also propose a new evaluation metric, 3D Intersection over Union, for 3D object detection. We demonstrate the usefulness of our dataset in 3D object detection tasks by providing baseline models trained on this dataset. Our dataset and evaluation source code are available online at http://www.objectron.dev | ['Matthias Grundmann', 'Artsiom Ablavatski', 'Jianing Wei', 'Liangkai Zhang', 'Adel Ahmadyan'] | 2020-12-18 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Ahmadyan_Objectron_A_Large_Scale_Dataset_of_Object-Centric_Videos_in_the_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Ahmadyan_Objectron_A_Large_Scale_Dataset_of_Object-Centric_Videos_in_the_CVPR_2021_paper.pdf | cvpr-2021-1 | ['3d-shape-representation', '3d-object-tracking'] | ['computer-vision', 'computer-vision'] | [-1.54207990e-01 -2.58425534e-01 -3.36068749e-01 -2.69282877e-01
-5.08812606e-01 -7.26198852e-01 7.66781807e-01 -6.18827641e-02
-3.05826247e-01 4.45443392e-02 2.98347056e-01 1.69288125e-02
2.34588474e-01 -1.63520411e-01 -9.93832588e-01 -1.89856932e-01
-3.24388385e-01 5.14314473e-01 6.28526628e-01 8.49501342e-02
1.93068817e-01 1.02448320e+00 -1.65223706e+00 -7.91437626e-02
1.58243626e-01 1.40330029e+00 2.69452304e-01 6.26066983e-01
1.81704700e-01 2.65573114e-01 -5.54636776e-01 -1.15218602e-01
6.01769030e-01 5.22045530e-02 -3.97845417e-01 4.08838630e-01
6.28847718e-01 -6.30576670e-01 -4.35496867e-01 1.01600182e+00
6.27161920e-01 6.78425506e-02 7.15753019e-01 -1.66722941e+00
-6.23084962e-01 8.89575481e-02 -7.28629112e-01 3.03174466e-01
6.90139413e-01 3.07823658e-01 7.33309090e-01 -1.36312449e+00
9.71413195e-01 1.50357842e+00 4.27373260e-01 6.91013336e-01
-9.40310895e-01 -5.59157372e-01 1.47522643e-01 1.73788127e-02
-1.41703212e+00 -4.72158074e-01 6.00890160e-01 -7.04651177e-01
1.20745969e+00 -1.23826172e-02 9.75992262e-01 1.04991865e+00
8.30773823e-03 1.32992971e+00 6.94850326e-01 -2.83667386e-01
-9.36095193e-02 -4.58903291e-04 1.84626821e-02 7.35707462e-01
5.54969430e-01 3.20034593e-01 -4.74326164e-01 7.25001544e-02
9.20860827e-01 1.14093482e-01 -4.16259058e-02 -1.23076630e+00
-1.49500418e+00 5.14458060e-01 6.12278104e-01 -3.29030156e-01
-2.09316969e-01 2.61818022e-01 3.93909961e-01 -8.00018609e-02
3.43268782e-01 2.55697906e-01 -4.47881818e-01 -7.82187805e-02
-7.17779994e-02 5.90096831e-01 5.71808815e-01 1.77333939e+00
4.46180820e-01 -4.35479470e-02 -4.25524190e-02 6.15293026e-01
5.76800883e-01 8.21292341e-01 1.21412314e-01 -1.22956181e+00
1.99669167e-01 9.42369163e-01 2.93225795e-01 -6.35301590e-01
-5.51678836e-01 -1.98363125e-01 -1.90931410e-01 2.71457195e-01
6.60592169e-02 3.80498886e-01 -1.00530028e+00 1.24727643e+00
7.51393378e-01 1.23483583e-03 -1.40406266e-01 1.32087886e+00
1.28610075e+00 3.49253476e-01 -2.54842222e-01 2.39753246e-01
1.41946459e+00 -8.74457598e-01 -3.11241955e-01 -7.99525529e-02
6.42570555e-01 -7.81894445e-01 7.31122851e-01 1.34541139e-01
-9.97156978e-01 -4.78730112e-01 -8.38567793e-01 -1.81496844e-01
-3.60774696e-01 1.48971826e-01 4.65097576e-01 3.83966655e-01
-5.90188563e-01 -2.67050505e-01 -8.06975842e-01 -7.02742696e-01
8.53726745e-01 2.70108521e-01 -5.76496303e-01 -8.23117644e-02
-5.83209038e-01 1.13415205e+00 5.30359089e-01 -2.13243306e-01
