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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
eff93855-ec54-4f50-a7d8-3d7fe82d8d74 | relaxing-the-additivity-constraints-in | 2305.19838 | null | https://arxiv.org/abs/2305.19838v1 | https://arxiv.org/pdf/2305.19838v1.pdf | Relaxing the Additivity Constraints in Decentralized No-Regret High-Dimensional Bayesian Optimization | Bayesian Optimization (BO) is typically used to optimize an unknown function $f$ that is noisy and costly to evaluate, by exploiting an acquisition function that must be maximized at each optimization step. Although provably asymptotically optimal BO algorithms are efficient at optimizing low-dimensional functions, scaling them to high-dimensional spaces remains an open research problem, often tackled by assuming an additive structure for $f$. However, such algorithms introduce additional restrictive assumptions on the additive structure that reduce their applicability domain. In this paper, we relax the restrictive assumptions on the additive structure of $f$, at the expense of weakening the maximization guarantees of the acquisition function, and we address the over-exploration problem for decentralized BO algorithms. To these ends, we propose DuMBO, an asymptotically optimal decentralized BO algorithm that achieves very competitive performance against state-of-the-art BO algorithms, especially when the additive structure of $f$ does not exist or comprises high-dimensional factors. | ['Thomas Begin', 'Patrick Thiran', 'Anthony Bardou'] | 2023-05-31 | null | null | null | null | ['bayesian-optimization'] | ['methodology'] | [-2.53559589e-01 1.72663778e-01 -1.92915455e-01 -2.32560202e-01
-9.55754697e-01 -6.39026999e-01 2.20925361e-01 6.57287240e-02
-4.88897145e-01 7.61754692e-01 1.60408422e-01 -2.83060998e-01
-5.60379624e-01 -7.67835855e-01 -1.06453109e+00 -8.77562821e-01
-3.36337715e-01 6.14487350e-01 -2.53097206e-01 -8.64101499e-02
2.15689447e-02 4.26953137e-01 -1.31193399e+00 -3.64999384e-01
7.71535575e-01 1.28910375e+00 1.58929810e-01 6.12341762e-01
1.58585280e-01 3.26636642e-01 -4.69250888e-01 -4.04611051e-01
5.97588122e-01 -2.80552059e-01 -6.82264268e-01 -1.23317037e-02
2.37709314e-01 -4.09234107e-01 -7.24291205e-02 1.47495675e+00
4.35555339e-01 1.40279159e-01 6.28131211e-01 -9.82723892e-01
-2.67755836e-01 9.40665245e-01 -2.56118327e-01 5.30510433e-02
1.10839847e-02 -8.69426653e-02 1.43963838e+00 -8.52990627e-01
4.84869927e-01 1.30422294e+00 5.60073495e-01 3.57826352e-01
-1.47517109e+00 -5.54923952e-01 4.69207406e-01 -1.17162652e-01
-1.69173992e+00 -5.07255435e-01 8.14672828e-01 -4.50252324e-01
5.25529146e-01 1.31242171e-01 3.78197849e-01 7.68742800e-01
-1.19243495e-01 1.04346192e+00 1.04254985e+00 -2.56568432e-01
6.97177231e-01 1.37146637e-01 2.57310480e-01 7.37198412e-01
5.64058959e-01 1.22403942e-01 -7.29732096e-01 -4.60407346e-01
5.08620083e-01 -2.75713056e-01 -4.72069502e-01 -8.34010959e-01
-8.27610612e-01 1.10521328e+00 4.36998039e-01 2.86667615e-01
-4.59631264e-01 4.50620264e-01 5.77401631e-02 4.34569269e-01
4.10419524e-01 6.50651217e-01 -5.69199026e-01 -2.68201083e-01
-1.07110333e+00 6.93569720e-01 1.01271236e+00 1.11313784e+00
9.94702935e-01 -9.87832099e-02 1.22777112e-02 6.17211342e-01
4.92766857e-01 5.86155176e-01 6.20833486e-02 -1.07735574e+00
6.24530017e-01 2.28553250e-01 4.76703465e-01 -1.17064738e+00
-3.29811752e-01 -7.84052193e-01 -8.45265031e-01 -1.61292732e-01
6.92040861e-01 -3.72492969e-01 -4.46309447e-01 1.96192658e+00
5.35280406e-01 -1.90812141e-01 6.53058365e-02 1.01972294e+00
9.24409851e-02 5.45470059e-01 -5.33595800e-01 -4.94220793e-01
1.00901091e+00 -7.62482643e-01 -7.36275554e-01 -3.48684132e-01
6.44369721e-01 -3.46216947e-01 1.16971636e+00 5.69720209e-01
-1.18193185e+00 7.44033456e-02 -1.14693785e+00 1.43596813e-01
-5.04186202e-04 -2.20135860e-02 7.46776462e-01 1.04513657e+00
-9.15745556e-01 4.91574943e-01 -8.61732781e-01 1.78704113e-01
4.31435615e-01 6.08694553e-01 -7.26279169e-02 -2.41406947e-01
-9.30295944e-01 6.55837953e-01 1.96951881e-01 3.73266965e-01
-1.15595949e+00 -7.32057571e-01 -6.84859216e-01 3.58419031e-01
9.72859442e-01 -7.76344121e-01 1.34675121e+00 -4.98219967e-01
-1.64095271e+00 2.32753426e-01 -1.85065284e-01 -6.59454107e-01
7.39325583e-01 -4.65499252e-01 3.53653371e-01 7.10570440e-02
-1.89872123e-02 -7.13363662e-03 1.06492054e+00 -1.12337220e+00
-4.48762149e-01 -7.65348077e-01 4.40333605e-01 3.67670923e-01
-4.43687022e-01 -2.30755582e-01 -5.30391514e-01 -4.22884494e-01
2.26852775e-01 -8.27435613e-01 -6.95692956e-01 -1.90522559e-02
-3.35917234e-01 -7.91877359e-02 4.85535949e-01 -3.43630195e-01
1.29605639e+00 -2.17803192e+00 4.99908775e-01 5.77329278e-01
3.73145550e-01 -1.77601665e-01 2.16783613e-01 2.17299625e-01
5.44145346e-01 3.13462645e-01 -4.51107234e-01 -6.89132452e-01
3.73751909e-01 2.50925362e-01 -2.01359913e-01 9.36712801e-01
-2.47941986e-01 4.90316778e-01 -8.81069422e-01 -2.45022699e-01
-4.77057621e-02 8.32369328e-02 -1.13154411e+00 5.79548478e-02
-5.39147794e-01 1.66408226e-01 -7.90443540e-01 5.92035353e-01
7.26327419e-01 -2.94114828e-01 3.70055944e-01 1.49968624e-01
6.29529431e-02 5.34662493e-02 -1.78981996e+00 1.72396159e+00
-2.55529463e-01 1.77709028e-01 9.11075830e-01 -1.37521827e+00
7.69071758e-01 4.36444022e-02 7.65041649e-01 -2.64437228e-01
2.64988869e-01 3.17787200e-01 -1.33483827e-01 4.08639312e-02
4.20381606e-01 -2.05904558e-01 -3.16853762e-01 3.71508777e-01
2.23657079e-02 -3.41909945e-01 2.27477044e-01 1.03604361e-01
1.11211312e+00 -2.71746188e-01 2.36838743e-01 -7.27428257e-01
3.74938875e-01 -2.43659735e-01 7.40854919e-01 1.27604413e+00
-1.77628309e-01 2.67058551e-01 6.24657571e-01 -8.51832852e-02
-8.49903286e-01 -8.31708789e-01 -3.71776849e-01 9.84185100e-01
2.41945863e-01 -6.72270656e-01 -7.61014521e-01 -8.00659597e-01
2.09634513e-01 3.76136750e-01 -5.70939243e-01 9.60235894e-02
-4.49242145e-01 -9.34876800e-01 2.70036042e-01 2.31239378e-01
3.62833828e-01 -1.18977629e-01 -5.28737962e-01 4.38404113e-01
-3.33142728e-02 -9.52974319e-01 -4.76616561e-01 3.61348927e-01
-8.43790710e-01 -9.65283036e-01 -6.01068258e-01 -2.19878152e-01
6.00317240e-01 1.38856679e-01 9.42819536e-01 -2.84382850e-01
2.59286631e-02 4.91189897e-01 -1.12925664e-01 -6.08093977e-01
4.18425426e-02 2.59819329e-01 1.63799822e-01 1.82768911e-01
-1.88694950e-02 -5.97176313e-01 -6.60625935e-01 4.15090621e-01
-8.98852229e-01 -3.80310088e-01 4.89438176e-01 9.83249545e-01
6.95396066e-01 2.66314924e-01 3.39743435e-01 -8.03295434e-01
6.51173294e-01 -4.25492376e-01 -1.21954584e+00 7.15838820e-02
-6.82459652e-01 5.79478621e-01 6.87499821e-01 -2.02130973e-01
-8.35700810e-01 1.13737598e-01 -2.09736656e-02 -4.72259134e-01
3.38870674e-01 8.16130340e-01 -3.52541775e-01 -4.01695073e-02
7.27022588e-01 -8.92781317e-02 1.12077735e-01 -7.94299304e-01
5.38084686e-01 3.97892475e-01 3.18186194e-01 -1.01670814e+00
7.53880978e-01 6.56837821e-01 3.15035433e-01 -6.46491349e-01
-1.10141158e+00 -5.35565257e-01 -3.68070424e-01 3.24712209e-02
3.08515042e-01 -8.38400662e-01 -7.37102866e-01 2.88846381e-02
-8.30340743e-01 -1.13161907e-01 -4.00904685e-01 5.34419537e-01
-8.87485445e-01 2.55715311e-01 -1.84901088e-01 -1.05675566e+00
-1.70013502e-01 -1.33031523e+00 8.85878265e-01 -2.27978542e-01
1.40381277e-01 -6.82211757e-01 9.38231871e-02 2.71791905e-01
4.58597273e-01 -2.97690742e-02 6.79630160e-01 -4.46480066e-01
-9.04364944e-01 -1.25201255e-01 1.49143869e-02 4.06352729e-01
-2.97220707e-01 -3.48919749e-01 -6.95535958e-01 -5.03286839e-01
3.47054809e-01 -1.56464383e-01 8.60820830e-01 6.41207039e-01
1.17201149e+00 -5.96888125e-01 -3.36202055e-01 7.78624535e-01
1.30216444e+00 -1.35776162e-01 5.51267788e-02 2.02440396e-01
2.77876586e-01 5.31884432e-01 6.20936573e-01 8.19453955e-01
3.32024217e-01 7.30508804e-01 6.85042918e-01 3.64638150e-01
4.06447202e-01 -8.77565816e-02 3.15426201e-01 6.94227219e-01
1.36951461e-01 -1.70933589e-01 -8.45714688e-01 5.95266879e-01
-2.04826927e+00 -5.61589420e-01 3.05016398e-01 2.26816535e+00
8.43912899e-01 1.06416168e-02 3.86943221e-02 9.11423340e-02
5.09349823e-01 2.34748825e-01 -6.38318956e-01 -1.16973601e-01
-1.14097536e-01 1.39194325e-01 7.70742536e-01 6.62402093e-01
-1.05231881e+00 7.20923066e-01 7.00984192e+00 8.83012354e-01
-8.51679564e-01 2.53919005e-01 2.83823252e-01 -3.43138039e-01
-3.28297824e-01 1.36594668e-01 -1.21976769e+00 3.88737053e-01
8.07229996e-01 -1.62938744e-01 7.87060559e-01 1.25593793e+00
1.14341930e-01 -7.02510327e-02 -1.27185917e+00 9.82017815e-01
-1.57157540e-01 -1.19533408e+00 -4.05630857e-01 4.58021730e-01
7.92190731e-01 1.28989115e-01 1.44146875e-01 2.69951522e-01
6.84180915e-01 -9.84806478e-01 7.71168649e-01 3.46436948e-01
3.02533507e-01 -9.37522531e-01 5.41248262e-01 6.02430701e-01
-8.70848179e-01 -4.33330476e-01 -5.71043491e-01 -1.29118320e-02
2.00076535e-01 9.42343295e-01 -6.53882384e-01 5.89332283e-01
9.18709934e-01 4.34004247e-01 -2.20130995e-01 9.66797054e-01
-8.14050138e-02 5.85608542e-01 -1.03074312e+00 -1.36509389e-01
4.37764883e-01 -2.94004053e-01 7.75577784e-01 8.88535261e-01
4.81103718e-01 -6.15757182e-02 2.86362231e-01 8.96314263e-01
-2.98700273e-01 2.04434320e-01 -6.69160903e-01 -1.41309202e-01
4.50668752e-01 7.78979242e-01 -1.43588096e-01 -1.61010906e-01
-3.36160123e-01 3.47815812e-01 4.48506474e-01 3.05989742e-01
-6.19084656e-01 -9.92448628e-03 8.29154015e-01 -4.14791005e-03
6.87428296e-01 -5.20058334e-01 -2.48042554e-01 -1.43308794e+00
3.33647281e-01 -9.70375597e-01 6.53219581e-01 1.44556509e-02
-1.20244074e+00 7.47547820e-02 2.10129574e-01 -7.95917153e-01
-4.22708631e-01 -5.51129758e-01 5.58166504e-02 5.44346273e-01
-1.39712620e+00 -6.67086542e-01 2.08282232e-01 7.59601295e-01
2.18764365e-01 1.05212599e-01 6.15480006e-01 2.52148688e-01
-5.58339834e-01 5.89365184e-01 7.35062063e-01 -2.26137757e-01
3.62463117e-01 -1.27976704e+00 -3.54767174e-01 8.96641910e-01
1.81113124e-01 8.13610077e-01 9.37916934e-01 -2.62973517e-01
-1.93598425e+00 -7.58785546e-01 4.07755911e-01 -3.37983578e-01
7.53300488e-01 -6.14722550e-01 -5.01434922e-01 6.10222518e-01
-2.46026039e-01 2.43865147e-01 5.48222363e-01 5.81758797e-01
-2.09558710e-01 -3.29761565e-01 -1.23666894e+00 5.74344754e-01
9.66499865e-01 -3.65083039e-01 -2.55738527e-01 3.21160376e-01
5.59848309e-01 -2.83001035e-01 -1.11210895e+00 5.46591580e-01
4.42382216e-01 -7.43048251e-01 1.07154679e+00 -4.53950554e-01
-1.56417221e-01 -3.42255026e-01 -6.02458358e-01 -1.11322129e+00
1.32598981e-01 -1.07441151e+00 -5.77036917e-01 9.03473675e-01
3.54027003e-01 -5.29280245e-01 8.95355284e-01 6.68381512e-01
3.98326889e-02 -6.58281922e-01 -1.41717875e+00 -9.47974086e-01
2.50583589e-01 -7.00938046e-01 6.23221755e-01 6.49739206e-01
-6.21029288e-02 2.76008070e-01 -6.59797370e-01 3.41165096e-01
8.93165767e-01 1.81401566e-01 1.07809901e+00 -1.01701510e+00
-8.09531093e-01 -5.38665771e-01 -1.19396731e-01 -1.74799132e+00
4.01674770e-02 -6.07196689e-01 3.23937118e-01 -8.84780824e-01
-5.80211282e-02 -6.54907048e-01 -2.01938614e-01 -1.63945686e-02
6.38317466e-02 -3.30324650e-01 2.29963720e-01 2.10827932e-01
-6.71167016e-01 8.83005202e-01 1.16697717e+00 -1.59867913e-01
-1.85068116e-01 2.03872561e-01 -9.29963887e-01 8.26444864e-01
3.75256389e-01 -5.96768796e-01 -3.40174109e-01 -5.49302697e-01
3.89858603e-01 1.17341712e-01 4.70619686e-02 -5.72131157e-01
4.34547096e-01 -2.63597846e-01 -2.36680016e-01 -5.61806560e-01
5.60391605e-01 -9.65406835e-01 -3.45031060e-02 2.35718653e-01
-2.14157686e-01 -2.52239823e-01 -3.07759374e-01 9.13907647e-01
-2.35324651e-01 -5.46625137e-01 7.62746036e-01 -1.54352516e-01
-6.35064542e-02 4.51804876e-01 -3.56152117e-01 7.33013079e-02
8.12845170e-01 1.54428914e-01 -3.24835703e-02 -5.20931900e-01
-8.28394473e-01 3.47980767e-01 2.77069032e-01 -1.21332407e-01
2.94577450e-01 -9.93611515e-01 -5.46172261e-01 -2.98745814e-04
-1.93005562e-01 5.71937680e-01 5.32577448e-02 8.48292649e-01
-2.28020608e-01 3.48564804e-01 5.12283981e-01 -6.63548887e-01
-8.00794601e-01 6.70894921e-01 1.86097011e-01 -6.69859111e-01
-4.25028861e-01 8.61144841e-01 1.50617763e-01 -3.43205541e-01
5.89637816e-01 -4.36737061e-01 2.16192231e-01 3.52439284e-02
2.69624650e-01 3.71444315e-01 -5.11308424e-02 -2.65633732e-01
-1.61463425e-01 3.54960501e-01 4.56136353e-02 -3.43016326e-01
1.34105158e+00 -3.46311688e-01 6.70692176e-02 1.57738760e-01
1.14802933e+00 1.66292712e-01 -1.19149625e+00 -6.38648868e-01
1.77668571e-01 -6.71823204e-01 3.47144037e-01 -3.66817504e-01
-1.03379476e+00 7.27434218e-01 2.70296335e-01 2.06612259e-01
8.92791927e-01 1.90732479e-01 4.52030092e-01 1.01109970e+00
7.03156471e-01 -1.16851389e+00 -5.83832897e-02 5.39170861e-01
8.02146316e-01 -1.05742753e+00 1.33670703e-01 -3.93633008e-01
-1.55222580e-01 7.74420321e-01 1.20942079e-01 -2.44637236e-01
9.58427846e-01 3.10911000e-01 -5.27275622e-01 -2.42548287e-01
-5.71599543e-01 -1.76036000e-01 1.73818991e-01 2.53553063e-01
-1.03899524e-01 1.09658718e-01 -6.08172894e-01 8.68624628e-01
-4.59200919e-01 -3.65645826e-01 1.20655552e-01 1.11930203e+00
-4.97857422e-01 -1.14656198e+00 -5.67280173e-01 3.77824634e-01
-7.81431854e-01 -4.41542864e-02 -2.72127360e-01 8.05691838e-01
-2.91893810e-01 8.73111129e-01 -3.16274315e-01 6.85966313e-02
1.52812764e-01 -1.43633872e-01 3.81487906e-01 -4.26538140e-01
-4.27859575e-01 4.07742321e-01 1.51900664e-01 -6.43146515e-01
-3.64759475e-01 -8.57999682e-01 -7.23188281e-01 -2.81599551e-01
-6.12931550e-01 5.31934559e-01 8.75355899e-01 9.67184544e-01
4.01755899e-01 2.90276725e-02 6.97032809e-01 -5.97809374e-01
-1.37345028e+00 -7.69212544e-01 -8.09426010e-01 -1.42969959e-03
5.37684739e-01 -9.39179242e-01 -5.87128341e-01 -3.18033516e-01] | [6.541970729827881, 4.224836349487305] |
a18cb741-b73c-4d82-aeff-086cc821fcde | an-effective-deep-network-for-head-pose | 2210.13705 | null | https://arxiv.org/abs/2210.13705v1 | https://arxiv.org/pdf/2210.13705v1.pdf | An Effective Deep Network for Head Pose Estimation without Keypoints | Human head pose estimation is an essential problem in facial analysis in recent years that has a lot of computer vision applications such as gaze estimation, virtual reality, and driver assistance. Because of the importance of the head pose estimation problem, it is necessary to design a compact model to resolve this task in order to reduce the computational cost when deploying on facial analysis-based applications such as large camera surveillance systems, AI cameras while maintaining accuracy. In this work, we propose a lightweight model that effectively addresses the head pose estimation problem. Our approach has two main steps. 1) We first train many teacher models on the synthesis dataset - 300W-LPA to get the head pose pseudo labels. 2) We design an architecture with the ResNet18 backbone and train our proposed model with the ensemble of these pseudo labels via the knowledge distillation process. To evaluate the effectiveness of our model, we use AFLW-2000 and BIWI - two real-world head pose datasets. Experimental results show that our proposed model significantly improves the accuracy in comparison with the state-of-the-art head pose estimation methods. Furthermore, our model has the real-time speed of $\sim$300 FPS when inferring on Tesla V100. | ['Hai Tran', 'Huong Ninh', 'Minh Bui', 'Viet Tran', 'Chien Thai'] | 2022-10-25 | null | null | null | null | ['head-pose-estimation', 'gaze-estimation'] | ['computer-vision', 'computer-vision'] | [-3.08643192e-01 1.47441342e-01 1.01822019e-01 -6.77693784e-01
-5.89701176e-01 2.52523065e-01 2.31287390e-01 -6.11653268e-01
-6.35549784e-01 4.94714171e-01 1.70258731e-01 -1.42229751e-01
1.57949641e-01 -3.44859123e-01 -4.99629647e-01 -6.75299406e-01
4.52766120e-02 3.62832516e-01 5.65790176e-01 -1.77802861e-01
7.03272074e-02 5.54075778e-01 -1.94797361e+00 -3.84800464e-01
6.67684436e-01 1.09282029e+00 1.50289208e-01 3.82384449e-01
2.12541491e-01 7.54552007e-01 -4.24297065e-01 -2.31454194e-01
5.21220453e-02 -2.38776132e-02 -5.32002866e-01 1.74296409e-01
7.28293717e-01 -6.86608851e-01 -2.90125310e-01 1.05015397e+00
1.01779604e+00 1.64188817e-01 2.98069954e-01 -1.50605607e+00
1.05222262e-01 1.16918907e-01 -1.01783705e+00 4.19243760e-02
3.22176546e-01 1.82967976e-01 3.06780875e-01 -9.24837589e-01
3.52490425e-01 1.47912872e+00 6.47678256e-01 8.54711652e-01
-5.25298059e-01 -1.24968410e+00 2.92150944e-01 7.14332938e-01
-1.66404712e+00 -1.03901827e+00 7.48889565e-01 -6.29695505e-02
5.84705055e-01 1.57524906e-02 5.21235406e-01 7.96072125e-01
-7.84304738e-02 9.56421971e-01 9.94753540e-01 -2.93428957e-01
-4.77532707e-02 -1.95264950e-01 2.83579320e-01 9.72543120e-01
-6.38207421e-03 -2.12285668e-01 -6.99282587e-01 5.09557799e-02
4.32501465e-01 -1.35654628e-01 -5.16486049e-01 -1.59842953e-01
-7.28991091e-01 6.06804550e-01 3.85008723e-01 -7.72346705e-02
-2.45965183e-01 1.07135139e-01 3.03215265e-01 -1.21838279e-01
5.22437513e-01 -2.76461512e-01 -2.96598732e-01 -4.02596556e-02
-9.55683589e-01 2.79964805e-01 7.05134392e-01 1.13640654e+00
5.73148608e-01 -1.07265608e-02 1.19629182e-01 8.76116574e-01
7.37812877e-01 7.10863531e-01 6.29632533e-01 -6.52951777e-01
3.06774050e-01 3.49686652e-01 -9.92604047e-02 -8.03310990e-01
-8.71148586e-01 -1.36290610e-01 -7.80187845e-01 2.47244947e-02
2.76843280e-01 -1.93596929e-01 -1.01094770e+00 1.87563944e+00
8.45877767e-01 5.46307027e-01 -3.06072623e-01 9.42698181e-01
1.20975339e+00 5.65953791e-01 1.03922546e-01 -3.92257869e-01
1.68840802e+00 -1.06261253e+00 -8.61265779e-01 -3.05937082e-01
4.81563121e-01 -7.90392995e-01 7.35260844e-01 2.63631642e-01
-9.50821042e-01 -5.51356494e-01 -8.43775332e-01 -2.06492245e-01
-1.90270580e-02 3.23107302e-01 3.92730355e-01 8.17534089e-01
-1.34577656e+00 1.61193162e-02 -8.80555153e-01 -5.51929951e-01
2.71019667e-01 7.57107615e-01 -3.69165361e-01 6.48602694e-02
-9.79572237e-01 9.11962509e-01 1.73303887e-01 4.15955126e-01
-5.79588652e-01 -4.49421883e-01 -1.07004046e+00 9.13241506e-02
4.23091173e-01 -4.21906322e-01 1.53600264e+00 -5.90785563e-01
-1.86735904e+00 7.79982507e-01 -6.16384745e-01 -8.16934481e-02
2.89665550e-01 -3.26160103e-01 -4.46903497e-01 -1.26921102e-01
-1.47217229e-01 8.73275340e-01 8.89925778e-01 -8.05122972e-01
-8.81378055e-01 -8.20030391e-01 -1.44386351e-01 2.27849945e-01
-3.38970244e-01 2.67383248e-01 -9.57450688e-01 -7.95494989e-02
1.17222138e-01 -1.21914363e+00 1.64819673e-01 1.64955948e-02
-2.04500616e-01 -6.73029184e-01 1.03859186e+00 -7.00792015e-01
1.17531931e+00 -2.02191138e+00 -1.17212079e-01 1.48865670e-01
2.74111599e-01 5.86312950e-01 -9.51268077e-02 -2.26563752e-01
-1.26927197e-01 -6.12431169e-01 2.53923923e-01 -8.00634205e-01
-1.12863734e-01 -1.34479016e-01 -9.71658751e-02 8.41990590e-01
-2.61429697e-01 7.17044532e-01 -4.09604549e-01 -7.48018444e-01
2.01979518e-01 7.28238165e-01 -6.20254695e-01 4.68675405e-01
9.03317630e-02 2.98088938e-01 -3.11078399e-01 5.23653328e-01
1.05664384e+00 1.03778820e-02 3.76534536e-02 -3.49656224e-01
-5.80281354e-02 -1.79223437e-02 -1.31204915e+00 1.49970770e+00
-3.04626077e-01 5.34124494e-01 2.16528997e-01 -5.72225988e-01
6.91501141e-01 3.15245628e-01 2.16703832e-01 -6.25376880e-01
6.15598261e-01 9.09229554e-03 -3.41016762e-02 -6.75519526e-01
4.12988931e-01 1.16372958e-01 2.35210419e-01 3.30640018e-01
1.95995629e-01 5.25235012e-02 -1.07320063e-01 -2.49773279e-01
6.34014785e-01 1.67178184e-01 2.46819675e-01 -1.87709853e-01
1.00985420e+00 -6.94041848e-01 8.20090473e-01 -6.58215433e-02
-5.65461516e-01 5.08227527e-01 3.03141087e-01 -4.35215145e-01
-4.45702702e-01 -4.21548814e-01 -6.08388707e-02 1.42625630e+00
1.39802229e-02 -3.38787884e-01 -1.28594303e+00 -5.23477912e-01
-2.49147221e-01 4.90888536e-01 -5.46024680e-01 -7.74883628e-02
-8.47905993e-01 -7.23468125e-01 4.39121962e-01 5.86700201e-01
8.22438121e-01 -1.11686933e+00 -5.78308046e-01 -2.74315387e-01
-1.72376856e-01 -1.28978229e+00 -6.88180923e-01 -4.60341632e-01
-4.79882985e-01 -1.02592826e+00 -8.12113762e-01 -9.26358998e-01
7.19874382e-01 3.70980203e-01 7.29630470e-01 2.53704578e-01
-5.88920824e-02 2.20812857e-01 -2.56050285e-02 -6.19716883e-01
2.27671355e-01 3.90072525e-01 3.86870325e-01 2.25372374e-01
6.98247433e-01 -3.38677913e-01 -7.49328136e-01 4.06189859e-01
-4.79771495e-01 5.42650633e-02 4.93252456e-01 5.30837417e-01
2.94606209e-01 -3.42327774e-01 5.49840808e-01 -7.11147547e-01
4.49673712e-01 -1.60226092e-01 -9.53949928e-01 2.61194795e-01
-3.58738601e-01 -3.59995104e-02 3.16120178e-01 -3.17476064e-01
-1.21541512e+00 2.32029483e-01 -5.82640231e-01 -4.13042665e-01
-1.30398929e-01 2.10630849e-01 -4.62827086e-01 -2.46901214e-01
1.88123107e-01 2.06295982e-01 3.16151917e-01 -3.84626001e-01
2.79523451e-02 1.05759597e+00 6.44283593e-01 -2.51573533e-01
6.09898686e-01 2.42634252e-01 8.53737891e-02 -1.20864987e+00
-7.85810828e-01 -5.88631988e-01 -4.94549692e-01 -2.12052941e-01
8.06561053e-01 -1.05692410e+00 -1.50090587e+00 9.57016647e-01
-1.27520311e+00 -1.16005860e-01 5.34262836e-01 4.84478712e-01
-4.16800469e-01 3.18179578e-01 -3.17900330e-01 -8.59478056e-01
-7.28432834e-01 -1.47776079e+00 1.34397674e+00 5.90316594e-01
2.29700822e-02 -5.87613225e-01 -1.08881362e-01 4.15640324e-01
4.01907384e-01 -2.22601324e-01 3.39456797e-01 -3.10808033e-01
-4.93571341e-01 -2.58798718e-01 -3.44758421e-01 -6.02237321e-02
-1.41835645e-01 -3.05954874e-01 -1.41635370e+00 -4.77357119e-01
-2.77296603e-02 -4.06720221e-01 5.31191885e-01 4.45848227e-01
1.46999907e+00 1.25270365e-02 -3.97941411e-01 9.04560030e-01
7.18388975e-01 4.36939001e-02 7.68117070e-01 7.37781525e-02
8.09711814e-01 6.46564305e-01 4.86216486e-01 3.93203974e-01
9.86854374e-01 8.84801805e-01 3.60203236e-01 5.19797988e-02
-2.09454685e-01 -2.70713598e-01 3.03140432e-01 9.45162117e-01
-1.94195509e-01 -9.19738784e-04 -7.87007332e-01 2.44442657e-01
-1.83703804e+00 -7.73786783e-01 2.13692427e-01 2.30052972e+00
7.42036223e-01 -2.74895411e-02 3.23987365e-01 7.13912945e-04
7.14243293e-01 1.20162554e-01 -5.15462399e-01 9.32643749e-03
5.95485449e-01 1.82937965e-01 3.27905327e-01 4.93379325e-01
-1.10573614e+00 1.16407275e+00 5.58736324e+00 7.59509265e-01
-1.46933544e+00 2.13386595e-01 5.05324721e-01 -1.49580136e-01
4.43023682e-01 -5.89966774e-01 -1.47497702e+00 4.43083435e-01
9.24019396e-01 7.29371011e-02 2.81299949e-01 9.90607977e-01
1.96315765e-01 -1.94522023e-01 -1.00087821e+00 1.41006362e+00
6.26044095e-01 -7.40987480e-01 -3.62183899e-01 1.04237825e-01
1.49120510e-01 1.85106846e-03 1.87037308e-02 5.60776949e-01
-1.62257597e-01 -1.04012477e+00 5.85688055e-01 2.73376644e-01
9.10625339e-01 -9.35086787e-01 8.77930105e-01 6.00423038e-01
-1.33075881e+00 1.66667238e-01 -3.79057854e-01 3.16626057e-02
1.88428476e-01 1.43411472e-01 -8.21952283e-01 1.75030138e-02
7.03837693e-01 3.94475967e-01 -6.49362981e-01 1.09270871e+00
-4.62388545e-01 4.43545103e-01 -5.15302241e-01 -9.71892551e-02
-1.35458961e-01 8.20064247e-02 1.40069976e-01 8.65883291e-01
2.03141376e-01 4.16145742e-01 5.18858619e-02 1.99513942e-01
-1.98430359e-01 3.93830016e-02 -3.84569198e-01 6.62822962e-01
5.32559633e-01 1.38445973e+00 -4.66153264e-01 -9.72283632e-02
-4.78918463e-01 6.83026314e-01 4.21712071e-01 1.92732707e-01
-1.06112671e+00 -4.16345388e-01 7.67526507e-01 2.45957881e-01
1.21623412e-01 -2.03418911e-01 2.36192346e-01 -1.06265092e+00
-4.44534644e-02 -9.03695762e-01 2.58233380e-02 -7.96980977e-01
-6.08378887e-01 9.11791921e-01 4.22346853e-02 -9.89535987e-01
-4.43564057e-01 -5.91685295e-01 -6.13613486e-01 8.48180234e-01
-1.69273424e+00 -1.51305366e+00 -8.39298368e-01 7.13767946e-01
6.37175918e-01 -1.03699312e-01 7.65670419e-01 4.67414409e-01
-1.01043653e+00 1.12785447e+00 -5.75153291e-01 2.18772769e-01
9.22769308e-01 -6.87461734e-01 5.41094244e-01 6.93511963e-01
-3.70589167e-01 6.11526966e-01 5.83516479e-01 -3.83197635e-01
-1.38616824e+00 -9.64520395e-01 9.80629146e-01 -2.06051052e-01
1.62131920e-01 -3.69177073e-01 -8.15657079e-01 7.78684556e-01
1.88719973e-01 3.51337403e-01 4.84584302e-01 8.42371210e-02
-4.34387252e-02 -3.87840420e-01 -1.13702095e+00 5.10387182e-01
1.04899561e+00 -2.40575939e-01 -5.38174212e-01 1.09903842e-01
4.76683199e-01 -8.01985025e-01 -4.06216800e-01 4.07677859e-01
7.74651408e-01 -8.79576206e-01 7.31512964e-01 -2.77321905e-01
-1.44920349e-01 -3.02615851e-01 1.52178586e-01 -1.17496657e+00
-1.78450599e-01 -6.03942811e-01 -1.43239230e-01 1.16998124e+00
9.76753980e-03 -8.10616672e-01 9.75742996e-01 9.57199395e-01
3.76653597e-02 -9.02372003e-01 -1.01239288e+00 -2.88275987e-01
-3.81083190e-01 -3.14554483e-01 9.33760524e-01 4.80086774e-01
-1.44590467e-01 6.19094908e-01 -3.66121948e-01 2.35587269e-01
7.73439527e-01 -2.20507801e-01 1.15756679e+00 -1.36729300e+00
2.06172079e-01 -1.86774805e-01 -5.99573910e-01 -1.26772320e+00
6.16098404e-01 -2.54100889e-01 1.11792430e-01 -1.10323679e+00
2.67126948e-01 -2.32851803e-01 3.29129025e-02 6.43681765e-01
-2.20920891e-01 1.57027155e-01 2.02856019e-01 -2.95280502e-03
-6.70999706e-01 7.58570075e-01 1.02327144e+00 1.50188074e-01
-1.46706685e-01 2.50141978e-01 -3.83624047e-01 1.21024859e+00
4.77878898e-01 -1.92969427e-01 -4.99721736e-01 -5.35883844e-01
-1.78120881e-01 9.44037139e-02 1.51514009e-01 -1.14029431e+00
6.88059092e-01 5.25901839e-02 3.71767014e-01 -8.53163719e-01
6.18992329e-01 -8.65788221e-01 -2.87133068e-01 3.45944315e-01
1.60462841e-01 1.39560491e-01 2.74567515e-01 7.17813000e-02
-4.98654321e-02 -1.10814631e-01 1.04860425e+00 3.01814616e-01
-9.34141457e-01 5.87003529e-01 -3.45096514e-02 -2.70258412e-02
1.12181640e+00 3.05542685e-02 -3.53899181e-01 -5.75945795e-01
-3.17431450e-01 5.29548228e-01 2.16230288e-01 5.51234663e-01
7.25023687e-01 -1.18219781e+00 -5.51240504e-01 6.74979627e-01
1.38407513e-01 1.60888270e-01 2.15865105e-01 9.69982445e-01
-6.00450218e-01 4.59126711e-01 -1.85937837e-01 -6.73275411e-01
-1.65476704e+00 2.50248343e-01 3.90194356e-01 8.62543359e-02
-4.58618701e-01 9.61274028e-01 4.28424507e-01 -4.91439015e-01
6.77901268e-01 -1.29567474e-01 -6.67286992e-01 5.36132529e-02
9.77266908e-01 4.52666879e-01 1.95819616e-01 -1.09589970e+00
-6.07719123e-01 8.87762070e-01 -1.90235242e-01 1.31015942e-01
1.27458990e+00 -2.25476429e-01 -1.30650952e-01 -1.89137757e-01
1.23061037e+00 -2.31188655e-01 -1.08166778e+00 -2.37368315e-01
-2.08236665e-01 -3.54259253e-01 3.31086010e-01 -2.98029184e-01
-1.41588056e+00 9.47552145e-01 9.91574109e-01 -5.90896964e-01
1.19713354e+00 -1.49306566e-01 9.80692625e-01 4.10826623e-01
5.99987805e-01 -1.05901122e+00 -2.15232447e-01 5.87151766e-01
7.50098348e-01 -1.34487724e+00 7.55505040e-02 -4.66015756e-01
-5.28473139e-01 8.28137398e-01 1.06041849e+00 2.09955454e-01
8.81021738e-01 2.13312596e-01 2.14597777e-01 -1.36059999e-01
-4.69303012e-01 -4.59524393e-01 4.23683524e-01 4.58576262e-01
5.10821342e-01 -7.94211254e-02 -1.35074899e-01 6.09888911e-01
-3.64320815e-01 3.67889047e-01 2.52486885e-01 7.42061138e-01
-5.81422746e-01 -8.34092379e-01 -5.38839817e-01 1.46728620e-01
-3.41455013e-01 2.07359686e-01 5.82249910e-02 7.31368303e-01
5.44490367e-02 8.25794041e-01 5.01674712e-02 -4.31841612e-01
4.35273618e-01 1.92970827e-01 5.09825885e-01 -4.79422480e-01
-1.11941144e-01 1.94893077e-01 -1.61019683e-01 -7.03227878e-01
-3.84501249e-01 -4.12993103e-01 -1.19756103e+00 -6.29344225e-01
-5.15914857e-01 -1.27913002e-02 7.41766572e-01 1.07296050e+00
3.92652065e-01 3.82595658e-01 5.55885136e-01 -1.23575878e+00
-5.88990390e-01 -1.31003153e+00 -5.98500133e-01 1.27074391e-01
3.11562479e-01 -1.04954195e+00 -1.00965522e-01 -8.49772524e-03] | [13.653663635253906, 0.303142249584198] |
7bd374ba-3055-44d8-8510-4ebf96a42ba0 | mara-net-single-image-deraining-network-with | 2009.13990 | null | https://arxiv.org/abs/2009.13990v4 | https://arxiv.org/pdf/2009.13990v4.pdf | MCW-Net: Single Image Deraining with Multi-level Connections and Wide Regional Non-local Blocks | A recent line of convolutional neural network-based works has succeeded in capturing rain streaks. However, difficulties in detailed recovery still remain. In this paper, we present a multi-level connection and wide regional non-local block network (MCW-Net) to properly restore the original background textures in rainy images. Unlike existing encoder-decoder-based image deraining models that improve performance with additional branches, MCW-Net improves performance by maximizing information utilization without additional branches through the following two proposed methods. The first method is a multi-level connection that repeatedly connects multi-level features of the encoder network to the decoder network. Multi-level connection encourages the decoding process to use the feature information of all levels. In multi-level connection, channel-wise attention is considered to learn which level of features is important in the decoding process of the current level. The second method is a wide regional non-local block. As rain streaks primarily exhibit a vertical distribution, we divide the grid of the image into horizontally-wide patches and apply a non-local operation to each region to explore the rich rain-free background information. Experimental results on both synthetic and real-world rainy datasets demonstrate that the proposed model significantly outperforms existing state-of-the-art models. Furthermore, the results of the joint deraining and segmentation experiment prove that our model contributes effectively to other vision tasks. | ['Myungjoo Kang', 'Myeongho Jeon', 'Yeachan Park', 'Junho Lee'] | 2020-09-29 | null | null | null | null | ['single-image-deraining'] | ['computer-vision'] | [ 1.89683527e-01 -2.43933842e-01 7.60771036e-02 -4.78520989e-01
-4.19747561e-01 4.32329699e-02 1.14725865e-01 -4.18810427e-01
-2.79321641e-01 9.61767256e-01 -8.28482360e-02 -4.18730259e-01
2.50610352e-01 -1.04320598e+00 -9.69901741e-01 -1.11364508e+00
-1.05956852e-01 -3.66869271e-02 5.60553491e-01 -4.64175940e-01
1.46354377e-01 5.97942293e-01 -1.50499332e+00 4.87332702e-01
1.33175862e+00 8.44570518e-01 5.84946275e-01 7.38254845e-01
-3.28893036e-01 9.64935184e-01 -5.26049197e-01 9.23817083e-02
3.68973374e-01 -7.30022311e-01 -3.75284553e-01 2.59448618e-01
9.12049770e-01 -6.76692367e-01 -2.91209072e-01 1.10200155e+00
4.09725636e-01 -9.20803770e-02 3.01583230e-01 -6.13117397e-01
-5.65074325e-01 1.99679852e-01 -9.33041155e-01 8.30855966e-01
-2.57387817e-01 1.44577935e-01 5.56627154e-01 -9.27018881e-01
3.70315611e-01 1.09002841e+00 6.31333828e-01 3.05397481e-01
-8.93846214e-01 -7.46330917e-01 4.13670421e-01 2.37431020e-01
-1.05916357e+00 -1.31436130e-02 6.43909931e-01 -1.91588834e-01
7.83677638e-01 1.84190527e-01 1.04074407e+00 5.22869945e-01
4.20828998e-01 9.17815745e-01 1.41564786e+00 -3.02514136e-01
-3.48529629e-02 -1.80643544e-01 3.48390847e-01 8.59865844e-01
5.43640554e-01 2.10913911e-01 -2.12828547e-01 3.71616960e-01
1.20320928e+00 -1.73430219e-02 -7.06018090e-01 -4.86489236e-02
-7.55943358e-01 8.00671220e-01 9.13301706e-01 3.70477140e-01
-4.36965376e-01 9.00272280e-02 -2.58477330e-01 3.65408599e-01
8.41862500e-01 9.44848731e-02 -5.13734221e-01 3.82447958e-01
-1.50590694e+00 2.70085394e-01 7.13223338e-01 4.90846246e-01
1.12205470e+00 5.25220037e-01 -2.50321120e-01 8.47959876e-01
3.45688343e-01 7.94156790e-01 1.26772150e-01 -7.43153393e-01
6.28366470e-01 3.76379222e-01 1.60153076e-01 -7.85128295e-01
-2.55058140e-01 -7.35448241e-01 -1.18761313e+00 5.81337452e-01
3.73768926e-01 -1.10357791e-01 -1.46910715e+00 1.19749308e+00
1.59830332e-01 5.58812022e-01 1.41121432e-01 1.24504066e+00
9.89846468e-01 9.97322083e-01 -9.91528481e-02 -3.20885867e-01
1.19472909e+00 -1.25231576e+00 -7.17472911e-01 -4.91463125e-01
-8.37076977e-02 -6.70775175e-01 6.49217963e-01 2.57535219e-01
-1.12291694e+00 -8.54755640e-01 -1.01074767e+00 9.65027511e-02
-2.67714679e-01 2.21055120e-01 6.55459166e-01 3.53644371e-01
-1.11765683e+00 5.19091845e-01 -5.88905036e-01 -1.91494063e-01
6.74336195e-01 2.34815516e-02 -1.22815050e-01 -5.38349628e-01
-1.16967940e+00 8.83033812e-01 2.46139035e-01 8.78399074e-01
-1.13540876e+00 -4.80545491e-01 -7.53422737e-01 5.63088618e-02
-2.27006469e-02 -7.56669462e-01 6.89324439e-01 -1.38438654e+00
-1.29749119e+00 6.33618236e-01 -3.21546137e-01 -5.00290632e-01
2.60435104e-01 -5.53894341e-01 -2.97940791e-01 4.18689162e-01
-1.03931949e-01 7.27331936e-01 1.27579677e+00 -1.65803254e+00
-9.16479170e-01 -1.19650409e-01 4.13662335e-03 3.72858435e-01
2.13794857e-01 -3.90974373e-01 -6.41731381e-01 -8.42535675e-01
2.94448256e-01 -3.94718856e-01 -3.10465276e-01 4.42188494e-02
-1.24266073e-01 3.94719094e-01 8.43805611e-01 -1.06714749e+00
1.14700580e+00 -1.99949610e+00 8.52024555e-02 1.97129156e-02
2.49640465e-01 7.83886433e-01 -3.00300151e-01 1.56889353e-02
4.09180578e-03 -1.42946213e-01 -7.08028734e-01 -9.80426893e-02
-6.74012721e-01 6.46474481e-01 -3.81930083e-01 5.29397547e-01
4.84919876e-01 8.30040276e-01 -5.85402727e-01 -3.62511456e-01
3.92265826e-01 6.63002133e-01 -3.34285110e-01 4.07514632e-01
-1.27941966e-01 5.62066317e-01 -2.78569102e-01 8.62141371e-01
1.26775014e+00 5.93585819e-02 -2.78248370e-01 -2.34135807e-01
-2.20537230e-01 -1.08914681e-01 -8.78825128e-01 1.12974143e+00
-4.60172176e-01 7.42426157e-01 3.82198036e-01 -1.07922757e+00
9.03080285e-01 -5.66974171e-02 -7.20465183e-02 -8.12014163e-01
-4.18599322e-02 1.86677232e-01 7.51172975e-02 -8.86647820e-01
3.97187471e-01 -1.72247455e-01 7.37214744e-01 -1.72470197e-01
-1.42894059e-01 -8.47934708e-02 7.69701749e-02 6.24772087e-02
7.19028413e-01 2.77920336e-01 7.33764842e-02 -1.43588975e-01
5.75074792e-01 -8.92399102e-02 7.22976089e-01 1.07881916e+00
-2.54329026e-01 9.21168208e-01 1.51670799e-01 -6.38376415e-01
-8.89085233e-01 -9.15642440e-01 -3.47041816e-01 8.22680056e-01
4.84001607e-01 3.08404505e-01 -6.78020656e-01 -5.46479464e-01
-8.51696059e-02 5.47631443e-01 -8.13526273e-01 3.03007811e-01
-8.80536556e-01 -1.36248374e+00 4.35207814e-01 4.97964174e-01
1.23036313e+00 -1.31770730e+00 -6.75083518e-01 3.51439983e-01
-4.48756874e-01 -1.06577003e+00 -1.26773924e-01 4.62265819e-01
-1.24342000e+00 -1.03389001e+00 -9.04783607e-01 -1.02587807e+00
8.55006337e-01 6.75403297e-01 1.35406339e+00 3.34321350e-01
-5.22224665e-01 -9.38347131e-02 -5.58422029e-01 -7.21133798e-02
1.60446450e-01 -3.54335546e-01 -8.70609939e-01 9.14959535e-02
1.74906597e-01 -6.14125669e-01 -9.02587116e-01 8.66880342e-02
-9.88744140e-01 1.50169194e-01 9.62507248e-01 1.03552210e+00
6.17273867e-01 2.15540081e-02 3.55504751e-01 -8.71568322e-01
2.83501774e-01 -4.70321715e-01 -6.04417443e-01 5.60849421e-02
-2.68509090e-01 -1.30081743e-01 3.61419022e-01 -1.22995339e-01
-1.34391451e+00 -9.35191661e-02 -2.82234967e-01 -2.92995244e-01
-3.08185428e-01 3.35664570e-01 -1.22856520e-01 -2.28953630e-01
3.06540757e-01 7.05216408e-01 -2.39095077e-01 -4.21438754e-01
3.91025156e-01 4.60167766e-01 5.51813364e-01 -1.85603753e-01
8.90671015e-01 6.07291877e-01 -3.21472108e-01 -9.68634784e-01
-1.13263595e+00 -5.15234590e-01 -5.64029872e-01 -3.58186871e-01
1.03380919e+00 -1.21322799e+00 -2.53093094e-01 8.59121859e-01
-9.89837885e-01 -6.44789159e-01 -2.44009376e-01 2.88140088e-01
-4.61643934e-02 3.54170114e-01 -7.14239180e-01 -9.14738894e-01
-4.78383660e-01 -1.03057671e+00 9.79823232e-01 6.26318157e-01
6.60367727e-01 -9.32424963e-01 -1.06329583e-01 3.47729176e-01
6.66667998e-01 1.79870531e-01 6.24009967e-01 1.00990817e-01
-9.45939124e-01 1.59512341e-01 -6.70330882e-01 7.13065088e-01
3.19067314e-02 -6.31387755e-02 -9.26419199e-01 -3.18083465e-01
1.77686393e-01 -1.79573447e-01 1.79863906e+00 8.35192502e-01
9.85545337e-01 -2.21163362e-01 -1.34891570e-01 1.05241096e+00
1.75599575e+00 8.18036199e-02 1.16113734e+00 6.24262691e-01
7.89971769e-01 3.76717985e-01 5.38554430e-01 1.25249758e-01
4.26336378e-01 1.96555436e-01 9.47251916e-01 -8.33458722e-01
-5.51503301e-01 2.33147889e-01 4.15339500e-01 5.74200332e-01
-4.03414160e-01 -3.19113731e-01 -4.62089777e-01 7.52413154e-01
-1.73091137e+00 -1.05312324e+00 -4.30309594e-01 1.90074253e+00
8.43841493e-01 -6.25845557e-03 -4.32253748e-01 -1.82104528e-01
5.43274164e-01 4.53613639e-01 -4.19635653e-01 -2.45798886e-01
-5.84092379e-01 4.62941438e-01 6.29420817e-01 7.79024601e-01
-1.27146053e+00 1.13365519e+00 5.99894047e+00 7.85053194e-01
-1.13116932e+00 3.59451175e-02 7.29793787e-01 9.28281322e-02
-3.23982425e-02 -1.57745525e-01 -1.04175913e+00 4.74716902e-01
3.93962562e-01 7.18950808e-01 4.03953582e-01 4.22436506e-01
3.19044352e-01 -6.58879578e-01 -3.25553924e-01 5.99260271e-01
2.55933732e-01 -1.17718649e+00 1.72568291e-01 -1.04006156e-01
9.74758267e-01 3.06808978e-01 -1.23969309e-01 3.00276369e-01
2.55045593e-01 -1.11377144e+00 4.58372951e-01 7.23340571e-01
5.80961585e-01 -6.25791848e-01 1.10916519e+00 2.81820983e-01
-1.26375353e+00 -2.08984260e-02 -7.37306416e-01 2.27933079e-02
2.55199578e-02 1.06917799e+00 -2.50488043e-01 7.73412943e-01
1.06651616e+00 7.93029189e-01 -5.79932153e-01 1.24496078e+00
-5.99514902e-01 1.00988519e+00 -2.49289840e-01 5.98622084e-01
5.70764244e-01 -5.82513690e-01 5.22213340e-01 1.65254939e+00
1.82514116e-01 2.82304496e-01 4.53517698e-02 6.13078237e-01
1.54658750e-01 -2.69184649e-01 -4.79067534e-01 5.97919703e-01
-7.00702593e-02 1.15192974e+00 -7.20250726e-01 -5.46220362e-01
-4.02055204e-01 1.21063614e+00 3.30720901e-01 7.59800375e-01
-9.26459789e-01 -3.60875785e-01 6.96314573e-01 9.96402428e-02
9.82706189e-01 -2.03462481e-01 -3.35912406e-01 -1.16492617e+00
-1.08749755e-01 -8.57965708e-01 6.87495396e-02 -9.21352208e-01
-1.20386887e+00 8.06091011e-01 -2.10286960e-01 -1.18630815e+00
3.27136219e-01 -3.98151606e-01 -6.78910315e-01 1.20922911e+00
-2.56260586e+00 -1.26969182e+00 -7.67135382e-01 7.07378864e-01
8.32201242e-01 5.94814532e-02 4.09463823e-01 3.23566705e-01
-5.47833622e-01 1.28128245e-01 2.92936563e-01 1.51133850e-01
5.75944662e-01 -1.28866684e+00 1.58369362e-01 1.34755278e+00
-3.20327953e-02 2.80225813e-01 6.18869424e-01 -7.78713048e-01
-8.98929477e-01 -1.31557262e+00 7.48939037e-01 1.34565905e-01
2.00982064e-01 2.17982125e-03 -1.33472276e+00 4.28552002e-01
3.48934382e-01 4.07496512e-01 1.98472321e-01 -3.32340449e-01
-2.82453269e-01 -4.17470545e-01 -1.03753948e+00 1.64219603e-01
6.15147948e-01 2.75739394e-02 -6.99460924e-01 1.89245209e-01
3.89142692e-01 -5.12903452e-01 -2.13645697e-01 5.91725886e-01
3.43033224e-01 -1.31161404e+00 1.04620886e+00 -2.10970536e-01
8.31430137e-01 -6.44911349e-01 -1.73550531e-01 -1.39753795e+00
-4.30580586e-01 -1.32851124e-01 -1.83511347e-01 8.66259933e-01
1.20642491e-01 -5.42281091e-01 5.64713120e-01 -3.45004588e-01
-2.68483549e-01 -8.25097561e-01 -7.93482542e-01 -4.21364427e-01
6.87143207e-02 -7.08981305e-02 1.16672933e-01 7.58583128e-01
-8.64110112e-01 1.10457025e-01 -7.85816491e-01 6.54382885e-01
8.74951184e-01 6.12775207e-01 5.43819904e-01 -1.01010573e+00
-2.29230568e-01 -2.46080875e-01 -9.53303929e-03 -1.53126371e+00
-1.72170907e-01 -5.44614434e-01 5.15911102e-01 -2.01469040e+00
1.30771428e-01 -2.50558227e-01 -2.96035111e-01 6.26751423e-01
-6.51961267e-01 7.05520988e-01 1.81767195e-01 3.90462160e-01
-3.63683343e-01 6.93139791e-01 1.79388893e+00 -1.77731454e-01
-2.04363376e-01 5.77943139e-02 -4.57137525e-01 8.85399342e-01
8.59631419e-01 -4.37240452e-01 -6.25996739e-02 -6.77020490e-01
-1.93208233e-01 5.56959845e-02 4.57014531e-01 -1.00601673e+00
5.51505946e-02 -1.31868705e-01 8.52189124e-01 -8.21609974e-01
2.92727172e-01 -7.39401519e-01 -3.54912341e-01 5.70122421e-01
1.10568427e-01 -4.32865769e-01 2.21637353e-01 7.00443566e-01
-6.81124151e-01 -3.23635042e-02 1.25605643e+00 -2.40197718e-01
-8.24028313e-01 2.09972516e-01 -6.87921226e-01 -1.51198000e-01
7.38548636e-01 -4.81044561e-01 -5.16138673e-01 -2.88855493e-01
-8.33009779e-01 2.61131942e-01 -6.72900537e-03 1.15170188e-01
8.44482660e-01 -6.74019098e-01 -1.02387190e+00 5.19571126e-01
-3.49318564e-01 1.54933900e-01 3.20230395e-01 9.34003830e-01
-1.06445730e+00 1.35180220e-01 -4.81376410e-01 -5.77650428e-01
-1.18277228e+00 -1.74040124e-02 6.95272923e-01 -4.49921489e-01
-1.07130122e+00 1.05967259e+00 6.47371531e-01 -1.14599534e-03
1.42658785e-01 -4.66095835e-01 -5.41997969e-01 8.19414761e-03
6.35692239e-01 1.47322953e-01 -5.43155484e-02 -4.91626024e-01
-5.13268672e-02 7.67473638e-01 -1.40546218e-01 1.94391742e-01
1.55467069e+00 -2.96446711e-01 -4.00463402e-01 2.48331964e-01
5.90917826e-01 -1.82622343e-01 -1.79794824e+00 -2.34838486e-01
-6.53247535e-01 -6.22821212e-01 3.20307106e-01 -1.02210844e+00
-1.74807858e+00 1.09341609e+00 8.06376934e-01 -6.92427158e-02
1.61010945e+00 -4.27467048e-01 7.44748235e-01 2.58687526e-01
7.24274740e-02 -5.69748819e-01 5.08926660e-02 8.51982474e-01
7.41964638e-01 -1.26325810e+00 8.43210295e-02 -4.93519187e-01
-5.75481355e-01 1.22171926e+00 9.08841908e-01 -4.56708223e-01
7.65107870e-01 4.30880904e-01 3.61253738e-01 -2.43480906e-01
-3.84636223e-01 -6.17100179e-01 -1.65958505e-03 6.31237864e-01
2.34442517e-01 -1.89772904e-01 -3.41180503e-01 2.56348193e-01
2.62947351e-01 4.44967933e-02 5.73025286e-01 8.76148999e-01
-1.13598812e+00 -6.91311121e-01 -7.62769639e-01 4.07094777e-01
-3.72986794e-01 -6.54138565e-01 1.04570843e-01 8.45787644e-01
5.46425819e-01 1.01474071e+00 1.90816239e-01 1.16789073e-01
6.66816980e-02 -2.48502553e-01 5.21204829e-01 -3.59565616e-01
-7.19925284e-01 3.78918350e-01 -1.35428295e-01 -5.41636467e-01
-8.06355834e-01 -2.60597736e-01 -1.03824222e+00 -2.90485639e-02
-3.11448872e-01 7.01989233e-02 4.27601397e-01 9.68852937e-01
1.77110489e-02 9.33070958e-01 5.14426887e-01 -1.12640834e+00
1.19616464e-02 -1.08563089e+00 -8.73204350e-01 2.27669552e-02
8.00336242e-01 -3.72854650e-01 -4.27266121e-01 3.43981355e-01] | [10.909379005432129, -3.227968215942383] |
2d61f716-987f-46b9-b5f2-054c50cf558c | building-change-detection-for-remote-sensing | 1909.07726 | null | https://arxiv.org/abs/1909.07726v1 | https://arxiv.org/pdf/1909.07726v1.pdf | Building Change Detection for Remote Sensing Images Using a Dual Task Constrained Deep Siamese Convolutional Network Model | In recent years, building change detection methods have made great progress by introducing deep learning, but they still suffer from the problem of the extracted features not being discriminative enough, resulting in incomplete regions and irregular boundaries. To tackle this problem, we propose a dual task constrained deep Siamese convolutional network (DTCDSCN) model, which contains three sub-networks: a change detection network and two semantic segmentation networks. DTCDSCN can accomplish both change detection and semantic segmentation at the same time, which can help to learn more discriminative object-level features and obtain a complete change detection map. Furthermore, we introduce a dual attention module (DAM) to exploit the interdependencies between channels and spatial positions, which improves the feature representation. We also improve the focal loss function to suppress the sample imbalance problem. The experimental results obtained with the WHU building dataset show that the proposed method is effective for building change detection and achieves a state-of-the-art performance in terms of four metrics: precision, recall, F1-score, and intersection over union. | ['Yi Liu', 'Xue Yang', 'Chao Pang', 'Zongqian Zhan', 'Xiaomeng Zhang'] | 2019-09-17 | null | null | null | null | ['change-detection-for-remote-sensing-images', 'building-change-detection-for-remote-sensing', 'extracting-buildings-in-remote-sensing-images'] | ['miscellaneous', 'miscellaneous', 'miscellaneous'] | [ 1.81523770e-01 -4.87835824e-01 -6.13733232e-02 -3.31504494e-01
-8.19912851e-01 -4.51521240e-02 4.08013761e-01 1.96220666e-01
-3.46680969e-01 4.12437260e-01 1.48475841e-01 2.47014016e-01
3.61860506e-02 -1.07776499e+00 -6.54571831e-01 -8.23510647e-01
-1.55015383e-03 -2.91826520e-02 7.55919695e-01 -1.72613636e-01
1.07377395e-01 2.33917519e-01 -1.60755789e+00 6.06730431e-02
1.23934066e+00 1.28774667e+00 3.07620794e-01 5.21620065e-02
-5.97259253e-02 5.03842592e-01 -3.85620415e-01 1.75043732e-01
1.65923581e-01 -3.52995694e-01 -5.48522115e-01 1.75307766e-01
2.40190312e-01 -3.92937332e-01 -5.05756259e-01 1.13648939e+00
6.69646740e-01 5.01125529e-02 3.87661010e-01 -1.05373836e+00
-5.02241194e-01 2.88876444e-01 -8.93175423e-01 4.40529674e-01
1.35322481e-01 1.90539196e-01 1.18298054e+00 -8.25671375e-01
2.66608119e-01 1.24857926e+00 7.95759737e-01 -7.14180851e-03
-1.00607955e+00 -8.64369810e-01 5.79432011e-01 4.49630022e-01
-1.60557246e+00 -1.05137460e-01 1.11843657e+00 -3.93690169e-01
6.66507781e-01 2.98270375e-01 8.46951246e-01 5.93832910e-01
-1.80257693e-01 1.35457110e+00 7.93731809e-01 -2.03787893e-01
1.54391557e-01 -2.94924051e-01 1.46183923e-01 8.07765007e-01
1.15438491e-01 -3.45363885e-01 -1.35918617e-01 2.94910640e-01
9.09983277e-01 3.77394140e-01 -5.87356329e-01 -5.01422524e-01
-1.13966262e+00 7.97210932e-01 1.10073698e+00 7.48077273e-01
-1.77600324e-01 6.87001497e-02 4.36061889e-01 -4.76921611e-02
6.16744220e-01 5.29284291e-02 -4.60998029e-01 1.63288489e-01
-9.51994658e-01 2.41262242e-01 2.08280057e-01 7.45150745e-01
9.34296012e-01 -3.12214553e-01 -5.68416238e-01 1.02646315e+00
2.93481410e-01 4.69583899e-01 5.96181214e-01 -5.96674323e-01
7.20538497e-01 1.07463944e+00 -1.36885881e-01 -1.19399250e+00
-4.72785711e-01 -6.88530803e-01 -1.04148996e+00 -1.81623697e-01
-5.92807159e-02 2.18169540e-01 -1.17548454e+00 1.71602142e+00
4.18568999e-01 1.44895256e-01 -5.15417576e-01 8.30491900e-01
7.52224147e-01 6.06855571e-01 -8.44810158e-02 -3.66923697e-02
1.14770925e+00 -1.19163644e+00 -7.33343720e-01 -3.59878331e-01
5.18873334e-01 -4.26443756e-01 1.25418508e+00 2.87455656e-02
-6.78855360e-01 -6.12800360e-01 -1.26008582e+00 -7.04520792e-02
-3.23213518e-01 2.42636308e-01 4.93649960e-01 2.55703390e-01
-5.98009706e-01 4.74397391e-01 -1.08151281e+00 -3.35450768e-01
9.27495837e-01 2.68397391e-01 2.83253472e-02 -3.83084893e-01
-1.10093486e+00 3.16481084e-01 5.10853946e-01 3.21752459e-01
-7.89495409e-01 -5.44483006e-01 -8.35388124e-01 3.02401453e-01
4.38887805e-01 -4.64839995e-01 8.67605209e-01 -9.22615230e-01
-1.11622775e+00 6.65844560e-01 4.79274392e-02 -5.81616871e-02
5.44314206e-01 -2.08033651e-01 -4.33107585e-01 -2.73606963e-02
3.76569450e-01 5.82171142e-01 4.41718876e-01 -1.17726409e+00
-1.04731560e+00 -6.25645697e-01 -7.76808187e-02 2.21589729e-01
-4.22168255e-01 -3.24354172e-01 -9.53025699e-01 -9.21809137e-01
4.78762448e-01 -7.93351471e-01 -2.61466861e-01 1.07065655e-01
-4.56974655e-01 -3.36880237e-01 1.14807510e+00 -1.08902252e+00
1.67664683e+00 -2.29832506e+00 1.12809479e-01 2.78998345e-01
-2.08871858e-03 3.73631597e-01 -1.25932947e-01 2.37834291e-03
9.78036132e-03 5.95368408e-02 -7.63194501e-01 -2.65047342e-01
-8.26319754e-02 1.77912146e-01 3.02290887e-01 4.23920125e-01
1.67445481e-01 8.04231286e-01 -8.28601718e-01 -5.09619117e-01
3.22633624e-01 2.94506729e-01 -5.37939787e-01 6.43070042e-02
-2.08197713e-01 4.88453865e-01 -8.07544887e-01 6.88123107e-01
8.44440997e-01 -1.79803640e-01 -1.81143895e-01 -1.24744430e-01
-1.26536131e-01 1.22269936e-01 -1.51153207e+00 1.95779800e+00
-4.06051487e-01 4.20738637e-01 -2.88009383e-02 -1.11179090e+00
8.15602124e-01 -2.06741512e-01 6.70941830e-01 -1.02505529e+00
-1.61855929e-02 1.63242891e-01 -2.80241936e-01 -6.00940764e-01
1.54002875e-01 2.00096726e-01 -2.24510014e-01 -1.46937929e-03
-4.08248216e-01 -2.84523889e-02 1.27838358e-01 -7.15900660e-02
1.16944385e+00 -9.48765967e-03 5.98941036e-02 -2.18539804e-01
7.37511575e-01 -2.46110752e-01 1.05367196e+00 3.68278891e-01
-2.37216040e-01 7.39646971e-01 2.80813962e-01 -3.75161946e-01
-7.19133556e-01 -9.16919529e-01 -1.27178952e-01 6.18597865e-01
5.81286967e-01 -1.81704655e-01 -7.33180821e-01 -8.36424410e-01
2.67275609e-02 5.75579226e-01 -5.70047379e-01 -3.86448860e-01
-8.76747608e-01 -8.89409006e-01 1.80310979e-01 8.01674783e-01
1.19922221e+00 -9.19133484e-01 -4.68916506e-01 3.82084042e-01
-5.93007028e-01 -9.25529957e-01 -6.40217841e-01 -6.49822061e-04
-7.11643040e-01 -1.09464777e+00 -7.79157460e-01 -9.61246908e-01
5.95240772e-01 4.41907227e-01 7.68104792e-01 1.04411140e-01
-3.76581073e-01 1.83059741e-02 -4.29771960e-01 -8.63660052e-02
2.01632455e-01 3.57558995e-01 -5.48157036e-01 6.38351366e-02
2.28376046e-01 -5.91905236e-01 -9.28940296e-01 3.84794235e-01
-1.05801761e+00 2.44616400e-02 6.58339858e-01 7.74508834e-01
7.31630325e-01 4.05720562e-01 5.15412569e-01 -4.64449227e-01
2.39842385e-01 -2.94932455e-01 -3.49809438e-01 3.41325343e-01
-7.20578730e-01 -1.38189927e-01 3.61731887e-01 9.99950431e-03
-1.13435912e+00 7.08871707e-02 -3.41932774e-01 -7.07246289e-02
-7.75267258e-02 4.31323230e-01 -6.70111418e-01 1.05804339e-01
1.70952871e-01 4.48827416e-01 -3.32769990e-01 -7.08071709e-01
5.02516665e-02 7.61503696e-01 4.09265667e-01 -1.56556323e-01
7.72700608e-01 6.42994106e-01 -3.62459391e-01 -6.02093458e-01
-1.05978441e+00 -7.88398802e-01 -6.29143953e-01 -8.83451253e-02
9.23626542e-01 -1.26267910e+00 -1.23382151e-01 1.03318739e+00
-8.36605132e-01 -1.03019446e-01 -3.79796952e-01 3.18877190e-01
-2.80185372e-01 4.86552715e-01 -5.00017881e-01 -5.17855048e-01
-4.57281083e-01 -1.13319278e+00 1.28984952e+00 5.13627827e-01
3.62513483e-01 -5.91681659e-01 -5.55497706e-02 3.97073269e-01
3.53301078e-01 5.38199365e-01 1.00706458e+00 -2.89655328e-01
-9.72882569e-01 -2.23397017e-01 -4.64426875e-01 5.28014243e-01
3.19552600e-01 -2.39643380e-01 -7.22597599e-01 -4.46091563e-01
-1.24647401e-01 2.04205699e-02 1.28201413e+00 4.25598294e-01
1.43253016e+00 -1.02335922e-02 -5.95902383e-01 6.50919735e-01
1.66169012e+00 2.13153467e-01 8.03661704e-01 5.76775849e-01
9.51945662e-01 2.00114489e-01 6.79261804e-01 5.05517781e-01
6.24495387e-01 7.96751797e-01 6.26215100e-01 -2.60997802e-01
-2.89095044e-01 -2.44850650e-01 1.16823904e-01 8.27346802e-01
2.32008487e-01 6.88024908e-02 -8.50178540e-01 9.11828756e-01
-1.98361957e+00 -7.05848038e-01 -2.05407679e-01 1.99618685e+00
7.42962062e-01 2.18442515e-01 7.40868747e-02 3.86298388e-01
9.14963007e-01 3.29171747e-01 -8.06373537e-01 4.15258944e-01
-1.04025207e-01 1.83179770e-02 3.05784017e-01 2.29813904e-02
-1.42821157e+00 7.07767665e-01 5.23378801e+00 1.14901614e+00
-1.01070476e+00 2.09173828e-01 6.15846276e-01 1.03378989e-01
-3.77181292e-01 -2.18084216e-01 -6.99369907e-01 8.10047507e-01
2.83672158e-02 3.10608059e-01 2.18644321e-01 7.83032238e-01
1.85876116e-01 -2.30320707e-01 -6.47899806e-01 1.01887178e+00
1.55973703e-01 -1.00296962e+00 -1.67023525e-01 -2.30292231e-01
9.58212912e-01 1.87138826e-01 -1.71843633e-01 5.01219869e-01
1.21295430e-01 -5.27772725e-01 6.55170202e-01 4.54065591e-01
6.09580100e-01 -8.80575120e-01 7.90862203e-01 3.86474460e-01
-1.60849154e+00 -3.74512136e-01 -1.59015745e-01 1.33965760e-01
4.91071939e-02 8.67765248e-01 -2.69707501e-01 8.49531651e-01
1.03895068e+00 1.00585186e+00 -7.88044035e-01 1.41851425e+00
-2.87425697e-01 5.35027206e-01 -5.65887153e-01 1.28433481e-01
3.70041788e-01 -1.98912367e-01 4.02675211e-01 1.07231927e+00
3.64805400e-01 -2.56406903e-01 5.48603296e-01 8.16218317e-01
-5.39628305e-02 1.50940344e-01 6.48736432e-02 3.09125245e-01
3.86043340e-01 1.10991323e+00 -1.05241144e+00 -1.47025958e-01
-4.44370002e-01 1.14260328e+00 2.17102021e-01 1.98698148e-01
-8.95273209e-01 -7.14888632e-01 5.58098435e-01 8.03337991e-02
8.40946078e-01 -6.82385042e-02 -3.32355529e-01 -1.26820409e+00
4.12934214e-01 -6.07625127e-01 4.20826674e-01 -3.94651771e-01
-9.91282642e-01 1.23247579e-01 -1.75938591e-01 -1.14904356e+00
4.18610483e-01 -2.63631213e-02 -7.53690541e-01 6.10814869e-01
-1.73738790e+00 -1.41476691e+00 -7.87634552e-01 5.41055918e-01
7.40167975e-01 3.27618361e-01 2.41969958e-01 9.01312113e-01
-1.15854692e+00 5.07995009e-01 3.81074071e-01 3.58294249e-01
4.73515034e-01 -1.14360559e+00 3.35153908e-01 1.10362864e+00
-1.89070448e-01 2.42474489e-02 2.90253371e-01 -5.63443601e-01
-7.44797051e-01 -1.46627665e+00 3.95312011e-01 1.47445336e-01
2.53547758e-01 -2.57756919e-01 -8.69466782e-01 5.00855923e-01
-2.97391504e-01 5.64944558e-02 2.28604406e-01 5.38148805e-02
-1.27178624e-01 -4.87026423e-01 -1.15431988e+00 2.50911921e-01
1.28147984e+00 -3.11634213e-01 -3.89327496e-01 2.58544922e-01
9.10156131e-01 -2.33252421e-01 -6.79404914e-01 6.51074350e-01
2.54073858e-01 -9.43599105e-01 8.18135202e-01 2.59314608e-02
2.67616272e-01 -6.04946971e-01 -2.26565123e-01 -1.18979502e+00
-6.25422120e-01 8.82526115e-02 1.96823414e-04 1.60940480e+00
-6.77493354e-03 -3.93133163e-01 6.66779399e-01 1.47875994e-01
-3.89573336e-01 -9.88643646e-01 -1.06959808e+00 -8.31758499e-01
-1.71289459e-01 -3.23139340e-01 8.01224113e-01 8.99594545e-01
-5.59376121e-01 4.15726006e-01 -5.32240532e-02 1.70176446e-01
3.56596708e-01 3.80875409e-01 5.14952481e-01 -1.24469733e+00
2.70326436e-02 -5.26044905e-01 -4.87513334e-01 -1.33699596e+00
-1.00798845e-01 -7.45633423e-01 2.70424426e-01 -1.99748540e+00
4.19487238e-01 -6.70563459e-01 -4.98098850e-01 4.80804652e-01
-5.64684749e-01 -8.72947741e-03 -4.52188104e-02 3.27054471e-01
-7.72481799e-01 1.04808879e+00 1.38695335e+00 -4.84980583e-01
-2.80275226e-01 3.19126733e-02 -6.05619490e-01 7.11634994e-01
5.72831988e-01 -4.08566654e-01 -1.54267207e-01 -4.41839486e-01
-1.40998244e-01 -4.26294118e-01 4.38231379e-01 -1.46407044e+00
1.07600819e-02 1.15344487e-02 4.52157676e-01 -9.62386012e-01
-4.20210250e-02 -8.93113256e-01 1.62310582e-02 7.23818481e-01
1.45246014e-02 -3.08045566e-01 1.78970397e-02 7.94639289e-01
-3.91712487e-01 -5.89078255e-02 1.00077271e+00 2.61491369e-02
-1.04510105e+00 5.64248145e-01 7.23873405e-03 5.61822429e-02
1.17532682e+00 -5.84568009e-02 -8.84164572e-02 -3.48572060e-02
-3.77419114e-01 6.35729551e-01 4.41679716e-01 4.99983042e-01
3.88032347e-01 -1.62674153e+00 -6.02970839e-01 2.89409965e-01
3.08322638e-01 5.87088466e-01 5.97917557e-01 7.39070535e-01
-5.73949575e-01 1.22326128e-01 -3.93064739e-03 -7.76649892e-01
-9.62312281e-01 4.78746176e-01 3.83349478e-01 -5.43954551e-01
-7.97321618e-01 7.93246746e-01 2.99589992e-01 -5.06701708e-01
1.64327949e-01 -5.94724953e-01 -2.01224849e-01 1.55564040e-01
3.33649993e-01 4.91916269e-01 1.56208336e-01 -4.87677783e-01
-5.24006546e-01 8.01657796e-01 -7.87997395e-02 5.23324549e-01
1.42996371e+00 -3.93576711e-01 3.31931598e-02 2.54805714e-01
1.29779768e+00 -3.00504476e-01 -1.39887464e+00 -5.47115147e-01
-8.17025825e-02 -4.81169045e-01 2.65431136e-01 -7.02128589e-01
-1.41182065e+00 7.24307299e-01 1.20290089e+00 2.35226169e-01
1.39955223e+00 1.19829737e-02 1.37804079e+00 1.66142508e-01
2.37670332e-01 -1.41174114e+00 2.91995525e-01 3.92045617e-01
8.13163459e-01 -1.36404860e+00 4.28984733e-03 -5.52750289e-01
-2.55499691e-01 7.74821699e-01 8.01223516e-01 -1.59688234e-01
8.08757842e-01 -8.72808471e-02 -3.50826412e-01 -1.40724376e-01
-4.86826012e-03 -4.85511035e-01 1.78016812e-01 3.43001693e-01
2.25794297e-02 1.20267913e-01 -4.03410017e-01 6.43175006e-01
3.44690353e-01 -1.20811872e-01 -6.70253038e-02 9.77961779e-01
-7.34831691e-01 -8.29049349e-01 -2.22707674e-01 6.30351305e-01
-2.87165999e-01 1.12770393e-01 -2.45789997e-02 8.78877461e-01
5.09348691e-01 7.43282914e-01 1.55067623e-01 -4.23569888e-01
6.46939278e-01 -2.96845257e-01 2.57143468e-01 -4.49872971e-01
-3.85824293e-01 1.26786381e-01 -2.12832183e-01 -5.39052606e-01
-5.40949464e-01 -7.68674791e-01 -1.25796902e+00 4.30256426e-02
-7.15399444e-01 -4.04428951e-02 3.94785315e-01 9.71904218e-01
3.12947601e-01 9.40063417e-01 9.49348032e-01 -5.90266883e-01
-3.94846231e-01 -1.12051976e+00 -6.79273367e-01 4.84627396e-01
2.35592708e-01 -7.12614000e-01 -2.14524701e-01 -1.67437539e-01] | [9.641438484191895, -1.0827559232711792] |
8ce2fe1f-71b1-4634-b6ea-1732aee8beb8 | self-supervised-representation-learning-from | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Li_Self-Supervised_Representation_Learning_From_Videos_for_Facial_Action_Unit_Detection_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Li_Self-Supervised_Representation_Learning_From_Videos_for_Facial_Action_Unit_Detection_CVPR_2019_paper.pdf | Self-Supervised Representation Learning From Videos for Facial Action Unit Detection | In this paper, we aim to learn discriminative representation for facial action unit (AU) detection from large amount of videos without manual annotations. Inspired by the fact that facial actions are the movements of facial muscles, we depict the movements as the transformation between two face images in different frames and use it as the self-supervisory signal to learn the representations. However, under the uncontrolled condition, the transformation is caused by both facial actions and head motions. To remove the influence by head motions, we propose a Twin-Cycle Autoencoder (TCAE) that can disentangle the facial action related movements and the head motion related ones. Specifically, TCAE is trained to respectively change the facial actions and head poses of the source face to those of the target face. Our experiments validate TCAE's capability of decoupling the movements. Experimental results also demonstrate that the learned representation is discriminative for AU detection, where TCAE outperforms or is comparable with the state-of-the-art self-supervised learning methods and supervised AU detection methods.
| [' Xilin Chen', ' Shiguang Shan', ' Jiabei Zeng', 'Yong Li'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 1.14565976e-01 2.74155289e-01 -2.19933614e-01 -3.97074312e-01
-2.07937032e-01 -1.44010961e-01 6.79870963e-01 -8.04190159e-01
-2.08373070e-01 3.10663134e-01 4.69888568e-01 6.23958409e-01
2.69496411e-01 -4.04476434e-01 -7.60280073e-01 -1.18571436e+00
2.24948321e-02 -5.94801269e-02 -9.46155190e-02 -1.16136834e-01
-1.69369817e-01 3.83740306e-01 -1.81544709e+00 2.01337636e-01
1.56915709e-01 1.01437831e+00 -2.22228661e-01 3.31879526e-01
4.06452268e-01 1.04563332e+00 -4.13804919e-01 -5.33491075e-02
2.99292356e-01 -8.31749678e-01 -6.09551132e-01 5.34054995e-01
6.29296899e-01 -6.66382134e-01 -7.04584718e-01 1.15512753e+00
5.02608538e-01 5.43916635e-02 7.01376617e-01 -1.43107557e+00
-4.60020721e-01 1.54675037e-01 -8.89657617e-01 1.57123253e-01
3.73544127e-01 -2.37622618e-04 7.82046497e-01 -9.15755868e-01
8.27483714e-01 1.41047311e+00 2.78151989e-01 1.10709572e+00
-1.01207173e+00 -8.84696424e-01 1.48381829e-01 4.44011718e-01
-1.42903996e+00 -1.06041288e+00 1.10418272e+00 -4.88032579e-01
5.61729550e-01 5.55886999e-02 6.33388162e-01 1.44609618e+00
7.19108209e-02 1.01198173e+00 7.27921903e-01 -4.46163744e-01
-6.86902087e-03 -1.45145908e-01 -3.20028067e-01 8.76591146e-01
-1.36944279e-01 3.09243172e-01 -5.04679084e-01 9.29078981e-02
8.79875779e-01 4.61211763e-02 -4.74968612e-01 -3.67625803e-01
-1.00952387e+00 7.35682964e-01 1.43259928e-01 4.78973210e-01
-4.78824884e-01 1.94587007e-01 3.68553132e-01 1.97961330e-01
2.53711313e-01 -2.65878618e-01 -1.63883179e-01 1.50447249e-01
-7.85334170e-01 -1.00267783e-01 5.20929396e-01 7.68691242e-01
6.93031132e-01 4.30784225e-01 -1.57913029e-01 5.04261196e-01
4.08411831e-01 2.98572361e-01 8.90537918e-01 -9.45919335e-01
-7.70808384e-02 5.22751272e-01 -1.36023119e-01 -1.12191820e+00
-3.47119778e-01 4.85277660e-02 -7.66803384e-01 1.06485248e-01
3.56418222e-01 -2.69853860e-01 -5.64902365e-01 1.95938945e+00
6.77726567e-01 5.03348053e-01 -3.61814126e-02 1.00915670e+00
9.86641407e-01 4.65780586e-01 -3.71866450e-02 -7.80905843e-01
1.18758571e+00 -9.15275395e-01 -1.17534244e+00 -2.62247264e-01
6.34938359e-01 -5.17857909e-01 3.30124497e-01 1.41211525e-01
-8.04099143e-01 -7.35603869e-01 -8.50283980e-01 1.55221090e-01
1.39119163e-01 6.03414178e-01 3.50150526e-01 1.46982312e-01
-8.01941693e-01 4.54491377e-01 -1.03195119e+00 -3.72785091e-01
4.16639239e-01 3.99291158e-01 -7.54308701e-01 3.33297133e-01
-1.07390332e+00 7.38616705e-01 2.05419675e-01 2.93383121e-01
-1.17836368e+00 -1.92899659e-01 -1.00468576e+00 -1.73221454e-01
5.11208832e-01 -1.88219056e-01 1.06960487e+00 -1.61688638e+00
-1.68289185e+00 1.02914000e+00 -1.85037956e-01 -1.54132843e-01
3.81523103e-01 -2.39464968e-01 -5.19421101e-01 3.70780170e-01
-3.81608680e-02 8.11113715e-01 1.56515074e+00 -9.20940936e-01
-4.64820802e-01 -5.97544551e-01 -3.17866653e-01 1.04401149e-01
-4.64918375e-01 2.53750920e-01 -5.82806468e-01 -6.01385474e-01
-3.21218297e-02 -1.08602297e+00 3.90492737e-01 3.14364940e-01
-1.34292409e-01 -4.41400886e-01 1.14550757e+00 -7.80956924e-01
1.06602550e+00 -2.50831652e+00 4.93599415e-01 -1.48768023e-01
2.44956002e-01 2.20487744e-01 -2.49537751e-01 5.78766465e-02
-3.81810099e-01 -5.15635788e-01 1.69245809e-01 -4.13155943e-01
-2.88412839e-01 4.89214152e-01 3.17376829e-03 9.76976395e-01
2.70554543e-01 7.56877124e-01 -7.82786012e-01 -6.26355708e-01
5.82613871e-02 5.71160257e-01 -3.28612119e-01 4.96409833e-01
1.60681069e-01 6.81568146e-01 -4.77942824e-01 6.48603797e-01
4.59137619e-01 1.26151383e-01 2.95778513e-01 -5.84927857e-01
1.96759343e-01 -2.04162509e-03 -1.07818723e+00 1.44997537e+00
6.55145347e-02 6.64265811e-01 2.70905405e-01 -1.21149647e+00
8.77104282e-01 6.72433496e-01 5.71632862e-01 -5.52699208e-01
7.12154865e-01 -2.18838558e-01 1.61466286e-01 -7.93231189e-01
-3.66989821e-01 -3.51759791e-02 4.54697162e-01 4.82009888e-01
5.43706238e-01 3.14151317e-01 1.04785971e-01 -2.43765578e-01
7.24715948e-01 3.66388857e-01 6.97745204e-01 -9.07655731e-02
7.49862790e-01 -7.18195617e-01 9.26309884e-01 -5.21141030e-02
-7.31738091e-01 3.88961911e-01 4.66586858e-01 -4.73045737e-01
-4.98198450e-01 -5.24558544e-01 6.29900172e-02 1.25743389e+00
2.51819361e-02 -2.05653116e-01 -1.15567040e+00 -8.65140796e-01
-8.64771828e-02 1.24837093e-01 -1.10458112e+00 -5.37279069e-01
-6.38897717e-01 -4.82487887e-01 4.28875208e-01 7.72775650e-01
4.74275857e-01 -1.17925596e+00 -6.87804163e-01 -7.22583523e-03
-2.09787324e-01 -1.38717401e+00 -7.91012406e-01 -2.19029173e-01
-6.95329964e-01 -1.32578087e+00 -5.03491759e-01 -8.91845822e-01
7.86709428e-01 2.58728683e-01 4.29031521e-01 -1.99521761e-02
-4.00301069e-01 4.17616010e-01 -2.87247747e-01 -2.90234476e-01
-3.59322935e-01 -7.11171687e-01 5.61398327e-01 8.53196800e-01
6.03429317e-01 -6.14398420e-01 -6.10085309e-01 3.74192595e-01
-7.51158476e-01 -1.34599343e-01 4.06053960e-01 8.82274151e-01
4.26179707e-01 -1.04462855e-01 2.51939297e-01 -3.83820683e-01
2.47807875e-01 -2.60291606e-01 -3.12041879e-01 7.52742290e-02
-2.19505981e-01 2.92021921e-03 3.08641106e-01 -7.88882613e-01
-9.73505020e-01 5.21750212e-01 1.16084412e-01 -9.64712560e-01
-4.26110923e-01 -8.24772418e-02 -3.12796116e-01 -2.22120747e-01
3.94072682e-01 4.15191650e-01 4.04306263e-01 -3.19928080e-01
1.96818277e-01 5.94609380e-01 7.26970911e-01 -1.90807760e-01
7.22371399e-01 8.34424019e-01 -1.44499689e-01 -8.57419968e-01
-7.39229500e-01 -2.95299768e-01 -1.02095640e+00 -5.43238759e-01
1.20066333e+00 -1.01854956e+00 -7.37418056e-01 4.32571411e-01
-1.14480436e+00 6.59857644e-03 4.76130582e-02 5.92504025e-01
-6.37184680e-01 5.14570057e-01 -6.57304704e-01 -7.37994850e-01
-3.84609342e-01 -1.03289592e+00 1.24161279e+00 2.49496534e-01
-2.33850658e-01 -8.20182085e-01 1.38230160e-01 1.71934873e-01
-2.01088026e-01 4.93416607e-01 4.78523403e-01 -4.59920555e-01
-1.35277197e-01 -3.24361980e-01 2.25237086e-01 5.32417893e-01
4.76934403e-01 2.33758226e-01 -1.16001940e+00 -4.48141575e-01
3.31969768e-01 -3.29846680e-01 5.18877923e-01 3.34275603e-01
8.50188196e-01 -6.34506762e-01 -2.88158059e-01 6.03417039e-01
8.33911538e-01 2.79611230e-01 5.70944488e-01 1.08228540e-02
7.82516122e-01 8.75958920e-01 4.85360146e-01 5.12090743e-01
1.39130121e-02 8.77169728e-01 6.27548158e-01 -2.79208217e-02
-1.43330082e-01 -1.48192853e-01 8.38073432e-01 8.07193100e-01
-5.16391635e-01 3.56178582e-01 -2.99187243e-01 3.34448427e-01
-1.87565732e+00 -1.02072883e+00 2.32962772e-01 1.90175915e+00
7.09730208e-01 -3.01724553e-01 2.47932941e-01 2.63547063e-01
7.74362564e-01 3.20572108e-01 -6.01419985e-01 -3.84782478e-02
9.04065818e-02 -4.32862900e-02 -6.88591623e-04 1.53352365e-01
-1.23688710e+00 1.03631651e+00 5.99761248e+00 4.99291301e-01
-1.35597217e+00 1.91394612e-01 1.88598484e-01 -8.50141943e-02
4.79033351e-01 -3.31733704e-01 -6.87551081e-01 3.72804046e-01
6.25004113e-01 -4.43915352e-02 2.76857615e-01 1.00512254e+00
2.23350197e-01 2.82894075e-01 -1.38754404e+00 1.09505057e+00
5.46450675e-01 -6.81745708e-01 2.24037766e-02 1.64281987e-02
4.67395097e-01 -3.87850761e-01 -1.95251301e-01 2.31596142e-01
-8.29471275e-02 -8.77030373e-01 6.48810327e-01 4.85671461e-01
6.86006546e-01 -4.54241365e-01 5.65782785e-01 3.50869745e-01
-1.29025745e+00 -8.32661837e-02 -2.63434321e-01 -1.57548919e-01
-1.20180763e-01 -1.01463571e-01 -5.50729752e-01 1.43447131e-01
6.81468487e-01 1.17768836e+00 -3.76212269e-01 2.44462550e-01
-5.59427023e-01 6.11680269e-01 -3.77813354e-02 3.58866096e-01
1.33904135e-02 -3.58281106e-01 5.85528135e-01 9.07508016e-01
-1.99865159e-02 3.05974245e-01 -2.23561022e-02 6.79829597e-01
-2.41898298e-02 5.66588640e-02 -7.38470852e-01 -1.83067292e-01
2.05570415e-01 1.29804194e+00 -1.31098777e-01 -2.03903750e-01
-7.04915762e-01 1.31812930e+00 2.92753726e-01 3.79556835e-01
-7.86167026e-01 5.55369221e-02 9.02997613e-01 -7.79409856e-02
4.07922685e-01 9.21151191e-02 6.80459559e-01 -1.22537100e+00
5.95368780e-02 -9.64805067e-01 3.56550217e-01 -6.98424995e-01
-7.68369853e-01 5.70912421e-01 -5.29622138e-02 -1.52306354e+00
-5.28253555e-01 -5.84723413e-01 -7.33620048e-01 1.96774095e-01
-1.22656012e+00 -1.22085321e+00 -3.65393311e-01 9.68046486e-01
5.45205414e-01 -2.44705006e-01 8.97871733e-01 2.02241525e-01
-9.45287108e-01 8.62472773e-01 -2.17085555e-01 6.14364684e-01
7.67698824e-01 -6.50047421e-01 -3.06321204e-01 7.97504485e-01
3.22675526e-01 4.36654776e-01 6.75979257e-01 -2.78200001e-01
-1.46473742e+00 -1.05056524e+00 6.22375071e-01 -1.78712815e-01
6.47516429e-01 -3.43753994e-01 -7.61406839e-01 9.27943766e-01
8.99674818e-02 5.43533266e-01 6.34154022e-01 -4.86059248e-01
-3.78783047e-01 -3.53808731e-01 -8.50614250e-01 4.74258482e-01
1.11410093e+00 -7.12401628e-01 -7.75950253e-01 1.39504597e-01
2.23895550e-01 -2.89934307e-01 -8.03130686e-01 5.14075160e-01
7.20256150e-01 -9.69813883e-01 7.43917227e-01 -7.56160855e-01
3.99287254e-01 -1.62519321e-01 -6.90976754e-02 -1.03535736e+00
-3.44555140e-01 -6.12501979e-01 -5.36298811e-01 1.13242853e+00
-4.00838763e-01 -2.60397702e-01 5.49591482e-01 2.13024795e-01
3.29582751e-01 -4.70516473e-01 -1.15417993e+00 -5.68437874e-01
-3.95020545e-01 1.18398719e-01 2.58661270e-01 1.08404684e+00
1.54965803e-01 4.11665261e-01 -7.79882789e-01 2.33184487e-01
5.36119580e-01 8.68296847e-02 8.38249326e-01 -1.11256790e+00
-1.71504319e-01 -2.36962929e-01 -8.55099916e-01 -8.06583464e-01
5.77932596e-01 -4.81827080e-01 1.08515583e-01 -7.94253826e-01
1.89316154e-01 6.26020074e-01 -2.60588199e-01 7.20698237e-01
-7.35517368e-02 2.17960626e-01 1.09914444e-01 3.45238179e-01
-3.60572159e-01 9.52924609e-01 1.38679159e+00 -1.22077651e-01
-1.37912989e-01 -1.41555563e-01 -3.59218299e-01 1.02241826e+00
3.99753273e-01 -3.64152580e-01 -1.98807791e-01 -2.41808578e-01
-5.11365175e-01 1.84194431e-01 2.00957343e-01 -8.75111103e-01
1.44277200e-01 -8.97764340e-02 4.87526715e-01 -1.02785923e-01
3.71709645e-01 -8.79450142e-01 -1.47884071e-01 4.84077841e-01
-1.19637258e-01 -3.68571162e-01 6.99259415e-02 6.88040137e-01
-4.45423573e-01 1.35410726e-01 9.96674478e-01 -4.58002761e-02
-7.41669416e-01 4.81197417e-01 -4.32166934e-01 -3.34824711e-01
1.31996667e+00 -2.14789286e-01 1.16090506e-01 -7.26686299e-01
-8.50200474e-01 -3.97313535e-02 3.14300656e-02 6.75949156e-01
5.86471736e-01 -1.48262894e+00 -5.52178979e-01 5.09018898e-01
1.50225252e-01 -2.61493504e-01 1.96668744e-01 1.15501523e+00
6.65769055e-02 2.18689457e-01 -4.43260223e-01 -5.75419843e-01
-1.78399026e+00 5.87828219e-01 5.16208827e-01 3.91009361e-01
-5.80521286e-01 7.39329875e-01 6.43322587e-01 -5.28204143e-02
4.15493727e-01 5.10263955e-03 -7.66566932e-01 3.37927312e-01
7.68020034e-01 2.95361549e-01 -2.43047252e-01 -1.28274095e+00
-3.58667254e-01 7.45233178e-01 -1.18792757e-01 3.99239659e-01
1.23237038e+00 -1.05334789e-01 -9.80619118e-02 3.22483152e-01
1.50309324e+00 -2.24391013e-01 -1.51453412e+00 -3.23952436e-01
-2.10533276e-01 -3.69179428e-01 1.60894513e-01 -4.13470306e-02
-1.57285595e+00 8.45180094e-01 8.98823440e-01 -4.24413502e-01
1.30890226e+00 1.52097598e-01 5.38267732e-01 2.19842821e-01
1.81035206e-01 -1.19504547e+00 3.92122477e-01 7.64864758e-02
1.09260654e+00 -1.29884386e+00 -1.56773612e-01 -3.91866624e-01
-6.99137688e-01 1.28567660e+00 8.45722497e-01 -2.83196986e-01
7.25977719e-01 9.34287086e-02 1.37424856e-01 -2.31690988e-01
-6.82842374e-01 -4.22696739e-01 4.51351047e-01 5.74090719e-01
3.53760123e-01 -7.44389221e-02 -2.18589753e-01 4.24817532e-01
1.17412657e-01 1.25237092e-01 1.79458067e-01 8.47919464e-01
-2.05558941e-01 -7.37394214e-01 -2.85734087e-01 3.93946655e-02
-2.97427088e-01 4.02302235e-01 -6.18047833e-01 9.62958694e-01
2.45250300e-01 6.93432152e-01 2.58088738e-01 -5.16784310e-01
2.07964286e-01 2.89216012e-01 7.09748089e-01 -4.41029042e-01
-7.59181334e-03 5.73459625e-01 -3.00954133e-01 -9.63093817e-01
-8.34185123e-01 -7.12752044e-01 -1.32282257e+00 8.40690956e-02
-3.67559820e-01 -2.10572481e-01 2.25139022e-01 1.15664124e+00
2.35228181e-01 2.82406032e-01 9.44791198e-01 -1.07959735e+00
-6.33162200e-01 -1.28557718e+00 -7.83529520e-01 7.38002241e-01
6.07622266e-01 -1.04403496e+00 -6.31140292e-01 4.81121361e-01] | [13.591639518737793, 1.5876882076263428] |
a15b31e6-e349-429a-a8ec-17f097bdb975 | shallow-encoder-deep-decoder-sedd-networks | 2001.03017 | null | https://arxiv.org/abs/2001.03017v2 | https://arxiv.org/pdf/2001.03017v2.pdf | Shallow Encoder Deep Decoder (SEDD) Networks for Image Encryption and Decryption | This paper explores a new framework for lossy image encryption and decryption using a simple shallow encoder neural network E for encryption, and a complex deep decoder neural network D for decryption. E is kept simple so that encoding can be done on low power and portable devices and can in principle be any nonlinear function which outputs an encoded vector. D is trained to decode the encodings using the dataset of image - encoded vector pairs obtained from E and happens independently of E. As the encodings come from E which while being a simple neural network, still has thousands of random parameters and therefore the encodings would be practically impossible to crack without D. This approach differs from autoencoders as D is trained completely independently of E, although the structure may seem similar. Therefore, this paper also explores empirically if a deep neural network can learn to reconstruct the original data in any useful form given the output of a neural network or any other nonlinear function, which can have very useful applications in Cryptanalysis. Experiments demonstrate the potential of the framework through qualitative and quantitative evaluation of the decoded images from D along with some limitations. | ['Chirag Gupta'] | 2020-01-09 | null | null | null | null | ['cryptanalysis'] | ['miscellaneous'] | [ 3.74554753e-01 2.42993310e-01 1.80248439e-01 -2.15530381e-01
-3.60081702e-01 -7.02894866e-01 6.52188599e-01 -2.25552678e-01
-6.36071801e-01 7.25118160e-01 -7.33697861e-02 -7.99424171e-01
2.26204723e-01 -1.10333180e+00 -8.36002767e-01 -1.05168247e+00
-2.50872552e-01 2.97327280e-01 -2.65555084e-01 -4.14898902e-01
2.26177812e-01 6.04914665e-01 -1.20321476e+00 6.40585661e-01
3.19318056e-01 1.07156062e+00 2.54509132e-02 1.14606404e+00
1.71877086e-01 1.10865641e+00 -9.52466309e-01 -7.55927920e-01
7.72121489e-01 -4.12662417e-01 -8.53452921e-01 -5.33060670e-01
1.25362083e-01 -1.08953595e+00 -8.36029172e-01 1.14190233e+00
5.15488863e-01 -4.97889698e-01 6.46425724e-01 -7.89486766e-01
-5.41143000e-01 4.99379098e-01 2.95951694e-01 -7.35799596e-02
2.50369400e-01 2.90078580e-01 6.94193602e-01 -3.96896333e-01
5.82199693e-01 9.35769022e-01 8.18205893e-01 5.65213382e-01
-1.19867361e+00 -7.69446194e-01 -8.32044959e-01 1.62402362e-01
-1.06409729e+00 -5.04019380e-01 6.24659181e-01 1.64281316e-02
1.33604574e+00 2.44480371e-01 7.31924415e-01 1.04322469e+00
5.40712655e-01 6.84034884e-01 1.12084770e+00 -4.84225541e-01
1.07878312e-01 5.22621870e-01 -3.43346834e-01 8.16066563e-01
1.79549322e-01 5.12717724e-01 -9.19875354e-02 -1.06231749e-01
5.96886098e-01 -9.90759432e-02 -5.06501913e-01 -1.47481218e-01
-9.70920742e-01 1.01449907e+00 6.75377011e-01 4.26474273e-01
-2.24941745e-01 3.31898332e-01 5.29885173e-01 1.21586728e+00
-8.47881485e-04 6.55816495e-01 -3.96914661e-01 -9.93949026e-02
-9.77313817e-01 2.13320747e-01 1.29171085e+00 4.56853896e-01
8.71417999e-01 3.58085006e-01 3.17553401e-01 3.40122163e-01
-1.89324185e-01 3.56747568e-01 7.48438060e-01 -9.13941741e-01
4.02239829e-01 2.18013957e-01 -3.17998648e-01 -1.03072357e+00
-1.28377587e-01 -1.00019470e-01 -1.27449989e+00 7.23952770e-01
4.30543929e-01 -2.60329068e-01 -7.57718146e-01 1.47418833e+00
-3.28453273e-01 -7.84824044e-02 6.37756884e-01 5.13258338e-01
8.03469837e-01 1.08570933e+00 -5.87768018e-01 -9.66053233e-02
8.11743021e-01 -4.85389560e-01 -5.45523882e-01 6.24420457e-02
7.19214559e-01 -6.38686419e-01 5.47141552e-01 6.85070395e-01
-1.43609536e+00 -4.80634928e-01 -1.51873469e+00 -3.92157882e-01
-5.60632825e-01 -8.13769400e-02 6.05066657e-01 8.63164186e-01
-1.49164736e+00 8.90135169e-01 -5.72616756e-01 2.09895313e-01
5.49419165e-01 1.05304909e+00 -6.82010710e-01 -1.49859721e-02
-1.42071950e+00 1.13213706e+00 7.88670838e-01 1.39616907e-01
-9.46014702e-01 -2.15821832e-01 -8.55504096e-01 4.67322648e-01
-4.23139006e-01 -7.78150380e-01 1.09887195e+00 -1.25428164e+00
-1.67933655e+00 8.07761371e-01 9.69861150e-02 -1.09573758e+00
3.89591694e-01 4.25843596e-01 -1.36144772e-01 5.15768826e-01
-5.44467807e-01 7.50551939e-01 1.24126065e+00 -1.16749465e+00
-3.12274307e-01 -3.24870050e-01 2.56700933e-01 -1.24776307e-02
-5.12548506e-01 -2.67397791e-01 8.82118493e-02 -6.03886843e-01
-1.62429526e-01 -8.21925759e-01 1.83822010e-02 6.68239743e-02
-1.85589567e-01 3.13068748e-01 1.03921986e+00 -9.07174349e-01
9.10948873e-01 -2.20758963e+00 2.06309613e-02 3.61097515e-01
4.08497691e-01 6.31514490e-01 -1.58215642e-01 5.90623081e-01
-3.77275705e-01 1.99979573e-01 -3.95043463e-01 -7.75594711e-02
-7.86394030e-02 5.11193573e-01 -4.93149489e-01 6.50203645e-01
1.03063576e-01 1.19184315e+00 -6.40429258e-01 5.18515818e-02
1.66706190e-01 8.27492476e-01 -6.55523241e-01 3.15733612e-01
1.26050189e-01 -2.52915043e-02 -2.54651089e-03 3.31203878e-01
8.51370811e-01 4.50570602e-03 2.11716056e-01 -9.57212299e-02
2.77717263e-01 4.53396201e-01 -7.15218842e-01 1.32900810e+00
-5.95472455e-01 1.27899480e+00 1.83613524e-01 -1.74726748e+00
8.14307749e-01 6.04298830e-01 7.84078091e-02 -9.38218355e-01
4.06011641e-01 3.68058801e-01 3.21050584e-01 -4.84969467e-01
4.19271708e-01 -4.98857468e-01 1.09197877e-01 8.51101339e-01
3.64104331e-01 -5.64579010e-01 -3.74441445e-01 1.91874430e-01
1.25793278e+00 -5.06808400e-01 3.41509461e-01 4.13138233e-02
5.40239036e-01 -3.13821226e-01 -4.88729998e-02 8.66788626e-01
2.02065796e-01 4.17705595e-01 6.46759868e-01 -8.74527931e-01
-1.39282429e+00 -8.53022635e-01 -1.73084289e-01 3.13502133e-01
2.22499054e-02 -4.70993310e-01 -7.65530348e-01 -5.56122780e-01
-2.15092987e-01 3.81317347e-01 -4.84260470e-01 -3.89182150e-01
-7.19553947e-01 -7.50581145e-01 1.04294872e+00 -8.07501003e-02
7.57771552e-01 -1.26453960e+00 -6.83677852e-01 8.77935961e-02
1.27721429e-01 -8.76470625e-01 3.07504952e-01 7.20313668e-01
-1.04729724e+00 -8.77685010e-01 -6.46611869e-01 -1.05621898e+00
8.36349487e-01 -7.19665885e-02 1.09579229e+00 4.42403615e-01
-8.58316049e-02 9.73844156e-02 3.88613120e-02 -3.47345322e-01
-1.07256210e+00 -1.51235193e-01 -1.07661195e-01 -2.44181857e-01
2.52465397e-01 -8.79633605e-01 -5.51275790e-01 -2.77518898e-01
-1.40626633e+00 4.57483046e-02 6.76695168e-01 1.22763669e+00
1.59369260e-02 6.70403421e-01 5.77620044e-02 -1.00647974e+00
8.61983895e-01 -2.17052102e-01 -5.69702506e-01 5.90160191e-02
-4.56614792e-01 4.78074104e-01 1.22729075e+00 -3.22663754e-01
-4.91546273e-01 -1.47265583e-01 -5.45274198e-01 -1.47553489e-01
5.98405488e-02 2.04446495e-01 5.77483186e-03 -5.32763302e-01
6.79904580e-01 6.09021544e-01 2.67686337e-01 -1.81456566e-01
9.39164385e-02 9.28717852e-01 6.84317827e-01 -2.73623735e-01
9.74330544e-01 4.10935789e-01 1.01667039e-01 -7.30056584e-01
8.87952223e-02 2.68850565e-01 -3.59845400e-01 3.61143321e-01
3.86495203e-01 -8.57676446e-01 -9.88266587e-01 6.81223810e-01
-1.45839894e+00 -3.09794188e-01 -4.69914466e-01 2.64494330e-01
-4.93006200e-01 3.97663295e-01 -9.22175407e-01 -4.89436537e-01
-3.98742408e-01 -1.33309054e+00 6.72658324e-01 -1.65804297e-01
-2.82939225e-02 -1.28316665e+00 -2.30263904e-01 -1.91970363e-01
4.64802980e-01 1.15388468e-01 1.30481434e+00 -5.17804921e-01
-7.13724315e-01 -5.19872785e-01 -1.81299075e-01 1.15477157e+00
-1.07383974e-01 -3.78699362e-01 -1.05175304e+00 -6.47535026e-01
7.95646608e-01 -5.40174067e-01 9.96722400e-01 6.96421266e-02
1.44431794e+00 -1.17664456e+00 1.19305491e-01 1.29713583e+00
1.72274160e+00 2.16021821e-01 1.08662188e+00 3.88700873e-01
2.47369215e-01 3.70021403e-01 -6.23558760e-01 2.76252240e-01
8.34611058e-02 3.06625128e-01 6.03826702e-01 -1.02602921e-01
4.70358506e-02 -1.15093961e-01 6.78187609e-01 1.05996430e+00
-1.72537670e-01 -2.73073465e-01 -5.83986223e-01 1.90953746e-01
-1.04845679e+00 -1.09258616e+00 2.68314749e-01 2.14800453e+00
9.65505421e-01 9.15594250e-02 -4.91050780e-01 6.40870869e-01
7.97180086e-02 1.26559108e-01 -3.63601238e-01 -1.06197786e+00
-3.54254365e-01 9.79919136e-01 6.52138710e-01 6.46348298e-01
-7.37087190e-01 4.62888688e-01 7.41912556e+00 7.94054925e-01
-1.35421395e+00 -1.49101853e-01 7.89461613e-01 1.24136373e-01
-5.37905037e-01 -1.73060223e-02 -2.88375765e-01 6.10999525e-01
1.28179765e+00 3.69886570e-02 9.10387337e-01 3.56497824e-01
-3.67712528e-01 -2.14360021e-02 -1.18687260e+00 1.30763066e+00
1.44993335e-01 -1.52184868e+00 3.12808365e-01 3.86216551e-01
6.27458811e-01 -1.24554776e-01 3.51652324e-01 2.51375824e-01
1.58276230e-01 -1.47227669e+00 2.31203794e-01 2.60885000e-01
9.06689584e-01 -9.70876813e-01 8.63607764e-01 6.46036208e-01
-2.72984803e-01 -3.29759568e-01 -7.09563375e-01 -3.37494731e-01
-3.48869801e-01 3.52494299e-01 -1.05342519e+00 2.82069176e-01
5.09881437e-01 6.06116712e-01 -3.41036290e-01 6.12114429e-01
-2.40282655e-01 6.09283447e-01 -3.66947204e-01 -3.27181444e-02
4.59953576e-01 -1.85479522e-01 4.29622084e-01 1.39432430e+00
5.82110465e-01 1.06767461e-01 -6.24476194e-01 5.48317790e-01
-4.14180338e-01 -3.44503880e-01 -1.06617427e+00 -5.58969751e-02
6.44476637e-02 8.68934929e-01 -3.67149204e-01 -4.55236018e-01
-1.38704151e-01 1.35361075e+00 1.38273627e-01 4.01911169e-01
-4.10700440e-01 -5.19107997e-01 5.02100408e-01 1.83325727e-02
4.91795897e-01 -1.38390586e-01 -3.15651357e-01 -1.37963009e+00
6.37490861e-03 -1.25789046e+00 -3.79214389e-03 -7.40591645e-01
-8.71785879e-01 7.72137642e-01 -3.38983595e-01 -1.27867091e+00
-8.03940654e-01 -9.81214881e-01 -4.44782019e-01 8.73797476e-01
-1.56755817e+00 -6.87822521e-01 1.21998854e-01 8.26655746e-01
-4.63601958e-04 -7.22617686e-01 1.35175037e+00 5.60473613e-02
1.21159345e-01 7.27800012e-01 5.42646646e-01 2.77753830e-01
2.23117635e-01 -1.16119659e+00 4.00597602e-01 6.34927392e-01
3.67613345e-01 7.04798281e-01 5.15509784e-01 3.74245807e-03
-1.56428444e+00 -4.97736573e-01 1.07076621e+00 -1.51660636e-01
4.09953922e-01 -5.70662141e-01 -7.91039228e-01 5.73449790e-01
6.09662950e-01 -2.92108636e-02 4.75997508e-01 -6.22368991e-01
-3.69191676e-01 -5.20686023e-02 -1.40614724e+00 3.72505903e-01
4.76538002e-01 -9.86302733e-01 -4.56912309e-01 3.78450572e-01
3.45717043e-01 -4.55063283e-01 -6.43082559e-01 1.57804623e-01
7.98981249e-01 -1.43373907e+00 1.28987968e+00 -3.79775405e-01
7.86375403e-01 6.38411641e-02 -1.32133633e-01 -1.25451136e+00
1.50381073e-01 -7.99123943e-01 -4.17632371e-01 4.72393215e-01
2.45489880e-01 -8.70997548e-01 8.49509656e-01 2.35049859e-01
3.28241587e-01 -6.97888851e-01 -1.00995529e+00 -5.97143292e-01
4.07786548e-01 -4.50489342e-01 7.19823658e-01 7.67094851e-01
-2.13011786e-01 1.46677375e-01 -6.63570046e-01 -2.60929000e-02
3.23789448e-01 -1.94065683e-02 7.96482027e-01 -9.29399192e-01
-6.09283924e-01 -4.71832454e-01 -8.56164396e-01 -1.42152309e+00
2.91491687e-01 -1.17773438e+00 -2.43894994e-01 -1.02145839e+00
-1.76396042e-01 -3.35273385e-01 -9.42772627e-02 4.85554129e-01
3.51334244e-01 5.43546557e-01 1.20949008e-01 2.48591065e-01
3.05196404e-01 2.72339672e-01 1.20327067e+00 -5.05896688e-01
1.11395106e-01 1.19923785e-01 -7.12277770e-01 5.04909575e-01
7.02448130e-01 -5.95675528e-01 -4.13581699e-01 -5.84151387e-01
6.04541183e-01 2.47392446e-01 4.55493033e-01 -9.06478047e-01
3.64564598e-01 6.56934977e-01 7.37257361e-01 -6.24174550e-02
4.37693238e-01 -1.10714471e+00 6.58460930e-02 8.72437239e-01
-2.13236898e-01 1.07051052e-01 3.96304205e-02 7.10190833e-02
-6.11582160e-01 -6.51551962e-01 6.52652383e-01 -4.68644589e-01
-3.56760472e-01 9.39064771e-02 -4.25203353e-01 -3.76515120e-01
6.05268419e-01 -7.24990904e-01 -1.76665872e-01 -9.15818632e-01
-4.94480163e-01 -3.55647892e-01 5.60560167e-01 -9.97156873e-02
9.22758222e-01 -1.14129138e+00 -6.16178215e-01 8.41139913e-01
-4.63831872e-01 -3.37895863e-02 -1.46905273e-01 2.91855425e-01
-1.06715143e+00 3.96447748e-01 -5.05080938e-01 -3.56178284e-01
-1.14498913e+00 4.86346900e-01 4.81804878e-01 -4.42685664e-01
-7.22037792e-01 7.37483382e-01 1.26735017e-01 -4.18889284e-01
1.28491908e-01 -2.55738556e-01 1.16556641e-02 -2.95484483e-01
6.53707027e-01 -1.99598968e-01 1.13964826e-01 -5.71829557e-01
5.20664230e-02 3.36879253e-01 5.70101850e-02 -1.80809587e-01
1.67162132e+00 2.90097684e-01 -5.56561291e-01 -5.28875105e-02
1.97179902e+00 -2.11040571e-01 -8.65802407e-01 5.15572391e-02
-4.97254938e-01 -3.01230490e-01 5.66278547e-02 -5.37723064e-01
-1.20298684e+00 1.33387816e+00 4.40830439e-01 4.11192805e-01
1.51582193e+00 -5.90384126e-01 1.09302759e+00 9.77699399e-01
2.18276381e-01 -7.66144753e-01 -3.40208933e-02 5.99793375e-01
5.68691552e-01 -1.02014315e+00 -2.89775780e-03 2.90109158e-01
-1.53919473e-01 1.68247235e+00 -2.71095008e-01 -4.88971144e-01
7.28359580e-01 7.79807448e-01 -9.16083753e-02 -3.84170786e-02
-6.17018461e-01 3.65179002e-01 8.61740634e-02 6.70289934e-01
7.72858039e-02 -1.91486016e-01 -8.93797800e-02 7.76072666e-02
-7.05411315e-01 2.44671050e-02 7.64864981e-01 8.13392043e-01
-4.05182876e-02 -1.51245284e+00 -3.03120881e-01 3.67728949e-01
-7.92293191e-01 -4.16578203e-01 -2.68366069e-01 7.78973520e-01
2.67352164e-01 3.46421391e-01 9.90822352e-03 -6.43008411e-01
-2.65812278e-01 -5.16790971e-02 7.16549873e-01 -2.11139292e-01
-9.06652987e-01 -5.44857800e-01 -2.75233924e-01 -2.81026781e-01
-3.64246428e-01 -1.35496438e-01 -9.16217983e-01 -7.81358778e-01
-1.22821234e-01 -1.39907077e-02 1.00231862e+00 7.65212417e-01
-6.68191984e-02 5.82162850e-02 9.02816355e-01 -7.90690422e-01
-6.97763205e-01 -6.01075351e-01 -4.48639363e-01 2.66841620e-01
1.02147007e+00 3.15673411e-01 -4.42394435e-01 1.33197680e-01] | [4.440208435058594, 7.98645544052124] |
a765c336-955f-45ae-aa74-cf09dc48c898 | attention-based-scaling-adaptation-for-target | 2010.10923 | null | https://arxiv.org/abs/2010.10923v3 | https://arxiv.org/pdf/2010.10923v3.pdf | Attention-based scaling adaptation for target speech extraction | The target speech extraction has attracted widespread attention in recent years. In this work, we focus on investigating the dynamic interaction between different mixtures and the target speaker to exploit the discriminative target speaker clues. We propose a special attention mechanism without introducing any additional parameters in a scaling adaptation layer to better adapt the network towards extracting the target speech. Furthermore, by introducing a mixture embedding matrix pooling method, our proposed attention-based scaling adaptation (ASA) can exploit the target speaker clues in a more efficient way. Experimental results on the spatialized reverberant WSJ0 2-mix dataset demonstrate that the proposed method can improve the performance of the target speech extraction effectively. Furthermore, we find that under the same network configurations, the ASA in a single-channel condition can achieve competitive performance gains as that achieved from two-channel mixtures with inter-microphone phase difference (IPD) features. | ['Yanhua Long', 'Wei Rao', 'Jiaen Liang', 'Jiangyu Han'] | 2020-10-19 | null | null | null | null | ['speech-extraction'] | ['speech'] | [ 7.75450049e-03 -2.11396471e-01 -9.55129117e-02 -1.57595173e-01
-1.12199807e+00 -3.23378354e-01 4.32814240e-01 -2.60394126e-01
-5.04914939e-01 2.59535104e-01 5.59739709e-01 -2.27267891e-01
-3.28381769e-02 -1.17935985e-01 -3.55486989e-01 -1.07757187e+00
-1.03235722e-01 -3.55731845e-01 1.51712596e-01 -1.24483436e-01
-8.82475227e-02 4.08312142e-01 -1.10203409e+00 -1.98714864e-02
7.93129504e-01 9.34690475e-01 6.71667635e-01 6.81380987e-01
1.82862669e-01 3.73172641e-01 -8.77845228e-01 -4.76480499e-02
3.84275794e-01 -4.17523772e-01 -1.51153222e-01 2.88162883e-02
2.06242323e-01 -1.53249511e-02 -5.32861173e-01 1.13067901e+00
1.18053734e+00 3.88965905e-01 4.89665002e-01 -8.03801358e-01
-2.06197783e-01 9.62290049e-01 -6.62103713e-01 6.89079702e-01
-1.32187560e-01 2.46140435e-01 9.80009794e-01 -1.34361660e+00
-1.42339796e-01 1.27970660e+00 4.46042895e-01 3.98589015e-01
-1.05321777e+00 -1.20817065e+00 4.14342046e-01 5.17142057e-01
-1.64091313e+00 -1.14015162e+00 1.22629404e+00 9.41864252e-02
9.10243809e-01 4.03044045e-01 2.38196328e-01 1.08759737e+00
-3.78166378e-01 9.62062120e-01 8.13603461e-01 -4.13038731e-01
1.62082255e-01 2.83784568e-01 3.21473666e-02 2.15589464e-01
-1.64431199e-01 -7.72570539e-03 -7.07089424e-01 -7.33747706e-02
4.73452210e-01 -4.20428365e-01 -7.28610635e-01 8.64669979e-02
-1.10433733e+00 5.83575904e-01 5.20359933e-01 6.30017877e-01
-3.09911281e-01 5.91013469e-02 2.11448818e-01 2.07306713e-01
7.11140037e-01 5.16124010e-01 -5.47008455e-01 -1.68103769e-01
-8.90045881e-01 -1.82477817e-01 6.33060873e-01 7.48979390e-01
4.31837767e-01 5.85322559e-01 -2.61790991e-01 1.30273104e+00
5.20156443e-01 6.87597156e-01 6.65542781e-01 -6.22917414e-01
7.64966547e-01 -1.62200943e-01 -1.56266585e-01 -8.10573578e-01
-2.84361333e-01 -1.23068309e+00 -8.49399388e-01 -1.33105993e-01
6.95972219e-02 -4.17785734e-01 -7.34952033e-01 2.04255128e+00
5.06531000e-01 6.23883188e-01 2.65745610e-01 8.61274719e-01
5.86702526e-01 8.26169968e-01 -7.54498467e-02 -4.25422460e-01
1.25823450e+00 -1.12035930e+00 -1.00186265e+00 -4.56896514e-01
2.57563621e-01 -9.40892816e-01 8.45012844e-01 2.31355622e-01
-1.02680635e+00 -8.17281008e-01 -1.19403803e+00 4.00605470e-01
-1.06645010e-01 2.09977284e-01 1.46547362e-01 8.19463909e-01
-7.76696265e-01 1.98328480e-01 -8.77332747e-01 -4.57464606e-02
2.92850137e-01 4.51791584e-01 -1.16175577e-01 5.96505851e-02
-1.20246506e+00 6.59408212e-01 1.68827116e-01 2.79064178e-01
-6.27933919e-01 -7.04615176e-01 -8.15930367e-01 4.20701861e-01
3.33733588e-01 -3.12152177e-01 1.24997115e+00 -6.82961643e-01
-1.85444331e+00 5.59354164e-02 -5.37278116e-01 -3.80725890e-01
1.11482210e-01 -3.09232026e-01 -8.58037889e-01 1.77478015e-01
-1.44679412e-01 2.76878476e-01 1.20691872e+00 -1.01765430e+00
-7.33907461e-01 -2.72914439e-01 -3.52024078e-01 4.54462320e-01
-8.78807783e-01 2.13204235e-01 -5.15569150e-01 -1.11241519e+00
2.31058508e-01 -7.64665067e-01 -2.30961561e-01 -4.84041959e-01
-4.57985520e-01 -1.32353857e-01 9.25027490e-01 -8.46730828e-01
1.46065056e+00 -2.73875022e+00 2.33418688e-01 1.13520592e-01
1.15976699e-01 6.34853303e-01 -2.71580249e-01 2.26103768e-01
-1.12108178e-01 -4.43415195e-02 -1.04232281e-01 -8.38001788e-01
4.45831977e-02 -3.00205588e-01 -2.26877600e-01 5.05888045e-01
3.97172719e-01 5.17601252e-01 -6.17776930e-01 -3.77125293e-01
9.22708288e-02 8.11729670e-01 -5.95744312e-01 3.94257933e-01
4.10537988e-01 5.79235971e-01 -3.67252648e-01 4.01964933e-01
8.94567668e-01 1.09602518e-01 1.52454793e-01 -3.17722708e-01
-2.05230881e-02 5.76074302e-01 -1.17854774e+00 1.58716953e+00
-8.65915477e-01 8.97736728e-01 6.62863791e-01 -9.14260089e-01
8.78674269e-01 5.46344399e-01 1.12895705e-01 -5.91686189e-01
3.98990437e-02 3.50721329e-01 6.66279972e-01 -3.79203200e-01
8.36076960e-02 -7.29882494e-02 1.08089447e-01 9.84191373e-02
1.47741809e-01 9.67190042e-02 -2.02207267e-01 -8.38012472e-02
1.03893244e+00 -5.21648943e-01 3.06605935e-01 -2.13648409e-01
6.45068228e-01 -8.67320061e-01 7.06455648e-01 6.68441236e-01
-3.43291551e-01 3.96309763e-01 -1.11723445e-01 3.58893573e-01
-6.62331998e-01 -1.09372461e+00 -1.03502102e-01 1.24939597e+00
-3.34305800e-02 -2.73170948e-01 -6.52774394e-01 -3.74214292e-01
-1.88012466e-01 7.57942200e-01 -1.19714186e-01 -2.55909711e-01
-9.54919457e-01 -9.03263509e-01 5.46631634e-01 6.26373231e-01
6.26966894e-01 -7.44485319e-01 -1.35829479e-01 4.41509902e-01
-3.19284946e-01 -1.33801377e+00 -9.22624588e-01 4.42072749e-01
-5.67749023e-01 -3.40388477e-01 -9.68210936e-01 -8.25258613e-01
3.14625740e-01 5.35741568e-01 3.22987258e-01 -4.51531500e-01
2.17570364e-01 1.18044712e-01 -3.14311773e-01 -3.84273916e-01
-3.68560433e-01 4.51325148e-01 3.09951305e-01 4.76094037e-01
2.19795957e-01 -8.99543762e-01 -6.11025870e-01 2.42495015e-01
-5.42360187e-01 -9.84129906e-02 8.36138785e-01 8.07081997e-01
-1.30699396e-01 4.31194037e-01 9.74562705e-01 -1.63295373e-01
6.73836172e-01 -4.59395617e-01 -2.04638913e-01 -2.26138070e-01
-3.32676828e-01 -1.41655263e-02 6.40739024e-01 -9.67171252e-01
-1.26187205e+00 -8.96131769e-02 -3.49189460e-01 -4.68165368e-01
-9.00070146e-02 4.15845960e-01 -8.03909719e-01 3.58923748e-02
2.81387866e-01 3.78420860e-01 -2.86277294e-01 -8.70994747e-01
1.96216285e-01 1.16942787e+00 3.59850377e-01 -2.02926069e-01
1.08609724e+00 2.47933090e-01 -3.12559336e-01 -1.19260454e+00
-4.54050511e-01 -8.67427289e-01 -3.38629931e-01 7.75962174e-02
5.37422419e-01 -1.07413065e+00 -5.02901793e-01 5.95498562e-01
-1.09656894e+00 -2.24833816e-01 -1.20754568e-02 9.89734590e-01
-6.26069382e-02 3.35173428e-01 -6.20451212e-01 -1.20776486e+00
-3.25918615e-01 -1.23620260e+00 8.77329051e-01 2.17445001e-01
-5.74127547e-02 -7.03624666e-01 -3.69779319e-02 2.46674195e-01
8.51735055e-01 -5.88615656e-01 6.66992426e-01 -1.06075096e+00
-5.15179992e-01 -1.25521839e-01 -9.48720798e-03 5.89461565e-01
6.60101652e-01 -4.66348469e-01 -1.50366485e+00 -3.99092466e-01
4.04541224e-01 4.54736233e-01 8.47999811e-01 4.32905287e-01
9.37333047e-01 -4.67249274e-01 -3.30335110e-01 7.08798826e-01
8.30203652e-01 5.02079129e-01 2.73026556e-01 1.29875401e-02
7.42960513e-01 3.56769592e-01 3.10599715e-01 2.39639387e-01
6.83522448e-02 8.70850623e-01 1.25065893e-01 -3.05748641e-01
-4.32032704e-01 6.88968822e-02 5.76852083e-01 1.52592325e+00
1.84067249e-01 -3.40304643e-01 -6.02569818e-01 6.50833726e-01
-1.55383193e+00 -7.45814145e-01 3.28185618e-01 2.29865026e+00
9.74201381e-01 2.05488190e-01 -7.30344728e-02 3.70551616e-01
8.08725893e-01 3.95055950e-01 -5.22665739e-01 6.36859834e-02
-2.91517764e-01 3.84228826e-01 3.39231998e-01 7.08343923e-01
-1.09744930e+00 7.49783278e-01 5.63837051e+00 1.38476968e+00
-1.45346200e+00 4.86460418e-01 3.53322625e-01 -3.49688351e-01
6.67695925e-02 -3.50726098e-01 -8.47951293e-01 5.21237731e-01
1.01150250e+00 -9.39481556e-02 5.19414485e-01 5.11788785e-01
3.21524888e-01 3.38107377e-01 -8.98772180e-01 1.05722713e+00
6.69034794e-02 -7.61618018e-01 -4.91506279e-01 1.04987286e-01
2.20461026e-01 9.85340402e-02 4.53858972e-01 4.91169840e-01
-2.46021837e-01 -7.03632653e-01 6.50193214e-01 1.04926899e-01
4.57220078e-01 -7.34293342e-01 5.78344584e-01 3.71450067e-01
-1.46707773e+00 -4.04284924e-01 -7.13749900e-02 1.67203769e-01
4.50243056e-02 6.04278207e-01 -1.30683553e+00 4.38382536e-01
5.86113334e-01 2.87983924e-01 -3.94527704e-01 1.28119671e+00
-2.10145295e-01 1.04903269e+00 -4.82678205e-01 -7.72552341e-02
-8.30254750e-04 1.72886744e-01 1.15082705e+00 1.55583608e+00
3.22079301e-01 -1.24792315e-01 -1.28848568e-01 6.12406969e-01
2.20729336e-02 2.38517940e-01 -1.93196386e-01 -2.06411276e-02
6.96283877e-01 1.17692649e+00 -4.77264702e-01 -2.03204647e-01
-2.74964511e-01 8.61840427e-01 6.35829493e-02 8.53356719e-01
-8.92938256e-01 -5.74232340e-01 1.07588077e+00 -1.22753598e-01
8.55622053e-01 -3.57611299e-01 1.00423999e-01 -1.04552412e+00
9.42813233e-02 -9.58800375e-01 -1.32110029e-01 -4.20611382e-01
-1.11028004e+00 7.05199659e-01 -1.62133679e-01 -1.20526135e+00
-8.83744359e-02 -4.36727881e-01 -6.94242060e-01 1.25671852e+00
-1.53025258e+00 -1.02077401e+00 2.07307294e-01 3.75361413e-01
8.39239717e-01 -2.05263957e-01 6.99277639e-01 6.36134565e-01
-9.09522831e-01 9.78361845e-01 1.40977845e-01 1.58917427e-01
8.05352926e-01 -1.08148730e+00 4.08476651e-01 1.07865465e+00
1.62922293e-01 7.06703424e-01 7.98097610e-01 -2.43784472e-01
-1.09845972e+00 -9.98903036e-01 6.21188760e-01 1.15829319e-01
5.68482339e-01 -7.73529768e-01 -1.00453520e+00 2.86603004e-01
5.15249312e-01 -3.02486140e-02 7.80013025e-01 1.48665652e-01
-3.28656793e-01 -4.49307889e-01 -7.61240721e-01 6.64400876e-01
8.46988678e-01 -6.60244346e-01 -4.87952679e-01 1.05959564e-01
1.07836533e+00 -2.63341308e-01 -4.61482316e-01 4.73396391e-01
2.97877014e-01 -4.91408944e-01 1.20940220e+00 -4.08043176e-01
-3.20056707e-01 -3.28866810e-01 -4.05078888e-01 -1.64861119e+00
-4.73563969e-01 -9.95906413e-01 -3.89636248e-01 1.56236267e+00
5.50780594e-01 -9.66067076e-01 3.65715504e-01 1.56515658e-01
-3.23721021e-01 -6.29247487e-01 -1.18449247e+00 -9.33993757e-01
-1.27066851e-01 -5.32829642e-01 7.06731856e-01 7.31700897e-01
-6.19975664e-02 4.70768362e-01 -4.00388032e-01 6.82118773e-01
5.44005334e-01 -2.89558202e-01 6.66086435e-01 -8.28415453e-01
-8.07965338e-01 -6.09011590e-01 -2.75314212e-01 -1.45080054e+00
2.28397518e-01 -6.97872937e-01 2.06777573e-01 -9.73522186e-01
-2.51173407e-01 -3.39429170e-01 -8.99087369e-01 1.43697649e-01
-6.91910923e-01 -3.67368348e-02 2.91884333e-01 4.56808172e-02
-2.36264929e-01 9.03019667e-01 9.90811110e-01 -3.32009047e-01
-2.98510820e-01 3.96971434e-01 -9.50977147e-01 6.84056282e-01
7.75772631e-01 -5.92193127e-01 -3.08232903e-01 -2.96173960e-01
-4.03416932e-01 -1.17740959e-01 7.48160258e-02 -1.08933449e+00
4.11295980e-01 2.45824859e-01 2.28964850e-01 -2.88289577e-01
9.01263595e-01 -7.83779085e-01 -1.98084503e-01 2.40489975e-01
-2.22548351e-01 -5.85537136e-01 4.78099614e-01 5.91105282e-01
-3.39147449e-01 -1.26400724e-01 7.50110865e-01 3.54271561e-01
-2.86343217e-01 1.20023318e-01 -3.74333352e-01 -2.18660831e-01
5.58216035e-01 1.05882123e-01 3.04961521e-02 -5.93517542e-01
-6.02394819e-01 -9.63005051e-02 -2.67320991e-01 3.81298453e-01
3.66255105e-01 -1.39652193e+00 -7.46507466e-01 2.75634408e-01
-2.45328233e-01 -2.74887532e-01 3.74258041e-01 1.02125275e+00
3.92764211e-01 3.59741956e-01 3.88016522e-01 -5.08951724e-01
-1.27569401e+00 4.14524585e-01 4.22136813e-01 -7.59866163e-02
-3.85207683e-01 1.24689221e+00 7.08611429e-01 -2.07536951e-01
4.33822900e-01 -4.15377796e-01 -2.03225136e-01 -5.57351075e-02
6.43412948e-01 3.88882160e-01 6.21960796e-02 -6.72134221e-01
-3.89832854e-01 3.79259616e-01 -6.54022768e-02 -3.99841756e-01
1.25989676e+00 -3.94131303e-01 3.02390128e-01 5.61397314e-01
1.41882360e+00 7.79326200e-01 -1.25443506e+00 -6.29733503e-01
-1.33762255e-01 -4.98540103e-01 4.06608850e-01 -4.81279373e-01
-1.17041850e+00 1.00138736e+00 7.81790495e-01 3.81614685e-01
1.41844714e+00 -3.67893279e-03 8.09616446e-01 1.78042367e-01
-1.17684655e-01 -7.51221895e-01 7.70981014e-02 2.29244336e-01
9.11474586e-01 -1.16828084e+00 -4.53835011e-01 -4.85537529e-01
-2.65019953e-01 9.11826134e-01 4.87530887e-01 2.39608750e-01
9.67640281e-01 4.27030951e-01 2.35217229e-01 1.99062303e-01
-4.99283969e-01 -3.34477872e-01 4.04802889e-01 4.78385657e-01
3.05764824e-01 1.14205092e-01 3.16101640e-01 8.04365873e-01
-2.00294882e-01 -7.87231505e-01 9.15954188e-02 5.85748136e-01
-5.21150053e-01 -1.04077756e+00 -6.45432591e-01 1.47293985e-01
-5.84649444e-01 -4.98903066e-01 -1.13728955e-01 4.82770830e-01
-3.38471621e-01 1.28323674e+00 -9.39694606e-03 -4.19170290e-01
5.32264948e-01 2.11426035e-01 1.74009547e-01 -5.58843970e-01
-4.93395895e-01 8.53400826e-01 -6.55744374e-02 -1.88455626e-01
-3.30732286e-01 -6.15649343e-01 -9.69508410e-01 1.49586499e-01
-8.39605331e-01 2.48219967e-01 7.94826448e-01 8.91630709e-01
4.73578095e-01 8.54648829e-01 1.01851273e+00 -9.79879558e-01
-7.40118206e-01 -1.59415627e+00 -6.12065673e-01 -5.66855557e-02
8.19103599e-01 -4.76525903e-01 -8.16663146e-01 -1.93989933e-01] | [14.81723690032959, 5.867146968841553] |
