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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
564547e4-7c95-41d3-bb95-da6e42abdf99 | risk-aware-and-multi-objective-decision | 2102.00966 | null | https://arxiv.org/abs/2102.00966v2 | https://arxiv.org/pdf/2102.00966v2.pdf | Risk Aware and Multi-Objective Decision Making with Distributional Monte Carlo Tree Search | In many risk-aware and multi-objective reinforcement learning settings, the utility of the user is derived from the single execution of a policy. In these settings, making decisions based on the average future returns is not suitable. For example, in a medical setting a patient may only have one opportunity to treat their illness. When making a decision, just the expected return -- known in reinforcement learning as the value -- cannot account for the potential range of adverse or positive outcomes a decision may have. Our key insight is that we should use the distribution over expected future returns differently to represent the critical information that the agent requires at decision time. In this paper, we propose Distributional Monte Carlo Tree Search, an algorithm that learns a posterior distribution over the utility of the different possible returns attainable from individual policy executions, resulting in good policies for both risk-aware and multi-objective settings. Moreover, our algorithm outperforms the state-of-the-art in multi-objective reinforcement learning for the expected utility of the returns. | ['Patrick Mannion', 'Enda Howley', 'Diederik M. Roijers', 'Mathieu Reymond', 'Conor F. Hayes'] | 2021-02-01 | null | null | null | null | ['multi-objective-reinforcement-learning'] | ['methodology'] | [ 7.15171248e-02 3.50658834e-01 -7.57763386e-01 -2.18464494e-01
-7.82559931e-01 -4.02305722e-01 2.43674308e-01 5.53814113e-01
-8.48378718e-01 1.27801633e+00 1.60977811e-01 -6.19536698e-01
-6.15127027e-01 -1.17262518e+00 -4.52577889e-01 -7.32312441e-01
-4.09597248e-01 9.10211861e-01 -1.93005294e-01 1.41072676e-01
8.86405557e-02 2.44504362e-01 -1.27150512e+00 3.30397226e-02
7.46679962e-01 1.16824865e+00 2.97471955e-02 6.81531072e-01
1.67329818e-01 7.68173754e-01 -7.90517449e-01 -2.70844430e-01
1.68313786e-01 -3.77326787e-01 -4.09317881e-01 -1.73560217e-01
-5.66641748e-01 -9.83748436e-01 1.24840826e-01 9.21536088e-01
3.97206783e-01 2.67168224e-01 9.45620358e-01 -1.18061972e+00
-6.37785718e-02 7.37434924e-01 -3.58976573e-01 1.83471367e-02
3.20299149e-01 4.16794628e-01 1.21907568e+00 2.78404027e-01
3.11776429e-01 1.31672120e+00 -4.13405709e-02 6.62466824e-01
-1.14835298e+00 -2.20559821e-01 4.90696311e-01 3.13369110e-02
-6.22287631e-01 8.79771337e-02 3.16154301e-01 -3.78043830e-01
8.80264819e-01 3.95683572e-02 7.49945462e-01 1.00292766e+00
7.22548425e-01 8.68812382e-01 1.26442301e+00 -2.08786637e-01
9.17880774e-01 -5.28790727e-02 -5.17072380e-01 9.45131779e-02
4.34946269e-01 8.05097997e-01 -1.67906865e-01 -6.03777528e-01
6.77400529e-01 4.30456519e-01 2.46106144e-02 -2.34662205e-01
-9.85060215e-01 1.00458527e+00 -4.73701842e-02 -2.50781924e-01
-9.62324321e-01 4.03783351e-01 3.62723976e-01 3.69313538e-01
1.34468317e-01 4.84751195e-01 -5.44175208e-01 -4.03427392e-01
-7.04454243e-01 6.39313936e-01 8.18140328e-01 4.29281592e-01
3.59718353e-01 2.91363779e-03 -8.12793791e-01 1.54573336e-01
2.11434484e-01 4.73598808e-01 1.42555103e-01 -1.05433083e+00
4.53419656e-01 2.39914834e-01 8.82009804e-01 -1.08070076e-01
-3.46367389e-01 -3.95700097e-01 -2.49951020e-01 7.00960338e-01
5.86232364e-01 -8.53553057e-01 -7.70285904e-01 1.78791833e+00
2.52234608e-01 -1.29927173e-01 2.72365659e-01 4.68752503e-01
-3.78439039e-01 4.52462107e-01 3.40097636e-01 -8.29925954e-01
1.07558608e+00 -2.14816242e-01 -6.87301099e-01 -1.90023035e-01
4.72390115e-01 -2.74949282e-01 7.03200340e-01 5.85437775e-01
-1.23848295e+00 2.15365678e-01 -7.71409392e-01 1.05696368e+00
1.10230081e-01 -3.52441549e-01 6.35356605e-01 5.90793371e-01
-4.05404687e-01 9.75121260e-01 -8.31182837e-01 2.45595619e-01
5.20275950e-01 2.57204592e-01 3.98217171e-01 9.18630064e-02
-1.18063164e+00 1.03813815e+00 3.83918434e-01 -4.22165513e-01
-1.56670308e+00 -6.86695576e-01 -2.73824930e-01 4.63368863e-01
1.19577527e+00 -7.57129848e-01 1.71419919e+00 -6.46431506e-01
-1.50661373e+00 2.33075812e-01 3.14167261e-01 -8.73700976e-01
1.02838063e+00 -1.22084960e-01 3.07403076e-02 5.37688360e-02
-5.03760725e-02 9.77082029e-02 9.29545045e-01 -7.82699704e-01
-1.00649023e+00 -3.08291286e-01 5.57114661e-01 5.21483660e-01
1.36459410e-01 -2.31883302e-01 4.74645853e-01 -4.33742613e-01
-7.79376805e-01 -8.81456316e-01 -7.02145934e-01 -3.70586872e-01
-1.82314798e-01 -2.94996142e-01 9.31066126e-02 -2.62530953e-01
1.34897387e+00 -1.79840946e+00 -1.13631122e-01 2.23060906e-01
-2.68939257e-01 -2.12405905e-01 2.66421556e-01 6.25013232e-01
2.84480244e-01 1.17493026e-01 -2.39583269e-01 2.23308101e-01
3.33601534e-01 4.01787519e-01 -3.01845819e-01 3.70543540e-01
1.26961246e-02 5.73068142e-01 -9.69548523e-01 -1.33795276e-01
1.40784040e-01 -2.63729155e-01 -4.98058081e-01 4.08684969e-01
-6.90344036e-01 2.68352985e-01 -9.97220397e-01 3.92473280e-01
3.47289383e-01 -1.90335348e-01 3.71253490e-01 6.11905873e-01
1.97188854e-01 1.03710599e-01 -1.19768667e+00 1.04141796e+00
-6.46442771e-01 -2.93183178e-01 -1.93705335e-01 -1.00446177e+00
2.11887062e-01 4.19463962e-01 7.34930336e-01 -3.89499485e-01
2.05544587e-02 2.20368914e-02 1.60289094e-01 -2.27538958e-01
1.37095541e-01 -7.73878396e-01 -1.80543110e-01 1.03163612e+00
-2.85700828e-01 -3.02725397e-02 1.06009863e-01 -1.46614565e-02
1.20332015e+00 1.14021935e-01 7.73133039e-01 -1.17585689e-01
3.71241570e-02 -2.30713099e-01 8.38173568e-01 9.92354989e-01
-1.08418308e-01 2.38379799e-02 1.30693579e+00 -2.11524498e-02
-7.19335198e-01 -1.36750448e+00 2.39719991e-02 7.83930421e-01
-2.19203487e-01 1.72122806e-01 -4.67671812e-01 -9.73856211e-01
4.43381339e-01 1.22137737e+00 -5.75879753e-01 -1.89200610e-01
-1.32474005e-01 -1.04837644e+00 -1.27370045e-01 4.20292258e-01
-1.39278099e-01 -1.07363045e+00 -1.33979547e+00 5.87576151e-01
1.84751302e-01 -4.44601744e-01 -3.46045583e-01 2.60853052e-01
-8.52206469e-01 -1.12177646e+00 -7.60804713e-01 3.03751498e-01
4.85473692e-01 -4.52588439e-01 1.08050513e+00 -3.15902263e-01
-1.80020481e-01 6.61178231e-01 -1.01250418e-01 -7.93525517e-01
-4.47902441e-01 -4.95953530e-01 5.61176129e-02 -5.84101081e-02
5.49508817e-02 -3.86719853e-01 -9.42059040e-01 -6.81941062e-02
-8.56903374e-01 -4.68774259e-01 4.86467957e-01 8.89556885e-01
6.58105493e-01 4.52754557e-01 1.03000700e+00 -9.20298457e-01
1.09449828e+00 -6.27009690e-01 -8.96141112e-01 6.25766814e-01
-8.95970762e-01 5.14182925e-01 6.58277273e-01 -6.30006969e-01
-1.23436952e+00 -4.42675114e-01 2.11091831e-01 -5.43563925e-02
-1.45819321e-01 5.15570879e-01 -8.52475390e-02 7.60677278e-01
2.85120487e-01 -2.14917585e-02 -6.92617819e-02 -3.30131263e-01
8.97465125e-02 4.97357219e-01 -1.23653568e-01 -1.12137496e+00
2.16412798e-01 2.26535738e-01 4.09316212e-01 -2.84606069e-01
-7.36827135e-01 -7.21711442e-02 2.42008984e-01 -3.07357401e-01
7.15371132e-01 -7.16850340e-01 -1.29800975e+00 1.46920159e-01
-6.81078196e-01 -5.23265123e-01 -8.43082011e-01 7.39342511e-01
-1.28892553e+00 6.94948807e-02 -5.66815808e-02 -1.39809203e+00
-9.02517233e-03 -1.32776082e+00 4.33686644e-01 3.67900491e-01
-1.16400056e-01 -1.15805101e+00 6.61992505e-02 -5.46109118e-02
2.66694754e-01 5.14004230e-01 1.17304134e+00 -7.09466577e-01
-6.16340697e-01 -1.45606592e-01 4.32530999e-01 2.26574674e-01
2.11370111e-01 -2.39498064e-01 -5.54163694e-01 -4.47633773e-01
6.70298859e-02 -5.48296988e-01 5.23617685e-01 9.26140606e-01
1.23061550e+00 -8.66925001e-01 -1.15652204e-01 -1.29521322e-02
1.45381570e+00 6.49561584e-01 3.44257563e-01 4.23065782e-01
-2.14864403e-01 6.77720845e-01 1.11584663e+00 1.28908479e+00
2.40556329e-01 3.77295136e-01 8.73334467e-01 5.47754824e-01
8.45000803e-01 -3.41456443e-01 4.66630280e-01 -5.92638612e-01
-9.44272131e-02 -2.84974486e-01 -5.59650838e-01 5.67584693e-01
-1.96657360e+00 -1.11518276e+00 8.35495234e-01 2.98725057e+00
9.54177618e-01 5.31843066e-01 5.28890431e-01 -1.52149022e-01
3.59813422e-01 5.33736385e-02 -1.30528271e+00 -7.41188288e-01
3.48302037e-01 1.48018494e-01 9.02340174e-01 4.67304945e-01
-6.50443912e-01 2.44747058e-01 6.43650389e+00 9.44247127e-01
-8.28996599e-01 -2.42711619e-01 1.05705810e+00 -6.29173696e-01
-4.30675834e-01 5.50614409e-02 -8.14295053e-01 5.65970123e-01
1.19500029e+00 -7.48040557e-01 4.83779430e-01 9.51297641e-01
4.14874554e-01 -6.29673064e-01 -1.40980995e+00 3.92501116e-01
-7.85689473e-01 -9.19042349e-01 -1.71578273e-01 4.11971539e-01
7.06072688e-01 -2.14464977e-01 2.36900732e-01 1.89237863e-01
1.02768707e+00 -1.12474811e+00 7.49984384e-01 8.34249496e-01
5.72817683e-01 -1.41503477e+00 6.19454741e-01 7.64342368e-01
-5.94503582e-01 -5.94272733e-01 -1.39662072e-01 -1.45576879e-01
2.63802409e-01 7.05473363e-01 -1.00016737e+00 4.33335453e-01
2.75187999e-01 2.27079198e-01 1.82977483e-01 8.50731134e-01
-3.63055617e-01 6.04606032e-01 -2.73302704e-01 -2.83304572e-01
4.60305959e-01 -3.08778703e-01 4.27094609e-01 6.10179484e-01
5.79141974e-01 -3.24385688e-02 3.96081239e-01 6.80925369e-01
2.32422546e-01 -1.32410184e-01 -5.13885736e-01 -3.16556841e-01
3.46221924e-01 8.79253685e-01 -4.63627577e-01 -2.93646634e-01
-2.19791412e-01 3.43076020e-01 -1.86480694e-02 3.98009539e-01
-6.47725165e-01 -1.46159306e-01 8.95059645e-01 -5.23173064e-02
3.79392713e-01 2.45377794e-01 -2.10112780e-01 -8.02325487e-01
1.97203010e-02 -8.45894098e-01 7.90375710e-01 -2.18007356e-01
-1.30894601e+00 -8.95643681e-02 3.88212562e-01 -1.09827983e+00
-9.67006505e-01 -5.91445565e-01 -7.81592369e-01 9.79966342e-01
-1.65453327e+00 -2.53976524e-01 8.41530561e-01 2.05061749e-01
3.59382361e-01 5.68681210e-03 6.89609349e-01 -5.00941575e-01
-2.94506848e-01 2.73283154e-01 4.13723022e-01 -3.03866029e-01
4.90451574e-01 -1.57372606e+00 -1.80986091e-01 3.13454896e-01
-5.59965014e-01 2.72589475e-01 8.62199008e-01 -6.81275725e-01
-1.12920368e+00 -6.97058916e-01 3.23527277e-01 -1.15851976e-01
7.25527287e-01 4.20757920e-01 -5.53226411e-01 5.41507006e-01
3.60426567e-02 -3.95534307e-01 5.26245058e-01 1.73038412e-02
2.97648549e-01 -1.31491914e-01 -1.53804839e+00 6.84010744e-01
5.07834852e-01 -1.22527130e-01 -5.18816471e-01 1.87956989e-01
5.51417708e-01 -9.98152867e-02 -1.00820720e+00 2.48957813e-01
5.42883098e-01 -9.83417809e-01 9.32340741e-01 -9.40862894e-01
4.24748123e-01 1.25440329e-01 -1.97810203e-01 -1.84501481e+00
2.39951257e-02 -8.92417729e-01 -2.44775802e-01 7.23287821e-01
4.80373263e-01 -8.41449440e-01 5.78797936e-01 8.72155964e-01
5.21300972e-01 -1.13659108e+00 -1.40755022e+00 -8.68265152e-01
3.73189032e-01 -2.79634714e-01 1.08093560e+00 3.53214234e-01
5.33291548e-02 -3.22848409e-01 -4.38108563e-01 5.02755344e-02
1.08209109e+00 2.77292401e-01 1.05787978e-01 -9.05220151e-01
-8.46306562e-01 -4.38589454e-01 2.01712817e-01 -6.94756389e-01
1.07051902e-01 -4.21229362e-01 -1.54267788e-01 -1.53966010e+00
4.16978896e-01 -3.65605116e-01 -6.47828937e-01 4.36978817e-01
-3.91300589e-01 -1.00449812e+00 3.21237892e-01 -4.18077856e-01
-4.99433666e-01 5.64644575e-01 1.43363154e+00 -5.62163442e-03
-1.92923501e-01 8.02030444e-01 -7.77993619e-01 7.72384644e-01
9.28755581e-01 -6.49180472e-01 -6.41993225e-01 2.75149733e-01
5.84117293e-01 1.07826626e+00 8.28480348e-02 -5.27778864e-01
-2.58155078e-01 -1.34437525e+00 1.30857423e-01 -5.60672760e-01
5.70575297e-02 -7.52017081e-01 8.11181441e-02 8.80391181e-01
-6.47924304e-01 -2.49468219e-02 -2.06354529e-01 9.81019378e-01
1.76347449e-01 -4.90315646e-01 6.92954361e-01 -7.18156338e-01
-5.07235080e-02 4.51497734e-01 -6.64175928e-01 4.28478569e-01
1.20724666e+00 2.20245510e-01 -9.29886997e-02 -7.49752223e-01
-7.89445221e-01 7.49363720e-01 3.48147124e-01 8.15955028e-02
6.07032537e-01 -1.00080585e+00 -7.40189314e-01 -2.71946162e-01
-5.84428757e-02 -6.26198798e-02 2.91425854e-01 3.81105363e-01
1.61217988e-01 1.58416718e-01 -2.08837867e-01 9.75262932e-03
-7.55706191e-01 5.39922357e-01 5.46864450e-01 -9.82707500e-01
-1.93934783e-01 1.63417444e-01 3.39430898e-01 6.30276129e-02
2.67637700e-01 -2.51174062e-01 -9.01719779e-02 3.49039823e-01
6.72840536e-01 7.33585835e-01 -2.76715189e-01 1.70307443e-01
-1.10259555e-01 9.90394056e-02 -1.42232031e-01 -6.65548861e-01
1.38525522e+00 -1.25819882e-02 3.87715489e-01 5.35163999e-01
4.24443454e-01 -4.07059669e-01 -1.85873091e+00 -1.29432052e-01
-1.15953349e-02 -4.94933784e-01 8.58095214e-02 -1.25054789e+00
-5.80060303e-01 8.48139763e-01 3.76444072e-01 2.75067389e-01
1.05006611e+00 -3.84474665e-01 3.92983854e-01 3.24032545e-01
6.66777551e-01 -1.45338786e+00 -1.29923403e-01 1.36804789e-01
5.22138298e-01 -1.12298381e+00 2.23773077e-01 5.91369450e-01
-1.10816491e+00 8.26984465e-01 2.27707610e-01 -1.80004220e-02
6.22398853e-01 2.57098436e-01 -2.22001776e-01 3.05218130e-01
-1.30102289e+00 -2.07718953e-01 -8.26134905e-02 5.90151906e-01
6.75975764e-03 7.51733243e-01 -3.47659111e-01 4.40948963e-01
1.37512714e-01 1.17574066e-01 8.71015429e-01 1.21273744e+00
-5.33501506e-01 -1.27823615e+00 -3.91035199e-01 9.21251297e-01
-9.49215114e-01 2.00535372e-01 1.26023740e-01 3.58592749e-01
-1.73303202e-01 1.00333381e+00 -4.73161228e-02 4.10260081e-01
3.99549156e-01 1.02387682e-01 5.53810120e-01 -6.97706342e-01
-3.05761576e-01 2.00747982e-01 1.01281889e-01 -5.79497039e-01
-1.20610036e-01 -8.35916340e-01 -1.39822268e+00 -1.91771016e-01
1.22414045e-01 1.70803800e-01 3.92518044e-01 1.01025498e+00
-5.62424771e-02 7.09728241e-01 9.85491574e-01 -3.50953847e-01
-1.73774385e+00 -5.47487497e-01 -9.58491504e-01 1.15739524e-01
6.92857921e-01 -7.72383392e-01 -5.86045146e-01 -7.13248491e-01] | [4.239803791046143, 2.644524335861206] |
3efd8ecc-056e-48b9-ae2a-c2b1290a4cd6 | can-a-simple-approach-identify-complex-nurse | null | null | https://doi.org/10.1145/3341162.3344859 | http://delivery.acm.org/10.1145/3350000/3344859/p736-kadir.pdf | Can a simple approach identify complex nurse care activity? | For the last two decades, more and more complex methods have been developed to identify human activities using various types of sensors, e.g., data from motion capture, accelerometer, and gyroscopes sensors. To date, most of the researches mainly focus on identifying simple human activities, e.g., walking, eating, and running. However, many of our daily life activities are usually more complex than those. To instigate research in complex activity recognition, the "Nurse Care Activity Recognition Challenge" [1] is initiated where six nurse activities are to be identified based on location, air pressure, motion capture, and accelerometer data. Our team, "IITDU", investigates the use of simple methods for this purpose. We first extract features from the sensor data and use one of the simplest classifiers, namely K-Nearest Neighbors (KNN). Experiment using an ensemble of KNN classifiers demonstrates that it is possible to achieve approximately 87% accuracy on 10-fold cross-validation and 66% accuracy on leave-one-subject-out cross-validation. | ['Sadia Sharmin', 'Mohammad Shoyaib', 'Md. Eusha Kadir', 'Pritom Saha Akash', 'Amin Ahsan Ali'] | 2019-09-09 | null | null | null | ubicompiswc-19-proceedings-of-the-2019-acm | ['multimodal-activity-recognition'] | ['computer-vision'] | [ 3.21957558e-01 -2.02821746e-01 -5.77717185e-01 -3.70920420e-01
-2.30042741e-01 -1.28640711e-01 2.03684688e-01 4.33550060e-01
-4.82881963e-01 8.75819445e-01 5.87274790e-01 -3.27282637e-01
-4.28829014e-01 -5.24474025e-01 -9.52161029e-02 -4.84766245e-01
-2.25950330e-02 -1.70487642e-01 1.77773193e-01 2.71949284e-02
3.80476445e-01 5.28164208e-01 -1.50904214e+00 8.97109415e-03
5.25725245e-01 8.57526720e-01 -2.26413891e-01 6.43016160e-01
1.78885087e-01 7.91102111e-01 -6.68083131e-01 3.98876965e-01
-3.07336841e-02 -6.52375400e-01 -7.85009921e-01 -1.41882198e-02
-2.71907985e-01 -2.09792390e-01 1.38809923e-02 3.97307992e-01
6.19017303e-01 1.98900759e-01 4.73493516e-01 -1.08076775e+00
-1.42954081e-01 1.84564009e-01 -7.82895759e-02 4.08957243e-01
1.11834466e+00 -1.92872226e-01 2.67281801e-01 -4.52909976e-01
4.85230908e-02 6.04381919e-01 1.06922662e+00 4.06291217e-01
-8.95241559e-01 -5.40180266e-01 -3.47525120e-01 1.19018763e-01
-1.48056984e+00 -1.84507579e-01 5.39533377e-01 -4.94671613e-01
1.01110733e+00 4.87027973e-01 1.08544612e+00 1.21777594e+00
6.03164732e-01 5.79540312e-01 1.39337313e+00 -6.40142083e-01
5.18086731e-01 1.57789871e-01 2.93187261e-01 5.76571703e-01
6.24337375e-01 -2.05250561e-01 -5.50532103e-01 -4.93364006e-01
8.64888549e-01 6.78122342e-01 -3.24000150e-01 -2.34406605e-01
-1.57859576e+00 4.55335528e-01 -4.77684997e-02 7.05888689e-01
-5.37721455e-01 -3.49453568e-01 1.91516846e-01 -1.60993591e-01
1.58858031e-01 4.55292702e-01 -4.52011168e-01 -9.07644033e-01
-7.36965060e-01 -7.33151063e-02 1.17393827e+00 6.59190297e-01
5.02165914e-01 -3.91203851e-01 2.83439774e-02 6.29267156e-01
2.93272495e-01 4.30038095e-01 1.04182696e+00 -6.32928610e-01
2.63914883e-01 9.86537218e-01 2.83535272e-01 -1.08491087e+00
-8.42734396e-01 -5.89348935e-02 -9.90932405e-01 -2.23182917e-01
2.38371089e-01 -4.04219538e-01 -6.43589377e-01 9.06164765e-01
4.55505520e-01 3.98680151e-01 -2.58951150e-02 6.39136612e-01
5.26447892e-01 2.09195659e-01 2.36514241e-01 -4.74620968e-01
1.44621778e+00 -7.55983412e-01 -9.19114530e-01 -1.59936935e-01
8.90863061e-01 -3.36729348e-01 9.10619497e-01 4.94622529e-01
-6.53746903e-01 -5.71975827e-01 -1.15683949e+00 4.34579611e-01
-5.89945257e-01 2.56967656e-02 5.35479963e-01 1.14440179e+00
-5.40100038e-01 7.81931162e-01 -1.60966873e+00 -8.04606557e-01
1.05324402e-01 3.51360857e-01 -5.33063114e-01 -2.03654263e-02
-8.02670181e-01 8.56872141e-01 1.64579377e-01 -1.16703272e-01
-2.85417467e-01 -3.71046543e-01 -8.12627137e-01 -4.92932320e-01
-7.98825622e-02 -4.76290435e-01 9.21487093e-01 -4.27214444e-01
-1.32162952e+00 5.05462945e-01 -3.21877211e-01 -2.37489492e-01
3.44595730e-01 -6.54885590e-01 -8.26373041e-01 -1.04358129e-01
8.84400085e-02 -2.18594566e-01 1.92844823e-01 -5.41894317e-01
-7.53780842e-01 -5.77191949e-01 -4.39422876e-01 2.56182045e-01
-5.62031507e-01 1.31914735e-01 5.89111000e-02 -5.44248521e-01
3.60522866e-01 -1.18933511e+00 -3.04339021e-01 -5.42344332e-01
-2.33156398e-01 -1.24331310e-01 6.77703619e-01 -7.28947759e-01
1.77674401e+00 -2.10161972e+00 -4.05218452e-01 2.76606768e-01
4.44450378e-02 1.72181383e-01 8.49655032e-01 4.58460510e-01
-8.49021599e-02 -1.89098224e-01 -9.72879902e-02 -2.61488892e-02
-2.83490032e-01 6.16924882e-01 4.60294485e-01 5.13637841e-01
-2.24232018e-01 5.81431746e-01 -1.00020981e+00 -5.30261040e-01
4.87969786e-01 5.44311583e-01 -1.05148301e-01 1.60154983e-01
7.52207696e-01 6.96275353e-01 -5.86562157e-01 7.37441361e-01
-5.14345169e-02 -3.20446491e-01 2.89356381e-01 -6.30169809e-02
-1.29305169e-01 2.54313588e-01 -1.33306265e+00 1.56535912e+00
-1.43607661e-01 4.42509860e-01 -5.23832381e-01 -1.11871886e+00
7.88686991e-01 5.85915148e-01 1.05777681e+00 -3.89504939e-01
1.23809814e-01 2.85195261e-01 -2.26226941e-01 -9.67080057e-01
4.40036133e-02 2.72391468e-01 -2.22227365e-01 2.96070218e-01
-5.39581418e-01 6.43207073e-01 6.53135851e-02 -4.54778105e-01
1.64010572e+00 1.88876271e-01 1.23554730e+00 -2.86424786e-01
5.38495660e-01 5.91697991e-02 6.02561593e-01 6.29489839e-01
-6.32323682e-01 5.82559168e-01 -1.89133614e-01 -7.24002659e-01
-3.52554470e-01 -8.47472787e-01 -6.58168346e-02 1.04900968e+00
-1.77834276e-02 -3.22350800e-01 -8.22209656e-01 -4.28165674e-01
5.48021495e-02 3.02267730e-01 -5.45772731e-01 -1.44469038e-01
-8.50031853e-01 -8.72560680e-01 7.21103311e-01 9.70346689e-01
6.82430327e-01 -1.01215434e+00 -1.29528534e+00 5.26136994e-01
-1.84338093e-01 -9.71331239e-01 -1.41928583e-01 2.03895956e-01
-1.40285909e+00 -1.12269926e+00 -5.80327153e-01 -5.17508388e-01
5.67329109e-01 2.22574607e-01 7.83942223e-01 -1.27329305e-01
-4.61807251e-01 6.79916918e-01 -5.46664953e-01 -6.49957359e-01
1.11255504e-01 2.05723405e-01 5.04781365e-01 -2.00462919e-02
1.07373643e+00 -6.76483691e-01 -9.18123901e-01 5.93391895e-01
-4.96396124e-01 -4.06029105e-01 9.01174843e-01 2.36861691e-01
5.88877976e-01 4.55910191e-02 3.46706361e-01 -5.59091151e-01
8.51293921e-01 -6.49480343e-01 2.36781582e-01 1.92321599e-01
-9.70718265e-01 -1.52261510e-01 3.06455493e-01 -6.09131217e-01
-4.36627448e-01 3.10410589e-01 -9.12901685e-02 9.21305045e-02
-6.99633956e-01 3.44123602e-01 -4.82811891e-02 1.41127650e-02
9.66550767e-01 3.34476680e-01 3.69740576e-02 -5.12026846e-01
-5.21643639e-01 1.07473171e+00 3.86957526e-01 -1.89619943e-01
2.17453107e-01 4.50605363e-01 -1.92970317e-02 -1.03401136e+00
-6.30788803e-01 -1.00339258e+00 -7.17388868e-01 -2.77532846e-01
1.11028457e+00 -7.78740227e-01 -7.62900054e-01 4.66775745e-01
-4.90792871e-01 6.97342679e-03 -7.80808702e-02 1.20722830e+00
-3.14111441e-01 2.69491524e-01 -2.69201726e-01 -1.00735247e+00
-3.36674988e-01 -7.79611588e-01 1.01983905e+00 4.87564564e-01
-9.55826938e-01 -9.37196195e-01 3.05147201e-01 7.77783334e-01
5.43141425e-01 7.00417459e-01 2.72421628e-01 -8.37701499e-01
3.12657803e-02 -5.69388568e-01 3.89280379e-01 2.56495476e-01
9.35637474e-01 -4.15099055e-01 -6.32168770e-01 -2.29177788e-01
2.15992779e-01 3.27993333e-01 -3.09159216e-02 3.25914472e-01
9.86580789e-01 -4.00293827e-01 -8.56324494e-01 2.47282848e-01
1.16027784e+00 6.42767668e-01 8.16441655e-01 4.66821998e-01
7.46300519e-01 1.41205907e-01 7.19219327e-01 5.53555250e-01
4.53406453e-01 4.07111347e-01 -4.81540337e-02 1.94225665e-02
4.68841136e-01 -7.75836259e-02 2.27627680e-01 9.68557537e-01
-7.48606980e-01 4.10593361e-01 -1.18333590e+00 4.99159515e-01
-1.86759520e+00 -7.87220478e-01 -8.05909038e-02 2.24380708e+00
6.37530625e-01 1.40968427e-01 3.49443704e-01 8.62850785e-01
3.96611094e-01 -1.46379009e-01 -3.81205082e-01 -1.31588295e-01
5.16505241e-01 2.10926160e-01 7.41540313e-01 -4.34219316e-02
-1.28009129e+00 4.95380089e-02 6.87913370e+00 2.65840814e-02
-9.28963959e-01 -4.76191975e-02 3.50874096e-01 3.18665802e-02
5.79132497e-01 -2.75173545e-01 -8.10636461e-01 8.12397838e-01
1.21117437e+00 3.30127567e-01 1.82983577e-01 1.21337092e+00
6.40461802e-01 -5.68255067e-01 -9.63440061e-01 1.25491929e+00
1.49167195e-01 -9.34767604e-01 -5.26540518e-01 1.27856642e-01
5.74033976e-01 -1.72957987e-01 -7.60255635e-01 2.02742919e-01
-8.85564908e-02 -1.03353655e+00 -1.06754623e-01 8.44292462e-01
2.82459795e-01 -5.32358050e-01 9.33081806e-01 6.29279733e-01
-1.31561530e+00 -3.29243690e-01 1.85079813e-01 -7.08534360e-01
1.81432441e-01 6.52906775e-01 -8.88778985e-01 4.31064874e-01
1.16780651e+00 6.62024379e-01 -3.61543089e-01 1.03078961e+00
1.88886508e-01 8.21663678e-01 -4.87023652e-01 -4.96447980e-01
-2.64343433e-02 -2.09764689e-01 2.88335923e-02 1.14638174e+00
4.87199843e-01 5.27017951e-01 3.55141670e-01 -1.96422547e-01
4.90545094e-01 1.45693764e-01 -8.10690522e-01 -4.05530538e-03
3.44628245e-01 9.07055378e-01 -7.11236238e-01 -2.25973606e-01
-4.49535787e-01 7.32660830e-01 -2.38488853e-01 -2.27853912e-03
-6.37614250e-01 -6.23576760e-01 7.71887541e-01 4.24979389e-01
-2.31088847e-01 -5.61204374e-01 -3.94280672e-01 -8.08809102e-01
1.97307065e-01 -8.83753717e-01 3.71531367e-01 -4.98491198e-01
-8.93687129e-01 -2.04339996e-02 2.52750635e-01 -1.38407493e+00
-3.96521240e-01 -5.85792065e-01 -4.92862374e-01 5.85504651e-01
-7.85102189e-01 -5.36383212e-01 -9.28143978e-01 8.54297578e-01
2.50192195e-01 1.19044811e-01 1.15583503e+00 5.03600955e-01
-6.06707931e-01 2.75473267e-01 7.68862888e-02 4.78423595e-01
4.53905761e-01 -9.46463764e-01 -5.70533946e-02 4.62432712e-01
-1.95142865e-01 9.21617866e-01 4.02816534e-01 -7.18859076e-01
-1.52603006e+00 -7.59709895e-01 1.05763483e+00 -7.80027926e-01
3.82798672e-01 -5.73358536e-02 -5.33176363e-01 7.83328950e-01
-2.60405451e-01 1.56580150e-01 1.29727912e+00 -1.46678627e-01
3.03367645e-01 -2.94750363e-01 -1.24757516e+00 4.09418851e-01
1.18302667e+00 -2.23236322e-01 -8.30200732e-01 1.94417715e-01
6.48697689e-02 -4.64839697e-01 -1.37647843e+00 7.69747436e-01
1.01375628e+00 -8.80564511e-01 9.60573137e-01 -6.70844555e-01
-1.32820070e-01 -4.75844860e-01 -2.00492188e-01 -9.38338578e-01
-2.14722201e-01 -4.73971575e-01 -2.96797305e-01 8.27793360e-01
8.35588202e-02 -8.43109429e-01 1.02928483e+00 7.28421271e-01
1.24147199e-01 -8.61117303e-01 -7.44496882e-01 -7.53727615e-01
-8.03781211e-01 -2.85955280e-01 5.39885283e-01 1.19532704e+00
5.10159969e-01 3.52496743e-01 -4.52380955e-01 -3.87364514e-02
2.80657768e-01 -3.56465369e-01 8.40486288e-01 -1.47815740e+00
-1.28393695e-01 1.34616777e-01 -8.72778535e-01 -8.99303496e-01
-6.96639895e-01 -1.05186798e-01 -1.31587207e-01 -1.84023380e+00
-1.08957760e-01 -1.45121247e-01 -4.52524036e-01 5.91830075e-01
-8.98065940e-02 8.14736411e-02 -4.93711561e-01 1.83661431e-01
-5.45901895e-01 -1.87546685e-01 8.72653484e-01 2.05304593e-01
-6.13634050e-01 4.83275950e-01 -5.57627797e-01 7.89535522e-01
1.04472375e+00 -4.48289365e-01 -4.70787197e-01 2.46729627e-02
-4.67719622e-02 1.82453781e-01 8.86924192e-02 -1.69038689e+00
2.82259494e-01 -5.24566174e-01 6.99582577e-01 -4.51545477e-01
3.39263350e-01 -1.04346263e+00 4.85009611e-01 8.55675042e-01
4.73242253e-02 3.18728447e-01 -1.46990567e-01 5.87248981e-01
-1.34035096e-01 -1.78530589e-02 2.29245275e-01 -3.91604245e-01
-5.61727345e-01 -2.44761527e-01 -6.83391631e-01 -2.08004087e-01
1.22304928e+00 -8.44568074e-01 -5.78286499e-02 -1.11683480e-01
-9.47373092e-01 -1.53375462e-01 1.39827043e-01 4.37795311e-01
4.28018898e-01 -1.35709679e+00 -2.05545858e-01 2.81103879e-01
3.38335156e-01 -8.16610157e-02 -1.19134061e-01 1.24170017e+00
-7.97845423e-01 4.71520901e-01 -2.61983424e-01 -6.56463206e-01
-1.38450134e+00 2.42620781e-01 2.48555511e-01 -2.42074803e-01
-6.09656394e-01 8.49036351e-02 -7.98878789e-01 -2.14676186e-01
2.62919933e-01 -6.50813401e-01 -4.96127278e-01 -4.94882688e-02
7.96554029e-01 1.04174137e+00 1.11526519e-01 -3.73255908e-01
-8.40054274e-01 6.43488169e-01 3.45607489e-01 2.44558841e-01
1.05830574e+00 -3.99244912e-02 1.89856768e-01 6.97156847e-01
1.13079751e+00 -8.93138722e-02 -7.25439906e-01 1.33670196e-01
4.60811019e-01 -3.73850763e-01 -4.39853638e-01 -5.02415419e-01
-4.65897590e-01 5.24462521e-01 1.08344591e+00 4.97275263e-01
1.24832070e+00 -4.25772846e-01 7.93909729e-01 6.77278161e-01
5.85394621e-01 -1.31096578e+00 5.45736812e-02 2.45461181e-01
1.65788069e-01 -1.12801981e+00 1.51245087e-01 -5.86633086e-02
-4.54707086e-01 7.62701869e-01 2.29021966e-01 -2.03031190e-02
1.08511651e+00 2.20127389e-01 1.39429688e-01 7.81856850e-02
-7.67427832e-02 6.94816336e-02 7.52353668e-02 7.61748135e-01
6.89527392e-01 4.04435962e-01 -6.23520136e-01 4.22760785e-01
-2.22953781e-01 6.64842367e-01 2.05068290e-01 1.63970244e+00
-7.15759516e-01 -8.16366076e-01 -6.56433880e-01 7.80975461e-01
-7.83126295e-01 3.83799434e-01 -1.83490992e-01 7.64567077e-01
1.36514693e-01 1.22658122e+00 7.32938293e-03 -4.46546227e-01
7.51248002e-01 4.26254839e-01 2.29045272e-01 -4.58942443e-01
-6.72077417e-01 -3.65363598e-01 2.63732880e-01 -7.79925168e-01
-8.74155462e-01 -8.16907346e-01 -1.24131775e+00 -2.70180821e-01
1.67897612e-01 1.82538271e-01 8.62112701e-01 1.20160699e+00
4.34486270e-01 5.57487249e-01 3.89269978e-01 -6.63596630e-01
-3.29684973e-01 -1.09924984e+00 -5.89539528e-01 4.39168513e-01
1.46531641e-01 -6.90903783e-01 -2.39611551e-01 1.53818145e-01] | [7.319064617156982, 0.6742533445358276] |
c5b643e9-c6fc-42fc-89a8-82e8f5bf00eb | fusing-video-and-inertial-sensor-data-for | 1802.07021 | null | http://arxiv.org/abs/1802.07021v1 | http://arxiv.org/pdf/1802.07021v1.pdf | Fusing Video and Inertial Sensor Data for Walking Person Identification | An autonomous computer system (such as a robot) typically needs to identify,
locate, and track persons appearing in its sight. However, most solutions have
their limitations regarding efficiency, practicability, or environmental
constraints. In this paper, we propose an effective and practical system which
combines video and inertial sensors for person identification (PID). Persons
who do different activities are easy to identify. To show the robustness and
potential of our system, we propose a walking person identification (WPID)
method to identify persons walking at the same time. By comparing features
derived from both video and inertial sensor data, we can associate sensors in
smartphones with human objects in videos. Results show that the correctly
identified rate of our WPID method can up to 76% in 2 seconds. | ['Yu-Chee Tseng', 'Yuehong Huang'] | 2018-02-20 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [-1.39622107e-01 -3.96219671e-01 9.38704982e-02 -1.29278839e-01
-1.85256004e-01 -2.64047444e-01 2.87490696e-01 -1.22580804e-01
-7.24677980e-01 8.80555093e-01 1.19955927e-01 2.94077486e-01
-4.25437801e-02 -7.40646243e-01 -2.62658536e-01 -4.73074824e-01
9.89534929e-02 2.88046092e-01 4.57889825e-01 1.60327375e-01
-1.15695409e-01 4.97655749e-01 -1.81798196e+00 -3.40371221e-01
7.94162035e-01 8.07906687e-01 2.87387639e-01 8.13464940e-01
3.92806411e-01 4.03725505e-01 -8.67969155e-01 -2.21617103e-01
1.11588806e-01 -2.70180125e-02 -3.21871370e-01 1.36751579e-02
1.40908688e-01 -7.56366432e-01 -5.02987206e-01 8.80340457e-01
9.10274088e-01 5.70616685e-02 5.66518843e-01 -1.64877176e+00
-2.49662012e-01 5.44533655e-02 -4.17017400e-01 -1.59030840e-01
1.37098265e+00 -6.89208731e-02 1.29001319e-01 -5.35615921e-01
1.79008484e-01 1.30277383e+00 1.15325332e+00 6.08336687e-01
-6.67414069e-01 -4.95121866e-01 -1.58433065e-01 3.11477661e-01
-1.81946015e+00 -5.54176748e-01 1.32741302e-01 -3.72189283e-01
7.82182574e-01 1.91339433e-01 7.22075284e-01 1.05136836e+00
-4.69235592e-02 7.65688002e-01 3.61662239e-01 -5.66509008e-01
1.80827826e-01 1.80470273e-01 5.06913900e-01 6.31183624e-01
9.63492334e-01 -8.89168605e-02 -4.42874640e-01 -1.29941821e-01
7.36091197e-01 4.33915377e-01 -3.51402760e-01 -2.35245287e-01
-1.37019205e+00 2.35250264e-01 9.30281449e-03 9.12836418e-02
-6.22970700e-01 -9.34433341e-02 2.18766421e-01 -1.85813099e-01
-2.91004598e-01 -4.39387001e-02 -9.34323296e-03 -5.78012109e-01
-2.56082028e-01 3.42330933e-02 8.41036677e-01 1.25497234e+00
3.96763176e-01 -3.74540955e-01 -1.32675394e-01 6.63546145e-01
4.55332100e-01 1.26460302e+00 6.01267934e-01 -7.62631238e-01
3.39314580e-01 7.07408905e-01 8.09026659e-01 -1.22503507e+00
-5.64833343e-01 4.40654725e-01 -7.29101241e-01 -8.12442303e-02
5.14419138e-01 -6.00307822e-01 -4.18952793e-01 1.34952939e+00
3.40865284e-01 1.63538516e-01 2.12940887e-01 9.28162098e-01
6.79251015e-01 4.01141882e-01 1.87040210e-01 -1.42345607e-01
1.55049229e+00 -6.28482819e-01 -9.19395387e-01 -2.27610409e-01
4.07294422e-01 -3.97749871e-01 7.09808826e-01 2.07138598e-01
-6.60943687e-01 -8.17731619e-01 -9.87746894e-01 3.66650701e-01
-4.39499199e-01 6.18263304e-01 3.74683410e-01 1.18522501e+00
-8.55029881e-01 1.42795414e-01 -7.97743738e-01 -1.26840603e+00
-2.41802394e-01 7.28255033e-01 -4.32165086e-01 5.46329990e-02
-1.12339532e+00 7.75690734e-01 2.54299343e-01 -7.08705783e-02
-1.34361759e-01 1.39921293e-01 -8.57946396e-01 -1.51329398e-01
-7.65997393e-04 -7.70502865e-01 1.07285869e+00 -3.81240398e-01
-1.43807268e+00 4.47576940e-01 -6.24875367e-01 -4.14467216e-01
6.43787801e-01 -5.89192331e-01 -8.14835370e-01 3.65008682e-01
4.73033577e-01 3.67750317e-01 5.70338130e-01 -8.26330066e-01
-1.13151979e+00 -5.33838212e-01 -5.16128689e-02 3.46154451e-01
-5.94420016e-01 9.64095667e-02 -6.60882950e-01 -1.97351538e-02
8.54260325e-02 -9.79829371e-01 1.89629897e-01 2.99058445e-02
-3.59238416e-01 -9.22402442e-02 8.77597034e-01 -7.87336349e-01
1.07898128e+00 -2.26944065e+00 -2.77719587e-01 1.29887149e-01
-1.01919368e-01 4.51870143e-01 4.44897771e-01 1.34493172e-01
5.85586846e-01 -4.42060649e-01 3.35596979e-01 -2.97214866e-01
1.29725650e-01 1.00919157e-01 1.93876237e-01 4.98195469e-01
-4.73308295e-01 4.97025728e-01 -8.77505004e-01 -5.46696782e-01
6.57875478e-01 6.26652181e-01 1.53711706e-01 2.35308126e-01
9.50650096e-01 1.31436214e-01 -4.47988510e-01 8.91270399e-01
6.32016063e-01 1.36384651e-01 1.46581292e-01 -3.00832033e-01
-1.56418622e-01 -1.78140447e-01 -1.67792416e+00 8.57207477e-01
-1.16647042e-01 5.48502743e-01 -4.23208345e-03 -6.15405977e-01
7.21821666e-01 5.59695601e-01 5.98877430e-01 -4.85831708e-01
1.20348074e-01 4.94626798e-02 -6.04120672e-01 -8.33031654e-01
5.91025770e-01 6.26535058e-01 -2.14432970e-01 2.08108246e-01
-2.66437829e-01 7.31364012e-01 1.66818932e-01 -9.90616828e-02
1.13999104e+00 -1.37887280e-02 6.28830433e-01 7.80467466e-02
6.98306859e-01 -1.80918291e-01 3.83718073e-01 9.86340344e-01
-8.14063191e-01 3.63416314e-01 -4.75637436e-01 -4.82041776e-01
-6.38047636e-01 -1.13792562e+00 6.26588240e-02 6.69538200e-01
6.87257051e-01 -2.18416244e-01 -8.47846031e-01 -3.22294772e-01
2.59061813e-01 1.27812147e-01 -3.70316952e-02 -6.45753443e-02
-4.95401382e-01 -5.93636990e-01 8.78655791e-01 7.30202317e-01
1.05729139e+00 -6.69575751e-01 -9.40351903e-01 4.02495354e-01
-5.56202888e-01 -1.46863282e+00 -3.74081731e-01 -4.87849325e-01
-3.26632947e-01 -1.15663493e+00 -9.46368396e-01 -8.29964519e-01
8.23459744e-01 1.02142155e+00 3.06435704e-01 -2.19506640e-02
-1.01108268e-01 9.81232107e-01 -3.14516366e-01 -3.85197341e-01
1.92930356e-01 -1.21566758e-01 1.05952024e+00 1.05840273e-01
9.82738376e-01 4.30341214e-02 -5.52363038e-01 8.27776909e-01
2.26862039e-02 -3.60991389e-01 2.16028348e-01 2.61946827e-01
-6.42095879e-02 3.76278281e-01 2.81912237e-01 4.56542261e-02
6.60479665e-01 -1.95112899e-01 -3.19848835e-01 5.27331948e-01
-3.37567002e-01 -4.09280568e-01 1.09747902e-01 -5.88800848e-01
-8.61832082e-01 5.26397407e-01 6.57858700e-03 8.59465823e-02
-6.39621198e-01 -5.35172857e-02 -4.48759705e-01 -2.12245256e-01
4.21010166e-01 2.48792440e-01 1.87024608e-01 -3.19297016e-01
-8.29888135e-02 1.39621580e+00 8.64989817e-01 -1.91729605e-01
5.55995166e-01 5.69376647e-01 -3.52251619e-01 -1.26395559e+00
-1.38101742e-01 -8.88049901e-01 -6.72132611e-01 -6.82474852e-01
8.12296391e-01 -1.25822592e+00 -1.51118612e+00 9.63811278e-01
-1.11386240e+00 1.65404737e-01 3.08287382e-01 9.19269383e-01
-2.34341085e-01 7.69924998e-01 -5.43424845e-01 -1.42172837e+00
-2.84116149e-01 -7.46073961e-01 1.14410162e+00 6.34235084e-01
-5.77111125e-01 -6.23375952e-01 -2.13929996e-01 2.54550964e-01
2.32286051e-01 1.54295266e-01 -3.42169613e-01 -1.10300645e-01
-1.96111754e-01 -8.40948999e-01 -3.10266227e-01 -3.48111272e-01
5.51807582e-01 -1.22955397e-01 -7.20006287e-01 -4.13402230e-01
-2.50864685e-01 1.87457383e-01 2.53730386e-01 5.92293203e-01
3.39441955e-01 -3.26973557e-01 -1.04432929e+00 3.09678018e-01
1.07646143e+00 5.32645166e-01 7.99593627e-01 6.30231321e-01
7.20225096e-01 5.37549734e-01 7.57959604e-01 5.88194549e-01
6.49577200e-01 9.70873415e-01 -1.30944565e-01 1.52671918e-01
7.27189332e-02 -1.83927700e-01 6.60983860e-01 3.31352592e-01
-5.61070621e-01 -4.13947523e-01 -7.46446788e-01 3.51103425e-01
-2.14124489e+00 -1.22344947e+00 -2.95823187e-01 2.45299387e+00
5.42773232e-02 3.05636507e-02 6.19623423e-01 5.16918659e-01
1.40413833e+00 -6.18375599e-01 -2.47047916e-01 2.46412262e-01
6.69361874e-02 -6.49607897e-01 9.16703045e-01 3.96046519e-01
-1.36333978e+00 4.54244316e-01 6.97362947e+00 1.88623190e-01
-6.25553370e-01 -1.74559608e-01 -9.00041610e-02 2.33285740e-01
6.82586849e-01 -6.54137611e-01 -1.35209358e+00 8.63030195e-01
7.67644405e-01 6.63414318e-03 2.40958646e-01 9.79952097e-01
3.14594269e-01 -5.21465182e-01 -8.44951391e-01 1.47673428e+00
3.45278680e-01 -4.26090568e-01 -3.98736238e-01 -1.60996069e-03
2.38202233e-03 -4.76298660e-01 -2.89071262e-01 1.75732896e-01
5.39469197e-02 -4.77824748e-01 6.07361853e-01 6.67360783e-01
5.96110046e-01 -5.03467262e-01 1.01272786e+00 3.36301953e-01
-1.72377610e+00 -3.07420462e-01 -4.44795877e-01 -5.20720363e-01
6.48196936e-01 1.44202560e-01 -8.37080538e-01 3.47621471e-01
1.02888930e+00 6.81282401e-01 -4.59712267e-01 1.41456974e+00
-8.55723694e-02 1.24587707e-01 -6.31946564e-01 -4.75517660e-01
-3.88048947e-01 -1.87853128e-01 3.22434843e-01 1.10979867e+00
9.76776898e-01 2.59299397e-01 1.79744080e-01 1.03448637e-01
4.97522771e-01 -2.38482863e-01 -9.14057195e-01 4.58963066e-01
8.97691309e-01 8.82729232e-01 -5.83610654e-01 -4.34745163e-01
-4.75812495e-01 1.30730259e+00 -2.22209617e-01 2.87781954e-01
-8.92382324e-01 -6.12471282e-01 6.72465265e-01 5.90687171e-02
-8.68756417e-03 -5.18722355e-01 -1.02313265e-01 -1.07106018e+00
1.67345434e-01 -3.86779547e-01 4.41766858e-01 -8.13304901e-01
-9.66634750e-01 1.85787618e-01 -2.36236639e-02 -1.56342256e+00
-3.33885074e-01 -7.55622447e-01 -2.29984283e-01 5.83998561e-01
-7.13606656e-01 -9.86254275e-01 -9.70980942e-01 8.44960570e-01
1.99728787e-01 -4.06299055e-01 8.11535656e-01 5.78287303e-01
-8.16205740e-01 6.79442167e-01 2.03553975e-01 4.53245580e-01
7.25258291e-01 -8.46817732e-01 3.51586163e-01 9.37679946e-01
-3.80849510e-01 7.66764760e-01 6.91444635e-01 -8.02457333e-01
-1.58341169e+00 -6.91341579e-01 1.05275917e+00 -5.64235389e-01
3.56999040e-02 -3.55349220e-02 -3.79034400e-01 6.18305624e-01
-2.45659128e-01 -4.58390713e-01 5.20220876e-01 -3.21417488e-02
2.49732718e-01 -2.60671318e-01 -1.53054726e+00 7.14946568e-01
1.11135185e+00 -3.51410538e-01 -6.27372026e-01 6.25102967e-02
2.18228146e-01 -2.25572601e-01 -5.96851349e-01 2.32505336e-01
1.19562078e+00 -7.88402498e-01 1.43745053e+00 1.26770616e-01
-7.85193145e-01 -5.21782458e-01 -1.81990564e-01 -8.08657467e-01
-4.49848801e-01 -4.27466452e-01 -2.03746989e-01 1.36428416e+00
-2.03460738e-01 -9.63581920e-01 6.73487127e-01 1.21169376e+00
4.74759251e-01 3.23098719e-01 -7.46128917e-01 -1.14153564e+00
-1.12411571e+00 -1.76052228e-01 8.18634629e-01 5.22884965e-01
6.06962919e-01 2.91246980e-01 -6.53601110e-01 7.80461431e-01
7.34199822e-01 -5.83881676e-01 1.08844697e+00 -1.36668372e+00
2.13210464e-01 -9.54712596e-05 -9.04983342e-01 -1.12905824e+00
-3.43332618e-01 2.32743099e-01 1.12876058e-01 -1.66474056e+00
1.79244801e-01 -2.13428080e-01 -2.37988941e-02 3.61857474e-01
-2.73568720e-01 1.64340779e-01 -6.77432790e-02 3.72096390e-01
-7.59754777e-01 1.38317525e-01 2.12348342e-01 -2.13176921e-01
-2.87529349e-01 2.95558006e-01 -2.63627231e-01 7.67809749e-01
8.62097859e-01 -9.68083441e-02 -1.66480895e-02 -5.03215313e-01
-3.53028625e-01 -7.23316974e-04 5.83648086e-01 -1.78681052e+00
6.44706190e-01 1.86775446e-01 6.81840122e-01 -5.20364106e-01
6.21750653e-01 -9.48312998e-01 3.97071004e-01 9.35721934e-01
4.08214003e-01 1.96000010e-01 -1.62511710e-02 5.02045095e-01
4.83448729e-02 -1.97972715e-01 3.50614637e-01 7.39808381e-03
-1.09450054e+00 -3.66709009e-02 -9.23908055e-01 -7.52786875e-01
1.30309701e+00 -6.89326942e-01 -4.73595738e-01 -6.31105900e-01
-4.38863337e-01 2.73193747e-01 6.58777714e-01 3.83214593e-01
7.34760582e-01 -1.41517591e+00 -1.52412608e-01 3.59682143e-01
1.43232182e-01 -5.97332239e-01 3.12909521e-02 6.08770907e-01
-7.01469004e-01 6.28609121e-01 -4.39549297e-01 -6.16970837e-01
-1.64738297e+00 5.70061147e-01 1.77773267e-01 4.50597048e-01
-5.12980402e-01 3.89502048e-01 -5.25225520e-01 -3.26154321e-01
5.80110610e-01 4.44689728e-02 -6.13645434e-01 -5.46229333e-02
9.99126971e-01 9.78206992e-01 -4.12831664e-01 -9.93764400e-01
-8.78110528e-01 9.86982465e-01 4.40148681e-01 -2.13782474e-01
5.66985786e-01 -7.03018010e-01 2.76945293e-01 1.15843445e-01
6.35678470e-01 -5.07882163e-02 -1.10787654e+00 1.46259954e-02
-4.93450835e-02 -4.88202929e-01 -5.28764129e-01 -2.11607680e-01
-4.16989625e-01 3.83584946e-01 1.23341572e+00 3.99572730e-01
9.05469656e-01 -2.95412987e-01 9.20104802e-01 7.47376263e-01
1.11933911e+00 -1.15050960e+00 -2.41244853e-01 2.50181288e-01
2.20718160e-01 -1.34658742e+00 1.65387709e-02 -5.74858367e-01
-6.27314091e-01 9.13293481e-01 6.11507177e-01 1.66300774e-01
4.35105264e-01 2.20885292e-01 8.23407471e-02 3.01103026e-01
1.58768430e-01 -5.59790611e-01 -1.81640550e-01 1.34886706e+00
-1.32851392e-01 2.81249493e-01 -1.93412229e-01 5.75607777e-01
-8.93455967e-02 1.65629193e-01 3.74028027e-01 1.04916072e+00
-7.61635542e-01 -6.98856771e-01 -9.79482353e-01 1.49524271e-01
-1.31006494e-01 5.62031388e-01 -2.02169701e-01 5.78963697e-01
1.52964786e-01 1.47080898e+00 2.03629788e-02 -7.08460450e-01
5.82455754e-01 -9.01537463e-02 2.62607247e-01 -1.00429812e-02
7.71388710e-02 -3.36709559e-01 2.38742754e-01 -4.43797439e-01
-7.32725501e-01 -9.32697713e-01 -1.10524523e+00 -6.66194975e-01
-9.59642082e-02 3.36265936e-02 4.09109950e-01 9.35070574e-01
3.54819506e-01 7.92076513e-02 3.06917101e-01 -8.91037285e-01
-1.78970233e-01 -8.43374312e-01 -7.31696069e-01 3.63199502e-01
1.05100594e-01 -9.89805877e-01 -1.08853951e-01 2.17313454e-01] | [7.083743572235107, 0.2069411277770996] |
7f9f7160-3d0b-4dff-86ec-29c543330808 | cost-minimization-predictive-energy | 2212.13661 | null | https://arxiv.org/abs/2212.13661v2 | https://arxiv.org/pdf/2212.13661v2.pdf | Cost-minimization predictive energy management of a postal-delivery fuel cell electric vehicle with intelligent battery State-of-Charge Planner | Fuel cell electric vehicles have earned substantial attentions in recent decades due to their high-efficiency and zero-emission features, while the high operating costs remain the major barrier towards their large-scale commercialization. In such context, this paper aims to devise an energy management strategy for an urban postal-delivery fuel cell electric vehicle for operating cost mitigation. First, a data-driven dual-loop spatial-domain battery state-of-charge reference estimator is designed to guide battery energy depletion, which is trained by real-world driving data collected in postal delivery missions. Then, a fuzzy C-means clustering enhanced Markov speed predictor is constructed to project the upcoming velocity. Lastly, combining the state-of-charge reference and the forecasted speed, a model predictive control-based cost-optimization energy management strategy is established to mitigate vehicle operating costs imposed by energy consumption and power-source degradations. Validation results have shown that 1) the proposed strategy could mitigate the operating cost by 4.43% and 7.30% in average versus benchmark strategies, denoting its superiority in term of cost-reduction and 2) the computation burden per step of the proposed strategy is averaged at 0.123ms, less than the sampling time interval 1s, proving its potential of real-time applications. | ['Ruiqing Ma', 'Marie-Cécile Péra', 'Alexandre Ravey', 'Zhen Zhang', 'Xianfeng Xu', 'Fuzeng Li', 'Yang Zhou'] | 2022-12-28 | null | null | null | null | ['energy-management'] | ['time-series'] | [-2.12769285e-02 -1.46804407e-01 -4.42522258e-01 -1.41892835e-01
-6.49699867e-01 -3.22832227e-01 5.90713263e-01 2.03866109e-01
-3.88343155e-01 9.44467545e-01 -3.88638735e-01 -3.90553296e-01
-4.55532581e-01 -8.29382241e-01 -5.42552173e-01 -1.01016557e+00
-1.44896489e-02 1.37712821e-01 5.28745120e-03 -5.52806035e-02
3.95500094e-01 6.02981210e-01 -1.80231893e+00 -5.05040467e-01
1.46088719e+00 1.30855703e+00 6.46429598e-01 2.62328625e-01
2.73882657e-01 3.01231712e-01 -2.19432279e-01 1.68465093e-01
-4.55915369e-02 -2.31114462e-01 -1.22571416e-01 -2.59296149e-01
-6.63789868e-01 -3.23733211e-01 -2.30873406e-01 9.39379990e-01
4.19343382e-01 3.81183237e-01 6.93220258e-01 -1.70252156e+00
3.70731264e-01 -2.07112417e-01 -2.49419004e-01 1.50352374e-01
-4.64764744e-01 3.49966139e-01 3.73111397e-01 -5.89949131e-01
1.23387836e-01 5.60985148e-01 3.20670962e-01 4.44073915e-01
-1.01940143e+00 -7.36989677e-01 -2.05527216e-01 5.01054525e-01
-1.60542321e+00 -4.46587175e-01 8.97022426e-01 -5.00615001e-01
1.21615589e+00 2.70675033e-01 1.07445157e+00 3.39707613e-01
7.12831497e-01 3.12928438e-01 1.13911664e+00 -1.73640758e-01
5.89827955e-01 1.18431278e-01 -2.93596208e-01 3.10969621e-01
5.20523489e-01 2.54535973e-01 -3.79351862e-02 5.11277542e-02
-6.42659292e-02 -3.66553590e-02 3.75167765e-02 -3.94717425e-01
-4.43789333e-01 6.63102925e-01 1.74684420e-01 1.51494339e-01
-4.97418553e-01 1.17539763e-01 3.93489748e-01 -6.02675319e-01
1.82437599e-01 -5.67962117e-02 -2.01574877e-01 -4.65551794e-01
-9.70239937e-01 1.98847622e-01 3.14506322e-01 9.52227712e-01
5.67657411e-01 3.24697584e-01 -3.14004272e-02 5.08856654e-01
4.38654065e-01 1.05883932e+00 2.63785690e-01 -1.01763523e+00
2.57898092e-01 6.48593664e-01 4.90325123e-01 -6.72602534e-01
-5.21460831e-01 -3.55793774e-01 -8.07385027e-01 1.88813075e-01
-1.68215245e-01 -2.12900773e-01 -5.65212667e-01 1.44819510e+00
3.13585162e-01 -1.36783853e-01 2.01128855e-01 7.44645894e-01
-1.56672493e-01 8.68762136e-01 2.98202068e-01 -6.87912107e-01
1.25968432e+00 -5.63189030e-01 -9.45133507e-01 -8.40780735e-02
4.50965285e-01 -8.26263651e-02 3.83886248e-01 1.43200278e-01
-1.13414872e+00 -3.95319194e-01 -1.35792637e+00 4.21876937e-01
-4.77342874e-01 2.19895944e-01 6.55140504e-02 5.46365201e-01
-7.09231794e-01 4.40228939e-01 -9.33702826e-01 -1.18687086e-01
3.69715303e-01 1.67189717e-01 3.55468422e-01 2.11932138e-01
-1.20213020e+00 1.16012692e+00 4.75606024e-01 2.72160321e-01
-1.14624453e+00 -8.68945599e-01 -5.35391450e-01 8.38871226e-02
3.34635288e-01 -2.41544098e-01 1.22515678e+00 -2.79731035e-01
-1.39281082e+00 1.07556738e-01 -4.61121589e-01 -5.89562118e-01
4.97292072e-01 8.57984424e-02 -7.08389699e-01 2.21015930e-01
2.53496831e-03 2.99217731e-01 4.00888771e-01 -1.14727998e+00
-1.19266117e+00 -3.82105678e-01 -5.57806194e-01 3.22373778e-01
-5.44311941e-01 -4.74202782e-01 -1.17899850e-01 -1.99104920e-01
-4.06701416e-01 -1.14003575e+00 -1.00567810e-01 -4.32931930e-01
-5.46522476e-02 -4.20992285e-01 9.72833216e-01 -7.73886263e-01
1.38195479e+00 -1.82751739e+00 6.83169672e-03 2.87868142e-01
-4.74908948e-01 2.89216071e-01 4.54601228e-01 3.81027520e-01
3.35979223e-01 -1.58661291e-01 -4.69564229e-01 -3.18258163e-03
4.61297408e-02 1.81513056e-01 1.64434630e-02 5.97659528e-01
2.68945396e-01 6.40506983e-01 -9.67966676e-01 -1.31024554e-01
6.67907476e-01 6.19237483e-01 -1.54107958e-01 1.05840310e-01
-1.09150141e-01 1.67370766e-01 -5.20299852e-01 4.90514606e-01
7.13998675e-01 3.72094393e-01 3.01086009e-01 -3.47236603e-01
-6.84322417e-01 -9.27892402e-02 -9.33654428e-01 1.31386483e+00
-7.70729303e-01 6.94492221e-01 4.24330622e-01 -1.08895123e+00
9.57348585e-01 -1.30579457e-01 8.01241040e-01 -1.45596647e+00
3.12312514e-01 4.78126407e-01 -3.43069106e-01 -4.42048371e-01
6.92128301e-01 -3.15649122e-01 -4.79139723e-02 -1.96957290e-01
-4.18677211e-01 -2.53878474e-01 3.13603163e-01 -1.79950431e-01
6.97142839e-01 1.26527771e-01 -5.91282267e-03 -8.18228304e-01
9.06863213e-01 2.35815093e-01 7.78162599e-01 -2.92851359e-01
-2.06467748e-01 -5.93276918e-01 1.28042296e-01 3.17790270e-01
-1.07692158e+00 -6.09208643e-01 -3.44934553e-01 5.14680147e-01
7.99426496e-01 1.00596860e-01 -8.36089551e-01 -1.73506901e-01
1.50766119e-01 1.50160861e+00 -4.00603414e-02 -4.85863566e-01
-5.96393526e-01 -6.86983466e-01 3.13659720e-02 6.61703646e-01
4.28390741e-01 -2.35411644e-01 -1.31934702e+00 3.64009947e-01
-1.04130521e-01 -8.38483691e-01 -1.91187173e-01 2.72878081e-01
-8.59756768e-01 -1.02045691e+00 -5.56509376e-01 -5.72552621e-01
7.12090552e-01 2.76101232e-01 6.09117329e-01 -4.96187247e-02
-1.08210132e-01 3.30283701e-01 2.23646343e-01 -6.72204971e-01
-3.26142281e-01 -1.15388542e-01 3.25245798e-01 -4.21038233e-02
3.78394216e-01 -5.77317700e-02 -8.57853711e-01 5.57498395e-01
-4.94194001e-01 1.67573512e-01 4.12893027e-01 6.12901092e-01
8.85878503e-01 8.58306050e-01 1.11548018e+00 -3.79988626e-02
4.15134668e-01 -6.07157230e-01 -1.21405292e+00 2.55841631e-02
-1.37126386e+00 -1.04322322e-01 7.38578022e-01 -3.97797003e-02
-1.29836023e+00 1.67210206e-01 1.46817928e-02 -2.89015979e-01
1.50463179e-01 1.69369847e-01 -5.34606814e-01 2.58301646e-01
-3.44158202e-01 5.03068209e-01 4.21812922e-01 -2.21609637e-01
-5.94391823e-02 7.35756338e-01 7.48085260e-01 -4.21751648e-01
8.60330582e-01 3.17676961e-01 5.12793601e-01 -6.95772290e-01
-3.57772559e-02 -3.24982554e-01 -8.89806077e-02 -7.18285263e-01
8.38325858e-01 -1.19805396e+00 -1.13901579e+00 5.55440784e-01
-6.13057733e-01 -3.03220153e-01 2.59818807e-02 5.05389273e-01
-6.41038179e-01 1.14493966e-02 1.40733212e-01 -1.31351066e+00
-5.40524006e-01 -1.21174324e+00 3.28358233e-01 6.52201056e-01
2.84991235e-01 -7.78798044e-01 -1.95421189e-01 2.32847169e-01
6.32825851e-01 4.43844229e-01 8.09124291e-01 -1.32760257e-02
-4.43188101e-01 -6.95928857e-02 -2.95864698e-02 5.78093231e-01
-1.28391773e-01 -3.62538129e-01 -7.84409225e-01 -6.94301486e-01
1.75100320e-03 2.80750751e-01 2.65300006e-01 3.33448589e-01
1.09079564e+00 -1.59076422e-01 -8.45258355e-01 2.50737101e-01
2.02793002e+00 1.03357446e+00 5.02114654e-01 3.72974515e-01
2.78576493e-01 4.99360919e-01 1.05399132e+00 6.63245499e-01
6.95030451e-01 6.78551376e-01 8.69564533e-01 6.09834604e-02
3.27326387e-01 -3.18326652e-01 2.90865242e-01 6.73431218e-01
9.20580328e-02 -1.79490015e-01 -7.99165905e-01 8.74817133e-01
-1.53084409e+00 -7.71049678e-01 -3.77433240e-01 2.51356745e+00
4.19213474e-01 2.46079072e-01 7.27802888e-03 3.03472519e-01
8.16138208e-01 -2.22337499e-01 -9.02169228e-01 -5.43618619e-01
1.19091451e-01 -3.32480133e-01 1.09552753e+00 2.14601412e-01
-5.52976191e-01 1.67010224e-03 4.57700539e+00 1.14864922e+00
-1.11625576e+00 7.07241222e-02 6.15352273e-01 -2.98548788e-01
-3.23205709e-01 -7.14176744e-02 -8.42214584e-01 1.05122817e+00
1.75286222e+00 -8.43125165e-01 5.94463289e-01 7.30032682e-01
7.61521161e-01 -6.70545101e-01 -7.25788653e-01 7.32435465e-01
-1.11199163e-01 -1.20531297e+00 -5.65225899e-01 4.04347986e-01
5.53575397e-01 -5.85778095e-02 -3.97269636e-01 5.04730225e-01
-4.00165498e-01 -6.62367344e-01 1.01632690e+00 6.67731583e-01
9.57917869e-01 -1.42371321e+00 8.37695599e-01 7.67221093e-01
-1.48040998e+00 -5.02919257e-01 3.80267911e-02 3.41524258e-02
6.46157265e-01 4.38887239e-01 -4.15253550e-01 6.28516436e-01
6.22843087e-01 2.77616322e-01 -1.35267287e-01 7.98867345e-01
2.17862502e-01 4.63531911e-01 -4.72119808e-01 -5.29864609e-01
5.23575991e-02 -5.91021538e-01 4.84188765e-01 1.01774085e+00
6.72514796e-01 1.26643077e-01 -1.48254827e-01 6.88958824e-01
2.77684629e-01 -2.92444140e-01 -3.61356080e-01 -6.18585162e-02
8.19337189e-01 1.46181941e+00 -5.00399351e-01 -2.66610801e-01
-2.32760265e-01 3.11358839e-01 -5.05336940e-01 1.51341170e-01
-1.37410259e+00 -7.89670885e-01 8.12398016e-01 5.53659618e-01
3.66326690e-01 -7.49160275e-02 -3.03125143e-01 -3.27069879e-01
4.55614831e-03 1.53702572e-01 -8.28562528e-02 -6.00367546e-01
-6.34143293e-01 1.60222501e-01 3.90102893e-01 -1.21215761e+00
-2.49148518e-01 -5.41048825e-01 -6.37270391e-01 1.00180376e+00
-1.99923670e+00 -5.31597137e-01 -3.98031324e-01 1.90521613e-01
6.27079129e-01 -1.34558335e-01 4.25064564e-01 6.87143862e-01
-9.67685163e-01 1.99291393e-01 7.33585477e-01 -4.96491879e-01
5.80910407e-02 -7.82422662e-01 -4.16097015e-01 8.59242499e-01
-1.20688570e+00 -1.35267535e-02 8.34929287e-01 -5.41902244e-01
-2.13705516e+00 -1.21044755e+00 6.09266639e-01 2.17212528e-01
6.43698096e-01 -2.00360164e-01 -7.34603763e-01 -1.63140029e-01
3.49508822e-01 -1.25908360e-01 7.28071108e-02 -9.14114475e-01
6.57097757e-01 -5.83216190e-01 -1.41769075e+00 4.12359089e-01
6.00032389e-01 -2.40118280e-01 -9.72956643e-02 -1.35831967e-01
2.44701684e-01 -2.00014472e-01 -1.04970586e+00 7.14239538e-01
5.54479241e-01 -4.63665754e-01 4.55148727e-01 9.14073586e-02
-1.54234007e-01 -5.40739477e-01 -1.11835189e-01 -1.26478982e+00
-1.25142813e-01 -5.88875055e-01 -4.57112372e-01 1.61385357e+00
2.01398715e-01 -7.61806607e-01 5.82174778e-01 5.81874490e-01
-5.81375062e-01 -1.11924720e+00 -1.45838559e+00 -9.00581360e-01
6.12445138e-02 -4.52398002e-01 4.89290357e-01 3.35225165e-01
1.94817632e-02 6.61982223e-02 2.02936992e-01 2.87178427e-01
8.88323367e-01 -2.22143322e-01 4.60392773e-01 -9.65753913e-01
4.88190502e-01 -4.88679320e-01 -9.42533463e-02 -5.56017518e-01
2.99145669e-01 -7.65506268e-01 5.04709005e-01 -1.84254181e+00
1.95142359e-01 -3.76266599e-01 -5.39938450e-01 5.16909957e-02
1.05926916e-02 -1.21958889e-01 4.57384810e-03 -1.59640044e-01
-2.35306948e-01 1.06386459e+00 7.30914533e-01 -3.17085236e-01
-6.91830590e-02 3.57622607e-03 -2.83654392e-01 5.77955723e-01
9.81161237e-01 -4.49225008e-01 -7.34475076e-01 5.85720576e-02
-8.08331892e-02 2.62217879e-01 3.75312626e-01 -1.38369572e+00
1.67133272e-01 -6.00487411e-01 3.12215853e-02 -1.04360819e+00
1.79094926e-01 -1.22103155e+00 6.63433254e-01 9.21030343e-01
3.85247707e-01 -1.70651078e-01 3.87163967e-01 7.63193488e-01
1.10559976e-02 -1.08207852e-01 9.49470639e-01 5.46521664e-01
-9.41879928e-01 6.75844075e-03 -7.25952804e-01 -2.03612298e-01
1.80900776e+00 -4.37167615e-01 -2.21692920e-01 2.85081744e-01
-1.98870134e-02 7.51162827e-01 4.45895255e-01 4.61067826e-01
2.78428108e-01 -1.35235989e+00 -4.00438309e-01 6.90829335e-03
1.01698719e-01 -9.30757523e-02 6.72432482e-01 9.70457613e-01
-2.12318182e-01 7.64729261e-01 -1.66164979e-01 -6.76185668e-01
-8.67248178e-01 5.24722397e-01 4.68757004e-01 4.12488952e-02
-5.39820552e-01 -1.66493729e-02 -3.58378798e-01 2.58073419e-01
-1.35880828e-01 -2.29478762e-01 7.83197209e-02 2.90666461e-01
3.62731814e-01 1.08293271e+00 3.11151236e-01 -7.88465381e-01
-6.44182861e-01 4.35812086e-01 7.08247483e-01 1.03880689e-01
1.27018380e+00 -3.67540121e-01 2.06216708e-01 3.54468346e-01
1.24251473e+00 -3.26422155e-01 -1.70074511e+00 4.02229816e-01
9.66861770e-02 -1.52857199e-01 6.69754028e-01 -9.32188988e-01
-9.41526592e-01 4.17330891e-01 8.54603469e-01 1.88800812e-01
1.62638831e+00 -3.69770110e-01 9.10582125e-01 -4.99666072e-02
5.77498078e-01 -1.77457309e+00 -6.80544496e-01 2.34913275e-01
6.03365123e-01 -7.58290946e-01 1.25554837e-02 2.45630130e-01
-5.48290849e-01 8.41082513e-01 5.26091158e-01 8.45664814e-02
4.89074528e-01 3.14600021e-01 -6.22902632e-01 2.17273131e-01
-7.01434493e-01 1.87818050e-01 -7.49868676e-02 3.27025145e-01
-2.01592326e-01 3.91727775e-01 -6.19621456e-01 7.76297152e-01
4.80141014e-01 2.13340729e-01 2.56442875e-01 1.07163906e+00
-6.07184470e-01 -4.61941510e-01 -2.98440069e-01 4.06991124e-01
-3.63087542e-02 6.63915455e-01 7.05706358e-01 7.10366607e-01
3.40792596e-01 1.16231072e+00 3.30389798e-01 -2.56186694e-01
7.97587097e-01 9.41613317e-02 3.15720551e-02 2.68849820e-01
-1.21540099e-01 -1.12993391e-02 2.02015594e-01 -5.63305140e-01
-4.00397062e-01 -6.18111789e-01 -1.84197104e+00 -2.97873735e-01
-5.27928412e-01 5.21371305e-01 1.48529959e+00 9.79478717e-01
6.71748340e-01 7.19268918e-01 1.03773677e+00 -1.02948737e+00
-5.83049119e-01 -7.74137437e-01 -6.34122550e-01 -3.73788439e-02
2.62992620e-01 -1.03343272e+00 -6.64794803e-01 -2.24333361e-01] | [5.612551212310791, 2.1952364444732666] |
3afa39f3-8c3f-487a-a1d3-690254328829 | assessment-of-cognitive-characteristics-in | 2209.11761 | null | https://arxiv.org/abs/2209.11761v1 | https://arxiv.org/pdf/2209.11761v1.pdf | Assessment of cognitive characteristics in intelligent systems and predictive ability | The article proposes a universal dual-axis intelligent systems assessment scale. The scale considers the properties of intelligent systems within the environmental context, which develops over time. In contrast to the frequent consideration of the 'mind' of artificial intelligent systems on a scale from 'weak' to 'strong', we highlight the modulating influences of anticipatory ability on their 'brute force'. In addition, the complexity, the 'weight' of the cognitive task and the ability to critically assess it beforehand determine the actual set of cognitive tools, the use of which provides the best result in these conditions. In fact, the presence of 'common sense' options is what connects the ability to solve a problem with the correct use of such an ability itself. The degree of 'correctness' and 'adequacy' is determined by the combination of a suitable solution with the temporal characteristics of the event, phenomenon, object or subject under study. | ['Nadezhda G. Bagdasaryan', 'Sergey V. Kovalchuk', 'Oleg V. Kubryak'] | 2022-09-16 | null | null | null | null | ['common-sense-reasoning'] | ['reasoning'] | [ 2.23952264e-01 5.42571861e-03 1.11518716e-02 -6.62318543e-02
1.92408010e-01 -6.28630936e-01 6.78441226e-01 6.53558373e-01
-6.93342388e-01 2.31646046e-01 2.48197734e-01 -5.13782442e-01
-1.02853656e+00 -8.80384207e-01 -9.62221324e-02 -4.40847516e-01
2.55512506e-01 1.70906559e-01 2.50969112e-01 -4.94224191e-01
5.67400157e-01 5.65810800e-01 -1.69341683e+00 -1.11317478e-01
6.87616050e-01 7.82375038e-01 6.31674170e-01 3.62007350e-01
-2.66866200e-02 6.12131834e-01 -4.68632787e-01 2.65708542e-03
-1.73921064e-01 -4.02491301e-01 -6.77013397e-01 2.68242341e-02
-3.32769573e-01 1.16048515e-01 1.90150768e-01 9.60414410e-01
3.99206579e-01 8.25175724e-04 4.81658429e-01 -8.16491067e-01
-6.50203288e-01 1.70120269e-01 4.81546760e-01 6.64922595e-01
7.71934152e-01 4.14249241e-01 3.17365408e-01 -3.62488121e-01
4.57560867e-01 1.16458440e+00 3.16045910e-01 6.61888951e-03
-1.19393444e+00 -3.91528815e-01 3.15461993e-01 1.63783446e-01
-1.07261801e+00 -6.28319740e-01 5.62973917e-01 -8.67489874e-01
8.56894016e-01 -1.86867325e-03 9.67196822e-01 6.07124925e-01
4.16032046e-01 -1.19608454e-01 1.27473950e+00 -6.38707578e-01
6.72869682e-01 3.80992293e-01 3.16983908e-01 -4.65820506e-02
4.29223061e-01 4.29566294e-01 -4.79259253e-01 9.54951122e-02
4.08625513e-01 -1.58563569e-01 -3.12180758e-01 7.89459888e-03
-1.18673909e+00 1.20352358e-01 1.40480831e-01 1.08936894e+00
-7.12109387e-01 -2.49985918e-01 2.91447461e-01 4.51685011e-01
-9.36173499e-02 8.33384573e-01 -3.87198687e-01 -2.84922391e-01
-6.13673389e-01 9.02196094e-02 6.00033283e-01 -9.83054787e-02
3.59411985e-01 1.63173378e-01 -2.31633827e-01 2.20224664e-01
3.33490491e-01 5.22436500e-01 5.35259604e-01 -6.87056780e-01
-8.49019513e-02 9.60575163e-01 2.19203502e-01 -1.09643269e+00
-5.84564388e-01 -7.99908280e-01 -1.64807271e-02 5.88247836e-01
4.90362763e-01 5.42738922e-02 -2.72618175e-01 2.09074783e+00
3.07553142e-01 -2.86006570e-01 9.84358713e-02 7.03585565e-01
7.86688700e-02 3.47327411e-01 6.18197799e-01 -6.09217048e-01
1.27047455e+00 4.49312705e-04 -9.36635673e-01 -4.54080284e-01
3.62840146e-01 -4.78880107e-01 1.01246393e+00 3.89506668e-01
-9.88006234e-01 -6.39631152e-01 -1.18927014e+00 3.36008459e-01
-6.63304687e-01 -4.50463921e-01 5.47698557e-01 6.19800568e-01
-1.14367807e+00 6.93308890e-01 -1.45562574e-01 -2.92820126e-01
-2.78944522e-01 5.82721643e-02 -1.66286781e-01 3.74629378e-01
-1.23598683e+00 1.47966576e+00 6.50646985e-01 3.52853298e-01
-4.91037756e-01 -4.33490396e-01 -3.95976096e-01 2.49799415e-01
3.78960192e-01 -6.26771092e-01 8.67642879e-01 -1.32928216e+00
-1.13594735e+00 6.85128689e-01 1.81204468e-01 -1.02165148e-01
4.15871441e-01 7.51171857e-02 -9.08893645e-01 1.32037848e-01
2.76228845e-01 1.77712634e-01 4.53400761e-01 -8.44814837e-01
-5.32563567e-01 -6.72657311e-01 3.94025773e-01 2.98220545e-01
-2.40031956e-03 1.95333079e-01 2.95600384e-01 -2.05993399e-01
4.00296718e-01 -7.02087760e-01 -1.38225764e-01 -1.29428923e-01
3.82169157e-01 -3.04732859e-01 4.02985454e-01 -2.97802120e-01
1.45649052e+00 -2.31415772e+00 3.25507149e-02 2.66138434e-01
2.98540685e-02 2.15707853e-01 1.72380850e-01 7.20766306e-01
-2.32610211e-01 -2.60176007e-02 1.01566806e-01 8.48745048e-01
3.70697528e-01 2.22973283e-02 8.95742550e-02 3.07932824e-01
-1.28150899e-02 5.11541307e-01 -9.17234659e-01 -7.02882782e-02
2.33790115e-01 2.86469609e-01 -2.14482576e-01 -2.13180184e-02
9.59038455e-03 1.39283687e-01 -8.13732743e-01 3.06496620e-01
3.87793817e-02 -2.74335593e-01 4.95949239e-01 1.73852101e-01
-7.22360551e-01 5.35168350e-01 -1.19269323e+00 9.36071336e-01
-1.61405340e-01 4.23339933e-01 -1.47265727e-02 -7.08350599e-01
8.94589722e-01 6.77577972e-01 7.86154624e-03 -1.39317870e+00
1.37698889e-01 2.28582025e-01 6.57226503e-01 -8.45777988e-01
4.71409969e-02 -1.82834402e-01 1.08872801e-01 5.06460249e-01
-2.93650120e-01 2.42799837e-02 4.38189358e-01 6.22034967e-02
1.01796246e+00 2.71767288e-01 6.56127930e-01 -6.44989133e-01
6.04579508e-01 -4.21665043e-01 6.23283029e-01 4.71980482e-01
-5.07492959e-01 -3.80782098e-01 5.77676892e-01 -2.04206347e-01
-5.73655605e-01 -7.73611188e-01 -3.50073069e-01 1.10536134e+00
2.35425487e-01 3.20705384e-01 -5.56271791e-01 2.76750717e-02
-1.17654540e-01 1.32018137e+00 -5.63203335e-01 -4.31375891e-01
-2.61912886e-02 -6.09386027e-01 -1.46947978e-02 1.69770896e-01
4.63175386e-01 -1.18278229e+00 -1.34931862e+00 4.95754927e-01
9.80368257e-02 -6.32030964e-01 1.84540048e-01 1.14350334e-01
-8.99980187e-01 -9.11811054e-01 3.32057029e-01 -2.78028935e-01
4.77467805e-01 1.56668723e-01 9.19412673e-01 4.13523495e-01
3.53397101e-01 6.61902606e-01 -3.00741374e-01 -4.75228101e-01
-4.02039319e-01 -4.47482347e-01 -3.16874832e-02 -3.45195860e-01
7.45823011e-02 -7.48918414e-01 -4.43414807e-01 -3.19677852e-02
-9.35946226e-01 -6.22528680e-02 5.17667353e-01 1.17427848e-01
2.99912333e-01 4.14306819e-01 8.31268907e-01 -3.54514509e-01
1.02078891e+00 -5.94429135e-01 -1.43693149e-01 3.66911441e-01
-1.10852695e+00 1.10916551e-02 3.29717174e-02 -4.36357081e-01
-1.05968034e+00 -4.46003526e-01 1.46637872e-01 4.15122956e-01
-4.17868614e-01 7.44712532e-01 -4.65071529e-01 9.68986824e-02
5.74238658e-01 3.41340899e-01 8.42313841e-02 -8.74931440e-02
-1.83141306e-01 1.55344695e-01 4.66871023e-01 -7.70313680e-01
4.12727326e-01 -7.69186988e-02 8.89409557e-02 -5.19578934e-01
-4.73691493e-01 7.87402615e-02 -5.25960088e-01 -7.75485635e-01
6.63870811e-01 -4.63061929e-01 -1.02031362e+00 2.94155478e-01
-7.14660645e-01 -1.71252310e-01 -4.11511771e-02 5.09314775e-01
-2.24426419e-01 2.04245836e-01 2.36490611e-02 -1.16106153e+00
-1.62488908e-01 -1.04069316e+00 3.21627051e-01 2.97860593e-01
-4.58793342e-01 -1.08212137e+00 -1.42708290e-02 1.84518639e-02
4.43188488e-01 3.50772291e-01 1.05294490e+00 -7.33080566e-01
-3.40713143e-01 -3.53906840e-01 1.56772524e-01 1.17749237e-01
1.38054237e-01 -1.20260455e-01 -6.98330104e-01 1.61745518e-01
7.52665341e-01 2.07977146e-01 1.03431657e-01 1.94016621e-01
1.35027155e-01 -3.35450739e-01 -1.03217922e-01 -2.03270748e-01
1.60239112e+00 1.00713575e+00 7.65527189e-01 7.71867096e-01
-1.56958699e-01 8.85351419e-01 7.14068770e-01 1.09040052e-01
3.77244413e-01 7.58169830e-01 1.78846076e-01 4.26695406e-01
1.63847670e-01 1.58950448e-01 3.03615332e-01 3.49981368e-01
-1.19625501e-01 -9.28109139e-02 -1.06792164e+00 2.45253757e-01
-1.41184580e+00 -1.03536785e+00 -1.29983440e-01 2.28935146e+00
5.15626490e-01 6.70269728e-01 -2.06038684e-01 5.37609220e-01
6.79176271e-01 1.72182675e-02 -3.54357123e-01 -6.34896934e-01
-2.20492389e-02 -2.15944737e-01 -6.88527524e-02 5.55647314e-01
-2.79142469e-01 4.89958227e-01 6.79426193e+00 3.97005409e-01
-9.86059427e-01 -5.78012019e-02 3.76436681e-01 1.08132690e-01
-4.68811661e-01 1.06517943e-02 -1.64136976e-01 9.27965164e-01
1.10394681e+00 -4.83201921e-01 4.83697414e-01 4.21579301e-01
7.13602602e-01 -7.86756992e-01 -9.53961015e-01 -4.32274789e-02
-4.73687261e-01 -6.06124163e-01 -1.59438953e-01 1.06636612e-02
-1.06549338e-01 -7.38497674e-01 1.64815187e-01 -5.89070655e-02
-3.73057537e-02 -5.68774402e-01 1.35108423e+00 8.44189644e-01
2.76017219e-01 -4.83317345e-01 5.00561774e-01 4.70181555e-01
-6.85145020e-01 -2.47211143e-01 2.34333366e-01 -6.70592010e-01
7.17845634e-02 5.65291226e-01 -3.42852294e-01 -5.50335553e-03
2.83811092e-01 -1.85214281e-01 -6.89681292e-01 5.71014702e-01
-3.79937083e-01 4.64181274e-01 -1.91323027e-01 -1.71515241e-01
-5.72902262e-02 -2.03264415e-01 7.25351274e-01 7.13014007e-01
6.66491389e-02 4.87373859e-01 -2.65610337e-01 8.11915159e-01
1.05785227e+00 -3.73682082e-02 -5.01215100e-01 -2.44726688e-01
8.53931487e-01 8.76143694e-01 -1.08580863e+00 -4.57869530e-01
-1.79746468e-03 2.33714908e-01 8.60317610e-03 4.00878400e-01
-2.16004014e-01 3.86551738e-01 3.68532538e-01 1.08340904e-01
-9.49531049e-02 -2.63944000e-01 -5.81275046e-01 -2.57451713e-01
-4.29823017e-03 -9.09535766e-01 2.07908452e-02 -8.83265972e-01
-5.48953056e-01 2.74669886e-01 -1.02132171e-01 -6.16356254e-01
-2.05385402e-01 -4.60597426e-01 -4.66311246e-01 8.99388075e-01
-9.49578047e-01 -6.99478090e-01 9.10890289e-03 2.09744021e-01
6.85071051e-02 1.40661642e-01 9.22734439e-01 4.76144589e-02
-4.55888093e-01 -4.56021912e-02 -8.62759128e-02 -6.28436387e-01
1.75571188e-01 -9.45733309e-01 -2.88937628e-01 8.42688978e-01
-6.82224274e-01 6.99156463e-01 1.09862995e+00 -8.87147486e-01
-1.22829652e+00 -1.95119202e-01 1.23450220e+00 -4.80850756e-01
6.47374153e-01 2.75010079e-01 -1.07060122e+00 3.28520209e-01
-6.95713460e-02 -9.44451928e-01 7.26099312e-01 1.26992345e-01
-1.63691580e-01 -2.19112128e-01 -1.14600813e+00 8.45890582e-01
1.00977933e+00 -5.59956670e-01 -9.94579554e-01 -1.06942192e-01
6.15587294e-01 6.71558082e-02 -9.31919158e-01 4.19676900e-01
6.86265528e-01 -1.21053958e+00 9.82509136e-01 -2.56246090e-01
7.64619112e-02 -4.29848075e-01 -2.68901773e-02 -8.78339529e-01
-8.84692788e-01 -3.44506390e-02 3.60372335e-01 1.14172935e+00
3.36383522e-01 -1.04756641e+00 -1.07998222e-01 1.09652984e+00
-2.01680511e-01 -4.05792594e-01 -8.67711604e-01 -2.30063856e-01
-2.74913341e-01 -6.77929759e-01 7.80043006e-01 9.06623244e-01
6.20646536e-01 1.34592339e-01 5.54127932e-01 1.03452370e-01
3.56239110e-01 -1.97362378e-01 -2.70531997e-02 -1.50794733e+00
-1.14913099e-01 -7.05675721e-01 -5.78221142e-01 1.16574541e-01
-1.55090079e-01 -5.68249702e-01 -2.48437792e-01 -1.30143929e+00
-1.67883575e-01 -1.82126641e-01 -4.98128772e-01 1.33379608e-01
-3.01199347e-01 -5.35880268e-01 2.89112091e-01 1.80920035e-01
-4.03624415e-01 -1.61289275e-02 9.77869630e-01 3.54792714e-01
-4.64607954e-01 -1.18247427e-01 -1.16856790e+00 8.62093151e-01
8.21257114e-01 -1.03611208e-01 -6.90050423e-01 -1.61334053e-01
8.75127673e-01 6.14882290e-01 4.32265192e-01 -1.17771220e+00
2.71000952e-01 -5.15984774e-01 4.44790781e-01 1.94602031e-02
5.01359366e-02 -1.08227789e+00 8.95390511e-01 9.39421952e-01
-5.06182790e-01 3.60643864e-01 3.12988460e-01 2.10168600e-01
1.38030037e-01 -4.52957511e-01 6.19177103e-01 -1.43661238e-02
-8.48508358e-01 -5.95640600e-01 -8.37955058e-01 -4.17187750e-01
8.61210167e-01 -5.31248212e-01 -1.78451493e-01 -1.21410072e-01
-9.05305088e-01 5.94728515e-02 4.48288858e-01 2.97976881e-01
2.56098241e-01 -7.73882508e-01 -1.62195817e-01 -2.00076029e-01
-2.54168421e-01 -6.29668713e-01 4.20001209e-01 9.11092818e-01
-2.84118086e-01 5.45299232e-01 -4.20818359e-01 2.25195498e-03
-6.91408396e-01 8.58630896e-01 4.36381400e-01 9.77087691e-02
-1.82850525e-01 1.76646009e-01 3.44352305e-01 4.43134248e-01
-6.18114918e-02 5.39518557e-02 -7.82753587e-01 4.85790938e-01
5.86267769e-01 7.61210680e-01 -4.23715375e-02 -5.73225915e-01
-4.01945055e-01 1.22806817e-01 2.32921615e-01 -4.52017456e-01
8.49040687e-01 -4.78992909e-01 -3.00272644e-01 6.38189018e-01
1.27822042e-01 -2.11453333e-01 -9.90723073e-01 8.00578147e-02
1.93788916e-01 -2.81862170e-01 1.25290439e-01 -1.37020946e+00
-2.67041832e-01 2.92580545e-01 8.60126257e-01 5.70297837e-01
1.19007134e+00 -2.86018550e-01 -2.79730022e-01 1.96184069e-01
3.80353540e-01 -1.40296197e+00 -8.21890011e-02 2.14733273e-01
1.08212030e+00 -6.05189621e-01 -7.21755251e-02 -2.43616015e-01
-3.33640337e-01 8.87514532e-01 4.82920766e-01 2.38871768e-01
6.82074368e-01 2.45276988e-01 -3.71074639e-02 -2.87601709e-01
-9.12033319e-01 -4.69486386e-01 -1.58863124e-02 6.17474139e-01
2.40148500e-01 2.95312077e-01 -1.07225454e+00 5.13920486e-01
-3.27382773e-01 3.35156918e-01 3.02700520e-01 7.46307433e-01
-9.49179411e-01 -6.42172217e-01 -7.30248213e-01 3.44042443e-02
-2.10452557e-01 1.18107922e-01 -4.44525599e-01 7.26336658e-01
6.14682317e-01 1.07651627e+00 2.16186270e-01 -1.40269905e-01
4.57561523e-01 3.46747309e-01 2.98929334e-01 -3.87514859e-01
-6.25430763e-01 9.58934501e-02 3.80653828e-01 -3.89769197e-01
-5.75149477e-01 -1.05180776e+00 -1.14980817e+00 -2.68612027e-01
3.60431895e-02 1.23602018e-01 6.74615383e-01 1.53313303e+00
1.57816589e-01 7.35981882e-01 2.43374184e-01 -4.35252845e-01
-3.68615687e-01 -8.66971970e-01 -5.21915078e-01 2.95439780e-01
3.13298665e-02 -7.11902618e-01 -4.13841486e-01 -2.50428736e-01] | [8.963595390319824, 6.362616062164307] |
6a316c68-1f63-4a46-8be3-12114c5c1817 | pcg-based-static-underground-garage-scenario | 2307.03988 | null | https://arxiv.org/abs/2307.03988v1 | https://arxiv.org/pdf/2307.03988v1.pdf | PCG-based Static Underground Garage Scenario Generation | Autonomous driving technology has five levels, from L0 to L5. Currently, only the L2 level (partial automation) can be achieved, and there is a long way to go before reaching the final level of L5 (full automation). The key to crossing these levels lies in training the autonomous driving model. However, relying solely on real-world road data to train the model is far from enough and consumes a great deal of resources. Although there are already examples of training autonomous driving models through simulators that simulate real-world scenarios, these scenarios require complete manual construction. Directly converting 3D scenes from road network formats will lack a large amount of detail and cannot be used as training sets. Underground parking garage static scenario simulation is regarded as a procedural content generation (PCG) problem. This paper will use the Sarsa algorithm to solve procedural content generation on underground garage structures. | ['Kai Li', 'Wenjin Li'] | 2023-07-08 | null | null | null | null | ['autonomous-driving'] | ['computer-vision'] | [ 2.80960321e-01 4.48288262e-01 2.11198255e-01 -5.14687777e-01
-4.56551403e-01 -4.17055577e-01 6.77824616e-01 -1.26023144e-01
-1.09372109e-01 8.31821442e-01 -2.73979753e-01 -1.04570401e+00
-1.53072253e-01 -1.42511284e+00 -7.28403270e-01 -2.61599511e-01
-5.80236688e-02 7.77026415e-01 7.44469464e-01 -7.70631313e-01
1.84116721e-01 6.88453615e-01 -2.03051353e+00 8.92572999e-02
1.17149329e+00 4.58388597e-01 6.45673990e-01 7.53834069e-01
-7.33472288e-01 5.66417038e-01 -2.57034630e-01 -7.64299035e-02
4.44339722e-01 -2.31876373e-01 -9.40639913e-01 1.91743940e-01
4.43087667e-02 -3.07381272e-01 -2.94496775e-01 9.00449812e-01
1.63469151e-01 -7.55323395e-02 5.34039915e-01 -1.40984631e+00
2.48063460e-01 2.86679626e-01 -2.02200130e-01 -4.46604937e-01
2.69649327e-01 2.19570816e-01 4.06607777e-01 -8.10621202e-01
6.44293427e-01 1.03392899e+00 5.35183907e-01 2.49646008e-01
-9.36299860e-01 -4.62684423e-01 -5.00310101e-02 -4.24764380e-02
-1.33966970e+00 -5.90025336e-02 7.23321557e-01 -4.73427296e-01
6.98931038e-01 1.68641448e-01 1.09813511e+00 3.99213344e-01
1.95719346e-01 7.27416158e-01 1.32364023e+00 -3.78169149e-01
2.31685758e-01 3.31156284e-01 4.03108336e-02 6.84944153e-01
3.91206801e-01 3.79732758e-01 -6.58266172e-02 3.66695762e-01
7.06901670e-01 -4.24269438e-01 5.94309717e-02 -7.63157964e-01
-9.13617194e-01 9.30921078e-01 2.34238893e-01 3.53178643e-02
-2.23084003e-01 2.22782299e-01 2.74186879e-01 4.02560413e-01
-1.68298557e-01 2.70750225e-01 -2.75406629e-01 -3.55142504e-01
-1.01016247e+00 8.35475624e-01 6.41646385e-01 1.33452976e+00
1.20031345e+00 3.19634169e-01 6.22755885e-01 6.47066414e-01
4.56877917e-01 6.22560263e-01 -1.45932347e-01 -1.09490216e+00
5.08262336e-01 6.24891937e-01 2.36138612e-01 -8.70849431e-01
-4.70515043e-01 -1.00195318e-01 -3.59279901e-01 9.31162536e-01
6.53183222e-01 -1.71394438e-01 -1.11360753e+00 1.22531486e+00
2.80260146e-01 -1.60157472e-01 7.90861920e-02 6.41221941e-01
6.41725063e-01 9.89951789e-01 5.51901851e-03 2.33962506e-01
1.16315532e+00 -7.86862612e-01 -5.34253597e-01 -5.68263888e-01
9.24587369e-01 -8.49720716e-01 1.27016509e+00 5.80277920e-01
-1.22238207e+00 -6.83335185e-01 -1.41095209e+00 5.59291244e-02
-8.45032930e-01 -2.38696337e-01 8.46564293e-01 9.96644557e-01
-1.03220248e+00 4.78159964e-01 -4.93660361e-01 -3.18459481e-01
1.44132823e-01 2.45717421e-01 -3.35401714e-01 -5.11208832e-01
-1.59107912e+00 1.32516444e+00 5.06365299e-01 3.04930300e-01
-9.51440215e-01 -5.75584650e-01 -1.01363778e+00 -3.44757557e-01
4.58000392e-01 -4.42408949e-01 1.25793922e+00 -1.86824888e-01
-1.44588101e+00 8.34829688e-01 2.91547775e-02 -1.68026641e-01
8.25081229e-01 6.25609905e-02 -4.04799879e-01 -3.12048346e-01
1.62285715e-01 8.83437037e-01 2.82871872e-01 -1.81876636e+00
-9.11299884e-01 1.19344629e-01 3.80118102e-01 2.12479353e-01
4.89725411e-01 -3.09031755e-01 -3.94052356e-01 -1.18257836e-01
2.50524610e-01 -1.13282549e+00 -7.41973519e-01 -9.21972319e-02
-1.96984604e-01 1.75591230e-01 1.07945108e+00 -6.23526454e-01
9.17165995e-01 -1.79104173e+00 -4.18275774e-01 6.15928173e-01
-3.11333597e-01 1.56615034e-01 1.43220201e-01 5.82370698e-01
1.48300320e-01 -7.50199333e-02 -5.32412171e-01 1.85607851e-01
2.08562687e-01 5.47640920e-01 -1.83190957e-01 -7.62178376e-02
-4.81997952e-02 7.78370202e-01 -9.14826274e-01 -6.23314261e-01
6.18571401e-01 2.22490191e-01 -5.14353096e-01 -3.42504352e-01
-2.39037171e-01 2.61579365e-01 -5.75987101e-01 3.68746340e-01
9.64245021e-01 2.97864795e-01 -5.07980073e-03 1.15367867e-01
-4.82219875e-01 3.10343266e-01 -1.59928620e+00 1.66341925e+00
-8.53184342e-01 5.86793959e-01 1.75840989e-01 -9.74213600e-01
1.18902493e+00 -7.83428773e-02 4.16539699e-01 -8.19709241e-01
3.25683579e-02 4.51774508e-01 7.37150013e-02 -5.11775672e-01
1.03705049e+00 -1.43828139e-01 -5.72620869e-01 3.95008713e-01
-3.86150122e-01 -1.13871086e+00 3.92813027e-01 1.55364618e-01
9.05860543e-01 5.02889216e-01 -1.87898412e-01 -4.16414708e-01
4.67596561e-01 7.76699364e-01 2.53387988e-01 5.52600503e-01
-8.94410536e-02 3.65565598e-01 4.24491674e-01 -4.53509897e-01
-1.54859829e+00 -1.11450279e+00 -1.27382085e-01 3.57932925e-01
4.27757978e-01 -3.26874405e-01 -9.22402918e-01 -3.57900590e-01
-1.19258456e-01 1.22888124e+00 -1.96217358e-01 4.71321009e-02
-6.35730207e-01 -4.19074714e-01 2.55965412e-01 2.36836180e-01
7.06495225e-01 -9.42561686e-01 -8.29415441e-01 4.42909867e-01
-1.55867606e-01 -1.24208939e+00 3.19459856e-01 2.40114138e-01
-9.23115075e-01 -8.26634169e-01 -1.89046279e-01 -5.98490596e-01
6.49221420e-01 4.94451076e-01 1.00133836e+00 8.90335888e-02
-4.23061311e-01 8.28951001e-02 -1.63555309e-01 -7.00066507e-01
-8.46615136e-01 -4.60050702e-02 -3.85242432e-01 -6.85924709e-01
1.56018436e-01 -4.89826828e-01 -3.09664845e-01 6.18545175e-01
-7.50746727e-01 5.59492707e-01 7.71847188e-01 3.96661699e-01
4.72868770e-01 7.91380346e-01 4.52418745e-01 -8.12449396e-01
2.19734699e-01 -3.15597832e-01 -6.00206852e-01 -1.51330873e-01
-5.28015196e-01 -4.59160209e-02 4.26521927e-01 1.94529280e-01
-1.21770251e+00 4.15697157e-01 -3.27885598e-01 4.58078906e-02
-5.16008854e-01 3.76263738e-01 -5.40950298e-01 1.16844326e-01
4.97031331e-01 4.71310914e-02 1.27703965e-01 -8.49305540e-02
5.17206252e-01 4.25767690e-01 5.22162437e-01 -7.14273334e-01
1.20412302e+00 5.29899955e-01 2.15421900e-01 -9.74654794e-01
-4.22634751e-01 -7.87415579e-02 -6.91533446e-01 -7.41734028e-01
7.12536871e-01 -6.65776134e-01 -3.58415216e-01 2.87860841e-01
-7.35444367e-01 -6.92522109e-01 -4.28326339e-01 3.11818868e-01
-9.29521441e-01 2.31936544e-01 -6.39420226e-02 -6.02429867e-01
3.60421270e-01 -1.41478598e+00 7.48540401e-01 -2.12193597e-02
-1.70726180e-01 -6.66594028e-01 -3.38828802e-01 4.21503007e-01
2.76902765e-01 3.64977241e-01 9.58143115e-01 1.54096171e-01
-8.51536632e-01 -4.34147954e-01 -1.77209377e-01 3.84967551e-02
7.61429369e-02 1.33175686e-01 -8.30426097e-01 3.23733240e-01
-1.17020816e-01 -4.84449901e-02 3.86129886e-01 3.10043603e-01
8.90854716e-01 3.96792263e-01 -3.99965793e-01 8.20426568e-02
1.51753294e+00 5.64376831e-01 1.41240776e+00 6.18772626e-01
5.53792000e-01 1.18993032e+00 1.26555669e+00 -1.00978158e-01
6.47086024e-01 7.21962094e-01 3.86723459e-01 -1.83205158e-01
-1.64561898e-01 -4.54848498e-01 9.15635899e-02 6.50837064e-01
-4.58622240e-02 2.26372182e-01 -1.33245862e+00 6.30969822e-01
-1.58822024e+00 -1.07874763e+00 -1.03632963e+00 2.04230547e+00
4.44269747e-01 7.87194014e-01 -2.79463449e-04 7.44385719e-01
5.36754310e-01 -2.48763785e-01 -2.09157631e-01 -8.39987040e-01
1.58453479e-01 4.32165675e-02 8.00872564e-01 5.95937550e-01
-9.97671306e-01 1.11399448e+00 6.69399691e+00 8.35512757e-01
-6.94569230e-01 -1.13889828e-01 2.65020788e-01 6.09422028e-01
-5.63241780e-01 4.81674880e-01 -8.13310564e-01 4.75077182e-01
9.93259668e-01 4.89209592e-02 1.54396251e-01 9.84450102e-01
7.07240343e-01 -6.87302649e-01 -5.30946851e-01 6.59090221e-01
-4.97531652e-01 -1.25739968e+00 -1.34799033e-01 2.88255036e-01
8.59244525e-01 -2.03960508e-01 -3.19703549e-01 4.33617026e-01
8.43505025e-01 -9.48050678e-01 9.03983414e-01 5.51621139e-01
6.24954045e-01 -1.00841606e+00 5.15589178e-01 5.03194511e-01
-1.29958200e+00 8.08696300e-02 -4.56701547e-01 9.60373506e-02
7.74027348e-01 6.52984560e-01 -7.54569352e-01 7.98998237e-01
5.96336246e-01 2.22451568e-01 -3.56292814e-01 1.08248937e+00
-1.98711708e-01 3.67132574e-01 -4.43031430e-01 4.37744148e-03
5.68030834e-01 -5.27288139e-01 1.52385190e-01 9.31274831e-01
5.46960890e-01 2.26799063e-02 1.31414279e-01 5.96412599e-01
6.39819324e-01 -3.96161154e-02 -1.18922806e+00 2.97710747e-01
2.04556420e-01 1.11251950e+00 -9.39692736e-01 -3.04371506e-01
-5.54387152e-01 4.67491686e-01 -4.81195271e-01 -3.85633297e-02
-1.03689039e+00 -6.85967147e-01 4.72264171e-01 5.70624530e-01
9.20164436e-02 -6.19193375e-01 -6.30970597e-01 -3.32620353e-01
-3.72405261e-01 -6.74525499e-01 -2.27082565e-01 -1.18013716e+00
-5.43991625e-01 1.89001963e-01 3.71681958e-01 -1.56778562e+00
-2.90975720e-01 -6.02571607e-01 -5.29710710e-01 7.69752979e-01
-1.54673338e+00 -1.23216617e+00 -6.24470472e-01 3.81332904e-01
5.87122142e-01 1.47652134e-01 6.18161559e-01 5.98343670e-01
1.52577743e-01 -4.74788286e-02 -1.44068390e-01 -2.98112094e-01
4.92582470e-02 -1.02856195e+00 5.99067330e-01 6.81038260e-01
-3.74641180e-01 1.10818841e-01 1.21627474e+00 -6.37355804e-01
-1.22843897e+00 -1.05756736e+00 7.50528336e-01 -3.77146065e-01
6.54955029e-01 -3.14994752e-01 -8.89050305e-01 3.47664863e-01
-2.97339708e-01 -6.02424622e-01 1.29419928e-02 -4.31444019e-01
3.51296842e-01 -4.48057204e-02 -1.03082645e+00 8.39852452e-01
1.27689040e+00 -1.95487231e-01 -4.04767334e-01 8.09522718e-02
4.38467294e-01 -3.76319259e-01 -6.94801688e-01 6.05706155e-01
4.64239389e-01 -8.44823539e-01 1.01389730e+00 6.72802776e-02
4.57326174e-01 -7.17804790e-01 -7.24437162e-02 -1.14243662e+00
1.99120954e-01 -4.51675385e-01 5.24586141e-01 8.96797478e-01
6.72424257e-01 -3.83580238e-01 1.17481291e+00 6.06219530e-01
-7.15434134e-01 -4.12004083e-01 -5.27313352e-01 -9.87666488e-01
2.79349059e-01 -1.13786936e+00 9.11220729e-01 7.75257349e-01
-2.46495306e-01 9.38783586e-02 -2.35910729e-01 1.05590895e-01
7.24810243e-01 5.14661260e-02 1.23585105e+00 -1.38871491e+00
1.87526330e-01 -3.44241917e-01 -5.89182496e-01 -9.84145522e-01
-1.47000439e-02 -8.15523267e-01 2.78118104e-01 -2.33155417e+00
-4.11302686e-01 -1.11836672e+00 6.14261031e-01 2.75135905e-01
3.76136988e-01 3.14348228e-02 -1.20500699e-01 -1.67230234e-01
2.31159270e-01 2.57012278e-01 1.52750599e+00 -1.60427421e-01
-1.93270728e-01 1.44811124e-01 -4.65020329e-01 8.38737726e-01
1.09662569e+00 -1.54602826e-01 -9.16703463e-01 -1.63743928e-01
4.02283102e-01 1.30313575e-01 4.47402030e-01 -1.21259975e+00
-1.54010393e-02 -3.29150319e-01 1.68460049e-02 -1.31306040e+00
5.23847520e-01 -1.11207712e+00 6.14780128e-01 6.27250612e-01
2.83063024e-01 -1.57973975e-01 3.65117729e-01 1.31749898e-01
-1.74500138e-01 -5.48853159e-01 8.68867338e-01 -2.86876619e-01
-1.28402984e+00 -1.22730481e-02 -1.04510689e+00 -3.29053760e-01
1.29391277e+00 -1.03423941e+00 -1.90809950e-01 -3.80999058e-01
-9.34197187e-01 4.50981677e-01 6.73255146e-01 3.80862564e-01
5.42530894e-01 -1.19721282e+00 -4.35640961e-01 2.44407147e-01
1.27319500e-01 3.68358791e-01 6.37052476e-01 4.08072680e-01
-1.05401325e+00 4.67664897e-01 -4.17105734e-01 -6.12301767e-01
-8.20360661e-01 4.19881016e-01 3.10124516e-01 -2.39841014e-01
-8.38154554e-01 2.41252959e-01 -2.40249578e-02 -9.77285087e-01
-4.08307910e-01 -1.73494667e-01 -1.93499148e-01 -1.52425751e-01
7.43259192e-02 3.36255282e-01 1.76735476e-01 -7.52448678e-01
5.51363342e-02 7.79763520e-01 4.19113070e-01 -3.53873909e-01
1.20834851e+00 -2.13790298e-01 2.42193848e-01 3.32211226e-01
7.35326886e-01 -3.19162272e-02 -1.14567101e+00 3.24088722e-01
1.21660814e-01 -5.66838503e-01 6.76235855e-02 -4.54737514e-01
-7.45521069e-01 1.08806634e+00 4.98649389e-01 2.27325127e-01
6.69121623e-01 -9.22352076e-02 8.71099651e-01 4.84348267e-01
9.86382902e-01 -1.54215443e+00 -3.39674562e-01 6.19167745e-01
7.36941755e-01 -9.13453996e-01 3.53976190e-02 -8.24751556e-01
-7.40495026e-01 8.66288185e-01 6.37091935e-01 -6.97261468e-02
8.59401643e-01 6.44739330e-01 3.87769565e-02 -4.26328748e-01
-3.88632625e-01 -3.90435070e-01 -3.71816427e-01 8.29478800e-01
-2.89225280e-02 1.65223628e-01 -2.42399305e-01 -8.13165158e-02
-8.76561105e-01 -1.55152613e-02 9.35825169e-01 1.37849438e+00
-9.63848829e-01 -1.40168881e+00 -4.77666527e-01 4.89872456e-01
1.46076642e-02 4.23048317e-01 2.55665570e-01 1.30743611e+00
1.90191209e-01 1.00275970e+00 5.66833019e-02 -3.47386986e-01
7.97430217e-01 4.46662679e-02 3.69385958e-01 -4.75409478e-01
1.08688042e-01 -2.62116313e-01 6.28338397e-01 -5.00821292e-01
-2.90196329e-01 -7.42960155e-01 -1.49120808e+00 -7.47009218e-01
-1.49375260e-01 1.38454020e-01 1.19544303e+00 6.76300347e-01
-1.13856219e-01 4.83558059e-01 6.72069550e-01 -9.69266951e-01
1.07311562e-01 -3.90012532e-01 -5.01207113e-01 -6.31702244e-02
-3.87384474e-01 -9.75571871e-01 -6.99476674e-02 1.41862780e-01] | [8.287989616394043, -2.115488290786743] |
6de3912b-d5a7-4314-9d32-ed055d6ccaef | source-free-domain-adaptation-via | 2204.11257 | null | https://arxiv.org/abs/2204.11257v1 | https://arxiv.org/pdf/2204.11257v1.pdf | Source-Free Domain Adaptation via Distribution Estimation | Domain Adaptation aims to transfer the knowledge learned from a labeled source domain to an unlabeled target domain whose data distributions are different. However, the training data in source domain required by most of the existing methods is usually unavailable in real-world applications due to privacy preserving policies. Recently, Source-Free Domain Adaptation (SFDA) has drawn much attention, which tries to tackle domain adaptation problem without using source data. In this work, we propose a novel framework called SFDA-DE to address SFDA task via source Distribution Estimation. Firstly, we produce robust pseudo-labels for target data with spherical k-means clustering, whose initial class centers are the weight vectors (anchors) learned by the classifier of pretrained model. Furthermore, we propose to estimate the class-conditioned feature distribution of source domain by exploiting target data and corresponding anchors. Finally, we sample surrogate features from the estimated distribution, which are then utilized to align two domains by minimizing a contrastive adaptation loss function. Extensive experiments show that the proposed method achieves state-of-the-art performance on multiple DA benchmarks, and even outperforms traditional DA methods which require plenty of source data. | ['DaCheng Tao', 'Yunhe Wang', 'Chao Xu', 'Yehui Tang', 'Yixing Xu', 'Ning Ding'] | 2022-04-24 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Ding_Source-Free_Domain_Adaptation_via_Distribution_Estimation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Ding_Source-Free_Domain_Adaptation_via_Distribution_Estimation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 1.33032486e-01 -2.09488586e-01 -4.68726784e-01 -6.82476819e-01
-1.02144659e+00 -7.31026888e-01 4.89845127e-01 3.59318554e-02
-3.67417485e-01 1.12348843e+00 2.67002322e-02 1.97838157e-01
1.33450702e-01 -6.36594594e-01 -7.42202580e-01 -1.08053815e+00
5.67054987e-01 6.00757957e-01 6.25877678e-02 1.77050635e-01
-3.69537137e-02 2.02612922e-01 -1.01287138e+00 -3.07181962e-02
1.05003572e+00 1.26166010e+00 -2.49515679e-02 -1.20529339e-01
-2.51769602e-01 3.37750018e-01 -6.68891847e-01 -4.10816789e-01
3.80217344e-01 -4.98810440e-01 -5.22642612e-01 2.95460552e-01
1.77921757e-01 -2.58873105e-01 -4.98929508e-02 1.32865822e+00
4.90643293e-01 2.47764409e-01 1.12270343e+00 -1.56159890e+00
-1.02798080e+00 9.55397934e-02 -7.34241188e-01 -1.01026766e-01
-5.98573275e-02 -1.87792912e-01 6.17287815e-01 -8.75240505e-01
6.57403469e-01 1.09078860e+00 2.74649531e-01 8.74644935e-01
-1.31480813e+00 -1.08845365e+00 1.72967240e-01 7.77202994e-02
-1.49173546e+00 -5.88181734e-01 1.00941801e+00 -4.32266444e-01
3.41672935e-02 -2.24546716e-01 -1.71731245e-02 1.36586845e+00
-3.27588022e-01 8.89590979e-01 1.11190975e+00 -2.06879288e-01
6.25424087e-01 9.05010402e-01 -6.66774660e-02 2.34233513e-01
2.36981079e-01 -1.03355326e-01 -3.50297958e-01 -6.41377628e-01
6.89101994e-01 1.32718712e-01 -3.84731203e-01 -1.28310859e+00
-1.31945908e+00 9.18383479e-01 2.63544828e-01 -7.72389397e-02
-2.62777954e-01 -4.97622699e-01 5.23290038e-01 3.78187269e-01
6.96477234e-01 -1.74555197e-01 -6.70058310e-01 4.72302765e-01
-6.23891354e-01 1.13849752e-01 7.15091288e-01 1.51571441e+00
8.16554546e-01 -2.08899990e-01 -7.60140046e-02 1.03932977e+00
4.91366208e-01 8.27825844e-01 7.06139147e-01 -7.03132570e-01
7.61472940e-01 4.96669799e-01 5.06911874e-01 -8.53813946e-01
1.40418380e-01 -3.34202379e-01 -1.03863001e+00 1.87894311e-02
7.81643450e-01 -3.33837330e-01 -6.08919322e-01 1.84386575e+00
8.17475915e-01 4.23303187e-01 5.28073728e-01 9.72836614e-01
3.89268160e-01 7.91811824e-01 1.22346066e-01 -4.26887780e-01
1.09482563e+00 -8.49266231e-01 -6.15476310e-01 -2.36732855e-01
4.05001938e-01 -4.90586847e-01 1.02130389e+00 3.59981954e-01
-5.28659403e-01 -4.08237547e-01 -1.03091514e+00 9.44312811e-02
-2.35091969e-01 4.32816446e-01 8.66595730e-02 5.91557384e-01
-2.97363639e-01 5.17276041e-02 -5.37423730e-01 -3.12587678e-01
9.87268567e-01 2.68322527e-01 -7.04692781e-01 -3.35995972e-01
-1.18121994e+00 4.60463822e-01 6.18263543e-01 -3.79856020e-01
-9.83208954e-01 -7.64993012e-01 -9.23345327e-01 -4.62006703e-02
4.52327073e-01 -5.76471686e-01 1.12234128e+00 -1.09705687e+00
-1.56555307e+00 1.07413936e+00 -2.82118499e-01 -3.37030530e-01
5.97933769e-01 -8.45877975e-02 -6.62969470e-01 -7.88471997e-02
3.46917629e-01 3.41855824e-01 1.24854743e+00 -1.40773857e+00
-8.86928082e-01 -5.99731863e-01 -5.82012534e-01 2.09083065e-01
-7.67826140e-01 -2.17612043e-01 -3.08170676e-01 -6.32041216e-01
-2.70878255e-01 -6.22201681e-01 -2.62032207e-02 2.79506058e-01
-2.82190055e-01 -2.09703729e-01 1.04270780e+00 -5.37667572e-01
1.04019141e+00 -2.47091961e+00 1.14804402e-01 3.80276859e-01
1.67090386e-01 4.37065363e-01 -1.84719378e-04 -4.00264449e-02
-3.67357619e-02 -3.68582994e-01 -6.63295805e-01 -2.96334386e-01
9.54040140e-02 1.00644305e-01 -6.96008742e-01 7.50461161e-01
-3.02798264e-02 2.85854816e-01 -9.16906118e-01 -5.72047055e-01
5.63057177e-02 3.41666192e-01 -3.05993944e-01 5.12513161e-01
-2.34968886e-01 7.84910142e-01 -8.07630956e-01 5.73381841e-01
1.16618192e+00 -3.06673497e-01 1.33906230e-02 2.18914822e-02
5.20720601e-01 -2.37937987e-01 -1.25338197e+00 1.91261673e+00
-2.68647343e-01 2.59899497e-02 1.90370768e-01 -1.27085173e+00
1.37937748e+00 2.27615535e-01 4.40808684e-01 -2.24031925e-01
2.21456319e-01 3.93715441e-01 -4.82848018e-01 -1.38695478e-01
-8.80294666e-02 -2.50227720e-01 -2.28750795e-01 3.58752400e-01
1.74724743e-01 1.09196335e-01 -3.45418394e-01 -1.14208050e-01
7.35258222e-01 8.00277889e-02 5.96417964e-01 -1.33816823e-01
1.06296432e+00 8.64413306e-02 9.13423836e-01 1.97487146e-01
-4.74488109e-01 6.57096207e-01 2.68534869e-01 -1.76579535e-01
-9.74127233e-01 -1.25976896e+00 -1.57591105e-01 1.00588882e+00
2.71690160e-01 2.59275287e-01 -9.12063777e-01 -1.37703204e+00
1.76485375e-01 8.30401778e-01 -5.00811338e-01 -3.37014973e-01
-2.76339024e-01 -5.75664580e-01 4.16476458e-01 2.90537089e-01
7.30312943e-01 -6.88866735e-01 5.93669564e-02 2.63152301e-01
-1.10840969e-01 -1.01832139e+00 -8.39816928e-01 -3.65551412e-02
-7.92399764e-01 -9.86227036e-01 -1.23980296e+00 -1.00635374e+00
9.47095692e-01 1.94930092e-01 6.89316213e-01 -7.79010117e-01
2.98907757e-01 2.50827283e-01 -1.64313048e-01 -4.58926201e-01
-2.48511076e-01 2.60713100e-01 3.02468419e-01 6.05778992e-01
9.84775364e-01 -6.18930280e-01 -4.58074868e-01 5.58723748e-01
-9.55653965e-01 -4.84432667e-01 6.53149188e-01 9.12126303e-01
9.73906338e-01 -9.70667321e-03 1.08662498e+00 -1.36338532e+00
6.31413281e-01 -1.06496143e+00 -7.81927049e-01 3.37212622e-01
-7.58638084e-01 -1.79022439e-02 9.89579856e-01 -6.83227837e-01
-1.39145279e+00 3.82447630e-01 3.97369623e-01 -8.51581097e-01
-5.79397500e-01 1.01875156e-01 -9.32236969e-01 2.77173340e-01
8.13776612e-01 4.77883041e-01 2.23336518e-01 -6.28992558e-01
3.99566263e-01 1.18697882e+00 7.57098317e-01 -8.39515746e-01
9.86189604e-01 5.91913879e-01 -3.55624676e-01 -2.85840034e-01
-9.92563248e-01 -6.14068866e-01 -7.15697348e-01 5.56442261e-01
5.69406748e-01 -1.21917331e+00 -1.00646280e-01 6.05154514e-01
-9.37338889e-01 -1.07008517e-01 -3.59072119e-01 5.58776855e-01
-6.33341312e-01 3.03793281e-01 1.08277515e-01 -5.23602605e-01
-2.57971734e-01 -7.94091642e-01 9.38128591e-01 3.48448575e-01
2.76670128e-01 -1.11794543e+00 1.21130712e-01 2.12190926e-01
1.78099275e-01 2.43777260e-01 7.21473157e-01 -1.31446290e+00
-1.90566033e-01 -1.71270221e-01 -3.08027357e-01 7.24961162e-01
6.71665072e-01 -6.57457948e-01 -9.76549566e-01 -4.62802798e-01
2.13087216e-01 -3.46063405e-01 4.15904313e-01 2.85282671e-01
1.34169614e+00 -4.54295486e-01 -5.99328399e-01 6.25557840e-01
1.22146463e+00 1.74216479e-01 2.98930317e-01 2.56295532e-01
6.12881899e-01 5.53190768e-01 1.14332438e+00 6.77933633e-01
4.19609010e-01 4.80889857e-01 1.46591544e-01 1.62521169e-01
2.14490443e-01 -5.56548655e-01 3.44959140e-01 3.96792710e-01
7.08217800e-01 -1.40639693e-01 -7.10595667e-01 8.54459584e-01
-1.97170556e+00 -6.00826085e-01 2.74920136e-01 2.56306434e+00
1.13567746e+00 -1.64584950e-01 3.63050431e-01 -3.85778517e-01
1.16120470e+00 -4.33798321e-02 -1.23977029e+00 1.50785238e-01
5.73334992e-02 -1.20394960e-01 6.39648914e-01 1.00817539e-01
-1.29885054e+00 8.37698996e-01 4.97454643e+00 1.20216596e+00
-9.04348254e-01 2.45257810e-01 5.99244535e-01 -4.31915512e-03
-2.45636642e-01 -1.30424574e-01 -8.62494349e-01 8.23330283e-01
6.77505255e-01 -5.88535011e-01 3.38309556e-01 1.42765331e+00
-8.96523520e-02 3.39343339e-01 -1.30184650e+00 1.14909673e+00
3.80121432e-02 -9.14114296e-01 -8.05403367e-02 1.64720327e-01
8.02920759e-01 -2.42752016e-01 2.06644267e-01 4.07387942e-01
5.30746639e-01 -6.52765512e-01 2.42057994e-01 2.40840226e-01
1.28225732e+00 -9.31078374e-01 6.44917309e-01 6.45512760e-01
-9.15874720e-01 -6.72729313e-02 -8.39335382e-01 3.93848568e-01
-1.90400019e-01 7.43841946e-01 -8.49989653e-01 4.85971421e-01
5.31135499e-01 9.19538736e-01 -1.22330479e-01 9.45176482e-01
-5.41784093e-02 7.10573792e-01 -3.09496373e-01 2.52728403e-01
-1.41824022e-01 -3.61829221e-01 4.21216875e-01 7.74384141e-01
4.66351986e-01 1.66714173e-02 1.93535864e-01 7.49557197e-01
-4.67803746e-01 3.90614748e-01 -7.19721735e-01 8.00874382e-02
1.02712405e+00 1.03819919e+00 -1.13041379e-01 -2.89577067e-01
-4.22815055e-01 1.28095579e+00 3.86472732e-01 4.81153071e-01
-7.46941924e-01 -4.51579094e-01 8.92890334e-01 -2.56121587e-02
3.60935956e-01 1.50124699e-01 -8.16532448e-02 -1.45441389e+00
1.12108365e-01 -8.35695326e-01 8.31467986e-01 -4.26449448e-01
-2.03903413e+00 3.27207506e-01 2.88744420e-02 -1.67326128e+00
-1.98216021e-01 -3.60591233e-01 -4.72547472e-01 1.10988331e+00
-1.71778727e+00 -1.14569724e+00 -1.81066468e-01 1.22436631e+00
3.17110151e-01 -7.03779817e-01 9.13295925e-01 4.10334527e-01
-4.85972077e-01 1.04964447e+00 8.84163737e-01 3.07138354e-01
1.49798226e+00 -1.11424196e+00 6.47647530e-02 5.34213841e-01
-2.28816539e-01 4.98057604e-01 3.43784750e-01 -4.56227303e-01
-9.17307556e-01 -1.57038546e+00 4.67806488e-01 -3.38814765e-01
4.44488764e-01 -3.33196342e-01 -1.11585414e+00 9.05543327e-01
-3.84914093e-02 5.97531199e-01 9.44274604e-01 -2.71178335e-01
-6.43655241e-01 -5.87795913e-01 -1.96338201e+00 9.34103578e-02
7.10384429e-01 -2.51475394e-01 -7.08547592e-01 3.09175551e-01
6.76375687e-01 -4.42460150e-01 -9.31806147e-01 -2.61367727e-02
1.56137347e-01 -3.92330736e-01 1.01592731e+00 -7.37409532e-01
2.13175058e-01 -5.75036883e-01 -2.05335125e-01 -1.70881534e+00
-2.03423291e-01 -3.52281690e-01 -1.46584347e-01 1.85740674e+00
1.03522129e-01 -1.00042295e+00 9.87398982e-01 7.07412481e-01
3.09079409e-01 -1.47285387e-01 -1.05083942e+00 -9.17074502e-01
3.91612917e-01 1.61107868e-01 9.84730899e-01 1.32845342e+00
-2.65544683e-01 3.40312481e-01 -4.11410719e-01 4.86201584e-01
1.01263034e+00 1.88059747e-01 9.81840611e-01 -1.33453536e+00
-3.98940630e-02 7.77032971e-02 -1.77616388e-01 -1.10448492e+00
5.91160059e-01 -9.81496453e-01 -2.78624035e-02 -1.04911673e+00
1.89269651e-02 -8.28939080e-01 -6.18096352e-01 4.84512299e-01
-1.70540869e-01 -2.04328045e-01 -1.39595225e-01 4.66535330e-01
-6.18361592e-01 8.77380550e-01 1.02439213e+00 -2.14902595e-01
-2.18990490e-01 2.03268349e-01 -1.05020261e+00 5.84612310e-01
8.62641454e-01 -7.89516628e-01 -7.92874336e-01 -3.65953147e-01
-6.44903839e-01 -1.09037116e-01 2.22361237e-01 -8.58376026e-01
1.67811304e-01 -5.99339545e-01 5.03000379e-01 -2.91212887e-01
1.22551024e-01 -1.39930224e+00 -5.80589436e-02 -7.53724203e-02
-4.13211852e-01 -7.25784183e-01 -8.93258080e-02 1.05867064e+00
-3.49970311e-01 -1.65019467e-01 1.08888960e+00 1.22018375e-01
-6.97404802e-01 5.91319501e-01 2.65564442e-01 3.50980431e-01
1.42094564e+00 9.57600623e-02 -1.83156237e-01 -1.56189382e-01
-5.62945366e-01 4.14529026e-01 6.03144586e-01 2.74295062e-01
4.35500383e-01 -1.78304970e+00 -7.80363142e-01 2.99075425e-01
5.41703343e-01 5.56672812e-01 1.56564236e-01 2.50932664e-01
1.36853889e-01 5.87508306e-02 -2.65155375e-01 -5.37407219e-01
-9.14580405e-01 9.65537071e-01 9.67730358e-02 -1.09210968e-01
-2.95286030e-01 8.48338962e-01 5.24591625e-01 -9.24599648e-01
1.70510590e-01 1.17146686e-01 -1.40681237e-01 5.95113449e-02
5.86703002e-01 3.29745889e-01 -2.08581999e-01 -6.13283455e-01
-4.71342683e-01 3.67134422e-01 -3.29014391e-01 5.91056384e-02
1.24879849e+00 -4.94462639e-01 2.14268029e-01 1.09859727e-01
1.53295624e+00 1.00629956e-01 -1.57667482e+00 -1.00845647e+00
-9.71617401e-02 -7.64567733e-01 -3.60734850e-01 -8.40325654e-01
-1.10253727e+00 9.27952588e-01 7.55352259e-01 -3.45390290e-01
1.30793357e+00 -6.61532804e-02 1.04376400e+00 2.72672892e-01
4.10965890e-01 -1.19811332e+00 -2.34429166e-01 1.34088099e-02
5.73295414e-01 -1.49531198e+00 -2.05836490e-01 -3.02715868e-01
-1.07549417e+00 8.25330973e-01 8.27058315e-01 -2.47798100e-01
6.44907892e-01 -1.57565132e-01 1.59491584e-01 3.38995159e-01
-3.68131012e-01 1.70549318e-01 3.68935466e-02 1.25209355e+00
-2.57535458e-01 8.58988017e-02 -2.03050207e-02 1.23799276e+00
2.53536075e-01 2.55832404e-01 2.08255380e-01 6.25615299e-01
-2.89481521e-01 -1.46468008e+00 -6.40101135e-01 4.00561154e-01
-2.91258067e-01 2.45584249e-01 -2.93256938e-01 4.88953680e-01
1.32682025e-01 7.46781886e-01 -2.03562379e-01 -1.14168070e-01
3.16319644e-01 2.48798043e-01 -3.79397571e-02 -6.69465303e-01
4.64012362e-02 -6.31928071e-02 -4.65265781e-01 -1.19438685e-01
-2.23086193e-01 -8.94983232e-01 -1.23919547e+00 -1.21934004e-01
-8.92716050e-02 4.45121348e-01 4.63158309e-01 7.14115500e-01
7.06727743e-01 -9.58489627e-02 9.56915259e-01 -2.85025060e-01
-1.03607810e+00 -7.73332894e-01 -9.27340508e-01 5.79208851e-01
3.88347983e-01 -7.72542715e-01 -3.05937260e-01 4.14022058e-01] | [10.490482330322266, 3.1490025520324707] |
ac6637aa-fbfc-4553-931f-d94829dc392d | design-pseudo-ground-truth-with-motion-cue | 1812.05206 | null | http://arxiv.org/abs/1812.05206v1 | http://arxiv.org/pdf/1812.05206v1.pdf | Design Pseudo Ground Truth with Motion Cue for Unsupervised Video Object Segmentation | One major technique debt in video object segmentation is to label the object
masks for training instances. As a result, we propose to prepare inexpensive,
yet high quality pseudo ground truth corrected with motion cue for video object
segmentation training. Our method conducts semantic segmentation using instance
segmentation networks and, then, selects the segmented object of interest as
the pseudo ground truth based on the motion information. Afterwards, the pseudo
ground truth is exploited to finetune the pretrained objectness network to
facilitate object segmentation in the remaining frames of the video. We show
that the pseudo ground truth could effectively improve the segmentation
performance. This straightforward unsupervised video object segmentation method
is more efficient than existing methods. Experimental results on DAVIS and FBMS
show that the proposed method outperforms state-of-the-art unsupervised
segmentation methods on various benchmark datasets. And the category-agnostic
pseudo ground truth has great potential to extend to multiple arbitrary object
tracking. | ['C. -C. Jay Kuo', 'Ming-Sui Lee', 'Yueru Chen', 'Ye Wang', 'Siyang Li', 'Qin Huang', 'Jongmoo Choi'] | 2018-12-13 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 5.01335025e-01 6.19855477e-03 -7.74477363e-01 -3.87277156e-01
-6.65445030e-01 -6.62724316e-01 1.37651205e-01 -4.58111793e-01
-4.49781746e-01 6.08584821e-01 -4.89154875e-01 -4.00898047e-02
1.70031309e-01 -7.28490055e-01 -9.43880498e-01 -9.14835930e-01
2.04071194e-01 6.76773608e-01 9.44866240e-01 3.32964987e-01
3.95424217e-01 3.94665241e-01 -1.31154132e+00 1.63994774e-01
9.96936500e-01 1.08990610e+00 6.33099198e-01 4.66332287e-01
-4.31456029e-01 8.47772717e-01 -6.22435927e-01 -1.93693265e-02
4.98134881e-01 -3.22911859e-01 -1.30829406e+00 8.59915912e-01
5.33735394e-01 -5.66712379e-01 -2.54426450e-01 1.37541890e+00
-1.33162111e-01 5.15264571e-01 4.65686411e-01 -1.28759062e+00
-3.71852428e-01 7.45817602e-01 -6.89881682e-01 4.39653397e-01
-1.21852934e-01 3.11747044e-01 6.79580450e-01 -8.17306340e-01
7.34170556e-01 1.01783216e+00 3.00285935e-01 8.06841910e-01
-9.39511836e-01 -6.36776388e-01 5.25386035e-01 8.98035616e-02
-1.34766924e+00 -3.01744848e-01 8.42753232e-01 -5.45049965e-01
3.21498543e-01 1.38839141e-01 8.43573809e-01 7.95031607e-01
-3.28296840e-01 1.32657027e+00 7.07886398e-01 -1.73385680e-01
2.85799384e-01 5.36245070e-02 5.31960964e-01 8.46519947e-01
4.40044016e-01 -1.75314486e-01 -8.01009089e-02 2.76547790e-01
9.82650697e-01 1.63499147e-01 -4.07329470e-01 -4.38934863e-01
-1.23299885e+00 4.92395192e-01 3.80262911e-01 3.51030350e-01
-3.22805971e-01 4.17089164e-01 1.92968145e-01 -4.33293790e-01
3.59611303e-01 1.75156593e-01 -6.00869536e-01 1.60250887e-01
-1.46077204e+00 2.41263341e-02 5.27570307e-01 1.19597280e+00
9.16772246e-01 2.51978606e-01 -4.55182016e-01 5.70606589e-01
4.04340118e-01 4.23491597e-01 3.25589627e-01 -1.44648862e+00
3.51433396e-01 6.92273498e-01 2.65647262e-01 -9.19226825e-01
-1.06170572e-01 -2.17187867e-01 -4.45248961e-01 -1.88374832e-01
6.95681989e-01 -5.62637188e-02 -1.65544558e+00 1.33581090e+00
5.54611385e-01 7.11923420e-01 -1.66099649e-02 1.33244097e+00
9.92466807e-01 7.49345958e-01 3.53873491e-01 -2.77982980e-01
1.04673135e+00 -1.35313821e+00 -6.76324785e-01 -4.94631588e-01
3.93258929e-01 -3.50774765e-01 7.27202356e-01 1.46319568e-01
-8.89145672e-01 -7.84772515e-01 -8.45980525e-01 3.92392844e-01
-6.76359683e-02 3.04311544e-01 8.11959326e-01 7.51654267e-01
-8.57557893e-01 4.69179720e-01 -1.09460914e+00 -2.46613115e-01
8.71389925e-01 5.73782146e-01 4.52705957e-02 -5.53476922e-02
-9.02390182e-01 1.51037917e-01 1.16703343e+00 3.09306771e-01
-1.35937893e+00 -4.24650520e-01 -7.04026699e-01 -2.72314042e-01
9.14851427e-01 -4.49809343e-01 1.21938550e+00 -1.60965145e+00
-1.33021784e+00 8.85735989e-01 -7.91097805e-02 -5.55035830e-01
5.36346257e-01 7.30359368e-03 -5.20343743e-02 6.63819015e-01
4.09971595e-01 1.33612275e+00 9.70449090e-01 -1.61713076e+00
-1.09170020e+00 1.74916571e-03 -3.20479134e-03 2.86137052e-02
-5.54724000e-02 -2.05056965e-01 -1.35262859e+00 -6.35492265e-01
2.77775019e-01 -9.66077030e-01 -4.98543411e-01 -2.38451347e-01
-7.50184119e-01 -2.76186675e-01 1.10007918e+00 -5.40907085e-01
1.10863876e+00 -1.90143108e+00 7.21372366e-02 2.50259221e-01
1.19480610e-01 5.06257355e-01 -8.89520049e-02 -5.31741858e-01
1.95458457e-01 2.32098609e-01 -6.44653499e-01 2.59068441e-02
-4.32856649e-01 2.92748809e-01 -2.22190879e-02 4.53449845e-01
1.66070908e-01 1.11330211e+00 -1.02144361e+00 -9.49952781e-01
2.67143250e-01 1.85034275e-02 -5.45457959e-01 2.50815630e-01
-6.41461134e-01 7.26936877e-01 -9.19290245e-01 9.31202769e-01
6.29690349e-01 -5.37763059e-01 1.41702399e-01 -2.92225271e-01
1.40117601e-01 -2.73434043e-01 -1.09199727e+00 1.65661407e+00
5.56959450e-01 6.55860305e-01 -9.49454401e-03 -1.18934882e+00
5.90906501e-01 7.52268806e-02 9.87229705e-01 -2.41506144e-01
4.44613338e-01 7.08458722e-02 -1.48205966e-01 -7.18209326e-01
6.59143150e-01 2.65750498e-01 8.50888938e-02 1.31042257e-01
3.20402533e-02 6.12736773e-03 5.40466011e-01 1.81456700e-01
3.78399193e-01 5.33470333e-01 -2.78398573e-01 -3.30048114e-01
4.53342378e-01 5.35949051e-01 8.92537117e-01 8.05086732e-01
-6.91261888e-01 6.79382086e-01 2.10764958e-03 -2.64935613e-01
-7.24073052e-01 -8.28252733e-01 -6.05929494e-02 9.02122855e-01
8.53009820e-01 2.22165301e-01 -1.41485023e+00 -1.14256048e+00
-1.86168939e-01 4.72905010e-01 -4.98717099e-01 1.06515385e-01
-7.30792522e-01 -8.11349988e-01 3.15016985e-01 7.91404009e-01
9.31031108e-01 -1.21114695e+00 -4.77558732e-01 2.42374361e-01
-5.43264866e-01 -1.43477380e+00 -6.72016263e-01 -8.95653889e-02
-1.06033814e+00 -1.23471725e+00 -8.25265706e-01 -1.10353827e+00
9.33568418e-01 7.10900962e-01 8.12096119e-01 3.07734579e-01
-1.05250612e-01 5.23082078e-01 -2.66271472e-01 9.59364027e-02
-2.02808082e-01 -6.03010990e-02 -1.30822167e-01 3.27732265e-01
3.26213390e-01 1.61817595e-01 -7.69030213e-01 5.18777668e-01
-1.12404764e+00 1.11680672e-01 2.97737837e-01 4.13862646e-01
1.08969188e+00 1.73717842e-01 5.09371400e-01 -1.17301929e+00
-2.40901932e-01 -2.97268778e-01 -8.44965935e-01 2.88672745e-01
-3.14770669e-01 -2.61765033e-01 9.41541940e-02 -5.81390858e-01
-1.09277070e+00 5.01906693e-01 2.68704474e-01 -7.03499794e-01
-3.13677073e-01 1.95013791e-01 -2.98826873e-01 -1.95437565e-01
2.40946457e-01 1.62717625e-01 -2.04854146e-01 -2.67880738e-01
4.75662440e-01 4.00235415e-01 7.99864709e-01 -4.95768845e-01
7.79793978e-01 6.15258157e-01 -3.56890827e-01 -7.25089133e-01
-9.77273226e-01 -7.63459265e-01 -7.85947025e-01 -5.45162737e-01
1.55357087e+00 -8.80460560e-01 -3.59649330e-01 4.83857960e-01
-1.04933226e+00 -6.76886976e-01 -1.74925074e-01 3.97079915e-01
-6.05382383e-01 4.01112884e-01 -7.16549754e-01 -7.56174862e-01
-1.03100494e-01 -1.44549191e+00 1.06200671e+00 7.21338809e-01
7.14451969e-02 -9.19106662e-01 -6.25938535e-01 8.41143847e-01
-1.63884655e-01 2.15245605e-01 3.42572391e-01 -6.50189698e-01
-1.53031480e+00 -5.32962754e-02 -4.10023391e-01 3.18916708e-01
1.95471764e-01 2.59535104e-01 -7.18448400e-01 -1.15741126e-01
-2.52755553e-01 -1.97689217e-02 1.00360537e+00 9.11187649e-01
1.57630885e+00 -1.61264092e-01 -7.41602719e-01 7.42169499e-01
1.31529856e+00 5.81932247e-01 5.23169339e-01 3.82284522e-01
1.28340173e+00 3.86293620e-01 1.16776514e+00 3.14050284e-03
1.79080233e-01 4.73389298e-01 3.38911772e-01 -2.51706630e-01
-1.62809327e-01 -2.28080545e-02 2.58599907e-01 5.06444395e-01
-1.07261702e-01 -3.56489241e-01 -8.62318099e-01 8.86262178e-01
-1.89769375e+00 -8.92662168e-01 -2.19359219e-01 1.81234694e+00
7.42810190e-01 2.08783299e-01 3.43101323e-01 -1.60922438e-01
1.17320788e+00 1.29152071e-02 -6.36745453e-01 4.65589762e-01
-1.66331269e-02 -1.90912291e-01 8.69363487e-01 2.99916953e-01
-1.58297682e+00 1.60305285e+00 5.74065399e+00 8.88224781e-01
-9.03247774e-01 1.02970809e-01 1.02024329e+00 2.72701561e-01
1.44531488e-01 -4.61274236e-02 -1.01775944e+00 4.29929137e-01
5.07718861e-01 1.17575742e-01 2.63621628e-01 9.77973044e-01
3.11252207e-01 -3.00082535e-01 -8.54768157e-01 6.93194151e-01
-2.29772609e-02 -1.41226399e+00 2.21172392e-01 -2.78294265e-01
1.19891322e+00 -1.50181457e-01 -1.01983368e-01 8.38990733e-02
1.32972792e-01 -5.53457797e-01 9.11140501e-01 3.56765717e-01
5.27592480e-01 -6.09595895e-01 5.26115358e-01 2.24359065e-01
-1.26608300e+00 -5.33122162e-04 -2.44070292e-01 5.72601557e-01
2.91205913e-01 7.70116374e-02 -7.80923963e-01 3.31717074e-01
7.32532442e-01 7.99599886e-01 -5.12716174e-01 1.40133345e+00
-1.10372171e-01 1.00170684e+00 -1.10324837e-01 1.56986594e-01
5.79573095e-01 -3.86700571e-01 5.56893945e-01 1.10483122e+00
8.70153308e-03 3.01011652e-01 8.38607788e-01 8.80627096e-01
-9.45446417e-02 -1.17722832e-01 -1.66716367e-01 -3.47881615e-01
3.09291661e-01 1.10633028e+00 -1.73526430e+00 -7.96922028e-01
-1.63678318e-01 1.04546964e+00 -1.64248288e-01 8.58696103e-01
-1.08568215e+00 -1.37296505e-02 2.45968163e-01 -6.19462803e-02
6.19652033e-01 -1.12759113e-01 -9.94903147e-02 -1.08187580e+00
-2.82307416e-01 -6.23000920e-01 3.01648945e-01 -6.81294024e-01
-9.04613912e-01 5.38935363e-01 2.03130290e-01 -1.21772635e+00
-3.89880836e-02 -6.85451388e-01 -3.28875780e-01 2.50562072e-01
-1.29154027e+00 -9.38523591e-01 -5.21522522e-01 4.97239590e-01
1.03045750e+00 3.95076424e-02 -1.59184877e-02 3.76251221e-01
-9.65823591e-01 2.14736149e-01 -1.84339389e-01 6.62961483e-01
1.88138872e-01 -1.04527175e+00 1.72306284e-01 1.16020441e+00
1.23020932e-01 3.52485031e-01 4.52780396e-01 -1.02868819e+00
-1.13660026e+00 -1.70981991e+00 -6.79173768e-02 -4.13930058e-01
4.02506441e-01 -2.58805137e-02 -9.69119072e-01 7.57873595e-01
1.63613893e-02 2.15978652e-01 2.76067853e-01 -7.73231983e-01
3.46942186e-01 5.29156253e-02 -9.91333008e-01 5.41082919e-01
1.02909350e+00 2.95707188e-03 -5.18022716e-01 5.59486747e-01
1.15255511e+00 -6.37119532e-01 -4.30279970e-01 4.95081753e-01
1.56577185e-01 -4.61814344e-01 1.11657143e+00 -6.89335465e-01
1.95832431e-01 -7.15978682e-01 -5.56808747e-02 -6.17676735e-01
-1.00351153e-02 -5.68774104e-01 -4.04307246e-02 1.40201521e+00
3.14422518e-01 -1.73719689e-01 1.38804007e+00 7.28722394e-01
-2.01931924e-01 -5.11153758e-01 -4.96842593e-01 -7.34021127e-01
-3.69512975e-01 -4.64337468e-01 3.40520293e-01 9.36078191e-01
-7.51217842e-01 -1.49254978e-01 -7.41878152e-02 3.17824423e-01
9.01997030e-01 4.48138744e-01 6.38328493e-01 -9.90511537e-01
-1.60299003e-01 -4.32064950e-01 -4.11770523e-01 -1.53335559e+00
4.15911496e-01 -9.55381632e-01 5.44533432e-01 -1.58428967e+00
3.68215054e-01 -5.97493052e-01 -3.21215361e-01 3.71689558e-01
-5.74022472e-01 6.88045919e-01 2.05764607e-01 2.88430423e-01
-1.13384128e+00 2.00891241e-01 1.71731722e+00 -5.29379308e-01
-5.62669873e-01 3.28096300e-01 -3.66007477e-01 8.19128513e-01
6.99801445e-01 -5.44535816e-01 -4.85156357e-01 -1.85561135e-01
-7.65256226e-01 6.35850728e-02 4.56134975e-01 -8.74459505e-01
1.77758038e-01 -6.75905585e-01 4.09904361e-01 -8.50661874e-01
7.85963684e-02 -9.20173347e-01 8.16105232e-02 3.61342311e-01
-1.24886371e-01 -6.13252044e-01 8.88713747e-02 8.76663446e-01
-1.94239959e-01 -4.75600451e-01 8.94723177e-01 -2.53733933e-01
-1.42570186e+00 7.92289376e-01 -3.72304618e-01 2.75469452e-01
1.33345783e+00 -6.61516428e-01 -2.24403962e-02 5.98223545e-02
-7.60004818e-01 4.33487952e-01 3.64667803e-01 3.74865472e-01
6.11097574e-01 -1.09855878e+00 -1.32694602e-01 -4.78377678e-02
-1.40914515e-01 4.18316156e-01 9.32673439e-02 8.39580774e-01
-6.86021268e-01 3.81183058e-01 -1.10296763e-01 -1.16120088e+00
-9.99182284e-01 7.10845292e-01 4.71577555e-01 3.98726583e-01
-5.52949905e-01 1.00114274e+00 5.01221299e-01 1.10670328e-01
3.81868631e-01 -5.36122501e-01 -2.51158863e-01 -7.76780769e-04
3.04391205e-01 2.87764221e-01 -3.78522098e-01 -9.07179892e-01
-3.39872748e-01 7.02857852e-01 -1.91445071e-02 1.38452068e-01
1.01402795e+00 -3.02843034e-01 2.08372492e-02 3.31595987e-01
1.01406693e+00 -4.31063622e-01 -1.80725980e+00 -1.94731466e-02
2.26974994e-01 -5.55903018e-01 3.14422920e-02 -3.97279918e-01
-1.73424160e+00 5.24057865e-01 5.77603459e-01 1.28730476e-01
1.14462793e+00 5.15682437e-02 8.79379749e-01 3.01817983e-01
5.68615735e-01 -1.21198642e+00 1.34950340e-01 1.89736709e-01
1.36574671e-01 -1.30853856e+00 -1.49593979e-01 -9.33907092e-01
-7.93476045e-01 9.01743889e-01 1.09489262e+00 -2.05821142e-01
2.98973769e-01 -5.72411567e-02 1.98891848e-01 -1.01766177e-01
-1.31045831e-02 -4.87185746e-01 5.72426498e-01 4.91762251e-01
2.20568478e-02 -1.70462187e-02 -1.35325432e-01 5.52557468e-01
4.18702394e-01 7.87031725e-02 3.81720871e-01 8.64173114e-01
-7.26326108e-01 -8.75382543e-01 -4.55489457e-01 4.58793759e-01
-5.66193163e-01 1.58134714e-01 -3.31258744e-01 8.99247110e-01
1.98496535e-01 9.25744534e-01 2.59581525e-02 -1.40567198e-01
-8.70622396e-02 -8.01479593e-02 4.43596601e-01 -6.00796402e-01
-3.66178662e-01 3.90059680e-01 -2.03318551e-01 -6.79749429e-01
-9.88226056e-01 -5.33145785e-01 -1.90877044e+00 3.33857417e-01
-5.81627965e-01 1.28441066e-01 3.42653751e-01 1.02317715e+00
-5.07280119e-02 6.46288157e-01 3.11865449e-01 -1.26845455e+00
1.52766988e-01 -5.64613461e-01 -3.61459494e-01 5.44514537e-01
2.49564081e-01 -8.17569673e-01 -1.22171767e-01 7.28477657e-01] | [9.152753829956055, -0.17368514835834503] |
7873cb16-9a8c-4341-abe9-13a9528e9db4 | cascadetabnet-an-approach-for-end-to-end | 2004.12629 | null | https://arxiv.org/abs/2004.12629v2 | https://arxiv.org/pdf/2004.12629v2.pdf | CascadeTabNet: An approach for end to end table detection and structure recognition from image-based documents | An automatic table recognition method for interpretation of tabular data in document images majorly involves solving two problems of table detection and table structure recognition. The prior work involved solving both problems independently using two separate approaches. More recent works signify the use of deep learning-based solutions while also attempting to design an end to end solution. In this paper, we present an improved deep learning-based end to end approach for solving both problems of table detection and structure recognition using a single Convolution Neural Network (CNN) model. We propose CascadeTabNet: a Cascade mask Region-based CNN High-Resolution Network (Cascade mask R-CNN HRNet) based model that detects the regions of tables and recognizes the structural body cells from the detected tables at the same time. We evaluate our results on ICDAR 2013, ICDAR 2019 and TableBank public datasets. We achieved 3rd rank in ICDAR 2019 post-competition results for table detection while attaining the best accuracy results for the ICDAR 2013 and TableBank dataset. We also attain the highest accuracy results on the ICDAR 2019 table structure recognition dataset. Additionally, we demonstrate effective transfer learning and image augmentation techniques that enable CNNs to achieve very accurate table detection results. Code and dataset has been made available at: https://github.com/DevashishPrasad/CascadeTabNet | ['Manish Visave', 'Kshitij Kapadni', 'Devashish Prasad', 'Ayan Gadpal', 'Kavita Sultanpure'] | 2020-04-27 | null | null | null | null | ['table-recognition', 'table-detection'] | ['computer-vision', 'miscellaneous'] | [ 9.09943972e-03 2.02529952e-01 -1.67931825e-01 -4.28389132e-01
-1.34652317e+00 -8.13085198e-01 4.11976844e-01 3.15059066e-01
-1.63897529e-01 3.58182102e-01 2.79419988e-01 -4.39033240e-01
3.00182134e-01 -8.92535865e-01 -1.10759234e+00 -1.27783924e-01
6.98731616e-02 8.77342999e-01 -1.00522913e-01 -2.73828059e-01
6.45538941e-02 5.95576227e-01 -1.02203548e+00 1.35619259e+00
2.41711929e-01 1.42383444e+00 -3.46332371e-01 1.11519718e+00
-9.79532152e-02 1.40293527e+00 -8.25050890e-01 -7.48639166e-01
5.25677085e-01 -3.09436798e-01 -9.91220236e-01 1.30147904e-01
9.00130987e-01 -3.72602552e-01 -5.17033756e-01 5.29681385e-01
4.99665409e-01 -1.74404547e-01 3.89100552e-01 -9.22602534e-01
-8.40134144e-01 1.05962050e+00 -7.28830338e-01 3.78008634e-01
3.45817357e-01 -1.67897925e-01 1.15850985e+00 -1.07576919e+00
7.27209330e-01 1.33026016e+00 7.45309949e-01 3.44764322e-01
-9.57187593e-01 -6.59926176e-01 -1.97940040e-02 2.87060052e-01
-1.52239454e+00 -5.24314344e-01 3.16109180e-01 -2.09846064e-01
1.39886630e+00 3.79806906e-01 1.26779646e-01 9.89794552e-01
2.91176558e-01 1.16500616e+00 4.49417323e-01 -2.88520485e-01
-6.51343241e-02 1.06751367e-01 -1.51060736e-02 9.99694824e-01
3.32748145e-01 -6.10313535e-01 -6.82134032e-01 1.21253178e-01
8.09019983e-01 -1.84336513e-01 1.56628355e-01 -1.32876813e-01
-1.35626495e+00 6.87707961e-01 6.04339600e-01 2.56736815e-01
-1.92246839e-01 5.89977130e-02 7.94969857e-01 1.21746890e-01
2.81606585e-01 3.61649960e-01 -8.27986598e-01 8.87322798e-02
-7.61619866e-01 3.82285774e-01 7.40816772e-01 1.25908363e+00
8.08141604e-02 -1.01020493e-01 -6.24179125e-01 8.25140357e-01
3.17010805e-02 1.75075695e-01 1.21488154e-01 -5.52418351e-01
1.07047403e+00 9.63610649e-01 -1.15428679e-01 -9.34827268e-01
-4.80178505e-01 -5.56837022e-01 -1.20880175e+00 1.06825773e-03
6.53162003e-01 1.78354844e-01 -1.13065243e+00 9.06345308e-01
1.01004831e-01 -7.35898554e-01 2.31492922e-01 6.66151881e-01
1.15467131e+00 7.70438969e-01 -2.39906952e-01 4.57781941e-01
1.83941114e+00 -1.00995862e+00 -8.21517646e-01 -1.17147863e-01
7.86312044e-01 -8.85948181e-01 1.09498286e+00 6.94729924e-01
-1.25503862e+00 -5.50593972e-01 -1.30958307e+00 -7.93659389e-01
-8.44206870e-01 7.47103512e-01 5.18991947e-01 6.59892499e-01
-8.41368258e-01 2.04163119e-01 -6.57495499e-01 -3.29468518e-01
7.79369056e-01 6.47293746e-01 -5.17435074e-01 -3.88338640e-02
-8.33104849e-01 6.77119434e-01 4.35865253e-01 4.28381532e-01
-5.72782636e-01 -5.89381456e-01 -8.71830463e-01 2.16365471e-01
5.69266438e-01 -3.37970287e-01 1.33198619e+00 -4.52554375e-01
-8.70820880e-01 1.34307754e+00 7.67700523e-02 -9.52708602e-01
6.21470571e-01 -9.10232216e-02 -2.80163646e-01 5.04739769e-02
8.28108639e-02 8.04089427e-01 2.43465289e-01 -9.82412159e-01
-6.07320368e-01 -4.56013203e-01 -1.62169039e-01 9.33521092e-02
6.60686493e-02 1.75437823e-01 -8.71379852e-01 -7.69572616e-01
1.94034144e-01 -5.41734457e-01 1.51641369e-01 -1.08075760e-01
-8.63450229e-01 -8.77135992e-02 6.15340531e-01 -1.13133347e+00
1.00708938e+00 -2.00812387e+00 -3.37634474e-01 9.75577254e-03
4.22676742e-01 1.77015841e-01 9.45731699e-02 1.49623618e-01
-2.37369239e-01 3.57749790e-01 -3.11962754e-01 -3.71873677e-01
-1.55461356e-02 -1.63229823e-01 -2.72626460e-01 3.68624240e-01
4.58761305e-01 1.20621860e+00 -1.22763313e-01 -5.59885323e-01
1.08876146e-01 6.03519619e-01 -5.48042953e-01 -2.79168151e-02
-2.14283362e-01 -1.77436974e-04 -7.78756686e-04 1.15696681e+00
7.98105717e-01 -5.05582809e-01 4.32453811e-01 -4.20827836e-01
2.13867024e-01 3.60745341e-01 -1.26603043e+00 1.38407779e+00
-8.77971500e-02 6.10992670e-01 1.06133342e-01 -7.62379408e-01
9.95437503e-01 2.57788658e-01 2.26690933e-01 -1.11393559e+00
2.83238709e-01 1.42221421e-01 -2.98110135e-02 5.54706715e-02
6.86822712e-01 3.57292920e-01 -4.24383223e-01 -2.50956882e-02
4.46193554e-02 2.15258688e-01 3.00851405e-01 4.12314206e-01
1.20270991e+00 -1.06057683e-02 4.49340343e-01 -2.26576596e-01
7.52273202e-01 3.01586807e-01 2.91084230e-01 9.60964501e-01
-1.10637415e-02 1.16660452e+00 7.84413636e-01 -1.06754994e+00
-1.24069881e+00 -8.56959581e-01 -5.13137616e-02 1.06398273e+00
-3.49582970e-01 -6.05145574e-01 -7.94184387e-01 -8.78056169e-01
6.55242428e-02 3.87220472e-01 -1.11690712e+00 1.57725871e-01
-9.47185338e-01 -7.98416436e-01 8.93769443e-01 9.61279273e-01
8.45694363e-01 -1.30293965e+00 -2.95279235e-01 2.26716205e-01
-3.25147331e-01 -1.56231844e+00 -5.37817597e-01 7.14769542e-01
-5.49860716e-01 -1.20113540e+00 -3.57325524e-01 -8.60967040e-01
5.00277340e-01 -2.76179671e-01 1.37011719e+00 1.56319052e-01
-6.28996670e-01 -1.03777416e-01 5.70901996e-03 -5.80492675e-01
-3.82889003e-01 4.49904561e-01 -4.26641613e-01 -1.51623145e-01
3.74106407e-01 1.65707007e-01 -2.71727085e-01 2.26561964e-01
-7.16080964e-01 1.75781980e-01 5.78237116e-01 6.87070191e-01
7.25674868e-01 -1.63949989e-02 1.35855526e-01 -1.14200413e+00
4.37608004e-01 -5.65120950e-02 -8.89814436e-01 4.42198932e-01
-5.78097165e-01 1.09702930e-01 8.02209079e-01 1.68723643e-01
-8.71040642e-01 3.56532097e-01 -3.10926676e-01 -3.17348838e-01
-1.11767843e-01 2.90207148e-01 -4.38921005e-01 4.78293002e-01
7.28151500e-01 1.50397211e-01 -2.63663083e-01 -5.03805161e-01
1.69440642e-01 6.28958523e-01 8.62969995e-01 -3.45346540e-01
6.09528542e-01 6.93954289e-01 -1.76860303e-01 -8.60501602e-02
-9.76274133e-01 -2.81938076e-01 -8.71669292e-01 1.36335373e-01
1.24961126e+00 -1.18987024e+00 -1.04939699e+00 3.34995478e-01
-9.54286575e-01 -2.58283466e-01 -7.31479526e-02 -1.31666988e-01
-3.83614510e-01 -9.40138921e-02 -9.85233128e-01 -5.53593159e-01
-7.04293370e-01 -1.11151195e+00 1.34213066e+00 -6.91966340e-02
-2.33141556e-01 -6.62797987e-01 -4.21157897e-01 8.40231717e-01
1.26792520e-01 4.06793624e-01 9.67585921e-01 -9.51285779e-01
-1.02671897e+00 -4.42274511e-01 -5.53339005e-01 -7.00199306e-02
-9.34812129e-02 5.28757162e-02 -1.02692628e+00 -7.66549772e-03
-5.50590217e-01 -3.62032413e-01 1.11264765e+00 3.53842050e-01
1.35584700e+00 -3.38171393e-01 -1.18301943e-01 7.24529326e-01
1.36900365e+00 4.89938885e-01 9.43316281e-01 6.90503061e-01
1.07538199e+00 4.68541741e-01 6.03688121e-01 3.32908630e-01
4.51970577e-01 6.35410130e-01 2.27056414e-01 -5.19362271e-01
-2.51344502e-01 -1.77194446e-01 3.54183018e-02 2.52799809e-01
3.89192402e-01 -6.91032767e-01 -1.28855669e+00 4.73353833e-01
-1.86018968e+00 -6.88953459e-01 -2.52446830e-01 1.88301885e+00
6.12743020e-01 6.64031327e-01 2.47727633e-01 4.31354076e-01
6.58785641e-01 -6.83401600e-02 -3.85333627e-01 -8.02444875e-01
-3.13341767e-01 2.81040460e-01 8.88498366e-01 2.01222241e-01
-1.49431908e+00 1.00438988e+00 6.12007236e+00 7.39488602e-01
-6.76931024e-01 -2.39483744e-01 1.51354909e+00 -7.33790174e-02
2.71896809e-01 -6.35492444e-01 -1.24561083e+00 -1.43805191e-01
1.10786903e+00 1.85249478e-01 3.53808790e-01 7.88922489e-01
-2.01206207e-01 -3.10400426e-02 -1.11714435e+00 1.18999028e+00
1.62166908e-01 -1.86173904e+00 2.72149712e-01 1.43763825e-01
2.36346036e-01 -2.13834926e-01 1.50380835e-01 4.57336843e-01
2.09891051e-01 -1.47335947e+00 8.83145511e-01 7.63040483e-02
9.75581527e-01 -1.07829285e+00 9.24909949e-01 -1.88182190e-01
-1.42107451e+00 -7.03403428e-02 -2.02459946e-01 1.06060535e-01
-6.78287148e-01 1.87303454e-01 -1.25248289e+00 5.63132524e-01
1.05767059e+00 7.10422933e-01 -1.08627355e+00 5.19985557e-01
1.22565560e-01 4.78182465e-01 -1.17483221e-01 1.82206392e-01
2.16929272e-01 2.70275295e-01 4.42633592e-03 1.41810679e+00
-2.53511161e-01 -1.76893726e-01 4.74559003e-03 9.86886084e-01
-8.13840806e-01 1.18509732e-01 -5.15626609e-01 -1.42182469e-01
7.90532976e-02 1.44890237e+00 -1.52167737e+00 -6.79175138e-01
-2.90503412e-01 9.03829455e-01 3.08945477e-01 -2.26528719e-01
-8.37427557e-01 -3.20335895e-01 3.05217415e-01 7.49078020e-02
8.34552824e-01 7.91969746e-02 -1.07476175e+00 -1.12309611e+00
3.03424090e-01 -1.37273860e+00 9.77264762e-01 -6.60839677e-01
-8.76036823e-01 8.11691284e-01 -2.29176596e-01 -9.24025238e-01
-1.38978690e-01 -9.19300079e-01 -1.04658969e-01 7.36929357e-01
-8.97538364e-01 -1.23074698e+00 -2.72452384e-01 5.48223495e-01
7.29788601e-01 -4.17745769e-01 9.10477161e-01 4.94314164e-01
-7.99595177e-01 1.12177801e+00 6.36414662e-02 9.50417101e-01
6.81560874e-01 -1.76495326e+00 1.02222788e+00 8.89938116e-01
1.77411363e-01 5.09447873e-01 3.75782311e-01 -6.89719915e-01
-1.57309258e+00 -1.20869315e+00 9.24959183e-01 -9.48361814e-01
1.88901708e-01 -1.29458332e+00 -8.78677607e-01 9.76198852e-01
3.69220287e-01 1.01874880e-01 3.21913898e-01 1.56057104e-01
-6.26149178e-01 -2.76592463e-01 -1.43825519e+00 4.52725023e-01
7.85082042e-01 -4.39547271e-01 -9.74202007e-02 5.29152930e-01
4.54061031e-01 -9.88971233e-01 -9.17545855e-01 3.94459575e-01
7.31566608e-01 -7.67686725e-01 1.01072204e+00 -2.68931508e-01
7.78224826e-01 -3.78622502e-01 -2.56482244e-01 -3.16031814e-01
-3.02782416e-01 -2.97254890e-01 -2.68184751e-01 1.13202894e+00
8.61727595e-01 -1.11049183e-01 1.29734993e+00 3.90624404e-01
1.33386599e-02 -7.95322180e-01 -6.89880490e-01 -5.61562598e-01
1.73377022e-01 -1.62020922e-01 5.76078355e-01 7.82757878e-01
-4.68685746e-01 5.79161167e-01 -3.28154087e-01 9.17098448e-02
3.63839746e-01 3.85169648e-02 8.04290414e-01 -7.38202989e-01
-4.98619154e-02 -3.91434222e-01 -2.53485531e-01 -5.41651726e-01
-2.49270752e-01 -1.05266309e+00 -9.25748795e-02 -1.78757727e+00
3.29752475e-01 1.32408202e-01 -2.26400629e-01 7.98125863e-01
6.05730377e-02 5.33975124e-01 3.23251814e-01 9.71523672e-02
-7.93704629e-01 -9.09756199e-02 9.13345218e-01 -4.75585371e-01
8.20696726e-02 -3.71226370e-01 -1.02676547e+00 1.61905378e-01
1.01654983e+00 -5.22189617e-01 -5.57855442e-02 -3.44814181e-01
2.34200627e-01 1.64078847e-01 1.17217757e-01 -1.14331460e+00
6.63136393e-02 4.72743154e-01 1.39105296e+00 -1.47806299e+00
7.07817450e-02 -5.38788497e-01 -2.65123338e-01 5.50568104e-01
-6.03808761e-01 5.79526722e-01 7.41542339e-01 2.03948304e-01
-1.42608196e-01 4.05966431e-01 7.21447587e-01 -4.67810810e-01
-4.66648281e-01 1.57170326e-01 -6.53183162e-01 -9.40739289e-02
5.23032486e-01 -2.99922258e-01 -3.69143605e-01 -2.51819044e-01
-6.53064787e-01 2.73985863e-01 8.87546837e-02 6.37738645e-01
6.52985394e-01 -1.01798153e+00 -8.86241436e-01 2.26622403e-01
3.86706889e-01 1.49019569e-01 -5.33042364e-02 4.01428103e-01
-1.02922964e+00 9.75984335e-01 -2.96957105e-01 -4.86182600e-01
-1.46641767e+00 6.42163336e-01 5.58704436e-01 -8.89022589e-01
-6.91595852e-01 1.10924780e+00 2.18752325e-01 -9.03419793e-01
6.19470954e-01 -6.74940109e-01 -2.09475130e-01 4.52365912e-03
6.85198069e-01 -3.84710450e-03 8.55605960e-01 -2.10745901e-01
-7.08357871e-01 -5.28072938e-04 -7.73838758e-01 1.31873637e-01
1.20061910e+00 2.85618812e-01 9.42109302e-02 2.02389956e-01
1.20946574e+00 -1.39968814e-02 -7.48765290e-01 1.02057755e-01
2.53808945e-01 -1.15695067e-01 -1.27630591e-01 -1.56797123e+00
-1.11657834e+00 5.91899574e-01 7.29858637e-01 -4.17615697e-02
1.16885173e+00 -1.32499605e-01 7.52237737e-01 7.14809775e-01
-1.83137253e-01 -8.12934101e-01 -2.39364021e-02 6.13589406e-01
1.06887853e+00 -1.45833123e+00 1.48649246e-01 -5.22941530e-01
-6.90868199e-01 1.31400073e+00 8.67553055e-01 -4.33977507e-02
2.02485263e-01 8.21415663e-01 2.20189616e-01 -4.20823961e-01
-9.35340762e-01 -6.20954223e-02 4.66682047e-01 3.16492945e-01
8.99896502e-01 -2.98494883e-02 1.26998186e-01 6.01927698e-01
-4.13003176e-01 -1.79510206e-01 6.10917807e-01 1.22896671e+00
1.74217790e-01 -1.06550395e+00 -7.34371662e-01 6.72532976e-01
-8.93754065e-01 -3.99929821e-01 -1.00337768e+00 1.06831622e+00
-5.91242090e-02 6.69964910e-01 3.12323600e-01 -4.54949737e-01
5.09182572e-01 7.65849724e-02 4.54570085e-01 -4.96726334e-01
-1.35035992e+00 1.61573663e-01 3.76918346e-01 -4.70602781e-01
2.69847780e-01 -5.33393860e-01 -1.33316314e+00 -5.84032118e-01
1.12793289e-01 -1.44339517e-01 3.99052620e-01 5.48814237e-01
5.31397700e-01 9.55542207e-01 8.31524059e-02 -1.08086474e-01
-1.03217885e-01 -9.81640577e-01 -4.58702803e-01 1.14297830e-01
2.96629041e-01 -2.27667734e-01 2.60460019e-01 2.75703758e-01] | [11.69408893585205, 3.0121941566467285] |
a4117d29-ad7a-44c5-9b1c-b928aa088926 | few-shot-parameter-efficient-fine-tuning-is | 2205.05638 | null | https://arxiv.org/abs/2205.05638v2 | https://arxiv.org/pdf/2205.05638v2.pdf | Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning | Few-shot in-context learning (ICL) enables pre-trained language models to perform a previously-unseen task without any gradient-based training by feeding a small number of training examples as part of the input. ICL incurs substantial computational, memory, and storage costs because it involves processing all of the training examples every time a prediction is made. Parameter-efficient fine-tuning (PEFT) (e.g. adapter modules, prompt tuning, sparse update methods, etc.) offers an alternative paradigm where a small set of parameters are trained to enable a model to perform the new task. In this paper, we rigorously compare few-shot ICL and PEFT and demonstrate that the latter offers better accuracy as well as dramatically lower computational costs. Along the way, we introduce a new PEFT method called (IA)$^3$ that scales activations by learned vectors, attaining stronger performance while only introducing a relatively tiny amount of new parameters. We also propose a simple recipe based on the T0 model called T-Few that can be applied to new tasks without task-specific tuning or modifications. We validate the effectiveness of T-Few on completely unseen tasks by applying it to the RAFT benchmark, attaining super-human performance for the first time and outperforming the state-of-the-art by 6% absolute. All of the code used in our experiments is publicly available. | ['Colin Raffel', 'Mohit Bansal', 'Tenghao Huang', 'Jay Mohta', 'Mohammed Muqeeth', 'Derek Tam', 'Haokun Liu'] | 2022-05-11 | null | null | null | null | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 2.08973870e-01 -2.45274916e-01 -1.44000202e-01 -4.21164125e-01
-7.09617615e-01 -3.12106758e-01 8.59392226e-01 1.79506809e-01
-9.03046608e-01 6.51624978e-01 -6.25754893e-02 -1.57686934e-01
1.08639456e-01 -5.59332311e-01 -8.34439993e-01 -5.37697196e-01
1.30282059e-01 5.07575095e-01 5.86532414e-01 -4.42412376e-01
1.02165021e-01 3.10010225e-01 -1.55728495e+00 3.43652636e-01
7.08071589e-01 9.72537160e-01 5.68524241e-01 4.50181276e-01
-2.27314144e-01 3.44644666e-01 -4.45972592e-01 -3.74193192e-01
1.65450528e-01 -7.80425817e-02 -6.43339455e-01 -1.88812867e-01
3.60471994e-01 -1.69218734e-01 -6.30452037e-02 5.85934520e-01
6.06053710e-01 6.43178761e-01 5.00745595e-01 -7.37436235e-01
-5.29389203e-01 5.65542459e-01 -2.50232428e-01 3.94061118e-01
3.23721692e-02 2.34568954e-01 8.87932718e-01 -1.50680530e+00
5.99690914e-01 8.59570384e-01 6.65907383e-01 7.83663988e-01
-1.32411575e+00 -4.94269460e-01 3.28528881e-01 3.03173035e-01
-1.22757745e+00 -6.54723048e-01 6.18605793e-01 -2.37703770e-01
1.53161752e+00 1.18054479e-01 4.10119563e-01 1.35766053e+00
-3.56391855e-02 7.42058456e-01 8.47395301e-01 -5.31593621e-01
6.63394749e-01 3.18922162e-01 3.32788914e-01 6.96057677e-01
-3.59791378e-03 1.24699868e-01 -6.24219418e-01 -9.26082805e-02
3.50342900e-01 1.26506448e-01 -1.66389987e-01 -3.81452829e-01
-1.18641353e+00 7.31051147e-01 5.32177091e-01 6.33444190e-01
-2.80973613e-01 1.37353539e-01 5.33799052e-01 4.52972621e-01
5.49236655e-01 5.84218025e-01 -6.75305665e-01 -1.71966791e-01
-1.05138779e+00 -6.59561306e-02 6.21218681e-01 7.04576790e-01
9.21223044e-01 2.45636299e-01 -4.73893434e-01 1.10779560e+00
-2.66431242e-01 6.77930787e-02 9.38353121e-01 -5.94318271e-01
3.20094913e-01 3.73237491e-01 6.07183315e-02 -4.81212586e-01
-3.75350922e-01 -7.06300855e-01 -6.92257285e-01 -1.35952637e-01
1.26916185e-01 -1.08003475e-01 -1.03642368e+00 1.82703519e+00
2.09894627e-01 4.80248272e-01 -2.41276044e-02 6.15200162e-01
5.58739603e-01 7.71280944e-01 1.52070194e-01 -4.03063983e-01
1.03725731e+00 -1.26408923e+00 -3.86835933e-01 -6.45278752e-01
7.44894266e-01 -5.28727174e-01 1.63930249e+00 3.03625613e-01
-8.93185198e-01 -7.48958588e-01 -9.24256027e-01 -1.90911237e-02
-5.72296262e-01 2.85559651e-02 6.98048174e-01 5.24917364e-01
-1.13496864e+00 8.41980934e-01 -7.62908518e-01 -4.77633893e-01
3.89874697e-01 3.23322713e-01 -3.00461441e-01 -1.79528609e-01
-1.19616210e+00 9.19800043e-01 5.86420774e-01 -2.09582344e-01
-1.12880468e+00 -9.31751728e-01 -8.02961648e-01 3.62591863e-01
7.41286099e-01 -6.14271879e-01 1.36169946e+00 -7.20346212e-01
-1.74221945e+00 5.72730005e-01 -3.20523709e-01 -5.92427492e-01
3.35882008e-01 -3.88128161e-01 -3.27293962e-01 -1.25538990e-01
-1.55400544e-01 7.69501567e-01 1.19292498e+00 -9.11539972e-01
-4.59539652e-01 -2.09819123e-01 1.13023259e-02 4.28606234e-02
-8.19788098e-01 -1.07190199e-01 -5.09811819e-01 -6.65701151e-01
-3.79468441e-01 -8.27096343e-01 -2.63199955e-01 -2.16964260e-01
-2.46602762e-03 -4.56692696e-01 7.00932980e-01 -3.04117173e-01
1.47071958e+00 -2.28042912e+00 1.98821902e-01 -1.40002728e-01
1.20913759e-01 7.16479540e-01 -3.37292016e-01 3.92989218e-01
1.03036903e-01 7.47032557e-03 -4.36645299e-01 -5.59487045e-01
4.96470323e-03 3.05580974e-01 -3.27626973e-01 -2.43377946e-02
3.06564331e-01 1.08716786e+00 -1.05203319e+00 -8.64181370e-02
1.81900054e-01 5.06537497e-01 -8.12469780e-01 1.49371088e-01
-3.49257976e-01 1.74468234e-01 -2.13753596e-01 2.89544344e-01
1.91521108e-01 -4.37249124e-01 7.76076689e-02 -6.08335920e-02
-1.20254606e-01 1.86147675e-01 -9.94385004e-01 1.84965622e+00
-8.79383385e-01 2.96765774e-01 -3.79763395e-01 -1.17157900e+00
8.52877796e-01 2.17720762e-01 -1.30618289e-01 -6.34454668e-01
1.77978694e-01 2.89464116e-01 -1.53344557e-01 -3.29643786e-01
3.52137625e-01 -1.31691188e-01 -9.61757749e-02 3.56270105e-01
5.22282183e-01 1.13884345e-01 4.41769212e-01 2.02022687e-01
1.25237250e+00 -6.95486069e-02 5.91590345e-01 -2.40532741e-01
5.73472202e-01 -2.16482341e-01 4.11365122e-01 9.37421203e-01
-1.28965288e-01 1.44386038e-01 2.55531315e-02 -4.71312642e-01
-1.03420711e+00 -9.39931273e-01 -2.88600288e-03 1.72801316e+00
-3.50177437e-01 -4.90651548e-01 -6.29568279e-01 -7.36608028e-01
-3.78778763e-02 1.04040611e+00 -7.83890009e-01 -3.41003895e-01
-6.05383098e-01 -6.97354853e-01 2.65400447e-02 5.87938368e-01
4.22298282e-01 -1.15804327e+00 -6.10528469e-01 4.12635535e-01
1.95222244e-01 -1.08742690e+00 -5.93322098e-01 5.78573167e-01
-1.02873039e+00 -4.47911024e-01 -8.32943618e-01 -8.18932533e-01
5.20423114e-01 3.62074614e-01 1.03545451e+00 5.94439209e-02
-1.84024483e-01 1.53757244e-01 -3.68207663e-01 -1.03461139e-01
-8.64859670e-02 4.07904804e-01 2.21695110e-01 1.70309186e-01
3.56559575e-01 -6.55863225e-01 -4.98705059e-01 1.96904749e-01
-7.82422125e-01 6.08376972e-03 6.45007610e-01 1.22873378e+00
6.80432677e-01 -1.99138686e-01 7.91990578e-01 -1.10806143e+00
6.50180876e-01 -4.42538261e-01 -4.83060956e-01 2.40035951e-01
-8.00993681e-01 2.92302400e-01 1.02588403e+00 -6.11302555e-01
-1.08351254e+00 7.89017528e-02 -1.27502456e-01 -5.19808233e-01
-4.58316393e-02 5.28726637e-01 1.78905293e-01 -3.94249670e-02
9.84310389e-01 3.31355542e-01 -3.95134211e-01 -8.49316955e-01
5.64170361e-01 5.14730930e-01 4.84575480e-01 -4.76710767e-01
6.74512565e-01 3.47451985e-01 -4.92004722e-01 -8.03801000e-01
-1.11098027e+00 -6.12295628e-01 -6.66532815e-01 6.20530285e-02
4.99279946e-01 -8.02200079e-01 -2.77628452e-01 1.35304496e-01
-8.00859392e-01 -6.71653688e-01 -4.30527091e-01 3.27194780e-01
-3.64664465e-01 7.43150562e-02 -5.15240848e-01 -5.11643231e-01
-5.35647810e-01 -9.92679477e-01 8.57253551e-01 1.57823823e-02
-1.82700008e-01 -8.96167397e-01 1.68844551e-01 1.59242168e-01
6.99704409e-01 -2.92796463e-01 1.01872849e+00 -8.66664767e-01
-2.13143736e-01 -1.99328169e-01 -1.12939000e-01 3.60827804e-01
-1.01621322e-01 -4.01818633e-01 -1.20146906e+00 -5.84848106e-01
1.16772242e-01 -4.14776683e-01 1.25963211e+00 1.40111119e-01
1.27156973e+00 -1.54654548e-01 -3.10946614e-01 5.31087339e-01
1.40821040e+00 1.12413876e-01 3.03019762e-01 2.91331083e-01
5.01046717e-01 2.07828820e-01 4.68099415e-01 4.35635239e-01
2.47898027e-02 8.46554637e-01 5.04901707e-02 2.41535917e-01
-4.83850166e-02 -2.00179458e-01 4.17205453e-01 1.09060597e+00
-1.39872544e-02 -5.11203753e-03 -9.18742359e-01 6.81948006e-01
-1.84144521e+00 -8.96306336e-01 5.82714379e-01 2.43090725e+00
1.04739404e+00 4.63038981e-01 -2.76362034e-03 1.54456809e-01
4.71915752e-01 1.30847186e-01 -7.05995679e-01 -5.68450034e-01
2.07911775e-01 7.03982294e-01 -1.89144798e-02 6.19934738e-01
-9.62696552e-01 1.32055235e+00 5.78187704e+00 1.08553660e+00
-1.28607643e+00 4.88075227e-01 5.89463890e-01 -4.30683613e-01
-1.40589178e-01 -9.09580737e-02 -1.04456258e+00 3.18196505e-01
1.24108398e+00 -2.06472889e-01 7.91004121e-01 1.06617200e+00
2.48725824e-02 4.25105840e-02 -1.17141211e+00 9.57895219e-01
3.54647338e-02 -1.48322403e+00 8.55151936e-02 -3.72311115e-01
7.36190557e-01 3.83933306e-01 -4.45320904e-02 9.84063923e-01
-6.81975111e-02 -5.71410418e-01 5.12849212e-01 2.85543889e-01
9.08324122e-01 -6.28298461e-01 4.57993776e-01 5.82218528e-01
-1.05344725e+00 -3.88178617e-01 -6.63667381e-01 -1.82363614e-01
-7.37114027e-02 7.00006187e-01 -9.26348388e-01 2.39938855e-01
6.63869262e-01 6.20254338e-01 -7.22650886e-01 8.93945098e-01
-2.19792128e-01 7.92702258e-01 -2.37134144e-01 -2.53764004e-01
3.91602457e-01 2.03332141e-01 5.15057445e-01 1.31246328e+00
3.33650678e-01 -3.03846914e-02 1.70368865e-01 6.85798764e-01
-2.93193102e-01 3.59801769e-01 -4.16298687e-01 6.53978586e-02
5.21028996e-01 1.33392382e+00 -5.20512283e-01 -7.11860657e-01
-4.96892273e-01 1.06517398e+00 8.48685086e-01 3.95382464e-01
-6.08446181e-01 -3.74343842e-01 3.06142896e-01 1.18952066e-01
6.52695894e-01 -1.65335953e-01 -1.21795610e-02 -1.31829798e+00
-1.03770616e-02 -6.57177329e-01 2.65329003e-01 -5.63961864e-01
-1.31985462e+00 9.46189463e-01 -1.59136400e-01 -8.79750133e-01
-3.68733913e-01 -5.78661740e-01 -6.21476829e-01 5.83012044e-01
-1.62593353e+00 -5.18961310e-01 -2.91323245e-01 6.13857210e-01
8.93004715e-01 -3.20949614e-01 9.98311102e-01 2.22761109e-01
-7.40488112e-01 7.26883411e-01 1.90547943e-01 -2.28363171e-01
7.81381786e-01 -1.15286589e+00 6.78763092e-01 7.21555829e-01
2.34892532e-01 6.67604327e-01 6.18057191e-01 -4.29247379e-01
-1.13861406e+00 -1.18329632e+00 1.23821414e+00 -3.00991863e-01
7.65604615e-01 -6.47817671e-01 -1.14071333e+00 5.50078809e-01
2.01795995e-02 3.23109090e-01 5.67959666e-01 2.45352745e-01
-4.43812907e-01 -3.22667718e-01 -8.95515203e-01 6.35063648e-01
1.17132747e+00 -5.10742426e-01 -8.84372175e-01 3.14161688e-01
1.07176304e+00 -5.82751632e-02 -5.93943775e-01 2.87013918e-01
3.32535952e-01 -8.74443531e-01 9.44478273e-01 -6.12412214e-01
1.58094600e-01 -2.95808911e-03 -2.50033796e-01 -1.46358120e+00
-5.70188820e-01 -6.75304353e-01 -3.63740027e-01 9.12946641e-01
6.35204911e-01 -7.30460703e-01 5.65646052e-01 3.55591714e-01
-2.81585753e-01 -1.03891575e+00 -8.78239095e-01 -1.09260798e+00
-3.79608274e-02 -5.97654581e-01 4.35973525e-01 9.04779255e-01
7.87045062e-02 7.34410167e-01 -4.00826961e-01 -3.18108559e-01
3.90110046e-01 6.10371307e-02 6.30615115e-01 -1.13518417e+00
-6.29810393e-01 -3.03598523e-01 -3.28881852e-02 -9.29922462e-01
1.39244661e-01 -1.11582196e+00 2.61964439e-03 -1.13578928e+00
2.41303250e-01 -3.85616511e-01 -7.39324749e-01 8.34774017e-01
-4.44543421e-01 1.75483301e-01 3.74036342e-01 2.27688268e-01
-7.02411652e-01 6.77346528e-01 9.65233803e-01 -1.08822957e-01
-3.81048232e-01 2.18474120e-02 -5.87434888e-01 5.89973211e-01
7.72850454e-01 -5.26853919e-01 -4.20786291e-01 -4.48289126e-01
1.06456943e-01 -2.06619322e-01 1.26625270e-01 -1.17683637e+00
2.00218037e-01 1.81821242e-01 3.84875774e-01 -1.51578352e-01
4.19108242e-01 -5.90515137e-01 -2.02092960e-01 4.88776445e-01
-4.07228202e-01 -5.82650006e-02 3.08221191e-01 6.25411630e-01
-4.50275168e-02 -3.71218652e-01 8.02907467e-01 -9.75201502e-02
-1.12243319e+00 3.77537549e-01 -1.71820477e-01 1.55539915e-01
8.08506429e-01 1.78141277e-02 -2.45208085e-01 -4.96264249e-02
-9.36056674e-01 7.29620680e-02 1.98643863e-01 4.26282972e-01
4.93171394e-01 -1.23985064e+00 -5.12010932e-01 3.65993261e-01
3.16038251e-01 -3.74288231e-01 2.69796908e-01 7.58201718e-01
-8.29952955e-03 5.81083238e-01 -1.77122131e-02 -4.48287845e-01
-9.56085920e-01 8.91874015e-01 1.18425362e-01 -4.96086001e-01
-6.66459262e-01 1.02471244e+00 2.34514415e-01 -5.65506637e-01
2.32731968e-01 -2.56577373e-01 -7.09001049e-02 9.76682603e-02
7.24751830e-01 2.33250260e-01 3.23368698e-01 -1.28075600e-01
-3.28203082e-01 3.04510564e-01 -4.54816133e-01 -2.48876829e-02
1.46089637e+00 1.21317618e-01 3.57735723e-01 7.09543526e-01
1.15091169e+00 -2.88184136e-01 -1.38144374e+00 -6.93837702e-01
9.96378288e-02 -3.25350940e-01 2.87379801e-01 -9.76304650e-01
-8.35346162e-01 1.07874572e+00 5.20261765e-01 -8.06823315e-04
1.15086377e+00 -5.37312739e-02 8.81182134e-01 7.55028546e-01
5.48469365e-01 -1.18606150e+00 3.10520738e-01 7.25043178e-01
8.02367449e-01 -1.20953858e+00 -2.63974160e-01 3.32346410e-02
-6.80147827e-01 8.39013755e-01 5.64766049e-01 -2.10609034e-01
6.74194694e-01 1.03538223e-01 -2.50882030e-01 1.04111165e-01
-1.26615381e+00 -2.39047825e-01 3.13946456e-01 3.31009239e-01
3.01170588e-01 -2.99203396e-02 -2.49703437e-01 8.06522429e-01
-9.37294513e-02 2.38768712e-01 5.30368909e-02 8.43073964e-01
-7.20345914e-01 -1.03284431e+00 1.17394388e-01 6.10215962e-01
-1.14271685e-01 -5.12976646e-01 -7.95320496e-02 6.58600092e-01
-2.48031970e-02 5.97956359e-01 -3.34496498e-02 -4.05784309e-01
4.38509673e-01 4.86570776e-01 3.50430667e-01 -9.07737017e-01
-7.83842325e-01 1.07371807e-01 -1.62815154e-01 -7.15373397e-01
-1.92548763e-02 -4.76115704e-01 -1.05755413e+00 -8.60228762e-02
-2.71378428e-01 2.20772736e-02 4.80076671e-01 1.11001563e+00
5.78580260e-01 6.11697674e-01 6.10332370e-01 -1.09399295e+00
-7.60385573e-01 -1.03414357e+00 -3.03111911e-01 2.88862020e-01
1.20918192e-01 -7.21682787e-01 -4.43484604e-01 -1.66967943e-01] | [9.892708778381348, 3.1234006881713867] |
33295f23-812b-404b-a4f0-f0377cdfbf81 | a-sequential-concept-drift-detection-method | 2212.09637 | null | https://arxiv.org/abs/2212.09637v2 | https://arxiv.org/pdf/2212.09637v2.pdf | A Sequential Concept Drift Detection Method for On-Device Learning on Low-End Edge Devices | A practical issue of edge AI systems is that data distributions of trained dataset and deployed environment may differ due to noise and environmental changes over time. Such a phenomenon is known as a concept drift, and this gap degrades the performance of edge AI systems and may introduce system failures. To address this gap, retraining of neural network models triggered by concept drift detection is a practical approach. However, since available compute resources are strictly limited in edge devices, in this paper we propose a fully sequential concept drift detection method in cooperation with an on-device sequential learning technique of neural networks. In this case, both the neural network retraining and the proposed concept drift detection are done only by sequential computation to reduce computation cost and memory utilization. Evaluation results of the proposed approach shows that while the accuracy is decreased by 3.8%-4.3% compared to existing batch-based detection methods, it decreases the memory size by 88.9%-96.4% and the execution time by 1.3%-83.8%. As a result, the combination of the neural network retraining and the proposed concept drift detection method is demonstrated on Raspberry Pi Pico that has 264kB memory. | ['Hiroki Matsutani', 'Takeya Yamada'] | 2022-12-19 | null | null | null | null | ['pico'] | ['natural-language-processing'] | [ 2.10854083e-01 -2.74320126e-01 -2.45857816e-02 -1.72644593e-02
2.21962094e-01 -3.96237612e-01 4.02436741e-02 3.14298242e-01
-6.61248863e-01 1.03513551e+00 -6.80397987e-01 -1.13894112e-01
-4.48205620e-02 -7.04916120e-01 -5.88018477e-01 -8.83266449e-01
1.41995534e-01 3.81773740e-01 5.47499716e-01 1.00267209e-01
2.59615719e-01 5.71859479e-01 -1.84740102e+00 2.62441814e-01
8.09333324e-01 1.20721579e+00 -6.74441010e-02 6.35948420e-01
-2.63473839e-01 6.88888848e-01 -1.03323698e+00 4.20150191e-01
1.60403773e-01 -1.76993042e-01 -8.36578459e-02 -2.72295684e-01
-1.89986646e-01 -2.84862757e-01 -3.74357104e-02 1.02768612e+00
5.95261335e-01 3.00374404e-02 4.56055790e-01 -1.41500342e+00
-4.65449728e-02 5.18587112e-01 -7.97068298e-01 5.14499784e-01
-1.22971117e-01 -1.69720963e-01 -7.81191066e-02 -5.44721603e-01
2.12952003e-01 5.25318146e-01 8.35890949e-01 7.04690397e-01
-7.74059653e-01 -9.73448038e-01 6.79586232e-02 2.48151988e-01
-1.66465962e+00 -3.88997167e-01 8.73957515e-01 -2.42190167e-01
1.10015464e+00 1.58163503e-01 5.21572113e-01 7.09021449e-01
4.29919720e-01 4.23672348e-01 9.01438057e-01 -4.83810991e-01
6.39115334e-01 4.16290224e-01 3.36315453e-01 1.41547397e-01
9.89758015e-01 7.98091367e-02 -5.24835825e-01 -1.39358386e-01
4.52633083e-01 2.91059524e-01 -9.47305262e-02 1.32437542e-01
-4.73945886e-01 2.13080272e-03 5.04436530e-02 5.76885521e-01
-5.68192780e-01 1.76728308e-01 6.75071239e-01 1.89058445e-02
4.51832741e-01 -1.72047645e-01 -5.54971695e-01 -5.10364890e-01
-9.72209871e-01 -1.36912372e-02 6.37325346e-01 9.62691069e-01
4.08074975e-01 5.17961502e-01 2.99977839e-01 4.73947227e-01
3.03186029e-02 4.49420959e-01 7.81396329e-01 2.23185699e-02
2.06993818e-01 8.55227649e-01 3.14726114e-01 -8.55362415e-01
-7.34363794e-01 -4.96316016e-01 -8.78711104e-01 4.42581773e-01
2.81163663e-01 -3.59987944e-01 -9.67037380e-01 1.29690588e+00
5.36050737e-01 4.29430246e-01 3.01775962e-01 6.73663855e-01
4.99849975e-01 6.33183002e-01 1.50371730e-01 -5.66349387e-01
1.05680490e+00 -5.04872203e-01 -9.87467051e-01 -6.73660338e-02
4.69431430e-01 -5.90061903e-01 6.26785994e-01 6.10394895e-01
-9.58562374e-01 -7.96479881e-01 -1.61284173e+00 7.34473825e-01
-3.10454339e-01 2.60418087e-01 2.62773633e-01 1.04082751e+00
-7.12353587e-01 4.41935509e-01 -1.17805707e+00 -4.63453084e-01
3.07922453e-01 6.86945677e-01 2.56441027e-01 7.83616826e-02
-8.94744992e-01 5.23238957e-01 8.35794091e-01 2.81513661e-01
-4.83424336e-01 -6.01525724e-01 -1.51761681e-01 5.78677133e-02
2.15730608e-01 -2.89171904e-01 1.07978940e+00 -1.23937857e+00
-1.60971558e+00 1.71509564e-01 5.36912158e-02 -6.02874458e-01
5.48100233e-01 -3.83788526e-01 -8.32475364e-01 -1.04588479e-01
-4.32568014e-01 1.87086836e-02 8.01204979e-01 -9.46756601e-01
-9.97772217e-01 -4.86930281e-01 -4.29895967e-01 1.67574912e-01
-6.34384215e-01 -1.47122681e-01 -1.29280791e-01 -3.13308924e-01
1.19600207e-01 -9.07662868e-01 3.71887535e-02 -3.85668695e-01
9.62354615e-02 -1.12271398e-01 1.24004078e+00 -4.08199489e-01
1.46412790e+00 -2.11854434e+00 -6.35081708e-01 1.09748349e-01
-1.17837146e-01 5.62779307e-01 6.00417554e-01 1.08383745e-01
5.32938354e-02 -2.13799238e-01 -3.69957872e-02 6.89127669e-02
-6.35009050e-01 9.03227255e-02 3.86374854e-02 2.48741820e-01
5.64142428e-02 1.29020199e-01 -6.54385388e-01 -9.93757844e-02
4.23418991e-02 6.18807018e-01 -1.38459563e-01 -3.81021611e-02
5.41156083e-02 1.61217585e-01 -4.53348786e-01 9.40650582e-01
8.97924483e-01 1.79189608e-01 2.38759577e-01 -1.16667368e-01
-2.06468835e-01 -1.11403696e-01 -1.41343617e+00 1.34305429e+00
-3.21322590e-01 5.59981883e-01 -1.18559681e-01 -1.04838502e+00
1.24738336e+00 5.90469003e-01 6.16346061e-01 -7.72709846e-01
3.62139225e-01 3.40017229e-01 7.84881040e-02 -4.69255239e-01
5.18679619e-01 9.84331891e-02 2.58757234e-01 2.55275249e-01
-3.75386924e-01 5.42155445e-01 -1.34037122e-01 -1.99479461e-01
1.21696436e+00 2.97097385e-01 2.31724888e-01 -4.04001117e-01
5.24953723e-01 2.29779720e-01 1.05732596e+00 5.97641110e-01
-4.05607700e-01 5.25184087e-02 -5.30121960e-02 -6.41685665e-01
-9.23672020e-01 -5.15374184e-01 -1.67217493e-01 6.08845592e-01
2.56008714e-01 -2.15731338e-01 -7.69300044e-01 -4.89788413e-01
-2.14037225e-02 6.01082623e-01 -2.66030580e-01 -5.25209188e-01
-5.82616985e-01 -1.19313657e+00 5.69910586e-01 8.35389495e-01
7.58776784e-01 -9.29472148e-01 -1.30326176e+00 5.93981266e-01
5.41168869e-01 -9.52167511e-01 2.03253001e-01 2.44492754e-01
-1.33096504e+00 -9.15912509e-01 -2.38798797e-01 -5.64077139e-01
7.91439354e-01 1.46241337e-01 8.11395049e-01 3.96320581e-01
-4.28558916e-01 7.15890899e-02 -4.26578194e-01 -9.05149758e-01
-3.91028523e-01 8.05611489e-04 4.23700243e-01 2.03755009e-03
6.55189037e-01 -5.95371902e-01 -7.62669444e-01 1.56062976e-01
-8.80745649e-01 -2.06511185e-01 4.84167695e-01 7.16401279e-01
3.81415516e-01 7.29495704e-01 9.57266748e-01 -9.65547085e-01
5.98174691e-01 -4.20750052e-01 -1.09615147e+00 1.76593214e-01
-1.35525382e+00 -1.93131372e-01 6.39480412e-01 -7.65427053e-01
-1.24130034e+00 3.30474496e-01 3.68709177e-01 -2.87515521e-01
-1.58660769e-01 3.14030081e-01 -1.52365953e-01 5.65949753e-02
8.69467318e-01 4.93430048e-02 -1.77816957e-01 -2.96179563e-01
-4.02734816e-01 1.12756407e+00 4.05254215e-01 -1.62710115e-01
3.83744240e-01 3.89364034e-01 5.22958189e-02 -7.64176488e-01
1.06763713e-01 -2.95009404e-01 -2.01200157e-01 -4.73310500e-01
4.25123453e-01 -9.37414408e-01 -8.52765381e-01 7.25143373e-01
-9.69435453e-01 -6.01987764e-02 8.38952232e-03 5.16562045e-01
8.50489810e-02 -1.14225186e-02 -4.97866780e-01 -1.29459834e+00
-9.33717906e-01 -7.90900767e-01 3.73295009e-01 7.17046201e-01
-1.25043169e-01 -5.73306799e-01 -2.72737592e-01 -1.07782051e-01
7.70387888e-01 2.25641847e-01 3.95608783e-01 -5.92012107e-01
-3.86193424e-01 -6.74277604e-01 -2.01026872e-02 2.90324986e-01
2.87313372e-01 1.08647928e-01 -9.38319087e-01 -3.35914254e-01
4.50304359e-01 3.55677485e-01 4.87908632e-01 3.34985316e-01
7.65937448e-01 4.12032045e-02 -6.59478009e-01 1.37799770e-01
1.73131347e+00 1.16243470e+00 5.97177744e-01 3.66408050e-01
6.24251008e-01 8.96386206e-02 7.40665019e-01 7.18482375e-01
-1.63005650e-01 4.51990604e-01 2.62730956e-01 2.14512572e-01
-4.37916443e-03 1.71674863e-01 3.88263166e-01 6.79585874e-01
-1.72917880e-02 -3.39088291e-01 -1.17604220e+00 5.25364697e-01
-1.96089697e+00 -6.38203561e-01 -2.81289428e-01 2.51806951e+00
6.51402593e-01 6.78958595e-01 9.08012316e-03 5.76619208e-01
9.63742077e-01 -5.22093713e-01 -9.65215445e-01 -3.93326432e-01
1.45594895e-01 4.67445329e-02 8.05677474e-01 2.81594545e-01
-6.86737061e-01 6.00436211e-01 5.27833843e+00 4.40483063e-01
-1.51505685e+00 1.39781833e-01 4.85998631e-01 -4.65475827e-01
3.74837250e-01 -4.09950353e-02 -9.97829497e-01 9.17305052e-01
1.13336134e+00 -1.05363145e-01 6.29783645e-02 1.01086390e+00
5.42880967e-02 -5.05975246e-01 -7.78874695e-01 1.33408332e+00
1.26384452e-01 -9.95111346e-01 -4.26177382e-01 -3.07358652e-01
6.00855529e-01 -1.85028732e-01 -3.72675776e-01 1.92647651e-01
-2.88303822e-01 -4.68521893e-01 4.99653935e-01 4.87590283e-01
4.70281482e-01 -9.45062935e-01 1.12334955e+00 4.01941478e-01
-1.00761378e+00 -2.27851614e-01 -3.86096090e-01 -3.14050704e-01
-6.16125166e-02 1.00126636e+00 -9.64979649e-01 3.42214882e-01
1.13582802e+00 2.11928442e-01 -2.79333711e-01 1.06720722e+00
1.75736547e-01 8.35885882e-01 -6.84359312e-01 -3.41375738e-01
-3.74175847e-01 7.07848370e-02 4.77467418e-01 8.89092743e-01
4.56491292e-01 4.99400236e-02 -1.70651838e-01 5.34823954e-01
9.03556570e-02 -4.79874834e-02 -4.62710351e-01 9.36203301e-02
8.11998129e-01 1.07089615e+00 -1.01633334e+00 -4.73589808e-01
-1.21585518e-01 9.57039893e-01 -2.26001263e-01 3.18351328e-01
-9.66306746e-01 -6.11699224e-01 1.69814244e-01 2.24385291e-01
1.77192986e-01 -1.07929170e-01 -4.99611169e-01 -6.79256976e-01
3.09507459e-01 -3.87044191e-01 3.74578446e-01 -5.90435803e-01
-7.61492431e-01 8.49174976e-01 -1.15893908e-01 -1.17924869e+00
-2.10105211e-01 -3.35839123e-01 -5.15865803e-01 6.21385694e-01
-1.10654902e+00 -5.65868318e-01 -7.22589970e-01 2.37773538e-01
4.78312969e-01 -3.39615434e-01 5.72347105e-01 5.90570033e-01
-9.71602678e-01 6.64801002e-01 1.83936268e-01 -3.88682306e-01
6.33753419e-01 -7.79181004e-01 6.41157031e-02 9.53220189e-01
-3.60003054e-01 2.70491898e-01 8.41620564e-01 -1.07898462e+00
-1.22461987e+00 -9.22321081e-01 5.09725034e-01 2.99915504e-02
1.58913031e-01 -3.06757241e-01 -1.18902516e+00 4.37708914e-01
1.58000305e-01 2.72523671e-01 4.42439705e-01 -1.64732695e-01
1.98461995e-01 -7.65582561e-01 -1.28660321e+00 5.09546757e-01
7.18144238e-01 -4.29221541e-02 -8.25995728e-02 4.46614511e-02
2.96268612e-01 -5.68326116e-01 -7.32601106e-01 6.16783381e-01
5.52958310e-01 -8.02343547e-01 3.61080140e-01 -1.66833326e-01
-3.73182327e-01 -7.01843262e-01 3.64023782e-02 -7.70857453e-01
-6.43211305e-02 -5.88764906e-01 -2.94798642e-01 1.45866799e+00
4.02609617e-01 -7.23755002e-01 1.18232155e+00 9.21033919e-01
-3.08537371e-02 -5.42372227e-01 -1.01780653e+00 -7.95752883e-01
-4.80521739e-01 -3.69683027e-01 5.90746999e-01 6.98118389e-01
8.46238658e-02 2.35375568e-01 -1.52004912e-01 2.61024952e-01
3.18141103e-01 -3.77346665e-01 4.53740239e-01 -1.26600385e+00
-1.19450107e-01 -2.29057856e-02 -6.34200215e-01 -3.27071279e-01
-4.86908376e-01 -3.15314144e-01 2.22163610e-02 -1.02268922e+00
-1.73816681e-01 -5.91551900e-01 -9.26713109e-01 2.44444638e-01
-5.94308116e-02 -6.19735047e-02 -2.13491067e-01 3.23375046e-01
-3.06979269e-01 2.78329611e-01 2.59570807e-01 7.44071677e-02
-7.47768342e-01 1.63117841e-01 -1.63332552e-01 5.62631369e-01
1.29679477e+00 -8.74877393e-01 -7.91523397e-01 -3.22738647e-01
3.30868274e-01 -1.17767572e-01 -7.06389919e-02 -1.69864154e+00
7.73218155e-01 1.16460592e-01 6.78010285e-01 -6.55268431e-01
3.00918589e-03 -1.25486672e+00 4.51484531e-01 8.67985725e-01
4.22466606e-01 3.29212308e-01 7.87564516e-01 6.42090917e-01
-3.87867466e-02 -2.88561970e-01 3.93710852e-01 2.01765209e-01
-6.21064365e-01 -1.45522788e-01 -4.62393880e-01 -6.03110731e-01
1.41798377e+00 -6.88430011e-01 -2.27832049e-01 1.39813021e-01
-3.62779140e-01 -4.27259579e-02 2.56129801e-01 4.56791431e-01
4.61988419e-01 -1.12670815e+00 -3.07289839e-01 3.09280068e-01
-7.26972520e-02 1.19689234e-01 5.35577953e-01 7.33662665e-01
-6.19821489e-01 2.64100552e-01 -4.07523900e-01 -6.13077819e-01
-1.36465764e+00 6.06928229e-01 3.24926078e-01 1.41581669e-01
-5.18915892e-01 5.36954880e-01 -4.82508272e-01 3.98500860e-01
4.24558073e-01 -3.81133974e-01 -2.78056383e-01 -1.46771356e-01
6.30244672e-01 6.17045403e-01 4.58825707e-01 5.05921096e-02
-7.20272124e-01 3.51921886e-01 -3.69014531e-01 -1.92610025e-02
1.03977156e+00 2.58755311e-02 8.01189244e-02 5.93290985e-01
4.88155723e-01 -6.92104474e-02 -1.22782874e+00 -9.53866243e-02
2.59244412e-01 -3.30612157e-03 1.67800501e-01 -8.96001816e-01
-9.67674732e-01 3.86507392e-01 1.30296218e+00 1.62051365e-01
1.60632050e+00 -7.01535523e-01 6.65616155e-01 3.01627934e-01
2.91086912e-01 -1.57487285e+00 -2.97134936e-01 5.35695553e-02
4.59831417e-01 -1.00581157e+00 1.47895619e-01 -1.44116357e-01
-3.94815087e-01 1.13382947e+00 1.04277897e+00 -7.10923895e-02
8.03898036e-01 6.86720967e-01 1.19884498e-01 1.33824065e-01
-8.48116219e-01 2.47569561e-01 -2.95110583e-01 4.37322617e-01
2.65484691e-01 1.38579840e-02 -6.92985952e-01 9.11881506e-01
3.08633626e-01 5.84469557e-01 5.94016492e-01 1.50406766e+00
-3.89689624e-01 -1.07140768e+00 -4.81407702e-01 3.16306263e-01
-5.99024653e-01 2.31997848e-01 -4.82682846e-02 7.64336884e-01
2.92916745e-01 1.05269539e+00 4.14394885e-01 -7.38315284e-01
3.94155294e-01 2.44589865e-01 1.92033678e-01 -1.70625508e-01
-6.85853899e-01 9.67053026e-02 -2.82659549e-02 -1.70535401e-01
-2.60291845e-01 -6.58886313e-01 -1.63112056e+00 -9.87910107e-02
-5.05413234e-01 1.18725091e-01 1.21082568e+00 6.95603848e-01
7.21708477e-01 9.49335277e-01 3.22862267e-01 -2.10463017e-01
-4.12057817e-01 -1.11787796e+00 -4.44420218e-01 1.22143611e-01
8.54220688e-02 -6.54662073e-01 -2.62225688e-01 -6.65707560e-03] | [7.689472198486328, 2.6403493881225586] |
7692b379-0cfb-49f6-8295-490d1aa0f474 | 190910351 | 1909.10351 | null | https://arxiv.org/abs/1909.10351v5 | https://arxiv.org/pdf/1909.10351v5.pdf | TinyBERT: Distilling BERT for Natural Language Understanding | Language model pre-training, such as BERT, has significantly improved the performances of many natural language processing tasks. However, pre-trained language models are usually computationally expensive, so it is difficult to efficiently execute them on resource-restricted devices. To accelerate inference and reduce model size while maintaining accuracy, we first propose a novel Transformer distillation method that is specially designed for knowledge distillation (KD) of the Transformer-based models. By leveraging this new KD method, the plenty of knowledge encoded in a large teacher BERT can be effectively transferred to a small student Tiny-BERT. Then, we introduce a new two-stage learning framework for TinyBERT, which performs Transformer distillation at both the pretraining and task-specific learning stages. This framework ensures that TinyBERT can capture he general-domain as well as the task-specific knowledge in BERT. TinyBERT with 4 layers is empirically effective and achieves more than 96.8% the performance of its teacher BERTBASE on GLUE benchmark, while being 7.5x smaller and 9.4x faster on inference. TinyBERT with 4 layers is also significantly better than 4-layer state-of-the-art baselines on BERT distillation, with only about 28% parameters and about 31% inference time of them. Moreover, TinyBERT with 6 layers performs on-par with its teacher BERTBASE. | ['Xin Jiang', 'Yichun Yin', 'Qun Liu', 'Fang Wang', 'Lifeng Shang', 'Xiaoqi Jiao', 'Xiao Chen', 'Linlin Li'] | 2019-09-23 | null | https://aclanthology.org/2020.findings-emnlp.372 | https://aclanthology.org/2020.findings-emnlp.372.pdf | findings-of-the-association-for-computational | ['linguistic-acceptability'] | ['natural-language-processing'] | [-3.94861162e-01 2.96371996e-01 -4.54401881e-01 -3.68308961e-01
-8.34414780e-01 -5.99639356e-01 4.13905352e-01 8.33936632e-02
-7.92147279e-01 7.50547349e-01 -1.75287947e-01 -7.34242737e-01
1.69177562e-01 -1.12513340e+00 -1.15330768e+00 -5.24880767e-01
2.86793709e-01 9.62771893e-01 4.83911902e-01 -1.91444457e-01
-5.71028113e-01 2.18439057e-01 -1.20449197e+00 3.07645738e-01
1.07216716e+00 1.00639284e+00 4.23773646e-01 7.19574749e-01
-1.09326988e-01 1.10687566e+00 -7.17087209e-01 -7.51289964e-01
-6.84323460e-02 2.61310935e-01 -8.50455284e-01 -7.23935127e-01
5.06230175e-01 -8.37769866e-01 -5.82978487e-01 7.84375429e-01
5.82381904e-01 2.31490821e-01 3.68015260e-01 -8.92275155e-01
-5.03359199e-01 1.40328908e+00 -2.66920090e-01 3.06971759e-01
-2.30026349e-01 1.71575040e-01 1.01612735e+00 -8.77062261e-01
1.01113535e-01 1.38741291e+00 4.44631219e-01 6.37405992e-01
-9.43072200e-01 -1.04759836e+00 2.01403499e-01 1.69595197e-01
-1.46045291e+00 -4.03692156e-01 3.50797266e-01 3.08570154e-02
1.27922344e+00 -4.86951917e-02 6.21674657e-01 8.36402655e-01
-1.63499549e-01 1.33658111e+00 8.15024316e-01 -3.10655534e-01
-1.56372413e-01 1.65842921e-01 1.02842636e-01 1.12412882e+00
1.22114733e-01 -5.15620224e-02 -5.52931070e-01 1.28380522e-01
7.48233140e-01 -1.50344282e-01 6.21679286e-03 5.53562529e-02
-1.08043003e+00 5.89098096e-01 7.12669790e-01 2.58953154e-01
-4.17923518e-02 5.60658395e-01 5.66473067e-01 3.37426513e-01
5.28516173e-01 3.70008260e-01 -9.59451616e-01 -3.74367237e-01
-9.18488324e-01 3.06386828e-01 9.23561156e-01 1.24263847e+00
7.90230513e-01 2.45789438e-01 -2.85956115e-01 7.95201540e-01
2.52872616e-01 8.98729563e-01 4.92654502e-01 -5.47559977e-01
8.65180373e-01 6.01509094e-01 -4.20849741e-01 -4.42698032e-01
-3.94631922e-02 -5.06395817e-01 -9.87764001e-01 -4.27478224e-01
3.73938978e-01 -2.91442722e-01 -1.02595127e+00 1.72244966e+00
4.40637290e-01 6.57759070e-01 1.49539605e-01 4.74428654e-01
1.29151988e+00 8.85468483e-01 2.18230978e-01 2.47558519e-01
1.48500228e+00 -1.35609841e+00 -3.81999761e-01 -4.61496472e-01
8.61541986e-01 -5.22706032e-01 1.19931602e+00 5.42088866e-01
-1.38588047e+00 -5.67577899e-01 -1.10778284e+00 -6.80215240e-01
-4.34339881e-01 2.52509952e-01 9.40915823e-01 3.31868768e-01
-9.88028646e-01 3.82331610e-01 -1.05706298e+00 3.50933969e-01
4.76012200e-01 5.73497474e-01 -6.48406669e-02 -2.93150872e-01
-1.35555053e+00 9.24226701e-01 7.87185192e-01 1.14867985e-02
-1.32327259e+00 -1.10290420e+00 -9.48298752e-01 3.67322505e-01
5.79452574e-01 -8.84061992e-01 1.88489556e+00 -1.75255775e-01
-1.86009526e+00 6.59871578e-01 -1.84379324e-01 -6.77841246e-01
4.84342843e-01 -6.18099749e-01 -8.81585404e-02 8.41455460e-02
-2.55816787e-01 7.38570809e-01 6.82727337e-01 -6.59862220e-01
-6.84647143e-01 -5.25553972e-02 4.84973192e-01 2.07128510e-01
-4.44333285e-01 -2.46876776e-01 -8.95254970e-01 -5.18670321e-01
-3.37434858e-01 -6.28618002e-01 -1.23418067e-02 -1.66243091e-01
-2.46322110e-01 -7.23437667e-01 7.60357738e-01 -4.68864262e-01
1.32179749e+00 -1.94726479e+00 2.82847695e-02 -9.54912305e-02
6.42305076e-01 1.00191593e+00 -7.64793456e-02 4.49767187e-02
3.59405458e-01 -6.12498932e-02 5.17139845e-02 -6.32292271e-01
2.07834050e-01 6.43767774e-01 -5.01485050e-01 -5.02065429e-03
5.25094289e-03 1.27867401e+00 -1.24362707e+00 -5.47565043e-01
1.96454272e-01 5.16199946e-01 -8.56531739e-01 3.50837320e-01
-5.55094182e-01 1.30116358e-01 -5.11951745e-01 5.60142636e-01
5.50113857e-01 -2.91650385e-01 1.83121786e-01 -7.89225399e-02
1.86244726e-01 1.01732874e+00 -6.20359302e-01 1.82787001e+00
-1.02449727e+00 5.42426348e-01 1.47654518e-01 -1.17010653e+00
7.90498614e-01 5.36030352e-01 -2.83068884e-02 -4.82358545e-01
2.66615361e-01 2.71787435e-01 -4.00635675e-02 -1.70301139e-01
2.93394476e-01 -7.41084889e-02 -8.93819258e-02 4.83344018e-01
4.05643165e-01 -5.19249737e-01 2.45367199e-01 3.75576735e-01
9.47940469e-01 -2.79471166e-02 9.17291343e-02 -5.95774651e-02
4.14234310e-01 -3.55245143e-01 5.37903845e-01 7.83840418e-01
4.21348587e-02 -1.06403716e-01 3.38415384e-01 -4.44583058e-01
-6.44143701e-01 -1.23193884e+00 3.18273269e-02 1.46427345e+00
-1.22925848e-01 -7.16738224e-01 -7.42425859e-01 -9.49620903e-01
2.35670716e-01 7.42152810e-01 -2.24777639e-01 -5.53379416e-01
-8.39768469e-01 -5.28945744e-01 1.16778362e+00 9.03586507e-01
1.19000936e+00 -8.29374135e-01 -4.10461605e-01 2.82679975e-01
-1.88563198e-01 -1.33478105e+00 -3.12916279e-01 2.88136005e-01
-8.80452573e-01 -7.72622108e-01 -3.90255094e-01 -8.61488044e-01
5.60624123e-01 1.43091038e-01 1.55367422e+00 2.10477963e-01
1.32910848e-01 -1.22969851e-01 -1.94113970e-01 -6.63062692e-01
-3.04507136e-01 4.62356299e-01 4.18887697e-02 -6.63188219e-01
6.19040549e-01 -6.08319581e-01 -3.44407380e-01 -4.32150578e-03
-7.55543351e-01 3.57526392e-01 8.26996624e-01 9.25318182e-01
5.24979234e-01 2.73270130e-01 3.54708463e-01 -1.00167966e+00
3.08481663e-01 -1.94419920e-01 -7.24970520e-01 4.56667930e-01
-5.68221152e-01 2.74421066e-01 1.01393664e+00 -5.82039952e-01
-1.03377402e+00 -3.24635595e-01 -4.37997431e-01 -4.95018780e-01
3.63384008e-01 7.06791341e-01 -1.93726346e-01 7.17761926e-03
5.02426445e-01 1.98548883e-01 -3.75366688e-01 -8.04749846e-01
5.64917386e-01 6.65320158e-01 6.95883095e-01 -1.22876036e+00
8.05378675e-01 8.33169520e-02 -4.23024148e-01 -4.87932533e-01
-1.48885679e+00 -3.80286098e-01 -5.69071114e-01 4.85988796e-01
4.47639436e-01 -1.44072366e+00 -9.55629051e-01 5.82109094e-01
-1.11956513e+00 -1.00776577e+00 -2.58514553e-01 4.41427737e-01
-1.86746031e-01 -4.21168990e-02 -9.90492523e-01 -2.70951837e-01
-8.23162675e-01 -1.31356215e+00 1.13427556e+00 1.45728707e-01
2.66404837e-01 -1.17637968e+00 -2.80846924e-01 5.94726503e-01
4.14756119e-01 -4.94844168e-01 1.03172183e+00 -6.04614794e-01
-5.61250329e-01 -4.65595871e-02 -3.66331547e-01 7.53557026e-01
-3.05447698e-01 -2.71784276e-01 -1.14379489e+00 -3.16108167e-01
-1.66805610e-01 -8.00292432e-01 1.08718610e+00 4.15185839e-02
1.48307681e+00 -4.74620998e-01 -3.89310539e-01 9.71534789e-01
1.08468831e+00 -1.67643681e-01 3.04633051e-01 1.53806746e-01
9.59982574e-01 -1.17895067e-01 4.06757444e-01 1.31561711e-01
7.84492493e-01 5.24889231e-01 2.53384322e-01 -2.84137353e-02
-2.22342297e-01 -6.39345527e-01 5.30657113e-01 1.17780781e+00
-6.18533194e-02 -1.27788454e-01 -1.11373699e+00 4.21504825e-01
-1.72202110e+00 -5.20025671e-01 3.44342917e-01 2.03915977e+00
1.60880041e+00 3.78291547e-01 -1.37793332e-01 3.37286182e-02
8.90498310e-02 2.58216448e-02 -6.30210936e-01 -5.08348584e-01
3.09750497e-01 7.97036946e-01 5.55176973e-01 6.05802894e-01
-1.04900336e+00 1.55476856e+00 5.94415236e+00 1.28777421e+00
-1.22842610e+00 2.16259331e-01 4.58525538e-01 -2.14440227e-01
-9.29219797e-02 -2.05295160e-01 -1.37975645e+00 2.22188592e-01
1.21398282e+00 -3.27734470e-01 5.30011714e-01 1.17099881e+00
-2.00757042e-01 1.43106952e-01 -1.33127034e+00 9.06400621e-01
-2.40644649e-01 -1.32632399e+00 2.32060388e-01 -2.15886176e-01
7.25351572e-01 3.45658869e-01 1.11925431e-01 1.22867489e+00
8.67477894e-01 -1.09050143e+00 5.47557056e-01 2.58845184e-02
9.78720546e-01 -9.18784380e-01 8.03095996e-01 8.44736993e-01
-1.33010960e+00 1.26978919e-01 -6.56765521e-01 -2.67574698e-01
-1.31524444e-01 7.97064483e-01 -1.20365906e+00 3.34766418e-01
6.61149085e-01 7.22395957e-01 -4.15797770e-01 6.80422544e-01
-1.04245985e+00 1.04866850e+00 -5.84578991e-01 -7.25434348e-02
4.61180836e-01 1.11536615e-01 4.22970988e-02 1.29216981e+00
-2.77868323e-02 1.74765319e-01 3.67889851e-01 7.51552701e-01
-6.91777050e-01 -3.69722843e-01 -3.26510787e-01 -1.86123103e-01
9.04670596e-01 1.22931337e+00 -7.49294832e-02 -1.02252507e+00
-4.49852437e-01 7.97809780e-01 9.37222004e-01 2.21494436e-01
-1.09793794e+00 -3.92882854e-01 6.53181195e-01 -1.93999588e-01
3.01690817e-01 -3.32033098e-01 -1.08629568e-02 -1.30752659e+00
-1.57150969e-01 -1.03517568e+00 4.23254073e-01 -6.08849347e-01
-9.22484457e-01 3.82733524e-01 1.87131748e-01 -5.37672877e-01
-2.87340969e-01 -6.86162949e-01 -5.12628615e-01 8.57508242e-01
-1.84002924e+00 -1.38203812e+00 -1.11332208e-01 7.26970196e-01
4.02161896e-01 -4.76280898e-02 9.23283279e-01 5.35117686e-01
-7.09612846e-01 1.15862858e+00 1.71987057e-01 4.53226119e-01
6.43751562e-01 -1.57310486e+00 5.53284466e-01 5.93791068e-01
1.97519347e-01 9.47559714e-01 8.47908258e-02 -2.58451462e-01
-1.59182513e+00 -1.30784869e+00 1.16192114e+00 -4.76978838e-01
8.19457769e-01 -6.18559837e-01 -7.90080249e-01 1.09592593e+00
-1.47668809e-01 2.20714658e-01 3.96266997e-01 4.06350493e-01
-6.68708324e-01 -3.80785882e-01 -9.48128998e-01 5.64483285e-01
9.30809259e-01 -7.88614690e-01 -8.46133828e-01 4.71994311e-01
1.12131751e+00 -9.40785885e-01 -9.47121084e-01 5.04795909e-01
4.26785618e-01 -5.07254660e-01 1.21088004e+00 -4.78720933e-01
4.30146873e-01 6.71546087e-02 1.63869426e-01 -1.33559358e+00
-1.54673830e-01 -5.60377121e-01 -7.24804997e-01 1.13279164e+00
3.62465709e-01 -6.36760890e-01 8.42225552e-01 4.54372525e-01
-3.56415600e-01 -1.07976079e+00 -7.20180333e-01 -8.35186481e-01
4.20404822e-01 -6.28291965e-01 8.34180534e-01 7.10995078e-01
-2.68334383e-03 7.34190643e-01 -7.98592046e-02 2.23642066e-02
1.90723971e-01 1.04662567e-01 1.00935733e+00 -9.69555378e-01
-5.80489516e-01 -2.95434535e-01 1.14183808e-02 -1.92595720e+00
4.45823252e-01 -1.22079694e+00 3.62834185e-02 -1.46573472e+00
1.38874918e-01 -9.32168245e-01 -3.61399412e-01 1.08048356e+00
-3.94973636e-01 7.19528347e-02 5.39823845e-02 -1.78709239e-01
-5.69006324e-01 6.53707027e-01 1.68116140e+00 -4.25971806e-01
-5.06658256e-02 4.94813360e-02 -5.56335032e-01 8.84066284e-01
6.18474424e-01 -2.92639911e-01 -7.45739460e-01 -1.10900831e+00
2.05415279e-01 -2.41503537e-01 2.28409305e-01 -8.88250470e-01
3.72563839e-01 8.56428444e-02 8.55006650e-02 -5.25975883e-01
3.93472850e-01 -6.86193943e-01 -5.68311512e-01 4.05332476e-01
-2.08670095e-01 -1.61092222e-01 6.14385307e-01 1.33704275e-01
-2.70800084e-01 -1.96855247e-01 6.63025081e-01 -3.24304283e-01
-7.02678382e-01 6.03267491e-01 -3.77330519e-02 3.58001500e-01
6.37768626e-01 5.47727764e-01 -4.27429348e-01 -6.64472729e-02
-3.60681057e-01 5.45826852e-01 -2.69621044e-01 1.14426121e-01
4.64815825e-01 -9.91178811e-01 -5.79204202e-01 2.47743547e-01
-3.01293701e-01 1.11170197e+00 3.21148895e-03 7.29336858e-01
-6.85113728e-01 7.14012980e-01 3.55380356e-01 -4.65588123e-01
-1.12135482e+00 4.29500550e-01 2.73327738e-01 -8.74515176e-01
-7.06402004e-01 1.56792140e+00 3.24483126e-01 -8.36899400e-01
4.47222441e-01 -9.48383808e-01 1.79453775e-01 -3.03973705e-01
7.77596354e-01 2.43960336e-01 1.74217746e-01 1.22468909e-02
-1.91587135e-01 1.38578385e-01 -4.36031073e-01 1.31624952e-01
1.16244817e+00 4.41213429e-01 -9.51882079e-02 2.04130113e-01
9.71401095e-01 -2.23948639e-02 -1.12810194e+00 -6.97924256e-01
-2.99426198e-01 -1.04850806e-01 3.71957541e-01 -1.04670024e+00
-1.18543613e+00 1.29003859e+00 -1.65681243e-01 -4.01741982e-01
1.04131365e+00 -1.33418180e-02 1.26248372e+00 1.01373112e+00
5.74339926e-01 -8.14002633e-01 2.47901361e-02 1.03708494e+00
4.57573354e-01 -1.12540865e+00 5.59477434e-02 -4.78488684e-01
-1.72919765e-01 8.33471358e-01 8.38404775e-01 1.31634418e-02
6.41853929e-01 6.39122009e-01 -1.27249628e-01 -5.13911061e-02
-1.13424599e+00 -1.35562733e-01 3.92641932e-01 3.26649934e-01
5.80227971e-01 3.66045266e-01 1.65726393e-01 8.05926740e-01
-6.78716362e-01 1.40892893e-01 -1.36775926e-01 8.42053831e-01
-4.74152327e-01 -1.09553647e+00 7.75895268e-02 3.37506533e-01
-4.25230950e-01 -5.99795103e-01 -3.42940763e-02 9.37310159e-01
1.58498317e-01 8.34082425e-01 -1.00907646e-01 -3.82609308e-01
2.67288893e-01 9.41141602e-03 7.29088485e-01 -9.66606796e-01
-7.52960682e-01 -1.92240521e-01 1.71034098e-01 -5.16683757e-01
5.63109741e-02 6.00368455e-02 -1.57362080e+00 -7.72030592e-01
-4.24423605e-01 3.74145776e-01 4.21112418e-01 1.18964732e+00
2.25981250e-01 6.96841121e-01 -7.60806650e-02 -4.61803496e-01
-9.75733876e-01 -1.09762347e+00 -4.20698911e-01 -2.03498602e-01
2.00145319e-01 -6.67665124e-01 -1.10876583e-01 -4.00612094e-02] | [8.796161651611328, 3.682978868484497] |
3d6f829b-5c49-4595-b848-329cb4515686 | a-search-without-expansions-learning | 2102.04518 | null | https://arxiv.org/abs/2102.04518v2 | https://arxiv.org/pdf/2102.04518v2.pdf | A* Search Without Expansions: Learning Heuristic Functions with Deep Q-Networks | Efficiently solving problems with large action spaces using A* search has been of importance to the artificial intelligence community for decades. This is because the computation and memory requirements of A* search grow linearly with the size of the action space. This burden becomes even more apparent when A* search uses a heuristic function learned by computationally expensive function approximators, such as deep neural networks. To address this problem, we introduce Q* search, a search algorithm that uses deep Q-networks to guide search in order to take advantage of the fact that the sum of the transition costs and heuristic values of the children of a node can be computed with a single forward pass through a deep Q-network without explicitly generating those children. This significantly reduces computation time and requires only one node to be generated per iteration. We use Q* search to solve the Rubik's cube when formulated with a large action space that includes 1872 meta-actions and find that this 157-fold increase in the size of the action space incurs less than a 4-fold increase in computation time and less than a 3-fold increase in number of nodes generated when performing Q* search. Furthermore, Q* search is up to 129 times faster and generates up to 1288 times fewer nodes than A* search. Finally, although obtaining admissible heuristic functions from deep neural networks is an ongoing area of research, we prove that Q* search is guaranteed to find a shortest path given a heuristic function that neither overestimates the cost of a shortest path nor underestimates the transition cost. | ['Pierre Baldi', 'Roy Fox', 'Stephen Mcaleer', 'Alexander Shmakov', 'Forest Agostinelli'] | 2021-02-08 | null | null | null | null | ['rubik-s-cube'] | ['graphs'] | [ 2.71979034e-01 5.97956002e-01 -2.04744503e-01 1.45412594e-01
-6.81745529e-01 -7.15050042e-01 2.85204165e-02 2.08958685e-01
-5.85515380e-01 1.02394712e+00 -1.24336220e-01 -5.52938879e-01
-5.63205004e-01 -1.38233387e+00 -9.52971041e-01 -6.25715792e-01
-4.58117098e-01 7.68628418e-01 1.02305217e-02 -3.20474684e-01
3.97298247e-01 6.11672342e-01 -1.48207808e+00 -4.85949926e-02
8.74540746e-01 1.27978599e+00 -9.38617997e-03 6.76862836e-01
-4.40881290e-02 3.81943285e-01 -6.43584549e-01 -5.12443557e-02
5.59189618e-01 -4.86704797e-01 -1.31619704e+00 -5.33543266e-02
1.69701979e-01 -2.73217976e-01 -3.00083220e-01 8.65149975e-01
3.64251614e-01 5.91902792e-01 2.98932821e-01 -1.19106388e+00
-1.93829954e-01 6.69365406e-01 -2.22337052e-01 9.32595506e-02
8.52854326e-02 2.91793585e-01 1.33267963e+00 -2.78078020e-01
5.75637996e-01 1.22787213e+00 7.09659636e-01 5.12841582e-01
-1.15723896e+00 -4.43147868e-01 -8.32263473e-03 2.64706612e-01
-1.28375983e+00 -4.90144268e-03 4.48563963e-01 2.14295551e-01
1.45333683e+00 2.45362118e-01 1.12028921e+00 3.38309228e-01
2.94674605e-01 5.64466357e-01 4.60308224e-01 -3.20657164e-01
6.77700877e-01 -4.77816284e-01 -3.13018650e-01 9.85345602e-01
1.10143267e-01 3.68323475e-01 -2.43903220e-01 -1.02340601e-01
9.05731380e-01 -1.96219712e-01 7.36363307e-02 -3.66220444e-01
-9.16338861e-01 1.29912937e+00 6.64360225e-01 9.63409916e-02
-3.94073874e-01 9.30478752e-01 5.03512084e-01 3.04567248e-01
7.80265406e-02 1.12234533e+00 -6.89851701e-01 -4.68598604e-01
-9.07284617e-01 5.01295030e-01 9.78034854e-01 5.41806698e-01
8.25262904e-01 1.72495514e-01 2.10248977e-01 4.05518532e-01
-2.48338372e-01 2.78029054e-01 1.30048901e-01 -1.79789054e+00
4.88177001e-01 6.98222578e-01 1.71763971e-01 -1.03659117e+00
-8.16075921e-01 -5.75480998e-01 -6.89034820e-01 4.17825043e-01
6.34765089e-01 -4.06989723e-01 -6.15299344e-01 1.71554089e+00
3.82738829e-01 -2.93984473e-01 -5.25692292e-02 4.92326558e-01
4.95307781e-02 9.11285222e-01 -4.21156585e-01 -3.89605045e-01
8.34876716e-01 -1.11954963e+00 -2.04833508e-01 -2.06210300e-01
1.12286627e+00 -9.29242074e-02 7.72838235e-01 5.66154718e-01
-1.72704685e+00 -1.35045797e-01 -1.10025656e+00 1.00329809e-01
-4.34118629e-01 -5.29873669e-01 9.03409362e-01 5.05290329e-01
-1.26447856e+00 1.00885558e+00 -7.64587104e-01 3.96916904e-02
5.16970932e-01 9.41385925e-01 -1.48775592e-01 -1.10309750e-01
-1.20730436e+00 1.00376236e+00 6.74532831e-01 2.73308903e-01
-7.58956492e-01 -7.50393450e-01 -7.78908610e-01 4.68754858e-01
8.78163517e-01 -5.10447502e-01 1.35194218e+00 -1.02193677e+00
-1.31383216e+00 3.17407884e-02 -1.36802360e-01 -6.35650992e-01
3.93339813e-01 2.16416150e-01 1.20447867e-01 2.02116162e-01
-3.92053090e-02 8.77600253e-01 2.86258101e-01 -6.93542600e-01
-7.21215665e-01 -2.74460673e-01 4.89983499e-01 2.39769503e-01
-1.67910844e-01 -3.81610215e-01 -1.77544832e-01 -2.54114538e-01
-3.50366929e-03 -1.08911252e+00 -8.09883952e-01 2.75142074e-01
-9.22241658e-02 -5.15472889e-01 2.28516519e-01 -3.23082536e-01
1.22542858e+00 -1.56537855e+00 2.86176324e-01 6.75426900e-01
1.82177186e-01 7.49783069e-02 -4.76295173e-01 5.09127438e-01