-1.22854733e+00 -5.00253320e-01 -1.02372897e+00 -3.88548672e-01
5.87573469e-01 -8.74111891e-01 1.49645805e+00 -4.89631221e-02
-1.01340270e+00 1.47133589e+00 3.32326978e-01 -4.43187028e-01
4.63756353e-01 -6.38188660e-01 -1.86089650e-01 1.15857609e-01
2.94655800e-01 1.07690263e+00 7.20704079e-01 -1.08755767e+00
-7.41930962e-01 -6.34099722e-01 2.74220295e-02 3.76038879e-01
9.53130648e-02 2.71794289e-01 -8.53928685e-01 -4.80516493e-01
3.38946313e-01 -1.12465453e+00 -1.16909027e-01 6.09767199e-01
-3.50371182e-01 -4.87034917e-01 1.11679649e+00 -7.85836130e-02
3.03603083e-01 -2.25496507e+00 2.20909595e-01 -2.84096599e-01
2.84130901e-01 1.10365301e-01 -1.36173561e-01 5.15405508e-03
2.92154431e-01 -3.42251100e-02 2.37192020e-01 -2.86727935e-01
1.18424676e-01 -3.09308637e-02 -9.91666168e-02 6.65910900e-01
3.44730645e-01 1.06909800e+00 -8.34142327e-01 -5.20769775e-01
5.26996434e-01 3.43798995e-01 -6.21448338e-01 2.29418918e-01
-3.23062688e-01 2.74282336e-01 -5.22170663e-01 1.24391031e+00
6.05545044e-01 -1.86652467e-01 -4.52015817e-01 -3.74593824e-01
-3.72432843e-02 1.77666068e-01 -1.14041984e+00 1.97024751e+00
1.26279548e-01 6.94433033e-01 -1.23694055e-01 -5.14414907e-01
1.00387323e+00 5.75149432e-03 6.74325883e-01 -3.90930593e-01
3.61300617e-01 2.17316583e-01 -2.94431329e-01 -4.61443067e-01
6.74743056e-01 4.02091891e-01 -2.24169850e-01 3.06588739e-01
2.67652184e-01 -7.38514841e-01 2.23079711e-01 2.74302065e-01
9.68389153e-01 4.14358765e-01 4.24251348e-01 3.39105502e-02
1.98327869e-01 2.78898537e-01 2.62670517e-01 6.09327793e-01
-5.84055781e-01 5.05307078e-01 9.31323916e-02 -4.72423255e-01
-1.09539235e+00 -1.22427642e+00 -3.61700028e-01 6.43974662e-01
6.55314207e-01 -3.22206736e-01 -1.71942562e-01 -9.54588592e-01
5.95912337e-01 2.20399842e-01 -3.51997554e-01 -1.39925167e-01
-4.81165230e-01 -1.46729261e-01 3.80887806e-01 6.94630682e-01
2.63403356e-01 -8.01905870e-01 -8.72287989e-01 -2.00207993e-01
4.22597304e-02 -1.54993045e+00 -5.47327399e-01 1.18200593e-01
-1.03256440e+00 -1.20795691e+00 -6.59032583e-01 -7.93583691e-01
6.14714444e-01 9.36528981e-01 1.07678151e+00 -2.05266818e-01
-7.07966089e-01 9.21249688e-01 -4.94987547e-01 -9.15627539e-01
-2.68899053e-01 -4.30613421e-02 6.66028023e-01 -5.65996587e-01
5.47995329e-01 2.95524001e-02 -5.62661827e-01 7.15640366e-01
-4.91054952e-01 5.67786815e-03 5.99008203e-01 5.28614044e-01
7.34566629e-01 -5.48809826e-01 2.04712823e-01 -3.00385922e-01
-9.98803154e-02 -2.05620483e-01 -1.01119018e+00 -1.81337357e-01
-2.13892013e-01 -2.89215416e-01 -4.43930954e-01 -4.68346059e-01
-6.31290078e-01 4.59685743e-01 1.21041939e-01 -1.03774095e+00
-4.00422573e-01 -7.65533745e-02 -2.62103975e-01 -3.12433094e-01
7.52164900e-01 -1.70070753e-02 1.08894549e-01 -5.70864260e-01
5.15580773e-01 5.61367810e-01 5.34870744e-01 -2.63382018e-01
8.89616311e-01 4.38376606e-01 6.70632645e-02 -7.57823706e-01
-9.32505548e-01 -8.09584498e-01 -7.92189777e-01 -5.42697728e-01
7.57846951e-01 -1.37716711e+00 -8.10418308e-01 2.84233063e-01
-1.35010600e+00 9.43027623e-03 -3.96471411e-01 8.69599938e-01
-8.51687670e-01 1.66185163e-02 -1.82376087e-01 -7.58709729e-01
-2.35990897e-01 -1.06746805e+00 1.68181527e+00 1.18119657e-01