993070fc-fce4-4e8f-9bc7-457c79d2e2ae | geometry-based-occlusion-aware-unsupervised | 2010.10700 | null | https://arxiv.org/abs/2010.10700v1 | https://arxiv.org/pdf/2010.10700v1.pdf | Geometry-based Occlusion-Aware Unsupervised Stereo Matching for Autonomous Driving | Recently, there are emerging many stereo matching methods for autonomous driving based on unsupervised learning. Most of them take advantage of reconstruction losses to remove dependency on disparity groundtruth. Occlusion handling is a challenging problem in stereo matching, especially for unsupervised methods. Previous unsupervised methods failed to take full advantage of geometry properties in occlusion handling. In this paper, we introduce an effective way to detect occlusion regions and propose a novel unsupervised training strategy to deal with occlusion that only uses the predicted left disparity map, by making use of its geometry features in an iterative way. In the training process, we regard the predicted left disparity map as pseudo groundtruth and infer occluded regions using geometry features. The resulting occlusion mask is then used in either training, post-processing, or both of them as guidance. Experiments show that our method could deal with the occlusion problem effectively and significantly outperforms the other unsupervised methods for stereo matching. Moreover, our occlusion-aware strategies can be extended to the other stereo methods conveniently and improve their performances. | ['Deng Cai', 'Dan Deng', 'Liang Peng'] | 2020-10-21 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [ 3.76261413e-01 1.57769442e-01 -1.23105496e-01 -6.14171565e-01
-3.63051981e-01 -1.78963572e-01 5.31286061e-01 1.45368904e-01
-5.05962849e-01 6.37154579e-01 1.67658553e-01 -2.49828532e-01
-3.94726694e-02 -1.15457869e+00 -6.41579807e-01 -7.45386243e-01
4.80396956e-01 6.53210640e-01 6.90080404e-01 -2.34222755e-01
5.64757049e-01 4.90563482e-01 -2.30531478e+00 -1.54789135e-01
1.28708458e+00 6.75200939e-01 4.77045029e-01 1.64325222e-01
-3.89575005e-01 6.04577839e-01 -7.49754012e-02 -1.09457925e-01
5.62846720e-01 -1.53816134e-01 -4.67816114e-01 4.19625938e-01
5.02148628e-01 -3.87940854e-01 -3.38453323e-01 9.81319308e-01
5.88687778e-01 2.55631387e-01 5.81384063e-01 -9.94258523e-01
6.87784493e-01 1.13381661e-01 -6.98108017e-01 -7.17674196e-02
2.85035431e-01 -1.07657336e-01 6.54921591e-01 -8.35231721e-01
8.12010527e-01 1.09681869e+00 5.23698032e-01 1.46446809e-01
-9.52934027e-01 -7.89687634e-01 9.49713364e-02 2.42432341e-01
-1.24792910e+00 -6.32529140e-01 1.13638794e+00 -6.70584023e-01
5.62463462e-01 6.38439357e-02 6.20292723e-01 3.76478165e-01
-6.28346950e-02 6.57723784e-01 1.23201680e+00 -6.94934130e-01
1.86386630e-01 1.28978744e-01 2.39255920e-01 7.53277540e-01
2.63056099e-01 4.03969944e-01 -4.77499872e-01 5.42344116e-02
5.43901622e-01 -4.09296155e-02 -1.28499717e-01 -8.91970634e-01
-9.07701850e-01 8.57454836e-01 3.28612596e-01 -1.27702681e-02
5.88074792e-04 -1.93769887e-01 3.43394488e-01 1.70983002e-01
5.98329961e-01 2.09486127e-01 -1.42633066e-01 9.51241851e-02
-8.59562039e-01 2.00313374e-01 6.27236545e-01 1.03023708e+00
1.66060007e+00 -2.64829278e-01 1.10514283e-01 9.49241281e-01
1.72294676e-01 4.10730332e-01 2.88959116e-01 -1.05287886e+00
6.67134106e-01 8.63968194e-01 -7.16902986e-02 -1.18149757e+00
-4.21658963e-01 -1.85200185e-01 -6.40622795e-01 4.11887228e-01
3.71147275e-01 1.29641533e-01 -1.06377614e+00 1.31899571e+00
6.72223210e-01 1.10944860e-01 7.74766207e-02 7.67194211e-01
9.05927122e-01 3.08293700e-01 -1.95603549e-01 -2.27692366e-01
9.70913827e-01 -9.37206328e-01 -5.96081078e-01 -4.40800875e-01
8.26183438e-01 -9.04469848e-01 5.19670188e-01 3.51163656e-01
-7.04848170e-01 -8.40635478e-01 -1.18130732e+00 -2.53506631e-01
-3.46705288e-01 1.09919332e-01 8.26487482e-01 6.01906240e-01
-8.00027668e-01 6.22668982e-01 -6.62219524e-01 -5.53393066e-01
2.31936499e-01 5.70305526e-01 -3.91045660e-01 -9.47594792e-02
-8.59673440e-01 7.48921454e-01 6.41526937e-01 1.31007046e-01
-2.91043550e-01 -5.05694635e-02 -1.29550683e+00 -3.88444871e-01
3.81507754e-01 -7.39803076e-01 8.46383870e-01 -8.28335166e-01
-1.48616648e+00 1.08573413e+00 -5.09931445e-01 -4.20090944e-01
7.66740680e-01 -2.46202588e-01 5.72660454e-02 3.61361615e-02
3.06969702e-01 8.83364677e-01 7.35506654e-01 -1.38046110e+00
-9.30690348e-01 -3.68142456e-01 2.12964088e-01 4.17315394e-01
-1.38235256e-01 -5.32007754e-01 -6.72884285e-01 -1.34805918e-01
9.71541166e-01 -7.58855283e-01 -5.14331460e-01 -3.41157392e-02
-3.21075976e-01 -1.08828329e-01 8.06542397e-01 -2.97856063e-01
9.59197819e-01 -2.02581358e+00 -4.71689031e-02 4.45485830e-01
2.04863206e-01 1.74671337e-01 2.97767967e-01 3.58707964e-01
1.66235775e-01 -5.09086907e-01 -4.64029908e-01 -5.98609805e-01
-1.97436228e-01 5.08902192e-01 -2.58157045e-01 5.72421849e-01
1.84904877e-02 4.44606513e-01 -9.70292270e-01 -1.07921612e+00
7.76575267e-01 1.55076548e-01 -6.52063251e-01 1.30672902e-01
-5.52678332e-02 9.06384170e-01 -6.20670199e-01 3.06680024e-01
1.15779626e+00 4.45393741e-01 5.76679632e-02 -1.67300075e-01
-5.58396399e-01 3.10451418e-01 -1.27867389e+00 1.87021089e+00
-4.60592210e-01 6.82794452e-01 -1.59464851e-01 -1.20685673e+00
1.30874801e+00 -4.34198193e-02 4.98354495e-01 -8.25430393e-01
1.99522167e-01 6.19167030e-01 -2.29056254e-01 -6.82239711e-01
5.39898932e-01 1.01437740e-01 2.02101454e-01 1.32837340e-01
-2.97134697e-01 -4.48258489e-01 2.88897216e-01 -8.10084045e-02
7.05940604e-01 4.95883465e-01 3.29703599e-01 -3.83330554e-01
9.86119270e-01 9.50215980e-02 9.24724936e-01 7.57457614e-01
2.62499545e-02 7.90046930e-01 5.37923127e-02 -7.07099080e-01
-1.01817822e+00 -6.99216247e-01 -3.47108394e-01 4.95244533e-01
7.98061132e-01 -3.93142521e-01 -5.79191029e-01 -3.58161092e-01
7.41489530e-02 3.22448730e-01 -2.28493989e-01 -1.80690661e-02
-7.90969908e-01 -3.64207536e-01 -4.92404439e-02 3.99981290e-01
8.44747424e-01 -9.90137696e-01 -6.40774667e-01 3.08122814e-01
-3.59582394e-01 -1.14324224e+00 -4.79933470e-02 2.11259142e-01
-1.26648307e+00 -9.12653625e-01 -3.13611686e-01 -9.85477686e-01
8.96075547e-01 6.62640274e-01 9.06444430e-01 3.23424995e-01
-2.11340174e-01 -5.80504313e-02 -1.43712312e-01 -3.05888057e-01
-2.64265120e-01 2.05146030e-01 -9.54599008e-02 1.35805845e-01
4.94021416e-01 -9.26078558e-01 -7.14170337e-01 7.29414761e-01
-6.43425524e-01 4.78313297e-01 3.59745741e-01 6.43366456e-01
8.17662299e-01 3.20093744e-02 1.45565465e-01 -1.19424665e+00
-1.37597203e-01 -1.37840658e-02 -9.66833353e-01 -2.65727907e-01
-4.09529746e-01 3.43986988e-01 3.46649587e-01 6.06557764e-02
-1.20379829e+00 5.37074685e-01 -3.10073435e-01 -3.66246283e-01
-2.40653723e-01 2.30482042e-01 -4.54620957e-01 -3.07072520e-01
4.33957279e-01 2.17625812e-01 9.47632566e-02 -5.24853110e-01
5.47203794e-02 7.67336547e-01 5.79778254e-01 -6.51443958e-01
1.11234808e+00 8.95456314e-01 2.21829638e-01 -8.46109748e-01
-7.94050753e-01 -1.03909862e+00 -9.95037198e-01 -2.84197479e-01
7.54121602e-01 -1.03937221e+00 -4.14921850e-01 4.93152291e-01
-1.10341227e+00 4.84950840e-02 -8.96053761e-02 7.37465143e-01
-7.25216866e-01 7.20306754e-01 -1.20459564e-01 -9.42426145e-01
4.18402478e-02 -1.37684047e+00 1.28588831e+00 2.09368482e-01
2.11098626e-01 -8.48350883e-01 1.76219463e-01 5.80826581e-01
1.11281648e-01 2.22767934e-01 5.12221754e-01 -1.39448598e-01
-8.67741644e-01 -1.79464743e-01 -2.67375648e-01 1.00335233e-01
2.52629459e-01 -1.64604962e-01 -1.29432714e+00 -1.81276336e-01
1.68472812e-01 -1.04464293e-01 1.17798579e+00 2.65821069e-01
1.13086963e+00 3.11184168e-01 -5.34867704e-01 8.94720674e-01
1.45335579e+00 1.01043135e-01 9.97991443e-01 7.73327112e-01
7.86417305e-01 9.25901532e-01 1.00949252e+00 3.01037103e-01
4.76836503e-01 8.56972873e-01 3.86305630e-01 -2.53525674e-01
1.09770345e-02 -4.00454938e-01 9.70917344e-02 9.92483139e-01
-2.72267967e-01 1.45781487e-01 -7.21837044e-01 3.71396333e-01
-2.15703130e+00 -8.56568575e-01 -4.16880250e-01 2.67649817e+00
5.08013368e-01 4.97714132e-01 -1.95614785e-01 4.88337785e-01
8.73151720e-01 1.74159661e-01 -3.09240222e-01 -2.90988624e-01
-1.34874955e-01 7.89538473e-02 7.23845661e-01 7.67447293e-01
-1.39696717e+00 1.10808527e+00 5.19650602e+00 8.29298198e-01
-8.08867574e-01 -1.88892201e-01 3.37901264e-01 5.80898106e-01
-4.01011854e-01 4.86894786e-01 -9.69549060e-01 2.48986691e-01
1.93018183e-01 1.40889496e-01 -1.19007481e-02 7.92705178e-01
3.29064220e-01 -6.71157479e-01 -8.33153307e-01 1.00426900e+00
-2.76598930e-02 -8.46366644e-01 -1.04963578e-01 1.48163408e-01
8.43744874e-01 -2.34463185e-01 -3.61307532e-01 5.90138622e-02
-1.13152780e-01 -5.38492203e-01 4.81315315e-01 3.83434951e-01
5.03562212e-01 -6.46892667e-01 7.75312066e-01 6.02210045e-01
-1.40921295e+00 2.06163123e-01 -7.89725125e-01 -4.05136704e-01
1.41891420e-01 8.68742824e-01 -5.70034564e-01 8.71063054e-01
2.79084265e-01 9.69128013e-01 -5.10549664e-01 1.45094967e+00
-2.87831694e-01 1.79961503e-01 -6.05326295e-01 3.64649624e-01
2.00542584e-01 -5.33262789e-01 6.30493760e-01 8.88240874e-01
3.06860834e-01 -1.02135494e-01 4.79412109e-01 4.24708456e-01
3.81704718e-01 3.74264777e-01 -9.35323000e-01 7.18123496e-01
3.17165971e-01 1.04470825e+00 -8.88573527e-01 -5.09452581e-01
-4.10311788e-01 6.84021592e-01 2.75297970e-01 1.34168848e-01
-6.32973194e-01 -2.95875877e-01 2.67758459e-01 2.57447451e-01
9.43062827e-02 -1.70642018e-01 -4.06111002e-01 -1.17613053e+00
3.09006631e-01 -4.51019019e-01 1.29340068e-01 -6.08019173e-01
-8.14985812e-01 5.47214568e-01 8.44008923e-02 -1.95389652e+00
-3.03307444e-01 -3.29144239e-01 -4.96059388e-01 6.10922277e-01
-1.95154274e+00 -8.67361248e-01 -6.90052748e-01 5.11799037e-01
6.83828235e-01 2.74315495e-02 4.64886427e-01 4.72765326e-01
-4.13621455e-01 2.25736201e-01 1.18368149e-01 -1.79067627e-01
8.63903522e-01 -1.09255171e+00 2.55416092e-02 8.70357692e-01
-9.56058726e-02 2.87328631e-01 8.19751978e-01 -6.22954726e-01
-8.88385713e-01 -1.00019252e+00 1.11359847e+00 1.26264784e-02
1.93536624e-01 -3.91155988e-01 -7.85772800e-01 3.77382368e-01
-6.06367178e-02 -1.51594669e-01 1.63413584e-01 9.52165425e-02
-1.32895976e-01 -4.93333220e-01 -9.82346058e-01 5.08093476e-01
1.45675027e+00 -3.24660957e-01 -7.09326804e-01 3.54130089e-01
3.86004239e-01 -6.33968174e-01 -2.78235078e-01 8.49927843e-01
5.06463408e-01 -1.41158319e+00 9.15052652e-01 1.23689935e-01
3.84561509e-01 -6.31116211e-01 9.01246350e-03 -7.93910325e-01
5.68220504e-02 -5.98196983e-01 4.40442532e-01 1.23744750e+00
1.22792266e-01 -9.28975761e-01 9.94814634e-01 2.97568709e-01
-5.27891994e-01 -2.79369116e-01 -9.19727743e-01 -6.21656716e-01
-4.02019769e-01 -4.12637085e-01 4.56700772e-01 5.88296652e-01
-1.18629336e-01 1.65073574e-02 -2.90620267e-01 1.91265613e-01
9.48741972e-01 6.26154840e-01 1.37360632e+00 -1.53100574e+00
-1.87766835e-01 -1.69067189e-01 -8.34319770e-01 -1.45054793e+00
1.63199320e-01 -5.49910009e-01 4.24976319e-01 -1.41311800e+00
2.83821952e-02 -7.81690776e-01 1.40478229e-02 1.67240217e-01
-2.93383449e-01 4.10371900e-01 -1.71621948e-01 4.44838196e-01
-4.02059585e-01 4.79255080e-01 1.26775789e+00 -1.42693043e-01
-4.03899610e-01 1.26673862e-01 -2.61056833e-02 1.03673315e+00
8.52583468e-01 -3.48578453e-01 -6.13177061e-01 -5.18125534e-01
-1.85717031e-01 -1.86905265e-02 1.05106279e-01 -1.52242982e+00
3.66715163e-01 -5.11583202e-02 5.65370321e-02 -1.02881384e+00
3.92331034e-01 -8.81809473e-01 1.91245392e-01 3.21091443e-01
1.45850137e-01 -2.37390190e-01 9.71408486e-02 4.32865292e-01
-4.53703701e-01 -4.42456692e-01 7.41683125e-01 -7.77383298e-02
-9.79711235e-01 2.80787319e-01 -2.85299838e-01 -2.30453044e-01
8.50569308e-01 -8.93059552e-01 6.11063428e-02 -2.57515818e-01
-5.93164563e-01 4.26164448e-01 7.78720379e-01 2.17362300e-01
4.77103055e-01 -1.21065640e+00 -5.04892707e-01 5.71914911e-01
5.02630770e-01 4.32324171e-01 1.68246746e-01 8.37574899e-01
-9.45619643e-01 3.02600831e-01 -1.37777254e-01 -1.05970621e+00
-1.09562874e+00 4.82289225e-01 1.93484485e-01 -3.69841725e-01
-8.82232666e-01 3.90033543e-01 6.92505717e-01 -4.04979736e-01
2.92598635e-01 -1.41900927e-01 -4.77746189e-01 -5.39621599e-02
1.26608953e-01 6.45413101e-02 1.85645074e-01 -8.66070569e-01
-1.58441484e-01 1.31399381e+00 9.65251625e-02 8.48372746e-03
1.06422973e+00 -3.82384300e-01 -6.66398555e-02 3.80482137e-01
1.19689274e+00 1.80575043e-01 -1.26785731e+00 -2.97889799e-01
-6.14194870e-02 -7.62283027e-01 -1.01304606e-01 7.55387470e-02
-1.23573649e+00 9.09243643e-01 5.25291085e-01 -1.53508902e-01
1.09256840e+00 -3.18408996e-01 7.03480840e-01 4.03848737e-01
8.97450864e-01 -1.19762897e+00 -4.93885040e-01 4.66252774e-01
3.40557367e-01 -1.50346947e+00 1.49382740e-01 -1.18465889e+00
-7.13409185e-02 1.31438088e+00 6.57588542e-01 -2.85240382e-01
6.08372509e-01 8.34926590e-02 1.52719572e-01 -1.20058611e-01
-2.22217605e-01 -7.92358994e-01 7.69961849e-02 5.62955320e-01
1.31650761e-01 -3.24981600e-01 -7.50338256e-01 -3.04348648e-01
-3.07092130e-01 -8.31147004e-03 2.13568598e-01 1.01537335e+00
-8.49618912e-01 -1.52200162e+00 -4.04222161e-01 3.38325769e-01
2.48014361e-01 7.41301775e-02 -1.20301414e-02 7.39880681e-01
6.15536511e-01 9.34317112e-01 2.36567736e-01 -4.52429593e-01
4.25123006e-01 -1.25491515e-01 4.77143437e-01 -5.49981177e-01
-9.37897936e-02 2.52380043e-01 2.40894929e-01 -6.33355081e-01
-7.88803041e-01 -6.47476792e-01 -1.08097398e+00 -4.02654782e-02
-5.75558186e-01 2.37345994e-01 4.65290874e-01 1.10903251e+00
7.63322115e-02 6.79344162e-02 8.78977001e-01 -1.09885561e+00
-8.34867079e-03 -6.72253668e-01 -4.20388401e-01 5.97388923e-01
2.11604282e-01 -1.19123292e+00 -6.73281610e-01 3.85552160e-02] | [8.7558012008667, -2.3744447231292725] |
67d372dd-c995-40c4-887e-a1faf99f0319 | deterministic-policy-optimization-by | 1711.08068 | null | http://arxiv.org/abs/1711.08068v1 | http://arxiv.org/pdf/1711.08068v1.pdf | Deterministic Policy Optimization by Combining Pathwise and Score Function Estimators for Discrete Action Spaces | Policy optimization methods have shown great promise in solving complex
reinforcement and imitation learning tasks. While model-free methods are
broadly applicable, they often require many samples to optimize complex
policies. Model-based methods greatly improve sample-efficiency but at the cost
of poor generalization, requiring a carefully handcrafted model of the system
dynamics for each task. Recently, hybrid methods have been successful in
trading off applicability for improved sample-complexity. However, these have
been limited to continuous action spaces. In this work, we present a new hybrid
method based on an approximation of the dynamics as an expectation over the
next state under the current policy. This relaxation allows us to derive a
novel hybrid policy gradient estimator, combining score function and pathwise
derivative estimators, that is applicable to discrete action spaces. We show
significant gains in sample complexity, ranging between $1.7$ and $25\times$,
when learning parameterized policies on Cart Pole, Acrobot, Mountain Car and
Hand Mass. Our method is applicable to both discrete and continuous action
spaces, when competing pathwise methods are limited to the latter. | ['Daniel Levy', 'Stefano Ermon'] | 2017-11-21 | null | null | null | null | ['acrobot'] | ['playing-games'] | [ 8.29619840e-02 -6.10918961e-02 -5.51411092e-01 4.33419943e-02
-9.00086999e-01 -6.26780570e-01 7.14336395e-01 -8.90718475e-02
-8.88110578e-01 1.23829436e+00 -2.65024662e-01 -5.15910447e-01
-3.39481443e-01 -4.11403626e-01 -6.01972044e-01 -7.09987760e-01
-1.87137723e-01 5.39949596e-01 2.64170676e-01 -2.42659315e-01
4.72558588e-01 5.67843974e-01 -1.30669165e+00 -4.43519056e-01
1.06206751e+00 8.82240474e-01 3.43572378e-01 7.10345268e-01
2.98920888e-02 7.28853226e-01 -4.66409236e-01 1.08629495e-01
3.86696726e-01 -6.42822981e-01 -5.22070169e-01 8.38199630e-02
7.12252557e-02 -3.95946831e-01 -2.49279186e-01 1.08020937e+00
5.96497357e-01 6.03386045e-01 6.22221708e-01 -1.37669444e+00
2.18819221e-03 2.47369632e-01 -4.71193731e-01 7.86543638e-02
1.28292009e-01 4.21064496e-01 8.35024357e-01 -4.50506508e-01
5.00659168e-01 1.32548654e+00 4.51485366e-01 6.07888460e-01
-1.39043331e+00 -4.78014499e-01 5.52488029e-01 2.24692538e-01
-8.95680845e-01 -2.44998589e-01 4.22852844e-01 -3.94394964e-01
1.11055756e+00 -6.75080866e-02 6.94138765e-01 9.62533772e-01
3.70873570e-01 9.21058774e-01 1.43624389e+00 -2.61200070e-01
6.29554152e-01 9.50435624e-02 -4.32820112e-01 7.84784734e-01
-4.34287600e-02 3.43706578e-01 -2.54151933e-02 -2.24559799e-01
9.16941226e-01 4.88311611e-02 -1.65386915e-01 -7.99429417e-01
-1.00369942e+00 1.02718651e+00 1.74723685e-01 -4.84264642e-02
-5.31698942e-01 3.66746575e-01 3.43266636e-01 5.29835999e-01
2.98422575e-01 7.12420881e-01 -4.89661843e-01 -6.31229818e-01
-7.42637694e-01 9.31641579e-01 9.83012259e-01 8.24758828e-01
5.54354310e-01 3.01714748e-01 -1.70328334e-01 7.17355549e-01
3.19938958e-02 5.80666602e-01 3.55858713e-01 -1.35395432e+00
6.03459299e-01 2.27587402e-01 8.63289475e-01 -5.22800624e-01
-4.48937833e-01 -3.27781856e-01 -3.81395638e-01 7.45716095e-01
6.99140131e-01 -4.50828403e-01 -7.50241280e-01 1.83999884e+00
5.47119796e-01 -1.48177892e-01 -6.83665425e-02 6.68646574e-01
-5.25702477e-01 5.73386371e-01 -7.34607205e-02 -5.06994605e-01
8.00025821e-01 -1.11686969e+00 -5.68490684e-01 -2.95291722e-01
4.81695652e-01 -4.68711793e-01 1.30559826e+00 5.59681356e-01
-1.25588918e+00 -2.39839286e-01 -9.07270670e-01 3.96456152e-01
-1.62917480e-01 9.37434658e-03 2.02476561e-01 4.45775867e-01
-9.06297803e-01 1.18260574e+00 -1.15536118e+00 -1.81345239e-01
3.43877107e-01 5.24099946e-01 2.36971572e-01 1.85240373e-01
-8.07382882e-01 1.29866266e+00 2.74103552e-01 -2.19335541e-01
-1.10185051e+00 -4.17780310e-01 -5.88562310e-01 8.66741017e-02
1.01698804e+00 -3.03694606e-01 1.82997453e+00 -7.57358432e-01
-2.29024553e+00 -8.38396102e-02 8.05723816e-02 -6.53208792e-01
9.34797943e-01 -3.24806333e-01 -1.18118249e-01 2.80458312e-02
9.64673832e-02 3.52731168e-01 1.05594659e+00 -8.17124188e-01
-9.05445337e-01 -1.42829388e-01 2.90817291e-01 4.53440011e-01
-1.11647487e-01 -1.96800888e-01 -2.76479833e-02 -5.04305542e-01
-3.62995058e-01 -1.33779657e+00 -5.40923953e-01 7.63124302e-02
1.59604058e-01 -1.90844148e-01 7.48039186e-01 -4.19953644e-01
1.05023956e+00 -1.72999346e+00 2.81677932e-01 8.08302090e-02
-3.23745459e-01 4.21724826e-01 -1.06716603e-01 5.66201389e-01
4.36262578e-01 -1.65675536e-01 -3.73424560e-01 -1.08493611e-01
2.70013094e-01 3.07177126e-01 -3.60127151e-01 5.35648525e-01
2.10213631e-01 7.14893937e-01 -1.12482691e+00 -3.12005132e-01
2.43960440e-01 9.78699178e-02 -7.87701726e-01 1.80307552e-02
-5.62283039e-01 5.27182162e-01 -7.02473342e-01 2.81030446e-01
2.20796704e-01 -1.75050762e-03 3.12379301e-01 5.93559384e-01
-3.36690992e-01 2.45886385e-01 -1.24608397e+00 1.41156423e+00
-6.91311061e-01 3.20558488e-01 4.37050313e-01 -1.38203645e+00
7.30086327e-01 1.87695503e-01 6.25864863e-01 -5.53733230e-01
9.64674056e-02 3.46887112e-01 9.60001573e-02 -3.79410684e-01
4.01535451e-01 -4.70083952e-01 6.76010326e-02 4.57251877e-01
-1.14390716e-01 -6.26435280e-01 3.54874492e-01 -2.32440248e-01
9.48679209e-01 6.55390918e-01 5.12729883e-01 -3.90218019e-01
4.14341301e-01 1.61228999e-01 4.74334329e-01 9.06719446e-01
-3.75025570e-01 -4.41813320e-02 7.12767184e-01 -1.07224606e-01
-1.06694078e+00 -8.84000897e-01 2.15215191e-01 9.66368675e-01
1.39545009e-01 -6.77899420e-02 -5.61933398e-01 -7.34273434e-01
3.33662480e-01 8.83625686e-01 -3.59312326e-01 -1.32926881e-01
-7.90439069e-01 -4.02656794e-01 9.40373242e-02 4.57618058e-01
3.97183627e-01 -9.64306056e-01 -9.33491886e-01 6.56411648e-01
2.71999687e-01 -8.94736767e-01 -6.06390953e-01 7.81162232e-02
-1.18725443e+00 -9.07996178e-01 -9.15891886e-01 -2.99482733e-01
3.85659337e-01 -1.13531627e-01 6.02784574e-01 -4.57963437e-01
-1.10333622e-01 7.09800839e-01 -2.14624614e-03 -5.06350815e-01
-3.64543557e-01 6.69468641e-02 4.84788030e-01 -1.49640188e-01
-2.08895788e-01 -3.57331008e-01 -5.86999357e-01 4.86695647e-01
-5.61159551e-01 -2.73987025e-01 4.59604502e-01 1.16794240e+00
4.77649987e-01 -1.51531026e-01 6.43811762e-01 -5.89695036e-01
9.27990794e-01 -2.76313454e-01 -1.16616845e+00 1.08804040e-01
-1.07656884e+00 5.09207129e-01 1.00915444e+00 -8.66684020e-01
-1.03451777e+00 2.39053443e-02 1.52773395e-01 -4.41533685e-01
2.13816226e-01 3.00350428e-01 2.66753078e-01 -1.68675154e-01
6.38036788e-01 3.04082096e-01 6.00844264e-01 -4.71986502e-01
2.86689073e-01 3.85855764e-01 1.00531697e-01 -9.06022370e-01
4.17651325e-01 2.73555279e-01 1.38545588e-01 -7.52526522e-01
-5.12414515e-01 -2.39407092e-01 -1.91660047e-01 -1.35003179e-01
4.59411353e-01 -5.32059848e-01 -9.71628249e-01 2.58241147e-01
-5.54913402e-01 -8.73751700e-01 -5.23349345e-01 7.18393147e-01
-1.24131572e+00 3.27365309e-01 -4.26521569e-01 -1.16853511e+00
2.67574769e-02 -1.23116219e+00 6.97059810e-01 2.61095971e-01
4.65924777e-02 -8.58833432e-01 1.88389122e-01 -2.22944602e-01
4.69887376e-01 1.44621730e-01 6.50239587e-01 -3.85995865e-01
-3.35426897e-01 -3.01923364e-01 1.61752209e-01 4.65179205e-01
7.21029863e-02 -3.43482196e-01 -3.56142074e-01 -8.36274385e-01
9.17870328e-02 -4.78056878e-01 5.43899179e-01 4.96407151e-01
1.10105217e+00 -6.00202918e-01 -4.06149328e-01 1.36882871e-01
1.37433982e+00 7.14425504e-01 1.75683245e-01 4.95404661e-01
2.58749217e-01 4.06104177e-01 1.04888940e+00 6.31250203e-01
3.58684659e-02 7.35972345e-01 3.16606432e-01 3.29435855e-01
4.62402612e-01 -3.63981009e-01 7.17655778e-01 2.89995641e-01
-1.39381394e-01 2.93199997e-02 -6.73679292e-01 4.67314094e-01
-2.06611991e+00 -1.03557932e+00 5.44595420e-01 2.44302082e+00
8.02000821e-01 2.95366198e-01 6.59624338e-01 -1.67579845e-01
5.51835299e-01 -3.46703604e-02 -1.11714470e+00 -4.65823770e-01
4.83326077e-01 2.72368044e-01 8.41138780e-01 7.75211155e-01
-9.08530176e-01 8.60749602e-01 6.55868292e+00 1.00519657e+00
-1.07300663e+00 -4.93482035e-03 2.01061547e-01 -4.13210094e-01
2.41816983e-01 2.20589101e-01 -8.66855145e-01 5.00337303e-01
9.12020445e-01 -3.13533902e-01 8.87198925e-01 1.12114310e+00
5.03558218e-01 -3.18723530e-01 -8.97410631e-01 7.71011829e-01
-5.24335027e-01 -9.55742836e-01 -5.46274722e-01 3.33901495e-01
8.13850999e-01 -9.67346951e-02 1.27169341e-01 7.60682821e-01
4.28683668e-01 -7.46972919e-01 6.09705329e-01 4.13778841e-01
5.35366237e-01 -1.01529026e+00 2.94184297e-01 7.21408725e-01
-9.21727896e-01 -5.92657089e-01 -3.43684167e-01 -4.12378520e-01
2.47423947e-01 -8.06735978e-02 -8.82488489e-01 1.71505406e-01
2.39963830e-01 3.95487130e-01 1.13238864e-01 9.30348098e-01
-1.15031727e-01 3.82730007e-01 -4.94583130e-01 -6.27580881e-01
7.65287936e-01 -5.71388125e-01 6.47538722e-01 7.88729250e-01
3.58354598e-01 6.61229119e-02 6.15015864e-01 7.21470118e-01
4.87996072e-01 1.06795758e-01 -6.71987057e-01 -3.19398224e-01
2.58673519e-01 9.47865367e-01 -6.05094969e-01 -3.54619205e-01
-2.88074166e-01 7.42212415e-01 5.22839189e-01 4.42189068e-01
-8.86824071e-01 -5.98499775e-01 7.53848672e-01 5.29835373e-02
5.43236136e-01 -6.46570861e-01 1.91040218e-01 -1.09416258e+00
-4.88711074e-02 -1.07561803e+00 1.69143856e-01 -8.33653063e-02
-8.43870759e-01 6.26396537e-02 3.52432519e-01 -1.32881427e+00
-9.12767529e-01 -7.98596382e-01 -3.39462221e-01 7.22114444e-01
-1.12008452e+00 -3.01615268e-01 4.53728765e-01 4.84694839e-01
7.58992851e-01 -2.76039481e-01 5.05499601e-01 1.04296841e-02
-5.24540544e-01 4.09646451e-01 7.93166578e-01 -4.58582431e-01
4.13722128e-01 -1.40223551e+00 1.73997581e-01 5.47600269e-01
-2.54058152e-01 2.53624409e-01 9.17081535e-01 -5.43348908e-01
-1.44055367e+00 -7.34725654e-01 2.25290656e-01 -1.23272873e-01
8.46523762e-01 -1.45086899e-01 -6.50214612e-01 4.76362944e-01
-1.05656259e-01 -1.77074373e-01 -6.01966903e-02 -7.02091753e-02
2.37893149e-01 -7.05072656e-02 -1.20708656e+00 9.39102709e-01
8.33297729e-01 -3.45958561e-01 -2.80925363e-01 2.25746363e-01
3.66178960e-01 -4.43900585e-01 -8.75124574e-01 2.42938772e-01
5.76393068e-01 -5.84701180e-01 8.15078378e-01 -8.54246974e-01
4.32257280e-02 -2.28005394e-01 1.15752965e-01 -1.55329120e+00
-1.70823902e-01 -1.12484276e+00 -4.20817256e-01 4.89471167e-01
3.25711608e-01 -1.01786411e+00 7.28594899e-01 5.19790888e-01
-2.72112014e-03 -1.09173822e+00 -1.03320467e+00 -1.41507232e+00
2.70093471e-01 -1.74034104e-01 2.93696076e-01 4.68376935e-01
2.56650478e-01 2.33002245e-01 -4.99094784e-01 -2.45774448e-01
6.29971206e-01 1.40074357e-01 7.70124733e-01 -7.15125263e-01
-7.20185399e-01 -8.21438432e-01 2.21804511e-02 -1.39991796e+00
2.24001259e-01 -4.17657226e-01 3.23823214e-01 -1.27182591e+00
-2.92883039e-01 -4.38359678e-01 -2.09580392e-01 2.41485491e-01
-2.11651921e-01 -5.03157794e-01 4.06459033e-01 6.09554127e-02
-3.44268233e-01 8.84670258e-01 1.49164891e+00 6.32172823e-02
-5.28863192e-01 5.10676563e-01 -1.80695444e-01 6.54412270e-01
1.12045300e+00 -3.95084918e-01 -7.44065225e-01 -8.09318796e-02
-1.72040999e-01 4.37511921e-01 1.43079355e-01 -9.46084023e-01
-1.10883387e-02 -6.10226691e-01 -4.57974970e-02 -3.39953572e-01
4.94056076e-01 -5.31623363e-01 -1.38049096e-01 9.23674226e-01
-5.25424182e-01 2.42432386e-01 2.20715210e-01 1.02036905e+00
5.85277332e-04 -3.97004873e-01 1.02078021e+00 -2.97112167e-01
-5.46247423e-01 3.79152089e-01 -5.43020129e-01 2.50155330e-01
1.08203268e+00 -8.83193538e-02 1.72773927e-01 -6.35915220e-01
-7.78427899e-01 4.25790519e-01 4.24604326e-01 2.00597063e-01
3.45406681e-01 -1.14475727e+00 -3.30353737e-01 -1.86165988e-01
-4.64067608e-01 -3.30803305e-01 -1.20844804e-01 9.50128019e-01
-3.20010960e-01 4.85251576e-01 -3.00880611e-01 -3.88056636e-01
-9.27067995e-01 7.26044476e-01 4.40253735e-01 -5.49709797e-01
-5.99635720e-01 5.27010620e-01 -1.13320045e-01 -4.00356412e-01
2.88888007e-01 -4.70081925e-01 1.56050161e-01 -7.28367046e-02
3.09462219e-01 7.70502567e-01 -3.79462868e-01 1.95989963e-02
-8.96748751e-02 4.17044312e-01 -1.91375439e-03 -7.20932841e-01
1.15523148e+00 9.54225212e-02 3.12398136e-01 4.76014018e-01
9.62146938e-01 -4.20376569e-01 -2.03338099e+00 -2.33380169e-01
2.32022986e-01 -4.92893070e-01 -1.19420975e-01 -6.30278230e-01
-5.58454752e-01 7.41326153e-01 6.30894601e-01 3.45436335e-01
9.04327929e-01 -5.34296453e-01 5.70803165e-01 6.18679702e-01
6.12881780e-01 -1.52406204e+00 7.57701024e-02 7.05592513e-01
7.28505433e-01 -1.09919357e+00 1.48543969e-01 2.29013175e-01
-8.29148948e-01 9.55042422e-01 5.50875604e-01 -5.12762845e-01
3.91919136e-01 1.31910229e-02 -3.23299885e-01 2.60300696e-01
-8.12291741e-01 -2.37463161e-01 2.86169667e-02 4.93454248e-01
8.65907371e-02 2.38823339e-01 -6.10888362e-01 -1.40362442e-01
1.19670413e-01 2.52990890e-02 3.47882926e-01 1.39135849e+00
-6.54071987e-01 -1.25915444e+00 -3.08581442e-01 5.48683226e-01
-4.37323779e-01 2.41209581e-01 2.34405454e-02 1.10003138e+00
-3.75631392e-01 7.05739200e-01 -1.34527385e-01 1.71813652e-01
3.85004818e-01 1.48153469e-01 8.53474855e-01 -3.13919187e-01
-2.74148345e-01 2.49207452e-01 1.31244972e-01 -8.98544371e-01
-8.97241011e-02 -9.09905612e-01 -1.02841043e+00 -7.50569925e-02
-2.75235176e-01 2.56965518e-01 7.34341681e-01 9.55822229e-01
3.22141826e-01 8.45924988e-02 7.93782175e-01 -1.12695789e+00
-1.66363740e+00 -9.57376420e-01 -6.03972435e-01 1.55206993e-02
5.43685615e-01 -1.05833292e+00 -2.13220030e-01 -4.74241495e-01] | [4.18212366104126, 2.261744499206543] |
204ed974-dd48-4bc2-8f0b-02d8e329d77e | hierarchical-personalized-federated-learning | 2303.10580 | null | https://arxiv.org/abs/2303.10580v1 | https://arxiv.org/pdf/2303.10580v1.pdf | Hierarchical Personalized Federated Learning Over Massive Mobile Edge Computing Networks | Personalized Federated Learning (PFL) is a new Federated Learning (FL) paradigm, particularly tackling the heterogeneity issues brought by various mobile user equipments (UEs) in mobile edge computing (MEC) networks. However, due to the ever-increasing number of UEs and the complicated administrative work it brings, it is desirable to switch the PFL algorithm from its conventional two-layer framework to a multiple-layer one. In this paper, we propose hierarchical PFL (HPFL), an algorithm for deploying PFL over massive MEC networks. The UEs in HPFL are divided into multiple clusters, and the UEs in each cluster forward their local updates to the edge server (ES) synchronously for edge model aggregation, while the ESs forward their edge models to the cloud server semi-asynchronously for global model aggregation. The above training manner leads to a tradeoff between the training loss in each round and the round latency. HPFL combines the objectives of training loss minimization and round latency minimization while jointly determining the optimal bandwidth allocation as well as the ES scheduling policy in the hierarchical learning framework. Extensive experiments verify that HPFL not only guarantees convergence in hierarchical aggregation frameworks but also has advantages in round training loss maximization and round latency minimization. | ['Tony Q. S. Quek', 'Howard H. Yang', 'Kun Guo', 'Chaoqun You'] | 2023-03-19 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [-6.33582056e-01 -7.62819499e-03 -3.17819029e-01 8.47553685e-02
-6.03191018e-01 -3.36590230e-01 -2.56585658e-01 -4.73947048e-01
-4.75794971e-02 1.07546294e+00 -2.75128186e-01 -5.76298058e-01
-5.32701671e-01 -8.20010662e-01 -4.64445442e-01 -9.97369528e-01
-5.00173688e-01 4.39786881e-01 1.72597721e-01 2.92113334e-01
-4.66245115e-01 6.60476923e-01 -1.34091914e+00 1.34818658e-01
1.01486301e+00 1.65298486e+00 3.69951844e-01 6.75653696e-01
-7.95131698e-02 1.14743757e+00 -1.23445697e-01 -6.36707902e-01
3.31491768e-01 -8.39816108e-02 -6.80447459e-01 1.06965549e-01
-3.04624200e-01 -1.87635019e-01 -4.75093007e-01 8.31087053e-01
8.19838941e-01 2.45845973e-01 4.72246530e-03 -1.77959108e+00
1.46876305e-01 5.12178600e-01 -3.68633926e-01 3.30365956e-01
-3.42017651e-01 -2.35606387e-01 7.62662768e-01 -8.03047478e-01
4.79035407e-01 5.91867983e-01 7.56471157e-01 5.01429558e-01
-5.20892382e-01 -7.74055064e-01 4.82933313e-01 7.13318050e-01
-1.29235184e+00 -8.76223564e-01 5.42203546e-01 -1.26425862e-01
5.85731566e-01 5.67577243e-01 5.41407764e-01 3.76074761e-01
2.25446131e-02 1.02477050e+00 4.61098731e-01 -3.77875388e-01
5.10188222e-01 5.16032517e-01 -2.21058622e-01 6.45843148e-01
3.46099108e-01 -1.05063297e-01 -5.36526620e-01 -2.78099984e-01
6.43804073e-01 2.18350347e-02 -3.82006824e-01 -5.19970715e-01
-7.37944782e-01 3.79273683e-01 3.00674140e-01 1.35874599e-01
-9.24742460e-01 5.58321737e-02 4.92255151e-01 4.83920246e-01
6.71804190e-01 -3.30635130e-01 -8.80688190e-01 -1.64806023e-01
-1.12838984e+00 -2.56356359e-01 1.06500185e+00 1.30048776e+00
7.32144058e-01 2.61183456e-02 -8.56350884e-02 5.08772850e-01
1.72460347e-01 3.25253159e-02 2.53732771e-01 -1.27864730e+00
4.17448133e-01 3.84782732e-01 3.18024844e-01 -7.62557030e-01
-1.48929462e-01 -1.16596723e+00 -1.06934428e+00 -2.23987207e-01
-2.12218493e-01 -1.03060329e+00 2.02325389e-01 1.85667706e+00
7.60816038e-01 7.66594112e-01 -1.42768044e-02 9.29216683e-01
3.80515397e-01 4.42868948e-01 -2.32943073e-01 -8.82672727e-01
9.18647230e-01 -1.52167988e+00 -6.15269303e-01 1.11137576e-01
9.68112528e-01 -5.64995348e-01 3.08132321e-01 3.39665443e-01
-1.49887121e+00 -1.80215254e-01 -7.04004228e-01 4.70379233e-01
3.61881666e-02 3.30169797e-01 6.04093969e-01 7.70596325e-01
-1.35005772e+00 3.99145365e-01 -7.12152839e-01 -2.06027046e-01
4.26624566e-01 6.33671761e-01 -6.14240253e-03 -5.11103123e-02
-9.30363774e-01 1.19542785e-01 9.03388411e-02 3.95391881e-02
-5.51099420e-01 -9.18701291e-01 -3.18499267e-01 5.32979846e-01
5.04536211e-01 -1.10329783e+00 1.40594327e+00 -9.91648793e-01
-1.30652905e+00 1.60655335e-01 -2.42683023e-01 -3.66223544e-01
7.95721829e-01 2.22434878e-01 -7.34570086e-01 6.26861751e-02
-8.78258497e-02 1.30805895e-02 5.11088133e-01 -1.29224229e+00
-1.43563437e+00 -2.76280761e-01 2.64669895e-01 3.17170531e-01
-7.79556692e-01 -9.38418582e-02 -8.01099241e-01 -2.85509139e-01
-9.49132368e-02 -8.43467176e-01 -3.47459227e-01 -1.56465724e-01
-2.37898733e-02 -1.86722443e-01 1.27325356e+00 -5.81878901e-01
1.71587086e+00 -2.14709854e+00 3.00714411e-02 2.50911355e-01
5.29442906e-01 1.09815218e-01 2.08509684e-01 2.13397443e-01
3.59898925e-01 -9.85632464e-02 4.26539391e-01 -7.23792434e-01
5.36849611e-02 1.69262156e-01 1.13469757e-01 2.29496464e-01
-7.82693267e-01 7.42554784e-01 -8.77941787e-01 -7.85638571e-01
-1.48054823e-01 3.91916186e-01 -8.05035174e-01 4.01841134e-01
7.44375065e-02 4.53580290e-01 -5.32571852e-01 7.68092752e-01
7.36797929e-01 -7.03847826e-01 4.32585388e-01 -3.96818280e-01
-1.15209147e-01 -1.56704739e-01 -1.40083456e+00 1.24242437e+00
-8.78663599e-01 2.17364892e-01 8.98682356e-01 -1.06792498e+00
3.00300807e-01 8.82061899e-01 1.12859273e+00 -3.00101459e-01
1.49353877e-01 4.14819151e-01 -4.32028949e-01 -3.61620516e-01
2.26148516e-01 2.55656168e-02 3.12974185e-01 5.84007561e-01
1.22811936e-01 9.95694041e-01 1.92057297e-01 3.18993151e-01
9.04733896e-01 -1.90755308e-01 -7.53683299e-02 -2.01280817e-01
6.98309183e-01 -4.99215722e-01 1.14772725e+00 4.76215333e-01
-3.88478696e-01 -4.54351194e-02 2.10622773e-01 -4.08662856e-01
-4.63919312e-01 -8.91301632e-01 2.76606202e-01 1.23817420e+00
3.53208035e-01 -4.73193347e-01 -7.84716368e-01 -8.46183300e-01
-2.29043067e-01 3.90404791e-01 -8.37768465e-02 -1.93903327e-01
-1.14692360e-01 -6.30164206e-01 9.22718495e-02 2.78357893e-01
9.10329878e-01 -5.90412021e-01 -4.36918080e-01 5.43822348e-01
-5.44786334e-01 -1.22827804e+00 -9.05478001e-01 7.29698613e-02
-7.89123237e-01 -9.99502957e-01 -3.75592351e-01 -8.10331762e-01
5.45857370e-01 6.72255516e-01 1.00928164e+00 1.71841700e-02
8.50914493e-02 6.64140761e-01 -3.32675904e-01 -1.00937106e-01
1.27890874e-02 3.79983753e-01 3.20530951e-01 6.79571629e-01
-6.17432296e-02 -8.08974266e-01 -8.05698991e-01 4.90359247e-01
-4.44641352e-01 8.27927291e-02 4.78528798e-01 6.47587001e-01
7.28152692e-01 7.47265399e-01 9.50361729e-01 -9.98701572e-01
3.74637514e-01 -1.00647438e+00 -2.89664656e-01 6.65361047e-01
-8.31573367e-01 -5.92041969e-01 7.67979980e-01 -8.08368921e-02
-1.29909027e+00 -1.61519259e-01 2.42916673e-01 -7.22543001e-01
2.19436064e-01 4.16715890e-01 -6.08911991e-01 -3.83617729e-01
1.74660280e-01 7.05184191e-02 -2.23066270e-01 -4.04465139e-01
9.08971727e-02 1.02491069e+00 2.29109704e-01 -5.50869048e-01
5.19112825e-01 2.88775295e-01 1.94474235e-02 -4.92264897e-01
-7.07328260e-01 -4.41440195e-01 -1.74064294e-01 -6.12618625e-01
2.26029024e-01 -1.21737015e+00 -1.02858496e+00 1.41281158e-01
-8.73252153e-01 -3.09662104e-01 -3.81052673e-01 5.80279469e-01
-6.20368183e-01 1.18494600e-01 -7.48566031e-01 -9.50910151e-01
-7.77153909e-01 -9.00098264e-01 4.59406316e-01 4.29622680e-01
4.54149574e-01 -1.19503272e+00 -5.60797036e-01 3.84099394e-01
8.57919812e-01 -1.99709386e-01 6.58569932e-01 -2.46753141e-01
-8.14862967e-01 -3.08721066e-01 -3.04102480e-01 2.36173853e-01
1.91886336e-01 -3.98614615e-01 -9.29441154e-01 -8.32399070e-01
1.95415527e-01 1.46269694e-01 1.27361357e-01 4.81860161e-01
1.33903491e+00 -7.75466621e-01 -5.32673240e-01 9.07770455e-01
1.58492112e+00 8.64527971e-02 1.81886464e-01 1.46385983e-01
5.57151139e-01 4.20857042e-01 4.37612742e-01 8.68946195e-01
6.68777883e-01 5.83800554e-01 6.43342197e-01 -1.87442750e-01
-1.76145621e-02 1.39085755e-01 4.20243263e-01 1.18892145e+00
-6.59310892e-02 -4.60333079e-01 -2.68814087e-01 2.70849317e-01
-2.30672288e+00 -9.35647130e-01 1.70589745e-01 2.23682976e+00
3.53798836e-01 -7.67085403e-02 2.31088668e-01 1.67923748e-01
8.32261324e-01 -3.09546113e-01 -8.42049420e-01 -1.29012451e-01
-4.98448983e-02 -4.97691840e-01 5.32502294e-01 1.21102229e-01
-7.84195364e-01 4.56737846e-01 4.76039362e+00 1.07552397e+00
-1.23056412e+00 6.34461284e-01 7.90756941e-01 -2.83975542e-01
-6.75388649e-02 -1.51717916e-01 -5.58209002e-01 7.69082785e-01
1.19890070e+00 -6.39434874e-01 8.90177727e-01 1.07017648e+00
4.41369325e-01 3.59150887e-01 -7.09389269e-01 1.24774051e+00
-4.96024996e-01 -1.55082583e+00 -1.83524996e-01 2.60803103e-01
9.91883278e-01 3.78844172e-01 -2.88691968e-01 3.22298169e-01
1.25993043e-01 -3.02637339e-01 6.25689030e-01 7.63945401e-01
8.09111357e-01 -1.19235647e+00 8.19764853e-01 6.49543345e-01
-1.68295491e+00 -5.52057028e-01 -1.07538238e-01 1.60156488e-01
3.29203725e-01 8.48470688e-01 -1.47365987e-01 9.68023658e-01
8.95173192e-01 5.46052396e-01 -3.72026348e-03 1.47923648e+00
5.42812407e-01 4.52340305e-01 -2.32107446e-01 5.18938661e-01
-1.21529050e-01 -3.75187218e-01 5.07126689e-01 9.81289566e-01
5.84730089e-01 -3.76304947e-02 3.19750667e-01 4.05749768e-01
-5.06300807e-01 2.72618771e-01 -6.15131147e-02 3.12561631e-01
9.56609130e-01 1.85079086e+00 -3.94068837e-01 -4.08330739e-01
-7.74213195e-01 9.99576271e-01 3.27481717e-01 5.34209907e-01
-8.17903101e-01 -4.41255182e-01 8.69858444e-01 3.45242992e-02
3.13013196e-01 6.97402284e-02 -9.76472199e-02 -1.17778397e+00
1.04359396e-01 -5.42505383e-01 7.14886129e-01 -4.10501212e-01
-1.30407417e+00 8.32900345e-01 -7.13277102e-01 -1.40599537e+00
4.45669778e-02 -3.53259966e-02 -9.29661632e-01 5.54114401e-01
-1.85920918e+00 -1.15664983e+00 -3.85177374e-01 8.86799812e-01
4.78292167e-01 -3.17566752e-01 7.14958966e-01 1.13133848e+00
-1.07722270e+00 1.02267683e+00 3.01239938e-01 -4.15565908e-01
2.98655063e-01 -8.91664505e-01 -3.84758979e-01 9.49456573e-01
-2.91869402e-01 8.64553973e-02 2.82715678e-01 -1.74252152e-01
-1.33246148e+00 -1.52807391e+00 9.19166386e-01 3.51344228e-01
3.30358148e-01 -3.44237052e-02 -4.55758005e-01 6.81712627e-01
-1.40577883e-01 4.38604623e-01 9.62618172e-01 -1.17721736e-01
3.02873462e-01 -5.98071396e-01 -1.35772896e+00 5.34927845e-01
1.13861442e+00 -3.03629398e-01 7.23651290e-01 5.75929642e-01
8.99233997e-01 -2.37914324e-01 -1.01729822e+00 3.46326649e-01
1.92442983e-01 -9.38633919e-01 5.48612893e-01 -6.17055058e-01
-5.15786886e-01 -1.70580298e-01 -2.83665150e-01 -1.24801314e+00
-5.69484949e-01 -1.10702670e+00 -8.97785127e-01 1.33763337e+00
2.38696977e-01 -7.62162089e-01 1.17229664e+00 6.84455335e-01
-4.99279380e-01 -1.11328232e+00 -1.20137537e+00 -8.55084121e-01
-4.25638109e-01 -4.19893712e-01 9.42867696e-01 8.19002569e-01
3.81262526e-02 2.15742692e-01 -4.15458322e-01 4.95971680e-01
9.13435936e-01 1.59422576e-01 5.01798272e-01 -1.38608074e+00
-5.65376699e-01 -3.43062013e-01 5.15472703e-02 -1.10062826e+00
3.41515355e-02 -9.05049801e-01 -3.97229522e-01 -1.47597945e+00
-8.31591487e-02 -1.13500333e+00 -6.38159394e-01 1.74047455e-01
1.71968743e-01 -8.55173618e-02 2.81375051e-01 4.61582899e-01
-1.35281897e+00 4.10208553e-01 1.00110817e+00 2.83774853e-01
-3.07836920e-01 7.53755808e-01 -7.01531589e-01 5.05174339e-01
8.80481362e-01 -1.73755705e-01 -6.65655494e-01 -5.76648474e-01
7.02870339e-02 5.33310056e-01 3.90048660e-02 -1.10736299e+00
9.35373545e-01 -1.23054693e-02 2.97240354e-03 -3.29097390e-01
4.78009582e-02 -1.38790727e+00 4.54889208e-01 2.84226120e-01
3.77943814e-01 -8.47478658e-02 -1.23119481e-01 6.10167861e-01
9.72837768e-03 2.93430239e-01 6.12182736e-01 1.25205621e-01
-5.66678166e-01 1.04591215e+00 -2.41919309e-01 -1.11900970e-01
1.33004141e+00 -8.44423398e-02 1.23883346e-02 -5.78326762e-01
-1.07381320e+00 7.86788285e-01 1.77619949e-01 8.69643986e-02
1.67243555e-01 -1.30476320e+00 -3.31541151e-01 2.55801290e-01
-3.61519605e-01 3.61697678e-03 7.33451247e-01 1.49872577e+00
-9.43710655e-02 5.11931062e-01 2.15052724e-01 -1.63735166e-01
-1.00955546e+00 5.39334714e-01 8.23464274e-01 -4.30093288e-01
-2.76216090e-01 8.23951244e-01 -1.13236427e-04 -4.38472956e-01
5.84399223e-01 6.24884248e-01 1.58875018e-01 -5.14707863e-02
4.24231678e-01 1.18823433e+00 4.71245497e-01 -5.19614995e-01
-1.32574797e-01 -2.82940939e-02 1.58201709e-01 4.88554090e-01
1.23212552e+00 -9.64694738e-01 -2.62734264e-01 -3.22731808e-02
1.08678818e+00 6.51954068e-03 -1.21109450e+00 -5.49937963e-01
-7.30308965e-02 -2.26304010e-01 6.22107148e-01 -6.06677353e-01
-1.85762060e+00 2.02370763e-01 5.33706844e-01 3.01963925e-01
1.47413921e+00 -2.19412237e-01 1.31208706e+00 -1.43351089e-02
9.72806036e-01 -1.14705884e+00 -4.12445933e-01 3.08791101e-01
6.97511481e-03 -7.09616184e-01 -5.11116982e-01 -4.43258196e-01
-3.45537424e-01 9.47632313e-01 8.50028872e-01 4.48146105e-01
1.12891555e+00 3.61142725e-01 -2.47841388e-01 4.29527611e-02
-1.17319202e+00 -1.26382694e-01 -2.09773391e-01 4.07147169e-01
1.19982578e-01 2.19946936e-01 -2.90951669e-01 1.29525912e+00
1.67542577e-01 3.40465337e-01 1.74748391e-01 7.23546505e-01
-4.30046946e-01 -1.20159280e+00 -9.47528332e-02 7.57377028e-01
-4.21042264e-01 2.68674374e-01 4.51058656e-01 1.81126490e-01
4.96794790e-01 1.14197755e+00 1.10230945e-01 -4.89658028e-01
2.25175083e-01 -8.56805220e-02 -4.75101024e-02 -2.75931656e-01
-3.71148378e-01 1.58971846e-01 1.22820856e-02 -5.79799533e-01
-5.58531173e-02 -4.53201383e-01 -1.37853658e+00 -7.75150120e-01
-5.37945569e-01 7.69465566e-01 5.22646606e-01 8.91246915e-01
9.22485948e-01 6.94622338e-01 1.51293218e+00 -7.09063709e-01
-3.61668885e-01 -4.13070410e-01 -1.04941618e+00 -2.13598058e-01
2.67341435e-01 -3.44972730e-01 -6.03050172e-01 -2.15972528e-01] | [5.965803146362305, 5.583015441894531] |
362de42d-0c07-4b95-800f-4cb0ca9bb529 | joint-rumour-stance-and-veracity-prediction | null | null | https://aclanthology.org/W19-6122 | https://aclanthology.org/W19-6122.pdf | Joint Rumour Stance and Veracity Prediction | The net is rife with rumours that spread through microblogs and social media. Not all the claims in these can be verified. However, recent work has shown that the stances alone that commenters take toward claims can be sufficiently good indicators of claim veracity, using e.g. an HMM that takes conversational stance sequences as the only input. Existing results are monolingual (English) and mono-platform (Twitter). This paper introduces a stance-annotated Reddit dataset for the Danish language, and describes various implementations of stance classification models. Of these, a Linear SVM provides predicts stance best, with 0.76 accuracy / 0.42 macro F1. Stance labels are then used to predict veracity across platforms and also across languages, training on conversations held in one language and using the model on conversations held in another. In our experiments, monolinugal scores reach stance-based veracity accuracy of 0.83 (F1 0.68); applying the model across languages predicts veracity of claims with an accuracy of 0.82 (F1 0.67). This demonstrates the surprising and powerful viability of transferring stance-based veracity prediction across languages. | ['Leon Derczynski', 'Emil Refsgaard Middelboe', 'Anders Edelbo Lillie'] | null | null | null | null | ws-nodalida-2019-9 | ['rumour-detection'] | ['natural-language-processing'] | [-3.70644182e-01 4.91816044e-01 -8.04844737e-01 4.41158394e-04
-1.06605482e+00 -8.60801339e-01 1.04462445e+00 2.97161549e-01
-2.92097211e-01 1.09141397e+00 6.14873052e-01 -6.72142029e-01
4.58861440e-01 -7.10544348e-01 -4.02074397e-01 -1.83255002e-01
3.01226467e-01 6.72786534e-01 3.50586176e-01 -6.28043413e-01
6.60396874e-01 -3.11154183e-02 -1.09079909e+00 6.34995103e-01
7.95624912e-01 7.81960607e-01 -5.40367901e-01 8.98426116e-01
-2.62952656e-01 1.86854076e+00 -1.14025557e+00 -9.86323416e-01
-1.22982085e-01 -5.48035264e-01 -9.27033365e-01 -2.48974755e-01
2.07367912e-01 -1.37182891e-01 -3.04112226e-01 7.29695380e-01
2.35621497e-01 -6.15039766e-01 5.53641140e-01 -1.33256292e+00
-1.07098067e+00 1.13465726e+00 -3.87129635e-01 4.90142763e-01
7.06903934e-01 1.93706438e-01 1.11223972e+00 -5.09496272e-01
1.05733776e+00 8.77218843e-01 1.21775019e+00 9.40176994e-02
-9.55786407e-01 -5.03833413e-01 -5.70081770e-01 1.25563711e-01
-7.68904626e-01 -5.37701964e-01 7.17634380e-01 -1.05909300e+00
9.29940999e-01 4.61321414e-01 7.44255543e-01 1.55138385e+00
3.21888506e-01 5.85287750e-01 1.73334825e+00 -1.71575040e-01
-1.53232828e-01 5.50424278e-01 3.83630991e-01 5.33290684e-01
3.29133123e-01 -3.30289721e-01 -7.57154524e-01 -6.87199891e-01
1.10799901e-01 -4.59945738e-01 -1.71303749e-02 8.08475852e-01
-1.34271705e+00 1.39301848e+00 1.36285767e-01 4.15735185e-01
-4.30216491e-01 -2.11009935e-01 8.27504158e-01 7.93410659e-01
1.07608724e+00 4.80156571e-01 -7.40816221e-02 -6.95663452e-01
-1.44647789e+00 2.59666860e-01 1.66833520e+00 3.96051794e-01
3.34745228e-01 1.36907380e-02 -8.98109823e-02 1.06567287e+00
3.64165604e-02 6.66377664e-01 5.12877107e-01 -1.05698860e+00
5.91473043e-01 3.88664991e-01 2.34871373e-01 -1.25897694e+00
-4.94439304e-01 -4.40064967e-01 -3.40491861e-01 -8.27795267e-02
5.47237992e-01 -3.27208042e-01 -2.30236769e-01 1.15139794e+00
-3.98677476e-02 -2.92283744e-01 2.76559174e-01 4.53452021e-01
8.25188577e-01 6.05828106e-01 -1.52456850e-01 -6.85439289e-01
1.06614316e+00 -7.83649266e-01 -8.49945307e-01 -2.46304810e-01
8.00001800e-01 -1.24349499e+00 8.94835472e-01 3.52020979e-01
-1.20548892e+00 1.03500612e-01 -9.84374583e-01 4.81724367e-02
-6.64689466e-02 -3.80033970e-01 2.82162279e-01 8.99638712e-01
-9.48734581e-01 5.23681819e-01 -4.44434732e-01 -1.73020661e-01
4.96033162e-01 -3.94966871e-01 -6.04955517e-02 2.80665845e-01
-1.45271206e+00 1.36059558e+00 7.46740773e-02 -4.55130190e-01
-4.48030740e-01 -3.55911106e-01 -4.23561513e-01 -6.10520720e-01
1.55742541e-01 -1.38920784e-01 1.35691381e+00 -9.08679783e-01
-1.28409588e+00 1.13336611e+00 -2.02487018e-02 -7.56270230e-01
1.06406236e+00 1.45782754e-01 -7.18577921e-01 1.12071127e-01
7.83276141e-01 -2.08036110e-01 3.26082379e-01 -1.02799845e+00
-3.89868855e-01 1.60987135e-02 -1.03137121e-01 -2.64699221e-01
-1.77336991e-01 5.69393992e-01 5.05531013e-01 -4.62526798e-01
8.71731639e-02 -9.56207216e-01 5.12094140e-01 -5.80157876e-01
-5.50409555e-01 -4.55894083e-01 6.04366124e-01 -1.39520919e+00
1.49742198e+00 -1.33920944e+00 -5.59578657e-01 5.31931892e-02
4.20042455e-01 -1.07462145e-01 4.26235378e-01 9.23353374e-01
4.29763228e-01 6.49575293e-01 -1.67965591e-02 -1.49840534e-01
7.93591887e-03 1.13088861e-01 -4.04738992e-01 7.09760070e-01
-1.11281089e-01 9.47653949e-01 -6.19150400e-01 -5.37182391e-01
-5.40510476e-01 1.02440462e-01 -4.03082997e-01 -2.42701009e-01
-2.36044586e-01 2.19308347e-01 -6.42965082e-04 4.92731929e-01
1.74607292e-01 -6.62094414e-01 3.36252153e-01 3.03298324e-01
-5.17619371e-01 8.49790812e-01 -1.74688205e-01 7.87540853e-01
-2.95135260e-01 1.24937034e+00 -2.95370251e-01 -6.84633017e-01
1.25256276e+00 5.02041399e-01 5.20382881e-01 -6.52623653e-01
1.83507577e-01 7.07116783e-01 2.14748636e-01 -7.71486223e-01
5.56889474e-01 -5.15053511e-01 -3.92468184e-01 1.12653911e+00
-5.23859978e-01 4.22417782e-02 2.41045818e-01 3.11501712e-01
1.03204381e+00 -5.28315008e-01 4.42462534e-01 -1.05060628e-02
2.26887688e-01 6.60294831e-01 3.02947879e-01 6.71129048e-01
-3.62771839e-01 3.25869769e-01 1.11137891e+00 -2.52563566e-01
-1.52241826e+00 -4.66314465e-01 -2.03193635e-01 1.08697581e+00
-3.98554057e-01 -2.02600986e-01 -5.39104402e-01 -4.94071156e-01
1.92237481e-01 1.01015401e+00 -6.44182324e-01 4.71712738e-01
-6.04241252e-01 -6.96535885e-01 1.18358779e+00 6.45248666e-02
5.44058740e-01 -1.06103408e+00 -2.22115412e-01 4.39773858e-01
-8.88144672e-01 -1.21329165e+00 -3.46085951e-02 -3.69919121e-01
-4.02010947e-01 -1.13035476e+00 -4.93059635e-01 -4.43525404e-01
-3.59913498e-01 -9.88234207e-02 9.68650103e-01 4.18995470e-01
5.81079304e-01 -2.88859546e-01 -3.85146290e-01 -2.50997722e-01
-1.28035378e+00 1.74514040e-01 9.13981870e-02 -2.41408139e-01
3.16066891e-01 -4.58552510e-01 1.39888441e-02 2.01010466e-01
-3.37115496e-01 7.99918640e-03 1.13923959e-01 7.96508372e-01
-2.19274521e-01 -6.08150661e-01 1.13926041e+00 -1.34678996e+00
1.03164434e+00 -1.20691144e+00 1.09809726e-01 5.12807891e-02
-8.35156918e-01 -5.94672918e-01 2.99628705e-01 -3.84227365e-01
-5.64009607e-01 -1.09753847e+00 -4.20285583e-01 1.87301323e-01
4.15451556e-01 1.04719532e+00 7.67126203e-01 3.85961086e-01
1.26067591e+00 7.76145831e-02 5.22056460e-01 -2.09617049e-01
6.95468411e-02 1.37913084e+00 4.19236422e-01 -1.60816565e-01
4.59728122e-01 3.95652711e-01 -7.32007623e-01 -9.17761803e-01
-1.27174830e+00 -4.20302629e-01 -1.92622200e-01 -4.71382529e-01
5.72580040e-01 -9.44545805e-01 -1.04757476e+00 4.92901653e-01
-1.27793288e+00 -5.05779862e-01 1.00054383e-01 3.45715672e-01
-7.21963942e-01 2.75285870e-01 -1.28110218e+00 -9.26542819e-01
-3.01958054e-01 -5.75344265e-01 2.07470939e-01 -2.13022634e-01
-9.39147711e-01 -1.21418929e+00 5.16367376e-01 1.08201671e+00
5.33845723e-01 7.12603152e-01 5.20188868e-01 -1.20585024e+00
2.08135948e-01 -3.33800644e-01 -2.16026530e-01 2.09722713e-01
-7.94269070e-02 1.17293373e-01 -9.07473564e-01 4.42889472e-03
2.02501044e-01 -5.88142216e-01 5.86673260e-01 8.06603208e-02
5.01492247e-02 -1.25689530e+00 -5.37925661e-02 -2.53760070e-01
1.11741352e+00 -3.13436538e-01 5.31972110e-01 9.28588867e-01
3.30822885e-01 5.47279119e-01 2.34649435e-01 5.67689002e-01
7.11157382e-01 4.82904792e-01 1.00930464e-02 3.67066503e-01
-1.10787779e-01 -3.76641423e-01 6.57338679e-01 1.18014801e+00
-1.45640358e-01 -2.53402025e-01 -1.32119560e+00 4.75372612e-01
-1.70009279e+00 -1.70594668e+00 -9.14082289e-01 1.51298273e+00
1.24980485e+00 6.56953633e-01 7.39479899e-01 2.75993317e-01
7.31680751e-01 4.82872397e-01 -2.40201920e-01 -8.95967245e-01
-6.45661891e-01 -3.52621347e-01 5.17439902e-01 9.08399284e-01
-6.15985453e-01 6.43808603e-01 6.97276926e+00 4.55884486e-01
-1.21499777e+00 7.70929515e-01 4.49838191e-01 -1.26681700e-01
-6.21156394e-01 3.86372246e-02 -4.24860001e-01 8.01092446e-01
1.51407146e+00 -4.90885854e-01 1.79885790e-01 6.46470428e-01
3.59664619e-01 -1.49045259e-01 -3.38804126e-01 6.15977049e-01
5.26907861e-01 -1.97989094e+00 -6.43310547e-01 1.94782481e-01
1.19073033e+00 6.32874072e-01 -6.47731721e-02 3.90245467e-01
5.88181019e-01 -8.36025596e-01 1.30452240e+00 6.15790963e-01
7.53232241e-01 -4.06800389e-01 8.70504916e-01 8.78023267e-01
-2.75448319e-02 -1.97743863e-01 -1.25882104e-01 -5.33180714e-01
4.82828319e-01 9.10919368e-01 -1.33344042e+00 3.26833837e-02
2.95000583e-01 9.17056501e-01 -3.09135437e-01 5.03624678e-01
-3.60273212e-01 1.14750552e+00 -1.17138810e-01 -3.90915841e-01
3.10599893e-01 2.09561586e-01 6.62069082e-01 1.39538407e+00
1.88810810e-01 1.13109760e-02 2.05106556e-01 5.22143424e-01
-2.38336653e-01 1.48907140e-01 -5.74335396e-01 -3.82174760e-01
6.30325437e-01 8.62250149e-01 -3.88649821e-01 -5.89948237e-01
-4.28974420e-01 6.76946044e-01 3.38482440e-01 -7.18559027e-02
-8.12233567e-01 2.78966814e-01 2.17246726e-01 5.92002392e-01
-1.61284551e-01 -2.01451004e-01 -8.57313395e-01 -1.07201552e+00
-1.93231016e-01 -1.22471011e+00 2.30378389e-01 -7.33006239e-01
-1.66073430e+00 6.98762000e-01 -5.27898192e-01 -9.80715036e-01
-4.27424133e-01 -1.73540488e-01 -6.28680766e-01 9.80386972e-01
-1.29573488e+00 -1.29174757e+00 -3.04408316e-02 8.49555135e-02
3.16703647e-01 -3.27048063e-01 4.47690696e-01 5.28404079e-02
-3.17072749e-01 2.70952404e-01 -6.00600541e-02 4.50142413e-01
8.33309293e-01 -8.94235671e-01 4.04776096e-01 3.37102175e-01
-4.78519350e-02 4.37027276e-01 1.02135074e+00 -1.13090885e+00
-6.05291784e-01 -6.26083195e-01 1.80574131e+00 -1.05906832e+00
1.55047333e+00 1.67180926e-01 -9.95490551e-01 8.24745893e-01
3.91100764e-01 -6.88719213e-01 1.08331323e+00 4.55037802e-01
-8.06681395e-01 6.45367622e-01 -1.03132069e+00 3.12410563e-01
6.03544891e-01 -9.99902070e-01 -8.38906229e-01 8.50017726e-01
4.96510118e-01 -2.83640653e-01 -1.20704162e+00 -3.54012787e-01
8.13449800e-01 -1.17391741e+00 3.80164504e-01 -7.25188613e-01
1.00123918e+00 1.29294008e-01 -1.97853357e-01 -9.74177599e-01
-1.00660734e-01 -7.65877962e-01 -1.28161713e-01 1.11506224e+00
8.98793578e-01 -9.75153804e-01 5.85732460e-01 2.19684988e-01
-1.25336125e-01 -6.10991836e-01 -8.23378921e-01 -7.81811774e-01
6.24970853e-01 -5.05088508e-01 2.54582822e-01 1.52288997e+00
6.27928555e-01 5.08989334e-01 -4.93725955e-01 -2.44503021e-01
3.40769202e-01 1.61269814e-01 6.83848381e-01 -1.44486523e+00
-2.33331740e-01 -6.62723541e-01 -2.09714830e-01 -4.94953841e-01
2.68859088e-01 -1.03541350e+00 -2.86217988e-01 -1.58842194e+00
4.19989049e-01 -7.82561541e-01 4.83097881e-01 4.94801015e-01
1.25509694e-01 7.49845207e-01 1.30381256e-01 1.12985981e+00
-1.81179583e-01 -1.96132854e-01 9.31411624e-01 -1.04727291e-01
1.96341686e-02 1.65919170e-01 -9.44398046e-01 7.30579674e-01
1.25518918e+00 -7.12342441e-01 2.09883362e-01 -4.49032756e-03
9.21349704e-01 4.86853451e-01 5.30670941e-01 -6.62789524e-01
2.98741251e-01 -2.40733892e-01 -1.95429713e-01 -4.86851990e-01
2.12805375e-01 2.41565943e-01 2.33581796e-01 5.82366645e-01
-4.74145859e-01 2.04000667e-01 -2.54998654e-01 5.17526507e-01
-1.46665752e-01 -2.34240577e-01 6.07829630e-01 -1.08587928e-01
-4.14308235e-02 -2.88386792e-01 -3.73143673e-01 6.52988732e-01
6.41604841e-01 -3.53425354e-01 -1.18515086e+00 -8.78466487e-01
-8.68737996e-01 -2.81642288e-01 7.87094057e-01 1.93021148e-01
-2.09962782e-02 -1.03334570e+00 -1.58255935e+00 -3.25318247e-01
2.02385746e-02 -1.28565788e+00 -9.43260416e-02 1.65179181e+00
-6.76846683e-01 3.45626026e-01 -1.56907856e-01 -5.06212294e-01
-1.00793648e+00 1.97634399e-01 7.89884031e-02 -2.12015212e-01
-4.41579252e-01 6.48074031e-01 -9.28366601e-01 -2.39490598e-01
-6.78415418e-01 3.17849845e-01 -2.52999634e-01 7.07149684e-01
4.69220459e-01 5.37791669e-01 1.07274942e-01 -1.11241031e+00
-1.39263660e-01 -6.55177608e-02 6.87909722e-02 -6.00301206e-01
1.34081089e+00 -1.16903402e-01 -3.89111906e-01 1.17330837e+00
1.24811089e+00 7.05221951e-01 -6.18242502e-01 -1.83106303e-01
2.86187381e-01 -2.59830415e-01 -1.88937500e-01 -9.59738970e-01
-3.97388846e-01 4.98503894e-01 -4.41947013e-01 1.29987979e+00
-7.03691021e-02 2.20054224e-01 8.47811580e-01 -5.82979992e-02
1.84963420e-01 -1.26326191e+00 1.32207125e-01 7.87552774e-01
1.10081434e+00 -1.11823761e+00 7.17789978e-02 -2.23709136e-01
-1.14401317e+00 9.73124266e-01 -1.37519240e-01 -7.65696019e-02
4.96769071e-01 1.75592482e-01 4.08232093e-01 -4.37264413e-01
-8.56394112e-01 2.44723529e-01 -1.74389139e-01 3.30124199e-01
7.72168756e-01 4.02563691e-01 -8.49996865e-01 4.03489083e-01
-9.76587594e-01 -1.21271074e-01 1.26689816e+00 5.39230704e-01
-8.76834989e-01 -7.36133993e-01 -4.22306150e-01 8.43329608e-01
-8.32603395e-01 -1.51275873e-01 -6.50328338e-01 8.30591023e-01
-1.30611822e-01 1.42953539e+00 -2.00764704e-02 -5.09841859e-01
-3.41844000e-02 3.56773436e-01 -1.18745036e-01 -6.62094891e-01
-9.98993695e-01 -2.68057585e-01 1.11834669e+00 -6.45126775e-02
-5.53731441e-01 -1.10841405e+00 -9.22591925e-01 -1.38603866e+00
-3.13860089e-01 4.59515691e-01 5.59494257e-01 9.86342251e-01
-1.00245796e-01 -5.80716040e-03 7.33989656e-01 -2.08806638e-02
-7.52562463e-01 -1.40074408e+00 -3.93800735e-01 2.33415633e-01
3.26137841e-01 -3.84040207e-01 -7.28421569e-01 2.78801639e-02] | [8.290670394897461, 10.088238716125488] |
c1cd3d7d-3cc9-4246-b8a0-5151fa1e4774 | the-manifold-hypothesis-for-gradient-based-1 | 2206.07387 | null | https://arxiv.org/abs/2206.07387v1 | https://arxiv.org/pdf/2206.07387v1.pdf | The Manifold Hypothesis for Gradient-Based Explanations | When do gradient-based explanation algorithms provide meaningful explanations? We propose a necessary criterion: their feature attributions need to be aligned with the tangent space of the data manifold. To provide evidence for this hypothesis, we introduce a framework based on variational autoencoders that allows to estimate and generate image manifolds. Through experiments across a range of different datasets -- MNIST, EMNIST, CIFAR10, X-ray pneumonia and Diabetic Retinopathy detection -- we demonstrate that the more a feature attribution is aligned with the tangent space of the data, the more structured and explanatory it tends to be. In particular, the attributions provided by popular post-hoc methods such as Integrated Gradients, SmoothGrad and Input $\times$ Gradient tend to be more strongly aligned with the data manifold than the raw gradient. As a consequence, we suggest that explanation algorithms should actively strive to align their explanations with the data manifold. In part, this can be achieved by adversarial training, which leads to better alignment across all datasets. Some form of adjustment to the model architecture or training algorithm is necessary, since we show that generalization of neural networks alone does not imply the alignment of model gradients with the data manifold. | ['Ulrike Von Luxburg', 'Zeynep Akata', 'Uddeshya Upadhyay', 'Sebastian Bordt'] | 2022-06-15 | the-manifold-hypothesis-for-gradient-based | https://openreview.net/forum?id=dmq_-R2LhQk | https://openreview.net/pdf?id=dmq_-R2LhQk | null | ['diabetic-retinopathy-detection'] | ['medical'] | [-4.04304974e-02 5.53087294e-01 -1.95672438e-01 -6.99338436e-01
-3.42090474e-03 -4.04601604e-01 6.13001943e-01 -1.18872963e-01
-1.95894778e-01 5.53737938e-01 5.01085520e-01 -4.48952496e-01
-3.58972967e-01 -4.93145764e-01 -9.44909096e-01 -5.09093165e-01
1.48414612e-01 3.77950519e-01 -3.50262910e-01 -2.69845068e-01
4.32927370e-01 4.30325836e-01 -1.25150990e+00 5.82513064e-02
9.06032622e-01 3.85330796e-01 -1.58581689e-01 5.86475074e-01
-8.11450705e-02 5.84703445e-01 -2.18714148e-01 -5.37763953e-01
3.46871585e-01 -6.63828492e-01 -8.93888712e-01 4.55117971e-01
8.78537118e-01 -2.84971118e-01 -3.54059279e-01 1.18465853e+00
-4.70632687e-02 5.66588230e-02 1.02848017e+00 -1.41279745e+00
-9.78845537e-01 4.30875719e-01 -1.71316981e-01 1.86064854e-01
-1.73229948e-01 2.34778076e-01 1.12882805e+00 -8.97619486e-01
8.80195022e-01 1.18396890e+00 6.14707351e-01 8.84233832e-01
-1.51296139e+00 -1.19921699e-01 2.22822249e-01 1.06683716e-01
-9.88351464e-01 -1.67859972e-01 9.97525036e-01 -4.65518862e-01
5.06124198e-01 3.52120012e-01 6.98633015e-01 1.10425007e+00
2.24989742e-01 6.09857619e-01 8.61504793e-01 -2.77398050e-01
1.05007984e-01 4.74811196e-01 2.50576794e-01 8.81876051e-01
3.21262985e-01 1.85342684e-01 -4.63889956e-01 -3.75938639e-02
1.12450600e+00 1.19928084e-02 -5.82859516e-01 -8.51563454e-01
-1.08911943e+00 1.09441698e+00 8.41190040e-01 4.34387207e-01
-5.25375545e-01 3.22477549e-01 3.82040106e-02 2.41173774e-01
4.36279595e-01 1.10146511e+00 -5.07158935e-01 2.48086020e-01
-6.40399098e-01 2.66130805e-01 4.78010952e-01 4.66519564e-01
8.84360015e-01 2.89365411e-01 1.87210545e-01 3.87558490e-01
5.74636161e-01 4.87119518e-02 6.49804592e-01 -1.32246184e+00
2.18475088e-01 8.29785943e-01 2.80477051e-02 -1.21719718e+00
-4.40818459e-01 -5.33649683e-01 -7.56027460e-01 5.88206112e-01
7.46394992e-01 -1.35261789e-01 -8.41626644e-01 2.07340622e+00
3.00981939e-01 2.18746603e-01 3.21848333e-01 1.20195508e+00
4.43811834e-01 2.75532812e-01 2.38885414e-02 1.35236517e-01
1.06004870e+00 -7.43534982e-01 -6.29883587e-01 -2.91773587e-01
6.39402151e-01 -4.66929704e-01 1.34816325e+00 -7.21728131e-02
-9.09503102e-01 -6.57684207e-01 -1.20742261e+00 -1.10790744e-01
-2.39407972e-01 -9.43768322e-02 5.77902317e-01 2.87856102e-01
-9.84555125e-01 1.21134245e+00 -9.94611561e-01 -4.79369551e-01
3.28381300e-01 2.13800982e-01 -4.37478691e-01 4.07201588e-01
-8.08164835e-01 1.03222072e+00 1.66060343e-01 -1.15750335e-01
-5.19843161e-01 -7.02104747e-01 -8.76534283e-01 1.10284992e-01
-3.67661595e-01 -1.02426136e+00 7.97350824e-01 -1.45347655e+00
-9.81620073e-01 7.50608563e-01 -7.06666037e-02 -6.31244540e-01
6.29431009e-01 -1.89397097e-01 -7.26178885e-02 2.30656937e-01
-1.25110149e-01 1.10762942e+00 1.13980162e+00 -1.44329154e+00
-2.52911329e-01 -5.09050906e-01 3.92315211e-03 2.23972052e-01
-5.34877658e-01 -5.18036783e-01 2.01227173e-01 -6.16752565e-01
4.43437248e-01 -9.85259771e-01 -3.19880962e-01 2.00552508e-01
-7.24038303e-01 -2.25141197e-01 8.00103188e-01 -4.80003744e-01
6.01283073e-01 -2.06042457e+00 3.84857118e-01 2.58344293e-01
4.90152210e-01 -7.09010512e-02 -1.77550137e-01 1.30688157e-02
-5.75038731e-01 3.59118074e-01 -1.94056675e-01 -3.91076714e-01
8.36134255e-02 3.83765101e-01 -5.27612448e-01 7.74939597e-01
4.52270746e-01 8.46735060e-01 -7.55138636e-01 -3.26959819e-01
2.64449477e-01 6.22580409e-01 -8.03046048e-01 8.54122415e-02
-2.44389594e-01 7.52067387e-01 -3.84679854e-01 -6.77930145e-03
3.39577705e-01 -4.32965189e-01 -1.55049756e-01 -1.49726883e-01
9.95114669e-02 2.81353086e-01 -8.96785080e-01 1.47512901e+00
-5.10124713e-02 1.06537461e+00 -2.78706759e-01 -1.13977790e+00
7.39023864e-01 2.59999603e-01 3.59674066e-01 -2.45296136e-01
9.21319723e-02 2.20841005e-01 2.24167660e-01 -5.54265857e-01
3.30388248e-01 -2.50616640e-01 5.76157510e-01 4.35180545e-01
3.30681615e-02 -9.69980285e-02 -1.28544778e-01 1.71720713e-01
5.58395267e-01 2.53520846e-01 -2.29596734e-01 -5.11544526e-01
2.27320835e-01 2.72562951e-01 5.02168834e-01 5.33260286e-01
-1.89926460e-01 7.50078321e-01 5.70562422e-01 -7.40048409e-01
-1.37531745e+00 -1.09762478e+00 -3.94072026e-01 4.71055776e-01
9.79142562e-02 4.53337990e-02 -1.04252815e+00 -8.40354979e-01
-7.50912400e-03 1.07249868e+00 -1.01216364e+00 -3.54377002e-01
-3.87247384e-01 -5.31545877e-01 1.87573999e-01 5.99541128e-01
3.50864083e-01 -9.59453404e-01 -5.99321127e-01 9.50729325e-02
3.22862109e-03 -7.79305458e-01 -5.38451672e-01 6.73706979e-02
-1.24257410e+00 -1.20487666e+00 -6.54401004e-01 -6.51117444e-01
1.19132507e+00 1.43789165e-02 9.75382984e-01 4.55359459e-01
-3.10431309e-02 5.67153096e-01 -5.71199059e-02 -4.03103024e-01
-5.77430665e-01 -1.26268134e-01 2.20746621e-01 1.33904606e-01
2.35987514e-01 -6.68814719e-01 -7.11835802e-01 3.52700591e-01
-8.64476085e-01 1.06421001e-01 3.48986983e-01 8.85826945e-01
4.99686033e-01 -1.75144985e-01 2.51269370e-01 -7.51154006e-01
6.02295578e-01 -3.44644636e-01 -3.47638071e-01 3.12118847e-02
-9.12582636e-01 5.08001566e-01 5.97045183e-01 -5.32941461e-01
-6.29529357e-01 2.89606322e-02 1.30989926e-03 -8.08450520e-01
-2.00376183e-01 4.31102455e-01 2.03702196e-01 1.12018086e-01
1.20169044e+00 1.03497751e-01 3.30898702e-01 -3.10330421e-01
5.67016363e-01 1.10623769e-01 5.71865499e-01 -2.09565386e-01
1.11334038e+00 7.33809590e-01 6.59820363e-02 -7.14025676e-01
-7.75812805e-01 7.30655044e-02 -6.32646441e-01 -9.91853625e-02
1.28401411e+00 -3.80933970e-01 -3.37551326e-01 1.15339108e-01
-1.21654475e+00 -3.88942122e-01 -4.51031625e-01 7.93610871e-01
-7.25454926e-01 2.05036759e-01 -4.93656635e-01 -4.00250345e-01
-1.25711650e-01 -1.31522048e+00 6.13812983e-01 2.60777861e-01
-6.70693755e-01 -1.40369844e+00 -9.41645876e-02 1.28809839e-01
4.00087595e-01 3.32939863e-01 1.15289438e+00 -7.69549787e-01
-5.44949591e-01 -1.86893463e-01 -8.24115798e-02 4.77215677e-01
2.69181430e-01 2.59560674e-01 -8.30448091e-01 -9.61999819e-02
2.15010330e-01 -1.36794999e-01 8.36801946e-01 6.87203288e-01
1.13814628e+00 -5.08135617e-01 -1.14621803e-01 8.05356443e-01
1.24435520e+00 -2.91937143e-01 4.73718762e-01 4.04403746e-01
8.59544516e-01 1.01222825e+00 1.26476094e-01 -1.35107696e-01
3.10247362e-01 4.39995229e-01 7.28827596e-01 -3.76707971e-01
-1.25364870e-01 -3.69961947e-01 2.21046671e-01 5.03643215e-01
5.20274565e-02 1.19623736e-01 -6.83722854e-01 5.02289176e-01
-1.81924105e+00 -9.45517898e-01 -5.46421289e-01 2.06028867e+00
4.20381933e-01 2.47711301e-01 9.16526541e-02 5.68866320e-02
5.44629097e-01 -2.46714540e-02 -8.30083787e-01 -5.54710627e-01
-8.93779472e-02 -2.35270813e-01 3.14720720e-01 8.18724453e-01
-8.35686445e-01 6.57604635e-01 6.77521801e+00 1.20988689e-01
-1.37701213e+00 -2.12813035e-01 7.83800125e-01 1.11832038e-01
-6.33780479e-01 2.34053545e-02 -2.32151821e-01 2.50287056e-01
7.53741443e-01 -9.16972309e-02 4.01118040e-01 9.34707284e-01
1.48379460e-01 3.53081435e-01 -1.51760006e+00 6.38473213e-01
-1.00650899e-01 -1.62867534e+00 1.52831405e-01 3.03952307e-01
7.17977464e-01 6.07306175e-02 3.99971694e-01 -5.49022406e-02
2.51065075e-01 -1.25424147e+00 7.10486233e-01 5.33917129e-01
3.17232937e-01 -4.80419576e-01 5.23283660e-01 2.65028298e-01
-5.52091479e-01 1.85874119e-01 -4.83299762e-01 7.70007223e-02
-1.07057199e-01 3.35979432e-01 -1.01603127e+00 1.03010289e-01
5.17946482e-01 7.29258001e-01 -5.75867653e-01 7.00234175e-01
-5.92738152e-01 5.14755070e-01 -1.49443537e-01 3.84476334e-02
2.56073862e-01 -3.69469613e-01 7.78213978e-01 7.83999562e-01
1.28478795e-01 -1.71909243e-01 -2.08860219e-01 1.33751082e+00
2.72412151e-02 5.71484603e-02 -7.79040098e-01 -8.00192058e-02
1.19965456e-01 1.03374279e+00 -4.70923692e-01 -2.99856961e-01
-2.69017607e-01 9.47481036e-01 3.83684486e-01 4.99837935e-01
-7.98520148e-01 -5.73500842e-02 1.02575278e+00 1.01654433e-01
7.42001981e-02 -1.79772153e-01 -7.87595749e-01 -1.23569357e+00
-4.77409214e-02 -7.74326921e-01 1.41627327e-01 -1.07384253e+00
-1.23609996e+00 6.53757989e-01 -2.76741505e-01 -1.08389521e+00
-3.42388749e-01 -6.59909487e-01 -9.80586410e-01 8.25856745e-01
-1.23471093e+00 -7.19670534e-01 -1.93612382e-01 6.59864962e-01
3.87169570e-01 -9.74367633e-02 7.85871923e-01 -1.11062229e-01
-4.67671871e-01 5.26322424e-01 -9.68097448e-02 2.68080503e-01
4.44917977e-01 -1.65445352e+00 3.45375121e-01 7.54384518e-01
7.48417795e-01 8.31893682e-01 1.35112143e+00 -3.73555362e-01
-1.20447385e+00 -1.05610156e+00 6.58004284e-01 -8.87151897e-01
6.40609741e-01 1.67160600e-01 -1.25567722e+00 9.52029586e-01
2.85881072e-01 -1.17190167e-01 4.38760310e-01 2.28498116e-01
-2.86702693e-01 1.25808105e-01 -1.09621918e+00 8.96827638e-01
8.55870128e-01 -5.07519364e-01 -8.09004962e-01 4.72319484e-01
6.27720237e-01 -3.00599694e-01 -7.54565239e-01 2.76878983e-01
2.11992934e-01 -8.75763297e-01 9.00191605e-01 -1.18326485e+00
8.39202523e-01 -1.89092979e-01 -1.21574245e-01 -1.62958241e+00
-3.19313169e-01 -3.18942040e-01 1.73760608e-01 8.89346838e-01
7.05787838e-01 -6.80526078e-01 1.04554486e+00 1.07569063e+00
-2.33313814e-01 -8.52854073e-01 -8.83339107e-01 -4.98479873e-01
3.55344951e-01 -4.14042085e-01 4.97936279e-01 1.26648235e+00
-3.21715921e-02 4.20784235e-01 -2.13515744e-01 2.12015107e-01
7.36381590e-01 -1.26361370e-01 8.63765419e-01 -1.12550318e+00
-1.31223604e-01 -6.85616672e-01 -5.84537327e-01 -8.53015482e-01
2.66627967e-01 -1.01297641e+00 -2.13661969e-01 -1.49668396e+00
-1.74621060e-01 -3.02765459e-01 -1.66479960e-01 5.01541436e-01
-2.34970361e-01 2.90733948e-02 -2.03110594e-02 4.60757494e-01
5.67333261e-03 6.12530649e-01 1.38898492e+00 4.12920266e-02
-2.66900092e-01 -1.92564949e-01 -8.32413733e-01 1.02502620e+00
9.32358563e-01 -4.66441363e-01 -4.05538201e-01 -6.03046536e-01
-2.31720656e-02 -8.08172077e-02 6.45880938e-01 -7.14756846e-01
-6.04743026e-02 -1.48905367e-01 6.25995457e-01 2.27497108e-02
3.35388362e-01 -7.47214377e-01 3.92775126e-02 6.38044834e-01
-5.93125880e-01 1.57673106e-01 -2.19859858e-03 4.82838094e-01
2.26602405e-02 -2.45131940e-01 8.52185845e-01 1.18923364e-02
-4.19853598e-01 2.49516189e-01 -2.49313459e-01 -1.51282728e-01
6.78514123e-01 -2.87047029e-01 -3.26678455e-01 -5.53732991e-01
-8.97647560e-01 1.78354189e-01 7.42379546e-01 4.71397579e-01
5.79120517e-01 -1.42126667e+00 -7.20074236e-01 2.75821656e-01
1.07024461e-02 -2.30189964e-01 -1.33419499e-01 8.67989838e-01
-3.99906367e-01 2.85122007e-01 -2.79138774e-01 -7.59853244e-01
-9.51075673e-01 5.74347138e-01 8.50276113e-01 2.36713752e-01
-7.82626212e-01 8.63389552e-01 3.07961255e-01 -4.74937081e-01
8.63973573e-02 -2.89196640e-01 -1.19703934e-01 -2.16379106e-01
2.79906273e-01 9.70619917e-02 -3.40325296e-01 -6.52365386e-01
-1.26299322e-01 4.08993483e-01 1.27995819e-01 -1.64653808e-01
1.46122169e+00 -3.57581377e-02 6.05393648e-02 2.59890825e-01
1.17486513e+00 -1.94611162e-01 -1.70557332e+00 -5.80771491e-02
7.51266023e-03 -3.68695915e-01 7.27017084e-03 -4.02105272e-01
-1.26185238e+00 9.32210028e-01 6.84884846e-01 5.97907603e-01
6.63755476e-01 5.13008162e-02 2.76024312e-01 2.88331747e-01
-1.29576236e-01 -9.16664243e-01 2.69534945e-01 -2.08057705e-02
1.17461789e+00 -1.37481451e+00 -2.09834173e-01 -1.01843178e-01
-7.30995834e-01 1.10586345e+00 6.71681464e-01 -5.62006056e-01
5.91089070e-01 -5.85645199e-01 3.94744992e-01 -4.69950140e-01
-5.16356170e-01 9.50738937e-02 6.73579037e-01 4.65276033e-01
3.91053766e-01 -9.28657278e-02 3.74861108e-03 1.87123731e-01
-5.40573597e-01 -3.66647214e-01 5.95037103e-01 2.92103678e-01
-4.43882763e-01 -9.41359758e-01 -5.13312876e-01 3.95559222e-01
-4.06748950e-01 1.54880375e-01 -4.52508599e-01 9.22250926e-01
-1.01325475e-01 7.71002233e-01 2.42161363e-01 -2.26489201e-01
2.42394820e-01 1.83023110e-01 3.50420743e-01 -3.59961450e-01
-1.65629581e-01 -9.14250538e-02 -1.76222369e-01 -5.73989749e-01
-4.13905978e-01 -6.97742403e-01 -1.51383185e+00 -2.77476817e-01
-1.65597692e-01 1.83937445e-01 8.30564559e-01 1.06635559e+00
3.68135631e-01 4.18906271e-01 4.89009082e-01 -7.82024682e-01
-7.03327537e-01 -8.68096352e-01 -3.61076862e-01 9.06264186e-01
5.02515137e-01 -5.85496068e-01 -7.39638209e-01 2.03016102e-01] | [8.740891456604004, 4.695564270019531] |
ae81898b-35af-47bc-b43a-0a281ecb8589 | comparative-analysis-of-segment-anything | 2306.12510 | null | https://arxiv.org/abs/2306.12510v1 | https://arxiv.org/pdf/2306.12510v1.pdf | Comparative Analysis of Segment Anything Model and U-Net for Breast Tumor Detection in Ultrasound and Mammography Images | In this study, the main objective is to develop an algorithm capable of identifying and delineating tumor regions in breast ultrasound (BUS) and mammographic images. The technique employs two advanced deep learning architectures, namely U-Net and pretrained SAM, for tumor segmentation. The U-Net model is specifically designed for medical image segmentation and leverages its deep convolutional neural network framework to extract meaningful features from input images. On the other hand, the pretrained SAM architecture incorporates a mechanism to capture spatial dependencies and generate segmentation results. Evaluation is conducted on a diverse dataset containing annotated tumor regions in BUS and mammographic images, covering both benign and malignant tumors. This dataset enables a comprehensive assessment of the algorithm's performance across different tumor types. Results demonstrate that the U-Net model outperforms the pretrained SAM architecture in accurately identifying and segmenting tumor regions in both BUS and mammographic images. The U-Net exhibits superior performance in challenging cases involving irregular shapes, indistinct boundaries, and high tumor heterogeneity. In contrast, the pretrained SAM architecture exhibits limitations in accurately identifying tumor areas, particularly for malignant tumors and objects with weak boundaries or complex shapes. These findings highlight the importance of selecting appropriate deep learning architectures tailored for medical image segmentation. The U-Net model showcases its potential as a robust and accurate tool for tumor detection, while the pretrained SAM architecture suggests the need for further improvements to enhance segmentation performance. | ['Abbas Sharifi', 'Ahmad Gholizadeh Lonbar', 'Elyas Irankhah', 'Kasra Danesh', 'Sara Asgarian', 'Masoumeh Farhadi Nia', 'Mohsen Ahmadi'] | 2023-06-21 | null | null | null | null | ['tumor-segmentation', 'medical-image-segmentation'] | ['computer-vision', 'medical'] | [ 5.46584189e-01 4.03737038e-01 -4.25143570e-01 -3.54922384e-01
-1.02888477e+00 -3.46588671e-01 3.37994039e-01 2.85127342e-01
-3.35999787e-01 2.75522947e-01 7.28238514e-03 -8.06616187e-01
-7.83089176e-02 -7.18519449e-01 -3.50589067e-01 -9.49843228e-01
-3.10745001e-01 7.35950708e-01 1.78572476e-01 -4.62385006e-02
-1.48210108e-01 8.54340553e-01 -7.42879689e-01 4.22954381e-01
7.68852413e-01 1.22490203e+00 2.84495056e-01 8.79279017e-01
-2.54505843e-01 6.63242340e-01 -4.60906833e-01 4.82106917e-02
1.11549921e-01 -4.27154005e-01 -9.87172902e-01 9.86104235e-02
5.29990017e-01 -3.99660587e-01 -1.92919776e-01 7.34693229e-01
5.05761981e-01 -4.15330797e-01 7.98066854e-01 -6.84268415e-01
-2.39027396e-01 5.53382456e-01 -5.25175929e-01 7.30575919e-01
-2.93800443e-01 2.33543754e-01 6.87631130e-01 -4.58106548e-01
6.06400847e-01 5.40833294e-01 1.18946075e+00 4.59594846e-01
-1.12904191e+00 -4.47565496e-01 -2.43718043e-01 -1.90682486e-01
-1.19281352e+00 -1.95614278e-01 3.90590012e-01 -5.25958717e-01
7.03030527e-01 5.61304152e-01 7.36659706e-01 8.46961975e-01
6.79693997e-01 1.11121631e+00 9.23655272e-01 -3.25252801e-01
1.23940483e-01 5.32131083e-03 3.68016362e-01 8.33994567e-01
3.64549845e-01 6.67645782e-02 6.97726980e-02 -9.50168818e-02
9.84981120e-01 -4.77264822e-02 -1.32631600e-01 -2.90426642e-01
-1.03122127e+00 6.86180651e-01 6.97770774e-01 8.92032504e-01
-5.11183858e-01 1.48166046e-01 7.31201053e-01 -1.05623774e-01
5.27472794e-01 4.65514392e-01 -1.62520900e-01 1.17880456e-01
-1.26306522e+00 -9.47599411e-02 5.64900219e-01 5.45917511e-01
1.05852872e-01 -4.11843210e-02 -4.46937054e-01 7.65996039e-01
5.19456714e-02 7.01288581e-02 7.76047945e-01 -4.25337791e-01
1.58526488e-02 6.55566216e-01 -2.65788525e-01 -8.07522237e-01
-1.02441120e+00 -9.26986337e-01 -9.62260544e-01 1.01153791e-01
6.32688820e-01 -5.08737415e-02 -1.66060662e+00 1.15198421e+00
1.23079389e-01 7.20127225e-02 1.80097055e-02 7.55266666e-01
1.32427931e+00 1.99996412e-01 2.76004970e-01 1.48593053e-01
1.36745381e+00 -8.13860238e-01 -5.54326832e-01 -2.58301675e-01
1.07348716e+00 -4.64068085e-01 7.56725490e-01 -6.66292310e-02
-1.19268668e+00 -2.85879135e-01 -9.26529229e-01 3.82323749e-02
-4.25952107e-01 4.72580343e-01 8.49521577e-01 8.54451239e-01
-1.35656846e+00 1.87356710e-01 -1.24529934e+00 -4.95412320e-01
9.62473392e-01 6.89917862e-01 -3.06176335e-01 -3.83394137e-02
-8.90654862e-01 9.89292979e-01 5.75517416e-01 2.82289743e-01
-1.09493220e+00 -8.97176325e-01 -9.02022243e-01 9.00636241e-03
1.19209260e-01 -5.25351703e-01 1.43073845e+00 -1.24432826e+00
-1.11889267e+00 1.12740326e+00 2.57471926e-03 -7.99025536e-01
7.03676760e-01 5.05525172e-01 -2.46480510e-01 5.18831015e-01
1.75333083e-01 8.08085203e-01 4.33268756e-01 -1.18646491e+00
-7.29582191e-01 -3.16295028e-01 -3.29895437e-01 3.28490995e-02
-7.05210716e-02 -3.67958307e-01 -4.06510949e-01 -7.24536896e-01
3.86166096e-01 -8.22455823e-01 -5.25814772e-01 -4.16294336e-02
-3.74485284e-01 1.00438319e-01 1.21167099e+00 -7.21520424e-01
1.04170406e+00 -1.94826198e+00 -2.55320191e-01 5.89772165e-01
4.60367143e-01 4.45539236e-01 -1.56948000e-01 -1.88980684e-01
-5.35209328e-02 2.53684193e-01 -3.69825304e-01 7.67034665e-02
-3.93218845e-01 3.04218143e-01 4.28404123e-01 4.99368995e-01
3.29818785e-01 1.51805091e+00 -7.10516214e-01 -6.56632304e-01
2.26783887e-01 3.62017840e-01 -3.16547006e-01 6.60051852e-02
-7.71796033e-02 6.34023070e-01 -4.34055209e-01 1.00351346e+00
5.89234650e-01 -4.57367420e-01 2.22499922e-01 -3.48747015e-01
1.57584071e-01 -2.02563658e-01 -4.97365177e-01 1.40466976e+00
-4.86037344e-01 9.47169423e-01 5.65234303e-01 -1.08994269e+00
7.23326743e-01 3.87415141e-01 1.06196022e+00 -6.96337283e-01
6.10173106e-01 4.07811612e-01 5.38215101e-01 -7.94919372e-01
2.72383213e-01 -2.55916536e-01 2.43693814e-01 3.38771015e-01
-1.66430380e-02 -1.13985255e-01 1.77363709e-01 -4.32251021e-02
1.22771764e+00 -3.02121848e-01 1.92435965e-01 -5.37070096e-01
4.27162200e-01 3.95314604e-01 2.48664960e-01 8.30490947e-01
-4.64678228e-01 7.15577006e-01 6.16621435e-01 -6.95685148e-01
-5.90159357e-01 -8.92788827e-01 -6.82226002e-01 8.68006527e-01
3.46453786e-02 1.27483234e-01 -6.67474389e-01 -9.32297587e-01
-8.97007063e-02 4.07241315e-01 -1.24701464e+00 1.98196173e-02
-7.49217093e-01 -1.07001925e+00 7.24846721e-01 8.24715793e-01
6.41134501e-01 -1.00767815e+00 -9.38459754e-01 3.43076795e-01
-1.71879098e-01 -8.62295806e-01 -2.39564016e-01 3.80229622e-01
-8.81786287e-01 -1.34406698e+00 -1.14083123e+00 -1.05417740e+00
1.16696239e+00 -1.11908235e-01 1.08611262e+00 1.81250364e-01
-8.34591627e-01 3.88076127e-01 -2.15836242e-01 -5.60884774e-01
-8.41078401e-01 3.01053822e-01 -7.99891114e-01 -2.31869057e-01
4.04215157e-01 -2.77973004e-02 -7.05966175e-01 3.06785047e-01
-1.10571027e+00 1.71502963e-01 1.09777999e+00 1.19315064e+00
6.54186606e-01 5.05058579e-02 3.53837013e-01 -1.06293869e+00
4.47806090e-01 -6.11430228e-01 -1.26383811e-01 5.13680317e-02
-1.25340387e-01 -5.03238440e-01 2.78902739e-01 -1.85698822e-01
-9.25662637e-01 6.89297765e-02 -4.20396179e-01 2.32327096e-02
-3.94872725e-01 8.01920593e-01 4.68917996e-01 -4.77881014e-01
7.78152525e-01 8.47988501e-02 4.53944296e-01 2.13887431e-02
-2.88693696e-01 5.02432585e-01 5.75746834e-01 -2.40444839e-01
2.49429762e-01 6.99640512e-01 2.78889835e-01 -8.48730862e-01
-5.87539315e-01 -5.04886031e-01 -8.02944541e-01 -1.70475826e-01
9.15796697e-01 -3.10881674e-01 -1.25158906e-01 4.66581315e-01
-6.27489924e-01 -7.61646509e-01 -4.32707012e-01 3.11297029e-01
-3.94947320e-01 -5.64331235e-03 -9.83868718e-01 -2.71000355e-01
-5.27175605e-01 -1.49068987e+00 1.20624363e+00 3.33665878e-01
-3.66767794e-01 -1.43027389e+00 -3.31264377e-01 3.39356363e-01
7.66026139e-01 6.59680903e-01 1.01327980e+00 -9.96786714e-01
-3.48040044e-01 -5.79857051e-01 -3.37692350e-01 4.08904478e-02
4.16198999e-01 5.09990416e-02 -7.57069945e-01 -2.79873371e-01
-3.01686645e-01 -1.95931822e-01 1.02382243e+00 1.13001800e+00
1.39989674e+00 1.12087898e-01 -7.02562988e-01 8.53253901e-01
1.10777080e+00 5.25137961e-01 3.53248507e-01 4.42508757e-01
5.50807774e-01 4.09987122e-01 4.05785024e-01 -1.96870536e-01
9.96697471e-02 1.66421279e-01 6.10149503e-01 -9.98790503e-01
-3.44442785e-01 3.36664319e-01 -3.68185133e-01 8.80634859e-02
5.67733161e-02 -9.61238891e-03 -1.60341036e+00 6.79882586e-01
-1.39982927e+00 -4.92609292e-01 -7.65234455e-02 1.67953765e+00
6.37697697e-01 1.72500044e-01 1.11605041e-03 -8.93298611e-02
3.80093157e-01 -9.23212245e-03 -5.38557053e-01 -3.94226938e-01
2.64131930e-02 4.00579840e-01 6.39638603e-01 1.76021695e-01
-1.34493589e+00 6.65616512e-01 7.34333229e+00 7.64575124e-01
-1.61271775e+00 7.93408379e-02 1.41745591e+00 1.91543952e-01
-6.22546673e-02 -6.24839783e-01 -4.42760766e-01 8.98379385e-02
8.23718071e-01 3.25298041e-01 -6.39818490e-01 4.70176131e-01
1.72868490e-01 -3.73577118e-01 -9.75901902e-01 5.28233171e-01
-3.37716080e-02 -1.86072898e+00 -9.21684876e-02 1.25885606e-01
7.46398628e-01 9.84762311e-02 2.59941250e-01 2.45701924e-01
1.33194998e-01 -1.48845196e+00 2.64550537e-01 2.89130807e-01
9.81835663e-01 -5.05337238e-01 1.41009903e+00 4.98104393e-02
-9.45964158e-01 4.64993007e-02 -1.87633827e-03 4.79509771e-01
-2.41784677e-01 2.20584989e-01 -1.52664435e+00 3.91568124e-01
5.97085178e-01 5.88234484e-01 -7.82615364e-01 1.15823436e+00
5.08702457e-01 8.12681675e-01 -2.18924627e-01 1.68618411e-01
7.80321002e-01 7.47700036e-02 3.11649859e-01 1.69925153e+00
2.01149821e-01 2.22597748e-01 2.49270275e-01 7.22386777e-01
2.53533632e-01 1.12936951e-01 -3.23031932e-01 -1.05905093e-01
9.13641602e-02 1.50536978e+00 -1.45101810e+00 -4.43202466e-01
-2.35636279e-01 4.30568784e-01 -1.30140185e-01 3.72621089e-01
-7.15017378e-01 -1.24238050e-02 7.69366920e-02 2.64767051e-01
1.00842871e-01 1.63286075e-01 -8.44059169e-01 -3.83109748e-01
-5.43301821e-01 -7.29429543e-01 6.31229877e-01 -3.25046808e-01
-8.37740600e-01 7.40761459e-01 -6.04475588e-02 -9.92334366e-01
-1.55220553e-01 -7.76979506e-01 -9.23292816e-01 5.65816522e-01
-1.47042000e+00 -1.54700303e+00 -6.55597568e-01 1.85858145e-01
5.79929948e-01 -2.49019906e-01 7.85391033e-01 -3.29532623e-02
-7.10948527e-01 6.67622089e-01 6.93087950e-02 4.13126916e-01
3.07695419e-01 -1.40145385e+00 -4.74477559e-03 4.56285238e-01
-4.42189634e-01 2.46760443e-01 3.71535391e-01 -5.39683282e-01
-1.01516569e+00 -1.40876210e+00 1.22278519e-01 -1.45550922e-01
4.79977459e-01 6.97083250e-02 -7.25183129e-01 7.88057983e-01
1.88411430e-01 4.28588182e-01 9.50795293e-01 -3.33981901e-01
3.26264083e-01 5.51871881e-02 -1.49831259e+00 4.01972771e-01
4.15837169e-01 2.68081147e-02 -1.51196256e-01 4.92550522e-01
5.96388765e-02 -1.01851928e+00 -1.23493123e+00 8.02627265e-01
4.69310850e-01 -9.07635272e-01 8.72701705e-01 -3.02218378e-01
5.64915121e-01 2.06453472e-01 3.77861798e-01 -1.15446854e+00
-5.06038249e-01 -1.07834935e-01 2.03622729e-01 5.93676150e-01
6.92748845e-01 -6.09695792e-01 1.37173128e+00 6.09409630e-01
-5.58683991e-01 -1.28989589e+00 -8.43613088e-01 -2.32751608e-01
4.15966064e-01 -1.97512627e-01 3.56839359e-01 9.21262801e-01
-4.94163692e-01 -3.85431170e-01 6.19688094e-01 1.08438665e-02
2.98193753e-01 -3.60387936e-02 3.24098527e-01 -8.58578205e-01
1.29554793e-01 -1.08243477e+00 -5.50853312e-01 -5.84564328e-01
3.54683027e-02 -1.01794410e+00 -2.95985155e-02 -1.76283777e+00
1.72028318e-01 -9.32799220e-01 -3.45297784e-01 6.21038556e-01
-1.25595018e-01 5.98849595e-01 -2.26462319e-01 1.89787611e-01
-1.42494500e-01 -1.38945192e-01 1.61360669e+00 -5.72213352e-01
-2.27496877e-01 3.23704958e-01 -5.69340825e-01 7.52095580e-01
9.05241311e-01 5.72063914e-03 -1.60725564e-01 -2.72149146e-01
-6.14588857e-01 6.15232810e-02 4.40378219e-01 -1.00016284e+00
1.81667879e-01 -3.66236642e-02 8.93867970e-01 -7.83863068e-01
5.38008921e-02 -8.83572161e-01 -1.21248685e-01 8.86004925e-01
-4.53292578e-01 -3.43326747e-01 7.02419400e-01 1.71443522e-01
-4.00117785e-01 -2.18111321e-01 9.95090425e-01 -4.26436245e-01
-7.55519748e-01 4.52738315e-01 -6.48965180e-01 -2.66959190e-01
1.45422947e+00 -8.30330133e-01 -9.83445272e-02 -9.68223363e-02
-1.06230736e+00 3.12974572e-01 7.95058310e-02 8.87615234e-02
5.35114408e-01 -1.00498998e+00 -8.74272764e-01 3.67002636e-01
2.02895235e-02 4.31847095e-01 3.61674964e-01 1.44546902e+00
-1.15197289e+00 6.58814251e-01 -2.64723152e-01 -1.16617751e+00
-1.33728313e+00 -1.47161528e-01 1.05535781e+00 -5.30869544e-01
-6.76782429e-01 1.15297592e+00 3.92311245e-01 -3.48428965e-01
3.10424447e-01 -7.69237936e-01 -3.71568173e-01 -9.05053094e-02
3.18723500e-01 1.93995938e-01 3.26621771e-01 -7.66721606e-01
-1.46551341e-01 9.83109996e-02 -5.13925910e-01 4.59309399e-01
1.16861951e+00 2.20454082e-01 -1.28127649e-01 3.08320746e-02
1.17341137e+00 -4.05713558e-01 -1.07131326e+00 -1.46735117e-01
1.39859766e-01 -1.24538340e-01 5.23205757e-01 -1.01215553e+00
-1.47285748e+00 6.13199592e-01 8.77520800e-01 2.39813730e-01
1.22774994e+00 6.00203723e-02 7.05860317e-01 -1.06761657e-01
-1.01306319e-01 -7.09950864e-01 -1.43717099e-02 2.79773086e-01
6.92029536e-01 -1.43292665e+00 -2.01924995e-01 -5.15190840e-01
-5.65785646e-01 1.44882596e+00 8.97026062e-01 -4.55133207e-02
5.74178696e-01 6.70249581e-01 5.60035110e-01 -6.00725234e-01
-2.62498915e-01 -1.73017919e-01 4.81606513e-01 5.64146757e-01
7.21067667e-01 2.00003088e-01 -8.70805457e-02 3.95919412e-01
-1.51962444e-01 -7.31993467e-02 4.10957247e-01 1.14611351e+00
-4.10429984e-01 -7.16868162e-01 -3.98656785e-01 1.26262379e+00
-7.16449440e-01 1.48152307e-01 -5.51373065e-01 1.23903608e+00
1.79013535e-01 4.67409283e-01 3.63800973e-01 7.59134218e-02
9.88698006e-02 -2.37571582e-01 3.18612456e-01 -6.15386367e-01
-1.05099189e+00 2.96203732e-01 5.22653162e-02 -3.64419311e-01
-2.44799480e-01 -5.38707793e-01 -1.30983436e+00 1.18946277e-01
-3.17779541e-01 2.84892377e-02 5.50046504e-01 8.64194512e-01
3.11448257e-02 1.02841985e+00 1.79836363e-01 -8.48849475e-01
-2.50634998e-01 -1.03038061e+00 -4.67898756e-01 2.49553725e-01
5.37102878e-01 -3.55451882e-01 2.13110209e-01 -1.96433902e-01] | [14.900250434875488, -2.5158543586730957] |
fcd9a9ed-57fa-4116-b3a4-fbc144194766 | codetrans-towards-cracking-the-language-of | 2104.02443 | null | https://arxiv.org/abs/2104.02443v2 | https://arxiv.org/pdf/2104.02443v2.pdf | CodeTrans: Towards Cracking the Language of Silicon's Code Through Self-Supervised Deep Learning and High Performance Computing | Currently, a growing number of mature natural language processing applications make people's life more convenient. Such applications are built by source code - the language in software engineering. However, the applications for understanding source code language to ease the software engineering process are under-researched. Simultaneously, the transformer model, especially its combination with transfer learning, has been proven to be a powerful technique for natural language processing tasks. These breakthroughs point out a promising direction for process source code and crack software engineering tasks. This paper describes CodeTrans - an encoder-decoder transformer model for tasks in the software engineering domain, that explores the effectiveness of encoder-decoder transformer models for six software engineering tasks, including thirteen sub-tasks. Moreover, we have investigated the effect of different training strategies, including single-task learning, transfer learning, multi-task learning, and multi-task learning with fine-tuning. CodeTrans outperforms the state-of-the-art models on all the tasks. To expedite future works in the software engineering domain, we have published our pre-trained models of CodeTrans. https://github.com/agemagician/CodeTrans | ['Burkhard Rost', 'Florian Matthes', 'Silvia Severini', 'Christoph Angerer', 'Tamas Feher', 'Tom Gibbs', 'Llion Jones', 'Wei Ding', 'Ahmed Elnaggar'] | 2021-04-06 | null | null | null | null | ['api-sequence-recommendation', 'code-summarization', 'contextual-embedding-for-source-code', 'git-commit-message-generation', 'code-documentation-generation', 'code-comment-generation', 'code-documentation-generation'] | ['computer-code', 'computer-code', 'computer-code', 'computer-code', 'computer-code', 'computer-code', 'natural-language-processing'] | [-2.15951893e-02 -1.66527912e-01 -3.15331757e-01 -4.22725201e-01
-1.08083403e+00 -3.23900461e-01 2.30640516e-01 3.03869769e-02
-1.47145446e-02 2.31852934e-01 1.20349094e-01 -6.83744907e-01
8.07898790e-02 -5.52538216e-01 -6.45251274e-01 -2.20699027e-01
8.98902193e-02 6.63949549e-02 3.27864110e-01 -2.44990081e-01
7.37851322e-01 -3.46558660e-01 -1.37569749e+00 7.41345108e-01
1.19046497e+00 5.26978135e-01 5.41913331e-01 5.67987144e-01
-5.64403415e-01 1.09522367e+00 -4.24261957e-01 -6.65986180e-01
-2.35514089e-01 -2.25904346e-01 -8.26421857e-01 -2.40974560e-01
4.17595170e-02 4.52684835e-02 2.40715548e-01 1.00375640e+00
5.36305070e-01 -4.97709364e-01 3.66087675e-01 -1.33287907e+00
-1.16631305e+00 8.19360614e-01 -7.40313530e-01 -3.66051532e-02
4.99632895e-01 -3.92276831e-02 8.68266463e-01 -1.31851530e+00
2.76922017e-01 1.01972508e+00 9.30377603e-01 4.30459827e-01
-8.20461929e-01 -8.74747574e-01 -2.53880680e-01 3.93001199e-01
-1.15811062e+00 -3.61867011e-01 8.97684097e-01 -9.99464273e-01
1.38318706e+00 -3.29120636e-01 2.34137595e-01 9.69516456e-01
8.93915832e-01 8.28681707e-01 8.63112152e-01 -7.59974003e-01
-1.56547278e-01 1.46897078e-01 1.88121304e-01 9.68307436e-01
4.34204414e-02 -3.17238458e-02 -6.40568733e-01 -1.28462762e-01
3.91577363e-01 8.92335474e-02 -6.36051670e-02 -3.52231443e-01
-1.20756364e+00 8.06586266e-01 1.08035721e-01 6.20421290e-01
3.44521971e-03 2.41257593e-01 7.52065599e-01 6.78311586e-01
7.36817777e-01 3.08287472e-01 -8.27978611e-01 -5.68339825e-01
-8.44063401e-01 -3.31006497e-02 8.42011154e-01 1.39201760e+00
8.64601016e-01 2.42222458e-01 3.36521827e-02 1.10098052e+00
4.73767787e-01 2.27814049e-01 6.79719090e-01 -5.18847466e-01
9.08815503e-01 9.56708610e-01 -3.32858711e-01 -6.33668303e-01
-3.13152485e-02 -2.20130891e-01 -6.97855115e-01 1.63439035e-01
6.13862872e-02 -3.95300537e-02 -5.93650758e-01 1.29908288e+00
-1.40335664e-01 -1.02541022e-01 6.29947754e-04 2.63896048e-01
7.22357750e-01 7.90541947e-01 6.72346577e-02 5.64114153e-02
1.45968843e+00 -1.23459721e+00 -6.77879035e-01 -4.41000909e-01
9.19252872e-01 -1.20782745e+00 1.39490974e+00 6.42421246e-01
-8.32588434e-01 -8.13631237e-01 -9.60970998e-01 -3.52636278e-01
-5.60647130e-01 5.37708700e-01 5.46449244e-01 6.60191655e-01
-9.85193908e-01 4.09446299e-01 -7.87400961e-01 -3.42835307e-01
3.09432715e-01 -7.23005459e-02 -2.79027820e-01 -1.38491243e-01
-1.00711226e+00 9.17544782e-01 4.11074907e-01 -1.86381087e-01
-1.15351021e+00 -9.00409400e-01 -9.10791695e-01 1.06009930e-01
3.63908470e-01 -5.16218066e-01 1.73958123e+00 -8.37406337e-01
-1.37252581e+00 9.51685071e-01 -1.30991861e-01 2.83410394e-04
1.18456148e-02 -5.10276258e-01 -4.12952393e-01 -5.02551675e-01
2.83856183e-01 1.11555988e-02 9.33988631e-01 -9.70690548e-01
-5.70172846e-01 -2.34584644e-01 -4.33902331e-02 -2.40539193e-01
-5.22473335e-01 4.97161031e-01 -3.34259421e-01 -6.59693897e-01
-4.07583892e-01 -6.81174695e-01 1.36776656e-01 3.41242962e-02
-8.63366872e-02 -7.12817490e-01 7.10669875e-01 -9.22324419e-01
1.56028390e+00 -2.44021535e+00 1.11366518e-01 -5.31487525e-01
2.25087479e-01 2.54348695e-01 -1.88824415e-01 7.37545073e-01
-2.71949828e-01 3.55960250e-01 -3.02756935e-01 -3.70631039e-01
1.06366128e-01 -1.70282900e-01 -2.39580914e-01 -1.43507458e-02
2.68135846e-01 8.66912425e-01 -9.43842411e-01 -4.99886960e-01
-7.67647922e-02 3.09019059e-01 -5.06568432e-01 4.01724637e-01
-2.69835472e-01 1.80589661e-01 -7.39588916e-01 7.43870556e-01
4.24105912e-01 -3.51269603e-01 -1.45618439e-01 9.89470333e-02
-3.90095353e-01 4.65216547e-01 -5.63409030e-01 2.23133898e+00
-1.04852808e+00 6.23196483e-01 -1.06435716e-01 -9.24162269e-01
1.09687030e+00 6.39675796e-01 2.03944415e-01 -5.48924983e-01
-1.18477428e-02 4.76265401e-01 3.95916775e-02 -1.14325786e+00
1.43232495e-01 -2.55075425e-01 -3.15548509e-01 6.50384784e-01
1.37481391e-01 -4.11090732e-01 2.59519547e-01 4.93314909e-03
1.08777666e+00 5.34841776e-01 4.07443643e-01 -2.30524734e-01
5.55716515e-01 -1.35104835e-01 2.82245725e-01 7.48832598e-02
6.69297297e-03 4.38927054e-01 5.49527168e-01 -4.33281869e-01
-9.71864283e-01 -7.97235191e-01 9.09171924e-02 1.60141563e+00
-3.03560197e-01 -7.32063115e-01 -7.21562624e-01 -6.93637788e-01
-1.85384929e-01 9.09508228e-01 -4.17604953e-01 -3.76403421e-01
-3.32336813e-01 -4.11463946e-01 6.76837027e-01 6.36458755e-01
6.12948596e-01 -1.28316987e+00 -4.39590394e-01 1.84882075e-01
-3.92097265e-01 -8.81475568e-01 -7.08551347e-01 1.11814760e-01
-8.78743112e-01 -1.18603134e+00 -6.10201180e-01 -1.10764360e+00
3.94189268e-01 1.80229187e-01 1.34113646e+00 2.58828551e-01
-2.09697828e-01 2.27443531e-01 -7.79874206e-01 -5.17970920e-01
-7.23585367e-01 4.81649220e-01 -4.90567267e-01 -4.40656602e-01
6.38736069e-01 -4.50438410e-01 -4.50358279e-02 5.46971932e-02
-6.75955236e-01 3.36211234e-01 9.28045273e-01 6.56752646e-01
-5.31076863e-02 9.07869264e-02 6.72424555e-01 -1.07409394e+00
9.16895807e-01 -6.65171146e-01 -2.98902303e-01 6.08530402e-01
-8.60421717e-01 1.86595857e-01 6.46987021e-01 -1.67320207e-01
-1.53483224e+00 -2.39288017e-01 -3.35116059e-01 -4.33202945e-02
1.64188415e-01 1.06726229e+00 -2.01867614e-02 9.38379019e-02
7.40089715e-01 4.28777069e-01 -4.66010012e-02 -6.75270677e-01
1.38144806e-01 1.14052296e+00 -2.84104254e-02 -9.51427400e-01
7.63151109e-01 -2.48196989e-01 -8.10918927e-01 -4.51321304e-01
-6.61733210e-01 -4.69549596e-01 -5.91783524e-01 -1.00722171e-01
8.90427709e-01 -9.29198027e-01 -1.05030455e-01 7.48798490e-01
-1.63830554e+00 -3.91824245e-01 2.25030735e-01 2.60882586e-01
-4.93643224e-01 3.05355161e-01 -6.17645681e-01 -4.62770492e-01
-4.62953359e-01 -1.48700130e+00 1.19249535e+00 -9.97928381e-02
-1.16726674e-01 -1.05116928e+00 1.60873696e-01 4.28563684e-01
6.24111652e-01 -8.38220865e-02 1.24815691e+00 -3.90215576e-01
-6.31084800e-01 -7.13836104e-02 -2.62900800e-01 6.00287914e-01
3.79456609e-01 3.01955491e-01 -7.76888847e-01 -2.08388761e-01
1.69885874e-01 -5.74510217e-01 5.79482436e-01 -1.09086357e-01
1.40437770e+00 -1.53549369e-02 -3.44093204e-01 4.16454881e-01
1.40623462e+00 3.24173689e-01 7.64547229e-01 3.39613259e-01
7.40153313e-01 4.40712571e-01 5.04231691e-01 4.62311804e-01
7.08780944e-01 2.98406720e-01 1.40876383e-01 1.32523805e-01
-7.09796995e-02 -3.06911707e-01 7.19380558e-01 1.55261266e+00
-1.11418612e-01 5.76215051e-02 -1.51257467e+00 6.71222508e-01
-1.68570375e+00 -6.30037367e-01 -3.69421214e-01 1.91285193e+00
1.17031777e+00 1.97191164e-01 -2.57304400e-01 3.99849489e-02
5.75099885e-01 5.43238968e-02 -2.81190664e-01 -5.48446655e-01
6.21386528e-01 2.56899565e-01 1.04788570e-02 1.32797778e-01
-9.09305394e-01 8.85603011e-01 5.75138235e+00 9.17216480e-01
-1.06778181e+00 5.99525690e-01 1.59603447e-01 6.27745152e-01
-4.08171177e-01 2.72315055e-01 -6.26148999e-01 5.28110206e-01
1.03176248e+00 -6.83982551e-01 3.11028183e-01 1.38917327e+00
4.30229791e-02 9.22622308e-02 -1.35334861e+00 7.52162337e-01
2.10642457e-01 -1.06404901e+00 -8.00126716e-02 -2.87917763e-01
5.91361463e-01 1.05232149e-01 -1.14393625e-02 9.16370869e-01
2.59004474e-01 -7.19896495e-01 6.76291645e-01 2.24315256e-01
1.01686990e+00 -3.02531004e-01 7.95880020e-01 5.37614286e-01
-1.34414530e+00 -1.76584661e-01 -2.98698366e-01 -1.35653138e-01
3.34549285e-02 7.64985621e-01 -6.79385483e-01 6.89436674e-01
8.68714154e-01 1.14833808e+00 -7.59266138e-01 8.11692059e-01
-3.57638747e-01 6.40215874e-01 3.38396132e-01 -1.08973950e-01
-1.51879042e-01 1.03013702e-01 -7.99161345e-02 1.42292142e+00
7.00154364e-01 -2.40836769e-01 2.18700126e-01 1.11367047e+00
4.39089425e-02 1.39743745e-01 -7.50794888e-01 -4.19382423e-01
2.94954270e-01 1.08044279e+00 -3.33572507e-01 -1.84187591e-01
-9.38480973e-01 5.70065796e-01 2.61610091e-01 2.05550089e-01
-8.40695381e-01 -7.50135303e-01 3.62677157e-01 2.40015402e-01
1.09415524e-01 -5.04412353e-01 -4.33726519e-01 -1.27414119e+00
2.69775987e-01 -1.20552766e+00 7.67582282e-02 -9.85583007e-01
-1.30739999e+00 6.77081645e-01 2.13189702e-02 -1.40597844e+00
-1.61207095e-01 -7.38468289e-01 -7.86266744e-01 6.95043206e-01
-1.79230380e+00 -1.30034721e+00 -3.16663623e-01 3.12763691e-01
1.05887294e+00 -5.13363183e-01 7.28808999e-01 6.10498130e-01
-4.32386965e-01 4.15564001e-01 1.45404218e-02 3.51094991e-01
9.64984059e-01 -1.19908702e+00 6.52882576e-01 8.36512089e-01
-1.52404234e-01 9.84752417e-01 2.64943480e-01 -5.42265832e-01
-1.52669716e+00 -1.06169176e+00 1.17682278e+00 -5.98343551e-01
9.00558710e-01 -5.24648249e-01 -1.05682456e+00 8.72932374e-01
6.03234470e-01 -4.22362030e-01 8.21132183e-01 2.30301276e-01
-6.28746450e-01 -2.20798165e-01 -7.12580502e-01 2.77186692e-01
8.75009120e-01 -8.60251546e-01 -8.96928012e-01 3.60102981e-01
8.22169065e-01 -2.98314422e-01 -1.07422769e+00 2.90449798e-01
4.00558263e-01 -1.01001751e+00 6.43445790e-01 -1.66532636e-01
1.24777198e+00 -2.30859831e-01 6.79869205e-03 -1.40946579e+00
-3.15537483e-01 -5.16135693e-01 9.31916535e-02 1.48793340e+00
5.34662366e-01 -5.75754464e-01 2.90340722e-01 3.09595376e-01
-4.99591023e-01 -6.30865037e-01 -6.91456318e-01 -6.65292621e-01
3.30837399e-01 -6.59260273e-01 4.96236891e-01 9.20138299e-01
3.74231070e-01 5.51670372e-01 -1.65328652e-01 -3.37953269e-01
2.45297223e-01 7.98487291e-02 7.67333269e-01 -1.19324744e+00
-3.85403484e-01 -3.77608597e-01 1.72669441e-01 -9.29645658e-01
4.06224042e-01 -1.09399080e+00 2.68582940e-01 -1.57777405e+00
2.86159664e-01 -3.66547614e-01 -1.08223595e-01 9.95675623e-01
-1.80235714e-01 -3.92892510e-01 8.79977345e-02 2.57415682e-01
-3.62927169e-01 5.30576348e-01 1.19889915e+00 -2.56346971e-01
2.36370385e-01 1.55551314e-01 -8.91610026e-01 6.02344453e-01
8.47198486e-01 -8.06092978e-01 -6.26016974e-01 -9.71342742e-01
5.20761609e-01 1.02129854e-01 1.53187020e-02 -8.79268229e-01
3.73701423e-01 -9.79904085e-02 -2.95416355e-01 -2.50466317e-01
-1.92776769e-01 -7.57967234e-01 -5.04729189e-02 5.23461699e-01
-2.73368537e-01 3.14179450e-01 2.54008800e-01 1.37684688e-01
-4.76583630e-01 -6.71213746e-01 6.19758487e-01 -3.34277093e-01
-9.92029071e-01 1.31362334e-01 -5.01406968e-01 2.39405096e-01
1.05927813e+00 -8.07829127e-02 -3.43086511e-01 8.65638629e-02
-2.17898473e-01 1.57078058e-01 2.37569794e-01 9.72216308e-01
7.83113539e-01 -1.06208110e+00 -8.73465538e-01 3.18392307e-01
5.76368988e-01 -6.02910332e-02 4.05497700e-02 7.21201837e-01
-5.72721839e-01 2.91527420e-01 -2.64704883e-01 -5.28670907e-01
-1.11094165e+00 5.67632198e-01 1.51599139e-01 -4.22703415e-01
-3.62437814e-01 8.08743536e-01 1.48066565e-01 -8.11166763e-01
5.30328192e-02 -5.54433584e-01 -4.20653448e-02 -1.94853574e-01
5.25085390e-01 1.88375920e-01 2.49519020e-01 -2.24171150e-02
-2.54343092e-01 7.79444993e-01 -3.28086019e-01 2.53166854e-01
1.41382635e+00 2.25084156e-01 -5.67167759e-01 7.85169005e-01
1.21114469e+00 -2.43127495e-01 -7.88482249e-01 -2.87917793e-01
5.90140104e-01 -4.04353619e-01 -1.63260594e-01 -8.38581979e-01
-9.29325700e-01 1.37540269e+00 3.44398558e-01 2.61060804e-01
1.21540892e+00 -2.60959193e-02 6.78052485e-01 3.19038063e-01
9.26425636e-01 -9.11505103e-01 5.09028554e-01 8.19938123e-01
1.19376993e+00 -1.31413722e+00 -1.56805545e-01 -4.36154753e-01
-6.06349468e-01 1.36806428e+00 8.76162410e-01 9.69896652e-03
9.43012834e-01 7.79066324e-01 1.92035615e-01 -1.25066265e-01
-1.04238236e+00 1.94073901e-01 2.50671152e-02 5.85376680e-01
1.35262895e+00 -2.34898493e-01 -3.23987573e-01 7.64809072e-01
-7.33926594e-02 6.13732338e-01 5.08837640e-01 1.18640351e+00
-3.77016574e-01 -1.53011203e+00 -1.48514718e-01 3.59472692e-01
-4.39429700e-01 -4.69622999e-01 -2.52597809e-01 6.43151224e-01
2.83014685e-01 8.18531156e-01 -4.06926244e-01 -4.84172851e-01
3.55244130e-01 3.63747686e-01 3.91494513e-01 -1.32022798e+00
-8.35954309e-01 -2.44029686e-01 -5.03702424e-02 -2.13314712e-01
-3.79229695e-01 -3.25781792e-01 -1.08761191e+00 -2.51934558e-01
-3.99675131e-01 2.25722879e-01 6.48723900e-01 9.20769334e-01
4.93547916e-01 8.09051752e-01 3.40112150e-01 -5.41266978e-01
-7.09405839e-01 -1.24397731e+00 -2.50114709e-01 4.32037981e-03
9.03404951e-02 -5.93431056e-01 -1.75581738e-01 7.77217507e-01] | [7.591330051422119, 7.953886985778809] |