1.34934008e-01 2.90155470e-01 -4.54903305e-01 3.94481152e-01
1.43307060e-01 5.60813904e-01 1.19349271e-01 2.38395438e-01
-3.26282568e-02 1.00205576e+00 -1.03975892e+00 -2.79287964e-01
-2.07549155e-01 1.69203207e-01 -1.09750485e+00 -3.86211306e-01
-4.90157396e-01 -2.93263435e-01 -3.89684319e-01 3.09815407e-01
3.48130047e-01 -3.61353755e-01 4.73015726e-01 8.58220160e-02
-1.33776188e-01 2.23166078e-01 -1.37689531e+00 1.56014347e+00
-7.72317410e-01 6.20627940e-01 2.29804423e-02 -1.33380270e+00
7.76204109e-01 9.18427110e-02 7.56634653e-01 -1.14221799e+00
2.48053521e-01 3.92576963e-01 8.93106684e-02 -8.72460827e-02
4.36896682e-01 -5.00242673e-02 -1.80364326e-01 7.34008074e-01
-3.67268413e-01 -1.55508786e-01 7.78222680e-01 -1.43733203e-01
1.41814768e+00 -1.30443111e-01 5.36532467e-03 -7.33061805e-02
4.07017648e-01 2.65341163e-01 6.94126010e-01 8.15839767e-01
8.41131359e-02 -6.73004091e-02 8.78445923e-01 -9.41171050e-01
-9.92854357e-01 -7.35035300e-01 2.52986073e-01 1.12696648e+00
6.17444292e-02 -3.68750483e-01 -8.17887068e-01 -5.37578225e-01
-3.25158518e-03 7.51097560e-01 -8.01979005e-01 -3.04394215e-01
-1.03693175e+00 -4.68720466e-01 4.15155470e-01 8.11243951e-01
6.05056524e-01 -1.35224950e+00 -1.12140906e+00 5.61769128e-01
9.72070824e-03 -5.42805374e-01 -3.80121350e-01 5.74881554e-01
-1.08092296e+00 -1.21382737e+00 -3.36777598e-01 -8.25284064e-01
8.37902606e-01 -2.68767118e-01 1.15714860e+00 3.25110674e-01
-4.20071244e-01 -2.62969937e-02 1.76245831e-02 -6.74549192e-02
-2.55996376e-01 2.98147082e-01 -2.07712218e-01 -8.86529684e-01
7.77346035e-03 -4.92064834e-01 -5.51129997e-01 1.61720291e-01
-8.19830000e-01 -1.59231663e-01 4.99392360e-01 8.78344953e-01
6.58339739e-01 6.13542438e-01 6.31756723e-01 -2.86852747e-01
6.51779771e-01 -2.79942989e-01 -1.15533113e+00 1.81901917e-01
-7.53090560e-01 5.83819866e-01 7.94940591e-01 -3.52108091e-01
-5.11119604e-01 2.12418735e-01 -4.63784672e-02 -1.93835180e-02
4.97326434e-01 6.21105313e-01 2.38935053e-01 -1.37291402e-01
5.86201370e-01 7.00051337e-02 1.31893560e-01 -1.58072889e-01
4.57800716e-01 3.45609449e-02 2.78665423e-01 -6.87649131e-01
5.26625991e-01 1.13602877e-01 5.68853378e-01 -3.43548656e-01
-7.25754023e-01 9.81757343e-02 -3.28096509e-01 -2.14540586e-02
7.71343052e-01 -9.39792767e-02 -1.53987753e+00 -4.80723307e-02
-1.31270158e+00 -6.58323646e-01 -6.50433898e-01 4.19641547e-02
-8.11676562e-01 5.01818489e-04 -3.98254305e-01 -6.68243527e-01
-4.56719935e-01 -1.15154850e+00 4.41658765e-01 8.71021301e-02
-2.78690696e-01 -8.96600783e-01 -1.27807945e-01 2.31186554e-01
3.02677006e-01 3.29574943e-01 1.08857548e+00 -4.12429750e-01
-7.27571785e-01 -1.79108366e-01 -1.77182585e-01 2.27023482e-01
-1.54741153e-01 -1.94917724e-01 -7.89707080e-02 -2.00246915e-01
-4.06651646e-01 -3.06988597e-01 4.54574078e-01 4.99866545e-01
1.43054020e+00 -8.01972687e-01 -1.60197273e-01 4.86823678e-01
1.50408578e+00 6.54007375e-01 4.97438222e-01 5.81328273e-01
3.47294658e-01 5.10993659e-01 4.88943815e-01 4.95586932e-01
1.11214325e-01 5.24819493e-01 7.99602389e-01 1.42873839e-01
2.04426005e-01 -6.57922402e-02 6.28355518e-02 3.35727960e-01
-6.34843856e-02 -2.37819031e-01 -1.09810436e+00 7.03942835e-01
-1.74305642e+00 -9.87091660e-01 2.56466717e-01 2.05641937e+00
9.84070837e-01 3.39589566e-01 2.10964695e-01 4.63260591e-01
3.46559316e-01 -1.06189109e-01 -1.03182721e+00 -1.23355186e+00
4.94150341e-01 7.84575403e-01 9.24135506e-01 7.39832520e-01
-5.43456316e-01 8.42715979e-01 7.30082703e+00 7.70557523e-01
-8.50235283e-01 -4.33003753e-01 5.22398293e-01 -5.13575196e-01
-2.00605139e-01 -2.00525895e-01 -4.11918223e-01 2.87357628e-01
1.09118509e+00 -3.41193855e-01 1.06979167e+00 9.76425588e-01
2.49870554e-01 -2.16585949e-01 -1.15210128e+00 6.56296790e-01
-3.86154443e-01 -1.80116773e+00 -1.88608870e-01 4.00563836e-01
8.87309790e-01 -1.37604505e-01 -9.13729593e-02 2.07029402e-01
6.27149463e-01 -1.41541433e+00 4.89009917e-01 -2.45383102e-02
5.06135225e-01 -1.59113050e+00 5.79422712e-01 3.86433482e-01
-1.18960667e+00 -5.75369537e-01 -3.84100974e-01 -3.50221187e-01
1.81200355e-01 2.52990514e-01 -7.87400723e-01 9.72963944e-02
5.93864620e-01 3.13571580e-02 3.80803570e-02 1.14438057e+00
-2.95300167e-02 2.20702976e-01 -6.61168575e-01 -4.04098660e-01
8.75137448e-01 -2.02889293e-01 6.99589103e-02 6.13058329e-01
3.52015048e-01 1.53887346e-01 1.11636072e-01 9.96345341e-01
-1.00778148e-01 -2.47368798e-01 -3.47350717e-01 -1.49461433e-01
6.05481744e-01 8.89929652e-01 -8.38156343e-01 -1.47779197e-01
1.19652674e-02 5.67766428e-01 1.83566228e-01 2.16697991e-01
-8.42344999e-01 -9.03123617e-01 5.56759357e-01 -3.44870165e-02
5.17991006e-01 -3.69521827e-01 -4.83219802e-01 -1.02312647e-01
-1.28101319e-01 -8.46851766e-01 1.52701929e-01 -6.00481212e-01
-4.64563638e-01 4.25028116e-01 -1.35375035e-03 -6.49488091e-01
-7.39042878e-01 -5.56858897e-01 -4.25620645e-01 6.55064762e-01
-1.03125358e+00 -3.91780496e-01 -9.54342932e-02 2.39219919e-01
5.47415376e-01 1.57193005e-01 7.02978492e-01 1.10874427e-02
-4.70990419e-01 5.82085669e-01 1.13302119e-01 -1.38561502e-01
-3.28804284e-01 -1.04729617e+00 5.61770260e-01 4.24622923e-01
-2.49041647e-01 2.72213817e-01 6.68922901e-01 -2.92354494e-01
-1.57054281e+00 -7.07019389e-01 8.35631609e-01 -1.13858357e-01
6.19208276e-01 9.82773528e-02 -4.47342604e-01 4.19942290e-01
-5.82141392e-02 7.40503669e-02 1.56884700e-01 5.73804928e-03
5.77149950e-02 -2.47622699e-01 -1.14987803e+00 5.08406103e-01
1.18885946e+00 -7.47541562e-02 -1.70263097e-01 3.57290030e-01
7.78871119e-01 -3.15813482e-01 -8.59151363e-01 1.34789437e-01
5.53057611e-01 -7.90904164e-01 9.71239984e-01 -4.44377691e-01
4.83952522e-01 3.75491777e-03 7.61363432e-02 -1.23952341e+00
-2.15676859e-01 -7.25827992e-01 -2.58574098e-01 2.63198733e-01
5.40636837e-01 -6.27843320e-01 1.24404848e+00 7.73909271e-01
-1.63962975e-01 -1.49729466e+00 -1.10403538e+00 -1.01578784e+00
4.65063602e-01 -1.93996534e-01 7.02365875e-01 5.31075835e-01
9.36592743e-02 2.57807272e-03 6.21735230e-02 -4.09563422e-01
3.83336365e-01 1.44821420e-01 3.31374705e-01 -1.25227034e+00
-3.09017777e-01 -6.78420603e-01 1.39599619e-02 -1.07886529e+00
2.03386694e-01 -8.41614366e-01 1.88231155e-01 -1.87443364e+00
-1.98935002e-01 -6.32036805e-01 9.52413864e-03 8.93953204e-01
6.25793338e-01 3.73092555e-02 7.51407817e-02 -4.80360389e-01
-4.95932430e-01 5.09380698e-01 1.53499508e+00 -1.86000943e-01
-3.68723303e-01 -1.06582470e-01 -6.61680341e-01 7.53469944e-01
8.76130342e-01 -4.89224136e-01 -6.20266378e-01 -4.36604381e-01
1.03592682e+00 4.57275391e-01 8.33726600e-02 -8.84882033e-01
2.34624386e-01 -5.23189187e-01 2.22165570e-01 -4.86109436e-01
3.82351965e-01 -6.46296144e-01 -4.60829660e-02 8.71282399e-01
-5.92195213e-01 5.25873184e-01 3.05102199e-01 2.27965534e-01
4.69032936e-02 -5.44510841e-01 6.02380037e-01 -3.94036919e-01
-4.67177540e-01 1.93175808e-01 -5.40198207e-01 1.08603567e-01
9.36456442e-01 -5.58953702e-01 -2.05260575e-01 -3.00706685e-01
-6.08952701e-01 3.83035034e-01 8.79861563e-02 1.38769923e-02
4.74352062e-01 -1.27077889e+00 -1.06847048e-01 -9.00418386e-02
-6.42903209e-01 4.59416986e-01 1.15250424e-02 5.45623362e-01
-8.05743933e-01 5.51741838e-01 -3.52728844e-01 7.91693330e-02
-1.01562202e+00 5.62149048e-01 6.27078891e-01 -4.89446670e-01
-5.94287634e-01 8.45282912e-01 -3.25167686e-01 -3.18547189e-01
3.72738242e-01 -4.37233776e-01 2.75970101e-01 4.02930491e-02
3.97976756e-01 1.06356204e+00 -6.19931668e-02 -7.18868747e-02
-4.35292840e-01 4.70837444e-01 1.70141265e-01 -1.81518137e-01
1.54856515e+00 4.56284970e-01 -2.14590073e-01 -4.01596129e-01
1.40958285e+00 -4.54236776e-01 -1.28625762e+00 2.71633208e-01
-1.25956178e-01 -1.60002336e-02 3.40747327e-01 -8.96161377e-01
-1.49266326e+00 6.61120355e-01 2.78185666e-01 3.36752683e-01
1.10614944e+00 -1.82915017e-01 1.11659753e+00 1.05847073e+00
3.37020278e-01 -1.70443273e+00 2.54385293e-01 7.05361843e-01
8.23920667e-01 -6.29570901e-01 7.56677762e-02 -7.82944635e-03
-1.25734806e-01 1.31726158e+00 7.25381672e-01 -2.59102464e-01
2.82096267e-01 1.65154442e-01 -4.96389508e-01 -1.38075918e-01
-8.53623807e-01 -8.81061703e-02 -2.07362808e-02 3.23589474e-01
-3.71782146e-02 -1.65037781e-01 -4.35874850e-01 1.73130661e-01
-4.55286533e-01 1.14161380e-01 5.07963300e-01 9.56219554e-01
-7.31727362e-01 -1.16618025e+00 -8.95759538e-02 3.84885818e-01
-4.98661846e-02 4.78834175e-02 -1.16086863e-01 7.97390938e-01
1.51214674e-01 9.06194091e-01 3.79642725e-01 -1.61035568e-01
1.21018395e-01 -2.38317758e-01 7.82290399e-01 -2.98765808e-01
-6.30278826e-01 -4.59153324e-01 2.09204927e-01 -1.05683279e+00
4.63172747e-03 -2.45490834e-01 -1.95830154e+00 -6.44398451e-01
-9.93556008e-02 2.16120094e-01 8.02618206e-01 1.04785049e+00
2.78835207e-01 5.65250814e-01 4.97513801e-01 -8.67735505e-01
-8.60212445e-01 -2.15046272e-01 -2.25212365e-01 -3.62866580e-01
2.23645642e-01 -5.65285623e-01 -2.90109128e-01 -5.88990033e-01] | [5.147482872009277, 2.9710733890533447] |
b0247371-7489-4b1e-9e36-fdd89d94e9f9 | automated-evaluation-of-scientific-writing | null | null | https://aclanthology.org/W15-0607 | https://aclanthology.org/W15-0607.pdf | Automated Evaluation of Scientific Writing: AESW Shared Task Proposal | null | ['Vidas Daudaravi{\\v{c}}ius'] | 2015-06-01 | null | null | null | ws-2015-6 | ['grammatical-error-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.175906181335449, 3.5512032508850098] |
9dd21300-270f-46e7-bc9d-033d38741c6a | detection-of-inferior-myocardial-infarction | 1710.01115 | null | http://arxiv.org/abs/1710.01115v4 | http://arxiv.org/pdf/1710.01115v4.pdf | Detection of Inferior Myocardial Infarction using Shallow Convolutional Neural Networks | Myocardial Infarction is one of the leading causes of death worldwide. This
paper presents a Convolutional Neural Network (CNN) architecture which takes
raw Electrocardiography (ECG) signal from lead II, III and AVF and
differentiates between inferior myocardial infarction (IMI) and healthy
signals. The performance of the model is evaluated on IMI and healthy signals
obtained from Physikalisch-Technische Bundesanstalt (PTB) database. A
subject-oriented approach is taken to comprehend the generalization capability
of the model and compared with the current state of the art. In a
subject-oriented approach, the network is tested on one patient and trained on
rest of the patients. Our model achieved a superior metrics scores (accuracy=
84.54%, sensitivity= 85.33% and specificity= 84.09%) when compared to the
benchmark. We also analyzed the discriminating strength of the features
extracted by the convolutional layers by means of geometric separability index
and euclidean distance and compared it with the benchmark model. | ['Tahsin Reasat', 'Celia Shahnaz'] | 2017-10-03 | null | null | null | null | ['electrocardiography-ecg'] | ['methodology'] | [ 2.29093477e-01 -5.67900538e-02 2.21512362e-01 -1.53920963e-01
-3.88514578e-01 -3.40330154e-01 8.74138549e-02 3.93815368e-01
-7.15982497e-01 7.28937328e-01 -1.05203986e-01 -4.62318718e-01
-5.10690451e-01 -6.04976594e-01 -1.68493614e-01 -6.62189841e-01
-7.55776525e-01 2.42816731e-01 -3.30124438e-01 5.23651913e-02
6.26066774e-02 8.07942390e-01 -8.53023887e-01 3.34874690e-01
6.23441637e-01 1.30847919e+00 -2.94377387e-01 8.24166179e-01
5.31067252e-01 3.72967362e-01 -8.04534137e-01 1.22658670e-01
3.62779200e-01 -7.20275044e-01 -9.30604696e-01 -2.74186939e-01
1.07919490e-02 -1.40910745e-01 -2.90781736e-01 6.48071468e-01
9.36865628e-01 -4.80271101e-01 7.77581036e-01 -8.80016923e-01
-1.23336293e-01 4.56430674e-01 -2.48534054e-01 8.37239742e-01
-5.35940006e-02 -1.86979659e-02 5.36469579e-01 -8.43219697e-01
6.76719308e-01 4.40422952e-01 1.11619461e+00 1.35468274e-01
-1.01429749e+00 -4.70299661e-01 -5.88757694e-01 3.75770122e-01
-1.50075698e+00 2.08961874e-01 6.83787942e-01 -4.72353995e-01
6.49906397e-01 3.51656646e-01 9.16154325e-01 7.73372293e-01
6.13855422e-01 3.45465899e-01 1.14322245e+00 -2.66624004e-01
9.90201756e-02 -1.48337811e-01 3.77726465e-01 4.21292722e-01
4.33617532e-01 2.04692721e-01 -5.80844469e-02 -6.07139543e-02
7.71586776e-01 -8.16984773e-02 -6.32219732e-01 -1.95630148e-01
-1.58515382e+00 5.69597304e-01 5.50657332e-01 5.99119842e-01
-8.18724394e-01 -1.58340156e-01 8.39619517e-01 4.71527845e-01
1.86587404e-02 6.19754076e-01 -5.55298269e-01 -2.09243119e-01
-8.83071244e-01 1.32962167e-01 5.55258632e-01 2.79302567e-01
1.26449794e-01 2.50500411e-01 -2.25438252e-01 4.41038251e-01
1.98686093e-01 4.22109127e-01 6.11200809e-01 -7.20980525e-01
2.09627271e-01 7.16804504e-01 -9.25564691e-02 -1.08635032e+00
-9.40790951e-01 -1.06522596e+00 -1.47892344e+00 3.44729245e-01
4.73506153e-01 -4.54684019e-01 -6.59534812e-01 1.19284868e+00
-1.89194843e-01 3.12204845e-02 2.57624745e-01 1.09514141e+00
1.07468009e+00 2.02016696e-01 -1.06938109e-01 -2.61423528e-01
1.26495147e+00 -1.92274734e-01 -6.49380028e-01 3.29143077e-01
7.94398129e-01 -3.40529859e-01 4.85189915e-01 6.26478672e-01
-9.16133106e-01 -6.57473981e-01 -1.36113703e+00 5.24124920e-01
-1.06371179e-01 5.43925464e-01 1.34972632e-01 7.00925112e-01
-9.19157267e-01 1.03521180e+00 -6.86596155e-01 -5.02751350e-01
6.44668102e-01 4.63119894e-01 -5.62181532e-01 3.01014096e-01
-1.52101004e+00 7.96325684e-01 4.59073752e-01 4.00096297e-01
-6.85980082e-01 -5.92188418e-01 -5.41945636e-01 2.72166915e-02
-4.69537824e-01 -7.00232863e-01 6.09736145e-01 -9.40542459e-01
-1.04918981e+00 8.53654623e-01 4.95876610e-01 -7.74580956e-01
7.76929677e-01 -3.86709422e-02 -3.41755658e-01 4.79475856e-01
-1.67035297e-01 1.99750900e-01 2.58519500e-01 -7.99856603e-01
-3.77267927e-01 -4.90310580e-01 -8.09568763e-02 -4.28193957e-02
-7.20758736e-02 -3.88857797e-02 3.41311246e-01 -5.91073334e-01
4.40651774e-01 -8.09265196e-01 -1.21728688e-01 -3.53307247e-01
-4.78419662e-01 1.57645434e-01 5.95363319e-01 -8.20793748e-01
1.22346985e+00 -2.21876764e+00 3.86000350e-02 5.22483468e-01
4.52455252e-01 5.58288276e-01 2.21628651e-01 4.37686294e-01
-5.87714970e-01 2.29927853e-01 -2.37296611e-01 4.21935499e-01
-4.30819333e-01 -1.23317987e-01 2.83107162e-01 7.63467848e-01
2.12635636e-01 8.82618010e-01 -4.90180820e-01 -3.13436955e-01
2.28050023e-01 5.50584078e-01 -1.16974570e-01 -4.41837311e-02
7.73611784e-01 6.87350988e-01 -2.36482129e-01 4.79508370e-01
5.65656185e-01 -1.24385662e-01 3.68770301e-01 -5.12722075e-01
1.50584668e-01 -6.12964444e-02 -1.12085164e+00 1.42700660e+00
-6.13138266e-02 7.37132728e-01 -3.13517541e-01 -1.36663401e+00
1.18044245e+00 7.05456018e-01 6.46312654e-01 -6.54621482e-01
4.78312761e-01 3.54634821e-01 6.27767503e-01 -6.96122706e-01
-2.85762608e-01 -1.33431166e-01 6.92064837e-02 3.28245938e-01
-1.13384306e-01 5.90758443e-01 -4.91213799e-02 -1.06339715e-01
1.10872424e+00 -3.04735135e-02 6.14250958e-01 -7.74176300e-01
6.57694578e-01 -1.08353652e-01 5.58217049e-01 7.48505235e-01
-5.21396935e-01 8.56996298e-01 7.79152572e-01 -1.14391720e+00
-7.99813271e-01 -1.05435288e+00 -4.83342975e-01 6.44545071e-03
-1.18887983e-01 -1.17729187e-01 -9.11609471e-01 -5.66229522e-01
-1.36858374e-01 -2.51761568e-03 -9.20180440e-01 -2.07425877e-01
-7.85501957e-01 -9.78735209e-01 9.47481990e-01 9.39541817e-01
7.79870570e-01 -1.09359670e+00 -1.36577892e+00 2.15399742e-01
-3.48245241e-02 -7.93860912e-01 2.45138124e-01 1.78289041e-01
-1.13040173e+00 -1.19922411e+00 -8.29160511e-01 -6.67877436e-01
2.33684123e-01 -4.62145925e-01 1.21200466e+00 2.13134885e-01
-6.87408268e-01 -2.23091796e-01 -2.30688855e-01 -6.74853802e-01
-4.41885918e-01 1.72096178e-01 -3.46738682e-03 2.30879351e-01
1.22307606e-01 -6.72969997e-01 -1.14903128e+00 1.70768142e-01
-7.02752352e-01 -2.08919376e-01 9.11903203e-01 7.90189266e-01
3.04525942e-01 -2.86542773e-01 9.85483527e-01 -7.03882754e-01
7.39864886e-01 -3.25813442e-01 -1.11961193e-01 -1.26248851e-01
-8.45424414e-01 -4.41770703e-01 4.63463902e-01 -1.81071267e-01
-2.67910480e-01 3.19401212e-02 -1.40312150e-01 -1.53590187e-01
-4.10990328e-01 5.97734571e-01 1.96142984e-03 2.12829083e-01
9.13489401e-01 1.71714500e-01 1.10598423e-01 -2.46807382e-01
-4.69565153e-01 7.35682666e-01 6.62912786e-01 -8.53071511e-02
2.98580706e-01 3.47114146e-01 4.02462155e-01 -8.10039997e-01
-2.73284525e-01 -3.21557134e-01 -8.79537225e-01 -3.27471763e-01
9.43371713e-01 -5.56308389e-01 -7.24916637e-01 5.91166317e-01
-1.12596357e+00 -4.70088571e-02 -3.50686044e-01 8.13836217e-01
-3.22557122e-01 3.28927070e-01 -6.02664292e-01 -5.46817124e-01
-9.29720700e-01 -9.69191849e-01 5.22423685e-01 -2.31271163e-01
-4.91415769e-01 -8.85949910e-01 8.27447325e-03 -8.34049508e-02
5.46258152e-01 1.11812174e+00 1.04246616e+00 -1.03492975e+00
1.45744113e-02 -7.28416264e-01 -2.80327618e-01 6.85548961e-01
-1.74041186e-02 -2.35770985e-01 -8.80202949e-01 -2.20632508e-01
-1.00588519e-02 2.03035161e-01 6.35095060e-01 6.07514918e-01
8.98828745e-01 9.69446152e-02 -1.48879707e-01 6.40702963e-01
1.46688592e+00 5.97530007e-01 1.01997721e+00 6.45964980e-01
3.90436500e-01 9.88485143e-02 1.47501841e-01 3.37737113e-01
-1.10818846e-02 3.21027786e-01 4.33999240e-01 -3.91846895e-01
2.02470813e-02 5.62322080e-01 -2.24352449e-01 5.49284220e-01
-5.25893211e-01 6.94759712e-02 -1.21933699e+00 5.55928767e-01
-1.56128740e+00 -7.60049999e-01 -5.41862011e-01 2.11309195e+00
1.90824628e-01 2.85056800e-01 2.15949193e-01 8.94108951e-01
6.25201404e-01 -3.48519385e-01 -3.60313326e-01 -4.96336848e-01
-1.77618071e-01 4.85022217e-01 2.19287083e-01 -6.21846616e-02
-1.25707531e+00 -8.58865976e-02 6.14539433e+00 1.37443185e-01
-1.44191802e+00 -6.70044050e-02 9.20872629e-01 2.18025327e-01
5.97286463e-01 -3.79323393e-01 2.61895135e-02 3.02732050e-01
1.17785871e+00 -2.39506289e-02 -1.98318064e-01 4.33611035e-01
1.69756398e-01 1.97757959e-01 -1.03058386e+00 1.07397234e+00
1.74662694e-02 -1.45096254e+00 -3.14075381e-01 1.12247374e-02
2.90052980e-01 1.11673318e-01 -1.65035456e-01 8.51487368e-02
-9.74572778e-01 -1.26954579e+00 4.51325208e-01 8.10789824e-01
9.67200756e-01 -7.93127000e-01 1.58156514e+00 1.59524575e-01
-9.34384048e-01 -1.90946162e-01 1.31389707e-01 -2.53854871e-01
-2.60628015e-01 3.75773042e-01 -7.87643254e-01 8.46053898e-01
9.37799275e-01 6.47635937e-01 -5.06986320e-01 1.02714109e+00
1.41504288e-01 8.40548217e-01 -1.50583476e-01 1.09412774e-01
1.84553668e-01 -2.71852557e-02 4.81079727e-01 1.28880310e+00
1.88435912e-01 -1.10785805e-01 -8.77397582e-02 7.32970536e-01
1.44699261e-01 4.37127680e-01 -5.51545262e-01 2.33949140e-01
4.65956777e-02 1.32725453e+00 -8.18234265e-01 -4.82321322e-01
7.06137531e-03 5.63862503e-01 -3.87390882e-01 2.68884838e-01
-5.74699759e-01 -8.88891876e-01 1.79623924e-02 3.40817720e-01
-1.01879239e-01 2.21517265e-01 -7.96962976e-01 -7.60917127e-01
1.27772853e-01 -8.29104424e-01 3.87835890e-01 -5.07987857e-01
-9.04240429e-01 1.06751120e+00 3.53294574e-02 -1.45406508e+00
-2.61332244e-01 -7.29009449e-01 -7.52654374e-01 1.32445419e+00
-1.18470979e+00 -6.92108214e-01 -5.69396794e-01 2.57513583e-01
-2.43009767e-03 -4.22297895e-01 1.17054141e+00 4.32513297e-01
-5.25942862e-01 5.99972606e-01 -1.24084041e-01 7.48478889e-01
2.41492406e-01 -1.27503073e+00 1.10690005e-01 7.02533662e-01
-4.75079447e-01 5.38636684e-01 4.72607404e-01 -1.19341888e-01
-7.41501510e-01 -9.76143062e-01 1.04291606e+00 -1.09666638e-01
2.69685715e-01 2.59388328e-01 -7.20299542e-01 2.50812113e-01
3.42473865e-01 2.00325429e-01 7.61812687e-01 -3.25452834e-01
-6.50181621e-03 -2.22957954e-01 -1.15525222e+00 1.42048404e-01
5.61637402e-01 -9.99140814e-02 -4.59402174e-01 -4.41324674e-02
-1.53454334e-01 -2.49522775e-01 -1.48462570e+00 8.64444554e-01
1.10155237e+00 -1.20476568e+00 8.94778371e-01 -7.14763701e-01
4.30545628e-01 -1.00969575e-01 4.57038879e-02 -1.14602423e+00
-3.82359684e-01 -4.99393493e-01 1.64044768e-01 4.96634275e-01
4.48117763e-01 -9.32051539e-01 5.68333626e-01 -1.09347455e-01
-1.52583584e-01 -1.39813948e+00 -8.76460493e-01 -6.54062927e-01
3.87184471e-01 -1.63391098e-01 3.98147672e-01 8.39871049e-01
-4.36557010e-02 2.96860695e-01 1.46906748e-02 -4.39728200e-02
3.94379675e-01 1.17157204e-02 3.00750196e-01 -1.66362178e+00
1.02041356e-01 -3.76004905e-01 -1.09565008e+00 1.81570873e-01
-4.00596738e-01 -9.67883706e-01 -5.04782796e-01 -1.69077957e+00
1.02599949e-01 -2.06071258e-01 -9.68027592e-01 3.77683222e-01
6.16139695e-02 7.51671612e-01 1.64924622e-01 2.34816790e-01
4.46194820e-02 -2.79500097e-01 1.10092771e+00 -1.92266470e-03
-2.35526353e-01 2.97982413e-02 -5.24905264e-01 6.77545726e-01
1.30974114e+00 -3.59979361e-01 -2.07528695e-01 5.63432314e-02
2.31614724e-01 3.34801495e-01 6.30132020e-01 -1.55091512e+00
-3.31203371e-01 5.09958088e-01 8.65234375e-01 -5.61120868e-01
-1.15819827e-01 -5.86635649e-01 3.44493419e-01 1.01017463e+00
-4.72953796e-01 4.05444235e-01 8.33885148e-02 1.73961833e-01
-2.97023445e-01 7.06789196e-02 9.12365496e-01 -1.49386674e-02
-1.53713793e-01 2.53494114e-01 -4.02942628e-01 6.61889091e-02
9.63045061e-01 -4.86530840e-01 1.07047092e-02 -3.49371210e-02
-1.05949342e+00 -4.22416568e-01 -6.97908625e-02 4.86677215e-02
8.22710574e-01 -1.35912275e+00 -1.14430666e+00 3.69563311e-01
1.90506622e-01 -3.59366894e-01 3.85377467e-01 1.65366781e+00
-1.21115446e+00 3.57510746e-01 -6.54953599e-01 -7.70417929e-01
-1.28728187e+00 1.96468845e-01 8.84923756e-01 -2.86322773e-01
-1.00749981e+00 3.34952176e-01 -1.34865895e-01 -1.73115972e-02
1.70173839e-01 -5.32397985e-01 -6.28742695e-01 4.08534035e-02
4.14585829e-01 5.55693984e-01 4.66589391e-01 -5.99242806e-01
-6.47783458e-01 5.63083410e-01 2.85628974e-01 1.99814200e-01
1.25742340e+00 3.64975393e-01 -9.29617435e-02 4.35017020e-01
1.36952734e+00 -6.10979319e-01 -6.32569849e-01 1.62235796e-01
-2.34668087e-02 2.99990121e-02 1.39062420e-01 -1.06750309e+00
-1.22567081e+00 1.30563343e+00 1.42873216e+00 4.36639875e-01
1.13654804e+00 -5.69854796e-01 8.08919668e-01 4.33008015e-01
-6.87942207e-02 -8.41309786e-01 -2.41216734e-01 1.06376581e-01
8.10602546e-01 -9.62527096e-01 -3.20106447e-02 1.54878814e-02
-6.35814667e-01 1.62066948e+00 -1.27996817e-01 -5.23243308e-01
1.01622248e+00 1.88965052e-01 5.34051120e-01 -3.65739197e-01
-1.57514840e-01 1.45806476e-01 3.56832355e-01 6.66985989e-01
7.15110004e-01 2.44317442e-01 -7.71221995e-01 7.52316594e-01
7.03107342e-02 3.38098139e-01 4.07251894e-01 9.87121761e-01
-1.74313903e-01 -7.03295410e-01 -1.49762288e-01 5.81784844e-01
-1.02384913e+00 2.33967099e-02 -3.02254051e-01 1.07997477e+00
2.88652241e-01 7.33469963e-01 -2.14915350e-01 -4.82185811e-01
4.52406704e-01 3.25455695e-01 2.34765813e-01 -1.66594848e-01
-1.03019691e+00 4.40849103e-02 7.47153684e-02 -2.58891821e-01
-4.88971561e-01 -5.27460754e-01 -1.26886261e+00 2.56461531e-01
9.26294699e-02 2.28286296e-01 7.48477101e-01 7.54298568e-01
4.39785153e-01 1.06288338e+00 6.53781056e-01 -5.42354524e-01
-5.17575264e-01 -1.30295491e+00 -8.26272309e-01 5.90085626e-01
4.10275191e-01 -3.14711004e-01 -2.42791384e-01 4.70851809e-02] | [14.32568645477295, 3.2830898761749268] |
0c35e7bf-664e-406f-87ee-2eac9e48055c | efficient-partial-credit-grading-of-proof | 2204.04196 | null | https://arxiv.org/abs/2204.04196v3 | https://arxiv.org/pdf/2204.04196v3.pdf | Efficient Feedback and Partial Credit Grading for Proof Blocks Problems | Proof Blocks is a software tool that allows students to practice writing mathematical proofs by dragging and dropping lines instead of writing proofs from scratch. Proof Blocks offers the capability of assigning partial credit and providing solution quality feedback to students. This is done by computing the edit distance from a student's submission to some predefined set of solutions. In this work, we propose an algorithm for the edit distance problem that significantly outperforms the baseline procedure of exhaustively enumerating over the entire search space. Our algorithm relies on a reduction to the minimum vertex cover problem. We benchmark our algorithm on thousands of student submissions from multiple courses, showing that the baseline algorithm is intractable, and that our proposed algorithm is critical to enable classroom deployment. Our new algorithm has also been used for problems in many other domains where the solution space can be modeled as a DAG, including but not limited to Parsons Problems for writing code, helping students understand packet ordering in networking protocols, and helping students sketch solution steps for physics problems. Integrated into multiple learning management systems, the algorithm serves thousands of students each year. | ['Matthew West', 'Geoffrey Herman', 'Shubhang Kulkarni', 'Seth Poulsen'] | 2022-04-08 | null | null | null | null | ['mathematical-proofs'] | ['miscellaneous'] | [ 2.12653711e-01 1.45273358e-01 -9.65353176e-02 -3.12042654e-01
-7.13779449e-01 -1.14231789e+00 -1.91356510e-01 8.57278347e-01
-1.50781378e-01 7.61683285e-01 -6.93639278e-01 -1.12609446e+00
-4.86614108e-01 -1.04319739e+00 -1.06345153e+00 -1.63884595e-01
-1.99042618e-01 3.57199013e-01 5.76602876e-01 -2.33553052e-01
9.01553035e-01 7.74533868e-01 -1.62889683e+00 -1.03818811e-01
1.22053111e+00 1.44614697e-01 2.63462543e-01 1.13896191e+00
-4.22819912e-01 7.97747731e-01 -7.94081569e-01 -5.17536163e-01
-4.31396365e-02 -3.21306556e-01 -1.10738075e+00 -4.06874090e-01
1.08335817e+00 -2.90418684e-01 -2.49854755e-02 1.22730660e+00
1.64920911e-01 2.27109194e-01 1.00082166e-01 -1.85716379e+00
-4.66035634e-01 8.84500921e-01 -5.78441679e-01 2.93670356e-01
7.71197677e-01 -4.40079719e-01 1.10202932e+00 -3.08632970e-01
7.19865501e-01 8.78068805e-01 4.17607963e-01 4.07538742e-01
-1.28555989e+00 -7.38677144e-01 3.27211380e-01 4.36140835e-01
-1.02090752e+00 1.18273959e-01 3.89723182e-01 -5.25786340e-01
7.87303984e-01 5.69915771e-01 8.96278501e-01 2.92047709e-01
2.40724862e-01 7.53054738e-01 7.50114560e-01 -5.06958246e-01
2.09541902e-01 2.85366446e-01 7.02041566e-01 1.22161746e+00
6.43540680e-01 -4.41310376e-01 -4.16227609e-01 -1.88974917e-01
2.77242839e-01 -1.24327503e-01 -3.79399002e-01 -6.72450840e-01
-1.08213413e+00 6.09043300e-01 -9.51829851e-02 -1.46473601e-01
4.43241060e-01 4.00584936e-01 -6.56970516e-02 8.27919602e-01
-4.40111458e-01 5.06439269e-01 -6.79926336e-01 -4.25163329e-01
-1.06767440e+00 5.88266551e-01 1.52626729e+00 1.30145717e+00
5.95589638e-01 -4.00985569e-01 1.68044105e-01 2.89004177e-01
2.99852461e-01 1.90743983e-01 -1.10891975e-01 -1.32656312e+00
5.25469482e-01 7.65489280e-01 -9.97433364e-02 -9.16826665e-01
1.49492949e-01 -1.75624579e-01 5.64039201e-02 5.25886416e-01
8.04956615e-01 -1.21840507e-01 -3.22362363e-01 1.78156853e+00
5.81004739e-01 4.69530761e-01 -2.37519518e-01 5.40613592e-01
7.68449068e-01 5.20617306e-01 -3.87844771e-01 -4.60727923e-02
1.26653838e+00 -1.06814659e+00 -3.77048165e-01 4.15383339e-01
1.01318383e+00 -1.15556252e+00 8.35516512e-01 1.15594864e+00
-1.52760339e+00 -1.66353285e-01 -1.17455077e+00 -3.51099849e-01
-4.63569641e-01 -4.64313179e-01 5.68004668e-01 1.05811238e+00
-1.34592640e+00 1.03071749e+00 -3.41989607e-01 -1.98675826e-01
2.50658721e-01 4.42795753e-01 -1.18416503e-01 -5.89636683e-01
-7.89725661e-01 5.37968993e-01 -1.63443744e-01 -6.06031120e-01
-6.48211956e-01 -1.35296035e+00 -5.70204020e-01 5.38141310e-01
3.42880577e-01 -5.96439958e-01 1.50956106e+00 -1.00708418e-01
-1.44804466e+00 7.47432113e-01 2.54579008e-01 1.66366901e-02
5.68032563e-01 8.32306817e-02 3.20325196e-01 4.41657417e-02
-2.15854660e-01 4.68269914e-01 2.60027200e-01 -6.19996965e-01
-5.47455668e-01 1.66928966e-03 6.90176904e-01 4.00527082e-02
-3.97369593e-01 -2.96817031e-02 -2.66315252e-01 -2.44864672e-01
2.04081967e-01 -8.75271440e-01 -2.92536706e-01 2.32291073e-01
-3.34938094e-02 -6.10795915e-01 7.59822309e-01 -1.44380152e-01
1.22503829e+00 -1.65608573e+00 3.20421122e-02 6.49986029e-01
4.91021335e-01 -3.51468064e-02 -1.69355378e-01 5.71176410e-01
-1.03942551e-01 3.24878007e-01 1.74447030e-01 8.06134865e-02
5.78384280e-01 6.23081923e-02 -2.19748318e-01 3.21877033e-01
-5.28558254e-01 4.49794441e-01 -1.31370389e+00 -6.46455467e-01
-2.68102903e-02 -7.53057227e-02 -9.14758921e-01 -1.54903885e-02
-2.47259989e-01 -2.61028618e-01 -1.83575302e-01 4.66149271e-01
9.10555959e-01 -2.43736014e-01 4.33200002e-01 8.87119591e-01
-2.46677741e-01 7.33458221e-01 -1.67142701e+00 1.85056388e+00
-3.47871006e-01 7.60498047e-01 5.42668760e-01 -1.01025534e+00
5.81966817e-01 4.74889018e-02 5.27168751e-01 -2.17796251e-01
-2.32457146e-01 1.91728994e-01 1.12617739e-01 -3.06971550e-01
4.41071123e-01 5.63431561e-01 6.96046203e-02 1.37095070e+00
-1.08810529e-01 -5.59033275e-01 9.24440861e-01 1.09122539e+00
1.62573493e+00 -8.51757545e-03 -4.12025511e-01 -5.42338014e-01
4.14843738e-01 -1.10276096e-01 3.40880781e-01 1.04458988e+00
-9.55414549e-02 -6.67364672e-02 9.21340287e-01 -1.36652321e-01
-6.33341670e-01 -1.13420236e+00 1.40442833e-01 1.17830920e+00
1.11723736e-01 -9.15830612e-01 -9.82142389e-01 -8.56397629e-01
2.31565535e-01 4.78023440e-01 2.36387048e-02 8.97384062e-03
-3.89608651e-01 1.83648616e-01 2.39421830e-01 3.08405191e-01
-1.17723010e-01 -6.34316504e-01 -6.57891512e-01 2.39507943e-01
5.63541763e-02 -9.28130329e-01 -6.76667035e-01 9.00893211e-02
-8.62282097e-01 -1.21488833e+00 -2.75809199e-01 -1.10587156e+00
1.13044024e+00 7.64359593e-01 1.22599018e+00 1.12440765e+00
-6.55240655e-01 9.07193542e-01 -8.14963430e-02 -5.68769157e-01
-3.20661187e-01 -3.57010253e-02 -1.46387190e-01 -1.15255356e+00
2.23280653e-01 -7.77128994e-01 -4.08918768e-01 3.42617370e-02
-6.61997497e-01 7.47600719e-02 4.30980064e-02 3.00865889e-01
2.09394470e-01 2.50194997e-01 2.34102502e-01 -1.11195886e+00
8.04403067e-01 -1.23974450e-01 -1.32798862e+00 5.74550629e-01
-7.89650738e-01 1.87816426e-01 6.72246516e-01 -1.53673217e-01
-2.52882630e-01 -1.97471976e-01 2.08336458e-01 1.90045554e-02
2.15302557e-02 3.21487993e-01 1.09729730e-01 -8.10159564e-01
1.01478949e-01 -1.47926554e-01 -1.72937587e-01 -5.70316277e-02
1.24721497e-01 2.27522049e-02 2.08792329e-01 -1.24230826e+00
9.72363710e-01 -3.67647201e-01 5.21808624e-01 -4.20941442e-01
-6.96135879e-01 -5.04363418e-01 -5.12478590e-01 -4.28448617e-01
1.14818893e-01 -4.12719727e-01 -1.69147646e+00 -4.96272184e-02
-1.25769043e+00 -6.02336526e-01 -9.19796973e-02 3.62559795e-01
-2.08726004e-01 5.82167983e-01 -8.39931250e-01 -4.47232008e-01
2.06113949e-01 -1.35502911e+00 4.19917971e-01 5.06330252e-01
-3.06428343e-01 -1.07471478e+00 2.17239872e-01 6.10295892e-01
1.91619858e-01 -3.14383395e-02 1.22813773e+00 -4.70473886e-01
-1.25001478e+00 -9.87179801e-02 -7.18010217e-02 -1.37876943e-01
-4.00020301e-01 5.61359286e-01 -5.17499089e-01 -3.81572932e-01
-5.34993768e-01 -1.62710786e-01 6.22135341e-01 5.33088446e-02
1.55004370e+00 -3.17378014e-01 -3.80164504e-01 4.12684381e-01
1.35832262e+00 -1.13667376e-01 1.83084682e-01 3.21337134e-01
4.69338357e-01 6.29318357e-01 3.48089665e-01 2.19902784e-01
5.60068607e-01 2.30674386e-01 3.52231830e-01 3.39020908e-01
1.84362352e-01 -1.65949196e-01 3.73623699e-01 1.07687354e+00
2.85039246e-01 -2.43430778e-01 -9.57259059e-01 6.05622351e-01
-1.79199743e+00 -9.68599021e-01 -7.07891583e-01 2.28245854e+00
9.43194926e-01 3.62430602e-01 1.13555431e-01 3.68106186e-01
3.39844912e-01 -5.12005687e-01 -1.15380110e-02 -1.31837332e+00
7.91368186e-01 8.24414253e-01 5.31692922e-01 1.00640440e+00
-4.12674129e-01 6.90269291e-01 6.24629593e+00 4.90611523e-01
-7.30116785e-01 -2.58317679e-01 -5.01594730e-02 -5.50422408e-02
-6.62170410e-01 2.79774487e-01 -9.02124763e-01 1.61329046e-01
1.04793131e+00 -6.61568463e-01 8.28653991e-01 7.94414699e-01
-2.61168815e-02 -1.10064350e-01 -1.46755564e+00 5.86472690e-01
-1.93751335e-01 -1.55631328e+00 -3.79120827e-01 2.08526462e-01
1.16537321e+00 -5.93201578e-01 3.06607075e-02 3.12065154e-01
1.15214217e+00 -8.54325891e-01 5.07758677e-01 1.90987606e-02
2.75011033e-01 -9.66521382e-01 1.65354908e-01 3.90061885e-01
-1.02706254e+00 2.07421165e-02 -2.30404243e-01 -4.10861164e-01
-5.59968114e-01 4.58565503e-01 -1.25961375e+00 2.95573205e-01
5.41536987e-01 2.26248369e-01 -5.59563279e-01 1.54244554e+00
-5.38347602e-01 5.54976046e-01 -3.63368541e-01 -6.15087748e-01
9.22147706e-02 -3.80157918e-01 4.59561855e-01 1.15787947e+00
5.36151230e-01 6.11017048e-01 3.79975528e-01 1.05620861e+00
-2.32832313e-01 1.54186234e-01 -4.96517599e-01 1.66558623e-02
7.88181067e-01 1.47486603e+00 -1.00399601e+00 -1.80238843e-01
-1.56006634e-01 6.83503628e-01 4.61213887e-01 3.01907063e-01
-7.94240952e-01 -1.15672433e+00 7.96783864e-01 1.19776271e-01
2.81529546e-01 -6.17321610e-01 -3.69011521e-01 -7.76208520e-01
-3.97186819e-03 -7.93922603e-01 3.28707516e-01 -7.34334052e-01
-7.36043751e-01 -2.86504805e-01 -2.02976063e-01 -7.15534925e-01
1.41081542e-01 -8.41038764e-01 -1.06753218e+00 8.88601184e-01
-1.57208097e+00 -1.17224120e-01 -5.30556142e-01 4.88306731e-01
1.09936133e-01 3.42759669e-01 8.28933358e-01 3.42227548e-01
-5.69885850e-01 9.34668899e-01 -1.66491792e-01 -2.69741595e-01
7.97263622e-01 -1.80219531e+00 2.86695898e-01 9.90014672e-01
2.38137394e-01 1.00287032e+00 8.79717588e-01 -3.53257954e-01
-1.79984021e+00 -4.94360238e-01 1.26861739e+00 -4.59876835e-01
8.58461320e-01 -3.57279867e-01 -7.31155157e-01 4.93990034e-01
3.25335562e-01 -1.11671075e-01 8.76337528e-01 2.17319578e-01
-3.95618170e-01 -1.53520808e-01 -1.03621149e+00 5.45804143e-01
1.13617814e+00 -2.12245673e-01 -3.52285683e-01 6.82711840e-01
6.98936582e-01 -8.32818866e-01 -8.12152922e-01 -3.41045469e-01
5.36356926e-01 -5.53853929e-01 7.87429512e-01 -8.94121647e-01
7.25850880e-01 -3.41555566e-01 1.38735265e-01 -1.14840746e+00
-7.08946064e-02 -1.12606430e+00 -1.90271080e-01 1.20406902e+00
3.26959997e-01 -2.89009899e-01 1.02381182e+00 9.29826438e-01
-2.42708847e-01 -7.66247153e-01 -3.86613786e-01 -7.18949914e-01
2.20195249e-01 -4.54493284e-01 7.45518148e-01 1.21863735e+00
5.74694693e-01 -1.00196622e-01 5.36403835e-01 3.07508141e-01
1.02930665e+00 6.53022349e-01 9.65051711e-01 -1.55174947e+00
-3.55443776e-01 -8.63171339e-01 -2.16762021e-01 -1.11304009e+00
3.77775788e-01 -1.36146438e+00 -1.75850064e-01 -1.76703334e+00
-3.41381738e-03 -9.00964141e-01 -2.97711253e-01 4.81650740e-01
1.12609632e-01 -1.82454824e-01 1.53825194e-01 -6.71300530e-01
-9.20127273e-01 -4.03339893e-01 1.40118659e+00 2.51581427e-02
1.26017675e-01 8.82621389e-03 -7.15761840e-01 6.93481565e-01
5.21036983e-01 -7.42381334e-01 -5.14112353e-01 -3.60059261e-01
8.39780748e-01 1.64294720e-01 3.36017519e-01 -9.40036714e-01
8.03568125e-01 -7.77555406e-01 -1.10318579e-01 -5.31845868e-01
-3.88408065e-01 -9.20663834e-01 -3.07464898e-01 5.68762302e-01
-5.74949741e-01 3.86269540e-01 4.51061755e-01 2.74090618e-01
2.43186593e-01 -7.77337372e-01 4.11113828e-01 -9.38994214e-02
-3.28214288e-01 1.87248021e-01 -4.38916296e-01 2.71672070e-01
1.26009297e+00 -1.80843264e-01 -7.55464077e-01 -2.88037300e-01
-4.74182755e-01 7.15600014e-01 4.86067086e-01 9.20731649e-02
5.97292781e-01 -1.06857967e+00 -2.32363284e-01 1.33093148e-01
-3.28918666e-01 -2.11695224e-01 1.31935682e-02 9.01218474e-01
-1.16060185e+00 5.21824837e-01 -2.13392824e-01 -3.54222417e-01
-2.08783460e+00 4.14281517e-01 -1.25146270e-01 -4.53094691e-01
-3.73275250e-01 1.33830726e+00 -4.71038997e-01 -6.52469099e-01
7.19424784e-01 -6.48391664e-01 3.09496999e-01 -1.32374406e-01
7.20288813e-01 6.16417170e-01 2.94509858e-01 6.42247975e-01
-3.34683090e-01 3.02217275e-01 -1.94391400e-01 1.53384194e-01
1.51588476e+00 7.38635333e-03 -2.25015789e-01 -1.90436803e-02
9.19880927e-01 5.91729760e-01 -4.65953112e-01 1.50149837e-01
1.95611473e-02 -4.70463753e-01 -1.16518877e-01 -8.72134984e-01
-7.35334396e-01 9.83737230e-01 7.72740394e-02 4.29378092e-01
7.88765490e-01 -5.46988010e-01 8.36473525e-01 7.17153132e-01
5.49493432e-01 -1.01644123e+00 1.40408009e-01 6.29107833e-01
1.63972244e-01 -8.72089446e-01 2.58987129e-01 -6.55448854e-01
2.43184924e-01 1.63099682e+00 7.26953149e-01 -1.12760374e-02
4.50540185e-01 6.16320372e-01 -4.45490986e-01 -8.77188072e-02
-1.31445622e+00 4.27617908e-01 1.71516277e-02 2.25030452e-01
6.71825051e-01 1.61239401e-01 -3.60905647e-01 3.99331838e-01
-5.57159066e-01 1.10880472e-01 1.39191580e+00 1.31982303e+00
-7.10722029e-01 -1.75869966e+00 -5.08435130e-01 1.48853824e-01
-4.58941072e-01 -2.90122122e-01 -6.10176146e-01 6.26847386e-01
-6.28902093e-02 8.90829682e-01 -2.54372925e-01 -8.91318321e-02
1.54443040e-01 2.00263903e-01 1.25359631e+00 -8.37948263e-01
-6.81691170e-01 -8.68901312e-01 -1.69559926e-01 -4.17486906e-01
1.37548417e-01 -6.61973178e-01 -1.57139456e+00 -1.22660220e+00
-3.87416214e-01 5.65639853e-01 8.18157971e-01 6.75363004e-01
2.09358707e-01 8.25024486e-01 3.42975706e-01 -3.94178890e-02
-7.73009241e-01 -6.16801046e-02 -3.29335064e-01 -1.68651983e-01
2.63790160e-01 -4.05502260e-01 -3.41726780e-01 6.34076213e-03] | [9.680899620056152, 7.324916362762451] |
11fc6db3-68fd-458d-902f-b0d197c45ded | a-recipe-for-efficient-sbir-models-combining | 2305.18988 | null | https://arxiv.org/abs/2305.18988v1 | https://arxiv.org/pdf/2305.18988v1.pdf | A Recipe for Efficient SBIR Models: Combining Relative Triplet Loss with Batch Normalization and Knowledge Distillation | Sketch-Based Image Retrieval (SBIR) is a crucial task in multimedia retrieval, where the goal is to retrieve a set of images that match a given sketch query. Researchers have already proposed several well-performing solutions for this task, but most focus on enhancing embedding through different approaches such as triplet loss, quadruplet loss, adding data augmentation, and using edge extraction. In this work, we tackle the problem from various angles. We start by examining the training data quality and show some of its limitations. Then, we introduce a Relative Triplet Loss (RTL), an adapted triplet loss to overcome those limitations through loss weighting based on anchors similarity. Through a series of experiments, we demonstrate that replacing a triplet loss with RTL outperforms previous state-of-the-art without the need for any data augmentation. In addition, we demonstrate why batch normalization is more suited for SBIR embeddings than l2-normalization and show that it improves significantly the performance of our models. We further investigate the capacity of models required for the photo and sketch domains and demonstrate that the photo encoder requires a higher capacity than the sketch encoder, which validates the hypothesis formulated in [34]. Then, we propose a straightforward approach to train small models, such as ShuffleNetv2 [22] efficiently with a marginal loss of accuracy through knowledge distillation. The same approach used with larger models enabled us to outperform previous state-of-the-art results and achieve a recall of 62.38% at k = 1 on The Sketchy Database [30]. | ['Thierry Dutoit', 'Stéphane Dupont', 'Nathan Hubens', 'Omar Seddati'] | 2023-05-30 | null | null | null | null | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 2.02255011e-01 -2.44670421e-01 -3.18722874e-01 -1.56839982e-01
-1.11337924e+00 -6.32340372e-01 9.21093643e-01 1.16568498e-01
-6.29469633e-01 3.96072447e-01 1.78942665e-01 -5.00364192e-02
-2.82795608e-01 -7.04796433e-01 -8.38355184e-01 -4.83978570e-01
-5.91553710e-02 2.20899194e-01 2.72342205e-01 -2.74314880e-01
4.52865660e-01 7.65216947e-01 -1.67425036e+00 3.15971136e-01
2.78612763e-01 1.17247188e+00 7.35886618e-02 6.93295360e-01
-2.26558790e-01 5.08552074e-01 -4.59044427e-01 -8.07219446e-01
4.08279657e-01 -4.72336337e-02 -7.40840197e-01 -1.91142514e-01
1.00541627e+00 -6.53477728e-01 -7.35137343e-01 5.69647253e-01
9.28972602e-01 8.31765905e-02 8.51176739e-01 -1.38269317e+00
-8.16061616e-01 1.86370850e-01 -3.36031139e-01 -1.34001806e-01
3.36540043e-01 -3.73104095e-01 1.27789652e+00 -1.38406920e+00
7.63919711e-01 1.24097729e+00 6.76206172e-01 5.88702023e-01
-1.15735483e+00 -6.20326638e-01 1.79976411e-02 5.97720504e-01
-1.85235190e+00 -5.06118834e-01 9.41169500e-01 -5.93352541e-02
8.67704868e-01 2.46818691e-01 2.74660230e-01 1.07099915e+00
-3.51568282e-01 1.07969439e+00 7.13436723e-01 -7.76580870e-01
-2.03739069e-02 2.84928769e-01 -1.15529865e-01 7.29833901e-01
1.73464939e-01 -1.54979378e-01 -5.57251930e-01 -2.62366235e-01
9.07845616e-01 1.21064588e-01 -2.48908028e-01 -8.29479575e-01
-1.10094166e+00 8.67804468e-01 5.21673560e-01 4.91644204e-01
-9.67227519e-02 4.75714743e-01 4.93528098e-01 4.33208108e-01
2.80595928e-01 3.27353776e-01 -2.18056858e-01 1.10948354e-01
-1.27503192e+00 1.42212078e-01 7.54056692e-01 1.14342797e+00
7.38119364e-01 -4.34831887e-01 -3.92277032e-01 1.14611447e+00
2.61166513e-01 4.59127367e-01 2.68291861e-01 -9.63876307e-01
4.40634847e-01 3.45226914e-01 8.61144799e-04 -1.04143620e+00
1.79637745e-01 -8.99510011e-02 -6.90684795e-01 1.35241672e-01
2.95703948e-01 5.08328557e-01 -1.03358662e+00 1.71431172e+00
-5.35318032e-02 2.23263547e-01 -8.30262080e-02 8.76688540e-01
8.61352563e-01 4.87709641e-01 -6.62661567e-02 1.96369782e-01
1.39518249e+00 -1.11809027e+00 -6.62015080e-01 2.27291331e-01
5.13462245e-01 -9.52032864e-01 1.24283195e+00 3.10602397e-01
-1.18799376e+00 -5.46879053e-01 -1.21898103e+00 -4.52873290e-01
-7.11417377e-01 4.12785560e-01 5.94217479e-01 7.74710596e-01
-1.42921782e+00 6.67758107e-01 -4.54719096e-01 -5.64707398e-01
3.85285825e-01 2.80228645e-01 -6.75927222e-01 -4.44929242e-01
-1.10222983e+00 8.61506402e-01 -3.09815485e-04 -2.30499879e-01
-6.26869678e-01 -7.27930546e-01 -7.98676193e-01 3.68290544e-01
3.36847216e-01 -8.17226231e-01 8.36999297e-01 -5.88174045e-01
-1.44058752e+00 9.64997768e-01 -2.00158745e-01 -3.43631715e-01
6.08238518e-01 -1.56918511e-01 -2.02857569e-01 4.52899992e-01
-2.22106934e-01 1.08001828e+00 8.99030507e-01 -1.35354042e+00
-2.65600145e-01 -2.50468343e-01 3.49811256e-01 5.73903173e-02
-7.38209903e-01 -4.99290340e-02 -1.08368838e+00 -8.44354510e-01
-1.67399153e-01 -9.33851063e-01 1.33689195e-01 5.91091454e-01
2.80135963e-02 -3.03478360e-01 7.49907911e-01 -5.27856827e-01
1.37934995e+00 -2.16346598e+00 2.21521743e-02 3.18358839e-01
-1.99848749e-02 4.98543978e-01 -5.97667694e-01 9.50989962e-01
-4.77497932e-04 3.62771332e-01 -2.83697754e-01 -7.05112636e-01
3.24442387e-01 2.16155782e-01 -4.35489237e-01 1.18392825e-01
2.97545373e-01 1.00865602e+00 -6.33507609e-01 -6.63933277e-01
2.25738525e-01 1.11061239e+00 -6.36722505e-01 1.51461378e-01
1.12622552e-01 -3.12766224e-01 -1.75462607e-02 6.90069973e-01
7.43884563e-01 -2.90828407e-01 4.94005829e-02 -4.82345670e-01
1.97160169e-01 3.09219975e-02 -1.29346347e+00 2.00536919e+00
-6.20821238e-01 7.88788915e-01 4.75243144e-02 -9.59042668e-01
8.55129659e-01 3.83259714e-01 4.10609424e-01 -8.03592086e-01
-2.03821182e-01 3.44346344e-01 -5.60409009e-01 -2.84056127e-01
7.70997584e-01 8.95359814e-02 1.52177483e-01 3.50103736e-01
2.78516591e-01 8.77292454e-02 1.75074577e-01 4.07004893e-01
1.14448285e+00 1.13588028e-01 1.46880827e-03 9.51750353e-02
6.88696265e-01 -4.40935493e-01 -1.61838323e-01 8.32042873e-01
8.29740055e-03 9.51550305e-01 2.44345292e-01 -2.25506708e-01
-1.33585572e+00 -1.07854080e+00 -7.86835775e-02 1.06706572e+00
1.60801589e-01 -6.70095205e-01 -3.50411743e-01 -7.11364627e-01
1.85880601e-01 3.94064099e-01 -6.01447046e-01 -5.54916728e-03
-5.68190873e-01 -3.99145037e-01 6.83175921e-01 5.04007578e-01
5.05539179e-01 -8.24932754e-01 -1.21361151e-01 -8.12896416e-02
-1.03206649e-01 -1.21808875e+00 -5.61336637e-01 -3.76830667e-01
-7.37193584e-01 -9.71801758e-01 -1.31453979e+00 -8.21776748e-01
6.17399454e-01 5.74280977e-01 1.10314095e+00 3.72706711e-01
-5.13141513e-01 1.02571058e+00 -4.84021872e-01 -1.47143289e-01
1.88384075e-02 2.24592566e-01 -3.10565919e-01 -1.29911199e-01
2.85881370e-01 -5.53466558e-01 -9.45588231e-01 4.09101933e-01
-1.31649041e+00 -3.14425915e-01 8.55002642e-01 8.32550526e-01
5.64766049e-01 -3.96686405e-01 4.10304368e-01 -4.93172348e-01
6.03582919e-01 -5.96737526e-02 -3.24134767e-01 6.08714581e-01
-7.82178044e-01 3.15687388e-01 2.86783904e-01 -5.31540334e-01
-6.17445230e-01 -8.37673694e-02 -2.03696162e-01 -7.76771367e-01
1.13593012e-01 1.91140890e-01 2.44170111e-02 -4.99412864e-01
3.87759954e-01 3.06674391e-01 1.53104961e-02 -6.10185802e-01
6.68575168e-01 6.58984780e-01 1.85883000e-01 -6.22261763e-01
7.97531724e-01 6.38268232e-01 1.92532778e-01 -8.48980784e-01
-4.29911345e-01 -6.71965003e-01 -4.89341021e-01 -7.63361948e-03
4.79196310e-01 -9.09436822e-01 -7.74233460e-01 9.27092955e-02
-1.08592820e+00 1.26789743e-02 -2.65820831e-01 3.49300236e-01
-6.01931036e-01 7.54352212e-01 -5.31315446e-01 -8.60760689e-01
-5.14488697e-01 -9.09685612e-01 1.43540490e+00 -2.03692988e-01
2.49701515e-01 -9.16362464e-01 5.98393269e-02 3.24267060e-01
6.88630581e-01 -5.37232645e-02 8.35072935e-01 -5.97621202e-01
-8.50767016e-01 -2.46759176e-01 -6.00349665e-01 4.33776319e-01
-1.12226672e-01 -1.26579583e-01 -1.03033721e+00 -5.39043009e-01
-5.36257386e-01 -4.10459667e-01 1.28665984e+00 -9.08794180e-02
1.27610147e+00 -2.20042005e-01 -2.25683615e-01 5.39860487e-01
1.77853572e+00 -2.85723209e-01 9.51256633e-01 3.29490393e-01
4.96185839e-01 4.54532504e-01 4.63792503e-01 2.03843832e-01
4.08574671e-01 1.07348359e+00 3.27765614e-01 -1.74256876e-01
-5.74696660e-01 -4.50276852e-01 2.45796755e-01 5.99992156e-01
-6.05434775e-02 -2.86456913e-01 -4.74802107e-01 6.90380156e-01
-1.94197750e+00 -8.78044367e-01 3.03304046e-01 2.46962261e+00
6.64337456e-01 -2.62328684e-01 1.67215541e-01 3.38104546e-01
4.53179240e-01 2.80563235e-01 -1.04933411e-01 -3.09292734e-01
-1.56651810e-01 4.31584924e-01 6.08827353e-01 5.52315533e-01
-9.78104293e-01 9.15383577e-01 6.56945705e+00 1.20965111e+00
-9.53425109e-01 -1.18297953e-02 3.07262152e-01 -1.00875214e-01
-4.24672693e-01 -2.18153764e-02 -6.85927808e-01 2.63631880e-01
6.15136921e-01 9.96120721e-02 4.95305181e-01 6.19135022e-01
-3.94013911e-01 1.48637921e-01 -1.28134632e+00 1.29169965e+00
6.06636345e-01 -1.28527653e+00 4.39230770e-01 -7.31441677e-02
4.46912050e-01 -2.19648182e-01 8.38666409e-02 3.43557149e-01
-3.41740161e-01 -1.02515304e+00 4.49426472e-01 5.89619458e-01
1.04534948e+00 -5.53112268e-01 6.15825891e-01 -1.36768699e-01
-1.30284405e+00 1.21262878e-01 -4.73851681e-01 2.76658744e-01
6.00434318e-02 2.96051353e-01 -7.59940803e-01 7.50318348e-01
4.57423508e-01 5.05947053e-01 -7.37784147e-01 1.11328876e+00
-1.00336358e-01 1.66300818e-01 -3.86024415e-01 -6.70319051e-03
2.06897900e-01 -2.37486623e-02 3.57756704e-01 1.43011165e+00
4.51361388e-01 -2.05086678e-01 -2.69686401e-01 5.61903715e-01
-4.03551131e-01 2.77949065e-01 -6.76993251e-01 6.13182783e-02
4.63479459e-01 1.31034815e+00 -3.40332180e-01 -3.69738489e-01
-5.46794057e-01 1.31072903e+00 3.92764956e-01 4.87473518e-01
-6.89166605e-01 -7.12792575e-01 4.92127836e-01 1.30774230e-01
8.02154422e-01 -2.56717026e-01 1.88891754e-01 -1.16025019e+00
3.86499226e-01 -5.53685188e-01 3.54514629e-01 -8.86445940e-01
-1.35407221e+00 3.37514758e-01 3.20482664e-02 -1.27794218e+00
-1.18115790e-01 -7.91761518e-01 -2.13553846e-01 7.15721130e-01
-2.14931250e+00 -1.44921172e+00 -2.72014439e-01 5.76867580e-01
3.29477698e-01 -5.40237315e-02 9.95296001e-01 9.04880881e-01
-9.54112485e-02 1.16836345e+00 9.95056778e-02 2.03694135e-01
1.23174250e+00 -1.06577361e+00 2.70769387e-01 3.69496942e-01
4.71010715e-01 6.25573218e-01 3.55505675e-01 -1.48624644e-01
-1.52821159e+00 -8.33763301e-01 1.17670262e+00 -4.23176408e-01
5.75553358e-01 -4.86555606e-01 -8.38032246e-01 4.38032687e-01
1.67138159e-01 1.81917846e-01 5.61883807e-01 -5.73996343e-02
-8.24406326e-01 -2.65482277e-01 -1.20947993e+00 6.30554616e-01
1.17478788e+00 -8.27950835e-01 -4.03467536e-01 1.68250307e-01
6.86087549e-01 -6.82264417e-02 -1.06856251e+00 4.85769033e-01
9.81612027e-01 -1.00053191e+00 1.53907812e+00 -5.11276782e-01
3.72387499e-01 -1.17314950e-01 -3.84728283e-01 -8.30919802e-01
-2.96698600e-01 -3.36535990e-01 -1.46278217e-01 1.29953599e+00
2.09388316e-01 -3.89152110e-01 8.77697825e-01 5.65876424e-01
2.57708818e-01 -8.05257440e-01 -8.70461881e-01 -1.00884819e+00
5.97658902e-02 -2.60752112e-01 4.04986501e-01 8.01725090e-01
-2.41085917e-01 1.14102013e-01 -3.99118721e-01 -4.94446270e-02
5.58683813e-01 1.91845763e-02 9.06974971e-01 -9.85783875e-01
3.41945291e-02 -6.24153674e-01 -6.66874707e-01 -1.26669145e+00
1.11109778e-01 -8.48998070e-01 -4.78730261e-01 -1.70888746e+00
3.60319108e-01 -5.24078488e-01 -5.03377438e-01 4.24593061e-01
-9.00949612e-02 6.74495518e-01 5.57148933e-01 1.84864193e-01
-7.79156089e-01 5.88593960e-01 9.78764534e-01 -3.22476208e-01
1.73741668e-01 -3.63770127e-01 -4.77811873e-01 2.29492739e-01
4.40285683e-01 -2.23230228e-01 -3.58570784e-01 -4.86778975e-01
2.76161253e-01 -2.91443467e-01 5.69347680e-01 -8.18081737e-01
3.89059782e-01 4.60272312e-01 2.61799961e-01 -5.01244366e-01
7.91015506e-01 -1.12414277e+00 -9.06000435e-02 1.20631613e-01
-3.88507813e-01 6.59578219e-02 2.26553321e-01 5.71584046e-01
-4.29603130e-01 -3.46546024e-01 3.55676383e-01 9.31448340e-02
-7.31506109e-01 3.94527942e-01 1.89091489e-01 -1.35861650e-01
7.83618212e-01 -1.47031754e-01 -2.83507615e-01 -5.26012421e-01
-5.22182286e-01 1.08161390e-01 5.26137114e-01 5.29421628e-01
8.34867239e-01 -1.74698603e+00 -7.42198229e-01 1.47606000e-01
4.69443709e-01 -5.01191974e-01 1.59178510e-01 6.05148077e-01
-3.74790967e-01 6.05626822e-01 -1.21420503e-01 -5.04272282e-01
-1.47862780e+00 7.40653515e-01 5.19553944e-02 -3.97300780e-01
-4.17817801e-01 6.25381947e-01 -1.07493907e-01 -3.44052583e-01
5.78685343e-01 -7.78137706e-03 -7.73573220e-02 1.90380514e-01
5.28530717e-01 4.64832127e-01 3.19137126e-01 -3.34430754e-01
-4.24810022e-01 9.88983035e-01 -8.41315091e-02 -2.92384446e-01
1.29513502e+00 -1.20864920e-02 1.51542231e-01 9.91679728e-02
1.67740870e+00 -1.87095664e-02 -9.47215438e-01 -3.93394977e-01
-1.35986671e-01 -7.24005103e-01 1.84757505e-02 -6.79327607e-01
-9.29062009e-01 1.04572928e+00 7.40092754e-01 1.05863467e-01
1.15149367e+00 1.01627558e-02 8.04354727e-01 6.06120527e-01
3.31714243e-01 -9.69376624e-01 2.78258771e-01 2.57331401e-01
1.09209859e+00 -1.33773541e+00 1.19209021e-01 -3.17123085e-01
-3.35330546e-01 1.08448625e+00 2.00740397e-02 -2.77973920e-01
6.45548046e-01 -8.72722715e-02 -1.82278886e-01 -5.17867394e-02
-6.43319428e-01 -3.48028034e-01 5.56168795e-01 3.55454504e-01
4.53741223e-01 -3.18981409e-01 -4.81114656e-01 8.81480873e-02
2.32847467e-01 1.02799512e-01 2.72537749e-02 8.37830365e-01
-1.20048963e-01 -1.58712101e+00 -2.49278530e-01 2.13036135e-01
-4.93070871e-01 -3.39715719e-01 -5.03276110e-01 9.82725620e-01
-2.57805139e-01 6.62452638e-01 1.54158762e-02 -3.53094637e-01
5.19169390e-01 2.25533649e-01 9.46424901e-01 -8.81411359e-02
-3.79333586e-01 -1.59983203e-01 -1.90118253e-02 -6.23448551e-01
-6.67549908e-01 -2.44410202e-01 -6.77266717e-01 -3.39568108e-01
-3.37732553e-01 5.46943815e-03 1.00374424e+00 4.66545314e-01
6.54426873e-01 1.97954983e-01 5.78982472e-01 -9.29334641e-01
-5.56994796e-01 -7.85570681e-01 -3.67598832e-01 6.06140316e-01
3.27304125e-01 -5.84920704e-01 -4.56475109e-01 -4.42032963e-02] | [11.413434982299805, 0.6259356141090393] |
af599f7c-d442-422f-b8d2-6e13de7d185f | geometric-feature-based-facial-expression | 1604.03225 | null | http://arxiv.org/abs/1604.03225v1 | http://arxiv.org/pdf/1604.03225v1.pdf | Geometric Feature-Based Facial Expression Recognition in Image Sequences Using Multi-Class AdaBoost and Support Vector Machines | Facial expressions are widely used in the behavioral interpretation of
emotions, cognitive science, and social interactions. In this paper, we present
a novel method for fully automatic facial expression recognition in facial
image sequences. As the facial expression evolves over time facial landmarks
are automatically tracked in consecutive video frames, using displacements
based on elastic bunch graph matching displacement estimation. Feature vectors
from individual landmarks, as well as pairs of landmarks tracking results are
extracted, and normalized, with respect to the first frame in the sequence. The
prototypical expression sequence for each class of facial expression is formed,
by taking the median of the landmark tracking results from the training facial
expression sequences. Multi-class AdaBoost with dynamic time warping similarity
distance between the feature vector of input facial expression and prototypical
facial expression, is used as a weak classifier to select the subset of
discriminative feature vectors. Finally, two methods for facial expression
recognition are presented, either by using multi-class AdaBoost with dynamic
time warping, or by using support vector machine on the boosted feature
vectors. The results on the Cohn-Kanade (CK+) facial expression database show a
recognition accuracy of 95.17% and 97.35% using multi-class AdaBoost and
support vector machines, respectively. | ['Joonwhoan Lee', 'Deepak Ghimire'] | 2016-04-12 | null | null | null | null | ['landmark-tracking'] | ['computer-vision'] | [-1.26297716e-02 -5.15876651e-01 -2.86835313e-01 -7.05294907e-01
-4.66477394e-01 -3.53444785e-01 4.26487803e-01 -2.06007466e-01
-7.11297333e-01 4.18582380e-01 8.22572485e-02 6.79253101e-01
-2.85624359e-02 -3.40742886e-01 -1.66353971e-01 -1.30634785e+00
-3.75496894e-01 1.20388001e-01 -1.85300454e-01 -3.79873484e-01
7.74160177e-02 1.14814591e+00 -1.64539123e+00 3.31508875e-01
-4.68052477e-02 1.27165151e+00 -4.60064888e-01 6.00504398e-01
-6.52316362e-02 7.18710244e-01 -5.05492687e-01 -2.33375326e-01
1.26353547e-01 -7.22395301e-01 -6.30019307e-01 3.12339157e-01
4.92958069e-01 -8.72964188e-02 -2.16716174e-02 1.12852049e+00
4.18406218e-01 4.22765583e-01 6.89062119e-01 -1.71628487e+00
-3.00994813e-01 -3.36936027e-01 -9.30992007e-01 2.96926886e-01
7.22684383e-01 -3.85820925e-01 5.52087426e-01 -1.19357955e+00
1.15361595e+00 1.25819528e+00 5.99071026e-01 7.84947276e-01
-1.18892896e+00 -9.91783977e-01 -1.28804550e-01 5.96626341e-01
-1.92019212e+00 -5.65368533e-01 1.13555014e+00 -6.28148198e-01
7.65327990e-01 3.24048847e-01 1.21213019e+00 7.40847111e-01
2.34745294e-01 6.38252258e-01 1.24349773e+00 -5.33129394e-01
1.72768116e-01 -1.74524590e-01 4.96881828e-02 1.02578890e+00
-8.52369070e-01 1.59590156e-04 -7.00830400e-01 -5.64168990e-01
4.39247161e-01 1.29293576e-01 -4.99122627e-02 -1.11773491e-01
-5.62593937e-01 7.85991728e-01 1.81880996e-01 8.26521933e-01
-7.14863837e-01 -4.93689906e-03 5.12369931e-01 3.97417992e-01
8.40598047e-01 -4.47152972e-01 -3.56145985e-02 -1.71476677e-01
-1.09392238e+00 2.88813144e-01 4.04638618e-01 5.10902107e-01
1.22004259e+00 6.26797527e-02 1.49344385e-01 8.84044588e-01
3.54538351e-01 5.83425939e-01 8.38737428e-01 -9.72315788e-01
-4.79079634e-01 5.53765893e-01 -3.47952813e-01 -1.59847891e+00
-5.89817345e-01 7.37379253e-01 -6.42036796e-01 5.90988338e-01
3.44628155e-01 -6.06032684e-02 -6.23096883e-01 1.76093936e+00
7.01157570e-01 5.05008936e-01 -6.00358360e-02 1.05598748e+00
5.22433519e-01 7.87853301e-01 2.18671888e-01 -1.02680194e+00
1.37136579e+00 -6.32876158e-01 -1.05550659e+00 4.91070867e-01
8.58438671e-01 -6.20653808e-01 4.26609427e-01 2.54699796e-01
-7.48921156e-01 -5.37866890e-01 -7.05378234e-01 2.61560023e-01
-2.65408814e-01 -9.66908336e-02 4.21052635e-01 5.09540439e-01
-1.11035848e+00 5.28125882e-01 -8.41043591e-01 -4.14008170e-01
2.93171704e-01 4.67139155e-01 -1.06963134e+00 3.58813286e-01
-1.03922522e+00 8.73295188e-01 -1.76781625e-01 1.83481574e-01
-4.18856114e-01 -2.66015649e-01 -1.06776023e+00 -4.26300228e-01
-3.87448013e-01 1.37323260e-01 9.59095895e-01 -2.15748549e+00
-1.75623500e+00 1.42588794e+00 -7.09401250e-01 -1.28037222e-02
1.28276914e-01 1.74957037e-01 -6.67382538e-01 4.25878793e-01
-8.51533785e-02 5.70163250e-01 1.16699100e+00 -7.61398375e-01
-3.72847319e-01 -9.78110015e-01 -7.13638067e-01 1.91921800e-01
-2.59548485e-01 7.85833895e-01 -1.91011980e-01 -2.93379128e-01
1.97219998e-01 -1.25459659e+00 -3.74532454e-02 3.66339982e-01
3.53078574e-01 -6.58686280e-01 1.22886002e+00 -6.59415543e-01
1.13654780e+00 -2.50229001e+00 2.59021461e-01 5.69594085e-01
-1.27322122e-01 2.63945702e-02 -1.78324699e-01 -1.50620550e-01
-5.47282279e-01 -3.52854520e-01 4.60566990e-02 -2.79762745e-01
-4.54140007e-01 3.14954937e-01 5.69849052e-02 1.11341739e+00
7.38206208e-02 6.18253827e-01 -8.78352880e-01 -7.72945046e-01
1.85449496e-02 5.49043238e-01 -6.82640746e-02 9.28093120e-02
3.49585742e-01 4.37173665e-01 -2.95703590e-01 8.81731212e-01
5.93580544e-01 5.08413136e-01 1.63006648e-01 -2.75975347e-01
-1.12745546e-01 -9.27718222e-01 -9.86909986e-01 1.63076925e+00
-4.22751196e-02 9.24070239e-01 1.61811918e-01 -1.02144539e+00
1.30227244e+00 5.97657502e-01 9.73152399e-01 -7.69350410e-01
4.47072148e-01 6.92904368e-02 -2.60596335e-01 -8.68690789e-01
1.58703849e-01 -5.58990717e-01 3.92000340e-02 3.58352542e-01
3.06776732e-01 1.19826213e-01 1.86014503e-01 -4.36539054e-01
5.13199389e-01 1.85136378e-01 6.71132028e-01 -1.99690491e-01
8.09399843e-01 -2.30656654e-01 6.01991296e-01 -2.76785821e-01
-6.59679353e-01 4.18575585e-01 3.97206694e-01 -1.03631210e+00
-7.68938243e-01 -6.27159655e-01 -2.03666210e-01 1.39254773e+00
-1.49649680e-01 -2.17951134e-01 -9.65010703e-01 -5.51803291e-01
3.18794996e-02 1.65971592e-01 -9.15284038e-01 -3.91803607e-02
-7.26538539e-01 -6.34317160e-01 6.74151838e-01 3.37432832e-01
2.26251185e-01 -1.13938439e+00 -5.10210752e-01 1.37315437e-01
4.40782718e-02 -8.53882194e-01 -3.23421091e-01 -3.87100905e-01
-5.99566638e-01 -9.69230771e-01 -8.58190000e-01 -8.51312160e-01
7.26605117e-01 2.10430026e-02 2.71105289e-01 6.58668131e-02
-5.28039098e-01 6.20921016e-01 -2.31825531e-01 -2.07790926e-01
-3.35134238e-01 -8.63061547e-01 5.41443527e-01 8.96836400e-01
7.78629124e-01 -5.42890668e-01 -3.04191738e-01 3.61775428e-01
-6.52642250e-01 -4.24963295e-01 -1.46982959e-02 7.44722903e-01
8.71458948e-01 -4.66368854e-01 7.65384436e-02 -1.58488095e-01
3.20385039e-01 -3.80353481e-01 -3.70815277e-01 1.05313249e-01
-2.43927002e-01 -3.95229399e-01 2.51859695e-01 -7.58212090e-01
-7.28669941e-01 5.23483574e-01 -2.67478883e-01 -9.25824463e-01
-1.69461027e-01 3.40588331e-01 2.12547183e-01 -4.59614545e-01
9.02654767e-01 3.45026106e-01 7.78137207e-01 1.50199383e-01
2.77184069e-01 5.96470535e-01 4.13060069e-01 -1.72296852e-01
1.77176580e-01 8.32282901e-01 2.20924526e-01 -1.33403420e+00
-3.22789669e-01 -6.58687770e-01 -9.82926011e-01 -9.45194125e-01
1.21174955e+00 -7.08216131e-01 -9.20030296e-01 8.07262897e-01
-1.19385600e+00 9.98362228e-02 9.74328369e-02 5.42271376e-01
-7.87922442e-01 5.42519033e-01 -4.78286773e-01 -1.01116228e+00
-3.86944979e-01 -9.40990984e-01 1.07209778e+00 3.19573611e-01
-7.01461494e-01 -9.63574886e-01 5.31202555e-01 9.02901590e-03
-5.42260967e-02 6.93911493e-01 4.51928943e-01 -5.58137894e-01
1.87330753e-01 -5.51241815e-01 3.90697867e-01 4.85236019e-01
1.85759157e-01 7.28656471e-01 -1.06098390e+00 -8.26470777e-02
1.48644775e-01 -3.12222838e-01 2.46855676e-01 4.23972696e-01
7.34935820e-01 -1.94288567e-01 -3.76875997e-01 5.83895087e-01
1.02355421e+00 6.34692252e-01 5.36270797e-01 8.05668384e-02
3.41676235e-01 8.67885947e-01 6.95906162e-01 5.70520043e-01
-9.69144031e-02 1.08035779e+00 1.84738524e-02 -5.17972447e-02
4.45073009e-01 1.77123636e-01 5.75982869e-01 5.99900901e-01
-6.49253368e-01 5.92615604e-01 -7.49191523e-01 1.54738694e-01
-1.81615305e+00 -1.38682020e+00 2.49535173e-01 1.86261439e+00
5.72121561e-01 -5.33785343e-01 3.62871319e-01 2.00546503e-01
7.58558154e-01 2.40873665e-01 -2.43019998e-01 -6.64417684e-01
-3.64645869e-01 3.19754362e-01 -6.33785650e-02 5.40436029e-01
-1.10092592e+00 8.94610345e-01 5.71287727e+00 7.40253627e-01
-1.62489629e+00 1.91869453e-01 5.78217149e-01 -2.02104896e-01
4.31387395e-01 -2.79505610e-01 -5.25219619e-01 2.08709091e-01
8.97740722e-01 -5.57972252e-01 1.61895454e-01 1.09841502e+00
3.90599519e-01 -1.89700007e-01 -7.74756908e-01 1.50919342e+00
5.92621803e-01 -1.04793680e+00 -1.13113314e-01 -2.22408548e-01
4.93084490e-01 -3.81022573e-01 -5.71308471e-02 -4.40381318e-02
-4.21860993e-01 -8.11294973e-01 6.17216587e-01 7.79048324e-01
7.00577855e-01 -8.55761111e-01 4.24387068e-01 -3.64251323e-02
-1.28839540e+00 1.12975627e-01 -1.60806224e-01 -1.57660365e-01
2.02153847e-01 6.89810142e-02 -4.13839191e-01 7.05901086e-02
7.64117122e-01 8.17853868e-01 -5.56736737e-02 3.91033947e-01
8.31930414e-02 2.76732862e-01 -3.06842774e-01 -2.66088784e-01
1.88359559e-01 -5.57154238e-01 4.54612464e-01 1.41464460e+00
3.95512193e-01 5.79859614e-01 -2.80802865e-02 2.49323621e-01
3.60851377e-01 6.97270036e-01 -9.30015504e-01 2.65545421e-03
3.85040827e-02 1.70301914e+00 -6.08381927e-01 -2.81077534e-01
-4.31592584e-01 1.15131974e+00 2.50253260e-01 3.14984918e-01
-6.66212142e-01 -9.13464725e-02 1.17042255e+00 -2.57434007e-02
-2.67184198e-01 -2.27421463e-01 4.29833561e-01 -7.61448622e-01
-1.75158009e-01 -5.40692389e-01 5.21137416e-01 -8.33367944e-01
-8.95008743e-01 7.96845913e-01 1.86526775e-02 -1.21136820e+00
-6.93792582e-01 -6.44664824e-01 -6.20860398e-01 7.04905808e-01
-9.99645531e-01 -1.09779096e+00 -3.49913418e-01 1.07571876e+00
2.15950459e-01 -1.98410973e-01 1.26620662e+00 1.47444934e-01
-3.57037544e-01 5.73942125e-01 -9.67514962e-02 3.86538118e-01
6.23491228e-01 -5.21609306e-01 -4.24743056e-01 2.03439623e-01
3.02570343e-01 3.10835958e-01 4.77317214e-01 -2.56902605e-01
-1.30190182e+00 -7.30448067e-01 1.13064849e+00 -1.28646374e-01
6.68445110e-01 -8.58957320e-02 -8.64628077e-01 6.09932840e-01
-7.93706775e-02 5.51837385e-01 1.09923506e+00 -2.75104553e-01
-2.42761970e-01 -3.83529007e-01 -1.35649037e+00 5.76377749e-01
7.36315846e-01 -7.51324117e-01 -2.62173057e-01 3.35771650e-01
-1.15234822e-01 -3.43515486e-01 -9.93697107e-01 3.34215850e-01
1.11744940e+00 -1.05524302e+00 6.34284079e-01 -8.20801795e-01
2.95275450e-02 -2.35076979e-01 -2.62610108e-01 -9.91695940e-01
-2.63700485e-01 -6.88092887e-01 4.19503480e-01 8.37178886e-01
-4.79818545e-02 -2.58636743e-01 9.47553813e-01 5.52645504e-01
4.16838914e-01 -7.58967876e-01 -1.58058274e+00 -6.22536719e-01
-2.89804071e-01 -2.87647992e-01 2.58566648e-01 1.12091649e+00
4.12886173e-01 4.74509746e-02 -3.42930943e-01 -1.61942363e-01
3.31750244e-01 1.62049413e-01 7.57468104e-01 -1.01964545e+00
3.69958848e-01 -5.07975399e-01 -1.37602198e+00 -2.57577866e-01
1.01520658e+00 -9.30569410e-01 -1.83575824e-01 -5.71722746e-01
2.67185997e-02 2.65207529e-01 -1.87989369e-01 5.33278108e-01
3.14424187e-01 4.83878076e-01 1.53585762e-01 2.05260307e-01
-1.44940898e-01 5.51560223e-01 9.51517224e-01 -2.61273570e-02
-2.47934237e-01 -3.21374349e-02 2.73702294e-01 1.10483325e+00
4.03525114e-01 -3.28320444e-01 -1.45860724e-02 3.19759607e-01
-3.90339255e-01 4.49347109e-01 1.40329644e-01 -7.42575109e-01
2.57780135e-01 -3.63130689e-01 5.46701372e-01 -1.85547128e-01
8.18699002e-01 -8.75628710e-01 4.20703113e-01 5.36715746e-01
-1.54150799e-01 4.76066083e-01 1.63831353e-01 2.01799557e-01
-6.08924270e-01 -1.41553208e-01 1.26736045e+00 1.07791096e-01
-1.27867007e+00 5.03347695e-01 -6.67048335e-01 -6.72994316e-01
1.59362650e+00 -4.85127836e-01 3.60051185e-01 -5.02535462e-01
-1.22936618e+00 -3.52659583e-01 2.49016792e-01 3.61399949e-01
8.19631159e-01 -1.89647186e+00 -6.24236166e-01 5.49883068e-01
2.46026039e-01 -8.92774224e-01 4.85731184e-01 1.07942700e+00
-4.05830115e-01 -2.34784171e-01 -7.19728947e-01 -7.93738902e-01
-2.21170974e+00 4.31625783e-01 6.04534507e-01 3.52581918e-01
-2.30676651e-01 8.31940591e-01 -1.28842324e-01 4.78574224e-02
-2.60346949e-01 1.70551971e-01 -5.56431830e-01 7.25693107e-01
8.94405127e-01 4.46744263e-01 -8.24453756e-02 -1.75858760e+00
-4.94412869e-01 1.15189993e+00 2.26739019e-01 -1.81382596e-01
1.00938284e+00 4.57933024e-02 -7.59083331e-01 6.51634037e-01
2.01692462e+00 -2.57739455e-01 -8.31771493e-01 -2.73193240e-01
1.76280409e-01 -4.99643177e-01 -2.29804307e-01 -4.63438146e-02
-1.10406530e+00 5.77625453e-01 9.96819258e-01 -1.24015868e-01
1.47759557e+00 -8.52725431e-02 2.63077140e-01 9.41266939e-02
2.91266173e-01 -1.09030747e+00 1.25392839e-01 3.27801973e-01
1.09392738e+00 -9.81928885e-01 -2.45827615e-01 -1.04818590e-01
-8.16311657e-01 1.54269612e+00 3.21738333e-01 -4.15182203e-01
1.06361067e+00 3.61634791e-01 6.14807606e-01 -3.03500593e-01
-4.48334634e-01 -7.42818341e-02 1.97684452e-01 5.05868256e-01
4.60188538e-01 5.77684008e-02 -4.41129208e-01 2.17162669e-01
-8.13826546e-02 2.73961157e-01 2.04218901e-03 1.03245127e+00
-5.09409368e-01 -6.22911990e-01 -4.67142433e-01 -2.16577221e-02
-4.18790370e-01 3.59608740e-01 -4.10710812e-01 8.01738262e-01
8.26501474e-02 4.41725224e-01 4.23497468e-01 -4.55571800e-01
4.61646944e-01 5.86048484e-01 7.35459626e-01 -2.49833241e-02
-3.54982585e-01 2.32282296e-01 -2.23147705e-01 -9.30092037e-01
-9.75787699e-01 -9.55453575e-01 -1.55452907e+00 -2.33984575e-01
2.72360109e-02 1.27136678e-01 7.26945341e-01 8.97096276e-01
9.35061201e-02 -5.98124802e-01 9.68186677e-01 -1.20650661e+00
-9.92471948e-02 -8.33520651e-01 -9.69335735e-01 7.75729775e-01
3.71910751e-01 -8.30393076e-01 -3.95912617e-01 4.31913435e-01] | [13.638944625854492, 1.8270667791366577] |
aab28b4f-5b37-4fc5-933f-fc42d4d45dc0 | bayesimp-uncertainty-quantification-for | 2106.03477 | null | https://arxiv.org/abs/2106.03477v1 | https://arxiv.org/pdf/2106.03477v1.pdf | BayesIMP: Uncertainty Quantification for Causal Data Fusion | While causal models are becoming one of the mainstays of machine learning, the problem of uncertainty quantification in causal inference remains challenging. In this paper, we study the causal data fusion problem, where datasets pertaining to multiple causal graphs are combined to estimate the average treatment effect of a target variable. As data arises from multiple sources and can vary in quality and quantity, principled uncertainty quantification becomes essential. To that end, we introduce Bayesian Interventional Mean Processes, a framework which combines ideas from probabilistic integration and kernel mean embeddings to represent interventional distributions in the reproducing kernel Hilbert space, while taking into account the uncertainty within each causal graph. To demonstrate the utility of our uncertainty estimation, we apply our method to the Causal Bayesian Optimisation task and show improvements over state-of-the-art methods. | ['Dino Sejdinovic', 'Yee Whye Teh', 'Javier González', 'Jean-François Ton', 'Siu Lun Chau'] | 2021-06-07 | null | http://proceedings.neurips.cc/paper/2021/hash/1ca5c750a30312d1919ae6a4d636dcc4-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/1ca5c750a30312d1919ae6a4d636dcc4-Paper.pdf | neurips-2021-12 | ['bayesian-optimisation'] | ['methodology'] | [ 4.11309630e-01 3.13403249e-01 -1.47643968e-01 -2.25298777e-01
-8.45178127e-01 -5.74220896e-01 9.06297505e-01 6.58676922e-01
-7.67921060e-02 8.40656698e-01 7.97300339e-01 -2.20539570e-01
-9.50088620e-01 -7.75584638e-01 -8.24561954e-01 -7.78957069e-01
-2.43384928e-01 3.75833035e-01 -3.37860435e-01 4.07773942e-01
-2.26149764e-02 2.59976387e-01 -1.00277936e+00 -6.00429550e-02
1.25418329e+00 6.55480444e-01 -2.57879198e-01 5.02966702e-01
1.92918241e-01 7.65631497e-01 -2.94955105e-01 -4.72759247e-01
-3.21845025e-01 -4.39275563e-01 -5.72450638e-01 -2.96229035e-01
1.37150034e-01 -1.74811799e-02 -4.46977645e-01 1.15447748e+00
5.34477890e-01 1.44246116e-01 1.38646472e+00 -1.35217011e+00
-8.67153287e-01 1.06517196e+00 -6.35140479e-01 -1.23402697e-03
2.92764544e-01 -1.32515922e-01 1.10332942e+00 -6.86529040e-01
3.05292040e-01 1.66688824e+00 4.46162552e-01 8.96847025e-02
-1.96985054e+00 -4.12013084e-01 -9.96446311e-02 1.08594717e-02
-1.11027563e+00 -4.74638939e-01 6.09933972e-01 -9.88920391e-01
2.31160581e-01 1.26066478e-02 3.90056014e-01 1.21351182e+00
6.92663252e-01 6.18219435e-01 1.18034756e+00 -3.00163895e-01
7.04396784e-01 -2.56026745e-01 7.25772977e-02 4.49372262e-01
4.90519166e-01 5.61219692e-01 -5.48435092e-01 -6.86652303e-01
5.33034503e-01 8.27904120e-02 -4.80086565e-01 -7.56513059e-01
-1.22859657e+00 1.34758449e+00 4.57134575e-01 7.24895298e-02
-8.30796242e-01 7.36894965e-01 2.85020471e-02 -3.08157146e-01
6.04665220e-01 2.91175306e-01 -2.72347569e-01 -5.54048195e-02
-6.98697031e-01 5.60298085e-01 8.33186269e-01 5.18505633e-01
1.81344360e-01 -3.03277224e-01 -5.05676150e-01 4.84287322e-01
7.47317135e-01 6.51738882e-01 -3.11161250e-01 -1.16586483e+00
-1.01305142e-01 4.04302150e-01 1.66706309e-01 -9.60418224e-01
-4.01684105e-01 -1.99192286e-01 -1.10103273e+00 1.93205744e-01
4.12028939e-01 -3.96838486e-01 -7.38933444e-01 1.93206227e+00
5.46713650e-01 5.76242983e-01 -1.18501171e-01 7.41702139e-01
4.57405627e-01 3.90029222e-01 3.91555578e-01 -5.07751405e-01
1.37118554e+00 -1.26317888e-01 -1.04737020e+00 1.67629480e-01
3.52995247e-01 -4.99534309e-01 2.54948527e-01 4.32739556e-01
-8.52787435e-01 -3.74594480e-02 -8.38144302e-01 1.83591217e-01
-2.99631119e-01 -4.18711275e-01 7.40006387e-01 6.20347977e-01
-7.34816551e-01 8.81081700e-01 -6.47219241e-01 -3.23389880e-02
6.13157272e-01 8.76379460e-02 -3.50577235e-01 -4.30515528e-01
-1.46533918e+00 9.16832864e-01 4.82059807e-01 -3.06718946e-02
-9.51528132e-01 -1.39293396e+00 -9.86471415e-01 1.63394839e-01
6.23401701e-01 -1.18717515e+00 1.12936127e+00 -4.46585938e-02
-1.07944250e+00 9.47917253e-02 1.34081751e-01 -4.20120806e-01
3.66549194e-01 -3.00255269e-01 -2.62838364e-01 -1.56236261e-01
7.32353255e-02 2.81226039e-01 1.00481844e+00 -1.10894954e+00
-4.62342083e-01 -5.90948641e-01 -1.81200579e-01 -1.69557527e-01
2.59876639e-01 -1.43176109e-01 1.91793755e-01 -5.76346219e-01
-1.21654160e-01 -7.13852286e-01 -3.77985269e-01 -2.54422724e-01
-4.15452302e-01 -4.32784498e-01 3.58923674e-01 -6.15654469e-01
1.29807329e+00 -2.12216139e+00 8.46138775e-01 1.56593516e-01
7.09200919e-01 -3.54730666e-01 2.13414729e-01 3.61399204e-01
-4.02948081e-01 2.53373951e-01 -7.94092476e-01 -2.18942128e-02
4.16773021e-01 1.63923934e-01 -2.50176042e-01 7.60634720e-01
5.62084734e-01 9.97657478e-01 -1.29236782e+00 -4.23024774e-01
3.20236981e-01 7.06832409e-01 -5.41956842e-01 2.20561638e-01
-3.07093710e-01 4.24911201e-01 -5.61199069e-01 2.37753615e-01
4.59776670e-01 -1.23199984e-01 1.72712952e-01 -3.91220659e-01
1.52886152e-01 1.62632857e-03 -1.48757863e+00 1.86895478e+00
-2.05047950e-01 1.57502383e-01 -4.81188223e-02 -9.93359745e-01
1.71628505e-01 4.86435175e-01 6.45275772e-01 -9.81300473e-02
3.06896746e-01 -3.69002763e-03 1.18994460e-01 -3.18691313e-01
3.01377568e-02 -4.18405890e-01 -4.75624561e-01 4.48930532e-01
2.31173605e-01 -4.72129643e-01 -7.53774196e-02 4.84580129e-01
1.40021789e+00 6.70662150e-02 5.79593837e-01 -3.84017527e-01
-7.32429847e-02 -2.59406209e-01 4.52953786e-01 8.29730093e-01
-1.68437332e-01 2.78379798e-01 1.01917708e+00 1.16378643e-01
-6.46096706e-01 -1.67379761e+00 -6.23422623e-01 6.04674518e-01
-3.82427067e-01 -1.57425895e-01 -6.46398544e-01 -7.65040278e-01
5.89114130e-01 1.07746828e+00 -1.01503110e+00 -3.59785646e-01
4.03383635e-02 -1.09495425e+00 1.79619372e-01 4.87676114e-01
-1.95037931e-01 -4.66285139e-01 -3.84706289e-01 3.81863713e-01
-9.82278064e-02 -5.95870078e-01 -3.46025556e-01 1.73140898e-01
-6.52708888e-01 -1.34196734e+00 -5.89298368e-01 3.65565717e-01
1.55854568e-01 -2.58018821e-01 1.19980884e+00 -8.15002441e-01
-5.31559765e-01 8.57478976e-01 -6.40934557e-02 -7.81949162e-01
-4.58871484e-01 -5.07515192e-01 1.70764372e-01 -1.25283217e-02
9.82433930e-03 -6.36968195e-01 -6.02180362e-01 -3.61567587e-01
-1.05865502e+00 -2.88935274e-01 5.73598564e-01 8.60852182e-01
3.90975714e-01 3.53248030e-01 7.41983354e-01 -8.36031377e-01
9.94562089e-01 -8.61136734e-01 -6.65372908e-01 1.74252391e-01
-6.28276587e-01 4.92931843e-01 -1.13169946e-01 -2.69359499e-01
-1.14255190e+00 -1.67847067e-01 5.77091575e-01 -3.25146317e-01
-1.16877340e-01 9.17021811e-01 -3.13849300e-01 4.37107414e-01
6.93954647e-01 -6.18157864e-01 3.84757575e-03 -3.25106531e-01
1.27339804e+00 5.02936482e-01 5.46606898e-01 -6.35439932e-01
3.97520512e-01 3.27409536e-01 4.83734220e-01 -4.06713754e-01
-8.26840341e-01 -2.00413600e-01 -5.26052177e-01 -1.66133150e-01
1.22420263e+00 -7.09108353e-01 -9.91894305e-01 -1.10428631e-02
-1.25162399e+00 -1.92364561e-03 -4.65284705e-01 7.39735126e-01
-7.25151181e-01 -4.57037427e-02 -2.45906100e-01 -1.13895309e+00
6.08705869e-03 -1.05446672e+00 1.21181750e+00 6.02766611e-02
-3.43565434e-01 -1.30535591e+00 4.59870249e-01 -3.61813605e-02
4.55588028e-02 6.79761589e-01 1.13608968e+00 -4.33183223e-01
-4.10727978e-01 -1.79820150e-01 -4.45290536e-01 7.46157765e-02
3.15191239e-01 -5.75908944e-02 -9.57819402e-01 1.75486282e-01
5.11173643e-02 1.42654441e-02 1.09251785e+00 9.81529415e-01
7.93882489e-01 -7.43383020e-02 -5.63093185e-01 -2.97979861e-02
1.37834656e+00 -3.10784057e-02 3.26430857e-01 -5.78893244e-01
8.10206950e-01 9.12919819e-01 3.55172634e-01 6.43899381e-01
3.67200732e-01 5.13559103e-01 4.47016031e-01 2.98121631e-01
4.06394415e-02 -2.54100174e-01 4.03642729e-02 3.44624817e-01
1.77800477e-01 -1.96614727e-01 -8.68700445e-01 5.78073800e-01
-2.14446759e+00 -1.00422537e+00 -4.75228697e-01 2.28598261e+00
9.52588260e-01 -1.71796098e-01 5.22854179e-02 -1.99728832e-02
6.92877173e-01 -2.34423839e-02 -5.05838394e-01 -1.09797247e-01
1.03768073e-01 2.58315742e-01 6.11093819e-01 5.81282556e-01
-1.19926083e+00 3.24225932e-01 6.43366718e+00 7.78446734e-01
-2.84409761e-01 2.51087010e-01 4.21535432e-01 2.20622029e-02
-5.58461726e-01 1.53682292e-01 -1.67998597e-01 4.31112617e-01
1.20415676e+00 -2.99872905e-01 4.30370867e-01 1.80374235e-01
4.08623129e-01 -4.07081217e-01 -1.42780101e+00 7.41241634e-01
-3.46586555e-01 -1.16163981e+00 -3.63541514e-01 4.83409941e-01
7.94428229e-01 -2.86550581e-01 -1.26747740e-02 1.18056335e-01
1.19515860e+00 -1.26524055e+00 3.77490669e-01 1.09598112e+00
5.11358261e-01 -7.56113648e-01 6.09651446e-01 7.01967478e-02
-6.10409021e-01 -7.71245509e-02 -1.42079681e-01 1.71721220e-01
4.92283702e-01 1.55036855e+00 -6.09305739e-01 9.59274888e-01
4.90565300e-01 8.01080942e-01 -6.19307458e-02 9.67216194e-01
-3.97402138e-01 7.15424657e-01 -2.88239628e-01 1.49944529e-01
-2.38801405e-01 -2.75484413e-01 5.37426531e-01 9.52178776e-01
3.83060515e-01 3.05793881e-01 -2.02214852e-01 1.40130603e+00
-1.04414806e-01 -2.91065067e-01 -8.98085952e-01 -4.61573541e-01
3.65360230e-01 1.00007498e+00 -2.48098060e-01 -1.88526496e-01
-2.06858858e-01 6.58697546e-01 2.07629621e-01 3.79369140e-01
-8.66583109e-01 -9.81465913e-03 7.56433427e-01 -4.30833369e-01
4.26655076e-02 9.27218348e-02 -3.34306270e-01 -9.56827939e-01
-4.19854999e-01 -5.34927726e-01 6.76228762e-01 -6.82737291e-01
-1.64916861e+00 -4.10871387e-01 3.77807796e-01 -5.85616589e-01
-3.64433825e-01 -5.94247580e-01 -2.35254765e-01 1.15183151e+00
-9.63309646e-01 -8.30038548e-01 3.49151045e-01 1.93300083e-01
-1.34669645e-02 3.40891808e-01 8.00940216e-01 -1.57025740e-01
-4.94928628e-01 -1.58985808e-01 3.68095309e-01 -2.71727175e-01
6.90592825e-01 -1.74248731e+00 -1.68200601e-02 6.55334115e-01
-5.73114865e-03 5.99544883e-01 9.07127202e-01 -9.70666647e-01
-1.63859844e+00 -9.49722111e-01 5.60964525e-01 -8.23827982e-01
1.19531155e+00 -2.41300896e-01 -8.12941074e-01 4.50436890e-01
2.09949881e-01 -7.38712475e-02 7.50466824e-01 4.63096559e-01
-3.69351566e-01 1.64605558e-01 -1.16351819e+00 5.86640775e-01
8.85820329e-01 -3.89944583e-01 -8.52778196e-01 2.14494556e-01
1.04803360e+00 1.88768283e-02 -1.48230934e+00 5.34918368e-01
4.54154521e-01 -5.22111118e-01 1.03350496e+00 -7.19384909e-01
5.83878696e-01 -3.30020726e-01 -3.62851620e-01 -1.98166573e+00
-5.82372606e-01 -4.44866240e-01 -3.63094568e-01 1.24937093e+00
1.58764109e-01 -5.04844844e-01 1.10242456e-01 8.40569615e-01
2.19063774e-01 -2.62654364e-01 -1.12294757e+00 -3.16228926e-01
2.71971345e-01 -8.35014701e-01 5.44399500e-01 9.27666605e-01
2.75887638e-01 6.37695909e-01 -4.07594711e-01 4.27580923e-01
1.31816411e+00 -1.40540615e-01 2.43005842e-01 -1.78251207e+00
-3.62212867e-01 -5.88543773e-01 -3.85581076e-01 -1.14296995e-01
2.36599013e-01 -7.79495776e-01 1.50942191e-01 -1.68002450e+00
4.01325524e-01 -6.46633580e-02 -3.68321955e-01 1.97554454e-01
-5.80142856e-01 -3.26105952e-01 -1.91687478e-03 -3.58663052e-01
-1.47001132e-01 8.56530666e-01 9.50164616e-01 -3.05070162e-01
5.67641966e-02 -2.17507452e-01 -8.33075404e-01 6.18699253e-01
4.13402468e-01 -7.05255330e-01 -5.64117432e-01 9.26968362e-03
4.11340535e-01 5.38412631e-01 7.01225519e-01 -4.19354588e-01
8.08618143e-02 -3.16652119e-01 2.96710223e-01 -2.80877173e-01
4.48375791e-02 -8.44648123e-01 3.44512463e-01 3.68893117e-01
-5.02919137e-01 -3.58992457e-01 1.99681461e-01 1.25964987e+00
1.32297471e-01 -3.13583463e-02 4.50104922e-01 3.01947981e-01
-6.00731149e-02 1.65083051e-01 -1.25789255e-01 3.09519768e-02
7.93668211e-01 6.28665209e-01 -3.36669147e-01 -1.70584887e-01
-9.28109288e-01 3.12912852e-01 2.05683596e-02 3.00997764e-01
6.39338195e-01 -1.51857650e+00 -1.08119380e+00 -4.34105992e-01
1.10402264e-01 -3.23846973e-02 4.63426143e-01 1.19387841e+00
4.01927501e-01 4.45760697e-01 4.10226256e-01 -6.32807076e-01
-5.42963743e-01 1.00696087e+00 1.93675682e-01 -1.43578753e-01
-1.43675312e-01 5.94564021e-01 5.36378622e-01 -4.07472491e-01
-9.45615545e-02 -3.53539526e-01 -2.02764094e-01 2.83088386e-01
5.69034517e-01 6.89148068e-01 -2.46245518e-01 -1.31232351e-01
-3.39666098e-01 8.46469700e-02 3.25379878e-01 -5.29181421e-01
1.24344182e+00 3.78845148e-02 -3.87041032e-01 8.34111929e-01
1.09232628e+00 -1.08967096e-01 -1.11331916e+00 -1.93534940e-01
2.00995877e-01 -2.46847272e-01 4.87025201e-01 -1.08453715e+00
-5.57400584e-01 7.97446489e-01 7.87039340e-01 3.97763163e-01
9.71127570e-01 3.44219208e-01 -9.29495022e-02 -2.74819821e-01
1.33246437e-01 -6.64203048e-01 -3.26959193e-01 -3.73805687e-02
1.10544407e+00 -1.26630020e+00 1.24930991e-02 -4.09768999e-01
-2.72146791e-01 7.24071264e-01 -2.79660106e-01 -8.37422609e-02
1.04343462e+00 2.34194547e-01 -5.31235039e-01 -4.66561228e-01
-7.67124891e-01 -1.56495929e-01 7.11232901e-01 7.09890425e-01
5.62012196e-01 6.12355113e-01 -5.04067183e-01 7.23399580e-01
1.62993252e-01 1.89620808e-01 5.09227633e-01 5.98423004e-01
-9.12207142e-02 -9.44780588e-01 -6.69996381e-01 7.24443436e-01
-5.37691712e-01 -2.26975694e-01 -1.66185185e-01 4.98897880e-01
5.82151040e-02 1.21860766e+00 -2.03827098e-01 -2.27219183e-02
4.96845782e-01 2.58626729e-01 5.65682292e-01 -4.74398017e-01
1.55413806e-01 2.95929193e-01 -2.89405957e-02 -6.49630189e-01
-4.83959734e-01 -1.11674404e+00 -9.37908232e-01 -2.27258146e-01
-2.54476160e-01 1.56135246e-01 8.18579376e-01 1.04589355e+00
2.81077415e-01 1.06745470e+00 2.09276900e-01 -5.71040034e-01
-8.31067324e-01 -1.02987039e+00 -8.04676950e-01 1.85916916e-01
4.61622268e-01 -1.10798573e+00 -4.06982660e-01 6.96296096e-02] | [7.849324703216553, 5.34016752243042] |
c81b29b3-f0a5-43ae-834f-f4987db87c68 | findings-of-the-2018-conference-on-machine | null | null | https://aclanthology.org/W18-6401 | https://aclanthology.org/W18-6401.pdf | Findings of the 2018 Conference on Machine Translation (WMT18) | This paper presents the results of the premier shared task organized alongside the Conference on Machine Translation (WMT) 2018. Participants were asked to build machine translation systems for any of 7 language pairs in both directions, to be evaluated on a test set of news stories. The main metric for this task is human judgment of translation quality. This year, we also opened up the task to additional test sets to probe specific aspects of translation. | ['Ond{\\v{r}}ej Bojar', 'Christof Monz', 'Barry Haddow', 'Mark Fishel', 'Yvette Graham', 'Philipp Koehn', 'Christian Federmann'] | 2018-10-01 | null | null | null | ws-2018-10 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 3.35952312e-01 7.37498477e-02 -4.73571032e-01 -5.85181892e-01
-1.69938660e+00 -9.39016879e-01 1.15444458e+00 -9.68473032e-02
-4.94481534e-01 1.09894836e+00 4.47093964e-01 -7.11024046e-01
4.14420962e-01 -2.67592698e-01 -8.23859572e-01 1.96255639e-01
3.92123669e-01 1.06781638e+00 -2.52873451e-01 -7.71999359e-01
1.91508487e-01 -1.92702338e-01 -5.27473032e-01 8.64418447e-01
1.05443609e+00 3.24566126e-01 -3.83000523e-02 5.82671165e-01
2.73262411e-01 1.42267160e-02 -6.19880557e-01 -1.27383316e+00
3.34291726e-01 -7.06489146e-01 -1.24883771e+00 -4.51606303e-01
6.21287405e-01 -5.46991117e-02 9.94834229e-02 9.48369443e-01
8.13622534e-01 -1.01145105e-02 4.92797107e-01 -1.05202973e+00
-1.28774476e+00 1.17160177e+00 -2.98813824e-02 6.46893203e-01
9.44464445e-01 2.01521531e-01 1.36563408e+00 -1.42750251e+00
1.03050125e+00 1.15617967e+00 6.46371424e-01 4.77567762e-01
-1.14237249e+00 -5.98193705e-01 -3.42404932e-01 3.66381198e-01
-8.63827944e-01 -6.17830038e-01 4.65237349e-02 -1.52751505e-01
1.45021129e+00 5.69218636e-01 5.63865423e-01 1.92478669e+00
3.67283165e-01 6.63154244e-01 1.75542653e+00 -5.30004859e-01
-2.19088361e-01 3.98882002e-01 -3.22744250e-02 2.81004366e-02
1.31496862e-01 2.08024487e-01 -8.82424355e-01 -1.41970411e-01
4.50757779e-02 -8.51161778e-01 -3.10565829e-01 5.02545357e-01
-2.00950980e+00 6.19034827e-01 1.62762076e-01 6.40722513e-01
-4.20166366e-02 -5.08398473e-01 3.49407881e-01 1.12779498e+00
8.52940798e-01 1.02338588e+00 -6.18739486e-01 -5.95152438e-01
-7.55353570e-01 3.81792665e-01 1.12709665e+00 9.64353740e-01
3.21671367e-01 -7.92352855e-01 -6.09800279e-01 8.87884378e-01
-7.34698726e-03 8.36126566e-01 5.86362481e-01 -2.61012673e-01
1.24953985e+00 2.86911309e-01 2.53480226e-01 -6.10427797e-01
-9.54340994e-02 -4.96250331e-01 -3.97294492e-01 -6.00514650e-01
2.50486106e-01 -3.67270768e-01 -6.08709753e-01 1.59680068e+00
-7.98540711e-02 -4.86095071e-01 5.02816588e-02 1.09558284e+00
8.46095681e-01 7.80510128e-01 -4.49694425e-01 -3.35035205e-01
1.18141675e+00 -1.47306061e+00 -9.16389644e-01 -4.71016437e-01
9.07106400e-01 -1.68394959e+00 1.26173294e+00 1.89868078e-01
-1.26319003e+00 -4.73437488e-01 -1.12161672e+00 -2.51631171e-01
-4.20467108e-01 1.65583849e-01 3.03944379e-01 4.37579066e-01
-1.37604272e+00 5.42971015e-01 -4.54202563e-01 -8.75346899e-01
-1.65316761e-01 9.13798586e-02 -4.86973941e-01 -1.54917136e-01
-1.84992743e+00 1.94947433e+00 -2.28425160e-01 2.50924937e-02
-5.01308441e-01 -2.72369891e-01 -3.30805123e-01 -4.96086031e-01
-2.00014547e-01 -9.20161009e-01 1.77061081e+00 -1.23725963e+00
-1.71966374e+00 1.56876576e+00 -1.76951990e-01 -4.08445537e-01
8.81562054e-01 -6.84175968e-01 -7.71361589e-01 -4.70872194e-01
6.48323298e-01 4.63316083e-01 4.02569860e-01 -6.75813913e-01
-2.48867750e-01 -5.67961968e-02 -1.30253851e-01 5.10209739e-01
-9.37220082e-02 7.52918839e-01 -1.76492497e-01 -7.37444282e-01
-1.16244070e-01 -1.12540185e+00 1.71683416e-01 -7.74178922e-01
-6.87126219e-01 -2.00434253e-01 2.80660719e-01 -1.05397511e+00
9.58329797e-01 -1.54021215e+00 6.05954766e-01 -1.80014387e-01
-2.94607550e-01 -7.60583505e-02 -5.21346331e-01 9.48426664e-01
2.51259536e-01 3.61493200e-01 -1.30666256e-01 -5.53095877e-01
-1.20161302e-01 -1.90577075e-01 -6.19301736e-01 2.84043819e-01
1.85413688e-01 1.52797246e+00 -7.86844015e-01 -2.52350062e-01
-5.98653436e-01 1.02464803e-01 6.52118586e-03 9.69720408e-02
-1.08295709e-01 6.79492593e-01 -7.48955160e-02 3.77661198e-01
3.90287846e-01 -1.01570696e-01 1.02285273e-01 3.14514995e-01
-1.17319159e-01 1.21672988e+00 -1.62740722e-01 1.80685043e+00
-5.04571676e-01 1.02348316e+00 -2.88921982e-01 -4.16395962e-01
8.09872806e-01 7.00457394e-01 -3.87721024e-02 -1.05788457e+00
2.50794530e-01 7.34759390e-01 3.38859588e-01 -3.33853543e-01
5.29320955e-01 2.64353529e-02 -1.91681877e-01 1.05438399e+00
1.04361296e-01 -4.48176384e-01 2.19167322e-01 6.91471919e-02
9.79332507e-01 1.44151971e-01 8.00529569e-02 -5.54934561e-01
2.13822320e-01 3.98800641e-01 3.11497003e-01 5.16226828e-01
-3.09346259e-01 6.32409632e-01 -2.23697051e-02 -4.74536031e-01
-1.26154101e+00 -9.08319652e-01 1.26779363e-01 1.16824961e+00
-3.12094688e-02 -6.14019752e-01 -8.80374193e-01 -5.56531012e-01
-4.63434875e-01 8.44087243e-01 -6.40942693e-01 -2.32654184e-01
-8.24212432e-01 -6.13372624e-01 7.20050335e-01 1.55687898e-01
4.11949068e-01 -6.50679410e-01 -2.42345527e-01 -1.04553893e-01
-1.16190767e+00 -1.25140691e+00 -9.49606121e-01 -2.94243544e-01
-8.08321178e-01 -4.55061466e-01 -8.90642703e-01 -1.10861778e+00
1.53463855e-01 3.22311401e-01 1.60954356e+00 -8.74622837e-02
2.51711875e-01 -5.50358891e-02 -6.12839937e-01 -5.05301178e-01
-7.43564963e-01 5.68145335e-01 5.33636846e-02 -3.76821071e-01
6.76035702e-01 -3.21871638e-01 -2.65337199e-01 7.19499648e-01
-2.07505479e-01 6.58095300e-01 6.49718106e-01 7.53720343e-01
1.91413179e-01 -1.05795085e+00 4.73173916e-01 -3.71413052e-01
1.35778117e+00 -4.08014208e-01 -9.54781696e-02 6.89076722e-01
-6.76233351e-01 -3.68351161e-01 3.05549026e-01 -4.38572109e-01
-6.18799627e-01 -6.43147230e-01 1.40426442e-01 4.55396205e-01
5.00287116e-01 7.14859843e-01 1.12074465e-01 3.02553207e-01
9.01964009e-01 8.64738300e-02 -1.28867626e-01 -3.16345930e-01
3.10720861e-01 8.76488626e-01 3.26657653e-01 -6.30812407e-01
8.07951331e-01 -3.26329708e-01 -5.29783189e-01 -5.08210761e-03
-7.39844918e-01 4.98283515e-03 -7.14469194e-01 -4.76208590e-02
7.93901980e-01 -9.92327511e-01 1.56556666e-01 2.15251744e-01
-1.52857411e+00 -5.07284522e-01 2.67909020e-01 8.39011133e-01
-7.07610071e-01 -3.37231129e-01 -8.47178400e-01 -2.30762392e-01
-7.26153314e-01 -1.39570498e+00 1.07643664e+00 -3.41313243e-01
-8.61939371e-01 -8.98483455e-01 5.87699711e-01 8.86396229e-01
6.96226656e-01 -1.44116133e-01 7.10300088e-01 -1.06186402e+00
-2.00644493e-01 -1.71604574e-01 -3.55087668e-02 3.38991522e-03
5.13879023e-02 3.52484435e-02 -5.50096631e-01 -2.67979831e-01
2.66946368e-02 -6.36732638e-01 4.88779426e-01 -1.84228554e-01
3.23237330e-02 -3.63783807e-01 -3.65844369e-01 2.11520016e-01
7.66386807e-01 -2.87867814e-01 4.01275814e-01 8.00326109e-01
1.26280949e-01 5.50322413e-01 6.72481954e-01 -4.43379313e-01
5.47257066e-01 1.35480225e+00 -4.20986652e-01 1.01846583e-01
-2.43445829e-01 -2.98220277e-01 9.16439474e-01 1.40830052e+00
-3.60550135e-01 -4.98500168e-01 -1.19420934e+00 4.41585362e-01
-1.73590434e+00 -6.80907190e-01 -1.86115041e-01 2.02405906e+00
1.11202466e+00 1.07701972e-01 2.10776657e-01 -4.75593358e-01
8.50420177e-01 -3.24531257e-01 -8.49519372e-02 -9.82083082e-01
-4.13644791e-01 3.15153211e-01 3.43915850e-01 6.64859712e-01
-7.35896707e-01 1.26914370e+00 7.65457010e+00 4.53213692e-01
-1.15564489e+00 5.55527270e-01 7.03796446e-01 -1.79191321e-01
-5.55208623e-01 1.31795600e-01 -5.99433541e-01 2.46189684e-01
1.44809127e+00 -6.97338700e-01 6.88417196e-01 2.33672053e-01
8.96376148e-02 2.47550830e-01 -1.44945478e+00 5.12632668e-01
2.33873188e-01 -1.01646149e+00 5.92552163e-02 -2.73289997e-02
9.53262746e-01 6.99702203e-01 2.40047157e-01 5.55775225e-01
3.41186255e-01 -1.28465986e+00 7.75124013e-01 1.12483628e-01
9.61210966e-01 -3.19780707e-01 8.32455575e-01 5.38698256e-01
-3.21781009e-01 3.03577513e-01 -2.04756454e-01 -5.17963052e-01
4.22692895e-01 3.27327222e-01 -1.09862220e+00 6.64841413e-01
6.09454393e-01 7.34779537e-01 -5.84323168e-01 6.32487237e-01
-5.57015955e-01 6.48959994e-01 -1.18400939e-01 -2.94672459e-01
2.75143981e-01 -2.18416601e-01 7.52339780e-01 1.25253201e+00
6.10403776e-01 -3.62617940e-01 -1.49934679e-01 5.54330409e-01
-6.53835595e-01 5.78333437e-01 -7.03620911e-01 -3.46844316e-01
5.02310693e-01 1.06131470e+00 -4.60134059e-01 -3.74240786e-01
-4.57264543e-01 1.46397400e+00 5.14242589e-01 2.64429241e-01
-7.43295312e-01 -7.33057212e-04 3.99562329e-01 -1.42816812e-01
-3.22780788e-01 -3.55417758e-01 -7.31566370e-01 -1.46072924e+00
4.32901025e-01 -1.44286048e+00 -5.85632445e-03 -8.25670421e-01
-1.51587093e+00 1.09973168e+00 -1.90431297e-01 -1.51151586e+00
-4.69465435e-01 -2.91323513e-01 -7.35850930e-01 1.35967624e+00
-9.66374040e-01 -1.14729488e+00 2.86450088e-01 1.95749417e-01
7.77907431e-01 -4.57649827e-01 1.13376987e+00 3.15548867e-01
-3.55765074e-01 1.05846488e+00 1.24045894e-01 9.42933783e-02
1.62196004e+00 -8.48260462e-01 1.44916904e+00 8.44406545e-01
4.43060964e-01 9.61047530e-01 9.98681664e-01 -7.99096525e-01
-1.48948491e+00 -8.09149504e-01 2.07953930e+00 -1.22777748e+00
9.08208072e-01 -5.54202020e-01 -4.65726882e-01 8.94832194e-01
1.12312913e+00 -5.47285438e-01 8.38068962e-01 3.70811105e-01
-5.42716444e-01 4.46798295e-01 -9.36701000e-01 6.24823570e-01
1.24749768e+00 -8.42048645e-01 -1.02986538e+00 8.10972989e-01
1.06814849e+00 -7.19296217e-01 -8.25593412e-01 4.59185332e-01
6.92090869e-01 -4.35973167e-01 5.12839258e-01 -9.31826174e-01
1.16351211e+00 1.51496425e-01 -2.71178693e-01 -2.01167464e+00
-3.09414268e-01 -9.44552660e-01 2.67881483e-01 8.19629252e-01
1.19150162e+00 -7.79404044e-01 -4.31638025e-03 3.88723016e-01
-2.38406792e-01 -8.92254233e-01 -1.12026262e+00 -1.03301036e+00
6.80775881e-01 -2.22804800e-01 5.06828547e-01 1.14426374e+00
4.82317656e-01 1.16291332e+00 -3.32305342e-01 -3.97581369e-01
1.04638644e-01 1.54755279e-01 7.71313608e-01 -6.63274944e-01
-4.21446621e-01 -7.67402291e-01 -1.67130068e-01 -8.78171384e-01
-1.67287160e-02 -1.41505992e+00 -7.14104846e-02 -1.74085879e+00
4.48055536e-01 2.11112127e-01 3.53369378e-02 1.42214790e-01
-5.77751577e-01 7.91154444e-01 2.87276089e-01 6.37061179e-01
-3.70610595e-01 3.94649237e-01 1.49111629e+00 -2.93718487e-01
2.21429095e-02 8.20582956e-02 -8.43322575e-01 -2.00059563e-01
8.05407763e-01 -3.09519619e-01 -1.72949821e-01 -1.30415618e+00
7.57390797e-01 -8.35318789e-02 2.57545430e-03 -5.91989219e-01
-1.33995358e-02 -4.34995554e-02 2.48157978e-01 -2.77905375e-01
1.73314422e-01 -2.11916700e-01 3.65574025e-02 3.27732533e-01
-8.20109427e-01 8.90213192e-01 1.98925778e-01 -2.24364847e-01
-3.75918388e-01 4.97237682e-01 2.42292151e-01 9.75521505e-02
9.05177891e-02 -1.83432624e-02 -2.76466489e-01 1.45300955e-01
8.73951495e-01 9.59930494e-02 -5.90017319e-01 -5.56074142e-01
-4.65781242e-01 3.96492660e-01 4.18112308e-01 1.07618761e+00
4.08249766e-01 -1.63415754e+00 -1.56010926e+00 -2.47504458e-01
3.91383857e-01 -7.98394561e-01 -3.28479022e-01 1.40866840e+00
-2.76387751e-01 7.73671746e-01 -3.93931568e-01 -3.48250508e-01
-1.08770573e+00 2.42740288e-01 1.67781681e-01 -5.35253584e-01
-4.27924514e-01 9.47886825e-01 -6.16324782e-01 -9.59450483e-01
-4.79810715e-01 -1.42293535e-02 8.51016864e-02 -1.91998437e-01
7.99877882e-01 2.85657316e-01 5.87905645e-01 -8.52484465e-01
-3.42959821e-01 3.20765227e-01 -2.59618104e-01 -9.06297147e-01
9.51410830e-01 -3.08641165e-01 -4.18531895e-01 8.23011398e-01
1.18262362e+00 9.71611738e-02 -1.29010484e-01 -4.54131126e-01
1.49700344e-01 -3.69209379e-01 -4.05420184e-01 -1.74090052e+00
-9.89413485e-02 7.21463382e-01 4.00402457e-01 -8.09748098e-02
6.32217348e-01 3.07270568e-02 1.18974185e+00 4.35094714e-01
6.03269041e-01 -1.25599098e+00 -2.82284707e-01 1.04649436e+00
1.38295579e+00 -1.54881668e+00 -3.13189864e-01 -2.40841210e-01
-5.41377246e-01 1.21577537e+00 4.23628986e-01 2.17586324e-01
1.21881485e-01 -1.11898400e-01 5.93256533e-01 7.75008202e-02
-1.19914377e+00 2.94411868e-01 6.69040918e-01 4.63441640e-01
1.10882688e+00 4.90971088e-01 -1.06618166e+00 5.41095555e-01
-9.92519200e-01 -1.18778162e-01 2.50832260e-01 5.51235974e-01
-1.65789142e-01 -1.38182557e+00 -1.69170037e-01 2.66868949e-01
-4.59966302e-01 -4.63736355e-01 -1.26337254e+00 3.98469836e-01
-2.79233694e-01 1.29291713e+00 -1.46027029e-01 -8.34668756e-01
3.99448991e-01 2.63930202e-01 6.71420753e-01 -6.68968558e-01
-8.93989921e-01 -4.94504541e-01 5.96947908e-01 -4.15228456e-01
-2.34106064e-01 -8.43701422e-01 -5.18648863e-01 -4.85129058e-01
-4.42345105e-02 4.57263947e-01 8.17748070e-01 1.16637027e+00
3.62787396e-01 -6.43058717e-02 6.64571285e-01 -5.17351508e-01
-7.62025297e-01 -1.69642568e+00 5.31283021e-01 3.47786635e-01
1.36062965e-01 -2.63105035e-01 -2.95235276e-01 -1.23464301e-01] | [11.544992446899414, 10.291520118713379] |
e4d9f44b-060e-4314-8c75-097fe839811b | timing-process-interventions-with-causal | 2306.04299 | null | https://arxiv.org/abs/2306.04299v1 | https://arxiv.org/pdf/2306.04299v1.pdf | Timing Process Interventions with Causal Inference and Reinforcement Learning | The shift from the understanding and prediction of processes to their optimization offers great benefits to businesses and other organizations. Precisely timed process interventions are the cornerstones of effective optimization. Prescriptive process monitoring (PresPM) is the sub-field of process mining that concentrates on process optimization. The emerging PresPM literature identifies state-of-the-art methods, causal inference (CI) and reinforcement learning (RL), without presenting a quantitative comparison. Most experiments are carried out using historical data, causing problems with the accuracy of the methods' evaluations and preempting online RL. Our contribution consists of experiments on timed process interventions with synthetic data that renders genuine online RL and the comparison to CI possible, and allows for an accurate evaluation of the results. Our experiments reveal that RL's policies outperform those from CI and are more robust at the same time. Indeed, the RL policies approach perfect policies. Unlike CI, the unaltered online RL approach can be applied to other, more generic PresPM problems such as next best activity recommendations. Nonetheless, CI has its merits in settings where online learning is not an option. | ['Jochen De Weerdt', 'Wouter Verbeke', 'Hans Weytjens'] | 2023-06-07 | null | null | null | null | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [ 2.77513444e-01 2.69899756e-01 -6.38742566e-01 1.49042040e-01
-2.60371238e-01 -3.25747877e-01 1.00661945e+00 7.08312035e-01
-3.45243126e-01 6.53025925e-01 1.59501404e-01 -5.91189981e-01
-8.66225600e-01 -8.05424035e-01 -5.36598742e-01 -4.91578639e-01
-4.39173013e-01 7.89476454e-01 1.54625311e-01 6.51329160e-02
6.30899787e-01 5.76683104e-01 -1.40873802e+00 6.36976212e-02
5.39499998e-01 9.55010593e-01 -5.29860444e-02 7.04594135e-01
-3.16749275e-01 1.13213611e+00 -4.49406862e-01 -1.68897912e-01
3.22400093e-01 -1.83379248e-01 -6.94580853e-01 1.97597146e-01
-2.46490449e-01 -1.51761383e-01 -5.25240079e-02 5.42943537e-01
1.39228836e-01 2.04766691e-01 3.87352705e-01 -1.44059873e+00
-4.04334635e-01 6.98605001e-01 -4.08781409e-01 3.13673556e-01
6.18424594e-01 4.10732955e-01 1.07442284e+00 -2.98990697e-01
4.68765974e-01 1.57187414e+00 4.52937633e-01 1.71128258e-01
-1.38562512e+00 -2.83967644e-01 6.07629538e-01 2.97378600e-01
-8.81466329e-01 -2.64204800e-01 4.03925300e-01 -4.76887077e-01
8.82953823e-01 2.13378936e-01 7.80645192e-01 1.08795762e+00
4.08177972e-01 9.12099719e-01 1.32276905e+00 -5.50490141e-01
7.69751251e-01 -1.35507071e-02 -1.13400251e-01 2.01932341e-01
5.21662474e-01 5.99639595e-01 -7.03205705e-01 -2.57022411e-01
8.58044147e-01 3.06982964e-01 -7.43384361e-02 -4.69949096e-01
-1.26220512e+00 6.66638851e-01 -3.26771110e-01 2.68065304e-01
-7.44395256e-01 1.04208581e-01 3.69744480e-01 6.08601213e-01
7.79095218e-02 8.29742730e-01 -5.91485977e-01 -6.19069040e-01
-8.28682184e-01 4.82506990e-01 1.47457886e+00 7.81496286e-01
3.17144513e-01 -1.11956701e-01 -7.24320471e-01 2.86171407e-01
3.30354542e-01 2.52196729e-01 3.05674970e-01 -1.23017931e+00
3.61592680e-01 4.79770005e-01 7.00910926e-01 -6.65726483e-01
-2.05299243e-01 -1.10719174e-01 -3.89893591e-01 1.42649665e-01
6.77057981e-01 -1.63645297e-01 -3.64646614e-01 1.15054536e+00
1.07799275e-02 1.34586960e-01 -1.04014970e-01 5.40498018e-01
-2.29242533e-01 6.01362109e-01 2.90622324e-01 -7.73826897e-01
1.08428824e+00 -8.41531575e-01 -9.22100127e-01 -3.72888863e-01
-1.16674341e-02 -4.52222914e-01 9.76854920e-01 8.90442431e-01
-9.35964108e-01 -3.61162663e-01 -7.43894160e-01 8.38327348e-01
-7.76403323e-02 -4.92205471e-01 7.81038404e-01 7.28477776e-01
-7.91377366e-01 1.21577370e+00 -1.18094468e+00 -3.27282399e-01
3.54582578e-01 3.69271517e-01 1.12083912e-01 -5.52337915e-02
-8.61584187e-01 1.01535201e+00 3.00647050e-01 -4.73651849e-02
-1.01539803e+00 -6.27297223e-01 -4.94565845e-01 2.87008911e-01
1.12235570e+00 -3.45149130e-01 1.75925589e+00 -9.15876985e-01
-1.87772405e+00 5.15962392e-02 -1.39924707e-02 -7.43807077e-01
8.30454171e-01 -4.57598627e-01 -3.14153999e-01 7.00728828e-03
-1.64968446e-01 4.37392928e-02 7.39865422e-01 -1.23723078e+00
-8.78056169e-01 -2.42418259e-01 -9.17331949e-02 3.86565924e-02
1.32451048e-02 -5.06504327e-02 -1.17191553e-01 -1.61473811e-01
-2.12965176e-01 -7.23232567e-01 -6.72341704e-01 -4.76763517e-01
-3.56607050e-01 -2.43451476e-01 5.26697397e-01 -3.88437182e-01
1.39099228e+00 -1.73569632e+00 -2.23443568e-01 3.40909213e-01
-6.32948056e-02 -1.02247857e-01 2.71485686e-01 1.05994725e+00
7.21436739e-02 3.23398262e-01 2.14552227e-02 -1.72608167e-01
3.04376990e-01 3.52294177e-01 -2.21570462e-01 2.57105023e-01
1.05614066e-01 7.50499785e-01 -1.03888786e+00 -1.98655158e-01
5.20525396e-01 -2.31784433e-01 -1.64541334e-01 3.46534550e-01
-4.18668687e-01 4.48758125e-01 -4.32742178e-01 7.22953260e-01
-1.15497224e-01 -3.27507794e-01 5.30153632e-01 4.12181586e-01
-4.75839853e-01 1.27381444e-01 -1.50287521e+00 8.89364123e-01
-4.70803618e-01 4.35014635e-01 -1.27358943e-01 -8.34135175e-01
7.95532286e-01 4.50689822e-01 1.04031897e+00 -7.79442728e-01
-2.55088121e-01 -1.47546589e-01 1.14088036e-01 -4.96891171e-01
2.25466982e-01 -8.91184136e-02 2.06670731e-01 6.72347069e-01
-1.62386462e-01 2.24347934e-01 5.68395257e-01 -3.09833914e-01
1.42703509e+00 2.80913979e-01 7.75569975e-01 -2.62688369e-01
3.29975247e-01 -5.56154810e-02 6.33675218e-01 1.17446864e+00
-2.73886055e-01 -1.09551653e-01 8.18506420e-01 -2.71586418e-01
-8.66288543e-01 -1.03077841e+00 1.82396889e-01 9.08360302e-01
1.07669286e-01 -4.38946486e-01 -3.06829721e-01 -5.09103179e-01
3.31002951e-01 9.43117023e-01 -3.64616424e-01 1.66253835e-01
-3.21299374e-01 -5.98837912e-01 -4.11929227e-02 5.35556018e-01
2.75722295e-01 -1.10065126e+00 -8.65754247e-01 8.57109129e-01
3.42575222e-01 -1.09780061e+00 3.30649316e-02 1.61366284e-01
-1.11014605e+00 -1.32670486e+00 -1.46924019e-01 7.33599141e-02
1.41820177e-01 -1.37030408e-01 1.21931720e+00 -3.86029363e-01
1.14226863e-01 7.57396519e-01 -2.40002304e-01 -9.52786446e-01
-6.34473324e-01 -1.20167941e-01 9.59711615e-03 8.15148652e-03
3.91976804e-01 -5.49271882e-01 -6.21794879e-01 3.56350094e-01
-5.64683378e-01 -3.12103122e-01 8.79511178e-01 6.59850180e-01
5.87180436e-01 2.90725410e-01 8.52770388e-01 -1.13279629e+00
1.01527393e+00 -3.46876770e-01 -8.04010570e-01 3.64320338e-01
-1.54534519e+00 2.76967287e-01 5.22622108e-01 -4.58237201e-01
-1.29660070e+00 -1.99459329e-01 4.06582743e-01 -2.92834550e-01
-4.20280129e-01 4.33417201e-01 -1.16738878e-01 5.61960161e-01
5.23914874e-01 -1.02668880e-02 1.53940246e-01 -2.97179937e-01
1.66862890e-01 2.12532386e-01 1.35883912e-01 -7.16914177e-01
4.74020451e-01 5.09016573e-01 1.39547423e-01 -6.08562946e-01
-4.37146425e-01 -5.86938798e-01 -2.51967877e-01 -4.53533530e-01
4.67323601e-01 -3.92872393e-01 -1.29614425e+00 -3.65225896e-02
-6.02832496e-01 -7.07420528e-01 -6.08636081e-01 3.40159297e-01
-9.69526112e-01 2.34923780e-01 -8.07924926e-01 -1.42780089e+00
-1.34146422e-01 -9.88511682e-01 8.36094320e-01 2.84183204e-01
-5.02458513e-01 -1.13509214e+00 2.70327851e-02 2.38095298e-01
2.61913061e-01 2.22541735e-01 6.36925876e-01 -9.85219955e-01
-8.14795852e-01 -3.23425025e-01 3.82801175e-01 1.92443848e-01
4.21662748e-01 3.39496583e-01 -8.49520981e-01 -4.60202619e-02
1.37005702e-01 2.56233633e-01 3.31151664e-01 5.10887861e-01
1.19640791e+00 -5.98087311e-01 -2.94632167e-01 -2.88317233e-01
1.41053247e+00 6.74344063e-01 6.13017440e-01 8.28413367e-01
2.49899492e-01 9.00686443e-01 1.03771079e+00 8.12889636e-01
1.56800911e-01 4.23919439e-01 4.80432063e-01 2.98172742e-01
5.83900988e-01 -3.90390396e-01 5.49983203e-01 2.76631415e-01
-4.13382918e-01 -1.89227864e-01 -1.02208459e+00 4.12001401e-01
-2.36258507e+00 -1.43408024e+00 -5.37995324e-02 2.61801887e+00
6.13629401e-01 5.58051467e-01 2.91265041e-01 3.30616266e-01
5.74082375e-01 -4.61391807e-02 -3.45269769e-01 -7.41662920e-01
3.66327256e-01 3.07884783e-01 8.41566682e-01 3.73362601e-01
-9.68237460e-01 4.08305049e-01 6.49723196e+00 4.34705943e-01
-6.15813911e-01 8.74598845e-05 5.96697748e-01 -1.75801903e-01
-1.10773690e-01 3.47681016e-01 -7.57302582e-01 4.23963964e-01
1.46940994e+00 -2.22964153e-01 6.62775934e-01 8.40823472e-01
1.01028681e+00 -3.68838578e-01 -1.65028334e+00 7.45374322e-01
-6.72096848e-01 -1.35281408e+00 -2.83159554e-01 3.22396040e-01
8.31775486e-01 -4.08515215e-01 -4.25926059e-01 4.32733536e-01
6.05857968e-01 -1.18048322e+00 7.43544638e-01 9.55770969e-01
-1.41590163e-01 -9.08816457e-01 7.22146869e-01 4.86749560e-01
-8.28499258e-01 -7.21818864e-01 4.69457917e-02 -4.22953367e-01
3.13516706e-01 7.21948862e-01 -1.04508018e+00 7.18304694e-01
5.48319101e-01 5.69554150e-01 -2.93264240e-01 1.01975584e+00
-3.84921342e-01 8.61231983e-01 -1.16114348e-01 -2.79344112e-01
1.64191455e-01 -4.97456282e-01 4.42018509e-01 1.15996301e+00
2.58257035e-02 -4.88593996e-01 4.72681761e-01 7.69393146e-01
4.62790877e-01 8.40000287e-02 -5.22027314e-01 -3.81333351e-01
7.33719885e-01 8.42340648e-01 -6.93001509e-01 -2.89553821e-01
-4.45476890e-01 3.59234989e-01 -6.07296340e-02 2.39076599e-01
-3.74510735e-01 2.60680228e-01 6.97118878e-01 3.22605342e-01
1.70874327e-01 -1.76722854e-01 -6.28057837e-01 -5.82779050e-01
-2.36726850e-01 -1.18253672e+00 4.69840676e-01 -2.21579567e-01
-1.44247282e+00 -3.35302234e-01 2.84422547e-01 -9.12071109e-01
-4.76268113e-01 -6.67471170e-01 -5.90526283e-01 4.73701924e-01
-1.40469074e+00 -5.25911570e-01 -9.74343792e-02 1.54501066e-01
7.56388068e-01 1.83432579e-01 6.55953348e-01 -7.56098777e-02
-6.99627817e-01 -2.32350126e-01 1.68452397e-01 -4.63529676e-01
5.21429539e-01 -1.61967981e+00 8.81499201e-02 7.61866808e-01
-1.26752071e-02 5.31227469e-01 9.67363894e-01 -7.88388371e-01
-1.58029056e+00 -7.95802832e-01 5.52497327e-01 -5.05410790e-01
9.47683334e-01 -4.28099819e-02 -7.36892581e-01 6.49469376e-01
3.07682216e-01 -6.73956573e-01 5.56136191e-01 3.89864624e-01
4.75992322e-01 -2.98518687e-01 -7.69938529e-01 8.56451511e-01
8.95402193e-01 -1.37352318e-01 -5.50605059e-01 3.19640726e-01
4.19782579e-01 -8.86708423e-02 -1.31295443e+00 2.70239532e-01
5.02541900e-01 -1.06736386e+00 7.58165717e-01 -5.17614961e-01
4.12391603e-01 -4.03504632e-02 1.74160466e-01 -1.10692155e+00
-1.67797536e-01 -1.01127374e+00 -6.11453772e-01 1.24362874e+00
3.21075290e-01 -6.76416278e-01 6.52852356e-01 7.76580632e-01
1.68134749e-01 -8.65982652e-01 -5.14977634e-01 -1.05485106e+00
-3.36184174e-01 -6.40942812e-01 6.53379679e-01 7.68268287e-01
4.50582318e-02 1.51181549e-01 -1.84653133e-01 2.50714850e-02
8.05498958e-01 1.99463278e-01 7.84011364e-01 -1.41232288e+00
-7.24839926e-01 -6.98038578e-01 -4.74294312e-02 -6.52499735e-01
-1.74057961e-01 -1.52557209e-01 2.89727803e-02 -1.66414607e+00
-2.17040390e-01 -2.81441540e-01 -3.80539954e-01 2.60236472e-01
-5.87577447e-02 -8.34927022e-01 2.17247143e-01 1.42351046e-01
-4.80760485e-01 4.72833931e-01 1.37962782e+00 3.49346995e-02
-5.62373340e-01 6.15702868e-01 -6.92927122e-01 6.30072832e-01
1.04119313e+00 -4.35129672e-01 -6.35157585e-01 4.07135069e-01
2.13361770e-01 3.78535211e-01 3.91913533e-01 -6.65074944e-01
3.52764964e-01 -7.42020130e-01 1.94899619e-01 -1.96216553e-01
-2.68897484e-03 -7.72429228e-01 4.45768327e-01 7.79694140e-01
-4.59809244e-01 2.46365160e-01 -3.08014434e-02 1.03685045e+00
-1.06652670e-01 -3.18555087e-01 3.46890211e-01 -5.12934804e-01
-7.33894110e-01 2.06373841e-01 -6.34187460e-01 -1.56104401e-01
1.11592972e+00 -4.81949598e-01 -1.52373493e-01 -4.57158566e-01
-8.74943435e-01 4.29565489e-01 2.10039914e-01 3.00057232e-01
1.88381642e-01 -7.10021138e-01 -3.90629739e-01 -2.82060713e-01
-1.11825414e-01 5.06025972e-04 -2.51447827e-01 1.07047558e+00
-3.01060200e-01 5.40400445e-01 1.15755480e-02 -5.41487038e-01
-6.89332843e-01 7.27297723e-01 3.32999676e-01 -8.38459909e-01
-6.19269073e-01 -1.28187597e-01 -2.27486104e-01 -1.49971083e-01
2.72002637e-01 -4.25432265e-01 -1.66700289e-01 7.93582126e-02
3.62846076e-01 7.91522741e-01 1.20804004e-01 4.81124640e-01
-6.42927215e-02 -3.22136015e-01 2.06531748e-01 -4.48340863e-01
1.41482949e+00 -1.41154051e-01 4.91046496e-02 9.34469521e-01
-1.33738515e-03 -5.93992881e-02 -1.68830049e+00 9.90117434e-03
9.98998702e-01 -5.68190455e-01 -1.19473651e-01 -8.17774534e-01
-6.22320712e-01 3.84212166e-01 4.83792126e-01 5.62359631e-01
8.47878754e-01 -2.50180006e-01 6.50191605e-02 3.51369143e-01
4.77400035e-01 -1.41919601e+00 2.36381888e-01 2.51793593e-01
6.44623637e-01 -1.08500278e+00 9.16754901e-02 -3.25975001e-01
-8.21754038e-01 1.03618228e+00 5.04744828e-01 -3.37864794e-02
5.68625629e-01 2.39764690e-01 -3.70375633e-01 -9.53453258e-02
-1.14479494e+00 -1.79270148e-01 -2.08706155e-01 5.37011385e-01
1.99795038e-01 1.67991221e-01 -4.19357121e-01 4.84848082e-01
-5.94645413e-03 2.23464668e-01 6.29950285e-01 1.30635381e+00
-3.20875376e-01 -1.38369477e+00 -5.16046286e-01 7.35923111e-01
-5.92949331e-01 2.25833148e-01 -2.73334295e-01 1.21507645e+00
-3.09195787e-01 1.36037838e+00 1.21828347e-01 2.54885703e-02
7.56940186e-01 3.31536382e-01 5.31540334e-01 -6.50580585e-01
-7.74232030e-01 1.27211526e-01 3.99359345e-01 -7.66786993e-01
-4.37749982e-01 -1.17956030e+00 -1.01024127e+00 -4.92884874e-01
-1.28041625e-01 9.41497982e-02 5.98793805e-01 1.08833194e+00
1.87084734e-01 7.86002219e-01 6.84204638e-01 -5.61619043e-01
-1.06852889e+00 -8.39040399e-01 -5.61633706e-01 3.80989760e-01
-5.87555990e-02 -7.13910043e-01 -1.19050078e-01 -4.09112684e-02] | [8.58629322052002, 5.956803321838379] |
45c29ed0-b4dd-49f8-9f00-7194a2cbb4f9 | combining-global-and-local-attention-with | null | null | https://www.iti.gr/~bmezaris/publications/ism2021a_preprint.pdf | https://www.iti.gr/~bmezaris/publications/ism2021a_preprint.pdf | Combining Global and Local Attention with Positional Encoding for Video Summarization | This paper presents a new method for supervised video summarization. To overcome drawbacks of existing RNN-based summarization architectures, that relate to the modeling of long-range frames' dependencies and the ability to parallelize the training process, the developed model relies on the use of self-attention mechanisms to estimate the importance of video frames. Contrary to previous attention-based summarization approaches that model the frames' dependencies by observing the entire frame sequence, our method combines global and local multi-head attention mechanisms to discover different modelings of the frames' dependencies at different levels of granularity. Moreover, the utilized attention mechanisms integrate a component that encodes the temporal position of video frames - this is of major importance when producing a video summary. Experiments on two datasets (SumMe and TVSum) demonstrate the effectiveness of the proposed model compared to existing attention-based methods, and its competitiveness against other state-of-the-art supervised summarization approaches. An ablation study that focuses on our main proposed components, namely the use of global and local multi-head attention mechanisms in collaboration with an absolute positional encoding component, shows their relative contributions to the overall summarization performance. | ['Ioannis Patras', 'Vasileios Mezaris', 'Georgios Balaouras', 'Evlampios Apostolidis'] | 2021-12-01 | null | null | null | ieee-international-symposium-on-multimedia-1 | ['supervised-video-summarization'] | ['computer-vision'] | [ 2.79014230e-01 3.94492656e-01 -2.54655689e-01 -1.22021019e-01
-8.71765733e-01 -1.39607102e-01 7.91803181e-01 3.88913780e-01
-5.86975873e-01 8.03394914e-01 9.54939485e-01 1.00226864e-01
-1.50716707e-01 -4.08546358e-01 -6.86322212e-01 -5.12989342e-01
4.18317728e-02 3.57305557e-01 4.60248739e-01 -3.03092659e-01
7.82098711e-01 2.34814078e-01 -1.79368520e+00 4.56741333e-01
8.63762200e-01 6.72612607e-01 4.59670693e-01 9.81448770e-01
-3.97863865e-01 1.32053232e+00 -1.07441270e+00 -1.32792592e-01
-3.45983773e-01 -9.03159738e-01 -8.75643909e-01 1.32150501e-01
6.69693291e-01 -4.94233549e-01 -2.64185131e-01 7.97766507e-01
6.37530386e-01 3.70320082e-01 6.96472168e-01 -6.50628567e-01
-5.91787159e-01 7.87634611e-01 -5.67320347e-01 7.81601191e-01
6.85873866e-01 -1.87480286e-01 1.18655562e+00 -4.60098952e-01
7.25542367e-01 9.91829395e-01 5.63659608e-01 4.01985019e-01
-6.84134543e-01 -1.63633272e-01 5.11863589e-01 7.85537601e-01
-8.79676998e-01 -6.28545463e-01 8.52735877e-01 -3.46601069e-01
1.30061805e+00 4.47489887e-01 6.44946277e-01 9.51471269e-01
2.60396123e-01 9.51490283e-01 4.34233040e-01 -5.32292485e-01
1.70847058e-01 -2.33198777e-01 4.70648110e-01 4.38742787e-01
1.70314431e-01 -5.94533086e-01 -7.56575584e-01 1.86921448e-01
6.31444216e-01 -2.71194607e-01 -4.29411709e-01 9.84811690e-03
-1.14887762e+00 5.64609647e-01 -6.32231124e-03 7.15700865e-01
-8.19777071e-01 2.55673289e-01 7.70158231e-01 -5.43780401e-02
8.38030040e-01 4.17172194e-01 -7.33640790e-02 -3.76850754e-01
-1.51537442e+00 1.72552973e-01 7.02192426e-01 8.92778277e-01
4.32286054e-01 2.66121507e-01 -8.21700215e-01 5.28409839e-01
6.04500473e-02 -1.92290902e-01 7.68266559e-01 -9.26378906e-01
5.89006007e-01 6.04513943e-01 9.41188931e-02 -9.56861377e-01
-2.18722135e-01 -4.20407504e-01 -7.15997517e-01 -1.90549046e-01
-1.81733817e-01 -1.17437027e-01 -6.93217814e-01 1.51167786e+00
2.25074459e-02 3.28387946e-01 9.40912962e-02 7.01840639e-01
1.21001554e+00 8.40083897e-01 7.93748125e-02 -6.60777450e-01
1.31931508e+00 -1.36981857e+00 -1.12442005e+00 1.65133551e-01
1.45564899e-01 -5.91229439e-01 6.42489135e-01 1.78014129e-01
-1.69673729e+00 -8.72116685e-01 -1.16790664e+00 -3.11056703e-01
-1.87732130e-01 1.41931653e-01 1.78188816e-01 2.53393382e-01
-1.42363536e+00 7.47538626e-01 -6.55387282e-01 -7.24492311e-01
2.38264874e-01 3.50347310e-01 -7.85487518e-02 4.22236264e-01
-8.16391110e-01 9.39201236e-01 6.41794801e-01 1.32814318e-01
-7.36612737e-01 -4.33245540e-01 -7.08264828e-01 6.36015117e-01
3.08427125e-01 -1.00755131e+00 1.26067901e+00 -1.29255116e+00
-1.62088680e+00 2.84694910e-01 -4.37603205e-01 -8.16517830e-01
3.40173423e-01 -6.96724951e-01 2.98207160e-02 7.20193207e-01
8.44954625e-02 6.35009885e-01 9.08604264e-01 -1.08660805e+00
-7.62570500e-01 -4.44956012e-02 1.98305383e-01 5.29142022e-01
-4.90975231e-01 1.61026791e-01 -6.31054282e-01 -8.26442063e-01
-3.06004018e-01 -2.98531145e-01 -1.41727582e-01 -8.08472753e-01
-3.81206274e-01 -2.49551147e-01 8.64221811e-01 -1.02390838e+00
1.86298192e+00 -1.85642648e+00 7.31688380e-01 -4.20156926e-01
1.13096975e-01 5.02579570e-01 -8.81479234e-02 6.95741534e-01
-7.88488612e-02 9.12748054e-02 -1.78263977e-01 -7.60477245e-01
-1.93434507e-01 1.08535461e-01 -7.67184496e-02 1.34554937e-01
2.13944986e-01 8.04603755e-01 -9.46696281e-01 -7.44570255e-01
4.11520630e-01 5.81784546e-01 -4.59638298e-01 3.01963121e-01
-3.53101999e-01 4.56995875e-01 -3.62228632e-01 2.01620609e-01
3.50200832e-01 5.71243875e-02 2.77272314e-01 -4.31089014e-01
-3.69197547e-01 3.58151257e-01 -7.72214174e-01 1.87996399e+00
-1.74975872e-01 7.78870702e-01 -1.32731691e-01 -1.26701653e+00
5.91967940e-01 6.87011838e-01 5.36166430e-01 -4.16559637e-01
3.40279907e-01 -1.85553402e-01 -2.80701339e-01 -7.36933172e-01
1.19750929e+00 3.85368913e-01 1.53833449e-01 3.70035052e-01
5.30589104e-01 3.30347925e-01 8.06094468e-01 4.54796523e-01
9.83318210e-01 5.66080332e-01 6.70733690e-01 -2.80017614e-01
9.26567376e-01 -1.79758638e-01 4.06302840e-01 7.61856139e-01
-9.37205628e-02 8.29015911e-01 7.82905281e-01 -3.37924600e-01
-1.01724124e+00 -3.60618889e-01 4.90134239e-01 1.34739220e+00
5.58453016e-02 -7.87549853e-01 -1.15827835e+00 -6.89158440e-01
-5.77740967e-01 8.71355772e-01 -7.80913055e-01 6.65760934e-02
-8.61574173e-01 -5.83501697e-01 3.38986993e-01 5.60913622e-01
5.40759265e-01 -1.27636182e+00 -1.14072585e+00 3.93149257e-01
-5.29677272e-01 -9.30801213e-01 -4.50850010e-01 -2.12295994e-01
-9.87550259e-01 -1.08212650e+00 -8.61257017e-01 -6.50070250e-01
4.43796545e-01 2.40427986e-01 1.23070180e+00 -1.34935081e-02
1.15634032e-01 6.72546268e-01 -6.98338985e-01 -4.56419438e-01
-4.88337517e-01 3.68114561e-01 -4.38885421e-01 6.34094030e-02
7.15866610e-02 -7.73446918e-01 -5.30847549e-01 -2.94239163e-01
-1.02702081e+00 2.35096589e-01 7.53113031e-01 5.33637762e-01
4.10031199e-01 -1.98843867e-01 7.37098932e-01 -9.22242939e-01
8.53213072e-01 -4.81579900e-01 -7.42473826e-02 3.61342460e-01
-1.20235860e-01 2.20377937e-01 4.74637866e-01 -1.90122947e-01
-1.38256049e+00 -3.02487135e-01 -1.54997528e-01 -3.21514010e-01
-1.87228292e-01 4.86780643e-01 -9.88713186e-03 4.50815678e-01
2.74730325e-01 4.24691230e-01 -7.01176673e-02 -4.38016057e-01
2.91960120e-01 3.21518183e-01 5.84832072e-01 -3.10567707e-01
1.34730533e-01 2.39587024e-01 -1.79126725e-01 -1.11455822e+00
-6.68826461e-01 -5.97742319e-01 -7.37573624e-01 -4.54407036e-01
1.23132896e+00 -8.00782323e-01 -4.24802899e-01 3.40010285e-01
-1.63670754e+00 1.35573536e-01 -5.08602083e-01 3.89403373e-01
-7.02532768e-01 8.87582123e-01 -6.07190490e-01 -8.06890249e-01
-8.11212122e-01 -1.00674832e+00 1.12011635e+00 2.84266919e-01
-5.00862300e-01 -9.84131932e-01 1.75210819e-01 4.12217736e-01
5.33707619e-01 2.50255704e-01 8.44107032e-01 -8.27522516e-01
-5.58627129e-01 4.83075380e-02 -5.30146398e-02 4.08213079e-01
-2.87570879e-02 2.32222915e-01 -7.50827491e-01 1.52693689e-02
6.71697631e-02 4.64786068e-02 1.12401009e+00 9.00233507e-01
8.79336655e-01 -4.56055611e-01 -1.94474645e-02 1.19758449e-01
1.32754707e+00 2.54316002e-01 9.06301260e-01 4.64968026e-01
6.62595034e-01 4.49733943e-01 5.82104981e-01 6.42118931e-01
3.98679942e-01 5.58714271e-01 6.73656285e-01 2.15633214e-02
-3.74736816e-01 7.47706443e-02 4.04761583e-01 1.15314484e+00
-7.47605741e-01 -5.78953147e-01 -2.52770752e-01 6.85074329e-01
-2.34228516e+00 -1.38809776e+00 -8.27426687e-02 1.97606504e+00
2.58272022e-01 1.06929921e-01 2.80539185e-01 1.14834376e-01
8.61403346e-01 7.14514613e-01 -2.25558698e-01 -8.30824971e-01
-1.20874740e-01 8.47281441e-02 2.59512097e-01 4.25646991e-01
-1.07304561e+00 7.60959625e-01 6.61398029e+00 7.97484398e-01
-8.41102540e-01 1.99279100e-01 3.60085309e-01 -2.63536245e-01
-2.33856037e-01 -2.36215740e-01 -5.71370184e-01 4.09772933e-01
1.05032301e+00 -7.47047961e-02 -6.27728701e-02 4.72691327e-01
4.83478189e-01 -3.11558038e-01 -1.00338244e+00 5.72134674e-01
7.14754403e-01 -1.47538960e+00 5.37722170e-01 -2.61340976e-01
7.52388120e-01 -2.31788665e-01 -2.68669993e-01 2.03404739e-01
-1.45139515e-01 -6.25659883e-01 9.78309095e-01 7.42671072e-01
3.32719415e-01 -7.39823103e-01 1.10678089e+00 2.25579023e-01
-1.21865416e+00 -8.82201791e-02 -3.08240771e-01 -2.08844423e-01
5.14808118e-01 2.40932927e-01 -6.01354778e-01 1.15076232e+00
4.79711026e-01 9.33324456e-01 -5.85028410e-01 8.71011436e-01
-3.87900889e-01 6.01210117e-01 1.67286783e-01 -3.17098037e-03
3.49814802e-01 -2.54152566e-01 9.51617837e-01 1.68090320e+00
3.78727674e-01 4.19361964e-02 -1.85515016e-01 4.11787868e-01
5.95957972e-02 2.42609009e-01 -3.30510765e-01 2.17462555e-02
1.58119231e-01 1.03571475e+00 -6.08338714e-01 -7.91976929e-01
-4.12743270e-01 1.08230913e+00 2.63117880e-01 3.49872977e-01
-1.05242753e+00 -2.91001111e-01 2.18235523e-01 1.06728226e-01
7.93062150e-01 -2.34232377e-02 -2.82308049e-02 -1.04867160e+00
7.16399774e-02 -7.93852806e-01 5.01677096e-01 -8.37389112e-01
-5.27896464e-01 8.30343902e-01 2.96954900e-01 -1.05589902e+00
-5.42445838e-01 1.49016857e-01 -9.57581520e-01 5.71895480e-01
-1.47373056e+00 -1.33583510e+00 -2.54930943e-01 3.54444057e-01
1.29914773e+00 -1.53225288e-01 6.32094741e-01 1.29092693e-01
-7.38866091e-01 2.24329337e-01 -1.05414279e-01 -2.97737688e-01
6.06206954e-01 -1.10925579e+00 2.00028151e-01 1.18193424e+00
4.51654643e-02 4.17840093e-01 9.84190464e-01 -5.56325793e-01
-8.86767447e-01 -8.01840127e-01 1.18083823e+00 -1.91539109e-01
3.70524794e-01 2.80676514e-01 -8.66840124e-01 8.57404649e-01
9.90154028e-01 -7.01786458e-01 5.98528683e-01 -1.15767762e-01
1.03979774e-01 4.60488796e-02 -7.66766250e-01 4.41004694e-01
8.81504297e-01 -1.10654987e-01 -1.14942527e+00 -5.25567308e-02
7.08820105e-01 -2.05384836e-01 -5.75971425e-01 3.93326432e-01
5.08749366e-01 -1.55736732e+00 7.71297395e-01 -5.37116170e-01
9.62738872e-01 -1.85003191e-01 1.53526500e-01 -1.11256456e+00
-5.87761343e-01 -5.63857913e-01 -5.96704125e-01 1.56174636e+00
1.46439029e-02 -1.96898803e-01 5.68724990e-01 1.58709317e-01
-5.95376194e-01 -5.68594456e-01 -8.79232228e-01 -2.29582593e-01
-5.71052015e-01 1.82200167e-02 2.23525450e-01 5.32513559e-01
3.23108546e-02 6.12276256e-01 -6.74926519e-01 -8.89641512e-03
2.96450794e-01 -1.02151304e-01 7.89788008e-01 -9.54478443e-01
-3.02181631e-01 -6.84017479e-01 -6.03950739e-01 -1.02555847e+00
2.56291568e-01 -3.63899589e-01 -1.18034586e-01 -2.11114025e+00
4.21813369e-01 6.11135304e-01 -4.35860127e-01 5.85280582e-02
-3.75951558e-01 -6.35533454e-03 4.63827103e-01 1.92455471e-01
-1.26156378e+00 6.59524381e-01 9.08754408e-01 -2.03817971e-02
-3.29270840e-01 -1.87675625e-01 -7.81380951e-01 7.71386147e-01
6.15670800e-01 -1.60902321e-01 -5.14800489e-01 -6.83297873e-01
-1.66275576e-01 1.04942635e-01 8.06023851e-02 -1.34134185e+00
4.95620519e-01 2.60260522e-01 5.70559278e-02 -1.01477826e+00
3.35989833e-01 -5.03486335e-01 1.99199542e-01 3.53118837e-01
-5.33675969e-01 2.12558746e-01 2.24434078e-01 6.41538084e-01
-5.53035200e-01 -4.78704482e-01 3.98561209e-01 -3.31138015e-01
-9.18409646e-01 -1.65247977e-01 -7.48535275e-01 -2.45711654e-01
1.06131041e+00 -4.50451374e-01 -2.11428836e-01 -6.45562470e-01
-6.04737937e-01 1.90315694e-02 2.48640522e-01 3.15701544e-01
5.05874872e-01 -8.23340297e-01 -8.99327815e-01 -2.33865529e-01
-2.75930107e-01 4.36989293e-02 5.54289877e-01 8.62015665e-01
-6.53143466e-01 5.98102093e-01 -4.57602531e-01 -3.92720580e-01
-1.54328871e+00 6.28167987e-01 -9.83970240e-02 -7.26723850e-01
-5.55829108e-01 5.99536836e-01 2.78095365e-01 5.51556528e-01
3.04678112e-01 -3.96980584e-01 -9.07933056e-01 5.11568606e-01
6.82132304e-01 7.71620870e-01 -7.03534558e-02 -8.60898137e-01
-6.31066933e-02 4.83032316e-01 -1.65122494e-01 -1.91539526e-02
1.39103353e+00 -4.74416792e-01 -2.34175965e-01 5.13777554e-01
7.47650802e-01 2.96655092e-02 -1.12628973e+00 2.12126528e-03
-2.18045544e-02 -8.28015655e-02 -1.17997557e-01 -4.84868526e-01
-8.56338799e-01 6.82855308e-01 4.27564532e-02 3.03672403e-01
1.34055185e+00 -1.37676418e-01 6.76237285e-01 -1.10874875e-02
-7.03609958e-02 -1.09509838e+00 1.27693251e-01 4.53768939e-01
9.47380006e-01 -8.03452134e-01 3.64162385e-01 -2.90713191e-01
-6.53406024e-01 1.27412093e+00 4.43222433e-01 -1.53750762e-01
-5.55456504e-02 -2.04430651e-02 -3.64577055e-01 -1.07667640e-01
-8.93555760e-01 -2.38825962e-01 3.65156561e-01 3.39147538e-01
7.08709776e-01 -3.55652064e-01 -8.39146197e-01 5.59295118e-01
1.91948131e-01 1.16644718e-01 7.12649643e-01 9.33902442e-01
-7.28290617e-01 -9.61937368e-01 -3.50797266e-01 2.67264128e-01
-7.67815232e-01 -8.91523734e-02 -3.93825740e-01 8.09446335e-01
7.15152696e-02 8.63582909e-01 1.85607851e-01 9.92070232e-03
3.45996320e-01 1.53253645e-01 4.86044049e-01 -5.66108167e-01
-9.71993387e-01 2.24534646e-01 3.06188792e-01 -4.63325530e-01
-1.27672696e+00 -6.00569069e-01 -8.75601172e-01 -1.02455139e-01
-7.20648915e-02 3.18701506e-01 5.65458357e-01 1.05092943e+00
5.67187905e-01 1.23339808e+00 2.62685508e-01 -1.49031925e+00
-2.06651822e-01 -1.25521815e+00 -2.69092053e-01 4.76543039e-01
3.94811928e-01 -2.75362343e-01 -1.29398197e-01 2.76871055e-01] | [10.418336868286133, 0.43520259857177734] |
27b2b8f1-1991-4c15-a99d-3a017f355066 | learning-latent-semantic-annotations-for | null | null | https://aclanthology.org/D18-1411 | https://aclanthology.org/D18-1411.pdf | Learning Latent Semantic Annotations for Grounding Natural Language to Structured Data | Previous work on grounded language learning did not fully capture the semantics underlying the correspondences between structured world state representations and texts, especially those between numerical values and lexical terms. In this paper, we attempt at learning explicit latent semantic annotations from paired structured tables and texts, establishing correspondences between various types of values and texts. We model the joint probability of data fields, texts, phrasal spans, and latent annotations with an adapted semi-hidden Markov model, and impose a soft statistical constraint to further improve the performance. As a by-product, we leverage the induced annotations to extract templates for language generation. Experimental results suggest the feasibility of the setting in this study, as well as the effectiveness of our proposed framework. | ['Chin-Yew Lin', 'Jin-Ge Yao', 'Guanghui Qin', 'Jinpeng Wang', 'Xuening Wang'] | 2018-10-01 | null | null | null | emnlp-2018-10 | ['grounded-language-learning'] | ['natural-language-processing'] | [ 2.14402750e-01 3.92959297e-01 -6.39584959e-01 -4.37185705e-01
-7.83315063e-01 -6.19703829e-01 1.02739680e+00 3.76782805e-01
-3.12104434e-01 8.92022371e-01 1.00777614e+00 -1.91755041e-01
1.35416672e-01 -9.24503922e-01 -8.02460611e-01 -3.03664088e-01
1.29791856e-01 3.88384432e-01 1.47437915e-01 -1.17790371e-01
3.96450236e-03 -2.25244045e-01 -1.11514902e+00 5.45975566e-01
7.98998833e-01 6.41914666e-01 5.41190915e-02 8.45204592e-02
-6.42618954e-01 1.29622483e+00 -5.44978797e-01 -7.37398803e-01
-5.64085320e-02 -5.35822809e-01 -9.15182173e-01 2.22521096e-01
-1.45698413e-01 -3.96087557e-01 -2.69696951e-01 8.67608964e-01
7.78600127e-02 2.15597421e-01 7.98208773e-01 -1.34125710e+00
-6.22302532e-01 1.60590327e+00 -2.33280852e-01 -8.57354179e-02
6.16974652e-01 -2.02119760e-02 1.49020576e+00 -8.28400612e-01
6.49369597e-01 1.37100947e+00 6.19829953e-01 3.83369923e-01
-1.18179333e+00 -4.34430927e-01 3.99989188e-01 -1.99502166e-02
-1.29042983e+00 -6.38837874e-01 9.26780820e-01 -3.53664398e-01
9.57029939e-01 -1.87310427e-01 5.10263920e-01 1.51416695e+00
-1.08549766e-01 9.62118447e-01 9.07105863e-01 -7.37900972e-01
3.53858769e-01 4.01002735e-01 1.97264254e-01 7.06210494e-01
4.40610647e-01 -2.48757720e-01 -8.60277176e-01 -3.27206790e-01
7.85712421e-01 -1.06592432e-01 1.36512890e-01 -3.61825645e-01
-1.36052859e+00 9.10351694e-01 -1.10115446e-01 4.29926276e-01
-2.91191667e-01 1.50675490e-01 3.69174242e-01 -1.39180347e-01
4.60804999e-01 2.29934961e-01 -4.88775283e-01 -1.59772173e-01
-7.89549410e-01 1.85219079e-01 9.30009842e-01 1.36171579e+00
8.20074499e-01 -1.13326445e-01 -2.70289451e-01 6.45543516e-01
7.19904482e-01 4.89594579e-01 4.44404185e-01 -6.87009931e-01
1.00937819e+00 6.91215098e-01 3.86998892e-01 -7.73281693e-01
-1.38050959e-01 6.61790222e-02 -4.26955730e-01 -7.98524857e-01
4.01647717e-01 -3.26814771e-01 -8.65453601e-01 1.91665912e+00
2.74217039e-01 2.68232763e-01 5.93533695e-01 5.13702869e-01
6.82674527e-01 8.45840216e-01 5.00783384e-01 -2.33497649e-01
1.39541924e+00 -5.43897152e-01 -1.10891759e+00 -2.67060816e-01
9.47073936e-01 -3.43672961e-01 1.18809891e+00 -3.87883365e-01
-1.04619098e+00 -2.91941315e-01 -7.94915557e-01 -2.47777060e-01
-3.46351206e-01 2.33344123e-01 6.35669231e-01 4.23502594e-01
-4.95330393e-01 2.78166503e-01 -1.26099515e+00 -2.85795510e-01
5.82878888e-02 -8.06291103e-02 -4.42804620e-02 3.08930516e-01
-1.74408710e+00 7.36348331e-01 9.80520129e-01 -3.77376117e-02
-5.23189902e-01 -5.36442995e-01 -1.33216369e+00 1.48280516e-01
6.01595759e-01 -7.17188954e-01 1.26479697e+00 -3.09254706e-01
-1.53725946e+00 5.66691518e-01 -4.53139246e-01 -6.89568758e-01
2.38194376e-01 -2.63508230e-01 -2.37450913e-01 -6.85630217e-02
2.30831787e-01 4.85053957e-01 4.82901871e-01 -1.22981381e+00
-6.25355899e-01 -2.17518896e-01 3.52433205e-01 1.92739040e-01
-5.16830683e-01 2.36173067e-02 -2.66348779e-01 -7.40447044e-01
1.35124400e-01 -7.64933646e-01 -2.56158203e-01 -4.52661872e-01
-7.27548599e-01 -4.44411546e-01 1.72368035e-01 -6.12164915e-01
1.45843780e+00 -2.04246879e+00 2.45633419e-03 2.50450879e-01
-1.38949513e-01 -3.35404724e-01 3.07144076e-02 6.47013724e-01
3.14493865e-01 2.48644873e-01 -3.66210252e-01 -4.43685114e-01
4.66382891e-01 4.16200399e-01 -9.29987967e-01 1.04811698e-01
3.41530263e-01 9.61359143e-01 -9.14119899e-01 -7.60918319e-01
2.62178276e-02 3.12324405e-01 -5.42522252e-01 2.52388954e-01
-6.18415236e-01 8.47782493e-02 -7.48295128e-01 1.70441061e-01
5.49911223e-02 -3.43920171e-01 6.55959308e-01 -1.80169925e-01
6.85936585e-02 1.09135866e+00 -1.37302303e+00 1.76922607e+00
-6.31981969e-01 1.04940481e-01 -5.36210954e-01 -6.98207259e-01
8.99599791e-01 5.59468806e-01 3.60007703e-01 -3.81644189e-01
-2.57417895e-02 -1.18485354e-01 -2.73340344e-01 -5.28331399e-01
6.00524604e-01 -3.87317777e-01 -4.80774581e-01 6.66921079e-01
5.99241070e-02 -1.58335716e-01 2.95102566e-01 4.12461936e-01
7.01683521e-01 4.71967280e-01 6.24384820e-01 7.46430978e-02
3.37555856e-01 -8.67405161e-03 5.07252753e-01 6.75942659e-01
2.96474576e-01 2.12099537e-01 7.09076524e-01 6.66281506e-02
-1.01258731e+00 -1.08789885e+00 -3.26545141e-03 1.09403276e+00
6.64844811e-02 -9.40273583e-01 -7.45158851e-01 -6.96275711e-01
-7.04375207e-02 1.32914925e+00 -4.41675872e-01 -9.18046013e-02
-5.64129174e-01 -5.35505295e-01 6.47853136e-01 9.48611856e-01
2.12542251e-01 -1.14904571e+00 -3.99029523e-01 3.12139481e-01
-5.90472043e-01 -1.59033954e+00 -2.78384835e-01 1.98143512e-01
-7.30840087e-01 -7.08127677e-01 -1.04505703e-01 -6.26409173e-01
5.24645388e-01 -2.35828727e-01 9.67132330e-01 -3.27119738e-01
4.48750556e-01 2.21698433e-01 -3.33591998e-01 -3.17933202e-01
-6.68886542e-01 3.34608912e-01 -4.92919609e-02 3.49079929e-02
2.94973224e-01 -2.37624481e-01 3.49103771e-02 -1.18233144e-01
-9.79834139e-01 3.50497782e-01 4.12361473e-01 6.88777208e-01
3.18909764e-01 -7.11850896e-02 2.17983916e-01 -1.17516601e+00
5.79108238e-01 -6.00449443e-01 -4.92418647e-01 4.85745162e-01
-3.85507494e-01 7.18277156e-01 5.49190879e-01 -2.75782317e-01
-1.49005961e+00 7.32955560e-02 -3.87177075e-04 -6.63849041e-02
-3.19419503e-01 1.02629375e+00 -5.23315728e-01 9.90510345e-01
2.77483940e-01 2.30318904e-01 -3.69988203e-01 -4.60045874e-01
7.55393565e-01 6.78512275e-01 3.87351304e-01 -9.87034976e-01
8.87724638e-01 3.61644059e-01 -2.43780538e-01 -6.40525579e-01
-1.37528205e+00 -2.84625798e-01 -8.00421834e-01 4.27163243e-01
9.18060601e-01 -1.10299528e+00 -1.72550380e-01 3.91016901e-02
-1.36551392e+00 -4.21214879e-01 -5.00382125e-01 5.42117476e-01
-6.03730619e-01 3.54672402e-01 -8.04457247e-01 -9.25592959e-01
-8.18045363e-02 -7.62448788e-01 1.20257604e+00 -1.21740781e-01
-5.71317434e-01 -1.47049773e+00 1.30606055e-01 3.30163807e-01
-1.03817135e-02 1.16944782e-01 1.25223470e+00 -8.77353489e-01
-5.08695602e-01 -2.15024520e-02 1.50061666e-03 -7.70687684e-02
4.31132853e-01 -7.44692907e-02 -6.68090284e-01 6.06980920e-02
-1.20915711e-01 -4.76777405e-01 6.92534864e-01 3.98629345e-02
7.59940207e-01 -7.50065327e-01 -3.12884063e-01 3.84957314e-01
1.11589801e+00 1.35795400e-02 3.86747301e-01 1.99154183e-01
8.08813393e-01 7.04583824e-01 5.33647120e-01 8.01462650e-01
8.11997533e-01 7.50705481e-01 -1.85064510e-01 1.48035794e-01
1.78322509e-01 -9.66221750e-01 6.10946655e-01 9.90259945e-01
3.62868607e-01 -3.13450545e-01 -1.13714492e+00 6.81287766e-01
-1.84072387e+00 -1.00806475e+00 1.42052576e-01 2.00685668e+00
1.38820958e+00 3.83253843e-01 -1.07210085e-01 -9.42489728e-02
7.53332019e-01 2.97310293e-01 -2.86548942e-01 7.19518214e-02
1.01511672e-01 1.72807023e-01 2.76172906e-01 5.31615317e-01
-1.01188743e+00 1.24865544e+00 6.37995338e+00 5.84958494e-01
-7.74874926e-01 -3.34544033e-02 1.93256974e-01 1.33607939e-01
-8.52192104e-01 3.78188133e-01 -1.23138487e+00 5.27209342e-01
1.16325223e+00 -4.73762989e-01 2.03486100e-01 7.38829374e-01
2.23697990e-01 3.33471090e-01 -1.28622317e+00 5.44241011e-01
3.10653001e-02 -1.30390716e+00 4.43202049e-01 -6.55144602e-02
6.57541275e-01 -3.94882739e-01 -2.11486459e-01 3.19964975e-01
8.57117116e-01 -6.12438142e-01 1.26598835e+00 3.04229856e-01
7.40595818e-01 -3.53884161e-01 5.88093102e-01 3.26214910e-01
-1.20373809e+00 -3.77775691e-02 -2.53223896e-01 -1.51187986e-01
2.29952022e-01 2.88602948e-01 -1.04252458e+00 5.30796111e-01
8.01604390e-02 9.12198782e-01 -3.85263354e-01 3.89498085e-01
-7.65298784e-01 9.72454011e-01 -2.85955280e-01 -1.75699934e-01
1.58492133e-01 1.66130457e-02 2.84970999e-01 1.27221727e+00
1.55709520e-01 1.93545148e-02 4.63945210e-01 1.20393753e+00
-1.55901879e-01 1.31886140e-01 -6.42751396e-01 -5.30625165e-01
8.84719014e-01 8.94134939e-01 -8.29083323e-01 -6.44772947e-01
-6.06287539e-01 4.34778899e-01 3.33808690e-01 5.22314131e-01
-9.89694953e-01 -1.26378179e-01 4.53300744e-01 1.12798177e-01
3.40392798e-01 -3.13062131e-01 -4.87553447e-01 -1.59279573e+00
9.64072049e-02 -6.05513215e-01 4.90143567e-01 -6.71832860e-01
-1.34327316e+00 2.17429131e-01 5.64608932e-01 -1.09404850e+00
-7.94270515e-01 -2.54573405e-01 -3.86758715e-01 6.41150057e-01
-1.30054259e+00 -1.32898271e+00 8.64187405e-02 5.71395993e-01
4.90112722e-01 9.34927091e-02 7.55091488e-01 -3.39903951e-01
-6.61542475e-01 4.45437431e-01 -1.52214929e-01 6.04359567e-01
4.38393652e-01 -1.27569044e+00 6.67420745e-01 9.86575246e-01
6.05053663e-01 1.11480427e+00 6.79128468e-01 -8.90800774e-01
-1.26575458e+00 -1.18705285e+00 1.07991564e+00 -5.72278619e-01
1.03312171e+00 -7.77269185e-01 -9.41914737e-01 1.09501088e+00
2.26185784e-01 -2.97700733e-01 8.28223825e-01 4.19292450e-01
-4.59422648e-01 3.01535219e-01 -6.64014935e-01 7.50082612e-01
1.03986967e+00 -9.70468223e-01 -1.04098427e+00 1.97297543e-01
1.10234833e+00 -4.72943842e-01 -7.60705888e-01 1.05830826e-01
2.64550537e-01 -3.89677167e-01 7.45699465e-01 -7.40446627e-01
6.31393135e-01 -1.05838194e-01 -2.55816013e-01 -1.18429828e+00
-1.57583266e-01 -5.45270920e-01 -4.15068477e-01 1.67341995e+00
8.01188827e-01 -3.94935489e-01 7.28235781e-01 8.22476804e-01
-8.88701528e-03 -4.03528869e-01 -6.45734251e-01 -5.61161876e-01
5.58412187e-02 -6.06535614e-01 8.71257842e-01 1.02276349e+00
5.05258739e-01 7.10055232e-01 -2.28145704e-01 3.01456124e-01
5.46942294e-01 1.91195518e-01 5.39557576e-01 -9.88656700e-01
-4.15865511e-01 2.12951005e-02 -5.53091951e-02 -9.67105031e-01
7.06307769e-01 -9.35631573e-01 1.97900131e-01 -1.76514637e+00
2.95523822e-01 -4.75285053e-01 -3.53603154e-01 8.79992723e-01
-4.62244183e-01 -2.46175632e-01 2.08719328e-01 2.30079547e-01
-7.28452206e-01 7.45638967e-01 7.76497543e-01 -3.60121988e-02
-3.49552006e-01 -2.16942221e-01 -6.64040506e-01 6.93278968e-01
7.18957007e-01 -6.08746469e-01 -7.34535992e-01 -3.85675132e-01
3.68357271e-01 3.09792846e-01 2.33144566e-01 -6.49827957e-01
1.10538289e-01 -5.05946279e-01 2.03743994e-01 -3.08437407e-01
2.53420770e-01 -7.56177902e-01 -1.30915821e-01 1.11987486e-01
-1.04037857e+00 -2.20670894e-01 4.61074971e-02 4.55882967e-01
-2.09838331e-01 -3.37431937e-01 1.60705313e-01 -1.17952749e-01
-5.70527494e-01 -2.70464122e-02 -3.24585646e-01 4.52797592e-01
7.89390862e-01 -7.11103231e-02 -1.53759882e-01 -3.71718138e-01
-6.29149199e-01 1.50608659e-01 3.41028988e-01 4.74861890e-01
4.37353373e-01 -1.30146050e+00 -4.48612958e-01 3.55869085e-01
1.86924174e-01 -1.16355658e-01 -2.34435961e-01 4.79277492e-01
8.61284211e-02 5.16761124e-01 1.77126061e-02 -2.56745994e-01
-7.80630827e-01 5.08278966e-01 -2.89761256e-02 -5.27763605e-01
-6.12724364e-01 3.07560503e-01 6.29053563e-02 -4.30141568e-01
3.94275516e-01 -7.52814054e-01 -3.15349460e-01 1.75176010e-01
2.46325269e-01 -3.48863862e-02 -2.36538202e-01 -4.67312455e-01
-8.12374279e-02 -3.78267653e-02 -1.42790467e-01 -5.45799255e-01
1.22267854e+00 -3.19768906e-01 5.66784702e-02 9.10491168e-01
8.43191683e-01 5.93944602e-02 -1.24059236e+00 -5.97507954e-01
6.41395688e-01 -2.08137147e-02 -9.48911905e-02 -5.15719354e-01
-4.95281637e-01 7.17571735e-01 -2.12306648e-01 4.73363064e-02
4.96059358e-01 3.13821286e-01 1.04369843e+00 4.54418451e-01
4.00696874e-01 -8.97689402e-01 8.54551122e-02 7.11397529e-01
4.78567302e-01 -9.24404085e-01 -1.95313379e-01 -5.78255117e-01
-8.63673568e-01 8.67582083e-01 6.30689740e-01 3.03753287e-01
3.80964905e-01 6.63064420e-01 7.18121277e-03 -8.77631456e-02
-1.05915427e+00 -2.76735932e-01 1.68573812e-01 4.26380724e-01
7.81811893e-01 -6.79127350e-02 -5.59220649e-02 7.94480145e-01
-4.01950151e-01 4.43446115e-02 5.96824646e-01 1.15221071e+00
-2.45157450e-01 -1.12955475e+00 -1.65657520e-01 2.54168123e-01
-4.48382169e-01 -3.82181078e-01 -3.26838404e-01 4.60964143e-01
-1.11270614e-01 7.64794588e-01 2.50325173e-01 -1.29179716e-01
2.91646391e-01 5.52104115e-01 3.72707784e-01 -9.91618395e-01
-3.00145686e-01 8.13566595e-02 1.91162378e-01 -3.52506191e-01
-4.36650902e-01 -8.21389198e-01 -1.69205844e+00 7.94773102e-02
-2.72657990e-01 4.03890342e-01 4.19205308e-01 1.28720951e+00
1.59702763e-01 2.77839810e-01 4.47374970e-01 -3.41366440e-01
-9.54966128e-01 -1.05379200e+00 -4.16490346e-01 6.86479032e-01
7.51212761e-02 -6.00704253e-01 -2.88308024e-01 5.58844924e-01] | [10.587059020996094, 8.761520385742188] |
051c339c-619c-4cee-8602-77b93a160822 | towards-learning-representations-of-binary | 2002.03388 | null | https://arxiv.org/abs/2002.03388v2 | https://arxiv.org/pdf/2002.03388v2.pdf | Bin2vec: Learning Representations of Binary Executable Programs for Security Tasks | Tackling binary program analysis problems has traditionally implied manually defining rules and heuristics, a tedious and time-consuming task for human analysts. In order to improve automation and scalability, we propose an alternative direction based on distributed representations of binary programs with applicability to a number of downstream tasks. We introduce Bin2vec, a new approach leveraging Graph Convolutional Networks (GCN) along with computational program graphs in order to learn a high dimensional representation of binary executable programs. We demonstrate the versatility of this approach by using our representations to solve two semantically different binary analysis tasks - functional algorithm classification and vulnerability discovery. We compare the proposed approach to our own strong baseline as well as published results and demonstrate improvement over state-of-the-art methods for both tasks. We evaluated Bin2vec on 49191 binaries for the functional algorithm classification task, and on 30 different CWE-IDs including at least 100 CVE entries each for the vulnerability discovery task. We set a new state-of-the-art result by reducing the classification error by 40% compared to the source-code-based inst2vec approach, while working on binary code. For almost every vulnerability class in our dataset, our prediction accuracy is over 80% (and over 90% in multiple classes). | ['Erik Kline', 'Sima Arasteh', 'Shushan Arakelyan', 'Aram Galstyan', 'Christophe Hauser'] | 2020-02-09 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [-7.19582438e-02 -1.17952816e-01 -4.88915533e-01 -3.79529536e-01
-6.49030507e-01 -9.05099213e-01 3.07369798e-01 6.77370071e-01
-2.16062427e-01 2.24102020e-01 1.14854440e-01 -9.35606956e-01
8.63250867e-02 -1.11490011e+00 -7.26723969e-01 -1.69378296e-01
-2.07688034e-01 4.63285178e-01 3.44410926e-01 -4.11378145e-01
4.19971138e-01 4.25531834e-01 -1.37300146e+00 4.33959424e-01
7.25611806e-01 8.58934522e-01 -1.60805136e-01 8.80527139e-01
-2.82793373e-01 1.09709430e+00 -6.29596233e-01 -9.26127613e-01
1.53582275e-01 7.00536668e-02 -8.76226306e-01 -3.94975752e-01
4.53702450e-01 4.34300192e-02 -6.09302521e-01 1.35622954e+00
1.10155970e-01 -8.50917175e-02 5.33933282e-01 -1.22934675e+00
-1.10657156e+00 9.15176272e-01 -4.96065170e-01 4.20199126e-01
1.76863983e-01 2.91411400e-01 1.46047509e+00 -2.02044487e-01
6.89057708e-01 9.10969317e-01 1.09583604e+00 4.03039932e-01
-1.62107706e+00 -3.28114212e-01 -8.55240077e-02 5.03611803e-01
-1.04350102e+00 -4.83633541e-02 5.24856389e-01 -8.76577079e-01
1.72142875e+00 2.54234463e-01 2.56448299e-01 1.06271064e+00
2.95164764e-01 4.50110853e-01 6.43784404e-01 -2.95692384e-01
1.34827837e-01 1.54923692e-01 1.09889925e+00 9.57031846e-01
4.75964814e-01 2.46439010e-01 2.79342741e-01 -5.79081893e-01
4.83969897e-02 9.30164158e-02 -2.42411867e-01 -4.20371830e-01
-8.39645624e-01 1.40052450e+00 6.17377341e-01 4.15473968e-01
-1.48923099e-01 4.29693252e-01 1.15878546e+00 5.13938189e-01
4.34190005e-01 7.99735606e-01 -5.58729649e-01 -3.84312153e-01
-1.02075911e+00 2.46893242e-01 1.08242035e+00 7.12848365e-01
7.48394966e-01 2.12995857e-01 -2.98248559e-01 6.73144341e-01
1.46636248e-01 1.56162649e-01 5.56218803e-01 4.08123620e-03
6.50545955e-01 9.29792881e-01 -6.10454261e-01 -1.16807544e+00
-3.80381316e-01 -2.93187767e-01 -4.01396334e-01 3.22235286e-01
2.09667563e-01 9.92203653e-02 -9.64642167e-01 1.50375664e+00
-2.02756867e-01 -2.47980878e-01 -4.50353883e-02 4.78758037e-01
8.14030826e-01 6.81405127e-01 9.91633087e-02 3.27046126e-01
1.20238304e+00 -1.08806586e+00 -1.05067253e-01 -6.48754090e-02
9.66746986e-01 -4.25112516e-01 9.88015056e-01 4.75441456e-01
-4.32277590e-01 -4.24503118e-01 -1.24533188e+00 -3.48806530e-02
-8.47732842e-01 -1.67576149e-02 1.05367541e+00 1.16300333e+00
-1.15479279e+00 7.53467202e-01 -7.79966474e-01 -3.42298180e-01
6.09700501e-01 4.41535711e-01 -4.01460737e-01 -3.09590735e-02
-7.93973446e-01 7.36309290e-01 4.00083005e-01 -5.99871457e-01
-1.14367568e+00 -9.43963349e-01 -1.11767864e+00 5.80200315e-01
3.74283612e-01 -2.30486557e-01 9.44903851e-01 -5.58909476e-01
-1.00843668e+00 8.85911286e-01 1.00849569e-01 -8.27775776e-01
-3.34973410e-02 6.79654330e-02 -5.57297051e-01 -2.17572913e-01
2.23742356e-03 -1.71076834e-01 5.10988295e-01 -7.79903531e-01
-1.89388871e-01 -2.44950905e-01 1.75193161e-01 -7.78193831e-01
-7.10607588e-01 3.34767580e-01 -2.81789780e-01 -5.26348054e-01
-5.18667400e-01 -7.71794021e-01 -3.08207244e-01 -5.29785097e-01
-5.84120989e-01 -3.37400645e-01 7.82126665e-01 -1.11453533e+00
1.63746405e+00 -2.06297398e+00 5.30542672e-01 2.26349950e-01
5.64043283e-01 3.68413031e-01 -2.39821389e-01 2.15470418e-01
-4.88063991e-01 5.03385365e-01 -4.43349272e-01 -1.73567370e-01
4.21785474e-01 -1.66681305e-01 -6.00666404e-01 6.64161861e-01
2.56051898e-01 1.08527029e+00 -6.58373356e-01 -1.77719817e-01
3.21155339e-02 3.10941547e-01 -8.86950195e-01 2.37983733e-01
-5.90892375e-01 -3.97874475e-01 -2.61721730e-01 8.72999787e-01
5.51668763e-01 -2.81313211e-01 2.17168331e-01 -6.75726533e-02
-1.13455756e-02 3.54747981e-01 -6.46798670e-01 1.43311393e+00
-7.78291225e-01 8.46925855e-01 -2.28463382e-01 -1.37238240e+00
8.49846542e-01 -3.94003317e-02 2.16843486e-01 -4.82321024e-01
2.26825297e-01 1.91866048e-02 1.40993312e-01 -4.68394488e-01
3.83527756e-01 3.87623787e-01 -6.17555082e-01 5.09622753e-01
4.49548811e-01 -1.52488351e-01 2.86110699e-01 3.25426579e-01
1.61464453e+00 -1.24233335e-01 4.12570596e-01 -5.58577538e-01
6.11203671e-01 3.13285530e-01 3.31283659e-01 6.81090951e-01
-1.69686198e-01 9.53104347e-02 1.21526647e+00 -7.03233719e-01
-9.85481441e-01 -5.46157718e-01 9.73643959e-02 1.26864886e+00
-4.02080536e-01 -9.29568291e-01 -7.13089287e-01 -1.30928171e+00
3.36340249e-01 9.48823869e-01 -9.67735291e-01 -3.44069451e-01
-6.45156562e-01 -9.39656675e-01 9.35249746e-01 7.21917212e-01
8.94142836e-02 -7.66612232e-01 -3.28200608e-01 1.85652599e-01
2.62596905e-01 -7.99137414e-01 -3.37306798e-01 5.55054247e-01
-3.75460178e-01 -1.32850969e+00 -1.66421711e-01 -4.33492094e-01
2.05556408e-01 -1.19215272e-01 1.60542607e+00 2.93001384e-01
-7.47301638e-01 1.17517948e-01 -4.36569512e-01 -9.78204757e-02
-6.98838234e-01 2.61130601e-01 -3.05696636e-01 -3.44114840e-01
4.54908997e-01 -4.69691813e-01 9.43883732e-02 -1.22156873e-01
-7.97177196e-01 -6.30920768e-01 3.01318973e-01 8.74190748e-01
2.74617612e-01 1.54470518e-01 1.83359280e-01 -1.00628257e+00
4.19186801e-01 -8.92616808e-01 -1.15186787e+00 2.83807307e-01
-7.70547688e-01 2.68647671e-01 9.79817212e-01 -3.15842986e-01
-5.55903018e-01 -1.41779855e-01 -2.69344270e-01 -6.64834678e-01
-2.43459679e-02 6.19442344e-01 -2.53876776e-01 -4.74867165e-01
9.94001746e-01 1.10352956e-01 -2.50600070e-01 -4.91223216e-01
6.85026646e-01 4.87718463e-01 3.95684898e-01 -6.34214759e-01
9.55766380e-01 -8.38869661e-02 6.23479746e-02 -4.65628237e-01
-3.54255438e-01 -4.52349991e-01 -4.26451623e-01 3.99628282e-01
1.09802663e+00 -4.95294094e-01 -7.48619378e-01 2.47761458e-01
-1.19981861e+00 -4.70689178e-01 1.29836664e-01 -1.36125356e-01
-3.01716536e-01 6.84633374e-01 -6.37443244e-01 -2.78346509e-01
-4.72336143e-01 -1.62889612e+00 8.66861522e-01 -1.60833374e-01
-8.80656689e-02 -1.07320487e+00 2.81085670e-01 1.48384333e-01
7.06085563e-01 4.74656999e-01 1.63682973e+00 -1.13311052e+00
-4.17144239e-01 -3.38074088e-01 -4.84952986e-01 3.22339922e-01
-1.27287790e-01 1.58866942e-01 -8.16549003e-01 -2.46102765e-01
-2.72546917e-01 -3.72961253e-01 1.28796124e+00 -5.17648295e-04
1.67086661e+00 -1.45762563e-01 -5.36326349e-01 9.82084513e-01
1.94625747e+00 7.21276775e-02 6.91377401e-01 4.61006105e-01
1.08626866e+00 2.30674535e-01 2.58567661e-01 3.15764368e-01
1.46482572e-01 8.27292025e-01 7.98274040e-01 2.35260457e-01
-1.63337976e-01 7.56154954e-02 3.41957331e-01 2.60347128e-01
1.29552022e-01 -4.74243790e-01 -1.46237803e+00 6.88430667e-01
-1.59022462e+00 -9.78151321e-01 -2.90138662e-01 1.99220026e+00
6.48878574e-01 8.04328769e-02 2.80667812e-01 1.60723016e-01
6.12374723e-01 3.05960655e-01 -2.55415589e-01 -9.93331552e-01
2.06272602e-01 8.40012074e-01 7.62566447e-01 2.84321308e-01
-1.51309395e+00 1.10613024e+00 6.24527025e+00 9.20451224e-01
-1.34599745e+00 3.61857653e-01 5.19893646e-01 1.11091837e-01
-2.92233139e-01 5.26494794e-02 -8.08032572e-01 4.66137379e-01
1.53217876e+00 -2.41718680e-01 7.19424963e-01 1.50409925e+00
-6.82807148e-01 4.38348770e-01 -1.22801137e+00 8.16041350e-01
2.44799510e-01 -1.69582069e+00 -3.44096750e-01 1.56887144e-01
5.71907580e-01 3.63743305e-01 8.38208571e-02 9.38124657e-01
6.57950282e-01 -1.29655731e+00 5.81045389e-01 8.50656442e-03
9.99781191e-01 -8.76733005e-01 5.88575900e-01 -1.66788056e-01
-1.42150879e+00 -5.42968273e-01 -3.84705842e-01 2.31324241e-01
-4.06554103e-01 6.68372929e-01 -7.57939160e-01 5.76700449e-01
7.38972962e-01 7.56069660e-01 -1.25335348e+00 7.96468616e-01
-2.15771720e-01 7.41952300e-01 5.78664914e-02 -1.73957527e-01
3.02903295e-01 2.58159518e-01 4.25550193e-01 1.61595941e+00
8.47632661e-02 -4.51219261e-01 1.71120256e-01 1.35584593e+00
-2.98867464e-01 2.94194818e-02 -8.02763700e-01 -5.65473795e-01
1.02131024e-01 1.31358969e+00 -6.79145634e-01 -4.89426136e-01
-6.57156765e-01 8.49976599e-01 8.06782186e-01 -6.98516294e-02
-1.34123862e+00 -9.35816705e-01 9.09036279e-01 -4.08951223e-01
7.78764725e-01 -9.74742845e-02 -3.67909551e-01 -1.36563659e+00
-2.48789713e-01 -1.27710879e+00 5.52104473e-01 -3.26195955e-02
-1.38117445e+00 9.32594121e-01 -2.45066792e-01 -7.21978486e-01
-3.29216003e-01 -8.90618384e-01 -9.34063494e-01 1.01025009e+00
-1.30155528e+00 -1.12670922e+00 -2.35175699e-01 3.96411449e-01
3.44823360e-01 -5.39104462e-01 8.82156134e-01 4.43188846e-01
-7.61499286e-01 1.18189800e+00 1.28075272e-01 2.85955489e-01
4.47882980e-01 -1.47578466e+00 8.83870482e-01 1.13157141e+00
1.75207570e-01 5.49662590e-01 4.32542384e-01 -8.09861958e-01
-1.81619954e+00 -1.56447375e+00 5.69978476e-01 -7.21509993e-01
1.08378971e+00 -5.91776669e-01 -9.50304031e-01 8.75607967e-01
3.55472974e-02 1.74038470e-01 7.98950672e-01 3.54719490e-01
-1.17026329e+00 3.62989426e-01 -1.28010464e+00 1.93745270e-01
9.51687694e-01 -7.46649861e-01 -4.86741692e-01 5.42787194e-01
1.06742454e+00 -1.49344698e-01 -1.02478874e+00 2.71015793e-01
1.43035084e-01 -9.65062559e-01 1.01531553e+00 -1.03725183e+00
7.89455771e-01 4.17837501e-02 -6.05227649e-01 -1.20334637e+00
-6.10429287e-01 -2.33269140e-01 -3.06328833e-01 1.26169395e+00
4.59719270e-01 -6.69622540e-01 6.01055980e-01 3.11046034e-01
-3.68222207e-01 -7.14618862e-01 -6.19468689e-01 -8.77008617e-01
3.26008886e-01 -5.90594769e-01 9.84877646e-01 9.85040486e-01
1.11024678e-01 4.27594688e-03 2.42644157e-02 3.74508113e-01
5.20193279e-01 5.29461861e-01 6.74443364e-01 -1.13009953e+00
-9.55976963e-01 -9.88469839e-01 -1.08819652e+00 -3.09531122e-01
8.45027328e-01 -1.48858011e+00 -2.62759537e-01 -1.05140758e+00
3.44606340e-01 -2.56224483e-01 -2.63541520e-01 7.26499856e-01
-2.16192901e-01 2.21960038e-01 1.27835318e-01 -5.66389747e-02
-2.72566199e-01 1.24462262e-01 1.34665608e-01 -6.95160329e-01
6.46548942e-02 -3.29409510e-01 -8.14476371e-01 3.84653330e-01
5.37925422e-01 -5.05325675e-01 -7.29133114e-02 -4.41856295e-01
6.85488954e-02 -4.70143370e-03 4.90720749e-01 -1.04790723e+00
-1.88786507e-01 1.13142475e-01 1.20421477e-01 -2.76327312e-01
-1.33576244e-01 -4.47824806e-01 -9.93813351e-02 6.54448807e-01
-1.71780184e-01 8.35592300e-02 5.99880040e-01 6.08658195e-01
-7.54952133e-02 -4.84813780e-01 7.12018251e-01 -8.27006549e-02
-1.06430745e+00 3.60368639e-01 -2.99917191e-01 -1.49917543e-01
1.14427948e+00 3.85869801e-01 -7.89093256e-01 3.40274796e-02
-5.52962661e-01 -1.80969298e-01 7.22061634e-01 4.32649553e-01
4.28021461e-01 -9.55973685e-01 -4.12157297e-01 3.74320954e-01
3.74456942e-01 -8.18770528e-01 -1.90486982e-02 5.50012767e-01
-6.44278169e-01 7.28365898e-01 -1.90770775e-01 -2.90608317e-01
-1.41118836e+00 1.26569438e+00 2.65120625e-01 -7.50166416e-01
-4.70722049e-01 1.03522491e+00 -8.84211957e-02 -5.53201497e-01
1.17989890e-01 -4.53759849e-01 -2.88687766e-01 -4.84598055e-02
4.86196816e-01 3.70985091e-01 5.30595064e-01 -4.47358012e-01
-5.21738112e-01 3.80531371e-01 -1.32974118e-01 5.32277763e-01
1.52828550e+00 8.10665369e-01 -6.88156605e-01 -9.75254849e-02
1.64206707e+00 9.30025708e-03 -3.48898768e-01 8.00677165e-02
2.96137482e-01 -5.81931114e-01 2.87440777e-01 -8.28580558e-01
-1.33365834e+00 1.01506996e+00 5.47809243e-01 4.51499224e-01
8.62384975e-01 1.73980534e-01 7.27052271e-01 4.33636129e-01
4.43739414e-01 -2.92585611e-01 -3.81743982e-02 7.20033407e-01
5.32171071e-01 -1.28355038e+00 -1.45298094e-01 -4.10286576e-01
-2.85349697e-01 1.33831549e+00 5.94253421e-01 -2.57659316e-01
5.48057258e-01 3.82333070e-01 -4.35613126e-01 -4.60209131e-01
-7.47166574e-01 -4.46732864e-02 3.21646333e-01 7.24732399e-01
5.60704291e-01 2.91555524e-01 -3.92420925e-02 9.58513021e-01
-1.55067474e-01 -5.08046508e-01 6.03713036e-01 9.12521899e-01
-5.28898202e-02 -1.30735648e+00 -2.07818344e-01 8.45839977e-01
-6.30123079e-01 -4.33716059e-01 -3.72378081e-01 7.42267370e-01
-2.73898467e-02 8.24828744e-01 -1.54842287e-01 -9.29836392e-01
1.78394213e-01 1.85464770e-01 2.48180807e-01 -9.96884108e-01
-9.73693967e-01 -6.53530240e-01 1.83730990e-01 -8.93739104e-01
4.02053058e-01 -5.15766263e-01 -8.45484555e-01 -5.10018587e-01
-2.53199667e-01 -5.12843020e-02 7.14234114e-01 3.70486110e-01
4.41343069e-01 9.81084943e-01 4.47703987e-01 -8.30265999e-01
-9.10032153e-01 -7.54284024e-01 -2.45735362e-01 4.76336628e-01
2.57375330e-01 -6.87735796e-01 -5.65687239e-01 -1.97458968e-01] | [7.185121059417725, 7.777173042297363] |
d117f151-24dd-4b36-87d8-01640c853ec5 | semantic-metadata-extraction-from-dense-video | 2211.02982 | null | https://arxiv.org/abs/2211.02982v1 | https://arxiv.org/pdf/2211.02982v1.pdf | Semantic Metadata Extraction from Dense Video Captioning | Annotation of multimedia data by humans is time-consuming and costly, while reliable automatic generation of semantic metadata is a major challenge. We propose a framework to extract semantic metadata from automatically generated video captions. As metadata, we consider entities, the entities' properties, relations between entities, and the video category. We employ two state-of-the-art dense video captioning models with masked transformer (MT) and parallel decoding (PVDC) to generate captions for videos of the ActivityNet Captions dataset. Our experiments show that it is possible to extract entities, their properties, relations between entities, and the video category from the generated captions. We observe that the quality of the extracted information is mainly influenced by the quality of the event localization in the video as well as the performance of the event caption generation. | ['Deepayan Bhowmik', 'Ansgar Scherp', 'Johannes Scherer'] | 2022-11-05 | null | null | null | null | ['dense-video-captioning'] | ['computer-vision'] | [ 2.03926265e-01 3.64661485e-01 2.79197432e-02 -3.03726673e-01
-1.06023967e+00 -7.85309196e-01 8.55747700e-01 4.22717035e-01
-3.92846256e-01 8.16054404e-01 7.19180524e-01 2.56981939e-01
3.92853796e-01 -5.58156252e-01 -1.24918342e+00 -2.22003162e-01
-7.38517940e-02 5.26572168e-01 6.39067352e-01 1.42965406e-01
-1.42910182e-01 -1.10127382e-01 -1.80855918e+00 7.23310232e-01
6.17837429e-01 1.17781365e+00 1.68232590e-01 7.09396183e-01
-1.31695792e-01 1.03209937e+00 -6.57678246e-01 -5.95884800e-01
-2.43257821e-01 -5.27281761e-01 -7.99045265e-01 1.80547565e-01
4.79101628e-01 -3.13440859e-01 -5.05690932e-01 8.42658460e-01
1.66964158e-01 -3.07464749e-02 5.75217366e-01 -1.58947575e+00
-4.68809903e-01 9.56840158e-01 5.67900799e-02 1.33508861e-01
6.36775136e-01 -2.55152378e-02 1.01663685e+00 -8.60715687e-01
1.11826766e+00 8.77378643e-01 3.52047741e-01 5.93880355e-01
-6.62905335e-01 -2.41019949e-01 1.84325650e-01 6.82814181e-01
-1.43165886e+00 -6.83976829e-01 4.44681168e-01 -6.52694464e-01
6.72468066e-01 2.02424228e-01 5.15865207e-01 1.41295075e+00
-3.72570902e-01 8.79587829e-01 2.92687595e-01 -3.10864210e-01
4.48889315e-01 1.30733564e-01 -7.96166286e-02 7.01558530e-01
2.37890720e-01 -5.96820474e-01 -8.72783303e-01 -1.15640417e-01
4.98972476e-01 -5.04160047e-01 -4.73655969e-01 -3.84539455e-01
-1.63691306e+00 3.88705522e-01 1.86035663e-01 2.50746310e-01
-6.08598232e-01 4.49572206e-01 4.67382073e-01 -4.19811517e-01
4.18278098e-01 3.29851240e-01 -3.07660699e-01 -2.84271359e-01
-1.16630590e+00 2.56738216e-01 7.53292382e-01 1.26083541e+00
6.36462271e-01 -2.48380408e-01 -7.34521031e-01 3.09400827e-01
2.83505887e-01 5.03112316e-01 3.21252346e-01 -8.53261650e-01
6.69846654e-01 4.44904745e-01 5.45556366e-01 -9.89709377e-01
-4.34349105e-03 -8.88989940e-02 -9.70243737e-02 -6.93854034e-01
2.79039115e-01 -1.61079362e-01 -9.48600411e-01 1.92028403e+00
2.36876264e-01 6.54231846e-01 7.20132887e-02 9.49424863e-01
1.12721694e+00 8.01123321e-01 6.97120726e-01 -1.96491241e-01
1.52447069e+00 -9.22078967e-01 -9.14579093e-01 -3.03341895e-01
5.79060137e-01 -3.98433954e-01 5.18201113e-01 -3.47985953e-01
-1.06897271e+00 -5.36524415e-01 -7.71901429e-01 -4.68600914e-02
-1.75872445e-01 4.56101149e-01 2.93423504e-01 4.13091742e-02
-9.54767585e-01 1.31202087e-01 -9.36088920e-01 -4.65721309e-01
2.67718464e-01 -2.04353668e-02 -4.84604508e-01 6.88090548e-02
-1.37640631e+00 7.11819351e-01 6.24778628e-01 -1.33436948e-01
-1.31045103e+00 -5.75334728e-01 -1.31503618e+00 3.55766743e-01
2.85481483e-01 -7.72859931e-01 1.26848543e+00 -1.15412319e+00
-7.81894207e-01 7.44962513e-01 -3.25012296e-01 -6.50798142e-01
3.79696667e-01 -1.19288921e-01 -5.36016822e-01 7.28675485e-01
3.46877515e-01 1.09217727e+00 6.06199503e-01 -1.14946413e+00
-8.88347983e-01 2.00798418e-02 6.47060871e-02 2.46000841e-01
-2.72217572e-01 3.67721356e-02 -9.58991110e-01 -4.16939318e-01
-3.62075508e-01 -7.84191251e-01 1.31548375e-01 -9.77197140e-02
-3.80345941e-01 -1.56344801e-01 6.93463862e-01 -1.11593986e+00
1.27784455e+00 -2.28145123e+00 1.39164045e-01 -5.96874319e-02
-1.02622531e-01 -6.30929843e-02 -3.30557734e-01 1.76137716e-01
1.65011823e-01 2.05574989e-01 -2.84417093e-01 -3.29793781e-01
1.43211246e-01 8.69813934e-02 -4.75288540e-01 2.92793289e-02
4.27972794e-01 9.14561510e-01 -1.15664959e+00 -8.04885387e-01
2.00292263e-02 7.06732690e-01 -4.61902946e-01 5.87259352e-01
-8.65369499e-01 4.02462184e-01 -5.76710820e-01 4.22141075e-01
1.49989799e-01 -5.01679897e-01 3.69749904e-01 -6.85014844e-01
9.51491892e-02 4.11684275e-01 -9.81103599e-01 1.76524794e+00
-2.76474327e-01 8.03162873e-01 -2.47370929e-01 -4.91022289e-01
4.60823685e-01 8.50727618e-01 6.46240234e-01 -3.99110615e-01
4.31808308e-02 3.77651565e-02 -7.49533355e-01 -8.49026382e-01
7.50402331e-01 2.78267086e-01 -2.62041509e-01 1.58704028e-01
4.87780243e-01 5.00827909e-01 7.05426693e-01 4.72433001e-01
1.29800737e+00 5.54482996e-01 2.95211654e-03 2.74309009e-01
3.49978775e-01 2.75574267e-01 4.39339489e-01 4.02959198e-01
1.02718197e-01 8.14207077e-01 6.02748871e-01 -1.28395423e-01
-1.08925521e+00 -1.14226198e+00 3.37000579e-01 9.32575345e-01
2.76001781e-01 -6.89435184e-01 -1.03028047e+00 -9.13281322e-01
-4.36965048e-01 8.12420368e-01 -4.42915410e-01 -6.25361502e-02
-6.13950074e-01 -4.74817097e-01 4.59001541e-01 5.49923122e-01
3.96697462e-01 -1.11996877e+00 -5.18677533e-01 2.28199318e-01
-1.13496149e+00 -2.02250004e+00 -4.15757775e-01 -4.61229235e-01
-3.74710023e-01 -1.21313262e+00 -6.17894471e-01 -7.92477846e-01
8.34272087e-01 -1.90802217e-01 1.42820561e+00 -1.89482309e-02
1.00160442e-01 6.07268631e-01 -5.69332898e-01 -1.06797352e-01
-5.47059894e-01 3.00252438e-02 -2.66723096e-01 3.81407410e-01
1.88397005e-01 -1.99857756e-01 -4.70512331e-01 2.71240562e-01
-1.09396565e+00 4.76228446e-01 2.75781482e-01 1.13007657e-01
6.88572288e-01 -2.98181862e-01 3.23844373e-01 -6.03773952e-01
2.34188139e-02 -7.30892003e-01 -4.61858213e-01 4.18419480e-01
3.70541662e-02 1.68564111e-01 3.43432367e-01 -3.35356086e-01
-1.18909311e+00 4.34302419e-01 1.04516849e-01 -4.85149205e-01
-2.71929115e-01 4.69851941e-01 -4.05613661e-01 4.29454654e-01
6.09479189e-01 1.51741371e-01 -6.29284263e-01 -3.33293676e-01
5.46786606e-01 5.99416196e-01 9.06354189e-01 -5.76993108e-01
6.41921043e-01 5.00237763e-01 -3.05179536e-01 -5.37811995e-01
-1.11216486e+00 -3.53890091e-01 -4.24630880e-01 -4.69511807e-01
1.26301253e+00 -1.41237235e+00 -1.63772509e-01 5.77711277e-02
-1.50831127e+00 -7.48736560e-02 -2.48250917e-01 5.13169408e-01
-7.02577293e-01 2.37903729e-01 -4.95317996e-01 -3.22831959e-01
-3.01618993e-01 -1.03687954e+00 1.38696086e+00 1.19944438e-01
-1.73398018e-01 -7.05965698e-01 -3.50686237e-02 6.88062549e-01
4.75836918e-02 4.84374374e-01 6.58917129e-01 -7.14130282e-01
-1.03792191e+00 -1.58241823e-01 -3.68327200e-01 -3.09007941e-03
-1.52644575e-01 6.02942146e-02 -8.70637596e-01 1.35956287e-01
-6.27275825e-01 -1.18016355e-01 5.54432392e-01 1.88185573e-01
8.38667035e-01 -5.75945795e-01 -5.46154320e-01 3.16766530e-01
1.26524484e+00 -6.50671646e-02 7.85723925e-01 1.63772359e-01
8.22407424e-01 3.59648168e-01 7.07133412e-01 5.09652019e-01
7.35887051e-01 8.19676697e-01 3.86363119e-01 2.85392344e-01
-3.60449523e-01 -6.41474783e-01 6.29129529e-01 6.04035914e-01
8.84474441e-02 -7.17434227e-01 -8.58660936e-01 1.06595445e+00
-2.15599179e+00 -1.18286204e+00 -1.95265338e-01 1.93402374e+00
8.40217769e-01 -6.02993555e-02 -4.65028174e-02 -1.83054358e-01
1.15698278e+00 -7.04786461e-03 4.01351526e-02 2.48101324e-01
-5.66058792e-03 -2.06234843e-01 3.75964850e-01 2.55031586e-01
-1.16701019e+00 9.69039917e-01 5.97709322e+00 4.69855905e-01
-6.90798938e-01 4.22316879e-01 3.97831589e-01 -8.80977437e-02
-3.30483884e-01 3.29017341e-02 -8.88215363e-01 8.58927250e-01
1.30815339e+00 -8.83088857e-02 1.07570872e-01 7.54576683e-01
1.75077036e-01 -1.15132309e-01 -1.39950764e+00 9.82902706e-01
4.39754337e-01 -1.56009257e+00 4.62376654e-01 -4.69619840e-01
6.72392070e-01 -1.07208118e-01 -4.81440485e-01 2.00167149e-01
-1.03629000e-01 -4.56250787e-01 1.29890859e+00 5.70900559e-01
9.06290174e-01 -3.16600353e-01 7.09623933e-01 -4.54246774e-02
-1.32972920e+00 2.36303449e-01 7.37552196e-02 3.98559332e-01
7.04208076e-01 6.85085773e-01 -1.00669885e+00 3.31910938e-01
5.74638247e-01 6.50384605e-01 -7.01154828e-01 1.21445489e+00
-6.19182110e-01 6.27229333e-01 -2.90594697e-01 2.23244309e-01
-3.43087642e-03 3.17602247e-01 5.67278206e-01 1.37453020e+00
5.50067425e-01 -1.19851269e-01 8.13532844e-02 8.01456094e-01
-5.48245311e-01 -7.55492076e-02 -2.75879264e-01 -4.41314161e-01
5.20861328e-01 1.25380528e+00 -7.13950634e-01 -6.26480758e-01
-3.51622611e-01 1.00403488e+00 4.24004570e-02 3.20311636e-01
-1.27031970e+00 -1.96224943e-01 5.24636269e-01 3.09780926e-01
5.52248120e-01 -2.43817791e-02 3.08360875e-01 -1.44379520e+00
3.96357983e-01 -6.43704474e-01 5.50806940e-01 -1.28788197e+00
-7.23380148e-01 7.16436386e-01 7.74099007e-02 -1.16442370e+00
-3.74546468e-01 -1.81580573e-01 -1.88143209e-01 1.46289423e-01
-1.39202571e+00 -1.26763058e+00 -6.66967869e-01 3.96583706e-01
4.89650369e-01 1.82576820e-01 5.64848244e-01 6.55820847e-01
-2.95841396e-01 9.40790549e-02 -5.22788763e-01 5.36446393e-01
6.01668477e-01 -8.52320254e-01 3.75835896e-01 9.52292562e-01
5.33417881e-01 -9.26518589e-02 8.85639191e-01 -8.24208498e-01
-9.94782507e-01 -1.49817836e+00 1.18884099e+00 -6.08698249e-01
6.01617992e-01 -4.63774055e-01 -6.40710652e-01 9.79656398e-01
1.63979739e-01 8.90983418e-02 6.03107572e-01 -3.45189452e-01
-3.53648096e-01 1.29222304e-01 -8.33521545e-01 4.88895416e-01
1.09002709e+00 -6.33337498e-01 -5.56948125e-01 5.58880985e-01
8.95159125e-01 -5.02743304e-01 -7.18685269e-01 3.00110132e-01
1.59292221e-01 -5.21049976e-01 8.28684688e-01 -5.34228802e-01
5.14429092e-01 -5.43344915e-01 -1.15118422e-01 -9.72561598e-01
6.39455914e-02 -3.49110246e-01 -5.06662488e-01 1.55271459e+00
6.76337183e-01 4.73400354e-01 5.99725187e-01 7.75738955e-01
-1.22004732e-01 3.45996916e-02 -8.69732797e-01 -4.25258636e-01
-9.03945088e-01 -3.40487123e-01 5.89664757e-01 8.00777972e-01
5.81486113e-02 5.03666282e-01 -3.69851917e-01 4.75771278e-01
4.78950769e-01 8.30461308e-02 4.70403492e-01 -1.05931890e+00
-1.80271983e-01 3.18022579e-01 -6.00932062e-01 -6.07731700e-01
3.68706554e-01 -6.91571116e-01 3.06564391e-01 -1.96921563e+00
3.35310608e-01 -7.75526017e-02 -1.52593732e-01 6.49254322e-01
-6.96428046e-02 3.52984488e-01 1.71749622e-01 2.00747833e-01
-1.37678206e+00 4.26546246e-01 8.01552296e-01 -1.00441828e-01
7.38332272e-02 -3.56059939e-01 -2.59709984e-01 5.38597405e-01
4.61479723e-01 -5.46659768e-01 -2.98903078e-01 -7.63789594e-01
5.80604911e-01 3.43421221e-01 6.09117031e-01 -1.35712576e+00
6.88917264e-02 -4.87142466e-02 2.17916265e-01 -3.80921692e-01
4.84356731e-01 -7.16284454e-01 4.05946374e-01 -7.69624338e-02
-4.92011338e-01 -6.38433322e-02 6.54144138e-02 5.80531001e-01
-4.30624187e-01 -2.00825781e-01 3.85340542e-01 -1.47250772e-01
-8.99682879e-01 2.63606966e-01 -3.74203950e-01 3.93060088e-01
1.08113587e+00 2.18679294e-01 -5.17568111e-01 -5.26991069e-01
-8.55673850e-01 2.03780353e-01 4.93209034e-01 6.42846882e-01
4.09151495e-01 -1.42548776e+00 -6.49343669e-01 -3.14064085e-01
4.17995811e-01 -1.54124409e-01 1.06550850e-01 5.91489017e-01
-4.16603178e-01 3.03937644e-01 -1.40870973e-01 -4.13949251e-01
-1.15764236e+00 5.43206811e-01 1.33937225e-02 -3.53341550e-02
-2.37687558e-01 4.23898637e-01 2.33340054e-03 2.73576081e-01
2.71287203e-01 -2.03162879e-01 -3.27034533e-01 2.38757387e-01
6.84983015e-01 1.04839072e-01 -9.97966272e-05 -9.13459063e-01
-4.14649695e-01 1.76468760e-01 3.39843035e-01 -2.33316034e-01
1.19679332e+00 -1.66519329e-01 8.65512863e-02 -1.37321167e-02
9.23774242e-01 -2.01840475e-01 -1.22762954e+00 1.47143630e-02
1.61368370e-01 -7.81432763e-02 -9.48058143e-02 -7.25095272e-01
-8.46384585e-01 4.95553702e-01 3.03646863e-01 5.60629815e-02
8.73795569e-01 2.79101044e-01 1.09139061e+00 1.53231978e-01
4.45162922e-01 -1.18183517e+00 2.66238786e-02 2.91174859e-01
6.25432849e-01 -1.02791774e+00 -2.75025427e-01 -6.89834595e-01
-8.37044835e-01 7.21913099e-01 6.06656671e-01 4.02219504e-01
2.46627718e-01 2.45891009e-02 -9.83730853e-02 -3.18507031e-02
-7.79698431e-01 -4.58848685e-01 4.35632885e-01 5.35859406e-01
2.92354107e-01 -1.07089102e-01 -1.17670715e-01 6.55201614e-01
1.46461293e-01 2.41956934e-01 6.18068635e-01 8.58292103e-01
-4.65212405e-01 -7.65839517e-01 -1.63493842e-01 8.68180208e-03
-5.35586059e-01 -1.02492310e-01 -2.09245726e-01 2.71698892e-01
2.82827824e-01 9.02729034e-01 3.73499066e-01 -1.92330226e-01
1.55379683e-01 3.24397177e-01 2.90445805e-01 -7.40911543e-01
-2.84880310e-01 -2.42823720e-01 5.55917621e-01 -6.47769034e-01
-7.19306409e-01 -5.05424798e-01 -1.54197717e+00 4.12699550e-01
7.40150036e-03 5.88428199e-01 8.12472403e-01 1.05624449e+00
8.97717416e-01 3.92059386e-01 1.25235990e-01 -6.01006985e-01
1.88650087e-01 -7.80642211e-01 5.36920577e-02 9.26044762e-01
2.48633567e-02 -5.58256626e-01 -2.03888804e-01 9.43248570e-01] | [10.50901985168457, 0.6712035536766052] |
076ba50e-c52a-4e19-872e-d8ac460f618a | unsupervised-classification-for-polarimetric | 2104.01656 | null | https://arxiv.org/abs/2104.01656v1 | https://arxiv.org/pdf/2104.01656v1.pdf | Unsupervised Classification for Polarimetric SAR Data Using Variational Bayesian Wishart Mixture Model with Inverse Gamma-Gamma Prior | Although various clustering methods have been successfully applied to polarimetric synthetic aperture radar (PolSAR) image clustering tasks, most of the available approaches fail to realize automatic determination of cluster number, nor have they derived an exact distribution for the number of looks. To overcome these limitations and achieve robust unsupervised classification of PolSAR images, this paper proposes the variational Bayesian Wishart mixture model (VBWMM), where variational Bayesian expectation maximization (VBEM) technique is applied to estimate the variational posterior distribution of model parameters iteratively. Besides, covariance matrix similarity and geometric similarity are combined to incorporate spatial information of PolSAR images. Furthermore, we derive a new distribution named inverse gamma-gamma (IGG) prior that originates from the log-likelihood function of proposed model to enable efficient handling of number of looks. As a result, we obtain a closed-form variational lower bound, which can be used to evaluate the convergence of proposed model. We validate the superiority of proposed method in clustering performance on four real-measured datasets and demonstrate significant improvements towards conventional methods. As a by-product, the experiments show that our proposed IGG prior is effective in estimating the number of looks. | ['Changlong Wang', 'Feng Zhou', 'Shijie Ren'] | 2021-04-04 | null | null | null | null | ['image-clustering'] | ['computer-vision'] | [-7.49910176e-02 -3.91057849e-01 3.18318307e-01 -6.11658454e-01
-9.57028985e-01 -4.13472086e-01 6.36717319e-01 -1.89380705e-01
-3.04047078e-01 6.17210031e-01 -4.32658801e-03 -2.31462628e-01
-7.09516048e-01 -4.57289308e-01 -4.68205698e-02 -1.24474382e+00
7.58570246e-03 6.27599895e-01 8.65852535e-02 1.09741427e-01
2.46904433e-01 5.71897328e-01 -1.19531357e+00 -4.33010936e-01
9.91718411e-01 8.05193961e-01 3.91825914e-01 3.16925317e-01
3.57653230e-01 3.42256874e-01 -4.86036360e-01 9.00585204e-02
3.00064683e-01 -4.38240975e-01 -2.88899243e-01 4.54161823e-01
-2.56974876e-01 4.71336357e-02 1.33221179e-01 1.33457065e+00
3.27640265e-01 4.83998328e-01 1.31382704e+00 -9.88262653e-01
-7.05039054e-02 2.90757120e-01 -1.02417254e+00 3.45631063e-01
-5.42647779e-01 -1.63616061e-01 8.58930767e-01 -7.44681060e-01
3.79901975e-01 9.89543736e-01 3.37603956e-01 -3.20809305e-01
-1.04966319e+00 -4.25413519e-01 -1.95934460e-01 1.95924535e-01
-1.76301551e+00 -4.24423933e-01 9.49110627e-01 -5.99480391e-01
3.14515680e-01 8.14084187e-02 3.69619191e-01 2.97518700e-01
1.85583949e-01 3.99197370e-01 1.24416327e+00 -3.67234290e-01
3.42193514e-01 1.79395080e-01 2.97319025e-01 3.71967673e-01
5.52807510e-01 -1.42995626e-01 2.29039416e-01 -3.01495969e-01
5.67068458e-01 -1.68473110e-01 -3.67167830e-01 -6.19365215e-01
-9.44468260e-01 1.01117110e+00 1.51910469e-01 4.14168924e-01
-5.53194880e-01 -1.85348153e-01 -8.94228593e-02 -4.20176774e-01
2.48490319e-01 6.14606850e-02 1.08318560e-01 3.07663023e-01
-1.24842596e+00 2.14401763e-02 3.48417640e-01 6.46626830e-01
7.40682960e-01 3.80647212e-01 -5.10928221e-02 8.38635385e-01
9.48185682e-01 1.03085589e+00 9.47906151e-02 -8.47991049e-01
1.65177867e-01 6.89798146e-02 3.85278672e-01 -1.25503194e+00
-2.61271715e-01 -1.03889525e+00 -9.24086154e-01 2.48689622e-01
1.41983226e-01 -1.77666470e-01 -8.54734957e-01 1.30844831e+00
3.48169744e-01 1.22695640e-01 4.30754036e-01 1.13152897e+00
3.09458047e-01 7.57181227e-01 6.53254464e-02 -5.79460144e-01
1.40459847e+00 -4.32916254e-01 -7.34355032e-01 -1.21210240e-01
9.60348547e-02 -1.00603652e+00 3.15377861e-01 6.29262984e-01
-5.83406866e-01 -4.55880463e-01 -1.27759206e+00 8.04449856e-01
2.36602947e-01 5.54329276e-01 5.14888763e-01 9.02124405e-01
-5.99589884e-01 1.59776628e-01 -1.11729896e+00 -2.52873480e-01
1.60657749e-01 2.14446902e-01 -8.32312256e-02 1.24734722e-01
-6.94019079e-01 6.01833105e-01 5.61997354e-01 5.99417210e-01
-8.94181490e-01 -2.66139418e-01 -5.29721022e-01 1.03670629e-02
-7.62116387e-02 -3.37587744e-01 7.26130188e-01 -6.70839250e-01
-1.15745020e+00 3.33490312e-01 -1.84606642e-01 -4.65269685e-01
3.26587819e-02 2.29689434e-01 -4.72877413e-01 4.46083724e-01
9.81019586e-02 2.06424251e-01 1.11221242e+00 -1.48774767e+00
-4.32976484e-01 -3.75559509e-01 -6.99397266e-01 1.86082706e-01
4.89414893e-02 -2.09026456e-01 -2.90059030e-01 -4.42483187e-01
6.93538129e-01 -7.69914508e-01 -3.65829825e-01 -7.70291030e-01
-2.20339581e-01 5.15873916e-02 7.99994946e-01 -7.23371029e-01
1.00943387e+00 -2.28857636e+00 -6.38518333e-02 7.34710455e-01
-7.16847405e-02 2.14136571e-01 2.70755321e-01 3.18436027e-01
8.54534581e-02 -2.81641036e-01 -6.15203261e-01 -4.04220708e-02
-6.60142582e-03 1.85750425e-01 -2.23276004e-01 8.62291515e-01
-3.37991677e-02 1.59798801e-01 -5.04198909e-01 -5.56655884e-01
4.80056882e-01 5.78432262e-01 -3.34270477e-01 -1.50629342e-01
-6.68515563e-02 7.60882676e-01 -6.78716421e-01 2.66581386e-01
1.43543422e+00 -2.77548879e-01 4.20128584e-01 -6.38869345e-01
-3.92428972e-02 -6.45615876e-01 -1.50382209e+00 1.12201977e+00
-4.74654734e-02 3.09888154e-01 2.74439096e-01 -1.30510926e+00
1.14577186e+00 2.31602877e-01 5.49469411e-01 -1.94228768e-01
3.47727001e-01 5.47800213e-02 1.31311283e-01 -4.67962801e-01
3.78706634e-01 -5.38056791e-01 -3.73997688e-02 1.46736994e-01
-3.53568718e-02 -1.91240191e-01 1.98765799e-01 1.13023460e-01
5.88925183e-01 -6.98186755e-02 4.15500104e-01 -4.46737915e-01
8.93175483e-01 1.56144902e-01 6.34607971e-01 5.42956889e-01
-2.22985998e-01 3.76351118e-01 1.33979067e-01 2.14664117e-02
-9.33769643e-01 -1.35859346e+00 -4.23620760e-01 1.04533516e-01
1.64458886e-01 1.83352992e-01 -5.07614017e-01 -1.90145805e-01
-4.20006245e-01 8.52594733e-01 -9.45615619e-02 4.39032584e-01
-2.54513413e-01 -1.55271924e+00 3.21440458e-01 -1.20581456e-01
7.45015740e-01 -3.30530703e-01 -5.25960982e-01 1.70210347e-01
-2.87612021e-01 -1.22955346e+00 1.29705146e-01 -1.54077426e-01
-9.61266160e-01 -1.01227307e+00 -4.76045340e-01 -5.29550910e-01
7.06821501e-01 2.65284956e-01 5.35280704e-01 -3.10993284e-01
-1.97063968e-01 3.31655681e-01 -3.64357561e-01 1.09039553e-01
-2.09425896e-01 -3.45897943e-01 2.23212987e-02 4.62336570e-01
1.96648762e-01 -6.98802412e-01 -7.02380598e-01 4.30235207e-01
-8.17449391e-01 -3.61369848e-01 8.32350373e-01 6.13446474e-01
6.45117164e-01 7.30424047e-01 4.30596262e-01 -5.84597349e-01
3.57178271e-01 -4.23077941e-01 -1.07575977e+00 2.20013902e-01
-4.83850092e-01 2.55131483e-01 3.62694502e-01 1.39243260e-01
-1.47897100e+00 6.01990372e-02 -1.28662288e-01 -4.22657192e-01
-2.31723487e-01 8.57046008e-01 2.34934948e-02 -5.24023594e-03
2.83973247e-01 3.97035301e-01 -3.71286012e-02 -4.64254916e-01
2.39663437e-01 8.11672330e-01 6.91180825e-01 -4.26323652e-01
9.11762118e-01 8.16148281e-01 2.19946831e-01 -1.19723880e+00
-5.74212015e-01 -7.93938935e-01 -4.29133505e-01 -3.10587704e-01
1.15212619e+00 -1.12601113e+00 -7.05779254e-01 1.65736929e-01
-6.65358722e-01 2.08646253e-01 3.78706157e-01 1.11957848e+00
-4.33433890e-01 7.70789206e-01 -1.80165723e-01 -1.41132891e+00
-1.20125428e-01 -1.12261498e+00 7.05799997e-01 1.20457701e-01
2.73181319e-01 -1.05680072e+00 4.32669446e-02 5.49204886e-01
1.33775860e-01 4.25768465e-01 8.42829466e-01 -3.42086732e-01
-7.06993103e-01 -1.11453816e-01 -3.06219965e-01 4.73607510e-01
-5.00984443e-03 -1.25405401e-01 -6.30568922e-01 -4.21993613e-01
4.34636027e-01 3.34154367e-01 6.96083546e-01 8.41393948e-01
4.55046415e-01 5.51576503e-02 -3.92818213e-01 5.52811027e-01
1.91076553e+00 5.59685409e-01 7.40665317e-01 8.17226805e-03
2.32714340e-01 3.29827398e-01 9.82255697e-01 8.67355347e-01
5.63151576e-02 4.26610410e-01 3.25334817e-01 1.45081848e-01
2.75971234e-01 2.34213650e-01 8.01376998e-02 1.11292756e+00
-1.14607781e-01 -3.57197344e-01 -1.04717660e+00 6.47662699e-01
-1.73091102e+00 -1.01963842e+00 -5.40407538e-01 2.15744925e+00
1.47684570e-02 -8.37380365e-02 -3.18930060e-01 6.86309561e-02
7.05982327e-01 2.12732017e-01 -9.41731036e-02 2.60809660e-01
-6.77596405e-02 2.56760091e-01 7.10077643e-01 8.10060859e-01
-1.21224523e+00 6.41499698e-01 5.44778967e+00 1.16909397e+00
-7.02393293e-01 2.34719470e-01 3.31243247e-01 3.81230086e-01
-3.27357382e-01 3.57254744e-01 -5.53151250e-01 4.20355886e-01
5.60641766e-01 -5.04583642e-02 -2.29744669e-02 5.40010989e-01
5.85014343e-01 -4.98530626e-01 -2.13952363e-01 1.02331746e+00
2.09989306e-02 -8.87131929e-01 3.12414989e-02 3.03983688e-01
7.44051337e-01 -1.55593514e-01 1.05379418e-01 -3.75644267e-02
3.54115784e-01 -6.25457346e-01 2.79801458e-01 8.48358929e-01
2.19556704e-01 -1.05599713e+00 1.14295411e+00 4.37795639e-01
-1.10131621e+00 3.45586799e-02 -3.92703056e-01 1.75607666e-01
4.81480628e-01 9.45266306e-01 -9.87122416e-01 9.70414698e-01
4.20974195e-01 3.36932480e-01 -3.33831698e-01 1.17116725e+00
-1.44861251e-01 7.27222979e-01 -5.23078859e-01 7.24874958e-02
4.47995782e-01 -7.95302093e-01 7.35423863e-01 7.34511256e-01
7.10865259e-01 2.80644804e-01 1.29192844e-01 6.80747747e-01
6.32794678e-01 1.82389900e-01 -1.89052254e-01 -1.44012481e-01
4.73333478e-01 1.46714783e+00 -1.13940787e+00 -6.65108114e-02
3.42262611e-02 4.24184442e-01 -2.83114433e-01 5.85771799e-01
-8.64826918e-01 -1.96132064e-01 2.36867055e-01 -3.93271782e-02
8.01021159e-01 -6.08938575e-01 -1.58632308e-01 -8.85021210e-01
-1.67854011e-01 -6.35992110e-01 3.63604873e-01 -8.64640355e-01
-1.05651665e+00 6.90691769e-01 4.43593353e-01 -1.51617348e+00
-2.46015370e-01 -3.58404100e-01 -5.10668874e-01 7.81697094e-01
-1.23683369e+00 -1.12742519e+00 -3.33990932e-01 4.53144342e-01
2.77137697e-01 -4.25317675e-01 4.08724993e-01 2.14890584e-01
-5.09985089e-01 2.45121755e-02 5.15925109e-01 4.75769378e-02
3.55375379e-01 -6.60942197e-01 -5.42747319e-01 1.24215043e+00
-4.62276675e-02 5.47821462e-01 1.22207594e+00 -5.47151744e-01