-3.34883365e-03 -3.51484418e-01 -2.07029119e-01 3.55480134e-01
-1.27046287e-01 2.83369631e-01 4.80020285e-01 -4.85799253e-01
-1.10082662e+00 -4.74130452e-01 -4.35734302e-01 2.22064018e-01
-9.03193746e-03 -1.21788025e-01 -7.45840311e-01 -2.07069233e-01
-4.64626759e-01 -4.41660076e-01 9.18752015e-01 3.21037829e-01
8.16913724e-01 7.49043673e-02 -6.21085882e-01 5.98681927e-01
1.04101562e+00 1.34457573e-01 6.85109571e-02 2.63905466e-01
7.49431670e-01 4.30156410e-01 1.00545299e+00 5.48473418e-01
2.73419082e-01 8.82075667e-01 8.89332116e-01 2.71552026e-01
-3.87692153e-01 -2.80034631e-01 2.58629173e-01 4.38444793e-01
-8.22187960e-02 -2.15450451e-01 -9.95858848e-01 5.36035955e-01
-1.66064417e+00 -7.65873432e-01 -5.67580238e-02 1.96561611e+00
2.88068444e-01 4.68777150e-01 5.10117948e-01 -1.75158009e-01
7.26964295e-01 -1.36777341e-01 -7.15864241e-01 3.36110234e-01
-1.41025886e-01 -3.88348192e-01 6.51873827e-01 -5.00507914e-02
-1.34852278e+00 9.27629590e-01 6.45910883e+00 2.48484731e-01
-7.39275277e-01 -4.47675511e-02 -1.20513111e-01 -2.84424633e-01
1.50314972e-01 -2.47969195e-01 -1.37133718e+00 9.65959057e-02
6.12561703e-02 -1.63261324e-01 -5.27440384e-02 1.25584054e+00
3.28255184e-02 -1.31832004e-01 -1.30345666e+00 1.21176207e+00
3.16015363e-01 -1.36184418e+00 2.34507650e-01 8.61676559e-02
7.73271024e-01 3.71587396e-01 -3.04259658e-02 1.90243036e-01
9.05512646e-02 -4.28050578e-01 1.19804156e+00 2.28974596e-01
7.38560498e-01 -3.67935121e-01 5.14125109e-01 3.70380044e-01
-1.11110246e+00 -4.11338098e-02 -3.75073195e-01 6.69871420e-02
2.01714426e-01 2.17125997e-01 -1.21850646e+00 1.87739283e-01
1.04437542e+00 1.15816808e+00 -6.70958698e-01 1.66794729e+00
-8.68184939e-02 -7.59896263e-03 -4.71080989e-01 -3.07077050e-01
-2.30247788e-02 1.86917230e-01 1.26174426e+00 8.96900654e-01
2.73574501e-01 1.16503298e-01 4.00657654e-01 6.95238709e-01
-2.92805523e-01 -1.92859367e-01 -1.06654131e+00 -3.05433925e-02
6.99940324e-01 1.09096909e+00 -9.00101364e-01 -2.17808947e-01
-5.27091801e-01 6.08796179e-01 -7.66117424e-02 2.48755924e-02
-8.05132985e-01 -3.24518025e-01 8.69918823e-01 1.24351144e-01
5.28781474e-01 -6.80891573e-01 5.26330881e-02 -1.13326430e+00
-1.95086980e-03 -5.45092046e-01 2.55650252e-01 -1.05828631e+00
-1.10339069e+00 2.05347985e-01 4.10344213e-01 -1.65568912e+00
-1.54630020e-01 -1.06329215e+00 8.70719627e-02 2.14950487e-01
-1.28669977e+00 -1.15890336e+00 -6.62660003e-01 4.07726854e-01
7.40548193e-01 -2.45177582e-01 4.90394443e-01 3.46899182e-01
-1.66311651e-01 1.62281051e-01 -2.58715838e-01 1.56472698e-01
7.64159024e-01 -9.51658785e-01 6.93374276e-01 4.95084375e-01
4.07709628e-01 1.79123268e-01 4.34805810e-01 -6.07189059e-01
-2.01581454e+00 -1.16485727e+00 3.33574414e-01 -1.13759339e+00
4.38177913e-01 -6.46956861e-01 -3.99717420e-01 7.90238917e-01
-3.93334419e-01 4.55141306e-01 2.18738675e-01 -1.94686651e-01
-4.14687812e-01 1.35772780e-03 -9.63386178e-01 4.00461286e-01
1.71248031e+00 -2.38880679e-01 -4.67295051e-01 4.56050754e-01
1.04873478e+00 -1.01857960e+00 -8.29472363e-01 8.06304336e-01
8.31738770e-01 -6.55301273e-01 1.31203115e+00 -5.40185571e-01
-1.20191127e-01 -5.83364546e-01 -5.19145072e-01 -8.07971299e-01
-2.05462351e-01 -3.34564954e-01 -7.14805663e-01 7.12939918e-01