2c130109-0b5e-47a2-88b1-50e1337fc026 | improving-neural-topic-models-with | 2303.15350 | null | https://arxiv.org/abs/2303.15350v1 | https://arxiv.org/pdf/2303.15350v1.pdf | Improving Neural Topic Models with Wasserstein Knowledge Distillation | Topic modeling is a dominant method for exploring document collections on the web and in digital libraries. Recent approaches to topic modeling use pretrained contextualized language models and variational autoencoders. However, large neural topic models have a considerable memory footprint. In this paper, we propose a knowledge distillation framework to compress a contextualized topic model without loss in topic quality. In particular, the proposed distillation objective is to minimize the cross-entropy of the soft labels produced by the teacher and the student models, as well as to minimize the squared 2-Wasserstein distance between the latent distributions learned by the two models. Experiments on two publicly available datasets show that the student trained with knowledge distillation achieves topic coherence much higher than that of the original student model, and even surpasses the teacher while containing far fewer parameters than the teacher's. The distilled model also outperforms several other competitive topic models on topic coherence. | ['Debarshi Kumar Sanyal', 'Suman Adhya'] | 2023-03-27 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-3.46182764e-01 5.39798439e-01 -3.05843890e-01 -2.77329147e-01
-1.21438408e+00 -3.06294501e-01 8.57377410e-01 2.08244607e-01
-1.27125636e-01 7.42426336e-01 5.00989497e-01 -1.09691292e-01
-8.91545713e-02 -7.83230543e-01 -7.19985664e-01 -9.17678237e-01
1.00346558e-01 8.48661840e-01 2.23386630e-01 3.06040913e-01
1.57704026e-01 -4.92281497e-01 -1.55285347e+00 -4.03490029e-02
1.28257227e+00 7.46985316e-01 7.09504962e-01 3.53615463e-01
-5.35671294e-01 2.24756598e-01 -5.92357218e-01 -2.17139050e-01
-3.03365290e-01 -1.95184559e-01 -7.73277342e-01 9.93697941e-02
5.85072398e-01 -3.01677197e-01 -4.50155497e-01 1.21163452e+00
1.41698465e-01 7.15910077e-01 8.18367720e-01 -9.11882997e-01
-8.65813553e-01 1.03634477e+00 -4.06133860e-01 -1.32123278e-02
-1.03948258e-01 -4.67627615e-01 1.15087950e+00 -1.00386763e+00
7.00505137e-01 1.43239546e+00 1.71139926e-01 3.66100132e-01
-1.44432545e+00 -7.28538394e-01 6.84872091e-01 1.95102155e-01
-1.51710880e+00 1.65821806e-01 8.08964431e-01 -5.11310279e-01
7.11389780e-01 -1.35792896e-01 9.22628939e-01 1.25337362e+00
1.92167148e-01 1.03910863e+00 8.99985433e-01 -2.49907181e-01
5.98292470e-01 8.13573003e-01 5.57932496e-01 4.48364347e-01
2.43594944e-01 -4.25486922e-01 -8.08080971e-01 -6.66688561e-01
4.85360086e-01 1.70128539e-01 -4.49772745e-01 -6.60753310e-01
-8.89834344e-01 1.20661557e+00 1.49649844e-01 2.27742076e-01
-3.57678205e-01 7.09955841e-02 1.55829135e-02 -1.57230526e-01
1.06231415e+00 3.03753912e-01 -3.60740870e-01 -3.66885811e-02
-1.50786698e+00 5.17040431e-01 9.90358829e-01 1.19636774e+00
8.06578279e-01 -1.34653017e-01 -7.52420574e-02 8.51750970e-01
7.67048895e-01 4.27333742e-01 7.27661431e-01 -8.31314147e-01
1.47677392e-01 3.77182752e-01 -1.07253179e-01 -8.01225126e-01
2.02671900e-01 -3.83390039e-01 -6.12598300e-01 -4.37327266e-01
-6.64409176e-02 -1.17665999e-01 -9.25093234e-01 1.70195591e+00
5.35477102e-01 4.48119104e-01 7.34236650e-03 5.36793172e-01
7.18224049e-01 1.20828831e+00 4.21590626e-01 -2.97392547e-01
1.32752335e+00 -9.35867250e-01 -1.05294287e+00 -9.14314613e-02
3.30631107e-01 -5.58314025e-01 9.97971177e-01 5.62137842e-01
-1.22789025e+00 -2.01139957e-01 -9.08769488e-01 -3.36312264e-01
-4.04867381e-01 -4.14870642e-02 5.07518172e-01 4.66670513e-01
-1.11738431e+00 3.93736362e-01 -1.24700177e+00 -3.23263168e-01
4.46382970e-01 -1.07201922e-03 2.15434566e-01 -5.91935813e-02
-1.14963913e+00 6.37933493e-01 6.91940367e-01 -5.72749197e-01
-1.37250388e+00 -1.17380607e+00 -6.79053605e-01 7.12577999e-01
4.38784510e-01 -5.11512756e-01 1.33169091e+00 -1.14811659e-01
-1.56721282e+00 3.64831150e-01 -3.87218416e-01 -5.68923235e-01
2.61343896e-01 -5.33210039e-01 2.68016532e-02 -1.88793633e-02
1.07530378e-01 9.12734926e-01 5.65463781e-01 -1.27829385e+00
-7.08523691e-01 -3.12843740e-01 -3.42264593e-01 3.54268909e-01
-8.74354899e-01 -5.64246058e-01 -8.07345152e-01 -6.09641671e-01
2.53568679e-01 -9.05969441e-01 -1.94014847e-01 -2.10888445e-01
-5.26547372e-01 -1.09307230e+00 1.25637829e+00 -6.51820123e-01
1.13722777e+00 -2.15891099e+00 2.53446251e-01 9.21967477e-02
2.79290557e-01 -1.59784839e-01 1.49119824e-01 2.08837166e-01
2.90892512e-01 2.45165661e-01 -1.08111598e-01 -6.23943090e-01
2.05506772e-01 1.14926457e-01 -9.76564288e-01 2.01252252e-01
-2.61274278e-01 4.77189630e-01 -8.39746177e-01 -6.77618384e-01
-1.67119130e-02 7.71975935e-01 -7.83707261e-01 2.69032121e-01
-7.63193846e-01 -8.42388049e-02 -5.09662867e-01 2.28476480e-01
5.79549074e-01 -5.36872685e-01 3.51523221e-01 3.44882160e-01
2.14421421e-01 7.63948023e-01 -1.07296097e+00 2.06298661e+00
-2.80424625e-01 1.12251914e+00 -1.82862714e-01 -7.19343245e-01
1.00973153e+00 6.03648722e-01 4.72981662e-01 -1.01013713e-01
-2.31740445e-01 -8.51982534e-02 -6.91067636e-01 -4.16542031e-02
1.09362090e+00 -2.00261511e-02 8.29898268e-02 7.70463109e-01
4.45127964e-01 -3.87197882e-01 8.71706083e-02 7.86716700e-01
2.22876653e-01 5.69460653e-02 -1.52897820e-01 -6.65734589e-01
-2.12389588e-01 -1.11134961e-01 4.60361749e-01 7.90534019e-01
7.02162683e-02 3.27761054e-01 5.55207491e-01 -4.52279933e-02
-9.90264058e-01 -1.37582612e+00 -3.53900731e-01 1.15352762e+00
2.71380961e-01 -6.90544605e-01 -8.44981790e-01 -4.12306845e-01
-6.08204044e-02 1.33891511e+00 -6.21680319e-01 -8.54356736e-02
-2.36864060e-01 -4.64608312e-01 1.20660610e-01 4.86418605e-01
3.73603612e-01 -4.42043602e-01 -2.68747717e-01 1.80999935e-01
-5.28863609e-01 -8.06429446e-01 -4.93562967e-01 1.14163004e-01
-1.20957100e+00 -6.65596187e-01 -9.42263365e-01 -6.86379015e-01
3.71124834e-01 3.12098771e-01 1.02077460e+00 -4.85705614e-01
-7.30469972e-02 3.37099403e-01 9.65748057e-02 -6.81143701e-01
1.38978943e-01 3.64839345e-01 1.30309492e-01 -4.41421837e-01
7.97350705e-01 -5.53752005e-01 -4.95970249e-01 -4.95411269e-02
-9.70875323e-01 -1.29660051e-02 4.28876579e-02 8.50385070e-01
7.64504731e-01 2.60997981e-01 4.34456319e-01 -7.78624237e-01
9.06549275e-01 -8.38263392e-01 -6.84356093e-01 3.52229774e-01
-1.07918239e+00 2.28882000e-01 -1.72487170e-01 -7.44961321e-01
-1.59086764e+00 -5.74815273e-01 6.22945070e-01 -5.05161405e-01
-5.15604876e-02 6.55207992e-01 1.23112611e-01 8.15444767e-01
3.71289819e-01 4.13824558e-01 -4.40730304e-01 -7.39931345e-01
5.79940200e-01 6.09159648e-01 3.58604431e-01 -8.13759863e-01
3.61111104e-01 5.32981396e-01 -6.92814469e-01 -1.10484028e+00
-9.84401047e-01 -5.76693356e-01 -3.64247978e-01 1.26781568e-01
8.42492580e-01 -1.27787983e+00 -3.15980524e-01 -7.30471173e-03
-1.25724602e+00 -1.40158027e-01 -4.97414559e-01 7.96375275e-01
-4.72915232e-01 2.43353009e-01 -4.24872249e-01 -7.89984703e-01
-3.80437523e-01 -1.02962470e+00 1.14086878e+00 6.57637119e-01
-1.18914150e-01 -1.36865819e+00 4.47695553e-01 1.89917341e-01
4.19327319e-01 -2.29110509e-01 1.25879991e+00 -9.97838318e-01
-7.55543590e-01 -2.91789044e-02 -1.12551950e-01 1.91749372e-02
-1.11087970e-01 -7.39899352e-02 -1.04141796e+00 -3.66620958e-01
2.08087172e-02 -2.85506397e-01 1.25289297e+00 8.18037152e-01
1.14713633e+00 -3.48408818e-01 -6.77698493e-01 3.33381295e-01
1.29507804e+00 3.55011038e-02 3.10387403e-01 2.10296348e-01
2.65361726e-01 5.54373682e-01 2.76118547e-01 4.32866067e-01
6.42648816e-01 3.77470493e-01 1.25131980e-01 4.60329592e-01
1.94685951e-01 -6.36790454e-01 3.48141313e-01 1.10996628e+00
3.76805812e-01 -3.66274834e-01 -9.38898504e-01 1.02948940e+00
-1.85462892e+00 -7.32404888e-01 2.61863232e-01 2.07627249e+00
1.15816545e+00 -2.06165045e-01 -2.11016163e-01 -6.09887540e-01
5.32452643e-01 2.78678238e-01 -5.16597092e-01 -1.24428146e-01
2.23105978e-02 -1.23502575e-01 -1.02101982e-01 6.80033445e-01
-1.08396947e+00 1.21029842e+00 6.51162624e+00 8.90191793e-01
-6.60616815e-01 3.01938295e-01 4.05945390e-01 -5.70274591e-01
-7.12546051e-01 1.38016060e-01 -1.21757185e+00 4.88920569e-01
1.21336532e+00 -7.32509434e-01 -1.33289099e-01 1.46732759e+00
-3.18047591e-02 -3.83372754e-01 -1.06721652e+00 5.30887604e-01
1.43772766e-01 -1.34269416e+00 2.54946411e-01 4.71249580e-01
1.26795030e+00 6.81886449e-02 3.68750066e-01 5.90461016e-01
7.86087215e-01 -7.72722363e-01 4.36552286e-01 5.65265894e-01
2.97923505e-01 -8.27993929e-01 5.54968417e-01 4.60128576e-01
-8.06063592e-01 2.88590819e-01 -8.57152700e-01 5.38528621e-01
1.67048633e-01 8.23047876e-01 -1.03068709e+00 7.03438148e-02
9.68561113e-01 5.21048546e-01 -2.14219421e-01 1.00878716e+00
-1.19332984e-01 1.04841292e+00 -5.05821049e-01 -1.81378245e-01
4.19202387e-01 -3.35217565e-01 7.70716786e-01 1.23553169e+00
3.86664152e-01 8.57411772e-02 4.75116670e-01 1.22369444e+00
-1.97589204e-01 1.12536877e-01 -4.43954587e-01 -2.89382309e-01
6.19043648e-01 8.16122472e-01 -4.06870395e-01 -9.17647421e-01
-1.32945003e-02 6.88408971e-01 3.18590879e-01 6.84953928e-01
-5.05315959e-01 -3.72973680e-01 7.79317439e-01 -1.94236532e-01
4.96823490e-01 -3.34949017e-01 -2.39431337e-01 -1.42309439e+00
-1.58491671e-01 -3.94516736e-01 3.64413977e-01 -6.88066125e-01
-9.52677131e-01 4.56732929e-01 5.39507627e-01 -4.41705763e-01
-5.00148118e-01 -1.05612926e-01 -6.76787913e-01 1.09144807e+00
-1.43555844e+00 -7.76210546e-01 -1.71747684e-01 4.24475044e-01
9.80388284e-01 -1.53506517e-01 8.87165964e-01 -2.04608068e-01
-4.30217922e-01 2.69296885e-01 7.75818706e-01 -3.66649121e-01
5.93776822e-01 -1.53349781e+00 8.26125965e-02 4.12150055e-01
2.23152712e-01 9.73584533e-01 9.29775000e-01 -9.20974016e-01
-8.66993189e-01 -9.72877145e-01 9.94371712e-01 -3.06425810e-01
5.37763238e-01 -3.91436249e-01 -1.41687107e+00 8.58023882e-01
6.70427620e-01 -7.71018445e-01 1.05434060e+00 6.28444910e-01
-4.84750748e-01 4.44680721e-01 -7.99072266e-01 4.93026614e-01
5.08207828e-02 -6.29038513e-01 -9.21485424e-01 6.61129773e-01
1.11067879e+00 -1.16375767e-01 -9.35748100e-01 -3.23171288e-01
4.01260048e-01 -4.20450807e-01 8.30601633e-01 -7.14057088e-01
5.61378181e-01 2.32776821e-01 -2.40853861e-01 -1.46041167e+00
-6.72089532e-02 -3.80258352e-01 -7.01183796e-01 1.30880153e+00
3.59599978e-01 -2.40869477e-01 1.15336680e+00 8.72039974e-01
1.97272822e-01 -7.74474263e-01 -8.10374320e-01 -7.47659385e-01
6.24774814e-01 -2.60682046e-01 4.16235119e-01 9.76481080e-01
3.41448456e-01 2.34055504e-01 -1.61006272e-01 1.29511863e-01
1.06893253e+00 2.01206222e-01 4.90253299e-01 -1.51614475e+00
-3.62327918e-02 -4.34560359e-01 2.28986114e-01 -1.35143292e+00
3.16582501e-01 -9.02784407e-01 1.95722118e-01 -1.57385707e+00
6.89880371e-01 -2.58127511e-01 -1.91178843e-01 1.90693960e-01
-4.34478790e-01 -5.18027484e-01 -4.25842628e-02 3.52991492e-01
-6.45646513e-01 1.18294132e+00 7.94447780e-01 -3.73547137e-01
-5.04094958e-01 -1.71698630e-01 -6.67421997e-01 7.88726628e-01
6.16102278e-01 -7.86762774e-01 -6.66190624e-01 -5.86600423e-01
2.41246037e-02 -7.06568211e-02 6.17858656e-02 -5.52469432e-01
7.39267588e-01 -1.67986751e-01 1.94025412e-01 -1.16903257e+00
5.11171341e-01 -4.37050164e-01 -2.41751194e-01 1.13598526e-01
-8.99796844e-01 -4.88491356e-01 8.72421786e-02 1.24529886e+00
-2.54417211e-01 -2.47799426e-01 5.46580672e-01 -1.30745173e-01
-2.81172901e-01 3.00784349e-01 -5.16230345e-01 1.70820013e-01
7.67500222e-01 2.60731429e-01 -1.94172248e-01 -5.72967947e-01
-7.66927600e-01 5.73796153e-01 1.36598602e-01 5.23073256e-01
6.24252141e-01 -1.15936363e+00 -6.75846100e-01 -1.18992098e-01
-8.96633863e-02 5.21752477e-01 3.66277575e-01 3.16034615e-01
5.95845319e-02 1.15149641e+00 3.78350198e-01 -7.17242777e-01
-1.02058136e+00 3.61500949e-01 -4.33417037e-02 -5.23895264e-01
-6.63025796e-01 1.03953111e+00 5.59615195e-01 -4.90804166e-01
7.91459680e-01 -1.89552352e-01 -6.74873665e-02 2.52772480e-01
5.95406115e-01 5.55109739e-01 -2.92419076e-01 8.62670317e-02
7.88024664e-02 1.87098578e-01 -7.52625763e-01 -5.19198895e-01
1.45742452e+00 -1.93480223e-01 1.04188256e-01 6.93216920e-01
1.35182202e+00 -6.13779008e-01 -1.43138981e+00 -9.15407419e-01
2.10175261e-01 -3.01287085e-01 6.34986341e-01 -5.03339171e-01
-6.16130531e-01 1.28618050e+00 5.52116632e-01 3.59917700e-01
5.20065248e-01 3.36275429e-01 8.18455517e-01 6.47583306e-01
-1.52115384e-03 -1.45510983e+00 3.06961387e-01 4.19943750e-01
6.54345930e-01 -1.18321884e+00 1.60642192e-01 -2.14737102e-01
-5.68953156e-01 9.65081930e-01 5.80395758e-01 -8.78758729e-02
9.91227329e-01 -1.71150938e-01 -3.40729088e-01 -4.19176877e-01
-1.33159697e+00 2.45223850e-01 5.37693679e-01 3.66601288e-01
4.53379422e-01 1.70577466e-01 -1.29906684e-01 6.46206796e-01
-4.60976064e-01 -2.97448784e-01 4.43550795e-01 6.58632398e-01
-9.27570939e-01 -6.68345094e-01 -2.29553401e-01 3.33846480e-01
-4.06380713e-01 -2.07826138e-01 -1.81598350e-01 5.24199128e-01
-4.63815868e-01 7.90588081e-01 4.26878035e-01 6.91950470e-02
-4.08829629e-01 6.30380511e-01 -1.87960789e-01 -7.79415667e-01
-7.19941407e-02 5.49702287e-01 -4.94996309e-01 -2.97104239e-01
1.01635732e-01 -8.04890096e-01 -9.12869394e-01 -3.60720120e-02
-8.77467930e-01 7.56011307e-01 1.21170533e+00 8.62659574e-01
4.01993006e-01 4.47349876e-01 2.66041338e-01 -4.33377832e-01
-7.31591642e-01 -1.35383880e+00 -8.28303635e-01 -1.35675311e-01
1.25009492e-01 -6.30621374e-01 -3.89718771e-01 2.11406782e-01] | [10.406427383422852, 6.93961238861084] |
88f8954e-d136-438e-b9f7-420f7fce564d | deep-meta-learning-for-real-time-visual | 1712.09153 | null | https://arxiv.org/abs/1712.09153v3 | https://arxiv.org/pdf/1712.09153v3.pdf | Deep Meta Learning for Real-Time Target-Aware Visual Tracking | In this paper, we propose a novel on-line visual tracking framework based on the Siamese matching network and meta-learner network, which run at real-time speeds. Conventional deep convolutional feature-based discriminative visual tracking algorithms require continuous re-training of classifiers or correlation filters, which involve solving complex optimization tasks to adapt to the new appearance of a target object. To alleviate this complex process, our proposed algorithm incorporates and utilizes a meta-learner network to provide the matching network with new appearance information of the target objects by adding target-aware feature space. The parameters for the target-specific feature space are provided instantly from a single forward-pass of the meta-learner network. By eliminating the necessity of continuously solving complex optimization tasks in the course of tracking, experimental results demonstrate that our algorithm performs at a real-time speed while maintaining competitive performance among other state-of-the-art tracking algorithms. | ['Janghoon Choi', 'Kyoung Mu Lee', 'Junseok Kwon'] | 2017-12-26 | deep-meta-learning-for-real-time-target-aware | http://openaccess.thecvf.com/content_ICCV_2019/html/Choi_Deep_Meta_Learning_for_Real-Time_Target-Aware_Visual_Tracking_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Choi_Deep_Meta_Learning_for_Real-Time_Target-Aware_Visual_Tracking_ICCV_2019_paper.pdf | iccv-2019-10 | ['real-time-visual-tracking'] | ['computer-vision'] | [-2.32519314e-01 -7.29998529e-01 -1.93848088e-01 -1.89283624e-01
-3.82725418e-01 -4.99821633e-01 5.14801860e-01 -1.24270417e-01
-9.65599418e-01 3.82686108e-01 -3.89644176e-01 1.12186246e-01
-2.01324020e-02 -4.67450261e-01 -6.86795831e-01 -4.86435890e-01
-1.55376419e-01 5.05559921e-01 6.62665248e-01 1.77790895e-01
1.62917078e-01 8.78076911e-01 -1.53856623e+00 -2.06044883e-01
6.70074642e-01 1.22288251e+00 2.74614424e-01 7.33484924e-01
-1.10111296e-01 3.57722729e-01 -4.10412341e-01 -2.73262233e-01
5.08013248e-01 -5.36682196e-02 -6.30450696e-02 1.36386365e-01
1.16084087e+00 -3.59464854e-01 -7.35795498e-01 1.27252316e+00
4.21585172e-01 3.31826180e-01 3.43236089e-01 -1.47949135e+00
-4.55050081e-01 -3.24647650e-02 -6.21136487e-01 5.31126797e-01
-6.72904849e-02 3.57465327e-01 7.11548686e-01 -1.05928683e+00
5.86093843e-01 1.07065356e+00 8.50987911e-01 5.87302804e-01
-9.96716142e-01 -1.06468356e+00 5.49682260e-01 2.65108854e-01
-1.24461186e+00 -2.55625546e-01 8.50155950e-01 -4.72798735e-01
5.38308442e-01 -1.78350076e-01 1.07829475e+00 7.96560466e-01
2.52588600e-01 7.35495865e-01 4.63693172e-01 -9.44508910e-02
1.45818055e-01 -1.33763090e-01 -6.78145746e-03 1.16607535e+00
5.15272081e-01 5.83104908e-01 -6.42424941e-01 -1.89111114e-01
1.00960362e+00 3.08839113e-01 -4.13468778e-02 -1.06770015e+00
-1.12628675e+00 7.77086020e-01 7.92769492e-01 2.02182233e-01
-3.85196358e-01 5.45543015e-01 3.90086383e-01 2.71324545e-01
2.80061215e-01 -5.41313831e-03 -3.98348510e-01 1.43406808e-01
-1.20209467e+00 3.09521437e-01 3.36403489e-01 1.10319722e+00
6.94816768e-01 2.54088312e-01 -3.67923170e-01 3.06370914e-01
6.16218984e-01 7.43491769e-01 4.90114868e-01 -8.41418862e-01
3.47286224e-01 6.82470560e-01 4.02126759e-01 -8.99815261e-01
-4.53882635e-01 -1.06558156e+00 -5.23963749e-01 4.97366488e-01
6.54041767e-01 -2.00819492e-01 -9.77899075e-01 1.74708986e+00
9.30031300e-01 6.28855407e-01 -3.69831651e-01 1.05319285e+00
4.83872622e-01 2.68196344e-01 1.04091227e-01 -1.81705937e-01
1.14443231e+00 -1.24486113e+00 -4.48273450e-01 -1.20854668e-01
4.19803202e-01 -6.75830245e-01 4.59317088e-01 -5.89780137e-02
-8.69492054e-01 -9.93690491e-01 -1.03337300e+00 3.13973427e-01
-3.44564587e-01 3.74883562e-01 7.47765839e-01 3.75935555e-01
-8.73057902e-01 6.22228563e-01 -1.02324986e+00 -2.80907750e-01
5.06226838e-01 6.94506764e-01 -3.84592444e-01 1.86293662e-01
-6.39679611e-01 8.34177077e-01 2.77791589e-01 4.10342842e-01
-1.06550133e+00 -1.03457797e+00 -8.05936217e-01 4.02815416e-02
5.49979389e-01 -8.65684032e-01 1.22415543e+00 -1.13244784e+00
-1.45716000e+00 5.31710863e-01 -1.54299170e-01 -3.42612177e-01
7.56852865e-01 -4.77581203e-01 -3.58674139e-01 -4.92385775e-02
-3.33183371e-02 5.76898515e-01 1.17143345e+00 -9.91641343e-01
-1.06854761e+00 -4.48820412e-01 -1.38761774e-01 7.89117664e-02
-2.74397820e-01 4.04164987e-03 -7.18130171e-01 -5.65897584e-01
4.46673669e-02 -9.18653548e-01 -1.98606715e-01 9.12001550e-01
1.94485888e-01 -3.35740685e-01 1.37304103e+00 -1.34755462e-01
8.97079825e-01 -1.98633695e+00 2.85361614e-02 1.44570768e-01
4.54522252e-01 6.23276651e-01 -4.11449522e-01 -5.67718334e-02
1.41779318e-01 -6.86629593e-01 2.85468638e-01 -3.46247911e-01
-7.44162425e-02 -1.55293584e-01 1.75218955e-02 9.16676402e-01
8.23658183e-02 1.06985652e+00 -1.12277639e+00 -7.64459610e-01
4.21340644e-01 4.90722775e-01 -4.65155989e-01 3.26052278e-01
-3.17713350e-01 4.07493025e-01 -5.89360774e-01 5.45288086e-01
6.39555812e-01 -3.93565893e-01 -7.26558343e-02 -4.14923787e-01
-3.67962390e-01 -4.25826281e-01 -1.22062564e+00 1.88890636e+00
-3.09487909e-01 6.84722722e-01 1.50959492e-01 -8.60946476e-01
8.83423328e-01 -1.60022840e-01 7.28309572e-01 -5.73967457e-01
3.15924883e-01 1.45992085e-01 1.64187759e-01 -3.07795155e-04
4.82365221e-01 8.82144570e-02 2.36872882e-01 1.54572606e-01
3.35777313e-01 5.63254178e-01 2.56056130e-01 -3.08594350e-02
9.42747355e-01 5.17340183e-01 2.02507731e-02 3.54631431e-02
7.98392117e-01 9.34749991e-02 7.95601904e-01 7.74218500e-01
-5.59303343e-01 -1.02459674e-03 -3.59315544e-01 -8.39155018e-01
-8.86561513e-01 -1.19356716e+00 1.19558014e-01 1.18764770e+00
4.12110031e-01 -8.87778550e-02 -1.90053880e-01 -8.13398123e-01
4.67392802e-01 -6.05716705e-02 -7.89981663e-01 -1.93717897e-01
-7.54257679e-01 -2.09357262e-01 2.57618845e-01 8.98407340e-01
5.98376453e-01 -7.70328522e-01 -1.00907314e+00 3.64997655e-01
4.07007962e-01 -9.94203150e-01 -8.86676729e-01 5.11235893e-02
-1.01848125e+00 -1.13140237e+00 -8.48002970e-01 -8.73962164e-01
8.56731892e-01 4.29228127e-01 5.40874660e-01 2.99297363e-01
-6.04597628e-01 3.90895009e-01 9.25012752e-02 -4.20961082e-02
7.42727369e-02 -6.71525598e-02 1.20631240e-01 1.58717975e-01
2.35555977e-01 -3.34276408e-01 -7.07040191e-01 3.48098099e-01
-3.75964463e-01 -2.76901692e-01 5.07524550e-01 8.87548506e-01
6.58008158e-01 -2.51377016e-01 2.36297026e-01 -3.77652347e-01
9.85022113e-02 -1.72079012e-01 -1.49266696e+00 3.38059604e-01
-3.72108549e-01 1.38242200e-01 7.18589127e-01 -9.96892989e-01
-9.43654180e-01 6.28166258e-01 4.17371571e-01 -1.09505188e+00
3.18965703e-01 1.69124827e-01 1.56566456e-01 -7.08228409e-01
3.85870308e-01 3.04167509e-01 1.18365310e-01 -3.12919050e-01
4.75988358e-01 1.31392747e-01 9.56392229e-01 -3.46190065e-01
1.57500911e+00 5.79617023e-01 2.46215299e-01 -2.42850304e-01
-9.81537938e-01 -8.14477324e-01 -1.00217187e+00 -7.18873382e-01
6.00449920e-01 -9.67526913e-01 -1.04493821e+00 4.98239189e-01
-8.32513213e-01 -1.19846031e-01 -1.83499485e-01 7.76379704e-01
-4.41897780e-01 2.87462205e-01 -1.99738562e-01 -6.48199975e-01
-3.89404923e-01 -7.86888540e-01 9.52624798e-01 7.46856511e-01
2.85460502e-01 -1.15355647e+00 2.94960797e-01 -1.32711038e-01
6.67837739e-01 2.64155954e-01 1.41094416e-01 -4.80811954e-01
-8.97172689e-01 -5.21326482e-01 -3.72745544e-01 -2.97492206e-01
1.12608194e-01 6.49519116e-02 -6.24367058e-01 -8.55019927e-01
-4.33742106e-01 -1.97896048e-01 7.90497959e-01 4.69714612e-01
8.78655970e-01 1.76754311e-01 -7.01148093e-01 1.01414669e+00
1.48670280e+00 1.49089560e-01 -2.13495195e-01 4.44842517e-01
5.56374490e-01 -7.00797811e-02 1.00113702e+00 5.41461110e-01
3.38860422e-01 8.53474140e-01 5.50911725e-01 -1.53032839e-01
-2.70754278e-01 -3.53181601e-01 2.15037629e-01 3.73112798e-01
1.32067084e-01 2.35650182e-01 -3.95846903e-01 5.44335425e-01
-2.13395166e+00 -1.15467989e+00 2.16913655e-01 2.30379033e+00
4.62026685e-01 1.06972702e-01 3.81311983e-01 -5.56827962e-01
8.48108232e-01 1.33232713e-01 -9.99901593e-01 3.87015641e-01
2.12272972e-01 2.62810923e-02 6.54374063e-01 2.91057795e-01
-1.41678929e+00 1.10445905e+00 5.65097427e+00 4.74722981e-01
-1.44061446e+00 5.76486886e-02 -3.15149516e-01 -3.59683484e-01
3.52095515e-01 -1.25164255e-01 -1.05062687e+00 4.18115586e-01
6.16668940e-01 -4.21762347e-01 2.53463656e-01 1.08544850e+00
-1.17822744e-01 6.49321675e-02 -1.06616306e+00 1.12599885e+00
8.38156193e-02 -1.42749822e+00 -2.88936734e-01 -1.42390355e-01
5.57683468e-01 2.89411575e-01 8.68958980e-02 4.05787379e-01
4.03007925e-01 -3.80541921e-01 7.45281696e-01 6.96498275e-01
4.74528342e-01 -6.28180623e-01 5.50751388e-01 2.15425640e-01
-1.85639822e+00 -1.71857134e-01 -5.01142323e-01 3.09012085e-01
9.50149633e-03 1.66652665e-01 -3.92609805e-01 3.99458021e-01
7.10386872e-01 7.42283702e-01 -6.80467844e-01 1.77341294e+00
1.16225272e-01 2.18341611e-02 -4.36451882e-01 -5.96313104e-02
3.13313395e-01 2.39562929e-01 6.26849413e-01 9.60620701e-01
2.87979841e-01 -3.58025014e-01 6.47132993e-01 6.74737334e-01
-2.21238751e-02 -1.88453764e-01 -3.57135445e-01 3.46885845e-02
5.32662928e-01 1.58191943e+00 -6.86092556e-01 -3.93639356e-01
-4.45146441e-01 8.96124542e-01 6.52487278e-01 1.79215863e-01
-1.09475553e+00 -1.59069732e-01 6.34274483e-01 -1.16430737e-01
8.08653533e-01 -4.52248991e-01 2.58587182e-01 -1.16906345e+00
8.89238808e-03 -3.45136970e-01 4.12077725e-01 -4.90002245e-01
-1.18185520e+00 3.67310613e-01 -2.82971203e-01 -1.46631289e+00
3.62588139e-03 -6.51243150e-01 -5.83549678e-01 6.40168011e-01
-1.58328295e+00 -1.30846345e+00 -6.06312692e-01 9.12450373e-01
3.88131768e-01 -3.39560419e-01 3.35237831e-01 4.19211835e-01
-3.99064511e-01 8.08367848e-01 1.54796854e-01 3.58974069e-01
8.24951530e-01 -9.10257578e-01 3.34154308e-01 8.92380774e-01
2.46913746e-01 6.28220916e-01 5.21609783e-01 -7.95884728e-01
-1.65810263e+00 -1.20202899e+00 5.58248341e-01 -1.49245247e-01
8.21281910e-01 -1.65822878e-01 -6.63817704e-01 6.21975422e-01
-1.50478259e-01 8.07129264e-01 3.54567826e-01 5.55577986e-02
-5.05402744e-01 -3.64212424e-01 -9.42466140e-01 4.57621247e-01
1.23360837e+00 -3.00354004e-01 -5.07723808e-01 2.00266466e-01
4.15162206e-01 -6.34292960e-01 -6.93470776e-01 3.17993820e-01
8.53968740e-01 -4.81492013e-01 1.08516848e+00 -8.46027017e-01
-4.00657088e-01 -8.48163188e-01 1.78450957e-01 -1.07365537e+00
-8.07623804e-01 -5.58622003e-01 -5.11118948e-01 7.46098161e-01
-6.05214275e-02 -4.36908394e-01 1.12853801e+00 3.66171867e-01
-1.73708536e-02 -4.86677498e-01 -1.21362066e+00 -1.07297909e+00
-4.18955028e-01 -1.49154231e-01 3.03599924e-01 6.70053959e-01
-4.31345493e-01 -1.44851338e-02 -4.35631126e-01 4.62638676e-01
1.19457853e+00 5.81498742e-01 1.04158652e+00 -1.52714121e+00
-2.24764198e-01 -3.97300333e-01 -6.74471915e-01 -1.19642079e+00
2.51050711e-01 -8.13963830e-01 1.10004567e-01 -1.04346931e+00
3.31511825e-01 -3.78253907e-01 -5.65077066e-01 4.37975436e-01
-3.15492928e-01 1.04069598e-01 5.61043680e-01 1.77876681e-01
-1.15536690e+00 6.91879988e-01 1.26424456e+00 -2.62488246e-01
-1.16928719e-01 2.53357708e-01 -1.50937080e-01 6.91980004e-01
4.13621664e-01 -7.08710492e-01 -2.06548363e-01 -5.01294076e-01
-3.41238588e-01 -3.99991162e-02 6.34608865e-01 -1.30168355e+00
8.49296093e-01 -9.38414708e-02 8.75577033e-01 -9.85027552e-01
5.02817810e-01 -1.21465480e+00 7.12862238e-02 8.93118262e-01
-1.22916363e-01 3.02506596e-01 3.40379179e-01 8.26975763e-01
4.84439693e-02 -7.11430535e-02 9.94060457e-01 1.65460780e-01
-9.49241877e-01 8.52680385e-01 -9.53886509e-02 7.88326468e-03
1.32171226e+00 -4.14154381e-01 -2.23998562e-01 -2.57983264e-02
-6.70883119e-01 5.42484045e-01 4.46069717e-01 6.72162473e-01
5.81410110e-01 -1.56635153e+00 -2.96745181e-01 1.58941761e-01
-4.09346819e-02 -3.52204323e-01 2.54990011e-01 8.27474415e-01
-2.51192302e-01 3.14226598e-01 -4.78845298e-01 -8.03794205e-01
-1.52312732e+00 7.41393149e-01 6.01972044e-01 -2.73293108e-01
-7.78302908e-01 8.99310708e-01 1.06807850e-01 -1.66546062e-01
5.93718410e-01 -2.99302749e-02 4.45314087e-02 -1.17125593e-01
5.65691411e-01 2.36778840e-01 -3.70264232e-01 -6.20273232e-01
-6.34034693e-01 8.45405936e-01 -2.63711810e-01 1.43644258e-01
1.21332252e+00 -5.61521482e-03 3.68339330e-01 1.61993027e-01
1.00834405e+00 -1.82933614e-01 -1.92479122e+00 -6.46518111e-01
1.11044623e-01 -7.82935202e-01 1.94482416e-01 -6.05697989e-01
-1.59438753e+00 5.60959995e-01 1.01949310e+00 -5.84966302e-01
9.51848388e-01 -2.01128542e-01 6.72725916e-01 4.86052215e-01
4.08197969e-01 -1.19606912e+00 3.14165950e-01 4.49105024e-01
4.57730949e-01 -1.21217549e+00 8.24119896e-02 -8.09993420e-04
-1.11507274e-01 1.22757542e+00 1.11128414e+00 -4.60699201e-01
7.28227973e-01 3.27806443e-01 2.01673761e-01 -1.69274315e-01
-6.91175640e-01 -3.20992857e-01 6.16245747e-01 6.83879614e-01
5.91909997e-02 -3.83268446e-01 -3.82483914e-03 -1.76078584e-02
2.64969289e-01 1.14982292e-01 -2.02286392e-01 1.04855657e+00
-7.13718772e-01 -1.04736745e+00 -3.87260020e-01 2.51753360e-01
-5.14734462e-02 3.05252492e-01 -1.81281090e-01 9.41517472e-01
8.82364623e-03 3.90582532e-01 2.40933701e-01 -2.47446656e-01
2.81265557e-01 -1.85360238e-01 7.26826370e-01 -2.89778888e-01
-7.19181895e-01 1.44159004e-01 -4.46840078e-01 -7.92056680e-01
-5.54544866e-01 -6.50167048e-01 -1.25391006e+00 -1.06774084e-01
-6.31216884e-01 5.66945188e-02 5.67693889e-01 8.65665257e-01
3.73457879e-01 5.10589361e-01 5.19007385e-01 -1.07994413e+00
-7.25502193e-01 -6.74982727e-01 -2.42946208e-01 1.06685266e-01
5.00656843e-01 -1.22045088e+00 3.50383408e-02 -2.71733552e-01] | [6.279449939727783, -2.1389145851135254] |
51eaad51-7462-4c43-9bbd-4fbcdecdf20d | it-is-ais-turn-to-ask-human-a-question | null | null | https://openreview.net/forum?id=kKUWbb_gzI0 | https://openreview.net/pdf?id=kKUWbb_gzI0 | It is AI’s Turn to Ask Human a Question: Question and Answer Pair Generation for Children Storybooks in FairytaleQA Dataset | Existing question answering (QA) techniques are created mainly to answer questions asked by humans. But in educational applications, teachers and parents sometimes may not know what questions they should ask best help children to develop their narrative understanding abilities. We design an automated question-answer generation (QAG) system for education purposes: given a storybook at the kindergarten to eighth-grade level, our system can automatically produce QA pairs that are capable of testing a variety of student comprehension skills. Using a new QA dataset FairytaleQA that has 278 child-friendly storybooks with 10,580 QA pairs labeled by experts, we design a novel QAG system architecture to generate QA pairs. Automatic and human evaluations show that our model outperforms state-of-the-art QAG systems. On top of our QAG system, we also build an interactive story-telling application for future real-world deployment. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['question-answer-generation'] | ['natural-language-processing'] | [-2.01771468e-01 7.10599065e-01 4.06264514e-01 -5.84389210e-01
-1.15426719e+00 -9.97439384e-01 2.87798941e-01 3.54743540e-01
3.58086079e-01 8.43913615e-01 4.84342694e-01 -8.02672446e-01
-6.50719181e-02 -1.42243648e+00 -5.27732611e-01 3.47819269e-01
4.48270708e-01 1.04563761e+00 8.99468184e-01 -8.55515778e-01
2.26079404e-01 -4.93106842e-02 -1.90365219e+00 8.24144125e-01
1.59980762e+00 2.92465866e-01 5.50031289e-02 1.28530133e+00
-5.86885035e-01 1.90049970e+00 -1.06911349e+00 -9.26665068e-01
-3.12224507e-01 -9.00333405e-01 -1.70875192e+00 -3.61091197e-01
8.37887108e-01 -9.17276502e-01 4.67691279e-04 5.80720305e-01
4.49329972e-01 2.14265004e-01 3.94021958e-01 -1.32256234e+00
-9.20199513e-01 9.64111686e-01 2.30944738e-01 -1.40102684e-01
1.34251940e+00 1.20459281e-01 9.13611352e-01 -5.55391490e-01
5.39119720e-01 1.04562378e+00 3.00318241e-01 1.05846453e+00
-8.51924300e-01 -4.60378498e-01 -2.68571228e-01 7.07688868e-01
-8.34003210e-01 -1.32726178e-01 7.50174046e-01 -4.34471518e-01
7.50856340e-01 5.14390647e-01 1.06807113e+00 7.05126643e-01
-5.11154011e-02 1.05469966e+00 9.54338610e-01 -5.66595078e-01
2.72178620e-01 6.31133243e-02 4.87025201e-01 9.70574856e-01
-1.73573688e-01 -6.02406740e-01 -7.39014328e-01 1.03690960e-02
5.82667232e-01 -7.99157202e-01 -1.43357635e-01 -1.65727556e-01
-7.76808083e-01 1.04002106e+00 -1.80287406e-01 1.28966823e-01
2.13312000e-01 -4.48929146e-03 1.39518246e-01 7.47986853e-01
-6.07705303e-02 1.13429701e+00 -3.16514432e-01 -6.82033956e-01
-6.58410847e-01 1.02632177e+00 1.44433022e+00 1.13161540e+00
1.40251040e-01 -2.70836234e-01 -5.18183947e-01 8.65878403e-01
2.31131300e-01 2.67874122e-01 3.84374976e-01 -1.55460060e+00
5.24303317e-01 1.05153787e+00 2.87572324e-01 -6.40334666e-01
-1.76871940e-01 -1.43788755e-01 1.97613407e-02 1.14688620e-01
8.57663512e-01 -3.06131095e-01 -3.80390257e-01 1.39549422e+00
5.31246960e-01 1.81069553e-01 2.99365789e-01 5.97239375e-01
1.96195805e+00 7.20645785e-01 1.02287091e-01 2.85546511e-01
1.69311595e+00 -1.21855569e+00 -7.30619192e-01 -1.01308689e-01
1.00886893e+00 -5.84791660e-01 1.54489803e+00 8.54582071e-01
-1.73547733e+00 -4.30335134e-01 -1.03118908e+00 -4.75526273e-01
-1.14336066e-01 -6.65669590e-02 3.99531543e-01 1.11165929e+00
-1.03296304e+00 1.03827551e-01 -4.61001694e-01 -1.32934049e-01
2.97674865e-01 -2.08209619e-01 -6.01995848e-02 -5.20980895e-01
-1.08643031e+00 9.57460821e-01 5.60355410e-02 -7.07548738e-01
-1.23253524e+00 -1.03218865e+00 -1.01463556e+00 4.31777865e-01
3.56423765e-01 -8.85196030e-01 2.15596890e+00 -3.57195169e-01
-1.94283843e+00 8.09071541e-01 1.96390375e-01 -1.79810524e-01
3.73092830e-01 -5.53935230e-01 -9.52101946e-02 5.91703713e-01
3.13611269e-01 8.32539380e-01 -1.24190345e-01 -7.79429555e-01
-4.47093487e-01 -2.01801173e-02 1.00223207e+00 5.10854483e-01
-1.31317779e-01 1.56814978e-01 4.69663180e-02 -3.82049173e-01
3.24475974e-01 -1.22731507e-01 -7.57919904e-03 -1.89811662e-01
6.81311190e-02 -7.25784063e-01 5.61800778e-01 -8.59851360e-01
1.25633347e+00 -1.36776733e+00 -4.56551045e-01 -3.01091582e-01
2.23359451e-01 3.60443443e-01 -2.79298723e-01 8.18784952e-01
3.00586838e-02 -1.08424522e-01 1.92820519e-01 8.71787444e-02
2.02242866e-01 7.18498742e-03 -4.29497272e-01 -4.36612457e-01
2.60408700e-01 9.36715782e-01 -1.57762301e+00 -7.65021503e-01
2.09767133e-01 -2.76469678e-01 -7.31072068e-01 1.06747127e+00
-7.35522568e-01 3.27381194e-01 -5.61214328e-01 5.99510133e-01
3.06819350e-01 -6.76450804e-02 -2.83836782e-01 8.78285885e-01
2.69943267e-01 1.05882299e+00 -8.49129438e-01 1.90154099e+00
-5.73306799e-01 7.77587235e-01 -2.02710494e-01 -5.64905405e-01
1.04329073e+00 6.22087538e-01 -1.30025625e-01 -6.16713464e-01
-5.40870354e-02 1.20545933e-02 -1.46262581e-02 -1.04268098e+00
6.02148175e-01 1.86487421e-01 -2.30055705e-01 8.89483035e-01
4.21550691e-01 -1.23821366e+00 6.31287158e-01 6.59276843e-01
1.54878008e+00 2.55505800e-01 6.19216189e-02 -1.13551117e-01
6.60990536e-01 3.88947606e-01 3.15764174e-02 9.52037513e-01
-9.98398513e-02 7.45105088e-01 7.07388401e-01 -4.08110231e-01
-4.95697677e-01 -1.33985448e+00 3.92605811e-01 1.45812857e+00
-2.50337124e-01 -6.37513757e-01 -1.21074069e+00 -7.69309223e-01
-8.47768009e-01 1.49740183e+00 -1.39230877e-01 7.73827881e-02
-6.03194296e-01 2.79627621e-01 8.73380959e-01 3.73686314e-01
7.40153909e-01 -1.17865562e+00 -1.05365074e+00 5.34157515e-01
-5.44574678e-01 -8.57188880e-01 -5.57579808e-02 -6.42013311e-01
-4.58714306e-01 -1.24626863e+00 -5.05656064e-01 -1.02326596e+00
4.46015239e-01 7.31756464e-02 1.84630311e+00 4.44821179e-01
1.10192627e-01 8.31173778e-01 -8.82788002e-01 -5.88351727e-01
-8.51208985e-01 3.77421454e-02 -6.73719585e-01 -1.01714969e+00
4.37083095e-01 -7.65607238e-01 -6.13697886e-01 1.83984771e-01
-6.65424645e-01 7.96631634e-01 -1.23643719e-01 5.79125404e-01
-3.11772853e-01 -1.22175694e-01 8.15352499e-01 -1.16347921e+00
1.03156698e+00 -3.80855620e-01 -6.06630147e-01 6.17746472e-01
-1.38327703e-01 -1.73545599e-01 6.30590916e-01 -3.80562335e-01
-1.47275901e+00 -3.63513350e-01 -8.31715405e-01 4.04258102e-01
-5.70987642e-01 3.50396007e-01 -3.15812021e-01 3.64659652e-02
9.94017363e-01 6.27060533e-02 -6.14225686e-01 -1.54155612e-01
7.33299553e-01 3.80724758e-01 1.06076169e+00 -1.28444171e+00
5.47428668e-01 -5.11303484e-01 -3.75743896e-01 -2.97075391e-01
-9.58550096e-01 -6.37567863e-02 -1.31293535e-01 -7.33039796e-01
7.28308082e-01 -7.41550148e-01 -1.18109417e+00 3.42484593e-01
-1.38854444e+00 -7.68167198e-01 -5.39500892e-01 6.94698617e-02
-7.10938156e-01 -1.09090783e-01 -7.68354416e-01 -7.44928837e-01
-3.41628402e-01 -9.79509175e-01 5.87869167e-01 8.76679778e-01
-6.17781341e-01 -8.68198097e-01 4.46846753e-01 1.25024462e+00
1.02776639e-01 4.66291755e-02 1.33397067e+00 -9.76615608e-01
-4.83382523e-01 1.36683658e-01 5.66175692e-02 2.94190310e-02
-4.83321190e-01 4.46982868e-02 -8.20740521e-01 4.17989105e-01
-2.09581241e-01 -1.03247190e+00 -4.88406830e-02 2.41609327e-02
9.37743545e-01 -4.72673595e-01 2.80879021e-01 -1.40528217e-01
8.11540127e-01 3.49347502e-01 7.79580057e-01 1.17095530e-01
3.43463361e-01 1.06296527e+00 7.40432918e-01 8.69555473e-02
1.31225693e+00 2.38560677e-01 1.96582437e-01 4.06823039e-01
-3.18313092e-01 -7.81611383e-01 3.68756026e-01 8.02455544e-01
2.10994035e-01 -3.60407084e-01 -1.29075062e+00 1.23188376e+00
-1.60945976e+00 -9.39927578e-01 -6.70236468e-01 1.61429548e+00
1.32250857e+00 -3.81139182e-02 7.94789195e-02 2.24540085e-01
1.81563795e-02 -2.78601885e-01 -2.15050951e-02 -8.43865752e-01
3.81487250e-01 9.39325869e-01 -5.44047594e-01 6.57345891e-01
-3.09166253e-01 1.04826927e+00 6.47638416e+00 8.02261412e-01
-1.63917974e-01 7.74302334e-02 7.24240065e-01 2.63744295e-01
-7.62212217e-01 -2.84820646e-02 -4.77076918e-01 -5.76995462e-02
1.02279389e+00 -4.46019083e-01 2.06994236e-01 7.17780113e-01
-1.12011507e-01 -4.08096224e-01 -1.30625463e+00 5.20711541e-01
1.70891792e-01 -1.42137170e+00 -7.24214837e-02 -8.70889783e-01
9.60818291e-01 -8.14119458e-01 -2.86437124e-01 6.98728383e-01
1.16528237e+00 -1.30038857e+00 6.96595252e-01 5.12342811e-01
4.36732024e-01 -8.38581204e-01 7.06299663e-01 6.35018289e-01
-9.09203351e-01 -4.15497124e-02 5.13026491e-02 -6.61367178e-01
5.80892041e-02 -2.48618852e-02 -1.46807623e+00 2.53001720e-01
6.88089490e-01 -1.98863044e-01 -1.01249576e+00 1.15063465e+00
-1.19220734e+00 1.16122758e+00 1.07382022e-01 -5.50392926e-01
2.63623483e-02 -7.55816549e-02 1.79263189e-01 6.31426692e-01
5.33690035e-01 1.24483681e+00 -1.78056061e-01 9.68400955e-01
7.06484467e-02 2.75689811e-01 -3.93910170e-01 8.69365931e-02
8.74292731e-01 1.13880956e+00 -3.83545578e-01 -4.15542156e-01
-4.24608946e-01 6.69647455e-01 2.97484696e-01 1.05113417e-01
-4.02990997e-01 -5.89693606e-01 3.83350313e-01 2.78571993e-01
-3.36411834e-01 -1.08510524e-01 -5.62270761e-01 -8.68508160e-01
-1.51850730e-01 -1.28687429e+00 5.21135688e-01 -1.69778061e+00
-9.76356864e-01 2.47942224e-01 2.48832703e-01 -7.76309967e-01
-9.73516166e-01 -4.78077203e-01 -1.31088245e+00 6.42668068e-01
-7.15778291e-01 -1.20079362e+00 -5.53539574e-01 4.44611669e-01
9.54603851e-01 -5.21687455e-02 1.00878060e+00 -1.46573097e-01
1.01934232e-01 5.61475217e-01 -7.43829131e-01 2.35445157e-01
5.45013607e-01 -1.72915900e+00 7.14124084e-01 7.71101654e-01
1.16381638e-01 3.14468294e-01 1.05130327e+00 -5.60016990e-01
-1.00252283e+00 -6.70300007e-01 1.16372859e+00 -8.43338370e-01
5.93421459e-01 -2.21916825e-01 -1.09030855e+00 4.81283307e-01
7.19176054e-01 -7.35432804e-01 9.85813856e-01 -1.25357732e-01
-1.47147030e-01 2.88725317e-01 -1.30073190e+00 6.93695247e-01
9.48504925e-01 -3.86938632e-01 -1.46946275e+00 4.39473838e-01
1.07769978e+00 -8.84384632e-01 -9.20544624e-01 8.33590031e-02
3.51863593e-01 -1.10600340e+00 8.34034085e-01 -7.78252602e-01
1.12411797e+00 -2.80655712e-01 3.55309695e-01 -1.27271402e+00
2.80119985e-01 -5.38167477e-01 -1.00226149e-01 1.65533614e+00
3.08045357e-01 -3.06951106e-01 1.05115819e+00 1.23717225e+00
-4.48080003e-01 -7.66063929e-01 -6.58463538e-01 -1.11324146e-01
4.74136531e-01 -6.45513833e-01 1.24482870e+00 8.88121724e-01
5.59699953e-01 5.18206418e-01 2.82508284e-01 1.36076540e-01
1.12571269e-01 4.78422642e-02 9.68604445e-01 -1.07927966e+00
-4.06584024e-01 -3.36196363e-01 -8.81256387e-02 -1.11920965e+00
-7.61987790e-02 -3.83194178e-01 1.39525622e-01 -1.95609975e+00
-1.31082341e-01 -1.34856299e-01 7.17912197e-01 4.91315573e-01
-3.48842114e-01 -2.42684465e-02 -2.73869000e-02 -8.09029996e-01
-8.70844424e-01 3.30924302e-01 1.56555498e+00 -4.21062633e-02
-9.47082415e-02 1.27786860e-01 -8.55502784e-01 8.15059781e-01
6.05880916e-01 -1.01502582e-01 -9.42930043e-01 -6.68130696e-01
6.91943169e-01 7.12062478e-01 2.26677358e-01 -1.16302788e+00
5.75147450e-01 -4.85201359e-01 8.29065070e-02 -7.46604741e-01
1.04310212e-03 -2.26252049e-01 -1.89782396e-01 4.72030640e-02
-5.08337498e-01 1.71802461e-01 1.16599783e-01 -5.34202397e-01
-3.27623159e-01 -1.04690015e+00 1.99951053e-01 -8.49305168e-02
-3.63637149e-01 -3.28171015e-01 -8.97501767e-01 4.62304503e-01
9.84422207e-01 -5.56053221e-03 -1.04890478e+00 -1.23405170e+00
-4.05368328e-01 1.01205146e+00 1.37168869e-01 1.71504021e-01
9.75129604e-01 -1.34868896e+00 -1.09454167e+00 -9.22028795e-02
2.57094592e-01 4.34788436e-01 3.19503129e-01 -1.98046997e-01
-1.35451365e+00 3.85153651e-01 -2.39934996e-01 -8.98833349e-02
-1.51410711e+00 -1.08961105e-01 3.22793394e-01 -4.79599386e-01
-7.73118958e-02 1.44592655e+00 -4.28118929e-02 -7.82216907e-01
-1.01039186e-02 -4.38614100e-01 -8.75871658e-01 -2.12654993e-01
1.01190734e+00 5.90232193e-01 4.92123365e-02 -2.17072964e-02
3.43515873e-01 -3.72887589e-02 2.85359323e-01 -5.08881092e-01
1.10709357e+00 1.09655268e-01 7.16617331e-02 1.27406806e-01
3.03852230e-01 1.56569704e-01 -6.88417375e-01 1.38913468e-01
-1.28025427e-01 -4.33565527e-01 -6.06418729e-01 -1.21945369e+00
-3.70565474e-01 9.73557293e-01 8.18300769e-02 6.02541685e-01
1.27004516e+00 3.53241891e-01 7.76674390e-01 5.65332651e-01
5.43660879e-01 -7.36338317e-01 4.28384006e-01 7.14040637e-01
1.03036153e+00 -8.63675058e-01 -2.74377495e-01 -5.64230502e-01
-4.96721864e-01 1.15993440e+00 1.40766752e+00 1.41506016e-01
-2.50530154e-01 -4.00573127e-02 2.90534317e-01 -3.30081284e-01
-1.29270816e+00 -6.46779779e-03 2.75028914e-01 8.12889397e-01
8.03021014e-01 1.56449601e-01 -4.17415470e-01 1.34029186e+00
-1.27954841e+00 -2.73137204e-02 1.20147359e+00 1.00655246e+00
-6.87267959e-01 -1.29627168e+00 -7.51960576e-01 3.22788507e-01
-3.26684326e-01 -9.02523845e-02 -7.30112016e-01 5.54579020e-01
9.52448323e-02 1.74799049e+00 -1.45146608e-01 -1.65630221e-01
4.38078701e-01 2.97469497e-01 9.33431149e-01 -1.12881911e+00
-9.88694966e-01 -1.05126929e+00 7.89567351e-01 -4.76193696e-01
1.03227206e-01 -5.51087499e-01 -1.25970888e+00 -4.02876735e-01
-2.46760175e-01 5.97433686e-01 9.16408300e-02 1.10577035e+00
-2.25020856e-01 4.69873160e-01 1.03820749e-01 1.67234004e-01
-1.74386472e-01 -1.03041732e+00 -7.44381547e-02 6.91426173e-02
-2.09440459e-02 -1.66880324e-01 2.83804357e-01 2.03027651e-01] | [11.513737678527832, 8.02258014678955] |
a2de8ec6-4579-4c08-b2a0-9faf70f3ea3e | t-fftradnet-object-detection-with-swin-vision | 2303.16940 | null | https://arxiv.org/abs/2303.16940v1 | https://arxiv.org/pdf/2303.16940v1.pdf | T-FFTRadNet: Object Detection with Swin Vision Transformers from Raw ADC Radar Signals | Object detection utilizing Frequency Modulated Continous Wave radar is becoming increasingly popular in the field of autonomous systems. Radar does not possess the same drawbacks seen by other emission-based sensors such as LiDAR, primarily the degradation or loss of return signals due to weather conditions such as rain or snow. However, radar does possess traits that make it unsuitable for standard emission-based deep learning representations such as point clouds. Radar point clouds tend to be sparse and therefore information extraction is not efficient. To overcome this, more traditional digital signal processing pipelines were adapted to form inputs residing directly in the frequency domain via Fast Fourier Transforms. Commonly, three transformations were used to form Range-Azimuth-Doppler cubes in which deep learning algorithms could perform object detection. This too has drawbacks, namely the pre-processing costs associated with performing multiple Fourier Transforms and normalization. We explore the possibility of operating on raw radar inputs from analog to digital converters via the utilization of complex transformation layers. Moreover, we introduce hierarchical Swin Vision transformers to the field of radar object detection and show their capability to operate on inputs varying in pre-processing, along with different radar configurations, i.e. relatively low and high numbers of transmitters and receivers, while obtaining on par or better results than the state-of-the-art. | ['Robert Laganiere', 'Martin Bouchard', 'James Giroux'] | 2023-03-29 | null | null | null | null | ['radar-object-detection'] | ['robots'] | [ 5.23603201e-01 -3.28915834e-01 5.34926593e-01 -4.24295753e-01
-5.20793438e-01 -4.35100138e-01 8.90339196e-01 -9.50779691e-02
-7.50170648e-01 5.81029773e-01 -4.23763841e-01 -2.90336192e-01
-4.25286531e-01 -1.23353028e+00 -3.84342223e-01 -7.00519264e-01
-4.30300951e-01 5.46946943e-01 1.78628176e-01 -2.97572613e-01
-6.10385127e-02 1.13287306e+00 -1.84153748e+00 -1.51114002e-01
5.24405301e-01 1.15995109e+00 1.24760177e-02 6.59981072e-01
-1.62778422e-01 1.41541898e-01 -8.93611312e-01 -1.55331895e-01
5.74089170e-01 -1.10213593e-01 3.11623484e-01 -2.59473711e-01
8.29036891e-01 -5.33526719e-01 -2.67938256e-01 9.83625472e-01
5.44640601e-01 -1.03053942e-01 7.85620928e-01 -1.12312317e+00
-2.59874195e-01 9.61684585e-02 -5.59720933e-01 2.44363040e-01
1.44146547e-01 2.94668287e-01 5.50705969e-01 -9.15855169e-01
2.36162364e-01 1.27177179e+00 6.81375563e-01 9.08257142e-02
-1.32434988e+00 -8.22565317e-01 -3.40812922e-01 1.92718416e-01
-1.25807238e+00 -4.90824848e-01 6.26796007e-01 -4.92581338e-01
1.15457189e+00 1.44291952e-01 6.71629727e-01 9.82022524e-01
3.12070668e-01 1.27592355e-01 1.20652044e+00 -3.48381191e-01
2.52133548e-01 -2.15205699e-04 1.97414290e-02 3.26430261e-01
7.70072401e-01 7.06757188e-01 -4.33852792e-01 9.51832235e-02
6.79325879e-01 -4.06068936e-02 -2.98585773e-01 -3.22886974e-01
-1.05891204e+00 9.63426769e-01 5.50403535e-01 3.83825451e-01
-6.94967687e-01 4.03732032e-01 1.32161453e-01 7.05943465e-01
2.74196535e-01 5.48847318e-01 -2.94664413e-01 2.02892587e-01
-1.27253938e+00 4.12692547e-01 6.59786463e-01 5.83292007e-01
8.67859125e-01 6.62693083e-01 2.92283952e-01 1.04075834e-01
5.11211812e-01 1.09831262e+00 7.76420757e-02 -4.88845080e-01
2.05936685e-01 2.90514082e-01 7.98317119e-02 -8.02248657e-01
-7.58678198e-01 -8.12043548e-01 -9.08475697e-01 9.74174321e-01
3.20759505e-01 -2.45249435e-01 -1.26010692e+00 1.19078052e+00
3.03056240e-02 1.01291329e-01 3.51384670e-01 1.02719331e+00
5.48858285e-01 5.96354425e-01 -1.28863811e-01 -2.14392543e-01
1.47208345e+00 4.23073620e-02 -5.91519713e-01 -4.53291655e-01
4.10685539e-02 -8.00010383e-01 5.73172748e-01 6.82688832e-01
-5.70769250e-01 -6.11927688e-01 -1.46082497e+00 2.63995230e-01
-6.23434246e-01 -1.13910869e-01 7.59868681e-01 1.10412753e+00
-7.88552105e-01 4.44044292e-01 -8.29762280e-01 -1.53076530e-01
3.92604828e-01 3.76499057e-01 -1.23487711e-01 5.51691838e-02
-1.17081070e+00 1.32951474e+00 9.68544632e-02 1.86893597e-01
-4.85915422e-01 -1.00193799e+00 -7.12108791e-01 1.56208009e-01
-8.72914568e-02 -6.14717484e-01 9.97348189e-01 -7.47490883e-01
-1.45251560e+00 4.20739740e-01 3.53506893e-01 -1.01557446e+00
6.12711251e-01 -5.15674174e-01 -7.49898136e-01 1.60477489e-01
-1.84060678e-01 5.41139245e-01 1.38507891e+00 -8.16737592e-01
-5.67447305e-01 -3.71158957e-01 -1.51769266e-01 -1.41930833e-01
-2.32250313e-03 -1.51172563e-01 4.17386830e-01 -3.39606762e-01
1.65895328e-01 -5.89766979e-01 -1.61001265e-01 2.39142790e-01
2.82035381e-01 2.21014082e-01 1.23421526e+00 -1.95056573e-01
4.54631358e-01 -2.04233599e+00 -3.04051906e-01 3.56232494e-01
3.92891243e-02 4.49172318e-01 6.03948422e-02 4.09719020e-01
-4.67461720e-02 -4.19345677e-01 -2.66297251e-01 1.77983999e-01
-6.23011738e-02 4.05316949e-02 -6.63229048e-01 6.72911406e-01
5.54287732e-01 5.31597137e-01 -4.74962354e-01 -7.74081051e-02
6.00624740e-01 7.98459113e-01 -9.19483528e-02 -2.31732473e-01
-1.98805571e-01 3.07942659e-01 -2.05714896e-01 5.50479174e-01
9.92370188e-01 4.05272275e-01 -2.48532951e-01 -2.44504333e-01
-4.72812355e-01 3.66191715e-01 -1.19451857e+00 1.20739830e+00
-5.31863272e-01 1.03876019e+00 4.13015038e-01 -1.03406179e+00
1.35834229e+00 1.69653386e-01 2.73805767e-01 -9.46715832e-01
1.58339098e-01 3.21019441e-01 3.38657379e-01 -2.75011063e-01
5.33601701e-01 -7.46705890e-01 2.38398574e-02 1.45889103e-01
5.80902323e-02 -5.64786077e-01 -1.31101031e-02 -2.83609867e-01
1.00190163e+00 2.36220155e-02 1.60997674e-01 -1.02658935e-01
3.33333284e-01 1.07793503e-01 2.25880276e-02 7.10641563e-01
2.87239403e-01 3.70803356e-01 2.92333923e-02 -4.12049562e-01
-7.30386078e-01 -1.27678692e+00 -5.72500408e-01 6.87812805e-01
-1.52037382e-01 8.09823275e-02 -4.89055105e-02 4.07787040e-03
2.19982818e-01 8.11515629e-01 -8.21966603e-02 -1.02905594e-01
-5.70153117e-01 -9.66933072e-01 8.55314732e-01 3.23526919e-01
4.13469881e-01 -8.29010725e-01 -1.66709268e+00 4.87796724e-01
4.60774034e-01 -1.02423131e+00 7.01103389e-01 6.40796363e-01
-1.17240322e+00 -7.20405281e-01 -5.38878798e-01 -1.41106889e-01
1.70758143e-01 3.94159466e-01 9.00997698e-01 -3.87907416e-01
-6.97530270e-01 3.64120334e-01 -3.21223050e-01 -1.01021993e+00
-1.08958013e-01 -2.86994576e-01 2.89521903e-01 -8.99217799e-02
6.63463175e-01 -8.85000467e-01 -3.80034447e-01 -1.67797282e-01
-8.53836060e-01 -3.67047459e-01 1.24237442e+00 5.40408969e-01
1.44517511e-01 1.01375535e-01 6.00177824e-01 -5.44933736e-01
3.88059556e-01 -7.18218014e-02 -1.10256684e+00 -3.82753700e-01
-5.11746824e-01 5.23622669e-02 6.88831389e-01 -2.34138533e-01
-6.19820416e-01 1.76661894e-01 -8.17506984e-02 -4.24056053e-01
-3.17071438e-01 5.10315120e-01 5.75130880e-02 -2.47007489e-01
9.52238739e-01 2.11894765e-01 1.67164251e-01 -3.38297635e-01
3.32396865e-01 6.20408833e-01 5.66824555e-01 1.10883825e-01
1.50794268e+00 6.24778032e-01 4.06450570e-01 -1.44367909e+00
-2.68678755e-01 -3.37424666e-01 -5.44170797e-01 -1.88951105e-01
5.60001135e-01 -9.65047657e-01 -4.77105707e-01 2.83181518e-01
-1.02996635e+00 2.34749466e-01 -3.86786669e-01 8.86091053e-01
-4.53460157e-01 1.24190524e-01 -4.10013311e-02 -9.99811530e-01
-4.44829434e-01 -6.20964289e-01 8.05218995e-01 2.78938890e-01
-7.00036436e-02 -4.63792801e-01 -1.65079549e-01 -1.86428979e-01
8.16569090e-01 4.20029223e-01 8.41989160e-01 -3.38737130e-01
-9.21474338e-01 -5.04380524e-01 -3.01782638e-01 1.91361606e-01
-6.33924529e-02 -2.14251429e-01 -1.16558218e+00 -4.43558007e-01
2.03596339e-01 -3.18528037e-04 1.06764734e+00 4.10040647e-01
2.54468679e-01 1.47571057e-01 -2.56880552e-01 7.24733353e-01
1.67711663e+00 2.49952435e-01 7.30764925e-01 4.34755087e-01
1.40699148e-01 5.18627226e-01 6.87182009e-01 3.08874995e-01
-4.10364181e-01 6.01935029e-01 6.46403134e-01 4.40312028e-02
-1.51586458e-01 3.29726934e-01 4.87997979e-01 5.24293780e-02
-2.34288260e-01 2.89665610e-02 -8.22585702e-01 3.53752732e-01
-1.08227384e+00 -1.02908242e+00 -3.07024866e-01 2.24637794e+00
1.97961211e-01 5.05080462e-01 -2.67190367e-01 4.73306209e-01
2.03054994e-01 1.90855846e-01 -3.09969246e-01 -3.61434400e-01
1.50726829e-03 9.53911781e-01 9.16112065e-01 2.94962913e-01
-1.09782624e+00 6.24088168e-01 5.60627985e+00 1.77356616e-01
-1.66915488e+00 -8.39359611e-02 -5.10716319e-01 -2.04852760e-01
-6.66142106e-02 -1.44653618e-01 -7.03218818e-01 1.65415809e-01
1.26567054e+00 8.63237754e-02 1.19037241e-01 6.13475025e-01
4.78709824e-02 -1.36995897e-01 -9.75485742e-01 9.84556437e-01
-2.51211673e-02 -1.02778518e+00 5.12527898e-02 1.26559794e-01
-9.28335562e-02 3.99720907e-01 1.33892879e-01 3.81953150e-01
-4.23960184e-04 -1.13400185e+00 6.63819909e-01 6.06746197e-01
8.39662552e-01 -6.54883146e-01 7.95083225e-01 3.13386977e-01
-9.01662171e-01 -6.97669089e-02 -5.74219704e-01 -4.19232070e-01
3.04520130e-01 8.73994172e-01 -1.06886351e+00 6.84754133e-01
6.67169333e-01 2.23071098e-01 -3.07913274e-01 1.07302511e+00
-1.40820578e-01 4.86559510e-01 -9.11405742e-01 -1.05161957e-01
3.17655861e-01 -3.00918460e-01 8.64998996e-01 1.36910200e+00
8.64412069e-01 -7.33144805e-02 -1.59212306e-01 7.67801344e-01
4.26043242e-01 -4.45777029e-01 -8.62023592e-01 8.80879983e-02
3.14488232e-01 1.36988270e+00 -5.60702205e-01 -6.66587278e-02
-5.67244649e-01 3.91420186e-01 -5.18229783e-01 2.19032392e-01
-7.91170001e-01 -7.10218668e-01 7.84097731e-01 4.70769852e-01
6.48324728e-01 -7.11969078e-01 -1.46969810e-01 -6.02946281e-01
2.38608625e-02 -5.65171540e-01 -5.02956100e-02 -6.64689124e-01
-1.12935817e+00 6.09791696e-01 7.85759315e-02 -1.66525626e+00
-3.18507284e-01 -9.21687722e-01 -4.83276129e-01 1.03788579e+00
-1.92565894e+00 -1.16041028e+00 -4.86097217e-01 4.90233958e-01
2.04583079e-01 -3.59995753e-01 8.79865527e-01 4.63838458e-01
2.44226053e-01 9.17655602e-02 -1.28783137e-01 -4.73082960e-02
6.97134197e-01 -1.08139765e+00 4.50630337e-01 9.34761643e-01
3.60056400e-01 3.07275534e-01 1.07625520e+00 -4.26725388e-01
-1.77748120e+00 -1.07801831e+00 4.86532748e-01 4.71183285e-03
6.60010338e-01 -4.61763263e-01 -7.77210176e-01 4.02893394e-01
-3.83589789e-02 -1.47039015e-02 3.97593588e-01 -5.93931787e-02
-3.08638662e-01 -4.00572658e-01 -1.02606559e+00 2.44301498e-01
6.20130360e-01 -3.48973244e-01 -9.54656363e-01 4.09908686e-03
2.64593393e-01 -1.91766813e-01 -4.14556891e-01 5.78668118e-01
5.41382730e-01 -1.02910471e+00 1.17108989e+00 -1.01068176e-01
1.72102032e-03 -6.33159935e-01 -2.44132593e-01 -1.16891921e+00
-4.46656734e-01 -3.05918247e-01 -1.13805784e-02 9.60185349e-01
3.14960301e-01 -1.01988065e+00 7.72309601e-01 -2.98001289e-01
-1.06102377e-01 -5.46525680e-02 -1.34191561e+00 -9.50676501e-01
-1.55718744e-01 -5.04175305e-01 3.44978333e-01 5.67284405e-01
-5.47943950e-01 6.34871960e-01 -8.66839066e-02 6.52550459e-01
8.17150891e-01 4.71457630e-01 8.77361000e-01 -1.97111607e+00
-2.42734075e-01 -3.64204198e-01 -8.05872262e-01 -7.92232513e-01
-3.58346134e-01 -7.87035048e-01 8.96284133e-02 -1.37534440e+00
-8.49727869e-01 -4.84858900e-01 -6.54526651e-02 1.46179914e-01
4.91711199e-01 4.03939575e-01 3.22626978e-01 1.73010409e-01
2.12537467e-01 4.81568038e-01 6.56518340e-01 -2.32790813e-01
-1.68145642e-01 2.19381928e-01 -3.00497323e-01 8.18131566e-01
7.49030054e-01 -5.53147137e-01 -1.02442406e-01 -4.56950068e-01
3.34189802e-01 -7.20995888e-02 7.52606332e-01 -1.80096042e+00
3.59117866e-01 2.68669397e-01 8.50874603e-01 -8.82188976e-01
7.69978344e-01 -1.25628495e+00 2.97813267e-01 7.60512590e-01
3.81579697e-01 -1.76370472e-01 2.97640651e-01 5.57878852e-01
-3.36765230e-01 -1.47004768e-01 9.05905128e-01 -5.77092171e-02
-7.77480185e-01 -1.39518321e-01 -9.30298328e-01 -6.12409651e-01
8.04628015e-01 -6.13780618e-01 -2.67276704e-01 -3.18619072e-01
-3.09752405e-01 -7.17626959e-02 4.35318500e-02 3.24922800e-01
7.11927831e-01 -7.09750295e-01 -9.21785593e-01 4.88638997e-01
-5.89106493e-02 -7.69795626e-02 -2.04794314e-02 8.81111145e-01
-7.26965666e-01 5.85692167e-01 -6.94278717e-01 -7.69755960e-01
-1.05670714e+00 3.32799554e-01 3.49984109e-01 8.79932661e-03
-1.01716018e+00 5.33627629e-01 -9.72862318e-02 -1.57738388e-01
-1.18189827e-01 -5.02196908e-01 -1.99546352e-01 5.71881771e-01
4.87161011e-01 1.69758335e-01 4.39499974e-01 -2.81314164e-01
-5.99145234e-01 7.58425236e-01 1.04103029e-01 -3.46348435e-01
1.45667839e+00 3.35657239e-01 2.12104172e-01 2.74742126e-01
6.37878418e-01 -3.14201862e-02 -1.06014907e+00 -9.66955647e-02
-3.61209805e-03 -3.22215348e-01 4.97400373e-01 -6.95412099e-01
-8.70269358e-01 1.20129478e+00 1.13415146e+00 3.37124437e-01
1.18087482e+00 -3.62997204e-01 4.74798113e-01 8.08420479e-01
5.76700985e-01 -7.15981722e-01 -4.40562278e-01 6.40166640e-01
7.96418071e-01 -6.79788351e-01 4.05176491e-01 -1.69824272e-01
-5.79566099e-02 1.37616777e+00 4.97790463e-02 -4.56278473e-01
6.07489944e-01 8.64765346e-01 3.49040598e-01 -4.20044512e-01
-4.38788861e-01 -6.52439177e-01 -1.13882288e-01 8.33314478e-01
2.55534321e-01 2.19311919e-02 -1.90661445e-01 1.72689587e-01
-6.32542789e-01 -4.23757508e-02 6.28315866e-01 9.42250431e-01
-7.53978372e-01 -8.18892717e-01 -8.39123845e-01 7.00901270e-01
-2.58481324e-01 -9.82641056e-02 -2.03051433e-01 1.03533745e+00
2.39264339e-01 8.06666911e-01 3.75850528e-01 -2.15288222e-01
5.80087006e-01 1.86201289e-01 5.98281443e-01 -4.56305861e-01
-5.72980225e-01 6.62532300e-02 -2.58862451e-02 -2.20549583e-01
-5.04448056e-01 -6.27729714e-01 -1.04691875e+00 9.78722796e-02
-4.11034495e-01 -1.88848615e-01 9.37210679e-01 7.43674994e-01
6.03379346e-02 7.58372366e-01 1.88408405e-01 -1.15697718e+00
-7.82735109e-01 -1.12451959e+00 -6.46785915e-01 -2.36222923e-01
6.54031813e-01 -7.97016084e-01 -4.69570577e-01 -5.92369996e-02] | [7.776658058166504, -1.2906707525253296] |
968a981e-71ae-4a00-aa3e-66d902efa8bd | safety-and-performance-why-not-both-bi | 2208.05969 | null | https://arxiv.org/abs/2208.05969v2 | https://arxiv.org/pdf/2208.05969v2.pdf | Safety and Performance, Why not Both? Bi-Objective Optimized Model Compression toward AI Software Deployment | The size of deep learning models in artificial intelligence (AI) software is increasing rapidly, which hinders the large-scale deployment on resource-restricted devices (e.g., smartphones). To mitigate this issue, AI software compression plays a crucial role, which aims to compress model size while keeping high performance. However, the intrinsic defects in the big model may be inherited by the compressed one. Such defects may be easily leveraged by attackers, since the compressed models are usually deployed in a large number of devices without adequate protection. In this paper, we try to address the safe model compression problem from a safety-performance co-optimization perspective. Specifically, inspired by the test-driven development (TDD) paradigm in software engineering, we propose a test-driven sparse training framework called SafeCompress. By simulating the attack mechanism as the safety test, SafeCompress can automatically compress a big model to a small one following the dynamic sparse training paradigm. Further, considering a representative attack, i.e., membership inference attack (MIA), we develop a concrete safe model compression mechanism, called MIA-SafeCompress. Extensive experiments are conducted to evaluate MIA-SafeCompress on five datasets for both computer vision and natural language processing tasks. The results verify the effectiveness and generalization of our method. We also discuss how to adapt SafeCompress to other attacks besides MIA, demonstrating the flexibility of SafeCompress. | ['Xiao Han', 'Leye Wang', 'Jie Zhu'] | 2022-08-11 | null | null | null | null | ['inference-attack', 'membership-inference-attack'] | ['adversarial', 'computer-vision'] | [ 2.73227721e-01 -1.49359424e-02 -2.06851110e-01 -1.31969929e-01
-3.84901017e-01 -2.51714498e-01 2.55229563e-01 4.85444553e-02
-4.37918194e-02 1.78950891e-01 -3.25671524e-01 -6.40048087e-01
-1.82386756e-01 -1.00426424e+00 -1.09365249e+00 -5.02931476e-01
-1.80557802e-01 7.01829940e-02 2.45919570e-01 1.52223691e-01
1.48163140e-01 2.36017063e-01 -1.59583330e+00 1.52326062e-01
1.18526530e+00 1.04165876e+00 2.06621155e-01 2.64262110e-01
1.57783896e-01 6.50753438e-01 -8.77083123e-01 -6.11933351e-01
3.67636859e-01 -1.66057363e-01 -5.02234340e-01 5.33855446e-02
4.44703661e-02 -4.80757952e-01 -3.20295900e-01 1.48351657e+00
1.03839286e-01 -4.06056792e-01 1.26920462e-01 -1.86482239e+00
-9.57227200e-02 1.03923738e+00 -5.63909948e-01 -1.83911860e-01
3.93611416e-02 2.71959811e-01 4.79386061e-01 -2.86641657e-01
9.61373746e-02 9.54168737e-01 4.06419367e-01 4.40633982e-01
-8.94334495e-01 -1.01719010e+00 1.24185011e-01 1.61680475e-01
-1.68007207e+00 -3.73152077e-01 8.96676660e-01 -2.87962556e-01
7.86500990e-01 4.88270462e-01 5.81268132e-01 1.09993494e+00
2.46923432e-01 8.24412346e-01 7.73949862e-01 -2.22771287e-01
6.93978786e-01 7.94225484e-02 2.18320966e-01 5.95284760e-01
8.48738313e-01 1.03596509e-01 -2.31825978e-01 -7.06254244e-01
2.63321787e-01 3.07746857e-01 -2.39161700e-01 -2.80933112e-01
-8.64836097e-01 7.50617266e-01 7.77115971e-02 1.36913151e-01
-2.40287811e-01 7.65389279e-02 7.32184172e-01 1.78935662e-01
2.53990702e-02 2.37246305e-01 -5.32888174e-01 -3.60417664e-01
-9.86365318e-01 1.27202883e-01 9.41428304e-01 1.19622838e+00
4.20539856e-01 4.48383570e-01 2.86153078e-01 3.28137428e-01
3.89781743e-01 4.25446033e-01 5.88328362e-01 -6.51752472e-01
4.91784930e-01 6.06403470e-01 -4.83661592e-01 -1.20231807e+00
1.09898098e-01 -5.56345165e-01 -1.06877291e+00 -3.84295397e-02
-1.27556145e-01 -1.20393567e-01 -4.61888552e-01 1.80207360e+00
2.58649737e-01 7.89230168e-01 1.30934283e-01 4.33780253e-01
2.13450700e-01 5.26340127e-01 -2.57174708e-02 -4.36026424e-01
1.25270867e+00 -6.60056531e-01 -3.67511779e-01 -2.67191529e-01
9.60143447e-01 -1.98710725e-01 1.26767516e+00 8.10973525e-01
-8.50323260e-01 -3.67325574e-01 -1.65160275e+00 5.36825478e-01
-1.30010813e-01 -1.26009196e-01 6.47203505e-01 1.01174581e+00
-5.22061348e-01 4.28097099e-01 -1.20805955e+00 1.92768425e-01
6.43702686e-01 3.33908021e-01 -1.56076059e-01 -1.45208761e-01
-1.03372371e+00 2.56217241e-01 5.27280092e-01 -1.99770615e-01
-1.20484328e+00 -7.77359068e-01 -1.06959927e+00 3.41919452e-01
6.32835388e-01 -5.46967328e-01 1.11547565e+00 -4.59293813e-01
-1.22475636e+00 3.03740352e-01 1.90672159e-01 -9.98097003e-01
2.96912640e-01 -2.08202496e-01 -5.84499955e-01 1.32236689e-01
-2.94017673e-01 5.72057813e-02 1.12862873e+00 -1.18931377e+00
-4.01247382e-01 -3.80202740e-01 3.32627237e-01 -4.85889465e-01
-9.74698961e-01 2.79517937e-02 -2.44989395e-01 -6.95542574e-01
-2.66672015e-01 -9.33637679e-01 -2.16083750e-01 -3.47583354e-01
-6.93346560e-01 1.03382118e-01 1.28819048e+00 -4.15102482e-01
1.71817017e+00 -2.42795944e+00 1.63199641e-02 4.37076032e-01
3.91123235e-01 6.59672439e-01 1.89968990e-03 3.97488177e-02
-1.34459580e-03 4.00979549e-01 -6.47430837e-01 -2.64313459e-01
3.13513763e-02 2.98250705e-01 -5.72674274e-01 3.17722738e-01
5.44047169e-02 6.07022643e-01 -5.85520566e-01 -5.00604033e-01
-4.98509556e-02 3.01915377e-01 -8.36261332e-01 1.38580829e-01
-4.39337313e-01 -1.71823546e-01 -6.55230343e-01 8.99410129e-01
8.83586347e-01 -2.69611984e-01 2.43414119e-01 -1.72527447e-01
2.46494800e-01 1.94555357e-01 -1.10028219e+00 1.46074402e+00
-6.16845846e-01 9.11985487e-02 1.36014819e-01 -1.10723436e+00
9.40027773e-01 1.45201325e-01 2.29680032e-01 -4.34985787e-01
2.79390872e-01 3.19246829e-01 2.06250101e-01 -3.22585225e-01
1.76024899e-01 1.99399024e-01 -3.51126701e-01 5.86317897e-01
-4.09924418e-01 1.00535221e-01 -8.91700834e-02 3.30506504e-01
1.32092702e+00 -1.97160810e-01 3.12148094e-01 -5.92733212e-02
6.60319269e-01 -2.24456638e-01 8.63416612e-01 5.93294024e-01
-1.79790687e-02 3.05119812e-01 6.67386711e-01 -1.71455950e-01
-7.82267988e-01 -5.51568866e-01 -2.89874803e-02 5.54149508e-01
6.69429079e-02 -1.07427108e+00 -1.11438739e+00 -1.04565620e+00
-1.03344381e-01 8.77652407e-01 -1.94776848e-01 -9.41928506e-01
-5.96061647e-01 -5.27065217e-01 9.13407385e-01 4.70399141e-01
7.66750693e-01 -7.77819812e-01 -8.92850935e-01 -6.44993261e-02
1.39126360e-01 -1.20722473e+00 -3.98060262e-01 -2.19499692e-03
-8.27941179e-01 -1.07164681e+00 -3.02063022e-02 -5.33145666e-01
5.70285380e-01 3.87708187e-01 7.84438431e-01 4.41468537e-01
-3.29796597e-02 1.01368658e-01 -4.17433590e-01 -5.50083399e-01
-7.38909125e-01 2.77080797e-02 3.63943309e-01 2.21378729e-02
5.46331048e-01 -8.36302042e-01 -1.76203877e-01 1.66175380e-01
-1.45912588e+00 7.71721005e-02 5.86328030e-01 6.50921583e-01
6.13172174e-01 7.75519252e-01 6.19898260e-01 -8.68915141e-01
6.53094590e-01 -6.16270125e-01 -9.41769421e-01 2.97849000e-01
-8.04129720e-01 -1.94052935e-01 1.15975416e+00 -8.09841216e-01
-4.88257080e-01 -3.86809856e-02 -2.67468512e-01 -1.00455499e+00
-8.20429474e-02 9.05712545e-01 -9.97484028e-01 -1.02598645e-01
3.33150506e-01 4.77671862e-01 1.34573221e-01 -3.48715305e-01
-1.63904518e-01 8.92704546e-01 4.67757404e-01 -6.36891246e-01
1.02167833e+00 2.94292450e-01 5.19893877e-02 -6.41477823e-01
-2.82329977e-01 2.84633577e-01 1.62814125e-01 2.22202808e-01
3.03325117e-01 -7.55510747e-01 -8.31320226e-01 6.45918787e-01
-9.48397815e-01 -2.06964239e-01 -2.12888509e-01 3.81590426e-01
-3.63089532e-01 6.08196139e-01 -3.83603990e-01 -6.42962277e-01
-5.85590482e-01 -1.61601925e+00 8.55136573e-01 1.12052739e-01
-3.89355198e-02 -4.65947747e-01 -3.72692108e-01 2.12751657e-01
2.81021982e-01 2.62352973e-01 1.19582975e+00 -1.20496023e+00
-4.96742547e-01 -4.45253223e-01 1.97578728e-01 7.28922248e-01
-9.98979285e-02 1.96047891e-02 -7.52296388e-01 -5.42156875e-01
7.55814254e-01 -2.08306894e-01 2.84070909e-01 6.55843839e-02
1.85632265e+00 -7.39551306e-01 -2.08011433e-01 8.83862376e-01
1.14962733e+00 4.36989129e-01 8.18694592e-01 3.14064413e-01
7.98344851e-01 2.39717245e-01 7.73222983e-01 7.07501292e-01
4.58812490e-02 5.80476105e-01 7.94082820e-01 4.20511842e-01
4.15441722e-01 -3.90137792e-01 5.52859962e-01 9.63935912e-01
4.74008083e-01 -5.96596375e-02 -8.72252345e-01 1.80213392e-01
-1.48780823e+00 -6.79542005e-01 2.01325595e-01 2.47157526e+00
8.73534977e-01 5.48360765e-01 2.91232280e-02 6.54545069e-01
5.88614881e-01 8.10933262e-02 -6.41543329e-01 -3.16308409e-01
1.88847259e-01 1.64510295e-01 2.31092021e-01 4.06448320e-02
-8.97547781e-01 5.51135302e-01 4.64704180e+00 1.36434698e+00
-1.31933200e+00 1.43602267e-01 6.64835811e-01 -5.09147644e-02
-4.17287171e-01 2.44069904e-01 -7.46422648e-01 9.36229825e-01
1.19889283e+00 -6.92073226e-01 3.32844615e-01 1.41741967e+00
-8.22932459e-03 2.83809513e-01 -1.12107825e+00 9.64675426e-01
1.09379455e-01 -1.26306438e+00 3.80389579e-02 4.17493284e-01
2.70111710e-01 -3.48678291e-01 1.29359931e-01 4.69903409e-01
-5.73768876e-02 -8.68819952e-01 6.69025660e-01 -1.37296081e-01
8.27943683e-01 -1.13990152e+00 6.34988725e-01 7.59090781e-01
-1.03254366e+00 -2.36272320e-01 -4.09374237e-01 1.33932129e-01
-6.44616559e-02 5.82729399e-01 -5.81086576e-01 5.35094202e-01
6.70150876e-01 6.14671648e-01 -7.31365502e-01 9.30777073e-01
-5.35289533e-02 9.33218479e-01 -3.97077531e-01 2.11480170e-01
-1.51605934e-01 -1.00694597e-01 6.69488192e-01 7.84151971e-01
4.93920922e-01 -1.80464819e-01 1.30021110e-01 8.75469625e-01
-2.38872886e-01 -2.76417613e-01 -7.20390320e-01 -2.31324077e-01
9.41627800e-01 1.08486283e+00 -4.08550680e-01 -2.83928722e-01
-3.51747781e-01 5.51107109e-01 -5.72662614e-02 -4.42850217e-03
-1.15773761e+00 -4.99991030e-01 6.57375395e-01 1.60512492e-01
1.43606141e-01 -1.19273685e-01 -3.87023002e-01 -1.18795538e+00
3.25578481e-01 -1.68762171e+00 3.20862085e-01 -3.25857013e-01
-1.00277340e+00 6.59025431e-01 2.33179614e-01 -1.52862346e+00
-2.44156092e-01 -1.10165514e-01 -7.46613562e-01 3.46821696e-01
-1.03467989e+00 -1.10104251e+00 -2.01450393e-01 6.12797439e-01
3.58331710e-01 -4.67800200e-01 7.14030087e-01 3.83931398e-01
-8.87857378e-01 1.15529156e+00 -3.29513252e-01 -1.55278131e-01
1.43351242e-01 -6.25426590e-01 4.19213027e-01 1.30507386e+00
6.99587241e-02 1.02476239e+00 6.25818133e-01 -7.44620979e-01
-1.88519561e+00 -1.26390433e+00 3.17781925e-01 2.08688036e-01
6.40182316e-01 -5.37187159e-01 -1.29619908e+00 6.58312440e-01
-1.48613110e-01 2.75603030e-02 6.54125214e-01 -3.58905137e-01
-6.87988758e-01 -3.04472536e-01 -1.37116337e+00 6.81670964e-01
7.95292199e-01 -5.76144278e-01 -3.26136887e-01 2.40401700e-01
1.36722744e+00 -2.26838306e-01 -8.29116523e-01 6.56283140e-01
2.16276094e-01 -9.12976027e-01 8.85781646e-01 -4.14569855e-01
4.34481412e-01 -2.76669383e-01 -3.54634672e-01 -8.82850409e-01
2.52311945e-01 -8.33570898e-01 -9.93743598e-01 1.34153283e+00
-5.40225804e-02 -8.03376377e-01 9.15190935e-01 7.50159442e-01
-2.29230419e-01 -9.20486927e-01 -9.78307545e-01 -1.15006602e+00
-8.32057372e-02 -6.97124720e-01 1.32452118e+00 9.89901483e-01
-1.41840074e-02 -1.38783380e-01 -4.48924989e-01 4.12200123e-01
7.01100230e-01 -3.44516039e-02 8.65287423e-01 -1.05572033e+00
-8.43578279e-01 -3.41244459e-01 -4.95125115e-01 -7.47216761e-01
3.33739579e-01 -7.00196266e-01 -3.70467514e-01 -6.02932155e-01
1.28844187e-01 -5.12530506e-01 -2.72236943e-01 4.94403839e-01
1.25457067e-02 -1.07538812e-01 3.69174778e-01 2.25205272e-01
-4.83919263e-01 6.21996164e-01 6.21508241e-01 -3.07906389e-01
-2.03605086e-01 1.63568914e-01 -9.65611994e-01 7.96107233e-01
9.19282019e-01 -6.52619362e-01 -7.58389115e-01 -2.79371977e-01
1.53931573e-01 7.69220591e-02 2.73447067e-01 -1.28761506e+00
2.84719557e-01 -1.45919979e-01 -4.28565472e-01 -3.48674744e-01
-2.02919021e-02 -1.22164416e+00 3.70854318e-01 8.80433083e-01
-1.61155820e-01 1.04926556e-01 1.96741194e-01 5.00084102e-01
-3.36880475e-01 -4.33004022e-01 5.34129918e-01 3.37634295e-01
-3.77086550e-01 4.88322377e-01 -1.17711350e-01 -1.39100611e-01
1.36995554e+00 -2.11996734e-01 -5.93667030e-01 -3.54969390e-02
-2.00984836e-01 8.11777636e-02 6.65448189e-01 3.63457441e-01
9.22381103e-01 -1.20835650e+00 -4.08471704e-01 7.35716283e-01
2.47245893e-01 3.25426549e-01 3.11627656e-01 9.01715875e-01
-3.47442359e-01 1.45364344e-01 6.79591373e-02 -4.16092932e-01
-1.44594252e+00 1.06284440e+00 5.96849993e-02 -2.33027771e-01
-6.22256160e-01 3.80691648e-01 1.89791963e-01 -1.30537912e-01
4.99166340e-01 -2.37391010e-01 1.05701484e-01 -5.17014205e-01
8.32277656e-01 3.17596525e-01 2.73396224e-01 -2.48447776e-01
-3.36537093e-01 1.92926847e-03 -3.02959561e-01 3.84477973e-01
1.22404361e+00 2.74075985e-01 -2.41523296e-01 -8.69640261e-02
1.19780374e+00 1.79055765e-01 -9.98333633e-01 -9.91273448e-02
-6.04345165e-02 -5.93755364e-01 1.10073730e-01 -2.31983230e-01
-1.34317470e+00 8.76124799e-01 3.69740307e-01 4.37659413e-01
1.37761474e+00 -2.81780303e-01 1.23487782e+00 4.48321462e-01
8.50921631e-01 -6.46647871e-01 9.82736144e-03 7.98234940e-02
5.24123192e-01 -7.75278747e-01 -7.32381493e-02 -4.02549267e-01
-5.59929788e-01 6.52193248e-01 8.97257745e-01 -4.12005521e-02
8.87977481e-01 7.74858832e-01 -5.54202020e-01 -1.21680964e-02
-6.95500970e-01 4.70296085e-01 -2.24667385e-01 6.65542424e-01
-3.03700268e-01 8.92530456e-02 -1.69086888e-01 1.30573869e+00
-1.46871030e-01 1.79382011e-01 7.39260375e-01 1.24892831e+00
-1.13018937e-01 -1.13682914e+00 -3.98070872e-01 6.29734993e-01
-4.84445423e-01 3.65151912e-02 -7.12846592e-02 5.22657096e-01
1.49153709e-01 9.77241576e-01 -1.47492990e-01 -1.12820423e+00
1.00123547e-01 -1.02084905e-01 -1.10011339e-01 -4.86616731e-01
-4.77354854e-01 6.85277060e-02 -1.77202851e-01 -6.28264010e-01
1.13000289e-01 -3.95721048e-01 -1.20405436e+00 -4.46523786e-01
-3.30212146e-01 1.50325879e-01 6.35181427e-01 7.63562441e-01
7.02328444e-01 5.97245753e-01 7.97441065e-01 -3.39417398e-01
-1.12580681e+00 -6.07502580e-01 -6.47853196e-01 1.48857012e-01
-1.59087237e-02 -4.84231472e-01 -5.08718312e-01 -1.86069846e-01] | [6.866187572479248, 7.761033535003662] |
2decfd4a-78ba-4586-8686-3ece19d5dcdd | simulation-based-bayesian-inference-for-multi | 2109.14275 | null | https://arxiv.org/abs/2109.14275v1 | https://arxiv.org/pdf/2109.14275v1.pdf | Simulation-based Bayesian inference for multi-fingered robotic grasping | Multi-fingered robotic grasping is an undeniable stepping stone to universal picking and dexterous manipulation. Yet, multi-fingered grippers remain challenging to control because of their rich nonsmooth contact dynamics or because of sensor noise. In this work, we aim to plan hand configurations by performing Bayesian posterior inference through the full stochastic forward simulation of the robot in its environment, hence robustly accounting for many of the uncertainties in the system. While previous methods either relied on simplified surrogates of the likelihood function or attempted to learn to directly predict maximum likelihood estimates, we bring a novel simulation-based approach for full Bayesian inference based on a deep neural network surrogate of the likelihood-to-evidence ratio. Hand configurations are found by directly optimizing through the resulting amortized and differentiable expression for the posterior. The geometry of the configuration space is accounted for by proposing a Riemannian manifold optimization procedure through the neural posterior. Simulation and physical benchmarks demonstrate the high success rate of the procedure. | ['Gilles Louppe', 'Olivier Brüls', 'Norman Marlier'] | 2021-09-29 | null | null | null | null | ['robotic-grasping'] | ['robots'] | [-3.88504043e-02 1.28804922e-01 1.98253274e-01 -2.31725261e-01
-7.10381448e-01 -6.39490783e-01 4.54791337e-01 -3.53646427e-01
-3.50493550e-01 7.69397497e-01 -3.42082530e-01 -6.53012320e-02
-8.04128230e-01 -3.91179621e-01 -9.80364561e-01 -9.17752206e-01
-1.85245380e-01 9.20470417e-01 -3.06850791e-01 4.80570868e-02
3.45533848e-01 6.73329413e-01 -1.31315327e+00 -4.92289335e-01
6.86928272e-01 8.28512192e-01 5.94683528e-01 6.72976136e-01
2.73059338e-01 2.84762293e-01 -1.05823688e-01 -1.99658453e-01
3.57954592e-01 -8.32909811e-03 -5.02460361e-01 -1.67062387e-01
-8.70494545e-02 -5.59217274e-01 -3.59513372e-01 1.17110300e+00
4.13480073e-01 3.71529162e-01 1.01754427e+00 -9.97133970e-01
-3.41020197e-01 5.58518589e-01 -1.98321268e-01 -5.21326363e-01
3.37983072e-01 4.55459744e-01 9.65982318e-01 -9.52309072e-01
8.03242981e-01 1.30057347e+00 3.89449626e-01 3.76161277e-01
-1.17621779e+00 -8.54400247e-02 1.91139281e-02 -7.79835209e-02
-1.40067399e+00 -3.25695388e-02 8.22112143e-01 -6.37919128e-01
5.71046710e-01 -1.06355190e-01 3.86194915e-01 1.35594010e+00
5.66828728e-01 6.01317644e-01 1.09328043e+00 -3.68347466e-01
4.59531993e-01 -1.00851253e-01 -1.70501217e-01 7.09767997e-01
4.28315401e-01 1.85405180e-01 -3.00543696e-01 -2.87129488e-02
1.01493609e+00 1.34478837e-01 -7.38590360e-02 -7.85949111e-01
-1.22816157e+00 5.14769912e-01 2.95407772e-01 -2.33897805e-01
-6.99098647e-01 4.16826248e-01 -3.94446515e-02 -1.63784623e-01
-1.10185735e-01 5.33420384e-01 -4.18333441e-01 -4.50729281e-01
-3.69421571e-01 8.07978272e-01 1.22347867e+00 1.13573456e+00
5.22510946e-01 -2.54691124e-01 -3.55316810e-02 5.08729339e-01
7.20795453e-01 8.78139496e-01 -4.35785711e-01 -1.28323388e+00
5.98201752e-01 2.27330282e-01 6.63896561e-01 -7.00674057e-01
-4.24844563e-01 -1.05982028e-01 -4.61815536e-01 7.54017711e-01
9.68556285e-01 -4.34234470e-01 -8.12109351e-01 1.70114958e+00
3.44027221e-01 -4.91715789e-01 -1.78088248e-01 1.25867343e+00
-2.84253955e-01 2.61575162e-01 -1.56878427e-01 8.31585526e-02
1.05723858e+00 -2.89989978e-01 -5.59771001e-01 1.83117446e-02
-1.51789576e-01 -5.92777371e-01 7.81509221e-01 7.19696403e-01
-1.16220236e+00 -1.90646704e-02 -1.06788492e+00 1.53242931e-01
-3.77548933e-02 1.59570441e-01 4.69063789e-01 2.98753560e-01
-4.28893328e-01 1.33098042e+00 -1.34245574e+00 3.18661593e-02
3.33693385e-01 3.96155149e-01 -1.50409967e-01 -1.58757731e-01
-6.16256416e-01 1.30210674e+00 3.78541559e-01 5.13695300e-01
-8.80092323e-01 -5.15003979e-01 -4.94890600e-01 -7.77058825e-02
5.93785822e-01 -6.71469688e-01 1.25648594e+00 2.19757613e-02
-2.09235501e+00 6.82020262e-02 1.21722579e-01 -6.73554372e-03
9.74661350e-01 -5.42812586e-01 4.83223110e-01 3.01877648e-01
-2.97324657e-01 3.26452374e-01 9.45069253e-01 -1.33488119e+00
2.26783663e-01 -4.14976865e-01 1.60731286e-01 1.44017756e-01
3.32206964e-01 -5.82061589e-01 -1.17979951e-01 -4.22243267e-01
4.24634010e-01 -1.22557938e+00 -4.30499315e-01 2.46804506e-01
-5.61559856e-01 -3.46458368e-02 4.51785266e-01 -9.27265406e-01
2.86212474e-01 -1.68291473e+00 7.43532240e-01 4.10297096e-01
-3.98417898e-02 -7.32265934e-02 1.19531795e-01 6.75312579e-01
5.63409925e-01 -7.54163861e-02 -2.16119871e-01 -2.26254836e-01
5.18987119e-01 1.94054767e-01 -4.18036908e-01 7.07274497e-01
4.89249676e-01 8.39909017e-01 -9.91036713e-01 -2.76251167e-01
4.31859553e-01 7.20130265e-01 -6.30989552e-01 2.22582206e-01
-6.23051286e-01 7.23804653e-01 -8.52416813e-01 6.46916211e-01
6.45636141e-01 1.12390690e-01 2.70416170e-01 -4.21270847e-01
-2.11244941e-01 4.06129025e-02 -1.39047027e+00 1.78623366e+00
-5.57421565e-01 -1.22854091e-01 5.29098511e-01 -7.29140878e-01
8.14901888e-01 3.12754139e-02 5.52209973e-01 -2.52999738e-02
5.00772536e-01 3.72763425e-01 2.78727561e-02 -6.38767004e-01
3.68209839e-01 -3.63268226e-01 -5.86834922e-02 2.45626867e-01
2.02184826e-01 -7.02133715e-01 -1.61484823e-01 -1.03444017e-01
9.97299910e-01 1.04609120e+00 -2.12982789e-01 -2.77049720e-01
-8.71022884e-03 -1.14782475e-01 1.73093364e-01 7.84215212e-01
1.45658419e-01 7.63415635e-01 5.28866410e-01 1.20997138e-01
-1.22282135e+00 -1.47322178e+00 -3.39783520e-01 1.19917288e-01
1.88934460e-01 8.63548145e-02 -8.35572422e-01 -1.21672086e-01
4.35787797e-01 6.91491187e-01 -3.60027820e-01 -9.31519717e-02
-5.61918199e-01 -5.79671741e-01 -3.26144206e-03 3.81473184e-01
1.14363782e-01 -8.20330739e-01 -8.90171051e-01 4.60997313e-01
1.59166425e-01 -1.09007168e+00 3.16940285e-02 1.80730104e-01
-7.46067584e-01 -1.02151799e+00 -7.35547543e-01 7.26146102e-02
4.99328911e-01 -3.11731219e-01 3.37191790e-01 -5.26621163e-01
-7.90845573e-01 7.65766084e-01 -1.37491718e-01 -3.61215740e-01
-4.20171618e-01 -1.73509449e-01 4.94518250e-01 -2.32393697e-01
-1.48176044e-01 -8.03766847e-01 -5.86680710e-01 2.35777348e-01
-4.15529191e-01 -1.44933254e-01 7.38761544e-01 7.96048880e-01
6.19477630e-01 -2.06772849e-01 4.19640243e-01 1.81113482e-02
4.97401476e-01 -4.84550476e-01 -9.74248886e-01 1.72675818e-01
-3.90835643e-01 5.67261100e-01 4.45703775e-01 -6.51510715e-01
-1.15327942e+00 2.97799945e-01 4.08117510e-02 -6.53137445e-01
-3.33604217e-02 4.18080866e-01 -7.53492415e-02 -1.00696273e-01
1.56845897e-01 -2.46602669e-01 4.16065514e-01 -7.28228569e-01
4.19370592e-01 4.33809876e-01 4.97522980e-01 -1.31195772e+00
5.85526884e-01 3.93254340e-01 6.51887119e-01 -6.44459784e-01
-4.11830127e-01 1.30959839e-01 -8.97616565e-01 -5.06258070e-01
8.76601636e-01 -2.93585390e-01 -1.53169692e+00 4.91649449e-01
-1.33720410e+00 -3.75226468e-01 -2.54060209e-01 9.13341403e-01
-1.16652060e+00 4.78345513e-01 -4.25469369e-01 -1.53831840e+00
-2.13744938e-01 -1.31943381e+00 1.09647739e+00 9.01547372e-02
-1.14734016e-01 -4.99174565e-01 -1.41352341e-01 -1.37132064e-01
3.90026987e-01 6.14779055e-01 6.92539513e-01 -1.79717764e-01
-9.13224280e-01 -6.20588064e-01 -1.51003793e-01 3.23534429e-01
-5.14617413e-02 9.82016921e-02 -7.48785257e-01 -2.65684664e-01
1.62227407e-01 -9.66987014e-02 5.27491450e-01 3.94958198e-01
1.13915980e+00 -1.80764526e-01 -3.70223880e-01 3.35603923e-01
1.61142969e+00 -6.00859150e-02 1.36865646e-01 -1.54381275e-01
7.00406492e-01 7.39103556e-01 8.16928566e-01 8.70889485e-01
8.17436650e-02 6.67071164e-01 9.16382730e-01 9.47660983e-01
2.22128332e-01 -3.78024369e-01 2.16854155e-01 5.60727239e-01
-6.04277134e-01 1.33605218e-02 -9.47844028e-01 2.13403016e-01
-1.84960616e+00 -5.48514307e-01 7.58055300e-02 2.54543614e+00
5.43574274e-01 1.77938879e-01 -2.24364638e-01 -2.07304046e-01
4.90549207e-01 -4.19259757e-01 -9.55678344e-01 2.33415868e-02
2.39787117e-01 1.61558136e-01 7.43198395e-01 6.05125129e-01
-7.61146367e-01 5.66588640e-01 6.07937574e+00 5.68672001e-01
-8.47106159e-01 -1.90441325e-01 -6.14273027e-02 -2.46577203e-01
8.31763223e-02 1.30725458e-01 -8.99098635e-01 3.61862868e-01
6.98280990e-01 4.32889722e-02 9.74954069e-01 9.10958767e-01
7.56846890e-02 -4.00579572e-01 -1.26816583e+00 6.99887395e-01
-3.70030493e-01 -1.03892553e+00 -1.53855279e-01 2.70459890e-01
3.35384697e-01 2.29724973e-01 -7.30972737e-02 -1.01394542e-01
2.94646293e-01 -8.09369087e-01 1.18033755e+00 1.25431883e+00
6.28938079e-01 -4.62613016e-01 4.18505788e-01 5.24895847e-01
-6.67425096e-01 -2.96177387e-01 -3.01133245e-01 -2.65769601e-01
5.74809313e-01 5.87938130e-01 -8.21826518e-01 4.32645708e-01
4.42053497e-01 3.02795053e-01 1.80501059e-01 9.05937791e-01
-9.56113413e-02 1.85008407e-01 -7.68147528e-01 -6.41257107e-01
-7.52419159e-02 -6.49306118e-01 1.08364546e+00 7.33481407e-01
4.91816998e-01 3.41103552e-03 1.38513580e-01 1.50967824e+00
2.62620330e-01 -4.52232510e-01 -5.53505719e-01 -3.77284020e-01
3.68328243e-01 1.28070140e+00 -6.55172169e-01 2.96722412e-01
2.26000339e-01 7.04098642e-01 3.37373823e-01 4.94148672e-01
-5.55179894e-01 -4.56095785e-01 7.41273642e-01 -1.35069847e-01
4.45935786e-01 -1.03463304e+00 -4.38516200e-01 -1.07527649e+00
6.41020834e-01 -3.03503364e-01 -3.54383886e-01 -7.32705593e-01
-1.43260860e+00 2.13370137e-02 4.74852085e-01 -9.11350429e-01
-6.08177304e-01 -1.31593490e+00 -2.99878210e-01 1.21129382e+00
-1.04840803e+00 -9.12579894e-01 9.62628722e-02 2.35017926e-01
8.26430023e-02 1.20589770e-01 8.27712119e-01 -2.26673111e-01
-3.00369292e-01 1.62793119e-02 2.04271436e-01 -2.44950473e-01
3.33778530e-01 -1.35667384e+00 9.15201753e-02 4.73996967e-01
-5.94767690e-01 9.16083276e-01 1.17070794e+00 -7.52425313e-01
-2.57056069e+00 -3.78294379e-01 -2.97602871e-03 -6.30619764e-01
9.26782668e-01 -4.84589010e-01 -7.00857818e-01 4.32179540e-01
-4.36356038e-01 -1.18262142e-01 -3.24852109e-01 -7.40087628e-02
-7.80583471e-02 7.06464052e-02 -1.33187187e+00 6.13272369e-01
9.58455980e-01 -3.84159297e-01 -5.20227671e-01 3.78076673e-01
4.57219481e-01 -5.97672224e-01 -1.12043190e+00 4.94229645e-01
1.03309679e+00 -2.77047306e-01 1.00813341e+00 -4.69363481e-01
4.93311048e-01 -1.49870530e-01 -4.67440188e-01 -1.13789666e+00
3.14222425e-02 -9.13859904e-01 -3.02176476e-01 1.07012248e+00
6.99844956e-02 -6.52617812e-01 5.85378528e-01 8.50683451e-01
-8.82787257e-02 -9.09354210e-01 -1.30801105e+00 -9.06206667e-01
4.24152017e-01 -6.00226879e-01 3.19208086e-01 3.21131825e-01
9.55193564e-02 -2.56081194e-01 -2.67390728e-01 5.16352713e-01
1.25176215e+00 6.33230284e-02 6.33458257e-01 -1.08122659e+00
-6.60950840e-01 -4.51294065e-01 2.09450107e-02 -9.67642188e-01
9.28287730e-02 -7.38497257e-01 6.43384337e-01 -1.40368104e+00
3.04458234e-02 -5.83514810e-01 1.83596775e-01 4.45301384e-02
7.84834027e-02 -3.91873181e-01 2.15050817e-01 -2.11187769e-02
-1.70219332e-01 5.59609354e-01 1.30345118e+00 1.17491968e-01
-4.36455719e-02 8.27737525e-02 -6.11597570e-05 7.04614222e-01
6.66614294e-01 -2.99290597e-01 5.60974814e-02 -3.38864088e-01
3.85541409e-01 5.36916554e-01 8.78295660e-01 -8.36613894e-01
2.51225084e-01 -3.94604236e-01 1.18868805e-01 -5.37085474e-01
8.03066552e-01 -8.15519631e-01 1.98617280e-01 3.71625900e-01
-2.14128241e-01 -3.98618579e-01 -4.32334542e-02 8.25313807e-01
3.10274452e-01 -5.74298739e-01 5.45462072e-01 -8.86737704e-02
-2.06100509e-01 2.33629301e-01 -2.74193019e-01 -2.66195953e-01
5.62318444e-01 9.10181627e-02 2.30042730e-03 -1.73326850e-01
-9.55660641e-01 7.78467059e-02 4.19465333e-01 2.81759441e-01
5.85449755e-01 -9.40765560e-01 -4.21752244e-01 5.33695929e-02
-3.62816036e-01 2.02856317e-01 2.14843825e-01 7.93201208e-01
-3.93134832e-01 1.44472435e-01 -3.29534650e-01 -6.89125776e-01
-4.59236413e-01 3.65134537e-01 3.93007904e-01 2.74602138e-03
-6.76738501e-01 6.28536820e-01 -4.05053139e-01 -6.90108418e-01
1.24118365e-01 -4.02354568e-01 5.03251791e-01 -3.85310650e-01
1.29703209e-01 6.60941780e-01 -1.79454863e-01 -2.76966900e-01
-2.72788644e-01 7.27139890e-01 2.87829489e-01 -5.28436959e-01
1.54425585e+00 -3.63316648e-02 -1.07281007e-01 6.03012443e-01
9.88302231e-01 -2.11421624e-01 -2.03715134e+00 1.09575503e-01
-4.46187034e-02 -3.72755319e-01 -8.63349438e-02 -9.01679873e-01
-4.11935925e-01 9.32608366e-01 2.96089083e-01 -2.01583132e-01
2.23829299e-01 9.68038384e-03 3.19857508e-01 9.79245126e-01
1.02641201e+00 -1.15185964e+00 -8.60105455e-02 6.19131327e-01
1.19994378e+00 -1.04815459e+00 6.71118349e-02 -3.81125480e-01
-2.47013420e-01 1.44045413e+00 7.90369809e-02 -5.41127741e-01
9.05489445e-01 5.26152074e-01 -3.64515573e-01 4.19232734e-02
-3.32486331e-01 3.14431220e-01 3.04842323e-01 4.78649884e-01
-5.88453189e-02 2.22140655e-01 -1.33822501e-01 6.94764256e-01
-1.64210215e-01 2.06095666e-01 3.24137688e-01 1.16211295e+00
-2.40470886e-01 -9.29896474e-01 -3.10817808e-01 3.62538785e-01
-3.86530817e-01 3.11496377e-01 1.61823630e-01 6.03965640e-01
-3.00579280e-01 5.96283436e-01 -2.67660290e-01 -2.62006491e-01
3.67080063e-01 1.68337524e-01 1.12719178e+00 -1.57510504e-01
7.29757324e-02 -9.47320089e-02 -1.12863243e-01 -7.80784011e-01
4.22223248e-02 -1.17654574e+00 -1.29631495e+00 6.11669086e-02
-3.32213968e-01 -2.41805077e-01 1.58878255e+00 1.22484863e+00
2.37416968e-01 3.62259835e-01 2.74790138e-01 -1.70509505e+00
-1.65365827e+00 -9.74543571e-01 -7.56918311e-01 1.69114217e-01
2.53172070e-01 -1.24814236e+00 -3.70265514e-01 -4.87388045e-01] | [5.658863067626953, -0.5981153845787048] |
a0ba26ef-dd4c-474a-bf8a-8445538b9702 | d-dpcc-deep-dynamic-point-cloud-compression | 2205.01135 | null | https://arxiv.org/abs/2205.01135v1 | https://arxiv.org/pdf/2205.01135v1.pdf | D-DPCC: Deep Dynamic Point Cloud Compression via 3D Motion Prediction | The non-uniformly distributed nature of the 3D dynamic point cloud (DPC) brings significant challenges to its high-efficient inter-frame compression. This paper proposes a novel 3D sparse convolution-based Deep Dynamic Point Cloud Compression (D-DPCC) network to compensate and compress the DPC geometry with 3D motion estimation and motion compensation in the feature space. In the proposed D-DPCC network, we design a {\it Multi-scale Motion Fusion} (MMF) module to accurately estimate the 3D optical flow between the feature representations of adjacent point cloud frames. Specifically, we utilize a 3D sparse convolution-based encoder to obtain the latent representation for motion estimation in the feature space and introduce the proposed MMF module for fused 3D motion embedding. Besides, for motion compensation, we propose a 3D {\it Adaptively Weighted Interpolation} (3DAWI) algorithm with a penalty coefficient to adaptively decrease the impact of distant neighbors. We compress the motion embedding and the residual with a lossy autoencoder-based network. To our knowledge, this paper is the first work proposing an end-to-end deep dynamic point cloud compression framework. The experimental result shows that the proposed D-DPCC framework achieves an average 76\% BD-Rate (Bjontegaard Delta Rate) gains against state-of-the-art Video-based Point Cloud Compression (V-PCC) v13 in inter mode. | ['Dong Wang', 'Zhu Li', 'Yiling Xu', 'Linyao Gao', 'Tingyu Fan'] | 2022-05-02 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [-2.78862923e-01 -4.00718898e-01 1.53669650e-02 1.64793665e-03
-3.66365731e-01 -8.47681612e-02 3.94119263e-01 -3.65665585e-01
-3.62055004e-01 1.26246527e-01 4.34474766e-01 -3.21870968e-02
-3.24041843e-02 -7.51882672e-01 -1.05252659e+00 -5.97234964e-01
-3.36958766e-01 -1.76414344e-02 2.10633084e-01 -9.80168805e-02
2.68528074e-01 7.69546509e-01 -1.47357380e+00 3.90965551e-01
5.77168643e-01 1.31861424e+00 5.88413656e-01 8.35457325e-01
-5.39711528e-02 7.01190472e-01 -2.54660219e-01 -2.53520072e-01
7.13988304e-01 -1.77989766e-01 -5.65419734e-01 1.84618950e-01
5.79865754e-01 -1.06701541e+00 -9.73017454e-01 7.19846249e-01
5.28891921e-01 2.36569017e-01 4.14686590e-01 -1.12101483e+00
-6.25655234e-01 -1.53565913e-01 -7.76196778e-01 3.48006338e-01
1.64191142e-01 2.03382939e-01 7.07651734e-01 -1.30555201e+00
7.22186565e-01 1.30435824e+00 8.30483675e-01 2.72781104e-01
-7.96264946e-01 -4.75227833e-01 -1.96012512e-01 5.02023399e-01
-1.61328828e+00 -1.93845093e-01 1.01933050e+00 -2.30978101e-01
1.25413203e+00 1.54073939e-01 9.03865457e-01 4.05665666e-01
3.74902189e-01 5.46500802e-01 3.38253200e-01 -2.86911190e-01
1.12021558e-01 -5.43940008e-01 -4.78866369e-01 7.32414842e-01
-4.37902138e-02 3.24844271e-01 -5.77190280e-01 7.41912704e-03
1.34769356e+00 4.26639259e-01 -4.61020917e-01 -2.41253704e-01
-1.14901209e+00 6.58221781e-01 6.13359630e-01 2.23844975e-01
-5.17397523e-01 7.18621433e-01 2.82520473e-01 2.12274536e-01
5.78953445e-01 -2.70414919e-01 -3.92185360e-01 -5.23190856e-01
-1.19382441e+00 4.72342879e-01 5.57586193e-01 1.21922982e+00
7.63622642e-01 2.19061404e-01 -6.35584909e-03 7.29742348e-01
4.77391571e-01 5.52218020e-01 5.14447093e-01 -1.82063723e+00
7.41418302e-01 3.17025483e-01 6.94959760e-02 -1.41205800e+00
2.20270142e-01 -3.44520807e-01 -1.01092327e+00 2.36500442e-01
-2.83941299e-01 1.28394857e-01 -4.66502815e-01 1.33072901e+00
4.44050282e-01 8.60834241e-01 3.34557444e-02 1.16508639e+00
6.70408309e-01 1.02894247e+00 -3.16708982e-01 -1.85125500e-01
8.47890019e-01 -9.73743439e-01 -5.71159542e-01 3.55777860e-01
6.10278308e-01 -8.83109689e-01 6.61251485e-01 -2.27953047e-02
-1.50344646e+00 -7.19691157e-01 -9.52360392e-01 -5.24751961e-01
2.13834673e-01 -1.07863702e-01 1.19882062e-01 4.55525815e-02
-1.12838638e+00 8.39171410e-01 -1.15163064e+00 2.76704073e-01
4.40707296e-01 2.65681952e-01 -3.14808398e-01 -5.34559488e-01
-8.96929502e-01 6.23655558e-01 5.75239994e-02 -2.00146753e-02
-6.99501932e-01 -1.07637572e+00 -8.51518631e-01 4.44048524e-01
-1.88489333e-01 -1.05184758e+00 8.11785460e-01 -5.81463635e-01
-1.24168682e+00 3.81934196e-01 -4.52708542e-01 -5.78381717e-01
3.17299813e-01 -4.05432999e-01 -7.82849342e-02 7.97662616e-01
-1.66858137e-02 9.00079012e-01 8.93861830e-01 -1.10140455e+00
-4.96459395e-01 -1.50723383e-01 -2.04181358e-01 4.12502825e-01
-1.92540407e-01 -1.75736636e-01 -5.43243051e-01 -1.03568542e+00
1.22142725e-01 -8.05108547e-01 -8.32566917e-02 7.25116849e-01
2.13723034e-01 1.23044841e-01 1.58235669e+00 -8.05991471e-01
1.16782117e+00 -2.46202636e+00 2.54021883e-01 -1.62818924e-01
4.45699066e-01 4.64165717e-01 -2.01874435e-01 2.84240782e-01
-4.82664211e-03 -1.92964122e-01 -4.08907264e-01 -8.49603534e-01
-2.04973206e-01 5.17261028e-01 -3.61772299e-01 5.49194455e-01
2.27454901e-01 8.02976906e-01 -7.72738993e-01 -3.90243411e-01
8.05134773e-01 1.23422039e+00 -1.18794835e+00 1.69416323e-01
1.94332093e-01 1.63924277e-01 -2.96628356e-01 4.95579511e-01
1.14038968e+00 -2.18989491e-01 -5.41163504e-01 -4.15130526e-01
-2.74856240e-01 -1.02142632e-01 -9.50263739e-01 2.14805055e+00
-5.91835201e-01 7.28243768e-01 2.48104230e-01 -7.74705410e-01
6.03193283e-01 4.01499599e-01 1.01989210e+00 -2.43911073e-01
1.57335728e-01 3.96019042e-01 -3.92081648e-01 -2.89751858e-01
7.38124430e-01 1.13028884e-01 4.17406529e-01 -1.22079603e-01
7.53508359e-02 -2.47107893e-01 -2.99339175e-01 2.29870260e-01
1.10180140e+00 1.61296353e-01 -1.65777847e-01 6.30333275e-02
7.47538328e-01 -1.28196046e-01 7.25400925e-01 2.15862598e-02
-3.36435109e-01 1.00955784e+00 -1.00956157e-01 -5.30129790e-01
-1.50556505e+00 -9.06382740e-01 4.79511032e-03 1.71652943e-01
3.21634501e-01 -5.21626532e-01 -6.02067351e-01 -7.44943991e-02
1.20330386e-01 4.72755939e-01 -8.93984437e-02 -2.27545291e-01
-9.93946731e-01 -4.17949539e-03 2.91789472e-01 5.32760143e-01
1.00640368e+00 -4.81279641e-01 -7.73007631e-01 4.09178883e-01
-4.26762223e-01 -1.41655850e+00 -8.59655857e-01 -3.49317551e-01
-1.25193501e+00 -7.12841928e-01 -9.63721871e-01 -8.82034600e-01
4.67341721e-01 8.15006912e-01 7.86761224e-01 5.49628995e-02
-2.80464422e-02 5.37307501e-01 -4.90443230e-01 2.06865072e-01
-1.38144821e-01 -4.84300166e-01 -7.16013461e-02 -1.14077695e-01
2.52094567e-01 -1.06326652e+00 -1.26353359e+00 2.05408156e-01
-1.02229071e+00 7.28425905e-02 4.38573629e-01 7.87435710e-01
9.54169333e-01 3.41411009e-02 -1.66138917e-01 9.44872126e-02
1.51451871e-01 -4.92370605e-01 -4.72421646e-01 -2.73945749e-01
-4.67325836e-01 -1.41629279e-01 6.22963369e-01 -2.09589377e-01
-7.24535167e-01 1.63359836e-01 -4.06677455e-01 -1.60622215e+00
2.46944070e-01 2.13542953e-01 1.20796956e-01 -5.07447362e-01
7.22588971e-02 5.95752239e-01 8.00780579e-02 -4.89155322e-01
4.53110993e-01 4.73104864e-01 7.83811510e-01 -2.22423449e-01
9.00929570e-01 7.93424070e-01 3.95903766e-01 -8.28188539e-01
-2.21508369e-01 -7.32371747e-01 -5.67736268e-01 -2.81852424e-01
1.03208160e+00 -1.38124442e+00 -6.61428154e-01 4.12706494e-01
-1.64640689e+00 -7.63334259e-02 -4.72916752e-01 8.12926292e-01
-9.40785110e-01 8.55997682e-01 -9.18824673e-01 -3.48132193e-01
-6.85366452e-01 -1.31154346e+00 1.37060809e+00 -3.20255756e-01
2.02498347e-01 -7.37426639e-01 -9.35044140e-03 2.91201383e-01
4.67235386e-01 3.51361394e-01 4.79199737e-01 3.58174384e-01
-1.12202692e+00 -1.47644147e-01 -2.56784230e-01 6.99782372e-01
-8.64585489e-02 -2.81486690e-01 -5.51547587e-01 -4.09207880e-01
3.69552940e-01 2.50925124e-01 7.65149295e-01 5.34216285e-01
1.36456466e+00 -6.01706386e-01 3.72008979e-02 1.41314411e+00
1.74375486e+00 -6.27134368e-02 7.91062355e-01 1.43509628e-02
1.02961624e+00 3.22080255e-02 4.19573069e-01 8.42190385e-01
4.55955595e-01 8.29451442e-01 7.32587218e-01 2.16081262e-01
-6.21316969e-01 -3.44254464e-01 4.73174900e-01 1.37194216e+00
-4.04227823e-01 -8.55629444e-02 -3.63894880e-01 5.93360305e-01
-1.68909001e+00 -1.12077153e+00 -1.31871119e-01 1.87242532e+00
5.43150067e-01 -3.42720926e-01 -4.27781552e-01 3.18135828e-01
5.81474960e-01 3.93260568e-01 -3.24586332e-01 -3.86025757e-01
-7.46020377e-02 3.24538887e-01 6.13761008e-01 6.57864630e-01
-7.83150673e-01 6.10371411e-01 4.38379002e+00 1.07085955e+00
-1.10549283e+00 4.00274217e-01 2.19036862e-01 -4.35229748e-01
-3.34648490e-01 -1.17442301e-02 -6.04276836e-01 7.48978555e-01
9.93265450e-01 5.40986983e-03 5.27828455e-01 8.36470544e-01
5.85880578e-01 2.60030687e-01 -6.37254000e-01 1.48123193e+00
3.51559371e-02 -1.89893055e+00 2.64015675e-01 4.42904204e-01
7.29999363e-01 2.58039415e-01 1.48654310e-03 -8.87822211e-02
-2.89782465e-01 -5.43698192e-01 9.39226329e-01 4.85657364e-01
1.04380238e+00 -1.00221550e+00 3.52039725e-01 2.81320006e-01
-1.47332060e+00 -1.38120949e-01 -6.40279651e-01 1.74134955e-01
5.02417982e-01 7.88656950e-01 -2.53552407e-01 6.09361589e-01
8.99936616e-01 1.13682842e+00 7.73325115e-02 9.43159342e-01
3.26814860e-01 3.62054110e-01 -3.88465375e-01 4.97400820e-01
4.17053580e-01 -1.37265608e-01 9.94928777e-01 9.66693103e-01
9.49870646e-01 5.12194633e-01 -2.58396089e-01 8.04226995e-01
-1.00406684e-01 -1.76256195e-01 -4.81315255e-01 3.91130716e-01
5.37268937e-01 7.61664152e-01 -6.70623183e-02 -3.27144802e-01
-6.26127362e-01 1.49067163e+00 -4.46110032e-02 1.95621982e-01
-6.62729859e-01 -2.78089404e-01 1.10209441e+00 3.95086646e-01
8.86708915e-01 -7.59174347e-01 -1.00965887e-01 -1.35396254e+00
3.03807914e-01 -2.49087840e-01 -8.92212540e-02 -9.73583639e-01
-8.60765517e-01 2.77343184e-01 -1.37058571e-02 -1.79221058e+00
-2.36993134e-01 -2.46550784e-01 -4.29670423e-01 9.27400351e-01
-1.93060219e+00 -9.46580350e-01 -4.73848403e-01 1.04268479e+00
7.38496661e-01 -1.05850697e-01 5.74653745e-01 7.67472565e-01
3.56562287e-02 4.33458298e-01 3.06363732e-01 -1.73982885e-02
3.68665695e-01 -5.21403551e-01 4.64942783e-01 9.49498415e-01
-3.19131404e-01 3.99865061e-01 2.43030310e-01 -5.99058092e-01
-1.77709842e+00 -1.56435740e+00 8.37481856e-01 1.50473788e-01
6.86604232e-02 4.65376414e-02 -9.42525446e-01 4.21110034e-01
1.74992606e-02 6.13669276e-01 3.80919456e-01 -1.15156591e+00
-1.77726939e-01 -2.18818665e-01 -1.39070320e+00 3.57703388e-01
1.19916177e+00 -5.71951032e-01 -3.06301355e-01 1.27371028e-01
1.39337933e+00 -5.45120060e-01 -1.30185664e+00 4.45862383e-01
4.29767787e-01 -1.03058374e+00 1.43325853e+00 1.96763203e-01
1.08134115e+00 -5.71280599e-01 -6.67252481e-01 -9.28988516e-01
-3.80009264e-01 -7.28755534e-01 -8.87222290e-01 7.89502919e-01
-4.82511550e-01 -1.37429714e-01 8.78578663e-01 4.74961102e-01
-7.52096474e-01 -9.53414917e-01 -1.32646251e+00 -5.91569901e-01
4.84830327e-02 -5.56572020e-01 6.03079319e-01 9.18945134e-01
-3.97224337e-01 -3.54750365e-01 -4.15691763e-01 2.39234373e-01
7.30192423e-01 -1.33448780e-01 5.74507117e-01 -5.39941788e-01
-4.16366637e-01 -1.91308618e-01 -9.31993783e-01 -1.72229445e+00
-5.34054562e-02 -8.19817722e-01 -3.22095186e-01 -1.19690394e+00
-3.66314948e-01 -2.03555331e-01 1.76594313e-02 -1.90203145e-01
1.47750214e-01 1.92281410e-01 4.88664180e-01 6.04165912e-01
-1.44209653e-01 1.17439377e+00 1.50287795e+00 -1.66649725e-02
-1.30112365e-01 -3.51540446e-01 -5.99994212e-02 3.28556269e-01
4.39553201e-01 -2.56221235e-01 -3.99948686e-01 -1.04004884e+00
-2.85489142e-01 5.00241339e-01 6.70736015e-01 -1.28336740e+00
3.79335582e-01 3.30405198e-02 4.67466712e-01 -1.07049334e+00
8.18526745e-01 -1.15380299e+00 1.39057770e-01 5.86234450e-01
4.61124666e-02 4.23628360e-01 -2.88089924e-02 8.65463257e-01
-4.57506031e-01 1.09710567e-01 7.75012612e-01 -2.01705188e-01
-7.75754869e-01 8.21726382e-01 -9.99774411e-02 -2.30053335e-01
8.92374873e-01 -4.65950102e-01 -4.96685393e-02 -3.05126250e-01
-4.40457463e-01 -4.10489812e-02 6.28672361e-01 1.59977615e-01
1.41204333e+00 -1.62261534e+00 -7.02921331e-01 3.34385514e-01
-2.79378682e-01 5.35871983e-01 6.02575362e-01 6.84934974e-01
-1.25871384e+00 4.05962467e-01 -1.79121420e-01 -7.40853667e-01
-1.02167523e+00 4.77264196e-01 2.07104877e-01 7.75263682e-02
-1.13509881e+00 7.74971008e-01 -1.21809177e-01 9.20763090e-02
1.38033763e-01 -4.50341821e-01 3.04450065e-01 -6.76780939e-01
5.03262162e-01 6.20176435e-01 1.85516514e-02 -9.31817710e-01
-6.33822232e-02 8.34741294e-01 2.51494735e-01 -4.84695323e-02
1.49571025e+00 -2.59634405e-01 -1.68310389e-01 -2.23582372e-01
1.98552680e+00 -3.52165073e-01 -1.72505653e+00 -1.45659238e-01
-8.32202315e-01 -1.11841297e+00 4.87765193e-01 4.81611118e-02
-1.47821987e+00 9.21594918e-01 7.05312073e-01 -5.16585410e-01
1.25026429e+00 -3.05368394e-01 1.70687246e+00 9.69881043e-02
3.30481976e-01 -7.50932992e-01 5.29512111e-03 5.79540133e-01
9.88923371e-01 -7.95694649e-01 2.77404636e-01 -5.49822569e-01
-3.17806453e-01 1.16921496e+00 3.09538096e-01 -6.20200813e-01
1.06441355e+00 -2.57908890e-04 -5.53200185e-01 -8.37703720e-02
-8.10086310e-01 2.44080812e-01 6.83184788e-02 5.84796309e-01
1.10083729e-01 -2.53051698e-01 -4.03175175e-01 -7.45770857e-02
-1.40814051e-01 2.15102196e-01 4.75578159e-01 9.29850757e-01
-3.33592117e-01 -9.33554590e-01 -2.74325252e-01 1.43623710e-01
-2.63068527e-01 -2.13144287e-01 3.86147261e-01 3.68600041e-01
2.92799681e-01 6.79581940e-01 4.22265053e-01 -6.92420840e-01
2.17683747e-01 -4.13745135e-01 4.11702305e-01 -5.75921908e-02
-3.65184605e-01 1.34667426e-01 -4.02612686e-01 -9.00933743e-01
-6.35668278e-01 -5.56479216e-01 -1.36844456e+00 -8.12983692e-01
1.28581434e-01 -1.43563613e-01 7.69034743e-01 4.08862174e-01
9.16805983e-01 2.81315029e-01 1.11635506e+00 -1.26706541e+00
-2.71403819e-01 -5.41931987e-01 -3.91245276e-01 4.06405389e-01
5.29895008e-01 -5.37270069e-01 -5.22780776e-01 1.86491281e-01] | [10.868632316589355, -1.7608287334442139] |
c0052fb4-f08a-49ba-b7e7-7a467d07908d | towards-poisoning-of-deep-learning-algorithms | 1708.08689 | null | http://arxiv.org/abs/1708.08689v1 | http://arxiv.org/pdf/1708.08689v1.pdf | Towards Poisoning of Deep Learning Algorithms with Back-gradient Optimization | A number of online services nowadays rely upon machine learning to extract
valuable information from data collected in the wild. This exposes learning
algorithms to the threat of data poisoning, i.e., a coordinate attack in which
a fraction of the training data is controlled by the attacker and manipulated
to subvert the learning process. To date, these attacks have been devised only
against a limited class of binary learning algorithms, due to the inherent
complexity of the gradient-based procedure used to optimize the poisoning
points (a.k.a. adversarial training examples). In this work, we rst extend the
de nition of poisoning attacks to multiclass problems. We then propose a novel
poisoning algorithm based on the idea of back-gradient optimization, i.e., to
compute the gradient of interest through automatic di erentiation, while also
reversing the learning procedure to drastically reduce the attack complexity.