-1.06052113e+00 -8.32897961e-01 2.16482535e-01 1.08253732e-01
5.76361835e-01 1.80896953e-01 -4.02974963e-01 6.05077088e-01
3.42895299e-01 -3.04469734e-01 6.28693342e-01 -1.86060101e-01
2.00554982e-01 -2.11651772e-01 -1.08884704e+00 2.31566712e-01
1.47779077e-01 1.01903737e-01 -4.86571282e-01 3.06954563e-01
4.17519957e-02 8.82247910e-02 -1.05849612e+00 7.21896470e-01
9.78900269e-02 -1.14186776e+00 9.43682730e-01 1.06588677e-01
-1.47847205e-01 -1.00509155e+00 -4.99879807e-01 -1.17926812e+00
-3.30335975e-01 -2.69307375e-01 5.61690629e-01 1.17469108e+00
3.09304416e-01 -6.15072072e-01 6.52025402e-01 -1.31731048e-01
-2.17486382e-01 -2.97477782e-01 -9.85771418e-01 -7.27105737e-01
-2.35331178e-01 -3.64773929e-01 2.77178526e-01 8.82648528e-01
-3.10932100e-01 2.35604092e-01 -4.61480707e-01 8.90340626e-01
1.40895677e+00 6.84677809e-02 5.84858716e-01 -1.25902069e+00
-5.62350035e-01 -6.36949241e-02 -5.92539966e-01 -8.56457829e-01
-2.67540980e-02 -7.13733912e-01 2.35817000e-01 -1.47223353e+00
2.64022619e-01 -6.29900157e-01 -1.61756411e-01 -2.21554592e-01
1.32816270e-01 3.00249875e-01 -2.86340006e-02 5.98742068e-01
-4.44869101e-01 6.22319281e-01 8.71568441e-01 -5.37643544e-02
2.95222793e-02 1.78363502e-01 -2.32910156e-01 7.47102797e-01
8.17332685e-01 -4.87649918e-01 -5.65260112e-01 -2.18566328e-01
1.02037318e-01 4.18439835e-01 4.21537817e-01 -1.36973798e+00
3.03343892e-01 1.10222232e-02 2.79214978e-01 -1.13653719e+00
4.82487738e-01 -7.84096301e-01 5.98661244e-01 8.07401538e-02
3.49971116e-01 -2.95730919e-01 -9.79651287e-02 7.91830420e-01
-4.64826763e-01 -5.79453826e-01 1.03136730e+00 1.92421824e-02
-6.37982070e-01 -3.98168154e-02 -6.33311510e-01 -2.12238148e-01
1.09621537e+00 -1.71751022e-01 -1.24938115e-02 -4.29305047e-01
-9.69383121e-01 2.12055668e-01 1.57455191e-01 -2.20409259e-01
5.47933340e-01 -1.19632828e+00 -8.76578689e-01 2.43925259e-01
-7.85167962e-02 -3.51829916e-01 7.15814114e-01 1.31242549e+00
-8.79754543e-01 3.91095787e-01 -1.28413374e-02 -9.19154584e-01
-1.18277502e+00 2.80693173e-01 1.76459149e-01 -2.50361443e-01
-2.97716081e-01 4.88927782e-01 -2.55042687e-02 -2.33228877e-01
-2.66544402e-01 1.04097627e-01 -3.06324989e-01 6.53858259e-02
1.48895741e-01 1.89350560e-01 -4.98048104e-02 -8.15333247e-01
-4.28989798e-01 7.06096649e-01 1.61769927e-01 -6.51652277e-01
1.41388226e+00 -3.35551351e-01 -3.14808935e-01 1.62109315e-01
1.02472949e+00 5.08351475e-02 -1.09374464e+00 -9.67437774e-02
2.42280774e-02 -3.73121530e-01 3.44822586e-01 -3.73801589e-01
-8.99217069e-01 7.30095565e-01 8.65294635e-01 -2.85251886e-02
1.14596450e+00 -1.29109129e-01 2.80499548e-01 3.27395797e-01
3.76425296e-01 -1.05676973e+00 -2.60772616e-01 1.48095772e-01
4.58786219e-01 -1.04133332e+00 3.33988577e-01 -4.76764500e-01
-8.09015214e-01 9.21977997e-01 1.95150208e-02 -1.52761951e-01
1.07909381e+00 8.56417492e-02 1.30630746e-01 -3.88829798e-01
-2.17502981e-01 -2.75269955e-01 1.29248261e-01 5.71615040e-01
1.78629711e-01 1.48930714e-01 -4.89549369e-01 3.97814006e-01
-1.47643179e-01 -3.23556960e-01 4.39054459e-01 6.11799061e-01
-8.73866916e-01 -8.53754461e-01 -8.21336746e-01 2.60959983e-01
-2.72657961e-01 9.11647677e-02 4.39266473e-01 7.54128933e-01
1.07469067e-01 1.25076425e+00 8.53249431e-02 -7.86006674e-02
-1.30939513e-01 -6.16614409e-02 3.47298115e-01 -1.62846848e-01
2.30812922e-01 6.35030389e-01 -1.11030571e-01 2.20834509e-01
-7.94896901e-01 -8.66393209e-01 -9.27323818e-01 1.46455318e-01
-4.70388472e-01 9.91964698e-01 7.58962274e-01 9.86540854e-01
-5.36303781e-02 3.45393211e-01 6.27493560e-01 -5.46539903e-01
-6.75031126e-01 -1.11906767e+00 -1.06420219e+00 1.47250980e-01
-1.33092180e-01 -7.06509471e-01 -6.37485683e-01 5.62313497e-02] | [10.022819519042969, -2.0491387844085693] |
064ebbd2-42b9-4af1-bc47-b6f494d33316 | hector-a-hybrid-text-simplification-tool-for | null | null | https://aclanthology.org/2022.lrec-1.493 | https://aclanthology.org/2022.lrec-1.493.pdf | HECTOR: A Hybrid TExt SimplifiCation TOol for Raw Texts in French | Reducing the complexity of texts by applying an Automatic Text Simplification (ATS) system has been sparking interest inthe area of Natural Language Processing (NLP) for several years and a number of methods and evaluation campaigns haveemerged targeting lexical and syntactic transformations. In recent years, several studies exploit deep learning techniques basedon very large comparable corpora. Yet the lack of large amounts of corpora (original-simplified) for French has been hinderingthe development of an ATS tool for this language. In this paper, we present our system, which is based on a combination ofmethods relying on word embeddings for lexical simplification and rule-based strategies for syntax and discourse adaptations. We present an evaluation of the lexical, syntactic and discourse-level simplifications according to automatic and humanevaluations. We discuss the performances of our system at the lexical, syntactic, and discourse levels | ['Núria Gala', 'Delphine Bernhard', 'Thomas François', 'Eva Rolin', 'Rodrigo Wilkens', 'Amalia Todirascu'] | null | null | null | null | lrec-2022-6 | ['lexical-simplification'] | ['natural-language-processing'] | [ 3.06579582e-02 4.45976645e-01 -1.80485234e-01 -2.87393332e-01
-6.50688469e-01 -4.79732990e-01 1.13708687e+00 8.15926790e-01
-8.96652281e-01 7.50106037e-01 7.78469563e-01 -3.98529589e-01
-8.10760409e-02 -7.85811961e-01 -2.42696375e-01 -2.68429846e-01
3.28047484e-01 7.42895067e-01 1.99214071e-01 -6.44214809e-01
1.83958963e-01 4.80042964e-01 -1.37504351e+00 6.54648960e-01
1.03769171e+00 3.20663899e-01 2.30732281e-02 6.14505887e-01
-5.12835205e-01 6.79430306e-01 -8.79230857e-01 -6.82271600e-01
3.83797586e-02 -4.02738452e-01 -9.97743845e-01 -3.19180518e-01
2.42947578e-01 1.10797130e-01 -6.60895333e-02 9.41345990e-01
5.90498507e-01 4.03008491e-01 4.81179386e-01 -6.22704327e-01
-6.78061664e-01 9.75458205e-01 3.59417312e-02 3.01486641e-01
3.72186154e-01 -1.44589484e-01 1.08608150e+00 -9.94901001e-01
1.02347350e+00 1.32007539e+00 6.54890180e-01 4.96714592e-01
-1.30461168e+00 -2.63157487e-01 1.37169197e-01 5.59143305e-01
-9.93548274e-01 -3.58029306e-01 8.20371389e-01 -3.21032196e-01
1.77512407e+00 3.15979034e-01 6.40884042e-01 9.39581811e-01
3.40685308e-01 7.07822561e-01 8.48211765e-01 -1.14776599e+00
5.05054602e-04 4.05043751e-01 3.24281603e-01 3.43571126e-01
2.81298369e-01 -1.64720684e-01 -2.59731501e-01 2.40272790e-01
2.31713444e-01 -6.01069570e-01 6.77117705e-02 -4.06907462e-02
-8.88089061e-01 1.15572631e+00 -3.00811473e-02 1.09609771e+00
-3.47949952e-01 -3.71357173e-01 1.03438234e+00 5.60341895e-01
8.37747455e-01 8.20526361e-01 -3.75746280e-01 -2.08970830e-01
-9.21610057e-01 7.51813531e-01 1.04142058e+00 6.64731860e-01
4.31060165e-01 1.30724847e-01 -2.37102568e-01 1.04596698e+00
-1.68804228e-01 4.75738168e-01 7.68715203e-01 -6.78309023e-01
6.50035739e-01 7.67726243e-01 -4.42275070e-02 -1.37324166e+00
-6.71804488e-01 1.61139831e-01 -7.30616510e-01 1.22174412e-01
2.60517299e-01 -2.50053853e-01 -5.89020014e-01 1.46853149e+00
2.98372298e-01 -7.40942895e-01 3.64838541e-02 4.00353491e-01
9.84072089e-01 7.26907730e-01 1.52881011e-01 -4.38034505e-01
9.39368129e-01 -9.65094686e-01 -1.22410238e+00 3.31456751e-01
1.14188206e+00 -1.10277033e+00 1.23280799e+00 4.67501283e-01
-1.38308406e+00 -6.29527032e-01 -8.63485634e-01 -5.18923044e-01
-6.84891224e-01 -6.00587353e-02 3.86721253e-01 6.84295416e-01
-9.64185238e-01 7.21385121e-01 -7.31341779e-01 -6.82737768e-01
4.63016331e-02 3.91723096e-01 -5.15055299e-01 2.59219319e-01
-1.35045362e+00 1.43651819e+00 5.78981936e-01 -3.07416439e-01
7.30232969e-02 -5.32949746e-01 -9.58375990e-01 1.80418398e-02
2.73061365e-01 -3.59712631e-01 1.24097037e+00 -1.03500688e+00
-1.94186854e+00 1.00401533e+00 -3.35125290e-02 -7.00221956e-01
4.50898290e-01 -5.08100212e-01 -4.33988184e-01 -2.01340973e-01
-8.59904103e-03 4.32487607e-01 2.99151719e-01 -4.79422510e-01
-7.63336420e-01 -1.05851203e-01 3.73627096e-01 4.18610543e-01
-5.53304434e-01 4.50496882e-01 -9.08209905e-02 -1.06607044e+00
-4.70041543e-01 -7.06386030e-01 -2.15663373e-01 -5.71016431e-01
-4.24524903e-01 -7.90454566e-01 5.01339495e-01 -9.54174042e-01
1.84068465e+00 -1.87833440e+00 7.65588999e-01 -1.04872234e-01
2.65131798e-02 1.00744581e+00 -1.97074845e-01 7.70835221e-01
-1.54352084e-01 4.10310954e-01 -1.47298262e-01 -6.41796470e-01
2.73582816e-01 1.95446119e-01 -1.96336761e-01 2.34798431e-01
1.08249590e-01 8.08229387e-01 -8.32719326e-01 -6.24763429e-01
8.48025978e-01 3.87446225e-01 -8.21975529e-01 2.18249366e-01
-3.78866106e-01 8.54714364e-02 -1.21466987e-01 5.00829741e-02
2.39070296e-01 6.93303347e-01 3.71718287e-01 8.82255845e-03
-4.82365251e-01 7.27645099e-01 -9.45884228e-01 1.56513584e+00
-7.49188721e-01 8.56325686e-01 -2.50853300e-01 -9.24803734e-01
6.65947974e-01 3.94798577e-01 2.86004901e-01 -8.45231771e-01
4.69143808e-01 3.05301640e-02 1.02170393e-01 -5.95147133e-01
7.15726316e-01 -1.76247388e-01 -3.42037082e-01 5.88804424e-01
2.17544377e-01 -3.63029063e-01 8.42486262e-01 -1.64818466e-01
1.03955042e+00 2.60440886e-01 7.35546410e-01 -2.92885303e-01
9.74708676e-01 3.07939738e-01 2.60373980e-01 3.96177441e-01
-2.42092907e-01 2.83976108e-01 4.42126542e-01 -8.35412145e-01
-1.38594985e+00 -4.50873584e-01 -4.28670272e-02 1.33490944e+00
-5.20427167e-01 -7.86225259e-01 -1.05182815e+00 -5.03409147e-01
-2.05246523e-01 1.25389218e+00 -5.41035116e-01 -2.46181358e-02
-1.21015227e+00 -6.29076958e-01 5.94822288e-01 3.14267427e-01
2.98899263e-01 -1.53927612e+00 -5.23885489e-01 4.64939535e-01
-2.01126456e-01 -1.25552714e+00 -7.02758553e-03 -3.03742401e-02
-6.29035830e-01 -1.04763138e+00 -2.54768401e-01 -8.37595344e-01
3.50561030e-02 -1.73153311e-01 1.07295978e+00 -6.37929440e-02
4.91334237e-02 -2.17289869e-02 -6.36577010e-01 -8.59916091e-01
-1.02870393e+00 6.21737361e-01 3.77591699e-03 -3.55516762e-01
7.94514060e-01 -4.06985611e-01 1.50316536e-01 -3.75657201e-01
-9.69895124e-01 1.77441258e-02 2.88312972e-01 5.75003922e-01
5.08303404e-01 -1.02066837e-01 3.40164989e-01 -1.19896030e+00
1.26717997e+00 1.98954027e-02 -4.46143568e-01 2.38780424e-01
-3.40439320e-01 1.44359246e-01 9.03975487e-01 -1.41166434e-01
-1.31429839e+00 -3.75171185e-01 -7.55131125e-01 1.47815853e-01
-4.50712532e-01 5.82223356e-01 -2.85981655e-01 2.81113774e-01
8.80493224e-01 -3.93284529e-01 -4.15709555e-01 -5.18660128e-01
7.76453555e-01 6.93200707e-01 8.90821368e-02 -2.90338337e-01
6.70095086e-01 8.12637955e-02 -2.26615727e-01 -1.28222573e+00
-9.15829182e-01 -1.67107180e-01 -9.40756142e-01 1.51551992e-01
9.85736132e-01 -3.64575535e-01 -2.43837342e-01 1.76844820e-01
-1.63167930e+00 -1.69788420e-01 -7.59934604e-01 4.91237223e-01
-2.69489408e-01 4.28389102e-01 -5.74899197e-01 -3.27789396e-01
-5.49221039e-01 -8.93364906e-01 6.64598703e-01 -1.12421885e-01
-9.77741778e-01 -1.35822403e+00 6.65995002e-01 5.03462516e-02
4.56906348e-01 3.34524304e-01 1.29401553e+00 -1.06826639e+00
-3.14295739e-02 -2.97504276e-01 3.99758220e-02 5.58202505e-01
-6.69656396e-02 1.02821358e-01 -7.89656341e-01 4.26941328e-02
2.10339546e-01 -1.98518351e-01 4.82013941e-01 4.03364420e-01
5.95949411e-01 -2.36765578e-01 -7.58248642e-02 4.21549886e-01
1.08378077e+00 1.51456241e-02 5.15927911e-01 7.75554836e-01
6.92751229e-01 7.49005795e-01 5.11864901e-01 2.31897920e-01
3.11903030e-01 9.07741666e-01 -1.29141003e-01 -8.37383047e-02
-4.66394901e-01 9.36995894e-02 3.05240363e-01 1.24916792e+00
-2.86406159e-01 -4.37557161e-01 -8.91235173e-01 4.55191463e-01
-1.67389619e+00 -7.91372240e-01 -1.61645964e-01 1.92187178e+00
1.03501296e+00 3.39082390e-01 -9.29948762e-02 2.92049617e-01
5.96472681e-01 3.43710542e-01 6.90012984e-03 -1.28954291e+00
-3.17330539e-01 5.98194480e-01 -9.10358056e-02 9.09573436e-01
-1.26016712e+00 1.36333430e+00 6.16186810e+00 8.14649820e-01
-1.05636585e+00 2.59624988e-01 2.40883872e-01 -1.41688390e-02
-3.49048495e-01 -3.55085701e-01 -9.54347968e-01 5.67627773e-02
1.23425615e+00 -4.01474535e-01 3.28286856e-01 6.32909417e-01
2.74404109e-01 1.03880115e-01 -1.06300247e+00 5.93170941e-01
4.35914665e-01 -1.43552434e+00 4.13787633e-01 -3.87260735e-01
6.44393682e-01 4.20449302e-02 -2.16689095e-01 6.66250885e-01
2.79156536e-01 -8.60835969e-01 5.07791877e-01 1.96399242e-01
5.58659017e-01 -9.22332585e-01 1.04558885e+00 2.46117875e-01
-7.62960911e-01 7.70727918e-02 -2.38594100e-01 -3.30409795e-01
2.84746200e-01 4.57533807e-01 -7.34726965e-01 6.37073517e-01
5.28715611e-01 5.59023023e-01 -5.05550206e-01 4.68534052e-01
-6.60329819e-01 4.26700056e-01 -2.32179627e-01 -2.84433931e-01
1.61294326e-01 -1.48331791e-01 8.18177164e-01 1.61958694e+00
-7.54678855e-03 -8.49940181e-02 2.15651408e-01 4.91208613e-01
-1.37979358e-01 8.75458539e-01 -6.01181984e-01 -2.09617838e-01
3.01781774e-01 1.09771645e+00 -6.60102427e-01 -4.46381837e-01
-5.01129091e-01 5.89116931e-01 5.08048475e-01 6.15573414e-02
-7.50856757e-01 -7.40579247e-01 3.24791223e-01 1.08168177e-01
1.56226441e-01 -3.37493330e-01 -6.53041482e-01 -1.23086393e+00
-3.02587822e-02 -1.12834001e+00 2.80437946e-01 -6.25026077e-02
-8.81742537e-01 9.81990278e-01 3.63193631e-01 -5.86711466e-01
-2.84422934e-01 -5.10321498e-01 -5.01303911e-01 7.46456802e-01
-1.50013971e+00 -9.46077883e-01 2.42982835e-01 2.02799752e-01
1.00369322e+00 -4.24583465e-01 1.24312603e+00 4.23735440e-01
-5.83920717e-01 5.06473243e-01 1.03259817e-01 -7.64207095e-02
7.98627138e-01 -1.18627131e+00 5.25329530e-01 8.94599617e-01
8.41737017e-02 4.54549670e-01 9.56588686e-01 -4.98680323e-01
-7.02878654e-01 -9.71324563e-01 2.08070302e+00 -5.32019317e-01
8.36325765e-01 -4.83328134e-01 -8.88475180e-01 6.71418846e-01
8.11838150e-01 -7.49381602e-01 6.75182343e-01 4.24365610e-01
1.41098022e-01 -2.42288094e-02 -9.50024128e-01 8.25002670e-01
7.11661339e-01 -3.86435091e-01 -1.33415127e+00 5.97117424e-01
7.36085832e-01 -4.10084903e-01 -6.71163023e-01 3.65685821e-01
1.63304120e-01 -9.44522381e-01 6.00612938e-01 -5.78017652e-01
4.33209419e-01 1.01315551e-01 9.47848186e-02 -1.48348510e+00
-3.76814812e-01 -6.26405418e-01 5.30766509e-02 1.09958816e+00
1.69505700e-01 -3.57896030e-01 6.18778825e-01 5.79194903e-01
-5.49885333e-01 -5.87643921e-01 -9.04977083e-01 -4.93194193e-01
6.36447489e-01 -4.43853557e-01 5.41636705e-01 7.71143019e-01
2.23080054e-01 7.17752516e-01 4.65799011e-02 -5.27004898e-01
-1.71452060e-01 -2.33561229e-02 9.57504869e-01 -1.26071393e+00
-5.62023632e-02 -8.34798396e-01 -3.22348535e-01 -5.66682935e-01
4.97631788e-01 -1.01503205e+00 -2.48464197e-01 -1.60952139e+00
-2.33118311e-01 2.14701042e-01 -3.62210944e-02 4.34583932e-01
1.14986375e-02 -8.82197171e-02 3.25346142e-01 -2.09802613e-01
-4.99193877e-01 7.01086700e-01 8.66305113e-01 -1.35614984e-02
-4.40256983e-01 -4.13184673e-01 -3.68925899e-01 1.03579664e+00
8.39916348e-01 -4.76166308e-01 -7.76598677e-02 -6.83314443e-01
3.46706510e-01 -5.84188282e-01 -3.63352686e-01 -9.19414639e-01
-2.06185901e-03 7.21082510e-03 8.52565318e-02 -5.68276227e-01
1.94866493e-01 -5.64872503e-01 -1.69738367e-01 4.40912187e-01
-4.49916989e-01 3.97446573e-01 4.88689542e-01 -7.75318295e-02
-4.28358227e-01 -3.37897867e-01 9.06938791e-01 -1.14407003e-01
-4.88394380e-01 -2.95283645e-01 -7.21827149e-01 1.66988432e-01
1.16294205e+00 -8.10656995e-02 -5.72559051e-02 -9.63935554e-02
-8.13907981e-01 -1.50559843e-01 2.42619425e-01 4.42052513e-01
2.28558272e-01 -1.09162271e+00 -8.06100368e-01 1.92074418e-01
-2.04507947e-01 -2.27249891e-01 -6.84541976e-03 6.64038062e-01
-9.43644822e-01 9.84361589e-01 -3.20014060e-01 -7.81124532e-02
-1.49238539e+00 6.30118608e-01 2.27112144e-01 -9.27970469e-01
-6.43149853e-01 6.83261752e-01 -6.16819002e-02 -6.79560721e-01
3.75748128e-01 -7.08810508e-01 -8.34206641e-01 3.40001553e-01
4.98725981e-01 5.84892809e-01 4.44019735e-01 -7.71237969e-01
-7.39294216e-02 2.49953613e-01 -3.51917565e-01 1.73888337e-02
1.57992065e+00 -1.49751320e-01 -2.14071900e-01 3.47738177e-01
9.26164329e-01 1.25334352e-01 -3.16088468e-01 -2.31260315e-01
4.89736795e-01 -1.86576828e-01 -3.25867273e-02 -4.82211858e-01
-4.31730866e-01 1.04403746e+00 2.56376445e-01 4.99331355e-01
8.42504084e-01 -4.37779963e-01 9.65107799e-01 8.74874115e-01
-1.78560168e-02 -1.65518844e+00 -2.32983559e-01 1.22334671e+00
1.11158061e+00 -9.65132713e-01 8.59292522e-02 -3.96785408e-01
-4.34992015e-01 1.29826331e+00 2.78118849e-01 -3.31646055e-01
6.36967838e-01 5.37404194e-02 9.53486338e-02 3.40573713e-02
-6.23145819e-01 -1.29485995e-01 1.96274951e-01 5.23538530e-01
8.50683331e-01 -2.50360183e-03 -9.99545515e-01 6.14606321e-01
-5.44100344e-01 -1.35676473e-01 5.21677375e-01 8.65473866e-01
-5.21529436e-01 -1.81511927e+00 -1.67149678e-01 5.44949919e-02
-5.58407366e-01 -1.96942747e-01 -7.24968255e-01 1.31850457e+00
3.56240243e-01 8.81221414e-01 -1.79874092e-01 -1.22782536e-01
9.28574741e-01 1.13350011e-01 4.96622503e-01 -1.13790929e+00
-1.05864382e+00 3.43009569e-02 5.16547680e-01 -4.04351175e-01
-5.75086296e-01 -7.13765860e-01 -1.11073875e+00 -5.17065585e-01
-1.49264500e-01 2.16754228e-01 5.44211924e-01 1.13533616e+00
1.36092812e-01 6.01666391e-01 1.97197050e-01 -1.05357146e+00
-3.55465084e-01 -1.17902124e+00 -1.10459313e-01 5.45331359e-01
-2.20481381e-01 -3.94648522e-01 -1.75807346e-03 6.67736754e-02] | [10.9152193069458, 10.403144836425781] |
19404aaa-7ce0-40c7-8716-93cddada7ff5 | drs-parsing-as-sequence-labeling | null | null | https://aclanthology.org/2022.starsem-1.19 | https://aclanthology.org/2022.starsem-1.19.pdf | DRS Parsing as Sequence Labeling | We present the first fully trainable semantic parser for English, German, Italian, and Dutch discourse representation structures (DRSs) that is competitive in accuracy with recent sequence-to-sequence models and at the same time {emph{compositional} in the sense that the output maps each token to one of a finite set of meaning {emph{fragments}, and the meaning of the utterance is a function of the meanings of its parts. We argue that this property makes the system more transparent and more useful for human-in-the-loop annotation. We achieve this simply by casting DRS parsing as a sequence labeling task, where tokens are labeled with both fragments (lists of abstracted clauses with relative referent indices indicating unification) and {emph{symbols} like word senses or names. We give a comprehensive error analysis that highlights areas for future work. | ['Kilian Evang', 'Minxing Shen'] | null | null | null | null | sem-naacl-2022-7 | ['drs-parsing'] | ['natural-language-processing'] | [ 6.51398778e-01 7.82113194e-01 -2.35838830e-01 -7.47865617e-01
-8.28059375e-01 -1.13426328e+00 4.98660147e-01 3.73340577e-01
-3.17684978e-01 8.86332273e-01 6.80956542e-01 -6.55001163e-01
1.20528139e-01 -7.00860918e-01 -5.31919777e-01 -3.38625640e-01
1.56137154e-01 7.07792580e-01 4.29324627e-01 -3.88506353e-01
1.13726906e-01 1.96646869e-01 -1.34128726e+00 7.69010305e-01
3.54275852e-01 6.34406269e-01 4.03160453e-01 5.52913427e-01
-6.60553217e-01 1.12076628e+00 -8.62472355e-01 -7.23351538e-01
-1.48679167e-01 -4.00960147e-01 -1.37215114e+00 -1.99227154e-01
2.37650827e-01 1.28408717e-02 8.31387341e-02 1.01818144e+00
-3.56142293e-03 6.45920113e-02 4.22221363e-01 -7.07514703e-01
-7.85429120e-01 8.52408886e-01 -1.89992204e-01 1.83055103e-01
6.98747754e-01 -1.46677583e-01 1.60530293e+00 -5.46220124e-01
1.04571819e+00 1.55146003e+00 4.77563620e-01 1.03256500e+00
-1.06092072e+00 -3.63451876e-02 5.94681501e-01 -1.59987286e-01
-7.80479670e-01 -3.91864926e-01 1.46515861e-01 -3.84573132e-01
1.70226347e+00 5.29293716e-01 2.12001204e-01 8.56839061e-01
-9.89973918e-02 7.48556495e-01 6.32494092e-01 -6.96161807e-01
2.70757675e-01 -1.96819589e-01 5.85388362e-01 6.26390398e-01
2.90636808e-01 -4.17533487e-01 -3.43903720e-01 -1.09154567e-01
5.74105978e-01 -5.63395381e-01 -6.65291771e-02 -4.19655703e-02
-1.11411655e+00 8.35305810e-01 1.28608301e-01 5.73659182e-01
-1.77304044e-01 3.66074979e-01 6.59069598e-01 2.58828670e-01
2.61218697e-01 3.84435058e-01 -8.83287191e-01 -1.47215873e-01
-3.41685891e-01 3.71592075e-01 8.39607358e-01 9.27780688e-01
3.38916302e-01 7.59493336e-02 -1.88254029e-01 7.61212826e-01
1.30147085e-01 2.50752896e-01 3.79259020e-01 -1.38054991e+00
4.80533749e-01 6.84885502e-01 2.95088828e-01 -3.17169815e-01
-4.12358284e-01 1.12482652e-01 -5.87970167e-02 -5.48363551e-02
6.63943887e-01 -1.28276259e-01 -4.63275790e-01 2.27255273e+00
2.26943836e-01 -2.02053204e-01 4.82082069e-01 7.13361502e-01
9.25860643e-01 6.20135069e-01 6.90087616e-01 -2.42107674e-01
1.95700967e+00 -6.37108564e-01 -7.99082696e-01 -5.72315753e-01
9.58432794e-01 -5.44956088e-01 1.09157658e+00 -4.89554554e-02
-1.32126558e+00 -2.35409975e-01 -8.03947151e-01 -4.33309466e-01
-1.77049786e-01 -1.21806152e-01 6.57080829e-01 4.60198462e-01
-1.04704273e+00 3.76727015e-01 -7.46038258e-01 -1.95873946e-01
-4.34416682e-02 1.99813485e-01 -3.82867843e-01 2.08579049e-01
-1.18529403e+00 1.11523652e+00 6.73642874e-01 -3.02913010e-01
-1.43797129e-01 -3.94547433e-01 -1.42912364e+00 2.13809505e-01
2.73260862e-01 -6.28024876e-01 1.74994564e+00 -1.36125600e+00
-1.25019038e+00 1.60284352e+00 -5.30150294e-01 -5.21897197e-01
-3.20436880e-02 -2.14198649e-01 -2.81923294e-01 1.69795960e-01
3.64437521e-01 6.65829718e-01 2.24289045e-01 -1.07941139e+00
-7.84907758e-01 -5.43915510e-01 5.82341015e-01 1.37982413e-01
1.00135557e-01 7.88782358e-01 -1.29131123e-01 -5.53691804e-01
2.81283021e-01 -7.37779379e-01 5.39374491e-03 -2.07301989e-01
-3.01449120e-01 -5.79325438e-01 2.67727077e-01 -8.65477562e-01
1.47402596e+00 -2.11935925e+00 2.75698036e-01 -1.52754486e-01
-1.19542798e-04 3.97614241e-01 -1.29169062e-01 5.63984931e-01
-4.73465502e-01 3.43463689e-01 -5.56916356e-01 -1.73190117e-01
2.45928153e-01 7.83152223e-01 -3.01777750e-01 2.03814223e-01
5.28165281e-01 1.05013752e+00 -1.07740390e+00 -3.11164379e-01
8.41434598e-02 1.41810074e-01 -4.49123323e-01 1.76491827e-01
-5.89615226e-01 2.26870492e-01 -6.26852810e-01 2.39794865e-01
3.43668222e-01 -8.85606855e-02 7.92742968e-01 3.33252907e-01
-1.23582289e-01 1.17203784e+00 -1.16464150e+00 1.49794805e+00
-2.35500798e-01 2.82793015e-01 3.53518695e-01 -9.63645339e-01
7.45710075e-01 5.84540308e-01 -1.64099354e-02 -4.29534793e-01
-1.41501173e-01 3.94372016e-01 9.44772437e-02 -5.29011130e-01
5.58627069e-01 -5.25226653e-01 -5.93773305e-01 4.18641597e-01
-1.90408516e-03 5.42579293e-02 2.44616762e-01 1.86510652e-01
1.21357071e+00 4.71943855e-01 7.41536379e-01 -3.45275819e-01
7.16393530e-01 1.71706468e-01 7.42910326e-01 5.60251415e-01
4.70553944e-03 4.63989913e-01 9.35493529e-01 -5.38267434e-01
-1.08725011e+00 -1.21476257e+00 -7.56828040e-02 1.30349946e+00
-3.96957733e-02 -4.76241261e-01 -9.34041083e-01 -6.88539505e-01
-2.89814621e-01 1.35260439e+00 -4.00598079e-01 2.78743267e-01
-1.32304955e+00 -1.87087968e-01 6.89948440e-01 1.00953841e+00
-1.17847927e-01 -1.66662920e+00 -8.32047820e-01 3.81026477e-01
-3.18717778e-01 -1.33106244e+00 -2.33819768e-01 4.29419547e-01
-5.75557530e-01 -1.24416268e+00 -2.36364920e-02 -1.18689191e+00
4.41053092e-01 -1.36211500e-01 1.37334716e+00 3.87775749e-01
1.05388246e-01 1.00990772e-01 -4.67036158e-01 -3.17602038e-01
-7.59746134e-01 -2.23381534e-01 -5.20231426e-01 -4.66862410e-01
4.52694923e-01 -2.25239918e-01 3.52570899e-02 -3.72649878e-01
-9.31482315e-01 4.12535388e-03 -1.40891373e-01 7.66641974e-01
5.31831801e-01 -5.48396528e-01 3.21716189e-01 -1.49718535e+00
5.88989854e-01 -3.95916373e-01 -3.69602412e-01 4.38595891e-01
2.06553921e-01 3.11257154e-01 5.97837925e-01 8.97890776e-02
-1.33611238e+00 1.42647922e-01 -6.25698566e-01 3.37374985e-01
-3.94304782e-01 5.00027001e-01 -3.54381233e-01 7.31118321e-01
5.02088428e-01 -3.36137228e-02 -1.66981900e-03 -4.74107325e-01
5.49328327e-01 3.74383420e-01 7.62723029e-01 -9.71740961e-01
3.89606804e-01 1.72186852e-01 -2.31247813e-01 -5.53544760e-01
-1.21519375e+00 -4.42610800e-01 -7.07023203e-01 2.99584955e-01
1.22019231e+00 -8.30073357e-01 -6.94187343e-01 -4.74535674e-03
-1.69953680e+00 -2.96471775e-01 -7.13076115e-01 2.29826216e-02
-5.82985222e-01 4.26504552e-01 -1.06035984e+00 -8.65896165e-01
-2.27244735e-01 -8.39956999e-01 1.17566264e+00 -6.00201599e-02
-7.82568514e-01 -1.11644697e+00 -3.49841475e-01 1.90131605e-01
8.55555311e-02 2.42829964e-01 1.51032972e+00 -1.16784072e+00
-7.09467456e-02 -2.18375977e-02 -2.18983516e-01 3.24493974e-01
-7.55066946e-02 -2.84096241e-01 -7.87117124e-01 1.20948792e-01
1.60442010e-01 -3.86010855e-01 6.19128346e-01 2.63070613e-01
8.07770550e-01 -4.61017936e-01 -9.03736651e-02 3.60242501e-02
1.38313782e+00 1.70146346e-01 6.86132431e-01 4.27036166e-01
3.56370032e-01 9.30713296e-01 4.53371793e-01 2.21169554e-02
6.14477813e-01 5.30807436e-01 3.75564307e-01 1.24394655e-01
-2.07317248e-01 -1.30272612e-01 4.17497307e-01 4.87904727e-01
2.40499511e-01 -4.61590976e-01 -7.26038158e-01 7.53353059e-01
-1.92892790e+00 -1.10091519e+00 -3.76724422e-01 2.00377893e+00
1.05558050e+00 2.45039612e-01 -5.64318569e-03 -2.57377960e-02
8.99491847e-01 3.08950454e-01 7.52983391e-02 -9.40539598e-01
-1.47497386e-01 7.81822860e-01 2.62821883e-01 8.56560707e-01
-8.74172211e-01 1.35106301e+00 6.46943140e+00 4.91975278e-01
-5.73714495e-01 2.41313919e-01 2.37221256e-01 4.96218860e-01
-6.26729965e-01 3.08911979e-01 -7.32919872e-01 8.08477923e-02
9.35203314e-01 1.99778959e-01 3.23268682e-01 7.26424634e-01
-9.80359092e-02 -1.27628343e-02 -1.22872567e+00 3.06556493e-01
-2.51407802e-01 -1.43088090e+00 1.06784664e-01 -4.77043629e-01
1.90312594e-01 -1.22663915e-01 -4.71212834e-01 1.76157221e-01
5.79917073e-01 -7.68690526e-01 1.16112638e+00 4.14462350e-02
7.26559043e-01 -3.59811485e-01 7.03596532e-01 3.22509557e-01
-1.40969896e+00 -1.77020449e-02 -3.26702237e-01 -4.78355885e-01
4.30423886e-01 -2.24922955e-01 -5.75302184e-01 6.37902439e-01
1.44197777e-01 2.99101889e-01 -1.16632946e-01 3.80942762e-01
-8.08027685e-01 7.69723713e-01 -5.17045148e-03 -2.21304655e-01
3.11031014e-01 -5.15382783e-03 6.86221242e-01 1.67727780e+00
1.04014473e-02 5.98366797e-01 3.52536738e-01 6.72040343e-01
1.57095894e-01 4.13925163e-02 -3.27638388e-01 4.91732322e-02
7.88874388e-01 8.50611329e-01 -8.97629559e-01 -5.09655297e-01
-4.99192685e-01 7.80494094e-01 4.86190438e-01 6.05715215e-02
-4.55047965e-01 -2.08543375e-01 9.29324806e-01 1.01801097e-01
4.69723821e-01 -1.93201736e-01 -3.48603308e-01 -9.42344010e-01
2.08320618e-01 -6.78247809e-01 7.41341293e-01 -8.61284792e-01
-1.33562279e+00 4.69237208e-01 6.50720745e-02 -5.15496552e-01
-3.83010954e-01 -1.02082765e+00 -4.87945557e-01 8.73560905e-01
-1.55932534e+00 -1.10665059e+00 3.76629591e-01 2.67230064e-01
6.87462866e-01 1.95614681e-01 1.44368160e+00 -1.59729958e-01
-3.08148354e-01 2.79256970e-01 -4.30252433e-01 4.88691002e-01
2.29246899e-01 -1.41321564e+00 8.14252019e-01 7.30604649e-01
1.51559561e-01 8.32830727e-01 7.68397868e-01 -4.74930167e-01
-9.55001891e-01 -7.68916845e-01 1.73924887e+00 -6.50186896e-01
8.58434260e-01 -3.25443983e-01 -1.05395114e+00 1.13838995e+00
3.27004582e-01 -1.70261830e-01 6.19476914e-01 1.96459517e-01
-4.62695956e-01 4.10903305e-01 -1.09601569e+00 3.14010352e-01
1.17340672e+00 -6.26708031e-01 -1.36957741e+00 2.52580732e-01
1.24321270e+00 -7.48011410e-01 -5.07427335e-01 1.46149471e-01
2.64974862e-01 -8.04000914e-01 8.23557258e-01 -1.07224739e+00
3.05550873e-01 -2.96690166e-01 -5.02910852e-01 -7.35632896e-01
-2.85593957e-01 -4.24941391e-01 2.63924867e-01 1.24729538e+00
6.06623828e-01 -5.16685009e-01 4.05198812e-01 8.45846117e-01
-7.12992966e-01 -2.23158956e-01 -1.13841915e+00 -4.84918863e-01
1.63793772e-01 -6.12286270e-01 6.28642917e-01 7.69961417e-01
4.13770288e-01 8.44526052e-01 1.89218536e-01 -1.71283066e-01
5.39917685e-02 1.44143388e-01 4.85263653e-02 -1.30024779e+00
-4.08591688e-01 -3.89818847e-01 -2.18403772e-01 -1.10768366e+00
7.69112468e-01 -1.15112638e+00 1.32531419e-01 -1.77401888e+00
5.32895066e-02 -3.86878788e-01 -2.86959335e-02 9.56498981e-01
-1.80693388e-01 -8.79875273e-02 2.43760943e-01 -2.74050105e-02
-7.02815175e-01 3.93435434e-02 9.10227358e-01 1.67124122e-01
9.98950675e-02 -1.67835206e-01 -7.99465179e-01 9.35581028e-01
6.83650732e-01 -5.81525266e-01 -1.21637352e-01 -7.53851295e-01
3.67211342e-01 4.44250792e-01 1.10806018e-01 -3.81196737e-01
-1.16728388e-01 -3.16521078e-01 -9.49500501e-02 -9.28549096e-02
2.79593170e-01 -4.40061539e-01 -1.35899559e-01 4.46224898e-01
-7.24115193e-01 4.48224425e-01 2.34696358e-01 1.50920123e-01
-9.40822735e-02 -6.94945335e-01 5.81482351e-01 -5.87943494e-01
-9.82709050e-01 -4.29375619e-01 -6.33556962e-01 4.63166684e-01
7.47023642e-01 -3.17422062e-01 -3.56861979e-01 -1.07693978e-01
-8.42469752e-01 1.03591211e-01 3.15832287e-01 3.59017611e-01
3.58863950e-01 -8.67881835e-01 -7.02717364e-01 -3.80188674e-02
-7.74763292e-03 3.58663499e-02 -1.27366289e-01 1.80892467e-01
-4.68438834e-01 5.44864416e-01 -5.28447740e-02 -2.00951368e-01
-1.41884410e+00 4.63370621e-01 1.11891091e-01 -3.86111051e-01
-5.59676945e-01 1.05499530e+00 3.71906072e-01 -6.83967650e-01
1.54115766e-01 -3.85871321e-01 -3.71680200e-01 -2.63596386e-01
6.01309657e-01 -7.98002332e-02 2.20917799e-02 -1.06616938e+00
-5.92933059e-01 1.89910933e-01 7.39070252e-02 -1.87201649e-02
1.39817727e+00 -1.49780035e-01 -6.36839032e-01 3.90621901e-01
8.40163410e-01 1.92130491e-01 -8.80408764e-01 -2.01010555e-01
6.54978871e-01 1.18812591e-01 -7.35847890e-01 -8.36361289e-01
-4.07408297e-01 5.06202459e-01 -1.86453119e-01 2.77437508e-01
7.83563972e-01 5.07914186e-01 8.86232138e-01 2.83844441e-01
2.41630271e-01 -9.65584099e-01 -1.19959503e-01 9.50504839e-01
6.02310181e-01 -5.39999664e-01 -4.48346227e-01 -7.71216929e-01
-8.66299927e-01 1.17701030e+00 4.23815072e-01 -2.15789095e-01
1.30554602e-01 5.71845770e-01 6.74813092e-02 -3.62779170e-01
-9.17044580e-01 -3.52410764e-01 -7.46427551e-02 6.20380461e-01
9.56830323e-01 3.19396555e-01 -9.77922201e-01 9.11997199e-01
-4.21422392e-01 -3.30012739e-01 3.52074683e-01 1.09129977e+00
-8.07451904e-01 -1.62236106e+00 -9.86579955e-02 4.22936566e-02
-7.60266781e-01 -3.46080571e-01 -5.59768379e-01 9.82396841e-01
1.71234503e-01 1.00941253e+00 3.48025292e-01 1.38771206e-01
4.32234496e-01 5.99398136e-01 6.13610327e-01 -1.30259383e+00
-9.37749267e-01 -9.08060446e-02 9.87241626e-01 -3.90459448e-01
-5.30067563e-01 -8.99247825e-01 -2.27577090e+00 2.50285100e-02
-1.66952610e-01 2.13217571e-01 2.85613835e-01 1.32804883e+00
-4.24349830e-02 4.03675497e-01 -1.48380119e-02 -2.99468130e-01
-6.72590017e-01 -7.01753139e-01 -3.40003997e-01 7.02141047e-01
1.10814825e-01 -2.44200468e-01 8.01768303e-02 2.00996727e-01] | [10.392426490783691, 9.251577377319336] |
6957cefd-bc10-4bf1-ba46-92858d9f190c | tan-without-a-burn-scaling-laws-of-dp-sgd | 2210.03403 | null | https://arxiv.org/abs/2210.03403v2 | https://arxiv.org/pdf/2210.03403v2.pdf | TAN Without a Burn: Scaling Laws of DP-SGD | Differentially Private methods for training Deep Neural Networks (DNNs) have progressed recently, in particular with the use of massive batches and aggregated data augmentations for a large number of training steps. These techniques require much more computing resources than their non-private counterparts, shifting the traditional privacy-accuracy trade-off to a privacy-accuracy-compute trade-off and making hyper-parameter search virtually impossible for realistic scenarios. In this work, we decouple privacy analysis and experimental behavior of noisy training to explore the trade-off with minimal computational requirements. We first use the tools of R\'enyi Differential Privacy (RDP) to highlight that the privacy budget, when not overcharged, only depends on the total amount of noise (TAN) injected throughout training. We then derive scaling laws for training models with DP-SGD to optimize hyper-parameters with more than a $100\times$ reduction in computational budget. We apply the proposed method on CIFAR-10 and ImageNet and, in particular, strongly improve the state-of-the-art on ImageNet with a +9 points gain in top-1 accuracy for a privacy budget epsilon=8. | ['Alexandre Sablayrolles', 'Pierre Stock', 'Tom Sander'] | 2022-10-07 | null | null | null | null | ['image-classification-with-dp'] | ['computer-vision'] | [ 1.28372863e-01 2.06822619e-01 2.54661918e-01 -7.82666981e-01
-9.20395076e-01 -6.71886444e-01 1.78162903e-01 -2.95989923e-02
-1.12878311e+00 8.77518654e-01 -2.75740862e-01 -5.59462547e-01
-4.50797789e-02 -5.75209856e-01 -8.64881992e-01 -1.15181029e+00
-1.62397802e-01 4.95088398e-02 -1.48882851e-01 1.69470787e-01
-1.68374151e-01 6.74065888e-01 -1.20301962e+00 -1.98923294e-02
5.99855661e-01 1.34460807e+00 -3.06539804e-01 4.60447937e-01
3.58901262e-01 3.83150309e-01 -6.39032423e-01 -9.72392857e-01
9.16004717e-01 -2.10213244e-01 -4.65788633e-01 -3.97900671e-01
5.34194410e-01 -6.82300746e-01 -6.25211835e-01 1.50596559e+00
7.95379400e-01 6.33493215e-02 3.36357415e-01 -1.25482559e+00
-3.57585371e-01 6.44131899e-01 -5.02537966e-01 1.14374541e-01
-6.91015482e-01 1.69389933e-01 8.32407415e-01 -3.61433268e-01
3.34771246e-01 7.77580082e-01 6.11949861e-01 8.46597672e-01
-1.26329446e+00 -1.19847190e+00 1.21139832e-01 -9.05332193e-02
-1.48380637e+00 -5.18603027e-01 4.28309292e-01 -2.35948950e-01
8.94598365e-01 4.64879453e-01 3.91282946e-01 1.19934058e+00
-6.76406249e-02 6.24168754e-01 1.06127870e+00 -9.97529104e-02
5.54531574e-01 5.19531012e-01 6.63796961e-02 3.83923322e-01
5.62488616e-01 1.42651215e-01 -2.58212000e-01 -3.60901862e-01
5.50145268e-01 8.01106468e-02 -4.27213252e-01 -3.18171859e-01
-5.01989543e-01 8.84938121e-01 2.75341600e-01 -2.55836230e-02
2.92110648e-02 5.82133770e-01 5.72189271e-01 6.30142808e-01
7.22713113e-01 3.39095950e-01 -7.39770830e-01 9.83489864e-03
-8.28767776e-01 4.48437244e-01 8.45156133e-01 1.08164144e+00
5.88957429e-01 -4.44092490e-02 -1.00715302e-01 6.64988756e-01
4.67154533e-02 1.94497108e-01 3.79888088e-01 -9.16523874e-01
8.05309236e-01 5.45289554e-03 2.62637343e-02 -7.47170568e-01
-1.26184642e-01 -7.96112955e-01 -1.06729662e+00 3.55573058e-01
6.11551404e-01 -7.29798257e-01 -6.50138199e-01 2.07055068e+00
3.16694945e-01 5.75681627e-02 1.67384848e-01 6.90493941e-01
7.36482367e-02 2.86210656e-01 6.49823919e-02 -1.63594812e-01
1.30910408e+00 -5.18828094e-01 -4.78917807e-01 -7.33333528e-02
1.05634332e+00 -2.51753807e-01 8.44211340e-01 3.66712719e-01
-9.28900957e-01 1.98579952e-01 -1.16576278e+00 -2.43950784e-01
-2.65516132e-01 4.70661074e-02 6.54869437e-01 1.24811506e+00
-1.10302413e+00 1.01196718e+00 -9.75226343e-01 4.58707921e-02
1.16347647e+00 8.33564639e-01 -5.54685354e-01 7.80532556e-03
-1.11401689e+00 4.41768408e-01 3.12253088e-01 9.33132172e-02
-7.87607908e-01 -1.10804725e+00 -3.64442468e-01 3.16364437e-01
2.09244668e-01 -5.92960238e-01 1.11711633e+00 -7.01162159e-01
-1.65779889e+00 8.87352824e-01 4.24275994e-01 -1.07453716e+00
1.16429973e+00 -3.03236693e-01 -1.17671430e-01 -1.19323522e-01
-5.11024594e-01 5.92648268e-01 6.17830515e-01 -9.36163366e-01
-6.43426716e-01 -7.84996629e-01 4.40864712e-02 8.55178833e-02
-7.33738303e-01 5.05760238e-02 -1.97657421e-01 -6.32241726e-01
-4.12110016e-02 -1.00102079e+00 -4.49232340e-01 3.37511808e-01
-2.90845364e-01 3.17508519e-01 1.01279283e+00 -4.06960338e-01
8.62083256e-01 -2.40113330e+00 -3.94613624e-01 2.62243241e-01
4.37158316e-01 3.95950854e-01 1.24495797e-01 -8.00647866e-03
-2.00005639e-02 3.74715060e-01 -3.40301543e-01 -9.23995197e-01
2.81803966e-01 3.01087260e-01 -4.30631161e-01 1.03257287e+00
-1.63535982e-01 5.97272992e-01 -4.95915234e-01 -6.21906668e-02
-2.26397574e-01 5.85621476e-01 -1.00387609e+00 -3.79148498e-02
-1.22928359e-01 2.55527675e-01 -4.65935618e-01 -5.65010123e-03
1.13528574e+00 -9.78974551e-02 1.76177204e-01 1.44445240e-01
2.99583338e-02 3.19153839e-03 -9.61185813e-01 1.41992903e+00
-3.11782956e-01 5.79219341e-01 3.53656501e-01 -7.90022671e-01
6.84803069e-01 3.18614215e-01 2.25238532e-01 -3.63733947e-01
5.55368364e-01 2.81879187e-01 -1.13750003e-01 -1.50996551e-01
1.15014173e-01 -7.90640116e-02 1.93045527e-01 3.92363787e-01
1.09987728e-01 3.12528729e-01 -3.72764319e-01 -1.38426244e-01
1.18372095e+00 -5.38515925e-01 -9.27380025e-02 -4.34288412e-01
1.14118949e-01 -5.87498724e-01 6.45913720e-01 7.38760054e-01
-2.93868005e-01 5.21529377e-01 8.53191555e-01 -2.15909988e-01
-1.10384178e+00 -4.50901806e-01 -3.84054214e-01 9.06268239e-01
-3.08311611e-01 -6.08405024e-02 -1.10157192e+00 -8.21011186e-01
1.35116354e-01 7.18186378e-01 -6.51258528e-01 -1.90332159e-01
-3.62097323e-01 -1.05009794e+00 9.35107887e-01 3.39490384e-01
8.22490752e-01 -4.11340594e-01 -7.18048334e-01 -2.72648215e-01
4.48272318e-01 -1.17198634e+00 -4.98011261e-01 6.18886769e-01
-8.64365399e-01 -4.77579623e-01 -9.54241812e-01 -4.70414430e-01
8.93413961e-01 -8.82253945e-02 5.51150203e-01 -2.19323188e-01
-1.04041576e-01 1.55515680e-02 1.78110823e-02 -5.96778750e-01
-1.04666136e-01 2.15650871e-01 1.36897996e-01 2.27950469e-01
2.26739526e-01 -9.39631581e-01 -8.56080353e-01 1.78941414e-01
-9.52044964e-01 -2.38408998e-01 3.93194139e-01 8.97981644e-01
6.74166322e-01 6.23503923e-02 2.83634126e-01 -1.16775393e+00
4.28162962e-01 -4.77976680e-01 -1.14034665e+00 5.89976124e-02
-7.52760887e-01 1.44261733e-01 7.91246116e-01 -4.85173702e-01
-9.02680695e-01 1.57181367e-01 -3.69570494e-01 -8.31678808e-01
7.38532916e-02 -1.98266041e-02 -4.30472285e-01 -5.32527626e-01
6.91662908e-01 -7.05339462e-02 -9.83763300e-03 -6.57872021e-01
2.56748766e-01 6.41754448e-01 4.00790960e-01 -5.16391218e-01
5.46205461e-01 5.16561151e-01 2.32023820e-01 -5.46235681e-01
-5.97072840e-01 1.28508464e-03 -1.34619609e-01 4.82165486e-01
6.63963795e-01 -1.05295336e+00 -1.11298561e+00 5.66906691e-01
-8.44467759e-01 -2.18250439e-01 -6.95202589e-01 5.75120628e-01
-3.90233248e-01 2.13523015e-01 -5.59036314e-01 -9.29517925e-01
-4.87232089e-01 -1.08264494e+00 5.59403300e-01 -6.48556789e-03
3.74589980e-01 -7.69354343e-01 -2.77336031e-01 2.08111435e-01
4.76108819e-01 3.80679607e-01 7.25546598e-01 -1.09338355e+00
-6.62306011e-01 -4.44424868e-01 -2.44498029e-01 7.75286794e-01
-4.75979537e-01 -5.58422506e-01 -1.50495648e+00 -5.64032912e-01
6.70029700e-01 -2.65757740e-01 7.92422116e-01 2.88878769e-01
1.80880785e+00 -9.91616726e-01 -1.84480641e-02 1.41972756e+00
1.61255443e+00 1.05010286e-01 6.35321021e-01 9.05426443e-02
6.78741813e-01 4.17623341e-01 8.12497362e-02 7.30399430e-01
-1.12776123e-01 4.11182284e-01 6.44104123e-01 3.91456448e-02
3.97777230e-01 -2.84756422e-01 1.10444678e-02 4.01241660e-01
2.44990081e-01 -3.91001374e-01 -6.64586842e-01 3.34613383e-01
-1.65681839e+00 -5.19228041e-01 6.02283999e-02 2.50207853e+00
9.95853186e-01 9.41832289e-02 -1.69801623e-01 -4.20550928e-02
4.07351643e-01 1.19957417e-01 -7.32397795e-01 -3.84880573e-01
-1.66105002e-01 2.19339862e-01 1.50833118e+00 3.44200999e-01
-1.06071866e+00 7.44638085e-01 5.78993988e+00 1.07139027e+00
-1.07798433e+00 3.84428918e-01 1.30318499e+00 -5.92458665e-01
-3.04754794e-01 -1.33411601e-01 -9.28795695e-01 4.41056520e-01
1.19336784e+00 -2.43278205e-01 4.27523136e-01 1.23613036e+00
-2.42035627e-01 2.46434808e-01 -1.31095004e+00 1.29415572e+00
-4.86101389e-01 -1.24809253e+00 -2.02800930e-01 4.66330171e-01
7.73442924e-01 3.69793475e-01 4.27487373e-01 1.55900093e-02
4.70212460e-01 -1.04964364e+00 6.02314711e-01 -4.46507940e-03
9.88771141e-01 -9.78557587e-01 8.04464698e-01 3.87436628e-01
-4.22615141e-01 -2.59199411e-01 -6.47724092e-01 1.88264146e-01
-2.30876714e-01 7.97937095e-01 -5.35870790e-01 1.88350320e-01
9.83280241e-01 4.71328832e-02 -1.68007284e-01 8.32999110e-01
-1.84138063e-02 5.78632772e-01 -7.40480244e-01 -1.35861218e-01
1.56937525e-01 -2.74210751e-01 4.12866741e-01 1.05652046e+00
4.07892704e-01 3.46345782e-01 -6.39100671e-01 8.33756745e-01
-6.79717004e-01 1.27442360e-01 -4.71099406e-01 8.63067247e-03
5.03504038e-01 1.02240276e+00 -3.21088135e-01 -4.96282056e-03
-7.62157589e-02 9.15305793e-01 3.79366755e-01 4.07704592e-01
-7.29922771e-01 -4.70447749e-01 1.21375775e+00 4.59851976e-03
5.13888299e-01 -1.80316672e-01 -5.13918161e-01 -1.06105995e+00
2.06163168e-01 -7.46847630e-01 2.51830310e-01 5.10564819e-02
-1.30090332e+00 5.62794566e-01 -8.53746012e-02 -9.86142218e-01
-6.00198358e-02 -5.71165383e-01 -2.25178450e-01 1.03896022e+00
-1.27764511e+00 -7.47896731e-01 1.37076721e-01 5.65645754e-01
-1.60204694e-01 -1.53959572e-01 8.57074499e-01 5.94206810e-01
-7.26577342e-01 1.57620227e+00 7.74793923e-01 2.15758830e-01
2.20385730e-01 -8.69785070e-01 5.95951796e-01 8.81751478e-01
-6.93718642e-02 6.34331286e-01 7.15408385e-01 -1.69519395e-01
-1.28958166e+00 -1.09080040e+00 6.91678047e-01 -2.06168741e-01
5.59545696e-01 -9.38083291e-01 -8.60644341e-01 7.26358831e-01
-1.61174476e-01 5.02790272e-01 7.79630005e-01 -1.28834471e-01
-5.22966027e-01 -5.59870303e-01 -1.77483594e+00 7.57416129e-01
1.19715023e+00 -4.17285621e-01 2.44094938e-01 4.62997586e-01
1.06263673e+00 -5.47846138e-01 -8.60972881e-01 2.30399191e-01
3.99558216e-01 -9.43610132e-01 6.41490579e-01 -6.12279356e-01
-6.41383976e-02 1.38596162e-01 -3.90487045e-01 -9.62074399e-01
2.29433417e-01 -1.21222389e+00 5.97836673e-02 1.15753841e+00
4.44602817e-01 -1.11955953e+00 1.28164637e+00 1.27983415e+00
1.90858752e-01 -1.00183427e+00 -1.35579526e+00 -7.33826995e-01
2.96958506e-01 -5.56018114e-01 7.77578413e-01 9.26280677e-01
-4.30278629e-01 -4.06897306e-01 -4.41869557e-01 2.57805526e-01
8.36853981e-01 -6.14575148e-01 5.73237479e-01 -8.10028493e-01
-6.44204736e-01 -2.00813189e-01 -5.48863649e-01 -1.02451146e+00
1.04815960e-01 -7.58685350e-01 -2.18649119e-01 -4.49115574e-01
-2.17374951e-01 -6.35299623e-01 -4.05804247e-01 6.47435248e-01
3.41443807e-01 3.64295915e-02 1.60921663e-01 -1.14291105e-02
-1.88880831e-01 6.49281800e-01 8.46322179e-01 1.58669323e-01
-9.66817141e-02 2.01125592e-01 -8.95907581e-01 4.78578806e-01
7.97432721e-01 -8.17078054e-01 -6.71550930e-01 -7.46009350e-01
1.80193335e-01 -1.91069081e-01 4.25283521e-01 -1.17252624e+00
2.76046485e-01 1.72324911e-01 2.70946473e-01 -6.82348460e-02
3.62368524e-01 -1.23839462e+00 2.87050754e-01 3.18290412e-01
-6.09141409e-01 -1.44518733e-01 2.90676177e-01 7.28180170e-01
9.10819620e-02 -2.36659333e-01 1.00633597e+00 3.33134085e-02
9.90844741e-02 6.95453227e-01 1.51978433e-01 1.49285093e-01
1.07288992e+00 -6.30996674e-02 -4.00632709e-01 -2.07954720e-01
-4.84398097e-01 8.64905491e-02 3.60866755e-01 -7.73573369e-02
2.07304850e-01 -9.82250750e-01 -4.31610614e-01 3.75382245e-01
-2.27716297e-01 2.06239402e-01 5.58330357e-01 4.97089416e-01
-6.71100795e-01 2.81313658e-01 4.42080107e-03 -2.60230392e-01
-1.20580268e+00 3.70367587e-01 4.68329817e-01 -2.75071889e-01
-7.53907144e-01 1.46304905e+00 3.46842557e-01 -4.66102846e-02
6.66666985e-01 -3.94473851e-01 4.04805601e-01 -2.30851173e-01
4.98625189e-01 2.58006573e-01 3.57542515e-01 -1.66348606e-01
-1.52936995e-01 1.03992820e-02 -3.83492172e-01 -2.14186594e-01
1.49908793e+00 1.14486381e-01 7.57605284e-02 -4.41699289e-02
1.89660895e+00 -2.10428998e-01 -1.75755966e+00 -1.75527856e-01
-3.04157287e-01 -6.69875085e-01 3.07274789e-01 -4.76752698e-01
-1.55660331e+00 8.90609801e-01 1.03780782e+00 8.28440674e-03
1.17056799e+00 -2.79866427e-01 8.06384683e-01 4.59823728e-01
5.59406161e-01 -1.05028868e+00 -6.90110564e-01 2.84759700e-01
5.76321542e-01 -9.58995700e-01 1.46948565e-02 -2.17123315e-01
-5.43986797e-01 6.03471935e-01 3.11937362e-01 -1.37588784e-01
1.11119950e+00 5.58089435e-01 -5.85620739e-02 1.77748978e-01
-7.12586284e-01 5.28602242e-01 -1.15830176e-01 3.22957635e-01
-1.64981216e-01 3.11598927e-03 -1.74847424e-01 8.78907204e-01
-3.69899422e-01 -6.01882231e-04 5.81757240e-02 9.31205273e-01
-2.60696765e-02 -1.08676064e+00 7.01437443e-02 5.33708334e-01
-1.03840232e+00 -1.44566104e-01 -1.07136346e-01 7.46674895e-01
8.55425745e-02 4.58443284e-01 1.66392997e-01 -5.28324306e-01
2.52934277e-01 4.20095660e-02 1.84896499e-01 -1.81227580e-01
-8.46829057e-01 -2.79990733e-01 -9.32242572e-02 -6.15375221e-01
7.63036013e-02 -6.30068302e-01 -1.00366294e+00 -7.31519163e-01
-2.85104215e-01 -1.51278637e-02 9.77151394e-01 6.56282306e-01
6.90069675e-01 5.57438396e-02 6.71849489e-01 -4.59516346e-01
-1.29589963e+00 -5.07234275e-01 -9.66419578e-01 2.06620142e-01
3.74623179e-01 -9.13422406e-02 -8.36022139e-01 -5.51182568e-01] | [5.9068121910095215, 6.8532490730285645] |
5e3fa616-acbd-41c5-b084-89623519380f | boosted-efficientnet-detection-of-lymph-node | 2010.05027 | null | https://arxiv.org/abs/2010.05027v1 | https://arxiv.org/pdf/2010.05027v1.pdf | Boosted EfficientNet: Detection of Lymph Node Metastases in Breast Cancer Using Convolutional Neural Network | In recent years, advances in the development of whole-slide images have laid a foundation for the utilization of digital images in pathology. With the assistance of computer images analysis that automatically identifies tissue or cell types, they have greatly improved the histopathologic interpretation and diagnosis accuracy. In this paper, the Convolutional Neutral Network (CNN) has been adapted to predict and classify lymph node metastasis in breast cancer. Unlike traditional image cropping methods that are only suitable for large resolution images, we propose a novel data augmentation method named Random Center Cropping (RCC) to facilitate small resolution images. RCC enriches the datasets while retaining the image resolution and the center area of images. In addition, we reduce the downsampling scale of the network to further facilitate small resolution images better. Moreover, Attention and Feature Fusion (FF) mechanisms are employed to improve the semantic information of images. Experiments demonstrate that our methods boost performances of basic CNN architectures. And the best-performed method achieves an accuracy of 97.96% and an AUC of 99.68% on RPCam datasets, respectively. | ['Hefeng Zhou', 'Zhaogang Yang', 'Haotian Xie', 'Qianying Liu', 'Jun Wang'] | 2020-10-10 | null | null | null | null | ['image-cropping'] | ['computer-vision'] | [ 3.48002732e-01 1.75000384e-01 -2.43210301e-01 -2.79813915e-01
-6.61622405e-01 -8.03287923e-02 3.85726422e-01 1.97579145e-01
-6.85805500e-01 5.31332195e-01 1.82058208e-03 -3.16346616e-01
6.89193085e-02 -9.63183880e-01 -2.84059227e-01 -1.03713858e+00
3.54982108e-01 -1.88873291e-01 1.73910975e-01 -1.88425049e-01
7.02770874e-02 8.70756924e-01 -1.19920218e+00 4.95562851e-01
9.16803181e-01 1.13823020e+00 3.85658979e-01 5.39367676e-01
-4.57970440e-01 6.90007269e-01 -3.96326572e-01 -3.12352836e-01
1.27198929e-02 -7.73044750e-02 -5.94558597e-01 -1.20087288e-01
1.10758945e-01 -2.31835559e-01 -4.20291692e-01 1.24384487e+00
5.05362749e-01 -4.08433497e-01 4.20570910e-01 -9.03938353e-01
-8.16756785e-01 6.32143795e-01 -8.81370902e-01 3.89788121e-01
-2.41389513e-01 -5.29397279e-02 4.71459508e-01 -8.36757481e-01
7.08647609e-01 9.12573099e-01 6.95299923e-01 6.58280611e-01
-8.12503040e-01 -8.10182452e-01 -2.77194560e-01 2.04271421e-01
-1.46595323e+00 -2.25620881e-01 8.08255792e-01 -3.02264448e-02
6.37737393e-01 2.21233904e-01 5.98348379e-01 7.25038528e-01
3.77687514e-01 6.40170634e-01 1.01660621e+00 -5.15833139e-01
-2.71312207e-01 2.51347542e-01 -1.56326905e-01 7.59865999e-01
5.45059144e-01 -1.55347094e-01 -8.94488320e-02 2.84039527e-01
1.16642809e+00 5.86447597e-01 -3.51221591e-01 -1.42072916e-01
-1.40308344e+00 6.44572735e-01 1.00189483e+00 7.84649849e-01
-2.38584861e-01 -1.52913705e-01 4.69184875e-01 -2.04628274e-01
2.85700321e-01 4.10834134e-01 8.46886821e-03 4.16153640e-01
-7.71918893e-01 -2.50140041e-01 1.72181591e-01 6.59056783e-01
5.55127680e-01 -7.43367225e-02 -3.21922451e-01 8.14367652e-01
-2.84028947e-01 4.64422822e-01 8.97655308e-01 -5.76740682e-01
1.40240476e-01 1.27977598e+00 -1.42634585e-01 -9.97356236e-01
-9.54224646e-01 -7.32251942e-01 -1.53379953e+00 2.42148563e-02
4.17246103e-01 1.97380990e-01 -1.12167645e+00 1.33013976e+00
2.05871865e-01 -3.34786475e-02 7.62219653e-02 8.26101243e-01
9.04068112e-01 2.62315333e-01 2.44257554e-01 -5.50937839e-02
1.62239504e+00 -9.13161635e-01 -1.02238321e+00 2.34647289e-01
7.65083373e-01 -5.91686904e-01 9.92939293e-01 8.99531841e-02
-9.21143413e-01 -6.83680177e-01 -1.17940128e+00 -2.27080151e-01
-6.96137309e-01 5.36647737e-01 7.68506229e-01 4.15931821e-01
-1.04017830e+00 3.14157575e-01 -8.82485867e-01 -3.35876405e-01
9.88117158e-01 4.74470526e-01 -6.95591271e-01 -1.61508292e-01
-9.65826452e-01 6.23984098e-01 4.10343945e-01 3.06462079e-01
-4.30141598e-01 -8.40499461e-01 -6.09496057e-01 1.79791570e-01
-1.77977253e-02 -4.62225616e-01 9.09229577e-01 -1.12411451e+00
-1.17959702e+00 9.42384243e-01 1.25790566e-01 -4.18493658e-01
4.20548350e-01 2.61349946e-01 -5.23608625e-01 5.54256141e-01
2.33292617e-02 1.20121574e+00 4.54325944e-01 -8.84973943e-01
-8.70253205e-01 -3.26298088e-01 -2.28885546e-01 1.09255582e-01
-8.08831751e-01 -2.87035733e-01 -6.17514133e-01 -5.55824697e-01
1.09272286e-01 -6.13693953e-01 -4.34393555e-01 2.53286988e-01
-3.02339941e-01 9.25312098e-03 9.60734427e-01 -5.91979682e-01
1.06253493e+00 -2.27916765e+00 -2.39650086e-01 1.44465163e-01
2.45110452e-01 4.57582355e-01 -7.74773061e-02 -2.23119900e-01
-1.48537353e-01 3.32558244e-01 -3.55913229e-02 -6.68966994e-02
-5.03371716e-01 7.88897127e-02 1.19290203e-01 4.54997271e-01
4.37519252e-01 1.24969602e+00 -5.69111705e-01 -1.01683891e+00
3.74258965e-01 7.75769830e-01 -3.92924428e-01 8.87580495e-03
2.12777361e-01 3.24785918e-01 -4.06249017e-01 8.10380042e-01
1.00646806e+00 -6.57873631e-01 1.29893929e-01 -5.58522284e-01
-6.10866807e-02 -4.27692384e-01 -6.90156996e-01 1.76421928e+00
-3.41138959e-01 5.90278268e-01 1.74077794e-01 -6.80871725e-01
8.47133338e-01 1.94610253e-01 4.46393102e-01 -8.96940708e-01
3.80447209e-01 9.45370197e-02 1.81969017e-01 -7.11604893e-01
5.07446945e-01 1.28767090e-02 2.10325569e-01 -5.85435070e-02
-2.93100089e-01 2.35638514e-01 1.37808040e-01 1.91520795e-01
8.85147691e-01 -3.53088140e-01 4.73628879e-01 -1.80597782e-01
8.39926243e-01 1.78051353e-01 4.58846927e-01 4.13443923e-01
-2.87430853e-01 6.96328163e-01 4.44507986e-01 -7.06770122e-01
-1.08756077e+00 -6.45208478e-01 -4.69324559e-01 5.54394782e-01
2.56224751e-01 2.27404699e-01 -7.83916712e-01 -7.09005356e-01
-1.58548981e-01 -5.66756688e-02 -1.07218134e+00 -9.48684290e-02
-7.48732507e-01 -9.97203529e-01 5.80293238e-01 8.20992649e-01
9.07797098e-01 -1.12473893e+00 -4.30775523e-01 5.68797924e-02
-3.26917142e-01 -1.09108305e+00 -2.03950882e-01 7.15003610e-02
-9.46024120e-01 -1.20965600e+00 -9.12265956e-01 -1.07809198e+00
1.22420704e+00 5.07698298e-01 6.22892857e-01 4.52705443e-01
-7.74842620e-01 -3.69462311e-01 -4.01731461e-01 -4.56674874e-01
-2.30321229e-01 3.75080228e-01 -4.03376430e-01 2.10830159e-02
4.24035341e-01 -6.00747988e-02 -9.13684428e-01 2.62420289e-02
-1.23900068e+00 4.09776419e-01 1.30736268e+00 1.10988092e+00
7.15703189e-01 8.75499770e-02 5.67653775e-01 -1.07963908e+00
2.50039786e-01 -1.36578858e-01 -2.83141941e-01 1.50089160e-01
-5.06203175e-01 -1.19037338e-01 8.45265031e-01 -4.60792035e-01
-1.17185676e+00 2.61116643e-02 -2.21752346e-01 -2.34866127e-01
-2.14662611e-01 2.78649777e-01 4.80018035e-02 -3.11300069e-01
5.41097581e-01 1.83265120e-01 3.97644103e-01 -1.77346483e-01
1.51492367e-02 7.33211756e-01 7.62278914e-01 2.02239156e-01
6.39082849e-01 1.01346016e+00 1.74788639e-01 -5.42072058e-01
-7.29138076e-01 -4.45751339e-01 -6.33347392e-01 -1.36989176e-01
8.75996470e-01 -9.42054272e-01 -7.37204194e-01 4.14000601e-01
-6.90419912e-01 9.22494233e-02 -1.46568194e-01 3.29730570e-01
-2.50229873e-02 1.97460279e-01 -8.53692234e-01 -2.19291031e-01
-6.30176425e-01 -1.02488220e+00 1.02849901e+00 8.09667230e-01
1.87771738e-01 -7.86116719e-01 -4.11441386e-01 2.17275143e-01
8.04821193e-01 3.36344540e-01 9.61757362e-01 -4.55068439e-01
-5.06225646e-01 -5.13655901e-01 -8.22279334e-01 3.37276161e-02
2.73857474e-01 9.77994725e-02 -1.05725336e+00 -3.12384576e-01
-3.95625740e-01 -8.40918906e-03 1.03068960e+00 6.00202322e-01
1.74263072e+00 7.74502754e-04 -8.73041749e-01 9.39686298e-01
1.56439865e+00 1.93140388e-01 7.89336383e-01 6.04524434e-01
5.81537485e-01 4.37745720e-01 5.97951472e-01 1.46014288e-01
-6.15944373e-05 3.00605118e-01 5.59060097e-01 -9.53691304e-01
-4.23511297e-01 -1.06201008e-01 -3.39866936e-01 4.27367091e-01
-3.90067324e-02 2.53342129e-02 -8.26705456e-01 6.09550118e-01
-1.40109336e+00 -6.42410338e-01 7.15105608e-02 1.56826663e+00
7.55227685e-01 1.10142967e-02 -3.63793761e-01 1.73097596e-01
8.74774992e-01 -1.04035422e-01 -5.99731684e-01 1.61462590e-01
-2.93921530e-01 1.82810768e-01 7.02822387e-01 1.83465078e-01
-1.17773199e+00 6.43883467e-01 6.06579590e+00 1.07428634e+00
-1.39874983e+00 -4.51890901e-02 1.12472498e+00 3.57318334e-02
2.10267361e-02 -6.66883051e-01 -8.20387065e-01 2.74094194e-01
5.52630723e-01 1.09655306e-01 -2.53175676e-01 8.76527846e-01
1.07891046e-01 2.84026051e-03 -7.02324748e-01 8.81459594e-01
6.94811866e-02 -1.69855988e+00 3.66128802e-01 2.37211287e-01
7.72558749e-01 -3.95258933e-01 2.76257277e-01 1.27146319e-01
-9.62680951e-02 -1.21025693e+00 1.21368980e-03 5.47574759e-01
1.26815534e+00 -9.62187231e-01 1.48266983e+00 2.13945750e-02
-1.10989332e+00 -2.75509834e-01 -5.99610507e-01 3.76572579e-01
-3.77146155e-01 4.51918989e-01 -1.09814608e+00 5.07921398e-01
7.49236345e-01 5.73551297e-01 -9.80084121e-01 8.43653798e-01
1.43820331e-01 1.71488151e-01 -9.49385669e-03 -6.16103671e-02
2.20618933e-01 3.20448756e-01 -1.11770794e-01 1.36259449e+00
3.99193019e-01 7.86255598e-02 -1.76337078e-01 5.63314021e-01
-2.53230602e-01 2.09392115e-01 -2.51013815e-01 1.27499983e-01
5.35926640e-01 1.88934445e+00 -1.08059907e+00 -3.25783491e-01
-3.35237414e-01 5.86952507e-01 3.00230473e-01 1.89636499e-01
-6.76129162e-01 -5.86755395e-01 2.47691751e-01 1.73868418e-01
2.11502716e-01 3.16399872e-01 -3.56938362e-01 -9.00398910e-01
-2.69980401e-01 -7.04524696e-01 3.59032720e-01 -5.86646080e-01
-1.00179613e+00 7.13414550e-01 -5.22161007e-01 -1.30725288e+00
2.54220188e-01 -7.54684508e-01 -5.64420521e-01 6.33102715e-01
-1.95170498e+00 -1.33455968e+00 -8.74548495e-01 5.33182502e-01
3.97114903e-01 -2.77195685e-02 8.15475881e-01 2.48679489e-01
-7.18787432e-01 6.65865481e-01 1.91208869e-01 6.14707351e-01
7.71900058e-01 -1.11827183e+00 -1.59090891e-01 6.87733889e-01
-5.27907848e-01 4.91696745e-01 7.56451413e-02 -3.27505589e-01
-1.16365290e+00 -1.26010704e+00 5.23709595e-01 9.74266902e-02
3.23636979e-01 -1.38612807e-01 -8.27306390e-01 3.09531808e-01
1.39042720e-01 4.06014264e-01 8.35317254e-01 -4.41202402e-01
-1.63350701e-01 -4.49373275e-01 -1.30020738e+00 6.47611678e-01
4.47192818e-01 -2.37384722e-01 5.88044301e-02 8.17857161e-02
6.83112860e-01 -4.28281337e-01 -9.84879613e-01 6.50264442e-01
6.04268134e-01 -8.11064243e-01 1.09288394e+00 -2.22058952e-01
5.37573040e-01 -3.72814208e-01 1.71439216e-01 -8.42721462e-01
-6.16494656e-01 2.09283918e-01 3.29170734e-01 9.28960979e-01
3.80779266e-01 -4.52460527e-01 1.00284922e+00 1.90423116e-01
-2.18201041e-01 -9.33711946e-01 -8.79206240e-01 -1.25684723e-01
5.38044050e-02 1.73236996e-01 7.11251080e-01 8.78587544e-01
2.31072288e-02 -2.09604070e-01 6.65037334e-02 5.94254918e-02
4.80238229e-01 1.47994503e-01 4.35819745e-01 -1.04234242e+00
2.96569884e-01 -7.20689535e-01 -5.35031319e-01 -4.85977024e-01
-2.08109245e-01 -9.60628331e-01 -3.81826818e-01 -1.38006103e+00
6.15594089e-01 -4.45894539e-01 -7.42093265e-01 5.26444077e-01
-4.17418689e-01 8.59610021e-01 -1.35754198e-01 3.59097183e-01
-6.02031529e-01 1.84106082e-01 1.78024900e+00 -2.48421818e-01
4.91882414e-02 -4.13142562e-01 -9.97421980e-01 6.67956769e-01
9.41821218e-01 -1.29097149e-01 4.25313553e-03 -1.38550282e-01
-8.67856517e-02 2.97235958e-02 2.48968497e-01 -1.16767740e+00
4.24832910e-01 8.31495896e-02 1.18182468e+00 -7.92612672e-01
9.04105678e-02 -8.53200018e-01 -1.98096391e-02 8.28315437e-01
-5.13536751e-01 -2.12984324e-01 2.08492428e-01 4.28539246e-01
-4.61278975e-01 1.16532341e-01 1.10886776e+00 -1.86427966e-01
-4.61270422e-01 3.80986214e-01 -1.89367399e-01 -7.03455269e-01
1.21326923e+00 -4.10394400e-01 -7.04374671e-01 1.58244878e-01
-4.14999157e-01 1.73238397e-01 5.22818923e-01 2.76403636e-01
7.20556378e-01 -1.35453260e+00 -5.27574122e-01 4.50961739e-01
3.01154584e-01 2.29297310e-01 7.03772068e-01 1.04236472e+00
-9.35445189e-01 7.59276748e-01 -5.76049149e-01 -6.19950175e-01
-1.33973420e+00 5.20785093e-01 5.11752546e-01 -5.88218093e-01
-6.40058875e-01 7.91442037e-01 4.58581716e-01 -1.29167929e-01
1.65901393e-01 -4.69165504e-01 -8.12718332e-01 -1.47214428e-01
9.24417317e-01 7.74957910e-02 1.51776373e-01 -3.86353225e-01
-1.95285410e-01 4.92896527e-01 -7.20656931e-01 5.16102552e-01
1.30031157e+00 -7.96653032e-02 -1.74901247e-01 -6.14025556e-02
1.07871842e+00 -1.16634950e-01 -1.13194084e+00 -3.11444432e-01
-3.39857012e-01 -3.65509927e-01 2.18927145e-01 -9.22850370e-01
-1.35688627e+00 8.67195606e-01 8.98668528e-01 6.75078928e-02
1.44247663e+00 -1.85923114e-01 6.82734489e-01 1.18125357e-01
-2.98374593e-02 -8.77205670e-01 8.24276265e-03 -5.28841540e-02
4.92019862e-01 -1.43552065e+00 2.44907606e-02 -5.66148341e-01
-4.07944053e-01 1.46116543e+00 8.32290232e-01 -3.45890895e-02
3.03792566e-01 6.36233449e-01 3.12682748e-01 -5.32823205e-02
-6.26342118e-01 -7.22942054e-02 1.85180344e-02 5.95629394e-01
5.08620322e-01 1.04724154e-01 -2.20366016e-01 6.00691080e-01
3.82523448e-03 1.98890701e-01 5.76133609e-01 7.62183547e-01
-5.71465731e-01 -7.11931288e-01 -4.20648724e-01 5.82337797e-01
-8.25957239e-01 -3.21795456e-02 -1.74913347e-01 1.19698393e+00
1.65510610e-01 6.03465140e-01 3.44443500e-01 -2.43369237e-01
5.64044602e-02 -2.53441006e-01 2.23195836e-01 -2.76505947e-01
-4.67682064e-01 1.72289953e-01 -3.93800288e-01 -3.11399072e-01
-4.51757699e-01 -2.00820893e-01 -1.40220988e+00 -2.94793069e-01
-5.22155821e-01 1.68225184e-01 8.25475633e-01 5.54702342e-01
2.70782411e-01 1.10127473e+00 4.99765545e-01 -5.69262028e-01
-1.32441625e-01 -1.22981274e+00 -6.15093052e-01 3.12812269e-01
2.94273794e-01 -3.94016385e-01 -9.11248475e-02 1.36318162e-01] | [15.032027244567871, -2.8839986324310303] |
22a3cc74-1ffe-48c1-88c8-a1509e21f69b | segmented-learning-for-class-of-service | 2208.01793 | null | https://arxiv.org/abs/2208.01793v1 | https://arxiv.org/pdf/2208.01793v1.pdf | Segmented Learning for Class-of-Service Network Traffic Classification | Class-of-service (CoS) network traffic classification (NTC) classifies a group of similar traffic applications. The CoS classification is advantageous in resource scheduling for Internet service providers and avoids the necessity of remodelling. Our goal is to find a robust, lightweight, and fast-converging CoS classifier that uses fewer data in modelling and does not require specialized tools in feature extraction. The commonality of statistical features among the network flow segments motivates us to propose novel segmented learning that includes essential vector representation and a simple-segment method of classification. We represent the segmented traffic in the vector form using the EVR. Then, the segmented traffic is modelled for classification using random forest. Our solution's success relies on finding the optimal segment size and a minimum number of segments required in modelling. The solution is validated on multiple datasets for various CoS services, including virtual reality (VR). Significant findings of the research work are i) Synchronous services that require acknowledgment and request to continue communication are classified with 99% accuracy, ii) Initial 1,000 packets in any session are good enough to model a CoS traffic for promising results, and we therefore can quickly deploy a CoS classifier, and iii) Test results remain consistent even when trained on one dataset and tested on a different dataset. In summary, our solution is the first to propose segmentation learning NTC that uses fewer features to classify most CoS traffic with an accuracy of 99%. The implementation of our solution is available on GitHub. | ['Xiao-Ping Zhang', 'Ramy Atawia', 'Akram Bin Sediq', 'Hatem Abou-zeid', 'Sihao Zhao', 'Yoga Suhas Kuruba Manjunath'] | 2022-08-03 | null | null | null | null | ['traffic-classification'] | ['miscellaneous'] | [ 1.67237237e-01 -2.88370430e-01 -7.19782770e-01 -5.36950231e-01
-3.39518249e-01 -4.63478416e-01 2.00474948e-01 -1.08673111e-01
-8.88891146e-02 7.54645407e-01 -6.04177415e-01 -1.03353143e+00
-4.45309043e-01 -6.95040286e-01 -1.88015655e-01 -3.38251412e-01
-3.17776352e-01 6.89415872e-01 6.89890265e-01 -1.80047333e-01
3.82747024e-01 1.13089943e+00 -1.61793137e+00 3.91238421e-01
7.93898046e-01 1.39737904e+00 2.84359366e-01 7.70628273e-01
-6.57041550e-01 8.72596264e-01 -9.06811774e-01 -3.06572109e-01
3.99937034e-01 -3.13778311e-01 -9.49926794e-01 2.00889841e-01
1.27810493e-01 4.54340763e-02 -2.42410004e-01 3.21266294e-01
2.00613812e-01 7.91060701e-02 5.95016956e-01 -1.87634397e+00
3.62462372e-01 4.90508258e-01 -2.33333930e-01 5.67538679e-01
3.70481461e-01 -1.22153915e-01 9.58703876e-01 -2.45111674e-01
5.88507295e-01 9.77890909e-01 6.77098870e-01 3.15831214e-01
-1.31286716e+00 -8.43863726e-01 8.27584863e-02 5.99980712e-01
-1.53823209e+00 -4.09593672e-01 6.53113723e-01 -3.79230350e-01
1.09394908e+00 9.23382938e-01 8.48310471e-01 9.64338124e-01
-7.74135161e-03 6.43131196e-01 6.75896227e-01 -3.99797559e-01
4.10876483e-01 7.01437116e-01 2.58865088e-01 4.64766473e-01
1.11705258e-01 -1.21210612e-01 1.58544168e-01 -1.86304286e-01
4.48434889e-01 6.68709427e-02 -6.03361242e-02 -2.31395602e-01
-6.43797576e-01 8.59713137e-01 3.70993204e-02 3.80660027e-01
-4.56258416e-01 -1.11779079e-01 6.81390464e-01 5.05844831e-01
3.81223112e-01 1.14338569e-01 -6.62350953e-01 -6.90073609e-01
-1.39575732e+00 -1.22747667e-01 1.12938368e+00 1.04918230e+00
5.44533432e-01 4.45630282e-01 2.77187020e-01 9.89463389e-01
8.45649373e-03 5.74136078e-01 3.91041726e-01 -1.05186486e+00
3.74399155e-01 5.28586447e-01 -4.57078665e-01 -1.12279487e+00
-4.20698017e-01 -5.82715273e-01 -4.45406497e-01 7.51200095e-02
2.59432286e-01 2.35359810e-04 -5.77254713e-01 1.30968499e+00
-3.29627804e-02 5.48509061e-01 -1.08106099e-01 5.07670045e-01
5.59939265e-01 7.55830348e-01 5.27877398e-02 -4.95583564e-01
1.02697599e+00 -6.61958814e-01 -3.89915645e-01 3.08558702e-01
8.62237513e-01 -8.52819145e-01 5.71268141e-01 3.65988880e-01
-8.03759038e-01 -6.78132653e-01 -7.53478467e-01 7.71472931e-01
-3.13596725e-01 -2.27042139e-01 9.01346862e-01 1.21743238e+00
-8.21326077e-01 4.31713700e-01 -5.23304760e-01 -6.09000444e-01
4.79565471e-01 5.79706371e-01 -2.41999313e-01 4.34463471e-02
-7.28458285e-01 6.33248687e-01 2.87706167e-01 -2.41978645e-01
-4.87397671e-01 -5.97217441e-01 -5.32100558e-01 2.06369579e-01
3.90145689e-01 -2.19909996e-01 9.13182080e-01 -1.24498630e+00
-1.32341576e+00 2.63710022e-01 -3.50584567e-01 -3.54730994e-01
5.31559050e-01 4.85344321e-01 -1.01279855e+00 3.68884534e-01
8.37812498e-02 2.54257500e-01 7.07300842e-01 -1.33508766e+00
-9.31650221e-01 1.05565749e-01 -2.08659545e-01 -4.65471566e-01
-3.04775745e-01 2.50630826e-01 -4.82456267e-01 -3.58706176e-01
2.60087550e-01 -8.33297253e-01 -2.06296697e-01 -4.34280038e-01
-2.61596322e-01 -3.37739438e-01 1.15973449e+00 -4.92050111e-01
1.42635226e+00 -1.94612122e+00 -3.94056290e-01 9.32212710e-01
1.77994341e-01 4.83074188e-01 -1.30893931e-01 4.30808425e-01
-2.80242562e-01 4.20879245e-01 1.86140209e-01 1.38695585e-02
-2.38758996e-01 2.37415075e-01 -2.85590023e-01 2.06346780e-01
-1.91971347e-01 4.86417443e-01 -5.33950448e-01 -7.33645201e-01
4.75398660e-01 1.72589451e-01 -7.34940946e-01 1.29837245e-01
2.76508242e-01 1.48142919e-01 -3.12994689e-01 7.64342844e-01
9.15400386e-01 -1.27057672e-01 2.63835996e-01 -3.06817502e-01
-1.94294423e-01 7.61204660e-02 -1.43215704e+00 7.02790737e-01
-7.82152712e-01 9.80854630e-01 -1.73040107e-01 -1.72255397e+00
1.08557487e+00 4.73034263e-01 1.01510608e+00 -7.02729702e-01
3.28174055e-01 4.40752059e-01 4.94446307e-02 -4.95688945e-01
1.16706684e-01 2.58660913e-01 1.86246917e-01 3.70626867e-01
1.59447119e-01 1.27886876e-01 6.22027516e-01 1.89818606e-01
1.12801802e+00 -2.91864872e-01 5.44081852e-02 2.12371852e-02
7.87721634e-01 -1.16266571e-02 7.44204819e-01 7.54632175e-01
-3.85742605e-01 2.26476401e-01 6.82109594e-01 -4.68235731e-01
-9.07817423e-01 -1.00173962e+00 -3.34440678e-01 9.20693576e-01
-2.69361008e-02 -2.46497005e-01 -5.38565159e-01 -8.02125871e-01
7.68353194e-02 6.62481248e-01 -1.13755316e-01 1.33991018e-01
-6.29585207e-01 -3.81013542e-01 4.53370214e-01 3.25648516e-01
1.52204260e-01 -9.71025765e-01 -1.84378460e-01 3.05737048e-01
-1.11984275e-01 -1.63414145e+00 -1.56231057e-02 9.37022492e-02
-9.69800234e-01 -1.18881631e+00 -3.63401622e-01 -8.98362458e-01
5.02663791e-01 6.02236927e-01 9.97395992e-01 2.58057475e-01
-4.73183423e-01 2.53791898e-01 -4.93392289e-01 -1.12152353e-01
-4.45735365e-01 2.67279953e-01 1.61625668e-01 1.84768230e-01
3.92111689e-01 -7.63759017e-01 -1.67033136e-01 7.97072113e-01
-3.98968548e-01 -2.37126917e-01 3.64676386e-01 3.73637974e-01
3.06457549e-01 4.31821495e-01 7.48183727e-01 -1.03617525e+00
4.52979952e-01 -8.59075963e-01 -5.44088781e-01 8.65658000e-02
-6.91409290e-01 -6.81324959e-01 9.75885630e-01 -5.65714836e-01
-6.15486145e-01 -3.76529336e-01 -2.89304763e-01 -4.34062630e-01
-2.37560049e-01 2.78923124e-01 1.62821431e-02 -2.05943674e-01
5.16143799e-01 1.71105832e-01 1.24445952e-01 -3.90855998e-01
-1.21934097e-02 1.21571016e+00 -2.23069385e-01 -4.72890198e-01
7.18754470e-01 2.81336755e-01 7.27502629e-02 -1.37852716e+00
-2.01795667e-01 -7.46667325e-01 -4.23685193e-01 -5.39134562e-01
3.03693473e-01 -4.16241914e-01 -1.07584584e+00 -5.55046946e-02
-8.19000781e-01 9.73103452e-04 -3.06847483e-01 6.84432626e-01
-4.82432276e-01 5.97279370e-01 -3.42389047e-01 -1.10915613e+00
-6.35689348e-02 -1.17432153e+00 4.84466463e-01 1.91190451e-01
-3.50690544e-01 -9.57246900e-01 -4.23619211e-01 6.02631629e-01
6.86867177e-01 -1.75911695e-01 9.50533569e-01 -1.01710820e+00
-5.63804626e-01 -5.90813577e-01 -4.10357147e-01 6.12402141e-01
9.92702320e-02 4.94703352e-01 -6.06743753e-01 -3.52011025e-01
-3.61984164e-01 4.01362404e-02 4.46792394e-01 4.67982471e-01
1.46846998e+00 2.12663412e-02 -4.52606410e-01 6.95530415e-01
1.31033373e+00 7.41814733e-01 5.99891663e-01 2.24411473e-01
5.12782872e-01 6.63237393e-01 4.81753230e-01 2.91528255e-01
1.61987424e-01 8.61580729e-01 2.45088860e-01 -2.39025921e-01
-6.05144389e-02 3.27187777e-01 1.52579039e-01 9.26504195e-01
-3.18522871e-01 -4.17178124e-01 -8.53961527e-01 2.27389768e-01
-1.53413582e+00 -1.37730694e+00 -4.64527458e-01 2.10730958e+00
1.04597114e-01 5.97422421e-01 5.19813895e-01 7.69469321e-01
7.34316409e-01 -1.60288483e-01 -1.66391023e-02 -9.27098930e-01
1.19412147e-01 4.89064544e-01 6.41900182e-01 4.47892576e-01
-7.72482812e-01 9.27023590e-01 6.55178213e+00 1.16396141e+00
-1.36232984e+00 -1.50089383e-01 5.10559976e-01 2.71014363e-01
-1.52395964e-01 2.06005797e-01 -6.65311933e-01 8.11538875e-01
1.53838289e+00 -1.47876218e-01 4.31429744e-01 9.53368962e-01
6.16410196e-01 -7.95093179e-03 -7.19596565e-01 1.07109964e+00
-1.49186730e-01 -1.43675840e+00 2.36827984e-01 1.02178589e-01
3.37893695e-01 -1.22153811e-01 -2.24527314e-01 5.96980691e-01
-1.12283379e-01 -7.85800874e-01 4.17943895e-01 5.16385674e-01
7.87166357e-01 -7.90771186e-01 1.07933307e+00 1.27150029e-01
-1.33335280e+00 -4.85747993e-01 -1.73043191e-01 1.65006608e-01
3.50622684e-01 3.77594441e-01 -9.57815647e-01 6.78409159e-01
3.90358776e-01 4.71780181e-01 -4.00299668e-01 1.37853682e+00
6.51110828e-01 1.04924130e+00 -2.74910152e-01 -1.41822770e-01
1.45582370e-02 -2.06087112e-01 3.46062660e-01 1.43047976e+00
3.55232656e-01 -2.00332463e-01 5.63440323e-01 2.57039011e-01
3.53996754e-01 6.17113829e-01 -3.11207294e-01 5.77906854e-02
7.24296093e-01 1.10398221e+00 -1.01902592e+00 -3.16080213e-01
-4.86448705e-01 6.65181458e-01 -3.12449336e-01 5.72363138e-01
-1.01239336e+00 -6.74297392e-01 5.14445603e-01 4.24433798e-01
3.59000623e-01 -1.94566652e-01 -2.93711722e-01 -8.11247647e-01
-1.58893347e-01 -7.14877307e-01 2.78729081e-01 -3.15954596e-01
-1.06575656e+00 8.98575783e-01 4.05904874e-02 -1.60640550e+00
-2.61682659e-01 -6.06394172e-01 -6.89571440e-01 6.82666481e-01
-1.40210843e+00 -8.55287373e-01 -5.09616077e-01 6.05946839e-01
6.63849056e-01 -8.99666190e-01 8.47047925e-01 7.24554539e-01
-8.48176658e-01 8.67073298e-01 -8.04451108e-02 -7.39604095e-03
5.07831573e-02 -7.84498990e-01 -3.70081738e-02 4.72348422e-01
2.76591666e-02 4.43367392e-01 5.20130575e-01 -3.52463007e-01
-8.69532049e-01 -8.36585999e-01 1.00760078e+00 -8.43338668e-02
5.73633790e-01 -1.95636332e-01 -6.21637285e-01 4.00449306e-01
-4.00375724e-01 1.53316692e-01 1.20134735e+00 7.17900172e-02
-1.59143656e-01 -5.08489311e-01 -1.33260286e+00 3.02908599e-01
8.72489393e-01 -2.80247003e-01 1.05284210e-02 1.70472115e-01
3.64676684e-01 2.31589198e-01 -8.89537871e-01 2.51905203e-01
6.48479879e-01 -9.29211259e-01 9.19573963e-01 -8.18942249e-01
-3.50307733e-01 -6.20837584e-02 -5.80219388e-01 -7.55299389e-01
-2.56991178e-01 -5.79700410e-01 -1.30573764e-01 1.26845908e+00
5.57459950e-01 -8.03808272e-01 1.16955912e+00 1.13046832e-01
-3.41954678e-02 -8.00174356e-01 -1.00750875e+00 -1.16886961e+00
-5.35005748e-01 -9.86029565e-01 5.41090548e-01 9.62345004e-01
-2.65635550e-01 2.11130857e-01 -1.28742635e-01 -4.99139242e-02
5.76816142e-01 1.14238001e-01 9.62797761e-01 -1.73533726e+00
-1.28950896e-02 -6.29256248e-01 -1.08623433e+00 -8.79703760e-01
4.19799536e-01 -1.10150611e+00 -6.69233918e-01 -1.28060126e+00
-7.56907985e-02 -1.18940568e+00 -2.53000230e-01 1.66234121e-01
4.86800075e-01 1.95925549e-01 2.02906847e-01 2.62985140e-01
-5.80698848e-01 1.79850861e-01 7.99278796e-01 1.38689220e-01
-2.82328963e-01 7.66084075e-01 -3.71912837e-01 3.51522893e-01
1.06183553e+00 -3.96529287e-01 -6.39912903e-01 3.22889984e-01
-4.43143606e-01 2.87470013e-01 4.12142090e-02 -1.11650181e+00
-3.97911295e-02 -3.67772222e-01 2.99526393e-01 -5.42851090e-01
3.13645393e-01 -1.32530034e+00 1.20456055e-01 4.15402174e-01
1.02094568e-01 -1.64216220e-01 7.16761593e-03 2.68747181e-01
4.55612410e-03 -4.05047268e-01 6.83586836e-01 2.34559283e-01
-8.37670922e-01 2.66810089e-01 -8.36500883e-01 -9.33259726e-02
1.31676567e+00 -8.38776052e-01 1.21316314e-01 -5.49959600e-01
-8.03920746e-01 2.61471383e-02 1.78682312e-01 4.62125659e-01
4.88064915e-01 -1.07374787e+00 -4.73317623e-01 4.22834992e-01
-9.45900828e-02 -8.75847161e-01 1.82447851e-01 6.80724084e-01
-9.74162042e-01 6.72637522e-01 -4.44390148e-01 -8.89996052e-01
-1.43208826e+00 3.79949450e-01 2.99165875e-01 9.87634528e-03
-4.39126015e-01 5.04799902e-01 -7.36590445e-01 -4.36114252e-01
4.35162961e-01 7.71949291e-02 -4.17006463e-01 2.79589772e-01
2.72401124e-01 8.68861675e-01 2.18360543e-01 -8.47128570e-01
-5.28161824e-01 4.65841681e-01 -1.63800269e-01 1.96331069e-01
1.19975781e+00 -8.09528008e-02 9.63300020e-02 2.95752138e-01
1.60957134e+00 5.08385040e-02 -5.66934705e-01 4.05961052e-02
2.22781692e-02 -8.17348123e-01 2.68849395e-02 -3.74959469e-01
-1.41823459e+00 6.30318522e-01 8.97496939e-01 8.25829148e-01
1.11792195e+00 -3.20358723e-01 8.52285683e-01 1.83729082e-01
5.84330678e-01 -9.42011476e-01 -2.07997173e-01 4.20670629e-01
1.81947440e-01 -1.01372671e+00 -1.37558743e-01 -8.48232985e-01
-5.76132178e-01 1.29599011e+00 5.05972147e-01 1.29335880e-01
1.00439239e+00 1.70000091e-01 1.28426388e-01 -7.65637457e-02
-7.71196365e-01 -3.49905014e-01 -1.06574565e-01 8.36733639e-01
3.15489739e-01 2.97807693e-01 -4.82357383e-01 2.52893448e-01
-3.40780705e-01 -6.04686029e-02 6.21652067e-01 8.56778443e-01
-4.90906388e-01 -1.19317698e+00 3.58060701e-03 8.04201782e-01
-4.28658575e-01 2.60061473e-01 1.86261684e-01 7.61419058e-01
1.63911711e-02 1.33613837e+00 3.06620449e-01 -7.90838420e-01
4.74944443e-01 -2.86357515e-02 3.74150015e-02 -2.40443781e-01
-2.73961902e-01 -1.88517243e-01 3.29731166e-01 -8.00699651e-01
-3.31654698e-02 -6.50357187e-01 -1.12329924e+00 -8.91372144e-01
-3.91528487e-01 6.36942148e-01 9.01730001e-01 8.73410523e-01
1.84350207e-01 7.26924241e-01 1.31645012e+00 -7.40961611e-01
-1.35875687e-01 -6.88604414e-01 -6.99681640e-01 4.14892286e-01
3.34261954e-02 -7.85710633e-01 -5.85647762e-01 -2.86508203e-01] | [5.067710876464844, 7.195638656616211] |
b65e6606-cc55-4a86-a4d4-f926caddaadc | think-twice-a-human-like-two-stage | 2301.04907 | null | https://arxiv.org/abs/2301.04907v2 | https://arxiv.org/pdf/2301.04907v2.pdf | Think Twice: A Human-like Two-stage Conversational Agent for Emotional Response Generation | Towards human-like dialogue systems, current emotional dialogue approaches jointly model emotion and semantics with a unified neural network. This strategy tends to generate safe responses due to the mutual restriction between emotion and semantics, and requires rare emotion-annotated large-scale dialogue corpus. Inspired by the "think twice" behavior in human dialogue, we propose a two-stage conversational agent for the generation of emotional dialogue. Firstly, a dialogue model trained without the emotion-annotated dialogue corpus generates a prototype response that meets the contextual semantics. Secondly, the first-stage prototype is modified by a controllable emotion refiner with the empathy hypothesis. Experimental results on the DailyDialog and EmpatheticDialogues datasets demonstrate that the proposed conversational outperforms the comparison models in emotion generation and maintains the semantic performance in automatic and human evaluations. | ['Yuexian Hou', 'Kun Huang', 'Dongming Zhao', 'Shuo Zhang', 'Wu Bin', 'Shangzhao Ma', 'Bo wang', 'Yushan Qian'] | 2023-01-12 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [-1.88669905e-01 9.58989084e-01 1.72672838e-01 -6.75198615e-01
-2.47478724e-01 -2.90826172e-01 7.56711066e-01 -1.69600174e-01
-2.87072599e-01 9.92248058e-01 8.63208234e-01 3.66094053e-01
6.29712164e-01 -7.31896043e-01 1.97520390e-01 -3.53714049e-01
3.65029603e-01 7.81719744e-01 -3.61615747e-01 -9.45115626e-01
1.16164371e-01 -1.23254344e-01 -1.15633142e+00 7.09387779e-01
1.11681056e+00 9.54554617e-01 -1.18636072e-01 7.03584373e-01
-5.26333272e-01 1.35782957e+00 -9.06313479e-01 -8.37873459e-01
-2.73764670e-01 -1.14026618e+00 -1.42993128e+00 8.53550285e-02
-7.58852899e-01 -4.95209873e-01 4.63601500e-02 8.32460165e-01
8.64985347e-01 5.59204638e-01 6.56921685e-01 -1.30525136e+00
-7.80942857e-01 9.30593252e-01 8.57229233e-02 -6.13164961e-01
9.95765507e-01 8.13428164e-02 9.14799452e-01 -6.17059052e-01
8.91751230e-01 1.53077209e+00 5.68086386e-01 1.50077760e+00
-8.55835915e-01 -2.13183552e-01 -1.36977047e-01 4.44444567e-02
-6.67276561e-01 -3.51472139e-01 1.07579100e+00 -2.52550066e-01
1.24484360e+00 2.00569391e-01 8.14815402e-01 1.61588597e+00
-1.16664059e-01 8.95254433e-01 1.04853630e+00 -2.82509267e-01
2.05619499e-01 7.18524337e-01 1.08646981e-01 3.68749112e-01
-8.18652689e-01 -8.20923075e-02 -5.73354006e-01 -3.23592514e-01
4.33512181e-01 -7.06323802e-01 -3.26058924e-01 -2.00605430e-02
-1.00617146e+00 1.29156590e+00 1.46443918e-01 2.78972059e-01
-6.63815618e-01 -4.27512795e-01 1.13723826e+00 6.18753731e-01
6.55394495e-01 1.08523881e+00 -2.79830068e-01 -6.03796005e-01
-3.46892476e-01 3.14653844e-01 1.65726554e+00 8.37702930e-01
3.95739883e-01 5.00704870e-02 -4.32590246e-01 1.39859557e+00
-3.28709930e-02 6.39095977e-02 8.01753640e-01 -1.16178310e+00
-2.32367832e-02 7.22020090e-01 4.93090510e-01 -9.43187177e-01
-7.30273843e-01 3.22462946e-01 -9.98140514e-01 -2.27526978e-01
7.20548183e-02 -8.57105196e-01 5.26671819e-02 1.83601928e+00
6.72434092e-01 -3.68365765e-01 9.46930110e-01 1.22119784e+00
1.38823795e+00 9.09605980e-01 3.88804615e-01 -5.39821386e-01
1.35773277e+00 -1.43857193e+00 -1.08098376e+00 -8.95172954e-02
7.61193573e-01 -4.95307982e-01 1.15364504e+00 2.98772782e-01
-1.12728822e+00 -4.48498189e-01 -6.76413894e-01 -1.08049296e-01
-2.25461319e-01 6.50470853e-02 7.17961550e-01 3.20857406e-01
-9.17831779e-01 1.46372095e-01 1.36429250e-01 -6.23150647e-01
-2.47485638e-01 -6.38832301e-02 -2.17207029e-01 5.08735120e-01
-1.90063083e+00 1.29884577e+00 5.13267040e-01 1.43059596e-01
-5.05854487e-01 -2.95103133e-01 -1.03577745e+00 4.25015716e-03
2.00163648e-02 -8.06698740e-01 1.58131111e+00 -1.50845313e+00
-2.50901842e+00 1.01414263e+00 2.37768903e-01 -4.03999507e-01
6.04097307e-01 -2.72290051e-01 -3.56089264e-01 3.57361168e-01
-2.46508330e-01 1.12245476e+00 4.19190288e-01 -1.25028896e+00
-4.27638769e-01 2.12585315e-01 1.96858227e-01 7.14980245e-01
-2.48399571e-01 2.74730951e-01 5.22689335e-02 -3.96821946e-01
-5.99303305e-01 -7.63900101e-01 -4.21987742e-01 -6.01933718e-01
-4.64250624e-01 -6.83885455e-01 3.51950705e-01 -4.75077599e-01
9.17289197e-01 -1.78871834e+00 2.97335684e-01 -7.39049315e-02
3.14173512e-02 6.22939244e-02 -2.87198901e-01 8.07498693e-01
-6.49647741e-03 -1.96912766e-01 8.31361935e-02 -3.28675926e-01
3.45900387e-01 7.85861835e-02 -4.33737010e-01 -1.89448416e-01
1.31182358e-01 9.20362890e-01 -1.18566453e+00 -6.82338595e-01
8.39017853e-02 2.11723849e-01 -7.20968902e-01 1.20058668e+00
-4.79085386e-01 7.72020102e-01 -6.11884534e-01 5.33989333e-02
1.71887472e-01 -1.20176665e-01 3.77649128e-01 -1.50739640e-01
1.99314088e-01 2.91828573e-01 -5.21292627e-01 1.72316301e+00
-7.62677789e-01 1.97954372e-01 9.62141454e-02 -7.13544548e-01
1.50545597e+00 8.50001752e-01 3.42929184e-01 -6.55780196e-01
4.48697537e-01 -5.96258491e-02 -9.10433233e-02 -1.03223550e+00
8.89123619e-01 -6.72397614e-01 -7.80695379e-01 9.45587933e-01
2.18409002e-01 -6.53327286e-01 -1.22133665e-01 3.82504851e-01
7.61306703e-01 3.02531570e-01 4.16175842e-01 7.16172531e-02
8.00893545e-01 2.42017999e-01 5.31091452e-01 5.07970154e-01
-5.25677741e-01 1.77614614e-01 7.98175454e-01 -5.02098203e-01
-7.93756962e-01 -4.52617496e-01 3.47465068e-01 1.41641343e+00
2.73114443e-01 -2.41904125e-01 -1.10359776e+00 -6.67621136e-01
-5.86936653e-01 1.06111801e+00 -6.01420820e-01 -2.49897704e-01
-2.02578366e-01 -4.62216139e-01 8.79958928e-01 1.63764343e-01
7.07782686e-01 -1.92602289e+00 -8.41332734e-01 5.63811243e-01
-8.38275135e-01 -1.09713495e+00 -1.57032907e-01 -1.92002848e-01
-2.25437865e-01 -7.44604945e-01 -5.03520310e-01 -8.60481381e-01
3.50685209e-01 -4.35313374e-01 1.25369334e+00 9.12809148e-02
1.50977299e-01 3.21033150e-01 -8.47674906e-01 -9.64944288e-02
-1.04628515e+00 -1.14646666e-01 -1.02816053e-01 -8.12624916e-02
5.22472560e-01 -3.41367602e-01 -4.50878084e-01 1.59029722e-01
-5.67796946e-01 5.67037046e-01 2.54364721e-02 1.17038155e+00
-2.68156081e-01 -4.35778946e-01 1.34874225e+00 -8.74710381e-01
1.75770533e+00 -6.22198522e-01 3.38555902e-01 3.13716441e-01
-4.22300071e-01 -1.26710132e-01 9.26936209e-01 -5.67730665e-01
-1.91965687e+00 -1.60996184e-01 -2.96846032e-01 9.10543054e-02
-3.88065666e-01 2.93543816e-01 -1.90146148e-01 5.04677117e-01
7.76471198e-01 4.22156639e-02 1.79479003e-01 -4.51168828e-02
7.98867524e-01 9.99409258e-01 7.32796967e-01 -9.95355189e-01
-9.81792063e-02 -1.86483916e-02 -8.77197146e-01 -5.33482015e-01
-8.19904685e-01 -1.86934397e-01 -2.56127894e-01 -7.27352917e-01
1.06278384e+00 -8.43461752e-01 -1.18182409e+00 4.73691881e-01
-1.79934692e+00 -5.61907291e-01 -2.42982268e-01 3.88711661e-01
-9.88089740e-01 2.64399350e-01 -1.20376110e+00 -9.93505299e-01
-8.84027362e-01 -7.45021582e-01 8.10308397e-01 5.33005416e-01
-1.13585234e+00 -1.06220591e+00 4.04211760e-01 4.37801033e-01
3.81581396e-01 3.45613331e-01 9.59111452e-01 -1.15472436e+00
5.94438314e-01 -1.47579862e-02 -4.86662649e-02 3.26541722e-01
-1.71032116e-01 -2.27486074e-01 -8.08160841e-01 3.98283601e-01
2.73602813e-01 -1.41063476e+00 7.12544695e-02 -3.39856833e-01
6.17589772e-01 -6.93466723e-01 1.73373207e-01 3.08265723e-02
7.13224351e-01 3.01849961e-01 6.50592506e-01 5.12414984e-02
2.01568320e-01 1.40259993e+00 9.37303543e-01 9.15687442e-01
8.02063763e-01 2.35733315e-01 -8.69974419e-02 -1.20123051e-01
3.78360480e-01 -3.82058084e-01 3.75345379e-01 9.99312043e-01
8.65415335e-02 -2.83174068e-01 -5.78644037e-01 4.43485677e-01
-2.18682146e+00 -1.12170410e+00 8.34259540e-02 1.24061072e+00
1.47447085e+00 -3.53533059e-01 1.10468455e-01 -5.57870448e-01
7.94248581e-01 3.01389635e-01 -5.21220803e-01 -1.42896569e+00
-1.31908253e-01 1.91316009e-02 -6.49495184e-01 6.28969729e-01
-4.69031066e-01 1.42401767e+00 5.92248535e+00 3.48785013e-01
-9.48492527e-01 1.33582175e-01 7.75104642e-01 2.96956822e-02
-3.71857375e-01 -2.21650347e-01 -1.53852105e-01 1.94390431e-01
9.08470809e-01 -4.56074029e-01 3.11236680e-01 1.02124965e+00
4.36578244e-01 -4.28444408e-02 -1.19100463e+00 7.75704563e-01
3.01533639e-01 -1.00486135e+00 3.22582549e-03 -5.53881288e-01
5.59380710e-01 -6.46804035e-01 -5.32495677e-01 9.07444119e-01
6.58410251e-01 -1.00998902e+00 3.64243388e-01 7.30408251e-01
4.39147294e-01 -8.29647839e-01 9.80129302e-01 5.38300514e-01
-4.15368527e-01 2.12006465e-01 -2.26294339e-01 -1.89286023e-01
6.23731852e-01 8.37046094e-03 -1.03778255e+00 2.86539793e-01
3.71684402e-01 4.37977463e-01 9.76498276e-02 1.42205089e-01
-5.94382226e-01 1.64991349e-01 2.12712094e-01 -5.23548782e-01
3.16646874e-01 -5.61077595e-01 4.39478189e-01 1.39204824e+00
-1.56720921e-01 6.52324557e-01 1.59561113e-01 9.42558289e-01
-1.51853487e-01 6.38309717e-01 -4.91506189e-01 -1.67532578e-01
5.06603479e-01 1.68591082e+00 -1.11973189e-01 -5.90995133e-01
-1.24087758e-01 1.37344515e+00 5.07255614e-01 2.79115617e-01
-8.26889575e-01 -4.68978971e-01 3.22813630e-01 -6.99653089e-01
-3.56981009e-01 5.34199893e-01 -1.79822698e-01 -9.96730208e-01
-4.93526310e-01 -1.23230374e+00 3.36660832e-01 -1.31348312e+00
-1.54907131e+00 1.17624414e+00 -3.51963460e-01 -7.49850810e-01
-9.68820632e-01 -1.19359992e-01 -1.01964879e+00 7.86475778e-01
-8.82211745e-01 -1.15928805e+00 -3.57768625e-01 5.46360552e-01
7.56866455e-01 -1.12014145e-01 1.48549438e+00 -2.41637483e-01
-4.87005532e-01 4.27124202e-01 -7.24277079e-01 1.63868278e-01
1.02590513e+00 -1.15075529e+00 -3.35011482e-02 -4.72073853e-02
-8.71628940e-01 3.16961497e-01 1.17021310e+00 -5.64328015e-01
-7.81113029e-01 -7.04024315e-01 1.11736810e+00 -1.15083762e-01
7.96839237e-01 -1.37691364e-01 -1.03835607e+00 2.76561946e-01
1.08330059e+00 -9.17341530e-01 1.13405895e+00 1.58041596e-01
-1.21145248e-01 5.66564083e-01 -1.50602436e+00 8.46481979e-01
7.11462259e-01 -4.29395705e-01 -1.20780456e+00 3.87464911e-01
9.51731563e-01 -4.16660815e-01 -1.20411384e+00 2.25021839e-01
5.70356309e-01 -1.08547950e+00 4.33048010e-01 -1.06450629e+00
1.02397716e+00 3.18092644e-01 2.64439046e-01 -1.78389144e+00
2.69519657e-01 -1.17253745e+00 1.38344809e-01 1.44444215e+00
3.50450158e-01 -4.53124940e-01 4.03912216e-01 1.25730932e+00
-1.18131958e-01 -6.26692653e-01 -5.01478672e-01 -1.10470809e-01
1.16312519e-01 -2.43396824e-03 7.64804900e-01 1.38245213e+00
1.17464340e+00 1.08460462e+00 -8.45851183e-01 -5.51011026e-01
8.89826100e-03 2.84815162e-01 1.02997947e+00 -9.86182868e-01
-2.01192319e-01 -5.17634571e-01 3.24734569e-01 -1.03701174e+00
9.85890150e-01 -4.91418064e-01 5.62758744e-01 -1.51641583e+00
1.80292521e-02 -1.89947218e-01 3.76402438e-01 2.84890950e-01
-3.42482001e-01 -2.78281599e-01 -4.51077186e-02 -7.37028942e-02
-8.13195884e-01 1.25287187e+00 1.33731782e+00 3.14462811e-01
-5.78493178e-01 -3.59511435e-01 -7.95828760e-01 8.09602141e-01
9.41695869e-01 -8.42456147e-02 -4.38203841e-01 9.99091789e-02
3.50783229e-01 7.55189776e-01 -1.58260465e-02 -3.56246114e-01
1.62738889e-01 -4.77438092e-01 -3.79219390e-02 -2.56899238e-01
5.09271622e-01 -3.98687661e-01 -1.60453111e-01 7.82199875e-02
-1.05580616e+00 -9.16219056e-02 -1.80206433e-01 1.77736908e-01
-4.41830158e-01 -3.03371578e-01 8.83536637e-01 -3.29147398e-01
-6.04172587e-01 -2.47448534e-01 -6.86886847e-01 2.71221548e-01
9.80300725e-01 -1.13647588e-01 -5.09193480e-01 -1.12077379e+00
-8.77212405e-01 6.60347700e-01 3.20612669e-01 7.08350539e-01
5.70060611e-01 -1.40386438e+00 -8.38011265e-01 -3.47849488e-01
1.84034705e-01 -1.09051667e-01 4.82554436e-01 5.52656472e-01
-1.56862423e-01 5.14653996e-02 -4.91977781e-01 6.52816845e-03
-9.61287677e-01 3.67861807e-01 6.06428266e-01 -3.70221913e-01
-5.33805072e-01 7.77496576e-01 2.12537363e-01 -9.91169870e-01
1.18177012e-01 3.48972470e-01 -6.69868767e-01 2.08716199e-01
5.00983179e-01 8.24967921e-02 -6.26523733e-01 -7.67937660e-01
1.23815112e-01 -2.82143593e-01 7.46749295e-03 -5.36670804e-01
1.20605993e+00 -3.70479912e-01 -4.00967300e-01 4.91569012e-01
9.22704220e-01 -1.61903784e-01 -8.76323938e-01 -1.20046191e-01
-3.68752680e-03 7.64349103e-02 -6.59063458e-01 -1.03713286e+00
-5.35845816e-01 5.94572365e-01 -1.81821704e-01 5.33540249e-01
9.42852557e-01 -1.10135429e-01 1.16959000e+00 6.17408693e-01
8.36411640e-02 -1.75803363e+00 5.43391645e-01 8.53537261e-01
1.26178467e+00 -1.16748989e+00 -4.87699717e-01 -1.31288618e-01
-1.85928869e+00 1.21858740e+00 1.41548085e+00 3.73496078e-02
1.03183992e-01 -1.12687470e-03 6.54576898e-01 -2.88431674e-01
-1.32939112e+00 7.16146752e-02 -2.52178341e-01 5.88145435e-01
6.31050885e-01 1.63699880e-01 -6.36894941e-01 1.23049843e+00
-6.97865546e-01 -2.60278732e-02 7.16351807e-01 4.57107216e-01
-4.15530026e-01 -9.21163499e-01 -2.68450845e-02 -4.69094329e-02
-1.42183259e-01 1.32480552e-02 -1.19158375e+00 5.22488236e-01
-3.10609877e-01 1.42194116e+00 3.22227962e-02 -3.59319687e-01
4.39560890e-01 5.92205942e-01 5.04357852e-02 -6.55118048e-01
-1.34814262e+00 -1.71968997e-01 9.70878243e-01 -5.74458420e-01
-6.35063052e-01 -1.01467088e-01 -1.78571761e+00 -2.32852533e-01
-1.04882702e-01 8.26036692e-01 4.45058554e-01 1.07976460e+00
4.35985804e-01 2.23594904e-01 1.03293037e+00 -5.70357621e-01
-6.13957524e-01 -1.42364395e+00 -3.44826818e-01 1.00960815e+00
-2.09159255e-01 -2.70281196e-01 -3.06952536e-01 -2.93459790e-03] | [13.104368209838867, 7.6696672439575195] |
bc44ecd0-9fde-46ff-a2d1-f4b8f3386cc3 | knowledge-graph-question-answering-datasets | 2205.06573 | null | https://arxiv.org/abs/2205.06573v1 | https://arxiv.org/pdf/2205.06573v1.pdf | Knowledge Graph Question Answering Datasets and Their Generalizability: Are They Enough for Future Research? | Existing approaches on Question Answering over Knowledge Graphs (KGQA) have weak generalizability. That is often due to the standard i.i.d. assumption on the underlying dataset. Recently, three levels of generalization for KGQA were defined, namely i.i.d., compositional, zero-shot. We analyze 25 well-known KGQA datasets for 5 different Knowledge Graphs (KGs). We show that according to this definition many existing and online available KGQA datasets are either not suited to train a generalizable KGQA system or that the datasets are based on discontinued and out-dated KGs. Generating new datasets is a costly process and, thus, is not an alternative to smaller research groups and companies. In this work, we propose a mitigation method for re-splitting available KGQA datasets to enable their applicability to evaluate generalization, without any cost and manual effort. We test our hypothesis on three KGQA datasets, i.e., LC-QuAD, LC-QuAD 2.0 and QALD-9). Experiments on re-splitted KGQA datasets demonstrate its effectiveness towards generalizability. The code and a unified way to access 18 available datasets is online at https://github.com/semantic-systems/KGQA-datasets as well as https://github.com/semantic-systems/KGQA-datasets-generalization. | ['Ricardo Usbeck', 'Longquan Jiang'] | 2022-05-13 | null | null | null | null | ['graph-question-answering'] | ['graphs'] | [-3.07859004e-01 4.33375180e-01 -1.42986804e-01 -4.27985817e-01
-8.23810220e-01 -9.13752615e-01 2.87100852e-01 3.68470848e-01
-8.44939873e-02 8.76328886e-01 1.12455972e-02 -4.14488137e-01
-5.48729122e-01 -1.30666411e+00 -7.71633744e-01 -1.92871168e-01
2.82021910e-01 8.88961077e-01 4.97410566e-01 -5.63078642e-01
1.26962960e-01 2.56684780e-01 -1.63911605e+00 5.00997543e-01
1.20978189e+00 8.38892698e-01 4.79939170e-02 5.04263878e-01
-4.46622550e-01 6.01482987e-01 -6.90854967e-01 -9.93128181e-01
3.49554241e-01 -4.14133102e-01 -1.39657044e+00 -2.64746428e-01
7.55539358e-01 3.11256535e-02 -2.18690470e-01 1.06410825e+00
4.18138623e-01 9.34781332e-04 4.15168107e-01 -1.85220492e+00
-1.16623318e+00 7.36174345e-01 7.26823509e-02 -2.81234123e-02
5.40482819e-01 -7.29624778e-02 1.10665774e+00 -5.79145968e-01
6.84258759e-01 1.15729213e+00 8.04534793e-01 6.23506904e-01
-7.85103500e-01 -6.41244233e-01 -6.68917373e-02 5.76880038e-01
-1.72985220e+00 -1.61073670e-01 6.10198200e-01 -2.00983301e-01
7.17707038e-01 4.37442183e-01 4.48070228e-01 1.18402469e+00
-2.86596537e-01 4.63996738e-01 1.11027467e+00 -4.62617695e-01
3.56571317e-01 3.06482673e-01 3.80540013e-01 7.15854287e-01
7.89415598e-01 -3.98436636e-01 -5.64671814e-01 -3.77670348e-01
3.83917004e-01 -4.59295988e-01 -3.88189077e-01 -5.74178696e-01
-6.59119189e-01 7.55544782e-01 3.24893147e-01 3.26528847e-01
-2.99815983e-01 -6.65784702e-02 3.96333307e-01 6.18379414e-01
1.40061364e-01 5.97338378e-01 -7.93626249e-01 -1.25791907e-01
-4.76467818e-01 6.24027193e-01 1.04141951e+00 1.39407325e+00
1.08687246e+00 -3.97600025e-01 1.01195030e-01 7.29742587e-01
2.66118288e-01 4.20586914e-01 5.23696065e-01 -8.33573699e-01
4.27407920e-01 1.15052736e+00 -1.70762837e-02 -8.00341904e-01
-4.48148578e-01 -3.66445750e-01 -4.09987777e-01 -4.11121577e-01
4.83062625e-01 1.13939770e-01 -8.63260448e-01 1.57916522e+00
4.88489926e-01 -2.62528788e-02 4.49887574e-01 6.98529303e-01
1.30392289e+00 2.90609986e-01 2.46820420e-01 1.82135537e-01
1.31741333e+00 -7.86658227e-01 -6.59012258e-01 7.64802173e-02
1.04176366e+00 -3.85042489e-01 1.52426636e+00 5.07946849e-01
-7.74429739e-01 -3.78320068e-01 -8.52720976e-01 -1.70116007e-01
-1.24168074e+00 -1.80378392e-01 6.59917891e-01 1.09603179e+00
-1.13245475e+00 4.00073230e-01 -4.10383672e-01 -7.25402057e-01
4.13265228e-01 1.14094235e-01 -5.18878579e-01 -4.17026669e-01
-1.69267571e+00 9.08094466e-01 9.47474062e-01 -2.67295033e-01
-8.84186089e-01 -9.94812667e-01 -6.44899964e-01 -7.21184686e-02
7.00576663e-01 -1.08746767e+00 1.20920014e+00 -8.13783288e-01
-1.20154476e+00 8.12157989e-01 1.57065421e-01 -3.97256166e-01
3.19223493e-01 -1.98716253e-01 -8.44043672e-01 7.78878778e-02
1.70120418e-01 4.14722502e-01 4.36871380e-01 -1.40640318e+00
-5.58258176e-01 -6.36924446e-01 6.22204125e-01 1.31469250e-01
-3.28846425e-01 -4.14269716e-01 -3.02789629e-01 -2.84056395e-01
-1.46312818e-01 -8.04158270e-01 1.93993926e-01 -4.70721871e-01
-3.13608587e-01 -4.98493671e-01 7.26513922e-01 -7.27677941e-01
1.47086811e+00 -1.65803742e+00 -2.03547925e-01 2.47912318e-01
-2.08451343e-03 4.29504544e-01 -1.32954046e-01 8.17275882e-01
3.95947583e-02 4.24526602e-01 -2.67913431e-01 1.93954870e-01
6.92285448e-02 4.72002596e-01 -2.53392875e-01 -5.12696654e-02
7.13523775e-02 1.06382835e+00 -1.26262724e+00 -4.03013945e-01
2.13808510e-02 5.30984588e-02 -4.71827447e-01 4.66837920e-02
-6.17699206e-01 6.15580268e-02 -5.10400116e-01 9.66217577e-01
6.89453423e-01 -1.84042886e-01 2.53680259e-01 -5.29908240e-01
3.98622364e-01 2.05832496e-01 -1.17648089e+00 2.05037689e+00
-4.13116574e-01 3.82618383e-02 -5.00974238e-01 -8.99480581e-01
9.74384785e-01 3.42260599e-01 9.37230587e-02 -6.67424083e-01
-1.02680780e-01 6.03814125e-01 -2.27674946e-01 -5.53123534e-01
7.10305214e-01 -1.18810385e-01 -1.95740104e-01 1.68241262e-01
3.56566578e-01 -3.21027309e-01 5.45740664e-01 4.34562474e-01
1.08090389e+00 2.18927458e-01 2.54244000e-01 -5.01207232e-01
4.80368167e-01 3.73616487e-01 3.39627832e-01 7.94876993e-01
-1.21838450e-01 4.05197352e-01 4.28659797e-01 -2.28358343e-01
-6.91109538e-01 -1.21041465e+00 5.66216260e-02 9.53481138e-01
9.09273252e-02 -8.03726077e-01 -8.01514149e-01 -1.14770937e+00
1.14706762e-01 1.20808351e+00 -5.77762067e-01 -3.67991805e-01
-2.73405667e-03 -3.73084068e-01 1.10198009e+00 4.79478747e-01
6.17379248e-01 -8.51978481e-01 -4.52754915e-01 4.73459810e-02
-3.41778189e-01 -9.38636899e-01 1.89122573e-01 -2.20694065e-01
-6.97958112e-01 -1.49311483e+00 -2.39163920e-01 -2.98767507e-01
5.65465569e-01 8.07060897e-02 1.58241987e+00 7.67036900e-02
-4.89133522e-02 8.63874853e-01 -8.47134888e-01 -6.44701779e-01
-4.77835000e-01 2.72071421e-01 -1.52711019e-01 -3.71155113e-01
6.88565910e-01 -3.36308777e-01 -4.92487192e-01 3.87685776e-01
-1.27463543e+00 -2.27965534e-01 4.86341327e-01 5.03413677e-01
6.50190473e-01 2.35390872e-01 1.14109039e+00 -1.26303685e+00
9.11701560e-01 -7.06000745e-01 -5.80993950e-01 8.81538570e-01
-8.50407481e-01 -1.36437202e-02 4.96923029e-01 -4.56555001e-02
-1.21313417e+00 -4.14746642e-01 -1.87301606e-01 -4.36220676e-01
-3.13168257e-01 7.52781272e-01 -4.85305399e-01 -5.63992234e-03
1.04288638e+00 -8.96151140e-02 -2.33804360e-01 -4.17276263e-01
9.14707363e-01 8.71825337e-01 3.61227483e-01 -9.84445691e-01
6.27879322e-01 3.37922513e-01 -2.19933286e-01 -6.13578379e-01
-8.82541895e-01 -5.73399127e-01 -3.68380457e-01 -1.24261618e-01
6.43259645e-01 -7.39046991e-01 -4.46712255e-01 3.47217172e-01
-8.15981805e-01 -3.74315023e-01 -4.94565070e-01 8.82071480e-02
-4.92839277e-01 4.13623512e-01 -2.20062926e-01 -4.94080663e-01
-4.51214105e-01 -5.71878374e-01 9.32849288e-01 2.77535737e-01
-2.64786035e-01 -1.18130243e+00 -3.06333359e-02 9.25791144e-01
4.11089301e-01 1.68294162e-01 1.09134579e+00 -8.92580390e-01
-3.07204843e-01 -1.16409818e-02 -1.02742158e-01 6.14727437e-01
3.74090165e-01 -4.45893668e-02 -9.22804594e-01 -1.95340708e-01
-4.12663817e-01 -5.65963566e-01 3.95449460e-01 -1.50892586e-01
1.25110996e+00 -3.08857560e-01 -1.09811220e-02 9.52620730e-02
1.63644457e+00 6.44737855e-02 7.91736603e-01 4.26021427e-01
6.83783412e-01 6.55110002e-01 8.73315275e-01 1.64567441e-01
7.88848758e-01 4.76223379e-01 5.34170449e-01 4.56349969e-01
-2.61884451e-01 -4.63461101e-01 1.71094835e-01 8.99551928e-01
-7.26953447e-02 -5.46210706e-01 -1.23839831e+00 1.03479707e+00
-1.80097818e+00 -6.46806180e-01 -4.04045373e-01 2.32072330e+00
9.00054038e-01 -7.71730766e-02 1.33480337e-02 1.10445730e-03
4.72201914e-01 -4.33664143e-01 -4.02358741e-01 -5.58622420e-01
-2.27917060e-01 4.56582844e-01 4.33065325e-01 3.30576181e-01
-6.33708894e-01 1.06219769e+00 5.11921549e+00 9.61056471e-01
-5.74968457e-01 1.15140811e-01 1.38034895e-01 2.78901309e-01
-6.26482785e-01 2.90576607e-01 -7.22663105e-01 2.41232380e-01
1.05290198e+00 -5.98803103e-01 2.62790143e-01 8.44510794e-01
-3.65421087e-01 9.19219106e-02 -9.53803599e-01 7.29556978e-01
4.23751399e-02 -1.29949415e+00 6.89466119e-01 -1.97758287e-01
7.40307271e-01 -1.58459187e-01 -2.34189421e-01 6.79880917e-01
3.72890651e-01 -8.66308153e-01 5.05891621e-01 4.86685842e-01
6.44507825e-01 -6.82225883e-01 1.00425446e+00 1.71904817e-01
-9.76405799e-01 2.87451670e-02 -4.58104670e-01 1.82143837e-01
-1.23110987e-01 5.08936048e-01 -8.77430260e-01 1.75255954e+00
1.01850677e+00 3.63057047e-01 -9.93626118e-01 7.39113629e-01
-3.44331801e-01 5.81483006e-01 -2.52219409e-01 1.96433887e-01
8.21376890e-02 -9.53748226e-02 1.91040784e-01 9.14244294e-01
5.59233367e-01 1.34376973e-01 -1.81752250e-01 7.58272290e-01
-9.03600901e-02 2.85825580e-01 -5.75994134e-01 -2.20002443e-01
5.76444089e-01 1.17033052e+00 -3.62417161e-01 -4.07236814e-01
-5.64948380e-01 7.03812301e-01 3.30992520e-01 3.19843352e-01
-7.21731365e-01 -5.15757382e-01 3.65211487e-01 2.33852714e-01
1.08138733e-01 2.75406480e-01 1.53327197e-01 -1.12532449e+00
1.35209933e-01 -1.09713590e+00 1.22786331e+00 -1.07079673e+00
-1.73859847e+00 3.87410492e-01 4.01058257e-01 -8.31143022e-01
-1.38794914e-01 -6.40898883e-01 -1.04690874e-02 6.94396198e-01
-1.42898667e+00 -1.39850259e+00 -7.53503025e-01 9.77733314e-01
1.10693045e-01 9.28756148e-02 9.38545883e-01 5.52049100e-01
-2.14661822e-01 5.82112074e-01 -2.97386497e-01 -1.24900714e-01
8.76384318e-01 -1.40636432e+00 2.81584680e-01 4.97660398e-01
2.83166438e-01 6.34787738e-01 6.41421497e-01 -7.55490422e-01
-1.44093466e+00 -1.28199518e+00 7.13872612e-01 -8.12063634e-01
8.54749084e-01 -1.51211008e-01 -1.42629969e+00 9.05182302e-01
1.99227169e-01 -1.29722625e-01 9.25328434e-01 2.14297608e-01
-6.25191510e-01 -3.55395585e-01 -1.33567488e+00 4.53249127e-01
1.29953301e+00 -5.45257628e-01 -9.39920962e-01 4.31928545e-01
6.48601830e-01 -2.24014178e-01 -1.37068534e+00 6.88745320e-01
2.23361999e-01 -1.03259265e+00 6.22100234e-01 -8.29790056e-01
5.25295921e-02 -5.99903703e-01 -3.27891201e-01 -1.36649156e+00
-3.01402994e-02 -1.96821749e-01 -1.96724266e-01 1.42892003e+00
5.52530229e-01 -9.70089972e-01 8.31959605e-01 6.05347812e-01
-4.16891187e-01 -7.10309744e-01 -8.45557034e-01 -1.11152184e+00
1.97337359e-01 -5.18784642e-01 9.23914731e-01 1.33376586e+00
-4.66126800e-02 3.32675666e-01 2.86009729e-01 4.92638499e-01
3.84461224e-01 1.64122134e-01 9.66048956e-01 -1.09064722e+00
-2.01965556e-01 -3.46487075e-01 -6.08727455e-01 -5.07234156e-01
-2.51083113e-02 -1.06585693e+00 -5.58729708e-01 -2.00636315e+00
-4.79706638e-02 -6.03019059e-01 -3.54093969e-01 7.10840285e-01
-7.21452087e-02 -9.89279244e-03 -2.69988626e-01 -1.08204305e-01
-6.52223587e-01 4.88941252e-01 9.48464751e-01 5.46384044e-03
-5.64862564e-02 -2.96760976e-01 -8.36510837e-01 3.74138504e-01
1.30457449e+00 -2.28046045e-01 -9.99413192e-01 -2.87805617e-01
6.42302155e-01 -4.18104798e-01 5.32127261e-01 -8.77447128e-01
1.20834619e-01 -8.55785329e-03 -2.33421654e-01 -4.27923441e-01
6.29121810e-02 -7.98967898e-01 6.04674518e-01 2.35452682e-01
8.51362422e-02 -4.23864983e-02 4.41273361e-01 3.04772109e-01
-5.01543105e-01 -4.85337853e-01 3.37490857e-01 -3.39228988e-01
-1.13470936e+00 2.46906921e-01 2.78667063e-01 4.57229614e-01
1.09234631e+00 -1.50335118e-01 -9.87483084e-01 -2.91108251e-01
-6.29848123e-01 3.33516091e-01 4.58308369e-01 6.76603734e-01
4.62209344e-01 -1.31187844e+00 -8.10486078e-01 -2.14364767e-01
8.15845728e-01 7.11857826e-02 5.76676428e-01 5.82808733e-01
-6.98389471e-01 5.12375534e-01 -2.76709050e-01 -1.70238957e-01
-1.04506624e+00 6.92824364e-01 3.92491400e-01 -1.53004780e-01
-1.62276402e-01 6.90051675e-01 -3.60250995e-02 -8.43640685e-01
-1.97624370e-01 -3.26654732e-01 -1.07624151e-01 4.18604678e-03
1.22841448e-01 5.82433045e-01 5.08643448e-01 -2.65952140e-01
-3.96080136e-01 3.11165690e-01 -7.73318810e-03 2.26060838e-01
9.08627570e-01 1.83988977e-02 -2.83554494e-01 4.91539210e-01
8.30767453e-01 4.95890249e-03 -3.77242088e-01 -6.46857247e-02
3.36810559e-01 -4.22460735e-01 -3.94314677e-01 -1.09882581e+00
-9.52593863e-01 4.97760236e-01 3.76652300e-01 6.31415725e-01
1.16147673e+00 3.50137830e-01 5.13417661e-01 5.29167831e-01
7.23492682e-01 -1.14954209e+00 -2.01354042e-01 2.19477296e-01
1.06959212e+00 -8.98280919e-01 1.66810229e-02 -7.75476933e-01
-6.47956610e-01 8.69066775e-01 8.41245174e-01 3.71950775e-01
5.04971981e-01 -4.68599230e-01 2.11266741e-01 -7.19345927e-01
-6.52348280e-01 -3.32935721e-01 2.44652495e-01 8.09104502e-01
2.80064046e-01 2.17516735e-01 -4.32619482e-01 8.54558945e-01
-5.12041330e-01 1.31000295e-01 6.06791496e-01 8.47448409e-01
-5.55788353e-02 -1.27426577e+00 -4.31707613e-02 5.76614976e-01
-2.94584244e-01 6.30100258e-03 -7.10582614e-01 1.15779519e+00
2.64074743e-01 1.15431821e+00 -4.09699082e-01 -2.64322042e-01
7.86656916e-01 4.72262651e-01 6.40579522e-01 -6.66824341e-01
-5.05796790e-01 -1.08692241e+00 5.04460216e-01 -5.19105673e-01
-5.01453340e-01 -3.36720914e-01 -1.08230329e+00 -2.72103250e-01
-5.03734350e-01 5.30136466e-01 5.55706024e-01 4.78443295e-01
7.09927857e-01 2.50432283e-01 -4.32659723e-02 3.30889165e-01
-4.83855873e-01 -9.10403728e-01 -5.57744205e-01 8.34365129e-01
-4.37538356e-01 -7.24910140e-01 -3.14507067e-01 5.12471013e-02] | [10.101189613342285, 8.019866943359375] |
d3109c66-3be3-4e2e-acdd-91393fc84f23 | chemcrow-augmenting-large-language-models | 2304.05376 | null | https://arxiv.org/abs/2304.05376v4 | https://arxiv.org/pdf/2304.05376v4.pdf | ChemCrow: Augmenting large-language models with chemistry tools | Over the last decades, excellent computational chemistry tools have been developed. Their full potential has not yet been reached as most are challenging to learn and exist in isolation. Recently, large-language models (LLMs) have shown strong performance in tasks across domains, but struggle with chemistry-related problems. Moreover, these models lack access to external knowledge sources, limiting their usefulness in scientific applications. In this study, we introduce ChemCrow, an LLM chemistry agent designed to accomplish tasks across organic synthesis, drug discovery, and materials design. By integrating 17 expert-designed tools, ChemCrow augments the LLM performance in chemistry, and new capabilities emerge. Our agent autonomously planned the syntheses of an insect repellent, three organocatalysts, as well as other relevant molecules. Our evaluation, including both LLM and expert assessments, demonstrates ChemCrow's effectiveness in automating a diverse set of chemical tasks. Surprisingly, we find that GPT-4 as an evaluator cannot distinguish between clearly wrong GPT-4 completions and Chemcrow's performance. There is a significant risk of misuse of tools like ChemCrow, and we discuss their potential harms. Employed responsibly, our work not only aids expert chemists and lowers barriers for non-experts, but also fosters scientific advancement by bridging the gap between experimental and computational chemistry. A subset of the code is publicly available at \url{https://github.com/ur-whitelab/chemcrow-public}. | ['Philippe Schwaller', 'Andrew D White', 'Sam Cox', 'Andres M Bran'] | 2023-04-11 | null | null | null | null | ['drug-discovery'] | ['medical'] | [-2.30804831e-02 2.69930456e-02 -2.42571145e-01 -9.80472891e-04
-1.20480895e+00 -1.18728375e+00 5.82324505e-01 6.77208662e-01
-2.92779237e-01 1.01542258e+00 -6.08155578e-02 -1.00892377e+00
7.17621371e-02 -4.99929547e-01 -6.91778600e-01 -5.71950972e-01
1.85182378e-01 4.95169520e-01 1.30870892e-02 -7.11583868e-02
3.71820331e-01 7.56938875e-01 -1.05231369e+00 2.52540559e-01
1.24070776e+00 2.92434663e-01 4.59066927e-01 3.99995148e-01
-2.06513867e-01 8.06001365e-01 -6.07684672e-01 -5.12614250e-01
-8.19864795e-02 -1.82384476e-01 -8.74003291e-01 -7.72139788e-01
2.65793264e-01 3.00297081e-01 1.86930120e-01 7.83630252e-01
8.81539464e-01 -5.65490685e-03 6.23409748e-01 -9.70881879e-01
-7.21965432e-01 4.62498099e-01 -1.61484271e-01 -1.92494303e-01
5.94769895e-01 7.00086951e-01 9.61503208e-01 -1.21237433e+00
7.28462517e-01 1.31958175e+00 6.02956057e-01 5.87971509e-01
-1.25565076e+00 -1.06033504e+00 -7.81490132e-02 -4.85857278e-02
-1.45576477e+00 -6.90858781e-01 1.20613456e-01 -8.86281013e-01
1.31092811e+00 8.25077295e-02 2.74580896e-01 1.22512734e+00
3.01336765e-01 2.44653150e-01 1.00273573e+00 -1.15379557e-01
6.14309371e-01 1.45091832e-01 -2.31303319e-01 6.19379282e-01
5.08726418e-01 -4.62995581e-02 -6.64076805e-01 -3.59138131e-01
3.31739098e-01 -1.47870630e-01 -1.63393408e-01 -6.68655038e-02
-1.06280553e+00 6.65201783e-01 9.60446224e-02 3.95498842e-01
-1.95464879e-01 2.67307200e-02 5.28763719e-02 1.30918428e-01
2.95036942e-01 1.32386947e+00 -7.49424040e-01 -1.59661263e-01
-5.65933168e-01 2.65848339e-01 1.09344900e+00 1.01295269e+00
5.44864476e-01 -1.72749937e-01 1.62986800e-01 4.54156041e-01
2.67761230e-01 3.58410209e-01 4.46785474e-03 -8.65308821e-01
3.22708666e-01 5.70921004e-01 4.02420968e-01 -6.24885976e-01
-5.85879505e-01 -2.88437963e-01 -3.93345863e-01 3.99632424e-01
5.91590106e-01 -2.49803364e-01 -6.73933625e-01 1.62188876e+00
3.38370323e-01 -4.87342209e-01 2.18635038e-01 5.16151488e-01
1.01436889e+00 7.78358042e-01 9.13346887e-01 -7.39650875e-02
1.17747915e+00 -7.03078628e-01 -4.10029292e-01 -1.27599671e-01
8.33121002e-01 -1.16843128e+00 9.73486125e-01 8.26726139e-01
-1.20299292e+00 -3.79406840e-01 -1.14000499e+00 -1.41526744e-01
-1.14284182e+00 -8.48116055e-02 9.27558184e-01 6.31417871e-01
-8.64930570e-01 1.06450224e+00 -6.86545193e-01 -2.41318688e-01
5.32744765e-01 5.29790044e-01 -4.91104513e-01 -2.85889536e-01
-1.05641782e+00 1.21096432e+00 4.13107216e-01 -1.72894791e-01
-1.24352455e+00 -1.02876484e+00 -7.53784597e-01 -1.43366903e-01
6.23859167e-01 -5.81010759e-01 1.35286868e+00 -2.16599047e-01
-1.22822952e+00 5.09073973e-01 -7.95384720e-02 1.37506604e-01
4.25294846e-01 3.80304642e-04 -6.05099380e-01 -2.52602734e-02
3.49261671e-01 7.37781167e-01 1.72885805e-01 -9.62339520e-01
-2.74717152e-01 -1.94747955e-01 8.54807645e-02 6.88619539e-02
-9.35311764e-02 1.91474065e-01 -2.17292026e-01 -2.15656817e-01
-5.42851090e-01 -6.56923950e-01 -5.32557130e-01 -6.49722889e-02
-5.04985929e-01 -3.43387783e-01 2.35245943e-01 -2.79378206e-01
1.10782874e+00 -1.70315540e+00 1.25789354e-02 1.08002841e-01
3.90143186e-01 3.20345283e-01 -2.64383912e-01 1.05410755e+00
-2.98657775e-01 7.77379990e-01 -2.20506668e-01 1.21241221e-02
2.17062086e-02 -2.37357125e-01 5.21402434e-02 3.19071293e-01
3.10693890e-01 8.45487535e-01 -1.25598407e+00 -2.55912274e-01
1.23238668e-01 5.03321648e-01 -3.34443420e-01 -8.50700587e-02
-8.71586204e-01 6.17538333e-01 -5.79102933e-01 9.98009562e-01
4.30142999e-01 -6.95192039e-01 4.69189793e-01 6.26517385e-02
-7.10854411e-01 5.12035787e-01 -9.40990269e-01 1.88895369e+00
-2.10202605e-01 4.34414716e-03 -2.32938509e-02 -2.98870832e-01
6.31006837e-01 4.80282187e-01 3.54995072e-01 -5.85855901e-01
-2.12533593e-01 5.14618874e-01 2.48702109e-01 -4.21407253e-01
4.14924212e-02 -4.60179150e-02 2.76419491e-01 3.79602313e-01
-6.58878535e-02 -3.39410096e-01 5.91414392e-01 2.82986432e-01
1.18418074e+00 4.13522631e-01 4.93748426e-01 -3.78520131e-01
4.48142141e-01 5.58177769e-01 3.51012409e-01 6.43257737e-01
2.19997272e-01 4.52322178e-02 3.66243660e-01 -3.66139829e-01
-1.04638445e+00 -7.83328354e-01 -2.35413238e-01 1.21844149e+00
-2.20197067e-01 -1.09849572e+00 -5.32272637e-01 -4.11455631e-01
2.78820768e-02 6.71067119e-01 -3.07185382e-01 3.17983665e-02
-5.62205128e-02 -7.59287715e-01 5.09926975e-01 1.62896439e-01
-1.62883073e-01 -9.99352694e-01 -9.21613276e-02 5.03690243e-01
2.19787791e-01 -9.22254443e-01 -3.79661471e-01 4.17524815e-01
-4.05519098e-01 -1.41577697e+00 -5.09491563e-01 -4.84228522e-01
3.49485546e-01 2.26343915e-01 1.03851855e+00 -1.95467383e-01
-5.66454232e-01 7.50186965e-02 5.82595244e-02 -8.47509563e-01
-8.89299691e-01 1.78967178e-01 -1.55484099e-02 -8.36377382e-01
3.22292000e-01 -6.12193406e-01 -7.29342043e-01 2.55893290e-01
-7.53289461e-01 -6.47662804e-02 7.01284051e-01 2.04998046e-01
4.79004800e-01 -2.48828575e-01 7.63681710e-01 -9.31864738e-01
7.64306605e-01 -5.78735352e-01 -7.98617303e-01 3.36658865e-01
-9.91705954e-01 1.63611293e-01 6.97771013e-01 -2.89682657e-01
-6.84458792e-01 6.16883263e-02 -2.64134616e-01 1.25394732e-01
-2.23497406e-01 8.60871732e-01 -4.14511323e-01 -1.98120296e-01
8.89171243e-01 -2.43034139e-01 -1.19367160e-01 -7.42435873e-01
2.26498306e-01 4.43902373e-01 9.13538877e-03 -1.04215157e+00
6.58260465e-01 -7.08187968e-02 8.08031186e-02 -8.16364706e-01
-6.84624612e-01 -3.57929796e-01 -8.27102289e-02 1.19445577e-01
8.30444455e-01 -1.15722537e+00 -1.27539158e+00 -1.10005267e-01
-1.32961988e+00 -5.28100193e-01 9.03761461e-02 3.63877565e-01
-2.16454156e-02 4.17151451e-01 -4.76624578e-01 -5.80396116e-01
-4.17714983e-01 -1.56401610e+00 9.32687581e-01 1.98193435e-02
-6.13363743e-01 -8.85553181e-01 6.67537078e-02 4.57294315e-01
1.87411308e-01 4.13255870e-01 1.24090898e+00 -7.97937572e-01
-7.55789638e-01 7.79480767e-03 -4.45738845e-02 8.15145858e-03
2.20769837e-01 1.41043350e-01 -1.07830989e+00 -1.50059342e-01
-6.89031303e-01 -5.39153934e-01 4.60738689e-01 1.43303964e-02
1.02377069e+00 -2.29212597e-01 -4.48715061e-01 1.93322703e-01
1.17231166e+00 7.46419549e-01 3.96079361e-01 2.89658695e-01
7.44498074e-01 5.83311677e-01 3.91828954e-01 4.09437448e-01
1.20762020e-01 2.74304688e-01 2.65535861e-01 -2.56854862e-01
1.78145885e-01 -4.53921676e-01 4.59251076e-01 4.20036435e-01
-2.86852807e-01 -4.20919389e-01 -1.14302135e+00 -9.17255506e-03
-1.50622082e+00 -6.71930075e-01 -1.99680746e-01 2.16814089e+00
1.32055902e+00 1.28445417e-01 5.03390543e-02 -2.67305523e-01
3.16472322e-01 -2.69555837e-01 -9.65499163e-01 -3.32047701e-01
-6.81675375e-02 7.08347142e-01 4.74376500e-01 6.83534682e-01
-8.07358265e-01 1.24426055e+00 6.59798908e+00 1.15514016e+00
-9.43725884e-01 -6.60237223e-02 5.51411927e-01 1.23044394e-01
-4.79801416e-01 2.91267008e-01 -8.71001303e-01 3.24526399e-01
1.04466307e+00 -2.91869044e-01 3.95519465e-01 8.58203471e-01
4.95884091e-01 7.66970813e-02 -1.49117780e+00 7.33350456e-01
-4.72956777e-01 -1.83536625e+00 1.70943618e-01 2.11204007e-01
4.70761538e-01 2.44656280e-02 3.38905007e-02 1.62070096e-01
7.55312800e-01 -1.55081499e+00 5.25182247e-01 2.32852131e-01
9.69242156e-01 -6.34251475e-01 1.11829378e-01 2.43288755e-01
-8.51534426e-01 2.65208036e-01 -2.25073293e-01 -6.27298728e-02
-2.10192591e-01 4.93257999e-01 -1.07263958e+00 5.15990555e-01
2.28861734e-01 6.83050275e-01 -5.63577771e-01 8.26327503e-01
-2.27939695e-01 2.78268278e-01 -1.00754388e-01 -2.08408788e-01
3.27854723e-01 -2.85430193e-01 2.08088011e-01 1.34658575e+00
1.12700537e-01 2.32595295e-01 4.76021290e-01 1.19893610e+00
-1.86457708e-01 3.13433647e-01 -6.83577657e-01 -7.76992738e-01
7.10281551e-01 1.26147437e+00 -4.87942755e-01 -2.11930305e-01
-3.02862674e-01 3.96721303e-01 9.38813314e-02 4.90805924e-01
-5.87778509e-01 -3.41697514e-01 7.73414016e-01 -7.44008087e-03
-1.46667406e-01 -3.13439518e-01 -1.24618948e-01 -7.33648181e-01
-2.90595651e-01 -1.35114503e+00 2.17003390e-01 -9.23617542e-01
-1.05796719e+00 4.16141748e-01 -2.25715101e-01 -8.01988542e-01
2.10916460e-01 -8.57558966e-01 -2.96585560e-01 9.92360651e-01
-1.29280066e+00 -9.13780212e-01 5.25949672e-02 2.70811081e-01
3.41915846e-01 -1.70348510e-01 1.12191856e+00 4.38616216e-01
-6.45562053e-01 2.74040163e-01 1.52833179e-01 -4.57809091e-01
1.17771900e+00 -1.13934815e+00 4.33508039e-01 3.60486329e-01
-2.38990784e-01 1.22950184e+00 8.43894362e-01 -8.38407755e-01
-1.62440419e+00 -1.05929315e+00 8.94067466e-01 -8.97877812e-01
8.31457555e-01 -5.62028706e-01 -7.03489959e-01 2.31955469e-01
3.31244282e-02 -5.42844951e-01 1.10945356e+00 1.34200707e-01
-7.36322105e-01 3.00034046e-01 -9.43295002e-01 8.50438595e-01
1.02663720e+00 -5.16924322e-01 3.48398909e-02 1.06746650e+00
9.52735662e-01 -3.06293964e-01 -1.30570984e+00 1.29145920e-01
5.37267685e-01 -5.26097476e-01 1.11036849e+00 -8.83739114e-01
5.78529060e-01 -3.74005377e-01 9.43497717e-02 -9.65042353e-01
-2.36239046e-01 -1.04072106e+00 1.79633990e-01 9.29355443e-01
1.09806633e+00 -7.60523140e-01 5.48753679e-01 9.19270515e-01
-4.14212286e-01 -7.31911659e-01 -3.35236192e-01 -8.91362071e-01
4.19527620e-01 -3.04875165e-01 4.94756222e-01 1.04145443e+00
3.76237154e-01 8.73240769e-01 -4.28690538e-02 4.72317599e-02
4.01724279e-01 -8.20553675e-03 5.82061052e-01 -1.34920037e+00
-4.24868673e-01 -5.81197739e-01 2.21429273e-01 -5.51635027e-01
-6.86390847e-02 -1.11573744e+00 -4.03188258e-01 -1.56425893e+00
2.05868691e-01 -4.59130406e-01 -2.29249354e-02 9.07984674e-01
-2.77303401e-02 -2.99305499e-01 5.23679927e-02 2.26931080e-01
-7.23745823e-01 1.02967471e-01 1.42039037e+00 -4.17204469e-01
-3.14787030e-01 -2.88617998e-01 -1.36646736e+00 5.34873307e-01
9.33755636e-01 -5.36138177e-01 -3.34891737e-01 -7.15845153e-02
5.54228663e-01 -1.32166699e-01 1.19752854e-01 -9.30531442e-01
3.22330475e-01 -4.77658659e-01 4.57490087e-01 -2.88761944e-01
3.33400160e-01 -4.48153704e-01 5.86354494e-01 4.82115507e-01
-2.90350854e-01 -1.96897276e-02 4.74175602e-01 3.46075237e-01
2.11302802e-01 -1.44006699e-01 5.33504307e-01 -6.96940005e-01
-1.76166281e-01 5.29968619e-01 -5.97838044e-01 -1.38488216e-02
7.73022473e-01 -7.74383843e-02 -7.12308407e-01 -1.95376948e-02
-7.03804910e-01 4.39275891e-01 6.29883826e-01 1.94577277e-01
3.27767193e-01 -6.67813122e-01 -3.73229474e-01 -3.84975702e-01
4.15907681e-01 1.24171406e-01 -2.88639264e-03 7.23687053e-01
-5.42847276e-01 8.25677991e-01 2.16383263e-01 -1.07637770e-01
-1.16207039e+00 7.87197173e-01 1.97618127e-01 -1.57085598e-01
-7.75771290e-02 6.02378964e-01 3.63706082e-01 -3.01850170e-01
5.00642717e-01 -2.49826670e-01 4.09754291e-02 -6.12468878e-03
6.35155022e-01 2.48545498e-01 2.76567817e-01 4.38975841e-02
-4.14199740e-01 3.05688709e-01 -4.30016518e-01 6.25440478e-02
1.46689939e+00 5.26146948e-01 4.27644439e-02 1.79036677e-01
9.04444635e-01 9.69419405e-02 -8.46325338e-01 1.00115530e-01
2.35926703e-01 1.40491694e-01 -2.73895502e-01 -1.31726682e+00
-2.72336602e-01 9.53499675e-01 1.59640312e-01 -1.50384471e-01
3.53689432e-01 5.80231249e-02 5.01998782e-01 8.34010243e-01
2.78423816e-01 -9.99341667e-01 1.29034117e-01 4.50735629e-01
8.17106426e-01 -1.31046867e+00 3.45785797e-01 -4.18692976e-01
-4.58215058e-01 9.63192344e-01 5.05621970e-01 7.38036394e-01
2.02427164e-01 2.39121959e-01 -1.82070121e-01 -6.18450880e-01
-9.03806448e-01 -2.17817232e-01 -1.66176647e-01 5.78757167e-01
9.38852251e-01 -4.49761003e-02 -2.68052816e-01 4.63875592e-01
5.06820567e-02 3.53937186e-02 3.50974977e-01 1.24597156e+00
-4.93632197e-01 -1.63174570e+00 -3.55340272e-01 -1.30130291e-01
-6.25297189e-01 -6.48461878e-01 -1.09155512e+00 7.43106902e-01
2.96934158e-01 1.18541539e+00 -7.59790540e-01 -1.60925582e-01
3.14298749e-01 2.36177012e-01 5.72704554e-01 -8.54594707e-01
-6.05041921e-01 1.73178732e-01 4.40569133e-01 -3.72853965e-01
-2.15163216e-01 -3.83745939e-01 -1.23399222e+00 -5.37904799e-01
-2.89749224e-02 7.06490636e-01 8.68380725e-01 6.69363260e-01
7.18346536e-01 3.81740183e-01 2.03388348e-01 -5.54832995e-01
-4.20342445e-01 -6.96545064e-01 -1.93721935e-01 -1.48101360e-01
1.01209007e-01 -5.19606888e-01 -1.48368761e-01 3.65919918e-02] | [4.847652912139893, 5.815973281860352] |
ce36dde6-1111-4e9e-b4f3-2429bd069bb8 | attentive-continuous-generative-self-training | 2305.14589 | null | https://arxiv.org/abs/2305.14589v1 | https://arxiv.org/pdf/2305.14589v1.pdf | Attentive Continuous Generative Self-training for Unsupervised Domain Adaptive Medical Image Translation | Self-training is an important class of unsupervised domain adaptation (UDA) approaches that are used to mitigate the problem of domain shift, when applying knowledge learned from a labeled source domain to unlabeled and heterogeneous target domains. While self-training-based UDA has shown considerable promise on discriminative tasks, including classification and segmentation, through reliable pseudo-label filtering based on the maximum softmax probability, there is a paucity of prior work on self-training-based UDA for generative tasks, including image modality translation. To fill this gap, in this work, we seek to develop a generative self-training (GST) framework for domain adaptive image translation with continuous value prediction and regression objectives. Specifically, we quantify both aleatoric and epistemic uncertainties within our GST using variational Bayes learning to measure the reliability of synthesized data. We also introduce a self-attention scheme that de-emphasizes the background region to prevent it from dominating the training process. The adaptation is then carried out by an alternating optimization scheme with target domain supervision that focuses attention on the regions with reliable pseudo-labels. We evaluated our framework on two cross-scanner/center, inter-subject translation tasks, including tagged-to-cine magnetic resonance (MR) image translation and T1-weighted MR-to-fractional anisotropy translation. Extensive validations with unpaired target domain data showed that our GST yielded superior synthesis performance in comparison to adversarial training UDA methods. | ['Jonghye Woo', 'Georges El Fakhri', 'Maureen Stone', 'Reese Timothy', 'Jiachen Zhuo', 'Fangxu Xing', 'Jerry L. Prince', 'Xiaofeng Liu'] | 2023-05-23 | null | null | null | null | ['value-prediction', 'unsupervised-domain-adaptation', 'pseudo-label'] | ['computer-code', 'methodology', 'miscellaneous'] | [ 8.28465164e-01 5.93568802e-01 -3.75963479e-01 -7.44961023e-01
-1.40636778e+00 -4.36484933e-01 6.11219943e-01 -3.28407943e-01
-2.58223683e-01 1.03383982e+00 2.03649089e-01 -1.96192399e-01
-1.00210207e-02 -5.57292759e-01 -9.97769892e-01 -9.32671010e-01
4.39760625e-01 8.06082785e-01 6.37218636e-03 7.70924240e-02
-1.02889419e-01 1.32761672e-01 -8.00652564e-01 3.20254534e-01
1.33635068e+00 7.64150679e-01 -2.82484898e-03 1.62278667e-01
2.12145105e-01 6.03912055e-01 -4.68236238e-01 -3.15626830e-01
1.01258166e-01 -7.94909656e-01 -9.92121816e-01 4.45243001e-01
3.75454277e-01 -2.53072321e-01 -4.12893854e-02 1.08015990e+00
8.47926319e-01 2.47843087e-01 1.02783024e+00 -9.40034688e-01
-9.51487482e-01 7.40796506e-01 -4.83022332e-01 3.32768977e-01
-1.00522563e-01 2.22791940e-01 6.37521207e-01 -7.89789915e-01
9.26248491e-01 1.05502808e+00 5.38649440e-01 8.99847448e-01
-1.73689020e+00 -6.41448557e-01 -9.40417051e-02 -1.92197293e-01
-9.99980867e-01 -4.05689240e-01 9.35818672e-01 -7.69017041e-01
6.06375992e-01 -8.54359642e-02 1.06595695e-01 1.55181050e+00
2.89076000e-01 6.50103629e-01 1.70993114e+00 -5.27784705e-01
4.91208822e-01 2.32920304e-01 -2.13554069e-01 4.21998888e-01
-2.58095145e-01 3.99127394e-01 -3.35976511e-01 -1.65632501e-01
8.63345981e-01 -4.53935355e-01 3.36919054e-02 -6.06732726e-01
-1.13943064e+00 9.61217880e-01 3.64349246e-01 3.32600176e-01
-4.49754775e-01 -1.46450460e-01 3.03623468e-01 1.69741452e-01
1.02864075e+00 4.53766763e-01 -3.47532272e-01 3.54393750e-01
-1.21881986e+00 6.38646409e-02 2.35345215e-01 8.17586541e-01
5.67898870e-01 4.74460125e-01 -4.48656768e-01 1.03946471e+00
3.05783570e-01 6.75992489e-01 7.13018000e-01 -9.18163359e-01
2.23918855e-01 2.49683693e-01 -2.32112840e-01 -3.66104066e-01
-1.89622760e-01 -6.10413551e-01 -7.06743300e-01 3.10780317e-01
2.96641797e-01 -3.20684552e-01 -1.35381281e+00 2.03601360e+00
5.15509129e-01 1.99670509e-01 7.52759576e-02 1.08603489e+00
5.66222727e-01 3.09554428e-01 4.32649493e-01 -4.45523530e-01
1.03313267e+00 -7.31446326e-01 -6.88673675e-01 -3.32951099e-01
4.71811742e-01 -5.76438963e-01 1.09372878e+00 -8.66583735e-03
-1.12263465e+00 -5.70750833e-01 -9.10875380e-01 9.44375098e-02
5.01409285e-02 -1.97866887e-01 2.16714442e-01 7.90161848e-01
-8.30035567e-01 5.37795544e-01 -1.05031788e+00 -1.44265503e-01
7.85281003e-01 4.32467371e-01 -2.42017999e-01 -3.70478928e-02
-1.38286877e+00 1.03333759e+00 2.45393574e-01 -3.75821143e-01
-1.16317713e+00 -9.97228086e-01 -9.18381095e-01 -4.46669459e-01
3.68290722e-01 -8.20245743e-01 1.16512740e+00 -1.28141248e+00
-1.69258392e+00 1.23938406e+00 2.46558897e-02 -6.69003010e-01
6.27547264e-01 5.53366914e-02 -4.17934835e-01 2.35828489e-01
4.95148778e-01 9.04113352e-01 1.36813939e+00 -1.29995191e+00
-3.18556353e-02 -5.63938498e-01 -4.51698124e-01 2.82934994e-01
1.33989468e-01 -1.96910445e-02 7.64729753e-02 -1.12503660e+00
1.10579610e-01 -1.14489675e+00 -1.61733910e-01 -3.56134862e-01
-3.58059496e-01 -3.29943979e-03 6.07688367e-01 -8.42135549e-01
7.63386130e-01 -2.11002493e+00 3.63194525e-01 2.44658038e-01
1.08261064e-01 4.69009504e-02 8.08806866e-02 -3.35315198e-01
-2.51855284e-01 -1.11767016e-01 -9.41130221e-01 -2.78027087e-01
-5.89236878e-02 3.70640546e-01 -4.40505564e-01 5.33996046e-01
4.69962299e-01 1.10219502e+00 -8.49040627e-01 -8.35972786e-01
1.34012520e-01 4.19810504e-01 -6.47408843e-01 1.50820479e-01
-4.73286659e-01 1.33363521e+00 -5.24782360e-01 6.98925376e-01
5.74051797e-01 -2.34254152e-01 2.31157899e-01 -1.54325739e-01
3.19915414e-01 1.91620942e-02 -6.02296412e-01 1.90139508e+00
-2.85981566e-01 2.78491020e-01 5.39578358e-03 -1.26960862e+00
8.23809803e-01 4.12713051e-01 8.60940039e-01 -8.77459884e-01
3.09299707e-01 3.64877582e-01 -4.42173705e-02 -4.19009000e-01
8.61447677e-02 -7.77902007e-01 -1.72007456e-01 4.87783343e-01
4.77926642e-01 -4.57463056e-01 -1.67521566e-01 7.95596093e-02
6.76200390e-01 5.88170290e-01 -1.94373522e-02 -3.48149180e-01
3.46592695e-01 1.48171932e-01 5.93114793e-01 5.95525026e-01
-3.32168311e-01 8.78825843e-01 2.44345635e-01 1.12424925e-01
-1.25148928e+00 -1.26445580e+00 -5.05594790e-01 9.80642259e-01
-1.99783191e-01 4.58241433e-01 -1.13037813e+00 -9.88115788e-01
-1.43307164e-01 9.93294239e-01 -7.03086972e-01 -4.01250899e-01
-5.47846675e-01 -9.59801435e-01 4.64311182e-01 5.59310198e-01
3.48265678e-01 -1.17301464e+00 -2.28327900e-01 3.50687444e-01
-3.30125570e-01 -1.12021244e+00 -6.62649930e-01 3.05345237e-01
-1.18285406e+00 -6.98037565e-01 -1.12649143e+00 -6.69956446e-01
8.10127974e-01 -4.83894229e-01 1.09348392e+00 -7.92670190e-01
3.23372968e-02 5.82212627e-01 -1.43052742e-01 -1.80719718e-01
-8.76183808e-01 8.32640454e-02 1.38210669e-01 8.92821997e-02
6.03792742e-02 -6.13732755e-01 -4.42021549e-01 4.41667914e-01
-1.03689909e+00 -6.18574545e-02 6.22187197e-01 1.19726288e+00
1.14766812e+00 -3.32126528e-01 7.98752010e-01 -1.16965055e+00
5.88945210e-01 -6.01964355e-01 -3.74142438e-01 2.70273536e-01
-9.49514866e-01 2.65544325e-01 2.79877186e-01 -7.73806751e-01
-1.34143364e+00 1.00239225e-01 1.02437194e-02 -5.85311651e-01
-1.36079550e-01 3.92725140e-01 -1.55974090e-01 -1.52987972e-01
1.11201560e+00 2.93659925e-01 1.62296727e-01 -1.77705005e-01
5.42927027e-01 5.05553246e-01 8.23030710e-01 -9.96526718e-01
5.71058452e-01 4.85497445e-01 -5.52581176e-02 -4.09714907e-01
-7.90198028e-01 -5.60690556e-03 -7.70979881e-01 -2.01697320e-01
1.06129587e+00 -7.93772817e-01 8.29067454e-02 4.48198855e-01
-6.70987189e-01 -6.69649541e-01 -6.40786052e-01 5.73145866e-01
-1.00452518e+00 2.93806970e-01 -3.67406219e-01 -4.41875368e-01
-4.35869813e-01 -1.42088163e+00 1.06910026e+00 3.53674218e-02
-2.94153959e-01 -1.19333076e+00 1.01076588e-01 8.10031414e-01
3.95213217e-01 5.28843880e-01 9.11749780e-01 -7.10592568e-01
-3.03973109e-01 2.00305268e-01 1.32476300e-01 6.65194690e-01
1.02121294e-01 -6.06661379e-01 -9.66520429e-01 -2.82784611e-01
2.08875209e-01 -4.69419062e-01 6.52511835e-01 8.78944159e-01
8.85523915e-01 -1.31336385e-02 -1.70975685e-01 5.22467375e-01
1.02458060e+00 1.62346482e-01 5.97290277e-01 1.64856464e-01
5.45643985e-01 4.90454704e-01 5.90950966e-01 1.19735532e-01
2.10333392e-01 7.15446472e-01 3.61878239e-02 -2.93174088e-01
-5.43844938e-01 -3.34537625e-01 2.83126473e-01 3.42830628e-01
1.49202392e-01 7.61439949e-02 -8.78188312e-01 5.81258655e-01
-1.52548528e+00 -6.36155009e-01 1.08027704e-01 2.02738261e+00
1.22091150e+00 1.18455626e-01 1.19298168e-01 -2.04245940e-01
8.01648915e-01 -1.07985780e-01 -9.98000085e-01 -2.81875916e-02
-2.53423125e-01 4.85082060e-01 4.70512480e-01 5.70134461e-01
-1.20943034e+00 9.97885883e-01 6.46213818e+00 8.76245618e-01
-1.21884465e+00 7.32181132e-01 8.76498222e-01 1.77692860e-01
-4.46386576e-01 -1.22578293e-01 -3.23471129e-01 5.08994699e-01
9.83096004e-01 8.82656053e-02 3.26712638e-01 7.68922031e-01
1.36003688e-01 -1.19661659e-01 -1.01812875e+00 5.38472235e-01
1.13770872e-01 -1.10866785e+00 -3.15911323e-01 -1.92707703e-02
1.18751001e+00 1.16265394e-01 4.35397804e-01 3.37013483e-01
5.12771666e-01 -8.15479696e-01 6.79092646e-01 2.85913557e-01
1.27102530e+00 -3.83666098e-01 3.35711449e-01 1.02071323e-01
-3.90398145e-01 2.90892839e-01 4.32795770e-02 7.11614132e-01
2.59750873e-01 5.97524107e-01 -9.98259962e-01 4.75987971e-01
4.30919528e-01 3.85257721e-01 -1.99524581e-01 4.00895476e-01
-3.99921000e-01 8.03204715e-01 -5.12342565e-02 6.72418654e-01
1.11110769e-01 -1.93760544e-01 7.30765402e-01 8.74305129e-01
5.90524562e-02 1.71793282e-01 3.12989578e-02 1.32780921e+00
-1.14036135e-01 -5.03150113e-02 -3.86875600e-01 3.78595926e-02
8.01831707e-02 8.01581323e-01 -7.46381819e-01 -3.78330141e-01
-1.15857184e-01 1.14742208e+00 1.67536419e-02 5.03161073e-01
-1.07721734e+00 1.82583824e-01 1.02617912e-01 3.52076381e-01
9.57598835e-02 -6.17699511e-02 -6.81204021e-01 -1.13517821e+00
-2.11087212e-01 -8.62190783e-01 4.76801544e-01 -8.62040579e-01
-1.44461107e+00 6.28399014e-01 3.70433509e-01 -1.02152264e+00
-4.46777970e-01 -2.30398387e-01 -2.06495196e-01 1.02321565e+00
-1.39186418e+00 -1.44960904e+00 1.37108251e-01 6.89631641e-01
4.86857235e-01 -3.15409601e-01 6.66414440e-01 4.36643541e-01
-3.45879793e-01 7.12117612e-01 3.00412476e-01 -2.51945714e-03
1.11324072e+00 -1.19689059e+00 7.15366453e-02 8.00225079e-01
-1.33923337e-01 3.94580632e-01 6.65607512e-01 -1.04757059e+00
-6.93128943e-01 -1.15783501e+00 5.20470619e-01 -5.86084366e-01
4.94943768e-01 -1.18319362e-01 -1.04731524e+00 8.01131725e-01
3.50208618e-02 2.48163074e-01 5.89788258e-01 -2.62304693e-01
-1.79360688e-01 1.63883969e-01 -1.63312531e+00 2.11441711e-01
8.17106962e-01 -5.81546605e-01 -7.99301624e-01 4.69561100e-01
6.47450089e-01 -7.33104646e-01 -1.16988206e+00 5.18450141e-01
2.33064279e-01 -5.16508460e-01 1.00205278e+00 -5.73527694e-01
5.87850034e-01 -8.04467276e-02 -7.71166803e-03 -1.52638865e+00
-2.64504731e-01 -4.84639615e-01 -2.56251749e-02 1.20669949e+00
3.66036594e-01 -6.29382670e-01 8.55953574e-01 8.43962371e-01
-4.34790432e-01 -3.45732301e-01 -1.06317520e+00 -6.80589914e-01
5.52196562e-01 -2.40709737e-01 2.73719370e-01 1.10986710e+00
-3.37057471e-01 2.29896516e-01 -4.19965208e-01 1.48143917e-01
1.04181099e+00 -8.10848847e-02 3.39165509e-01 -9.09442961e-01
-3.39657575e-01 -2.35591665e-01 -5.46201691e-03 -5.13909698e-01
5.77279687e-01 -1.23971689e+00 1.48152262e-01 -1.21400702e+00
1.37911603e-01 -4.40116167e-01 -2.81730741e-01 5.19346952e-01
-1.23163030e-01 2.92744815e-01 -2.68579245e-01 4.35555786e-01
-2.45680332e-01 5.98301768e-01 1.62108672e+00 -3.46346557e-01
-1.94345549e-01 7.96740502e-02 -6.07920766e-01 3.20099205e-01
5.59014499e-01 -8.41444433e-01 -5.72120368e-01 -3.08437437e-01
-3.83703798e-01 2.56612092e-01 4.18345422e-01 -7.19698787e-01
-1.56967729e-01 -2.88715661e-01 3.47197264e-01 -1.84115574e-01
-1.16472167e-03 -6.17347538e-01 2.32083872e-01 3.50247651e-01
-5.71465969e-01 -6.08837724e-01 1.06677771e-01 4.89460915e-01
-1.08533353e-01 -1.00242399e-01 1.32393563e+00 -1.08212709e-01
-5.03796935e-01 2.11310789e-01 -2.08574474e-01 4.32739615e-01
9.74393547e-01 -1.00286238e-01 3.86177748e-02 -1.94514334e-01
-1.30483556e+00 -1.10418228e-02 3.21607023e-01 2.78066933e-01
4.78002250e-01 -1.35847867e+00 -8.58852983e-01 2.99953759e-01
3.00196167e-02 8.32482129e-02 5.39713621e-01 1.07998431e+00
2.79614162e-02 2.43701622e-01 -3.07811290e-01 -8.25319946e-01
-7.96680808e-01 5.39413095e-01 5.25700152e-01 -2.91247994e-01
-5.55528700e-01 7.79147983e-01 3.54873538e-01 -6.49577498e-01
-1.04012504e-01 -1.75584584e-01 1.29518509e-01 -1.81611627e-01
-1.74530037e-02 2.17006430e-01 1.73096612e-01 -8.48884583e-01
-2.42160171e-01 4.55344051e-01 3.73048335e-02 -4.97673154e-01
1.18527484e+00 -2.00449720e-01 1.84959710e-01 1.71311796e-01
9.55381513e-01 -3.14323455e-01 -1.59743953e+00 -5.49739957e-01
-1.08454809e-01 -1.63448557e-01 3.20815027e-01 -1.24315965e+00
-1.02316463e+00 7.62606859e-01 1.01779127e+00 -3.07845592e-01
1.04979110e+00 1.96183413e-01 6.95022106e-01 -3.91665906e-01
2.40551621e-01 -1.27643776e+00 2.51003504e-01 1.90316483e-01
8.76747489e-01 -1.52042890e+00 -1.36653379e-01 -2.53344834e-01
-1.16911614e+00 5.84761977e-01 5.23996890e-01 -3.72133218e-02
4.71787512e-01 2.07693204e-01 4.37457077e-02 -6.39614686e-02
-1.41672984e-01 -2.03320738e-02 5.95670879e-01 8.82363796e-01
2.75040686e-01 1.58262998e-01 -2.62237459e-01 6.83331966e-01
1.53666183e-01 3.32915217e-01 6.20975047e-02 8.79788280e-01
-6.66693822e-02 -1.23885119e+00 -5.93263209e-01 2.77437389e-01
-6.18931472e-01 -8.58120546e-02 -1.52991340e-01 4.83658433e-01
2.11634383e-01 5.95933676e-01 -1.60858333e-01 -6.92778304e-02
1.85302824e-01 4.68541682e-01 7.62528718e-01 -6.92682028e-01
-4.59598988e-01 4.45978582e-01 -1.98708251e-01 -2.41530731e-01
-6.24872804e-01 -8.37977946e-01 -1.33615398e+00 1.51766688e-01
-2.40963042e-01 -1.08616605e-01 6.30341411e-01 1.04382193e+00
3.09429854e-01 7.68457711e-01 5.74023962e-01 -6.22056842e-01
-7.52788186e-01 -1.06926489e+00 -5.04274607e-01 6.61317647e-01
9.12270769e-02 -8.84084344e-01 -1.95515990e-01 3.82953316e-01] | [14.573246002197266, -1.9723126888275146] |
098d2457-7880-432c-8572-6e3c6aff5633 | onsets-and-frames-dual-objective-piano | 1710.11153 | null | http://arxiv.org/abs/1710.11153v2 | http://arxiv.org/pdf/1710.11153v2.pdf | Onsets and Frames: Dual-Objective Piano Transcription | We advance the state of the art in polyphonic piano music transcription by
using a deep convolutional and recurrent neural network which is trained to
jointly predict onsets and frames. Our model predicts pitch onset events and
then uses those predictions to condition framewise pitch predictions. During
inference, we restrict the predictions from the framewise detector by not
allowing a new note to start unless the onset detector also agrees that an
onset for that pitch is present in the frame. We focus on improving onsets and
offsets together instead of either in isolation as we believe this correlates
better with human musical perception. Our approach results in over a 100%
relative improvement in note F1 score (with offsets) on the MAPS dataset.
Furthermore, we extend the model to predict relative velocities of normalized
audio which results in more natural-sounding transcriptions. | ['Jesse Engel', 'Adam Roberts', 'Douglas Eck', 'Sageev Oore', 'Jialin Song', 'Ian Simon', 'Erich Elsen', 'Curtis Hawthorne', 'Colin Raffel'] | 2017-10-30 | null | null | null | null | ['music-transcription'] | ['music'] | [ 4.69559819e-01 -2.93350574e-02 -1.37735307e-02 -1.68791618e-02
-8.07084680e-01 -7.77003944e-01 3.03433836e-01 7.64264166e-02
-2.11037099e-01 4.49788004e-01 7.76829898e-01 1.43119037e-01
7.08223283e-02 -4.44865853e-01 -6.51676297e-01 -5.77320695e-01
-1.83075413e-01 1.98799111e-02 1.17469475e-01 -2.03514785e-01
2.99813777e-01 2.85841286e-01 -1.60745859e+00 7.33476222e-01
1.87595502e-01 8.58860791e-01 2.12313235e-02 1.32706261e+00
5.41036248e-01 8.61051977e-01 -7.19697773e-01 -1.98039323e-01
4.75834459e-01 -8.46709728e-01 -7.49171317e-01 -3.48039657e-01
6.35073125e-01 -4.30829465e-01 -1.87805653e-01 5.90310872e-01
7.24547625e-01 4.19452161e-01 6.05573118e-01 -7.07915783e-01
-2.05447391e-01 1.03029287e+00 -3.12389046e-01 1.84073940e-01
3.93869787e-01 1.08902164e-01 1.55280733e+00 -7.66431212e-01
6.00263298e-01 6.84439719e-01 9.66386735e-01 2.27622852e-01
-1.17570233e+00 -5.86440921e-01 -1.60539851e-01 2.75960296e-01
-1.26900876e+00 -6.81898117e-01 8.48964095e-01 -3.09868813e-01
1.38271272e+00 2.78909236e-01 7.42738485e-01 7.66266942e-01
3.91296484e-02 5.43865979e-01 4.71117586e-01 -7.38851547e-01
5.69346435e-02 -7.54905522e-01 -4.35246199e-01 4.85771634e-02
-6.14595175e-01 2.95687199e-01 -1.00014794e+00 3.70795280e-01
1.04514539e+00 -3.38337362e-01 -1.31489694e-01 3.26776654e-01
-1.45426631e+00 3.50639313e-01 3.83012235e-01 2.31514648e-01
-7.20291972e-01 5.29999793e-01 5.88437259e-01 2.14445695e-01
7.81360567e-02 8.59850943e-01 -5.42743921e-01 -8.02362859e-01
-1.58203983e+00 5.38385391e-01 6.30199730e-01 4.87211466e-01
3.22835177e-01 3.45641226e-01 -5.00100732e-01 8.77555192e-01
-1.91501439e-01 2.60234118e-01 2.65092641e-01 -1.60669863e+00
2.41950259e-01 -8.11100453e-02 2.22286433e-01 -7.27054477e-01
-4.02149439e-01 -6.88866556e-01 -2.89165229e-01 2.00511351e-01
5.66959321e-01 -2.64295131e-01 -7.22777426e-01 1.85556364e+00
-1.31978408e-01 6.54635131e-01 -5.80654033e-02 1.08585095e+00
3.83606404e-01 9.26351964e-01 -2.40844432e-02 -4.87337679e-01
1.24966693e+00 -9.08327401e-01 -7.80316830e-01 1.59570768e-01
3.67291868e-01 -1.18923974e+00 1.12243772e+00 7.72935092e-01
-1.25978577e+00 -9.30366635e-01 -1.06406963e+00 -2.92116880e-01
4.35347319e-01 3.44427764e-01 3.02726388e-01 1.56325176e-01
-8.22055101e-01 1.25847411e+00 -8.44697595e-01 6.73002303e-02
-8.79232585e-02 4.20872897e-01 9.70380306e-02 5.76308310e-01
-1.28330564e+00 4.11158651e-01 1.67125762e-01 -3.55182104e-02
-6.72480822e-01 -1.02494907e+00 -5.06591320e-01 3.40628833e-01
9.80319083e-02 -4.72902328e-01 1.84485269e+00 -1.03941679e+00
-1.79098856e+00 4.84256685e-01 -3.83330137e-01 -6.96707666e-01
2.65566230e-01 -5.96983492e-01 -4.31904197e-01 2.27453619e-01
-1.32208988e-01 1.07868493e+00 6.80611789e-01 -7.14450538e-01
-7.87022233e-01 3.44064981e-01 2.69639138e-02 5.07750809e-01
-1.50051247e-02 4.33204360e-02 -2.48213366e-01 -9.74735737e-01
5.24295960e-03 -1.17771029e+00 1.45946890e-01 -2.93064952e-01
-3.44454527e-01 -8.77852067e-02 4.28583860e-01 -8.77959728e-01
1.57751667e+00 -2.30291128e+00 3.15992862e-01 -5.67952022e-02
-1.56839520e-01 -1.31067354e-02 -3.17200534e-02 4.39691603e-01
-1.80577844e-01 -1.91581100e-01 -7.66214207e-02 -3.62517387e-01
6.62445650e-02 1.10513151e-01 -8.84235978e-01 4.09774967e-02
1.05486102e-01 6.41081095e-01 -8.29202592e-01 -2.01847449e-01
1.71666056e-01 6.48912728e-01 -1.09004343e+00 5.73114716e-02
-3.80239278e-01 6.09763920e-01 5.06934464e-01 3.52911085e-01
6.94468245e-02 2.54387230e-01 1.46253854e-01 -3.74283195e-01
-4.54697609e-01 1.21512628e+00 -1.16805696e+00 2.07773495e+00
-5.07450998e-01 9.24370289e-01 -2.86774695e-01 -3.61260206e-01
9.82894897e-01 7.92865634e-01 5.15508056e-01 -3.81925941e-01
-8.37188661e-02 1.87867865e-01 6.87288940e-01 -1.77012980e-01
7.93091178e-01 -4.34415877e-01 7.64303580e-02 3.91556531e-01
9.79232565e-02 -3.26179177e-01 2.29893148e-01 -2.47761235e-01
9.30477202e-01 6.32861733e-01 1.17160633e-01 2.64068060e-02
3.53732966e-02 -3.35989386e-01 5.85496485e-01 5.18128872e-01
-8.37892145e-02 1.13333094e+00 5.33233523e-01 -3.25444162e-01
-1.28934062e+00 -1.33262706e+00 6.07588552e-02 1.55051064e+00
-3.18389267e-01 -8.80168617e-01 -6.08831227e-01 -2.35829040e-01
-3.04474235e-01 8.10773015e-01 -3.49726260e-01 8.99527818e-02
-1.02596974e+00 1.48115391e-02 8.33348811e-01 8.27810287e-01
8.22174400e-02 -1.55110943e+00 -6.67529821e-01 5.21366417e-01
-4.11924511e-01 -7.44409025e-01 -8.38957846e-01 5.02510965e-01
-6.78221941e-01 -6.12911940e-01 -6.08748555e-01 -7.26191342e-01
-2.48586535e-01 -4.06123251e-01 1.02962434e+00 -2.16530815e-01
-5.71639054e-02 -2.07855433e-01 -4.32211399e-01 -4.92643714e-01
-3.70299757e-01 3.87964636e-01 3.00399423e-01 -2.95279324e-01
4.96095978e-02 -9.30097997e-01 -8.16883504e-01 9.56409425e-02
-6.08257234e-01 1.38804227e-01 2.86562443e-01 6.40504897e-01
6.66480124e-01 -2.28160053e-01 8.29184830e-01 -4.38287497e-01
4.08897668e-01 4.25592288e-02 -2.61167496e-01 -2.82613814e-01
-2.69511998e-01 -1.11111954e-01 7.70200908e-01 -5.61904073e-01
-6.83113217e-01 2.40552142e-01 -4.81751233e-01 -7.35943377e-01
-2.59432703e-01 2.90706128e-01 2.59284079e-01 5.99446952e-01
8.13800693e-01 -1.76554680e-01 -4.09761220e-01 -3.72482061e-01
3.06014448e-01 4.48779553e-01 1.07580841e+00 -3.95742476e-01
4.98495728e-01 2.01326627e-02 1.20356180e-01 -6.73662007e-01
-8.98068726e-01 -3.29699814e-01 -8.12817216e-01 -3.45054507e-01
8.30235839e-01 -8.67257714e-01 -9.63011324e-01 1.22499980e-01
-1.25850248e+00 -4.33340698e-01 -6.23591483e-01 7.59334326e-01
-9.04387712e-01 7.26918727e-02 -1.00748205e+00 -8.65834177e-01
-4.42437261e-01 -9.89951432e-01 1.02696204e+00 1.83812678e-01
-1.02182794e+00 -5.01657844e-01 3.83358687e-01 -1.47863969e-01
1.70640916e-01 3.34347747e-02 6.87989950e-01 -4.53710794e-01
-3.70334387e-01 1.80323683e-02 3.29498023e-01 3.02011937e-01
2.85832286e-01 3.70985299e-01 -1.43900490e+00 9.39080343e-02
-2.83596992e-01 -3.58304791e-02 9.11852360e-01 7.01417744e-01
8.12073290e-01 -3.12923789e-01 2.71720469e-01 6.31284058e-01
8.39965343e-01 3.05324614e-01 8.00076604e-01 9.64953974e-02
6.76629126e-01 4.14036959e-01 5.72591186e-01 5.85871518e-01
-8.66370350e-02 9.48050916e-01 1.86163083e-01 -3.53161208e-02
-4.50513721e-01 -5.16361654e-01 4.55120504e-01 1.08385313e+00
-5.06661117e-01 -1.67901099e-01 -7.12808073e-01 6.63947165e-01
-1.68343639e+00 -1.27949500e+00 -6.42299801e-02 2.20816588e+00
1.18270743e+00 2.89498925e-01 4.52551693e-01 6.70625865e-01
6.59428358e-01 1.87253833e-01 -4.50781763e-01 -8.46675813e-01
-1.28488140e-02 8.69763851e-01 2.08111495e-01 6.22240067e-01
-1.09430444e+00 1.08888304e+00 7.20926380e+00 6.79741800e-01
-1.27491617e+00 -2.46402785e-01 3.29855829e-01 -7.93149650e-01
-9.26768184e-02 1.02196075e-01 -5.67193210e-01 1.17520623e-01
1.06304336e+00 1.77180380e-01 8.21981609e-01 4.67557877e-01
4.75329757e-01 1.06619895e-01 -1.32969928e+00 8.67125094e-01
-2.22140461e-01 -1.47091746e+00 -8.30825344e-02 -1.33584678e-01
8.30673516e-01 -1.52682140e-01 1.14303105e-01 2.44718120e-01
-5.01556732e-02 -1.18422842e+00 1.09305382e+00 6.42918527e-01
8.60478580e-01 -1.03733647e+00 2.86872447e-01 5.19318245e-02
-1.37582350e+00 1.47672653e-01 -5.66205755e-02 -7.98760474e-01
3.27861190e-01 2.92406440e-01 -1.25783241e+00 3.00034344e-01
5.00477672e-01 8.99866581e-01 -7.39817768e-02 1.10698569e+00
-5.03122807e-01 1.22226334e+00 -3.60883504e-01 3.76691252e-01
-6.21140823e-02 1.82920784e-01 6.17880404e-01 1.40241456e+00
5.05270123e-01 2.64518987e-02 1.87045410e-01 8.21666777e-01
-3.97179388e-02 7.22570717e-03 -2.19558999e-01 4.77189906e-02
7.17536449e-01 1.03936291e+00 -6.81064427e-01 -2.45509550e-01
-9.82675999e-02 1.02719176e+00 -3.22609060e-02 3.15186650e-01
-9.04248536e-01 -6.02721035e-01 8.52146208e-01 -4.96515026e-03
5.51398873e-01 -2.42683411e-01 -5.20655692e-01 -6.20293379e-01
-2.58060396e-01 -7.30409741e-01 2.18776330e-01 -1.03087449e+00
-9.24125135e-01 5.45759559e-01 -3.04973334e-01 -1.43994367e+00
-1.02576065e+00 -2.77294517e-01 -7.61018574e-01 9.78561401e-01
-1.06590915e+00 -7.98373759e-01 3.65084201e-01 1.78441033e-01
6.54240966e-01 3.12584192e-01 1.03405702e+00 2.87493914e-01
-8.72411579e-02 6.12374306e-01 -5.30316532e-01 2.39547729e-01
1.20788348e+00 -1.34822226e+00 6.20505452e-01 8.16037834e-01
8.06601822e-01 5.81035197e-01 1.09776974e+00 -5.25098205e-01
-6.42977774e-01 -7.08100677e-01 1.07568955e+00 -3.85006338e-01
5.60547769e-01 -2.67880410e-01 -7.88860440e-01 6.52928710e-01
3.39871556e-01 -2.45823637e-01 6.25154674e-01 4.21516061e-01
-2.99230784e-01 -9.12064910e-02 -2.95938969e-01 7.74829447e-01
9.49000895e-01 -9.25804436e-01 -6.97990417e-01 -1.13772722e-02
8.06854963e-01 -6.35678470e-01 -8.31323802e-01 5.54258347e-01
9.76563334e-01 -1.05174410e+00 9.26220834e-01 -3.29507023e-01
7.00245678e-01 -5.60530484e-01 -1.29003614e-01 -1.23161614e+00
-4.80718583e-01 -9.34771419e-01 -1.95318297e-01 1.15774024e+00
5.10388434e-01 3.10093939e-01 6.66439354e-01 1.10525921e-01
-5.82149386e-01 -5.18757343e-01 -9.68029261e-01 -5.27678192e-01
-1.24880746e-01 -7.01353967e-01 4.21856105e-01 7.10674405e-01
3.16219062e-01 2.59023368e-01 -6.98288083e-01 1.14612944e-01
-1.59855038e-02 2.63866603e-01 5.22770882e-01 -1.08633327e+00
-8.71957004e-01 -5.21995842e-01 -2.00523511e-01 -1.06004798e+00
-1.15488797e-01 -8.59216690e-01 4.02874947e-01 -1.17552149e+00
-2.13276640e-01 6.21416606e-02 -7.12086618e-01 7.19403863e-01
-1.07529655e-01 8.37112606e-01 5.61533391e-01 2.84214258e-01
-3.78940165e-01 1.82663187e-01 1.09468663e+00 9.06868577e-02
-7.19947934e-01 9.36131328e-02 -3.07781786e-01 7.96346903e-01
7.12055087e-01 -4.07665491e-01 -1.64390787e-01 -2.47586980e-01
2.75883466e-01 4.02086496e-01 2.68329114e-01 -1.46228290e+00
1.58024311e-01 1.18080989e-01 6.58548236e-01 -6.35051250e-01
5.79567254e-01 -2.84288466e-01 2.33344898e-01 3.13219339e-01
-9.33119833e-01 7.43948892e-02 4.56307501e-01 -3.44242640e-02
-1.50543481e-01 -1.34975895e-01 5.54799080e-01 1.76087111e-01
-4.74452287e-01 -1.69167295e-01 -4.04445976e-01 -1.02091528e-01
3.06243747e-01 -2.06592858e-01 4.25169431e-02 -6.59694314e-01
-1.18894458e+00 -5.39142132e-01 2.48441771e-01 4.07064885e-01
9.46843997e-02 -1.30119860e+00 -6.62451684e-01 7.76366368e-02
-1.58800170e-01 -3.25389922e-01 2.36903727e-01 6.74628079e-01
-4.66465145e-01 2.92002022e-01 -2.39949465e-01 -6.87906802e-01
-1.16656673e+00 2.73805559e-01 3.01240474e-01 -6.05591238e-02
-7.50316322e-01 9.92031395e-01 3.86848859e-02 7.32071102e-02
3.70992959e-01 -8.27423275e-01 -1.09067149e-01 3.71143296e-02
4.85973507e-01 3.27584624e-01 5.37709333e-02 -5.33044279e-01
-2.90274531e-01 4.83510226e-01 2.06529230e-01 -6.93942845e-01
1.15114582e+00 8.32248032e-02 1.96680158e-01 1.09422767e+00
7.84065783e-01 7.05023050e-01 -1.75193930e+00 2.24473447e-01
-1.44344151e-01 -1.01486966e-01 1.12771250e-01 -1.13231504e+00
-6.25389457e-01 8.50441456e-01 3.42366099e-01 -1.82287842e-02
1.22288406e+00 -2.14943349e-01 9.05319691e-01 2.75429845e-01
-8.59808326e-02 -1.06199884e+00 5.70947640e-02 7.90352762e-01
8.68809819e-01 -4.53742057e-01 -2.02864796e-01 2.10697111e-02
-7.24821270e-01 1.31061459e+00 2.71995544e-01 -5.41078210e-01
3.11515659e-01 5.60484111e-01 2.63134956e-01 2.90486842e-01
-9.24238622e-01 -1.67491689e-01 5.83731472e-01 2.61773109e-01
1.11007428e+00 1.24378420e-01 -5.10004126e-02 6.51548564e-01
-1.00059450e+00 1.35558799e-01 3.96438926e-01 4.64976162e-01
-4.70931649e-01 -1.13486218e+00 -3.09829533e-01 6.48499280e-02
-6.90394998e-01 -5.88596642e-01 -3.81394923e-01 4.55976725e-01
3.74444574e-01 9.40564990e-01 4.96232629e-01 -6.45266354e-01
3.21232915e-01 4.66817111e-01 6.59439266e-01 -6.19296670e-01
-9.84225988e-01 9.14561272e-01 1.28688455e-01 -3.35698843e-01
-2.15580449e-01 -7.74273872e-01 -1.76002336e+00 4.58079390e-02
-4.90010493e-02 2.05433667e-02 4.37024355e-01 7.30472445e-01
3.05692226e-01 1.11140275e+00 4.29193348e-01 -1.25666630e+00
-1.24793671e-01 -1.16602457e+00 -5.09131670e-01 2.58328497e-01
4.63191122e-01 -2.61786163e-01 -1.63148150e-01 5.77325404e-01] | [15.849663734436035, 5.427392482757568] |
ace753cd-b5d2-465c-89ee-cf920109640b | sphere-embedding-an-application-to-part-of | null | null | http://papers.nips.cc/paper/3979-sphere-embedding-an-application-to-part-of-speech-induction | http://papers.nips.cc/paper/3979-sphere-embedding-an-application-to-part-of-speech-induction.pdf | Sphere Embedding: An Application to Part-of-Speech Induction | Motivated by an application to unsupervised part-of-speech tagging, we present an algorithm for the Euclidean embedding of large sets of categorical data based on co-occurrence statistics. We use the CODE model of Globerson et al. but constrain the embedding to lie on a high-dimensional unit sphere. This constraint allows for efficient optimization, even in the case of large datasets and high embedding dimensionality. Using k-means clustering of the embedded data, our approach efficiently produces state-of-the-art results. We analyze the reasons why the sphere constraint is beneficial in this application, and conjecture that these reasons might apply quite generally to other large-scale tasks. | ['Yariv Maron', 'Michael Lamar', 'Elie Bienenstock'] | 2010-12-01 | null | null | null | neurips-2010-12 | ['unsupervised-part-of-speech-tagging'] | ['natural-language-processing'] | [ 3.01183164e-02 5.09605050e-01 -6.34027198e-02 -2.10130006e-01
-7.50174463e-01 -7.62923539e-01 5.74016273e-01 3.95262182e-01
-6.90833032e-01 3.12996149e-01 6.17031991e-01 -5.34209430e-01
-2.36417770e-01 -6.39627814e-01 -4.00300086e-01 -7.21899927e-01
-4.97673273e-01 5.07542431e-01 2.88612306e-01 2.97259204e-02
2.27070287e-01 4.49451149e-01 -1.39379692e+00 3.92499454e-02
3.98874700e-01 3.69186133e-01 1.12774342e-01 5.82674921e-01
-2.30691552e-01 2.72024095e-01 -4.42287832e-01 -7.03804553e-01
2.77548432e-01 -2.58346051e-01 -9.16000962e-01 2.08234504e-01
-4.50722948e-02 1.81038439e-01 -2.38941342e-01 1.02510631e+00
4.30820614e-01 2.77534574e-01 9.14061904e-01 -1.21250916e+00
-7.82210290e-01 5.67122281e-01 -9.32846218e-02 1.73989160e-03
1.86865479e-01 -6.69952571e-01 1.47188604e+00 -8.37528050e-01
6.32593274e-01 9.86015558e-01 6.33835495e-01 3.12198251e-01
-1.34211135e+00 1.71777643e-02 -1.98107455e-02 4.26096329e-03
-1.44289029e+00 -3.41588020e-01 7.55102575e-01 -5.47987223e-01
1.14969170e+00 2.37943813e-01 5.43620944e-01 6.41521633e-01
1.13498837e-01 5.97349346e-01 8.31115961e-01 -7.12355912e-01
4.11008865e-01 1.80653468e-01 1.22947907e-02 6.75023317e-01
3.31056029e-01 -1.33310601e-01 -1.45658329e-01 -7.22698927e-01
4.44037199e-01 3.05100977e-02 2.38249257e-01 -8.97949338e-01
-1.60526061e+00 1.18803668e+00 1.91140383e-01 6.15987778e-01
-2.34140232e-01 -1.32814497e-02 3.80256981e-01 3.48574698e-01
6.18846178e-01 6.21833444e-01 -6.73077822e-01 -3.90098184e-01
-6.80747628e-01 -9.78023279e-04 1.03712344e+00 8.45246434e-01
5.91988981e-01 -3.72842073e-01 6.31474912e-01 9.62674737e-01
3.95650834e-01 1.23541906e-01 3.61251205e-01 -1.10220861e+00
3.56004894e-01 3.70643914e-01 -6.76728040e-02 -8.31725299e-01
-3.60914558e-01 1.59514137e-02 -6.16944253e-01 -9.28140357e-02
3.80672902e-01 -1.68898597e-01 -3.01313072e-01 1.53274488e+00
3.15446436e-01 -8.02837461e-02 1.88627511e-01 5.75410962e-01
2.01871693e-01 3.82743090e-01 3.56779881e-02 -2.08120570e-01
1.49414611e+00 -6.38398111e-01 -6.46106601e-01 -1.09825850e-01
1.19128966e+00 -6.10725760e-01 8.47710907e-01 7.54443035e-02
-4.98788744e-01 -2.00954631e-01 -7.93146551e-01 -5.29373996e-02
-6.61715150e-01 -1.51718194e-02 9.43072975e-01 6.74248457e-01
-1.08953631e+00 5.50151527e-01 -9.50686574e-01 -7.32968450e-01
-5.50034381e-02 4.19504881e-01 -6.95153177e-01 1.49205148e-01
-1.04030895e+00 7.99991310e-01 4.45668072e-01 -2.38419518e-01
1.31811336e-01 -5.31494170e-02 -1.05713642e+00 5.78652471e-02
9.78772715e-02 -2.13467002e-01 1.02942717e+00 -4.97032791e-01
-1.24945962e+00 9.40473557e-01 -2.77297080e-01 -2.60285497e-01
-6.97643831e-02 1.04278520e-01 -4.50255871e-01 1.55186817e-01
1.22659523e-02 5.63854039e-01 3.94251853e-01 -8.43535841e-01
-1.85139492e-01 -3.25559616e-01 3.00019141e-02 1.40174180e-01
-7.22557843e-01 3.72211486e-01 -2.33714178e-01 -7.17853546e-01
5.49468935e-01 -1.31621671e+00 -2.90728390e-01 -6.55943826e-02
1.06838495e-02 -3.99055183e-01 3.92956853e-01 -5.24419308e-01
1.26378143e+00 -2.47961020e+00 2.91829616e-01 2.53559411e-01
2.17864737e-01 -1.68041378e-01 -1.69379890e-01 8.80687535e-01
-2.98049569e-01 3.08080584e-01 -3.00639868e-01 -3.80692273e-01
3.99394870e-01 4.50664967e-01 -3.43275778e-02 7.43341506e-01
1.66224808e-01 6.72431588e-01 -9.53395188e-01 -7.67118514e-01
1.43968225e-01 3.48421574e-01 -6.93185627e-01 3.67588662e-02
1.63273200e-01 -7.20244721e-02 -3.78105581e-01 4.05066073e-01
2.11152539e-01 -3.36199343e-01 3.87445182e-01 -4.03358527e-02
-1.31634101e-01 4.65855658e-01 -1.27479815e+00 1.67217731e+00
-3.04117113e-01 8.06520283e-01 2.16015056e-03 -1.06551194e+00
7.32636869e-01 4.67364281e-01 8.90710115e-01 -7.67372176e-02
-3.09073329e-02 2.00308129e-01 2.86772817e-01 -3.41943532e-01
6.26609504e-01 -3.72946054e-01 -3.97064567e-01 5.81794918e-01
5.84908240e-02 -6.21696301e-02 1.81095853e-01 3.34007055e-01
1.35384095e+00 -2.23935455e-01 4.03692365e-01 -4.21114564e-01
6.33249283e-02 -7.80350864e-02 4.86787468e-01 3.58056694e-01
-2.28839964e-01 5.87290823e-01 7.60210395e-01 -2.65898913e-01
-1.26067352e+00 -9.74476933e-01 -4.57476377e-01 1.24556601e+00
-2.46164948e-01 -8.38817596e-01 -5.13612211e-01 -8.31485450e-01
1.30562767e-01 4.87032205e-01 -7.60037541e-01 1.29542917e-01
-2.29370296e-01 -7.03826427e-01 4.38376606e-01 4.60929632e-01
-2.69950539e-01 -9.71627772e-01 -2.43705690e-01 3.00148934e-01
-1.41316634e-02 -1.13130665e+00 -2.88754463e-01 6.26043200e-01
-9.54449773e-01 -9.74577248e-01 -6.13956571e-01 -9.95449901e-01
8.27979743e-01 1.28435612e-01 1.09370077e+00 2.54450808e-03
-4.66837958e-02 4.23029959e-01 -8.07802022e-01 -4.70160097e-02
-4.71845269e-01 2.29022548e-01 2.15465739e-01 -2.17075944e-01
8.68453562e-01 -4.11238164e-01 -1.98488295e-01 4.11605835e-01
-1.06352115e+00 -4.03906971e-01 5.13025105e-01 6.31714106e-01
3.99689257e-01 1.96391135e-01 1.68688983e-01 -8.14767778e-01
5.04136562e-01 -6.16875947e-01 -5.03308773e-01 5.19014746e-02
-5.10760844e-01 3.86882663e-01 4.52527314e-01 -5.08419991e-01
-2.05966718e-02 1.82945311e-01 -2.23690152e-01 -1.92378029e-01
-1.44391507e-01 4.73114014e-01 -7.47138187e-02 5.76444715e-03
3.41775179e-01 8.28299001e-02 -8.81484523e-03 -6.30116701e-01
6.06530607e-01 1.17020667e+00 -2.50167940e-02 -4.58503962e-01
6.95098639e-01 3.68818104e-01 -2.32761223e-02 -1.03646576e+00
-5.05831838e-01 -8.57657611e-01 -1.12244999e+00 3.52549493e-01
1.00230753e+00 -9.30429578e-01 -4.89746153e-01 -1.62304491e-01
-1.06335437e+00 -1.89355046e-01 -2.57759660e-01 7.05675066e-01
-6.19750619e-01 6.08989179e-01 -5.03464997e-01 -7.67961323e-01
2.52707988e-01 -9.25221980e-01 1.14171410e+00 -5.00823200e-01
-5.15180886e-01 -1.35309303e+00 4.80269819e-01 7.39843920e-02
1.59457773e-02 -1.50019154e-01 9.12179768e-01 -1.09941578e+00
-1.29978031e-01 -2.47567967e-01 1.14619568e-01 2.79909492e-01
1.92700744e-01 -4.49057333e-02 -5.89213133e-01 -2.12946236e-01
-2.68916219e-01 -4.80822586e-02 6.01702690e-01 5.76373786e-02
1.06015193e+00 -3.33362579e-01 -3.50891858e-01 2.88829744e-01
1.36015403e+00 -1.48566410e-01 4.32236105e-01 3.75626177e-01
6.43140316e-01 8.36112738e-01 6.09595537e-01 3.82387012e-01
4.86062050e-01 7.47292101e-01 1.36756331e-01 1.82719246e-01
4.03108269e-01 -1.96366161e-01 4.05918747e-01 1.42996323e+00
3.07969093e-01 -1.19603001e-01 -1.02426398e+00 1.01233947e+00
-1.74347126e+00 -1.00455511e+00 -1.32047430e-01 2.06126523e+00
6.42101705e-01 1.14478640e-01 3.84126127e-01 1.25535488e-01
6.58324897e-01 4.04099673e-01 -5.40110609e-03 -6.52637839e-01
-1.15179673e-01 -1.78706534e-02 8.07694376e-01 5.90240657e-01
-1.05129325e+00 7.46191978e-01 7.47470617e+00 4.86469567e-01
-5.66987753e-01 3.19461197e-01 1.93933636e-01 1.13195926e-01
-4.61179376e-01 1.79998755e-01 -5.13863087e-01 4.90217358e-01
1.18072069e+00 1.80061497e-02 3.89697105e-01 7.60051906e-01
-1.83538988e-01 4.44332734e-02 -9.86295998e-01 8.36418808e-01
9.02616903e-02 -8.63309443e-01 -4.15074706e-01 5.99203110e-01
6.72200799e-01 2.18670145e-01 -1.95891261e-01 6.81859329e-02
4.63264525e-01 -6.97460353e-01 3.42598557e-01 5.27628418e-03
6.79697514e-01 -6.14236534e-01 7.74526775e-01 2.37998366e-01
-1.17661881e+00 1.17191887e-02 -6.88847542e-01 -3.71595621e-02
1.08485565e-01 8.73617351e-01 -7.66022861e-01 1.68645978e-01
3.86469007e-01 6.63470805e-01 -4.80522394e-01 8.88044238e-01
-1.15702255e-02 6.72520638e-01 -5.48921347e-01 -3.40149641e-01
2.63930380e-01 -4.04650718e-01 4.24664110e-01 1.41552281e+00
2.84174740e-01 1.76449016e-01 5.36423223e-03 2.74876654e-01
1.11839660e-01 2.54679143e-01 -1.02638495e+00 -5.74982762e-01
7.29916692e-01 1.11595476e+00 -9.30195093e-01 -2.62164384e-01
-8.02597046e-01 1.02896726e+00 4.17726576e-01 8.24062154e-02
-4.13553774e-01 -7.00876117e-01 9.38845754e-01 7.66895115e-02
8.13792646e-01 -7.29850590e-01 -2.33896166e-01 -1.22335410e+00
1.17401674e-01 -6.50641739e-01 3.14213544e-01 -4.30301845e-01
-1.22428608e+00 4.33174461e-01 4.84318892e-03 -1.18692935e+00
-6.14063740e-01 -8.82196009e-01 -2.67560393e-01 5.53529859e-01
-9.79383767e-01 -6.97401106e-01 4.32934672e-01 3.47067147e-01
1.63688883e-01 -2.22693294e-01 1.16398299e+00 2.02953696e-01
-2.37645864e-01 3.27330917e-01 6.01290226e-01 5.17708421e-01
5.94001770e-01 -1.61234856e+00 6.45461023e-01 6.17069662e-01
8.46829176e-01 7.74926007e-01 8.29030454e-01 -4.84931707e-01
-1.46560919e+00 -7.27523685e-01 1.45554364e+00 -8.58880103e-01
1.12711060e+00 -8.30793679e-01 -7.19041407e-01 7.50381351e-01
-5.50251454e-02 1.95989966e-01 1.26750207e+00 7.08430052e-01
-5.22560477e-01 2.59822041e-01 -9.30747211e-01 5.18178761e-01
1.09910655e+00 -9.31324720e-01 -8.94414186e-01 5.80465972e-01
6.58084750e-01 1.22765653e-01 -1.35376096e+00 -1.16802221e-02
4.33598965e-01 -7.27268636e-01 8.94113481e-01 -6.69923365e-01
1.49127588e-01 -1.69961959e-01 -5.88249981e-01 -1.33267796e+00
-6.33169055e-01 -6.71750784e-01 -8.81676748e-02 1.06126666e+00
4.78504628e-01 -6.83456302e-01 6.22029960e-01 6.25912249e-01
9.47654769e-02 -6.96346462e-01 -1.09403920e+00 -8.99091244e-01
2.43380249e-01 -4.98398781e-01 5.52167058e-01 1.12835610e+00
5.79137504e-01 2.12516233e-01 -8.95848349e-02 1.75167710e-01
6.06747389e-01 2.87511311e-02 4.78088051e-01 -1.47600174e+00
-3.80535573e-01 -2.51936942e-01 -8.79711211e-01 -9.52764630e-01
6.22903407e-01 -9.35272217e-01 3.36123891e-02 -1.32205999e+00
1.04108840e-01 -6.05540276e-01 -3.21348697e-01 4.73582774e-01
-3.54329832e-02 2.81072021e-01 2.89679170e-01 2.71963447e-01
-8.72903287e-01 4.63324875e-01 5.33025742e-01 1.55203432e-01
5.17307296e-02 -2.42191255e-01 -7.73967028e-01 5.38975716e-01
6.94898367e-01 -8.73115361e-01 -8.70470181e-02 -3.74252707e-01
4.64828998e-01 -3.05586785e-01 1.22460365e-01 -6.11584485e-01
4.88018431e-02 -5.73928542e-02 -4.86968681e-02 -2.84369320e-01
3.74828577e-01 -1.14891458e+00 5.41288555e-02 1.84369951e-01
-3.46186846e-01 3.35753292e-01 -5.79281524e-02 5.55912435e-01
-2.71798402e-01 -4.95885015e-01 4.23604816e-01 4.26192507e-02
-4.93860036e-01 8.46644416e-02 -4.30256218e-01 9.08726230e-02
7.98748314e-01 -2.48210967e-01 1.21344469e-01 -2.41220847e-01
-8.32569063e-01 -1.93246454e-01 8.46885204e-01 4.75398421e-01
2.38471672e-01 -1.72248030e+00 -6.09858811e-01 2.49537632e-01
4.11020070e-01 -5.14164805e-01 -4.50042129e-01 7.40634501e-01
-2.47493148e-01 5.46184003e-01 2.58639038e-01 -3.24615330e-01
-1.18818700e+00 6.92707956e-01 -1.83596820e-01 -1.19253948e-01
-4.98176455e-01 4.75792795e-01 -1.43195495e-01 -7.27181852e-01
1.99475959e-02 -2.56025046e-01 -8.10660571e-02 2.85670936e-01
2.34847024e-01 2.04461426e-01 9.04642195e-02 -8.90189767e-01
-6.34662926e-01 5.23821831e-01 2.47292936e-01 -4.96320009e-01
1.57076979e+00 -7.86623955e-02 -2.60660946e-01 7.53800690e-01
1.91177487e+00 3.94064605e-01 -8.08323085e-01 -2.09476843e-01
1.21088527e-01 -4.62730736e-01 -1.10381812e-01 -1.14933923e-01
-5.65459132e-01 8.25423658e-01 1.61427379e-01 6.13100767e-01
6.66052282e-01 4.84219909e-01 4.29644108e-01 6.35418117e-01
3.69298398e-01 -1.10505354e+00 -2.94261187e-01 6.85218155e-01
3.98580164e-01 -1.12003541e+00 2.85408534e-02 -2.86912829e-01
-7.50523806e-01 9.38364923e-01 -2.16530323e-01 -2.49067783e-01
9.71719086e-01 2.02475995e-01 -2.45369766e-02 -1.55008867e-01
-7.54535496e-01 -4.26924944e-01 3.01072419e-01 6.53006196e-01
7.22207487e-01 2.84772903e-01 -5.04819870e-01 1.71031788e-01
-4.04731423e-01 -4.81498957e-01 3.35465550e-01 9.41096544e-01
-4.35158730e-01 -1.37507236e+00 -3.19156736e-01 5.95117807e-01
-5.00451505e-01 -2.76909441e-01 -4.79007334e-01 7.43936896e-01
-3.07583455e-02 1.00430655e+00 2.65924603e-01 -3.44294876e-01
7.64089003e-02 4.06347245e-01 3.71198744e-01 -8.16384614e-01
-3.70831460e-01 2.29918644e-01 2.12963626e-01 -4.25356388e-01
-4.23725307e-01 -1.15394092e+00 -1.16162539e+00 -1.97596267e-01
-4.21066761e-01 6.39994919e-01 9.38800156e-01 8.30940962e-01
6.07087433e-01 2.61209048e-02 7.24826157e-01 -4.71966863e-01
-7.31490970e-01 -1.03142965e+00 -1.02683949e+00 5.77277184e-01
2.26942599e-01 -6.15079224e-01 -6.65311038e-01 -8.39733556e-02] | [10.31196403503418, 8.404583930969238] |
1fc38147-6c16-4752-8005-86069ea83626 | generalized-transition-based-dependency | null | null | https://aclanthology.org/P16-1015 | https://aclanthology.org/P16-1015.pdf | Generalized Transition-based Dependency Parsing via Control Parameters | null | ['Emily Pitler', 'Ryan Mcdonald', 'Bernd Bohnet', 'Ji Ma'] | 2016-08-01 | null | null | null | acl-2016-8 | ['transition-based-dependency-parsing'] | ['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] | [-7.282987117767334, 3.749140977859497] |
f9ebba1f-e0b4-48a1-8769-22c5af17f6f1 | model-extraction-attacks-on-split-federated | 2303.08581 | null | https://arxiv.org/abs/2303.08581v1 | https://arxiv.org/pdf/2303.08581v1.pdf | Model Extraction Attacks on Split Federated Learning | Federated Learning (FL) is a popular collaborative learning scheme involving multiple clients and a server. FL focuses on protecting clients' data but turns out to be highly vulnerable to Intellectual Property (IP) threats. Since FL periodically collects and distributes the model parameters, a free-rider can download the latest model and thus steal model IP. Split Federated Learning (SFL), a recent variant of FL that supports training with resource-constrained clients, splits the model into two, giving one part of the model to clients (client-side model), and the remaining part to the server (server-side model). Thus SFL prevents model leakage by design. Moreover, by blocking prediction queries, it can be made resistant to advanced IP threats such as traditional Model Extraction (ME) attacks. While SFL is better than FL in terms of providing IP protection, it is still vulnerable. In this paper, we expose the vulnerability of SFL and show how malicious clients can launch ME attacks by querying the gradient information from the server side. We propose five variants of ME attack which differs in the gradient usage as well as in the data assumptions. We show that under practical cases, the proposed ME attacks work exceptionally well for SFL. For instance, when the server-side model has five layers, our proposed ME attack can achieve over 90% accuracy with less than 2% accuracy degradation with VGG-11 on CIFAR-10. | ['Chaitali Chakrabarti', 'Deliang Fan', 'Zhezhi He', 'Li Yang', 'Xing Chen', 'Adnan Siraj Rakin', 'Jingtao Li'] | 2023-03-13 | null | null | null | null | ['blocking'] | ['natural-language-processing'] | [-1.41553521e-01 -5.90795092e-02 -3.65490407e-01 -1.64579481e-01
-8.46108556e-01 -1.06827784e+00 2.47602656e-01 -2.27244318e-01
-3.19220781e-01 7.30714560e-01 -2.64060766e-01 -9.23489809e-01
-1.13064125e-01 -9.18586135e-01 -8.47305298e-01 -7.89331257e-01
-2.54088104e-01 2.01547489e-01 7.26854622e-01 9.80523378e-02
8.45342129e-02 8.99585307e-01 -1.04875064e+00 5.43611765e-01
5.98349869e-01 1.26353157e+00 -4.04916495e-01 4.95118022e-01
-2.26042822e-01 9.97866452e-01 -8.44597101e-01 -7.74109244e-01
4.82519865e-01 -6.19212762e-02 -9.47058320e-01 -5.73026061e-01
5.36769688e-01 -6.18123591e-01 -3.31752509e-01 1.04847515e+00
3.99441928e-01 -3.97874951e-01 1.36886388e-01 -1.88878560e+00
-9.49282497e-02 9.13942516e-01 -2.76814640e-01 -1.84726417e-02
-1.28022075e-01 1.13839246e-01 8.72202754e-01 -5.07309437e-01
5.70365429e-01 7.88290739e-01 7.36376584e-01 8.24474454e-01
-1.14731717e+00 -1.48742235e+00 1.46619499e-01 2.53555298e-01
-1.19202375e+00 -5.09274840e-01 6.77996576e-01 -9.99408886e-02
4.83375072e-01 6.07209682e-01 -5.16071245e-02 1.11162937e+00
1.76007047e-01 8.13529909e-01 1.05718815e+00 2.56941691e-02
6.12709820e-01 5.26954710e-01 3.14037412e-01 2.60931671e-01
3.33423167e-01 6.08228743e-02 -6.07101023e-01 -1.14977825e+00
3.29913884e-01 1.93094388e-02 -4.70298856e-01 -5.73827147e-01
-4.40946937e-01 8.29074860e-01 3.05216730e-01 3.21065746e-02
-2.09760144e-02 1.49098352e-01 5.60603976e-01 5.73976696e-01
1.43124446e-01 -1.78533807e-01 -8.82489264e-01 2.42092311e-01
-8.49204361e-01 1.34249017e-01 1.17535114e+00 7.19175696e-01
5.70522070e-01 -1.44557208e-02 1.77877948e-01 3.46769214e-01
4.44606394e-01 4.12028313e-01 3.05459201e-01 -9.66550052e-01
5.70019960e-01 2.17086568e-01 7.62065947e-02 -8.49592984e-01
-6.07967973e-02 -7.25278080e-01 -5.69478691e-01 4.50604081e-01
5.96181870e-01 -3.87288898e-01 -1.71279341e-01 2.01763129e+00
5.01991451e-01 3.38233739e-01 1.54782653e-01 5.47599316e-01
3.56354803e-01 3.58879238e-01 1.27973840e-01 -1.32338628e-01
1.08469367e+00 -7.76573241e-01 -3.09061319e-01 6.01699725e-02
8.04811656e-01 -3.14420015e-01 4.81126547e-01 7.50886977e-01
-8.20391417e-01 9.01513025e-02 -1.03247559e+00 4.37613457e-01
-3.45547467e-01 -4.38222766e-01 5.00703871e-01 1.30015087e+00
-9.23458993e-01 6.54039145e-01 -7.86027193e-01 2.31056005e-01
9.64557350e-01 5.28637946e-01 -5.44926941e-01 1.16529442e-01
-1.41424537e+00 3.34445357e-01 -1.11100696e-01 -5.42534292e-01
-1.04277778e+00 -9.26528394e-01 -3.18026781e-01 3.82387400e-01
3.17449778e-01 -3.46009284e-01 1.26807427e+00 -9.56453919e-01
-1.25514829e+00 3.97791535e-01 1.91356868e-01 -7.88899004e-01
8.25753272e-01 -5.58281988e-02 -5.27120173e-01 2.97987133e-01
-3.31986040e-01 -5.75351454e-02 1.00678325e+00 -1.24086845e+00
-6.94818199e-01 -6.42028511e-01 -3.74315605e-02 -5.13368905e-01
-7.61520028e-01 2.97596574e-01 -5.57946451e-02 -4.14211690e-01
-7.37784356e-02 -7.70435214e-01 1.79868732e-02 5.22152126e-01
-4.50358391e-01 -1.10467141e-02 1.52587211e+00 -4.97619003e-01
1.31276035e+00 -2.25121665e+00 -7.49197483e-01 5.82700014e-01
3.69744539e-01 6.93956017e-01 9.38935056e-02 4.90970939e-01
5.56380935e-02 3.52605551e-01 -7.71867558e-02 -4.23080146e-01
-3.40641104e-02 1.30219739e-02 -7.17912972e-01 6.55791640e-01
-6.66318953e-01 5.45283258e-01 -4.27944392e-01 -3.44322890e-01
-3.55873764e-01 5.45736611e-01 -6.94156289e-01 1.53810084e-02
-2.01909468e-01 3.00173819e-01 -6.50249958e-01 5.28249502e-01
1.11952889e+00 -4.17978674e-01 4.56585258e-01 -3.06345616e-02
8.79305229e-02 3.80919069e-01 -1.25803030e+00 9.85955477e-01
-3.85891050e-01 4.00656134e-01 5.81973076e-01 -6.66729808e-01
7.14132547e-01 5.00184894e-01 3.84539306e-01 -3.88292909e-01
-2.68597179e-03 3.13410103e-01 -3.88370723e-01 -3.69205996e-02
-2.49532580e-01 1.49008304e-01 2.31725708e-01 9.22362268e-01
-1.83091953e-01 9.07836080e-01 -7.45491683e-01 4.21641380e-01
1.37586677e+00 -2.41414592e-01 -1.43116698e-01 -6.50310814e-02
6.50463521e-01 -3.42915922e-01 8.32570314e-01 9.00344312e-01
-5.15959978e-01 1.45791760e-02 6.82489395e-01 -1.96846291e-01
-5.57987213e-01 -1.22395957e+00 -1.75185233e-01 1.15565205e+00
1.04731910e-01 -5.29428482e-01 -8.56885672e-01 -1.43332589e+00
4.17774200e-01 7.44667232e-01 -1.68089002e-01 -2.68476695e-01
-3.44840050e-01 -3.64262074e-01 1.21484172e+00 3.00171614e-01
8.89718592e-01 -7.00278580e-01 -4.63375211e-01 1.38556004e-01
-5.65454662e-02 -9.53868747e-01 -5.40528297e-01 1.34214774e-01
-9.62935567e-01 -1.25454032e+00 -3.04496586e-01 -2.84093231e-01
2.49106377e-01 3.53447765e-01 6.37136936e-01 8.58050510e-02
-1.19537022e-02 1.37906879e-01 -1.45036737e-02 -2.82043755e-01
-6.43121481e-01 4.51372424e-03 -6.33336529e-02 5.07271409e-01
2.29894862e-01 -8.15825224e-01 -5.27722597e-01 6.20038390e-01
-1.00406468e+00 -4.26624298e-01 1.32191122e-01 4.79428858e-01
1.87246829e-01 1.67971551e-01 8.15429568e-01 -1.39623988e+00
1.53382972e-01 -6.77537858e-01 -7.38415658e-01 3.62678587e-01
-9.54035521e-01 -1.69433147e-01 9.90799248e-01 -5.90556145e-01
-9.14878607e-01 -4.79695313e-02 -1.24912545e-01 -7.20976770e-01
-1.49076238e-01 -1.02762528e-01 -6.86283886e-01 -5.16496837e-01
8.13118041e-01 1.07335165e-01 1.31891578e-01 -9.82744694e-01
1.49477065e-01 1.08981359e+00 2.90106058e-01 -3.77767920e-01
8.56804729e-01 6.92424953e-01 -5.72711863e-02 -6.47306621e-01
-4.70704824e-01 -9.41809863e-02 8.14887136e-03 -1.46332635e-02
2.67679960e-01 -6.21925712e-01 -1.03123009e+00 8.64472032e-01
-9.68395412e-01 -2.36221910e-01 6.52560294e-02 2.88974851e-01
-3.12430024e-01 4.02661324e-01 -9.06547785e-01 -7.77927816e-01
-7.54340768e-01 -8.18265140e-01 1.48545116e-01 -2.60552038e-02
2.24422172e-01 -8.70380938e-01 -1.41875982e-01 6.54626846e-01
8.72631490e-01 8.89282376e-02 9.79037106e-01 -1.42290890e+00
-5.73611617e-01 -3.82822931e-01 -1.21169470e-01 5.80350935e-01
-1.33761123e-01 -2.00487792e-01 -1.35039556e+00 -5.42234778e-01
2.18938276e-01 -2.48573378e-01 4.78514075e-01 -1.29546553e-01
1.32094836e+00 -8.64275455e-01 -3.83866280e-01 7.88164794e-01
1.49491048e+00 1.48058355e-01 6.58194602e-01 2.88052350e-01
4.53011036e-01 3.75621527e-01 8.30238406e-03 5.22806466e-01
9.54569876e-02 5.49431980e-01 7.60037243e-01 1.67500183e-01
1.80196136e-01 -3.81526351e-01 5.53134620e-01 1.68751869e-02
6.11033201e-01 -1.65807888e-01 -6.51661873e-01 2.80610751e-02
-1.58187020e+00 -1.08516371e+00 -1.18634753e-01 2.47461438e+00
8.34858596e-01 2.36417785e-01 2.84464419e-01 3.00670058e-01
4.94340450e-01 9.36462265e-03 -8.95872593e-01 -4.54613715e-01
-6.64348453e-02 2.01333210e-01 9.90117490e-01 2.03867555e-01
-1.02225733e+00 6.59532070e-01 5.39353228e+00 9.53764617e-01
-1.21866679e+00 6.31632090e-01 5.79418898e-01 -1.14163667e-01
-2.08798155e-01 1.85349539e-01 -1.03068817e+00 6.43995404e-01
1.25641143e+00 -8.89339447e-02 2.81023264e-01 1.12467706e+00
-1.38236329e-01 2.80550867e-01 -8.41922462e-01 7.75170684e-01
-1.78119749e-01 -1.42390871e+00 4.20877673e-02 3.85386914e-01
1.78647280e-01 3.01350445e-01 3.47833745e-02 1.01454608e-01
2.74126023e-01 -8.30252647e-01 6.03411853e-01 2.75122255e-01
6.73507094e-01 -1.13178921e+00 3.82368237e-01 8.31786692e-01
-7.82762170e-01 -5.33296645e-01 -2.04598770e-01 4.29208368e-01
-2.45744199e-01 3.56422096e-01 -2.32169747e-01 4.13374692e-01
8.68320048e-01 1.18295997e-01 -5.05676568e-01 9.84959304e-01
8.95687416e-02 1.14998364e+00 -5.45622230e-01 3.55639100e-01
1.67688981e-01 5.95820099e-02 6.14473999e-01 8.56769264e-01
-1.34928316e-01 -2.10730493e-01 -2.87287179e-02 8.83341551e-01
-3.27120245e-01 1.35999635e-01 -5.28592110e-01 3.65244269e-01
8.44079077e-01 9.98186588e-01 -1.89366445e-01 -1.26687154e-01
-4.14055377e-01 9.22251403e-01 2.01480106e-01 3.75640392e-01
-6.99413717e-01 -5.22805333e-01 1.02528238e+00 2.40052566e-01
3.41256469e-01 1.04466766e-01 -1.72930956e-01 -1.02840126e+00
7.19239120e-04 -1.19331396e+00 7.54481196e-01 -2.67589808e-01
-1.31829381e+00 6.08708799e-01 -3.30363184e-01 -1.01716733e+00
1.14313945e-01 -2.53495365e-01 -6.05903268e-01 8.07266057e-01
-1.38247490e+00 -1.18894601e+00 2.17813969e-01 1.24589550e+00
-6.49149567e-02 -2.26146832e-01 8.94028246e-01 4.20250237e-01
-6.10121548e-01 1.34869492e+00 4.09973741e-01 4.26976383e-01
6.57487571e-01 -6.15855336e-01 2.67756432e-01 6.35091543e-01
3.31232965e-01 6.33886516e-01 3.40634882e-01 -5.05449831e-01
-1.36348677e+00 -1.05942917e+00 8.52753937e-01 -1.89363360e-01
5.90024769e-01 -4.67801452e-01 -1.19513643e+00 7.16127634e-01
-3.63623857e-01 3.30514133e-01 1.08829975e+00 -3.69725019e-01
-1.08742368e+00 -4.46751386e-01 -1.86568010e+00 2.96420485e-01
6.24156475e-01 -6.34209454e-01 1.86952967e-02 3.13885868e-01
6.48473442e-01 4.00087312e-02 -7.87209392e-01 6.39193133e-02
7.14068592e-01 -1.36307049e+00 8.40696335e-01 -9.84990537e-01
-5.23071706e-01 1.52892843e-02 -3.13482434e-01 -6.64173424e-01
-1.38141951e-02 -9.78926480e-01 -5.36954284e-01 1.30663931e+00
4.39142346e-01 -1.20045352e+00 1.28015900e+00 8.60562503e-01
5.57620883e-01 -6.27946913e-01 -1.19860995e+00 -1.13453865e+00
5.65917790e-01 -7.22132683e-01 9.36941028e-01 8.46885979e-01
-1.23497203e-01 -2.60342538e-01 -3.26813251e-01 5.35591722e-01
1.19180667e+00 -6.17769323e-02 7.60588646e-01 -1.29514861e+00
-4.60891664e-01 -1.66833088e-01 -1.28150761e-01 -8.65985096e-01
2.65359610e-01 -9.42029595e-01 -8.17595065e-01 -7.33054578e-01
-9.46005061e-02 -8.02422941e-01 -5.62821150e-01 8.53524029e-01
5.38783252e-01 2.51247197e-01 5.46863914e-01 6.06119454e-01
-2.82643288e-01 -2.08629649e-02 5.53611159e-01 -2.40762569e-02
-2.27384970e-01 6.49963498e-01 -6.12504005e-01 5.27496278e-01
1.08674610e+00 -7.99003780e-01 -5.08682251e-01 1.92843471e-02
2.31343508e-02 2.07276851e-01 5.93424976e-01 -8.32793474e-01
5.22544742e-01 -8.48475769e-02 2.00301915e-01 -4.00444835e-01
-7.22031370e-02 -1.27969551e+00 4.71662879e-01 6.01256013e-01
-3.74458462e-01 -5.11910975e-01 -3.00909136e-03 6.33105516e-01
2.31223300e-01 -1.00735925e-01 1.12243688e+00 1.61684081e-01
1.87844075e-02 4.79604393e-01 -3.42334956e-01 -9.13310573e-02
1.10441947e+00 -3.99970412e-02 -5.88976145e-01 -4.62586790e-01
-7.82731116e-01 2.45308265e-01 5.62898338e-01 1.05835363e-01
4.25369442e-01 -9.44302738e-01 -3.67221087e-01 5.52966297e-01
-2.62044162e-01 -5.24731934e-01 2.80274421e-01 8.28654051e-01
-2.99564749e-01 3.80245715e-01 1.22706823e-01 -9.24910754e-02
-1.44243145e+00 7.83304453e-01 6.79894626e-01 -1.60008624e-01
-8.42271149e-01 6.54077113e-01 7.94646367e-02 -3.04213077e-01
8.19302499e-01 4.75930870e-01 1.34053111e-01 -8.56409669e-02
9.81733024e-01 7.43754804e-01 2.05389142e-01 -4.08735007e-01
-6.26254857e-01 -5.37059046e-02 -2.12323785e-01 2.19442453e-02
1.01207316e+00 -7.95032550e-03 -7.12548494e-02 -5.17264828e-02
1.53982353e+00 3.41295838e-01 -1.18145645e+00 -5.57209671e-01
1.26324356e-01 -4.64538574e-01 2.09049985e-01 -9.94561255e-01
-1.49571836e+00 6.46421134e-01 5.70251882e-01 1.22133799e-01
1.08731222e+00 -2.23270491e-01 1.27675545e+00 2.46659786e-01
9.03196990e-01 -7.08020985e-01 -3.25401008e-01 1.72373161e-01
2.34055862e-01 -6.11890733e-01 -3.08612168e-01 -3.20394665e-01
-4.97253716e-01 9.79697526e-01 5.03722548e-01 1.05157740e-01
1.21260321e+00 4.99904394e-01 1.73107103e-01 1.23527981e-01
-1.01069713e+00 5.92706800e-01 -2.61063933e-01 6.03050530e-01
-3.20790410e-01 -1.89395010e-01 1.02499135e-01 1.04849494e+00
8.27106312e-02 3.88925374e-02 3.57144415e-01 9.68347251e-01
-5.68144143e-01 -1.29871082e+00 -5.44718683e-01 4.19038981e-01
-1.06590593e+00 2.30067149e-01 -3.58567804e-01 3.91149253e-01
-6.82312474e-02 1.14012825e+00 -1.22915603e-01 -3.80796164e-01
1.42291309e-02 4.21738446e-01 -2.03725442e-01 -1.40225887e-01
-1.04105473e+00 -1.12785414e-01 -9.77862030e-02 -9.73851204e-01
1.98316962e-01 -2.48565495e-01 -1.29454899e+00 -6.71853304e-01
-5.10486603e-01 3.98355395e-01 8.35731089e-01 5.67535818e-01
6.79950178e-01 -3.13840806e-01 1.32865632e+00 -8.28547031e-03
-1.26021361e+00 -4.02306169e-01 -1.15334511e+00 9.69974771e-02
4.07919735e-01 -3.13364238e-01 -9.88332748e-01 -4.39936161e-01] | [5.794227123260498, 6.840498924255371] |
13508250-8a43-41ca-ab3c-9e818e4f2b37 | clear-a-dataset-for-compositional-language | 1811.10561 | null | http://arxiv.org/abs/1811.10561v1 | http://arxiv.org/pdf/1811.10561v1.pdf | CLEAR: A Dataset for Compositional Language and Elementary Acoustic Reasoning | We introduce the task of acoustic question answering (AQA) in the area of
acoustic reasoning. In this task an agent learns to answer questions on the
basis of acoustic context. In order to promote research in this area, we
propose a data generation paradigm adapted from CLEVR (Johnson et al. 2017). We
generate acoustic scenes by leveraging a bank elementary sounds. We also
provide a number of functional programs that can be used to compose questions
and answers that exploit the relationships between the attributes of the
elementary sounds in each scene. We provide AQA datasets of various sizes as
well as the data generation code. As a preliminary experiment to validate our
data, we report the accuracy of current state of the art visual question
answering models when they are applied to the AQA task without modifications.