1.08835950e-01 -1.85916394e-01 9.10270095e-01 1.94622472e-01
-4.27919686e-01 -5.92105150e-01 -9.12209868e-01 -1.14014006e+00
-5.65120459e-01 -4.78597879e-01 5.24355352e-01 3.68415266e-01
-5.32601953e-01 3.97659421e-01 -1.29236072e-01 2.31095433e-01
8.48658323e-01 4.87334818e-01 1.42324054e+00 -1.43715835e+00
1.66210458e-01 -6.41446769e-01 -1.11673439e+00 -1.63189316e+00
-3.38192917e-02 -9.53956068e-01 6.87332600e-02 -1.43368459e+00
2.78302670e-01 -4.29514676e-01 -1.84912700e-02 3.90559405e-01
2.20251381e-01 6.92617834e-01 4.93901193e-01 2.64235914e-01
-1.05727077e+00 7.17588127e-01 1.46914732e+00 -2.05098212e-01
-1.61879599e-01 1.15452819e-01 -2.51834214e-01 6.77006781e-01
5.98174810e-01 -3.79032105e-01 1.46662012e-01 -5.44847071e-01
-1.87271491e-01 -1.59094393e-01 8.03919971e-01 -9.73879278e-01
1.51439775e-02 -7.55055994e-02 6.96982622e-01 -1.26414263e+00
7.41343260e-01 -1.01514935e+00 -5.60596585e-02 4.73982662e-01
-7.42341802e-02 -4.22808975e-02 3.19020241e-01 6.75923228e-01
1.64209381e-01 -1.71039836e-03 6.30493999e-01 -1.64327428e-01
-1.32970452e+00 5.89832723e-01 2.34167986e-02 1.34321034e-01
1.36399758e+00 -4.12749559e-01 -2.07590938e-01 -1.06430195e-01
-5.46056032e-01 4.81345415e-01 6.32678807e-01 9.92878258e-01
9.70835745e-01 -1.61299455e+00 -6.44787908e-01 2.81704545e-01
7.19259381e-01 1.91162720e-01 -1.64431587e-01 6.54162109e-01
-4.39841598e-01 4.51279461e-01 -3.27415526e-01 -1.32987809e+00
-1.30951977e+00 6.42427564e-01 6.40591830e-02 7.12423384e-01
-6.49788260e-01 9.26504254e-01 3.01090240e-01 -5.46685457e-01
6.54293835e-01 -4.94452029e-01 1.15681462e-01 -2.14339063e-01
2.89178491e-01 5.13113379e-01 -1.67494267e-01 -7.26689994e-01
-6.57622993e-01 8.53054702e-01 1.10014556e-02 2.31675208e-01
1.18307960e+00 -1.88821882e-01 3.19354266e-01 4.85198647e-01
1.09393954e+00 -4.67432648e-01 -1.39594936e+00 -4.24806029e-01
8.61081555e-02 -8.56577814e-01 -8.70170444e-02 -5.28173387e-01
-6.23555183e-01 6.53491616e-01 7.74631441e-01 2.49060709e-02
8.03618729e-01 8.02958131e-01 5.11688769e-01 7.97801971e-01
7.19483316e-01 -7.33485281e-01 4.65870976e-01 7.59060919e-01
1.00043643e+00 -1.70126903e+00 2.87417322e-01 -4.67954874e-01
-4.37122852e-01 7.74661541e-01 9.03675377e-01 -1.70847028e-01
6.24483287e-01 2.27159649e-01 -3.49101238e-02 -2.74243712e-01
-6.48774028e-01 -5.58127224e-01 4.96630311e-01 8.07363749e-01
9.09230560e-02 -1.07692154e-02 2.48739809e-01 2.81983554e-01
5.08545935e-02 -2.65213400e-01 2.69702941e-01 1.08962297e+00
-5.62070310e-01 -6.50948763e-01 -3.48550439e-01 4.22242463e-01
-8.27104673e-02 4.51576769e-01 -4.80240911e-01 9.96493161e-01
-8.08797702e-02 3.72967452e-01 1.27057821e-01 -2.88946390e-01
6.70795441e-01 -1.70499235e-01 9.38368499e-01 -8.32463324e-01
1.28950253e-01 1.38474897e-01 -3.98159884e-02 -5.38514614e-01
-6.22170329e-01 -8.90736341e-01 -1.20552194e+00 8.42899382e-02
-7.69406974e-01 -2.94179976e-01 1.08172774e+00 4.13463771e-01
6.65085435e-01 8.63660276e-02 5.23469388e-01 -1.52744687e+00
-5.67128003e-01 -8.39312553e-01 -4.12952542e-01 1.03778705e-01
4.37095106e-01 -1.38900530e+00 -8.70793127e-03 7.05244169e-02] | [7.519432544708252, -2.6433167457580566] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.