Compared to current poisoning strategies, our approach is able to target a
wider class of learning algorithms, trained with gradient- based procedures,
including neural networks and deep learning architectures. We empirically
evaluate its e ectiveness on several application examples, including spam
ltering, malware detection, and handwritten digit recognition. We nally show
that, similarly to adversarial test examples, adversarial training examples can
also be transferred across di erent learning algorithms. | ['Andrea Paudice', 'Luis Muñoz-González', 'Battista Biggio', 'Fabio Roli', 'Vasin Wongrassamee', 'Ambra Demontis', 'Emil C. Lupu'] | 2017-08-29 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 3.75525653e-01 -7.65124410e-02 -1.87456664e-02 1.78287178e-01
-5.29378116e-01 -1.26233554e+00 1.02465641e+00 3.20032090e-01
-7.44556427e-01 6.70944810e-01 -3.42979431e-01 -5.87660372e-01
2.36252998e-03 -9.40697610e-01 -8.68371725e-01 -9.66694653e-01
-3.91871601e-01 6.21090353e-01 2.63635039e-01 -3.51207197e-01
4.17930841e-01 1.16651332e+00 -1.21632516e+00 2.51411170e-01
6.21077895e-01 8.41926634e-01 -5.51871181e-01 8.96399796e-01
3.74998599e-02 8.59976292e-01 -1.12360048e+00 -6.84356093e-01
4.23727959e-01 -4.11632121e-01 -7.16402233e-01 4.80964929e-02
4.86408502e-01 -3.19482893e-01 -3.61409009e-01 1.40661395e+00
3.75401616e-01 -1.57877743e-01 6.19747639e-01 -1.53632188e+00
-1.35444641e-01 5.10462463e-01 -1.40585482e-01 8.17216039e-02
1.34753302e-01 4.11477149e-01 5.89265525e-01 -4.73512113e-01
4.74272221e-01 1.16255772e+00 5.85929036e-01 9.76520419e-01
-1.42921364e+00 -8.53184879e-01 -8.77439901e-02 1.31453842e-01
-8.41475308e-01 -2.40109891e-01 8.94056857e-01 -2.34907269e-01
4.45250303e-01 4.92586583e-01 3.32667291e-01 1.49855256e+00
2.28520855e-01 7.60683179e-01 1.23895419e+00 -2.00053602e-01
6.02817833e-01 3.51728171e-01 -5.18183298e-02 6.35330975e-01
3.59214336e-01 4.35304880e-01 -2.56943703e-01 -7.99519122e-01
2.07813904e-01 -4.08440828e-02 -3.01568508e-01 -5.92513442e-01
-6.17437243e-01 9.90362883e-01 5.43492913e-01 2.04609469e-01
-2.34203741e-01 2.41535038e-01 8.20446730e-01 6.81703866e-01
1.99003413e-01 6.59555316e-01 -4.27316993e-01 1.92347467e-01
-6.14152133e-01 2.88809538e-01 1.31147921e+00 1.69647262e-01
6.54121697e-01 2.13285327e-01 2.08202794e-01 3.03495526e-01
1.12070829e-01 4.01854008e-01 7.08929539e-01 -4.45102155e-01
5.10813296e-01 4.63700771e-01 -4.52044569e-02 -8.14467072e-01
-2.35676125e-01 -3.94775957e-01 -8.51338983e-01 8.30735207e-01
8.88917744e-01 -2.71421373e-01 -8.32971811e-01 1.54965675e+00
5.60004830e-01 2.20605135e-01 2.40039423e-01 6.56289458e-01
-1.98550783e-02 2.19024539e-01 6.54132813e-02 4.94009182e-02
9.65957522e-01 -5.76545060e-01 -1.24034569e-01 -1.76911037e-02
8.08087826e-01 -3.17271799e-01 1.09329736e+00 9.36713874e-01
-7.50768363e-01 -3.45142968e-02 -1.25319815e+00 5.58464587e-01
-9.12611365e-01 -5.08426547e-01 3.93042266e-01 1.06759143e+00
-5.70145369e-01 1.06602693e+00 -8.49776208e-01 9.23120901e-02
1.02815163e+00 7.37800419e-01 -3.55457962e-01 9.93196741e-02
-1.21172559e+00 6.68718755e-01 4.10143733e-01 -2.38998249e-01
-1.31174242e+00 -6.25892222e-01 -4.12544042e-01 -8.72320533e-02
3.38980138e-01 -2.95789838e-01 9.93546903e-01 -1.26865709e+00
-1.43827665e+00 8.69160771e-01 6.26127303e-01 -1.07192981e+00
1.02937198e+00 -2.87593275e-01 -1.13194600e-01 3.18682969e-01
-3.24452966e-01 1.12397067e-01 1.40935874e+00 -1.32848620e+00
-4.31777865e-01 -6.02884769e-01 2.52053618e-01 -2.91866213e-01
-8.96304727e-01 1.49235532e-01 3.06968063e-01 -7.90714800e-01
-3.04195136e-01 -9.75087762e-01 -4.27640229e-01 5.99962240e-03
-6.07553363e-01 9.00209025e-02 1.39184177e+00 -3.56797993e-01
9.02420104e-01 -2.04533434e+00 2.96633393e-01 4.55549300e-01
3.44485104e-01 7.63716638e-01 2.65622884e-02 5.17845929e-01
-1.93687961e-01 1.62435859e-01 -7.19777584e-01 -2.33077779e-01
4.66834158e-02 2.40462840e-01 -8.50866199e-01 9.30423677e-01
1.99491978e-01 8.18615437e-01 -9.64344561e-01 -1.67834274e-02
1.96574762e-01 1.84768975e-01 -5.58804035e-01 1.90433562e-01
-4.44789886e-01 3.64272326e-01 -4.93912488e-01 5.06210804e-01
4.73248035e-01 8.27033073e-02 -9.41194501e-03 2.42843404e-01
3.18659276e-01 6.00645132e-02 -1.04589772e+00 1.08664811e+00
-4.90583390e-01 3.44900191e-01 2.35908881e-01 -1.18748415e+00
6.59056485e-01 1.49664328e-01 2.68395543e-01 -1.19164646e-01
4.28319156e-01 4.55184191e-01 1.78798363e-01 -2.15678871e-01
-1.24625184e-01 -9.04576033e-02 -2.51492679e-01 6.29157603e-01
-2.09270731e-01 -9.21944082e-02 -1.02089874e-01 1.28242180e-01
1.52799332e+00 -1.23987287e-01 2.37908751e-01 -1.54391870e-01
9.30491269e-01 1.94501951e-01 -2.74651814e-02 1.06199801e+00
-1.70276865e-01 9.98323560e-02 6.54016614e-01 -5.10902107e-01
-1.06710041e+00 -9.82392251e-01 1.30999666e-02 8.23144436e-01
-3.20268869e-02 -3.02131891e-01 -1.04431486e+00 -1.56414616e+00
1.99754387e-01 6.49785876e-01 -7.59286046e-01 -7.73303628e-01
-7.92965770e-01 -7.96532989e-01 1.28532028e+00 1.51693538e-01
7.70115972e-01 -1.08876848e+00 -8.84158492e-01 1.10596016e-01
6.29446924e-01 -6.89379692e-01 -6.99508637e-02 4.87172455e-01
-9.52429533e-01 -1.41248333e+00 -1.95622876e-01 -3.22500557e-01
5.45982242e-01 -2.73405075e-01 8.11328888e-01 2.75034457e-01
-4.87940520e-01 3.11544478e-01 -2.50053704e-01 -4.73633915e-01
-1.09117520e+00 1.44536242e-01 4.88344133e-02 2.78368324e-01
2.99810588e-01 -7.20494986e-01 -1.95811212e-01 2.97164097e-02
-1.35361063e+00 -5.86187601e-01 4.38789308e-01 1.09034884e+00
7.67107159e-02 2.65219510e-01 3.45164716e-01 -1.19925344e+00
7.03084409e-01 -5.27253807e-01 -8.78289819e-01 1.15986437e-01
-3.26556772e-01 2.28470594e-01 1.54148149e+00 -1.04010820e+00
-3.60097617e-01 2.29860231e-01 -2.69925416e-01 -7.38156319e-01
-3.49181622e-01 -2.39131488e-02 -4.42772835e-01 -5.24380922e-01
1.09786487e+00 3.87710631e-01 8.22258890e-02 -4.85287845e-01
4.07022804e-01 6.33450270e-01 4.51612771e-01 -4.87931103e-01
1.34283257e+00 7.13267744e-01 4.11203086e-01 -7.48814940e-01
-3.78227681e-01 -1.89862549e-02 -4.32157695e-01 3.76518033e-02
4.88277465e-01 -2.25359544e-01 -8.49188328e-01 7.86387503e-01
-1.05934453e+00 -4.15652215e-01 -5.12960017e-01 1.27017602e-01
-6.38571918e-01 4.28406805e-01 -6.80173278e-01 -6.58176780e-01
-4.30726647e-01 -1.07500648e+00 7.69176364e-01 -2.29660124e-01
4.11690883e-02 -1.01261199e+00 6.33867905e-02 7.13206222e-03
3.63749474e-01 5.34519613e-01 1.02247846e+00 -1.44342673e+00
-3.47307622e-01 -6.49694026e-01 2.65415847e-01 6.75069988e-01
1.39265908e-02 -3.34234148e-01 -1.02034688e+00 -4.88088459e-01
5.80226600e-01 -5.11278510e-01 8.12948346e-01 -4.63357478e-01
1.31201124e+00 -8.91289949e-01 -2.22727358e-01 6.70381248e-01
1.32992458e+00 1.26702050e-02 5.88934779e-01 6.75096810e-01
8.28399599e-01 5.98298907e-01 3.27211529e-01 3.14611435e-01
-5.41127086e-01 6.41044855e-01 9.05191064e-01 5.12035303e-02
1.23929344e-01 -8.22978541e-02 5.11841476e-01 -2.34367043e-01
3.88749838e-01 -1.58352461e-02 -8.81977737e-01 1.80002954e-02
-1.48727453e+00 -9.86463368e-01 1.26169818e-02 2.41069674e+00
9.01954293e-01 4.38549012e-01 3.77733320e-01 5.76769352e-01
6.32248700e-01 -4.06665169e-02 -9.45645213e-01 -5.11120081e-01
6.66336790e-02 4.67357159e-01 8.48275781e-01 4.56183016e-01
-1.41321766e+00 9.55561697e-01 5.23048925e+00 9.46228266e-01
-1.39321947e+00 1.70545369e-01 4.17585403e-01 -3.03661730e-02
1.90573782e-01 -1.83944792e-01 -5.57032049e-01 5.14130414e-01
1.11714458e+00 -4.33237180e-02 7.44738340e-01 1.00811946e+00
-1.76754951e-01 4.64131862e-01 -1.06735229e+00 6.33604109e-01
1.22946486e-01 -1.15383518e+00 1.49275333e-01 3.37836981e-01
3.12421292e-01 2.89798100e-02 8.97165239e-02 1.21934041e-01
6.93113267e-01 -9.71839130e-01 4.79151100e-01 -1.40491286e-02
2.61542201e-01 -1.11296737e+00 3.79896462e-01 6.36316657e-01
-5.11027515e-01 -5.32871127e-01 -9.08382982e-02 1.72158837e-01
-3.82758111e-01 4.12452251e-01 -8.04786205e-01 2.92904645e-01
5.00289559e-01 3.60961318e-01 -6.72338128e-01 7.79837430e-01
-4.37925339e-01 6.99628234e-01 -3.99104923e-01 -1.36801183e-01
4.85472143e-01 -2.08444342e-01 8.03610384e-01 1.06210947e+00
-2.68329859e-01 -3.82889420e-01 1.77237660e-01 6.39551580e-01
-3.50026697e-01 -1.60585418e-02 -1.08400691e+00 -1.09479930e-02
3.36304158e-01 1.13625109e+00 -7.77610838e-01 -3.25079441e-01
-1.39652118e-01 1.22766614e+00 3.87536824e-01 1.52070582e-01
-8.59492600e-01 -5.54205179e-01 7.25499928e-01 5.57107963e-02
2.85474569e-01 -1.60487697e-01 -6.91257566e-02 -1.03615713e+00
-1.29535615e-01 -1.50146019e+00 4.68432158e-01 1.24855374e-03
-1.37441778e+00 5.51975369e-01 -1.52356654e-01 -1.10709035e+00
-3.06083351e-01 -8.54187250e-01 -6.46618009e-01 5.04106104e-01
-1.24279988e+00 -7.77078390e-01 1.73100784e-01 7.85299480e-01
2.96368122e-01 -3.46677423e-01 7.65488446e-01 1.57804769e-02
-5.87173104e-01 6.62353396e-01 3.50412756e-01 3.51318568e-01
3.89670819e-01 -1.36047029e+00 4.59783912e-01 9.83207166e-01
2.15619564e-01 5.49850225e-01 7.79842377e-01 -5.61526954e-01
-1.49275362e+00 -1.23971927e+00 2.89867073e-01 -5.39761186e-01
1.08973336e+00 -7.51362026e-01 -9.05067146e-01 4.72673565e-01
-1.34465784e-01 1.21987738e-01 4.61826891e-01 -3.73476893e-01
-6.64233685e-01 -6.26725778e-02 -1.65413475e+00 7.95877576e-01
6.46863580e-01 -5.44187248e-01 -4.36731458e-01 5.94221294e-01
4.91619706e-01 -1.41006097e-01 -5.92666447e-01 2.24464715e-01
2.69965559e-01 -8.49759161e-01 1.21295989e+00 -1.22551119e+00
3.28122258e-01 -1.29564092e-01 -1.07882157e-01 -1.12429357e+00
3.74624372e-01 -8.75655770e-01 -7.34733343e-01 9.31935906e-01
4.49762270e-02 -9.43894744e-01 9.56829667e-01 2.01788947e-01
2.31320351e-01 -7.05667734e-01 -1.11331332e+00 -9.88614082e-01
4.52162355e-01 -2.35862792e-01 5.65107763e-01 9.42307532e-01
-2.44251773e-01 4.21132036e-02 -3.03062171e-01 3.84957731e-01
9.35212076e-01 -9.41556245e-02 9.73369837e-01 -1.15168846e+00
-7.48766184e-01 -6.11155272e-01 -6.20025277e-01 -5.81609309e-01
4.47244138e-01 -8.99362803e-01 -1.73193857e-01 -4.98579562e-01
-3.54232937e-01 -3.01488608e-01 -1.55459240e-01 5.09337783e-01
-4.83218879e-02 5.26531935e-01 2.83045739e-01 3.65410775e-01
-1.47835344e-01 1.89563438e-01 7.77489126e-01 -2.56761938e-01
-1.37115687e-01 3.93914312e-01 -4.04900521e-01 8.68756115e-01
1.12574315e+00 -1.01762390e+00 -1.82133853e-01 1.89467356e-01
-7.66945481e-02 -4.94376957e-01 7.32141614e-01 -1.06259668e+00
2.92562902e-01 1.01491556e-01 3.29367816e-01 -5.20343147e-02
1.27661511e-01 -1.03706598e+00 -3.75691056e-01 1.30820179e+00
-3.94243389e-01 -9.45451707e-02 1.20264418e-01 5.93134403e-01
2.06129149e-01 -5.24353981e-01 1.04148185e+00 -2.59519130e-01
-1.34014055e-01 2.32695118e-01 -4.22108024e-01 6.85536955e-03
1.28197074e+00 3.33281830e-02 -3.16509128e-01 3.87300812e-02
-5.50743639e-01 -1.59331813e-01 6.21347189e-01 1.03051305e-01
4.58293110e-01 -9.93346691e-01 -5.45411050e-01 3.75724554e-01
-1.50672883e-01 -3.99513781e-01 -4.66907531e-01 4.94578987e-01
-5.75976729e-01 4.72963741e-03 -1.14093997e-01 -3.30255300e-01
-1.47313309e+00 1.15706635e+00 7.24618137e-01 -4.25492704e-01
-6.81321204e-01 4.42713916e-01 -1.40712678e-01 -3.49810094e-01
6.19066119e-01 2.50758350e-01 1.13529190e-02 -2.95984279e-02
6.73206508e-01 4.13089335e-01 2.78234631e-01 -2.61332810e-01
-3.18409503e-01 1.18750587e-01 -2.41641909e-01 5.60466237e-02
1.13640893e+00 5.05877435e-01 -2.11568519e-01 7.01150000e-02
1.29789817e+00 8.63474533e-02 -1.21750367e+00 -1.61187295e-02
2.65017182e-01 -4.80615288e-01 -2.25704700e-01 -7.47330844e-01
-9.52921569e-01 9.32416320e-01 6.81167901e-01 5.60730875e-01
1.21433437e+00 -2.77071208e-01 5.77455163e-01 8.34934890e-01
4.12191570e-01 -7.19052196e-01 2.89775252e-01 2.33125940e-01
6.24330103e-01 -9.89228904e-01 -1.65636718e-01 -1.79815799e-01
-2.26735577e-01 1.27347147e+00 3.28768700e-01 -5.85433602e-01
5.17751694e-01 5.37028968e-01 7.06178462e-03 -1.63965985e-01
-4.80569124e-01 9.04561058e-02 -2.03175515e-01 5.96741498e-01
-2.99170017e-01 -3.21298152e-01 -1.89695969e-01 8.42418149e-02
-9.97112170e-02 -3.44837129e-01 5.75011551e-01 1.11518371e+00
-3.97967219e-01 -1.47781122e+00 -6.67268932e-01 3.81575286e-01
-6.97641134e-01 -5.92150260e-03 -9.68921065e-01 8.98662925e-01
-1.02732731e-02 8.64655912e-01 -3.77972484e-01 -4.61111546e-01
3.83607298e-01 8.18154067e-02 3.88132095e-01 -5.20324528e-01
-1.23517513e+00 -4.81932461e-01 -1.95824608e-01 -6.52237713e-01
9.34591517e-03 -5.70044518e-01 -1.00567484e+00 -4.64482874e-01
-3.34889501e-01 2.20715523e-01 8.17977786e-01 9.12681699e-01
-8.44277162e-03 2.54863352e-01 1.07731521e+00 -9.53414798e-01
-1.55285656e+00 -4.89679605e-01 -3.73992264e-01 7.07553983e-01
5.31016111e-01 -2.84202904e-01 -8.61534119e-01 -2.97786623e-01] | [5.685928821563721, 7.6347551345825195] |
c405efe7-204f-44a4-9674-39486f357665 | masksketch-unpaired-structure-guided-masked | 2302.05496 | null | https://arxiv.org/abs/2302.05496v1 | https://arxiv.org/pdf/2302.05496v1.pdf | MaskSketch: Unpaired Structure-guided Masked Image Generation | Recent conditional image generation methods produce images of remarkable diversity, fidelity and realism. However, the majority of these methods allow conditioning only on labels or text prompts, which limits their level of control over the generation result. In this paper, we introduce MaskSketch, an image generation method that allows spatial conditioning of the generation result using a guiding sketch as an extra conditioning signal during sampling. MaskSketch utilizes a pre-trained masked generative transformer, requiring no model training or paired supervision, and works with input sketches of different levels of abstraction. We show that intermediate self-attention maps of a masked generative transformer encode important structural information of the input image, such as scene layout and object shape, and we propose a novel sampling method based on this observation to enable structure-guided generation. Our results show that MaskSketch achieves high image realism and fidelity to the guiding structure. Evaluated on standard benchmark datasets, MaskSketch outperforms state-of-the-art methods for sketch-to-image translation, as well as unpaired image-to-image translation approaches. | ['Irfan Essa', 'Kate Saenko', 'Kihyuk Sohn', 'Jose Lezama', 'Dina Bashkirova'] | 2023-02-10 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Bashkirova_MaskSketch_Unpaired_Structure-Guided_Masked_Image_Generation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Bashkirova_MaskSketch_Unpaired_Structure-Guided_Masked_Image_Generation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['sketch-to-image-translation', 'conditional-image-generation'] | ['computer-vision', 'computer-vision'] | [ 8.55902314e-01 3.70506316e-01 -4.37167138e-02 -4.23045307e-01
-6.59137964e-01 -7.22325087e-01 1.11655259e+00 -3.06846172e-01
1.04318976e-01 6.33341253e-01 2.00929761e-01 -1.45838067e-01
4.64834571e-01 -9.58236217e-01 -1.28341854e+00 -5.14335632e-01
5.18543124e-01 5.04918993e-01 1.00583859e-01 -1.55022383e-01
3.62693101e-01 4.51654613e-01 -1.52387440e+00 7.56778896e-01
8.46131504e-01 9.73115861e-01 5.81813753e-01 7.89900362e-01
-1.91876113e-01 7.94923007e-01 -5.25933027e-01 -4.41100717e-01
2.76395053e-01 -8.06704700e-01 -5.47990203e-01 1.73079044e-01
9.12708938e-01 -6.53959036e-01 -3.59499305e-01 7.07524419e-01
4.26968098e-01 -1.99357912e-01 8.95048678e-01 -1.15726542e+00
-1.15250468e+00 5.85336924e-01 -2.96763062e-01 -4.97098595e-01
4.62297529e-01 5.06302714e-01 9.25139904e-01 -1.18426573e+00
1.00392950e+00 1.31913102e+00 4.15073901e-01 6.66458011e-01
-1.90090191e+00 -7.53628731e-01 1.66613623e-01 -2.90781111e-01
-1.32838857e+00 -6.42971873e-01 9.37043369e-01 -5.02219200e-01
6.58311009e-01 2.67733097e-01 7.70920455e-01 1.33276653e+00
-7.87576735e-02 8.83647263e-01 1.36542702e+00 -5.68143547e-01
2.76776880e-01 2.82061517e-01 -6.69647217e-01 9.23848927e-01
-9.12062824e-02 3.25210214e-01 -7.87479758e-01 3.68182175e-02
1.33908677e+00 -1.35988325e-01 -3.31186116e-01 -4.78414536e-01
-1.36641026e+00 7.45181799e-01 6.08664870e-01 -1.21151768e-01
-1.97101668e-01 5.25321245e-01 -9.29464102e-02 1.05158038e-01
3.78878653e-01 7.10986197e-01 4.22655903e-02 1.48196727e-01
-1.19488466e+00 3.28997314e-01 4.82696712e-01 1.31817234e+00
9.70757902e-01 1.63148493e-01 -8.67230356e-01 5.24863362e-01
-7.71597847e-02 8.02999794e-01 1.03253752e-01 -1.10300326e+00
4.63446110e-01 4.31872666e-01 1.67864084e-01 -8.64087224e-01
3.51681441e-01 -1.64248481e-01 -8.20375264e-01 4.92038071e-01
2.16893852e-01 1.66605011e-01 -1.11122513e+00 1.90922594e+00
3.32322605e-02 6.22937307e-02 -2.77720988e-01 8.08931947e-01
7.15051413e-01 6.76047802e-01 -5.29139675e-02 1.88467249e-01
9.82249558e-01 -1.02248931e+00 -5.82912207e-01 -8.62750709e-02
9.58323479e-02 -8.16636920e-01 1.60292852e+00 3.79399508e-01
-1.40542758e+00 -8.67412508e-01 -1.04765344e+00 -2.77404934e-01
-2.24023685e-01 3.54387462e-01 5.94517469e-01 4.34649169e-01
-1.30318189e+00 7.73387909e-01 -5.12324631e-01 -5.78717478e-02
6.07050717e-01 2.04245094e-02 -2.91831315e-01 -1.55123249e-01
-8.38574052e-01 5.81736684e-01 6.53018057e-02 -2.13496044e-01
-1.13062549e+00 -9.31971371e-01 -9.50652003e-01 6.77415505e-02
5.60294501e-02 -1.06270695e+00 1.17339921e+00 -1.04366887e+00
-1.96753728e+00 9.04725552e-01 -8.82526338e-02 -7.09560513e-02
7.33632922e-01 -4.52882759e-02 1.66258171e-01 2.71228582e-01
1.37963325e-01 1.35528743e+00 1.25117183e+00 -1.71166039e+00
-1.22475401e-01 -2.87464894e-02 5.52295782e-02 3.35674291e-03
-5.90172112e-02 -3.66813660e-01 -5.03170550e-01 -8.20713758e-01
-1.90742493e-01 -8.12958062e-01 -8.34735706e-02 5.28667808e-01
-6.86586380e-01 1.56173497e-01 7.84465253e-01 -3.50265503e-01
8.85474980e-01 -2.08850217e+00 2.99176723e-01 6.30702004e-02
6.89600557e-02 1.03766777e-01 -4.65044945e-01 6.03092074e-01
3.88560370e-02 2.64786094e-01 -3.01078647e-01 -8.78499568e-01
2.83630759e-01 2.15888381e-01 -8.37472737e-01 1.19747438e-01
6.34347558e-01 1.28460157e+00 -9.69045937e-01 -5.61692357e-01
5.18707156e-01 7.22753108e-01 -9.14043903e-01 6.33633733e-01
-6.02336228e-01 6.58520341e-01 -8.94659013e-02 4.06796187e-01
6.75201535e-01 -3.55810046e-01 -1.53847188e-02 -4.82624501e-01
-7.63610797e-03 2.46980965e-01 -7.07713783e-01 2.17878032e+00
-7.21525848e-01 7.51249671e-01 -1.34476930e-01 -5.45180082e-01
9.68314528e-01 1.59955338e-01 -6.13363609e-02 -7.82122374e-01
-2.50650030e-02 1.61664244e-02 -5.03967047e-01 -9.47136804e-02
5.65733075e-01 -1.92013010e-01 4.27163206e-02 5.74894965e-01
1.02438055e-01 -8.45366001e-01 6.93544792e-03 5.22118092e-01
8.15893352e-01 6.31935596e-01 -3.68278585e-02 -1.97603300e-01
1.08951703e-01 -3.19935858e-01 3.84139940e-02 9.45071340e-01
4.98288393e-01 1.21551573e+00 4.53080714e-01 -2.52731651e-01
-1.48746026e+00 -1.45171738e+00 2.43728589e-02 8.63137543e-01
1.23589769e-01 -4.64860678e-01 -9.28009152e-01 -5.46413600e-01
-1.09850757e-01 8.76503944e-01 -9.58090305e-01 -1.35479078e-01
-4.46255356e-01 -5.29988185e-02 5.45114696e-01 5.44165492e-01
3.84272277e-01 -1.33757031e+00 -6.36440217e-01 7.73924068e-02
-1.14233382e-01 -1.19590545e+00 -8.17707419e-01 -1.51651114e-01
-7.25941777e-01 -6.43491387e-01 -8.41506481e-01 -6.52605355e-01
1.12420237e+00 2.47206893e-02 1.36428189e+00 8.74994621e-02
-4.59628642e-01 2.62703449e-01 -9.66430157e-02 -1.20376885e-01
-5.86827755e-01 -3.80329415e-02 -4.47742194e-01 1.88912079e-01
-5.61316609e-01 -8.74497116e-01 -8.29306781e-01 2.43045971e-01
-1.11479700e+00 9.54985559e-01 8.01232994e-01 1.09781277e+00
5.27520835e-01 -5.53161502e-01 1.81543097e-01 -8.54383051e-01
5.24155378e-01 1.65979877e-01 -4.71677810e-01 1.56838983e-01
-3.91167432e-01 4.37665999e-01 7.93858826e-01 -6.91332817e-01
-1.15310514e+00 2.36940145e-01 2.05102652e-01 -6.74880624e-01
-9.38152224e-02 9.57153887e-02 -2.31175557e-01 -5.87418349e-03
7.83196270e-01 4.42688793e-01 5.03858067e-02 -3.85172576e-01
7.68521428e-01 1.97782829e-01 7.71675885e-01 -9.68955398e-01
8.19195867e-01 6.91027403e-01 5.13003506e-02 -5.89366317e-01
-6.57787681e-01 3.38994265e-01 -5.55862010e-01 -1.30255431e-01
7.03889489e-01 -7.03494489e-01 -5.32965600e-01 3.37755471e-01
-1.50858963e+00 -7.96887159e-01 -5.16072035e-01 -8.07538405e-02
-8.72317135e-01 -2.17665881e-02 -6.05154514e-01 -6.46205187e-01
-2.84429848e-01 -1.15443075e+00 1.59511948e+00 -1.64485648e-01
-1.70987368e-01 -5.98076165e-01 -7.04768598e-02 2.55601425e-02
6.04111433e-01 4.17245120e-01 9.05620039e-01 2.52929926e-01
-1.13599122e+00 4.66824770e-02 -4.53654230e-01 2.88883686e-01
1.66259587e-01 9.37369615e-02 -1.05682552e+00 -2.25880906e-01
-4.67434317e-01 -5.45521796e-01 8.59372139e-01 1.88864693e-01
1.36018062e+00 -5.52074850e-01 -2.48806804e-01 8.82938802e-01
1.40941119e+00 -1.20747909e-01 9.96707439e-01 -2.26451755e-01
8.72265041e-01 4.62050050e-01 1.48707718e-01 3.97505641e-01
9.14918557e-02 8.49987864e-01 3.63266736e-01 -5.70421696e-01
-6.41990483e-01 -1.00616300e+00 3.76623750e-01 3.59828532e-01
4.86123562e-02 -2.84610510e-01 -3.29338878e-01 4.08657759e-01
-1.59205544e+00 -9.59076166e-01 1.63134128e-01 2.12182760e+00
1.14417231e+00 -6.51583374e-02 -1.77967340e-01 -8.13295767e-02
6.15429521e-01 2.02680439e-01 -4.61842805e-01 -3.70651454e-01
-1.96493372e-01 6.60154462e-01 2.03871489e-01 6.73551559e-01
-7.19171584e-01 1.16855288e+00 6.61211538e+00 9.55373168e-01
-1.25166428e+00 -1.22753292e-01 7.63782561e-01 -2.89256535e-02
-7.79061913e-01 1.25125915e-01 -4.27758306e-01 4.27931815e-01
2.34902069e-01 1.24857426e-01 7.38686442e-01 5.89623034e-01
3.53236273e-02 2.25735195e-02 -1.49262750e+00 9.13802445e-01
1.56176791e-01 -1.72987974e+00 5.81338346e-01 7.91598633e-02
1.04481483e+00 -4.44298953e-01 4.34506744e-01 2.10613292e-02
4.03208077e-01 -1.30548775e+00 1.22796404e+00 7.28376389e-01
1.53502083e+00 -5.51286995e-01 6.11328036e-02 1.45226061e-01
-1.06047702e+00 3.72063428e-01 -1.46951050e-01 -5.93972020e-02
3.11261356e-01 4.42369998e-01 -8.10489237e-01 3.43583852e-01
2.20239654e-01 5.05572379e-01 -5.86554825e-01 3.96175653e-01
-6.48090422e-01 5.25003850e-01 -1.35638058e-01 1.07265949e-01
9.25932899e-02 -1.09282255e-01 2.16530159e-01 1.31249022e+00
4.21038896e-01 -4.36870717e-02 1.45803094e-01 1.84667742e+00
-1.05683953e-01 -6.24801554e-02 -8.48582804e-01 -1.45453870e-01
4.01572436e-01 1.03453982e+00 -5.66590667e-01 -6.53798938e-01
7.08988635e-03 1.51600003e+00 3.70955825e-01 5.33945441e-01
-8.84513438e-01 -2.71710068e-01 4.40466791e-01 4.64280367e-01
4.71713245e-01 -1.74268857e-01 -5.75549960e-01 -1.07470584e+00
3.10996436e-02 -6.78007483e-01 -4.01088476e-01 -1.27010834e+00
-1.08231699e+00 7.03052163e-01 -2.41507869e-02 -1.03784657e+00
-2.94716120e-01 -5.31582773e-01 -5.98000348e-01 1.04324639e+00
-1.30189490e+00 -1.66105080e+00 -5.99917769e-01 5.87120235e-01
3.65862936e-01 2.15033926e-02 9.48010564e-01 -2.02602427e-02
-2.68633496e-02 7.24826217e-01 -4.35773641e-01 1.34822860e-01
8.19928110e-01 -1.24778199e+00 7.32096851e-01 6.51690125e-01
2.67567456e-01 6.93993688e-01 5.91454804e-01 -6.00359142e-01
-1.50492871e+00 -1.18152320e+00 6.03377402e-01 -5.35774887e-01
1.75694704e-01 -8.44380915e-01 -5.19395888e-01 4.77614105e-01
5.46037495e-01 1.24350064e-01 2.19860300e-01 -3.26758593e-01
-7.02675700e-01 -1.39540406e-02 -1.04805112e+00 9.07711148e-01
1.33776462e+00 -7.74072826e-01 -2.69051492e-01 8.33363831e-02
8.71170342e-01 -4.47050065e-01 -4.27650630e-01 2.42232069e-01
7.16412425e-01 -1.17261398e+00 1.06008339e+00 -2.90828496e-01
1.00258875e+00 -4.39669579e-01 -1.82737157e-01 -1.33463740e+00
-4.87676233e-01 -8.98316681e-01 4.57950728e-03 1.14861512e+00
4.53761607e-01 -2.90354431e-01 7.09628105e-01 4.26648408e-01
-2.55173445e-01 -5.80937088e-01 -4.48846519e-01 -7.19497919e-01
-1.55396638e-02 -2.66347647e-01 7.94834137e-01 7.41982520e-01
-2.33487263e-01 3.87755483e-01 -6.00002229e-01 -2.08762378e-01
7.26781666e-01 3.89580280e-01 1.14028132e+00 -7.96736777e-01
-5.12809634e-01 -6.07789218e-01 -1.38853848e-01 -1.32400870e+00
5.49861006e-02 -9.29969192e-01 1.70518622e-01 -1.51451206e+00
2.65167385e-01 -3.56123388e-01 2.48953685e-01 4.49516267e-01
-1.52398273e-01 6.93153560e-01 4.68018383e-01 4.50070053e-02
-2.84034342e-01 8.44591975e-01 1.79961741e+00 -3.29358280e-01
-3.18649411e-02 -4.92406696e-01 -6.99662268e-01 2.65478522e-01
4.07847077e-01 -1.45432040e-01 -6.35482907e-01 -4.37403023e-01
5.74164502e-02 3.36699598e-02 7.62244344e-01 -8.13313961e-01
6.46066340e-03 -2.67488688e-01 7.05152810e-01 -4.85627681e-01
5.79010785e-01 -4.08870071e-01 4.27292436e-01 2.43589818e-01
-6.43698037e-01 -1.51466340e-01 1.15895592e-01 3.47435921e-01
-5.53390756e-02 1.50268614e-01 7.79278576e-01 -1.28527701e-01
-1.81160748e-01 4.33886498e-01 2.19051968e-02 7.58062229e-02
5.44211030e-01 -2.89282948e-01 -2.41431668e-01 -5.96367478e-01
-4.00056303e-01 -3.43094945e-01 9.09049153e-01 3.14434022e-01
8.59421790e-01 -1.73894429e+00 -7.85740614e-01 5.92508495e-01
3.14390182e-01 1.60026088e-01 -1.36255389e-02 3.46090972e-01
-3.77646983e-01 2.38761827e-01 -2.46729940e-01 -7.96050012e-01
-9.21910644e-01 7.41340578e-01 7.59912748e-03 7.49040220e-04
-6.79472685e-01 8.15415502e-01 8.82755995e-01 -2.89539635e-01
6.48268610e-02 -5.79890668e-01 6.49563789e-01 -3.78590673e-01
4.96530116e-01 -2.93261766e-01 -2.85505235e-01 -2.95824349e-01
1.68240249e-01 6.22151196e-01 1.14422485e-01 -5.68242550e-01
1.02496052e+00 1.58075273e-01 4.64056544e-02 2.41806671e-01
1.06598628e+00 -1.98657941e-02 -1.84523523e+00 -2.09290266e-01
-6.47351563e-01 -8.54746282e-01 -9.43755358e-02 -1.08863556e+00
-9.62768972e-01 1.08740234e+00 1.93743646e-01 -1.59310579e-01
1.03770041e+00 -3.16586555e-03 5.79910457e-01 2.33795661e-02
5.88933349e-01 -7.47478426e-01 6.00807965e-01 2.39124775e-01
1.56703866e+00 -1.06639254e+00 -2.39145994e-01 -4.14948672e-01
-6.54253602e-01 7.91612506e-01 6.17174447e-01 -3.31623852e-01
2.94744283e-01 3.84702593e-01 -1.46687746e-01 -3.64778787e-02
-8.18291962e-01 -6.62161112e-02 4.75498468e-01 7.61171162e-01
4.00922269e-01 1.08318828e-01 1.48743480e-01 3.05116743e-01
-3.59355956e-01 8.39308798e-02 1.44496709e-01 7.33801782e-01
-2.88544558e-02 -1.18884993e+00 -3.17584485e-01 1.14413366e-01
1.37503535e-01 -4.58197743e-01 -6.86790645e-01 5.91625452e-01
2.25949451e-01 5.66734076e-01 1.17127798e-01 -2.82834381e-01
1.90022036e-01 -1.21110857e-01 1.04407918e+00 -6.49215817e-01
-3.64266604e-01 8.46884549e-02 -1.25407323e-01 -7.96431839e-01
-1.29614770e-01 -3.70990723e-01 -8.99502635e-01 -2.56567508e-01
-1.60401449e-01 -1.99040890e-01 6.63021803e-01 6.02280259e-01
5.86000323e-01 5.64535201e-01 6.23694956e-01 -1.42954254e+00
-1.79074734e-01 -1.00001872e+00 -2.83620179e-01 7.08720565e-01
3.18006217e-01 -6.07041597e-01 -1.92828014e-01 4.30702627e-01] | [11.466694831848145, -0.3659377694129944] |
42e3f44b-f5cd-4c78-93e6-2fc6ee4444cf | 3d-object-detection-method-based-on-yolo-and | 2005.02132 | null | https://arxiv.org/abs/2005.02132v1 | https://arxiv.org/pdf/2005.02132v1.pdf | 3D Object Detection Method Based on YOLO and K-Means for Image and Point Clouds | Lidar based 3D object detection and classification tasks are essential for autonomous driving(AD). A lidar sensor can provide the 3D point cloud data reconstruction of the surrounding environment. However, real time detection in 3D point clouds still needs a strong algorithmic. This paper proposes a 3D object detection method based on point cloud and image which consists of there parts.(1)Lidar-camera calibration and undistorted image transformation. (2)YOLO-based detection and PointCloud extraction, (3)K-means based point cloud segmentation and detection experiment test and evaluation in depth image. In our research, camera can capture the image to make the Real-time 2D object detection by using YOLO, we transfer the bounding box to node whose function is making 3d object detection on point cloud data from Lidar. By comparing whether 2D coordinate transferred from the 3D point is in the object bounding box or not can achieve High-speed 3D object recognition function in GPU. The accuracy and precision get imporved after k-means clustering in point cloud. The speed of our detection method is a advantage faster than PointNet. | ['Xuanyu YIN', 'Kentaro SHIMIZU', 'Weimin WANG', 'Yoko SASAKI'] | 2020-04-21 | null | null | null | null | ['3d-object-recognition'] | ['computer-vision'] | [-3.41881216e-01 -6.56730890e-01 1.96046144e-01 -3.69048089e-01
-2.28784874e-01 -4.01441365e-01 2.89552659e-01 2.98325509e-01
-6.64329886e-01 -1.29487246e-01 -8.25815797e-01 -5.92682958e-01
1.96436018e-01 -1.22141016e+00 -6.76877081e-01 -3.49285662e-01
5.00168316e-02 1.09317136e+00 1.04475892e+00 -2.78895378e-01
7.15189099e-01 1.30877578e+00 -2.00237274e+00 -2.96055853e-01
7.17700660e-01 1.17090344e+00 6.09377027e-01 6.54530883e-01
-8.23576152e-01 -4.12593752e-01 -5.10771751e-01 4.07545596e-01
5.71300089e-01 2.69017279e-01 2.81340536e-03 4.27580178e-02
3.29433352e-01 -5.81384182e-01 3.54157299e-01 1.17341220e+00
4.83804733e-01 -1.78847507e-01 8.32727492e-01 -1.83083785e+00
-3.62544581e-02 -2.29441702e-01 -9.32043076e-01 2.97426492e-01
-9.57714543e-02 1.22497074e-01 1.12242997e-02 -1.10495150e+00
3.42915654e-01 1.58863544e+00 7.22067177e-01 1.22523598e-01
-5.95910788e-01 -1.20315409e+00 -2.70489782e-01 4.16676313e-01
-1.75325513e+00 5.44744097e-02 8.07381809e-01 -5.49755454e-01
9.81477678e-01 3.05942565e-01 7.91688859e-01 -1.22969195e-01
2.80671656e-01 3.82862151e-01 1.03210354e+00 -2.82195061e-01
1.78476304e-01 2.63891995e-01 4.57890302e-01 4.28535998e-01
5.26451409e-01 2.16602832e-01 -4.32722643e-02 -3.34717333e-02
8.65701139e-01 3.97626221e-01 2.07260713e-01 -3.85168999e-01
-1.03384817e+00 7.31124461e-01 5.34055948e-01 -1.11912735e-01
-8.51031393e-02 1.92458451e-01 3.30263078e-01 9.10454765e-02
8.01977292e-02 -3.04926485e-01 -2.35135764e-01 -6.92536011e-02
-7.45713949e-01 1.52080044e-01 2.54445612e-01 1.57800376e+00
1.21206152e+00 -1.75218493e-01 5.00845432e-01 3.39282751e-01
4.58801687e-01 1.16958153e+00 3.77191693e-01 -8.01710546e-01
3.57207239e-01 1.13577855e+00 2.08919868e-02 -1.03155267e+00
-3.88402760e-01 1.30946159e-01 -5.26391089e-01 1.02166367e+00
-1.71428695e-01 2.78828949e-01 -1.04147470e+00 4.46273535e-01
1.08959603e+00 3.27309161e-01 -9.23542231e-02 1.02940512e+00
1.22815561e+00 8.37588251e-01 -4.30062026e-01 5.66758066e-02
1.66194141e+00 -3.94890547e-01 -3.81757587e-01 -1.97391629e-01
6.63535237e-01 -8.80328417e-01 6.99011207e-01 1.44623429e-01
-6.32090926e-01 -9.03274596e-01 -1.20656824e+00 -3.40528011e-01
-5.10744989e-01 1.72382697e-01 1.88810572e-01 6.52461588e-01
-6.38339221e-01 -7.39120319e-02 -8.54030728e-01 -3.27983052e-01
4.34396625e-01 6.27016604e-01 -3.92503172e-01 -6.06642924e-02
-5.90720952e-01 1.03052592e+00 6.50550604e-01 -7.78365508e-02
-4.33432400e-01 -6.09972358e-01 -5.18090069e-01 -3.36218625e-01
2.54307449e-01 -5.52907169e-01 7.86544621e-01 -2.46445164e-01
-9.51835930e-01 1.21804368e+00 -4.26395833e-01 -5.00249267e-01
5.16192615e-01 -7.95102641e-02 -6.95924237e-02 1.12805866e-01
4.55567390e-01 7.90671051e-01 6.79739535e-01 -1.25022364e+00
-1.13508534e+00 -8.95974934e-01 -7.42431700e-01 3.63012671e-01
3.18905324e-01 -7.98479915e-02 -4.90043670e-01 5.09999275e-01
7.08125412e-01 -7.69616544e-01 -1.60004482e-01 2.95013964e-01
-2.08999902e-01 -6.28897548e-01 1.74622905e+00 1.33035421e-01
4.74623203e-01 -2.29246569e+00 -8.61867428e-01 3.05984288e-01
2.20301569e-01 3.11903715e-01 3.91178757e-01 5.05522676e-02
2.17524171e-01 6.32170513e-02 4.40507419e-02 -6.91632330e-02
-2.39539668e-01 2.88987517e-01 -2.91393071e-01 5.71577966e-01
2.28888050e-01 5.66640198e-01 -5.50331235e-01 -9.65980589e-01
7.24567473e-01 4.07700598e-01 -2.68734187e-01 -6.84213862e-02
1.62343904e-01 5.59553020e-02 -7.73306549e-01 7.64075696e-01
1.45263934e+00 5.10709763e-01 -8.83483112e-01 -1.51117712e-01
-7.24923193e-01 -7.36221224e-02 -1.55745077e+00 1.11423242e+00
-2.10630465e-02 6.14736140e-01 2.41611928e-01 -5.72367132e-01
1.70068455e+00 -2.80653864e-01 3.61322582e-01 -5.10137141e-01
4.84808147e-01 4.05967683e-01 -2.16766506e-01 -3.78566474e-01
5.71280718e-01 4.33280431e-02 1.91247672e-01 1.28799915e-01
-6.14795923e-01 -1.00982964e+00 -3.01031858e-01 7.66664336e-04
6.36261940e-01 -1.24617405e-01 -3.17160040e-02 -2.88117349e-01
5.72355807e-01 9.48681593e-01 3.80672038e-01 5.89471400e-01
-2.86826700e-01 4.40383255e-01 -2.81180558e-03 -2.78725296e-01
-1.15877366e+00 -8.42835963e-01 -7.30734289e-01 5.56587242e-02
9.47686791e-01 -1.49455918e-02 -4.95955884e-01 -1.37903169e-01
6.26893520e-01 4.81827140e-01 -5.72235622e-02 4.67349682e-03
-7.38625646e-01 -1.95009366e-01 1.34982318e-01 4.27417666e-01
7.80789852e-01 -6.28610611e-01 -1.22378552e+00 -3.33597586e-02
4.45472151e-01 -1.09606540e+00 -1.16324604e-01 2.16175646e-01
-1.34768105e+00 -1.01421523e+00 -2.21742950e-02 -8.76156807e-01
6.72492087e-01 9.46359158e-01 6.84156954e-01 2.91310221e-01
-4.94917363e-01 2.31880382e-01 -1.93405658e-01 -1.36171782e+00
-3.58469412e-02 -2.72595197e-01 2.77272135e-01 -5.28867900e-01
1.15087783e+00 -4.72820759e-01 -5.47141969e-01 6.17387891e-01
-3.89846176e-01 -1.58111244e-01 4.34059411e-01 -3.80194597e-02
1.07974005e+00 3.63269091e-01 -1.90708831e-01 -2.98103750e-01
2.03478962e-01 -1.98924437e-01 -1.27864981e+00 -6.18453920e-01
-5.55690706e-01 -4.80663031e-01 -5.74914441e-02 -2.36811191e-01
-2.50051558e-01 5.10038793e-01 -1.23442514e-04 -9.76896882e-01
-5.00446975e-01 -1.66202039e-01 -9.14958790e-02 -2.11959213e-01
5.94379544e-01 3.44043702e-01 2.56568521e-01 -4.61155951e-01
2.75009871e-01 1.17588329e+00 6.46832108e-01 -3.15970987e-01
1.17729497e+00 8.70430827e-01 4.36680585e-01 -1.12841582e+00
1.46702752e-01 -9.44454908e-01 -1.27707791e+00 -3.75117838e-01
1.00700974e+00 -1.07016683e+00 -1.10657871e+00 1.74295306e-01
-1.57265675e+00 1.92721769e-01 -1.28655359e-01 6.26200616e-01
-2.50058562e-01 2.51574636e-01 -5.28200641e-02 -1.16136849e+00
-3.38545889e-01 -1.24000525e+00 1.33664572e+00 1.26994208e-01
4.48918819e-01 -2.86449283e-01 -2.00943321e-01 1.81192949e-01
-2.48256743e-01 3.49187404e-01 6.08762562e-01 -1.72841981e-01
-1.09723353e+00 -5.95035613e-01 -5.02369761e-01 -2.34035868e-02
-1.94786102e-01 3.23118538e-01 -7.68278658e-01 8.98752660e-02
2.71764547e-01 3.80176097e-01 6.59452260e-01 1.96897641e-01
8.85651588e-01 3.80976170e-01 -9.75587130e-01 6.54670298e-01
1.57444310e+00 4.15126175e-01 5.31951666e-01 4.61671531e-01
6.49764240e-01 2.78549135e-01 1.31335330e+00 -5.82405441e-02
5.46131074e-01 5.21608472e-01 9.18008626e-01 -6.39982820e-02
-2.96555441e-02 -1.81313813e-01 2.10806593e-01 7.37945676e-01
-3.08893509e-02 3.26103032e-01 -1.29412961e+00 4.07351077e-01
-1.60269976e+00 -7.20135152e-01 -1.30004060e+00 2.21716523e+00
3.20215672e-01 5.26934147e-01 2.54722863e-01 3.85447919e-01
9.29356456e-01 -6.62974179e-01 -6.53704882e-01 -3.15527707e-01
2.45791301e-01 -1.06598362e-02 1.03927994e+00 4.88464028e-01
-7.68400133e-01 1.05926096e+00 4.91428137e+00 8.94096613e-01
-1.20524883e+00 1.27857566e-01 -1.24135679e-02 9.83621925e-02
1.48482546e-01 8.26763287e-02 -1.67854881e+00 5.34578621e-01
3.89295429e-01 -1.27055466e-01 -3.65190834e-01 1.31748712e+00
4.00940776e-01 -4.32638466e-01 -7.77515709e-01 1.52031350e+00
-9.13152993e-02 -1.03321886e+00 5.13180457e-02 4.21193242e-01
1.54620320e-01 4.05359626e-01 -3.15730304e-01 6.98666647e-02
-5.45575172e-02 -7.17013538e-01 9.63496566e-01 1.41972110e-01
9.59107518e-01 -8.81778479e-01 6.43271685e-01 1.07036793e+00
-1.53082979e+00 1.90146133e-01 -1.13303268e+00 -2.47889608e-01
6.28331304e-02 8.35074127e-01 -1.29783154e+00 2.95333862e-01
9.59372520e-01 4.94447947e-01 -3.79701793e-01 1.37497926e+00
1.06099851e-01 1.98937356e-01 -9.68500853e-01 -3.42358917e-01
2.53049403e-01 -6.71191812e-01 7.73453116e-01 1.01885819e+00
6.23690188e-01 4.58471298e-01 3.48670453e-01 1.00834835e+00
3.49945724e-01 1.83320969e-01 -8.96976829e-01 5.97595751e-01
1.09119749e+00 1.19805622e+00 -1.04058397e+00 -3.68574113e-01
-3.12884785e-02 3.30710411e-01 -2.29524523e-01 -3.60839307e-01
-7.38817155e-01 -8.28090191e-01 6.57445610e-01 6.91052735e-01
3.48312050e-01 -9.87020373e-01 -7.67107427e-01 -4.57968444e-01
-4.51421365e-02 3.90175395e-02 -1.33826196e-01 -1.17296624e+00
-6.06167257e-01 3.24589759e-01 1.91224247e-01 -1.54451156e+00
3.62384588e-01 -8.44054997e-01 -6.58464730e-01 9.97536421e-01
-1.75133777e+00 -9.17801321e-01 -7.96202123e-01 7.59710133e-01
4.94155467e-01 1.78280815e-01 3.21153194e-01 4.44558889e-01
-1.81387275e-01 -1.95109084e-01 -5.84206544e-02 8.07221308e-02
3.15101087e-01 -8.85897160e-01 3.98546755e-01 4.83336478e-01
-2.64234483e-01 2.18641162e-01 3.14118087e-01 -1.14859545e+00
-1.60478604e+00 -1.10870969e+00 3.80304575e-01 -7.93253541e-01
1.22178242e-01 -4.58343893e-01 -8.45180452e-01 5.44556856e-01
-4.15920109e-01 6.45143762e-02 3.92926037e-02 -5.15753567e-01
1.19798601e-01 -4.02687937e-01 -1.43507004e+00 1.34740353e-01
1.02180254e+00 -6.55090362e-02 -5.89374363e-01 4.62041855e-01
7.11237848e-01 -9.10962343e-01 -4.88398463e-01 5.40859938e-01
2.66174287e-01 -8.50881338e-01 1.26037681e+00 -1.05041593e-01
-1.06385067e-01 -1.07685339e+00 1.45007484e-02 -3.76752049e-01
-1.77608635e-02 1.73857920e-02 4.83382314e-01 1.01672566e+00
-9.65349898e-02 -6.86665833e-01 1.01783288e+00 2.23462954e-01
-5.21890342e-01 -2.24756360e-01 -1.24015129e+00 -8.76331031e-01
-4.80855480e-02 -9.42257166e-01 1.00701869e+00 8.41415584e-01
-6.74256027e-01 4.83627617e-01 5.19214213e-01 7.02589571e-01
7.36488163e-01 3.28886360e-01 1.48669827e+00 -1.70067096e+00
8.04152966e-01 -4.94735539e-01 -1.08107519e+00 -1.31704152e+00
-2.32537046e-01 -9.84349608e-01 -1.25727460e-01 -1.60130906e+00
-2.98855454e-01 -1.04612756e+00 5.48452020e-01 1.44293711e-01
2.87620783e-01 3.39291453e-01 1.74245581e-01 7.58458495e-01
-1.20869339e-01 3.03670377e-01 1.20918441e+00 1.57150328e-02
-3.19404602e-01 1.06369495e-01 1.96946681e-01 6.94826484e-01
6.23269200e-01 -6.56930685e-01 -1.38911322e-01 -6.19125426e-01
-3.81686166e-02 -2.66515277e-02 4.53580528e-01 -1.27224243e+00
5.67779303e-01 -2.73115009e-01 5.96315682e-01 -1.84067559e+00
7.61446774e-01 -1.38792682e+00 -5.31793945e-02 8.74753416e-01
8.21202517e-01 1.58116236e-01 4.44639653e-01 3.71006697e-01
6.31124079e-02 -4.99104738e-01 8.99342060e-01 -4.08357829e-01
-8.08153331e-01 3.91609430e-01 -2.41149008e-01 -5.32815278e-01
1.56597006e+00 -1.19335163e+00 -1.20990150e-01 1.45226702e-01
-3.06928694e-01 6.11784577e-01 5.82424760e-01 1.34905055e-01
1.04493713e+00 -1.41563511e+00 -6.74905360e-01 5.45256793e-01
1.06312193e-01 8.58431637e-01 -2.06462711e-01 6.46875143e-01
-1.29375446e+00 3.82490009e-01 -1.20618612e-01 -1.61623406e+00
-1.64891243e+00 5.72687328e-01 4.35029358e-01 8.79634440e-01
-7.35547483e-01 7.10105062e-01 -1.56135425e-01 -6.21094525e-01
-1.75519049e-01 -6.75130010e-01 -1.71841398e-01 -9.38473940e-02
2.94329524e-01 5.56693077e-01 1.74200431e-01 -8.96215916e-01
-6.44335747e-01 1.43287885e+00 3.16508740e-01 6.30931975e-03
8.96871984e-01 -1.17246404e-01 -2.14187995e-01 4.99204785e-01
1.06256783e+00 -4.87075150e-02 -7.64703631e-01 1.04142226e-01
-2.32832119e-01 -8.71178806e-01 3.01734626e-01 -7.73489177e-02
-6.37069285e-01 1.35467684e+00 1.14063334e+00 2.16325507e-01
5.73353648e-01 9.09028202e-03 8.76939714e-01 5.09547234e-01
6.74487293e-01 -9.18165863e-01 -4.05391872e-01 5.83401442e-01
5.68611383e-01 -1.29422069e+00 3.05945337e-01 -7.41360605e-01
-1.26269519e-01 1.44583523e+00 7.19566584e-01 -5.01144111e-01
8.16797018e-01 4.62313890e-01 1.96039572e-01 -5.90231597e-01
-7.24317804e-02 -4.15956289e-01 -9.65665728e-02 9.24090564e-01
-5.07675350e-01 6.83942959e-02 -4.22366485e-02 1.59811035e-01
-7.18111515e-01 -1.99560389e-01 3.49376887e-01 9.95694220e-01
-1.48234236e+00 -6.33525789e-01 -1.22349286e+00 4.55030888e-01
3.93123537e-01 3.01382154e-01 -2.18670413e-01 9.00189281e-01
5.70167959e-01 7.65041888e-01 8.39155972e-01 -3.04530293e-01
5.72492182e-01 -6.94609359e-02 2.22376302e-01 -7.04247296e-01
-2.39812620e-02 -8.40855762e-03 -5.78201234e-01 -3.54011863e-01
1.27646670e-01 -6.15770042e-01 -1.93976855e+00 -3.23053718e-01
-6.67647958e-01 2.21528843e-01 1.57333720e+00 4.94830579e-01
4.46820498e-01 -8.79520550e-02 7.12879419e-01 -9.02640462e-01
-1.16170458e-01 -5.89336932e-01 -5.54860473e-01 5.97106945e-03
1.63799405e-01 -7.56366372e-01 -3.25580060e-01 -2.57934332e-01] | [7.7261061668396, -2.635901927947998] |
91068085-b6ba-4dc5-a731-eb133a0fdcb9 | systematic-analysis-of-the-impact-of-label | 2306.15994 | null | https://arxiv.org/abs/2306.15994v1 | https://arxiv.org/pdf/2306.15994v1.pdf | Systematic analysis of the impact of label noise correction on ML Fairness | Arbitrary, inconsistent, or faulty decision-making raises serious concerns, and preventing unfair models is an increasingly important challenge in Machine Learning. Data often reflect past discriminatory behavior, and models trained on such data may reflect bias on sensitive attributes, such as gender, race, or age. One approach to developing fair models is to preprocess the training data to remove the underlying biases while preserving the relevant information, for example, by correcting biased labels. While multiple label noise correction methods are available, the information about their behavior in identifying discrimination is very limited. In this work, we develop an empirical methodology to systematically evaluate the effectiveness of label noise correction techniques in ensuring the fairness of models trained on biased datasets. Our methodology involves manipulating the amount of label noise and can be used with fairness benchmarks but also with standard ML datasets. We apply the methodology to analyze six label noise correction methods according to several fairness metrics on standard OpenML datasets. Our results suggest that the Hybrid Label Noise Correction method achieves the best trade-off between predictive performance and fairness. Clustering-Based Correction can reduce discrimination the most, however, at the cost of lower predictive performance. | ['R. Ghani', 'I. Sousa', 'C. Soares', 'I. Oliveira e Silva'] | 2023-06-28 | null | null | null | null | ['fairness', 'fairness', 'decision-making'] | ['computer-vision', 'miscellaneous', 'reasoning'] | [ 2.48491541e-01 1.40476480e-01 -6.04046345e-01 -8.72686744e-01
-6.66304708e-01 -4.75478858e-01 4.72258717e-01 7.10566938e-01
-6.80441797e-01 9.70018089e-01 1.62898347e-01 -2.71952540e-01
-1.96683675e-01 -7.04833567e-01 -3.15273970e-01 -4.80037481e-01
4.88463104e-01 5.03054440e-01 -2.48165086e-01 1.95470765e-01
6.17678285e-01 2.93514818e-01 -1.50646591e+00 1.69642165e-01
1.36820412e+00 7.88401842e-01 -6.88239574e-01 2.78269529e-01
-1.56878859e-01 1.06281662e+00 -5.87888837e-01 -1.13755357e+00
2.87920266e-01 -5.05936384e-01 -8.34186435e-01 -4.85998064e-01
7.33537734e-01 -2.22051337e-01 2.96246648e-01 1.38480699e+00
5.38518846e-01 -5.07385246e-02 9.27139401e-01 -1.45514524e+00
-6.94347501e-01 9.52222288e-01 -6.45033360e-01 9.17525589e-02
1.83121639e-03 -7.48762190e-02 1.00206745e+00 -3.08265299e-01
5.07973433e-01 1.38824201e+00 7.88209379e-01 8.51367712e-01
-1.59172058e+00 -1.04611290e+00 -2.24517033e-01 8.64064470e-02
-1.20895994e+00 -7.56041169e-01 4.83255565e-01 -6.10005796e-01
1.35461101e-02 6.47720218e-01 1.66551739e-01 1.07009113e+00
1.14686400e-01 1.66318908e-01 1.57848072e+00 -4.79476064e-01
3.53995234e-01 4.18195635e-01 6.01902366e-01 5.29013872e-01
7.75461137e-01 2.88216710e-01 -3.06932271e-01 -6.56614482e-01
9.33180451e-02 -1.25879467e-01 7.93521404e-02 -3.31474781e-01
-7.34466434e-01 9.25682843e-01 2.38106310e-01 9.00430009e-02
6.31035417e-02 -1.35277873e-02 3.77401590e-01 2.92845279e-01
7.82805264e-01 7.27882206e-01 -2.99034417e-01 1.12047484e-02
-1.12894344e+00 4.51309562e-01 6.96367502e-01 3.63466829e-01
7.10229218e-01 -3.33353847e-01 -6.40567541e-01 1.06739032e+00
1.26844674e-01 3.60282362e-01 4.32456493e-01 -1.43079972e+00
2.68511504e-01 6.38212740e-01 2.53527790e-01 -1.12132931e+00
-3.83560568e-01 -4.07860100e-01 -8.70864689e-01 2.92308390e-01
8.95658433e-01 -1.13557033e-01 -6.38771236e-01 1.98916626e+00
2.95521736e-01 -3.92263591e-01 -2.70201653e-01 7.59621859e-01
4.62690592e-01 1.24298811e-01 6.89916253e-01 -3.23761284e-01
1.01626945e+00 -4.74092901e-01 -7.82684863e-01 -9.41459015e-02
9.17678297e-01 -7.41210759e-01 8.42422962e-01 1.23886272e-01
-8.62473130e-01 -2.48907298e-01 -5.71290672e-01 -2.11764351e-01
-1.78870276e-01 -5.74665107e-02 5.55664837e-01 1.12333953e+00
-5.52755773e-01 7.93998837e-01 -2.23111823e-01 -2.45016932e-01
6.53399050e-01 8.28402638e-02 -1.50463045e-01 -1.49795145e-01
-1.14468122e+00 1.07886481e+00 1.66680470e-01 -3.95716459e-01
-3.89793336e-01 -9.43241298e-01 -6.36820376e-01 1.64146170e-01
2.97851235e-01 -4.98518735e-01 1.01297367e+00 -1.42668521e+00
-8.80790651e-01 1.22566605e+00 -3.67028117e-02 -3.66279602e-01
1.02356684e+00 1.27170861e-01 -3.55357260e-01 -5.50862968e-01
3.90941590e-01 5.12013853e-01 5.40504515e-01 -1.43824565e+00
-6.16644979e-01 -6.70387447e-01 -1.02075852e-01 -1.04421778e-02
-2.35647038e-01 3.37482363e-01 4.34127092e-01 -6.16501272e-01
-8.62536281e-02 -8.76848459e-01 -2.93707073e-01 2.68708598e-02
-3.76922995e-01 -4.51214388e-02 1.94290370e-01 -6.12759113e-01
1.46777976e+00 -2.17054439e+00 -3.81433755e-01 4.34186816e-01
2.84541458e-01 2.44904026e-01 1.17974304e-01 -2.75394380e-01
-3.34603451e-02 6.47466481e-01 -4.11594689e-01 -2.22097874e-01
6.90954551e-02 1.40191749e-01 -6.15950078e-02 4.67469037e-01
-2.33059391e-01 5.66084445e-01 -7.07266033e-01 -8.15309346e-01
-1.19573712e-01 1.40233442e-01 -7.29847252e-01 -3.50453635e-03
9.20463055e-02 2.96750993e-01 -6.38299361e-02 5.14249802e-01
7.96023071e-01 5.52416630e-02 4.24475342e-01 3.77396084e-02
1.61010474e-02 2.53498197e-01 -9.93050456e-01 9.76990521e-01
-1.30723715e-01 3.72803569e-01 5.64031862e-02 -7.83439815e-01
1.03045344e+00 -1.09402828e-01 2.58491755e-01 -7.51689196e-01
2.56826818e-01 3.67411017e-01 2.54606575e-01 -3.65651667e-01
4.95007694e-01 -4.01750207e-01 -2.28512749e-01 7.02956378e-01
-4.17694390e-01 2.11019218e-01 5.50802797e-02 8.55526552e-02
6.18050635e-01 -3.93605560e-01 4.97408271e-01 -5.61800182e-01
3.53750616e-01 2.92119961e-02 1.01342249e+00 8.49299312e-01
-7.31889784e-01 3.79629403e-01 7.74226308e-01 -3.95424008e-01
-9.52810049e-01 -6.82688653e-01 -5.10487199e-01 1.33989978e+00
-6.59534484e-02 -2.22468898e-01 -9.12967741e-01 -9.64669406e-01
4.06168222e-01 1.09099448e+00 -9.07310247e-01 -4.77068305e-01
-2.05940790e-02 -1.02385437e+00 9.27281678e-01 2.15222895e-01
3.81447911e-01 -5.02332926e-01 -4.27849472e-01 -2.05853522e-01
-3.22682321e-01 -5.74625075e-01 -3.57510984e-01 -2.15397719e-02
-8.50009024e-01 -1.28427839e+00 -3.22313637e-01 -1.97213188e-01
7.68450201e-01 -6.23349063e-02 1.23271418e+00 4.26430941e-01
-6.88616559e-02 -2.39292279e-01 -1.60899445e-01 -5.08176506e-01
-7.60673106e-01 -5.37810810e-02 -1.18525857e-02 8.45496655e-02
6.63381040e-01 -9.54503268e-02 -3.28918844e-01 4.06389654e-01
-7.67956913e-01 -2.04871267e-01 2.55541086e-01 7.01489568e-01
3.40706527e-01 -1.13645837e-01 6.08809233e-01 -1.65443659e+00
8.34106088e-01 -4.72876042e-01 -3.84110749e-01 4.92778599e-01
-1.31439936e+00 1.55617952e-01 4.86148179e-01 -2.05230147e-01
-1.22152829e+00 -3.36818427e-01 7.37061277e-02 9.84572470e-02
-2.91632742e-01 2.37436399e-01 -2.55092114e-01 -1.33887753e-01
1.07351196e+00 -6.40047848e-01 3.74424011e-02 -4.56513643e-01
1.74877763e-01 8.74168158e-01 1.73530728e-01 -7.77565598e-01
2.85166234e-01 2.82478660e-01 1.07844602e-02 -6.27346262e-02
-1.22713375e+00 -9.02762264e-02 -5.09164333e-01 -1.34027496e-01
4.08150584e-01 -4.53891844e-01 -6.13234222e-01 2.69448042e-01
-6.87988162e-01 -1.46310493e-01 -2.82744288e-01 3.72553915e-01
-1.01464793e-01 3.10627878e-01 -5.68585992e-01 -8.41524720e-01
-2.96401381e-01 -8.80029440e-01 5.00137329e-01 1.43936515e-01
-7.14994252e-01 -8.81965280e-01 2.21459821e-01 8.00340354e-01
4.67164338e-01 2.91199595e-01 1.27042484e+00 -8.61231804e-01
1.09692767e-01 -1.97629362e-01 -4.92362767e-01 3.33686709e-01
6.66159764e-02 3.12261045e-01 -1.06558716e+00 -1.01306155e-01
-1.98725417e-01 -3.75777602e-01 8.77580881e-01 4.32955921e-01
1.47659028e+00 -4.35052842e-01 -1.05109259e-01 4.05625105e-01
1.36040068e+00 1.21819880e-02 4.37940240e-01 2.07789481e-01
7.45586097e-01 1.06812716e+00 5.94501555e-01 4.51876342e-01
4.54986513e-01 5.37583649e-01 1.42606154e-01 1.15573972e-01
1.03371076e-01 -2.39762783e-01 -8.20555091e-02 2.25090489e-01
-1.88293487e-01 1.86565429e-01 -1.03466761e+00 1.27949998e-01
-1.76728225e+00 -1.00261164e+00 -3.20836455e-01 2.50284696e+00
1.05524147e+00 -8.95215292e-03 1.62499249e-01 3.64733875e-01
9.99604046e-01 -8.70128870e-02 -4.26715732e-01 -8.17757666e-01
-4.99859825e-02 -6.28532767e-02 6.42554224e-01 6.03992760e-01
-8.85867715e-01 5.51489651e-01 6.71552944e+00 7.70859540e-01
-8.75061810e-01 1.70905441e-01 1.21378386e+00 -2.13501751e-01
-5.96347272e-01 9.33574215e-02 -4.27595049e-01 7.30949521e-01
8.95726323e-01 -4.75150943e-01 3.09481531e-01 8.29181015e-01
3.34463000e-01 -3.10252160e-01 -1.10639489e+00 7.36639321e-01
6.91917762e-02 -7.78221726e-01 7.09386766e-02 1.64031684e-01
8.75923097e-01 -6.85257375e-01 1.53641393e-02 3.06349069e-01
6.26720846e-01 -1.16249776e+00 7.51911581e-01 5.41611254e-01
7.05434620e-01 -9.99136448e-01 9.22855139e-01 3.92811656e-01
-1.93837225e-01 -3.00142497e-01 -4.18456703e-01 -3.38008255e-01
-4.00677979e-01 1.16586328e+00 -3.74455452e-01 1.96787760e-01
5.88973761e-01 3.96877378e-01 -8.36102009e-01 8.77887905e-01
-6.00664280e-02 7.09074795e-01 1.58138081e-01 1.70445561e-01
-2.99578071e-01 -1.45443425e-01 1.25792474e-02 1.02303481e+00
1.70551628e-01 -1.11085828e-02 -1.61245629e-01 9.53665078e-01
-3.86199743e-01 4.51086968e-01 -4.99856830e-01 2.32908338e-01
6.39175534e-01 1.09641337e+00 -5.30780613e-01 -2.44216815e-01
-1.44530773e-01 4.32310343e-01 5.10024548e-01 9.48254615e-02
-7.35967159e-01 5.31223305e-02 7.10439920e-01 2.84079432e-01
-6.75502300e-01 3.68747622e-01 -9.62324142e-01 -9.75803018e-01
-3.71930659e-01 -1.08138096e+00 7.54659891e-01 -3.25148493e-01
-1.55875480e+00 -6.57884078e-03 -1.88831970e-01 -7.19789565e-01
4.80133817e-02 -2.33705565e-01 -2.38031134e-01 9.58237231e-01
-1.25356555e+00 -8.40688288e-01 -3.59733909e-01 2.77244955e-01
-1.43909916e-01 2.14724299e-02 7.30423927e-01 4.68139470e-01
-7.21390843e-01 9.21746314e-01 8.67807865e-02 2.01410741e-01
1.32408190e+00 -1.15663421e+00 -1.22020371e-01 6.86134934e-01
-1.68277800e-01 6.35434628e-01 6.47615790e-01 -7.20196426e-01
-2.97909290e-01 -1.09667242e+00 1.31421292e+00 -5.19902170e-01
1.50222480e-01 -5.24393544e-02 -1.03057253e+00 6.14756048e-01
-3.09509844e-01 -5.93985766e-02 1.25472832e+00 4.78493959e-01
-7.71344244e-01 -3.42946202e-01 -1.70734441e+00 3.91253948e-01
9.92654622e-01 -4.34717804e-01 -3.87298316e-01 -6.24495223e-02
1.24487616e-01 -2.02991694e-01 -7.96461344e-01 3.91666323e-01
7.38623202e-01 -1.27963936e+00 6.13648772e-01 -9.29182172e-01
7.15241432e-01 4.31890339e-02 -1.82270244e-01 -1.48587668e+00
-6.74817622e-01 -3.10446485e-03 4.09528375e-01 1.63861752e+00
6.16573155e-01 -6.32201493e-01 7.35546470e-01 1.51648998e+00
5.18667102e-01 -3.25470090e-01 -7.81878948e-01 -4.35180575e-01
4.63997811e-01 -2.24989578e-01 8.79205585e-01 1.57366836e+00
-1.20378949e-01 1.05739318e-01 -5.79104483e-01 -2.18353122e-01
1.04767060e+00 1.57626897e-01 6.49938524e-01 -1.62886143e+00
1.28557637e-01 -6.00738287e-01 -2.46210426e-01 2.21514061e-01
6.18831336e-01 -1.03298378e+00 -1.17646709e-01 -1.00520360e+00
7.39971101e-01 -8.93618762e-01 -2.63417095e-01 5.47564209e-01
-6.39077187e-01 2.04208374e-01 2.37667188e-01 2.51034737e-01
-3.14149678e-01 1.22185148e-01 9.90924776e-01 -1.69295982e-01
1.16456911e-01 7.76696280e-02 -1.22177267e+00 7.44626760e-01
9.45641160e-01 -8.94748986e-01 -2.08576471e-01 -2.50098765e-01
2.75231779e-01 -3.21455359e-01 4.67775285e-01 -8.16805243e-01
4.65367883e-02 -5.81327140e-01 4.78675932e-01 3.43125582e-01
-2.64902025e-01 -7.86818683e-01 3.08260113e-01 4.38618869e-01
-1.02851069e+00 -1.44091114e-01 -2.20110714e-01 3.64318520e-01
2.11400539e-02 -3.99372965e-01 1.18410218e+00 -1.30734026e-01
-2.93439448e-01 4.13847081e-02 -4.29710299e-02 4.65278119e-01
9.41491187e-01 8.82070065e-02 -7.54179955e-01 -1.95570201e-01
-5.48055351e-01 1.03934906e-01 9.12426770e-01 2.32173011e-01
3.05528678e-02 -1.29243338e+00 -9.38478708e-01 1.05026506e-01
2.65599549e-01 -6.89729929e-01 4.21053022e-02 6.18406236e-01
-2.77683675e-01 8.67044702e-02 -5.03274441e-01 -1.83168426e-01
-1.51849568e+00 5.18841565e-01 4.60596174e-01 -1.72271997e-01
1.72091290e-01 5.77090502e-01 -5.88424504e-02 -6.44369781e-01
1.70238808e-01 5.87909259e-02 -2.90722400e-01 3.23852211e-01
3.93205702e-01 9.05217290e-01 2.78879572e-02 -7.39218116e-01
-3.46405745e-01 3.14968526e-01 6.49239421e-02 7.43163452e-02
8.50076556e-01 -2.34680206e-01 -4.98499960e-01 3.95785064e-01
8.56506348e-01 2.80991733e-01 -6.93139076e-01 -6.11427315e-02
2.17547253e-01 -1.02639401e+00 -6.78587100e-03 -1.07781518e+00
-1.04384196e+00 6.46115243e-01 5.93215883e-01 2.82692492e-01
7.96234369e-01 -5.06590724e-01 1.53759003e-01 -1.44481257e-01
3.78992021e-01 -1.33988023e+00 -5.24733841e-01 1.25210434e-01
3.31963480e-01 -1.55221236e+00 1.39483318e-01 -4.76688623e-01
-6.93841398e-01 6.56341493e-01 6.86998069e-01 2.82048494e-01
5.60161948e-01 1.28003091e-01 4.31508303e-01 8.83479565e-02
-5.37151277e-01 1.06883168e-01 2.29910031e-01 4.24908251e-01
8.22537303e-01 6.50759816e-01 -1.01528454e+00 8.35858166e-01
-2.72780299e-01 7.48168305e-02 5.45735240e-01 5.13308108e-01
-2.08297029e-01 -1.15482891e+00 -5.85300803e-01 1.04581320e+00
-9.42169070e-01 -5.71919885e-03 -8.44527125e-01 3.76929760e-01
4.52898055e-01 1.08106411e+00 5.80971204e-02 -3.16244245e-01
2.56421000e-01 2.56171674e-01 2.40802333e-01 -4.12572891e-01
-7.97190964e-01 -6.63114607e-01 4.21867311e-01 -5.84098756e-01
-6.54547036e-01 -7.72883236e-01 -7.92509139e-01 -1.08680534e+00
-1.31471232e-01 3.24603409e-01 4.43731815e-01 8.59024465e-01
2.82247514e-01 2.84350049e-02 4.93718266e-01 -1.22635692e-01
-8.37952197e-01 -7.17903554e-01 -7.18635798e-01 9.98629808e-01
1.47499278e-01 -7.07025468e-01 -5.96414745e-01 -2.81722784e-01] | [8.885669708251953, 5.362449645996094] |
8a91177f-c9a2-4456-b3ff-c899b135c009 | accelerated-primal-dual-methods-with-enlarged | 2307.00296 | null | https://arxiv.org/abs/2307.00296v1 | https://arxiv.org/pdf/2307.00296v1.pdf | Accelerated primal-dual methods with enlarged step sizes and operator learning for nonsmooth optimal control problems | We consider a general class of nonsmooth optimal control problems with partial differential equation (PDE) constraints, which are very challenging due to its nonsmooth objective functionals and the resulting high-dimensional and ill-conditioned systems after discretization. We focus on the application of a primal-dual method, with which different types of variables can be treated individually and thus its main computation at each iteration only requires solving two PDEs. Our target is to accelerate the primal-dual method with either larger step sizes or operator learning techniques. For the accelerated primal-dual method with larger step sizes, its convergence can be still proved rigorously while it numerically accelerates the original primal-dual method in a simple and universal way. For the operator learning acceleration, we construct deep neural network surrogate models for the involved PDEs. Once a neural operator is learned, solving a PDE requires only a forward pass of the neural network, and the computational cost is thus substantially reduced. The accelerated primal-dual method with operator learning is mesh-free, numerically efficient, and scalable to different types of PDEs. The acceleration effectiveness of these two techniques is promisingly validated by some preliminary numerical results. | ['Hangrui Yue', 'Xiaoming Yuan', 'Yongcun Song'] | 2023-07-01 | null | null | null | null | ['operator-learning'] | ['miscellaneous'] | [-2.42457256e-01 -8.44085123e-03 5.92109933e-02 3.09245795e-01
-8.43325615e-01 -2.56965935e-01 2.22982504e-04 8.69219527e-02
-2.39825845e-01 1.06982505e+00 -2.81818390e-01 -3.23204637e-01
-2.19734579e-01 -6.63901687e-01 -7.83706248e-01 -8.90832484e-01
-1.87617078e-01 6.99656844e-01 -1.13972075e-01 -4.84179705e-01
1.56718895e-01 5.74636400e-01 -1.15961325e+00 -2.30713546e-01
1.07090044e+00 1.52452147e+00 -6.90281540e-02 7.92021871e-01
1.75582260e-01 6.28055930e-01 3.06284010e-01 1.96335405e-01
6.11529112e-01 -3.46690118e-01 -9.29058373e-01 2.82959193e-01
2.47202396e-01 -2.84183055e-01 -1.28049135e-01 1.16305625e+00
6.02749050e-01 5.34424782e-01 6.06250465e-01 -1.10708380e+00
-7.48918951e-01 -5.36081195e-02 -7.09711969e-01 -5.30552901e-02
4.08595353e-02 1.71583146e-01 1.04721034e+00 -1.32013476e+00
3.25358868e-01 1.19009042e+00 1.15422928e+00 4.57459301e-01
-1.60117030e+00 -2.64947355e-01 -7.05893189e-02 -5.96654527e-02
-1.32848370e+00 -1.95480250e-02 7.94275165e-01 -7.31427073e-01
4.53427464e-01 3.17751735e-01 8.24479222e-01 4.95993972e-01
2.45466232e-01 6.89998567e-01 1.12699616e+00 -1.64608747e-01
5.24268806e-01 2.15748817e-01 8.45593587e-02 9.86183465e-01
7.24454923e-03 3.64081562e-02 1.31837785e-01 -4.53709096e-01
9.59922731e-01 -8.54148567e-02 -5.19457757e-01 -6.06097162e-01
-1.06692874e+00 1.32798290e+00 3.64254624e-01 1.41030803e-01
-6.38032794e-01 3.75004709e-02 6.12335920e-01 2.64807284e-01
8.72542918e-01 5.73239326e-01 -5.27546883e-01 1.33961171e-01
-7.14212775e-01 6.64808631e-01 1.09667027e+00 7.56697297e-01
8.14736664e-01 2.05097690e-01 -3.46370526e-02 8.58416200e-01
1.70368955e-01 5.14488518e-01 3.63211147e-02 -1.47464120e+00
1.37846157e-01 3.35754097e-01 5.16650200e-01 -1.19523871e+00
-4.84946281e-01 -2.90045857e-01 -1.41260767e+00 7.27051198e-01
5.64615905e-01 -5.49891233e-01 -5.53296447e-01 1.48572171e+00
7.31607378e-01 1.85752675e-01 7.50319660e-02 1.10878241e+00
4.30597141e-02 1.06802833e+00 -3.02351147e-01 -5.70617557e-01
1.05271745e+00 -1.07603049e+00 -7.38883913e-01 6.42214045e-02
9.80153143e-01 -6.26535892e-01 9.45450664e-01 4.50885534e-01
-1.44233251e+00 -4.92844790e-01 -8.65776300e-01 -3.23407114e-01
-8.83999914e-02 2.74247527e-01 5.49263299e-01 -2.38056239e-02
-1.06963861e+00 1.15169132e+00 -8.12580585e-01 2.61379659e-01
3.30112517e-01 3.96654755e-01 -1.33818612e-01 2.11747721e-01
-1.24257267e+00 8.21045578e-01 2.16855064e-01 4.02810276e-01
-6.01336896e-01 -1.43110383e+00 -8.59715343e-01 9.67384577e-02
3.19276005e-01 -8.33246708e-01 9.52841938e-01 -8.66526544e-01
-1.72569335e+00 6.68671191e-01 5.01909805e-03 -2.12318048e-01
8.77049327e-01 3.04518491e-02 1.96134806e-01 1.06134251e-01
3.14890862e-01 1.12502687e-01 9.54717636e-01 -1.19443250e+00
-6.27396703e-01 -3.21409822e-01 9.38451663e-02 8.87009278e-02
-2.71495908e-01 -4.01716471e-01 2.44895536e-02 -7.07298219e-01
-6.18960075e-02 -1.25882220e+00 -6.94030225e-01 6.59171402e-01
-1.40038515e-02 -2.13655263e-01 8.64068985e-01 -8.25613737e-01
9.89537716e-01 -2.14618349e+00 6.10898614e-01 -2.23332029e-02
3.56475234e-01 3.14151645e-01 5.78161217e-02 2.25120917e-01
-1.04871079e-01 -1.13296516e-01 -8.41643453e-01 -3.94961566e-01
-9.61615071e-02 3.19162220e-01 -2.60282069e-01 8.61056209e-01
2.85169184e-01 5.81582963e-01 -1.02539492e+00 -3.68018627e-01
2.48495340e-01 5.60166955e-01 -9.99779582e-01 2.44657725e-01
-1.59997299e-01 7.94596851e-01 -6.36603713e-01 2.80086368e-01
1.02897954e+00 -3.41886550e-01 -1.19230464e-01 -2.64847547e-01
-5.03610909e-01 -2.70068884e-01 -1.68976045e+00 1.71648622e+00
-1.06269169e+00 4.44185764e-01 1.00403118e+00 -1.87358546e+00
6.00228906e-01 4.05853480e-01 6.84764504e-01 -3.58647734e-01
1.36632621e-01 8.42477262e-01 -3.66581500e-01 -6.38649046e-01
1.33007020e-01 -5.79556942e-01 1.46416798e-01 1.88744307e-01
-3.09308738e-01 -3.12192827e-01 4.29093093e-01 -1.38271660e-01
6.77718461e-01 -8.78006890e-02 5.45933135e-02 -1.02854264e+00
1.04487562e+00 2.39296019e-01 8.34253669e-01 3.00752163e-01
-3.80374976e-02 4.55526888e-01 7.52767682e-01 -7.71596730e-01
-1.26227140e+00 -5.86073935e-01 -5.68878055e-01 8.11195552e-01
1.14840038e-01 2.03032434e-01 -5.16385317e-01 -3.60083282e-01
3.27187926e-01 4.29151744e-01 -7.00612247e-01 -7.07202181e-02
-8.57239187e-01 -7.62920141e-01 3.69888321e-02 6.51323259e-01
5.05784571e-01 -5.57338417e-01 -1.79859146e-01 3.85790020e-01
1.92980528e-01 -9.00657594e-01 -5.60717046e-01 9.99828801e-02
-1.03219867e+00 -1.15461183e+00 -1.20894051e+00 -1.01319039e+00
7.25693941e-01 -1.35391712e-01 8.04603696e-01 1.34861879e-02
-1.13470711e-01 3.33673656e-01 2.30741903e-01 -2.30764467e-02
-4.45548892e-01 -2.23093212e-01 2.90235668e-01 4.80481356e-01
-6.73895419e-01 -6.56796932e-01 -5.64962864e-01 1.64486796e-01
-7.71000266e-01 -1.26280328e-02 1.88705146e-01 1.36463785e+00
7.32467890e-01 8.81781876e-02 4.09029186e-01 -8.59395087e-01
6.54884040e-01 -5.58521748e-01 -1.13314378e+00 -7.64198601e-02
-4.55995053e-01 2.23871246e-01 1.26633942e+00 -5.56794643e-01
-9.85335112e-01 1.94409296e-01 -1.78163067e-01 -8.14418316e-01
4.22645599e-01 6.22058034e-01 2.66512632e-01 -5.91016352e-01
6.39196336e-01 3.07626091e-02 3.25840235e-01 -5.77315688e-01
4.43476886e-02 2.22206280e-01 2.75093794e-01 -9.13257360e-01
6.38271093e-01 6.58413351e-01 6.75336123e-01 -1.05314136e+00
-9.15851057e-01 -3.76351148e-01 -5.07285476e-01 -4.14914778e-03
7.47353554e-01 -6.22973084e-01 -1.15735543e+00 5.40688038e-01
-1.41392457e+00 -6.29603863e-01 -4.97673839e-01 4.45621729e-01
-7.30151236e-01 4.02491182e-01 -8.69082987e-01 -8.92163157e-01
-1.51484340e-01 -1.25977969e+00 1.02036107e+00 -1.91502959e-01
1.63488045e-01 -1.78036761e+00 1.46730095e-01 -4.08468163e-03
4.26685482e-01 6.07584357e-01 8.65239859e-01 1.08532114e-02
-1.09746590e-01 -1.83623686e-01 -3.50846171e-01 7.87857056e-01
-1.56701595e-01 -9.80813280e-02 -5.64165711e-01 -5.59373260e-01
7.60019183e-01 -3.43219995e-01 4.83236909e-01 4.74399090e-01
1.19398510e+00 -6.12337410e-01 -1.65026158e-01 8.98213863e-01
1.74327099e+00 -1.31435126e-01 -1.32346466e-01 -1.30671427e-01
1.04183602e+00 6.11949384e-01 4.97612923e-01 5.62867820e-01
1.21858343e-01 5.02532482e-01 3.89962405e-01 -4.31844592e-01
3.87946367e-01 3.76420915e-01 1.01351745e-01 8.26643825e-01
-2.62103140e-01 2.54989773e-01 -9.28216279e-01 6.49678469e-01
-2.13518190e+00 -5.23380101e-01 -7.19280183e-01 2.09707236e+00
9.64979887e-01 -4.55554545e-01 7.98618123e-02 3.69165272e-01
5.99085629e-01 -1.64087601e-02 -6.89530611e-01 -7.41210401e-01
1.26575783e-01 1.96180686e-01 5.34049273e-01 9.72134173e-01
-1.11655605e+00 4.24671024e-01 6.17156744e+00 7.67641485e-01
-1.22586763e+00 2.04734668e-01 7.08377004e-01 6.25714734e-02
6.38751984e-02 -1.23673111e-01 -5.45983851e-01 4.23346043e-01
6.84732795e-01 -2.66548693e-01 6.62315845e-01 1.10178971e+00
5.26530862e-01 -1.61728542e-02 -1.18039382e+00 1.02969182e+00
-4.14433271e-01 -1.65542972e+00 -3.96083683e-01 1.54595867e-01
1.28146219e+00 -1.51194200e-01 2.08370928e-02 3.22464824e-01
7.15266690e-02 -7.34457612e-01 4.18265194e-01 2.28094548e-01
6.43105686e-01 -8.11984360e-01 5.76401293e-01 3.79280090e-01
-1.13882208e+00 -2.29831114e-01 -4.65649277e-01 -5.38218617e-01
3.05344731e-01 9.87695634e-01 1.26448825e-01 2.76787162e-01
4.96932536e-01 9.21070695e-01 2.89522767e-01 7.20200002e-01
1.66559637e-01 1.85385332e-01 -4.81860995e-01 1.07065305e-01
5.64396143e-01 -8.95613968e-01 8.15424860e-01 9.48549449e-01
3.90118122e-01 3.99236262e-01 5.64812660e-01 1.17562187e+00
1.57661796e-01 3.23963016e-01 -4.95793521e-01 3.44133615e-01
-2.74913073e-01 1.31195641e+00 -4.42574620e-01 -4.89194483e-01
-4.92480755e-01 9.83492851e-01 3.89975339e-01 5.19065917e-01
-7.80458450e-01 -2.47159943e-01 8.90713990e-01 1.27433062e-01
2.43493974e-01 -3.41631025e-01 -5.80591977e-01 -1.31093550e+00
2.68386126e-01 -4.85554636e-01 4.18909788e-01 -2.74214536e-01
-1.43599832e+00 2.80951738e-01 -2.18498528e-01 -9.91826117e-01
-3.03380210e-02 -1.02883160e+00 -5.76262474e-01 1.12922645e+00
-1.51821041e+00 -7.26506054e-01 -1.28718421e-01 6.61331236e-01
3.77786666e-01 4.31720465e-01 6.97362602e-01 6.45675898e-01
-6.79226577e-01 1.22411318e-01 5.57094991e-01 -5.24946898e-02
1.59569412e-01 -1.60714126e+00 -1.88039437e-01 6.17879748e-01
-8.21518421e-01 8.16532522e-02 8.34602237e-01 -4.05700415e-01
-1.83709109e+00 -1.02489042e+00 7.05612957e-01 2.11258158e-02
1.17809319e+00 -2.49345794e-01 -1.30864799e+00 5.04674494e-01
-4.66213971e-02 7.29679585e-01 1.02073155e-01 -2.30093449e-02
3.93273503e-01 -1.59810677e-01 -1.13114071e+00 3.30004543e-01
5.79827249e-01 -4.04797673e-01 -1.78820536e-01 8.20457518e-01
4.06808645e-01 -6.50080681e-01 -1.14725196e+00 4.59848702e-01
1.49729267e-01 -6.70190811e-01 1.01589525e+00 -6.93482101e-01
6.00085080e-01 -1.71069801e-01 1.03092834e-01 -1.44712567e+00
-3.15500736e-01 -9.50578570e-01 -3.56919825e-01 7.11950898e-01
3.42853636e-01 -9.28256631e-01 4.92884904e-01 8.16405714e-01
-3.36610258e-01 -1.32549596e+00 -9.68746841e-01 -1.03829169e+00
5.21567822e-01 -2.61686444e-01 -1.77541763e-01 1.28721249e+00
-5.60347363e-02 1.95502490e-01 -5.62716365e-01 3.35825354e-01
7.15083778e-01 1.26331270e-01 4.43459630e-01 -1.15596175e+00
-4.99562144e-01 -4.38600570e-01 4.47531231e-02 -1.12127829e+00
2.65744895e-01 -9.65495884e-01 2.24041402e-01 -1.23146605e+00
-4.26801950e-01 -7.37357616e-01 3.47177079e-03 1.78790033e-01
-2.28440076e-01 1.84634939e-01 -1.39302194e-01 6.45430386e-02
-4.34266292e-02 8.22895169e-01 1.63117278e+00 -1.85577601e-01
-4.22144502e-01 9.75299627e-02 -2.11952150e-01 9.62444723e-01
5.24445236e-01 -2.52045810e-01 -3.38614494e-01 -3.67529929e-01
2.77006298e-01 4.91764009e-01 4.89995778e-01 -9.02467370e-01
1.38399109e-01 -1.79705128e-01 -1.04552187e-01 -6.08197525e-02
2.25841627e-01 -7.66477644e-01 -1.86282948e-01 7.18938589e-01
-2.24474892e-01 -1.29766567e-02 4.03459996e-01 3.58944684e-01
-2.93932974e-01 -3.36392403e-01 1.43962240e+00 -8.17182437e-02
-3.34586263e-01 6.52266800e-01 -3.80242109e-01 5.06962061e-01
1.00373828e+00 1.47727415e-01 6.82622612e-01 -8.30830485e-02
-1.16115499e+00 3.66716981e-01 1.73672050e-01 -1.83219790e-01
2.60859609e-01 -1.64999521e+00 -8.24150801e-01 2.06151158e-01
-5.37695348e-01 4.54500854e-01 4.03016329e-01 1.37388742e+00
-8.00125539e-01 1.82579339e-01 -3.67896669e-02 -6.59103811e-01
-6.70854211e-01 8.78862679e-01 5.48231900e-01 -4.79419500e-01
-6.83464825e-01 8.80330682e-01 3.51495773e-01 -3.60984266e-01
2.56667472e-02 -4.44953918e-01 1.82112411e-01 1.14685386e-01
3.24189544e-01 1.03635180e+00 -1.49912164e-01 -4.09035861e-01
-5.02238795e-02 7.80218542e-01 3.88741195e-01 8.91147777e-02
1.47702944e+00 4.07752246e-02 -5.21449685e-01 4.06011403e-01
1.81197727e+00 -1.70254141e-01 -1.49684060e+00 -6.38265833e-02
-4.21763569e-01 -3.15217316e-01 4.29302543e-01 1.10927694e-01
-1.42556477e+00 9.68666792e-01 2.47073114e-01 5.05659103e-01
1.07344210e+00 -5.30681431e-01 1.21688998e+00 3.41860294e-01
9.84029770e-02 -1.23596942e+00 -1.45348668e-01 6.76872134e-01
1.25878823e+00 -1.47869658e+00 8.99566412e-02 -5.91796935e-01
-2.31135860e-01 1.31116450e+00 4.22743261e-01 -6.41161561e-01
8.77292991e-01 1.52601123e-01 -2.08205849e-01 -4.50988002e-02
-3.33031625e-01 1.89025983e-01 3.32628012e-01 7.94488788e-02
2.58399427e-01 -2.06490383e-01 -6.65802896e-01 1.07446343e-01
3.09217632e-01 2.31986754e-02 3.94156277e-01 6.71070457e-01
-3.15452814e-02 -8.66024911e-01 -2.69085467e-01 1.68143868e-01
-3.32732081e-01 8.73409882e-02 1.93424091e-01 8.51708591e-01
-2.48846523e-02 3.79404545e-01 -1.07764848e-01 3.45782578e-01
3.15353483e-01 -1.43436998e-01 3.16583246e-01 -4.05224890e-01
-3.00231665e-01 -7.92875662e-02 -1.09277412e-01 -6.82499468e-01
-2.71866113e-01 -6.87594056e-01 -1.45961285e+00 -5.67105472e-01
-1.74216151e-01 3.47989649e-01 3.79791588e-01 9.39816952e-01
4.34492826e-01 4.05242562e-01 9.88709927e-01 -1.28171206e+00
-8.44342530e-01 -5.53081155e-01 -7.20450640e-01 3.28527391e-01
6.58053160e-01 -7.79903293e-01 -6.60177827e-01 -2.70778686e-01] | [6.577547550201416, 3.4518189430236816] |
6e285c7e-541a-4671-baae-593f6ef7fe8f | temporal-lift-pooling-for-continuous-sign | 2207.08734 | null | https://arxiv.org/abs/2207.08734v1 | https://arxiv.org/pdf/2207.08734v1.pdf | Temporal Lift Pooling for Continuous Sign Language Recognition | Pooling methods are necessities for modern neural networks for increasing receptive fields and lowering down computational costs. However, commonly used hand-crafted pooling approaches, e.g., max pooling and average pooling, may not well preserve discriminative features. While many researchers have elaborately designed various pooling variants in spatial domain to handle these limitations with much progress, the temporal aspect is rarely visited where directly applying hand-crafted methods or these specialized spatial variants may not be optimal. In this paper, we derive temporal lift pooling (TLP) from the Lifting Scheme in signal processing to intelligently downsample features of different temporal hierarchies. The Lifting Scheme factorizes input signals into various sub-bands with different frequency, which can be viewed as different temporal movement patterns. Our TLP is a three-stage procedure, which performs signal decomposition, component weighting and information fusion to generate a refined downsized feature map. We select a typical temporal task with long sequences, i.e. continuous sign language recognition (CSLR), as our testbed to verify the effectiveness of TLP. Experiments on two large-scale datasets show TLP outperforms hand-crafted methods and specialized spatial variants by a large margin (1.5%) with similar computational overhead. As a robust feature extractor, TLP exhibits great generalizability upon multiple backbones on various datasets and achieves new state-of-the-art results on two large-scale CSLR datasets. Visualizations further demonstrate the mechanism of TLP in correcting gloss borders. Code is released. | ['Wei Feng', 'Zekang Liu', 'Liqing Gao', 'Lianyu Hu'] | 2022-07-18 | null | null | null | null | ['sign-language-recognition'] | ['computer-vision'] | [ 3.07214797e-01 -7.85502613e-01 -2.33694777e-01 -3.30006927e-01
-6.96773469e-01 -3.99426669e-01 2.90851295e-01 -6.89169705e-01
-6.97673678e-01 6.63002789e-01 5.57033837e-01 -2.28784252e-02
-1.99817479e-01 -4.98281509e-01 -5.22539973e-01 -9.43065584e-01
-1.39232472e-01 -6.58488393e-01 8.66978705e-01 -4.22234058e-01
4.02497232e-01 6.54856086e-01 -1.77315724e+00 6.33946061e-01
7.34750807e-01 1.01168716e+00 4.43026036e-01 4.13942188e-01
-8.90350863e-02 3.02629977e-01 -5.15797138e-01 -2.01226529e-02
5.01646161e-01 -1.81274727e-01 -4.13549602e-01 -9.67981517e-02
4.68005031e-01 -2.14054316e-01 -4.21673357e-01 9.51628268e-01
7.77117908e-01 3.10373694e-01 3.25798303e-01 -1.04486644e+00
-6.51003063e-01 4.58253473e-01 -8.03790510e-01 3.58454317e-01
9.98904333e-02 1.81096002e-01 9.49000716e-01 -1.02973306e+00
5.78050733e-01 1.19415319e+00 8.84211123e-01 5.49284637e-01
-1.20758367e+00 -7.50263095e-01 3.76645237e-01 1.75923958e-01
-1.41165018e+00 -3.32060844e-01 7.13395655e-01 -1.60136715e-01
9.78415728e-01 4.00863051e-01 5.96251488e-01 8.38840127e-01
2.01672912e-01 1.08666599e+00 1.46519935e+00 -2.10531399e-01
-8.31476226e-02 -2.05009758e-01 4.38597411e-01 5.54543555e-01
7.40133449e-02 -8.09151232e-02 -9.48399007e-01 7.07138032e-02
7.07311153e-01 1.51230380e-01 -6.04834259e-01 1.75511651e-02
-1.44328249e+00 4.02169317e-01 6.04644954e-01 3.52863848e-01
-4.15288001e-01 1.69850998e-02 3.96162808e-01 3.73701036e-01
1.83049500e-01 8.70733112e-02 -5.08440435e-01 2.28603333e-02
-9.47885990e-01 4.93328333e-01 3.06524187e-01 9.11557198e-01
4.96314526e-01 3.03134061e-02 -5.13417304e-01 1.12761390e+00
6.50851130e-02 5.03620803e-01 9.32698905e-01 -6.83879435e-01
4.59023088e-01 5.56102991e-01 -1.30843073e-01 -9.19645548e-01
-5.41062951e-01 -3.41137290e-01 -9.38765645e-01 3.08185488e-01
4.51284438e-01 4.40232316e-03 -1.25038838e+00 1.84958351e+00
-1.25591550e-02 1.98771700e-01 -3.98435108e-02 1.14381468e+00
7.39999771e-01 5.88650525e-01 -4.54803370e-02 -1.34439051e-01
1.59556377e+00 -9.66986656e-01 -6.56731844e-01 -3.66003886e-02
3.04103494e-01 -7.32271612e-01 1.23790348e+00 5.94545364e-01
-9.31752861e-01 -6.45515144e-01 -1.22162080e+00 -1.62933812e-01
-4.64387983e-01 4.22411948e-01 8.34961355e-01 5.90760648e-01
-9.78643179e-01 6.89872444e-01 -8.93380046e-01 -2.62162864e-01
4.16103750e-01 4.68164861e-01 -6.17926478e-01 -1.45766418e-02
-1.06791103e+00 7.00114310e-01 2.61377513e-01 4.16684896e-01
-2.67899662e-01 -6.24903083e-01 -5.97069442e-01 -1.34127140e-01
2.63436675e-01 -3.41851264e-01 9.89193618e-01 -5.13790727e-01
-1.47398984e+00 5.18618345e-01 -4.64077711e-01 -3.79597813e-01
4.29698378e-01 -2.73128778e-01 -4.97900873e-01 -8.40796679e-02
-3.34399901e-02 6.66192770e-01 1.04663169e+00 -8.19166005e-01
-8.01818132e-01 -3.16386789e-01 -1.74084470e-01 1.57293126e-01
-4.40798163e-01 2.85391718e-01 -5.18625021e-01 -1.05404186e+00
6.89397216e-01 -8.60102713e-01 -4.02741507e-02 -5.49776368e-02
-1.88020632e-01 -2.22776696e-01 9.33634162e-01 -7.80404150e-01
1.48214936e+00 -2.27575541e+00 1.65998295e-01 1.67271450e-01
-8.15981627e-02 3.97056907e-01 -1.86797157e-01 1.98668346e-01
6.99061677e-02 2.26562452e-02 -2.56997347e-01 -3.75677139e-01
1.88558530e-02 1.88019484e-01 -6.56029224e-01 4.12678659e-01
2.02765062e-01 6.83185935e-01 -6.87375367e-01 -4.19719100e-01
4.16746400e-02 6.07065380e-01 -6.54758990e-01 -2.47098669e-01
1.98455125e-01 1.99740767e-01 -4.28330898e-01 9.15592313e-01
7.47006714e-01 9.23839584e-02 -1.62212163e-01 -5.94377875e-01
-3.49569410e-01 8.51053074e-02 -1.28784597e+00 1.83191943e+00
-2.35097095e-01 8.12065959e-01 -1.00045823e-01 -6.26181841e-01
8.43998671e-01 1.37479067e-01 2.71651089e-01 -6.77527726e-01
1.13615118e-01 4.39093381e-01 -9.18899253e-02 -6.19697213e-01
6.46540821e-01 -1.31075636e-01 -1.31418213e-01 1.89788654e-01
1.82293296e-01 1.86673656e-01 2.50915885e-01 -2.39425287e-01
1.12214434e+00 2.77506292e-01 2.53650904e-01 -1.51281789e-01
5.36092699e-01 -3.27706695e-01 8.17016006e-01 6.56789720e-01
-5.40444255e-01 8.28890681e-01 2.57835865e-01 -3.14704955e-01
-7.19965518e-01 -1.07211363e+00 -2.86620855e-01 1.39618838e+00
1.29072905e-01 -2.75653362e-01 -4.35855240e-01 -5.40418148e-01
-6.25769198e-02 5.29424585e-02 -4.74323481e-01 5.54868802e-02
-9.22427237e-01 -8.67736936e-01 6.57777429e-01 7.20772028e-01
9.42211509e-01 -1.17147470e+00 -8.81837487e-01 6.96259513e-02
-2.46673092e-01 -1.02132165e+00 -7.55744159e-01 1.01067042e-02
-7.91481316e-01 -6.23637497e-01 -1.02725661e+00 -9.66118634e-01
4.58135873e-01 6.09417975e-01 3.42308939e-01 -1.62632227e-01
-4.19106543e-01 -1.86354458e-01 -4.58451778e-01 -2.66805410e-01
4.27869052e-01 -2.19690558e-02 2.54096359e-01 2.67236769e-01
3.82902265e-01 -8.32370281e-01 -7.54701257e-01 5.24676859e-01
-9.60991263e-01 8.87112226e-03 8.40730846e-01 1.10187352e+00
6.18646681e-01 5.38603626e-02 4.89491343e-01 -1.20660074e-01
8.79760921e-01 3.83380651e-02 -3.53194565e-01 3.32003444e-01
-3.20372693e-02 2.43042231e-01 5.97276330e-01 -8.64456058e-01
-1.17499268e+00 -9.07691717e-02 -4.84655872e-02 -2.46764451e-01
-5.99873662e-02 2.97618270e-01 -4.50798944e-02 -2.69771576e-01
5.96809387e-01 5.51130354e-01 -5.16121015e-02 -6.85082018e-01
1.90194309e-01 7.16326952e-01 6.21324182e-01 -4.82712001e-01
5.93135178e-01 6.28852189e-01 -3.51361069e-03 -8.25021625e-01
-6.10844076e-01 -3.90318453e-01 -5.93048394e-01 -7.23565370e-02
7.46903896e-01 -7.20412254e-01 -8.05951297e-01 7.26956189e-01
-9.23672080e-01 -1.49452582e-01 -4.42088917e-02 6.34628654e-01
-3.67966592e-01 2.23779932e-01 -5.26657999e-01 -6.41376138e-01
-2.09091485e-01 -1.16269851e+00 1.04501712e+00 3.63802254e-01
1.75327044e-02 -2.96032697e-01 -1.59517437e-01 1.14118665e-01
6.26698494e-01 1.59379408e-01 5.94596326e-01 -2.61781931e-01
-5.58396816e-01 -2.10785031e-01 -3.54316950e-01 5.43773532e-01
1.25981301e-01 -1.29899368e-01 -1.24040854e+00 -3.77029300e-01
-1.50268897e-01 -2.45641679e-01 1.37208450e+00 4.63338077e-01
1.14964616e+00 -1.00472279e-01 -2.29309648e-01 7.51176894e-01
1.25269103e+00 2.31860831e-01 6.48357630e-01 2.92110533e-01
3.42144221e-01 5.65849960e-01 5.54806948e-01 3.00655097e-01
2.65376810e-02 8.71569872e-01 -1.48672387e-01 1.01790838e-01
-3.91933858e-01 -2.50850827e-01 5.25536299e-01 8.23415399e-01
-5.76677859e-01 2.42039874e-01 -5.68300784e-01 4.57836568e-01
-1.87414610e+00 -1.07884455e+00 3.54795247e-01 1.96199059e+00
9.74455237e-01 1.57188922e-01 5.90068735e-02 4.28005934e-01
4.92908150e-01 3.63041878e-01 -4.15921926e-01 -2.20427811e-02
-4.51608390e-01 3.20011944e-01 6.15015566e-01 2.58478105e-01
-1.08094335e+00 1.06775594e+00 5.94226837e+00 1.04113281e+00
-1.40788102e+00 3.58259715e-02 5.61507652e-03 -2.46717244e-01
1.33114040e-01 -2.91127890e-01 -1.01881635e+00 3.57376009e-01
2.90636152e-01 7.32249534e-03 4.19847935e-01 6.61367953e-01
2.57808626e-01 8.27806741e-02 -7.23184168e-01 1.05707455e+00
4.76849228e-02 -1.15203035e+00 2.75523383e-02 -1.32871553e-01
5.60096443e-01 9.25710648e-02 8.90450031e-02 4.16578203e-01
-6.49602488e-02 -8.29571486e-01 7.52292216e-01 4.60654199e-01
6.35465443e-01 -4.35298234e-01 6.61136448e-01 5.91529049e-02
-1.56290352e+00 -2.75612980e-01 -3.68322968e-01 1.03202006e-02
2.96182066e-01 2.38118112e-01 -2.25312456e-01 4.36954498e-01
1.02424300e+00 6.26380920e-01 -4.70400363e-01 1.17813456e+00
-1.96649268e-01 4.19741750e-01 -5.26226759e-01 -4.82380614e-02
2.08148673e-01 6.76675141e-02 6.33020818e-01 1.40677273e+00
3.59564066e-01 1.64590701e-01 -1.25282973e-01 4.98808384e-01
1.43814147e-01 6.53913319e-02 -4.17557299e-01 1.91198945e-01
3.48752886e-01 1.08484185e+00 -6.90342128e-01 -1.36453673e-01
-4.53966141e-01 9.52630043e-01 6.45971671e-02 6.40645921e-01
-7.25755870e-01 -9.07032073e-01 8.55055451e-01 -1.41484648e-01
5.21142304e-01 -4.12611991e-01 -2.36659557e-01 -1.24834931e+00
4.01645422e-01 -8.21549296e-01 2.82492548e-01 -6.85614407e-01
-1.16404939e+00 7.29243636e-01 -2.41948459e-02 -1.57610559e+00
1.78503528e-01 -9.09135699e-01 -3.99297982e-01 9.52294052e-01
-1.55964935e+00 -1.17459714e+00 -4.21492457e-01 9.53762949e-01
6.18880808e-01 -1.92278698e-01 6.20509982e-01 3.12204868e-01
-6.07810795e-01 7.00355768e-01 -2.09758744e-01 5.92875294e-02
7.16308475e-01 -8.84464264e-01 1.12610362e-01 1.10643780e+00
2.51337942e-02 9.21112239e-01 5.31572640e-01 -3.67097676e-01
-1.32689905e+00 -8.02943707e-01 7.59190440e-01 1.16218910e-01
6.31799102e-01 -3.61107796e-01 -9.48963821e-01 5.35614431e-01
-1.16036028e-01 2.51284838e-01 4.58956510e-01 -1.01576053e-01
-5.91626406e-01 -4.70069259e-01 -1.04601467e+00 9.15443122e-01
1.39412320e+00 -5.20518243e-01 -8.16962659e-01 -5.17433584e-02
7.50903070e-01 -3.46616536e-01 -7.25293636e-01 5.94410777e-01
1.04260826e+00 -1.13524449e+00 1.07592690e+00 -5.42951167e-01
8.88766721e-02 -6.16649389e-01 -4.90661383e-01 -1.04937327e+00
-2.64556259e-01 -6.69703782e-01 2.30181903e-01 9.68903840e-01
3.08250487e-01 -8.99583459e-01 5.61892509e-01 3.57927471e-01
-2.00402722e-01 -8.60945702e-01 -1.05170143e+00 -9.00617957e-01
-3.78368020e-01 -6.72845602e-01 4.58257020e-01 5.92190683e-01
1.18448436e-01 -1.32265672e-01 -4.69314933e-01 2.05391794e-01
5.57879686e-01 1.69424698e-01 6.19925320e-01 -8.64441991e-01
-2.64636755e-01 -7.42921472e-01 -6.80706918e-01 -1.20152831e+00
-5.20799048e-02 -6.85308754e-01 1.08199559e-01 -1.29908192e+00
-8.32341537e-02 -2.07597792e-01 -6.93666339e-01 8.03029239e-01
-5.24962246e-02 5.05749345e-01 2.63251215e-01 2.60924071e-01
-2.08114207e-01 4.31146055e-01 1.36424994e+00 2.89367605e-02
-5.69464862e-01 5.71325235e-02 -4.83601660e-01 8.36220980e-01
8.22107375e-01 -1.33282468e-01 -3.54959756e-01 -4.36545461e-01
-1.46841183e-01 -2.73751169e-01 3.67575556e-01 -1.20906842e+00
3.37255329e-01 -1.33987620e-01 3.17728519e-01 -6.44862354e-01
4.81877208e-01 -6.03439271e-01 -1.47530884e-01 3.26798707e-01
-2.58821189e-01 -9.88052413e-02 4.24703032e-01 3.22920442e-01
-5.56853712e-01 1.45681202e-01 7.23770499e-01 -3.56566161e-02
-1.14053512e+00 2.48004019e-01 -9.10935998e-02 -1.74175322e-01
7.08419085e-01 -4.92502302e-01 -2.70443380e-01 3.68434042e-02
-7.42068768e-01 3.48745286e-02 -6.37787059e-02 5.35996020e-01
7.93780029e-01 -1.38297260e+00 -5.43242216e-01 6.12481117e-01
-1.44820837e-02 -2.50399858e-01 5.47750235e-01 1.12041688e+00
-2.97287822e-01 4.85196680e-01 -4.68842596e-01 -5.57305753e-01
-1.53538537e+00 2.20063359e-01 2.33461291e-01 -2.11947247e-01
-9.06213284e-01 1.04140711e+00 1.22239828e-01 -6.50046766e-02
4.62658614e-01 -8.59685659e-01 -2.07434669e-01 1.75346524e-01
8.87567163e-01 2.91630924e-01 2.51745861e-02 -5.13421834e-01
-4.27697539e-01 1.04871368e+00 -1.11414663e-01 -2.16303349e-01
1.40971434e+00 -5.28458282e-02 6.76444406e-03 3.30712676e-01
1.21084142e+00 1.12339154e-01 -1.40764034e+00 -5.95719695e-01
-4.50776033e-02 -6.02410138e-01 1.93857756e-02 -6.25383496e-01
-1.06947863e+00 9.18952107e-01 6.22879922e-01 -7.50087854e-03
1.66957819e+00 -2.64589280e-01 8.58499885e-01 4.77908999e-01
6.20435894e-01 -1.05310106e+00 -5.61050698e-02 4.59537506e-01
1.27643251e+00 -9.80364561e-01 -9.61924195e-02 -2.90011883e-01
-7.28993773e-01 1.03744364e+00 4.34209645e-01 -2.61198103e-01
7.22845614e-01 3.59723181e-01 1.78545211e-02 1.65165126e-01
-5.67109346e-01 -3.69643956e-01 5.63890159e-01 3.70106488e-01
2.89620101e-01 1.16174310e-01 -4.55129743e-01 8.16657722e-01
-3.32505882e-01 2.72948682e-01 2.68698689e-02 1.08636844e+00
-5.17078400e-01 -9.80599284e-01 -3.85250807e-01 2.88838536e-01
-4.22279984e-01 -8.51329565e-02 6.60394132e-02 7.86829650e-01
4.08618689e-01 6.71532750e-01 -1.50158510e-01 -6.91456556e-01
6.41212761e-01 1.70716807e-01 5.49836755e-01 -1.12501092e-01
-5.59809327e-01 1.23080619e-01 -1.57828301e-01 -8.57370138e-01
-5.39179742e-01 -5.46132684e-01 -1.25365663e+00 -6.51716292e-02
-8.53685662e-02 -4.79994982e-01 4.54685330e-01 7.76793480e-01
2.70858139e-01 5.51429451e-01 1.95408180e-01 -1.13365245e+00
-6.43027961e-01 -1.16430354e+00 -6.87475443e-01 4.85336512e-01
6.12529933e-01 -8.56667101e-01 -3.55268925e-01 8.55927765e-02] | [9.219744682312012, -6.47070837020874] |
6c07fe6d-609b-46be-b4cd-95f9fef8d910 | language-models-can-see-plugging-visual | 2205.02655 | null | https://arxiv.org/abs/2205.02655v2 | https://arxiv.org/pdf/2205.02655v2.pdf | Language Models Can See: Plugging Visual Controls in Text Generation | Generative language models (LMs) such as GPT-2/3 can be prompted to generate text with remarkable quality. While they are designed for text-prompted generation, it remains an open question how the generation process could be guided by modalities beyond text such as images. In this work, we propose a training-free framework, called MAGIC (iMAge-Guided text generatIon with CLIP), for plugging in visual controls in the generation process and enabling LMs to perform multimodal tasks (e.g., image captioning) in a zero-shot manner. MAGIC is a simple yet efficient plug-and-play framework, which directly combines an off-the-shelf LM (i.e., GPT-2) and an image-text matching model (i.e., CLIP) for image-grounded text generation. During decoding, MAGIC influences the generation of the LM by introducing a CLIP-induced score, called magic score, which regularizes the generated result to be semantically related to a given image while being coherent to the previously generated context. Notably, the proposed decoding scheme does not involve any gradient update operation, therefore being computationally efficient. On the challenging task of zero-shot image captioning, MAGIC outperforms the state-of-the-art method by notable margins with a nearly 27 times decoding speedup. MAGIC is a flexible framework and is theoretically compatible with any text generation tasks that incorporate image grounding. In the experiments, we showcase that it is also capable of performing visually grounded story generation given both an image and a text prompt. | ['Nigel Collier', 'Lingpeng Kong', 'Yan Wang', 'Dani Yogatama', 'Fangyu Liu', 'Yahui Liu', 'Tian Lan', 'Yixuan Su'] | 2022-05-05 | null | null | null | null | ['story-generation'] | ['natural-language-processing'] | [ 7.88834751e-01 6.14759982e-01 1.22305751e-01 -7.66228884e-02
-9.81939375e-01 -6.16855264e-01 1.20903826e+00 -5.38243689e-02
-2.14720234e-01 5.07696033e-01 1.79507583e-01 -4.20064270e-01
4.25589085e-01 -8.94122481e-01 -1.11035252e+00 -6.73133850e-01
4.87015694e-01 6.38164818e-01 8.00034925e-02 -2.73649096e-01
1.93209887e-01 1.24189444e-01 -1.70864105e+00 4.65404332e-01
7.77619004e-01 8.29858422e-01 7.85886824e-01 9.29842949e-01
-3.92515540e-01 7.19636500e-01 -5.32805741e-01 -5.61188340e-01
-9.36059654e-02 -8.39803696e-01 -5.42698443e-01 2.83925354e-01
5.46395540e-01 -2.94102699e-01 -6.24747127e-02 5.29035985e-01
7.67516136e-01 1.60079300e-01 6.55802250e-01 -1.26924181e+00
-7.21883297e-01 6.64872587e-01 -3.10331464e-01 -4.68297333e-01
8.03541064e-01 4.07509595e-01 8.02626252e-01 -1.14748812e+00
1.10825038e+00 1.21791995e+00 1.36558682e-01 8.90921533e-01
-1.24033391e+00 -2.27289051e-01 6.55194074e-02 -1.54947087e-01
-1.15162206e+00 -6.94158554e-01 6.69459522e-01 -4.48376387e-01
7.99162447e-01 4.93114978e-01 5.82785070e-01 1.38886547e+00
-1.73956454e-01 9.52697158e-01 8.91019523e-01 -8.36058319e-01
3.00303131e-01 3.27655554e-01 -7.51818597e-01 6.52721047e-01
-8.90731066e-02 4.51971963e-02 -8.38213921e-01 1.15064591e-01
7.34216034e-01 -4.96766001e-01 -4.01047140e-01 -4.50422555e-01
-1.54266512e+00 8.25654924e-01 2.94394404e-01 2.33225822e-01
-3.75567019e-01 4.67027158e-01 2.22464919e-01 1.16351654e-03
4.09201026e-01 3.64521891e-01 1.08906440e-01 -1.48483053e-01
-1.38192976e+00 3.78525883e-01 7.44743764e-01 1.14773643e+00
7.77934492e-01 1.73256233e-01 -1.02222919e+00 6.09285295e-01
3.54421943e-01 6.53549612e-01 4.64542091e-01 -6.60006046e-01
5.91937304e-01 3.58460635e-01 1.82813779e-01 -7.32135832e-01
1.08741196e-02 -1.24888085e-01 -5.89292526e-01 2.45367825e-01
3.62387240e-01 -2.73450643e-01 -1.09059942e+00 1.89144635e+00
3.34132284e-01 1.86360329e-01 2.38124758e-01 9.37371969e-01
1.22617126e+00 8.80547643e-01 1.55662507e-01 -9.51035842e-02
1.52049434e+00 -1.10855079e+00 -5.85772038e-01 -3.63848478e-01
5.06730080e-01 -9.44730937e-01 1.32697344e+00 -6.15898799e-03
-1.40814531e+00 -5.59960425e-01 -8.86617303e-01 -1.64036140e-01
-2.97886848e-01 1.90892980e-01 4.16798204e-01 5.03770888e-01
-1.39042091e+00 2.86468118e-01 -4.04027373e-01 -6.74306750e-01
7.76521116e-02 -7.34479493e-03 -3.02509725e-01 3.49266059e-03
-1.16321433e+00 6.75399423e-01 5.55061817e-01 -1.88269287e-01
-1.11879206e+00 -4.76855069e-01 -1.07866681e+00 7.44548142e-02
4.52306956e-01 -1.24851167e+00 1.38904440e+00 -1.14862788e+00
-1.85145235e+00 1.00478745e+00 -2.93236464e-01 -4.36954051e-01
7.98604846e-01 -1.04699098e-02 -2.11786143e-02 4.84614313e-01
7.62840435e-02 1.50715184e+00 1.22681081e+00 -1.63703263e+00
-2.09752440e-01 1.71416521e-01 9.29905400e-02 4.49360758e-01
-2.23850071e-01 -4.63306122e-02 -8.97314489e-01 -7.38517344e-01
-2.65132785e-01 -9.10897553e-01 -2.15198889e-01 9.98451188e-02
-7.54516721e-01 1.68927774e-01 6.76373065e-01 -4.95342374e-01
1.02147090e+00 -1.84005594e+00 1.82876989e-01 2.90664472e-02
-2.27958992e-01 2.12290376e-01 -4.37761128e-01 8.61274719e-01
6.30179048e-02 1.48277685e-01 -2.76342779e-01 -8.38653743e-01
2.56497949e-01 3.63660790e-03 -4.25183326e-01 -3.17373537e-02
3.43830734e-01 1.40074742e+00 -9.06108558e-01 -6.52673423e-01
4.82603729e-01 6.54437125e-01 -3.88660848e-01 4.30633247e-01
-8.23735833e-01 5.35287440e-01 -3.04792732e-01 3.03538471e-01
4.36065674e-01 -3.39932442e-01 8.02027956e-02 -7.43546486e-02
-1.98276028e-01 2.41673496e-02 -8.40234399e-01 2.13676834e+00
-6.83501422e-01 7.42013633e-01 -1.21886894e-01 -5.09849548e-01
9.64518249e-01 6.00137889e-01 3.63151580e-02 -7.71951795e-01
6.30766824e-02 4.49657664e-02 -5.01400650e-01 -5.31973362e-01
9.34215665e-01 -3.10893320e-02 -7.98283815e-02 6.50981188e-01
1.10943392e-01 -4.10636932e-01 3.79173964e-01 5.24999797e-01
8.30587685e-01 6.21012509e-01 7.64900260e-03 1.87822595e-01
4.08243507e-01 1.09262295e-01 -2.89816141e-01 9.58643019e-01
5.41204095e-01 1.13432062e+00 3.61977071e-01 1.74341887e-01
-1.24317777e+00 -9.12695765e-01 2.87673205e-01 1.06916511e+00
9.88579541e-02 -5.43560386e-01 -1.18007183e+00 -2.93225735e-01
-4.18626040e-01 1.03389859e+00 -5.07503569e-01 2.00100127e-03
-3.97884578e-01 -3.76411319e-01 5.14590085e-01 2.55479723e-01
4.00001705e-01 -1.41308367e+00 -8.60458493e-01 3.58124554e-01
-4.69549149e-01 -1.34136713e+00 -6.02878869e-01 -1.05580360e-01
-6.75047755e-01 -6.28999054e-01 -1.29316032e+00 -8.55414152e-01
9.14673924e-01 1.82153940e-01 1.28413630e+00 5.55270389e-02
-1.25206590e-01 8.04777026e-01 -5.61329067e-01 -1.16315559e-01
-8.49189997e-01 5.78293055e-02 -6.49365544e-01 1.93567976e-01
-4.15853232e-01 -3.96011502e-01 -7.37603426e-01 7.71368593e-02
-1.31978440e+00 9.82094288e-01 7.45029092e-01 7.48007357e-01
5.90176940e-01 -5.74509382e-01 3.00382018e-01 -7.52724051e-01
7.70227790e-01 -2.74915189e-01 -3.76782686e-01 4.57982928e-01
-2.70836622e-01 1.61521256e-01 5.48159599e-01 -5.51502585e-01
-1.20840228e+00 2.97237962e-01 -1.62133291e-01 -3.21011722e-01
-2.22839504e-01 4.68809605e-01 -1.34591863e-01 9.91991442e-03
5.94850779e-01 6.29625440e-01 -1.96042776e-01 -2.48533234e-01
9.41052377e-01 4.62287009e-01 6.94670081e-01 -5.00539422e-01
9.48778033e-01 3.56712908e-01 -1.15742810e-01 -5.86379468e-01
-4.68509316e-01 -1.83285788e-01 -3.00999939e-01 -4.36288536e-01
1.02908802e+00 -9.25250590e-01 -4.04331058e-01 3.75731617e-01
-1.42272198e+00 -6.01042092e-01 -3.22385252e-01 5.26106590e-03
-1.03037179e+00 2.06635237e-01 -3.89428407e-01 -8.47640455e-01
-5.81607521e-01 -1.01379550e+00 1.65664744e+00 2.66488045e-01
-1.86380416e-01 -9.45162892e-01 -3.68887186e-02 2.86597311e-01
4.27306652e-01 4.95635033e-01 7.30472624e-01 -3.28357637e-01
-8.14181089e-01 3.02028619e-02 -3.04536641e-01 -8.21035169e-03
-3.47986609e-01 -3.05754300e-02 -1.19411588e+00 -1.71889037e-01
-4.05211538e-01 -3.37251425e-01 7.83611774e-01 2.55334824e-01
7.39719629e-01 -4.01701331e-01 -2.95420736e-01 4.60540831e-01
1.42437100e+00 8.07478130e-02 1.06277096e+00 2.90954262e-01
7.17606127e-01 5.97802460e-01 5.67914009e-01 4.53778088e-01
3.62187803e-01 1.09018517e+00 5.44304669e-01 -4.56753612e-01
-5.19278646e-01 -8.09586704e-01 5.48761725e-01 3.99810612e-01
1.08473673e-01 -8.84477019e-01 -6.94817483e-01 4.65497315e-01
-2.06751752e+00 -1.03353047e+00 -3.04978728e-01 2.11210847e+00
7.77194023e-01 -1.43346742e-01 -1.67317674e-01 -1.18674628e-01
8.40650618e-01 2.10778415e-01 -2.77881235e-01 -5.18929839e-01
-3.35829854e-01 1.68716282e-01 9.91291106e-02 4.81538892e-01
-8.22220445e-01 1.22020781e+00 5.60018682e+00 1.05502665e+00
-1.09935296e+00 1.58571556e-01 7.09524274e-01 -1.93912704e-02
-6.25850439e-01 1.04153916e-01 -6.28878891e-01 4.68111545e-01
8.55968297e-01 -1.51660770e-01 3.82674307e-01 6.47372842e-01
5.82224846e-01 -3.01225454e-01 -1.04133809e+00 9.30831790e-01
4.35855329e-01 -1.55526257e+00 4.14994210e-01 -7.46357888e-02
7.55942762e-01 -4.00226593e-01 1.54822096e-01 1.07556127e-01
-4.07078192e-02 -9.44058120e-01 1.24950695e+00 5.86823583e-01
1.22655940e+00 -4.56977904e-01 3.64781618e-01 3.88896942e-01
-1.08270276e+00 3.86509836e-01 -9.96803120e-03 2.62343049e-01
6.45092905e-01 5.16426861e-01 -1.07025766e+00 5.12449026e-01
1.54127955e-01 2.14622900e-01 -5.95519662e-01 9.28944468e-01
-5.53934693e-01 4.49448317e-01 -3.06981802e-02 2.44385544e-02
4.89869088e-01 5.07447049e-02 6.78569734e-01 1.35180199e+00
8.07824850e-01 -2.13070065e-01 4.75482717e-02 1.25669467e+00
-2.86003910e-02 3.10000300e-01 -6.46029234e-01 -2.31080711e-01
2.47422755e-01 1.40988147e+00 -8.59899580e-01 -5.49702764e-01
-4.99093160e-02 1.60384464e+00 -3.54992002e-02 5.06046057e-01
-9.71275270e-01 -2.67357051e-01 3.19968686e-02 2.85680026e-01
5.25600493e-01 -5.63157722e-02 3.03284861e-02 -9.85309601e-01
6.34748116e-02 -7.37616539e-01 -1.13588773e-01 -1.58730602e+00
-8.77384424e-01 8.62635374e-01 -5.39156422e-02 -1.13805437e+00
-7.84187317e-01 -3.51520121e-01 -8.63895833e-01 9.04123604e-01
-1.47482324e+00 -1.49532497e+00 -5.28401554e-01 7.32718289e-01
6.31148934e-01 1.48140341e-01 8.64538729e-01 -7.37146810e-02
-1.81495890e-01 5.23557067e-01 -1.73139691e-01 -2.12773010e-01
7.44936943e-01 -1.15197480e+00 6.12958431e-01 8.48146915e-01
3.77217710e-01 2.58352816e-01 8.79609227e-01 -6.94684744e-01
-1.43405426e+00 -1.14595747e+00 9.90927696e-01 -1.89543098e-01
3.28701973e-01 -7.45588243e-01 -5.28197408e-01 2.47285813e-01
5.98767877e-01 -4.64443713e-01 3.79922390e-01 -5.83117187e-01
-1.69789139e-02 3.81076217e-01 -9.76979852e-01 9.16541755e-01
1.13270938e+00 -4.57246125e-01 -2.00756341e-01 4.71072614e-01
9.78412032e-01 -6.40822172e-01 -3.55440795e-01 -1.09853048e-03
3.18845451e-01 -9.55005527e-01 9.20889020e-01 -1.10441625e-01
7.03633726e-01 -3.73495370e-01 6.20882288e-02 -1.11562240e+00
5.86038344e-02 -1.19959378e+00 -1.23680659e-01 1.52290940e+00
5.62479138e-01 -2.30267495e-01 5.96407056e-01 4.45800871e-01
-2.87897378e-01 -5.17548680e-01 -7.45473266e-01 -6.57969952e-01
-3.21516752e-01 -5.88857710e-01 5.14495373e-01 4.35214847e-01
-3.51465382e-02 4.13669437e-01 -5.79463661e-01 -1.95971951e-01
2.73614317e-01 1.86639160e-01 9.64647770e-01 -6.93684697e-01
-4.19206768e-01 -3.32129598e-01 -8.46083090e-02 -1.33755732e+00
-5.98803721e-02 -9.97540295e-01 3.34951699e-01 -1.90975475e+00
1.92221612e-01 -2.14280859e-01 2.99615175e-01 4.20034319e-01
-1.91885501e-01 4.29917783e-01 6.00103796e-01 1.57523826e-01
-6.70184672e-01 6.59107685e-01 1.51274824e+00 -3.54250930e-02
-1.74298733e-01 -3.14957291e-01 -7.05851972e-01 3.71150941e-01
6.17750704e-01 -3.68566334e-01 -5.01481295e-01 -3.84331793e-01
3.13088506e-01 3.35872680e-01 6.40966833e-01 -8.91582191e-01
3.07889253e-01 1.28973663e-01 2.60649562e-01 -4.82997864e-01
5.42885423e-01 -4.62232471e-01 4.16713476e-01 1.23696648e-01
-4.79186147e-01 -7.76958615e-02 2.06865355e-01 3.80911142e-01
-1.40266702e-01 -4.75079268e-01 4.09121335e-01 -2.06941426e-01
-6.69054389e-01 1.25848398e-01 -4.50252503e-01 -3.16513479e-02
9.46169853e-01 -4.04571921e-01 -4.23210740e-01 -7.93429792e-01
-6.42754495e-01 -8.00286513e-03 6.86983526e-01 4.92212713e-01
7.15291739e-01 -1.50112545e+00 -7.94860721e-01 8.77571329e-02
3.16155761e-01 -9.22394544e-02 2.96173096e-01 6.87031746e-01
-3.42471451e-01 5.17487586e-01 1.25738397e-01 -7.09523439e-01
-1.19215739e+00 5.10787189e-01 2.98397150e-02 -4.36833292e-01
-6.30229771e-01 7.36543000e-01 4.56346035e-01 -1.79945841e-01
9.25970003e-02 -8.05180743e-02 9.02111754e-02 3.77161498e-03
6.63570523e-01 -1.50553077e-01 -1.44432038e-02 -6.97521627e-01
-1.78569537e-02 5.31882763e-01 3.54664594e-01 -7.89058745e-01
1.04727232e+00 -2.30289310e-01 1.70165107e-01 2.46828198e-01
8.17028940e-01 -1.88497961e-01 -1.39763367e+00 7.55540133e-02
-2.81142414e-01 -2.54819870e-01 -6.25835508e-02 -1.16393948e+00
-9.16345417e-01 8.52519631e-01 4.11618203e-01 1.30300775e-01
1.14971423e+00 1.49466693e-01 8.15066576e-01 1.49321109e-01
3.51347953e-01 -8.55681241e-01 4.52492833e-01 2.79536635e-01
1.18389928e+00 -1.06506741e+00 -4.72575009e-01 -3.41234118e-01
-9.24154103e-01 1.00882971e+00 3.92311513e-01 2.93914020e-01
-1.88178539e-01 1.99039742e-01 -1.14639727e-02 -1.13431454e-01
-8.93867731e-01 -4.73504156e-01 5.73516071e-01 7.65859127e-01
5.01387179e-01 -8.54093730e-02 -2.13029236e-01 3.37959319e-01
-2.04924881e-01 1.20116293e-01 4.34191108e-01 6.76120818e-01
-3.98734748e-01 -1.24023461e+00 -5.51205337e-01 9.66340769e-03
-7.77643248e-02 -5.29442906e-01 -4.96478826e-01 6.55261993e-01
2.24506840e-01 1.16347206e+00 -2.96441205e-02 -1.61016151e-01
2.08925128e-01 1.64377660e-01 5.61959028e-01 -6.79152131e-01
-6.56284392e-01 2.15582594e-01 1.72991246e-01 -5.83486974e-01
-4.14357364e-01 -4.68828261e-01 -1.17375720e+00 -1.05205774e-01
-2.68862933e-01 -7.74969012e-02 9.54702854e-01 7.64431238e-01
5.04540086e-01 5.08138955e-01 2.77168214e-01 -1.24282515e+00
1.30027100e-01 -7.94354439e-01 -1.32308841e-01 4.60617781e-01
-7.00427666e-02 -2.86442518e-01 -1.11585647e-01 4.29577738e-01] | [11.218948364257812, 0.321459025144577] |
2013239e-bfcf-4336-a9c0-a417d2627ed3 | coupling-knowledge-based-and-data-driven | null | null | https://aclanthology.org/W12-0510 | https://aclanthology.org/W12-0510.pdf | Coupling Knowledge-Based and Data-Driven Systems for Named Entity Recognition | null | ['Jean-Yves Antoine', 'Nathalie Friburger', 'Damien Nouvel', 'Arnaud Soulet'] | 2012-04-01 | null | null | null | ws-2012-4 | ['sequential-pattern-mining'] | ['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] | [-7.257062911987305, 3.695284128189087] |
4be4d0b9-c988-413e-ad38-0ccd33c8eeac | 4d-x-ray-ct-reconstruction-using-multi-slice | 1906.06601 | null | https://arxiv.org/abs/1906.06601v1 | https://arxiv.org/pdf/1906.06601v1.pdf | 4D X-Ray CT Reconstruction using Multi-Slice Fusion | There is an increasing need to reconstruct objects in four or more dimensions corresponding to space, time and other independent parameters. The best 4D reconstruction algorithms use regularized iterative reconstruction approaches such as model based iterative reconstruction (MBIR), which depends critically on the quality of the prior modeling. Recently, Plug-and-Play methods have been shown to be an effective way to incorporate advanced prior models using state-of-the-art denoising algorithms designed to remove additive white Gaussian noise (AWGN). However, state-of-the-art denoising algorithms such as BM4D and deep convolutional neural networks (CNNs) are primarily available for 2D and sometimes 3D images. In particular, CNNs are difficult and computationally expensive to implement in four or more dimensions, and training may be impossible if there is no associated high-dimensional training data. In this paper, we present Multi-Slice Fusion, a novel algorithm for 4D and higher-dimensional reconstruction, based on the fusion of multiple low-dimensional denoisers. Our approach uses multi-agent consensus equilibrium (MACE), an extension of Plug-and-Play, as a framework for integrating the multiple lower-dimensional prior models. We apply our method to the problem of 4D cone-beam X-ray CT reconstruction for Non Destructive Evaluation (NDE) of moving parts. This is done by solving the MACE equations using lower-dimensional CNN denoisers implemented in parallel on a heterogeneous cluster. Results on experimental CT data demonstrate that Multi-Slice Fusion can substantially improve the quality of reconstructions relative to traditional 4D priors, while also being practical to implement and train. | ['Craig A. J. Kemp', 'Thilo Balke', 'Gregery T. Buzzard', 'Charles A. Bouman', 'Soumendu Majee'] | 2019-06-15 | null | null | null | null | ['low-dose-x-ray-ct-reconstruction'] | ['medical'] | [-8.02307501e-02 -1.83760524e-01 4.17543977e-01 -1.20928630e-01
-1.20277131e+00 -1.79270193e-01 3.76210153e-01 -2.32665129e-02
-5.79262555e-01 6.55213654e-01 2.46048748e-01 -1.50082469e-01
-5.08135080e-01 -7.28853047e-01 -5.29491425e-01 -1.10732937e+00
-2.32196093e-01 8.83063257e-01 2.63204992e-01 -1.67749494e-01
-1.98818609e-01 9.13993061e-01 -1.17613220e+00 3.05554956e-01
3.85209382e-01 8.58588159e-01 1.95905849e-01 1.02427518e+00
1.55916125e-01 8.01574230e-01 -2.75968581e-01 3.62297371e-02
2.40311682e-01 -4.74187523e-01 -5.84104776e-01 -3.81049141e-02
-1.72566727e-01 -5.76299131e-01 -3.04251552e-01 8.12111974e-01
1.04294086e+00 1.00203745e-01 6.31254435e-01 -6.53280318e-01
-1.64043218e-01 1.80398792e-01 -5.43490767e-01 1.98944002e-01
1.04753479e-01 1.19899854e-01 1.52346268e-01 -8.00461531e-01
5.85489452e-01 1.18579948e+00 1.13909459e+00 6.78850949e-01
-1.42084587e+00 -3.20756108e-01 -4.94823664e-01 1.41808391e-01
-1.07760167e+00 -2.95161486e-01 8.05217505e-01 -3.24181885e-01
9.29540336e-01 4.20738846e-01 5.95683038e-01 1.08163118e+00
5.21082580e-01 4.47082520e-01 1.41630995e+00 -4.87542152e-01
5.57491064e-01 -5.55763662e-01 -2.26280466e-01 4.00461555e-01
1.54477105e-01 2.70211071e-01 -4.36066955e-01 -4.47577000e-01
1.17796969e+00 -1.00542553e-01 -3.85358751e-01 -2.69226342e-01
-1.02137351e+00 8.56228650e-01 2.70804524e-01 3.94487172e-01
-1.03186512e+00 4.38989908e-01 4.56613541e-01 1.74131095e-01
9.29644525e-01 7.97932521e-02 -9.34969932e-02 -6.75330758e-02
-1.17693245e+00 3.78466278e-01 7.93131530e-01 1.77413896e-01
5.95069110e-01 -2.21241936e-02 5.35577200e-02 7.49930978e-01
5.81403017e-01 3.04505140e-01 4.03839558e-01 -1.46688557e+00
-1.35381356e-01 -1.41689867e-01 -5.73938861e-02 -5.82845390e-01
-5.84694266e-01 -5.66368699e-01 -1.56570756e+00 8.20107222e-01
1.36925295e-01 -5.58835752e-02 -1.17824268e+00 1.55629170e+00
7.78556287e-01 5.10758579e-01 1.54269084e-01 1.20896554e+00
5.96887529e-01 4.35262412e-01 -2.20182836e-01 -5.02480805e-01
1.02244663e+00 -4.31070417e-01 -7.97818005e-01 1.53941233e-02
3.31005096e-01 -9.85024095e-01 2.11925521e-01 7.43477464e-01
-1.69064522e+00 -4.16502714e-01 -9.87043023e-01 1.51511386e-01
2.82562971e-01 -5.08904815e-01 3.01592022e-01 5.38348556e-01
-1.34070015e+00 1.10094559e+00 -1.50380611e+00 2.84271128e-02
3.48478854e-01 5.26162624e-01 -3.63982916e-01 -4.53851104e-01
-9.87289131e-01 1.11599433e+00 4.68802340e-02 4.45519328e-01
-1.20069146e+00 -8.55734348e-01 -5.88008761e-01 -2.51394242e-01
1.53868021e-02 -1.07117891e+00 1.36538315e+00 -6.17568254e-01
-1.72838402e+00 5.55290759e-01 2.00892270e-01 -3.49177331e-01
8.49567592e-01 1.27057925e-01 -1.72544122e-01 4.57506418e-01
6.98106065e-02 1.98509514e-01 8.33562613e-01 -1.56547737e+00
-2.65127160e-02 -3.83415580e-01 -2.48076335e-01 2.25334927e-01
3.22513372e-01 9.84349847e-02 -3.24708581e-01 -5.14762938e-01
6.61548615e-01 -7.28803277e-01 -7.77527213e-01 1.34025365e-01
-2.11460561e-01 2.79879034e-01 6.51255906e-01 -1.01752543e+00
6.42447054e-01 -1.87682414e+00 5.30577242e-01 4.64115649e-01
4.15007502e-01 2.18983088e-03 8.52742270e-02 3.03158969e-01
-2.97463208e-01 -1.53531551e-01 -5.90670288e-01 -7.62714446e-01
-1.70467660e-01 7.08760738e-01 4.32010591e-01 8.23452055e-01
-2.76908606e-01 4.85238671e-01 -7.30040669e-01 -4.34034556e-01
3.52051705e-01 9.34443414e-01 -3.27906370e-01 1.15682974e-01
1.08247273e-01 8.60442102e-01 -2.83043802e-01 3.39485526e-01
9.59415019e-01 -2.72435755e-01 2.11610764e-01 -4.46336478e-01
-9.47711915e-02 -1.24211803e-01 -1.49093175e+00 1.90093815e+00
-6.05121493e-01 1.23500191e-01 9.56446767e-01 -1.10875952e+00
3.91299456e-01 8.89891505e-01 9.58317935e-01 -7.28023887e-01
1.16390169e-01 6.60617113e-01 5.04407380e-03 -4.40816283e-01
2.15821683e-01 -9.71467793e-01 2.82606840e-01 6.12222135e-01
-6.84750825e-02 -5.08492291e-01 -9.34668779e-02 2.78635830e-01
1.60689473e+00 -3.11000235e-02 -2.23265700e-02 -2.29162589e-01
2.02670172e-01 -6.56606853e-02 4.29316521e-01 8.36756229e-01
2.12225672e-02 1.19392228e+00 1.12470023e-01 -3.39168936e-01
-1.27005255e+00 -9.51731563e-01 -1.69753656e-01 2.32588857e-01
-1.59206137e-01 1.03534507e-02 -8.04693162e-01 -9.59608853e-02
-3.04473341e-01 6.36952043e-01 -5.94900906e-01 2.58850187e-01
-9.19785142e-01 -1.18904197e+00 3.57587844e-01 4.23557311e-01
3.38023543e-01 -8.31890643e-01 -5.69086134e-01 7.85748243e-01
-2.06524953e-01 -8.50946009e-01 1.97288811e-01 6.73236489e-01
-1.19003153e+00 -9.44636881e-01 -9.79559660e-01 -1.71547651e-01
4.87225860e-01 9.59378257e-02 1.27928531e+00 1.82712674e-01
-2.31825083e-01 8.29426050e-01 -2.03246251e-01 -1.26542434e-01
-1.03810191e+00 -3.54639471e-01 1.90259293e-01 -1.78416833e-01
-3.92950565e-01 -1.15682411e+00 -8.39218020e-01 2.60140985e-01
-1.35171294e+00 1.35093287e-01 5.14535308e-01 9.42167282e-01
9.09146130e-01 2.94025213e-01 3.02597106e-01 -5.15082896e-01
7.21078694e-01 -4.26063597e-01 -3.88135284e-01 -1.78685668e-03
-3.04719865e-01 4.60931659e-02 4.04328585e-01 -2.44447768e-01
-1.16433656e+00 4.60230932e-03 -7.25343645e-01 -7.46754527e-01
-2.66372263e-01 7.37419426e-01 2.37212375e-01 -4.19339925e-01
8.60618412e-01 -9.92150232e-03 4.17325467e-01 -7.22646058e-01
2.18532830e-01 4.41275746e-01 6.26157403e-01 -6.83274388e-01
5.69101334e-01 7.99672902e-01 5.13069570e-01 -6.86241090e-01
-5.86977839e-01 -4.69901145e-01 -6.39348924e-01 -5.41740596e-01
1.09289169e+00 -8.34953308e-01 -6.65863752e-01 7.07217395e-01
-1.35186338e+00 -4.83863294e-01 -5.18889964e-01 8.09436321e-01
-7.53042400e-01 6.33928418e-01 -9.26012576e-01 -6.19118154e-01
-4.59437639e-01 -1.47000825e+00 1.03507817e+00 -2.36457750e-01
1.48661388e-02 -1.06305993e+00 4.50543523e-01 2.22137704e-01
7.65637219e-01 3.65083992e-01 8.04072917e-01 -2.35180959e-01
-5.76088488e-01 -1.64155409e-01 9.77192521e-02 6.70085609e-01
-2.71405756e-01 -4.73185390e-01 -8.00175846e-01 -3.23317111e-01
8.71503592e-01 -2.33446121e-01 5.93615592e-01 8.80707681e-01
9.82318521e-01 1.50722548e-01 -1.27308488e-01 5.56745946e-01
1.70251679e+00 -1.00543335e-01 9.03333306e-01 3.45697373e-01
4.73905623e-01 2.04635918e-01 5.85019216e-02 3.52611482e-01
-5.29801613e-03 5.96717834e-01 6.60545528e-01 -2.98568070e-01
-4.28667724e-01 6.74009919e-01 -9.58352685e-02 1.25772858e+00
-5.39035618e-01 -1.24461658e-01 -1.00503850e+00 3.65204096e-01
-1.76296735e+00 -9.54577923e-01 -4.75276649e-01 1.96767294e+00
8.48663628e-01 -6.66281879e-02 -3.21867794e-01 2.61823803e-01
5.24234414e-01 -3.02970439e-01 -3.97154391e-01 -4.80947308e-02
-1.72482673e-02 7.53416598e-01 5.71147323e-01 5.99000573e-01
-7.84243822e-01 -8.04759115e-02 6.02499247e+00 9.62787986e-01
-9.34939981e-01 7.95409799e-01 4.93168503e-01 -1.12393081e-01
-1.47413850e-01 -1.94951922e-01 -1.90201864e-01 2.35096648e-01
1.01639795e+00 3.78812641e-01 5.37378371e-01 4.37192351e-01
4.71316308e-01 -3.98234129e-01 -9.10466671e-01 9.49261725e-01
-8.97435769e-02 -1.49835384e+00 -6.06502056e-01 1.67285249e-01
7.82848358e-01 4.68977928e-01 -4.20208454e-01 -9.94112417e-02
4.97596502e-01 -8.33711267e-01 6.27338111e-01 7.40490615e-01
4.81876582e-01 -5.86450815e-01 8.25212538e-01 5.39108634e-01
-5.92607260e-01 3.70815367e-01 -2.08485514e-01 2.84510016e-01
7.36979425e-01 9.00704324e-01 -2.99821287e-01 1.04904604e+00
9.21781957e-01 2.43351340e-01 1.27908677e-01 1.18344879e+00
7.08649904e-02 4.48886573e-01 -8.14423800e-01 5.11790991e-01
2.71804690e-01 -1.07993118e-01 6.81400359e-01 9.26873088e-01
6.16461754e-01 4.70355660e-01 -3.69245745e-02 4.50269043e-01
1.37855455e-01 -4.72369492e-01 -1.19447149e-01 8.44327927e-01
-2.99987853e-01 1.19340038e+00 -8.60516131e-01 -3.08810055e-01
-2.86788583e-01 9.21722233e-01 -1.05711661e-01 3.59678566e-01
-6.97374046e-01 3.63171905e-01 4.47812110e-01 3.25717241e-01
3.48200560e-01 -4.26711529e-01 -2.95593351e-01 -9.53344107e-01
-1.07866511e-01 -9.19620395e-01 3.17766070e-01 -9.84458327e-01
-1.83014333e+00 5.79098463e-01 1.75492808e-01 -1.12104452e+00
-4.09796536e-01 -4.77666914e-01 -4.96877491e-01 9.96082008e-01
-1.40927625e+00 -1.05274832e+00 -3.08545679e-01 7.15199053e-01
2.15597108e-01 4.06690449e-01 9.82293546e-01 6.94077611e-01
-3.49488929e-02 -2.24309012e-01 6.16391540e-01 -2.24437386e-01
5.07635057e-01 -1.17274034e+00 -8.79081935e-02 7.25483298e-01
-3.68624926e-01 1.08629957e-01 9.23956215e-01 -8.09440374e-01
-1.55096602e+00 -7.23534703e-01 1.91784427e-01 -5.00475913e-02
5.23661137e-01 -7.87652880e-02 -1.03244209e+00 5.96485019e-01
4.26722974e-01 2.15322688e-01 5.10896981e-01 -2.38628671e-01
3.06209147e-01 4.25092950e-02 -1.40185845e+00 2.05102876e-01
6.26537681e-01 -3.55703205e-01 -2.99068779e-01 6.11995637e-01
1.40509740e-01 -8.23988557e-01 -1.11896563e+00 5.00355482e-01
2.04049081e-01 -1.19145274e+00 1.40775895e+00 -1.58886071e-02
3.76561642e-01 -2.92717636e-01 -3.43015730e-01 -1.50626743e+00
-3.91292572e-01 -5.31773031e-01 -9.82826501e-02 5.95261097e-01
-1.47182336e-02 -4.42940652e-01 7.77482748e-01 6.49941802e-01
-6.26569152e-01 -6.10757291e-01 -1.66844499e+00 -6.19963288e-01
2.34309375e-01 -9.48698759e-01 1.52147681e-01 8.22103977e-01
-7.59558022e-01 -2.59596445e-02 -3.77251327e-01 4.52292413e-01
1.11838257e+00 -2.88676739e-01 4.02370661e-01 -1.18870020e+00
-6.89248562e-01 -1.19684257e-01 -2.57602245e-01 -7.36006916e-01
-1.16405226e-01 -7.97591686e-01 2.01153681e-01 -1.95153844e+00
1.24928698e-01 -5.27384281e-01 -1.08481847e-01 2.15828568e-01
4.09557909e-01 4.84047741e-01 -8.32202435e-02 4.66142356e-01
-2.69419730e-01 4.79074895e-01 1.40144348e+00 -4.01014946e-02
7.92903155e-02 -1.35368735e-01 7.86512345e-03 7.86644936e-01
4.79687661e-01 -8.71044278e-01 -4.66855653e-02 -5.94969153e-01
2.71388948e-01 6.64308548e-01 9.28794146e-01 -1.15689981e+00
6.72798574e-01 2.97176749e-01 6.10978484e-01 -8.75784278e-01
8.10334980e-01 -1.07320213e+00 1.08529961e+00 5.49700201e-01
9.94310156e-02 -4.13080379e-02 1.83752537e-01 5.51702082e-01
-2.73889571e-01 -5.62721670e-01 1.01733243e+00 -6.84357584e-01
-1.23042040e-01 1.28525838e-01 -5.12076318e-01 -4.03017163e-01
6.85922444e-01 -9.93344635e-02 1.22212656e-01 -3.47994328e-01
-1.36194050e+00 -3.71674180e-01 2.75309920e-01 -4.36734408e-01
6.60257995e-01 -1.26785970e+00 -1.03083253e+00 8.21692124e-02
-6.66640759e-01 3.84501725e-01 8.51915896e-01 1.12225080e+00
-9.77338195e-01 -3.00247014e-01 -5.20246252e-02 -9.05543506e-01
-1.07151759e+00 2.39528343e-01 8.24413955e-01 -7.46358156e-01
-8.35995078e-01 6.44761264e-01 -1.87356457e-01 -5.11942685e-01
-2.23623648e-01 -8.67398530e-02 4.56283718e-01 -1.58187479e-01
3.39563727e-01 4.75525349e-01 6.75059915e-01 -4.64745879e-01
-1.62727475e-01 6.11352026e-01 1.05178699e-01 -5.39391398e-01
1.88527608e+00 -1.08157828e-01 -5.01350880e-01 -3.27714160e-02
1.07665718e+00 -4.29127246e-01 -1.12219286e+00 -2.78978553e-02
-5.22815526e-01 -2.26199687e-01 7.21169412e-01 -8.38819742e-01
-1.17226231e+00 7.97732055e-01 8.99831295e-01 2.94721723e-01
1.27040005e+00 -6.73107430e-02 7.88041294e-01 2.75508091e-02
3.59371871e-01 -9.46103930e-01 -1.22076655e-02 2.44314656e-01
9.33847249e-01 -9.59748685e-01 4.63084728e-01 -2.43549287e-01
2.79478636e-02 1.23694229e+00 8.69031772e-02 -1.60043567e-01
1.05880165e+00 6.34150922e-01 1.21195346e-01 -5.06239176e-01
-3.70132715e-01 1.32286504e-01 -6.00451566e-02 5.02379715e-01
1.20614521e-01 -1.52534619e-01 -1.67008713e-01 1.12809122e-01
3.73023868e-01 1.89012364e-01 5.19035518e-01 1.23177648e+00
-5.45406416e-02 -1.28757882e+00 -8.79265845e-01 2.52241701e-01
-7.24282384e-01 1.12219935e-03 4.15184051e-01 8.02949369e-01
-5.23782745e-02 6.44632816e-01 -1.07060678e-01 1.40359834e-01
2.06025913e-01 -1.55231550e-01 8.65329146e-01 -2.76381403e-01
-9.12867904e-01 6.22224271e-01 5.33529632e-02 -5.58431923e-01
-8.81054938e-01 -6.24849796e-01 -1.12391329e+00 -4.67303246e-01
-2.16172501e-01 -1.54862925e-03 1.16497695e+00 9.19735491e-01
1.77177846e-01 7.91439772e-01 3.16903740e-01 -1.59195757e+00
-6.36244595e-01 -8.84415805e-01 -6.85624421e-01 1.91838771e-01
2.98954993e-01 -5.16045690e-01 -4.12720919e-01 -2.47576535e-01] | [13.234664916992188, -2.507720470428467] |
7a118f25-5fc4-4366-9777-e2788b8c6cc0 | deep-de-aliasing-for-fast-compressive-sensing | 1705.07137 | null | http://arxiv.org/abs/1705.07137v1 | http://arxiv.org/pdf/1705.07137v1.pdf | Deep De-Aliasing for Fast Compressive Sensing MRI | Fast Magnetic Resonance Imaging (MRI) is highly in demand for many clinical
applications in order to reduce the scanning cost and improve the patient
experience. This can also potentially increase the image quality by reducing
the motion artefacts and contrast washout. However, once an image field of view
and the desired resolution are chosen, the minimum scanning time is normally
determined by the requirement of acquiring sufficient raw data to meet the
Nyquist-Shannon sampling criteria. Compressive Sensing (CS) theory has been
perfectly matched to the MRI scanning sequence design with much less required
raw data for the image reconstruction. Inspired by recent advances in deep
learning for solving various inverse problems, we propose a conditional
Generative Adversarial Networks-based deep learning framework for de-aliasing
and reconstructing MRI images from highly undersampled data with great promise
to accelerate the data acquisition process. By coupling an innovative content
loss with the adversarial loss our de-aliasing results are more realistic.