Although there is a plethora of question answering tasks based on text, image
or video data, to our knowledge, we are the first to propose answering
questions directly on audio streams. We hope this contribution will facilitate
the development of research in the area. | ['Jerome Abdelnour', 'Jean Rouat', 'Giampiero Salvi'] | 2018-11-26 | null | null | null | null | ['acoustic-question-answering'] | ['speech'] | [ 4.31738555e-01 1.97258398e-01 6.85514212e-01 -4.79682893e-01
-1.09953082e+00 -6.68971598e-01 7.40524590e-01 1.51965499e-01
-3.21242332e-01 6.04372248e-02 4.49731827e-01 -4.37690794e-01
-1.12060905e-01 -1.02585447e+00 -8.07638347e-01 -2.19163522e-01
6.45997524e-02 3.57881397e-01 4.36778009e-01 -4.68900770e-01
1.96769878e-01 -1.30397305e-01 -2.10617280e+00 7.22026467e-01
3.47523272e-01 1.21654546e+00 3.98090273e-01 1.21712828e+00
-5.02942681e-01 1.12295377e+00 -8.78325522e-01 -6.22815967e-01
-2.43419018e-02 -7.99039066e-01 -1.22068417e+00 -2.85158306e-01
7.80603170e-01 -3.54256004e-01 -1.64465249e-01 6.27075315e-01
4.69230533e-01 3.73402089e-01 5.26203454e-01 -1.31987536e+00
-8.62478018e-01 8.52672219e-01 3.16972584e-01 3.12887102e-01
9.49394345e-01 3.18336487e-01 1.62713063e+00 -1.01643360e+00
3.76831859e-01 1.29069424e+00 5.06890714e-01 8.09433401e-01
-1.03720605e+00 -6.46207750e-01 2.94618458e-01 6.20352447e-01
-9.08925474e-01 -5.07666469e-01 9.31980848e-01 -4.14674968e-01
1.09399712e+00 4.70514774e-01 8.84102285e-01 1.34252846e+00
-3.77063930e-01 8.49706888e-01 8.87439072e-01 -5.27700186e-01
6.26918912e-01 -2.29582205e-01 1.79070383e-01 7.70986438e-01
-4.11839038e-01 5.73088266e-02 -1.20007634e+00 -1.86251149e-01
3.08425039e-01 -8.11222851e-01 -6.71257824e-02 -1.84423313e-01
-1.25785708e+00 1.12909722e+00 3.35326433e-01 2.17897788e-01
-2.01971471e-01 5.48484981e-01 2.70766288e-01 4.43474829e-01
1.26837837e-02 6.89523041e-01 -1.36088341e-01 -2.81202286e-01
-4.96796817e-01 6.55486643e-01 9.95189548e-01 7.47388065e-01
5.10441422e-01 3.32074404e-01 -3.62152755e-01 7.51071632e-01
6.14333808e-01 5.70905030e-01 4.45937902e-01 -1.38726652e+00
2.91313052e-01 1.76018327e-01 -6.67656586e-02 -8.49357188e-01
-2.70890027e-01 1.15353838e-01 5.72564080e-02 3.96010801e-02
6.49761677e-01 -3.56288217e-02 -7.08561301e-01 1.90638077e+00
3.06331515e-01 4.32096213e-01 3.02475154e-01 9.03144360e-01
1.52727318e+00 9.22456861e-01 2.61250734e-01 2.06717908e-01
1.80412447e+00 -8.33351076e-01 -6.74569905e-01 -2.31439769e-01
3.17579746e-01 -6.87115014e-01 1.56710625e+00 4.61263806e-01
-1.00125527e+00 -8.40122104e-01 -9.14593279e-01 -3.21607381e-01
-3.41479510e-01 -1.97967783e-01 7.08006084e-01 7.92897999e-01
-1.15078580e+00 -1.69041321e-01 -5.49356163e-01 -4.42858547e-01
2.10183173e-01 -1.10457003e-01 1.53433368e-01 2.18803585e-01
-1.22187448e+00 6.30898774e-01 -4.37374599e-02 -1.19380839e-01
-1.20452273e+00 -7.69509077e-01 -1.02447915e+00 -1.31474912e-01
4.79378939e-01 -1.01608300e+00 1.89386487e+00 -8.19346905e-01
-1.69781089e+00 7.02308774e-01 -3.86908501e-02 -6.41959727e-01
-2.25192562e-01 -2.33734787e-01 -2.54294217e-01 4.99139965e-01
4.27236520e-02 1.02496541e+00 8.61438274e-01 -1.14560592e+00
-6.77347779e-01 -3.60296667e-01 6.46586418e-01 1.13991633e-01
-1.93709910e-01 7.53574371e-02 -8.65007043e-02 -7.99163163e-01
-3.82717431e-01 -7.32194304e-01 7.92970434e-02 7.06627369e-02
-1.76254421e-01 -4.73340720e-01 6.80385530e-01 -6.92058086e-01
9.29676533e-01 -2.13297534e+00 1.88850954e-01 1.19513355e-03
-5.04704900e-02 -1.46167710e-01 -5.76255441e-01 3.34608406e-01
1.62513375e-01 2.40466110e-02 -4.27328557e-01 -3.15319896e-01
3.92221719e-01 2.18885034e-01 -9.68597710e-01 -1.95072487e-01
4.20594364e-01 8.47125649e-01 -9.00723934e-01 -3.66401643e-01
2.22262349e-02 3.26387554e-01 -8.71544898e-01 6.43553257e-01
-8.72215867e-01 2.99055338e-01 -5.40668488e-01 4.11323518e-01
6.45122910e-03 6.48150891e-02 -3.02051723e-01 -1.19787939e-01
1.47607327e-01 7.24459708e-01 -1.00447512e+00 2.05102420e+00
-6.09255493e-01 8.35065484e-01 1.23323470e-01 -1.07517636e+00
8.94101322e-01 4.75205421e-01 1.64268643e-01 -7.35478282e-01
6.79711625e-02 3.89650352e-02 1.37582840e-02 -9.75934744e-01
6.91202104e-01 -3.10547084e-01 -1.99701414e-01 4.78407800e-01
3.91234666e-01 -9.39056516e-01 3.35188627e-01 3.88588935e-01
1.16130364e+00 2.37397775e-01 3.80091462e-03 8.61732811e-02
4.58745450e-01 1.61882058e-01 -3.16887438e-01 9.38106835e-01
-1.66826531e-01 6.95034862e-01 2.15209812e-01 -2.51869500e-01
-7.90919840e-01 -1.31134009e+00 -2.50090435e-02 1.51462293e+00
-1.96964905e-01 -6.71293497e-01 -8.58377576e-01 -3.43717873e-01
-1.25439286e-01 8.87109220e-01 -6.14917934e-01 3.57773006e-02
-3.94534856e-01 -2.93785840e-01 8.28782141e-01 6.11403108e-01
1.95671514e-01 -1.59171391e+00 -1.00074458e+00 1.10809319e-01
-4.40829843e-01 -1.27517581e+00 -2.30245888e-01 7.60555342e-02
-4.04456913e-01 -1.03515482e+00 -4.84989911e-01 -9.29484785e-01
-4.51102927e-02 -1.84026957e-02 1.33710182e+00 -3.38619314e-02
-2.98742682e-01 1.20542371e+00 -6.06959879e-01 -7.83590078e-01
-5.57412684e-01 -1.06679507e-01 -3.40110004e-01 7.07432404e-02
2.92786717e-01 -4.15974438e-01 -3.48085552e-01 -1.40597969e-01
-1.11618221e+00 2.29893830e-02 -7.20912765e-04 3.57822865e-01
5.19342661e-01 -3.43618035e-01 7.45699883e-01 -6.04817212e-01
1.05335534e+00 -4.91989493e-01 -4.61578161e-01 7.82941878e-02
7.25474283e-02 1.40659750e-01 5.48342705e-01 -3.19174796e-01
-1.04928970e+00 -6.31666137e-03 -6.84321880e-01 1.78234430e-03
-5.25901973e-01 5.56193948e-01 1.01080395e-01 1.60950631e-01
8.44023943e-01 1.36078507e-01 -3.10489774e-01 -1.59562245e-01
8.34711254e-01 5.14588714e-01 8.84656131e-01 -8.60034406e-01
5.92788100e-01 4.90128517e-01 -1.86497316e-01 -9.26841259e-01
-8.20106626e-01 -3.44866157e-01 -6.00271896e-02 -5.91114521e-01
1.27672470e+00 -6.50031865e-01 -8.00255418e-01 1.86944038e-01
-1.21160519e+00 -5.09260893e-01 -6.00423813e-01 3.27977091e-01
-9.22694623e-01 3.52635771e-01 -4.02100593e-01 -1.00758934e+00
-1.91722274e-01 -1.05264926e+00 1.16864550e+00 1.50777027e-01
-6.60897076e-01 -6.40970767e-01 3.46441984e-01 7.90700614e-01
4.18018639e-01 1.28412684e-02 1.11028516e+00 -5.72129250e-01
-6.70770764e-01 4.70718682e-01 2.21517310e-01 1.03928633e-01
-2.15908945e-01 -4.92805569e-03 -1.41735625e+00 2.75746316e-01
7.53757432e-02 -9.29940045e-01 9.56001103e-01 2.12473705e-01
1.23186755e+00 -3.66680883e-02 3.75379562e-01 1.13764897e-01
8.39949906e-01 3.19406182e-01 4.90894824e-01 2.51101196e-01
3.84787053e-01 9.55505908e-01 3.41352224e-01 4.68721151e-01
8.40666831e-01 6.65441930e-01 5.84289432e-01 2.76353568e-01
-5.38343310e-01 -4.12352741e-01 4.98388141e-01 7.87763953e-01
2.06129789e-01 -1.80273831e-01 -9.09791768e-01 8.27643216e-01
-1.50151217e+00 -1.07542634e+00 -2.79344052e-01 1.68320239e+00
9.17096078e-01 -2.12745428e-01 3.03948343e-01 3.44267190e-01
9.37459469e-02 3.00965846e-01 -2.27961823e-01 -6.64697170e-01
1.06524527e-01 7.91113317e-01 -5.23054838e-01 6.85338557e-01
-1.05437624e+00 7.23685503e-01 6.75909519e+00 4.70875055e-01
-5.89719176e-01 -1.32347465e-01 1.65130064e-01 2.59717237e-02
-8.12700212e-01 2.07061395e-02 -2.97165602e-01 5.64554557e-02
1.10939980e+00 3.11878830e-01 7.15358377e-01 5.68170786e-01
-1.68953970e-01 -2.08323568e-01 -1.32894886e+00 9.54430223e-01
4.91150856e-01 -1.15980160e+00 2.16273695e-01 -6.34061575e-01
2.63576239e-01 -2.32953563e-01 2.42080986e-01 4.88311052e-01
4.19503599e-01 -1.01447070e+00 1.14420712e+00 6.52319789e-01
2.84671813e-01 -4.24050808e-01 2.66332418e-01 1.56794228e-02
-1.22956049e+00 -1.43138811e-01 1.37017062e-02 -2.71552771e-01
2.50713527e-01 1.15800194e-01 -7.80565023e-01 2.61365980e-01
1.07754147e+00 1.41701579e-01 -6.87168837e-01 8.03812146e-01
-3.64667237e-01 9.58706498e-01 -3.73819292e-01 -2.90022880e-01
5.16207144e-02 1.26236513e-01 4.52944845e-01 9.87605333e-01
3.22840005e-01 1.97298110e-01 -4.98334244e-02 1.15429914e+00
2.53448938e-03 2.70808786e-01 -5.68151653e-01 -1.53016135e-01
2.73026228e-01 9.08407688e-01 -3.64020765e-01 -1.71038285e-01
-6.65809989e-01 5.99849582e-01 1.25578597e-01 3.79296541e-01
-6.38743579e-01 -3.78783166e-01 3.77304584e-01 -1.89153820e-01
3.53163272e-01 -3.17061573e-01 -1.27626568e-01 -8.66558611e-01
1.25936242e-02 -1.11292458e+00 6.80478334e-01 -1.37458706e+00
-1.46884263e+00 5.19134223e-01 1.86857834e-01 -6.32879972e-01
-5.80807090e-01 -6.39234602e-01 -5.18350422e-01 4.46213931e-01
-1.52046955e+00 -1.06928742e+00 -4.83420193e-01 6.33234918e-01
7.51156986e-01 -1.92380235e-01 1.22616708e+00 -1.04532801e-02
1.22847065e-01 1.90523520e-01 -7.43206978e-01 1.55657515e-01
5.41099608e-01 -1.23818195e+00 6.26977384e-01 5.51956058e-01
9.78263617e-01 3.46422821e-01 8.03645909e-01 -2.63694763e-01
-1.51730335e+00 -5.86680830e-01 7.60718703e-01 -7.62455523e-01
9.04225767e-01 -5.46225429e-01 -9.30853367e-01 5.64020157e-01
5.69938064e-01 -2.82543391e-01 1.03847730e+00 5.42902313e-02
-6.90742671e-01 -1.01863891e-02 -7.25365460e-01 5.47523856e-01
8.25727403e-01 -1.02754140e+00 -1.03659892e+00 2.91215628e-02
9.40940619e-01 -3.23177576e-01 -5.69366992e-01 2.93465286e-01
5.25461972e-01 -9.86155987e-01 1.04128563e+00 -6.14641488e-01
4.70938057e-01 -5.39803624e-01 -6.49454653e-01 -1.09536171e+00
1.39432520e-01 -3.74545008e-01 -1.91226140e-01 1.19500530e+00
6.02399051e-01 -1.73381492e-01 5.08172512e-01 2.60177314e-01
-2.58224875e-01 -3.06122541e-01 -1.06886411e+00 -3.67318839e-01
9.53306109e-02 -1.26161480e+00 6.07251704e-01 4.73788768e-01
6.86298609e-02 6.05887532e-01 -9.82790217e-02 2.31982589e-01
2.59540528e-01 2.64441818e-01 9.26461697e-01 -9.97956336e-01
-7.94784009e-01 -3.32856953e-01 -1.61121041e-01 -1.02074003e+00
3.08100134e-01 -9.93941724e-01 3.84927124e-01 -1.56131494e+00
-8.09345767e-02 -1.60602704e-02 8.41758549e-02 3.13580751e-01
-1.47223145e-01 3.46866041e-01 5.56393445e-01 -2.14880258e-01
-5.17524242e-01 6.14903271e-01 1.19347048e+00 -2.94028163e-01
3.36826630e-02 -1.45239502e-01 -7.71400273e-01 5.53604841e-01
9.12158310e-01 -1.45528376e-01 -6.76807463e-01 -7.67557323e-01
5.93206167e-01 1.35736331e-01 7.11701691e-01 -1.05965698e+00
2.66174495e-01 -1.79250259e-02 -1.50621608e-01 -3.75201881e-01
7.94823527e-01 -5.88404775e-01 -2.05184832e-01 6.95002973e-02
-8.06952655e-01 6.49378896e-02 3.44506651e-01 5.55927694e-01
-4.34050560e-01 -4.94693190e-01 3.18796486e-01 -3.10030989e-02
-8.30599248e-01 -1.37108907e-01 -7.63510168e-01 3.98023099e-01
6.50036097e-01 2.31172860e-01 -3.18400174e-01 -1.07654119e+00
-6.23236239e-01 1.69821605e-01 -3.25750917e-01 7.03218579e-01
8.54558289e-01 -1.34614348e+00 -7.11077690e-01 9.08849947e-03
4.99235779e-01 -1.63626567e-01 2.64041066e-01 2.47599080e-01
-3.50076556e-01 3.59322131e-01 1.27931297e-01 -4.59536254e-01
-1.11237514e+00 4.82551843e-01 2.83090115e-01 2.91316032e-01
-4.03416216e-01 9.56484973e-01 2.46889949e-01 -5.02301037e-01
4.56799895e-01 -6.62094474e-01 -4.66918081e-01 5.73028205e-03
6.76767588e-01 3.52213867e-02 -1.05087779e-01 -5.51032484e-01
-1.28884733e-01 4.98156577e-01 4.55325037e-01 -7.81061351e-01
1.01481450e+00 9.54690203e-02 2.51317620e-01 6.43539071e-01
8.04899096e-01 1.39628768e-01 -9.07420695e-01 -3.93748805e-02
-1.82282582e-01 -2.62096137e-01 -1.33908197e-01 -7.74038851e-01
-6.42042398e-01 1.14027262e+00 5.96146166e-01 7.10235715e-01
1.11195338e+00 6.58537626e-01 7.94924438e-01 5.44060111e-01
1.39523298e-01 -1.06322050e+00 8.46528411e-01 6.78361237e-01
1.27927613e+00 -9.18360174e-01 -5.93161464e-01 -1.69195756e-01
-6.00793898e-01 9.93154585e-01 5.39710760e-01 -1.55436061e-02
4.16040629e-01 3.00159186e-01 3.03629100e-01 -3.67016435e-01
-8.37577045e-01 -6.32273138e-01 2.73073077e-01 1.05298293e+00
3.65526855e-01 -2.41154013e-03 1.03527464e-01 8.13856781e-01
-9.20684695e-01 -1.93594888e-01 5.01784921e-01 7.62310863e-01
-6.34931445e-01 -1.06957364e+00 -5.66839278e-01 2.16908470e-01
-3.64715904e-01 -1.93622082e-01 -6.59465790e-01 4.70400214e-01
-6.34664223e-02 1.56402612e+00 5.42601338e-04 -4.01335955e-01
4.57498044e-01 4.33251351e-01 6.39449477e-01 -7.53956199e-01
-5.53122222e-01 -3.32274228e-01 4.43115354e-01 -4.13871884e-01
-7.59516835e-01 -6.80421531e-01 -1.52366757e+00 1.86435372e-01
-2.36024801e-03 2.87624206e-02 7.03403771e-01 8.71084988e-01
6.02801032e-02 7.40286231e-01 7.66568258e-02 -3.37813854e-01
-3.34484696e-01 -9.14691806e-01 -2.38587782e-01 4.40341741e-01
4.28496510e-01 -3.60669672e-01 -2.02424973e-01 4.15818423e-01] | [15.143771171569824, 5.101315975189209] |
f460cc5a-7605-4625-8d7d-f88f691a0539 | identifying-nonlinear-dynamical-systems-from | 2111.02922 | null | https://arxiv.org/abs/2111.02922v3 | https://arxiv.org/pdf/2111.02922v3.pdf | Reconstructing Nonlinear Dynamical Systems from Multi-Modal Time Series | Empirically observed time series in physics, biology, or medicine, are commonly generated by some underlying dynamical system (DS) which is the target of scientific interest. There is an increasing interest to harvest machine learning methods to reconstruct this latent DS in a data-driven, unsupervised way. In many areas of science it is common to sample time series observations from many data modalities simultaneously, e.g. electrophysiological and behavioral time series in a typical neuroscience experiment. However, current machine learning tools for reconstructing DSs usually focus on just one data modality. Here we propose a general framework for multi-modal data integration for the purpose of nonlinear DS reconstruction and the analysis of cross-modal relations. This framework is based on dynamically interpretable recurrent neural networks as general approximators of nonlinear DSs, coupled to sets of modality-specific decoder models from the class of generalized linear models. Both an expectation-maximization and a variational inference algorithm for model training are advanced and compared. We show on nonlinear DS benchmarks that our algorithms can efficiently compensate for too noisy or missing information in one data channel by exploiting other channels, and demonstrate on experimental neuroscience data how the algorithm learns to link different data domains to the underlying dynamics. | ['Daniel Kramer', 'Daniel Durstewitz', 'Georgia Koppe', 'Carlo Tombolini', 'Philine Lou Bommer'] | 2021-11-04 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [ 4.76246864e-01 -4.39086668e-02 -7.02930912e-02 -1.96971223e-01
-7.85621762e-01 -6.45348072e-01 7.39014030e-01 -1.11741468e-01
-3.11903894e-01 6.99270606e-01 2.54888773e-01 -1.74288198e-01
-3.97199124e-01 -1.84445053e-01 -9.84218597e-01 -1.11998367e+00
2.56277502e-01 5.89230895e-01 -2.38579009e-02 -2.19097584e-02
-5.81509946e-03 5.32239795e-01 -1.13300538e+00 1.28678411e-01
6.14380538e-01 8.88793409e-01 3.28206152e-01 6.44626975e-01
1.50938111e-03 4.97563273e-01 -1.33901656e-01 1.22298211e-01
7.85701573e-02 -5.65365732e-01 -5.55933475e-01 -2.58646280e-01
-3.67600173e-02 -3.88127342e-02 -5.72546542e-01 1.06983829e+00
4.95415837e-01 -1.56818688e-01 7.99656332e-01 -1.04987681e+00
-4.43944961e-01 7.48438299e-01 -8.88508782e-02 3.63177985e-01
-1.59836844e-01 1.34019271e-01 7.58442760e-01 -4.81197864e-01
9.90818202e-01 9.33851838e-01 4.45391089e-01 8.27371120e-01
-2.10823393e+00 -2.45732978e-01 -6.40831813e-02 5.71771041e-02
-9.24274445e-01 -6.02548897e-01 1.09871483e+00 -6.73110366e-01
7.74810016e-01 9.43583772e-02 7.43785024e-01 1.66106343e+00
5.19857705e-01 6.63696170e-01 1.22911060e+00 -2.34753191e-01
4.90793824e-01 -1.47907063e-02 1.88789785e-01 1.56490624e-01
-2.94197232e-01 2.84074038e-01 -6.78661287e-01 -2.90526181e-01
9.13296342e-01 3.07959262e-02 -2.80684322e-01 -4.98273343e-01
-1.35005212e+00 5.80030382e-01 9.98950005e-02 6.10441327e-01
-4.60499942e-01 2.33420789e-01 4.38510686e-01 4.14168447e-01
3.75961810e-01 4.29099172e-01 -6.91357255e-01 -1.06021032e-01
-8.93954873e-01 3.41761559e-01 6.77200615e-01 3.31636757e-01
5.02618849e-01 1.11252040e-01 1.94097608e-01 6.19744718e-01
3.06345165e-01 7.82015681e-01 6.78330958e-01 -1.23814154e+00
-1.20703354e-01 3.19785088e-01 -1.25019565e-01 -5.62837481e-01
-6.90686226e-01 -4.57913488e-01 -1.19412553e+00 -1.69126838e-01
5.63834727e-01 5.50407283e-02 -8.33300591e-01 2.27449417e+00
4.39415723e-02 5.85134745e-01 9.29536447e-02 7.38819957e-01
4.84152526e-01 8.14317822e-01 -1.25710011e-01 -8.22159588e-01
9.09095824e-01 -4.29123156e-02 -9.82451737e-01 1.43867716e-01
4.63601917e-01 -3.13705772e-01 6.39782310e-01 4.81715351e-01
-1.38273048e+00 -2.94233590e-01 -6.17696643e-01 -2.30545133e-01
-2.80803025e-01 -9.14767608e-02 1.71988800e-01 -8.65925550e-02
-1.04595423e+00 8.65256548e-01 -1.31509459e+00 -2.33088315e-01
1.89490989e-01 4.23825592e-01 -3.23104143e-01 4.86677259e-01
-1.04896617e+00 9.05438006e-01 1.32392898e-01 3.09430391e-01
-9.48390245e-01 -9.72292960e-01 -4.71277177e-01 -2.03137383e-01
-1.03947394e-01 -7.76253819e-01 1.00291860e+00 -7.84020066e-01
-1.65841722e+00 8.76151204e-01 -3.43968362e-01 -7.04737961e-01
1.46228105e-01 2.66692907e-01 -5.51798403e-01 1.21227562e-01
-1.46165535e-01 5.14856100e-01 1.13784373e+00 -1.17794669e+00
9.75440145e-02 -4.90974009e-01 -3.99842322e-01 -2.23097757e-01
-2.18008578e-01 -2.99636245e-01 -1.06979506e-02 -5.76323688e-01
4.26898271e-01 -1.00547552e+00 -2.68273652e-01 2.15734188e-02
-4.77051169e-01 1.26977507e-02 6.90865099e-01 -6.91677570e-01
1.02258694e+00 -2.07878566e+00 1.13927817e+00 3.21374863e-01
5.36408901e-01 7.37845823e-02 -1.63787872e-01 3.33158582e-01
-4.97461379e-01 -3.09858263e-01 -6.18795633e-01 -4.92980778e-01
-1.83002688e-02 4.56330806e-01 -7.70520091e-01 5.04163146e-01
-3.48741375e-02 1.13288844e+00 -5.65591514e-01 5.05577736e-02
2.24920109e-01 6.30824625e-01 -4.61029530e-01 1.61549836e-01
-4.56675917e-01 1.18954587e+00 -8.59080404e-02 3.02291840e-01
3.71269703e-01 -6.25440478e-01 2.57249206e-01 -6.38916135e-01
-1.84220672e-01 3.03910464e-01 -9.16883945e-01 2.22484064e+00
-3.46747309e-01 7.65300632e-01 3.73490639e-02 -1.78044581e+00
4.50092912e-01 5.97516537e-01 8.57509732e-01 -8.03092420e-01
2.57598728e-01 5.21550953e-01 1.36780217e-01 -6.87148511e-01
-7.88402036e-02 -3.93320709e-01 -9.80049819e-02 5.64550400e-01
5.41985691e-01 -3.21661122e-02 -1.59655467e-01 7.81619996e-02
1.10177934e+00 5.20355105e-02 5.36315069e-02 -2.84984767e-01
1.82869405e-01 -3.10064048e-01 3.85165155e-01 6.07353449e-01
8.19495022e-02 5.01926899e-01 5.70841670e-01 -2.28606462e-01
-1.32160187e+00 -1.32109177e+00 -6.20830894e-01 5.12687385e-01
-2.30000913e-01 -1.07485484e-02 -5.75563192e-01 2.01494515e-01
-2.27543205e-01 4.84839082e-01 -5.68362892e-01 -4.92754519e-01
-4.44158643e-01 -9.22465682e-01 6.08063757e-01 7.39152059e-02
-1.07495263e-01 -1.00226712e+00 -4.24163520e-01 4.30984199e-01
-2.35984966e-01 -1.19423711e+00 -4.97940835e-03 5.11704028e-01
-1.21176505e+00 -4.71924603e-01 -6.48814082e-01 -3.81027669e-01
4.98701513e-01 -2.62721717e-01 7.27557719e-01 -5.50452709e-01
-4.21461552e-01 6.78822935e-01 2.32218742e-01 -1.90053299e-01
-6.85023367e-01 -4.30894941e-02 4.31200504e-01 2.55158514e-01
-3.34685072e-02 -1.32348704e+00 -4.09396321e-01 -9.66768786e-02
-1.34838462e+00 1.78488597e-01 4.22426224e-01 8.34289253e-01
7.37256289e-01 -3.41078281e-01 6.16724670e-01 -6.73163235e-01
4.92912441e-01 -6.74565852e-01 -6.24889195e-01 1.28922179e-01
-4.09247547e-01 6.24925733e-01 5.41966856e-01 -7.37997770e-01
-7.86427319e-01 2.47579142e-01 -1.54965535e-01 -5.73034286e-01
-1.49222195e-01 8.94081652e-01 -9.42819193e-02 6.77327886e-02
7.46919513e-01 6.92713678e-01 3.39460552e-01 -4.16253626e-01
3.29161823e-01 3.26679885e-01 7.09740043e-01 -5.97068667e-01
5.04261017e-01 7.29798377e-01 2.13170454e-01 -9.85057652e-01
-5.34564316e-01 -1.59185335e-01 -6.19359851e-01 -2.16883704e-01
7.64444411e-01 -7.58401930e-01 -7.59077907e-01 7.28465557e-01
-1.15290082e+00 -5.11540651e-01 -4.71762061e-01 5.72295010e-01
-8.97377133e-01 6.52819946e-02 -7.99870193e-01 -6.29068315e-01
1.56724080e-02 -1.03794920e+00 1.10108101e+00 -1.72134954e-02
-1.58632234e-01 -1.41992831e+00 6.00623071e-01 -3.81847508e-02
3.58336836e-01 9.09572244e-02 1.13700891e+00 -4.34035808e-01
-5.48117340e-01 -9.87321511e-02 2.19850853e-01 2.92696714e-01
-8.73887390e-02 9.29186419e-02 -1.17429948e+00 1.65152624e-02
4.57182407e-01 -1.17851667e-01 8.93724859e-01 1.07258153e+00
1.14395559e+00 -4.15756032e-02 -4.20161128e-01 5.21023035e-01
1.29944849e+00 2.96531077e-02 5.23656428e-01 -4.01028931e-01
6.62596464e-01 4.70438510e-01 -4.34171021e-01 1.74438044e-01
2.94199347e-01 7.17863142e-01 2.57238865e-01 2.03792170e-01
3.30330908e-01 -7.33608603e-02 6.07201636e-01 1.35412467e+00
-4.90150340e-02 4.61059473e-02 -7.66971171e-01 5.39575934e-01
-2.04065251e+00 -1.08692122e+00 -1.28005713e-01 2.14994931e+00
9.79145467e-01 1.21844552e-01 9.42487642e-02 1.53912038e-01
4.20515627e-01 -1.70716435e-01 -1.05013299e+00 1.32619709e-01
-5.36271036e-01 1.72279447e-01 2.06588283e-01 3.83811802e-01
-5.38843453e-01 3.65130633e-01 6.43711185e+00 6.53062403e-01
-1.48267853e+00 1.54940188e-01 2.75663465e-01 -1.98084220e-01
-6.59808695e-01 -1.11108106e-02 -4.82546151e-01 4.79992867e-01
1.55332100e+00 -1.76425725e-01 9.03641701e-01 -8.16981867e-02
5.78868449e-01 4.89333086e-02 -1.27932799e+00 1.20799863e+00
-2.84692466e-01 -1.71766233e+00 -1.14644729e-01 1.69389799e-01
5.99813223e-01 2.63064742e-01 4.01285350e-01 -1.19057313e-01
4.12989184e-02 -8.91390979e-01 6.41030669e-01 1.19217169e+00
3.65916282e-01 -1.22670783e-02 3.12370900e-02 7.70789742e-01
-6.59768760e-01 -8.78173858e-03 -9.75802839e-02 3.37420590e-02
4.43646431e-01 9.37931001e-01 -6.50630146e-02 2.81181067e-01
5.04594862e-01 1.21016693e+00 -8.24719220e-02 7.64677048e-01
3.71512920e-01 7.26822615e-01 -6.90430760e-01 4.12012517e-01
-3.28384966e-01 -2.44508833e-01 9.09755051e-01 6.16268098e-01
4.68030572e-01 -2.18099300e-02 -5.72687268e-01 1.30583620e+00
8.28227624e-02 -4.14803565e-01 -6.59494698e-01 -4.70128328e-01
1.40465140e-01 1.18737936e+00 -6.69474602e-01 -7.89842010e-02
-2.20355496e-01 7.63355196e-01 3.18785310e-01 5.90896547e-01
-7.91072488e-01 4.90116954e-01 4.57532078e-01 3.30904499e-03
1.54204352e-03 -5.39985180e-01 -3.71069610e-01 -1.60845661e+00
7.70288333e-03 -5.91939986e-01 1.18426703e-01 -1.00327396e+00
-1.44133759e+00 2.48940632e-01 8.50172788e-02 -1.31078553e+00
-4.62961882e-01 -6.21133089e-01 -3.63366663e-01 7.26706803e-01
-1.16871881e+00 -8.73920500e-01 3.70088071e-01 7.60343850e-01
6.26500323e-02 -6.08402640e-02 9.28835809e-01 4.51253265e-01
-6.15855575e-01 -2.66002547e-02 4.72084552e-01 -3.14125150e-01
5.08643329e-01 -1.15870929e+00 2.05932446e-02 6.34204686e-01
3.34240645e-01 8.42862546e-01 1.06630206e+00 -3.70465577e-01
-1.64258599e+00 -6.18428290e-01 6.42101705e-01 -3.47253442e-01
9.44573879e-01 -5.62856734e-01 -1.20083153e+00 7.91542530e-01
3.44322361e-02 2.06740171e-01 6.67329669e-01 -1.12026729e-01
-2.24433988e-01 -7.80165344e-02 -7.64272630e-01 6.37867928e-01
9.22995150e-01 -1.01333630e+00 -4.29814667e-01 3.21670264e-01
3.84086818e-01 -5.32833040e-01 -1.00731158e+00 1.33961454e-01
6.63541377e-01 -5.65050602e-01 9.70600665e-01 -7.55438626e-01
4.59087044e-01 -3.56824487e-01 -1.75265804e-01 -1.49580240e+00
6.24367967e-02 -7.38770545e-01 -5.52116275e-01 8.04395199e-01
3.62141848e-01 -6.16617024e-01 3.20578724e-01 7.23684490e-01
-7.64538050e-02 -2.91405022e-01 -1.26271260e+00 -4.60395664e-01
2.99962312e-01 -6.92594767e-01 -1.05097495e-01 8.53409410e-01
9.57611278e-02 4.92247760e-01 -4.06198651e-01 1.74369037e-01
7.70700693e-01 1.06284052e-01 2.90444523e-01 -1.41884935e+00
-4.18479383e-01 -5.03329217e-01 -3.80899191e-01 -1.21907270e+00
4.31917906e-01 -1.08875918e+00 3.75956558e-02 -1.08901989e+00
6.13292158e-02 -1.64474487e-01 -3.70853812e-01 3.46480533e-02
3.46349478e-01 1.19916527e-02 -1.42360076e-01 4.79704857e-01
-2.70789176e-01 6.25240624e-01 1.09400082e+00 -5.73458634e-02
-3.10189486e-01 5.09509817e-02 -3.35922062e-01 5.41321814e-01
4.83409792e-01 -6.09284103e-01 -4.43022221e-01 -3.09178382e-01
7.95345783e-01 6.56643450e-01 9.19185519e-01 -7.87653685e-01
5.76902151e-01 1.50372563e-02 1.52011499e-01 -2.78393358e-01
4.26450521e-01 -9.84858215e-01 6.23677373e-01 3.01997364e-01
-7.01570690e-01 -3.51026207e-01 2.47129813e-01 7.31594205e-01
-2.13021055e-01 1.20654970e-01 6.93538547e-01 1.20850772e-01
-2.99886078e-01 3.08229208e-01 -6.42544806e-01 4.92834561e-02
5.20187199e-01 1.82707578e-01 -2.07548097e-01 -3.99918407e-01
-1.42659605e+00 -1.71761826e-01 4.22725733e-03 1.47213116e-01
4.03436303e-01 -1.39759088e+00 -5.06693065e-01 3.51218641e-01
-1.24078140e-01 -2.23252594e-01 7.03310847e-01 1.40679276e+00
2.05348447e-01 3.25676084e-01 -4.19479311e-01 -1.03028524e+00
-7.37657726e-01 3.33937526e-01 6.77071333e-01 -2.35630065e-01
-5.06085932e-01 4.76640195e-01 2.10756674e-01 -5.79186976e-01
-1.18610568e-01 -7.44280994e-01 9.31272376e-03 6.11813515e-02
1.91190258e-01 8.38126019e-02 -1.16185127e-02 -6.81790888e-01
-4.50480310e-03 5.90348840e-01 1.05509691e-01 -5.36644995e-01
1.48287225e+00 -2.79017925e-01 -4.26424623e-01 1.33785117e+00
1.39447284e+00 -3.68534476e-01 -1.35005474e+00 -5.59741735e-01
-1.30764261e-01 4.51717228e-01 9.59320068e-02 -5.71707904e-01
-9.20380056e-01 1.08530366e+00 6.37663245e-01 6.93651736e-01
1.19551551e+00 2.78649867e-01 5.73637486e-01 3.28278303e-01
1.50612295e-01 -1.03719258e+00 -1.00760505e-01 3.95367444e-01
8.04177761e-01 -9.50443208e-01 -3.33345175e-01 8.53521451e-02
-1.87127426e-01 1.26528895e+00 -1.76208794e-01 -3.45678151e-01
1.11740351e+00 4.89691526e-01 -2.28079125e-01 -1.80415362e-01
-1.00404811e+00 1.54868126e-01 3.68215173e-01 3.52157414e-01
7.45573640e-02 -5.37698567e-02 2.23351568e-02 7.35585928e-01
1.54828280e-01 3.43832225e-01 4.69987184e-01 7.39341974e-01
8.78957193e-03 -1.11836529e+00 -2.62483135e-02 6.31980360e-01
-3.02201807e-01 4.55052182e-02 4.94062901e-02 3.80606860e-01
-1.55467778e-01 3.37302417e-01 9.15766358e-02 -7.01910108e-02
4.79463600e-02 3.11425626e-01 7.68310428e-01 -2.08059013e-01
3.37795843e-03 5.01361549e-01 -5.65515757e-01 -6.08793139e-01
-1.07980919e+00 -1.07723069e+00 -1.41790283e+00 -7.98277929e-02
-5.22941127e-02 -2.94984698e-01 7.14124978e-01 1.47998452e+00
4.98259038e-01 6.82588875e-01 2.73762077e-01 -8.97943258e-01
-3.64683867e-01 -7.14586854e-01 -7.49181747e-01 2.42020905e-01
8.97357404e-01 -5.46852767e-01 -5.14957368e-01 5.59019327e-01] | [6.713119983673096, 3.543926477432251] |
49b15fdd-3806-488d-8b0b-20e6a1cb9f80 | referring-expressions-with-rational-speech | 2205.07795 | null | https://arxiv.org/abs/2205.07795v1 | https://arxiv.org/pdf/2205.07795v1.pdf | Referring Expressions with Rational Speech Act Framework: A Probabilistic Approach | This paper focuses on a referring expression generation (REG) task in which the aim is to pick out an object in a complex visual scene. One common theoretical approach to this problem is to model the task as a two-agent cooperative scheme in which a `speaker' agent would generate the expression that best describes a targeted area and a `listener' agent would identify the target. Several recent REG systems have used deep learning approaches to represent the speaker/listener agents. The Rational Speech Act framework (RSA), a Bayesian approach to pragmatics that can predict human linguistic behavior quite accurately, has been shown to generate high quality and explainable expressions on toy datasets involving simple visual scenes. Its application to large scale problems, however, remains largely unexplored. This paper applies a combination of the probabilistic RSA framework and deep learning approaches to larger datasets involving complex visual scenes in a multi-step process with the aim of generating better-explained expressions. We carry out experiments on the RefCOCO and RefCOCO+ datasets and compare our approach with other end-to-end deep learning approaches as well as a variation of RSA to highlight our key contribution. Experimental results show that while achieving lower accuracy than SOTA deep learning methods, our approach outperforms similar RSA approach in human comprehension and has an advantage over end-to-end deep learning under limited data scenario. Lastly, we provide a detailed analysis on the expression generation process with concrete examples, thus providing a systematic view on error types and deficiencies in the generation process and identifying possible areas for future improvements. | ['Sang Chin', 'Elizabeth Coppock', 'Derry Wijaya', 'Fabian Zhafransyah', 'Taufiq Daryanto', 'Hieu Le'] | 2022-05-16 | null | null | null | null | ['referring-expression-generation'] | ['computer-vision'] | [ 2.73661584e-01 7.51440823e-01 1.23209447e-01 -7.45112658e-01
-1.26760876e+00 -3.53639513e-01 8.57839525e-01 -1.62933230e-01
-2.80643761e-01 7.63126075e-01 6.34473026e-01 -1.19035408e-01
-1.21299192e-01 -5.01008928e-01 -6.39946699e-01 -8.17107737e-01
1.16229683e-01 9.20911670e-01 -1.93083972e-01 -3.39343697e-01
8.29240829e-02 2.31132433e-01 -1.25275409e+00 7.96631038e-01
1.51947826e-01 8.99477839e-01 2.31387421e-01 7.58644760e-01
-1.27490684e-01 1.52651441e+00 -8.88734698e-01 -6.45926893e-01
-8.76988247e-02 -4.08543140e-01 -1.31185162e+00 2.41696745e-01
1.51489213e-01 -1.69582188e-01 1.30451903e-01 8.52778912e-01
6.58595920e-01 2.45183423e-01 9.27253425e-01 -1.30059731e+00
-7.01814890e-01 9.78307903e-01 -4.30914581e-01 -1.01678371e-01
7.87359536e-01 2.91998088e-01 1.32749915e+00 -4.29558814e-01
5.58240950e-01 1.80193269e+00 4.67520267e-01 7.78782368e-01
-1.14048719e+00 -3.36898178e-01 1.29738808e-01 2.85347611e-01
-1.15515769e+00 -4.80239868e-01 9.98854637e-01 -3.24650586e-01
1.28882575e+00 6.66137338e-02 3.00958365e-01 1.53790045e+00
-9.40745920e-02 1.42316365e+00 1.50768709e+00 -5.68725407e-01
3.72006685e-01 6.05758727e-02 -3.56423523e-04 6.22792304e-01
-5.63957989e-01 3.15562516e-01 -6.30727410e-01 -2.29568660e-01
3.37224305e-01 -7.43568361e-01 3.82585917e-04 -3.14857006e-01
-1.11192775e+00 1.16827047e+00 7.48989701e-01 2.60753185e-01
-7.77262688e-01 5.74138045e-01 3.98009211e-01 -2.56149340e-02
2.80331641e-01 4.91500407e-01 -4.56822097e-01 -4.05257821e-01
-5.93548536e-01 7.88909793e-01 9.22902286e-01 9.57471967e-01
4.68641907e-01 1.11643456e-01 -5.77521026e-01 9.51469183e-01
7.23585367e-01 1.04334034e-01 2.69982517e-01 -1.43811619e+00
3.38141918e-01 4.19138521e-01 1.72467902e-01 -7.17250109e-01
-4.88210052e-01 -1.80750936e-01 -6.07363641e-01 3.68748277e-01
3.98714602e-01 -2.80294955e-01 -7.96723843e-01 2.05686259e+00
1.86114460e-01 -2.19229102e-01 5.43521702e-01 9.59423602e-01
8.87776732e-01 8.34459305e-01 6.95075691e-01 -1.62011534e-02
1.58786607e+00 -1.05774450e+00 -7.91169584e-01 -3.15374911e-01
6.16026759e-01 -6.00966096e-01 1.09610343e+00 5.21438539e-01
-1.22997952e+00 -3.34143251e-01 -6.48245633e-01 -2.52282917e-01
-2.97731370e-01 1.75577909e-01 1.01477802e+00 4.40027863e-01
-1.13838434e+00 8.08332413e-02 -6.54894054e-01 -2.58743942e-01
7.55467534e-01 3.00054431e-01 -2.18316823e-01 3.09386790e-01
-1.17440903e+00 9.83804107e-01 4.28097337e-01 4.76699769e-02
-1.23811114e+00 -3.06738168e-01 -1.00816882e+00 1.42099753e-01
3.19306016e-01 -8.36324096e-01 2.04063344e+00 -1.48774111e+00
-1.98950946e+00 1.07573545e+00 -2.90733695e-01 -9.35170472e-01
4.61931974e-01 -3.60407606e-02 2.07870677e-01 5.55951893e-02
1.67330042e-01 1.23501825e+00 5.47194600e-01 -1.57318544e+00
-6.55516744e-01 -3.62015545e-01 4.75111604e-01 4.42715317e-01
6.35253727e-01 3.34480554e-01 2.69835871e-02 -4.16877091e-01
-2.63406187e-01 -8.40182483e-01 -1.71088576e-01 -2.77336150e-01
-3.30185622e-01 -1.00109220e+00 4.82489645e-01 -3.94928426e-01
6.73171639e-01 -1.95647740e+00 2.90914297e-01 -2.41135597e-01
1.17350928e-01 9.40384790e-02 3.66339497e-02 4.52621102e-01
-1.49690285e-01 1.43667668e-01 -3.29625010e-01 -7.31857002e-01
4.37963158e-01 2.98244119e-01 -3.94873500e-01 2.64298111e-01
2.16805115e-01 9.93865788e-01 -9.56793606e-01 -5.62869966e-01
3.07200283e-01 4.41962153e-01 -3.98574322e-01 5.42846024e-01
-7.03658640e-01 2.72920370e-01 -6.12164438e-01 5.06414950e-01
2.68553317e-01 -1.57389179e-01 1.00372918e-01 -3.07709388e-02
3.16810250e-01 3.31952602e-01 -7.67159641e-01 1.58274710e+00
-7.88985789e-01 8.44489753e-01 9.48760211e-02 -1.08660400e+00
9.47657764e-01 5.10056198e-01 5.70777580e-02 -5.37035167e-01
3.25914085e-01 6.78828359e-02 6.65752366e-02 -6.85937226e-01
2.16509566e-01 -5.99438310e-01 -5.30737042e-01 3.13663274e-01
-1.02282874e-02 -4.08187270e-01 -1.41887397e-01 2.44783163e-01
9.82765913e-01 6.97197020e-01 7.51848221e-01 2.60623973e-02
5.29261053e-01 2.13332832e-01 2.62722343e-01 9.15760219e-01
-6.31293952e-01 3.85649353e-01 7.93517888e-01 -4.90136653e-01
-8.20873320e-01 -8.63281429e-01 2.65463740e-01 1.38321674e+00
-8.65840539e-02 -9.18035358e-02 -8.05793941e-01 -7.37141430e-01
-4.89807487e-01 1.39852595e+00 -7.53474712e-01 8.62126574e-02
-5.04200757e-01 -6.63461268e-01 7.49583781e-01 6.30483866e-01
7.83431888e-01 -1.79498863e+00 -9.60600555e-01 1.37761176e-01
-4.72197443e-01 -1.27913094e+00 1.95959508e-01 2.00234458e-01
-2.30357111e-01 -8.16734672e-01 -4.56098109e-01 -8.37295413e-01
3.30236733e-01 -1.88063964e-01 1.31943774e+00 -2.22600415e-01
1.33288249e-01 4.78718460e-01 -5.00708878e-01 -7.66767263e-01
-9.12332535e-01 -1.71554551e-01 -3.68132174e-01 -3.72760631e-02
6.55183673e-01 -2.23947555e-01 -2.81169146e-01 -9.57428440e-02
-5.56082368e-01 4.33255851e-01 5.17496467e-01 8.55126977e-01
2.08278075e-01 -1.74214527e-01 6.22198641e-01 -7.19603062e-01
1.05056655e+00 -4.71132040e-01 -4.46190238e-01 1.32797182e-01
-1.16450980e-01 2.22352281e-01 4.49178994e-01 -1.60344243e-01
-1.65261376e+00 3.62284571e-01 -3.40083241e-01 -4.60650437e-02
-6.88411891e-01 2.79484391e-01 -1.99791580e-01 5.05076408e-01
8.06088805e-01 1.26089692e-01 -7.44475722e-02 -1.56805396e-01
5.82750320e-01 8.56992483e-01 4.51205194e-01 -8.30421686e-01
8.24787170e-02 3.82438183e-01 2.40191855e-02 -6.14315093e-01
-1.15322375e+00 -2.36367479e-01 -3.39436829e-01 -2.50335038e-01
1.22022545e+00 -1.00227511e+00 -1.15710235e+00 3.87466997e-01
-1.72113276e+00 -6.26064777e-01 -9.59842056e-02 8.35819542e-02
-1.33322453e+00 4.31793332e-02 -3.75909925e-01 -1.30735409e+00
-2.41235524e-01 -1.26935077e+00 1.61401653e+00 2.74952084e-01
-6.59602523e-01 -1.11188829e+00 -8.47493112e-02 8.90038013e-01
2.11843506e-01 3.55070800e-01 9.92412686e-01 -9.25696552e-01
-5.41928232e-01 1.10099204e-01 -1.15182295e-01 2.04040185e-01
-1.66118100e-01 -3.32138330e-01 -1.15598798e+00 1.53169975e-01
-4.16075694e-04 -9.49884236e-01 5.86732149e-01 4.04822111e-01
1.08766890e+00 -3.29046607e-01 -1.72897264e-01 1.54737487e-01
1.12716985e+00 3.56371462e-01 6.05300426e-01 3.47691923e-01
3.41665387e-01 1.07947540e+00 6.99349225e-01 1.96696803e-01
8.57822061e-01 1.01231205e+00 6.61324859e-01 -7.15917870e-02
-2.83895761e-01 -2.24750683e-01 5.72117627e-01 1.11091253e-03
7.94710964e-02 -5.97600877e-01 -1.02649760e+00 7.23550558e-01
-2.12265849e+00 -1.06511664e+00 8.87204707e-03 1.56469464e+00
8.36769998e-01 -1.28381431e-01 1.85626388e-01 -1.08924806e-01
4.34865147e-01 2.78097451e-01 -2.09400997e-01 -8.85017633e-01
-2.47050539e-01 9.22476873e-02 -5.27616926e-02 8.52141619e-01
-9.94569421e-01 1.45860839e+00 6.15373659e+00 7.06584573e-01
-8.29518974e-01 2.15114340e-01 9.11628067e-01 2.63475239e-01
1.75970029e-02 8.32466502e-03 -6.60762727e-01 -1.98137183e-02
9.98584986e-01 1.24966145e-01 4.85816777e-01 9.71546054e-01
5.63850701e-01 -3.26918513e-01 -1.21395636e+00 1.02958453e+00
3.10774803e-01 -1.23332512e+00 2.34618671e-02 -1.26903743e-01
5.37536442e-01 -6.11571409e-03 -1.68942243e-01 5.81901014e-01
8.48127663e-01 -1.38575780e+00 8.74999046e-01 4.41138744e-01
6.28913566e-02 -5.49085438e-01 9.19967234e-01 5.75858951e-01
-4.93187755e-01 -1.92499414e-01 -3.44912671e-02 -2.79200613e-01
4.49430019e-01 -1.54082179e-01 -1.36234689e+00 2.97099769e-01
7.50716209e-01 2.17179567e-01 -2.30422288e-01 5.71889400e-01
-6.04691863e-01 5.76389670e-01 -1.81401804e-01 -5.33796966e-01
7.48625457e-01 2.45422363e-01 6.28419638e-01 1.28297210e+00
-1.12560980e-01 1.52576536e-01 9.06254258e-03 1.28085661e+00
7.19333664e-02 7.86993727e-02 -5.97364545e-01 1.63485125e-01
1.76672786e-01 1.22866285e+00 -3.81245762e-01 -4.31010455e-01
-6.88896030e-02 7.81827986e-01 6.72492325e-01 3.90531212e-01
-8.59287441e-01 1.49979785e-01 3.30400646e-01 -9.39152241e-02
2.80528516e-01 1.16370812e-01 -1.65664643e-01 -6.36479557e-01
-2.50773996e-01 -1.24858260e+00 3.13168108e-01 -1.45608401e+00
-1.17736423e+00 7.42408335e-01 4.57266271e-01 -7.66821444e-01
-7.92593777e-01 -8.59620094e-01 -4.38032240e-01 7.96834886e-01
-1.35297549e+00 -1.70719099e+00 -2.82356799e-01 4.11359191e-01
1.00566268e+00 -3.10481638e-01 1.03128707e+00 -4.19200540e-01
-1.79630265e-01 1.65963128e-01 -6.23890281e-01 2.71092281e-02
1.52836844e-01 -1.52363575e+00 -3.21530402e-02 3.69611353e-01
4.35921222e-01 4.51691747e-01 1.07019043e+00 -3.01993102e-01
-8.73427629e-01 -7.71486223e-01 1.04827034e+00 -6.47866488e-01
4.69495922e-01 -3.31711560e-01 -6.78778648e-01 9.50883508e-01
8.39367151e-01 -2.91866630e-01 6.03459001e-01 1.70562752e-02
-2.12409031e-02 2.24425554e-01 -1.20569146e+00 7.47506499e-01
7.44855106e-01 -3.83430928e-01 -1.11877632e+00 5.86251915e-01
4.83148128e-01 -5.36287665e-01 -4.36335385e-01 3.16048622e-01
2.35988304e-01 -1.27708995e+00 5.94429016e-01 -8.83424044e-01
5.94282150e-01 -2.53472000e-01 -4.02568668e-01 -1.30580890e+00
-7.38162249e-02 -9.82719421e-01 9.74624306e-02 1.28685415e+00
3.89204174e-01 -2.27417365e-01 7.24283099e-01 5.54167509e-01
-3.16866748e-02 -8.39086592e-01 -1.09834468e+00 -1.25338584e-01
2.52209008e-01 -8.62954557e-01 5.17370880e-01 3.60771984e-01
-1.50943011e-01 8.33150983e-01 -4.62385565e-01 9.21750888e-02
3.75868320e-01 -4.94313836e-02 1.02183998e+00 -7.65199602e-01
-3.44209522e-01 -5.28554857e-01 -3.20630938e-01 -1.26998520e+00
9.02005970e-01 -8.09345722e-01 4.27972764e-01 -1.70698488e+00
1.29459560e-01 -1.28875464e-01 1.45676866e-01 5.46190262e-01
-2.10986622e-02 -1.99213400e-01 2.85502881e-01 4.88974014e-03
-6.59652054e-01 7.48095274e-01 9.91765738e-01 -2.27103546e-01
-2.29422692e-02 2.79580295e-01 -9.46625054e-01 1.09651077e+00
6.48605764e-01 -3.82311463e-01 -5.49484611e-01 -4.00532573e-01
1.93895593e-01 2.30973244e-01 7.32723594e-01 -4.88562971e-01
-1.65421180e-02 -1.59349278e-01 9.01993737e-02 -4.66623276e-01
7.04081833e-01 -7.71283746e-01 -2.18128487e-01 3.49765830e-02
-8.63894165e-01 2.21567936e-02 9.92858484e-02 4.15386230e-01
-2.18167543e-01 -3.38182718e-01 8.26130688e-01 -4.27931458e-01
-9.16954517e-01 -3.51686686e-01 -6.19751573e-01 -1.83391467e-01
1.22847426e+00 -4.21305485e-02 -2.03166887e-01 -1.05521035e+00
-1.05582082e+00 3.19341451e-01 -3.75664011e-02 5.42810202e-01
5.77605247e-01 -1.21351004e+00 -8.21928322e-01 -4.05797303e-01
1.54202640e-01 1.22950271e-01 7.01754764e-02 5.58133721e-01
-5.19998491e-01 5.29184818e-01 8.56831893e-02 -7.65868187e-01
-1.25952899e+00 4.29793566e-01 8.00615251e-01 -4.72095907e-01
-3.67464364e-01 9.51558709e-01 3.38221014e-01 -5.54436386e-01
3.83904606e-01 -1.40103355e-01 -4.34229314e-01 -1.51997328e-01
3.90024781e-01 -3.76109742e-02 -4.12840813e-01 -1.18981445e+00
-2.71881253e-01 2.47837991e-01 -3.36577408e-02 -4.01998878e-01
1.24295425e+00 -3.10949683e-01 -1.21407267e-02 4.25592661e-01
9.93100643e-01 -3.10234219e-01 -1.17968798e+00 -2.19034567e-01
1.69214427e-01 -5.71962819e-02 1.61985695e-01 -1.16264808e+00
-6.86373234e-01 9.12152946e-01 5.23009002e-01 2.06737623e-01
1.02584529e+00 5.80435872e-01 2.79708564e-01 4.69687909e-01
3.48028600e-01 -9.68272150e-01 3.86560917e-01 3.80454630e-01
1.38471377e+00 -1.49871135e+00 -2.80384898e-01 -2.00021699e-01
-1.31913972e+00 1.02131951e+00 8.12871516e-01 -4.55714166e-02
1.23685107e-01 4.09443229e-02 4.98419762e-01 -5.45057893e-01
-1.12059081e+00 -2.58633912e-01 -1.34318113e-01 9.05061901e-01
6.21304929e-01 2.65225857e-01 3.02903038e-02 5.26665747e-01
-5.01037121e-01 2.38406472e-03 4.59426433e-01 6.88661039e-01
-1.22196011e-01 -9.26800966e-01 -2.48363882e-01 -7.69997984e-02
-5.28667748e-01 5.34223951e-02 -7.15457618e-01 9.97002184e-01
4.92563508e-02 1.16079795e+00 -6.01333147e-03 1.62666813e-01
3.70610446e-01 1.81207940e-01 4.30054307e-01 -6.61465466e-01
-6.99347079e-01 8.16900879e-02 5.27473450e-01 -5.52424610e-01
-1.03424048e+00 -7.47369587e-01 -1.53227568e+00 1.43538564e-01
3.72883044e-02 -4.50022258e-02 5.33251405e-01 1.31484210e+00
1.63590372e-01 6.25629306e-01 3.36672217e-01 -9.34079468e-01
-8.77759755e-01 -1.14825583e+00 -1.16684131e-01 7.44818568e-01
2.97521174e-01 -6.41945183e-01 -2.33649835e-01 1.69171691e-01] | [10.789072036743164, 1.6665418148040771] |
7dad1ae1-d6e9-476b-a8a0-e31e608dd7a3 | offline-reinforcement-learning-with-6 | 2305.12679 | null | https://arxiv.org/abs/2305.12679v2 | https://arxiv.org/pdf/2305.12679v2.pdf | Offline Reinforcement Learning with Additional Covering Distributions | We study learning optimal policies from a logged dataset, i.e., offline RL, with function approximation. Despite the efforts devoted, existing algorithms with theoretic finite-sample guarantees typically assume exploratory data coverage or strong realizable function classes, which is hard to be satisfied in reality. While there are recent works that successfully tackle these strong assumptions, they either require the gap assumptions that only could be satisfied by part of MDPs or use the behavior regularization that makes the optimality of learned policy even intractable. To solve this challenge, we provide finite-sample guarantees for a simple algorithm based on marginalized importance sampling (MIS), showing that sample-efficient offline RL for general MDPs is possible with only a partial coverage dataset and weak realizable function classes given additional side information of a covering distribution. Furthermore, we demonstrate that the covering distribution trades off prior knowledge of the optimal trajectories against the coverage requirement of the dataset, revealing the effect of this inductive bias in the learning processes. | ['Chenjie Mao'] | 2023-05-22 | null | null | null | null | ['offline-rl'] | ['playing-games'] | [ 1.36485800e-01 7.32216835e-01 -8.06127965e-01 -4.11503250e-04
-1.17791140e+00 -9.22696054e-01 1.99493334e-01 2.21380249e-01
-3.06598127e-01 1.30055678e+00 6.24695458e-02 -4.83512700e-01
-5.53552270e-01 -9.26570833e-01 -1.56623733e+00 -8.27062845e-01
-2.69037545e-01 8.14089477e-01 -1.23933502e-01 2.42854565e-01
8.02739561e-02 3.88888627e-01 -1.64199269e+00 -1.59380153e-01
9.51388359e-01 1.01498318e+00 2.12349474e-01 6.49486065e-01
7.24000558e-02 6.54633284e-01 -2.91655093e-01 2.37551317e-01
6.50123358e-01 -3.54796350e-01 -4.74700093e-01 1.85092583e-01
6.10441640e-02 -8.18974435e-01 -1.49054527e-01 1.18220329e+00
2.56504446e-01 1.49614364e-01 4.23858643e-01 -1.27569997e+00
-2.15412602e-01 9.16406333e-01 -2.70723790e-01 -1.58671260e-01
1.28112838e-01 2.90405959e-01 1.10357261e+00 -1.04278229e-01
5.88066816e-01 1.03139865e+00 4.27728564e-01 3.78819674e-01
-1.28770173e+00 -5.38880944e-01 4.30088401e-01 -3.92174393e-01
-9.11043286e-01 -1.92023769e-01 4.86591607e-01 -3.18321884e-01
3.42204869e-01 2.18627915e-01 7.30556607e-01 1.07276011e+00
-3.70034277e-01 9.61475432e-01 1.27419806e+00 -1.19845286e-01
7.39435673e-01 4.64333385e-01 1.07013144e-01 5.43866396e-01
8.26787472e-01 4.41820592e-01 -2.41287425e-01 -3.62563014e-01
7.84367561e-01 3.02716970e-01 -4.65771526e-01 -7.89577663e-01
-7.81511724e-01 9.80810940e-01 6.29449412e-02 -1.89618133e-02
-3.42147440e-01 3.42559338e-01 1.37479648e-01 6.05273247e-01
3.86304140e-01 4.82630193e-01 -6.40082002e-01 -3.26865375e-01
-7.93857157e-01 5.32965779e-01 1.06816792e+00 1.34363008e+00
7.40928710e-01 -6.08606450e-02 -3.38744372e-01 1.25455279e-02
-1.21534184e-01 9.09511745e-01 -2.12952062e-01 -1.10503960e+00
7.55563498e-01 4.01185304e-01 1.03693557e+00 -5.14596403e-01
-1.55939981e-01 -6.58548355e-01 -1.30582258e-01 2.25260720e-01
1.07553399e+00 -4.57214117e-01 -6.06816173e-01 2.09654784e+00
4.75953221e-01 -7.00573325e-02 -2.76978426e-02 7.86164105e-01
-3.40414941e-01 6.22033000e-01 -4.85538512e-01 -7.15517342e-01
7.71829844e-01 -4.79632705e-01 -5.81111968e-01 -4.45531830e-02
7.25003123e-01 2.53570855e-01 1.50366640e+00 4.62751895e-01
-1.36530960e+00 6.79932684e-02 -1.03935182e+00 3.36005390e-01
1.19201779e-01 -1.78991228e-01 6.86360776e-01 7.01966047e-01
-5.92882693e-01 6.52213991e-01 -9.86155272e-01 -1.11159883e-01
7.16981232e-01 2.70339012e-01 6.18096534e-03 -2.78129548e-01
-6.46013081e-01 3.64330381e-01 2.47860029e-01 -3.48870337e-01
-1.67899144e+00 -1.04414463e+00 -3.46971929e-01 2.71663308e-01
1.15961444e+00 -3.35863441e-01 1.32188570e+00 -8.95211399e-01
-1.25785100e+00 2.45213419e-01 -8.35761279e-02 -8.44150066e-01
1.28165078e+00 -3.27131897e-01 4.04728830e-01 2.85034031e-01
-2.62297094e-01 3.33807617e-02 9.05256391e-01 -1.28977823e+00
-8.02320004e-01 -3.74367982e-01 7.62364686e-01 9.75604802e-02
-3.38915348e-01 -8.13126147e-01 9.79019627e-02 3.86324674e-02
-3.84676397e-01 -8.93121302e-01 -4.09733564e-01 8.70575197e-04
-3.65636110e-01 -4.42990139e-02 3.78027469e-01 -3.98617148e-01
1.04077697e+00 -1.92712963e+00 -8.30119029e-02 3.14617574e-01
-1.50119781e-01 -2.95746416e-01 1.91946495e-02 6.37191057e-01
4.90039319e-01 2.32746273e-01 -2.63450712e-01 -3.28959376e-01
3.42055261e-01 3.71806055e-01 -8.11052620e-01 8.54721248e-01
-2.47953966e-01 6.79964066e-01 -1.04608929e+00 -7.07436875e-02
-3.55969071e-02 -1.02972917e-01 -8.72505546e-01 2.18207866e-01
-9.28170860e-01 7.29898155e-01 -7.39691198e-01 3.57249916e-01
5.96605957e-01 -4.04924154e-01 5.10353386e-01 2.91145533e-01
-8.26252066e-03 1.83404177e-01 -1.31531465e+00 1.41530585e+00
-5.52436531e-01 2.74364427e-02 4.29027528e-01 -1.29855502e+00
3.36146563e-01 1.24271624e-01 7.17317045e-01 -2.51712233e-01
1.09867807e-02 3.58400375e-01 -2.99846917e-01 -5.44116557e-01
1.15516469e-01 -2.68418759e-01 4.84415852e-02 6.79248095e-01
-2.13293359e-01 1.71959534e-01 -3.16664055e-02 -7.55100697e-02
1.22040486e+00 3.70178759e-01 2.27164030e-01 -4.76113766e-01
-1.18060604e-01 2.15570077e-01 6.84696555e-01 1.30179882e+00
9.53752324e-02 2.90298641e-01 1.02536190e+00 -6.72076792e-02
-1.31317258e+00 -9.49225247e-01 4.66307998e-02 8.32804203e-01
1.18546382e-01 -5.06613357e-03 -7.63766527e-01 -8.84278715e-01
2.95438737e-01 9.41586912e-01 -9.68078434e-01 5.16887829e-02
-3.05435598e-01 -5.02894878e-01 4.13513966e-02 3.66275668e-01
3.05359419e-02 -4.85628366e-01 -8.43037486e-01 2.06204578e-01
1.70839369e-01 -9.23572242e-01 -4.04843658e-01 3.26600730e-01
-9.57664728e-01 -1.24337447e+00 -5.45919359e-01 -1.96460679e-01
9.26998794e-01 1.42257288e-01 8.41860712e-01 -3.38735014e-01
5.03389239e-02 7.74495006e-01 -6.20262362e-02 -5.15006840e-01
-8.81216824e-02 1.81616738e-01 5.79389818e-02 -1.44021809e-01
-1.41874626e-01 -7.60560632e-01 -6.83074892e-01 -3.32768224e-02
-8.44088674e-01 -1.52488992e-01 6.09895647e-01 5.82594573e-01
8.15008461e-01 3.28425318e-01 7.28158355e-01 -1.05920601e+00
3.32421958e-01 -7.76907861e-01 -1.34348392e+00 3.29309553e-01
-7.62790620e-01 4.55182999e-01 8.25406551e-01 -7.16502249e-01
-8.00200880e-01 8.22043791e-02 3.42173994e-01 -6.51385427e-01
4.45410125e-02 2.33558699e-01 -3.90879631e-01 2.41911814e-01
5.55411816e-01 3.86760890e-01 1.83930412e-01 -4.67593104e-01
3.12598288e-01 2.29909316e-01 1.84569433e-01 -1.16178250e+00
8.07734191e-01 9.37586904e-01 2.99408436e-01 -6.65394485e-01
-1.20969832e+00 -9.87618864e-02 -2.39612892e-01 -8.83906614e-03
4.33977604e-01 -7.79613793e-01 -1.10173309e+00 -2.68471450e-01
-6.17596388e-01 -6.87248170e-01 -1.00980365e+00 4.55908537e-01
-1.04954195e+00 5.58391139e-02 -8.16123262e-02 -1.65513623e+00
1.73451543e-01 -8.44858050e-01 7.08051443e-01 -6.55884892e-02
3.38717461e-01 -9.95921969e-01 -1.01290934e-03 5.37961051e-02
3.19865018e-01 5.52374244e-01 8.73959303e-01 -6.16962194e-01
-1.11426032e+00 -1.16427936e-01 1.92415759e-01 1.19222119e-01
3.71605940e-02 -4.13775563e-01 -8.40820491e-01 -5.91805875e-01
8.88707936e-02 -6.85400546e-01 7.81879127e-01 7.45398164e-01
1.37298989e+00 -1.11603582e+00 -4.86962974e-01 5.31815529e-01
1.52284276e+00 6.10297360e-02 2.63063282e-01 -1.84920244e-02
2.80984819e-01 6.68086529e-01 9.32663441e-01 1.04635262e+00
2.17285886e-01 2.13962287e-01 5.74326396e-01 3.83374065e-01
5.67031860e-01 -8.84386003e-01 3.93083334e-01 7.22918846e-03
1.05955563e-01 -2.30253175e-01 -3.35311204e-01 7.22123265e-01
-1.87432528e+00 -9.14247513e-01 3.72349799e-01 2.87580824e+00
1.02990127e+00 2.19516665e-01 5.67279696e-01 -3.47873494e-02
3.57225269e-01 -7.82603249e-02 -1.26084805e+00 1.19211413e-02
3.20039801e-02 -5.20126075e-02 1.23662496e+00 7.34599411e-01
-5.97552598e-01 3.47758710e-01 6.22913837e+00 8.88104320e-01
-8.19714963e-01 2.94473022e-01 5.12644887e-01 -6.25367165e-01
-8.60755146e-01 3.24927449e-01 -8.97301495e-01 5.31311989e-01
9.49228704e-01 -2.97440320e-01 8.89221132e-01 1.15357089e+00
3.31548363e-01 -1.91105813e-01 -1.36999941e+00 6.15103304e-01
-5.32526851e-01 -1.11923659e+00 -3.58104050e-01 6.82015181e-01
8.13067555e-01 -1.87374890e-01 1.30419597e-01 3.73234361e-01
7.08865285e-01 -1.01254070e+00 7.66265929e-01 6.08175278e-01
8.45046639e-01 -1.07022583e+00 2.53266603e-01 8.80706728e-01
-8.44619036e-01 -5.94560266e-01 -4.69454914e-01 -1.29757121e-01
2.12838454e-03 8.39153171e-01 -7.08818376e-01 3.48840564e-01
1.59277499e-01 4.06599492e-01 2.63205469e-01 7.71788597e-01
-7.88776949e-03 9.61904526e-01 -8.10107887e-01 -1.06321372e-01
1.43361121e-01 -3.20836246e-01 6.33295238e-01 5.98942101e-01
4.26710576e-01 -1.71416268e-01 5.37886798e-01 1.05082023e+00
-5.04829511e-02 -2.12328404e-01 -7.68707871e-01 -5.78113377e-01
6.12057686e-01 9.02461648e-01 -4.56461370e-01 -1.16767399e-01
-1.75290138e-01 4.50254738e-01 4.32654113e-01 4.04043913e-01
-1.00122595e+00 1.67708755e-01 7.07774401e-01 4.09309149e-01
5.42730749e-01 -2.83418238e-01 -3.14363390e-01 -9.80691314e-01
2.77058333e-01 -7.37539649e-01 5.02034187e-01 8.58660191e-02
-1.08755434e+00 -3.55471075e-01 2.14896113e-01 -1.03119147e+00
-2.92096019e-01 -2.51501918e-01 -2.29492143e-01 3.88493925e-01
-1.47717988e+00 -8.25129807e-01 1.75303102e-01 4.45080310e-01
3.41059685e-01 3.18051666e-01 3.72818768e-01 -6.08154163e-02
-2.95472622e-01 6.02434814e-01 4.84819502e-01 -5.81685364e-01
2.76366830e-01 -1.42114329e+00 -3.19449186e-01 4.65347052e-01
-2.38720968e-01 3.79144698e-01 9.34425652e-01 -6.47266090e-01
-2.00274301e+00 -1.09750700e+00 4.19727005e-02 -4.66146052e-01
6.33008063e-01 -6.77687883e-01 -6.56790376e-01 7.96041906e-01
-3.29952121e-01 1.57070220e-01 1.03085086e-01 1.17220722e-01
-6.67799711e-02 -3.19647521e-01 -1.45271599e+00 3.56747508e-01
1.29186893e+00 -2.04852283e-01 5.45582268e-03 5.38300931e-01
1.06448233e+00 -2.16831267e-01 -8.59901488e-01 2.31737316e-01
5.42970717e-01 -5.90418816e-01 7.17727721e-01 -9.86323833e-01
1.55370757e-01 -1.34516627e-01 -3.11881870e-01 -1.00435448e+00
2.02085570e-01 -9.91076469e-01 -7.16005087e-01 1.12318921e+00
4.44533527e-01 -7.27797329e-01 7.72054493e-01 5.93119562e-01
1.24902405e-01 -9.24843192e-01 -9.16797519e-01 -1.16326988e+00
3.75280172e-01 -5.11527479e-01 6.84402943e-01 6.09750032e-01
1.72680598e-02 -1.43818140e-01 -6.68123484e-01 3.27324212e-01
7.66347945e-01 4.58012104e-01 8.22279394e-01 -1.01083803e+00
-9.26481307e-01 -9.74240378e-02 4.65825945e-01 -1.24105346e+00
1.88254073e-01 -5.43813109e-01 1.37948126e-01 -1.30205333e+00
2.55131274e-01 -9.45402205e-01 -2.73797009e-02 2.74578065e-01
1.01927437e-01 -7.33624160e-01 4.18620259e-02 1.55767068e-01
-7.50184894e-01 6.65330052e-01 1.40545654e+00 4.70707379e-02
-3.63243014e-01 3.79549444e-01 -8.37673485e-01 4.56313252e-01
6.40454054e-01 -5.13100386e-01 -9.94120896e-01 -6.39357492e-02
5.76643050e-01 5.32239497e-01 1.74375385e-01 -8.11541498e-01
3.13603389e-03 -6.62133276e-01 -1.14428259e-01 -6.04643524e-01
2.41208658e-01 -9.65944767e-01 1.07693300e-01 6.67971313e-01
-6.87900901e-01 -5.78855038e-01 -1.85716301e-01 1.21922922e+00
5.13301671e-01 -3.26090038e-01 6.84407413e-01 -1.66044414e-01
2.76252747e-01 7.22963274e-01 -7.76994303e-02 5.62128484e-01
1.02091658e+00 1.17392875e-01 -1.73954785e-01 -7.93251872e-01
-4.41969126e-01 4.12181884e-01 5.74042618e-01 -2.98499286e-01
2.02242345e-01 -9.86539125e-01 -4.90194619e-01 -1.22007370e-01
-1.92624107e-01 3.85512769e-01 1.25478759e-01 9.27736998e-01
2.25484580e-01 3.80219549e-01 2.35361740e-01 -3.32483321e-01
-5.34850478e-01 9.96975362e-01 2.60636002e-01 -5.31713724e-01
-7.52515078e-01 3.44161510e-01 2.50740886e-01 -2.47016236e-01
4.30848241e-01 -7.24857271e-01 3.28887552e-01 -7.18299523e-02
5.35696566e-01 5.20807147e-01 -3.10766637e-01 4.45105344e-01
7.08291680e-02 1.47969484e-01 1.54127494e-01 -4.17041898e-01
1.37873983e+00 -2.99744904e-01 5.22437572e-01 7.26466000e-01
9.01661098e-01 3.02581131e-01 -2.14152098e+00 -2.73281306e-01
6.47255406e-03 -4.78251338e-01 -2.38231853e-01 -6.52937174e-01
-8.31797838e-01 7.86744833e-01 3.32636118e-01 3.43220800e-01
8.81386280e-01 3.70494127e-02 6.74634218e-01 6.97493732e-01
8.75657678e-01 -1.11754489e+00 -2.88437515e-01 2.14468911e-01
6.03183806e-01 -1.02330065e+00 -7.35288784e-02 -7.44146947e-03
-4.93531108e-01 7.81705737e-01 2.11323068e-01 -3.65106910e-01
4.62100416e-01 4.24270809e-01 -8.45158577e-01 2.42888272e-01
-1.03916001e+00 -3.62865329e-01 -2.66580164e-01 5.30719161e-01
-2.97129810e-01 1.74089476e-01 -2.48238072e-01 8.67969930e-01
-5.82120195e-02 1.77005768e-01 4.14927244e-01 9.59789574e-01
-8.46674919e-01 -8.19928944e-01 -1.76358640e-01 6.86848402e-01
-6.91245556e-01 2.96466440e-01 -1.61652118e-02 9.60295081e-01
-2.21655726e-01 8.43154907e-01 8.84649809e-03 1.08426608e-01
4.76816595e-02 -1.88785270e-01 7.71395683e-01 -3.65281433e-01
-4.10383865e-02 7.30246827e-02 1.20926104e-01 -6.74100637e-01
-6.70585036e-02 -7.24047124e-01 -1.24023438e+00 -3.68556798e-01
-2.69378543e-01 3.12781364e-01 4.35993314e-01 1.22772801e+00
3.17637056e-01 1.95204243e-01 8.27795923e-01 -4.05023783e-01
-1.44469333e+00 -9.31355774e-01 -8.91178489e-01 1.28132135e-01
7.86967576e-01 -6.73623741e-01 -7.42884934e-01 -3.69255066e-01] | [4.374209403991699, 2.8535308837890625] |
aa9de4a5-6c59-4272-912c-61fa705bfd1c | what-can-an-accent-identifier-learn-probing | 2306.06524 | null | https://arxiv.org/abs/2306.06524v1 | https://arxiv.org/pdf/2306.06524v1.pdf | What Can an Accent Identifier Learn? Probing Phonetic and Prosodic Information in a Wav2vec2-based Accent Identification Model | This study is focused on understanding and quantifying the change in phoneme and prosody information encoded in the Self-Supervised Learning (SSL) model, brought by an accent identification (AID) fine-tuning task. This problem is addressed based on model probing. Specifically, we conduct a systematic layer-wise analysis of the representations of the Transformer layers on a phoneme correlation task, and a novel word-level prosody prediction task. We compare the probing performance of the pre-trained and fine-tuned SSL models. Results show that the AID fine-tuning task steers the top 2 layers to learn richer phoneme and prosody representation. These changes share some similarities with the effects of fine-tuning with an Automatic Speech Recognition task. In addition, we observe strong accent-specific phoneme representations in layer 9. To sum up, this study provides insights into the understanding of SSL features and their interactions with fine-tuning tasks. | ['John H. L. Hansen', 'Okim Kang', 'Ram C. M. C. Shekar', 'Mu Yang'] | 2023-06-10 | null | null | null | null | ['prosody-prediction'] | ['natural-language-processing'] | [ 8.58604684e-02 1.48084044e-01 -4.15547192e-01 -5.92758954e-01
-7.44877696e-01 -8.11568677e-01 5.81579149e-01 8.02633986e-02
-3.84845287e-01 3.12547445e-01 9.11516011e-01 -4.89471033e-02
1.28749490e-01 -3.80483776e-01 -5.98635674e-01 -5.04876912e-01
1.26988798e-01 2.10299343e-01 -9.94179398e-02 -4.10251647e-01
7.54900947e-02 3.03812265e-01 -1.25121009e+00 6.84613585e-01
4.49554563e-01 1.09386301e+00 2.11788177e-01 4.36329156e-01
-2.82033682e-01 7.37318099e-01 -5.02119303e-01 -2.66391844e-01
-9.73811299e-02 -3.02046865e-01 -8.89489770e-01 -2.33782381e-01
5.74773252e-01 2.07705498e-01 3.73318009e-02 9.07934844e-01
5.71403325e-01 1.14382699e-01 4.16538179e-01 -6.33389413e-01
-6.41400099e-01 1.43820298e+00 -2.30308369e-01 5.85973799e-01
9.56047028e-02 6.56931698e-02 1.29071045e+00 -1.03114593e+00
2.84613788e-01 1.41763127e+00 1.08119452e+00 5.74144304e-01
-1.50847042e+00 -8.67403448e-01 3.86884034e-01 5.33558019e-02
-1.16518164e+00 -9.57853198e-01 1.07688701e+00 -7.22493887e-01
1.07369065e+00 1.52049512e-01 3.05918306e-01 1.23639393e+00
5.77326678e-02 5.27369618e-01 1.31552494e+00 -5.20846307e-01
1.43159196e-01 5.02203286e-01 4.23606426e-01 3.86401594e-01
-4.34663862e-01 5.96070647e-01 -1.11588049e+00 1.49811447e-01
6.58782542e-01 -5.98334849e-01 -1.45637408e-01 -1.50995433e-01
-1.16818988e+00 6.97656453e-01 1.93845853e-01 6.22682929e-01
-3.73537838e-01 -7.04645365e-02 6.54832661e-01 3.25342268e-01
5.13049960e-01 8.65203321e-01 -1.13661635e+00 -2.89738744e-01
-9.70850587e-01 -3.54892761e-01 5.81754506e-01 3.84945422e-01
7.89351523e-01 4.15211797e-01 -2.93696254e-01 1.42319858e+00
2.15734154e-01 2.95204036e-02 9.53270376e-01 -6.87782347e-01
4.24015462e-01 -2.42621042e-02 -3.49012315e-01 -5.45922041e-01
-5.56207716e-01 -8.88193846e-01 -5.21016240e-01 -8.34007189e-03
3.07493865e-01 -2.34368011e-01 -3.04361582e-01 2.35543251e+00
-1.74887925e-01 -6.62154332e-02 -2.13459238e-01 5.45432687e-01
5.18563688e-01 5.53658247e-01 5.61513543e-01 -4.30307269e-01
1.33091831e+00 -1.07323360e+00 -8.85208666e-01 -3.89476508e-01
5.73613405e-01 -8.40324044e-01 1.53003442e+00 1.67487726e-01
-1.22027862e+00 -1.29994571e+00 -9.78182495e-01 1.00345314e-01
-5.07667720e-01 1.24320246e-01 4.02935475e-01 8.59418452e-01
-1.04248750e+00 6.20997190e-01 -6.22635067e-01 -2.49466881e-01
7.69494250e-02 2.71174103e-01 -1.99423552e-01 6.38343751e-01
-1.38512135e+00 1.02266026e+00 4.78761196e-01 -1.73415780e-01
-5.13333201e-01 -1.29350543e+00 -8.62944782e-01 3.13201725e-01
-2.23677307e-01 -3.30266178e-01 1.48031104e+00 -1.09367633e+00
-1.78090620e+00 1.10036397e+00 -2.15898842e-01 -4.81240064e-01
-3.57315913e-02 -2.16442823e-01 -6.19736433e-01 -4.55222964e-01
-1.88409135e-01 7.03882694e-01 7.92574823e-01 -1.02366197e+00
-3.80676597e-01 -3.70923549e-01 -3.67627680e-01 1.38017088e-01
-3.99317622e-01 2.01142639e-01 1.76755831e-01 -1.22483921e+00
-1.15704440e-01 -7.05370426e-01 2.35153005e-01 -7.32257128e-01
-3.24186713e-01 -2.82106459e-01 1.83843225e-01 -7.64280796e-01
1.55032027e+00 -2.57132649e+00 -1.21740162e-01 -5.74851446e-02
-2.00193614e-01 2.85738498e-01 -3.81079465e-01 4.04542625e-01
-5.63209295e-01 2.10414290e-01 2.24299449e-02 -5.31107187e-01
3.29102516e-01 -1.35164112e-01 -7.02115834e-01 -1.51592260e-02
4.15299714e-01 8.79489124e-01 -4.60389555e-01 -1.02238342e-01
2.33480483e-01 3.55364442e-01 -7.21109748e-01 1.81107804e-01
-1.01477578e-01 5.26750326e-01 2.43883461e-01 2.53901929e-01
5.72496653e-01 2.52505124e-01 2.60718763e-01 -5.90184629e-01
-5.26819527e-01 1.10408998e+00 -7.36859858e-01 1.55463648e+00
-8.62638056e-01 6.62529171e-01 1.77086070e-01 -7.84833908e-01
1.17427862e+00 3.17837030e-01 9.15773585e-02 -9.14971709e-01
-9.11266729e-02 8.59730542e-02 2.12861747e-01 -3.50506678e-02
3.41282755e-01 -6.88258767e-01 -2.99997866e-01 2.74315447e-01
4.68184024e-01 1.44933671e-01 -3.43875349e-01 -3.54990721e-01
6.95246041e-01 -9.41507444e-02 4.74943221e-01 -6.03166401e-01
5.33495247e-01 -3.53307426e-01 6.95771873e-01 6.89263165e-01
-4.21758294e-01 3.70021492e-01 1.31030247e-01 -1.10737666e-01
-6.67611659e-01 -1.44001365e+00 -3.11693996e-01 1.91615701e+00
-4.76603955e-01 -5.97796559e-01 -8.61471295e-01 -3.34184527e-01
-6.94419593e-02 7.97808409e-01 -6.58111751e-01 -5.08827031e-01
-7.23938704e-01 -5.11972487e-01 6.55805886e-01 7.41773367e-01
4.35059428e-01 -1.50820458e+00 3.00133578e-03 4.11578447e-01
-1.58745036e-01 -1.07312262e+00 -7.45521903e-01 6.78902209e-01
-8.33400309e-01 -4.26871806e-01 -3.05127650e-01 -1.17185199e+00
-9.72132832e-02 -4.00349230e-01 1.24698675e+00 -5.40439606e-01
1.14139676e-01 3.09019327e-01 -1.00701645e-01 -2.22264394e-01
-7.09894478e-01 6.64514005e-01 3.29152703e-01 1.07531406e-01
4.92974401e-01 -6.27640784e-01 -1.46976337e-01 2.99002469e-01
-4.49738592e-01 -1.08049594e-01 5.49044609e-01 7.69577742e-01
5.75594008e-01 6.37405459e-03 1.10231471e+00 -7.52700388e-01
7.26363361e-01 -2.90419608e-01 -2.42456973e-01 3.45122479e-02
-5.74297488e-01 3.02211851e-01 5.91506958e-01 -3.98177475e-01
-1.33786714e+00 8.44474137e-02 -4.75563645e-01 -1.11876734e-01
-4.43774611e-01 4.49732929e-01 -4.32221591e-01 2.66408980e-01
6.65316761e-01 3.49477232e-02 -1.96661711e-01 -1.07200217e+00
4.63773400e-01 6.06523693e-01 8.57917368e-01 -5.52654564e-01
6.11452699e-01 -9.83887911e-02 -8.25105906e-01 -8.53538156e-01
-1.19453728e+00 -3.70354831e-01 -7.73142934e-01 9.98371914e-02
8.57515931e-01 -1.01217365e+00 -4.77975428e-01 7.13093519e-01
-9.40723121e-01 -7.78899968e-01 -5.78996778e-01 5.72144091e-01
-7.84564912e-01 1.54219379e-04 -8.35887015e-01 -6.30058169e-01
-1.47664815e-01 -9.06712055e-01 7.55368233e-01 -4.35806476e-02
-8.24562430e-01 -1.30819821e+00 4.81153011e-01 3.69298339e-01
6.88338459e-01 -4.00209785e-01 1.23081136e+00 -8.46549392e-01
-2.59996764e-02 6.17361665e-01 9.62509513e-02 6.18383408e-01
3.56027544e-01 -4.59901631e-01 -1.70931709e+00 -1.14592724e-01
3.39663774e-01 -3.01065892e-01 1.15387189e+00 6.31104052e-01
1.15818894e+00 -3.71912539e-01 1.03101984e-01 8.17346036e-01
8.78892362e-01 4.84704636e-02 2.94621289e-01 3.15274239e-01
3.86537492e-01 7.87889659e-01 4.87257183e-01 2.61670291e-01
2.42181271e-01 8.56604278e-01 -1.65311471e-01 -8.07178244e-02
-6.79817259e-01 -6.97097063e-01 6.81694150e-01 1.33428526e+00
4.23027366e-01 2.53564358e-01 -8.64637196e-01 5.57731569e-01
-1.37264168e+00 -8.60487998e-01 4.75340843e-01 2.17800260e+00
1.45127511e+00 3.21421653e-01 1.98735625e-01 2.58495301e-01
7.96582937e-01 5.05444527e-01 -4.23703134e-01 -9.24177051e-01
-2.51177788e-01 4.86644417e-01 1.14117458e-01 8.21831346e-01
-1.13870716e+00 1.40866518e+00 7.03733635e+00 7.97309220e-01
-1.32033181e+00 1.01344086e-01 4.37714428e-01 1.38101116e-01
-3.34241152e-01 -4.00279135e-01 -9.94467795e-01 3.62059861e-01
1.34497488e+00 1.39507398e-01 7.30614305e-01 7.06561804e-01
2.62874186e-01 3.93042684e-01 -1.47019851e+00 9.68490362e-01
-1.74623340e-01 -1.45634866e+00 -6.43308684e-02 -1.51331663e-01
4.43832725e-01 2.52349168e-01 3.89296353e-01 7.37693250e-01
1.74509659e-01 -9.85354125e-01 9.18497622e-01 5.03771663e-01
8.29866350e-01 -6.48137331e-01 4.08766389e-01 8.37481469e-02
-1.13921499e+00 -1.54349089e-01 8.29117596e-02 -2.23346189e-01
1.49405316e-01 3.99146020e-01 -1.09119201e+00 -3.20428908e-01
7.09082663e-01 4.48297799e-01 -5.02769411e-01 5.57334542e-01
-8.72851983e-02 1.15883541e+00 1.64336428e-01 4.24202085e-01
-1.27736449e-01 1.64985657e-01 3.40620190e-01 1.77187109e+00
-1.30319566e-01 -3.75324816e-01 -2.32115999e-01 9.57948506e-01
-1.42203227e-01 3.67579199e-02 -1.23539411e-01 -2.72715211e-01
8.51495683e-01 9.47536707e-01 -2.14915514e-01 1.78509916e-03
-2.09376469e-01 7.36605346e-01 6.25380099e-01 2.69135803e-01
-5.47914326e-01 -2.64604669e-02 1.32637596e+00 1.39869871e-02
4.17002559e-01 -2.55683869e-01 -6.95123792e-01 -8.28205347e-01
-4.23999816e-01 -8.14825714e-01 1.55601367e-01 -6.91806495e-01
-1.57048583e+00 4.94837105e-01 -3.30640137e-01 -4.01297808e-01
-5.73406935e-01 -5.89342594e-01 -7.28772044e-01 1.03382719e+00
-1.62899494e+00 -8.16601634e-01 1.52859300e-01 5.65814495e-01
7.24118710e-01 -3.69864494e-01 1.22562861e+00 2.09793657e-01
-6.73085392e-01 1.12816715e+00 -6.60132393e-02 3.17475200e-01
9.89441872e-01 -1.26080728e+00 6.90111279e-01 4.00778472e-01
3.45269084e-01 7.02902019e-01 4.38800633e-01 -2.04203054e-01
-6.07505560e-01 -1.00024784e+00 1.27827978e+00 -5.21341443e-01
8.44106138e-01 -6.14165604e-01 -1.04941654e+00 6.69926405e-01
3.28765988e-01 -2.10056603e-01 1.08768785e+00 7.17761993e-01
-7.39766598e-01 -2.95740694e-01 -7.62219667e-01 2.22564414e-01
1.07009852e+00 -1.31703675e+00 -1.09503400e+00 -3.28578621e-01
1.04249358e+00 8.27720016e-02 -9.23261881e-01 4.19251055e-01
5.21503448e-01 -1.06636870e+00 1.02753210e+00 -7.86884546e-01
-2.49698269e-03 1.69697732e-01 -4.82349098e-01 -1.80381858e+00
-8.06083739e-01 -5.29895306e-01 1.10709302e-01 1.79881465e+00
6.09879851e-01 -4.63396251e-01 3.90714467e-01 8.49445537e-02
-4.50272411e-01 -3.60132337e-01 -7.97843754e-01 -8.14319074e-01
4.79085267e-01 -5.38756073e-01 5.29569745e-01 1.20195377e+00
3.11841965e-01 6.12378597e-01 4.16437536e-02 7.29773790e-02
1.30154610e-01 -8.24836269e-02 1.71503976e-01 -1.30442441e+00
-5.01659513e-01 -6.58969641e-01 -4.55517247e-02 -7.14704156e-01
4.48703140e-01 -1.00516403e+00 1.07674524e-01 -8.24772775e-01
-1.07585207e-01 -4.77472037e-01 -9.82054412e-01 5.25096953e-01
5.28266281e-02 5.29971831e-02 5.91905057e-01 1.32318556e-01
-2.09766567e-01 5.21968663e-01 7.19557524e-01 5.13209365e-02
-4.34324801e-01 2.06476346e-01 -8.66825223e-01 9.22929347e-01
1.27966225e+00 -2.00945452e-01 -2.16244847e-01 -3.18058461e-01
1.64486930e-01 -7.44293258e-02 1.19434200e-01 -9.75578010e-01
-6.63648769e-02 8.28509927e-02 2.56663144e-01 -3.89990449e-01
4.03259784e-01 -3.01363468e-01 -2.71604538e-01 1.23665184e-01
-9.23623919e-01 -1.65297225e-01 6.62482619e-01 3.70156078e-04
-4.58182365e-01 -4.52495143e-02 1.10921288e+00 4.68790047e-02
-8.48109007e-01 -1.68609932e-01 -5.09812891e-01 3.39583218e-01
2.22114325e-01 -5.57889231e-02 -3.16722572e-01 -5.41433766e-02
-1.19045722e+00 -3.27417493e-01 -1.64835881e-02 8.69469821e-01
-1.26048655e-03 -1.46218622e+00 -5.14476717e-01 6.34115338e-01
7.21113086e-02 -6.74255848e-01 1.36927903e-01 6.60302281e-01
4.69789684e-01 6.36520445e-01 -3.95728827e-01 -2.89431304e-01
-9.74683344e-01 4.31534022e-01 5.91457725e-01 -3.72382879e-01
1.29359394e-01 1.06473982e+00 7.71882951e-01 -9.19516861e-01
4.76714075e-01 -6.27198756e-01 -2.99932063e-01 4.72932696e-01
2.80156642e-01 2.62180477e-01 1.01184838e-01 -5.12111127e-01
-4.96514648e-01 4.58912551e-01 -1.00996166e-01 -3.06788653e-01
1.18690121e+00 -3.72981220e-01 1.35131553e-01 1.00241196e+00
1.19127941e+00 5.94228506e-01 -1.27438283e+00 -4.59093630e-01
2.88928568e-01 3.10805768e-01 1.64949015e-01 -1.27072227e+00
-8.18015695e-01 1.10834718e+00 7.21445918e-01 1.40873879e-01
9.48418200e-01 3.57338548e-01 5.80787480e-01 9.86600965e-02
-4.70216461e-02 -1.38536668e+00 7.00390413e-02 9.89951551e-01
1.06190372e+00 -1.04862499e+00 -8.58789921e-01 -2.08784640e-01
-6.90397143e-01 8.56690705e-01 7.15547323e-01 6.24341480e-02
8.51396561e-01 4.39767569e-01 2.99745977e-01 2.10838631e-01
-7.40280688e-01 -8.42686966e-02 2.61097938e-01 7.52798975e-01
9.01256680e-01 2.48501360e-01 1.74709201e-01 1.19429946e+00
-1.04434192e+00 -4.00612235e-01 -8.58447179e-02 1.36682525e-01
-5.66313088e-01 -1.13234127e+00 -3.06999594e-01 -1.64651871e-02
-2.63941377e-01 -5.23856819e-01 -6.22795224e-01 4.22548860e-01
2.89033115e-01 8.88734519e-01 3.99291903e-01 -6.79601789e-01
3.00332457e-01 8.41133654e-01 1.92663729e-01 -8.29035282e-01
-9.95866716e-01 -9.45585892e-02 2.09788352e-01 -5.94242990e-01
-2.50311702e-01 -7.38773346e-01 -1.12154710e+00 2.17843264e-01
-9.49432179e-02 7.81645551e-02 6.15109324e-01 9.17063773e-01
3.17336410e-01 7.53091753e-01 8.42679620e-01 -7.00488210e-01
-7.32723236e-01 -1.18624222e+00 -6.39799535e-01 2.69272178e-01
6.05091691e-01 -4.62366939e-01 -3.96413803e-01 6.45229667e-02] | [14.405254364013672, 6.842644691467285] |
b8ead6c2-32b5-45e9-98f5-5c8856166e3a | latent-tree-decomposition-parsers-for-amr-to | 2108.12304 | null | https://arxiv.org/abs/2108.12304v2 | https://arxiv.org/pdf/2108.12304v2.pdf | Latent Tree Decomposition Parsers for AMR-to-Text Generation | Graph encoders in AMR-to-text generation models often rely on neighborhood convolutions or global vertex attention. While these approaches apply to general graphs, AMRs may be amenable to encoders that target their tree-like structure. By clustering edges into a hierarchy, a tree decomposition summarizes graph structure. Our model encodes a derivation forest of tree decompositions and extracts an expected tree. From tree node embeddings, it builds graph edge features used in vertex attention of the graph encoder. Encoding TD forests instead of shortest-pairwise paths in a self-attentive baseline raises BLEU by 0.7 and chrF++ by 0.3. The forest encoder also surpasses a convolutional baseline for molecular property prediction by 1.92% ROC-AUC. | ['Daniel Gildea', 'Lisa Jin'] | 2021-08-27 | null | null | null | null | ['tree-decomposition'] | ['graphs'] | [ 6.30074680e-01 1.10168505e+00 -4.08142298e-01 -2.64894694e-01
-6.52549028e-01 -7.47551382e-01 6.56237900e-01 5.36680341e-01
7.59030506e-02 8.55892122e-01 5.50455570e-01 -8.05978775e-01
2.97013491e-01 -1.43559754e+00 -9.46375668e-01 -3.37743163e-01
-2.83488631e-01 5.48767328e-01 -7.64573291e-02 -2.83684265e-02
-1.29622012e-01 4.46676195e-01 -7.53827453e-01 6.19194269e-01
6.80067718e-01 6.88277781e-01 -2.20318772e-02 1.18138337e+00
-4.06107396e-01 9.30833697e-01 -4.99509633e-01 -9.59390223e-01
-3.25656757e-02 -5.55684805e-01 -1.02244389e+00 -7.83236697e-03
6.79095209e-01 -2.38724157e-01 -8.00630569e-01 6.83088601e-01
3.32576513e-01 -2.02407643e-01 9.10460055e-01 -1.08965349e+00
-1.10133100e+00 1.12650979e+00 -5.16869843e-01 8.39310959e-02
3.52159768e-01 1.38214335e-01 1.72239304e+00 -4.44068313e-01
1.14714551e+00 1.15359163e+00 8.59466136e-01 5.92932165e-01
-1.66445541e+00 -2.50916988e-01 1.52790532e-01 -2.16666922e-01
-9.95201230e-01 -2.38744885e-01 3.10118705e-01 -4.14048404e-01
1.57055867e+00 1.53817773e-01 8.17511499e-01 1.30311608e+00
5.92485189e-01 4.30917233e-01 6.43097460e-01 -7.50495493e-02
-5.43188006e-02 -4.21892345e-01 1.15446582e-01 1.18297458e+00
5.76574028e-01 -6.11194894e-02 -5.56288302e-01 -1.35423437e-01
8.62207770e-01 -2.82600820e-01 -1.52291358e-01 -1.84854880e-01
-9.95038092e-01 9.39820707e-01 8.88841808e-01 -1.12451516e-01
-2.68054157e-01 7.44460166e-01 4.15917337e-01 2.77120709e-01
4.81138706e-01 6.39729798e-01 -3.28530937e-01 2.56357998e-01
-7.98810959e-01 2.96458989e-01 7.86989868e-01 1.29993463e+00
8.22204053e-01 9.16185752e-02 -5.99475741e-01 2.89308876e-01
8.30607712e-02 4.10286747e-02 -1.47881573e-02 -5.43118119e-01
3.05665940e-01 7.28860795e-01 -3.03388864e-01 -7.23006785e-01
-3.73214036e-01 -7.81803727e-01 -9.00462985e-01 -2.46072799e-01
2.43209094e-01 -8.12136680e-02 -1.40602589e+00 1.68699312e+00
-3.66058433e-04 -9.99864265e-02 -3.34425047e-02 3.12717944e-01
1.19692612e+00 6.49116635e-01 2.55365789e-01 2.92115770e-02
1.40316546e+00 -9.70617890e-01 -5.07142127e-01 -2.46728033e-01
8.27542484e-01 -3.22290182e-01 8.02494526e-01 9.52696949e-02
-1.03959072e+00 -3.65792811e-01 -8.50383222e-01 -4.55866039e-01
-3.63473743e-01 -1.96104750e-01 1.01644838e+00 5.72550237e-01
-1.53454471e+00 9.31116104e-01 -5.78158975e-01 -4.68163043e-01
6.55407071e-01 2.75061607e-01 -4.03219372e-01 -6.76333085e-02
-1.07485521e+00 7.71290243e-01 4.03871179e-01 -3.34956735e-01
-1.13060415e+00 -7.11694419e-01 -1.02763247e+00 2.34611586e-01
2.13065408e-02 -1.29546034e+00 1.25352407e+00 -5.12091339e-01
-1.16218543e+00 7.73851454e-01 -3.08992803e-01 -9.09499049e-01
1.82869583e-01 1.72006637e-01 -1.73191905e-01 1.21957183e-01
1.82048023e-01 8.68642151e-01 6.48914456e-01 -7.80251026e-01
-1.40334830e-01 -9.01260972e-02 2.15015754e-01 1.34506479e-01
-1.86088420e-02 -2.82215357e-01 -2.80797988e-01 -6.01771295e-01
-8.06261674e-02 -7.08652198e-01 -5.32191992e-01 -1.51164234e-01
-9.12319720e-01 -1.70551553e-01 3.89263660e-01 -6.86672926e-01
1.22418308e+00 -1.38754213e+00 1.07344761e-01 7.22528175e-02
9.77783203e-01 -2.33702376e-01 -4.33472216e-01 8.88277829e-01
-3.63894105e-01 6.43965125e-01 -2.48176083e-01 -9.51425731e-03
-1.48140818e-01 -6.65995702e-02 -1.57413259e-01 1.40359446e-01
6.92645192e-01 1.47436726e+00 -1.10042131e+00 -3.62191886e-01
-2.13229209e-01 4.66714054e-01 -8.27254355e-01 2.44126827e-01
-6.45652890e-01 -1.02716073e-01 -2.31467754e-01 5.31559885e-01
5.00023365e-01 -6.53132737e-01 5.48001111e-01 -1.32617846e-01
2.41082773e-01 8.54550660e-01 -4.18157756e-01 1.60781658e+00
-3.88825864e-01 6.96637630e-01 -2.83248067e-01 -8.26391637e-01
9.39312339e-01 1.58979192e-01 2.04354167e-01 -2.08175257e-01
-2.81403333e-01 -6.43784404e-02 2.99926251e-01 7.36467093e-02
4.94637042e-01 -2.10973285e-02 -2.04054154e-02 2.89270043e-01
5.21755338e-01 -1.30846247e-01 2.61909038e-01 8.41497123e-01
1.83818495e+00 4.54561412e-01 4.14392114e-01 -9.70328897e-02
-6.51010200e-02 6.74931556e-02 2.27600291e-01 7.39184201e-01
3.21562439e-01 5.86807787e-01 1.05989337e+00 -6.42932832e-01
-1.23088098e+00 -1.32808232e+00 1.00855835e-01 9.16682184e-01
-6.34582102e-01 -1.25408757e+00 -7.00119793e-01 -9.10302699e-01
-4.34031896e-03 8.40324819e-01 -7.91056335e-01 -3.73035073e-01
-3.64368021e-01 -8.09262335e-01 8.59946907e-01 5.25543094e-01
3.90214584e-04 -8.55596364e-01 -3.76811028e-02 4.73543048e-01
-1.57039866e-01 -1.22001314e+00 -6.59872413e-01 6.41793430e-01
-8.87571037e-01 -1.02987671e+00 -5.18018425e-01 -6.82322323e-01
6.97625995e-01 -7.16566518e-02 1.74074411e+00 -1.40478685e-02
-4.06107396e-01 -4.61080447e-02 -3.37789178e-01 -2.53151178e-01
-6.73239827e-01 5.78510702e-01 -3.09675097e-01 -3.36975396e-01
2.53518879e-01 -6.66856229e-01 -4.81626689e-01 -3.18627238e-01
-3.84414732e-01 3.57058585e-01 6.68426692e-01 8.92307222e-01
6.31654978e-01 -3.63425553e-01 4.90631402e-01 -1.21414399e+00
7.18133152e-01 -4.42006886e-01 -4.54065800e-01 1.97962612e-01
-8.26145232e-01 4.27318066e-01 7.86160648e-01 -6.37325691e-03
-7.09892154e-01 1.51255399e-01 -1.00454569e-01 -1.16430923e-01
-5.35330661e-02 5.44906437e-01 -8.52851048e-02 2.54210532e-01
8.44589114e-01 3.14274520e-01 -1.74089357e-01 -2.97782570e-02
8.96862864e-01 3.33227426e-01 3.23184341e-01 -4.28419143e-01
8.66772115e-01 2.21567191e-02 2.45172307e-01 -8.19168806e-01
-8.12277496e-01 1.13796212e-01 -5.49184740e-01 2.51680970e-01
1.24216986e+00 -9.63627815e-01 -7.31196702e-01 -8.10416788e-02
-1.45452082e+00 -4.14309829e-01 -4.53401625e-01 -1.04431227e-01
-5.33699155e-01 3.37966770e-01 -1.01704049e+00 -7.13935375e-01
-6.74749613e-01 -7.86684036e-01 1.19028759e+00 -1.70546725e-01
-3.86969179e-01 -8.22666109e-01 -2.90733557e-02 4.41491082e-02
1.50592014e-01 6.00974262e-01 1.22481978e+00 -5.00344694e-01
-9.50520098e-01 8.15965980e-02 -3.78101766e-01 -1.02465913e-01
1.45097181e-01 1.50501058e-01 -9.31921422e-01 -9.06012282e-02
-1.02242339e+00 -2.39981070e-01 1.36795497e+00 4.42171216e-01
1.22566485e+00 -6.35386169e-01 -6.55532539e-01 8.18711221e-01
1.24863183e+00 -1.16399564e-01 6.27163589e-01 -1.26450941e-01
9.74147677e-01 3.95256251e-01 -1.40538946e-01 3.46995555e-02
3.64006132e-01 3.02988946e-01 5.56606412e-01 -2.12628797e-01
-4.90068793e-01 -1.01446986e+00 5.74096322e-01 5.33403456e-01
-1.01220556e-01 -6.78430319e-01 -8.58016074e-01 5.13852179e-01
-1.46862769e+00 -9.96438384e-01 -6.78880632e-01 1.87147963e+00
9.02143419e-01 3.32122922e-01 1.35461807e-01 -3.01952869e-01
7.06290722e-01 3.91210109e-01 -4.96458858e-01 -1.00147951e+00
-1.70921430e-01 6.13061786e-01 7.38438606e-01 5.89042187e-01
-8.27395618e-01 1.29805541e+00 6.92034674e+00 4.52116370e-01
-5.29150844e-01 -2.78997153e-01 8.64267707e-01 1.66301981e-01
-8.81819010e-01 2.80717373e-01 -7.35439420e-01 1.30252987e-01
1.38766468e+00 -2.50442445e-01 5.52494466e-01 7.14916527e-01
-1.72567189e-01 4.50980455e-01 -1.32588840e+00 6.32631361e-01
-3.08982491e-01 -1.94645238e+00 5.33004940e-01 5.25279164e-01
5.82166612e-01 2.02116847e-01 -1.62852257e-01 2.98027486e-01
1.16765857e+00 -1.80857158e+00 3.12983096e-01 3.36766928e-01
1.29095805e+00 -7.88814604e-01 4.07488167e-01 -5.29475287e-02
-1.38551962e+00 3.20240468e-01 -5.26800454e-01 -1.37447163e-01
1.16171859e-01 7.48210192e-01 -1.51098073e+00 7.30444849e-01
2.66525060e-01 7.36627460e-01 -5.92532754e-01 5.54768026e-01
-4.90361214e-01 7.29847729e-01 7.15789711e-03 -1.77556843e-01
3.51699322e-01 -2.28531674e-01 3.85193020e-01 1.55580759e+00
9.43836644e-02 -1.02447338e-01 -3.28537412e-02 1.09337234e+00
-6.91724479e-01 -4.48403247e-02 -1.17449617e+00 -7.71847188e-01
3.84476542e-01 1.32444584e+00 -5.54458022e-01 -4.60576952e-01
-2.99921125e-01 1.30605876e+00 8.33103716e-01 2.51424432e-01
-8.39664817e-01 -5.98445356e-01 7.60976076e-01 4.10360515e-01
4.52892214e-01 -1.22069471e-01 -1.42148629e-01 -9.85472262e-01
-4.06185061e-01 -8.77019227e-01 3.52122188e-01 -8.63179564e-01
-1.16025138e+00 8.32319796e-01 -3.64043832e-01 -5.33104122e-01
-2.33025193e-01 -6.20317340e-01 -7.25352705e-01 1.02133691e+00
-1.10114229e+00 -1.44566965e+00 -1.39885053e-01 4.42644730e-02
2.64885336e-01 6.84284568e-02 1.30089712e+00 -3.77532601e-01
-4.17647421e-01 6.73904836e-01 -2.77389079e-01 3.53948563e-01
2.27632493e-01 -1.64356053e+00 1.53963089e+00 7.32309461e-01
4.90416795e-01 8.94821286e-01 5.90047538e-01 -8.90088558e-01
-1.30022347e+00 -1.58464026e+00 1.29519153e+00 -6.13138735e-01
7.58680522e-01 -7.38134265e-01 -7.01253593e-01 1.05105913e+00
5.08070827e-01 -1.08631641e-01 6.83094800e-01 5.03431022e-01
-9.43005025e-01 2.85707980e-01 -9.44472849e-01 7.89540291e-01
1.68063080e+00 -5.72538376e-01 -2.31435359e-01 5.52083611e-01
1.10360491e+00 -3.35186094e-01 -1.12861073e+00 1.45759597e-01
4.49830830e-01 -6.76381826e-01 7.90863931e-01 -1.22160852e+00
9.30258214e-01 -9.54568014e-02 6.13273643e-02 -1.50308311e+00
-8.57096016e-01 -9.19397116e-01 -4.48036581e-01 9.04697537e-01
8.61142457e-01 -3.76681298e-01 1.00351191e+00 -1.58609688e-01
-1.67046174e-01 -8.12485814e-01 -4.98929799e-01 -5.31001031e-01
2.65587032e-01 -2.94162303e-01 6.37902915e-01 7.96093404e-01
3.47050130e-02 1.15377438e+00 -1.04923062e-01 1.52951460e-02
6.63764834e-01 6.72406182e-02 6.27469480e-01 -1.30995178e+00
-5.62352061e-01 -4.60024208e-01 -3.60803545e-01 -1.10350883e+00
1.59538597e-01 -1.69157326e+00 -3.18438977e-01 -2.28234172e+00
3.25689107e-01 -3.99780348e-02 -5.81181757e-02 5.56110263e-01
-2.57007509e-01 2.71494072e-02 -3.85578419e-03 -4.57963228e-01
-2.56091923e-01 4.06901270e-01 1.25303292e+00 -3.06705117e-01
9.57022607e-02 -2.99837887e-01 -1.10643280e+00 3.59210134e-01
1.05667996e+00 -5.41372538e-01 -5.30224562e-01 -4.22518104e-01
4.65380669e-01 2.07567483e-01 2.46506900e-01 -6.13736212e-01
-2.64630318e-01 -7.19354546e-04 5.61719120e-01 -4.51900691e-01
1.77633032e-01 -9.05200839e-02 2.15399504e-01 5.89667916e-01
-5.52001655e-01 1.07641317e-01 5.74113317e-02 8.11858296e-01
2.75314093e-01 3.32647413e-02 5.72633147e-01 -5.88992357e-01
-6.13794029e-02 6.54416740e-01 -5.57315707e-01 1.75373480e-01
4.53904897e-01 -3.70764852e-01 -5.23251235e-01 -6.18196487e-01
-8.35585356e-01 -1.27639743e-02 4.70392287e-01 1.72360748e-01
6.48228228e-01 -1.12825501e+00 -8.97280097e-01 -1.02567844e-01
2.14197412e-01 6.47439584e-02 -1.95041478e-01 3.91545504e-01
-5.94530284e-01 3.64159018e-01 5.33575518e-03 -2.77692527e-01
-1.15719402e+00 5.28095067e-01 4.21984285e-01 -5.21353960e-01
-6.78397596e-01 1.18729961e+00 4.32592958e-01 -4.68775749e-01
-6.37849867e-02 -5.08346081e-01 1.28090069e-01 -1.16205081e-01
2.09738046e-01 1.09093720e-02 -3.62510756e-02 -2.91262001e-01
-4.44674164e-01 1.32079288e-01 -2.29451567e-01 1.00472361e-01
1.20831513e+00 4.01554704e-01 -1.47369429e-01 1.30933464e-01
1.14307570e+00 1.46116659e-01 -1.05296791e+00 1.32037133e-01
9.84192416e-02 1.84142843e-01 -1.22883119e-01 -8.28977585e-01
-7.05888152e-01 1.00659633e+00 -1.13100842e-01 2.48716101e-01
6.39902174e-01 1.82580531e-01 7.63851404e-01 4.25499856e-01
9.53444913e-02 -4.39528674e-01 7.19730854e-02 5.91034710e-01
8.28946412e-01 -7.76876688e-01 6.43388852e-02 -7.27832735e-01
-5.61988413e-01 1.22881019e+00 5.58875799e-01 -1.24568939e-01
2.75799990e-01 4.87731040e-01 -5.76975346e-01 -2.92125493e-01
-1.25200450e+00 -4.58546638e-01 2.22342163e-01 1.07822633e+00
1.04146338e+00 3.68028998e-01 -1.47263631e-01 4.09180999e-01
-5.71580172e-01 -2.52575517e-01 8.00282776e-01 3.16657990e-01
-5.25695086e-01 -1.13633871e+00 3.05186003e-01 9.17223930e-01
-4.71484452e-01 -7.58038580e-01 -9.33442354e-01 6.53936207e-01
-1.15094692e-01 8.17686856e-01 2.00001806e-01 -6.11718893e-01
-1.92583036e-02 2.71520942e-01 7.03747451e-01 -1.07028759e+00
-6.72260463e-01 -1.05923861e-01 8.19253027e-01 -4.42647874e-01
1.70156926e-01 -3.17333460e-01 -1.32628143e+00 -5.43748438e-01
-2.18457833e-01 1.34958223e-01 2.01314703e-01 1.61990717e-01
5.43893337e-01 9.98869479e-01 2.51799732e-01 -2.76314318e-01
-1.97490260e-01 -9.80180204e-01 -3.05832952e-01 2.61734128e-02
2.52373189e-01 -6.95879459e-02 1.00931495e-01 -6.41824082e-02] | [10.201780319213867, 8.226901054382324] |
829b9569-5681-4b5d-89a7-75cd0c891e6f | perceptual-loss-based-speech-denoising-with | 2010.1186 | null | http://arxiv.org/abs/2010.11860v1 | http://arxiv.org/pdf/2010.11860v1.pdf | Perceptual Loss based Speech Denoising with an ensemble of Audio Pattern Recognition and Self-Supervised Models | Deep learning based speech denoising still suffers from the challenge of
improving perceptual quality of enhanced signals. We introduce a generalized
framework called Perceptual Ensemble Regularization Loss (PERL) built on the
idea of perceptual losses. Perceptual loss discourages distortion to certain
speech properties and we analyze it using six large-scale pre-trained models:
speaker classification, acoustic model, speaker embedding, emotion
classification, and two self-supervised speech encoders (PASE+, wav2vec 2.0).
We first build a strong baseline (w/o PERL) using Conformer Transformer
Networks on the popular enhancement benchmark called VCTK-DEMAND. Using
auxiliary models one at a time, we find acoustic event and self-supervised
model PASE+ to be most effective. Our best model (PERL-AE) only uses acoustic
event model (utilizing AudioSet) to outperform state-of-the-art methods on
major perceptual metrics. To explore if denoising can leverage full framework,
we use all networks but find that our seven-loss formulation suffers from the
challenges of Multi-Task Learning. Finally, we report a critical observation
that state-of-the-art Multi-Task weight learning methods cannot outperform hand
tuning, perhaps due to challenges of domain mismatch and weak complementarity
of losses. | [] | 2020-10-22 | perceptual-loss-based-speech-denoising-with-1 | https://ieeexplore.ieee.org/abstract/document/9413555 | https://arxiv.org/pdf/2010.11860.pdf | null | ['speech-denoising'] | ['speech'] | [ 4.13062423e-01 -3.13085131e-02 2.29952350e-01 -3.76461923e-01
-1.43417931e+00 -3.01290780e-01 6.70164227e-01 2.34481692e-02
-6.37351990e-01 3.41231138e-01 7.80225694e-01 -1.78344250e-01
-5.62733486e-02 -1.92875654e-01 -7.55279243e-01 -6.36893630e-01
-5.13510592e-02 -5.48919439e-02 4.79480512e-02 -5.15595675e-01
-1.77520543e-01 8.19789767e-02 -1.45241976e+00 6.20679379e-01
4.32780832e-01 1.18843067e+00 3.32149640e-02 8.45188022e-01
2.00331092e-01 7.66031265e-01 -7.06062555e-01 -6.64163768e-01
1.82341054e-01 -1.32458702e-01 -4.63775456e-01 -3.05599749e-01
5.42650580e-01 -3.72999072e-01 -4.77791756e-01 9.39292848e-01
1.18750238e+00 3.36461514e-01 5.30397177e-01 -1.27715635e+00
-8.01052988e-01 7.47349977e-01 -3.48410517e-01 2.75556147e-01
-2.11038422e-02 2.38403350e-01 1.37075102e+00 -1.15572429e+00
4.17954087e-01 1.54886758e+00 1.12342095e+00 4.72164214e-01
-1.47923017e+00 -7.31313169e-01 2.56178081e-01 4.86030072e-01
-1.07535708e+00 -1.16950536e+00 8.29111814e-01 -1.18816249e-01
1.48109663e+00 1.97954923e-01 -1.49655081e-02 1.91517162e+00
6.18903302e-02 7.08898783e-01 1.02760828e+00 -3.91400963e-01
1.93242237e-01 1.36126712e-01 2.29557320e-01 3.08067024e-01
-4.65457678e-01 5.47499597e-01 -1.00103951e+00 -2.05849022e-01
6.85132667e-02 -4.39130008e-01 -2.88454175e-01 -2.41589844e-02
-8.88943553e-01 8.11627448e-01 1.00547718e-02 2.96008557e-01
-3.49976718e-01 3.24137688e-01 8.28588665e-01 7.98534989e-01
7.81788170e-01 1.55739278e-01 -8.15716505e-01 -3.95207435e-01
-1.12087774e+00 1.87071919e-01 6.35694802e-01 5.06932139e-01
4.55071390e-01 5.62264740e-01 -4.09301877e-01 1.43019533e+00
2.22031116e-01 2.01903746e-01 4.66595113e-01 -1.04221630e+00
3.86074722e-01 -5.14809191e-01 -2.71167189e-01 -5.75340629e-01
-3.00227433e-01 -7.43448079e-01 -7.32683837e-01 3.95330399e-01
-7.38896653e-02 -3.44213009e-01 -1.00233710e+00 2.08317351e+00
-1.39133796e-01 3.66104066e-01 1.44504011e-01 5.63887477e-01
9.53517616e-01 8.45250905e-01 3.74615848e-01 -1.43641666e-01
1.14297426e+00 -1.14079475e+00 -8.87038112e-01 -4.10086662e-01
2.04685763e-01 -8.02790761e-01 1.09173703e+00 7.75615513e-01
-1.29282916e+00 -7.76635170e-01 -1.23901772e+00 -1.90843865e-01
-2.85601437e-01 5.96544743e-02 3.20575863e-01 8.27739775e-01
-1.30913484e+00 7.82436192e-01 -8.26336324e-01 -1.11011177e-01
2.39220262e-01 1.05908036e-01 -2.72873998e-01 2.94452179e-02
-1.41498685e+00 1.20297897e+00 4.45133977e-04 -2.15872139e-01
-1.33984423e+00 -1.19240892e+00 -1.00661123e+00 2.86246836e-01
1.33159533e-01 -5.51033020e-01 1.37241518e+00 -1.02252579e+00
-1.66291451e+00 6.29074156e-01 -1.82600915e-01 -7.98098922e-01
2.81792313e-01 -4.51858461e-01 -9.89277244e-01 1.32269144e-01
-2.59179652e-01 6.44426703e-01 1.38392437e+00 -1.45934021e+00
-4.03893769e-01 -6.10647760e-02 -3.32693338e-01 8.98103938e-02
-7.31042087e-01 3.94238591e-01 -1.62150204e-01 -1.10694826e+00
-3.89271826e-01 -3.96154195e-01 4.60960083e-02 -1.05042264e-01
-2.97154456e-01 -1.87752545e-01 9.52918053e-01 -1.18167448e+00
1.22438252e+00 -2.65172696e+00 1.84239730e-01 -1.21586854e-02
1.02822006e-01 4.20549691e-01 -8.04328978e-01 4.72841710e-01
-5.98774314e-01 2.24748313e-01 -1.74473569e-01 -1.18196416e+00
4.19674665e-01 8.48987922e-02 -4.63328153e-01 2.77251482e-01
3.55536669e-01 5.25204659e-01 -5.02705336e-01 -1.99917898e-01
3.07908893e-01 1.09535921e+00 -8.16935718e-01 6.97612241e-02
7.58288577e-02 -1.04484275e-01 2.97310740e-01 4.16046560e-01
7.41378963e-01 3.35687310e-01 -2.41929337e-01 -7.12625742e-01
1.68565422e-01 7.06211686e-01 -1.24980772e+00 1.91348171e+00
-7.87607253e-01 6.90333605e-01 6.33867204e-01 -1.17814434e+00
5.44888794e-01 6.84211433e-01 4.38800961e-01 -6.19111955e-01
-2.60601997e-01 9.92489681e-02 1.17811561e-02 -2.51597553e-01
4.81883258e-01 -4.62406039e-01 2.43065983e-01 2.04588264e-01
9.16551828e-01 -1.26315966e-01 -2.32681349e-01 1.95659027e-01
1.47948050e+00 -8.83777216e-02 -1.24307334e-01 -2.85911597e-02
7.90491998e-02 -6.64872944e-01 5.93540549e-01 8.35633755e-01
-5.53417623e-01 6.73282385e-01 2.68487543e-01 1.56987831e-01
-1.04782903e+00 -1.42705882e+00 -2.12094054e-01 1.78145933e+00
-4.52618957e-01 -6.45816922e-01 -4.74122465e-01 -4.33376700e-01
-7.66657293e-03 1.13383651e+00 -4.63948667e-01 -3.54369700e-01
-4.35354441e-01 -7.78363407e-01 9.13556039e-01 5.12516618e-01
5.70123158e-02 -9.51025963e-01 -1.23230189e-01 4.46727753e-01
-7.72344172e-02 -1.16083181e+00 -4.73229289e-01 8.09429467e-01
-4.10385907e-01 -2.22208098e-01 -6.11132324e-01 -6.14217460e-01
-1.66120589e-01 5.12639135e-02 1.36963892e+00 -2.18457744e-01
8.54915902e-02 7.41214871e-01 -4.51749176e-01 -5.10847390e-01
-6.94529951e-01 -4.48991582e-02 2.10059434e-01 -1.40985221e-01
2.12660715e-01 -1.08386898e+00 -5.27148843e-01 4.45685610e-02
-8.51602316e-01 -3.50724101e-01 4.54258740e-01 1.05910265e+00
3.37510407e-01 1.85778394e-01 9.96489167e-01 -2.89883554e-01
8.29737782e-01 -4.21501577e-01 5.98433279e-02 4.36180942e-02
-5.37181318e-01 5.28128743e-02 4.79307413e-01 -4.87608492e-01
-1.20280337e+00 -3.89727592e-01 -8.57177496e-01 -5.58619738e-01
-1.36521682e-01 3.24840456e-01 -1.97817892e-01 8.93843472e-02
7.81854868e-01 1.04946576e-01 -1.80586763e-02 -6.58196270e-01
5.35287976e-01 6.59090281e-01 6.48566067e-01 -6.53416514e-01
5.86494207e-01 2.66666681e-01 -5.06257892e-01 -9.69507337e-01
-8.00882041e-01 -3.26838046e-01 -3.38853709e-02 8.78781453e-02
7.66519129e-01 -1.35211635e+00 -3.70825440e-01 3.77489299e-01
-1.34746039e+00 -4.95566994e-01 -5.54799139e-01 2.88426071e-01
-4.60463345e-01 4.25077528e-01 -9.64710474e-01 -8.06506872e-01
-4.38948691e-01 -1.13293469e+00 1.03099251e+00 -6.23891950e-01
-5.93204014e-02 -9.08382893e-01 6.60528764e-02 2.46872455e-01
9.49132025e-01 -8.61774385e-02 8.33540738e-01 -8.74885440e-01
1.27054378e-01 3.20378572e-01 -5.13512753e-02 1.19841814e+00
-8.81429091e-02 -1.06033936e-01 -1.62535453e+00 -4.93499756e-01
3.48076671e-01 -6.34943604e-01 1.57530665e+00 6.40697658e-01
1.28049254e+00 -5.22939079e-02 1.61714196e-01 7.81562567e-01
1.12045205e+00 5.55276796e-02 6.07310116e-01 9.27099064e-02
4.05330271e-01 4.83752191e-01 -1.06528923e-01 4.10629690e-01
2.80239820e-01 6.38166308e-01 3.74196500e-01 -2.13721037e-01
-5.65644026e-01 -8.35717916e-02 9.75920200e-01 1.20024848e+00
2.17169300e-01 -3.39025706e-01 -5.75470090e-01 6.12456203e-01
-1.38968718e+00 -1.07764542e+00 3.36863756e-01 1.83465457e+00
1.07062149e+00 4.58067000e-01 9.86967087e-02 4.14535522e-01
4.80903715e-01 5.72593868e-01 -4.04665291e-01 -5.89778185e-01
-4.55954820e-01 7.17102051e-01 2.28653282e-01 6.83125973e-01
-1.26861191e+00 6.57488406e-01 6.38506556e+00 1.26666617e+00
-9.64700699e-01 7.29186058e-01 6.20692372e-01 -4.12295252e-01
-3.53652209e-01 -3.79800171e-01 -6.80609584e-01 3.01506579e-01
1.49365735e+00 1.55723944e-01 6.89861000e-01 5.58029592e-01
2.26080582e-01 4.86456782e-01 -1.26067579e+00 1.02698302e+00
2.25841686e-01 -1.03822064e+00 -8.74443203e-02 -2.29287729e-01
4.24127609e-01 5.41639090e-01 4.28550512e-01 7.61614382e-01
4.64346230e-01 -9.84242797e-01 9.18448150e-01 1.89741060e-01
8.42225194e-01 -6.93063676e-01 4.63046819e-01 -6.84583932e-02
-1.04351187e+00 -1.80252239e-01 -2.15680435e-01 2.26685822e-01
4.93125409e-01 7.67974138e-01 -3.40865672e-01 4.41664279e-01
8.78510654e-01 7.35138297e-01 -3.87652934e-01 6.54693484e-01
-3.58191580e-02 1.07801080e+00 -3.70917976e-01 5.95248580e-01
2.07692638e-01 3.06067556e-01 8.48613203e-01 1.66557276e+00
3.70909572e-01 -3.18442285e-01 -1.42633587e-01 5.32752037e-01
-2.84635663e-01 -2.48487458e-01 -3.25714231e-01 1.13228314e-01
4.67304230e-01 1.09040380e+00 6.56927377e-02 -2.91074574e-01
-6.07566059e-01 1.17047989e+00 1.88509360e-01 5.48749983e-01
-9.03294563e-01 -2.76918709e-01 1.19355965e+00 -2.64449239e-01
6.76272988e-01 -6.86122552e-02 -2.50846356e-01 -9.56525624e-01
-1.02188654e-01 -1.30540812e+00 1.90781534e-01 -7.50259459e-01
-1.66685045e+00 5.99510550e-01 -3.54610860e-01 -7.84866631e-01
-9.67522040e-02 -6.83225811e-01 -6.50617659e-01 8.06771696e-01
-1.63552868e+00 -1.22327197e+00 8.01907331e-02 7.94821203e-01
7.97808766e-01 -3.01539361e-01 9.50077474e-01 7.95904636e-01
-4.28500473e-01 9.93890464e-01 1.11939207e-01 -1.23135291e-01
1.28665209e+00 -1.44710016e+00 5.22892118e-01 8.07201147e-01
2.70733476e-01 2.55031317e-01 9.09072459e-01 -2.44180292e-01
-1.02066672e+00 -1.07074666e+00 6.84014797e-01 -4.28002179e-01
8.11473787e-01 -5.78158021e-01 -9.15842414e-01 7.65172005e-01
7.92722642e-01 6.73330063e-03 7.69833744e-01 3.82920086e-01
-6.87292874e-01 -3.01673859e-01 -1.02047312e+00 4.73363012e-01
1.03317583e+00 -9.25111115e-01 -6.11601591e-01 2.30473071e-01
1.30879211e+00 -8.86774287e-02 -8.92819285e-01 3.92105460e-01
3.78139019e-01 -1.05100143e+00 1.40592313e+00 -5.49053013e-01
5.00530422e-01 1.44470006e-01 -6.20947540e-01 -1.94047832e+00
-2.97627032e-01 -8.55086982e-01 -1.04725108e-01 1.65051627e+00
6.03470862e-01 -5.24362266e-01 1.58676267e-01 2.46401414e-01
-7.30488777e-01 -4.48159903e-01 -1.27036488e+00 -8.26163471e-01
3.32824171e-01 -1.23516631e+00 3.20921361e-01 8.52407634e-01
-3.15122604e-01 4.44724530e-01 -6.94580436e-01 7.02382550e-02
6.96013808e-01 -9.13959563e-01 3.48663568e-01 -9.01582956e-01
-6.92775488e-01 -6.29919827e-01 -3.70441750e-02 -8.35431039e-01
3.78763467e-01 -7.81451046e-01 5.22383265e-02 -1.22688568e+00
-6.24628328e-02 3.01608555e-02 -8.73220026e-01 5.14294028e-01
-5.98012023e-02 1.17579594e-01 2.91406751e-01 -3.66610229e-01
-5.73046088e-01 9.13724422e-01 6.43338203e-01 -3.48925889e-01
-1.10410132e-01 -2.67706156e-01 -8.92211556e-01 7.36579120e-01
5.71160614e-01 -4.12816793e-01 -2.74299473e-01 -6.96592748e-01
5.81724346e-02 -7.98501670e-02 5.54320157e-01 -1.04208767e+00
1.93412244e-01 5.21225989e-01 2.20452160e-01 -2.90513158e-01
9.22244966e-01 -7.25201786e-01 -1.29046902e-01 1.00340493e-01
-6.39081180e-01 3.65483314e-02 4.63714719e-01 4.89341795e-01
-5.10674298e-01 -1.21000158e-02 8.97240043e-01 2.53952235e-01
-6.24618471e-01 -2.55332142e-02 -3.51969302e-01 2.90339917e-01
1.82296664e-01 2.17938766e-01 -3.67906630e-01 -6.48449004e-01
-1.01459706e+00 -6.80573657e-02 -2.37495869e-01 4.99021888e-01
6.81384265e-01 -1.37407207e+00 -1.05837262e+00 1.44096032e-01
-6.47810195e-03 -8.11522365e-01 4.32727337e-01 7.74771094e-01
3.38263422e-01 -7.54860416e-02 1.52498171e-01 -3.21494460e-01
-1.14990282e+00 3.14533561e-01 3.55611056e-01 -3.27994853e-01
-3.47285599e-01 1.13036203e+00 1.04608491e-01 -6.24504328e-01
6.79796994e-01 -4.32800829e-01 2.07403168e-01 1.56074241e-01
4.08735484e-01 5.19286156e-01 5.54898739e-01 -4.41889167e-01
-4.19677854e-01 4.97670025e-02 -5.42689823e-02 -5.57202816e-01
1.73868072e+00 -7.47578442e-02 2.90436745e-01 3.95565331e-01
1.45053744e+00 5.20931445e-02 -1.30520189e+00 -3.86012286e-01
-2.37278327e-01 3.49852741e-02 6.10767126e-01 -1.24691474e+00
-1.02034342e+00 1.16043317e+00 9.99782681e-01 1.64737582e-01
1.44617939e+00 -2.59184062e-01 9.85578835e-01 2.01983675e-01
-1.18612066e-01 -1.47335017e+00 2.13839039e-01 5.75185955e-01
1.19197643e+00 -1.32210195e+00 -2.83020973e-01 2.64610704e-02
-8.40013206e-01 7.81079233e-01 2.74219155e-01 5.30291982e-02
1.14416587e+00 6.53653502e-01 4.10168394e-02 1.06598154e-01
-1.06546557e+00 -1.88940108e-01 2.21229672e-01 8.28932583e-01
4.55076784e-01 -1.81429550e-01 1.44748077e-01 9.31636572e-01
-2.72176594e-01 -3.73717517e-01 1.08853735e-01 4.08272624e-01
-2.98278898e-01 -1.10479903e+00 -2.21256822e-01 3.86793911e-01
-8.38475525e-01 -6.14132285e-01 2.75520608e-02 5.01231790e-01
1.61074907e-01 1.37765265e+00 -6.35837466e-02 -5.11534393e-01
5.88120341e-01 2.67422378e-01 1.51216567e-01 -4.24847722e-01
-1.03676152e+00 3.53766322e-01 4.82015491e-01 -5.99167228e-01
-4.03291404e-01 -5.92949152e-01 -6.34924173e-01 -1.24857880e-01
-1.07420817e-01 -1.67384937e-01 8.07267547e-01 7.73342490e-01
4.12149340e-01 9.46437120e-01 5.31883299e-01 -1.05653048e+00
-1.06824303e+00 -1.25620198e+00 -5.97060680e-01 6.20700479e-01
8.05633605e-01 -5.69144785e-01 -8.43865097e-01 1.39006227e-01] | [15.078836441040039, 5.848421573638916] |
67574efd-0b84-4a1f-873c-0e0a97a4c0a4 | pmatch-paired-masked-image-modeling-for-dense | 2303.17342 | null | https://arxiv.org/abs/2303.17342v1 | https://arxiv.org/pdf/2303.17342v1.pdf | PMatch: Paired Masked Image Modeling for Dense Geometric Matching | Dense geometric matching determines the dense pixel-wise correspondence between a source and support image corresponding to the same 3D structure. Prior works employ an encoder of transformer blocks to correlate the two-frame features. However, existing monocular pretraining tasks, e.g., image classification, and masked image modeling (MIM), can not pretrain the cross-frame module, yielding less optimal performance. To resolve this, we reformulate the MIM from reconstructing a single masked image to reconstructing a pair of masked images, enabling the pretraining of transformer module. Additionally, we incorporate a decoder into pretraining for improved upsampling results. Further, to be robust to the textureless area, we propose a novel cross-frame global matching module (CFGM). Since the most textureless area is planar surfaces, we propose a homography loss to further regularize its learning. Combined together, we achieve the State-of-The-Art (SoTA) performance on geometric matching. Codes and models are available at https://github.com/ShngJZ/PMatch. | ['Xiaoming Liu', 'Shengjie Zhu'] | 2023-03-30 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhu_PMatch_Paired_Masked_Image_Modeling_for_Dense_Geometric_Matching_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhu_PMatch_Paired_Masked_Image_Modeling_for_Dense_Geometric_Matching_CVPR_2023_paper.pdf | cvpr-2023-1 | ['geometric-matching'] | ['computer-vision'] | [ 5.31338453e-01 1.10285193e-01 -2.75875837e-01 -5.23569942e-01
-8.12814116e-01 -2.32054561e-01 5.04905045e-01 -3.98661256e-01
4.04212587e-02 3.37998927e-01 1.29905239e-01 -1.11923911e-01
2.62623131e-01 -8.37565780e-01 -1.17568493e+00 -5.48126042e-01
5.62717140e-01 6.13409467e-02 2.10307673e-01 1.14694238e-01
1.88356996e-01 4.29433078e-01 -1.54131591e+00 5.09960234e-01
1.01273811e+00 1.22421288e+00 4.03849185e-01 2.56153584e-01
-2.44979635e-02 7.02055871e-01 -1.14524234e-02 -4.98341620e-01
6.25656247e-01 -3.48350108e-01 -5.17923772e-01 3.98140103e-01
1.16468465e+00 -7.03705907e-01 -7.74219275e-01 1.23211396e+00
2.08129913e-01 -1.97945684e-01 3.16977590e-01 -1.13334513e+00
-6.50666893e-01 2.09532782e-01 -7.45563626e-01 -2.77370661e-01
4.05869782e-01 -3.07249129e-02 8.87271643e-01 -1.22989082e+00
6.60124540e-01 1.34005249e+00 4.94832635e-01 2.92182833e-01
-1.12980831e+00 -8.27881157e-01 2.00649887e-01 2.27074951e-01
-1.45314610e+00 -7.89730668e-01 1.09965467e+00 -3.78600508e-01
5.76747537e-01 7.07960203e-02 4.60382938e-01 6.62500322e-01
1.30739883e-01 8.14824760e-01 1.09775972e+00 -4.54377204e-01
-2.18768060e-01 -3.22026879e-01 -2.47812361e-01 8.57479334e-01
1.23681337e-01 5.04350066e-01 -6.48710787e-01 9.88666415e-02
1.17118549e+00 3.26335609e-01 -5.26706338e-01 -7.00247288e-01
-1.03964376e+00 4.87160951e-01 6.49127543e-01 9.80496407e-03
-8.01937133e-02 1.71199128e-01 -1.27023041e-01 3.31687957e-01
4.93405640e-01 -1.28018586e-02 -2.89155543e-01 3.78840238e-01
-9.62415695e-01 4.68455702e-02 3.02612782e-01 1.26450992e+00
1.43370199e+00 -1.01793557e-01 8.95202607e-02 8.69383037e-01
4.25130874e-01 5.86168528e-01 1.94479346e-01 -1.18484628e+00
7.13723898e-01 7.25771904e-01 -1.09823018e-01 -9.98271465e-01
-4.36929762e-02 -5.27893245e-01 -7.47959495e-01 2.43402913e-01
1.94771484e-01 2.94870466e-01 -1.01339543e+00 1.65595078e+00
4.32348490e-01 4.94062841e-01 -2.66038567e-01 1.08208704e+00
7.04954684e-01 4.44252312e-01 -4.68522549e-01 1.83292121e-01
1.21926868e+00 -1.15782332e+00 -4.26101476e-01 -5.69668412e-01
4.16015565e-01 -1.25330186e+00 8.48160863e-01 2.18219645e-02
-1.27758110e+00 -7.03541577e-01 -1.18502164e+00 -6.45839274e-01
7.26697668e-02 2.42983863e-01 4.88874108e-01 2.11471185e-01
-9.17301118e-01 4.99381155e-01 -9.81028199e-01 2.23981533e-02
4.74823147e-01 3.03072602e-01 -3.97828817e-01 -8.03202808e-01
-8.97013903e-01 5.32645524e-01 1.17471172e-02 -9.71906446e-03
-5.88952243e-01 -7.76551247e-01 -1.19241226e+00 -1.23855002e-01
3.12781364e-01 -8.68062019e-01 1.02196062e+00 -8.84488285e-01
-1.34042144e+00 1.23959005e+00 -4.98297900e-01 -1.11525513e-01
4.93968546e-01 -2.13039756e-01 -1.29175797e-01 2.55311817e-01
1.84691086e-01 8.91633987e-01 1.03376544e+00 -1.43566787e+00
-5.69327474e-01 -3.63322765e-01 4.93312255e-02 2.21741691e-01
2.79134661e-02 -2.16874644e-01 -8.86742532e-01 -8.47261190e-01
6.22350872e-01 -7.90975332e-01 3.43375374e-03 4.19518054e-01
-4.63488489e-01 3.87726039e-01 6.81447685e-01 -7.62860835e-01
9.25479949e-01 -2.27685833e+00 1.66416764e-01 1.04024179e-01
2.02984467e-01 -2.17062943e-02 -2.60824621e-01 3.89382020e-02
-1.21748865e-01 -3.48083943e-01 -3.02114576e-01 -8.55346024e-01
-4.21767980e-02 2.28683233e-01 -2.89935350e-01 6.69247806e-01
1.85536951e-01 9.61007297e-01 -5.63349605e-01 -4.01153564e-01
4.52085644e-01 4.58257556e-01 -8.11857641e-01 3.89758289e-01
-3.48084830e-02 6.11692965e-01 -3.10196429e-01 8.77566040e-01
1.19580269e+00 -4.86513525e-01 2.84327511e-02 -5.91513276e-01
-1.14710957e-01 4.24507856e-01 -1.09541690e+00 2.15100479e+00
-4.87658143e-01 5.05027413e-01 2.91878223e-01 -8.00591409e-01
9.95356441e-01 -2.11249050e-02 6.62114859e-01 -8.38747621e-01
2.15080053e-01 4.40332592e-01 -2.25827649e-01 1.12452758e-02
3.17529619e-01 8.30464736e-02 2.57120460e-01 1.86147898e-01
-2.51493081e-02 -9.19714049e-02 -1.96433112e-01 -2.58721784e-02
8.00148547e-01 4.02432203e-01 1.01162933e-01 -1.52370185e-01
5.91587007e-01 -2.53914684e-01 7.42570996e-01 3.30652922e-01
7.78478682e-02 1.24642086e+00 9.46834013e-02 -3.18482012e-01
-1.15867484e+00 -1.12559187e+00 -2.73708820e-01 5.07817030e-01
7.05102503e-01 -3.00387979e-01 -7.10048378e-01 -2.42828265e-01
7.80242458e-02 1.20729379e-01 -2.46440008e-01 -1.89417437e-01
-8.70267570e-01 -2.05849439e-01 4.32262495e-02 3.92521203e-01
8.62428427e-01 -4.70485210e-01 -3.60482812e-01 7.03871772e-02
-3.11986715e-01 -1.44623470e+00 -1.04800951e+00 -1.26241609e-01
-8.98723245e-01 -1.15886414e+00 -6.48470700e-01 -1.07064331e+00
9.55626249e-01 7.05001950e-01 1.16007018e+00 3.75341505e-01
-1.69351816e-01 5.78885749e-02 -1.53516918e-01 1.91554904e-01
-1.11551687e-01 1.10883843e-02 -2.12043181e-01 2.51839638e-01
2.07586810e-01 -7.68616378e-01 -9.10158157e-01 7.22925305e-01
-8.79591167e-01 7.31906235e-01 7.16649711e-01 9.03838217e-01
8.63802254e-01 -3.33527833e-01 -1.20644242e-01 -7.14435697e-01
-4.22442704e-01 -1.46924868e-01 -7.43510723e-01 2.35600322e-01
-5.20492136e-01 6.77239299e-02 3.38044494e-01 -2.54657745e-01
-1.01693285e+00 3.46375793e-01 -1.82839960e-01 -9.48946238e-01
-2.98623927e-02 1.28386647e-01 -6.08851194e-01 -4.48569387e-01
2.09246740e-01 2.60773659e-01 1.00237064e-01 -6.67618811e-01
1.94129601e-01 4.53015983e-01 7.41096973e-01 -6.12155318e-01
1.15456176e+00 6.29754126e-01 2.52243038e-02 -4.81607497e-01
-1.06890869e+00 -4.73550230e-01 -6.93211973e-01 -3.23775481e-03
6.70255601e-01 -1.40612984e+00 -3.17464620e-01 5.96899211e-01
-1.13294971e+00 -4.25545305e-01 9.03735310e-02 4.74065930e-01
-7.19288588e-01 4.83910143e-01 -6.17756367e-01 -1.38444886e-01
-1.70994475e-01 -1.23423076e+00 1.50768399e+00 9.27752256e-02
1.39041245e-01 -7.23339498e-01 -2.51197129e-01 6.07237816e-01
3.46799046e-01 4.17408608e-02 6.87685072e-01 2.05007166e-01
-1.25288653e+00 -9.85567831e-03 -5.84697783e-01 3.00669104e-01
3.21825415e-01 -3.44046384e-01 -1.06664705e+00 -4.79364991e-01
1.84954241e-01 -1.44931704e-01 8.83466125e-01 3.53802919e-01
1.08327115e+00 -1.84225768e-01 -3.23133290e-01 1.37178075e+00
1.48765326e+00 7.76333734e-02 8.34853172e-01 4.03010994e-01
1.12077832e+00 4.75677490e-01 6.21875286e-01 2.06005022e-01
6.04094803e-01 1.05301929e+00 3.32875222e-01 -4.24188316e-01
-5.77721596e-01 -6.17990553e-01 3.51253927e-01 8.07511449e-01
3.07648182e-01 -7.71014066e-03 -6.53700352e-01 1.27902672e-01
-1.77129722e+00 -6.46136999e-01 1.31371349e-01 2.28083467e+00
7.91131377e-01 4.35609727e-05 -3.97413343e-01 -2.14208905e-02
7.94497311e-01 3.59207183e-01 -5.98719358e-01 3.39070946e-01
-4.24416691e-01 2.04191312e-01 6.98231101e-01 1.04376948e+00
-1.14259195e+00 1.02476919e+00 5.15854120e+00 9.22646105e-01
-1.26272476e+00 5.28395474e-02 7.46510327e-01 1.39446529e-02
-5.59553623e-01 3.61790746e-01 -9.06212389e-01 4.93276387e-01
2.92146474e-01 1.18226551e-01 3.94261658e-01 5.40954411e-01
5.69828935e-02 6.48088171e-04 -1.09639215e+00 1.23191702e+00
2.25430503e-01 -1.35293448e+00 1.97990760e-01 1.22302033e-01
9.97318208e-01 -9.82342567e-03 5.88787198e-02 -1.76439032e-01
-6.51175156e-02 -7.93832183e-01 8.89963746e-01 3.97400588e-01
1.09269440e+00 -4.63638157e-01 4.24983412e-01 1.27882659e-01
-1.46786308e+00 2.33362600e-01 -5.11244237e-01 1.17580950e-01
2.27390528e-01 7.42748141e-01 -1.78467423e-01 8.55363965e-01
5.07989526e-01 1.12641573e+00 -4.67930675e-01 9.69309866e-01
-3.22608411e-01 5.90209439e-02 -2.62446880e-01 8.24678898e-01
-1.11091346e-01 -3.82369041e-01 3.90446365e-01 5.73848784e-01
4.42821145e-01 6.75000018e-03 3.46505076e-01 9.21397269e-01
-2.17150986e-01 -1.61522523e-01 -4.52715963e-01 4.00950104e-01
6.42314434e-01 9.55364108e-01 -3.10571551e-01 -2.37358212e-01
-8.78456295e-01 1.29349506e+00 3.34482253e-01 3.61378670e-01
-6.32265210e-01 -2.42005922e-02 9.83667433e-01 4.51620668e-01
3.65448713e-01 -2.95445263e-01 -5.57458639e-01 -1.57684624e+00
3.72247875e-01 -9.02669191e-01 1.16928242e-01 -7.33707786e-01
-1.13961864e+00 4.68782753e-01 -1.93317339e-01 -1.62892473e+00
-9.50096454e-03 -5.72044671e-01 -4.72646922e-01 9.83813107e-01
-1.89807332e+00 -1.34633601e+00 -6.15738988e-01 6.92934036e-01
3.62581134e-01 4.24282961e-02 3.74943584e-01 7.76177883e-01
-3.36705714e-01 7.39612281e-01 -1.54604614e-01 9.73253101e-02
9.23368871e-01 -7.60663986e-01 6.12362266e-01 1.06207848e+00
-7.65862986e-02 4.85863000e-01 2.61543512e-01 -6.03437185e-01
-1.58075869e+00 -1.32417774e+00 7.32337058e-01 -2.08437592e-01
2.60051906e-01 -4.15728956e-01 -9.70764101e-01 6.68075740e-01
-5.07401675e-02 1.92551643e-01 1.50361016e-01 -4.23525214e-01
-5.32149792e-01 -2.47641400e-01 -8.67988944e-01 5.40713608e-01
1.43006885e+00 -8.40896726e-01 -1.81900412e-01 1.78881979e-03
6.81357741e-01 -9.13834274e-01 -8.64641726e-01 5.65350294e-01
5.20481229e-01 -1.18988252e+00 1.23603296e+00 7.91682154e-02
7.57910013e-01 -5.81710517e-01 -4.96233702e-01 -9.12046134e-01
-2.19272271e-01 -6.37113154e-01 -1.85738012e-01 1.12235570e+00
6.93470389e-02 -6.58901632e-01 1.12169981e+00 5.72393715e-01
-4.87834215e-01 -8.53024423e-01 -8.61685574e-01 -7.13887870e-01
-1.12497784e-01 -1.40915468e-01 7.14087725e-01 9.12532330e-01
-4.62548584e-01 9.18794498e-02 -6.52412236e-01 4.70798075e-01
9.39629436e-01 6.15273476e-01 9.49170649e-01 -8.45442772e-01
-4.27141339e-01 -3.94560963e-01 -4.88747478e-01 -1.88778067e+00
1.52462929e-01 -9.75215554e-01 2.92952489e-02 -1.18549967e+00
1.62149146e-01 -5.75525045e-01 3.68565507e-02 3.30123752e-01
-2.94442683e-01 4.72567409e-01 1.00843512e-01 4.02574748e-01
-2.17294469e-01 6.58845425e-01 1.59802175e+00 2.59278738e-03
8.70600417e-02 -1.94481939e-01 -6.12245560e-01 5.13546169e-01
6.73913479e-01 -3.57444078e-01 -3.35532874e-01 -9.04438257e-01
-2.10840687e-01 1.12130582e-01 3.61185253e-01 -9.87539113e-01
3.47100914e-01 -1.55942559e-01 5.13543129e-01 -6.41963482e-01
5.41498780e-01 -8.39897513e-01 4.44505543e-01 4.21291798e-01
-2.83747725e-02 -9.67092291e-02 -3.84381115e-02 2.90356994e-01
-6.18242979e-01 -3.33233401e-02 9.70935225e-01 1.14567615e-01
-6.06175303e-01 8.53976548e-01 3.13910574e-01 -7.57936314e-02
6.73656285e-01 -4.05868113e-01 -4.61354673e-01 -3.01882923e-01
-1.37538448e-01 2.56553203e-01 1.19719481e+00 3.81543458e-01
7.97154367e-01 -1.51521778e+00 -5.40215373e-01 7.14218259e-01
8.77211019e-02 4.27204490e-01 4.09875423e-01 7.47690737e-01
-6.61759794e-01 2.81337172e-01 -2.40733191e-01 -6.63184106e-01
-1.15183222e+00 3.89487565e-01 4.80501235e-01 5.51999547e-02
-7.83874393e-01 7.05285430e-01 8.48511040e-01 -3.71581823e-01
2.53814280e-01 -2.41071537e-01 3.76105964e-01 -4.26550299e-01
3.61740530e-01 6.23905882e-02 1.52640417e-01 -7.79770851e-01
-2.29788005e-01 1.20528173e+00 -1.21498488e-01 1.30499169e-01
1.00377858e+00 -2.96330243e-01 -1.05428986e-01 -1.00435451e-01
1.67514515e+00 1.29185662e-01 -1.65260196e+00 -6.69892192e-01
-2.99239039e-01 -1.00387716e+00 1.22799389e-01 -2.09434226e-01
-1.31875265e+00 8.49133968e-01 5.03756404e-01 -5.89953959e-01
1.29514527e+00 -5.12137748e-02 1.01912403e+00 1.56848934e-02
5.49016893e-01 -5.59391499e-01 -4.11310382e-02 5.04247427e-01
8.35038841e-01 -1.37121654e+00 1.32957414e-01 -1.10853410e+00
-1.54073194e-01 1.05383551e+00 7.13024855e-01 -2.00212494e-01
7.71127224e-01 2.26387262e-01 1.32571384e-01 -1.90554604e-01
-5.67433596e-01 -1.18984327e-01 5.68692505e-01 3.05131167e-01
4.13741022e-01 -2.05483939e-02 8.68093874e-03 1.53699145e-01
-2.35786542e-01 -1.22664496e-01 2.13753790e-01 9.51233506e-01
-2.39033908e-01 -1.35322797e+00 -3.41503382e-01 2.89854020e-01
-2.33819097e-01 -1.81102753e-01 -1.74536392e-01 4.90937293e-01
2.17096340e-02 7.72492528e-01 2.51266539e-01 -5.49338639e-01
3.37976635e-01 -3.40448588e-01 6.27523005e-01 -7.14193404e-01
3.82852480e-02 2.69491613e-01 -7.81860277e-02 -8.84762287e-01
-4.49127346e-01 -4.91151690e-01 -9.02828395e-01 -4.86862659e-01
-2.33645067e-01 -2.68278778e-01 2.50042051e-01 7.60792971e-01
5.02522349e-01 1.06259331e-01 9.38703239e-01 -1.13596475e+00
-2.05705971e-01 -6.30655050e-01 -3.70285600e-01 4.99148250e-01
4.21277285e-01 -8.22073221e-01 -3.88072103e-01 8.38520899e-02] | [8.730473518371582, -2.3664588928222656] |
4b9d676d-86b9-4f6f-ac98-2fc2f548925f | competition-based-resilience-in-distributed | 2203.14099 | null | https://arxiv.org/abs/2203.14099v3 | https://arxiv.org/pdf/2203.14099v3.pdf | Competition-Based Resilience in Distributed Quadratic Optimization | This paper proposes a novel approach to resilient distributed optimization with quadratic costs in a networked control system (e.g., wireless sensor network, power grid, robotic team) prone to external attacks (e.g., hacking, power outage) that cause agents to misbehave. Departing from classical filtering strategies proposed in literature, we draw inspiration from a game-theoretic formulation of the consensus problem and argue that adding competition to the mix can enhance resilience in the presence of malicious agents. Our intuition is corroborated by analytical and numerical results showing that i) our strategy highlights the presence of a nontrivial tradeoff between blind collaboration and full competition, and ii) such competition-based approach can outperform state-of-the-art algorithms based on Mean Subsequence Reduced. | ['Luca Schenato', 'Jeff S. Shamma', 'Giacomo Como', 'Luca Ballotta'] | 2022-03-26 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-7.77391940e-02 2.42808089e-01 3.43397826e-01 5.70311129e-01
-2.09793165e-01 -1.17685568e+00 4.73897427e-01 2.36968920e-01
-2.74461687e-01 7.79811680e-01 -1.07531194e-02 -4.38617527e-01
-8.13982725e-01 -7.56814659e-01 -4.02561128e-01 -9.75377262e-01
-7.30997145e-01 7.18374876e-03 2.29588240e-01 -6.59964979e-01
2.14877024e-01 2.62836099e-01 -9.24887598e-01 -6.95398927e-01
1.05015457e+00 1.00312328e+00 -3.34798135e-02 6.22519672e-01
5.54852307e-01 9.10654008e-01 -9.64922488e-01 -4.15819287e-01
9.36006427e-01 -3.33810538e-01 -4.46962476e-01 2.04640388e-01
-3.44117403e-01 -3.47981274e-01 -4.57430035e-01 1.76166916e+00
5.96002221e-01 -1.00213066e-01 3.07537019e-01 -1.70105970e+00
-7.09047377e-01 1.04108298e+00 -6.04134262e-01 -3.50427479e-02
1.88172549e-01 3.89033765e-01 7.54729509e-01 -1.79444864e-01
6.27983749e-01 8.88992429e-01 6.99195921e-01 5.75963259e-01
-1.06880164e+00 -7.63505518e-01 2.82081336e-01 3.40393209e-03
-1.26577389e+00 -1.62185773e-01 5.73949873e-01 -1.72263116e-01
3.69941205e-01 3.25414240e-01 5.75540543e-01 7.58195341e-01
4.23596293e-01 1.81387275e-01 1.11219800e+00 -1.57587186e-01
8.20490479e-01 -3.70978564e-01 -2.39986420e-01 3.00228059e-01
1.01223934e+00 4.30410713e-01 -3.94176245e-01 -6.42360389e-01
4.00681645e-01 -2.38179609e-01 -5.26213109e-01 -5.15108287e-01
-9.67404664e-01 8.01038265e-01 4.59883273e-01 1.15903080e-01
-7.50557303e-01 6.39541969e-02 4.01828557e-01 9.99579728e-01
1.54967725e-01 4.88715082e-01 -3.61828715e-01 1.73563868e-01
-5.47293782e-01 1.29642710e-01 1.14944589e+00 7.75036573e-01
2.71342605e-01 4.14255857e-01 1.26312479e-01 -2.25004330e-02
4.88204271e-01 5.60036123e-01 -9.82666165e-02 -1.16546774e+00
2.04259276e-01 2.52693087e-01 6.03151798e-01 -1.16467798e+00
-4.54234511e-01 -5.74911952e-01 -1.36140347e+00 6.46518171e-01
3.60188276e-01 -9.76054132e-01 -1.27310395e-01 1.78536332e+00
3.61765623e-01 1.60180211e-01 3.78841877e-01 1.02936935e+00
-2.36787573e-02 9.45242345e-02 -3.44198614e-01 -7.32832074e-01
8.25782895e-01 -4.60487425e-01 -8.07380259e-01 5.93140498e-02
6.05366379e-02 -4.70126599e-01 -4.49863113e-02 7.51542389e-01
-9.42364097e-01 3.40103656e-01 -1.21601367e+00 1.05148852e+00
-1.86519504e-01 -6.71371520e-01 4.01385069e-01 1.14109600e+00
-1.39586949e+00 5.85977793e-01 -6.41200244e-01 -6.07577443e-01
2.32061327e-01 5.02655745e-01 -1.69923320e-01 2.77233392e-01
-9.73710775e-01 8.69033992e-01 2.13011414e-01 4.33974480e-03
-1.19685411e+00 -4.24758285e-01 -2.02866480e-01 -8.41779858e-02
9.97008085e-01 -8.16325665e-01 9.38112736e-01 -9.83071089e-01
-1.65351880e+00 3.93870585e-02 8.05355370e-01 -8.57864261e-01
8.52448106e-01 6.06829599e-02 -5.84722534e-02 4.06370789e-01
-1.36217624e-01 -2.36660182e-01 9.56895113e-01 -1.31357682e+00
-4.52091426e-01 -2.78889447e-01 4.46639091e-01 8.65420699e-02
-6.30039811e-01 2.27039725e-01 6.38477921e-01 -6.93844140e-01
-1.28956780e-01 -9.02038574e-01 -9.52607691e-01 8.39562416e-02
-5.46445072e-01 1.08732417e-01 8.77006888e-01 -1.71469405e-01
8.22769582e-01 -1.85761094e+00 3.82986486e-01 5.38571954e-01
6.46016240e-01 2.64867663e-01 -1.28952980e-01 1.10379326e+00
1.64000511e-01 3.79804552e-01 1.35685489e-01 -1.62474532e-02
4.33463812e-01 2.18519941e-01 -1.09426484e-01 1.09733021e+00
-3.88644546e-01 3.53436291e-01 -1.07264030e+00 2.07028762e-01
1.41813368e-01 -2.78520770e-02 -3.47950459e-01 1.74388915e-01
4.18153703e-02 4.72571552e-01 -6.13396049e-01 4.19136882e-01
7.54146338e-01 -2.15630904e-01 6.20255172e-01 1.29548296e-01
-2.61056304e-01 -2.38417104e-01 -1.70017028e+00 1.21831965e+00
4.23174538e-02 8.39191303e-02 1.24304998e+00 -1.16727698e+00
5.77198148e-01 4.80903029e-01 5.90510607e-01 -1.99957326e-01
6.35806739e-01 -6.68673369e-04 2.02150136e-01 -1.94329291e-03
2.78547764e-01 2.40237519e-01 -2.35402063e-01 7.92444110e-01
3.79354395e-02 -6.49143010e-04 6.05949014e-02 6.19979978e-01
1.80178833e+00 -7.29024053e-01 5.51322043e-01 -7.48773277e-01
3.92535895e-01 -7.62246400e-02 6.36525869e-01 1.38469696e+00
-5.33687770e-01 -1.85276583e-01 5.73237538e-01 3.15788835e-02
-7.47172594e-01 -8.60274255e-01 5.13346732e-01 6.35565758e-01
5.38918257e-01 -5.61890364e-01 -7.21289694e-01 -3.88083190e-01
8.38806480e-02 4.75164503e-02 -4.76953238e-01 -1.35243043e-01
8.93400982e-03 -9.09107804e-01 5.71171105e-01 -2.97166631e-02
5.48466027e-01 -2.83798516e-01 -7.99629390e-01 4.58323181e-01
2.45420277e-01 -1.01524091e+00 -4.23405409e-01 2.25772828e-01
-4.72734034e-01 -1.52057362e+00 -1.67577773e-01 -1.89027384e-01
6.28015041e-01 6.23923302e-01 8.45634043e-01 4.43424016e-01
-5.92837408e-02 1.09990406e+00 -6.34875178e-01 -3.36746842e-01
-6.34830415e-01 -3.25697541e-01 5.48511624e-01 -4.10983935e-02
-4.05289292e-01 -7.72880197e-01 -8.04190695e-01 4.09130573e-01
-9.81528282e-01 -5.67893565e-01 2.83045352e-01 6.99670494e-01
-2.95527995e-01 5.25541842e-01 8.83207440e-01 -2.42770121e-01
1.23918366e+00 -6.72669172e-01 -1.32321584e+00 1.56065226e-01
-7.05847740e-01 -3.49347442e-01 1.00798321e+00 -5.16498864e-01
-4.42460001e-01 1.00755170e-01 6.31615996e-01 -2.72597760e-01
1.04860321e-01 4.43809569e-01 5.66395409e-02 -9.23342526e-01
5.97676754e-01 -4.11737524e-02 4.75009143e-01 2.01985184e-02
7.09901869e-01 4.75273192e-01 3.34919423e-01 -3.72204334e-01
1.38885355e+00 6.03266537e-01 2.50848055e-01 -5.78292727e-01
4.54022698e-02 6.80266470e-02 -1.43670291e-01 -4.88452941e-01
3.87204230e-01 -9.01007295e-01 -1.42791247e+00 6.67059302e-01
-1.03897107e+00 -1.12688862e-01 -2.04070166e-01 2.02325970e-01
-2.85778403e-01 6.56294763e-01 -6.84586465e-01 -1.27686155e+00
-4.09614921e-01 -7.68022895e-01 1.85389236e-01 5.42406499e-01
2.23294437e-01 -9.21354234e-01 1.17405675e-01 -8.20618793e-02
9.70316529e-01 4.03943598e-01 1.37315542e-01 -8.50839138e-01
-8.05324137e-01 -2.32113943e-01 2.05855042e-01 1.88394144e-01
1.58294261e-01 1.09830104e-01 -2.28971913e-01 -8.83820176e-01
1.83022246e-01 -2.28769660e-01 6.08816221e-02 1.53069630e-01
3.02977622e-01 -8.89325321e-01 -1.34125784e-01 1.20911583e-01
1.66625702e+00 6.07032795e-03 2.24579424e-01 1.92290217e-01
7.69569799e-02 4.11891878e-01 3.22646707e-01 9.90628242e-01
4.52350110e-01 3.46345842e-01 1.21137822e+00 2.90300041e-01
5.07593393e-01 2.44390160e-01 5.04614055e-01 6.81494415e-01
8.74857724e-05 -3.98941725e-01 -4.86409009e-01 4.98569429e-01
-2.14260221e+00 -8.47851098e-01 -1.88643355e-02 2.38213158e+00
3.21218133e-01 -1.90883724e-03 4.44038033e-01 7.46870860e-02
9.87519085e-01 1.07079409e-02 -4.49806571e-01 -1.07308984e-01
-4.11490709e-01 -2.00073704e-01 1.19781971e+00 3.60893011e-01
-8.93576920e-01 4.62451607e-01 6.61232185e+00 6.13088548e-01
-8.58731389e-01 4.32025731e-01 1.15318075e-01 4.78263721e-02
9.86470655e-02 4.47417498e-02 9.77674648e-02 4.13630337e-01
6.85890019e-01 -8.61724973e-01 1.07537878e+00 3.66524637e-01
3.61751169e-01 -1.37662724e-01 -4.65597034e-01 5.51340401e-01
-8.63981694e-02 -1.20233011e+00 -6.60218537e-01 3.83465171e-01
9.89957750e-01 2.93370873e-01 -1.87215671e-01 -4.46097493e-01
1.48594558e+00 -6.75588012e-01 6.74215019e-01 1.73875421e-01
-1.61269322e-01 -6.95140719e-01 5.87541938e-01 4.06893075e-01
-1.07266891e+00 -7.20197201e-01 -1.44831941e-01 -4.17292237e-01
2.29319841e-01 5.42864501e-01 -1.38740703e-01 1.14294076e+00
7.15186536e-01 3.35854918e-01 -2.63749540e-01 1.35389304e+00
-3.79823267e-01 4.81890857e-01 -6.67078018e-01 -1.77908003e-01
3.26112568e-01 -3.57064635e-01 1.20331776e+00 3.33048522e-01
2.16792449e-01 2.24710777e-01 5.23125410e-01 6.65601492e-01
-2.66399477e-02 -1.82975665e-01 -8.10499191e-01 2.00422183e-01
9.34745669e-01 1.42012107e+00 -7.89313912e-01 -8.69074091e-02
-3.46265197e-01 8.33237469e-01 -6.64553866e-02 4.02987331e-01
-4.99693602e-01 -3.56043845e-01 8.20387602e-01 -5.10713041e-01
5.06752670e-01 -4.40402120e-01 -2.81222582e-01 -9.65246260e-01
-9.73028988e-02 -1.11087227e+00 3.21954995e-01 -4.11821336e-01
-1.69325054e+00 2.51399815e-01 -4.28680807e-01 -1.36853600e+00
-2.43145227e-02 -1.44042954e-01 -8.78834784e-01 4.02817398e-01
-1.04233873e+00 -6.45855784e-01 1.20291725e-01 7.60148883e-01
-2.28517935e-01 -2.46520996e-01 6.89018250e-01 1.17126308e-01
-4.96108025e-01 2.30809420e-01 5.78949213e-01 6.26245812e-02
3.52890790e-01 -1.23368239e+00 -5.13390824e-02 1.41851413e+00
-3.87258470e-01 5.81966043e-01 1.10154974e+00 -6.93320513e-01
-2.09824491e+00 -8.92938733e-01 6.20955937e-02 -2.70478521e-02
1.44415700e+00 -3.14305991e-01 -1.85080498e-01 1.98433518e-01
8.00835729e-01 5.18192835e-02 4.47361499e-01 -4.21205580e-01
-1.73117548e-01 -1.53076664e-01 -1.61916482e+00 7.42850542e-01
8.90216172e-01 -1.65036201e-01 -2.59028047e-01 4.50762361e-01
4.51814085e-01 -1.74905524e-01 -8.56921017e-01 2.79088016e-03
2.28711456e-01 -8.16740394e-01 7.60286510e-01 -3.56885791e-01
-4.31901515e-01 -6.62704647e-01 -1.59591630e-01 -1.77304506e+00
-1.67208225e-01 -1.75206399e+00 -4.91096117e-02 8.93157363e-01
-1.90823689e-01 -1.17425680e+00 4.63298291e-01 3.79779547e-01
2.82379687e-01 1.27373710e-01 -1.45943069e+00 -1.10739458e+00
3.02677929e-01 1.37804464e-01 4.69195753e-01 9.51795816e-01
5.06452322e-01 -1.57777388e-02 -6.08300269e-01 9.43264186e-01
1.26049662e+00 -3.55756760e-01 7.05433607e-01 -1.23499012e+00
-1.80756927e-01 -4.34148967e-01 -4.67496306e-01 -5.74667454e-01
-9.54110175e-02 -2.24978238e-01 -4.58117872e-02 -1.16389894e+00
-2.84336299e-01 -2.41556078e-01 -2.63422310e-01 3.86205256e-01
5.06544486e-02 2.15694681e-01 7.13071227e-01 8.63736670e-04
-1.13686728e+00 3.69640052e-01 6.26190126e-01 -1.97619677e-01
5.18322922e-02 -1.11470573e-01 -1.00514138e+00 3.90267909e-01
9.07606184e-01 -7.22687364e-01 -2.72298217e-01 -1.31264612e-01
4.22884822e-01 4.87467080e-01 5.08923233e-01 -8.20551515e-01
1.00360525e+00 -2.49555826e-01 -5.57680607e-01 1.53976217e-01
-2.24809662e-01 -1.29443562e+00 2.64424354e-01 1.11892581e+00
4.45522703e-02 1.11460999e-01 -2.35381976e-01 1.00549018e+00
3.17911118e-01 -5.52034453e-02 5.58646739e-01 -1.70577019e-01
-8.72691870e-02 8.32505599e-02 -1.15491152e+00 -1.02084979e-01
1.18272066e+00 1.93979010e-01 -8.97169948e-01 -8.06271136e-01
-6.60534799e-01 8.06714475e-01 6.33517683e-01 3.84978741e-01
2.86423564e-02 -8.26434553e-01 -8.15926909e-01 -3.46969604e-01
-4.31498766e-01 -5.50373614e-01 3.10308523e-02 1.07997274e+00
-1.55203849e-01 -1.42712384e-01 -1.73849002e-01 -2.22998217e-01
-1.11680067e+00 5.81400096e-01 5.80550075e-01 -6.63023442e-02
-4.00332630e-01 3.03046286e-01 -5.64432859e-01 -2.29548097e-01
3.79429162e-01 9.09148827e-02 2.50294089e-01 -3.91157866e-02
4.58022565e-01 7.28332400e-01 4.56339605e-02 -2.63448894e-01
-5.51434040e-01 1.84271827e-01 2.63462126e-01 -5.52695319e-02
1.21916819e+00 -6.02904558e-01 -5.54642737e-01 -3.67282391e-01
1.97542027e-01 8.54759961e-02 -1.00609624e+00 -1.42880172e-01
-1.10705066e-02 -4.51864421e-01 8.97940248e-02 -8.19772959e-01
-1.36609316e+00 -6.63865656e-02 5.64857662e-01 1.08298326e+00
1.10265183e+00 -3.48281890e-01 1.02397211e-01 5.06308973e-01
9.82308447e-01 -1.05780280e+00 5.71771823e-02 3.80669773e-01
4.25367534e-01 -7.79004157e-01 6.21701777e-02 -3.70191902e-01
-2.83892035e-01 9.82644379e-01 1.78216696e-01 -6.10185325e-01
8.47843468e-01 6.89145625e-01 1.61721632e-02 -9.10515934e-02
-8.94389093e-01 -2.01855570e-01 -5.22262096e-01 1.01312530e+00
-5.26879370e-01 2.40223497e-01 -6.80552244e-01 6.92692757e-01
2.11567104e-01 -2.28431851e-01 1.45021737e+00 1.13798869e+00
-5.16298950e-01 -9.64371741e-01 -6.45337462e-01 3.29367608e-01
-7.57683694e-01 1.28075734e-01 -5.64602554e-01 3.85871261e-01
-2.11311087e-01 1.81511784e+00 -3.91667753e-01 -3.57302815e-01
1.61133677e-01 -6.78947151e-01 1.30525485e-01 -1.00268967e-01
-1.09362578e+00 9.29446518e-02 -1.12806730e-01 -5.73635042e-01
-6.68249011e-01 -5.34508586e-01 -5.81150949e-01 -5.20788491e-01
-6.97868109e-01 5.22679269e-01 4.02785212e-01 7.97132015e-01
5.08113921e-01 4.41375494e-01 1.02757335e+00 -6.30820513e-01
-1.14485097e+00 -6.03827596e-01 -8.43289256e-01 -1.93354160e-01
4.50803250e-01 -5.18826902e-01 -8.98439050e-01 -4.96132731e-01] | [4.605893611907959, 2.795121669769287] |
40726cd6-f2fa-4eac-ac7f-9179d3f76759 | a-multimodal-dynamical-variational | 2305.03582 | null | https://arxiv.org/abs/2305.03582v1 | https://arxiv.org/pdf/2305.03582v1.pdf | A Multimodal Dynamical Variational Autoencoder for Audiovisual Speech Representation Learning | In this paper, we present a multimodal \textit{and} dynamical VAE (MDVAE) applied to unsupervised audio-visual speech representation learning. The latent space is structured to dissociate the latent dynamical factors that are shared between the modalities from those that are specific to each modality. A static latent variable is also introduced to encode the information that is constant over time within an audiovisual speech sequence. The model is trained in an unsupervised manner on an audiovisual emotional speech dataset, in two stages. In the first stage, a vector quantized VAE (VQ-VAE) is learned independently for each modality, without temporal modeling. The second stage consists in learning the MDVAE model on the intermediate representation of the VQ-VAEs before quantization. The disentanglement between static versus dynamical and modality-specific versus modality-common information occurs during this second training stage. Extensive experiments are conducted to investigate how audiovisual speech latent factors are encoded in the latent space of MDVAE. These experiments include manipulating audiovisual speech, audiovisual facial image denoising, and audiovisual speech emotion recognition. The results show that MDVAE effectively combines the audio and visual information in its latent space. They also show that the learned static representation of audiovisual speech can be used for emotion recognition with few labeled data, and with better accuracy compared with unimodal baselines and a state-of-the-art supervised model based on an audiovisual transformer architecture. | ['Renaud Séguier', 'Xavier Alameda-Pineda', 'Laurent Girin', 'Simon Leglaive', 'Samir Sadok'] | 2023-05-05 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [ 1.3401705e-01 -1.5493090e-02 -1.5233368e-01 -3.1358910e-01
-6.7539191e-01 -6.3466507e-01 7.4985093e-01 -2.1589148e-01
-2.9266748e-01 1.2199735e-01 5.2126205e-01 -4.5715183e-02
2.4410667e-01 -2.8039744e-01 -6.2905073e-01 -1.1606362e+00
-7.7685136e-03 2.1014677e-01 -1.8055092e-01 -1.5898213e-02
-3.9095971e-01 4.9547133e-01 -1.9038861e+00 7.0918757e-01
2.9555899e-01 1.2461278e+00 1.7399552e-01 9.4421387e-01
-4.1286311e-01 9.8104692e-01 -4.9582106e-01 -8.7611035e-02
-7.6345988e-02 -4.6060005e-01 -5.1827198e-01 3.6404747e-01
3.0910733e-01 -1.6782250e-01 -6.5579385e-01 9.3065906e-01
3.4531432e-01 2.8906080e-01 1.0254096e+00 -1.6005757e+00
-5.2773196e-01 2.9743466e-01 -2.5084409e-01 4.8276227e-02
2.6109731e-01 5.9222419e-02 1.0204535e+00 -1.0600290e+00
7.6544833e-01 1.5711617e+00 1.1413860e-01 7.5588703e-01
-1.4958022e+00 -5.7796502e-01 3.6281192e-01 4.3350431e-01
-1.3684999e+00 -9.3406326e-01 1.1253520e+00 -8.9415640e-01
1.0909907e+00 3.3747745e-01 7.4978375e-01 1.6316363e+00
3.5639140e-01 1.1560302e+00 9.4514710e-01 -3.7497351e-01
3.9754435e-01 1.6577901e-01 -4.3418281e-02 5.1157773e-01
-1.0987586e+00 4.1789323e-01 -9.4739050e-01 -2.1218343e-01
4.3915159e-01 -2.3857257e-01 -2.3995756e-01 -6.1718237e-01
-7.2647476e-01 1.0918466e+00 -1.1154353e-02 3.3495253e-01
-5.4061836e-01 1.8716894e-01 7.9431361e-01 6.9472504e-01
4.6813297e-01 -3.5378128e-01 -3.6709240e-01 -3.0207852e-01
-1.0549634e+00 -3.2073465e-01 4.8143089e-01 5.7642597e-01
5.2856565e-01 6.5464944e-01 -2.6447225e-01 9.1450435e-01
7.1095830e-01 7.4087882e-01 5.9061444e-01 -9.9471492e-01
2.2529215e-01 2.0206720e-01 -3.2239702e-01 -9.0603542e-01
3.0052722e-02 -2.8518716e-02 -8.8957471e-01 3.3738291e-01
-2.7879202e-01 2.4356183e-01 -1.2942853e+00 2.0700216e+00
7.9331122e-02 2.3991394e-01 3.1366962e-01 6.2511927e-01
1.0871305e+00 1.1661317e+00 2.8389719e-01 -7.0382506e-01
1.2759813e+00 -7.6213616e-01 -1.3120909e+00 1.0974018e-02
4.3546744e-02 -6.4287841e-01 8.6648571e-01 4.7285613e-01
-1.0099052e+00 -5.8077949e-01 -8.6193949e-01 -1.6175723e-01
-4.1271290e-01 2.6924106e-01 3.1448171e-01 4.7237998e-01
-1.2811407e+00 -1.5822344e-01 -1.0794238e+00 -1.6197039e-01
3.4223523e-02 1.9799641e-01 -6.9036317e-01 3.0275589e-01
-1.3349324e+00 4.9733362e-01 1.5549532e-01 1.0485261e-01
-1.6532753e+00 -3.7173185e-01 -1.3594698e+00 1.8780698e-01
7.5891100e-02 -4.2399946e-01 1.1783570e+00 -1.4976970e+00
-1.7692875e+00 6.8253869e-01 -6.8340003e-01 -1.1789062e-01
1.3704413e-01 2.8578258e-01 -5.2132416e-01 5.9341836e-01
-2.5046015e-01 8.5922527e-01 1.6357142e+00 -1.6209767e+00
-2.2050461e-01 -3.5246080e-01 -4.4765443e-01 1.3016728e-01
-3.7525243e-01 2.2049017e-01 -6.5504682e-01 -8.0029088e-01
7.6037566e-03 -8.2265198e-01 4.7717503e-01 7.5809181e-02
1.9422486e-02 -2.3368721e-01 1.1808567e+00 -7.9116488e-01
1.1165364e+00 -2.5282955e+00 8.9673287e-01 1.5588385e-01
1.2316330e-01 -1.9261112e-02 -4.8078367e-01 4.6216902e-01
-4.9711207e-01 -1.5646459e-01 -8.1164666e-02 -8.6997473e-01
1.7865506e-01 5.4388940e-01 -8.7627280e-01 3.9852941e-01
2.3539396e-01 8.3226800e-01 -6.2338364e-01 -3.6389303e-01
3.2723963e-01 1.1131285e+00 -3.7468874e-01 2.7092001e-01
-1.6570921e-01 6.4342153e-01 5.7865474e-02 7.1403545e-01
5.2783132e-01 1.6135629e-01 1.0023411e-01 -4.2178780e-01
-2.2882406e-02 3.0351644e-02 -9.4935572e-01 1.6508522e+00
-4.3221945e-01 1.0206857e+00 5.3537518e-01 -1.0074853e+00
6.2421924e-01 9.4814759e-01 6.4498246e-01 -7.8539777e-01
3.2111579e-01 -2.5977084e-01 -4.0571955e-01 -5.5425251e-01
3.3610979e-01 -4.8470226e-01 -1.1790510e-01 1.0531602e-01
8.9128619e-01 1.0415272e-01 -1.7836562e-01 4.2262781e-01
6.7278659e-01 -1.0012652e-01 -8.0324244e-03 2.5474888e-01
6.3294560e-01 -6.2718660e-01 5.1904398e-01 2.7878729e-01
-4.0993923e-01 1.8471219e-01 6.2780690e-01 1.4143360e-01
-8.2678616e-01 -1.4455156e+00 -1.2368236e-01 1.3522087e+00
-1.7439684e-01 -5.9185487e-01 -3.6968124e-01 -3.9929250e-01
-2.3072955e-01 6.6195458e-01 -8.4280539e-01 -4.3936604e-01
4.1181739e-02 -4.9857248e-02 6.4397442e-01 5.3996718e-01
1.1752116e-01 -1.0187918e+00 -2.0910262e-01 -3.6729950e-02
-4.7850621e-01 -1.0845367e+00 -3.0113742e-01 3.8151556e-01
-4.5571870e-01 -3.7173527e-01 -6.3186407e-01 -6.6698158e-01
3.6412215e-01 -1.4863305e-01 7.5161803e-01 -5.8031052e-01
-1.4518054e-01 1.1174653e+00 -4.2372444e-01 -1.8238726e-01
-5.9536463e-01 -5.6255645e-01 3.1568095e-01 6.0031295e-01
3.4801129e-01 -5.5242050e-01 -9.0129673e-02 8.1426568e-02
-1.2316663e+00 -2.1869110e-01 2.1389279e-01 1.0379391e+00
6.8973613e-01 2.6755333e-02 5.4346234e-02 3.8731616e-02
3.6105585e-01 -4.0885159e-01 -3.5276908e-01 2.5745657e-01
2.5010752e-02 1.9236290e-01 2.9352641e-01 -9.0156323e-01
-1.0676912e+00 3.2884905e-01 -2.7825636e-01 -1.2505579e+00
-1.6405427e-01 6.5307117e-01 -5.0137454e-01 2.2724570e-01
1.6277154e-01 5.3518623e-01 2.3871045e-01 -3.3301350e-01
6.1662143e-01 7.2684449e-01 6.3051277e-01 -3.6225718e-01
5.1161104e-01 5.8003545e-01 -3.6094072e-01 -1.3561887e+00
-1.8780659e-01 -5.7008326e-01 -6.3662344e-01 -5.2207762e-01
1.2770821e+00 -1.1789478e+00 -6.5828842e-01 6.3724482e-01
-1.1987286e+00 -4.0931529e-01 -4.4797212e-01 5.7146931e-01
-6.4288139e-01 4.6878302e-01 -8.3043110e-01 -1.0816725e+00
5.2687004e-02 -1.3883246e+00 1.3999646e+00 -2.2965293e-01
-1.7187396e-01 -1.0961312e+00 2.0200829e-01 1.8741623e-01
4.9786914e-02 -4.0021028e-02 9.1980147e-01 -2.6141244e-01
-2.7910024e-01 2.2297522e-01 2.3284985e-01 6.9719905e-01
1.7671683e-01 2.6020375e-01 -1.4639426e+00 -5.0674045e-01
1.4341307e-01 -6.9509220e-01 1.2453015e+00 4.4128898e-01
7.3329586e-01 -3.3131179e-01 -3.9821930e-02 6.7723376e-01
9.3839800e-01 3.3718720e-01 4.9852338e-01 -2.8217429e-01
4.6706456e-01 6.3488948e-01 3.2203987e-01 2.9383552e-01
3.1849471e-01 8.1591409e-01 5.4188848e-01 -8.3339617e-02
-1.5383968e-01 -2.0407391e-01 1.0871739e+00 1.3778138e+00
2.3208134e-01 -3.0913937e-01 -7.0846069e-01 5.2719784e-01
-1.6927434e+00 -1.2571766e+00 5.0554454e-01 1.8181573e+00
6.2690777e-01 -3.4407851e-01 2.4311323e-02 2.0490403e-01
4.0569094e-01 4.4456410e-01 -4.7629732e-01 -6.0904324e-01
-5.0971878e-01 9.4712645e-02 -2.9438314e-01 8.9397800e-01
-1.2052832e+00 1.1212907e+00 6.2326493e+00 8.0474192e-01
-1.6338868e+00 3.1749138e-01 2.2283979e-02 -1.9938889e-01
-5.2331430e-01 -1.7394370e-01 -4.1789448e-01 2.6305771e-01
1.1729416e+00 1.7918199e-01 4.7711182e-01 5.2157021e-01
2.4317573e-01 3.0468005e-01 -1.0315775e+00 1.3018614e+00
3.5681796e-01 -9.2338461e-01 4.7104308e-01 1.7622890e-01
2.9247236e-01 -1.9679913e-02 5.7646924e-01 5.2030200e-01
7.2930805e-02 -1.0688334e+00 1.0575837e+00 7.5453079e-01
9.4786608e-01 -6.7205375e-01 3.5596317e-01 2.9604462e-01
-1.5559447e+00 -1.5065721e-01 -1.3284531e-01 4.5540091e-01
2.3970781e-01 9.0302685e-03 -3.8310766e-01 3.9758015e-01
7.9371506e-01 7.1651208e-01 -2.7167293e-01 2.7347827e-01
-3.5874400e-01 7.5804645e-01 -7.8662328e-02 5.2552468e-01
1.8883635e-01 -1.5915310e-01 8.7237287e-01 1.2019135e+00
1.6833098e-01 -1.1981452e-02 8.4286436e-02 6.8265301e-01
1.7484570e-01 7.1943656e-02 -7.1299970e-01 -5.0867051e-01
2.4998547e-01 1.0309985e+00 -2.9782456e-01 -3.5187575e-01
-3.0462682e-01 1.4151973e+00 8.7108538e-02 7.8979921e-01
-8.6387056e-01 8.6297609e-02 8.2020760e-01 -5.1073867e-01
6.8598628e-01 -3.4849653e-01 4.9274072e-01 -1.3866305e+00
-1.6524541e-01 -9.0358466e-01 4.1545510e-01 -1.0020479e+00
-1.1595738e+00 8.9164001e-01 6.0683358e-02 -1.1448327e+00
-7.4470747e-01 -5.9931809e-01 -2.5513580e-01 8.1412113e-01
-1.4165223e+00 -1.3859390e+00 -3.4331623e-02 1.1391103e+00
5.5927688e-01 -5.3721416e-01 1.0740482e+00 1.9437039e-01
-2.7894557e-01 7.2913694e-01 1.7158622e-01 9.4979994e-02
6.6654027e-01 -9.7130191e-01 -2.6702273e-01 6.5253514e-01
4.5291507e-01 6.5174848e-01 6.7725825e-01 -4.8065403e-01
-1.4113560e+00 -8.6333990e-01 7.7945405e-01 -2.3726153e-01
6.4814013e-01 -8.5282767e-01 -9.8510063e-01 7.7912307e-01
4.7061428e-01 4.8851524e-02 1.1037805e+00 -1.1111013e-01
-8.3372897e-01 -3.9690226e-02 -7.2916740e-01 3.3046740e-01
4.1471192e-01 -1.5261991e+00 -6.8125457e-01 -1.4998859e-01
9.1519624e-01 -1.1799529e-01 -8.0749178e-01 3.4642813e-01
5.9127599e-01 -5.0679147e-01 9.3972874e-01 -6.8546706e-01
2.2805515e-01 -2.1542780e-01 -7.2151369e-01 -1.4589200e+00
-1.8649822e-01 -6.0043377e-01 -3.9215806e-01 1.4030699e+00
4.7377922e-02 -2.7106538e-01 2.3850721e-01 -4.2821202e-02
7.7303708e-02 -2.8844860e-01 -1.4226978e+00 -7.1022326e-01
6.7007490e-02 -7.6089352e-01 5.4455101e-02 1.1596886e+00
-1.3568527e-01 5.5868983e-01 -5.5973774e-01 3.7903485e-01
4.2594838e-01 -1.2652926e-01 4.7938538e-01 -1.0922472e+00
-3.3443484e-01 -3.4268141e-01 -7.3211068e-01 -1.0187558e+00
7.6905119e-01 -9.8720300e-01 8.5510701e-02 -1.1623162e+00
1.6211776e-01 4.7918856e-01 -4.8132095e-01 7.2153097e-01
3.0477202e-01 4.9618436e-03 3.3066589e-01 7.4303769e-02
-3.4097868e-01 1.2854437e+00 8.0133229e-01 -5.9773439e-01
-2.9209971e-01 -5.0646156e-01 -4.3010104e-02 6.5545499e-01
2.2742923e-01 -2.4527100e-01 -7.7397722e-01 -1.6677238e-01
-6.7116760e-02 4.8262158e-01 5.4387546e-01 -4.8279434e-01
2.5152558e-01 3.9296277e-02 1.6351652e-01 -6.6787386e-01
1.0629815e+00 -9.1473001e-01 1.5369038e-01 6.7596495e-02
-4.3644306e-01 -1.9927847e-01 4.9679419e-01 6.3143337e-01
-8.1766963e-01 2.9948062e-01 7.8171295e-01 3.1196335e-01
-7.1607250e-01 3.7926143e-01 -1.0039113e+00 -3.4248161e-01
7.9594499e-01 -7.3944494e-02 1.1326550e-01 -8.9488435e-01
-1.4869877e+00 -6.0422592e-02 1.5715379e-01 7.1189708e-01
9.0179753e-01 -1.6154805e+00 -5.1763290e-01 5.1604724e-01
4.2489159e-01 -7.4846047e-01 6.4760303e-01 7.2124422e-01
2.3934947e-01 3.0887654e-01 -1.9306932e-01 -1.1196920e+00
-1.7882903e+00 7.2457427e-01 3.5961404e-01 -1.2101802e-02
-4.1410172e-01 8.5690290e-01 5.1650566e-01 -4.7272885e-01
7.0166975e-01 -1.9673637e-01 -3.2392055e-01 7.4410021e-01
4.5249876e-01 -1.4646769e-01 -1.7798840e-01 -1.4858712e+00
-3.9036268e-01 5.7022065e-01 1.6890396e-01 -8.5179985e-01
1.2152860e+00 -3.3581057e-01 -1.6162835e-01 1.1310179e+00
1.6013255e+00 7.1180232e-02 -1.2662728e+00 -2.7432090e-01
-5.2225471e-01 -1.3500267e-01 3.7596855e-01 -5.7088798e-01
-1.1651911e+00 1.4172240e+00 9.1395420e-01 3.2953106e-02
1.4417766e+00 1.9996199e-01 3.0030376e-01 1.0638861e-02
-2.0174177e-02 -9.5822990e-01 4.3138450e-01 6.9927007e-01
1.2041725e+00 -1.0013126e+00 -5.6199235e-01 -2.0838696e-01
-1.0428662e+00 1.0638976e+00 1.4055401e-01 4.3838975e-01
8.5144532e-01 3.5917529e-01 3.7010673e-01 -3.0347964e-01
-1.1629393e+00 -1.7103848e-01 6.2596709e-01 4.5344341e-01
1.8084849e-01 -6.5380357e-02 3.3988819e-01 5.2454275e-01
-4.1290473e-02 -3.6386779e-01 7.7366635e-02 8.6478847e-01
-3.0829176e-02 -1.0147883e+00 -3.4461480e-01 -3.6350599e-01
-1.7046928e-01 1.3468954e-02 -6.6975749e-01 5.1194370e-01
8.5762531e-02 1.0154910e+00 3.1947055e-01 -5.7756317e-01
6.1386369e-02 5.8170277e-01 5.1717246e-01 -2.0529570e-01
-3.2850292e-01 8.1197071e-01 -2.1578179e-01 -6.9795287e-01
-5.0147361e-01 -6.7134166e-01 -1.2709470e+00 2.4981961e-01
2.2925327e-02 2.2582594e-01 8.2262135e-01 8.2354808e-01
2.5952607e-01 6.8342263e-01 7.4077922e-01 -9.2104012e-01
-2.1196514e-01 -6.8779683e-01 -6.4110744e-01 4.1396698e-01
9.7132963e-01 -8.2562965e-01 -7.8393924e-01 4.3361375e-01] | [14.207903861999512, 5.182390213012695] |
f9486f6f-896f-45a6-81ef-0589e7acae4b | 191013439 | 1910.13439 | null | https://arxiv.org/abs/1910.13439v2 | https://arxiv.org/pdf/1910.13439v2.pdf | Learning to Manipulate Deformable Objects without Demonstrations | In this paper we tackle the problem of deformable object manipulation through model-free visual reinforcement learning (RL). In order to circumvent the sample inefficiency of RL, we propose two key ideas that accelerate learning. First, we propose an iterative pick-place action space that encodes the conditional relationship between picking and placing on deformable objects. The explicit structural encoding enables faster learning under complex object dynamics. Second, instead of jointly learning both the pick and the place locations, we only explicitly learn the placing policy conditioned on random pick points. Then, by selecting the pick point that has Maximal Value under Placing (MVP), we obtain our picking policy. This provides us with an informed picking policy during testing, while using only random pick points during training. Experimentally, this learning framework obtains an order of magnitude faster learning compared to independent action-spaces on our suite of deformable object manipulation tasks with visual RGB observations. Finally, using domain randomization, we transfer our policies to a real PR2 robot for challenging cloth and rope coverage tasks, and demonstrate significant improvements over standard RL techniques on average coverage. | ['Pieter Abbeel', 'Yilin Wu', 'Thanard Kurutach', 'Wilson Yan', 'Lerrel Pinto'] | 2019-10-29 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [ 2.55947381e-01 1.61738291e-01 -4.35463935e-01 -3.43176462e-02
-7.70789802e-01 -9.24942851e-01 3.08751017e-01 1.17438152e-01
-5.81657231e-01 9.29536104e-01 -6.15102500e-02 1.90730747e-02
-3.48510481e-02 -7.01865613e-01 -1.52549982e+00 -8.78616750e-01
-2.96541154e-01 7.82981277e-01 2.19765365e-01 1.18011674e-02
1.53482929e-01 5.66012323e-01 -1.54985464e+00 1.23023599e-01
6.68742716e-01 9.72612381e-01 6.28793240e-01 9.00114477e-01
2.35843286e-01 8.77365649e-01 -3.02884430e-01 3.74017656e-01
5.11647344e-01 -1.27793372e-01 -8.12285841e-01 3.87132496e-01
4.63421255e-01 -9.10082698e-01 -1.50494024e-01 7.28186965e-01
3.41382056e-01 3.62776399e-01 6.77698851e-01 -1.16760325e+00
-7.21168697e-01 7.12417006e-01 -5.73068023e-01 -2.36709565e-01
3.92535478e-01 6.02094173e-01 9.90880251e-01 -4.10402000e-01
8.53857040e-01 1.25881934e+00 2.77372211e-01 8.43435466e-01
-1.40098429e+00 -3.00305009e-01 5.84945261e-01 -1.05392352e-01
-7.15513706e-01 -2.54416585e-01 5.64953148e-01 -4.36098814e-01
7.89182425e-01 7.99623430e-02 9.77917671e-01 1.20086074e+00
-5.41440435e-02 1.33019304e+00 1.15872014e+00 -2.78599203e-01
5.22983015e-01 -3.14754874e-01 -2.55966574e-01 1.07826746e+00
2.70466000e-01 1.44095004e-01 -3.24900001e-01 -2.05164254e-01
1.25202465e+00 1.55508012e-01 -3.04414541e-01 -1.10476351e+00
-1.30406892e+00 7.34880805e-01 6.41954124e-01 -2.24252209e-01
-3.69581878e-01 8.88713062e-01 7.67223239e-02 1.66674882e-01
8.80631804e-03 5.27910948e-01 -6.82668865e-01 3.81627828e-02
-3.73272240e-01 6.86711311e-01 9.01817381e-01 1.11490691e+00
7.88200557e-01 -6.87136278e-02 -3.20813924e-01 5.78698099e-01
3.09519589e-01 8.12190592e-01 -1.27586275e-01 -1.38006711e+00
4.75708008e-01 2.34871224e-01 6.97405040e-01 -4.89605546e-01
-3.08698624e-01 1.47899970e-01 -3.47409159e-01 6.23378873e-01
6.18559778e-01 -2.97337651e-01 -1.35620284e+00 1.76002073e+00
5.57872772e-01 1.87965687e-02 -2.43642610e-02 9.71881986e-01
2.77411431e-01 4.87077415e-01 7.11984485e-02 1.49358567e-02
9.61102962e-01 -1.03839231e+00 -3.58023971e-01 -1.74768642e-01
3.75163943e-01 -1.72051758e-01 1.16087079e+00 4.66015995e-01
-1.20058274e+00 -2.23960459e-01 -9.27844882e-01 -1.52312564e-02
-4.06472757e-02 2.71429658e-01 8.84301484e-01 2.94570867e-02
-9.20703292e-01 9.63145792e-01 -1.28305304e+00 -2.22815230e-01
7.08805382e-01 4.95355397e-01 -2.79156238e-01 -2.61431813e-01
-4.94408578e-01 7.68100500e-01 1.98441312e-01 -1.32313654e-01
-1.69932818e+00 -5.26777983e-01 -8.42575788e-01 -2.26861909e-01
8.61808717e-01 -5.52159727e-01 1.48402381e+00 -6.88118458e-01
-1.80756629e+00 5.68948567e-01 -1.41132483e-02 -4.38935876e-01
6.62728310e-01 -5.79916418e-01 6.63329065e-01 2.45262682e-01
-2.20846273e-02 1.03658688e+00 1.00504994e+00 -1.81352484e+00
-6.26142740e-01 -2.00253755e-01 5.76485097e-01 4.62996542e-01
1.48455441e-01 -6.72344506e-01 -3.91249269e-01 -3.73897582e-01
-2.79947054e-02 -1.29796708e+00 -3.56995374e-01 6.54046535e-01
-3.29438865e-01 -3.08996141e-01 7.53516972e-01 -4.24554527e-01
3.62582535e-01 -1.88929629e+00 7.51029372e-01 1.22469343e-01
2.11828575e-01 -1.43973500e-01 -2.66627938e-01 2.60612220e-01
5.04764676e-01 -1.44790128e-01 -3.23982835e-01 -3.27096075e-01
1.04535088e-01 6.73092902e-01 -2.98660219e-01 5.95092952e-01
2.35159397e-01 1.13739097e+00 -1.23102009e+00 -2.94452161e-01
2.42706925e-01 5.83174936e-02 -9.66178477e-01 2.81450868e-01
-9.75396037e-01 7.81039655e-01 -8.13106298e-01 9.85200644e-01
5.25854945e-01 -3.91225159e-01 5.16899228e-01 1.30694397e-02
-1.13969552e-03 -4.11251746e-02 -1.20015347e+00 1.94917142e+00
-3.21782321e-01 2.29921699e-01 3.78125876e-01 -8.81937861e-01
5.26715636e-01 -4.75853420e-04 6.54992998e-01 -2.23651692e-01
4.64706263e-03 1.06530227e-01 -2.83931255e-01 -6.76822007e-01
2.23864242e-01 1.27770245e-01 7.46590691e-03 3.87074858e-01
9.50177759e-02 -4.04276967e-01 -5.44231609e-02 -4.85228822e-02
1.33481276e+00 9.13993716e-01 6.10278174e-02 -3.60980257e-02
-1.66783065e-01 8.40208232e-02 3.64576846e-01 8.78077030e-01
-1.52013451e-01 3.35043281e-01 5.65981627e-01 -3.30284864e-01
-1.02818871e+00 -1.24897587e+00 2.20245257e-01 1.23868299e+00
4.18815613e-01 1.57408789e-01 -4.69523221e-01 -8.75537455e-01
5.67586184e-01 4.60697979e-01 -8.53365541e-01 1.31679967e-01
-1.06126904e+00 -2.92788506e-01 1.57736480e-01 7.91163623e-01
2.93618292e-01 -1.39773679e+00 -1.11130595e+00 8.97557512e-02
-4.05682921e-02 -8.50462079e-01 -3.88278514e-01 5.83217919e-01
-8.72534871e-01 -1.17592061e+00 -7.61365592e-01 -8.13005269e-01
9.15561318e-01 1.91349953e-01 8.76825571e-01 9.26846489e-02
-4.61156428e-01 8.32078755e-01 -3.93270820e-01 -2.31798396e-01
-1.04911275e-01 -1.09386528e-02 -1.50728654e-02 -5.50287127e-01
-4.82358098e-01 -3.80831271e-01 -7.05249369e-01 -2.08715778e-02
-6.44380391e-01 -1.45395491e-02 6.77333891e-01 8.45458269e-01
7.99272954e-01 -4.01382923e-01 2.05294251e-01 -5.66198349e-01
2.88948476e-01 -3.15361708e-01 -8.95862758e-01 2.86295086e-01
-7.51775550e-03 5.42630732e-01 2.95354813e-01 -8.35100770e-01
-6.43305182e-01 5.48747182e-01 2.97669381e-01 -6.81187034e-01
-1.29098073e-01 -1.08072467e-01 -1.76552758e-02 -2.53040254e-01
4.71649885e-01 -1.15433792e-02 4.19762582e-02 -4.07206774e-01
5.64705491e-01 8.58075079e-03 2.16798097e-01 -1.21969140e+00
6.69802368e-01 6.06019557e-01 8.61716177e-03 -3.93868446e-01
-7.21309900e-01 -1.11873649e-01 -5.75245738e-01 -2.35279724e-01
1.01316440e+00 -6.95505023e-01 -1.29822493e+00 2.86418319e-01
-1.06072199e+00 -1.19383860e+00 -5.88275194e-01 3.02955478e-01
-1.15263879e+00 8.74633938e-02 -4.71285671e-01 -8.92298579e-01
1.21536814e-01 -1.22380221e+00 1.36378920e+00 1.03904687e-01
1.46083802e-01 -7.42005229e-01 5.94211463e-03 5.37586845e-02
8.62300098e-02 6.39612198e-01 7.75265515e-01 -1.82562515e-01
-1.06843090e+00 1.62541986e-01 1.33563560e-02 5.92087582e-02
1.99212223e-01 -2.98469692e-01 -6.09087467e-01 -7.13616729e-01
-4.26708311e-01 -9.13232088e-01 9.71543431e-01 4.91996616e-01
1.46587944e+00 -4.63086098e-01 -7.21980214e-01 5.78408897e-01
1.41476166e+00 1.50684878e-01 2.03888565e-01 3.20048004e-01
7.94930816e-01 3.28209907e-01 9.94263411e-01 6.61508501e-01
3.99150401e-01 5.67196250e-01 9.34647858e-01 6.46169186e-02
-6.26436025e-02 -4.77135241e-01 2.48076171e-01 -2.24083707e-01
-3.66108000e-01 -3.85318249e-01 -6.04206920e-01 5.56887507e-01
-2.00372100e+00 -7.10555732e-01 5.65896213e-01 2.23796248e+00
1.02082527e+00 -1.35598509e-02 2.73071319e-01 -2.41133377e-01
3.62944126e-01 1.32358566e-01 -1.10852134e+00 -5.11320867e-03
2.02443227e-01 1.92610323e-01 1.02688205e+00 7.55802214e-01
-1.18499231e+00 1.03857195e+00 6.18914890e+00 4.26468045e-01
-1.01228702e+00 -1.33589357e-01 2.58874565e-01 -3.63064498e-01
-3.30168098e-01 -6.41411245e-02 -7.53305376e-01 1.86048761e-01
8.75362381e-02 5.30904949e-01 1.07896042e+00 9.87892091e-01
-3.49491052e-02 -2.56610125e-01 -1.39972460e+00 5.95084250e-01
-1.92470685e-01 -1.12407279e+00 -1.27298564e-01 5.29511198e-02
7.17464030e-01 8.46880823e-02 1.38629554e-02 2.29188770e-01
1.11229026e+00 -9.49427962e-01 9.56866860e-01 4.78888959e-01
7.27673590e-01 -5.25318027e-01 1.39098307e-02 4.73841220e-01
-9.99103129e-01 -3.88870776e-01 -3.86066020e-01 4.24054451e-02
7.23511651e-02 1.21910505e-01 -9.79943156e-01 1.69903174e-01
6.71900868e-01 6.72127664e-01 1.91893756e-01 8.21052372e-01
-4.96937513e-01 1.77467495e-01 -4.24062997e-01 -2.66623408e-01
1.97338343e-01 6.03203960e-02 6.16837025e-01 8.53593647e-01
-1.02398787e-02 1.95438951e-01 7.12450981e-01 8.02145898e-01
-1.34433165e-01 -4.91384149e-01 -6.57531857e-01 1.62825212e-02
5.06165683e-01 1.09556377e+00 -7.67594635e-01 -2.44601309e-01
1.52463377e-01 1.14753461e+00 8.51921678e-01 5.34682810e-01
-1.02260673e+00 7.61061460e-02 6.66741848e-01 1.09709963e-01
9.05143440e-01 -6.79356217e-01 8.97524878e-02 -9.57239628e-01
1.54305190e-01 -7.84032583e-01 -7.82115757e-03 -5.24127305e-01
-1.02451539e+00 -1.53938562e-01 2.62999833e-01 -9.54037368e-01
-6.47114441e-02 -7.45942235e-01 -1.77269951e-01 4.19093817e-01
-1.68604565e+00 -1.06005704e+00 -3.05641681e-01 4.59222585e-01
6.69538677e-01 3.82498354e-01 6.13827169e-01 -2.91817129e-01
-6.74901083e-02 2.89338201e-01 -3.50836106e-02 -2.21453216e-02
5.77807188e-01 -1.53124368e+00 1.95931360e-01 1.78301767e-01
-1.23979211e-01 4.21933621e-01 5.54291964e-01 -8.44767511e-01
-1.82260370e+00 -9.57558274e-01 -2.14453712e-01 -5.68331659e-01
4.08068538e-01 -4.84354734e-01 -5.94209671e-01 9.06647444e-01
-4.06427011e-02 3.59841943e-01 -3.23371664e-02 -2.79066741e-01
-2.42706671e-01 1.26431942e-01 -1.48883986e+00 5.98125398e-01
1.29949677e+00 1.52639868e-02 -4.28061903e-01 6.09877706e-01
9.34653342e-01 -8.84434819e-01 -7.44095683e-01 5.15896022e-01
6.79708004e-01 -3.88588846e-01 1.07277739e+00 -6.86753690e-01
4.32485640e-01 -4.40606236e-01 -3.56019378e-01 -1.30289900e+00
-3.17738861e-01 -5.99452376e-01 -5.44397175e-01 7.40900278e-01
2.42486179e-01 -3.85265112e-01 8.51582885e-01 4.03053015e-01
1.05591342e-01 -1.03188252e+00 -7.71541536e-01 -6.95987165e-01
4.88575436e-02 7.15690553e-02 3.46609652e-01 5.95647454e-01
-3.33468676e-01 -2.88931191e-01 -4.63697642e-01 3.16619515e-01
7.56181300e-01 3.29171330e-01 7.75934279e-01 -8.02511334e-01
-8.90010118e-01 -1.09916702e-01 1.03649668e-01 -1.56181145e+00
9.05485898e-02 -6.53332412e-01 7.34467149e-01 -1.75088489e+00
2.92891800e-01 -8.97711337e-01 -4.07459289e-02 1.02045548e+00
-1.11205369e-01 -9.48048607e-02 4.68748719e-01 1.63365856e-01
-8.93411934e-01 4.15439188e-01 1.82782722e+00 -1.78931326e-01
-3.29263359e-01 -1.44252598e-01 -3.10071915e-01 7.31267035e-01
6.96715474e-01 -3.16388786e-01 -3.31345409e-01 -7.63770580e-01
-4.55084480e-02 2.35796019e-01 6.82565093e-01 -7.63758481e-01
-5.47299571e-02 -7.34861910e-01 6.16327524e-01 -3.86353582e-01
4.58023816e-01 -8.46048713e-01 -4.09772038e-01 8.05288494e-01
-5.99804699e-01 -1.61381528e-01 2.31468558e-01 9.66642678e-01
7.40654171e-01 -1.78997666e-01 7.90829420e-01 -4.52144414e-01
-6.09159172e-01 4.09666538e-01 -2.79778153e-01 -4.52519581e-02
1.24217319e+00 1.30239308e-01 -1.95084378e-01 -1.28780141e-01
-9.17260945e-01 5.94523609e-01 8.45246136e-01 3.25798929e-01
6.69679880e-01 -1.09881735e+00 -3.45608056e-01 4.98336889e-02
-1.44061416e-01 5.71833134e-01 -1.27983734e-01 4.92519617e-01
-3.97119910e-01 -6.38654456e-02 -2.61103958e-01 -6.93919718e-01
-1.09763980e+00 6.75017297e-01 2.78300792e-01 -2.73507386e-01
-7.49276578e-01 1.02196717e+00 6.17341734e-02 -5.65097868e-01
5.46616495e-01 -7.54001260e-01 2.43344214e-02 -3.45008671e-01
1.13742083e-01 4.14771616e-01 -5.01815617e-01 2.10645795e-01
-2.33333558e-01 6.65800989e-01 -5.87900206e-02 -2.72111416e-01
1.51159990e+00 1.80161327e-01 4.47480939e-02 4.80586022e-01
7.36998975e-01 1.39776347e-02 -2.30669475e+00 1.27741277e-01
-2.42537022e-01 -5.07728398e-01 -3.25572759e-01 -9.05973613e-01
-8.90512049e-01 5.32437444e-01 3.96396577e-01 -4.93673384e-02
7.13166773e-01 2.85175145e-01 5.11703134e-01 8.65475357e-01
7.69171417e-01 -1.01269543e+00 3.98618072e-01 3.65968496e-01
1.05137432e+00 -1.32430184e+00 9.56452191e-02 -2.05302089e-01
-6.24171555e-01 9.92100477e-01 8.35455000e-01 -6.36603057e-01
2.76942909e-01 6.35763407e-01 -3.38418812e-01 -7.07034171e-02
-7.45922744e-01 -2.77424216e-01 -8.12330917e-02 7.26687074e-01
-5.01800254e-02 2.30312422e-01 1.58417717e-01 -1.68967515e-01
2.38166139e-01 1.06021211e-01 2.52954215e-01 1.55725133e+00
-9.05786753e-01 -1.04468191e+00 -2.15259105e-01 3.04290622e-01
-8.61683562e-02 3.31836194e-01 -2.45434970e-01 8.78506780e-01
5.22372685e-02 5.30964851e-01 -7.38941878e-03 -8.64363611e-02
2.05504164e-01 -2.88037539e-01 1.27699459e+00 -6.72536075e-01
-1.59639776e-01 -5.59227131e-02 -1.34275272e-01 -8.77700031e-01
-4.20507491e-01 -7.73841441e-01 -1.71856987e+00 8.19212291e-03
-1.89670384e-01 -2.06013203e-01 6.60768449e-01 9.79877651e-01
1.02827966e-01 5.88335037e-01 4.77417588e-01 -1.56558669e+00
-9.39310133e-01 -6.14586651e-01 -3.43106806e-01 2.30187505e-01
9.19102550e-01 -9.68230247e-01 -1.98284566e-01 3.42052206e-02] | [4.726578235626221, 0.6230683326721191] |
86c38389-d929-442c-8d6b-decc58235347 | combining-sequence-distillation-and-transfer | null | null | https://aclanthology.org/2020.wmt-1.61 | https://aclanthology.org/2020.wmt-1.61.pdf | Combining Sequence Distillation and Transfer Learning for Efficient Low-Resource Neural Machine Translation Models | In neural machine translation (NMT), sequence distillation (SD) through creation of distilled corpora leads to efficient (compact and fast) models. However, its effectiveness in extremely low-resource (ELR) settings has not been well-studied. On the other hand, transfer learning (TL) by leveraging larger helping corpora greatly improves translation quality in general. This paper investigates a combination of SD and TL for training efficient NMT models for ELR settings, where we utilize TL with helping corpora twice: once for distilling the ELR corpora and then during compact model training. We experimented with two ELR settings: Vietnamese–English and Hindi–English from the Asian Language Treebank dataset with 18k training sentence pairs. Using the compact models with 40% smaller parameters trained on the distilled ELR corpora, greedy search achieved 3.6 BLEU points improvement in average while reducing 40% of decoding time. We also confirmed that using both the distilled ELR and helping corpora in the second round of TL further improves translation quality. Our work highlights the importance of stage-wise application of SD and TL for efficient NMT modeling for ELR settings. | ['Atsushi Fujita', 'Raj Dabre'] | null | null | null | null | wmt-emnlp-2020-11 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 2.90713400e-01 5.08700078e-03 -3.76106948e-01 -3.48245770e-01
-1.63549638e+00 -7.47948289e-01 4.46855009e-01 -3.28978747e-01
-9.55110848e-01 1.22406662e+00 2.37270162e-01 -1.13488638e+00
4.22375590e-01 -2.69322693e-01 -8.28993022e-01 -3.86164576e-01
4.01218295e-01 9.46429610e-01 -4.07185495e-01 -4.97302353e-01
-4.40678522e-02 1.52283698e-01 -5.89124382e-01 4.24116164e-01
1.27379751e+00 2.38887146e-01 6.95070028e-01 7.32110023e-01
-1.57206461e-01 5.72918057e-01 -7.17728794e-01 -7.69181669e-01
5.91845989e-01 -6.92261755e-01 -7.22421527e-01 -3.13246667e-01
3.39743108e-01 -4.05702025e-01 -1.08904190e-01 6.62449181e-01
9.50006664e-01 6.74667433e-02 3.91426057e-01 -6.35737717e-01
-8.71739328e-01 1.47815633e+00 -6.51528239e-01 3.03059965e-01
-1.65520757e-02 1.33911386e-01 1.17314029e+00 -1.33877182e+00
7.87980795e-01 1.15986729e+00 5.84660649e-01 8.78732264e-01
-1.28606594e+00 -8.93270671e-01 -3.05249244e-01 1.54643044e-01
-1.17839217e+00 -8.05433571e-01 3.08575153e-01 1.54401988e-01
1.68970346e+00 3.06227475e-01 3.65432769e-01 1.13973367e+00
1.24432653e-01 1.04404771e+00 1.17957044e+00 -9.17260170e-01
-2.86737047e-02 3.04885298e-01 -1.24862209e-01 4.33228910e-01
7.34870583e-02 -1.12526439e-01 -6.35738552e-01 -5.27585782e-02
4.50577497e-01 -6.57048821e-01 -1.88373506e-01 4.62325782e-01
-1.41177702e+00 8.84718299e-01 -9.52615812e-02 5.07932186e-01
-3.00211012e-01 5.07202186e-03 6.32577538e-01 8.61955583e-01
7.39664555e-01 8.52031469e-01 -8.78933549e-01 -7.23571241e-01
-1.17723894e+00 -1.46655858e-01 7.48310864e-01 1.30868244e+00
6.22523844e-01 2.30167851e-01 -9.38975811e-02 1.44006300e+00
-2.03230038e-01 7.86358356e-01 7.47055054e-01 -6.98059678e-01
1.32409549e+00 1.50080353e-01 -1.96656272e-01 -9.80166271e-02
1.19563397e-02 -5.43832600e-01 -7.17662692e-01 -4.45650369e-01
4.46037054e-01 -5.43059766e-01 -9.60594654e-01 1.87244439e+00
-2.09056586e-02 -3.37863207e-01 5.15435338e-01 6.30106568e-01
4.28782701e-01 1.15904176e+00 -2.21043631e-01 -4.26261008e-01
1.04419780e+00 -1.21786880e+00 -7.79019892e-01 -4.18227136e-01
1.36650181e+00 -1.13212872e+00 1.38381374e+00 1.18114926e-01
-1.48230338e+00 -3.83490801e-01 -8.96312535e-01 -4.85932350e-01
-7.57175162e-02 6.38202488e-01 4.02959764e-01 5.32466710e-01
-1.33176208e+00 5.60256839e-01 -8.78914177e-01 -4.96512622e-01
1.07496262e-01 4.82323259e-01 -3.89709175e-01 -5.02066553e-01
-1.29691136e+00 1.34126580e+00 2.55599827e-01 2.29619101e-01
-5.69694877e-01 -7.00760901e-01 -6.97907150e-01 -1.54961467e-01
1.49281666e-01 -5.71592450e-01 1.41816425e+00 -9.20097709e-01
-1.88727188e+00 8.17266285e-01 -2.74144828e-01 -6.24277472e-01
7.11119831e-01 -5.11871994e-01 -3.96523252e-02 -2.95226097e-01
-6.02332354e-02 6.68948948e-01 3.58213395e-01 -6.95990205e-01
-4.09600884e-01 -1.73446253e-01 -3.17961514e-01 5.29518068e-01
-4.66119230e-01 5.12061596e-01 -4.35220420e-01 -7.41776407e-01
-1.21263437e-01 -1.21256721e+00 -2.11711824e-01 -7.74438977e-01
-2.56163985e-01 -9.99501571e-02 3.48454118e-01 -1.39902759e+00
1.43404794e+00 -1.67962801e+00 4.53916371e-01 -9.16648507e-02
-2.42586434e-01 5.44739902e-01 -6.11021042e-01 6.63297534e-01
1.71247542e-01 2.55993485e-01 -2.71456808e-01 -6.92821980e-01
-1.04950450e-01 4.46952045e-01 -2.20621869e-01 1.27618909e-01
1.76501423e-01 1.33137751e+00 -6.87344909e-01 -5.20794868e-01
-4.88798656e-02 2.75750071e-01 -5.05840302e-01 2.55908817e-01
-9.83931869e-02 4.59703416e-01 1.73082352e-02 6.15012527e-01
3.91532034e-01 1.52791455e-01 4.22500789e-01 1.61636263e-01
-1.96737811e-01 8.78258109e-01 -4.11655933e-01 1.96058846e+00
-9.06554937e-01 8.48877609e-01 -1.81702241e-01 -6.82193935e-01
9.80384886e-01 4.81444359e-01 1.24006532e-01 -9.65130329e-01
8.88755843e-02 9.14441407e-01 4.46806848e-01 -2.14535564e-01
8.38269711e-01 -1.44270658e-01 -4.55788597e-02 7.87591517e-01
1.28911838e-01 -2.19500631e-01 2.97532350e-01 1.93842292e-01
1.11445916e+00 2.85481185e-01 1.83373258e-01 -1.52767092e-01
2.26886451e-01 2.10562080e-01 6.91077232e-01 5.65198243e-01
3.96876447e-02 4.55139220e-01 1.25338379e-02 -8.60015154e-02
-1.66585302e+00 -6.64782405e-01 1.31949544e-01 1.26372039e+00
-6.25432014e-01 -5.51639080e-01 -9.35378015e-01 -6.07481301e-01
-3.37443888e-01 1.15204144e+00 -2.82516420e-01 -1.41663700e-01
-1.56165707e+00 -1.00836444e+00 9.80963647e-01 4.44818884e-01
4.26791847e-01 -9.00022089e-01 -6.18974902e-02 4.58716631e-01
-6.70920014e-01 -1.19690144e+00 -7.77413189e-01 3.69216532e-01
-1.09609652e+00 -2.41833866e-01 -1.07689595e+00 -8.45293343e-01
5.27336359e-01 1.65980712e-01 1.24175382e+00 -4.86697927e-02
2.44899303e-01 -3.66744727e-01 -5.17670512e-01 -1.92753524e-01
-1.05756664e+00 7.06459701e-01 1.34806335e-01 -6.23419881e-01
3.90252173e-01 -4.65180010e-01 -6.45821029e-03 2.99284220e-01
-5.05162060e-01 4.97041136e-01 1.04761124e+00 1.04946303e+00
4.71759379e-01 -7.48249829e-01 6.24122322e-01 -9.76050377e-01
7.77716815e-01 -3.31324697e-01 -3.29539508e-01 6.40410483e-01
-6.36121869e-01 2.73020864e-01 8.44278216e-01 -6.91164672e-01
-1.07698762e+00 -4.72809523e-01 -4.02450949e-01 -2.63445944e-01
5.04859626e-01 4.65086937e-01 -1.54018745e-01 2.70010233e-01
6.78810418e-01 4.35195595e-01 -1.04159862e-01 -7.82456875e-01
4.87651855e-01 1.10413182e+00 2.09419042e-01 -8.71542454e-01
6.98027432e-01 -3.93187255e-01 -3.54095697e-01 -6.06011808e-01
-3.06909174e-01 -1.68596148e-01 -7.09446549e-01 2.41401449e-01
5.91703594e-01 -9.66089785e-01 1.33062899e-01 3.40107381e-01
-1.22585285e+00 -8.38656366e-01 -1.68789029e-01 8.39123368e-01
-5.81924319e-01 2.16056436e-01 -1.14436948e+00 -5.56453943e-01
-9.82390821e-01 -1.41560376e+00 9.60668623e-01 -4.12036359e-01
-2.65152335e-01 -8.55369687e-01 1.51121199e-01 6.62502408e-01
5.46916783e-01 -4.30651903e-01 1.13179564e+00 -7.09524155e-01
-2.76497155e-01 4.93899360e-02 -8.90084952e-02 6.05449378e-01
1.50711566e-01 -2.09442332e-01 -5.71472168e-01 -5.54845095e-01
-1.83469042e-01 -5.14924943e-01 5.79379261e-01 1.55249044e-01
3.41068715e-01 -3.90345097e-01 6.13632388e-02 5.14983416e-01
1.15884018e+00 2.69285560e-01 5.47221661e-01 4.41487640e-01
6.31520391e-01 2.70566344e-01 7.85286248e-01 6.50062710e-02
3.59711587e-01 8.76493573e-01 -3.64544511e-01 -2.94094592e-01
-3.49831343e-01 -3.12707335e-01 9.27350640e-01 1.79891610e+00
-3.08984280e-01 -2.62373865e-01 -9.69874084e-01 4.51327741e-01
-1.63583994e+00 -5.83237052e-01 8.83753002e-02 2.18129230e+00
1.59192002e+00 -9.08827484e-02 -1.73091870e-02 -2.98181593e-01
6.47780716e-01 -3.14300716e-01 -4.44710165e-01 -8.51937175e-01
-3.75178277e-01 4.79666024e-01 7.32019782e-01 6.09166563e-01
-3.05792242e-01 1.51173532e+00 5.71418810e+00 1.09555793e+00
-1.06272745e+00 5.20695627e-01 6.07059419e-01 -4.18885499e-01
-4.79068041e-01 4.01676558e-02 -1.15886831e+00 3.84530932e-01
1.56766617e+00 -3.18478495e-01 7.27788925e-01 5.48663974e-01
3.30981523e-01 2.19748303e-01 -1.17623639e+00 8.24624062e-01
-1.25440896e-01 -1.23247075e+00 1.20910630e-01 1.99540943e-01
8.44093502e-01 4.77212697e-01 -7.37232938e-02 7.30528891e-01
5.64506829e-01 -9.43928301e-01 8.27015221e-01 -2.08868891e-01
1.17886424e+00 -7.97591686e-01 8.02436233e-01 6.36940360e-01
-8.11652839e-01 2.61058718e-01 -7.73071349e-01 1.21741816e-01
2.79589027e-01 3.43195349e-01 -1.38327610e+00 7.25986421e-01
2.82195091e-01 6.07470095e-01 -3.95903617e-01 4.49165016e-01
-2.81337768e-01 1.24352217e+00 -3.75515670e-01 -8.78298059e-02
4.73719776e-01 -3.76975864e-01 5.14450312e-01 1.60718179e+00
5.01375794e-01 -2.44344756e-01 -2.39973232e-01 4.95288908e-01
-4.54531699e-01 5.95913053e-01 -4.98949587e-01 -1.86157316e-01
7.34862566e-01 9.65516329e-01 -2.90299445e-01 -4.67140168e-01
-3.70015889e-01 1.15197170e+00 7.68623054e-01 2.74277985e-01
-7.70089328e-01 -1.96465209e-01 6.12187207e-01 -1.41496867e-01
1.16787599e-02 -5.48925400e-01 -3.33909869e-01 -1.36928141e+00
6.79036975e-02 -1.38715518e+00 1.79803327e-01 -5.64374626e-01
-8.80757034e-01 9.63997960e-01 -8.03053603e-02 -1.02747548e+00
-6.10124886e-01 -3.91706914e-01 -2.01486707e-01 1.28325701e+00
-1.52199566e+00 -1.23936963e+00 5.36489367e-01 3.31731647e-01
1.16784811e+00 -2.85928547e-01 8.92558873e-01 6.17140412e-01
-8.35533738e-01 1.15823829e+00 6.33191228e-01 1.69683024e-01
9.51541305e-01 -1.16167009e+00 1.04179943e+00 9.03178751e-01
3.73418659e-01 9.33002353e-01 5.03721297e-01 -6.10382438e-01
-1.67588389e+00 -1.14904106e+00 1.47810090e+00 -4.13585007e-01
5.50505400e-01 -6.54436111e-01 -7.72951424e-01 9.17124927e-01
5.47608554e-01 -5.89727640e-01 7.65692234e-01 1.64142951e-01
-2.40979105e-01 1.52129158e-01 -8.99008870e-01 8.96827877e-01
1.10555232e+00 -4.68092352e-01 -5.83533049e-01 3.72944772e-01
9.87825036e-01 -5.06640911e-01 -9.73411441e-01 2.91821897e-01
4.86195207e-01 -2.21422598e-01 6.60466731e-01 -7.20440269e-01
4.90875602e-01 2.09503010e-01 -4.62687612e-01 -1.78923512e+00
-1.10377319e-01 -9.27961409e-01 2.82566212e-02 1.29425061e+00
1.06876123e+00 -6.09469593e-01 4.46672529e-01 4.22237635e-01
-4.46914434e-01 -9.55934882e-01 -9.75168288e-01 -1.07634282e+00
7.13391364e-01 -3.59776556e-01 4.85813946e-01 9.26818252e-01
7.54128322e-02 7.01338768e-01 -6.10254705e-01 -4.80949819e-01
3.05314749e-01 -1.15864925e-01 8.23723614e-01 -4.98077273e-01
-5.45015097e-01 -1.63064465e-01 3.07869732e-01 -1.22317433e+00
2.24055886e-01 -1.33880949e+00 1.05548255e-01 -1.43901110e+00
3.41219068e-01 -7.40496218e-01 4.13123034e-02 6.67734325e-01
-4.33572888e-01 3.24999064e-01 4.14932489e-01 4.69500601e-01
-7.89867267e-02 4.75732952e-01 1.27383518e+00 1.54376090e-01
-2.57623732e-01 -2.19716340e-01 -5.20484507e-01 5.29128499e-02
1.00353122e+00 -5.67429900e-01 -2.23845765e-01 -1.18192065e+00
9.66946557e-02 1.35591015e-01 -5.62527597e-01 -5.03538728e-01
-2.42631063e-02 -4.47102860e-02 -1.49525786e-02 -4.86767679e-01
3.49241763e-01 -4.05477673e-01 6.96603507e-02 2.94566691e-01
-5.59535563e-01 6.10441267e-01 4.04541194e-01 -1.46814585e-01
9.25708264e-02 -2.48928875e-01 7.18208313e-01 -2.82243222e-01
-2.46372506e-01 -3.56862657e-02 -3.69769186e-01 3.18590730e-01
4.83664185e-01 -8.55864286e-02 -1.44742474e-01 -2.41154075e-01
-2.72148073e-01 8.38076919e-02 3.24861586e-01 4.11740869e-01
3.50706339e-01 -1.11649919e+00 -1.41191745e+00 1.12882055e-01
-1.39211312e-01 -2.45971680e-01 -2.96364009e-01 1.02336192e+00
-4.60346699e-01 6.72114551e-01 -1.38181567e-01 -2.92374462e-01
-1.49167275e+00 1.05733410e-01 5.42215332e-02 -7.64134705e-01
-5.45050442e-01 1.01234353e+00 -2.47731894e-01 -8.12101722e-01
2.38537081e-02 -1.96641266e-01 3.65997195e-01 -2.24661857e-01
2.79198259e-01 5.14520407e-01 4.03038889e-01 -6.57247245e-01
-9.62989405e-02 1.81806967e-01 -6.14209235e-01 -7.35425055e-01
1.34987271e+00 -2.57750690e-01 -8.97347406e-02 3.32382202e-01
1.28917325e+00 2.73507386e-01 -7.72858560e-01 -5.70936561e-01
2.01465592e-01 -3.39115560e-01 -3.73599678e-02 -9.97824192e-01
-7.45222628e-01 1.06193626e+00 1.58233009e-02 -6.57638073e-01
1.03970242e+00 -3.31654161e-01 1.36793458e+00 6.18363082e-01
5.95770061e-01 -1.14803684e+00 -2.39838347e-01 9.00457561e-01
7.42523491e-01 -1.11125600e+00 -3.61583114e-01 -2.74699274e-02
-7.22708702e-01 1.04405248e+00 4.57437634e-01 2.49041453e-01
-7.97030404e-02 5.40871024e-01 3.37191850e-01 5.38086653e-01
-1.05379272e+00 7.26822838e-02 5.85436076e-02 2.27853909e-01
8.01264226e-01 1.66093230e-01 -7.16256380e-01 3.54029417e-01
-6.17823601e-01 -1.43206835e-01 4.22020376e-01 9.76804256e-01
-2.74332583e-01 -1.65878332e+00 -2.15957388e-01 4.08276647e-01
-7.84049690e-01 -9.06366825e-01 -3.89322639e-01 6.79845393e-01
-3.39529932e-01 9.64763105e-01 -1.88508689e-01 -5.11415899e-01
1.02027781e-01 3.59851897e-01 6.61323667e-01 -7.21666753e-01
-9.26523387e-01 1.31554380e-01 5.56610346e-01 -2.45126605e-01
3.50490846e-02 -6.42405033e-01 -1.06491065e+00 -5.83048999e-01
-5.83475530e-01 5.02523959e-01 9.33941126e-01 9.06467080e-01
3.22139412e-01 2.08740965e-01 6.67263448e-01 -5.79534352e-01
-1.02138507e+00 -1.57646728e+00 -9.99127701e-02 -3.02090645e-02
-3.04776765e-02 -9.33976322e-02 -1.50596976e-01 3.99409346e-02] | [11.639324188232422, 10.283758163452148] |
ca3fea96-f7d2-4508-986e-7ca0e542dc5a | uzbek-affix-finite-state-machine-for-stemming | 2205.10078 | null | https://arxiv.org/abs/2205.10078v1 | https://arxiv.org/pdf/2205.10078v1.pdf | Uzbek affix finite state machine for stemming | This work presents a morphological analyzer for the Uzbek language using a finite state machine. The proposed methodology is a morphologic analysis of Uzbek words by using an affix striping to find a root and without including any lexicon. This method helps to perform morphological analysis of words from a large amount of text at high speed as well as it is not required using of memory for keeping vocabulary. According to Uzbek, an agglutinative language can be designed with finite state machines (FSMs). In contrast to the previous works, this study modeled the completed FSMs for all word classes by using the Uzbek language's morphotactic rules in right to left order. This paper shows the stages of this methodology including the classification of the affixes, the generation of the FSMs for each affix class, and the combination into a head machine to make analysis a word. | ['Ulugbek Salaev', 'Maksud Sharipov'] | 2022-05-20 | null | null | null | null | ['morphological-analysis'] | ['natural-language-processing'] | [-2.29840167e-02 -2.95413714e-02 8.50833654e-02 -1.41619250e-01
8.38246942e-02 -7.93912292e-01 6.59356534e-01 7.35391796e-01
-8.34146738e-01 7.96074092e-01 -1.62121162e-01 -8.98822069e-01
9.09583718e-02 -1.18085134e+00 -1.89809263e-01 -5.48821807e-01
2.53184229e-01 7.23809421e-01 4.97779489e-01 -5.56642711e-01
4.95662123e-01 9.70967948e-01 -1.57829082e+00 1.00666493e-01
8.50713313e-01 1.15067281e-01 4.37231600e-01 7.20343828e-01
-3.54719758e-01 6.01143897e-01 -7.97911763e-01 -2.25776777e-01
1.85054138e-01 -6.25757575e-01 -9.55680430e-01 -9.20503363e-02
3.22517782e-01 -9.72780138e-02 2.00233698e-01 1.34966898e+00
9.46136191e-02 5.05736768e-01 7.36285925e-01 -3.28129530e-01
-2.96910793e-01 9.45611477e-01 -1.20214624e-02 3.13601851e-01
2.18583927e-01 -1.00540169e-01 8.31683576e-01 -6.57410920e-01
5.98756790e-01 1.25450242e+00 1.69268236e-01 3.58191580e-01
-8.18381131e-01 -4.51657981e-01 -1.12735599e-01 2.90207386e-01
-1.30220079e+00 -4.52170551e-01 4.59125906e-01 -4.36744392e-01
1.45447159e+00 1.93927944e-01 8.92101645e-01 2.81333793e-02
4.07384068e-01 1.30013332e-01 1.53900063e+00 -1.24590290e+00
4.78211224e-01 -1.08599290e-02 9.83093083e-01 8.57530415e-01
8.19099486e-01 -1.51036695e-01 -2.20244497e-01 -2.77224742e-02
4.52513844e-01 -4.72293139e-01 4.29998249e-01 3.41650635e-01
-6.71469629e-01 6.35786593e-01 -2.54307598e-01 9.43011999e-01
-4.16425854e-01 -4.31147158e-01 2.99207956e-01 1.13983497e-01
-2.28232220e-02 3.49239379e-01 -2.90696055e-01 1.73839539e-01
-9.60877240e-01 1.09568469e-01 8.92287433e-01 4.91842628e-01
6.63423777e-01 3.88263345e-01 3.83328080e-01 4.78413284e-01
3.32426876e-01 9.30755258e-01 7.18527615e-01 -1.78776845e-01
1.69774532e-01 8.45366240e-01 -5.19446433e-02 -4.98223454e-01
-3.59763145e-01 -8.74224529e-02 -2.51127303e-01 6.13169134e-01
6.54465556e-01 -2.17154130e-01 -1.17015731e+00 1.62099361e+00
5.35185039e-01 -1.35745540e-01 4.86590981e-01 3.48867089e-01
5.14605999e-01 9.63004529e-01 2.00956449e-01 -5.41959405e-01
1.79252195e+00 -4.38881606e-01 -1.04861057e+00 -7.88105931e-03
6.17253363e-01 -1.04240894e+00 7.94175565e-01 5.67076504e-01
-1.11123240e+00 -4.12407577e-01 -1.49056149e+00 8.81798789e-02
-7.91750193e-01 1.46814555e-01 3.19521129e-01 9.93749022e-01
-1.05300915e+00 4.66693699e-01 -1.15190959e+00 -6.18528545e-01
-6.34682715e-01 6.49361014e-01 -3.76731962e-01 6.78662479e-01
-1.05521619e+00 1.31108081e+00 9.38811481e-01 2.89459646e-01
-3.74574840e-01 3.20952266e-01 -9.35095727e-01 -9.41564292e-02
-4.35154513e-02 -4.41920310e-02 9.32892680e-01 -9.15543199e-01
-1.66538334e+00 1.05017710e+00 -2.54032165e-01 -6.01733148e-01
-4.70753163e-02 3.54876146e-02 -6.67036653e-01 5.23390919e-02
-9.80763808e-02 -5.09176254e-02 5.20121992e-01 -8.04384053e-01
-8.46023023e-01 -6.25802457e-01 -1.32371247e-01 -1.50144957e-02
-1.73275530e-01 4.36664373e-01 5.67478165e-02 -4.34174448e-01
2.71542549e-01 -8.59541237e-01 -1.16305597e-01 -1.31475413e+00
-1.19515374e-01 -3.33244979e-01 2.44237021e-01 -9.68789160e-01
1.52795422e+00 -1.93605661e+00 -1.32952079e-01 4.52201128e-01
-2.52935261e-01 6.81410134e-01 1.28750443e-01 4.63968337e-01
2.66851019e-02 -2.36574575e-01 -2.40331158e-01 -3.88437975e-03
-1.85860455e-01 3.70106041e-01 -1.12368613e-01 5.88978648e-01
9.55452994e-02 3.77789259e-01 -7.37794578e-01 -6.87419057e-01
5.00204861e-01 9.08574834e-02 -3.36642921e-01 -4.36550267e-02
-2.64322460e-02 -2.77467489e-01 -1.97669059e-01 4.98915076e-01
8.09637368e-01 6.59074187e-01 9.08853590e-01 1.37377650e-01
-7.87394226e-01 4.55521405e-01 -1.47001171e+00 1.00385582e+00
-4.68026161e-01 1.25825673e-01 7.08064511e-02 -8.17471981e-01
1.31310749e+00 2.20297739e-01 -1.47676557e-01 -5.07013917e-01
4.84523386e-01 7.37597287e-01 2.90857106e-01 -1.43458918e-01
9.04439867e-01 -6.39534235e-01 -2.42078453e-01 3.45299870e-01
3.38115275e-01 1.06389457e-02 7.48878062e-01 -7.02721998e-02
6.11125827e-01 1.24042481e-01 9.85672712e-01 -8.46392691e-01
9.47412670e-01 1.86506078e-01 6.70797050e-01 3.87334287e-01
1.16279483e-01 -1.05176516e-01 6.18020818e-02 -4.25377369e-01
-6.96436167e-01 -1.13650000e+00 -2.83499181e-01 8.14310670e-01
4.02210094e-02 -1.37597144e-01 -1.15945935e+00 -2.83013850e-01
-3.70907724e-01 9.32313144e-01 -4.12280858e-01 1.28927007e-01
-1.13984358e+00 -8.28792095e-01 7.37868249e-01 8.20276067e-02
2.36884430e-01 -1.62350202e+00 -1.05558586e+00 3.97498131e-01
2.12917134e-01 -8.26618910e-01 1.66562095e-01 4.56910521e-01
-1.08016324e+00 -9.24992263e-01 1.50254518e-01 -1.35693634e+00
6.28816366e-01 -2.78150082e-01 5.66957533e-01 1.98322594e-01
1.10985577e-01 -5.81048787e-01 -3.84659111e-01 -5.71533918e-01
-9.05513465e-01 2.55125165e-01 3.96969318e-01 -3.71995792e-02
7.39169061e-01 -4.09531415e-01 1.44360557e-01 -3.39793295e-01
-1.06808269e+00 -2.16225877e-01 4.86415207e-01 2.72770494e-01
4.67150778e-01 5.26760444e-02 2.82157391e-01 -1.04416311e+00
4.94787365e-01 -1.50626063e-01 -9.75796878e-01 3.55256766e-01
-5.80060065e-01 3.85118723e-01 9.94662166e-01 -1.57230496e-01
-1.10445499e+00 -2.48395256e-03 -4.24182624e-01 5.37231326e-01
-4.44357276e-01 3.41059893e-01 -3.96217704e-01 3.06839347e-01
5.12673855e-01 3.01057488e-01 1.09262295e-01 -6.87204838e-01
1.22292854e-01 7.59523869e-01 7.09077299e-01 -5.15087783e-01
5.74792683e-01 3.37407410e-01 3.00571024e-01 -1.06729972e+00
-2.41313785e-01 -3.35589647e-01 -1.02070403e+00 -6.63669407e-02
9.36279833e-01 -3.53306711e-01 -4.14068580e-01 7.22142935e-01
-1.23492694e+00 -3.74578387e-02 -3.18357229e-01 7.34842598e-01
-8.94907489e-02 5.60524821e-01 -9.92419720e-01 -9.85197783e-01
-5.71335793e-01 -9.20841515e-01 2.33670443e-01 3.27495009e-01
-4.41393465e-01 -8.71150732e-01 5.54406762e-01 -1.61906078e-01
-2.46139631e-01 -5.97002730e-02 1.33611429e+00 -1.10543168e+00
1.12637401e-01 -8.47357810e-02 4.58604962e-01 4.05406058e-01
2.75773734e-01 3.35919827e-01 -5.20519555e-01 -1.14929095e-01
3.15359265e-01 1.73206046e-01 4.39914227e-01 2.06915438e-01
-4.13267404e-01 -2.92125672e-01 1.92501366e-01 2.91246861e-01
1.65025377e+00 9.25932527e-01 4.84693795e-01 7.23189890e-01
2.66959339e-01 5.77722788e-01 7.68595755e-01 1.59396052e-01
1.82885870e-01 1.87624857e-01 5.91217913e-02 4.34007309e-02
-1.13590881e-01 -1.47877291e-01 7.45360315e-01 1.55374384e+00
9.83902602e-04 -6.86058179e-02 -9.53419566e-01 7.15619147e-01
-1.41622293e+00 -7.98144042e-01 -3.98580909e-01 2.39279151e+00
6.99150503e-01 1.60518646e-01 4.03098296e-03 6.04257762e-01
1.01003015e+00 1.29236430e-01 3.11682850e-01 -1.15524173e+00
-3.62091422e-01 1.09841168e+00 4.46644515e-01 1.39678729e+00
-6.80368125e-01 1.57064617e+00 5.98258162e+00 7.52366543e-01
-1.36767316e+00 1.38222829e-01 -6.76695779e-02 1.01757392e-01
-1.37471408e-01 5.04083455e-01 -1.09869981e+00 2.91201949e-01
1.09777081e+00 1.43103570e-01 2.92938411e-01 1.55704975e-01
3.00473511e-01 -4.83533055e-01 -4.03203219e-01 5.94067276e-01
1.21298395e-01 -9.08885896e-01 5.57126999e-01 3.55023611e-03
3.70386064e-01 -1.47112086e-01 -6.70738935e-01 5.10991402e-02
3.12990546e-01 -6.65304720e-01 1.13085186e+00 2.85455942e-01
3.64406556e-01 -9.31010306e-01 7.83937752e-01 2.79796571e-01
-1.20641148e+00 1.07824519e-01 -5.50793052e-01 -5.62445402e-01
2.75608838e-01 2.91923046e-01 -7.43801177e-01 5.05672991e-01
-9.35090333e-02 -2.47831926e-01 -5.12316883e-01 5.97846389e-01
-4.14621413e-01 9.46831167e-01 -4.29695994e-01 -5.65185070e-01
3.65556657e-01 -8.74987543e-01 5.47769010e-01 1.37875044e+00
3.33058268e-01 2.59864539e-01 1.01713009e-01 5.56698069e-02
5.34887552e-01 1.03874743e+00 -3.06044906e-01 9.03521255e-02
5.59350252e-01 1.00168788e+00 -1.41547680e+00 -7.53859162e-01
-4.65984866e-02 5.60577154e-01 4.26818162e-01 -1.11647807e-01
-2.02822998e-01 -6.62150383e-01 3.08827728e-01 3.43185693e-01
1.28189683e-01 -5.59281588e-01 -4.22980160e-01 -8.98249090e-01
-1.88830882e-01 -9.84984338e-01 6.75858855e-01 -2.14157879e-01
-3.17414910e-01 9.75922048e-01 2.58035343e-02 -6.35239065e-01
-4.19516712e-01 -9.53313410e-01 -3.44609052e-01 9.69042480e-01
-9.52739418e-01 -1.08016706e+00 4.39716130e-01 3.55039924e-01
3.94957393e-01 -2.83521682e-01 1.00499785e+00 2.58678477e-02
-4.98181850e-01 1.09919347e-01 -3.36543955e-02 2.85817534e-01
3.93054485e-01 -1.21713734e+00 3.14629018e-01 1.64638114e+00
4.05669391e-01 1.03312409e+00 8.60516310e-01 -1.13064563e+00
-1.06726336e+00 -3.52799028e-01 1.65234065e+00 -1.11957826e-01
6.53873980e-01 -2.33525902e-01 -6.57528996e-01 5.54369152e-01
5.19887507e-01 -4.54577357e-01 6.17544651e-01 9.92189050e-02
4.75305282e-02 -1.40013427e-01 -9.80759025e-01 6.74988449e-01
4.79647487e-01 -3.97772282e-01 -1.23488283e+00 -1.87303603e-01
1.93234265e-01 -4.57071979e-03 -5.05350590e-01 2.41473559e-02
6.08667552e-01 -8.07828426e-01 1.58580303e-01 -6.85617268e-01
-1.42263606e-01 -6.71711922e-01 -4.14787568e-02 -1.06611657e+00
-4.03308481e-01 -7.01528430e-01 2.82566041e-01 1.22126579e+00
4.47573543e-01 -7.06076324e-01 5.05119503e-01 8.46573487e-02
-1.56715497e-01 -8.41732696e-02 -8.29463065e-01 -6.91124380e-01
2.29118690e-01 -1.80103391e-01 7.53938138e-01 9.47977722e-01
3.68465066e-01 4.90882188e-01 9.71133355e-03 2.14654520e-01
4.69002038e-01 1.53797090e-01 2.13122338e-01 -1.35659468e+00
-3.69887590e-01 -4.21134949e-01 -6.59672499e-01 -3.22509289e-01
3.89342964e-01 -1.02492356e+00 7.12538809e-02 -1.29502547e+00
-3.20432305e-01 -3.10739785e-01 -4.74932641e-01 4.22620207e-01
-7.12334737e-02 5.82930818e-02 1.63330629e-01 -3.77883650e-02
7.05226958e-02 -3.96465734e-02 6.27591431e-01 1.03458062e-01
-5.11554182e-01 -2.01485798e-01 -2.13445038e-01 8.02030861e-01
8.19621205e-01 -6.16008222e-01 -2.93721408e-01 -3.39881927e-01
1.79277599e-01 6.46821572e-04 -3.71952623e-01 -1.09806502e+00
1.08929932e-01 -4.06338632e-01 -7.21768290e-02 -6.58950090e-01
-6.83323294e-02 -4.74580586e-01 2.69348532e-01 9.97512817e-01
2.12155893e-01 6.97826326e-01 2.56874263e-01 -2.24549890e-01
-2.14005589e-01 -9.91199613e-01 1.08664370e+00 -1.84913546e-01
-8.15863132e-01 -3.54813814e-01 -1.13332939e+00 -3.09704155e-01
1.07811689e+00 -3.93819839e-01 1.91460755e-02 4.13896382e-01
-7.02819407e-01 -4.14261281e-01 5.73833227e-01 -1.51236340e-01
2.51531035e-01 -7.25820661e-01 -5.11968791e-01 4.19572681e-01
-2.81306624e-01 -4.74663734e-01 -2.30625775e-02 3.35210621e-01
-1.45993030e+00 4.51802105e-01 -7.88287044e-01 1.63644165e-01
-1.58128643e+00 5.36321998e-01 1.69461086e-01 -4.87951726e-01
-3.43576998e-01 3.19612861e-01 -1.94980577e-01 -4.06546772e-01
-4.84400988e-01 -5.35569370e-01 -8.04149151e-01 3.90794992e-01
4.75026637e-01 4.33028579e-01 5.37687600e-01 -1.06159163e+00
-5.94721258e-01 4.50510770e-01 6.43843934e-02 -8.01393390e-01
1.02150798e+00 1.50885195e-01 -7.82340050e-01 5.68252146e-01
5.33638239e-01 9.78346229e-01 -1.85166195e-01 1.57199264e-01
2.75555104e-01 7.63590559e-02 -2.24616662e-01 -2.95151442e-01
-4.01821226e-01 6.17684901e-01 3.69100899e-01 -1.34573448e-02
8.93237829e-01 -5.29757857e-01 5.67284226e-01 6.27807617e-01
4.91287142e-01 -1.39826679e+00 -8.98373008e-01 9.80055928e-01
2.60528438e-02 -2.58057654e-01 -1.22162044e-01 -3.21821004e-01
-3.37234527e-01 1.22918820e+00 3.18996400e-01 -3.90491575e-01
8.34870100e-01 6.73038602e-01 4.30061072e-01 -1.07243955e-01
-4.99118239e-01 -4.69738573e-01 -1.01851560e-01 2.99677700e-01
4.77258414e-01 5.69503903e-01 -1.57474685e+00 6.45198226e-01
-6.86841905e-01 -4.04391527e-01 8.41661870e-01 1.25701082e+00
-9.99578059e-01 -1.62819529e+00 -7.33813345e-01 3.17849427e-01
-8.46395791e-01 -3.71014982e-01 -4.32001621e-01 7.73017287e-01
4.61304218e-01 1.04430270e+00 2.62742758e-01 -3.53114456e-01
2.34491363e-01 3.78877908e-01 7.36783683e-01 -8.15403283e-01
-1.11727309e+00 -1.87145546e-02 3.30752373e-01 2.18467325e-01
-9.93883684e-02 -9.42642927e-01 -1.71330070e+00 -1.90849543e-01
-3.02121073e-01 8.58501494e-01 6.62317514e-01 1.25717270e+00
-3.99837017e-01 -8.26017931e-02 2.59289831e-01 -1.45851821e-01
-4.77720529e-01 -1.13617384e+00 -1.01776469e+00 2.76532620e-01
-3.16294469e-02 -3.56179863e-01 -1.75775737e-01 2.07903311e-02] | [10.39012622833252, 10.205150604248047] |
7edfd40f-f384-4209-95dc-8dd6733f90c4 | online-multi-object-tracking-with-delta-glmb | 2011.10111 | null | https://arxiv.org/abs/2011.10111v2 | https://arxiv.org/pdf/2011.10111v2.pdf | Online Multi-Object Tracking with delta-GLMB Filter based on Occlusion and Identity Switch Handling | In this paper, we propose an online multi-object tracking (MOT) method in a delta Generalized Labeled Multi-Bernoulli (delta-GLMB) filter framework to address occlusion and miss-detection issues, reduce false alarms, and recover identity switch (ID switch). To handle occlusion and miss-detection issues, we propose a measurement-to-disappeared track association method based on one-step delta-GLMB filter, so it is possible to manage these difficulties by jointly processing occluded or miss-detected objects. This part of proposed method is based on a proposed similarity metric which is responsible for defining the weight of hypothesized reappeared tracks. We also extend the delta-GLMB filter to efficiently recover switched IDs using the cardinality density, size and color features of the hypothesized tracks. We also propose a novel birth model to achieve more effective clutter removal performance. In both occlusion/miss-detection handler and newly-birthed object detector sections of the proposed method, unassigned measurements play a significant role, since they are used as the candidates for reappeared or birth objects. In addition, we perform an ablation study which confirms the effectiveness of our contributions in comparison with the baseline method. We evaluate the proposed method on well-known and publicly available MOT15 and MOT17 test datasets which are focused on pedestrian tracking. Experimental results show that the proposed tracker performs better or at least at the same level of the state-of-the-art online and offline MOT methods. It effectively handles the occlusion and ID switch issues and reduces false alarms as well. | ['Mohammad Ali Masnadi-Shirazi', 'Mohammadjavad Abbaspour'] | 2020-11-19 | null | null | null | null | ['online-multi-object-tracking'] | ['computer-vision'] | [-1.01628572e-01 -5.79341114e-01 3.81792672e-02 -6.37477934e-02
-5.84933221e-01 -4.46812510e-01 3.80823821e-01 3.08797956e-01
-5.45721292e-01 9.52611983e-01 -2.04423681e-01 -2.23371666e-03
-7.41056800e-02 -6.95552111e-01 -7.83162355e-01 -8.79835963e-01
-1.31928071e-01 5.56233466e-01 1.00300133e+00 2.18011335e-01
1.96850467e-02 5.21808088e-01 -1.99965417e+00 -4.67882752e-02
7.30712771e-01 9.27136302e-01 3.02602261e-01 7.40551114e-01
3.41330357e-02 4.18242782e-01 -7.08800495e-01 -3.82657886e-01
3.45233142e-01 9.25413072e-02 3.93933281e-02 -7.31633902e-02
9.00663733e-01 -3.40634227e-01 -1.24992058e-01 1.02587855e+00
6.49170220e-01 2.22678423e-01 5.28383255e-01 -1.40525854e+00
-8.51238146e-02 2.26269662e-01 -1.08229220e+00 7.65238822e-01
8.72014463e-02 1.06792408e-03 5.07142007e-01 -9.62177038e-01
2.10489169e-01 1.53047848e+00 8.60917270e-01 2.21834198e-01
-1.05627334e+00 -1.11201751e+00 3.41781586e-01 3.32845867e-01
-1.45057666e+00 -4.16105151e-01 3.20948064e-01 -5.08072376e-01
2.20002905e-01 5.50669432e-01 5.32284439e-01 8.57369483e-01
3.68368238e-01 7.93659806e-01 8.51801813e-01 -3.59968036e-01
1.92273613e-02 2.23302692e-01 5.01183093e-01 6.92053258e-01
9.85404730e-01 3.13175559e-01 -2.52600819e-01 -3.68281990e-01
5.73178828e-01 1.13297723e-01 -4.74446826e-02 -4.72771317e-01
-9.57220852e-01 5.14803588e-01 2.48619854e-01 4.00294140e-02
-2.56998926e-01 2.10839603e-02 2.14659542e-01 -1.92337245e-01
3.67022514e-01 -3.96878958e-01 -1.93571925e-01 1.22134864e-01
-8.84606004e-01 2.68069476e-01 4.46681678e-01 1.16816890e+00
5.71665347e-01 -9.54045262e-03 -7.30232835e-01 5.57355642e-01
5.50423861e-01 1.08220518e+00 -6.52230307e-02 -3.69497657e-01
3.99885416e-01 4.81683016e-01 6.04623854e-01 -1.05682003e+00
-4.66369808e-01 -8.85759294e-01 -6.10026538e-01 3.13184857e-01
4.01179612e-01 -3.10434643e-02 -9.12980378e-01 1.63885379e+00
9.28164601e-01 7.98295915e-01 -3.21232289e-01 8.93067956e-01
7.34460533e-01 3.88518959e-01 3.08191240e-01 -4.47080940e-01
1.77208185e+00 -7.07780778e-01 -7.83521175e-01 -1.64595112e-01
1.65047839e-01 -9.82881010e-01 3.27215016e-01 4.14368689e-01
-7.64383495e-01 -9.76881683e-01 -1.06940675e+00 6.17247105e-01
-1.18305683e-01 5.75030863e-01 3.53972763e-01 9.64969933e-01
-5.98259866e-01 2.45291844e-01 -8.50545228e-01 -4.85141486e-01
3.53420049e-01 3.73771250e-01 7.84236714e-02 -6.86667114e-02
-6.51111722e-01 7.64448464e-01 2.81206280e-01 1.82000041e-01
-8.95431399e-01 -6.68961704e-01 -5.10545492e-01 -8.36498737e-02
4.70965892e-01 -5.83420098e-01 8.45674157e-01 -4.78214324e-01
-8.19633245e-01 3.85906547e-01 -3.87298137e-01 -6.57300830e-01
4.39877301e-01 -4.57308352e-01 -7.79547751e-01 -6.34375289e-02
3.46112520e-01 3.64036143e-01 8.99154663e-01 -1.33695030e+00
-1.30690777e+00 -3.30347180e-01 -3.15872759e-01 -1.11042097e-01
-1.58176333e-01 9.78582576e-02 -5.86450160e-01 -7.28140652e-01
5.12861684e-02 -9.18429315e-01 5.41626662e-02 5.76178730e-02
-2.29527935e-01 -1.10370323e-01 9.80460286e-01 -4.89091307e-01
1.36078155e+00 -2.16395378e+00 -3.44368666e-01 9.54187810e-02
1.46938905e-01 6.32729471e-01 -2.94747353e-02 1.40344843e-01
3.04574102e-01 -5.07791221e-01 2.41227061e-01 -6.30419612e-01
-1.71733499e-01 8.62929597e-02 -2.45473415e-01 7.88006544e-01
3.89979989e-03 2.04822540e-01 -9.02628958e-01 -6.01557434e-01
5.84795177e-01 5.34488440e-01 -1.08971864e-01 5.42486347e-02
2.17419371e-01 5.57006538e-01 -4.55444902e-01 8.34681749e-01
1.30541575e+00 1.36284932e-01 -2.33644307e-01 -6.00703180e-01
-5.05992591e-01 -2.88702458e-01 -1.81336451e+00 9.35131252e-01
-6.66825473e-02 3.18338841e-01 1.48867741e-01 -6.59994543e-01
8.37468863e-01 7.66122565e-02 5.89931726e-01 -5.53143501e-01
2.45479941e-01 8.83003846e-02 -3.49456482e-02 -1.42909274e-01
6.13888860e-01 1.84766233e-01 2.37451181e-01 -4.37372625e-02
-1.37495130e-01 1.10891628e+00 4.07216668e-01 1.66226611e-01
1.08222842e+00 1.82534128e-01 1.84482530e-01 -3.11824501e-01
9.28057015e-01 5.47478572e-02 1.09241295e+00 1.21334803e+00
-6.00311637e-01 2.58016467e-01 -3.54011536e-01 -3.30222905e-01
-5.97817779e-01 -1.33164465e+00 -3.04218173e-01 9.61110830e-01
7.33624220e-01 -1.86869219e-01 -3.17159027e-01 -3.94343346e-01
3.05740416e-01 5.37525833e-01 -4.73916978e-01 -8.35447460e-02
-6.98837161e-01 -1.12281609e+00 4.77291763e-01 5.03820360e-01
3.58260840e-01 -7.42482126e-01 -8.15193415e-01 5.34923077e-01
-1.17141508e-01 -1.17567885e+00 -4.10306782e-01 -1.51152670e-01
-7.14856803e-01 -1.28733754e+00 -5.12738347e-01 -4.31501746e-01
6.40595555e-01 8.78056347e-01 7.01162934e-01 1.32752746e-01
-5.74559093e-01 4.18108106e-01 -5.30314565e-01 -5.06598890e-01
-1.03712924e-01 -4.53116447e-01 3.59773993e-01 4.59652871e-01
2.38534704e-01 -3.02120030e-01 -8.31856012e-01 7.87217796e-01
-4.12600487e-01 -2.73195118e-01 6.26382709e-01 5.56517243e-01
7.61905193e-01 7.86230564e-02 6.03813231e-01 -4.22932982e-01
2.31979620e-02 -4.22608197e-01 -1.04286885e+00 3.22497398e-01
-4.64420259e-01 -1.31541252e-01 2.74579853e-01 -7.71049738e-01
-1.05560136e+00 3.49650942e-02 1.69135451e-01 -5.23891449e-01
-8.86297673e-02 -2.92264074e-01 -2.21912593e-01 -4.05435234e-01
2.61595815e-01 -3.07984706e-02 -3.37004513e-01 -6.99938893e-01
8.94800127e-02 4.26659822e-01 6.50818467e-01 -5.44853091e-01
9.50241089e-01 9.96047378e-01 2.63001561e-01 -6.29863620e-01
-6.97562099e-01 -1.01883900e+00 -3.52549851e-01 -5.34106970e-01
6.63865328e-01 -1.11367393e+00 -1.01149142e+00 3.33948612e-01
-1.00746965e+00 4.41407204e-01 5.06865419e-02 7.09873259e-01
8.87168273e-02 6.09855592e-01 -2.36465245e-01 -1.53996325e+00
-3.51399392e-01 -1.04144049e+00 1.16675913e+00 4.79309887e-01
4.83501226e-01 -5.44529557e-01 3.03244088e-02 1.15988314e-01
2.43291646e-01 3.73966783e-01 1.10186681e-01 -5.28448641e-01
-9.51020718e-01 -2.68192261e-01 -3.38110387e-01 -1.37169823e-01
8.26033577e-02 -1.56465396e-01 -6.82121456e-01 -8.20225954e-01
-2.73228437e-01 4.02939707e-01 1.05158532e+00 7.14265823e-01
6.38725638e-01 1.90104485e-01 -9.71267402e-01 3.37884665e-01
1.38674402e+00 2.55605966e-01 4.88721490e-01 4.58500117e-01
6.37904346e-01 1.87736243e-01 1.32201147e+00 7.05898881e-01
4.30153817e-01 9.37836349e-01 6.06877565e-01 2.89844237e-02
-4.25061107e-01 1.12865707e-02 4.15728599e-01 4.05661613e-01
1.01818345e-01 -3.87013882e-01 -3.39651734e-01 5.67811847e-01
-2.19613194e+00 -1.00955796e+00 -8.22024047e-01 2.70614481e+00
1.69133946e-01 4.20970261e-01 3.84036511e-01 2.76093278e-02
1.21599567e+00 -2.56077617e-01 -3.31397474e-01 3.56299549e-01
-4.52836826e-02 1.11979984e-01 9.03374851e-01 3.61372232e-01
-1.45917749e+00 5.47313273e-01 4.90741253e+00 9.47253168e-01
-4.64282066e-01 4.00821865e-01 -1.05158582e-01 6.54693646e-03
5.75439572e-01 2.63305213e-02 -1.71786439e+00 6.89001322e-01
6.97260916e-01 3.42014521e-01 -1.09403254e-02 5.96856296e-01
2.71611720e-01 -5.13299942e-01 -6.96068048e-01 1.16509259e+00
1.13694556e-01 -8.32591712e-01 -1.27696559e-01 4.73937020e-02
1.72280282e-01 -3.24699461e-01 -1.30761176e-01 3.98049563e-01
1.50336459e-01 -1.39724866e-01 8.93367529e-01 6.43289089e-01
2.92721480e-01 -6.19439781e-01 6.60610259e-01 3.39090019e-01
-2.06714916e+00 -2.67908245e-01 -5.30274868e-01 1.29211739e-01
4.64169115e-01 7.60423958e-01 -7.07973361e-01 9.60169196e-01
7.96540797e-01 4.07923579e-01 -7.72592425e-01 1.81063712e+00
2.54599899e-01 5.04921377e-01 -4.99582052e-01 1.02742545e-01
-2.50608712e-01 1.38628587e-01 1.06462336e+00 1.17684376e+00
3.86559129e-01 -5.23520224e-02 6.56783044e-01 5.47292054e-01
5.50477445e-01 -7.65029863e-02 -3.28736678e-02 6.30439401e-01
8.31995726e-01 1.38115752e+00 -9.16373491e-01 -4.83569413e-01
-4.19119120e-01 3.19925666e-01 -5.65316044e-02 1.09659977e-01
-1.33429575e+00 -1.06492758e-01 5.94378769e-01 2.03139737e-01
7.97439635e-01 -2.43836958e-02 2.03020751e-01 -1.00571096e+00
-4.66268370e-03 -4.94351089e-01 7.02549636e-01 -3.93241316e-01
-1.22813022e+00 3.49428803e-01 2.21639931e-01 -1.56250858e+00
3.70096654e-01 -3.73123407e-01 -4.62902129e-01 6.96406722e-01
-1.48479152e+00 -1.31242394e+00 -5.39010584e-01 5.21030366e-01
5.82319081e-01 -6.17630780e-02 1.72109306e-01 1.07422864e+00
-6.83148146e-01 7.77970135e-01 3.00140865e-02 -1.63781554e-01
8.05679440e-01 -7.64369011e-01 6.30194917e-02 1.22963536e+00
-2.76830569e-02 4.88687485e-01 9.84579861e-01 -1.19150782e+00
-1.32602024e+00 -1.43099558e+00 1.67420208e-01 -2.62820691e-01
2.96647996e-01 -4.02044207e-01 -7.07570314e-01 5.51007628e-01
-3.11865747e-01 1.16359785e-01 4.17160183e-01 3.74313518e-02
3.82410586e-02 -4.87317741e-01 -1.16194284e+00 1.89134449e-01
1.06733310e+00 4.91517544e-01 -4.81859803e-01 2.31569201e-01
3.55158061e-01 -3.78494799e-01 -5.10014057e-01 6.76721215e-01
6.89495206e-01 -8.02367508e-01 1.28254271e+00 -2.13479158e-02
-9.46527064e-01 -1.02607024e+00 -7.43993074e-02 -6.98230147e-01
-6.46465361e-01 -3.59691739e-01 -4.21195507e-01 1.60474479e+00
-1.40680671e-01 -6.07048035e-01 4.98469651e-01 7.51739088e-03
-2.10626632e-01 -1.92842856e-01 -1.18143487e+00 -1.09627950e+00
-7.70406067e-01 -4.00665924e-02 3.65965009e-01 3.78901958e-01
-8.02140653e-01 -6.58127144e-02 -6.96917653e-01 8.11700046e-01
1.43563032e+00 6.67246357e-02 1.12674320e+00 -1.56773150e+00
-3.11256349e-01 1.07547112e-01 -5.06722689e-01 -9.06876206e-01
-3.73313069e-01 -5.16400635e-01 2.03609273e-01 -1.25816846e+00
4.09172297e-01 -6.49715900e-01 -5.84589899e-01 8.57786462e-02
-3.55112612e-01 2.06973672e-01 2.23195195e-01 2.48686373e-01
-9.56243813e-01 3.90574366e-01 7.45239735e-01 6.63552666e-03
-2.20106781e-01 3.94052744e-01 -3.07411611e-01 5.16563177e-01
4.23369437e-01 -8.33206654e-01 5.97935207e-02 -2.44565517e-01
-3.64951670e-01 -1.25092879e-01 8.82145524e-01 -1.57932675e+00
3.91375333e-01 1.42188728e-01 5.93037426e-01 -1.38440502e+00
4.95362759e-01 -8.84721279e-01 5.07016242e-01 7.18769550e-01
3.15746993e-01 7.23483711e-02 5.25121272e-01 1.03246117e+00
3.27506781e-01 -7.02995956e-02 8.47497880e-01 1.33095458e-01
-5.81245065e-01 2.18914658e-01 -2.24371046e-01 -3.59872162e-01
1.20309770e+00 -3.86531472e-01 -5.07415652e-01 1.64914265e-01
-6.65634453e-01 5.69983900e-01 4.04149406e-02 6.06948078e-01
4.19632196e-01 -1.36109793e+00 -7.18838334e-01 8.99252892e-02
1.34131074e-01 -5.05588353e-01 3.24490935e-01 1.14387655e+00
1.83077878e-03 3.44080031e-01 -1.96201101e-01 -9.58967090e-01
-1.77248585e+00 6.13698900e-01 -3.53127543e-04 -2.74836332e-01
-7.21678615e-01 5.34425199e-01 3.42428803e-01 2.15874106e-01
4.39109474e-01 -3.03456157e-01 -2.75402129e-01 -1.15099348e-01
7.99921393e-01 7.39914596e-01 -9.44705606e-02 -8.48641634e-01
-7.49552310e-01 5.59072375e-01 -1.65552422e-01 1.26906171e-01
8.39792132e-01 -3.19808334e-01 2.23463193e-01 2.01299921e-01
2.90375441e-01 2.12912306e-01 -1.26347077e+00 -1.11057021e-01
-1.43230408e-01 -6.56608641e-01 -6.79939091e-02 -5.33527076e-01
-8.36185753e-01 4.63010043e-01 1.36953259e+00 -6.77527413e-02
8.31740499e-01 -1.92074731e-01 8.48384380e-01 -1.63620114e-01
6.43771231e-01 -7.89965570e-01 -7.60833845e-02 1.64834991e-01
2.77537853e-01 -1.09393179e+00 2.96977252e-01 -6.23538315e-01
2.62011606e-02 7.29899883e-01 9.19749618e-01 -1.33139923e-01
5.67772627e-01 2.99730510e-01 -4.18965071e-01 3.09526315e-03
-3.61150444e-01 -6.72337651e-01 2.35962778e-01 6.55702710e-01
-1.41959861e-01 1.43099636e-01 -6.10072196e-01 5.64783394e-01
4.43927020e-01 -1.24542527e-01 2.47851685e-01 1.03189671e+00
-9.80385244e-01 -1.05388403e+00 -1.22939098e+00 4.15443122e-01
-2.20764011e-01 2.72307068e-01 2.16828972e-01 8.09849203e-01
5.41183293e-01 1.21376920e+00 7.72599056e-02 -3.41527969e-01
5.00627816e-01 -3.51670325e-01 6.41448796e-01 -2.90528774e-01
-5.47023177e-01 3.92634034e-01 4.63487506e-02 -3.04541260e-01
-5.94455183e-01 -8.14936399e-01 -1.10884273e+00 -9.73398536e-02
-9.14174616e-01 1.80322498e-01 5.16534448e-01 8.27670097e-01
4.97781038e-01 8.19601476e-01 3.05802196e-01 -7.84651935e-01
-5.06612420e-01 -8.98584247e-01 -4.55068141e-01 3.63546997e-01
3.09338301e-01 -1.63274729e+00 -1.66400433e-01 -1.75700471e-01] | [6.47983455657959, -2.0075182914733887] |
6ca56e04-59a6-4c34-93f9-35a27b6a759f | a-deep-learning-approach-for-digital | 2202.0527 | null | https://arxiv.org/abs/2202.05270v2 | https://arxiv.org/pdf/2202.05270v2.pdf | A Deep Learning Approach for Digital Color Reconstruction of Lenticular Films | We propose the first accurate digitization and color reconstruction process for historical lenticular film that is robust to artifacts. Lenticular films emerged in the 1920s and were one of the first technologies that permitted to capture full color information in motion. The technology leverages an RGB filter and cylindrical lenticules embossed on the film surface to encode the color in the horizontal spatial dimension of the image. To project the pictures the encoding process was reversed using an appropriate analog device. In this work, we introduce an automated, fully digital pipeline to process the scan of lenticular films and colorize the image. Our method merges deep learning with a model-based approach in order to maximize the performance while making sure that the reconstructed colored images truthfully match the encoded color information. Our model employs different strategies to achieve an effective color reconstruction, in particular (i) we use data augmentation to create a robust lenticule segmentation network, (ii) we fit the lenticules raster prediction to obtain a precise vectorial lenticule localization, and (iii) we train a colorization network that predicts interpolation coefficients in order to obtain a truthful colorization. We validate the proposed method on a lenticular film dataset and compare it to other approaches. Since no colored groundtruth is available as reference, we conduct a user study to validate our method in a subjective manner. The results of the study show that the proposed method is largely preferred with respect to other existing and baseline methods. | ['Jan Dirk Wegner', 'David Pfluger', 'Giorgio Trumpy', "Stefano D'Aronco"] | 2022-02-10 | null | null | null | null | ['colorization'] | ['computer-vision'] | [ 2.95841336e-01 -3.48043650e-01 2.72355348e-01 -2.00160176e-01
-5.05961299e-01 -8.97031665e-01 3.92646194e-01 -4.35477167e-01
-3.06333393e-01 6.04443252e-01 -1.73034176e-01 -2.76648462e-01
3.23527485e-01 -7.40968406e-01 -9.59489763e-01 -3.40451717e-01
3.11702251e-01 7.52911195e-02 3.97631794e-01 -2.80402601e-02
5.12414813e-01 5.79939961e-01 -1.52057052e+00 2.97665149e-01
9.57837999e-01 1.30758047e+00 2.17086107e-01 8.80681992e-01
-1.66565981e-02 5.24714530e-01 -3.97149205e-01 -4.85092878e-01
7.95672476e-01 -4.41094995e-01 -3.70035112e-01 9.46040973e-02
7.37907708e-01 -7.87730038e-01 -1.71211854e-01 1.07292449e+00
8.25217590e-02 -3.45362306e-01 6.95677102e-01 -8.24368894e-01
-6.57837391e-01 -1.31067127e-01 -9.58519459e-01 -4.52850938e-01
4.11764830e-01 2.54311144e-01 6.52637482e-01 -6.76759899e-01
9.80159521e-01 8.34203005e-01 7.36564696e-01 3.90918940e-01
-1.14970207e+00 -7.97492683e-01 -4.01930839e-01 1.12275936e-01
-1.16112423e+00 -1.81839719e-01 1.11468887e+00 -5.29612422e-01
8.28399584e-02 4.07367796e-01 1.16455936e+00 8.18260491e-01
4.03785586e-01 5.79467595e-01 1.67965364e+00 -6.93004966e-01
3.01925510e-01 6.96851239e-02 -2.71270245e-01 8.22630882e-01
1.32481039e-01 2.51778543e-01 -3.95827621e-01 2.93278426e-01
1.30489945e+00 -8.33226666e-02 -3.35482508e-01 -2.71615446e-01
-7.76721835e-01 3.77164274e-01 4.09999996e-01 1.17084280e-01
-3.08347702e-01 3.06441963e-01 -1.96670458e-01 1.19471690e-02
2.95607656e-01 5.23735344e-01 3.17869857e-02 -9.02100801e-02
-1.46372211e+00 1.26487404e-01 5.43720961e-01 7.47374773e-01
8.88047755e-01 1.24915587e-02 9.10869688e-02 6.63352311e-01
4.81724858e-01 6.31824911e-01 2.61573225e-01 -1.31066549e+00
3.61942142e-01 6.19593740e-01 3.52519184e-01 -1.01488638e+00
-1.93788081e-01 2.06911564e-01 -5.29659331e-01 9.92931604e-01
6.07380509e-01 -1.52678996e-01 -1.21653295e+00 1.06839883e+00
2.65101194e-02 2.90727615e-01 -2.57854551e-01 1.23637235e+00
3.74910802e-01 8.61567020e-01 -3.44542861e-01 2.39400510e-02
1.10697162e+00 -5.55048585e-01 -7.33540595e-01 4.26333658e-02
-6.79606721e-02 -1.15516508e+00 1.44758177e+00 7.71100163e-01
-1.28844535e+00 -4.44599301e-01 -1.63409698e+00 -2.32193843e-01
-1.33483514e-01 3.19124669e-01 7.03045607e-01 8.77493799e-01
-1.04108608e+00 6.75055504e-01 -9.05751228e-01 -1.83643252e-01
2.37426609e-01 2.44451631e-02 -3.03614795e-01 -9.11881626e-02
-8.96914542e-01 5.31276107e-01 8.66076648e-02 1.05653495e-01
-5.84450364e-01 -7.67419577e-01 -4.40665364e-01 -2.37535685e-01
-7.01268241e-02 -4.03464526e-01 8.75592232e-01 -1.37770855e+00
-2.01069927e+00 8.94475758e-01 5.11626415e-02 -1.64318457e-01
8.02882731e-01 -1.86246440e-01 -2.77386695e-01 4.60756302e-01
-2.90496975e-01 5.53952873e-01 8.65549445e-01 -1.70954537e+00
-5.91990471e-01 -1.17498770e-01 -4.94605079e-02 -6.38965815e-02
-2.11121574e-01 -3.83043081e-01 -1.06233561e+00 -5.99686503e-01
2.56265819e-01 -7.86951423e-01 2.29838252e-01 3.94214958e-01
-5.89337170e-01 6.11644804e-01 7.53026366e-01 -1.01804376e+00
1.15275633e+00 -1.79819739e+00 -2.48130068e-01 5.34915864e-01
1.96521670e-01 2.64607280e-01 -2.58720033e-02 2.67264485e-01
2.25470111e-01 -8.47344473e-03 -3.39878529e-01 -4.77968335e-01
-2.43517488e-01 -2.79609054e-01 -1.24093987e-01 5.21843314e-01
9.31652077e-03 7.58704960e-01 -5.17969072e-01 -4.55225676e-01
4.42542791e-01 6.91437364e-01 -4.60473180e-01 2.94487923e-01
-3.26432645e-01 3.21166635e-01 -1.31505638e-01 1.08736300e+00
1.04564261e+00 -3.68257724e-02 2.06155941e-01 -6.83702052e-01
-3.96772563e-01 -4.12793845e-01 -1.11954892e+00 1.78795886e+00
-4.24285471e-01 1.03973258e+00 -1.39508548e-03 -9.48458724e-03
9.85354364e-01 2.27484442e-02 5.30710161e-01 -1.15526938e+00
2.95006692e-01 3.24297547e-01 -3.13371241e-01 -5.28988957e-01
8.57370079e-01 -1.40531629e-01 2.62294263e-01 4.00141299e-01
-3.97213072e-01 -4.62180763e-01 6.47723377e-02 7.87714720e-02
6.86920226e-01 6.97696507e-01 -1.95565179e-01 8.63291509e-03
2.81908691e-01 1.39407113e-01 4.48525310e-01 4.44299012e-01
-2.25372892e-02 1.29880702e+00 4.23634052e-01 -4.68656421e-01
-1.55994165e+00 -1.14382041e+00 -5.43315187e-02 2.01722220e-01
4.93726760e-01 -6.95670992e-02 -1.12470067e+00 -1.78319126e-01
-1.68007851e-01 6.91997945e-01 -6.74629867e-01 3.66095066e-01
-4.40435886e-01 -4.41298485e-01 2.15863824e-01 3.74981731e-01
7.27450252e-01 -7.37878501e-01 -8.39263141e-01 -8.52911323e-02
1.86820462e-01 -8.59062374e-01 -4.37220216e-01 -2.04107225e-01
-6.10150337e-01 -1.42247391e+00 -8.70620906e-01 -5.36198914e-01
6.72503233e-01 -2.91054090e-03 7.26653755e-01 -1.77237481e-01
-5.69183052e-01 3.01075488e-01 -2.87220597e-01 -1.13961942e-01
-3.38232905e-01 -3.90443206e-01 -3.03695828e-01 2.46988073e-01
-3.85095505e-03 -4.45254743e-01 -1.24348497e+00 4.45717834e-02
-1.08646810e+00 5.72786391e-01 6.38276696e-01 2.20005780e-01
7.02937186e-01 -2.62235224e-01 -1.06990688e-01 -1.01683116e+00
5.41858137e-01 -2.45526507e-02 -1.19907415e+00 3.20165306e-01
-6.63392007e-01 -1.32098705e-01 6.23539865e-01 -2.70828635e-01
-1.21284485e+00 2.87181079e-01 8.34359601e-03 -6.17685258e-01
1.56302109e-01 2.80641139e-01 7.93295354e-02 -3.44072282e-01
5.98544419e-01 7.64746740e-02 -4.23344858e-02 -4.13590997e-01
4.99706715e-01 6.10475242e-01 7.67809153e-01 -3.52995336e-01
9.00646448e-01 8.28832269e-01 2.96624377e-02 -6.94212139e-01
-1.67757690e-01 -5.26940858e-04 -7.56151378e-01 -8.66254330e-01
1.03601050e+00 -5.58947384e-01 -1.02053475e+00 5.25329888e-01
-1.10277104e+00 -6.16740286e-01 1.27654403e-01 2.54314274e-01
-5.20632386e-01 3.30482483e-01 -7.67816246e-01 -8.62868905e-01
-2.77671427e-01 -9.69009459e-01 9.04769838e-01 7.28819370e-01
3.20868753e-03 -8.32185984e-01 2.39243567e-01 2.25253969e-01
3.79945010e-01 5.77685237e-01 7.96984076e-01 5.20556033e-01
-9.07847583e-01 -3.74008417e-01 -5.38057506e-01 3.75794441e-01
1.56226203e-01 5.22032082e-01 -1.10937524e+00 9.19248611e-02
-2.98022896e-01 -2.38348931e-01 6.15186930e-01 3.72500271e-01
1.00552034e+00 7.57599995e-02 3.76011468e-02 1.05733120e+00
2.01663089e+00 4.53782499e-01 1.14179957e+00 6.30254686e-01
7.52551615e-01 2.86628842e-01 4.37774599e-01 2.63122588e-01
2.55798846e-01 5.85042596e-01 3.47919434e-01 -4.93045002e-01
-5.87790132e-01 -4.50216353e-01 1.62636757e-01 6.47373140e-01
-3.40692222e-01 -1.52483806e-01 -7.40764856e-01 1.73359841e-01
-1.30369782e+00 -6.14284575e-01 -1.72183067e-01 2.48076606e+00
8.51860344e-01 -3.37104164e-02 -1.38113841e-01 -3.23298797e-02
5.01096427e-01 -1.56158626e-01 -5.13832629e-01 -4.03895319e-01
-2.64175057e-01 4.11685109e-01 8.35848749e-01 6.47307813e-01
-8.62865269e-01 1.04119122e+00 6.28538895e+00 3.82147312e-01
-1.73573327e+00 -1.13556899e-01 7.43542373e-01 -9.09832865e-02
-5.31208158e-01 3.74436602e-02 -2.56792396e-01 5.35139024e-01
3.30649585e-01 2.97309488e-01 7.10756302e-01 3.32955897e-01
4.18594420e-01 -4.24581498e-01 -7.35822499e-01 1.11728168e+00
1.40596494e-01 -1.45955479e+00 -1.11323722e-01 -2.86655366e-01
1.06458509e+00 -2.84518212e-01 1.72900572e-01 -5.11404693e-01
5.93838207e-02 -8.46336663e-01 1.06866753e+00 1.17704535e+00
1.14593899e+00 -4.68929231e-01 2.60717332e-01 -2.47568265e-01
-8.98329139e-01 2.65827209e-01 -2.16096029e-01 1.98912963e-01
2.85917073e-01 3.16289425e-01 -7.44689465e-01 3.57607096e-01
5.70212007e-01 4.43926632e-01 -6.77146196e-01 1.29015768e+00
-2.75347203e-01 6.06449664e-01 -2.47523800e-01 9.01736617e-02
9.73605141e-02 -7.75052249e-01 1.86822146e-01 9.62315500e-01
4.82201517e-01 -1.24230729e-02 -4.24465179e-01 1.25572956e+00
-8.66503045e-02 6.52002990e-02 -1.66067362e-01 -3.07107233e-02
4.04837221e-01 1.37040603e+00 -8.68263066e-01 -2.64758784e-02
-4.41845924e-01 1.34402502e+00 4.75731166e-03 7.80321956e-01
-7.90370226e-01 -5.16618609e-01 2.15141907e-01 5.94162405e-01
-1.86303388e-02 -3.79892439e-01 -6.75027549e-01 -8.64856660e-01
8.98482427e-02 -3.63984555e-01 -1.86026245e-01 -1.39245772e+00
-1.01229239e+00 7.19008267e-01 -4.16271091e-01 -1.44825733e+00
9.06939991e-03 -7.64093101e-01 -5.67956626e-01 9.91486788e-01
-1.63764501e+00 -1.35564184e+00 -5.36822617e-01 3.96511883e-01
1.66577637e-01 -2.11194828e-02 5.14243603e-01 3.38625908e-01
-3.96625906e-01 3.54454249e-01 2.51374632e-01 -5.94190545e-02
8.20021451e-01 -1.16064548e+00 3.05944961e-02 1.06649673e+00
-2.02801764e-01 4.94500488e-01 6.86915338e-01 -9.03939128e-01
-1.66944671e+00 -9.23914313e-01 2.34910548e-02 -1.54388547e-01
4.15835023e-01 -4.21851277e-01 -7.12209761e-01 4.84662056e-01
2.88661242e-01 -2.36482859e-01 2.44665384e-01 -4.97267872e-01
-9.25881267e-02 -4.47950512e-01 -1.01807916e+00 6.49700344e-01
3.78824979e-01 -4.30583984e-01 -4.97062728e-02 -1.67068299e-02
1.95483699e-01 -5.85549295e-01 -6.39712691e-01 1.17067784e-01
1.17309332e+00 -1.26932728e+00 6.21347666e-01 7.10246414e-02
7.82920599e-01 -6.06197715e-01 -9.67528969e-02 -8.70797217e-01
1.67689323e-02 -6.21595025e-01 2.49117076e-01 1.17149734e+00
4.90176558e-01 -3.23133051e-01 1.17332745e+00 9.45671976e-01
-1.30389407e-01 -5.61999679e-01 -7.27906883e-01 -3.62250149e-01
-8.18058029e-02 -5.29407680e-01 4.16953951e-01 7.19974637e-01
-4.26155508e-01 -3.77818853e-01 -6.15299463e-01 1.23625852e-01
6.76401794e-01 2.86124855e-01 6.48809135e-01 -9.05336082e-01
-1.17519960e-01 -3.77969027e-01 -2.59533882e-01 -7.52661586e-01
-2.75346249e-01 -6.29226863e-01 6.44128844e-02 -1.69881392e+00
3.13364491e-02 -6.62386954e-01 -1.29840197e-02 1.60280064e-01
-7.54020587e-02 8.18929195e-01 3.72868747e-01 3.03880751e-01
-1.19922802e-01 2.13463530e-01 1.42520392e+00 4.04779725e-02
-5.02147317e-01 -2.32719928e-01 -4.37044054e-01 6.64710641e-01
6.51166439e-01 1.69830509e-02 -3.02471906e-01 -5.18351793e-01
4.34434861e-01 1.34805040e-02 3.57431501e-01 -1.16307688e+00
2.78942108e-01 -9.53042805e-02 9.99456227e-01 -5.14073014e-01
5.42194009e-01 -1.04385269e+00 2.52251565e-01 3.04143965e-01
4.97092269e-02 -2.15272412e-01 1.35028452e-01 2.62331784e-01
2.84194797e-02 -4.62187119e-02 1.00154603e+00 1.92213375e-02
-7.83168912e-01 1.84203848e-01 -1.29013836e-01 -5.18073082e-01
1.06633377e+00 -4.99409527e-01 -2.99890429e-01 -4.91058648e-01
-2.65243769e-01 -2.32689127e-01 1.18555093e+00 1.83161438e-01
8.43344450e-01 -1.22506225e+00 -3.21611047e-01 4.25520301e-01
-1.14552788e-01 -4.73668754e-01 1.56993002e-01 4.76986408e-01
-1.60133266e+00 1.42033353e-01 -4.97987151e-01 -4.87649351e-01
-8.82677138e-01 3.17055523e-01 3.40548664e-01 2.83062279e-01
-7.09582627e-01 5.12164235e-01 1.66500825e-02 9.66140777e-02
-4.36200686e-02 -3.60954970e-01 3.41551527e-02 -2.04242244e-01
4.42823499e-01 3.49344075e-01 -1.17821433e-01 -3.35431397e-01
3.76033895e-02 1.10814178e+00 1.42739847e-01 -5.60155809e-01
1.24845779e+00 -1.44910857e-01 -3.48956473e-02 4.39856946e-01
1.02267444e+00 2.86291808e-01 -1.91349971e+00 3.44020426e-01
-4.18528229e-01 -7.73173928e-01 2.14427099e-01 -1.11766315e+00
-1.38391209e+00 8.91544342e-01 1.06231773e+00 3.53966607e-03
1.42965889e+00 -4.53366220e-01 8.78441930e-01 -3.57545823e-01
2.93837011e-01 -1.22923422e+00 -1.71464887e-02 1.45221995e-02
7.13879228e-01 -9.36322391e-01 1.51250511e-01 -4.52616572e-01
-6.17477596e-01 1.53856289e+00 5.34177244e-01 -2.80844659e-01
3.26948494e-01 3.64310443e-01 3.82111877e-01 -1.21441655e-01
-8.19308683e-02 1.04693331e-01 5.59788764e-01 3.95642966e-01
5.41116714e-01 -4.04556468e-02 -3.06167424e-01 4.99957085e-01
-2.00169683e-01 3.29223037e-01 6.58622444e-01 7.24937379e-01
-2.33089581e-01 -1.09781897e+00 -6.57109439e-01 2.99680561e-01
-4.28810678e-02 -5.73185179e-03 -3.88070703e-01 8.02903712e-01
1.42959476e-01 6.50549889e-01 1.82407707e-01 -3.34774077e-01
3.86539906e-01 -1.72661811e-01 7.48776197e-01 -5.48192188e-02
-2.29956493e-01 3.43789637e-01 -2.01392353e-01 -8.38006854e-01
-1.11089811e-01 -3.81957918e-01 -1.24646258e+00 -2.03197017e-01
2.09701568e-01 -4.30691510e-01 1.17017126e+00 5.71574748e-01
1.03353158e-01 4.17358518e-01 6.60406232e-01 -1.06547678e+00
3.48137468e-01 -6.47050619e-01 -7.93680251e-01 5.95159352e-01
2.01835185e-01 -3.58045667e-01 -6.00702725e-02 4.10112530e-01] | [10.89099407196045, -1.5974209308624268] |
d47b00cd-734c-43c9-8ff7-e1584fad0e6e | generalizing-adam-to-manifolds-for | 2305.16901 | null | https://arxiv.org/abs/2305.16901v1 | https://arxiv.org/pdf/2305.16901v1.pdf | Generalizing Adam To Manifolds For Efficiently Training Transformers | One of the primary reasons behind the success of neural networks has been the emergence of an array of new, highly-successful optimizers, perhaps most importantly the Adam optimizer. It is wiedely used for training neural networks, yet notoriously hard to interpret. Lacking a clear physical intuition, Adam is difficult to generalize to manifolds. Some attempts have been made to directly apply parts of the Adam algorithm to manifolds or to find an underlying structure, but a full generalization has remained elusive. In this work a new approach is presented that leverages the special structure of the manifolds which are relevant for optimization of neural networks, such as the Stiefel manifold, the symplectic Stiefel manifold, the Grassmann manifold and the symplectic Grassmann manifold: all of these are homogeneous spaces and as such admit a global tangent space representation. This global tangent space representation is used to perform all of the steps in the Adam optimizer. The resulting algorithm is then applied to train a transformer for which orthogonality constraints are enforced up to machine precision and we observe significant speed-ups in the training process. Optimization of neural networks where they weights do not lie on a manifold is identified as a special case of the presented framkework. This allows for a flexible implementation in which the learning rate is adapted simultaneously for all parameters, irrespective of whether they are an element of a general manifold or a vector space. | ['Benedikt Brantner'] | 2023-05-26 | null | null | null | null | ['physical-intuition'] | ['reasoning'] | [-6.31938875e-02 2.45077625e-01 -7.65343010e-02 -1.19512998e-01
7.66822994e-02 -6.34482265e-01 7.65338719e-01 -1.22545625e-03
-6.01576746e-01 5.97836971e-01 -1.67437896e-01 -4.28350955e-01
-2.96517134e-01 -4.92748320e-01 -7.45814502e-01 -1.10295737e+00
-3.06477368e-01 3.83468896e-01 3.23697999e-02 -6.23432755e-01
2.31486902e-01 7.78819561e-01 -1.37316763e+00 -4.23870802e-01
6.66826129e-01 7.71417797e-01 -1.55444384e-01 6.95216715e-01
1.48480441e-02 2.65574664e-01 -2.52349079e-01 -3.81284624e-01
3.44172657e-01 -2.70701885e-01 -1.03199220e+00 8.75634551e-02
4.44196284e-01 2.14727208e-01 -2.26141542e-01 9.43933785e-01
1.16918489e-01 3.34434807e-01 6.86931431e-01 -1.14606798e+00
-3.12997997e-01 2.61219174e-01 -6.47748858e-02 2.04364598e-01
-1.84136614e-01 -1.15459479e-01 1.31382215e+00 -9.59696174e-01
5.44203162e-01 9.86936390e-01 6.29322529e-01 5.01197457e-01
-1.24195743e+00 3.42387334e-02 -1.03814811e-01 2.33136594e-01
-1.16770446e+00 -3.49858791e-01 7.04893172e-01 -4.47529465e-01
8.17251742e-01 3.39197844e-01 7.30488002e-01 8.61293793e-01
4.75084394e-01 4.47383463e-01 9.08801019e-01 -6.04642272e-01
3.09435219e-01 3.29998672e-01 3.05749625e-01 8.53488028e-01
2.17641830e-01 5.17426170e-02 -2.29626447e-01 9.34712812e-02
8.73358548e-01 -1.79801181e-01 -2.72670656e-01 -6.30172908e-01
-1.19746411e+00 1.05491161e+00 5.72645247e-01 5.34922540e-01
-2.51088113e-01 7.96181441e-04 2.21245810e-01 5.10267138e-01
2.88803160e-01 6.53924882e-01 -2.20055029e-01 -9.70147848e-02
-6.43737257e-01 2.09000319e-01 1.06072307e+00 4.15339679e-01
8.38289797e-01 1.68013543e-01 5.37549078e-01 4.34692442e-01
2.49029949e-01 5.16655780e-02 4.73317802e-01 -9.59030628e-01
4.38716024e-01 6.70286655e-01 -3.74077484e-02 -1.20278621e+00
-6.95140541e-01 -6.31340623e-01 -1.01296079e+00 5.14772236e-01
6.77104712e-01 -2.06234738e-01 -4.21030045e-01 1.69302404e+00
6.10114098e-01 -6.32875934e-02 2.90924847e-01 1.02698147e+00
1.61937952e-01 4.47548509e-01 -3.95356953e-01 2.39900261e-01
1.20873106e+00 -5.67077696e-01 -2.34269544e-01 6.12546690e-03
9.05596495e-01 -5.77668786e-01 8.55810106e-01 3.97674024e-01
-1.08955657e+00 -2.31922403e-01 -1.52772093e+00 5.59000596e-02
-5.80217898e-01 1.49706230e-01 5.62233865e-01 6.03641629e-01
-1.13000834e+00 1.34471703e+00 -9.02834475e-01 -4.11581844e-01
1.18715495e-01 6.94171429e-01 -6.73598528e-01 5.03441453e-01
-1.00838184e+00 1.25579047e+00 6.35188103e-01 3.03335935e-01
-2.95498282e-01 -3.58711839e-01 -7.35360146e-01 -1.32821694e-01
2.20139742e-01 -7.54906893e-01 9.52664971e-01 -1.11614251e+00
-1.61980987e+00 8.34009647e-01 5.61582148e-02 -6.01485312e-01
5.53622007e-01 -9.92693529e-02 -1.26734734e-01 7.82693848e-02
-3.58629167e-01 3.31304163e-01 1.11426497e+00 -8.06502998e-01
-3.67041081e-01 -3.29271913e-01 3.51806462e-01 1.50817245e-01
-4.13212806e-01 -1.45528674e-01 1.30885169e-01 -5.62318623e-01
3.37421656e-01 -1.16523194e+00 -1.93456948e-01 -1.22303143e-01
-6.36976004e-01 -3.87971401e-01 8.01524341e-01 -3.89040649e-01
1.02321661e+00 -1.94502032e+00 8.18544745e-01 3.39327395e-01
3.11787039e-01 2.74795622e-01 1.09030716e-01 5.49672067e-01
-4.04463649e-01 1.42405331e-01 -3.26536685e-01 -3.82140368e-01
7.61151090e-02 3.01484555e-01 -3.34054142e-01 1.01384246e+00
5.38965464e-01 6.17115200e-01 -6.62388265e-01 -1.45974353e-01
1.33199021e-01 5.51008105e-01 -3.65398198e-01 -3.66321653e-02
-8.23178291e-02 4.56938833e-01 -4.48091149e-01 1.37726858e-01
3.76520336e-01 -2.20062286e-01 -2.48056725e-01 -1.75051630e-01
-4.19117272e-01 4.40642118e-01 -1.50115383e+00 1.34639859e+00
-2.28447795e-01 7.64737189e-01 3.12742889e-01 -1.44197428e+00
7.74888337e-01 3.00925374e-01 5.57108819e-01 -5.62035479e-02
4.13573802e-01 3.23430926e-01 1.21175282e-01 -3.34996879e-01
4.72313941e-01 -3.96882683e-01 3.26790422e-01 4.86894071e-01
2.09304690e-01 1.70888543e-01 2.36646116e-01 -3.18823420e-02
8.23181272e-01 9.94286761e-02 1.85490847e-01 -6.60153985e-01
8.32422972e-01 1.79642849e-02 2.18347192e-01 3.08076411e-01
9.03643444e-02 6.60121918e-01 4.93513376e-01 -6.65664375e-01
-1.20044875e+00 -9.73616183e-01 -5.19403040e-01 7.06217527e-01
-1.34298027e-01 -3.70935529e-01 -9.07683849e-01 -3.62807423e-01
-2.83087015e-01 3.03111047e-01 -6.11954749e-01 -3.18783015e-01
-7.71723211e-01 -6.82671189e-01 3.14487666e-01 1.34026304e-01
3.37559044e-01 -8.15535843e-01 -8.40500951e-01 1.92131668e-01
4.40816045e-01 -9.98958409e-01 -2.74924010e-01 2.51998454e-01
-1.28046608e+00 -1.20346737e+00 -5.15786052e-01 -5.44346869e-01
6.32642806e-01 -4.92552249e-03 8.71087193e-01 2.39859700e-01
-2.27554485e-01 4.88523066e-01 4.30891812e-02 -3.95348430e-01
-5.86579800e-01 3.10196519e-01 5.05068183e-01 4.03299570e-01
-1.01279775e-02 -8.86323214e-01 -4.58767325e-01 4.13174897e-01
-9.88312006e-01 -1.22657612e-01 2.99562126e-01 7.60868907e-01
1.98954895e-01 -3.03493184e-03 3.12150121e-01 -4.04247820e-01
3.43076378e-01 -3.82180154e-01 -7.71681845e-01 -8.40805322e-02
-5.26573598e-01 5.56751132e-01 8.99140537e-01 -4.13840562e-01
-3.43441427e-01 -1.37977144e-02 -2.66447842e-01 -2.62297183e-01
1.86166186e-02 4.48930293e-01 2.01399364e-02 -4.76031005e-01
6.57251537e-01 4.04827250e-03 3.37834418e-01 -5.84611535e-01
2.29685500e-01 2.87465245e-01 4.13498461e-01 -4.69781339e-01
1.10133505e+00 5.32572031e-01 5.85873246e-01 -1.32921541e+00
-6.57536447e-01 -3.15045923e-01 -9.74838734e-01 -2.94550769e-02
9.19494867e-01 -3.67951632e-01 -7.65189886e-01 2.08024457e-01
-1.22265041e+00 -1.67898372e-01 -4.12459522e-01 5.65369308e-01
-6.79191351e-01 2.71995395e-01 -4.55964327e-01 -7.54944980e-01
-2.66880542e-01 -1.10686266e+00 5.90364277e-01 2.04097107e-01
-2.39682078e-01 -1.57222605e+00 1.05168402e-01 -9.31539312e-02
4.57952589e-01 4.13018554e-01 1.01247048e+00 -7.00208247e-01
-4.80481446e-01 -3.41095895e-01 1.89528137e-01 5.33960879e-01
-8.44645053e-02 3.53324503e-01 -7.57979631e-01 -2.97548771e-01
4.97722536e-01 4.39906344e-02 6.36818290e-01 1.08414322e-01
6.10222995e-01 -3.90074402e-01 -7.33910054e-02 7.40931690e-01
1.32832098e+00 -1.02552839e-01 3.47311854e-01 6.14407659e-01
7.34415829e-01 7.92236328e-01 3.55609059e-02 -9.54678096e-03
2.58102655e-01 8.89258206e-01 7.03124166e-01 2.61686035e-02
4.27549362e-01 1.23107001e-01 4.88922834e-01 9.79343534e-01
-4.43129510e-01 4.01368260e-01 -7.65889585e-01 2.87805945e-01
-1.75751126e+00 -8.15047264e-01 -3.35691988e-01 2.44998550e+00
5.57248712e-01 1.08426556e-01 2.53954172e-01 3.94009918e-01
3.89764160e-01 8.85201618e-02 -4.14707482e-01 -5.51090419e-01
-2.45475799e-01 1.61645293e-01 4.29807246e-01 6.82135940e-01
-1.22874629e+00 7.12940931e-01 6.34628963e+00 4.32583958e-01
-1.32127976e+00 -1.95121784e-02 2.90763080e-01 1.84014291e-01
1.53091162e-01 2.10774243e-01 -8.21409523e-01 7.64518380e-02
9.59802151e-01 -6.89848959e-02 5.99627614e-01 7.66240418e-01
1.05424993e-01 7.89630935e-02 -1.26457882e+00 9.09125507e-01
-9.76111963e-02 -1.47131860e+00 -2.49361515e-01 3.92846376e-01
4.57002580e-01 1.41750559e-01 1.76943660e-01 1.76814534e-02
-1.09685205e-01 -1.17958558e+00 6.83866739e-01 4.24737811e-01
2.27314085e-01 -6.17167354e-01 4.37000304e-01 4.65871423e-01
-8.28071654e-01 4.36603799e-02 -5.13259053e-01 -5.21491170e-02
1.04876198e-01 3.12342674e-01 -6.32279813e-01 4.40782189e-01
2.93927491e-01 7.32333541e-01 -4.28712010e-01 9.94323969e-01
-1.46151930e-02 4.82153952e-01 -6.52774990e-01 -9.05231610e-02
6.87166393e-01 -8.47323895e-01 1.07199228e+00 9.89598274e-01
3.71034026e-01 -1.44307569e-01 -2.91938066e-01 9.18403387e-01
7.16505870e-02 3.00942540e-01 -7.42073357e-01 -1.34690315e-01
-2.45907277e-01 1.58780861e+00 -6.15887523e-01 2.37525757e-02
-1.94303110e-01 7.24226356e-01 4.12152171e-01 3.68492901e-01
-5.89503050e-01 -2.76320577e-01 1.02636063e+00 4.23832610e-02
3.18959028e-01 -5.48019230e-01 -2.31885210e-01 -1.30524445e+00
3.12398463e-01 -7.29234576e-01 1.33070216e-01 -3.28394234e-01
-9.69495893e-01 6.14044189e-01 2.74569988e-02 -1.00613129e+00
-4.55412716e-01 -1.06899071e+00 -1.04281843e+00 9.68773782e-01
-1.15547347e+00 -5.55839896e-01 5.75696863e-02 5.84299028e-01
2.24106442e-02 -4.05645311e-01 8.44378293e-01 -1.76841468e-02
-7.05670536e-01 2.46327296e-01 2.79933244e-01 -2.62697358e-02
5.59096523e-02 -1.65377104e+00 2.87668943e-01 7.60660708e-01
4.27489191e-01 8.88430893e-01 1.04146481e+00 -2.94169667e-03
-1.79596794e+00 -7.97820568e-01 7.40167379e-01 -5.51004291e-01
1.10364199e+00 -4.23330873e-01 -1.04462063e+00 6.13224328e-01
1.35284513e-01 -1.87111035e-01 4.06073451e-01 1.19249687e-01
-1.37441710e-01 -2.10476860e-01 -7.39217579e-01 8.83769274e-01
6.95506155e-01 -3.51173550e-01 -5.30539691e-01 4.43631709e-01
5.19063532e-01 -3.63359153e-01 -9.22096550e-01 6.51789596e-03
3.88553113e-01 -1.18151510e+00 1.00202489e+00 -9.72403944e-01
2.08861187e-01 -2.55966783e-01 -1.12676263e-01 -1.37498546e+00
4.99965437e-02 -1.16494370e+00 -2.30765134e-01 9.19912100e-01
3.24744672e-01 -1.02387214e+00 7.21377313e-01 4.55964118e-01
-2.87720501e-01 -1.15182829e+00 -1.41970503e+00 -7.91816711e-01
3.71134818e-01 -2.31113851e-01 2.79918522e-01 7.95794368e-01
2.88967937e-02 5.79298675e-01 -8.00879002e-02 7.14671835e-02
6.05466783e-01 -9.36006829e-02 6.86836660e-01 -1.36169398e+00
-4.64585334e-01 -8.99572313e-01 -7.56233215e-01 -9.49400902e-01
1.89091727e-01 -1.08107448e+00 -4.03004766e-01 -9.65553582e-01
-5.19136488e-01 -5.12555540e-01 -7.10418299e-02 1.82001993e-01
2.10871845e-01 4.24075611e-02 2.02398062e-01 3.08630794e-01
-7.56625682e-02 5.91402113e-01 1.03355241e+00 1.68696120e-01
-3.52108508e-01 2.68529922e-01 -4.48904663e-01 7.56412387e-01
8.96466136e-01 -3.06320637e-01 -1.78056300e-01 -2.20968932e-01
4.55806851e-01 -2.95235902e-01 6.08367801e-01 -1.04495013e+00
1.53235346e-01 9.17433053e-02 1.42658753e-02 -1.08045392e-01
5.49622893e-01 -8.90972555e-01 9.53224078e-02 3.82837653e-01
-8.70763697e-03 3.93966556e-01 -4.38891798e-02 3.19407165e-01
-1.16330743e-01 -6.25764966e-01 8.90479565e-01 7.37146586e-02
-2.09000692e-01 3.66310090e-01 -4.41727340e-02 3.42952535e-02
9.32528019e-01 -3.05911899e-01 -4.01935093e-02 -1.47476763e-01
-7.75772274e-01 -9.89246070e-02 5.30609787e-01 2.42509514e-01
3.23810726e-01 -1.34754944e+00 -4.91689771e-01 1.77784741e-01
-2.76127130e-01 6.76118508e-02 -1.71987101e-01 1.41101825e+00
-7.20895886e-01 4.93465990e-01 -1.17516041e-01 -9.00016904e-01
-1.10266638e+00 4.14213240e-01 8.62483919e-01 -1.46816924e-01
-9.38179314e-01 4.62854415e-01 4.42799646e-03 -2.24210396e-01
1.71569988e-01 -4.21774656e-01 -2.00247794e-01 1.44921299e-02
4.88182902e-01 2.64756113e-01 1.74010888e-01 -1.06912160e+00
-1.36110976e-01 6.81342900e-01 1.72587752e-01 -8.53026286e-02
1.66443717e+00 1.38628647e-01 -3.70263547e-01 6.87961757e-01
1.41324914e+00 -2.90722400e-01 -1.24186540e+00 -1.08988248e-01
1.60593212e-01 1.50611820e-02 -6.98863119e-02 1.34558734e-02
-9.52919066e-01 1.18674564e+00 3.29121560e-01 6.44554436e-01
7.21443236e-01 -1.20321147e-01 5.41653574e-01 6.28040493e-01
2.72088915e-01 -9.68879342e-01 -8.84866640e-02 7.47069418e-01
1.01660454e+00 -1.10494053e+00 -1.13196708e-01 -8.84928480e-02
-2.90475965e-01 1.64899075e+00 1.40008330e-02 -6.33664250e-01
6.52021170e-01 -2.12075293e-01 -8.85614157e-02 -2.25126654e-01
-4.76808369e-01 2.21510809e-02 6.36924744e-01 3.56787950e-01
3.98127854e-01 -1.63334161e-01 -1.98411196e-01 1.24335781e-01
-5.21318316e-01 -6.34520292e-01 6.36267602e-01 6.11613870e-01
-5.10585010e-01 -1.08127260e+00 -4.12948012e-01 1.79767624e-01
-4.65544879e-01 1.79032296e-01 -3.00446957e-01 1.00157475e+00
-1.05129652e-01 7.02594399e-01 -1.67082846e-01 -3.61054331e-01
3.22915077e-01 2.60515749e-01 5.17581344e-01 -3.31595600e-01
-6.30245686e-01 -4.10864085e-01 -1.21930949e-01 -5.45041680e-01
-3.11423242e-01 -7.19866991e-01 -1.29740739e+00 -3.07922959e-01
-1.66320115e-01 3.68028760e-01 1.09361982e+00 1.36875463e+00
9.87009183e-02 1.26883626e-01 5.99780858e-01 -1.20827436e+00
-9.87276316e-01 -7.28666604e-01 -6.52396858e-01 3.05287212e-01
7.73913085e-01 -7.88913250e-01 -5.21606147e-01 -1.40464902e-01] | [7.719864845275879, 3.812159538269043] |
fbf2d7d6-077b-4466-bbb4-a2cdee173d17 | is-your-code-generated-by-chatgpt-really | 2305.0121 | null | https://arxiv.org/abs/2305.01210v2 | https://arxiv.org/pdf/2305.01210v2.pdf | Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generation | Program synthesis has been long studied with recent approaches focused on directly using the power of Large Language Models (LLMs) to generate code. Programming benchmarks, with curated synthesis problems and test-cases, are used to measure the performance of various LLMs on code synthesis. However, these test-cases can be limited in both quantity and quality for fully assessing the functional correctness of the generated code. Such limitation in the existing benchmarks begs the following question: In the era of LLMs, is the code generated really correct? To answer this, we propose EvalPlus -- a code synthesis benchmarking framework to rigorously evaluate the functional correctness of LLM-synthesized code. EvalPlus augments a given evaluation dataset with large amounts of test-cases newly produced by an automatic test input generator, powered by both LLM- and mutation-based strategies. While EvalPlus is general, we extend the test-cases of the popular HUMANEVAL benchmark by 81x to build HUMANEVAL+. Our extensive evaluation across 19 popular LLMs (e.g., GPT-4 and ChatGPT) demonstrates that HUMANEVAL+ is able to catch significant amounts of previously undetected wrong code synthesized by LLMs, reducing the pass@k by 13.6-15.3% on average. Our work not only indicates that prior popular code synthesis evaluation results do not accurately reflect the true performance of LLMs for code synthesis, but also opens up a new direction to improve such programming benchmarks through automated testing. | ['Lingming Zhang', 'Yuyao Wang', 'Chunqiu Steven Xia', 'Jiawei Liu'] | 2023-05-02 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 1.28083527e-01 1.55557036e-01 -2.97283232e-01 -1.82288080e-01
-9.97056425e-01 -9.04624224e-01 4.43280011e-01 2.45410368e-01
1.94316015e-01 6.17231071e-01 -1.31886348e-01 -8.91298354e-01
2.93208569e-01 -1.15667403e+00 -1.10003686e+00 -3.98793183e-02
1.31251231e-01 1.84513003e-01 4.65167165e-01 -4.29179311e-01
6.10495806e-01 -1.06745780e-01 -1.77435863e+00 6.68311656e-01
1.38903713e+00 3.24112892e-01 3.42311040e-02 1.00496101e+00
-3.71017307e-01 6.30410254e-01 -9.34675336e-01 -5.93611658e-01
1.31485566e-01 -6.86029196e-01 -7.55811930e-01 -4.26858962e-01
2.62095869e-01 -5.97661510e-02 3.71597975e-01 1.18394470e+00
3.55419159e-01 -5.26323676e-01 2.17820033e-01 -1.53336465e+00
-4.11279410e-01 1.22240067e+00 -4.71607447e-01 -2.58317590e-01
5.54115415e-01 6.70067132e-01 9.77497220e-01 -8.02932441e-01
6.53254628e-01 1.28674531e+00 7.21980035e-01 5.67410290e-01
-1.46219587e+00 -5.72494507e-01 -4.94728982e-01 -5.03151476e-01
-1.41378200e+00 -2.46382684e-01 1.38166502e-01 -7.84101784e-01
1.35812652e+00 6.21273756e-01 5.16612887e-01 1.05937195e+00
4.67761546e-01 5.47948778e-01 1.04644656e+00 -6.63397610e-01
3.18416387e-01 3.07371199e-01 1.55928820e-01 1.02016711e+00
6.64985120e-01 3.23877633e-01 -1.74613670e-01 -5.84201097e-01
2.48498172e-01 -6.48255765e-01 -1.45039529e-01 -2.13503063e-01
-1.36575866e+00 9.12811995e-01 4.92305681e-02 2.47939125e-01
3.45060319e-01 5.73117256e-01 6.34644568e-01 4.74443108e-01
-6.49926066e-02 1.08644438e+00 -6.05194032e-01 -6.49239898e-01
-1.20428061e+00 6.08736932e-01 1.12927639e+00 1.36173844e+00
7.89720893e-01 4.19632554e-01 -5.38476110e-01 5.16734660e-01
2.91429162e-01 6.74673676e-01 4.89100963e-01 -6.49760604e-01
7.32429683e-01 1.19022799e+00 -1.74466759e-01 -7.66262829e-01
-9.37729180e-02 -4.61078882e-01 -1.37806609e-01 5.41436791e-01
2.41025358e-01 -3.13708723e-01 -5.91821074e-01 1.57759178e+00
-1.02543138e-01 -3.76742691e-01 1.29757002e-01 2.82505244e-01
6.38465881e-01 7.13964820e-01 -4.36702341e-01 1.39088467e-01
7.96209633e-01 -1.04027450e+00 -7.03711212e-02 -2.36818582e-01
1.28801978e+00 -8.77871811e-01 1.54434943e+00 3.62058133e-01
-1.14696944e+00 -5.41672051e-01 -1.29348123e+00 4.34464186e-01
-2.34921739e-01 7.66682401e-02 5.42406619e-01 1.09481478e+00
-1.25856042e+00 5.36239028e-01 -6.65934980e-01 -5.98691963e-02
2.44698614e-01 2.93802023e-01 8.58597923e-03 -1.17113315e-01
-5.65782249e-01 5.82962513e-01 4.40985680e-01 -3.93986523e-01
-1.15442157e+00 -1.08013511e+00 -8.68015230e-01 3.64798540e-03
3.66270423e-01 -4.67340261e-01 1.44634044e+00 -8.51278722e-01
-1.23558545e+00 6.01444781e-01 1.80614144e-02 -3.29479069e-01
6.64201438e-01 7.91372433e-02 -4.62131947e-01 -5.97450435e-01
2.71377534e-01 5.19022882e-01 4.22813803e-01 -1.33298540e+00
-4.83252376e-01 4.20178205e-01 1.22823700e-01 -6.89010382e-01
5.35257943e-02 1.73571184e-01 -2.38742277e-01 -5.88566065e-01
-6.06819749e-01 -1.03146148e+00 -9.24286991e-02 -3.80915105e-01
-5.72909653e-01 2.07523242e-01 4.89055425e-01 -3.07797760e-01
1.55550134e+00 -2.19000125e+00 -4.82814945e-02 4.14455682e-01
6.68745786e-02 4.75184143e-01 -4.88837510e-01 6.59988403e-01
-1.15112476e-01 7.94622064e-01 -6.44626856e-01 3.54869664e-01
3.73317420e-01 5.54759093e-02 -4.28904831e-01 1.07871490e-02
4.76901859e-01 1.29808664e+00 -1.00905418e+00 -2.99582064e-01
-8.24303105e-02 6.49689287e-02 -1.22406900e+00 8.86451080e-02
-7.87387133e-01 1.06389575e-01 -3.17449063e-01 8.86345208e-01
3.39365929e-01 -1.58506930e-01 -6.11479357e-02 4.20834720e-01
-3.17536891e-01 2.59760112e-01 -9.93581593e-01 1.54648495e+00
-6.38765514e-01 7.26223528e-01 -4.55231905e-01 -4.01967347e-01
1.09561551e+00 7.77712986e-02 -2.55550176e-01 -7.54428387e-01
-1.50411606e-01 8.00031066e-01 4.65409666e-01 -6.59776092e-01
4.90472913e-01 3.90941322e-01 -5.78354657e-01 5.64356744e-01
-1.30477667e-01 -7.10204720e-01 7.92476416e-01 -5.15837781e-02
1.62979352e+00 4.08120781e-01 2.36469120e-01 -3.83707404e-01
6.62902415e-01 4.99093384e-01 5.45086622e-01 9.80213284e-01
3.00341576e-01 6.20954156e-01 1.08014011e+00 -3.77111956e-02
-1.30606699e+00 -7.90877581e-01 -1.34173304e-01 8.56159925e-01
-3.77975672e-01 -1.02642071e+00 -1.13641298e+00 -9.16367173e-01
3.54693644e-02 1.23623562e+00 -4.33566153e-01 -4.61173773e-01
-6.85462534e-01 -9.52505529e-01 1.08808482e+00 3.26017588e-01
3.52442414e-01 -1.12772465e+00 -7.64645755e-01 3.41868728e-01
9.94515046e-02 -6.27919078e-01 -3.24757129e-01 2.14480571e-02
-6.51043713e-01 -1.23528838e+00 -2.48748899e-01 -5.06332934e-01
7.77234077e-01 -4.19813335e-01 1.73031652e+00 6.79596186e-01
-4.43432987e-01 -2.58766767e-02 -6.07710719e-01 -2.48458296e-01
-1.60202432e+00 2.90552199e-01 -5.85039020e-01 -5.72636008e-01
3.93839702e-02 -4.51192886e-01 -9.90341082e-02 5.01772523e-01
-1.04506183e+00 5.03594354e-02 7.83288419e-01 8.30862343e-01
2.08335131e-01 -6.63414374e-02 6.05920672e-01 -1.08238840e+00
7.06880450e-01 -3.63573581e-01 -1.15417957e+00 4.46058065e-01
-6.89455092e-01 4.60891426e-01 6.87153101e-01 -3.33300024e-01
-7.05439746e-01 -3.52010757e-01 -5.34939766e-01 3.06807667e-01
2.32162401e-01 8.09638798e-01 -1.38810784e-01 -3.29048604e-01
1.15030754e+00 2.95230806e-01 -2.94504672e-01 -2.94465311e-02
6.04993105e-02 3.48685533e-01 3.60317200e-01 -1.10675955e+00
9.60150480e-01 -3.05212826e-01 -1.87502891e-01 -2.24497333e-01
-4.39631790e-02 4.04632121e-01 -4.60752062e-02 9.39858332e-02
3.68108779e-01 -5.33356130e-01 -4.50010568e-01 2.32103869e-01
-1.07204354e+00 -8.24698985e-01 -4.30620372e-01 -7.08693191e-02
-4.68311608e-01 1.59112036e-01 -3.76729310e-01 -5.08041561e-01
-2.95578510e-01 -2.05466032e+00 1.19930947e+00 -2.00273648e-01
-7.25437999e-01 -5.90274930e-01 2.24862278e-01 -6.88479934e-03
6.76177800e-01 4.68385875e-01 1.57030642e+00 -4.70463246e-01
-6.55337870e-01 -2.87742198e-01 -3.79892066e-02 4.45739746e-01
-1.52420521e-01 6.00963354e-01 -5.75447798e-01 -2.50423461e-01
-3.87974381e-01 -1.95626318e-01 2.88743258e-01 -1.35120824e-01
8.47961545e-01 -3.54052812e-01 -2.44916990e-01 5.26636481e-01
1.65475857e+00 9.03208479e-02 7.89016843e-01 3.59159112e-01
4.02438492e-01 2.15733096e-01 5.90299726e-01 4.22723264e-01
7.52404844e-03 5.34243226e-01 5.00683248e-01 2.66710937e-01
-2.67120361e-01 -3.72064084e-01 9.26018476e-01 7.41357863e-01
3.56292039e-01 -3.39534551e-01 -1.42671406e+00 5.38332164e-01
-1.54857659e+00 -5.66270411e-01 -6.52916729e-01 2.26998711e+00
1.24351597e+00 4.34405327e-01 -2.28961781e-02 3.59324902e-01
3.94033134e-01 -4.08534408e-01 -2.07317352e-01 -8.06932449e-01
3.17061767e-02 4.82515156e-01 2.61608064e-01 3.15148115e-01
-3.77563357e-01 7.94204831e-01 6.24455452e+00 8.81852388e-01
-1.16474223e+00 -1.99943017e-02 2.80545712e-01 9.91983265e-02
-7.54010797e-01 4.18879718e-01 -9.61877227e-01 6.10733092e-01
1.34394574e+00 -6.54325187e-01 4.08975482e-01 1.06876349e+00
9.18479040e-02 -2.61513650e-01 -1.41063666e+00 4.10064936e-01
-8.72463137e-02 -1.53059995e+00 1.36737078e-02 4.83479239e-02
1.26417625e+00 -3.11105847e-02 1.07562440e-02 8.35973382e-01
6.22542322e-01 -1.14809215e+00 1.16036201e+00 3.84666950e-01
9.07625675e-01 -8.31909955e-01 8.19898903e-01 3.44590962e-01
-8.31283152e-01 -4.35554720e-02 -1.09980077e-01 6.89944625e-02
-2.04650924e-01 7.94356585e-01 -1.20925701e+00 4.05956656e-01
5.10942042e-01 2.07743287e-01 -1.33596420e+00 1.28606665e+00
-3.92085433e-01 8.66734087e-01 3.75063941e-02 -2.43611947e-01
1.55035555e-01 2.93030947e-01 4.09016997e-01 1.42967582e+00
6.80633366e-01 -6.97291017e-01 3.31954025e-02 1.63892651e+00
-5.55121563e-02 4.76459414e-02 -5.95730424e-01 -4.51701760e-01
3.16291928e-01 1.05048442e+00 -6.22992218e-01 -4.49182987e-01
-4.73850995e-01 4.64574397e-01 1.41739827e-02 3.73844914e-02
-1.05847669e+00 -5.92981100e-01 3.95404458e-01 2.31865525e-01
1.45571381e-01 -2.09249452e-01 -4.23334688e-01 -1.00113702e+00
2.74254680e-01 -1.62575281e+00 -1.62125185e-01 -6.77074969e-01
-6.37616217e-01 8.36783588e-01 8.07095170e-02 -1.05539405e+00
-6.22939467e-01 -4.06922251e-01 -7.23728597e-01 8.44096780e-01
-1.10489786e+00 -7.37801373e-01 -2.67640114e-01 -1.23896785e-01
4.09991503e-01 -2.39305601e-01 7.85919070e-01 2.92604089e-01
-5.21277487e-01 9.93150592e-01 -7.30460584e-02 -4.56641503e-02
4.99691308e-01 -1.22790945e+00 9.94795561e-01 1.12735176e+00
-2.51135379e-01 7.78466940e-01 8.76133382e-01 -7.93423235e-01
-1.68187571e+00 -1.38065743e+00 7.84426212e-01 -7.43335605e-01
7.27970541e-01 -6.61921501e-01 -8.60376954e-01 4.53934908e-01
1.31439082e-02 1.94794014e-02 2.62294859e-01 -3.88372421e-01
-4.41243798e-01 5.21474443e-02 -1.11924636e+00 6.79657459e-01
8.99631381e-01 -2.83050001e-01 -3.68894011e-01 2.19601870e-01
9.95257497e-01 -5.24948299e-01 -9.19001997e-01 4.98640776e-01
3.48058432e-01 -1.18377769e+00 5.08803904e-01 -1.96235999e-01
1.02315652e+00 -5.88820100e-01 -3.26848745e-01 -1.28528666e+00
3.72218698e-01 -6.22405469e-01 1.89557239e-01 1.64321721e+00
9.50980842e-01 -6.68592274e-01 4.80927944e-01 5.44518411e-01
-3.87477279e-01 -6.40977681e-01 -2.47725666e-01 -9.94099319e-01
3.17446589e-01 -7.71875679e-01 1.09925270e+00 6.80858314e-01
3.07700951e-02 -3.83509956e-02 1.78265691e-01 -1.96345448e-01
1.56641006e-01 2.18343914e-01 1.15478837e+00 -8.28383386e-01
-7.65382886e-01 -7.80701399e-01 -3.41429949e-01 -3.17792624e-01
2.18435362e-01 -1.32527399e+00 1.56251773e-01 -1.07292676e+00
1.94790646e-01 -4.99141306e-01 4.81559157e-01 7.22792387e-01
-6.40317723e-02 1.19068764e-01 -2.33042371e-02 -2.90845692e-01
-2.05386043e-01 1.80466965e-01 9.42772090e-01 -2.57544070e-01
-2.35582381e-01 -1.04580641e-01 -7.17352092e-01 4.59876209e-01
6.34935081e-01 -7.22447753e-01 -2.90939391e-01 -4.00464743e-01
1.12315261e+00 1.97399080e-01 4.39517587e-01 -1.33987343e+00
-2.08856240e-01 -1.28564015e-01 -2.31518686e-01 -3.37451577e-01
-5.97283125e-01 -4.41573709e-01 4.82513130e-01 8.04676473e-01
-4.42027986e-01 5.07200956e-01 5.66536546e-01 6.63113892e-02
-3.26112241e-01 -7.11992025e-01 5.17628849e-01 -2.23010838e-01
-7.43932188e-01 -1.62735447e-01 -3.43345135e-01 4.81890589e-01
9.51470733e-01 -8.86148885e-02 -5.81130564e-01 3.23709428e-01
-1.20603032e-01 7.46017229e-03 8.89622867e-01 5.93300223e-01
3.16581666e-01 -1.09820354e+00 -9.03510928e-01 5.18185675e-01
6.01704061e-01 -2.80627161e-01 -3.88586402e-01 6.49453878e-01
-8.53940070e-01 2.84067959e-01 -4.65445220e-02 -6.97195530e-01
-1.04064023e+00 6.34287834e-01 2.21524894e-01 -2.52627730e-01
-3.31558615e-01 6.33325458e-01 -4.15390544e-02 -8.42981517e-01
-3.61718208e-01 -7.88920224e-01 5.75905263e-01 -5.37517488e-01
3.20408970e-01 1.46507964e-01 4.10196811e-01 -9.52376351e-02
-3.31925958e-01 1.64648712e-01 2.78429151e-01 -1.82792336e-01
1.31090105e+00 7.02668011e-01 -5.31420529e-01 3.02992284e-01
1.29108512e+00 4.55738515e-01 -6.59183204e-01 3.10367376e-01
3.68079841e-01 -4.26693499e-01 -3.66992265e-01 -1.06626046e+00
-8.54110122e-01 7.22480595e-01 1.64050370e-01 1.16202071e-01
7.83912718e-01 -3.56907785e-01 3.49905372e-01 2.98675328e-01
8.04107547e-01 -5.41638792e-01 9.09698755e-03 6.08294010e-01
9.11805570e-01 -1.05872631e+00 -3.63589227e-01 -3.11886668e-01
-1.40960872e-01 1.16571677e+00 7.79696763e-01 -9.12689418e-03
7.24260136e-03 8.33337784e-01 -4.57153082e-01 4.76609655e-02
-8.69641960e-01 2.71920592e-01 1.54438406e-01 2.84021586e-01
7.13664889e-01 -4.89597917e-02 -4.33994234e-01 5.52194238e-01
-6.54923558e-01 1.98490068e-01 9.87202406e-01 1.28334463e+00
-3.05888414e-01 -1.60448778e+00 -6.09199822e-01 5.60271204e-01
-2.33640671e-01 -4.17123467e-01 -3.50283831e-01 1.03649509e+00
2.27592915e-01 9.45477962e-01 -4.46643382e-01 -4.97330815e-01
4.13506359e-01 5.77670746e-02 4.51661944e-01 -9.80595171e-01
-1.07294381e+00 -4.45158809e-01 4.22327757e-01 -5.58136106e-01
2.00482607e-01 -5.30014694e-01 -1.18560946e+00 -4.14587349e-01
-2.64033884e-01 2.35546708e-01 5.92115939e-01 6.39852047e-01
4.85884041e-01 9.07304823e-01 2.72885293e-01 -3.58920723e-01
-5.28123498e-01 -5.49881697e-01 1.03717409e-01 -4.46396619e-02
1.75812140e-01 -3.14742804e-01 -1.99080631e-01 1.90570299e-02] | [7.79813814163208, 7.636441707611084] |
6fb38d9d-19d9-4c63-8d3b-274ba24531b5 | cross-modal-contrastive-learning-for-1 | 2302.14057 | null | https://arxiv.org/abs/2302.14057v1 | https://arxiv.org/pdf/2302.14057v1.pdf | Cross-modal Contrastive Learning for Multimodal Fake News Detection | Automatic detection of multimodal fake news has gained a widespread attention recently. Many existing approaches seek to fuse unimodal features to produce multimodal news representations. However, the potential of powerful cross-modal contrastive learning methods for fake news detection has not been well exploited. Besides, how to aggregate features from different modalities to boost the performance of the decision-making process is still an open question. To address that, we propose COOLANT, a cross-modal contrastive learning framework for multimodal fake news detection, aiming to achieve more accurate image-text alignment. To further improve the alignment precision, we leverage an auxiliary task to soften the loss term of negative samples during the contrast process. A cross-modal fusion module is developed to learn the cross-modality correlations. An attention mechanism with an attention guidance module is implemented to help effectively and interpretably aggregate the aligned unimodal representations and the cross-modality correlations. Finally, we evaluate the COOLANT and conduct a comparative study on two widely used datasets, Twitter and Weibo. The experimental results demonstrate that our COOLANT outperforms previous approaches by a large margin and achieves new state-of-the-art results on the two datasets. | ['Siqi Wang', 'Xiaohan Xu', 'Shuai Zhang', 'Hongbo Xu', 'Chuang Zhang', 'Longzheng Wang'] | 2023-02-25 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 4.89407107e-02 -3.46579701e-01 -2.58207321e-01 -4.27770704e-01
-1.34101367e+00 -3.23510557e-01 1.12963939e+00 1.76517442e-01
-2.80644089e-01 2.64504254e-01 4.00640070e-01 1.10540632e-02
2.74943709e-01 -3.13038230e-01 -5.90118289e-01 -7.26618528e-01
3.84106755e-01 1.21371582e-01 -4.25158776e-02 -3.64304394e-01
4.54608232e-01 -5.69403917e-02 -1.28009748e+00 9.27294016e-01
1.12206137e+00 1.10254037e+00 -3.72444570e-01 3.03342372e-01
6.74742833e-02 9.32715237e-01 -5.42438567e-01 -1.03392363e+00
1.06724240e-02 -5.83272099e-01 -5.44901192e-01 9.32110175e-02
5.37646055e-01 -2.29889810e-01 -4.30304855e-01 1.27868986e+00
4.83910918e-01 -1.55752391e-01 8.91096175e-01 -1.41775119e+00
-8.00135851e-01 5.94770849e-01 -1.07895184e+00 2.97511548e-01
3.00119400e-01 3.38160038e-01 1.09025419e+00 -1.16529679e+00
3.81413102e-01 1.46131623e+00 3.99548739e-01 1.99872673e-01
-1.09316397e+00 -7.86544740e-01 6.91124573e-02 4.44990605e-01
-1.25946689e+00 -4.40285891e-01 1.04272318e+00 -4.05397922e-01
2.82313913e-01 2.89632171e-01 4.66584951e-01 1.51372349e+00
1.82162419e-01 1.24956512e+00 1.33182359e+00 -3.61863792e-01
-2.80815482e-01 3.21672499e-01 2.76025683e-02 8.12551558e-01
2.83963941e-02 -1.47879124e-01 -8.41815114e-01 -2.44515643e-01
2.42103934e-01 2.71187499e-02 -2.76854366e-01 -2.88680624e-02
-1.35068417e+00 1.09074628e+00 7.56107748e-01 4.86560762e-01
-2.90636331e-01 6.14658184e-02 6.36517882e-01 2.72293568e-01
7.95129716e-01 5.64116418e-01 1.06239416e-01 1.10381491e-01
-7.72302210e-01 9.05233100e-02 2.91398138e-01 4.18521464e-01
6.66892886e-01 -2.22078234e-01 -5.24021745e-01 9.93129611e-01
3.76871705e-01 6.81567073e-01 5.01072466e-01 -4.95663613e-01
9.76935744e-01 7.62270033e-01 9.98816863e-02 -1.60465991e+00
-1.22709543e-01 -3.64688575e-01 -8.93142998e-01 -1.89961001e-01
2.79420555e-01 1.24490097e-01 -6.27165437e-01 1.42902946e+00
2.52848417e-01 2.21339658e-01 2.76682395e-02 1.33181167e+00
8.27242017e-01 7.23921835e-01 8.40376988e-02 -2.03160010e-02
1.61162031e+00 -1.22823060e+00 -8.33394349e-01 -4.56720650e-01
5.31610131e-01 -1.06461298e+00 1.14700782e+00 1.51378214e-01
-8.03754866e-01 -1.54360205e-01 -1.07871997e+00 -2.72239774e-01
-3.38297099e-01 3.42148274e-01 4.24981236e-01 3.77461642e-01
-2.09968820e-01 1.88701883e-01 -5.40035963e-01 4.61075604e-02
6.51299000e-01 -1.24365337e-01 -4.47178543e-01 -2.34323218e-01
-1.46366990e+00 1.03615141e+00 3.41004312e-01 1.48156941e-01
-6.50725961e-01 -2.96268612e-01 -8.22420776e-01 -1.15703583e-01
3.32140923e-01 -7.25981712e-01 9.93902981e-01 -1.39943016e+00
-1.34318304e+00 9.91191089e-01 7.52686188e-02 -2.21564382e-01
7.34105706e-01 -3.16825211e-01 -4.51364905e-01 3.79030108e-01
2.25143656e-01 5.28508544e-01 1.30820525e+00 -1.58978117e+00
-5.30362666e-01 -4.30445403e-01 -1.53691411e-01 5.06494343e-01
-5.52029073e-01 -1.24337133e-02 -6.25307322e-01 -9.66769457e-01
6.09583892e-02 -9.27852213e-01 2.68928558e-01 -2.08684117e-01
-6.23063028e-01 -1.77337732e-02 8.63329709e-01 -8.66928697e-01
1.09417140e+00 -2.23451805e+00 3.80203694e-01 8.34372714e-02
2.72456527e-01 1.49321005e-01 -2.15471566e-01 3.09109420e-01
1.29320353e-01 -1.55685902e-01 -2.05485389e-01 -7.64298320e-01
4.02457342e-02 -1.64872050e-01 -4.90090460e-01 7.75501430e-01
4.77884442e-01 1.05023038e+00 -8.46247613e-01 -6.12281322e-01
1.13872878e-01 5.70582449e-01 -2.99771369e-01 1.57709181e-01
-6.18847013e-02 6.00494266e-01 -2.82512993e-01 8.66173029e-01
7.39353955e-01 -3.61576855e-01 -4.21952941e-02 -4.73674387e-01
2.52881050e-01 7.70098343e-02 -5.32196581e-01 1.36689174e+00
-2.66317070e-01 7.29793549e-01 -2.19661176e-01 -1.11085665e+00
6.81245804e-01 7.48204365e-02 1.74439028e-01 -1.02138948e+00
6.32956445e-01 2.73963183e-01 -2.48903662e-01 -7.10167885e-01
6.22817993e-01 -3.42502505e-01 -3.62738252e-01 3.98027509e-01
-1.74512982e-01 -4.37016971e-02 -2.57571846e-01 3.73199642e-01
4.80631411e-01 -2.53943622e-01 1.68243740e-02 1.68409824e-01
5.39505005e-01 5.94943529e-03 2.10559592e-01 5.32248020e-01
-1.93946868e-01 6.20159566e-01 6.46896720e-01 -2.00005412e-01
-7.71682322e-01 -6.92866504e-01 1.00650929e-01 1.17379725e+00
7.37392128e-01 -2.88292527e-01 -6.62756622e-01 -9.59967315e-01
2.10657972e-03 6.32247686e-01 -7.84753621e-01 -4.41865742e-01
-3.69575649e-01 -1.11203802e+00 5.66528857e-01 4.79146801e-02
7.46665239e-01 -6.44872606e-01 -1.52319804e-01 -1.90849021e-01
-9.61969972e-01 -1.25615633e+00 -6.20344341e-01 -4.03724313e-01
-4.96330261e-01 -9.37015414e-01 -8.51225734e-01 -6.82916224e-01
6.54108286e-01 6.57768726e-01 7.79803276e-01 8.27585310e-02
-2.56174300e-02 1.96411714e-01 -4.62980241e-01 -1.83410496e-01
-5.14158130e-01 -1.04535714e-01 -2.06782192e-01 6.14080787e-01
1.51906222e-01 -4.28644232e-02 -6.35283351e-01 3.62550467e-01
-1.08204758e+00 3.73143762e-01 8.15738499e-01 1.18570530e+00
1.40705481e-01 -2.56697208e-01 4.70391572e-01 -5.99971473e-01
6.55099571e-01 -7.88914502e-01 -2.87857831e-01 2.09666342e-01
-3.45088810e-01 1.24298178e-01 3.78926069e-01 -7.07893133e-01
-1.02281308e+00 -3.69577199e-01 8.79846215e-02 -5.06597638e-01
3.46980721e-01 8.05665970e-01 -8.05348977e-02 7.53595382e-02
5.54938674e-01 2.63982087e-01 1.81709398e-02 -1.53418541e-01
5.93109787e-01 9.33264673e-01 5.15920937e-01 -3.02215248e-01
7.43151546e-01 7.28073835e-01 -3.76854509e-01 -5.42509079e-01
-1.20627451e+00 -5.06117344e-01 -1.34781405e-01 -3.77153099e-01
7.05457926e-01 -1.15686345e+00 -5.78135073e-01 7.50893772e-01
-1.22376513e+00 1.39684647e-01 5.11863530e-01 4.37434852e-01
-3.51145595e-01 5.91808617e-01 -7.33155847e-01 -6.98025286e-01
-3.84367794e-01 -1.25839043e+00 1.42508507e+00 -6.59797387e-03
1.12210214e-01 -7.16109693e-01 9.51548014e-03 1.03630626e+00
2.82561004e-01 2.01167539e-01 7.45890975e-01 -6.40160859e-01
-4.85667348e-01 -3.59754056e-01 -6.71118200e-01 2.31329903e-01
-1.60859808e-01 -1.78476095e-01 -1.02330482e+00 -2.50028908e-01
-2.15170532e-01 -5.54940403e-01 1.28163409e+00 -1.30798161e-01
9.15467560e-01 -5.26538908e-01 -2.48145834e-01 1.47471696e-01
1.02160466e+00 -4.30155784e-01 6.20830357e-01 4.30944383e-01
8.85221660e-01 7.25326121e-01 8.65016282e-01 4.25866425e-01
6.99376822e-01 8.10114920e-01 6.51356041e-01 -1.01933628e-01
8.11800808e-02 -1.36554226e-01 5.38636923e-01 8.77913833e-01
2.22078055e-01 -2.98477918e-01 -8.25803995e-01 4.74625558e-01
-2.05698943e+00 -1.17915583e+00 -1.32654935e-01 1.97588503e+00
7.81281650e-01 -5.78668825e-02 1.42318696e-01 -1.97364520e-02
9.32356834e-01 3.44546437e-01 -2.28925958e-01 3.37829068e-02
-5.05002975e-01 -6.27895415e-01 1.25912026e-01 4.53006953e-01
-1.56001604e+00 1.00292075e+00 5.10739040e+00 1.06402755e+00
-1.53853035e+00 4.05497968e-01 9.67177629e-01 4.23551127e-02
-4.30026531e-01 -1.93247080e-01 -2.40984604e-01 6.97875798e-01
4.09792453e-01 2.69625813e-01 2.74433821e-01 5.45346081e-01
9.18790698e-02 -1.60934299e-01 -6.54336870e-01 1.36596107e+00
5.73825598e-01 -1.25210428e+00 6.63229227e-02 -1.09646255e-02
1.01462996e+00 5.93550093e-02 5.33641219e-01 4.15836722e-01
-1.65642679e-01 -9.22798395e-01 9.69076872e-01 6.59165204e-01
4.49288368e-01 -8.69326174e-01 1.02538371e+00 3.77113968e-01
-7.05888748e-01 -2.18824044e-01 4.04133461e-04 2.98988253e-01
1.67979106e-01 7.12586880e-01 -6.07530296e-01 6.27161384e-01
4.64924544e-01 7.46934414e-01 -8.01618993e-01 8.27809095e-01
-1.25483483e-01 3.49880159e-01 7.08647743e-02 -1.23651763e-02
3.63728553e-01 -6.84060305e-02 6.05520725e-01 1.31273186e+00
1.01065122e-01 -2.60123849e-01 6.30299225e-02 6.81645453e-01
-3.54118049e-01 3.83997440e-01 -3.22876632e-01 -3.18040550e-01
1.60415545e-01 1.32266021e+00 -3.76607925e-01 -5.66127121e-01
-4.15668249e-01 1.27137125e+00 4.39397663e-01 1.66351512e-01
-1.27400541e+00 -9.31964368e-02 4.32404608e-01 -1.29086822e-01
1.35513946e-01 7.95206055e-02 -4.75852758e-01 -1.70514429e+00
8.66866708e-02 -1.15794075e+00 4.79156613e-01 -7.53885627e-01
-1.77686512e+00 4.93355781e-01 -1.40677020e-01 -1.46535552e+00
7.66178593e-02 -4.03540522e-01 -4.41710025e-01 6.13352835e-01
-1.67237329e+00 -1.56062603e+00 -4.28244889e-01 5.30624390e-01
3.13889384e-01 -1.52795330e-01 3.48566473e-01 4.11014587e-01
-6.89472318e-01 7.51199126e-01 1.16357796e-01 2.23469526e-01
1.17324543e+00 -8.26194704e-01 -2.04162061e-01 7.31277168e-01
6.58003911e-02 5.33519268e-01 7.46994555e-01 -5.75063407e-01
-1.41963434e+00 -9.56986666e-01 6.80529118e-01 -4.00469482e-01
7.59750187e-01 -1.89798072e-01 -9.64330912e-01 3.24747175e-01
4.63656873e-01 3.83209959e-02 6.25283360e-01 -1.33010611e-01
-9.38924491e-01 -9.70595852e-02 -1.01057231e+00 6.62757039e-01
4.10587519e-01 -8.26509356e-01 -5.46405077e-01 3.68619651e-01
3.26063395e-01 -3.91820759e-01 -5.65980017e-01 3.22711855e-01
6.42265558e-01 -1.08124650e+00 7.60734797e-01 -5.28189719e-01
8.76027107e-01 -3.05399567e-01 -1.09369159e-01 -1.34416151e+00
-4.23306674e-02 -3.59721899e-01 -5.06698489e-02 1.17283976e+00
3.92995924e-01 -6.38272345e-01 3.34082752e-01 5.12642004e-02
3.63768004e-02 -7.04195023e-01 -9.81986821e-01 -3.28536004e-01
-8.96987226e-03 -4.32883590e-01 2.26009592e-01 1.39361465e+00
3.04113597e-01 6.71819568e-01 -8.78382862e-01 3.65860164e-01
4.99899268e-01 3.55605513e-01 8.65502656e-01 -8.87275040e-01
-2.01147422e-01 -6.82668269e-01 -3.39532018e-01 -1.03410590e+00
2.65144408e-01 -8.18808138e-01 2.60989927e-02 -1.21769452e+00
7.80432343e-01 -5.83059825e-02 -1.62223816e-01 3.73973101e-01
-6.25457764e-01 6.89871550e-01 3.21239769e-01 5.52101314e-01
-9.27860737e-01 1.01365745e+00 1.30966783e+00 -5.37156403e-01
2.35462621e-01 -3.56635630e-01 -7.93232918e-01 5.82631648e-01
6.32953346e-01 -4.73497003e-01 5.35459742e-02 -4.11573529e-01
2.91147918e-01 -1.49653545e-02 6.81893826e-01 -5.41873395e-01
9.16175023e-02 -6.35421947e-02 3.01159382e-01 -5.03166795e-01
5.93339205e-01 -5.80562115e-01 -2.54621476e-01 3.63284528e-01
-3.65809351e-01 7.16154128e-02 -6.76297843e-02 7.95595944e-01
-5.41349769e-01 1.23819470e-01 1.01603270e+00 1.67934850e-01
-2.53831953e-01 8.03275332e-02 -2.74755061e-01 7.13724270e-02
9.50031936e-01 2.23388314e-01 -9.00346577e-01 -7.17707396e-01
-3.52162182e-01 2.89736003e-01 4.28918689e-01 7.27819979e-01
7.27549791e-01 -1.65042782e+00 -8.79063666e-01 -9.40462947e-02
6.80432677e-01 -5.59877634e-01 3.96861851e-01 1.48451340e+00
-2.99961478e-01 1.03741489e-01 -3.42428088e-02 -7.36952901e-01
-1.29805088e+00 6.09008789e-01 3.28062892e-01 -2.61679381e-01
-1.65984824e-01 6.93719983e-01 2.08122164e-01 -3.08335185e-01
-3.54669876e-02 1.27453297e-01 -2.00376824e-01 4.24768418e-01
6.76492989e-01 1.67720839e-01 -6.24472499e-02 -1.30257797e+00
-3.35658908e-01 3.84116024e-01 -1.31895974e-01 -2.53943443e-01
1.08493555e+00 -4.24672455e-01 -2.37575233e-01 4.00257528e-01
1.39846814e+00 1.39800414e-01 -1.01488113e+00 -5.04499435e-01
-6.54388517e-02 -7.53550828e-01 1.05586750e-02 -8.34140480e-01
-1.04578710e+00 8.74667525e-01 5.55808008e-01 2.63236821e-01
1.02570927e+00 1.87129259e-01 8.45702767e-01 6.63137212e-02
-6.07543625e-02 -9.50308442e-01 8.99540842e-01 4.02881324e-01
1.10691309e+00 -1.88962626e+00 -4.53680791e-02 -4.54596072e-01
-1.16681373e+00 1.02709579e+00 4.99462545e-01 4.66154218e-02
1.11134432e-01 -2.31449172e-01 3.05802107e-01 -3.97285461e-01
-3.30457360e-01 3.05543039e-02 7.98094094e-01 4.40737568e-02
2.75921941e-01 -2.47290097e-02 -3.22660625e-01 5.25792956e-01
1.37524039e-01 -6.22089148e-01 2.02036455e-01 6.11472547e-01
-3.07711154e-01 -7.07464397e-01 -7.10168660e-01 1.75944209e-01
-5.82407475e-01 -1.65528327e-01 -6.65377676e-01 5.18314242e-01
1.07995532e-02 9.82617736e-01 -1.10609636e-01 -5.30244350e-01
5.85018620e-02 -1.05040118e-01 2.87967503e-01 -1.10708117e-01
-5.00681341e-01 2.21005559e-01 9.73236859e-02 -4.22461659e-01
-6.04341924e-01 -5.22502959e-01 -7.48004973e-01 -3.51831585e-01
-7.17020154e-01 -1.00826181e-01 6.95700645e-01 1.17875421e+00
4.39851016e-01 2.03137323e-01 8.77310634e-01 -9.30781543e-01
-5.41294515e-01 -1.05385566e+00 -3.43258977e-02 7.82422006e-01
4.92602646e-01 -6.99849725e-01 -4.97253299e-01 -1.27464086e-01] | [8.165167808532715, 10.301095008850098] |
bf2bb55b-ce3c-4158-88c3-e6ec46d12792 | causal-discovery-in-probabilistic-networks | 2208.04627 | null | https://arxiv.org/abs/2208.04627v2 | https://arxiv.org/pdf/2208.04627v2.pdf | Causal Discovery in Probabilistic Networks with an Identifiable Causal Effect | Causal identification is at the core of the causal inference literature, where complete algorithms have been proposed to identify causal queries of interest. The validity of these algorithms hinges on the restrictive assumption of having access to a correctly specified causal structure. In this work, we study the setting where a probabilistic model of the causal structure is available. Specifically, the edges in a causal graph are assigned probabilities which may, for example, represent degree of belief from domain experts. Alternatively, the uncertainly about an edge may reflect the confidence of a particular statistical test. The question that naturally arises in this setting is: Given such a probabilistic graph and a specific causal effect of interest, what is the subgraph which has the highest plausibility and for which the causal effect is identifiable? We show that answering this question reduces to solving an NP-hard combinatorial optimization problem which we call the edge ID problem. We propose efficient algorithms to approximate this problem, and evaluate our proposed algorithms against real-world networks and randomly generated graphs. | ['Negar Kiyavash', 'Jalal Etesami', 'Matthew J. Vowels', 'Ehsan Mokhtarian', 'Fateme Jamshidi', 'Sina Akbari'] | 2022-08-09 | null | null | null | null | ['causal-identification'] | ['reasoning'] | [ 5.66764235e-01 6.71391666e-01 -5.50844312e-01 -3.00008565e-01
-4.93880540e-01 -6.88179016e-01 4.53640580e-01 5.74939907e-01
1.44039486e-02 9.63384092e-01 2.53638387e-01 -6.77713454e-01
-8.02126765e-01 -1.17060256e+00 -1.04362142e+00 -5.90145767e-01
-4.95526969e-01 5.74525952e-01 2.31115282e-01 3.55598599e-01
3.34759146e-01 2.20949650e-01 -1.00129294e+00 -4.95599210e-02
7.40349829e-01 7.12724864e-01 -4.85995784e-02 5.54951727e-01
2.05853924e-01 5.63973129e-01 -2.93889701e-01 -4.20167476e-01
-6.95078447e-02 -3.05043399e-01 -1.09991717e+00 -6.85741231e-02
1.24757834e-01 -6.93486035e-02 -3.43878567e-01 1.25397909e+00
1.02715641e-01 -1.30823985e-01 8.18877935e-01 -1.66407537e+00
-3.70083004e-01 1.14685380e+00 -7.54186571e-01 3.08824569e-01
6.27187729e-01 -3.75824064e-01 1.58290720e+00 -3.35023433e-01
4.75164264e-01 1.30659485e+00 1.31601036e-01 -8.83601308e-02
-1.43076992e+00 -5.09084225e-01 2.04870179e-01 2.29276955e-01
-1.40218234e+00 -1.18930951e-01 6.21126354e-01 -6.10183239e-01
5.08549400e-02 2.63817817e-01 2.56424516e-01 8.17397892e-01
-6.17871173e-02 3.29468220e-01 1.06938672e+00 -3.79604787e-01
5.76172471e-01 -3.97149213e-02 2.93042690e-01 6.44109786e-01
7.64398813e-01 2.08453193e-01 -5.09414792e-01 -5.99991858e-01
4.28891182e-01 -2.70046920e-01 -5.58184266e-01 -3.28135669e-01
-8.60807359e-01 9.30768907e-01 3.85719657e-01 -1.04295634e-01
-3.56549829e-01 4.22689468e-01 -7.93219060e-02 7.00898655e-03
1.69692993e-01 1.61717057e-01 -4.39479858e-01 2.99561352e-01
-4.44335103e-01 2.22108230e-01 1.02988052e+00 8.20689380e-01
6.41686857e-01 -5.27582526e-01 -8.75863060e-02 2.55918860e-01
5.39739430e-01 3.73854637e-01 -5.36314189e-01 -6.98671937e-01
5.21834195e-01 5.37598550e-01 4.67132568e-01 -1.26986003e+00
-1.58348620e-01 -2.57170409e-01 -6.13610506e-01 -2.09315985e-01
7.34812319e-01 -4.60395813e-01 -5.11396170e-01 2.23740482e+00
6.71301603e-01 7.40793169e-01 -2.43189082e-01 9.27154899e-01
3.62977475e-01 4.80573952e-01 2.08127439e-01 -3.87290657e-01
1.32172918e+00 -1.48615530e-02 -4.60249215e-01 -9.50360969e-02
1.97235063e-01 -4.22423363e-01 5.69747984e-01 2.56275553e-02
-4.42607284e-01 1.65575325e-01 -8.21941555e-01 4.14106965e-01
-1.35279922e-02 -3.49517375e-01 7.74144292e-01 8.08314502e-01
-6.44044399e-01 3.91041577e-01 -4.93258208e-01 -4.54638988e-01
5.15091904e-02 3.33276987e-01 -2.99390018e-01 -4.62605089e-01
-1.50338733e+00 5.03705740e-01 5.38923025e-01 1.56268165e-01
-1.11878681e+00 -6.15124643e-01 -5.13997316e-01 2.48243108e-01
1.12675679e+00 -7.97987282e-01 9.01215374e-01 -6.01125121e-01
-6.98596478e-01 4.80576545e-01 -3.52351695e-01 -5.02022505e-01
3.72997880e-01 5.43020517e-02 -3.67577910e-01 9.31350961e-02
2.64877975e-01 -1.09278589e-01 7.03553081e-01 -1.28721738e+00
-8.38947833e-01 -4.32830036e-01 7.08804905e-01 -1.07914798e-01
3.74062955e-02 -2.11619567e-02 -2.70792156e-01 -1.93969995e-01
6.74389228e-02 -1.02839172e+00 -2.36439109e-01 -1.55588150e-01
-1.02538812e+00 -3.45557302e-01 5.28170228e-01 -3.16327065e-01
1.37952852e+00 -1.86252439e+00 2.98465267e-02 7.32242048e-01
3.55380893e-01 -6.22526944e-01 3.37984860e-01 5.20526111e-01
-1.94103539e-01 4.87871319e-01 -4.28120196e-01 5.54210663e-01
-2.68884748e-02 2.30889857e-01 -4.31390673e-01 8.32655668e-01
1.62307933e-01 4.10918146e-01 -9.92149711e-01 -7.03675985e-01
-1.87646747e-01 -2.55538709e-02 -6.07188940e-01 2.24042296e-01
-4.36258316e-01 1.79606616e-01 -9.27468717e-01 2.80218005e-01
5.80730379e-01 -6.43031776e-01 7.50475466e-01 -2.60272235e-01
2.41456032e-02 2.75654018e-01 -1.68937838e+00 8.78490448e-01
-9.92219150e-02 2.52174854e-01 1.17216617e-01 -1.11017871e+00
3.70162368e-01 4.09478694e-01 4.23858553e-01 4.24591973e-02
3.10717709e-02 -1.84444204e-01 7.47358203e-02 -3.33425879e-01
3.90984006e-02 -1.71142429e-01 -3.00561517e-01 5.12766838e-01
-3.42093229e-01 4.40423787e-01 1.95247769e-01 4.50263143e-01
1.47995722e+00 -3.48424673e-01 5.69127738e-01 -4.03216332e-01
1.92831665e-01 9.84059051e-02 7.99063921e-01 1.02672887e+00
1.61139131e-01 1.97308809e-01 1.27676892e+00 1.24070622e-01
-6.14903629e-01 -1.30569732e+00 -1.43259987e-01 6.50312483e-01
2.43828923e-01 -1.46030679e-01 -5.45364201e-01 -6.92486882e-01
1.54432490e-01 5.46853423e-01 -8.49438787e-01 1.71427608e-01
-1.67200521e-01 -8.18050027e-01 7.37530440e-02 6.04533106e-02
1.16600551e-01 -5.32947838e-01 -1.97669834e-01 1.59649074e-01
-3.48788381e-01 -9.51530397e-01 -4.18585926e-01 6.69385344e-02
-7.07575738e-01 -1.80665028e+00 -3.94975916e-02 -2.16176465e-01
9.10122454e-01 6.81739580e-03 1.06710958e+00 -1.55338896e-02
-5.42208068e-02 2.20054567e-01 -1.05449446e-01 -2.12788731e-01
-2.23750919e-01 -7.56556392e-02 -1.44935802e-01 2.68324107e-01
-2.22343449e-02 -6.10598922e-01 -4.82219398e-01 2.78248698e-01
-9.48707879e-01 -2.22180068e-01 6.71185553e-01 4.86516118e-01
5.37536979e-01 6.95717156e-01 6.30510926e-01 -1.52266049e+00
6.16998494e-01 -1.08441758e+00 -9.21778619e-01 5.22464514e-01
-5.05373597e-01 3.45928282e-01 4.65951622e-01 -7.35474452e-02
-1.12504661e+00 1.70157835e-01 4.01261508e-01 2.00276837e-01
-2.10343987e-01 1.09015822e+00 -6.66096926e-01 2.86980063e-01
3.13554347e-01 -4.19587851e-01 -5.84614456e-01 -2.25519940e-01
4.14914846e-01 1.85580522e-01 4.13226485e-01 -9.26676095e-01
8.93403947e-01 5.02952814e-01 5.22354782e-01 -5.53852439e-01
-1.12636840e+00 -4.47928160e-01 -2.53936678e-01 -1.16696738e-01
6.46705627e-01 -4.89112884e-01 -1.20970106e+00 -2.36129984e-01
-1.07995462e+00 6.28167465e-02 2.17334405e-01 4.46760595e-01
-2.79308110e-01 2.10036367e-01 -6.59985617e-02 -1.10536993e+00
2.82499373e-01 -7.63618946e-01 5.89116037e-01 9.79259089e-02
-2.96417147e-01 -1.05009580e+00 1.65669650e-01 1.43937767e-01
-2.95977563e-01 6.17889524e-01 1.46817017e+00 -5.14423966e-01
-8.87701869e-01 -2.70342439e-01 -5.57155609e-01 -4.92391199e-01
2.20195845e-01 4.20742519e-02 -6.34627521e-01 8.60168263e-02
-4.04688835e-01 1.29016355e-01 6.04471803e-01 6.94066763e-01
1.11391008e+00 -6.86131716e-01 -6.94436729e-01 -5.28095812e-02
1.61500823e+00 -1.06790848e-02 2.31440425e-01 -2.13164285e-01
4.76057321e-01 8.85229409e-01 6.08641803e-01 5.00552654e-01
4.72139269e-01 6.18517697e-01 6.07189655e-01 -1.69054493e-02
5.12310088e-01 -5.81614852e-01 -1.30281880e-01 1.45038694e-01
1.71908375e-03 -7.02059925e-01 -7.76314974e-01 6.84015453e-01
-1.84746730e+00 -9.43427145e-01 -4.45741355e-01 2.46427512e+00
8.67177188e-01 1.88676283e-01 1.18837841e-01 1.29219353e-01
1.24348783e+00 -1.25897393e-01 -4.63800192e-01 -1.73660833e-02
-4.14933264e-02 -9.13443193e-02 6.81517959e-01 7.17828691e-01
-9.66273367e-01 4.33806658e-01 6.04158688e+00 5.36893308e-01
-5.52408218e-01 6.73663151e-03 6.85271502e-01 3.49950075e-01
-4.97134566e-01 6.59624338e-01 -6.34814978e-01 5.07074952e-01
8.98189366e-01 -6.53582931e-01 2.62726426e-01 4.67100173e-01
5.49289465e-01 -4.50854987e-01 -1.34770930e+00 2.19870225e-01
-6.08420789e-01 -1.12583721e+00 -1.52309120e-01 4.52482313e-01
7.35532820e-01 -4.50484097e-01 -7.07205236e-02 -3.45654190e-01
1.07951951e+00 -1.04983854e+00 4.00320649e-01 6.23961866e-01
6.87555254e-01 -7.76382267e-01 4.92603064e-01 1.62622035e-01
-1.11157274e+00 -5.08266650e-02 -1.00388333e-01 3.74851786e-02
2.84967721e-01 1.17415333e+00 -1.18043089e+00 6.71613991e-01
4.52847064e-01 2.24653527e-01 -1.24173850e-01 1.22997034e+00
-7.36772180e-01 1.28984940e+00 -5.55651426e-01 -6.99857026e-02
-6.55514821e-02 2.65485328e-02 5.02532482e-01 1.04081583e+00
1.80256426e-01 3.56191605e-01 7.91459382e-02 8.91087651e-01
-4.10380483e-01 -1.14517376e-01 -5.37760139e-01 -1.81786805e-01
8.03569734e-01 9.85643923e-01 -7.55998373e-01 3.82227227e-02
-3.88471186e-01 4.01304632e-01 2.60298759e-01 4.69286412e-01
-8.01658154e-01 -1.15248570e-02 4.58124757e-01 3.10723603e-01
-4.82294708e-02 1.41644761e-01 -2.57633448e-01 -6.80049598e-01
-7.34496266e-02 -3.89736265e-01 9.75609601e-01 -5.74365795e-01
-1.28113949e+00 -2.63337940e-02 2.27703050e-01 -5.93525708e-01
-1.48353443e-01 -2.47548565e-01 -6.44923270e-01 9.13721919e-01
-1.32022572e+00 -6.57562852e-01 -1.51631162e-01 4.58302736e-01
-1.85629502e-01 5.48719764e-01 5.37378669e-01 8.81964266e-02
-5.46697021e-01 1.82018459e-01 -1.91400126e-01 6.30389005e-02
4.24297094e-01 -1.51691043e+00 -1.07842080e-01 1.00605106e+00
-4.44347486e-02 5.41627169e-01 1.23859453e+00 -9.29516137e-01
-1.58347058e+00 -1.13681328e+00 9.79982436e-01 -2.72614777e-01
1.03385723e+00 -2.10541382e-01 -7.36926258e-01 7.26841807e-01
-7.73338303e-02 -3.34192067e-02 4.64539886e-01 6.07282162e-01
-2.69682974e-01 -9.14552249e-03 -1.13328850e+00 5.09675205e-01
1.14365530e+00 -1.97925046e-01 -3.98734272e-01 3.71865422e-01
6.91514909e-01 1.01138800e-01 -7.30792701e-01 2.37225965e-01
3.19726676e-01 -5.54003596e-01 7.68770158e-01 -7.89151967e-01
5.53425610e-01 -5.77822089e-01 -1.22154824e-01 -1.27405214e+00
-2.25836366e-01 -3.75487059e-01 1.08667739e-01 1.19075847e+00
8.10210049e-01 -5.20500779e-01 8.46564651e-01 8.02657485e-01
5.45157433e-01 -3.99213970e-01 -9.77419198e-01 -3.98755342e-01
-3.85408401e-01 -5.49658418e-01 5.91852009e-01 1.10418093e+00
1.58065513e-01 5.92768312e-01 -4.41459119e-01 1.10898292e+00
1.09831011e+00 4.90197361e-01 4.44356263e-01 -1.63478851e+00
-5.29079139e-01 -1.24262283e-02 -3.48233134e-01 -6.45647824e-01
2.94303447e-01 -7.67961025e-01 3.12984698e-02 -1.56531358e+00
5.71615160e-01 -6.70340359e-01 -4.01699431e-02 4.03186649e-01
-4.25320357e-01 -3.71997446e-01 -3.64876926e-01 -1.88020974e-01
-3.62815022e-01 1.65918469e-01 8.05682361e-01 -1.00959770e-01
5.38598150e-02 2.41303846e-01 -8.81726503e-01 6.81867540e-01
5.62647343e-01 -7.74020076e-01 -5.71120143e-01 -9.36944559e-02
7.36213386e-01 8.40746760e-01 7.38407493e-01 -2.87413955e-01
5.24081409e-01 -7.39228368e-01 -1.66779190e-01 -3.46994847e-01
-1.30962804e-01 -9.00792360e-01 5.31309187e-01 2.50538141e-01
-5.15222907e-01 -1.36917070e-01 -2.57192343e-01 1.13680506e+00
-1.41227916e-01 -2.89516956e-01 4.06448066e-01 1.51279777e-01
-4.58564371e-01 4.25121069e-01 -1.37034342e-01 3.20631325e-01
9.17923748e-01 2.10066020e-01 -3.90667558e-01 -5.33010960e-01
-7.78158128e-01 3.95416796e-01 1.35450736e-01 9.05936062e-02
4.95071977e-01 -1.01345384e+00 -8.16481769e-01 -6.31147206e-01
1.68334633e-01 -2.34241500e-01 5.66735268e-02 8.00864458e-01
1.07636541e-01 4.56917882e-01 4.97224420e-01 -2.31749117e-01
-1.07750010e+00 7.05360889e-01 1.46403670e-01 -1.27573431e-01
-8.78340825e-02 5.41410744e-01 5.08769274e-01 9.83484238e-02
-2.67065018e-02 -1.88748408e-02 -1.45064026e-01 -2.26523966e-01
7.68005103e-02 5.62746108e-01 -3.17026585e-01 -4.12856102e-01
-5.15745699e-01 7.05418587e-02 2.18661457e-01 -3.08254808e-01
1.12429142e+00 -1.60795942e-01 -3.86041880e-01 2.48825192e-01
8.42783689e-01 3.03773105e-01 -8.64446402e-01 -3.28471839e-01
3.86931956e-01 -4.91117507e-01 5.84654100e-02 -8.05618644e-01
-9.11082506e-01 3.92931670e-01 5.63093722e-02 5.06119609e-01
1.00268948e+00 4.57879156e-01 -2.00260221e-03 8.82823542e-02
4.87212747e-01 -5.81752539e-01 -5.44286966e-01 1.19818099e-01
6.10466361e-01 -1.08783400e+00 -5.56267537e-02 -1.01709414e+00
-1.08658768e-01 7.73736358e-01 3.29436839e-01 4.58838195e-02
7.99384236e-01 2.17390627e-01 -7.03261137e-01 -4.05563444e-01
-8.73979390e-01 -2.03944832e-01 2.52534151e-01 2.38659292e-01
3.94751161e-01 6.42657042e-01 -4.95950609e-01 5.01818478e-01
-1.21284306e-01 -2.89277844e-02 5.97512186e-01 5.55771708e-01
-3.68177265e-01 -8.69041920e-01 -5.12169123e-01 5.72167218e-01
-6.31706715e-01 -1.45733178e-01 -5.63610911e-01 5.46086013e-01
1.88097209e-02 1.41601169e+00 -2.05696985e-01 -5.15749902e-02
3.56120676e-01 -3.56163800e-01 3.60234410e-01 -6.18457079e-01
1.89741343e-01 -2.40313917e-01 4.23823386e-01 -3.13768655e-01
-3.65896910e-01 -6.82201087e-01 -1.12031066e+00 -4.68907446e-01
-3.62636060e-01 5.38434148e-01 2.88392484e-01 1.19485271e+00
8.72764364e-02 3.15227538e-01 6.62790656e-01 3.42073530e-01
-3.27780902e-01 -6.71174645e-01 -7.84163117e-01 7.35084787e-02
2.33937532e-01 -8.09002638e-01 -4.14940566e-01 6.22416772e-02] | [7.727264404296875, 5.3139238357543945] |
089f0f74-a743-405a-b3db-49b70fe9ed61 | metabolic-regulatory-network-kinetic-modeling | 2305.00165 | null | https://arxiv.org/abs/2305.00165v1 | https://arxiv.org/pdf/2305.00165v1.pdf | Metabolic Regulatory Network Kinetic Modeling with Multiple Isotopic Tracers for iPSCs | The rapidly expanding market for regenerative medicines and cell therapies highlights the need to advance the understanding of cellular metabolisms and improve the prediction of cultivation production process for human induced pluripotent stem cells (iPSCs). In this paper, a metabolic kinetic model was developed to characterize underlying mechanisms of iPSC culture process, which can predict cell response to environmental perturbation and support process control. This model focuses on the central carbon metabolic network, including glycolysis, pentose phosphate pathway (PPP), tricarboxylic acid (TCA) cycle, and amino acid metabolism, which plays a crucial role to support iPSC proliferation. Heterogeneous measures of extracellular metabolites and multiple isotopic tracers collected under multiple conditions were used to learn metabolic regulatory mechanisms. Systematic cross-validation confirmed the model's performance in terms of providing reliable predictions on cellular metabolism and culture process dynamics under various culture conditions. Thus, the developed mechanistic kinetic model can support process control strategies to strategically select optimal cell culture conditions at different times, ensure cell product functionality, and facilitate large-scale manufacturing of regenerative medicines and cell therapies. | ['Sarah W. Harcum', 'Wei Xie', 'Keqi Wang'] | 2023-04-29 | null | null | null | null | ['culture'] | ['speech'] | [ 5.50723784e-02 -6.38342917e-01 -4.04022783e-01 4.36393052e-01
1.84404224e-01 -9.09105957e-01 5.39379060e-01 5.63637257e-01
-4.33782227e-02 9.27035213e-01 3.08487147e-01 -2.15626478e-01
1.41072392e-01 -7.59430349e-01 -1.85546443e-01 -1.02762747e+00
3.82568479e-01 5.98210990e-01 -2.71089762e-01 2.11507201e-01
5.51963151e-01 9.02607024e-01 -9.97581780e-01 -3.74783725e-02
7.79076815e-01 8.22171807e-01 4.70485270e-01 6.05898857e-01
-3.95880789e-01 4.48459953e-01 -9.28816944e-02 1.54120266e-01
-2.95231249e-02 -8.20007026e-01 -2.09063590e-01 -1.22550517e-01
-1.08168566e+00 3.52916867e-01 1.08457699e-01 5.58233976e-01
7.56472468e-01 -1.42396256e-01 9.40350652e-01 -1.19314063e+00
-6.53192937e-01 9.52858329e-02 1.02466173e-01 4.89510596e-02
2.65781134e-01 6.28838241e-01 1.90694258e-01 -7.64393389e-01
8.02553892e-01 8.46114039e-01 4.13436741e-01 4.84265864e-01
-1.12840772e+00 -5.02237916e-01 -1.64180830e-01 -9.83425509e-03
-1.22387981e+00 -3.84185731e-01 6.71818405e-02 -7.95808017e-01
7.32647657e-01 2.93667525e-01 1.39410603e+00 6.46933675e-01
9.03803349e-01 2.41621867e-01 1.25262797e+00 -1.64370723e-02
7.64745712e-01 -1.42709278e-02 -5.00720382e-01 4.51379687e-01
5.49563289e-01 -4.97295447e-02 -7.80812681e-01 1.86415941e-01
1.06321442e+00 -7.34008178e-02 -3.40332985e-01 -1.81552187e-01
-1.41683042e+00 3.13344330e-01 -3.41222703e-01 4.30816352e-01
-5.59454679e-01 2.15362653e-01 3.39361995e-01 -2.05793872e-01
1.68132037e-02 4.67699826e-01 -8.76967072e-01 -2.57487416e-01
-5.31211734e-01 -3.44078243e-01 8.99879992e-01 7.03685999e-01
2.73963034e-01 -1.45317279e-02 -3.18848878e-01 6.98686957e-01
4.04601097e-01 5.09424627e-01 5.23028076e-01 -1.07872856e+00
-3.73409539e-01 6.91431761e-01 1.65077582e-01 -4.51567113e-01
-7.17986763e-01 -4.70633239e-01 -8.51333141e-01 -4.32824433e-01
4.57332522e-01 -1.43401995e-01 -7.74987161e-01 1.24931788e+00
4.54445004e-01 3.04814905e-01 1.50100097e-01 8.40642154e-01
3.04818451e-01 7.89979100e-01 4.36636925e-01 -1.04526925e+00
1.25760615e+00 -3.55274767e-01 -8.71054351e-01 2.59705693e-01
7.63184309e-01 -5.62434793e-01 5.07183075e-01 1.94103852e-01
-1.16053832e+00 -1.00501850e-01 -8.22025359e-01 1.89806998e-01
-5.86692393e-01 -8.41938108e-02 8.13018799e-01 4.31659400e-01
-6.80581272e-01 1.14742553e+00 -9.29307222e-01 -7.62445867e-01
3.06292653e-01 3.52446586e-01 -7.71340355e-02 -2.63061643e-01
-7.79815972e-01 1.19494641e+00 3.14303964e-01 3.62265676e-01
-1.10961282e+00 -1.06188381e+00 -5.20682454e-01 1.37966692e-01
-5.21598697e-01 -1.46371591e+00 6.85805619e-01 -1.24542840e-01
-2.10582685e+00 4.62610155e-01 -2.21843660e-01 -7.23228753e-02
2.57164389e-01 4.09786880e-01 -2.71087706e-01 1.38114721e-01
-1.44378385e-02 4.36977118e-01 -2.81884998e-01 -1.03933156e+00
-2.21674964e-01 -4.26345080e-01 -9.81008828e-01 1.81818083e-01
1.68867692e-01 8.19261521e-02 -2.33629540e-01 -2.73163348e-01
4.11796749e-01 -1.06669545e+00 -3.89186591e-01 4.85378243e-02
-1.14335649e-01 4.20277327e-01 1.12048730e-01 -5.97678781e-01
6.87399268e-01 -1.63713086e+00 5.32037139e-01 8.36386308e-02
-5.16634345e-01 1.69456750e-02 1.31657913e-01 8.18586767e-01
2.60353506e-01 6.16464615e-01 -7.04418495e-02 2.60696113e-01
-2.89225429e-01 -1.11864105e-01 5.51986814e-01 5.95396638e-01
2.33508334e-01 9.57582772e-01 -1.23129630e+00 -3.44550192e-01
3.78663689e-01 5.45962334e-01 2.21965406e-02 -7.16775283e-02
-5.08942008e-01 1.16416442e+00 -4.77816790e-01 1.21425331e+00
3.51257354e-01 -3.08497380e-02 5.04021943e-01 -1.15108751e-01
-4.05002385e-01 -1.45334110e-01 -7.02463329e-01 1.70387292e+00
-3.58648151e-01 1.92655250e-01 1.64444506e-01 -2.22979441e-01
8.13772857e-01 4.82270062e-01 1.10340476e+00 -8.70575249e-01
3.76061201e-01 5.23451090e-01 -2.98343934e-02 -5.11257410e-01
-3.07484120e-01 -8.33160698e-01 3.83793980e-01 6.48454726e-02
-7.42995515e-02 -2.49556378e-01 3.86808991e-01 -2.38278463e-01
6.80187225e-01 3.46322358e-01 1.08479470e-01 -6.74711823e-01
7.86524832e-01 5.28237484e-02 9.22029316e-01 -2.42935538e-01
-5.16607940e-01 2.70639211e-01 5.34943402e-01 -3.14153754e-03
-1.03239417e+00 -7.26307213e-01 -3.84430178e-02 2.42620155e-01
3.54123354e-01 -2.49192882e-02 -4.64998752e-01 4.19629484e-01
-3.96968126e-02 7.26373971e-01 -3.64889979e-01 -3.14313620e-01
-1.98985219e-01 -1.21173537e+00 2.57363468e-01 3.09111059e-01
2.31710479e-01 -4.95926291e-01 6.10223413e-03 6.41536057e-01
-1.41285211e-01 -8.00279617e-01 -1.63212270e-01 3.18911910e-01
-1.10186279e+00 -1.51162708e+00 -7.16188669e-01 -5.43417752e-01
4.71572131e-01 -9.90648381e-03 5.71086705e-01 -4.79266793e-02
-3.55754763e-01 5.63051812e-02 -1.14843613e-02 -5.09156346e-01
-6.49084687e-01 -4.08978671e-01 3.98903400e-01 -2.58851051e-01
-4.44056801e-02 -8.01373243e-01 -1.15894878e+00 5.32336533e-01
-6.15965784e-01 2.39398539e-01 6.00859225e-01 3.67577493e-01
1.25330722e+00 -6.78266585e-02 1.08075881e+00 -4.83456254e-01
3.61740440e-01 -7.70954430e-01 -4.19352263e-01 4.47282940e-01
-9.33379531e-01 -9.25818756e-02 6.97293401e-01 -1.28325731e-01
-1.16390169e+00 3.69538777e-02 2.28171438e-01 1.30642503e-01
1.09735228e-01 5.57717383e-01 -6.10307455e-01 1.22397125e-01
1.19746430e-02 6.64100528e-01 2.21809030e-01 -2.26344615e-01
-6.87894002e-02 1.46109238e-01 1.50109142e-01 -6.48898005e-01
1.99805304e-01 4.09094572e-01 7.21133530e-01 -5.82663476e-01
-6.34233803e-02 -6.04244232e-01 -4.40987527e-01 -5.20866990e-01
1.00405228e+00 -1.15295458e+00 -9.88588989e-01 5.46572626e-01
-7.19129384e-01 -5.55033505e-01 2.08643749e-01 7.48430789e-01
-8.46476734e-01 1.63758844e-01 -1.11587286e+00 -7.95528054e-01
-5.90152502e-01 -8.93905699e-01 4.78850454e-01 3.18427145e-01
-2.93281496e-01 -9.46704149e-01 4.37128693e-01 2.65369803e-01
3.80887389e-01 8.33824635e-01 1.26388383e+00 -1.37537792e-01
-6.95868075e-01 1.93123788e-01 3.01366955e-01 -3.98617499e-02
4.66033995e-01 4.90130275e-01 -5.30485511e-01 4.93866690e-02
-9.88516659e-02 3.47221017e-01 4.25473899e-01 6.54481053e-01
5.95643938e-01 1.02091849e-01 -7.12176383e-01 8.66947651e-01
1.85175931e+00 6.84680462e-01 8.96210015e-01 4.22828645e-01
3.81048173e-01 5.69108486e-01 6.53561831e-01 4.84386116e-01
-4.39354479e-02 -1.11357393e-02 3.75813514e-01 2.64302939e-01
-1.20847397e-01 -3.76437396e-01 4.09677774e-01 8.10687363e-01
-3.47375810e-01 -2.64499903e-01 -7.63934791e-01 5.28699636e-01
-1.28955483e+00 -6.34089649e-01 -4.93090123e-01 2.09650993e+00
6.67076647e-01 -4.01032537e-01 5.08806743e-02 -2.44310439e-01
8.07334363e-01 -7.46762395e-01 -8.67084503e-01 -3.36818427e-01
-4.44139808e-01 1.10289671e-01 6.86775982e-01 9.90554318e-02
-5.99629879e-02 6.44002020e-01 7.20011282e+00 1.07133374e-01
-1.15038586e+00 -2.10511938e-01 7.02200890e-01 -4.53816742e-01
-2.25594461e-01 3.01837265e-01 -4.97914135e-01 6.98347926e-01
1.20788670e+00 -6.60079360e-01 3.64143193e-01 1.40736416e-01
1.02819550e+00 -3.45570117e-01 -1.18891823e+00 6.67035341e-01
-6.62910521e-01 -1.83123362e+00 -7.58103058e-02 4.30025995e-01
9.02475715e-01 -1.17777750e-01 -5.03172159e-01 -1.55120060e-01
6.63347989e-02 -7.55071104e-01 6.03487194e-01 9.69381690e-01
6.02364540e-01 -6.58759117e-01 7.76853323e-01 3.53623658e-01
-1.06220353e+00 -1.57295227e-01 -2.88264364e-01 1.14420548e-01
4.07988191e-01 7.34848082e-01 -1.01234913e+00 5.39239049e-01
1.19806848e-01 6.36525035e-01 -1.13413759e-01 1.22039580e+00
8.24121535e-02 2.15618119e-01 -2.71300077e-01 1.49142668e-02
-4.00517851e-01 -7.61381924e-01 -4.15684283e-02 1.03897607e+00
8.79212379e-01 3.47713351e-01 -2.42819265e-01 7.87486732e-01
-4.79307305e-03 3.75118405e-01 -4.02720243e-01 -6.35583282e-01
8.29485416e-01 9.82287109e-01 -1.30549920e+00 -4.98080179e-02
5.40689193e-02 7.59611547e-01 -4.35253590e-01 2.80610174e-01
-8.19489300e-01 4.83131170e-01 9.58828926e-01 2.92734087e-01
-1.23689519e-02 -3.10170323e-01 -5.39360702e-01 -8.86024296e-01
-5.74894845e-01 -2.63351709e-01 2.24632159e-01 -7.04553723e-01
-6.85083389e-01 -4.00136441e-01 -4.49561030e-01 -9.32255864e-01
3.63939106e-01 -5.83682835e-01 -7.28222370e-01 9.50373471e-01
-1.57188320e+00 -9.52350616e-01 -3.40791931e-03 2.39800103e-02
4.20225471e-01 3.50591332e-01 9.16756809e-01 3.94462347e-02
-1.22718132e+00 -3.84085417e-01 6.75999463e-01 -6.60599828e-01
6.02115333e-01 -8.01056027e-01 -4.67482328e-01 7.07962871e-01
-8.61477971e-01 6.20786846e-01 7.04995573e-01 -1.10967493e+00
-1.89115644e+00 -1.24532723e+00 3.10603976e-01 -1.63356345e-02
3.56905431e-01 4.50931750e-02 -3.21741313e-01 1.76309898e-01
-2.19477918e-02 -4.43895161e-01 1.44065976e+00 -4.08718497e-01
3.56925219e-01 9.21887532e-02 -1.25851500e+00 5.93739569e-01
7.17363834e-01 -9.14903581e-02 4.35767949e-01 4.68351990e-01
3.43287259e-01 -2.94395328e-01 -1.67887914e+00 1.99843287e-01
6.72288120e-01 -4.57428306e-01 5.87318301e-01 -6.00258708e-01
3.94385606e-01 -7.88210928e-01 -1.54426575e-01 -1.26032341e+00
-4.56680447e-01 -5.79059064e-01 -4.75326329e-01 1.21618342e+00
5.00280201e-01 -4.44650590e-01 5.60454845e-01 7.45756745e-01
-4.26446676e-01 -8.54080498e-01 -7.17321992e-01 -7.93279648e-01
2.89359272e-01 1.82666272e-01 6.93828046e-01 8.77656519e-01
3.97991985e-01 1.36973947e-01 3.75853628e-02 1.76831372e-02
1.49559632e-01 -1.82973742e-01 3.24767321e-01 -1.10069752e+00
9.18169171e-02 -5.50883591e-01 -1.73557162e-01 -2.00966876e-02
-1.15200222e-01 -8.74406815e-01 -1.45757064e-01 -2.10352659e+00
1.43270746e-01 -3.05211574e-01 -4.77200150e-01 -1.66504353e-01
2.88504269e-03 -7.70750567e-02 1.68538764e-01 3.21111053e-01
9.07081291e-02 6.73005998e-01 1.53971481e+00 -4.53214943e-02
-2.69120365e-01 -1.98518857e-01 -8.34504604e-01 1.59620509e-01
8.87598038e-01 -2.86890984e-01 -4.00878906e-01 -6.37204424e-02
3.54773223e-01 4.85668391e-01 -6.24316856e-02 -9.91907537e-01
1.50665268e-01 -7.15312421e-01 7.09459901e-01 -2.61345416e-01
2.05123927e-02 -6.51905358e-01 1.17727089e+00 1.09159684e+00
-7.59894252e-02 -2.22464725e-01 1.79154411e-01 7.02637076e-01
2.13701889e-01 1.15797266e-01 7.86246002e-01 -4.17305589e-01
-1.59572363e-01 3.84226114e-01 -8.13347161e-01 -5.60590804e-01
1.54334891e+00 -6.46426618e-01 -4.28544015e-01 2.71354169e-01
-1.02066779e+00 4.04651910e-01 1.07289279e+00 -3.63979414e-02
1.06547557e-01 -1.43056035e+00 -5.89967489e-01 -3.88462216e-01
-1.21822409e-01 5.21722669e-03 5.68405032e-01 1.15992689e+00
-1.19621825e+00 5.59471905e-01 -5.65505445e-01 -5.60897529e-01
-5.40699482e-01 8.28345418e-01 6.16875291e-01 -1.45601153e-01
-2.49174442e-02 4.37523395e-01 -7.08791688e-02 1.70546591e-01
-5.13245881e-01 -3.09793353e-01 6.07108995e-02 2.01672167e-01
3.83496135e-01 7.76444554e-01 3.45406272e-02 -3.54664624e-01
-4.43715423e-01 5.71662605e-01 6.24987662e-01 2.86952436e-01
1.44578993e+00 -5.31101763e-01 -5.03214359e-01 6.97062314e-01
8.94397736e-01 -3.31446320e-01 -1.36112416e+00 5.61227739e-01
-1.93605833e-02 -2.26750076e-01 -3.08175292e-02 -1.01340425e+00
-1.00881934e+00 3.85423183e-01 5.79875290e-01 -3.15652400e-01
1.05999017e+00 -3.13128233e-01 7.22849905e-01 -1.56906284e-02
4.64342713e-01 -1.45278060e+00 -1.71688989e-01 9.81176496e-02
6.93915427e-01 -6.42150760e-01 7.35597014e-02 -7.49023139e-01
-3.62258047e-01 1.17188919e+00 3.82866234e-01 4.20359284e-01
4.46115404e-01 1.35854587e-01 -1.09141521e-01 -6.87030517e-03
-1.01282561e+00 1.64002672e-01 -7.49398112e-01 7.16390133e-01
7.45461404e-01 1.34697586e-01 -6.78166449e-01 3.38908225e-01
5.03688574e-01 5.00247121e-01 7.16507196e-01 7.61377156e-01
-4.21027452e-01 -1.16925645e+00 -3.15959781e-01 2.55142242e-01
-3.82632107e-01 1.83900461e-01 -5.10879397e-01 6.86298072e-01
3.58704180e-01 8.91404033e-01 -1.80444241e-01 1.49492785e-01
1.21658646e-01 7.56734550e-01 5.76118708e-01 -4.85295892e-01
-5.89184165e-01 4.36393768e-01 6.54844716e-02 -3.83342326e-01
-6.68880343e-01 -9.84697223e-01 -1.63295364e+00 -3.82673442e-01
-3.09784293e-01 1.66586503e-01 1.21857071e+00 9.04276371e-01
6.79225147e-01 6.66676581e-01 7.49433041e-01 -6.40697777e-01
3.30721647e-01 -7.06080735e-01 -8.01115811e-01 1.96997020e-02
-3.29669178e-01 -4.84341055e-01 -1.45883679e-01 4.45333451e-01] | [5.8332438468933105, 4.33102560043335] |
049ac2e5-47fc-4b8f-bf06-fb1832ba8b97 | qu-apporte-bert-a-l-analyse-syntaxique-en | null | null | https://aclanthology.org/2020.jeptalnrecital-taln.17 | https://aclanthology.org/2020.jeptalnrecital-taln.17.pdf | Qu'apporte BERT \`a l'analyse syntaxique en constituants discontinus ? Une suite de tests pour \'evaluer les pr\'edictions de structures syntaxiques discontinues en anglais (What does BERT contribute to discontinuous constituency parsing ? A test suite to evaluate discontinuous constituency structure predictions in English) | Cet article propose d{'}analyser les apports d{'}un mod{\`e}le de langue pr{\'e}-entra{\^\i}n{\'e} de type BERT (bidirectional encoder representations from transformers) {\`a} l{'}analyse syntaxique en constituants discontinus en anglais (PTB, Penn Treebank). Pour cela, nous r{\'e}alisons une comparaison des erreurs d{'}un analyseur syntaxique dans deux configurations (i) avec un acc{\`e}s {\`a} BERT affin{\'e} lors de l{'}apprentissage (ii) sans acc{\`e}s {\`a} BERT (mod{\`e}le n{'}utilisant que les donn{\'e}es d{'}entra{\^\i}nement). Cette comparaison s{'}appuie sur la construction d{'}une suite de tests que nous rendons publique. Nous annotons les phrases de la section de validation du Penn Treebank avec des informations sur les ph{\'e}nom{\`e}nes syntaxiques {\`a} l{'}origine des discontinuit{\'e}s. Ces annotations nous permettent de r{\'e}aliser une {\'e}valuation fine des capacit{\'e}s syntaxiques de l{'}analyseur pour chaque ph{\'e}nom{\`e}ne cible. Nous montrons que malgr{\'e} l{'}apport de BERT {\`a} la qualit{\'e} des analyses (jusqu{'}{\`a} 95 en F1 ), certains ph{\'e}nom{\`e}nes complexes ne sont toujours pas analys{\'e}s de mani{\`e}re satisfaisante. | ['Maximin Coavoux'] | 2020-06-01 | null | null | null | jeptalnrecital-2020-6 | ['constituency-parsing'] | ['natural-language-processing'] | [ 1.69605955e-01 5.89755952e-01 3.58089358e-01 -3.81891996e-01
-6.91698790e-01 -1.15383506e+00 4.61797714e-01 5.09046912e-01
-7.30283976e-01 9.52090204e-01 -2.28629321e-01 -5.43127418e-01
-4.01234865e-01 -1.08617783e+00 -7.34059870e-01 -4.77577031e-01
-3.66381854e-01 5.60198963e-01 4.16306674e-01 -7.22652614e-01
-1.69892590e-02 3.41708481e-01 -1.47468245e+00 3.52047205e-01
5.39020717e-01 1.21218479e+00 3.09237719e-01 5.14557064e-01
3.41668874e-02 7.71482766e-01 -6.33858204e-01 -9.60206032e-01
5.33921242e-01 -7.42900789e-01 -8.68526995e-01 -4.55718577e-01
-4.92849853e-03 4.06009763e-01 -9.97352004e-02 1.69359708e+00
-7.18762204e-02 1.67693689e-01 1.56249690e+00 -8.32354009e-01
-5.48115671e-01 8.61865401e-01 -1.30111262e-01 9.74707603e-01
3.48136663e-01 6.48214817e-02 1.48098397e+00 -6.36150599e-01
4.56483543e-01 4.62290049e-01 1.00108027e+00 4.65588391e-01
-9.37155902e-01 -8.10650408e-01 1.98147133e-01 -3.90210748e-01
-1.14542997e+00 -2.82021403e-01 9.63146538e-02 -5.24220586e-01
6.25057101e-01 4.24834639e-01 2.86940753e-01 9.05030012e-01
1.05708018e-01 6.77277029e-01 1.42761016e+00 -2.44828507e-01
1.80784017e-01 4.44628626e-01 -9.76588577e-02 9.29535806e-01
-1.17376871e-01 1.04144134e-01 8.43825936e-02 -1.59405470e-01
1.03712761e+00 -7.75059909e-02 -8.98828879e-02 8.42070103e-01
-9.07961011e-01 1.14547718e+00 2.69069493e-01 5.96698403e-01
-1.86304390e-01 8.32339972e-02 4.83583063e-01 2.61569113e-01
-4.27006297e-02 4.33947235e-01 -4.52056676e-01 -6.42604947e-01
-5.54916680e-01 5.72887421e-01 9.59290504e-01 1.30867314e+00
4.86786693e-01 1.00784183e-01 3.31449240e-01 7.77638674e-01
1.38781384e-01 8.21185559e-02 -2.93022245e-02 -9.84152019e-01
4.75237608e-01 2.98856229e-01 3.54842059e-02 -1.31956294e-01
-4.95236576e-01 -5.70761383e-01 -8.84016097e-01 3.04633200e-01
7.70496964e-01 -8.53451043e-02 -7.89941922e-02 1.94729209e+00
-5.71242392e-01 -4.45021540e-01 1.18176416e-02 7.42038488e-01
1.63306102e-01 4.02644813e-01 6.86707675e-01 -7.48384953e-01
1.92429388e+00 1.47747412e-01 3.12813707e-02 1.16905205e-01
6.26894891e-01 -1.09055305e+00 1.11721778e+00 7.13344693e-01
-1.51681602e+00 -3.51882309e-01 -9.53192234e-01 4.66945730e-02
-1.11550456e-02 2.91874111e-02 2.87970388e-03 7.88673759e-01
-8.42114329e-01 6.54189706e-01 -3.14476103e-01 1.05205722e-01
3.48019227e-02 4.25697505e-01 -3.13070476e-01 5.25994062e-01
-1.39805102e+00 8.03381503e-01 7.77520239e-01 3.64401132e-01
-8.05822670e-01 -5.69232479e-02 -9.83940542e-01 2.54348069e-01
8.57522249e-01 -8.78402516e-02 1.45194936e+00 -9.62401390e-01
-8.24416518e-01 9.57333982e-01 -1.77476704e-01 -3.34344447e-01
8.57351363e-01 4.26174879e-01 -7.75059700e-01 1.86937481e-01
1.99035361e-01 3.14165682e-01 2.26300620e-02 -1.24009538e+00
-1.28679121e+00 -1.89563721e-01 8.18894029e-01 1.04438961e-01
-2.46994168e-01 6.71349406e-01 2.07715064e-01 -1.29044008e+00
3.04753721e-01 -1.06190026e+00 3.90270710e-01 -1.83573753e-01
-5.92446148e-01 -1.10905834e-01 5.83722651e-01 -4.10089880e-01
1.46846652e+00 -1.87668931e+00 2.54002213e-01 2.87349790e-01
7.33677447e-01 7.44993508e-01 -3.24553363e-02 2.99481571e-01
-1.47272572e-01 3.25818092e-01 -9.23331082e-01 8.78919587e-02
2.65688419e-01 2.56353438e-01 2.36404374e-01 1.04328036e-01
-3.64539295e-01 2.74265021e-01 -7.33480334e-01 -2.66969562e-01
-3.35602790e-01 1.02650508e-01 -8.23244989e-01 1.26517555e-02
-7.46915817e-01 4.19802427e-01 -8.18322301e-01 -1.02178812e-01
3.98771048e-01 -2.38879576e-01 3.52366745e-01 -1.68896168e-01
-7.53253102e-01 4.67199445e-01 -1.67028594e+00 1.05481052e+00
-1.89823844e-02 -1.69674143e-01 5.88027239e-01 -1.15403283e+00
1.45288002e+00 6.11158609e-01 7.38983631e-01 -3.01874816e-01
4.45076525e-01 6.32437944e-01 9.29689705e-02 2.81593174e-01
2.35465229e-01 -6.37597680e-01 -1.63605645e-01 1.12050557e+00
-4.08457845e-01 -3.09535474e-01 6.78727984e-01 -4.38167006e-01
1.35020459e+00 3.66420984e-01 -4.90012690e-02 -6.86882079e-01
1.09240711e+00 -4.18386430e-01 7.28498697e-01 4.76850808e-01
-1.74186349e-01 1.05322385e-03 9.90850389e-01 -4.70755160e-01
-1.77716792e+00 -9.54984009e-01 -3.30108911e-01 1.15575516e+00
-4.64733869e-01 -4.71619785e-01 -9.11750555e-01 -9.06703115e-01
-5.17662346e-01 1.03148496e+00 -6.10926449e-01 1.87506229e-01
-7.48560369e-01 -8.96408319e-01 1.21570098e+00 9.60784316e-01
7.72077084e-01 -1.07079637e+00 -8.32751036e-01 3.57089788e-01
-7.11391091e-01 -9.02771413e-01 -1.47133559e-01 4.46757615e-01
-3.95658851e-01 -7.82029033e-01 -5.61580777e-01 -2.61007875e-01
3.46340418e-01 -6.55731618e-01 1.37408841e+00 -2.70896554e-01
-2.42796063e-01 -2.95334160e-01 -5.74188650e-01 -8.80332291e-01
-1.08329809e+00 1.20159693e-01 1.00574316e-02 -2.78511196e-01
4.78158027e-01 -5.71120620e-01 -9.17542696e-01 8.64895105e-01
-9.20523345e-01 -3.67286265e-01 3.46466601e-01 4.94980931e-01
6.04478419e-01 1.50636584e-02 3.66771132e-01 -9.87758100e-01
6.91523194e-01 -1.56729534e-01 -5.11469007e-01 -3.86036575e-01
-3.52410257e-01 -9.32327732e-02 1.33564639e+00 3.81168187e-01
-7.55909622e-01 -7.19700813e-01 -7.52112627e-01 -1.08030647e-01
-7.67697930e-01 5.83631396e-01 -4.37248915e-01 1.15234399e+00
8.96776617e-01 2.23624900e-01 -7.40211248e-01 -7.43677378e-01
-7.38516748e-02 3.07632655e-01 6.72012806e-01 -9.43337679e-01
4.63822693e-01 -2.57301666e-02 -1.16727494e-01 -3.65763485e-01
-3.42238367e-01 3.27541441e-01 -8.13979208e-01 7.24933669e-02
1.39565146e+00 -5.43384969e-01 -1.27317595e+00 -3.52049798e-01
-9.84009326e-01 -2.80696243e-01 -5.44788480e-01 8.32930207e-01
-8.85503948e-01 -1.07722118e-01 -9.76156771e-01 -6.78504646e-01
1.26510963e-01 -1.17519581e+00 8.50090563e-01 -9.78373066e-02
-5.65424860e-01 -7.44045496e-01 -4.20469642e-02 4.19579268e-01
-2.91796505e-01 -2.39914060e-01 1.40300310e+00 -5.39546907e-01
-7.52737284e-01 -9.56354812e-02 -2.58168012e-01 6.79544687e-01
5.94216697e-02 -1.18533492e-01 -7.60050416e-01 -2.19063088e-01
-4.33157623e-01 4.99885857e-01 1.64370060e-01 -3.00290324e-02
9.70182061e-01 -7.73389220e-01 2.74416953e-01 2.41342098e-01
1.58475637e+00 7.09631264e-01 7.29731739e-01 2.42044345e-01
-4.01348740e-01 5.96026897e-01 6.57282472e-01 7.22056702e-02
3.93814027e-01 8.43274772e-01 1.92717105e-01 5.81084311e-01
6.95700824e-01 1.14916369e-01 5.28811276e-01 9.81262803e-01
-1.12119603e+00 -3.00156951e-01 -1.16060221e+00 7.65476644e-01
-1.45754373e+00 -5.55979550e-01 -5.40615499e-01 2.16591263e+00
1.03199351e+00 6.78002000e-01 4.26225305e-01 4.60796088e-01
6.64029539e-01 1.67943120e-01 2.51570821e-01 -9.73253250e-01
-3.81216615e-01 7.84037948e-01 5.97905554e-02 5.69326699e-01
-5.44859946e-01 8.36206853e-01 3.20590162e+00 9.34165001e-01
-6.88886702e-01 1.60388052e-01 5.96925437e-01 2.60997206e-01
-7.75069237e-01 1.18496448e-01 -9.82316256e-01 8.34976971e-01
1.19041240e+00 1.42610535e-01 5.28880656e-02 5.26811182e-01
3.94425215e-03 -1.90489769e-01 -1.26915061e+00 3.98551494e-01
1.05915248e-01 -1.31335104e+00 -9.42734182e-01 -7.56342784e-02
2.17033237e-01 -1.09955490e-01 6.20234460e-02 5.51569521e-01
1.16753244e+00 -1.21544898e+00 1.08779538e+00 1.98436111e-01
1.43252885e+00 -6.57850981e-01 -4.81489599e-02 4.66405869e-01
-1.30542994e+00 -6.72848672e-02 -1.89982444e-01 -2.29989737e-02
5.16348004e-01 2.32334062e-01 -4.23498191e-02 1.11651707e+00
7.30027378e-01 -9.89905521e-02 -1.83934972e-01 4.96961474e-01
-6.20507598e-01 9.99132156e-01 -4.79130208e-01 -1.73681632e-01
6.71522737e-01 -7.45148599e-01 9.04894173e-01 1.29166424e+00
3.19134831e-01 1.09016514e+00 -7.45610967e-02 9.27405596e-01
-1.55933440e-01 3.34812582e-01 -9.15431380e-02 1.60770535e-01
-3.62319872e-02 1.12128280e-01 -1.22048938e+00 9.71901491e-02
-2.95953214e-01 4.53842461e-01 8.01810473e-02 -4.01293449e-02
-9.03580606e-01 -7.73324728e-01 5.53487360e-01 4.06023353e-01
3.06473393e-02 -3.16557229e-01 -1.23249963e-01 -6.78009152e-01
-1.70582548e-01 -8.75156045e-01 5.47375917e-01 -7.70398259e-01
-8.27070296e-01 7.50246823e-01 4.82682526e-01 -7.92460263e-01
-5.51131606e-01 -4.56847072e-01 7.15436488e-02 1.07636642e+00
-7.48036683e-01 -5.85291028e-01 6.14681840e-01 7.93694437e-01
2.03412712e-01 -4.57084686e-01 1.44583523e+00 4.65942562e-01
-2.49713197e-01 2.01249078e-01 5.21811806e-02 5.66567957e-01
2.20920116e-01 -1.35256124e+00 1.55943647e-01 5.98523676e-01
2.75156736e-01 4.74478960e-01 5.66571712e-01 -7.11537480e-01
-3.26823056e-01 -9.72417235e-01 1.33331668e+00 -4.24469858e-01
6.37135029e-01 -4.81317461e-01 -9.44710970e-01 1.84546340e+00
5.95209282e-03 -3.12736720e-01 7.57652819e-01 1.10822394e-01
-2.94072598e-01 -1.74473614e-01 -8.72743785e-01 7.07359076e-01
4.53896493e-01 -4.25328046e-01 -5.01950502e-01 5.47774173e-02
-5.70567548e-02 -1.98202312e-01 -1.47702491e+00 7.63511002e-01
1.93644106e-01 -1.46294427e+00 1.98732555e-01 -4.26281214e-01
6.30349696e-01 -4.70716476e-01 3.35698128e-02 -6.60023928e-01
8.46419334e-02 -7.25055456e-01 4.36773300e-01 1.15873373e+00
3.94535750e-01 -5.26086211e-01 5.23143411e-01 3.39579284e-01
-4.61125255e-01 1.93344653e-02 -1.47311687e+00 -1.88914910e-01
9.01165903e-01 -1.15731883e+00 5.12722075e-01 7.80020297e-01
-3.44147533e-01 5.19050611e-03 4.73810919e-02 -2.14540675e-01
2.67355829e-01 -1.75203472e-01 1.73606381e-01 -1.06557941e+00
-7.63122499e-01 -6.82986379e-01 -2.23064214e-01 -1.00833106e+00
7.40386397e-02 -1.05371916e+00 -6.90474749e-01 -1.14408290e+00
-5.79233319e-02 -9.28850174e-01 -2.81091750e-01 6.99203014e-01
3.65475416e-01 5.11819482e-01 6.08038366e-01 3.04058611e-01
-3.40774626e-01 -3.43563370e-02 1.36650789e+00 3.57861340e-01
4.36596274e-01 8.51516053e-02 -6.45835340e-01 1.26955891e+00
3.82932067e-01 -6.16298139e-01 -4.56239134e-01 -3.36436898e-01
1.02827609e+00 4.61176097e-01 1.88910440e-01 -7.72404075e-01
1.19228981e-01 -1.20786034e-01 5.51280975e-01 -3.91440451e-01
7.77704954e-01 -5.56367815e-01 2.78307140e-01 -6.76196516e-02
-3.11118364e-01 5.18258154e-01 2.78340340e-01 -1.14019573e-01
-2.06644237e-01 -5.24462640e-01 1.14271677e+00 -7.38260090e-01
-3.75199080e-01 5.02512157e-01 -8.61609340e-01 5.11768699e-01
1.04008019e+00 -4.53549445e-01 1.66497648e-01 -4.30356599e-02
-1.28797841e+00 -1.09315291e-02 1.78127617e-01 -1.00695722e-01
-1.45642519e-01 -7.79377639e-01 -1.25838137e+00 2.23381639e-01
-1.67173311e-01 4.25287187e-01 7.53609419e-01 5.08454025e-01
-1.02356088e+00 1.52068436e-01 -3.93626839e-01 -6.69837236e-01
-1.16632390e+00 2.52945185e-01 3.31508487e-01 -1.25263199e-01
-6.95025504e-01 1.47534919e+00 -2.21862510e-01 -2.00436875e-01
-4.55126077e-01 -1.58496439e-01 -2.32127439e-02 4.79208939e-02
-2.23482922e-01 5.78954160e-01 -6.87098876e-02 -1.01745331e+00
-2.15832610e-02 2.56198704e-01 -3.72704715e-02 -5.42557180e-01
8.26832235e-01 -8.26483369e-02 -6.15904212e-01 3.46972138e-01
9.69901741e-01 3.40482116e-01 -1.23266065e+00 7.59702861e-01
-2.67917126e-01 1.00504339e-01 -1.05404103e+00 -1.03904283e+00
-6.28074288e-01 8.98714602e-01 3.58210623e-01 2.47811183e-01
5.94965875e-01 -1.38303474e-01 5.05263865e-01 1.88227236e-01
6.50518477e-01 -1.52344465e+00 -5.64072490e-01 8.66530418e-01
8.14544559e-01 -5.32678187e-01 -3.71578872e-01 -3.32347840e-01
-7.18835711e-01 1.13412189e+00 -3.37844491e-01 6.87512904e-02
8.10366035e-01 3.88004690e-01 2.58202553e-01 -2.68172771e-01
-1.71910658e-01 -2.12608904e-01 -7.08636716e-02 2.48498440e-01
7.82718003e-01 7.09190816e-02 -1.04339516e+00 1.12485480e+00
-1.17733324e+00 3.11138984e-02 5.25955617e-01 1.01460958e+00
-9.22597051e-02 -1.46115685e+00 -3.32214057e-01 -3.12808007e-02
-7.59790838e-01 -1.02191716e-01 -5.96802607e-02 9.29405808e-01
7.96211183e-01 7.72639751e-01 -1.72687490e-02 -4.67530228e-02
8.81558776e-01 4.23539132e-01 4.23260361e-01 -6.80518031e-01
-1.49610722e+00 6.79163158e-01 2.94788927e-01 5.30437604e-02
-6.24281466e-01 -8.22324574e-01 -1.10270464e+00 -7.70641625e-01
2.44332358e-01 7.60891318e-01 2.14303866e-01 1.13834643e+00
-4.52835649e-01 6.09678030e-01 -5.32303117e-02 -1.35111123e-01
-6.84290886e-01 -6.09058440e-01 -1.38162899e+00 5.25409698e-01
-3.85904074e-01 -5.43216467e-01 -1.44587234e-01 1.50781617e-01] | [14.100536346435547, 13.315065383911133] |
4b09bf06-d26c-4bbf-a410-90b6ce500069 | confit-toward-faithful-dialogue-summarization-1 | null | null | https://openreview.net/forum?id=cjUocqbaCcF | https://openreview.net/pdf?id=cjUocqbaCcF | CONFIT: Toward Faithful Dialogue Summarization with Linguistically-Informed Contrastive Fine-tuning | Factual inconsistencies in generated summaries severely limit the practical applications of abstractive dialogue summarization. Although significant progress has been achieved by using pre-trained neural language models, substantial amounts of hallucinated content are found during the human evaluation. In this work, we first devised a typology of factual errors to better understand the types of hallucinations generated by current models and conducted human evaluation on popular dialog summarization dataset. We further propose a training strategy that improves the factual consistency and overall quality of summaries via a novel contrastive fine-tuning, called CONFIT. To tackle top factual errors from our annotation, we introduce additional contrastive loss with carefully designed hard negative samples and self-supervised dialogue-specific loss to capture the key information between speakers. We show that our model significantly reduces all kinds of factual errors on both SAMSum dialogue summarization and AMI meeting summarization. On both datasets, we achieve significant improvements over state-of-the-art baselines using both automatic metrics, ROUGE and BARTScore, and human evaluation. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['meeting-summarization'] | ['natural-language-processing'] | [ 1.62491024e-01 7.87275195e-01 1.92316975e-02 -5.23137629e-01
-1.40154517e+00 -4.85079050e-01 8.58423531e-01 4.00922090e-01
-2.87253112e-01 1.17591465e+00 1.03715253e+00 1.30332038e-01
3.35369855e-01 -2.98102349e-01 -4.79725391e-01 -7.91259184e-02
1.69353336e-01 6.70609117e-01 -4.54731612e-03 -5.35247445e-01
5.18660247e-01 -3.08536381e-01 -1.03581238e+00 7.58101404e-01
1.49298692e+00 5.84337831e-01 -9.90732685e-02 8.07577550e-01
-4.52110358e-02 1.18192124e+00 -1.51244450e+00 -8.32481384e-01
-2.84938395e-01 -1.01060426e+00 -1.09354687e+00 2.29390457e-01
7.33653963e-01 -4.76135403e-01 -1.96229398e-01 9.04747725e-01
1.00292754e+00 1.54811919e-01 6.81576848e-01 -9.48645532e-01
-5.65898120e-01 1.04761910e+00 -3.15142304e-01 1.55020520e-01
8.35851490e-01 2.35802993e-01 1.26805139e+00 -7.86318362e-01
5.61923206e-01 1.46932566e+00 8.49915981e-01 8.18703353e-01
-1.00540531e+00 -4.42655534e-01 -4.41820771e-02 1.94202587e-01
-7.64754117e-01 -8.45376074e-01 7.37062275e-01 -1.31932631e-01
1.18707120e+00 6.00428760e-01 3.20570976e-01 1.42945039e+00
1.14005052e-01 1.00972044e+00 6.02784157e-01 -2.95821071e-01
8.69222134e-02 3.29847872e-01 3.57727379e-01 4.82981205e-01
1.46546632e-01 -5.41308522e-01 -7.74939060e-01 -3.93643588e-01
1.09525602e-02 -5.96289814e-01 -6.20598078e-01 1.73464149e-01
-1.10910165e+00 8.98173213e-01 9.45429876e-02 1.24546271e-02
-2.90660322e-01 -3.19298148e-01 8.63944709e-01 2.23512858e-01
7.40684450e-01 9.91662204e-01 -2.98514426e-01 -4.07560557e-01
-1.20205367e+00 6.03067100e-01 1.30329609e+00 8.29968750e-01
2.03595847e-01 9.41215754e-02 -9.40629065e-01 1.29067433e+00
-1.63743615e-01 3.77427906e-01 7.30007410e-01 -1.22496009e+00
9.20730174e-01 5.45469403e-01 2.82254755e-01 -9.70777988e-01
-4.34013367e-01 -5.12690246e-01 -9.59391892e-01 -4.95609850e-01
1.93513855e-01 -3.78623813e-01 -3.29126775e-01 1.73318422e+00
-1.45873293e-01 -3.55715990e-01 4.09416407e-01 6.62999630e-01
1.21531904e+00 7.65658557e-01 -1.09531693e-01 -6.25894248e-01
1.13285542e+00 -1.38885152e+00 -1.26603186e+00 -2.78183281e-01
5.69385290e-01 -7.99317122e-01 1.44568300e+00 3.28149468e-01
-1.63575888e+00 -3.16986024e-01 -1.20169258e+00 -2.05004200e-01
2.74005115e-01 2.16704994e-01 2.66956359e-01 4.22303259e-01
-9.48444784e-01 6.94401443e-01 -4.63591367e-01 -3.19260865e-01
2.67027736e-01 -2.00081058e-02 1.97057407e-02 3.15850168e-01
-1.29813719e+00 1.02476728e+00 3.46688688e-01 -1.74462080e-01
-7.17365026e-01 -6.36353970e-01 -9.21510994e-01 2.02914774e-01
5.07168472e-01 -8.91797483e-01 1.93310452e+00 -5.72495699e-01
-1.66111529e+00 7.78249681e-01 -2.70941317e-01 -8.22886288e-01
7.44491816e-01 -5.80592811e-01 -8.91404897e-02 7.53967986e-02
2.31806636e-01 5.18971622e-01 4.52582210e-01 -1.30708945e+00
-3.28176379e-01 -1.74833480e-02 -8.00883472e-02 6.47601008e-01
-2.66750127e-01 -6.82966858e-02 -4.01403382e-02 -7.05845952e-01
-3.18272173e-01 -6.68180346e-01 -4.85140458e-02 -7.07964182e-01
-1.20137870e+00 -4.40331042e-01 4.46945012e-01 -1.01478505e+00
1.54096174e+00 -1.78907013e+00 2.69095171e-02 -5.20551503e-01
3.21997613e-01 4.44159597e-01 -8.90764371e-02 7.67070115e-01
4.12647754e-01 1.59597471e-01 -5.16867340e-01 -7.35650420e-01
9.07802582e-02 -1.53998822e-01 -7.18517065e-01 -7.48123750e-02
1.34866923e-01 8.12150717e-01 -1.04153883e+00 -4.33009028e-01
-7.90342987e-02 7.84836113e-02 -7.31144190e-01 6.25976920e-01
-4.79212493e-01 3.13669324e-01 -1.42877907e-01 2.01047003e-01
3.91165525e-01 -2.16384515e-01 2.08719429e-02 -9.73484293e-02
1.14309497e-01 1.11104703e+00 -4.84205872e-01 1.78818882e+00
-3.49897265e-01 6.93817377e-01 -9.05671269e-02 -5.98636031e-01
7.60618150e-01 4.85063195e-01 -8.28769058e-02 -5.40333092e-01
8.72230530e-02 3.44042554e-02 -1.36409357e-01 -5.72719097e-01
1.26273501e+00 -1.67669192e-01 -3.92690480e-01 5.81607521e-01
2.00636312e-01 -4.15649414e-01 1.54957056e-01 6.63536072e-01
1.09599388e+00 -3.81916225e-01 6.22767389e-01 7.90184084e-03
4.67729330e-01 1.93550467e-01 4.90009099e-01 1.08570874e+00
-2.75804102e-01 8.82790625e-01 8.07764709e-01 3.45445797e-02
-9.11885560e-01 -8.04181933e-01 1.59508348e-01 9.23305511e-01
-1.43326363e-02 -6.75806165e-01 -1.04970181e+00 -8.35886478e-01
-3.41082096e-01 1.42923951e+00 -3.53464812e-01 -2.68912435e-01
-6.11298740e-01 -8.60057414e-01 1.03341079e+00 3.19427907e-01
8.30326200e-01 -1.30375195e+00 -4.48551923e-01 2.37962618e-01
-8.73813331e-01 -1.10522509e+00 -7.44303882e-01 -2.10270435e-01
-8.15643907e-01 -8.65476191e-01 -6.17173970e-01 -4.36784357e-01
4.53325398e-02 1.80658266e-01 1.48088169e+00 -1.14987921e-02
1.25147760e-01 2.38990515e-01 -3.25748414e-01 -5.40861547e-01
-9.58771884e-01 3.00886661e-01 -9.37049612e-02 -4.33722436e-01
9.54869464e-02 -3.60038191e-01 -3.75489950e-01 -1.80964507e-02
-5.71519375e-01 1.27311975e-01 3.95162821e-01 1.07702219e+00
-1.35396540e-01 -3.27899456e-01 1.07983530e+00 -1.08450711e+00
1.59183776e+00 -3.02396983e-01 2.15489447e-01 2.99856037e-01
-4.52757359e-01 1.06998764e-01 6.58238113e-01 -1.42188415e-01
-1.37279952e+00 -7.99354494e-01 -2.94279039e-01 1.48103222e-01
5.71086369e-02 4.55543667e-01 -1.89009830e-01 7.52663016e-01
9.40212309e-01 1.26603886e-01 -7.39458203e-02 -2.80601442e-01
3.41530681e-01 1.06570065e+00 7.49613643e-01 -4.26000059e-01
2.97368199e-01 -3.96511406e-02 -8.14916015e-01 -1.04378784e+00
-1.56536639e+00 -2.90306866e-01 -1.24882005e-01 -2.15391368e-02
6.47147536e-01 -1.02468061e+00 -5.61409771e-01 3.88867080e-01
-1.70508730e+00 -2.74902672e-01 -3.50056618e-01 5.94922043e-02
-5.16602933e-01 7.99521387e-01 -8.76912951e-01 -9.53586519e-01
-9.49664295e-01 -6.72723114e-01 9.60723639e-01 1.07736573e-01
-9.72156763e-01 -8.64871025e-01 4.28896755e-01 7.24145234e-01
3.31829488e-01 1.05667293e-01 8.03151906e-01 -1.26839662e+00
-6.26880303e-02 -1.37035266e-01 -7.88441598e-02 4.72072095e-01
-1.21648526e-02 -2.95608938e-01 -9.74618673e-01 -1.87936425e-01
3.40146482e-01 -8.68583620e-01 1.01839948e+00 2.32654572e-01
8.82778943e-01 -1.11069143e+00 -5.36387926e-03 1.62863582e-02
6.69209123e-01 -8.01393464e-02 6.79530859e-01 1.22741871e-01
5.28648555e-01 9.01542187e-01 5.54233193e-01 6.83320403e-01
6.77289128e-01 6.19261444e-01 2.55441554e-02 1.80990055e-01
-1.31268471e-01 -5.27592719e-01 5.27265966e-01 1.12175632e+00
2.38320544e-01 -7.54127502e-01 -5.15538812e-01 6.76395118e-01
-2.02400374e+00 -1.22732627e+00 8.79562870e-02 2.01190233e+00
1.33025146e+00 2.87346214e-01 8.89790356e-02 -1.07959583e-01
7.84958541e-01 5.61086237e-01 -4.30893153e-01 -6.20681167e-01
-3.41799110e-01 -2.43532434e-01 -1.40889481e-01 7.16697097e-01
-9.09978271e-01 1.00095642e+00 6.23034906e+00 7.18925476e-01
-7.08537817e-01 1.16554506e-01 6.07905090e-01 -2.46093944e-01
-3.97212356e-01 -3.42446268e-01 -7.26009250e-01 4.58389670e-01
1.03273439e+00 -4.03016061e-01 6.31910115e-02 6.42358601e-01
4.32323247e-01 -6.85019325e-03 -1.17981863e+00 6.88376665e-01
6.76218212e-01 -1.21039999e+00 2.37919033e-01 -3.21498156e-01
8.94177854e-01 -2.23227605e-01 -2.35813335e-01 6.29265964e-01
3.78642976e-01 -9.20645356e-01 6.07812405e-01 2.85060942e-01
5.48371017e-01 -5.51935077e-01 9.65870500e-01 5.54789543e-01
-1.86542153e-01 1.22766785e-01 -3.90746236e-01 -1.64470211e-01
3.40847373e-01 5.77903688e-01 -1.35743594e+00 4.71948117e-01
1.74870417e-01 6.11281037e-01 -5.15013993e-01 8.20615113e-01
-4.77286816e-01 7.21348286e-01 7.52181262e-02 -2.02703804e-01
1.26661554e-01 2.29133025e-01 1.03261828e+00 1.51410055e+00
1.68393850e-02 3.84103209e-02 1.28641307e-01 9.08652544e-01
-5.36604345e-01 6.84754774e-02 -4.88965362e-01 2.68016178e-02
7.16574669e-01 1.05402398e+00 5.95795065e-02 -6.72854781e-01
5.68983592e-02 1.14886928e+00 5.34493089e-01 2.03909069e-01
-6.88419819e-01 -3.90218735e-01 2.05352664e-01 -1.15858369e-01
-2.91006826e-02 1.43012822e-01 -4.87805486e-01 -1.38083863e+00
1.55342072e-01 -1.17506468e+00 3.41936439e-01 -5.75503051e-01
-1.38902915e+00 7.92323709e-01 -1.45908713e-01 -9.75105882e-01
-6.73018694e-01 1.80243149e-01 -1.09062552e+00 5.36932588e-01
-1.26125121e+00 -7.99708724e-01 -3.19094330e-01 4.28341143e-02
1.17218137e+00 -2.39699215e-01 8.61946464e-01 -8.70421007e-02
-7.00123191e-01 9.03777957e-01 -1.44740641e-01 -4.18945514e-02
1.26001346e+00 -1.38594544e+00 4.18135226e-01 6.57215178e-01
-2.07934812e-01 5.49869061e-01 1.35417986e+00 -7.78603852e-01
-5.92734873e-01 -9.84944046e-01 1.21166623e+00 -5.48616707e-01
4.03314710e-01 -1.62070364e-01 -1.06838131e+00 6.64693892e-01
9.88972545e-01 -1.05102837e+00 7.24494338e-01 3.17004085e-01
-1.61307171e-01 3.36004943e-01 -1.10417628e+00 7.88426101e-01
8.92369151e-01 -3.00709516e-01 -1.32868135e+00 5.92871785e-01
9.33032036e-01 -5.62585056e-01 -3.91591489e-01 4.20465887e-01
2.35694304e-01 -1.15963590e+00 6.44715965e-01 -7.10109770e-01
9.89672542e-01 1.93929419e-01 1.70208588e-01 -1.73430467e+00
1.68973491e-01 -9.66786504e-01 -1.99891210e-01 1.49128699e+00
5.75515628e-01 -2.40769506e-01 5.60057819e-01 6.96297348e-01
-6.66732490e-01 -5.65977216e-01 -6.69010103e-01 -5.20697713e-01
2.03847051e-01 4.34119292e-02 2.79827565e-01 9.41740572e-01
7.45958567e-01 1.13300157e+00 -6.93024814e-01 -3.33778679e-01
5.42845845e-01 -1.02723569e-01 9.83101010e-01 -9.90137637e-01
-1.68550819e-01 -5.31680286e-01 1.65959284e-01 -1.26976967e+00
3.99380147e-01 -6.53748155e-01 4.08365130e-01 -1.74222577e+00
5.24433196e-01 2.56365627e-01 3.05836916e-01 2.64392555e-01
-5.33799469e-01 -1.43439874e-01 1.96578145e-01 2.48902068e-01
-1.10505128e+00 1.17020965e+00 1.16943419e+00 -2.27171436e-01
-5.19688785e-01 2.32648272e-02 -1.11671472e+00 7.59945512e-01
7.68246949e-01 -2.84950137e-01 -2.29540914e-01 -3.18246216e-01
-2.36025974e-02 5.47020316e-01 2.55653679e-01 -1.01233625e+00
3.11977714e-01 1.69110909e-01 -5.55735268e-02 -8.63257945e-01
4.09586191e-01 8.34149420e-02 -5.30104697e-01 1.73559219e-01
-1.13168800e+00 -1.10857584e-01 9.23468322e-02 4.97718811e-01
-4.37477440e-01 -3.44223380e-01 5.68190217e-01 -2.66916335e-01
-4.24846299e-02 -4.78971899e-01 -3.82356405e-01 7.68378615e-01
4.47910428e-01 1.03781752e-01 -8.13056886e-01 -1.05786335e+00
-4.43662763e-01 4.95141864e-01 3.52499664e-01 2.49477133e-01
5.86346269e-01 -9.33765531e-01 -1.17632556e+00 -3.31363529e-01
3.31809558e-02 -2.71584801e-02 3.56809229e-01 6.98017478e-01
-2.01456398e-01 6.65188134e-01 -2.12514400e-02 -2.92987496e-01
-1.24908078e+00 6.55488251e-03 2.41655946e-01 -7.66926050e-01
-6.73064590e-01 7.84505963e-01 1.70083359e-01 -6.86877549e-01
5.58566689e-01 -2.36929014e-01 -3.05848420e-01 1.81018904e-01
7.71750331e-01 7.17965782e-01 1.31968617e-01 -3.46102864e-01
-3.52382809e-02 -1.89945683e-01 -5.21332264e-01 -3.68807524e-01
9.60218906e-01 -3.86757582e-01 3.69592607e-02 5.62002420e-01
8.84078562e-01 3.23829144e-01 -1.13430274e+00 -1.81980744e-01
5.58051206e-02 -1.55467866e-02 -3.69835913e-01 -1.33729577e+00
-3.38696927e-01 8.65743577e-01 -2.53903538e-01 4.74728435e-01
7.43950725e-01 -1.26642734e-01 1.18659949e+00 8.42196047e-01
-1.11853182e-01 -1.13646996e+00 4.43369806e-01 9.63744342e-01
1.38346910e+00 -1.39187813e+00 7.99512863e-02 -3.74008238e-01
-1.35383964e+00 7.39010334e-01 7.23710179e-01 4.25582714e-02
-3.04267555e-01 -1.07491747e-01 2.06811622e-01 -2.20906496e-01
-1.02264106e+00 2.94748574e-01 3.15298378e-01 1.59323752e-01
7.54224956e-01 -3.53529640e-02 -5.27372003e-01 1.04954338e+00
-7.92769372e-01 -3.42620045e-01 9.87610877e-01 4.81844753e-01
-6.71435356e-01 -6.13853335e-01 -6.86726570e-02 5.54121196e-01
-6.27357721e-01 -2.52030909e-01 -9.38661993e-01 5.45166969e-01
-7.47700810e-01 1.30717695e+00 -1.21104024e-01 -2.23706231e-01
6.39299095e-01 2.21239373e-01 3.33785117e-01 -8.64492178e-01
-1.03973794e+00 -1.69418797e-01 8.14824939e-01 -2.26452321e-01
-2.02594995e-01 -5.09390533e-01 -1.12618518e+00 -3.21527869e-01
-4.03162688e-01 5.00007391e-01 2.82885343e-01 9.36156511e-01
4.48316544e-01 6.89019561e-01 5.25713384e-01 -7.56715834e-01
-1.12502372e+00 -1.67653239e+00 -1.81520745e-01 6.18719637e-01
4.12564993e-01 -2.19451323e-01 -7.17992663e-01 -1.74657419e-01] | [12.349945068359375, 9.224235534667969] |
bd933a94-69c7-4561-a669-fb57eff41a63 | emocaps-emotion-capsule-based-model-for | 2203.13504 | null | https://arxiv.org/abs/2203.13504v1 | https://arxiv.org/pdf/2203.13504v1.pdf | EmoCaps: Emotion Capsule based Model for Conversational Emotion Recognition | Emotion recognition in conversation (ERC) aims to analyze the speaker's state and identify their emotion in the conversation. Recent works in ERC focus on context modeling but ignore the representation of contextual emotional tendency. In order to extract multi-modal information and the emotional tendency of the utterance effectively, we propose a new structure named Emoformer to extract multi-modal emotion vectors from different modalities and fuse them with sentence vector to be an emotion capsule. Furthermore, we design an end-to-end ERC model called EmoCaps, which extracts emotion vectors through the Emoformer structure and obtain the emotion classification results from a context analysis model. Through the experiments with two benchmark datasets, our model shows better performance than the existing state-of-the-art models. | ['Yusen Zhu', 'Ming Zhao', 'Fengxiao Tang', 'Zaijing Li'] | 2022-03-25 | null | https://aclanthology.org/2022.findings-acl.126 | https://aclanthology.org/2022.findings-acl.126.pdf | findings-acl-2022-5 | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [-3.02941978e-01 -3.67461979e-01 3.36577356e-01 -8.29036117e-01
-2.48042047e-01 -2.25515798e-01 1.47852406e-01 -1.52192116e-01
-3.85906488e-01 1.22195534e-01 8.38123322e-01 9.35123339e-02
5.25641561e-01 -3.25890571e-01 2.17964396e-01 -4.73128200e-01
3.80474061e-01 -4.61458266e-01 -2.65493840e-01 -4.71305072e-01
2.54890531e-01 -5.92516549e-02 -1.54832554e+00 7.84560442e-01
8.45867634e-01 1.43314672e+00 -1.27879754e-01 7.93937624e-01
-6.74622297e-01 1.19916701e+00 -7.14311242e-01 -4.57172364e-01
-6.77529037e-01 -5.03975928e-01 -7.19395339e-01 2.53788620e-01
-5.03699183e-01 1.33357272e-01 -4.95233759e-02 9.20339882e-01
8.21459055e-01 4.27433670e-01 5.25510192e-01 -1.12638879e+00
-4.88733500e-01 2.66588271e-01 -2.48402745e-01 7.32518807e-02
8.07482541e-01 -3.35378945e-01 5.01351058e-01 -1.11138701e+00
5.37382662e-01 1.27098858e+00 5.48902035e-01 8.21412444e-01
-4.23216701e-01 -3.47415268e-01 4.87223923e-01 4.62989748e-01
-9.89129364e-01 -7.19425976e-01 1.49045134e+00 -3.54153402e-02
1.04163980e+00 6.23028576e-01 4.69895005e-01 1.29353583e+00
1.56817347e-01 1.23363042e+00 1.36939764e+00 -3.84893268e-01
3.79469007e-01 4.47094977e-01 3.96546990e-01 4.06043857e-01
-1.09301364e+00 -4.13995355e-01 -5.65265000e-01 -2.60751516e-01
-2.25994974e-01 1.61335766e-01 -2.85255939e-01 4.22565877e-01
-6.35130167e-01 6.67712331e-01 1.54228439e-03 4.96512085e-01
-7.57660329e-01 -2.38607004e-01 1.09090984e+00 4.65489447e-01
8.30342352e-01 -1.43281505e-01 -5.70680141e-01 -7.92963326e-01
-2.84885049e-01 -3.38602215e-01 1.05078733e+00 7.26218224e-01
4.04316425e-01 1.56634301e-01 -2.37075955e-01 1.29755020e+00
5.08248150e-01 4.09135312e-01 7.56123126e-01 -6.22359216e-01
4.03030038e-01 8.78642201e-01 -7.93699846e-02 -1.31952894e+00
-5.53371489e-01 -1.53718129e-01 -7.14683890e-01 -2.99524575e-01
-6.20266259e-01 -8.86120796e-01 -3.13301027e-01 1.69799709e+00
5.85253596e-01 3.78983736e-01 6.43787920e-01 8.40811670e-01
1.44149899e+00 9.87360775e-01 5.13990700e-01 -4.75726604e-01
1.74970782e+00 -9.39905584e-01 -1.46856773e+00 -2.46429756e-01
6.25180364e-01 -7.72817075e-01 1.03961635e+00 4.51635748e-01
-7.04656124e-01 -4.52717900e-01 -8.31908762e-01 -4.84313583e-03
-7.30463266e-01 4.10062730e-01 6.69274926e-01 7.50521362e-01
-5.11824906e-01 5.11741713e-02 -5.51571965e-01 -1.15515850e-01
-1.55292032e-02 -2.49919575e-02 -4.18443531e-01 3.52174848e-01
-1.44799888e+00 9.44588065e-01 9.79960486e-02 5.33423960e-01
-1.77292243e-01 -1.15417026e-01 -1.29084599e+00 2.13220313e-01
1.92152008e-01 -1.00446865e-01 1.06917799e+00 -1.37899244e+00
-2.19690013e+00 3.87028068e-01 -7.15696871e-01 4.51550603e-01
-2.20081016e-01 -1.27580062e-01 -1.20015812e+00 6.01557270e-02
-6.84867918e-01 1.00261629e-01 5.55594265e-01 -1.18759406e+00
-6.00676239e-01 -4.04486626e-01 -2.90548772e-01 4.79128391e-01
-7.73290575e-01 8.06537747e-01 -7.04452157e-01 -3.58732462e-01
-8.24626684e-02 -6.09494090e-01 -2.44130269e-01 -8.87444198e-01
-1.50187314e-01 -5.20787776e-01 1.18672502e+00 -6.55734718e-01
1.63811266e+00 -2.37418580e+00 2.48945028e-01 -8.13352168e-02
-1.79180671e-02 1.19316027e-01 -1.99069753e-01 4.59485054e-01
-2.32349783e-01 -1.57841444e-02 1.42605394e-01 -8.21672082e-01
2.86754638e-01 1.35901332e-01 -2.22365871e-01 1.63297430e-01
1.90789625e-01 7.89563894e-01 -6.28438771e-01 -8.44547391e-01
2.97473490e-01 6.74555004e-01 -4.13249761e-01 7.09273994e-01
2.84256220e-01 2.32463524e-01 -6.19670451e-01 5.65286279e-01
7.08112955e-01 1.97591782e-01 1.79515839e-01 -2.64194369e-01
7.21667660e-03 -3.67714055e-02 -1.11581397e+00 1.56443226e+00
-9.57666397e-01 4.83096808e-01 3.73794585e-01 -7.21269250e-01
1.10211349e+00 7.57292151e-01 2.89796859e-01 -4.57872450e-01
8.51636648e-01 -2.70962387e-01 -3.08016062e-01 -1.15437663e+00
6.27460718e-01 -3.21151644e-01 -6.24798119e-01 2.93764800e-01
2.93670803e-01 1.54339626e-01 -3.08232099e-01 3.01968046e-02
7.45216906e-01 -2.45594487e-01 9.58334208e-02 1.26137093e-01
1.08739984e+00 -6.53116167e-01 9.23683584e-01 -1.32755548e-01
-5.17258823e-01 7.61754587e-02 5.04205406e-01 -9.22532752e-02
-1.18564531e-01 -4.21444595e-01 8.43718573e-02 1.34045827e+00
9.25576501e-03 -7.59242833e-01 -7.70634949e-01 -8.67107749e-01
-5.72946727e-01 7.56487370e-01 -8.34468544e-01 -3.06735337e-01
-2.32246108e-02 -7.11888969e-01 2.45389044e-01 6.24276459e-01
5.21843672e-01 -1.31034243e+00 -3.77003849e-01 2.14302436e-01
-6.80806816e-01 -1.25050974e+00 -5.21378338e-01 7.84312710e-02
-4.23010439e-01 -5.51639855e-01 -1.71972364e-01 -7.35087216e-01
3.80733699e-01 -2.71363765e-01 9.38822091e-01 -1.27879798e-01
2.12277249e-02 6.73240542e-01 -8.48206580e-01 -6.82199419e-01
-1.49691895e-01 -2.04521999e-01 2.66139917e-02 6.71052694e-01
9.27559137e-01 -5.01794755e-01 -5.44324934e-01 1.98248327e-01
-7.00663686e-01 -1.51073352e-01 8.45348388e-02 5.66751719e-01
2.87888497e-01 1.52152643e-01 9.71223116e-01 -7.06640482e-01
1.15505874e+00 -7.24665105e-01 2.49260023e-01 3.54006499e-01
-1.47272751e-01 -3.51409316e-01 7.99934149e-01 -5.40094078e-01
-1.78517282e+00 -1.79464894e-03 -5.88174343e-01 -4.91957277e-01
-4.34064895e-01 9.15200531e-01 -5.44557393e-01 2.72094935e-01
-2.70412099e-02 3.40678543e-01 -2.70626754e-01 -4.40514147e-01
4.97405320e-01 1.51041615e+00 4.91150588e-01 -5.48150361e-01
-2.21787378e-01 1.92742258e-01 -6.08475864e-01 -9.38357055e-01
-7.21317947e-01 -6.07131600e-01 -1.12565123e-01 -7.71268666e-01
1.02794755e+00 -8.22100341e-01 -1.34852350e+00 4.40817058e-01
-1.24295318e+00 2.18252018e-01 2.59935588e-01 6.40569210e-01
-3.09340447e-01 4.80444491e-01 -8.79007638e-01 -1.57162893e+00
-7.32291043e-01 -9.31356311e-01 1.04426718e+00 6.98148668e-01
-3.68068963e-01 -1.11475456e+00 3.28482427e-02 1.77753225e-01
5.33144176e-01 9.27702785e-02 3.91516954e-01 -7.10413456e-01
6.77079320e-01 -4.85948384e-01 1.41257912e-01 5.89404702e-01
4.55522351e-02 1.03288591e-01 -1.20493782e+00 3.58912796e-01
5.70501626e-01 -4.41171557e-01 5.62460840e-01 -2.05542088e-01
1.36300123e+00 -6.84461743e-02 -2.22837995e-03 3.00273776e-01
1.09289849e+00 5.43756485e-01 8.10927927e-01 -3.33427936e-01
3.31298888e-01 9.22725081e-01 6.30574942e-01 6.82297766e-01
6.93738759e-01 2.57443100e-01 1.78412721e-01 1.16275378e-01
5.79803765e-01 -3.32324989e-02 6.06476247e-01 1.73189151e+00
3.45677510e-02 -3.45193297e-01 -4.04977798e-01 2.56078988e-01
-1.95018029e+00 -1.12264085e+00 -1.36470973e-01 1.25815642e+00
7.29727209e-01 -1.96068287e-01 -2.07477614e-01 -4.06580754e-02
6.31196082e-01 3.46799940e-01 -3.79347175e-01 -1.10234058e+00
-1.34983525e-01 -5.92579246e-02 -4.71115410e-01 4.12982166e-01
-1.26469469e+00 9.66746449e-01 6.11224937e+00 8.92027915e-01
-1.34571242e+00 2.93255568e-01 7.33614206e-01 -1.33347929e-01
-3.07601720e-01 -4.04062122e-01 -4.49575365e-01 6.40845835e-01
1.25001085e+00 8.43579993e-02 2.55118012e-01 1.23582506e+00
2.63823569e-01 8.26722011e-03 -6.36887610e-01 1.56663990e+00
6.10462606e-01 -6.63143814e-01 -6.72607541e-01 -4.41638321e-01
2.73389250e-01 -2.95886427e-01 -1.11626409e-01 9.90870833e-01
-5.30512116e-05 -7.00336754e-01 2.95469701e-01 9.52997684e-01
4.36425894e-01 -9.98204410e-01 1.14785516e+00 1.93572551e-01
-1.25855553e+00 -7.93913528e-02 -2.20113739e-01 -2.51553595e-01
3.44734997e-01 2.36333817e-01 -1.28125817e-01 6.76632166e-01
5.69111705e-01 6.78260088e-01 -4.34912145e-01 3.38655442e-01
-7.71525577e-02 7.74524748e-01 -2.45293900e-01 -6.85420811e-01
1.39752418e-01 -3.43479216e-01 2.56519824e-01 1.76193273e+00
1.76912278e-01 5.86173296e-01 -3.48873599e-03 4.33531940e-01
-1.07405610e-01 7.64571488e-01 -2.06318811e-01 -1.78253144e-01
4.40430909e-01 1.84968650e+00 -2.73220301e-01 -4.44881231e-01
-5.67629039e-01 1.44443798e+00 2.33381256e-01 3.29423964e-01
-9.57992077e-01 -1.03125584e+00 7.11120963e-01 -1.21853101e+00
2.04991132e-01 2.08212182e-01 -2.15078756e-01 -1.43255210e+00
5.75270578e-02 -8.21704507e-01 4.18259263e-01 -1.06031406e+00
-1.47445786e+00 1.00337195e+00 -5.80375493e-01 -1.10674536e+00
-1.38531402e-01 -6.07318103e-01 -1.36795187e+00 1.02479923e+00
-1.17201614e+00 -9.58232045e-01 -3.14755797e-01 6.98209703e-01
6.11593723e-01 -7.50139207e-02 1.11540627e+00 3.22170466e-01
-9.30186808e-01 5.84165752e-01 -1.60724610e-01 2.94580221e-01
6.25573695e-01 -1.10100162e+00 -3.70358586e-01 5.08148432e-01
-3.76942247e-01 7.50953317e-01 7.04253733e-01 -4.39619482e-01
-1.67285192e+00 -7.42499650e-01 1.00878155e+00 -4.01449919e-01
5.01055539e-01 -4.46424216e-01 -7.30063140e-01 2.73414105e-01
7.80831516e-01 -4.77146320e-02 1.42303669e+00 6.24486148e-01
-8.54134932e-02 3.52825485e-02 -1.02416337e+00 6.97460532e-01
5.11639893e-01 -8.65270138e-01 -8.91137719e-01 -3.56129199e-01
7.42726803e-01 -2.76226997e-01 -1.07358849e+00 3.03084433e-01
5.74219048e-01 -7.40523100e-01 5.17514348e-01 -1.03506052e+00
6.94990873e-01 -8.88842270e-02 -5.18881738e-01 -1.56599391e+00
1.42281211e-03 -4.56635803e-01 -1.83569089e-01 1.82392323e+00
1.93421721e-01 -4.49714899e-01 4.41508740e-01 1.02305913e+00
-2.24878639e-01 -9.76034224e-01 -8.83750796e-01 -1.03226006e-01
-2.93864846e-01 -9.29885983e-01 7.73510039e-01 1.11795342e+00
8.61317813e-01 9.33396637e-01 -6.36443973e-01 -8.34742859e-02
-1.77344233e-01 2.15464041e-01 5.16635954e-01 -8.85991633e-01
7.44098946e-02 -4.23675746e-01 -2.68229723e-01 -8.90229523e-01
7.52523541e-01 -5.54523885e-01 1.69772848e-01 -1.28804421e+00
4.21908088e-02 2.53723294e-01 -4.87042814e-01 2.28180319e-01
-5.97643375e-01 -2.42807180e-01 1.60848334e-01 -5.86764932e-01
-1.09881163e+00 1.26462340e+00 9.44069326e-01 5.55663556e-02
-3.83178771e-01 -3.24383259e-01 -8.68206859e-01 8.88588011e-01
6.50227427e-01 -4.51565087e-02 -2.06311643e-01 2.83933878e-02
2.37394273e-01 4.32137430e-01 -1.17011592e-01 -6.01581514e-01
2.90658534e-01 -1.77493036e-01 4.87388760e-01 -8.27861190e-01
8.16208482e-01 -1.12592638e+00 -3.39164287e-01 -2.63307184e-01
-4.17791933e-01 2.59034242e-02 1.28820509e-01 4.34322774e-01
-5.73782980e-01 -3.21592182e-01 2.44507670e-01 2.19268277e-01
-7.62834191e-01 5.43610230e-02 -8.39974761e-01 -1.22443661e-01
6.79594874e-01 3.43165517e-01 -3.10750961e-01 -6.36762142e-01
-9.93397355e-01 5.78445613e-01 -1.83191076e-01 6.79474950e-01
7.52380848e-01 -1.59360528e+00 -4.65108722e-01 -8.93608332e-02
2.24868625e-01 -7.38538265e-01 1.10223436e+00 8.34309459e-01
1.84785992e-01 1.13179255e-02 1.22794181e-01 -2.22902857e-02
-1.61115873e+00 7.33946979e-01 4.44985360e-01 -1.97799690e-02
-1.05146103e-01 6.55113697e-01 -1.46399423e-01 -7.39540517e-01
2.29542732e-01 5.01580834e-02 -9.04313624e-01 3.65288705e-01
9.13395762e-01 1.38892189e-01 -1.98360048e-02 -1.00783777e+00
-5.78122258e-01 3.84362608e-01 1.65232554e-01 -2.27089033e-01
1.26755154e+00 -5.51142991e-01 -3.63449931e-01 7.89165854e-01
1.61706972e+00 3.09064656e-01 -6.00122273e-01 -1.90661624e-01
-1.52586058e-01 -2.16168404e-01 2.71676362e-01 -9.28255320e-01
-1.09186876e+00 1.09375274e+00 7.02269733e-01 1.68592602e-01
1.43968344e+00 -1.60479739e-01 9.07152534e-01 5.00503242e-01
-1.28248602e-01 -1.74417603e+00 1.97917270e-03 7.90817261e-01
1.04940999e+00 -1.19329500e+00 -5.51418722e-01 -3.27972531e-01
-1.44753373e+00 1.14989483e+00 7.86529183e-01 1.92108914e-01
1.18960202e+00 2.91289538e-01 6.23153210e-01 -2.78223068e-01
-1.14528644e+00 -3.28154743e-01 1.85008541e-01 3.44133139e-01
6.27253354e-01 8.84177536e-02 -5.29505670e-01 1.80301595e+00
-6.67788088e-02 7.22769462e-03 1.90502316e-01 8.65815282e-01
-2.34071985e-01 -1.02665842e+00 -2.56355822e-01 1.21264711e-01
-6.93829894e-01 4.25134897e-02 -5.97262204e-01 -5.30267544e-02
3.77771109e-02 1.84476829e+00 -9.62993689e-03 -1.10337484e+00
6.06829107e-01 8.31433892e-01 -1.79157540e-01 -2.21442804e-01
-9.53035712e-01 3.36870402e-01 5.84252119e-01 -5.27428389e-01
-6.31676793e-01 -5.37894011e-01 -1.28707111e+00 2.11553951e-03
-3.04616183e-01 5.70504665e-01 8.60793173e-01 1.01765358e+00
7.17759430e-01 8.08879256e-01 1.30151367e+00 -6.29907548e-01
1.15520079e-02 -1.40473223e+00 -5.59593022e-01 7.23962784e-01
-3.78228053e-02 -5.03045976e-01 -4.16641027e-01 -2.14891151e-01] | [13.069226264953613, 5.930931091308594] |
44a1e089-012b-4f7e-b1ed-9d0ec5e01d2b | automating-cell-counting-in-fluorescent | null | null | https://www.nature.com/articles/s41598-021-01929-5#Abs1 | https://rdcu.be/cB1Ds | Automating cell counting in fluorescent microscopy through deep learning with c-ResUnet | Counting cells in fluorescent microscopy is a tedious, time-consuming task that researchers have to accomplish to assess the effects of different experimental conditions on biological structures of interest. Although such objects are generally easy to identify, the process of manually annotating cells is sometimes subject to fatigue errors and suffers from arbitrariness due to the operator’s interpretation of the borderline cases. We propose a Deep Learning approach that exploits a fully-convolutional network in a binary segmentation fashion to localize the objects of interest. Counts are then retrieved as the number of detected items. Specifically, we introduce a Unet-like architecture, cell ResUnet (c-ResUnet), and compare its performance against 3 similar architectures. In addition, we evaluate through ablation studies the impact of two design choices, (i) artifacts oversampling and (ii) weight maps that penalize the errors on cells boundaries increasingly with overcrowding. In summary, the c-ResUnet outperforms the competitors with respect to both detection and counting metrics (respectively, F1 score = 0.81 and MAE = 3.09). Also, the introduction of weight maps contribute to enhance performances, especially in presence of clumping cells, artifacts and confounding biological structures. Posterior qualitative assessment by domain experts corroborates previous results, suggesting human-level performance inasmuch even erroneous predictions seem to fall within the limits of operator interpretation. Finally, we release the pre-trained model and the annotated dataset to foster research in this and related fields. | ['Antonio', 'Fabio and Zoccoli', 'Lorenzo and Squarcio', 'Marco and Rinaldi', 'Timna and Luppi', 'Matteo and Hitrec', 'Roberto and Cerri', 'Luca and Amici', 'Roberto and Clissa', 'Morelli'] | 2021-11-25 | null | null | null | scientific-reports-2021-11 | ['object-counting', '2d-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 3.38311911e-01 -1.11658446e-01 5.07264495e-01 -1.67781487e-01
-5.36181390e-01 -7.52817929e-01 4.46706831e-01 6.44048929e-01
-1.14357626e+00 9.95538712e-01 -3.23076159e-01 -1.98848903e-01
1.87833190e-01 -4.17680591e-01 -7.85528481e-01 -9.84989762e-01
7.87339360e-02 4.84369338e-01 2.39821985e-01 3.32675695e-01
5.46292126e-01 7.21697152e-01 -1.37976718e+00 2.82614045e-02
8.09291184e-01 9.02019501e-01 3.34923357e-01 7.72758067e-01
-1.23257218e-02 5.22848308e-01 -8.92258286e-01 -4.31319088e-01
2.32278649e-02 -1.19008921e-01 -6.43283963e-01 3.62175936e-03
4.80539322e-01 -1.83486685e-01 4.02319759e-01 9.34712589e-01
5.71787834e-01 -2.75371104e-01 8.60982478e-01 -9.48370993e-01
-4.40695405e-01 1.48429051e-01 -6.98576868e-01 6.61089301e-01
-2.46297233e-02 7.01661229e-01 5.76498568e-01 -8.47109735e-01
7.00017512e-01 9.50748026e-01 8.05203021e-01 4.16250437e-01
-1.55398905e+00 -4.75714236e-01 -1.15022846e-02 -1.99056581e-01
-1.46653819e+00 -3.03244650e-01 4.38765362e-02 -7.49013543e-01
7.35020876e-01 2.36411497e-01 4.96476889e-01 1.14617813e+00
2.19927818e-01 3.91196311e-01 1.30929244e+00 -2.07820237e-01
4.32074100e-01 2.39088356e-01 -1.14661314e-01 4.39094752e-01
7.36547470e-01 -2.44749263e-01 -1.35719746e-01 -6.55451119e-02
8.87497246e-01 -5.31048886e-02 -2.63957173e-01 -1.55444801e-01
-1.36041093e+00 4.94489074e-01 2.32556701e-01 5.73403358e-01
-2.19746992e-01 1.62600592e-01 4.63891745e-01 -3.05333853e-01
3.00498605e-01 8.65351439e-01 -4.64042962e-01 1.37169808e-01
-1.10293615e+00 2.08095193e-01 4.21414137e-01 5.61422229e-01
3.90959024e-01 -2.28879094e-01 -4.49629724e-01 6.60271466e-01
-1.68562625e-02 1.08986668e-01 4.14949208e-01 -9.65345800e-01
-1.97584722e-02 8.75106633e-01 4.57720995e-01 -7.26185322e-01
-6.91031039e-01 -7.00258851e-01 -8.22918177e-01 3.93290579e-01
1.17307699e+00 -1.32744193e-01 -9.43183482e-01 1.59957802e+00
1.22997157e-01 2.43408140e-02 -3.33947480e-01 1.08133292e+00
5.82344532e-01 1.76749527e-01 4.92868036e-01 -1.08156204e-01
1.57529485e+00 -6.24065220e-01 -6.41082108e-01 -1.25154689e-01
6.96708262e-01 -6.67621374e-01 1.28234994e+00 3.98315281e-01
-1.04091418e+00 -3.81484330e-01 -1.07085896e+00 -8.30945969e-02
-6.69002771e-01 4.46538717e-01 5.02865255e-01 5.71388364e-01
-1.00973606e+00 6.91911578e-01 -7.72452533e-01 -6.01274252e-01
1.06874514e+00 5.44349551e-01 -3.15805048e-01 3.78998339e-01
-6.21142566e-01 9.48524296e-01 1.82260081e-01 2.78992265e-01
-8.42675030e-01 -8.07243228e-01 -4.32784945e-01 1.09780133e-01
1.07560985e-01 -7.29318202e-01 8.20829332e-01 -6.12884998e-01
-1.08122277e+00 1.42081773e+00 -2.05350161e-01 -5.63959599e-01
9.36841965e-01 -5.41307218e-02 1.29032969e-01 1.05806172e-01
1.32587075e-01 8.86397004e-01 2.12743953e-01 -1.22458231e+00
-7.00139284e-01 -4.20531511e-01 -7.29742199e-02 -1.32905766e-01
-1.76338792e-01 -7.09330142e-02 -2.13720024e-01 -4.56737220e-01
-3.11022520e-01 -6.37758970e-01 -3.73740137e-01 4.09398586e-01
-3.11927468e-01 -5.08975871e-02 5.73092639e-01 -3.19433600e-01
9.53036666e-01 -2.07834458e+00 -3.55440438e-01 -4.92001958e-02
4.68740374e-01 3.77532601e-01 4.23731245e-02 1.01430111e-01
1.40656069e-01 4.29297388e-01 -3.05169672e-01 -4.36783552e-01
-2.06722662e-01 -2.06637189e-01 7.13450462e-02 7.90543735e-01
5.87934792e-01 8.93016398e-01 -9.06541944e-01 -6.91295624e-01
2.96513885e-01 6.88761711e-01 -3.71241808e-01 -8.39582980e-02
-2.07131580e-02 6.97226167e-01 -6.18278980e-02 7.14133084e-01
6.88460112e-01 -6.29945219e-01 -4.72646952e-03 -2.99273193e-01
-3.67239535e-01 -7.03009367e-02 -9.68230188e-01 1.24530268e+00
-2.48440221e-01 5.75456858e-01 1.96912259e-01 -6.72840655e-01
7.30316281e-01 1.23518743e-01 2.99300045e-01 -4.82896507e-01
5.81752598e-01 5.46105981e-01 2.40030795e-01 -4.23008353e-01
2.99423963e-01 -2.06836656e-01 2.43058458e-01 1.76938206e-01
1.06913544e-01 -1.07003953e-02 5.18272579e-01 1.15491465e-01
1.10725677e+00 -4.20632325e-02 3.93700242e-01 -7.14517713e-01
5.15456975e-01 -2.08179966e-01 4.17911977e-01 7.06314862e-01
-5.63604355e-01 8.91945302e-01 7.58984208e-01 -6.15934968e-01
-1.02346158e+00 -7.13264942e-01 -5.06260037e-01 7.74323285e-01
1.72067747e-01 2.01056570e-01 -9.82579291e-01 -4.89407182e-01
-1.50315529e-02 5.28783560e-01 -1.05551672e+00 1.78581432e-01
-3.76701593e-01 -1.09494746e+00 6.70487881e-01 5.50108612e-01
3.48200887e-01 -1.03423297e+00 -1.36656141e+00 1.63608730e-01
-2.51291096e-02 -1.21015966e+00 -2.63706803e-01 5.07471740e-01
-6.41405225e-01 -1.17538846e+00 -9.71642137e-01 -5.94405413e-01
8.54027808e-01 -3.78198735e-02 1.01781166e+00 4.39831674e-01
-4.54757750e-01 -1.87289685e-01 -1.83757156e-01 -6.86819315e-01
-1.31591961e-01 8.08131844e-02 -1.99248821e-01 -2.11803764e-01
5.78829288e-01 -3.77723038e-01 -9.69937861e-01 4.71431822e-01
-8.83322239e-01 -1.57859042e-01 5.50169706e-01 8.42591107e-01
7.78808475e-01 -2.02644944e-01 4.71784592e-01 -1.08143771e+00
3.87843281e-01 -1.72224671e-01 -7.21645534e-01 2.78047696e-02
-3.33602965e-01 -6.13121353e-02 7.09769607e-01 -5.39085209e-01
-6.06659412e-01 4.07301895e-02 1.03962161e-01 -2.82933652e-01
-4.73390877e-01 3.39666158e-02 9.40534174e-02 -1.94922403e-01
8.16253483e-01 -1.67031333e-01 3.35269771e-03 -2.52169687e-02
-4.37064916e-01 3.31879377e-01 5.06751835e-01 -4.68365312e-01
3.58684003e-01 8.66344154e-01 1.33956239e-01 -4.71433640e-01
-7.58959174e-01 -3.94371718e-01 -4.52749163e-01 -2.74250895e-01
8.95594776e-01 -6.26167476e-01 -1.03190911e+00 4.71821129e-01
-1.23587906e+00 -5.94024956e-01 -2.03619629e-01 3.29412580e-01
-3.05135101e-01 3.57472658e-01 -7.18926132e-01 -9.68575120e-01
-2.96246380e-01 -1.29324901e+00 1.18431997e+00 4.60346222e-01
-5.87019384e-01 -9.38891351e-01 -2.52839804e-01 3.43199462e-01
4.56447214e-01 6.64539278e-01 8.56272876e-01 -8.50034297e-01
-4.30711597e-01 -2.17214942e-01 -4.71738607e-01 1.23757450e-02
-5.92681244e-02 3.73046160e-01 -1.38843668e+00 -2.03961551e-01
-2.21067384e-01 -3.24840784e-01 8.05428565e-01 7.31052160e-01
1.44136226e+00 3.57226320e-02 -5.41476786e-01 4.22804892e-01
1.44472790e+00 9.64522362e-02 7.14573085e-01 3.90985698e-01
3.12487543e-01 8.14519107e-01 3.29614878e-01 3.72795075e-01
-4.06828849e-03 3.28415215e-01 6.40194893e-01 -5.36812007e-01
-1.36317611e-02 9.00751427e-02 -2.18686655e-01 1.33651987e-01
-2.31028810e-01 -5.44008732e-01 -8.21871579e-01 7.20589399e-01
-1.39205348e+00 -5.81762731e-01 -4.35984164e-01 2.32469273e+00
6.76761746e-01 4.09758419e-01 2.15584785e-01 2.27625728e-01
8.44735801e-01 -3.22503626e-01 -5.76258123e-01 -2.73199141e-01
-3.72344315e-01 1.79401398e-01 7.05424070e-01 2.43081242e-01
-9.83140171e-01 7.20816314e-01 6.10518026e+00 8.17685068e-01
-1.25089288e+00 -1.48152774e-02 1.32949722e+00 -1.89534679e-01
8.10474753e-02 -2.05332085e-01 -7.46637583e-01 8.22035968e-01
5.52245319e-01 3.61933351e-01 1.21591546e-01 2.90776014e-01
4.44723040e-01 -5.23109138e-01 -1.12724829e+00 7.79256821e-01
-2.38583967e-01 -1.28167367e+00 -2.45920420e-02 3.19273949e-01
4.48958993e-01 -1.19288668e-01 5.80281839e-02 1.19861625e-01
2.07754970e-01 -1.30294132e+00 8.68673146e-01 4.74263549e-01
8.88163090e-01 -3.92837375e-01 1.21322799e+00 1.05531946e-01
-7.50983059e-01 7.94020295e-02 -3.88246387e-01 -3.67479436e-02
6.24371320e-02 7.93115735e-01 -9.08051550e-01 -9.13622603e-02
8.24678719e-01 1.44721732e-01 -8.29968274e-01 1.30744159e+00
1.80923447e-01 4.59451020e-01 -4.06719267e-01 -2.03752160e-01
1.67260066e-01 -9.45512205e-03 2.42814317e-01 1.57610512e+00
4.01381135e-01 -1.07147440e-01 -4.05778974e-01 1.30960405e+00
-1.97189644e-01 9.41817909e-02 -2.57989287e-01 -7.70759434e-02
4.35168624e-01 1.66308510e+00 -1.57768595e+00 -1.93538904e-01
-1.80708885e-01 6.56809807e-01 4.26794261e-01 3.85496140e-01
-8.92877698e-01 -2.71967649e-01 2.55966216e-01 5.35530925e-01
3.44893277e-01 1.80638820e-01 -7.57275760e-01 -6.62446797e-01
-4.16438840e-02 -5.18289208e-01 2.70091355e-01 -8.00654650e-01
-1.20317280e+00 4.35321808e-01 -4.25988197e-01 -1.00509083e+00
4.46484357e-01 -9.83461082e-01 -5.43876767e-01 7.11486161e-01
-1.39229679e+00 -6.59392595e-01 -4.83364314e-01 -1.31914869e-01
3.70069444e-01 3.40930223e-01 6.73935652e-01 3.67951393e-01
-6.87434971e-01 6.38466656e-01 -1.52045131e-01 1.32015422e-01
6.63151503e-01 -1.36820459e+00 6.12708144e-02 5.93156338e-01
-1.89137116e-01 8.31868947e-01 9.56597805e-01 -2.71557719e-01
-4.71608460e-01 -8.64765584e-01 8.17714870e-01 -6.47971630e-01
5.26584685e-01 -4.83059734e-01 -9.58087146e-01 3.14886332e-01
-1.83559898e-02 2.96626776e-01 7.83515453e-01 -1.01753138e-01
-1.39249578e-01 8.57662186e-02 -1.37951422e+00 7.48030961e-01
7.90686548e-01 -8.50553364e-02 4.21766303e-02 2.17163041e-01
1.27892479e-01 -2.82148540e-01 -7.81347096e-01 3.18307251e-01
6.17301285e-01 -1.23959887e+00 6.45748317e-01 -1.89247221e-01
4.54091430e-01 -5.04312038e-01 2.96887487e-01 -8.27139497e-01
-2.64237702e-01 -2.11576894e-01 3.62567991e-01 1.13441181e+00
6.05497777e-01 -5.21763980e-01 9.18535411e-01 5.63545644e-01
4.39052097e-02 -1.02272582e+00 -9.55301762e-01 -6.65977597e-01
1.99833289e-01 1.38826743e-01 4.82293069e-01 9.01386082e-01
-2.27351785e-01 3.57366264e-01 1.29392684e-01 -9.00787711e-02
3.65707546e-01 -2.25829974e-01 6.64430976e-01 -1.29400885e+00
-7.40290154e-03 -7.90395558e-01 -4.45112735e-01 -6.63885295e-01
-2.12332368e-01 -5.68736255e-01 9.98884141e-02 -1.28288996e+00
3.88265014e-01 -3.36213350e-01 -3.54214400e-01 2.23992854e-01
-3.59972894e-01 5.46754956e-01 -4.04146910e-02 1.03562690e-01
-7.53507793e-01 -5.97874112e-02 1.35472178e+00 -8.30805674e-02
9.71798375e-02 -2.95707136e-01 -7.33915329e-01 7.34129250e-01
7.47605503e-01 -5.08294344e-01 1.01459287e-01 -4.54219609e-01
3.29202265e-01 -4.07429844e-01 7.77728260e-01 -1.19786453e+00
1.09699890e-01 2.41255537e-02 7.51748741e-01 -2.88524598e-01
1.26109749e-01 -7.68211782e-01 1.06211409e-01 5.71029961e-01
-3.00139040e-01 -1.55419469e-01 4.10947293e-01 4.09332484e-01
1.30495876e-01 -3.98441702e-01 1.07014334e+00 -3.33340555e-01
-4.21735272e-03 -3.52236964e-02 -7.16000855e-01 -3.28954346e-02
9.47516978e-01 -8.91237438e-01 -5.14269888e-01 -7.26296157e-02
-4.61069167e-01 1.53852269e-01 9.03297603e-01 -3.46015275e-01
8.75774771e-02 -6.69916332e-01 -5.23880422e-01 -4.49744612e-02
2.53421277e-01 3.25888634e-01 2.33849972e-01 1.29207790e+00
-9.83397961e-01 3.46609205e-01 -1.92412600e-01 -7.38527179e-01
-9.79797661e-01 4.44887549e-01 4.14797097e-01 -5.34166396e-01
-1.92825019e-01 1.06070745e+00 3.66815358e-01 -1.38209879e-01
2.00176135e-01 -7.38395095e-01 -3.24201882e-01 2.99800863e-03
4.09459025e-01 5.20736396e-01 1.44550323e-01 -3.40305030e-01
-4.42127615e-01 3.85431111e-01 -2.29246244e-01 2.35693857e-01
1.06015122e+00 -2.45253295e-02 8.13960433e-02 4.73785847e-01
8.30164671e-01 -4.45306348e-03 -1.57942998e+00 3.18126529e-01
9.04054642e-02 -2.57108480e-01 -2.65273511e-01 -1.04538476e+00
-9.41215038e-01 9.77445483e-01 6.51970625e-01 1.89042062e-01
8.15467298e-01 -2.08538517e-01 4.63069707e-01 3.73842604e-02
2.62306362e-01 -1.05979311e+00 -1.22232974e-01 2.04934716e-01
5.43468356e-01 -1.21706772e+00 1.01462029e-01 -3.36165220e-01
-3.37503463e-01 9.44727063e-01 7.70075262e-01 -7.88312778e-02
5.65223508e-02 5.89333713e-01 2.28411376e-01 -2.68476069e-01
-6.15768969e-01 -9.74979177e-02 -2.67800510e-01 6.00842237e-01
6.67364299e-01 -2.10879952e-01 -5.56365609e-01 6.12553835e-01
-2.98216790e-02 1.75075829e-01 5.21817684e-01 6.67861581e-01
-3.02623421e-01 -5.03456295e-01 -3.97524267e-01 7.11562693e-01
-8.38388264e-01 2.93188654e-02 -4.94180381e-01 9.32838440e-01
3.92385811e-01 6.75529003e-01 3.59915167e-01 6.66533634e-02
2.27420911e-01 -2.89859474e-01 4.13886279e-01 -4.19493735e-01
-8.50582242e-01 1.84225932e-01 -3.26978296e-01 -2.86933810e-01
-5.78995347e-01 -5.50932050e-01 -1.34775293e+00 -1.57854035e-01
-5.23430347e-01 -1.87166296e-02 2.64631033e-01 8.27322006e-01
3.45411837e-01 5.85316718e-01 -1.45737857e-01 -7.90771782e-01
-6.67589381e-02 -1.05689740e+00 -5.83988965e-01 5.56547463e-01
1.05196171e-01 -7.65733540e-01 -6.07364655e-01 3.62558812e-01] | [14.685964584350586, -3.1470234394073486] |
f515fdf9-c671-46f1-b5a9-5196c859c046 | attribution-scores-and-causal-counterfactuals | 2303.02829 | null | https://arxiv.org/abs/2303.02829v2 | https://arxiv.org/pdf/2303.02829v2.pdf | Attribution-Scores and Causal Counterfactuals as Explanations in Artificial Intelligence | In this expository article we highlight the relevance of explanations for artificial intelligence, in general, and for the newer developments in {\em explainable AI}, referring to origins and connections of and among different approaches. We describe in simple terms, explanations in data management and machine learning that are based on attribution-scores, and counterfactuals as found in the area of causality. We elaborate on the importance of logical reasoning when dealing with counterfactuals, and their use for score computation. | ['Leopoldo Bertossi'] | 2023-03-06 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [ 4.08600628e-01 9.34198737e-01 -6.84230328e-01 -6.27917767e-01
9.62898806e-02 -2.76998967e-01 8.92481089e-01 3.49936843e-01
-9.41232890e-02 1.22144961e+00 7.28123426e-01 -8.08744073e-01
-9.77247238e-01 -6.52705431e-01 -4.85303581e-01 -3.64699781e-01
-3.14561307e-01 5.97153604e-01 -5.70166349e-01 -3.84510234e-02
8.28960240e-01 4.11255598e-01 -1.72265029e+00 2.41992339e-01
7.38005817e-01 7.29922891e-01 -4.62753654e-01 4.81135815e-01
-3.16967517e-01 1.26116610e+00 -9.59356368e-01 -1.03261137e+00
-3.53694320e-01 -5.68575799e-01 -9.66777623e-01 -9.13922861e-02
-1.52060390e-01 -1.15472171e-02 -5.09587377e-02 5.79378188e-01
-2.01366425e-01 -1.09060239e-02 1.25336230e+00 -1.99908721e+00
-1.03855574e+00 1.19763100e+00 -3.17202620e-02 4.54141907e-02
5.90336859e-01 -7.75118098e-02 1.12030256e+00 -5.28792620e-01
2.09798425e-01 1.54248989e+00 3.97942275e-01 8.06391358e-01
-1.41322207e+00 -4.12831813e-01 3.72802347e-01 7.19819486e-01
-4.87551272e-01 -1.72713503e-01 5.39713621e-01 -5.45517445e-01
8.59892726e-01 8.60045254e-01 4.22309339e-01 1.06339610e+00
3.45658183e-01 6.18624389e-01 1.21444964e+00 -7.54678726e-01
5.54352701e-01 1.75651744e-01 5.02326012e-01 3.94266129e-01
9.02313054e-01 6.84836268e-01 -9.40228403e-01 -2.48968557e-01
5.60494363e-01 -6.05858043e-02 -1.72083870e-01 -2.55366236e-01
-1.41728210e+00 1.31130767e+00 4.60121840e-01 1.57343432e-01
-5.55843174e-01 5.10720134e-01 8.13891962e-02 1.94520801e-01
4.05742645e-01 9.77997303e-01 -7.37680435e-01 1.65146217e-01
-5.38901925e-01 3.94289255e-01 9.36709285e-01 5.52806556e-01
2.83602804e-01 2.45879471e-01 1.72531456e-02 1.19806185e-01
4.62391138e-01 2.07122520e-01 4.01255041e-01 -1.35387540e+00
1.16715424e-01 5.72493196e-01 2.35758498e-01 -7.16898859e-01
-4.45202261e-01 -2.17199475e-01 -6.33718193e-01 4.41793889e-01
2.71946222e-01 -1.46800242e-02 -4.85291451e-01 1.77714777e+00
3.11893076e-02 1.31609693e-01 4.44338083e-01 5.90243161e-01
7.80524194e-01 -6.45767301e-02 3.24671179e-01 -7.75977910e-01
9.29728687e-01 -5.59061885e-01 -1.09872699e+00 -2.82681167e-01
6.47024333e-01 -2.90178657e-01 8.42079341e-01 5.93096077e-01
-1.04017222e+00 -1.95531800e-01 -1.06706834e+00 -3.57202739e-02
-5.00631392e-01 -4.81067121e-01 1.51167059e+00 6.42483294e-01
-7.26098657e-01 8.68856370e-01 -3.60345691e-01 -1.67680964e-01
3.44797313e-01 3.94844353e-01 -3.68314832e-02 4.12784278e-01
-1.20515418e+00 1.28386915e+00 4.01485145e-01 -2.25938842e-01
-4.57521737e-01 -6.29607141e-01 -9.10743117e-01 1.78069532e-01
4.94129270e-01 -8.55770350e-01 1.16855752e+00 -8.96062791e-01
-1.18024850e+00 6.69214189e-01 -9.19743329e-02 -8.22819948e-01
3.47760081e-01 -1.85549274e-01 -5.05435646e-01 -3.97434533e-01
8.85760337e-02 4.49055851e-01 2.32857361e-01 -1.10090685e+00
-3.98819804e-01 -4.65196013e-01 2.97735453e-01 -4.45511714e-02
1.71859816e-01 2.19571404e-02 8.11806619e-01 -5.95841348e-01
7.98924565e-02 -7.91377425e-01 -3.62714112e-01 -1.46268934e-01
-6.24926209e-01 -6.07127368e-01 1.72857583e-01 -1.38006359e-01
1.25510871e+00 -1.58405066e+00 1.83098227e-01 2.55189002e-01
2.31022492e-01 -3.88319790e-01 3.09960216e-01 2.39767820e-01
-7.31091321e-01 4.80702311e-01 -2.49306381e-01 6.48741275e-02
4.60360944e-01 4.42714661e-01 -7.40540445e-01 2.45435491e-01
2.45755896e-01 9.09887254e-01 -7.08264470e-01 -4.16804820e-01
4.58097458e-01 -8.62722993e-02 -2.83885360e-01 7.05581978e-02
-3.14636379e-01 1.33555401e-02 -4.05546546e-01 2.52059907e-01
1.38626732e-02 -7.95322582e-02 4.27921206e-01 3.17819208e-01
-4.01740372e-01 8.12284708e-01 -8.89848173e-01 1.14844191e+00
-8.79146978e-02 7.60837257e-01 -7.75758684e-01 -9.51635122e-01
8.19538057e-01 5.72176516e-01 -6.18174188e-02 -1.36827141e-01
1.95659757e-01 3.13229799e-01 2.07447007e-01 -4.27813709e-01
2.19173983e-01 -6.18751228e-01 -1.99247405e-01 7.63707519e-01
-3.24887276e-01 -2.09667146e-01 7.48521537e-02 1.17169015e-01
7.56117463e-01 7.51081407e-02 1.21703184e+00 -2.98213422e-01
4.99817997e-01 3.68863225e-01 2.03742981e-01 6.96290135e-01
2.35456407e-01 2.43936792e-01 7.98667431e-01 -7.07281590e-01
-6.69142365e-01 -9.47633207e-01 -3.22445363e-01 8.26992035e-01
-1.59988582e-01 5.28181065e-03 -5.35124600e-01 -6.96686268e-01
4.00282741e-01 1.80373716e+00 -1.04397440e+00 -2.03377858e-01
-1.85068175e-01 -9.50061321e-01 2.54490882e-01 6.42210543e-01
-1.99553490e-01 -1.16342163e+00 -7.69485533e-01 -1.07862214e-02
-9.52870399e-02 -3.74716729e-01 4.17717814e-01 4.14503694e-01
-1.09217179e+00 -1.28248739e+00 2.71864962e-02 2.90904462e-01
3.74279290e-01 -9.87447724e-02 1.29971910e+00 4.29277033e-01
1.30086839e-01 1.96335375e-01 -8.82166065e-03 -1.15791452e+00
-5.14489233e-01 -5.86524725e-01 2.59576708e-01 -5.21722078e-01
7.00465977e-01 -5.45777678e-01 -1.68150485e-01 1.15233153e-01
-7.56291807e-01 1.14364021e-01 4.44887400e-01 7.81563282e-01
1.29316255e-01 -2.95864403e-01 5.50367594e-01 -1.17126071e+00
6.66315913e-01 -7.41409004e-01 -2.81857431e-01 4.41904306e-01
-1.21483457e+00 5.56332469e-01 1.63705274e-01 -1.13329336e-01
-1.19896090e+00 -7.56615460e-01 4.50946122e-01 1.14985049e-01
-3.36731493e-01 4.92681652e-01 -2.36960292e-01 2.89969563e-01
8.32345426e-01 -5.80343068e-01 -5.62139414e-02 -3.30896080e-01
8.10558796e-01 4.75697935e-01 4.41617280e-01 -4.38477278e-01
5.31030118e-01 5.33078730e-01 6.20829880e-01 -1.45832887e-02
-1.10128176e+00 2.76729345e-01 -3.92380446e-01 -2.20798299e-01
6.49718404e-01 -1.78289622e-01 -1.16998518e+00 -4.19925243e-01
-1.28341150e+00 1.62243083e-01 -7.30782986e-01 9.35320497e-01
-1.05106795e+00 -2.35367775e-01 -4.58150096e-02 -1.13330197e+00
9.21978503e-02 -8.19798112e-01 2.66500920e-01 1.18138731e-01
-9.75852370e-01 -1.32122076e+00 7.96284899e-02 4.34591144e-01
-1.71687379e-02 6.01799726e-01 1.29998446e+00 -1.00705481e+00
-3.17911446e-01 -1.08570039e-01 1.33214854e-02 -2.43971676e-01
-1.09417267e-01 9.19352323e-02 -1.00922394e+00 6.54264808e-01
1.11676045e-01 -1.47020549e-01 7.71606445e-01 4.58150744e-01
1.09796286e+00 -7.96221495e-01 -4.08655852e-01 5.99184521e-02
1.28216922e+00 2.30916560e-01 5.08044124e-01 4.83637154e-01
2.23844513e-01 1.30698848e+00 8.60965669e-01 5.48359394e-01
4.20338035e-01 4.32890475e-01 7.47767210e-01 6.25047088e-02
6.45311177e-02 -1.29556879e-01 2.73745563e-02 1.79165348e-01
-6.01881504e-01 -2.10237503e-01 -5.39760768e-01 3.75702053e-01
-2.20299315e+00 -1.32278335e+00 -8.63656163e-01 2.21647573e+00
5.60711265e-01 1.78049430e-01 -5.52176014e-02 4.31573331e-01
5.74857950e-01 -1.03785366e-01 -5.71962416e-01 -1.00102067e+00
-2.76858240e-01 1.90513805e-01 4.32477504e-01 6.98111176e-01
-6.03152633e-01 3.44093323e-01 8.25375748e+00 3.50497693e-01
-4.70414817e-01 4.55637313e-02 6.64724112e-01 -2.82504439e-01
-9.56265271e-01 3.81127477e-01 -3.76083888e-02 2.25367323e-01
1.09444141e+00 -5.50895154e-01 3.32145393e-01 8.06638777e-01
3.31079036e-01 -2.52299190e-01 -1.68846679e+00 3.75446916e-01
-9.45107490e-02 -1.41737568e+00 4.32743043e-01 7.69971088e-02
7.51938641e-01 -5.09561241e-01 6.54749870e-02 -1.87896550e-01
5.90426087e-01 -1.31211078e+00 1.12009227e+00 6.90055072e-01
2.45222151e-01 -6.98521733e-01 9.44272757e-01 1.12170830e-01
-2.28934988e-01 -2.42356822e-01 -3.91242146e-01 -9.17798936e-01
-9.61584002e-02 6.67449594e-01 -8.37935090e-01 6.86037004e-01
2.03313097e-01 3.97502363e-01 -1.01779833e-01 5.35525680e-01
-9.97161388e-01 4.22770798e-01 5.81821352e-02 -3.74897093e-01
2.65093632e-02 -1.20580783e-02 3.97717059e-01 9.28446054e-01
1.55578867e-01 2.71684945e-01 -6.63782299e-01 1.28260565e+00
2.65812248e-01 -4.10647422e-01 -6.41462862e-01 1.45561844e-01
5.84853709e-01 6.76684260e-01 -3.24903786e-01 -4.73763883e-01
-2.99166650e-01 4.09944534e-01 1.41301781e-01 1.75278708e-01
-9.69408095e-01 5.83189502e-02 8.43822956e-01 -1.10493049e-01
-3.93461347e-01 2.25881994e-01 -1.12531459e+00 -9.91368234e-01
-2.97959358e-01 -6.74377441e-01 7.38046885e-01 -1.02375615e+00
-1.28383529e+00 2.46106893e-01 5.72166800e-01 -7.41428733e-01
-8.25851381e-01 -8.44240248e-01 -8.58212113e-01 6.56642735e-01
-1.32272553e+00 -6.42729938e-01 1.18148319e-01 2.19086945e-01
3.36111993e-01 1.10416643e-01 9.50413644e-01 -6.71210468e-01
-4.02283370e-01 1.65807307e-01 -9.74486321e-02 -6.22850597e-01
4.07877207e-01 -1.67367244e+00 4.05888438e-01 5.13207674e-01
3.61471683e-01 6.95857704e-01 1.30119443e+00 -4.25958037e-01
-7.35374510e-01 -5.62230229e-01 1.39892673e+00 -7.22354591e-01
5.66762328e-01 2.38833964e-01 -5.41600585e-01 9.59193110e-01
3.16678733e-01 -8.31145525e-01 1.02432430e+00 5.50521195e-01
-2.93895185e-01 2.87507355e-01 -1.22816479e+00 8.89343739e-01
1.14994276e+00 -6.58776686e-02 -1.24338567e+00 3.04706275e-01
6.75166190e-01 2.09306806e-01 -6.06754661e-01 2.45858535e-01
7.30230570e-01 -1.24981141e+00 9.89028156e-01 -1.37458634e+00
6.11395359e-01 -2.25993887e-01 -2.08060499e-02 -1.27717495e+00
-4.32783484e-01 -5.14987588e-01 -1.52997021e-02 8.59024942e-01
6.12463534e-01 -7.85038173e-01 5.58339775e-01 1.27525210e+00
1.11702316e-01 -5.33690333e-01 -9.72494483e-01 -4.81211662e-01
2.06740759e-02 -7.19340086e-01 1.07478392e+00 1.23569536e+00
7.37238169e-01 3.00081193e-01 -1.85571864e-01 9.08589829e-03
1.09861982e+00 2.92246640e-01 4.89582062e-01 -1.66685724e+00
-1.62671477e-01 -5.45611560e-01 -4.15514141e-01 -2.67705828e-01
2.95456916e-01 -6.16823018e-01 -3.40426326e-01 -1.48590577e+00
2.54296064e-01 -2.56163459e-02 -2.78706819e-01 5.58533847e-01
-3.65349054e-01 4.46307473e-02 1.98647276e-01 2.88938969e-01
-2.24566072e-01 2.56734848e-01 6.78775072e-01 2.30523989e-01
2.50755310e-01 -1.02335587e-01 -1.13243139e+00 1.25531209e+00
9.34303641e-01 -8.07884276e-01 -3.39437336e-01 -2.39766926e-01
6.11261606e-01 9.39365178e-02 7.74672091e-01 -4.85356510e-01
-1.25888407e-01 -9.46998119e-01 3.44717860e-01 -2.41154373e-01
1.45740226e-01 -6.36478066e-01 5.84553838e-01 7.46805906e-01
-9.76171136e-01 1.50287479e-01 -2.10213199e-01 4.45068240e-01
2.60811928e-03 -5.65908134e-01 1.89914092e-01 -2.46169999e-01
-3.71775478e-01 -3.88241798e-01 -2.72593379e-01 -2.83041596e-01
9.65938091e-01 -3.77625823e-01 -5.89260161e-01 -5.31772554e-01
-6.70793772e-01 1.08783111e-01 1.46033898e-01 3.60739887e-01
5.09186268e-01 -1.23160672e+00 -6.57851994e-01 -4.56181109e-01
2.26819944e-02 -6.69812799e-01 -2.34114714e-02 7.87476897e-01
-9.37680230e-02 8.77083421e-01 -3.10085028e-01 3.00247729e-01
-9.78847206e-01 8.55160058e-01 1.64585367e-01 -1.47026420e-01
1.07285865e-01 6.28898442e-01 6.00228727e-01 -1.14566453e-01
9.00688395e-02 -3.21670532e-01 -4.90527153e-01 -2.23179191e-01
4.41132575e-01 8.60934317e-01 -3.73667479e-01 6.99048042e-02
-6.04029715e-01 3.41684893e-02 3.13664585e-01 -3.75020593e-01
1.33180094e+00 -2.09605284e-02 -4.80710268e-01 9.97290790e-01
3.65570843e-01 -3.10800076e-01 -7.70871103e-01 1.56943142e-01
4.81453389e-01 -3.18639606e-01 -1.47457659e-01 -1.24318373e+00
-3.67252231e-01 7.35286117e-01 9.31687728e-02 6.37034953e-01
8.12825024e-01 3.19803357e-01 -3.01615864e-01 3.78180861e-01
2.67126054e-01 -7.04558730e-01 -3.98783863e-01 -1.91494644e-01
1.27154207e+00 -9.33609545e-01 4.18130040e-01 -4.04417604e-01
-5.58361590e-01 1.24498975e+00 2.93100208e-01 1.29277378e-01
2.02960536e-01 7.05804080e-02 -9.42999423e-02 -3.99311453e-01
-1.15134919e+00 -1.39302388e-01 2.81846493e-01 7.17962623e-01
7.69633472e-01 6.43433154e-01 -9.59116220e-01 7.87989199e-01
-6.64977849e-01 1.86328311e-02 8.92519057e-01 5.35660088e-01
-3.49846959e-01 -1.01686656e+00 -7.73634672e-01 5.67462087e-01
-4.21809703e-01 -2.35361323e-01 -1.03747952e+00 1.10661983e+00
3.13886285e-01 1.36618686e+00 1.87198356e-01 -2.67989874e-01
2.78703421e-01 1.17516249e-01 4.48500097e-01 -3.36846173e-01
-4.95742708e-01 -6.27542555e-01 4.35107619e-01 -5.45306385e-01
-9.07177448e-01 -9.84213710e-01 -1.19682693e+00 -7.18554676e-01
-4.69386667e-01 5.98501146e-01 7.75475919e-01 1.15895534e+00
-8.18977654e-02 5.66416502e-01 3.92679811e-01 -3.98829222e-01
-6.31577015e-01 -9.20714974e-01 -5.32453477e-01 8.46905410e-02
2.46633649e-01 -8.06634068e-01 -8.01297605e-01 -1.29128620e-01] | [8.664872169494629, 5.682901382446289] |
158e8cf6-0750-4d6d-a80d-515dbc2fc454 | dissecting-arbitrary-scale-super-resolution | 2306.00714 | null | https://arxiv.org/abs/2306.00714v1 | https://arxiv.org/pdf/2306.00714v1.pdf | Dissecting Arbitrary-scale Super-resolution Capability from Pre-trained Diffusion Generative Models | Diffusion-based Generative Models (DGMs) have achieved unparalleled performance in synthesizing high-quality visual content, opening up the opportunity to improve image super-resolution (SR) tasks. Recent solutions for these tasks often train architecture-specific DGMs from scratch, or require iterative fine-tuning and distillation on pre-trained DGMs, both of which take considerable time and hardware investments. More seriously, since the DGMs are established with a discrete pre-defined upsampling scale, they cannot well match the emerging requirements of arbitrary-scale super-resolution (ASSR), where a unified model adapts to arbitrary upsampling scales, instead of preparing a series of distinct models for each case. These limitations beg an intriguing question: can we identify the ASSR capability of existing pre-trained DGMs without the need for distillation or fine-tuning? In this paper, we take a step towards resolving this matter by proposing Diff-SR, a first ASSR attempt based solely on pre-trained DGMs, without additional training efforts. It is motivated by an exciting finding that a simple methodology, which first injects a specific amount of noise into the low-resolution images before invoking a DGM's backward diffusion process, outperforms current leading solutions. The key insight is determining a suitable amount of noise to inject, i.e., small amounts lead to poor low-level fidelity, while over-large amounts degrade the high-level signature. Through a finely-grained theoretical analysis, we propose the Perceptual Recoverable Field (PRF), a metric that achieves the optimal trade-off between these two factors. Extensive experiments verify the effectiveness, flexibility, and adaptability of Diff-SR, demonstrating superior performance to state-of-the-art solutions under diverse ASSR environments. | ['Zhenhua Han', 'Yifei Shen', 'Xinyang Jiang', 'Jingcai Guo', 'Jie Zhang', 'Song Guo', 'Qihua Zhou', 'Ruibin Li'] | 2023-06-01 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [ 6.63917065e-01 -1.07152656e-01 2.67022736e-02 -1.61001086e-01
-8.83479714e-01 -4.01643455e-01 7.10330307e-01 -2.57884651e-01
-2.53876895e-01 6.75860465e-01 1.81565911e-01 -8.96033123e-02
-2.52165973e-01 -8.21940899e-01 -6.06699228e-01 -8.11899245e-01
4.90807146e-02 1.57097250e-01 4.83914673e-01 -3.91254187e-01
1.78662732e-01 6.57539845e-01 -1.55418956e+00 2.49395519e-01
1.08532643e+00 8.36078703e-01 6.22297049e-01 6.40588343e-01
-6.15842976e-02 5.19088447e-01 -4.41981733e-01 -3.81453067e-01
3.91357988e-01 -7.36544073e-01 -5.37794054e-01 1.24915846e-01
4.85229492e-01 -3.40960741e-01 -2.96656191e-01 1.20553422e+00
4.67486411e-01 1.64051615e-02 4.73350346e-01 -6.76126838e-01
-9.58211482e-01 5.46670914e-01 -7.31865942e-01 5.51712394e-01
5.98065816e-02 2.81878173e-01 8.36533070e-01 -8.66550803e-01
6.15126252e-01 1.17156065e+00 6.17104709e-01 6.80399954e-01
-1.58703744e+00 -5.11669576e-01 2.97028217e-02 -9.88983735e-03
-1.48812270e+00 -5.12107670e-01 7.47498333e-01 -3.11411470e-01
5.85345745e-01 1.35793105e-01 3.69615316e-01 1.15477157e+00
1.10827893e-01 2.01199263e-01 1.48490047e+00 -3.60103607e-01
3.91946614e-01 2.04941705e-01 -3.14308703e-01 3.76553833e-01
4.25735027e-01 1.48409799e-01 -5.48397541e-01 4.30436768e-02
1.46208954e+00 -2.06808105e-01 -4.24168408e-01 -2.80163437e-01
-1.11661136e+00 6.36404812e-01 4.36505646e-01 5.47758877e-01
-3.22233051e-01 -1.21437863e-01 -7.96815902e-02 2.67322451e-01
3.65193248e-01 6.47623122e-01 -1.53596416e-01 1.41092330e-01
-1.38678551e+00 2.91338153e-02 2.50264406e-01 7.65928328e-01
8.18850636e-01 3.57320368e-01 -1.68378547e-01 8.61714244e-01
-1.55917019e-01 2.28747010e-01 4.29079264e-01 -9.43055511e-01
1.79205462e-01 2.23307520e-01 1.76151440e-01 -8.11671138e-01
-1.46753252e-01 -9.80666459e-01 -1.09406424e+00 3.86594772e-01
4.90260124e-01 1.01360485e-01 -1.00929499e+00 1.81457448e+00
2.78023154e-01 1.80207506e-01 2.11890396e-02 1.10657501e+00
3.90346438e-01 5.37428558e-01 -1.03194341e-02 -4.17315662e-01
1.33524716e+00 -6.94178462e-01 -5.18119752e-01 -1.90734535e-01
-1.16148308e-01 -7.43326545e-01 1.26582682e+00 4.29225206e-01
-1.37265909e+00 -8.81186426e-01 -1.34409320e+00 -8.66973922e-02
-8.40993971e-02 -5.16998880e-02 4.44232017e-01 6.58008933e-01
-1.37414372e+00 8.58080566e-01 -5.66365242e-01 -2.58588910e-01
3.44294578e-01 2.49535277e-01 -1.98627576e-01 -1.58901840e-01
-1.09415388e+00 7.46874273e-01 6.50479943e-02 4.47001345e-02
-8.95355701e-01 -7.54931629e-01 -3.74477506e-01 1.93528719e-02
4.04634982e-01 -8.31545234e-01 7.55972803e-01 -9.69410121e-01
-1.66239166e+00 7.58100867e-01 -1.38628185e-01 -5.41215956e-01
5.77988803e-01 6.35604560e-02 -6.57450616e-01 3.21211904e-01
-6.68518469e-02 5.18478453e-01 1.38648689e+00 -1.51650167e+00
-4.67604101e-01 -3.25816602e-01 1.35610238e-01 2.29324233e-02
-2.98559487e-01 7.13528767e-02 -2.64715105e-01 -7.97338247e-01
2.55862266e-01 -6.67480171e-01 -4.28097010e-01 -2.10976243e-01
-8.34216475e-02 3.14754456e-01 4.96282667e-01 -4.76125091e-01
1.22412932e+00 -2.12660742e+00 2.04645067e-01 -2.94659752e-02
4.68982875e-01 4.65801001e-01 -3.70146841e-01 1.90400600e-01
4.95136194e-02 1.48649588e-01 -3.87905836e-01 -2.06645340e-01
-2.44078010e-01 4.05141897e-02 -4.66010422e-01 2.63466150e-01
5.00088573e-01 7.42198348e-01 -8.54862869e-01 -3.82235736e-01
1.05085470e-01 7.80721903e-01 -5.57671368e-01 1.62910655e-01
-1.05939150e-01 7.01607108e-01 -3.24573547e-01 3.11476141e-01
8.63152206e-01 -5.08300602e-01 2.32960477e-01 -4.21428114e-01
-2.80571938e-01 1.83187485e-01 -1.38796091e+00 1.53788722e+00
-3.84498358e-01 3.93031567e-01 1.28742516e-01 -8.75766814e-01
1.07765150e+00 7.91888684e-02 3.79128128e-01 -8.16688538e-01
-1.32233575e-01 4.96527284e-01 -3.68081592e-02 -1.17487095e-01
6.88937306e-01 -4.87643957e-01 1.31308913e-01 1.86606407e-01
2.30488241e-01 -6.34069294e-02 3.24913822e-02 1.93871617e-01
1.09609425e+00 1.84603203e-02 3.09244066e-01 -2.91071981e-01
5.53431988e-01 -3.36097896e-01 5.06264210e-01 9.37187552e-01
-1.31618276e-01 1.07890701e+00 2.36602217e-01 -1.36679128e-01
-1.38650703e+00 -1.31909871e+00 -2.55255699e-01 7.77429342e-01
2.81086296e-01 -1.92923799e-01 -9.01107609e-01 -4.11374420e-01
-4.15065795e-01 3.26529711e-01 -5.69469392e-01 4.42063808e-02
-7.14589477e-01 -9.97785985e-01 4.09128398e-01 3.40909511e-01
6.78782880e-01 -6.75879538e-01 -5.32211125e-01 2.58655250e-01
-7.03699812e-02 -1.40878499e+00 -2.49475062e-01 2.87083052e-02
-9.55529034e-01 -5.35938799e-01 -1.02238441e+00 -5.08706629e-01
6.07651889e-01 5.69122076e-01 1.13052297e+00 -3.74155566e-02
-1.45195484e-01 5.12792952e-02 -2.25494057e-01 1.46802515e-01
-5.69696009e-01 1.07229792e-01 6.07158504e-02 1.58249393e-01
-9.29080416e-03 -1.19382262e+00 -8.61036718e-01 3.06658775e-01
-1.10733068e+00 1.93182722e-01 9.46915567e-01 8.65697086e-01
8.04647088e-01 2.88968593e-01 8.79401803e-01 -6.36337757e-01
4.83330846e-01 -2.64105231e-01 -5.66515148e-01 1.47704989e-01
-7.75852025e-01 2.27154210e-01 8.07381034e-01 -6.75413489e-01
-1.08550990e+00 -2.33469546e-01 -1.58611253e-01 -4.77517664e-01
-4.82641859e-03 1.05156898e-01 -2.97064573e-01 -2.28580192e-01
8.06854010e-01 4.72608656e-01 -2.35653549e-01 -5.98830044e-01
5.78334510e-01 4.26069915e-01 8.36742580e-01 -5.33145487e-01
1.12338412e+00 6.61136270e-01 -8.26563500e-03 -7.71366894e-01
-8.31987977e-01 -1.49497986e-01 -5.29569983e-01 9.12063289e-03
8.45672429e-01 -9.78431523e-01 -1.64885834e-01 2.85261899e-01
-7.80393660e-01 -3.33340824e-01 -5.57142854e-01 2.01953977e-01
-5.59364378e-01 2.94415146e-01 -6.05159461e-01 -6.69130206e-01
-1.93397135e-01 -1.09834337e+00 9.08859491e-01 3.48554283e-01
1.55354273e-02 -6.42705083e-01 -1.63390175e-01 3.06312114e-01
9.64537799e-01 5.81084192e-02 8.74381125e-01 -1.92312777e-01
-8.75591934e-01 2.96881497e-01 -5.60510993e-01 4.19016808e-01
1.33009285e-01 -2.64224976e-01 -9.20190871e-01 -4.44851309e-01
3.18179488e-01 -1.11212075e-01 8.66997242e-01 3.55289578e-01
9.52720582e-01 -2.21881330e-01 1.79719049e-02 8.30473602e-01
1.75904310e+00 -3.72149572e-02 8.74597311e-01 3.68265569e-01
6.01047873e-01 3.34083587e-01 2.07236916e-01 3.06622952e-01
1.80106640e-01 8.20773721e-01 2.29359031e-01 -1.64055035e-01
-6.74570024e-01 -4.03105080e-01 2.75308639e-01 6.74897790e-01
-3.77671361e-01 -1.81177128e-02 -4.29426938e-01 4.19598132e-01
-1.35955822e+00 -9.78166342e-01 1.75862521e-01 2.37724495e+00
1.02860260e+00 2.52566993e-01 1.48401693e-01 1.47844434e-01
6.00951433e-01 5.00571430e-01 -6.26445830e-01 -1.25037104e-01
-4.69764680e-01 5.26339710e-01 3.91786337e-01 4.62890059e-01
-7.99059510e-01 8.72455776e-01 6.22482443e+00 1.12160850e+00
-1.41068697e+00 3.08234274e-01 7.60325074e-01 -1.13796972e-01
-5.60517788e-01 7.73403049e-02 -8.68378341e-01 4.29631025e-01
1.08677816e+00 -1.30473241e-01 7.61818111e-01 4.98753130e-01
2.49548391e-01 8.29856321e-02 -8.34468007e-01 8.56694341e-01
-8.65176767e-02 -1.44562066e+00 1.79757789e-01 1.18742839e-01
8.62051368e-01 -2.91616712e-02 3.67180824e-01 6.08891733e-02
2.28573337e-01 -1.01025820e+00 6.93992198e-01 4.67990547e-01
1.03919911e+00 -5.00266552e-01 2.34601498e-01 2.09169671e-01
-1.20036530e+00 -7.02522844e-02 -6.09243989e-01 2.49918565e-01
2.07420066e-01 8.88034642e-01 -3.50601405e-01 6.05905712e-01
5.74552119e-01 2.61600375e-01 -5.53833187e-01 7.21790075e-01
-8.42464566e-02 5.38228452e-01 -1.17377155e-01 5.66032410e-01
1.59689169e-02 -2.39261210e-01 7.80417919e-01 1.20687127e+00
6.00618839e-01 2.84243137e-01 -2.06313759e-01 1.24338412e+00
-1.48317125e-02 -1.64786354e-01 -2.57045567e-01 1.12189353e-01
4.95663434e-01 1.34882522e+00 -7.50332296e-01 -2.73177236e-01
-2.67767429e-01 1.10485232e+00 2.21303046e-01 6.02772772e-01
-7.87891746e-01 -5.63747473e-02 5.27818739e-01 5.56002915e-01
5.23626983e-01 -3.84302169e-01 -4.49422836e-01 -1.31657624e+00
9.70537215e-02 -1.04959607e+00 4.28064056e-02 -6.25515103e-01
-1.23283863e+00 8.67273986e-01 -2.60155499e-01 -1.32283437e+00
-2.05754772e-01 -1.57002226e-01 -2.54917711e-01 1.00698566e+00
-1.89890373e+00 -1.06873858e+00 -3.43699843e-01 4.89465684e-01
5.24800420e-01 8.78032818e-02 5.43132603e-01 4.39648181e-01
-4.24367905e-01 6.53468072e-01 3.52228098e-02 -2.41286695e-01
6.16495311e-01 -1.01155210e+00 4.16292101e-01 1.27193904e+00
2.04918280e-01 6.56325221e-01 9.68394578e-01 -4.50512916e-01
-1.36286116e+00 -9.81003284e-01 6.06209159e-01 -3.97351980e-01
6.43395662e-01 -3.89018029e-01 -1.26098299e+00 2.38549218e-01
2.79085487e-02 2.56425381e-01 2.25617617e-01 -1.70593753e-01
-5.03684402e-01 -2.25068882e-01 -1.20608723e+00 5.85220337e-01
1.26911688e+00 -5.73891282e-01 -3.80818546e-01 -1.33472413e-01
8.34627330e-01 -2.62878984e-01 -9.44705307e-01 5.09387553e-01
3.66656095e-01 -1.29353797e+00 1.24511242e+00 -1.93188727e-01
6.82128489e-01 -5.21384060e-01 -3.42407316e-01 -1.12616801e+00
-5.55486321e-01 -8.02708030e-01 -1.32147983e-01 1.37795866e+00
2.81891644e-01 -5.36077499e-01 4.94586200e-01 4.73657191e-01
-9.38818499e-04 -7.12754130e-01 -8.33385527e-01 -9.77730274e-01
8.36121216e-02 -5.79145066e-02 7.19478786e-01 9.45126414e-01
-4.28037733e-01 3.28370780e-01 -5.08387089e-01 4.49048579e-01
8.91193867e-01 8.69279653e-02 6.02114856e-01 -1.04248524e+00
-6.70846701e-01 -4.24211830e-01 -2.45109141e-01 -1.18832612e+00
-3.72509390e-01 -4.90308464e-01 -9.73045379e-02 -1.29267299e+00
2.22254544e-01 -6.16830289e-01 -4.14315462e-01 -1.72532145e-02
-3.26693773e-01 4.13529962e-01 2.24871546e-01 5.31383336e-01
-3.65070313e-01 4.22398031e-01 1.44322169e+00 1.36017054e-01
-2.98178762e-01 -2.47634023e-01 -9.98052180e-01 5.98958492e-01
5.51892757e-01 -2.70983845e-01 -5.69644928e-01 -3.91511798e-01
1.34342417e-01 1.88563406e-01 4.54388827e-01 -1.26018429e+00
8.60200673e-02 -2.70244926e-01 4.05793309e-01 -1.24586985e-01
3.83159667e-01 -5.18626869e-01 3.73876899e-01 2.54023671e-02
-3.46146911e-01 -2.24568084e-01 -5.43287434e-02 5.48955262e-01
-2.39959434e-01 -1.12697780e-01 1.24207997e+00 -1.93465188e-01
-6.63598597e-01 2.57417470e-01 -9.31375101e-02 -1.30939456e-02
6.46191895e-01 -4.11464870e-01 -3.03424925e-01 -2.02340290e-01
-6.14990890e-01 -4.95971769e-01 7.02541530e-01 3.25108409e-01
5.36187351e-01 -1.07153118e+00 -8.20065856e-01 3.57717097e-01
-2.31196940e-01 -1.38964295e-01 5.40709615e-01 7.32862413e-01
4.04770067e-03 5.66109158e-02 -2.36214012e-01 -4.15635139e-01
-8.18190753e-01 7.53007948e-01 3.23013008e-01 -4.92918015e-01
-8.97492468e-01 7.65781581e-01 4.13234055e-01 1.88004911e-01
-2.69130826e-01 -1.64472848e-01 -4.58448976e-02 -1.52216911e-01
7.76401877e-01 1.31724313e-01 6.46768436e-02 -5.51959097e-01
-4.23509032e-02 6.55213416e-01 -1.25203609e-01 -2.41371438e-01
1.35788524e+00 -4.58918065e-01 1.29963621e-01 1.67447031e-01
8.88109446e-01 1.38098717e-01 -1.68436468e+00 -4.17962283e-01
-1.84331894e-01 -6.96172357e-01 2.33510882e-01 -6.51410222e-01
-1.12291086e+00 8.13731015e-01 6.36080801e-01 2.90095210e-01
1.46708179e+00 1.78790838e-02 8.91377509e-01 -3.04623783e-01
6.13395929e-01 -8.93145204e-01 3.28726977e-01 8.65283981e-02
8.46670091e-01 -1.02929258e+00 2.13533174e-02 -4.07046944e-01
-4.52934802e-01 8.46121013e-01 4.07394588e-01 -2.42109343e-01
2.92676419e-01 4.73281741e-01 -2.17111915e-01 -4.25671431e-04
-6.84042513e-01 -3.16991448e-01 2.87600338e-01 5.75057566e-01
2.06685290e-01 -2.27342397e-02 -2.52324730e-01 6.22269571e-01
-2.31962949e-01 1.86699107e-01 5.81726909e-01 5.18923938e-01
-6.20297372e-01 -1.04966700e+00 -3.43524277e-01 2.27807671e-01
-5.45108676e-01 -2.52831429e-01 1.39062037e-03 5.80448389e-01
1.31368011e-01 7.81344056e-01 -9.36228707e-02 -4.07483757e-01
2.36799136e-01 -2.95920581e-01 6.74125910e-01 -3.93429518e-01
-2.75545329e-01 2.74808168e-01 -1.78585514e-01 -5.40324152e-01
-5.06621838e-01 -4.63524044e-01 -9.18292522e-01 -2.86452085e-01
-2.55322695e-01 -8.31879079e-02 4.43839818e-01 8.56457889e-01
5.00749767e-01 5.26776612e-01 6.92147076e-01 -9.69874680e-01
-6.42150819e-01 -7.39567161e-01 -6.24657035e-01 3.06751549e-01
4.16133881e-01 -4.93274510e-01 -4.29468334e-01 4.79242653e-02] | [11.208745002746582, -1.9269622564315796] |
33799f05-bea7-489f-af75-3485b8887941 | reflectance-hashing-for-material-recognition | 1502.02092 | null | http://arxiv.org/abs/1502.02092v1 | http://arxiv.org/pdf/1502.02092v1.pdf | Reflectance Hashing for Material Recognition | We introduce a novel method for using reflectance to identify materials.