Furthermore, we propose a refinement learning procedure for training the
generator network, which can stabilise the training with fast convergence and
less parameter tuning. We demonstrate that the proposed framework outperforms
state-of-the-art CS-MRI methods, in terms of reconstruction error and
perceptual image quality. In addition, our method can reconstruct each image in
0.22ms--0.37ms, which is promising for real-time applications. | ['Yike Guo', 'Greg Slabaugh', 'Simon Arridge', 'Jennifer Keegan', 'Hao Dong', 'Pier Luigi Dragotti', 'Guang Yang', 'Fangde Liu', 'David Firmin', 'Xujiong Ye', 'Simiao Yu'] | 2017-05-19 | null | null | null | null | ['de-aliasing'] | ['computer-vision'] | [ 8.23682964e-01 9.37182605e-02 8.52696821e-02 -2.75673896e-01
-9.46860015e-01 -1.80329531e-01 3.07414889e-01 -1.51674449e-01
-6.89458907e-01 7.17366159e-01 1.08015031e-01 -1.98970646e-01
-2.98023403e-01 -6.16531312e-01 -7.24704266e-01 -1.03541303e+00
-1.36737019e-01 2.38427907e-01 1.11209519e-01 8.93715322e-02
3.12399529e-02 4.77019697e-01 -9.48153198e-01 1.41475722e-01
9.70158577e-01 1.09073865e+00 6.97582901e-01 4.56361026e-01
3.47896546e-01 9.19961274e-01 -2.28348121e-01 -5.45779951e-02
2.56647199e-01 -7.11963236e-01 -7.17676103e-01 -9.88619216e-03
-7.23435660e-04 -4.51173604e-01 -6.26588225e-01 1.20179534e+00
9.42864001e-01 1.51569545e-01 3.68945688e-01 -5.77023983e-01
-3.56024235e-01 5.74608207e-01 -5.89313030e-01 3.52168649e-01
-1.05335668e-01 1.53594807e-01 2.64256328e-01 -7.49329329e-01
5.89998722e-01 5.71744025e-01 6.06214821e-01 5.97964704e-01
-1.24067473e+00 -6.24126673e-01 -4.68507737e-01 3.02839339e-01
-1.27841377e+00 -5.31097651e-01 8.70440125e-01 -3.24702591e-01
2.67085850e-01 3.85053992e-01 5.46810746e-01 9.50071037e-01
4.14130002e-01 5.99293232e-01 1.31111944e+00 -4.47427243e-01
2.55684108e-01 -2.90753454e-01 -6.27585232e-01 4.35855091e-01
6.16711006e-02 1.07493304e-01 -2.18579486e-01 1.49272710e-01
1.28897202e+00 1.80920199e-01 -6.70687616e-01 -4.35788304e-01
-1.45768023e+00 7.36216247e-01 6.70461774e-01 5.68102360e-01
-7.29180098e-01 2.10954115e-01 3.84497851e-01 6.27592877e-02
1.73113331e-01 4.70641077e-01 1.44891575e-01 2.08864361e-01
-1.29499948e+00 -9.65647399e-02 2.42213085e-01 4.76453066e-01
2.35618670e-02 3.92686367e-01 -1.48709223e-01 8.95416200e-01
5.90362139e-02 6.46033764e-01 8.37653697e-01 -1.13967121e+00
1.33827835e-01 -1.75235853e-01 2.26555001e-02 -6.97441101e-01
-4.18186188e-01 -8.50600719e-01 -1.39271438e+00 9.64436308e-02
3.07613790e-01 2.97437701e-02 -7.92887807e-01 1.67038906e+00
3.37313414e-01 4.43014354e-01 -1.01692230e-01 1.21502328e+00
5.39945483e-01 4.84334439e-01 -1.06152132e-01 -6.13816500e-01
1.25651121e+00 -6.79000914e-01 -9.78188276e-01 -3.62045079e-01
2.45298848e-01 -7.41690814e-01 9.60083902e-01 5.61870039e-01
-1.45374990e+00 -3.97676200e-01 -1.21661949e+00 3.79155099e-01
6.00846291e-01 -2.48637035e-01 4.63712364e-01 5.98741114e-01
-8.69844258e-01 7.01393902e-01 -1.13615572e+00 4.11064714e-01
6.28833115e-01 3.99512053e-01 -2.19233096e-01 -4.82855111e-01
-1.26848423e+00 8.24850380e-01 3.38927567e-01 1.63034156e-01
-9.37057674e-01 -1.08590853e+00 -7.16428578e-01 -1.79397285e-01
2.99186081e-01 -6.33438647e-01 1.07361257e+00 -7.91814387e-01
-1.51132762e+00 7.76214063e-01 3.34117681e-01 -6.57684684e-01
7.73197770e-01 -9.38604847e-02 -4.59781080e-01 6.64525032e-01
4.74127010e-02 4.67563599e-01 1.16216731e+00 -1.02097631e+00
-2.54019517e-02 -3.47884148e-01 -3.54669064e-01 1.41639069e-01
1.79088693e-02 -1.57064632e-01 -3.00235212e-01 -9.22699749e-01
4.00594473e-01 -1.01041031e+00 -5.45722663e-01 1.26324845e-02
-1.06849879e-01 7.31834948e-01 3.74896228e-01 -9.22493577e-01
9.82235193e-01 -2.10985923e+00 4.32836078e-02 1.22468509e-01
3.40157151e-01 3.65235597e-01 -1.95903052e-02 1.17065106e-02
-3.59494269e-01 -3.16351056e-01 -5.49364805e-01 9.36989561e-02
-5.87536573e-01 5.53548038e-02 -1.41580388e-01 9.32458699e-01
-2.05716893e-01 9.13170576e-01 -1.04391026e+00 -4.20493066e-01
4.35826451e-01 7.00196683e-01 -6.48989499e-01 3.78839165e-01
3.87495935e-01 1.17502403e+00 -3.20683360e-01 1.10793583e-01
8.57475579e-01 -4.13272232e-01 3.76381785e-01 -3.93257737e-01
6.48708642e-02 5.39946137e-03 -1.00384891e+00 1.99500883e+00
-7.56108046e-01 4.08669651e-01 3.88744265e-01 -1.36591113e+00
5.93446553e-01 4.93669868e-01 9.84314740e-01 -1.17361975e+00
1.57622963e-01 5.07611394e-01 2.05493242e-01 -6.26139700e-01
-1.37374684e-01 -7.86493719e-01 2.39347234e-01 3.07221472e-01
-1.94720984e-01 -3.81908715e-01 -2.36076713e-01 5.06968908e-02
9.79659438e-01 -2.07861692e-01 1.05263032e-01 -3.56354415e-01
5.34365952e-01 -4.84285563e-01 4.51231748e-01 6.85714304e-01
-1.23987980e-01 8.44558835e-01 -2.17476152e-02 -2.46984005e-01
-1.36704302e+00 -1.10061455e+00 -5.04776359e-01 4.49741751e-01
2.26653680e-01 3.51575583e-01 -8.90505910e-01 -1.99949920e-01
-6.16690338e-01 5.86442769e-01 -3.00756842e-01 -3.01158547e-01
-9.39118862e-01 -7.68143237e-01 3.86961371e-01 2.86012650e-01
3.57680172e-01 -1.14113081e+00 -9.34026122e-01 4.59576309e-01
-6.81755900e-01 -1.24245894e+00 -5.75356245e-01 1.09542578e-01
-1.02773619e+00 -7.98767090e-01 -1.17466915e+00 -7.19076872e-01
7.53622711e-01 3.08620661e-01 9.03314769e-01 -7.76800662e-02
-4.43813324e-01 9.11005735e-02 -2.34132171e-01 -2.85696760e-02
-6.52303517e-01 -2.47677609e-01 1.54940002e-02 -2.32210495e-02
-4.05346662e-01 -7.54292190e-01 -1.24754977e+00 1.17534250e-01
-1.45728314e+00 2.89308578e-01 7.54921675e-01 1.09690464e+00
6.82736158e-01 1.69427574e-01 5.72261512e-01 -7.58047760e-01
2.69279510e-01 -4.44813997e-01 -3.53859425e-01 -9.43467468e-02
-4.75920022e-01 2.59330481e-01 7.45471060e-01 -6.09145820e-01
-8.88550043e-01 -5.47510572e-02 -4.33305264e-01 -4.57664639e-01
2.34264523e-01 4.63848054e-01 2.10423529e-01 -3.29260528e-01
6.83537245e-01 5.63986480e-01 4.73116666e-01 -9.85845923e-02
1.26348361e-01 3.23009342e-01 8.46945107e-01 -2.25328848e-01
5.02263784e-01 6.71068907e-01 2.77415514e-01 -8.57038796e-01
-7.10312188e-01 -3.13196003e-01 -3.82591039e-01 -3.55545193e-01
8.40955138e-01 -8.73340964e-01 -7.01994538e-01 4.70499277e-01
-7.93829620e-01 -2.34510437e-01 -3.09099495e-01 8.59045744e-01
-8.60408187e-01 7.03146696e-01 -7.57981360e-01 -4.79556262e-01
-6.18956864e-01 -1.61319637e+00 6.36108041e-01 -1.20239787e-01
8.02093819e-02 -8.05561662e-01 -2.10503474e-01 3.85346055e-01
8.06391001e-01 4.97065306e-01 7.36567616e-01 -1.58814803e-01
-5.54074347e-01 -6.15952797e-02 3.52769382e-02 5.31427681e-01
1.50188273e-02 -1.10383594e+00 -6.58255279e-01 -6.14936829e-01
7.61260688e-01 -2.84839541e-01 6.19942963e-01 1.01259995e+00
1.48352182e+00 -3.04052204e-01 3.91093530e-02 9.34352815e-01
1.54168940e+00 3.74387026e-01 9.13230062e-01 1.90982446e-01
5.06593049e-01 3.86485755e-01 3.55835170e-01 5.46554804e-01
-1.29933968e-01 6.11913502e-01 5.23937523e-01 -2.70596951e-01
-3.42859119e-01 -4.55540530e-02 -1.46736419e-02 1.03504789e+00
9.52490941e-02 7.52943605e-02 -8.31862569e-01 5.35930276e-01
-1.39341128e+00 -9.80860054e-01 -2.83480100e-02 2.46454549e+00
1.08415616e+00 -2.87267920e-02 -1.51642546e-01 3.64109784e-01
6.51834965e-01 1.25170037e-01 -6.92826569e-01 1.97461829e-01
2.03442559e-01 6.19855821e-01 7.22249329e-01 5.70875406e-01
-8.72831762e-01 2.91178048e-01 5.88967180e+00 9.29338574e-01
-1.49703729e+00 4.32137132e-01 7.53945172e-01 -1.29316360e-01
-3.03445786e-01 -4.04763669e-01 -2.86159627e-02 5.36748827e-01
9.93053555e-01 8.78679156e-02 7.59616792e-01 5.13365746e-01
3.34987998e-01 -3.49172251e-03 -7.11187065e-01 1.34105146e+00
-1.00134723e-01 -1.45972514e+00 -2.65864015e-01 1.14401588e-02
6.73721850e-01 4.41267863e-02 2.06627682e-01 -2.20790327e-01
-2.55548090e-01 -1.13767934e+00 7.25614429e-01 4.41522956e-01
1.25681591e+00 -8.05719197e-01 5.87349534e-01 4.60243016e-01
-6.00571871e-01 1.27880603e-01 -1.64305329e-01 4.17107940e-01
3.96127552e-01 1.00161982e+00 -7.28816211e-01 3.80289912e-01
4.63728845e-01 2.69101501e-01 1.10087721e-02 1.04588974e+00
-2.43683755e-02 4.46329951e-01 -1.44340709e-01 4.92432863e-01
2.88187712e-01 -3.19444150e-01 5.87861300e-01 8.28598320e-01
4.95013475e-01 3.59236598e-01 -5.25523685e-02 7.31000841e-01
4.60012667e-02 -1.40047222e-01 -3.02024513e-01 2.26503298e-01
3.77262801e-01 1.02549314e+00 -7.85914898e-01 -1.06512189e-01
-1.85160890e-01 1.09624934e+00 -2.52396673e-01 2.57618338e-01
-9.35781896e-01 -2.42455557e-01 2.21959464e-02 5.88725209e-01
2.19098210e-01 -1.72244787e-01 -2.58784175e-01 -1.11655724e+00
9.26245824e-02 -1.01635456e+00 4.09122370e-03 -5.65990925e-01
-9.77077842e-01 7.74484336e-01 -2.48595610e-01 -1.31017339e+00
-3.49895805e-01 -1.80471316e-01 -1.91192150e-01 8.29311073e-01
-1.62128532e+00 -7.58531332e-01 -2.82101512e-01 6.38388991e-01
3.50147635e-01 7.34707564e-02 6.48336649e-01 7.56303608e-01
-4.35342714e-02 3.86565149e-01 2.66903907e-01 -2.53431983e-02
5.98119438e-01 -9.21011448e-01 6.05702884e-02 8.79906476e-01
-1.97261706e-01 4.92375255e-01 8.84782374e-01 -4.69512731e-01
-1.42309988e+00 -8.86358798e-01 2.80577600e-01 2.09786579e-01
5.73839188e-01 -6.66352585e-02 -1.00175750e+00 2.10658044e-01
1.14125811e-01 3.66819084e-01 5.89392543e-01 -6.68966711e-01
1.60111666e-01 -1.97468907e-01 -1.43399298e+00 3.41400385e-01
7.61249721e-01 -4.63248968e-01 -2.97192365e-01 3.38652164e-01
6.30358696e-01 -6.97785079e-01 -1.09813857e+00 4.56208438e-01
4.56578314e-01 -9.48473811e-01 1.34903097e+00 -7.03207478e-02
7.50290275e-01 -5.64134419e-02 3.31866816e-02 -1.35811615e+00
-4.48402554e-01 -6.91887498e-01 -1.85188639e-03 5.48580110e-01
-5.85909933e-03 -5.36005795e-01 8.25611353e-01 3.87067527e-01
-2.05959097e-01 -7.92478025e-01 -1.09417641e+00 -6.09434307e-01
9.28799659e-02 -3.75714570e-01 2.81969935e-01 8.96326005e-01
-2.75070250e-01 -2.82257497e-02 -6.76438451e-01 1.33128166e-01
1.13260210e+00 -7.75730684e-02 8.26093485e-04 -7.42394805e-01
-5.64004660e-01 -2.89642781e-01 -2.30870515e-01 -1.14933932e+00
-1.83625832e-01 -9.81849134e-01 1.22736566e-01 -1.28738105e+00
2.12452844e-01 -7.12806225e-01 -4.76443380e-01 -2.70382091e-02
-2.24664301e-01 4.38609898e-01 1.77096009e-01 3.52864534e-01
-1.70301154e-01 5.20601869e-01 1.92365551e+00 -3.22620310e-02
1.23560250e-01 7.78068379e-02 -5.36152363e-01 4.68415260e-01
6.39166355e-01 -4.71969664e-01 -5.70152342e-01 -4.56447780e-01
1.28780469e-01 7.81364143e-01 3.29682052e-01 -1.14435196e+00
1.85193896e-01 6.96924403e-02 4.50537831e-01 -2.55285084e-01
1.81016088e-01 -9.94482219e-01 4.90942776e-01 8.82885396e-01
-4.56604779e-01 -3.56169999e-01 -8.53080899e-02 3.72064590e-01
-2.87702382e-01 -4.12234068e-01 1.42250717e+00 -2.55638391e-01
-2.49709338e-01 4.85878527e-01 -1.74338624e-01 1.74257338e-01
8.40876698e-01 -7.47547075e-02 3.21604997e-01 -4.51204717e-01
-8.78714502e-01 -2.39338458e-01 1.56265199e-01 9.15252790e-02
8.99358451e-01 -1.34858727e+00 -7.99785316e-01 4.21296328e-01
-3.77123147e-01 -4.44811489e-03 8.27168643e-01 1.24446356e+00
-8.69886100e-01 3.08375359e-01 -3.80237758e-01 -6.66454673e-01
-6.01857364e-01 6.27936840e-01 3.55966508e-01 -5.47210932e-01
-9.08962846e-01 7.03769088e-01 2.43875533e-01 -1.28468156e-01
-1.30762709e-02 -1.01932334e-02 9.45755392e-02 -3.80652755e-01
8.30877244e-01 1.79899171e-01 2.32217744e-01 -3.73823404e-01
-2.76013285e-01 3.08075905e-01 4.00779536e-03 -2.44908750e-01
1.52875662e+00 -1.08669475e-01 7.89337605e-02 -2.88814865e-02
1.25354755e+00 -1.33860588e-01 -1.41816103e+00 -3.69267642e-01
-3.13496500e-01 -4.77022707e-01 6.18401110e-01 -7.03372777e-01
-1.40518701e+00 9.25712347e-01 9.19772148e-01 7.28205740e-02
1.43046498e+00 -2.45555535e-01 1.17982042e+00 -3.05533379e-01
5.64933240e-01 -6.98162377e-01 2.42970914e-01 -5.01412153e-02
9.84356999e-01 -1.16794837e+00 6.88531175e-02 -2.65570521e-01
-7.50380516e-01 9.10399497e-01 -8.07371810e-02 -2.32188448e-01
4.87586349e-01 5.05486608e-01 -1.07715078e-01 -1.08396076e-01
-8.06477740e-02 3.25161368e-01 2.21482009e-01 5.24661243e-01
4.34321940e-01 1.88804999e-01 -4.32610869e-01 2.85650671e-01
7.38632977e-02 2.45314419e-01 4.52489078e-01 5.98836482e-01
-2.69726634e-01 -9.37015176e-01 -4.44236219e-01 3.68981153e-01
-9.52020288e-01 -2.70043015e-01 6.45610213e-01 4.05533463e-01
-1.75679937e-01 6.57522321e-01 -2.39827543e-01 1.45157441e-01
9.75861400e-02 -3.41186345e-01 8.59867573e-01 -3.05577368e-01
-2.93831557e-01 4.32621717e-01 -3.77462655e-01 -5.77467263e-01
-5.08206666e-01 -4.79803234e-01 -1.35898256e+00 -2.46755198e-01
-2.41408497e-01 2.04945244e-02 7.75650859e-01 7.80251443e-01
2.05044746e-01 7.72769570e-01 8.54515493e-01 -8.15195978e-01
-7.17306197e-01 -7.32771575e-01 -6.36965454e-01 5.32722592e-01
3.68184179e-01 -3.30735356e-01 -1.62641838e-01 8.27202722e-02] | [13.552154541015625, -2.392425537109375] |
ccd5d32c-6c9f-47b4-8284-5b198284e242 | modality-based-factorization-for-multimodal | 1811.12624 | null | https://arxiv.org/abs/1811.12624v3 | https://arxiv.org/pdf/1811.12624v3.pdf | Modality-based Factorization for Multimodal Fusion | We propose a novel method, Modality-based Redundancy Reduction Fusion (MRRF), for understanding and modulating the relative contribution of each modality in multimodal inference tasks. This is achieved by obtaining an $(M+1)$-way tensor to consider the high-order relationships between $M$ modalities and the output layer of a neural network model. Applying a modality-based tensor factorization method, which adopts different factors for different modalities, results in removing information present in a modality that can be compensated by other modalities, with respect to model outputs. This helps to understand the relative utility of information in each modality. In addition it leads to a less complicated model with less parameters and therefore could be applied as a regularizer avoiding overfitting. We have applied this method to three different multimodal datasets in sentiment analysis, personality trait recognition, and emotion recognition. We are able to recognize relationships and relative importance of different modalities in these tasks and achieves a 1\% to 4\% improvement on several evaluation measures compared to the state-of-the-art for all three tasks. | ['Pascale Fung', 'Peyman Momeni', 'Elham J. Barezi'] | 2018-11-30 | modality-based-factorization-for-multimodal-1 | https://aclanthology.org/W19-4331 | https://aclanthology.org/W19-4331.pdf | ws-2019-8 | ['personality-trait-recognition'] | ['computer-vision'] | [ 3.50691557e-01 -2.82965805e-02 5.89543991e-02 -4.35745209e-01
-3.50007802e-01 -2.62721390e-01 6.64945841e-01 2.76509613e-01
-6.81791246e-01 6.54318333e-01 2.78227299e-01 1.73934162e-01
-3.50876778e-01 -5.43822825e-01 -5.24654329e-01 -7.33747244e-01
7.72799626e-02 1.93434596e-01 -2.62609780e-01 -3.80316496e-01
2.37959181e-03 3.38563740e-01 -1.79508972e+00 7.61084855e-01
5.38980603e-01 1.17400825e+00 -9.15054791e-03 4.19230342e-01
4.29213978e-03 8.99111450e-01 -2.36601830e-01 -6.31835401e-01
1.32233417e-02 -2.78490484e-01 -8.74447167e-01 9.40769315e-02
3.88727754e-01 1.88642055e-01 5.23279561e-03 9.45563018e-01
2.32657239e-01 2.96352535e-01 6.90763831e-01 -8.98472369e-01
-1.24044083e-01 8.28187585e-01 -5.76357603e-01 -1.60819307e-01
5.12011945e-01 -3.48398745e-01 1.07371497e+00 -8.05864096e-01
7.32536018e-01 1.28007030e+00 3.37795109e-01 4.53285992e-01
-1.45272112e+00 -2.80545950e-01 1.19828589e-01 1.99066475e-01
-1.21800411e+00 -4.78313297e-01 1.08476162e+00 -5.06993949e-01
9.09241617e-01 5.30412972e-01 4.20823693e-01 8.95877302e-01
5.72765060e-02 6.44521832e-01 1.39967597e+00 -5.83580554e-01
-2.74877436e-02 4.81420577e-01 2.47018844e-01 7.22544849e-01
-2.26320758e-01 -2.41403505e-01 -9.26886380e-01 -6.11628070e-02
1.65722460e-01 -1.16071766e-02 -9.15180519e-02 -7.48817101e-02
-1.38280976e+00 5.06429017e-01 4.82656062e-01 8.21745694e-01
-6.69906199e-01 9.24279839e-02 3.20319384e-01 3.46557021e-01
3.45001280e-01 3.89286935e-01 -5.30789793e-01 -1.39936917e-02
-7.07435250e-01 5.75444922e-02 6.50977850e-01 1.43302202e-01
8.32493365e-01 5.31466976e-02 -1.64863765e-01 1.15357459e+00
4.82557654e-01 4.26447988e-01 4.29774046e-01 -1.05081928e+00
3.62399042e-01 1.17723298e+00 -1.30077243e-01 -1.20066547e+00
-6.17935538e-01 -3.78142476e-01 -1.08469307e+00 1.98830381e-01
3.66961837e-01 -7.22967833e-02 -6.80569708e-01 1.97300577e+00
1.91195399e-01 -1.83451623e-01 7.64812380e-02 8.94178033e-01
7.40931571e-01 4.23450887e-01 1.21153988e-01 -2.63304859e-01
1.67382193e+00 -4.97990012e-01 -5.83284378e-01 -3.72086614e-02
5.17412603e-01 -7.88347125e-01 7.36129224e-01 7.39128232e-01
-1.14919007e+00 -6.51572168e-01 -8.88828397e-01 -1.47856185e-02
-6.06329262e-01 4.62760687e-01 6.76208198e-01 6.02854609e-01
-9.67690885e-01 6.67578220e-01 -6.39674127e-01 -2.27391377e-01
1.62692174e-01 6.65347934e-01 -7.82257199e-01 -1.80725798e-01
-1.12107074e+00 1.12111616e+00 2.02954501e-01 2.32916847e-01
-3.96565795e-01 -5.09744942e-01 -8.29268575e-01 2.66889464e-02
3.19552183e-01 -7.83421993e-01 5.24394751e-01 -1.01372421e+00
-1.34137940e+00 6.24657393e-01 -5.99444449e-01 -2.01875374e-01
3.65929939e-02 8.56391992e-03 -4.00782585e-01 2.24857971e-01
-3.70127857e-01 7.00654209e-01 1.01704574e+00 -1.31136072e+00
-4.18822557e-01 -6.71919525e-01 1.71910256e-01 3.29230636e-01
-4.90620732e-01 1.71904769e-02 -3.36358458e-01 -3.03454995e-01
3.88630062e-01 -9.60450113e-01 -4.19472158e-02 -3.79144162e-01
-3.08411330e-01 -1.28576115e-01 4.86599028e-01 -7.46447980e-01
9.70167518e-01 -1.94259703e+00 1.20019901e+00 3.93770665e-01
2.74942994e-01 8.42052028e-02 -1.37820497e-01 3.84741038e-01
-2.58999020e-01 7.34537169e-02 -2.44100168e-01 -8.64854574e-01
3.91530842e-02 2.52431989e-01 -4.36954945e-02 1.48622453e-01
3.17618191e-01 6.19064689e-01 -4.25720245e-01 -2.67164558e-01
4.11491960e-01 8.47708464e-01 -2.39968956e-01 -9.15732235e-02
3.04095708e-02 5.24164379e-01 -7.81019256e-02 6.96321130e-01
5.80257833e-01 1.99342165e-02 3.74235928e-01 -5.95523715e-01
-1.21073145e-03 -4.11285833e-02 -1.21195543e+00 1.72041655e+00
-6.63200259e-01 4.60664392e-01 1.14181533e-01 -1.15828943e+00
8.79575074e-01 3.21331650e-01 4.96291190e-01 -7.95563579e-01
3.87224197e-01 1.44594878e-01 1.16593488e-01 -5.48023283e-01
7.08753705e-01 -3.09405595e-01 -1.24877423e-01 3.54127347e-01
3.80196691e-01 2.14816973e-01 4.86674488e-01 1.68908224e-01
7.54784167e-01 7.13705570e-02 -7.50349537e-02 -5.34235984e-02
1.11020386e+00 -6.08541131e-01 2.43106633e-01 4.43084508e-01
1.41467214e-01 1.69721127e-01 7.38149285e-01 -2.67028958e-01
-6.28878057e-01 -5.16014397e-01 3.31932716e-02 1.18714345e+00
-3.00238818e-01 -3.13316405e-01 -4.42810506e-01 -5.22939980e-01
-2.68190224e-02 6.06165588e-01 -8.51949275e-01 -1.19171821e-01
-2.20058441e-01 -9.27772045e-01 4.34526503e-01 2.71454126e-01
2.04464495e-01 -7.97067761e-01 -3.36595386e-01 -1.95149392e-01
-5.15054464e-01 -1.10057557e+00 1.32471830e-01 2.67535657e-01
-9.70922947e-01 -9.43709075e-01 -3.94013226e-01 -2.04824865e-01
7.09463894e-01 -5.98340435e-03 8.93057346e-01 -1.15845978e-01
5.41261435e-02 5.46898246e-01 -3.53893548e-01 -1.78150594e-01
-2.56931454e-01 -4.84298281e-02 8.71057138e-02 6.22505844e-01
2.69216061e-01 -4.20938611e-01 -3.15696746e-01 1.07191883e-01
-1.16618061e+00 -2.00557962e-01 6.48634493e-01 9.26272094e-01
3.19410890e-01 3.58327664e-02 2.35464230e-01 -7.53176212e-01
6.89499259e-01 -2.92080820e-01 -3.37506175e-01 2.20473304e-01
-6.25975311e-01 4.62170154e-01 5.61709642e-01 -3.11540842e-01
-1.04771948e+00 1.13827229e-01 -2.77563445e-02 -3.86159718e-01
-1.21171951e-01 9.07200873e-01 -1.64096639e-01 -2.24247247e-01
6.01311266e-01 1.10903017e-01 1.81425646e-01 -6.36764288e-01
3.71257991e-01 3.05165708e-01 1.85090750e-02 -5.50560892e-01
4.21653897e-01 5.81887841e-01 5.40596008e-01 -8.87337983e-01
-3.55495900e-01 -4.37394381e-01 -6.25477672e-01 -4.11425710e-01
6.08444333e-01 -8.62470269e-01 -1.14437163e+00 5.58629870e-01
-1.01605153e+00 4.05190825e-01 -2.33708024e-01 5.92191637e-01
-8.48490372e-02 4.16564435e-01 -4.50796634e-01 -9.56765771e-01
-9.26167369e-02 -1.33207333e+00 8.47887814e-01 2.04729155e-01
-4.33789194e-02 -1.02904570e+00 -1.71477627e-02 7.74389744e-01
4.21900302e-01 4.76036407e-02 9.55935001e-01 -3.89700472e-01
-2.00170830e-01 -2.88737565e-01 -2.19339967e-01 7.07042933e-01
-1.76723182e-01 -6.63066283e-02 -1.21850264e+00 -6.33043870e-02
-3.62850726e-03 -1.97271407e-01 1.33275938e+00 1.67558894e-01
6.99650943e-01 -1.61002800e-01 1.99138876e-02 1.35278061e-01
1.31525671e+00 -6.23891987e-02 6.70573473e-01 -1.67430326e-01
8.36073816e-01 1.03344798e+00 3.38129282e-01 2.83100247e-01
5.84255576e-01 8.51320744e-01 5.16136348e-01 -4.49608900e-02
1.67694315e-01 2.59990662e-01 4.67214912e-01 7.83415616e-01
-6.24248803e-01 6.08235504e-03 -5.49919426e-01 3.83317262e-01
-1.92960584e+00 -8.97984028e-01 -3.73263389e-01 2.30550718e+00
5.44497907e-01 -1.16913967e-01 1.82740316e-01 4.01015729e-01
2.65855908e-01 1.05350144e-01 -1.87919214e-02 -6.74735546e-01
-2.92828828e-01 3.37335587e-01 1.87061653e-01 5.74519336e-01
-1.04367399e+00 5.30278683e-01 5.94904995e+00 5.84600210e-01
-1.27084947e+00 1.95594914e-02 4.29384500e-01 -6.81658611e-02
-5.19702077e-01 1.10422019e-02 -4.77160424e-01 2.37904340e-01
9.53574598e-01 3.35550755e-01 8.01775396e-01 4.41940695e-01
1.66968480e-02 -4.64662910e-01 -9.65409696e-01 1.00690269e+00
3.37032378e-01 -9.29567873e-01 5.73518462e-02 -1.95094175e-03
4.18934077e-01 -1.73218131e-01 1.45659506e-01 1.99383244e-01
-3.72809649e-01 -8.44070733e-01 4.51770991e-01 1.06626880e+00
1.98920250e-01 -8.05015147e-01 9.18503702e-01 2.17574865e-01
-9.42823470e-01 -1.29210413e-01 -1.78819731e-01 -1.75328746e-01
-4.66641039e-02 8.60567451e-01 -5.95034122e-01 9.69381034e-01
5.83638847e-01 5.32707393e-01 -5.96653759e-01 2.81302184e-01
5.11365719e-02 2.16252863e-01 -3.67439270e-01 1.04423948e-02
-1.97391376e-01 -3.36936951e-01 5.75582266e-01 1.01656711e+00
2.01106384e-01 -2.06178859e-01 -2.99386173e-01 7.42129266e-01
-7.42362887e-02 2.86623746e-01 -5.26487589e-01 -3.36279333e-01
-1.78778633e-01 1.67160928e+00 -4.42245126e-01 -2.35951766e-01
-2.64249861e-01 9.53853846e-01 2.43529111e-01 2.76507884e-01
-5.38084030e-01 -1.63401112e-01 5.13228536e-01 -3.40166241e-01
2.97651827e-01 -2.79170573e-01 -2.36885577e-01 -1.38279903e+00
1.05854668e-01 -8.06871772e-01 4.02376741e-01 -7.50614226e-01
-1.11980605e+00 7.99086630e-01 -2.04634145e-02 -1.20237029e+00
-4.41259831e-01 -8.76668394e-01 -2.30866093e-02 1.22837269e+00
-1.39422810e+00 -1.37837112e+00 -2.07384564e-02 8.03110659e-01
-1.48882657e-01 -1.77253753e-01 1.08495533e+00 5.25787234e-01
-6.15950108e-01 5.13488293e-01 -3.07114065e-01 -2.17579335e-01
6.04054749e-01 -1.19068289e+00 -6.71130121e-01 9.52189744e-01
1.45491540e-01 8.16504180e-01 6.60886824e-01 -2.65668899e-01
-1.74045610e+00 -5.51857173e-01 1.26725996e+00 -4.58387226e-01
5.13459206e-01 -1.05971426e-01 -6.78414106e-01 2.77912796e-01
3.25342596e-01 -3.04533452e-01 1.05103505e+00 6.22850895e-01
-7.43050039e-01 -4.84968364e-01 -1.17572773e+00 4.10138726e-01
5.25251508e-01 -7.38840044e-01 -4.21005398e-01 -1.63918287e-01
3.63673955e-01 -2.27725245e-02 -1.17398071e+00 4.22010481e-01
8.90620172e-01 -1.07084191e+00 8.75929773e-01 -6.90474927e-01
7.19042897e-01 -3.89698237e-01 -4.82662827e-01 -1.33540726e+00
1.20939373e-03 -1.42617330e-01 -1.71540156e-01 1.09067965e+00
6.43238723e-01 -5.73842764e-01 3.69729251e-01 9.22246516e-01
2.69221455e-01 -6.01264834e-01 -1.03477073e+00 -1.25247955e-01
-3.54023546e-01 -6.38428152e-01 2.60867417e-01 1.06145287e+00
3.44686151e-01 4.99867827e-01 -7.56095946e-01 -9.08888206e-02
4.58167791e-01 -9.59981605e-03 4.39152002e-01 -1.13170874e+00
-4.00650442e-01 -4.34386641e-01 -4.82021272e-01 -4.05216157e-01
1.16576493e-01 -8.73284400e-01 -3.89587045e-01 -1.15250289e+00
1.63739517e-01 -5.22543937e-02 -6.75893903e-01 7.87298977e-01
-5.50448895e-02 6.60031617e-01 6.06940389e-01 -1.91108382e-03
-3.73810649e-01 3.95699471e-01 1.08733022e+00 -2.56620020e-01
-1.29251733e-01 -2.18190774e-01 -9.35015857e-01 6.95337355e-01
5.10857224e-01 -2.23057494e-01 -1.68093607e-01 -2.34065130e-01
6.51887000e-01 1.05988733e-01 4.07260954e-01 -8.59565854e-01
8.97536799e-02 8.19525793e-02 5.04896820e-01 -3.44366074e-01
9.07165825e-01 -1.08163846e+00 3.09606880e-01 2.86957532e-01
-3.81988525e-01 -1.99896812e-01 2.74492502e-01 1.93534181e-01
-4.13080335e-01 -3.12260240e-01 4.25776422e-01 -6.56633079e-02
-5.75827956e-01 -3.15234542e-01 -2.96778351e-01 -6.95531070e-01
3.86967838e-01 5.89799620e-02 -3.62510622e-01 -2.90540159e-01
-1.03464222e+00 -7.20099956e-02 9.31526795e-02 4.75971222e-01
5.54797053e-01 -1.21028674e+00 -5.77271342e-01 1.99908257e-01
1.02637574e-01 -6.44617856e-01 7.36405969e-01 1.35810947e+00
2.40577068e-02 2.93845713e-01 -4.31520581e-01 -5.83660662e-01
-1.57827532e+00 3.79442841e-01 2.54990041e-01 -5.80556810e-01
3.06673318e-01 6.79432988e-01 -3.61936569e-01 -4.25163150e-01
-2.90305838e-02 -2.39163503e-01 -8.04630756e-01 6.34016275e-01
3.81202728e-01 5.14794588e-01 2.72395879e-01 -1.16118646e+00
-5.73067427e-01 7.68729150e-01 -6.12226427e-02 -4.66894537e-01
1.15981209e+00 -7.42487535e-02 -6.37118220e-01 7.77610362e-01
1.21447980e+00 1.62757128e-01 -5.00208139e-01 -2.64948070e-01
-2.51379520e-01 -3.96590471e-01 2.03589752e-01 -1.12107158e+00
-1.25818717e+00 1.02600539e+00 7.34322608e-01 3.91097128e-01
1.50280452e+00 -1.59188867e-01 9.58239958e-02 4.07205939e-01
2.63890892e-01 -9.90954816e-01 -1.72103867e-01 5.01568735e-01
9.56623018e-01 -1.25807476e+00 1.55241266e-01 -3.33566040e-01
-9.56310570e-01 1.17213881e+00 1.09098151e-01 1.03793040e-01
6.58335209e-01 -1.71499535e-01 6.97948635e-02 -2.35461131e-01
-7.31416285e-01 -4.34292912e-01 7.80773580e-01 2.31141597e-01
6.50911689e-01 2.19864562e-01 -6.70395732e-01 5.44233441e-01
3.00466210e-01 4.02116738e-02 1.61593184e-01 6.49406731e-01
2.53718700e-02 -1.60715044e+00 -3.97605836e-01 4.00241047e-01
-4.65563357e-01 -1.39429979e-02 -4.56342191e-01 5.17290175e-01
4.31240618e-01 1.17282248e+00 -3.56285870e-01 -9.51560199e-01
3.52470070e-01 5.14713764e-01 6.54700816e-01 -1.68520138e-01
-8.32998037e-01 2.13732466e-01 2.67138124e-01 -6.65522277e-01
-9.91180837e-01 -8.73555660e-01 -9.06313002e-01 -1.38585865e-01
-1.69273183e-01 -7.92836100e-02 1.10785079e+00 1.11064219e+00
5.33871114e-01 6.85495138e-01 6.61902308e-01 -9.92081046e-01
-2.57343709e-01 -1.14888370e+00 -5.51097333e-01 6.88490570e-01
1.19791299e-01 -7.86305606e-01 -4.02765632e-01 1.51131451e-01] | [13.21065616607666, 5.1414794921875] |
f72e84d5-516c-4255-ab37-99f7fd1fc312 | fundus-image-analysis-for-age-related-macular | 2009.01548 | null | https://arxiv.org/abs/2009.01548v1 | https://arxiv.org/pdf/2009.01548v1.pdf | Fundus Image Analysis for Age Related Macular Degeneration: ADAM-2020 Challenge Report | Age related macular degeneration (AMD) is one of the major causes for blindness in the elderly population. In this report, we propose deep learning based methods for retinal analysis using color fundus images for computer aided diagnosis of AMD. We leverage the recent state of the art deep networks for building a single fundus image based AMD classification pipeline. We also propose methods for the other directly relevant and auxiliary tasks such as lesions detection and segmentation, fovea detection and optic disc segmentation. We propose the use of generative adversarial networks (GANs) for the tasks of segmentation and detection. We also propose a novel method of fovea detection using GANs. | ['Sharath M. Shankaranarayana'] | 2020-09-03 | null | null | null | null | ['fovea-detection'] | ['medical'] | [ 9.97654498e-02 2.58617967e-01 3.70496005e-01 -2.18926013e-01
-6.12418950e-01 -9.96993482e-02 2.17159480e-01 -4.38104600e-01
-5.90764344e-01 8.85888994e-01 1.05571404e-01 -6.96965635e-01
4.68502969e-01 -8.34973037e-01 -4.27942097e-01 -3.74022543e-01
1.28261641e-01 2.10511714e-01 2.43808880e-01 2.23914951e-01
1.47675797e-01 6.46129310e-01 -1.63897073e+00 5.61686337e-01
1.22267532e+00 1.06958389e+00 -3.07470262e-01 1.22916985e+00
9.67833921e-02 7.51326799e-01 -6.18787289e-01 -4.40421611e-01
5.01540303e-01 -7.33557522e-01 -9.13525343e-01 1.62226602e-01
1.08247983e+00 -1.06643319e+00 -3.35199356e-01 9.58556294e-01
1.07759106e+00 -3.32108885e-01 8.76342475e-01 -7.61824608e-01
-5.23680210e-01 -2.69523621e-01 -7.85704732e-01 6.10059381e-01
-2.20006526e-01 7.24582493e-01 1.41200304e-01 -5.81357598e-01
3.96293581e-01 1.25231731e+00 7.67649651e-01 9.54767644e-01
-8.08976173e-01 -1.78102478e-01 -2.77514815e-01 3.70690972e-01
-7.26596475e-01 -6.18840992e-01 6.66026995e-02 -9.40525472e-01
9.26326871e-01 -6.14175051e-02 1.28229964e+00 6.02721035e-01
2.46378675e-01 6.78176999e-01 1.37679315e+00 -5.37947536e-01
2.92747647e-01 -5.14006257e-01 1.76912788e-02 9.84347463e-01
6.34518981e-01 3.13820541e-01 2.26068467e-01 -1.71284378e-01
1.23288071e+00 -4.07445729e-01 2.09913161e-02 1.71545312e-01
-6.47016764e-01 8.79350245e-01 5.98139584e-01 -4.33890939e-01
-4.92143869e-01 4.87067282e-01 1.82387099e-01 -5.74924238e-03
4.23508674e-01 3.64872605e-01 -1.36810303e-01 1.34309009e-01
-7.31250048e-01 2.01791078e-01 1.97165146e-01 4.42148358e-01
4.05294836e-01 -3.22281457e-02 -8.18190694e-01 6.79403126e-01
5.67123532e-01 2.30350643e-01 5.04492342e-01 -1.51076472e+00
-4.80125993e-02 7.90129781e-01 3.72414172e-01 1.34396762e-01
-4.39592123e-01 -4.28083897e-01 -6.34842038e-01 1.35016334e+00
7.42054820e-01 -5.74401975e-01 -1.61831784e+00 9.90693986e-01
2.16992572e-01 1.28368974e-01 -1.90323278e-01 8.94757509e-01
8.92427683e-01 -2.77931631e-01 2.60914952e-01 1.75598085e-01
1.63037753e+00 -1.09968019e+00 -3.05285603e-01 -6.05970621e-01
4.83290076e-01 -7.46387184e-01 7.59804666e-01 2.37675667e-01
-1.41319859e+00 -3.87298614e-01 -7.59665072e-01 -7.87203610e-01
-2.57596105e-01 9.79731739e-01 7.53700614e-01 9.81788099e-01
-1.56510854e+00 1.81318223e-02 -9.38331425e-01 -4.91631269e-01
1.32340682e+00 5.34453034e-01 -1.30936459e-01 7.22105103e-03
-4.39111710e-01 1.04291952e+00 1.38917575e-02 4.40087989e-02
-6.48806036e-01 -6.23378456e-01 -3.75581324e-01 -4.97643709e-01
-4.32705909e-01 -1.94567978e+00 1.49102664e+00 -7.70982981e-01
-1.34125149e+00 1.57094669e+00 -5.89992464e-01 -1.04437077e+00
7.76032269e-01 -6.46379173e-01 -1.30770609e-01 5.55054247e-01
1.07268728e-01 1.07826149e+00 1.00556338e+00 -5.72860062e-01
-9.96655583e-01 -6.54833853e-01 3.40940952e-02 -2.16893166e-01
1.99975982e-01 2.87869930e-01 -7.57082328e-02 -4.81081784e-01
-4.04818326e-01 -6.56130791e-01 -2.71046013e-01 9.82877493e-01
-6.78344846e-01 -2.60265797e-01 6.08364463e-01 -1.13563371e+00
8.95914078e-01 -1.78211212e+00 -3.99300307e-01 -1.28077865e-01
1.03071153e+00 8.57538998e-01 -1.17210746e-01 -7.03171611e-01
5.92467934e-02 1.11782528e-01 -1.29725635e-01 -5.62789500e-01
-5.61304629e-01 -1.96256056e-01 2.70062447e-01 3.50705087e-01
5.91027319e-01 1.29225588e+00 -6.40061438e-01 -3.17180604e-01
2.00465888e-01 5.52611411e-01 -4.98779833e-01 2.34954402e-01
-1.22814462e-01 5.30878901e-01 -1.48284048e-01 8.21070135e-01
5.99209428e-01 -3.08776021e-01 -4.22208756e-01 -2.83727348e-01
-1.07802115e-01 1.78733587e-01 -5.23124516e-01 1.25722694e+00
-1.14448667e-01 9.25499022e-01 -2.20618874e-01 -3.02247286e-01
5.09342611e-01 -3.20764557e-02 1.61252066e-01 -4.08263683e-01
4.79598701e-01 2.83425957e-01 4.51846898e-01 -5.95862508e-01
-1.58641130e-01 2.02491015e-01 1.02899551e+00 4.71288532e-01
-1.72974646e-01 2.93052316e-01 1.78259045e-01 -1.33824170e-01
1.23625672e+00 9.84652638e-02 4.58841085e-01 3.00878108e-01
3.99677575e-01 -1.01841532e-01 1.78632081e-01 6.30706549e-01
-7.79191613e-01 1.05893731e+00 6.92865908e-01 -6.96755886e-01
-1.33465493e+00 -1.16266716e+00 -1.00420341e-01 1.97874904e-01
-4.65536147e-01 8.93380940e-02 -9.95137036e-01 -6.78219318e-01
1.59131512e-01 9.84466299e-02 -1.07321620e+00 -1.85285863e-02
-3.96214336e-01 -9.55139339e-01 6.28737271e-01 7.45302558e-01
1.04236603e+00 -8.18764746e-01 -7.79920936e-01 4.08673547e-02
1.25094444e-01 -8.36380422e-01 -8.45487118e-02 -6.44075274e-01
-1.03068757e+00 -1.66581726e+00 -1.54648864e+00 -9.04170454e-01
8.20360541e-01 1.22141195e-02 9.71100152e-01 -1.12042025e-01
-1.11920822e+00 4.00441498e-01 1.14818878e-01 -8.73969913e-01
-4.66797113e-01 -2.90897101e-01 -2.69081384e-01 7.15493485e-02
6.96579158e-01 -3.40300977e-01 -1.60948193e+00 -1.24646001e-01
-1.80590078e-01 -1.84090495e-01 9.85927880e-01 5.77780068e-01
8.90798688e-01 -4.53353226e-01 3.58976722e-01 -7.19831228e-01
7.58360267e-01 -4.84172106e-02 -4.65901822e-01 1.73290625e-01
-4.87470597e-01 -2.27036387e-01 -2.17592254e-01 -1.32829100e-01
-9.37069178e-01 6.88475519e-02 1.96813475e-02 -1.66351914e-01
-5.05224705e-01 -1.52731821e-01 2.73793489e-01 -6.33486807e-01
1.17295134e+00 -2.33475447e-01 7.71343172e-01 -5.34099400e-01
3.19395602e-01 9.10217524e-01 7.68933952e-01 2.57325798e-01
2.33091444e-01 7.95977175e-01 3.68887454e-01 -6.69217229e-01
-7.79221714e-01 -3.43371004e-01 -4.13342834e-01 -4.50923890e-02
1.23157120e+00 -1.13892734e+00 -4.60587949e-01 1.13696027e+00
-1.21771955e+00 -5.36007702e-01 -4.05842841e-01 3.88266712e-01
-6.60634458e-01 4.90548849e-01 -5.49710333e-01 -6.22628033e-01
-9.54754293e-01 -1.10258496e+00 1.08780003e+00 6.48423076e-01
-2.04975188e-01 -7.84909368e-01 1.07916363e-01 4.94296521e-01
5.23707449e-01 3.75083029e-01 1.00868857e+00 5.84496409e-02
-5.87566674e-01 -1.00598946e-01 -8.50307286e-01 7.62186468e-01
2.63528198e-01 1.53053239e-01 -1.10494423e+00 -1.01790212e-01
-3.86403590e-01 -8.03877190e-02 1.28137970e+00 1.35273230e+00
9.55551267e-01 3.23881917e-02 -4.50787812e-01 8.19558918e-01
1.06198573e+00 1.90825939e-01 1.53145564e+00 6.45136774e-01
5.84475935e-01 4.40225691e-01 8.25722739e-02 1.75592989e-01
2.01731190e-01 1.78955778e-01 3.51757884e-01 -6.19860649e-01
-1.09466767e+00 3.30335170e-01 -8.41940492e-02 -5.58830440e-01
-6.31722450e-01 6.58531263e-02 -1.00852668e+00 6.96365356e-01
-1.47897744e+00 -5.67970514e-01 -5.25524259e-01 1.92003500e+00
5.52388072e-01 -1.95002913e-01 6.17914498e-01 -2.29684472e-01
8.16471756e-01 -7.21592546e-01 -9.45904255e-01 -1.08205535e-01
-2.15581000e-01 8.27307165e-01 4.52567756e-01 1.95117101e-01
-1.57138813e+00 7.33427763e-01 7.07060003e+00 -1.81327481e-02
-8.70550036e-01 1.18550025e-01 9.51973021e-01 -8.40069056e-02
3.95713270e-01 -2.38041282e-01 -6.50440514e-01 4.13018465e-01
5.48448622e-01 3.37537229e-01 1.15938641e-01 3.99490029e-01
4.32634681e-01 -4.09430504e-01 -7.92575121e-01 9.44254279e-01
-3.34367514e-01 -1.38364100e+00 2.00811058e-01 2.76670575e-01
7.69404948e-01 2.08629608e-01 4.11060065e-01 -4.92232412e-01
1.09317213e-01 -1.23073924e+00 -1.49234772e-01 1.14167798e+00
1.29653621e+00 -4.15512919e-01 8.23811233e-01 -6.95419133e-01
-6.07827246e-01 -1.95988625e-01 -2.98962981e-01 7.72419050e-02
-2.06795603e-01 3.77800703e-01 -8.24393094e-01 -1.77232802e-01
6.74814761e-01 8.73505354e-01 -1.16703999e+00 2.23455095e+00
-5.17979920e-01 5.01219869e-01 -6.62675723e-02 3.90867531e-01
-6.26831129e-02 -1.16387874e-01 6.40526831e-01 6.62816584e-01
4.68110979e-01 9.68463067e-03 -5.86621881e-01 1.24830997e+00
8.58670846e-02 -1.61555782e-01 -2.57928252e-01 4.63305712e-02
3.01040918e-01 8.04083169e-01 -3.51574808e-01 -1.70174688e-01
-6.79470003e-01 8.66664708e-01 1.80876721e-02 6.50036395e-01
-9.71950218e-02 -4.81565535e-01 9.06022906e-01 4.87932652e-01
2.84718424e-01 1.51884988e-01 -7.08209038e-01 -8.57986987e-01
1.17095597e-01 -7.57442296e-01 4.60633606e-01 -9.99001563e-01
-1.38427758e+00 3.12013090e-01 -8.33029091e-01 -1.30492163e+00
-4.46613133e-02 -1.11719275e+00 -1.07497108e+00 1.40280366e+00
-1.80882490e+00 -1.38212490e+00 -6.88937366e-01 7.29885399e-01
1.02667920e-01 -8.76660824e-01 7.26756275e-01 2.48343106e-02
-7.78446257e-01 5.63205481e-01 -1.23925515e-01 3.94921392e-01
9.86326039e-01 -1.60013103e+00 9.49173450e-01 1.16538918e+00
-7.84991145e-01 5.38014829e-01 1.70386642e-01 -8.39149535e-01
-5.38425744e-01 -1.54934096e+00 4.60802257e-01 -3.40889573e-01
4.53957558e-01 5.84207177e-01 -5.40587783e-01 6.03552282e-01
4.98829335e-02 1.12674423e-01 5.75163960e-01 -3.63697827e-01
-5.88897057e-02 4.79227491e-02 -1.35821319e+00 4.93793428e-01
7.97328532e-01 -3.72298181e-01 -4.01699066e-01 4.19393271e-01
2.04239845e-01 -3.69941175e-01 -5.26092470e-01 2.65580714e-01
5.88501871e-01 -1.12491620e+00 1.14400148e+00 -8.48497748e-01
5.80229938e-01 -3.12012464e-01 5.74170947e-01 -9.31386828e-01
-2.05947578e-01 -7.37364709e-01 -5.24901986e-01 6.17105544e-01
9.39740986e-02 -1.03828394e+00 1.11469913e+00 6.04205132e-01
-2.59680808e-01 -5.87266922e-01 -8.60657096e-01 -3.36253405e-01
2.61842728e-01 2.46727020e-01 1.84413791e-01 5.38444743e-02
-8.87046158e-01 -6.17140457e-02 6.20548129e-02 1.20228477e-01
9.71739173e-01 -3.61434281e-01 6.87289655e-01 -1.39648497e+00
-1.66745409e-01 -6.48438454e-01 -1.07904506e+00 -5.33074498e-01
-3.16958100e-01 -4.82868165e-01 -5.19892633e-01 -2.17198277e+00
-9.71544683e-02 2.25715805e-02 -1.39823407e-01 5.18308759e-01
-3.65755409e-01 6.82201862e-01 -3.92646968e-01 1.84880614e-01
7.73618743e-02 -5.18956557e-02 1.50276601e+00 -2.17196196e-01
-4.24775928e-01 5.90545416e-01 -9.24771667e-01 7.62681961e-01
1.06189537e+00 9.15389508e-02 -1.93939477e-01 -4.42627251e-01
-4.65139896e-02 -4.91078824e-01 1.23966503e+00 -1.20997572e+00
6.08442798e-02 5.79520762e-01 8.21741164e-01 -2.97471702e-01
2.24112451e-01 1.14964463e-01 -5.29686570e-01 7.40406752e-01
4.16768603e-02 -5.53169906e-01 3.17514330e-01 4.54590261e-01
2.38091424e-02 1.78176522e-01 1.24135780e+00 -2.58047551e-01
-6.65195465e-01 4.44378644e-01 -6.84984088e-01 1.88090891e-01
9.62843657e-01 -4.16494220e-01 -1.00189960e+00 -2.07367122e-01
-1.25822425e+00 9.62913930e-02 3.34066987e-01 -6.44374192e-02
9.38274443e-01 -8.39344323e-01 -1.17545056e+00 2.36887306e-01
5.72154596e-02 -4.96022627e-02 1.42665684e-01 1.09966266e+00
-9.96623456e-01 2.62742549e-01 -7.87575245e-01 -3.31410497e-01
-1.43284929e+00 2.01104283e-02 1.25595069e+00 1.13332465e-01
-6.99964106e-01 1.09882665e+00 1.86295599e-01 6.87298417e-01
2.40114257e-01 -4.21034157e-01 -4.72056657e-01 -1.68409288e-01
9.41270113e-01 9.12747025e-01 1.27138510e-01 -6.22426420e-02
2.25781500e-02 5.42543173e-01 -4.54479367e-01 3.33330035e-01
9.63435769e-01 -1.55196995e-01 -4.84147578e-01 -2.34629929e-01
6.04379237e-01 -5.20538628e-01 -1.25365031e+00 -6.35129958e-02
-5.37278056e-01 -2.50035763e-01 4.66286600e-01 -1.30744028e+00
-1.03585017e+00 1.11687934e+00 1.66993582e+00 -1.37999222e-01
1.31716275e+00 -7.09014684e-02 8.83086562e-01 6.21823370e-02
1.47540048e-01 -6.43152893e-01 -1.48276985e-01 2.30960138e-02
7.04319894e-01 -1.29412365e+00 1.82430819e-02 -5.54428279e-01
-1.24152839e-01 1.39575899e+00 7.83846259e-01 -2.81712562e-01
3.96881431e-01 -1.04300216e-01 4.59213853e-01 -9.55384672e-02
-9.06422809e-02 -9.74737227e-01 7.61268795e-01 1.26571083e+00
3.26796234e-01 8.81109238e-02 -3.78229409e-01 3.17351371e-01
9.65691358e-02 3.97807032e-01 7.14168906e-01 8.51993740e-01
-6.73694491e-01 -1.23380923e+00 -1.28143221e-01 1.21678758e+00
-5.81768274e-01 -3.43190581e-01 -8.44377697e-01 5.23532987e-01
4.73669410e-01 6.67431533e-01 2.16748938e-01 2.81374633e-01
1.49347046e-02 2.93816738e-02 7.85626113e-01 -8.55630457e-01
-2.32280135e-01 8.77881609e-03 2.95084447e-01 -6.33379102e-01
-5.82099736e-01 -6.39739335e-01 -7.41932809e-01 2.29110748e-01
3.41253579e-01 -8.67663503e-01 2.44400769e-01 8.38824809e-01
8.19213986e-01 5.51091552e-01 -3.61739919e-02 -4.46602106e-01
-1.91227626e-02 -1.10430038e+00 -6.56333029e-01 -1.43801436e-01
6.92198515e-01 -5.81939340e-01 -1.80303708e-01 3.32287371e-01] | [15.8577241897583, -4.0324296951293945] |
52c36a91-1780-41e2-b9a0-2c285fc64303 | marf-the-medial-atom-ray-field-object | 2307.00037 | null | https://arxiv.org/abs/2307.00037v1 | https://arxiv.org/pdf/2307.00037v1.pdf | MARF: The Medial Atom Ray Field Object Representation | We propose Medial Atom Ray Fields (MARFs), a novel neural object representation that enables accurate differentiable surface rendering with a single network evaluation per camera ray. Existing neural ray fields struggle with multi-view consistency and representing surface discontinuities. MARFs address both using a medial shape representation, a dual representation of solid geometry that yields cheap geometrically grounded surface normals, in turn enabling computing analytical curvature despite the network having no second derivative. MARFs map a camera ray to multiple medial intersection candidates, subject to ray-sphere intersection testing. We illustrate how the learned medial shape quantities applies to sub-surface scattering, part segmentation, and aid representing a space of articulated shapes. Able to learn a space of shape priors, MARFs may prove useful for tasks like shape retrieval and shape completion, among others. Code and data can be found at https://github.com/pbsds/MARF. | ['Theoharis Theoharis', 'Peder Bergebakken Sundt'] | 2023-06-30 | null | null | null | null | ['retrieval'] | ['methodology'] | [ 2.82847613e-01 4.53028917e-01 3.54599625e-01 -3.98113728e-01
-1.13561261e+00 -6.52609646e-01 5.84654868e-01 -1.41884722e-02
-1.43800765e-01 3.09507579e-01 -1.86592769e-02 -1.92183867e-01
-1.10767774e-01 -1.04584146e+00 -1.03067815e+00 -3.26086909e-01
-2.15022024e-02 8.84537637e-01 2.61714578e-01 -2.61057734e-01
1.54947847e-01 1.12797403e+00 -1.24779499e+00 3.47821712e-01
3.89431149e-01 1.04120696e+00 -9.06458125e-02 9.19620097e-01
-1.60286963e-01 3.71295735e-02 -1.46043092e-01 -7.18570709e-01
5.25688350e-01 1.67859495e-01 -6.65467501e-01 -4.88881692e-02
1.03187668e+00 -5.89124739e-01 1.58288684e-02 7.97254860e-01
3.55901212e-01 2.69840807e-01 9.35827374e-01 -1.03230786e+00
-6.70917630e-01 -1.37096629e-01 -5.28979123e-01 -3.80310386e-01
5.17607450e-01 4.05205786e-02 1.16255248e+00 -1.45535898e+00
8.75782788e-01 1.47468877e+00 1.13597012e+00 8.10171843e-01
-1.23375392e+00 -1.04909405e-01 -7.07428604e-02 -5.59961259e-01
-1.07723844e+00 -2.38115743e-01 8.38829637e-01 -5.35624325e-01
8.51440907e-01 5.84477723e-01 8.71009231e-01 7.02506065e-01
1.60117745e-01 7.38627136e-01 6.92986012e-01 -5.28358519e-02
2.52323449e-01 -3.73013675e-01 -3.34794004e-03 1.00697446e+00
6.94006402e-03 -9.75180417e-02 -4.38529342e-01 -5.57777226e-01
1.57639039e+00 9.68501717e-02 -4.28664297e-01 -5.57987213e-01
-1.03188801e+00 6.67460084e-01 4.94832188e-01 -3.65934700e-01
-3.70546609e-01 6.72125459e-01 -5.53414784e-03 -2.11432558e-02
6.38385773e-01 3.75474066e-01 -4.73356307e-01 2.06136387e-02
-6.23646379e-01 6.07542336e-01 7.72876740e-01 1.02768171e+00
8.36930811e-01 1.38770580e-01 2.27907091e-01 6.53039515e-01
6.29009366e-01 6.63885593e-01 -4.79235709e-01 -1.62583721e+00
2.60094196e-01 5.19944668e-01 3.66833732e-02 -8.31552804e-01
-5.34548998e-01 1.70978140e-02 -5.78511834e-01 7.81325400e-01
6.18972600e-01 1.19205818e-01 -1.00760174e+00 1.28768718e+00
6.25554323e-01 2.01551408e-01 -1.79214895e-01 9.90698457e-01
9.58697557e-01 6.36768401e-01 -5.02913773e-01 4.99666721e-01
1.42941368e+00 -4.09720242e-01 -7.75106158e-03 1.01072915e-01
1.31008178e-01 -8.26551735e-01 1.29554832e+00 6.09266639e-01
-1.83076215e+00 2.21177563e-02 -8.24723959e-01 -5.41308522e-01
-1.27870515e-01 -1.13594785e-01 7.09439576e-01 2.51183987e-01
-1.34707808e+00 8.21837187e-01 -1.09105563e+00 1.25350028e-01
8.32913756e-01 5.47827721e-01 -2.45397925e-01 7.02046603e-02
-5.46133935e-01 3.41646314e-01 -4.28482860e-01 1.00905493e-01
-8.41585934e-01 -1.09844553e+00 -8.24739039e-01 -3.48048024e-02
4.94412817e-02 -1.15567231e+00 1.41908836e+00 -6.73907578e-01
-1.57006931e+00 1.07992125e+00 -3.47314402e-02 -5.12133278e-02
7.75982141e-01 -4.24074113e-01 3.30171645e-01 3.70869070e-01
-1.75284252e-01 8.41677368e-01 9.03522074e-01 -1.62653077e+00
-2.96333786e-02 -5.56697667e-01 3.00331026e-01 3.34096819e-01
2.65335828e-01 -2.36182168e-01 -4.41910714e-01 -6.53869689e-01
6.43227339e-01 -6.91199720e-01 -3.22879612e-01 8.86564851e-01
-5.29912591e-01 -1.82918951e-01 6.23624444e-01 -8.01598310e-01
2.18355432e-01 -1.98374844e+00 2.91571081e-01 6.40336931e-01
3.14988315e-01 -2.97225863e-01 -1.39634892e-01 2.72667408e-01
7.91222900e-02 1.89233497e-01 -7.16689646e-01 -6.13584340e-01
9.67530236e-02 -1.67624075e-02 -2.45622769e-01 6.67800725e-01
3.04454207e-01 9.28468406e-01 -5.76683521e-01 -4.70966920e-02
-3.24139632e-02 1.08171105e+00 -8.68840933e-01 1.33830115e-01
-6.52571857e-01 4.60984349e-01 -3.56429756e-01 7.23273933e-01
8.70581627e-01 -3.34994823e-01 -3.86221290e-01 -3.50018352e-01
-1.06775705e-02 1.08884178e-01 -1.26467526e+00 1.88137913e+00
-3.12037766e-01 4.88463938e-01 6.45011961e-01 -1.65404752e-01
7.69193232e-01 2.88861066e-01 7.67174244e-01 -3.49842101e-01
9.53250006e-02 1.34075150e-01 -5.91089606e-01 -6.76045194e-02
4.29255843e-01 -1.46588504e-01 4.59901333e-01 5.61598480e-01
-3.15578938e-01 -7.25209713e-01 -4.59346831e-01 2.89547354e-01
9.26531196e-01 5.79802155e-01 -2.50927418e-01 -2.03654826e-01
1.83348864e-01 -9.81470346e-02 3.02127868e-01 1.89761475e-01
4.69407558e-01 1.33865476e+00 2.84975499e-01 -6.15586340e-01
-1.36778569e+00 -1.76588130e+00 -4.34434742e-01 7.96704412e-01
1.43513158e-01 -2.60483503e-01 -9.17132199e-01 -1.77301064e-01
2.41488382e-01 4.39608037e-01 -4.16197658e-01 3.50776970e-01
-9.11176145e-01 -2.90270090e-01 5.36716044e-01 6.01123989e-01
6.84216321e-02 -9.02548194e-01 -7.85655856e-01 -1.36014298e-01
2.57973164e-01 -6.71329200e-01 -5.50995409e-01 3.13993394e-02
-1.16617644e+00 -1.23661828e+00 -1.10507262e+00 -4.28555250e-01
9.59351838e-01 6.13903888e-02 1.37157357e+00 4.81305450e-01
-4.30607736e-01 1.07983494e+00 1.65661260e-01 -2.91759044e-01
-4.59697783e-01 -3.21203798e-01 -1.50860325e-01 -1.66103974e-01
-2.39535764e-01 -8.37209761e-01 -8.19496334e-01 1.06235288e-01
-1.03902769e+00 2.50931174e-01 -7.29648909e-03 4.71912086e-01
9.79962111e-01 -8.54495287e-01 -3.90707664e-02 -7.97751367e-01
4.61806327e-01 -2.90442258e-01 -8.37397993e-01 3.71793658e-02
-1.49891406e-01 -7.28724971e-02 2.64410019e-01 2.09996328e-02
-9.96317208e-01 -7.98664168e-02 -5.19340038e-01 -5.54803669e-01
-3.10053259e-01 5.26365116e-02 2.07149029e-01 -1.50591329e-01
7.28638887e-01 -1.56572580e-01 2.15477452e-01 -7.01219261e-01
5.25050581e-01 -9.62981135e-02 4.39209998e-01 -9.00243223e-01
8.55394900e-01 1.06167769e+00 3.60270262e-01 -9.81175005e-01
-5.28360009e-01 -4.13996458e-01 -7.99780369e-01 -3.94438773e-01
1.06347001e+00 -6.31153107e-01 -9.47184980e-01 3.58184218e-01
-1.62949789e+00 -6.86466515e-01 -5.47387481e-01 1.73163995e-01
-1.00729191e+00 3.53828013e-01 -8.64544749e-01 -7.67127872e-01
-3.37085634e-01 -1.01108885e+00 1.63810110e+00 -1.69605352e-02
-1.70170411e-01 -1.08060014e+00 -1.48947462e-02 3.21823210e-01
8.24052617e-02 4.53161478e-01 9.52403426e-01 -7.28787854e-02
-1.11883998e+00 1.97381690e-01 -1.16720438e-01 2.71858200e-02
-1.64717421e-01 2.44047135e-01 -1.15589452e+00 -3.63832474e-01
-1.34792417e-01 -1.96262971e-01 6.13330543e-01 6.40854061e-01
1.17039514e+00 -2.53697217e-01 -4.15234789e-02 1.35088432e+00
1.37385881e+00 -9.22464803e-02 5.74327290e-01 1.73829466e-01
1.06280208e+00 6.15200996e-01 2.67662723e-02 5.13243318e-01
3.11808646e-01 5.57530701e-01 8.33364189e-01 -2.42807984e-01
-2.23196521e-01 -2.94455811e-02 2.00878829e-01 6.77230060e-01
-5.57489872e-01 1.77735556e-02 -1.11186421e+00 1.36656836e-01
-1.36671567e+00 -5.75779259e-01 -5.81852794e-01 2.10297871e+00
5.47964990e-01 5.20356605e-03 1.81426644e-01 -2.32556969e-01
3.28687519e-01 -5.47601096e-02 -6.74310029e-01 -3.92059922e-01
-2.91891042e-02 3.69259685e-01 4.72448140e-01 9.74447310e-01
-6.59732103e-01 6.73920095e-01 5.97350073e+00 4.79853362e-01
-8.56534362e-01 1.87084392e-01 4.79947537e-01 -5.28215095e-02
-1.11524439e+00 -1.39595360e-01 -5.65336049e-01 -2.27186531e-01
2.87370265e-01 3.57707083e-01 5.92048883e-01 7.18651414e-01
-1.16679564e-01 2.00565115e-01 -1.07796168e+00 9.69084203e-01
3.56813744e-02 -1.65622330e+00 2.10572913e-01 1.05189532e-02
7.18421996e-01 3.57617229e-01 1.00490592e-01 -3.60294402e-01
2.95231134e-01 -1.00032008e+00 9.24739242e-01 7.84908414e-01
1.10131550e+00 -5.37214696e-01 -9.86810923e-02 1.79396495e-01
-1.15388632e+00 4.29839581e-01 -2.14826673e-01 2.94497997e-01
3.47407222e-01 3.14169616e-01 -8.33939433e-01 5.09387970e-01
6.82758391e-01 3.39595228e-01 -3.04964900e-01 1.06549549e+00
2.50382036e-01 2.48121455e-01 -7.56530941e-01 3.82240415e-01
1.86563686e-01 -6.66051924e-01 9.47117269e-01 8.96773458e-01
3.31954241e-01 2.19502062e-01 8.26324895e-02 1.38281655e+00
-4.16433476e-02 9.88999456e-02 -6.92909598e-01 3.96445632e-01
2.35125721e-01 1.06801188e+00 -1.06771553e+00 2.03011860e-03
-2.73110449e-01 9.69599485e-01 3.55105281e-01 6.20783031e-01
-5.69927216e-01 2.68151127e-02 6.38700306e-01 5.53818226e-01
-1.11634238e-02 -5.38552642e-01 -7.02738464e-01 -1.02239120e+00
1.52729705e-01 -4.13308978e-01 1.16557375e-01 -1.18121994e+00
-1.38752747e+00 4.83106315e-01 1.27869278e-01 -1.11911941e+00
6.74213171e-02 -1.04616547e+00 -6.50209963e-01 7.52970397e-01
-1.31429636e+00 -1.42589819e+00 -3.00218731e-01 5.84665120e-01
4.76976156e-01 9.17555988e-02 6.84638381e-01 1.28394261e-01
8.50487202e-02 2.98652023e-01 -8.79267827e-02 1.64436903e-02
1.92557573e-01 -1.61039424e+00 9.47216928e-01 4.26194400e-01
8.79209340e-02 6.98585331e-01 3.67807984e-01 -9.03506100e-01
-1.62352717e+00 -8.68077457e-01 -3.33554149e-02 -7.10941732e-01
1.08687960e-01 -3.51473212e-01 -1.10600686e+00 9.00471330e-01
-1.35759264e-01 -7.29850307e-02 3.49026173e-01 -1.30383804e-01
-3.97191942e-01 3.70688677e-01 -1.11809862e+00 8.81179214e-01
1.24981618e+00 -5.63119709e-01 -1.90535039e-01 4.47062939e-01
5.17718196e-01 -1.01620972e+00 -1.01272821e+00 3.66139740e-01
4.73364741e-01 -1.12876320e+00 1.44964421e+00 -5.99621236e-01
3.40254694e-01 -2.19418094e-01 -3.25726211e-01 -1.15899694e+00
-1.08324021e-01 -8.39239061e-01 -2.11766392e-01 6.44516170e-01
2.47923806e-01 -5.17940819e-01 1.04595780e+00 7.38371968e-01
-6.94244623e-01 -1.34847319e+00 -1.00938129e+00 -5.25889516e-01
3.23052883e-01 -7.63878345e-01 6.79927528e-01 8.08254361e-01
-7.88919926e-01 -1.70896202e-01 2.43674725e-01 3.12748045e-01
8.04013789e-01 1.82061698e-02 7.19756246e-01 -1.61036885e+00
-2.93836176e-01 -7.19617307e-01 -7.57174715e-02 -1.30107570e+00
-7.80068710e-02 -1.20001781e+00 4.37748171e-02 -1.84629214e+00
-2.82760441e-01 -6.95192695e-01 5.42333066e-01 2.74114937e-01
1.87615633e-01 4.36176568e-01 1.12535216e-01 8.08485970e-02
-8.39417428e-02 5.82516849e-01 1.69911337e+00 9.94220972e-02
-1.40132040e-01 5.78330606e-02 -2.35652566e-01 1.29961038e+00
7.11040914e-01 -2.68758357e-01 -3.10825974e-01 -7.87835300e-01
8.34177315e-01 2.00034067e-01 8.86074722e-01 -7.21782327e-01
1.60402060e-01 -1.71467125e-01 4.41842616e-01 -7.69599319e-01
1.12238002e+00 -9.40660298e-01 3.88925344e-01 6.60680905e-02
-9.41433534e-02 3.31649184e-01 2.08225518e-01 5.29218554e-01
3.62391323e-01 -3.16457212e-01 7.47107446e-01 -5.24881959e-01
-1.28686816e-01 7.23888099e-01 3.66716906e-02 2.28072450e-01
7.30663538e-01 -5.86139441e-01 -1.69956744e-01 -3.62329870e-01
-9.19343293e-01 -6.73821941e-02 9.01519537e-01 1.45401852e-02
1.24527967e+00 -1.48428762e+00 -6.67753696e-01 3.54239136e-01
-3.32983345e-01 8.36220145e-01 1.58796638e-01 5.54039776e-01
-1.23064756e+00 -2.25620285e-01 -7.77237713e-02 -9.37404871e-01
-1.03866446e+00 -2.12156057e-01 7.98085451e-01 3.73367220e-01
-1.13568437e+00 1.26710117e+00 4.30726022e-01 -6.88219726e-01
2.06949458e-01 -4.67940032e-01 3.15826386e-01 -2.67286628e-01
2.09674165e-01 6.25295103e-01 2.00642109e-01 -5.74902773e-01
-8.34456682e-02 1.09578168e+00 3.25365424e-01 -3.98294091e-01
1.49298024e+00 2.91549981e-01 -2.86186695e-01 4.21467185e-01
1.05715287e+00 1.95578367e-01 -1.73819041e+00 1.53877005e-01
-2.10396186e-01 -4.36866015e-01 -3.49964760e-02 -4.73381251e-01
-9.92862403e-01 1.07844508e+00 3.46547186e-01 7.20342696e-02
5.87295353e-01 1.84640601e-01 1.00014269e+00 5.06518364e-01
2.24995390e-01 -7.77652919e-01 2.60628134e-01 6.50285780e-01
1.40846276e+00 -8.41361880e-01 8.28854069e-02 -6.99315965e-01
-2.19448701e-01 1.42106354e+00 6.37520432e-01 -4.96700019e-01
8.84725928e-01 5.94846904e-01 1.01419941e-01 -8.93228829e-01
-3.30711126e-01 2.82143772e-01 6.32715166e-01 5.90659559e-01
4.10738617e-01 -5.48966713e-02 4.55669552e-01 9.39474031e-02
-2.94436276e-01 -5.39152205e-01 4.31398809e-01 7.87316978e-01
-2.57102340e-01 -8.42019856e-01 -6.25822246e-01 5.75785816e-01
-3.67018998e-01 -6.14324696e-02 -2.94951677e-01 5.25749683e-01
-1.98335215e-01 -1.15109971e-02 5.05079925e-01 2.52272487e-01
5.65535903e-01 -3.49655971e-02 6.93731725e-01 -7.31853068e-01
-5.52341759e-01 2.37908706e-01 -2.03445256e-02 -8.39151323e-01
-2.16366023e-01 -7.44132221e-01 -1.87463796e+00 -2.12621510e-01
-6.94608241e-02 -2.17444539e-01 7.65094101e-01 5.01081824e-01
3.05243284e-01 3.58946919e-01 1.91146687e-01 -1.51266515e+00
-4.78339761e-01 -3.55842471e-01 -3.36125135e-01 6.11499250e-01
5.73047936e-01 -6.07285917e-01 -1.61865205e-01 1.74772963e-01] | [8.72671890258789, -3.609804391860962] |
567e9922-1717-4aac-86d7-9cc38176558f | building-scalable-video-understanding | 2301.06866 | null | https://arxiv.org/abs/2301.06866v3 | https://arxiv.org/pdf/2301.06866v3.pdf | Building Scalable Video Understanding Benchmarks through Sports | Existing benchmarks for evaluating long video understanding falls short on two critical aspects, either lacking in scale or quality of annotations. These limitations arise from the difficulty in collecting dense annotations for long videos, which often require manually labeling each frame. In this work, we introduce an automated Annotation and Video Stream Alignment Pipeline (abbreviated ASAP). We demonstrate the generality of ASAP by aligning unlabeled videos of four different sports with corresponding freely available dense web annotations (i.e. commentary). We then leverage ASAP scalability to create LCric, a large-scale long video understanding benchmark, with over 1000 hours of densely annotated long Cricket videos (with an average sample length of ~50 mins) collected at virtually zero annotation cost. We benchmark and analyze state-of-the-art video understanding models on LCric through a large set of compositional multi-choice and regression queries. We establish a human baseline that indicates significant room for new research to explore. Our human studies indicate that ASAP can align videos and annotations with high fidelity, precision, and speed. The dataset along with the code for ASAP and baselines can be accessed here: https://asap-benchmark.github.io/. | ['Yash Kant', 'Vishvak Murahari', 'Igor Gilitschenski', 'Karthik Narasimhan', 'Alex Zhang', 'Aniket Agarwal'] | 2023-01-17 | null | null | null | null | ['video-understanding'] | ['computer-vision'] | [ 1.44582719e-01 -3.37128848e-01 -5.03675222e-01 -3.35173696e-01
-1.28153312e+00 -8.78243506e-01 3.19995373e-01 -8.66669342e-02
-4.48492467e-01 3.52590263e-01 6.32021308e-01 2.07307190e-01
3.53773355e-01 -1.30697101e-01 -1.01384914e+00 -2.12450415e-01
-1.99579626e-01 4.82626587e-01 3.33353192e-01 1.43118147e-02
1.91622600e-01 -1.13177307e-01 -1.64940870e+00 6.86501801e-01
4.80807394e-01 9.00515854e-01 1.47036657e-01 1.06501615e+00
3.52248490e-01 1.04880798e+00 -3.52949500e-01 -6.13755882e-01
4.81801689e-01 -2.76414990e-01 -1.07380831e+00 -5.84746152e-03
9.53591704e-01 -7.82521188e-01 -6.28383338e-01 6.53766453e-01
5.04795849e-01 1.81847140e-01 2.19542325e-01 -1.65050435e+00
-3.84411365e-01 5.83439887e-01 -6.18344247e-01 4.81027454e-01
8.63451898e-01 4.41257596e-01 1.36380291e+00 -8.46253693e-01
9.38438475e-01 9.34998035e-01 8.89302969e-01 5.42935669e-01
-9.26865399e-01 -9.21715081e-01 8.09811652e-02 5.40674686e-01
-1.50884521e+00 -8.47440541e-01 2.25245520e-01 -6.63268268e-01
9.03522670e-01 1.26972824e-01 7.81067729e-01 1.52586961e+00
-3.83612543e-01 1.10628653e+00 5.77224970e-01 -1.15725547e-01
-1.65361881e-01 -5.22720039e-01 3.35798264e-01 6.53269827e-01
-2.73704994e-02 1.84114799e-02 -1.12925327e+00 1.82384983e-01
5.87228775e-01 -2.65307993e-01 -3.49749595e-01 -2.33551517e-01
-1.61130130e+00 5.51951706e-01 1.98038444e-02 -1.07640572e-01
-1.36514843e-01 4.24606502e-01 8.84576797e-01 3.28564346e-01
2.15867981e-01 3.95461470e-01 -4.63441730e-01 -1.05211771e+00
-9.25486267e-01 4.85394388e-01 9.77025390e-01 1.28983247e+00
4.01263446e-01 -3.16169411e-01 -1.08217448e-01 7.29420364e-01
-2.62336731e-02 4.01048362e-01 2.55348802e-01 -1.69172168e+00
6.84005916e-01 9.07222927e-03 6.32882416e-02 -1.00571001e+00
-1.45112097e-01 -1.46270543e-02 -2.46345803e-01 -3.16127121e-01
6.30397081e-01 8.52042157e-03 -5.45261979e-01 1.75166738e+00
1.71235114e-01 7.54991233e-01 -1.57318950e-01 1.04666960e+00
1.03325510e+00 6.55298352e-01 6.13535382e-02 3.81513722e-02
1.45197332e+00 -1.60855806e+00 -4.84203130e-01 -3.37419719e-01
7.12799191e-01 -8.05571556e-01 1.45356882e+00 4.55081075e-01
-1.08075023e+00 -5.72460234e-01 -9.55272496e-01 -3.71126354e-01
6.09157793e-03 -1.48396939e-01 4.65229332e-01 2.56578028e-01
-8.91992331e-01 4.18692291e-01 -1.11203802e+00 -5.92767000e-01
5.16566634e-01 5.64204752e-02 -6.34484351e-01 -3.91995728e-01
-9.35892224e-01 4.44911420e-01 3.04268032e-01 -1.23540118e-01
-1.24161482e+00 -1.00769615e+00 -1.05115378e+00 -1.72670171e-01
6.71040237e-01 -3.77688199e-01 1.74662006e+00 -1.17712235e+00
-1.05875647e+00 9.28728461e-01 -1.39036015e-01 -4.89288330e-01
5.16092539e-01 -9.62732852e-01 -2.77914017e-01 5.76068461e-01
3.90991956e-01 1.09973371e+00 5.18503129e-01 -8.61849248e-01
-6.88155234e-01 -5.11344932e-02 4.49809968e-01 3.29166144e-01
-2.95798272e-01 3.09625775e-01 -1.20926023e+00 -7.77044296e-01
-4.42231625e-01 -1.15511644e+00 1.95015535e-01 -5.72085157e-02
-2.26457447e-01 1.64062530e-01 8.45980585e-01 -9.86545563e-01
1.34319997e+00 -2.19453669e+00 3.90986025e-01 -2.47852683e-01
3.04829180e-01 -6.83604628e-02 -4.61731017e-01 4.24996912e-01
1.41765373e-02 2.19582900e-01 -7.57163912e-02 -3.72065067e-01
-1.03218574e-02 1.77353978e-01 -2.54536182e-01 3.19872111e-01
-7.53014386e-02 9.98546183e-01 -1.08571708e+00 -6.44787133e-01
-1.80231333e-02 2.76299298e-01 -9.99197245e-01 4.17073131e-01
-3.25283080e-01 5.27007163e-01 -1.04879953e-01 9.84215677e-01
1.87021002e-01 -5.32212734e-01 4.85731363e-02 -6.39609337e-01
4.47752774e-02 8.58299211e-02 -9.47157562e-01 2.37985206e+00
-1.38139948e-01 1.16126299e+00 -9.30379853e-02 -7.50491083e-01
1.78391665e-01 3.91967654e-01 7.74769127e-01 -5.41055620e-01
-6.63401186e-02 9.39122215e-02 -2.56738961e-01 -8.40911746e-01
7.27601290e-01 3.98357540e-01 -2.01002613e-01 4.18617219e-01
3.48133624e-01 1.06570996e-01 6.55957580e-01 5.87779939e-01
1.34378886e+00 5.63862264e-01 1.57938212e-01 -1.70559101e-02
9.90253314e-02 2.91754484e-01 6.50446892e-01 6.49890184e-01
-5.18325090e-01 9.65522587e-01 5.23147285e-01 -4.42108512e-01
-1.29062784e+00 -8.68760407e-01 2.33307213e-01 1.46199298e+00
2.90234029e-01 -1.11452377e+00 -8.92974138e-01 -5.17268717e-01
-2.21649960e-01 2.04413071e-01 -5.55167139e-01 2.79014975e-01
-7.54748702e-01 -3.68416429e-01 9.57133949e-01 7.69866824e-01
5.48066974e-01 -8.31197917e-01 -5.70113838e-01 -4.46480587e-02
-1.00871527e+00 -1.79287040e+00 -9.61405396e-01 -4.65639591e-01
-4.34570312e-01 -1.53221834e+00 -4.16563123e-01 -6.05101049e-01
2.34188944e-01 4.58911747e-01 1.67602146e+00 2.81138390e-01
-3.22054058e-01 8.56899261e-01 -7.78854311e-01 4.55638301e-03
-1.15504466e-01 3.05616707e-01 1.46205453e-02 -2.69923091e-01
4.67779249e-01 -4.07593101e-01 -7.60125339e-01 7.32982576e-01
-7.26393521e-01 5.11664987e-01 3.01820010e-01 6.12327337e-01
7.07600534e-01 -6.46752894e-01 4.94019777e-01 -7.38923371e-01
-2.18380354e-02 -6.73768818e-01 -4.36828136e-01 1.60230830e-01
-9.78686586e-02 -3.13273907e-01 2.32303366e-01 -2.35152021e-01
-6.98502779e-01 7.80522078e-02 -1.68631703e-01 -7.47427225e-01
-1.60222784e-01 3.61344486e-01 -8.98697414e-03 3.08222678e-02
5.72951853e-01 -6.64849803e-02 -9.27458480e-02 -3.71867448e-01
4.15033817e-01 4.56525385e-01 1.10929799e+00 -9.48681176e-01
6.00423813e-01 3.40714723e-01 -5.67418337e-01 -6.83921218e-01
-1.08020759e+00 -7.30546951e-01 -7.38972008e-01 -5.11893034e-01
1.10072494e+00 -1.41274071e+00 -7.32391477e-01 4.76475000e-01
-8.62594128e-01 -7.34897375e-01 -9.14073288e-02 4.96707976e-01
-9.02199030e-01 5.14808476e-01 -9.24309850e-01 -9.61223915e-02
-2.63938695e-01 -1.20440400e+00 1.23241603e+00 6.32091090e-02
-5.68069041e-01 -6.44932151e-01 1.42124668e-01 1.06450045e+00
8.40978473e-02 3.03245455e-01 1.06514014e-01 -6.43041015e-01
-8.54213357e-01 -9.45597738e-02 -3.37411970e-01 2.24346027e-01
-3.85197401e-01 1.55159280e-01 -8.52876127e-01 -5.21523297e-01
-5.64389169e-01 -1.04564500e+00 7.25115180e-01 1.86154664e-01
1.35642099e+00 -1.75034732e-01 -9.06029195e-02 9.54813600e-01
1.24860036e+00 -2.78865725e-01 6.25647902e-01 4.63312387e-01
9.40607965e-01 3.34959090e-01 1.03352487e+00 4.96082574e-01
6.09721780e-01 8.81204069e-01 2.77253360e-01 3.29810768e-01
-2.99800515e-01 -4.60422367e-01 5.63001871e-01 1.11871445e+00
-4.40202862e-01 -3.86236787e-01 -1.08403158e+00 7.12284982e-01
-1.89259362e+00 -1.22229099e+00 1.55143186e-01 2.06476498e+00
8.56015384e-01 -2.00348664e-02 2.86209464e-01 -1.56993851e-01
5.81628799e-01 3.95900398e-01 -4.09772098e-01 2.04925537e-01
-8.00471455e-02 -8.77629444e-02 3.87140930e-01 5.57800055e-01
-1.24258125e+00 8.65645587e-01 6.18742609e+00 7.62972295e-01
-6.74055099e-01 2.58416563e-01 6.03480995e-01 -6.39113367e-01
3.59950773e-02 2.63285227e-02 -8.46812963e-01 4.99070674e-01
1.16150129e+00 -3.74398082e-02 6.08047843e-01 7.89233267e-01
1.69030100e-01 -7.19183162e-02 -1.43503129e+00 1.30524909e+00
3.57873440e-01 -1.41345382e+00 -5.28713949e-02 -1.57152325e-01
6.85545504e-01 5.01083076e-01 -3.55994672e-01 4.07999933e-01
1.35452032e-01 -9.46564198e-01 8.99176598e-01 2.81771213e-01
1.07559264e+00 -3.12025309e-01 5.91051221e-01 -1.45839676e-01
-1.55405807e+00 -7.87706487e-03 2.26866137e-02 -9.19186473e-02
4.90424335e-01 1.64064616e-02 -3.43240827e-01 3.93726617e-01
1.08999944e+00 1.35924721e+00 -7.49768138e-01 9.47300315e-01
2.27760221e-03 8.22929263e-01 -4.34157193e-01 5.34874797e-01
1.90099850e-01 -2.30421349e-02 4.41451162e-01 1.55343151e+00
3.08801830e-01 3.35737824e-01 5.31964302e-01 2.42848083e-01
-3.89376342e-01 1.17891096e-02 -4.38110948e-01 -2.72561312e-01
7.52917349e-01 1.12483919e+00 -6.61634803e-01 -5.13987422e-01
-8.43927324e-01 1.00630271e+00 1.51790023e-01 2.37021178e-01
-1.34199619e+00 -1.08942632e-02 8.27139497e-01 1.38419196e-01
1.95664763e-01 -2.55607933e-01 1.09896965e-01 -1.64132535e+00
5.89445122e-02 -1.44210303e+00 7.98866391e-01 -9.78198111e-01
-1.06981683e+00 3.81601781e-01 2.78521657e-01 -1.27881742e+00
-2.15984643e-01 -2.97728330e-01 -2.14364603e-01 8.77532363e-02
-1.15210116e+00 -1.24549615e+00 -1.06254220e+00 6.82325363e-01
1.18825984e+00 -1.33374631e-02 6.09347939e-01 8.27278495e-01
-6.99210703e-01 6.00451112e-01 -3.00136060e-01 4.26560223e-01
1.09468019e+00 -1.04864001e+00 5.49418092e-01 8.56482685e-01
3.67652148e-01 1.31277144e-01 7.46435225e-01 -5.54709733e-01
-1.46240747e+00 -9.83909130e-01 4.18893993e-01 -7.94909000e-01
9.13469434e-01 -3.93762231e-01 -7.54904389e-01 1.17020869e+00
4.30211276e-01 1.76728055e-01 9.54549670e-01 1.49269477e-01
-6.60251260e-01 -2.05766037e-02 -4.67904389e-01 5.28507710e-01
1.48170626e+00 -6.30054593e-01 -3.11207533e-01 5.14435053e-01
9.02320385e-01 -7.67293036e-01 -1.24520516e+00 4.37334329e-01
1.05505311e+00 -1.01448679e+00 1.12693357e+00 -5.70058823e-01
8.12289834e-01 -2.60102361e-01 -4.29986537e-01 -8.06494653e-01
-2.29598302e-02 -6.48312092e-01 -2.23780379e-01 1.20263767e+00
2.40833044e-01 4.08883840e-02 7.46461987e-01 6.63425088e-01
-2.56095976e-01 -6.16309702e-01 -4.86197174e-01 -6.89047456e-01
-4.31695342e-01 -7.07625806e-01 2.17307374e-01 1.01778126e+00
8.30729902e-02 2.97781557e-01 -6.39886916e-01 7.81758502e-02
5.71721315e-01 2.20371429e-02 1.17821229e+00 -7.61623740e-01
-5.67985058e-01 -1.69819653e-01 -5.11729777e-01 -1.26543260e+00
2.20224008e-01 -5.83133519e-01 4.81050946e-02 -1.08877468e+00
6.32597625e-01 -1.47713140e-01 -5.35562150e-02 6.19830787e-01
-1.43382803e-01 8.14381301e-01 4.29083824e-01 5.71717799e-01
-1.30886173e+00 2.62457550e-01 7.82636106e-01 1.62354931e-02
9.31613520e-02 -5.21159470e-01 -4.26959097e-01 8.09797168e-01
6.73149645e-01 -1.81589246e-01 -4.18482482e-01 -8.40582430e-01
1.93985060e-01 1.33668706e-01 4.28563237e-01 -1.34919119e+00
1.56705931e-01 2.50273291e-02 2.25124937e-02 -4.43797469e-01
4.39419717e-01 -4.69557703e-01 3.74056280e-01 1.02408275e-01
-5.23232818e-01 4.76163507e-01 2.66978860e-01 6.02010012e-01
-5.17937243e-01 7.90805090e-03 4.75034654e-01 -9.64422897e-02
-1.32598031e+00 5.18855453e-01 2.42051799e-02 6.72120452e-01
1.00456440e+00 -7.61780441e-02 -6.22383475e-01 -6.68218136e-01
-6.91789031e-01 5.50037861e-01 7.74422646e-01 7.80797303e-01
4.06488568e-01 -1.29465544e+00 -8.19908798e-01 -3.18036705e-01
4.16500598e-01 -9.00840759e-02 3.12979937e-01 9.51963365e-01
-9.75427151e-01 1.95744172e-01 -3.72765303e-01 -8.26896191e-01
-1.44251347e+00 4.21541154e-01 7.53658712e-02 -1.07324250e-01
-7.72394598e-01 8.06362569e-01 2.23253816e-01 -1.47827268e-01
3.39262456e-01 -1.31523162e-01 1.75866205e-02 1.36263832e-01
7.55162358e-01 4.46474075e-01 -2.11491480e-01 -7.89014757e-01
-2.42169753e-01 7.12451279e-01 9.83751938e-02 -8.54721069e-02
1.24195099e+00 -2.87038773e-01 1.98240474e-01 4.04171854e-01
1.23608398e+00 -2.02597156e-02 -1.59696698e+00 -2.40018293e-02
-1.45386085e-01 -7.92619467e-01 -4.10862684e-01 -5.97346008e-01
-1.20749629e+00 5.79360187e-01 4.12943989e-01 -3.18967223e-01
1.09286404e+00 1.08465724e-01 1.09278631e+00 4.00868446e-01
3.47097307e-01 -1.06369114e+00 3.87841642e-01 4.12489533e-01
8.27151000e-01 -1.47107160e+00 1.45544149e-02 -4.72926527e-01
-8.21216404e-01 8.92207861e-01 9.59227741e-01 3.20292227e-02
2.73782074e-01 4.49584633e-01 8.96121114e-02 -1.31232038e-01
-8.51241529e-01 -2.27501914e-02 7.99444318e-02 3.30340296e-01
5.62001467e-01 -2.42131293e-01 -3.29096876e-02 5.14476001e-01
-2.47369707e-01 2.36237466e-01 6.24140322e-01 8.13307524e-01
-1.24909200e-01 -8.67029190e-01 -1.70708895e-01 2.98128456e-01
-5.96856415e-01 -1.19561203e-01 -1.34091945e-02 7.84312487e-01
-1.19738616e-01 8.79043579e-01 1.27584696e-01 -5.42109907e-01
1.43002808e-01 -2.45837551e-02 4.55671072e-01 -4.03498203e-01
-2.61604100e-01 -1.18440554e-01 5.73269069e-01 -1.27094638e+00
-8.35092306e-01 -9.31937218e-01 -1.03827667e+00 -5.28340399e-01
-6.19510785e-02 9.69921872e-02 2.42338404e-01 7.80160904e-01
5.13609886e-01 2.37005994e-01 3.18586198e-03 -1.09487188e+00
2.21701637e-02 -9.18035328e-01 -4.34365496e-02 8.98414195e-01
1.85611218e-01 -5.63708901e-01 -1.76806733e-01 7.79628456e-01] | [10.209755897521973, 0.8023236393928528] |
d6c5d69f-f15c-4965-ace6-0bc75bbbdcaa | learning-to-extend-program-graphs-to-work-in | 2105.14038 | null | https://arxiv.org/abs/2105.14038v1 | https://arxiv.org/pdf/2105.14038v1.pdf | Learning to Extend Program Graphs to Work-in-Progress Code | Source code spends most of its time in a broken or incomplete state during software development. This presents a challenge to machine learning for code, since high-performing models typically rely on graph structured representations of programs derived from traditional program analyses. Such analyses may be undefined for broken or incomplete code. We extend the notion of program graphs to work-in-progress code by learning to predict edge relations between tokens, training on well-formed code before transferring to work-in-progress code. We consider the tasks of code completion and localizing and repairing variable misuse in a work-in-process scenario. We demonstrate that training relation-aware models with fine-tuned edges consistently leads to improved performance on both tasks. | ['Daniel Tarlow', 'Chris J. Maddison', 'Xuechen Li'] | 2021-05-28 | null | null | null | null | ['variable-misuse'] | ['computer-code'] | [ 2.16588125e-01 3.86242241e-01 -7.91649163e-01 -4.55861479e-01
-5.83093762e-01 -5.85063696e-01 2.81061888e-01 9.06113625e-01
3.44161242e-01 1.14060035e-02 5.63415512e-02 -8.97227764e-01
-1.78347036e-01 -8.83545041e-01 -8.63290489e-01 3.41613501e-01
-6.09849811e-01 6.29625618e-02 1.51534438e-01 2.67128218e-02
3.89624774e-01 2.08678052e-01 -1.18654954e+00 4.50118750e-01
8.04449379e-01 -3.11714765e-02 -4.31480967e-02 1.11866128e+00
-6.03130996e-01 1.20791674e+00 -5.21787226e-01 -5.85602939e-01
-3.60048795e-03 -2.35177606e-01 -1.09566987e+00 5.77567779e-02
2.94629961e-01 2.21414968e-01 -2.23221451e-01 1.17179394e+00
-3.69616061e-01 -1.76871255e-01 4.50973511e-01 -1.51476169e+00
-7.98260093e-01 1.20811582e+00 -9.44141030e-01 2.81203449e-01
5.51729500e-01 -1.22708653e-03 1.20489275e+00 -6.85511529e-01
7.90876269e-01 9.64669287e-01 1.04973781e+00 3.76233816e-01
-1.60738504e+00 -2.25124225e-01 1.89885780e-01 -1.28215579e-02
-1.08085608e+00 -3.82566124e-01 7.22607911e-01 -1.06806111e+00
1.84410608e+00 -1.33736935e-02 2.85409033e-01 3.97917449e-01
6.80866420e-01 5.20557940e-01 2.15405136e-01 -6.32938266e-01
-1.25933960e-01 -3.92157495e-01 8.81315649e-01 1.44818342e+00
6.14157081e-01 -2.81445265e-01 -3.98056388e-01 -6.15422189e-01
3.58871996e-01 2.91515142e-01 -1.74483024e-02 -6.23108268e-01
-1.08828366e+00 6.27858341e-01 2.16224402e-01 3.56920183e-01
-6.40471950e-02 8.89742494e-01 6.06609702e-01 7.53788888e-01
4.65033919e-01 7.71410465e-01 -7.14107454e-01 -6.24821007e-01
-9.65099692e-01 8.88410732e-02 1.07780898e+00 1.27057660e+00
1.28007841e+00 1.65661231e-01 -1.43456096e-02 4.15503979e-01
2.00554505e-01 -8.07662532e-02 1.22868001e-01 -5.45269489e-01
8.43277991e-01 1.43599308e+00 -1.98954314e-01 -7.00737417e-01
-3.20335805e-01 -4.09253985e-01 -1.51848912e-01 4.24286067e-01
2.02947855e-01 -8.39332119e-02 -8.39494467e-01 1.30494213e+00
-3.25291269e-02 -5.33352308e-02 -1.19504280e-01 -3.87644880e-02
3.97109210e-01 3.28726739e-01 4.23206622e-03 3.46910097e-02
9.44723010e-01 -1.05586612e+00 -5.41951060e-01 -5.31042039e-01
1.36103499e+00 -3.80314559e-01 7.95083225e-01 2.19419539e-01
-1.04482830e+00 -4.30987597e-01 -1.03160143e+00 -2.59673238e-01
-3.34475696e-01 -9.61728580e-03 9.31160629e-01 7.11631358e-01
-1.40494812e+00 8.25327933e-01 -1.23243105e+00 -3.94727774e-02
4.80135173e-01 4.26067472e-01 -6.80061162e-01 -2.57511526e-01
-2.53736943e-01 7.72563756e-01 2.32831150e-01 -1.58009857e-01
-1.09929490e+00 -8.99319589e-01 -1.46157217e+00 4.42814857e-01
4.47102398e-01 -5.92150152e-01 1.32205975e+00 -8.00726473e-01
-5.95069170e-01 8.59918416e-01 -2.81085193e-01 -3.10824752e-01
-7.22646788e-02 -1.71890184e-01 -2.76489824e-01 -6.06974185e-01
1.06417157e-01 -2.77099669e-01 7.71042526e-01 -1.06808162e+00
-4.89687204e-01 -4.40423876e-01 4.14163232e-01 -5.52515686e-01
-2.44728118e-01 3.91992390e-01 -3.65007818e-01 -2.68116802e-01
7.12244436e-02 -7.84202933e-01 -2.75479704e-01 -1.27133235e-01
-3.85888517e-01 -3.20141613e-01 6.24819458e-01 -9.55847263e-01
1.75524592e+00 -2.01511478e+00 4.63382393e-01 3.29331696e-01
6.63971603e-01 -1.21062852e-01 -2.30630204e-01 7.86337793e-01
-4.86769050e-01 5.71859598e-01 -3.92858744e-01 -2.44445011e-01
6.17635548e-02 1.13298431e-01 3.01138870e-02 6.19619071e-01
4.21844840e-01 1.09060967e+00 -1.18828046e+00 -4.71557558e-01
-2.64624089e-01 -2.38691065e-02 -7.18522668e-01 2.30993018e-01
-3.89825016e-01 -1.99903697e-01 -4.50312436e-01 9.18213546e-01
2.56376386e-01 -2.83419907e-01 4.72467303e-01 4.34499443e-01
-5.58706336e-02 2.54479408e-01 -6.93782032e-01 2.01254582e+00
-1.05645669e+00 7.96712101e-01 1.42918885e-01 -7.96567798e-01
8.94643843e-01 1.48013353e-01 1.59436509e-01 -2.64605075e-01
-4.15172011e-01 -1.52481347e-01 2.50913799e-01 -9.71397877e-01
6.43881619e-01 3.10478330e-01 -4.00403976e-01 7.56434023e-01
4.71343212e-02 -3.55523378e-01 4.34595615e-01 3.73362005e-01
2.00849700e+00 3.89573187e-01 4.70275044e-01 -2.42466241e-01
2.70638734e-01 9.70125943e-02 5.38178504e-01 5.28371334e-01
2.59072870e-01 1.10623553e-01 1.41110098e+00 -4.48048979e-01
-1.02661347e+00 -7.63103366e-01 3.44484210e-01 1.37278998e+00
-3.35065395e-01 -1.11981893e+00 -6.80534601e-01 -1.30544734e+00
1.37260079e-01 8.57193649e-01 -6.60041034e-01 -4.96062905e-01
-8.05168986e-01 -2.17617795e-01 5.90518415e-01 9.09233868e-01
-3.73384029e-01 -7.69104660e-01 -4.64345157e-01 2.17853427e-01
3.85962069e-01 -3.14142048e-01 -4.06187028e-01 4.61183101e-01
-9.65852201e-01 -1.62950301e+00 -1.89340293e-01 -8.96311164e-01
1.16254377e+00 -9.49807167e-02 1.68834996e+00 9.72574174e-01
-3.30358505e-01 4.51652944e-01 -2.89709449e-01 -2.07904711e-01
-9.70620632e-01 2.51148015e-01 -7.45679140e-01 -5.51563144e-01
3.01246077e-01 -4.99304354e-01 2.31964216e-01 -7.06759170e-02
-8.74062598e-01 -1.32185847e-01 3.39872003e-01 9.60515082e-01
5.66509515e-02 2.05152288e-01 9.70100984e-02 -1.40648115e+00
7.78856695e-01 -7.65609324e-01 -8.75774384e-01 5.76268673e-01
-6.80327117e-01 5.80921650e-01 5.52546263e-01 -1.38532341e-01
-9.56988394e-01 1.68822065e-01 1.42295629e-01 -1.96508527e-01
3.18902284e-01 1.10417378e+00 1.25593781e-01 -1.83909923e-01
8.87182832e-01 -1.35545194e-01 -2.57017612e-01 -3.25676978e-01
2.17486128e-01 3.06015193e-01 3.53922874e-01 -9.98963237e-01
1.02318609e+00 -1.04589924e-01 5.91069050e-02 -3.00840825e-01
-1.58573747e-01 -4.62879062e-01 -7.24783778e-01 1.30121633e-01
5.33368468e-01 -5.99036813e-01 -3.78947347e-01 1.51803911e-01
-1.16310310e+00 -8.37894440e-01 -2.72131711e-01 -8.45918953e-02
-2.43255094e-01 4.56349283e-01 -6.21892273e-01 -7.56655753e-01
-4.30760719e-02 -1.26925898e+00 1.01108241e+00 -1.07051849e-01
-5.69078624e-01 -1.34033418e+00 4.87737834e-01 -2.64196713e-02
4.18388933e-01 3.32646668e-01 1.61935711e+00 -4.92983073e-01
-6.89067662e-01 -4.98833805e-01 -1.22162916e-01 1.89721331e-01
3.90958369e-01 4.47204053e-01 -5.09858131e-01 -3.22505862e-01
-4.27227259e-01 -1.21863373e-01 7.07869768e-01 -6.98591322e-02
1.24319470e+00 -1.22611023e-01 -6.59344673e-01 4.85878050e-01
1.41321838e+00 -3.39873023e-02 4.30747241e-01 8.52565616e-02
9.08272982e-01 5.63736379e-01 2.91765898e-01 3.20963562e-01
4.96371776e-01 2.60682255e-01 6.00913346e-01 9.03398544e-02
-1.16048619e-01 -3.44407797e-01 3.38360250e-01 6.13929570e-01
-2.64351610e-02 -3.41505371e-02 -1.71513081e+00 1.14261413e+00
-1.97029376e+00 -7.16741443e-01 -6.59955442e-01 2.13059258e+00
8.19827378e-01 1.49294153e-01 -1.76183864e-01 1.45341367e-01
6.47506058e-01 5.41025288e-02 -2.07206443e-01 -6.55594409e-01
6.94746792e-01 3.32951933e-01 4.45521384e-01 6.13929868e-01
-8.94434512e-01 6.44939661e-01 6.58695698e+00 1.40248641e-01
-7.92793214e-01 2.09678680e-01 2.08259299e-01 3.97766441e-01
-7.35955179e-01 7.84515500e-01 -3.69340628e-01 1.05355136e-01
1.25234807e+00 -5.69459617e-01 6.42742157e-01 1.34652269e+00
-3.01172346e-01 2.03370620e-02 -1.94520605e+00 5.10139167e-01
2.33440418e-02 -1.19146419e+00 -4.33677107e-01 -2.06268251e-01
9.32204485e-01 -1.42066097e-02 -3.54559690e-01 6.43979371e-01
8.16602588e-01 -1.11190486e+00 5.19805074e-01 4.73432034e-01
7.67047465e-01 -5.56188822e-01 4.83630419e-01 2.05865145e-01
-1.39879847e+00 -2.08058506e-01 -2.24008814e-01 -2.31042057e-01
-2.43800148e-01 6.02911890e-01 -1.19582772e+00 6.04579568e-01
3.00501019e-01 9.22218561e-01 -1.08378422e+00 1.00584149e+00
-4.55081075e-01 5.46897233e-01 2.92764306e-01 1.41325906e-01
-2.58592278e-01 1.30122766e-01 1.25174820e-01 1.63595712e+00
3.38790327e-01 -5.90088546e-01 2.58469194e-01 1.18045259e+00
-3.24726611e-01 -3.80637556e-01 -9.37548697e-01 -3.69154930e-01
1.64084569e-01 1.20624447e+00 -7.64578044e-01 -4.42455381e-01
-8.10080886e-01 4.30726349e-01 6.32231176e-01 4.15606469e-01
-5.63677371e-01 -7.25406826e-01 5.27647138e-01 3.51710439e-01
2.00379446e-01 -5.31021774e-01 -3.56707811e-01 -1.15493119e+00
3.51285428e-01 -7.27733374e-01 3.53819907e-01 -6.67284191e-01
-8.73792768e-01 3.55641723e-01 5.59788793e-02 -6.03964984e-01
-4.81083304e-01 -3.76262665e-01 -1.08799708e+00 8.32461238e-01
-1.34074807e+00 -1.10654020e+00 -3.17304790e-01 2.13327274e-01
5.30608416e-01 -7.72077516e-02 7.68086374e-01 2.07459867e-01
-4.71876591e-01 6.40718043e-01 -2.20805779e-01 3.11249346e-01
2.18171790e-01 -1.55248296e+00 1.18558955e+00 1.37544119e+00
2.54647523e-01 1.16245353e+00 6.16374135e-01 -9.49475408e-01
-2.05444431e+00 -1.22664404e+00 9.14990783e-01 -5.65017402e-01
1.01505864e+00 -5.40593505e-01 -1.29194057e+00 1.56077933e+00
1.49948508e-01 3.28198880e-01 2.85348207e-01 7.33330667e-01
-8.69680882e-01 2.30125248e-01 -7.81085432e-01 3.21241140e-01
1.28091490e+00 -1.07013845e+00 -7.66890705e-01 5.34176409e-01
5.62305927e-01 -4.99193221e-01 -1.11136067e+00 -9.16516185e-02
1.01982705e-01 -6.37443125e-01 6.17597580e-01 -8.72904778e-01
8.14101875e-01 -2.43517756e-01 2.33840421e-01 -1.42018032e+00
-3.29546839e-01 -8.51020455e-01 -3.96550030e-01 1.31975770e+00
7.32076526e-01 -4.34374511e-01 9.02687848e-01 8.15755606e-01
-6.48451388e-01 -4.89861637e-01 -4.09885406e-01 -6.11712396e-01
-4.22297493e-02 -3.92214328e-01 6.83597684e-01 1.15137482e+00
7.82596886e-01 7.40765482e-02 2.36726254e-01 1.36262566e-01
4.23326522e-01 2.83436835e-01 8.87086034e-01 -1.37005782e+00
-6.06928647e-01 -4.10041869e-01 -4.70664501e-01 -5.55148840e-01
8.30428243e-01 -1.37682092e+00 1.19566821e-01 -1.84684980e+00
4.86125648e-01 -5.19679546e-01 -6.65896088e-02 9.43625808e-01
-3.12460899e-01 -4.85975713e-01 -2.64738828e-01 -1.97945684e-02
-5.74468970e-01 -9.88519043e-02 6.20186090e-01 -6.28297389e-01
-9.38533545e-02 1.58900321e-01 -5.94627976e-01 5.19214392e-01
5.71222425e-01 -9.62456763e-01 -4.88716751e-01 -7.32358336e-01
8.47168028e-01 6.46905124e-01 1.25716090e-01 -7.42878079e-01
2.82618046e-01 -2.28988618e-01 -2.86788166e-01 -3.51618528e-02
-1.42269194e-01 -7.13343799e-01 3.45540047e-01 5.77278674e-01
-5.74256957e-01 5.45159996e-01 3.02224457e-01 5.53220451e-01
-2.77545571e-01 -9.22731817e-01 2.85678416e-01 -3.60433519e-01
-7.88300216e-01 1.81322932e-01 -2.83220053e-01 1.06620796e-01
1.02677035e+00 -2.19227344e-01 -5.77855587e-01 2.07764313e-01
-5.69753826e-01 6.12973720e-02 9.94900107e-01 6.20301008e-01
4.51210171e-01 -7.19937921e-01 -5.86455464e-01 3.07178468e-01
5.37321329e-01 -2.07613129e-02 2.05760095e-02 5.91382146e-01
-8.68452430e-01 4.78319041e-02 -5.66940159e-02 -3.34729135e-01
-1.51268589e+00 8.59495938e-01 2.40831047e-01 -6.44231141e-01
-4.21955973e-01 6.92175806e-01 -3.85008976e-02 -5.53853929e-01
-1.16004981e-01 -8.44646454e-01 2.75228828e-01 -4.34952050e-01
1.96135432e-01 3.74684393e-01 4.48076665e-01 -7.11952820e-02
-4.32775050e-01 2.50205219e-01 -2.17134863e-01 3.77407044e-01
1.47502983e+00 4.53010976e-01 -7.24450469e-01 3.32283080e-01
1.11130202e+00 3.78402650e-01 -1.02606118e+00 -1.18564367e-01
7.34652877e-01 -6.27276599e-01 -9.54364799e-03 -4.54851359e-01
-1.00769663e+00 8.89185846e-01 -1.74227983e-01 4.13916767e-01
5.35346746e-01 1.46530882e-01 1.14577539e-01 6.49578094e-01
6.25113547e-01 -6.66413724e-01 1.16037987e-01 4.82612073e-01
5.97938359e-01 -1.04727197e+00 2.70070583e-01 -4.41099435e-01
-4.63761352e-02 1.33299625e+00 6.07912362e-01 -1.29440412e-01
4.59280699e-01 7.76809156e-01 -4.75788087e-01 -5.56645811e-01
-8.91600490e-01 3.65658611e-01 7.28333816e-02 9.14593101e-01
8.57503653e-01 3.87976579e-02 1.15277268e-01 3.49222779e-01
2.49349877e-01 3.09273843e-02 1.13388014e+00 1.66457152e+00
-2.02160269e-01 -1.46245563e+00 -1.88744515e-01 8.19548607e-01
-5.89548588e-01 -3.54250997e-01 -4.27060097e-01 9.75927114e-01
-6.87192976e-02 8.10202241e-01 -7.42454156e-02 -3.47899944e-01
4.36571360e-01 2.95417666e-01 5.91264188e-01 -1.33937669e+00
-1.05545795e+00 -6.69261634e-01 4.67978418e-01 -5.67634761e-01
4.09907140e-02 -6.53168023e-01 -1.37543654e+00 -2.55369753e-01
-4.48509574e-01 7.18155727e-02 8.28929186e-01 5.41874349e-01
5.75352728e-01 1.13713253e+00 8.71578678e-02 -5.16318679e-01
-4.71755624e-01 -6.67551458e-01 -3.74921143e-01 4.89938885e-01
5.41008413e-01 -4.92946118e-01 -1.12554312e-01 3.68948579e-01] | [7.56673002243042, 7.8383941650390625] |
a57a271f-eacd-4ec7-8614-ea9de0140663 | weisfeiler-and-lehman-go-cellular-cw-networks | 2106.12575 | null | https://arxiv.org/abs/2106.12575v3 | https://arxiv.org/pdf/2106.12575v3.pdf | Weisfeiler and Lehman Go Cellular: CW Networks | Graph Neural Networks (GNNs) are limited in their expressive power, struggle with long-range interactions and lack a principled way to model higher-order structures. These problems can be attributed to the strong coupling between the computational graph and the input graph structure. The recently proposed Message Passing Simplicial Networks naturally decouple these elements by performing message passing on the clique complex of the graph. Nevertheless, these models can be severely constrained by the rigid combinatorial structure of Simplicial Complexes (SCs). In this work, we extend recent theoretical results on SCs to regular Cell Complexes, topological objects that flexibly subsume SCs and graphs. We show that this generalisation provides a powerful set of graph "lifting" transformations, each leading to a unique hierarchical message passing procedure. The resulting methods, which we collectively call CW Networks (CWNs), are strictly more powerful than the WL test and not less powerful than the 3-WL test. In particular, we demonstrate the effectiveness of one such scheme, based on rings, when applied to molecular graph problems. The proposed architecture benefits from provably larger expressivity than commonly used GNNs, principled modelling of higher-order signals and from compressing the distances between nodes. We demonstrate that our model achieves state-of-the-art results on a variety of molecular datasets. | ['Michael Bronstein', 'Guido Montúfar', 'Pietro Liò', 'Yu Guang Wang', 'Nina Otter', 'Fabrizio Frasca', 'Cristian Bodnar'] | 2021-06-23 | null | http://proceedings.neurips.cc/paper/2021/hash/157792e4abb490f99dbd738483e0d2d4-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/157792e4abb490f99dbd738483e0d2d4-Paper.pdf | neurips-2021-12 | ['graph-property-prediction', 'graph-regression'] | ['graphs', 'graphs'] | [ 5.63935697e-01 5.11638939e-01 -1.87364057e-01 -2.07594335e-02
3.96320187e-02 -6.69037998e-01 9.83763516e-01 2.87724465e-01
-3.99019539e-01 8.43237102e-01 -6.95317611e-02 -5.25703549e-01
-6.88900232e-01 -1.10212672e+00 -1.07294595e+00 -9.49803293e-01
-7.57213652e-01 7.37857044e-01 2.27385119e-01 -5.28457105e-01
6.39611036e-02 8.08034003e-01 -1.36269391e+00 6.66284934e-02
7.03060091e-01 6.08646631e-01 -1.00841653e-02 7.72845328e-01
-9.71381292e-02 5.67730010e-01 1.22670196e-02 -3.06024164e-01
1.79989159e-01 -4.05630499e-01 -7.30836451e-01 -9.76321101e-02
4.72452372e-01 3.15705657e-01 -4.98785526e-01 1.01327825e+00
3.38107824e-01 -1.89872496e-02 5.83858490e-01 -1.17282832e+00
-3.88905585e-01 9.27486956e-01 -1.34076819e-01 -1.35166079e-01
3.01921200e-02 -5.79900369e-02 1.28548098e+00 -6.02559268e-01
1.04940629e+00 1.22141695e+00 8.08918178e-01 7.65176892e-01
-1.92040396e+00 -1.55374214e-01 9.96587705e-03 5.95564246e-02
-1.32513249e+00 -2.66446680e-01 9.49412584e-01 -3.39191258e-01
7.79197514e-01 6.89088821e-01 7.68201828e-01 1.04195356e+00
1.37673944e-01 3.01019341e-01 1.11885107e+00 -4.01434302e-01
3.29567462e-01 -4.40238804e-01 3.22452575e-01 1.11432719e+00
4.88671184e-01 5.36585264e-02 -3.12090248e-01 -3.63171548e-01
8.42437744e-01 5.91230802e-02 -5.02625406e-01 -8.09503198e-01
-1.25558364e+00 9.55524921e-01 5.82991004e-01 5.19742310e-01
-6.80560619e-02 4.42856133e-01 3.51010650e-01 2.40081713e-01
4.24926281e-02 4.15147603e-01 -1.96042404e-01 5.05844414e-01
-5.74553609e-01 2.42074728e-01 1.14721751e+00 7.19164729e-01
6.66005373e-01 -1.01847306e-01 9.14637595e-02 5.22525907e-01
2.17541844e-01 3.07822563e-02 -5.46165043e-03 -6.67937338e-01
8.85155201e-02 5.31881869e-01 -5.05563557e-01 -1.23729753e+00
-1.03027463e+00 -8.02688003e-01 -1.57278943e+00 8.13439488e-02
5.46635568e-01 4.26840335e-01 -8.81324470e-01 2.02339697e+00
3.09822172e-01 1.53480038e-01 -2.14003455e-02 5.47860861e-01
6.94521368e-01 3.37553382e-01 2.96877995e-02 -2.09121212e-01
1.08291686e+00 -3.60162705e-01 -4.05000210e-01 2.60006547e-01
7.69690156e-01 -2.30730083e-02 6.47266984e-01 4.99307662e-01
-1.33013666e+00 -2.13634819e-01 -1.05870736e+00 -6.03840090e-02
-5.42870283e-01 -3.86813074e-01 1.07235563e+00 5.76393664e-01
-1.31278050e+00 1.16108787e+00 -7.66352952e-01 -2.32901409e-01
5.73855340e-01 7.43233323e-01 -5.65844655e-01 -2.17634765e-03
-1.24647236e+00 5.73917508e-01 5.17166674e-01 4.13377494e-01
-6.66704834e-01 -4.89132285e-01 -7.64982998e-01 1.26753405e-01
4.15466249e-01 -9.24646378e-01 4.22364384e-01 -4.64663804e-01
-1.32566488e+00 9.30913270e-01 2.70883799e-01 -8.39700699e-01
5.77922642e-01 6.25327766e-01 -4.76794727e-02 2.27863327e-01
-4.75578964e-01 5.97081244e-01 5.15190482e-01 -1.07931352e+00
5.32416925e-02 -3.87445927e-01 3.82275075e-01 -2.24527165e-01
-1.41075969e-01 -5.36107838e-01 -1.45325109e-01 -4.95200098e-01
2.95363218e-01 -1.10951185e+00 -5.20013273e-01 -1.29219973e-02
-7.28637338e-01 -2.42150038e-01 6.62069976e-01 9.27624479e-03
1.10043299e+00 -1.89978385e+00 7.75421917e-01 5.99305570e-01
9.24118936e-01 1.05129026e-01 -2.74688542e-01 6.81007802e-01
-1.61612734e-01 3.66553098e-01 -4.81990308e-01 -3.65923764e-03
2.99819261e-01 7.01229572e-01 -6.43591955e-02 8.05934191e-01
2.37971563e-02 1.10092735e+00 -7.41936684e-01 -3.37556303e-01
6.11612871e-02 6.35404289e-01 -7.85622418e-01 -4.16442335e-01
-4.72111553e-01 2.60060757e-01 -2.25098491e-01 1.44045293e-01
8.20311487e-01 -6.64864302e-01 6.64158940e-01 -1.80653796e-01
1.52843773e-01 1.99560940e-01 -1.15872335e+00 1.57792115e+00
1.68146454e-02 3.08956325e-01 3.75363231e-01 -1.45761681e+00
7.00653970e-01 -8.46952014e-03 4.28720206e-01 -3.26098949e-01
-8.50843906e-04 1.41117483e-01 3.65370721e-01 -1.27898365e-01
3.80620211e-02 -1.75524652e-01 1.06784604e-01 1.12254664e-01
1.76429138e-01 5.19610867e-02 4.96416479e-01 4.37831670e-01
1.23267066e+00 -1.06888689e-01 9.75141525e-02 -7.06503570e-01
6.10692322e-01 -4.51942891e-01 3.59374553e-01 9.17181790e-01
2.20607892e-01 3.32164943e-01 1.12282157e+00 -6.25489116e-01
-1.01852572e+00 -1.08731651e+00 -1.28113657e-01 1.02680254e+00
-1.50234848e-02 -6.49696708e-01 -7.97955036e-01 -4.70475793e-01
-6.71417732e-03 9.66154113e-02 -7.85913229e-01 -2.56272584e-01
-7.31546879e-01 -1.10995746e+00 9.65599656e-01 4.51511592e-02
1.25885174e-01 -6.07543230e-01 -2.72393525e-01 2.85710573e-01
1.44614518e-01 -1.04864430e+00 -1.43824905e-01 4.17935848e-01
-1.01534021e+00 -1.30865037e+00 -3.08352143e-01 -6.62995517e-01
6.85447812e-01 -3.99512835e-02 1.05793667e+00 5.05235672e-01
-3.22727889e-01 1.50791466e-01 1.31187409e-01 -9.01372209e-02
-6.20584667e-01 2.38410905e-01 1.28278822e-01 1.42162099e-01
-2.43315935e-01 -1.01296258e+00 -4.03548092e-01 1.27445966e-01
-1.35588729e+00 3.70438725e-01 4.94016945e-01 1.00054610e+00
5.54452121e-01 3.20432261e-02 4.68539983e-01 -1.15759671e+00
5.07741272e-01 -3.49799335e-01 -6.92564309e-01 1.24430850e-01
-5.29954910e-01 5.12617826e-01 8.88107479e-01 -4.04953063e-01
-2.73707926e-01 9.14282501e-02 -8.93138945e-02 4.67636250e-03
2.91745327e-02 6.61542714e-01 -7.15977550e-02 -5.96757889e-01
6.44131064e-01 1.70080677e-01 7.04013631e-02 -2.79017508e-01
5.17108500e-01 -6.66089356e-02 4.81341451e-01 -7.85377681e-01
7.08767653e-01 8.20238769e-01 1.20315623e+00 -1.20019877e+00
-2.23686904e-01 9.00017284e-03 -6.53615654e-01 -8.06354880e-02
6.37741745e-01 -3.45685303e-01 -1.19091868e+00 3.90517890e-01
-1.20733130e+00 -2.37137854e-01 -1.51281610e-01 2.75668561e-01
-6.79355621e-01 6.86542928e-01 -9.96605754e-01 -6.32088840e-01
-3.31618264e-02 -9.35331404e-01 7.76624680e-01 -3.32165390e-01
1.47752568e-01 -1.27246940e+00 9.21334475e-02 -3.77792567e-02
4.05324996e-01 8.05398524e-01 1.45734489e+00 -7.64580011e-01
-8.09449494e-01 1.20227793e-02 -8.79134983e-02 -2.61464920e-02
-3.84836346e-01 -6.76128566e-02 -6.59211040e-01 -3.38412821e-01
-1.59636527e-01 -1.37377784e-01 1.23745990e+00 3.40179503e-01
1.15675879e+00 -5.37871480e-01 -3.89493108e-01 7.21946657e-01
1.54416347e+00 -5.09730399e-01 6.76352680e-01 -1.13392219e-01
1.04652309e+00 7.30967581e-01 -5.48979461e-01 3.87466773e-02
5.73754497e-02 6.77293539e-01 7.42039084e-01 -1.75084651e-01
2.61858776e-02 -2.06315219e-01 5.21233976e-02 1.29826796e+00
-6.15032256e-01 -2.18850732e-01 -7.80702770e-01 6.59283325e-02
-1.94234872e+00 -1.04641426e+00 -6.57765508e-01 2.12335682e+00
8.54301095e-01 3.21026713e-01 9.99164507e-02 3.02590430e-01
6.90797746e-01 2.74153739e-01 -1.62388697e-01 -2.86928564e-01
-5.89812756e-01 4.47099179e-01 6.95937634e-01 7.41310537e-01
-8.27061832e-01 5.50224423e-01 6.33252716e+00 1.03366518e+00
-8.58633876e-01 8.15514252e-02 3.82405728e-01 9.38436911e-02
-5.36840796e-01 4.77181152e-02 -5.38878381e-01 2.03924820e-01
9.45316970e-01 1.86483771e-01 7.43888557e-01 4.62475777e-01
-6.24320619e-02 4.90768731e-01 -1.39840114e+00 7.13505507e-01
-1.16133817e-01 -1.97280002e+00 3.37470591e-01 5.27156115e-01
5.33115625e-01 1.43849909e-01 -1.69178084e-01 -6.32887781e-02
3.17344248e-01 -1.38585281e+00 3.72560501e-01 5.44903457e-01
7.30101526e-01 -6.89724505e-01 1.73098132e-01 3.42023641e-01
-1.11514056e+00 1.05252095e-01 -4.54792470e-01 -1.29008129e-01
-1.43389106e-01 6.81499600e-01 -3.82114053e-01 6.87429190e-01
5.38846627e-02 4.63579029e-01 -4.23691303e-01 7.22945213e-01
1.37569636e-01 3.92746091e-01 -5.71213961e-01 -2.02540711e-01
3.62629622e-01 -3.59201580e-01 7.22604811e-01 1.34437013e+00
-4.77256477e-02 4.94622365e-02 9.40247625e-02 1.04792392e+00
-4.29699332e-01 -6.63142651e-02 -1.04023588e+00 -4.29464802e-02
6.08986244e-02 1.35421336e+00 -1.03780973e+00 -1.44563287e-01
-1.55737713e-01 5.06917357e-01 6.07678950e-01 3.88567299e-01
-5.58145523e-01 -2.91562170e-01 4.11078870e-01 1.75982520e-01
3.75866622e-01 -4.43814993e-01 -6.10873941e-03 -1.14560831e+00
-3.63306589e-02 -8.98336828e-01 3.03604782e-01 -2.81132311e-01
-1.37716341e+00 3.25855941e-01 -1.46337569e-01 -6.30027413e-01
1.14294074e-01 -1.15801680e+00 -3.03347260e-01 3.73464346e-01
-1.29098117e+00 -1.21171832e+00 -1.91058661e-03 6.57323062e-01
-2.51157433e-01 2.68061787e-01 9.20926392e-01 2.81279624e-01
-1.89988568e-01 3.17634016e-01 1.46527871e-01 7.29570389e-02
-5.90001754e-02 -1.40980208e+00 3.44106972e-01 4.73342448e-01
3.76015544e-01 7.86802292e-01 6.46564066e-01 -4.45770204e-01
-1.97428000e+00 -1.04632270e+00 7.68649518e-01 -4.98807460e-01
7.86929429e-01 -1.04664695e+00 -1.01094592e+00 5.07938921e-01
-2.04001024e-01 2.01907054e-01 5.36526322e-01 2.37223983e-01
-5.46994627e-01 7.53267258e-02 -9.11125541e-01 7.35390604e-01
1.64981365e+00 -6.15889609e-01 -2.39345953e-01 6.16521060e-01
7.41932213e-01 -2.28294805e-01 -9.54492509e-01 5.35727441e-01
4.93910432e-01 -1.02289951e+00 1.19893205e+00 -9.84793723e-01
3.20829272e-01 -2.62405485e-01 -1.40513197e-01 -9.35633123e-01
-4.68994111e-01 -9.14974630e-01 -2.46511489e-01 5.83746552e-01
3.63131553e-01 -9.12217140e-01 7.29429662e-01 1.24122530e-01
-1.23497844e-01 -9.38009560e-01 -1.19156277e+00 -7.69975662e-01
2.29405791e-01 -3.22435111e-01 4.31702018e-01 1.26940000e+00
1.71886772e-01 5.50406754e-01 -3.11348259e-01 -3.80823351e-02
9.25851524e-01 -1.84475109e-01 5.19549012e-01 -1.75460851e+00
-6.91922486e-01 -8.43084157e-01 -7.68155396e-01 -8.60529125e-01
1.43731341e-01 -1.48130584e+00 -3.25222999e-01 -1.21645558e+00
1.91855237e-01 -4.43772405e-01 -2.37719849e-01 2.17323169e-01
4.13506806e-01 2.48318315e-01 1.96431085e-01 6.83803782e-02
-6.62574232e-01 2.65512705e-01 1.31297755e+00 -2.03853682e-01
-3.66711318e-02 -3.58481884e-01 -5.29584289e-01 7.84358203e-01
6.38821006e-01 -3.59804332e-01 -2.66848236e-01 -1.14835121e-01
8.64679992e-01 7.44406059e-02 7.84999967e-01 -8.00142348e-01
5.28309107e-01 7.53113925e-02 1.35653704e-01 -2.14466289e-01
2.36789778e-01 -6.01705432e-01 5.40385127e-01 8.56496334e-01
-6.37069881e-01 -2.16515943e-01 4.21489067e-02 8.26298356e-01
2.72081494e-01 -1.22476541e-01 8.02067459e-01 -2.67931193e-01
6.87483773e-02 4.74312454e-01 -1.73112780e-01 -5.80626242e-02
5.56761026e-01 -1.10927001e-01 -5.48462629e-01 -9.70283747e-02
-1.01264739e+00 -2.37838887e-02 4.01819199e-01 -7.69730210e-02
3.86126518e-01 -1.31964970e+00 -6.45816326e-01 2.10663930e-01
-1.27737075e-01 3.65922414e-03 1.23793922e-01 1.07401013e+00
-5.34014404e-01 5.96572280e-01 -1.68396384e-01 -7.00345576e-01
-1.11014199e+00 6.87716067e-01 4.45710510e-01 -6.35171592e-01
-6.42405152e-01 4.96803194e-01 5.47519922e-01 -6.34614646e-01
1.44418433e-01 -5.36569953e-01 -1.99829228e-02 -2.39674032e-01
1.03669047e-01 4.24772143e-01 4.65026386e-02 -5.56768477e-01
-2.39945009e-01 4.94626015e-01 9.58899781e-02 2.46002689e-01
1.37300050e+00 2.67951578e-01 -7.32715309e-01 2.83664197e-01
1.16127813e+00 -9.60206613e-03 -8.93663764e-01 -6.92920908e-02
4.14679870e-02 2.20120758e-01 -2.96341568e-01 -3.53520811e-01
-7.36028016e-01 7.73906350e-01 1.62837178e-01 8.27006638e-01
7.36025870e-01 1.61920160e-01 5.25360048e-01 7.69004941e-01
4.60155666e-01 -8.30248654e-01 -2.79362738e-01 5.07880449e-01
7.75816858e-01 -6.78802848e-01 -8.30531418e-02 -6.52819455e-01
3.80407482e-01 1.18432689e+00 -2.29686368e-02 -3.46277356e-01
4.91695523e-01 9.07220915e-02 -7.58914113e-01 -5.56221068e-01
-7.69445777e-01 -1.63432047e-01 2.38004759e-01 5.77248514e-01
1.58126324e-01 2.16716826e-01 -5.98632991e-01 3.65173280e-01
-7.74733722e-02 -1.61636293e-01 6.83524132e-01 6.49831831e-01
-4.21898156e-01 -1.15530908e+00 -9.06990319e-02 3.96166176e-01
-5.10187805e-01 -1.58606738e-01 -6.03224933e-01 1.05016923e+00
2.04766124e-01 5.22461295e-01 -2.59166956e-01 -2.52732635e-01
1.29697919e-01 -1.58496127e-01 9.06143486e-01 -3.37900192e-01
-2.49057353e-01 -8.71165842e-02 1.61152855e-01 -6.05650902e-01
-6.56895041e-01 -3.07948172e-01 -1.17682207e+00 -6.66574001e-01
-2.96546608e-01 1.44106597e-01 6.06212258e-01 8.41253817e-01
4.76070702e-01 5.22354007e-01 2.14505836e-01 -9.32760537e-01
-5.50531685e-01 -5.68754017e-01 -6.33093357e-01 3.64164710e-01
4.71161067e-01 -5.42379975e-01 -4.66860265e-01 -2.29283869e-01] | [6.805246829986572, 6.111117839813232] |
28090b96-7c4e-422c-9bc5-244256aee313 | deep-learning-for-optimal-volt-var-control | 2211.09557 | null | https://arxiv.org/abs/2211.09557v1 | https://arxiv.org/pdf/2211.09557v1.pdf | Deep Learning for Optimal Volt/VAR Control using Distributed Energy Resources | Given their intermittency, distributed energy resources (DERs) have been commissioned with regulating voltages at fast timescales. Although the IEEE 1547 standard specifies the shape of Volt/VAR control rules, it is not clear how to optimally customize them per DER. Optimal rule design (ORD) is a challenging problem as Volt/VAR rules introduce nonlinear dynamics, require bilinear optimization models, and lurk trade-offs between stability and steady-state performance. To tackle ORD, we develop a deep neural network (DNN) that serves as a digital twin of Volt/VAR dynamics. The DNN takes grid conditions as inputs, uses rule parameters as weights, and computes equilibrium voltages as outputs. Thanks to this genuine design, ORD is reformulated as a deep learning task using grid scenarios as training data and aiming at driving the predicted variables being the equilibrium voltages close to unity. The learning task is solved by modifying efficient deep-learning routines to enforce constraints on rule parameters. In the course of DNN-based ORD, we also review and expand on stability conditions and convergence rates for Volt/VAR rules on single-/multi-phase feeders. To benchmark the optimality and runtime of DNN-based ORD, we also devise a novel mixed-integer nonlinear program formulation. Numerical tests showcase the merits of DNN-based ORD. | ['Vassilis Kekatos', 'Spyros Chatzivasileiadis', 'Sarthak Gupta'] | 2022-11-17 | null | null | null | null | ['unity'] | ['computer-vision'] | [-5.87126911e-01 -3.40900093e-01 1.61588229e-02 1.02022197e-02
-5.22099316e-01 -8.88206720e-01 3.03689867e-01 1.27343997e-01
1.14313319e-01 1.17044723e+00 -2.91793376e-01 -5.71554244e-01
-6.46544814e-01 -8.90257537e-01 -5.01266301e-01 -9.95167971e-01
-4.83230650e-01 5.09196043e-01 -5.80888391e-01 -5.39325356e-01
-1.01583757e-01 9.70521152e-01 -1.05468774e+00 -1.19165950e-01
1.21723700e+00 1.35026205e+00 -1.16581306e-01 5.39451480e-01
4.24858123e-01 6.37400806e-01 -8.63499463e-01 1.42305508e-01
5.52691162e-01 -4.54037666e-01 -4.00226563e-01 -3.84695688e-03
-2.36880317e-01 -3.44112486e-01 -4.34337139e-01 1.01040900e+00
9.20435131e-01 4.73768741e-01 6.82334542e-01 -1.35356152e+00
-7.48707652e-01 6.11718893e-01 -1.87552929e-01 3.92436117e-01
9.07359272e-02 5.45341134e-01 9.42232788e-01 -7.54032373e-01
1.67128474e-01 6.92675471e-01 7.69995570e-01 -8.40205848e-02
-1.64920187e+00 -2.40009710e-01 2.47558147e-01 5.20125628e-01
-1.60470605e+00 -1.37884080e-01 7.96125472e-01 -4.87444669e-01
1.48743796e+00 1.79851234e-01 1.02486861e+00 5.05829930e-01
5.37730813e-01 4.26113129e-01 8.29881191e-01 -1.22086726e-01
6.55326724e-01 -2.34417289e-01 -1.08470410e-01 -1.37239194e-03
1.93813935e-01 2.74479657e-01 -4.48283516e-02 7.77518824e-02
6.09276235e-01 -5.31983078e-01 -6.36966467e-01 -5.17669499e-01
-5.71628988e-01 1.04110253e+00 5.91710985e-01 4.24790949e-01
-6.61838889e-01 -9.17828232e-02 5.90987206e-01 5.91593444e-01
2.23656863e-01 6.30633235e-01 -6.50044560e-01 -2.51687877e-02
-7.77880609e-01 4.60789412e-01 9.88900483e-01 5.44149041e-01
2.20624104e-01 1.10386288e+00 2.46423502e-02 6.17279410e-01
-6.44084066e-02 5.34123898e-01 3.61310929e-01 -8.83162916e-01
1.53058648e-01 4.64296222e-01 2.94147998e-01 -7.88680315e-01
-6.86858654e-01 -7.29964495e-01 -1.30110037e+00 6.44210577e-01
3.44414353e-01 -6.64883852e-01 -4.14212912e-01 1.47391880e+00
2.31080577e-01 -4.66573358e-01 -1.51074275e-01 9.06215906e-01
-1.17977343e-01 1.16045022e+00 -4.14495558e-01 -7.22518981e-01
8.57441247e-01 -4.79615897e-01 -1.10408044e+00 3.62097204e-01
6.15358293e-01 -4.24663186e-01 6.75024629e-01 7.18751788e-01
-1.34185827e+00 -2.78279036e-01 -1.18043363e+00 3.42234194e-01
-7.31438696e-01 1.23522148e-01 1.87623501e-01 3.77223074e-01
-1.37207687e+00 8.72636378e-01 -7.46185303e-01 1.57634437e-01
1.46474525e-01 5.34829915e-01 1.84824660e-01 7.28675246e-01
-1.52302349e+00 1.59770513e+00 6.96350098e-01 8.28236997e-01
-6.01824105e-01 -1.12126374e+00 -8.05522203e-01 3.77555639e-01
2.42380962e-01 -4.95435774e-01 1.16519892e+00 -8.99959981e-01
-1.84066558e+00 -4.75260317e-02 4.24623787e-01 -8.14806879e-01
5.03684044e-01 1.36338279e-01 -6.05841219e-01 -5.72436415e-02
-3.42400640e-01 1.44435009e-02 5.44186711e-01 -9.77844477e-01
-3.76951844e-01 1.87350750e-01 -2.62924563e-02 -1.73866805e-02
-1.05165713e-01 -3.43339413e-01 5.02381146e-01 -8.25283289e-01
-4.84933883e-01 -3.13099474e-01 -2.59378016e-01 -2.36025974e-01
-2.39279270e-01 -4.98814881e-01 7.89422512e-01 -1.06702626e+00
1.41748202e+00 -1.65994561e+00 2.80676067e-01 4.77626234e-01
-1.88213482e-01 3.49413306e-01 1.85697809e-01 5.82592487e-01
-6.15155637e-01 -3.04216862e-01 -2.64831513e-01 2.82857865e-01
6.06989741e-01 3.28286409e-01 -3.46444607e-01 5.48181832e-01
4.08254802e-01 7.91861117e-01 -6.91256762e-01 2.55169928e-01
6.49032295e-01 6.16874695e-01 -4.35090512e-01 3.74464951e-02
-3.34424943e-01 1.16557255e-01 -1.89062878e-01 3.89092326e-01
6.08665705e-01 -1.73930883e-01 4.55349326e-01 -5.10635912e-01
-4.43188757e-01 1.13867521e-01 -1.52077270e+00 1.03469956e+00
-7.93458045e-01 6.14508390e-01 6.99676037e-01 -1.77773070e+00
8.20475399e-01 3.04134011e-01 7.61249900e-01 -1.10169780e+00
2.79044628e-01 2.61271179e-01 7.76996538e-02 -3.32297087e-01
7.65526444e-02 -5.27771935e-02 8.21059570e-02 3.20798516e-01
1.81901623e-02 -5.01408696e-01 6.33154452e-01 -1.82532981e-01
6.23089552e-01 -4.96268012e-02 3.56623203e-01 -8.65328312e-01
8.12759936e-01 5.56873903e-02 8.13338697e-01 1.34501532e-01
1.95204481e-01 9.77082923e-03 9.77949440e-01 -5.90431690e-01
-1.14495349e+00 -9.59908664e-01 -5.31001687e-01 5.34799635e-01
-3.49138826e-01 1.76004365e-01 -4.24500793e-01 -1.93806782e-01
2.17722714e-01 1.42962492e+00 -4.40977186e-01 -1.08381175e-01
-8.20751905e-01 -1.09634519e+00 -8.19655731e-02 5.44574022e-01
3.04473817e-01 -6.06627882e-01 -6.39861345e-01 7.16538429e-01
4.19558972e-01 -7.32752800e-01 -4.35722679e-01 9.94054794e-01
-3.45825166e-01 -1.11133540e+00 -5.92556775e-01 -7.35427141e-01
4.92997050e-01 -6.26373947e-01 1.18933058e+00 -1.19207375e-01
-1.49515077e-01 1.42442629e-01 4.73088697e-02 7.44696930e-02
-2.69169539e-01 8.01299512e-02 4.75342989e-01 -2.48322353e-01
-3.40723455e-01 -8.96529555e-01 -3.62665594e-01 2.37700120e-01
-7.53974557e-01 -1.76664338e-01 1.35610595e-01 1.10564804e+00
4.55254078e-01 4.84925091e-01 1.13592279e+00 -1.42826691e-01
9.11940277e-01 -4.82531995e-01 -1.45157182e+00 3.31188858e-01
-8.80844772e-01 4.87772748e-02 1.52508795e+00 -2.95651138e-01
-6.69607401e-01 7.81313516e-03 -1.20137431e-01 -4.77683544e-01
2.44957387e-01 4.72769469e-01 -4.22544360e-01 -2.75592268e-01
2.72528321e-01 1.52933881e-01 -1.79918483e-02 -2.79314280e-01
3.59137148e-01 4.13573891e-01 5.06372750e-01 -7.27338135e-01
7.99933732e-01 -2.16904506e-01 -2.97577064e-02 -4.43646133e-01
-4.10826623e-01 3.91011208e-01 -5.29594421e-01 -3.53022754e-01
5.08923709e-01 -6.89777195e-01 -1.29081047e+00 5.89562535e-01
-1.09170830e+00 -7.11152613e-01 -6.23219967e-01 3.91770117e-02
-6.32758200e-01 -4.64282781e-02 -6.56193137e-01 -6.92104042e-01
-5.12658715e-01 -1.15226460e+00 2.86755860e-01 2.63616085e-01
-3.97293605e-02 -1.39122951e+00 -6.79698959e-02 -3.22468489e-01
7.00571895e-01 6.57634020e-01 1.20110774e+00 -4.83471513e-01
-3.22369695e-01 1.06168807e-01 2.38512531e-02 6.60655141e-01
-2.82447655e-02 1.20902145e-02 -3.93825859e-01 -6.45667195e-01
2.69584358e-01 -1.98118687e-02 -9.13053676e-02 6.45376861e-01
1.03940177e+00 -7.20136821e-01 6.32074624e-02 8.77063334e-01
1.92888665e+00 8.55588019e-01 6.86750337e-02 3.53907913e-01
2.63191581e-01 4.56799231e-02 -4.25441563e-02 6.32252514e-01
4.58465159e-01 6.59123421e-01 4.62172210e-01 4.42477949e-02
4.45547014e-01 4.89729077e-01 3.84686589e-01 7.91101277e-01
6.54077828e-02 -1.66392118e-01 -8.58220458e-01 5.57580471e-01
-1.56262624e+00 -8.23418856e-01 -5.21238893e-02 1.85412169e+00
8.60028684e-01 1.88007757e-01 1.11553252e-01 5.01710534e-01
5.46909094e-01 2.18115281e-02 -9.77276444e-01 -1.02003777e+00
-4.58336055e-01 2.47625545e-01 4.41110790e-01 6.20869279e-01
-8.43557060e-01 2.36907490e-02 6.26825809e+00 7.01496542e-01
-1.45361638e+00 -5.44908307e-02 7.78698802e-01 -3.36275280e-01
-7.12590367e-02 -5.04610837e-01 -5.12901962e-01 6.52515709e-01
1.11610436e+00 -7.96275020e-01 1.01142311e+00 7.17834473e-01
1.03239489e+00 -2.42748726e-02 -1.07663679e+00 7.80489445e-01
-3.11311424e-01 -1.46476460e+00 -3.32130224e-01 -7.12771565e-02
1.24635100e+00 -1.01331860e-01 -2.38131568e-01 4.96833324e-01
5.07039428e-01 -1.08180380e+00 8.63134384e-01 4.62075144e-01
4.29407716e-01 -1.08213711e+00 9.01449263e-01 8.81359503e-02
-1.09517121e+00 -5.72368681e-01 2.33655404e-02 -2.38421395e-01
5.30043662e-01 7.69343197e-01 -2.71130621e-01 8.97817910e-01
6.25964582e-01 6.89787149e-01 8.47234763e-03 8.57433498e-01
-2.07256615e-01 3.69523168e-01 -6.24673903e-01 -9.53196585e-02
2.28017434e-01 -7.07466304e-01 2.92855173e-01 9.01389241e-01
3.58956903e-01 2.19786361e-01 1.80905253e-01 1.14236140e+00
1.73858076e-01 -2.30464280e-01 -1.04919754e-01 2.25078642e-01
5.40272534e-01 1.19815743e+00 -4.25804287e-01 -1.04614653e-01
-2.00917616e-01 2.81852543e-01 3.67733650e-02 7.16852069e-01
-9.37964201e-01 -3.48260522e-01 9.06505704e-01 -1.33667812e-01
4.72937554e-01 -3.38161379e-01 -7.60154903e-01 -9.39572394e-01
5.40078580e-02 -1.05793488e+00 3.98212790e-01 -8.53285551e-01
-1.42224181e+00 3.95300090e-01 -2.92950794e-02 -1.15469372e+00
-6.48581445e-01 -6.64852262e-01 -9.71279800e-01 1.14585304e+00
-1.59719372e+00 -4.56723958e-01 2.02257559e-01 5.92652798e-01
5.36202848e-01 -2.72009429e-02 6.19487405e-01 4.38243330e-01
-1.13053274e+00 2.43407249e-01 8.19598734e-01 5.49444035e-02
-1.53686479e-01 -1.54376125e+00 -9.00065824e-02 1.01586509e+00
-6.95330441e-01 4.75853719e-02 1.01048386e+00 -1.75839856e-01
-1.80370724e+00 -7.98290074e-01 4.04282361e-01 3.82594883e-01
1.31029761e+00 -2.45134249e-01 -9.82855082e-01 4.99701530e-01
7.94560254e-01 1.56550720e-01 -2.62301881e-02 -4.18348312e-01
3.92693490e-01 -5.68041205e-01 -1.12135208e+00 2.10815310e-01
2.38056034e-01 -2.73277521e-01 -4.97458309e-01 3.74111056e-01
2.15863064e-01 -4.65934038e-01 -1.39267313e+00 3.14853996e-01
-7.86500722e-02 -6.15851581e-01 9.03195739e-01 -3.17387432e-01
1.92057565e-02 -4.63609904e-01 -2.70876940e-03 -2.23206735e+00
-1.81683645e-01 -1.03076017e+00 -6.84280694e-01 1.02169156e+00
1.46228895e-01 -1.01544392e+00 1.72168046e-01 4.49301392e-01
-3.31643522e-01 -1.14621639e+00 -1.28637850e+00 -8.90410662e-01
6.93325639e-01 -2.74064004e-01 9.52000439e-01 1.11984158e+00
1.77425280e-01 5.08761890e-02 1.12914644e-01 5.70294917e-01
6.15191162e-01 2.95375824e-01 4.51698005e-02 -7.72722721e-01
-1.12827800e-01 -1.01068437e+00 3.64934430e-02 -5.73830605e-01
1.46348700e-01 -6.29120588e-01 -1.72740281e-01 -1.58806288e+00
-9.32839692e-01 -2.32130572e-01 -3.15356314e-01 6.05526865e-01
2.29923010e-01 -4.07150000e-01 3.26822549e-01 -3.58888060e-01
4.43323851e-02 8.60599577e-01 8.56044114e-01 -4.06040609e-01
-2.71699905e-01 -1.49011567e-01 -1.47765741e-01 4.16978329e-01
1.02717638e+00 2.45861664e-01 -3.07894647e-01 -3.45402032e-01
4.99985754e-01 4.42201346e-01 3.02127570e-01 -1.01291013e+00
3.52968097e-01 -1.71419799e-01 5.31551003e-01 -8.00398350e-01
-9.29512754e-02 -1.03748679e+00 4.57991153e-01 6.79555297e-01
-3.78445052e-02 6.25243366e-01 2.90655434e-01 -1.63966551e-01
-2.27379322e-01 5.35052903e-02 8.68784785e-01 3.61823291e-01
-5.78176558e-01 8.75088349e-02 -8.32170546e-01 8.35365057e-02
1.16374540e+00 -4.25128415e-02 -1.40784085e-01 -2.69050330e-01
-9.79143322e-01 1.10452890e+00 1.26945451e-01 1.66914925e-01
3.94708514e-02 -1.45183587e+00 -7.01073349e-01 2.00279295e-01
-7.83560574e-01 -5.41068555e-04 3.19892824e-01 9.26937699e-01
-5.91914773e-01 5.24083078e-01 -9.09934640e-02 -5.95643640e-01
-3.53497207e-01 5.71074307e-01 1.24221623e+00 -4.28915232e-01
-4.67938572e-01 2.15576515e-01 -5.76033592e-01 -3.37890804e-01
2.25187354e-02 -5.53747416e-01 -4.86271642e-02 6.41166568e-01
2.58177429e-01 5.88397503e-01 5.50416410e-01 -1.46676870e-02
-1.32894769e-01 2.62463897e-01 4.78504807e-01 3.69431406e-01
1.82769847e+00 -6.99720085e-02 -1.64528877e-01 2.22404063e-01
1.11062610e+00 -7.07548916e-01 -1.47489095e+00 3.32151532e-01
-1.69800952e-01 2.39350542e-01 4.17296410e-01 -9.55472827e-01
-1.64244318e+00 4.49897856e-01 4.86131668e-01 9.95392382e-01
1.45939112e+00 -9.23582911e-01 6.68960929e-01 2.89064407e-01
2.92454779e-01 -1.61364627e+00 -6.86440110e-01 9.30103004e-01
1.05741811e+00 -5.86777687e-01 -2.46189032e-02 2.57621258e-01
-8.93645510e-02 1.43035662e+00 3.95151496e-01 -3.92562151e-01
7.72936463e-01 9.52621520e-01 -1.82397261e-01 2.45490834e-01
-8.63955200e-01 2.28698358e-01 -1.42571451e-02 4.75290835e-01
8.86439085e-02 5.10905907e-02 -5.00147939e-01 5.89911401e-01
-2.77339518e-01 1.43805131e-01 5.50115168e-01 6.45812273e-01
-6.99341893e-02 -6.27274811e-01 -6.02018535e-01 4.47595835e-01
-1.94471732e-01 1.84916615e-01 5.07665873e-01 1.01504505e+00
1.09208144e-01 6.65323853e-01 2.01597378e-01 2.00236872e-01
7.12682426e-01 6.48754016e-02 1.91384539e-01 1.69691101e-01
-1.06662059e+00 -1.24167450e-01 2.08059680e-02 -4.16663915e-01
2.27713630e-01 -6.95475221e-01 -1.34077168e+00 -5.85343182e-01
-2.54246086e-01 3.55737954e-01 4.20231611e-01 1.19428945e+00
-8.86595994e-03 9.50014472e-01 9.67391253e-01 -9.89554346e-01
-1.13643563e+00 -5.56290865e-01 -7.60186434e-01 -1.10126071e-01
3.57805490e-01 -4.96440709e-01 -5.86562037e-01 -2.36255854e-01] | [5.762209415435791, 2.600282669067383] |
d731e221-8dfc-4811-9159-71d084e04dbc | gda-generative-data-augmentation-techniques | 2305.16663 | null | https://arxiv.org/abs/2305.16663v2 | https://arxiv.org/pdf/2305.16663v2.pdf | GDA: Generative Data Augmentation Techniques for Relation Extraction Tasks | Relation extraction (RE) tasks show promising performance in extracting relations from two entities mentioned in sentences, given sufficient annotations available during training. Such annotations would be labor-intensive to obtain in practice. Existing work adopts data augmentation techniques to generate pseudo-annotated sentences beyond limited annotations. These techniques neither preserve the semantic consistency of the original sentences when rule-based augmentations are adopted, nor preserve the syntax structure of sentences when expressing relations using seq2seq models, resulting in less diverse augmentations. In this work, we propose a dedicated augmentation technique for relational texts, named GDA, which uses two complementary modules to preserve both semantic consistency and syntax structures. We adopt a generative formulation and design a multi-tasking solution to achieve synergies. Furthermore, GDA adopts entity hints as the prior knowledge of the generative model to augment diverse sentences. Experimental results in three datasets under a low-resource setting showed that GDA could bring {\em 2.0\%} F1 improvements compared with no augmentation technique. Source code and data are available. | ['Philip S. Yu', 'Irwin King', 'Chenwei Zhang', 'Xin Zhang', 'Zeqi Tan', 'Aiwei Liu', 'Xuming Hu'] | 2023-05-26 | null | null | null | null | ['relation-extraction'] | ['natural-language-processing'] | [ 4.03561860e-01 8.59522402e-01 -1.52681366e-01 -5.08969903e-01