Reflectance offers a unique signature of the material but is challenging to
measure and use for recognizing materials due to its high-dimensionality. In
this work, one-shot reflectance is captured using a unique optical camera
measuring {\it reflectance disks} where the pixel coordinates correspond to
surface viewing angles. The reflectance has class-specific stucture and angular
gradients computed in this reflectance space reveal the material class.
These reflectance disks encode discriminative information for efficient and
accurate material recognition. We introduce a framework called reflectance
hashing that models the reflectance disks with dictionary learning and binary
hashing. We demonstrate the effectiveness of reflectance hashing for material
recognition with a number of real-world materials. | ['Kristin Dana', 'Ko Nishino', 'Hang Zhang'] | 2015-02-07 | reflectance-hashing-for-material-recognition-1 | http://openaccess.thecvf.com/content_cvpr_2015/html/Zhang_Reflectance_Hashing_for_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Zhang_Reflectance_Hashing_for_2015_CVPR_paper.pdf | cvpr-2015-6 | ['material-recognition'] | ['computer-vision'] | [ 6.25516176e-01 -6.77266777e-01 -2.18695581e-01 -1.11769237e-01
-7.05942631e-01 -6.01984739e-01 2.41184130e-01 -1.96854040e-01
2.01868221e-01 1.26415659e-02 9.63480622e-02 3.13137323e-01
4.69493791e-02 -1.00480580e+00 -4.71528679e-01 -8.78451109e-01
3.40138555e-01 3.37967277e-01 1.04969777e-01 1.23111516e-01
6.18066549e-01 1.00249803e+00 -1.60540855e+00 5.43633759e-01
1.40581960e-02 1.34386063e+00 3.31839055e-01 8.01739991e-01
-1.68454852e-02 1.17364883e-01 -4.28776026e-01 -2.26549760e-01
7.96633184e-01 2.37490073e-01 -2.82850981e-01 3.34800124e-01
9.09308136e-01 -8.67377400e-01 -6.59545898e-01 9.36260462e-01
3.35400045e-01 -9.07345191e-02 1.10811174e+00 -6.43344998e-01
-1.19331563e+00 5.57181891e-03 -7.96495199e-01 -1.29850730e-01
1.05756187e+00 -9.78096426e-02 1.34797990e+00 -1.10937250e+00
4.94345188e-01 1.10381973e+00 6.33826256e-01 4.47329789e-01
-1.13322544e+00 -1.84089601e-01 -4.85796034e-01 -7.48076523e-03
-1.68117118e+00 -4.03746158e-01 1.24175560e+00 -5.76495290e-01
7.97201276e-01 4.03430730e-01 7.92104781e-01 8.62890244e-01
5.22536971e-02 4.89499241e-01 1.26710844e+00 -2.81470984e-01
-3.26353800e-03 6.47691935e-02 2.73756146e-01 1.06661093e+00
4.11720008e-01 1.31942987e-01 -7.97013700e-01 -2.25484744e-01
9.97128069e-01 6.39369309e-01 -3.35293233e-01 -4.38189894e-01
-1.12804163e+00 3.59923899e-01 3.12361121e-01 -3.26372832e-01
-2.73699850e-01 2.01707631e-01 -4.23581749e-02 3.71791720e-01
2.24943057e-01 3.03387761e-01 1.15057997e-01 -4.15234305e-02
-4.46238041e-01 -2.80581951e-01 7.56125748e-01 8.54650497e-01
1.08845913e+00 -1.63028777e-01 4.48923379e-01 1.28300238e+00
3.89687747e-01 1.34133840e+00 1.12120204e-01 -7.23497748e-01
2.17483580e-01 6.57182097e-01 5.97387366e-02 -1.31249535e+00
-1.49138970e-02 7.06092954e-01 -4.97265697e-01 -3.90193518e-03
1.49176747e-01 9.15655136e-01 -6.97924137e-01 8.50347400e-01
2.60896564e-01 -1.25736803e-01 6.54344931e-02 1.08870518e+00
1.01231360e+00 6.05186045e-01 -8.16992521e-01 2.05673829e-01
1.40348709e+00 -6.05336785e-01 -3.82357180e-01 1.63781807e-01
-1.20020546e-01 -1.04853630e+00 1.36951530e+00 7.23293006e-01
-9.00544822e-01 -3.18718851e-01 -1.24071467e+00 -2.31530458e-01
-3.87934089e-01 2.68424630e-01 6.24652624e-01 9.50233042e-01
-5.78449011e-01 2.22921506e-01 -8.20994616e-01 -7.08884969e-02
2.79635638e-01 2.16296434e-01 -2.58885503e-01 -6.17657661e-01
-4.96883869e-01 2.36615598e-01 -4.58348632e-01 4.76724170e-02
-8.95243764e-01 -3.92523825e-01 -7.03078687e-01 -5.28142750e-01
-3.89562324e-02 -2.01005757e-01 5.73949039e-01 -3.74637157e-01
-1.64359641e+00 1.32381594e+00 -1.69984475e-01 3.19914460e-01
6.88450113e-02 -2.70114571e-01 -4.68556941e-01 8.08985710e-01
-3.68197471e-01 1.04496658e-01 1.44048452e+00 -1.47030127e+00
2.32695341e-01 -5.93609273e-01 -4.77879532e-02 9.01012272e-02
-6.08906329e-01 6.84965327e-02 -2.64690816e-01 -4.53574121e-01
6.72858000e-01 -8.39981318e-01 4.65125829e-01 5.23282588e-01
-5.44514537e-01 2.50542253e-01 1.01703012e+00 -4.42506582e-01
8.96921992e-01 -2.44003582e+00 -1.35654822e-01 3.24653387e-01
3.97475034e-01 -2.50002861e-01 -1.01703048e-01 6.48897350e-01
3.08224559e-01 -1.33444458e-01 1.31668329e-01 -2.89556552e-02
-1.03083827e-01 -1.50937932e-02 -6.20019257e-01 1.04911113e+00
5.85960820e-02 4.60803241e-01 -6.63236201e-01 -1.69558033e-01
3.28673989e-01 8.83994818e-01 -9.48509723e-02 2.84333229e-01
-1.05743222e-02 -9.75627974e-02 -4.17302191e-01 1.38274407e+00
8.17055464e-01 -1.86450481e-01 -2.51580983e-01 -1.00031614e+00
3.90435942e-02 2.80749261e-01 -1.09760594e+00 1.32016540e+00
-2.37011164e-01 5.89634061e-01 1.13808764e-02 -4.06605721e-01
1.55053496e+00 -1.60055906e-01 7.97483444e-01 -5.47183275e-01
-2.77308822e-01 2.37527311e-01 -7.47363985e-01 -5.70440352e-01
9.63807940e-01 -2.02785269e-03 6.05178289e-02 9.66248810e-01
-6.44125462e-01 -4.90720540e-01 -4.88462031e-01 2.08833683e-02
1.20129418e+00 -1.33447245e-01 -2.65843794e-02 7.68954158e-02
2.48421848e-01 -1.91704407e-01 2.49189995e-02 4.84400898e-01
1.38031906e-02 1.11639071e+00 -4.08386737e-01 -6.92812204e-01
-1.17104769e+00 -1.54291487e+00 -4.31187391e-01 8.12863588e-01
7.84342706e-01 -3.37843776e-01 -1.58528030e-01 -7.92133138e-02
4.23308015e-01 -8.65124986e-02 -4.60516930e-01 -3.41836512e-01
-5.56300163e-01 -3.48572671e-01 2.60683149e-01 4.65403765e-01
3.59904557e-01 -7.52433777e-01 -6.58605456e-01 -4.45081174e-01
1.36313155e-01 -1.11046851e+00 -7.11634636e-01 -2.34619930e-01
-7.82543600e-01 -1.22291708e+00 -4.46516991e-01 -7.20365465e-01
9.36590612e-01 1.32971084e+00 1.12410712e+00 -3.89274617e-04
-1.18113089e+00 1.29269850e+00 -3.72442126e-01 3.05306315e-02
-1.62130430e-01 -4.14816469e-01 1.52146623e-01 2.88522810e-01
5.09849846e-01 -2.71251082e-01 -9.53396499e-01 6.43875360e-01
-7.05845594e-01 -1.52434662e-01 4.26161736e-01 5.17473757e-01
1.05476677e+00 4.60460261e-02 -1.55080035e-01 -5.06926060e-01
3.32410008e-01 -1.98883012e-01 -4.91796821e-01 3.63429546e-01
-3.24534804e-01 -1.79419518e-01 5.33241987e-01 -5.32158434e-01
-5.54450452e-01 -7.04853563e-03 5.56484401e-01 -5.87510526e-01
1.26111627e-01 -4.13393378e-02 2.31638514e-02 -4.96731341e-01
5.79814613e-01 3.76312643e-01 6.43337443e-02 -5.24088323e-01
2.79305905e-01 9.17229533e-01 4.63955998e-01 -9.64365959e-01
9.13382411e-01 1.04832757e+00 2.46679708e-02 -1.33525765e+00
-6.92241251e-01 -8.37970197e-01 -6.47885442e-01 -4.50762331e-01
4.87345845e-01 -9.46806610e-01 -1.07030821e+00 6.04125857e-01
-8.37560415e-01 -1.47305861e-01 -4.56525475e-01 1.93697423e-01
-6.98101342e-01 5.11943161e-01 -8.69027376e-01 -1.05724096e+00
-5.24790525e-01 -9.50489342e-01 1.76923335e+00 -1.51052654e-01
2.08089083e-01 -8.71409655e-01 1.50439844e-01 5.47296524e-01
2.37098530e-01 -1.03292346e-01 8.21823001e-01 2.69044638e-01
-1.05643713e+00 -2.90346652e-01 -5.68595767e-01 2.98588812e-01
6.10417604e-01 4.82207149e-01 -1.08690238e+00 -4.26348150e-01
7.67845986e-03 -5.06157339e-01 7.97855079e-01 -1.06533065e-01
1.34471166e+00 -1.23700842e-01 1.00337192e-02 7.59453118e-01
1.72147322e+00 -4.66424227e-02 7.28728116e-01 7.45634213e-02
1.07705665e+00 3.82385105e-01 5.39061010e-01 6.59095109e-01
3.38123828e-01 7.89212465e-01 1.70192584e-01 1.68752238e-01
-4.58434939e-01 -1.39149293e-01 8.74614954e-01 1.21105587e+00
-6.68577775e-02 6.96505681e-02 -7.43704796e-01 2.01509550e-01
-8.90915334e-01 -1.04331911e+00 1.89946234e-01 2.50767136e+00
8.69133830e-01 -3.61314923e-01 2.43427143e-01 1.09864928e-01
7.13424504e-01 4.28229392e-01 -7.12255239e-01 -3.01272362e-01
-4.16848630e-01 6.27756938e-02 6.18291020e-01 2.77488977e-01
-7.06168532e-01 5.99746466e-01 7.62154627e+00 3.26170564e-01
-1.38244045e+00 -3.18541646e-01 8.77457410e-02 1.06000900e-01
-6.83122694e-01 -4.11126077e-01 -1.07757807e+00 2.33828858e-01
6.11095667e-01 2.89594650e-01 9.43896830e-01 7.03197420e-01
-2.15978384e-01 -1.52224302e-01 -1.08206129e+00 1.59355056e+00
6.47548318e-01 -9.68594849e-01 1.87310129e-01 -9.25644417e-04
5.03853559e-01 -1.57898758e-02 6.44858897e-01 -6.48835182e-01
-5.00079663e-03 -8.59382987e-01 5.72231412e-01 7.58836508e-01
1.49877906e+00 -1.57464594e-01 -2.17501670e-01 -3.74392182e-01
-1.47136927e+00 1.57837793e-02 -8.85390282e-01 3.49131852e-01
-2.65761137e-01 6.69137120e-01 -1.16473913e+00 3.45433950e-02
6.51773930e-01 1.19475222e+00 -3.69707048e-01 7.62050509e-01
2.24442989e-01 2.74129778e-01 -4.95449662e-01 -8.20789784e-02
-4.16543245e-01 -8.25542212e-01 5.25049746e-01 1.01771021e+00
4.81675357e-01 4.06539291e-02 3.60523731e-01 8.92851532e-01
-2.63264835e-01 -5.72832227e-02 -1.00226843e+00 -3.22703242e-01
7.39361107e-01 1.17165124e+00 -6.46878481e-01 1.91816613e-02
-6.09720469e-01 1.03923690e+00 1.08740300e-01 1.71088070e-01
-3.36637229e-01 -1.90932021e-01 8.07064950e-01 3.46721113e-01
3.86960596e-01 -8.75755072e-01 -3.83639693e-01 -1.09993601e+00
2.72203803e-01 -5.62601030e-01 -7.96126500e-02 -9.51067626e-01
-1.82244813e+00 3.72134964e-04 -3.41230601e-01 -1.49908113e+00
7.74278998e-01 -1.19118476e+00 4.26380821e-02 5.66760302e-01
-1.48486125e+00 -9.52107310e-01 -7.98642933e-01 8.46373856e-01
2.10483938e-01 -1.97494566e-01 9.91624594e-01 1.73142076e-01
-4.42409664e-01 3.97934437e-01 4.28088725e-01 1.34957582e-01
7.54988432e-01 -1.18259668e+00 2.86938608e-01 1.45684406e-01
3.31165940e-01 8.25737000e-01 3.23827744e-01 -6.26561821e-01
-2.82481503e+00 -8.28257859e-01 -1.17318764e-01 -4.69620436e-01
6.28855765e-01 -5.01710176e-01 -6.19840205e-01 2.29451954e-01
-5.13875246e-01 1.52838200e-01 1.11978102e+00 -1.14938684e-01
-1.42873621e+00 -6.38084650e-01 -1.48327470e+00 3.26802641e-01
1.10348737e+00 -1.53998041e+00 -3.74787837e-01 5.82884312e-01
4.87503171e-01 -2.43187219e-01 -1.31681395e+00 7.73880724e-03
1.04954314e+00 -7.34702408e-01 1.47555864e+00 2.11927295e-01
3.67419183e-01 -2.42760018e-01 -8.16942096e-01 -5.93167841e-01
-1.31090239e-01 -5.21193445e-01 -2.48119354e-01 8.41033876e-01
-3.02593671e-02 -7.15510488e-01 9.44228053e-01 5.28951824e-01
-9.56513360e-03 -4.42864150e-01 -4.77206558e-01 -8.92368853e-01
-6.92589998e-01 -2.67197967e-01 5.93945801e-01 7.34803915e-01
-3.37924987e-01 -5.26674278e-03 -3.05746049e-01 1.89904109e-01
9.91112828e-01 8.01114500e-01 6.12105012e-01 -1.15020061e+00
-2.97942549e-01 -1.92937076e-01 -6.66947186e-01 -1.44543386e+00
-7.75327682e-02 -8.00423384e-01 1.22111244e-02 -1.06156111e+00
5.47387362e-01 -9.44537699e-01 -1.12500124e-01 1.73170909e-01
2.28854075e-01 5.48648238e-01 -1.26676112e-01 6.80896401e-01
-7.28775203e-01 4.66528475e-01 1.25814188e+00 -5.57300866e-01
9.07668844e-03 -2.83899337e-01 -3.54246199e-01 1.51790395e-01
5.21590710e-01 -1.62556209e-02 -1.83580816e-01 -6.22096002e-01
5.25838017e-01 -1.67674646e-01 3.65830570e-01 -8.58443201e-01
-1.81375612e-02 -2.26738274e-01 2.43570164e-01 -8.17442656e-01
8.98817778e-01 -1.04755306e+00 2.99329549e-01 1.03290826e-01
-1.85844451e-01 2.25463495e-01 -3.86449218e-01 8.10621142e-01
7.24149272e-02 -1.05505474e-01 6.91239357e-01 -3.05852219e-02
-5.65838516e-01 5.40394962e-01 1.62396938e-01 -3.20837975e-01
7.49507368e-01 -7.68395185e-01 -8.40115309e-01 -6.32125512e-02
-1.09828778e-01 -6.80037141e-01 1.35308218e+00 3.80919576e-02
1.47649908e+00 -1.49626613e+00 -2.55636811e-01 5.45395851e-01
4.71048951e-01 -2.18524978e-01 -5.63219711e-02 3.14061314e-01
-8.68021846e-01 1.24440910e-02 -2.07748786e-01 -8.77032042e-01
-1.33279109e+00 5.93997002e-01 1.62728146e-01 4.38066065e-01
-9.37470436e-01 7.62819171e-01 -2.30115931e-02 -1.48095787e-01
6.95715398e-02 2.17984989e-02 3.46174687e-01 -2.10896760e-01
7.07187414e-01 6.04788184e-01 1.86545670e-01 -7.32808828e-01
-1.62140131e-01 1.30924559e+00 -9.51266661e-02 -2.55687274e-02
1.25926614e+00 -1.59038350e-01 -4.32778597e-01 8.97660375e-01
1.48397386e+00 3.77487570e-01 -1.13606644e+00 -4.66584057e-01
-4.73712087e-01 -1.17563820e+00 -7.30602369e-02 -4.26291019e-01
-1.29969120e+00 9.02935743e-01 4.64543134e-01 4.01503026e-01
1.01455855e+00 -5.00205159e-02 1.03930449e+00 6.63396299e-01
8.88463676e-01 -1.02394736e+00 6.66994214e-01 2.14791670e-01
8.87678742e-01 -1.04883313e+00 3.83617401e-01 -6.72009766e-01
-2.33014643e-01 1.57473564e+00 1.61033809e-01 -2.89958477e-01
6.71234429e-01 3.79781038e-01 -6.12819102e-03 -6.96522176e-01
-3.10198903e-01 1.70035735e-01 3.21422607e-01 7.19143867e-01
1.96472242e-01 4.71633017e-01 4.09392625e-01 -1.78718299e-01
-1.67045623e-01 -7.19123125e-01 3.25042069e-01 7.83671618e-01
-6.07588649e-01 -9.91283655e-01 -7.47753799e-01 5.07322907e-01
-6.40775487e-02 8.02873373e-02 -5.64969659e-01 -1.86955594e-02
-6.78769648e-01 6.86850309e-01 1.76746532e-01 -8.26790392e-01
7.61512592e-02 -3.80508512e-01 1.00290537e+00 -7.60294259e-01
-5.24404980e-02 -1.39213905e-01 -1.66451558e-01 -7.87684739e-01
-5.60766280e-01 -8.85572970e-01 -1.00508523e+00 -2.47042865e-01
-2.53013253e-01 -4.63514477e-01 9.46624458e-01 1.78051770e-01
2.00019106e-01 -2.89516121e-01 1.43911660e+00 -7.32059658e-01
-5.55902421e-01 -2.25094259e-01 -9.90173578e-01 7.07190394e-01
4.39805359e-01 -8.12623024e-01 -5.11452913e-01 1.05914317e-01] | [9.766813278198242, -2.840467929840088] |
a097db13-5b3b-4a11-9c1b-b5c19b111b99 | contextual-blocking-bandits | 2003.03426 | null | https://arxiv.org/abs/2003.03426v2 | https://arxiv.org/pdf/2003.03426v2.pdf | Contextual Blocking Bandits | We study a novel variant of the multi-armed bandit problem, where at each time step, the player observes an independently sampled context that determines the arms' mean rewards. However, playing an arm blocks it (across all contexts) for a fixed and known number of future time steps. The above contextual setting, which captures important scenarios such as recommendation systems or ad placement with diverse users, invalidates greedy solution techniques that are effective for its non-contextual counterpart (Basu et al., NeurIPS19). Assuming knowledge of the context distribution and the mean reward of each arm-context pair, we cast the problem as an online bipartite matching problem, where the right-vertices (contexts) arrive stochastically and the left-vertices (arms) are blocked for a finite number of rounds each time they are matched. This problem has been recently studied in the full-information case, where competitive ratio bounds have been derived. We focus on the bandit setting, where the reward distributions are initially unknown; we propose a UCB-based variant of the full-information algorithm that guarantees a $\mathcal{O}(\log T)$-regret w.r.t. an $\alpha$-optimal strategy in $T$ time steps, matching the $\Omega(\log(T))$ regret lower bound in this setting. Due to the time correlations caused by blocking, existing techniques for upper bounding regret fail. For proving our regret bounds, we introduce the novel concepts of delayed exploitation and opportunistic subsampling and combine them with ideas from combinatorial bandits and non-stationary Markov chains coupling. | ['Sanjay Shakkottai', 'Constantine Caramanis', 'Soumya Basu', 'Orestis Papadigenopoulos'] | 2020-03-06 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [ 2.81812310e-01 2.19021961e-01 -7.60964811e-01 -7.91047662e-02
-9.21300352e-01 -1.03594661e+00 -1.15255676e-01 4.73028794e-02
-6.29437447e-01 1.00779545e+00 -6.57084584e-02 -7.69105673e-01
-9.14880812e-01 -8.25281322e-01 -1.12737501e+00 -9.82698739e-01
-2.24564731e-01 1.00714421e+00 -5.03575467e-02 1.05524138e-02
8.87411907e-02 4.59365934e-01 -1.05679536e+00 1.68180406e-01
3.80072713e-01 1.13132107e+00 7.72827789e-02 8.18775237e-01
-3.54357928e-01 5.89200079e-01 -1.56322017e-01 -5.94902813e-01
7.60044277e-01 -6.61243677e-01 -9.58542168e-01 4.88662906e-02
-1.64403573e-01 -3.67821038e-01 -4.36305881e-01 1.06136906e+00
2.65713781e-01 2.96005845e-01 2.99869895e-01 -1.20725501e+00
-1.17177688e-01 1.28684831e+00 -1.05837345e+00 3.23255450e-01
1.66584715e-01 -2.85371244e-01 1.31825745e+00 2.99385548e-01
5.20780623e-01 9.48789299e-01 2.40897551e-01 5.54862440e-01
-1.54624355e+00 -6.68785453e-01 6.67218149e-01 -7.18093887e-02
-8.05218816e-01 -1.68037385e-01 5.72523475e-01 -7.15380162e-02
4.64100003e-01 8.00446570e-01 7.62728453e-01 7.71991968e-01
-4.14778829e-01 1.24992824e+00 1.13088298e+00 -6.79149210e-01
4.42779601e-01 -9.99827534e-02 3.88928354e-01 1.27533257e-01
5.55844545e-01 3.18869352e-01 -2.64581233e-01 -3.29122812e-01
5.77888787e-01 5.25932312e-01 -1.99533671e-01 -5.76550424e-01
-7.43464410e-01 7.68123984e-01 2.54308313e-01 1.69809759e-01
-5.53689718e-01 6.13849282e-01 8.10394622e-03 6.48828268e-01
3.11596721e-01 -1.64483190e-02 -5.17502725e-01 -2.42146969e-01
-7.82199919e-01 2.69569963e-01 9.62043822e-01 1.30620849e+00
6.62804902e-01 -6.30795300e-01 -5.45279741e-01 5.12496650e-01
-5.05398288e-02 7.43978500e-01 -3.38970602e-01 -8.35118234e-01
9.86541092e-01 1.39173329e-01 9.54973698e-01 -6.08274043e-01
-3.72809112e-01 -6.51440978e-01 -5.21428049e-01 -2.16568738e-01
8.13145757e-01 -4.08110321e-01 -7.30030119e-01 2.00850892e+00
3.45073611e-01 -1.19543217e-01 -3.81781816e-01 8.75675440e-01
-4.44050789e-01 4.82447475e-01 -4.12749588e-01 -8.10896933e-01
9.81657088e-01 -8.56796324e-01 -3.87365729e-01 -3.54895920e-01
4.22309160e-01 -5.31142712e-01 4.27835107e-01 5.41424096e-01
-1.48750496e+00 3.38338882e-01 -5.32812774e-01 5.41763365e-01
1.44995451e-01 -5.58539093e-01 9.19081748e-01 1.15649855e+00
-7.65925825e-01 5.32401979e-01 -7.06988096e-01 -5.55817336e-02
4.92098719e-01 5.00428081e-01 2.41648898e-01 -4.94386256e-01
-5.77906013e-01 1.88620597e-01 -4.40331213e-02 3.00742537e-01
-9.89651918e-01 -4.01149869e-01 3.85049992e-04 1.71395272e-01
1.08488071e+00 -6.78385198e-01 1.57389963e+00 -1.33876669e+00
-1.30845129e+00 4.46787924e-01 -2.34594166e-01 -4.16646749e-01
9.27271307e-01 1.39076123e-02 2.47564793e-01 2.61394214e-02
5.43440841e-02 -2.15070665e-01 6.44656301e-01 -1.01395500e+00
-1.04349935e+00 -8.19135606e-01 6.97093844e-01 1.08328998e-01
-3.16087641e-02 -1.08997911e-01 -4.72543210e-01 -2.51551837e-01
2.18283147e-01 -1.39892125e+00 -6.72206759e-01 -4.65568662e-01
-3.35449576e-01 2.25993305e-01 -2.13166296e-01 -3.95228900e-02
1.10425377e+00 -1.94121921e+00 2.11256757e-01 6.35806024e-01
-1.23697899e-01 -2.31555074e-01 -2.42909908e-01 7.12775648e-01
8.15479606e-02 3.13425392e-01 1.73589230e-01 -3.45485091e-01
1.85803175e-01 2.74060249e-01 -3.96930456e-01 7.07793236e-01
-8.63062501e-01 7.63471007e-01 -8.46805274e-01 1.42791122e-01
-2.35468492e-01 -4.10270959e-01 -8.07665586e-01 -1.74505617e-02
-5.62863231e-01 2.40944445e-01 -8.09467375e-01 3.63889515e-01
9.28868175e-01 -3.61258805e-01 7.03282952e-01 5.33844471e-01
8.93138051e-02 3.14586144e-03 -1.53387594e+00 1.33796740e+00
-7.03402102e-01 3.05482931e-02 5.58140874e-01 -1.13872230e+00
-2.18466204e-02 1.79177716e-01 4.96107280e-01 -7.04874754e-01
1.92269072e-01 2.42397130e-01 -1.57045424e-01 -2.59770811e-01
2.26245075e-01 -5.05587339e-01 -2.25017473e-01 8.56817484e-01
-4.36928242e-01 6.87218666e-01 1.54371589e-01 3.60484570e-01
1.40789235e+00 -1.89295337e-01 -5.34919603e-03 -7.13844672e-02
-1.39280200e-01 -3.19188554e-03 6.23088419e-01 1.67551351e+00
2.62686666e-02 2.69228399e-01 9.11167145e-01 -2.87100464e-01
-8.63015413e-01 -8.86395752e-01 3.34204942e-01 1.47094524e+00
5.25758028e-01 1.86073512e-01 -3.22809398e-01 -5.91354132e-01
3.75139505e-01 6.18590593e-01 -1.01602638e+00 3.19796890e-01
-1.09948523e-01 -6.67409956e-01 -2.07619011e-01 2.42274091e-01
9.17543769e-02 -5.79336941e-01 -5.65500855e-01 6.10851824e-01
-1.28653646e-01 -7.13767350e-01 -5.85208535e-01 4.46906745e-01
-9.58686411e-01 -1.04349637e+00 -7.71357834e-01 -2.02931181e-01
7.51609921e-01 7.96660662e-01 8.26747715e-01 -1.05430439e-01
2.81518400e-02 5.86596012e-01 -5.03580749e-01 -3.51769358e-01
1.47476032e-01 6.31536618e-02 -2.95896858e-01 1.96391210e-01
3.34716514e-02 -5.30528426e-01 -9.33175206e-01 4.29641992e-01
-9.69677031e-01 1.24076545e-01 6.19691074e-01 8.21943939e-01
6.25866055e-01 -1.59891173e-02 4.16752279e-01 -1.23726940e+00
2.69705743e-01 -5.97733438e-01 -1.08546591e+00 5.19507110e-01
-3.50283265e-01 1.64147198e-01 5.35389721e-01 -5.92022121e-01
-1.00193298e+00 2.30996800e-03 4.71703351e-01 -1.58006936e-01
2.14957103e-01 4.57729071e-01 -4.86883335e-04 3.65347601e-02
2.57059216e-01 1.70059904e-01 -1.92001715e-01 -5.84927082e-01
3.59472305e-01 6.32766545e-01 -1.50455207e-01 -1.01105702e+00
5.35563648e-01 6.97225571e-01 2.51263201e-01 -2.09795177e-01
-1.11530483e+00 -5.75231969e-01 1.99051976e-01 -1.40963778e-01
5.64861223e-02 -6.24525070e-01 -1.57085943e+00 9.82239768e-02
-9.26847517e-01 -5.15569091e-01 -4.00485039e-01 4.76107597e-01
-9.12431419e-01 1.10582508e-01 -2.72866398e-01 -1.81063497e+00
1.17086880e-01 -7.47330785e-01 4.76807386e-01 2.56605268e-01
4.62143451e-01 -5.15328169e-01 2.07498409e-02 6.46780550e-01
3.89864326e-01 -1.69800296e-01 7.65621841e-01 -5.86190701e-01
-1.17102766e+00 -3.09178859e-01 -9.28659812e-02 -2.56942600e-01
-2.41432026e-01 -7.50659227e-01 -3.76283795e-01 -5.04867733e-01
-2.75443673e-01 -6.34705946e-02 8.25595498e-01 7.50876069e-01
1.15849829e+00 -7.85391569e-01 -7.90408075e-01 1.67331249e-01
1.60961008e+00 5.37486970e-01 3.69616657e-01 3.75366479e-01
-1.04239024e-01 2.73478419e-01 5.24554014e-01 8.90575647e-01
8.56908876e-03 6.68763578e-01 6.74414814e-01 2.70742804e-01
6.33074105e-01 -1.87373459e-01 1.10065542e-01 -2.35157609e-01
-1.52498811e-01 -5.29961467e-01 -4.03064936e-01 9.29562807e-01
-2.27789783e+00 -1.15227020e+00 7.95249566e-02 2.83861804e+00
8.43769670e-01 2.84107417e-01 6.87556505e-01 -1.33224294e-01
8.58303666e-01 -2.14262992e-01 -8.57878625e-01 -5.24101496e-01
9.63103697e-02 2.07065850e-01 1.34074533e+00 4.99971062e-01
-5.82007110e-01 4.48418617e-01 5.41445494e+00 9.73142326e-01
-5.26162863e-01 3.64548504e-01 8.27987909e-01 -1.07101250e+00
-6.31632626e-01 3.80271554e-01 -6.55331790e-01 5.35438538e-01
8.78232002e-01 -2.48238891e-01 1.26985061e+00 8.60540807e-01
5.17037697e-02 -4.50548768e-01 -1.23464072e+00 9.10342157e-01
-4.71458733e-01 -1.34773660e+00 -5.67538559e-01 5.08045435e-01
1.00430620e+00 -9.41782221e-02 9.79977623e-02 1.96133479e-01
9.79435802e-01 -6.09736621e-01 8.66795480e-01 2.96273381e-01
4.88514721e-01 -1.15874946e+00 4.55105573e-01 6.12213254e-01
-8.49572361e-01 -8.78675759e-01 -1.05814047e-01 -3.17064404e-01
2.98491567e-01 7.25133717e-01 -2.65783399e-01 8.51600528e-01
5.53790510e-01 -9.67767239e-02 4.39921796e-01 1.31472933e+00
1.32509485e-01 5.91749072e-01 -7.97977984e-01 -2.93011099e-01
3.71415883e-01 -5.78388274e-01 5.46422660e-01 7.44413912e-01
2.42523372e-01 4.29045796e-01 5.15539162e-02 4.09199297e-01
-2.04421759e-01 7.06174076e-02 -2.19977960e-01 -6.06237128e-02
4.18975413e-01 8.42315793e-01 -9.08023179e-01 -1.01172686e-01
-3.57137054e-01 8.33262146e-01 2.37706840e-01 4.70619649e-01
-8.56289029e-01 -1.82482243e-01 6.23084784e-01 3.26570310e-02
7.77038395e-01 1.72961175e-01 -4.42297846e-01 -8.26895654e-01
2.34775990e-01 -3.82479042e-01 7.06097424e-01 -1.90111548e-01
-1.25016546e+00 -6.91829547e-02 1.17848366e-02 -7.95084596e-01
1.52805438e-02 -1.83465436e-01 -2.13753626e-01 6.66804612e-01
-1.49528122e+00 -7.82122195e-01 4.02350247e-01 5.64012408e-01
1.60302773e-01 2.77254403e-01 3.64816070e-01 1.52768284e-01
-6.15646958e-01 6.88319743e-01 9.39242542e-01 -2.81634986e-01
-3.48698907e-02 -1.08004653e+00 -2.35847503e-01 7.20603108e-01
1.02850407e-01 4.48283046e-01 8.02000284e-01 -3.34345967e-01
-1.90619206e+00 -6.82814181e-01 4.27373320e-01 -7.58612826e-02
7.14050114e-01 -2.78383523e-01 6.30798116e-02 8.53017986e-01
-1.96406245e-01 -2.64883765e-05 6.28027022e-01 5.66576302e-01
-1.42245606e-01 -5.11078000e-01 -1.16440213e+00 7.00375021e-01
1.51770294e+00 -8.00767466e-02 -2.43873764e-02 3.44253361e-01
4.91001248e-01 -2.65264720e-01 -4.11316067e-01 -1.61459744e-02
8.25200856e-01 -9.18031633e-01 6.49533510e-01 -1.03988719e+00
7.49957468e-03 1.48855194e-01 -2.71724522e-01 -9.52269018e-01
-2.86161095e-01 -1.28199685e+00 -1.20051488e-01 8.67689312e-01
7.39693820e-01 -7.68250883e-01 1.09261811e+00 9.83636737e-01
4.26033080e-01 -5.75750530e-01 -1.42710376e+00 -1.01666820e+00
1.02591217e-01 -5.67837596e-01 6.09680355e-01 3.64450574e-01
2.55135030e-01 -6.67218072e-03 -6.49802625e-01 4.74905707e-02
7.00128555e-01 9.15684223e-01 6.32601082e-01 -8.36972713e-01
-1.02889061e+00 -5.01617730e-01 3.32114369e-01 -1.54166365e+00
-1.82123601e-01 -5.64907908e-01 -1.83411948e-02 -1.29544175e+00
8.61800075e-01 -9.90381658e-01 -6.16261601e-01 4.72847432e-01
1.81854323e-01 -2.21459344e-01 4.91809309e-01 -8.09601694e-03
-9.80392516e-01 4.17524651e-02 1.02544212e+00 -2.01556087e-01
-3.74635696e-01 8.54853511e-01 -1.02588236e+00 6.69747517e-02
5.24891198e-01 -8.06556702e-01 -3.93251538e-01 -3.08021665e-01
8.60925078e-01 9.95544791e-01 1.08819054e-02 -3.29963118e-01
2.28260592e-01 -8.07571113e-01 -1.35954529e-01 -6.61930978e-01
1.37291506e-01 -1.27036178e+00 6.16322875e-01 6.05567694e-01
-6.82782769e-01 -3.97380143e-01 -1.35913521e-01 1.31298208e+00
6.46048248e-01 -4.98350859e-01 4.87878531e-01 -1.97260916e-01
3.27030957e-01 5.06224036e-01 -2.97618181e-01 3.40409353e-02
1.34294629e+00 -2.75181860e-01 -1.22202657e-01 -7.80490100e-01
-9.21165943e-01 3.55567575e-01 3.01394612e-01 -2.05428824e-01
7.33477250e-02 -9.36065793e-01 -3.10361028e-01 -1.92552730e-01
-1.10325009e-01 -7.04737306e-02 7.34350383e-01 1.08486032e+00
1.38246134e-01 5.29104471e-01 1.73665062e-01 -1.58407390e-01
-1.00620115e+00 1.03112078e+00 1.75285488e-01 -5.42954981e-01
-1.61000878e-01 8.58509839e-01 6.39253482e-02 3.57789129e-01
3.07132095e-01 -1.66466963e-02 5.24728060e-01 7.23635778e-02
4.76898372e-01 5.33707440e-01 -9.93040726e-02 2.08944231e-01
-1.88816965e-01 3.58894542e-02 -4.80088711e-01 -6.04390442e-01
1.47967088e+00 -5.04308879e-01 1.87575091e-02 1.62159741e-01
9.16804373e-01 1.62340432e-01 -1.19733310e+00 -5.28882086e-01
4.86547574e-02 -9.11300361e-01 3.27607021e-02 -9.99157190e-01
-1.29189372e+00 4.23511535e-01 3.23714972e-01 7.73535609e-01
1.16580331e+00 1.70223624e-01 5.51537991e-01 2.89152890e-01
9.89191949e-01 -1.03845263e+00 -3.27528298e-01 3.06453705e-01
3.39804888e-01 -6.46198511e-01 -3.00485224e-01 -7.41463667e-03
-3.11338961e-01 7.32127309e-01 1.43632423e-02 1.40178293e-01
4.77652371e-01 9.86186638e-02 -6.04674041e-01 2.08063602e-01
-8.59257102e-01 -5.20016313e-01 -4.12777334e-01 2.19314501e-01
-1.13820188e-01 4.80537295e-01 -6.03732884e-01 8.53593349e-01
3.17881197e-01 1.74036130e-01 5.57935417e-01 1.15855134e+00
-3.99038583e-01 -1.39677250e+00 -4.76750374e-01 7.24780023e-01
-8.80276680e-01 1.81269366e-02 -2.10331514e-01 3.63901883e-01
-4.35445070e-01 1.11272502e+00 4.88434657e-02 1.65778473e-01
2.75843799e-01 -1.28654405e-01 8.53826523e-01 -3.36185753e-01
-5.09219170e-01 3.87778342e-01 1.44995213e-01 -4.77179259e-01
-4.55773622e-01 -6.22533262e-01 -6.14320159e-01 -6.37980223e-01
-5.88472009e-01 5.45894206e-01 5.90634704e-01 1.02401936e+00
4.00544316e-01 1.27823591e-01 1.20967579e+00 -5.48118889e-01
-9.40765500e-01 -6.26970112e-01 -1.01072454e+00 1.35194868e-01
3.70102704e-01 -4.77779120e-01 -4.19447333e-01 -5.33792436e-01] | [4.565964221954346, 3.350005626678467] |
f798f114-497f-45c8-9635-58be42c6e5c0 | entity-type-prediction-leveraging-graph-walks | 2207.14094 | null | https://arxiv.org/abs/2207.14094v2 | https://arxiv.org/pdf/2207.14094v2.pdf | Entity Type Prediction Leveraging Graph Walks and Entity Descriptions | The entity type information in Knowledge Graphs (KGs) such as DBpedia, Freebase, etc. is often incomplete due to automated generation or human curation. Entity typing is the task of assigning or inferring the semantic type of an entity in a KG. This paper presents \textit{GRAND}, a novel approach for entity typing leveraging different graph walk strategies in RDF2vec together with textual entity descriptions. RDF2vec first generates graph walks and then uses a language model to obtain embeddings for each node in the graph. This study shows that the walk generation strategy and the embedding model have a significant effect on the performance of the entity typing task. The proposed approach outperforms the baseline approaches on the benchmark datasets DBpedia and FIGER for entity typing in KGs for both fine-grained and coarse-grained classes. The results show that the combination of order-aware RDF2vec variants together with the contextual embeddings of the textual entity descriptions achieve the best results. | ['Mehwish Alam', 'Harald Sack', 'Heiko Paulheim', 'Jan Portisch', 'Russa Biswas'] | 2022-07-28 | null | null | null | null | ['type-prediction', 'entity-typing'] | ['computer-code', 'natural-language-processing'] | [-5.71636379e-01 4.46493477e-01 -2.23540097e-01 -4.84343320e-01
-1.18044987e-01 -6.61867619e-01 6.99182391e-01 8.20117652e-01
-6.71147645e-01 8.80244315e-01 4.51382577e-01 -1.44215003e-01
-2.02452421e-01 -1.61129713e+00 -8.70139837e-01 -2.99312472e-01
-4.02313471e-01 7.67821312e-01 4.59243625e-01 -5.69970369e-01
-1.98752284e-01 6.28049150e-02 -1.63683224e+00 2.05604315e-01
1.01844645e+00 9.48896766e-01 1.95719808e-01 4.60552067e-01
-8.42293918e-01 8.17659020e-01 -4.14027274e-01 -9.77115214e-01
1.13295615e-02 4.07928199e-01 -9.68077362e-01 -7.73800015e-01
5.14901042e-01 7.21959844e-02 -4.41875190e-01 9.97825563e-01
7.42842674e-01 1.04320586e-01 7.47925103e-01 -1.33592081e+00
-9.68380570e-01 9.62487757e-01 -2.67370522e-01 -7.92463496e-02
6.07431829e-01 -4.16456372e-01 1.43665349e+00 -9.28020537e-01
1.33497918e+00 1.12554419e+00 9.77423608e-01 3.03020328e-01
-9.54087973e-01 -2.50002742e-01 -7.50889676e-03 6.91734552e-01
-1.65133929e+00 1.31946594e-01 3.30029726e-01 -6.53419018e-01
1.31240320e+00 1.14567891e-01 3.82339895e-01 9.27631438e-01
-3.31802964e-02 3.31218213e-01 6.85415149e-01 -4.42853034e-01
1.31678075e-01 5.01012683e-01 7.85324872e-01 8.48538458e-01
9.89143074e-01 -3.50936502e-01 -2.90647805e-01 -3.29658210e-01
2.25127131e-01 -3.22183341e-01 -1.70628101e-01 -4.42549825e-01
-9.08034205e-01 8.26269686e-01 6.79591537e-01 1.78515419e-01
-5.94311893e-01 2.87882864e-01 7.86464274e-01 6.95857778e-02
3.59487563e-01 6.08377337e-01 -6.27753437e-01 2.31155157e-01
-3.90266597e-01 6.13646567e-01 1.24094033e+00 1.47783625e+00
9.95378435e-01 -5.88648356e-02 -4.53344315e-01 9.58163023e-01
3.98861200e-01 5.39532959e-01 3.72846685e-02 -2.18781650e-01
7.15425789e-01 9.98454094e-01 3.94869208e-01 -1.27287710e+00
-4.41220492e-01 -1.81226254e-01 -4.45146620e-01 -1.23313442e-01
2.35874832e-01 -3.42909306e-01 -8.72862101e-01 1.50636804e+00
8.87698412e-01 4.25887406e-02 2.85238236e-01 7.00039327e-01
1.55256355e+00 4.30848390e-01 5.27404070e-01 3.87489855e-01
1.70808518e+00 -7.36492455e-01 -9.86130595e-01 2.13807985e-01
7.19471157e-01 -3.44578743e-01 7.10793853e-01 -2.30861247e-01
-4.27335829e-01 -3.34819049e-01 -1.16810441e+00 -4.38771963e-01
-1.42422605e+00 -3.82665433e-02 7.54927814e-01 5.31596124e-01
-7.63225079e-01 6.59686148e-01 -5.48079610e-01 -4.65279460e-01
3.27165455e-01 1.08398803e-01 -7.33039021e-01 -3.22153687e-01
-1.79482472e+00 9.13072348e-01 9.32483554e-01 -1.95714701e-02
-2.86408275e-01 -9.05657113e-01 -1.25350761e+00 6.38699392e-03
5.16212821e-01 -1.07120085e+00 5.10864198e-01 -1.44336492e-01
-3.29904914e-01 5.42616069e-01 -7.90384933e-02 -6.47933722e-01
3.11122924e-01 -1.69323608e-01 -8.13024819e-01 -3.77781034e-01
3.47224832e-01 2.15648726e-01 3.59895110e-01 -1.20925212e+00
-8.25757086e-01 -5.96632123e-01 4.57756191e-01 5.93053177e-02
-3.29391986e-01 -3.79685372e-01 -5.86588085e-01 -3.82804900e-01
-4.56100643e-01 -7.42277265e-01 -5.49589396e-02 -6.52498484e-01
-7.83085823e-01 -7.67484903e-01 4.89523768e-01 -9.20512974e-01
1.76846218e+00 -1.74838209e+00 2.09898666e-01 2.92300642e-01
4.14221019e-01 1.42950520e-01 1.82035759e-01 9.50466096e-01
1.15067244e-01 4.38636541e-01 2.35609099e-01 9.81026590e-02
5.66024125e-01 4.31124032e-01 1.19650446e-01 1.72384921e-02
-2.06990600e-01 1.00784075e+00 -1.08739173e+00 -6.15973532e-01
-3.28996867e-01 6.34183884e-01 -4.90505725e-01 1.37739271e-01
-5.66489398e-01 -4.95668977e-01 -5.11496782e-01 3.76876742e-01
8.23996425e-01 -1.70172840e-01 7.42223263e-01 -8.04975390e-01
-6.36761412e-02 2.29304343e-01 -1.46667385e+00 1.50628471e+00
-5.17451227e-01 2.16233075e-01 -5.27268052e-01 -3.90364826e-01
6.26222491e-01 1.85000643e-01 2.36065045e-01 -5.03471494e-01
-3.02121222e-01 1.86437517e-01 -3.74295175e-01 -8.65713716e-01
1.11687541e+00 3.25658232e-01 -4.27023798e-01 6.41584098e-02
3.95190924e-01 6.50529206e-01 6.26102805e-01 6.66108847e-01
1.10880494e+00 2.88394302e-01 2.41572365e-01 -3.43349963e-01
3.15224171e-01 1.29876107e-01 6.61267817e-01 6.54690802e-01
6.69595420e-01 -9.97234229e-03 5.83505511e-01 -5.54718554e-01
-1.00668418e+00 -9.64379072e-01 -1.54834032e-01 1.12153101e+00
2.13331565e-01 -9.12381411e-01 -4.51048642e-01 -1.16213810e+00
7.24037886e-01 7.83313274e-01 -9.58757579e-01 2.18721062e-01
-3.15324605e-01 -7.51929104e-01 6.84555829e-01 7.98247337e-01
2.42467895e-01 -9.26967263e-01 3.94914374e-02 3.46807450e-01
-2.22307131e-01 -1.38021016e+00 -6.82684556e-02 -1.69228986e-02
-1.32648781e-01 -1.39524674e+00 -2.27735966e-01 -9.32988822e-01
6.51333869e-01 -4.62180346e-01 1.25154710e+00 3.48411947e-02
-2.95294821e-01 1.96535289e-01 -7.84661412e-01 -4.10110772e-01
-3.30278240e-02 4.28160846e-01 -1.77074969e-02 -7.79229328e-02
5.94097376e-01 -3.57012540e-01 -3.09727639e-01 -1.91690162e-01
-9.00785506e-01 -2.83371150e-01 1.26189783e-01 7.71166623e-01
5.75311005e-01 2.71263629e-01 5.12325883e-01 -1.93984520e+00
6.47004426e-01 -8.55941176e-01 -4.33308721e-01 7.17221320e-01
-1.04453039e+00 5.51097691e-01 4.73883569e-01 1.38111323e-01
-1.03129363e+00 -3.82002056e-01 -1.79982692e-01 -2.60983445e-02
7.48179331e-02 1.04903948e+00 -3.49928856e-01 4.10426930e-02
6.66596830e-01 -2.03223135e-02 -7.60025203e-01 -7.23259151e-01
7.17239082e-01 4.22844946e-01 1.65804893e-01 -6.86605215e-01
8.70142460e-01 1.79864485e-02 -4.87439521e-02 -7.39333510e-01
-4.31003988e-01 -4.99114215e-01 -6.93256080e-01 2.12797076e-02
1.03152215e+00 -1.00601196e+00 -4.47778106e-01 3.07690471e-01
-1.16103327e+00 3.02401814e-03 -2.15045556e-01 2.52472490e-01
6.72012642e-02 2.81294167e-01 -5.32880485e-01 -3.61651957e-01
-3.72642934e-01 -6.34764493e-01 8.57641220e-01 9.57026407e-02
1.39752403e-01 -1.36247885e+00 2.74347365e-01 1.62187263e-01
3.04911613e-01 6.87167704e-01 1.45300984e+00 -1.09019196e+00
-5.34782171e-01 -2.20508575e-01 -4.19216067e-01 -2.73383800e-02
5.74041232e-02 -6.74969852e-02 -7.01565504e-01 1.35036841e-01
-1.32657540e+00 1.02294862e-01 7.72796273e-01 -1.89902753e-01
5.59966683e-01 -6.53354049e-01 -4.71250296e-01 6.98312938e-01
2.22501206e+00 -1.40291497e-01 8.10198247e-01 5.04860282e-01
1.38467741e+00 4.31773484e-01 7.22921669e-01 5.26469469e-01
1.01702738e+00 7.91810989e-01 3.22265863e-01 2.96665817e-01
-1.95868790e-01 -6.66353881e-01 -1.43459797e-01 7.33096778e-01
-4.10065055e-01 -6.07856512e-01 -1.02033103e+00 7.11111665e-01
-1.83603239e+00 -1.05815828e+00 -5.69758534e-01 2.17043233e+00
7.99396098e-01 -3.56823981e-01 -9.17800665e-02 -8.14062729e-02
6.75232887e-01 2.28248149e-01 -1.52488306e-01 -5.08117855e-01
-1.42533064e-01 1.99974507e-01 1.06018937e+00 7.13866055e-01
-1.01501000e+00 1.01240098e+00 4.89772129e+00 7.07537353e-01
-5.05964518e-01 2.06569433e-01 -3.56121182e-01 4.86824006e-01
-8.77041280e-01 -1.03831552e-01 -1.31613314e+00 6.49484038e-01
6.69431925e-01 -5.55494130e-01 2.59626895e-01 7.86296129e-01
-4.05272514e-01 2.03756407e-01 -9.69265878e-01 4.81022298e-01
-4.93390635e-02 -1.52977896e+00 1.94813892e-01 1.09048188e-01
6.14009738e-01 5.38105629e-02 -4.32056546e-01 7.23242223e-01
7.32580781e-01 -8.63328636e-01 4.68379945e-01 6.86510861e-01
1.01042116e+00 -7.04465747e-01 1.11357808e+00 -2.11802632e-01
-1.51612282e+00 2.08571315e-01 -4.57778543e-01 6.71373606e-01
1.35933729e-02 5.98582089e-01 -1.22962594e+00 1.26973963e+00
6.16416991e-01 5.69389224e-01 -7.65242219e-01 8.11612844e-01
-4.08801734e-01 2.48950556e-01 -5.97434081e-02 -2.59402126e-01
1.44689232e-02 -2.24110335e-02 5.19454479e-01 1.61916900e+00
1.37944058e-01 -1.84667438e-01 1.34186253e-01 5.88540256e-01
-4.47763622e-01 4.57605571e-01 -7.84237325e-01 -2.20052347e-01
8.05734515e-01 1.29412436e+00 -1.12393193e-01 -5.12028635e-01
-5.31410277e-01 7.02066362e-01 7.06074119e-01 6.28894210e-01
-8.93989563e-01 -9.99947727e-01 7.19269097e-01 4.32831049e-01
8.09218764e-01 -8.31338018e-02 1.16884388e-01 -1.12750566e+00
1.44557238e-01 -3.82663161e-01 8.83761764e-01 -6.16676927e-01
-1.26759064e+00 6.41721487e-01 2.57095620e-02 -7.06550837e-01
-3.00314963e-01 -3.79002541e-01 -1.92681327e-01 8.37328494e-01
-1.67333400e+00 -1.43243372e+00 -5.83442748e-01 5.16966462e-01
-2.18052436e-02 -4.79339436e-02 9.88004506e-01 8.29012036e-01
-4.79378134e-01 9.49173152e-01 1.66668549e-01 6.74581528e-01
6.12101257e-01 -1.95606089e+00 4.70427513e-01 4.99360442e-01
9.33926478e-02 8.14158142e-01 7.99276888e-01 -9.74424362e-01
-1.70969057e+00 -1.53389466e+00 1.48615956e+00 -3.72948855e-01
7.04676628e-01 -5.05995452e-01 -9.17984247e-01 9.44585025e-01
3.23039651e-01 2.64728010e-01 7.28738666e-01 7.29393959e-01
-9.24620867e-01 -2.38333151e-01 -1.28831351e+00 2.82349378e-01
1.21180618e+00 -4.80113864e-01 -6.30476594e-01 3.49517316e-01
8.45073998e-01 -3.44332546e-01 -1.61984622e+00 3.82185042e-01
6.44570172e-01 -4.31985050e-01 9.58797038e-01 -1.06144333e+00
2.91109651e-01 -4.36533988e-01 -3.33443642e-01 -1.58228290e+00
-5.33939779e-01 1.13289505e-01 -4.43424135e-01 1.51215398e+00
9.72808778e-01 -7.33537257e-01 5.82890749e-01 5.87315500e-01
6.15265779e-02 -5.74175656e-01 -5.28860629e-01 -8.41518998e-01
-2.41248012e-01 -5.66671528e-02 1.12407565e+00 1.26510489e+00
-1.25354556e-02 4.17796582e-01 -1.38283119e-01 5.60088634e-01
6.76098883e-01 2.30036318e-01 7.68895984e-01 -1.29807723e+00
-1.25662580e-01 1.09066695e-01 -1.11295485e+00 -3.19696814e-01
1.93766430e-01 -1.27634263e+00 -4.29075748e-01 -2.03525138e+00
-9.24854651e-02 -8.44667614e-01 -3.31587344e-01 4.32540506e-01
-3.86508524e-01 -1.24834254e-01 -1.15861990e-01 -3.86626273e-01
-6.27355337e-01 1.94726542e-01 7.21057475e-01 -3.93653452e-01
1.73861429e-01 -5.70920050e-01 -5.50961852e-01 3.16107303e-01
4.70143914e-01 -4.33197945e-01 -4.22595322e-01 -7.73557246e-01
9.14913297e-01 -1.80924423e-02 2.07300425e-01 -6.12404346e-01
4.61445868e-01 6.81872964e-02 1.27211869e-01 -3.11836600e-01
6.76160529e-02 -7.92984307e-01 4.36893642e-01 -6.48842528e-02
-1.39293760e-01 9.09947008e-02 -3.09482515e-02 1.00006795e+00
-2.89150119e-01 -2.62215674e-01 1.50682837e-01 -8.82751420e-02
-1.41191328e+00 6.59950197e-01 2.49269009e-01 5.77862620e-01
8.55331719e-01 5.52360713e-02 -5.20560265e-01 1.66277379e-01
-1.00746834e+00 3.13125581e-01 4.61003423e-01 6.55431569e-01
3.57633650e-01 -1.70431197e+00 -7.14122832e-01 -3.94939631e-03
5.86884558e-01 1.73067804e-02 3.09742868e-01 4.61784393e-01
-6.63718879e-01 2.80419141e-01 -1.48104250e-01 -2.96071507e-02
-1.33054435e+00 5.03694773e-01 4.73256297e-02 -4.78143930e-01
-4.44623083e-01 8.91595840e-01 -2.27404550e-01 -5.78755915e-01
1.78506777e-01 -1.36162519e-01 -7.99654901e-01 6.14883602e-01
3.30831170e-01 5.87549031e-01 3.30753356e-01 -7.20637262e-01
-4.77569044e-01 3.22781354e-01 -2.21605182e-01 3.31283808e-01
1.49932325e+00 1.34623617e-01 -1.60919964e-01 3.76359344e-01
1.28176963e+00 3.55347902e-01 -2.27864027e-01 -2.16155797e-01
3.66423815e-01 -3.91627997e-01 -2.03012198e-01 -8.08505297e-01
-1.07805121e+00 3.68054479e-01 2.30710179e-01 3.61698925e-01
3.94892365e-01 -3.79242785e-02 8.42215776e-01 3.99992675e-01
8.44562054e-01 -1.23193634e+00 -8.89946342e-01 5.08590102e-01
6.49278104e-01 -9.99393523e-01 8.57012793e-02 -5.77970266e-01
-8.19762409e-01 8.53450894e-01 5.05721211e-01 -1.71441548e-02
8.42260420e-01 1.75749421e-01 -1.86433986e-01 -6.22815311e-01
-7.28686571e-01 -4.73914057e-01 3.97362828e-01 9.15433764e-01
4.61644322e-01 4.50279385e-01 -5.54140747e-01 8.38278234e-01
-1.60796613e-01 -3.70925739e-02 3.51751775e-01 7.68200338e-01
-3.00455898e-01 -1.29076946e+00 3.25424284e-01 6.32488072e-01
-4.35263276e-01 -3.32042664e-01 -2.03781858e-01 9.79364574e-01
5.63845098e-01 5.26966393e-01 -1.34927601e-01 -6.34211123e-01
6.51335835e-01 2.79309928e-01 4.21523362e-01 -7.34438956e-01
-6.34596646e-01 -8.50583136e-01 9.74876404e-01 -5.17553687e-01
-1.09198198e-01 -4.46920305e-01 -1.25971639e+00 -5.04027903e-01
-4.19960976e-01 4.69063610e-01 6.34425282e-01 4.20932889e-01
6.64746344e-01 5.01252830e-01 2.87209362e-01 5.45841269e-02
-1.39698163e-01 -9.00401473e-01 -1.00865352e+00 9.04168963e-01
-8.15941393e-02 -8.85075748e-01 -1.43729690e-02 -1.52252242e-01] | [9.103018760681152, 8.127575874328613] |
2a2d6d16-b53b-44a8-9cc3-5a15bbb7da96 | principal-component-classification | 2210.12746 | null | https://arxiv.org/abs/2210.12746v2 | https://arxiv.org/pdf/2210.12746v2.pdf | Principal Component Classification | We propose to directly compute classification estimates by learning features encoded with their class scores using PCA. Our resulting model has a encoder-decoder structure suitable for supervised learning, it is computationally efficient and performs well for classification on several datasets. | ['Rozenn Dahyot'] | 2022-10-23 | null | null | null | null | ['component-classification'] | ['natural-language-processing'] | [ 4.60674852e-01 1.34135932e-01 -7.63161421e-01 -1.05721414e+00
-1.41808152e+00 -3.50175947e-01 8.28986883e-01 1.66832402e-01
-4.10951912e-01 8.10131371e-01 2.80003279e-01 -5.62691949e-02
-7.63429701e-02 -5.84005892e-01 -5.97683132e-01 -6.28756762e-01
-2.75308639e-01 6.75104439e-01 -7.40156099e-02 5.16240358e-01
3.63644511e-01 2.29186043e-01 -1.70159459e+00 6.43649817e-01
3.84473771e-01 1.30144489e+00 1.93153061e-02 8.66111338e-01
1.19295940e-01 1.20050967e+00 -4.27342713e-01 -7.42327332e-01
6.36880621e-02 -3.99454027e-01 -5.93718290e-01 2.60295451e-01
3.13811272e-01 -1.03168301e-01 -1.39067486e-01 8.50191653e-01
1.27948090e-01 -2.63705611e-01 1.26442993e+00 -7.92508125e-01
-7.65565574e-01 2.31372029e-01 -1.61956340e-01 1.39690667e-01
6.80240631e-01 -3.83213133e-01 1.42739499e+00 -1.27608681e+00
4.39265847e-01 1.03859210e+00 9.39146817e-01 3.59450072e-01
-1.50295663e+00 -2.05805227e-01 -3.35797876e-01 2.27166742e-01
-1.39972866e+00 -7.94227779e-01 5.57593107e-01 -4.77064729e-01
1.14498079e+00 3.22561115e-01 5.79961956e-01 1.00831521e+00
4.60092783e-01 1.05753648e+00 1.22542155e+00 -3.50601077e-01
3.72251391e-01 3.15661848e-01 1.12552583e-01 9.80194569e-01
-3.20279673e-02 3.11853439e-01 -7.51371324e-01 -3.87951463e-01
4.99126107e-01 -8.67316872e-02 1.16371132e-01 -4.59189147e-01
-1.12906206e+00 1.22337651e+00 1.36056662e-01 -2.65616447e-01
-5.06222069e-01 4.90956008e-01 4.29187208e-01 4.06966507e-01
7.12456167e-01 1.73503891e-01 -9.18915212e-01 -3.60796988e-01
-6.59334302e-01 2.56049544e-01 9.01665926e-01 9.83971477e-01
8.02533805e-01 1.10006407e-01 -1.70395244e-02 1.00212681e+00
3.07907104e-01 5.27777553e-01 8.33244503e-01 -9.78017986e-01
2.22249418e-01 9.35038701e-02 -1.89266250e-01 -5.66792905e-01
-5.33425212e-01 -1.72521383e-01 -4.80577916e-01 -3.41733918e-02
2.71690432e-02 -9.26903486e-02 -5.81249714e-01 1.29949760e+00
5.05750347e-03 3.30258548e-01 2.83343405e-01 4.39426333e-01
6.81740642e-01 6.72671556e-01 1.00797825e-01 -3.53732049e-01
1.10178542e+00 -7.41525590e-01 -8.34774256e-01 -3.13398957e-01
6.41199410e-01 -7.14286566e-01 6.28843904e-01 7.34767199e-01
-9.02244747e-01 -4.14683729e-01 -9.44218278e-01 -9.99513790e-02
-3.11981738e-01 7.89412796e-01 1.39109266e+00 1.03407109e+00
-9.51973796e-01 6.75761759e-01 -1.16588342e+00 5.65515682e-02
5.70204854e-01 6.86134160e-01 -5.99485576e-01 4.49291579e-02
-5.09564281e-01 9.39682245e-01 6.26019597e-01 -1.19779944e-01
-7.77574480e-01 -1.74549833e-01 -1.22315860e+00 -1.10558227e-01
-3.41739178e-01 -3.97171319e-01 1.42904544e+00 -9.34722066e-01
-1.92858589e+00 6.37267888e-01 -3.37083042e-01 -4.96129006e-01
-5.00521697e-02 -7.00179935e-02 -4.08422709e-01 2.81031430e-02
5.33589162e-02 6.34941101e-01 1.04478562e+00 -7.18245566e-01
-6.80876374e-01 -3.35227937e-01 -5.14958620e-01 9.01743174e-02
-3.39632779e-01 3.04457337e-01 -9.12036840e-03 -7.26715386e-01
2.53052175e-01 -7.56876707e-01 -2.05388010e-01 4.21457551e-02
-1.64951384e-02 -7.98731446e-02 4.35208797e-01 -6.45499051e-01
9.47501838e-01 -2.00637031e+00 5.28690256e-02 1.69978082e-01
-4.85854670e-02 -2.73361616e-02 -8.29520747e-02 3.46585423e-01
4.88025695e-02 -3.25279027e-01 -3.92672926e-01 -3.76861125e-01
-6.35933131e-02 5.95062256e-01 -2.64262974e-01 7.85372615e-01
6.72993660e-01 8.65155160e-01 -8.25686514e-01 -3.84977907e-01
3.19635481e-01 5.74271441e-01 -9.29781497e-01 4.48506653e-01
1.25710085e-01 -1.22603334e-01 -4.75325495e-01 7.68542647e-01
2.74814487e-01 -2.09273398e-01 2.35871151e-01 1.02286540e-01
2.61295378e-01 6.58812106e-01 -8.44131291e-01 1.65017760e+00
-5.34662306e-01 1.03164351e+00 -1.11864686e-01 -1.16683924e+00
9.99922335e-01 3.54673594e-01 1.85420498e-01 1.66387279e-02
6.49571195e-02 2.27487415e-01 -3.31573308e-01 -5.64714730e-01
-7.34579936e-02 -2.72630811e-01 -2.42338404e-01 2.05925286e-01
7.60950029e-01 -3.54421377e-01 -9.47335735e-02 -3.80144656e-01
1.14472771e+00 2.55011339e-02 7.46439993e-01 -1.04865402e-01
5.54313958e-01 -2.92818040e-01 3.02210271e-01 2.45501176e-01
2.59906594e-02 6.22025132e-01 4.59662080e-01 -7.26181209e-01
-6.24053836e-01 -1.23724091e+00 -6.00263357e-01 1.10181618e+00
-5.51138699e-01 -6.93846047e-01 -4.29840058e-01 -6.90693557e-01
1.26587495e-01 5.26223183e-01 -5.84039867e-01 -2.08924249e-01
-7.93007761e-02 -1.04742301e+00 4.40097779e-01 8.98656964e-01
-5.21227606e-02 -7.11909592e-01 -3.06914687e-01 2.37054139e-01
1.26873881e-01 -9.83440161e-01 1.14631347e-01 7.84242511e-01
-1.15957522e+00 -1.14577889e+00 -3.38468224e-01 -1.05792880e+00
6.46112561e-01 -1.74352407e-01 9.27720070e-01 -3.85180622e-01
-1.68022230e-01 1.97690561e-01 -2.35577583e-01 -3.76591295e-01
-2.64814407e-01 1.10074125e-01 1.15450889e-01 8.04884732e-03
7.73996413e-01 -4.08587366e-01 -5.37430160e-02 1.75787598e-01
-5.41153669e-01 -2.84454435e-01 8.51982594e-01 1.26782489e+00
6.29561901e-01 -4.83420417e-02 5.55953264e-01 -1.25392139e+00
8.08152318e-01 -4.33046550e-01 -4.41175908e-01 7.06722867e-03
-6.29951537e-01 1.94999844e-01 7.19590783e-01 -1.76805541e-01
-8.81778479e-01 7.53832877e-01 -6.70611501e-01 -4.40195948e-03
-2.29094729e-01 4.31591600e-01 1.47790611e-01 -2.15445116e-01
7.28911459e-01 3.66564721e-01 -3.29141319e-02 -6.51240766e-01
3.87985528e-01 9.93701994e-01 3.15182894e-01 -2.36926183e-01
3.83836865e-01 1.92694768e-01 -1.39123917e-01 -6.91415191e-01
-6.89515054e-01 -1.42185584e-01 -7.90274918e-01 1.76794559e-01
6.42354012e-01 -1.25775921e+00 -6.22255623e-01 4.27979752e-02
-1.19126570e+00 5.69424927e-02 -8.55677351e-02 9.75476205e-01
-8.85070562e-01 -5.14819883e-02 -8.82291079e-01 -8.34373653e-01
-2.91185826e-01 -9.39086735e-01 1.37445927e+00 -1.79728568e-01
-2.69203275e-01 -1.23270833e+00 1.56235710e-01 1.70257278e-02
8.49653035e-02 -1.28069744e-01 8.02210689e-01 -7.26248980e-01
-2.49325246e-01 -7.69475341e-01 -1.46032453e-01 3.94827664e-01
2.29673758e-01 -1.36456162e-01 -1.26127481e+00 -2.02340513e-01
-1.81131929e-01 -5.46773791e-01 9.99208033e-01 3.20742995e-01
1.62483251e+00 -4.25152749e-01 -3.52321476e-01 5.58291256e-01
1.35564840e+00 -7.17773959e-02 2.31638685e-01 -1.83738917e-01
3.45131963e-01 2.86450237e-01 2.98394561e-01 4.24429506e-01
4.56576735e-01 4.63204443e-01 -8.04094374e-02 3.85702014e-01
-3.57448123e-02 -2.50193894e-01 6.15379989e-01 1.26252508e+00
2.10012905e-02 2.47185275e-01 -7.94116080e-01 3.70456278e-01
-1.79309976e+00 -7.85744369e-01 -9.58920270e-02 1.85618758e+00
8.96878481e-01 1.28308281e-01 -3.44735272e-02 3.54893178e-01
2.77052522e-01 1.17185339e-01 -2.14477647e-02 -5.83471477e-01
1.64759368e-01 5.72941124e-01 4.41636741e-01 4.78800297e-01
-1.24464393e+00 8.70179355e-01 9.75615025e+00 7.14403987e-01
-5.95437527e-01 2.62192428e-01 5.47797143e-01 2.90913275e-03
-2.67685186e-02 -2.09174559e-01 -6.41596138e-01 4.22159404e-01
1.43241656e+00 1.96599737e-01 3.28332812e-01 1.02565932e+00
-2.80270785e-01 7.49006495e-02 -1.38046086e+00 1.10670662e+00
2.75196463e-01 -1.21596098e+00 -8.01384002e-02 -3.67389023e-01
4.72074866e-01 1.55068710e-01 2.08791435e-01 4.22911465e-01
3.19504976e-01 -7.75466204e-01 4.11816627e-01 2.00716168e-01
1.01708782e+00 -6.94998324e-01 8.19734395e-01 2.52253383e-01
-9.18818831e-01 -2.00472429e-01 -8.25218379e-01 -4.86192226e-01
-2.19792828e-01 5.38014531e-01 -1.15791845e+00 1.26543716e-01
1.21608838e-01 1.15297925e+00 -8.21655512e-01 1.02076375e+00
-4.75089252e-01 8.46188486e-01 -1.07169181e-01 -3.07330221e-01
-3.14842984e-02 -1.30331907e-02 -1.13289028e-01 1.42927969e+00
3.37899238e-01 -1.01716317e-01 2.32721731e-01 1.84579805e-01
1.06853567e-01 1.07297234e-01 -8.40360761e-01 -8.35007876e-02
1.77521020e-01 9.75605667e-01 -7.49131858e-01 -4.86623019e-01
-6.08279526e-01 9.08551157e-01 7.49347627e-01 4.60248701e-02
-6.00457788e-01 -6.10642195e-01 6.41579390e-01 -3.98161590e-01
4.30154383e-01 -2.02006847e-01 -3.12156707e-01 -1.44682431e+00
1.74139408e-04 -4.78934497e-01 2.99458832e-01 -6.88700795e-01
-1.31268549e+00 7.95236468e-01 -3.00768673e-01 -1.27434993e+00
-7.47334898e-01 -1.23031318e+00 -1.73530519e-01 4.82463926e-01
-1.56743634e+00 -8.21761012e-01 2.36883372e-01 3.88949931e-01
6.00835085e-01 -5.78086138e-01 1.39072466e+00 2.79379487e-01
-4.09229040e-01 5.81855714e-01 4.52659518e-01 2.11845741e-01
4.40821230e-01 -1.36437762e+00 3.25494915e-01 3.89483631e-01
8.05766225e-01 3.78456354e-01 3.42322558e-01 -2.50598103e-01
-1.62470973e+00 -9.04797196e-01 1.26628006e+00 -7.23639727e-01
5.71995437e-01 -5.03315926e-01 -5.06912827e-01 7.84896016e-01
-4.47374955e-03 2.72642374e-01 1.35609496e+00 4.51521128e-01
-5.55865288e-01 -2.37552375e-01 -1.07416189e+00 -1.92394838e-01
6.23764336e-01 -1.02914572e+00 -7.17224300e-01 6.71289682e-01
2.58841068e-01 -4.75535214e-01 -8.76320481e-01 2.33458616e-02
7.70443797e-01 -4.93594468e-01 9.49991643e-01 -7.50851989e-01
6.33215010e-01 -3.13336737e-02 -3.52185339e-01 -1.40097034e+00
-3.82167011e-01 -2.57186472e-01 -5.75088501e-01 6.87884867e-01
8.08805764e-01 -6.57712102e-01 8.43073487e-01 6.52882218e-01
1.14901036e-01 -7.11374164e-01 -1.21421218e+00 -4.91587818e-01
-2.48848811e-01 -8.01813066e-01 3.66058320e-01 7.23192751e-01
3.75526905e-01 4.31818932e-01 -4.14790452e-01 8.13608170e-02
4.27185178e-01 1.04321420e-01 2.75717527e-01 -1.29639578e+00
-5.42266130e-01 -3.30203891e-01 -9.38073158e-01 -1.02986813e+00
5.62705576e-01 -1.21106267e+00 4.24890667e-02 -1.06576180e+00
2.64589220e-01 -2.29678139e-01 -2.56616741e-01 7.87110209e-01
1.66474283e-02 6.93640292e-01 -1.44426480e-01 1.16847396e-01
-7.25356758e-01 5.31316578e-01 5.00939310e-01 -1.19020149e-01
3.39144468e-01 2.67368644e-01 -4.58778620e-01 7.14528620e-01
5.90825379e-01 -7.36298740e-01 -9.93508697e-02 -4.33770776e-01
1.88329086e-01 2.67513931e-01 1.47797475e-02 -1.00377703e+00
-1.76109880e-01 1.30227938e-01 9.10380661e-01 -4.32788759e-01
6.34424329e-01 -8.70189369e-01 -1.00400902e-01 4.27546889e-01
-7.84031153e-01 7.65388235e-02 -2.04591587e-01 8.49176109e-01
-5.54166079e-01 -3.74709725e-01 4.40083593e-01 1.44839272e-01
-6.94723189e-01 -5.66529110e-02 -9.60716546e-01 -5.13087809e-01
9.28142726e-01 8.25858787e-02 2.72598058e-01 -3.12389016e-01
-6.99619353e-01 -1.73035949e-01 -7.19034160e-03 2.58016646e-01
8.04461241e-01 -1.57113326e+00 -6.75504386e-01 6.85532570e-01
2.12645248e-01 -5.73350549e-01 -6.87535286e-01 2.98427105e-01
-5.85764945e-01 7.29513526e-01 -1.66220963e-01 -5.83075762e-01
-1.09734881e+00 4.73247379e-01 -6.78887963e-02 1.10143334e-01
-2.09037155e-01 9.44595337e-01 -4.05872047e-01 -5.55839956e-01
9.76331905e-02 -3.33059877e-01 -3.34467083e-01 7.05941617e-02
8.40187728e-01 3.76793116e-01 2.60265380e-01 -7.04106927e-01
-3.53535235e-01 3.94392133e-01 1.10972099e-01 -1.75293616e-03
1.42276549e+00 9.24899057e-02 -9.14905518e-02 5.21153688e-01
1.85885262e+00 -3.93974721e-01 -1.14131355e+00 -2.86242783e-01
2.84252644e-01 -6.87739730e-01 2.70897567e-01 -5.90956569e-01
-9.11648929e-01 8.67174029e-01 5.38560450e-01 -7.51385614e-02
1.16399384e+00 -2.79808659e-02 3.71643603e-01 8.30736876e-01
5.59827745e-01 -9.23503816e-01 -2.02744544e-01 5.46758235e-01
3.76273543e-01 -1.34048581e+00 3.27087402e-01 -6.53320611e-01
-6.29307806e-01 1.48002458e+00 -1.32863954e-01 -4.91682917e-01
9.09841001e-01 3.84139776e-01 2.03380059e-03 -3.26149277e-02
-1.27456164e+00 -2.61525929e-01 6.23980820e-01 9.38328743e-01
8.44131052e-01 3.67034763e-01 -4.58010733e-01 6.71812534e-01
-3.11495602e-01 -2.33304556e-02 4.77240920e-01 7.64209032e-01
-3.70000690e-01 -1.23445320e+00 -7.16694631e-03 9.32919621e-01
-6.26334190e-01 -1.28226485e-02 -2.57106066e-01 1.59243986e-01
-7.62139559e-02 7.50620723e-01 1.99894920e-01 -6.17574811e-01
-1.94483683e-01 4.42960531e-01 5.37997961e-01 -7.82433033e-01
-4.10501689e-01 3.38950716e-02 4.31403548e-01 -4.95370299e-01
-6.19366229e-01 -9.96006012e-01 -6.88117027e-01 1.66124851e-01
-4.18994039e-01 1.54846951e-01 8.07612717e-01 8.53280604e-01
2.60082543e-01 2.50301838e-01 9.37383592e-01 -9.36485469e-01
-6.69280529e-01 -8.24487090e-01 -5.92441142e-01 1.13909297e-01
3.58397901e-01 -8.07078660e-01 -1.70000479e-01 2.23584950e-01] | [9.425410270690918, 2.8440749645233154] |
74c9da47-fefd-4fda-8ff7-1be80300055a | mulda-a-multilingual-data-augmentation | null | null | https://aclanthology.org/2021.acl-long.453 | https://aclanthology.org/2021.acl-long.453.pdf | MulDA: A Multilingual Data Augmentation Framework for Low-Resource Cross-Lingual NER | Named Entity Recognition (NER) for low-resource languages is a both practical and challenging research problem. This paper addresses zero-shot transfer for cross-lingual NER, especially when the amount of source-language training data is also limited. The paper first proposes a simple but effective labeled sequence translation method to translate source-language training data to target languages and avoids problems such as word order change and entity span determination. With the source-language data as well as the translated data, a generation-based multilingual data augmentation method is introduced to further increase diversity by generating synthetic labeled data in multiple languages. These augmented data enable the language model based NER models to generalize better with both the language-specific features from the target-language synthetic data and the language-independent features from multilingual synthetic data. An extensive set of experiments were conducted to demonstrate encouraging cross-lingual transfer performance of the new research on a wide variety of target languages. | ['Chunyan Miao', 'Luo Si', 'Shafiq Joty', 'Lidong Bing', 'Bosheng Ding', 'Linlin Liu'] | 2021-08-01 | null | null | null | acl-2021-5 | ['cross-lingual-ner'] | ['natural-language-processing'] | [-1.75652251e-01 -3.25902194e-01 -4.22032565e-01 -5.83271921e-01
-1.36816919e+00 -7.43343592e-01 4.97155637e-01 -3.62298876e-01
-9.25741553e-01 1.42238665e+00 2.94169992e-01 -2.76198298e-01
6.57650113e-01 -6.41628385e-01 -6.18760586e-01 -2.15312317e-01
3.20703447e-01 6.70655072e-01 -8.57153386e-02 -5.15942931e-01
-1.79040298e-01 3.34859163e-01 -9.21528339e-01 2.60360450e-01
1.44380355e+00 1.21069528e-01 3.83451790e-01 3.11227232e-01
-6.33068979e-01 2.97514468e-01 -5.91782153e-01 -6.03789866e-01
3.74454647e-01 -6.80707633e-01 -7.38216102e-01 -8.14200193e-02
2.42696628e-01 1.88671976e-01 1.72419380e-02 1.02831697e+00
8.15518200e-01 2.20389575e-01 4.76441920e-01 -9.99552250e-01
-1.06858861e+00 7.58132935e-01 -4.76816803e-01 1.14911966e-01
2.19632626e-01 -6.83164820e-02 6.86579704e-01 -1.32097745e+00
1.04842567e+00 1.20940614e+00 7.27269650e-01 8.90478790e-01
-1.02472460e+00 -8.49955559e-01 -2.17124492e-01 3.48114111e-02
-1.65136981e+00 -5.21229982e-01 4.68986988e-01 -1.06260022e-02
1.04630804e+00 -1.04966216e-01 1.37743756e-01 1.28179669e+00
-7.14879832e-04 6.41616166e-01 1.42102730e+00 -8.41868222e-01
2.68006306e-02 6.10429943e-01 -7.90853947e-02 5.93554914e-01
3.17241341e-01 1.02850378e-01 -4.76878643e-01 1.12354672e-02
6.31357789e-01 -4.91795093e-01 -1.33728072e-01 -1.44099697e-01
-1.41348767e+00 8.91441405e-01 1.44900814e-01 7.01819837e-01
-2.95941383e-01 -5.16976237e-01 6.53622448e-01 4.86208469e-01
6.62315071e-01 6.32863104e-01 -1.02278876e+00 6.12057894e-02
-8.22778761e-01 -2.93212354e-01 9.98984098e-01 1.41822255e+00
9.51413929e-01 6.10326588e-01 -1.58660561e-01 1.22886872e+00
1.79693513e-02 9.00518596e-01 1.03474891e+00 -1.83608204e-01
8.48844826e-01 4.39245522e-01 1.13851599e-01 -3.96097004e-01
-1.53820619e-01 -3.44270736e-01 -7.13028848e-01 -2.45438546e-01
2.23831788e-01 -6.92872465e-01 -1.02386343e+00 2.01570034e+00
3.27607602e-01 -5.76374829e-02 7.53190935e-01 5.56430817e-01
7.09327996e-01 9.16009426e-01 4.16950315e-01 -2.22458497e-01
1.39379489e+00 -1.09916496e+00 -6.37681484e-01 -3.17761987e-01
1.13204122e+00 -9.52824533e-01 1.27486420e+00 -4.23255831e-01
-6.33661449e-01 -6.27035618e-01 -7.92646110e-01 -1.34621292e-01
-8.09758604e-01 4.31413263e-01 4.32902902e-01 6.93676651e-01
-8.89768958e-01 5.23642227e-02 -5.38984835e-01 -8.86406660e-01
-4.95170876e-02 1.31817430e-01 -7.40757942e-01 -4.68875587e-01
-1.72247553e+00 1.20030403e+00 8.17563951e-01 -1.51297033e-01
-6.01443112e-01 -6.64187551e-01 -1.12552464e+00 -1.72325835e-01
3.52284834e-02 -4.67543900e-01 1.15532804e+00 -8.88610184e-01
-1.29221332e+00 9.70349610e-01 -1.70627043e-01 -1.26168951e-01
3.95722151e-01 -1.10147707e-01 -9.31322277e-01 -4.14226919e-01
6.10777199e-01 6.89807296e-01 1.78592160e-01 -1.04801428e+00
-6.97516859e-01 -2.02279553e-01 -5.23406506e-01 5.25727630e-01
-5.02998769e-01 3.25119197e-01 -3.23405176e-01 -9.27863240e-01
-4.94854122e-01 -1.01543653e+00 -2.99109131e-01 -6.86619163e-01
-2.46011719e-01 -2.32488260e-01 3.70376438e-01 -8.99643421e-01
9.36740279e-01 -1.88056386e+00 -2.12074235e-01 -9.12355408e-02
-5.89346230e-01 4.82057750e-01 -7.16373920e-01 6.63619280e-01
-1.39429659e-01 3.42318982e-01 -2.41401672e-01 -9.53315422e-02
-2.44886324e-01 2.15304747e-01 -1.02578744e-01 7.26234838e-02
3.50858897e-01 9.61773753e-01 -9.58491743e-01 -5.89306951e-01
-5.03234342e-02 4.35977042e-01 -2.12666795e-01 2.79097974e-01
7.53261149e-02 8.24941814e-01 -4.83231813e-01 5.19040287e-01
4.98337001e-01 2.03111276e-01 1.48346916e-01 -1.31036386e-01
-2.77589291e-01 3.23607177e-01 -8.78399909e-01 1.85444570e+00
-9.10021722e-01 2.99752861e-01 -2.83443362e-01 -6.01308823e-01
1.18253398e+00 6.85719550e-01 1.38447553e-01 -7.38527477e-01
-1.91639923e-02 6.25067294e-01 -1.25178620e-01 -4.09321904e-01
6.84427559e-01 -5.63408673e-01 -6.20968461e-01 4.27506953e-01
5.19091785e-01 4.90719266e-02 5.15267968e-01 -1.65836737e-01
5.23295760e-01 3.20967287e-01 5.31271815e-01 -3.98236662e-01
4.30552244e-01 5.66848159e-01 1.00657165e+00 3.02534133e-01
-1.78835064e-01 1.54988885e-01 -3.83803308e-01 -2.78723150e-01
-1.28785872e+00 -8.78229201e-01 -1.12159103e-01 1.41413498e+00
-1.73753321e-01 -7.06990734e-02 -8.81369472e-01 -9.57981467e-01
-4.03082401e-01 1.15431523e+00 -2.05061540e-01 -2.27257013e-02
-7.21250176e-01 -8.90675843e-01 1.02254140e+00 4.50940281e-01
6.27122283e-01 -1.17215741e+00 3.90665203e-01 2.91380316e-01
-5.59418261e-01 -1.38227546e+00 -9.50370014e-01 1.00858353e-01
-7.72503495e-01 -6.38380170e-01 -1.05252242e+00 -1.45403445e+00
7.86726356e-01 1.02221794e-01 1.06953967e+00 -6.04282439e-01
8.97349641e-02 -3.36833932e-02 -4.08636332e-01 -4.00788337e-01
-8.02549005e-01 5.21677017e-01 4.66194749e-01 -1.11631505e-01
8.15681100e-01 -1.86695173e-01 2.33657509e-01 4.11954314e-01
-7.13300824e-01 1.49701312e-02 8.52094114e-01 1.00312471e+00
6.08856738e-01 -2.76514858e-01 9.53528583e-01 -9.86891806e-01
7.62067378e-01 -4.90918189e-01 -3.63803148e-01 7.74151206e-01
-4.51386482e-01 3.33852142e-01 9.62957978e-01 -7.39891887e-01
-1.57089341e+00 4.05894443e-02 -3.38436775e-02 -1.14100218e-01
-3.02926719e-01 6.14672542e-01 -5.23048580e-01 1.25692591e-01
9.42340672e-01 4.51407045e-01 -5.69578111e-01 -7.20390856e-01
6.59616232e-01 1.01714396e+00 5.26645303e-01 -6.54002547e-01
7.88510144e-01 -2.04201639e-01 -4.94522661e-01 -9.36929882e-01
-7.80965030e-01 -4.08824950e-01 -1.02944481e+00 2.39025861e-01
8.10545921e-01 -1.33381987e+00 3.07252645e-01 4.11500156e-01
-1.32423365e+00 -2.27854699e-01 -2.73982882e-01 1.00335777e+00
-2.56437063e-01 5.34379110e-02 -8.57476354e-01 -4.05734032e-01
-5.71970046e-01 -9.24001217e-01 8.01127255e-01 2.53463477e-01
-1.75095215e-01 -1.31942534e+00 5.66119492e-01 2.80900449e-01
2.99034327e-01 1.23391319e-02 9.44987655e-01 -1.29868770e+00
-1.35523394e-01 -2.76269913e-01 -2.21118703e-03 4.24263805e-01
4.28157389e-01 -4.07248527e-01 -4.76916611e-01 -4.00518268e-01
-1.44023359e-01 -5.02003968e-01 9.20758694e-02 -2.86444604e-01
1.11672487e-02 -3.07277441e-01 -1.48395717e-01 5.20558238e-01
1.47467685e+00 1.09753825e-01 2.74420053e-01 2.95452595e-01
9.67278063e-01 6.02935076e-01 6.68161392e-01 -2.18990192e-01
6.06849372e-01 4.94286627e-01 -7.52690256e-01 -4.19058532e-01
-2.84000695e-01 -5.97165763e-01 5.74625969e-01 1.73314822e+00
1.51362970e-01 -2.73412347e-01 -1.17507446e+00 7.89111078e-01
-1.36622202e+00 -7.37898409e-01 -1.21133318e-02 2.21769571e+00
1.29975247e+00 -4.60877836e-01 -1.40315667e-01 -6.89264059e-01
1.26246607e+00 -3.22940081e-01 -4.36887622e-01 -4.22412932e-01
-3.73873115e-01 3.10216218e-01 7.41383731e-01 4.26968843e-01
-9.08206165e-01 1.68379867e+00 6.18012524e+00 1.03228450e+00
-1.24527371e+00 3.76973242e-01 3.89437199e-01 4.38462794e-01
-2.51536220e-01 -6.86356202e-02 -1.27938187e+00 4.07471746e-01
1.28631866e+00 -6.07472599e-01 2.97196239e-01 8.57475102e-01
1.12198569e-01 4.89791781e-01 -9.32990015e-01 6.79606795e-01
1.74215555e-01 -8.46545696e-01 3.77181351e-01 -1.09357312e-01
1.11565340e+00 5.03154755e-01 -4.71693337e-01 7.68386364e-01
7.69206047e-01 -6.49405599e-01 3.37425649e-01 1.78723767e-01
1.39816046e+00 -9.10077989e-01 8.02703679e-01 6.36673450e-01
-1.28102779e+00 5.18618882e-01 -3.98558199e-01 3.80153179e-01
3.72818589e-01 1.36412710e-01 -1.16316557e+00 8.56624722e-01
3.32826972e-01 3.95272195e-01 -4.85793918e-01 7.06448078e-01
-4.05487388e-01 6.32011473e-01 -2.48837262e-01 -9.04473737e-02
2.08095595e-01 -2.84457535e-01 4.31732982e-01 1.50269735e+00
6.62210464e-01 -2.27507398e-01 4.14648503e-01 4.24273789e-01
-4.19159204e-01 9.56236243e-01 -8.98356795e-01 -3.20693135e-01
5.14780462e-01 1.16726351e+00 -4.16885614e-01 -4.67510581e-01
-6.66769028e-01 1.25705421e+00 6.43691361e-01 5.13954580e-01
-4.64522660e-01 -6.78573072e-01 3.27595651e-01 -3.00629884e-01
7.32366815e-02 -3.71834964e-01 -1.80608109e-01 -1.69817114e+00
-1.08036660e-01 -9.73871946e-01 3.56986642e-01 -5.67754328e-01
-1.71883667e+00 1.02324951e+00 -1.73229083e-01 -1.42241240e+00
-5.63432336e-01 -4.46355581e-01 -4.39910561e-01 1.34279227e+00
-1.48050058e+00 -1.55197382e+00 2.77917355e-01 6.23426855e-01
8.87526095e-01 -7.02795029e-01 1.23974669e+00 5.05660355e-01
-5.79432189e-01 9.88195837e-01 4.08693135e-01 6.57132208e-01
1.17629778e+00 -8.47153425e-01 6.94554389e-01 1.07901597e+00
3.17492664e-01 7.78949559e-01 3.82354081e-01 -9.46826816e-01
-1.05480099e+00 -1.53989685e+00 1.64882660e+00 -4.28001374e-01
7.27315366e-01 -4.36053008e-01 -9.65914130e-01 8.20275426e-01
4.26082700e-01 -4.05870192e-02 1.16902447e+00 1.05999112e-01
-3.14380318e-01 1.22309439e-01 -1.22540915e+00 8.41896117e-01
8.80191684e-01 -6.75878048e-01 -8.81412745e-01 2.60505110e-01
7.76952028e-01 -1.43809974e-01 -1.09354138e+00 3.37978750e-01
1.63291663e-01 -3.48137207e-02 5.21765351e-01 -1.02156901e+00
-3.23793553e-02 -3.26709867e-01 -3.09868425e-01 -1.76393318e+00
-1.94887981e-01 -4.87554967e-01 6.05579376e-01 1.74590015e+00
8.94811213e-01 -6.13476694e-01 3.02719712e-01 3.85515660e-01
-1.28466249e-01 -2.57836103e-01 -7.93226182e-01 -1.12506664e+00
3.35478127e-01 -1.20232657e-01 5.33223450e-01 1.53694344e+00
-1.84429586e-02 1.08034289e+00 -5.45761824e-01 1.77335218e-02
3.93211812e-01 -1.76434189e-01 6.67449594e-01 -8.69949639e-01
1.57780558e-01 1.91223606e-01 -9.20674726e-02 -6.64796472e-01
4.62187231e-01 -1.42297137e+00 3.12205940e-01 -1.39230525e+00
1.49826050e-01 -5.55948257e-01 -3.65260363e-01 6.50038898e-01
-3.77864301e-01 1.83744729e-01 1.65236399e-01 2.99022466e-01
-3.08039039e-01 7.19052255e-01 1.21889615e+00 2.06197754e-01
-3.35474104e-01 -1.96202934e-01 -5.47401667e-01 4.38065261e-01
8.47841322e-01 -7.67982066e-01 -3.24550480e-01 -6.70179009e-01
-2.48714998e-01 7.87873715e-02 -5.12812138e-01 -7.85015523e-01
-7.04970285e-02 -3.26901346e-01 3.17187846e-01 -1.88069090e-01
-8.56307223e-02 -5.62643409e-01 -1.50662009e-02 2.11426795e-01
-4.09503460e-01 3.27638358e-01 3.28548282e-01 3.47825378e-01
-3.11567158e-01 -1.19319461e-01 8.63495171e-01 -3.10453176e-01
-1.04913962e+00 4.04049337e-01 -2.58229256e-01 4.51846242e-01
1.00772047e+00 -6.97701797e-02 -1.68260336e-01 -8.10588226e-02
-5.50823987e-01 1.90957695e-01 6.15960360e-01 8.53744447e-01
2.77366489e-02 -1.76102018e+00 -1.28156006e+00 2.61908650e-01
4.52854306e-01 -5.02354205e-01 1.36865079e-01 4.16734397e-01
-3.06891173e-01 5.42899668e-01 -5.62306464e-01 -1.42150089e-01
-9.74070251e-01 5.67723453e-01 1.40875220e-01 -5.61727107e-01
-1.48444399e-01 5.63077807e-01 9.59748775e-02 -1.32807660e+00
-3.09280664e-01 2.23363236e-01 -1.71871305e-01 -7.25806430e-02
3.13351661e-01 2.78449684e-01 5.07651828e-02 -1.25097549e+00
-2.66934395e-01 4.30028200e-01 -1.61091805e-01 -5.21183968e-01
1.17851603e+00 -3.67518157e-01 9.19223055e-02 6.20766521e-01
1.39906275e+00 3.53939801e-01 -5.47772586e-01 -6.52235091e-01
4.10909623e-01 -7.98793733e-02 -4.11482036e-01 -8.68419826e-01
-6.63902462e-01 7.37327754e-01 4.90071863e-01 -5.24885416e-01
7.72111654e-01 -2.17207268e-01 1.07641649e+00 5.46916068e-01
7.84586668e-01 -1.34651947e+00 -3.87582839e-01 9.96546686e-01
6.21694267e-01 -1.43087637e+00 -6.18295133e-01 -2.75797993e-01
-9.50584769e-01 7.83230722e-01 8.15497160e-01 1.31243452e-01
2.25554124e-01 2.47568712e-01 5.67579210e-01 3.54317486e-01
-4.38017756e-01 -3.21672827e-01 2.96971679e-01 8.27989876e-01
8.49717259e-01 1.89313740e-01 -6.58229589e-01 6.79755569e-01
-3.04062605e-01 4.64680977e-02 3.29739392e-01 9.02021945e-01
-1.79529339e-01 -1.75564909e+00 -3.33189398e-01 3.57670197e-03
-6.50247693e-01 -5.57896256e-01 -3.42687458e-01 9.65784550e-01
1.24867797e-01 8.27929795e-01 -1.67916954e-01 -2.08420992e-01
4.44246858e-01 5.84313750e-01 1.88271940e-01 -1.00214255e+00
-4.91324455e-01 -8.36064219e-02 4.00216728e-01 4.17650007e-02
-2.79565960e-01 -3.84889305e-01 -1.33125699e+00 -5.56638390e-02
-3.59363288e-01 6.30751908e-01 8.76543164e-01 9.13968205e-01
5.01449883e-01 1.63779989e-01 8.18273127e-01 -4.45585698e-01
-5.37670732e-01 -1.27712572e+00 -4.77351993e-01 4.88389730e-01
-2.04414859e-01 -2.04635859e-01 2.21588500e-02 3.11329424e-01] | [10.032968521118164, 9.720654487609863] |
86d5d5bc-d4e9-40a4-a940-177c68e7cb02 | analysis-of-optimal-portfolios-on-finite-and | 2302.06778 | null | https://arxiv.org/abs/2302.06778v1 | https://arxiv.org/pdf/2302.06778v1.pdf | Analysis of optimal portfolios on finite and small-time horizons for a multi-dimensional correlated stochastic volatility model | In this paper, we consider the portfolio optimization problem in a financial market where the underlying stochastic volatility model is driven by n-dimensional Brownian motions. At first, we derive a Hamilton-Jacobi-Bellman equation including the scaled covariances between the standard Brownian motions. We use an approximation method for the optimization of portfolios. With such approximation, the value function is analyzed using the first-order terms of expansion of the utility function in the powers of time to the horizon. The error of this approximation is controlled using the second-order terms of expansion of the utility function. It is also shown that the one-dimensional version of this analysis corresponds to a known result in the literature. We also generate a close-to-optimal portfolio near the time to horizon using the first-order approximation of the utility function. It is shown that the error is controlled by the square of the time to the horizon. Finally, we provide an approximation scheme to the value function for all times and generate a close-to-optimal portfolio. | ['Indranil SenGupta', 'Minglian Lin'] | 2023-02-14 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-4.35890645e-01 3.01160902e-01 3.22876155e-01 7.51851201e-02
-4.13005501e-01 -8.33793163e-01 3.43951195e-01 -1.19560502e-01
-6.00147545e-01 8.00707519e-01 -1.11360930e-01 -3.92100692e-01
-5.97952247e-01 -9.38101530e-01 -5.00756323e-01 -9.37123299e-01
-1.81912556e-01 4.18012619e-01 1.28878146e-01 -3.07849884e-01
3.48855585e-01 4.76187617e-01 -9.69390571e-01 -5.35832942e-01
4.84189391e-01 1.51983702e+00 -3.23240727e-01 9.34991241e-01
-1.73889086e-01 4.81378525e-01 -3.85516465e-01 -3.93616378e-01
1.04069304e+00 -4.68387693e-01 -3.04856747e-01 9.71292146e-03
-4.86725450e-01 -4.54829514e-01 -6.87887147e-02 1.39561367e+00
3.15525591e-01 5.06253302e-01 7.38766611e-01 -1.01221347e+00
-3.58534724e-01 2.43049130e-01 -4.74855423e-01 3.26703012e-01
-2.08820760e-01 -3.56003582e-01 8.83646488e-01 -6.79608345e-01
4.62147027e-01 1.01828063e+00 6.43274784e-01 5.17007113e-01
-9.74059701e-01 -1.05229013e-01 -1.82506099e-01 -3.80824298e-01
-9.24743235e-01 4.29297201e-02 4.19153273e-01 -5.81341684e-01
6.36469603e-01 4.50176746e-01 6.01723850e-01 4.72003743e-02
6.00545824e-01 2.43559077e-01 8.15420330e-01 -4.71170366e-01
6.13167286e-01 1.18613228e-01 1.69071436e-01 3.29638273e-01
5.19166231e-01 4.67112809e-01 3.17298770e-01 -5.46007335e-01
9.81250286e-01 2.86178261e-01 -1.86042190e-01 -3.85908425e-01
-6.31282508e-01 1.16852999e+00 -2.80554950e-01 3.45155507e-01
-6.66762650e-01 2.45020211e-01 -8.16076528e-03 2.96627849e-01
5.80497980e-01 4.00975913e-01 -6.37425035e-02 -2.04755366e-01
-8.52357268e-01 4.86294955e-01 1.29721987e+00 8.77695262e-01
1.36071116e-01 2.25614190e-01 -2.46214598e-01 2.05502346e-01
3.26698184e-01 7.99897730e-01 4.51977313e-01 -1.34366214e+00
5.49732208e-01 7.13392273e-02 8.46958458e-01 -3.01082730e-01
-3.85608315e-01 -5.57232559e-01 -5.70545793e-01 5.56337893e-01
9.31450903e-01 -7.22003818e-01 -8.20943192e-02 1.64076567e+00
3.13214958e-01 -1.34351939e-01 3.25690389e-01 5.40937722e-01
-5.37270248e-01 6.10860109e-01 -7.70151377e-01 -9.63962555e-01
1.04995430e+00 -6.15937233e-01 -7.80069053e-01 4.52513844e-01
2.86032319e-01 -6.83979571e-01 4.79323775e-01 1.41278401e-01
-1.43918896e+00 -3.57204676e-02 -8.99794042e-01 6.85204744e-01
1.02735348e-01 -3.70331317e-01 -2.03751013e-01 7.95551836e-01
-1.11483467e+00 1.24350333e+00 -6.17999256e-01 1.07642415e-03
-2.66318470e-01 1.64793670e-01 3.08988899e-01 9.26233828e-01
-9.64530706e-01 7.86570370e-01 2.26344272e-01 1.02502964e-01
-4.47337031e-01 -8.84107232e-01 -3.43234301e-01 1.98093444e-01
3.49462837e-01 -4.45815891e-01 1.79256713e+00 -6.87759399e-01
-1.90511394e+00 2.99401611e-01 1.99939087e-01 -6.62195563e-01
1.14617431e+00 -6.46492988e-02 -3.09149697e-02 1.48840442e-01
1.57062292e-01 -4.46681380e-01 5.72560251e-01 -6.18982136e-01
-5.45238018e-01 -2.95953900e-01 3.61200087e-02 -7.20817922e-03
-2.22552270e-01 -1.40566275e-01 3.28271478e-01 -8.88584495e-01
-1.50508955e-01 -1.09631896e+00 -5.08021235e-01 -1.57765314e-01
3.23188258e-03 2.08375841e-01 2.82721162e-01 -6.87963784e-01
1.33796930e+00 -1.93555784e+00 -1.46378860e-01 2.82530576e-01
-2.13665232e-01 -1.88330874e-01 4.00802314e-01 6.61105692e-01
-1.64729312e-01 2.73564458e-01 -2.93947697e-01 -1.93781778e-01
3.37560534e-01 -2.82522384e-02 -7.33866513e-01 5.77729166e-01
-2.08481312e-01 5.98881781e-01 -7.62499988e-01 -1.68516804e-02
-1.53252065e-01 -1.28665328e-01 -4.01062608e-01 9.28327888e-02
-1.11632250e-01 5.19542769e-02 -7.29802072e-01 -6.67070448e-02
5.32789290e-01 -1.48957986e-02 -5.16289622e-02 4.59096849e-01
-7.22146988e-01 -2.48334363e-01 -1.26507008e+00 6.80548549e-01
-5.12293041e-01 2.59661108e-01 1.66580752e-01 -8.66110861e-01
8.22333753e-01 5.42237878e-01 7.19296515e-01 -3.74814153e-01
2.28516728e-01 6.27814531e-01 -1.52065799e-01 4.49026823e-02
2.45929644e-01 -1.16176605e+00 4.57259230e-02 7.94148326e-01
-4.78076518e-01 -1.30671054e-01 4.88472193e-01 -2.20409244e-01
8.47560406e-01 -2.69726545e-01 4.22410101e-01 -7.56482542e-01
4.37342852e-01 -3.03263038e-01 3.28361541e-01 5.07080019e-01
-1.64109319e-01 2.69956738e-01 1.05289590e+00 -2.71378845e-01
-1.36228025e+00 -9.71040606e-01 -3.52218181e-01 3.02197069e-01
-2.75826544e-01 4.49715182e-02 -9.70840096e-01 -1.63665041e-01
3.49263132e-01 1.22481382e+00 -8.94174099e-01 6.04246072e-02
-3.15460354e-01 -8.47324073e-01 -5.27275950e-02 3.48893583e-01
5.44846892e-01 -7.24667728e-01 -9.86679435e-01 4.49777663e-01
4.61092591e-01 -6.63447857e-01 -7.41125643e-01 5.61523773e-02
-9.46117580e-01 -1.02420092e+00 -1.29809952e+00 1.74775749e-01
3.47692847e-01 -1.86054990e-01 6.08128548e-01 -5.47906518e-01
1.25683025e-01 7.43610740e-01 1.66988820e-01 -8.62214446e-01
-3.84990662e-01 -6.89666152e-01 1.79446608e-01 4.48355705e-01
-1.75924480e-01 -3.29202712e-01 -6.00121081e-01 2.70695180e-01
-8.36054325e-01 -7.47521579e-01 -1.84217706e-01 5.51665545e-01
7.43272722e-01 2.97996581e-01 4.51630592e-01 -1.29701659e-01
9.27243233e-01 -3.35516959e-01 -1.76433527e+00 2.05510288e-01
-7.73144722e-01 5.50050616e-01 6.67258859e-01 -4.54313993e-01
-1.01915395e+00 -3.04702550e-01 4.96268749e-01 -2.86157668e-01
7.52083600e-01 1.20605893e-01 8.08808059e-02 -9.66202766e-02
3.91800515e-02 -2.06244905e-02 1.02622546e-01 -7.28572190e-01
2.63281316e-01 3.79010379e-01 3.01852822e-01 -5.63266933e-01
6.92305386e-01 6.86514437e-01 5.98932028e-01 -6.63352072e-01
-9.34089541e-01 -3.28551173e-01 -4.49108452e-01 -2.04874337e-01
8.53464127e-01 -1.15041919e-01 -1.04552960e+00 3.34629953e-01
-1.07454562e+00 -2.87350357e-01 -1.14165020e+00 7.26458549e-01
-1.22270870e+00 2.52423137e-01 -4.83271420e-01 -1.81729400e+00
-4.82727885e-01 -7.24262774e-01 4.82190669e-01 2.93108702e-01
-1.08818607e-02 -1.28876054e+00 6.58692896e-01 -4.19750899e-01
3.33259076e-01 5.88383377e-01 8.01239908e-01 -7.49688506e-01
-2.83656806e-01 -4.39692318e-01 2.24108085e-01 5.46239316e-01
-2.03256384e-01 -1.84372470e-01 -2.95565069e-01 -5.12516014e-02
1.17702353e+00 5.83510995e-01 3.81057531e-01 7.41244793e-01
4.21862811e-01 -7.17705190e-01 5.51459827e-02 5.82839429e-01
1.61036754e+00 7.88277209e-01 8.82106721e-02 7.53447890e-01
-1.09608978e-01 6.62340164e-01 6.42219663e-01 8.96176815e-01
-2.51857162e-01 4.81832057e-01 3.35463822e-01 6.04520738e-01
7.94272184e-01 1.29042799e-02 4.11808848e-01 3.92065912e-01
-5.22611141e-02 1.73747182e-01 -7.56882668e-01 4.40272599e-01
-1.96885133e+00 -1.24434352e+00 -2.02011079e-01 2.94331503e+00
5.12870431e-01 2.38541692e-01 4.23995823e-01 -1.27500474e-01
5.95120549e-01 -1.84702978e-01 -6.72801435e-01 -7.75901377e-01
-9.65530574e-02 1.14959605e-01 1.19015527e+00 9.16393340e-01
-8.59732866e-01 1.02205994e-02 7.27333498e+00 5.81272542e-01
-6.61588788e-01 6.59969971e-02 5.03790319e-01 -3.57807696e-01
-3.85019392e-01 1.41023517e-01 -8.90481234e-01 8.01164567e-01
1.10437107e+00 -1.22472131e+00 3.70277822e-01 1.03236532e+00
3.52938533e-01 -5.47046512e-02 -7.49103189e-01 5.42459846e-01
-4.36945528e-01 -9.46790338e-01 -4.55141962e-01 7.57453382e-01
8.63864481e-01 -4.30556923e-01 1.24131702e-01 -7.39656910e-02
2.52582937e-01 -5.51813006e-01 9.09608126e-01 9.60714221e-01
2.09114447e-01 -1.08771729e+00 9.04400229e-01 5.21636844e-01
-1.34485042e+00 -3.73891592e-01 -4.96017635e-01 -2.85463691e-01
7.26376355e-01 6.96130753e-01 -4.80583459e-01 4.30221677e-01
2.00159341e-01 6.22690795e-03 1.48228258e-01 1.23692095e+00
2.05570713e-01 1.67328835e-01 -7.08523273e-01 -2.67108202e-01
5.45494914e-01 -1.21682179e+00 8.13066542e-01 5.72236121e-01
1.03771698e+00 3.12941551e-01 -2.48744175e-01 7.84157276e-01
3.15950036e-01 1.63019210e-01 -4.36176628e-01 -2.31989473e-01
-1.87745094e-01 7.21778452e-01 -5.84644675e-01 -3.02175343e-01
-3.90256017e-01 7.55109012e-01 -2.19442502e-01 4.61796343e-01
-4.58048344e-01 -5.43902457e-01 7.71497548e-01 1.69473901e-01
4.85545844e-01 -1.98295429e-01 -2.89426029e-01 -7.17842579e-01
4.74853367e-01 -1.57333478e-01 4.22901481e-01 -2.87943214e-01
-9.89131331e-01 4.46619809e-01 3.21989983e-01 -1.06801558e+00
-7.94806063e-01 -6.83959901e-01 -9.12904441e-01 1.16499364e+00
-9.88542795e-01 1.41052648e-01 7.56613314e-01 3.62646967e-01
1.72371596e-01 -4.65658963e-01 5.54137766e-01 -1.75663874e-01
-4.69766945e-01 3.84269208e-02 1.17922091e+00 -2.12204367e-01
-3.98761928e-02 -1.42763650e+00 3.61111820e-01 9.18381035e-01
-3.79530936e-01 5.12715161e-01 1.13044250e+00 -8.32247615e-01
-9.20869231e-01 -7.57321119e-01 8.52021515e-01 -2.37711877e-01
1.39759493e+00 1.57932922e-01 -6.67022645e-01 6.81074917e-01
-2.08830251e-03 -1.44589216e-01 4.26897228e-01 -5.49303055e-01
3.75928789e-01 -2.24346235e-01 -1.22356784e+00 4.37538117e-01
2.14241669e-01 -3.31581622e-01 -5.29171944e-01 3.12625617e-01
6.58846378e-01 -7.93880522e-02 -8.03466499e-01 3.73232886e-02
7.44370461e-01 -9.14614379e-01 6.74027741e-01 -6.85120344e-01
-7.40462691e-02 1.72571819e-02 -3.13413203e-01 -1.13841915e+00
-1.68356046e-01 -1.47080839e+00 -2.33764753e-01 9.77531075e-01
2.43395343e-01 -1.04277742e+00 6.43695831e-01 9.89261925e-01
4.82410520e-01 -9.15069938e-01 -1.42201471e+00 -1.42443764e+00
5.94511032e-01 2.21799724e-02 3.42974395e-01 2.06359565e-01
1.23512119e-01 -2.71956325e-01 -3.12190950e-01 -5.51753603e-02
1.21170855e+00 1.60869882e-01 6.60151020e-02 -1.14657426e+00
-6.72472239e-01 -6.48408234e-01 -1.75097492e-02 -8.49330127e-01
4.28099409e-02 -4.82217759e-01 -2.29831010e-01 -1.16059351e+00
5.92188677e-03 -6.02618046e-03 -3.20263714e-01 -4.53318864e-01
2.49693349e-01 -5.39400935e-01 4.65335488e-01 8.48043934e-02
4.87925811e-03 4.70105976e-01 1.18157232e+00 2.56968021e-01
-2.31594250e-01 7.88175642e-01 -3.90219122e-01 9.71287966e-01
8.62613976e-01 -5.55845797e-01 -5.23968816e-01 2.44104683e-01
1.63557425e-01 5.90522826e-01 2.24541128e-01 -6.09647214e-01
5.11760078e-02 -4.07875806e-01 -1.81760594e-01 -8.28382432e-01
3.66625667e-01 -6.12979710e-01 5.03554106e-01 9.16848421e-01
-2.23663181e-01 3.81618857e-01 4.69143689e-02 7.23798513e-01
6.94416743e-03 -1.13818896e+00 8.96423876e-01 -2.46556759e-01
2.77939975e-01 2.27586016e-01 -3.01682234e-01 3.35397154e-01
1.25315118e+00 1.30456060e-01 1.22737959e-01 -8.41337144e-01
-8.19956362e-01 1.70940161e-01 3.32362384e-01 -2.31276363e-01
9.24852416e-02 -1.46127629e+00 -3.72885138e-01 -6.30119219e-02
-8.40761483e-01 -5.13588130e-01 1.35472059e-01 9.36649024e-01
-5.59345603e-01 7.66084671e-01 1.21062197e-01 9.70925763e-02
-8.93394530e-01 8.60137641e-01 7.60116816e-01 -6.11845434e-01
-1.71783552e-01 4.18797612e-01 2.77563125e-01 1.92171529e-01
3.66099365e-02 -4.22356308e-01 1.99903041e-01 2.50420898e-01
9.15015697e-01 9.25222218e-01 -2.75109410e-01 -3.62306386e-01
-4.25138734e-02 1.03295088e+00 4.58360016e-01 -8.70709121e-01
1.15573764e+00 -1.93003833e-01 -2.37852067e-01 7.47083187e-01
1.41834080e+00 -1.96433254e-02 -1.42247438e+00 1.54169332e-02
3.27236503e-01 -2.76116699e-01 -3.37497532e-01 -1.91873014e-01
-8.55874240e-01 8.68662596e-01 4.30677444e-01 7.50575483e-01
8.05518985e-01 -5.34633696e-01 5.73786259e-01 3.45813543e-01
4.52218026e-01 -1.32172132e+00 -3.21055681e-01 4.94619519e-01
1.08461475e+00 -5.51069379e-01 -1.62524834e-01 -2.76692230e-02
-5.61469078e-01 1.43712807e+00 -2.31976956e-01 -5.48786044e-01
1.25874293e+00 4.19095308e-01 -6.67572543e-02 3.49168271e-01
-5.72196543e-01 5.95386215e-02 1.06621362e-01 1.90424532e-01
2.03023091e-01 1.80248499e-01 -8.07217062e-01 7.50459254e-01
-3.41354251e-01 -9.31573734e-02 6.52441502e-01 7.43543863e-01
-5.78026950e-01 -1.20758545e+00 -6.60774589e-01 7.18240142e-02
-7.73944199e-01 -4.59853075e-02 -1.24198765e-01 7.57671416e-01
-5.71340561e-01 7.96313286e-01 2.48637751e-01 5.19751132e-01
6.60262406e-01 4.50622410e-01 2.25052357e-01 -1.82327032e-01
-2.66533405e-01 3.22178632e-01 -2.24084973e-01 -3.90642554e-01
2.27877662e-01 -8.96744132e-01 -1.20017934e+00 -3.14387709e-01
-5.88366836e-02 6.73964739e-01 5.65496981e-01 6.96545601e-01
-6.98373988e-02 1.26954049e-01 8.71278465e-01 -5.75173616e-01
-1.77948618e+00 -4.93178338e-01 -1.36286652e+00 -1.15981768e-03
4.30303037e-01 -3.93905848e-01 -8.02521586e-01 -4.60784078e-01] | [4.936863422393799, 3.9603147506713867] |
e091f354-0237-48f5-9764-efd13cf4e0ba | validate-on-sim-detect-on-real-model | 2111.00765 | null | https://arxiv.org/abs/2111.00765v3 | https://arxiv.org/pdf/2111.00765v3.pdf | Validate on Sim, Detect on Real -- Model Selection for Domain Randomization | A practical approach to learning robot skills, often termed sim2real, is to train control policies in simulation and then deploy them on a real robot. Popular techniques to improve the sim2real transfer build on domain randomization (DR) -- training the policy on a diverse set of randomly generated domains with the hope of better generalization to the real world. Due to the large number of hyper-parameters in both the policy learning and DR algorithms, one often ends up with a large number of trained policies, where choosing the best policy among them demands costly evaluation on the real robot. In this work we ask - can we rank the policies without running them in the real world? Our main idea is that a predefined set of real world data can be used to evaluate all policies, using out-of-distribution detection (OOD) techniques. In a sense, this approach can be seen as a `unit test' to evaluate policies before any real world execution. However, we find that by itself, the OOD score can be inaccurate and very sensitive to the particular OOD method. Our main contribution is a simple-yet-effective policy score that combines OOD with an evaluation in simulation. We show that our score - VSDR - can significantly improve the accuracy of policy ranking without requiring additional real world data. We evaluate the effectiveness of VSDR on sim2real transfer in a robotic grasping task with image inputs. We extensively evaluate different DR parameters and OOD methods, and show that VSDR improves policy selection across the board. More importantly, our method achieves significantly better ranking, and uses significantly less data compared to baselines. Project website is available at https://sites.google.com/view/vsdr/home. | ['Aviv Tamar', 'Gal Novik', 'Shadi Endrawis', 'Guy Jacob', 'Gal Leibovich'] | 2021-11-01 | null | null | null | null | ['robotic-grasping'] | ['robots'] | [ 1.89628135e-02 -5.42163178e-02 -3.89632702e-01 -1.60578370e-01
-9.44990039e-01 -1.07183242e+00 6.47884488e-01 -1.67439610e-01
-7.34489918e-01 8.19378674e-01 -1.40808821e-01 -6.05643332e-01
-9.11306739e-02 -5.62840641e-01 -1.15196705e+00 -7.46037781e-01
-2.41138682e-01 9.78837132e-01 5.24341106e-01 -4.62320894e-01
1.11813113e-01 5.14538109e-01 -1.55923414e+00 8.65600929e-02
7.34415829e-01 8.31953585e-01 5.35026312e-01 5.87884724e-01
3.08164418e-01 6.77088201e-01 -9.93271470e-01 2.07641646e-01
6.42357111e-01 -1.93095833e-01 -6.66867733e-01 -3.76604617e-01
2.38997847e-01 -5.17438591e-01 -1.00497544e-01 1.06259525e+00
6.77584767e-01 1.93188369e-01 6.49146378e-01 -1.43585050e+00
-2.13534907e-02 6.82852387e-01 -2.58268625e-01 -1.85994953e-01
3.49840879e-01 4.62201178e-01 6.46175086e-01 -1.93355188e-01
6.82089031e-01 1.45670617e+00 2.89102226e-01 7.45629609e-01
-1.38041306e+00 -8.77905011e-01 1.85897529e-01 -1.27293855e-01
-7.16234148e-01 -9.80318040e-02 2.78273523e-01 -5.59166789e-01
7.33496070e-01 -4.25302163e-02 4.21107084e-01 1.58074617e+00
5.03166839e-02 9.39329743e-01 1.41170657e+00 -3.27945918e-01
6.90366983e-01 2.10675523e-01 -2.50965804e-01 4.78995323e-01
2.49839455e-01 7.16153502e-01 2.79752128e-02 -2.72000521e-01
8.53941143e-01 -2.21178770e-01 -2.76748061e-01 -9.98450577e-01
-1.20793223e+00 6.70596719e-01 3.10649842e-01 1.72501784e-02
-3.15775454e-01 4.77762789e-01 6.91193879e-01 7.78519869e-01
5.37417270e-03 9.88367319e-01 -8.27922642e-01 -3.42132032e-01
-1.69939995e-01 7.24610269e-01 1.00643253e+00 8.99503946e-01
5.74669242e-01 -1.45125851e-01 -1.44461036e-01 7.88679898e-01
-6.69725239e-02 7.20614612e-01 6.73606992e-01 -1.33516729e+00
4.51206923e-01 8.89723077e-02 6.73933923e-01 -5.19059181e-01
-1.90623492e-01 -1.40055731e-01 -1.55407384e-01 1.07522273e+00
6.96845114e-01 -4.44345802e-01 -9.13708508e-01 2.12216043e+00
2.94476807e-01 -7.99106956e-02 1.21774703e-01 1.04113925e+00
-9.22673270e-02 4.18391198e-01 7.08804801e-02 1.60492033e-01
8.92231464e-01 -8.19178820e-01 -2.40992710e-01 -3.07759345e-01
6.04809284e-01 -5.04133821e-01 1.50885677e+00 6.12401485e-01
-7.74471283e-01 -3.73854727e-01 -9.51675236e-01 7.12223887e-01
-2.89678037e-01 1.75674930e-02 5.25519133e-01 2.58975118e-01
-9.88032401e-01 1.07205582e+00 -1.04432762e+00 -4.54152822e-01
2.55647331e-01 5.55275619e-01 -3.49397451e-01 -1.60203665e-01
-1.08724988e+00 1.27826941e+00 5.69906592e-01 -5.88943720e-01
-1.59169412e+00 -5.08872330e-01 -5.63399076e-01 -1.93423286e-01
7.22001195e-01 -3.66300255e-01 2.03445244e+00 -1.02626061e+00
-1.95645905e+00 5.53230882e-01 3.07096750e-01 -4.04831618e-01
8.77782941e-01 -6.52213395e-01 1.78097114e-01 -5.24836183e-02
1.42539695e-01 6.35007977e-01 8.59049678e-01 -1.58187461e+00
-7.05108643e-01 -8.96367207e-02 3.86534423e-01 2.79797196e-01
4.22911905e-02 -7.58321136e-02 -2.72200286e-01 -4.81842488e-01
-3.37340057e-01 -1.28162253e+00 -2.96239018e-01 -2.07269639e-01
-7.83682242e-02 -3.29087794e-01 5.66154540e-01 -2.10354552e-01
5.84385455e-01 -2.09381390e+00 2.60356933e-01 1.68787405e-01
-2.84073800e-01 3.57623637e-01 -3.66057754e-01 4.15224969e-01
-1.25417769e-01 -8.06441084e-02 -8.17733333e-02 1.56825230e-01
3.40843976e-01 3.59986991e-01 -6.19456351e-01 4.62015986e-01
4.71780300e-02 6.02965117e-01 -1.36232984e+00 -1.88643843e-01
3.38279247e-01 1.76639184e-02 -6.50231361e-01 4.77984220e-01
-6.82631731e-01 6.22131824e-01 -6.64116025e-01 2.54047543e-01
3.15274864e-01 2.85783187e-02 3.57143521e-01 2.67139614e-01
2.30455771e-02 4.29948419e-01 -1.13743210e+00 1.59935927e+00
-5.31907260e-01 4.54716802e-01 5.99005036e-02 -1.11701369e+00
6.39914572e-01 1.75834849e-01 5.42125404e-01 -6.92169726e-01
2.07695186e-01 4.33313519e-01 7.35267252e-02 -4.70719159e-01
-2.95784026e-02 1.04033381e-01 -3.08055639e-01 3.06267381e-01
1.55480653e-01 -4.09317464e-01 2.52338856e-01 -6.41327128e-02
1.52751374e+00 7.24470258e-01 8.06990415e-02 -3.33097816e-01
1.36930244e-02 3.86109501e-01 5.27698815e-01 1.02985585e+00
-2.64678746e-01 1.02598481e-01 7.66356051e-01 -1.68220013e-01
-1.20853329e+00 -1.14565110e+00 1.14880651e-01 1.32574475e+00
2.94629633e-01 4.58634309e-02 -7.63397574e-01 -1.02428472e+00
4.81137425e-01 9.36764598e-01 -5.98220766e-01 -2.19910100e-01
-7.22256899e-01 -2.93848425e-01 4.48884577e-01 4.22582865e-01
1.56373918e-01 -1.23078442e+00 -1.06947196e+00 1.35794237e-01
1.78252101e-01 -9.03529465e-01 -1.37387738e-01 6.10707581e-01
-7.34693944e-01 -1.13473177e+00 -7.86222100e-01 -6.61835790e-01
4.86560673e-01 1.42209977e-01 1.16348755e+00 -4.11020130e-01
7.27486610e-02 3.79104793e-01 -5.22266150e-01 -5.18168092e-01
-7.81538963e-01 -3.53539325e-02 4.32184190e-01 -7.71592855e-01
7.88114145e-02 -6.56976998e-01 -5.87448359e-01 6.21862710e-01
-6.81069434e-01 -3.29744846e-01 8.16183925e-01 9.69297230e-01
3.30200106e-01 -2.88942605e-02 4.18780774e-01 -7.84003735e-01
8.24954152e-01 -3.40344369e-01 -1.15166068e+00 2.20960319e-01
-5.11098564e-01 4.85827386e-01 8.01351666e-01 -8.61814320e-01
-5.71316242e-01 -1.53296264e-02 1.26419097e-01 -7.49868870e-01
-3.04349542e-01 4.93061058e-02 -4.36275117e-02 6.37406483e-02
1.02869940e+00 -1.94633052e-01 2.36995131e-01 -4.55738276e-01
4.41814482e-01 5.22887230e-01 3.72954965e-01 -1.22703350e+00
7.06147552e-01 1.23280860e-01 -3.81402075e-01 -3.21045786e-01
-6.15590513e-01 -2.96919912e-01 -5.77455014e-02 -5.97735755e-02
4.85898972e-01 -7.64436901e-01 -9.37826455e-01 3.27157944e-01
-8.78911018e-01 -1.36720538e+00 -4.56588745e-01 6.51917458e-01
-9.38472569e-01 5.37093803e-02 -3.59387934e-01 -6.96663976e-01
1.21923588e-01 -1.54226148e+00 1.08860004e+00 1.04241356e-01
-1.57286987e-01 -6.20217204e-01 2.16350377e-01 -2.71034032e-01
3.06646913e-01 2.29249984e-01 7.95108795e-01 -8.17602992e-01
-1.74810871e-01 -9.35903639e-02 3.60781364e-02 5.23822129e-01
7.46939108e-02 -3.56656373e-01 -8.32267523e-01 -7.26810992e-01
-2.41892606e-01 -8.39208484e-01 5.04885972e-01 3.12254220e-01
1.20668876e+00 -1.42056510e-01 -4.03079212e-01 2.86115229e-01
1.38376319e+00 5.08995295e-01 4.07338649e-01 8.41445148e-01
2.23515555e-01 4.53632116e-01 1.18136072e+00 3.75480860e-01
-1.07922114e-01 7.89820910e-01 5.58717847e-01 1.14980347e-01
1.68578178e-01 -3.13495040e-01 7.95922935e-01 8.07633251e-02
9.60334092e-02 -2.96473026e-01 -8.15639496e-01 4.25275147e-01
-1.91835296e+00 -8.42365801e-01 7.04479456e-01 2.49095488e+00
8.74341667e-01 3.56328934e-01 4.59291220e-01 -2.09017441e-01
5.68147957e-01 -4.10317212e-01 -9.17909622e-01 -2.28863999e-01
2.98201412e-01 2.90017664e-01 8.55485380e-01 4.15708482e-01
-1.06697369e+00 1.06536686e+00 5.89831066e+00 8.49228263e-01
-1.37125885e+00 -2.31667273e-02 1.63305834e-01 -3.64425555e-02
8.94000381e-02 -3.10243163e-02 -6.12826645e-01 5.51439524e-01
6.45213246e-01 -6.10877834e-02 7.10251212e-01 1.37608933e+00
2.06549272e-01 -2.77593851e-01 -1.30398178e+00 6.91854835e-01
-4.99266267e-01 -7.65812099e-01 -3.48315299e-01 2.05021854e-02
6.45478368e-01 4.82254148e-01 1.09092118e-02 7.26521671e-01
1.29750931e+00 -9.00936961e-01 8.69566381e-01 1.45493615e-02
8.71533394e-01 -5.74994624e-01 6.72008216e-01 4.97974366e-01
-5.39574981e-01 -3.63475919e-01 -4.34673995e-01 1.71138957e-01
-1.70916334e-01 1.31673068e-01 -1.04665923e+00 4.05549318e-01
7.81874657e-01 2.83529282e-01 1.59640033e-02 1.00723672e+00
-4.72181141e-01 6.01269007e-01 -5.34095883e-01 -2.75767207e-01
2.90295094e-01 7.02622980e-02 5.03062308e-01 9.66704607e-01
2.04212099e-01 -1.76048338e-01 5.52063346e-01 6.24316990e-01
2.62324631e-01 -2.86540955e-01 -8.60111952e-01 4.74124635e-03
6.08479798e-01 7.83105195e-01 -3.83214623e-01 -2.22434998e-01
8.89812708e-02 7.97215402e-01 5.25026798e-01 4.31819767e-01
-8.65358174e-01 -3.34939659e-01 9.44151461e-01 -1.54126331e-01
2.94845998e-01 -2.83085853e-01 1.83167204e-01 -9.39570487e-01
-4.29735407e-02 -1.44604766e+00 2.15126947e-01 -5.40872812e-01
-1.25333536e+00 4.11635906e-01 3.43162030e-01 -1.54529893e+00
-4.65745181e-01 -1.04317153e+00 -2.59648919e-01 6.02576077e-01
-1.39116955e+00 -4.70349640e-01 -1.51207969e-01 3.72144550e-01
5.76109648e-01 -2.70067394e-01 8.34941030e-01 1.54501423e-01
-3.41035813e-01 5.80103397e-01 5.11696756e-01 -1.31071443e-02
1.09963942e+00 -1.48428679e+00 2.22192138e-01 2.87628502e-01
-3.14047635e-01 5.11352479e-01 1.10751271e+00 -5.35027206e-01
-1.32517827e+00 -9.27969992e-01 -1.48783177e-01 -4.82810408e-01
8.59700680e-01 -3.14252347e-01 -6.76682830e-01 7.11252272e-01
-6.36818856e-02 -1.42039254e-01 -1.33908875e-02 -4.83615920e-02
-2.08124742e-01 -4.62564081e-02 -1.07066524e+00 7.82670379e-01
1.00950873e+00 -1.15081526e-01 -6.36206746e-01 4.25926328e-01
8.41543078e-01 -5.97111404e-01 -7.94707477e-01 6.05575621e-01
6.63194895e-01 -7.58422136e-01 7.91302145e-01 -8.35857987e-01
3.59675735e-01 -3.30385268e-01 -1.86623007e-01 -1.84878719e+00
-3.63123529e-02 -7.12177098e-01 2.23939598e-01 7.75279105e-01
3.79111648e-01 -9.00421739e-01 6.60738111e-01 2.48479739e-01
-8.37820861e-03 -7.40008891e-01 -5.56730986e-01 -1.36293542e+00
3.40079963e-01 -3.20024848e-01 5.00958562e-01 7.57008493e-01
1.30990952e-01 1.79859579e-01 -7.17765614e-02 1.66496992e-01
5.25411606e-01 1.77498031e-02 1.04714727e+00 -9.47197020e-01
-6.72928095e-01 -4.60412651e-01 -2.49476194e-01 -1.19180751e+00
3.40694308e-01 -7.28005469e-01 5.37351668e-01 -1.20297885e+00
-1.44933030e-01 -9.74022150e-01 -3.82798016e-01 6.68356717e-01
2.49381806e-03 -4.88967031e-01 3.11588377e-01 2.55156815e-01
-5.99868536e-01 5.80742478e-01 1.34926116e+00 -2.08933782e-02
-2.20332265e-01 1.10568292e-01 -2.17662483e-01 6.48038745e-01
1.28462243e+00 -6.25180125e-01 -6.00245357e-01 -4.21650231e-01
-7.27900118e-02 6.41451031e-02 3.91217709e-01 -1.14941788e+00
-1.91710472e-01 -4.09170389e-01 1.19815312e-01 -3.58463451e-02
2.90409550e-02 -7.44445622e-01 -2.84646749e-01 7.49796867e-01
-4.73313481e-01 2.07784534e-01 4.18502897e-01 5.80149353e-01
1.47031054e-01 -2.25604251e-01 8.53053331e-01 -2.93343127e-01
-8.51475477e-01 -1.89484973e-02 -3.50742579e-01 2.39823610e-01
1.16790354e+00 1.83460668e-01 -3.91249299e-01 -2.99741179e-01
-4.73245025e-01 5.39122403e-01 8.37926626e-01 4.91247773e-01
1.04917519e-01 -1.03149891e+00 -4.56874430e-01 -9.58173499e-02
1.22615553e-01 5.20185530e-02 -4.32163209e-01 4.06759024e-01
-4.21726257e-01 1.18701458e-01 -4.09022272e-01 -6.90227151e-01
-9.59834337e-01 6.43653154e-01 5.26403368e-01 -4.05354887e-01
-5.49918890e-01 7.73229957e-01 3.28351706e-01 -9.11183059e-01
5.62228620e-01 -3.10667694e-01 -1.25723351e-02 -4.53408957e-01
1.92213833e-01 1.95725620e-01 -1.58719793e-01 1.64194152e-01
-2.48632073e-01 3.60525876e-01 1.80876330e-02 -6.32611334e-01
1.23962748e+00 4.86313522e-01 2.33749494e-01 3.69803637e-01
9.35254693e-01 -3.11003208e-01 -1.71711552e+00 1.18431635e-01
1.39213413e-01 -4.65960532e-01 -3.70825112e-01 -1.07439649e+00
-6.82957470e-01 6.14572346e-01 8.74849439e-01 8.68405253e-02
8.05027664e-01 -9.97119322e-02 3.39189798e-01 7.94800937e-01
9.62395370e-01 -1.29059827e+00 2.48616308e-01 5.68521619e-01
9.15416598e-01 -1.39215767e+00 -2.37423316e-01 4.58346978e-02
-8.67169023e-01 8.80495250e-01 9.09546316e-01 -5.85169852e-01
2.20264167e-01 4.49093878e-01 2.06496075e-01 7.92883113e-02
-6.90762579e-01 -1.88178062e-01 -2.75957257e-01 7.31813610e-01
1.41829737e-02 1.55159518e-01 -3.13061699e-02 6.42422140e-02
-2.55369127e-01 1.99213341e-01 1.98202521e-01 1.24570763e+00
-5.46333551e-01 -1.38150167e+00 -3.08324248e-01 2.61298925e-01
-2.21742094e-01 4.97109562e-01 -2.40705043e-01 9.66671467e-01
-1.14837669e-01 7.75184035e-01 -2.80743033e-01 -5.99531889e-01
6.41483605e-01 -4.74864729e-02 7.11161852e-01 -5.59149921e-01
-4.35088366e-01 -1.93146497e-01 2.18182221e-01 -1.02889657e+00
-7.97319487e-02 -6.05161786e-01 -1.27247500e+00 -1.59184441e-01
-1.37025610e-01 2.59953260e-01 7.38143206e-01 6.54720843e-01
3.80366772e-01 5.26769161e-01 6.78220034e-01 -1.14719844e+00
-1.44074404e+00 -8.58417988e-01 -3.28365177e-01 5.46575248e-01
4.20893610e-01 -1.18794954e+00 -3.71961147e-01 -3.69650841e-01] | [4.374575614929199, 1.4457868337631226] |
63472f63-d2bd-4d9b-bc7a-e9821dce352e | data-quality-as-predictor-of-voice-anti | 2103.14602 | null | https://arxiv.org/abs/2103.14602v2 | https://arxiv.org/pdf/2103.14602v2.pdf | Data Quality as Predictor of Voice Anti-Spoofing Generalization | Voice anti-spoofing aims at classifying a given utterance either as a bonafide human sample, or a spoofing attack (e.g. synthetic or replayed sample). Many anti-spoofing methods have been proposed but most of them fail to generalize across domains (corpora) -- and we do not know \emph{why}. We outline a novel interpretative framework for gauging the impact of data quality upon anti-spoofing performance. Our within- and between-domain experiments pool data from seven public corpora and three anti-spoofing methods based on Gaussian mixture and convolutive neural network models. We assess the impacts of long-term spectral information, speaker population (through x-vector speaker embeddings), signal-to-noise ratio, and selected voice quality features. | ['Tomi Kinnunen', 'Md Sahidullah', 'Rosa González Hautamäki', 'Bhusan Chettri'] | 2021-03-26 | null | null | null | null | ['voice-anti-spoofing'] | ['audio'] | [ 2.84461081e-01 -1.55231813e-02 -1.62294418e-01 -2.38671571e-01
-7.47235179e-01 -7.81295478e-01 9.56636548e-01 4.91261482e-03
-3.40188533e-01 3.69413108e-01 7.73571014e-01 -5.87460637e-01
-3.37386578e-02 -2.74625510e-01 -3.68526310e-01 -6.62401855e-01
-2.36438699e-02 2.53475726e-01 -1.10515423e-01 -2.25809410e-01
6.58078641e-02 3.68659675e-01 -1.13069558e+00 1.94284409e-01
6.01955295e-01 5.84577978e-01 -1.82000712e-01 1.07327211e+00
2.01036096e-01 5.25722861e-01 -1.23103690e+00 -3.76737535e-01
-1.28263040e-02 -5.19324183e-01 -7.11328685e-01 -7.77848959e-02
3.06732267e-01 4.14732769e-02 -4.46937203e-01 1.24362004e+00
8.84238660e-01 1.18995108e-01 6.25286937e-01 -1.21338999e+00
-7.41087675e-01 7.37787902e-01 9.78702605e-02 4.86449301e-01
4.83531505e-01 3.00691277e-01 8.62350345e-01 -5.34465909e-01
4.19156611e-01 1.63593197e+00 6.52988195e-01 7.19105780e-01
-1.44304872e+00 -8.34854245e-01 -4.50542182e-01 -3.61059122e-02
-7.69304156e-01 -1.28182089e+00 1.05464458e+00 -4.39520597e-01
6.64856195e-01 4.33647960e-01 1.73257798e-01 1.82347834e+00
9.26968381e-02 5.45425117e-01 1.01175928e+00 -5.46267807e-01
1.03557736e-01 3.41670394e-01 2.32718855e-01 1.96058542e-01
-5.58206402e-02 4.22178626e-01 -8.47508669e-01 -7.54937053e-01
1.00443386e-01 -5.07969618e-01 -3.74360681e-01 -8.41385797e-02
-1.20404029e+00 9.20163989e-01 -1.80320576e-01 5.33371270e-01
-2.27137119e-01 -1.68773249e-01 7.76340842e-01 6.40398741e-01
6.05415046e-01 6.65516198e-01 -6.75816953e-01 -1.17110409e-01
-8.32688212e-01 4.64146674e-01 8.75290632e-01 2.26405054e-01
2.67897546e-01 6.17102087e-01 8.11214969e-02 1.05025697e+00
3.16753805e-01 8.90988648e-01 8.88939798e-01 -6.69596076e-01
2.70786762e-01 -4.29093212e-01 1.48189425e-01 -1.07246017e+00
-4.70543176e-01 -3.01156968e-01 -3.59768212e-01 -1.38004676e-01
4.13379997e-01 -3.95402312e-01 -5.61403930e-01 1.87274933e+00
1.85025573e-01 1.21913433e-01 1.64976388e-01 5.45032918e-01
5.31442821e-01 5.75744212e-01 2.04616487e-02 -5.54135740e-01
1.29294729e+00 -7.00028300e-01 -1.13377547e+00 -2.00043365e-01
3.91806453e-01 -1.02161336e+00 1.14631188e+00 4.90917742e-01
-7.87737548e-01 -6.26373768e-01 -8.04156184e-01 4.82148498e-01
-2.95530051e-01 -4.34875578e-01 1.02056200e-02 1.52771270e+00
-7.71078408e-01 5.63339114e-01 -4.38963741e-01 -1.52204990e-01
-9.06918645e-02 1.15797460e-01 -4.33498830e-01 4.30884957e-01
-1.38580489e+00 7.29754329e-01 2.67637938e-01 -6.55006051e-01
-1.05260956e+00 -7.13613987e-01 -7.76377201e-01 -1.54931426e-01
-1.62657604e-01 -4.38951582e-01 1.34233034e+00 -8.63265634e-01
-1.65360284e+00 8.11261296e-01 -1.67261332e-01 -7.50647604e-01
2.67034024e-01 -4.47452813e-01 -1.36526203e+00 1.56309500e-01
-1.06263056e-01 6.88383356e-02 1.46896982e+00 -1.18198228e+00
-2.13706881e-01 -4.64518219e-01 -5.85633159e-01 -2.15679213e-01
-4.74669605e-01 6.30072534e-01 5.32782793e-01 -1.03655851e+00
4.06005085e-02 -1.08593929e+00 2.17760637e-01 -6.95081592e-01
-7.98624337e-01 -7.80091211e-02 1.09073198e+00 -1.04927075e+00
1.21816206e+00 -2.48799682e+00 -2.50183433e-01 -8.78054649e-02
-1.57617807e-01 8.76851380e-01 -6.67064339e-02 3.81553590e-01
-4.32119876e-01 4.84403074e-01 -5.53286374e-02 -4.76467311e-01
1.58303231e-01 -1.50543004e-01 -5.69903016e-01 9.45050180e-01
1.12718478e-01 6.22686334e-02 -9.75769937e-01 -3.42351586e-01
4.48427260e-01 8.29536259e-01 -8.13876569e-01 3.18972379e-01
9.44175124e-02 5.14953673e-01 3.55155729e-02 3.27843755e-01
6.11693978e-01 6.80985510e-01 -5.01675270e-02 5.08925803e-02
1.85185999e-01 1.02973342e+00 -8.31852853e-01 1.21009326e+00
-4.22285438e-01 8.34615827e-01 4.24870551e-01 -8.00676227e-01
1.02501011e+00 8.17445099e-01 3.39767277e-01 -2.39681408e-01
2.14495480e-01 3.24780107e-01 2.62342155e-01 -5.19738734e-01
4.32705522e-01 -5.86821258e-01 -2.21435484e-02 4.13729906e-01
4.98573065e-01 -2.62230635e-01 -4.60877657e-01 -2.36484751e-01
1.09932423e+00 -5.28909206e-01 6.35378212e-02 -4.64286119e-01
5.01887560e-01 -5.32067537e-01 5.15891552e-01 7.76772261e-01
-8.86229575e-01 5.62361181e-01 1.85095891e-01 -3.47355939e-02
-9.10995305e-01 -1.27825928e+00 -3.47945601e-01 1.40367019e+00
-4.15523976e-01 -2.10835546e-01 -9.67393219e-01 -5.89473426e-01
4.95510213e-02 1.31135130e+00 -2.45080158e-01 -4.39732611e-01
-5.13285518e-01 -4.68853951e-01 1.24023569e+00 -1.78987876e-01
-3.18725444e-02 -1.26532328e+00 -1.36558741e-01 3.05212438e-01
-3.18609595e-01 -1.04140949e+00 -5.96804082e-01 2.17369869e-01
-5.70336282e-01 -7.19783604e-01 -3.38204652e-01 -4.67503220e-01
-2.56836355e-01 1.60852313e-01 1.02602935e+00 -3.54156256e-01
2.19581559e-01 3.28330636e-01 -4.01933849e-01 -7.19804883e-01
-1.25537360e+00 -1.44846141e-01 1.00039113e+00 2.18070343e-01
5.57893872e-01 -5.75788498e-01 -1.67910948e-01 3.55092376e-01
-6.75015450e-01 -8.58505428e-01 -7.93134868e-02 9.15938437e-01
-2.34118819e-01 3.06361914e-01 7.75580466e-01 -8.04074347e-01
1.12590301e+00 -5.05915165e-01 2.61940777e-01 -1.78958714e-01
-4.67889339e-01 -7.67752305e-02 4.88552153e-01 -6.88705325e-01
-8.64050925e-01 -4.66227472e-01 -4.77455407e-01 -5.08276522e-01
-6.62179053e-01 8.94856919e-03 -4.16226536e-01 3.84963095e-01
1.09144926e+00 1.55289292e-01 2.04159960e-01 -7.73602068e-01
4.55986708e-01 1.16235518e+00 7.63860703e-01 -4.86352921e-01
9.74978507e-01 2.74991661e-01 -7.42421091e-01 -1.47701383e+00
-3.47910821e-01 -5.65608561e-01 -4.40780044e-01 1.07826695e-01
6.28496826e-01 -5.41493475e-01 -7.06073403e-01 6.53840184e-01
-1.32336366e+00 -1.54053345e-02 -9.34355631e-02 7.20168412e-01
-5.64821124e-01 4.20731992e-01 -6.56655371e-01 -1.31255567e+00
-3.00153643e-01 -1.09828210e+00 6.48131251e-01 -2.24406287e-01
-7.37789512e-01 -9.96173501e-01 3.08409452e-01 4.89006460e-01
5.45933664e-01 1.26671284e-01 7.00813591e-01 -1.39734054e+00
6.42028332e-01 -3.45386505e-01 4.34452236e-01 7.27406085e-01
3.18211496e-01 1.30508110e-01 -1.57125604e+00 -3.05016756e-01
6.90491378e-01 -2.14554042e-01 6.34746671e-01 4.78767037e-01
6.38665140e-01 -8.42981040e-01 -2.21940316e-03 3.78356516e-01
6.50362849e-01 2.64553398e-01 2.96992928e-01 3.00662145e-02
4.76271480e-01 9.33724999e-01 2.54387528e-01 2.93121994e-01
-3.58484447e-01 5.85071743e-01 1.89279124e-01 4.16070163e-01
-1.82440970e-02 -5.56864321e-01 8.37614179e-01 1.05112052e+00
4.21154529e-01 -6.30434155e-01 -8.62491786e-01 6.76028371e-01
-8.49454522e-01 -1.24612939e+00 5.33550233e-02 2.14166665e+00
7.31937230e-01 2.69068718e-01 6.77097321e-01 8.95580828e-01
1.07585144e+00 7.85712719e-01 -4.88300100e-02 -8.52324724e-01
-1.98497817e-01 9.09822807e-02 3.59852880e-01 7.75021851e-01
-1.18634593e+00 8.29079330e-01 6.78475380e+00 7.87736773e-01
-1.19725740e+00 2.74841547e-01 3.69123548e-01 4.19498943e-02
-3.83721799e-01 -2.87590951e-01 -3.80826563e-01 5.76314330e-01
1.54062545e+00 -1.02799304e-01 5.82787216e-01 8.46916497e-01
3.57626528e-01 4.78391409e-01 -9.16181743e-01 8.29233885e-01
1.03977099e-01 -9.43874538e-01 -3.27353716e-01 1.15387134e-01
1.55385524e-01 2.80934066e-01 4.59431976e-01 1.70713589e-01
4.28120524e-01 -9.11942542e-01 9.91590381e-01 -5.21477051e-02
7.33127058e-01 -6.50124848e-01 4.08400029e-01 3.73084158e-01
-5.36655903e-01 -1.04594007e-01 2.31430102e-02 1.18726663e-01
1.97647065e-01 4.50500995e-01 -9.98538852e-01 1.66086480e-01
4.58902597e-01 4.23734300e-02 -5.42543158e-02 3.91765416e-01
1.41955132e-03 1.35316360e+00 -1.09146707e-01 8.58922601e-02
-1.22954212e-01 3.35509598e-01 1.21314764e+00 1.62807822e+00
1.29771665e-01 -3.17999065e-01 -1.57608718e-01 6.01056337e-01
1.63251087e-01 -7.07046390e-02 -8.23661089e-01 -5.09776115e-01
8.20926964e-01 6.16425633e-01 -9.06204507e-02 3.82605344e-02
-2.01995336e-02 7.93321490e-01 -3.20457667e-01 4.00207430e-01
-4.55897093e-01 -4.87428218e-01 1.25374854e+00 2.52707243e-01
-1.62671641e-01 -2.63027966e-01 -3.21250975e-01 -9.42987800e-01
-4.92824554e-01 -1.44783223e+00 2.76709527e-01 -3.14297259e-01
-1.29315042e+00 5.35462677e-01 -3.06099534e-01 -7.65240788e-01
-5.18595994e-01 -5.53382874e-01 -6.25037730e-01 9.01686072e-01
-9.95062709e-01 -6.00376010e-01 4.47054595e-01 5.48987865e-01
4.84302342e-01 -6.12465382e-01 1.01915824e+00 2.49175206e-01
-4.73753303e-01 7.64833152e-01 5.71152084e-02 4.38381016e-01
7.87542462e-01 -1.06155288e+00 8.00548017e-01 6.37124002e-01
3.74169976e-01 8.30458343e-01 1.46834433e+00 -2.56366819e-01
-1.18458033e+00 -7.21666455e-01 1.09467769e+00 -7.18185246e-01
8.38213086e-01 -2.00448439e-01 -9.25925672e-01 3.49534899e-01
2.97744036e-01 8.12953264e-02 9.80206430e-01 2.37602875e-01
-7.28635252e-01 1.38656557e-01 -1.51746607e+00 2.94419229e-01
7.51058936e-01 -1.35341477e+00 -8.87158990e-01 1.20787844e-01
1.12858915e+00 1.04355767e-01 -6.54612124e-01 2.81997293e-01
4.33292568e-01 -1.28710210e+00 1.08331132e+00 -9.84291971e-01
4.14174423e-03 1.39708772e-01 -6.28895283e-01 -1.68325698e+00
-3.11592042e-01 -1.27664924e+00 -9.22613870e-03 1.57790589e+00
2.67097741e-01 -7.91180372e-01 3.93125206e-01 1.56033128e-01
-1.79755658e-01 9.69818234e-03 -1.07653117e+00 -1.09377813e+00
2.85147429e-01 -9.26206291e-01 7.88613081e-01 1.27860963e+00
1.82633132e-01 3.82487625e-01 -6.31795526e-01 3.81554335e-01
5.32204747e-01 -7.72076845e-01 9.35678959e-01 -1.00172472e+00
-5.01208305e-01 -4.70770419e-01 -2.84993887e-01 -6.12260818e-01
5.02364218e-01 -7.32307732e-01 3.71570094e-03 -6.20279431e-01
-5.16945004e-01 -2.54578054e-01 -4.54200149e-01 -1.34742305e-01
-1.56470209e-01 -8.74390230e-02 2.91593224e-01 5.16885184e-02
1.39605790e-01 3.97602528e-01 6.07564092e-01 -1.07558295e-01
-3.07922035e-01 3.35729897e-01 -3.07431817e-01 8.72837245e-01
9.30257916e-01 -7.55676508e-01 -1.33924097e-01 -7.14704022e-02
-4.64933306e-01 4.96099830e-01 3.47595930e-01 -1.03723037e+00
-3.07819128e-01 -2.37486288e-01 -1.76096007e-01 -1.96157932e-01
4.36274588e-01 -5.49022794e-01 -3.98356877e-02 6.11961186e-01
-5.41383266e-01 -7.13602901e-02 -2.54742876e-02 5.98044097e-01
-1.26855448e-01 -2.44277358e-01 1.07395995e+00 1.99729741e-01
3.33435973e-03 -1.56479239e-01 -4.46714759e-01 3.79410207e-01
2.78455406e-01 7.20408335e-02 -4.09003675e-01 -5.62551379e-01
-6.24291003e-01 -7.39480674e-01 3.36258411e-01 6.94759488e-01
2.54581034e-01 -1.09671688e+00 -9.30935800e-01 5.66988528e-01
-3.35324034e-02 -8.36224318e-01 8.99565518e-02 5.18327594e-01
-2.47322395e-01 5.90028584e-01 3.69710922e-02 -3.78918588e-01
-1.41286266e+00 7.78041124e-01 4.20630366e-01 1.47785351e-01
-1.48362339e-01 1.03689706e+00 -3.56294103e-02 -7.17459261e-01
3.43260825e-01 1.51254728e-01 2.09139481e-01 7.82497898e-02
6.21350944e-01 6.27917111e-01 1.94357540e-02 -1.22339380e+00
-4.81398523e-01 -2.19850227e-01 2.61229783e-01 -6.35190010e-01
1.02138042e+00 -6.31774738e-02 2.09093362e-01 7.58899093e-01
1.48482537e+00 5.26190877e-01 -7.77763367e-01 -4.09773499e-01
1.04064010e-01 -5.95301986e-01 2.27277242e-02 -4.48463887e-01
-5.12529850e-01 1.07282460e+00 6.84982777e-01 8.39864016e-01
6.93473339e-01 -1.12754107e-01 8.37516963e-01 9.89327803e-02
1.05248079e-01 -1.07723224e+00 1.83971196e-01 3.90276462e-01
8.68387878e-01 -1.00707412e+00 -4.16370124e-01 -1.28575638e-01
-6.97937131e-01 7.51161873e-01 -1.26759320e-01 1.87433213e-02
9.32971954e-01 8.77942219e-02 5.26369929e-01 8.81784558e-02
-3.74126196e-01 2.26651922e-01 8.42361748e-02 1.12720096e+00
5.74115098e-01 5.43307006e-01 1.05469070e-01 5.30135393e-01
-8.84423733e-01 -6.04527175e-01 2.95717001e-01 3.17283660e-01
-5.74412942e-01 -9.29970682e-01 -9.23150003e-01 2.33930543e-01
-1.02682161e+00 2.55558640e-02 -5.90619504e-01 2.78961420e-01
8.78835171e-02 1.68199754e+00 -6.61331937e-02 -9.52847779e-01
1.75250351e-01 5.43943226e-01 -8.28633681e-02 -5.62787473e-01
-7.41203189e-01 1.64901868e-01 4.10567820e-01 -5.11359200e-02
-3.56415570e-01 -8.99617493e-01 -6.58403337e-01 -7.15365887e-01
-6.80096567e-01 2.15537846e-01 9.15915728e-01 8.69655430e-01
1.06934592e-01 3.56768787e-01 9.71628487e-01 -7.99318194e-01
-1.00657797e+00 -1.37267148e+00 -6.64502263e-01 6.73966348e-01
1.01456630e+00 -4.34395343e-01 -8.99001777e-01 3.33788097e-02] | [14.078413963317871, 5.864184856414795] |
4846a4ad-de7f-48ee-af2d-984fdb6428ac | audio-driven-co-speech-gesture-video | 2212.0235 | null | https://arxiv.org/abs/2212.02350v1 | https://arxiv.org/pdf/2212.02350v1.pdf | Audio-Driven Co-Speech Gesture Video Generation | Co-speech gesture is crucial for human-machine interaction and digital entertainment. While previous works mostly map speech audio to human skeletons (e.g., 2D keypoints), directly generating speakers' gestures in the image domain remains unsolved. In this work, we formally define and study this challenging problem of audio-driven co-speech gesture video generation, i.e., using a unified framework to generate speaker image sequence driven by speech audio. Our key insight is that the co-speech gestures can be decomposed into common motion patterns and subtle rhythmic dynamics. To this end, we propose a novel framework, Audio-driveN Gesture vIdeo gEneration (ANGIE), to effectively capture the reusable co-speech gesture patterns as well as fine-grained rhythmic movements. To achieve high-fidelity image sequence generation, we leverage an unsupervised motion representation instead of a structural human body prior (e.g., 2D skeletons). Specifically, 1) we propose a vector quantized motion extractor (VQ-Motion Extractor) to summarize common co-speech gesture patterns from implicit motion representation to codebooks. 2) Moreover, a co-speech gesture GPT with motion refinement (Co-Speech GPT) is devised to complement the subtle prosodic motion details. Extensive experiments demonstrate that our framework renders realistic and vivid co-speech gesture video. Demo video and more resources can be found in: https://alvinliu0.github.io/projects/ANGIE | ['Ziwei Liu', 'Dahua Lin', 'Wayne Wu', 'Yuanqi Du', 'Hang Zhou', 'Qianyi Wu', 'Xian Liu'] | 2022-12-05 | null | null | null | null | ['video-generation'] | ['computer-vision'] | [ 3.15242767e-01 -1.40419900e-01 -2.20145881e-01 2.98188590e-02
-8.54953170e-01 -5.16271830e-01 8.66977274e-01 -8.24663758e-01
1.23571530e-01 1.92033753e-01 8.46169651e-01 -3.55730802e-02
1.16126053e-01 -3.72479171e-01 -5.64283252e-01 -8.17287445e-01
1.72457144e-01 2.88855489e-02 1.34518400e-01 -2.24124491e-01
5.57626225e-02 1.66110516e-01 -1.48391163e+00 1.97205782e-01
4.61604238e-01 4.43356097e-01 3.01549762e-01 1.09931791e+00
-5.95412403e-02 6.14494681e-01 -5.67195594e-01 -7.92596862e-02
1.49683490e-01 -8.49360943e-01 -6.18451715e-01 3.34901452e-01
2.36862749e-01 -5.81508458e-01 -5.17544806e-01 9.04159844e-01
4.98617381e-01 3.35637599e-01 7.21027434e-01 -1.46563220e+00
-4.02047753e-01 5.00296116e-01 -5.49265504e-01 -2.52065629e-01
7.46178389e-01 5.39325178e-01 9.18712914e-01 -9.56663132e-01
9.31789398e-01 1.37826717e+00 2.69956738e-01 9.54909801e-01
-9.34278905e-01 -8.87013376e-01 1.34440809e-01 4.84500453e-02
-1.51839256e+00 -6.12176955e-01 1.00635660e+00 -4.93724346e-01
4.53132570e-01 5.12796938e-01 9.41800773e-01 1.52957797e+00
-2.43104577e-01 1.06356537e+00 4.52671736e-01 -2.91679770e-01
6.09105043e-02 -7.33567774e-01 -4.39291567e-01 5.68588555e-01
-3.39642555e-01 1.67741835e-01 -8.99401903e-01 7.67288432e-02
1.31130028e+00 1.60170466e-01 -4.54858810e-01 -2.38851637e-01
-1.62901771e+00 6.12518191e-01 -7.44029507e-02 2.62082011e-01
-3.80433232e-01 5.40686011e-01 1.93597779e-01 -1.93355158e-02
-1.98947471e-02 -1.92522004e-01 1.11585028e-01 -5.31274319e-01
-1.11293674e+00 5.56927204e-01 5.29888093e-01 1.12462854e+00
4.75209951e-01 3.42687607e-01 -2.18056530e-01 6.40687346e-01
5.47643363e-01 7.15304673e-01 6.76631272e-01 -1.18961132e+00
4.84390855e-01 1.96405649e-01 8.97332951e-02 -9.91457343e-01
-6.33785948e-02 2.21376553e-01 -1.03692091e+00 5.85118169e-03
3.38739127e-01 -2.29023546e-01 -9.06988204e-01 1.65533400e+00
6.21276498e-01 7.17515469e-01 5.67862689e-02 1.29672027e+00
9.47111845e-01 8.28260601e-01 3.47679034e-02 -2.16786236e-01
1.42097139e+00 -1.05212593e+00 -8.49162579e-01 2.72025108e-01
2.69936740e-01 -7.58854985e-01 1.23698819e+00 2.86405116e-01
-1.17481542e+00 -6.01850748e-01 -7.53130853e-01 -9.69358720e-03
4.24353719e-01 -6.32849559e-02 2.59225249e-01 3.61122668e-01
-7.81182885e-01 2.04993263e-01 -1.24944615e+00 -2.99399644e-01
7.37785250e-02 1.30354658e-01 -2.38655582e-01 2.33246967e-01
-1.01885891e+00 -8.19423124e-02 1.47732154e-01 1.05377071e-01
-1.06349492e+00 -4.82716084e-01 -1.14175797e+00 -2.19588533e-01
5.58983326e-01 -8.75197649e-01 1.64941204e+00 -8.79999518e-01
-2.08374238e+00 4.87329364e-01 -4.83177930e-01 -1.75190449e-01
7.04458892e-01 -3.93637270e-01 -1.96802825e-01 5.80336273e-01
-6.73350990e-02 9.14224267e-01 1.18060780e+00 -1.27150691e+00
-6.64328396e-01 6.46847039e-02 -3.63691777e-01 4.52193022e-01
-2.12421492e-01 1.78288609e-01 -9.58665252e-01 -1.41071987e+00
2.53393829e-01 -1.08455288e+00 -1.03976332e-01 -1.00927830e-01
-6.10579967e-01 -1.11760959e-01 9.44410264e-01 -7.71359324e-01
1.48951507e+00 -2.03933024e+00 5.01751840e-01 2.01084286e-01
2.06522152e-01 2.76544858e-02 -2.23382682e-01 4.64706719e-01
9.63147581e-02 7.03458488e-02 -3.78526032e-01 -3.97069395e-01
1.69980168e-01 2.68395782e-01 -4.13117111e-01 1.58149302e-01
9.70449671e-02 1.23219848e+00 -1.03579414e+00 -5.58120608e-01
3.82262051e-01 8.04630458e-01 -6.53845608e-01 2.20629320e-01
-1.52834073e-01 1.10302246e+00 -7.62557864e-01 6.91175103e-01
2.77636409e-01 -2.51992303e-03 1.92393169e-01 -1.92317560e-01
-9.49520245e-02 1.77763328e-01 -1.33689654e+00 2.10233831e+00
-2.86702096e-01 4.42364752e-01 2.49798551e-01 -5.93048573e-01
6.95701718e-01 6.81475222e-01 7.28566229e-01 -4.55647796e-01
1.87640842e-02 -6.29427703e-03 -3.12261879e-01 -7.27349520e-01
3.88002187e-01 -1.63333848e-01 -2.72578686e-01 5.83404243e-01
-1.24646321e-01 -2.67589092e-01 -9.16195437e-02 1.23937123e-01
9.14726794e-01 4.63745922e-01 2.56877065e-01 2.80260533e-01
4.40113634e-01 -1.02096424e-01 5.52901447e-01 2.70465314e-01
-5.98607287e-02 1.09767246e+00 9.53527838e-02 -1.08231016e-01
-9.09645617e-01 -1.18466890e+00 5.34491956e-01 9.48692143e-01
2.70531893e-01 -6.80549800e-01 -9.60060358e-01 -3.01481545e-01
-3.52942646e-01 1.73697785e-01 -2.83467412e-01 2.41977021e-01
-1.11945343e+00 -2.15720236e-01 7.45594025e-01 6.48558199e-01
3.53995919e-01 -1.41790986e+00 -7.89535105e-01 2.46641800e-01
-5.52554250e-01 -1.12648821e+00 -1.25163293e+00 -5.84564030e-01
-5.61057866e-01 -7.67249107e-01 -1.30455315e+00 -8.42005908e-01
2.77183354e-01 3.53589416e-01 7.21211851e-01 -2.51548551e-02
-3.88868630e-01 7.87313640e-01 -6.62730396e-01 -1.38809169e-02
-3.99912179e-01 -2.12091580e-01 1.56943098e-01 2.51130402e-01
1.05502933e-01 -8.43148053e-01 -9.02892172e-01 4.02665228e-01
-1.03610122e+00 5.15942931e-01 4.71713483e-01 8.30529749e-01
7.59664476e-01 -3.21560919e-01 2.02552423e-01 -1.46043912e-01
5.59900343e-01 -2.79908985e-01 -1.78059161e-01 -1.54046372e-01
1.75763980e-01 -1.29912734e-01 2.77686387e-01 -8.15317035e-01
-1.01465845e+00 3.46731693e-01 -2.45164528e-01 -9.34315026e-01
-4.27298039e-01 2.24208564e-01 -4.29620683e-01 3.54977965e-01
1.66290089e-01 6.09695196e-01 -3.03681064e-02 -4.31380779e-01
6.58370495e-01 6.86450541e-01 1.02217400e+00 -7.22047091e-01
1.07186067e+00 5.24729848e-01 -2.20645860e-01 -1.26047111e+00
-7.79218227e-02 -6.12825155e-01 -6.25663698e-01 -4.35494125e-01
1.23140776e+00 -9.37771142e-01 -7.60578096e-01 5.46446502e-01
-1.12953055e+00 -8.11021566e-01 -1.45881966e-01 6.18139505e-01
-9.14286375e-01 6.81991100e-01 -7.76415408e-01 -8.68984640e-01
-3.73262346e-01 -1.20297456e+00 1.47702694e+00 1.58112019e-01
-6.65623963e-01 -6.49951935e-01 1.79939002e-01 3.50027740e-01
-2.38451883e-01 5.72124362e-01 2.59364694e-01 1.03895299e-01
-7.58718133e-01 2.63568342e-01 2.25613311e-01 -5.19097559e-02
3.15350890e-01 8.34139436e-02 -5.97164869e-01 -5.99632189e-02
-2.61136383e-01 -5.16775176e-02 6.50008082e-01 5.09294868e-01
8.28323722e-01 -5.63164949e-01 -1.97351307e-01 7.76816845e-01
7.12368548e-01 3.73245388e-01 6.04693055e-01 -4.95685413e-02
1.06272721e+00 7.03775525e-01 7.31481493e-01 6.42124176e-01
4.41227049e-01 1.16520727e+00 7.86876455e-02 2.30980217e-01
-5.82631946e-01 -7.62109041e-01 7.22515404e-01 1.24057662e+00
-4.98356998e-01 -5.38658574e-02 -8.22683930e-01 5.25267541e-01
-1.95071232e+00 -1.13086092e+00 -8.85200128e-02 1.83806193e+00
6.47006750e-01 -2.50055015e-01 5.44596434e-01 2.83324778e-01
7.79402435e-01 2.65266806e-01 -4.06999052e-01 1.45463586e-01
7.82873780e-02 2.31872365e-01 -4.37229872e-02 5.90447247e-01
-1.03546774e+00 1.13666987e+00 4.56923342e+00 1.02602422e+00
-1.28688478e+00 1.85316920e-01 5.84481321e-02 -3.42365354e-01
-4.70037580e-01 -2.09330887e-01 -5.20024240e-01 5.18524826e-01
4.89967287e-01 -1.18467964e-01 3.31722230e-01 4.80650812e-01
7.62533307e-01 5.13218045e-01 -7.87482977e-01 1.37171495e+00
1.45665964e-03 -1.31197250e+00 4.60314095e-01 1.40701041e-01
6.22849286e-01 -4.88204658e-01 8.07246566e-02 7.03393146e-02
2.37008184e-01 -9.36162531e-01 1.20373392e+00 3.92105848e-01
1.20602536e+00 -6.63079739e-01 -9.44350660e-02 3.28141749e-01
-1.85801911e+00 1.55738011e-01 2.92106748e-01 1.55782282e-01
7.21430957e-01 -8.30203071e-02 -5.37525177e-01 6.71797752e-01
5.97784638e-01 8.27913880e-01 1.41658112e-01 6.67091072e-01
-4.81053919e-01 7.81623006e-01 -3.03731322e-01 1.00101925e-01
3.20278913e-01 -1.06932350e-01 9.39947188e-01 1.36265993e+00
6.84648991e-01 5.43876708e-01 1.97652236e-01 7.15497077e-01
2.47381270e-01 2.19930764e-02 -3.85002792e-01 -7.40862489e-02
5.44806957e-01 1.06711054e+00 -7.06190526e-01 -2.63370603e-01
-3.32445771e-01 1.35224247e+00 -4.78418380e-01 6.28933489e-01
-8.33541572e-01 -1.62848428e-01 9.85601306e-01 1.51980102e-01
4.47198719e-01 -5.73230207e-01 1.79756194e-01 -1.30571198e+00
9.66273621e-02 -9.12785053e-01 1.37771159e-01 -6.86630905e-01
-8.08649123e-01 4.10002351e-01 2.88483147e-02 -1.66859722e+00
-7.46850073e-01 -2.98601687e-01 -8.60421002e-01 5.29102623e-01
-1.06686878e+00 -1.39906180e+00 -5.48449278e-01 9.68708992e-01
1.08365619e+00 1.04267143e-01 5.44218600e-01 2.53216505e-01
-4.20497507e-01 6.72943115e-01 -3.55100095e-01 4.36639339e-01
5.87322652e-01 -8.06435227e-01 6.28010154e-01 7.98344851e-01
5.62762916e-01 5.76802552e-01 5.40639043e-01 -7.41676271e-01
-1.55267930e+00 -1.21080172e+00 4.32389617e-01 -4.17803854e-01
5.74794471e-01 -3.36693347e-01 -6.57567382e-01 6.01261199e-01
-5.42000160e-02 -2.23217085e-01 5.37640989e-01 -7.69630671e-01
-1.14238910e-01 3.97308677e-01 -5.08802116e-01 1.03830600e+00
1.53240848e+00 -5.54560363e-01 -5.71934342e-01 -2.08328843e-01
8.42310071e-01 -5.48819184e-01 -5.97019553e-01 3.26545626e-01
9.08060491e-01 -7.12227404e-01 1.16788363e+00 -2.86325574e-01
4.89659578e-01 -6.18698776e-01 -1.07895195e-01 -9.19154942e-01
1.08446792e-01 -1.42551661e+00 -5.50873220e-01 1.23675346e+00
-1.31764367e-01 -4.79118451e-02 9.97757435e-01 2.05102727e-01
-2.16395408e-01 -5.85264921e-01 -9.11745787e-01 -6.14436448e-01
-1.64090797e-01 -7.46016085e-01 6.01863563e-01 9.09202397e-01
8.67300332e-02 1.25006452e-01 -7.40816593e-01 2.12175012e-01
6.01433873e-01 1.26342967e-01 1.23903358e+00 -7.51958013e-01
-4.46695060e-01 -5.95737159e-01 -3.75531942e-01 -1.68405068e+00
1.32478699e-01 -7.01826990e-01 3.19676429e-01 -1.34884977e+00
-2.62447800e-02 -9.80385765e-02 2.20291629e-01 2.95536429e-01
-1.58079684e-01 3.96450341e-01 5.02420247e-01 5.71851969e-01
-4.15766895e-01 8.32112312e-01 1.64584112e+00 5.57126813e-02
-6.13172412e-01 -1.46590516e-01 -2.81105399e-01 7.95301855e-01
5.57496905e-01 -2.09962934e-01 -5.35987556e-01 -3.00251037e-01
-2.77507156e-01 6.37658656e-01 7.17853844e-01 -9.34687793e-01
3.26748610e-01 -3.74862283e-01 3.30691859e-02 -5.91573060e-01
5.04196882e-01 -4.01151776e-01 3.74617398e-01 5.09601176e-01
-2.21415401e-01 -7.31015354e-02 -1.48731813e-01 6.17885053e-01
-4.08119589e-01 2.00037792e-01 1.67370975e-01 -1.22061178e-01
-7.28102207e-01 4.92759019e-01 -5.37988663e-01 -3.76256704e-02
8.23422074e-01 -4.49697465e-01 1.98381305e-01 -8.24315965e-01
-1.04316521e+00 2.12008163e-01 3.04100513e-01 6.97593987e-01
9.20825839e-01 -1.70734322e+00 -7.39402711e-01 1.87854990e-01
-3.51028331e-02 2.74993151e-01 5.92457891e-01 8.49646270e-01
-4.79701966e-01 2.87632853e-01 -4.43398133e-02 -7.44288504e-01
-1.55293500e+00 1.86822295e-01 -6.99048489e-02 1.04845531e-01
-1.18348086e+00 7.26867199e-01 5.57401955e-01 -1.60066769e-01
1.91879526e-01 -4.76968914e-01 -6.86625019e-02 -1.47986248e-01
7.54008055e-01 3.04586232e-01 -6.39059246e-01 -1.07559514e+00
-1.07428834e-01 1.02621114e+00 4.74542350e-01 -7.40464628e-01
1.04130840e+00 -2.89849937e-01 4.24572974e-01 3.58449340e-01
1.05482912e+00 2.01918781e-01 -1.71140826e+00 5.28564826e-02
-1.50478438e-01 -3.25636685e-01 -3.87743711e-01 -1.38237968e-01
-1.03128886e+00 1.03684473e+00 1.74534887e-01 -3.22451055e-01
1.18710053e+00 6.79313987e-02 1.27611625e+00 1.11189850e-01
2.82088697e-01 -6.76656485e-01 5.20774007e-01 2.70952523e-01
1.09705043e+00 -6.38610244e-01 -4.71011400e-01 -5.50615191e-01
-9.29521024e-01 1.13598716e+00 4.23803031e-01 -2.62474958e-02
5.09649932e-01 3.04582328e-01 2.48542979e-01 -7.69204870e-02
-4.75331217e-01 -2.97480822e-01 4.44620878e-01 6.83245957e-01
2.75984824e-01 1.44147560e-01 -1.58559531e-01 8.43433261e-01
-6.12598181e-01 1.19318292e-01 2.88599998e-01 8.61770213e-01
-1.14418156e-01 -1.04341555e+00 -6.03631616e-01 -2.61161715e-01
-1.30696163e-01 -1.59428976e-02 -3.58972877e-01 7.29967892e-01
1.12448623e-02 9.11280215e-01 7.07189292e-02 -6.24088466e-01
2.23125488e-01 6.65082857e-02 4.74483967e-01 -5.79186976e-01
-2.93166816e-01 9.07543123e-01 -2.34920874e-01 -7.95489073e-01
-6.12840414e-01 -8.32812190e-01 -1.57972240e+00 -6.54492974e-02
2.84367502e-01 -2.50015431e-03 3.30586851e-01 6.73212469e-01
2.26632088e-01 5.76484501e-01 2.91702867e-01 -1.52765191e+00
-1.57511741e-01 -7.99687386e-01 -3.76339525e-01 6.55808032e-01
4.64070201e-01 -6.20156109e-01 -2.38074660e-01 7.16984034e-01] | [5.743698596954346, -0.1982276886701584] |
5b4f2bf7-d80c-415e-8f52-5ef8e1011b02 | semi-supervised-clustering-via-dynamic-graph | 2209.02513 | null | https://arxiv.org/abs/2209.02513v1 | https://arxiv.org/pdf/2209.02513v1.pdf | Semi-Supervised Clustering via Dynamic Graph Structure Learning | Most existing semi-supervised graph-based clustering methods exploit the supervisory information by either refining the affinity matrix or directly constraining the low-dimensional representations of data points. The affinity matrix represents the graph structure and is vital to the performance of semi-supervised graph-based clustering. However, existing methods adopt a static affinity matrix to learn the low-dimensional representations of data points and do not optimize the affinity matrix during the learning process. In this paper, we propose a novel dynamic graph structure learning method for semi-supervised clustering. In this method, we simultaneously optimize the affinity matrix and the low-dimensional representations of data points by leveraging the given pairwise constraints. Moreover, we propose an alternating minimization approach with proven convergence to solve the proposed nonconvex model. During the iteration process, our method cyclically updates the low-dimensional representations of data points and refines the affinity matrix, leading to a dynamic affinity matrix (graph structure). Specifically, for the update of the affinity matrix, we enforce the data points with remarkably different low-dimensional representations to have an affinity value of 0. Furthermore, we construct the initial affinity matrix by integrating the local distance and global self-representation among data points. Experimental results on eight benchmark datasets under different settings show the advantages of the proposed approach. | ['Zuoqiang Shi', 'Xin Liang', 'Chenglong Bao', 'Huaming Ling'] | 2022-09-06 | null | null | null | null | ['graph-structure-learning'] | ['graphs'] | [-2.08062604e-01 3.91810201e-02 -4.06372488e-01 -4.67367470e-01
-4.40614432e-01 -5.14107764e-01 5.46802320e-02 2.87064701e-01
-1.92289904e-01 2.41494223e-01 1.88708022e-01 1.98174670e-01
-7.14260161e-01 -5.52752495e-01 -5.42768836e-01 -1.07023311e+00
6.63045198e-02 7.73911059e-01 -3.44239548e-02 9.09060463e-02
3.07515800e-01 3.81283849e-01 -1.08774400e+00 -2.20309660e-01
1.23528624e+00 7.74230778e-01 2.31505901e-01 1.61070481e-01
1.51148839e-02 2.02925593e-01 -2.12495163e-01 2.12892339e-01
4.21215922e-01 -4.30785418e-01 -6.22026742e-01 8.05098653e-01
1.39223337e-01 2.85347819e-01 -3.92838806e-01 1.26835835e+00
2.91175455e-01 4.13229823e-01 6.33530557e-01 -1.23234749e+00
-7.64936924e-01 4.30886269e-01 -1.00955224e+00 -5.78619866e-03
8.63256603e-02 -1.63116857e-01 1.08341503e+00 -9.19980049e-01
5.16232789e-01 9.26035285e-01 3.80504996e-01 1.56614408e-01
-1.58674908e+00 -4.59131658e-01 4.81977820e-01 1.85295761e-01
-1.93225932e+00 -9.42791849e-02 1.26967144e+00 -5.31092107e-01
2.15460733e-01 8.63805041e-02 7.37304807e-01 3.54888946e-01
-2.25787058e-01 4.74401444e-01 8.44472647e-01 -1.76672548e-01
4.59944040e-01 -4.65609357e-02 4.30842787e-01 8.86781991e-01
3.55015785e-01 -3.45735848e-01 -2.93913305e-01 -4.01789725e-01
7.21559405e-01 2.53918320e-01 -3.87955546e-01 -1.15572143e+00
-1.24814665e+00 9.18688953e-01 6.21658921e-01 2.86609858e-01
-3.81648123e-01 -2.01014742e-01 1.07155167e-01 3.77399400e-02
2.88795084e-01 1.49172574e-01 -9.95541811e-02 2.32984722e-01
-7.72993505e-01 -1.70329496e-01 5.62280059e-01 9.81529176e-01
1.33287978e+00 -2.06493348e-01 7.36533031e-02 9.54415679e-01
6.38354301e-01 2.34092936e-01 5.35052419e-01 -7.97527075e-01
5.76305687e-01 1.19668949e+00 -1.20065799e-02 -1.58158612e+00
-4.71719742e-01 -2.78436035e-01 -1.12914109e+00 -2.67734021e-01
3.57849509e-01 -2.27268994e-01 -9.54589069e-01 1.49500203e+00
7.20369995e-01 4.35772151e-01 -4.12937440e-02 1.29614496e+00
4.89000678e-01 6.00740194e-01 -3.15918237e-01 -5.07334054e-01
6.56475842e-01 -8.11682522e-01 -7.64648736e-01 6.87487125e-02
7.56043196e-01 -4.88646984e-01 1.09260285e+00 1.60281789e-02
-9.10652041e-01 -3.47503752e-01 -1.02975786e+00 1.92901880e-01
-4.03782874e-02 2.68158853e-01 4.29181159e-01 3.28302950e-01
-7.34869182e-01 3.80714238e-01 -9.58327353e-01 -9.92934927e-02
1.81875303e-01 7.74158478e-01 -2.44973749e-01 -9.38786485e-04
-8.48518074e-01 1.21419899e-01 5.95219970e-01 3.31697971e-01
-4.62016225e-01 -5.78945756e-01 -7.79274940e-01 -5.83988093e-02
5.64526975e-01 -3.83406043e-01 2.53117591e-01 -8.44140589e-01
-1.43134582e+00 7.56187558e-01 -1.30187348e-01 2.43492611e-02
2.91711569e-01 4.47462164e-02 -1.13267012e-01 1.72386959e-01
5.39641120e-02 3.04286122e-01 1.00922239e+00 -1.57970011e+00
-3.55917454e-01 -5.14698684e-01 -1.34025335e-01 5.25097311e-01
-6.83708251e-01 -5.78319430e-01 -9.37876225e-01 -4.98669952e-01
7.75217831e-01 -1.17005706e+00 -5.05615413e-01 -2.93201625e-01
-5.74109614e-01 -1.92614958e-01 9.35380995e-01 -2.09857538e-01
1.38154984e+00 -2.43592644e+00 8.57690632e-01 9.65298355e-01
4.30373460e-01 -3.38439569e-02 -8.35670978e-02 1.97020859e-01
-6.50104210e-02 -1.19655125e-01 -4.88347828e-01 -2.82727391e-01
-1.44631580e-01 5.27881205e-01 2.15643689e-01 7.97367513e-01
-9.33996364e-02 6.32307351e-01 -1.08355498e+00 -6.32245600e-01
3.15002024e-01 5.56936622e-01 -7.38013387e-01 2.70776749e-01
1.98187545e-01 5.00267923e-01 -7.36648381e-01 3.19708556e-01
8.12246323e-01 -4.96708810e-01 6.31255746e-01 -5.27884066e-01
4.56219912e-02 -2.51084417e-01 -1.74693596e+00 1.96336699e+00
-1.20252214e-01 -4.31700284e-03 2.90322155e-01 -1.24296033e+00
1.12463856e+00 -5.84294386e-02 1.03403449e+00 -1.06670447e-01
1.82935044e-01 6.95616156e-02 -7.85286538e-03 -7.66613781e-02
2.36682862e-01 1.26276642e-01 1.58849910e-01 4.01768386e-01
-1.61482349e-01 1.14073165e-01 3.23910475e-01 3.61080974e-01
5.39922237e-01 -1.62300795e-01 -6.61641010e-04 -5.11670649e-01
9.19282377e-01 -1.05952121e-01 7.88788795e-01 1.93194821e-01
8.97509903e-02 7.46448398e-01 4.87734020e-01 -2.56358683e-01
-6.91760778e-01 -1.00489342e+00 -1.10330664e-01 7.45860040e-01
5.67829669e-01 -5.97102344e-01 -8.90174508e-01 -8.98404956e-01
6.52603060e-02 9.98874083e-02 -7.05543399e-01 -4.09970403e-01
-5.69473624e-01 -9.98257995e-01 -1.72316536e-01 2.35266164e-01
3.19063574e-01 -3.94475758e-01 1.26682839e-03 3.14814821e-02
-7.53843561e-02 -7.16152370e-01 -1.01763535e+00 1.53673589e-01
-9.10418987e-01 -1.18217659e+00 -5.60304880e-01 -1.12863314e+00
1.58583546e+00 4.98675793e-01 5.20763516e-01 3.63807410e-01
-4.89395577e-03 4.28799301e-01 -3.20694685e-01 4.14756954e-01
2.24905815e-02 2.66604692e-01 1.67688206e-01 6.50472820e-01
4.37787771e-02 -6.31963789e-01 -6.52931690e-01 7.88101614e-01
-8.67383361e-01 -1.25762001e-01 4.45651501e-01 8.58029127e-01
1.27588904e+00 3.46739978e-01 5.35695672e-01 -9.97828186e-01
5.90671062e-01 -5.65073669e-01 -6.75280511e-01 2.27734640e-01
-7.99348354e-01 1.31785095e-01 6.55507386e-01 -6.02675796e-01
-6.98350132e-01 5.98126233e-01 3.71067286e-01 -8.84239197e-01
2.47371957e-01 7.15139866e-01 -4.48952049e-01 -2.11560875e-01
3.40195417e-01 1.80032715e-01 1.05507955e-01 -4.15583491e-01
5.67818642e-01 4.01773036e-01 3.99600416e-01 -7.06484556e-01
1.08324277e+00 6.47948980e-01 5.52606769e-02 -5.88024378e-01
-8.16880465e-01 -8.66798639e-01 -1.05178320e+00 -1.27070308e-01
7.28536248e-01 -7.71296382e-01 -6.30012512e-01 1.49995118e-01
-4.07927394e-01 -1.17847465e-01 -2.94400305e-01 5.95097184e-01
-4.85196203e-01 7.53324628e-01 -2.58974642e-01 -3.75863135e-01
-6.41583949e-02 -1.07986212e+00 8.94039690e-01 2.98118889e-02
-6.96211262e-03 -1.30676055e+00 2.60901093e-01 2.61838049e-01
-1.96974814e-01 3.36660415e-01 7.58863151e-01 -4.22840714e-01
-3.34503710e-01 -2.53753781e-01 -2.36606807e-01 2.35215724e-02
6.55322194e-01 -6.04023822e-02 -1.75278649e-01 -5.58697045e-01
-6.56266436e-02 -5.93097322e-02 5.82030892e-01 4.55408752e-01
1.21967614e+00 -4.15936023e-01 -4.00469005e-01 8.87703240e-01
1.35315919e+00 3.54268923e-02 2.05661386e-01 1.30296960e-01
1.33714402e+00 6.67931020e-01 6.59278035e-01 6.52356207e-01
4.83635634e-01 5.54406226e-01 3.62151325e-01 -7.84749687e-02
3.24438274e-01 -3.27221662e-01 -3.12997736e-02 1.29523158e+00
-1.16125412e-01 1.60864577e-01 -9.17429447e-01 4.81165767e-01
-2.30636573e+00 -4.54270601e-01 -3.74178559e-01 2.37669325e+00
7.04829454e-01 -3.03196367e-02 2.10472062e-01 1.77518100e-01
1.01613450e+00 1.54129148e-01 -7.93739855e-01 1.70288265e-01
-6.04380667e-02 -4.17493135e-01 4.70391810e-01 6.72045767e-01
-1.05383229e+00 7.73586392e-01 5.77361584e+00 5.53513587e-01
-1.04070735e+00 -2.42869228e-01 3.75717193e-01 -1.34846926e-01
-3.60890925e-01 3.69954072e-02 -6.51907802e-01 5.32617450e-01
3.77471268e-01 -3.43926966e-01 6.78946733e-01 8.16928923e-01
4.66263354e-01 1.84974283e-01 -9.72602427e-01 1.10537422e+00
9.33042169e-02 -1.22154188e+00 2.13695720e-01 2.75353163e-01
1.12043762e+00 -2.56341159e-01 -1.35254785e-02 -1.83238357e-01
3.66695195e-01 -5.34760475e-01 3.56252104e-01 3.83215547e-01
4.87427115e-01 -1.08123887e+00 4.59844202e-01 3.12083304e-01
-1.39450860e+00 1.58912223e-02 -6.04234278e-01 2.91727304e-01
2.40470525e-02 7.39445686e-01 -5.35497367e-01 5.84059417e-01
4.47883695e-01 1.28546774e+00 -5.97923100e-01 8.04660022e-01
-1.34870177e-02 3.90510529e-01 -4.22953725e-01 2.32670292e-01
3.69923353e-01 -9.91622150e-01 5.06927907e-01 8.65144908e-01
-4.07928899e-02 3.16408664e-01 7.24223256e-01 7.90285587e-01
-4.23965901e-02 4.21453476e-01 -3.42904866e-01 -4.65370603e-02
4.53526914e-01 1.35049295e+00 -8.88916373e-01 -9.26569328e-02
-2.71556556e-01 8.97548616e-01 6.58221364e-01 5.55633485e-01
-6.88272834e-01 -1.95697799e-01 4.89389837e-01 1.16184987e-01
1.70265108e-01 -5.56814015e-01 -3.05715680e-01 -1.16107440e+00
5.32326661e-02 -6.43439114e-01 7.14508891e-01 -1.29658669e-01
-1.21251619e+00 2.85453647e-01 -3.21970247e-02 -1.22899938e+00
-4.26856391e-02 -1.02457292e-01 -5.57042301e-01 4.99766350e-01
-1.24690652e+00 -9.13843930e-01 -4.25754547e-01 1.02000558e+00
1.80558667e-01 1.12012491e-01 3.12697619e-01 2.56430775e-01
-8.37332368e-01 3.91837448e-01 4.46525007e-01 1.29929647e-01
5.78379035e-01 -1.30175447e+00 -3.15847963e-01 5.53091645e-01
9.23409462e-02 8.56714547e-01 3.50688279e-01 -7.39130259e-01
-1.71183252e+00 -1.17242932e+00 2.14010775e-01 -4.99246642e-02
7.83084095e-01 -3.02889824e-01 -1.20754313e+00 6.64687216e-01
-2.90406287e-01 2.38168716e-01 8.83365035e-01 1.26339421e-01
-1.01924367e-01 -1.99879438e-01 -1.00303912e+00 3.94207478e-01
9.44170296e-01 -3.81420583e-01 -2.68598735e-01 6.70202196e-01
5.12168050e-01 -4.14364427e-01 -1.13790500e+00 2.98725665e-01
2.26357967e-01 -6.28756225e-01 8.97435367e-01 -5.13374150e-01
-9.74551588e-02 -8.35933983e-01 -8.23893473e-02 -1.39429975e+00
-6.52727664e-01 -5.44283390e-01 -1.79261178e-01 1.27818429e+00
2.57552743e-01 -6.01506591e-01 1.17458701e+00 7.10105896e-01
-4.10485081e-02 -9.53371227e-01 -7.64686108e-01 -6.60939395e-01
-2.66660035e-01 5.27201593e-02 4.53520030e-01 1.30929101e+00
1.45809099e-01 3.49047124e-01 -3.47672164e-01 5.56285143e-01
9.76503015e-01 4.17613894e-01 8.16585958e-01 -1.17746758e+00
-1.39525577e-01 -3.22808713e-01 -5.00788033e-01 -1.18814290e+00
3.65767717e-01 -1.01477754e+00 -5.96907735e-02 -1.49067688e+00
2.63168156e-01 -7.09458292e-01 -4.54433382e-01 2.76402533e-01
-3.01193655e-01 1.70149878e-02 4.98060174e-02 6.27351582e-01
-7.97691643e-01 7.40602791e-01 1.29172814e+00 -8.21301341e-02
-7.72256136e-01 -2.61613578e-01 -5.86595178e-01 7.63472021e-01
6.42632902e-01 -4.56114560e-01 -8.05046856e-01 -4.22257185e-01
3.81038897e-02 -2.37590730e-01 -9.74577516e-02 -6.40068054e-01
4.27174598e-01 -4.36418265e-01 2.27117702e-01 -5.38507164e-01
1.73400745e-01 -1.13280892e+00 2.06188589e-01 3.30345124e-01
-3.68223459e-01 -2.29079962e-01 -3.54016513e-01 1.03092647e+00
-2.57771671e-01 -9.01201218e-02 8.99074435e-01 2.03012243e-01
-3.73450190e-01 6.46390319e-01 2.30402648e-02 9.61405635e-02
1.31149817e+00 -3.76487583e-01 3.18188399e-01 -1.54871330e-01
-1.10854852e+00 7.44792700e-01 7.35028565e-01 3.01587611e-01
7.74597585e-01 -1.61677182e+00 -5.30084670e-01 3.57781082e-01
1.16481960e-01 4.66297448e-01 6.73086941e-02 9.60780084e-01
-2.19347075e-01 7.88808540e-02 -3.08908639e-03 -8.80501032e-01
-1.10301769e+00 7.15227544e-01 2.82826841e-01 -5.55005409e-02
-5.61740279e-01 4.22364622e-01 3.06408435e-01 -5.23849308e-01
1.82104170e-01 -5.23699299e-02 -3.65536243e-01 2.62095720e-01
4.03583981e-02 5.38003445e-01 -2.60921925e-01 -9.61532116e-01
-3.80899459e-01 1.02159643e+00 -1.97336644e-01 2.49935642e-01
1.32464671e+00 -4.97909248e-01 -2.77941346e-01 5.12420952e-01
1.51019335e+00 -3.86806242e-02 -1.32133269e+00 -3.15933734e-01
-1.47516765e-02 -7.42339969e-01 1.47526979e-01 9.21065062e-02
-1.51910436e+00 4.18757796e-01 2.70448297e-01 1.01821601e-01
1.14451599e+00 4.97761816e-02 5.30675828e-01 4.26456064e-01
2.24935755e-01 -1.29707277e+00 3.34539831e-01 1.42720699e-01
6.54509842e-01 -1.09118652e+00 2.34549329e-01 -7.33165681e-01
-7.29106486e-01 8.81489277e-01 6.70524240e-01 -4.16129559e-01
8.67932916e-01 -2.96687841e-01 -4.43560667e-02 -4.53213423e-01
-3.50904912e-01 6.46458045e-02 5.47443748e-01 4.09760237e-01
1.62386835e-01 5.16936220e-02 -3.99275929e-01 2.63269037e-01
7.02867135e-02 -3.76524657e-01 3.20326298e-01 6.82609558e-01
-3.12210917e-01 -9.05497551e-01 -5.05069196e-01 4.19265866e-01
1.59167141e-01 2.73075849e-01 -6.48534596e-01 4.56642658e-01
-5.58872297e-02 9.73465681e-01 -6.68631820e-03 -3.61854285e-01
3.73982966e-01 -2.92408347e-01 5.89125380e-02 -6.76826060e-01
-2.04381809e-01 3.48443985e-01 -4.61229354e-01 -4.71562564e-01
-5.37454247e-01 -7.76481152e-01 -1.75397575e+00 -7.96049535e-02
-5.59662938e-01 6.44102931e-01 4.55375522e-01 6.66885555e-01
5.33823013e-01 2.73045242e-01 1.14680326e+00 -7.90751338e-01
-2.31131539e-01 -4.07807380e-01 -7.97734559e-01 5.83605468e-01
9.75359455e-02 -7.96613276e-01 -5.53198040e-01 9.98583138e-02] | [7.769883632659912, 4.52731466293335] |
1b9ed547-b907-4d0b-ae18-155975800d2c | safeguarding-data-in-multimodal-ai-a | 2306.08173 | null | https://arxiv.org/abs/2306.08173v1 | https://arxiv.org/pdf/2306.08173v1.pdf | Safeguarding Data in Multimodal AI: A Differentially Private Approach to CLIP Training | The surge in multimodal AI's success has sparked concerns over data privacy in vision-and-language tasks. While CLIP has revolutionized multimodal learning through joint training on images and text, its potential to unintentionally disclose sensitive information necessitates the integration of privacy-preserving mechanisms. We introduce a differentially private adaptation of the Contrastive Language-Image Pretraining (CLIP) model that effectively addresses privacy concerns while retaining accuracy. Our proposed method, Dp-CLIP, is rigorously evaluated on benchmark datasets encompassing diverse vision-and-language tasks such as image classification and visual question answering. We demonstrate that our approach retains performance on par with the standard non-private CLIP model. Furthermore, we analyze our proposed algorithm under linear representation settings. We derive the convergence rate of our algorithm and show a trade-off between utility and privacy when gradients are clipped per-batch and the loss function does not satisfy smoothness conditions assumed in the literature for the analysis of DP-SGD. | ['Wanrong Zhang', 'Linjun Zhang', 'Ryumei Nakada', 'Peihan Liu', 'Alyssa Huang'] | 2023-06-13 | null | null | null | null | ['visual-question-answering-1', 'question-answering'] | ['computer-vision', 'natural-language-processing'] | [ 5.22253990e-01 1.99862927e-01 -1.81663454e-01 -7.22536683e-01
-1.38439739e+00 -8.99052680e-01 6.16923094e-01 2.50369668e-01
-8.23664904e-01 6.05473816e-01 1.74990982e-01 -4.01706398e-01
1.99674755e-01 -1.36950582e-01 -1.03605938e+00 -9.06248271e-01
1.39226586e-01 -6.73571229e-02 -4.44488108e-01 6.01986229e-01
-7.63979629e-02 3.43339801e-01 -1.19706392e+00 2.69568533e-01
7.63989508e-01 1.24194241e+00 -4.34792489e-01 7.15069413e-01
3.28210741e-01 8.99512291e-01 -2.71994233e-01 -1.07965827e+00
7.37098396e-01 -3.71910185e-01 -7.79269457e-01 2.00678170e-01
1.03807569e+00 -6.32862210e-01 -6.28521979e-01 1.28528047e+00
5.16194165e-01 -8.25313628e-02 5.09822190e-01 -1.63861001e+00
-1.01473880e+00 1.44783780e-01 -7.10126758e-01 -1.66056901e-01
3.80872265e-02 1.59169346e-01 1.23580742e+00 -6.17742896e-01
5.98400950e-01 1.20598292e+00 5.81490636e-01 6.83848917e-01
-1.49180400e+00 -5.69727778e-01 1.68848112e-02 -1.08895048e-01
-1.34931850e+00 -6.63746715e-01 4.56462473e-01 -4.23823148e-01
4.20303196e-01 6.11645043e-01 8.36012587e-02 1.17849469e+00
-1.34176444e-02 1.24929845e+00 1.10321128e+00 -2.86631614e-01
2.74251789e-01 5.20278454e-01 -7.29755545e-03 7.29566932e-01
2.60794163e-01 -8.48325118e-02 -6.71289563e-01 -7.54661739e-01
3.73583376e-01 -1.07476607e-01 -5.25089085e-01 -7.68683314e-01
-8.25470448e-01 9.20122087e-01 7.77446330e-02 -3.31802100e-01
-1.49014860e-01 3.41672897e-01 5.43258071e-01 4.84715432e-01
6.03734612e-01 9.02957320e-02 -3.04355949e-01 3.12283546e-01
-7.83267736e-01 4.53232199e-01 8.53065968e-01 1.16506660e+00
5.12045383e-01 -3.48004013e-01 -5.70942879e-01 6.54068828e-01
5.43644950e-02 5.12016773e-01 1.79795459e-01 -1.21361256e+00
6.48080885e-01 1.67161092e-01 2.30489969e-01 -9.83139634e-01
1.97696537e-01 -1.31290331e-01 -7.49024153e-01 1.53076515e-01
6.52797878e-01 -3.27195913e-01 -6.65048242e-01 2.35207129e+00
1.69690564e-01 -2.25228176e-01 3.33149105e-01 9.15094793e-01
3.74871343e-01 5.51648974e-01 2.86175162e-01 -9.66109633e-02
1.22706842e+00 -7.24449933e-01 -6.50239646e-01 -6.35147542e-02
5.00362337e-01 -3.31539571e-01 1.12809741e+00 1.13785744e-01
-1.15883648e+00 1.71485156e-01 -8.55772078e-01 -7.35490680e-01
-1.16591454e-01 8.18122476e-02 5.47189593e-01 9.01118040e-01
-1.07662416e+00 1.78085595e-01 -6.78493202e-01 -3.85660678e-01
1.04911137e+00 5.12701213e-01 -7.29491889e-01 -1.69503003e-01
-8.22775781e-01 4.54912543e-01 1.14043113e-02 -1.03420787e-01
-9.63412404e-01 -1.01130927e+00 -9.07605708e-01 1.77694365e-01
4.32140887e-01 -1.04748940e+00 1.21268725e+00 -1.25119185e+00
-1.19452572e+00 1.40088308e+00 -9.31795835e-02 -8.79938960e-01
1.16836667e+00 -1.48766056e-01 1.68476507e-01 3.75540793e-01
-6.00402020e-02 8.62598658e-01 1.13674736e+00 -1.19532096e+00
-6.38958573e-01 -7.33104944e-01 1.25798881e-01 3.57965052e-01
-4.64003593e-01 -1.01216570e-01 -5.56661427e-01 -5.08480787e-01
-5.21781206e-01 -1.01089394e+00 -5.09907268e-02 8.16624403e-01
-3.94751251e-01 1.97091028e-01 8.64328325e-01 -7.26515651e-01
6.92507207e-01 -2.51656008e+00 -1.89824507e-01 1.55036986e-01
2.59430379e-01 9.61499289e-02 -1.68390870e-01 2.54587531e-01
1.03197724e-01 1.13138177e-01 -6.22133911e-01 -9.58104491e-01
3.42237502e-01 2.24195361e-01 -6.15690172e-01 1.06350970e+00
-2.75367368e-02 9.55067694e-01 -5.10751605e-01 -4.38153565e-01
-1.34903848e-01 5.15459895e-01 -7.28244543e-01 1.88600138e-01
-3.04346591e-01 1.45342201e-01 -5.05668879e-01 6.65909290e-01
1.02654362e+00 -2.64768302e-01 -1.05827730e-02 -4.92891185e-02
8.74034241e-02 -2.82371610e-01 -5.16955137e-01 1.88877773e+00
-2.00340245e-02 7.38373578e-01 7.24559247e-01 -8.27799678e-01
3.12993288e-01 3.62316430e-01 2.61537910e-01 -5.60311556e-01
2.01427475e-01 -4.74069938e-02 -5.74398160e-01 -5.54818153e-01
4.50019717e-01 -1.41147733e-01 -9.51439887e-02 3.01122010e-01
-2.27648895e-02 2.78622746e-01 -4.23524857e-01 3.49385172e-01
8.83404016e-01 -2.39242747e-01 -3.44365649e-02 -3.16451877e-01
5.58062732e-01 -8.80951658e-02 4.26832706e-01 9.24905896e-01
-5.21771789e-01 6.97183251e-01 7.27363884e-01 4.05215984e-03
-8.70756209e-01 -1.03071642e+00 -2.37841591e-01 1.17713833e+00
-8.05987939e-02 -1.06146879e-01 -8.93159032e-01 -8.52676690e-01
4.35671151e-01 6.12336159e-01 -7.65237331e-01 -1.05772816e-01
6.58727512e-02 -7.00879395e-01 8.55645180e-01 1.21655598e-01
5.21088898e-01 -3.41003060e-01 -5.61475277e-01 -4.55155820e-01
4.85144816e-02 -1.13654852e+00 -9.70943868e-01 -1.53477877e-01
-4.42861468e-01 -8.59521806e-01 -1.00425375e+00 -6.09260619e-01
9.36356306e-01 3.24581504e-01 6.93066180e-01 -4.56145465e-01
-3.09116483e-01 1.21262646e+00 9.37441811e-02 -5.55732012e-01
-1.91302553e-01 -3.93451080e-02 -3.74347091e-01 7.16141999e-01
3.69549036e-01 -4.33819443e-01 -7.90556788e-01 -2.07168698e-01
-1.21881878e+00 -2.04521075e-01 3.78265291e-01 7.67728329e-01
7.43341386e-01 -4.86393273e-01 3.49773526e-01 -1.13618064e+00
7.65674829e-01 -4.54776734e-01 -8.52491021e-01 5.63456178e-01
-6.33168876e-01 7.54830837e-02 2.36715302e-01 -2.95626879e-01
-1.09479308e+00 3.30789059e-01 2.27592707e-01 -7.39250243e-01
6.98539764e-02 2.18024641e-01 -5.22960663e-01 -5.28955102e-01
2.20235303e-01 2.23583728e-01 4.08627421e-01 -2.98742652e-01
6.81488097e-01 6.25922084e-01 9.60784614e-01 -6.25658989e-01
5.31132698e-01 8.14327955e-01 7.40035204e-03 -8.03461015e-01
-5.22204041e-01 -2.37440482e-01 -7.54493326e-02 2.00096816e-01
8.88752222e-01 -1.22392917e+00 -1.19451153e+00 4.30808932e-01
-1.04479814e+00 1.02231272e-01 -3.81118029e-01 1.96346104e-01
-7.04785228e-01 6.61045492e-01 -4.82040495e-01 -1.13629615e+00
-5.19240797e-01 -9.21363235e-01 9.02213097e-01 1.26830220e-01
1.65920213e-01 -9.22705293e-01 -7.35921636e-02 5.04415929e-01
2.38897279e-01 4.24727619e-01 9.08872783e-01 -7.06702232e-01
-8.09102297e-01 -3.96178186e-01 -3.65800500e-01 6.06379151e-01
-2.24682942e-01 -5.43107271e-01 -1.31684375e+00 -5.95683455e-01
1.97434381e-01 -7.34060347e-01 9.88501728e-01 2.80039161e-01
1.59910595e+00 -9.65235472e-01 2.54690181e-02 9.59588587e-01
1.62103415e+00 -1.34830281e-01 3.76575887e-01 5.80743439e-02
6.09272838e-01 7.48223424e-01 1.32817835e-01 6.76492572e-01
4.48356628e-01 2.22354963e-01 5.83451033e-01 -1.25575811e-01
4.23605233e-01 -4.46604550e-01 2.90736854e-01 -3.20440046e-02
4.89009321e-01 -4.93281007e-01 -5.73097646e-01 7.83341885e-01
-2.04963446e+00 -7.03382671e-01 1.01089805e-01 2.52112103e+00
8.21280181e-01 -5.27484953e-01 1.05459474e-01 -4.48265612e-01
4.27914679e-01 2.73570389e-01 -8.91514838e-01 -5.93714535e-01
-4.59030539e-01 -1.35633841e-01 1.16125250e+00 5.37597477e-01
-1.31567669e+00 7.20833123e-01 6.37652969e+00 5.08385599e-01
-8.83075476e-01 3.02168369e-01 9.62654114e-01 -4.41257089e-01
-5.76946318e-01 -1.59155965e-01 -4.51801687e-01 2.18798429e-01
6.97691798e-01 -5.93713403e-01 3.49527210e-01 8.30715597e-01
-1.33769363e-01 9.23280939e-02 -1.49356019e+00 1.31295991e+00
2.62265861e-01 -1.22318685e+00 1.61953270e-01 4.06533539e-01
6.92125440e-01 3.81592885e-02 7.49785602e-01 -2.55032659e-01
2.11633474e-01 -1.02439940e+00 7.04188347e-01 3.26165736e-01
9.27774012e-01 -8.61819148e-01 2.85373479e-01 2.42420688e-01
-3.83726716e-01 -1.64122209e-01 -2.88958460e-01 2.97122747e-01
-1.56981215e-01 2.28920043e-01 -5.60975969e-01 4.64966536e-01
4.70529050e-01 4.14180756e-01 -4.63749319e-01 8.35058689e-01
2.24047989e-01 5.31769037e-01 -3.30332786e-01 2.72286803e-01
3.79891038e-01 -3.82531852e-01 6.40779674e-01 1.27648997e+00
1.33010685e-01 7.02626109e-02 -2.46834934e-01 8.81619692e-01
-6.21520698e-01 2.93355793e-01 -8.01080346e-01 -3.37897480e-01
2.86007702e-01 8.30346167e-01 1.71122730e-01 3.64874117e-02
-6.27826214e-01 1.36975825e+00 2.30205521e-01 6.23572290e-01
-5.75580657e-01 -1.21434696e-01 1.05981290e+00 -1.93262205e-01
3.71306002e-01 -8.11130553e-02 -4.17971283e-01 -1.24258709e+00
3.21637064e-01 -8.91489089e-01 7.16205180e-01 -3.44889879e-01
-1.56522667e+00 2.21638799e-01 -1.02452099e-01 -7.13401198e-01
-6.95764273e-02 -4.17504162e-01 -1.48347095e-01 9.56690729e-01
-1.47768247e+00 -1.37311316e+00 1.41834274e-01 1.06004226e+00
-6.52116537e-02 -8.18860307e-02 8.47626686e-01 2.92489529e-01
-5.33288896e-01 1.41995192e+00 5.45431197e-01 1.15072601e-01
6.91950142e-01 -9.58427489e-01 -3.11585981e-02 9.51196313e-01
9.68874712e-03 4.76931393e-01 6.01264238e-01 -1.90226361e-01
-1.83349323e+00 -1.18772376e+00 8.31212938e-01 -3.94030809e-01
5.63706160e-01 -7.42920458e-01 -8.70948732e-01 9.88640428e-01
3.39351773e-01 1.71962142e-01 9.10749555e-01 -2.96139121e-01
-8.11331451e-01 -3.34758610e-01 -1.61020851e+00 5.23189962e-01
6.42473280e-01 -9.77055430e-01 -2.07351610e-01 3.52101654e-01
7.39273489e-01 -2.39263460e-01 -5.77808619e-01 4.90911454e-02
6.02829278e-01 -7.80457795e-01 8.74954939e-01 -8.44862103e-01
2.75511384e-01 7.02378005e-02 -5.61629176e-01 -6.06811941e-01
1.43939421e-01 -1.15968382e+00 5.29079288e-02 1.38675880e+00
3.78247708e-01 -7.83696711e-01 9.65464473e-01 1.55277514e+00
4.40797210e-01 -4.00710672e-01 -1.28924417e+00 -6.00142002e-01
1.86603934e-01 -3.01231742e-01 1.85050324e-01 1.00424075e+00
-2.37966970e-01 -3.11997961e-02 -8.04302275e-01 3.70284736e-01
1.08316803e+00 -1.43078476e-01 7.84595072e-01 -5.82565784e-01
-3.68749708e-01 -2.65748829e-01 -3.36785227e-01 -1.03877044e+00
4.50860560e-01 -1.02225232e+00 -1.90274104e-01 -8.45611155e-01
4.28428888e-01 -5.07738926e-02 -2.26900175e-01 5.72627842e-01
4.79476005e-02 5.55254333e-02 3.33413184e-01 6.58755377e-02
-4.30948168e-01 6.92405879e-01 9.34093416e-01 -2.26943612e-01
8.18632394e-02 2.33601462e-02 -9.96365309e-01 3.57431561e-01
4.73926514e-01 -4.33504701e-01 -6.89959288e-01 -7.49107599e-01
1.63884372e-01 -9.98680443e-02 7.83486187e-01 -4.92833674e-01
4.39339668e-01 2.79007293e-02 3.03280191e-03 -3.07114363e-01
4.16170090e-01 -1.19042373e+00 -2.29213629e-02 1.58080727e-01
-1.02242780e+00 -2.96923190e-01 1.67329282e-01 1.00788641e+00
-5.20253517e-02 5.09215854e-02 8.33185613e-01 1.26618445e-01
-2.73777127e-01 7.29957879e-01 -1.41487375e-01 2.19507813e-01
1.05594492e+00 1.51918903e-01 -3.39931726e-01 -6.26351893e-01
-4.67139065e-01 5.80559790e-01 5.63016772e-01 2.27536842e-01
5.22290766e-01 -1.14554346e+00 -7.32684851e-01 1.42316431e-01
2.93512762e-01 -2.05577657e-01 3.14404905e-01 7.38703847e-01
-3.49470615e-01 4.20123935e-01 1.41522408e-01 -3.65851969e-01
-1.54712737e+00 7.07858205e-01 2.41028935e-01 1.73019916e-01
-6.77349150e-01 1.00742579e+00 6.92519367e-01 -7.42850900e-02
8.43277752e-01 -6.25258684e-02 4.78836656e-01 -1.51051402e-01
6.95673108e-01 1.30570471e-01 -2.73662776e-01 -4.47338641e-01
-3.15129817e-01 2.75267637e-04 -3.40637177e-01 -4.37516510e-01
1.16263819e+00 -4.77831900e-01 -1.30320087e-01 2.82903135e-01
1.83678174e+00 -6.53342754e-02 -1.62776303e+00 -2.93301165e-01
-3.09499264e-01 -7.62468159e-01 7.15895370e-02 -7.49810517e-01
-1.03303897e+00 8.99609089e-01 8.30822170e-01 -8.17272142e-02
1.15453231e+00 3.78920347e-03 7.66463399e-01 3.26621145e-01
1.16561197e-01 -8.07923615e-01 -5.56771100e-01 5.19019738e-02
8.48047376e-01 -1.36798751e+00 -8.34475756e-02 -2.34742150e-01
-9.43678558e-01 4.47097540e-01 1.49501815e-01 1.76241949e-01
5.73061109e-01 9.25472602e-02 7.44645372e-02 2.30576649e-01
-8.00146580e-01 1.98838294e-01 3.05863231e-01 7.07490623e-01
1.33042140e-02 -3.44475312e-03 -3.45667154e-01 5.61582088e-01
2.85364147e-02 9.08116996e-02 3.03411841e-01 1.04747403e+00
9.50662196e-02 -9.89598930e-01 -1.47210300e-01 1.85630172e-01
-9.52049553e-01 -5.83577268e-02 -6.62091851e-01 6.01784348e-01
-8.55719894e-02 7.74504006e-01 8.25404152e-02 -2.20450964e-02
-1.72049850e-02 8.73519778e-02 5.25222957e-01 -1.86348595e-02
-5.62758625e-01 -1.09914519e-01 -1.17023349e-01 -6.49040163e-01
-4.88741219e-01 -9.18961048e-01 -8.48328054e-01 -3.96720648e-01
1.40646115e-01 8.96389410e-02 8.22867393e-01 6.90279305e-01
7.89215147e-01 -3.69097888e-01 6.70594752e-01 -5.27741015e-02
-9.78599250e-01 -1.58384874e-01 -7.02754557e-01 5.34902394e-01
8.81017387e-01 1.14667915e-01 -4.95957017e-01 2.29434118e-01] | [5.920570373535156, 6.749251365661621] |
55e2217c-59cc-44aa-891e-38611fbd8ecc | emotional-talking-head-generation-based-on | 2306.03594 | null | https://arxiv.org/abs/2306.03594v1 | https://arxiv.org/pdf/2306.03594v1.pdf | Emotional Talking Head Generation based on Memory-Sharing and Attention-Augmented Networks | Given an audio clip and a reference face image, the goal of the talking head generation is to generate a high-fidelity talking head video. Although some audio-driven methods of generating talking head videos have made some achievements in the past, most of them only focused on lip and audio synchronization and lack the ability to reproduce the facial expressions of the target person. To this end, we propose a talking head generation model consisting of a Memory-Sharing Emotion Feature extractor (MSEF) and an Attention-Augmented Translator based on U-net (AATU). Firstly, MSEF can extract implicit emotional auxiliary features from audio to estimate more accurate emotional face landmarks.~Secondly, AATU acts as a translator between the estimated landmarks and the photo-realistic video frames. Extensive qualitative and quantitative experiments have shown the superiority of the proposed method to the previous works. Codes will be made publicly available. | ['Sen Li', 'Qi Li', 'Tianyi Xu', 'Li Liu', 'Yaxin Zhao', 'Jianrong Wang'] | 2023-06-06 | null | null | null | null | ['talking-head-generation'] | ['computer-vision'] | [-6.89536706e-02 2.32736394e-01 1.59111068e-01 -5.86955249e-01
-7.72862375e-01 5.69483899e-02 4.18966830e-01 -8.19117010e-01
6.27112314e-02 6.12438679e-01 5.89517951e-01 5.17244816e-01
3.73709202e-01 -2.84658074e-01 -5.72964311e-01 -8.20298135e-01
1.50600731e-01 1.61037734e-03 -3.14394861e-01 -2.34490391e-02
2.16528416e-01 3.10353011e-01 -1.93648398e+00 2.22506657e-01
6.19913280e-01 1.24791944e+00 2.74066269e-01 5.32166600e-01
-2.04326026e-03 8.77652526e-01 -5.20041466e-01 -4.40567434e-01
-1.24326311e-01 -5.58680236e-01 -4.75865126e-01 4.08087432e-01
2.32109651e-01 -4.15133059e-01 -5.77975094e-01 1.06125045e+00
1.08128250e+00 2.64844507e-01 5.40433884e-01 -1.61575449e+00
-5.26597857e-01 4.50363159e-01 -5.11911213e-01 -6.61732629e-02
9.75206137e-01 6.82112649e-02 4.08579528e-01 -1.25019169e+00
8.61459613e-01 1.45821738e+00 5.88732958e-01 9.53775287e-01
-7.88874686e-01 -1.08463609e+00 -9.77414548e-02 6.36656225e-01
-1.86324239e+00 -1.24930871e+00 1.30885649e+00 -1.62299767e-01
3.68227154e-01 1.15247242e-01 7.12398827e-01 1.13931322e+00
1.99440736e-02 8.96508515e-01 8.54153454e-01 -3.92175227e-01
-2.34567001e-02 4.07654971e-01 -5.63963175e-01 6.87905312e-01
-7.30935752e-01 1.62956901e-02 -8.56428266e-01 4.58167046e-02
5.87872207e-01 -2.26378709e-01 -8.49355996e-01 -6.63699135e-02
-8.68007421e-01 7.03523219e-01 2.81215817e-01 3.26958954e-01
-4.86610830e-01 2.27330849e-02 2.90048033e-01 -3.25210765e-02
5.91208398e-01 -1.82872936e-01 3.46609771e-01 -2.11221293e-01
-1.14111757e+00 2.30059605e-02 5.02085567e-01 1.02652752e+00
6.02071464e-01 2.19889373e-01 -1.58094659e-01 7.81256974e-01
4.05332237e-01 6.15245342e-01 5.25581539e-01 -1.05189764e+00
1.00431539e-01 1.81803972e-01 2.71791853e-02 -1.27355790e+00
-1.35720894e-01 -1.09266408e-01 -6.40828788e-01 1.96973840e-03
-1.97512120e-01 -4.48173851e-01 -3.60061020e-01 1.92098188e+00
5.28564095e-01 6.51054382e-01 -3.66419479e-02 1.12003279e+00
1.03284287e+00 8.90225947e-01 -1.22644529e-01 -6.16173029e-01
1.11928701e+00 -8.14973593e-01 -1.36774337e+00 3.80685292e-02
1.54573098e-01 -8.94120753e-01 8.29201996e-01 1.82581395e-01
-1.43501937e+00 -7.79118776e-01 -7.72124767e-01 -5.84055036e-02
-2.68295710e-03 2.56059766e-01 2.36065537e-01 5.10699987e-01
-1.30446470e+00 6.27500415e-02 -3.51186097e-01 -2.56081134e-01
2.72535235e-01 3.93208146e-01 -5.49771011e-01 8.95666704e-03
-1.18525290e+00 6.79516315e-01 1.10429481e-01 4.03474361e-01
-9.48114574e-01 -2.86380082e-01 -9.83840287e-01 1.15326889e-01
6.74400404e-02 -6.64327502e-01 1.38706505e+00 -1.53980601e+00
-2.01171732e+00 6.06542468e-01 -6.31298006e-01 -2.00919565e-02
4.31086034e-01 -1.71471789e-01 -4.79090601e-01 4.63818640e-01
-6.97167143e-02 1.08300984e+00 1.28130436e+00 -1.36527669e+00
-5.87865412e-01 -3.45324814e-01 -3.98047298e-01 4.75241512e-01
-4.58733618e-01 3.62280071e-01 -7.62673497e-01 -5.68342984e-01
-1.27955422e-01 -7.97703326e-01 3.64534855e-01 4.40075211e-02
-4.36985224e-01 -1.10934690e-01 1.18796659e+00 -8.43681991e-01
1.23422146e+00 -2.41972733e+00 2.31740892e-01 -9.00288448e-02
3.03883348e-02 2.05067337e-01 -1.01177663e-01 2.07958892e-01
-2.19518825e-01 -3.06786388e-01 1.64394855e-01 -7.54008055e-01
-9.47136059e-02 -1.84292451e-01 -2.95515418e-01 6.02102160e-01
1.29354790e-01 7.66300678e-01 -7.71675766e-01 -1.05822599e+00
2.53514022e-01 1.15649199e+00 -5.49378514e-01 5.63349187e-01
1.21457323e-01 6.00341320e-01 -2.97337949e-01 6.08552396e-01
7.19371676e-01 3.96332741e-01 -1.21007897e-01 -3.13344210e-01
-1.53004318e-01 -1.63120046e-01 -1.00571632e+00 1.62717628e+00
-5.13217688e-01 7.18670666e-01 5.22425413e-01 -4.13435370e-01
1.02302551e+00 8.68795931e-01 4.76231843e-01 -5.11538684e-01
5.91951668e-01 -2.72330921e-02 -3.29228342e-01 -8.19765031e-01
3.59372705e-01 -3.04061860e-01 2.92738646e-01 1.89468801e-01
1.64486647e-01 -1.30502433e-01 -2.76679188e-01 -3.51047926e-02
5.39181352e-01 2.22927079e-01 -5.31343557e-03 1.14515148e-01
9.75808680e-01 -6.42850459e-01 6.24681175e-01 -3.00477203e-02
-3.36222082e-01 7.51282394e-01 1.46960527e-01 -5.28042018e-02
-8.15468192e-01 -7.38015831e-01 4.36835140e-02 1.07299542e+00
7.63288662e-02 -3.58392745e-01 -1.20287573e+00 -1.76066265e-01
-5.12738943e-01 6.87798560e-01 -5.35419822e-01 -1.70097530e-01
-3.29651177e-01 -2.15718951e-02 5.93725264e-01 4.64335084e-01
7.13748693e-01 -1.26037979e+00 -3.81868005e-01 9.23006758e-02
-7.98497856e-01 -1.07931530e+00 -8.75996828e-01 -5.64530671e-01
-2.68860251e-01 -6.52165830e-01 -1.01702416e+00 -1.02604091e+00
7.59139836e-01 9.81863439e-02 5.42736709e-01 -2.53986984e-01
-1.56943232e-01 4.69171703e-01 -1.98854789e-01 -6.09756231e-01
-1.86041042e-01 -2.85846740e-01 3.24793547e-01 6.03442788e-01
2.61841565e-01 -5.48020065e-01 -5.84973454e-01 3.26037258e-01
-6.43333197e-01 1.77813023e-01 4.25069243e-01 7.21228063e-01
3.18062961e-01 -8.81032720e-02 7.16874182e-01 -2.18358934e-01
7.13817239e-01 -2.67033964e-01 -2.22781092e-01 -8.10687542e-02
-8.05154964e-02 -3.30701858e-01 5.63976824e-01 -5.86430788e-01
-1.63380313e+00 2.83562571e-01 -4.56509143e-01 -9.61029887e-01
-1.38554767e-01 9.81214121e-02 -6.25402927e-01 -5.22916727e-02
1.06306791e-01 4.57760900e-01 9.76339877e-02 -1.45567715e-01
2.71304160e-01 1.18948412e+00 9.08902287e-01 -2.74889171e-01
5.51036298e-01 4.41274375e-01 -3.39080155e-01 -9.92679954e-01
-5.06087363e-01 -2.74165571e-01 -4.78255630e-01 -8.88966799e-01
9.51514959e-01 -9.53356087e-01 -1.15517056e+00 6.77987993e-01
-1.27030432e+00 1.64709631e-02 1.61746200e-02 5.51595926e-01
-8.66565406e-01 3.75746757e-01 -5.30871391e-01 -1.12718236e+00
-4.33323294e-01 -1.19705236e+00 1.25874794e+00 4.28156078e-01
-7.52923489e-02 -7.11679161e-01 1.12118199e-02 4.64958936e-01
4.59484637e-01 3.66659835e-02 4.67745692e-01 -1.87840536e-02
-3.28232139e-01 -2.68594295e-01 -1.07945599e-01 2.66467303e-01
5.16405739e-02 1.42921150e-01 -1.43536925e+00 -2.06960097e-01
3.35931987e-01 -4.08181667e-01 3.12167168e-01 5.98653257e-01
1.00766599e+00 -4.66508508e-01 -3.81984532e-01 6.30246818e-01
8.76683235e-01 4.76386577e-01 8.52551222e-01 -3.13959956e-01
4.50554848e-01 8.47825587e-01 6.79272890e-01 5.95698953e-01
4.18773681e-01 8.99393439e-01 3.94211143e-01 -8.09258781e-03
-1.45462930e-01 -4.09599632e-01 7.22308755e-01 1.11986840e+00
-8.77287760e-02 -1.39405102e-01 -5.34835458e-01 5.27970254e-01
-1.70771289e+00 -1.21017516e+00 3.42120498e-01 2.01436830e+00
7.49645352e-01 -1.99450687e-01 1.33789377e-03 1.57601267e-01
1.07509339e+00 9.09269229e-02 -4.90235716e-01 -2.66024232e-01
-3.42029780e-02 -1.07828990e-01 -3.99981141e-01 4.97469604e-01
-7.65351355e-01 8.82400930e-01 5.75874901e+00 7.63836563e-01
-1.46981156e+00 2.16755886e-02 5.85320294e-01 -1.47027701e-01
-8.68143439e-02 -2.81186819e-01 -6.55029356e-01 5.13843179e-01
8.00865948e-01 -2.97971487e-01 3.58436495e-01 9.82506216e-01
6.16464317e-01 5.79515807e-02 -9.55337048e-01 1.41941166e+00
6.51363611e-01 -9.19743359e-01 -9.48836934e-03 -7.35463575e-02
3.35269719e-01 -5.25815427e-01 2.30323538e-01 2.01862708e-01
-4.25426871e-01 -1.02452576e+00 8.19628179e-01 9.82851386e-01
1.05939341e+00 -1.03782213e+00 8.74315262e-01 2.25729972e-01
-1.35804296e+00 -7.05157593e-02 -2.69905925e-01 4.48882356e-02
3.97401989e-01 1.00998804e-01 -9.75487411e-01 3.78643006e-01
6.78516746e-01 3.96843791e-01 -2.37627760e-01 9.27499294e-01
-1.46442577e-01 3.39720100e-01 -1.12217203e-01 1.47925496e-01
-1.28143236e-01 -3.42414677e-02 6.62097037e-01 1.12070489e+00
4.57243621e-01 1.97033346e-01 -5.56525663e-02 7.69677222e-01
-2.18082309e-01 3.88212860e-01 -7.40155160e-01 7.61901960e-02
4.58611131e-01 1.38376224e+00 -1.80392101e-01 -2.07450151e-01
-3.36211234e-01 1.11742496e+00 -1.27762958e-01 2.21782520e-01
-1.00783682e+00 -6.76232576e-01 6.95433199e-01 1.72134176e-01
2.76742298e-02 2.35450119e-01 3.72449845e-01 -1.11683440e+00
-7.00411201e-02 -7.86284626e-01 -1.00606807e-01 -1.43342626e+00
-9.11916137e-01 8.96238029e-01 -2.03848541e-01 -1.22757733e+00
-6.03019774e-01 -7.72670731e-02 -7.70613611e-01 8.71791124e-01
-1.20256329e+00 -1.17939973e+00 -6.62779927e-01 1.03101408e+00
7.54870415e-01 -6.19460419e-02 7.59655058e-01 4.83556807e-01
-6.02481127e-01 7.00204432e-01 -2.19228238e-01 1.49931684e-01
9.86697197e-01 -5.44215381e-01 -2.39760533e-01 4.87698406e-01
-2.40433365e-01 4.12621081e-01 9.11753714e-01 -4.92187589e-01
-1.33984816e+00 -1.06380081e+00 8.99928153e-01 3.00676245e-02
2.73412526e-01 -3.75709921e-01 -7.50067770e-01 7.04066694e-01
5.01999676e-01 -8.57865065e-02 5.58686316e-01 -5.51247120e-01
1.52391508e-01 -2.50274420e-01 -1.21012402e+00 5.26167154e-01
8.32931995e-01 -7.25828707e-01 -4.75445956e-01 -6.10694140e-02
3.74076158e-01 -3.59208494e-01 -7.40981579e-01 2.74423301e-01
7.27736592e-01 -9.66641247e-01 6.25783145e-01 -7.19181225e-02
3.36374730e-01 -2.50609279e-01 1.18974280e-02 -1.26596236e+00
-8.27678144e-02 -7.54317403e-01 1.39674529e-01 1.77624679e+00
-1.69961572e-01 -4.02109623e-01 5.77138364e-01 4.14195925e-01
-1.08340628e-01 -4.60240215e-01 -1.00503147e+00 -3.27881545e-01
-5.11667848e-01 -2.96093524e-01 5.34618497e-01 8.53317201e-01
4.08630759e-01 5.09494543e-01 -7.31485546e-01 1.61743294e-02
4.81121570e-01 -2.82904990e-02 8.24218392e-01 -1.01833022e+00
9.69595090e-02 -1.55693427e-01 -4.38920736e-01 -8.73410046e-01
6.90596998e-01 -5.20976961e-01 3.58193547e-01 -1.09725356e+00
2.43602753e-01 9.37582180e-02 7.06416741e-02 2.65988588e-01
-5.11648890e-04 2.97420740e-01 2.05155134e-01 3.77341770e-02
-4.53577846e-01 1.06445384e+00 1.24764657e+00 5.70168979e-02
-1.87404335e-01 2.48309392e-02 -4.24965918e-01 8.76905560e-01
5.79730093e-01 -2.26881921e-01 -4.91223097e-01 -7.23326728e-02
-3.67871255e-01 7.64339209e-01 2.38640189e-01 -1.15165091e+00
4.05986756e-01 6.47946373e-02 4.70909625e-01 -5.81492424e-01
8.75654936e-01 -8.10918391e-01 3.11463475e-01 1.77735969e-01
-2.59436011e-01 1.13230892e-01 2.94571500e-02 3.30271840e-01
-5.74342668e-01 -7.51368478e-02 8.42083275e-01 3.08741406e-02
-4.52816904e-01 3.12916011e-01 -4.00404334e-01 -3.00256699e-01
1.02337372e+00 -3.13385099e-01 -8.67183656e-02 -1.17395222e+00
-7.25072026e-01 3.82462218e-02 4.36432719e-01 4.46361095e-01
1.05441499e+00 -1.40667760e+00 -7.15601563e-01 2.60252953e-01
-1.04690669e-02 -2.60822922e-01 4.76787597e-01 9.03004587e-01
-1.71538889e-01 4.48739201e-01 -4.97396410e-01 -4.31807548e-01
-1.55283296e+00 5.95711410e-01 3.77904356e-01 4.18715835e-01
-4.06931192e-01 7.81076431e-01 4.03748304e-01 -7.55553171e-02
5.37487149e-01 2.81710565e-01 -3.29825938e-01 1.53621778e-01
6.92932129e-01 3.00244123e-01 -2.99993455e-01 -1.33980238e+00
-3.00713122e-01 5.55543602e-01 2.26609349e-01 -2.92479366e-01
1.25963449e+00 -5.08894384e-01 8.82084668e-03 4.20277148e-01
1.28399980e+00 1.62993193e-01 -1.24549031e+00 -6.66948557e-02
-5.65256119e-01 -4.61664051e-01 1.41103834e-01 -3.09520125e-01
-1.30647361e+00 1.11640549e+00 7.72472858e-01 -3.95253032e-01
1.55727541e+00 -1.75533667e-01 9.11288559e-01 7.58706480e-02
3.03247690e-01 -1.08102739e+00 2.58741349e-01 6.07188605e-02
1.05858946e+00 -8.44680429e-01 -4.46238577e-01 -5.70845902e-01
-8.33289623e-01 1.04104650e+00 7.23929703e-01 2.71872252e-01
5.54049015e-01 2.65065253e-01 2.31394500e-01 8.12567621e-02
-8.09969544e-01 -1.33335128e-01 8.72699469e-02 7.99406469e-01
4.51914102e-01 -4.09323543e-01 -2.50535440e-02 5.86544335e-01
-4.05583471e-01 4.06147450e-01 4.03356433e-01 5.94211161e-01
-2.95927405e-01 -6.12590194e-01 -6.80831492e-01 -1.26939788e-01
-4.20632154e-01 3.11740469e-02 -3.15275222e-01 5.07102609e-01
1.84498280e-01 1.06933010e+00 9.97955650e-02 -5.90902567e-01
2.11594850e-01 5.33836722e-01 3.10866296e-01 -1.83602050e-01
-3.61682296e-01 3.51963788e-01 -1.01868570e-01 -5.30332029e-01
-5.19922078e-01 -5.67178845e-01 -1.09814727e+00 -3.25858295e-01
-4.56219107e-01 3.09009165e-01 7.26987720e-01 6.82023764e-01
4.82709855e-01 3.64439070e-01 9.56420481e-01 -1.45035493e+00
-9.43633243e-02 -1.10019052e+00 -6.08822644e-01 4.52948511e-01
2.56459653e-01 -8.27710688e-01 -3.97632390e-01 3.30451131e-01] | [13.246480941772461, -0.38782984018325806] |
22031173-176b-4cae-b2c3-617cef8c174a | imbedding-deep-neural-networks-1 | 2202.00113 | null | https://arxiv.org/abs/2202.00113v2 | https://arxiv.org/pdf/2202.00113v2.pdf | Imbedding Deep Neural Networks | Continuous-depth neural networks, such as Neural ODEs, have refashioned the understanding of residual neural networks in terms of non-linear vector-valued optimal control problems. The common solution is to use the adjoint sensitivity method to replicate a forward-backward pass optimisation problem. We propose a new approach which explicates the network's `depth' as a fundamental variable, thus reducing the problem to a system of forward-facing initial value problems. This new method is based on the principle of `Invariant Imbedding' for which we prove a general solution, applicable to all non-linear, vector-valued optimal control problems with both running and terminal loss. Our new architectures provide a tangible tool for inspecting the theoretical--and to a great extent unexplained--properties of network depth. They also constitute a resource of discrete implementations of Neural ODEs comparable to classes of imbedded residual neural networks. Through a series of experiments, we show the competitive performance of the proposed architectures for supervised learning and time series prediction. | ['Dmitry Kangin', 'Andrew Corbett'] | 2022-01-31 | imbedding-deep-neural-networks | https://openreview.net/forum?id=yKIAXjkJc2F | https://openreview.net/pdf?id=yKIAXjkJc2F | iclr-2022-4 | ['time-series-prediction'] | ['time-series'] | [ 2.21756086e-01 5.61952353e-01 2.70685758e-02 -1.09207556e-01