-7.41269231e-01 -4.34928566e-01 5.17817140e-01 1.39339074e-01
-3.99674624e-01 9.92086709e-01 4.76330698e-01 -3.20164323e-01
5.34119718e-02 -8.67937684e-01 -7.24125862e-01 -2.42373690e-01
1.90662727e-01 4.53723758e-01 8.64395965e-03 -5.03123581e-01
-1.85060039e-01 -8.04511979e-02 -1.17851484e+00 5.58279157e-01
1.17195249e+00 5.67038596e-01 3.62944543e-01 4.30933148e-01
-5.70306480e-01 9.84526217e-01 -7.37601101e-01 -9.17890668e-01
1.37275040e-01 -5.18839121e-01 -1.11332405e+00 2.14227531e-02
4.41640690e-02 -9.93414223e-02 -3.58943045e-01 8.26049209e-01
5.21489143e-01 1.61562040e-01 2.93974251e-01 -1.02754629e+00
-7.47062624e-01 1.41989696e+00 -4.65392768e-01 1.79991916e-01
4.27353144e-01 -2.40619034e-01 1.22916245e+00 -7.71999896e-01
6.58570766e-01 1.08841264e+00 5.62111974e-01 6.63362205e-01
-1.15357542e+00 -5.55401683e-01 2.89237350e-01 4.25357148e-02
-1.40042067e+00 -6.57441139e-01 7.75907993e-01 1.27690621e-02
1.33599305e+00 4.20712918e-01 2.80688703e-01 1.17501140e+00
-2.90033430e-01 8.35445464e-01 7.61426687e-01 -6.40920758e-01
-7.68076852e-02 1.37879983e-01 1.38577968e-01 5.56188703e-01
2.70929605e-01 -5.82967699e-01 -6.39331520e-01 -1.46249076e-02
4.14727300e-01 -3.47162187e-01 -2.19671607e-01 1.04721718e-01
-1.12881172e+00 6.33234203e-01 2.60418326e-01 4.66073990e-01
-3.53308588e-01 -1.50448963e-01 4.70765382e-01 1.68971136e-01
7.68851757e-01 6.22052729e-01 -7.37992704e-01 -3.32317390e-02
-5.94375312e-01 1.94833800e-01 7.59111226e-01 1.47646391e+00
6.25732064e-01 -9.73958746e-02 -6.30426407e-01 9.52973545e-01
1.49435624e-01 3.37431848e-01 5.63102305e-01 -7.33463407e-01
1.10129189e+00 7.86363959e-01 -8.30949694e-02 -9.01616395e-01
-5.05586147e-01 -6.00304842e-01 -9.90740716e-01 -6.72332823e-01
2.45592698e-01 -4.41783756e-01 -7.06689358e-01 1.91863787e+00
4.46585715e-01 1.17792562e-02 4.74481434e-01 5.08909464e-01
1.13573885e+00 4.80591983e-01 1.43463790e-01 -2.72388250e-01
1.47428572e+00 -9.56738710e-01 -1.18074822e+00 -5.23451090e-01
1.27348113e+00 -7.37438619e-01 1.22203159e+00 -1.18887611e-01
-1.05222023e+00 -6.26982689e-01 -9.60982859e-01 -1.73712596e-01
-2.51758546e-01 2.87845671e-01 6.92067027e-01 6.06188178e-01
-7.12728322e-01 4.24410909e-01 -8.16675842e-01 -2.17624933e-01
4.03330177e-01 2.08400324e-01 -3.08019280e-01 2.00572222e-01
-1.48519647e+00 8.42002094e-01 7.52803564e-01 2.60938346e-01
-1.49404511e-01 -6.77662253e-01 -1.16971719e+00 1.01215601e-01
8.04738224e-01 -8.07422101e-01 1.31976974e+00 -5.10752499e-01
-1.36609483e+00 5.69320202e-01 -4.17501837e-01 -5.90337574e-01
2.52948701e-01 -5.04008055e-01 -2.91710407e-01 -2.09439829e-01
1.33776337e-01 5.95766664e-01 2.91188836e-01 -1.08027112e+00
-3.86613071e-01 -1.57967016e-01 1.92224219e-01 3.29843462e-01
-5.58176398e-01 4.44871522e-02 -4.48161930e-01 -8.63564134e-01
9.23952758e-02 -9.48738694e-01 -3.56163502e-01 -8.08767498e-01
-9.28035021e-01 -3.96152526e-01 5.96669614e-01 -7.83639312e-01
1.54584098e+00 -1.86894393e+00 1.11781836e-01 2.53401157e-02
2.22937256e-01 5.20948589e-01 -1.72917485e-01 4.80255932e-01
-4.46671993e-03 4.29787815e-01 -3.38824302e-01 -5.67416430e-01
-9.81119499e-02 3.91970187e-01 -2.64693230e-01 -2.46092394e-01
6.39517426e-01 1.16051519e+00 -8.83637905e-01 -6.12302244e-01
-2.13896722e-01 3.83970439e-01 -3.89087409e-01 3.39484930e-01
-3.42482716e-01 4.13587898e-01 -5.37544549e-01 2.36786366e-01
5.03815770e-01 -3.36137027e-01 4.71913993e-01 -2.89390415e-01
2.85973221e-01 7.22434819e-01 -1.04716945e+00 1.82898581e+00
-7.06910610e-01 2.34342307e-01 -3.04789215e-01 -8.62642348e-01
1.12597668e+00 4.84343976e-01 1.13090865e-01 -6.26550972e-01
1.31469533e-01 6.27948344e-02 2.30238602e-01 -6.92734003e-01
7.44402289e-01 -4.73042428e-02 -1.76445961e-01 2.63492823e-01
1.53075501e-01 -4.63729016e-02 5.83644569e-01 4.98600394e-01
1.24824679e+00 3.68235260e-01 2.54004210e-01 8.45028087e-02
3.86994421e-01 -7.19511583e-02 7.45358169e-01 6.56317592e-01
3.94819081e-01 4.93034393e-01 4.64157164e-01 -1.09739423e-01
-1.00270939e+00 -4.84695375e-01 7.82324076e-02 9.18466866e-01
-1.87442638e-02 -8.83571446e-01 -6.70570552e-01 -1.07588291e+00
-3.43890667e-01 1.01233470e+00 -5.27537107e-01 -2.04828143e-01
-7.71893919e-01 -9.31592941e-01 7.28992581e-01 7.97904789e-01
6.16965711e-01 -9.96356130e-01 -2.10056901e-01 1.39245376e-01
-7.07152188e-01 -1.60801923e+00 -3.24959338e-01 2.30898678e-01
-6.18429184e-01 -8.14053416e-01 -3.66581172e-01 -5.50666392e-01
7.28304029e-01 2.08054576e-02 1.10414708e+00 -3.13140117e-02
2.25072965e-01 -2.61796892e-01 -7.23032713e-01 -3.65034848e-01
-4.91226643e-01 5.24349511e-01 -2.05444366e-01 -1.62829220e-01
3.40324640e-01 -5.43141425e-01 -6.28532842e-02 -4.32162359e-02
-8.75874877e-01 4.76707160e-01 7.98986077e-01 8.41678262e-01
4.06676501e-01 4.13823165e-02 9.25819993e-01 -1.51779640e+00
7.25133717e-01 -4.10929739e-01 -1.20263882e-01 4.54920858e-01
-5.17896056e-01 3.64531010e-01 8.07951152e-01 -3.52799654e-01
-1.55060077e+00 8.96980707e-03 -2.06464902e-01 -1.95705384e-01
-8.53464454e-02 8.22809517e-01 -5.61665773e-01 6.48871243e-01
5.57125807e-01 9.24450755e-02 -2.28818446e-01 -5.57402849e-01
6.24256313e-01 6.14336491e-01 4.60061967e-01 -6.96949720e-01
7.97968149e-01 -3.66675705e-02 -5.39408950e-03 -4.82696146e-01
-1.31083775e+00 -2.33931899e-01 -8.20318580e-01 3.03030610e-01
6.35684192e-01 -1.03466821e+00 -2.99657732e-01 6.84232637e-02
-1.35789478e+00 -2.11152673e-01 -5.39336860e-01 4.00791615e-01
-1.24912642e-01 3.15528482e-01 -5.68585813e-01 -6.94338679e-01
-5.97347617e-01 -7.48787999e-01 9.51354504e-01 2.20933273e-01
-6.01803005e-01 -8.59421551e-01 -5.13277277e-02 6.49443984e-01
2.59558558e-01 6.98586926e-02 8.31763685e-01 -1.14055562e+00
-3.32169563e-01 -8.39755312e-02 -1.79876432e-01 2.06495345e-01
5.06282985e-01 -2.30651155e-01 -9.71343875e-01 3.25736217e-02
-1.61143556e-01 -3.84464324e-01 5.95276415e-01 -8.64465386e-02
1.12218225e+00 -7.68734455e-01 -3.28199118e-01 3.28233182e-01
7.92980313e-01 1.97675437e-01 5.73733509e-01 2.04830155e-01
1.11090469e+00 8.42763901e-01 7.18724668e-01 4.17003870e-01
7.90181220e-01 7.81045139e-01 -1.05539106e-01 -1.62520990e-01
-3.95870835e-01 -4.29829568e-01 1.38002872e-01 1.16266751e+00
2.75159981e-02 -4.38631743e-01 -8.75593662e-01 5.79144895e-01
-1.89340687e+00 -7.94406593e-01 -3.09010208e-01 1.87879205e+00
1.56042218e+00 2.31027618e-01 -9.27364454e-02 1.82792082e-01
6.38959646e-01 -6.43727705e-02 -2.29789257e-01 -2.36508757e-01
-3.40814829e-01 4.41791087e-01 2.07468301e-01 4.13721681e-01
-1.10155296e+00 1.11734641e+00 5.14353275e+00 8.76407504e-01
-6.40566409e-01 1.50731534e-01 5.95834434e-01 -5.81242926e-02
-6.54005468e-01 2.20953524e-01 -1.08722353e+00 3.35083514e-01
1.01888013e+00 -1.95478588e-01 -1.72918379e-01 5.57908297e-01
1.48875713e-01 5.31809740e-02 -1.07283366e+00 6.17375672e-01
-1.85609423e-02 -1.21258390e+00 4.47984487e-02 -2.08641663e-01
6.00431263e-01 -3.76583070e-01 -3.05851460e-01 5.84103167e-01
5.43856204e-01 -8.74017775e-01 5.20325541e-01 4.19751167e-01
6.01647317e-01 -6.91761494e-01 1.12905896e+00 3.60094488e-01
-1.14459288e+00 3.41546446e-01 -7.29603693e-02 -6.07384071e-02
3.78803015e-01 7.48883784e-01 -1.18086183e+00 1.13366234e+00
4.92141366e-01 5.07362008e-01 -9.06835318e-01 4.91932303e-01
-5.50554693e-01 6.02912188e-01 -3.12241793e-01 -3.90655920e-02
-9.16507319e-02 5.32179289e-02 4.38046932e-01 1.37536132e+00
1.89001441e-01 2.19011754e-01 2.59273499e-01 7.57337987e-01
-3.91162395e-01 2.60268241e-01 -5.61822772e-01 -1.88624933e-01
6.91547394e-01 1.38901639e+00 -6.17818534e-01 -3.95907789e-01
-5.04344404e-01 8.74696076e-01 5.43116331e-01 1.65000081e-01
-6.99912846e-01 -6.10867560e-01 2.03869149e-01 -2.69151991e-04
2.11359218e-01 -1.24725461e-01 -5.33550560e-01 -1.29898524e+00
3.35732758e-01 -8.79672825e-01 4.04031992e-01 -6.70138836e-01
-1.06039619e+00 8.60600710e-01 2.35445842e-01 -9.31650043e-01
-4.69132364e-01 6.25586957e-02 -3.23957980e-01 9.41652834e-01
-1.28590357e+00 -1.46334910e+00 -2.70467043e-01 1.31395623e-01
5.94806612e-01 -1.46981496e-02 8.08947086e-01 4.02765632e-01
-1.01028800e+00 8.72255743e-01 -5.22501767e-01 3.90304267e-01
6.47734463e-01 -1.27319252e+00 5.77137291e-01 1.14416838e+00
3.34447503e-01 7.32342601e-01 5.87412715e-01 -9.35160339e-01
-1.11371684e+00 -1.28926730e+00 1.50158715e+00 -2.86205530e-01
5.47754288e-01 -4.97535348e-01 -1.24762797e+00 8.52607250e-01
4.55539942e-01 -1.70286834e-01 9.79825437e-01 5.68007350e-01
-3.19008559e-01 5.18440232e-02 -8.51117194e-01 6.41373634e-01
1.25521696e+00 -4.09849852e-01 -6.28854752e-01 2.31934041e-01
1.16819525e+00 -6.83294058e-01 -9.94729996e-01 7.06339598e-01
1.21584199e-01 -3.01153839e-01 6.74905658e-01 -6.93311930e-01
5.48242390e-01 -3.91772598e-01 -7.55876154e-02 -1.31055081e+00
-1.19178660e-01 -7.23056495e-01 -4.83950019e-01 2.03590059e+00
9.24508691e-01 -3.98962587e-01 5.79647481e-01 8.32600355e-01
-3.86491597e-01 -9.42233682e-01 -5.76548576e-01 -8.37824523e-01
-2.02987477e-01 -3.41787487e-01 8.12392533e-01 1.11815608e+00
2.46258557e-01 9.90675807e-01 -3.26716214e-01 1.74126878e-01
1.38486207e-01 -2.04988588e-02 7.70130038e-01 -8.45868051e-01
-3.87716860e-01 -1.21605657e-01 4.75041643e-02 -7.61726618e-01
4.75457788e-01 -1.15701103e+00 1.33596882e-02 -1.70017624e+00
1.73684046e-01 -7.34604836e-01 5.27563505e-02 1.06460190e+00
-7.08494365e-01 -5.65828085e-02 8.90545249e-02 4.10405919e-02
-5.35957217e-01 9.55647528e-01 1.07054532e+00 -4.24753428e-02
-3.83590192e-01 9.29162838e-03 -1.09586692e+00 5.07425904e-01
9.32489455e-01 -4.77438569e-01 -8.19616258e-01 -6.22862041e-01
4.17833596e-01 -7.56720304e-02 -8.37319717e-02 -6.99202836e-01
1.43853694e-01 3.01017966e-02 2.19735190e-01 -3.96246672e-01
2.03769535e-01 -5.87172389e-01 1.54801875e-01 2.22863518e-02
-6.88351095e-01 6.53898111e-03 4.00517851e-01 3.08706552e-01
-2.89877921e-01 -3.79781485e-01 3.13847482e-01 7.42704719e-02
-3.56238693e-01 -3.65374088e-02 -9.67116430e-02 3.36887300e-01
8.97591054e-01 1.90838799e-01 -4.73697186e-01 -2.95484841e-01
-6.42360032e-01 2.90926248e-01 2.65502427e-02 5.15719533e-01
4.59781021e-01 -1.27994049e+00 -8.73501658e-01 -1.22514032e-01
1.30985290e-01 5.63267231e-01 2.52449214e-01 7.15716243e-01
-1.34048257e-02 4.21710074e-01 2.84930557e-01 -1.34433374e-01
-1.34942162e+00 4.57666457e-01 -4.68162335e-02 -6.91405535e-01
-4.47596550e-01 9.50969815e-01 -2.06872765e-02 -6.39288604e-01
2.91131139e-02 -3.68089259e-01 -3.91472340e-01 1.67391405e-01
3.70595425e-01 5.61977625e-02 4.56403404e-01 -4.53603625e-01
-2.48227030e-01 -1.68844815e-02 -4.48194444e-01 3.07935115e-04
1.29063880e+00 -2.98561752e-01 -7.26014301e-02 3.95821512e-01
8.22631538e-01 2.25963354e-01 -8.20224464e-01 -6.87419176e-01
4.24671978e-01 -2.49451086e-01 -1.02310978e-01 -8.69854510e-01
-9.77627933e-01 5.33472836e-01 -1.74986124e-01 2.77241141e-01
1.12334681e+00 2.42431536e-01 9.89634454e-01 5.24466217e-01
4.68066111e-02 -7.82821596e-01 -8.66285115e-02 5.81375659e-01
7.84666240e-01 -1.21486139e+00 -5.04551195e-02 -9.96209741e-01
-9.46425617e-01 7.43535221e-01 9.73510027e-01 4.74279404e-01
2.75773674e-01 4.47521508e-01 -1.43977627e-01 -4.84677367e-02
-9.47569788e-01 -5.16566336e-01 3.78032744e-01 5.33309579e-01
8.90657425e-01 -1.23800993e-01 -4.44957435e-01 1.01780832e+00
-4.75493073e-01 -3.15681458e-01 4.89422470e-01 1.01750350e+00
9.18490887e-02 -1.60405743e+00 1.30488783e-01 4.37273145e-01
-5.97060084e-01 -5.49193025e-01 -4.77947444e-01 6.58404469e-01
8.81442130e-02 1.15167451e+00 -2.83404477e-02 -4.79597777e-01
4.69059199e-01 3.04863513e-01 4.79110688e-01 -9.93428826e-01
-6.35744214e-01 5.90081001e-03 8.30810964e-01 -2.29483292e-01
-6.25688910e-01 -6.73883796e-01 -1.40100443e+00 -4.08144295e-02
-5.70952475e-01 3.17193121e-01 3.55514735e-01 1.19186544e+00
6.65708125e-01 9.66342509e-01 5.02045095e-01 -2.02036902e-01
-2.17592254e-01 -1.46130693e+00 -5.81634380e-02 3.31753492e-01
-1.40384406e-01 -4.00121331e-01 1.16725937e-01 2.57724553e-01] | [9.461145401000977, 8.683223724365234] |
b4420966-9b3c-4baa-809b-b92cf7a015f5 | revisiting-data-free-knowledge-distillation | 2306.02368 | null | https://arxiv.org/abs/2306.02368v1 | https://arxiv.org/pdf/2306.02368v1.pdf | Revisiting Data-Free Knowledge Distillation with Poisoned Teachers | Data-free knowledge distillation (KD) helps transfer knowledge from a pre-trained model (known as the teacher model) to a smaller model (known as the student model) without access to the original training data used for training the teacher model. However, the security of the synthetic or out-of-distribution (OOD) data required in data-free KD is largely unknown and under-explored. In this work, we make the first effort to uncover the security risk of data-free KD w.r.t. untrusted pre-trained models. We then propose Anti-Backdoor Data-Free KD (ABD), the first plug-in defensive method for data-free KD methods to mitigate the chance of potential backdoors being transferred. We empirically evaluate the effectiveness of our proposed ABD in diminishing transferred backdoor knowledge while maintaining compatible downstream performances as the vanilla KD. We envision this work as a milestone for alarming and mitigating the potential backdoors in data-free KD. Codes are released at https://github.com/illidanlab/ABD. | ['Jiayu Zhou', 'Ruoxi Jia', 'Lingjuan Lyu', 'Shuyang Yu', 'Yi Zeng', 'Junyuan Hong'] | 2023-06-04 | null | null | null | null | ['backdoor-defense-for-data-free-distillation'] | ['adversarial'] | [-7.90544078e-02 5.16044259e-01 -2.92559683e-01 1.98162273e-01
-8.86593878e-01 -1.48472452e+00 5.49680352e-01 1.10013127e-01
-2.26009637e-01 7.24615872e-01 -2.57210821e-01 -1.12678742e+00
-1.98050458e-02 -9.22218621e-01 -1.29987288e+00 -6.42392933e-01
4.67022844e-02 2.64498964e-02 2.96856761e-01 -1.60986274e-01
7.54401833e-02 6.24500751e-01 -9.85810518e-01 9.37555358e-02
1.02245164e+00 6.39546156e-01 -8.40255097e-02 5.43479681e-01
2.46749237e-01 6.24436796e-01 -7.60153890e-01 -8.54774952e-01
7.73058355e-01 -1.37432292e-01 -7.95679688e-01 -7.57326841e-01
5.04101992e-01 -6.32057250e-01 -6.88013494e-01 1.10756373e+00
4.10454988e-01 -7.94074759e-02 4.82587457e-01 -1.44355273e+00
-7.72812068e-01 8.15704107e-01 -5.38870811e-01 2.83545017e-01
-2.89324611e-01 1.58028379e-01 7.46550560e-01 -8.68280232e-01
5.71461976e-01 9.08397257e-01 3.14150602e-01 7.32973158e-01
-1.15344644e+00 -1.24852180e+00 -1.72008961e-01 -2.05596201e-02
-1.55098045e+00 -5.70737839e-01 5.93966007e-01 -4.90835398e-01
5.06976545e-01 2.04880580e-01 2.17364609e-01 1.28383100e+00
-1.02583490e-01 8.71376216e-01 1.11639404e+00 -4.57052499e-01
2.24473864e-01 9.36785936e-01 3.12173218e-01 8.69464219e-01
6.02761269e-01 8.01738679e-01 -6.22429371e-01 -6.89339459e-01
3.44007432e-01 -1.87447205e-01 -3.28391969e-01 -6.10361218e-01
-2.68203825e-01 8.93005848e-01 3.82437050e-01 -1.21052504e-01
1.49806947e-01 1.62201319e-02 2.39985630e-01 5.66940486e-01
2.76411504e-01 5.31998694e-01 -8.78686786e-01 -3.32419313e-02
-8.67583811e-01 1.91317037e-01 7.43608773e-01 9.38559532e-01
9.22161043e-01 1.51023671e-01 1.04033552e-01 1.46442235e-01
2.04041854e-01 4.08094198e-01 2.51517981e-01 -4.70881581e-01
7.29729235e-01 2.78175324e-01 -7.56248040e-03 -6.96460187e-01
2.94517189e-01 -4.91707742e-01 -3.66148412e-01 1.26531512e-01
4.99232084e-01 -6.34623528e-01 -9.83477950e-01 1.82404673e+00
7.74596035e-01 4.31382805e-01 4.68551815e-01 5.38172424e-01
3.86883199e-01 7.06790030e-01 -2.37295181e-01 1.50991067e-01
1.22800422e+00 -7.96489000e-01 -1.56214952e-01 8.09616745e-02
1.10457230e+00 -3.94661784e-01 1.06574011e+00 5.32611907e-01
-7.90934801e-01 -1.57708228e-01 -1.22550213e+00 -2.06514467e-02
-6.83603287e-01 -1.62789762e-01 3.67764026e-01 1.42729783e+00
-7.76702821e-01 3.42972636e-01 -7.89676011e-01 2.45824724e-01
8.65708113e-01 3.19155335e-01 -4.77866232e-01 2.62075830e-02
-1.53734136e+00 7.19533920e-01 4.39529628e-01 -6.76044002e-02
-1.79048932e+00 -1.33583117e+00 -5.53163290e-01 -8.46320167e-02
8.50624800e-01 -1.92580655e-01 1.11696327e+00 -4.44947422e-01
-1.34849679e+00 6.05075181e-01 3.81236941e-01 -6.20073855e-01
5.89307189e-01 -5.38708270e-01 -1.53008789e-01 9.03148763e-03
-3.80174965e-01 3.73798460e-01 1.05859888e+00 -1.47034800e+00
-4.15320367e-01 -2.94794142e-01 1.54724061e-01 -2.10873663e-01
-6.75454974e-01 -1.72111392e-01 -1.30628422e-01 -6.84471786e-01
-6.08596742e-01 -1.13504601e+00 8.70366469e-02 -1.32176682e-01
-9.45329249e-01 1.53969064e-01 1.29334557e+00 -7.02364862e-01
1.19411981e+00 -2.29570389e+00 -1.72253564e-01 5.91320157e-01
3.67388725e-01 1.07002163e+00 -7.00725242e-02 5.39670348e-01
-6.41762987e-02 5.94433069e-01 -9.40382183e-02 -1.07772499e-01
-9.59594175e-02 2.63149858e-01 -1.07534039e+00 4.05603975e-01
1.78921625e-01 7.22305775e-01 -6.89878941e-01 -5.11338860e-02
-1.60030976e-01 3.21107805e-01 -6.84698224e-01 3.98649991e-01
-3.54555607e-01 5.39256394e-01 -6.46884501e-01 5.42675614e-01
1.03390467e+00 1.29102677e-01 8.01070482e-02 -2.52755284e-02
3.95463966e-02 3.05489630e-01 -8.70485902e-01 1.27566433e+00
-1.29340753e-01 4.20579791e-01 1.15368292e-01 -7.47872889e-01
7.58481383e-01 1.48542911e-01 -2.82670915e-01 -3.39539737e-01
1.89497575e-01 3.33596766e-02 2.69643515e-02 -2.03636497e-01
6.87174201e-01 -6.04054183e-02 -8.45853463e-02 5.07148921e-01
1.04849674e-01 1.17287531e-01 -5.20440757e-01 6.73528910e-01
1.04691970e+00 -1.38297558e-01 -9.15385261e-02 -2.55749494e-01
2.67163903e-01 9.34457872e-03 5.47803879e-01 8.66496980e-01
-2.38815606e-01 -2.06099004e-02 7.47692168e-01 8.35956857e-02
-8.47682655e-01 -1.35001445e+00 -5.77375665e-02 9.09027219e-01
1.83770895e-01 -6.50509655e-01 -8.27142358e-01 -1.32058215e+00
4.38356012e-01 1.01198113e+00 -5.08670628e-01 -7.38462925e-01
-2.22500667e-01 -1.30785942e-01 1.62188911e+00 2.76735872e-01
3.84955615e-01 -2.76547462e-01 -1.33521467e-01 -3.90908778e-01
2.46548623e-01 -8.22548449e-01 -3.65603089e-01 4.43007618e-01
-6.08221471e-01 -1.16063738e+00 -3.40110272e-01 -1.92046136e-01
9.06039417e-01 2.21176282e-01 2.33764231e-01 -5.32334633e-02
-2.12355971e-01 2.32405514e-01 -2.11042956e-01 -7.55995035e-01
-6.78157270e-01 5.51011786e-02 2.56336272e-01 -7.83560351e-02
3.13594013e-01 -1.91284940e-01 -5.62112391e-01 3.04661602e-01
-1.14730048e+00 -4.07336324e-01 3.14272583e-01 7.13640153e-01
3.53060871e-01 2.10638702e-01 2.40712315e-01 -1.17544305e+00
5.72455645e-01 -7.96844304e-01 -1.01229119e+00 2.21406817e-01
-6.95248663e-01 3.04683030e-01 8.42869520e-01 -7.98610270e-01
-1.02409208e+00 -2.61537999e-01 1.70159694e-02 -9.78581190e-01
7.71223754e-02 3.09588641e-01 -1.87778443e-01 -3.82287502e-01
8.49005044e-01 9.76683944e-02 -1.12038061e-01 -7.20872104e-01
7.76128113e-01 5.11505663e-01 3.68867427e-01 -9.24485028e-01
1.55074978e+00 3.96594852e-01 -1.01612248e-01 -6.50872409e-01
-8.08238089e-01 1.75599363e-02 -3.40140432e-01 1.90402061e-01
3.87578845e-01 -9.93911982e-01 -4.31337208e-01 7.23960698e-01
-7.90894151e-01 -5.93484998e-01 -2.43303537e-01 2.10642070e-01
1.59050420e-01 4.12114024e-01 -3.31952900e-01 -7.98325002e-01
-3.13581914e-01 -1.06843126e+00 4.24230695e-01 3.02716583e-01
1.36387393e-01 -9.63086784e-01 2.05466822e-01 7.08480299e-01
2.65600830e-01 1.09904939e-02 1.10656261e+00 -1.16465366e+00
-8.51624548e-01 -3.10563594e-01 5.30047268e-02 7.69114316e-01
-1.29958212e-01 9.31552276e-02 -1.38207269e+00 -5.79534888e-01
-9.71276611e-02 -7.77159691e-01 1.02976394e+00 -2.10954443e-01
1.18884504e+00 -5.54510593e-01 -3.09476346e-01 8.22257757e-01
1.15508771e+00 -5.94231710e-02 4.85248089e-01 8.91382620e-03
7.83117473e-01 4.52331126e-01 5.27270675e-01 3.97285283e-01
2.36423805e-01 3.69918615e-01 5.47876120e-01 3.07961907e-02
3.99356067e-01 -1.02827859e+00 6.91292524e-01 3.20273638e-01
5.60466707e-01 -3.63707155e-01 -9.52520490e-01 4.98806804e-01
-1.28809237e+00 -6.00242198e-01 1.01220352e-03 2.37125230e+00
1.13069367e+00 1.96276873e-01 -6.06076531e-02 -8.79952088e-02
5.53227425e-01 -7.13156313e-02 -6.99181378e-01 -6.19512141e-01
1.56348914e-01 4.24502820e-01 8.98724258e-01 5.45892239e-01
-9.70775127e-01 1.26546407e+00 5.19394016e+00 1.21725476e+00
-1.14126861e+00 2.90783703e-01 4.66637135e-01 -1.51612222e-01
-5.43034911e-01 5.22243738e-01 -1.17726207e+00 4.61810797e-01
1.14148664e+00 -2.43098751e-01 4.10739094e-01 1.02015734e+00
-2.46254832e-01 3.81814949e-02 -1.02328563e+00 3.39963198e-01
-2.07175493e-01 -1.42457652e+00 -4.11140062e-02 5.63666940e-01
5.99185824e-01 -2.34341826e-02 4.95809287e-01 6.23722672e-01
6.66537225e-01 -7.30803370e-01 7.64089882e-01 1.29117459e-01
8.81903231e-01 -8.66205513e-01 1.38928100e-01 6.93481684e-01
-6.58408582e-01 -1.83252454e-01 -4.49124664e-01 2.04661816e-01
-3.93753618e-01 5.71894944e-01 -1.16543078e+00 5.50775170e-01
4.35841233e-01 -1.52263433e-01 -4.07028705e-01 3.50806564e-01
-4.32529330e-01 1.29071009e+00 -5.08173287e-01 4.25257236e-01
3.28959703e-01 8.63358304e-02 6.99957013e-01 8.78012955e-01
2.90807426e-01 9.85975862e-02 -3.82591672e-02 1.07355022e+00
-4.93770957e-01 -3.48321468e-01 -9.16872621e-01 -5.20213306e-01
9.94946122e-01 9.54389453e-01 -1.11803360e-01 -9.87873152e-02
-1.29426286e-01 7.96054125e-01 4.97931480e-01 3.09146553e-01
-7.74426103e-01 -4.47569668e-01 9.48194027e-01 1.00414425e-01
4.29256409e-01 -1.55542091e-01 1.42828420e-01 -1.14413011e+00
-1.25112340e-01 -9.15897250e-01 6.56887710e-01 -3.41202408e-01
-1.28290188e+00 7.02454671e-02 3.36134881e-02 -8.54817152e-01
2.46150326e-02 -5.25739551e-01 -6.93493426e-01 9.50370014e-01
-1.66387260e+00 -1.08565080e+00 2.75484979e-01 8.57974887e-01
-2.15244457e-01 -3.76884162e-01 6.04315579e-01 1.54642209e-01
-7.61162639e-01 1.43987405e+00 6.47490680e-01 3.72197926e-01
7.99224079e-01 -9.61873531e-01 2.63760954e-01 1.00431514e+00
7.95877874e-02 9.89278615e-01 4.59132880e-01 -8.09304118e-01
-1.80882251e+00 -1.18158066e+00 3.03406000e-01 -7.17746437e-01
8.80503893e-01 -7.41728306e-01 -1.06946909e+00 7.82080293e-01
-2.48596191e-01 -5.58620691e-02 1.04217708e+00 -4.61780243e-02
-1.07811666e+00 -1.19272515e-01 -1.50750029e+00 5.34978867e-01
4.73326504e-01 -8.08403134e-01 -4.54327047e-01 -5.47889695e-02
1.13499153e+00 -4.37965035e-01 -7.03427255e-01 1.72891840e-02
4.12137449e-01 -6.48144901e-01 9.26930368e-01 -8.31952333e-01
9.38516557e-02 -1.17928609e-01 -1.33778840e-01 -1.25694788e+00
3.84046346e-01 -1.14189136e+00 -5.31941056e-01 1.54813445e+00
4.93338794e-01 -8.92995000e-01 9.79442120e-01 4.99134809e-01
5.23421317e-02 -4.07776922e-01 -9.39722836e-01 -1.02088988e+00
6.98383570e-01 -4.14132327e-01 5.07942736e-01 1.33196771e+00
-7.51455575e-02 -5.62154176e-03 -6.37839556e-01 8.35178137e-01
7.87370443e-01 -2.89167434e-01 1.17411137e+00 -8.17490876e-01
-4.88739252e-01 1.88321859e-01 -4.72104214e-02 -9.98725414e-01
-9.82095022e-03 -1.29557085e+00 -4.64949399e-01 -4.31024611e-01
2.28178520e-02 -7.33216166e-01 -4.74806905e-01 8.26416433e-01
-1.47218645e-01 -4.78683859e-02 3.87658447e-01 3.14269885e-02
4.62769940e-02 5.17524838e-01 1.08870959e+00 7.89112225e-02
-2.22809598e-01 5.43454364e-02 -1.24233258e+00 3.83751661e-01
1.03915429e+00 -1.03191805e+00 -8.61993849e-01 -1.12803191e-01
8.73147044e-03 -1.25355661e-01 5.55249512e-01 -3.78283173e-01
2.08200797e-01 -1.79449707e-01 -1.13376923e-01 -5.50511003e-01
1.54992014e-01 -8.22642148e-01 3.64996716e-02 5.33280909e-01
-3.53362232e-01 -5.35125494e-01 4.92821366e-01 7.91101754e-01
1.74448967e-01 -3.42713147e-01 7.42955446e-01 1.88271210e-01
-3.38408172e-01 2.96201140e-01 -1.45169213e-01 3.73443544e-01
1.23584092e+00 1.10320346e-02 -1.05991495e+00 -6.58887997e-02
-4.24556315e-01 3.92896384e-01 3.42131525e-01 3.03504765e-01
6.21118248e-01 -7.41520464e-01 -3.67510051e-01 6.10684454e-01
5.41665498e-03 3.65999751e-02 4.10605043e-01 5.16277194e-01
-3.61614615e-01 3.57714623e-01 -9.53566981e-04 1.04645325e-03
-1.41552687e+00 8.68860960e-01 2.67892778e-01 -3.43061954e-01
-4.64587152e-01 1.22258091e+00 3.77577305e-01 -8.27104986e-01
4.68814284e-01 6.57571852e-03 5.75697780e-01 -5.74579388e-02
3.51925880e-01 4.00725603e-01 -3.15028355e-02 -1.09333090e-01
-1.99441403e-01 -6.73035160e-02 -6.44628286e-01 -5.67731969e-02
9.04249966e-01 1.36769533e-01 1.90611467e-01 -3.33746187e-02
1.15946627e+00 5.03852904e-01 -1.15040839e+00 -4.48354691e-01
-1.21445432e-01 -6.14286065e-01 1.92383796e-01 -1.04850936e+00
-8.91505837e-01 1.18080389e+00 4.58748430e-01 1.75057575e-02
7.75777638e-01 9.66840796e-03 9.08165097e-01 6.38174593e-01
3.02938521e-01 -9.09901738e-01 1.01241991e-01 2.95626283e-01
2.85462826e-01 -8.15189242e-01 -6.37344345e-02 -2.52049774e-01
-7.23675907e-01 5.89452028e-01 8.01228464e-01 6.68876618e-02
1.07292938e+00 4.42052603e-01 1.00648671e-01 -7.53080323e-02
-9.16264236e-01 1.77688614e-01 -4.93155047e-02 7.16559529e-01
-3.72586757e-01 2.13501919e-02 3.05013955e-01 1.00686240e+00
-2.29387328e-01 -2.22394526e-01 6.41571581e-01 1.19920814e+00
-3.59771162e-01 -1.30194104e+00 -5.77264845e-01 3.19676638e-01
-7.27596283e-01 -1.75607726e-01 -8.15094233e-01 8.10265064e-01
1.81119442e-01 8.74474466e-01 -4.11929846e-01 -7.97212005e-01
1.44718587e-01 3.03725660e-01 1.64817572e-01 -4.45399880e-01
-7.65040159e-01 -2.73620576e-01 4.29296792e-02 -2.72117257e-01
5.94579637e-01 -1.25664547e-01 -1.04690886e+00 -7.57396400e-01
-7.63175309e-01 4.26257044e-01 5.51215768e-01 5.58578670e-01
6.94155633e-01 1.00714102e-01 8.56818676e-01 -1.35992572e-01
-1.13264644e+00 -3.72304261e-01 -6.62901342e-01 -9.21435375e-03
5.15335441e-01 -5.05934477e-01 -5.18349767e-01 -1.92765772e-01] | [5.823169231414795, 7.69917106628418] |
90286745-0124-484a-8a40-7c24de6fdd0b | artificial-intelligence-control-in-4d | null | null | https://doi.org/10.1109/ACCESS.2020.3026193 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9204607 | Artificial Intelligence Control in 4D Cylindrical Space for Industrial Robotic Applications | This article argues that an efficient artificial intelligence control algorithm needs the built-in symmetries of an industrial robot manipulator to be further characterized and exploited. The product of this enhancement is a four-dimensional (4D) discrete cylindrical grid space that can directly replace complex robot models. A* is chosen for its wide use among such algorithms to study the advantages and disadvantages of steering the robot manipulator within the 4D cylindrical discrete grid. The study shows that this approach makes it possible to control a robot without any specific knowledge of the robot kinematic and dynamic models at planning and execution time. In fact, the robot joint positions for each grid cell are pre-calculated and stored as knowledge, then quickly retrieved by the pathfinding algorithm when needed. The 4D cylindrical discrete space has both the advantages of the configuration space and the three-dimensional Cartesian workspace of the robot. Since path optimization is the core of any search algorithms, including A*, the 4D cylindrical grid provides for a search space that can embed further knowledge in form of cell properties, including the presence of obstacles and volumetric occupancy of the entire industrial robot body for obstacle avoidance applications. The main trade-off is between a limited capacity for pre-computed grid knowledge and the path search speed. This innovative approach encourages the use of search algorithms for industrial robotic applications, opens up to the study of other robot symmetries present in different robot models and lays a foundation for the application of dynamic obstacle avoidance algorithms. | ['Lihui Wang', 'Andrea de Giorgio'] | 2020-09-23 | null | null | null | null | ['industrial-robots'] | ['robots'] | [-5.72014898e-02 4.23351824e-01 -1.04320228e-01 3.62329632e-01
2.54981995e-01 -6.91714346e-01 4.55165803e-01 2.08397627e-01
-4.96724933e-01 7.72800922e-01 -7.60775030e-01 -4.31796163e-01
-1.02353251e+00 -9.67078745e-01 -3.15897644e-01 -9.13968980e-01
-3.02637577e-01 1.12845325e+00 8.93222615e-02 -3.88231367e-01
5.14467239e-01 1.21061361e+00 -1.53428555e+00 -4.76034641e-01
7.90884197e-01 7.85327196e-01 9.53121006e-01 1.60093874e-01
-6.95338994e-02 3.88950527e-01 -2.00544566e-01 3.81054372e-01
4.73966628e-01 -1.14400782e-01 -8.47042561e-01 2.03737602e-01
-7.24131584e-01 2.89186507e-01 6.96521476e-02 8.51559341e-01
6.95949867e-02 4.16821212e-01 8.54494691e-01 -1.23460567e+00
2.37505943e-01 3.02331239e-01 -1.68657348e-01 -3.07993591e-01
3.49090807e-02 -1.64756961e-02 3.14139932e-01 -6.96650565e-01
1.01797843e+00 9.43813682e-01 3.52015585e-01 -1.44613124e-02
-7.92895138e-01 -1.20669201e-01 -1.86734229e-01 4.48073864e-01
-1.50646913e+00 3.55953723e-01 7.90728569e-01 -4.16014820e-01
9.04379189e-01 3.15118879e-01 6.82749450e-01 4.18875724e-01
7.50196517e-01 -1.22653201e-01 7.69049585e-01 -6.80149317e-01
6.23060286e-01 1.59272313e-01 -2.33473238e-02 5.31118274e-01
6.33128107e-01 1.94498226e-01 2.01573014e-01 1.88675195e-01
1.24322045e+00 -9.13623199e-02 -1.17843919e-01 -1.14943612e+00
-1.14847910e+00 7.83727288e-01 3.88203681e-01 6.07574403e-01
-5.32259047e-01 -1.62180498e-01 5.06969750e-01 3.25090855e-01
-2.71750063e-01 1.12739742e+00 -6.85966253e-01 -1.68627694e-01
-8.17396715e-02 3.88738692e-01 1.09115088e+00 9.29581165e-01
8.79138649e-01 -1.28750764e-02 6.36111557e-01 4.89130288e-01
1.63899928e-01 2.14222461e-01 1.97445691e-01 -1.17199266e+00
3.18994164e-01 8.17980051e-01 1.59312487e-01 -1.02306938e+00
-7.61838615e-01 -3.82495880e-01 -8.34114313e-01 9.73193586e-01
5.46100914e-01 4.32088152e-02 -3.23583871e-01 1.00089598e+00
6.65808022e-01 -6.34647310e-01 2.65451819e-01 7.11334944e-01
-3.11584830e-01 6.29422545e-01 -4.32966709e-01 -5.06413102e-01
1.34089971e+00 -7.72642255e-01 -6.47748291e-01 2.33728603e-01
8.41138780e-01 -6.50751650e-01 5.37124991e-01 7.29509473e-01
-8.74162376e-01 -6.57617867e-01 -1.07933128e+00 2.85638094e-01
-6.41411126e-01 2.86280606e-02 6.38170779e-01 4.57529068e-01
-6.69232965e-01 8.02339315e-01 -6.34218454e-01 -3.90300751e-01
-3.24215144e-01 5.92134118e-01 -4.35782254e-01 -1.23784594e-01
-8.30573440e-01 1.56311619e+00 7.44735360e-01 3.87933671e-01
-1.26809299e-01 -3.33879560e-01 -6.11187816e-01 -5.27382456e-03
7.35686779e-01 -5.42875886e-01 8.77730846e-01 -3.70855957e-01
-1.70144570e+00 2.26439357e-01 1.48309425e-01 -3.28688249e-02
6.61788940e-01 2.92875946e-01 8.68726000e-02 1.59414813e-01
7.09542334e-02 1.27182573e-01 4.63268429e-01 -1.30005491e+00
-4.96273965e-01 -3.97785395e-01 -4.10680659e-03 3.04917425e-01
2.03973442e-01 -6.11818373e-01 -1.86539516e-01 -3.11477989e-01
4.57223177e-01 -1.19885457e+00 -7.63193130e-01 -3.14497799e-01
-1.82250261e-01 -4.18658346e-01 9.28934932e-01 -1.65057018e-01
6.42867386e-01 -2.02792978e+00 3.93318534e-01 9.04610634e-01
-3.45857561e-01 -1.59934759e-01 1.25055164e-01 9.05463278e-01
-2.58630246e-01 -1.71310350e-01 1.02432802e-01 4.69373465e-01
-1.43361047e-01 4.49859470e-01 3.73150669e-02 5.18297851e-01
-4.05103192e-02 3.10724407e-01 -6.30424142e-01 -3.87473166e-01
5.77361524e-01 2.80735880e-01 -4.05910671e-01 -4.20905411e-01
-1.20749481e-01 5.42249441e-01 -8.85875463e-01 2.17522874e-01
6.07567251e-01 3.97664040e-01 3.82553786e-01 -1.06885284e-01
-9.10408914e-01 -2.18176506e-02 -1.57281017e+00 1.44099188e+00
-6.56817436e-01 9.37256664e-02 3.40581000e-01 -1.05433095e+00
1.33749378e+00 3.44886780e-01 7.49836743e-01 -7.74050474e-01
3.40817779e-01 5.47469854e-01 1.88950915e-02 -3.41571242e-01
5.27717173e-01 1.19514085e-01 5.23806214e-02 2.82735471e-02
-4.46724027e-01 -8.23756874e-01 3.96383315e-01 -4.99290556e-01
8.30970526e-01 1.80441126e-01 3.70143503e-01 -7.59524763e-01
8.77297461e-01 6.68417275e-01 3.79598469e-01 4.17888999e-01
5.09676456e-01 3.02840453e-02 5.68557560e-01 -5.83221853e-01
-1.28782141e+00 -7.25757718e-01 -2.69754946e-01 2.70301908e-01
6.45439327e-01 1.97724272e-02 -4.94354457e-01 -1.08819433e-01
6.45446405e-02 7.75967479e-01 -3.28249007e-01 1.78225324e-01
-1.05671299e+00 -3.33725959e-01 -3.44803602e-01 3.04907337e-02
1.22379005e-01 -1.02636325e+00 -1.25612056e+00 4.55624074e-01
1.97093770e-01 -7.90859163e-01 3.92356843e-01 7.21761405e-01
-1.30661237e+00 -1.50109816e+00 -2.43281931e-01 -1.05261195e+00
8.20029795e-01 1.24611467e-01 4.54982787e-01 1.13420740e-01
-5.54574192e-01 2.51318276e-01 -4.69594419e-01 -4.20480430e-01
-5.45087278e-01 1.83305219e-01 6.49608225e-02 -7.22483695e-01
-3.38340342e-01 -6.21559978e-01 -2.36733511e-01 6.45709455e-01
-5.84726155e-01 -1.31259456e-01 6.97451472e-01 6.83121920e-01
6.25743330e-01 9.38821197e-01 6.56954288e-01 -2.83763081e-01
4.77510124e-01 -2.47775331e-01 -1.03126526e+00 -1.16152219e-01
-6.53559983e-01 9.98528898e-02 7.35748708e-01 -1.03915140e-01
-8.91090930e-01 4.47884887e-01 1.16293915e-01 -8.80804807e-02
-1.89538851e-01 5.54991663e-01 -3.56953174e-01 -6.25736862e-02
2.86401749e-01 3.53133865e-02 5.45455813e-01 -5.96763611e-01
2.04832599e-01 3.34951490e-01 2.90025055e-01 -5.01116276e-01
7.08136082e-01 2.61732340e-01 8.67230892e-01 -9.72437680e-01
4.43640649e-01 -4.95292962e-01 -1.10305202e+00 -1.97498769e-01
7.82948852e-01 -9.68825743e-02 -9.96239543e-01 1.85583923e-02
-1.33311462e+00 -2.09248707e-01 -6.05920494e-01 5.24699390e-01
-1.08380187e+00 4.70168501e-01 -1.96467087e-01 -1.04984665e+00
2.71003731e-02 -1.44409013e+00 4.35187519e-01 -2.40161404e-01
-3.36031467e-01 -7.27943242e-01 -1.35624960e-01 -2.15663731e-01
4.33535203e-02 3.96478802e-01 1.34679425e+00 -2.98045039e-01
-8.72613668e-01 -2.84567863e-01 2.54977286e-01 -1.95012808e-01
1.53445482e-01 -2.06817448e-01 -1.26459673e-01 -1.20006315e-01
4.19099361e-01 2.36370698e-01 2.03833044e-01 1.44783139e-01
7.48357117e-01 -1.52073577e-01 -7.27130353e-01 7.33455792e-02
1.66432369e+00 9.45301473e-01 7.67880797e-01 9.90675986e-01
1.26057595e-01 9.87422466e-01 1.23817432e+00 3.67863089e-01
-1.59665361e-01 1.07223630e+00 9.82271671e-01 7.76018500e-02
3.04209173e-01 4.80753422e-01 -4.69475724e-02 4.96002406e-01
-5.81540108e-01 1.17737919e-01 -8.44046474e-01 3.35211664e-01
-1.83578098e+00 -8.49165559e-01 -4.74170655e-01 2.15188241e+00
2.93358773e-01 -2.48877448e-03 -4.74076578e-03 8.54168892e-01
6.05754733e-01 -5.49498856e-01 -2.63368905e-01 -8.60832453e-01
2.91057169e-01 6.46449402e-02 8.08887899e-01 7.04398334e-01
-8.40827465e-01 3.36120427e-01 4.82356834e+00 7.33394325e-01
-8.97177160e-01 -3.64572614e-01 -9.02052224e-02 3.81002277e-01
2.32308403e-01 -3.84630337e-02 -5.91278374e-01 2.81401962e-01
5.17968893e-01 -3.18044447e-03 6.51919127e-01 1.25114167e+00
3.33369315e-01 -5.12616336e-01 -8.34809303e-01 6.91617012e-01
-4.09161061e-01 -1.11711359e+00 -1.44325703e-01 4.98295277e-01
4.35376048e-01 -2.91367203e-01 -1.74180970e-01 -7.50880167e-02
-1.27601147e-01 -8.23414266e-01 8.77185762e-01 2.32192650e-01
4.43191260e-01 -1.14717507e+00 9.67524171e-01 6.68109536e-01
-1.17958534e+00 -4.62159157e-01 -5.23692429e-01 -3.46195608e-01
4.32501197e-01 3.28861147e-01 -1.42591441e+00 1.19783199e+00
3.73178869e-01 6.80568144e-02 6.31978139e-02 1.16051567e+00
3.50364029e-01 -4.20323431e-01 -5.05128384e-01 -4.84272540e-01
4.41089422e-01 -7.73023725e-01 5.54999411e-01 6.37700200e-01
7.31294394e-01 -2.95969937e-02 1.03626989e-01 7.91511118e-01
9.77921605e-01 2.29573876e-01 -7.93158054e-01 3.62144381e-01
4.80752677e-01 8.81147981e-01 -1.22831833e+00 1.79599673e-01
-4.78117727e-04 5.44618368e-01 -2.60087311e-01 1.85909316e-01
-6.11437738e-01 -7.11990118e-01 4.55769241e-01 2.23712668e-01
4.30797994e-01 -8.04849505e-01 -6.07050538e-01 -2.09097132e-01
-1.43331245e-01 -4.62019414e-01 -9.95130017e-02 -5.97321808e-01
-7.56843567e-01 4.78746772e-01 2.88169771e-01 -1.25779116e+00
-7.53219843e-01 -1.25183439e+00 -3.29728007e-01 8.83565664e-01
-1.27728999e+00 -7.47883081e-01 -1.29866481e-01 4.71319199e-01
6.85684443e-01 -1.58626154e-01 1.04701066e+00 -3.09438288e-01
-9.76433009e-02 -3.62297297e-01 4.08277690e-01 -4.23375547e-01
-6.80016577e-02 -1.08037770e+00 -1.46045849e-01 7.71103576e-02
-5.58658898e-01 4.64312106e-01 8.60503912e-01 -7.38240659e-01
-1.65290368e+00 -7.79867768e-01 7.69394934e-01 -1.86280996e-01
5.20521343e-01 4.79402719e-03 -5.48983097e-01 3.34962875e-01
6.02355786e-02 -6.21366143e-01 -1.65039986e-01 2.31792908e-02
5.38666368e-01 -2.04211920e-01 -1.04951358e+00 6.64888799e-01
7.27918684e-01 2.19982088e-01 -4.08696532e-01 3.11501652e-01
3.16594005e-01 -3.04180592e-01 -9.49098170e-01 5.60186565e-01
4.24990594e-01 -7.49438345e-01 1.09597051e+00 1.27864301e-01
-1.26533449e-01 -4.35510874e-01 3.13813150e-01 -1.15125227e+00
-3.71180475e-01 -4.98187393e-01 6.44012749e-01 7.70117879e-01
3.32294434e-01 -9.17452574e-01 9.01998937e-01 3.15329015e-01
-4.95800048e-01 -9.25541222e-01 -1.33096588e+00 -1.10988092e+00
1.26006216e-01 -1.27674103e-01 3.93465221e-01 9.36938941e-01
4.85181868e-01 5.87541163e-02 3.45493436e-01 2.46029809e-01
2.78676391e-01 2.13304430e-01 6.42690122e-01 -1.56963813e+00
1.26701638e-01 -2.83230722e-01 -4.58111852e-01 -5.32359838e-01
5.96438684e-02 -7.03701377e-01 -3.00628524e-02 -1.53987682e+00
-6.37616813e-01 -1.20367742e+00 1.45933822e-01 1.86009973e-01
7.81944335e-01 -3.17210674e-01 3.26326311e-01 3.53325397e-01
8.31562206e-02 2.42055729e-01 1.66100442e+00 2.18118593e-01
-5.67754269e-01 1.91523612e-01 9.16846395e-02 6.56558514e-01
1.11151719e+00 -2.52777845e-01 -6.69910669e-01 -1.34894261e-02
3.05226445e-01 2.97581136e-01 1.86586678e-01 -1.11279476e+00
3.93448412e-01 -3.41046482e-01 1.85441196e-01 -8.24001431e-01
5.56435943e-01 -1.63151836e+00 6.44360542e-01 9.16144848e-01
3.67869765e-01 3.50172430e-01 1.02964886e-01 3.83796334e-01
-1.45006344e-01 -8.50947201e-01 5.24744928e-01 -4.73225296e-01
-7.25472391e-01 -4.17654753e-01 -7.57161260e-01 -8.85909855e-01
1.49768019e+00 -8.86650980e-01 1.28503010e-01 1.82701603e-01
-8.71507704e-01 -1.91217661e-02 6.81181192e-01 1.29167989e-01
2.20790699e-01 -8.13965857e-01 -6.38838038e-02 2.49010384e-01
-3.15055817e-01 4.60651457e-01 1.45943359e-01 7.55096674e-01
-1.10310364e+00 8.36175323e-01 -7.29371488e-01 -5.71429193e-01
-9.48250175e-01 9.93361235e-01 2.38573074e-01 -3.23267668e-01
-6.83440745e-01 1.48339465e-01 -7.16424808e-02 -3.29360992e-01
-1.35800987e-01 -6.40894413e-01 -4.30456787e-01 2.29881648e-02
-3.68901603e-02 8.10703397e-01 2.34953910e-01 -3.21939617e-01
-2.60717571e-01 8.12386036e-01 5.46510994e-01 -1.01201450e-02
1.31110954e+00 -2.76449263e-01 -5.24355233e-01 9.23839584e-02
7.16417670e-01 1.17493778e-01 -7.78655469e-01 6.29457116e-01
2.67415106e-01 -2.72439659e-01 -2.68702984e-01 -6.67998791e-01
-5.94441235e-01 2.73219854e-01 2.38005653e-01 4.47612315e-01
9.21174467e-01 -1.50834054e-01 1.68717168e-02 7.31177270e-01
9.77738202e-01 -1.16119230e+00 -1.14239916e-01 8.72308314e-01
1.22994876e+00 -4.17121381e-01 8.36079847e-03 -1.30418134e+00
-3.71296257e-01 1.71232259e+00 3.25589567e-01 -1.90042064e-01
4.14025575e-01 4.61715341e-01 -3.08014397e-02 -3.42442989e-02
-1.34194463e-01 -7.53917545e-03 -2.67789185e-01 8.89057696e-01
-1.92466468e-01 -2.25228935e-01 -6.97466135e-01 4.25149471e-01
-3.48200023e-01 2.29935031e-02 6.04378521e-01 1.15313959e+00
-6.87314808e-01 -1.43709624e+00 -7.62722790e-01 -1.72203988e-01
8.84607583e-02 6.18749678e-01 1.83353703e-02 1.66874349e+00
5.49408197e-01 7.74711788e-01 1.68006733e-01 4.82581258e-02
6.19989693e-01 1.79580189e-02 6.63587451e-01 -3.87873739e-01
-3.41043621e-01 -9.21739638e-02 1.69783771e-01 -4.48525667e-01
4.13815603e-02 -6.72641635e-01 -1.64660776e+00 5.34993708e-02
-4.87732261e-01 7.10020661e-01 1.28370392e+00 6.78769290e-01
2.33259514e-01 4.75438535e-01 4.28364247e-01 -1.27580869e+00
-6.02587759e-01 -6.55127883e-01 -9.04185951e-01 -1.78942993e-01
-3.42934281e-02 -1.14170861e+00 6.88456203e-05 -1.38301656e-01] | [4.949461460113525, 1.4949122667312622] |
29b166a8-f073-42e5-8c6c-7f61627f8a85 | knowing-when-to-quit-selective-cascaded | 2108.00377 | null | https://arxiv.org/abs/2108.00377v2 | https://arxiv.org/pdf/2108.00377v2.pdf | Knowing When to Quit: Selective Cascaded Regression with Patch Attention for Real-Time Face Alignment | Facial landmarks (FLM) estimation is a critical component in many face-related applications. In this work, we aim to optimize for both accuracy and speed and explore the trade-off between them. Our key observation is that not all faces are created equal. Frontal faces with neutral expressions converge faster than faces with extreme poses or expressions. To differentiate among samples, we train our model to predict the regression error after each iteration. If the current iteration is accurate enough, we stop iterating, saving redundant iterations while keeping the accuracy in check. We also observe that as neighboring patches overlap, we can infer all facial landmarks (FLMs) with only a small number of patches without a major accuracy sacrifice. Architecturally, we offer a multi-scale, patch-based, lightweight feature extractor with a fine-grained local patch attention module, which computes a patch weighting according to the information in the patch itself and enhances the expressive power of the patch features. We analyze the patch attention data to infer where the model is attending when regressing facial landmarks and compare it to face attention in humans. Our model runs in real-time on a mobile device GPU, with 95 Mega Multiply-Add (MMA) operations, outperforming all state-of-the-art methods under 1000 MMA, with a normalized mean error of 8.16 on the 300W challenging dataset. | ['Roy J. Jevnisek', 'Ishay Goldin', 'Noga Levy', 'Gil Shapira'] | 2021-08-01 | null | null | null | null | ['face-alignment'] | ['computer-vision'] | [ 1.47366181e-01 1.13769092e-01 -1.07018553e-01 -4.60660875e-01
-7.42924869e-01 -2.85602242e-01 4.97841060e-01 -5.11821508e-02
-3.76774609e-01 2.00701058e-01 1.62406817e-01 1.15553908e-01
2.76933342e-01 -6.50818884e-01 -6.49234176e-01 -6.03210211e-01
-1.26089171e-01 2.63576180e-01 -1.67241111e-01 -1.18453950e-01
2.86770791e-01 7.56775558e-01 -1.89481044e+00 2.83451289e-01
4.77807641e-01 1.40550745e+00 -3.38090181e-01 5.38888276e-01
2.30667055e-01 4.19307053e-01 -3.84522855e-01 -6.76231980e-01
3.74409229e-01 -5.20821400e-02 -6.08527482e-01 6.69309571e-02
1.12215340e+00 -3.33723992e-01 4.53998931e-02 9.19016838e-01
7.11013436e-01 4.23427448e-02 5.44542313e-01 -1.20013499e+00
-2.70822734e-01 1.56511031e-02 -1.08164275e+00 1.98429808e-01
5.04978001e-01 2.05118299e-01 1.04236257e+00 -1.48499870e+00
5.68042934e-01 1.35675263e+00 9.23421025e-01 5.28201878e-01
-1.39176428e+00 -8.14056873e-01 1.93485841e-01 8.57845172e-02
-1.80639410e+00 -1.05295336e+00 6.29420042e-01 -3.19530308e-01
9.83690679e-01 3.23398501e-01 7.05117285e-01 7.16113985e-01
2.76001036e-01 4.48433399e-01 7.59974778e-01 -1.18969880e-01
1.22991718e-01 -9.24428254e-02 -2.94995785e-01 1.04230917e+00
-2.09499709e-02 -5.80577366e-02 -8.38302195e-01 -2.81855315e-01
6.75153196e-01 6.20413460e-02 -1.95694238e-01 3.57711427e-02
-7.24149883e-01 7.28703737e-01 5.92912555e-01 -5.10417449e-04
-6.03160381e-01 3.00714970e-01 6.72671720e-02 2.12427199e-01
6.79441571e-01 1.46192983e-01 -4.42687869e-01 -2.59877797e-02
-1.23357785e+00 1.87920362e-01 5.68723917e-01 5.74212611e-01
1.23787463e+00 -1.63717642e-01 -2.12199181e-01 8.43110502e-01
2.44201571e-01 4.98021126e-01 1.64048463e-01 -9.67809975e-01
-3.35799269e-02 5.88116527e-01 -2.07371995e-01 -1.32592726e+00
-5.51778674e-01 -2.59712607e-01 -7.14038014e-01 3.87709141e-01
4.61919218e-01 -1.06719844e-01 -7.81234145e-01 1.76430023e+00
6.89420581e-01 3.75723243e-01 -6.87080324e-01 8.45009506e-01
6.31318867e-01 4.15190399e-01 1.82242528e-01 -1.46673039e-01
1.67733121e+00 -7.61182010e-01 -2.82085419e-01 -3.16093057e-01
4.25284654e-01 -8.98376703e-01 1.11517775e+00 3.46647531e-01
-1.13573956e+00 -7.04790533e-01 -8.98191094e-01 -3.09892476e-01
-1.02283135e-01 3.23498756e-01 5.77344775e-01 6.66027904e-01
-1.27149391e+00 8.03893924e-01 -8.54377568e-01 -2.80519813e-01
9.03276324e-01 7.81448424e-01 -6.25910461e-01 3.77675705e-02
-5.20814180e-01 7.39362776e-01 -3.51293921e-01 1.95283800e-01
-6.44994080e-01 -1.26212811e+00 -8.53561878e-01 2.12834030e-01
2.82874584e-01 -4.76981640e-01 8.84082496e-01 -1.33846247e+00
-1.48845911e+00 1.02010036e+00 -7.01474071e-01 -2.93214936e-02
4.12341207e-01 -2.09280401e-01 -2.10883886e-01 1.10732146e-01
1.34797215e-01 1.08049703e+00 1.27530897e+00 -8.98065627e-01
-5.40738106e-01 -6.83923841e-01 -1.37007609e-01 2.41312776e-02
-3.65115434e-01 -8.48589651e-03 -6.91321671e-01 -4.51114625e-01
-3.50938812e-02 -9.77804542e-01 5.33843674e-02 5.92035711e-01
-1.26401177e-02 -3.73910666e-01 6.14791811e-01 -5.57221830e-01
9.83692646e-01 -2.39178014e+00 1.80361178e-02 4.85150516e-01
4.73707289e-01 1.67128608e-01 -3.14287364e-01 -4.56415266e-02
-3.48342210e-02 -6.83359429e-02 -6.05496839e-02 -7.24605143e-01
-1.18583582e-01 -7.26308003e-02 -1.30376965e-01 8.32520068e-01
6.64424419e-01 9.20372427e-01 -5.77351749e-01 -4.73554701e-01
-6.34350022e-03 1.00279760e+00 -1.14877808e+00 -9.61713493e-02
8.51721987e-02 3.41462791e-01 -9.54870805e-02 9.04648125e-01
8.41974735e-01 -1.38694599e-01 1.21456437e-01 -4.55948293e-01
6.38194680e-02 -5.54509135e-03 -1.23311794e+00 1.53507102e+00
-6.94616556e-01 8.12132239e-01 1.85575157e-01 -4.73054409e-01
7.35952079e-01 8.32207054e-02 4.60409909e-01 -7.58898020e-01
2.45063871e-01 1.55105621e-01 -4.08185907e-02 -1.59272492e-01
3.56671780e-01 6.08210564e-02 4.78752643e-01 3.84485334e-01
2.20926404e-01 2.23375574e-01 -8.15562531e-02 -1.45729348e-01
9.64590430e-01 9.07641277e-02 2.87085205e-01 -5.01834452e-01
5.02441168e-01 -5.05335093e-01 3.51586848e-01 1.58887118e-01
-1.64749861e-01 6.39023423e-01 7.09820807e-01 -7.33170211e-01
-6.82328641e-01 -6.83012187e-01 -2.28275284e-01 1.59473515e+00
-1.20742112e-01 -7.67514050e-01 -1.00051033e+00 -6.33172750e-01
1.48603782e-01 1.58303797e-01 -1.07243335e+00 -5.31594530e-02
-5.87872565e-01 -6.02528214e-01 2.78020054e-01 5.38937032e-01
2.76475251e-01 -1.00363171e+00 -7.62970507e-01 -1.68302193e-01
1.28791779e-01 -8.88656557e-01 -8.09335113e-01 -2.62925953e-01
-4.95551378e-01 -1.11277807e+00 -4.80407715e-01 -5.42628586e-01
1.10410917e+00 1.50144249e-01 1.21147966e+00 4.91212249e-01
-5.51481843e-01 2.71821558e-01 1.29787520e-01 -2.69458622e-01
3.31781298e-01 2.39392370e-02 -2.70647779e-02 4.78650063e-01
4.25700903e-01 -5.80057561e-01 -9.25087452e-01 3.55596304e-01
-4.00394738e-01 -1.62289500e-01 5.77289283e-01 8.38720620e-01
7.25421071e-01 -3.25868130e-01 2.26562917e-01 -7.97094107e-01
2.52638608e-01 -4.51280624e-01 -5.57347775e-01 -1.42179048e-02
-3.27277839e-01 -1.06754815e-02 5.49216330e-01 -5.76096714e-01
-6.54486477e-01 2.91237056e-01 -3.43525410e-01 -4.93117243e-01
5.80917336e-02 1.54044032e-01 -1.82518303e-01 -4.53833699e-01
6.98250055e-01 -8.24520960e-02 1.88759595e-01 -2.08221585e-01
3.61842513e-01 2.61821628e-01 4.31025594e-01 -6.65206075e-01
5.86872280e-01 6.16634250e-01 2.02945113e-01 -1.03956521e+00
-6.81863725e-01 -2.33854175e-01 -6.11670196e-01 -4.05038863e-01
5.04985511e-01 -8.41673553e-01 -1.34562147e+00 1.59470275e-01
-9.70552385e-01 -3.80510837e-01 -9.89326015e-02 -2.43745488e-03
-3.52754503e-01 -5.64219384e-03 -2.68333405e-01 -7.05427945e-01
-4.19021040e-01 -1.19995773e+00 1.57931960e+00 1.54358700e-01
-5.33962965e-01 -6.39126658e-01 -1.72188982e-01 1.14249609e-01
4.66330469e-01 2.37661526e-01 5.84053516e-01 -2.69363493e-01
-3.16873550e-01 -1.88745990e-01 -3.83137137e-01 -6.17743172e-02
1.44033059e-02 1.36442155e-01 -1.37767947e+00 -4.13931668e-01
-1.60983264e-01 -1.79339796e-01 6.83393598e-01 2.76132166e-01
1.25806928e+00 -4.77774233e-01 -2.05416441e-01 9.78966594e-01
1.26679015e+00 -3.43058765e-01 6.12315953e-01 -1.88464060e-01
7.21059024e-01 6.63345635e-01 3.99748474e-01 5.57261944e-01
2.30194807e-01 7.66792417e-01 4.51637447e-01 -1.86116844e-01
-2.38526478e-01 -2.04053119e-01 3.37690830e-01 1.20563850e-01
-4.06110257e-01 3.26643467e-01 -8.11890006e-01 3.77666742e-01
-1.46464622e+00 -8.09793890e-01 2.14870960e-01 2.27855706e+00
8.07586432e-01 -1.53130993e-01 2.10742936e-01 8.93168524e-02
3.42687935e-01 1.62777096e-01 -4.01252598e-01 -2.88680971e-01
2.13645980e-01 8.63707423e-01 2.19323710e-01 8.27969432e-01
-1.02091944e+00 1.11407995e+00 5.95211840e+00 9.15198207e-01
-1.58655488e+00 5.76559305e-02 1.09826314e+00 -4.38892841e-01
-7.25767314e-02 -2.26342887e-01 -7.90375769e-01 3.60862106e-01
8.15477788e-01 1.49646252e-01 6.95391297e-01 1.09818387e+00
-1.03244536e-01 -1.69918954e-01 -1.33897328e+00 1.17628098e+00
2.89656252e-01 -1.17642832e+00 -6.73588440e-02 2.30153710e-01
5.64078689e-01 -1.07494593e-01 3.70503813e-01 5.26648052e-02
-3.58693123e-01 -1.49349523e+00 7.44966269e-01 4.04757202e-01
9.98782933e-01 -9.37755108e-01 5.37419856e-01 -4.54802774e-02
-1.27952945e+00 1.12892710e-01 -4.11920786e-01 -1.20567158e-01
-3.21135312e-01 4.31878418e-01 -6.96184576e-01 1.07366359e-02
7.70619154e-01 5.59818149e-01 -7.36468673e-01 5.41396797e-01
-6.55509233e-02 2.83662260e-01 -6.44439757e-01 1.15359560e-01
9.40449536e-03 -7.96816275e-02 2.05313295e-01 1.18855548e+00
3.90668362e-01 1.66293994e-01 -7.19254315e-02 7.52306879e-01
-2.85713464e-01 3.62131178e-01 -2.36044958e-01 4.28349763e-01
4.04640108e-01 1.61155736e+00 -7.53781140e-01 -1.16572089e-01
-4.63997066e-01 1.14042854e+00 5.37275851e-01 8.19223747e-02
-8.24160814e-01 -1.20391957e-01 1.16670775e+00 5.50355017e-01
4.98961419e-01 -7.16900826e-02 -3.44880253e-01 -8.54324698e-01
1.37183234e-01 -8.84548187e-01 1.29353270e-01 -3.79997164e-01
-1.02434909e+00 7.94665456e-01 -4.68745530e-01 -8.32976162e-01
-9.58688110e-02 -6.43679261e-01 -5.72610557e-01 8.15669537e-01
-1.59663928e+00 -1.31919897e+00 -5.43745935e-01 7.43152678e-01
3.25817645e-01 4.42732833e-02 8.61633182e-01 4.20950890e-01
-6.65230870e-01 1.21460032e+00 -5.23746371e-01 2.51613557e-01
7.39100754e-01 -7.81208336e-01 4.86941695e-01 5.86996973e-01
4.79209989e-01 8.41472328e-01 5.20176709e-01 -3.72374952e-01
-1.62131166e+00 -1.05634046e+00 9.34221387e-01 -4.27955478e-01
3.27262044e-01 -5.07376611e-01 -7.75204480e-01 5.98063052e-01
-1.29395902e-01 7.77264118e-01 6.39602482e-01 3.83919507e-01
-7.05076396e-01 -3.56148720e-01 -1.25044167e+00 6.95787489e-01
1.06845689e+00 -5.57012796e-01 1.49262816e-01 1.07692562e-01
2.04678718e-02 -4.77589607e-01 -8.36536646e-01 2.60690004e-01
9.08751011e-01 -1.18365574e+00 9.99943972e-01 -3.02400529e-01
3.51946265e-01 -2.42850184e-01 -1.17647968e-01 -9.73694384e-01
-4.35440123e-01 -7.80804455e-01 -1.35233209e-01 1.08619916e+00
3.12994570e-01 -5.99216282e-01 9.22045827e-01 6.79750502e-01
2.91238695e-01 -1.22821045e+00 -1.06843352e+00 -1.79358482e-01
-2.75419116e-01 -4.26594645e-01 7.77180851e-01 7.30590820e-01
-8.81525278e-02 2.71270216e-01 -2.62227148e-01 1.45159572e-01
3.95436138e-01 -6.27795160e-02 9.23680544e-01 -1.20192397e+00
-2.53114142e-02 -6.73038006e-01 -6.23516977e-01 -9.18037236e-01
3.47398400e-01 -7.43193030e-01 -1.95823520e-01 -7.56318331e-01
1.09863497e-01 -3.95530999e-01 -4.48174477e-02 9.26423967e-01
-3.87127280e-01 9.98490870e-01 2.14102119e-01 9.45404451e-03
-3.67614299e-01 3.20905834e-01 9.15535867e-01 4.53417674e-02
-1.32982776e-01 -1.68645471e-01 -8.18709075e-01 8.46036673e-01
4.30883557e-01 -2.88803667e-01 -1.14047833e-01 -4.11784470e-01
1.82418525e-01 -3.56240690e-01 3.90846103e-01 -9.85973060e-01
2.55005211e-01 6.65798560e-02 7.85483718e-01 -2.10060969e-01
6.58899128e-01 -7.06165254e-01 4.58860993e-02 2.59165138e-01
-7.90608004e-02 5.03830984e-02 3.29334229e-01 1.82988018e-01
-4.85686846e-02 2.99512088e-01 1.05081367e+00 2.17402175e-01
-5.18539786e-01 6.80315316e-01 1.12303130e-01 -4.40526865e-02
9.83053327e-01 -2.31604710e-01 2.17512883e-02 -3.01109821e-01
-6.16939962e-01 -1.53778553e-01 5.28913498e-01 2.68293023e-01
5.10932088e-01 -1.43540180e+00 -7.98356712e-01 7.77735114e-01
2.79263835e-02 -1.24372050e-01 3.07225615e-01 1.04739225e+00
-5.07579625e-01 4.18584011e-02 -2.28382364e-01 -7.55339503e-01
-1.40982938e+00 3.72967422e-01 3.30104172e-01 1.40503228e-01
-4.95648235e-01 1.09417236e+00 2.49458209e-01 3.55564356e-02
1.60684466e-01 -4.49307077e-02 -5.47072627e-02 4.22652185e-01
9.95308280e-01 3.81885588e-01 3.74416053e-01 -9.64068890e-01
-5.99002004e-01 1.07606399e+00 1.81717109e-02 7.51950815e-02
1.30958760e+00 2.61758447e-01 -3.15888703e-01 -7.53806606e-02
1.51639307e+00 4.23758626e-01 -1.50776923e+00 -6.68751001e-02
-2.76469857e-01 -7.76746094e-01 1.62531540e-01 -4.56177980e-01
-1.48358750e+00 8.49413276e-01 6.12219572e-01 -2.55385160e-01
1.27757156e+00 -6.67725801e-02 5.47335088e-01 2.11247406e-03
2.70362675e-01 -8.87229264e-01 6.06500097e-02 1.70226887e-01
1.11239719e+00 -1.32665968e+00 2.75959313e-01 -4.84243423e-01
-5.19643545e-01 1.06449902e+00 6.39830768e-01 -3.79000962e-01
8.01726878e-01 4.52147126e-01 -6.17636591e-02 -2.01284811e-01
-7.14073062e-01 -1.44213825e-01 5.65645516e-01 4.54220146e-01
5.40513396e-01 -9.89958048e-02 1.31426692e-01 4.23222035e-01
-5.20079553e-01 -1.90941900e-01 -2.24765733e-01 5.29871464e-01
-3.04054439e-01 -8.20041895e-01 -3.41717750e-01 3.47679973e-01
-5.48051596e-01 -2.04627886e-01 -2.41177827e-01 6.13200605e-01
2.55232602e-01 7.41977334e-01 5.50713301e-01 -3.01280439e-01
2.19175190e-01 3.61415301e-03 6.29148841e-01 -3.45767468e-01
-6.86350763e-01 7.04526305e-02 -2.26073831e-01 -1.18541932e+00
-1.52128652e-01 -7.35172391e-01 -9.42331493e-01 -7.45018005e-01
-1.02667652e-01 -2.86109775e-01 5.53210557e-01 6.60766065e-01
6.97734773e-01 1.25632286e-01 7.25186050e-01 -1.37592542e+00
-1.26625508e-01 -8.53303909e-01 -1.67402342e-01 2.60241002e-01
4.43889946e-01 -6.69229925e-01 -2.21966743e-01 -1.47869373e-02] | [13.460769653320312, 0.3886638283729553] |
a448443b-f787-4805-a2ff-0ace11619b8b | coronary-artery-semantic-labeling-using-edge | 2305.12327 | null | https://arxiv.org/abs/2305.12327v1 | https://arxiv.org/pdf/2305.12327v1.pdf | Coronary Artery Semantic Labeling using Edge Attention Graph Matching Network | Coronary artery disease (CAD) is one of the primary causes leading deaths worldwide. The presence of atherosclerotic lesions in coronary arteries is the underlying pathophysiological basis of CAD, and accurate extraction of individual arterial branches using invasive coronary angiography (ICA) is crucial for stenosis detection and CAD diagnosis. We propose an innovative approach called the Edge Attention Graph Matching Network (EAGMN) for coronary artery semantic labeling. By converting the coronary artery semantic segmentation task into a graph node similarity comparison task, identifying the node-to-node correspondence would assign semantic labels for each arterial branch. More specifically, The EAGMN utilizes the association graph constructed from the two individual graphs as input. Experimental results indicate the EAGMN achieved a weighted accuracy of 0.8653, a weighted precision of 0.8656, a weighted recall of 0.8653 and a weighted F1-score of 0.8643. Furthermore, we employ ZORRO to provide interpretability and explainability of the graph matching for artery semantic labeling. These findings highlight the potential of the EAGMN for accurate and efficient coronary artery semantic labeling using ICAs. By leveraging the inherent characteristics of ICAs and incorporating graph matching techniques, our proposed model provides a promising solution for improving CAD diagnosis and treatment | ['Weihua Zhou', 'Guang-Uei Hung', 'Zhihui Xu', 'Chen Zhao'] | 2023-05-21 | null | null | null | null | ['graph-matching'] | ['graphs'] | [-7.56152943e-02 4.29592520e-01 -6.13785803e-01 -5.07244587e-01
-5.28496087e-01 -5.11243105e-01 -1.93994552e-01 2.14573145e-01
1.94537804e-01 3.25572997e-01 -2.31174175e-02 -8.49310219e-01
-3.63337368e-01 -9.48613346e-01 -1.04195019e-03 -2.37671599e-01
-9.24781412e-02 5.32672703e-01 2.25057676e-01 2.25507542e-01
1.33148968e-01 6.33666217e-01 -8.94119084e-01 -5.91200730e-03
1.09070373e+00 1.16952991e+00 3.73746119e-02 5.32420218e-01
-4.52222586e-01 5.69788456e-01 -2.97933370e-01 -5.85847616e-01
1.84349000e-01 -7.77223408e-01 -7.70404220e-01 2.50260025e-01
1.19465977e-01 1.60598561e-01 -2.72371501e-01 1.23641002e+00
4.89471823e-01 -1.11388169e-01 5.60316563e-01 -1.22730505e+00
-7.05508709e-01 5.04278004e-01 -6.51971698e-01 7.04461396e-01
-1.79843590e-01 -1.15828810e-03 1.14384091e+00 -5.97727954e-01
4.84404176e-01 1.29888844e+00 5.01602530e-01 2.62109101e-01
-9.79512513e-01 -6.27066314e-01 1.14463784e-01 2.88436651e-01
-1.35412002e+00 -8.05799365e-02 9.49085653e-01 -3.84385943e-01
3.11677605e-01 2.13387072e-01 7.20176876e-01 2.96975613e-01
6.97223842e-02 4.16112393e-01 7.64151514e-01 -3.38081509e-01
2.09550068e-01 -3.19273248e-02 5.32885432e-01 1.04706156e+00
6.06221855e-01 -1.04391910e-01 1.23159446e-01 -1.70163020e-01
1.15405142e+00 1.88173175e-01 1.22504756e-02 -3.06135714e-01
-8.89744043e-01 8.39411318e-01 1.09024739e+00 3.04771483e-01
-3.96949172e-01 1.83673173e-01 5.99790692e-01 1.30898934e-02
2.12616563e-01 5.00372171e-01 -1.47255272e-01 4.65358168e-01
-4.82602030e-01 -2.74230659e-01 4.38363552e-01 7.08259523e-01
4.97463316e-01 3.39880213e-02 -2.41714660e-02 5.72483480e-01
5.00841796e-01 4.77004707e-01 4.04480278e-01 -1.16156745e+00
3.72253865e-01 1.16439509e+00 -4.85692203e-01 -1.60337913e+00
-5.30428410e-01 -8.72355163e-01 -1.01808512e+00 2.71420442e-02
5.14529347e-01 4.45076451e-02 -7.63260126e-01 1.50648725e+00
3.69619071e-01 6.19550765e-01 -1.00750990e-01 1.15977705e+00
9.91696000e-01 -4.78252023e-02 8.44152272e-01 7.03773350e-02
1.93736780e+00 -9.22196746e-01 -7.20935047e-01 -3.74865621e-01
9.18560684e-01 -4.83327836e-01 8.16855192e-01 -4.29057986e-01
-6.71756625e-01 -6.64135754e-01 -8.82055640e-01 1.54641420e-01
-4.30748053e-02 3.33699256e-01 6.89236283e-01 7.38289893e-01
-7.60002196e-01 2.18930468e-01 -7.64717579e-01 -1.06072232e-01
9.95746374e-01 1.81516543e-01 -3.04198802e-01 -7.30682388e-02
-1.13269067e+00 6.92335963e-01 3.30725372e-01 3.14704962e-02
-1.56607732e-01 -8.66115272e-01 -9.05698061e-01 1.59863546e-01
3.56971115e-01 -1.13817275e+00 8.03494811e-01 -8.75474215e-01
-9.74020779e-01 1.11589420e+00 -3.08556765e-01 -6.00791156e-01
2.40854487e-01 1.01549976e-01 -6.61526263e-01 7.85682857e-01
4.38046962e-01 4.73766208e-01 2.80681968e-01 -9.12000418e-01
-6.21022165e-01 -6.37469769e-01 -2.66599536e-01 -5.37166968e-02
-2.14965586e-02 -9.65603441e-02 -2.62015760e-01 -8.60767782e-01
6.52955055e-01 -7.65100837e-01 -7.26817131e-01 3.57097030e-01
-3.22145045e-01 -2.79305995e-01 7.39784777e-01 -1.07792199e+00
1.52615881e+00 -2.03014421e+00 -1.03565380e-01 7.90769756e-01
1.07207251e+00 4.16101664e-01 -3.67379077e-02 -3.78863722e-01
-2.62291789e-01 6.19102836e-01 -3.35888982e-01 3.35779428e-01
-4.55804974e-01 -6.31087646e-02 1.24381460e-01 2.65192062e-01
2.18904018e-01 1.54566109e+00 -1.03475666e+00 -7.65041769e-01
3.95195544e-01 3.77678275e-01 -3.90282631e-01 -5.83049990e-02
2.33048975e-01 7.20364332e-01 -1.20223808e+00 4.81997490e-01
3.94126564e-01 -9.42878425e-01 5.17496884e-01 -4.79576588e-01
4.83336508e-01 -8.18021875e-03 -1.01679552e+00 1.51047063e+00
-2.29589984e-01 4.10871476e-01 -3.95261317e-01 -1.26314104e+00
1.33940327e+00 3.88951629e-01 6.42472863e-01 -8.48915577e-01
4.15091157e-01 4.83257264e-01 2.78404355e-01 -5.81293285e-01
-4.65824813e-01 1.61435723e-01 7.70208389e-02 2.99666643e-01
-3.98788482e-01 3.99132907e-01 2.62512527e-02 3.25029254e-01
8.72746587e-01 -3.09753418e-01 8.43263268e-01 -3.61896783e-01
9.43560719e-01 1.42185003e-01 5.42628229e-01 4.59215432e-01
-5.57348907e-01 5.08898497e-01 6.04020357e-01 -7.54800260e-01
-8.99417818e-01 -8.15531790e-01 -9.28176194e-02 3.13767374e-01
3.99882913e-01 -3.29249412e-01 -7.44921625e-01 -9.71068740e-01
1.98577959e-02 4.22542542e-01 -3.71592909e-01 -4.75539088e-01
-9.44923043e-01 -6.43567681e-01 3.79450917e-01 9.07750428e-01
9.01597738e-01 -7.14014113e-01 -5.61275423e-01 2.74914533e-01
-4.19489086e-01 -1.28230679e+00 -4.27978337e-01 -7.05598354e-01
-1.24242938e+00 -1.58162820e+00 -6.46814704e-01 -1.05851293e+00
8.70462835e-01 3.11744571e-01 1.30982697e+00 5.66326976e-01
-7.07285762e-01 1.01468861e-01 -1.95954069e-01 9.60974842e-02
-3.84481102e-01 1.20744638e-01 -6.42843664e-01 2.22466409e-01
4.03592020e-01 -5.77150762e-01 -1.11651099e+00 4.50199246e-01
-1.95529342e-01 2.11445019e-02 7.02513874e-01 4.36550647e-01
8.00013006e-01 -6.00684620e-03 1.03880668e+00 -1.16805470e+00
2.92899281e-01 -5.50475836e-01 -4.47294682e-01 3.62427890e-01
-8.38513374e-01 -1.63267836e-01 4.08852845e-01 -1.66767947e-02
-9.50588822e-01 5.42172119e-02 -2.38522273e-02 -1.84951499e-01
-2.86131829e-01 5.48195422e-01 -3.43109101e-01 1.66154802e-02
6.02947593e-01 -2.55220681e-01 2.19217435e-01 -5.07984698e-01
5.69749057e-01 5.65735459e-01 7.15368032e-01 -3.48901868e-01
3.89372945e-01 4.03764367e-01 5.80549240e-01 -3.21915239e-01
-8.59477580e-01 -6.56385958e-01 -5.49734116e-01 -2.19997793e-01
1.29276109e+00 -8.04285049e-01 -8.25034201e-01 -1.02222458e-01
-8.96884561e-01 2.83230156e-01 -1.87765643e-01 5.39118767e-01
-1.44901276e-01 6.73568547e-01 -4.11250859e-01 -4.54285502e-01
-5.89674354e-01 -1.09511030e+00 9.29700851e-01 2.47515097e-01
-3.86717707e-01 -1.35318065e+00 -3.96824598e-01 5.62691331e-01
3.10804754e-01 5.23441494e-01 1.51374209e+00 -7.62625217e-01
-5.65282702e-01 -4.12489504e-01 -8.87801766e-01 2.03287639e-02
4.47611064e-01 -3.46809804e-01 -4.81147677e-01 1.85228363e-01
-2.57559001e-01 5.12502491e-01 6.83955014e-01 8.84038806e-01
1.21470344e+00 5.08662872e-02 -6.91574454e-01 4.86359537e-01
1.29006219e+00 4.55323637e-01 5.69700301e-01 2.82607526e-01
1.01864934e+00 6.06929541e-01 3.48033637e-01 -1.71958450e-02
4.60939765e-01 5.50075233e-01 4.67814207e-01 -3.95858407e-01
-9.09698009e-01 -2.49968953e-02 -4.32886273e-01 5.90531707e-01
-7.60540552e-03 -2.07348578e-02 -1.06174493e+00 3.94979447e-01
-1.74418557e+00 -5.10484874e-01 -7.82160163e-01 1.90140665e+00
3.42773259e-01 -1.76477451e-02 6.01550266e-02 6.59654737e-02
1.24385679e+00 -1.78943753e-01 -5.72064757e-01 1.28916591e-01
-3.53152566e-02 4.46911365e-01 2.99652874e-01 4.89929318e-01
-1.07343364e+00 8.67293954e-01 5.43958139e+00 6.48529291e-01
-8.28421175e-01 -5.25727775e-03 1.22704172e+00 9.15660918e-01
-1.70995131e-01 1.49486929e-01 -5.38771152e-01 4.45710480e-01
7.79020309e-01 -3.31495345e-01 -5.09619303e-02 8.15790236e-01
3.11730087e-01 4.05292183e-01 -5.29361844e-01 8.48996878e-01
-2.68097878e-01 -1.40714109e+00 2.95350313e-01 -1.91292450e-01
4.13395137e-01 -6.14632428e-01 -4.68849950e-03 -1.85422868e-01
1.76484868e-01 -9.55382347e-01 3.29403728e-02 2.94239789e-01
8.99973512e-01 -5.23306608e-01 8.79311800e-01 -2.82767206e-01
-1.71058214e+00 -1.92671400e-02 5.79968318e-02 3.70061606e-01
4.12507027e-01 7.16671884e-01 -7.95941472e-01 8.20982218e-01
4.25098360e-01 1.09123695e+00 -6.34175003e-01 1.20373058e+00
-4.23030287e-01 7.84187376e-01 2.41181001e-01 4.49922949e-01
1.14953602e-02 -4.04955298e-01 5.50580502e-01 8.22478056e-01
3.71974021e-01 1.65950269e-01 4.92154956e-01 9.16079402e-01
-1.17864460e-01 4.77917969e-01 -1.84984401e-01 8.18253160e-02
7.36241162e-01 1.15909791e+00 -1.35093486e+00 -6.53851569e-01
-3.17013413e-01 5.42321265e-01 -6.88776597e-02 3.75024348e-01
-9.48194444e-01 -2.88765192e-01 4.93905634e-01 1.59904748e-01
-2.94709802e-02 2.49081805e-01 -9.61894095e-01 -8.46772373e-01
-1.27227217e-01 -5.59525609e-01 8.93889546e-01 -6.64134204e-01
-1.36814797e+00 6.67231619e-01 -3.00842583e-01 -1.14054906e+00
1.96174204e-01 -4.74065989e-01 -5.99758327e-01 9.77461934e-01
-1.72635269e+00 -1.14953983e+00 -6.93806767e-01 2.81651109e-01
5.17626286e-01 -1.80082723e-01 8.80433500e-01 5.46442151e-01
-7.17751205e-01 3.63374829e-01 -7.32043624e-01 5.04816592e-01
2.99882114e-01 -1.16747200e+00 6.38953269e-01 6.79942369e-01
-9.01859626e-02 4.37612116e-01 1.78440914e-01 -8.37622583e-01
-5.55843472e-01 -1.55182612e+00 1.03648055e+00 -1.93998069e-01
5.15286505e-01 4.87298936e-01 -1.03047287e+00 7.01032877e-01
-4.51965034e-01 3.12284648e-01 7.94993222e-01 -4.10447776e-01
-4.24229920e-01 -1.09950276e-02 -1.07834494e+00 5.35712898e-01
1.34212995e+00 -3.29512179e-01 -4.14719105e-01 3.22272629e-01
8.24346006e-01 -2.09222361e-01 -1.12110603e+00 4.51581627e-01
4.23059613e-01 -4.57060158e-01 1.53418863e+00 -9.79541659e-01
3.32879156e-01 -4.77218390e-01 -2.20158827e-02 -8.20252717e-01
-6.38086140e-01 -9.11653340e-02 -1.94011740e-02 9.18301225e-01
4.03256118e-01 -9.35756147e-01 8.65455866e-01 5.47340155e-01
-3.30067575e-01 -5.06269336e-01 -6.04866922e-01 -6.19557261e-01
-1.09286726e-01 -2.98364133e-01 6.86953843e-01 1.01488519e+00
-5.26184142e-01 3.50300461e-01 8.49701911e-02 3.69661689e-01
8.43693078e-01 3.02988440e-01 3.63910884e-01 -1.69691741e+00
-8.87358785e-02 -7.88899839e-01 -8.22550416e-01 -7.47919261e-01
1.81316629e-01 -1.52659166e+00 -8.47782016e-01 -1.72192264e+00
1.25877485e-01 -7.75992930e-01 -5.26713312e-01 3.75697643e-01
-5.73857903e-01 4.68182713e-01 -2.21426375e-02 3.65555465e-01
-2.95111358e-01 -1.15186796e-01 1.60693169e+00 4.01039384e-02
-3.22459728e-01 1.07325651e-01 -1.11735594e+00 8.04644763e-01
1.10447204e+00 -5.20685971e-01 -5.74475884e-01 -1.70943618e-01
-1.15787879e-01 1.76132128e-01 7.18163073e-01 -5.75093508e-01
6.62964508e-02 1.18861206e-01 8.83321986e-02 2.90044528e-02
-2.73093224e-01 -9.74717617e-01 3.87960285e-01 8.99003148e-01
-4.36356902e-01 -2.54915264e-02 -3.21477540e-02 7.46689558e-01
-3.20811659e-01 9.48143899e-02 5.83213508e-01 -2.24757522e-01
-9.72049892e-01 3.08058441e-01 -2.65030563e-01 3.97442311e-01
1.22253275e+00 -2.79656708e-01 -3.36021781e-01 -2.30888631e-02
-1.17450571e+00 1.85121186e-02 -2.32963443e-01 3.38145941e-01
6.48526967e-01 -1.33591938e+00 -7.91121900e-01 1.00402482e-01
2.33230799e-01 -2.12854460e-01 4.08331335e-01 1.11433303e+00
-7.78469682e-01 4.04187471e-01 -8.80723670e-02 -7.08627939e-01
-1.32877886e+00 3.39756310e-01 5.79409361e-01 -3.11106831e-01
-1.06495774e+00 5.81063569e-01 3.41278613e-01 -1.15326814e-01
-1.39770553e-01 -1.14123225e-01 -4.74707842e-01 -3.41686487e-01
4.00762707e-01 6.27707124e-01 -9.08514932e-02 -7.09987342e-01
-4.19662297e-01 1.04884171e+00 2.30060652e-01 4.50273663e-01
8.31656933e-01 -2.46752694e-01 -2.05885842e-01 -3.02489609e-01
1.06369936e+00 -3.25585365e-01 -5.11310637e-01 -3.83959919e-01
2.14156270e-01 -4.79892164e-01 2.63666689e-01 -8.06215286e-01
-1.65688920e+00 1.00221884e+00 8.21847975e-01 -4.21293406e-03
1.07666790e+00 2.32551068e-01 1.03633988e+00 -1.90035969e-01
7.32600540e-02 -1.51577160e-01 -7.48573542e-02 -3.90331000e-01
4.25672859e-01 -1.25308430e+00 -7.15795010e-02 -1.34872460e+00
-6.88763142e-01 1.25629759e+00 3.14812064e-01 -2.68123686e-01
9.47793245e-01 -3.51543576e-01 2.39580140e-01 -5.80452085e-01
1.01361059e-01 -1.64079770e-01 7.79933512e-01 4.15860146e-01
2.82934159e-01 3.97712290e-01 -4.63900715e-01 7.04489470e-01
4.45453040e-02 -1.00242995e-01 -1.79920286e-01 4.68102038e-01
-5.81231654e-01 -1.08318102e+00 -1.16543835e-02 5.14151454e-01
-5.31494021e-01 -3.75195146e-02 -3.11816722e-01 6.61286831e-01
-1.35434583e-01 9.40269172e-01 4.04601975e-04 -1.10129006e-01
3.83703768e-01 9.00965482e-02 -4.78569269e-02 -4.15789127e-01
-4.05469388e-01 8.77139196e-02 -5.64183220e-02 -5.78150570e-01
-4.22791898e-01 -2.00463146e-01 -1.79897130e+00 -3.41055728e-02
-2.65381873e-01 9.37855765e-02 3.90498489e-01 9.38815355e-01
9.26881075e-01 1.04966033e+00 5.77946603e-01 5.31824864e-02
-7.40845799e-02 -3.00367862e-01 -3.69053781e-01 8.35900962e-01
-1.12560458e-01 -5.89492917e-01 -2.82541096e-01 3.50903988e-01] | [14.56943416595459, -2.450978994369507] |
2004dd8c-7924-47da-9bab-2c50e7cec19a | single-uhd-image-dehazing-via-interpretable | 2202.08589 | null | https://arxiv.org/abs/2202.08589v1 | https://arxiv.org/pdf/2202.08589v1.pdf | Single UHD Image Dehazing via Interpretable Pyramid Network | Currently, most single image dehazing models cannot run an ultra-high-resolution (UHD) image with a single GPU shader in real-time. To address the problem, we introduce the principle of infinite approximation of Taylor's theorem with the Laplace pyramid pattern to build a model which is capable of handling 4K hazy images in real-time. The N branch networks of the pyramid network correspond to the N constraint terms in Taylor's theorem. Low-order polynomials reconstruct the low-frequency information of the image (e.g. color, illumination). High-order polynomials regress the high-frequency information of the image (e.g. texture). In addition, we propose a Tucker reconstruction-based regularization term that acts on each branch network of the pyramid model. It further constrains the generation of anomalous signals in the feature space. Extensive experimental results demonstrate that our approach can not only run 4K images with haze in real-time on a single GPU (80FPS) but also has unparalleled interpretability. The developed method achieves state-of-the-art (SOTA) performance on two benchmarks (O/I-HAZE) and our updated 4KID dataset while providing the reliable groundwork for subsequent optimization schemes. | ['Tao Wang', 'Yunliang Zhuang', 'Chen Lv', 'Xiang Chen', 'Zhuoran Zheng', 'Boxue Xiao'] | 2022-02-17 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 1.58665940e-01 -2.03745201e-01 5.92378438e-01 -3.48869264e-02
-3.00901473e-01 1.34257630e-01 1.93607062e-01 -1.82781547e-01
-2.25497022e-01 2.29949251e-01 -3.09451312e-01 -3.23299050e-01
1.28063247e-01 -9.46955264e-01 -8.87393057e-01 -1.08334732e+00
-8.43756273e-02 -4.00510803e-02 5.78633547e-01 -4.32482213e-01
3.56698930e-01 6.15287662e-01 -2.31818795e+00 5.77943921e-01
1.08667159e+00 1.21628428e+00 1.87831610e-01 9.82886016e-01
2.02329800e-01 1.01960814e+00 -5.57110667e-01 -1.97584126e-02
5.89224994e-01 -9.19850245e-02 -3.40449512e-01 6.37022555e-02
9.97677624e-01 -6.41528070e-01 -4.16103899e-01 1.24641013e+00
3.77729148e-01 1.09877750e-01 3.83990824e-01 -1.02188325e+00
-7.63909996e-01 -3.37170035e-01 -9.85665560e-01 2.28967115e-01
1.38052925e-01 4.67013091e-01 6.04628861e-01 -9.77202117e-01
4.10920054e-01 1.05338693e+00 6.06104910e-01 1.44459248e-01
-1.15281415e+00 -3.25170010e-01 -2.27874115e-01 2.89189488e-01
-1.40154910e+00 -8.57518539e-02 5.84345162e-01 -3.47962886e-01
1.00159359e+00 6.20569706e-01 7.58356392e-01 4.71027374e-01
7.74760485e-01 3.38843584e-01 1.38134158e+00 -4.43651795e-01
1.44894823e-01 -2.59677600e-02 2.64433116e-01 8.37797523e-01
2.16002494e-01 3.39575857e-01 -9.43163753e-01 -3.07916284e-01
1.01964152e+00 -2.20295802e-01 -4.57085937e-01 -1.68878457e-03
-8.88808370e-01 6.21913671e-01 4.19503033e-01 -2.74595385e-03
-3.36400568e-01 1.96524918e-01 -5.76100349e-02 4.00172263e-01
6.57945633e-01 1.96166351e-01 -1.82810247e-01 2.45578378e-01
-7.94454694e-01 2.99264163e-01 6.49897933e-01 8.53496730e-01
1.23889244e+00 2.60727555e-01 -6.17091022e-02 5.06913066e-01
2.01119483e-01 7.49488175e-01 1.38656601e-01 -1.27434909e+00
3.29405591e-02 2.33262211e-01 1.06194019e-01 -1.17936969e+00
-3.10684472e-01 -2.41896287e-01 -1.14735579e+00 9.18110371e-01
3.69595200e-01 2.15228289e-01 -9.92922604e-01 9.41608131e-01
5.51733792e-01 5.82680941e-01 -4.79048453e-02 1.04314101e+00
7.53957152e-01 1.05936122e+00 -6.17939889e-01 -3.09149951e-01
1.46228826e+00 -9.53433454e-01 -6.34437859e-01 6.48186207e-02
3.42671275e-01 -1.00670266e+00 1.18954313e+00 7.44852304e-01
-1.02002323e+00 -5.37776172e-01 -1.13036084e+00 -4.01108831e-01
-2.58769244e-01 -3.69626403e-01 6.09559178e-01 4.62209523e-01
-1.34377801e+00 6.95864379e-01 -8.20331991e-01 1.18121825e-01
7.37110153e-02 3.48342150e-01 -1.87153459e-01 -1.15798280e-01
-9.05032337e-01 5.98684967e-01 1.60000384e-01 3.21705937e-01
-6.25851810e-01 -1.07983470e+00 -5.81797719e-01 -1.45837530e-01
9.56636295e-02 -6.01351321e-01 5.42058885e-01 -1.00545907e+00
-1.68588400e+00 1.03942585e+00 -3.33403312e-02 -3.70363176e-01
3.97953629e-01 -3.27848703e-01 -3.37905318e-01 3.66923720e-01
-2.73433208e-01 3.04379702e-01 1.48283589e+00 -1.46528983e+00
-5.71538627e-01 -4.17460561e-01 -1.16946012e-01 2.22333908e-01
-1.43420398e-01 -5.06528765e-02 -6.38768077e-01 -5.61865509e-01
6.72160760e-02 -9.00055349e-01 -1.88572973e-01 1.29536912e-01
-3.02636474e-01 3.38486671e-01 1.02048326e+00 -7.71731794e-01
1.02496421e+00 -2.52019906e+00 2.39564311e-02 3.94817173e-01
3.98435295e-01 3.39170784e-01 9.39597487e-02 -4.74120378e-02
3.32678668e-02 -3.89202923e-01 -2.19441772e-01 -1.89444691e-01
-4.24399078e-01 4.30267751e-01 -4.08471823e-01 7.79567420e-01
1.23601640e-02 4.59160119e-01 -5.63086092e-01 -3.16961259e-01
4.53378409e-01 1.05461109e+00 -7.73978591e-01 4.23487395e-01
-1.00886002e-01 4.67130065e-01 -1.37593523e-01 3.99729282e-01
1.25145030e+00 -2.19162822e-01 -4.55327034e-01 -4.08570915e-01
-5.76701820e-01 -3.76947880e-01 -1.08692122e+00 1.40579414e+00
-4.43396032e-01 6.07927263e-01 1.84605911e-01 -4.55625027e-01
7.44514763e-01 -3.34104151e-02 2.17568830e-01 -9.39447343e-01
-7.62223452e-02 2.64252573e-01 -3.15534383e-01 -5.28055429e-01
8.02788436e-01 5.09262197e-02 4.00806218e-01 1.57202393e-01
-2.41739824e-01 -5.20331860e-01 -1.99185476e-01 2.40869392e-02
9.55088198e-01 1.97581854e-02 -9.27675068e-02 -6.50839627e-01
5.54019034e-01 3.73491123e-02 2.61221141e-01 9.76605773e-01
-6.39082340e-04 8.08794916e-01 4.51294154e-01 -8.45177948e-01
-1.28734696e+00 -7.74929523e-01 -3.66743386e-01 9.49101925e-01
3.22487414e-01 -4.01967973e-01 -8.73131692e-01 1.21193990e-01
-6.85202554e-02 6.78546727e-01 -7.84777761e-01 -7.13901371e-02
-5.50311327e-01 -1.06181073e+00 3.65758508e-01 -1.97188571e-01
6.90019369e-01 -6.87342048e-01 -1.16809356e+00 -7.20575154e-02
1.24134541e-01 -1.20798206e+00 -7.61572793e-02 2.13777468e-01
-7.48920918e-01 -9.41949189e-01 -3.98704797e-01 -5.47066510e-01
5.75303197e-01 5.82741320e-01 1.28118610e+00 3.79447430e-01
-5.43380201e-01 3.18321019e-01 -2.24600092e-01 -2.51816779e-01
-3.42392236e-01 -6.35626972e-01 -1.92185819e-01 2.47482643e-01
3.48676741e-02 -6.82209313e-01 -7.61451185e-01 3.28290224e-01
-1.12689459e+00 3.20308834e-01 2.28139877e-01 9.38414097e-01
1.03325331e+00 4.54108983e-01 -5.00741601e-01 -7.70859003e-01
1.11497320e-01 -8.41762275e-02 -1.12854946e+00 -6.67395163e-03
-4.33181942e-01 1.15572199e-01 7.99317777e-01 -3.09737802e-01
-1.35078430e+00 -1.86319575e-02 -2.91663781e-02 -6.78685188e-01
-4.99159284e-02 3.49475518e-02 1.76984578e-01 -8.46115410e-01
7.20637023e-01 3.41507167e-01 -5.25505841e-02 -4.29533094e-01
2.49593526e-01 4.56060141e-01 1.02175295e+00 -7.67877460e-01
1.00127137e+00 9.80673552e-01 4.52018082e-01 -1.45599437e+00
-9.22685504e-01 -5.61669290e-01 -4.42780763e-01 -2.28721097e-01
9.53625917e-01 -1.14977026e+00 -9.44716692e-01 9.41256344e-01
-1.14472044e+00 -4.55357760e-01 -2.12585211e-01 5.18485755e-02
-5.13267398e-01 5.29088080e-01 -8.00770879e-01 -8.46097589e-01
-4.17811185e-01 -1.03597498e+00 1.38334012e+00 7.93471709e-02
6.23936951e-01 -6.97028935e-01 -7.15827793e-02 2.75041223e-01
4.56602842e-01 4.22329158e-01 8.47461343e-01 4.72566098e-01
-1.00318980e+00 2.12838903e-01 -5.61831892e-01 6.90306067e-01
-3.93321931e-01 2.74789780e-01 -1.28433132e+00 -2.52981722e-01
6.32013381e-01 -1.13581143e-01 8.80210102e-01 3.98002058e-01
1.40486145e+00 -2.58342236e-01 4.65497196e-01 1.40471804e+00
1.94456255e+00 -2.80657113e-01 1.07425725e+00 4.43072289e-01
1.02843606e+00 4.18318897e-01 7.04754829e-01 6.24775469e-01
1.07963488e-01 9.58182871e-01 7.41099536e-01 -3.64670098e-01
-2.02038348e-01 2.22320676e-01 3.66903096e-01 7.40749538e-01
-5.36122024e-01 2.17660982e-02 -7.01963484e-01 6.03506640e-02
-1.72976971e+00 -5.57278335e-01 -8.92867744e-01 2.33024931e+00
5.43485463e-01 -9.98284295e-02 -4.15426046e-01 -1.79658327e-02
2.98040807e-01 2.36278698e-01 -3.88727844e-01 -5.30509830e-01
-4.90030080e-01 3.24143559e-01 7.34829485e-01 7.12010741e-01
-8.14592004e-01 9.10884619e-01 6.34735346e+00 9.45558667e-01
-1.22343922e+00 9.57817361e-02 6.12613440e-01 1.76799297e-02
-3.54722798e-01 -1.72362760e-01 -6.57633841e-01 5.39653242e-01
9.89129961e-01 1.21795870e-01 7.68066883e-01 5.54138422e-01
2.66068786e-01 -4.66724724e-01 -6.22480154e-01 1.19816005e+00
1.96195573e-01 -1.19077802e+00 1.22650638e-01 2.95356542e-01
8.35129917e-01 2.10731223e-01 4.01497841e-01 -2.92209089e-01
-4.64712456e-02 -9.09762084e-01 7.95279086e-01 4.45056707e-01
9.85541642e-01 -7.30516434e-01 5.79723954e-01 2.70204604e-01
-9.77419436e-01 2.80847728e-01 -8.10717940e-01 -1.29960418e-01
-1.97496042e-01 9.07962918e-01 -2.97211766e-01 6.02466643e-01
1.13678741e+00 4.52410638e-01 -5.84464312e-01 6.39248133e-01
-1.20194711e-01 3.88838440e-01 -6.29284620e-01 5.86206436e-01
2.48294175e-01 -5.99533081e-01 4.39419419e-01 9.93821383e-01
5.36407471e-01 5.23046374e-01 -3.26847471e-02 5.75775862e-01
3.34831208e-01 -7.44577795e-02 -4.02111113e-01 5.17270803e-01
-2.98329234e-01 1.04189241e+00 -5.74485540e-01 -1.66914478e-01
-5.46642125e-01 1.29238749e+00 1.71880960e-01 5.16243517e-01
-8.94738913e-01 -1.76314667e-01 8.34048986e-01 3.45419407e-01
3.33067328e-01 -2.57881939e-01 -4.12843883e-01 -1.45300603e+00
9.72184539e-02 -9.17894185e-01 1.12363152e-01 -1.03257883e+00
-1.08757854e+00 4.47882235e-01 -3.13747823e-01 -9.84751642e-01
3.17567497e-01 -9.91854012e-01 -3.45503598e-01 1.03704500e+00
-2.01755643e+00 -1.07828355e+00 -8.29544246e-01 1.07579780e+00
7.92129263e-02 3.24237347e-01 8.96034241e-01 3.08480442e-01
-4.13101077e-01 3.05440545e-01 4.10026282e-01 -3.40861410e-01
5.54037392e-01 -1.18579316e+00 2.72699744e-01 9.51828122e-01
-2.79161274e-01 3.61652434e-01 1.11633492e+00 -4.91086721e-01
-1.91766679e+00 -7.67533481e-01 2.06894979e-01 -3.26120168e-01
6.24617696e-01 -2.70704538e-01 -1.43552923e+00 5.02634466e-01
4.74049598e-02 6.55096710e-01 5.59865296e-01 -3.06633115e-01
-5.99466741e-01 -2.60460645e-01 -1.11314893e+00 2.75124103e-01
7.19757915e-01 -5.29526293e-01 -2.06301257e-01 5.45241594e-01
8.06209326e-01 -9.56148505e-01 -8.59896243e-01 1.24362290e-01
4.27830309e-01 -1.65235925e+00 1.11861014e+00 -1.09752871e-01
4.63211626e-01 -6.29537880e-01 -2.06677079e-01 -9.23065841e-01
-3.34275872e-01 -9.79151070e-01 -3.81861001e-01 5.57318270e-01
-1.95344225e-01 -6.31708324e-01 6.89739764e-01 5.35200119e-01
-3.17918301e-01 -7.23641038e-01 -1.01250637e+00 -6.78471565e-01
-2.17529699e-01 -4.65100378e-01 4.37511057e-01 7.33153760e-01
-5.06931663e-01 -8.78689438e-02 -8.53863060e-01 8.32593203e-01
1.22079873e+00 1.37652442e-01 8.24395120e-01 -1.03095555e+00
-5.08665621e-01 2.71207821e-02 -5.94581723e-01 -8.37299764e-01
-5.10741258e-03 -2.53890246e-01 5.91323804e-03 -1.00378823e+00
7.22873658e-02 -3.76733571e-01 -1.32218003e-01 2.80616373e-01
-1.53724521e-01 5.15039742e-01 7.60723576e-02 3.64206582e-01
-4.05713230e-01 3.72107267e-01 1.32264042e+00 1.29074708e-01
-1.45784855e-01 -2.84364998e-01 -2.96816468e-01 7.89345682e-01
3.84003609e-01 -2.98763007e-01 -1.50356129e-01 -6.98412776e-01
4.13880497e-01 -2.04933909e-04 6.20979548e-01 -1.04628134e+00
2.51513243e-01 -4.03704457e-02 3.39184850e-01 -4.45463419e-01
4.61436450e-01 -9.67005372e-01 3.92891765e-01 4.18302715e-01
3.40248317e-01 -1.46889046e-01 2.25930512e-01 4.95671034e-01
-3.93072665e-01 1.54116154e-01 1.20351183e+00 -4.19565663e-02
-7.80554116e-01 2.98535526e-01 -6.17796183e-02 -1.28711805e-01
8.18174422e-01 -2.23286390e-01 -6.75557315e-01 -9.96859595e-02
-1.79516837e-01 -7.86841512e-02 1.07106280e+00 -2.32262257e-02
7.60166407e-01 -9.37159061e-01 -8.15384686e-01 7.19731987e-01
7.39360154e-02 3.55506629e-01 6.38820291e-01 7.25857556e-01
-1.31824243e+00 -2.19139948e-01 -4.28561687e-01 -8.32813084e-01
-1.29775429e+00 5.59730113e-01 4.38534886e-01 -1.43912196e-01
-1.15970969e+00 8.61674249e-01 5.55869818e-01 1.04464412e-01
-5.38664125e-02 -1.37518525e-01 1.57364026e-01 -3.88817430e-01
9.20363605e-01 3.26471895e-01 2.32378691e-01 -7.85158932e-01
-1.70475096e-01 8.79631460e-01 8.48092139e-02 1.39707059e-01
1.46051204e+00 -7.05030793e-03 -8.27029347e-01 3.10001284e-01
1.16512036e+00 9.81438458e-02 -1.44104743e+00 -4.72663194e-02
-5.80084741e-01 -1.05514479e+00 5.52507102e-01 -2.84738243e-01
-1.01840198e+00 8.87133241e-01 7.85807014e-01 2.42706314e-01
1.53503573e+00 -5.09360671e-01 8.35507154e-01 2.94379741e-01
4.69208300e-01 -9.84636545e-01 -1.18056796e-02 6.27014458e-01
5.95560014e-01 -1.04794705e+00 3.42960358e-01 -8.02277327e-01
-3.14156085e-01 1.24901617e+00 4.77660507e-01 -2.59523451e-01
6.88836575e-01 5.09889305e-01 8.94270614e-02 -4.29295897e-01
-5.90763807e-01 1.30438849e-01 2.80441582e-01 5.31566560e-01
1.19328931e-01 -5.85977882e-02 1.31008044e-01 -1.50585651e-01
-1.89058661e-01 -1.80988699e-01 7.88810730e-01 7.11715162e-01
-5.46394169e-01 -5.18097043e-01 -8.39428902e-01 2.44681999e-01
-3.34012091e-01 -2.14090958e-01 2.06148237e-01 4.87643510e-01
2.82221049e-01 7.04292655e-01 2.04450637e-01 -2.35354930e-01
3.83554190e-01 -3.74293745e-01 5.08673429e-01 -2.18551248e-01
-5.52175760e-01 2.36176237e-01 -3.11309844e-01 -1.00568712e+00
-2.84982413e-01 -3.34451646e-01 -9.99711752e-01 -7.65596509e-01
-4.83631045e-02 2.42205169e-02 6.35073960e-01 4.63363290e-01
3.83633435e-01 4.59195971e-01 6.37089491e-01 -1.14755273e+00
-2.45088607e-01 -3.89389217e-01 -1.07469845e+00 4.71548617e-01
6.52264714e-01 -2.63499051e-01 -7.70778894e-01 4.69602942e-02] | [10.88658618927002, -2.88037109375] |
ed7fe6e9-6d0f-4164-a807-549363411c7b | fine-grained-analysis-of-propaganda-in-news-1 | null | null | https://aclanthology.org/D19-1565 | https://aclanthology.org/D19-1565.pdf | Fine-Grained Analysis of Propaganda in News Article | Propaganda aims at influencing people{'}s mindset with the purpose of advancing a specific agenda. Previous work has addressed propaganda detection at document level, typically labelling all articles from a propagandistic news outlet as propaganda. Such noisy gold labels inevitably affect the quality of any learning system trained on them. A further issue with most existing systems is the lack of explainability. To overcome these limitations, we propose a novel task: performing fine-grained analysis of texts by detecting all fragments that contain propaganda techniques as well as their type. In particular, we create a corpus of news articles manually annotated at fragment level with eighteen propaganda techniques and propose a suitable evaluation measure. We further design a novel multi-granularity neural network, and we show that it outperforms several strong BERT-based baselines. | ['Rostislav Petrov', "Alberto Barr{\\'o}n-Cede{\\~n}o", 'Giovanni Da San Martino', 'Preslav Nakov', 'Seunghak Yu'] | 2019-11-01 | null | null | null | ijcnlp-2019-11 | ['propaganda-detection'] | ['natural-language-processing'] | [ 3.06207180e-01 3.14150155e-01 -6.63282156e-01 -1.66472763e-01
-8.68560076e-01 -5.45740902e-01 1.29061043e+00 4.42079663e-01
-4.33457494e-01 7.75245130e-01 1.13366163e+00 -6.18060470e-01
3.39698851e-01 -8.22624266e-01 -6.87000871e-01 -4.25261706e-01
3.31316829e-01 4.30793524e-01 -3.51081975e-02 -3.24543059e-01
5.59821010e-01 -2.39226699e-01 -7.39359379e-01 7.24439502e-01
9.76659238e-01 2.18414947e-01 -4.38830763e-01 6.29918396e-01
-1.70173600e-01 1.48340976e+00 -1.25621581e+00 -5.43771148e-01
-1.66937098e-01 -7.62841403e-01 -1.19651437e+00 -2.72190161e-02
5.91598213e-01 -4.38241750e-01 -9.04103816e-02 1.03036392e+00
2.04468951e-01 -2.85420656e-01 1.02484143e+00 -5.23772240e-01
-1.05995297e+00 1.40087223e+00 -6.49482012e-01 4.95127738e-01
3.93194497e-01 -2.02120885e-01 1.14711928e+00 -4.46572870e-01
8.69816422e-01 1.50585294e+00 7.26735592e-01 6.95074797e-01
-1.12125981e+00 -3.16793352e-01 3.85628253e-01 6.92116693e-02
-7.89336622e-01 -2.31656536e-01 9.25216019e-01 -7.66198337e-01
6.45065904e-01 2.67750770e-01 6.28250599e-01 1.82396150e+00
3.25148433e-01 8.74169827e-01 1.21479237e+00 -6.80251300e-01
4.76846583e-02 -8.95743519e-02 5.58537722e-01 6.69122875e-01
3.90973836e-01 -3.33611131e-01 -3.69556636e-01 -2.95120060e-01
3.58651817e-01 -3.97971421e-01 -2.17905939e-01 3.71684372e-01
-1.24166274e+00 1.33112299e+00 4.55859095e-01 6.44740999e-01
-2.92445242e-01 3.99521142e-01 4.37712044e-01 1.50741249e-01
1.08014894e+00 7.07615733e-01 -7.45814294e-02 -1.00617573e-01
-9.84884560e-01 7.20770121e-01 9.59893167e-01 3.28362495e-01
2.09376484e-01 -2.51905233e-01 -6.39228940e-01 4.43956673e-01
1.83366641e-01 4.25270170e-01 2.48357907e-01 -3.88465613e-01
6.23036027e-01 5.17275333e-01 1.71586186e-01 -1.35936952e+00
-7.63721824e-01 -6.34667397e-01 -7.20267892e-01 -2.59150326e-01
5.92059910e-01 -3.50251913e-01 -8.23151946e-01 1.52368498e+00
3.44326824e-01 -2.61157453e-01 -4.45357502e-01 8.99842143e-01
9.98326361e-01 7.85488427e-01 3.72660607e-01 -2.79761076e-01
1.45331419e+00 -1.10387087e+00 -9.50706363e-01 -2.60392100e-01
9.63211536e-01 -8.07806849e-01 9.18812692e-01 1.91899478e-01
-6.56589985e-01 -6.69672415e-02 -7.48049974e-01 -1.94292307e-01
-1.87247351e-01 1.29005099e-02 6.57869935e-01 6.37373149e-01
-4.11253482e-01 3.44465822e-01 -6.43999934e-01 -1.93295613e-01
4.65904564e-01 -3.93265694e-01 2.13766575e-01 3.59728336e-01
-1.40271831e+00 1.22477949e+00 3.02321672e-01 -2.00325936e-01
-7.78095961e-01 -5.98969877e-01 -7.00940192e-01 -6.51946813e-02
4.88389164e-01 -7.02693343e-01 1.43935919e+00 -1.12856412e+00
-9.82716203e-01 7.86581218e-01 1.00399606e-01 -7.16074347e-01
5.49694121e-01 -6.06515110e-01 -5.64956486e-01 -3.01962905e-03
5.21228135e-01 1.13236651e-01 1.09953201e+00 -1.05095863e+00
-5.47297239e-01 7.76491463e-02 3.20527315e-01 -2.63340329e-03
-3.07076901e-01 5.53023577e-01 3.71742666e-01 -1.01487017e+00
-3.15278292e-01 -7.11778879e-01 -9.10981596e-02 -6.82493329e-01
-8.38936567e-01 -5.95919907e-01 5.64188838e-01 -7.75681734e-01
1.60034442e+00 -1.63760209e+00 -3.22996378e-02 -4.45961654e-02
6.90486252e-01 3.04837883e-01 -1.30104152e-02 5.75824142e-01
3.20394903e-01 6.05375171e-01 1.86871678e-01 9.89006981e-02
6.85819238e-02 -9.94763821e-02 -5.37228227e-01 6.83646262e-01
2.04932064e-01 9.06404257e-01 -1.03899574e+00 -4.59096611e-01
-2.18692109e-01 3.46315801e-01 -6.83783352e-01 -2.14693040e-01
-6.97505772e-01 4.21096444e-01 -8.06169331e-01 3.29840422e-01
2.52492607e-01 -6.81794047e-01 1.49655357e-01 1.62828073e-01
-3.07678074e-01 1.11613524e+00 -4.17443842e-01 1.12107265e+00
-2.87784636e-01 7.94995487e-01 -7.89204761e-02 -6.82391763e-01
2.84169585e-01 2.97389716e-01 1.93090364e-01 -4.79822606e-01
4.79501814e-01 -9.19849500e-02 1.10394932e-01 -4.63331193e-01
5.68721890e-01 -2.64928699e-01 -5.50953984e-01 1.05035460e+00
-3.59235674e-01 2.45834857e-01 2.21433565e-01 6.48931861e-01
1.24464536e+00 -1.13104433e-01 6.35245383e-01 -3.34597498e-01
4.06737268e-01 4.33064312e-01 2.23809168e-01 1.27405322e+00
-2.47750580e-01 3.48336369e-01 5.26160181e-01 -8.92759919e-01
-9.34349656e-01 -6.84403300e-01 -1.63507134e-01 1.61379504e+00
-1.08855985e-01 -7.24998355e-01 -8.51953506e-01 -1.26496458e+00
-2.27443203e-01 9.42209303e-01 -1.14076126e+00 3.00872624e-01
-9.77772295e-01 -8.53281677e-01 5.68280101e-01 1.72063336e-01
4.09486383e-01 -1.10969913e+00 -5.94685316e-01 4.70212996e-01
-5.68221748e-01 -7.79850364e-01 -3.99433523e-01 -9.68275964e-02
-9.30769220e-02 -1.21783352e+00 -3.55012566e-01 -6.42648518e-01
5.23172855e-01 1.45806074e-01 1.30498707e+00 3.72367084e-01
1.43898845e-01 -1.84445158e-01 -6.42687142e-01 -5.12587786e-01
-1.11251521e+00 3.86810243e-01 -1.01585262e-01 -4.01951283e-01
4.41159099e-01 -8.76705945e-02 -2.96459228e-01 -2.63442934e-01
-6.79943860e-01 4.20099974e-01 4.22803044e-01 8.88803482e-01
-2.02428490e-01 -1.12520300e-01 6.86212420e-01 -1.42350209e+00
1.14547563e+00 -5.30628145e-01 -1.11753859e-01 1.55130535e-01
-2.84433663e-01 3.79295051e-02 6.71454430e-01 -3.76046211e-01
-1.14778519e+00 -5.29257238e-01 -2.57981598e-01 7.33331501e-01
-2.63285637e-01 6.67389452e-01 4.66678768e-01 4.64923233e-01
1.47319043e+00 -8.16444233e-02 -4.45991874e-01 -5.29061139e-01
7.47744679e-01 7.64853477e-01 4.06821042e-01 -4.43785042e-01
8.54535460e-01 3.54568481e-01 -5.47201097e-01 -5.95354080e-01
-1.90508199e+00 -6.15142167e-01 -3.56575757e-01 -7.40078762e-02
7.61963367e-01 -7.52064824e-01 -4.70306665e-01 1.39997050e-01
-1.55941749e+00 -8.17899629e-02 1.31115973e-01 3.87057006e-01
-2.34841436e-01 3.15441787e-01 -1.08337831e+00 -7.57738352e-01
-5.94170392e-01 -4.42384392e-01 8.45600307e-01 -2.30858102e-01
-7.32133508e-01 -1.11024320e+00 3.83957773e-01 5.92821240e-01
2.81673789e-01 4.04703856e-01 1.02952993e+00 -9.23225343e-01
2.93394458e-02 -4.77327742e-02 -2.38353103e-01 -8.26748088e-02
1.54726267e-01 -9.29505080e-02 -6.67579234e-01 5.60782813e-02
1.40441790e-01 -4.84065473e-01 1.13589895e+00 4.75864112e-01
5.32136381e-01 -1.36395645e+00 -3.87691557e-01 -2.45675468e-03
8.82746935e-01 -2.80822754e-01 2.07495347e-01 8.33524585e-01
7.64274061e-01 4.93577808e-01 3.10832500e-01 4.05219346e-01
3.36679876e-01 3.90838027e-01 2.33766630e-01 -8.77529010e-02
-1.62389740e-01 -4.01023865e-01 4.86998349e-01 6.14358604e-01
-1.42907664e-01 -5.80522418e-01 -9.98168409e-01 6.87496006e-01
-1.86174774e+00 -1.61440313e+00 -4.97139037e-01 1.31175363e+00
1.20254290e+00 2.67077833e-01 3.27353269e-01 1.08159803e-01
7.78132021e-01 5.52016735e-01 2.76220560e-01 -5.96951306e-01
1.42702088e-02 -3.65841947e-02 1.10238761e-01 7.57357717e-01
-1.74652123e+00 1.00977850e+00 6.87717962e+00 7.78250694e-01
-9.39161360e-01 2.82490492e-01 5.36081374e-01 -2.54689723e-01
-4.59667861e-01 -3.01546127e-01 -8.43732715e-01 5.05433083e-01
8.77242386e-01 -7.77099133e-02 8.15906301e-02 8.84453773e-01
4.74052370e-01 1.05967365e-01 -7.92936087e-01 2.96278834e-01
3.29834253e-01 -1.89922023e+00 2.49737412e-01 -2.33500600e-02
1.16229451e+00 1.29496690e-03 -1.11620307e-01 3.37155879e-01
6.81504607e-01 -1.07180691e+00 1.10790157e+00 -1.96422964e-01
2.58968472e-01 -4.71766859e-01 6.39806926e-01 6.37940407e-01
-4.64818686e-01 1.54139936e-01 3.87373269e-02 -6.56778336e-01
4.84471262e-01 7.50353932e-01 -1.08329749e+00 1.92094103e-01
8.97744820e-02 5.52168429e-01 -5.49831986e-01 4.91079122e-01
-9.25745249e-01 1.38202643e+00 -1.11253157e-01 -5.84431589e-01
7.29142547e-01 4.79754329e-01 6.60509527e-01 1.61950886e+00
-2.04213157e-01 1.25278234e-01 3.38874757e-01 8.99361253e-01
-2.83382297e-01 2.45752394e-01 -3.39073658e-01 -3.22692513e-01
2.22796723e-01 1.09650707e+00 -6.17251873e-01 -5.81581414e-01
-3.08311284e-01 5.67449510e-01 4.80359674e-01 1.18276328e-01
-9.85354960e-01 2.16956958e-01 2.06948131e-01 4.45445746e-01
-7.29721561e-02 -8.86809826e-02 -3.79967988e-01 -1.19076669e+00
-1.37124538e-01 -9.21145976e-01 5.40621996e-01 -1.71123743e-01
-1.45357740e+00 4.82113093e-01 -2.79245466e-01 -6.24932289e-01
-4.14852023e-01 -3.44443202e-01 -7.17289567e-01 4.53541785e-01
-1.23548770e+00 -1.51141572e+00 3.25251706e-02 1.05734624e-01
7.40577579e-01 2.67138779e-01 5.72753489e-01 -8.15366954e-02
-3.08432728e-01 2.20627755e-01 -1.46769822e-01 4.94874865e-01
7.35890627e-01 -1.12665391e+00 6.45476580e-01 1.02973294e+00
3.49276781e-01 8.89117062e-01 1.27215600e+00 -9.53327537e-01
-7.11180329e-01 -1.14085162e+00 1.52702272e+00 -9.48446453e-01
1.13999796e+00 -3.61405849e-01 -7.91549981e-01 6.97701514e-01
4.40622509e-01 -6.41701519e-01 6.51821077e-01 7.74777055e-01
-6.96347177e-01 7.76557505e-01 -7.83455193e-01 6.14609122e-01
9.73524988e-01 -3.77467722e-01 -1.11009026e+00 9.54200685e-01
5.74613452e-01 -2.51045078e-01 -1.35530338e-01 -1.06138317e-02
4.31056619e-01 -5.46259880e-01 5.21698654e-01 -1.09695673e+00
1.10986865e+00 7.24675134e-03 4.49762285e-01 -1.29086530e+00
-6.61425054e-01 -6.25465631e-01 -2.48203918e-01 1.06182957e+00
4.03219879e-01 -2.14450970e-01 6.52292967e-01 8.66793767e-02
1.69424294e-03 -3.86518210e-01 -6.75498664e-01 -6.46576405e-01
4.04846668e-01 -3.33784878e-01 1.06219225e-01 1.26160955e+00
5.00440180e-01 9.60470319e-01 -9.71307933e-01 -3.77180390e-02
4.78999645e-01 2.21228763e-01 7.10246027e-01 -1.04875243e+00
-2.63943285e-01 -5.98524928e-01 7.54507035e-02 -1.02544963e+00
1.04542404e-01 -8.20957780e-01 2.17388272e-01 -1.92408895e+00
5.55165410e-01 -9.67875961e-03 1.06545582e-01 4.60436910e-01
-3.47072721e-01 2.03040108e-01 -8.59189853e-02 3.02531898e-01
-8.14540327e-01 2.27091834e-01 1.22997177e+00 -3.72988880e-01
-2.14253783e-01 -1.33413300e-01 -9.81443107e-01 1.17007756e+00
8.81645620e-01 -7.68630683e-01 1.24759279e-01 -6.60094559e-01
5.93784094e-01 -3.94980133e-01 4.91923392e-01 -5.87883949e-01
-3.70813757e-02 -4.88595128e-01 1.88736945e-01 -6.10089302e-01
-2.63158172e-01 -9.01640281e-02 -2.88444340e-01 6.84840739e-01
-8.35408509e-01 -2.61761606e-01 -4.49772030e-01 6.13718092e-01
6.11268692e-02 -1.81527615e-01 5.80355108e-01 -1.78973705e-01
-2.94690162e-01 -1.66992903e-01 -9.29068983e-01 3.65968823e-01
5.24318159e-01 4.13548917e-01 -1.30637813e+00 -4.16934848e-01
-4.98576134e-01 -2.06109777e-01 3.25678647e-01 4.43433821e-01
6.64418750e-03 -1.19098973e+00 -1.20524704e+00 -3.17758977e-01
4.18128148e-02 -3.95060837e-01 -1.44069567e-01 8.65741432e-01
-6.62869394e-01 5.81345141e-01 1.26533881e-01 -6.33759499e-02
-1.17290461e+00 5.38612485e-01 -7.72105157e-02 -5.97236693e-01
-8.29241395e-01 8.17256689e-01 3.21273983e-01 5.92571609e-02
-6.60900101e-02 -2.30966598e-01 -5.38748443e-01 9.66923460e-02
9.47552025e-01 2.15107217e-01 -2.49958917e-01 -7.43197799e-01
-3.97028685e-01 -1.11553170e-01 -4.01305497e-01 -5.20459656e-03
1.27982783e+00 -2.57962942e-02 -1.91871926e-01 1.81956977e-01
7.65538514e-01 5.80710828e-01 -6.38884902e-01 -2.40941972e-01
1.52432844e-01 -2.87862241e-01 -2.13565156e-02 -1.09147322e+00
-2.85763502e-01 5.35950005e-01 -2.17718318e-01 7.38074660e-01
3.79318625e-01 3.73012275e-01 7.44818270e-01 4.33336288e-01
7.55716162e-03 -1.27596819e+00 2.05508634e-01 8.67760122e-01
1.05016816e+00 -1.14819396e+00 2.12894365e-01 -4.95382041e-01
-4.16692108e-01 8.59962881e-01 3.10427189e-01 -1.12225860e-01
2.58498669e-01 9.83071551e-02 4.26070303e-01 -6.51479185e-01
-6.55368626e-01 -1.65806144e-01 6.05481446e-01 3.34949642e-01
8.93174171e-01 2.36278370e-01 -1.14529073e+00 6.27187490e-01
-3.77799898e-01 -4.82412636e-01 6.69161499e-01 7.36960292e-01
-9.87204909e-01 -7.30030298e-01 -4.80733961e-01 4.20030683e-01
-1.03820264e+00 -3.96109879e-01 -1.05329835e+00 6.31794393e-01
2.03752548e-01 1.14470506e+00 -2.65438884e-01 -1.45719528e-01
-2.91035205e-01 -1.91576540e-01 4.13284302e-01 -9.20369864e-01
-1.09190977e+00 2.00895503e-01 8.97962451e-01 -9.45305973e-02
-6.73872948e-01 -4.57785875e-01 -9.14105415e-01 -6.75318658e-01
-4.13831621e-01 4.84667003e-01 4.38657254e-01 1.30828917e+00
-1.70081869e-01 4.03869480e-01 3.24542850e-01 -4.37605232e-01
-7.28290856e-01 -1.26927006e+00 -6.07267655e-02 8.22148085e-01
4.38121438e-01 -4.28892761e-01 -3.51110935e-01 1.62677854e-01] | [8.541291236877441, 10.598146438598633] |
4eeb8d98-91a2-48ba-a6b9-95f95a8a58ea | robustly-pre-trained-neural-model-for-direct | 2004.06216 | null | https://arxiv.org/abs/2004.06216v1 | https://arxiv.org/pdf/2004.06216v1.pdf | Robustly Pre-trained Neural Model for Direct Temporal Relation Extraction | Background: Identifying relationships between clinical events and temporal expressions is a key challenge in meaningfully analyzing clinical text for use in advanced AI applications. While previous studies exist, the state-of-the-art performance has significant room for improvement. Methods: We studied several variants of BERT (Bidirectional Encoder Representations using Transformers) some involving clinical domain customization and the others involving improved architecture and/or training strategies. We evaluated these methods using a direct temporal relations dataset which is a semantically focused subset of the 2012 i2b2 temporal relations challenge dataset. Results: Our results show that RoBERTa, which employs better pre-training strategies including using 10x larger corpus, has improved overall F measure by 0.0864 absolute score (on the 1.00 scale) and thus reducing the error rate by 24% relative to the previous state-of-the-art performance achieved with an SVM (support vector machine) model. Conclusion: Modern contextual language modeling neural networks, pre-trained on a large corpus, achieve impressive performance even on highly-nuanced clinical temporal relation tasks. | ['Hua Xu', 'Jianfu Li', 'Hong Guan', 'Murthy Devarakonda'] | 2020-04-13 | null | null | null | null | ['temporal-relation-extraction'] | ['natural-language-processing'] | [ 5.91057479e-01 6.00824654e-01 -7.26653039e-01 -4.23189372e-01
-1.07504845e+00 -3.06035280e-01 6.92873895e-01 7.00506151e-01
-6.47553205e-01 8.48913252e-01 6.14956439e-01 -6.38528168e-01
-5.40279925e-01 -3.90937090e-01 -3.59185100e-01 -4.18537140e-01
-6.14351034e-01 8.54973137e-01 2.39135206e-01 -4.95369881e-01
-2.99001366e-01 1.51907668e-01 -6.49249196e-01 8.70844245e-01
2.27536663e-01 1.12839282e+00 -5.65930784e-01 7.15382457e-01
1.78277507e-01 1.55131161e+00 -3.99122089e-01 -4.69799876e-01
-1.22199975e-01 -3.41943502e-01 -1.38666010e+00 -6.70570076e-01
4.59201857e-02 -1.23363659e-02 -4.49413568e-01 4.19253945e-01
5.69327772e-01 -7.79210925e-02 3.13751668e-01 -6.70060754e-01
-4.92743254e-01 9.52862203e-01 -1.57339558e-01 9.04431403e-01
4.78874087e-01 -1.06802195e-01 1.05922425e+00 -1.69227526e-01
9.29801106e-01 8.21988702e-01 1.21606815e+00 7.22493887e-01
-1.33364582e+00 -7.67422557e-01 -1.42667904e-01 4.26321238e-01
-1.14846087e+00 -4.75960135e-01 1.92516223e-01 -4.31473851e-01
1.89194918e+00 3.46084803e-01 5.75375617e-01 1.34538782e+00
6.59695923e-01 5.01522183e-01 9.27006900e-01 -4.46872294e-01
-2.11220771e-01 4.62422986e-03 3.21907252e-01 7.03179240e-01
-3.34866762e-01 2.29626581e-01 -5.14316499e-01 -3.85564983e-01
3.68163049e-01 -2.29739442e-01 -1.93554580e-01 2.11651698e-01
-1.46633744e+00 7.97633588e-01 5.58717668e-01 7.41984129e-01
-4.61015284e-01 2.52797067e-01 1.07003951e+00 2.76798278e-01
6.40698612e-01 6.82333231e-01 -7.31408954e-01 -3.93236578e-01
-8.95460188e-01 4.78021055e-02 5.62940419e-01 6.74803257e-01
-5.14121115e-01 -4.28604960e-01 -3.32847387e-01 7.69703209e-01
-1.00739971e-01 -2.68848926e-01 8.73825967e-01 -4.69349980e-01
5.75099707e-01 6.19087100e-01 -4.86039490e-01 -7.81523526e-01
-9.33263123e-01 -4.84839678e-01 -7.93332338e-01 -4.37457204e-01
4.13419127e-01 -2.31368199e-01 -1.08795309e+00 1.83761454e+00
1.90471218e-03 1.97990537e-01 4.09518689e-01 3.98361772e-01
1.00100732e+00 3.27324837e-01 6.23513162e-01 -3.60516995e-01
1.65666449e+00 -8.71044636e-01 -1.12671793e+00 -2.16590181e-01
1.30495203e+00 -7.31117547e-01 3.38425040e-01 4.28639174e-01
-1.10093379e+00 3.79391164e-02 -1.11234462e+00 -3.74915063e-01
-5.00492811e-01 -6.57379925e-02 9.62894201e-01 3.20106447e-01
-1.07689226e+00 7.56644845e-01 -1.17094791e+00 -7.02725291e-01
5.12269080e-01 6.17592156e-01 -6.94438875e-01 2.01881856e-01
-1.75299072e+00 1.58961463e+00 5.61583459e-01 3.19319889e-02
-4.88543391e-01 -1.17499125e+00 -7.63946474e-01 -1.84641525e-01
2.19470590e-01 -6.96131825e-01 1.39442670e+00 -4.64889467e-01
-1.09757340e+00 1.09016335e+00 -2.45280247e-02 -1.07136762e+00
4.22175229e-01 -2.36440793e-01 -8.36440325e-01 2.55265325e-01
-2.13268995e-02 4.92586136e-01 -1.47395387e-01 -2.91834414e-01
-4.88988817e-01 -2.34684303e-01 -1.57674514e-02 -2.92498972e-02
-3.33769679e-01 3.44921321e-01 -3.13845277e-01 -7.37261891e-01
-2.66998559e-01 -1.07105124e+00 -4.83700722e-01 -7.75365978e-02
-3.82664502e-01 -5.01683414e-01 4.23953384e-01 -9.77361023e-01
1.66887808e+00 -1.89188337e+00 -1.12521071e-02 -7.28650466e-02
3.29835534e-01 2.78398126e-01 2.40949169e-01 5.32926440e-01
-8.75154436e-01 1.61441520e-01 -3.55797589e-01 -2.23395616e-01
-6.14031076e-01 2.45299175e-01 -1.38221964e-01 5.05068362e-01
4.46084917e-01 1.04169643e+00 -9.29366350e-01 -7.56604671e-01
-7.55689740e-02 5.64369261e-01 -5.39736867e-01 3.53565381e-04
1.63579136e-02 2.89887965e-01 -2.85890460e-01 3.47536713e-01
-1.25584632e-01 -6.19679272e-01 5.37599087e-01 -3.47719193e-01
2.64908612e-01 7.77764559e-01 -4.31144297e-01 2.05433607e+00
-4.04069573e-01 7.92708814e-01 -6.36352122e-01 -1.22755980e+00
6.38061285e-01 9.75457370e-01 1.06110024e+00 -8.34768772e-01
1.60796136e-01 1.40648052e-01 3.43553036e-01 -6.12188995e-01
1.13878548e-01 -7.34103024e-01 -2.88758613e-02 1.12509049e-01
5.03231548e-02 3.30734015e-01 -5.06883534e-03 2.58619547e-01
1.63056278e+00 6.22425564e-02 7.79481947e-01 -3.65888149e-01
4.07193244e-01 4.52088952e-01 5.94295204e-01 2.71528870e-01
-3.77974749e-01 4.50149596e-01 6.64599717e-01 -8.57094884e-01
-8.28415930e-01 -6.11236989e-01 -7.66369700e-01 7.14916229e-01
-4.22368169e-01 -8.25680852e-01 -3.22161824e-01 -8.77534449e-01
-2.00978950e-01 8.61328363e-01 -1.30127585e+00 -3.10118407e-01
-9.82040107e-01 -1.10078168e+00 1.10723507e+00 9.70149696e-01
2.45651603e-03 -1.04964578e+00 -6.96683228e-01 5.68096519e-01
-2.07410678e-01 -1.40736783e+00 -2.30698869e-01 6.46987677e-01
-9.17800367e-01 -1.13066268e+00 -5.00321686e-01 -5.60083389e-01
1.03866860e-01 -6.98351026e-01 1.24207592e+00 -2.66699642e-01
-5.23972452e-01 2.56282482e-02 -3.22172493e-01 -5.45712590e-01
-4.16676551e-01 1.53885171e-01 -1.38528213e-01 -6.76334679e-01
6.35057509e-01 -2.95107424e-01 -5.85417509e-01 -5.35086952e-02
-7.99755216e-01 1.82848483e-01 4.76253450e-01 1.13739157e+00
4.61404234e-01 -2.58999854e-01 6.57193482e-01 -1.30929029e+00
6.13807380e-01 -7.32132614e-01 6.25360832e-02 3.11766177e-01
-1.07495093e+00 2.23816171e-01 4.75948542e-01 -3.43322814e-01
-9.13209438e-01 -2.15418816e-01 -1.55618593e-01 -1.63038671e-01
2.02002183e-01 8.15195858e-01 7.76909590e-01 4.53419566e-01
1.11872876e+00 -9.80252847e-02 9.93043929e-02 -2.43073851e-01
3.22401196e-01 4.89623517e-01 4.96759206e-01 -3.21829408e-01
5.56637906e-02 4.28846687e-01 3.27724457e-01 -2.15092093e-01
-1.08614635e+00 -6.68021321e-01 -5.89188635e-01 3.41890603e-01
1.23850226e+00 -8.44110548e-01 -6.13499761e-01 -2.77766019e-01
-1.00742137e+00 -3.77960235e-01 -3.10787469e-01 6.07798100e-01
-4.14724112e-01 -1.35294348e-01 -9.73958552e-01 -1.99730054e-01
-7.17735827e-01 -1.11727667e+00 1.00030375e+00 -4.21277404e-01
-1.01284730e+00 -1.36706567e+00 3.70912045e-01 4.26322877e-01
5.21277249e-01 7.42418408e-01 1.20930457e+00 -1.10874414e+00
1.42858267e-01 -2.63146043e-01 -2.91961670e-01 -1.73924863e-01
2.89625287e-01 -3.14648032e-01 -7.34720170e-01 -5.17880507e-02
-1.49272382e-01 -4.46733981e-01 8.08222950e-01 2.74747640e-01
9.48166907e-01 -2.51821280e-01 -8.98498237e-01 3.74898344e-01
1.25094247e+00 7.08630681e-01 5.28425515e-01 4.90000486e-01
3.82717699e-01 4.82917041e-01 4.50434446e-01 3.01710129e-01
4.10771906e-01 8.65627706e-01 5.95329478e-02 -2.85287380e-01
-2.79512346e-01 9.98949856e-02 1.97732225e-02 6.64526522e-01
-1.86765417e-01 3.31465565e-02 -1.47899449e+00 8.54977190e-01
-1.96316922e+00 -7.84412086e-01 -1.19158803e-02 1.53263068e+00
1.49496913e+00 3.69568437e-01 -3.96237187e-02 1.60708293e-01
7.76861534e-02 9.06679372e-04 -3.95978957e-01 -7.15759158e-01
3.32305990e-02 7.61364758e-01 4.78364348e-01 3.13234538e-01
-1.17856967e+00 8.33649814e-01 6.63256025e+00 7.81135023e-01
-1.34184515e+00 2.62535036e-01 7.82935977e-01 -4.13473070e-01
1.07332282e-01 -2.53708780e-01 -6.50372982e-01 7.55588934e-02
1.73759854e+00 -2.89398819e-01 -3.46301161e-02 4.86716688e-01
5.65024130e-02 2.26897240e-01 -1.54659331e+00 8.95400167e-01
1.13277905e-01 -1.89317572e+00 -2.82907635e-01 -3.27477977e-02
4.53449428e-01 1.98013306e-01 -2.71103829e-02 4.95853990e-01
3.62998456e-01 -1.55108845e+00 2.93060154e-01 5.48546195e-01
1.03611839e+00 -3.41722369e-01 1.11802435e+00 -9.56007466e-02
-1.02414942e+00 -3.28251836e-03 3.28263521e-01 2.65548736e-01
2.34188735e-01 3.77383411e-01 -1.49059856e+00 7.70619452e-01
8.63640845e-01 1.09010935e+00 -4.49644059e-01 4.20789152e-01
1.54409483e-01 8.26977789e-01 -3.31707925e-01 6.03796393e-02
5.04905939e-01 4.88106281e-01 3.06291580e-01 1.71121895e+00
-1.87821779e-02 5.77145875e-01 -8.34994838e-02 2.47167662e-01
-4.85324562e-02 1.99722216e-01 -5.70976734e-01 -1.63032919e-01
9.25586373e-03 9.62148547e-01 -4.35787380e-01 -5.60398757e-01
-3.21933985e-01 5.46603680e-01 3.54930639e-01 -1.61308944e-02
-1.11338723e+00 -9.65808928e-02 3.97238702e-01 2.33706757e-01
-7.96805769e-02 1.70657158e-01 -4.62092966e-01 -7.63387620e-01
-1.58901528e-01 -1.00056016e+00 1.17308629e+00 -5.67612946e-01
-1.12639940e+00 1.06709909e+00 1.57682732e-01 -1.09942162e+00
-5.56887746e-01 -6.12987638e-01 6.44736290e-02 7.92014778e-01
-1.16500807e+00 -1.39744794e+00 2.03160375e-01 5.40212095e-01
4.60221857e-01 -1.65217116e-01 1.49306166e+00 6.23746216e-01
-5.51649392e-01 7.86188900e-01 -2.53858209e-01 2.31993705e-01
9.57188487e-01 -1.19534612e+00 -5.57116838e-03 3.61126214e-01
-1.04313411e-01 9.96837497e-01 4.54484165e-01 -5.21647990e-01
-9.68786657e-01 -1.10715616e+00 1.42903137e+00 -7.20878184e-01
9.88909960e-01 8.06618258e-02 -7.54671097e-01 1.21291339e+00
4.92419988e-01 1.04231626e-01 1.10112858e+00 4.33464676e-01
-3.68335009e-01 3.28565948e-02 -1.14844525e+00 4.71393526e-01
1.02394724e+00 -6.27306104e-01 -9.28819895e-01 6.64185584e-01
9.11811888e-01 -5.69219470e-01 -1.63775492e+00 8.67509484e-01
6.07519627e-01 -4.97803271e-01 9.92209256e-01 -1.39144218e+00
8.98016453e-01 -7.72042386e-03 -2.13324770e-01 -1.09591568e+00
-3.77160460e-01 -6.66068614e-01 -9.22026262e-02 7.40588725e-01
8.81201088e-01 -4.85737562e-01 6.05651081e-01 6.24210775e-01
-2.40375161e-01 -1.51076615e+00 -1.17147267e+00 -5.36055684e-01
4.16958362e-01 -5.16879797e-01 1.19338170e-01 1.46499121e+00
8.26749325e-01 6.33250356e-01 -5.60410507e-02 -2.22837746e-01
5.58863319e-02 -3.03859770e-01 -4.07045409e-02 -7.11293697e-01
-3.02082002e-01 -5.76107919e-01 -5.79156160e-01 -3.74661952e-01
-5.79676824e-03 -1.27942598e+00 -3.87263924e-01 -1.59625804e+00
2.85933256e-01 -5.43595135e-01 -6.43167138e-01 1.17479837e+00
-4.77080792e-02 1.70257390e-01 -1.88888267e-01 5.49159758e-02
-1.99748322e-01 4.45296317e-02 8.90254438e-01 -2.89275527e-01
-1.19838655e-01 -3.95158350e-01 -7.27791071e-01 4.08289313e-01
5.36235929e-01 -5.46270251e-01 -4.68817562e-01 -2.92646706e-01
1.89492539e-01 4.83443648e-01 9.74917039e-02 -7.98384905e-01
3.74694258e-01 1.15303904e-01 1.93464145e-01 -3.00485820e-01
4.49891776e-01 -6.91789985e-01 3.51576924e-01 9.38314915e-01
-1.01544416e+00 3.60611796e-01 8.01744163e-01 4.68454927e-01
-3.27267975e-01 1.94260702e-01 7.51478553e-01 -3.47262695e-02
-4.16755885e-01 2.47974228e-02 -1.49507105e-01 1.88001990e-01
1.23045051e+00 2.46108830e-01 -5.46076834e-01 -1.14343129e-01
-1.11752522e+00 1.94483325e-02 -5.04119456e-01 4.31138366e-01
2.66296476e-01 -1.15689600e+00 -7.52197206e-01 -4.08501863e-01
3.36967081e-01 -1.28246784e-01 7.09804222e-02 1.27783918e+00
-6.03419960e-01 1.10360408e+00 7.21137151e-02 -6.28184736e-01
-1.56538415e+00 6.80715501e-01 5.43662488e-01 -1.06485009e+00
-1.00640893e+00 1.08030438e+00 -4.59576212e-02 -1.27206057e-01
7.48367459e-02 -6.85239315e-01 -4.44932878e-01 2.23919094e-01
3.52205336e-01 -4.04859986e-03 5.56322336e-01 -3.91504079e-01
-9.15472746e-01 3.14232141e-01 -3.99997503e-01 -7.22200945e-02
1.59541881e+00 5.32943666e-01 -3.48083973e-02 3.83246243e-01
1.43179095e+00 -3.33509654e-01 -3.00035685e-01 -5.20588934e-01
3.36833626e-01 3.33202153e-01 2.30511099e-01 -1.31395042e+00
-8.81936371e-01 7.36676276e-01 7.85295844e-01 4.38780263e-02
1.08367682e+00 1.57452330e-01 8.75276685e-01 3.04736584e-01
3.49804610e-02 -7.34761298e-01 -1.08975284e-01 4.70673442e-01
9.54728365e-01 -1.01117432e+00 3.34738910e-01 -2.83510774e-01
-8.28318119e-01 1.10654056e+00 3.19017738e-01 -1.06172648e-03
9.47816014e-01 5.22024393e-01 2.78000981e-01 -5.62692583e-01
-1.22612965e+00 1.04604244e-01 4.64612961e-01 1.16858587e-01
1.05986464e+00 1.72426954e-01 -4.38807696e-01 5.95537722e-01
-3.44277501e-01 3.15722287e-01 1.25334471e-01 7.87704170e-01
4.37279552e-01 -1.19214678e+00 2.12885290e-01 5.61157346e-01
-1.05518425e+00 -6.54457748e-01 -9.72205251e-02 1.00308967e+00
5.75944595e-02 7.91766226e-01 8.95310100e-03 -3.12933952e-01
5.73351741e-01 2.80031741e-01 4.31367099e-01 -6.36563599e-01
-1.07003832e+00 -3.97072136e-02 9.38192248e-01 -8.55448484e-01
-7.05701590e-01 -7.91056335e-01 -1.50752783e+00 8.64160508e-02
-2.01282680e-01 -4.98469099e-02 4.09439653e-01 9.54760730e-01
5.09619832e-01 1.24497736e+00 -1.64104044e-01 1.25051156e-01
-3.47489029e-01 -1.14343131e+00 3.62770036e-02 6.03728890e-01
3.89946610e-01 -6.77299559e-01 1.37064964e-01 2.90968031e-01] | [8.501730918884277, 8.971166610717773] |
1005319c-5fb0-422d-b7d5-03bee6b0faa2 | typography-with-decor-intelligent-text-style | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Wang_Typography_With_Decor_Intelligent_Text_Style_Transfer_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Wang_Typography_With_Decor_Intelligent_Text_Style_Transfer_CVPR_2019_paper.pdf | Typography With Decor: Intelligent Text Style Transfer | Text effects transfer can dramatically make the text visually pleasing. In this paper, we present a novel framework to stylize the text with exquisite decor, which is ignored by the previous text stylization methods. Decorative elements pose a challenge to spontaneously handle basal text effects and decor, which are two different styles. To address this issue, our key idea is to learn to separate, transfer and recombine the decors and the basal text effect. A novel text effect transfer network is proposed to infer the styled version of the target text. The stylized text is finally embellished with decor where the placement of the decor is carefully determined by a novel structure-aware strategy. Furthermore, we propose a domain adaptation strategy for decor detection and a one-shot training strategy for text effects transfer, which greatly enhance the robustness of our network to new styles. We base our experiments on our collected topography dataset including 59,000 professionally styled text and demonstrate the superiority of our method over other state-of-the-art style transfer methods.
| [' Zongming Guo', ' Shuai Yang', ' Jiaying Liu', 'Wenjing Wang'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['text-effects-transfer'] | ['natural-language-processing'] | [ 4.09698814e-01 -2.11181313e-01 1.67909041e-01 -1.93135977e-01
-2.58314639e-01 -8.48166347e-01 6.64591730e-01 -4.78789032e-01
-8.32846239e-02 8.34700525e-01 5.24627268e-01 -6.12397045e-02
3.24001133e-01 -5.51824927e-01 -8.37474167e-01 -6.51991129e-01
7.56066442e-01 1.29618675e-01 2.80206829e-01 -4.00765598e-01
4.01974618e-01 3.91342103e-01 -1.23593330e+00 3.58254462e-01
1.04812777e+00 5.54099202e-01 4.56481516e-01 6.75924718e-01
-5.09424806e-01 5.65345287e-01 -9.74493623e-01 -5.81482768e-01
2.07277550e-03 -6.53485477e-01 -5.68415463e-01 2.36762539e-01
5.72535217e-01 -3.66494715e-01 -2.92469084e-01 9.44309711e-01
9.03283596e-01 8.89617652e-02 1.11144495e+00 -8.52084100e-01
-1.29496062e+00 6.18447542e-01 -9.27416205e-01 -1.23699479e-01
3.16213548e-01 1.34974644e-01 6.75521851e-01 -9.20818806e-01
7.80233622e-01 1.60731018e+00 5.70853174e-01 8.59837770e-01
-1.31189990e+00 -7.40593910e-01 3.89790326e-01 -2.33430743e-01
-8.00904095e-01 -3.58021885e-01 1.19919765e+00 -3.87575567e-01
1.53555796e-01 6.44139349e-02 5.89562476e-01 1.55785739e+00
1.60244703e-01 1.03203475e+00 1.24310613e+00 -5.28344393e-01
1.28647670e-01 1.64659336e-01 -2.92471141e-01 5.62865257e-01
9.66117084e-02 -4.31403816e-02 -6.92027986e-01 3.33795249e-01
1.10608244e+00 -6.61042705e-02 -3.11311960e-01 -3.77623916e-01
-1.24898958e+00 4.11017656e-01 4.29355770e-01 5.99426031e-02
-1.70741789e-02 2.63415515e-01 4.90196943e-01 2.45449692e-01
6.15174711e-01 5.25259912e-01 -2.60180742e-01 -1.51959434e-01
-5.89230537e-01 9.01676193e-02 6.95872784e-01 1.18527412e+00
5.97174168e-01 2.34757990e-01 -6.91259503e-01 1.14164400e+00
-1.25209363e-02 6.75943792e-01 2.17799172e-01 -4.18261439e-01
4.61659312e-01 4.70892727e-01 2.63836563e-01 -6.98691487e-01
2.31452342e-02 -1.87015861e-01 -8.90270412e-01 3.55970234e-01
4.25607741e-01 -6.35525405e-01 -9.93237495e-01 1.58080316e+00
2.41368622e-01 -2.73266528e-02 -1.78538471e-01 8.67953598e-01
8.60453546e-01 5.52637637e-01 4.82222475e-02 3.01988989e-01
1.38443589e+00 -1.15743220e+00 -1.02794015e+00 -2.24345133e-01
4.30863708e-01 -1.05556071e+00 1.76033938e+00 2.49740764e-01
-8.99647415e-01 -7.14706421e-01 -1.21703660e+00 -2.91362941e-01
-3.13720822e-01 4.00158048e-01 4.56006914e-01 5.27157605e-01
-7.59589672e-01 5.88877499e-01 -3.06730181e-01 -3.49391550e-01
6.58717155e-01 -2.24715304e-02 -9.57022011e-02 1.64620459e-01
-1.06231678e+00 6.14430428e-01 6.44387826e-02 -1.02326013e-01
-4.49514955e-01 -7.95469880e-01 -7.11014092e-01 -6.81901798e-02
2.31418759e-01 -8.43553483e-01 1.07810783e+00 -1.36267579e+00
-2.04303813e+00 9.11267519e-01 -2.12381244e-01 1.61596820e-01
7.41759837e-01 -5.41423798e-01 -3.60159606e-01 -1.20498732e-01
4.90973368e-02 7.46409893e-01 1.47960770e+00 -1.76665723e+00
-4.71080601e-01 1.41441487e-02 -1.13380611e-01 4.32964861e-01
-5.94239831e-01 -6.28667399e-02 -5.89969456e-01 -1.44426548e+00
-3.47579390e-01 -7.70134985e-01 2.71892011e-01 3.80199611e-01
-5.45853317e-01 1.91724729e-02 1.04676604e+00 -6.32280529e-01
1.08614588e+00 -2.30880499e+00 4.21492934e-01 -1.80115312e-01
2.72343606e-01 1.10034503e-01 -1.26872689e-01 4.63853210e-01
1.59213226e-02 1.48431048e-01 -1.32095963e-01 -6.66810751e-01
1.71773076e-01 6.49211407e-02 -5.80422521e-01 3.86804715e-02
3.71550232e-01 9.63905275e-01 -8.59918475e-01 -4.44296628e-01
1.08368143e-01 5.20877004e-01 -3.67860109e-01 3.71791959e-01
-3.54643494e-01 7.25476861e-01 -5.67009926e-01 4.84655470e-01
8.82377863e-01 -2.34062131e-02 -5.26966378e-02 -4.12440181e-01
-5.82515001e-02 3.99201028e-02 -9.47593331e-01 1.88053143e+00
-5.41515052e-01 6.65908337e-01 -1.51870161e-01 -3.31409633e-01
1.12720942e+00 7.24821165e-02 -1.57573089e-01 -7.14642525e-01
2.99342901e-01 -3.52076665e-02 -3.24368477e-01 -4.72975999e-01
6.66282415e-01 -4.21887696e-01 -5.87901883e-02 7.39281356e-01
-2.84595758e-01 -2.99005240e-01 2.02658605e-02 9.44839045e-02
6.50481701e-01 5.84693968e-01 2.57097296e-02 -4.47778583e-01
3.16716194e-01 -3.86250585e-01 3.22329134e-01 5.66201448e-01
8.04333910e-02 8.31685126e-01 7.08840251e-01 -3.01437616e-01
-1.09605300e+00 -9.77221906e-01 8.29814821e-02 1.37694144e+00
3.80953074e-01 -4.57982942e-02 -1.00782204e+00 -1.07201636e+00
5.21188155e-02 6.76473737e-01 -9.29685652e-01 -2.38392204e-01
-6.88637257e-01 -3.73555303e-01 5.33573568e-01 7.49912560e-01
7.82159269e-01 -1.34737766e+00 -1.85904026e-01 -3.32789379e-03
-9.94277373e-02 -8.18798304e-01 -1.13882565e+00 9.99903753e-02
-6.24536693e-01 -5.82879961e-01 -1.30806267e+00 -1.03008795e+00
8.10718358e-01 2.01231360e-01 1.02842999e+00 3.86327393e-02
4.75180857e-02 -5.01567759e-02 -4.88426596e-01 -5.38484514e-01
-5.10659754e-01 2.06533015e-01 -1.08487070e-01 2.27505729e-01
-1.74915977e-02 -5.65795839e-01 -6.74074352e-01 3.73397171e-01
-8.79647970e-01 4.81456012e-01 7.84210443e-01 8.93155336e-01
3.81444693e-01 -2.50074387e-01 6.85188770e-01 -1.19246244e+00
8.48960400e-01 -2.93397810e-02 -1.59507751e-01 1.33153215e-01
-1.51163861e-01 4.17856365e-01 8.41630340e-01 -8.68258893e-01
-1.61694121e+00 -1.09825365e-01 1.69859137e-02 -3.95197362e-01
-2.91513294e-01 -7.49636143e-02 -5.53016186e-01 -1.09766252e-01
7.33073592e-01 9.25856158e-02 -1.98356569e-01 -7.49215245e-01
6.62165105e-01 7.06454575e-01 4.78119314e-01 -9.47286189e-01
1.04087257e+00 6.24807358e-01 -1.74554124e-01 -6.89538777e-01
-9.60887432e-01 1.08291976e-01 -8.63230467e-01 -1.03697196e-01
7.88040638e-01 -6.50038660e-01 -4.18843359e-01 8.26864719e-01
-1.27600074e+00 -8.42631996e-01 -3.13250095e-01 -1.98882878e-01
-4.35072541e-01 4.36928838e-01 -6.30634546e-01 -4.34144199e-01
-3.29476178e-01 -9.54951286e-01 1.26870334e+00 2.82776177e-01
-3.47976834e-01 -1.02426541e+00 2.06891336e-02 3.56229693e-02
3.74496818e-01 2.78842717e-01 1.09619808e+00 -5.71379252e-03
-1.61335886e-01 2.19572976e-01 -6.67331159e-01 1.86006337e-01
6.60824597e-01 -1.47525072e-02 -1.05554426e+00 -1.99138924e-01
-3.33682179e-01 -1.44291669e-01 9.15302753e-01 1.85738817e-01
1.27405667e+00 -1.57181844e-01 -3.90538722e-01 8.40837538e-01
1.06769013e+00 2.77160794e-01 7.16301739e-01 4.13082719e-01
1.23325098e+00 4.87561882e-01 3.67840528e-01 4.82019722e-01
6.56047389e-02 5.23937881e-01 -3.33321430e-02 -4.68359768e-01
-7.67611086e-01 -6.35336518e-01 2.77949303e-01 5.90659440e-01
-6.06549159e-02 -5.68428099e-01 -3.59222949e-01 3.80847454e-01
-1.56014121e+00 -5.88815629e-01 8.79850686e-02 1.82807934e+00
1.08453441e+00 3.30493182e-01 3.16187032e-02 3.48954983e-02
9.71811593e-01 2.13457763e-01 -7.08065569e-01 -4.66308892e-01
-2.95748532e-01 2.03887925e-01 2.88529098e-01 3.15285504e-01
-1.09836602e+00 1.38880849e+00 6.32074308e+00 8.98020208e-01
-1.30924475e+00 -2.34148577e-01 4.46581334e-01 1.54924959e-01
-5.23174703e-01 -2.68522561e-01 -6.95876658e-01 6.42775178e-01
1.16157085e-01 4.93778251e-02 5.24520099e-01 5.99882066e-01
2.66754270e-01 1.46909967e-01 -1.16205430e+00 9.82863545e-01
2.47752935e-01 -1.16080928e+00 4.95927781e-01 -2.49366432e-01
1.02424896e+00 -8.04125130e-01 4.39895213e-01 1.71060562e-01
2.51071662e-01 -8.67869496e-01 9.91963923e-01 5.39553642e-01
1.35703456e+00 -6.33200884e-01 2.30381683e-01 -1.67547926e-01
-1.11900103e+00 1.79758891e-01 -2.86813796e-01 -9.44612101e-02
8.55196491e-02 4.93358374e-01 -4.22800362e-01 3.90232563e-01
5.88296711e-01 1.08672822e+00 -6.18340313e-01 7.46385813e-01
-6.58369839e-01 5.75708151e-01 2.14697689e-01 -2.87838340e-01
-4.58713621e-02 -1.01953007e-01 6.40150070e-01 1.46464324e+00
3.47373843e-01 -1.81420445e-01 -1.83454305e-01 1.15190959e+00
-5.12743354e-01 7.67245069e-02 -5.68931639e-01 -1.18164951e-02
3.71049553e-01 1.23333442e+00 -7.07803547e-01 -4.05826807e-01
-3.17149580e-01 1.69802344e+00 4.37138319e-01 6.69256926e-01
-7.70463586e-01 -7.27882147e-01 5.14725327e-01 9.89200398e-02
5.24814367e-01 2.71987694e-04 -8.88153076e-01 -1.27298963e+00
-5.79967024e-03 -7.53441811e-01 -8.94734338e-02 -1.18726766e+00
-1.64109492e+00 3.78540456e-01 -3.10934126e-01 -1.36664498e+00
3.45099300e-01 -7.75851667e-01 -1.07050550e+00 1.11993659e+00
-1.38549662e+00 -1.32818484e+00 -4.57008392e-01 3.65596652e-01
7.58866787e-01 -3.18849161e-02 4.73094523e-01 1.45823270e-01
-5.02756715e-01 7.82740355e-01 2.06659943e-01 9.68593359e-02
1.50589883e+00 -1.54940367e+00 7.10875154e-01 7.88153470e-01
-3.72059196e-01 5.45042813e-01 7.96652675e-01 -9.20030236e-01
-1.23835194e+00 -1.21983039e+00 3.39772880e-01 -4.39591140e-01
5.60856521e-01 -7.17192054e-01 -9.99739766e-01 6.07703030e-01
5.00238657e-01 -4.75933969e-01 4.36616629e-01 1.26329035e-01
-4.37084913e-01 -8.72791335e-02 -9.40087855e-01 1.20767665e+00
1.31534469e+00 -2.78844446e-01 -7.49868870e-01 -8.39028973e-03
7.56599069e-01 -5.03991842e-01 -3.89887512e-01 -4.88030445e-03
6.46768749e-01 -8.18985701e-01 6.68852329e-01 -3.62887174e-01
8.67048621e-01 -3.83871943e-01 3.43239546e-01 -1.81767738e+00
-4.82790858e-01 -8.58695686e-01 -3.02049201e-02 1.54156578e+00
2.18346804e-01 -5.42182088e-01 6.41153514e-01 1.62830353e-01
-3.52062970e-01 -3.61017168e-01 -2.00069696e-01 -5.03068745e-01
2.89716512e-01 1.90758392e-01 8.34685564e-01 9.65350211e-01
-2.72667408e-01 6.48695469e-01 -6.87508583e-01 -2.36528516e-01
5.65574706e-01 2.24159330e-01 9.89884496e-01 -1.13955057e+00
-3.17092746e-01 -6.17354035e-01 1.83568314e-01 -1.38439417e+00
1.27519220e-01 -5.64149261e-01 1.19047530e-01 -1.49650216e+00
1.89062059e-01 -3.40682685e-01 -6.92600459e-02 3.61464292e-01
-5.31751752e-01 2.23416045e-01 6.58955500e-02 6.53282627e-02
-2.16949090e-01 8.94882798e-01 2.12504053e+00 -1.70317501e-01
-3.75474006e-01 -1.98445618e-01 -1.10573053e+00 5.79993188e-01
7.11081505e-01 -1.40023351e-01 -4.76928055e-01 -6.23621702e-01
1.97119460e-01 -3.36361974e-01 2.30539799e-01 -7.35943735e-01
-1.49642855e-01 -1.41623110e-01 7.57153094e-01 -5.64582407e-01
1.73666656e-01 -5.48983693e-01 -3.04150790e-01 2.31942281e-01
-3.88132691e-01 -2.42862001e-01 4.24114048e-01 5.96771002e-01
1.61743596e-01 -5.01102209e-02 9.64418769e-01 9.34477225e-02
-6.19419217e-01 1.99206829e-01 -2.59342372e-01 1.65454879e-01
6.83613777e-01 -2.91373044e-01 -4.27703679e-01 -3.65482450e-01
-4.37564105e-01 -2.79629137e-02 7.64416933e-01 5.89786947e-01
4.35142279e-01 -1.46036220e+00 -5.36364138e-01 2.89020747e-01
1.74900353e-01 3.41758411e-03 3.91641855e-01 3.54807109e-01
-4.78454024e-01 -1.45477220e-01 -4.42032248e-01 -1.46868870e-01
-1.09624386e+00 7.56061733e-01 1.91973224e-01 8.39892924e-02
-9.45167601e-01 7.67853379e-01 8.46852541e-01 -2.43281990e-01
1.05547287e-01 -2.99542159e-01 -1.66355744e-01 -9.47090015e-02
4.93477374e-01 2.19875008e-01 -3.35067391e-01 -2.22727656e-01
2.23116681e-01 8.33625138e-01 -3.14119726e-01 -5.50079383e-02
1.02554417e+00 -3.44318092e-01 -1.24842371e-03 5.89359224e-01
8.83706987e-01 2.75950611e-01 -1.79046464e+00 -1.66730538e-01
-4.31959063e-01 -4.45330322e-01 -2.53379047e-01 -9.94884312e-01
-8.89709771e-01 1.06798720e+00 3.72826576e-01 -6.95330277e-02
1.06064439e+00 -1.73388228e-01 9.36664045e-01 2.15558589e-01
-9.32885632e-02 -1.19480228e+00 5.45244575e-01 5.25119662e-01
1.16008389e+00 -1.09189725e+00 -3.11995327e-01 -5.50556481e-01
-8.88704360e-01 1.02476084e+00 8.55217695e-01 -2.59840459e-01
3.83677959e-01 4.15919274e-01 2.15282366e-01 -2.85905451e-02
-4.05504256e-01 1.98750407e-03 4.56064373e-01 6.37516320e-01
5.31307399e-01 3.83018143e-02 -1.04578488e-01 6.62845910e-01
-2.89063513e-01 -7.82196373e-02 4.09952044e-01 8.66052806e-01
-3.71518940e-01 -1.12365782e+00 -4.14115548e-01 1.83926791e-01
-2.56725937e-01 -2.72022754e-01 -7.54343927e-01 5.99488437e-01
1.39372703e-02 5.17383993e-01 -4.39314805e-02 -3.08760017e-01
5.48821032e-01 1.11616008e-01 5.84760308e-01 -5.12162268e-01
-4.64590758e-01 2.63430715e-01 -1.33013964e-01 -8.20323005e-02
-1.64781153e-01 -4.34342742e-01 -1.07983816e+00 -3.30727100e-01
-2.43667495e-02 -3.01673889e-01 3.68667811e-01 8.94254804e-01
3.22336465e-01 9.81759787e-01 6.49037540e-01 -1.10829985e+00
-1.32777065e-01 -1.15627933e+00 -7.15130746e-01 6.42944574e-01
2.91865826e-01 -8.53771448e-01 -2.86877453e-01 5.51696420e-01] | [11.591793060302734, -0.4471655786037445] |
1ee51f1c-47db-4fe6-8abd-f412b285736b | low-power-object-counting-with-hierarchical | 2007.01369 | null | https://arxiv.org/abs/2007.01369v1 | https://arxiv.org/pdf/2007.01369v1.pdf | Low-Power Object Counting with Hierarchical Neural Networks | Deep Neural Networks (DNNs) can achieve state-of-the-art accuracy in many computer vision tasks, such as object counting. Object counting takes two inputs: an image and an object query and reports the number of occurrences of the queried object. To achieve high accuracy on such tasks, DNNs require billions of operations, making them difficult to deploy on resource-constrained, low-power devices. Prior work shows that a significant number of DNN operations are redundant and can be eliminated without affecting the accuracy. To reduce these redundancies, we propose a hierarchical DNN architecture for object counting. This architecture uses a Region Proposal Network (RPN) to propose regions-of-interest (RoIs) that may contain the queried objects. A hierarchical classifier then efficiently finds the RoIs that actually contain the queried objects. The hierarchy contains groups of visually similar object categories. Small DNNs are used at each node of the hierarchy to classify between these groups. The RoIs are incrementally processed by the hierarchical classifier. If the object in an RoI is in the same group as the queried object, then the next DNN in the hierarchy processes the RoI further; otherwise, the RoI is discarded. By using a few small DNNs to process each image, this method reduces the memory requirement, inference time, energy consumption, and number of operations with negligible accuracy loss when compared with the existing object counters. | ['Yung-Hsiang Lu', 'George K. Thiruvathukal', 'Sara Aghajanzadeh', 'Caleb Tung', 'Shreya Ghosh', 'Isha Ghodgaonkar', 'Abhinav Goel'] | 2020-07-02 | null | null | null | null | ['object-counting'] | ['computer-vision'] | [ 2.06487969e-01 -1.82048649e-01 -2.85599321e-01 -3.04439485e-01
-1.37977526e-01 -2.80341804e-01 7.13636950e-02 4.77574706e-01
-8.16978276e-01 4.33103681e-01 -3.59875530e-01 -8.95519033e-02
3.16787064e-01 -1.28060937e+00 -6.86976731e-01 -5.48401654e-01
1.32697806e-01 4.32596058e-01 1.18589318e+00 5.53383827e-01
2.30487227e-01 8.68375480e-01 -1.83662176e+00 3.60816121e-01
3.17656755e-01 1.55477154e+00 4.82510865e-01 6.21948361e-01
-2.90864408e-01 1.07005203e+00 -9.58434224e-01 -4.56831492e-02
2.68450260e-01 -8.80649388e-02 -5.83997428e-01 -3.22516739e-01
8.16163421e-01 -9.40834641e-01 -5.64731956e-01 1.13849664e+00
2.90854305e-01 1.59187213e-01 3.21166784e-01 -1.16793525e+00
-2.70361364e-01 4.47913259e-01 -6.63401008e-01 7.53750026e-01
-2.85905600e-01 4.05626111e-02 8.95258963e-01 -7.56971717e-01
2.57956862e-01 1.20962620e+00 4.92466390e-01 5.68120360e-01
-9.52207446e-01 -9.45753038e-01 1.06962338e-01 1.54481292e-01
-1.33422720e+00 -3.56970221e-01 2.59241998e-01 -6.96584433e-02
1.06632137e+00 1.06573537e-01 8.80905330e-01 3.52564275e-01
1.29741475e-01 6.66749060e-01 4.89737809e-01 -2.35105038e-01
6.57691300e-01 -2.41902828e-01 3.72102022e-01 7.85705149e-01
9.42154408e-01 -2.77463555e-01 -6.35833859e-01 2.64657736e-02
9.48930502e-01 5.50184071e-01 3.50267798e-01 -5.29960729e-02
-1.03798699e+00 8.58948052e-01 9.70847011e-01 1.73861831e-01
-5.46589971e-01 8.99287045e-01 5.83949924e-01 -4.45559382e-01
1.13579370e-01 7.84776360e-02 -2.27755234e-01 1.91063777e-01
-1.14594984e+00 1.97123364e-01 5.15678883e-01 9.17572320e-01
7.32609808e-01 -1.31431073e-01 -2.92112708e-01 6.40724897e-01
2.33826593e-01 3.52740973e-01 -1.04887299e-02 -1.20511913e+00
4.53385204e-01 1.24592543e+00 -1.05386890e-01 -9.51175809e-01
-3.86447579e-01 -2.05367610e-01 -8.48234057e-01 3.10393333e-01
3.59710008e-01 2.63981014e-01 -1.04514635e+00 1.33170128e+00
5.36991060e-01 -8.87195393e-02 -3.07400763e-01 7.81339586e-01
9.75880921e-01 6.46655858e-01 3.79710525e-01 2.36526623e-01
1.80726707e+00 -8.81132185e-01 -2.67692566e-01 -4.58692908e-01
1.30562559e-01 -3.43305349e-01 7.58210063e-01 1.05890244e-01
-1.23433626e+00 -7.42719114e-01 -1.19670558e+00 -5.85881591e-01
-4.13988471e-01 1.91009119e-01 7.75699139e-01 3.78352761e-01
-1.03263128e+00 4.12426800e-01 -1.05391347e+00 -2.28867382e-01
1.01396382e+00 6.59778476e-01 1.79155022e-01 -1.81425035e-01
-7.21186280e-01 5.68343401e-01 6.33851171e-01 -1.13138901e-02
-1.18646193e+00 -5.24102271e-01 -8.56669486e-01 4.82961774e-01
2.86780924e-01 -4.98879850e-01 1.36950684e+00 -6.98931336e-01
-7.01172292e-01 6.82266116e-01 -3.91819835e-01 -6.21122420e-01
2.21306801e-01 5.89185730e-02 -1.15227722e-01 5.20982146e-01
3.53574365e-01 1.31910574e+00 6.96293771e-01 -8.17093253e-01
-1.18637693e+00 -5.66407025e-01 2.44513318e-01 -3.09102368e-02
-3.77495080e-01 8.58919397e-02 -8.68678331e-01 -2.36313567e-01
4.51240301e-01 -6.50202811e-01 -1.35720938e-01 6.86816990e-01
-3.05056095e-01 -6.25894606e-01 1.16355169e+00 -2.04611674e-01
1.02014101e+00 -2.13575768e+00 -8.03848743e-01 -4.50484008e-02
6.33576214e-01 2.03891501e-01 9.83479917e-02 -4.58751649e-01
4.38507229e-01 -2.61601317e-03 5.61567992e-02 -2.09286198e-01
-4.49448317e-01 3.32617313e-01 -1.16596140e-01 4.06290054e-01
1.60371423e-01 7.21618593e-01 -8.42751861e-01 -8.57212007e-01
3.43233138e-01 3.25768769e-01 -5.74471056e-01 -1.30628377e-01
-2.42620990e-01 -1.66183382e-01 -3.38164270e-01 7.79295743e-01
7.27065027e-01 -4.00040329e-01 1.54956337e-02 -5.22168517e-01
-2.28728369e-01 5.50591886e-01 -1.09648132e+00 8.58900428e-01
-1.13314018e-01 8.25954378e-01 9.63421017e-02 -9.36775625e-01
7.96187222e-01 -1.58922836e-01 4.25676852e-01 -8.88805091e-01
3.16156864e-01 1.04697131e-01 9.01720524e-02 -6.57922328e-02
5.88056087e-01 1.47477046e-01 -1.66554838e-01 3.35703194e-01
-6.38262331e-02 1.72545403e-01 4.08500016e-01 1.18268371e-01
1.39052033e+00 -4.35319960e-01 5.38688481e-01 -8.84351414e-03
2.19799057e-01 1.65631890e-01 7.17545033e-01 1.12373006e+00
-5.40811658e-01 1.69455379e-01 3.26104909e-01 -8.33092690e-01
-1.01391244e+00 -1.16677058e+00 -1.25115648e-01 1.46458244e+00
3.93587112e-01 -2.57997680e-02 -7.66762733e-01 -4.81067210e-01
1.49630040e-01 3.00796270e-01 -4.72550035e-01 3.05241141e-02
-6.85873866e-01 -4.31142181e-01 6.57836676e-01 9.86407578e-01
1.05419064e+00 -1.29643524e+00 -1.66580403e+00 1.70449719e-01
7.27783144e-02 -1.33064270e+00 -2.89895862e-01 5.09438634e-01
-1.23664999e+00 -1.01248527e+00 -2.99985796e-01 -9.45520043e-01
9.55122054e-01 5.46575904e-01 1.07104230e+00 3.40012997e-01
-5.80367029e-01 -1.39450982e-01 5.96690066e-02 -6.51632845e-01
-5.05808294e-02 -6.72011822e-02 -7.26535870e-03 -5.38966060e-01
8.02429438e-01 -9.64528471e-02 -9.11722362e-01 3.57288092e-01
-6.74599111e-01 -1.92802727e-01 5.36849678e-01 3.76468450e-01
9.98810768e-01 3.33774567e-01 3.68204653e-01 -6.01008415e-01
3.96695407e-03 -2.38748550e-01 -1.06337082e+00 -1.81003466e-01
-1.30713107e-02 -8.47117901e-02 6.85032904e-01 -5.57725608e-01
-6.04653776e-01 3.77654821e-01 3.15005124e-01 -4.92127806e-01
-5.16086183e-02 -1.52910948e-01 -7.98543245e-02 1.50753826e-01
4.37773317e-01 1.03469454e-01 -4.39447850e-01 -1.91740423e-01
6.47139400e-02 4.74174589e-01 6.86950028e-01 -7.09937364e-02
2.31725082e-01 8.17463994e-01 1.94831401e-01 -6.06752515e-01
-8.50081146e-01 -4.79926437e-01 -5.53115427e-01 -2.97017753e-01
1.03660178e+00 -9.51674283e-01 -1.24911761e+00 2.48041153e-01
-1.40728176e+00 -1.78005606e-01 -3.91304046e-01 3.56057018e-01
1.40967285e-02 -1.04456134e-01 -7.59041131e-01 -1.00022960e+00
-5.18031180e-01 -1.09813178e+00 8.65344107e-01 7.12902188e-01
-1.75017059e-01 -2.41696805e-01 -5.96102417e-01 1.70475557e-01
3.10668886e-01 -1.14460863e-01 1.06517863e+00 -5.82484782e-01
-1.00934803e+00 -3.55753273e-01 -8.26021016e-01 2.05108777e-01
-1.16507575e-01 -1.35775656e-01 -9.56899107e-01 -6.97598904e-02
-1.14632510e-01 -1.99347779e-01 1.07958400e+00 7.39561379e-01
1.79783440e+00 -2.56278157e-01 -7.84162641e-01 3.28758478e-01
1.52970183e+00 5.28279305e-01 4.63386714e-01 1.31522253e-01
6.45700336e-01 2.02276871e-01 2.44182840e-01 3.12676519e-01
3.03218096e-01 2.98551530e-01 7.69649267e-01 -2.39286184e-01
-2.99685329e-01 -1.57682952e-02 6.15274645e-02 3.88659835e-01
2.50956655e-01 -3.34464699e-01 -8.78396153e-01 8.25887322e-01
-1.41543663e+00 -8.85499597e-01 -9.26516652e-02 2.25308943e+00
6.77219570e-01 6.09878361e-01 6.06430732e-02 1.16227470e-01
9.74089205e-01 1.43821687e-01 -1.17024457e+00 -2.66153336e-01
2.06164956e-01 2.99608469e-01 8.79484653e-01 -4.83069532e-02
-1.06910956e+00 9.06598091e-01 6.20181084e+00 5.13464212e-01
-8.76751482e-01 1.97644874e-01 8.68311882e-01 -5.34021974e-01
4.36012596e-01 -1.97130889e-01 -1.40010965e+00 3.54446024e-01
7.74630964e-01 3.88598353e-01 3.05711538e-01 1.23722565e+00
1.52971536e-01 -8.59384716e-01 -1.34534013e+00 9.05720472e-01
3.71587500e-02 -1.35220301e+00 4.42295447e-02 9.87784714e-02
4.89176959e-01 3.46280456e-01 -2.37178490e-01 2.48858780e-01
4.29205000e-01 -7.99945712e-01 8.14236224e-01 9.89657715e-02
7.95765162e-01 -7.49159217e-01 7.26037502e-01 3.77634645e-01
-1.44100785e+00 -2.85843819e-01 -9.12226737e-01 -1.15329728e-01
-1.32688984e-01 9.73787248e-01 -6.86487377e-01 -4.40397471e-01
1.21772349e+00 1.71174742e-02 -4.12817061e-01 1.13966787e+00
4.87544872e-02 4.71330971e-01 -6.51899457e-01 -3.76774371e-01
1.63838550e-01 2.12315217e-01 3.19778919e-02 1.07285464e+00
1.66133314e-01 2.68357724e-01 2.68748552e-01 1.08466709e+00
-6.99753225e-01 -3.91008705e-01 -3.55617791e-01 2.65958637e-01
1.00273693e+00 1.28541219e+00 -1.42700040e+00 -9.15309429e-01
-4.33332950e-01 8.43185067e-01 4.18298244e-01 -1.07976213e-01
-7.87565231e-01 -5.99968255e-01 4.49232817e-01 2.40020007e-01
6.43888116e-01 -6.90202937e-02 -6.21237040e-01 -3.84031326e-01
1.16740309e-01 -3.50176215e-01 4.19986993e-01 -6.54182017e-01
-9.11260426e-01 3.70407790e-01 -2.06308886e-01 -9.42495465e-01
3.11188221e-01 -6.95146680e-01 -4.81205553e-01 5.57059169e-01
-1.20888913e+00 -6.45891786e-01 -7.43813813e-01 4.36179459e-01
5.58810651e-01 1.77472681e-01 5.08432508e-01 3.98111463e-01
-5.29839218e-01 4.39406425e-01 -2.92146176e-01 6.66260660e-01
9.41429511e-02 -8.71141195e-01 4.00942504e-01 7.14574337e-01
-9.57316160e-02 6.16860032e-01 9.14938524e-02 -7.76992202e-01
-1.07297862e+00 -1.37299478e+00 7.87248194e-01 -1.97643310e-01
2.03198731e-01 -6.15667164e-01 -7.88215101e-01 3.95634890e-01
-3.15666348e-01 7.03861475e-01 4.03411418e-01 -3.01647902e-01
-3.19840729e-01 -4.01385456e-01 -1.42458534e+00 5.33609927e-01
1.02819419e+00 -4.02022928e-01 -3.03044409e-01 2.41836920e-01
6.98443472e-01 -2.35217050e-01 -3.42603028e-01 1.83788449e-01
6.10748529e-01 -9.09789801e-01 9.61744905e-01 -1.56514540e-01
4.05757368e-01 -5.41969121e-01 -2.23791435e-01 -4.62278038e-01
-3.48816842e-01 2.90432572e-01 -4.55894917e-01 8.03090334e-01
8.48889202e-02 -2.27758601e-01 1.09148657e+00 5.15777051e-01
9.23552439e-02 -7.69217789e-01 -1.11679828e+00 -8.00338507e-01
-3.98585439e-01 -4.61424500e-01 5.46826899e-01 3.17180157e-01
-3.15250129e-01 4.17425960e-01 3.74522537e-01 2.74620324e-01
7.33631253e-01 -1.04475588e-01 4.38001454e-01 -1.52323878e+00
3.36251915e-01 -4.87547219e-01 -6.20179176e-01 -1.21211183e+00
-2.95935065e-01 -5.93905509e-01 3.49266559e-01 -1.88184559e+00
4.58499521e-01 -5.98908544e-01 -2.80436277e-01 8.01642835e-01
-3.38219218e-02 5.30394435e-01 2.85990268e-01 2.67187953e-01
-9.99336302e-01 -1.32395402e-01 9.04197931e-01 -1.99588254e-01
-2.13507026e-01 -3.47056352e-02 -4.45674837e-01 1.05697894e+00
7.60371387e-01 -7.96194434e-01 -1.09815232e-01 -5.42062402e-01
7.44247437e-03 -1.47092432e-01 6.26228690e-01 -1.63558412e+00
6.47399724e-01 9.46488380e-02 1.11350143e+00 -1.30669093e+00
3.35517466e-01 -9.60008502e-01 -3.53841931e-01 9.31031764e-01
-2.27841616e-01 1.00542970e-01 4.80786830e-01 3.94523591e-01
4.94253077e-02 -4.16827530e-01 1.12378299e+00 -3.63009363e-01
-6.26996756e-01 3.41053009e-01 -5.94641984e-01 -1.45471618e-01
1.13448203e+00 -6.37398839e-01 -3.41468573e-01 1.64302036e-01
-1.95491597e-01 1.33328915e-01 1.89528704e-01 6.57225624e-02
6.41483366e-01 -1.10525465e+00 -1.95275739e-01 1.15605310e-01
-7.90899619e-02 6.71099961e-01 1.17978014e-01 2.94858426e-01
-6.51448965e-01 6.13961935e-01 -1.59080774e-01 -7.08140194e-01
-1.23072338e+00 6.19038284e-01 4.27090406e-01 -3.19137901e-01
-5.95398247e-01 1.07273126e+00 3.35773110e-01 8.23767763e-03
4.34842199e-01 -8.70818853e-01 -5.17900214e-02 1.20159656e-01
9.09316421e-01 6.92104816e-01 6.16800413e-02 -3.80509496e-01
-7.55612135e-01 2.76867747e-01 -3.65903467e-01 1.07012749e-01
1.08433008e+00 2.60513961e-01 -2.05699161e-01 2.43152127e-01
1.07423580e+00 -3.37123573e-01 -1.43714654e+00 -2.09634781e-01
-7.75728226e-02 -3.05497408e-01 4.40340817e-01 -5.99035978e-01
-1.40871584e+00 7.76685894e-01 7.33227015e-01 1.10560536e-01
1.18971968e+00 2.93685079e-01 8.46039236e-01 4.32505518e-01
5.11935353e-01 -1.21748066e+00 2.87782252e-01 5.86826980e-01
3.08105350e-01 -9.72883046e-01 2.90945500e-01 -2.56042272e-01
-1.69796750e-01 1.02736640e+00 1.16987014e+00 -2.07290888e-01
4.37801629e-01 5.70336342e-01 -4.62791473e-01 -4.71748114e-01
-5.03009975e-01 -1.19614065e-01 -1.47593617e-01 4.18935359e-01
-2.17803959e-02 1.49917841e-01 8.07691216e-02 3.10129195e-01
-1.27829323e-02 3.06135744e-01 2.58775741e-01 9.78450954e-01
-1.09300232e+00 -2.39917576e-01 -4.74169910e-01 9.83227730e-01
-5.07154286e-01 -2.11848348e-01 -9.12800059e-02 4.97536629e-01
4.96634245e-01 8.36647272e-01 1.09252095e+00 -4.74354327e-02
2.69276828e-01 -4.10579801e-01 3.03832650e-01 -8.04199994e-01
-6.05525732e-01 -2.04880401e-01 -3.29098314e-01 -6.21596396e-01
-1.80876285e-01 -2.96839982e-01 -1.83686590e+00 -4.12753284e-01
-3.67328644e-01 -5.68949997e-01 7.38564789e-01 6.70351505e-01
2.62917966e-01 7.07006633e-01 1.44627035e-01 -7.87560940e-01
-1.35816976e-01 -7.45160282e-01 -3.60518068e-01 -1.64278910e-01
4.13042516e-01 -6.49501443e-01 -9.03936289e-03 -1.12084076e-01] | [8.838510513305664, -0.09952455759048462] |
51fe47d1-8104-41d4-b995-c4e4c98bc94a | automotive-radar-mutual-interference | 2307.04326 | null | https://arxiv.org/abs/2307.04326v1 | https://arxiv.org/pdf/2307.04326v1.pdf | Automotive Radar Mutual Interference Mitigation Based on Hough Transform in Time-Frequency Domain | With the development of autonomous driving technology, automotive radar has received unprecedented attention due to its day-and-night and all-weather working capability. It is worthwhile to note that more and more vehicles are equipped with automotive radars, resulting in mutual interference between radars. The interference reduces radar target detection performance, making perception information unreliable. In this paper, a novel interference mitigation method based on power-weighted Hough transform is proposed for solving the radar mutual interference and improving the safety of autonomous driving systems. Firstly, the frequency modulation characteristics of interference signals and target echo signals are analyzed, and differences between the two signals are introduced. Secondly, based on the straight line detection technique, the power of the mutual interference signal in time-frequency domain is accumulated, and the accurate position of the interference is located. Finally, the target echo is recovered by autoregressive model. Compared with existing state-of-the-art methods, the proposed method has the ability to retain more useful signals after the interference mitigation, and achieve better interference detection robustness under low signal-to-noise ratio conditions. Simulation experiments and real scenario experiments verify the effectiveness of the proposed method and show its superiority. | ['Lianying Ji', 'Weichuan Zhang', 'Yanbing Li'] | 2023-07-10 | null | null | null | null | ['autonomous-driving', 'line-detection'] | ['computer-vision', 'computer-vision'] | [ 3.07061702e-01 -4.59663242e-01 1.37000546e-01 -1.00744404e-01
-2.55783558e-01 -1.01075910e-01 4.83712345e-01 -6.33496523e-01
-3.51432920e-01 6.62708938e-01 -8.26399103e-02 -2.09166706e-01
-5.20005047e-01 -7.48102248e-01 1.64750759e-02 -9.91343737e-01
3.54756676e-02 -2.84619540e-01 4.31829989e-01 -3.15794587e-01
2.20241904e-01 5.08777738e-01 -1.68938911e+00 -4.37432468e-01
1.21549010e+00 1.13962138e+00 3.20726067e-01 1.87966809e-01
1.38387576e-01 5.67279696e-01 -9.36156094e-01 2.38919966e-02
3.87679607e-01 -3.16143185e-01 5.42957664e-01 -1.65699884e-01
-1.33115724e-01 -2.28813171e-01 -8.72194946e-01 1.27342844e+00
4.08797055e-01 1.41304389e-01 5.58164597e-01 -1.28276420e+00
-3.49086702e-01 -6.19952530e-02 -8.17951262e-01 4.90872741e-01
-1.08142667e-01 3.10143251e-02 1.70633227e-01 -9.24539328e-01
-6.99763075e-02 1.06533957e+00 3.41454476e-01 1.28117651e-01
-7.06987262e-01 -1.43821025e+00 -1.86890632e-01 8.15118313e-01
-1.58764446e+00 -4.21942562e-01 1.02931106e+00 -3.79811227e-01
3.09088528e-01 3.01404804e-01 5.68226516e-01 4.31161731e-01
1.09575331e+00 3.95703077e-01 1.24002421e+00 -1.88762397e-01
-3.08604926e-01 1.07485183e-01 3.23601753e-01 3.37047487e-01
6.12532198e-01 9.96353984e-01 -1.94632411e-01 8.23851824e-02
2.68745303e-01 3.26108575e-01 -6.09607816e-01 -1.24822790e-02
-9.46408570e-01 5.93581140e-01 2.37176433e-01 5.97694218e-01
-3.94087464e-01 -3.44107896e-01 7.51872137e-02 2.94591546e-01
2.77275503e-01 -7.69722238e-02 1.76415011e-01 1.32209480e-01
-8.33509505e-01 7.82616511e-02 5.82192719e-01 7.58466005e-01
6.83547378e-01 6.42005622e-01 -1.49675058e-02 6.03551984e-01
5.10158718e-01 1.34690738e+00 1.89451739e-01 -5.08473337e-01
1.46991061e-02 1.83246017e-01 -4.23171297e-02 -1.43713796e+00
-5.38940907e-01 -9.71680164e-01 -1.08136702e+00 5.18333733e-01
1.01984836e-01 -3.08398336e-01 -7.02516258e-01 1.39547241e+00
1.22834660e-01 4.68181849e-01 4.97012585e-01 8.97536397e-01
6.04334414e-01 9.31439698e-01 -2.83500701e-01 -9.08764660e-01
1.41040838e+00 -4.00490165e-01 -1.47066939e+00 -4.05571073e-01
-1.90733671e-01 -1.15397263e+00 9.91406757e-03 4.03119892e-01
-5.08527160e-01 -8.94494891e-01 -1.53172970e+00 8.85200322e-01
2.81426474e-04 -1.80206269e-01 2.53127635e-01 8.39059412e-01
-2.53316343e-01 -1.00464299e-01 -1.96666554e-01 2.23946050e-01
-6.31489381e-02 -1.08591408e-01 -4.81421091e-02 -3.97329509e-01
-1.56452954e+00 1.25704312e+00 1.13016039e-01 5.88454962e-01
-3.54459286e-01 -7.75802612e-01 -7.96375453e-01 -2.86635578e-01
3.50524724e-01 -1.04908973e-01 1.03031957e+00 -5.99131405e-01
-1.19233263e+00 4.36775684e-02 -2.76298434e-01 -5.02541006e-01
9.79472548e-02 -1.49694264e-01 -1.49375093e+00 1.35498494e-02
1.67447209e-01 -6.50144741e-02 8.95353436e-01 -1.13972402e+00
-9.90699351e-01 -3.93420070e-01 -5.73618293e-01 2.01872978e-02
1.71528190e-01 -6.59853294e-02 5.69806658e-02 -3.58015597e-01
2.26115167e-01 -6.64235830e-01 -9.61189196e-02 -6.02158308e-01
9.16502103e-02 2.91848592e-02 1.36555171e+00 -4.38633978e-01
1.18622482e+00 -2.63538218e+00 -4.85819459e-01 3.98469657e-01
-5.09548793e-03 4.96700436e-01 -8.04303680e-03 2.74958670e-01
1.62217207e-02 -6.88774526e-01 -7.15898573e-02 6.53587401e-01
-4.11615551e-01 4.94982973e-02 -4.33613658e-01 7.33916223e-01
2.31069922e-02 4.92372513e-01 -5.55792511e-01 -2.65627742e-01
4.80573267e-01 5.95360458e-01 2.14970037e-01 3.70957889e-02
5.90196431e-01 4.85276222e-01 -4.17762995e-01 5.04619598e-01
1.30037308e+00 6.68630302e-01 -3.57169360e-01 -4.97326821e-01
-5.71510971e-01 -3.44628453e-01 -1.14761806e+00 5.71176291e-01
-2.47864157e-01 1.00680506e+00 2.92488635e-01 -8.46171856e-01
1.54350245e+00 1.50766551e-01 4.64294076e-01 -1.54068828e+00
2.71626979e-01 2.97191262e-01 6.60921633e-01 -3.79866362e-01
2.47961938e-01 -4.03818905e-01 -3.80286835e-02 -1.17573716e-01
-4.57355618e-01 -1.73206702e-01 1.04843974e-01 -1.24330059e-01
9.28941190e-01 -4.14554238e-01 3.39419276e-01 -3.42742860e-01
1.06703639e+00 -2.53173530e-01 8.96262109e-01 3.97832453e-01
-2.54730374e-01 -2.34429121e-01 -3.54042709e-01 -7.65066892e-02
-4.17565852e-01 -1.01077199e+00 -5.58329403e-01 2.86518753e-01
8.14552546e-01 2.54658729e-01 -2.58209974e-01 9.75826848e-03
2.08841190e-01 1.03940558e+00 -1.94653511e-01 -7.63625741e-01
-5.50320148e-01 -4.87324446e-01 4.51489359e-01 1.63795985e-02
8.61159682e-01 -8.24010074e-01 -6.89913332e-01 3.48223090e-01
-9.57094058e-02 -1.03628588e+00 -2.19855264e-01 -1.63913801e-01
-4.05134171e-01 -1.08357823e+00 -4.46774960e-01 -5.56367636e-01
3.89112711e-01 1.22792077e+00 4.93274540e-01 2.49557439e-02
-5.17203987e-01 3.54982793e-01 -6.33722693e-02 -1.01653636e+00
-2.06465676e-01 -8.30274642e-01 2.04517916e-01 2.61757135e-01
7.77252674e-01 -3.74808997e-01 -3.79457444e-01 6.88026011e-01
-6.41960561e-01 -3.33359689e-01 1.12374651e+00 7.27863789e-01
8.01763237e-02 7.89596915e-01 7.29602695e-01 -2.80194372e-01
6.24380291e-01 -2.13799000e-01 -1.00768113e+00 -3.70682329e-01
-5.61522663e-01 -3.00204754e-01 5.02024233e-01 -2.01576650e-01
-1.39015281e+00 -4.06324625e-01 1.69458866e-01 -3.42101276e-01
-1.27001524e-01 4.96368289e-01 -3.99998844e-01 -3.08649480e-01
3.44344318e-01 6.20755196e-01 4.76114541e-01 6.75235391e-02
-1.13290874e-03 7.64249623e-01 8.55980754e-01 3.58892024e-01
1.35721505e+00 3.02995324e-01 3.02432239e-01 -1.32059574e+00
-6.73039436e-01 -6.49201572e-01 -1.60089433e-01 -6.75299108e-01
6.35956883e-01 -9.10803556e-01 -9.11189914e-01 5.95085025e-01
-1.06464791e+00 5.64612329e-01 1.74078807e-01 1.22114134e+00
-9.49630663e-02 5.30111074e-01 -1.43416420e-01 -1.45381773e+00
-2.55790293e-01 -8.73156607e-01 3.84800255e-01 5.70260465e-01
2.57581532e-01 -4.26649272e-01 1.40928775e-02 2.36853525e-01
6.42545700e-01 -4.07823101e-02 5.25164425e-01 -2.89155811e-01
-7.79165506e-01 -5.92296481e-01 -2.56534994e-01 3.33670050e-01
1.96467429e-01 -3.65501434e-01 -5.74716628e-01 -1.24196380e-01
7.53361940e-01 5.86626112e-01 7.24013269e-01 5.40403128e-01
2.18162060e-01 -3.61945145e-02 -5.52333653e-01 3.99508953e-01
1.34207428e+00 1.08334398e+00 1.17725885e+00 3.45459640e-01
1.31285548e-01 6.19516134e-01 1.31429505e+00 2.26373121e-01
-9.74359810e-02 5.56464612e-01 4.84346002e-01 -9.34794694e-02
1.72136039e-01 3.07079881e-01 3.83651465e-01 8.28241646e-01
9.85114649e-02 7.25412816e-02 -2.69100875e-01 3.38288695e-01
-1.63019800e+00 -1.44959462e+00 -6.40630603e-01 2.33331537e+00
3.17520022e-01 4.69119310e-01 -4.18025285e-01 2.82990158e-01
7.57437646e-01 1.32879525e-01 -3.55776280e-01 -1.12202950e-01
-6.09438196e-02 -9.46542900e-03 8.37871253e-01 5.79892814e-01
-8.57734144e-01 2.27155313e-01 5.87684441e+00 9.45653081e-01
-1.00547826e+00 -1.45315886e-01 -8.62882733e-02 2.18403786e-01
-1.49771124e-02 -1.23985983e-01 -8.67873192e-01 6.40855610e-01
8.51009309e-01 -7.35684812e-01 5.05258515e-02 4.71862435e-01
3.72306079e-01 -4.84425932e-01 -2.63639361e-01 9.35607910e-01
3.29956889e-01 -6.75377727e-01 -3.93965602e-01 3.13806593e-01
1.86681494e-01 -3.58010918e-01 3.93310338e-02 4.85518366e-01
-1.95515484e-01 -8.04321408e-01 3.74488711e-01 1.06170905e+00
4.32283968e-01 -1.37318408e+00 1.22689450e+00 5.85854590e-01
-1.35339582e+00 -1.80810407e-01 -5.90936720e-01 -3.51000875e-01
3.96736979e-01 1.20970082e+00 -4.65819150e-01 1.01851225e+00
2.71604419e-01 4.22899783e-01 -1.97377682e-01 1.13400590e+00
-2.98638314e-01 4.59371388e-01 -1.70274913e-01 -1.11052684e-01
-4.68213595e-02 -6.58299446e-01 9.73887742e-01 1.09224617e+00
6.56565726e-01 5.27292132e-01 3.10460776e-01 6.43485427e-01
7.57069528e-01 -1.53254226e-01 -9.84510362e-01 2.12368608e-01
6.69228435e-01 1.45347452e+00 -1.07663520e-01 -1.15048863e-01
-5.38852215e-01 1.17782928e-01 -7.72923231e-01 3.87572378e-01
-1.08657253e+00 -9.39760089e-01 7.55549490e-01 9.66960005e-03
2.04026401e-01 -4.86085266e-01 -6.00651838e-03 -2.80937344e-01
5.82113571e-04 -6.68212891e-01 -5.93262725e-02 -6.30285740e-01
-1.08630753e+00 4.81164873e-01 -1.89421065e-02 -1.81744337e+00
-1.72491595e-01 -3.28719825e-01 -7.46559978e-01 1.06357193e+00
-1.65809727e+00 -6.79403186e-01 -3.99287879e-01 5.83179176e-01
4.25842255e-01 -5.87868810e-01 5.23082495e-01 3.79877716e-01
-3.45424354e-01 2.98548430e-01 1.34019107e-01 -1.45270720e-01
5.89352250e-01 -3.71283561e-01 -3.23601425e-01 1.06204772e+00
-4.21578586e-01 3.38916004e-01 1.08295918e+00 -8.07811797e-01
-1.43637049e+00 -9.77659822e-01 5.21485806e-01 3.42059195e-01
7.00687170e-01 -1.41205549e-01 -7.61399269e-01 3.34178098e-02
6.14233017e-01 -2.79059738e-01 5.70388436e-01 -3.09669673e-01
-7.30574429e-02 -5.26099265e-01 -9.01217043e-01 5.90663493e-01
5.31296670e-01 -4.73872349e-02 -1.01479280e+00 7.59919779e-03
4.83285308e-01 -9.93397161e-02 -4.13686067e-01 1.00760698e+00
6.40502691e-01 -9.82684851e-01 7.71468163e-01 2.62742877e-01
-2.60060340e-01 -9.78620231e-01 -6.38426319e-02 -1.37149310e+00
-8.71053755e-01 -4.18363452e-01 1.91493809e-01 1.04880250e+00
1.90615684e-01 -1.13343060e+00 9.43370834e-02 -3.40690374e-01
-2.35008046e-01 -3.31788301e-01 -1.04892814e+00 -1.16616130e+00
-5.58282614e-01 -3.63765925e-01 6.26062378e-02 4.15976584e-01
1.03467837e-01 6.99397147e-01 -6.08032942e-01 5.57578087e-01
1.04981983e+00 1.86954528e-01 9.21174467e-01 -1.54850948e+00
-1.17516415e-02 -3.10499907e-01 -7.92987049e-01 -1.03736460e+00
1.02972113e-01 -4.01622117e-01 4.10182029e-01 -1.53323627e+00
-2.06247345e-01 6.82412609e-02 -3.19135517e-01 -2.61734128e-01
-1.34868711e-01 2.67698050e-01 6.40380308e-02 1.56363800e-01
-9.68222618e-02 7.35285163e-01 1.22644949e+00 -2.96116561e-01
2.72251721e-02 4.01706100e-01 -5.45160949e-01 7.81295717e-01
6.62443638e-01 -2.32329056e-01 -3.68979216e-01 1.36154607e-01
-4.02081907e-01 1.14460215e-01 2.73977011e-01 -1.52605522e+00
5.62268913e-01 -1.49592221e-01 4.27595973e-01 -1.18150210e+00
4.36736554e-01 -1.31866479e+00 2.82600254e-01 8.19870591e-01
5.75582743e-01 -3.34898770e-01 3.97106528e-01 7.47667074e-01
-4.85620528e-01 -2.13739425e-01 9.34290886e-01 3.29631895e-01
-8.85333896e-01 -8.71492848e-02 -1.02181077e+00 -3.75800908e-01
1.37389779e+00 -3.96643400e-01 -6.06967986e-01 -4.61141199e-01
4.51497082e-03 2.97977895e-01 -2.92546719e-01 6.69546068e-01
7.96278417e-01 -1.52873278e+00 -9.59866166e-01 6.40677512e-01
3.78157683e-02 -6.00402594e-01 6.41293645e-01 1.14522731e+00
1.18053801e-01 7.01620877e-01 -2.68334627e-01 -6.89665854e-01
-1.40958643e+00 6.26348078e-01 1.53565899e-01 9.13014263e-02
-6.39769614e-01 9.16301981e-02 5.22844613e-01 1.55211866e-01
-6.71136379e-02 4.30922627e-01 -5.42187810e-01 -7.68823400e-02
1.08001578e+00 4.46875989e-01 -6.81575313e-02 -9.42577183e-01
-4.53618497e-01 8.05368721e-01 -1.19928852e-01 -5.26816063e-02
7.78468549e-01 -1.15885258e-01 -2.19833091e-01 3.62126291e-01
8.42659235e-01 5.96277952e-01 -8.68044078e-01 -2.24528313e-01
-3.17088962e-01 -7.15840399e-01 2.67416745e-01 -6.10101521e-01
-9.39615905e-01 5.60550988e-01 7.93164790e-01 4.55508590e-01
1.27786648e+00 -5.45803010e-01 7.69460618e-01 1.68175638e-01
6.91182554e-01 -9.06602800e-01 -3.76684725e-01 7.69017518e-01
7.08048582e-01 -7.02440202e-01 1.64422750e-01 -6.84139848e-01
-2.97967315e-01 1.01493073e+00 5.71881235e-01 -2.77830750e-01
7.10775971e-01 6.83806360e-01 3.47023994e-01 -4.47021099e-03
-4.78434622e-01 -4.42529291e-01 3.48391920e-01 9.21234488e-01
8.97669271e-02 -9.37095471e-03 -6.76981986e-01 5.88357985e-01
-1.54499978e-01 -3.01586419e-01 4.55726743e-01 7.27521420e-01
-1.13556492e+00 -6.80784523e-01 -1.03239894e+00 4.07286555e-01
-2.19255567e-01 1.37672648e-01 3.22367176e-02 8.32643986e-01
6.83874637e-02 1.64756906e+00 1.35418564e-01 -5.92303038e-01
9.08374846e-01 -8.15204382e-02 2.96696782e-01 -5.04468493e-02
3.54040205e-01 3.86915088e-01 -1.70172095e-01 -2.18925610e-01
-3.82058382e-01 -3.81318092e-01 -1.26357973e+00 -9.09789801e-02
-7.59506166e-01 5.29659688e-01 5.48707604e-01 1.13049889e+00
2.33751133e-01 1.05991089e+00 1.07751298e+00 -6.12401426e-01
-4.26673144e-01 -9.71841156e-01 -8.71837735e-01 -2.21284404e-01
2.46389538e-01 -1.06928301e+00 -6.99594855e-01 -4.20647711e-01] | [6.492772102355957, 1.1723220348358154] |
7545b150-a404-4082-b45a-439db1c62d02 | multi-view-learning-with-privileged-weighted | 2201.11306 | null | https://arxiv.org/abs/2201.11306v1 | https://arxiv.org/pdf/2201.11306v1.pdf | Multi-view learning with privileged weighted twin support vector machine | Weighted twin support vector machines (WLTSVM) mines as much potential similarity information in samples as possible to improve the common short-coming of non-parallel plane classifiers. Compared with twin support vector machines (TWSVM), it reduces the time complexity by deleting the superfluous constraints using the inter-class K-Nearest Neighbor (KNN). Multi-view learning (MVL) is a newly developing direction of machine learning, which focuses on learning acquiring information from the data indicated by multiple feature sets. In this paper, we propose multi-view learning with privileged weighted twin support vector machines (MPWTSVM). It not only inherits the advantages of WLTSVM but also has its characteristics. Firstly, it enhances generalization ability by mining intra-class information from the same perspective. Secondly, it reduces the redundancy constraints with the help of inter-class information, thus improving the running speed. Most importantly, it can follow both the consensus and the complementarity principle simultaneously as a multi-view classification model. The consensus principle is realized by minimizing the coupling items of the two views in the original objective function. The complementary principle is achieved by establishing privileged information paradigms and MVL. A standard quadratic programming solver is used to solve the problem. Compared with multi-view classification models such as SVM-2K, MVTSVM, MCPK, and PSVM-2V, our model has better accuracy and classification efficiency. Experimental results on 45 binary data sets prove the effectiveness of our method. | ['Huiru Wang', 'Ruxin Xu'] | 2022-01-27 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [ 7.20534995e-02 -1.14831299e-01 -7.34184802e-01 -3.13141942e-01
-3.08564544e-01 -3.39483917e-01 3.34882915e-01 -3.03183168e-01
-3.73400971e-02 6.81115687e-01 -2.73899406e-01 -1.69053540e-01
-7.15165317e-01 -8.85061443e-01 -3.24372262e-01 -1.05193162e+00
4.44721192e-01 3.14285338e-01 1.69736832e-01 -2.42698044e-01
4.42117542e-01 2.55320966e-01 -1.60287023e+00 5.54265678e-01
1.10181761e+00 1.20871878e+00 9.59918648e-02 -2.72002071e-02
-4.37557921e-02 6.67645335e-01 -5.77169508e-02 -3.81332040e-01
3.36728096e-01 -4.61717770e-02 -5.38677216e-01 2.12496519e-01
4.68465030e-01 4.68166433e-02 -6.02772534e-02 9.74804938e-01
2.34006047e-01 -1.15603998e-01 4.45647597e-01 -1.85433626e+00
-4.29221660e-01 1.27325371e-01 -1.14613342e+00 -6.82398453e-02
-5.67517243e-02 -3.07420731e-01 9.28308845e-01 -1.23283136e+00
3.69001627e-01 8.91901135e-01 4.34482038e-01 3.52262855e-01
-1.10168076e+00 -6.23405635e-01 1.96089774e-01 7.67071307e-01
-1.22486484e+00 8.88561457e-02 9.62898731e-01 -5.26244342e-01
7.31928527e-01 7.74778903e-01 6.97616518e-01 8.09940338e-01
3.98655981e-01 9.02657688e-01 1.33203685e+00 -3.63518715e-01
9.50248167e-02 6.27163827e-01 4.67894703e-01 6.94538713e-01
5.01214743e-01 -4.61092442e-02 -7.42171645e-01 -4.29800689e-01
3.74308944e-01 4.00474936e-01 -4.91418868e-01 -1.20140302e+00
-1.22373390e+00 9.99557078e-01 -2.65276637e-02 2.43249953e-01
6.99303448e-02 -7.64425337e-01 6.12478316e-01 2.59064734e-01
5.33488154e-01 1.80003762e-01 -6.61334693e-01 3.03217024e-01
-7.43853629e-01 -1.95520729e-01 5.92683196e-01 9.23801780e-01
5.77152014e-01 -6.67961733e-03 3.60007226e-01 7.80424833e-01
2.14570537e-01 5.68907619e-01 6.08547807e-01 -5.66798210e-01
8.01244080e-01 1.09858680e+00 -2.66861260e-01 -1.14827657e+00
-3.67217392e-01 -3.88048470e-01 -1.06410480e+00 3.63818884e-01
2.67335236e-01 1.78986400e-01 -4.11072731e-01 1.21430385e+00
4.22663122e-01 -2.08226629e-02 3.29167694e-01 8.36588740e-01
7.42813051e-01 4.78803933e-01 -5.55716932e-01 -6.30442381e-01
1.28627419e+00 -1.00982654e+00 -5.97531796e-01 -8.65041539e-02
6.06455266e-01 -6.15882397e-01 5.29186249e-01 8.00698459e-01
-5.87386847e-01 -4.91799504e-01 -1.26574159e+00 4.49170470e-01
-3.87859046e-01 1.87932476e-01 8.17509949e-01 6.93196535e-01
-3.22344959e-01 4.61254537e-01 -6.27862632e-01 1.20834798e-01
4.42982793e-01 5.26071131e-01 -8.95951331e-01 -7.89972618e-02
-1.04053354e+00 9.05955374e-01 5.64278781e-01 2.50888895e-02
-6.94227889e-02 -8.91359806e-01 -9.04716253e-01 -9.85797346e-02
6.68887079e-01 -3.41612369e-01 5.30306578e-01 -1.02107167e+00
-1.36625576e+00 6.92385614e-01 -2.19467163e-01 1.85132071e-01
4.99578297e-01 6.95538595e-02 -6.92452431e-01 9.14385319e-02
1.95146516e-01 6.61198748e-03 8.99962604e-01 -1.18309200e+00
-7.50390649e-01 -8.51621568e-01 -3.53018641e-01 2.40270495e-01
-6.70315146e-01 -3.08022678e-01 -1.37438342e-01 -3.05957556e-01
5.82533896e-01 -7.91037858e-01 -7.19726607e-02 -1.81679145e-01
-3.87664407e-01 -2.92972893e-01 1.57641685e+00 -3.80342782e-01
1.14444709e+00 -2.09602404e+00 6.92180276e-01 5.00749648e-01
3.93402189e-01 3.47719729e-01 2.13272169e-01 4.36585605e-01
-4.01445061e-01 -2.68037975e-01 -2.71396756e-01 2.87229091e-01
-6.01937771e-01 3.37574750e-01 -2.49461055e-01 7.37328112e-01
-6.00379556e-02 6.28769875e-01 -7.00995624e-01 -6.53612852e-01
2.56955504e-01 5.20634018e-02 -3.92982274e-01 -1.89198613e-01
3.51300687e-01 3.93608026e-02 -4.63947862e-01 5.18762946e-01
9.69653368e-01 -4.69361037e-01 4.89426851e-01 -3.46519381e-01
-1.86321989e-01 -4.80041742e-01 -1.57501042e+00 1.19475198e+00
-3.72189283e-01 2.08983779e-01 -9.95091200e-02 -1.58426440e+00
9.65327740e-01 3.20758790e-01 7.78768480e-01 -4.31142420e-01
-1.58493996e-01 3.04127753e-01 -1.20478638e-01 -6.90415144e-01
-7.38241896e-02 -2.98618823e-01 8.44013616e-02 2.29685113e-01
2.32147872e-01 2.12591782e-01 -1.56070828e-01 1.04522750e-01
1.54317766e-01 4.68950383e-02 7.32353806e-01 -4.14007634e-01
9.52855051e-01 6.01025820e-02 7.98026145e-01 2.18033660e-02
7.08310157e-02 2.32504204e-01 6.32177413e-01 -6.40462935e-01
-7.36008584e-01 -9.26632524e-01 -5.61521769e-01 5.71324587e-01
5.00293314e-01 -4.02282029e-01 -3.54861081e-01 -1.13211858e+00
2.88414270e-01 5.56059837e-01 -5.99335194e-01 -1.89742044e-01
-5.17661095e-01 -8.03337932e-01 -5.53107113e-02 7.05597758e-01
4.65336829e-01 -3.22859883e-01 -5.38880289e-01 -2.52582461e-01
-3.10496300e-01 -9.86346781e-01 -2.92328209e-01 2.37406209e-01
-9.73600507e-01 -1.41375637e+00 -4.82079357e-01 -7.03556240e-01
7.20226228e-01 8.18115830e-01 5.58264792e-01 -1.79960564e-01
-3.38754743e-01 1.78359538e-01 -2.35701114e-01 -4.14836258e-01
1.32750407e-01 -1.05495602e-01 5.15833020e-01 4.28687006e-01
6.09937489e-01 -4.71131682e-01 -1.22710146e-01 8.62188816e-01
-6.29428029e-01 3.49895090e-01 4.19694543e-01 1.39322627e+00
7.33562648e-01 1.66391939e-01 5.51652014e-01 -8.09179664e-01
9.03801993e-03 -3.45084250e-01 -7.30932176e-01 5.56057870e-01
-1.04694152e+00 -2.52780139e-01 7.74729848e-01 -4.32906955e-01
-9.70034122e-01 -1.07700646e-01 6.32440865e-01 -7.35283673e-01
1.42262161e-01 3.72515947e-01 -4.68683183e-01 -3.22637618e-01
3.41288567e-01 5.82720101e-01 4.86583799e-01 -2.88000554e-01
9.03589576e-02 8.59562278e-01 -6.67382777e-02 -2.09782526e-01
7.78108537e-01 7.26396561e-01 4.20045495e-01 -8.40373993e-01
-8.63015771e-01 -7.05296159e-01 -9.62267280e-01 -2.27937967e-01
7.11821139e-01 -6.63271248e-01 -1.03607571e+00 4.62696344e-01
-7.22058892e-01 7.06490040e-01 -7.08830804e-02 7.26690054e-01
-4.98115689e-01 7.45921135e-01 -3.30451503e-02 -6.15296602e-01
-2.49792740e-01 -1.08478069e+00 7.16189742e-01 7.60645717e-02
9.73005071e-02 -1.00105977e+00 -1.99687734e-01 7.88434386e-01
3.84242423e-02 1.87692076e-01 9.59357679e-01 -8.54964733e-01
-2.54859865e-01 -1.81070447e-01 -5.53201549e-02 4.93324786e-01
2.74491072e-01 -1.52300566e-01 -8.55122566e-01 -5.06262183e-01
4.97261941e-01 -2.20636413e-01 6.73171103e-01 2.10004792e-01
1.28358769e+00 -2.40134999e-01 -5.55895865e-01 7.57497430e-01
1.57791793e+00 3.29656005e-01 2.76399016e-01 5.39639711e-01
7.92784691e-01 8.87440979e-01 7.90998697e-01 3.89453351e-01
2.36646891e-01 8.34214330e-01 5.66716373e-01 -3.23700421e-02
7.17646182e-01 -4.32536155e-02 1.52424008e-01 1.01647711e+00
-2.73637652e-01 2.12707326e-01 -7.75359035e-01 2.80843787e-02
-2.08947635e+00 -1.41079533e+00 -4.98349249e-01 2.18823171e+00
2.30520710e-01 -8.00954998e-02 1.92125179e-02 5.14001071e-01
6.71306968e-01 1.47508353e-01 -5.36597371e-01 -2.63347864e-01
-1.98518634e-01 -2.07751706e-01 3.79541934e-01 2.69953549e-01
-1.25237417e+00 3.14650029e-01 4.98140526e+00 9.58583355e-01
-1.30632901e+00 2.17264127e-02 4.79512334e-01 -2.95248747e-01
-2.32105777e-01 -1.81171402e-01 -9.31173682e-01 4.10930574e-01
2.55067408e-01 -1.92449108e-01 1.96197197e-01 1.22197115e+00
-1.77616045e-01 -2.73056999e-02 -1.11879969e+00 1.20790720e+00
5.33372760e-01 -1.25906765e+00 1.21007361e-01 2.97886908e-01
5.12248337e-01 -2.84654409e-01 1.19984649e-01 5.31534478e-02
-4.99395430e-01 -7.44472921e-01 4.59237456e-01 2.95345545e-01
6.96310222e-01 -9.85863984e-01 8.08866322e-01 7.74290979e-01
-1.13767171e+00 -3.33633393e-01 -3.80086750e-01 6.73604682e-02
-1.10655792e-01 6.09617174e-01 -2.77161628e-01 1.33690822e+00
5.29823065e-01 9.51828182e-01 -3.41622740e-01 6.01380169e-01
1.32536069e-01 5.04198037e-02 -9.95012075e-02 1.23870373e-01
5.04380688e-02 -7.51083553e-01 5.92636406e-01 5.41203916e-01
8.36781040e-02 1.64363921e-01 4.05787945e-01 4.34922367e-01
7.56287217e-01 3.30507308e-01 -8.45640898e-01 4.82799351e-01
1.20649822e-01 1.36859286e+00 -3.76394123e-01 -2.39197776e-01
-8.09356213e-01 8.35284650e-01 3.45177323e-01 2.67531946e-02
-6.63145006e-01 -3.61203611e-01 4.52062815e-01 -8.01225454e-02
3.97266954e-01 6.16336353e-02 -5.07925868e-01 -1.59157252e+00
4.37660366e-01 -1.05073309e+00 5.50982296e-01 -2.86899328e-01
-1.36089194e+00 3.61012667e-01 2.18383893e-01 -1.96433115e+00
7.14264140e-02 -1.06623530e+00 -4.31686640e-01 8.33950460e-01
-1.42512584e+00 -1.33307528e+00 -1.51527807e-01 7.54313350e-01
4.13284868e-01 -5.27751148e-01 9.07613635e-01 1.13198422e-01
-8.45115483e-01 5.73394597e-01 2.83360422e-01 -1.12608178e-02
6.38968945e-01 -1.01140916e+00 -4.18711036e-01 4.48836178e-01
-1.52543902e-01 4.78617698e-01 1.71676993e-01 -4.43666905e-01
-1.58620036e+00 -8.55480492e-01 7.55563676e-01 -4.47689891e-01
5.53687572e-01 -1.73563644e-01 -8.44018281e-01 4.65367585e-01
-2.03091085e-01 3.11181575e-01 1.26528811e+00 1.75436825e-01
-6.01782203e-01 -4.04722780e-01 -1.10947728e+00 1.29719481e-01
6.35935664e-01 -2.97685921e-01 -5.92207551e-01 3.77785861e-01
3.63019168e-01 -3.12542588e-01 -1.03709710e+00 6.05470121e-01
7.51492500e-01 -1.18857825e+00 1.08022094e+00 -9.51299131e-01
5.65590739e-01 -2.68192053e-01 -4.44825768e-01 -1.37551475e+00
-4.67472315e-01 5.91083616e-02 -2.18165949e-01 9.73363757e-01
3.11367363e-01 -1.08539581e+00 8.84333670e-01 3.73259187e-01
9.45798680e-02 -1.34776235e+00 -1.35599339e+00 -1.10290802e+00
-2.99012456e-02 -6.36691973e-02 4.32713836e-01 1.45359409e+00
5.30571699e-01 2.81899869e-01 -6.41412020e-01 2.50266254e-01
1.07377994e+00 7.34200895e-01 5.05922198e-01 -1.50386631e+00
-5.03194451e-01 -2.15857506e-01 -6.29202962e-01 -2.51474679e-01
1.78206295e-01 -1.13057792e+00 -7.08986461e-01 -1.13853621e+00
5.19954979e-01 -3.19551796e-01 -4.26039159e-01 4.58341122e-01
-1.36281580e-01 -1.70093030e-01 1.33916020e-01 2.22092196e-01
-3.45778763e-01 4.90931332e-01 1.44744897e+00 -8.30827206e-02
-1.26344025e-01 3.15732718e-01 -6.25229299e-01 9.99992192e-01
7.05874205e-01 -2.45910645e-01 -4.81131524e-01 -3.59684825e-02
-4.68749590e-02 3.51158351e-01 2.91209996e-01 -6.41204238e-01
1.90432325e-01 -5.98339379e-01 6.71551943e-01 -6.52355492e-01
4.85558808e-01 -1.22109246e+00 3.00849397e-02 7.51880944e-01
3.04622650e-01 -1.05687790e-02 -4.21340652e-02 6.57891512e-01
-3.97894681e-01 -1.19539946e-01 9.18379545e-01 -2.50938125e-02
-7.76565850e-01 1.58374265e-01 -1.54377148e-02 -2.33044699e-01
1.71912098e+00 -6.20468616e-01 -4.04729307e-01 1.99526519e-01
-4.06401783e-01 5.33184111e-01 1.94248691e-01 5.57652116e-01
8.45785618e-01 -1.44598818e+00 -4.48788971e-01 6.76846743e-01
5.14811695e-01 -2.47949287e-01 3.88000041e-01 1.15071487e+00
5.73414713e-02 5.76199889e-01 -4.26742941e-01 -8.52407992e-01
-1.74000943e+00 8.51485312e-01 3.14718157e-01 -2.75084972e-01
-5.98697007e-01 6.86657846e-01 2.08274990e-01 -5.37652016e-01
-8.85828212e-02 4.14676845e-01 -6.17168665e-01 4.08306539e-01
5.84421813e-01 8.17243695e-01 1.58004209e-01 -6.92759693e-01
-5.79309821e-01 8.85684073e-01 -3.33903491e-01 3.11023325e-01
1.45310807e+00 3.12780380e-01 -4.85500962e-01 4.29840893e-01
1.33172750e+00 -1.25671580e-01 -6.76655114e-01 -2.02200666e-01
-1.12273172e-01 -8.22358549e-01 5.96368872e-02 -3.75637442e-01
-1.10206509e+00 7.85724878e-01 4.99600768e-01 6.70707971e-02
9.88945544e-01 -2.68023819e-01 4.82670069e-01 2.77328581e-01
7.27389574e-01 -1.06747711e+00 -1.38613414e-02 2.33523205e-01
7.99266756e-01 -1.42304254e+00 2.84004152e-01 -9.25255895e-01
-8.88372779e-01 1.46838903e+00 1.13708675e+00 1.98548928e-01
8.68169844e-01 1.13284476e-01 -6.05736524e-02 -2.75467515e-01
-7.60606945e-01 4.52810049e-01 5.25417149e-01 4.29234028e-01
3.41176125e-03 2.13913366e-01 -3.97455752e-01 6.28522992e-01
9.94075760e-02 -1.92825407e-01 2.93884248e-01 8.62327337e-01
-3.82962644e-01 -1.21378481e+00 -4.43028182e-01 7.16434658e-01
-1.86579913e-01 2.38237098e-01 -2.11085156e-01 7.84167707e-01
2.15683430e-01 9.05348837e-01 -2.26588324e-01 -7.31895387e-01
5.08932710e-01 1.02923378e-01 3.86022925e-01 -3.97729367e-01
-2.35695928e-01 -1.81749135e-01 -1.57714754e-01 -5.36923349e-01
-3.67552370e-01 -7.10379004e-01 -7.65271306e-01 -3.13616872e-01
-8.44468951e-01 2.53219038e-01 4.81629461e-01 1.01618671e+00
2.67646104e-01 2.85445035e-01 1.11555886e+00 -6.30880356e-01
-1.05452693e+00 -4.45243478e-01 -7.80137360e-01 1.27258345e-01
1.43223152e-01 -8.08980405e-01 -4.14463222e-01 -4.02487099e-01] | [8.368111610412598, 4.5124077796936035] |
776057f0-735e-4a73-a400-963b15ed3cd8 | shallow-pooling-for-sparse-labels | 2109.00062 | null | https://arxiv.org/abs/2109.00062v2 | https://arxiv.org/pdf/2109.00062v2.pdf | Shallow pooling for sparse labels | Recent years have seen enormous gains in core IR tasks, including document and passage ranking. Datasets and leaderboards, and in particular the MS MARCO datasets, illustrate the dramatic improvements achieved by modern neural rankers. When compared with traditional test collections, the MS MARCO datasets employ substantially more queries with substantially fewer known relevant items per query. Given the sparsity of these relevance labels, the MS MARCO leaderboards track improvements with mean reciprocal rank (MRR). In essence, a relevant item is treated as the "right answer", with rankers scored on their ability to place this item high in the ranking. In working with these sparse labels, we have observed that the top items returned by a ranker often appear superior to judged relevant items. To test this observation, we employed crowdsourced workers to make preference judgments between the top item returned by a modern neural ranking stack and a judged relevant item. The results support our observation. If we imagine a perfect ranker under MRR, with a score of 1 on all queries, our preference judgments indicate that a searcher would prefer the top result from a modern neural ranking stack more frequently than the top result from the imaginary perfect ranker, making our neural ranker "better than perfect". To understand the implications for the leaderboard, we pooled the top document from available runs near the top of the passage ranking leaderboard for over 500 queries. We employed crowdsourced workers to make preference judgments over these pools and re-evaluated the runs. Our results support our concerns that current MS MARCO datasets may no longer be able to recognize genuine improvements in rankers. In future, if rankers are measured against a single "right answer", this answer should be the best answer or most preferred answer, and maintained with ongoing judgments. | ['Charles L. A. Clarke', 'Xinyi Yan', 'Alexandra Vtyurina', 'Negar Arabzadeh'] | 2021-08-31 | null | null | null | null | ['passage-ranking'] | ['natural-language-processing'] | [ 2.02263948e-02 -1.86671332e-01 -4.32328105e-01 -4.82841939e-01
-1.51705217e+00 -1.05612588e+00 6.20278656e-01 4.24361438e-01
-7.98998296e-01 6.73631608e-01 8.69620025e-01 -3.22999388e-01
-6.36148036e-01 -6.54656410e-01 -8.48348260e-01 -2.60647535e-01
-1.23073600e-01 8.77588272e-01 2.06373170e-01 -6.00903749e-01
6.62140250e-01 4.15166579e-02 -1.59071708e+00 7.88975656e-01
7.09278226e-01 8.18922937e-01 8.16375315e-02 7.93826938e-01
5.45997620e-02 7.92816162e-01 -1.03534746e+00 -5.46812177e-01
5.02866685e-01 1.81847177e-02 -1.07196724e+00 -8.16480041e-01
1.05616486e+00 -4.85292166e-01 -4.80625868e-01 8.69795680e-01
6.58016086e-01 5.66355824e-01 5.48350513e-01 -8.73459518e-01
-1.07319820e+00 1.04741287e+00 -2.52124608e-01 4.14688349e-01
6.96236551e-01 -2.82904245e-02 1.84695160e+00 -1.04460156e+00
8.24406028e-01 1.09154940e+00 5.22464812e-01 3.84257883e-01
-1.07197356e+00 -5.37463844e-01 7.52932578e-02 -8.94542709e-02
-1.20929945e+00 -4.95216340e-01 2.93267667e-02 -2.80590355e-01
1.06573009e+00 5.08207321e-01 2.18789756e-01 7.66330421e-01
3.27547453e-02 7.76659846e-01 8.29251409e-01 -3.06281865e-01
1.40694842e-01 -4.17756289e-02 5.00896513e-01 3.87652725e-01
2.84217089e-01 2.56884098e-01 -9.25444901e-01 -6.04147255e-01
1.70018420e-01 2.18088683e-02 -2.21812531e-01 2.05714151e-01
-1.25783634e+00 5.95370471e-01 8.12392175e-01 2.15122581e-01
-2.69593477e-01 2.90722638e-01 1.48427024e-01 6.43907011e-01
3.86843264e-01 1.23829925e+00 -5.94799697e-01 -2.43529513e-01
-8.83710444e-01 6.24057651e-01 8.50855291e-01 7.84361899e-01
7.77498364e-01 -6.36476517e-01 -7.61667669e-01 1.18332493e+00
2.77433284e-02 4.56865489e-01 7.90278673e-01 -1.46365654e+00
7.27426410e-01 6.79297328e-01 5.67655325e-01 -7.75597215e-01
-3.46262932e-01 -6.97241127e-01 -2.09772661e-01 8.07372481e-02
5.63049555e-01 1.55809829e-02 -7.23100364e-01 1.78756392e+00
-3.09916556e-01 -5.32405257e-01 1.64039493e-01 1.10846043e+00
5.91453910e-01 5.54768145e-01 1.37474379e-02 1.79408446e-01
1.30159783e+00 -9.23212409e-01 -1.49033725e-01 -2.67031729e-01
9.86194968e-01 -9.65817153e-01 1.65476000e+00 3.92624348e-01
-1.04062390e+00 -5.30530393e-01 -1.05917120e+00 -2.65858740e-01
-2.87658572e-01 -8.78400356e-03 6.22314930e-01 3.88245732e-02
-1.37391150e+00 8.80184472e-01 -1.79324597e-01 -3.23041111e-01
-7.29960427e-02 4.38291192e-01 -3.21902692e-01 -1.63165480e-01
-1.47638440e+00 8.49472046e-01 -2.62724818e-03 -4.52595763e-02
-9.52960730e-01 -5.27798593e-01 -2.74418205e-01 2.76463985e-01
3.33550185e-01 -6.04925573e-01 1.76721323e+00 -6.41193449e-01
-8.98844779e-01 8.48126650e-01 -3.30394715e-01 -3.53910059e-01
3.66061032e-01 -5.70143342e-01 -3.97437871e-01 -1.22067213e-01
4.90627408e-01 6.05587244e-01 1.80824623e-01 -1.02069473e+00
-1.03590214e+00 -3.65274101e-01 5.88192165e-01 6.27208114e-01
-2.28168711e-01 4.87031043e-02 -3.07270467e-01 -4.29246575e-01
2.39149317e-01 -9.66490090e-01 7.61148408e-02 -5.17694354e-01
-2.11409390e-01 -7.39007890e-01 -1.58433408e-01 -1.44393519e-01
1.30462313e+00 -2.02361894e+00 -4.16738242e-01 2.83719420e-01
3.19826096e-01 -8.23214184e-03 -6.31134808e-01 4.82612997e-01
1.45887747e-01 4.04698879e-01 2.91016757e-01 -1.08382300e-01
2.87161946e-01 -2.05133736e-01 -6.16133273e-01 1.13132276e-01
-2.53927767e-01 9.68491495e-01 -1.17869616e+00 -2.44066179e-01
-5.45958877e-01 -1.38538450e-01 -5.80382824e-01 5.82276955e-02
-2.98470497e-01 -2.15161964e-01 -4.36764777e-01 6.01225436e-01
9.27964747e-02 -5.65611959e-01 -1.73760355e-01 2.51320228e-02
2.88238225e-04 7.89615571e-01 -9.50553834e-01 1.42231202e+00
-2.81264037e-01 6.23671174e-01 -2.37901837e-01 -3.10655832e-01
9.27247286e-01 1.94419697e-01 1.41349688e-01 -1.16108990e+00
-4.54544812e-01 6.10651612e-01 5.79596199e-02 -2.45342374e-01
1.40171802e+00 1.76831633e-01 -3.82993706e-02 9.52708840e-01
-3.11463356e-01 -1.80372372e-02 4.31903005e-01 6.97392881e-01
1.48445785e+00 -1.37753472e-01 -2.01944917e-01 -2.35501304e-01
1.22514203e-01 4.27684903e-01 2.97100067e-01 1.54704535e+00
-1.66361272e-01 7.89876103e-01 5.08210123e-01 -5.09155333e-01
-9.40930367e-01 -1.12225187e+00 1.57929063e-02 2.19557285e+00
6.08603396e-02 -3.24607372e-01 -3.15585971e-01 -6.87143743e-01
1.76696360e-01 4.04291511e-01 -5.65106094e-01 -1.25993431e-01
-6.00350320e-01 -4.49065149e-01 8.29165101e-01 5.59481263e-01
3.13966200e-02 -1.17042828e+00 -1.09151863e-01 1.40744492e-01
-3.26862156e-01 -3.42019588e-01 -9.04186726e-01 4.49851900e-01
-7.56309748e-01 -9.99468207e-01 -1.02383518e+00 -7.39217043e-01
5.41624725e-01 5.65704525e-01 1.51577485e+00 4.15853679e-01
3.49520177e-01 2.78250307e-01 -4.17799324e-01 -7.26573020e-02
-1.91547647e-01 4.71463561e-01 2.60686338e-01 -4.94605839e-01
5.33082426e-01 -1.51222140e-01 -8.36055875e-01 5.56354582e-01
-8.04877400e-01 -5.90683758e-01 6.75216913e-01 8.09708714e-01
3.63610059e-01 -4.38858777e-01 9.41239417e-01 -9.85170960e-01
1.29687476e+00 -5.45495570e-01 -5.20571351e-01 5.15600502e-01
-9.62852657e-01 5.11527479e-01 5.26475906e-01 -4.25676107e-01
-7.12741196e-01 -2.63393074e-01 1.23676129e-01 1.35488555e-01
3.46711129e-01 6.82422101e-01 4.74576622e-01 4.34623480e-01
1.54590082e+00 -1.46039873e-01 -3.88862818e-01 -6.01076424e-01
3.49767715e-01 8.63045156e-01 6.93655133e-01 -1.07743561e+00
5.52382231e-01 1.83279917e-01 -5.83895147e-01 2.09429557e-03
-1.21642435e+00 -7.61473417e-01 7.66245425e-02 5.08306213e-02
4.30839449e-01 -9.38422740e-01 -9.56276536e-01 -1.56474318e-02
-1.08103490e+00 -3.82905841e-01 -4.16614115e-01 6.07427537e-01
-2.17170969e-01 1.43121823e-03 -8.43816459e-01 -4.42082793e-01
-5.51793933e-01 -1.11613142e+00 9.21707571e-01 4.09655422e-01
-6.43610775e-01 -6.80282831e-01 4.28016454e-01 5.95963299e-01
7.32113838e-01 -6.05348110e-01 1.12212968e+00 -1.17020273e+00
-4.68373805e-01 -5.34421146e-01 -2.14882150e-01 1.48534611e-01
-1.53737441e-01 -1.82956219e-01 -1.02956319e+00 -3.30572069e-01
-4.85300839e-01 -5.88643491e-01 1.01669979e+00 1.88201383e-01
7.08588898e-01 -1.41511559e-01 -2.26105615e-01 1.41298701e-03
1.08605170e+00 3.47861238e-02 1.41928732e-01 6.24510348e-01
2.61008501e-01 5.81717610e-01 5.72481275e-01 1.12982072e-01
2.84348130e-01 6.25669181e-01 7.50244036e-02 2.40346357e-01
2.30085086e-02 -6.30313993e-01 5.36406696e-01 6.78228140e-01
-1.68807674e-02 -5.68030834e-01 -9.18069780e-01 5.44468462e-01
-1.70918930e+00 -8.85178208e-01 4.12461534e-02 2.58107376e+00
1.01227641e+00 3.90525728e-01 -9.92694050e-02 -3.17034453e-01
7.75439501e-01 1.77295998e-01 -6.49543881e-01 -2.98375517e-01
-3.06305736e-01 7.23765939e-02 5.71860373e-01 7.79818714e-01
-7.05261052e-01 6.98456168e-01 7.39427853e+00 7.12515354e-01
-7.94911623e-01 -1.25722572e-01 5.45634747e-01 -5.24565339e-01
-6.13146782e-01 7.48299882e-02 -1.00606441e+00 3.65611225e-01
9.60618734e-01 -4.70258445e-01 8.31768095e-01 8.48993421e-01
-3.87883671e-02 -1.44609958e-01 -1.48894644e+00 6.29715145e-01
-8.27743635e-02 -1.28207839e+00 1.08540796e-01 1.45361751e-01
9.72222805e-01 5.97600639e-01 2.63216972e-01 6.63567603e-01
1.08750534e+00 -1.22928667e+00 7.24579215e-01 5.71327806e-01
5.38416266e-01 -4.14529413e-01 9.55807686e-01 3.41058344e-01
-7.42925644e-01 -2.02176109e-01 -7.42699862e-01 -3.01716059e-01
-1.19312264e-01 5.71217954e-01 -9.88268256e-01 1.51378512e-01
6.49443209e-01 2.35306039e-01 -7.97267675e-01 1.06792295e+00
-2.06274912e-01 4.42632794e-01 -3.77473950e-01 -2.59936631e-01
2.91496992e-01 3.49749476e-01 4.21850324e-01 9.53757465e-01
2.78869599e-01 -1.38284728e-01 -1.39529761e-02 4.50697601e-01
-7.56089926e-01 1.18091449e-01 -4.99894649e-01 -5.14784753e-02
8.71573508e-01 1.03459215e+00 -2.42812559e-01 -7.20821500e-01
-1.48677409e-01 5.55053651e-01 6.41488612e-01 5.47417760e-01
-2.61481315e-01 -4.52015370e-01 3.87169123e-01 7.28442892e-02
-4.21577871e-01 1.66790560e-01 -2.05454141e-01 -9.00176823e-01
3.33236367e-01 -1.04374731e+00 8.18616986e-01 -1.03657615e+00
-1.51029098e+00 7.73692548e-01 -1.79521248e-01 -1.06371737e+00
-3.94352376e-01 -5.25142729e-01 -3.97036880e-01 1.04593444e+00
-1.13547313e+00 -3.39128733e-01 -1.71900943e-01 3.27668160e-01
4.39817995e-01 -3.51091832e-01 6.33283913e-01 3.72943342e-01
6.79389015e-02 8.05304885e-01 3.25696796e-01 1.56735063e-01
1.32928145e+00 -1.37693536e+00 3.70338202e-01 3.63564283e-01
5.51620483e-01 1.40863740e+00 5.98644018e-01 -6.94406688e-01
-1.05021715e+00 -5.32668889e-01 1.25764632e+00 -9.86580372e-01
5.08029461e-01 1.08540967e-01 -9.17890787e-01 7.19842434e-01
2.97508389e-02 -1.74292549e-01 6.19781852e-01 8.37424457e-01
-6.45992339e-01 -2.35364363e-01 -7.61360586e-01 6.53897107e-01
1.14546382e+00 -9.24452543e-01 -1.01611352e+00 5.71489275e-01
8.81860852e-01 -3.16466838e-01 -6.70032978e-01 2.95070201e-01
8.92996311e-01 -6.51887536e-01 8.40638578e-01 -8.39672387e-01
5.64379156e-01 -5.11523426e-01 -4.67847675e-01 -1.35774243e+00
-4.45849836e-01 -3.79761666e-01 6.29548728e-02 9.84386384e-01
9.98460233e-01 -4.80948478e-01 9.14919853e-01 1.01708591e+00
-2.04678550e-01 -6.74447119e-01 -7.96392500e-01 -6.96072638e-01
2.88718015e-01 -2.94350058e-01 7.50868380e-01 7.43729234e-01
5.38521223e-02 5.72458386e-01 -1.84061915e-01 -1.34831388e-02
1.31734505e-01 -6.14689663e-02 6.24583364e-01 -1.32287431e+00
-4.97698635e-01 -6.26045823e-01 2.03887224e-01 -1.41729772e+00
-1.65703490e-01 -1.23281765e+00 3.92459661e-01 -1.63013470e+00
4.46873456e-01 -7.99312353e-01 -7.98641682e-01 5.55899799e-01
-4.25411016e-01 4.12838429e-01 1.18656754e-01 8.06724131e-01
-1.17130125e+00 -2.89326925e-02 1.00339389e+00 -2.98858255e-01
-3.16316217e-01 -8.57461691e-02 -1.26164436e+00 4.69365716e-01
4.48068142e-01 -7.46168554e-01 -3.04259270e-01 -9.39344406e-01
1.13074148e+00 3.59989740e-02 7.93587118e-02 -8.64839137e-01
7.67530560e-01 4.04542051e-02 3.52420717e-01 -4.11709011e-01
1.41598895e-01 -2.90301740e-01 -1.65964350e-01 3.04054737e-01
-1.24616504e+00 5.18433869e-01 -1.99365675e-01 3.30716193e-01
-1.63535088e-01 -4.89620715e-01 3.01789105e-01 -2.03839362e-01
-6.39000416e-01 -9.32033062e-02 -1.29126519e-01 3.91730428e-01
1.47664696e-01 2.77172495e-02 -1.09779310e+00 -5.37571788e-01
-5.96104562e-01 4.14314359e-01 4.83392864e-01 5.40705085e-01
4.15118456e-01 -1.27361035e+00 -9.75211024e-01 -2.50072688e-01
4.78760362e-01 -1.21844731e-01 -6.26226664e-02 3.47946584e-01
-2.42183313e-01 6.43146813e-01 3.22876453e-01 -2.39462256e-01
-7.98167288e-01 2.19244018e-01 3.49065959e-01 -5.34359753e-01
-2.15563029e-01 9.89749253e-01 1.64093748e-01 -7.03071952e-01
4.02895570e-01 -1.57676294e-01 -1.89057857e-01 1.22643732e-01
7.80931056e-01 5.10920942e-01 2.47781172e-01 -1.84549898e-01
-2.14736924e-01 3.67419243e-01 -4.92528379e-01 -5.47263563e-01
1.03178024e+00 1.14387594e-01 -4.15290259e-02 3.25611889e-01
1.20714808e+00 4.81733054e-01 -8.90039802e-01 -4.60525542e-01
3.16497833e-01 -4.26944405e-01 -3.14966351e-01 -1.37897360e+00
-5.22573590e-01 4.26719427e-01 4.07863647e-01 2.51749933e-01
5.75091660e-01 1.82095021e-01 5.34720778e-01 1.19817817e+00
5.10963500e-01 -1.25996888e+00 1.86042786e-01 7.16736019e-01
9.72551465e-01 -1.28021181e+00 2.62788497e-03 5.30220091e-01
-6.66785717e-01 7.07895279e-01 6.56569302e-01 -9.09426361e-02
-4.34746780e-02 -3.14090908e-01 3.80838662e-01 -3.52556169e-01
-1.12734950e+00 6.15600497e-02 5.32480597e-01 1.15335025e-01
8.62309277e-01 1.65959433e-01 -3.68273377e-01 5.01431525e-01
-5.46157181e-01 -4.86242631e-03 3.88958186e-01 6.72832370e-01
-8.30612540e-01 -9.64157760e-01 -3.81849617e-01 8.74326944e-01
-6.62255585e-01 -4.42569524e-01 -4.47050661e-01 2.63545841e-01
-3.09072644e-01 1.22662961e+00 2.51504360e-03 -6.05856061e-01
5.62752783e-01 1.04155548e-01 -6.19424731e-02 -7.83815801e-01
-8.79038393e-01 -4.45084125e-01 1.72310427e-01 -6.15790248e-01
1.12718478e-01 -1.95046097e-01 -1.24232078e+00 -5.08219838e-01
-2.00445905e-01 6.95440710e-01 4.16494906e-01 8.26810062e-01
4.52862680e-01 -2.52834725e-04 4.89098340e-01 -3.67948711e-01
-1.07487404e+00 -1.14379668e+00 -3.67225438e-01 6.10246360e-01
4.39344972e-01 -2.94044524e-01 -5.09723604e-01 -4.05350894e-01] | [11.48799991607666, 7.579523086547852] |
4f3fa442-1712-4363-ab6c-66b58fdde7dc | cross-clinic-de-identification-of-swedish | null | null | https://aclanthology.org/2022.legal-1.10 | https://aclanthology.org/2022.legal-1.10.pdf | Cross-Clinic De-Identification of Swedish Electronic Health Records: Nuances and Caveats | Privacy preservation of sensitive information is one of the main concerns in clinical text mining. Due to the inherent privacy risks of handling clinical data, the clinical corpora used to create the clinical Named Entity Recognition (NER) models underlying clinical de-identification systems cannot be shared. This situation implies that clinical NER models are trained and tested on data originating from the same institution since it is rarely possible to evaluate them on data belonging to a different organization. These restrictions on sharing make it very difficult to assess whether a clinical NER model has overfitted the data or if it has learned any undetected biases. This paper presents the results of the first-ever cross-institution evaluation of a Swedish de-identification system on Swedish clinical data. Alongside the encouraging results, we discuss differences and similarities across EHR naming conventions and NER tagsets. | ['Marina Santini', 'Thomas Vakili', 'Olle Bridal'] | null | null | null | null | legal-lrec-2022-6 | ['de-identification'] | ['natural-language-processing'] | [ 2.16299355e-01 3.37963998e-01 -2.89261080e-02 -4.71606731e-01
-8.54724407e-01 -6.98994040e-01 2.29132578e-01 9.75838006e-01
-9.01358366e-01 9.45909202e-01 6.37933612e-01 -5.67613482e-01
-3.96903247e-01 -3.07412505e-01 -1.93526506e-01 -6.60152256e-01
1.34954154e-01 6.50555015e-01 -1.57850131e-01 3.81241918e-01
1.16207423e-02 6.89697862e-01 -7.09516406e-01 5.43392301e-01
6.41095579e-01 3.49196523e-01 -3.37825239e-01 6.21674776e-01
4.07139808e-02 6.94742203e-01 -7.29178488e-01 -7.44086206e-01
2.98016697e-01 -2.82646537e-01 -1.03598976e+00 -4.21386153e-01
4.80237752e-02 -1.20256562e-02 -8.31393823e-02 1.29389334e+00
6.42597854e-01 -2.74310708e-01 3.47446203e-01 -8.73245776e-01
-3.78103614e-01 6.18171155e-01 2.99057197e-02 1.29779860e-01
3.29254419e-01 -2.30853707e-01 8.11123192e-01 -2.33109802e-01
1.30189252e+00 3.23652625e-01 9.65921164e-01 6.43278241e-01
-1.25136006e+00 -6.30973637e-01 -4.82783675e-01 -3.56867284e-01
-1.57837152e+00 -5.59966683e-01 -4.46174331e-02 -3.94529015e-01
7.61744440e-01 6.56412661e-01 4.61199880e-01 1.09981132e+00
4.43550766e-01 9.62191746e-02 1.02991581e+00 -4.09073144e-01
2.55657613e-01 7.43306816e-01 1.87375173e-01 3.73134136e-01
7.38539696e-01 6.41656807e-03 -1.38619974e-01 -1.10963380e+00
1.24152780e-01 -1.14289053e-01 -3.21589351e-01 -3.59231263e-01
-1.11143029e+00 6.16475523e-01 8.16627815e-02 8.12610924e-01
-4.60687697e-01 -5.10495901e-01 7.55596578e-01 2.54886985e-01
5.86867929e-02 8.43674421e-01 -6.05905235e-01 -1.86140895e-01
-1.10607898e+00 -2.89838165e-01 1.05564976e+00 1.02626336e+00
2.30442733e-02 -6.10029638e-01 2.01578185e-01 6.35222435e-01
2.08605736e-01 -1.39322191e-01 7.82912374e-01 -4.37917560e-01
2.71052033e-01 5.43373764e-01 1.65115982e-01 -8.97030830e-01
-4.73114580e-01 -3.52693081e-01 -8.04719865e-01 -7.87813365e-02
5.72883427e-01 -4.11172181e-01 -6.91344261e-01 1.54599130e+00
2.10448295e-01 -2.15815619e-01 5.88574648e-01 4.21085685e-01
7.34675348e-01 -6.80839643e-02 7.77979851e-01 -3.29608560e-01
1.55331862e+00 -1.78566054e-01 -9.80516195e-01 1.83104128e-01
1.28316057e+00 -7.91593254e-01 6.55439198e-02 2.53595263e-01
-6.27091646e-01 1.42109215e-01 -7.54422188e-01 4.87063155e-02
-6.15763247e-01 -2.92995721e-01 5.27728736e-01 1.14420784e+00
-8.59997630e-01 5.54844499e-01 -9.42696929e-01 -1.01573753e+00
4.74390537e-01 5.20429432e-01 -9.98731673e-01 5.03283925e-02
-1.26491737e+00 8.94608021e-01 5.39466679e-01 8.54282156e-02
-8.83450061e-02 -7.12093771e-01 -4.53119427e-01 -2.50302911e-01
8.81072581e-02 -5.77533841e-01 9.81080234e-01 -7.66921520e-01
-5.85558474e-01 1.35712767e+00 2.07332037e-02 -4.32518005e-01
6.88005567e-01 1.17218807e-01 -1.24810922e+00 -2.15821594e-01
2.16385290e-01 -3.67102325e-02 -1.53378621e-01 -9.29001033e-01
-8.49898934e-01 -6.55382693e-01 -6.31388664e-01 -1.00650750e-01
-1.09319352e-01 5.27818263e-01 1.27434030e-01 -6.25737786e-01
-6.73709065e-02 -9.50040102e-01 -6.08400881e-01 -2.91107446e-01
-7.40632474e-01 3.85440737e-01 3.05318803e-01 -9.95751798e-01
1.59821320e+00 -2.34744978e+00 -4.91741747e-01 5.80449462e-01
4.52106833e-01 4.48421657e-01 3.82058501e-01 5.31975150e-01
-3.89704645e-01 8.35482538e-01 -5.64632751e-02 1.86400071e-01
-3.51251364e-01 9.13507268e-02 -1.26409354e-02 5.49602330e-01
-2.46369570e-01 5.09017348e-01 -8.75593603e-01 -7.49541104e-01
-2.06289783e-01 3.76321852e-01 -2.66673982e-01 -2.33063605e-02
4.84599143e-01 3.38340372e-01 -6.70468688e-01 3.38152319e-01
5.47023773e-01 -1.44312397e-01 8.20505857e-01 -1.55889556e-01
-4.97400761e-02 4.41783637e-01 -1.19509614e+00 1.49385226e+00
2.33412231e-03 2.97446787e-01 -1.97580885e-02 -2.86855131e-01
6.17107272e-01 8.98604870e-01 7.87005007e-01 -2.06896380e-01
1.29075021e-01 4.74096626e-01 1.33026928e-01 -5.94630301e-01
5.62492549e-01 -4.23933387e-01 -2.78810561e-01 4.41258639e-01
-1.37269735e-01 6.04859412e-01 -3.15196514e-01 6.26975447e-02
1.37359941e+00 -4.95968372e-01 7.67431736e-01 -2.87466198e-01
4.29794639e-01 3.34185302e-01 1.01598251e+00 5.38208961e-01
-5.01198769e-01 5.75271130e-01 3.44206542e-01 -3.37530971e-01
-1.02551675e+00 -8.57918739e-01 -7.57319689e-01 3.30294341e-01
-5.22601783e-01 -6.59814894e-01 -6.75139844e-01 -1.02262211e+00
-7.30889738e-02 8.21106493e-01 -6.60758018e-01 -1.33818954e-01
-1.92338496e-01 -7.31598318e-01 1.20603120e+00 2.51318753e-01
3.55168134e-02 -8.04399908e-01 -7.57011414e-01 5.78090191e-01
5.11212125e-02 -1.04793251e+00 -5.35394788e-01 4.62538183e-01
-8.13351512e-01 -1.40579844e+00 -2.15164348e-01 -6.98850811e-01
7.41019487e-01 -5.06305337e-01 8.00379515e-01 -2.51522437e-02
-4.76582885e-01 2.15465173e-01 -1.23608284e-01 -6.92886174e-01
-9.27642167e-01 2.34737188e-01 -3.88833545e-02 -1.36465997e-01
1.21624947e+00 -9.33121368e-02 -3.90658647e-01 1.19994722e-01
-1.02526248e+00 -6.22497559e-01 3.94131273e-01 7.30179727e-01
4.52866167e-01 5.71506545e-02 3.57796520e-01 -1.74227965e+00
5.83899200e-01 -5.15503585e-01 -1.71295509e-01 7.79119730e-01
-9.80824351e-01 -1.82178896e-02 3.93464684e-01 -1.23964809e-01
-1.06134093e+00 3.99910241e-01 -1.18077189e-01 1.24560498e-01
-5.55319250e-01 5.24589419e-01 -2.94663489e-01 1.93164319e-01
6.96754754e-01 -2.89052486e-01 -1.26070842e-01 -5.44614792e-01
-7.67994300e-02 1.17180860e+00 3.11358631e-01 -7.29202926e-02
3.78670633e-01 2.84572482e-01 -3.51829350e-01 -8.63288581e-01
-3.69499892e-01 -7.46976316e-01 -6.38832450e-01 4.23589528e-01
1.27828884e+00 -7.66196012e-01 -4.78153139e-01 2.02524424e-01
-7.90931404e-01 3.17128897e-01 -4.48830962e-01 6.31673694e-01
-6.55873641e-02 3.76491904e-01 -5.30135036e-01 -6.45904422e-01
-4.22479391e-01 -8.30972254e-01 3.16953629e-01 5.12386411e-02
-1.03898871e+00 -1.10729647e+00 4.89167273e-01 3.61155957e-01
1.68196440e-01 4.40680623e-01 8.93801630e-01 -1.53810513e+00
5.36494143e-02 -7.66339064e-01 1.79971576e-01 9.23458580e-03
4.48370814e-01 1.55634418e-01 -9.53025639e-01 -3.63382012e-01
1.96023390e-01 1.80549815e-01 1.18335024e-01 -1.72683503e-02
6.47626042e-01 -3.82544577e-01 -6.99880183e-01 4.77383077e-01
1.42602491e+00 4.99403894e-01 7.98460424e-01 4.90871370e-01
5.35208344e-01 5.70523500e-01 2.72624701e-01 2.98132092e-01
6.07800409e-02 5.55293262e-01 -3.56709868e-01 -9.91389379e-02
6.16209626e-01 -3.62950176e-01 -1.53666139e-01 4.67619687e-01
1.24101616e-01 -1.66641980e-01 -1.21580446e+00 6.36849344e-01
-1.48525012e+00 -8.25978339e-01 -3.44172508e-01 2.52013731e+00
8.76546085e-01 -2.86021471e-01 -6.49620593e-02 -2.14225322e-01
8.48042727e-01 -4.54142630e-01 -3.70759994e-01 -7.50166416e-01
-1.11665659e-01 2.55988538e-01 1.00734091e+00 1.27043441e-01
-9.45451319e-01 4.33758259e-01 6.36732149e+00 1.05223380e-01
-7.96450078e-01 1.89409286e-01 6.75134897e-01 -1.08583696e-01
1.23916601e-04 3.01531870e-02 -6.11813426e-01 3.46278250e-01
1.40232313e+00 -4.98712599e-01 -2.47917786e-01 6.71082973e-01
1.48357809e-01 2.37974562e-02 -1.21655309e+00 7.75006533e-01
-3.19076419e-01 -1.18115914e+00 -1.66736946e-01 7.19334483e-01
3.75351429e-01 1.20169103e-01 -2.09096402e-01 -1.97691917e-01
7.05981910e-01 -1.11643553e+00 4.37064290e-01 6.03981137e-01
8.62030864e-01 -6.75462008e-01 1.17316461e+00 7.37260133e-02
-5.74180484e-01 8.18736479e-02 -3.64316106e-01 7.87155628e-01
1.58543825e-01 3.87061417e-01 -1.23431754e+00 7.45913684e-01
7.55046725e-01 2.56346047e-01 -5.68732083e-01 1.30053043e+00
1.54944316e-01 4.93025154e-01 -2.59370714e-01 2.40664139e-01
3.35826986e-02 3.64510082e-02 3.93087715e-01 1.25619447e+00
4.16967750e-01 3.18792582e-01 -2.98835099e-01 4.45931137e-01
-1.77044228e-01 5.99891603e-01 -8.54023635e-01 -5.59313595e-01
4.91734207e-01 1.15429032e+00 -5.20922959e-01 -1.58124954e-01
-5.80580056e-01 8.52112114e-01 -3.29283960e-02 -1.62011422e-02
-2.88097531e-01 -3.49129051e-01 8.83135259e-01 2.40210280e-01
2.86154598e-02 3.81014973e-01 -1.99269459e-01 -1.04866385e+00
-1.94006130e-01 -1.10166097e+00 1.35412824e+00 -3.42151135e-01
-1.30271113e+00 7.98405528e-01 -4.05595362e-01 -1.12941408e+00
-3.60151321e-01 -2.93351322e-01 -3.90034579e-02 1.08811057e+00
-8.18736851e-01 -7.29706168e-01 2.06890851e-01 4.87611264e-01
-5.76217115e-01 -1.56946376e-01 1.56744289e+00 3.50505233e-01
-7.42821395e-01 9.28079307e-01 5.87183058e-01 6.64436221e-01
1.07156014e+00 -1.10577214e+00 1.17407970e-01 5.84075570e-01
9.43616107e-02 1.15767944e+00 4.99509871e-01 -9.21855569e-01
-6.95545912e-01 -9.45982099e-01 1.72121000e+00 -8.85024548e-01
6.48786843e-01 -7.12449849e-02 -1.30986106e+00 9.17368174e-01
9.24401879e-02 -1.67507753e-01 1.65777683e+00 1.71838522e-01
-3.42794001e-01 1.32240057e-01 -1.74531889e+00 3.56663018e-01
7.06038773e-01 -7.77598381e-01 -6.62286341e-01 1.94618881e-01
9.39708650e-02 -7.94586316e-02 -1.52219331e+00 -1.23386234e-01
5.04280448e-01 -6.69287920e-01 5.79071522e-01 -1.39463186e+00
-1.75110161e-01 -2.47632399e-01 -1.11838378e-01 -9.62282717e-01
-2.52206951e-01 -7.29904950e-01 7.31754959e-01 1.56574118e+00
8.64742756e-01 -9.06590879e-01 8.21440518e-01 1.55617058e+00
4.33304846e-01 -2.84824427e-02 -1.09197199e+00 -5.03497183e-01
1.12984344e-01 -1.04559548e-01 8.02925646e-01 1.66664350e+00
4.98749465e-01 -2.75040828e-02 -4.02226634e-02 4.63318527e-01
4.55018103e-01 -4.31799620e-01 2.96063691e-01 -1.33159256e+00
-9.50482562e-02 2.79757567e-02 -7.95830786e-01 2.75290996e-01
-1.74922973e-01 -1.02629519e+00 -1.94896296e-01 -1.30504906e+00
3.54402214e-01 -7.13670850e-01 -7.42064297e-01 6.16943061e-01
-7.20352903e-02 -1.26474127e-01 -8.50975979e-03 4.20105219e-01
-3.06022972e-01 -3.50934803e-01 3.56801569e-01 3.82757097e-01
-3.17386091e-01 3.72834951e-02 -1.05489314e+00 5.28453529e-01
8.75293195e-01 -1.21630692e+00 -1.94466501e-01 -1.07151337e-01
2.44099006e-01 6.18287139e-02 3.80363241e-02 -7.88005114e-01
5.74154019e-01 1.17080674e-01 3.79404426e-01 -1.04720742e-01
-4.42191005e-01 -1.30610216e+00 1.04137015e+00 6.81565821e-01
-5.79346478e-01 1.59135878e-01 2.94584006e-01 6.50434554e-01
-1.79687902e-01 -4.55319047e-01 6.59375429e-01 -3.05860519e-01
-3.52361262e-01 1.24370433e-01 -6.51143193e-01 1.71415344e-01
1.21687078e+00 -5.07837832e-01 -5.50543144e-02 1.41956896e-01
-1.17124915e+00 9.52160079e-03 8.51299465e-01 2.51189888e-01
1.35228693e-01 -1.00109029e+00 -5.47199249e-01 1.76549762e-01
5.71678698e-01 -3.82301688e-01 3.40071946e-01 8.03915143e-01
-5.90925753e-01 4.81270403e-01 -2.63240725e-01 -1.51756185e-03
-1.80485427e+00 8.11250269e-01 4.99540478e-01 -5.24461806e-01
-7.96125472e-01 2.73147225e-01 -8.84830505e-02 -5.01665711e-01
-1.92716978e-02 4.05765586e-02 -1.35721460e-01 1.47760168e-01
5.79213798e-01 8.01180899e-02 5.10670424e-01 -7.16647387e-01
-6.81451559e-01 -9.36124101e-02 -5.71445465e-01 -8.98870155e-02
1.51429296e+00 -6.34430116e-03 -1.02584757e-01 2.10429445e-01
1.26740229e+00 3.10873300e-01 -3.62692207e-01 -1.61631286e-01
6.36135280e-01 -2.56433547e-01 -3.47485207e-02 -1.06164789e+00
-8.54173243e-01 3.01270008e-01 7.46322811e-01 9.29801390e-02
8.79337192e-01 -3.20269495e-01 3.35842639e-01 3.52399886e-01
4.47212815e-01 -1.19255030e+00 -1.23280561e+00 -3.65239121e-02
3.03478926e-01 -9.73291159e-01 4.03895825e-02 -2.76218057e-01
-8.38057756e-01 9.69506025e-01 -6.83726594e-02 6.45507038e-01
9.40147698e-01 2.68534005e-01 4.97902930e-01 -3.92327756e-01
-4.08423156e-01 4.42887694e-01 -5.38430549e-02 7.41935313e-01
5.62455595e-01 1.98494837e-01 -6.74471498e-01 6.84436977e-01
-1.33952484e-01 3.53941858e-01 7.89150953e-01 1.13542736e+00
3.94359022e-01 -1.58098209e+00 -1.85083091e-01 6.82881713e-01
-1.27917445e+00 -1.84251323e-01 -7.81129062e-01 9.55665231e-01
3.63174118e-02 8.25508654e-01 -2.09256038e-01 -2.79521465e-01
5.80139637e-01 5.99489033e-01 -1.01510979e-01 -7.50744820e-01
-1.21959853e+00 -2.30738759e-01 6.11785710e-01 -3.46815556e-01
-3.48842889e-01 -1.18528569e+00 -1.22175026e+00 -3.08461070e-01
-2.32509106e-01 7.68153489e-01 6.44898117e-01 5.06622493e-01
6.75075531e-01 2.24851191e-01 1.89114675e-01 8.22065890e-01
-5.31203210e-01 -4.08689648e-01 -9.07711864e-01 6.37823284e-01
1.03262655e-01 8.92237127e-02 -1.43119022e-01 1.58365086e-01] | [6.809483051300049, 7.027385234832764] |
c94fe118-e2da-4e09-8dea-994459e18fae | retico-an-incremental-framework-for-spoken | null | null | https://aclanthology.org/2020.sigdial-1.6 | https://aclanthology.org/2020.sigdial-1.6.pdf | Retico: An incremental framework for spoken dialogue systems | In this paper we present the newest version of retico - a python-based incremental dialogue framework to create state-of-the-art spoken dialogue systems and simulations. Retico provides a range of incremental modules that are based on services like Google ASR, Google TTS and Rasa NLU. Incremental networks can be created either in code or with a graphical user interface. In this demo we present three use cases that are implemented in retico: a spoken translation tool that translates speech in real-time, a conversation simulation that models turn-taking and a spoken dialogue restaurant information service. | ['Thilo Michael'] | null | null | null | null | sigdial-acl-2020-7 | ['spoken-dialogue-systems'] | ['speech'] | [-4.79898900e-01 7.00786352e-01 3.23479801e-01 -7.37054527e-01
-4.55352634e-01 -7.34606445e-01 9.19690669e-01 -1.27581581e-01
-1.04838185e-01 6.98707998e-01 7.73652136e-01 -9.07915652e-01
2.31094271e-01 -7.25875735e-01 -1.40365958e-01 1.31461844e-01
-1.34570867e-01 1.09488583e+00 3.80889773e-01 -1.27939868e+00
1.10241212e-01 3.10852915e-01 -1.40314186e+00 5.46817422e-01
6.61720157e-01 1.97189957e-01 1.20574549e-01 1.56393480e+00
-6.58455014e-01 1.15475941e+00 -9.87250566e-01 1.99217603e-01
9.67642888e-02 -8.31349015e-01 -1.11532295e+00 -3.43244851e-01
-2.26116300e-01 -4.79949653e-01 -2.73197711e-01 4.42822248e-01
1.03243256e+00 1.78408667e-01 -1.00428134e-01 -1.39212453e+00
-8.79874080e-02 1.07319546e+00 5.53180933e-01 1.06419593e-01
1.50682306e+00 5.45845687e-01 2.69995421e-01 -5.71894765e-01
6.47987723e-01 1.84331906e+00 7.06566095e-01 7.50881553e-01
-1.28669262e+00 -4.92781281e-01 -4.43453968e-01 -3.76043200e-01
-9.40303147e-01 -9.36596096e-01 2.36889943e-01 2.12725215e-02
1.68596041e+00 7.86940694e-01 8.82379413e-01 1.17948794e+00
-1.98134780e-02 7.07187176e-01 1.34985256e+00 -7.13994622e-01
2.95839429e-01 3.01966131e-01 4.30246741e-01 5.92642903e-01
-1.01601493e+00 -1.74337178e-01 -9.30652082e-01 -4.47239667e-01
7.63633788e-01 -8.88868093e-01 4.83505428e-03 4.56628859e-01
-1.30395401e+00 6.09641016e-01 -1.88294321e-01 2.92906195e-01
-2.84588963e-01 -3.21139544e-01 7.03808308e-01 1.04823279e+00
4.85014826e-01 4.83836293e-01 -3.02187473e-01 -1.09998918e+00
-5.58944941e-01 5.46556532e-01 2.00304437e+00 1.09520578e+00
5.02720952e-01 1.63052246e-01 -2.46376216e-01 1.20955777e+00
6.13002658e-01 9.16096717e-02 6.59317434e-01 -1.28860712e+00
2.85525978e-01 3.74950171e-01 4.03500110e-01 -3.28404188e-01
-5.79042852e-01 4.10343319e-01 -1.56999126e-01 1.43669307e-01
3.47266108e-01 -7.86067426e-01 -5.06353140e-01 1.07743764e+00
5.07108510e-01 -1.21264800e-01 6.32460833e-01 7.26102769e-01
1.21869242e+00 9.44257617e-01 -2.14140207e-01 1.06874667e-02
1.10870337e+00 -1.37613535e+00 -1.08772254e+00 -4.47799452e-02
7.92672336e-01 -1.15661538e+00 1.38788831e+00 1.94573894e-01
-1.39448714e+00 -3.06918532e-01 -7.97105372e-01 -6.01535216e-02
-6.56686485e-01 -5.40370584e-01 4.69439238e-01 6.92273021e-01
-1.77714312e+00 4.41932261e-01 -8.15830350e-01 -9.42475140e-01
-8.05399656e-01 2.78968453e-01 1.59471095e-01 4.83213425e-01
-1.66242659e+00 1.20173025e+00 1.40459538e-01 7.61977509e-02
-6.65071726e-01 -4.60942656e-01 -9.35065091e-01 -6.38791248e-02
7.54790008e-02 -6.68868959e-01 2.48973536e+00 -7.86593676e-01
-2.81531477e+00 4.20484960e-01 -1.10208422e-01 -6.23194396e-01
7.59326100e-01 -2.09688589e-01 -5.53408742e-01 -1.39152169e-01
4.59337495e-02 4.42593127e-01 -2.27366179e-01 -9.15155590e-01
-5.21635711e-01 5.20262821e-03 3.52686763e-01 7.41412044e-01
6.30604923e-01 5.45245111e-01 -2.60784537e-01 -9.91915073e-03
-4.49553281e-01 -6.71795249e-01 -3.26450914e-01 -3.22631776e-01
-5.05909801e-01 -2.45729640e-01 8.30977798e-01 -7.04446673e-01
1.13362741e+00 -1.47329915e+00 -3.62256974e-01 -2.95549482e-02
-4.71215218e-01 5.42974651e-01 -5.16882353e-02 1.47716868e+00
2.71905482e-01 -2.85369754e-02 2.53964011e-02 -4.73658413e-01
5.51252484e-01 3.77549231e-01 -4.26797494e-02 -3.06668133e-01
-3.54036391e-01 6.76766455e-01 -9.61742997e-01 -2.63354152e-01
8.25019062e-01 3.35609674e-01 1.34451156e-02 6.13740027e-01
-5.02838850e-01 5.55559695e-01 -6.20950460e-02 2.08267510e-01
3.19989115e-01 2.77880818e-01 4.02005255e-01 7.26800740e-01
-9.95266795e-01 1.22705150e+00 -1.13310075e+00 2.08066678e+00
-9.77084637e-01 7.33799994e-01 6.56802773e-01 -2.58553743e-01
1.10051632e+00 8.70284259e-01 -2.83593416e-01 -6.70535564e-01
-5.23551330e-02 2.57712424e-01 -1.30325496e-01 -8.11871588e-01
9.84910548e-01 3.25500607e-01 -1.79210529e-02 9.14615273e-01
1.85948476e-01 -6.58466339e-01 1.84281290e-01 4.43158597e-01
9.17590857e-01 1.40233040e-01 5.28729498e-01 -3.89139056e-01
5.40655911e-01 3.30276072e-01 -1.14507511e-01 8.08326483e-01
-5.62685654e-02 3.06664497e-01 5.87994099e-01 -1.12210922e-01
-1.17912543e+00 -6.50115311e-01 1.69993252e-01 1.48215544e+00
-4.99437749e-01 -5.42301059e-01 -1.13144124e+00 -1.43146306e-01
-5.49553275e-01 1.32626688e+00 -7.30740651e-02 5.31860828e-01
-3.57999712e-01 -1.12189367e-01 1.04954147e+00 -6.38016313e-02
6.80688083e-01 -1.42101765e+00 -4.25099313e-01 4.49030548e-01
-6.69399202e-02 -8.81669700e-01 -6.63547695e-01 -1.09296612e-01
-2.02951580e-01 -6.14824951e-01 -6.09809041e-01 -8.30871165e-01
-1.14878900e-01 1.39299855e-01 1.28712857e+00 -1.77654177e-01
-5.22334911e-02 6.60807550e-01 -4.94444758e-01 -3.99665564e-01
-1.63181984e+00 2.02783704e-01 -9.35695022e-02 -4.41243440e-01
1.37201130e-01 -6.59225702e-01 -3.07643205e-01 4.36333656e-01
-5.64877391e-01 4.85886663e-01 -2.89700538e-01 3.88925493e-01
-6.02277398e-01 -8.23649943e-01 7.64489293e-01 -9.20706391e-01
1.60982144e+00 -4.24859673e-01 -4.77514952e-01 2.92612702e-01
-2.96896160e-01 -2.00549707e-01 5.65787137e-01 -2.08217520e-02
-1.20418942e+00 3.31562124e-02 -8.38816702e-01 3.94198090e-01
-4.41070616e-01 7.42624581e-01 1.73382938e-01 2.16322079e-01
8.16039145e-01 5.14767408e-01 6.20595336e-01 -1.90287396e-01
4.68824089e-01 1.54516602e+00 4.08712387e-01 -2.74340421e-01
-9.38673317e-02 -2.45833084e-01 -9.21584547e-01 -1.29145885e+00
-2.23831534e-02 -5.98256052e-01 -4.84354645e-01 -6.02611184e-01
4.19112831e-01 -7.75288761e-01 -9.54086363e-01 8.48578691e-01
-1.34141910e+00 -1.31876755e+00 -1.85474515e-01 2.36100197e-01
-6.92246497e-01 2.10996166e-01 -8.49361062e-01 -9.53483880e-01
-8.24451268e-01 -9.68581080e-01 7.21928775e-01 4.42995369e-01
-9.76173997e-01 -1.09925950e+00 5.80366254e-01 1.09908037e-01
1.00019550e+00 -1.75706953e-01 2.56856352e-01 -9.59356844e-01
7.18226358e-02 2.21081674e-01 1.57903269e-01 1.87289983e-01
1.09989941e-01 5.48021495e-01 -9.62508559e-01 8.53055567e-02
-4.21698391e-01 -4.85096961e-01 -2.33908877e-01 -1.05489008e-01
-1.12296581e-01 -5.95193326e-01 3.62314321e-02 -1.27977848e-01
5.76008976e-01 5.71917951e-01 4.93855119e-01 2.71903187e-01
2.94618923e-02 9.00285065e-01 5.56665063e-01 5.13711035e-01
8.66695940e-01 7.66193509e-01 -6.69161826e-02 -2.98399329e-02
-5.53099951e-03 -3.14103037e-01 7.74953902e-01 1.43732882e+00
5.02417982e-01 -1.21920936e-01 -1.38858235e+00 3.45335722e-01
-2.21163726e+00 -9.31917787e-01 -4.46661979e-01 2.02486706e+00
1.28623056e+00 -1.52433470e-01 4.65947509e-01 -3.45291585e-01
5.84536374e-01 2.17229098e-01 -1.96302850e-02 -1.47010338e+00
2.17097551e-01 2.03726292e-01 -3.80905680e-02 1.38207805e+00
-4.17851955e-01 1.14970529e+00 7.33607626e+00 1.86104253e-01
-1.22385502e+00 1.90440670e-01 1.72116265e-01 6.20988123e-02
-8.62586871e-02 -1.20140925e-01 -5.85318029e-01 1.88568890e-01
1.88279414e+00 -6.88372374e-01 7.79721320e-01 7.32667744e-01
1.03591454e+00 -2.95755506e-01 -9.03332114e-01 3.41242999e-01
-5.92069663e-02 -1.36162519e+00 -1.51599675e-01 -2.77936786e-01
9.94354263e-02 4.23200071e-01 -5.81281364e-01 6.82404160e-01
1.21159494e+00 -8.55012596e-01 5.81090391e-01 5.38391113e-01
6.04511023e-01 -6.70022547e-01 4.77474958e-01 6.78679705e-01
-7.66758204e-01 4.25852865e-01 7.23469779e-02 -3.62958163e-01
6.34000301e-01 -1.96430329e-02 -1.70130944e+00 2.80776739e-01
4.60673302e-01 2.86210165e-03 -5.26119210e-02 9.51280296e-01
-1.64399892e-01 7.82389104e-01 -6.03039324e-01 -4.61964279e-01
4.26122487e-01 -2.42901593e-01 6.76493764e-01 1.75328624e+00
7.27872849e-02 1.91265449e-01 2.98469305e-01 4.33564037e-01
2.99641073e-01 1.98774278e-01 -6.72339916e-01 3.29638302e-01
7.94377506e-01 1.18954217e+00 -3.52182299e-01 -5.63969016e-01
-2.20427051e-01 1.21517110e+00 -2.61799842e-01 2.58210152e-01
-6.30857885e-01 -6.79963887e-01 6.53896987e-01 1.92242451e-02
-4.52087522e-01 -5.57752728e-01 3.02201062e-01 -6.68352544e-01
-4.93453622e-01 -1.34712577e+00 -4.27400380e-01 -1.09843338e+00
-6.65379465e-01 8.99695635e-01 8.16910788e-02 -5.60414612e-01
-1.05338192e+00 -4.78388800e-04 -1.14206553e+00 1.25529134e+00
-8.00034702e-01 -9.26009774e-01 -4.20000702e-01 5.01035273e-01
1.18188131e+00 -5.73508814e-02 1.59726000e+00 1.88896045e-01
-4.48804498e-01 8.17684084e-02 1.20781630e-01 3.53049650e-03
6.97619259e-01 -1.62937367e+00 1.40441430e+00 3.32476348e-02
-3.45375866e-01 6.20106161e-01 1.20265496e+00 -4.11113858e-01
-1.19245052e+00 -6.18434072e-01 1.01522279e+00 -1.23169608e-02
9.19552445e-01 -7.92751431e-01 -8.60029936e-01 8.33186269e-01
1.12549436e+00 -5.22957683e-01 6.60554349e-01 2.43434124e-03
2.42787614e-01 2.51950055e-01 -9.82528508e-01 8.40992212e-01
5.86230755e-01 -8.58987868e-01 -7.21177876e-01 4.47311759e-01
9.12676990e-01 -1.01866591e+00 -9.07517374e-01 -3.66001487e-01
8.74234021e-01 -1.15023983e+00 2.98198462e-01 -3.37340742e-01
-8.75380859e-02 -5.70561737e-02 9.24340338e-02 -1.77076972e+00
6.49475574e-01 -1.85819459e+00 2.55995959e-01 1.39068139e+00
7.49555290e-01 -1.23447394e+00 4.42170680e-01 9.19938207e-01
-3.74333382e-01 -1.39191538e-01 -1.01713109e+00 -3.52599680e-01
-2.56077915e-01 -2.82742232e-01 8.25269818e-01 9.65955436e-01
7.20369041e-01 9.65881824e-01 -1.41664324e-02 -5.33993430e-02
-2.26203606e-01 -6.13856196e-01 1.31264567e+00 -7.71014392e-01
-3.34009320e-01 -1.24727145e-01 2.08646804e-01 -1.43616736e+00
-1.91791400e-01 -3.55391771e-01 3.82411778e-01 -1.82970178e+00
-7.33336747e-01 -2.43312687e-01 6.75392628e-01 3.82981479e-01
5.78839719e-01 -5.27910411e-01 2.88326651e-01 3.63013521e-02
-3.96138579e-01 5.66169739e-01 8.78794372e-01 2.31186539e-01
-8.46642673e-01 2.92214423e-01 -1.53086632e-01 4.21662867e-01
1.07921255e+00 -2.13232324e-01 -4.38641042e-01 -7.60046542e-02
-3.60335298e-02 8.60014558e-01 9.53709930e-02 -9.26562011e-01
5.85249841e-01 2.67674983e-01 -6.68622673e-01 -6.01015151e-01
4.98188645e-01 -2.56510645e-01 2.20738798e-02 3.23185027e-01
-7.11235225e-01 3.43147457e-01 3.71848077e-01 -1.01333849e-01
-1.42225042e-01 -2.64168501e-01 3.55406493e-01 -5.33271134e-01
-2.52868533e-01 -6.48043990e-01 -1.52395868e+00 -3.11261266e-01
8.42401445e-01 -5.55230193e-02 -7.55104721e-01 -1.23992050e+00
-8.40643287e-01 6.58374369e-01 2.57877588e-01 6.26721740e-01
2.64293492e-01 -5.97155511e-01 -7.48638749e-01 -8.80610547e-04
-2.86597550e-01 -3.32220714e-03 -1.68841258e-02 5.08924246e-01
-1.25689018e+00 7.19587862e-01 -2.19862416e-01 -3.83416116e-01
-1.55870664e+00 -1.87612221e-01 8.28796923e-01 -1.23359978e-01
-4.63664025e-01 3.85901242e-01 -6.72590196e-01 -1.69462097e+00
2.05043986e-01 -3.78740758e-01 -5.94856739e-02 -3.53050493e-02
7.67132342e-01 4.95334268e-01 5.74557930e-02 -4.07043099e-01
5.43469396e-05 -3.64910066e-01 3.06975484e-01 -1.12685752e+00
1.08359754e+00 -5.32638907e-01 -1.79368183e-01 1.32625699e+00
4.92898941e-01 -2.44780779e-02 -8.20901036e-01 5.11941649e-02
-8.57913047e-02 2.93057472e-01 -3.16485733e-01 -1.38737285e+00
4.90713455e-02 6.37385428e-01 3.76425087e-01 9.03829932e-01
3.20292324e-01 -3.72308433e-01 6.07439816e-01 7.43281186e-01
4.45764244e-01 -1.15792060e+00 -4.59234416e-01 1.02299595e+00
1.24446213e+00 -8.84465516e-01 -4.99566168e-01 -3.06535780e-01
-8.66930783e-01 1.32501447e+00 3.83263856e-01 1.58552006e-01
5.72240531e-01 5.51946461e-01 9.75792825e-01 -1.22468948e-01
-1.24872065e+00 5.32128336e-03 -6.43062651e-01 7.10139096e-01
8.65126431e-01 2.01057941e-01 -4.11464870e-01 1.62636533e-01
-7.99517274e-01 2.85950124e-01 1.02405536e+00 1.11557746e+00
-4.78925586e-01 -1.31684780e+00 -5.33929348e-01 -7.05692545e-02
-2.04460502e-01 -3.03103358e-01 -1.03698027e+00 9.11678970e-01
-7.29023993e-01 1.38413489e+00 1.84114173e-01 -2.84664810e-01
5.67395985e-01 4.10606593e-01 -9.54919904e-02 -9.88950789e-01
-1.62348795e+00 -2.31695727e-01 1.30091071e+00 -4.40726727e-01
-1.46349877e-01 -5.15361667e-01 -1.41371238e+00 -8.16228092e-01
-7.49711245e-02 6.33925796e-01 1.12100565e+00 5.78719974e-01
6.41827285e-01 3.79599512e-01 6.27405584e-01 -9.93680358e-01
-4.62389052e-01 -1.68284285e+00 -2.69705623e-01 -6.58642113e-01
1.31933942e-01 3.90986532e-01 -4.67079207e-02 -3.98897111e-01] | [12.886287689208984, 7.935369968414307] |
43847a2c-27c0-4d31-9f3d-1ac4db4f287e | interpretable-machine-learning-accelerated | 2304.03928 | null | https://arxiv.org/abs/2304.03928v1 | https://arxiv.org/pdf/2304.03928v1.pdf | Interpretable machine learning-accelerated seed treatment by nanomaterials for environmental stress alleviation | Crops are constantly challenged by different environmental conditions. Seed treatment by nanomaterials is a cost-effective and environmentally-friendly solution for environmental stress mitigation in crop plants. Here, 56 seed nanopriming treatments are used to alleviate environmental stresses in maize. Seven selected nanopriming treatments significantly increase the stress resistance index (SRI) by 13.9% and 12.6% under salinity stress and combined heat-drought stress, respectively. Metabolomics data reveals that ZnO nanopriming treatment, with the highest SRI value, mainly regulates the pathways of amino acid metabolism, secondary metabolite synthesis, carbohydrate metabolism, and translation. Understanding the mechanism of seed nanopriming is still difficult due to the variety of nanomaterials and the complexity of interactions between nanomaterials and plants. Using the nanopriming data, we present an interpretable structure-activity relationship (ISAR) approach based on interpretable machine learning for predicting and understanding its stress mitigation effects. The post hoc and model-based interpretation approaches of machine learning are combined to provide complementary benefits and give researchers or policymakers more illuminating or trustworthy results. The concentration, size, and zeta potential of nanoparticles are identified as dominant factors for correlating root dry weight under salinity stress, and their effects and interactions are explained. Additionally, a web-based interactive tool is developed for offering prediction-level interpretation and gathering more details about specific nanopriming treatments. This work offers a promising framework for accelerating the agricultural applications of nanomaterials and may profoundly contribute to nanosafety assessment. | ['Fang Cheng', 'Yingchao He', 'Da Liu', 'Maozhen Qu', 'Sam F. Y. Li', 'Dan Luo', 'Hengjie Yu'] | 2023-04-08 | null | null | null | null | ['interpretable-machine-learning'] | ['methodology'] | [ 6.28679633e-01 -4.35149223e-01 -4.45011288e-01 2.80311882e-01
7.89478123e-02 -1.01229954e+00 -9.42205265e-02 1.03761101e+00
2.32886627e-01 7.62597084e-01 -8.03840384e-02 -8.14693391e-01
-3.49494904e-01 -1.11121583e+00 -5.86347520e-01 -1.07474029e+00
-3.36489119e-02 -3.58198255e-01 2.20242068e-02 -6.94494724e-01
4.55816478e-01 6.92027211e-01 -1.62874246e+00 2.58224726e-01
1.37105572e+00 7.15952516e-01 1.05265331e+00 3.19032848e-01
-2.08301917e-01 -3.98564577e-01 -2.41403952e-01 3.90379131e-01
-9.37913507e-02 -2.83199728e-01 -1.70724452e-01 -1.20677352e-01
-3.84045839e-01 1.01173297e-01 6.21187925e-01 1.14914095e+00
5.78043044e-01 -2.46173948e-01 7.23980069e-01 -7.34508812e-01
-1.06108439e+00 7.11371899e-01 -6.98395967e-01 -2.58381426e-01
2.62015820e-01 2.01205030e-01 3.47062975e-01 -7.41838396e-01
1.98742062e-01 9.97631192e-01 2.02960953e-01 3.41163337e-01
-1.07782400e+00 -4.81351882e-01 8.31191465e-02 2.89972782e-01
-8.62317443e-01 -9.81986374e-02 6.17015734e-02 -5.34875810e-01
9.14569318e-01 3.44746351e-01 7.79000342e-01 5.36794543e-01
9.23254490e-01 4.02540341e-02 7.74000585e-01 -3.99087936e-01
5.27243733e-01 -3.67526025e-01 -1.24628007e-01 1.49673894e-01
1.02520704e+00 -1.57062858e-02 -2.41601989e-01 -2.72909738e-02
2.33669594e-01 5.82685880e-02 -2.99929380e-01 1.18600409e-02
-8.89433086e-01 4.26859289e-01 4.99880165e-01 2.52700001e-01
-6.65169120e-01 -1.61101669e-01 4.54619050e-01 -2.87461549e-01
5.56488574e-01 6.73950791e-01 -1.27147102e+00 3.73036951e-01
-9.44441855e-02 -1.02334552e-01 5.97995520e-01 5.16787052e-01
4.82356220e-01 2.46482462e-01 1.09848477e-01 7.32818425e-01
4.43361610e-01 1.19207799e+00 1.38192296e-01 -5.83926558e-01
-2.21428648e-01 6.26228809e-01 3.80466342e-01 -1.05473065e+00
-4.67295200e-01 -1.45158350e-01 -6.86177671e-01 1.82542220e-01
2.46509090e-01 -3.07453036e-01 -7.39836037e-01 1.65049362e+00
3.53832930e-01 -1.14177056e-01 4.05125648e-01 7.47069061e-01
9.67151523e-01 7.69598663e-01 9.48469698e-01 -4.10807461e-01
1.90390110e+00 -8.40433165e-02 -9.10201192e-01 5.63226491e-02
9.03700292e-01 -6.82461858e-01 8.75998437e-01 2.89332122e-01
-8.90482068e-01 2.83071138e-02 -1.26444077e+00 5.17213106e-01
-1.10517156e+00 3.21391165e-01 7.47484148e-01 8.44798386e-01
-4.78810817e-01 7.78083324e-01 -9.41349983e-01 -7.88452685e-01
3.53928775e-01 3.43181670e-01 -2.48991609e-01 -1.16965614e-01
-1.15266538e+00 1.24713850e+00 4.69065517e-01 4.94378746e-01
-5.25319219e-01 -9.03295159e-01 -7.85578310e-01 3.73244733e-01
8.70511979e-02 -4.12108094e-01 6.41604424e-01 5.65444604e-02
-1.69162965e+00 5.03621638e-01 -1.98363677e-01 -1.36704976e-03
-5.22997499e-01 -2.96461403e-01 -3.56578290e-01 -5.08426130e-02
3.12049627e-01 1.87076822e-01 -1.86899126e-01 -1.16608703e+00
-2.04217240e-01 -7.33070970e-01 -1.59767911e-01 -1.10196900e-02
-5.06929278e-01 7.74282962e-02 7.72447944e-01 -2.93553978e-01
6.77237809e-01 -6.04489982e-01 -6.13946378e-01 -1.72413811e-01
-5.36340952e-01 2.03679621e-01 8.32975566e-01 -4.50333923e-01
5.73882878e-01 -1.78642416e+00 -2.32876819e-02 8.13097730e-02
5.21279238e-02 6.81063831e-01 -4.02917892e-01 8.14321041e-01
-2.89396614e-01 8.21285725e-01 -2.61904389e-01 8.82314503e-01
-3.54396075e-01 -2.56630361e-01 1.15623496e-01 4.75234210e-01
5.44093788e-01 8.24353397e-01 -1.26556516e+00 4.06662941e-01
5.31657755e-01 5.28046012e-01 1.09115906e-01 -3.71132314e-01
-4.14157391e-01 3.98819774e-01 -6.44821525e-01 1.12378478e+00
1.41664922e+00 1.72319099e-01 4.31543082e-01 -4.73618418e-01
-8.15974295e-01 -1.60105124e-01 -7.08441794e-01 1.21678412e+00
-1.46867186e-01 1.36574000e-01 1.72392175e-01 -7.02246964e-01
1.07494974e+00 2.08388984e-01 2.94096947e-01 -6.07539415e-01
2.82918394e-01 4.49391097e-01 -1.42715499e-01 -8.28950584e-01
-5.66537157e-02 -1.41409740e-01 4.37254757e-01 -1.00118198e-01
-4.50307548e-01 -9.96002704e-02 3.27544510e-01 -2.34246418e-01
8.48108768e-01 2.05068678e-01 4.78553742e-01 -8.85618925e-01
5.75862408e-01 -1.52937695e-01 6.09389484e-01 2.07015365e-01
-1.42779842e-01 -4.90267612e-02 5.35184562e-01 2.03684177e-02
-7.48213232e-01 -8.27729106e-01 -4.01619673e-01 1.13727546e+00
4.68087494e-01 -7.39861056e-02 -6.68669045e-01 2.18661591e-01
9.47235301e-02 7.74808168e-01 -5.61733782e-01 -1.65501952e-01
1.79066375e-01 -1.40719891e+00 3.51872653e-01 4.11178827e-01
5.39354980e-01 -8.46969068e-01 -4.43897337e-01 4.06637669e-01
1.54772252e-02 -1.00590861e+00 3.98334831e-01 6.37704849e-01
-8.92183542e-01 -1.17677522e+00 -4.45692033e-01 -4.21110272e-01
5.68223894e-01 7.96040595e-01 5.39976835e-01 1.29635736e-01
-3.26919407e-01 -3.16711873e-01 -6.82250142e-01 -1.28517067e+00
-5.52946270e-01 -2.54939646e-02 6.35675415e-02 -6.43015027e-01
4.31008577e-01 -6.26121044e-01 -8.18042219e-01 1.27357662e-01
-1.03937566e+00 -2.23332420e-01 6.45703614e-01 5.40140033e-01
5.49957991e-01 -6.63459525e-02 9.96651590e-01 -7.87817121e-01
4.81593996e-01 -8.90189052e-01 -4.96192515e-01 5.95977247e-01
-6.25707090e-01 -4.37631488e-01 4.52973425e-01 -2.80848473e-01
-9.27459896e-01 7.42851943e-03 1.43411994e-01 8.31250906e-01
-5.87725282e-01 1.17308080e+00 -9.60210264e-01 2.22863797e-02
8.79143119e-01 -1.87828332e-01 -2.51987414e-03 -2.05205187e-01
8.73880982e-02 2.01053560e-01 6.89303651e-02 -3.96780640e-01
5.10887504e-01 1.07744895e-01 5.04472971e-01 -1.26029932e+00
-6.45604134e-01 -1.35311872e-01 -4.79371637e-01 -1.00879088e-01
9.61772323e-01 -7.44999766e-01 -1.24373591e+00 5.88155985e-01
-1.06470454e+00 -1.82910025e-01 4.04364228e-01 6.59795403e-01
3.61491442e-02 2.98283607e-01 -5.49301088e-01 -7.43899167e-01
-6.47078991e-01 -9.28045630e-01 5.61983824e-01 6.66358590e-01
1.27406806e-01 -7.71624804e-01 -1.51116207e-01 2.05955282e-02
5.62070191e-01 8.96561146e-01 1.46211231e+00 -2.24152774e-01
-1.27271056e-01 -2.08031628e-02 -4.03699130e-01 1.58062458e-01
7.53497124e-01 5.79120815e-01 -9.94947135e-01 8.92175958e-02
-2.84154743e-01 -6.74015805e-02 5.48485219e-01 1.06229687e+00
8.64359379e-01 1.02331311e-01 -4.08600509e-01 4.16601717e-01
1.75673735e+00 6.50688410e-01 7.63630807e-01 5.14384449e-01
4.64542121e-01 9.69719946e-01 9.28507328e-01 5.78894019e-01
-1.45739019e-01 -1.76038295e-01 1.24669099e+00 -3.88431221e-01
2.95192719e-01 3.01750809e-01 3.35493207e-01 2.90452689e-01
-4.01957095e-01 -6.89839900e-01 -9.70535457e-01 2.98916370e-01
-1.53265548e+00 -1.03053963e+00 -8.34480941e-01 1.85742855e+00
4.91155326e-01 -2.86658168e-01 -4.29719448e-01 1.71640396e-01
8.65224004e-01 -2.96401709e-01 -8.42554927e-01 -5.98966181e-01
-7.68983543e-01 1.90467656e-01 8.71417165e-01 1.16031483e-01
-7.75430918e-01 1.01025307e+00 6.34057522e+00 2.38815710e-01
-1.34930301e+00 -5.49359918e-01 3.29791844e-01 4.54414755e-01
-3.30814958e-01 2.19354138e-01 -5.34777164e-01 -1.13721816e-02
9.00186002e-01 -3.35770160e-01 5.12423925e-02 5.13129652e-01
1.08655250e+00 -6.52797699e-01 -4.94286418e-01 1.65860921e-01
-5.57414949e-01 -1.42855799e+00 2.55154260e-02 4.97823581e-02
7.18819439e-01 -1.55647250e-03 -1.40072688e-01 -2.42105708e-01
-5.02691977e-02 -7.45203555e-01 3.03793252e-01 4.94586617e-01
6.44112527e-01 -7.08644927e-01 8.86485994e-01 1.38582140e-01
-1.33529341e+00 -1.01880848e-01 -8.05089474e-01 -4.49228495e-01
-1.74478397e-01 1.12690914e+00 -8.09452295e-01 9.25440490e-01
5.98207593e-01 7.61637211e-01 -5.93966067e-01 7.61207163e-01
-2.24194601e-01 6.20878100e-01 -6.70513362e-02 -1.27512276e-01
-2.03234628e-01 -5.20850480e-01 9.37775150e-02 9.89027739e-01
5.64193547e-01 5.84906042e-01 1.82241015e-02 8.47627103e-01
4.15452123e-01 5.34030974e-01 -6.29938543e-01 -4.81032670e-01
8.61967862e-01 1.25855541e+00 -1.43420553e+00 5.22748120e-02
-2.06547782e-01 2.90619016e-01 -6.25909865e-01 3.82682651e-01
-4.58212972e-01 -5.30598104e-01 9.07063007e-01 -6.34509400e-02
1.24199949e-01 -2.85720676e-01 -7.59715378e-01 -7.64564991e-01
-4.91055548e-01 -5.61297894e-01 -2.66870975e-01 -7.53911674e-01
-1.06607068e+00 -3.00308943e-01 -8.05210322e-02 -7.37298727e-01
5.90700328e-01 -8.77217770e-01 -6.92597330e-01 1.08583665e+00
-1.51197052e+00 -1.02423704e+00 -5.17016411e-01 -4.88886684e-01
1.61407515e-01 1.77428663e-01 1.50696754e+00 -3.01857233e-01
-9.80939209e-01 -9.88958329e-02 6.95086777e-01 -6.76771224e-01
5.31670690e-01 -8.80570829e-01 1.07315913e-01 6.22639418e-01
-1.08569157e+00 5.41044950e-01 9.79679346e-01 -9.96286809e-01
-1.49519455e+00 -1.20048714e+00 7.76371777e-01 7.02494802e-03
4.60692942e-01 3.26294638e-02 -8.25783014e-01 -1.02848463e-01
7.97156468e-02 -5.38063705e-01 1.45895624e+00 2.87858956e-02
1.17534131e-01 5.16305715e-02 -1.28689587e+00 7.84460485e-01
5.26760697e-01 1.26758441e-02 2.32335269e-01 4.24714863e-01
8.51468921e-01 -2.39173844e-01 -1.28872144e+00 9.09481823e-01
8.04592371e-01 -5.97609103e-01 9.27589774e-01 -7.20180571e-01
6.48006618e-01 -4.83795106e-01 -3.60133141e-01 -1.13833916e+00
-5.87455034e-01 -1.03946999e-01 2.38335192e-01 1.32758963e+00
4.97775167e-01 -5.48050523e-01 3.71599883e-01 5.98509133e-01
-1.18813157e-01 -7.60776579e-01 -2.45306101e-02 -6.73260987e-01
2.08197966e-01 -9.86721739e-02 9.51357186e-01 1.02503097e+00
1.66180566e-01 -1.28662676e-01 2.81222820e-01 6.49778962e-01
4.78047013e-01 -2.96404600e-01 1.88237637e-01 -1.44673967e+00
6.30332053e-01 -3.22395474e-01 -2.39373699e-01 -3.34921628e-02
1.51959866e-01 -7.17763424e-01 -7.89415613e-02 -1.83025205e+00
-1.01027554e-02 1.16776926e-02 -3.55650842e-01 5.32304943e-01
-4.60981697e-01 -1.44423172e-01 3.33766565e-02 -3.27926993e-01
3.11464429e-01 4.62010473e-01 1.27001309e+00 -2.55543262e-01
-6.57923460e-01 -4.80320007e-02 -1.23150980e+00 5.25067389e-01
1.39824378e+00 -3.67190927e-01 -5.61919451e-01 -3.81327927e-01
4.80076998e-01 5.12610897e-02 1.51178256e-01 -6.13045931e-01
-3.62893492e-01 -9.36908901e-01 6.08519614e-01 -6.37520909e-01
-2.48178422e-01 -5.65426052e-01 1.11282371e-01 9.43178892e-01
-1.13363728e-01 -1.95888132e-01 8.58084857e-01 4.16872561e-01
5.40386021e-01 -3.75082135e-01 5.28903484e-01 4.73665074e-02
-3.55370164e-01 1.94804490e-01 -1.20295286e+00 -5.71138680e-01
9.98295963e-01 -4.69693720e-01 -6.47706985e-01 1.79996982e-01
-7.07175195e-01 2.40643248e-01 3.76970857e-01 1.48088589e-01
4.56516296e-01 -1.09006870e+00 -7.49064863e-01 2.08938606e-02
9.46928859e-02 -3.35633278e-01 5.60010374e-01 6.22803152e-01
-7.73949802e-01 4.07022893e-01 -7.68729329e-01 -4.87171859e-01
-1.13813579e+00 3.50975990e-01 -4.38841879e-02 2.51545995e-01
2.20848873e-01 4.85108435e-01 2.36100286e-01 -1.87903538e-01
-3.39485914e-01 -6.44786596e-01 -7.13634193e-01 1.92732409e-01
4.99878019e-01 7.59435058e-01 2.81656146e-01 -1.20870851e-01
-6.00450993e-01 5.29984713e-01 2.10209370e-01 4.42046404e-01
1.57024777e+00 -1.72410250e-01 -6.52094543e-01 5.55181742e-01
3.53312135e-01 -1.64037883e-01 -8.73169959e-01 4.59226400e-01
2.02478886e-01 -3.03386673e-02 -2.24179015e-01 -1.01024842e+00
-5.58353066e-01 5.72014213e-01 9.06012118e-01 3.98144096e-01
1.30040514e+00 -4.41448718e-01 3.90709698e-01 4.76959139e-01
-7.62269124e-02 -8.29313636e-01 -2.24181920e-01 6.06424212e-01
8.83595824e-01 -1.19465733e+00 1.45769119e-01 -1.04397273e+00
9.94644687e-02 1.36198199e+00 6.41608477e-01 2.96499014e-01
7.25606859e-01 3.81385267e-01 2.30813384e-01 -6.52994495e-03
-6.66235089e-01 -1.91160560e-01 -2.95663029e-01 1.16391850e+00
1.03879237e+00 3.65103543e-01 -9.45530176e-01 6.81276739e-01
2.19498187e-01 -5.84046245e-02 6.89489067e-01 1.05941033e+00
-9.14213359e-01 -1.43438113e+00 -6.00017130e-01 4.32146758e-01
-2.90376157e-01 -2.82783866e-01 -4.07509387e-01 2.31200263e-01
3.42932314e-01 1.37442446e+00 -7.63803363e-01 -2.96691746e-01
5.03816426e-01 -9.20418799e-02 3.49055797e-01 -7.43652165e-01
-4.15172994e-01 3.27325732e-01 -1.21659644e-01 -2.16773450e-01
-6.61483407e-01 -5.58400989e-01 -1.53930795e+00 -4.83649075e-01
-9.62053180e-01 9.23999920e-02 1.34287548e+00 1.08074903e+00
4.90191638e-01 6.83445573e-01 7.59081185e-01 -8.90823185e-01
2.62211636e-02 -8.10450912e-01 -8.14504445e-01 -2.30149046e-01
-1.77166179e-01 -7.76334405e-01 -1.88336492e-01 2.52570778e-01] | [9.269200325012207, -1.5725524425506592] |
2b2243e4-232f-4984-ab62-9ecfa5c92429 | home-activity-monitoring-using-low-resolution | 1811.05416 | null | http://arxiv.org/abs/1811.05416v1 | http://arxiv.org/pdf/1811.05416v1.pdf | Home Activity Monitoring using Low Resolution Infrared Sensor | Action monitoring in a home environment provides important information for
health monitoring and may serve as input into a smart home environment. Visual
analysis using cameras can recognise actions in a complex scene, such as
someones living room. However, although there the huge potential benefits and
importance, specifically for health, cameras are not widely accepted because of
privacy concerns. This paper recognises human activities using a sensor that
retains privacy. The sensor is not only different by being thermal, but it is
also of low resolution: 8x8 pixels. The combination of the thermal imaging, and
the low spatial resolution ensures the privacy of individuals. We present an
approach to recognise daily activities using this sensor based on a discrete
cosine transform. We evaluate the proposed method on a state-of-the-art dataset
and experimentally confirm that our approach outperforms the baseline method.