-2.41599053e-01 -5.66939116e-01 4.01111692e-01 -3.83724302e-01
-5.20873487e-01 7.23347247e-01 -3.85638028e-02 -5.83890557e-01
-4.18235511e-01 -4.59730864e-01 -8.43359232e-01 -1.05031705e+00
-4.68950689e-01 8.24788809e-02 -4.67311814e-02 -5.79523861e-01
1.72184452e-01 7.29675591e-01 -1.45763946e+00 -2.01355442e-01
4.84606147e-01 1.42400157e+00 -1.67171568e-01 6.20574355e-01
1.53638735e-01 9.73270237e-01 -1.39866516e-01 -2.05985799e-01
5.77428579e-01 -4.01064485e-01 -8.15762460e-01 -1.31803334e-01
1.39134109e-01 -3.04012261e-02 -3.32637012e-01 1.07084179e+00
3.91858190e-01 4.57034767e-01 6.69495523e-01 -1.33164477e+00
-6.71317220e-01 1.73998341e-01 -1.21836476e-01 2.39107355e-01
-3.77842456e-01 -7.77052939e-02 1.04852092e+00 -8.22518289e-01
4.98887569e-01 9.79865134e-01 1.29435754e+00 7.29771495e-01
-1.56578505e+00 -8.80349055e-03 1.75559521e-01 -5.72184287e-02
-9.35177326e-01 -5.20627558e-01 8.75802517e-01 -6.93894804e-01
9.38285291e-01 3.19418430e-01 5.95492363e-01 9.03882265e-01
2.51966089e-01 5.80497324e-01 7.91530788e-01 -4.14437264e-01
3.91219825e-01 1.63867280e-01 1.14038169e-01 8.02601397e-01
1.42848343e-02 4.86726016e-01 1.00930937e-01 2.31852368e-01
1.06032181e+00 -2.31663641e-02 -5.72140515e-01 -6.47353768e-01
-1.15548599e+00 1.14327133e+00 5.66715240e-01 4.21812147e-01
-3.42402250e-01 3.13497692e-01 6.07479453e-01 5.77131748e-01
6.72294021e-01 7.11613715e-01 -6.54786527e-01 2.53241688e-01
-7.39678562e-01 3.19921553e-01 1.22614062e+00 6.11052215e-01
5.71314156e-01 5.98082960e-01 1.92072317e-02 5.85565865e-01
4.01069015e-01 1.98741242e-01 6.45128191e-01 -1.40100598e+00
2.24371210e-01 4.08967674e-01 2.37471256e-02 -1.16348803e+00
-7.07799315e-01 -7.26497829e-01 -1.34397519e+00 7.56745398e-01
5.76514602e-01 -4.24476683e-01 -5.28851390e-01 1.96158087e+00
9.68692377e-02 5.01042977e-02 2.98993587e-01 1.00534630e+00
1.05609931e-01 7.36845911e-01 -5.30819654e-01 -3.97587031e-01
8.31428766e-01 -8.11221480e-01 -5.76227367e-01 9.63713750e-02
6.53142273e-01 -1.90134794e-02 7.82670915e-01 2.94508129e-01
-1.34240913e+00 -3.25060964e-01 -1.07635510e+00 2.76817679e-02
-4.75221783e-01 -1.29587531e-01 2.67659485e-01 3.65303367e-01
-1.46678376e+00 1.58544660e+00 -8.08010697e-01 2.64118966e-02
1.50357053e-01 4.54747111e-01 -4.10045415e-01 7.31570303e-01
-1.38613760e+00 9.68539596e-01 3.34503621e-01 8.86662304e-01
-5.50869286e-01 -8.90322626e-01 -8.11163962e-01 1.81595124e-02
2.50086039e-01 -5.08916736e-01 1.31812418e+00 -1.03164566e+00
-1.77575302e+00 8.69210780e-01 1.88404620e-01 -7.50229955e-01
7.73324966e-01 6.42318204e-02 6.03472143e-02 -1.13746813e-02
-2.89043039e-01 2.82796383e-01 9.72826242e-01 -1.04268420e+00
-1.56855330e-01 -3.75613064e-01 8.01123306e-02 -5.12869470e-02
-3.40120167e-01 -4.33333904e-01 3.58397931e-01 -7.59149373e-01
2.07067907e-01 -8.12281311e-01 -6.58055723e-01 4.42727864e-01
-1.71310753e-01 7.00261444e-02 7.35452950e-01 -5.65505743e-01
9.68653500e-01 -1.86367929e+00 7.29616404e-01 1.98377505e-01
4.06182021e-01 2.74367571e-01 4.40542400e-02 4.06618387e-01
-5.52809775e-01 5.99245280e-02 -5.99729776e-01 -4.37243730e-01
1.32038027e-01 3.34804416e-01 -5.39157212e-01 8.32457602e-01
2.77790397e-01 7.91553974e-01 -6.98692799e-01 -1.28932655e-01
1.07597530e-01 5.96421421e-01 -6.27558649e-01 2.19732255e-01
-3.23798448e-01 3.81015241e-01 -3.38055283e-01 5.31059392e-02
2.54618734e-01 -5.55433556e-02 -2.18788370e-01 -1.27754488e-03
-5.03349364e-01 -2.11796150e-01 -1.29359913e+00 1.60612845e+00
-4.35059309e-01 9.44708645e-01 6.73854649e-01 -1.67172909e+00
8.56394887e-01 2.86212862e-01 4.65574324e-01 -5.19235849e-01
2.97600389e-01 3.54639888e-01 -3.75236660e-01 -5.66973865e-01
2.08730489e-01 -5.84745944e-01 9.38543230e-02 1.35857061e-01
-2.73759197e-02 2.58434325e-01 -4.01710197e-02 -2.75673926e-01
9.86127377e-01 3.79971653e-01 -5.01561202e-02 -7.44500875e-01
8.54652643e-01 1.09424300e-01 3.58346462e-01 5.93420446e-01
-6.80723339e-02 4.82551962e-01 9.83962059e-01 -5.71136892e-01
-1.23147309e+00 -7.80684054e-01 -3.59187067e-01 8.45238328e-01
-3.06014806e-01 4.42346066e-01 -8.27539265e-01 6.01000562e-02
4.50841559e-04 3.62793177e-01 -9.33656693e-01 -2.41654187e-01
-7.55106211e-01 -4.91105050e-01 5.26932955e-01 5.51248789e-01
4.90584016e-01 -1.04369652e+00 -8.83383274e-01 3.11030328e-01
2.29469955e-01 -6.28274143e-01 -2.78753817e-01 7.63187289e-01
-1.15394580e+00 -8.27105343e-01 -1.22671199e+00 -8.15335512e-01
3.82870972e-01 -3.79646689e-01 8.46354961e-01 -9.83314663e-02
-1.91343024e-01 3.26875538e-01 2.07126752e-01 -1.44492118e-02
-4.61196870e-01 1.55429497e-01 3.54577541e-01 1.94532439e-01
-5.11124372e-01 -8.90159249e-01 -5.65303445e-01 1.49464116e-01
-9.14666295e-01 -2.49186307e-01 3.84460598e-01 1.08659375e+00
3.85395616e-01 -8.05092007e-02 5.97315609e-01 -5.96809030e-01
7.15435684e-01 -5.00233054e-01 -9.99802291e-01 2.96947844e-02
-7.62555718e-01 8.22761655e-01 1.14749646e+00 -5.41714966e-01
-7.83198059e-01 -1.25376799e-03 -2.92798609e-01 -7.33069539e-01
2.35798448e-01 5.34352660e-01 6.32785186e-02 -3.20946813e-01
7.45063543e-01 1.00447990e-01 5.45060635e-01 -5.51109910e-01
3.18339437e-01 2.72733837e-01 8.68387043e-01 -6.23250425e-01
7.75048316e-01 2.39536449e-01 6.55531108e-01 -7.93790698e-01
-6.45317972e-01 -3.97504538e-01 -6.39252365e-01 -2.10351929e-01
7.70950258e-01 -5.40660858e-01 -1.28181839e+00 6.44407988e-01
-1.29358876e+00 -7.32206821e-01 -7.72924125e-01 2.96617627e-01
-8.84840250e-01 7.32641295e-02 -7.71676123e-01 -1.04613066e+00
-1.65209591e-01 -9.29935575e-01 6.64032280e-01 1.41794030e-02
1.79633334e-01 -1.63186336e+00 4.35862809e-01 -5.24931908e-01
6.11154377e-01 7.14151025e-01 8.05015981e-01 -4.88018125e-01
-1.22585997e-01 -1.91642955e-01 -4.73858789e-02 8.56074989e-01
-5.49655735e-01 -1.90134600e-01 -1.07010448e+00 -2.94247597e-01
6.66577280e-01 -7.03351349e-02 8.94561172e-01 4.71402735e-01
7.85749495e-01 -8.55360091e-01 -1.24919258e-01 9.75436211e-01
1.76094663e+00 7.23094568e-02 4.58491772e-01 2.96377003e-01
6.48816586e-01 9.74185050e-01 -8.18166211e-02 2.09676102e-01
-1.75817743e-01 4.56512511e-01 7.83481896e-01 -1.42395362e-01
5.90884984e-01 6.81887642e-02 4.16111857e-01 8.01138103e-01
-3.39800924e-01 1.02701947e-01 -7.92662978e-01 5.19836724e-01
-2.01261139e+00 -1.05007565e+00 -2.71675140e-01 2.22363329e+00
6.42890215e-01 8.33223015e-03 2.81030655e-01 4.30300623e-01
6.95225298e-01 2.31990144e-01 -6.29652023e-01 -6.52589798e-01
4.75180772e-04 -3.89004219e-03 6.91423655e-01 7.38984168e-01
-1.04636574e+00 3.04486156e-01 6.38383579e+00 3.12409520e-01
-1.23009491e+00 5.09657636e-02 3.86802703e-01 6.24048114e-02
-7.06446543e-02 -2.79128343e-01 -6.17287695e-01 3.19479823e-01
1.29478943e+00 -2.03495517e-01 7.11690962e-01 1.03686810e+00
2.05497399e-01 3.70396584e-01 -1.25771272e+00 6.62106872e-01
-1.60802886e-01 -1.50727952e+00 -4.55476612e-01 2.05156237e-01
6.35984719e-01 -5.63129224e-02 2.29672194e-01 3.18790823e-01
2.25545410e-02 -1.04277253e+00 8.56719077e-01 6.30218208e-01
6.79514468e-01 -6.24572814e-01 4.67346668e-01 4.68653589e-01
-1.00453413e+00 -3.42998147e-01 -2.70022541e-01 -4.23857331e-01
-2.43362132e-02 3.33387524e-01 -2.49086842e-01 1.48862630e-01
4.52296853e-01 9.08961773e-01 -8.66823420e-02 6.95921123e-01
2.77797699e-01 1.51567217e-02 -3.49639446e-01 -7.67125562e-02
5.50304592e-01 -4.07091290e-01 5.83479762e-01 1.03872311e+00
2.66809553e-01 5.14838088e-04 -3.64201665e-01 1.15218067e+00
1.16734736e-01 -1.11546382e-01 -8.74510527e-01 7.73806497e-02
-1.97376058e-01 1.05725443e+00 -6.62826538e-01 1.37872934e-01
-2.21537396e-01 8.19426060e-01 3.91916931e-01 5.13687849e-01
-5.90932727e-01 -5.88704884e-01 7.73819566e-01 -3.14574502e-02
3.16916436e-01 -1.50364220e-01 -1.88956350e-01 -1.16956794e+00
3.18434268e-01 -5.32917082e-01 7.05951899e-02 -4.78487521e-01
-9.74882424e-01 7.05748796e-01 5.41138425e-02 -1.10190344e+00
-7.54571497e-01 -1.25929308e+00 -5.96784234e-01 1.01355064e+00
-1.31487453e+00 -5.02849698e-01 8.11906084e-02 6.61279559e-01
1.07637718e-01 -6.07735403e-02 8.07487309e-01 1.11734837e-01
-6.96046829e-01 3.70287210e-01 6.81037128e-01 9.39501151e-02
-2.14123741e-01 -1.48123240e+00 2.20009550e-01 9.00500715e-01
-3.95974308e-01 3.66452515e-01 1.02974057e+00 5.56235900e-03
-1.48248816e+00 -9.57502246e-01 6.11636639e-01 -2.60577381e-01
1.12863290e+00 -3.47497940e-01 -1.18071353e+00 8.27690721e-01
-1.31616056e-01 1.78555846e-01 -1.06183559e-01 -9.35135186e-02
-5.49628474e-02 -2.47918859e-01 -1.07302892e+00 5.22629499e-01
1.02312696e+00 -3.94607723e-01 -6.74713552e-01 1.21700279e-01
8.74308884e-01 -3.00448328e-01 -1.09694970e+00 3.52160007e-01
6.18399143e-01 -9.89768267e-01 9.99282241e-01 -7.48361826e-01
5.85185409e-01 1.87856972e-01 -5.59476092e-02 -1.37378883e+00
-7.29645938e-02 -1.11316121e+00 -2.52536952e-01 8.07707071e-01
4.20236826e-01 -1.06161129e+00 7.99857080e-01 6.37865126e-01
-4.85576332e-01 -1.21971869e+00 -1.19933522e+00 -9.10517812e-01
5.87855935e-01 -3.65127683e-01 -8.41184333e-02 8.00275862e-01
5.82553111e-02 8.12244490e-02 -2.23738313e-01 1.45596126e-02
5.72687745e-01 -3.17080736e-01 -5.13040833e-02 -1.43970191e+00
-4.93433088e-01 -9.77495432e-01 -5.11625230e-01 -1.04413891e+00
4.73636270e-01 -9.45386827e-01 4.05032098e-01 -1.02980506e+00
-6.47579372e-01 -2.76820719e-01 -2.68480778e-01 3.27662796e-01
5.85003316e-01 4.02351767e-02 -1.75450802e-01 1.83954954e-01
-9.80680287e-02 4.65733498e-01 1.02661753e+00 1.60868272e-01
-4.09849256e-01 1.54734269e-01 -3.47678304e-01 7.35252857e-01
7.31029749e-01 -3.02129209e-01 -5.44258237e-01 -3.36641818e-01
4.71903056e-01 3.80161971e-01 7.00842679e-01 -1.00027490e+00
3.84261370e-01 8.05494189e-02 2.36016139e-02 -1.30791888e-01
3.91705960e-01 -8.79993796e-01 5.01010641e-02 7.38373578e-01
-6.66268229e-01 2.52718806e-01 2.25360692e-01 5.44914603e-01
-8.58553275e-02 -4.70274627e-01 9.07290816e-01 -2.86582232e-01
-5.10676920e-01 2.16709405e-01 -2.20524698e-01 3.66766959e-01
8.15505028e-01 -2.61738330e-01 -1.39386967e-01 -1.24806531e-01
-1.04826689e+00 4.54055853e-02 3.42439204e-01 -9.64659750e-02
4.73500013e-01 -1.26309633e+00 -4.33726907e-01 4.77106631e-01
-4.17724609e-01 1.14236534e-01 6.60127923e-02 1.05690885e+00
-6.55967295e-01 5.32935321e-01 -1.93102673e-01 -6.08163297e-01
-5.62341034e-01 7.24053919e-01 1.09689069e+00 -2.54882276e-01
-8.36120009e-01 8.02482188e-01 2.46029645e-02 -3.90605718e-01
6.79976523e-01 -6.11555696e-01 -3.93301025e-02 2.21506402e-01
1.93152681e-01 4.38367367e-01 3.66328321e-02 -4.84759420e-01
-1.09714620e-01 5.47931850e-01 4.00498956e-01 -3.14250976e-01
1.61100924e+00 -5.85589092e-04 -2.97178239e-01 9.81392503e-01
1.83841348e+00 -7.01064825e-01 -1.64178443e+00 -1.61180928e-01
1.30025953e-01 2.99064845e-01 1.92276359e-01 -2.90290833e-01
-1.24401069e+00 1.03843415e+00 3.91277105e-01 6.50821507e-01
1.04546452e+00 -3.35382283e-01 4.49565351e-01 7.71419227e-01
4.76743355e-02 -8.03256810e-01 -1.65600866e-01 6.58790708e-01
1.21609139e+00 -9.84466136e-01 -5.20436347e-01 2.12448999e-01
-4.88064066e-02 1.38662207e+00 1.97830424e-01 -5.89406133e-01
8.23763788e-01 3.33990872e-01 -9.55815315e-02 5.92561923e-02
-8.11168790e-01 1.19501770e-01 2.27201656e-01 4.42900211e-01
1.89433903e-01 -4.43387628e-01 1.68205351e-02 3.64387691e-01
-1.05141565e-01 -1.07150264e-01 5.13363779e-01 6.03999138e-01
-2.77351201e-01 -6.53575242e-01 -2.23353520e-01 1.56605199e-01
-5.51122367e-01 4.41835523e-02 -2.56803725e-03 1.11791837e+00
-2.48712078e-01 3.48535478e-01 8.89704973e-02 -2.39063352e-01
5.07973373e-01 2.59185195e-01 4.01737928e-01 -2.89200008e-01
-4.72768903e-01 -3.05034786e-01 -2.75389463e-01 -5.10831535e-01
-2.93948054e-01 -5.63912809e-01 -1.04347849e+00 -4.29527938e-01
-4.83879484e-02 1.67963311e-01 8.14215839e-01 9.58792448e-01
-1.66873094e-02 6.96794987e-01 7.77635872e-01 -1.42132974e+00
-1.26507449e+00 -6.50880158e-01 -7.79727340e-01 6.74377382e-02
1.09057283e+00 -5.11255503e-01 -9.81759965e-01 -1.12426549e-01] | [7.212666034698486, 3.548041582107544] |
b57e59cf-9479-4351-835c-734733f52502 | learning-to-recognize-3d-human-action-from-a | 1812.1055 | null | http://arxiv.org/abs/1812.10550v1 | http://arxiv.org/pdf/1812.10550v1.pdf | Learning to Recognize 3D Human Action from A New Skeleton-based Representation Using Deep Convolutional Neural Networks | Recognizing human actions in untrimmed videos is an important challenging
task. An effective 3D motion representation and a powerful learning model are
two key factors influencing recognition performance. In this paper we introduce
a new skeleton-based representation for 3D action recognition in videos. The
key idea of the proposed representation is to transform 3D joint coordinates of
the human body carried in skeleton sequences into RGB images via a color
encoding process. By normalizing the 3D joint coordinates and dividing each
skeleton frame into five parts, where the joints are concatenated according to
the order of their physical connections, the color-coded representation is able
to represent spatio-temporal evolutions of complex 3D motions, independently of
the length of each sequence. We then design and train different Deep
Convolutional Neural Networks (D-CNNs) based on the Residual Network
architecture (ResNet) on the obtained image-based representations to learn 3D
motion features and classify them into classes. Our method is evaluated on two
widely used action recognition benchmarks: MSR Action3D and NTU-RGB+D, a very
large-scale dataset for 3D human action recognition. The experimental results
demonstrate that the proposed method outperforms previous state-of-the-art
approaches whilst requiring less computation for training and prediction. | ['Sergio A. Velastin', 'Pablo Zegers', 'Huy-Hieu Pham', 'Alain Crouzil', 'Louahdi Khoudour'] | 2018-12-26 | null | null | null | null | ['3d-human-action-recognition'] | ['computer-vision'] | [ 3.82283449e-01 -3.74822378e-01 -3.82590085e-01 -2.65944172e-02
-1.90911844e-01 -1.82926789e-01 6.24706626e-01 -7.01447308e-01
-6.60158098e-01 2.75509626e-01 4.10017610e-01 -4.54245508e-02
1.75828710e-01 -4.52423781e-01 -8.11361015e-01 -7.31924474e-01
-2.10498959e-01 1.76479742e-01 5.47332168e-01 1.11651458e-02
2.71684259e-01 1.05834460e+00 -1.60929191e+00 2.43377388e-01
2.33303413e-01 1.21094835e+00 -3.39878380e-01 1.05330861e+00
7.79870749e-02 1.28635621e+00 -5.84621131e-01 4.22698781e-02
5.20364463e-01 -7.52477169e-01 -8.08616519e-01 6.69300854e-01
3.65694135e-01 -4.74193215e-01 -8.16203117e-01 6.66578829e-01
3.30124259e-01 4.28244323e-01 8.20381522e-01 -1.05269504e+00
-5.01737833e-01 -7.84888417e-02 -6.24560595e-01 3.95085186e-01
6.81060076e-01 2.06176639e-01 6.50848627e-01 -5.66172838e-01
8.00106287e-01 1.48280287e+00 3.98723334e-01 6.76925540e-01
-8.23812962e-01 -2.12950945e-01 6.21310063e-02 7.37367868e-01
-1.05753171e+00 -2.04766840e-01 8.37745428e-01 -4.92581606e-01
1.33277452e+00 1.94644108e-02 1.22144413e+00 1.27517354e+00
4.59609568e-01 1.25598621e+00 7.15858698e-01 -3.63843888e-01
1.00788571e-01 -8.82847190e-01 -1.81662843e-01 8.71312559e-01
-6.60932958e-02 -6.21238574e-02 -5.96989572e-01 2.99429476e-01
1.23293042e+00 4.40917462e-01 -2.27636859e-01 -8.13101768e-01
-1.48233664e+00 4.69984829e-01 4.85096455e-01 3.48281175e-01
-6.16013885e-01 5.81861675e-01 5.81410110e-01 1.77373767e-01
3.15655738e-01 -1.66940853e-01 -2.59522021e-01 -5.77368021e-01
-6.73816383e-01 1.24451801e-01 3.52456063e-01 5.00674069e-01
3.45610052e-01 2.97088683e-01 -1.40697852e-01 6.26989007e-01
2.90282577e-01 5.48566699e-01 9.65737283e-01 -1.02059877e+00
4.21935499e-01 9.71842051e-01 3.15540433e-02 -1.08233464e+00
-5.41906953e-01 6.21504188e-02 -9.44651008e-01 4.84341085e-01
5.97313821e-01 2.45259225e-01 -1.09713757e+00 1.25732112e+00
3.89614731e-01 3.22942227e-01 5.49479313e-02 1.10787833e+00
5.34980237e-01 6.73468530e-01 5.81696257e-02 1.69780359e-01
1.13599443e+00 -1.19274688e+00 -4.97439176e-01 -3.42193134e-02
6.12646401e-01 -3.69125754e-01 6.65768683e-01 2.70669460e-01
-1.10258925e+00 -9.34244275e-01 -1.12207997e+00 -9.89206955e-02
-3.19615841e-01 3.80959600e-01 3.73245835e-01 3.76897097e-01
-8.36378634e-01 7.35001504e-01 -1.37118602e+00 -3.39782804e-01
3.73874038e-01 1.85282812e-01 -7.39283681e-01 1.12372264e-02
-1.05382884e+00 8.61416101e-01 3.27113599e-01 3.86434406e-01
-9.71957803e-01 7.94793367e-02 -1.05746984e+00 -2.33073696e-01
1.61173195e-01 -7.05466032e-01 9.34606194e-01 -1.06877291e+00
-1.75499094e+00 1.11665416e+00 1.62198335e-01 -5.12406528e-01
6.28246903e-01 -4.16942358e-01 -2.55410731e-01 5.20530879e-01
-2.07214519e-01 5.47472477e-01 1.24103761e+00 -5.72988987e-01
-5.70218563e-01 -6.56945884e-01 -2.78395135e-02 3.56104791e-01
-1.58736855e-01 5.92389423e-03 -7.89935350e-01 -8.58937144e-01
2.97209978e-01 -1.05801058e+00 -2.18104452e-01 3.91552061e-01
-1.00690849e-01 -2.42129266e-01 8.52738500e-01 -7.85186827e-01
8.60675395e-01 -2.03037953e+00 9.07918394e-01 -1.38990313e-01
-1.89058453e-01 5.05080819e-01 -2.26192296e-01 1.02297604e-01
-2.44012550e-01 -4.40152079e-01 -2.11185217e-01 -1.45628497e-01
-5.52602187e-02 4.37520504e-01 1.56773627e-01 6.88202262e-01
3.33357811e-01 1.01809776e+00 -7.70890176e-01 -4.20090288e-01
6.63234115e-01 7.82049000e-01 -3.28370035e-01 4.11304414e-01
-5.16849868e-02 5.94614208e-01 -5.41486859e-01 7.19751656e-01
2.19150916e-01 -8.51007476e-02 3.26030776e-02 -2.43183389e-01
2.22885922e-01 -8.57915431e-02 -1.26437998e+00 2.05717134e+00
-1.88683331e-01 4.38154727e-01 -4.01503652e-01 -1.43518519e+00
1.03555477e+00 1.62744582e-01 8.82232130e-01 -6.27076089e-01
2.62282073e-01 1.26503125e-01 -2.90542305e-01 -7.05754340e-01
2.51619875e-01 1.92771316e-01 -1.49185479e-01 5.74423194e-01
1.98113665e-01 1.17361043e-02 2.20334396e-01 -2.68972963e-01
1.24191463e+00 7.45113671e-01 4.68493819e-01 3.95919561e-01
1.01944637e+00 -2.69119084e-01 4.55220550e-01 2.66043872e-01
-5.20396650e-01 5.02285242e-01 6.39085948e-01 -9.21791255e-01
-9.62617397e-01 -7.99948335e-01 5.01561344e-01 8.36670160e-01
1.26382103e-03 1.79646872e-02 -5.87110698e-01 -1.01782513e+00
-5.24994172e-02 1.81061551e-01 -8.94866109e-01 -3.54267359e-01
-9.38474953e-01 -3.53952885e-01 5.48421979e-01 9.16569054e-01
7.87670016e-01 -1.31268394e+00 -1.36204827e+00 1.48480475e-01
-2.99016219e-02 -1.16555262e+00 -3.10359627e-01 -1.10921822e-03
-1.17539144e+00 -1.37210739e+00 -1.27728522e+00 -6.44155622e-01
5.55889249e-01 2.96787232e-01 6.84004009e-01 -4.72005829e-02
-4.78118777e-01 7.05533981e-01 -6.95182502e-01 3.41100633e-01
-3.49520057e-01 -3.70178491e-01 3.34051400e-02 4.41772521e-01
4.43679571e-01 -6.08804226e-01 -7.27825224e-01 4.12116259e-01
-1.09729218e+00 -8.10096264e-02 7.54517615e-01 8.47113073e-01
7.37383842e-01 -2.23958880e-01 -6.91761002e-02 -1.40545756e-01
9.50364098e-02 -4.81962692e-04 -3.08425963e-01 2.98222929e-01
2.85469025e-01 2.45314553e-01 4.06617284e-01 -5.45215607e-01
-7.99844444e-01 7.21943378e-01 -1.70156181e-01 -9.04936910e-01
-4.32633132e-01 -1.72250774e-02 1.40557275e-03 -8.41884390e-02
4.62665647e-01 5.48612773e-01 1.07907429e-01 -5.65428913e-01
4.76543039e-01 3.78721416e-01 7.01472759e-01 -1.96772024e-01
5.12037992e-01 5.63660383e-01 2.76590288e-01 -8.60540926e-01
-6.47261083e-01 -4.23805386e-01 -1.38822973e+00 -6.86822236e-01
1.44897115e+00 -8.41661870e-01 -4.97490823e-01 1.04146898e+00
-1.04394829e+00 -3.54843318e-01 -3.00715655e-01 7.48817265e-01
-1.09928501e+00 7.75857210e-01 -6.70347512e-01 -6.29715025e-01
-2.03099772e-01 -1.19581974e+00 1.23455858e+00 6.97288513e-02
-4.59126420e-02 -8.15170705e-01 2.74117202e-01 4.99495953e-01
-5.16392961e-02 7.07913756e-01 7.81032860e-01 -2.50137001e-01
-3.30571294e-01 -5.08915365e-01 2.07210314e-02 6.74196124e-01
1.49337456e-01 7.85423368e-02 -4.93154049e-01 5.00046387e-02
-7.61822537e-02 -5.50606787e-01 1.01166606e+00 3.86786073e-01
1.16309416e+00 -2.55182479e-02 -1.19619325e-01 6.40417159e-01
9.63320255e-01 2.16420278e-01 8.32080364e-01 4.23723906e-01
8.60158384e-01 2.43551686e-01 4.58679706e-01 6.38297677e-01
-1.60857588e-02 9.06339169e-01 4.86096650e-01 6.06882982e-02
-2.74219215e-01 -1.93044901e-01 7.61526883e-01 8.23886096e-01
-8.46593022e-01 2.98998207e-01 -7.06613243e-01 9.43752676e-02
-1.93250835e+00 -1.11215878e+00 1.51044965e-01 1.99190986e+00
4.67545360e-01 1.92240924e-01 3.95887882e-01 5.33524334e-01
5.34157038e-01 5.12803793e-01 -6.53763235e-01 -2.82359600e-01
-4.56654020e-02 3.52096766e-01 2.85342783e-01 6.81920573e-02
-1.43432605e+00 8.53386223e-01 5.62294769e+00 5.66115081e-01
-1.06503308e+00 -2.96914905e-01 3.15615028e-01 -6.92363381e-02
5.82522571e-01 -5.24550140e-01 -3.53471726e-01 7.89026171e-02
5.81484854e-01 3.17889959e-01 6.66426793e-02 9.49548900e-01
1.59800917e-01 -6.87561482e-02 -1.14318585e+00 1.17239678e+00
3.75619739e-01 -1.10911620e+00 3.18981200e-01 -2.13547930e-01
6.77518606e-01 -1.97588027e-01 -1.80129379e-01 1.46032214e-01
-3.32444794e-02 -9.28091228e-01 8.44563723e-01 8.49891365e-01
5.77347815e-01 -6.79471970e-01 5.44756532e-01 1.79273710e-01
-1.30669689e+00 -2.57606953e-01 -4.07620907e-01 -1.05780348e-01
2.70121634e-01 -3.67283337e-02 -1.82578892e-01 5.23225427e-01
8.13865423e-01 1.52957594e+00 -6.64840996e-01 8.67904246e-01
-5.34169436e-01 2.74201840e-01 1.04521699e-01 -6.93972856e-02
5.09148180e-01 -2.10726529e-01 4.29139644e-01 1.09816241e+00
3.60814780e-01 1.66650444e-01 -7.06363022e-02 4.94158536e-01
1.81747824e-01 -1.70720667e-01 -6.17667556e-01 -2.43650451e-01
-3.27848703e-01 9.21720386e-01 -7.80528247e-01 -3.66309732e-01
-5.28495908e-01 1.48982620e+00 1.50382698e-01 3.08099270e-01
-8.02566588e-01 -2.34138295e-01 7.27318168e-01 -2.47961774e-01
8.58540773e-01 -7.20128953e-01 3.95785064e-01 -1.29196179e+00
1.81195326e-02 -9.06215250e-01 5.40499568e-01 -8.65388989e-01
-8.58840525e-01 4.05674368e-01 -1.76724009e-02 -1.65356612e+00
-5.39948583e-01 -1.09820151e+00 -2.91083544e-01 2.73378760e-01
-1.16421795e+00 -9.68969882e-01 -3.10863435e-01 1.08308578e+00
7.12472439e-01 -2.53102094e-01 7.63883829e-01 1.30563036e-01
-5.78892648e-01 1.68333396e-01 3.61109972e-02 4.72058862e-01
2.66399324e-01 -9.57645833e-01 5.96719742e-01 7.42422402e-01
4.71602231e-01 -5.61092608e-02 4.54423539e-02 -4.06308830e-01
-1.75577855e+00 -8.47817779e-01 5.89699864e-01 -4.34439957e-01
5.50942063e-01 -7.72147179e-02 -5.72209835e-01 6.39937103e-01
-2.11766109e-01 2.35544741e-01 4.35687333e-01 -6.58994853e-01
-3.87821734e-01 1.20505709e-02 -7.41600454e-01 5.11864483e-01
1.39105785e+00 -3.35833400e-01 -9.28878367e-01 2.29371548e-01
2.06722036e-01 -4.18702811e-01 -8.95340502e-01 3.48812997e-01
8.10429335e-01 -1.19292581e+00 1.22294569e+00 -9.07455862e-01
5.95943034e-01 -3.57130587e-01 -1.53396204e-01 -1.03013480e+00
-1.82180256e-01 -2.69333690e-01 -5.29962540e-01 2.78643996e-01
-2.81597942e-01 -6.49381950e-02 9.85396683e-01 1.49902374e-01
3.45047861e-02 -7.42483795e-01 -1.15982985e+00 -7.37264395e-01
-2.68816948e-01 -4.18921381e-01 2.29036555e-01 4.87888545e-01
-2.41039082e-01 -1.02303317e-02 -6.10694230e-01 -3.00374985e-01
5.24392664e-01 1.34841919e-01 9.84236062e-01 -8.52001488e-01
-2.69094497e-01 -5.54077804e-01 -1.37428534e+00 -1.46497142e+00
3.16676974e-01 -6.05262637e-01 -1.42354608e-01 -1.37974858e+00
-5.16386516e-02 3.10509533e-01 -2.87589639e-01 5.52726984e-01
1.16467096e-01 4.62837785e-01 3.03557813e-01 2.02344105e-01
-7.61122823e-01 9.26207900e-01 1.51736379e+00 -3.64942908e-01
-6.63603097e-02 2.15697795e-01 4.38584775e-01 7.58986056e-01
4.28606093e-01 -6.55581206e-02 -1.23876221e-01 -3.80424172e-01
-2.58655578e-01 1.80820808e-01 6.09351575e-01 -1.41819406e+00
-8.19287598e-02 -4.01378758e-02 8.83308649e-01 -7.50662088e-01
5.65734148e-01 -9.24056947e-01 1.31805807e-01 9.13804770e-01
-4.21795279e-01 9.43686627e-03 -4.25533988e-02 6.82063997e-01
-3.81635875e-01 -2.09236406e-02 6.78541720e-01 -3.67027611e-01
-1.20380843e+00 3.21748078e-01 -6.29270911e-01 -1.40525773e-01
1.15209603e+00 -5.61324179e-01 3.51286292e-01 -2.98161447e-01
-9.69432175e-01 -3.11519086e-01 3.32157403e-01 6.49097800e-01
9.76479411e-01 -1.69596744e+00 -4.54668671e-01 3.27560961e-01
-1.96762057e-03 -2.61856139e-01 3.45702291e-01 8.17753732e-01
-9.21949744e-01 4.88828272e-01 -1.02492690e+00 -7.78314650e-01
-1.28543079e+00 4.89324659e-01 5.92441022e-01 -2.92490423e-01
-8.92356038e-01 6.83588624e-01 -2.38569826e-01 -2.67439097e-01
6.02549791e-01 -5.47794044e-01 -4.05545264e-01 -1.09238185e-01
5.28771579e-01 5.90348840e-01 -2.31438890e-01 -1.12198877e+00
-4.01867956e-01 1.10369647e+00 3.92673284e-01 9.18814316e-02
1.40150380e+00 1.77214608e-01 -5.43171912e-03 3.64146024e-01
1.43034494e+00 -8.71748447e-01 -1.64290643e+00 -2.06976309e-01
-5.33407964e-02 -6.29482806e-01 -3.23460639e-01 -1.53949216e-01
-1.45675385e+00 1.12793577e+00 6.98781848e-01 -1.64622322e-01
1.23317301e+00 -1.13449330e-02 7.54347146e-01 6.14073515e-01
1.78399071e-01 -1.08148265e+00 6.90624475e-01 5.53575277e-01
1.05092740e+00 -8.84541392e-01 1.85857430e-01 1.46669045e-01
-7.22470641e-01 1.52749372e+00 5.19664943e-01 -4.03778076e-01
5.58218539e-01 -4.64664251e-01 5.03427163e-03 -1.56810507e-01
-3.19552153e-01 -2.54636139e-01 5.61767280e-01 6.97605550e-01
2.98482090e-01 -1.93789750e-01 -2.73160011e-01 2.40258485e-01
3.32856625e-01 5.79933189e-02 2.33115181e-01 1.11218905e+00
-3.29883248e-01 -9.89151299e-01 -3.97080123e-01 -6.44831881e-02
-1.46099865e-01 6.80680811e-01 -5.97322822e-01 8.43999088e-01
7.01179504e-02 4.61707175e-01 2.38505714e-02 -5.31899571e-01
6.20110929e-01 2.83189446e-01 8.11809659e-01 -2.08255947e-01
-2.61476666e-01 -2.01629214e-02 -1.22551493e-01 -1.12921000e+00
-1.17194355e+00 -7.39590764e-01 -1.22911489e+00 1.32583484e-01
2.60861844e-01 -4.35743034e-01 5.04593015e-01 8.46940100e-01
1.60841063e-01 4.55045849e-01 6.72695696e-01 -1.50013101e+00
-6.85281098e-01 -7.97885478e-01 -6.35643363e-01 8.47078860e-01
2.14389920e-01 -1.06117845e+00 -2.22619995e-01 3.81912708e-01] | [7.858481407165527, 0.38466569781303406] |
5c6f0b43-53db-4bfc-9cb0-a1ccaf67d705 | human-centric-spatio-temporal-video-grounding | 2011.05049 | null | https://arxiv.org/abs/2011.05049v2 | https://arxiv.org/pdf/2011.05049v2.pdf | Human-centric Spatio-Temporal Video Grounding With Visual Transformers | In this work, we introduce a novel task - Humancentric Spatio-Temporal Video Grounding (HC-STVG). Unlike the existing referring expression tasks in images or videos, by focusing on humans, HC-STVG aims to localize a spatiotemporal tube of the target person from an untrimmed video based on a given textural description. This task is useful, especially for healthcare and security-related applications, where the surveillance videos can be extremely long but only a specific person during a specific period of time is concerned. HC-STVG is a video grounding task that requires both spatial (where) and temporal (when) localization. Unfortunately, the existing grounding methods cannot handle this task well. We tackle this task by proposing an effective baseline method named Spatio-Temporal Grounding with Visual Transformers (STGVT), which utilizes Visual Transformers to extract cross-modal representations for video-sentence matching and temporal localization. To facilitate this task, we also contribute an HC-STVG dataset consisting of 5,660 video-sentence pairs on complex multi-person scenes. Specifically, each video lasts for 20 seconds, pairing with a natural query sentence with an average of 17.25 words. Extensive experiments are conducted on this dataset, demonstrating the newly-proposed method outperforms the existing baseline methods. | ['Dong Xu', 'Qian Yu', 'Hongxu Jiang', 'Xiaojie Jin', 'Guanbin Li', 'Si Liu', 'Yue Liao', 'Zongheng Tang'] | 2020-11-10 | null | null | null | null | ['video-grounding', 'spatio-temporal-video-grounding'] | ['computer-vision', 'computer-vision'] | [ 4.63041104e-02 -4.01387602e-01 -1.84677899e-01 -3.09940845e-01
-9.32510734e-01 -3.87003690e-01 5.52851260e-01 1.73302472e-01
-4.46468830e-01 4.93019462e-01 3.68894279e-01 8.14791024e-02
1.73152000e-01 -4.72521007e-01 -7.22730219e-01 -5.53867817e-01
-6.73289225e-02 -1.62360460e-01 3.50165784e-01 -3.15259956e-02
6.68674782e-02 -5.38023748e-02 -1.28349650e+00 5.81688523e-01
3.51551741e-01 1.28778327e+00 3.30258042e-01 3.75396848e-01
1.79454535e-01 9.75186408e-01 -6.04477167e-01 -5.01965284e-01
3.12360786e-02 -5.44573247e-01 -6.63773954e-01 3.19899887e-01
7.04269886e-01 -4.12107885e-01 -6.58513963e-01 1.03614950e+00
4.17104274e-01 5.48292577e-01 3.33271831e-01 -1.49249625e+00
-5.63624620e-01 1.86923042e-01 -6.77706420e-01 5.03034234e-01
1.08431733e+00 2.43153661e-01 7.25852966e-01 -7.86548555e-01
7.82784283e-01 1.26691639e+00 5.48781097e-01 4.14549649e-01
-7.14548290e-01 -6.13019168e-01 4.47260588e-01 4.43668306e-01
-1.76147938e+00 -3.75979245e-01 7.84748912e-01 -5.94589889e-01
8.16701829e-01 1.56981722e-01 8.51078212e-01 1.35673726e+00
1.15653194e-01 7.47357428e-01 7.69479930e-01 -1.00185990e-01
3.91428322e-02 -1.53063461e-01 -1.79408684e-01 7.42365658e-01
-1.94429636e-01 -1.93787724e-01 -9.20604706e-01 -3.32491696e-02
7.05040693e-01 2.88640529e-01 -6.79158092e-01 -2.23226562e-01
-1.80820322e+00 6.70980632e-01 5.10343313e-01 6.15913928e-01
-3.85182053e-01 2.17594504e-01 6.93504274e-01 1.33078024e-01
4.71647084e-01 1.57891717e-02 1.62141129e-01 -1.94922701e-01
-1.18194807e+00 3.71524692e-01 2.46113464e-01 1.21983635e+00
5.60789883e-01 -1.38125181e-01 -7.40449369e-01 4.06136066e-01
2.36629099e-02 5.49275577e-01 4.08977002e-01 -7.91286707e-01
7.83526897e-01 4.80241269e-01 3.30117881e-01 -1.64396846e+00
-2.54615754e-01 -1.55318022e-01 -9.15736616e-01 -7.06249833e-01
2.38355204e-01 5.98953702e-02 -4.19790387e-01 1.80892527e+00
2.83097446e-01 6.13612771e-01 -7.25563914e-02 1.34296215e+00
1.19951212e+00 8.47834826e-01 2.89405555e-01 -5.51379263e-01
1.59135032e+00 -9.40451622e-01 -1.02339494e+00 -2.51015931e-01
6.47914171e-01 -4.38847929e-01 1.09339130e+00 5.15267700e-02
-9.34181213e-01 -5.25532603e-01 -7.98808753e-01 -1.35380467e-02
-3.35224450e-01 1.68322533e-01 2.35025436e-01 1.29275441e-01
-9.69358742e-01 6.34564459e-02 -4.83545512e-01 -7.54327476e-01
1.54197052e-01 -1.10323258e-01 -6.87998116e-01 -2.65446812e-01
-1.43989813e+00 6.25416517e-01 3.80745143e-01 3.23851109e-01
-9.01914775e-01 -2.53838450e-01 -1.17824090e+00 -3.61586899e-01
5.47728002e-01 -8.20721090e-01 9.46294904e-01 -8.94843102e-01
-8.77659261e-01 1.18157887e+00 -5.90434551e-01 -4.58347559e-01
6.01594388e-01 -2.23558173e-01 -6.02975607e-01 7.32384086e-01
4.88747507e-01 6.51456058e-01 1.00886428e+00 -8.55791867e-01
-3.34036738e-01 -2.49910593e-01 3.38614553e-01 2.78833985e-01
-4.08795953e-01 3.48785192e-01 -8.82620096e-01 -9.69026446e-01
8.35215002e-02 -7.03459740e-01 2.04491943e-01 7.62296990e-02
-4.48046178e-01 -2.17275709e-01 1.04223406e+00 -8.70580077e-01
1.38797379e+00 -2.18786478e+00 3.83883029e-01 -3.09234858e-01
1.35079131e-01 -1.46152321e-02 -6.03774190e-03 5.40747344e-01
7.60431886e-02 -8.70264396e-02 -3.95806730e-02 -4.16644901e-01
-2.07758233e-01 -1.06393166e-01 -3.51019442e-01 7.12295234e-01
-1.10872224e-01 1.13141096e+00 -1.14764881e+00 -1.02651381e+00
2.52522856e-01 4.24557477e-01 -1.88474178e-01 4.37368751e-01
-6.22803010e-02 6.32546306e-01 -4.75226462e-01 7.91684449e-01
3.03463787e-01 -3.15390617e-01 -1.91004112e-01 -6.01783812e-01
4.16095778e-02 -3.22742283e-01 -7.05518663e-01 2.29043627e+00
-1.87029868e-01 7.95978785e-01 -2.57688046e-01 -9.87701416e-01
6.47969663e-01 5.58429480e-01 7.45114505e-01 -7.17543185e-01
1.67173624e-01 -1.32216662e-01 -6.65375710e-01 -1.09349656e+00
4.70929861e-01 -4.62210178e-02 -3.16279382e-01 3.18649441e-01
3.97945605e-02 2.32034102e-01 3.18543881e-01 4.91588712e-01
9.89949107e-01 1.25997066e-01 3.22713226e-01 6.19760565e-02
6.87174201e-01 -3.45070884e-02 5.34265995e-01 4.86980319e-01
-6.44747078e-01 7.10000932e-01 3.43536377e-01 -4.08193946e-01
-6.45807981e-01 -9.51774240e-01 2.69048214e-01 9.68933702e-01
6.94754124e-01 -7.89802551e-01 -7.81845391e-01 -6.73437297e-01
-2.92366505e-01 4.01885360e-01 -6.36823475e-01 -1.68849662e-01
-5.92624366e-01 -1.98387757e-01 4.38351721e-01 3.96226197e-01
9.73245442e-01 -1.03822231e+00 -8.01156878e-01 -3.12490650e-02
-1.11545503e+00 -1.78479123e+00 -1.07187068e+00 -7.48830259e-01
-4.79455948e-01 -1.14824975e+00 -1.17874599e+00 -9.04851019e-01
5.82303822e-01 6.99952066e-01 9.46254134e-01 1.12319574e-01
-1.75848007e-01 7.16029108e-01 -6.09653413e-01 1.43750623e-01
3.26233834e-01 -1.63807169e-01 1.14901915e-01 4.36664939e-01
4.69889611e-01 -1.47472933e-01 -7.68484652e-01 5.34712613e-01
-8.69498968e-01 3.11327904e-01 2.60043859e-01 8.34530830e-01
4.81385171e-01 7.74993154e-04 3.44452947e-01 -1.96719483e-01
3.67396832e-01 -5.74780226e-01 -1.73438117e-01 4.47897434e-01
3.60848010e-01 -5.80183864e-01 3.22436452e-01 -6.25244558e-01
-6.62051201e-01 -5.04234061e-02 2.45901138e-01 -9.70229626e-01
1.11073507e-02 4.78203386e-01 -1.98408708e-01 -1.07042007e-02
3.16472530e-01 4.37562525e-01 -2.29495481e-01 3.86858694e-02
1.43860757e-01 5.27315557e-01 9.18838739e-01 -4.45785284e-01
6.70176387e-01 4.85738516e-01 -1.38509721e-01 -9.57827389e-01
-9.52881575e-01 -8.07222903e-01 -5.79386294e-01 -7.65993178e-01
1.29516792e+00 -1.16349065e+00 -8.78221810e-01 4.10791755e-01
-1.37437403e+00 -1.09925039e-01 1.89226225e-01 5.90965867e-01
-7.06651151e-01 6.39395177e-01 -2.97521979e-01 -7.55324543e-01
-1.21236078e-01 -1.05641091e+00 1.51738620e+00 -6.87237158e-02
-1.53022528e-01 -9.01383042e-01 -3.04399371e-01 5.71819186e-01
1.50212035e-01 6.19613290e-01 2.62900174e-01 -3.05354685e-01
-6.11096025e-01 -2.20034197e-01 -3.64730865e-01 -8.81867856e-02
1.47255614e-01 -4.46344763e-01 -7.14419723e-01 -4.60054547e-01
1.17035456e-01 -4.45785850e-01 7.49654889e-01 3.93395901e-01
1.14759994e+00 -3.98690879e-01 -4.39072162e-01 5.41868985e-01
1.13897228e+00 1.37408718e-01 5.11015177e-01 4.29566920e-01
7.51317441e-01 8.25133562e-01 1.15159714e+00 4.03493404e-01
5.70986807e-01 1.13087857e+00 4.63902116e-01 -3.78559418e-02
1.26322210e-01 -3.96010697e-01 4.45522666e-01 6.37184978e-01
-1.56669348e-01 -4.22752678e-01 -8.59706879e-01 6.01046503e-01
-2.18036079e+00 -1.35602379e+00 1.20867111e-01 2.04713058e+00
4.76225436e-01 -2.39289224e-01 2.90411472e-01 1.86191185e-03
1.03100169e+00 5.61968803e-01 -3.65312010e-01 3.88813138e-01
-2.65093029e-01 -5.77701807e-01 3.10565494e-02 1.32004589e-01
-1.47381246e+00 8.39891136e-01 5.25354195e+00 7.73700237e-01
-1.11968446e+00 3.97797883e-01 5.12630820e-01 -1.99251339e-01
1.29583076e-01 -3.20321530e-01 -4.38105553e-01 7.07806110e-01
5.52327931e-01 -3.06647778e-01 1.80995286e-01 5.82566440e-01
5.85793734e-01 -1.91149622e-01 -1.09049869e+00 1.70690668e+00
6.05946600e-01 -1.13020837e+00 8.28172192e-02 -4.21694517e-01
4.69216734e-01 -6.13982439e-01 8.03245306e-02 1.63954794e-01
-6.20540321e-01 -9.78073597e-01 9.94913399e-01 6.42910302e-01
9.81978238e-01 -5.33403397e-01 8.13433230e-01 2.15245143e-01
-1.72116590e+00 1.82854943e-02 -1.11013226e-01 1.95923597e-01
5.60307562e-01 2.10361987e-01 -1.77639320e-01 8.02377999e-01
9.16013002e-01 1.23410821e+00 -5.41650414e-01 9.30142522e-01
-8.55047852e-02 1.74782827e-01 5.52962124e-02 3.37163955e-02
4.08246130e-01 -1.72411613e-02 6.86625898e-01 1.30031824e+00
6.09148562e-01 3.41120005e-01 3.37227166e-01 6.68685257e-01
1.14007935e-01 2.26265073e-01 -9.12396848e-01 7.62224663e-03
3.55868250e-01 1.01599932e+00 -7.23169923e-01 -3.52806956e-01
-4.24491405e-01 1.39912844e+00 8.45347345e-03 7.32093036e-01
-1.23059309e+00 -4.21881080e-01 3.52378964e-01 2.61981279e-01
2.10235566e-01 -2.28436559e-01 6.05538249e-01 -1.51908791e+00
4.28155750e-01 -6.61883950e-01 5.43750763e-01 -1.31153893e+00
-1.09352505e+00 7.21103072e-01 3.06150317e-01 -1.64854026e+00
-3.61413240e-01 -1.31683901e-01 -4.71504509e-01 4.18602347e-01
-1.21663833e+00 -1.56206584e+00 -8.54999006e-01 1.06239438e+00
6.60766006e-01 -2.72987923e-03 3.84406865e-01 4.13163036e-01
-5.28181732e-01 4.69757080e-01 -5.24297953e-01 2.81736702e-01
9.32358086e-01 -6.60323858e-01 2.63551879e-03 9.81786549e-01
1.43670693e-01 6.92952216e-01 7.89756536e-01 -6.50889397e-01
-1.42167890e+00 -1.48523951e+00 1.02138424e+00 -3.84325147e-01
5.86491466e-01 -4.36614573e-01 -7.70225883e-01 7.85222054e-01
3.36482406e-01 3.16758722e-01 3.65372539e-01 -5.18757463e-01
-2.88710505e-01 -7.44791552e-02 -9.59094465e-01 5.15570998e-01
1.33446646e+00 -9.57420588e-01 -6.34375513e-01 5.56848049e-01
8.18105876e-01 -6.36631131e-01 -6.77922130e-01 2.43692517e-01
4.12495166e-01 -1.10202265e+00 1.06935382e+00 -4.28565979e-01
3.80905956e-01 -5.31544685e-01 -4.08145666e-01 -9.61860240e-01
-8.47128257e-02 -7.40596414e-01 -7.94408545e-02 1.25798357e+00
-3.80743861e-01 -2.84549534e-01 4.27212477e-01 1.94368362e-01
-7.09114745e-02 -6.34981811e-01 -1.07188737e+00 -1.02251923e+00
-5.59840143e-01 -4.70126987e-01 4.21002597e-01 9.97957885e-01
1.07600443e-01 6.72608018e-02 -8.60004783e-01 1.74940348e-01
6.03216708e-01 1.84768945e-01 7.59383798e-01 -5.93401194e-01
2.86659934e-02 -8.10963437e-02 -9.52029705e-01 -1.12619460e+00
3.35884511e-01 -5.49628496e-01 2.03495920e-02 -1.49034154e+00
4.21164006e-01 3.77876163e-01 -1.99316069e-01 2.93497562e-01
-1.59684241e-01 4.18038636e-01 3.32420588e-01 1.75361603e-01
-1.26890802e+00 6.41267180e-01 1.15965152e+00 -3.89572054e-01
1.29580528e-01 -2.93389440e-01 -2.33129531e-01 5.62709272e-01
2.93306857e-01 -2.19912469e-01 -5.04848599e-01 -3.10427576e-01
4.42171171e-02 5.40506303e-01 9.21385050e-01 -9.43565726e-01
5.11419475e-01 -1.53561413e-01 1.22246310e-01 -8.19606543e-01
6.80594802e-01 -7.30951488e-01 3.02081078e-01 3.64747375e-01
-3.31999183e-01 3.91156882e-01 4.13292162e-02 8.59057128e-01
-6.80256724e-01 1.81837320e-01 4.58364189e-01 -1.38663173e-01
-1.17311394e+00 4.36715037e-01 -2.89142132e-01 2.73543209e-01
1.60341859e+00 -2.99876422e-01 -4.70505118e-01 -7.54663885e-01
-7.59118676e-01 3.84257853e-01 4.83872801e-01 5.35327315e-01
1.05569530e+00 -1.74053764e+00 -5.11952162e-01 -1.90964386e-01
5.70282102e-01 -2.02481121e-01 7.04372585e-01 1.20820844e+00
-2.67479807e-01 6.90518081e-01 -1.02516696e-01 -8.77100170e-01
-1.43683398e+00 9.62702334e-01 2.63215512e-01 1.88396558e-01
-8.45380127e-01 7.00564206e-01 5.25967598e-01 3.90883595e-01
3.35479409e-01 -2.57727712e-01 -2.98570454e-01 1.55576795e-01
6.89415395e-01 1.58133045e-01 -2.98127532e-01 -1.40623462e+00
-5.36701322e-01 9.49299812e-01 3.33391160e-01 -5.41010983e-02
8.45029891e-01 -4.79201376e-01 -7.56357387e-02 6.76076591e-01
1.36929107e+00 -4.07057166e-01 -9.72883761e-01 -3.86390507e-01
-1.99555740e-01 -8.20406318e-01 -1.27353936e-01 -1.38810337e-01
-1.13290107e+00 9.03617024e-01 3.66391957e-01 -3.30998451e-02
1.30891311e+00 1.50820613e-01 9.73086298e-01 2.46957660e-01
6.96755886e-01 -6.30375206e-01 6.94770515e-01 5.94409555e-02
1.19285429e+00 -1.36492002e+00 -1.05094388e-02 -3.96665663e-01
-8.43575716e-01 9.69684482e-01 6.36616766e-01 2.37979576e-01
3.07872891e-01 -3.75161946e-01 -1.21576600e-01 -3.19795966e-01
-6.15498245e-01 -2.68476665e-01 4.76089090e-01 5.29156864e-01
1.49487391e-01 -2.15537310e-01 -9.12346989e-02 6.28372669e-01
9.19522494e-02 -1.13509912e-02 1.92042232e-01 9.59528089e-01
-1.25590384e-01 -4.80052471e-01 -5.06089270e-01 9.27491710e-02
-3.36429417e-01 4.64487411e-02 -2.82085299e-01 7.09277034e-01
1.61860287e-01 1.20523119e+00 1.35406807e-01 -5.75956762e-01
2.88995624e-01 -2.56127566e-01 3.63640338e-01 -3.99044335e-01
-4.55005765e-01 1.82441249e-01 4.00607064e-02 -9.89911854e-01
-1.03058946e+00 -6.93843484e-01 -1.02749681e+00 -1.46002129e-01
8.06538835e-02 1.32301331e-01 1.80096179e-01 9.20156479e-01
1.06242426e-01 3.53986382e-01 4.35941130e-01 -1.00293124e+00
6.89135939e-02 -7.86390662e-01 -4.97501135e-01 6.69774234e-01
4.79664713e-01 -1.01625896e+00 -2.68280715e-01 4.05757219e-01] | [10.060369491577148, 0.7528606057167053] |
4cc3ad60-ba1b-4977-a688-febcf3c80aa8 | improving-a-sequence-to-sequence-nlp-model | 2212.14117 | null | https://arxiv.org/abs/2212.14117v1 | https://arxiv.org/pdf/2212.14117v1.pdf | Improving a sequence-to-sequence nlp model using a reinforcement learning policy algorithm | Nowadays, the current neural network models of dialogue generation(chatbots) show great promise for generating answers for chatty agents. But they are short-sighted in that they predict utterances one at a time while disregarding their impact on future outcomes. Modelling a dialogue's future direction is critical for generating coherent, interesting dialogues, a need that has led traditional NLP dialogue models that rely on reinforcement learning. In this article, we explain how to combine these objectives by using deep reinforcement learning to predict future rewards in chatbot dialogue. The model simulates conversations between two virtual agents, with policy gradient methods used to reward sequences that exhibit three useful conversational characteristics: the flow of informality, coherence, and simplicity of response (related to forward-looking function). We assess our model based on its diversity, length, and complexity with regard to humans. In dialogue simulation, evaluations demonstrated that the proposed model generates more interactive responses and encourages a more sustained successful conversation. This work commemorates a preliminary step toward developing a neural conversational model based on the long-term success of dialogues. | ['El ouaazizi Aziza', 'Aboulbichr Ahmed', 'Jabri Ismail'] | 2022-12-28 | null | null | null | null | ['policy-gradient-methods'] | ['methodology'] | [-1.77933320e-01 7.62249947e-01 1.97200805e-01 -4.84220266e-01
-2.32751131e-01 -4.37867552e-01 1.13157713e+00 -3.16825621e-02
-3.21944624e-01 1.30179393e+00 6.68956935e-01 -2.34712198e-01
1.31948784e-01 -9.55945849e-01 -8.29870850e-02 -3.45681518e-01
-1.63433865e-01 6.83306336e-01 -1.16129711e-01 -1.02324617e+00
4.96519268e-01 1.63218334e-01 -1.29486489e+00 5.78814864e-01
8.74966979e-01 3.06307584e-01 2.97666043e-01 1.08959079e+00
-4.42570657e-01 1.58530211e+00 -1.04174137e+00 -5.42601705e-01
-3.00024450e-01 -9.85987484e-01 -1.41097319e+00 -9.49820802e-02
-5.85139811e-01 -7.52834618e-01 -2.97451288e-01 3.62398416e-01
5.16312301e-01 3.90042752e-01 7.30862916e-01 -1.28817201e+00
-5.31370521e-01 9.74969447e-01 3.43879491e-01 -5.03024235e-02
8.78157198e-01 6.32217407e-01 1.09908247e+00 -2.53781557e-01
7.75253594e-01 1.49653471e+00 5.22847772e-01 1.21820247e+00
-9.63363826e-01 -1.43045977e-01 3.40990759e-02 1.85262516e-01
-3.94204289e-01 -3.73498917e-01 6.56915963e-01 -3.58744562e-01
1.29853523e+00 3.01575661e-01 9.43837821e-01 1.35616124e+00
1.46002769e-01 9.71652806e-01 1.05225277e+00 -4.64061111e-01
3.78764421e-01 4.97622222e-01 -6.49338439e-02 3.14019173e-01
-7.07669258e-01 2.56020784e-01 -5.50537050e-01 -1.39646798e-01
6.61522329e-01 -5.06964266e-01 -1.11102592e-02 1.63097933e-01
-1.00585282e+00 1.29189467e+00 2.83830911e-01 5.58099866e-01
-6.84482574e-01 -2.40964293e-02 4.41755354e-01 6.38323128e-01
3.40442032e-01 9.76347804e-01 -2.44191721e-01 -7.88394630e-01
-3.40672046e-01 7.11608946e-01 1.42044151e+00 5.96589983e-01
4.63020086e-01 2.46853366e-01 -7.23643780e-01 1.00093532e+00
1.99492767e-01 1.42628923e-01 6.53656542e-01 -1.21024990e+00
1.43505365e-01 5.46233177e-01 4.36515599e-01 -7.27167010e-01
-6.27974689e-01 5.87010197e-02 -5.88964105e-01 2.87740231e-01
5.15688717e-01 -8.39493275e-01 -1.35499954e-01 1.67936099e+00
1.17329307e-01 -4.36999351e-01 5.65308690e-01 7.49705493e-01
1.05150044e+00 9.51678157e-01 1.07417874e-01 -3.33884388e-01
8.72578084e-01 -1.05855858e+00 -9.60150599e-01 1.03568926e-01
7.71836877e-01 -5.52456558e-01 1.13006651e+00 2.23639563e-01
-1.40643692e+00 -4.35765952e-01 -5.07564545e-01 3.04347217e-01
-7.67839327e-02 -3.24704617e-01 7.35961914e-01 7.08417594e-01
-1.28992569e+00 8.30950201e-01 -2.63802677e-01 -4.29605275e-01
-1.74032435e-01 2.77725756e-01 1.95461884e-01 6.39140427e-01
-1.63235569e+00 1.36297059e+00 1.01869501e-01 1.67528484e-02
-7.27606893e-01 -1.09975569e-01 -5.48994839e-01 2.01500729e-01
8.24295953e-02 -6.40501678e-01 1.93301308e+00 -9.48687673e-01
-2.45228195e+00 3.88271302e-01 8.90828148e-02 -7.71250188e-01
7.20940650e-01 4.51565199e-02 5.83685003e-02 1.16555378e-01
-2.35457137e-01 1.01056993e+00 2.50553399e-01 -1.12343323e+00
-6.53684735e-01 2.03558922e-01 6.15305841e-01 7.34144211e-01
-1.94365650e-01 -8.56425017e-02 4.09405440e-01 -1.14616394e-01
-6.91461444e-01 -7.35746264e-01 -4.75874871e-01 -5.24751902e-01
-1.92661256e-01 -7.23980606e-01 3.69642943e-01 -4.41388577e-01
9.68619823e-01 -1.38611436e+00 2.66554598e-02 -2.14612916e-01
1.50277957e-01 3.52117717e-01 -2.47832611e-01 1.24189937e+00
4.84321058e-01 4.24149819e-02 1.46246359e-01 -1.45936996e-01
2.29161054e-01 1.02507629e-01 -2.31556565e-01 -2.12567329e-01
2.73073852e-01 1.02779555e+00 -1.18499720e+00 -3.26749176e-01
2.18720064e-01 1.70505553e-01 -7.48433650e-01 9.86474156e-01
-5.72031975e-01 7.42533684e-01 -4.63249475e-01 -6.74413145e-02
1.10162325e-01 -7.71605819e-02 3.54953259e-01 4.86609697e-01
-3.21644306e-01 6.98655248e-01 -5.23470819e-01 1.04588068e+00
-7.43913531e-01 6.65930629e-01 7.93271735e-02 -7.04165399e-01
1.31517696e+00 6.50001585e-01 2.39252925e-01 -7.82250762e-01
1.31698996e-01 -3.89745948e-03 4.26616907e-01 -7.15246499e-01
8.66605818e-01 -2.78831214e-01 7.57936388e-02 9.75117147e-01
8.35540965e-02 -6.03458226e-01 4.59578246e-01 3.69525492e-01
8.11930776e-01 1.34026438e-01 2.37727746e-01 5.79779707e-02
5.45787930e-01 3.62577178e-02 9.31114778e-02 9.26447332e-01
-2.72151709e-01 1.68798134e-01 9.26737905e-01 -4.57398713e-01
-1.04524100e+00 -5.61677754e-01 5.29552341e-01 1.14823830e+00
-1.92812890e-01 -3.82581577e-02 -9.62107360e-01 -4.75322217e-01
-4.64196116e-01 1.18382239e+00 -4.53115910e-01 -9.63900238e-02
-5.98058164e-01 -4.34921414e-01 5.75877309e-01 7.21548200e-02
3.84952962e-01 -1.98497474e+00 -9.38984454e-01 5.71739197e-01
-5.51983714e-01 -6.06147170e-01 -1.82704180e-01 -1.34671390e-01
-7.17324317e-01 -8.47908854e-01 -9.46338713e-01 -7.32127786e-01
1.58782646e-01 2.73617413e-02 1.17664731e+00 3.13921243e-01
2.23908558e-01 5.16818821e-01 -6.31323516e-01 -4.63253826e-01
-1.29320586e+00 1.22716255e-01 -9.99260396e-02 -3.61254960e-01
4.12110686e-02 -5.40564239e-01 -5.49584985e-01 2.48961076e-01
-5.38994849e-01 3.32692474e-01 2.28634760e-01 1.24787009e+00
-5.69343686e-01 -5.91015518e-01 1.25468528e+00 -8.55664074e-01
1.84558463e+00 -6.45210564e-01 -1.25037163e-01 2.07336918e-01
-5.87266028e-01 1.67603679e-02 7.84687579e-01 -3.24387491e-01
-1.59857261e+00 -2.63580531e-01 -5.44773638e-01 4.91963536e-01
-2.63366431e-01 3.61952722e-01 5.04106939e-01 3.11068535e-01
9.03719723e-01 4.19638216e-01 5.49517810e-01 -1.53629463e-02
4.77729738e-01 8.33977103e-01 4.43185270e-02 -6.86212540e-01
9.02798399e-02 -2.42773980e-01 -5.86937070e-01 -9.62846458e-01
-1.42822325e-01 -1.21968463e-01 -3.39005947e-01 -7.97559261e-01
5.73590159e-01 -4.12453800e-01 -1.56610072e+00 4.94938195e-01
-1.50376678e+00 -7.22699463e-01 -3.29726219e-01 2.62065977e-01
-9.78878021e-01 3.22668016e-01 -8.52005482e-01 -1.46404624e+00
-5.66034377e-01 -9.90064561e-01 3.00651282e-01 5.69720805e-01
-8.04057002e-01 -1.16678500e+00 3.65861833e-01 3.28083307e-01
7.85607815e-01 -6.98939338e-02 8.05275917e-01 -9.06609952e-01
-2.20503360e-01 1.27482295e-01 1.38195306e-01 2.90521085e-01
-2.42810752e-02 -3.74808572e-02 -8.62127900e-01 1.48881644e-01
6.98169023e-02 -9.77075696e-01 1.63299412e-01 1.88360974e-01
4.09965932e-01 -8.48771095e-01 2.37185329e-01 -2.45915905e-01
6.37936294e-01 6.64313257e-01 6.25783622e-01 1.94188803e-01
-9.28437263e-02 1.26322305e+00 6.48772597e-01 8.28695416e-01
6.29623413e-01 5.02922118e-01 5.44618964e-01 2.02906672e-02
2.09185839e-01 -3.44369262e-01 4.07255977e-01 7.25584865e-01
-2.21542731e-01 -4.46533382e-01 -7.12610126e-01 5.53324699e-01
-1.89846873e+00 -1.28950763e+00 -1.20469943e-01 1.64539874e+00
1.02187037e+00 7.43336827e-02 5.24834216e-01 -1.88262731e-01
6.08631015e-01 2.31043711e-01 -3.82377535e-01 -1.22322929e+00
5.94431050e-02 3.28677036e-02 -2.90717065e-01 8.21605384e-01
-3.10875148e-01 1.00771356e+00 6.68295908e+00 2.60330260e-01
-8.72139096e-01 -1.96001500e-01 9.16362286e-01 1.49985418e-01
-4.87690419e-01 -1.60458833e-01 -4.08391684e-01 3.16321522e-01
1.20897925e+00 -4.77070510e-01 6.07150912e-01 7.09195673e-01
6.78786159e-01 -2.46465772e-01 -1.00134969e+00 2.98539579e-01
-2.96323806e-01 -1.61883605e+00 -1.31280702e-02 -1.80768773e-01
5.21188974e-01 -4.31539118e-01 -2.18022317e-01 7.49123871e-01
8.97985935e-01 -1.16430938e+00 4.11077172e-01 6.08786583e-01
1.01012684e-01 -7.54512906e-01 8.46240878e-01 7.90955126e-01
-4.62121904e-01 -1.48987874e-01 -1.78424761e-01 -7.33312428e-01
4.32935923e-01 -2.58682936e-01 -1.90747464e+00 2.12427303e-01
1.02179699e-01 1.24876484e-01 1.16104402e-01 6.36740148e-01
-2.93069839e-01 6.16095960e-01 1.11359037e-01 -1.15512788e+00
5.67723274e-01 -3.35872233e-01 3.61468107e-01 1.21269929e+00
6.94430321e-02 2.44445637e-01 9.26661342e-02 9.91886020e-01
1.16047688e-01 3.29831272e-01 -6.78521156e-01 -8.86427686e-02
6.09631419e-01 1.03968811e+00 -3.62181336e-01 -2.26502985e-01
4.72570583e-02 7.86371768e-01 2.47878313e-01 1.14471756e-01
-5.17748594e-01 -2.26486892e-01 3.84678692e-01 -1.50212184e-01
-1.73652619e-01 1.04846224e-01 -3.53323370e-01 -5.99814832e-01
-4.27482843e-01 -1.18713021e+00 -3.83451842e-02 -7.56028354e-01
-1.13319850e+00 8.66026461e-01 -1.74571633e-01 -8.01568985e-01
-1.20664680e+00 -2.47848988e-01 -1.03662241e+00 1.09927428e+00
-1.04127133e+00 -9.85615313e-01 1.16006499e-02 2.69417077e-01
9.25383151e-01 -3.72896761e-01 1.20848846e+00 -3.27668130e-01
-1.65053397e-01 4.49899882e-01 -1.57034189e-01 -1.30413910e-02
4.31295455e-01 -1.23231864e+00 4.91237134e-01 -4.75980639e-02
-2.55471438e-01 5.00073016e-01 1.08076441e+00 -3.81921023e-01
-8.47038627e-01 -3.71959180e-01 1.13024223e+00 -1.29791632e-01
5.80711663e-01 1.42176021e-02 -7.02823222e-01 1.42530546e-01
8.01223218e-01 -1.06825495e+00 5.18745482e-01 3.29995692e-01
3.76876295e-01 3.59936506e-01 -1.15907669e+00 9.17676806e-01
5.48520148e-01 -2.42279455e-01 -7.18484461e-01 4.44775164e-01
7.25731969e-01 -2.47693658e-01 -6.35078430e-01 -1.07479736e-01
5.87908387e-01 -1.50812852e+00 6.05802059e-01 -7.18127906e-01
8.01674545e-01 4.93117422e-01 4.46773201e-01 -1.72100186e+00
-1.66457109e-02 -1.19890678e+00 9.38480422e-02 1.08866656e+00
4.67480242e-01 -5.14108121e-01 1.01233125e+00 7.61776626e-01
-1.32164061e-01 -8.29300582e-01 -6.10274851e-01 -3.64085406e-01
2.83542246e-01 1.04645625e-01 6.56103671e-01 7.77063370e-01
7.98525691e-01 8.33162904e-01 -8.37473989e-01 -6.29627526e-01
-3.91424708e-02 -1.65968344e-01 1.07274187e+00 -1.12177598e+00
-4.00094241e-01 -6.72203898e-01 2.49397442e-01 -1.19894910e+00
4.89867665e-02 -4.29241687e-01 3.68986279e-01 -1.74477613e+00
-1.32053658e-01 -3.43790948e-01 4.22974050e-01 8.59975442e-02
-1.06063113e-01 -4.94693786e-01 4.39842165e-01 3.93416211e-02
-4.49270308e-01 8.67367446e-01 1.63126624e+00 4.15112190e-02
-5.86723745e-01 5.33203006e-01 -5.61510503e-01 5.95354617e-01
1.18382335e+00 -1.08417273e-01 -5.90817988e-01 3.98947597e-02
6.74309954e-02 9.30219233e-01 -6.50863126e-02 -6.79574966e-01
2.38110170e-01 -4.48832542e-01 -1.98425323e-01 -2.83996463e-01
6.10320389e-01 -5.34831621e-02 -3.21425468e-01 7.15266347e-01
-1.02627873e+00 7.04941228e-02 -1.02169462e-01 3.42658997e-01
-1.96885109e-01 -5.99009216e-01 5.78865826e-01 -5.67469060e-01
-4.32728112e-01 -2.44593859e-01 -1.21238220e+00 -5.09617291e-02
9.09269392e-01 -2.13243529e-01 -1.29304737e-01 -1.44278753e+00
-6.35657847e-01 4.91667122e-01 -1.07869937e-03 4.45754677e-01
7.11808562e-01 -7.67927289e-01 -9.39776242e-01 -3.17852974e-01
-2.51342237e-01 -3.37168545e-01 3.86681497e-01 2.28986517e-01
-6.43764853e-01 5.89480460e-01 -6.85733676e-01 -2.28037596e-01
-1.27582228e+00 1.73067115e-02 4.71117288e-01 -7.68116832e-01
-4.30861682e-01 8.64174366e-01 -9.85479914e-03 -7.96306491e-01
4.37713414e-01 1.08629532e-01 -8.65905881e-01 1.33246392e-01
4.51887786e-01 4.44866717e-01 -3.52991551e-01 -1.65483981e-01
3.20068210e-01 -4.35569346e-01 -1.88892812e-01 -6.19522512e-01
1.19615304e+00 -1.40450001e-01 -4.67322767e-04 4.96725380e-01
4.90092605e-01 -3.40255409e-01 -1.36426866e+00 2.98532937e-02
3.84553596e-02 -1.54889107e-01 -6.24027252e-01 -1.14273107e+00
-3.30776185e-01 8.90919626e-01 1.06540695e-01 8.50283980e-01
5.10649979e-01 -3.07059914e-01 7.76951671e-01 8.03452194e-01
2.88563877e-01 -1.37658906e+00 5.72939575e-01 1.07859170e+00
1.14116633e+00 -1.17512894e+00 -4.25813496e-01 1.03847869e-01
-1.31020772e+00 1.36350060e+00 9.32463944e-01 1.00012207e-02
1.85558833e-02 -1.20907001e-01 3.47223908e-01 -3.89995910e-02
-1.35974634e+00 -5.47904968e-02 -3.75306755e-01 9.21570837e-01
8.79198134e-01 8.95555019e-02 -7.79383361e-01 2.93490350e-01
-6.08501136e-01 -7.11062923e-02 1.04061222e+00 7.35753357e-01
-6.29011154e-01 -1.34669554e+00 -2.02226520e-01 2.35074490e-01
-1.46043941e-01 4.76194620e-02 -8.37253511e-01 6.04235530e-01
-4.95023251e-01 1.50971520e+00 -1.77343160e-01 -3.59253168e-01
2.76351541e-01 1.61900118e-01 2.51168281e-01 -6.06193006e-01
-1.35271525e+00 -4.57698047e-01 8.49040151e-01 -1.17137261e-01
-3.32505375e-01 -4.15114194e-01 -1.15684068e+00 -5.99111080e-01
-1.56157836e-01 6.45885110e-01 7.34804273e-01 9.08662140e-01
1.05959393e-01 5.59383214e-01 1.06125426e+00 -8.18858027e-01
-9.86484170e-01 -1.42027342e+00 -1.67943522e-01 3.64843994e-01
1.31249025e-01 -2.22173959e-01 -1.01488121e-01 -2.29941279e-01] | [12.762392044067383, 8.187705039978027] |
31f34d26-e4e3-4e57-b773-0af131e200ae | facial-emotion-recognition-using-deep | 1910.11113 | null | https://arxiv.org/abs/1910.11113v1 | https://arxiv.org/pdf/1910.11113v1.pdf | Facial Emotion Recognition Using Deep Learning | We aim to construct a system that captures real-world facial images through the front camera on a laptop. The system is capable of processing/recognizing the captured image and predict a result in real-time. In this system, we exploit the power of deep learning technique to learn a facial emotion recognition (FER) model based on a set of labeled facial images. Finally, experiments are conducted to evaluate our model using largely used public database. | ['Li-Heng Chen', 'Ching-Da Wu'] | 2019-10-19 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [ 1.73087671e-01 -2.57196128e-01 1.72334403e-01 -1.23999166e+00
-1.52705327e-01 -1.69745460e-01 3.57565045e-01 -5.34918070e-01
-4.51692313e-01 4.93086755e-01 -3.00967366e-01 2.76778847e-01
4.08931643e-01 -8.69054556e-01 -5.66856146e-01 -3.11475843e-01
-2.50221759e-01 1.07375227e-01 -5.97722948e-01 -9.92862061e-02
1.57506958e-01 9.30528402e-01 -1.90479195e+00 6.93148971e-01
-7.70175755e-02 1.55442083e+00 -1.59296572e-01 7.27281213e-01
1.35986134e-01 9.86325681e-01 -3.68253112e-01 -5.07342637e-01
3.65053147e-01 2.26453736e-01 -8.33384633e-01 5.59262633e-01
2.91732013e-01 -4.74159330e-01 -1.11510180e-01 1.14140654e+00
3.40483844e-01 1.78017505e-02 5.59059560e-01 -1.51836276e+00
-4.70747709e-01 -3.30345660e-01 -5.06209016e-01 1.48819685e-01
6.32418752e-01 4.09162343e-02 2.82324970e-01 -1.20113420e+00
4.86588597e-01 1.30330670e+00 3.82711709e-01 6.08966172e-01
-7.35474229e-01 -9.37453687e-01 -3.09259117e-01 3.51915836e-01
-1.64909470e+00 -8.46019149e-01 8.85697067e-01 -9.72059146e-02
1.06168604e+00 -2.65902914e-02 7.25135028e-01 8.36223006e-01
5.48767805e-01 5.97970307e-01 1.40470695e+00 -4.63238090e-01
5.42831644e-02 2.57228583e-01 1.02600858e-01 1.13503981e+00
-4.86627996e-01 1.42457321e-01 -5.93840122e-01 -1.04994804e-01
5.49931586e-01 3.38297427e-01 1.89354457e-02 5.08279741e-01
-1.29084721e-01 6.31774545e-01 4.42610942e-02 1.68372288e-01
-9.28290963e-01 -7.54861161e-02 4.68872875e-01 7.64829159e-01
6.14255846e-01 -2.88306266e-01 -4.70775783e-01 -9.93632004e-02
-9.89706099e-01 -1.96578637e-01 8.39939654e-01 5.58090985e-01
9.40481246e-01 1.00833207e-01 4.38168764e-01 7.26146877e-01
6.39093101e-01 5.36709666e-01 5.60123444e-01 -6.84686303e-01
-3.13874096e-01 6.88790858e-01 -4.85657491e-02 -1.05086839e+00
-2.92684913e-01 2.72442579e-01 -5.79147100e-01 6.38187170e-01
-3.02253246e-01 -4.29318845e-01 -9.11426187e-01 1.30505729e+00
2.84312725e-01 5.44670522e-01 2.43637547e-01 9.31246519e-01
6.93970144e-01 7.47710109e-01 4.19004977e-01 -2.53795713e-01
1.40841985e+00 -6.32492423e-01 -8.12821686e-01 -1.09086949e-02
6.67337924e-02 -8.22555304e-01 7.42214441e-01 8.67861509e-01
-8.83573771e-01 -8.03590894e-01 -7.20796943e-01 1.48299471e-01
-6.86844885e-01 6.08946383e-01 8.05913210e-01 7.00785875e-01
-1.31483817e+00 2.28797019e-01 -5.90221405e-01 -4.23740476e-01
5.53670883e-01 1.00627220e+00 -9.92045105e-01 1.64566264e-01
-1.01163769e+00 6.99841976e-01 3.26443344e-01 5.00667810e-01
-1.04297650e+00 -7.47326314e-02 -7.51586378e-01 3.76645476e-02
-1.86398149e-01 -1.95030406e-01 1.06148481e+00 -2.03734946e+00
-1.89299560e+00 1.50119460e+00 -2.67766386e-01 -2.76125938e-01
9.79517866e-03 -2.40658402e-01 -1.08624113e+00 3.61650139e-01
-5.13678551e-01 7.96977997e-01 1.25727475e+00 -9.27105904e-01
-5.98388910e-01 -7.28679419e-01 -1.07905932e-01 -1.50283560e-01
-4.15054351e-01 4.69613850e-01 -1.47116244e-01 1.16910962e-02
-3.59122932e-01 -8.45251620e-01 2.50728101e-01 -7.56693408e-02
8.04100409e-02 -3.32595110e-01 1.30844486e+00 -4.57712382e-01
6.40294611e-01 -2.32491899e+00 -4.15401429e-01 4.23449367e-01
-1.16263121e-01 7.08595932e-01 -9.73815694e-02 1.74784034e-01
-5.35798073e-01 -2.69169152e-01 3.96471649e-01 -3.10639858e-01
-4.16074038e-01 3.28902662e-01 -2.71372527e-01 4.33879972e-01
5.77826619e-01 6.42894983e-01 -1.69464827e-01 -6.81704104e-01
3.92734706e-01 7.52554595e-01 -3.87620091e-01 6.88802660e-01
9.14867371e-02 -2.05824040e-02 -3.40144783e-01 9.89617407e-01
9.35564339e-01 2.10727841e-01 7.00329542e-02 -2.71845102e-01
5.58756180e-02 -6.05669498e-01 -1.03825510e+00 1.38370228e+00
-5.41921914e-01 6.56616628e-01 1.52332619e-01 -1.14560175e+00
1.24286413e+00 6.17395639e-01 2.81906545e-01 -4.10205930e-01
8.16451669e-01 -2.07080022e-01 -5.06271601e-01 -1.18863106e+00
2.11364955e-01 -5.87005317e-01 3.25889558e-01 5.35929799e-01
7.06042588e-01 4.77260262e-01 -2.66334563e-01 -5.12500823e-01
8.15397143e-01 -4.68764640e-02 2.40636021e-01 -4.26403247e-02
9.36846912e-01 -2.57718831e-01 2.51866996e-01 9.60639566e-02
-6.38869166e-01 3.33373472e-02 1.75966546e-01 -1.12168074e+00
-5.81893325e-01 -6.70945883e-01 -7.21621364e-02 1.29718816e+00
-3.71362001e-01 2.26473920e-02 -8.76004100e-01 -6.19233549e-01
-1.80320591e-01 8.46034139e-02 -7.77917504e-01 -2.03091979e-01
-7.46145174e-02 -5.71845949e-01 6.80980623e-01 3.36884916e-01
6.65874064e-01 -1.49814773e+00 -8.85495007e-01 -1.27657549e-02
3.15648854e-01 -1.07452035e+00 1.31575376e-01 -1.09601706e-01
-6.45297945e-01 -1.08304799e+00 -2.43433803e-01 -1.11694014e+00
7.51455367e-01 -1.71476658e-02 9.89938021e-01 1.47615388e-01
-4.82642621e-01 5.30257881e-01 -2.53816307e-01 -7.81062067e-01
1.89023446e-02 -5.11016250e-01 1.95477694e-01 8.22897077e-01
1.15690446e+00 -4.38359886e-01 -4.33375716e-01 -1.04271822e-01
-1.09897828e+00 -2.27068648e-01 4.99585390e-01 5.91554701e-01
4.10027921e-01 4.08963829e-01 3.74042690e-01 -7.52905130e-01
6.10889852e-01 -5.72609186e-01 -6.79240823e-01 3.68605793e-01
-2.57988691e-01 -2.81205237e-01 5.98927915e-01 -3.05481344e-01
-1.30888891e+00 5.47188461e-01 -4.43626285e-01 -6.53891742e-01
-6.78512037e-01 5.33442616e-01 -6.34852871e-02 -3.00117224e-01
1.84339985e-01 2.44359925e-01 2.66356140e-01 -2.78994054e-01
-7.57616684e-02 1.32791388e+00 5.59589148e-01 -2.80834675e-01
9.56368744e-02 6.76080108e-01 -1.70264751e-01 -7.11549044e-01
-5.02655625e-01 -1.11120455e-01 -6.99556112e-01 -8.50557685e-01
6.42186940e-01 -1.19436789e+00 -1.31220198e+00 6.60065532e-01
-1.03179300e+00 1.23192832e-01 3.70026469e-01 3.64255816e-01
-5.14485657e-01 -4.25009787e-01 -7.35570073e-01 -1.12098539e+00
-5.79296350e-01 -8.72571647e-01 1.34011149e+00 5.08502185e-01
-8.99701286e-03 -8.33851457e-01 1.18422374e-01 2.92369068e-01
1.33225366e-01 1.85055718e-01 4.07892197e-01 -5.45549273e-01
-1.18323443e-02 -5.78745961e-01 -2.35877931e-01 7.62358963e-01
2.29775533e-01 5.85770726e-01 -1.42444086e+00 3.38859819e-02
1.66960716e-01 -8.35556030e-01 4.72985715e-01 1.22814029e-01
1.52901065e+00 -8.05783048e-02 -5.36281578e-02 6.01464272e-01
1.74415100e+00 3.99698079e-01 7.47915626e-01 -1.51363596e-01
4.94567677e-02 7.27326632e-01 5.96070588e-01 6.19507492e-01
1.87724471e-01 1.18143231e-01 3.28799635e-01 -7.85218552e-02
4.18164819e-01 8.15698877e-03 4.10035491e-01 2.99691737e-01
-2.64693141e-01 -1.41233742e-01 -9.52853739e-01 1.39992731e-02
-1.44398737e+00 -1.17132318e+00 4.34873700e-01 1.45149231e+00
4.93204623e-01 -2.17853040e-01 1.21651992e-01 1.26948744e-01
7.25644350e-01 -1.49840698e-01 -2.20567152e-01 -8.52980912e-01
2.25431606e-01 7.30619788e-01 -1.28283441e-01 4.02197212e-01
-1.20510507e+00 1.27012360e+00 7.46764326e+00 3.94687474e-01
-2.14355087e+00 -9.49547142e-02 1.05826616e+00 -6.45367131e-02
4.84904319e-01 -6.04345560e-01 -4.98994708e-01 2.65907913e-01
1.49072218e+00 -4.42894101e-02 4.53287750e-01 1.23504329e+00
2.71403641e-01 -1.25919074e-01 -9.07822907e-01 1.35910225e+00
4.07718390e-01 -9.29506183e-01 2.67415732e-01 -5.19328006e-02
1.63310647e-01 -1.00683719e-01 2.33429715e-01 3.83403659e-01
-3.00422031e-02 -1.52119851e+00 2.77369291e-01 9.06961501e-01
8.79510939e-01 -1.05062461e+00 8.59960079e-01 2.19902396e-01
-8.20898533e-01 -1.19147323e-01 -4.67327207e-01 -4.57685471e-01
-3.31892490e-01 1.06230840e-01 -9.08533037e-01 -3.32240388e-03
7.93357015e-01 6.02283239e-01 -4.65775639e-01 2.22268492e-01
2.64044344e-01 5.19273102e-01 -2.14537665e-01 1.03179999e-01
1.59615099e-01 -1.95829496e-01 -2.56160229e-01 1.28389490e+00
2.79480755e-01 7.26981878e-01 -2.36467123e-01 4.87624615e-01
-2.99674064e-01 1.77091658e-01 -9.06157792e-01 -1.57578766e-01
-5.31826764e-02 1.65946186e+00 -4.33752269e-01 -4.23300862e-01
-4.75555718e-01 1.07713556e+00 4.34760720e-01 1.26929015e-01
-8.19267929e-01 -1.93052381e-01 6.39361799e-01 -2.14099586e-01
5.04294708e-02 1.16378039e-01 2.64853328e-01 -1.10673046e+00
-1.68692142e-01 -8.70869935e-01 2.93851823e-01 -1.29102135e+00
-1.17423642e+00 1.05386198e+00 -3.96424770e-01 -8.42568040e-01
-1.02735341e-01 -1.04138052e+00 -6.82972372e-01 6.06891453e-01
-1.56682742e+00 -1.27773535e+00 -4.84912187e-01 1.44738483e+00
3.85217071e-01 -6.29598916e-01 1.18932498e+00 5.93705714e-01
-7.83535063e-01 3.59646857e-01 -3.81986797e-01 4.44683492e-01
6.98790669e-01 -7.38380015e-01 -4.15659338e-01 3.21146518e-01
2.22952098e-01 5.49999177e-01 5.33795357e-01 -2.62869656e-01
-1.61891854e+00 -1.08220959e+00 8.39016020e-01 -6.37873560e-02
1.84123039e-01 -2.33862624e-01 -6.33819818e-01 9.18339610e-01
3.90649319e-01 7.12057650e-01 1.23969841e+00 -1.36771187e-01
-2.12009117e-01 -5.58079123e-01 -1.78468287e+00 1.53645864e-02
4.17680860e-01 -7.52768278e-01 -5.19180536e-01 8.29161406e-02
-2.39596933e-01 -1.78666636e-02 -8.07735801e-01 2.72733659e-01
9.69145179e-01 -1.03146493e+00 6.27730668e-01 -9.61657405e-01
5.85745990e-01 1.43442348e-01 -2.39713579e-01 -1.10348630e+00
6.78333715e-02 -3.51656169e-01 1.60576656e-01 1.15715742e+00
-8.55720863e-02 -4.80712116e-01 9.27054703e-01 9.74804878e-01
5.45958519e-01 -6.84864998e-01 -7.34400153e-01 -1.44298673e-01
-4.98450637e-01 -4.55976456e-01 8.36961687e-01 8.87293994e-01
-2.01609619e-02 7.86502212e-02 -4.72598135e-01 2.70124197e-01
3.03727299e-01 -5.55341095e-02 4.66190010e-01 -1.06157327e+00
3.01986217e-01 3.01360220e-01 -9.63199556e-01 -6.97633177e-02
8.88490140e-01 -4.79874879e-01 -2.32918948e-01 -6.31990075e-01
1.58188552e-01 1.74499825e-02 -6.66008174e-01 6.74924195e-01
3.36450964e-01 8.15938056e-01 2.93319754e-04 -1.04029030e-01
-7.79437244e-01 4.67755973e-01 5.48457384e-01 2.08668470e-01
2.09594682e-01 -1.53454378e-01 -3.45153660e-01 1.07154858e+00
8.34950924e-01 -4.47598189e-01 -2.22434744e-01 -4.15126979e-01
-2.98347678e-02 3.74060780e-01 4.63161200e-01 -1.01689744e+00
5.29396869e-02 -2.80763716e-01 1.03765154e+00 -2.36826107e-01
5.14150321e-01 -1.37849414e+00 2.80804157e-01 2.87319213e-01
-2.65434742e-01 2.34636605e-01 2.28626966e-01 1.15434095e-01
-6.76611900e-01 5.75343519e-02 8.96348596e-01 -1.38455302e-01
-1.11207104e+00 4.49957013e-01 -5.14022470e-01 -6.75225019e-01
1.40864372e+00 -8.89232457e-02 8.57657939e-02 -3.38953555e-01
-9.20186162e-01 -1.99464113e-01 1.74841642e-01 4.03291434e-01
9.45388019e-01 -1.38340366e+00 -4.11347061e-01 8.70873034e-01
1.03733316e-01 -8.40570331e-01 2.21791327e-01 1.38201371e-01
-8.38228106e-01 3.53827208e-01 -1.08108115e+00 -2.15386704e-01
-1.86500311e+00 4.04646039e-01 5.72231054e-01 1.69985086e-01
-1.53729543e-01 8.51415694e-01 -3.66656572e-01 -2.04635784e-01
1.07567161e-01 3.36243212e-01 -5.93519926e-01 -1.19308151e-01
9.20995772e-01 -1.20944545e-01 4.67900485e-02 -1.05147624e+00
-4.58234608e-01 4.22331065e-01 1.14180773e-01 -1.79611817e-01
1.69407094e+00 -4.36437279e-02 -3.63235384e-01 1.88341394e-01
1.63071287e+00 -4.64837998e-01 -1.00323749e+00 -6.62370995e-02
-2.48669714e-01 -6.12585783e-01 3.85217369e-01 -7.02867866e-01
-1.62644851e+00 9.93170619e-01 1.35059249e+00 -1.99524388e-01
1.84030068e+00 -5.42318344e-01 7.80982196e-01 5.77302098e-01
4.34078395e-01 -1.25201607e+00 -7.82151520e-03 2.22527713e-01
7.02493906e-01 -1.61853468e+00 -3.03320765e-01 -1.13981217e-01
-7.97258317e-01 1.52701199e+00 6.39805377e-01 -5.77591121e-01
1.35570109e+00 6.10317349e-01 4.27150309e-01 -5.98596334e-01
-1.14954329e+00 5.64270765e-02 3.85137647e-02 4.34154034e-01
4.64309692e-01 7.39828497e-02 1.49955735e-01 5.49290061e-01
-1.26553997e-01 1.00132668e+00 2.57789642e-01 9.98526752e-01
-3.84385437e-01 -8.93904626e-01 -4.57394630e-01 2.97967404e-01
-9.53706503e-01 2.41193384e-01 -4.72962826e-01 4.31077838e-01
4.49497730e-01 1.06491613e+00 2.81962454e-01 -6.35826945e-01
1.37848616e-01 5.75566530e-01 4.61118370e-01 -2.57089257e-01
-6.18972659e-01 -7.42197111e-02 -1.99810326e-01 -6.45315349e-01
-7.79667914e-01 -2.87941277e-01 -9.79430318e-01 -3.77854407e-01
8.82253051e-02 -4.22193669e-02 8.66326034e-01 9.97839808e-01
5.44975519e-01 -3.72221954e-02 9.54769731e-01 -7.25348473e-01
-1.59464195e-01 -7.76519597e-01 -8.54447067e-01 6.09010816e-01
3.74880224e-01 -6.17354572e-01 -3.05775143e-02 6.36625826e-01] | [13.511208534240723, 1.7822997570037842] |
af3ba28d-08bb-481a-b7f3-29e2fb92dd1d | non-asymptotic-performance-of-social-machine | 2306.09397 | null | https://arxiv.org/abs/2306.09397v1 | https://arxiv.org/pdf/2306.09397v1.pdf | Non-Asymptotic Performance of Social Machine Learning Under Limited Data | This paper studies the probability of error associated with the social machine learning framework, which involves an independent training phase followed by a cooperative decision-making phase over a graph. This framework addresses the problem of classifying a stream of unlabeled data in a distributed manner. We consider two kinds of classification tasks with limited observations in the prediction phase, namely, the statistical classification task and the single-sample classification task. For each task, we describe the distributed learning rule and analyze the probability of error accordingly. To do so, we first introduce a stronger consistent training condition that involves the margin distributions generated by the trained classifiers. Based on this condition, we derive an upper bound on the probability of error for both tasks, which depends on the statistical properties of the data and the combination policy used to combine the distributed classifiers. For the statistical classification problem, we employ the geometric social learning rule and conduct a non-asymptotic performance analysis. An exponential decay of the probability of error with respect to the number of unlabeled samples is observed in the upper bound. For the single-sample classification task, a distributed learning rule that functions as an ensemble classifier is constructed. An upper bound on the probability of error of this ensemble classifier is established. | ['Ali H. Sayed', 'Mert Kayaalp', 'Virginia Bordignon', 'Ping Hu'] | 2023-06-15 | null | null | null | null | ['classification-1'] | ['methodology'] | [ 3.60697627e-01 5.40166318e-01 -3.70080471e-01 -5.09404600e-01
-4.34203267e-01 -1.97222427e-01 3.60962868e-01 5.63911140e-01
-4.03163195e-01 8.61571789e-01 -5.55717647e-01 -2.74928451e-01
-4.73600298e-01 -7.56527603e-01 -7.98332810e-01 -1.20915103e+00
-1.06094368e-01 4.43478405e-01 6.78462610e-02 2.23841637e-01
7.76873529e-02 1.21457331e-01 -1.57124078e+00 -8.60803574e-02
1.11020303e+00 1.29425740e+00 -1.95818469e-01 6.48775458e-01
3.06534339e-02 7.23596871e-01 -6.87377036e-01 -3.60582292e-01
2.91395068e-01 -7.84180462e-01 -6.27692521e-01 3.01881343e-01
-1.12757474e-01 1.36360958e-01 3.13596606e-01 1.25936937e+00
1.98940650e-01 2.73128271e-01 1.34280813e+00 -1.72677791e+00
-2.82552809e-01 7.24811196e-01 -5.06182134e-01 -1.66432634e-02
-2.32174452e-02 -5.61207116e-01 9.59581375e-01 -4.41310436e-01
3.39129388e-01 7.30769157e-01 4.87820357e-01 5.66389680e-01
-9.01617646e-01 -6.65031016e-01 1.02285795e-01 1.44790843e-01
-1.29568756e+00 -6.73000887e-02 6.84124470e-01 -6.11398697e-01
1.11247271e-01 -5.21704974e-03 3.34814429e-01 7.42540658e-01
4.41385567e-01 6.50296688e-01 1.02340424e+00 -8.87541592e-01
9.16702271e-01 5.38318872e-01 6.81866586e-01 6.16264760e-01
7.18924344e-01 1.30707309e-01 -3.98202032e-01 -3.97509605e-01
1.00405663e-01 1.27327845e-01 -1.04338765e-01 -4.56983954e-01
-3.84664983e-01 9.97377336e-01 2.24614009e-01 2.28956595e-01
-3.55435163e-01 -2.02289775e-01 7.25747049e-02 5.04341364e-01
1.04119837e+00 -1.81413755e-01 -2.87306100e-01 4.45988536e-01
-4.88087416e-01 -1.51841059e-01 1.12079298e+00 6.20913923e-01
7.42387414e-01 -4.61798936e-01 6.62962496e-02 6.06301665e-01
5.08661389e-01 6.65906310e-01 1.32566750e-01 -5.72343051e-01
3.75843793e-01 6.17246747e-01 7.64227360e-02 -6.98971510e-01
-2.11909100e-01 -7.03507483e-01 -8.76721084e-01 3.24916363e-01
8.57506216e-01 -6.13610029e-01 -2.55116343e-01 1.88935149e+00
5.44193625e-01 4.46542315e-02 2.40493312e-01 5.20509481e-01
-5.46742342e-02 4.35305566e-01 -1.01036578e-02 -7.83635616e-01
8.16941559e-01 -7.05338657e-01 -6.43611431e-01 2.78880358e-01
1.03481770e+00 -9.39774290e-02 4.52251971e-01 3.37686092e-01
-6.07023895e-01 -2.02613786e-01 -1.19142425e+00 5.84096491e-01
-1.94429010e-01 -2.47846115e-02 -1.60037801e-02 8.74693274e-01
-7.43580282e-01 7.50068724e-01 -6.28786564e-01 -3.89309675e-01
3.33230019e-01 3.44784558e-01 -2.11611837e-01 -1.55511186e-01
-8.56674850e-01 5.64900815e-01 2.93004364e-01 -5.90803511e-02
-2.24011347e-01 -2.98355162e-01 -4.94313568e-01 2.87333447e-02
2.18840897e-01 -6.34596765e-01 8.14004183e-01 -1.43529940e+00
-1.18835676e+00 5.77244580e-01 -1.07961558e-01 -4.28140312e-01
6.74547136e-01 1.89523250e-01 -1.77149296e-01 1.11721486e-01
6.74101636e-02 -1.73339665e-01 9.94428039e-01 -1.24890316e+00
-8.73188078e-01 -8.22452486e-01 -2.65524685e-01 1.15402855e-01
-5.09958327e-01 -2.92402238e-01 4.37065750e-01 -2.78654307e-01
-1.55385956e-01 -1.06222963e+00 -1.79707602e-01 -1.55406207e-01
-1.07282780e-01 -6.41723752e-01 7.50259280e-01 -2.11125270e-01
1.07625151e+00 -2.39065266e+00 8.21304843e-02 5.57538092e-01
2.46445358e-01 -1.61900952e-01 1.12067088e-01 2.83859253e-01
2.58670539e-01 1.72689885e-01 -4.10922527e-01 -5.93975961e-01
-2.41961047e-01 1.38303980e-01 -1.24203742e-01 7.31097460e-01
3.90605442e-02 1.99814424e-01 -9.36849952e-01 -2.98862815e-01
-3.00895870e-01 -3.20293382e-02 -4.20295656e-01 3.58395725e-01
-1.40877590e-01 4.13840681e-01 -6.12251461e-01 9.46960598e-02
5.54830432e-01 -6.70751154e-01 5.13043761e-01 2.89451092e-01
4.33233231e-01 -3.20181519e-01 -1.14746940e+00 8.70086193e-01
-2.83463448e-01 2.26325765e-01 1.10321783e-01 -1.46873701e+00
9.06942427e-01 1.81556582e-01 4.59472775e-01 1.88846782e-01
4.43267405e-01 4.69088614e-01 -5.59565844e-03 -2.65007377e-01
-7.23614618e-02 -3.18523407e-01 -3.58147100e-02 9.40014660e-01
1.98955521e-01 4.38897491e-01 -1.18714357e-02 1.73085898e-01
1.10682392e+00 -3.15805107e-01 5.03109753e-01 -5.84196568e-01
4.02920127e-01 -2.98085213e-01 3.97069961e-01 8.43433738e-01
-2.26034194e-01 -4.19068746e-02 7.03979254e-01 -2.20196679e-01
-6.97299302e-01 -7.55323410e-01 -2.52628237e-01 1.05571079e+00
4.36258197e-01 -3.34617831e-02 -7.68302739e-01 -1.19126022e+00
1.80936098e-01 7.32436299e-01 -8.61084044e-01 -4.43381190e-01
1.54185519e-01 -9.32586014e-01 -3.19453180e-02 1.81113347e-01
5.36383986e-01 -5.91760814e-01 -2.71963716e-01 -2.16759592e-01
1.25471979e-01 -9.32660937e-01 -3.27669054e-01 4.10048783e-01
-7.73401618e-01 -1.33971691e+00 -4.20690715e-01 -6.27015293e-01
9.89104211e-01 1.40735079e-02 6.06335998e-01 8.22607577e-02
3.87291908e-01 7.16327250e-01 -5.06901801e-01 -7.32596576e-01
-7.09406555e-01 1.99365765e-01 2.15841725e-01 7.75100350e-01
2.79914409e-01 -3.08854789e-01 -4.32355165e-01 4.82095033e-01
-7.82041550e-01 -2.35557973e-01 4.08389688e-01 6.91033363e-01
2.42746472e-01 2.14383930e-01 1.11723948e+00 -8.62429202e-01
5.74610531e-01 -1.00150120e+00 -4.34621513e-01 6.26280665e-01
-7.82841682e-01 3.01241964e-01 6.60493374e-01 -5.72047472e-01
-8.96616399e-01 1.00936919e-01 6.26475334e-01 -2.15535685e-01
5.53364009e-02 5.37540793e-01 -4.52639312e-02 -9.18944478e-02
7.05905020e-01 -1.10204287e-01 4.60082322e-01 -2.17489988e-01
5.92066236e-02 1.23116112e+00 -7.65181035e-02 -5.80751538e-01
5.62247276e-01 3.22068334e-01 3.64052176e-01 -9.27046359e-01
-1.19757426e+00 -3.65250319e-01 -5.72316170e-01 -7.28979707e-01
6.82084620e-01 -5.13966858e-01 -7.36446023e-01 5.39741933e-01
-8.74422789e-01 -4.50586438e-01 -4.87049222e-01 7.19958365e-01
-5.55132627e-01 2.08880678e-01 -2.78925627e-01 -1.50997210e+00
-1.89197347e-01 -6.44000828e-01 6.76135480e-01 1.77489698e-01
1.22723386e-01 -1.28521371e+00 3.44687216e-02 2.45310843e-01
9.86495335e-03 2.85665005e-01 8.96678746e-01 -1.30014789e+00
-2.79639244e-01 -5.87999940e-01 -5.27404286e-02 6.33078814e-01
8.34292248e-02 -6.39532357e-02 -7.28947103e-01 -2.93287873e-01
2.17852280e-01 -3.57882589e-01 7.86858439e-01 3.87440592e-01
9.35109138e-01 -2.86708415e-01 -4.76107597e-01 2.71584583e-03
1.19533348e+00 2.11517677e-01 6.95858523e-02 -1.69523209e-01
2.85566896e-01 9.36922669e-01 6.48029029e-01 5.92965424e-01
2.85180271e-01 1.56726822e-01 2.19157085e-01 4.36745614e-01
4.29581791e-01 -2.42084205e-01 3.69202673e-01 8.07811975e-01
-6.06129318e-02 -6.29595160e-01 -6.68278337e-01 -1.84076279e-02
-2.12553573e+00 -7.81166255e-01 3.04679177e-03 2.68081236e+00
5.77872396e-01 -7.84105659e-02 3.21250826e-01 4.62438375e-01
1.23962820e+00 -2.20550120e-01 -6.40618145e-01 -1.64060965e-01
4.18842249e-02 -1.59883767e-01 5.27201593e-01 5.88935971e-01
-8.84936452e-01 1.26328781e-01 6.11950302e+00 8.75539482e-01
-9.91330743e-01 1.13946669e-01 9.12541568e-01 2.20834747e-01
-2.10944027e-01 -3.20783332e-02 -8.76922488e-01 7.20996320e-01
1.13426745e+00 -4.52935964e-01 2.38424420e-01 8.21998656e-01
-1.43531010e-01 -4.10038024e-01 -1.15001667e+00 5.60713053e-01
6.37569502e-02 -8.34938467e-01 -1.39030397e-01 4.28580999e-01
7.23456323e-01 -2.10460067e-01 -4.98015285e-02 1.40350044e-01
4.66158777e-01 -4.99179602e-01 6.03586376e-01 5.82128108e-01
5.84740341e-01 -6.54125750e-01 8.68879855e-01 1.02270532e+00
-9.23989594e-01 -5.76614499e-01 -2.79604644e-02 -2.89491922e-01
-2.16715321e-01 1.10398126e+00 -3.64749432e-01 6.45334601e-01
1.69611618e-01 8.04054260e-01 -2.53890276e-01 8.50799322e-01
-1.35320546e-02 7.98589647e-01 -5.32356024e-01 -5.26290119e-01
-4.04182941e-01 -3.96204233e-01 3.57891440e-01 5.23758531e-01
4.35255349e-01 1.69561446e-01 3.17316145e-01 3.91399086e-01
-2.50791878e-01 3.82240295e-01 -6.04848087e-01 1.76682085e-01
5.80990732e-01 9.94099021e-01 -9.13801968e-01 -3.22666019e-01
-2.32890740e-01 5.58662891e-01 6.76677465e-01 3.75606835e-01
-5.85832119e-01 -2.31221527e-01 -2.78867800e-02 7.80787691e-02
1.71420410e-01 2.41621006e-02 -4.40934837e-01 -1.16514599e+00
2.29533777e-01 -1.88627049e-01 5.96916318e-01 -3.94558646e-02
-1.69836581e+00 3.23514640e-01 1.74017966e-01 -1.23770821e+00
-2.47691184e-01 -4.21097159e-01 -8.28109860e-01 4.16784048e-01
-1.12562978e+00 -6.29270196e-01 -8.08281079e-02 4.48955208e-01
-8.09505060e-02 -3.09720367e-01 5.12917757e-01 8.46797973e-02
-8.63407850e-01 6.29034638e-01 5.24097741e-01 3.97491176e-03
5.65766811e-01 -1.13962853e+00 -6.72307670e-01 5.60529113e-01
-1.43016621e-01 3.82028110e-02 4.58782971e-01 -6.46494091e-01
-8.90173972e-01 -1.15660989e+00 9.22894120e-01 -3.13531220e-01
6.37337923e-01 -2.93956846e-01 -7.02429533e-01 5.13070643e-01
-2.64432132e-01 4.24748302e-01 9.62141991e-01 2.13119298e-01
-2.09540591e-01 -2.64341384e-01 -1.42069232e+00 5.71609251e-02
9.18527424e-01 -2.29294911e-01 -2.43716568e-01 4.66606766e-01
4.09606993e-01 3.53488892e-01 -8.48030388e-01 3.59899610e-01
4.50320274e-01 -7.88801730e-01 1.20395817e-01 -6.32789552e-01
4.80324477e-01 -2.68380102e-02 -2.90760517e-01 -1.29728723e+00
-3.89911910e-03 -4.67183262e-01 -2.27759749e-01 1.06037152e+00
5.36291242e-01 -1.17986333e+00 7.03905642e-01 5.85578799e-01
3.99208963e-01 -9.03533757e-01 -1.08115840e+00 -9.97754157e-01
2.84085244e-01 -1.65557370e-01 -1.13413101e-02 7.54791081e-01
3.26533169e-01 5.50510585e-01 -7.00816363e-02 1.48190022e-01
1.13737655e+00 7.44802654e-02 4.96870190e-01 -1.54967666e+00
-3.35798949e-01 -1.91813067e-01 -4.48444068e-01 -6.32642448e-01
5.49248934e-01 -1.02797019e+00 8.89915973e-02 -8.52663279e-01
3.17938358e-01 -7.38563836e-01 -3.60679626e-01 1.32322237e-01
-2.64125496e-01 -2.74326950e-01 1.21870369e-01 4.39240038e-01
-8.18352222e-01 5.00141263e-01 7.59190381e-01 1.69240624e-01
-2.39358142e-01 5.27043164e-01 -4.17999595e-01 6.47489011e-01
7.89269567e-01 -6.38763070e-01 -5.06546915e-01 2.57007450e-01
9.51438397e-02 3.27737361e-01 1.98585138e-01 -8.39740992e-01
2.76700050e-01 -6.84031174e-02 8.78536701e-02 -6.22185692e-02
-1.58490524e-01 -8.09695303e-01 -7.30003566e-02 6.70684814e-01
-7.73151696e-01 -6.19500399e-01 -6.04935527e-01 1.18609118e+00
5.37232384e-02 -5.81609428e-01 9.74464595e-01 4.49781239e-01
2.35733226e-01 2.21319661e-01 -3.47746491e-01 1.06689617e-01
1.50249517e+00 1.89279199e-01 -1.77564114e-01 -4.68914330e-01
-9.71935332e-01 2.84231633e-01 1.41131505e-01 -4.89197634e-02
1.63594976e-01 -1.13497961e+00 -7.59762049e-01 2.14388162e-01
8.58896375e-02 -1.33999184e-01 7.35488161e-02 9.59196210e-01
1.19979873e-01 -2.42581069e-01 1.09627068e-01 -6.70274496e-01
-1.03327858e+00 5.84646165e-01 4.66172427e-01 -3.25449139e-01
8.65910668e-03 4.61142570e-01 -2.29788143e-02 -7.42503479e-02
2.48850301e-01 -4.48596627e-02 -2.11657315e-01 2.22970858e-01
3.46038252e-01 5.62178195e-01 -5.32870069e-02 -1.99842677e-01
-9.02165025e-02 3.70723188e-01 1.24190450e-01 -3.46094668e-02
9.52369571e-01 -2.39189252e-01 -7.80882239e-02 9.10514653e-01
1.24513865e+00 -1.25000685e-01 -1.00776196e+00 -6.32555485e-01
4.94726673e-02 -1.05249643e-01 -1.99356437e-01 -4.99365628e-01
-9.30425346e-01 5.45458972e-01 4.60758984e-01 8.71030927e-01
9.83486593e-01 1.91340491e-01 1.09999523e-01 2.80958563e-01
4.65083867e-01 -9.95817721e-01 -9.29994360e-02 4.16455477e-01
6.41827643e-01 -1.37527442e+00 -1.64633036e-01 -6.12226605e-01
-3.89545679e-01 1.10021293e+00 3.79828155e-01 -2.83527225e-01
1.36991525e+00 2.52303392e-01 -2.64680475e-01 2.39454269e-01
-9.32837069e-01 -3.67534421e-02 3.70148838e-01 4.23003197e-01
1.59725070e-01 2.37111673e-01 -6.16252720e-01 7.48596907e-01
1.36701062e-01 5.31508513e-02 3.11361969e-01 7.51995862e-01
-5.76636195e-01 -8.65580022e-01 -1.65493637e-01 6.54888690e-01
-1.78855777e-01 4.53911513e-01 -4.11082447e-01 2.90740430e-01
-1.72406584e-02 1.23579645e+00 2.62117624e-01 -5.62568545e-01
-1.72585919e-01 3.11505318e-01 3.96112412e-01 -5.50932825e-01
-1.46291777e-01 -3.82734239e-01 -5.17872907e-02 7.13046193e-02
-7.26172149e-01 -5.70561051e-01 -9.53578353e-01 -3.66954088e-01
-7.42557347e-01 6.56596422e-01 6.28495395e-01 1.34352195e+00
3.06123704e-01 1.16476774e-01 1.25103164e+00 -5.56230664e-01
-1.16235423e+00 -8.86918962e-01 -1.12803745e+00 3.51907849e-01
1.58898160e-01 -5.88155270e-01 -1.18174398e+00 -8.26980993e-02] | [9.085372924804688, 4.129692077636719] |
8d462eed-9669-4f09-a150-9102d5349bba | addsl-hand-gesture-detection-and-sign | 2305.09736 | null | https://arxiv.org/abs/2305.09736v1 | https://arxiv.org/pdf/2305.09736v1.pdf | ADDSL: Hand Gesture Detection and Sign Language Recognition on Annotated Danish Sign Language | For a long time, detecting hand gestures and recognizing them as letters or numbers has been a challenging task. This creates communication barriers for individuals with disabilities. This paper introduces a new dataset, the Annotated Dataset for Danish Sign Language (ADDSL). Annota-tions for the dataset were made using the open-source tool LabelImg in the YOLO format. Using this dataset, a one-stage ob-ject detector model (YOLOv5) was trained with the CSP-DarkNet53 backbone and YOLOv3 head to recognize letters (A-Z) and numbers (0-9) using only seven unique images per class (without augmen-tation). Five models were trained with 350 epochs, resulting in an average inference time of 9.02ms per image and a best accu-racy of 92% when compared to previous research. Our results show that modified model is efficient and more accurate than existing work in the same field. The code repository for our model is available at the GitHub repository https://github.com/s4nyam/pvt-addsl. | ['Sanyam Jain'] | 2023-05-16 | null | null | null | null | ['sign-language-recognition'] | ['computer-vision'] | [-1.43044502e-01 -1.73284918e-01 -1.50480211e-01 -2.23143116e-01
-5.92471659e-01 -5.03253400e-01 3.15864742e-01 -5.27750194e-01
-7.53657103e-01 8.29616785e-01 4.14064020e-01 -1.70306712e-02
2.51990378e-01 -2.18217254e-01 -5.38461030e-01 -5.56110919e-01
1.73165023e-01 4.86531287e-01 6.06383264e-01 1.57787457e-01
3.11529160e-01 6.69862151e-01 -1.64500153e+00 4.69073266e-01
6.97591603e-01 6.28060818e-01 2.18986824e-01 1.08032155e+00
1.51761547e-01 9.00865376e-01 -5.93123198e-01 -4.65167284e-01
3.64500314e-01 -3.43903422e-01 -8.96210551e-01 -4.57615018e-01
9.57544386e-01 -8.55013549e-01 -7.97761738e-01 8.99757147e-01
1.21955562e+00 -2.49059405e-02 5.84477842e-01 -1.25482571e+00
-5.54457426e-01 4.32657301e-01 -2.57162333e-01 3.60466018e-02
3.84123504e-01 4.15323168e-01 6.59868598e-01 -1.01257467e+00
8.98103833e-01 1.34574437e+00 5.70039988e-01 1.16411901e+00
-6.47462368e-01 -9.84064460e-01 -6.79625059e-03 6.11927450e-01
-1.44926369e+00 -6.27055883e-01 2.21442670e-01 -5.85175633e-01
1.16760671e+00 4.07456279e-01 8.69249523e-01 1.44738567e+00
-3.33902240e-01 1.22188711e+00 1.32471561e+00 -7.28918433e-01
-5.95825352e-02 -4.01074171e-01 2.51770288e-01 7.19666600e-01
2.94449985e-01 -2.29599793e-02 -8.01793158e-01 3.32819849e-01
8.57039869e-01 -2.40863472e-01 -1.74939498e-01 1.96486011e-01
-1.03247845e+00 1.87402263e-01 4.04234290e-01 2.41412982e-01
-3.61984342e-01 2.83350557e-01 3.50283921e-01 1.99520230e-01
-1.14212401e-01 -7.10470378e-02 -4.10483152e-01 -5.98985255e-01
-7.73220241e-01 3.44927520e-01 7.49560416e-01 1.37448967e+00
-9.56268311e-02 -3.73518467e-03 -3.97595167e-01 8.43490362e-01
2.99896032e-01 6.59229219e-01 3.28552455e-01 -9.24492717e-01
6.27943814e-01 3.89631599e-01 -2.30240449e-02 -3.56510043e-01
-4.39190894e-01 1.74149558e-01 -5.03925264e-01 5.46891510e-01
1.13661659e+00 -3.88893515e-01 -1.54170012e+00 1.34758866e+00
-1.08499736e-01 1.97428297e-02 -2.96710968e-01 1.07079709e+00
1.28869843e+00 1.22118078e-01 1.21146180e-01 4.17827517e-01
1.30323517e+00 -1.09955204e+00 -7.19855130e-01 -1.10552557e-01
5.03943384e-01 -7.69822478e-01 1.03317750e+00 5.91668963e-01
-8.92181516e-01 -2.37163514e-01 -5.94963789e-01 -3.97380918e-01
-4.62549746e-01 5.90718329e-01 5.72331786e-01 7.00487614e-01
-1.15073681e+00 2.11726457e-01 -9.30474997e-01 -8.51656139e-01
8.20157826e-01 4.59979028e-01 -3.78985077e-01 -5.88707440e-02
-6.59377396e-01 1.12587869e+00 2.63783067e-01 3.16739053e-01
-4.94881034e-01 -1.50367901e-01 -4.41251427e-01 -5.21535516e-01
1.27617002e-01 -2.41411492e-01 1.35124815e+00 -7.20470428e-01
-1.62538874e+00 1.12534118e+00 -3.99327636e-01 -9.74579975e-02
1.05349970e+00 -4.64075297e-01 -4.08270270e-01 1.98455915e-01
-5.50354831e-02 9.56066132e-01 5.62255800e-01 -9.37114954e-01
-7.86225855e-01 -4.66913164e-01 -1.58017546e-01 1.96446225e-01
1.04002088e-01 6.52622700e-01 -7.19279051e-01 -6.98105633e-01
1.05982147e-01 -1.01898241e+00 3.49338263e-01 3.93315285e-01
-5.74215353e-01 -3.23297888e-01 5.44002593e-01 -1.39350545e+00
1.09131610e+00 -1.74971128e+00 -7.22903609e-02 5.48110753e-02
1.70722201e-01 7.91511595e-01 -8.58292803e-02 3.80560249e-01
1.12262033e-01 -1.17442660e-01 -2.27194592e-01 -2.14873746e-01
-8.20385963e-02 2.94833541e-01 2.14421779e-01 4.58572179e-01
-8.93237591e-02 9.97321546e-01 -8.72522414e-01 -4.71323550e-01
3.65001231e-01 6.18849516e-01 -3.33046883e-01 5.26572093e-02
1.62337609e-02 6.93017364e-01 7.01363310e-02 1.13888705e+00
5.53804874e-01 1.03652947e-01 6.30309060e-02 1.13999777e-01
-3.54596764e-01 2.07123104e-02 -1.28734875e+00 1.57418716e+00
-1.25633404e-01 1.22898090e+00 -3.73382270e-02 -3.40090632e-01
4.84199941e-01 3.83360028e-01 7.23766536e-02 -4.81876582e-01
5.49010694e-01 5.65176189e-01 3.26455623e-01 -9.32143807e-01
6.86409175e-02 6.32930875e-01 4.47533041e-01 3.46050151e-02
7.66561627e-02 2.51299918e-01 4.48735237e-01 -1.36240736e-01
1.04053926e+00 3.32491428e-01 1.87017769e-01 2.22555101e-01
3.31517667e-01 -1.36456937e-01 3.64957213e-01 9.94724691e-01
-5.56159735e-01 9.07303870e-01 3.68763030e-01 -2.29033172e-01
-1.00554955e+00 -8.82501543e-01 -3.21243033e-02 8.94287646e-01
-3.84170592e-01 -1.23808645e-01 -7.29351163e-01 -4.58680123e-01
1.24481231e-01 4.02028352e-01 -3.94208044e-01 6.04837298e-01
-8.33107233e-01 -1.06929287e-01 1.19654453e+00 1.00389111e+00
7.52274990e-01 -1.67452264e+00 -8.74454618e-01 -1.38595015e-01
-1.16616189e-02 -1.15002370e+00 -3.29291523e-01 -8.38263426e-03
-5.68114817e-01 -1.21750247e+00 -1.49823809e+00 -1.22813129e+00
7.51423657e-01 -3.78784150e-01 5.13109148e-01 8.44528675e-02
-5.65435886e-01 1.74573869e-01 -6.61557496e-01 -6.08860552e-01
-7.50577971e-02 -6.38873950e-02 -1.94421429e-02 -4.28619385e-01
8.50293219e-01 -2.42329642e-01 -6.19953215e-01 -1.03187196e-01
-3.59660178e-01 1.00610964e-01 6.41286433e-01 8.00478578e-01
8.72115642e-02 -9.61284876e-01 1.31749690e-01 -3.47758889e-01
4.67120141e-01 1.82356015e-01 -6.25455797e-01 1.19017363e-01
-3.22524250e-01 -1.67058244e-01 1.20478414e-01 -5.08227944e-01
-8.64089072e-01 4.23615634e-01 -4.36995447e-01 -1.84868425e-01
-6.78220153e-01 -2.00862586e-01 1.55577194e-02 -2.88385063e-01
5.32173097e-01 2.23730832e-01 1.33068906e-03 -7.70003915e-01
5.18982470e-01 1.33077395e+00 8.24858189e-01 -2.54342735e-01
2.59159029e-01 4.95582253e-01 -3.22924614e-01 -1.03008187e+00
-3.14915478e-01 -3.81965071e-01 -1.14642644e+00 -5.45293868e-01
8.33868682e-01 -7.42897272e-01 -9.84718740e-01 1.55198491e+00
-1.41394067e+00 -6.22038066e-01 1.35323107e-02 8.15301716e-01
-2.93725967e-01 3.65799934e-01 -5.82907379e-01 -9.28521037e-01
-4.71979767e-01 -9.56675887e-01 7.53885984e-01 3.34679306e-01
-3.73149157e-01 -3.63764405e-01 -2.46509671e-01 2.90192366e-01
3.04343164e-01 2.70806760e-01 3.51912677e-01 -4.62386638e-01
-3.79672319e-01 -3.55780691e-01 -7.01865673e-01 5.10781288e-01
-3.03233918e-02 2.18792371e-02 -1.17565024e+00 -7.67518952e-02
-9.11790252e-01 -2.95193076e-01 9.32452381e-01 5.31889975e-01
9.84552383e-01 -2.71260828e-01 -1.14625193e-01 4.63638842e-01
1.04942715e+00 4.88404185e-01 7.34031856e-01 4.69202638e-01
1.03936684e+00 3.29309523e-01 1.53883561e-01 4.68645662e-01
6.01859331e-01 7.39234209e-01 5.65511100e-02 4.51896973e-02
-8.05917978e-01 -2.10847542e-01 1.68206200e-01 6.48591816e-01
-8.16654623e-01 -2.08800212e-01 -1.32283413e+00 6.92077100e-01
-1.82932413e+00 -1.05655408e+00 -5.17867863e-01 1.94478059e+00
9.69945312e-01 -2.14221671e-01 4.47057337e-01 4.16946411e-01
8.42116296e-01 -5.15221320e-02 -5.02119541e-01 -3.35385859e-01
-1.87115893e-01 5.77177584e-01 8.24349463e-01 6.30945265e-01
-1.07086062e+00 1.46388173e+00 6.19937706e+00 4.64644432e-01
-1.24953854e+00 2.02007920e-01 -1.17493339e-01 -2.74427980e-01
7.14005768e-01 -2.82928675e-01 -9.87428606e-01 7.35770106e-01
5.49972594e-01 4.07064766e-01 6.81688309e-01 7.59105742e-01
2.41914943e-01 -2.91181326e-01 -9.13591385e-01 1.21017802e+00
2.61970222e-01 -9.03786242e-01 -1.52488351e-01 -2.66788363e-01
4.09918219e-01 4.99521136e-01 -3.53499144e-01 -1.77959818e-02
2.28557467e-01 -1.03876805e+00 9.08761382e-01 7.47858942e-01
1.40615714e+00 -2.60224283e-01 6.99012518e-01 1.58738196e-01
-1.00027490e+00 -1.30100578e-01 -3.05442680e-02 -3.39252204e-01
8.84666741e-02 -2.33974606e-01 -8.29559088e-01 -2.83260457e-02
9.98934627e-01 7.73977995e-01 -6.42984927e-01 1.58390498e+00
-8.77591074e-01 6.24368012e-01 -4.86283988e-01 -4.25288856e-01
-4.37716953e-02 1.86940849e-01 6.32818937e-01 1.56126380e+00
2.63132125e-01 1.55789107e-01 -4.20527548e-01 1.36482388e-01
9.35522765e-02 -1.06932446e-01 -3.47771138e-01 1.44556627e-01
6.96641505e-01 6.51699483e-01 -6.62326396e-01 -3.50513041e-01
-4.84208107e-01 1.36781967e+00 1.16387039e-01 4.64674205e-01
-4.85382736e-01 -7.77367234e-01 3.27931970e-01 -1.88444227e-01
2.03963771e-01 -4.30006862e-01 -6.17503464e-01 -1.02622259e+00
4.35788512e-01 -9.70454931e-01 3.34910035e-01 -8.92312288e-01
-9.62125063e-01 4.88548666e-01 -1.68494910e-01 -1.30282414e+00
-8.83789062e-02 -1.17300344e+00 -2.23998994e-01 9.48769033e-01
-1.25594640e+00 -1.43971038e+00 -7.68266559e-01 6.93249881e-01
5.56110144e-01 -3.72908294e-01 9.24789488e-01 5.22208273e-01
-6.14367664e-01 9.03573096e-01 9.29180626e-03 8.35758448e-01
7.86625504e-01 -1.17229855e+00 4.15708482e-01 1.01096070e+00
-7.89498463e-02 4.96493071e-01 3.26680928e-01 -8.21793377e-01
-9.87879634e-01 -7.04084516e-01 1.21156323e+00 -6.40420079e-01
4.69830394e-01 -3.76472473e-01 -4.58385378e-01 8.40279162e-01
1.69952527e-01 6.67290911e-02 2.94289947e-01 -4.21478063e-01
-3.28501940e-01 2.46997848e-01 -1.17583752e+00 6.41272306e-01
1.67725194e+00 -4.56416100e-01 -5.95112622e-01 3.09206337e-01
4.22491953e-02 -1.04319990e+00 -5.22281528e-01 4.73135598e-02
1.33398855e+00 -6.24087274e-01 6.14943027e-01 -6.00817204e-01
2.15964645e-01 -2.36109570e-01 -8.77082422e-02 -8.23649764e-01
-3.37496325e-02 -4.34945613e-01 -3.45653653e-01 8.30264926e-01
2.56095320e-01 -7.41890013e-01 7.31561244e-01 6.69842958e-01
-7.69872889e-02 -4.57829297e-01 -1.24228251e+00 -1.12480044e+00
-9.80846956e-02 -6.87980890e-01 2.75526881e-01 6.04472578e-01
-1.54010013e-01 -2.47614518e-01 -4.98759776e-01 1.37618810e-01
5.03671408e-01 -4.98125672e-01 7.18832850e-01 -1.08401144e+00
1.65501490e-01 -7.07978189e-01 -7.74158657e-01 -8.99162889e-01
-4.71887328e-02 -9.38593626e-01 -1.63709279e-02 -1.91487288e+00
-3.00309379e-02 -1.15315832e-01 8.30989853e-02 1.13359487e+00
2.08002850e-01 4.34358150e-01 5.06883621e-01 3.51520091e-01
-2.57769614e-01 5.21832071e-02 1.18117511e+00 1.90923922e-02
-2.20112175e-01 7.60041624e-02 -7.36323819e-02 8.67725790e-01
8.59466672e-01 -3.33687395e-01 3.14751655e-01 -7.75604486e-01
-2.16125831e-01 -4.26236391e-01 5.86805999e-01 -1.17028594e+00
4.07013714e-01 1.38390318e-01 4.54752803e-01 -9.33930337e-01
4.76803809e-01 -5.32597005e-01 -2.67717808e-01 8.23802412e-01
-1.30690008e-01 -2.40969688e-01 8.38692784e-02 -7.07981884e-02
2.32560150e-02 -1.96422115e-01 7.91332245e-01 -1.53431073e-01
-1.22689867e+00 9.65741947e-02 -4.50256109e-01 -1.13271950e-02
9.08243299e-01 -4.66280758e-01 -4.88701373e-01 -2.82280028e-01
-8.44392955e-01 2.19658941e-01 1.72596171e-01 6.79221630e-01
6.65603518e-01 -1.12423027e+00 -7.45152175e-01 1.44454092e-01
1.27044827e-01 -7.65698031e-02 5.18763624e-02 8.16363811e-01
-1.14567804e+00 6.22452378e-01 -6.69484198e-01 -3.08646291e-01
-1.70189953e+00 -4.01933342e-01 2.07440957e-01 4.61094767e-01
-9.71166670e-01 1.27752709e+00 -8.42842638e-01 -4.77072030e-01
9.11818147e-01 -4.97086823e-01 -3.51871938e-01 1.25196785e-01
5.24254620e-01 9.72167909e-01 -7.00706318e-02 -8.17620814e-01
-7.43212461e-01 6.82002366e-01 1.96243450e-02 -3.55526507e-01
1.40935254e+00 3.81945789e-01 3.24151851e-03 2.12905318e-01
7.63539374e-01 -1.14446834e-01 -1.16838467e+00 -4.36616428e-02
1.19686015e-01 -6.21682286e-01 -2.09788918e-01 -1.28671038e+00
-9.05204296e-01 9.41606343e-01 1.08812952e+00 -5.22262454e-01
8.56489778e-01 1.78463422e-02 6.98406994e-01 4.91845071e-01
3.98490191e-01 -1.36065567e+00 -2.79992998e-01 9.11662936e-01
1.29694521e+00 -1.27464247e+00 -1.97300181e-01 -2.46036232e-01
-7.72941768e-01 1.11085463e+00 7.82991469e-01 -2.03232005e-01
3.42279822e-01 3.59017283e-01 4.44091409e-01 5.97175397e-02
1.12420455e-01 -6.87156558e-01 3.53246927e-01 9.09880698e-01
5.35814881e-01 3.61128241e-01 -7.41967976e-01 5.06863713e-01
-2.66370863e-01 5.42490184e-01 3.09802622e-01 1.26959729e+00
-3.22953202e-02 -1.04996860e+00 -3.62276286e-01 6.47687197e-01
-3.66541803e-01 -1.97256386e-01 -7.07973599e-01 8.73258233e-01
2.69256353e-01 7.68532455e-01 -8.66482779e-02 -3.64339501e-01
4.03872758e-01 2.99876928e-01 8.36006582e-01 -5.88242471e-01
-3.61149013e-01 -1.29459515e-01 3.84252280e-01 -5.33703268e-01
-4.25522804e-01 -7.76733160e-01 -1.39216471e+00 -2.45494545e-01
-1.68784782e-01 -8.73788238e-01 6.64674222e-01 9.30558801e-01
2.62381017e-01 3.39314550e-01 -3.27431440e-01 -8.94344211e-01
-1.99243188e-01 -1.30796969e+00 -5.06162763e-01 -1.91511004e-03
1.81944028e-01 -6.03870273e-01 -7.23524168e-02 4.32226025e-02] | [9.107240676879883, -6.414835453033447] |
b8f2f9e3-04ed-4418-84bb-c2c2b6ebd3a9 | regularizing-contrastive-predictive-coding | 2304.05974 | null | https://arxiv.org/abs/2304.05974v2 | https://arxiv.org/pdf/2304.05974v2.pdf | Regularizing Contrastive Predictive Coding for Speech Applications | Self-supervised methods such as Contrastive predictive Coding (CPC) have greatly improved the quality of the unsupervised representations. These representations significantly reduce the amount of labeled data needed for downstream task performance, such as automatic speech recognition. CPC learns representations by learning to predict future frames given current frames. Based on the observation that the acoustic information, e.g., phones, changes slower than the feature extraction rate in CPC, we propose regularization techniques that impose slowness constraints on the features. Here we propose two regularization techniques: Self-expressing constraint and Left-or-Right regularization. We evaluate the proposed model on ABX and linear phone classification tasks, acoustic unit discovery, and automatic speech recognition. The regularized CPC trained on 100 hours of unlabeled data matches the performance of the baseline CPC trained on 360 hours of unlabeled data. We also show that our regularization techniques are complementary to data augmentation and can further boost the system's performance. In monolingual, cross-lingual, or multilingual settings, with/without data augmentation, regardless of the amount of data used for training, our regularized models outperformed the baseline CPC models on the ABX task. | ['Najim Dehak', 'Laureano Moro-Velazquez', 'Piotr Żelasko', 'Jesús Villalba', 'Saurabhchand Bhati'] | 2023-04-12 | null | null | null | null | ['acoustic-unit-discovery'] | ['speech'] | [ 4.38232213e-01 3.25195044e-01 -3.61945122e-01 -7.05470622e-01
-1.03282320e+00 -4.94654864e-01 7.65589297e-01 -3.42774380e-04
-5.27425826e-01 5.72843015e-01 5.55752873e-01 -5.71210027e-01
4.45441067e-01 -2.69898683e-01 -9.17355359e-01 -6.77616894e-01
-1.01374783e-01 2.98298329e-01 -6.55075014e-02 2.78146155e-02
1.99591294e-02 1.82530090e-01 -1.75777602e+00 5.32980680e-01
6.73466265e-01 9.66919303e-01 3.81373495e-01 7.23621905e-01
-1.73275903e-01 8.48065197e-01 -1.87603444e-01 8.73972401e-02
1.04854211e-01 -3.36353123e-01 -7.55292714e-01 1.75454751e-01
1.83590442e-01 6.73281699e-02 -4.14225847e-01 7.58013427e-01
2.26984411e-01 4.54764932e-01 5.94534039e-01 -8.91389608e-01
-6.93558872e-01 7.98453510e-01 -2.56858647e-01 4.84137148e-01
1.10698335e-01 -9.25675109e-02 9.63017166e-01 -1.34249151e+00
5.12636781e-01 1.36034036e+00 5.27360797e-01 8.80652308e-01
-1.20862937e+00 -5.71330488e-01 6.18650198e-01 3.04104924e-01
-1.26976669e+00 -1.33110499e+00 5.77872336e-01 -4.19738263e-01
1.45156097e+00 2.18041450e-01 1.46317542e-01 1.30914664e+00
-3.90432268e-01 7.90795207e-01 1.00949144e+00 -9.90766644e-01
2.54093319e-01 1.20698571e-01 2.93992877e-01 6.10849917e-01
-4.88990664e-01 4.12793338e-01 -8.64519894e-01 6.03012852e-02
3.46100926e-01 -2.30301678e-01 -3.64281923e-01 1.55540332e-01
-1.33605444e+00 7.89588690e-01 -3.27399299e-02 3.88371795e-01
-1.66761428e-01 1.24289609e-01 5.47317863e-01 2.76720375e-01
8.73700440e-01 2.05678850e-01 -8.47936094e-01 -5.53924739e-01
-7.66331851e-01 -4.02688503e-01 5.62685966e-01 1.11604691e+00
7.31157064e-01 5.36894917e-01 -5.26796505e-02 1.19944453e+00
3.54934067e-01 6.41708374e-01 9.73473370e-01 -9.79584992e-01
8.02564621e-01 1.10057764e-01 -3.76472145e-01 -4.71714407e-01
-1.95349947e-01 -1.13448016e-01 -6.49432898e-01 -3.15696985e-01
1.02686815e-01 -2.41773471e-01 -1.17413366e+00 1.99325073e+00
-1.63715348e-01 3.79274666e-01 3.78611147e-01 5.47075808e-01
4.35380131e-01 1.31090009e+00 8.08434933e-02 -7.69962966e-01
9.41004634e-01 -1.17361295e+00 -1.01233256e+00 -4.04424608e-01
8.75043213e-01 -7.85376847e-01 1.11638987e+00 2.65477985e-01
-8.44402432e-01 -7.11132824e-01 -8.45634580e-01 9.40513622e-04
-3.54080945e-01 2.87005931e-01 4.60120797e-01 6.63393378e-01
-1.09650087e+00 4.43316013e-01 -1.16984963e+00 -2.37499878e-01
1.13375396e-01 4.22448725e-01 -4.05625522e-01 -3.34760658e-02
-1.02237082e+00 6.99831128e-01 4.12317067e-01 1.57718346e-01
-8.31908405e-01 -5.93215048e-01 -9.61409211e-01 1.21286236e-01
1.37616158e-01 1.47325560e-01 1.34360814e+00 -1.10932052e+00
-1.86745143e+00 7.26586759e-01 -4.56415862e-01 -8.32181394e-01
-8.48324820e-02 -4.98387784e-01 -6.55509293e-01 -1.34783015e-01
-6.85447529e-02 9.77684915e-01 8.82597983e-01 -9.37807858e-01
-7.13924170e-01 -1.96671441e-01 -4.29027706e-01 3.63331020e-01
-3.71836662e-01 -1.19862728e-01 -7.30211854e-01 -7.13742852e-01
1.02569625e-01 -1.23719430e+00 -1.85082987e-01 -4.92822379e-01
-2.44323999e-01 -3.63038331e-01 9.86887157e-01 -7.39755988e-01
1.12623703e+00 -2.64683962e+00 1.36500493e-01 6.14369549e-02
-3.84073257e-01 3.73057097e-01 -2.05370814e-01 9.79514122e-02
-3.12527210e-01 1.08972006e-01 -2.40859941e-01 -7.66253769e-01
-2.49552250e-01 7.21632779e-01 -5.39537489e-01 3.90043586e-01
3.27697814e-01 4.99773800e-01 -7.36956656e-01 -1.72069043e-01
1.76496342e-01 5.38454831e-01 -7.13051856e-01 2.64487296e-01
-2.50545293e-01 7.58906245e-01 7.22932369e-02 3.38592738e-01
2.11705402e-01 -8.59555528e-02 3.00812781e-01 4.39923182e-02
-2.31959924e-01 8.74054015e-01 -8.39555383e-01 1.79905069e+00
-7.71723032e-01 9.49018419e-01 -1.31772563e-01 -1.30880034e+00
9.27346051e-01 7.00290084e-01 1.38982207e-01 -7.50012219e-01
-2.69678384e-01 3.19398195e-01 1.98831018e-02 -1.51853219e-01
1.82258338e-01 -1.79592259e-02 1.05812959e-01 3.07434440e-01
4.01622325e-01 2.02777073e-01 1.75623029e-01 9.61084664e-02
8.46326649e-01 1.34930804e-01 1.49934813e-01 -3.24826926e-01
6.93510652e-01 -2.83150226e-01 9.50003207e-01 5.30971527e-01
-8.16187412e-02 6.08129561e-01 1.74302861e-01 -2.41595656e-01
-9.38270211e-01 -7.28026450e-01 -1.88909531e-01 1.57146096e+00
-3.51988465e-01 -5.87749124e-01 -6.01281285e-01 -5.88041544e-01
-3.01814377e-01 7.14675903e-01 -3.46589297e-01 -1.73251495e-01
-8.84702802e-01 -6.84727252e-01 3.57572615e-01 7.04779267e-01
1.40948206e-01 -1.16260111e+00 1.23110101e-01 3.47056121e-01
-3.78396630e-01 -1.53642404e+00 -6.14980638e-01 6.62918806e-01
-8.90000165e-01 -5.87806165e-01 -2.53232211e-01 -9.78852451e-01
6.35493100e-01 1.75642088e-01 8.81977022e-01 -1.30249336e-01
2.13891938e-01 5.17154872e-01 -6.45986736e-01 -1.80941552e-01
-6.39858007e-01 1.24448337e-01 6.09783828e-01 2.37009764e-01
2.29063034e-01 -5.80698848e-01 -3.51834565e-01 1.15964986e-01
-3.96281898e-01 1.20796233e-01 4.43722785e-01 9.12525952e-01
9.10241544e-01 -3.88063639e-01 7.45178282e-01 -1.07258022e+00
2.08076894e-01 -3.15897524e-01 -5.07428885e-01 8.51301029e-02
-8.59674454e-01 3.52347344e-01 6.25639021e-01 -6.51356936e-01
-1.35830772e+00 3.36848438e-01 -2.88852125e-01 -4.92585838e-01
-8.72480720e-02 5.19416451e-01 2.84161568e-02 3.90465707e-01
5.30531108e-01 2.75331616e-01 -6.54739747e-03 -7.66578138e-01
6.31540656e-01 7.79004812e-01 6.81564510e-01 -4.68836218e-01
4.41713333e-01 1.20462857e-01 -6.67290986e-01 -1.09865117e+00
-8.12391639e-01 -4.97131407e-01 -7.34141111e-01 3.57269682e-02
8.15085649e-01 -1.25047052e+00 -2.01276571e-01 2.11468458e-01
-1.15070927e+00 -4.44663882e-01 -5.65165102e-01 8.44302833e-01
-6.67357862e-01 2.67219931e-01 -7.11986899e-01 -9.98342156e-01
-3.35178822e-01 -1.10800886e+00 8.94203246e-01 -2.26011306e-01
-9.68365446e-02 -9.08181608e-01 1.59852728e-01 3.80732924e-01
3.39006245e-01 -5.08002579e-01 1.00613463e+00 -9.30067897e-01
-3.18747401e-01 -4.01150174e-02 2.46043369e-01 8.29461932e-01
2.22660303e-01 -1.19749699e-02 -1.45207322e+00 -3.78019452e-01
-1.00643553e-01 -2.82944798e-01 8.96259069e-01 3.94851863e-01
1.32736468e+00 -3.96884292e-01 -2.01308459e-01 6.39427602e-01
8.28263283e-01 4.70752656e-01 5.07286191e-01 -5.35646603e-02
6.80280566e-01 5.75416148e-01 3.98468524e-01 2.75705755e-01
2.12889537e-01 6.96114242e-01 9.24236476e-02 4.09364663e-02
-4.07718778e-01 -3.83232653e-01 8.10411215e-01 1.83399689e+00
-1.74111873e-01 -1.79894835e-01 -9.32988107e-01 5.24009466e-01
-1.82192838e+00 -6.56848133e-01 5.06126322e-02 2.21926045e+00
9.99418914e-01 1.75958097e-01 -2.40909353e-01 2.71307170e-01
8.00875545e-01 1.23744838e-01 -3.74439120e-01 -4.88652617e-01
-1.39611810e-01 2.44799390e-01 4.19446647e-01 7.62789786e-01
-1.14678192e+00 1.18069875e+00 6.65719795e+00 7.35813498e-01
-1.22914004e+00 4.86003757e-01 8.00462067e-01 -3.73493619e-02
-8.68866593e-02 4.66052582e-03 -1.01665616e+00 4.81735826e-01
1.65752304e+00 1.92065179e-01 5.92941999e-01 8.89287293e-01
2.72839487e-01 2.81359315e-01 -1.38602495e+00 1.11489666e+00
7.07485005e-02 -1.43115437e+00 -5.90051003e-02 5.28709441e-02
6.82197034e-01 4.47085559e-01 2.52456844e-01 8.45946848e-01
2.78105110e-01 -8.23118806e-01 6.87032461e-01 2.33943909e-01
1.03639340e+00 -4.92515087e-01 5.17183006e-01 4.38510895e-01
-1.09628570e+00 -1.10919677e-01 -2.46616840e-01 -7.40791112e-02
1.77273497e-01 3.60168844e-01 -8.91696155e-01 1.00077726e-01
5.71942687e-01 8.14691484e-01 -4.82331604e-01 3.60317737e-01
-2.77219087e-01 1.25141692e+00 -3.87028575e-01 4.47537541e-01
2.12326035e-01 -6.96448311e-02 2.74645507e-01 1.41505289e+00
2.71642119e-01 9.01933238e-02 2.04905093e-01 2.97656119e-01
-2.25083128e-01 3.24212223e-01 -5.28633893e-01 -1.58473417e-01
6.45697296e-01 7.30255902e-01 -4.09979403e-01 -4.76297051e-01
-7.79270172e-01 7.94028699e-01 5.94990075e-01 5.21951616e-01
-5.67584753e-01 2.29512259e-01 4.49855894e-01 -6.04225881e-02
2.69043148e-01 -3.48477423e-01 -5.36637791e-02 -1.29053557e+00
-9.56556052e-02 -9.15321946e-01 1.65771380e-01 -3.66331965e-01
-1.05758333e+00 8.65790188e-01 -1.99597970e-01 -1.23631489e+00
-6.99261785e-01 -6.36281312e-01 -4.45346385e-01 6.83633149e-01
-1.61926079e+00 -8.21302950e-01 2.71297842e-01 6.16879463e-01
1.15023065e+00 -6.38598680e-01 1.20819128e+00 3.55645657e-01
-7.27387130e-01 5.55760086e-01 3.54325920e-01 1.16302997e-01
6.72108829e-01 -9.47704673e-01 4.61363733e-01 1.08151734e+00
6.61874831e-01 4.67707247e-01 2.32667103e-01 -3.35029721e-01
-1.11549640e+00 -1.14512813e+00 1.03132570e+00 -2.00517476e-01
5.72693765e-01 -6.06709540e-01 -1.04511034e+00 1.04396427e+00
9.34530422e-02 4.94026959e-01 9.50020730e-01 4.43622470e-01
-4.77135390e-01 -2.14382231e-01 -5.61716437e-01 2.33190343e-01
8.87042761e-01 -9.24782336e-01 -5.96432149e-01 5.11923730e-01
1.13322937e+00 -1.82778671e-01 -5.28012633e-01 3.93294364e-01
2.44403392e-01 -5.10517776e-01 8.16945016e-01 -6.15585923e-01
7.38832541e-03 1.28975675e-01 -7.42503107e-01 -1.17394614e+00
-2.00957999e-01 -7.54603684e-01 -3.61821234e-01 1.36027539e+00
8.24007988e-01 -4.04851794e-01 6.79490983e-01 4.51479793e-01
-6.45645678e-01 -4.91426826e-01 -1.28888726e+00 -8.66145551e-01
1.66763946e-01 -8.31874907e-01 1.37466684e-01 8.46432447e-01
1.26731381e-01 5.41666985e-01 -5.22910535e-01 1.94779202e-01
1.26779407e-01 -2.25987360e-01 3.39424819e-01 -1.08295333e+00
-3.17963749e-01 1.09074113e-03 -2.89210498e-01 -1.31972837e+00
5.03404379e-01 -1.06526136e+00 3.69719088e-01 -7.88150549e-01
-1.90170512e-01 -6.83519483e-01 -3.87251973e-01 6.54248059e-01
-6.77104387e-03 1.34711526e-02 2.16141462e-01 5.16959071e-01
-4.76859689e-01 6.53318763e-01 6.64706051e-01 -1.50749281e-01
-4.86166388e-01 3.97540107e-02 -2.08700210e-01 8.01381946e-01
7.65964150e-01 -3.92738014e-01 -6.02016211e-01 -6.75969839e-01
-2.19620496e-01 7.88324326e-02 -2.98218280e-01 -9.59545791e-01
2.23423153e-01 1.57194108e-01 4.18557487e-02 -4.77864206e-01
6.04540408e-01 -5.44835091e-01 -3.32799882e-01 2.49366701e-01
-6.65185571e-01 -1.08511997e-02 1.67211309e-01 7.50361383e-01
-4.82353151e-01 -2.40891442e-01 8.38445842e-01 2.06985399e-02
-8.48995209e-01 1.11995362e-01 -7.81507134e-01 6.38660863e-02
6.04751229e-01 1.68886691e-01 -1.85178116e-01 -4.70304549e-01
-1.22516429e+00 -6.07683733e-02 -1.82134375e-01 6.75789118e-01
6.65299356e-01 -1.45682573e+00 -5.50285876e-01 5.50758421e-01
1.30240887e-01 -1.38482541e-01 -4.87928502e-02 8.07514966e-01
-5.56508079e-02 7.99521387e-01 2.63515621e-01 -7.50210643e-01
-1.06630623e+00 6.35057390e-01 1.13890782e-01 -1.75509498e-01
-6.89890146e-01 9.18526351e-01 4.45069253e-01 -5.91754556e-01
6.64025187e-01 -4.81840700e-01 -2.35220283e-01 9.29909274e-02
4.13840979e-01 5.32593690e-02 2.37358943e-01 -8.70876372e-01
-4.45934832e-01 2.41826758e-01 -4.50285405e-01 -3.59050930e-01
1.38305306e+00 -3.78872842e-01 3.20171058e-01 8.70016992e-01
1.30669749e+00 -9.35965925e-02 -1.35923016e+00 -5.50738752e-01
1.92150056e-01 -7.46645257e-02 2.34000221e-01 -5.06581008e-01
-1.00120211e+00 1.05987513e+00 8.53537560e-01 -7.92581886e-02
8.81368995e-01 7.66681880e-02 4.96876597e-01 4.79966104e-01
2.36811623e-01 -1.32580256e+00 8.19306355e-03 8.52084696e-01
7.97380686e-01 -1.43528092e+00 -4.12734836e-01 -5.64280808e-01
-7.58005142e-01 1.00523031e+00 4.31515396e-01 1.26616359e-01
8.07069719e-01 2.77637333e-01 1.18048698e-01 1.95974857e-01
-1.22563982e+00 -2.23919377e-01 2.74995208e-01 7.47217059e-01
6.98504388e-01 1.36642843e-01 8.28812644e-02 5.58734536e-01
-1.48019850e-01 -4.68116552e-01 3.68377358e-01 8.76703143e-01
-3.86389732e-01 -1.12655389e+00 -2.29560196e-01 2.12127507e-01
-4.23570007e-01 -5.03411531e-01 5.15092611e-02 4.76607412e-01
9.62289125e-02 1.15575719e+00 3.82403702e-01 -3.26282114e-01
1.80826351e-01 6.70790374e-01 1.50861740e-01 -9.26223218e-01
-1.82366937e-01 4.88821208e-01 2.49684528e-01 -4.21917409e-01
-7.13240445e-01 -8.82193387e-01 -1.33191264e+00 1.47540599e-01
-4.44216043e-01 3.12683582e-01 6.74490213e-01 1.04572642e+00
4.50562686e-01 4.38954681e-01 8.10613275e-01 -7.93877840e-01
-3.31596553e-01 -1.27998126e+00 -2.98826396e-01 3.46802264e-01
5.06228387e-01 -5.98682046e-01 -5.43215692e-01 7.44473636e-01] | [14.451542854309082, 6.532403469085693] |
fca12cf1-f039-4f28-819f-dab66a660476 | web-scale-photo-hash-clustering-on-a-single | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Gong_Web_Scale_Photo_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Gong_Web_Scale_Photo_2015_CVPR_paper.pdf | Web Scale Photo Hash Clustering on A Single Machine | This paper addresses the problem of clustering a very large number of photos (i.e. hundreds of millions a day) in a stream into millions of clusters. This is particularly important as the popularity of photo sharing websites, such as Facebook, Google, and Instagram. Given large number of photos available online, how to efficiently organize them is an open problem. To address this problem, we propose to cluster the binary hash codes of a large number of photos into binary cluster centers. We present a fast binary k-means algorithm that works directly on the similarity-preserving hashes of images and clusters them into binary centers on which we can build hash indexes to speedup computation. The proposed method is capable of clustering millions of photos on a single machine in a few minutes. We show that this approach is usually several magnitude faster than standard k-means and produces comparable clustering accuracy. In addition, we propose an online clustering method based on binary k-means that is capable of clustering large photo stream on a single machine, and show applications to spam detection and trending photo discovery. | ['Yunchao Gong', 'Marcin Pawlowski', 'Louis Brandy', 'Fei Yang', 'Rob Fergus', 'Lubomir Bourdev'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['online-clustering', 'spam-detection'] | ['computer-vision', 'natural-language-processing'] | [-1.39085859e-01 -2.96373636e-01 -2.17949212e-01 -2.53110021e-01
-6.11717880e-01 -9.13291276e-01 4.71680641e-01 8.07499766e-01
-4.89381015e-01 3.73878516e-02 -7.43143782e-02 8.90398175e-02
1.65440962e-02 -1.19581902e+00 -7.44425356e-01 -6.23701930e-01
-4.16174471e-01 6.06002510e-01 6.44389331e-01 2.52714157e-01
6.60698056e-01 3.69557381e-01 -1.98636520e+00 2.48199388e-01
5.05302131e-01 1.16354895e+00 1.11848958e-01 8.30678403e-01
-1.59834534e-01 3.75169158e-01 -3.55361521e-01 -3.93117815e-01
6.47060215e-01 -3.52122992e-01 -8.20706844e-01 5.04842281e-01
7.44217277e-01 -3.08850378e-01 -3.46635103e-01 1.26293862e+00
3.64900172e-01 1.22515984e-01 4.50686961e-01 -1.74563551e+00
-4.62924391e-01 8.70722055e-01 -7.80456901e-01 5.24909347e-02
2.54370987e-01 -3.28912556e-01 7.89781570e-01 -3.42110664e-01
5.88207781e-01 1.08377862e+00 7.18776882e-01 -9.15880278e-02
-1.01900589e+00 -5.45078278e-01 -2.09987909e-01 5.14551044e-01
-1.89662158e+00 -1.62533492e-01 3.82903636e-01 -4.62341905e-01
5.48944235e-01 4.95683789e-01 7.77854800e-01 -9.70418528e-02
-4.74916250e-01 5.16297579e-01 9.17750478e-01 -2.03276530e-01
6.92603767e-01 3.21091890e-01 2.78638780e-01 3.58389705e-01
4.20496434e-01 -7.96286047e-01 -3.01105678e-01 -5.82369745e-01
3.23970348e-01 4.05512631e-01 7.56102800e-02 -2.80855417e-01
-1.26408482e+00 9.68749344e-01 6.87977374e-01 1.89568877e-01
-3.88559401e-01 2.70656675e-01 4.33467537e-01 1.54622376e-01
2.41925582e-01 3.71910073e-02 2.16120914e-01 8.96319672e-02
-1.28573537e+00 -2.28845254e-02 9.86579716e-01 9.72678304e-01
1.37525380e+00 -7.80838072e-01 3.83562744e-01 6.01815999e-01
4.75080907e-02 6.79735482e-01 6.07794702e-01 -1.19281995e+00
1.47153184e-01 7.40636587e-01 2.22830236e-01 -1.57285404e+00
-1.88944966e-03 4.03969914e-01 -1.15464485e+00 -3.26916963e-01
2.81344622e-01 2.56761819e-01 -6.70374870e-01 1.19804811e+00
8.69774342e-01 3.23420554e-01 -2.92731583e-01 7.84977496e-01
2.70782024e-01 9.02993262e-01 -1.11156553e-01 -3.96551102e-01
1.42937326e+00 -7.02255130e-01 -2.74798095e-01 1.29596457e-01
3.27037126e-01 -7.62825847e-01 7.19816148e-01 2.88589746e-01
-9.91394162e-01 -5.37489414e-01 -6.69023812e-01 1.92724243e-02
-6.41284585e-01 -9.33720842e-02 3.81848574e-01 6.33023262e-01
-1.37964153e+00 8.69825065e-01 -7.52722621e-01 -1.16898537e+00
3.91821653e-01 3.50956380e-01 -2.88139552e-01 -1.57388046e-01
-6.17339194e-01 2.26539880e-01 3.86561006e-01 -4.91371483e-01
-2.90068984e-01 -3.77353698e-01 -4.59744066e-01 1.63842410e-01
-4.70439717e-02 -3.27678114e-01 8.58455658e-01 -1.04389060e+00
-9.38639164e-01 9.71062422e-01 -2.28579462e-01 -6.48305953e-01
2.04973191e-01 2.52895057e-01 -2.38458544e-01 8.51214290e-01
3.84950280e-01 9.28476036e-01 1.06619811e+00 -9.60782468e-01
-9.38576639e-01 -5.73837996e-01 -4.07810152e-01 1.06238142e-01
-1.15613866e+00 5.71977869e-02 -5.24456799e-01 -1.43861502e-01
2.24294454e-01 -1.22771180e+00 -4.30687487e-01 7.26456791e-02
-3.08786422e-01 -4.84360099e-01 8.23846281e-01 -1.53381273e-01
1.25823438e+00 -2.32181048e+00 -3.22214097e-01 6.12031937e-01
3.24037194e-01 5.38270809e-02 1.12913370e-01 8.34569752e-01
8.73766318e-02 3.20374757e-01 -1.57882616e-01 -5.33599369e-02
1.06786042e-02 -6.32139668e-02 -2.97801167e-01 8.34397078e-01
-5.13638258e-01 5.00730991e-01 -8.99847984e-01 -7.84565866e-01
1.82363883e-01 1.90266445e-01 -3.18750352e-01 4.02820632e-02
5.83772846e-02 -2.80537158e-01 -2.28299171e-01 4.50926036e-01
9.66701090e-01 -8.28422725e-01 3.12064409e-01 -2.23731011e-01
-2.60483831e-01 -1.72391191e-01 -1.42397428e+00 1.03159583e+00
2.04927102e-01 4.49042827e-01 -1.10196300e-01 -1.02842212e+00
7.18989909e-01 1.69057064e-02 8.55912983e-01 -1.68753594e-01
2.56014522e-02 -2.17750520e-02 -6.87869191e-01 -1.98789075e-01
7.78920531e-01 6.36203066e-02 -1.39353022e-01 1.06286812e+00
-4.58466113e-01 4.62479480e-02 5.53757071e-01 7.10705042e-01
1.24024868e+00 -1.09106636e+00 3.11363816e-01 -6.27390742e-02
3.57481003e-01 4.19141710e-01 1.10010736e-01 6.73892975e-01
-2.69473404e-01 6.10750258e-01 1.41615659e-01 -5.69128215e-01
-1.46922350e+00 -8.24669361e-01 6.60904199e-02 7.94174433e-01
5.63282013e-01 -7.56896138e-01 -9.36536252e-01 -3.28383058e-01
3.24694663e-01 -2.87642963e-02 -3.15075338e-01 2.41217822e-01
-2.99367040e-01 -8.47361326e-01 1.82024971e-01 1.31869957e-01
5.57278037e-01 -7.79580057e-01 -2.77016640e-01 1.42890021e-01
-4.04363900e-01 -1.14808345e+00 -4.95406121e-01 -5.84849954e-01
-9.42282677e-01 -1.21395957e+00 -6.94983602e-01 -9.97185826e-01
9.66927946e-01 1.20777285e+00 8.90916526e-01 3.13316226e-01
-3.70830059e-01 6.77947640e-01 -7.35548913e-01 1.82092898e-02
-2.67183334e-01 7.06502572e-02 2.47733936e-01 4.29445624e-01
7.26581931e-01 -8.74997616e-01 -9.54462767e-01 4.69796658e-01
-1.15353966e+00 -4.96457994e-01 4.25135672e-01 -4.10224795e-02
6.32310033e-01 7.52934933e-01 2.09751636e-01 -6.35970414e-01
2.56133467e-01 -9.54171360e-01 -7.31578767e-01 1.12600334e-01
-6.91713035e-01 -5.48554599e-01 7.03632951e-01 -4.89936858e-01
-1.28062800e-01 7.09636927e-01 4.24292982e-01 -3.53364080e-01
-2.32798740e-01 1.33361548e-01 3.67198408e-01 -2.62985826e-01
4.95399356e-01 3.96343917e-01 1.11061990e-01 -3.57921660e-01
8.17891598e-01 1.17775416e+00 5.70652127e-01 9.67347547e-02
1.09801853e+00 1.03832519e+00 -1.41340196e-01 -1.25174606e+00
-2.28630483e-01 -1.43458354e+00 -6.56230688e-01 -2.56547779e-01
7.19825387e-01 -1.14213514e+00 -1.06984746e+00 6.25442624e-01
-6.86797738e-01 -1.32582799e-01 -6.76757544e-02 1.03924431e-01
-4.37830865e-01 1.12799430e+00 -7.50702202e-01 -5.08923531e-01
-4.19666976e-01 -3.98761451e-01 9.71752405e-01 2.81426162e-01
-1.27030939e-01 -4.71960723e-01 2.22426623e-01 2.66162723e-01
1.44714490e-01 -7.35825226e-02 4.34729517e-01 -6.05192423e-01
-8.62756729e-01 -6.69374466e-01 -4.86904025e-01 9.84351262e-02
-7.55589232e-02 1.68586150e-01 -5.19299686e-01 -4.69764441e-01
-3.47257495e-01 -2.30535164e-01 6.67158425e-01 3.31053644e-01
1.17673731e+00 -7.23987520e-01 -6.82471514e-01 2.01604187e-01
1.69480693e+00 -2.69229323e-01 8.55583072e-01 2.53734291e-01
6.01289928e-01 4.65744585e-01 6.24210358e-01 8.58822882e-01
7.17011750e-01 2.86811471e-01 3.09303492e-01 1.64038703e-01
2.61080295e-01 -3.46131772e-01 7.72864148e-02 1.07586670e+00
3.37877899e-01 -2.14262784e-01 -6.83535397e-01 1.06476092e+00
-1.86021686e+00 -1.29826534e+00 -4.66857880e-01 2.38002801e+00
5.87755740e-01 -3.76205474e-01 7.68949866e-01 3.62920761e-01
1.47892380e+00 -8.98639187e-02 -3.93585950e-01 5.27585596e-02
1.82752654e-01 -3.14122289e-01 9.84166026e-01 2.57838964e-01
-1.20691240e+00 8.00181627e-01 6.24564600e+00 8.03289592e-01
-9.24296141e-01 8.08904618e-02 7.37566710e-01 2.22505108e-01
1.49687842e-01 2.74316758e-01 -8.82923126e-01 9.09908295e-01
1.15859020e+00 -5.98320603e-01 5.17456353e-01 1.15818155e+00
1.66326597e-01 -4.84748989e-01 -8.72181356e-01 1.51039422e+00
3.04749459e-01 -1.29977548e+00 1.28227979e-01 3.16806644e-01
1.02824819e+00 1.07568584e-01 -1.10310853e-01 -4.02938217e-01
3.57497662e-01 -4.87547845e-01 4.39892918e-01 -5.72796091e-02
5.89635134e-01 -6.31649792e-01 2.82366395e-01 4.17125732e-01
-1.39546144e+00 -3.28233153e-01 -8.60041320e-01 -6.86826706e-02
2.92791843e-01 1.12570488e+00 -8.26239526e-01 -2.79903635e-02
1.13903308e+00 8.48778784e-01 -7.90969372e-01 1.73544049e+00
4.93870109e-01 5.34800708e-01 -8.70892346e-01 -2.36638993e-01
2.02260762e-01 -3.88475060e-01 8.74969140e-02 1.14232087e+00
7.54602551e-01 4.78107780e-01 3.15297782e-01 1.50778711e-01
-2.59160221e-01 4.20388669e-01 -6.74061716e-01 -2.27137551e-01
8.67849410e-01 1.64873815e+00 -1.50825691e+00 -1.02671683e+00
-6.13166355e-02 1.40356505e+00 1.25324488e-01 -1.61741689e-01
-6.27119720e-01 -6.04224145e-01 5.48214018e-01 6.41269028e-01
6.64835155e-01 -2.57881314e-01 2.01274112e-01 -9.08891797e-01
-1.24994762e-01 -5.85040510e-01 6.25701964e-01 -6.73267007e-01
-1.57311201e+00 3.74737591e-01 -3.06783378e-01 -1.55394125e+00
6.69757947e-02 8.12803023e-03 -5.94999552e-01 5.41794114e-02
-1.23891163e+00 -6.88137352e-01 -6.16343141e-01 8.46526206e-01
2.90946122e-02 3.78127486e-01 4.17735249e-01 5.19075930e-01
-1.69149935e-01 2.49945045e-01 6.13258719e-01 1.14474334e-01
8.91044259e-01 -1.22652268e+00 6.38960481e-01 5.76555312e-01
1.80023327e-01 1.96238503e-01 5.57040811e-01 -6.93084061e-01
-1.30265820e+00 -1.21670461e+00 1.23343360e+00 -2.09233612e-01
6.88673317e-01 -2.11636782e-01 -6.22139335e-01 3.74188483e-01
7.81470686e-02 1.48252636e-01 7.84543455e-01 -4.82473761e-01
-4.64657962e-01 -4.28466022e-01 -1.30498397e+00 2.26863459e-01
7.49274075e-01 -4.58892763e-01 -5.39045669e-02 1.16122139e+00
3.95151973e-01 2.83475518e-01 -9.18346167e-01 -3.83392215e-01
3.32822949e-01 -1.09784663e+00 1.08941984e+00 -3.45429964e-03
3.24859321e-02 -6.42557859e-01 -4.40973230e-02 -9.34644163e-01
-4.60945934e-01 -8.06599140e-01 2.10221291e-01 1.24302864e+00
-1.70561060e-01 -4.17658478e-01 1.03660440e+00 5.04851341e-01
7.60270476e-01 1.29132017e-01 -6.52969658e-01 -7.95919240e-01
-5.15428901e-01 1.86968192e-01 4.83292967e-01 1.11946750e+00
4.18788940e-01 1.46448374e-01 -2.15688497e-01 3.50529730e-01
1.17280746e+00 4.72739935e-01 1.00203872e+00 -1.39164042e+00
1.87108859e-01 -2.17503801e-01 -9.03237820e-01 -9.65485692e-01
-2.57204741e-01 -7.82135069e-01 -4.97938469e-02 -1.31547892e+00
5.62953115e-01 -7.29251146e-01 1.35045841e-01 3.68555963e-01
-5.96012250e-02 9.61423278e-01 2.60999084e-01 8.36956382e-01
-1.15726995e+00 -3.72197595e-03 2.68606842e-01 -3.83109897e-02
-1.04040630e-01 -5.75818270e-02 -4.05410975e-01 5.41535556e-01
7.65556812e-01 -7.68381476e-01 -1.73130974e-01 4.25973274e-02
3.40696633e-01 -2.95630157e-01 2.57216215e-01 -1.27253401e+00
8.31743598e-01 8.34037289e-02 9.16681066e-02 -8.37407112e-01
5.70854843e-02 -9.36944067e-01 3.33907694e-01 3.62712801e-01
-1.84896901e-01 1.73991412e-01 -3.28597754e-01 9.34378088e-01
-3.62662554e-01 -8.46966282e-02 1.02320850e+00 -5.08228362e-01
-6.20276809e-01 4.44638938e-01 -5.33343852e-01 -8.59360397e-02
1.46707308e+00 -2.09772930e-01 -2.72887588e-01 -7.28760958e-01
-4.23604459e-01 3.14612716e-01 9.31421399e-01 1.18789032e-01
3.90241921e-01 -1.49704087e+00 -4.73656148e-01 -1.03134453e-01
8.79972503e-02 -2.66537786e-01 3.44763190e-01 5.87934315e-01
-7.97731161e-01 1.11348234e-01 -1.00430302e-01 -7.99898207e-01
-1.67964506e+00 1.03115833e+00 -3.96210104e-01 3.46836418e-01
-5.67831933e-01 5.03628731e-01 -2.91294754e-01 -1.19293956e-02
2.63721734e-01 1.18040495e-01 6.56639636e-02 4.32989597e-01
1.11276221e+00 7.04457045e-01 -1.24799527e-01 -7.63276875e-01
-2.72655517e-01 8.87264967e-01 1.95766538e-01 -1.71401538e-02
1.43991184e+00 -6.28336251e-01 -7.39864111e-01 1.24151766e-01
1.61213458e+00 -2.09614679e-01 -7.49495685e-01 -3.99154425e-02
-5.88005781e-02 -7.68844843e-01 -2.85178572e-01 6.34426251e-02
-9.67096448e-01 6.01172030e-01 5.96761405e-01 9.27700281e-01
1.07163870e+00 1.52648464e-01 1.47941768e+00 6.14198327e-01
3.13695341e-01 -1.20332193e+00 1.19783103e-01 -2.49769241e-01
7.73621947e-02 -1.13467860e+00 2.52497375e-01 -5.42452157e-01
-4.01683539e-01 9.17244673e-01 -4.59412709e-02 -4.32274967e-01
9.55470920e-01 -1.34961545e-01 -8.65070224e-02 -1.61186084e-01
-4.01293904e-01 -3.14492464e-01 -2.59398907e-01 4.07762796e-01
-3.97273600e-01 1.82746485e-01 -3.14919233e-01 -2.61771947e-01
-1.71558812e-01 -8.60848576e-02 7.74960577e-01 9.33400154e-01
-1.18744802e+00 -9.97745335e-01 -7.86659181e-01 6.12515986e-01
-1.65310383e-01 -3.46558169e-02 -5.72690785e-01 2.63294458e-01
-1.20279342e-01 1.06910503e+00 4.48070467e-01 -4.45903361e-01
-1.76746160e-01 -3.03858127e-02 -5.38888834e-02 -3.19081753e-01
-3.58752906e-01 -9.34868529e-02 -4.29146051e-01 -5.54230332e-01
-6.15589201e-01 -5.64537406e-01 -9.56100821e-01 -9.48535562e-01
-1.66530326e-01 5.04875541e-01 9.49251711e-01 2.40506217e-01
6.42353952e-01 -9.20714557e-01 1.40708423e+00 -9.26116526e-01
-4.38229293e-01 -5.11554658e-01 -8.93888831e-01 8.36093187e-01
3.50429751e-02 1.77812666e-01 -7.39388883e-01 5.60559452e-01] | [7.44149112701416, 4.730705738067627] |
4b5c5ba6-72a5-4525-999f-9f2823126576 | a-benchmark-for-automatic-medical | 2204.08997 | null | https://arxiv.org/abs/2204.08997v3 | https://arxiv.org/pdf/2204.08997v3.pdf | A Benchmark for Automatic Medical Consultation System: Frameworks, Tasks and Datasets | In recent years, interest has arisen in using machine learning to improve the efficiency of automatic medical consultation and enhance patient experience. In this article, we propose two frameworks to support automatic medical consultation, namely doctor-patient dialogue understanding and task-oriented interaction. We create a new large medical dialogue dataset with multi-level finegrained annotations and establish five independent tasks, including named entity recognition, dialogue act classification, symptom label inference, medical report generation and diagnosis-oriented dialogue policy. We report a set of benchmark results for each task, which shows the usability of the dataset and sets a baseline for future studies. Both code and data is available from https://github.com/lemuria-wchen/imcs21. | ['Jiajie Peng', 'Zhongyu Wei', 'Xuanjing Huang', 'Qi Zhang', 'Jianye Hao', 'Cheng Zhong', 'Qianyuan Yao', 'Hongyi Fang', 'Zhiwei Li', 'Wei Chen'] | 2022-04-19 | null | null | null | null | ['medical-report-generation', 'dialogue-understanding', 'dialogue-act-classification'] | ['medical', 'natural-language-processing', 'natural-language-processing'] | [ 2.55786479e-01 1.05359542e+00 -3.98473531e-01 -7.55730808e-01
-1.07972550e+00 -5.57318747e-01 5.91525137e-01 4.79928881e-01
-3.22381020e-01 1.14428735e+00 8.32834899e-01 -4.55164075e-01
-6.49692118e-02 -2.37175286e-01 3.58695328e-01 -3.86871278e-01
1.94118649e-01 1.08980596e+00 -1.80004552e-01 -1.33704841e-01
-4.66367155e-02 7.21954629e-02 -6.57100797e-01 8.33340466e-01
9.67532456e-01 6.07684612e-01 -2.64910877e-01 1.16783619e+00
4.77858335e-02 1.24733949e+00 -7.18308628e-01 -6.63525879e-01
-3.38701159e-01 -7.73381472e-01 -1.66190028e+00 8.00712332e-02
-3.88065338e-01 -3.53876084e-01 4.80228662e-02 5.30399978e-01
1.07337844e+00 5.80872968e-02 5.79579771e-01 -8.59604478e-01
-2.59978712e-01 6.29559755e-01 2.11797640e-01 -8.24896172e-02
9.53519285e-01 2.36208603e-01 7.14565456e-01 -1.57107741e-01
8.04804325e-01 1.08067286e+00 4.69985962e-01 1.34957528e+00
-9.20839250e-01 -1.34784177e-01 -2.96127737e-01 -2.64366746e-01
-1.00534713e+00 -6.74941719e-01 2.43843034e-01 -3.23127151e-01
1.07382977e+00 8.17162335e-01 5.80630124e-01 1.34999514e+00
-8.19930211e-02 1.09386075e+00 1.16763949e+00 -4.32668447e-01
1.24172300e-01 5.24314225e-01 2.51075774e-01 7.23173559e-01
-4.41910356e-01 -2.23445684e-01 -1.73053607e-01 -8.71158421e-01
4.18092042e-01 -3.72654378e-01 -4.89967614e-01 2.94333160e-01
-1.19905555e+00 9.45635974e-01 -1.53256571e-02 3.82546067e-01
-3.43070835e-01 -4.53088433e-01 8.32818747e-01 1.86288223e-01
5.62799454e-01 7.91676223e-01 -9.13190007e-01 -6.52770340e-01
-3.29141527e-01 3.85179609e-01 1.75187719e+00 8.04350793e-01
-1.90730751e-01 -7.18813360e-01 -7.66115010e-01 1.13721347e+00
3.66116390e-02 1.00247234e-01 6.23567641e-01 -1.21271527e+00
3.03100735e-01 6.80550456e-01 2.52440453e-01 -5.43981910e-01
-8.88077378e-01 1.93295434e-01 -7.93597341e-01 -6.17069066e-01
3.57046902e-01 -7.56112576e-01 -5.52332580e-01 1.45023680e+00
5.93890190e-01 -1.52278602e-01 6.02418125e-01 6.85256183e-01
1.31115794e+00 2.39913642e-01 3.73809159e-01 -3.48375857e-01
1.59111893e+00 -1.16310406e+00 -1.06856155e+00 1.73303515e-01
1.30744624e+00 -7.97208428e-01 6.41812027e-01 3.52591187e-01
-1.32150328e+00 -5.67571335e-02 -6.95927441e-02 -2.00919658e-01
-2.22070590e-01 2.15216294e-01 8.63911450e-01 5.15614629e-01
-1.07121646e+00 2.19314650e-01 -7.68559098e-01 -6.78714752e-01
1.17774025e-01 4.17835057e-01 -1.82091773e-01 2.66509295e-01
-1.42860389e+00 1.04799473e+00 4.83704835e-01 -5.35326898e-02
-3.07979167e-01 -4.96692300e-01 -8.28821182e-01 -2.63193190e-01
2.41342366e-01 -1.14448822e+00 1.99575436e+00 -6.47376835e-01
-1.82348001e+00 1.46433020e+00 -2.23967270e-03 -4.87057328e-01
7.14798331e-01 -6.86400980e-02 -5.35598576e-01 1.95522159e-01
-1.60650715e-01 8.40453684e-01 6.07699947e-03 -8.32344234e-01
-8.76938701e-01 -2.12126598e-01 2.31310934e-01 5.63369155e-01
2.09571794e-01 1.79173574e-01 -4.61581469e-01 -3.03651720e-01
-6.57239556e-01 -1.06641662e+00 -5.56674242e-01 -4.79437470e-01
-9.16376770e-01 -4.99368042e-01 8.23874697e-02 -7.92014062e-01
1.47780728e+00 -1.75138974e+00 1.03533506e-01 -8.87917578e-02
4.11574453e-01 4.03507918e-01 8.41639861e-02 6.26046836e-01
-6.88149333e-02 2.19611719e-01 -3.45003247e-01 -2.09300771e-01
-1.38797462e-01 2.15148062e-01 1.06941171e-01 -7.44437426e-02
2.23596603e-01 1.17275929e+00 -9.63444233e-01 -6.66651905e-01
3.33490252e-01 2.93515891e-01 -5.57971299e-01 6.92456841e-01
-3.76191556e-01 1.19030821e+00 -9.49517488e-01 6.01886928e-01
2.97214668e-02 -5.96787095e-01 4.97310698e-01 -1.40050188e-01
2.15774640e-01 6.19571030e-01 -4.12920773e-01 1.75062966e+00
-5.02472103e-01 3.19454819e-02 2.30961129e-01 -5.77221990e-01
3.96681070e-01 9.19064105e-01 7.42712796e-01 -3.47924292e-01
2.71480590e-01 3.79350930e-02 9.91854593e-02 -1.10740006e+00
3.67234051e-01 7.17426687e-02 -3.91541958e-01 4.65879917e-01
-1.29471689e-01 -2.46223554e-01 1.07793525e-01 2.74674654e-01
1.29494953e+00 -1.59546003e-01 9.13333952e-01 2.39601661e-03
7.64726281e-01 3.27545881e-01 1.80630907e-01 5.56586683e-01
-4.04282033e-01 8.93809348e-02 4.49699819e-01 -4.25839186e-01
-5.47983289e-01 -6.10815644e-01 -5.36440253e-01 1.07955670e+00
-3.49168301e-01 -4.27570075e-01 -8.59477341e-01 -9.71770227e-01
-8.57877135e-02 9.77210104e-01 -5.87852359e-01 5.00915535e-02
-4.74328399e-01 -7.56185055e-01 8.44781637e-01 2.91793942e-01
3.73231351e-01 -1.41140592e+00 -4.78469253e-01 3.08112025e-01
-6.62534833e-01 -1.21957076e+00 -4.37038779e-01 -9.01735574e-03
-6.90341949e-01 -1.26962030e+00 -7.95984268e-01 -6.44222736e-01
6.80263519e-01 -6.81892216e-01 1.55391800e+00 1.32679477e-01
-6.26104236e-01 8.63087177e-01 -4.61740047e-01 -3.44925612e-01
-9.44233000e-01 3.49978179e-01 -3.63778949e-01 -5.23419499e-01
5.64971149e-01 1.90051238e-03 -8.54799390e-01 1.79365799e-01
-7.29409814e-01 3.98479730e-01 3.00344318e-01 1.03545856e+00
2.31594190e-01 -7.33292162e-01 5.67624688e-01 -1.53903234e+00
1.36157286e+00 -4.79454547e-01 1.06412344e-01 2.73318201e-01
-3.63283694e-01 -9.61944610e-02 1.72148705e-01 4.91250819e-03
-1.24892139e+00 8.52562338e-02 -7.45041192e-01 4.71882164e-01
-8.75490308e-01 5.50351977e-01 2.42002159e-01 3.38370472e-01
6.15402699e-01 -1.51386067e-01 -3.15164179e-02 -3.71298820e-01
4.90624338e-01 1.25172675e+00 3.44658852e-01 -4.79608566e-01
-1.96103469e-01 7.10058883e-02 -5.24992168e-01 -4.82015163e-01
-1.15778327e+00 -6.51105165e-01 -5.51749587e-01 6.56786337e-02
1.20735753e+00 -6.89284980e-01 -1.12907720e+00 -6.32997528e-02
-1.17418528e+00 -6.84252799e-01 -4.30542499e-01 3.76236260e-01
-6.23950064e-01 1.21174611e-01 -1.02622974e+00 -9.79443789e-01
-8.64447832e-01 -9.60614443e-01 1.15366673e+00 7.43783191e-02
-1.08328998e+00 -1.54886281e+00 3.55364978e-01 8.24690163e-01
1.11610562e-01 4.03706461e-01 8.60419095e-01 -1.43276978e+00
1.91494271e-01 -2.24661633e-01 3.63858975e-03 2.54629776e-02
3.92282724e-01 -3.35171461e-01 -9.19603467e-01 -1.14652380e-01
1.25195980e-01 -8.57574642e-01 3.97495210e-01 3.48868966e-01
1.25514078e+00 -6.54961824e-01 -5.69349647e-01 1.71007812e-01
5.74750185e-01 5.38624287e-01 3.34722459e-01 -1.54763952e-01
4.39905614e-01 9.29258406e-01 7.46133864e-01 7.95395494e-01
8.99017453e-01 5.90010047e-01 -1.91085383e-01 -4.65193480e-01
2.32565105e-01 2.50552356e-01 -2.37662658e-01 9.94943976e-01
-1.27216518e-01 -5.10465264e-01 -1.22866738e+00 5.01468301e-01
-1.94696569e+00 -5.32882869e-01 1.14443026e-01 1.77792728e+00
1.71705401e+00 -2.80137897e-01 3.29887033e-01 -3.75551194e-01
3.75280142e-01 -3.02522629e-01 -4.30160254e-01 -6.56333029e-01
4.83173519e-01 1.37770161e-01 1.36101723e-01 7.75921404e-01
-1.18436396e+00 9.20897722e-01 6.40453863e+00 5.33740044e-01
-6.86387181e-01 1.78981498e-01 1.18130147e+00 1.82559237e-01
-3.22459228e-02 -5.61958611e-01 -7.05373049e-01 2.49090418e-01
1.24178278e+00 -4.77365367e-02 2.66360100e-02 6.26352310e-01
2.51661241e-01 -1.76851466e-01 -1.40269232e+00 6.82662666e-01
-9.38074887e-02 -1.13863480e+00 -2.04928219e-01 -3.18052806e-02
5.00365019e-01 -1.72072932e-01 -2.02787369e-01 3.43838334e-01
8.48077416e-01 -1.01963186e+00 -4.48656350e-01 6.60104752e-01
9.38079298e-01 -4.11967665e-01 9.20872629e-01 3.15461487e-01
-6.10774040e-01 1.90407202e-01 3.15705448e-01 2.81938016e-01
3.42516452e-01 2.21128061e-01 -1.56016612e+00 6.50004625e-01
3.23196769e-01 5.22613645e-01 -2.72254348e-01 8.05088997e-01
-1.74740806e-01 5.70594609e-01 -4.60527539e-02 1.38322795e-02
1.67442188e-01 2.81637185e-03 4.06901300e-01 1.55560207e+00
-1.88310057e-01 6.98674619e-01 5.51247299e-01 4.69855845e-01
-6.99685887e-02 4.75377798e-01 -4.39179838e-01 -3.38357121e-01
1.78382069e-01 1.43900263e+00 -4.98859972e-01 -6.31610274e-01
-2.55387813e-01 1.18748927e+00 1.31826118e-01 1.43419161e-01
-6.67562783e-01 1.16328113e-02 2.86045581e-01 -2.74871260e-01
-4.72967744e-01 4.86036271e-01 -9.56630930e-02 -1.10662103e+00
-5.03639221e-01 -1.36877811e+00 7.51973152e-01 -2.44471207e-01
-1.25709045e+00 8.47938299e-01 -2.11367995e-01 -8.16238105e-01
-9.51383054e-01 -4.50871378e-01 -3.59581351e-01 6.66606724e-01
-1.19271505e+00 -1.07547486e+00 -2.50136197e-01 6.64233983e-01
6.09380484e-01 3.69520783e-02 1.79334354e+00 2.54283190e-01
-5.27813911e-01 6.07786059e-01 -2.44922891e-01 3.46756577e-01
1.27300334e+00 -1.47294211e+00 1.83739170e-01 -3.86021107e-01
-4.17250097e-01 5.34493566e-01 4.60202336e-01 -8.18040907e-01
-8.71976495e-01 -9.26301658e-01 1.14116049e+00 -9.07889783e-01
4.92499232e-01 -7.49855414e-02 -6.14219666e-01 5.56642473e-01
5.24490535e-01 -5.31112850e-01 1.40078378e+00 1.69710696e-01
3.58306438e-01 5.52707195e-01 -1.45497906e+00 6.26235962e-01
7.56782174e-01 -4.24912602e-01 -4.40161228e-01 1.01655030e+00
7.37421095e-01 -1.01758111e+00 -1.60201907e+00 5.80661237e-01
4.48416859e-01 -6.35358453e-01 7.45280921e-01 -1.10910642e+00
4.58169043e-01 4.47585553e-01 3.59503895e-01 -1.17859352e+00
6.80876821e-02 -9.62699294e-01 -8.74556676e-02 9.78417158e-01
7.82568455e-01 -6.65620208e-01 5.10839820e-01 1.39615762e+00
-6.91121742e-02 -1.21803951e+00 -5.44315517e-01 2.14557111e-01
-6.92969337e-02 -4.32240032e-02 2.29622528e-01 1.37247956e+00
6.71811104e-01 6.07972145e-01 -3.08357686e-01 -2.43149266e-01
-7.47207552e-02 5.21989539e-02 5.88952661e-01 -1.04153299e+00
-5.54553628e-01 -3.32103431e-01 1.25147641e-01 -9.21297610e-01
7.02577271e-03 -8.14182818e-01 -7.39369467e-02 -1.81615269e+00
2.50982523e-01 -2.99618483e-01 -2.15873495e-02 8.82070541e-01
-4.76909220e-01 -1.57610849e-01 -2.32784048e-01 1.23954520e-01
-9.74610567e-01 1.44190565e-01 1.37197673e+00 5.56888580e-02
-5.58388472e-01 4.86255229e-01 -7.16071784e-01 7.93951809e-01
9.81704831e-01 -2.47749031e-01 -1.72612891e-01 -6.34061843e-02
-6.71779737e-02 7.80633450e-01 -1.04801536e-01 -3.21792543e-01
7.38085657e-02 -9.50874090e-02 1.45292252e-01 -2.25903571e-01
3.56529385e-01 -5.63691616e-01 -2.55633473e-01 7.19957530e-01
-1.18648136e+00 -1.69583127e-01 3.46363723e-01 1.79625675e-01
-1.81196436e-01 -1.62899196e-01 4.76679802e-01 -4.91125196e-01
-1.80561349e-01 9.05479491e-02 -5.02533078e-01 3.71173650e-01
9.66854334e-01 4.26207364e-01 -2.40323633e-01 -5.60142577e-01
-1.40478659e+00 6.70611978e-01 -3.62891681e-03 2.96964049e-01
3.93256873e-01 -8.76504302e-01 -9.40565407e-01 -3.66914004e-01
2.03184545e-01 -1.74641922e-01 3.28026623e-01 1.15365982e+00
-3.79213214e-01 8.74414086e-01 1.66432604e-01 -2.72401214e-01
-1.68943238e+00 2.38856331e-01 4.74117219e-01 -9.59110916e-01
-4.80626225e-01 8.91981006e-01 1.14900157e-01 -9.24050093e-01
4.71305043e-01 -2.56674767e-01 -5.55815756e-01 -1.46255001e-01
5.09688437e-01 1.28845930e-01 -1.33654937e-01 -3.30814958e-01
-2.30535761e-01 -4.84984703e-02 -4.48306382e-01 -6.24448769e-02
9.04586792e-01 -3.55245143e-01 -2.41097406e-01 4.09794301e-01
7.76068032e-01 -2.19581604e-01 -3.73524278e-01 -1.88039482e-01
3.64470482e-01 6.28504083e-02 -2.46126711e-01 -1.64052331e+00
-5.28371930e-01 5.03337681e-01 6.01969481e-01 5.33882141e-01
1.12362838e+00 2.24455163e-01 8.34511936e-01 6.22393370e-01
-8.51415321e-02 -1.06905258e+00 -1.11290082e-01 5.05893171e-01
8.33070815e-01 -1.51773143e+00 -2.01447338e-01 -5.64917088e-01
-1.38197958e+00 8.19944561e-01 5.95551431e-01 5.28051078e-01
3.99538517e-01 4.05941069e-01 5.43565750e-01 -4.35226530e-01
-1.01770675e+00 -2.37158358e-01 4.58611667e-01 3.81465793e-01
1.21128786e+00 2.48139113e-01 -5.89492977e-01 6.53722584e-01
-1.12782821e-01 2.02952534e-01 1.46716118e-01 1.07869732e+00
9.90742296e-02 -1.58246064e+00 5.51111139e-02 6.01564109e-01
-9.27008688e-01 -3.41475993e-01 -1.07195699e+00 4.85911965e-01
-2.46528223e-01 1.18677413e+00 -3.12589735e-01 -1.47476852e-01
3.75160128e-01 4.39365119e-01 3.17344815e-01 -1.01429760e+00
-1.13408303e+00 2.35661641e-01 1.06662416e+00 -6.29470468e-01
-7.09183931e-01 -2.70485908e-01 -1.43455887e+00 2.99741309e-02
-2.34653115e-01 5.94988227e-01 2.73992300e-01 8.29358935e-01
6.00612700e-01 8.16121876e-01 3.30696613e-01 -8.81907940e-02
-4.52175498e-01 -1.15037739e+00 -9.77003202e-02 4.85476166e-01
1.97625965e-01 8.28195736e-02 3.85034718e-02 1.22192532e-01] | [12.408304214477539, 8.37809944152832] |
b4f40ab3-29ab-438e-afe5-4aa542192337 | chain-based-discriminative-autoencoders-for | 2203.13687 | null | https://arxiv.org/abs/2203.13687v3 | https://arxiv.org/pdf/2203.13687v3.pdf | Chain-based Discriminative Autoencoders for Speech Recognition | In our previous work, we proposed a discriminative autoencoder (DcAE) for speech recognition. DcAE combines two training schemes into one. First, since DcAE aims to learn encoder-decoder mappings, the squared error between the reconstructed speech and the input speech is minimized. Second, in the code layer, frame-based phonetic embeddings are obtained by minimizing the categorical cross-entropy between ground truth labels and predicted triphone-state scores. DcAE is developed based on the Kaldi toolkit by treating various TDNN models as encoders. In this paper, we further propose three new versions of DcAE. First, a new objective function that considers both categorical cross-entropy and mutual information between ground truth and predicted triphone-state sequences is used. The resulting DcAE is called a chain-based DcAE (c-DcAE). For application to robust speech recognition, we further extend c-DcAE to hierarchical and parallel structures, resulting in hc-DcAE and pc-DcAE. In these two models, both the error between the reconstructed noisy speech and the input noisy speech and the error between the enhanced speech and the reference clean speech are taken into the objective function. Experimental results on the WSJ and Aurora-4 corpora show that our DcAE models outperform baseline systems. | ['Hsin-Min Wang', 'Yao-Fei Cheng', 'Pin-Tuan Huang', 'Hung-Shin Lee'] | 2022-03-25 | null | null | null | null | ['robust-speech-recognition'] | ['speech'] | [ 1.07504338e-01 5.97608760e-02 3.51943135e-01 -6.35348618e-01
-1.09020698e+00 -3.19783390e-01 5.53764999e-01 -3.67641926e-01
-3.13485861e-01 3.90131444e-01 3.97847146e-01 -3.09511513e-01
2.21153826e-01 -4.96607602e-01 -5.43537796e-01 -8.06377590e-01
7.26486072e-02 1.93679165e-02 -6.78484365e-02 1.35075852e-01
-2.63193220e-01 1.89734519e-01 -1.55514622e+00 2.96802908e-01
8.23842525e-01 1.17342377e+00 4.86322552e-01 8.96272779e-01
1.65070862e-01 7.94026971e-01 -7.16325402e-01 -6.38619661e-01
1.10018075e-01 -8.28398287e-01 -5.51718473e-01 1.19432792e-01
9.41163525e-02 -1.03324473e-01 -6.36302590e-01 1.29667580e+00
7.93189406e-01 4.17299241e-01 5.94931483e-01 -9.34755683e-01
-7.49541461e-01 7.18117833e-01 1.05310269e-01 5.05509228e-02
1.04970060e-01 -1.47964597e-01 1.04659235e+00 -1.25296366e+00
1.95136845e-01 1.22184408e+00 6.71337366e-01 6.33797228e-01
-1.14908779e+00 -5.25483668e-01 -1.00408241e-01 4.84713554e-01
-1.48945510e+00 -9.37557817e-01 7.51802385e-01 -2.44109452e-01
1.33281231e+00 4.32680905e-01 4.17265266e-01 1.21079111e+00
2.98133399e-02 9.36642051e-01 1.08065879e+00 -7.64867127e-01
4.53692049e-01 4.75015044e-02 -2.11534232e-01 5.81918597e-01
-4.19367343e-01 5.13566554e-01 -4.12506431e-01 3.49094979e-02
3.56614769e-01 -3.72883558e-01 -4.12052929e-01 1.94819897e-01
-9.60333288e-01 7.08502769e-01 2.63385661e-03 5.61697721e-01
-3.22465092e-01 -2.23121151e-01 3.63047928e-01 3.10729802e-01
3.66632968e-01 -3.98146063e-02 -2.55513340e-01 -5.34759939e-01
-9.20240045e-01 -1.39993981e-01 9.37225044e-01 7.61734664e-01
6.46142602e-01 5.86329877e-01 1.95089672e-02 1.37305081e+00
5.97574770e-01 6.25459135e-01 8.67686570e-01 -9.20120478e-01
4.84089971e-01 -5.18855490e-02 -4.45452869e-01 -8.53877544e-01
1.89790085e-01 -5.74434996e-01 -9.53237057e-01 -1.69196740e-01
-2.24713013e-01 -1.96025327e-01 -9.77641284e-01 1.80615830e+00
8.51922929e-02 3.28594774e-01 5.75878739e-01 8.60843956e-01
6.74276650e-01 1.18537366e+00 -2.33063266e-01 -4.15235758e-01
9.22927141e-01 -1.13270855e+00 -1.29723024e+00 -3.90717924e-01
3.43181610e-01 -8.24296892e-01 8.67521346e-01 2.58694261e-01
-1.26082253e+00 -8.34894121e-01 -1.14956510e+00 -4.21287073e-03
-2.14597911e-01 5.69113374e-01 -2.61732906e-01 6.40479505e-01
-1.06886113e+00 3.76983106e-01 -8.74984384e-01 -4.40223841e-03
-1.97177172e-01 1.46537453e-01 -2.50064462e-01 1.82756141e-01
-1.51559734e+00 9.54473376e-01 4.14394796e-01 2.15618983e-01
-1.04365194e+00 -2.23779231e-01 -1.04916084e+00 2.65260339e-01
7.42462426e-02 -1.65942982e-01 1.66312575e+00 -7.82704771e-01
-2.14149642e+00 4.26196039e-01 -3.21130037e-01 -5.69133580e-01
-6.11642823e-02 3.02258832e-03 -1.14613545e+00 -1.17740363e-01
-2.29053929e-01 4.07990962e-01 8.96134436e-01 -1.27225149e+00
-5.30932784e-01 -6.96521848e-02 -4.47618216e-01 3.08059841e-01
-2.89065301e-01 -1.47418544e-01 -5.94785273e-01 -1.00911021e+00
1.16896771e-01 -9.53070402e-01 -2.49357428e-03 -5.66787004e-01
-5.27289689e-01 -3.82610291e-01 1.03614271e+00 -1.09590387e+00
1.55236626e+00 -2.50201201e+00 3.09328347e-01 1.55283749e-01
-3.59760404e-01 6.74595535e-01 -1.53137311e-01 4.55849081e-01
-5.72738424e-02 -1.39865473e-01 -7.30580270e-01 -6.78397834e-01
2.96524465e-01 6.50627434e-01 -4.25265878e-01 1.93108737e-01
3.24764282e-01 4.32659507e-01 -7.18164086e-01 -3.17942739e-01
2.60653257e-01 9.14798737e-01 -5.12372553e-01 6.00627363e-01
1.05099350e-01 1.72076002e-01 1.68070272e-01 3.48695934e-01
6.69295073e-01 3.90610188e-01 1.39787629e-01 -4.92757484e-02
-1.26758724e-01 9.21411633e-01 -9.87631083e-01 1.63288009e+00
-6.35050952e-01 6.43396616e-01 1.51699632e-01 -9.91442561e-01
1.07207882e+00 7.89200842e-01 1.19271941e-01 -6.98908508e-01
1.00715803e-02 4.82335597e-01 1.06322281e-01 -3.45633596e-01
3.95908624e-01 -3.07363093e-01 1.14975661e-01 2.29105413e-01
4.30909812e-01 -3.14315438e-01 -9.65145305e-02 -7.12538362e-02
7.61580408e-01 -7.24130943e-02 2.91445464e-01 1.00006118e-01
7.51884937e-01 -6.25879765e-01 8.90259385e-01 4.26245332e-01
-2.89509475e-01 8.67004156e-01 2.53192514e-01 1.63780555e-01
-9.60764587e-01 -1.23641491e+00 -2.75091380e-01 7.15852678e-01
-2.40071028e-01 -5.02310872e-01 -1.18341589e+00 -6.51073039e-01
-3.08246464e-01 9.20249164e-01 -1.84390068e-01 -3.48660648e-01
-4.93009746e-01 -3.59746367e-01 8.87022078e-01 5.36184192e-01
5.43681800e-01 -9.05197680e-01 1.16900407e-01 3.31848502e-01
-4.59512711e-01 -1.27229595e+00 -8.38217497e-01 4.49486047e-01
-4.42033321e-01 -3.76084924e-01 -6.85463548e-01 -1.12807369e+00
2.69424289e-01 -1.04205301e-02 7.04467535e-01 -3.57172608e-01
1.82615176e-01 5.38155101e-02 -5.00876546e-01 -2.06194460e-01
-8.76838148e-01 -2.95161396e-01 4.28571492e-01 3.72206092e-01
3.20045114e-01 -5.12764573e-01 -2.30969891e-01 2.89821297e-01
-8.19719911e-01 -1.73156187e-01 5.09932578e-01 1.10034657e+00
8.93935084e-01 1.50405169e-01 5.98180771e-01 -3.20531458e-01
6.14757895e-01 -3.01419139e-01 -4.93615240e-01 1.75600991e-01
-6.12784207e-01 1.21764854e-01 7.83514380e-01 -4.54123318e-01
-1.25396383e+00 6.37948066e-02 -9.67693150e-01 -7.28638649e-01
-5.77875078e-02 6.43179774e-01 -5.54384768e-01 4.43808675e-01
2.91518658e-01 6.73355997e-01 -1.18829556e-01 -7.03051567e-01
3.01515132e-01 1.54662442e+00 9.49021995e-01 -2.59558618e-01
4.68072623e-01 -2.39120439e-01 -8.04054558e-01 -1.03565204e+00
-7.11961985e-01 -4.03844446e-01 -3.53694916e-01 -3.43473516e-02
8.47578585e-01 -9.05716538e-01 -3.53120208e-01 5.66505194e-01
-1.36288154e+00 -6.78698570e-02 -3.39787662e-01 9.21204448e-01
-6.75657809e-01 5.56300521e-01 -6.23049259e-01 -1.05951798e+00
-2.84023523e-01 -1.43689489e+00 9.54919457e-01 -8.19017515e-02
4.51733731e-02 -9.03457999e-01 3.16992521e-01 8.63214210e-02
3.92486870e-01 -3.25977653e-01 9.11217213e-01 -8.77207518e-01
1.39508592e-02 -2.17472777e-01 2.37048015e-01 1.33548045e+00
2.71993071e-01 -1.20608367e-01 -1.43513465e+00 -3.52066875e-01
4.95336443e-01 -1.17757194e-01 6.07683837e-01 2.59097636e-01
1.26868677e+00 -6.48771286e-01 1.53341591e-01 5.77729404e-01
1.11164427e+00 8.11668873e-01 8.11165929e-01 -1.63721472e-01
4.03551996e-01 3.07809174e-01 1.64814427e-01 1.37284562e-01
3.70530456e-01 9.28186536e-01 -3.96793075e-02 2.42139950e-01
-4.49587941e-01 -4.94162709e-01 9.74722505e-01 2.41858387e+00
2.14590833e-01 -4.08258706e-01 -6.35599911e-01 6.62860632e-01
-1.58817446e+00 -1.00735569e+00 2.04930127e-01 2.14258718e+00
1.02219522e+00 -8.98275822e-02 -2.16267303e-01 5.51272869e-01
8.35822225e-01 2.66236037e-01 -4.21169996e-01 -7.65943170e-01
-1.54459611e-01 3.80108863e-01 -1.61663935e-01 6.94167972e-01
-1.11933160e+00 6.57757759e-01 6.00096512e+00 1.07294273e+00
-1.07444680e+00 3.10730159e-01 3.43645841e-01 2.08267406e-01
-1.30820230e-01 -7.56124631e-02 -6.13918662e-01 6.39207065e-01
1.64806974e+00 -4.36758585e-02 7.19450653e-01 9.27134931e-01
4.83082309e-02 3.74402791e-01 -1.01039875e+00 9.98000324e-01
1.68345109e-01 -8.73520017e-01 -1.55062705e-01 -3.44653502e-02
7.35177636e-01 5.36788478e-02 2.59765536e-01 5.05559504e-01
2.98882961e-01 -6.90903127e-01 8.78069997e-01 3.32683861e-01
9.81808543e-01 -7.94924140e-01 7.72466063e-01 3.28853756e-01
-1.19672835e+00 5.18886037e-02 -4.57541972e-01 4.45136636e-01
1.93089798e-01 6.50401533e-01 -7.64703870e-01 5.28515100e-01
5.82240164e-01 6.28274381e-01 1.27847135e-01 8.02662551e-01
-3.84128898e-01 9.52816010e-01 -9.66114849e-02 1.57045200e-01
2.28263453e-01 -1.94907963e-01 6.95598125e-01 1.58684766e+00
5.94191074e-01 4.94043678e-02 -1.72339112e-01 6.88898981e-01
-1.43085167e-01 -1.77381039e-01 -3.80431086e-01 -2.62804389e-01
7.55754709e-01 8.13123882e-01 9.96132493e-02 -3.70472819e-01
-3.92849684e-01 1.13158679e+00 2.42013350e-01 5.28361917e-01
-6.41906798e-01 -7.74310112e-01 9.73627090e-01 -4.89920706e-01
6.53903484e-01 -2.83678889e-01 3.40853669e-02 -1.26636767e+00
2.26984639e-02 -1.17426455e+00 -5.51150404e-02 -6.51267707e-01
-1.16324329e+00 1.23069000e+00 -4.49410170e-01 -1.28594422e+00
-6.21165514e-01 -4.95039970e-01 -3.46395195e-01 1.02032959e+00
-1.46228778e+00 -6.44356012e-01 2.58389235e-01 5.00086963e-01
8.21327865e-01 -4.70042855e-01 1.27274084e+00 5.36001086e-01
-7.77821541e-01 7.91007161e-01 5.08507073e-01 3.00265640e-01
3.30122918e-01 -1.21324217e+00 6.81227922e-01 1.11615539e+00
5.49551606e-01 4.07404721e-01 4.47175205e-01 -3.88500601e-01
-1.13296056e+00 -1.21622527e+00 1.21955466e+00 1.03824018e-02
4.23755020e-01 -4.31735784e-01 -1.16418433e+00 5.99245906e-01
3.54078084e-01 1.02297328e-01 7.48225451e-01 -3.02406996e-01
-3.69941711e-01 -1.85690165e-01 -7.97974110e-01 2.71594703e-01
8.03308249e-01 -1.34582961e+00 -7.71576166e-01 -1.03034519e-01
8.86630476e-01 -5.24147451e-01 -9.74238217e-01 3.57790768e-01
3.42069566e-01 -9.12760317e-01 6.70313179e-01 -1.88738063e-01
2.24973410e-01 -3.31335127e-01 -7.28828847e-01 -1.80160046e+00
-1.67983010e-01 -5.57149887e-01 -4.52560097e-01 1.30126607e+00
5.46972036e-01 -3.34885627e-01 6.16225749e-02 -3.51810530e-02
-7.30094075e-01 -6.14894211e-01 -1.30009234e+00 -1.08936203e+00
-6.91474527e-02 -7.40947247e-01 6.36152923e-01 7.78426826e-01
-9.78431255e-02 1.81017533e-01 -4.50295746e-01 4.64428335e-01
3.20549935e-01 -2.63150781e-01 2.72797704e-01 -6.89144075e-01
-5.48907518e-01 -1.96101949e-01 -4.31713015e-01 -1.33296359e+00
3.09733868e-01 -1.10570991e+00 6.24307752e-01 -1.23582196e+00
-2.28868157e-01 -1.67053565e-01 -3.90736103e-01 3.44724834e-01
-7.76991919e-02 -1.43730804e-01 1.84582993e-01 8.18393156e-02
-1.02087118e-01 1.04118049e+00 8.67461503e-01 -7.15636611e-02
-1.47209093e-01 1.13795444e-01 -1.35707974e-01 5.96643865e-01
5.79624653e-01 -5.31476378e-01 -2.84824848e-01 -5.48960865e-01
-4.67952877e-01 2.23994926e-01 3.32745425e-02 -1.27833712e+00
3.06721807e-01 2.70266265e-01 -1.81006819e-01 -4.60799813e-01
6.74665630e-01 -7.10852027e-01 2.20374703e-01 3.21207047e-01
-4.83988315e-01 -1.57308415e-01 4.09349352e-02 6.86390638e-01
-9.35385704e-01 -2.22871050e-01 8.84909630e-01 2.67290503e-01
-4.70906645e-01 -7.98059255e-02 -6.80077016e-01 -7.27984086e-02
5.77958524e-01 6.67246282e-02 -4.62767109e-02 -5.92365801e-01
-8.76753867e-01 -3.56999159e-01 -8.64571184e-02 5.41450918e-01
8.24112236e-01 -1.76165545e+00 -7.19377697e-01 6.41756237e-01
1.63137782e-02 -1.74880862e-01 8.47782791e-02 6.20694280e-01
-1.07603826e-01 6.20193362e-01 4.29792762e-01 -4.16600287e-01
-1.08691037e+00 2.40546897e-01 5.09215057e-01 4.53848206e-02
-3.83607686e-01 9.17497754e-01 1.11499444e-01 -7.86557317e-01
6.50267243e-01 -5.47958314e-01 8.66544917e-02 -2.98024386e-01
5.19746423e-01 2.90146887e-01 4.16377842e-01 -1.09888756e+00
-3.46842974e-01 2.04351127e-01 7.68104494e-02 -4.66588616e-01
1.30607533e+00 -2.36070052e-01 3.98379676e-02 5.50533712e-01
1.67065203e+00 9.24304724e-02 -1.29584277e+00 -5.27815402e-01
1.14779972e-01 -2.36352950e-01 4.00521398e-01 -8.62508237e-01
-1.12611961e+00 1.18517101e+00 7.86300421e-01 3.91440809e-01
1.36274159e+00 -1.54478773e-01 1.05537045e+00 1.65163964e-01
-9.23877582e-02 -1.18627357e+00 -7.34718889e-02 8.18577766e-01
8.74438882e-01 -8.87461901e-01 -6.50164008e-01 -1.12626143e-01
-9.12470222e-01 9.76611614e-01 3.37257773e-01 3.17545205e-01
7.60352492e-01 4.66856092e-01 2.04203323e-01 3.36408138e-01
-8.18426490e-01 -2.69820780e-01 4.46262449e-01 5.32213211e-01
3.60205084e-01 2.57106990e-01 -1.11321419e-01 9.37520683e-01
-5.17574549e-01 -2.25322261e-01 2.29164496e-01 6.85557783e-01
-4.15912837e-01 -1.24383211e+00 -3.73912364e-01 9.16118361e-03
-4.23460543e-01 -2.47678027e-01 -3.21281791e-01 1.51303038e-01
7.94644430e-02 1.21138704e+00 2.74160177e-01 -9.22451377e-01
4.52789813e-01 3.77811790e-01 -5.51694296e-02 -5.35658956e-01
-3.39453042e-01 3.69874477e-01 1.16322912e-01 -2.61345834e-01
-3.67936611e-01 -5.29213727e-01 -1.20851576e+00 9.09262300e-02
-7.07155287e-01 4.58352298e-01 1.08259940e+00 9.67867494e-01
4.37229186e-01 5.79983234e-01 1.07769275e+00 -5.86660266e-01
-7.59468973e-01 -1.17584658e+00 -5.13203204e-01 7.62797594e-02
7.25995421e-01 -2.77003646e-01 -5.82414031e-01 2.60510564e-01] | [14.797351837158203, 6.400570392608643] |
5d05fa79-42cc-423f-881b-b456ca1ecc11 | simultaneous-iris-and-periocular-region | 1908.00069 | null | https://arxiv.org/abs/1908.00069v1 | https://arxiv.org/pdf/1908.00069v1.pdf | Simultaneous Iris and Periocular Region Detection Using Coarse Annotations | In this work, we propose to detect the iris and periocular regions simultaneously using coarse annotations and two well-known object detectors: YOLOv2 and Faster R-CNN. We believe coarse annotations can be used in recognition systems based on the iris and periocular regions, given the much smaller engineering effort required to manually annotate the training images. We manually made coarse annotations of the iris and periocular regions (122K images from the visible (VIS) spectrum and 38K images from the near-infrared (NIR) spectrum). The iris annotations in the NIR databases were generated semi-automatically by first applying an iris segmentation CNN and then performing a manual inspection. These annotations were made for 11 well-known public databases (3 NIR and 8 VIS) designed for the iris-based recognition problem and are publicly available to the research community. Experimenting our proposal on these databases, we highlight two results. First, the Faster R-CNN + Feature Pyramid Network (FPN) model reported an Intersection over Union (IoU) higher than YOLOv2 (91.86% vs 85.30%). Second, the detection of the iris and periocular regions being performed simultaneously is as accurate as performed separately, but with a lower computational cost, i.e., two tasks were carried out at the cost of one. | ['Diego R. Lucio', 'Rayson Laroca', 'Gladston Moreira', 'Luiz A. Zanlorensi', 'David Menotti'] | 2019-07-31 | null | null | null | null | ['iris-segmentation'] | ['medical'] | [ 9.76007134e-02 2.58813322e-01 -1.33084401e-01 -1.65472731e-01
-5.53533733e-01 -5.94775200e-01 3.61012995e-01 1.61024798e-02
-4.62196708e-01 1.97674543e-01 -1.35549977e-01 -1.46685034e-01
-1.42970890e-01 -3.69967580e-01 -3.51003855e-01 -7.69578695e-01
1.35170072e-01 2.88075387e-01 -1.30795255e-01 1.19223885e-01
1.72821105e-01 9.34823394e-01 -2.02590895e+00 4.92386287e-03
9.38120186e-01 1.17385173e+00 -5.09358346e-01 9.55741107e-01
5.09383798e-01 5.24966121e-01 -5.61063051e-01 -4.44756091e-01
7.65371680e-01 -3.56627494e-01 -8.18318427e-01 4.42166805e-01
1.10986686e+00 -3.61157387e-01 -8.39183703e-02 1.19635141e+00
5.98029077e-01 -1.31433249e-01 3.23870450e-01 -6.94424570e-01
-7.97699690e-01 9.19560343e-02 -8.36891353e-01 -3.19559649e-02
2.41045132e-01 6.08487427e-01 8.11948419e-01 -7.23181605e-01
5.00535607e-01 6.54340982e-01 7.03071594e-01 3.54478836e-01
-1.23739934e+00 -4.82634366e-01 -6.19819820e-01 -4.63896543e-02
-1.90401149e+00 -6.12569630e-01 2.10635245e-01 -7.27318347e-01
9.14609611e-01 4.91358429e-01 8.31126869e-01 3.71801913e-01
-1.67715788e-01 4.42855716e-01 1.58463657e+00 -6.48163974e-01
-2.94302344e-01 4.57050920e-01 4.41758819e-02 8.81111741e-01
2.76716709e-01 5.97733259e-01 -2.22747788e-01 -1.09811485e-01
9.36329126e-01 -2.00474322e-01 -2.87285715e-01 -1.06874943e-01
-1.07963336e+00 4.60835963e-01 5.14520466e-01 3.14957917e-01
-3.05183440e-01 -2.91261524e-01 2.16119215e-01 2.44835779e-01
5.16108394e-01 7.36694753e-01 -4.32415128e-01 1.89541534e-01
-1.11477399e+00 -2.37547219e-01 4.80540782e-01 8.55634630e-01
8.73872459e-01 -3.59078944e-01 -2.61162519e-01 6.72295630e-01
2.46803313e-01 2.53418088e-01 3.00613701e-01 -5.47660768e-01
-1.07900985e-01 8.68750691e-01 2.98450470e-01 -7.55342603e-01
-5.68570077e-01 -5.56686997e-01 -9.38804984e-01 3.16173524e-01
8.21751714e-01 -1.51317224e-01 -1.26216555e+00 8.52476597e-01
4.69887584e-01 9.10766423e-02 3.94803211e-02 1.16982949e+00
1.23051524e+00 9.66060907e-03 -1.09907180e-01 -1.48873255e-01
1.66355669e+00 -9.33837652e-01 -4.47048187e-01 4.18457925e-01
7.31121957e-01 -1.26667976e+00 6.41422212e-01 4.23617125e-01
-7.65448213e-01 -9.19877648e-01 -6.73415184e-01 -2.41464943e-01
-4.82312918e-01 1.17125750e+00 5.24110436e-01 7.45424986e-01
-1.30787516e+00 4.01047140e-01 -4.84419823e-01 -4.03440595e-01
2.51385838e-01 7.52041996e-01 -4.64904636e-01 3.13503385e-01
-7.02573895e-01 7.49885857e-01 3.44748795e-01 4.73551929e-01
-4.99511838e-01 -3.28672379e-01 -8.99359107e-01 -1.14394858e-01
2.72002935e-01 -4.05083001e-01 9.61710751e-01 -1.29715979e+00
-1.78376770e+00 1.72641802e+00 -1.00451127e-01 -4.28377301e-01
3.93062025e-01 -1.36843294e-01 -5.54209232e-01 1.57604605e-01
-3.83355439e-01 4.22014087e-01 7.92970419e-01 -7.72955358e-01
-6.84483528e-01 -5.61797678e-01 1.22153468e-01 -1.14378713e-01
1.55475974e-01 5.03024399e-01 -7.30884612e-01 -3.62964660e-01
1.07589707e-01 -1.11805081e+00 -1.40098915e-01 -4.91021536e-02
-7.43452787e-01 -5.41686654e-01 1.97025210e-01 -9.01714802e-01
8.76026928e-01 -2.22334409e+00 -1.85379133e-01 4.15572613e-01
6.07470393e-01 9.17516708e-01 -1.24856405e-01 -7.51576647e-02
-3.50757331e-01 -7.82418400e-02 2.85354048e-01 -2.83186376e-01
-3.65204245e-01 -1.80409059e-01 6.71060905e-02 9.73562717e-01
4.83652912e-02 7.46416271e-01 -6.61850214e-01 -4.67690527e-01
4.90090638e-01 5.67320287e-01 -2.78645068e-01 6.88854381e-02
2.13358104e-01 5.18930137e-01 8.30342248e-03 1.13686979e+00
6.92471087e-01 -2.35322267e-01 -1.57644629e-01 -5.13365448e-01
-5.12017488e-01 -1.96334660e-01 -1.22725511e+00 1.39248538e+00
-2.19641879e-01 8.94154131e-01 8.37260708e-02 -5.28071284e-01
1.09575188e+00 4.89296168e-01 3.93126935e-01 -4.23246711e-01
4.27137971e-01 3.60459059e-01 8.09844732e-02 -5.29126823e-01
5.58723748e-01 2.86016077e-01 7.06436872e-01 3.54385786e-02
3.24670643e-01 5.00791222e-02 9.45091695e-02 -3.56448263e-01
3.46691608e-01 1.60518467e-01 5.19470453e-01 -1.62036568e-01
7.80013382e-01 1.42205313e-01 4.23857898e-01 6.68725908e-01
-4.55842137e-01 5.89383185e-01 3.79927337e-01 -8.93356502e-01
-9.18599844e-01 -4.99148011e-01 -6.68903410e-01 6.19536936e-01
1.31278321e-01 -4.32900548e-01 -6.99618518e-01 -5.31935096e-01
2.24271100e-02 -2.44930506e-01 -7.26294637e-01 3.85627389e-01
-8.58559012e-02 -7.04544902e-01 5.66301286e-01 1.51678383e-01
3.74901801e-01 -6.95915341e-01 -3.06407928e-01 -3.00398529e-01
1.79270908e-01 -9.95225549e-01 -2.33932078e-01 -3.05546373e-01
-5.30525386e-01 -1.68763971e+00 -7.89542019e-01 -7.11277008e-01
9.71764088e-01 -1.26018882e-01 9.44485307e-01 2.01630577e-01
-1.08223569e+00 2.97260314e-01 -9.91894454e-02 -5.77283561e-01
-2.18384430e-01 -2.09186032e-01 2.29731649e-01 4.95275110e-01
8.43354046e-01 6.49680123e-02 -7.30443656e-01 4.67341721e-01
-4.61169720e-01 1.98263917e-02 6.33275568e-01 7.63757169e-01
9.00040627e-01 -2.26473749e-01 -1.66812927e-01 -6.48310900e-01
-6.23944774e-02 1.23641118e-01 -1.40411532e+00 1.66736990e-01
-6.24450147e-01 -3.80187035e-01 4.02587324e-01 -2.58376122e-01
-5.68429172e-01 6.29938543e-01 -5.51347733e-02 -6.18263245e-01
-6.93320036e-01 -9.20822546e-02 4.74593192e-01 -8.33034098e-01
8.65816593e-01 1.82588138e-02 1.22415178e-01 -6.92997396e-01
2.99171597e-01 1.18487132e+00 6.48537874e-01 -1.16103128e-01
7.45231986e-01 4.84326333e-01 1.44299805e-01 -1.01129127e+00
-9.64579284e-01 -1.03019309e+00 -9.20968533e-01 -1.25486553e-01
1.04333997e+00 -1.09937108e+00 -1.28245330e+00 6.72334492e-01
-8.85414779e-01 1.18968725e-01 -4.07213688e-01 8.26988101e-01
-3.96053553e-01 4.20548260e-01 -6.71804905e-01 -7.80442476e-01
-4.86647189e-01 -1.30674183e+00 1.16203153e+00 6.02903903e-01
3.44749987e-02 -5.73976398e-01 4.58006524e-02 7.38470078e-01
2.85268039e-01 3.88286650e-01 2.42486298e-01 -5.52010894e-01
-6.26242936e-01 -4.56671566e-01 -7.09665954e-01 4.78352875e-01
1.34301484e-01 6.26050532e-01 -1.56646335e+00 -3.76621813e-01
-3.79836708e-01 -1.99678332e-01 7.67336845e-01 5.68309247e-01
1.28851974e+00 -1.45501077e-01 -5.50311804e-02 1.05801415e+00
1.54103541e+00 -6.46812990e-02 6.04667902e-01 1.19024314e-01
6.24878585e-01 7.49692440e-01 5.27345240e-01 2.56544590e-01
-3.38355638e-02 8.60711396e-01 4.17031765e-01 -9.32386458e-01
-3.78234297e-01 1.65895432e-01 -6.82659745e-02 2.48800084e-01
-9.26905811e-01 3.79502475e-01 -1.05104256e+00 7.39565611e-01
-1.52228606e+00 -6.14070296e-01 -4.01225716e-01 2.40280962e+00
9.97029722e-01 -5.02868652e-01 2.80366659e-01 -1.52618855e-01
6.38477445e-01 -3.01697165e-01 -2.30405495e-01 -2.44003177e-01
-2.33625233e-01 5.23178041e-01 9.66075003e-01 4.88317996e-01
-1.33028448e+00 8.50519300e-01 6.68877792e+00 6.98247492e-01
-1.33837748e+00 -2.36153796e-01 6.70644641e-01 -2.03528494e-01
6.81594253e-01 -1.30268931e-01 -9.79105353e-01 1.52604431e-01
9.03037012e-01 3.17316592e-01 3.99999887e-01 7.88504422e-01
7.69190192e-02 -3.12622815e-01 -9.49438155e-01 1.40464675e+00
5.35801100e-03 -1.39604855e+00 -3.82357240e-01 2.86162853e-01
7.59233892e-01 4.00959074e-01 1.95240807e-02 -2.33955637e-01
-3.98325548e-02 -1.30527604e+00 5.91344526e-03 7.39692032e-01
1.37830317e+00 -5.90113223e-01 1.07315373e+00 -1.58467945e-02
-9.81968105e-01 1.08101957e-01 -3.93639237e-01 8.64498615e-02
-6.68226600e-01 5.64078391e-01 -9.59418416e-01 5.61598003e-01
8.81088734e-01 8.18509221e-01 -8.45450938e-01 1.30747676e+00
-2.08927467e-01 4.20045227e-01 -2.25807771e-01 2.47606605e-01
4.56551574e-02 -4.49135870e-01 5.27768254e-01 9.87894475e-01
4.48791459e-02 1.91323534e-02 4.61854674e-02 1.00942314e+00
-1.05293818e-01 4.43652570e-01 -3.95633131e-01 -4.92348261e-02
-1.66553348e-01 1.61409450e+00 -4.27755862e-01 -3.81972760e-01
-4.86496150e-01 6.41627789e-01 -9.76068079e-02 3.49979371e-01
-4.57800925e-01 -5.96237779e-01 9.00661588e-01 6.96833804e-02
3.75508443e-02 2.02900425e-01 -8.67392793e-02 -1.16851795e+00
-1.19110815e-01 -7.85401285e-01 2.27933913e-01 -7.35478938e-01
-1.00196064e+00 7.03849673e-01 -5.38904011e-01 -1.38424468e+00
1.56648546e-01 -9.79403257e-01 -1.73322529e-01 1.40512323e+00
-1.78621149e+00 -1.28960979e+00 -5.70945084e-01 5.94959557e-01
-4.00715582e-02 -2.65043169e-01 9.90981877e-01 2.95747787e-01
-8.37252617e-01 7.32685030e-01 4.96184640e-02 6.31580710e-01
8.76221716e-01 -1.25508857e+00 3.31973955e-02 8.10125887e-01
2.21981704e-01 6.23651981e-01 2.92200804e-01 -2.46658385e-01
-1.16846275e+00 -9.31377530e-01 1.06007242e+00 -5.96374393e-01
4.46008384e-01 1.18498921e-01 -5.25087714e-01 5.67303300e-01
1.27287582e-01 4.33798969e-01 9.20497537e-01 2.84497589e-01
-3.13480586e-01 -6.62799329e-02 -1.14929080e+00 2.46126711e-01
5.37135005e-01 -7.99520493e-01 -2.23652527e-01 7.71658301e-01
2.66590238e-01 -9.12154078e-01 -1.29208028e+00 4.30839837e-01
5.54631770e-01 -1.14529920e+00 8.10509562e-01 -4.14344996e-01
8.94231424e-02 -7.71436751e-01 3.24750125e-01 -7.37737417e-01
-1.19852796e-01 -8.84244919e-01 7.97926486e-02 9.72763002e-01
2.73678869e-01 -8.82831037e-01 6.47855103e-01 3.20621997e-01
1.09634653e-01 -6.65597558e-01 -7.45791972e-01 -5.91646194e-01
-6.33575082e-01 1.13867745e-02 5.85350156e-01 1.02199876e+00
-3.73139888e-01 -1.83904648e-01 -3.35377961e-01 6.22851610e-01
5.78310251e-01 3.83416206e-01 1.01190770e+00 -1.40472782e+00
-1.87993422e-01 -3.60245585e-01 -9.22061324e-01 -6.19552791e-01
-4.19428378e-01 -6.41207516e-01 -2.65299171e-01 -9.19560790e-01
1.20549738e-01 -3.70318204e-01 -3.01908553e-01 9.07458007e-01
6.79798871e-02 9.48393464e-01 -1.00654870e-01 5.62319160e-01
-1.84368610e-01 -2.49736801e-01 1.27569973e+00 1.40496355e-03
-4.41191792e-01 2.56503820e-01 -4.03707355e-01 9.15873647e-01
6.71062410e-01 -3.49339917e-02 3.34140658e-01 7.33726919e-02
1.23300545e-01 -6.77247941e-02 5.29731452e-01 -8.34609151e-01
3.11713189e-01 1.23711146e-01 5.17296314e-01 -5.06244183e-01
8.34194571e-02 -7.43749082e-01 4.87733865e-03 4.59143132e-01
-3.10856421e-02 -5.94968498e-01 3.02851826e-01 2.07838431e-01
-5.48435032e-01 9.47155058e-02 1.23624051e+00 3.49045098e-02
-5.55228055e-01 3.62071067e-01 8.58279690e-02 -4.68838573e-01
1.11662734e+00 -4.77728188e-01 -4.03636068e-01 7.89390653e-02
-1.09333372e+00 -4.67487127e-02 6.50273204e-01 2.22093254e-01
3.04559767e-01 -7.10962355e-01 -7.92476952e-01 8.21512341e-01
5.95574021e-01 -1.36582702e-01 2.21420169e-01 1.42956710e+00
-8.73817325e-01 6.78801596e-01 -2.15449437e-01 -9.38964725e-01
-1.96225691e+00 4.60409224e-01 9.06981111e-01 6.88914806e-02
-5.21079004e-01 9.35272336e-01 -9.48104858e-02 -3.24973404e-01
5.03859758e-01 -6.10935509e-01 -5.68353832e-01 1.07515343e-01
6.86604202e-01 3.66467834e-01 2.19453603e-01 -8.09070230e-01
-1.93841353e-01 1.08285165e+00 4.16567996e-02 7.93350279e-01
1.02013385e+00 6.48301840e-02 -6.48971379e-01 -2.27520868e-01
8.95222843e-01 2.79879838e-01 -5.71141720e-01 -5.60330689e-01
-3.09369564e-01 -6.91682696e-01 1.95068091e-01 -8.58286619e-01
-9.89001632e-01 7.10504949e-01 1.04261875e+00 2.43538842e-01
1.48944294e+00 6.47429302e-02 2.53421366e-01 -1.03519121e-02
6.06835447e-03 -1.12236869e+00 -7.23949552e-01 2.00315222e-01
6.34279072e-01 -1.51331973e+00 3.58683914e-02 -6.29617691e-01
-3.29024911e-01 1.30075860e+00 5.28850675e-01 1.00623474e-01
5.02204776e-01 -2.78158069e-01 3.96409959e-01 -4.45510745e-01
-2.04255618e-02 -8.09124112e-01 1.12827504e+00 5.18693507e-01
6.75449967e-01 2.58289754e-01 -1.06864825e-01 -6.15857132e-02
-1.91845357e-01 1.70440674e-01 5.02236068e-01 2.42930815e-01
-1.68991238e-01 -8.52207780e-01 -5.15246093e-01 5.29830754e-01
-6.50235891e-01 -2.06785545e-01 -5.84617853e-01 6.81258082e-01
5.36911309e-01 9.29110289e-01 1.16450496e-01 -2.28253752e-01
1.53716862e-01 -2.17239782e-01 3.37470114e-01 -6.73249662e-01
-8.98701847e-01 4.03071523e-01 3.12373284e-02 -9.19225037e-01
-7.35744655e-01 -1.64425999e-01 -5.92250466e-01 -2.13728860e-01
-6.05588317e-01 4.91407737e-02 6.39599562e-01 6.44598663e-01
4.50569212e-01 2.08848044e-01 5.84885597e-01 -6.45669520e-01
-3.20008963e-01 -1.01975024e+00 -1.06684399e+00 2.13759333e-01
7.77514577e-01 -2.63042241e-01 -5.10142028e-01 1.17687248e-01] | [3.743572235107422, -3.631558895111084] |
1906437b-de2a-4436-9edc-6f9a4fb6faf4 | skinnet-a-deep-learning-framework-for-skin | 1806.09522 | null | http://arxiv.org/abs/1806.09522v1 | http://arxiv.org/pdf/1806.09522v1.pdf | SkinNet: A Deep Learning Framework for Skin Lesion Segmentation | There has been a steady increase in the incidence of skin cancer worldwide,
with a high rate of mortality. Early detection and segmentation of skin lesions
are crucial for timely diagnosis and treatment, necessary to improve the
survival rate of patients. However, skin lesion segmentation is a challenging
task due to the low contrast of lesions and their high similarity in terms of
appearance, to healthy tissue. This underlines the need for an accurate and
automatic approach for skin lesion segmentation. To tackle this issue, we
propose a convolutional neural network (CNN) called SkinNet. The proposed CNN
is a modified version of U-Net. We compared the performance of our approach
with other state-of-the-art techniques, using the ISBI 2017 challenge dataset.
Our approach outperformed the others in terms of the Dice coefficient, Jaccard
index and sensitivity, evaluated on the held-out challenge test data set,
across 5-fold cross validation experiments. SkinNet achieved an average value
of 85.10, 76.67 and 93.0%, for the DC, JI, and SE, respectively. | ['Sulaiman Vesal', 'Nishant Ravikumar', 'Andreas Maier'] | 2018-06-25 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 3.02753478e-01 -2.51480252e-01 -2.88602799e-01 8.11423436e-02
-6.73008442e-01 -4.83919322e-01 5.42303860e-01 4.95555013e-01
-6.73566878e-01 8.23063076e-01 3.56711377e-03 2.79462188e-02
-1.94134101e-01 -7.04133332e-01 -8.37151557e-02 -9.68391359e-01
6.34065345e-02 4.24599051e-02 3.89209211e-01 8.44804421e-02
1.26141146e-01 4.98312891e-01 -1.12777233e+00 9.83893946e-02
1.18995738e+00 1.22397685e+00 -5.10231331e-02 5.49825013e-01
-1.61538467e-01 5.73817551e-01 -3.36037874e-01 -6.11515701e-01
8.68148506e-02 -4.83226597e-01 -8.85108173e-01 1.67462658e-02
2.95807242e-01 -2.89493836e-02 -3.03827003e-02 9.90665317e-01
5.90189517e-01 -7.85612240e-02 9.36084867e-01 -7.78075099e-01
-3.70772690e-01 6.27875179e-02 -7.60101378e-01 2.43271235e-02
-6.90609887e-02 -6.17044345e-02 7.39531934e-01 -4.09941614e-01
7.89829493e-01 4.98675168e-01 8.96051526e-01 6.93224430e-01
-8.39624226e-01 -3.73422086e-01 -3.77171636e-01 2.80791849e-01
-1.38536084e+00 -2.89342888e-02 2.14176968e-01 -4.59772587e-01
4.56485003e-01 4.59007412e-01 5.62897086e-01 9.78954494e-01
1.23084657e-01 6.68733537e-01 1.36857677e+00 -3.98578465e-01
3.53423238e-01 -3.03450208e-02 -9.69032049e-02 5.86132884e-01
3.26192915e-01 -4.21615332e-01 1.22344457e-02 3.82805467e-02
5.86349845e-01 1.31659331e-02 -2.84765393e-01 -6.54937550e-02
-8.35525453e-01 6.16397500e-01 6.91877663e-01 6.26134813e-01
-5.95849454e-01 -1.70037985e-01 5.97400069e-01 -2.29724959e-01
6.75188839e-01 1.87660620e-01 -1.98408395e-01 1.51328463e-02
-9.09977794e-01 -1.81131244e-01 6.15503550e-01 -2.55260766e-02
-5.19045033e-02 -4.55206215e-01 -4.25468296e-01 9.23851371e-01
-2.06952058e-02 1.88428417e-01 5.31999350e-01 -4.36545700e-01
6.08481318e-02 7.79928088e-01 -2.66774982e-01 -8.95352721e-01
-5.70368588e-01 -7.80842483e-01 -1.45385349e+00 8.50092620e-02
5.93684256e-01 -1.27563760e-01 -1.18961632e+00 1.47533751e+00
4.02986676e-01 2.76743442e-01 -9.34252422e-03 9.64879930e-01
8.29408705e-01 2.58179784e-01 3.62770289e-01 -1.53607786e-01
1.28628993e+00 -8.82430315e-01 -6.69110477e-01 1.39762759e-01
4.67462897e-01 -6.76203787e-01 5.96433580e-01 4.38758165e-01
-7.61290550e-01 -1.38829306e-01 -8.79256248e-01 2.46551976e-01
-3.13492090e-01 4.93773937e-01 4.54236150e-01 6.45345151e-01
-1.05610335e+00 5.17321229e-01 -7.33914733e-01 -1.10965133e+00
8.89050484e-01 1.38329744e-01 -6.68647766e-01 -2.14097440e-01
-9.45024073e-01 9.03444648e-01 4.65251923e-01 4.03076373e-02
-4.53447104e-01 -6.66388988e-01 -3.63374203e-01 -2.03437835e-01
2.41481885e-01 -5.49492002e-01 9.97744620e-01 -8.90113831e-01
-1.15161824e+00 9.28905785e-01 1.67657062e-01 -5.40596008e-01
8.33431721e-01 -7.44714588e-02 -2.53447205e-01 3.73977810e-01
-1.59428224e-01 5.04916966e-01 3.27132910e-01 -1.02598834e+00
-7.59759605e-01 -4.03311491e-01 -2.58858442e-01 1.37362525e-01
-4.79234606e-01 -4.86205478e-04 -5.97516060e-01 -3.12229723e-01
-1.92901120e-01 -8.55997622e-01 -4.56480980e-01 2.92982340e-01
-6.15808308e-01 -1.72588050e-01 6.35678053e-01 -1.20293319e+00
1.23727906e+00 -2.08333254e+00 -1.52820647e-02 4.02128875e-01
3.42682362e-01 8.58504951e-01 4.81471457e-02 3.64875466e-01
-1.74321551e-02 4.70093250e-01 -5.90431333e-01 -7.40785748e-02
-3.64095777e-01 -1.98736295e-01 5.84649384e-01 5.52481830e-01
3.17510486e-01 6.61073744e-01 -8.65117192e-01 -5.81156671e-01
1.94468334e-01 7.64681101e-01 8.16159397e-02 8.49242657e-02
-8.00044388e-02 2.98431337e-01 -2.09241256e-01 6.99479461e-01
6.90337181e-01 -3.53732526e-01 2.29447186e-01 -2.50510454e-01
2.56973449e-02 -3.46109122e-01 -8.58226538e-01 1.46001983e+00
-2.23917291e-01 6.52837515e-01 6.82800170e-03 -7.57125139e-01
7.06926823e-01 4.91369814e-01 7.16118097e-01 -8.36705208e-01
2.45193899e-01 3.39469671e-01 2.61144996e-01 -6.57186270e-01
-3.06272600e-02 -9.55545995e-03 4.46689576e-01 9.61471424e-02
-3.12662035e-01 2.19579995e-01 5.61392844e-01 1.75526679e-01
1.12977469e+00 -2.97558993e-01 6.64766967e-01 -1.06104158e-01
6.46724522e-01 -4.84420955e-02 3.69579494e-01 2.20833153e-01
-5.01749575e-01 7.74706066e-01 6.59750819e-01 -2.33057842e-01
-7.62406170e-01 -9.27864075e-01 -4.28406119e-01 2.44232669e-01
-3.28183398e-02 -1.06141277e-01 -1.19373953e+00 -9.38396513e-01
-1.18640043e-01 2.83029765e-01 -9.13337827e-01 6.95212111e-02
-1.51821807e-01 -1.02867591e+00 5.67067981e-01 4.14278567e-01
8.70267510e-01 -9.42188084e-01 -4.63915586e-01 4.08351012e-02
-2.39380673e-01 -1.15315771e+00 -1.63773179e-01 -1.39541179e-01
-6.69308841e-01 -1.50087690e+00 -1.25770617e+00 -7.37828672e-01
8.77691567e-01 4.91088741e-02 7.99043059e-01 3.20643902e-01
-9.40613568e-01 -1.50319040e-01 -4.27843422e-01 -2.18012229e-01
-2.48035401e-01 3.15229923e-01 -3.96323293e-01 2.96013296e-01
1.50009617e-01 -1.50678620e-01 -9.16821659e-01 2.29618296e-01
-1.27315485e+00 -1.29235554e-02 1.03468382e+00 7.71265566e-01
7.85084128e-01 2.15095654e-01 5.17964005e-01 -8.82235527e-01
6.97978020e-01 -5.71426690e-01 -2.22706124e-01 3.87025476e-01
-4.54965085e-01 -5.18244267e-01 5.28269708e-01 -5.67638576e-02
-9.02321398e-01 1.86933711e-01 -2.82418251e-01 1.94800064e-01
-3.38936836e-01 6.06401980e-01 2.27275044e-01 -1.66380420e-01
7.45106816e-01 -1.12802602e-01 2.13662103e-01 -3.93122762e-01
-2.57661595e-04 8.01449001e-01 3.99804920e-01 7.13689700e-02
4.86536115e-01 4.73588586e-01 3.54869455e-01 -9.13201272e-01
-8.94538224e-01 -6.50765240e-01 -6.85238361e-01 -2.26395667e-01
1.29253340e+00 -5.09601772e-01 -3.07803869e-01 9.32858109e-01
-1.01049614e+00 -2.37168700e-01 1.11987628e-01 1.64466381e-01
5.36771864e-03 4.96282488e-01 -4.62808937e-01 -6.48250937e-01
-7.74844944e-01 -9.14980292e-01 6.24005258e-01 6.65025771e-01
-2.02080190e-01 -1.24225688e+00 1.81109756e-01 4.43009615e-01
6.36766672e-01 8.93841147e-01 8.03400278e-01 -6.93989336e-01
9.86808091e-02 -6.64030612e-01 -7.44450510e-01 6.09066725e-01
4.01394933e-01 4.49328303e-01 -8.62930834e-01 -2.33669668e-01
-7.06494391e-01 -2.49949574e-01 1.01067126e+00 5.23687422e-01
1.39788425e+00 -4.52099331e-02 -3.79225850e-01 4.42054272e-01
1.79012990e+00 1.21067211e-01 7.87171304e-01 1.55966133e-01
4.13771540e-01 6.22522116e-01 4.56066489e-01 2.74034739e-01
1.47277445e-01 4.28843260e-01 6.90826952e-01 -6.14616871e-01
-3.94036502e-01 1.28018528e-01 -2.51748502e-01 4.94969219e-01
-3.02109808e-01 -3.12748909e-01 -1.09251416e+00 7.89521098e-01
-1.58899593e+00 -7.96716750e-01 -4.42766815e-01 2.22417593e+00
6.88216209e-01 1.63510479e-02 3.94260138e-01 2.61193603e-01
8.99316192e-01 -1.65617466e-01 -4.22425538e-01 -3.98449808e-01
-7.80604333e-02 3.73016089e-01 5.99261105e-01 7.54432157e-02
-1.36440480e+00 6.63363934e-01 5.61126900e+00 9.95705307e-01
-1.43486321e+00 9.42755267e-02 1.07843959e+00 2.97624409e-01
2.89638847e-01 -5.77439606e-01 -3.97390097e-01 5.29093564e-01
6.54478312e-01 6.74932674e-02 8.06548446e-02 3.66853058e-01
8.22522417e-02 -4.36174095e-01 -5.66449940e-01 6.35664523e-01
2.45916680e-01 -1.26502860e+00 -4.32898849e-01 8.19454119e-02
9.68016624e-01 -3.06344908e-02 3.27804983e-02 -2.47487053e-01
-9.45401564e-02 -1.33578432e+00 -8.40491131e-02 6.17028534e-01
1.00905871e+00 -7.34165549e-01 1.39377165e+00 1.81793213e-01
-9.81697142e-01 1.82058811e-01 2.85448525e-02 4.75469321e-01
-9.48691592e-02 8.59546781e-01 -9.52835500e-01 7.22795427e-01
6.23623073e-01 5.18455029e-01 -7.77968049e-01 1.74916816e+00
-1.23830155e-01 7.42522001e-01 -2.74082124e-01 -2.46233419e-01
3.14277261e-01 -8.59197751e-02 8.84290934e-02 1.24676275e+00
2.45942771e-01 -1.78309932e-01 -1.33516535e-01 4.83539253e-01
-7.73980469e-02 5.57230353e-01 -1.59669653e-01 -1.73799947e-01
1.25787944e-01 1.73817503e+00 -1.01975596e+00 -4.59827222e-02
-1.99350044e-01 1.01414669e+00 5.28074466e-02 3.49769667e-02
-7.41286576e-01 -6.38921976e-01 3.77572656e-01 1.13746142e-02
5.64486720e-03 2.03349441e-01 -2.69568115e-01 -6.81240559e-01
1.31075874e-01 -7.70411372e-01 5.69822192e-01 -2.95652330e-01
-1.48698688e+00 6.39515877e-01 -3.89233708e-01 -1.07806420e+00
1.57862887e-01 -6.82831764e-01 -9.36154544e-01 8.80067885e-01
-1.78828895e+00 -1.18259633e+00 -8.68315935e-01 3.29180241e-01
2.23120421e-01 9.25123245e-02 8.89469504e-01 2.33161598e-01
-8.78885806e-01 6.16981745e-01 2.05827922e-01 3.24385464e-01
7.60013700e-01 -1.33081126e+00 1.41628325e-01 6.55251861e-01
-4.54733819e-01 1.09901324e-01 2.48752892e-01 -5.51042795e-01
-8.21549177e-01 -1.21697259e+00 6.82405114e-01 3.64209227e-02
6.12168491e-01 1.67171165e-01 -7.67994285e-01 -1.18345670e-01
2.22932518e-01 1.50618210e-01 1.05264759e+00 -1.38529763e-01
-1.52570128e-01 -1.85900852e-01 -1.61214614e+00 5.52419186e-01
5.58924973e-01 -1.29515126e-01 1.52371496e-01 5.02492249e-01
2.36553133e-01 -2.03442320e-01 -1.10872126e+00 4.88221437e-01
7.28880584e-01 -1.05469894e+00 8.50629866e-01 -5.04750431e-01
7.60842443e-01 -3.59042063e-02 4.21943255e-02 -1.39258397e+00
-4.18121904e-01 -1.22745931e-02 1.79024518e-01 1.19803715e+00
4.91994888e-01 -4.48142260e-01 1.02268815e+00 3.44533682e-01
7.24276081e-02 -1.32491815e+00 -8.02455664e-01 -5.30992270e-01
2.24169239e-01 5.97554035e-02 1.85233131e-01 8.51289093e-01
-3.50288659e-01 -1.01150848e-01 -7.50366822e-02 -1.93717569e-01
7.01468647e-01 -4.35854375e-01 3.64555985e-01 -1.31979179e+00
9.30750445e-02 -7.28673398e-01 -5.67151010e-01 1.45141520e-02
-3.14425409e-01 -8.69374990e-01 -2.16280550e-01 -2.17669344e+00
5.21671593e-01 -2.32830957e-01 -5.57739496e-01 5.72975278e-01
-4.29528266e-01 7.53748119e-01 1.38843194e-01 1.39012663e-02
-4.85588521e-01 2.83047371e-03 1.37500179e+00 -1.58166170e-01
2.31813509e-02 -6.88099712e-02 -5.66478491e-01 5.95257103e-01
1.25130832e+00 -4.78945598e-02 1.71936862e-02 -1.34590358e-01
-1.90072809e-03 -1.93237722e-01 2.93179214e-01 -1.29609573e+00
1.84254080e-01 -2.61461198e-01 5.40271282e-01 -4.94865566e-01
1.64164975e-01 -7.46351898e-01 3.28494646e-02 9.14672911e-01
-3.64392757e-01 -2.69526571e-01 1.38387203e-01 3.15027326e-01
-1.63564861e-01 -2.22142667e-01 1.15324402e+00 2.10808478e-02
-6.69785619e-01 4.75518644e-01 -1.92566797e-01 -1.96961179e-01
1.57750845e+00 -1.42221123e-01 -4.94616866e-01 -1.00730054e-01
-4.56315875e-01 9.43496749e-02 2.65248030e-01 1.83334589e-01
4.17241991e-01 -1.07988083e+00 -9.23241615e-01 -1.65354848e-01
2.94386864e-01 -6.99923001e-03 5.00596941e-01 1.35872126e+00
-9.56398010e-01 2.69976497e-01 -3.88269514e-01 -4.81905222e-01
-1.47411656e+00 -2.22352054e-02 4.98104841e-01 -5.56377351e-01
-3.00231516e-01 8.88765693e-01 -2.05139577e-01 -2.19861418e-01
3.00103903e-01 1.05429832e-02 -4.84166771e-01 1.05067685e-01
5.60526967e-01 5.82083821e-01 2.80275196e-01 -4.71543550e-01
-3.29593062e-01 6.39954031e-01 -2.19633788e-01 3.65415514e-01
1.18064487e+00 3.09683830e-01 -4.52215701e-01 3.40367248e-03
1.20311797e+00 -1.66480392e-01 -8.51842880e-01 -1.00042466e-02
1.17888093e-01 -4.01352376e-01 2.26942390e-01 -1.47990179e+00
-1.23494506e+00 8.68231058e-01 1.04672813e+00 2.99867451e-01
1.26678848e+00 -2.35608131e-01 1.00954962e+00 -6.02291478e-03
-2.25729365e-02 -1.09447706e+00 -6.99363044e-03 7.44660050e-02
6.05135977e-01 -1.41543639e+00 5.26290946e-02 -6.01097584e-01
-5.98480105e-01 1.06018555e+00 6.38230622e-01 -2.48659495e-02
4.33918536e-01 -2.98404656e-02 3.35910380e-01 7.10936487e-02
-4.81581300e-01 -3.55193317e-01 4.96838212e-01 5.89724541e-01
6.26480818e-01 2.60761142e-01 -8.15177798e-01 3.61237645e-01
3.63647193e-01 9.42999795e-02 2.00399578e-01 6.30661845e-01
-3.94777477e-01 -1.02848160e+00 -2.36378729e-01 7.46099055e-01
-9.42440689e-01 1.40934855e-01 -9.29543793e-01 9.78324592e-01
2.05284521e-01 8.14130783e-01 -5.30709662e-02 -2.45056480e-01
-5.13865752e-03 -1.21759556e-01 2.99306452e-01 -2.29681835e-01
-6.85579121e-01 -1.66898951e-01 1.97261587e-01 -2.57420838e-01
-4.52445447e-01 -4.57676470e-01 -1.06551814e+00 -1.13828450e-01
-3.04879367e-01 4.77670468e-02 1.31627798e+00 8.45993996e-01
1.75863370e-01 6.24738634e-01 6.52299106e-01 -1.76587030e-01
-5.30859590e-01 -9.14743662e-01 -5.20438612e-01 6.17500782e-01
3.31599824e-03 -3.45284134e-01 -1.02789171e-01 -1.88099891e-01] | [15.630043983459473, -2.9466097354888916] |
e385fdca-417e-43b1-9fa2-71595cc5ab4b | unifying-bayesian-inference-and-vector-space | null | null | https://aclanthology.org/P15-1081 | https://aclanthology.org/P15-1081.pdf | Unifying Bayesian Inference and Vector Space Models for Improved Decipherment | null | ['Qing Dou', 'Chris Dyer', 'Ashish Vaswani', 'Kevin Knight'] | 2015-07-01 | unifying-bayesian-inference-and-vector-space-1 | https://aclanthology.org/P15-1081 | https://aclanthology.org/P15-1081.pdf | ijcnlp-2015-7 | ['decipherment'] | ['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] | [-7.327815532684326, 3.644780158996582] |
2df71a64-08b5-4190-a280-aded6c857936 | sars-cov-2-s-closest-relative-ratg13-was | 2111.09469 | null | https://arxiv.org/abs/2111.09469v2 | https://arxiv.org/pdf/2111.09469v2.pdf | SARS-CoV-2's closest relative, RaTG13, was generated from a bat transcriptome not a fecal swab: implications for the origin of COVID-19 | RaTG13 is the closest related coronavirus genome phylogenetically to SARS-CoV-2, consequently understanding its provenance is of key importance to understanding the origin of the COVID-19 pandemic. The RaTG13 NGS dataset is attributed to a fecal swab from the intermediate horseshoe bat Rhinolophus affinis. However, sequence analysis reveals that this is unlikely. Metagenomic analysis using Metaxa2 shows that only 10.3 % of small subunit (SSU) rRNA sequences in the dataset are bacterial, inconsistent with a fecal sample, which are typically dominated by bacterial sequences. In addition, the bacterial taxa present in the sample are inconsistent with fecal material. Assembly of mitochondrial SSU rRNA sequences in the dataset produces a contig 98.7 % identical to R.affinis mitochondrial SSU rRNA, indicating that the sample was generated from this or a closely related species. 87.5 % of the NGS reads map to the Rhinolophus ferrumequinum genome, the closest bat genome to R.affinis available. In the annotated genome assembly, 62.2 % of mapped reads map to protein coding genes. These results clearly demonstrate that the dataset represents a Rhinolophus sp. transcriptome, and not a fecal swab sample. Overall, the data show that the RaTG13 dataset was generated by the Wuhan Institute of Virology (WIV) from a transcriptome derived from Rhinolophus sp. tissue or cell line, indicating that RaTG13 was in live culture. This raises the question of whether the WIV was culturing additional unreported coronaviruses closely related to SARS-CoV-2 prior to the pandemic. The implications for the origin of the COVID-19 pandemic are discussed. | ['Steven E Massey'] | 2021-11-18 | null | null | null | null | ['virology'] | ['miscellaneous'] | [ 7.33312309e-01 -4.26030725e-01 3.70296091e-02 7.12875463e-03
-3.71159226e-01 -1.03039563e+00 3.71573597e-01 2.91549712e-01
-3.01894277e-01 8.42054963e-01 3.58163387e-01 -3.74094397e-01
2.73268223e-01 -6.42746449e-01 -8.49068999e-01 -1.06265461e+00
-1.87753752e-01 8.86990666e-01 -5.38030148e-01 -1.56383649e-01
-1.41931206e-01 4.70057726e-01 -1.24902606e+00 2.23797843e-01
9.96652901e-01 6.31362498e-02 7.73019195e-01 8.51624727e-01
-1.47554994e-01 -1.21232934e-01 -4.85137284e-01 2.62193292e-01
5.93221426e-01 -5.33527255e-01 -2.50545621e-01 -4.64806378e-01
3.54429126e-01 -6.95542574e-01 5.51526070e-01 8.06858659e-01
1.06763653e-01 -5.82684875e-01 6.99819028e-01 -1.13767469e+00
8.14609453e-02 1.08031355e-01 -4.02525306e-01 2.48747557e-01
5.45622408e-02 3.02049458e-01 7.58462667e-01 -3.72847050e-01
1.49402940e+00 1.15651393e+00 8.56339216e-01 3.17751944e-01
-1.35582972e+00 -5.08226931e-01 -5.75840473e-01 -5.34082390e-03
-1.37006032e+00 -7.25602284e-02 -2.69207537e-01 -1.24605155e+00
1.11631382e+00 4.11403328e-01 1.43027735e+00 8.13334644e-01
7.94713259e-01 1.58852592e-01 1.21959186e+00 3.99501085e-01
3.56876403e-01 -5.72953105e-01 1.59442410e-01 3.53389323e-01
1.44275331e+00 2.91635156e-01 2.46770442e-01 -9.55598295e-01
1.26139238e-01 5.11808097e-01 -6.29232883e-01 -8.46170858e-02
-9.44925785e-01 1.17187512e+00 1.40385851e-01 -7.69860968e-02
-7.66795516e-01 4.44128923e-02 5.56921899e-01 1.64323226e-02
-1.01837397e-01 -3.25870723e-03 -7.56736517e-01 -9.98995081e-02
-4.69895601e-01 -9.69790667e-02 7.88842857e-01 5.51024795e-01
5.69116771e-01 -8.39680135e-02 6.21106148e-01 3.89932245e-01
5.71539998e-01 1.26231861e+00 -2.01513588e-01 -5.70871055e-01
-1.61595017e-01 4.82853353e-01 4.48752791e-01 -6.00634456e-01
-4.61121887e-01 -8.94694030e-02 -3.32551807e-01 -7.63476919e-03
4.63690460e-01 -1.25435576e-01 -8.66169930e-01 1.63997650e+00
5.59776723e-01 1.15796350e-01 2.15020820e-01 1.00936210e+00
4.18982118e-01 9.94229257e-01 -7.23199844e-02 -5.28596938e-01
2.09909081e+00 -2.28677645e-01 -3.27317566e-01 -2.73058355e-01
2.26473019e-01 -7.93082893e-01 3.85579586e-01 8.57343152e-02
-8.40906501e-02 -2.31979117e-02 -1.23500097e+00 5.36621094e-01
1.72546059e-01 -3.75494778e-01 -3.75112221e-02 1.02806723e+00
-5.39059222e-01 1.88898936e-01 -1.00599766e+00 -1.24150062e+00
9.79258195e-02 -1.74581423e-01 -1.73162207e-01 -3.58529866e-01
-5.27652264e-01 6.82158172e-01 3.98967505e-01 3.13231140e-01
-9.81582284e-01 -3.75828534e-01 -2.16440365e-01 -1.54945999e-01
-1.66567788e-01 -1.08679307e+00 7.70423472e-01 -3.44180197e-01
-8.85203063e-01 8.34799886e-01 -4.66584295e-01 -4.35266525e-01
1.15865111e-01 -2.37857237e-01 -4.01366204e-01 5.21837234e-01
2.22585171e-01 2.11132228e-01 2.87715882e-01 -1.17776883e+00
-6.69820130e-01 -7.39991546e-01 -4.34238762e-01 -5.43446243e-01
9.88459885e-01 2.61552513e-01 7.44848073e-01 -6.21513605e-01
-1.13476701e-01 -1.48398030e+00 -2.53674835e-01 -3.26371402e-01
1.05311595e-01 4.13248837e-02 4.85886991e-01 -6.86141789e-01
3.29807103e-01 -2.03617239e+00 -5.36066353e-01 1.68474883e-01
-6.74034506e-02 4.00564253e-01 -6.60570860e-01 1.07543123e+00
2.41383195e-01 5.67514479e-01 -3.52384329e-01 8.48394215e-01
-2.93104261e-01 4.60300416e-01 -4.35166657e-01 1.27089179e+00
2.75780750e-03 7.52660036e-01 -1.43537629e+00 2.02231497e-01
-3.11193258e-01 8.27814460e-01 -4.36155856e-01 1.23914883e-01
-4.51546967e-01 5.36666036e-01 -3.06364357e-01 6.30808055e-01
1.42176497e+00 6.57969043e-02 7.72691727e-01 -2.47250870e-01
-4.12555158e-01 3.32421005e-01 -3.47328961e-01 1.36763895e+00
3.90480340e-01 3.66374165e-01 5.34815907e-01 -5.06878905e-02
5.14782488e-01 3.32162768e-01 2.41399780e-01 -1.92707896e-01
3.22814554e-01 3.11543703e-01 3.82943720e-01 -6.38352394e-01
2.86977202e-01 -7.04814851e-01 3.70975852e-01 4.00672048e-01
-4.13259566e-01 2.09914565e-01 3.15593123e-01 8.20209607e-02
9.77300465e-01 2.22454324e-01 7.01954901e-01 -6.27429247e-01
1.51707930e-03 7.37106681e-01 1.12381887e+00 4.71959144e-01
-3.03912699e-01 4.58071679e-01 9.34723988e-02 -2.12888390e-01
-1.22107232e+00 -1.32989693e+00 -6.44852459e-01 4.44917113e-01
-1.16327479e-01 -5.07418275e-01 -5.50938666e-01 7.84596354e-02
6.95602745e-02 7.22251952e-01 -3.81455898e-01 3.99931729e-01
-4.35884029e-01 -8.99478316e-01 7.90963888e-01 5.97547833e-03
-5.64102866e-02 -3.96746874e-01 -8.94269347e-01 3.58348250e-01
-7.82970250e-01 -8.41534138e-01 -4.72012967e-01 -5.19864038e-02
-9.66481507e-01 -1.60050356e+00 4.10940647e-02 -4.55862105e-01
4.49118316e-01 4.35632646e-01 4.88517880e-01 1.85923859e-01
-4.54811484e-01 2.23502934e-01 -4.18622047e-01 -4.41399127e-01
-7.16624022e-01 -5.37028193e-01 1.56365588e-01 -7.67299354e-01
7.81555891e-01 -2.38225237e-01 -5.96594095e-01 2.46877789e-01
-5.11446357e-01 -3.03907275e-01 1.77157134e-01 8.36910427e-01
6.29514992e-01 -5.53123593e-01 5.78958869e-01 -5.12787998e-01
-8.58462695e-03 -1.02039242e+00 -6.28160357e-01 4.07187417e-02
-2.04515159e-01 -4.06027496e-01 6.04896069e-01 1.43528268e-01
-8.83661628e-01 -3.53940934e-01 -4.24788445e-01 3.52222323e-01
-3.82693887e-01 4.46446776e-01 1.80321038e-01 8.30417454e-01
5.54574311e-01 7.17616454e-02 5.90560079e-01 -1.84195951e-01
-3.71483117e-02 9.41767573e-01 3.01724613e-01 -1.11690953e-01
3.78844172e-01 1.04514837e+00 -1.56802222e-01 -1.48492122e+00
-3.66585255e-01 -8.95827293e-01 -1.26324952e-01 1.08413957e-01
1.52208710e+00 -1.01537704e+00 -1.15331328e+00 2.77852684e-01
-1.08906078e+00 -3.74071121e-01 1.36709660e-01 8.26346278e-01
-3.28482896e-01 2.60858446e-01 -1.01187837e+00 -7.71905959e-01
-8.17298472e-01 -1.00736761e+00 7.76846230e-01 -2.84364253e-01
-5.71510375e-01 -3.38517606e-01 8.79839361e-01 7.18825340e-01
2.69783080e-01 8.02099526e-01 9.44706798e-01 -7.76552558e-01
-6.47686958e-01 1.74698874e-01 -1.03798145e-02 -9.55492556e-02
4.46107954e-01 3.04217011e-01 -6.02817178e-01 -6.65842772e-01
5.18273592e-01 3.89738418e-02 5.90958953e-01 2.01256827e-01
-5.22463083e-01 -5.65385997e-01 -5.17609656e-01 5.19343138e-01
1.73783851e+00 6.33085847e-01 5.30717313e-01 3.80309016e-01
2.83789605e-01 4.17230695e-01 7.90342510e-01 3.71248841e-01
5.65274134e-02 3.01464468e-01 6.04740024e-01 6.56890094e-01
1.69193909e-01 -2.28741206e-02 7.35805273e-01 7.16827154e-01
-2.01894775e-01 -4.65003669e-01 -9.66461897e-01 2.51879513e-01
-1.40965104e+00 -1.33840334e+00 -7.18129635e-01 1.96436477e+00
2.97279239e-01 -8.66895795e-01 4.71266024e-02 -7.96181142e-01
6.83624327e-01 -3.32282186e-01 -4.74619538e-01 -6.62336111e-01
-2.00791270e-01 6.74711317e-02 5.24158239e-01 4.01268870e-01
-4.73810226e-01 -5.05831267e-05 5.87095928e+00 6.71207556e-04
-9.56485271e-01 -2.74162944e-02 -2.62494415e-01 -1.65025577e-01
-7.05794513e-01 3.02433729e-01 -5.84151745e-01 4.90272045e-01
1.11274350e+00 -3.22282583e-01 6.33231103e-01 3.46695900e-01
4.17140156e-01 -3.51850651e-02 -6.65386200e-01 4.34231341e-01
2.42080301e-01 -1.22259343e+00 -2.92062789e-01 5.00771999e-01
6.02581024e-01 1.19255161e+00 -9.77990985e-01 -3.65434945e-01
1.84764996e-01 -8.16637278e-01 6.14230752e-01 3.86333585e-01
6.26746595e-01 -8.77403677e-01 1.18981624e+00 5.65845788e-01
-1.21284568e+00 3.49768043e-01 -6.43514276e-01 6.20242618e-02
8.87342811e-01 5.67543387e-01 -1.71400750e+00 1.74595222e-01
5.70213616e-01 2.09839210e-01 9.94370878e-02 8.79847288e-01
-3.08707982e-01 9.97069538e-01 -5.97434580e-01 -1.00294381e-01
-1.42728254e-01 -8.06263208e-01 1.04698014e+00 1.36364996e+00
5.84340513e-01 4.14773941e-01 2.63534367e-01 5.85768700e-01
2.52143800e-01 1.02338949e-02 -5.40281713e-01 -7.20071733e-01
5.60675561e-01 1.02136111e+00 -8.08830321e-01 -6.38063490e-01
-1.81360543e-01 7.11778522e-01 -1.32222340e-01 2.93144405e-01
-4.50160563e-01 -2.40224842e-02 1.56243587e+00 2.84722537e-01
8.31761241e-01 -1.68779686e-01 3.39689583e-01 -8.32308650e-01
-5.40774047e-01 -8.33064198e-01 5.19616663e-01 -6.19326949e-01
-1.49431753e+00 5.44082820e-01 -2.03440130e-01 -1.23706424e+00
-3.64274204e-01 -3.13526809e-01 -3.12301636e-01 7.26695418e-01
-8.82084906e-01 -8.03546965e-01 -2.79270262e-01 -3.22645366e-01
-1.47693530e-01 3.76534283e-01 9.30865109e-01 -2.26238728e-01
-5.46697415e-02 -4.58655059e-01 8.89667988e-01 -2.73380697e-01
5.60150981e-01 -5.85325062e-01 5.01728475e-01 6.37123048e-01
-6.25531733e-01 1.46014369e+00 8.87101531e-01 -1.38526750e+00
-1.81040430e+00 -1.51454151e+00 8.26787472e-01 -4.22972254e-02
6.04776740e-01 -3.63634527e-01 -6.13570511e-01 1.10938311e+00
7.52270281e-01 -6.73205912e-01 1.48647714e+00 -6.22940481e-01
-4.41076934e-01 3.42032015e-01 -1.76086259e+00 4.15702164e-01
9.09685493e-01 -4.90678042e-01 -8.02949011e-01 2.65196592e-01
6.39757872e-01 1.19690225e-01 -8.46328497e-01 3.13155055e-01
1.10414636e+00 -5.76811552e-01 7.88640797e-01 -6.18337333e-01
-4.73357961e-02 -1.45509470e+00 -5.07097423e-01 -1.32093263e+00
-4.19526607e-01 -1.79027691e-01 7.76132882e-01 9.03115034e-01
-3.16652983e-01 -7.88566470e-01 -9.98315513e-02 -4.78908211e-01
-2.87150711e-01 4.48712781e-02 -7.77822733e-01 -1.07070446e+00
-1.41164884e-01 1.39593273e-01 6.81123853e-01 1.12495327e+00
-1.39869317e-01 6.48104548e-01 -1.66170716e-01 1.31588280e-01
8.04593384e-01 6.39054775e-01 8.83793056e-01 -1.31970942e+00
-1.64807253e-02 1.52996168e-01 -2.65693247e-01 -6.28422618e-01
-3.25174630e-01 -1.45026147e+00 4.39931393e-01 -1.96489930e+00
4.65604454e-01 -3.11701059e-01 3.61134589e-01 1.49304152e-01
4.82595414e-02 2.81713009e-01 3.10571194e-01 9.28837955e-02
4.92823660e-01 2.09369045e-02 7.90644944e-01 3.89919616e-02
-4.44939593e-03 -5.11728287e-01 -3.40897501e-01 7.73595095e-01
1.00292540e+00 -7.42255926e-01 -1.61925465e-01 -1.57470256e-01
5.66320240e-01 -2.96561737e-02 9.76566598e-02 -2.56212711e-01
-4.29919749e-01 -6.55967951e-01 2.97353007e-02 -1.06265652e+00
8.82942155e-02 -8.85619938e-01 1.41692531e+00 1.57261217e+00
7.86380649e-01 1.38189122e-01 2.62776673e-01 7.49331117e-01
5.27118087e-01 -2.93719918e-01 6.39225185e-01 -2.17728138e-01
-2.21870378e-01 -1.26665458e-01 -1.39902925e+00 7.83610791e-02
8.45883131e-01 -3.95131171e-01 -8.86363685e-01 4.19220775e-01
-1.70354709e-01 -9.08594951e-02 1.23957491e+00 8.44807252e-02
1.54764473e-01 -8.47957313e-01 -1.31948817e+00 2.08121419e-01
4.02314126e-01 -5.51434517e-01 2.51871616e-01 9.95944440e-01
-1.37472570e+00 4.61798161e-01 -5.64560950e-01 -1.18945968e+00
-1.32141149e+00 8.78849208e-01 -1.79012343e-01 5.65999806e-01
-5.03514230e-01 2.17991680e-01 1.19667970e-01 -5.64863861e-01
-6.61607206e-01 -3.20948571e-01 2.63303556e-02 2.92324364e-01
5.98783135e-01 4.71959770e-01 -1.43035069e-01 -9.81176019e-01
-7.48002648e-01 2.40452528e-01 4.70662534e-01 4.72767383e-01
1.41676664e+00 -3.49100307e-02 -8.32290411e-01 4.24580753e-01
1.36604106e+00 5.09353638e-01 -3.93209159e-01 5.18543661e-01
-3.25309485e-01 -5.70861101e-01 -8.59174132e-01 -7.49203503e-01
-4.55360115e-01 1.71708271e-01 8.59584391e-01 -5.03424883e-01
7.29476750e-01 5.46265393e-02 1.02425194e+00 2.48224422e-01
6.70345604e-01 -4.82261539e-01 -7.56951630e-01 4.90908951e-01
9.63465810e-01 -5.78129470e-01 -1.89186484e-01 -3.74068230e-01
-2.96701547e-02 9.42569435e-01 7.48914033e-02 -7.94726759e-02
1.06298298e-01 4.51243728e-01 2.58553982e-01 -2.68338948e-01
-9.47079420e-01 1.04316063e-01 -4.35095578e-01 6.03614867e-01
6.41767263e-01 7.55963147e-01 -1.09334850e+00 3.43371704e-02
-7.22662508e-01 2.05914989e-01 7.82768190e-01 1.08917689e+00
-7.46754706e-01 -9.43238854e-01 -6.63223803e-01 7.34114468e-01
-1.97958216e-01 -1.88825667e-01 -2.27432728e-01 4.54534382e-01
2.45954052e-01 1.00201964e+00 4.99162227e-01 5.91205433e-02
-1.16116747e-01 2.06487209e-01 5.23591876e-01 -3.16267580e-01
-7.35531211e-01 6.03177011e-01 4.68798459e-01 -2.16569841e-01
-4.79596227e-01 -1.04805410e+00 -1.83318281e+00 -5.75220644e-01
-3.64832699e-01 4.78036404e-01 7.67612755e-01 5.82505822e-01
2.29891628e-01 -1.55361325e-01 2.62922585e-01 -1.06370918e-01
-2.38633454e-01 -9.95213389e-01 -4.99160945e-01 2.88278341e-01
5.22755563e-01 6.66044652e-02 -4.23567921e-01 6.15066364e-02] | [4.858774662017822, 5.081958293914795] |
ad74bd19-f17d-489a-bbb4-02d8a083e689 | when-sam-meets-shadow-detection | 2305.11513 | null | https://arxiv.org/abs/2305.11513v1 | https://arxiv.org/pdf/2305.11513v1.pdf | When SAM Meets Shadow Detection | As a promptable generic object segmentation model, segment anything model (SAM) has recently attracted significant attention, and also demonstrates its powerful performance. Nevertheless, it still meets its Waterloo when encountering several tasks, e.g., medical image segmentation, camouflaged object detection, etc. In this report, we try SAM on an unexplored popular task: shadow detection. Specifically, four benchmarks were chosen and evaluated with widely used metrics. The experimental results show that the performance for shadow detection using SAM is not satisfactory, especially when comparing with the elaborate models. Code is available at https://github.com/LeipingJie/SAMSh. | ['HUI ZHANG', 'Leiping Jie'] | 2023-05-19 | null | null | null | null | ['shadow-detection'] | ['computer-vision'] | [ 3.19440722e-01 -1.14358000e-01 -2.90174961e-01 -2.18299165e-01
-5.68965256e-01 -2.75469691e-01 2.89374739e-01 -1.17531635e-01
-3.97732943e-01 7.15575993e-01 -2.73513824e-01 -3.65406334e-01
3.82534236e-01 -1.59892425e-01 -2.09170848e-01 -9.53614950e-01
2.36695305e-01 1.36588678e-01 9.30096507e-01 3.84856462e-02
3.00587445e-01 3.85841370e-01 -1.26397192e+00 -1.86463863e-01
1.10121095e+00 9.31705415e-01 4.66236621e-01 5.06576717e-01
-1.75310485e-02 5.37014186e-01 -6.81539834e-01 -2.63554394e-01
1.43998871e-02 -3.70654225e-01 -6.36864364e-01 1.03669524e-01
2.32632488e-01 -2.23691970e-01 -1.50693208e-01 1.16165030e+00
3.78405809e-01 8.01602006e-03 3.26673687e-01 -1.57387757e+00
-3.37637246e-01 3.67717147e-01 -7.41932988e-01 3.36250186e-01
2.03643307e-01 2.79264390e-01 7.79620230e-01 -1.03566420e+00
4.79140967e-01 1.18955541e+00 5.40631831e-01 4.03062195e-01
-8.13930869e-01 -6.76910043e-01 3.01190495e-01 3.35552216e-01
-1.36256027e+00 -2.33889669e-01 6.55367613e-01 -7.89372399e-02
2.22216919e-01 7.53668666e-01 4.81880188e-01 9.82951939e-01
3.25789452e-01 1.52604437e+00 1.41838050e+00 2.38369772e-04
1.03222556e-01 1.70062482e-01 4.66399968e-01 6.62633121e-01
3.13501626e-01 -1.32895648e-01 -1.71548262e-01 -2.03374267e-01
4.09662157e-01 -3.01735643e-02 -6.79849029e-01 4.42682989e-02
-1.14615297e+00 5.60479522e-01 4.62518126e-01 3.14539909e-01
1.84890077e-01 9.62392315e-02 2.40840614e-01 -1.28089502e-01
4.97320354e-01 2.01208860e-01 -3.64469230e-01 2.25463845e-02
-9.35489714e-01 1.28659949e-01 6.51299417e-01 9.77846324e-01
4.81881738e-01 4.30348106e-02 -3.43220830e-01 6.00446999e-01
1.77322194e-01 7.37461686e-01 3.10761571e-01 -7.43869781e-01
8.25960934e-02 3.08195382e-01 2.02015787e-01 -1.07214892e+00
-5.67941904e-01 -3.14612389e-01 -9.49883521e-01 1.00274086e-01
3.17370743e-01 -9.20288563e-02 -1.19605446e+00 1.20329082e+00
5.61614454e-01 6.22264802e-01 -1.89633846e-01 1.03800321e+00
1.33336425e+00 6.60827518e-01 6.04982600e-02 -3.48253042e-01
1.28380358e+00 -1.36784506e+00 -8.89162719e-01 -4.22189087e-01
1.00558035e-01 -1.11270010e+00 1.22556376e+00 5.63690364e-01
-9.35971677e-01 -5.75024009e-01 -1.02169776e+00 4.02922218e-04
-3.11819971e-01 9.66784656e-02 8.82848442e-01 7.32988477e-01
-7.18348145e-01 2.99262166e-01 -8.68627489e-01 -2.79050678e-01
7.49723792e-01 1.48524597e-01 1.92956045e-01 -1.07655145e-01
-1.01042593e+00 6.19560778e-01 3.77035141e-01 3.34027469e-01
-8.76633763e-01 -2.39804313e-01 -4.52081203e-01 -1.66492462e-01
8.06568563e-01 -3.42703223e-01 1.37048328e+00 -7.13289797e-01
-1.45664489e+00 6.77224040e-01 -1.17787234e-01 -2.38855392e-01
8.47321093e-01 -4.66041118e-01 -4.06319737e-01 5.06097898e-02
-5.15068956e-02 3.50219399e-01 9.01236176e-01 -1.54515779e+00
-6.39597356e-01 -1.46297172e-01 7.48523548e-02 2.46291570e-02
-1.01345986e-01 2.17848957e-01 -9.07115161e-01 -8.42230320e-01
1.93919554e-01 -1.16119206e+00 -4.78880554e-01 7.13734403e-02
-9.72308517e-01 -2.02126548e-01 1.32160521e+00 -6.18950427e-01
1.42537057e+00 -2.31820941e+00 -3.04296672e-01 8.41096789e-02
2.58256614e-01 5.74237466e-01 1.98011428e-01 1.46107435e-01
4.33035761e-01 6.04908429e-02 -8.69262516e-01 -3.44314724e-01
-1.29626274e-01 1.53370023e-01 -1.56841308e-01 7.02695966e-01
-3.62115465e-02 1.04744506e+00 -8.73404145e-01 -9.53527927e-01
2.83837020e-01 3.11653078e-01 2.21466664e-02 2.85077393e-01
-2.35096395e-01 6.86311781e-01 -7.95854270e-01 1.22214830e+00
1.04880559e+00 -4.56750304e-01 -9.08769816e-02 -2.90557053e-02
-8.57113525e-02 -3.19592327e-01 -1.18798125e+00 1.27920747e+00
-6.18336387e-02 7.53615201e-01 7.02451244e-02 -6.02484584e-01
6.66494489e-01 3.26806121e-02 4.07398522e-01 -5.06217122e-01
4.16246176e-01 3.76987666e-01 5.85391968e-02 -5.34577250e-01
5.70531189e-01 2.95965701e-01 -5.01770340e-02 7.66637251e-02
-5.73142290e-01 -3.22514027e-01 1.18857451e-01 1.77893713e-01
8.95889819e-01 3.33669782e-02 4.58279938e-01 -4.17283922e-01
5.39957345e-01 2.50972688e-01 6.65598214e-01 7.26420224e-01
-4.98461008e-01 9.34130251e-01 2.80409276e-01 1.19371125e-02
-3.46248418e-01 -1.02989817e+00 -4.68490720e-01 8.83618653e-01
1.11656559e+00 -1.82107568e-01 -8.69714975e-01 -8.14568102e-01
-3.70434336e-02 6.06927752e-01 -5.04563510e-01 2.28413157e-02
-6.86402142e-01 -9.89875019e-01 4.76908863e-01 4.84971464e-01
8.57133627e-01 -1.25698149e+00 -9.10201788e-01 -5.50400056e-02
-2.20115855e-01 -1.20154643e+00 -6.30224347e-01 -8.46669450e-02
-5.99162877e-01 -1.07932937e+00 -1.06891167e+00 -8.09665322e-01
6.00917399e-01 9.22079980e-01 9.51642692e-01 5.37939668e-01
-6.29239619e-01 1.39145866e-01 -5.18972874e-01 -6.77351117e-01
-8.27018246e-02 1.10048108e-01 -3.15869331e-01 -1.20167760e-02
-1.27634630e-01 -1.18329309e-01 -9.14103985e-01 6.80267990e-01
-1.11317897e+00 2.80996740e-01 7.45460629e-01 6.44744217e-01
6.34761870e-01 -5.67202903e-02 3.99587542e-01 -1.36913991e+00
3.56940717e-01 -3.26254100e-01 -6.19207501e-01 2.46728122e-01
-5.32032371e-01 -6.04272604e-01 4.23359454e-01 -3.36240590e-01
-1.22903025e+00 -1.04034878e-01 -2.82824546e-01 -1.78854659e-01
-3.47699314e-01 3.82358767e-02 -3.80775332e-01 -1.51127353e-01
2.09900260e-01 1.94406450e-01 -3.07239711e-01 -5.29706955e-01
1.08534947e-01 6.57817543e-01 5.73935568e-01 -1.68552220e-01
1.02681315e+00 7.08261549e-01 -2.67646223e-01 -9.10089314e-01
-9.67163622e-01 -7.14628041e-01 -3.42415035e-01 -1.16725825e-01
7.92602837e-01 -4.71421897e-01 -5.72484195e-01 8.38689446e-01
-1.08416545e+00 -5.89741349e-01 5.04574068e-02 1.41371161e-01
-2.06736580e-01 5.63452065e-01 -4.44025785e-01 -7.37112403e-01
-3.84907544e-01 -1.21620691e+00 1.05955565e+00 6.86868191e-01
1.58091381e-01 -8.05800319e-01 -1.58086494e-01 3.14132303e-01
3.41088772e-01 4.96549964e-01 4.93704081e-01 -5.06466091e-01
-8.01103473e-01 -1.48924738e-01 -5.69333255e-01 3.15531522e-01
1.58878416e-01 1.78437114e-01 -1.11818540e+00 -3.06455344e-01
-1.74165808e-03 4.13327478e-02 1.05399704e+00 4.85625893e-01
1.66607976e+00 -9.96322557e-02 -7.03122675e-01 5.57098508e-01
1.35045779e+00 4.54177648e-01 6.43944025e-01 3.31161886e-01
7.97364473e-01 2.13049933e-01 1.20670545e+00 4.19767678e-01
2.27700517e-01 5.35477817e-01 5.62442422e-01 -5.60444713e-01
-2.94638276e-01 2.22677574e-01 6.70808032e-02 6.10219777e-01
-1.55014470e-02 -5.81391156e-01 -9.61902261e-01 3.42582732e-01
-2.04369664e+00 -4.94979978e-01 -6.33022606e-01 1.83335018e+00
5.86726308e-01 4.72692341e-01 -7.35463807e-03 3.46456692e-02
6.39437437e-01 5.06783307e-01 -8.47199619e-01 5.48657514e-02
-2.67721862e-01 1.05939195e-01 5.24139345e-01 3.72370780e-01
-1.35131621e+00 1.27123821e+00 5.71867132e+00 1.10422838e+00
-1.04582858e+00 4.38625902e-01 9.72536743e-01 4.44207102e-01
-1.37998387e-01 -2.88356338e-02 -7.66546369e-01 7.45132685e-01
2.14506567e-01 -1.34383768e-01 -1.17105626e-01 8.20544481e-01
1.85375959e-01 -6.35306656e-01 -4.57692802e-01 9.00016785e-01
7.52574429e-02 -1.02698743e+00 -3.57146442e-01 -2.09582254e-01
8.50674331e-01 -3.31432708e-02 3.05301100e-01 1.15787283e-01
-1.38762698e-01 -1.00394964e+00 5.80057919e-01 2.17599064e-01
5.06021917e-01 -4.90732044e-01 1.01243377e+00 2.90409565e-01
-1.27229333e+00 7.79649839e-02 -2.00466439e-01 3.72017801e-01
3.20429415e-01 7.74416685e-01 -7.35595703e-01 4.66294169e-01
6.71669006e-01 4.94220674e-01 -7.41504848e-01 1.55343056e+00
-3.68646324e-01 7.27815747e-01 -2.66433835e-01 -6.65561035e-02
3.03893983e-01 -1.37065247e-01 6.12790644e-01 1.53426456e+00
-7.43966084e-03 2.69818991e-01 5.05130053e-01 4.77306813e-01
1.69632852e-01 2.52142429e-01 -1.50383830e-01 2.26284266e-01
2.85548657e-01 1.45924819e+00 -1.45734787e+00 -4.26551551e-01
-3.15396160e-01 1.03074825e+00 -3.07663560e-01 4.75187540e-01
-1.44109929e+00 -3.63937050e-01 4.34794515e-01 -5.98266833e-02
3.30593325e-02 -2.73457468e-01 -4.51939464e-01 -9.58765864e-01
-6.90077469e-02 -6.76796734e-01 2.29438558e-01 -6.69718444e-01
-7.91692495e-01 6.42529845e-01 9.81978774e-02 -1.18066287e+00
5.20619869e-01 -5.31998217e-01 -8.15550864e-01 2.80662149e-01
-1.69546711e+00 -9.69173908e-01 -9.36721385e-01 5.11025608e-01
9.02387559e-01 2.14366198e-01 2.91383654e-01 3.17238361e-01
-9.77818906e-01 5.74599087e-01 6.70776293e-02 4.85148281e-02
6.16124272e-01 -1.19448352e+00 5.37052751e-01 9.56746876e-01
3.48016694e-02 2.21263200e-01 9.35023785e-01 -5.71409643e-01
-1.30173123e+00 -1.15150476e+00 2.49549389e-01 -2.41648361e-01
5.03570437e-01 -2.05053583e-01 -1.07344675e+00 3.87307793e-01
2.05322132e-01 9.90207493e-02 2.54122943e-01 -4.85072374e-01
1.82899356e-01 1.94750447e-02 -1.02930796e+00 7.32129037e-01
1.10741651e+00 9.80021730e-02 -5.41032627e-02 6.95536971e-01
7.96350718e-01 -7.95043528e-01 -4.25886780e-01 5.93503952e-01
3.75307560e-01 -1.10855389e+00 8.39345157e-01 -4.64892713e-03
1.52786568e-01 -3.99090827e-01 -1.52144004e-02 -8.75878394e-01
-9.84311476e-03 -7.51993358e-01 -3.22925240e-01 1.14178085e+00
2.33795494e-01 -8.13681901e-01 9.87458050e-01 3.23058635e-01
-3.47018868e-01 -1.17387390e+00 -8.73446763e-01 -8.72864306e-01
-3.72365445e-01 -2.82178044e-01 4.35495645e-01 6.46447897e-01
-5.51423788e-01 8.61882344e-02 -4.13005173e-01 3.21669400e-01
6.47817612e-01 5.52383840e-01 9.01565015e-01 -1.00215042e+00
-3.18864524e-01 -5.21070302e-01 -1.15494117e-01 -1.29469180e+00
5.70359789e-02 -5.83789587e-01 4.75843281e-01 -1.60059726e+00
3.39701444e-01 -6.31387472e-01 -2.92537987e-01 5.00147045e-01
-6.93095028e-01 5.28148592e-01 4.50957984e-01 2.48720556e-01
-8.41214597e-01 4.09642756e-01 1.68629146e+00 -2.37698272e-01
-1.67360976e-01 6.06184244e-01 -6.49071515e-01 1.03436351e+00
1.44375253e+00 -4.45388645e-01 -2.01655850e-01 -1.65863872e-01
-4.71731663e-01 -1.74882859e-01 5.36001444e-01 -9.99023855e-01
6.99924380e-02 -5.30437291e-01 6.84239119e-02 -6.43648267e-01
5.13211787e-01 -7.81403124e-01 -1.49668977e-02 8.80360484e-01
1.06833957e-01 -1.94253400e-01 1.88419297e-01 4.87103343e-01
-1.17489301e-01 -2.03403249e-01 1.06382656e+00 1.02231035e-03
-1.10531235e+00 4.92296249e-01 -1.04088299e-02 2.02936679e-01
1.20101094e+00 -3.36681962e-01 -5.06484091e-01 -1.15134403e-01
-5.37840784e-01 3.98228854e-01 4.04242605e-01 4.38814282e-01
8.13945413e-01 -9.04405177e-01 -5.40395916e-01 -6.85876608e-02
9.20486003e-02 2.96778437e-02 3.06888402e-01 1.22155941e+00
-5.96172035e-01 2.96268851e-01 7.33718872e-02 -8.05716336e-01
-1.56023777e+00 4.56114143e-01 3.39248553e-02 -1.34504512e-01
-9.16361511e-01 9.71964955e-01 6.24772668e-01 1.24983139e-01
3.64218950e-01 -6.30295932e-01 -3.23601365e-02 -2.71193802e-01
4.22721297e-01 6.30999684e-01 -1.46982446e-01 -4.63920921e-01
-4.99010772e-01 4.44646478e-01 2.73752899e-04 3.37914437e-01
8.45087528e-01 -1.18287817e-01 1.74916387e-02 4.67644393e-01
9.12016034e-01 -2.22717170e-02 -1.34995794e+00 -1.07473314e-01
6.39170334e-02 -7.72702098e-01 -9.31606293e-02 -5.63004792e-01
-1.24574804e+00 6.57759488e-01 7.83143044e-01 4.78407472e-01
1.15962040e+00 9.58404914e-02 1.31485128e+00 8.22710022e-02
5.50943911e-01 -9.36352074e-01 1.19145893e-01 3.10892671e-01
7.80558050e-01 -1.56976056e+00 2.93825448e-01 -8.92075479e-01
-7.32150197e-01 6.67669773e-01 7.25592256e-01 -8.80944729e-02
8.55816960e-01 3.83889854e-01 1.66875929e-01 1.18308933e-02
-1.72671735e-01 -4.53203976e-01 2.93680280e-01 3.61987740e-01
7.07366019e-02 3.20075780e-01 -5.99501252e-01 2.72990435e-01
-1.91985797e-02 -2.43051112e-01 5.03632903e-01 9.05724704e-01
-7.27757692e-01 -7.91404605e-01 -5.57303727e-01 4.60701257e-01
-4.73462045e-01 -9.03027207e-02 -4.85244244e-01 9.70830083e-01
2.10159972e-01 1.00152361e+00 -3.84488761e-01 -1.52466893e-01
1.91118360e-01 -6.45350575e-01 1.53791308e-01 -5.90834737e-01
-4.04077798e-01 1.92816511e-01 -4.45741937e-02 -6.15421593e-01
-6.07012808e-01 -5.74410617e-01 -1.46074843e+00 -1.02933832e-01
-5.03622055e-01 1.60943240e-01 4.82861280e-01 7.64647305e-01
-1.01353951e-01 6.04276359e-01 5.75452983e-01 -8.99419129e-01
-1.28246516e-01 -6.83832228e-01 -6.05598450e-01 1.06064461e-01
3.90830308e-01 -8.02229166e-01 -2.66746134e-01 -8.47325996e-02] | [9.642765045166016, -0.18088886141777039] |
41d2cf87-4228-4865-8955-adaefd3b91ca | independent-prototype-propagation-for-zero | 2106.00305 | null | https://arxiv.org/abs/2106.00305v2 | https://arxiv.org/pdf/2106.00305v2.pdf | Independent Prototype Propagation for Zero-Shot Compositionality | Humans are good at compositional zero-shot reasoning; someone who has never seen a zebra before could nevertheless recognize one when we tell them it looks like a horse with black and white stripes. Machine learning systems, on the other hand, usually leverage spurious correlations in the training data, and while such correlations can help recognize objects in context, they hurt generalization. To be able to deal with underspecified datasets while still leveraging contextual clues during classification, we propose ProtoProp, a novel prototype propagation graph method. First we learn prototypical representations of objects (e.g., zebra) that are conditionally independent w.r.t. their attribute labels (e.g., stripes) and vice versa. Next we propagate the independent prototypes through a compositional graph, to learn compositional prototypes of novel attribute-object combinations that reflect the dependencies of the target distribution. The method does not rely on any external data, such as class hierarchy graphs or pretrained word embeddings. We evaluate our approach on AO-Clever, a synthetic and strongly visual dataset with clean labels, and UT-Zappos, a noisy real-world dataset of fine-grained shoe types. We show that in the generalized compositional zero-shot setting we outperform state-of-the-art results, and through ablations we show the importance of each part of the method and their contribution to the final results. | ['Gertjan Burghouts', 'Doina Bucur', 'Frank Ruis'] | 2021-06-01 | null | http://proceedings.neurips.cc/paper/2021/hash/584b98aac2dddf59ee2cf19ca4ccb75e-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/584b98aac2dddf59ee2cf19ca4ccb75e-Paper.pdf | neurips-2021-12 | ['compositional-zero-shot-learning'] | ['computer-vision'] | [ 3.92454296e-01 1.52568623e-01 -1.21707052e-01 -5.23937762e-01
-2.54394621e-01 -6.03388011e-01 7.88805902e-01 3.08960319e-01
-3.31582844e-01 5.39858401e-01 -1.99887902e-03 2.50993297e-03
-3.17038178e-01 -9.40784335e-01 -8.91734004e-01 -6.74414098e-01
1.26569793e-02 9.80179191e-01 5.07245660e-01 -3.93502891e-01
1.08404376e-01 2.74086058e-01 -1.98757517e+00 4.69321847e-01
5.34222484e-01 8.58232319e-01 -8.95926282e-02 5.23231804e-01
8.31353944e-04 9.13021266e-01 -5.52577794e-01 -6.98278725e-01
2.77650267e-01 -4.15469080e-01 -6.06816649e-01 1.64795280e-01
8.98080230e-01 1.88401062e-02 -2.00176775e-01 9.84169543e-01
2.32758611e-01 3.60963732e-01 1.09283698e+00 -1.35728312e+00
-1.12590003e+00 8.54296923e-01 -5.44961333e-01 4.21546176e-02
4.03200462e-02 4.82565999e-01 1.48704326e+00 -9.11511898e-01
8.85916173e-01 1.37815154e+00 7.15171456e-01 6.76568747e-01
-1.84578013e+00 -5.24369597e-01 3.21394891e-01 5.37268937e-01
-1.39276576e+00 -1.84129983e-01 8.04642975e-01 -6.54589474e-01
7.26082325e-01 2.39358857e-01 5.93831301e-01 1.40499830e+00
-1.58284307e-01 7.69652665e-01 1.30257237e+00 -3.55272979e-01
4.44858342e-01 2.45703697e-01 5.37262022e-01 8.07650983e-01
4.49871212e-01 1.65754840e-01 -6.20986760e-01 -1.35687724e-01
3.67377698e-01 8.59474838e-02 -6.48516044e-02 -9.09722626e-01
-1.12576199e+00 8.11751604e-01 7.25182712e-01 8.69880319e-02
-2.18988299e-01 2.25933746e-01 1.49698719e-01 2.13342801e-01
1.62893504e-01 6.21999145e-01 -2.63637781e-01 3.01188648e-01
-7.62768269e-01 3.79732251e-01 7.08791435e-01 1.02615130e+00
8.94744337e-01 -1.82339415e-01 -1.94123730e-01 9.17334795e-01
-2.40437798e-02 4.91554350e-01 2.04541624e-01 -7.31386185e-01
2.17432790e-02 5.64853191e-01 -5.16791493e-02 -7.73415267e-01
-4.55745578e-01 -4.07983869e-01 -4.77550864e-01 4.84750897e-01
6.12710893e-01 1.38026595e-01 -1.46696222e+00 1.88660777e+00
2.64388412e-01 1.94612488e-01 2.90090386e-02 1.14045846e+00
5.87364137e-01 3.04210335e-01 1.08686462e-01 3.86957198e-01
1.46060634e+00 -8.58509898e-01 -3.53668958e-01 -4.85819101e-01
3.59080106e-01 -3.10386181e-01 1.47550714e+00 3.98658514e-01
-5.63989103e-01 -5.36949277e-01 -1.28750551e+00 -1.96852297e-01
-6.29826367e-01 -2.15880677e-01 6.65441394e-01 5.90456545e-01
-4.31550831e-01 6.97539091e-01 -6.99711919e-01 -5.04519582e-01
6.83144033e-01 1.08243398e-01 -3.63260299e-01 -3.16812217e-01
-9.50921118e-01 1.09723091e+00 2.80295819e-01 -2.36735374e-01
-1.09179270e+00 -6.82413220e-01 -9.05813932e-01 1.77629396e-01
8.02706003e-01 -8.05791914e-01 8.96781266e-01 -1.00458324e+00
-9.28756475e-01 9.86317933e-01 5.67468209e-03 -4.02548701e-01
4.37619925e-01 -1.28030896e-01 -3.27742189e-01 5.74225113e-02
5.93967028e-02 6.19460583e-01 1.06973636e+00 -1.60067630e+00
-6.63563550e-01 -4.53770429e-01 1.59012988e-01 -1.71941649e-02
-1.43501796e-02 -3.70223254e-01 6.15331121e-02 -4.68121350e-01
7.33918101e-02 -1.02011824e+00 7.16923326e-02 1.32161453e-01
-5.05146265e-01 -1.52869776e-01 5.80031633e-01 -1.07979126e-01
7.14450657e-01 -2.26478028e+00 3.06675762e-01 3.22253942e-01
2.50419527e-01 -2.39183810e-02 -6.34813607e-02 2.88854033e-01
1.10484481e-01 9.83572006e-02 -4.95268285e-01 -1.97245434e-01
2.81972349e-01 6.81844532e-01 -5.26855886e-01 5.43844402e-01
5.21738887e-01 8.65866959e-01 -1.18968558e+00 -2.58333117e-01
1.85775027e-01 3.16858053e-01 -4.58944827e-01 3.92501503e-02
-3.95809531e-01 -3.14496942e-02 1.31485254e-01 5.86063623e-01
3.50546926e-01 -3.02257776e-01 1.67767853e-01 -1.97076082e-01
4.39219385e-01 2.91854721e-02 -1.05528212e+00 1.35384142e+00
-3.77264261e-01 6.81659937e-01 -4.30757791e-01 -8.41800034e-01
9.36866581e-01 -1.30510345e-01 -3.66302699e-01 -3.46685469e-01
-1.68917347e-02 2.31718853e-01 3.38388115e-01 -5.85670650e-01
3.86431217e-01 -6.15660369e-01 -1.02381870e-01 2.28596285e-01
3.99970293e-01 -1.45384312e-01 1.58152103e-01 4.07772332e-01
1.08976293e+00 2.11362049e-01 2.49936923e-01 -1.18450165e-01
-1.87121645e-01 1.40557855e-01 4.92237359e-01 1.11369860e+00
1.98349971e-02 8.91837955e-01 7.75557578e-01 -4.68449861e-01
-1.21252227e+00 -1.46701145e+00 -1.44314334e-01 1.41710937e+00
4.21325713e-01 -2.85845876e-01 -1.44257233e-01 -8.54391515e-01
3.33261937e-01 1.24253547e+00 -1.26870251e+00 -3.94602507e-01
-2.43086517e-01 -5.69363356e-01 5.29264390e-01 6.26956582e-01
-7.81641006e-02 -1.01065135e+00 -9.10323203e-01 -7.85000101e-02
2.96486467e-01 -8.77110898e-01 -1.70950145e-02 6.17082357e-01
-4.93425757e-01 -1.25602102e+00 -3.39311302e-01 -5.20665109e-01
7.37324834e-01 1.17930450e-01 1.26840830e+00 1.48363277e-01
-5.61375260e-01 3.16676885e-01 -3.95396024e-01 -3.12265128e-01
-1.77445918e-01 -2.54375875e-01 -1.50611177e-02 1.22215912e-01
6.28203511e-01 -6.82715833e-01 -2.19131231e-01 2.13972688e-01
-7.72776425e-01 9.47422683e-02 5.31410992e-01 1.37841475e+00
4.88442570e-01 -1.70940951e-01 2.73297071e-01 -1.63487399e+00
4.56355065e-01 -5.82257628e-01 -2.95047313e-01 5.20733297e-01
-5.41925430e-01 3.98563087e-01 5.78805208e-01 -7.99692035e-01
-8.58307958e-01 -5.41404523e-02 3.78634900e-01 -6.10008538e-01
-4.09463912e-01 2.54414588e-01 -9.18704569e-02 2.79568762e-01
1.25680697e+00 -9.07225627e-03 -2.47663230e-01 -3.96783859e-01
9.36915159e-01 2.70763069e-01 7.69010723e-01 -7.55327106e-01
7.76802480e-01 6.32648230e-01 7.81327784e-02 -7.46861875e-01
-9.07520235e-01 -3.37214977e-01 -6.43160820e-01 1.88395043e-03
7.06661940e-01 -5.79056263e-01 -4.39438522e-01 1.17625698e-01
-8.85907590e-01 -3.99664730e-01 -6.06229901e-01 2.55749315e-01
-5.55954814e-01 -5.93608245e-02 -4.44697320e-01 -7.88339972e-01
2.90584862e-01 -7.46594250e-01 8.83848369e-01 1.57280266e-01
-4.61530149e-01 -7.01757848e-01 -1.14129737e-01 1.32187724e-01
1.40643314e-01 3.06373626e-01 1.36559618e+00 -1.11470735e+00
-4.14346159e-01 -1.26503199e-01 -1.97884724e-01 3.26497287e-01
5.84582426e-02 8.97480994e-02 -1.19625902e+00 -4.02225740e-02
-3.05184245e-01 -6.72744751e-01 1.22442067e+00 -1.79414019e-01
9.22254503e-01 -2.41731018e-01 -1.77637964e-01 5.17919481e-01
1.31443775e+00 -1.84518158e-01 3.59500915e-01 6.55723587e-02
8.98557782e-01 7.77935147e-01 4.40712124e-01 2.69445866e-01
2.51996726e-01 5.42048633e-01 3.99678677e-01 1.47567481e-01
-4.15508926e-01 -6.46756887e-01 9.49640013e-03 1.84781462e-01
-1.08672976e-01 -1.62427306e-01 -9.56671834e-01 7.80886352e-01
-1.88059604e+00 -9.87339854e-01 1.69009343e-02 2.10638571e+00
8.44184518e-01 3.47968400e-01 -6.05058484e-03 7.41046742e-02
6.25431597e-01 1.09992005e-01 -7.12571979e-01 -1.75652295e-01
-3.28329921e-01 3.89160961e-01 3.58540684e-01 5.38938642e-01
-9.94036615e-01 1.13017821e+00 5.48479748e+00 6.82013512e-01
-9.20238018e-01 1.31043941e-01 3.55029434e-01 -3.12728286e-01
-5.41076899e-01 1.80400416e-01 -5.31767607e-01 2.06731960e-01
4.67344284e-01 8.79825652e-02 6.09109581e-01 1.02044833e+00
-7.22807229e-01 -1.39846921e-01 -1.54936838e+00 8.72972131e-01
4.46666539e-01 -1.16544902e+00 6.62910491e-02 -2.08959356e-01
7.43044734e-01 -6.06071725e-02 1.51891038e-01 5.46738565e-01
6.54562831e-01 -1.13045025e+00 9.34154749e-01 4.51850176e-01
5.97966313e-01 -5.61957955e-01 6.50704205e-01 2.40207002e-01
-5.69657564e-01 -1.08608000e-01 -5.96148312e-01 -1.92581460e-01
-5.69996797e-02 3.61564279e-01 -9.81652260e-01 1.94294587e-01
4.95382011e-01 4.39886957e-01 -8.88082206e-01 9.33129907e-01
-8.51385057e-01 6.50837123e-01 -3.75171661e-01 -2.48032898e-01
9.34052542e-02 2.25684807e-01 4.65570152e-01 1.14042234e+00
2.58961786e-03 1.40874967e-01 1.98408458e-02 1.21714234e+00
-4.74376604e-02 -2.54489392e-01 -6.95900619e-01 -6.73954142e-03
4.99739379e-01 1.15549886e+00 -8.73341620e-01 -5.80540359e-01
-2.39035279e-01 8.59311819e-01 5.41009068e-01 5.59833229e-01
-6.98125839e-01 -4.84148949e-01 6.31438673e-01 1.71819434e-01
5.35223544e-01 5.33047505e-02 -5.32657743e-01 -1.10071182e+00
-3.25565897e-02 -7.81579018e-01 4.23657954e-01 -8.65942776e-01
-1.97847521e+00 4.19244826e-01 6.81054220e-02 -8.42823088e-01
-2.17333302e-01 -7.57644773e-01 -3.82826924e-01 6.12631559e-01
-1.20418668e+00 -1.31381106e+00 -3.39761287e-01 4.43881243e-01
3.64740998e-01 -6.71143308e-02 8.15697491e-01 -1.83935791e-01
-2.13557407e-01 4.14354742e-01 -2.76266187e-01 7.22718537e-02
8.52498949e-01 -1.45773113e+00 5.27064502e-01 7.88006604e-01
7.34807372e-01 8.82260323e-01 9.78336453e-01 -6.48778737e-01
-1.13983035e+00 -8.88178825e-01 5.37483573e-01 -7.53439665e-01
8.32663059e-01 -7.84082353e-01 -1.16461802e+00 8.11619341e-01
-7.84449056e-02 4.79490489e-01 6.39839411e-01 6.36586785e-01
-1.16686428e+00 7.70589709e-02 -9.73749042e-01 7.13039815e-01
1.22931838e+00 -6.91789687e-01 -1.16351938e+00 1.55385971e-01
6.10748112e-01 -8.07389021e-02 -3.33017647e-01 3.50337803e-01
6.45488441e-01 -1.04790127e+00 8.75212789e-01 -1.11542642e+00
5.63882887e-01 -5.15560806e-01 -4.04759377e-01 -1.58690190e+00
-5.55375278e-01 -3.13738063e-02 -7.92514905e-02 9.66260433e-01
6.08103335e-01 -5.18624902e-01 5.56489229e-01 6.69101834e-01
-8.84626359e-02 -7.01562762e-01 -7.35783577e-01 -1.04285777e+00
-1.27992824e-01 -4.45681095e-01 5.66798866e-01 1.04399192e+00
5.12007140e-02 7.07325339e-01 -4.97734725e-01 9.40733328e-02
7.77482748e-01 3.56021285e-01 8.80004585e-01 -1.28305113e+00
-6.51824176e-01 -3.88203740e-01 -8.24861705e-01 -5.59334695e-01
6.33855769e-03 -9.93063688e-01 2.55175948e-01 -1.32276750e+00
3.96287411e-01 -5.76782465e-01 -3.48474741e-01 8.12570035e-01
-4.07871991e-01 3.48937124e-01 4.90842015e-01 -3.72584611e-02
-5.96154690e-01 3.85069191e-01 9.40260589e-01 -4.63147730e-01
4.06965651e-02 -4.10028875e-01 -6.62564993e-01 7.82000721e-01
5.73982060e-01 -7.29236662e-01 -6.13900602e-01 -3.60805750e-01
4.26317751e-01 -5.17580688e-01 8.15108538e-01 -9.86917019e-01
1.92604780e-01 -1.29066125e-01 4.94557649e-01 -1.03440285e-01
5.75885832e-01 -7.48245418e-01 1.05639324e-01 2.51430005e-01
-6.13915741e-01 -3.47397923e-01 -7.59995207e-02 7.95267344e-01
1.01108029e-01 -2.20483318e-01 7.49463856e-01 -9.98044983e-02
-8.70903015e-01 -5.18182889e-02 3.70567441e-02 3.60390633e-01
8.61924469e-01 -4.73806895e-02 -9.96394873e-01 -8.91099051e-02
-8.11192751e-01 7.05531016e-02 9.15518939e-01 5.16093314e-01
6.53336704e-01 -1.11130261e+00 -5.98887920e-01 3.02964747e-01
6.10134184e-01 -2.43341506e-01 2.10546732e-01 6.07790649e-01
-2.68723875e-01 -2.81565815e-01 -4.29093391e-01 -5.78248560e-01
-1.09167910e+00 1.03670037e+00 1.02430180e-01 3.24022889e-01
-8.33067656e-01 1.07873392e+00 4.67071593e-01 -2.91493207e-01
1.83515608e-01 -3.36868137e-01 2.67708339e-02 1.57580197e-01
5.18459082e-01 4.17107567e-02 -1.18573658e-01 -5.10323882e-01
-4.12279069e-01 3.63762766e-01 -3.99823524e-02 -2.11440161e-01
1.31975889e+00 2.10320801e-01 1.35856140e-02 1.12615836e+00
8.51585686e-01 3.67693901e-02 -1.36480069e+00 -2.54330635e-01
1.51937127e-01 -7.75142848e-01 -3.56717587e-01 -1.10677707e+00
-7.57723629e-01 1.17276537e+00 4.54148233e-01 1.31340042e-01
6.44894481e-01 4.10264909e-01 3.29239011e-01 5.23575544e-01
4.62885201e-01 -8.46441329e-01 1.39150470e-01 2.90006459e-01
5.87112963e-01 -1.27205944e+00 6.02805391e-02 -3.81875604e-01
-1.00484490e+00 9.19762373e-01 5.21586537e-01 -3.71363938e-01
1.13356002e-01 4.00715955e-02 5.42887300e-03 -4.27708745e-01
-1.01783454e+00 -5.15970767e-01 3.33489120e-01 8.72745156e-01
1.52647952e-02 4.15938079e-01 7.34025240e-02 8.01942348e-01
-2.22609505e-01 -3.39590460e-01 5.04813075e-01 7.56989062e-01
-4.86838460e-01 -7.85610914e-01 -2.63322830e-01 6.08790576e-01
-3.14116627e-02 -1.05233841e-01 -6.35732710e-01 8.67552757e-01
4.62609649e-01 6.45346522e-01 1.11082971e-01 -6.29890263e-01
4.92172509e-01 3.20207298e-01 8.01773787e-01 -8.00498545e-01
-4.13681358e-01 -3.81093532e-01 1.96600512e-01 -2.56363779e-01
-7.78768137e-02 -5.48419297e-01 -1.24487162e+00 -1.38737485e-01
-1.20956764e-01 -1.67036474e-01 4.00296092e-01 8.25522900e-01
7.74325356e-02 5.26352406e-01 2.31139027e-02 -6.58205152e-01
-5.55157840e-01 -8.39711487e-01 -7.00426340e-01 1.10488236e+00
3.39365691e-01 -1.06699240e+00 -4.60607380e-01 -9.06977896e-03] | [10.250089645385742, 2.303956985473633] |
7740d8eb-0ff9-4d6c-8510-08baab61bb78 | smaragd-synthesized-smatch-for-accurate-and | 2203.13226 | null | https://arxiv.org/abs/2203.13226v2 | https://arxiv.org/pdf/2203.13226v2.pdf | SMARAGD: Learning SMatch for Accurate and Rapid Approximate Graph Distance | The similarity of graph structures, such as Meaning Representations (MRs), is often assessed via structural matching algorithms, such as Smatch (Cai and Knight, 2013). However, Smatch involves a combinatorial problem that suffers from NP-completeness, making large-scale applications, e.g., graph clustering or search, infeasible. To alleviate this issue, we learn SMARAGD: Semantic Match for Accurate and Rapid Approximate Graph Distance. We show the potential of neural networks to approximate Smatch scores, i) in linear time using a machine translation framework to predict alignments, or ii) in constant time using a Siamese CNN to directly predict Smatch scores. We show that the approximation error can be substantially reduced through data augmentation and graph anonymization. | ['Anette Frank', 'Philipp Meier', 'Juri Opitz'] | 2022-03-24 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [ 5.42026162e-01 4.64016527e-01 -3.99182320e-01 -2.77546167e-01
-1.01598978e+00 -9.52691376e-01 2.06004784e-01 5.71604013e-01
-1.97696090e-01 5.62764704e-01 1.53632417e-01 -5.23201108e-01
-3.33434999e-01 -1.00520778e+00 -9.79790866e-01 -1.99729413e-01
-9.70950872e-02 8.34851801e-01 -2.24017084e-01 2.01909654e-02
4.38247859e-01 5.20703793e-01 -1.11727333e+00 1.24212243e-01
9.55170035e-01 6.49601400e-01 -1.97703540e-01 6.69267297e-01
-2.17979863e-01 5.98824918e-01 -2.83306569e-01 -8.66713524e-01
5.24019480e-01 -5.13109982e-01 -1.23944199e+00 -2.46964365e-01
9.56171572e-01 -1.76979840e-01 -6.05988860e-01 1.44548714e+00
2.08966464e-01 2.39489991e-02 4.43201482e-01 -1.63718653e+00
-8.89565110e-01 1.05958927e+00 -6.30209327e-01 -1.63841829e-01
5.43479919e-01 -8.71575922e-02 1.58302093e+00 -6.23339355e-01
7.76741683e-01 1.27741623e+00 9.89113927e-01 2.34924898e-01
-1.53023541e+00 -6.39755070e-01 -2.46915504e-01 1.71074256e-01
-1.45663559e+00 -2.49274120e-01 6.41189694e-01 -9.73990709e-02
8.98984313e-01 3.41045380e-01 6.49345398e-01 7.79127598e-01
-2.65013784e-01 7.41650403e-01 6.98223650e-01 -1.22440033e-01
2.24877894e-01 -5.75623631e-01 9.88555551e-02 1.04833615e+00
4.89406198e-01 -7.20134154e-02 -3.60579878e-01 -5.37159622e-01
6.41120315e-01 -1.37232572e-01 -6.89159557e-02 -6.12589777e-01
-1.32498729e+00 1.02052248e+00 9.08877194e-01 1.63693517e-01
4.23056185e-02 5.48660159e-01 4.93331343e-01 7.35180855e-01
2.87834227e-01 9.49963272e-01 -3.15156996e-01 -6.01461381e-02
-7.59191573e-01 1.77955627e-01 8.71321678e-01 1.15196598e+00
1.03383219e+00 -2.56956547e-01 3.09189647e-01 5.21456301e-01
-2.10148871e-01 4.09496397e-01 2.50402480e-01 -1.20167792e+00
7.47503459e-01 7.33028829e-01 -2.61975080e-01 -1.49361551e+00
-3.00293595e-01 -1.45596907e-01 -9.43859637e-01 -2.42384195e-01
6.12726510e-01 1.13514140e-01 -6.20270133e-01 2.08072972e+00
3.47717226e-01 3.95214051e-01 -1.96009371e-02 7.31442630e-01
4.39333797e-01 4.50126350e-01 -3.61797899e-01 1.76904827e-01
1.19249582e+00 -9.33122873e-01 -1.88092947e-01 -3.14112872e-01
1.26580501e+00 -5.96433222e-01 1.30070555e+00 6.74763173e-02
-1.12705588e+00 7.75813833e-02 -1.03492892e+00 -4.28806454e-01
-3.00765038e-01 -4.64208871e-01 8.35746646e-01 5.84989250e-01
-1.24707866e+00 1.09013557e+00 -9.26109791e-01 -5.84342062e-01
7.00509310e-01 5.66679657e-01 -7.16756880e-01 -3.22769552e-01
-9.06110942e-01 6.69284105e-01 4.29212630e-01 -2.00084656e-01
-3.25952470e-01 -8.29747081e-01 -1.02004910e+00 3.81269842e-01
4.19227362e-01 -8.31862926e-01 9.38937247e-01 -1.20530438e+00
-1.18818069e+00 8.20915222e-01 -8.24397728e-02 -7.08958149e-01
3.04373205e-01 -8.06142983e-04 -5.58983050e-02 3.73090625e-01
9.12953448e-03 5.01604795e-01 5.33913910e-01 -6.95812225e-01
-2.42949322e-01 -5.53471744e-01 1.40769929e-01 2.62460321e-01
-4.44084108e-01 7.41219074e-02 -1.44735694e-01 -6.15265548e-01
2.89043754e-01 -1.03750086e+00 -3.47533286e-01 3.03515941e-01
-5.43792427e-01 -1.10463157e-01 5.70684135e-01 -9.07373548e-01
8.41527164e-01 -1.77786756e+00 2.77097881e-01 7.82312989e-01
6.97020352e-01 8.12520310e-02 -7.70928502e-01 5.41945219e-01
8.07692632e-02 3.47099572e-01 -7.35642910e-01 4.29010168e-02
1.72864050e-01 2.95722395e-01 1.58112254e-02 6.58485770e-01
3.13058883e-01 1.33544457e+00 -1.18466902e+00 -3.54671806e-01
-2.24089533e-01 1.00159489e-01 -8.10479462e-01 5.33386320e-02
-1.15862705e-01 1.25048041e-01 -2.66004354e-01 5.20424604e-01
6.56961203e-01 -6.40981078e-01 6.74586236e-01 -1.48023456e-01
4.81611222e-01 4.86461073e-01 -9.57344234e-01 2.06088662e+00
-4.91905898e-01 7.79006004e-01 -4.09030467e-02 -1.38610864e+00
8.84316921e-01 -1.20082587e-01 3.91983658e-01 -4.80670542e-01
4.82108034e-02 4.63964134e-01 -5.27133644e-02 8.77775848e-02
6.40855491e-01 3.31562638e-01 -2.10159734e-01 8.34724128e-01
-2.65977859e-01 1.90228981e-04 2.06663553e-02 5.58585048e-01
1.56448615e+00 -1.47009403e-01 3.17406684e-01 -2.13685572e-01
1.92179959e-02 2.38357395e-01 3.25945646e-01 5.21446943e-01
1.06852062e-01 5.47261596e-01 7.91881979e-01 -5.45695066e-01
-1.48616862e+00 -1.06104302e+00 4.30372983e-01 8.13544750e-01
2.99569398e-01 -4.07825917e-01 -1.05409372e+00 -8.57183933e-01
2.86789060e-01 3.34468156e-01 -3.13426048e-01 -5.11552155e-01
-9.33947802e-01 -3.77111077e-01 8.37413549e-01 5.42356193e-01
1.17442265e-01 -7.33706713e-01 -1.04845360e-01 1.73279032e-01
-1.82742923e-01 -1.19452691e+00 -9.65422630e-01 -1.31685376e-01
-1.06904578e+00 -1.30286920e+00 -3.36978674e-01 -9.47124064e-01
1.01170695e+00 3.68554533e-01 1.28517365e+00 5.25099814e-01
-1.64348647e-01 2.47657765e-02 -2.01605678e-01 4.39686686e-01
-6.87494278e-01 4.68589097e-01 -1.26743928e-01 -2.54375160e-01
1.86623469e-01 -1.11023390e+00 -6.44866586e-01 2.89789319e-01
-7.73420334e-01 3.47163565e-02 6.54265523e-01 8.54656756e-01
6.73771679e-01 -3.29698443e-01 4.60861683e-01 -1.25871253e+00
6.69185102e-01 -3.51461858e-01 -9.71177697e-01 3.69473815e-01
-9.49453413e-01 4.16468471e-01 8.84708643e-01 -4.31700975e-01
-1.20400153e-01 7.09152371e-02 -7.82923698e-02 -6.00639820e-01
2.74973392e-01 6.89499021e-01 -1.30693614e-01 -6.21444285e-01
6.81004465e-01 4.30662856e-02 3.15204769e-01 -3.04650992e-01
7.65563309e-01 4.80799556e-01 8.18800449e-01 -5.83543658e-01
1.16108584e+00 2.95048326e-01 3.41675550e-01 -2.91051477e-01
-8.58288646e-01 -2.71845043e-01 -5.58929682e-01 3.77980769e-01
5.83149672e-01 -7.40282476e-01 -7.59785533e-01 -9.96236354e-02
-1.08246303e+00 -3.53413373e-01 -1.95466891e-01 2.86040574e-01
-6.70930862e-01 7.53487766e-01 -7.50977457e-01 -2.12465942e-01
-6.80944979e-01 -7.70846784e-01 9.77237165e-01 -2.72521794e-01
-5.52586138e-01 -8.63399267e-01 -1.16008939e-02 5.97533584e-01
2.39405379e-01 4.93176907e-01 1.27856445e+00 -9.66303945e-01
-8.64786088e-01 -3.67199630e-01 -6.35495842e-01 -8.90981033e-02
5.63062094e-02 -1.89870954e-01 -4.30628389e-01 -5.66148698e-01
-6.67017639e-01 -4.56437707e-01 4.08396751e-01 -3.91117623e-03
1.42821360e+00 -9.84714866e-01 -2.36350924e-01 1.00074339e+00
1.28930533e+00 -5.26356280e-01 6.24959946e-01 7.15316683e-02
1.07978845e+00 3.42784613e-01 3.20325285e-01 1.01127133e-01
4.42994714e-01 5.86126924e-01 4.56414044e-01 -5.56850322e-02
-3.08685362e-01 -8.67974997e-01 1.26488611e-01 1.25528491e+00
2.00635016e-01 -1.40034392e-01 -8.90208006e-01 6.22084498e-01
-1.93962121e+00 -8.47993076e-01 -6.45928532e-02 2.23348093e+00
7.91776896e-01 -1.69173017e-01 1.53009802e-01 1.48126371e-02
9.66388643e-01 -2.53160391e-03 -6.89721525e-01 -6.75203323e-01
-1.26639768e-01 4.36766356e-01 8.54282320e-01 5.57298958e-01
-6.76550090e-01 1.14971626e+00 5.90048504e+00 8.54154348e-01
-6.64926946e-01 -8.74548778e-02 5.52512109e-01 3.89497057e-02
-6.91116154e-01 4.22280073e-01 1.77708387e-01 2.86978602e-01
1.03654790e+00 -4.41855907e-01 1.19900239e+00 7.76536882e-01
-2.70108908e-01 6.60381973e-01 -1.28046501e+00 1.10113811e+00
5.38085513e-02 -1.65363693e+00 1.87811628e-01 1.70168012e-01
8.48548770e-01 1.52558163e-01 -1.62768453e-01 -6.37291968e-02
9.17288780e-01 -1.21014535e+00 3.49541813e-01 8.27617124e-02
1.04778159e+00 -1.06226361e+00 4.37689036e-01 -2.72030719e-02
-1.37918317e+00 3.21515724e-02 -6.35331154e-01 1.97113901e-01
6.63590431e-03 5.16445875e-01 -9.18972075e-01 6.76264346e-01
2.88948536e-01 7.31394827e-01 -3.90545964e-01 9.12109554e-01
-2.91722268e-01 4.41986173e-01 -4.00346100e-01 -1.87803090e-01
2.54401058e-01 -6.18239284e-01 5.49160540e-01 6.78048313e-01
6.76946342e-01 -1.82890758e-01 1.56163588e-01 1.02180600e+00
-8.22126389e-01 1.33729443e-01 -8.48338425e-01 -5.97606897e-01
8.58414292e-01 1.06998420e+00 -7.43960440e-01 -1.92366511e-01
-1.76628307e-01 1.37807906e+00 8.94652665e-01 1.41923964e-01
-6.31245852e-01 -6.98930502e-01 8.39009345e-01 4.17619273e-02
1.56272739e-01 -3.20373952e-01 -4.92342800e-01 -1.19850492e+00
7.89870471e-02 -1.08990347e+00 6.49745047e-01 -8.08460593e-01
-1.44110548e+00 5.03936112e-01 -3.53380144e-01 -8.97100985e-01
-1.81828737e-01 -3.92783284e-01 -5.60537398e-01 4.83175457e-01
-1.14285624e+00 -1.11307681e+00 -1.02587543e-01 5.46373606e-01
2.68662279e-03 3.24883685e-02 7.45132387e-01 3.63978654e-01
-3.88196468e-01 1.18337202e+00 7.03063831e-02 2.49140531e-01
4.40639049e-01 -1.34117293e+00 1.13883030e+00 7.98877597e-01
6.19379044e-01 4.84108210e-01 5.01419365e-01 -4.52728719e-01
-1.72943211e+00 -1.29282713e+00 1.06773472e+00 -2.81086832e-01
9.87897277e-01 -4.80569273e-01 -9.84958291e-01 8.11590552e-01
-1.50988489e-01 1.26660854e-01 4.51381445e-01 9.78370234e-02
-8.43265176e-01 4.20621000e-02 -1.30774832e+00 1.01824713e+00
1.82326269e+00 -9.05965567e-01 -2.61404157e-01 5.62384784e-01
1.24046099e+00 -3.98555070e-01 -1.19366264e+00 1.90072984e-01
4.99508202e-01 -4.86783952e-01 9.38501775e-01 -1.13529229e+00
3.76461238e-01 -5.09499796e-02 -1.97819069e-01 -1.27670717e+00
-1.94294006e-01 -1.05471897e+00 -2.41144791e-01 8.69388819e-01
7.70112693e-01 -8.81039679e-01 1.19682288e+00 5.39387047e-01
7.09818453e-02 -5.48688173e-01 -1.06135964e+00 -9.38631475e-01
1.50676072e-01 -5.13987131e-02 1.22750378e+00 1.43738937e+00
1.78250208e-01 3.91901851e-01 -2.52824217e-01 2.51630127e-01
6.47414505e-01 5.57104647e-01 8.92732084e-01 -1.36424541e+00
-1.66683361e-01 -6.37418985e-01 -7.27077842e-01 -8.97047818e-01
5.66401601e-01 -1.56574512e+00 -3.13319921e-01 -1.57225871e+00
3.26531619e-01 -4.45096701e-01 -6.61531612e-02 5.56059062e-01
-3.85183096e-02 2.94473976e-01 -2.93035563e-02 2.11779624e-01
-5.66342056e-01 5.27804136e-01 1.04049873e+00 -2.03498483e-01
1.22384906e-01 -2.94440866e-01 -7.89875984e-01 4.68585134e-01
9.87549901e-01 -8.54976416e-01 -4.58095491e-01 -6.38322353e-01
7.03462601e-01 1.14130042e-01 2.71650791e-01 -6.52534187e-01
2.32159242e-01 -1.19159333e-01 -2.35598311e-01 -2.64185995e-01
6.22775964e-02 -6.45947039e-01 1.97012722e-01 6.25084877e-01
-6.87647462e-01 6.22452736e-01 -1.41041964e-01 6.72035098e-01
-1.27037331e-01 -1.46888599e-01 5.47059536e-01 1.92503646e-01
-1.39648214e-01 7.36121953e-01 1.19593315e-01 4.55012649e-01
5.82214832e-01 -1.94136709e-01 -5.17087698e-01 -5.98634958e-01
-1.67697594e-01 8.37156773e-02 8.95690918e-01 3.22953425e-02
6.81307614e-01 -1.55365193e+00 -5.97275257e-01 -1.27613991e-01
1.56808898e-01 1.08497694e-01 -1.87920779e-01 8.62541318e-01
-7.85112977e-01 -2.02778876e-01 -1.27475321e-01 -1.89447746e-01
-1.17673576e+00 6.44632280e-01 1.57396510e-01 -4.13892299e-01
-7.78234482e-01 4.40264583e-01 -9.17421579e-02 -8.79216611e-01
-5.89579679e-02 -1.98143318e-01 4.59340602e-01 -4.75556284e-01
1.97253957e-01 4.36408520e-01 2.22278256e-02 -4.34568673e-01
-3.00977290e-01 6.28133953e-01 1.62043706e-01 9.41255018e-02
1.33922195e+00 2.49063019e-02 -4.79707778e-01 -3.73869896e-01
1.61719370e+00 -2.29269322e-02 -8.59992206e-01 -4.37565029e-01
5.45722723e-01 -5.99525690e-01 -2.63660401e-01 -3.59280467e-01
-1.17788923e+00 5.83563626e-01 1.46443456e-01 8.15188289e-02
9.99320388e-01 9.69224796e-03 1.42731404e+00 1.01473916e+00
3.83530825e-01 -8.70472789e-01 5.83690144e-02 2.85989791e-01
6.48186147e-01 -9.59388852e-01 -1.47325203e-01 -6.43352509e-01
-4.17504221e-01 1.02009451e+00 1.95529655e-01 -4.30778980e-01
2.55243927e-01 1.44806862e-01 -3.53192478e-01 -2.06244305e-01
-5.98871350e-01 6.15722574e-02 2.66266435e-01 6.95245206e-01
-4.32768129e-02 1.29157171e-01 -1.25686228e-01 1.44609272e-01
-6.57663524e-01 -3.91184181e-01 5.62562764e-01 5.79495907e-01
-2.51960576e-01 -1.10558367e+00 8.05244595e-02 6.49250984e-01
-2.63420701e-01 -4.19244945e-01 -8.67450297e-01 6.32851422e-01
-7.05105245e-01 7.97520339e-01 1.63728058e-01 -5.53411841e-01
8.79209116e-02 -1.82126656e-01 5.15372813e-01 -3.52327496e-01
-4.49146271e-01 -6.29119992e-01 4.70343590e-01 -8.28240097e-01
2.08730176e-02 -2.78081834e-01 -1.38166511e+00 -9.69604969e-01
-1.75713107e-01 9.67350751e-02 4.64169174e-01 7.59342611e-01
8.90716195e-01 -5.35884686e-02 6.73227549e-01 -1.63493171e-01
-7.65138745e-01 -6.28861129e-01 -4.89737988e-01 7.96640933e-01
-1.54056475e-01 -1.62137896e-01 -4.57176119e-01 -3.46858442e-01] | [7.098536968231201, 6.190821647644043] |
fbd1f793-aeb3-465d-805f-9c2ed1a0eaa4 | uet-headpose-a-sensor-based-top-view-head | 2111.07039 | null | https://arxiv.org/abs/2111.07039v1 | https://arxiv.org/pdf/2111.07039v1.pdf | UET-Headpose: A sensor-based top-view head pose dataset | Head pose estimation is a challenging task that aims to solve problems related to predicting three dimensions vector, that serves for many applications in human-robot interaction or customer behavior. Previous researches have proposed some precise methods for collecting head pose data. But those methods require either expensive devices like depth cameras or complex laboratory environment setup. In this research, we introduce a new approach with efficient cost and easy setup to collecting head pose images, namely UET-Headpose dataset, with top-view head pose data. This method uses an absolute orientation sensor instead of Depth cameras to be set up quickly and small cost but still ensure good results. Through experiments, our dataset has been shown the difference between its distribution and available dataset like CMU Panoptic Dataset \cite{CMU}. Besides using the UET-Headpose dataset and other head pose datasets, we also introduce the full-range model called FSANet-Wide, which significantly outperforms head pose estimation results by the UET-Headpose dataset, especially on top-view images. Also, this model is very lightweight and takes small size images. | ['Long Tran Quoc', 'Duc Tran Minh', 'Hoang Nguyen Viet', 'Tuan Nguyen Dinh', 'Linh Nguyen Viet'] | 2021-11-13 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [-5.90142429e-01 -1.13576509e-01 1.23565741e-01 -6.71482027e-01
-5.93920410e-01 -1.57481492e-01 1.62052959e-01 -2.27108851e-01
-5.54987609e-01 5.32841384e-01 3.56885642e-01 2.46181190e-01
-9.55966581e-03 -4.40813512e-01 -5.55605531e-01 -6.11747622e-01
3.40701416e-02 7.04917371e-01 1.98002532e-01 -3.75539869e-01
2.79739857e-01 3.86255682e-01 -1.76470304e+00 -4.13511604e-01
4.82944459e-01 9.40797567e-01 5.36580205e-01 4.74020362e-01
2.60889292e-01 2.19002709e-01 -5.39993703e-01 -1.63533241e-01
1.88404739e-01 1.37871966e-01 -3.52019340e-01 2.95791715e-01
4.51831222e-01 -7.41666257e-01 5.13115190e-02 8.53488684e-01
1.28808308e+00 4.17306311e-02 4.55784380e-01 -1.54591203e+00
-7.44447112e-02 2.35336170e-01 -1.14108872e+00 5.18559851e-02
1.10664415e+00 -1.97798714e-01 1.51963010e-01 -9.15867746e-01
4.92741436e-01 1.39133382e+00 7.54158139e-01 6.90869391e-01
-3.57549995e-01 -9.97605681e-01 1.90267250e-01 3.80911559e-01
-1.64183533e+00 -4.70567107e-01 7.05688655e-01 -1.90830126e-01
4.39228863e-01 4.31507587e-01 6.51387930e-01 1.03412461e+00
4.56572650e-03 9.32347476e-01 1.13627148e+00 -2.22891167e-01
1.46940351e-01 2.60471523e-01 4.23610628e-01 3.04549962e-01
5.68340957e-01 -3.40502113e-01 -4.59802508e-01 2.84163933e-02
5.46480834e-01 3.18375766e-01 -7.94145286e-01 -5.22024691e-01
-9.87715423e-01 6.22489095e-01 3.99585456e-01 1.92149267e-01
-2.22717941e-01 -6.49651408e-01 3.50583494e-01 -1.41332477e-01
2.53370166e-01 -1.23062931e-01 -6.49049103e-01 -5.96663579e-02
-5.01312673e-01 3.53337556e-01 8.76985431e-01 1.54808629e+00
5.16092181e-01 -2.84027666e-01 2.95357317e-01 9.25811052e-01
5.27310014e-01 8.61423314e-01 7.94552207e-01 -4.57628429e-01
6.44820690e-01 4.19620246e-01 5.60699739e-02 -1.10035753e+00
-1.15985382e+00 5.96410967e-02 -6.69653416e-01 -2.62516260e-01
2.19562754e-01 -3.23989034e-01 -6.72352433e-01 1.46654058e+00
8.41151297e-01 -2.05458552e-01 -1.42696306e-01 1.28503358e+00
1.27321267e+00 3.59574676e-01 -3.24455231e-01 -4.30715382e-01
1.66604447e+00 -7.43468404e-01 -9.60339427e-01 -2.08132565e-01
6.08954370e-01 -8.59281659e-01 8.69497061e-01 8.04950356e-01
-7.58206189e-01 -4.18583930e-01 -9.62695718e-01 -1.54917240e-01
-2.80156225e-01 2.43891701e-01 6.19066834e-01 7.51140177e-01
-1.04698849e+00 -9.98072550e-02 -5.94190896e-01 -8.88579488e-01
-2.23330259e-01 6.50601566e-01 -8.05944443e-01 -9.76138562e-02
-9.14552808e-01 9.43325937e-01 3.43478590e-01 3.47361088e-01
-3.90711784e-01 -8.89528617e-02 -1.09228683e+00 -3.15561384e-01
3.79933417e-01 -3.59284610e-01 1.32799840e+00 -6.35990202e-02
-1.72592366e+00 7.28427291e-01 -4.07065541e-01 5.23278862e-02
5.13356149e-01 -7.25799978e-01 -3.00621331e-01 -1.53020218e-01
7.81475529e-02 5.78096569e-01 5.21554112e-01 -1.16199720e+00
-5.19452393e-01 -1.28415430e+00 -3.29851240e-01 6.17115676e-01
-3.13818276e-01 9.95530412e-02 -8.28304231e-01 -3.08430661e-02
6.67607427e-01 -1.18743110e+00 1.86174177e-02 -3.40259194e-01
-5.71979165e-01 -2.55706906e-01 9.01198328e-01 -7.87976980e-01
1.06999457e+00 -1.75334013e+00 -6.89977854e-02 1.39010370e-01
1.59561366e-01 3.27785499e-02 4.88965809e-01 9.51693058e-02
-7.22611919e-02 -4.50397402e-01 3.50867420e-01 -5.75661719e-01
-9.51017812e-02 1.49143293e-01 2.35571161e-01 8.36625814e-01
-7.73959041e-01 2.98790425e-01 -5.54554880e-01 -7.09377646e-01
4.59041119e-01 4.92110342e-01 -5.59419751e-01 3.51177514e-01
5.86559057e-01 6.05607390e-01 -4.02946055e-01 7.91295767e-01
1.37402952e+00 2.87948698e-01 -3.20244278e-03 -2.65820473e-01
-2.25085229e-01 -2.08826870e-01 -1.51703584e+00 1.56685030e+00
-3.80058676e-01 2.52640069e-01 4.09719259e-01 -5.80031037e-01
9.84451115e-01 2.41163984e-01 3.55935901e-01 -6.35984004e-01
7.14726627e-01 2.28094965e-01 -3.70678306e-01 -1.04412675e+00
5.34197748e-01 8.92578736e-02 -1.19236514e-01 1.01340145e-01
-1.36259332e-01 -1.94605187e-01 -8.33668187e-02 -2.23342761e-01
2.96907842e-01 2.45326415e-01 5.03892183e-01 -2.05278486e-01
6.70849442e-01 -5.95555186e-01 6.10771298e-01 5.36984997e-03
-3.70045781e-01 9.95421648e-01 9.67857428e-03 -2.29649797e-01
-6.39153063e-01 -5.34883738e-01 -4.61887419e-01 1.02130365e+00
2.99635530e-01 -3.40356052e-01 -1.08152211e+00 -3.76735449e-01
2.02761535e-02 3.00499201e-01 -4.07859147e-01 2.65963137e-01
-8.01175177e-01 -8.22699726e-01 1.06767453e-01 5.67069232e-01
7.31413364e-01 -6.28122330e-01 -5.66018939e-01 -2.65183091e-01
-2.99634606e-01 -1.12624419e+00 -6.58807993e-01 -2.39750117e-01
-7.14407444e-01 -1.09333575e+00 -1.22704995e+00 -7.88985610e-01
8.04366648e-01 5.77378035e-01 7.79799581e-01 -3.18416238e-01
-5.16900681e-02 3.10552180e-01 -3.72049868e-01 -8.72104883e-01
6.60843730e-01 3.02863885e-02 5.68278074e-01 -8.69180560e-02
7.71798253e-01 -4.11904126e-01 -8.23480666e-01 6.98422849e-01
-2.69167334e-01 -2.53316700e-01 5.49889088e-01 4.08551931e-01
2.95922279e-01 -3.20326030e-01 1.53359458e-01 -6.07691467e-01
5.34188509e-01 -5.14544845e-01 -5.03933668e-01 1.07384529e-02
-3.55554700e-01 -2.33328670e-01 2.25943759e-01 -2.65297234e-01
-9.37419057e-01 1.59136638e-01 -4.86528367e-01 -3.78527910e-01
-3.57610404e-01 6.92864209e-02 -7.88438499e-01 5.76099046e-02
3.72692734e-01 2.55657643e-01 1.61639065e-01 -8.41360927e-01
-9.60806087e-02 1.26315606e+00 6.06771350e-01 -2.76729196e-01
4.26395237e-01 4.32487667e-01 -7.68691525e-02 -1.19358635e+00
-4.67993766e-01 -8.66304636e-01 -9.82225299e-01 -3.22732806e-01
8.36352706e-01 -1.04103339e+00 -1.23267925e+00 8.37125182e-01
-1.14425039e+00 4.75493133e-01 7.58075535e-01 1.00615287e+00
-3.49855632e-01 4.34855253e-01 -3.80412340e-01 -1.04855275e+00
-5.30859709e-01 -1.22852063e+00 1.67518032e+00 4.82512921e-01
-1.68897003e-01 -6.47238195e-01 -1.41153678e-01 4.74417537e-01
1.45381421e-01 2.73803562e-01 -6.63844720e-02 -3.01859736e-01
-2.13820517e-01 -5.24050057e-01 2.88280342e-02 -2.60483362e-02
5.65330777e-03 -3.83151948e-01 -9.85126197e-01 -5.63455164e-01
4.00821388e-01 -2.20512316e-01 1.56621993e-01 5.40905774e-01
1.12097812e+00 -1.36091143e-01 -6.00667953e-01 7.50098467e-01
1.07730627e+00 3.41241211e-01 6.40154421e-01 5.67574680e-01
8.05295765e-01 8.42530787e-01 9.33194518e-01 6.87546611e-01
1.05638874e+00 1.00251353e+00 1.94584325e-01 8.79884958e-02
3.31848145e-01 -2.93737352e-01 2.47005716e-01 1.22263992e+00
-2.41050467e-01 -1.42947450e-01 -6.86840892e-01 1.11260392e-01
-1.73133373e+00 -6.69794679e-01 -3.78461510e-01 2.45407486e+00
4.55284327e-01 -1.04555450e-01 3.68395567e-01 4.55298007e-01
7.43289590e-01 -1.69713229e-01 -5.06229222e-01 -7.30583295e-02
2.98108637e-01 -2.56345421e-01 5.74267685e-01 3.09425443e-01
-9.77889657e-01 4.20468956e-01 5.86460686e+00 3.25561076e-01
-1.39465451e+00 8.86529237e-02 1.53734654e-01 -9.99742150e-02
3.36077422e-01 -4.43490446e-01 -1.46660268e+00 6.97333813e-01
6.19357288e-01 1.89812675e-01 -2.08452970e-01 1.25925136e+00
2.46300027e-01 -6.08783543e-01 -1.07506657e+00 1.67818666e+00
7.15743005e-01 -1.16456889e-01 -4.23637211e-01 2.55699784e-01
3.58567247e-03 -3.07577103e-01 -1.00037940e-01 4.56626922e-01
-4.03380454e-01 -5.90079129e-01 7.25091100e-01 2.81648189e-01
5.22141635e-01 -6.90758467e-01 1.03190434e+00 7.82656431e-01
-1.31987643e+00 1.15640618e-01 -6.01180136e-01 -2.50480831e-01
3.11622471e-01 4.50372398e-01 -9.28922057e-01 4.50903922e-01
1.01655233e+00 5.42485833e-01 -6.32835031e-01 1.17727625e+00
-1.23744942e-01 -7.19950646e-02 -7.28125274e-01 -2.97200024e-01
-3.68276477e-01 -1.37241557e-01 3.58599424e-01 9.34394717e-01
5.76628506e-01 3.05125594e-01 2.50558317e-01 -1.11857457e-02
3.17419946e-01 2.66724557e-01 -8.49070549e-01 1.04132426e+00
5.44307292e-01 1.36301768e+00 -4.13543344e-01 -4.89693843e-02
-2.90289670e-01 8.30548346e-01 -1.13399187e-02 9.65272356e-03
-9.55039799e-01 -4.01278824e-01 3.45539212e-01 2.27271125e-01
1.28111420e-02 -2.47864008e-01 -3.83499041e-02 -1.34728050e+00
2.51642793e-01 -7.06497550e-01 2.51541972e-01 -1.09453952e+00
-7.11446285e-01 6.54259026e-01 4.51665550e-01 -1.61515486e+00
-4.05039847e-01 -8.18557918e-01 -3.16130131e-01 6.40348852e-01
-1.19862556e+00 -1.12487364e+00 -9.58331764e-01 8.83326232e-01
6.39668643e-01 1.96367323e-01 6.33256137e-01 5.99947095e-01
-1.05223358e+00 9.56540644e-01 -1.66642696e-01 -7.53595307e-02
9.17666495e-01 -1.05264235e+00 -1.41971871e-01 2.79835463e-01
-5.06590188e-01 9.70021605e-01 9.46374536e-01 -5.15942633e-01
-1.68839037e+00 -5.04407704e-01 7.02056468e-01 -6.55137956e-01
-2.02813029e-01 -5.57545066e-01 -4.85940427e-01 8.16952646e-01
1.50850460e-01 1.58559978e-02 5.11563897e-01 1.66587099e-01
1.15714639e-01 -3.28612864e-01 -1.43215668e+00 1.91386193e-01
1.12437212e+00 8.53221118e-02 -7.03763545e-01 1.90224245e-01
2.80675858e-01 -9.76215303e-01 -7.03274667e-01 5.29513478e-01
8.00788939e-01 -1.06370175e+00 8.04523170e-01 1.81637421e-01
-2.13238016e-01 -3.50400418e-01 -2.01258808e-01 -1.24126184e+00
6.67115971e-02 -4.24373209e-01 5.72695136e-02 1.30627501e+00
6.50041923e-02 -7.97498703e-01 9.42709804e-01 8.24512303e-01
-5.95653709e-03 -6.72375560e-01 -6.93282902e-01 -6.61834598e-01
-3.04939657e-01 -1.94277644e-01 9.26615417e-01 5.69154382e-01
1.96829677e-01 5.70526481e-01 -5.60037196e-01 2.76051968e-01
7.46421754e-01 -1.66048363e-01 1.42063141e+00 -1.39024258e+00
1.69118941e-01 1.97648093e-01 -7.17918932e-01 -1.37432837e+00
-2.25327447e-01 3.57595715e-03 9.68910828e-02 -1.45720744e+00
3.03040296e-01 -1.52256221e-01 3.97385359e-01 8.99333507e-02
-1.18677348e-01 2.62798388e-02 5.04168272e-02 2.57722944e-01
-4.76456970e-01 4.51627105e-01 1.32563877e+00 3.01778376e-01
-1.66189000e-01 2.44123548e-01 -5.16647160e-01 1.15097940e+00
4.75501329e-01 -9.60723907e-02 -5.11205435e-01 -4.96419489e-01
-2.62303025e-01 3.14485997e-01 -1.30625546e-01 -1.36892819e+00
4.94997919e-01 1.08185075e-01 7.54417181e-01 -1.24206161e+00
7.59226680e-01 -9.29474413e-01 4.73435000e-02 3.33098114e-01
4.57705319e-01 5.21697938e-01 -1.35856923e-02 1.68082789e-01
-2.08831370e-01 -2.30683893e-01 6.18774235e-01 -3.00615191e-01
-8.67199659e-01 3.91952664e-01 5.96066341e-02 -1.63879588e-01
1.41329861e+00 -4.54515159e-01 -2.23241344e-01 -6.80277288e-01
-5.95089674e-01 5.10141671e-01 4.26952094e-01 6.00104034e-01
6.99080229e-01 -1.34253168e+00 -3.35496128e-01 5.21920204e-01
1.93239450e-01 3.72659475e-01 2.95446336e-01 1.07305610e+00
-6.38987422e-01 7.35764146e-01 -2.14718744e-01 -8.90322626e-01
-1.65176988e+00 4.36289668e-01 5.53713404e-02 2.17053950e-01
-5.41103721e-01 8.23566079e-01 2.25200355e-01 -9.34862912e-01
4.96349365e-01 -4.30778712e-01 -9.55395639e-01 2.50475526e-01
8.11569154e-01 5.30851781e-01 1.83041468e-01 -1.27126408e+00
-5.22983670e-01 1.14259040e+00 -2.83740759e-02 -4.60484549e-02
1.20266354e+00 -6.68502331e-01 1.83184054e-02 2.19779685e-01
1.50312257e+00 1.03640631e-01 -7.93826580e-01 1.49223417e-01
-1.13243759e-01 -6.46200418e-01 -1.49777502e-01 -2.47216597e-01
-1.08911777e+00 8.76522899e-01 1.11524391e+00 -2.91032404e-01
1.11290014e+00 -1.59281313e-01 1.04054093e+00 4.68674392e-01
1.28799117e+00 -1.00557888e+00 -1.06122792e-01 4.46181297e-01
1.04215539e+00 -1.52883625e+00 2.76735187e-01 -6.28450274e-01
-7.24003792e-01 8.46861541e-01 1.17198646e+00 1.78789616e-01
7.85623789e-01 2.83639371e-01 1.78133309e-01 -1.75991893e-01
-1.74701467e-01 -2.70948082e-01 1.56246182e-02 8.11951101e-01
4.77407545e-01 1.24833450e-01 -5.68908095e-01 6.03158176e-01
-8.27466071e-01 1.38016731e-01 4.48091805e-01 1.16993153e+00
-6.20472133e-01 -8.69779587e-01 -1.01919150e+00 2.99257368e-01
-3.55157405e-01 4.54011291e-01 4.53572683e-02 1.01991022e+00
1.13183275e-01 1.06171024e+00 -1.43505454e-01 -6.32322133e-01
8.18261564e-01 -2.28861809e-01 5.08128405e-01 -3.80135030e-01
1.52658764e-03 3.12219411e-01 -9.98349786e-02 -4.88764346e-01
-3.29886734e-01 -8.14330935e-01 -1.13795590e+00 -4.85089332e-01
-6.65851653e-01 2.17129216e-01 8.00357044e-01 7.63962150e-01
2.01127946e-01 2.62197945e-02 8.38538766e-01 -1.51670098e+00
-4.99953210e-01 -1.49506807e+00 -9.13940251e-01 4.63097751e-01
2.12697461e-01 -1.07015872e+00 -2.96068072e-01 -2.61881411e-01] | [13.801108360290527, 0.2518424689769745] |
606e20ef-5177-4a3d-82cd-042f17b8cc9e | fully-unsupervised-training-of-few-shot | 2210.02732 | null | https://arxiv.org/abs/2210.02732v2 | https://arxiv.org/pdf/2210.02732v2.pdf | Fully Unsupervised Training of Few-shot Keyword Spotting | For training a few-shot keyword spotting (FS-KWS) model, a large labeled dataset containing massive target keywords has known to be essential to generalize to arbitrary target keywords with only a few enrollment samples. To alleviate the expensive data collection with labeling, in this paper, we propose a novel FS-KWS system trained only on synthetic data. The proposed system is based on metric learning enabling target keywords to be detected using distance metrics. Exploiting the speech synthesis model that generates speech with pseudo phonemes instead of texts, we easily obtain a large collection of multi-view samples with the same semantics. These samples are sufficient for training, considering metric learning does not intrinsically necessitate labeled data. All of the components in our framework do not require any supervision, making our method unsupervised. Experimental results on real datasets show our proposed method is competitive even without any labeled and real datasets. | ['Nam Soo Kim', 'Min Hyun Han', 'Sung Hwan Mun', 'Minchan Kim', 'Dongjune Lee'] | 2022-10-06 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 3.47197235e-01 1.50550306e-01 -2.49288812e-01 -5.17263234e-01
-1.16064608e+00 -5.94378412e-01 7.16292560e-01 -2.61845943e-02
-4.23914462e-01 7.80327082e-01 -7.70868585e-02 -1.84335992e-01
5.98506033e-02 -5.75788498e-01 -7.29740202e-01 -4.80185181e-01
5.24412692e-01 4.84228939e-01 2.50186056e-01 -1.34894669e-01
1.11627303e-01 1.09569080e-01 -1.54456162e+00 1.63569853e-01
1.16870391e+00 9.57720220e-01 5.09571731e-01 7.32828021e-01
-2.77005315e-01 5.41823328e-01 -6.32690787e-01 -3.72214168e-01
2.80040234e-01 -5.93029976e-01 -4.56124991e-01 3.93528193e-01
4.02837813e-01 -3.74794215e-01 -3.04623753e-01 9.89998639e-01
5.95389783e-01 1.62343606e-01 6.63514018e-01 -1.21570182e+00
-6.55643940e-01 1.01020706e+00 -1.34530053e-01 -1.99566171e-01
2.84887195e-01 -1.71544868e-02 1.08541977e+00 -1.46756136e+00
4.90852952e-01 9.91220713e-01 3.93254668e-01 5.63919902e-01
-1.08348513e+00 -6.06312931e-01 1.05074048e-01 3.79146188e-01
-1.76530683e+00 -5.95302939e-01 9.39864516e-01 -5.63161038e-02
4.23899829e-01 3.55184734e-01 3.68837774e-01 1.33901322e+00
-7.63591826e-01 1.00897861e+00 1.11445236e+00 -5.87278187e-01
3.73536438e-01 7.62428641e-01 -7.18914438e-03 6.38393104e-01
1.58863395e-01 -2.26549685e-01 -6.65226698e-01 -8.50418061e-02
-1.56139927e-02 9.32240784e-02 -5.00604689e-01 -4.81448323e-01
-1.48716927e+00 7.16397762e-01 -1.55016810e-01 1.87867492e-01
9.88558680e-02 -2.75940359e-01 3.20510685e-01 4.55544114e-01
4.46517467e-01 2.69701749e-01 -5.69520414e-01 -2.31141642e-01
-1.30565178e+00 6.11089319e-02 7.24986196e-01 1.44181931e+00
8.90737295e-01 4.44814935e-02 -6.64842799e-02 1.02307940e+00
-4.43066321e-02 9.09656286e-01 9.07145977e-01 -5.25638223e-01
6.98323131e-01 6.48472130e-01 2.63687104e-01 -8.07131290e-01
-1.69531889e-02 -2.75774121e-01 -6.87639594e-01 -4.57729518e-01
3.26675445e-01 -8.48928168e-02 -8.41786444e-01 1.57889998e+00
3.67115915e-01 4.86910492e-01 3.33430499e-01 5.82359493e-01
6.06478989e-01 8.09327960e-01 -4.27587748e-01 -4.05205429e-01
1.01525283e+00 -1.09185374e+00 -6.92260504e-01 4.19469886e-02
9.73743200e-01 -7.57123053e-01 1.85181594e+00 3.92456651e-01
-6.43683672e-01 -4.20361042e-01 -1.27766180e+00 2.25534379e-01
-6.00377858e-01 6.22775316e-01 2.15575352e-01 8.93647790e-01
-6.70849979e-01 4.29949641e-01 -4.10145491e-01 -4.19022858e-01
1.56885490e-01 8.56753513e-02 -2.34307155e-01 -1.82900563e-01
-1.21648824e+00 4.47662234e-01 7.65924752e-01 -5.92116304e-02
-1.16964757e+00 -7.12406158e-01 -7.55495310e-01 1.39649421e-01
7.71375477e-01 -1.62745878e-01 1.41843867e+00 -5.71497619e-01
-1.57782245e+00 4.66642559e-01 -8.10594633e-02 -3.20668995e-01
6.21566832e-01 8.18334287e-04 -6.30657911e-01 3.47257555e-01
1.29107296e-01 3.55424136e-01 1.04877043e+00 -1.25895476e+00
-6.44272029e-01 -2.22846836e-01 -2.62831271e-01 2.79867142e-01
-1.17472184e+00 -3.26335877e-01 -5.09667456e-01 -7.93909669e-01
-6.63520843e-02 -7.88710833e-01 -9.71926190e-03 -3.32047641e-01
-7.56759465e-01 -2.58553803e-01 8.29642296e-01 -4.06744719e-01
1.24227667e+00 -1.96328390e+00 -2.10226268e-01 2.18426824e-01
-9.94828269e-02 3.61001432e-01 -1.49297073e-01 5.47047138e-01
2.79048800e-01 3.34340706e-02 -4.13025171e-01 -4.72734421e-01
8.96832868e-02 -2.30377223e-02 -5.16870975e-01 2.94237584e-01
-1.42908156e-01 7.55754650e-01 -9.67838585e-01 -6.46708786e-01
1.24955840e-01 -1.54941469e-01 -3.51661980e-01 4.30816054e-01
-4.44464952e-01 2.55233794e-01 -4.28800344e-01 5.86900890e-01
5.49366534e-01 -2.00571761e-01 -1.80577729e-02 2.98397453e-03
2.38661170e-01 1.38329089e-01 -1.32703424e+00 2.06727004e+00
-8.21805656e-01 8.21885765e-02 -3.72912705e-01 -1.39123976e+00
1.01354265e+00 4.33976114e-01 2.82761961e-01 -3.83656442e-01
-8.73559620e-03 5.23417354e-01 -2.76291847e-01 -4.84560877e-01
3.64058733e-01 1.60785928e-01 -2.83621341e-01 6.67518198e-01
4.39496487e-01 -1.62088498e-01 4.56718802e-01 4.60391879e-01
8.73135507e-01 -1.46560147e-01 1.82612054e-02 -1.11466870e-01
8.09548378e-01 2.77310498e-02 5.04203439e-01 9.10426795e-01
1.09542757e-01 6.29328012e-01 1.46028042e-01 5.07275984e-02
-1.15127492e+00 -1.09164953e+00 1.54362544e-01 8.47826660e-01
2.07251430e-01 -5.05332410e-01 -1.00392532e+00 -9.52960074e-01
-4.06102806e-01 8.43110502e-01 -3.43062550e-01 -2.97121674e-01
-1.90448329e-01 -4.43832308e-01 7.31230140e-01 1.11629710e-01
2.12101415e-01 -9.24100101e-01 -3.07242751e-01 1.28330186e-01
-2.63633519e-01 -1.58082187e+00 -8.66983593e-01 1.12990132e-02
-6.27923071e-01 -1.06057823e+00 -9.58315074e-01 -8.32141638e-01
8.78587723e-01 4.39609885e-01 6.43220425e-01 -3.76342654e-01
-3.05358022e-01 3.37509662e-01 -6.70921862e-01 -2.62183845e-01
-6.65367484e-01 3.17078441e-01 3.57899755e-01 4.70224530e-01
3.18636864e-01 -6.34220660e-01 -5.57081521e-01 2.25099578e-01
-8.43929529e-01 1.60332024e-01 6.92607701e-01 1.05434608e+00
3.41003299e-01 -1.09696291e-01 9.25552428e-01 -1.03627622e+00
6.81170166e-01 -3.69412512e-01 -4.91827130e-01 6.81649327e-01
-9.24736977e-01 -4.28486243e-02 1.23540175e+00 -8.96959126e-01
-9.90014255e-01 -5.03584258e-02 3.21453184e-01 -5.71264923e-01
-1.31436214e-01 2.92117268e-01 -5.74575603e-01 1.50278032e-01
6.04500055e-01 8.90676796e-01 -1.37512982e-01 -6.15394652e-01
7.74774551e-01 1.14315534e+00 4.57332343e-01 -4.34591144e-01
9.92315054e-01 3.89591247e-01 -3.96829933e-01 -8.75206411e-01
-8.20051908e-01 -5.97020805e-01 -5.37834585e-01 -7.37703219e-02
2.70424604e-01 -9.14356351e-01 -5.48101604e-01 2.89392799e-01
-8.71936679e-01 4.51507084e-02 -4.27478194e-01 6.78109050e-01
-5.37468255e-01 4.80856061e-01 -2.61611402e-01 -8.44156146e-01
-3.86212468e-01 -1.02238786e+00 9.85485315e-01 9.34522748e-02
-2.49824710e-02 -7.43486881e-01 -1.17117763e-01 3.69875908e-01
1.52722046e-01 -2.17907086e-01 7.60796368e-01 -1.41120875e+00
-4.75780517e-01 -3.77618730e-01 7.73622543e-02 5.64437628e-01
4.58194554e-01 -3.03364575e-01 -1.02765512e+00 -1.60262093e-01
2.16824949e-01 -7.79297113e-01 5.97584128e-01 -2.58864224e-01
1.54621887e+00 -4.40601319e-01 -1.75577760e-01 4.74263877e-01
1.07011688e+00 1.94698289e-01 -4.18738686e-02 -1.47245809e-01
9.09683466e-01 6.10944152e-01 1.06315005e+00 6.39575720e-01
2.65331298e-01 4.83093083e-01 -5.13858423e-02 1.93733796e-01
1.30597930e-02 -7.49649048e-01 4.49356437e-01 1.35161543e+00
7.10935295e-01 -4.14367855e-01 -8.37978721e-01 8.80671978e-01
-1.67636120e+00 -7.09506273e-01 3.80728215e-01 2.36587048e+00
1.21813738e+00 8.69875997e-02 -5.95121235e-02 4.43040401e-01
8.11991632e-01 8.60807002e-02 -6.67989910e-01 -2.07063835e-02
6.49098083e-02 2.05121323e-01 4.79578763e-01 3.02151114e-01
-7.94103444e-01 1.09462512e+00 5.05592346e+00 1.43891203e+00
-1.00537097e+00 1.51632115e-01 6.18252099e-01 -1.54343888e-01
-7.29811192e-01 2.39343345e-02 -9.52195942e-01 7.16218948e-01
1.13090587e+00 -4.88344520e-01 4.38095868e-01 9.92270827e-01
3.58578086e-01 8.75883773e-02 -1.11588061e+00 1.22606277e+00
3.24004024e-01 -1.06949663e+00 2.52432734e-01 -2.92160273e-01
6.12756252e-01 -3.66496354e-01 -1.57826655e-02 3.83087188e-01
2.02885255e-01 -7.79705763e-01 6.57362521e-01 2.12066740e-01
1.10370231e+00 -8.49993765e-01 3.09874952e-01 8.30564022e-01
-8.84096444e-01 6.53898567e-02 -4.84161794e-01 5.23893058e-01
2.98758578e-02 7.11416721e-01 -1.46550405e+00 5.27427137e-01
3.30332249e-01 5.18017411e-01 -5.89183569e-01 6.28832281e-01
-3.90925258e-02 6.90410733e-01 -3.97903025e-01 -4.13392842e-01
1.79803252e-01 -3.48461866e-01 4.25109386e-01 1.06321871e+00
6.71214759e-01 -1.51224777e-01 3.72764587e-01 6.42540574e-01
-3.42710227e-01 6.46603107e-01 -8.45210016e-01 -4.18934762e-01
8.74538779e-01 1.36113846e+00 -7.44764328e-01 -6.23396814e-01
-5.40616453e-01 1.28897083e+00 1.08836442e-01 3.90830517e-01
-7.57538557e-01 -6.51007891e-01 2.86187708e-01 -1.22963853e-01
2.94187993e-01 -4.19975333e-02 -1.62258655e-01 -1.38955414e+00
3.60009283e-01 -9.36894476e-01 7.33871758e-02 -4.02142644e-01
-1.09500241e+00 3.38819325e-01 -1.39435068e-01 -1.60595751e+00
-3.39109242e-01 -2.49705151e-01 -3.20351273e-01 6.44470453e-01
-1.56718051e+00 -1.26459408e+00 -2.47782201e-01 8.56550217e-01
9.19177949e-01 -3.89850646e-01 8.24104548e-01 3.61311913e-01
-3.89186144e-01 1.04884648e+00 2.45473772e-01 4.13851887e-02
9.89722967e-01 -1.23148251e+00 3.90336275e-01 9.73359168e-01
6.40637577e-01 5.52820504e-01 4.60205674e-01 -5.10100126e-01
-1.38923013e+00 -1.18079042e+00 8.30628276e-01 -1.05895042e-01
7.56130934e-01 -8.72472286e-01 -7.04218209e-01 2.63916731e-01
-2.93471385e-02 1.41076893e-01 8.71670008e-01 -2.46168911e-01
-3.55898738e-01 -3.79000455e-01 -1.03839684e+00 7.02586114e-01
1.19580245e+00 -7.29788959e-01 -4.20790553e-01 6.13901377e-01
1.10496473e+00 -9.82545987e-02 -5.66488862e-01 4.77699548e-01
1.74191400e-01 -6.39424086e-01 6.77845836e-01 -5.46513677e-01
1.92199796e-02 -3.95731807e-01 -2.39278108e-01 -1.36129630e+00
4.89599168e-01 -6.12971783e-01 -3.65401864e-01 1.55647564e+00
6.72829390e-01 -3.62483025e-01 8.96641135e-01 3.35329115e-01
-1.85229704e-01 -6.71834588e-01 -8.28774929e-01 -1.11342478e+00
-4.81890380e-01 -5.05542815e-01 5.79677761e-01 1.07165670e+00
1.47381471e-02 3.77956688e-01 -6.43805623e-01 1.88497692e-01
6.52025878e-01 3.82790565e-01 8.96082819e-01 -9.92584229e-01
-3.06762338e-01 7.50854760e-02 -1.29265115e-01 -1.12979424e+00
2.71677732e-01 -1.10839927e+00 5.62476777e-02 -1.13030314e+00
1.14990570e-01 -6.32465601e-01 -2.47287706e-01 2.69642144e-01
-3.37325364e-01 1.72015559e-02 -1.20936714e-01 4.07397933e-02
-5.99290490e-01 1.01992261e+00 1.00726759e+00 -3.33798110e-01
-1.16305940e-01 2.01507315e-01 -6.00183010e-01 6.41579807e-01
6.97813630e-01 -5.90606272e-01 -1.10123181e+00 -2.37819348e-02
-1.99418664e-01 -4.37936150e-02 -1.93303436e-01 -8.96778464e-01
2.91967660e-01 -2.22435728e-01 1.85627509e-02 -4.81599659e-01
3.69929403e-01 -9.65886176e-01 -3.59001428e-01 1.77866131e-01
-5.39016545e-01 -2.50907868e-01 -3.79062444e-01 8.37134719e-01
-1.94629192e-01 -5.29913545e-01 7.14120328e-01 -8.07242244e-02
-7.54625320e-01 4.36072946e-01 1.44830644e-02 4.06100452e-01
1.26426768e+00 -3.74294892e-02 1.56446189e-01 -4.57002372e-01
-7.50649273e-01 2.76977748e-01 4.20998305e-01 6.68809712e-01
7.19045401e-01 -1.39077997e+00 -4.68826205e-01 7.25249797e-02
6.35037720e-01 1.25087246e-01 9.03536305e-02 4.47759777e-01
-1.82493269e-01 4.74726140e-01 5.42094290e-01 -5.50529957e-01
-1.04266822e+00 8.94973218e-01 -9.76672247e-02 -1.11432925e-01
-4.79483992e-01 7.38637447e-01 -1.05687343e-01 -7.67790735e-01
4.12564307e-01 -2.97319770e-01 -4.09506485e-02 2.05552474e-01
5.53534448e-01 2.50242651e-01 2.12420449e-01 -2.66771346e-01
-2.46417113e-02 2.89517730e-01 -2.62947351e-01 -4.22626376e-01
1.08446574e+00 -3.15103114e-01 3.54428530e-01 8.65643442e-01
1.38025773e+00 1.00468293e-01 -9.31702375e-01 -6.33417726e-01
3.33314836e-01 -3.96428108e-01 -3.21067840e-01 -5.24535537e-01
-7.14048326e-01 9.51588154e-01 4.78477865e-01 -1.29417768e-02
1.08482838e+00 -1.75115839e-01 1.27533233e+00 7.53420472e-01
5.11462033e-01 -1.46951997e+00 3.11157346e-01 1.17771581e-01
4.91893113e-01 -1.41186965e+00 -4.27844346e-01 -4.11994368e-01
-6.62643492e-01 9.78574514e-01 5.68482161e-01 2.75913417e-01
5.59236407e-01 6.29007518e-02 7.44430954e-03 2.12829500e-01
-6.85079634e-01 -2.09367007e-01 6.01034006e-03 4.16113347e-01
-1.44770831e-01 -8.84315073e-02 -3.54969859e-01 6.63806558e-01
-2.27339804e-01 -1.30011097e-01 6.76714540e-01 6.48696184e-01
-5.25786817e-01 -1.38095021e+00 2.01620888e-02 5.11396289e-01
-2.91898996e-01 -3.72036934e-01 -3.28709632e-01 3.72665972e-01
-1.96389064e-01 1.03224421e+00 -3.44294935e-01 -4.72877771e-01
2.50073880e-01 5.25555551e-01 1.87888652e-01 -8.65274489e-01
3.32272090e-02 -8.31348225e-02 -1.09228238e-01 -2.26880044e-01
-1.59382269e-01 -4.64046955e-01 -1.12895596e+00 1.08793043e-01
-4.88424927e-01 5.73861659e-01 6.36358082e-01 9.24711525e-01
3.32135499e-01 2.52376080e-01 1.36471641e+00 -2.44375721e-01
-1.08770895e+00 -1.27916098e+00 -7.32871175e-01 4.62681502e-01
1.23690270e-01 -3.58509958e-01 -4.57269818e-01 3.36646318e-01] | [14.096206665039062, 6.421380996704102] |
e63963e8-2565-4aca-adb6-03c74b697c1e | learning-to-prune-instances-of-steiner-tree | 2208.11985 | null | https://arxiv.org/abs/2208.11985v2 | https://arxiv.org/pdf/2208.11985v2.pdf | Learning to Prune Instances of Steiner Tree Problem in Graphs | We consider the Steiner tree problem on graphs where we are given a set of nodes and the goal is to find a tree sub-graph of minimum weight that contains all nodes in the given set, potentially including additional nodes. This is a classical NP-hard combinatorial optimisation problem. In recent years, a machine learning framework called learning-to-prune has been successfully used for solving a diverse range of combinatorial optimisation problems. In this paper, we use this learning framework on the Steiner tree problem and show that even on this problem, the learning-to-prune framework results in computing near-optimal solutions at a fraction of the time required by commercial ILP solvers. Our results underscore the potential of the learning-to-prune framework in solving various combinatorial optimisation problems. | ['Deepak Ajwani', 'Jiwei Zhang'] | 2022-08-25 | null | null | null | null | ['steiner-tree-problem'] | ['graphs'] | [ 6.88212454e-01 7.83915579e-01 -6.07765675e-01 -1.21217221e-01
-6.45603836e-01 -6.49144292e-01 1.11141624e-02 4.21213746e-01
-3.72125232e-03 7.56517828e-01 -5.06195188e-01 -5.45728326e-01
-7.15000868e-01 -1.00988090e+00 -6.25692010e-01 -5.85195601e-01
-7.16702461e-01 9.72511411e-01 2.11242199e-01 -4.21544164e-02
9.77474749e-02 4.09400642e-01 -1.44048560e+00 1.58770502e-01
7.47746348e-01 9.91752326e-01 1.05183735e-01 6.51092172e-01
-7.60388076e-02 2.29303196e-01 -2.69274622e-01 -5.21877229e-01
7.12429285e-01 7.09042996e-02 -1.23298216e+00 3.79120439e-01
2.34807998e-01 5.25724947e-01 -4.35898714e-02 9.75406885e-01
2.06913143e-01 -1.88993290e-01 4.38185722e-01 -1.67531264e+00
6.60155490e-02 6.93034112e-01 -8.40410650e-01 1.88876837e-01
4.51592445e-01 -4.17330489e-02 1.72627378e+00 -2.37935990e-01
4.67665851e-01 1.25162780e+00 3.50964248e-01 1.05986200e-01
-1.47969282e+00 -3.01111042e-01 2.83619434e-01 3.70099455e-01
-1.27452767e+00 -1.00106457e-02 5.32488763e-01 -9.00124107e-03
1.21276677e+00 7.15189397e-01 6.98304594e-01 2.40814745e-01
4.18799281e-01 8.91668022e-01 1.05604804e+00 -8.15421939e-01
3.34429532e-01 -2.67850459e-01 -8.63528550e-02 8.86060297e-01
2.30721638e-01 3.23987007e-01 -1.15580887e-01 -1.17007449e-01
2.40350574e-01 -2.37036690e-01 6.50226995e-02 -5.84611893e-01
-7.59475052e-01 1.22116995e+00 5.46382844e-01 7.21728504e-02
-1.77670829e-02 1.61958799e-01 1.89724579e-01 5.01216114e-01
4.12710130e-01 9.68050599e-01 -9.49441671e-01 2.40842387e-01
-7.36401916e-01 1.37510344e-01 1.26704097e+00 9.27797556e-01
8.47369730e-01 -3.52499872e-01 5.51090725e-02 6.25684023e-01
1.70357257e-01 5.73631190e-02 -3.55200887e-01 -1.04757714e+00
5.67805409e-01 8.12954664e-01 -2.88987309e-01 -8.28459144e-01
-6.92821801e-01 -6.23767614e-01 -6.36288404e-01 3.58923793e-01
2.04046220e-01 -2.56864876e-01 -9.83414829e-01 1.47998631e+00
5.51501215e-01 1.13322802e-01 -1.76321059e-01 5.51217377e-01
2.88732350e-01 8.63545835e-01 -2.03743964e-01 -5.97691596e-01
9.09369349e-01 -1.38855612e+00 -1.69137836e-01 -5.92354894e-01
8.87058914e-01 -5.75041056e-01 3.57955307e-01 7.31934547e-01
-1.07050920e+00 1.13890983e-01 -1.02002156e+00 2.88186699e-01
-2.62065887e-01 -5.75966895e-01 1.08400750e+00 7.94507205e-01
-1.09020925e+00 8.30993056e-01 -5.93621671e-01 -3.05263221e-01
4.46280092e-01 8.67659926e-01 -2.48160064e-01 -5.88359296e-01
-5.98745704e-01 1.05205810e+00 7.98123181e-01 -1.32174551e-01
-5.37184000e-01 -5.02249897e-01 -8.58603001e-01 1.55171216e-01
1.13139474e+00 -7.61838317e-01 1.22051096e+00 -8.98468912e-01
-1.01420915e+00 8.07709992e-01 -1.18713766e-01 -4.22163188e-01
1.06986001e-01 3.66631597e-01 -1.80297434e-01 5.45082195e-03
-3.83448601e-02 5.22616804e-01 6.32942796e-01 -1.15470397e+00
-8.97467494e-01 -4.46733594e-01 4.47068900e-01 9.76995304e-02
8.13541189e-02 -8.59366804e-02 -5.22087574e-01 -2.81701028e-01
2.98024952e-01 -1.12112629e+00 -9.54054534e-01 -2.82068729e-01
-5.46836913e-01 -4.95429456e-01 5.70587635e-01 -1.66630298e-01
1.25815070e+00 -1.59607422e+00 6.50566041e-01 7.08156943e-01
4.95058000e-01 5.66071607e-02 -5.95015824e-01 6.32661045e-01
-2.71005601e-01 2.98844308e-01 -2.76158631e-01 -1.38992980e-01
1.98740140e-03 8.52770746e-01 3.10266823e-01 3.97957534e-01
3.63803297e-01 9.02746558e-01 -9.64647889e-01 -5.62645555e-01
1.13231517e-01 -3.68362904e-01 -6.96540058e-01 -1.69184119e-01
-6.56608105e-01 -4.65065008e-03 -4.11019713e-01 7.75836647e-01
6.33694291e-01 -2.40004912e-01 5.33619225e-01 3.32963973e-01
1.32887319e-01 1.39781028e-01 -1.25942481e+00 1.34012151e+00
-4.70064044e-01 2.62520939e-01 2.98562050e-01 -1.40055335e+00
6.06567979e-01 -1.17086485e-01 7.47608900e-01 -7.10892856e-01
1.75438389e-01 2.34563753e-01 2.85961151e-01 -3.66403162e-01
-5.51567450e-02 -3.55616689e-01 -1.85104385e-01 4.81686950e-01
-6.44083694e-02 -3.52937728e-01 6.61839843e-01 1.88393723e-02
1.62308323e+00 -3.54728878e-01 3.45676184e-01 -2.30719894e-01
4.07344580e-01 1.59510821e-01 7.57397711e-01 6.84904754e-01
3.76778215e-01 2.25368351e-01 9.46972013e-01 -5.52568257e-01
-9.67785001e-01 -6.91546500e-01 -3.88193354e-02 1.00603664e+00
-2.61312071e-02 -5.72858274e-01 -4.34840471e-01 -7.68023074e-01
8.48757252e-02 6.03082955e-01 -3.48977864e-01 -4.35010754e-02
-3.95793736e-01 -6.72493398e-01 -1.40992880e-01 7.39894733e-02
-3.14317763e-01 -7.85756052e-01 -4.78632569e-01 4.20495033e-01
5.72176203e-02 -1.06977725e+00 -2.29474276e-01 7.60779262e-01
-1.20642567e+00 -1.32438290e+00 -1.10186227e-01 -1.01327646e+00
7.60571837e-01 1.07255213e-01 1.34255052e+00 3.91062260e-01
-6.23645663e-01 1.03128165e-01 -4.50607151e-01 -4.51822460e-01
-2.44820386e-01 6.21051729e-01 -3.88979912e-01 -3.08419734e-01
8.73040967e-03 -6.46522880e-01 6.31265342e-02 3.84990692e-01
-8.30675244e-01 -8.44470505e-03 8.34978878e-01 6.53149962e-01
7.87019491e-01 9.82223988e-01 4.11699653e-01 -1.00556874e+00
3.63042235e-01 -7.45580733e-01 -9.35693681e-01 4.62714285e-01
-8.50552499e-01 4.47874814e-01 5.63031137e-01 -1.55691028e-01
5.16771758e-03 3.93501908e-01 1.37905657e-01 -3.17875624e-01
9.88289043e-02 8.34630072e-01 -3.83935273e-01 -5.62930405e-01
8.52558389e-02 -1.46490127e-01 -1.47670969e-01 -1.83020294e-01
1.61779627e-01 1.39622480e-01 1.47928447e-01 -4.97419775e-01
1.07286406e+00 -8.61928612e-02 9.02518153e-01 -4.37847078e-01
-1.21437562e+00 -6.07792795e-01 -6.17036819e-01 -3.56684290e-02
2.81713039e-01 -3.00453871e-01 -6.63081348e-01 -2.14358419e-01
-8.99130702e-01 -1.80395707e-01 -3.18583131e-01 -1.50077403e-01
-6.99168324e-01 4.30390328e-01 5.78393638e-02 -8.84374678e-01
-1.69561729e-01 -1.00771165e+00 9.13447678e-01 1.33668125e-01
-8.12113658e-02 -1.24503267e+00 1.45394504e-01 5.02606392e-01
-1.17717743e-01 7.10724413e-01 1.37328875e+00 -6.58683598e-01
-7.92948008e-01 -2.95488596e-01 -1.03941292e-01 -5.49332537e-02
1.51217595e-01 -1.88923433e-01 -3.36327374e-01 -5.52833200e-01
-2.61047423e-01 -2.97195822e-01 8.14998388e-01 3.71642709e-01
1.31302798e+00 -8.21264148e-01 -6.43799007e-01 5.62013149e-01
1.82771242e+00 5.74899763e-02 4.63456303e-01 3.05539280e-01
3.68360579e-01 7.95331717e-01 7.45321631e-01 3.61346990e-01
3.37025374e-01 5.64840317e-01 1.00985968e+00 -5.70499338e-02
4.28489685e-01 -6.19260632e-02 -2.08829641e-01 3.02595824e-01
9.06857699e-02 -5.67848206e-01 -1.08540225e+00 6.75190985e-01
-1.96947622e+00 -6.89538181e-01 -2.46313706e-01 2.17445374e+00
8.82160068e-01 5.20982981e-01 2.51079738e-01 6.08648062e-01
5.76853573e-01 1.21049009e-01 -4.58795488e-01 -1.03960395e+00
6.27845973e-02 7.03820467e-01 7.65433908e-01 6.80854142e-01
-1.05577159e+00 8.61902058e-01 7.17517567e+00 9.10296261e-01
-6.45565569e-01 -2.10745052e-01 4.99796033e-01 -2.16080055e-01
-3.25272113e-01 2.94435084e-01 -5.19732773e-01 1.84232831e-01
9.38836694e-01 -3.02550435e-01 8.88869464e-01 7.75315702e-01
-4.54322249e-02 -2.71202028e-01 -1.41330111e+00 6.41990483e-01
-1.04091123e-01 -1.23227489e+00 -3.09693933e-01 4.58207726e-01
1.05791020e+00 -8.23307782e-02 -1.32381581e-02 1.41196877e-01
6.52658224e-01 -1.46721184e+00 1.70967877e-01 -3.88490483e-02
6.18104279e-01 -1.16329467e+00 6.47381485e-01 6.32468998e-01
-1.13952458e+00 -5.26648223e-01 -3.35121334e-01 -2.43186474e-01
2.39511654e-01 9.20955122e-01 -1.11883819e+00 8.53700459e-01
5.83702147e-01 4.93274242e-01 -4.35342461e-01 1.72295773e+00
-3.02851766e-01 4.53460604e-01 -6.66984022e-01 4.64677103e-02
4.58371997e-01 -1.67364776e-01 6.50540352e-01 9.15054917e-01
-2.74261646e-02 1.24533810e-01 6.43096447e-01 4.39191878e-01
3.54679627e-03 -2.08615698e-02 -6.27638400e-01 -8.15266892e-02
2.17915356e-01 1.47330713e+00 -8.94686460e-01 2.58601815e-01
-1.53254732e-01 4.05201793e-01 5.94689608e-01 -1.01781590e-02
-5.77118635e-01 -3.36381704e-01 2.84719527e-01 8.00480023e-02
6.14719927e-01 -2.95949370e-01 -4.18374747e-01 -5.12538671e-01
1.15786307e-02 -8.32712710e-01 7.39697278e-01 -4.38125789e-01
-1.24124861e+00 2.24906251e-01 6.60772994e-02 -8.26404214e-01
-1.02023389e-02 -7.76112556e-01 -7.62452602e-01 5.09426653e-01
-1.52139819e+00 -8.86990905e-01 1.13268137e-01 2.94654310e-01
5.93890011e-01 7.18791261e-02 6.82941258e-01 -1.29761726e-01
-5.61201453e-01 3.79370809e-01 1.41711980e-01 -4.74707842e-01
-9.49967206e-02 -1.53967381e+00 1.93634078e-01 6.41464591e-01
4.03225631e-01 3.75659503e-02 7.92068064e-01 -4.29185510e-01
-1.75203812e+00 -1.03401768e+00 9.82088327e-01 -2.89353877e-01
6.77358091e-01 -1.44172251e-01 -4.54874843e-01 6.61423624e-01
1.41412809e-01 -1.10860139e-01 6.25029445e-01 4.83449250e-01
-1.79467544e-01 -1.04623869e-01 -1.26598179e+00 3.21358562e-01
1.15769660e+00 2.43650123e-01 -5.19254744e-01 7.67187953e-01
7.14045823e-01 -4.13148433e-01 -8.18750978e-01 5.83180070e-01
8.22335705e-02 -5.68036616e-01 8.57375801e-01 -8.22373688e-01
5.47898591e-01 -6.44494593e-02 -1.07357884e-02 -1.47954953e+00
-5.40498614e-01 -9.55574989e-01 -3.32052618e-01 8.61091137e-01
7.20679581e-01 -5.26402414e-01 1.03980899e+00 4.16066796e-01
-7.57154897e-02 -1.34519422e+00 -1.39483702e+00 -1.00694239e+00
5.81988320e-02 -3.83938640e-01 5.42249560e-01 6.34396493e-01
8.87245834e-02 5.49709976e-01 -2.04020411e-01 2.72651792e-01
9.03193116e-01 4.87470895e-01 4.95831549e-01 -1.61490202e+00
-6.07007682e-01 -4.24893260e-01 -3.40392739e-01 -6.33318782e-01
3.74558836e-01 -1.39842987e+00 -1.17555872e-01 -1.83573425e+00
1.80243134e-01 -9.31191087e-01 -7.01353103e-02 8.98711205e-01
1.28367171e-02 -1.16672203e-01 8.12778249e-02 -2.31251046e-01
-9.35787737e-01 1.55970231e-01 1.37973368e+00 -4.28809941e-01
-2.79756021e-02 5.07909238e-01 -7.85685599e-01 5.57180345e-01
8.28577459e-01 -9.46409404e-01 -2.26150706e-01 -3.27777386e-01
6.20999277e-01 2.08341956e-01 -4.33399752e-02 -6.27131879e-01
1.81232408e-01 -6.54500127e-01 -1.45118043e-01 -5.40285230e-01
9.12542269e-02 -1.19286203e+00 1.50549352e-01 8.72902811e-01
-1.07318610e-01 1.26990482e-01 2.20467001e-01 6.26476526e-01
8.28630626e-02 -6.69777095e-01 7.45475769e-01 -7.85815418e-02
-6.31894350e-01 4.34985965e-01 -1.70268849e-01 1.73910052e-01
1.55169201e+00 -3.44127834e-01 -1.20136529e-01 -1.51229665e-01
-7.54167080e-01 9.50522304e-01 2.95043647e-01 2.15117589e-01
4.58934069e-01 -9.83370900e-01 -6.85737014e-01 -5.23047745e-02
1.08150214e-01 3.36934716e-01 -1.74671084e-01 6.41181946e-01
-4.00065422e-01 5.74579060e-01 -8.74636918e-02 -5.01947880e-01
-1.45920229e+00 9.21177924e-01 1.99139789e-01 -8.52583289e-01
-3.83069485e-01 8.91699791e-01 -3.65376085e-01 -5.40034831e-01
2.12356329e-01 -1.35036469e-01 7.03655854e-02 -1.47515044e-01
1.56680718e-02 5.87293983e-01 1.40907571e-01 -1.64818227e-01
-4.39425409e-01 6.08053327e-01 -9.49109048e-02 1.95144683e-01
1.62748587e+00 1.79709613e-01 -4.80873495e-01 -7.99652562e-02
1.05205894e+00 -1.64488301e-01 -8.09018195e-01 -3.23897690e-01
5.08210659e-01 -4.70263720e-01 4.36465256e-02 -8.68569553e-01
-1.38260043e+00 3.72432232e-01 5.52662462e-02 5.57249069e-01
1.38830590e+00 3.77364576e-01 5.89566231e-01 5.85446775e-01
8.55033576e-01 -1.05762410e+00 -8.08444768e-02 6.25253916e-01
6.59612119e-01 -1.08650529e+00 3.23822379e-01 -9.40966070e-01
-3.73476595e-01 1.20265663e+00 4.14868802e-01 -1.59525424e-01
4.76088703e-01 4.78307784e-01 -6.44425094e-01 -1.08194843e-01
-1.37428641e+00 -4.12388295e-01 2.83023149e-01 7.04915285e-01
-2.47728035e-01 1.96734950e-01 -2.49532536e-01 7.37322196e-02
-4.16285992e-01 -2.72465020e-01 5.38795650e-01 1.06250000e+00
-6.53306723e-01 -1.75261867e+00 -2.65209317e-01 6.65675461e-01
-3.89535725e-01 -3.80956903e-02 -6.80865467e-01 5.08035004e-01
2.56976038e-01 1.19742632e+00 -1.81324512e-01 -3.21022958e-01
1.65953174e-01 -2.42859617e-01 8.80084336e-01 -1.06108558e+00
-6.64671123e-01 -1.25844955e-01 5.21051228e-01 -6.41987383e-01
-2.54439980e-01 -4.98437226e-01 -1.06149459e+00 -3.19483280e-01
-6.43573880e-01 2.38094866e-01 5.27713716e-01 1.14875364e+00
6.27952293e-02 4.20579940e-01 1.03851247e+00 -4.98942077e-01
-6.00167096e-01 -3.59101057e-01 -6.52888894e-01 -3.58680427e-01
1.82313189e-01 -4.84611511e-01 -3.37262303e-01 -5.72874308e-01] | [5.241411209106445, 3.025007724761963] |
b4bbcec4-790a-4580-ab9e-1622539d083e | divide-and-denoise-learning-from-noisy-labels-1 | null | null | https://aclanthology.org/2022.acl-long.141 | https://aclanthology.org/2022.acl-long.141.pdf | Divide and Denoise: Learning from Noisy Labels in Fine-Grained Entity Typing with Cluster-Wise Loss Correction | Fine-grained Entity Typing (FET) has made great progress based on distant supervision but still suffers from label noise. Existing FET noise learning methods rely on prediction distributions in an instance-independent manner, which causes the problem of confirmation bias. In this work, we propose a clustering-based loss correction framework named Feature Cluster Loss Correction (FCLC), to address these two problems. FCLC first train a coarse backbone model as a feature extractor and noise estimator. Loss correction is then applied to each feature cluster, learning directly from the noisy labels. Experimental results on three public datasets show that FCLC achieves the best performance over existing competitive systems. Auxiliary experiments further demonstrate that FCLC is stable to hyperparameters and it does help mitigate confirmation bias. We also find that in the extreme case of no clean data, the FCLC framework still achieves competitive performance. | ['Ting Wang', 'Jie zhou', 'Haoyu Zhang', 'Kunyuan Pang'] | null | null | null | null | acl-2022-5 | ['entity-typing'] | ['natural-language-processing'] | [-1.56960618e-02 -2.31699586e-01 -3.08788866e-01 -6.99950576e-01
-1.31504869e+00 -2.41499797e-01 4.47302401e-01 1.99519843e-01
-6.87975109e-01 9.14040148e-01 1.38877451e-01 -7.37584308e-02
7.50092268e-02 -6.19040489e-01 -7.32122660e-01 -7.81364024e-01
2.54225284e-01 4.13040072e-01 3.36326927e-01 1.86435327e-01
-1.00977048e-02 8.25885609e-02 -1.48764932e+00 4.99105990e-01
1.13865960e+00 9.24271524e-01 -3.76977660e-02 4.24002409e-01
-1.50189206e-01 1.03915405e+00 -5.75341821e-01 -6.59871578e-01
8.96074474e-02 -1.54456526e-01 -7.84431636e-01 -5.61442137e-01
3.46421391e-01 -2.14394912e-01 -1.57324150e-01 1.16335690e+00
7.29368210e-01 2.33371779e-02 6.32444084e-01 -1.41330874e+00
-5.36202192e-01 9.03718293e-01 -5.18278539e-01 -9.04523805e-02
-2.61149388e-02 -2.05732256e-01 1.33960640e+00 -1.10688818e+00
5.47357082e-01 1.28280556e+00 1.31549120e+00 7.06724823e-01
-1.48783123e+00 -9.03920472e-01 3.93994838e-01 2.02592388e-01
-1.60904038e+00 -2.50090599e-01 4.87963438e-01 -2.04878584e-01
8.67546499e-01 1.56728432e-01 -1.91442251e-01 1.14536321e+00
-4.84194756e-02 1.12093365e+00 1.08263600e+00 -4.63130385e-01
3.17766309e-01 3.27751160e-01 4.41184461e-01 5.82623780e-01
4.62893426e-01 7.52472132e-02 -5.55122077e-01 -5.94080389e-01
1.43207431e-01 9.10520628e-02 -1.43986553e-01 -3.35981071e-01
-7.40094960e-01 7.64417470e-01 4.59688902e-01 2.36212105e-01
3.69271412e-02 2.22925976e-01 4.85328227e-01 4.78651077e-01
6.84409022e-01 1.47061214e-01 -9.23973143e-01 -1.28348455e-01
-8.07231665e-01 3.64533484e-01 9.03832078e-01 1.03847885e+00
8.60901058e-01 -5.45255363e-01 -3.79117429e-01 9.70925272e-01
2.14219496e-01 4.82296914e-01 3.59394222e-01 -8.49603593e-01
5.35395324e-01 5.23918509e-01 1.01986021e-01 -8.39605331e-01
-5.86764455e-01 -5.60510039e-01 -7.67886519e-01 -4.07159328e-02
4.62021261e-01 -2.43883938e-01 -7.99473405e-01 1.98523951e+00
4.61917758e-01 3.18738222e-01 -2.87952602e-01 5.22324145e-01
6.97291732e-01 1.37992859e-01 4.38177139e-01 -5.78409731e-02
9.77540374e-01 -8.78421485e-01 -8.11619997e-01 1.30465522e-01
1.18576384e+00 -5.08224308e-01 1.15976250e+00 3.91283363e-01
-6.77078545e-01 -1.65480211e-01 -8.91743004e-01 -1.31615415e-01
-3.13943505e-01 1.24500953e-01 5.38253784e-01 6.94171965e-01
-6.70594156e-01 6.33638740e-01 -1.04157162e+00 -2.25230247e-01
6.34923458e-01 2.25183398e-01 -4.27773237e-01 -4.16492164e-01
-1.17021430e+00 7.91895211e-01 1.58210278e-01 -7.95686841e-02
-4.13946092e-01 -1.02006781e+00 -7.75443256e-01 1.22118346e-01
4.76625890e-01 -5.63590944e-01 1.27112937e+00 -7.12456524e-01
-1.11795759e+00 6.68415308e-01 -5.12576878e-01 -4.29193109e-01
7.74712920e-01 -5.43582320e-01 -3.49260747e-01 -4.21267629e-01
3.68751615e-01 1.70070827e-01 4.32893962e-01 -1.42671657e+00
-1.09240401e+00 -3.13632190e-01 -3.79215449e-01 -2.97550149e-02
-3.34068060e-01 3.13269719e-02 -5.97566366e-01 -5.93914151e-01
1.97112113e-01 -8.17990184e-01 -2.68423229e-01 -2.86482304e-01
-5.32341242e-01 -6.21240139e-01 5.00497103e-01 -3.76138806e-01
1.60316944e+00 -2.05942345e+00 -2.81285435e-01 3.55566621e-01
2.66523331e-01 1.93353608e-01 -1.44712841e-02 1.95336699e-01
2.09771797e-01 2.05481648e-01 -2.09571227e-01 -6.56844616e-01
1.42810583e-01 1.77656695e-01 5.04001416e-02 3.86755198e-01
1.12296984e-01 8.29096854e-01 -9.63463247e-01 -5.41369975e-01
-2.77764410e-01 3.49856764e-01 -9.12826896e-01 1.98503777e-01
-1.04929626e-01 1.49333596e-01 -3.95409077e-01 7.41550148e-01
9.13182735e-01 -4.08280730e-01 3.42582732e-01 -2.20110312e-01
1.85799658e-01 4.07036364e-01 -1.26336884e+00 1.59317267e+00
-3.42318267e-01 2.38622487e-01 2.85579916e-02 -7.75175691e-01
7.41397560e-01 4.30882983e-02 3.90176326e-01 -5.22872686e-01
4.62486688e-03 3.50850463e-01 -3.57152402e-01 -3.66258085e-01
3.16382885e-01 -2.81292617e-01 -3.60271335e-01 3.84620309e-01
1.10779136e-01 6.56812131e-01 -1.52047411e-01 3.77236903e-01
1.48377061e+00 6.17242008e-02 7.21564144e-02 -1.38093963e-01
2.99764097e-01 -2.48680115e-01 1.31273913e+00 1.23959732e+00
-5.19843876e-01 7.74249017e-01 3.99941266e-01 -2.13183433e-01
-7.23833561e-01 -9.99172449e-01 -4.01043385e-01 1.40958130e+00
2.73176223e-01 -7.68513143e-01 -8.51508856e-01 -1.43721974e+00
3.99896771e-01 5.20622253e-01 -6.84558570e-01 -1.31043226e-01
-5.10813594e-01 -1.05584276e+00 7.99363613e-01 7.63388872e-01
2.54975945e-01 -8.40025902e-01 3.16050857e-01 3.36163074e-01
-4.54830199e-01 -9.48623598e-01 -4.42521513e-01 5.43204129e-01
-4.89745736e-01 -1.09558618e+00 -2.23255098e-01 -7.58457601e-01
6.14008486e-01 2.75611520e-01 1.25318813e+00 3.59270841e-01
-1.73207924e-01 -1.00626253e-01 -3.52114230e-01 -4.19376940e-01
-2.34305233e-01 4.75283295e-01 1.90459132e-01 -1.83015347e-01
9.78814125e-01 -4.54083145e-01 -4.60005254e-01 3.80639881e-01
-5.92988789e-01 -4.57067192e-01 3.64776373e-01 1.26768446e+00
7.14405656e-01 1.65718436e-01 9.29414749e-01 -1.58233452e+00
3.61749440e-01 -6.07789934e-01 -3.10212046e-01 4.66303945e-01
-1.01505888e+00 1.99763373e-01 5.42833090e-01 -4.66718711e-02
-1.33989108e+00 4.04413380e-02 -3.87043476e-01 -1.10808440e-01
-2.06523806e-01 3.02491099e-01 -6.06799364e-01 8.91668648e-02
7.54625976e-01 -2.64320344e-01 -4.26585585e-01 -9.84214962e-01
2.96387345e-01 8.66325617e-01 5.48697352e-01 -7.24338651e-01
7.30628252e-01 4.18134004e-01 -4.08213198e-01 -2.47465774e-01
-1.28331995e+00 -7.51235604e-01 -5.97097397e-01 3.80510539e-01
3.08181703e-01 -1.20219076e+00 -6.20873392e-01 5.37677526e-01
-7.74864137e-01 -1.92976370e-01 -1.89421669e-01 3.14169973e-01
-3.04806620e-01 3.22639763e-01 -9.39745784e-01 -6.69200480e-01
-3.00605118e-01 -8.03185225e-01 1.02312088e+00 7.56041557e-02
3.53765562e-02 -9.75595415e-01 1.63210273e-01 3.03318411e-01
1.91668913e-01 4.89837453e-02 7.69538105e-01 -8.47302735e-01
-4.59303886e-01 -3.59755933e-01 -3.83639991e-01 4.19976592e-01
7.03241974e-02 -2.22560748e-01 -1.18828487e+00 -3.92510384e-01
-3.28898609e-01 -4.31694120e-01 1.35011041e+00 1.34098008e-01
1.19962299e+00 -5.48327230e-02 -5.97357810e-01 7.88823247e-01
1.53752720e+00 -3.51883978e-01 4.83291715e-01 4.23143893e-01
9.79670465e-01 2.46321425e-01 8.49048495e-01 3.61795962e-01
7.57886052e-01 5.77981472e-01 3.97159085e-02 -1.01981372e-01
-2.36969516e-01 -6.28099680e-01 2.73395509e-01 7.11392403e-01
2.51153409e-01 -1.42116517e-01 -7.55530894e-01 5.03311157e-01
-2.21256113e+00 -8.80490720e-01 -2.17659846e-01 2.03450274e+00
1.24736619e+00 1.70242846e-01 2.78890487e-02 1.93080902e-01
7.68788397e-01 -3.50422293e-01 -5.83634198e-01 -8.02915916e-02
-3.39271247e-01 6.44550249e-02 5.94769597e-01 3.95320058e-01
-1.28774440e+00 1.08003962e+00 6.03236675e+00 1.23253047e+00
-6.68488443e-01 4.30863082e-01 5.67945659e-01 -2.35369836e-04
-3.32165301e-01 -8.24531987e-02 -1.25296354e+00 7.42589653e-01
7.41336703e-01 5.75688444e-02 1.37692615e-01 9.04635489e-01
-7.62809739e-02 -8.54868628e-03 -1.07653606e+00 8.26288700e-01
-2.03338802e-01 -9.72251475e-01 -2.21988901e-01 -1.74010724e-01
9.80539501e-01 3.06922019e-01 -2.35090718e-01 5.80766499e-01
8.74016225e-01 -8.39224637e-01 4.27281648e-01 4.36576068e-01
8.09379816e-01 -1.08083010e+00 1.16556394e+00 5.23335755e-01
-9.42099929e-01 -1.67442560e-01 -5.79507113e-01 1.22249231e-01
-9.08494368e-02 1.05373371e+00 -6.66240096e-01 4.90845382e-01
9.77999032e-01 6.23385787e-01 -5.43282986e-01 1.18251586e+00
-3.19137186e-01 9.38502848e-01 -3.48651707e-01 1.73477173e-01
-1.66209534e-01 3.30271333e-01 1.74783841e-01 1.59122896e+00
8.92644152e-02 -8.11530948e-02 3.17846328e-01 4.89782840e-01
-5.70559740e-01 1.29904255e-01 -2.78402120e-01 5.60036778e-01
9.79412735e-01 1.23271525e+00 -3.98815900e-01 -2.16031924e-01
-5.73725283e-01 9.11868215e-01 9.66712952e-01 2.44753018e-01
-7.11659253e-01 -5.46248317e-01 8.56626749e-01 -1.37867313e-02
4.75297153e-01 3.19665402e-01 -5.96482277e-01 -1.49751770e+00
1.81903049e-01 -7.64723182e-01 6.67800665e-01 -2.53074002e-02
-1.87752807e+00 3.41677457e-01 -5.70636809e-01 -1.05135536e+00
2.15418369e-01 -3.15194666e-01 -4.34293717e-01 6.00469410e-01
-1.79560184e+00 -1.17270935e+00 -1.96463183e-01 5.25723934e-01
9.39750001e-02 9.04959813e-02 9.07784402e-01 6.67436540e-01
-7.28976130e-01 1.60752714e+00 6.79388285e-01 2.29562581e-01
1.25779891e+00 -1.64739072e+00 2.38668010e-01 6.42286003e-01
-4.51485813e-02 5.92946172e-01 5.20107508e-01 -7.05572307e-01
-8.93607497e-01 -1.54133451e+00 1.32282674e+00 -8.39437723e-01
3.93020809e-01 -6.12009823e-01 -1.15875804e+00 6.87077820e-01
-4.11490738e-01 3.50119025e-01 9.52933908e-01 7.12963045e-01
-8.00102055e-01 -2.08810478e-01 -1.47258854e+00 1.69474527e-01
1.33924198e+00 -4.58051473e-01 -3.59895736e-01 2.94238091e-01
8.09878826e-01 -2.47058541e-01 -9.23493981e-01 4.77421194e-01
5.78841269e-01 -8.70573401e-01 7.79601693e-01 -8.53335321e-01
4.79357988e-02 -3.26219261e-01 -2.40273550e-01 -1.30006742e+00
-6.10139668e-01 -4.00483757e-01 -1.87269762e-01 1.80673075e+00
6.59055352e-01 -4.83247340e-01 8.07077408e-01 8.17211628e-01
1.38924763e-01 -5.54675102e-01 -8.58148336e-01 -9.30057287e-01
3.33552629e-01 -6.18031561e-01 7.48948097e-01 1.11022496e+00
4.68829162e-02 3.81879658e-01 -5.05009532e-01 2.05360189e-01
9.30579007e-01 -4.33738902e-02 6.96789742e-01 -1.45734882e+00
-2.57485658e-01 9.49117634e-03 -2.23708928e-01 -8.23784471e-01
3.66912961e-01 -1.05523503e+00 3.73245150e-01 -1.15196419e+00
6.51088595e-01 -1.17189860e+00 -8.03644538e-01 6.26310706e-01
-8.82798314e-01 3.22487593e-01 -1.18931029e-02 2.40077361e-01
-1.24023318e+00 4.47957546e-01 5.30381262e-01 1.14247352e-01
-9.67587009e-02 1.96720317e-01 -9.83269691e-01 7.86775947e-01
6.67097092e-01 -8.93357754e-01 -2.28214096e-02 -3.56371880e-01
1.67295232e-01 -6.10854924e-01 -3.31112929e-02 -7.14620113e-01
3.39502931e-01 7.08198845e-02 3.94282639e-01 -5.61405420e-01
-2.84548253e-01 -7.35986829e-01 -1.85416028e-01 1.30540207e-01
-5.51168680e-01 -3.58016163e-01 -2.44285956e-01 9.56451774e-01
-1.18984878e-01 -9.83071923e-02 8.38049531e-01 1.46772996e-01
-4.99780267e-01 3.25007647e-01 -6.74012527e-02 5.36986351e-01
5.81852973e-01 2.76245832e-01 -3.47815812e-01 1.33982571e-02
-7.04017878e-01 5.40445924e-01 5.15181959e-01 2.80147493e-01
1.25416934e-01 -1.61941183e+00 -8.16553414e-01 1.52024537e-01
3.62986624e-01 2.95267850e-02 1.19404495e-01 7.45311499e-01
1.17565736e-01 1.97900131e-01 4.94893223e-01 -5.46201229e-01
-1.37271416e+00 5.04540324e-01 2.18679801e-01 -6.01807177e-01
-5.88803172e-01 1.15549052e+00 2.48640656e-01 -8.83980393e-01
6.53658092e-01 -1.78763241e-01 8.64851624e-02 -1.61324352e-01
7.13993013e-01 4.94623601e-01 5.19929230e-01 -3.76971990e-01
-4.21759993e-01 2.69934803e-01 -5.10222137e-01 3.56408000e-01
1.37120533e+00 -2.71726996e-01 1.47419617e-01 3.41175944e-01
1.35315192e+00 5.35427183e-02 -1.23550820e+00 -9.88337040e-01
6.05614305e-01 -5.18684447e-01 1.03160385e-02 -1.06057596e+00
-9.11971331e-01 5.93387187e-01 5.62495053e-01 3.94471362e-02
1.04392886e+00 -5.16212732e-02 9.52035308e-01 4.99973834e-01
5.18660426e-01 -1.33586550e+00 -2.55127281e-01 6.89891338e-01
1.84835792e-01 -1.57281542e+00 -1.91215515e-01 -5.32205582e-01
-4.97708321e-01 4.81196433e-01 7.67135024e-01 7.47409463e-02
9.22711968e-01 6.08911037e-01 1.75726905e-01 8.31144005e-02
-1.04803586e+00 -2.54444540e-01 9.90680084e-02 7.10964620e-01
5.78652680e-01 1.53700382e-01 -4.49983299e-01 1.33077121e+00
8.27765241e-02 6.17664382e-02 -3.67485150e-03 8.76043022e-01
-4.35091734e-01 -1.45447087e+00 -9.92736593e-02 6.14012361e-01
-8.53755295e-01 -1.92515060e-01 -2.98582047e-01 5.47555506e-01
4.93073344e-01 9.29052234e-01 -2.54765272e-01 -5.34277141e-01
4.73830879e-01 2.27505431e-01 2.48446241e-01 -5.19813120e-01
-8.86243463e-01 -3.13533135e-02 1.85676560e-01 -7.31870472e-01
-1.36960790e-01 -7.76920736e-01 -1.41588843e+00 -4.65557843e-01
-8.03786278e-01 3.86056483e-01 4.41260248e-01 9.19226646e-01
5.59227705e-01 2.32681781e-01 8.24853301e-01 -2.12976694e-01
-8.08778107e-01 -8.79533112e-01 -7.72965729e-01 8.10246706e-01
3.95269781e-01 -6.92708611e-01 -5.47449231e-01 -2.20648572e-01] | [9.54353141784668, 8.844675064086914] |
4db446fa-b833-457f-a6fb-940718306460 | llmzip-lossless-text-compression-using-large | 2306.0405 | null | https://arxiv.org/abs/2306.04050v2 | https://arxiv.org/pdf/2306.04050v2.pdf | LLMZip: Lossless Text Compression using Large Language Models | We provide new estimates of an asymptotic upper bound on the entropy of English using the large language model LLaMA-7B as a predictor for the next token given a window of past tokens. This estimate is significantly smaller than currently available estimates in \cite{cover1978convergent}, \cite{lutati2023focus}. A natural byproduct is an algorithm for lossless compression of English text which combines the prediction from the large language model with a lossless compression scheme. Preliminary results from limited experiments suggest that our scheme outperforms state-of-the-art text compression schemes such as BSC, ZPAQ, and paq8h. | ['Srinivas Shakkottai', 'Jean-Francois Chamberland', 'Dileep Kalathil', 'Krishna Narayanan', 'Chandra Shekhara Kaushik Valmeekam'] | 2023-06-06 | null | null | null | null | ['text-compression'] | ['natural-language-processing'] | [ 5.62123023e-02 1.54385373e-01 -3.74895513e-01 -1.31473914e-01
-1.11951518e+00 -1.85271144e-01 5.94277382e-01 6.26255333e-01
-9.41662371e-01 1.21103227e+00 5.41216493e-01 -6.91525280e-01
-3.13868700e-03 -8.87941241e-01 -8.17990005e-01 -2.53054231e-01
-4.01096523e-01 4.01164711e-01 3.60434979e-01 -3.51942718e-01
6.88695371e-01 2.41960138e-01 -1.32472277e+00 3.17380488e-01
7.14031994e-01 8.59544098e-01 3.37373704e-01 1.06587255e+00
-2.66892701e-01 1.00601983e+00 -5.49191296e-01 -8.26401353e-01
2.16556177e-01 -5.70571959e-01 -1.35100043e+00 -5.84050715e-01
-1.51979119e-01 -8.85457695e-01 -1.03680289e+00 9.76558566e-01
3.80610526e-01 -1.37780473e-01 5.68574250e-01 -6.84451997e-01
-1.38328850e-01 1.43076539e+00 -6.01657212e-01 5.39218485e-01
5.89273870e-01 -2.90498793e-01 1.03537714e+00 -3.32320035e-01
8.52142334e-01 1.00237155e+00 7.93265879e-01 1.28891915e-01
-8.62801015e-01 -6.61512375e-01 -6.48943186e-01 3.25709581e-01
-1.83399498e+00 -9.58056331e-01 1.02620810e-01 3.89966160e-01
1.51273644e+00 5.44856906e-01 6.48736596e-01 4.89490867e-01
6.78623497e-01 1.02738214e+00 5.50357103e-01 -8.23758304e-01
7.29019791e-02 2.05034584e-01 -3.27376872e-01 6.58455133e-01
5.48781872e-01 3.77040915e-02 -8.11548889e-01 -4.19125199e-01
2.45048717e-01 -3.81650925e-01 -2.14777410e-01 5.77198088e-01
-8.10102224e-01 5.89410067e-01 -3.13415885e-01 2.76545316e-01
-2.58991301e-01 6.12593591e-01 6.55311048e-01 6.25241995e-01
8.80342662e-01 4.63014096e-02 -4.02410597e-01 -9.41609323e-01
-1.65997970e+00 5.87271333e-01 1.19252932e+00 1.35499930e+00
5.49891829e-01 -1.75659463e-01 1.64768949e-01 6.36735022e-01
3.19113344e-01 5.70865929e-01 5.79995513e-01 -1.14181876e+00
9.75081027e-01 -4.70279187e-01 6.90877810e-02 -6.95412397e-01
6.70347735e-02 -1.05118737e-01 -6.50677502e-01 -4.93875116e-01
1.31528169e-01 2.69121472e-02 -3.02745551e-01 1.24125230e+00
-2.41129696e-01 -2.19701052e-01 2.17780858e-01 -1.08181722e-01
1.52217045e-01 1.21470606e+00 -2.00693518e-01 -6.30067885e-01
7.02747345e-01 -4.06672567e-01 -8.56348038e-01 1.32454246e-01
1.09703195e+00 -1.08140206e+00 5.11226058e-01 5.68150580e-01
-1.82963049e+00 1.14430150e-03 -1.07018340e+00 -3.14471275e-01
-1.09776378e-01 -3.75098974e-01 5.22787988e-01 8.25795770e-01
-1.26484895e+00 1.05884016e+00 -8.06355894e-01 -1.89845279e-01
1.53504118e-01 8.30820426e-02 -3.45334373e-02 -9.21498090e-02
-1.44317853e+00 8.88879478e-01 9.56222177e-01 -6.39117777e-01
-4.20109272e-01 -4.30639595e-01 -6.07536018e-01 4.36813802e-01
-6.92085326e-02 -1.27286524e-01 1.35465991e+00 -9.07793082e-03
-1.21980846e+00 7.42746294e-01 -3.20014387e-01 -1.23750794e+00
7.43906140e-01 -1.43712923e-01 -3.26870561e-01 7.38907397e-01
-3.15314353e-01 5.09614646e-01 6.12125218e-01 -9.40365911e-01
-7.87490487e-01 1.66973203e-01 -3.73164713e-01 2.40819007e-01
-5.20297945e-01 1.28476515e-01 -4.32431310e-01 -8.38376641e-01
-3.66322547e-02 -5.22250950e-01 2.78246887e-02 1.00450970e-01
-3.59429985e-01 -2.18481258e-01 6.73718512e-01 -1.12036717e+00
1.87630117e+00 -2.02917552e+00 -3.32900435e-01 1.98247895e-01
-8.64057094e-02 1.44296542e-01 1.18551463e-01 1.04052937e+00
3.91087681e-01 6.62309766e-01 -1.11316204e-01 -4.53893423e-01
1.45487323e-01 1.12799726e-01 -6.94723427e-01 7.60128140e-01
-7.59816587e-01 6.29246354e-01 -6.83474720e-01 -1.10628664e+00
-4.41418737e-02 8.69920626e-02 -5.95478952e-01 -5.74252158e-02
-1.23266093e-01 -7.58218646e-01 -2.80138135e-01 5.17381370e-01
6.75778866e-01 5.77055961e-02 -1.28202006e-01 5.61331272e-01
-7.12331012e-02 7.61564374e-01 -8.97940099e-01 1.73288381e+00
-3.14521939e-01 7.55701602e-01 1.30670369e-01 -6.89025581e-01
7.24232852e-01 5.80064237e-01 6.17774546e-01 -3.32671851e-01
2.14314908e-01 4.46555406e-01 -5.55486143e-01 -3.39785129e-01
1.42908263e+00 -1.79816157e-01 -2.30520945e-02 9.50308561e-01
2.66279187e-02 -7.71997094e-01 4.16338444e-01 7.45268047e-01
1.11969399e+00 -1.27383903e-01 3.87977093e-01 -4.01458949e-01
1.84612587e-01 -3.99669111e-01 4.66608964e-02 1.01895463e+00
-2.80774832e-01 4.60226148e-01 2.24458292e-01 -1.17981909e-02
-1.65150523e+00 -5.86884737e-01 -5.03387272e-01 7.12423384e-01
1.62925363e-01 -1.13364375e+00 -9.07035172e-01 -1.70972586e-01
4.57995608e-02 1.20227301e+00 -9.32420120e-02 -8.36653113e-02
-4.02324408e-01 -5.60550928e-01 1.10637140e+00 1.27955884e-01
6.36051834e-01 -9.57476556e-01 -6.05686545e-01 1.81780800e-01
-6.45456851e-01 -1.01931894e+00 -5.63601136e-01 2.21754596e-01
-1.01622057e+00 -4.12048399e-01 -6.23535395e-01 -4.91580814e-01
6.46632016e-02 1.39176637e-01 1.04695678e+00 3.33939672e-01
-8.41360316e-02 3.86764050e-01 -5.88912070e-01 -5.23227632e-01
-9.96848762e-01 2.89881706e-01 -4.37896252e-02 -8.13942254e-01
4.16241050e-01 -6.19399250e-01 -4.30200398e-01 -4.75540996e-01
-1.07110763e+00 2.01420188e-02 4.56773579e-01 5.80210090e-01
5.17267585e-01 5.65710783e-01 1.27213404e-01 -6.17333412e-01
8.61061811e-01 -6.08961642e-01 -3.64866108e-01 2.67975420e-01
-1.06259251e+00 3.64385813e-01 6.95809126e-01 -1.54532110e-02
-1.05038655e+00 -3.33447546e-01 -6.29250288e-01 1.19165480e-01
3.20659220e-01 4.95962411e-01 4.25981998e-01 1.08053394e-01
4.79057133e-01 7.60865033e-01 -2.15335682e-01 -4.38671380e-01
2.40923330e-01 1.13054872e+00 7.71436036e-01 -5.92943013e-01
5.44982970e-01 2.33456746e-01 -1.28814712e-01 -1.00960577e+00
-1.74958438e-01 -4.92310464e-01 -3.66883397e-01 1.11770034e-01
1.65875748e-01 -9.25096333e-01 -5.57396650e-01 3.70808601e-01
-1.23119569e+00 -3.25737029e-01 -7.29844928e-01 5.30305445e-01
-1.08011544e+00 9.98389661e-01 -1.03101599e+00 -1.01531816e+00
-8.08604658e-01 -2.91083813e-01 9.15018022e-01 -2.44481549e-01
-2.24723950e-01 -8.70990098e-01 -7.49596879e-02 1.37503728e-01
2.12999240e-01 -3.17506462e-01 7.47458458e-01 -6.21713281e-01
-4.20402735e-01 -6.10878408e-01 -3.67639959e-02 3.64984214e-01
-4.77801412e-01 -3.96667570e-02 -4.16957825e-01 -6.05720818e-01
1.63277522e-01 -6.77725792e-01 1.14652705e+00 1.96426675e-01
1.50146222e+00 -8.49533856e-01 -4.04231280e-01 8.01874995e-01
1.65901983e+00 1.82698533e-01 1.10941243e+00 3.31607282e-01
-1.85168236e-01 5.30995578e-02 3.63024950e-01 1.20753229e+00
3.45197141e-01 9.12130326e-02 1.32457763e-01 5.77419281e-01
-1.65162049e-02 -7.59132683e-01 4.30009365e-01 1.44614935e+00
4.08535860e-02 -8.50879550e-01 -5.69259107e-01 6.47508025e-01
-1.16988325e+00 -1.46138942e+00 3.29954386e-01 2.18802428e+00
1.46141243e+00 5.39644301e-01 -2.76897073e-01 6.38165891e-01
3.74470025e-01 4.91806149e-01 -2.68203795e-01 -7.69024253e-01
-1.79711014e-01 4.02724981e-01 1.27533281e+00 7.58871853e-01
-6.46886706e-01 7.80805469e-01 8.23261166e+00 1.40197527e+00
-4.80233312e-01 3.57489549e-02 7.57860303e-01 -2.64335304e-01
-7.25343168e-01 9.58925262e-02 -1.10526562e+00 7.17244327e-01
1.78335667e+00 -1.15357649e+00 8.68727028e-01 3.96589309e-01
1.14185428e-02 -4.11142886e-01 -5.73224664e-01 8.25828969e-01
3.21984470e-01 -1.46172535e+00 7.75722563e-02 3.59633684e-01
5.69263399e-01 1.88155800e-01 -7.68221691e-02 1.49319097e-01
5.10060608e-01 -8.03208053e-01 9.48795259e-01 5.39242208e-01
1.41520679e+00 -1.01898050e+00 6.25085890e-01 8.84684384e-01
-9.26022768e-01 -1.83635522e-02 -9.88923728e-01 1.07734136e-01
3.65633696e-01 5.21320105e-01 -8.31521153e-01 5.00229359e-01
6.49259269e-01 4.77755636e-01 -2.02191368e-01 9.18515980e-01
2.14729711e-01 9.91377771e-01 -6.98604643e-01 -2.75389194e-01
1.30286723e-01 1.29445195e-01 5.88134170e-01 1.58327401e+00
8.06719422e-01 4.44009334e-01 -3.94039810e-01 4.25447464e-01
-5.43005466e-01 2.60114253e-01 -5.12959838e-01 -3.06825876e-01
7.72813797e-01 3.03708732e-01 -4.88642484e-01 -5.34361899e-01
-2.44720116e-01 1.03645968e+00 -6.27698153e-02 8.21667239e-02
-5.41375339e-01 -1.14860952e+00 -8.51041824e-03 2.42476568e-01
5.95938325e-01 -4.61434841e-01 -4.35399175e-01 -1.32402444e+00
1.32676987e-02 -6.67993248e-01 2.21472964e-01 -5.61352968e-01
-5.05684972e-01 2.57629633e-01 4.46082264e-01 -9.14526761e-01
-6.48428261e-01 6.68847114e-02 -2.02781320e-01 9.04833198e-01
-1.32684040e+00 -3.59978557e-01 5.06203532e-01 4.32523757e-01
4.56469178e-01 -2.99125493e-01 9.57324564e-01 2.64776926e-02
2.19658747e-01 1.01864815e+00 1.18251038e+00 -2.57556707e-01
4.30777192e-01 -9.79256868e-01 4.65476424e-01 9.17491674e-01
7.36832470e-02 2.34681696e-01 1.06381464e+00 -7.92006969e-01
-1.39133441e+00 -6.28404737e-01 1.68726838e+00 6.00543283e-02
5.07049203e-01 -1.94870397e-01 -7.84154594e-01 6.14093423e-01
3.06283772e-01 -3.81137669e-01 3.62203032e-01 -4.05583739e-01
-1.41844094e-01 -1.67933092e-01 -1.43142033e+00 2.86237925e-01
8.59161377e-01 -6.99379921e-01 -6.26329899e-01 4.58213836e-01
1.04007745e+00 -4.69539285e-01 -1.03142476e+00 -9.77495983e-02
6.45200372e-01 -8.67188990e-01 8.13412428e-01 -6.27178103e-02
6.05295897e-01 6.18017018e-01 -5.42120934e-01 -7.55450666e-01
1.60406932e-01 -1.30410743e+00 -6.07975006e-01 1.06750560e+00
1.92601144e-01 -3.20379704e-01 7.90792644e-01 6.59168065e-01
1.35079727e-01 -6.39261723e-01 -1.37209630e+00 -9.94593620e-01
7.03861177e-01 -6.18159413e-01 6.55220628e-01 3.38742286e-01
5.65397024e-01 -4.32003826e-01 -6.11921847e-01 -5.49636066e-01
6.26502395e-01 -2.67596424e-01 2.94495016e-01 -7.64984190e-01
-4.07065660e-01 -2.81702787e-01 -2.29202688e-01 -1.70048976e+00
4.15489376e-02 -9.73802269e-01 7.62515143e-02 -1.22480333e+00
4.08271849e-01 -4.51447606e-01 -6.89806789e-02 2.33324170e-01
3.89678031e-01 -7.80935958e-02 2.31206715e-01 3.70162487e-01
-6.33106947e-01 7.35993207e-01 8.52971852e-01 -1.94432810e-01
2.98905894e-02 -4.69888262e-02 -5.35704792e-01 2.24661991e-01
6.42979503e-01 -6.87270880e-01 -3.06369990e-01 -3.90244842e-01
4.00958687e-01 7.02795565e-01 -3.58259290e-01 -9.95106995e-01
4.43729460e-01 -7.96703026e-02 5.92846796e-03 -9.41168666e-01
1.49283677e-01 -6.33366823e-01 -1.18143141e-01 7.59099901e-01
-7.12153673e-01 1.34747401e-01 2.18191758e-01 6.77913547e-01
-1.81750357e-01 -9.70458090e-01 6.24308825e-01 -3.03901523e-01
-1.27606094e-01 3.41591448e-01 -7.71324933e-01 2.64349133e-01
4.98138547e-01 -7.82237723e-02 -2.87227780e-01 -1.03832948e+00
-1.60122544e-01 3.62220183e-02 4.21559423e-01 -1.22783184e-01
7.08033800e-01 -8.64194274e-01 -1.02977180e+00 7.29100481e-02
-4.82703418e-01 -3.07215780e-01 -1.38028547e-01 2.88743496e-01
-1.11944938e+00 8.25294077e-01 6.73545301e-02 1.36769041e-01
-1.31162453e+00 4.79558319e-01 -6.15172498e-02 -7.10094631e-01
-7.17440784e-01 7.31577456e-01 -1.01357710e+00 3.22480172e-01
4.54349548e-01 -1.22458637e-01 4.83855128e-01 -4.68688458e-01
7.85568774e-01 6.47986948e-01 -1.79567203e-01 -5.40312171e-01
1.04142606e-01 -1.32502854e-01 -2.42173716e-01 -6.08824372e-01
1.46244478e+00 -7.53371656e-01 -3.55408520e-01 3.50411713e-01
1.61713552e+00 1.33279949e-01 -7.67927468e-01 -3.88854206e-01
-2.20936269e-01 -5.75781167e-01 1.61129847e-01 -5.98308623e-01
-4.28761721e-01 7.73613751e-01 2.10940897e-01 3.98174942e-01
1.32781827e+00 -2.40703389e-01 1.56480348e+00 9.13192570e-01
7.45279253e-01 -1.37492251e+00 -4.24450845e-01 6.62067950e-01
7.69360363e-01 -6.18740976e-01 7.94184208e-01 1.58740237e-01
-1.40764922e-01 1.31791770e+00 -3.00533772e-01 2.35419810e-01
1.01226699e+00 5.36247551e-01 -6.31623268e-01 3.77998084e-01
-1.03255093e+00 5.08368373e-01 -4.33981657e-01 4.20608111e-02
4.22108501e-01 -8.23571905e-02 -8.92674565e-01 2.60972083e-01
-8.22412789e-01 2.52149671e-01 8.58556926e-01 1.35702562e+00
-8.26904178e-01 -1.04484284e+00 -2.46055767e-01 7.32147694e-01
-1.17844248e+00 -5.78896642e-01 1.15729772e-01 5.16900718e-01
-2.29001418e-01 8.60047579e-01 1.00168481e-01 -3.72609377e-01
-4.81787503e-01 1.86434746e-01 5.19448042e-01 -1.12237893e-02
-3.49551618e-01 -3.67460139e-02 2.58434266e-01 -2.69520611e-01
-8.62638280e-02 -8.90796721e-01 -1.20348668e+00 -1.41406548e+00
-5.65197170e-01 5.03447652e-01 6.16260409e-01 6.95542991e-01
1.37511626e-01 -3.84749621e-01 6.41175807e-01 -4.55573022e-01
-1.01564026e+00 -1.08904636e+00 -1.08863401e+00 7.56225064e-02
3.39284718e-01 4.25078541e-01 -3.87247264e-01 3.76104355e-01] | [8.539252281188965, 3.467489004135132] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.