We also introduce a new dataset, and evaluate the method on it. Here we show
that the sensor is considered better at detecting the occurrence of falls and
Activities of Daily Living. Our method achieves an overall accuracy of 87.50%
across 7 activities with a fall detection sensitivity of 100% and specificity
of 99.21%. | ['Timothy Volonakis', 'Lili Tao', 'Melvyn Smith', 'Kevin Chetty', 'Bo Tan', 'Yanguo Jing'] | 2018-11-13 | null | null | null | null | ['home-activity-monitoring'] | ['miscellaneous'] | [ 6.52284563e-01 2.67090220e-02 2.63874352e-01 -3.61283392e-01
-6.85508788e-01 -1.52110577e-01 3.28212202e-01 -8.65854695e-02
-8.99302900e-01 8.22099447e-01 6.36479855e-01 2.04674825e-01
1.10836394e-01 -6.37175739e-01 -2.76601791e-01 -8.86265218e-01
-1.07062191e-01 -4.96251099e-02 6.36800945e-01 1.36806592e-01
-1.73765212e-01 3.89260739e-01 -1.78320289e+00 5.98392546e-01
4.25727248e-01 9.62209821e-01 -1.81848854e-01 6.19174540e-01
7.54430830e-01 6.65297568e-01 -7.02774763e-01 -2.13583410e-01
3.43712091e-01 -4.30903703e-01 -5.75961471e-01 3.46608162e-01
4.25059974e-01 -7.22956300e-01 -2.08462372e-01 5.84249556e-01
8.53389323e-01 1.23228073e-01 4.87844199e-01 -1.22967422e+00
-1.23742096e-01 -3.76964837e-01 -4.92885321e-01 2.98033923e-01
1.03870976e+00 1.91519573e-01 3.41793418e-01 -2.07736447e-01
2.12925315e-01 8.43655646e-01 7.95355618e-01 6.25680089e-01
-9.77103293e-01 -4.40881401e-01 -1.72798723e-01 4.36557561e-01
-1.42324460e+00 -7.27870703e-01 3.31807226e-01 -2.54901141e-01
1.26069045e+00 6.28355563e-01 8.71742487e-01 1.31001306e+00
2.49211773e-01 7.22369075e-01 1.25385487e+00 -4.99110252e-01
4.25597727e-01 -3.62191088e-02 2.38030516e-02 4.66438293e-01
5.64879298e-01 1.03951670e-01 -5.59398890e-01 -2.79746950e-01
4.77686316e-01 6.36921823e-01 -6.36597037e-01 -3.20766568e-01
-1.14905131e+00 4.61559534e-01 2.53576994e-01 3.27328235e-01
-4.87524092e-01 -1.19660888e-02 3.06568027e-01 -2.68382207e-02
2.47233585e-02 -2.02079087e-01 6.67798799e-03 -5.85619569e-01
-8.33134174e-01 -9.16388035e-02 6.47910953e-01 6.74582005e-01
1.84751704e-01 -5.99733829e-01 -2.74611652e-01 5.65434158e-01
3.57718378e-01 8.29154789e-01 4.43186015e-01 -9.17074382e-01
4.81998444e-01 7.16698587e-01 2.47170478e-01 -6.88919485e-01
-5.22139192e-01 3.33695322e-01 -9.45591152e-01 5.78311920e-01
4.41232413e-01 -1.56691998e-01 -7.48300076e-01 1.19970822e+00
2.47909725e-01 -8.83656666e-02 -9.75311100e-02 1.07277906e+00
4.28982317e-01 1.75150454e-01 1.04356825e-01 -3.94806772e-01
1.60833740e+00 -5.29913187e-01 -9.35900927e-01 -5.46113908e-01
4.24216986e-01 -3.21596324e-01 6.63606524e-01 5.86428046e-01
-7.21526921e-01 -2.44187087e-01 -1.13421261e+00 1.92534342e-01
-2.41019905e-01 1.21573396e-01 1.24221832e-01 1.21725583e+00
-1.04308844e+00 3.14618945e-01 -1.27890146e+00 -1.16232395e+00
6.63206279e-01 3.74523759e-01 -7.91551590e-01 -1.42544493e-01
-8.09358656e-01 1.05514133e+00 1.16804019e-01 -1.36787310e-01
-2.91576475e-01 6.43365011e-02 -8.82008433e-01 -2.66111344e-01
4.28529121e-02 -6.55355811e-01 9.43083286e-01 -5.57741880e-01
-1.15438306e+00 9.87759113e-01 -3.64691556e-01 -7.24520087e-01
1.00964880e+00 -5.06570995e-01 -6.38420761e-01 5.46560764e-01
1.37053639e-01 1.58949167e-01 6.42703950e-01 -8.09955597e-01
-1.03219581e+00 -8.62806916e-01 -1.63574055e-01 2.19193101e-01
-4.12705213e-01 2.47620255e-01 -1.29515365e-01 -2.79096574e-01
-1.46200939e-04 -8.65054548e-01 -6.96689188e-02 2.82965690e-01
-7.39844963e-02 3.86780053e-01 8.35388422e-01 -8.66032541e-01
1.34941709e+00 -2.15256643e+00 -6.47513568e-01 2.60729402e-01
-1.83132626e-02 3.66457641e-01 7.63870656e-01 4.16291744e-01
1.50317952e-01 -3.69047880e-01 -6.19537175e-01 -3.36234540e-01
-2.06337571e-01 2.73208499e-01 3.72074068e-01 9.33462262e-01
-1.78261891e-01 7.27834821e-01 -7.30768621e-01 -6.20061874e-01
7.74370313e-01 8.49328935e-01 -4.14492674e-02 1.29949033e-01
8.60203803e-01 4.29690778e-01 -2.05807000e-01 6.32257640e-01
4.24409926e-01 1.80821478e-01 8.73076320e-02 9.69102047e-03
-2.45216545e-02 -2.58466452e-01 -1.51188481e+00 1.35889781e+00
1.43043652e-01 4.78346318e-01 -2.67092399e-02 -8.69055688e-01
7.10322440e-01 6.10200882e-01 6.06131077e-01 -8.72280538e-01
2.53488645e-02 -1.71479639e-02 -4.64714170e-01 -8.70479763e-01
3.09804846e-02 -2.51445204e-01 1.03709303e-01 4.65967149e-01
-4.87214655e-01 6.28269911e-01 -1.81103081e-01 -2.81455398e-01
1.71268094e+00 1.78053811e-01 7.98677087e-01 -1.75409596e-02
4.61745173e-01 -3.89977336e-01 2.82272369e-01 6.13465071e-01
-8.88087988e-01 9.65626419e-01 -2.73150384e-01 -5.24291098e-01
-6.27300501e-01 -1.08716428e+00 -9.16766301e-02 7.33373404e-01
7.85921067e-02 -2.83019155e-01 -7.53388226e-01 -5.36673069e-01
-4.83632181e-03 4.60649759e-01 -6.84571624e-01 -2.90431350e-01
-6.85163677e-01 -7.94862211e-01 6.43677533e-01 1.00093317e+00
1.04352307e+00 -1.15131140e+00 -1.91041553e+00 8.37796032e-02
-3.92562896e-01 -1.19180381e+00 -2.93655604e-01 1.44833773e-01
-9.35084939e-01 -1.42141473e+00 -9.07548964e-01 -5.87598860e-01
6.54028952e-01 4.04014975e-01 7.76472926e-01 -8.09128135e-02
-5.17671227e-01 9.52972412e-01 -5.26988864e-01 -4.21765953e-01
1.54318377e-01 -3.85322928e-01 7.35972002e-02 3.79255027e-01
1.00601065e+00 -6.50830746e-01 -1.05864513e+00 3.95356596e-01
-7.06587672e-01 -3.56097788e-01 4.99234825e-01 3.38322729e-01
3.63304019e-01 5.83854690e-02 -2.67443508e-02 -3.66151720e-01
5.13186991e-01 -7.65728354e-02 1.22385159e-01 2.60983050e-01
-4.28704441e-01 -1.27328917e-01 1.33544981e-01 -1.83971867e-01
-1.09137499e+00 6.41802669e-01 5.97352460e-02 2.91681737e-01
-8.38234186e-01 -4.70559329e-01 -2.80716538e-01 7.14325458e-02
7.64459014e-01 8.85723904e-02 1.12124018e-01 -5.10077775e-01
-1.32407933e-01 1.14160812e+00 6.36920333e-01 -5.88139445e-02
3.91601861e-01 1.22149205e+00 6.27237745e-03 -1.15839767e+00
-1.76640853e-01 -1.09464490e+00 -9.25434113e-01 -4.83439118e-01
1.18668735e+00 -8.05661082e-01 -8.22310388e-01 6.72165334e-01
-8.07094157e-01 1.11710496e-01 -2.42480859e-01 6.82127118e-01
-6.39235139e-01 6.45138860e-01 -1.88068211e-01 -1.38377368e+00
-5.00106215e-01 -6.04680657e-01 1.34850705e+00 4.40904610e-02
-6.50199771e-01 -7.14145720e-01 6.36833310e-02 5.98783314e-01
3.59119505e-01 1.02377033e+00 -9.59274024e-02 -2.08466023e-01
-9.62721258e-02 -4.86251235e-01 2.79045943e-02 1.47671267e-01
5.91524184e-01 -6.43449903e-01 -1.30150485e+00 -4.05031741e-01
2.73684442e-01 -1.48121314e-02 7.73742795e-01 4.68265742e-01
4.84911054e-01 -2.16401458e-01 -6.32260144e-01 2.90930122e-01
1.32027078e+00 4.72880572e-01 1.35995996e+00 8.59161377e-01
4.30280775e-01 5.31039953e-01 5.66849649e-01 6.31867230e-01
2.64545947e-01 7.15509593e-01 4.77294475e-01 -1.02612272e-01
1.01023279e-01 2.36508012e-01 5.84303260e-01 -1.11618079e-02
-6.47872031e-01 9.19595212e-02 -8.47426534e-01 5.43793142e-01
-2.02265835e+00 -1.27852452e+00 -2.95575202e-01 2.55654001e+00
4.67646331e-01 -3.43052447e-02 5.80602825e-01 8.44458342e-01
6.54940844e-01 -5.14069311e-02 -2.44286433e-01 -1.88876644e-01
1.36273695e-04 2.80815840e-01 8.36935699e-01 3.06759000e-01
-1.38564539e+00 5.10171475e-03 6.18410301e+00 -1.15348995e-01
-5.63676238e-01 1.85636610e-01 1.95850253e-01 -3.19097459e-01
7.20426500e-01 -6.28931284e-01 -5.18695116e-01 5.72249472e-01
8.70733976e-01 2.48096585e-01 3.47501896e-02 7.23353684e-01
4.41742390e-01 -6.88386977e-01 -1.12314272e+00 1.28103554e+00
4.69878793e-01 -4.11685646e-01 -4.10690337e-01 2.67479539e-01
9.96910930e-02 -3.95246446e-01 -4.71171290e-01 -2.28414282e-01
-2.01339453e-01 -1.04145539e+00 6.18528545e-01 6.60098493e-01
6.30705535e-01 -8.15663457e-01 9.46684241e-01 4.74108070e-01
-1.36932147e+00 -2.06483454e-01 -4.58299071e-02 -4.49253500e-01
3.90611589e-01 1.40994504e-01 -5.33614278e-01 2.25423664e-01
1.42264509e+00 7.28271723e-01 -7.57459402e-01 1.22427857e+00
-1.34338766e-01 3.51810962e-01 -5.83915591e-01 -1.97794847e-02
-1.96873277e-01 -6.14863969e-02 3.29284251e-01 1.47234845e+00
4.03736562e-01 5.58825672e-01 3.63004133e-02 -7.71466829e-03
6.66995347e-01 -8.72549713e-02 -8.65091145e-01 6.18805349e-01
-1.74276419e-02 6.38759971e-01 -6.23601735e-01 -1.22669689e-01
-6.41370952e-01 1.51387811e+00 -1.90463632e-01 1.33604243e-01
-5.97668588e-01 -3.17373931e-01 6.99096024e-01 6.09367609e-01
3.96240860e-01 -5.36256619e-02 -1.76024824e-01 -8.93139422e-01
6.80857301e-01 -8.05014670e-01 6.88552618e-01 -7.44495273e-01
-9.06943262e-01 2.27179736e-01 2.56497681e-01 -1.54304588e+00
-1.78731695e-01 -6.46883368e-01 -2.89459288e-01 4.98290300e-01
-1.02294540e+00 -1.02670729e+00 -8.10106754e-01 1.05071473e+00
1.90588474e-01 1.13427117e-01 1.20834363e+00 4.04842764e-01
-1.59064323e-01 5.02574801e-01 1.04483455e-01 2.84136146e-01
7.28001118e-01 -1.18200839e+00 1.70469388e-01 1.02754474e+00
-1.88763022e-01 4.29949999e-01 6.10390246e-01 -6.63887918e-01
-1.01629937e+00 -9.80177701e-01 9.18687105e-01 -8.82160962e-01
2.32277848e-02 -1.85773715e-01 -6.72147095e-01 6.97435796e-01
1.77390859e-01 -1.45805418e-01 8.24234366e-01 -3.72911572e-01
6.07029833e-02 -1.70991912e-01 -1.77973509e+00 3.73838097e-01
1.37328422e+00 -3.06337744e-01 -9.61538494e-01 1.48002908e-01
-5.10686375e-02 -9.16458741e-02 -6.92350328e-01 1.78464368e-01
1.00268102e+00 -1.54472899e+00 1.09351254e+00 -1.07374415e-01
-2.93612897e-01 -4.10195589e-01 -3.04388672e-01 -7.57245481e-01
-3.76404196e-01 -3.51023048e-01 -1.74267739e-01 9.94851887e-01
-1.11617699e-01 -8.38927865e-01 9.33893025e-01 1.03863049e+00
3.81515831e-01 -2.11766243e-01 -1.30173969e+00 -8.43244076e-01
-6.92871511e-01 -6.03265047e-01 2.88650841e-01 4.12101358e-01
3.30381364e-01 1.09156063e-02 -5.82292199e-01 2.05785066e-01
9.04128492e-01 -5.36088526e-01 5.63799977e-01 -1.27895379e+00
-2.58742124e-02 2.10695416e-01 -1.20679438e+00 -4.90126282e-01
-6.23667479e-01 -1.22769371e-01 1.23284832e-02 -1.88093722e+00
3.24426442e-01 4.88313407e-01 -2.13150114e-01 6.56807780e-01
-1.18213696e-02 4.88197356e-01 -8.60253349e-02 6.07785322e-02
-6.27022564e-01 1.78645700e-01 5.81244171e-01 -1.23215146e-01
-1.65793061e-01 2.03036666e-01 -4.74005938e-01 7.53032327e-01
8.75772953e-01 -3.58766735e-01 -2.46335685e-01 -7.32757896e-02
-3.06610078e-01 -5.30398548e-01 7.00794756e-01 -1.69545746e+00
2.15638161e-01 1.29276529e-01 6.55088067e-01 -4.26409364e-01
5.46968520e-01 -1.43849635e+00 5.12625992e-01 8.89797986e-01
1.23012871e-01 -8.29429403e-02 -1.18806101e-01 7.15779364e-01
1.19243048e-01 6.34126067e-02 8.17774713e-01 -4.19297576e-01
-7.72576809e-01 -1.95941240e-01 -4.73137826e-01 -1.66281536e-01
1.34451723e+00 -1.10158932e+00 -1.84995383e-01 -4.77581888e-01
-7.60823071e-01 5.08369543e-02 6.91551328e-01 3.32837492e-01
7.29193449e-01 -1.41032803e+00 -4.69796270e-01 3.82440656e-01
4.90166485e-01 -3.84878695e-01 -1.24082536e-01 8.97407413e-01
-6.02224290e-01 4.05881822e-01 -5.00421405e-01 -6.03955388e-01
-1.90124977e+00 2.88446397e-01 3.71317029e-01 -4.17569699e-03
-1.10907519e+00 3.36936265e-01 -1.63844317e-01 1.83622792e-01
5.47283053e-01 -5.31016231e-01 -4.05636758e-01 -3.60303000e-02
1.13737464e+00 9.52144980e-01 1.57084674e-01 -1.02462554e+00
-9.03114080e-01 7.32557893e-01 4.76873666e-01 -2.83018887e-01
1.33286464e+00 -3.52628708e-01 2.43193954e-01 3.39004904e-01
1.06834102e+00 -3.34736913e-01 -1.26191974e+00 1.16759218e-01
5.46364486e-02 -6.45883203e-01 -2.91978300e-01 -7.91872025e-01
-7.55588114e-01 7.70290911e-01 1.48705983e+00 2.20189855e-01
1.47885966e+00 -1.88502848e-01 7.59634614e-01 3.00228983e-01
7.86731184e-01 -1.23684537e+00 -1.87502801e-01 -1.02507412e-01
6.61183953e-01 -1.44919813e+00 2.80591369e-01 -2.19610184e-01
-7.48790979e-01 8.51514757e-01 1.35044202e-01 8.12533200e-02
5.90057254e-01 3.16949487e-01 2.42179662e-01 -9.09248739e-02
-6.15791939e-02 -4.50962067e-01 -1.44868596e-02 1.36299264e+00
1.27863273e-01 2.05567956e-01 -3.53775084e-01 3.73395056e-01
-4.70237397e-02 3.56603175e-01 1.87817931e-01 1.64993882e+00
-6.72245681e-01 -7.22724140e-01 -9.28623021e-01 4.49202418e-01
-6.98415935e-01 4.21820015e-01 -4.88335729e-01 6.49190724e-01
3.33115339e-01 1.42355692e+00 -2.46978030e-01 -3.06056052e-01
9.79959667e-01 3.21874112e-01 4.62473601e-01 -2.78118014e-01
-5.22301197e-01 -2.06003696e-01 2.45365992e-01 -9.34777617e-01
-1.02334952e+00 -1.09380460e+00 -1.00422776e+00 -7.66594857e-02
2.57625371e-01 -2.18219668e-01 3.12876642e-01 9.47947800e-01
1.07629173e-01 2.40796417e-01 2.11805046e-01 -6.80804968e-01
-2.35908866e-01 -7.93473721e-01 -8.19436669e-01 8.69945586e-01
5.03118634e-01 -5.78446209e-01 -3.80667627e-01 6.14902377e-01] | [7.264708042144775, 0.5409824252128601] |
56706ea5-eb26-4bf1-a595-46ef6e6d8f1e | brain-decoding-from-functional-mri-using-long | 1809.05561 | null | http://arxiv.org/abs/1809.05561v1 | http://arxiv.org/pdf/1809.05561v1.pdf | Brain decoding from functional MRI using long short-term memory recurrent neural networks | Decoding brain functional states underlying different cognitive processes
using multivariate pattern recognition techniques has attracted increasing
interests in brain imaging studies. Promising performance has been achieved
using brain functional connectivity or brain activation signatures for a
variety of brain decoding tasks. However, most of existing studies have built
decoding models upon features extracted from imaging data at individual time
points or temporal windows with a fixed interval, which might not be optimal
across different cognitive processes due to varying temporal durations and
dependency of different cognitive processes. In this study, we develop a deep
learning based framework for brain decoding by leveraging recent advances in
sequence modeling using long short-term memory (LSTM) recurrent neural networks
(RNNs). Particularly, functional profiles extracted from task functional
imaging data based on their corresponding subject-specific intrinsic functional
networks are used as features to build brain decoding models, and LSTM RNNs are
adopted to learn decoding mappings between functional profiles and brain
states. We evaluate the proposed method using task fMRI data from the HCP
dataset, and experimental results have demonstrated that the proposed method
could effectively distinguish brain states under different task events and
obtain higher accuracy than conventional decoding models. | ['Yong Fan', 'Hongming Li'] | 2018-09-14 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [ 7.04621732e-01 -6.37366951e-01 -5.16891852e-02 -5.99617064e-01
-3.07800710e-01 -2.81143844e-01 6.94334149e-01 -1.75409645e-01
-6.39696240e-01 5.68152785e-01 3.71932834e-01 -1.32396206e-01
-4.77831423e-01 -3.63883346e-01 -3.47749382e-01 -5.91271460e-01
-4.09840524e-01 1.83186501e-01 -6.75675422e-02 1.33520648e-01
5.55028915e-01 2.77080238e-01 -1.32735288e+00 5.49751222e-01
1.02405846e+00 1.00090575e+00 6.23348653e-01 2.66044617e-01
-6.44969568e-02 5.45838237e-01 -4.16326791e-01 8.96921158e-02
-5.34577444e-02 -6.15902781e-01 -5.48080802e-01 -8.16664249e-02
-1.24141917e-01 -3.67998071e-02 -6.54638946e-01 1.13904703e+00
6.21575236e-01 3.98481190e-01 7.38237917e-01 -5.25609136e-01
-7.21327364e-01 5.00781059e-01 -3.48283291e-01 9.20910954e-01
1.15205660e-01 1.74857333e-01 7.16133296e-01 -7.86660612e-01
1.15753330e-01 9.54705894e-01 2.81267434e-01 5.82726002e-01
-1.28163314e+00 -9.07036722e-01 2.24230960e-01 6.69995964e-01
-1.10690284e+00 -6.46702588e-01 4.98500556e-01 -7.99079478e-01
1.30144536e+00 -8.91032964e-02 9.66274798e-01 1.47210467e+00
1.06302845e+00 5.46845078e-01 1.44811857e+00 2.79885940e-02
6.55755820e-03 -4.99787331e-01 6.71000361e-01 5.95537007e-01
-1.28381029e-01 2.27420330e-01 -5.28055310e-01 1.01294890e-02
8.35423768e-01 2.06842929e-01 -4.72094089e-01 2.19373628e-01
-1.75923061e+00 5.74258506e-01 4.97064322e-01 6.33464158e-01
-8.41381907e-01 2.52178740e-02 6.71504021e-01 3.08890134e-01
6.26146615e-01 -9.70497802e-02 -3.58628422e-01 -3.46377157e-02
-1.18216443e+00 -3.51024687e-01 2.23347813e-01 1.97255343e-01
4.26768869e-01 3.29015225e-01 -5.86573303e-01 1.20351422e+00
4.06649172e-01 1.98706314e-01 1.29845047e+00 -4.94857311e-01
3.57953012e-01 2.90075362e-01 -2.94618875e-01 -1.20468247e+00
-6.93065524e-01 -4.74851876e-01 -1.02999401e+00 -3.03763777e-01
-7.18347356e-02 -8.94007832e-02 -9.49369848e-01 1.91724885e+00
-3.68932009e-01 4.36124116e-01 -1.94351986e-01 9.60305750e-01
5.35501838e-01 5.14029682e-01 3.59925121e-01 -5.28486550e-01
1.42203438e+00 -6.33630097e-01 -8.01357567e-01 -4.83608246e-01
5.45462728e-01 3.80959846e-02 5.57131708e-01 2.80169189e-01
-9.14575458e-01 -5.83615303e-01 -9.89038050e-01 2.35826746e-01
-1.37820482e-01 7.34204380e-03 6.24897957e-01 4.74779338e-01
-1.34169447e+00 4.43948954e-01 -1.19049704e+00 -3.03306639e-01
6.11809731e-01 5.28485596e-01 -4.06246692e-01 -8.29055719e-03
-1.50005054e+00 1.04466915e+00 6.25426590e-01 7.40584016e-01
-1.35441649e+00 -2.87911206e-01 -6.43344104e-01 1.72535226e-01
-1.23608597e-01 -7.42203116e-01 7.63365865e-01 -1.07717884e+00
-1.44298160e+00 7.90884972e-01 -6.63088143e-01 -3.35323870e-01
-1.14093721e-01 1.61250114e-01 -7.95495689e-01 -1.15734385e-02
5.45602338e-03 4.86724436e-01 8.84658873e-01 -4.53971833e-01
8.45384076e-02 -8.10962081e-01 -4.62295115e-01 3.60700071e-01
-3.86410803e-01 4.15894687e-01 -3.55526805e-03 -5.79959095e-01
3.93624276e-01 -6.90304875e-01 -5.32254316e-02 -5.82905114e-01
-1.66631088e-01 -1.83116779e-01 3.36447448e-01 -1.08864415e+00
1.14297569e+00 -1.99170816e+00 5.49741685e-01 1.94481999e-01
4.60167259e-01 4.69670892e-02 -3.01754057e-01 -1.11672424e-01
-4.53184873e-01 1.70850649e-01 -3.90565813e-01 1.83699131e-01
-1.97542265e-01 1.52383000e-01 -3.57987806e-02 7.95391619e-01
1.01390168e-01 1.08936656e+00 -6.52021706e-01 -1.10675260e-01
-6.26914129e-02 4.03656214e-01 -1.95013955e-01 3.93543206e-02
6.57616034e-02 8.69000793e-01 -3.95510197e-01 5.44943392e-01
3.31365377e-01 -2.53934443e-01 4.24565166e-01 -1.64751068e-01
-8.89897421e-02 3.80128354e-01 -2.54552305e-01 1.92402780e+00
-2.47731075e-01 7.85235345e-01 -1.89934835e-01 -1.77794659e+00
8.70282233e-01 6.22687757e-01 5.89009404e-01 -1.24296868e+00
4.41360176e-01 -1.35319810e-02 8.35307598e-01 -7.54192770e-01
-2.64405400e-01 -2.80270368e-01 1.08130656e-01 6.26283646e-01
3.10708672e-01 7.34339714e-01 -3.07956394e-02 -3.62433970e-01
1.36074936e+00 -9.28302482e-02 1.40465111e-01 -4.23745573e-01
6.89332783e-01 -6.12111509e-01 6.25093043e-01 4.99667674e-01
-5.44519782e-01 1.65346831e-01 4.96378422e-01 -1.94959626e-01
-5.80709994e-01 -9.02574837e-01 -2.29858086e-01 1.30346560e+00
-3.34736675e-01 -4.87315878e-02 -5.30200839e-01 -2.38553599e-01
-4.46424425e-01 5.74058473e-01 -8.53117824e-01 -5.64600170e-01
-7.34264672e-01 -1.22428346e+00 6.35725439e-01 4.64050084e-01
5.64375997e-01 -1.34820998e+00 -6.86117768e-01 4.11797851e-01
-1.53649896e-01 -9.49211657e-01 -3.64836365e-01 2.36566320e-01
-1.27696908e+00 -8.80776703e-01 -7.53579795e-01 -8.74426305e-01
6.02319181e-01 2.55769104e-01 6.34652317e-01 -1.80330947e-01
-2.22665787e-01 1.00735180e-01 -5.81628643e-02 1.73392743e-01
9.85413268e-02 8.08344856e-02 2.30979562e-01 3.31492722e-01
3.29159111e-01 -7.35329032e-01 -7.23078191e-01 3.46769542e-01
-8.71441722e-01 1.54908672e-01 7.79802978e-01 9.21254992e-01
4.42335427e-01 -1.04235001e-01 8.31930280e-01 -4.78500575e-01
1.06029665e+00 -9.35005963e-01 -2.50561059e-01 4.20358658e-01
-4.75272804e-01 1.27654985e-01 4.93541032e-01 -4.68715131e-01
-1.35233128e+00 -5.80697834e-01 -3.49840559e-02 -2.47850537e-01
-1.25298068e-01 9.85597193e-01 -3.92322131e-02 2.51692291e-02
3.21288437e-01 8.31096947e-01 -1.48080122e-02 -2.40747795e-01
-1.58015504e-01 4.65595663e-01 2.86060125e-01 -6.09475970e-01
-1.85139105e-01 1.42477304e-01 -2.95263380e-01 -5.71231484e-01
-6.49371684e-01 -6.19850866e-02 -9.28041101e-01 -4.49508488e-01
1.15250099e+00 -8.45137775e-01 -8.28937113e-01 6.18492126e-01
-9.77352798e-01 -3.22607845e-01 6.03566825e-01 8.20395112e-01
-4.01240557e-01 2.93133259e-01 -9.94100869e-01 -5.97313285e-01
-4.42704260e-01 -1.28969157e+00 6.01828158e-01 5.91321774e-02
-1.80827647e-01 -1.21432233e+00 6.93293065e-02 8.01291466e-02
6.26062810e-01 2.33205780e-01 1.31096971e+00 -5.68212390e-01
-1.27861455e-01 -2.31252946e-02 -2.24880084e-01 1.74760669e-01
1.99756175e-01 -5.48658252e-01 -6.36790633e-01 -1.95358962e-01
5.80211937e-01 -1.01343289e-01 1.16443789e+00 5.56789577e-01
1.56521499e+00 -8.64173919e-02 -4.36027676e-01 5.63977599e-01
1.05313265e+00 5.10876179e-01 8.01943660e-01 6.61637262e-02
6.43401980e-01 8.16018522e-01 -7.57190958e-02 1.33027360e-01
1.86496571e-01 3.08146566e-01 9.97619405e-02 4.98834133e-01
2.62559503e-01 2.66095638e-01 8.91359985e-01 1.18183005e+00
-2.44913816e-01 6.15916736e-02 -1.15945280e+00 4.20540661e-01
-1.86350822e+00 -1.24427068e+00 -1.63008466e-01 2.07039285e+00
6.82551384e-01 2.02028360e-02 -3.15467298e-01 -3.67928445e-01
8.61106873e-01 3.94552648e-01 -9.85633373e-01 -2.95462132e-01
-6.71110377e-02 3.36642534e-01 4.84474786e-02 -1.62523717e-01
-6.39935315e-01 7.40969777e-01 6.76982021e+00 4.70480710e-01
-1.31848371e+00 7.14012027e-01 6.30291581e-01 -4.75479603e-01
-3.86850052e-02 -3.51234466e-01 -2.63592452e-01 5.78159869e-01
1.60337269e+00 -2.46026158e-01 8.04696620e-01 6.42759874e-02
6.46583974e-01 9.12764072e-02 -9.62238610e-01 1.13770366e+00
1.97375640e-01 -9.38918114e-01 -1.16188727e-01 6.01636693e-02
2.84382075e-01 4.46617544e-01 3.06510270e-01 4.22864586e-01
-1.28090501e-01 -1.38879049e+00 3.71787637e-01 1.23653901e+00
8.41088593e-01 -5.39951980e-01 3.81624222e-01 6.31515682e-01
-1.01428497e+00 -2.23254591e-01 -4.60512966e-01 -1.83381081e-01
2.01794386e-01 5.54710746e-01 -3.80227506e-01 1.88264474e-01
3.94184858e-01 1.27194202e+00 -5.30416429e-01 8.37259769e-01
2.48314179e-02 7.61423409e-01 2.09207565e-01 7.70045519e-02
4.37277295e-02 -2.65400201e-01 3.51165920e-01 1.10679579e+00
5.36899686e-01 1.43817693e-01 -4.97593209e-02 1.08959615e+00
-9.16747674e-02 8.81221145e-02 -3.88349205e-01 -4.90286499e-01
1.99989099e-02 1.41678262e+00 -8.67823541e-01 -2.83360422e-01
-3.51973593e-01 9.46795166e-01 5.37770391e-01 4.80873317e-01
-8.93999040e-01 5.30003645e-02 5.85610271e-01 -3.13398361e-01
-1.60588637e-01 -6.30125940e-01 -3.17679554e-01 -1.56022668e+00
-1.13778293e-01 -6.17552698e-01 2.67419487e-01 -7.18438208e-01
-1.39692581e+00 7.65782237e-01 6.78842291e-02 -7.00599730e-01
-2.00179502e-01 -5.45659184e-01 -6.74942851e-01 1.37810814e+00
-1.25921035e+00 -5.54167092e-01 -1.42834872e-01 9.14912701e-01
7.04263985e-01 -2.32144281e-01 9.08347964e-01 3.39805335e-01
-1.10561359e+00 7.32084140e-02 7.27582499e-02 3.46531831e-02
5.18958867e-01 -5.73211551e-01 -8.67775525e-04 7.86943793e-01
-7.98081011e-02 1.06713021e+00 9.25300568e-02 -8.70941699e-01
-1.26088202e+00 -9.07323897e-01 5.41758955e-01 -1.48658767e-01
6.85379922e-01 -4.76201802e-01 -1.26482022e+00 7.76792645e-01
3.06121498e-01 -1.47024482e-01 9.60165918e-01 -2.17385199e-02
-1.80334281e-02 1.77436143e-01 -9.88991857e-01 4.29760218e-01
1.33936858e+00 -7.74560273e-01 -8.46844018e-01 3.57848704e-01
2.12159067e-01 1.66863114e-01 -1.00809693e+00 4.66259778e-01
7.23784983e-01 -7.58916199e-01 6.99438632e-01 -6.66796565e-01
1.57114387e-01 -2.19268110e-02 -2.63654105e-02 -1.51568496e+00
-7.88159966e-01 7.22858161e-02 -1.97764114e-01 7.71020472e-01
1.82931766e-01 -9.72587883e-01 3.06779325e-01 7.21230865e-01
-3.13120574e-01 -5.72267056e-01 -8.90900731e-01 -3.95383835e-01
-1.05171902e-02 -4.83860075e-01 2.33943239e-01 9.86490190e-01
9.85114574e-02 3.57702464e-01 -3.50945681e-01 6.13978086e-03
3.11753571e-01 -1.43378213e-01 -2.46538132e-01 -1.22505486e+00
-1.65978879e-01 -6.98466301e-01 -2.78704226e-01 -7.50906408e-01
9.12285984e-01 -1.48753798e+00 4.46366146e-02 -1.77290344e+00
5.89089453e-01 7.87908360e-02 -1.01203847e+00 6.72117114e-01
-1.06801018e-01 -2.36241557e-02 -2.44535014e-01 4.82436240e-01
-3.78473014e-01 6.95622563e-01 1.01292825e+00 -1.24710701e-01
-2.16686595e-02 -3.31824243e-01 -6.18631363e-01 4.35409099e-01
9.53450143e-01 -3.82993549e-01 -4.61973518e-01 -6.31300926e-01
-2.23726369e-02 4.66443211e-01 3.70404273e-01 -1.10976768e+00
1.59454554e-01 4.19932529e-02 1.00759256e+00 -1.61213905e-01
2.31374368e-01 -4.38140243e-01 2.30613843e-01 6.06731236e-01
-4.87959713e-01 2.94383466e-01 1.30512297e-01 6.57895446e-01
-1.71142131e-01 -3.06364633e-02 5.14693618e-01 -3.14440995e-01
-8.95904720e-01 5.75340450e-01 -8.64781499e-01 -3.58636320e-01
7.73956478e-01 -3.48068655e-01 -2.37939194e-01 9.07045975e-02
-1.06432319e+00 2.42905226e-02 -2.88149089e-01 5.74650347e-01
1.02549314e+00 -1.38145339e+00 -6.57782078e-01 5.22174478e-01
-2.78487027e-01 -8.15576553e-01 7.54189610e-01 1.46108603e+00
2.56428309e-02 6.30437255e-01 -8.66221011e-01 -6.51629329e-01
-8.62574160e-01 3.19820821e-01 5.42541146e-01 -8.56400505e-02
-6.28219724e-01 6.37826800e-01 4.47658449e-01 -1.38367906e-01
-3.39031160e-01 -1.31317988e-01 -6.94160461e-01 4.61901933e-01
5.85191131e-01 7.79138580e-02 2.22060412e-01 -9.15427029e-01
-3.83046091e-01 6.24816157e-02 -2.57275760e-01 -2.10297212e-01
1.48613000e+00 -1.69691116e-01 -4.81453955e-01 7.84000397e-01
1.23106015e+00 -6.95424020e-01 -1.09489369e+00 -3.08581978e-01
2.18761727e-01 -2.74179846e-01 3.82625878e-01 -8.37598383e-01
-1.33503401e+00 1.10275304e+00 8.67443383e-01 -2.01630935e-01
1.24346519e+00 -3.22635710e-01 8.74541938e-01 1.92180499e-01
5.94613612e-01 -9.29233670e-01 -1.11047681e-02 7.08174825e-01
7.95808792e-01 -9.87595499e-01 -4.44460124e-01 4.09888059e-01
-5.53052723e-01 1.27658808e+00 5.63925922e-01 -3.61444741e-01
7.39762008e-01 -2.60582209e-01 -5.02469897e-01 -4.95297492e-01
-9.01169896e-01 -1.70801729e-02 5.82164466e-01 1.97500914e-01
7.82365680e-01 1.79539189e-01 -5.62993526e-01 8.83960307e-01
1.19935408e-01 -3.84482928e-02 1.10846438e-01 7.16075480e-01
-5.08808911e-01 -9.34116840e-01 -2.48536035e-01 1.34554172e+00
-6.33020937e-01 -2.34907225e-01 8.03898945e-02 2.72910446e-01
-1.50846705e-01 7.07372665e-01 7.43480250e-02 -4.34948385e-01
2.64298767e-02 7.64414787e-01 6.78087533e-01 -7.93099403e-01
-5.60367823e-01 1.04204692e-01 -8.70142058e-02 -6.13784790e-01
-6.41334653e-01 -1.01479018e+00 -1.36546922e+00 2.04469915e-02
-8.02679434e-02 -2.37533197e-01 5.07129967e-01 1.40933740e+00
4.80142862e-01 9.53213274e-01 3.34931523e-01 -9.17109609e-01
-8.15655366e-02 -1.27198708e+00 -6.35163605e-01 6.02191389e-02
1.21572159e-01 -1.09871042e+00 1.20797053e-01 2.72480734e-02] | [12.607540130615234, 3.389878034591675] |
3fd0ab4e-4632-42c2-9296-b5311c31c25d | learning-the-dimensionality-of-word | 1511.05392 | null | http://arxiv.org/abs/1511.05392v3 | http://arxiv.org/pdf/1511.05392v3.pdf | Learning the Dimensionality of Word Embeddings | We describe a method for learning word embeddings with data-dependent
dimensionality. Our Stochastic Dimensionality Skip-Gram (SD-SG) and Stochastic
Dimensionality Continuous Bag-of-Words (SD-CBOW) are nonparametric analogs of
Mikolov et al.'s (2013) well-known 'word2vec' models. Vector dimensionality is
made dynamic by employing techniques used by Cote & Larochelle (2016) to define
an RBM with an infinite number of hidden units. We show qualitatively and
quantitatively that SD-SG and SD-CBOW are competitive with their
fixed-dimension counterparts while providing a distribution over embedding
dimensionalities, which offers a window into how semantics distribute across
dimensions. | ['Sachin Ravi', 'Eric Nalisnick'] | 2015-11-17 | null | null | null | null | ['learning-word-embeddings'] | ['methodology'] | [-6.38069332e-01 -5.47728762e-02 -5.69101334e-01 -2.27937549e-01
-3.80505055e-01 -7.79500008e-01 1.12951946e+00 1.01427577e-01
-5.70378602e-01 3.68524790e-01 1.01037896e+00 -7.99175560e-01
-1.75488248e-01 -7.77082562e-01 -1.28416657e-01 -7.27618694e-01
-4.07665670e-01 4.40448672e-01 4.11042683e-02 -3.48298371e-01
2.91741699e-01 4.65317488e-01 -1.38737309e+00 -1.14324108e-01
1.70715302e-02 5.79786181e-01 -9.43673495e-03 9.55247402e-01
-4.13682550e-01 1.24296054e-01 -1.41794220e-01 -5.53763986e-01
3.05602163e-01 9.14681107e-02 -5.91262519e-01 -1.97448939e-01
3.52350138e-02 -3.20198119e-01 -6.41590953e-01 7.33576238e-01
4.37993526e-01 4.44261044e-01 1.26874828e+00 -1.11068094e+00
-1.59142733e+00 5.70805967e-01 -3.93349677e-01 4.48440284e-01
-1.48688808e-01 2.25645870e-01 1.58494437e+00 -1.25159705e+00
5.56248307e-01 1.65878260e+00 6.56048417e-01 7.82416284e-01
-1.50061119e+00 -2.54216522e-01 2.98388600e-01 -1.31723419e-01
-1.22647154e+00 -2.56290939e-02 4.35614735e-01 -8.29708755e-01
1.30735385e+00 2.38440216e-01 6.29832625e-01 1.40070689e+00
3.04482490e-01 5.52530944e-01 9.21404898e-01 -6.47619545e-01
5.35801768e-01 1.11252591e-01 6.39400721e-01 2.91198909e-01
7.43446648e-01 2.43338376e-01 -2.85096616e-01 -5.73434532e-01
8.29160213e-01 3.30900043e-01 1.56162336e-01 -9.17645514e-01
-1.01585793e+00 1.55501020e+00 4.04617861e-02 4.54058379e-01
-2.19293684e-01 4.50632244e-01 6.99433327e-01 3.57146889e-01
6.62513018e-01 4.89657462e-01 -5.91653705e-01 -4.51316178e-01
-3.57235938e-01 4.06919479e-01 6.87862456e-01 1.01757455e+00
5.57763040e-01 3.39453071e-01 -2.41641048e-02 9.16339219e-01
5.31111658e-01 3.63070607e-01 1.15721774e+00 -7.92107821e-01
1.64524034e-01 5.82829602e-02 2.04187900e-01 -7.65095949e-01
-3.23933482e-01 -2.65421662e-02 -6.49267435e-01 -1.41276261e-02
2.12887079e-01 -6.49645999e-02 -8.64096582e-01 1.75623262e+00
-3.24778892e-02 -1.66520998e-01 1.26291081e-01 6.68655097e-01
3.50943744e-01 6.49593592e-01 2.65297592e-01 2.98115015e-02
1.38495934e+00 -7.34537959e-01 -7.43785024e-01 6.63123131e-02
9.80583489e-01 -2.92013466e-01 1.43365002e+00 2.51782894e-01
-7.89139628e-01 -4.24796611e-01 -8.09074759e-01 -3.55311483e-01
-9.74958599e-01 -3.72859120e-01 9.57096517e-01 1.04162097e+00
-1.30230260e+00 4.42502171e-01 -8.65827858e-01 -5.67374647e-01
2.17945024e-01 9.47035775e-02 -4.17106777e-01 2.77691722e-01
-1.25710404e+00 9.64527845e-01 1.07792355e-01 -5.56900680e-01
-8.05770576e-01 -6.73254728e-01 -1.03493023e+00 -9.87999327e-03
-3.08205962e-01 -5.45251429e-01 1.12186134e+00 -1.60677344e-01
-1.34422266e+00 6.45318329e-01 -1.48221537e-01 -4.44624841e-01
2.27485560e-02 -1.67026252e-01 -5.28968990e-01 -2.07462743e-01
-3.72877046e-02 5.11060238e-01 7.62357712e-01 -1.14197314e+00
-1.70555919e-01 -4.19628143e-01 3.87529582e-02 1.03534579e-01
-1.05609441e+00 -1.51104048e-01 -1.10279731e-02 -8.93230438e-01
-1.99458048e-01 -6.74562335e-01 -2.78499633e-01 -4.37285975e-02
-8.19399655e-02 -5.24729311e-01 5.05588651e-01 -4.21939343e-01
1.75944591e+00 -2.18019581e+00 3.33969861e-01 1.21853173e-01
6.58583820e-01 1.06251188e-01 -4.68575448e-01 1.06059551e+00
-2.80533850e-01 6.63190544e-01 -1.33096855e-02 -5.53310335e-01
5.49223125e-01 6.01940632e-01 -3.93692464e-01 6.51581228e-01
4.34901416e-02 9.11081970e-01 -8.27921450e-01 -5.91121987e-02
2.36564204e-01 6.26865983e-01 -7.68189430e-01 2.94721387e-02
5.63378632e-02 -3.92483503e-01 -2.53269762e-01 1.97578162e-01
4.37633276e-01 -1.24468513e-01 8.78518447e-02 2.15453908e-01
-2.34818935e-01 3.68865579e-01 -1.15268874e+00 1.26651049e+00
-6.29239678e-01 6.61787450e-01 -2.32141748e-01 -7.47450829e-01
7.66909242e-01 7.10594580e-02 2.20405146e-01 -1.98077321e-01
1.07817024e-01 1.44450828e-01 -1.37468562e-01 -3.56782794e-01
8.83900106e-01 -5.86529374e-01 -2.82061249e-01 7.79419601e-01
3.58109117e-01 1.48596510e-01 1.60064891e-01 3.72186810e-01
1.01868606e+00 -2.74808288e-01 3.41893882e-01 -6.95792317e-01
9.71921384e-02 -4.29637134e-01 9.85055193e-02 6.76159382e-01
-2.46697277e-01 5.38343489e-01 7.67735720e-01 -3.54826957e-01
-1.55743694e+00 -1.40383935e+00 -4.29356724e-01 1.48934913e+00
-2.95492679e-01 -7.46707618e-01 -3.11257899e-01 -4.15285170e-01
4.92677182e-01 1.01305747e+00 -1.19369209e+00 -1.98137596e-01
-3.08274589e-02 -5.43173790e-01 5.15936136e-01 7.70352781e-01
-6.27319992e-01 -5.71481168e-01 -3.39013278e-01 4.55385186e-02
6.72425508e-01 -4.12927181e-01 -7.19063640e-01 5.13836741e-01
-8.35485816e-01 -6.28061771e-01 -9.31721985e-01 -5.58638811e-01
4.61414158e-01 3.12348068e-01 1.02456343e+00 -5.19995153e-01
-3.56647253e-01 5.64468503e-01 -5.83284080e-01 -2.68418074e-01
-2.39943534e-01 9.87610072e-02 7.28035152e-01 -2.84120113e-01
8.70470107e-01 -6.78185523e-01 -6.70670033e-01 -2.20266998e-01
-1.33894181e+00 -5.76172411e-01 1.48280427e-01 1.00399721e+00
2.91230053e-01 -2.68787980e-01 3.60721707e-01 -7.26161122e-01
1.18326104e+00 -7.69133866e-01 -4.18804109e-01 -2.45296881e-01
-9.95236874e-01 4.63349134e-01 4.73225325e-01 -7.66062081e-01
-2.55267918e-01 -7.14823902e-01 -1.28446996e-01 -5.28602898e-01
1.55698672e-01 2.25102261e-01 1.02604829e-01 4.15432036e-01
5.39345503e-01 1.35363594e-01 -2.29587946e-02 -6.50987208e-01
1.17761314e+00 1.01743007e+00 -1.54849395e-01 -5.56819320e-01
7.90587664e-01 4.60543483e-01 -3.08311015e-01 -8.57891202e-01
-3.63299072e-01 -7.21125901e-01 -8.12094569e-01 4.26222086e-01
1.11349034e+00 -8.47636998e-01 -4.30974185e-01 -2.23011784e-02
-1.01506102e+00 -1.94546446e-01 -7.83069491e-01 6.64114118e-01
-5.36590874e-01 2.70641744e-01 -6.37655497e-01 -1.09737277e+00
1.53800650e-02 -8.00318182e-01 1.03432775e+00 -3.09772998e-01
-6.40354156e-01 -1.71827388e+00 6.56307399e-01 -3.79173398e-01
6.26110017e-01 -2.35173330e-01 1.25189769e+00 -1.06376135e+00
3.96730572e-01 -1.62620693e-01 -2.24884972e-01 6.25772476e-01
2.07840234e-01 -1.11589290e-01 -8.20552826e-01 -4.33063447e-01
-2.15315670e-01 -2.12831602e-01 8.31979930e-01 4.47958231e-01
1.03280938e+00 -5.33404768e-01 -1.24243811e-01 6.12505078e-01
1.64024055e+00 -1.08536907e-01 3.06173563e-01 4.40606296e-01
7.94965506e-01 5.05650163e-01 6.93182275e-02 7.11056352e-01
5.06360292e-01 2.55395144e-01 1.43697962e-01 1.91358313e-01
2.59578556e-01 -4.90950078e-01 3.90131295e-01 1.19002676e+00
1.10915698e-01 -2.65202135e-01 -8.49942803e-01 9.69878912e-01
-1.53712630e+00 -9.38580036e-01 -5.78776747e-02 2.18304443e+00
6.44917488e-01 3.96650955e-02 4.99844611e-01 -2.12654807e-02
4.28383142e-01 6.63421094e-01 -3.33730996e-01 -1.24374282e+00
-1.94164693e-01 3.56222153e-01 9.55523312e-01 7.97306478e-01
-7.17431545e-01 9.65904534e-01 7.04673672e+00 6.45013034e-01
-4.31504846e-01 2.84077972e-01 2.20799983e-01 -2.87304640e-01
-1.09777451e+00 -9.30035114e-02 -8.45994651e-01 4.83769953e-01
1.39602041e+00 -3.49950194e-01 2.99512506e-01 8.77080441e-01
1.04341105e-01 3.69847953e-01 -1.10373974e+00 9.77724016e-01
1.87758446e-01 -1.14936197e+00 3.24330419e-01 5.29131293e-01
6.96388602e-01 1.24560758e-01 4.26256001e-01 5.04582703e-01
9.33239341e-01 -1.14599323e+00 2.01155782e-01 2.75616795e-01
9.72357571e-01 -8.97041142e-01 4.37815696e-01 4.94572856e-02
-1.00831091e+00 -3.86348486e-01 -7.94814289e-01 -1.57975435e-01
3.56629752e-02 4.96945590e-01 -2.39979357e-01 6.44493401e-02
4.40742046e-01 5.19231081e-01 -4.74034131e-01 1.67766199e-01
2.82420278e-01 6.91401780e-01 -1.55798405e-01 -3.57683599e-01
6.41603529e-01 -3.73845160e-01 3.10767114e-01 1.45473289e+00
3.22820246e-01 -1.26875818e-01 -2.97255814e-01 6.84768379e-01
9.37412307e-02 7.76406676e-02 -9.46780562e-01 -6.93577647e-01
6.41646028e-01 8.98292840e-01 -4.01292801e-01 -2.46960342e-01
-5.01375914e-01 8.27939391e-01 2.03974411e-01 4.72557396e-01
-3.05139899e-01 -5.54157674e-01 1.54697061e+00 1.72645897e-01
6.90621257e-01 -7.95676827e-01 -3.05129319e-01 -1.21178114e+00
-4.03655797e-01 -4.51612264e-01 2.57814974e-01 -3.94966006e-01
-1.82866657e+00 3.61592561e-01 1.33617178e-01 -1.10776031e+00
-3.24391842e-01 -1.16228795e+00 -3.27909321e-01 1.02488697e+00
-1.07457864e+00 -1.01497936e+00 4.85258937e-01 3.34690213e-01
5.22063136e-01 -4.10395414e-01 1.27513087e+00 -7.22670481e-02
-2.92939365e-01 5.99418581e-01 1.00010407e+00 9.34764519e-02
4.38605398e-01 -1.58427048e+00 1.01460874e+00 2.01288864e-01
2.45179579e-01 1.01062489e+00 8.81209552e-01 -4.20014352e-01
-1.70902097e+00 -8.97047877e-01 9.36639965e-01 -8.11028659e-01
1.38649428e+00 -1.01469600e+00 -7.40548074e-01 8.66175652e-01
2.26536885e-01 2.01629885e-02 1.15829802e+00 3.68220419e-01
-8.91080260e-01 3.56388003e-01 -9.06318367e-01 8.85434091e-01
9.17742550e-01 -8.54660988e-01 -8.68978500e-01 3.21236312e-01
1.27186549e+00 4.61831808e-01 -1.03879654e+00 -1.40083656e-01
6.99999630e-01 -8.40798914e-01 9.35695589e-01 -1.23045290e+00
3.41583997e-01 4.10933703e-01 -7.36824274e-01 -1.44413722e+00
-6.32083118e-01 -5.01157403e-01 -2.90751159e-01 1.13271189e+00
1.88077912e-01 -9.77824330e-01 3.16191375e-01 3.47845793e-01
2.45895550e-01 -9.68786120e-01 -8.16646695e-01 -1.13408601e+00
1.09302354e+00 -6.89444244e-01 6.53396428e-01 1.08273673e+00
1.88111857e-01 1.42267704e-01 -1.30419821e-01 -2.39833564e-01
3.53755206e-01 -4.67964888e-01 4.46103781e-01 -1.16550982e+00
-4.12527382e-01 -7.41124034e-01 -7.84151435e-01 -1.10703945e+00
2.43435264e-01 -1.04540491e+00 -5.59592903e-01 -1.32493114e+00
5.33453710e-02 -4.99291793e-02 -6.43605530e-01 4.90115471e-02
1.41695678e-01 7.36429915e-02 4.17365022e-02 1.11656241e-01
-2.35734627e-01 7.78872907e-01 7.18213320e-01 3.19893360e-01
-1.41572475e-01 -5.27374566e-01 -9.93636966e-01 4.07794654e-01
6.97117329e-01 -2.42013916e-01 -6.71668410e-01 -3.88262182e-01
2.23675966e-01 -3.95316750e-01 1.31377708e-02 -2.49088094e-01
-1.72080770e-01 -2.95121044e-01 1.40177622e-01 -3.34057570e-01
4.35260147e-01 -4.42579776e-01 -4.11281526e-01 2.73056686e-01
-7.06704736e-01 6.26009405e-01 7.29360357e-02 7.90669024e-01
2.15148211e-01 -2.13411838e-01 4.98497456e-01 9.67030525e-02
-4.22021836e-01 3.12372029e-01 -6.99968100e-01 1.60950959e-01
8.33991945e-01 -2.65171647e-01 -1.30931571e-01 -4.10455108e-01
-7.33780980e-01 5.24357185e-02 4.14104760e-01 6.34731472e-01
6.73064888e-01 -1.83165467e+00 -4.65052068e-01 4.39364523e-01
4.62880224e-01 -4.53601390e-01 -3.44366059e-02 3.47862840e-01
-2.63887018e-01 6.48796916e-01 1.14178121e-01 -1.10574424e-01
-7.25551069e-01 7.88890004e-01 -9.23881754e-02 -1.12409011e-01
-6.17413580e-01 9.32760894e-01 1.72362223e-01 -6.28498256e-01
2.15228349e-01 -2.72213608e-01 -4.52953912e-02 5.19000530e-01
4.94624138e-01 4.65917230e-01 -4.84784216e-01 -5.99284470e-01
-1.95481271e-01 3.43995273e-01 -1.73232228e-01 -6.27759695e-01
1.55577528e+00 -3.49026859e-01 6.59688711e-02 1.38844979e+00
1.80310535e+00 -9.39496756e-02 -9.20717061e-01 -1.41631976e-01
9.41315070e-02 -4.69425648e-01 4.13983092e-02 -1.20357066e-01
-5.09560466e-01 1.28622472e+00 7.99664915e-01 6.50561512e-01
3.19682002e-01 1.42197460e-01 5.98707914e-01 7.28284940e-02
1.83861032e-01 -1.16814864e+00 -2.50516715e-03 7.23812342e-01
6.71674013e-01 -8.49418938e-01 -7.63157532e-02 3.86427760e-01
-7.85628378e-01 1.03023887e+00 2.17522636e-01 -4.22818273e-01
1.16090834e+00 2.21054241e-01 -9.84018147e-02 -9.93512571e-02
-1.01794767e+00 -3.12827319e-01 1.73563153e-01 8.19550931e-01
5.11634707e-01 2.93622971e-01 -5.39408922e-01 7.66028762e-01
-4.48181421e-01 -5.57835817e-01 5.16425788e-01 7.70199299e-01
-5.50784171e-01 -1.27455378e+00 -1.69931188e-01 8.09916615e-01
-1.11894310e-01 -4.12640482e-01 -2.02337548e-01 9.07958329e-01
-2.74390638e-01 6.50418222e-01 4.76741880e-01 -5.34026146e-01
7.47685805e-02 4.73256171e-01 7.55591094e-02 -1.05596924e+00
-1.85284153e-01 -3.25401910e-02 -1.44475549e-01 -3.07386816e-01
1.88460141e-01 -6.86989129e-01 -1.06841147e+00 -4.87498343e-01
-2.00802699e-01 1.12115346e-01 8.16492379e-01 7.04114377e-01
3.37215871e-01 3.19333255e-01 8.14136982e-01 -6.16867602e-01
-1.16648006e+00 -1.15380752e+00 -1.22985196e+00 4.84781474e-01
5.45400620e-01 -7.07619727e-01 -9.85411108e-01 -2.57207721e-01] | [10.461138725280762, 8.685569763183594] |
3638bdab-0802-4e61-aa23-41c68046dfb7 | word-level-persian-lipreading-dataset | 2304.04068 | null | https://arxiv.org/abs/2304.04068v1 | https://arxiv.org/pdf/2304.04068v1.pdf | Word-level Persian Lipreading Dataset | Lip-reading has made impressive progress in recent years, driven by advances in deep learning. Nonetheless, the prerequisite such advances is a suitable dataset. This paper provides a new in-the-wild dataset for Persian word-level lipreading containing 244,000 videos from approximately 1,800 speakers. We evaluated the state-of-the-art method in this field and used a novel approach for word-level lip-reading. In this method, we used the AV-HuBERT model for feature extraction and obtained significantly better performance on our dataset. | ['Nasser Mozayani', 'Hossein Zeinali', 'Samin Heydarian', 'Ali Lashini', 'Javad Peymanfard'] | 2023-04-08 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [-1.21879233e-02 1.12052642e-01 -6.85339510e-01 -4.41669710e-02
-1.43368840e+00 8.54782537e-02 5.53332448e-01 -2.32588485e-01
-5.52264094e-01 7.49058664e-01 4.75470692e-01 -2.25500949e-02
4.19528872e-01 -1.92797974e-01 -4.88549858e-01 -6.32311106e-01
9.93064865e-02 7.99202546e-02 2.87623703e-01 2.92089768e-02
4.15527165e-01 3.04700315e-01 -1.81138086e+00 2.16510028e-01
6.72350705e-01 1.20130312e+00 1.44314453e-01 8.19381058e-01
-2.61963606e-01 6.37774467e-01 -5.69126785e-01 -6.79625928e-01
-2.87766308e-02 -4.98545945e-01 -8.09726536e-01 -1.92090914e-01
8.02553713e-01 -3.14141542e-01 -5.04745603e-01 7.66104043e-01
1.22868884e+00 5.33581600e-02 7.11495996e-01 -1.43147552e+00
-6.12840056e-01 6.35657609e-01 -6.23469949e-01 2.09474921e-01
5.07145882e-01 2.87930280e-01 9.79409575e-01 -1.08110034e+00
5.06660461e-01 1.42065835e+00 7.06813693e-01 9.91018057e-01
-8.55129957e-01 -7.34552503e-01 -4.58508760e-01 8.04874241e-01
-1.44692075e+00 -1.26880693e+00 1.03797293e+00 -1.27550632e-01
9.15872753e-01 -1.86593875e-01 6.78847075e-01 1.34901929e+00
-4.84649725e-02 1.23175347e+00 1.18427598e+00 -8.28372657e-01
1.10365249e-01 -1.47766145e-02 -2.52381206e-01 5.70647299e-01
-8.38208571e-02 -2.63479333e-02 -1.17053330e+00 2.66022533e-01
1.60002500e-01 -7.03345358e-01 -3.29654932e-01 -2.03828722e-01
-1.04603684e+00 8.06483448e-01 -9.09510255e-02 3.00367296e-01
-3.02399963e-01 -2.47229729e-03 5.65208554e-01 -1.28193423e-01
4.83756870e-01 -9.80356708e-02 -4.30494398e-01 -6.47514164e-01
-1.35678613e+00 3.96993607e-02 7.36675143e-01 5.45088589e-01
2.47303739e-01 1.69065803e-01 -2.01823473e-01 1.23221779e+00
5.54051876e-01 8.10771585e-01 9.38218534e-01 -9.19252336e-01
5.23518264e-01 1.75435077e-02 -3.07896733e-01 -6.40079618e-01
-2.67603904e-01 2.05237433e-01 -7.50261962e-01 2.31305197e-01
6.30336940e-01 -9.59042087e-03 -9.17218208e-01 1.64412069e+00
1.16339557e-01 -8.20499193e-03 3.61966491e-01 4.63193834e-01
1.17316878e+00 5.42682230e-01 1.71038166e-01 -2.78746873e-01
1.34730399e+00 -1.09302521e+00 -9.67255712e-01 8.21628571e-02
4.45825495e-02 -9.83362675e-01 1.19100308e+00 5.57226002e-01
-1.41628206e+00 -5.94817460e-01 -9.07861471e-01 -2.25491345e-01
-4.38200295e-01 1.74632475e-01 2.47299165e-01 7.64290035e-01
-1.34856284e+00 3.15272927e-01 -2.40789816e-01 -7.45169580e-01
6.89127564e-01 2.41648406e-01 -6.09454036e-01 2.57562459e-01
-1.20915544e+00 8.78119230e-01 7.11289793e-02 -2.21383661e-01
-7.69718289e-01 -4.37805951e-01 -9.70011413e-01 -1.93983033e-01
3.09988111e-01 -2.39809126e-01 1.46295774e+00 -5.84990084e-01
-2.19934940e+00 1.31253648e+00 -5.31534135e-01 -6.37067795e-01
7.44674146e-01 -2.65396416e-01 -4.76009011e-01 4.71619636e-01
-2.64696032e-01 1.32827055e+00 1.42673028e+00 -1.06118751e+00
-4.92333204e-01 -2.13434175e-01 -3.65977198e-01 1.68464817e-02
-1.26498744e-01 2.17969045e-01 -3.53636205e-01 -5.46352386e-01
-1.57157034e-01 -6.69181526e-01 6.45703495e-01 2.39867285e-01
-3.58111143e-01 -8.17280591e-01 7.74190664e-01 -9.69570935e-01
8.45600426e-01 -2.25605178e+00 -1.93600431e-01 -5.79121828e-01
1.06551878e-01 6.86341703e-01 -2.07576588e-01 2.72717297e-01
5.16498908e-02 2.55536467e-01 -8.44483152e-02 -8.32821608e-01
1.92149103e-01 -1.85658097e-01 -6.58112988e-02 7.30767310e-01
2.40426473e-02 9.65650320e-01 -5.02425611e-01 -1.19731808e+00
4.36993718e-01 8.25458884e-01 -2.60750026e-01 4.20526415e-01
8.32568388e-03 -6.99028820e-02 2.32786924e-01 8.94920409e-01
8.85897458e-01 3.63931328e-01 -3.85311306e-01 -2.18328208e-01
-1.64185286e-01 1.08242884e-01 -7.16484129e-01 1.86815083e+00
-5.34863591e-01 1.39947677e+00 2.60785103e-01 -9.42809403e-01
7.77319849e-01 5.50295174e-01 5.33463836e-01 -7.69082010e-01
3.20959300e-01 2.21815273e-01 -1.61213160e-01 -9.58173752e-01
2.73026139e-01 -1.41345367e-01 4.59457606e-01 8.05881619e-02
3.01820606e-01 -2.72307664e-01 1.24787793e-01 -2.35967115e-01
2.69707084e-01 7.61502162e-02 4.40813720e-01 -3.03006709e-01
9.59898353e-01 -4.72648442e-01 2.08520874e-01 3.67063314e-01
-9.25479233e-01 7.98066020e-01 4.39508021e-01 -9.97769833e-02
-6.94379985e-01 -1.06050110e+00 -3.85078460e-01 1.10352874e+00
-3.18847060e-01 -2.29518995e-01 -1.24612963e+00 -5.50216794e-01
-2.37161204e-01 5.83781600e-01 -5.37117183e-01 2.88181245e-01
-4.82210100e-01 -2.04340473e-01 1.15473819e+00 4.28002208e-01
9.69381273e-01 -1.60682130e+00 -6.28467143e-01 -9.37050432e-02
-5.20657420e-01 -1.23231411e+00 -6.09795272e-01 -2.35158697e-01
-4.34802562e-01 -1.10452092e+00 -1.45597959e+00 -1.05143666e+00
-3.49426158e-02 -7.89101720e-02 1.04530716e+00 -2.19962999e-01
-3.04529518e-01 2.26960912e-01 -2.46548444e-01 -7.16123998e-01
-5.97368240e-01 1.02600440e-01 1.92874268e-01 4.83072642e-03
4.72456723e-01 -9.81829911e-02 -3.39693010e-01 1.12955496e-01
-3.82418126e-01 -2.96314746e-01 6.52886868e-01 6.87974095e-01
2.97540426e-01 -3.68302047e-01 7.91569710e-01 2.80603886e-01
6.91515028e-01 1.27966553e-01 -5.33989847e-01 1.40405148e-01
-4.97846067e-01 1.07652701e-01 3.36154252e-01 -5.67256808e-01
-9.24382329e-01 -1.84043467e-01 -8.56910110e-01 -2.81506330e-01
-4.64161754e-01 -1.21470474e-01 -4.49557990e-01 -7.11373389e-02
2.35028431e-01 4.34567481e-01 3.87853831e-01 -6.66778028e-01
3.21135610e-01 1.47790122e+00 7.77740955e-01 -2.29333758e-01
3.05419505e-01 2.07193524e-01 -3.76598397e-03 -1.59945035e+00
-3.72574031e-01 -4.47417766e-01 -7.00249910e-01 -3.52517486e-01
1.08418083e+00 -8.29799831e-01 -1.22546518e+00 1.43946397e+00
-1.11731637e+00 -5.04965782e-01 -3.09586763e-01 4.40065145e-01
-9.21509743e-01 5.19201815e-01 -3.51828992e-01 -1.08089507e+00
-5.52522242e-01 -1.26088357e+00 1.09639370e+00 3.17965478e-01
7.70872310e-02 -8.44900370e-01 3.65278274e-01 5.04625440e-01
6.50056958e-01 -3.14535886e-01 6.12814426e-01 -6.24122083e-01
-1.34701684e-01 4.98674810e-02 -3.68074536e-01 6.36901319e-01
5.65724075e-02 1.10166311e-01 -1.54464245e+00 -3.36610600e-02
-2.51679361e-01 -7.35386550e-01 1.09156990e+00 7.38570333e-01
1.19209731e+00 -1.78960025e-01 -5.03976531e-02 5.45851767e-01
1.03905642e+00 -7.32651576e-02 9.21134412e-01 8.99523199e-02
2.66481042e-01 6.28868699e-01 3.14838499e-01 1.63512513e-01
5.25496125e-01 9.41020191e-01 1.71510398e-01 -1.38331324e-01
-8.89168322e-01 -5.90864003e-01 4.49375868e-01 5.61468542e-01
-9.17594433e-02 -3.51249903e-01 -8.85590792e-01 7.50009358e-01
-1.39585555e+00 -1.10173762e+00 4.25863922e-01 2.00972080e+00
8.76473010e-01 7.68594667e-02 5.61447024e-01 5.55628359e-01
7.79597402e-01 5.29912949e-01 -4.21108812e-01 -5.02017498e-01
-2.14293405e-01 2.15573370e-01 1.27971932e-01 6.18871510e-01
-1.06882250e+00 1.19046724e+00 7.28480101e+00 1.07262075e+00
-1.45280600e+00 1.52863696e-01 4.37843144e-01 2.67398924e-01
2.55405605e-01 -8.48681092e-01 -1.02523839e+00 4.24551845e-01
1.00906885e+00 2.47974414e-03 2.73471743e-01 7.54430354e-01
2.35233381e-01 -3.62279803e-01 -6.24469459e-01 1.50252867e+00
8.49810958e-01 -1.13259184e+00 -2.00201437e-01 1.00317016e-01
1.83773339e-01 1.59904972e-01 2.78795391e-01 2.78998494e-01
-5.30133963e-01 -1.13164437e+00 5.94389260e-01 4.47409868e-01
1.30787849e+00 -5.94491959e-01 6.85024858e-01 1.72699496e-01
-1.19144249e+00 1.28034711e-01 -2.62941569e-01 3.86524171e-01
3.28738056e-02 9.97592211e-02 -8.35144639e-01 9.07685235e-03
6.35993123e-01 6.19256377e-01 -4.29489911e-01 1.03697288e+00
-1.85659781e-01 7.52089918e-01 -3.14676464e-01 -2.80495495e-01
-3.91104445e-02 4.69449878e-01 5.52402139e-01 1.37495136e+00
1.12390719e-01 -3.14528555e-01 -3.59174013e-01 4.10632044e-01
-3.72475535e-01 4.74276245e-01 -6.04342639e-01 -2.14479938e-02
1.02497809e-01 1.07367539e+00 -2.62949675e-01 -1.64932042e-01
-4.29505497e-01 6.64274096e-01 1.04025781e-01 2.06947356e-01
-7.38573253e-01 -4.47876751e-01 7.17383742e-01 2.61550754e-01
2.75571764e-01 -3.34633812e-02 4.53355685e-02 -1.05817950e+00
-1.04879886e-01 -9.42574263e-01 1.39223978e-01 -8.19682181e-01
-1.05453181e+00 5.07384241e-01 2.45148875e-02 -7.85309434e-01
-3.29783559e-01 -8.25661838e-01 -4.65878636e-01 6.77748919e-01
-2.02285314e+00 -1.23754585e+00 -3.34135890e-01 9.22155976e-01
1.23398471e+00 -5.40598691e-01 7.61770248e-01 5.40890917e-02
-3.85217577e-01 1.00094712e+00 -1.17362067e-01 2.65805572e-01
1.09413028e+00 -9.73667920e-01 2.80042201e-01 6.39575362e-01
2.76565522e-01 -1.07892910e-02 4.66655165e-01 -1.58908606e-01
-1.15387261e+00 -4.40459669e-01 1.45036793e+00 -1.51597001e-02
4.79687661e-01 -4.58758295e-01 -6.06275022e-01 1.95962563e-01
8.63399029e-01 6.58816472e-03 5.70410252e-01 -1.16520576e-01
-3.07579458e-01 -3.68074208e-01 -1.55716586e+00 4.72259730e-01
8.32449019e-01 -6.46093190e-01 -9.35233891e-01 -1.45205762e-02
3.80470872e-01 -1.80730537e-01 -5.57215571e-01 3.60565662e-01
9.30752397e-01 -1.08475196e+00 9.71642435e-01 -2.44550869e-01
8.93689692e-02 2.53186017e-01 -2.19505310e-01 -1.24783790e+00
3.04206312e-01 -9.84051168e-01 -3.11133921e-01 1.33238697e+00
2.18592092e-01 -5.50688505e-01 8.43933523e-01 1.05960071e-01
1.14589565e-01 -8.31775486e-01 -1.24785602e+00 -7.19348729e-01
5.41052282e-01 -4.29193944e-01 4.33527529e-01 1.90131888e-01
2.43689746e-01 2.50342578e-01 -4.82113719e-01 -5.64909279e-01
7.65485227e-01 -4.33840156e-01 7.12518096e-01 -1.29308951e+00
2.98412263e-01 -6.77297115e-01 -5.76493680e-01 -1.02169144e+00
8.92520845e-01 -5.19375980e-01 3.67834568e-01 -1.63301528e+00
1.81745604e-01 1.02706216e-01 -1.27306566e-01 5.25274754e-01
-9.58057344e-02 4.03763294e-01 5.25286734e-01 -6.72885822e-03
-2.80345559e-01 6.08915508e-01 1.10212004e+00 -4.00405467e-01
7.89914187e-03 1.29574656e-01 -1.67497844e-01 7.72795260e-01
9.46849167e-01 -8.36779475e-02 -1.35398790e-01 -1.92158222e-02
-3.38743836e-01 -3.29368524e-02 2.30926663e-01 -1.30389309e+00
2.44329169e-01 3.06029350e-01 2.20677704e-01 -9.34929013e-01
8.14736009e-01 -4.41402406e-01 -6.71872914e-01 2.83468783e-01
-5.13137698e-01 -1.73284113e-01 2.93563336e-01 6.70066252e-02
-3.29434812e-01 -2.19812930e-01 1.37617409e+00 1.48857459e-01
-8.79772127e-01 1.78244114e-01 -4.21979547e-01 6.80504143e-01
8.94515038e-01 -1.55131519e-01 -2.63933331e-01 -5.89530051e-01
-5.07796884e-01 -4.57659572e-01 3.59842271e-01 4.97395515e-01
8.05534422e-01 -9.19228971e-01 -1.03898835e+00 4.82327193e-01
4.96629626e-02 -7.01466084e-01 5.45395985e-02 9.17087257e-01
-3.30946535e-01 7.57234573e-01 -4.70247924e-01 -6.91111326e-01
-1.58529937e+00 4.22316253e-01 3.87068450e-01 7.17458725e-02
-6.16666555e-01 8.93616438e-01 -4.35900182e-01 5.44442795e-02
7.21821606e-01 -4.32933532e-02 -4.36458319e-01 5.08744478e-01
6.94275320e-01 6.83878899e-01 -4.75021377e-02 -1.11620712e+00
-6.42386258e-01 9.49235618e-01 2.90211409e-01 -4.92521912e-01
8.81105959e-01 -1.51843816e-01 3.59072417e-01 5.32711387e-01
1.37709975e+00 4.28926587e-01 -1.12426996e+00 2.89221644e-01
-3.44752729e-01 -3.66942018e-01 2.36924335e-01 -7.56148398e-01
-1.02895176e+00 1.42741656e+00 8.48728895e-01 2.12999955e-02
9.89568710e-01 1.62413433e-01 9.47094798e-01 1.32342800e-01
2.49333009e-01 -1.21574581e+00 8.57921392e-02 3.06419939e-01
8.27722192e-01 -1.61531484e+00 -2.98195571e-01 -3.64776254e-02
-7.10114479e-01 1.22621500e+00 1.90401211e-01 4.11677748e-01
8.17467093e-01 3.16308826e-01 4.91882473e-01 4.39336807e-01
-4.36575651e-01 -5.29080927e-01 3.27904016e-01 1.19611096e+00
5.61847806e-01 8.51827636e-02 -3.41205865e-01 1.18864261e-01
-1.78802833e-01 3.48385900e-01 2.26307794e-01 3.98023635e-01
-3.96537840e-01 -1.02918327e+00 -2.68091053e-01 1.33920449e-03
-7.09129393e-01 -2.89712965e-01 -3.23158532e-01 9.74121988e-01
-4.46675085e-02 1.44086480e+00 -1.55288324e-01 -1.31886780e-01
1.87636837e-01 5.01411855e-01 5.95557809e-01 6.46968931e-02
-1.22645490e-01 5.44249825e-02 5.90372682e-02 -5.90352714e-01
-5.81077635e-01 -9.19099808e-01 -1.00908339e+00 -2.82467753e-01
-1.38265014e-01 -7.25268126e-02 1.09969950e+00 8.12352419e-01
1.40165195e-01 -6.75023198e-02 3.98043126e-01 -1.06285810e+00
-6.20173752e-01 -1.38406050e+00 -5.96471786e-01 8.75361115e-02
7.93450832e-01 -7.44497895e-01 -8.38807106e-01 1.23474926e-01] | [14.32300853729248, 5.0104756355285645] |
5db3251c-aea1-4c0b-910b-40cf0b061011 | e-2-net-an-edge-enhanced-network-for-accurate | 2007.09791 | null | https://arxiv.org/abs/2007.09791v1 | https://arxiv.org/pdf/2007.09791v1.pdf | E$^2$Net: An Edge Enhanced Network for Accurate Liver and Tumor Segmentation on CT Scans | Developing an effective liver and liver tumor segmentation model from CT scans is very important for the success of liver cancer diagnosis, surgical planning and cancer treatment. In this work, we propose a two-stage framework for 2D liver and tumor segmentation. The first stage is a coarse liver segmentation network, while the second stage is an edge enhanced network (E$^2$Net) for more accurate liver and tumor segmentation. E$^2$Net explicitly models complementary objects (liver and tumor) and their edge information within the network to preserve the organ and lesion boundaries. We introduce an edge prediction module in E$^2$Net and design an edge distance map between liver and tumor boundaries, which is used as an extra supervision signal to train the edge enhanced network. We also propose a deep cross feature fusion module to refine multi-scale features from both objects and their edges. E$^2$Net is more easily and efficiently trained with a small labeled dataset, and it can be trained/tested on the original 2D CT slices (resolve resampling error issue in 3D models). The proposed framework has shown superior performance on both liver and liver tumor segmentation compared to several state-of-the-art 2D, 3D and 2D/3D hybrid frameworks. | ['Yu-Xing Tang', 'Ronald M. Summers', 'Yingying Zhu', 'Youbao Tang', 'Jing Xiao'] | 2020-07-19 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [-3.47916722e-01 2.58729249e-01 -2.65568256e-01 -2.53677696e-01
-3.35943848e-01 -1.97172180e-01 2.38272205e-01 -2.36961190e-02
-9.00290906e-02 2.69189596e-01 3.04308981e-01 -3.76084954e-01
1.32663712e-01 -8.16773713e-01 -1.95562795e-01 -8.17439079e-01
-4.62205142e-01 5.44918954e-01 3.68874818e-01 2.39078432e-01
-2.58766919e-01 8.42339873e-01 -6.02270067e-01 -1.33442357e-01
9.44857717e-01 1.17981875e+00 1.85156718e-01 5.13983428e-01
-3.03875625e-01 5.17326891e-01 2.79969871e-02 1.56815737e-01
7.24141777e-01 -6.60854399e-01 -7.19409347e-01 4.51581329e-01
2.00442486e-02 -2.90153563e-01 -4.13428575e-01 1.25245655e+00
7.46625006e-01 -2.96292067e-01 7.12161005e-01 -7.87130713e-01
-4.27318960e-01 6.25199378e-01 -7.49839664e-01 2.80257583e-01
-1.05875671e-01 2.79462308e-01 3.15717191e-01 -1.07496858e+00
4.01046336e-01 5.00668406e-01 9.11556304e-01 4.62173730e-01
-7.89365292e-01 -4.96331304e-01 -2.64250994e-01 -1.39121458e-01
-1.47478616e+00 -4.40984145e-02 7.71606207e-01 -5.63587308e-01
6.43661678e-01 8.69945437e-02 1.28628266e+00 2.22605199e-01
5.57087302e-01 9.90990222e-01 1.03237212e+00 -3.21361244e-01
-1.40176311e-01 1.90821916e-01 1.94538191e-01 1.31325591e+00
1.05075411e-01 3.84142965e-01 2.82037199e-01 1.59571990e-01
1.30867958e+00 3.54756922e-01 -7.41085351e-01 -7.04186320e-01
-1.36616576e+00 6.42038345e-01 1.13805926e+00 5.76805532e-01
-6.11599982e-01 -5.49357273e-02 4.01630700e-01 -4.90590511e-03
3.79996151e-01 2.41286494e-02 -2.77535379e-01 5.14954448e-01
-1.05324900e+00 -4.73659933e-01 8.42522025e-01 8.84866238e-01
5.75548708e-01 5.61053827e-02 -4.28323030e-01 4.59715605e-01
6.51637137e-01 1.54568091e-01 8.43208671e-01 -5.65997735e-02
-2.28131220e-01 9.03284967e-01 -2.41292298e-01 -4.73022938e-01
-8.14150512e-01 -8.50872576e-01 -1.43504798e+00 3.14399153e-01
3.11106920e-01 -3.49525623e-02 -1.36180425e+00 1.13571465e+00
7.56496608e-01 6.35559142e-01 -9.12870094e-02 1.08460462e+00
1.32998526e+00 1.04383163e-01 6.27700016e-02 -2.31919274e-01
1.63106632e+00 -1.52039719e+00 -4.42095131e-01 -9.16902125e-02
8.52328718e-01 -7.39538848e-01 5.68451762e-01 -3.67189854e-01
-1.22389090e+00 -2.64274001e-01 -7.08426595e-01 4.24965397e-02
-1.13213539e-01 3.45469147e-01 7.34078646e-01 5.57397187e-01
-1.25952613e+00 3.27412605e-01 -1.16039026e+00 -2.52264529e-01
6.30477071e-01 6.64613485e-01 -4.12748426e-01 -1.52475327e-01
-8.77482772e-01 1.23506975e+00 4.23950016e-01 3.23797822e-01
-1.17065132e+00 -9.01237905e-01 -1.19238424e+00 1.46173507e-01
1.65582389e-01 -9.61394250e-01 9.44582582e-01 -6.33800030e-01
-1.55975461e+00 1.14328551e+00 -9.32437368e-03 -3.71122569e-01
6.27958596e-01 5.18010080e-01 -2.63617113e-02 1.87222585e-01
-7.70744085e-02 6.67763948e-01 4.86406207e-01 -1.09900844e+00
-4.80164081e-01 -4.28530425e-01 -5.47929704e-01 4.25417274e-01
2.98551768e-01 -1.60325110e-01 -5.19293368e-01 -7.38766611e-01
4.62659359e-01 -8.51868510e-01 -4.59225059e-01 2.83151418e-01
-4.76305813e-01 -1.79177169e-02 9.27420020e-01 -1.07511175e+00
8.56529951e-01 -1.81095290e+00 -2.25058850e-02 3.51890117e-01
5.95505834e-01 2.23475829e-01 7.56866038e-02 -5.17032087e-01
-2.20599085e-01 -3.74393500e-02 -3.04529965e-01 -2.28646040e-01
-2.16364950e-01 6.34223670e-02 7.04767644e-01 7.37671852e-01
3.42078833e-03 1.44398570e+00 -8.22119892e-01 -7.05782950e-01
4.49739188e-01 5.89238107e-01 -2.74519444e-01 2.17940137e-01
3.33597362e-01 7.74770141e-01 -4.06810492e-01 1.04587388e+00
8.14039767e-01 -5.50399840e-01 -1.97948962e-02 -5.51072001e-01
-1.41314253e-01 -2.91615129e-02 -9.28052902e-01 1.68784201e+00
-3.48973751e-01 1.56009227e-01 4.09936160e-01 -9.31885839e-01
7.39362895e-01 7.96507835e-01 9.72911835e-01 -3.54583979e-01
5.71184218e-01 3.64918768e-01 1.38632864e-01 -5.57966590e-01
-2.16245234e-01 -3.43439370e-01 2.76677072e-01 4.38941628e-01
9.48266909e-02 -5.10270119e-01 -8.61275122e-02 -1.70214564e-01
7.67455339e-01 -4.94392365e-02 5.36880612e-01 -7.72508264e-01
7.43502080e-01 1.46161839e-01 7.00400114e-01 3.40479404e-01
-7.09074974e-01 5.48833072e-01 1.65234819e-01 -7.87589848e-01
-8.80863428e-01 -8.65784764e-01 -3.60353619e-01 2.09642947e-01
5.36409736e-01 1.63819388e-01 -7.26532280e-01 -1.40285683e+00
-9.82113853e-02 1.92660704e-01 -7.05650985e-01 -7.41066039e-02
-6.96948588e-01 -1.08476138e+00 2.38781035e-01 6.11338556e-01
9.59478259e-01 -6.72705233e-01 -4.91810828e-01 2.55600154e-01
6.33778125e-02 -5.71803629e-01 -1.02796853e+00 4.14457172e-01
-1.04845083e+00 -1.29540002e+00 -1.09063041e+00 -1.36763811e+00
1.19379330e+00 3.07029188e-01 1.24170756e+00 4.51509356e-01
-6.24101937e-01 1.69591039e-01 -8.81994218e-02 -2.39460260e-01
-3.61616015e-01 -1.21843033e-01 -1.64829642e-01 -3.20929915e-01
2.58707047e-01 -2.52437294e-01 -1.00667024e+00 4.36562330e-01
-6.97545648e-01 3.45990211e-01 8.39641988e-01 1.27585506e+00
7.79858589e-01 2.38009870e-01 1.73408106e-01 -5.55524349e-01
1.58652693e-01 -3.03248107e-01 -5.02448142e-01 4.05112892e-01
-6.14498913e-01 -9.93401930e-02 4.21404541e-01 -2.62554705e-01
-7.71050751e-01 4.53056842e-01 -2.93817103e-01 -5.17980695e-01
-1.44936861e-02 5.76483190e-01 1.52078748e-01 -7.41501093e-01
2.99545199e-01 3.22528392e-01 2.24153951e-01 -2.96893477e-01
-2.02615689e-02 2.88355231e-01 3.19481611e-01 7.33017176e-02
6.50513232e-01 3.59220892e-01 2.52087861e-01 -3.25594693e-01
-5.79640150e-01 -5.15405416e-01 -9.65615392e-01 -2.58346591e-02
8.84215772e-01 -9.68523979e-01 -3.16309392e-01 5.40663958e-01
-5.61805427e-01 -5.11905372e-01 -6.89155519e-01 1.05193102e+00
-2.67881960e-01 4.03716117e-01 -9.37386096e-01 -1.31997243e-01
-7.59149492e-01 -1.70526409e+00 9.73960876e-01 6.20667160e-01
3.95052522e-01 -1.45072699e+00 -2.40844816e-01 -1.56761765e-01
6.93148315e-01 4.65850145e-01 8.26255500e-01 -7.94405639e-01
-7.70156920e-01 -3.54219526e-01 -5.31020761e-01 1.63229004e-01
3.64124298e-01 -1.87603131e-01 -3.96957636e-01 -5.65731227e-01
3.11267555e-01 7.63323009e-02 8.20620835e-01 1.00242937e+00
7.63819158e-01 -1.43883899e-01 -6.86411738e-01 9.67939913e-01
1.44401264e+00 1.25242546e-01 2.20789820e-01 -1.49883315e-01
5.46832919e-01 1.04500046e-02 6.16198219e-02 5.88556826e-02
3.72560889e-01 2.38907561e-01 4.42458391e-01 -8.43250155e-01
-6.90351248e-01 1.20448112e-01 -5.03776371e-02 9.53226149e-01
1.49072975e-01 2.54976898e-01 -1.06722081e+00 6.87262237e-01
-1.36690938e+00 -4.86352682e-01 3.42323887e-03 2.01162362e+00
8.19466949e-01 -1.28564745e-01 -1.48067012e-01 -2.25957379e-01
7.15748787e-01 -3.25002283e-01 -4.64382887e-01 3.22033823e-01
2.08859131e-01 2.46671692e-01 4.67738301e-01 6.24399304e-01
-1.24445653e+00 5.20400703e-01 6.14262247e+00 7.41795838e-01
-1.52481937e+00 1.39848500e-01 9.16449845e-01 3.23119074e-01
-5.19755594e-02 -1.97919577e-01 -5.34575343e-01 2.17895627e-01
4.46592122e-02 -2.75262356e-01 9.05350670e-02 9.49046731e-01
1.38743669e-01 -2.12403014e-01 -1.21154857e+00 8.74933183e-01
4.95490357e-02 -1.48868167e+00 -3.70392382e-01 5.34419343e-02
8.56250405e-01 3.32229435e-01 -4.07085121e-01 2.31094077e-01
2.49461025e-01 -1.13386178e+00 1.36660412e-01 3.83585304e-01
9.72827017e-01 -3.51903617e-01 1.19410241e+00 4.97718245e-01
-1.55124378e+00 3.16641480e-01 -1.81356445e-01 5.40886641e-01
1.45438358e-01 7.46246457e-01 -1.33493125e+00 8.06339443e-01
4.79929745e-01 7.90230751e-01 -5.09542584e-01 1.53146493e+00
-1.60968095e-01 1.40685365e-01 -4.79201019e-01 2.38404125e-01
2.10213542e-01 -2.44453743e-01 5.30684710e-01 1.22434819e+00
6.67184591e-01 2.93870538e-01 6.69120073e-01 7.97281265e-01
-2.22461149e-01 7.06206039e-02 -2.44900107e-01 3.37491333e-01
1.17885560e-01 1.75220001e+00 -1.24158013e+00 -7.11782455e-01
-4.87521142e-01 8.52328598e-01 -9.48753506e-02 -1.13905534e-01
-7.96853364e-01 -1.25134945e-01 4.02321815e-02 4.12651300e-02
1.61443334e-02 8.63565132e-02 -2.55540878e-01 -1.60631430e+00
-5.22787035e-01 -2.79795825e-01 5.18355072e-01 -4.26129848e-01
-1.40298975e+00 7.77480781e-01 -3.08720946e-01 -1.21033001e+00
1.64067850e-01 -6.28945112e-01 -9.64688778e-01 1.02583992e+00
-1.93157983e+00 -1.27466452e+00 -8.22621942e-01 6.80623889e-01
3.76681596e-01 1.50629282e-01 7.48284876e-01 3.04845840e-01
-4.94576484e-01 3.71138960e-01 -1.62469730e-01 6.91118956e-01
3.65424067e-01 -1.44931555e+00 -8.39639008e-02 7.53367424e-01
-3.00039351e-01 1.65822148e-01 7.76836351e-02 -7.60841072e-01
-1.35971856e+00 -1.36715734e+00 5.52507579e-01 -2.15561651e-02
1.10232264e-01 1.96625516e-01 -6.72612190e-01 9.06467974e-01
3.18783432e-01 9.31579113e-01 7.07514644e-01 -7.22388864e-01
3.88859093e-01 1.34859040e-01 -1.79477239e+00 3.28171015e-01
7.08956003e-01 -5.82469553e-02 -5.38283169e-01 5.21203339e-01
4.49987978e-01 -1.01934409e+00 -1.34637129e+00 8.59833598e-01
2.85015196e-01 -8.28262746e-01 1.04372406e+00 -6.66062674e-03
-1.57260403e-01 -4.39899355e-01 3.48454446e-01 -1.47745264e+00
-6.23120964e-01 -2.67105997e-01 -5.90474419e-02 4.36132431e-01
2.79093355e-01 -5.28298855e-01 1.06426466e+00 5.64429283e-01
-7.18307316e-01 -1.03488708e+00 -9.06922102e-01 -3.53678405e-01
1.04912229e-01 2.81057090e-01 5.49405694e-01 1.09443367e+00
-1.44174974e-02 -1.19839728e-01 1.91511527e-01 2.89137602e-01
7.18110025e-01 2.90182114e-01 2.49403074e-01 -1.01851320e+00
1.13044903e-01 -8.22428942e-01 -3.89350265e-01 -1.21132088e+00
-2.17654616e-01 -1.36731398e+00 -2.68084481e-02 -1.72647381e+00
3.63713950e-01 -6.81508601e-01 -3.00290197e-01 6.38856769e-01
-1.74579129e-01 4.77940202e-01 -1.04384936e-01 2.50582933e-01
-1.66161329e-01 5.15670121e-01 1.96573877e+00 -2.52958000e-01
-2.98353910e-01 1.84704080e-01 -3.48304182e-01 8.16854298e-01
5.26746094e-01 -1.58964470e-01 9.92078856e-02 -2.63025910e-01
-7.06740916e-01 4.37046707e-01 3.88614208e-01 -9.27146316e-01
5.65095782e-01 5.18099144e-02 9.99816418e-01 -7.04970300e-01
-1.83091909e-01 -1.28003335e+00 2.75240034e-01 1.00030088e+00
7.80896395e-02 -1.32263318e-01 9.73430648e-02 7.47621730e-02
-3.61125976e-01 -3.48028868e-01 1.24502850e+00 -8.07625890e-01
-4.54264939e-01 1.01091564e+00 -1.23336818e-02 -3.71017158e-01
1.61925912e+00 -2.55983979e-01 1.43574283e-01 1.72811344e-01
-1.29475141e+00 5.30731440e-01 3.79873812e-01 -1.40776888e-01
7.38879561e-01 -1.49304855e+00 -7.77567267e-01 6.94284201e-01
-3.14727485e-01 6.69672668e-01 3.12017858e-01 1.60222685e+00
-9.01050627e-01 3.44143957e-01 -7.02969879e-02 -1.01061285e+00
-1.05226350e+00 5.25453508e-01 1.10297716e+00 -8.20896208e-01
-8.23799253e-01 1.00830781e+00 2.57614911e-01 -5.02748728e-01
1.88593924e-01 -6.91552341e-01 6.21754788e-02 -3.78363580e-01
3.32721770e-01 7.73537392e-03 8.08353275e-02 -5.40495515e-01
-4.05056030e-01 6.44408107e-01 3.71623300e-02 5.31599224e-01
1.22897136e+00 -1.86160550e-01 -4.93278444e-01 -3.09896499e-01
1.06267512e+00 -1.17485665e-01 -1.29840207e+00 -4.30129766e-01
-1.64545342e-01 -3.93165261e-01 5.38066149e-01 -9.92251337e-01
-1.61620343e+00 7.05708861e-01 8.00411642e-01 1.01936102e-01
1.32092285e+00 4.28764299e-02 9.33579206e-01 -3.83316517e-01
1.25112310e-01 -3.96675736e-01 -1.03505462e-01 4.00768012e-01
5.83769083e-01 -1.40982974e+00 1.17583364e-01 -5.76291263e-01
-4.73954558e-01 1.37126493e+00 5.99469185e-01 -2.73308843e-01
1.11676335e+00 5.77760100e-01 1.84462711e-01 -3.48984957e-01
-2.20503882e-01 -1.69729888e-01 4.80667770e-01 3.79120708e-01
4.60684359e-01 1.41289279e-01 -8.40509012e-02 5.55492580e-01
2.85192281e-01 3.78497154e-01 2.33507216e-01 8.99285972e-01
-4.77798939e-01 -8.06630254e-01 -1.45249486e-01 5.07577479e-01
-3.45726520e-01 -2.69908816e-01 1.74303412e-01 1.06817853e+00
2.31893763e-01 9.12193358e-02 -3.54887210e-02 -8.97946116e-03
9.28157941e-02 1.26500562e-01 5.49978554e-01 -6.16225779e-01
-1.02374268e+00 4.91712302e-01 -4.66032118e-01 -2.26478949e-01
-3.40343207e-01 -3.64919126e-01 -1.44518614e+00 -6.82273135e-02
-7.59295166e-01 3.53105396e-01 7.03693628e-01 6.49337590e-01
2.65126467e-01 6.01693988e-01 9.21276510e-01 -1.08079338e+00
-5.29512405e-01 -8.41058433e-01 -6.94381475e-01 1.19894713e-01
3.44895720e-01 -4.72472906e-01 -3.20033729e-01 -1.40581629e-03] | [14.496053695678711, -2.719008207321167] |
d5942524-ecdd-4a4e-978a-773328b70f68 | rethinking-pseudo-labels-for-semi-supervised | 2106.00168 | null | https://arxiv.org/abs/2106.00168v2 | https://arxiv.org/pdf/2106.00168v2.pdf | Rethinking Pseudo Labels for Semi-Supervised Object Detection | Recent advances in semi-supervised object detection (SSOD) are largely driven by consistency-based pseudo-labeling methods for image classification tasks, producing pseudo labels as supervisory signals. However, when using pseudo labels, there is a lack of consideration in localization precision and amplified class imbalance, both of which are critical for detection tasks. In this paper, we introduce certainty-aware pseudo labels tailored for object detection, which can effectively estimate the classification and localization quality of derived pseudo labels. This is achieved by converting conventional localization as a classification task followed by refinement. Conditioned on classification and localization quality scores, we dynamically adjust the thresholds used to generate pseudo labels and reweight loss functions for each category to alleviate the class imbalance problem. Extensive experiments demonstrate that our method improves state-of-the-art SSOD performance by 1-2% AP on COCO and PASCAL VOC while being orthogonal and complementary to most existing methods. In the limited-annotation regime, our approach improves supervised baselines by up to 10% AP using only 1-10% labeled data from COCO. | ['Larry S. Davis', 'Abhinav Shrivastava', 'Zuxuan Wu', 'Hengduo Li'] | 2021-06-01 | null | null | null | null | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 4.42969710e-01 1.61793441e-01 -4.42193091e-01 -6.82342827e-01
-1.47041333e+00 -8.04249704e-01 6.86074793e-01 3.14813018e-01
-7.10822582e-01 6.95657611e-01 -1.72076434e-01 5.46464510e-02
1.21594518e-01 -1.17779918e-01 -8.36222172e-01 -7.44248867e-01
2.89635122e-01 4.96940017e-01 4.89750057e-01 3.42179269e-01
2.31333837e-01 3.17976713e-01 -1.62319815e+00 4.30540085e-01
6.78491950e-01 1.27852201e+00 1.53717458e-01 7.25348830e-01
1.07900076e-01 7.83770740e-01 -6.50058687e-01 -2.48161450e-01
2.65690714e-01 -3.06223303e-01 -9.03788745e-01 3.55625629e-01
9.55224752e-01 -6.66103810e-02 2.11063370e-01 1.19654417e+00
3.91718477e-01 -2.37262502e-01 8.64215076e-01 -1.37822545e+00
-4.24322993e-01 5.40545285e-01 -6.68935180e-01 1.52449414e-01
4.67016734e-02 1.76417157e-01 1.25336993e+00 -1.14450312e+00
5.03807127e-01 1.01826727e+00 8.41455698e-01 4.91468191e-01
-1.72670722e+00 -7.16434419e-01 1.02726467e-01 -7.66277313e-02
-1.48016357e+00 -7.00811982e-01 3.57585460e-01 -6.48006201e-01
6.02187097e-01 1.47733405e-01 1.09515548e-01 9.88740861e-01
-1.26382515e-01 9.19867635e-01 1.36844218e+00 -6.69268370e-01
3.18936586e-01 5.93408287e-01 2.46701345e-01 6.13241017e-01
2.80735970e-01 -3.45328934e-02 -6.79630995e-01 -9.81051549e-02
2.44451210e-01 -4.94519114e-01 -1.86677158e-01 -6.32381678e-01
-1.33975136e+00 6.14664674e-01 5.63151777e-01 -1.55060560e-01
-1.27053991e-01 3.01925868e-01 2.84617126e-01 -2.94536240e-02
6.59699619e-01 6.98117733e-01 -5.04907966e-01 1.88224316e-01
-1.18375409e+00 1.73916861e-01 5.61036050e-01 1.13057745e+00
6.90916240e-01 -2.77137935e-01 -5.61686754e-01 9.88437772e-01
2.94633597e-01 5.95837653e-01 2.07321167e-01 -1.21822238e+00
2.91963607e-01 5.09446800e-01 4.34392631e-01 -4.94933575e-01
-3.98680866e-01 -7.16716588e-01 -2.04930693e-01 4.01445746e-01
7.03068316e-01 1.72667801e-01 -1.24808741e+00 1.79317427e+00
4.28232849e-01 -2.37460155e-02 -2.36446261e-01 9.12883162e-01
4.56373513e-01 1.98136225e-01 2.62869596e-01 -3.55610102e-02
1.41191554e+00 -1.12696767e+00 -4.79865640e-01 -4.82151747e-01
6.92762733e-01 -7.41282105e-01 1.22040796e+00 3.58739883e-01
-5.99015534e-01 -6.15406215e-01 -1.23865449e+00 3.11610587e-02
-1.24626808e-01 6.12464309e-01 2.73191154e-01 6.03877187e-01
-9.16747928e-01 3.97261590e-01 -8.63503575e-01 -2.55086601e-01
8.49812806e-01 2.76963651e-01 -3.38370740e-01 -5.21017797e-02
-6.98231578e-01 7.21252978e-01 4.38609779e-01 -1.08366758e-01
-1.02515292e+00 -6.91626489e-01 -7.20509052e-01 -2.45341793e-01
4.76469368e-01 -1.19129092e-01 1.35413682e+00 -9.80671823e-01
-1.06782365e+00 1.21983433e+00 3.84173580e-02 -5.33940792e-01
6.08992457e-01 -1.49041265e-01 -1.90619752e-01 1.00337610e-01
6.29713535e-01 1.33168077e+00 9.27133799e-01 -1.49494183e+00
-1.09360111e+00 -2.17274860e-01 -1.44471660e-01 2.91787446e-01
-1.68197900e-01 -3.35338265e-02 -5.27759910e-01 -4.13538843e-01
2.10434839e-01 -1.13947558e+00 -1.41151264e-01 4.31882948e-01
-3.79087389e-01 -2.86664575e-01 5.52187264e-01 -3.39177847e-01
6.27437174e-01 -2.29742336e+00 -1.69770390e-01 -2.86529157e-02
2.39783183e-01 1.63656220e-01 -2.78506815e-01 -2.33824417e-01
2.56290197e-01 -6.11584671e-02 -2.59146541e-01 -7.50103414e-01
-1.80261414e-02 1.02553323e-01 -2.20118016e-01 6.45266294e-01
5.62019110e-01 7.50108838e-01 -1.09703279e+00 -6.54623985e-01
6.40671626e-02 3.87548596e-01 -3.18152577e-01 1.67624369e-01
-3.50425720e-01 4.36349958e-01 -6.92334026e-02 8.10225427e-01
5.25021911e-01 -4.59277928e-01 -6.60870671e-02 -2.02524915e-01
-4.67170775e-03 3.48880112e-01 -1.15366912e+00 1.59231877e+00
-3.30471188e-01 7.55346894e-01 5.03685698e-02 -7.65693128e-01
7.04570413e-01 -1.41533129e-02 1.46655291e-01 -4.50013191e-01
4.54090349e-02 3.57955724e-01 -2.31258377e-01 -7.77906254e-02
4.45921451e-01 1.40187696e-01 5.28113879e-02 3.91881317e-01
4.12462443e-01 -2.61123389e-01 3.14850062e-01 1.63019449e-01
9.36405957e-01 4.58981782e-01 1.44065648e-01 -4.73125190e-01
2.78827161e-01 3.59357983e-01 6.35935307e-01 9.74277973e-01
-4.75129336e-01 9.70117271e-01 5.05310833e-01 -5.22352271e-02
-9.59474325e-01 -9.92026329e-01 -6.95854545e-01 1.32062411e+00
1.67466506e-01 -1.17371678e-01 -8.17295253e-01 -1.14385128e+00
1.21865109e-01 5.96312404e-01 -7.75036812e-01 -8.10616016e-02
-1.45635769e-01 -1.03838122e+00 7.60510504e-01 7.70793378e-01
2.98427254e-01 -8.21723223e-01 -4.97925520e-01 5.64679317e-02
-1.55516803e-01 -1.38374126e+00 -5.74904501e-01 6.76706016e-01
-5.83529234e-01 -1.08492064e+00 -6.43727243e-01 -7.82269537e-01
8.87478709e-01 3.57598573e-01 1.10155964e+00 -2.37549886e-01
-4.56437349e-01 1.99426562e-01 -2.83543259e-01 -4.74455267e-01
-4.48251545e-01 2.06045657e-01 9.14418772e-02 2.18424931e-01
2.86415845e-01 -1.86266042e-02 -6.74233437e-01 6.01841211e-01
-5.54840922e-01 -7.70771429e-02 6.07222497e-01 9.72293258e-01
8.96016479e-01 -3.04957151e-01 6.49804175e-01 -1.14838147e+00
1.28655165e-01 -2.44356066e-01 -8.60779107e-01 2.46992126e-01
-1.14747083e+00 2.28236720e-01 2.36136332e-01 -6.94572926e-01
-1.02773213e+00 5.60372829e-01 2.10915193e-01 -3.02918255e-01
-2.37095580e-01 -1.32779044e-03 6.46150634e-02 -3.39282840e-01
1.25975382e+00 -1.82802871e-01 -1.62430584e-01 -3.09119940e-01
5.60313642e-01 9.50737476e-01 7.70386875e-01 -4.66017455e-01
4.73768800e-01 5.39142489e-01 -1.92921266e-01 -3.07100534e-01
-1.52744949e+00 -7.11768627e-01 -8.07130456e-01 -1.50849596e-01
7.70009637e-01 -1.22119594e+00 -1.36496753e-01 3.66607517e-01
-7.53144860e-01 -4.38546807e-01 -3.72031927e-01 3.69456619e-01
-3.64103317e-01 6.69797435e-02 -4.82413083e-01 -6.96424365e-01
-1.56260654e-01 -1.24428666e+00 1.48103380e+00 1.21535458e-01
-1.53750420e-01 -5.33059597e-01 -1.28486276e-01 6.30042851e-01
3.23261529e-01 6.62623718e-02 3.20292532e-01 -7.38568187e-01
-4.80477035e-01 -4.34212476e-01 -6.22875035e-01 5.85830450e-01
-8.26251283e-02 -3.04438651e-01 -1.49474096e+00 -3.03895086e-01
-4.04225677e-01 -8.23299170e-01 9.28440213e-01 2.08157659e-01
9.85980213e-01 -2.42100121e-03 -4.22304720e-01 4.27772582e-01
1.13726413e+00 -2.09324598e-01 1.34526655e-01 2.86459059e-01
6.33140743e-01 7.21921206e-01 8.84204209e-01 5.71752638e-02
3.82761478e-01 8.01061451e-01 4.08619225e-01 -5.98144680e-02
-6.99368119e-01 -1.74199402e-01 1.41395971e-01 4.03844081e-02
4.68252569e-01 -6.25874326e-02 -9.72194970e-01 7.02294827e-01
-1.79135752e+00 -4.10505950e-01 -2.56602198e-01 2.25547528e+00
1.18836856e+00 4.74410385e-01 7.60748759e-02 1.77164093e-01
7.12280631e-01 -1.12468757e-01 -5.01251996e-01 3.43116760e-01
-9.85191613e-02 -7.23551065e-02 8.75632584e-01 4.26830590e-01
-1.42063773e+00 9.41835880e-01 6.42970991e+00 8.68030787e-01
-1.05063987e+00 4.31814671e-01 8.82799625e-01 -3.94752249e-02
2.73713559e-01 1.43123977e-02 -1.13436365e+00 3.85815024e-01
6.83216333e-01 5.14453769e-01 1.66395605e-01 1.06342793e+00
-1.82352543e-01 -4.33536202e-01 -1.33493054e+00 9.56719339e-01
1.85608104e-01 -9.15569305e-01 -3.39458287e-01 -1.33981079e-01
1.03835213e+00 3.21519941e-01 1.38204962e-01 1.26127407e-01
2.28515357e-01 -7.68575430e-01 1.19185722e+00 2.46697702e-02
1.15466332e+00 -3.60724866e-01 8.25583935e-01 1.94815874e-01
-8.77587259e-01 -6.55556621e-04 -2.78516114e-01 1.71345800e-01
-4.50341702e-02 7.92376101e-01 -9.98218477e-01 1.37231909e-02
7.68164456e-01 4.12307322e-01 -9.39723194e-01 1.03177226e+00
-6.55854821e-01 8.69437337e-01 -3.56046081e-01 1.39668226e-01
1.41034439e-01 3.00888330e-01 2.30347767e-01 1.31735957e+00
-2.22499907e-01 -2.13684648e-01 2.55084157e-01 9.82340872e-01
-2.60207176e-01 -1.55115679e-01 -7.60044828e-02 2.04389542e-01
7.13930070e-01 1.48018873e+00 -1.03943038e+00 -2.79108584e-01
-9.74353477e-02 9.27243412e-01 4.83533293e-01 1.94755420e-01
-8.83808613e-01 -3.29724610e-01 1.38400361e-01 1.66300267e-01
1.95974573e-01 -6.83576614e-03 -5.59811354e-01 -9.26677346e-01
1.54407322e-01 -5.66323519e-01 3.11100811e-01 -7.53249288e-01
-1.05200148e+00 5.39042115e-01 1.28120324e-02 -1.13837612e+00
-5.78392372e-02 -7.34438837e-01 -9.15682688e-02 6.18158698e-01
-1.59555113e+00 -1.33616424e+00 -4.34089333e-01 -6.89259842e-02
4.83823091e-01 1.49191663e-01 6.10054970e-01 4.16307837e-01
-4.37056571e-01 7.12263763e-01 3.36138643e-02 1.24113284e-01
1.14335358e+00 -1.54257858e+00 3.13346237e-01 9.45496261e-01
4.78723198e-01 1.70569107e-01 6.25165403e-01 -3.81832749e-01
-7.10413933e-01 -1.33482087e+00 8.81596744e-01 -9.85167027e-01
4.37238425e-01 -6.43376946e-01 -7.39673257e-01 6.04833245e-01
-3.88665766e-01 6.99904740e-01 3.56091201e-01 1.79449394e-01
-6.62371159e-01 -2.08332166e-01 -1.25628376e+00 2.32519209e-01
1.12876022e+00 -6.19487166e-01 -3.34210992e-01 6.13121510e-01
6.93365872e-01 -5.63970447e-01 -5.07452309e-01 5.50633371e-01
4.81886566e-01 -6.11113429e-01 7.79420376e-01 -1.94620997e-01
5.89370281e-02 -7.90619314e-01 -1.91512749e-01 -1.10131454e+00
-2.63718367e-01 -4.66827825e-02 2.31616259e-01 1.42473769e+00
6.99917555e-01 -3.23508829e-01 8.02555501e-01 4.26290005e-01
-8.07429552e-02 -5.02538621e-01 -9.40852880e-01 -7.83715963e-01
-4.71367687e-01 -5.59800386e-01 9.57453176e-02 8.19600761e-01
-4.25887734e-01 3.90063345e-01 -1.09768398e-01 3.64401549e-01
1.02599990e+00 -2.18479540e-02 6.82321012e-01 -1.03906167e+00
-3.61080617e-01 -2.87523985e-01 -3.85582715e-01 -1.04645550e+00
1.22560941e-01 -8.68207455e-01 7.00030506e-01 -1.06460094e+00
3.22310895e-01 -9.78175700e-01 -4.80874658e-01 7.44721234e-01
-2.74166286e-01 1.10384142e+00 1.48435757e-02 4.35266882e-01
-1.05130577e+00 2.58345246e-01 6.07906342e-01 -1.78774297e-01
-8.23085085e-02 -1.05161831e-01 -6.80476725e-01 5.99921167e-01
4.48400259e-01 -8.76734376e-01 -1.95896566e-01 -5.04126847e-01
9.42557231e-02 -4.29131716e-01 5.53510249e-01 -1.07353032e+00
1.85876772e-01 6.76979199e-02 3.58241946e-01 -3.95884007e-01
2.33621091e-01 -5.60119271e-01 -2.04864368e-01 3.41012567e-01
-7.32208133e-01 -5.37066221e-01 6.18053367e-03 7.27017283e-01
-7.23489374e-02 -2.05455884e-01 1.13979423e+00 2.05088496e-01
-6.84441864e-01 -1.16550110e-01 3.55609111e-03 1.74741656e-01
9.69737411e-01 7.57297203e-02 -3.02425504e-01 -4.81352359e-02
-6.01642370e-01 5.18755257e-01 3.35691035e-01 3.98577243e-01
1.81409016e-01 -1.21431577e+00 -7.42832601e-01 1.04983024e-01
6.72621191e-01 3.28504443e-01 -3.00373793e-01 8.72370422e-01
-3.41626167e-01 3.48591745e-01 7.13723525e-02 -1.18067646e+00
-1.11879241e+00 2.95979619e-01 3.52697849e-01 -1.59426946e-02
-1.84424862e-01 1.19151342e+00 1.43309638e-01 -5.05116642e-01
6.11585319e-01 -3.15271646e-01 -4.28261847e-04 9.34917405e-02
5.37168562e-01 2.51919627e-01 2.04702884e-01 -4.62697268e-01
-5.02603710e-01 4.30206835e-01 -1.85991600e-01 -2.20117852e-01
9.89014506e-01 -1.69176519e-01 1.52232334e-01 5.80877900e-01
8.58794868e-01 -1.30282089e-01 -1.66634214e+00 -3.27006847e-01
3.11286896e-01 -4.65419352e-01 1.94721133e-01 -1.12975156e+00
-7.39528537e-01 6.28842413e-01 8.64271581e-01 -1.52646258e-01
8.20697606e-01 4.05088514e-01 2.66476065e-01 1.94980711e-01
4.27714854e-01 -1.10374188e+00 2.37772793e-01 1.16249457e-01
5.94013572e-01 -1.66085136e+00 -9.49146971e-03 -5.65431178e-01
-5.38256943e-01 6.60056829e-01 7.62158275e-01 1.49180805e-02
2.23517075e-01 3.32714349e-01 2.55378157e-01 1.76639715e-03
-5.38882077e-01 -2.53205061e-01 5.30457199e-01 5.89034200e-01
3.22519779e-01 4.47465219e-02 -1.78100556e-01 4.09245461e-01
2.22083479e-01 -1.31218210e-01 3.44474673e-01 9.15100634e-01
-5.26293397e-01 -9.40355718e-01 -4.89330530e-01 5.39095283e-01
-4.99510020e-01 -9.15072933e-02 -2.22264484e-01 4.56760317e-01
3.42180461e-01 9.74792063e-01 7.83122256e-02 -2.17921034e-01
2.19014987e-01 1.03141911e-01 3.25623542e-01 -9.21742439e-01
-2.61889368e-01 2.07634047e-02 1.14716358e-01 -5.35606027e-01
-5.39440334e-01 -7.38176525e-01 -1.20652056e+00 4.46897656e-01
-1.05065334e+00 8.36415440e-02 8.46850753e-01 8.50460052e-01
3.33566308e-01 4.16672111e-01 4.44785058e-01 -8.39999199e-01
-9.59750295e-01 -1.07398021e+00 -4.14337933e-01 4.58202392e-01
3.84146273e-01 -8.98863077e-01 -5.20291984e-01 2.87619650e-01] | [9.248537063598633, 1.2629212141036987] |
94f7c48e-4087-4739-8e5f-36d3041d7802 | double-graph-based-reasoning-for-document | 2009.13752 | null | https://arxiv.org/abs/2009.13752v1 | https://arxiv.org/pdf/2009.13752v1.pdf | Double Graph Based Reasoning for Document-level Relation Extraction | Document-level relation extraction aims to extract relations among entities within a document. Different from sentence-level relation extraction, it requires reasoning over multiple sentences across a document. In this paper, we propose Graph Aggregation-and-Inference Network (GAIN) featuring double graphs. GAIN first constructs a heterogeneous mention-level graph (hMG) to model complex interaction among different mentions across the document. It also constructs an entity-level graph (EG), based on which we propose a novel path reasoning mechanism to infer relations between entities. Experiments on the public dataset, DocRED, show GAIN achieves a significant performance improvement (2.85 on F1) over the previous state-of-the-art. Our code is available at https://github.com/DreamInvoker/GAIN . | ['Lei LI', 'Runxin Xu', 'Shuang Zeng', 'Baobao Chang'] | 2020-09-29 | null | https://aclanthology.org/2020.emnlp-main.127 | https://aclanthology.org/2020.emnlp-main.127.pdf | emnlp-2020-11 | ['document-level-relation-extraction'] | ['natural-language-processing'] | [-1.39994860e-01 7.07922578e-01 -3.38362634e-01 -3.05972576e-01
-7.37852633e-01 -6.72047853e-01 6.75883770e-01 7.29243696e-01
-2.93790195e-02 8.63659859e-01 4.46987629e-01 -6.97513282e-01
-2.21861050e-01 -1.32322216e+00 -6.02530241e-01 4.26116735e-02
-3.52130353e-01 3.52510780e-01 4.53228831e-01 -1.61107615e-01
-2.07951471e-01 2.57664829e-01 -7.19415605e-01 4.53345150e-01
8.70074391e-01 6.68308139e-01 -2.91106671e-01 5.43172121e-01
-4.20525998e-01 1.38637114e+00 -5.62238276e-01 -9.90891159e-01
-1.51979044e-01 -2.20652491e-01 -1.29119432e+00 -9.47977975e-02
9.02750343e-02 -9.84502677e-03 -6.79531038e-01 1.01828599e+00
1.55798374e-02 -4.82444838e-02 5.20544112e-01 -1.25414848e+00
-6.23016238e-01 1.36887467e+00 -6.62405252e-01 1.99894726e-01
6.40514195e-01 -3.83107930e-01 1.68932366e+00 -9.36743677e-01
9.76294637e-01 9.85118747e-01 5.39612889e-01 1.63351089e-01
-1.02299917e+00 -4.63938177e-01 2.96464384e-01 2.11139262e-01
-1.59678018e+00 -1.83453605e-01 5.04506707e-01 -2.97656119e-01
1.58594084e+00 4.83323455e-01 4.54318941e-01 5.35010040e-01
1.34992808e-01 8.34269643e-01 6.52059793e-01 -4.01015669e-01
-2.06122220e-01 -1.56987309e-01 6.98236346e-01 1.01854634e+00
8.66792619e-01 -5.76107383e-01 -4.18586254e-01 -1.74587175e-01
1.90275267e-01 -2.85738736e-01 -2.92835325e-01 2.08161771e-01
-7.31522799e-01 4.55491215e-01 6.10730052e-01 4.07260746e-01
-3.32763493e-01 -8.60679150e-02 3.28965187e-01 7.70155415e-02
5.36264360e-01 5.80645800e-01 -6.56838179e-01 1.28078759e-01
-6.90726399e-01 1.12261876e-01 1.20371568e+00 1.32059014e+00
5.52590370e-01 -5.14453709e-01 -3.78994226e-01 6.39260352e-01
4.00684267e-01 7.06398338e-02 -1.31850660e-01 -3.13113809e-01
1.05256295e+00 1.35733008e+00 -2.41094872e-01 -1.31365585e+00
-5.67837954e-01 -6.02490485e-01 -8.94091189e-01 -6.70129001e-01
1.03608221e-01 -4.59014714e-01 -6.09996319e-01 1.30890000e+00
5.05416870e-01 7.05412626e-02 1.97718829e-01 3.68361861e-01
1.50159013e+00 5.25425673e-01 7.66403303e-02 -2.54274458e-01
1.77918267e+00 -1.04798019e+00 -1.00324917e+00 -3.05046856e-01
1.06244004e+00 -3.80627960e-01 4.70345110e-01 1.59517840e-01
-1.02649963e+00 5.49218268e-04 -1.06088877e+00 -2.47043490e-01
-7.54099190e-01 5.14092334e-02 9.52842832e-01 2.25456789e-01
-8.89511287e-01 4.69673425e-01 -7.90295839e-01 -1.24772675e-01
4.31115687e-01 2.85386890e-01 -6.14314735e-01 1.25284225e-01
-1.61146152e+00 8.76970589e-01 7.69731283e-01 8.73992965e-02
-2.22643197e-01 -7.01527953e-01 -1.15551925e+00 3.38141978e-01
1.05355811e+00 -9.66069877e-01 1.04695237e+00 2.34343231e-01
-9.03036714e-01 7.77828932e-01 -4.20517623e-01 -4.94444788e-01
-5.67945093e-02 -2.71538943e-01 -7.95301497e-01 4.67655584e-02
1.36679307e-01 -1.96760073e-02 -9.26232561e-02 -1.08181703e+00
-6.43150389e-01 -3.52886736e-01 4.59885627e-01 -1.13757201e-01
-2.14457855e-01 2.61517227e-01 -1.06889689e+00 -3.32106471e-01
1.92972948e-03 -7.38727570e-01 -7.30826110e-02 -8.08333814e-01
-1.39711905e+00 -6.08326316e-01 5.22529781e-01 -9.46998239e-01
2.12437725e+00 -1.49479318e+00 5.90229146e-02 4.26275015e-01
8.35580051e-01 2.13924736e-01 1.36619598e-01 9.54279542e-01
-2.67352343e-01 5.08377433e-01 -1.88377410e-01 -2.88494051e-01
2.84879729e-02 5.49530350e-02 -6.71039894e-02 5.03660478e-02
3.63629073e-01 1.32691073e+00 -9.85355377e-01 -7.57894397e-01
-3.23364496e-01 2.28717625e-01 -3.12997729e-01 1.42905407e-03
-3.68500143e-01 -1.67966858e-01 -5.87339342e-01 6.26352847e-01
7.38208234e-01 -7.62913942e-01 8.48982155e-01 -3.79239291e-01
1.46018684e-01 7.66706407e-01 -1.09857154e+00 1.46885800e+00
-2.81999409e-01 4.84756380e-01 -2.38958970e-01 -6.77407742e-01
8.51874828e-01 2.33013853e-01 2.86491215e-01 -2.52166599e-01
1.10547557e-01 -2.36505806e-01 -5.26549518e-02 -5.37344813e-01
6.11965179e-01 3.91215712e-01 -3.52825135e-01 1.88613549e-01
2.18426406e-01 1.60798281e-01 8.27444136e-01 1.17190588e+00
1.78652811e+00 -6.15576021e-02 5.99261522e-01 -3.13136764e-02
7.28090882e-01 1.79318096e-02 6.01716876e-01 5.11169314e-01
4.37511623e-01 -9.11199301e-02 1.15265977e+00 -1.16824582e-01
-4.66533244e-01 -9.31153297e-01 5.88707924e-02 3.84566665e-01
-1.36023602e-02 -1.60175896e+00 -5.24926960e-01 -1.10160506e+00
-2.24406295e-03 7.33555019e-01 -5.82246482e-01 1.83091879e-01
-6.21541440e-01 -7.68156946e-01 6.38127387e-01 5.86274981e-01
5.36346197e-01 -6.38096392e-01 2.49348715e-01 4.62822989e-02
-5.49740314e-01 -1.76332390e+00 -2.25817278e-01 -1.82535633e-01
-5.41728258e-01 -1.37235594e+00 2.82961965e-01 -5.43320656e-01
7.72838891e-01 -1.40130281e-01 1.54748309e+00 4.05492693e-01
-1.75191954e-01 6.16513798e-03 -4.32302207e-01 -3.62015098e-01
-1.85032487e-01 4.53592986e-01 -2.76273906e-01 -3.43819171e-01
5.16286910e-01 -5.81963837e-01 -3.12904865e-01 -1.18170157e-01
-7.20730126e-01 2.59684712e-01 5.83432615e-01 4.58196968e-01
5.62264860e-01 5.32895446e-01 4.11529720e-01 -1.52423251e+00
9.53396380e-01 -7.24156499e-01 -5.95284164e-01 6.24268591e-01
-9.17453945e-01 1.72327235e-01 5.16643584e-01 2.86361098e-01
-9.90806401e-01 -2.64013708e-01 -3.13013867e-02 3.56732517e-01
-2.78972797e-02 1.15614545e+00 -3.25170666e-01 5.73087990e-01
2.51969695e-01 -2.39849642e-01 -5.08559227e-01 -3.00796598e-01
6.26882970e-01 6.47144556e-01 5.05523682e-01 -5.95914960e-01
1.01368976e+00 1.92992344e-01 2.93023080e-01 -3.83994758e-01
-1.25316870e+00 -6.06979370e-01 -7.75529861e-01 -7.79527947e-02
8.42588127e-01 -9.84192252e-01 -9.74007010e-01 -5.83275482e-02
-1.37154222e+00 -1.32669985e-01 1.06022637e-02 2.02983439e-01
2.14618638e-01 1.49032086e-01 -1.07925940e+00 -6.47199512e-01
-6.84883118e-01 -5.46128273e-01 8.00836623e-01 2.25209817e-01
-4.40740168e-01 -1.21582317e+00 4.82189283e-02 4.82403636e-01
-2.27490664e-01 5.67997456e-01 7.69840837e-01 -7.70592451e-01
-7.17409194e-01 -2.88453072e-01 -5.16487598e-01 -1.12476043e-01
4.16948080e-01 2.75980413e-01 -4.48758841e-01 4.63825539e-02
-7.06085503e-01 1.30422533e-01 7.40913928e-01 -1.65873021e-01
1.04765511e+00 -6.87374294e-01 -9.19251323e-01 2.31938243e-01
1.39748275e+00 -6.08069152e-02 7.49494016e-01 2.90142298e-01
9.93163049e-01 6.31530225e-01 5.49854398e-01 2.01031640e-01
8.89668226e-01 6.77574039e-01 1.58353060e-01 -7.74902627e-02
-2.30430722e-01 -5.09105623e-01 6.91072792e-02 1.04750335e+00
-2.27010354e-01 -5.05426526e-01 -1.17488563e+00 5.58319390e-01
-1.98856592e+00 -8.96380365e-01 -8.53890300e-01 1.53981483e+00
1.07012796e+00 3.77424926e-01 9.23102275e-02 7.35497475e-02
5.54200470e-01 5.32285348e-02 -1.13623351e-01 -2.51290202e-01
-2.76765004e-02 2.16990918e-01 4.26695228e-01 7.88247526e-01
-9.44665492e-01 1.16259634e+00 5.04203224e+00 8.05596352e-01
-4.47875977e-01 2.83742533e-03 1.98044941e-01 1.33400977e-01
-4.28934157e-01 3.96408319e-01 -1.24648833e+00 1.82089150e-01
1.12027264e+00 -5.57358086e-01 -3.17799784e-02 4.10085440e-01
-7.09785819e-02 -3.12638208e-02 -9.30239797e-01 4.23600137e-01
-1.48272008e-01 -1.62132990e+00 5.21688573e-02 1.32561833e-01
5.23935497e-01 2.51678322e-02 -5.85881293e-01 3.95645499e-01
8.60034645e-01 -7.80293226e-01 1.89357758e-01 5.82554579e-01
5.86639881e-01 -8.06086540e-01 8.57174635e-01 3.08598965e-01
-1.76752710e+00 2.23885298e-01 2.12286226e-02 1.56206805e-02
3.86452198e-01 1.20726156e+00 -1.10193098e+00 1.52060831e+00
5.38091540e-01 8.22630703e-01 -7.00996876e-01 4.96988863e-01
-9.60201919e-01 6.87796414e-01 -3.95317189e-02 -1.79767430e-01
2.07237676e-02 -1.76330179e-01 4.83769268e-01 1.72433281e+00
-9.20587778e-02 4.91061836e-01 3.31876911e-02 7.41750479e-01
-6.42987013e-01 1.87738448e-01 -6.28785968e-01 -1.39555603e-01
6.91846073e-01 1.74254572e+00 -7.15896308e-01 -5.51686108e-01
-6.09642684e-01 7.42116153e-01 8.31977129e-01 2.72076130e-01
-8.09666514e-01 -8.03045154e-01 4.46136296e-01 5.69014251e-02
3.50702435e-01 -2.50423014e-01 -6.86070770e-02 -1.28734696e+00
3.59629363e-01 -5.19041121e-01 7.92949200e-01 -6.73382998e-01
-1.13852549e+00 7.21253395e-01 1.41487256e-01 -9.46519613e-01
-7.99408630e-02 -4.28562075e-01 -5.18185496e-01 6.69556379e-01
-1.23406875e+00 -1.25229490e+00 -1.17377907e-01 3.56787831e-01
3.82991834e-03 2.90280968e-01 8.31542611e-01 3.77341866e-01
-8.83809626e-01 5.49694657e-01 -6.37885749e-01 8.42455626e-01
2.75688320e-01 -1.51760924e+00 7.72172213e-01 1.12853897e+00
4.37398016e-01 1.11909842e+00 4.49283093e-01 -1.04103816e+00
-1.40008211e+00 -1.33209097e+00 1.59691787e+00 -7.08898783e-01
1.16490698e+00 -4.99351144e-01 -8.97898257e-01 1.21882331e+00
4.87302423e-01 -2.20987294e-02 9.40252542e-01 7.07090080e-01
-6.50825560e-01 -5.33363409e-02 -8.87949705e-01 7.54633129e-01
1.40711570e+00 -4.31902498e-01 -5.73377669e-01 4.51842606e-01
1.10639250e+00 -6.53987527e-01 -1.46279931e+00 4.91906106e-01
2.58707047e-01 -5.36192000e-01 8.53455544e-01 -7.65474439e-01
6.79254115e-01 -4.70481813e-01 1.27605841e-01 -1.26287365e+00
-4.87427235e-01 -6.56371117e-01 -1.02059245e+00 1.72363389e+00
1.18246067e+00 -7.34255433e-01 5.52185714e-01 6.88033998e-01
-1.73728429e-02 -1.23602486e+00 -2.93673486e-01 -8.82767737e-01
-1.89741507e-01 -4.55479145e-01 9.04891431e-01 1.09182715e+00
7.03461051e-01 1.06700027e+00 2.53006257e-02 6.89330697e-01
5.96086621e-01 3.30416471e-01 6.00243747e-01 -1.25373220e+00
-4.06802565e-01 -3.34434599e-01 -3.05179954e-01 -8.20439398e-01
4.31434542e-01 -1.26296723e+00 -4.02933121e-01 -2.41954947e+00
3.59500259e-01 -2.19405666e-01 -6.39595687e-02 8.58110964e-01
-3.10128748e-01 -1.93210930e-01 5.66432066e-02 -1.09171823e-01
-9.33092654e-01 2.64121026e-01 1.15840268e+00 -2.98969328e-01
-1.38660565e-01 -1.49286136e-01 -1.13887548e+00 6.26817644e-01
9.89401281e-01 -5.98238945e-01 -4.33492810e-01 -3.43627453e-01
7.10224688e-01 2.19631106e-01 -8.07033256e-02 -5.44988930e-01
5.80608547e-01 3.26486081e-02 -1.71918310e-02 -6.80907547e-01
9.09727812e-02 -5.33588111e-01 1.98713481e-01 3.29956055e-01
-3.20959896e-01 5.59431389e-02 3.06367934e-01 3.86155188e-01
-3.93505096e-01 -1.14232622e-01 -5.60681894e-02 -1.18779792e-02
-4.38752145e-01 3.74026150e-01 1.01960510e-01 2.81016648e-01
8.67417872e-01 5.30058980e-01 -9.66832697e-01 -2.51653135e-01
-8.26260090e-01 5.91470599e-01 -1.83056340e-01 3.62036824e-01
6.27315164e-01 -1.13702834e+00 -9.61505771e-01 -4.19687122e-01
2.02402517e-01 2.31946617e-01 3.27900834e-02 9.79989231e-01
-2.64437795e-01 6.60200775e-01 6.47853374e-01 3.71383466e-02
-1.77106833e+00 4.51798469e-01 9.08588096e-02 -1.07894826e+00
-6.73115134e-01 9.09512937e-01 -2.16814950e-01 -4.14195925e-01
-2.51767427e-01 -3.54953527e-01 -4.37450767e-01 -3.53792608e-02
5.99462390e-01 2.50294328e-01 2.49808148e-01 -4.45652246e-01
-7.05497742e-01 1.95523664e-01 -3.28929096e-01 1.37129992e-01
1.30770063e+00 -2.12107394e-02 -8.23134363e-01 2.10069105e-01
1.05523217e+00 4.91325796e-01 -3.73897225e-01 -4.18891072e-01
5.57538688e-01 -2.10275769e-01 -1.04815573e-01 -8.60995114e-01
-9.74786043e-01 1.89685941e-01 -6.65344536e-01 7.76514947e-01
9.66748953e-01 5.42334318e-01 7.69227087e-01 4.92115676e-01
3.05651098e-01 -6.53514206e-01 -3.89992923e-01 6.55489624e-01
9.25731003e-01 -1.00481939e+00 4.66438115e-01 -1.26695645e+00
-5.23540974e-01 8.11545670e-01 7.45625257e-01 1.43378004e-01
8.49435389e-01 7.82283902e-01 -4.41532314e-01 -7.47084916e-01
-1.15272129e+00 -4.55231309e-01 5.81927419e-01 1.88274324e-01
7.96881676e-01 3.39807421e-01 -4.62319344e-01 8.70906174e-01
-3.35328251e-01 -8.43988881e-02 5.58157742e-01 7.80446470e-01
-7.16758445e-02 -1.47324646e+00 2.11148292e-01 6.73697412e-01
-6.09993935e-01 -5.06103337e-01 -8.02131772e-01 8.51694345e-01
1.60381813e-02 1.29918981e+00 -1.96487725e-01 -7.03050613e-01
6.65567636e-01 -4.15622443e-02 3.58515888e-01 -8.68664742e-01
-5.66728950e-01 -4.57479388e-01 1.02696419e+00 -4.97709125e-01
-3.45465541e-01 -4.65732068e-01 -1.77655900e+00 -4.26021636e-01
-4.91697252e-01 4.86670852e-01 2.42838696e-01 8.62702131e-01
6.43966556e-01 1.03610504e+00 2.95801669e-01 2.12339520e-01
3.11594486e-01 -1.01548302e+00 -6.44583762e-01 2.51248598e-01
4.14298922e-02 -3.70386779e-01 -1.31892607e-01 -4.26552407e-02] | [9.210741996765137, 8.584807395935059] |
26770479-baf1-45bb-bd60-aa475dd9f475 | i-can-read-your-mind-control-mechanism | 2205.03556 | null | https://arxiv.org/abs/2205.03556v1 | https://arxiv.org/pdf/2205.03556v1.pdf | I Can Read Your Mind: Control Mechanism Secrecy of Networked Dynamical Systems under Inference Attacks | Recent years have witnessed the fast advance of security research for networked dynamical system (NDS). Considering the latest inference attacks that enable stealthy and precise attacks into NDSs with observation-based learning, this article focuses on a new security aspect, i.e., how to protect control mechanism secrets from inference attacks, including state information, interaction structure and control laws. We call this security property as control mechanism secrecy, which provides protection of the vulnerabilities in the control process and fills the defense gap that traditional cyber security cannot handle. Since the knowledge of control mechanism defines the capabilities to implement attacks, ensuring control mechanism secrecy needs to go beyond the conventional data privacy to cover both transmissible data and intrinsic models in NDSs. The prime goal of this article is to summarize recent results of both inference attacks on control mechanism secrets and countermeasures. We first introduce the basic inference attack methods on the state and structure of NDSs, respectively, along with their inference performance bounds. Then, the corresponding countermeasures and performance metrics are given to illustrate how to preserve the control mechanism secrecy. Necessary conditions are derived to guide the secrecy design. Finally, thorough discussions on the control laws and open issues are presented, beckoning future investigation on reliable countermeasure design and tradeoffs between the secrecy and control performance. | ['Xinping Guan', 'Lin Cai', 'Yushan Li', 'Jianping He'] | 2022-05-07 | null | null | null | null | ['inference-attack'] | ['adversarial'] | [-7.46132014e-03 -1.41321095e-02 -4.89491105e-01 2.08645985e-01
-4.81628329e-02 -1.19608223e+00 5.43437898e-01 -1.93175673e-01
-1.50270626e-01 8.79743516e-01 -1.84952602e-01 -7.50229418e-01
-6.63370252e-01 -8.24577510e-01 -4.84267890e-01 -1.08276141e+00
-6.51057601e-01 -3.04342538e-01 7.20261633e-02 -3.51315647e-01
1.27164230e-01 7.37010479e-01 -8.95423114e-01 -5.96355915e-01
4.39794749e-01 1.23541820e+00 -4.63668972e-01 7.41246462e-01
4.81311232e-01 5.90197682e-01 -8.08198750e-01 -1.85979977e-01
6.47570968e-01 -2.77221411e-01 -3.81991178e-01 -3.16687584e-01
-1.19660974e-01 -5.45570374e-01 -7.80202389e-01 1.31422412e+00
3.19394290e-01 -1.54529467e-01 4.07602578e-01 -1.78395808e+00
-5.33253193e-01 6.37268126e-01 -1.76208064e-01 1.33876532e-01
-2.31395826e-01 1.50192603e-01 7.35016108e-01 4.61188518e-02
3.27361733e-01 1.12174499e+00 3.70209724e-01 9.74121094e-01
-1.00440967e+00 -1.27455270e+00 3.44588757e-01 -1.47765040e-01
-1.20024812e+00 -3.30359459e-01 5.72917283e-01 -2.05859229e-01
3.37544829e-01 7.35560596e-01 6.68543816e-01 1.21413827e+00
6.15979433e-01 5.13691485e-01 1.14848089e+00 -4.48657311e-02
3.85525942e-01 2.24001721e-01 3.38576257e-01 6.88516736e-01
1.04981697e+00 8.85122180e-01 -3.69460043e-03 -6.57045305e-01
9.94922698e-01 -1.03131011e-01 -6.21595025e-01 -6.17785573e-01
-7.93955624e-01 7.48854816e-01 1.75566792e-01 -6.94860220e-02
1.12361833e-01 2.25252703e-01 4.65599775e-01 6.93225205e-01
-3.31758559e-02 4.58750814e-01 -8.13285232e-01 3.62163484e-01
-1.03835657e-01 2.04386637e-01 1.33855438e+00 8.48300397e-01
2.16705725e-01 4.87696022e-01 6.15471676e-02 -2.93815523e-01
4.87757564e-01 1.09970355e+00 -4.35534805e-01 -8.86839867e-01
4.65714157e-01 1.83827773e-01 2.71563947e-01 -1.27115107e+00
-2.95136780e-01 -5.27239978e-01 -1.41741157e+00 1.62304327e-01
2.72504807e-01 -8.52098823e-01 -5.19734561e-01 2.07331967e+00
5.30295491e-01 3.48303258e-01 4.19090927e-01 5.79870164e-01
5.56386225e-02 7.62024164e-01 -4.35009688e-01 -7.18767285e-01
1.06300616e+00 -7.79580548e-02 -1.01309073e+00 2.11699769e-01
3.18605781e-01 -2.43480906e-01 3.96297425e-01 -1.11420706e-01
-8.79349172e-01 2.09573284e-01 -1.15936494e+00 7.69411683e-01
-1.87931806e-01 -3.49041522e-01 5.00972033e-01 1.14555585e+00
-7.09403515e-01 2.14898631e-01 -1.26743090e+00 8.12514350e-02
1.21811114e-01 6.78218246e-01 -5.29919118e-02 6.39251292e-01
-1.78921986e+00 7.60856509e-01 3.17738056e-01 2.76638418e-01
-1.29803431e+00 -9.93739605e-01 -8.17994237e-01 2.95738995e-01
9.27512407e-01 -6.02153301e-01 9.76927340e-01 3.59009802e-02
-1.66258276e+00 3.66458856e-02 4.79429126e-01 -8.18991303e-01
3.44267279e-01 2.07660440e-02 -5.56962371e-01 1.49060950e-01
-3.75017583e-01 -4.44912583e-01 9.44783449e-01 -1.33510506e+00
-4.98457640e-01 -2.13671207e-01 3.97438079e-01 -2.92565763e-01
-5.14180839e-01 -8.27981234e-02 2.63849825e-01 -5.29111803e-01
6.61184788e-02 -1.18658519e+00 -3.48760337e-01 7.74887577e-02
-8.91619802e-01 3.44918013e-01 1.42986310e+00 -1.71034023e-01
1.42311239e+00 -1.76239729e+00 1.60099596e-01 6.55991614e-01
3.20530087e-01 5.64278424e-01 2.85325140e-01 9.49764490e-01
6.97740614e-02 5.47052324e-01 8.66724085e-03 1.63090341e-02
1.99183270e-01 1.26690328e-01 -1.16106617e+00 8.58381987e-01
-2.85408348e-01 5.80447435e-01 -4.40007210e-01 1.24808349e-01
1.98210552e-01 3.75403970e-01 -4.51722056e-01 2.80316800e-01
1.48948487e-02 7.72315741e-01 -1.16280377e+00 4.37139779e-01
8.00900996e-01 -2.47568920e-01 4.09796238e-01 -1.65531844e-01
-3.51822853e-01 1.76619552e-02 -1.35045695e+00 4.52679127e-01
-2.61514634e-01 4.18097712e-03 6.74493253e-01 -8.05388033e-01
6.77633762e-01 4.03281480e-01 3.21981847e-01 8.61071423e-03
4.52909738e-01 5.05997129e-02 2.85055131e-01 -1.86671332e-01
-1.36905104e-01 4.52722013e-02 -5.69866717e-01 5.83638608e-01
-3.70110333e-01 -1.14980236e-01 -4.35332447e-01 3.79302502e-01
8.98900330e-01 -6.99032068e-01 5.44545531e-01 -6.84813380e-01
9.24046040e-01 -4.36290622e-01 8.05935681e-01 1.01007330e+00
-3.45627308e-01 -6.08049572e-01 7.42621422e-01 -1.88362628e-01
-7.93491185e-01 -1.19228923e+00 -3.47666442e-02 2.81940132e-01
5.64176142e-01 -3.15199882e-01 -6.20127439e-01 -7.02850938e-01
1.74136981e-01 2.26339400e-01 -7.50576079e-01 -8.45285714e-01
-3.46754819e-01 -6.93776965e-01 1.03126013e+00 1.53183803e-01
1.05054128e+00 -1.38108775e-01 -1.77630097e-01 -2.05720142e-01
1.94965169e-01 -1.08270490e+00 -3.47004324e-01 -1.08335443e-01
-5.76084435e-01 -1.37483883e+00 5.81270792e-02 -2.89291233e-01
6.22756302e-01 2.20156327e-01 1.84592858e-01 3.10225308e-01
3.30893211e-02 6.22798383e-01 1.04851700e-01 -6.20205224e-01
-5.84357917e-01 4.20591757e-02 8.49522591e-01 4.39271778e-02
-3.87440562e-01 -4.85116810e-01 -4.02529597e-01 5.11275172e-01
-8.57156277e-01 -3.87681633e-01 3.44967782e-01 8.53284657e-01
2.74350762e-01 4.98246104e-01 4.74337190e-01 -4.16042566e-01
6.00771487e-01 -4.14417177e-01 -1.50712717e+00 4.76691723e-01
-6.84357762e-01 1.16010241e-01 9.21089947e-01 -3.51701826e-01
-9.56831753e-01 -3.46921623e-01 2.60819435e-01 -3.15865308e-01
3.10016304e-01 1.68493822e-01 -5.23237824e-01 -6.12686396e-01
1.21767603e-01 4.19516474e-01 4.55247015e-01 -2.77538180e-01
-1.78900233e-03 4.04229701e-01 1.36727795e-01 -7.26910293e-01
1.53885090e+00 8.05939794e-01 7.39043593e-01 -1.02512693e+00
-1.08152330e+00 1.41291901e-01 -2.44872063e-01 -8.41335431e-02
6.02721334e-01 -5.01470089e-01 -1.74256349e+00 7.74976730e-01
-9.70953047e-01 -5.65961450e-02 1.38564929e-01 5.62238991e-01
-5.09455323e-01 3.63795459e-01 -9.98866320e-01 -1.50709581e+00
-2.90638357e-01 -8.41675878e-01 4.67053980e-01 3.08459193e-01
5.65984905e-01 -1.31598341e+00 2.03013405e-01 -1.27249748e-01
4.98424351e-01 4.74079609e-01 6.52809501e-01 -6.67963326e-01
-1.05226719e+00 -3.68714511e-01 3.74715403e-02 4.44492728e-01
1.78486526e-01 1.39943780e-02 -5.36109865e-01 -8.96777630e-01
7.14018464e-01 -2.90334038e-02 5.00194192e-01 3.40364546e-01
1.02014959e+00 -1.02856374e+00 -4.86471474e-01 7.39834607e-01
1.38818336e+00 3.21253568e-01 3.96195799e-01 2.48661507e-02
5.70424736e-01 5.78632772e-01 3.52679342e-01 5.46793103e-01
7.19462931e-02 2.06576437e-01 7.83313870e-01 1.09803118e-01
8.16942096e-01 -3.82178307e-01 5.22185087e-01 8.34758103e-01
1.16623938e-02 -5.51673412e-01 -3.45561326e-01 -3.07150185e-01
-1.69297552e+00 -1.10233891e+00 5.45632020e-02 2.35975933e+00
6.92652583e-01 1.75102234e-01 -1.11646064e-01 9.98441800e-02
9.57006156e-01 2.96011806e-01 -7.13625848e-01 -6.32464960e-02
-8.67373273e-02 -3.93930674e-01 1.01797986e+00 6.70250237e-01
-1.27419829e+00 6.35269701e-01 6.93118668e+00 6.21973634e-01
-1.04135215e+00 -1.72305807e-01 2.93180853e-01 1.90762833e-01
9.89458058e-03 1.86156556e-01 -1.20326698e+00 4.91128027e-01
8.47029328e-01 -4.17456090e-01 5.39229333e-01 5.13312817e-01
4.18687999e-01 4.33065146e-01 -5.23377597e-01 4.16529775e-01
-3.82002056e-01 -1.45711195e+00 8.43895301e-02 5.52742541e-01
6.89265311e-01 -4.73741412e-01 3.49563479e-01 -3.25021930e-02
5.25786757e-01 -5.60304821e-01 2.24340498e-01 4.78664339e-01
5.89084983e-01 -1.00360548e+00 5.79446971e-01 5.44848084e-01
-1.29032135e+00 -5.72909653e-01 -2.00002491e-01 -2.14436606e-01
2.78560758e-01 3.01672250e-01 -9.52161327e-02 9.11868632e-01
2.05757633e-01 5.92786312e-01 1.76278964e-01 6.50187850e-01
-4.58950907e-01 6.96760297e-01 -5.25788963e-01 -1.76453769e-01
1.75660849e-01 -1.32408351e-01 1.05882275e+00 5.50732374e-01
1.35586515e-01 4.20532405e-01 4.40328628e-01 8.06148529e-01
2.59979069e-01 -6.97330475e-01 -6.78099632e-01 -2.74439126e-01
1.06158125e+00 9.75241661e-01 -3.28271836e-01 -8.88970345e-02
-1.42600209e-01 2.09492028e-01 -3.36726069e-01 3.94852966e-01
-9.28459346e-01 -4.40034509e-01 1.52114666e+00 -1.25936285e-01
1.73547685e-01 -6.64063334e-01 -1.71380982e-01 -1.46594858e+00
-3.09815437e-01 -8.19308579e-01 4.94407147e-01 1.89189687e-01
-1.42077565e+00 1.73583150e-01 2.78128982e-01 -1.41423154e+00
-1.88910477e-02 -4.69923854e-01 -6.83726609e-01 4.41527516e-01
-1.33953285e+00 -1.05901885e+00 3.78699362e-01 7.49639869e-01
-3.42068523e-01 -2.28238255e-01 7.07083046e-01 5.50812073e-02
-9.73744571e-01 6.35928690e-01 4.09994483e-01 4.00603592e-01
1.01579659e-01 -7.26629317e-01 2.20807567e-01 1.15410805e+00
-2.66170263e-01 1.21657360e+00 8.37686837e-01 -9.30982113e-01
-1.99701369e+00 -9.75639224e-01 2.59003669e-01 -2.75012404e-01
1.35993242e+00 -6.96740627e-01 -6.53543115e-01 7.07787097e-01
-1.64240986e-01 3.02524604e-02 4.05651182e-01 -2.64768630e-01
-3.41696054e-01 -1.24609172e-01 -1.27689373e+00 7.69797146e-01
9.95273113e-01 -5.28498292e-01 -3.50498706e-01 6.86720982e-02
1.07869172e+00 -4.14263964e-01 -8.18951547e-01 3.97459090e-01
5.31858742e-01 -4.02463764e-01 1.09607315e+00 -1.04718268e+00
-4.60112572e-01 -4.49225843e-01 -1.83133617e-01 -9.81339693e-01
2.12091003e-02 -1.35248661e+00 -6.53372169e-01 1.14802516e+00
1.19575717e-01 -1.17217696e+00 5.45019925e-01 5.53152919e-01
2.27660537e-01 -5.59775949e-01 -1.01374424e+00 -1.41623092e+00
1.66439027e-01 -1.01879053e-01 7.12753952e-01 1.06294274e+00
4.10221666e-02 8.75308439e-02 -9.12803650e-01 1.09412384e+00
1.09265172e+00 7.28258640e-02 6.64933801e-01 -1.07754087e+00
-1.37682438e-01 -2.82513984e-02 -2.97113329e-01 -9.51613724e-01
3.54831219e-01 -4.83416557e-01 -5.03368556e-01 -5.89217722e-01
-4.60369699e-02 -3.85036170e-01 -2.56368101e-01 2.29637310e-01
3.57587397e-01 -2.00518250e-01 3.95573884e-01 2.55381912e-01
-3.30991060e-01 6.62546873e-01 1.46913254e+00 1.16979271e-01
-1.49864689e-01 5.90242445e-01 -6.41514838e-01 6.42193556e-01
9.50237811e-01 -4.61858660e-01 -8.32073510e-01 8.86907578e-02
-1.64044965e-02 4.56841677e-01 7.50499427e-01 -6.67591751e-01
4.27646041e-01 -6.86706960e-01 -1.86325312e-01 -4.60007459e-01
2.67034322e-01 -1.29732263e+00 7.04960153e-02 1.27979171e+00
-3.35565835e-01 -2.24794313e-01 4.72479053e-02 1.19711244e+00
2.59870231e-01 2.12736979e-01 9.65420961e-01 2.48014763e-01
-9.24563184e-02 7.73835301e-01 -6.15520656e-01 2.42210496e-02
1.26816058e+00 3.41214895e-01 -6.64393783e-01 -6.75661564e-01
-8.73763740e-01 6.85596526e-01 5.23234718e-02 1.39523327e-01
2.94419944e-01 -1.15188003e+00 -3.08170080e-01 5.69269478e-01
-5.09990513e-01 -5.39787173e-01 4.21588272e-01 7.42079079e-01
1.04372025e-01 7.30360746e-01 -9.32051241e-02 -2.66305983e-01
-1.16112304e+00 7.52696812e-01 7.28729069e-01 -2.12271303e-01
-2.82436728e-01 3.79300654e-01 3.37807238e-01 -1.56503633e-01
3.40198100e-01 -3.09862971e-01 2.25208271e-02 -5.38207889e-01
7.19937146e-01 5.36722720e-01 -5.51477909e-01 -7.45233148e-02
-3.84921044e-01 3.70899200e-01 5.45910653e-03 4.96052727e-02
8.88914466e-01 -7.45691478e-01 -1.77618533e-01 -3.81612442e-02
9.68357384e-01 6.06503561e-02 -1.37760043e+00 -1.62271187e-01
-9.97583941e-02 -5.12676120e-01 2.28713211e-02 -7.21664667e-01
-1.35625196e+00 4.95364755e-01 2.52905816e-01 9.07762527e-01
8.55464876e-01 -1.56480610e-01 7.98045516e-01 7.36397982e-01
7.54503369e-01 -7.04131484e-01 -2.84953207e-01 8.42802107e-01
5.76848626e-01 -7.65745699e-01 -1.44605383e-01 -7.78647363e-01
-1.19734839e-01 8.68048966e-01 5.93784273e-01 -4.94696379e-01
1.27134490e+00 8.06541383e-01 -4.22991246e-01 -7.43147954e-02
-7.98765063e-01 1.85193509e-01 4.52444330e-02 6.01343453e-01
-4.53209966e-01 -1.10356420e-01 -3.59533429e-01 8.19349527e-01
4.91956770e-02 -4.55116659e-01 7.84828424e-01 9.55647409e-01
-3.52998406e-01 -1.21517301e+00 -4.48903352e-01 2.30700538e-01
-6.17755532e-01 2.38398969e-01 -5.81483305e-01 1.14637971e+00
-4.36996579e-01 9.18451607e-01 -1.93797469e-01 -5.48288584e-01
3.23901117e-01 -5.34505606e-01 2.01990437e-02 -2.80205905e-01
-2.54295826e-01 -1.61746144e-01 2.87169386e-02 -4.80398029e-01
-1.92979291e-01 -4.44970816e-01 -1.13296843e+00 -9.86294448e-01
-7.85512388e-01 6.26861691e-01 4.03119117e-01 9.78998125e-01
3.90556306e-01 3.12712401e-01 1.19464743e+00 -3.39986593e-01
-1.43554962e+00 -2.20557079e-01 -8.61528218e-01 -6.10854685e-01
8.81995976e-01 -8.21857512e-01 -1.06333482e+00 -5.53808749e-01] | [5.800440311431885, 7.15212345123291] |
738b62b2-8f87-4725-86ba-c4e06ed18a41 | recovering-surveillance-video-using-rf-cues | 2212.13340 | null | https://arxiv.org/abs/2212.13340v1 | https://arxiv.org/pdf/2212.13340v1.pdf | Recovering Surveillance Video Using RF Cues | Video capture is the most extensively utilized human perception source due to its intuitively understandable nature. A desired video capture often requires multiple environmental conditions such as ample ambient-light, unobstructed space, and proper camera angle. In contrast, wireless measurements are more ubiquitous and have fewer environmental constraints. In this paper, we propose CSI2Video, a novel cross-modal method that leverages only WiFi signals from commercial devices and a source of human identity information to recover fine-grained surveillance video in a real-time manner. Specifically, two tailored deep neural networks are designed to conduct cross-modal mapping and video generation tasks respectively. We make use of an auto-encoder-based structure to extract pose features from WiFi frames. Afterward, both extracted pose features and identity information are merged to generate synthetic surveillance video. Our solution generates realistic surveillance videos without any expensive wireless equipment and has ubiquitous, cheap, and real-time characteristics. | ['Rabih Younes', 'Xiang Li'] | 2022-12-27 | null | null | null | null | ['video-generation'] | ['computer-vision'] | [ 5.94553709e-01 -2.57719547e-01 3.09508648e-02 -3.89741451e-01
-7.40212262e-01 -8.78980994e-01 4.46848154e-01 -5.22683144e-01
-1.25700414e-01 7.11420238e-01 2.32595518e-01 -8.04100856e-02
2.24894620e-02 -7.55701602e-01 -1.06506610e+00 -6.03967845e-01
1.98665962e-01 -3.35378647e-01 -8.75381976e-02 3.82203870e-02
-2.19156131e-01 1.81440964e-01 -1.69865525e+00 -6.91822469e-02
6.53048694e-01 1.43252158e+00 1.97803766e-01 8.28308761e-01
4.81985480e-01 6.65246248e-01 -4.97620195e-01 -3.50029230e-01
6.24858320e-01 -8.49168599e-02 2.00715363e-02 1.88722402e-01
7.01085329e-01 -9.47909653e-01 -6.47135437e-01 9.42392647e-01
5.39830267e-01 -8.55157301e-02 6.06301129e-02 -1.28596306e+00
-4.19033945e-01 2.36836866e-01 -5.73135257e-01 -2.20327184e-01
1.09512269e+00 2.94206500e-01 2.54456043e-01 -5.30661047e-01
4.32853490e-01 8.21889341e-01 9.05554652e-01 5.23421824e-01
-6.00164890e-01 -9.71355796e-01 -2.92629041e-02 -2.08312348e-01
-1.37195921e+00 -7.13556826e-01 8.74593258e-01 -3.00246835e-01
2.65194118e-01 3.15045536e-01 8.08587432e-01 1.94940925e+00
1.00534663e-01 3.93143475e-01 8.34163189e-01 -5.06973825e-02
4.95951325e-02 1.96497496e-02 -4.91569966e-01 7.24799275e-01
4.88602966e-01 2.08717018e-01 -6.35962188e-01 -2.15217754e-01
1.01660752e+00 6.84381366e-01 -8.63413572e-01 -3.44122231e-01
-1.52240503e+00 2.96635479e-01 1.28227472e-01 -1.38662234e-01
-6.40066266e-01 2.52745688e-01 2.15382814e-01 1.84117660e-01
-4.15811501e-02 3.55452478e-01 -1.84062794e-01 -4.41971183e-01
-9.32909608e-01 -8.53036866e-02 6.01863325e-01 1.27830696e+00
5.52246332e-01 1.24516442e-01 3.16080870e-03 3.61689776e-01
2.88152575e-01 1.15262508e+00 1.87214687e-01 -1.29111826e+00
5.74606657e-01 1.42639175e-01 4.45893019e-01 -1.56897879e+00
-1.38664544e-01 -5.25824666e-01 -7.86308348e-01 -2.15103492e-01
2.03665853e-01 -7.08322823e-01 -5.48350990e-01 1.95008039e+00
4.48406607e-01 7.43559659e-01 -4.58314531e-02 1.09773946e+00
6.25725925e-01 5.26930451e-01 -2.96365172e-01 -1.51458353e-01
1.27821696e+00 -6.41751409e-01 -5.29883087e-01 -3.55007350e-02
-1.05151907e-01 -6.19448721e-01 7.28063703e-01 4.67607617e-01
-7.35114515e-01 -7.44372487e-01 -1.07339001e+00 3.39414984e-01
1.51873250e-02 2.59067744e-01 6.16055906e-01 8.87087524e-01
-7.50523806e-01 4.25027758e-02 -7.27450252e-01 -4.13609147e-01
1.82465106e-01 3.03122133e-01 -6.82423770e-01 -2.48503059e-01
-1.14776337e+00 1.36303231e-01 3.61534730e-02 3.20420623e-01
-1.21195161e+00 -5.00571430e-01 -9.70901728e-01 -1.51696861e-01
6.62428856e-01 -9.65378046e-01 9.79527056e-01 -1.07977283e+00
-1.66746724e+00 4.19066668e-01 9.39221680e-03 -1.57093301e-01
5.07270873e-01 -5.25003076e-01 -5.98434269e-01 4.36274946e-01
1.48608401e-01 4.24745828e-01 1.23285580e+00 -1.22471440e+00
-6.55860364e-01 -3.71395528e-01 4.60057020e-01 7.05846399e-02
-5.48741639e-01 -3.69628251e-01 -4.47823822e-01 -6.41491532e-01
-1.59051791e-01 -1.01290607e+00 8.22068527e-02 1.43862650e-01
-4.48194563e-01 6.00060463e-01 7.98290730e-01 -7.73563027e-01
9.72125590e-01 -2.14710212e+00 -1.56776696e-01 1.20496318e-01
4.43404704e-01 7.56958723e-02 -4.19022050e-03 1.65949464e-01
2.61642307e-01 -3.09420168e-01 6.21313117e-02 -9.36408192e-02
-6.65890798e-02 -2.53714919e-01 -2.69303828e-01 4.34661776e-01
-2.60786712e-02 4.37588960e-01 -1.09840333e+00 -3.05682570e-01
3.73952478e-01 6.43800914e-01 -6.43544137e-01 4.53372777e-01
8.74659140e-03 6.91189885e-01 -6.42435074e-01 8.98721874e-01
9.06768501e-01 -8.16975906e-02 1.90070823e-01 -5.88038206e-01
-1.81081504e-01 -2.01827243e-01 -1.05608296e+00 2.03213763e+00
-6.39445066e-01 5.58640599e-01 4.87888157e-01 -6.27699375e-01
8.46039116e-01 5.59382975e-01 6.00580275e-01 -3.97871107e-01
3.94342035e-01 3.02457940e-02 -6.10928655e-01 -8.49629343e-01
4.38055843e-01 3.91146511e-01 -3.81573677e-01 1.46764085e-01
-1.08659580e-01 3.99653763e-01 -2.59375364e-01 6.30580336e-02
1.44891691e+00 4.64724511e-01 2.74065137e-01 5.84106967e-02
3.81940097e-01 -2.63007909e-01 8.01432788e-01 7.76853204e-01
-1.40251413e-01 7.80064642e-01 -2.08404258e-01 -4.66793567e-01
-8.88658702e-01 -1.15657163e+00 4.46015775e-01 8.17074299e-01
4.55610484e-01 -2.34422833e-01 -9.19851661e-01 -4.69293535e-01
-2.23187566e-01 1.48826599e-01 -5.54563224e-01 -1.30272165e-01
-6.16562128e-01 -1.42654508e-01 7.82475710e-01 3.63373786e-01
7.46127307e-01 -4.94770765e-01 -8.69314671e-01 5.72785437e-02
-4.11494344e-01 -1.35044873e+00 -8.06524515e-01 -5.23363650e-01
-1.79276839e-01 -1.29382002e+00 -7.59846091e-01 -6.31022811e-01
5.47143996e-01 7.99840450e-01 8.23204100e-01 -2.87783116e-01
-1.99881181e-01 5.92725337e-01 -2.52748400e-01 -3.99863645e-02
1.26375034e-01 -2.87934858e-02 4.49164510e-01 5.62476516e-01
3.55720282e-01 -6.75486028e-01 -9.10786808e-01 2.36409023e-01
-8.17274630e-01 1.60612643e-01 5.94367802e-01 6.90685928e-01
2.30113149e-01 -1.08020179e-01 3.73583347e-01 -3.96275461e-01
3.19945067e-01 -6.98310792e-01 -6.80908620e-01 2.43723586e-01
2.55128741e-01 -4.15035635e-01 9.80578363e-01 -4.67433959e-01
-1.18145037e+00 2.25213081e-01 2.26099685e-01 -8.51339161e-01
-4.68646735e-01 9.02370661e-02 -6.66659057e-01 -1.96050882e-01
4.59045947e-01 3.71926129e-01 -1.68458045e-01 -1.30050763e-01
7.44779781e-02 1.06638110e+00 9.11748648e-01 -5.18963218e-01
9.73268807e-01 6.22075975e-01 -2.08767489e-01 -9.02927279e-01
-6.46606624e-01 -2.05748558e-01 -9.03084949e-02 -5.29422045e-01
8.74322057e-01 -1.24424577e+00 -1.12100840e+00 5.94280243e-01
-1.17923164e+00 1.76736027e-01 3.63882512e-01 8.13804090e-01
-2.61584520e-01 3.12613875e-01 -3.47888201e-01 -7.48911560e-01
-3.25692207e-01 -1.24290395e+00 1.40835214e+00 4.64004189e-01
3.75021771e-02 -5.17956913e-01 -5.79433851e-02 5.65207183e-01
4.43054914e-01 7.81820655e-01 -1.38499707e-01 2.18530357e-01
-9.22074437e-01 -2.20397115e-01 -3.02493095e-01 -3.07314079e-02
5.21746278e-01 -4.46500517e-02 -1.12954974e+00 -4.10449713e-01
2.04388872e-02 -2.03533277e-01 2.85866261e-01 3.46832007e-01
1.13045228e+00 -4.52850431e-01 -4.96960282e-01 1.20646524e+00
1.26341486e+00 3.14709246e-01 2.83330113e-01 5.91252334e-02
1.02081132e+00 4.39251006e-01 6.10282063e-01 7.81909704e-01
5.30362785e-01 6.02467716e-01 5.61216474e-01 -2.00013004e-04
2.43973762e-01 -4.41944271e-01 4.75942612e-01 7.54958034e-01
-1.46928340e-01 -4.45721656e-01 -5.44259071e-01 3.16087484e-01
-1.57336974e+00 -1.28173590e+00 3.44755232e-01 2.38182211e+00
4.42002445e-01 -1.14995152e-01 -4.61602882e-02 -2.36048847e-01
9.98375297e-01 7.90348276e-02 -5.48276186e-01 3.20173264e-01
-4.25783172e-02 -2.43526220e-01 8.43265891e-01 1.50223091e-01
-1.23706114e+00 3.85815024e-01 5.46618319e+00 5.28575122e-01
-1.23684168e+00 -9.83549654e-02 2.97574192e-01 -4.01390046e-01
-4.55872059e-01 -3.01907301e-01 -3.23974878e-01 7.08028913e-01
7.32528210e-01 4.06939864e-01 5.92540979e-01 8.55447412e-01
3.07301342e-01 2.18636304e-01 -9.53806162e-01 1.50144494e+00
1.48319811e-01 -1.36206448e+00 -2.21685231e-01 -4.55215760e-02
5.92290580e-01 -1.67431757e-01 1.34892985e-01 -2.61127856e-02
6.16967231e-02 -6.62429214e-01 8.75668645e-01 7.93683589e-01
1.26010823e+00 -6.47783577e-01 5.63298047e-01 1.47426650e-01
-1.51293111e+00 -1.62838668e-01 -1.78361416e-01 -7.68565089e-02
3.52883756e-01 5.32045722e-01 -4.11389709e-01 6.98665082e-01
8.55822027e-01 7.42473960e-01 -2.61669993e-01 8.72114897e-01
9.90799814e-02 5.01497865e-01 -4.27258074e-01 1.26673684e-01
1.45332724e-01 -9.57571864e-02 6.36023104e-01 1.02887273e+00
8.01397264e-01 1.74050238e-02 5.62869430e-01 4.50976908e-01
-1.21733539e-01 -5.34587145e-01 -1.10369384e+00 1.82721511e-01
1.00898051e+00 1.20916903e+00 -2.05478251e-01 -1.12620771e-01
-6.65014207e-01 1.08553886e+00 -2.49485686e-01 3.05168509e-01
-1.34013104e+00 -6.45453334e-01 9.18156028e-01 1.15245372e-01
2.33712614e-01 -1.92230567e-01 2.83484757e-01 -1.59715796e+00
3.18725735e-01 -1.09666085e+00 -1.30765811e-01 -7.06376553e-01
-1.12712061e+00 8.05206835e-01 -1.96984366e-01 -1.63005829e+00
-4.54341114e-01 -4.41006780e-01 -2.93415874e-01 5.45760155e-01
-1.15552449e+00 -1.41624844e+00 -1.09660172e+00 9.25292909e-01
2.81101823e-01 -2.98648685e-01 7.52627254e-01 6.65878594e-01
-6.21326268e-01 6.98774695e-01 2.60259509e-02 3.88019741e-01
7.13896155e-01 -6.88498855e-01 3.22378844e-01 1.08734190e+00
-3.57171923e-01 8.38292122e-01 5.46585977e-01 -5.83313942e-01
-2.05900955e+00 -1.16722214e+00 9.99822468e-02 -3.59247595e-01
3.52339804e-01 -5.89506626e-01 -2.32021272e-01 5.64703465e-01
6.95895106e-02 1.28288046e-01 6.40896261e-01 -5.15082717e-01
-3.80904347e-01 -5.57074189e-01 -1.20324755e+00 5.46256959e-01
1.37620902e+00 -6.56855941e-01 -2.16364980e-01 -7.53494054e-02
9.38712239e-01 -3.42473119e-01 -8.33677113e-01 3.08690429e-01
1.13854146e+00 -1.20474041e+00 9.15152729e-01 1.44823249e-02
1.26495585e-01 -6.50502026e-01 -4.60313410e-01 -8.43994856e-01
-1.49033889e-01 -1.24805903e+00 5.05403019e-02 1.21314728e+00
4.10588421e-02 -7.65150368e-01 7.93623269e-01 4.51787889e-01
-5.18480949e-02 -3.98842126e-01 -5.78533769e-01 -5.43384850e-01
-1.12975240e+00 -4.66838777e-01 1.06522214e+00 9.56925511e-01
-3.25884342e-01 5.94705231e-02 -1.10807824e+00 5.71356058e-01
8.86933982e-01 2.24178135e-01 1.10740089e+00 -1.05906284e+00
-5.05521119e-01 3.98084730e-01 -3.87649268e-01 -1.30526900e+00
-1.16430975e-01 -4.70179804e-02 1.51428431e-01 -9.48360145e-01
-2.85434667e-02 -4.81184214e-01 -5.21544693e-03 -4.19174545e-02
1.15860693e-01 4.49613392e-01 4.01427448e-02 2.47769225e-02
-6.96451962e-01 3.59023809e-01 9.04547155e-01 -1.05621004e-02
9.12710801e-02 -5.48103414e-02 -6.41376019e-01 6.77405953e-01
7.41795123e-01 -1.64519355e-01 -7.33770072e-01 -9.92734969e-01
1.93859562e-01 5.74974000e-01 6.05390012e-01 -1.41891015e+00
2.71586090e-01 -1.52558357e-01 5.49036920e-01 -1.30224496e-01
6.26514494e-01 -1.16763306e+00 6.43958151e-01 2.01839164e-01
1.03495289e-02 2.59914190e-01 -8.79068822e-02 8.03638458e-01
-5.24645709e-02 3.08079720e-01 2.69977182e-01 -1.37871459e-01
-4.39718425e-01 5.95258355e-01 -2.16993183e-01 -1.54908910e-01
1.11947966e+00 -5.99571049e-01 -2.84262270e-01 -6.71379209e-01
-1.73777387e-01 -1.26971111e-01 8.88517737e-01 5.91794193e-01
6.53979659e-01 -1.40286398e+00 -4.28444028e-01 4.28725541e-01
2.70916875e-02 -9.49238390e-02 4.90543991e-01 5.68020761e-01
-5.72260499e-01 4.51565683e-01 -3.64003092e-01 -6.05776846e-01
-9.77647245e-01 6.38223171e-01 5.10176308e-02 3.81662279e-01
-3.83219481e-01 6.33185565e-01 2.48329803e-01 -3.61493260e-01
1.13747351e-01 -1.53701961e-01 1.15667082e-01 -3.31267387e-01
9.00925219e-01 3.13372344e-01 -2.84301102e-01 -7.15304911e-01
-4.81373340e-01 7.19879985e-01 4.49008346e-01 -3.98350023e-02
1.07393110e+00 -4.23844755e-01 2.85161018e-01 -2.10564900e-02
1.21850431e+00 3.57125580e-01 -1.79299390e+00 2.81424820e-02
-5.78677773e-01 -7.75309265e-01 -1.49564788e-01 -3.56926113e-01
-1.24072683e+00 5.43938220e-01 6.07932210e-01 1.31784484e-01
1.43532920e+00 -5.41328669e-01 1.18366301e+00 3.41544956e-01
9.40742314e-01 -7.34593451e-01 -1.83102354e-01 -3.93445119e-02
3.51352036e-01 -1.21671855e+00 -4.42950130e-01 -4.27153140e-01
-2.83879608e-01 1.03834724e+00 6.57339215e-01 -3.18064019e-02
3.65765274e-01 4.26268309e-01 6.02590069e-02 7.30868876e-02
-4.48059887e-01 1.91493750e-01 -4.54381406e-02 1.13136578e+00
1.84801415e-01 2.69062631e-02 5.04520059e-01 6.57920897e-01
-2.63928264e-01 2.01618329e-01 3.44241470e-01 8.36306632e-01
-1.52409866e-01 -4.84601498e-01 -7.15716183e-01 2.70249456e-01
-5.00522375e-01 7.29233772e-02 6.19466864e-02 4.03219074e-01
4.72473264e-01 1.16866696e+00 -5.57478741e-02 -8.26652467e-01
5.90562671e-02 -6.68136537e-01 4.25368130e-01 -1.16691425e-01
-3.17700654e-01 -1.14703640e-01 1.60448387e-01 -9.86486256e-01
-5.23753107e-01 -4.39946115e-01 -6.06937289e-01 -4.91560429e-01
-1.98319569e-01 6.03588633e-02 5.68040371e-01 7.79945970e-01
7.80223608e-01 4.06135768e-01 7.37834454e-01 -1.09594953e+00
-1.77587405e-01 -6.61449432e-01 -2.16603845e-01 3.79000753e-01
6.41464770e-01 -7.31490195e-01 -3.65429223e-02 2.14056715e-01] | [10.6797456741333, -1.4250322580337524] |
d1b178f8-d548-4377-8feb-e8427ad6e237 | handsoff-labeled-dataset-generation-with-no | 2212.12645 | null | https://arxiv.org/abs/2212.12645v2 | https://arxiv.org/pdf/2212.12645v2.pdf | HandsOff: Labeled Dataset Generation With No Additional Human Annotations | Recent work leverages the expressive power of generative adversarial networks (GANs) to generate labeled synthetic datasets. These dataset generation methods often require new annotations of synthetic images, which forces practitioners to seek out annotators, curate a set of synthetic images, and ensure the quality of generated labels. We introduce the HandsOff framework, a technique capable of producing an unlimited number of synthetic images and corresponding labels after being trained on less than 50 pre-existing labeled images. Our framework avoids the practical drawbacks of prior work by unifying the field of GAN inversion with dataset generation. We generate datasets with rich pixel-wise labels in multiple challenging domains such as faces, cars, full-body human poses, and urban driving scenes. Our method achieves state-of-the-art performance in semantic segmentation, keypoint detection, and depth estimation compared to prior dataset generation approaches and transfer learning baselines. We additionally showcase its ability to address broad challenges in model development which stem from fixed, hand-annotated datasets, such as the long-tail problem in semantic segmentation. Project page: austinxu87.github.io/handsoff. | ['Arjun Seshadri', 'Achal Dave', 'Mariya I. Vasileva', 'Austin Xu'] | 2022-12-24 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xu_HandsOff_Labeled_Dataset_Generation_With_No_Additional_Human_Annotations_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xu_HandsOff_Labeled_Dataset_Generation_With_No_Additional_Human_Annotations_CVPR_2023_paper.pdf | cvpr-2023-1 | ['keypoint-detection'] | ['computer-vision'] | [ 7.45290637e-01 4.98439580e-01 -2.85779964e-02 -4.80767846e-01
-1.29324436e+00 -8.86404455e-01 6.88552260e-01 -6.32900894e-01
-2.91023374e-01 8.93331170e-01 -1.21631168e-01 -2.01789916e-01
6.27577305e-01 -7.60127902e-01 -1.02308166e+00 -3.56233150e-01
5.05835712e-01 8.45211446e-01 1.28481448e-01 -1.49392396e-01
-5.06996061e-04 2.52066761e-01 -1.49628949e+00 2.59704888e-01
7.74888873e-01 8.71950686e-01 -4.88739870e-02 7.54387200e-01
1.87982470e-01 7.20911443e-01 -6.96421683e-01 -6.74429476e-01
5.20304799e-01 -6.88393474e-01 -7.79909670e-01 3.98451507e-01
7.73217916e-01 -5.68117499e-01 -1.56961188e-01 8.73593628e-01
5.30716896e-01 2.01540869e-02 7.02219129e-01 -1.76512849e+00
-5.27398705e-01 3.12097102e-01 -6.80488348e-01 -1.98750019e-01
2.12146401e-01 6.19262516e-01 6.93789423e-01 -8.26752901e-01
9.71809924e-01 1.17329419e+00 7.57830203e-01 1.07456696e+00
-1.26609218e+00 -9.45062459e-01 -8.69459435e-02 -3.87013048e-01
-1.10179043e+00 -6.33474529e-01 6.23547018e-01 -7.00025618e-01
2.09438279e-01 -1.25859436e-02 5.44827342e-01 1.66708517e+00
-4.57202077e-01 9.80081081e-01 1.17607713e+00 -2.85304993e-01
2.34417632e-01 -1.13201365e-01 -5.18741071e-01 6.79521322e-01
1.16015635e-01 2.75636017e-01 -5.11626899e-01 1.25837758e-01
9.40731406e-01 -4.22048122e-01 -9.04107839e-02 -5.59959948e-01
-1.41189063e+00 8.48405421e-01 2.57944077e-01 -4.27786946e-01
-2.06339777e-01 5.26098371e-01 2.74103880e-01 -6.60776272e-02
3.89181495e-01 6.58428192e-01 -4.20507193e-01 2.25938950e-02
-1.04618013e+00 6.61616623e-01 5.18106282e-01 1.29542160e+00
6.95339441e-01 2.19573662e-01 -2.78047055e-01 6.05103970e-01
2.88546141e-02 7.49217331e-01 2.02274740e-01 -1.30542862e+00
4.93617356e-01 2.90971577e-01 3.02847773e-01 -3.88648242e-01
6.42206669e-02 -1.67751759e-01 -4.93481129e-01 4.12900865e-01
6.72831178e-01 -3.84351701e-01 -1.46981585e+00 1.87408710e+00
6.23260736e-01 1.12275690e-01 7.37509802e-02 7.22122788e-01
8.70057106e-01 3.45495224e-01 2.23038226e-01 3.83965552e-01
1.06304407e+00 -1.31504393e+00 -3.54337811e-01 -5.12660921e-01
1.86604261e-01 -7.10055530e-01 1.30412865e+00 3.11473221e-01
-1.22861934e+00 -5.89937091e-01 -7.96720326e-01 -2.35878095e-01
-2.52626806e-01 -2.92721167e-02 6.70174599e-01 7.08529472e-01
-1.04203176e+00 3.92760903e-01 -7.69996822e-01 6.57299254e-03
9.52318311e-01 1.09268941e-01 -2.41222546e-01 -2.87816972e-01
-9.33841944e-01 5.25188208e-01 2.87515938e-01 -1.18135244e-01
-1.43788707e+00 -9.02475595e-01 -1.00182664e+00 -4.58318800e-01
3.75133246e-01 -9.66771066e-01 1.64521527e+00 -1.22548163e+00
-1.27550948e+00 1.19499278e+00 6.73645362e-02 -2.56867260e-01
1.05256462e+00 -1.69898883e-01 -1.83527786e-02 1.68694466e-01
5.21970809e-01 1.58216202e+00 9.71216381e-01 -1.37552464e+00
-4.76745665e-01 -6.20702058e-02 -3.49256806e-02 8.89229700e-02
2.11443827e-01 -1.99414536e-01 -5.80990613e-01 -8.32681000e-01
-2.82222122e-01 -1.19972730e+00 -4.33080316e-01 1.33385062e-01
-7.08080232e-01 2.12650478e-01 8.10769618e-01 -5.04063487e-01
3.19466531e-01 -1.98822188e+00 -3.51003651e-03 -6.85392618e-02
9.74487886e-02 1.63438439e-01 -3.33416879e-01 8.19031224e-02
4.75709029e-02 2.04144895e-01 -4.62347716e-01 -6.13444805e-01
8.84548500e-02 1.06978245e-01 -4.14886087e-01 1.60069123e-01
2.78849304e-01 1.37568879e+00 -1.00460768e+00 -7.01580763e-01
1.96480587e-01 4.35607702e-01 -5.72489679e-01 2.77409732e-01
-7.20980287e-01 1.08157527e+00 -3.61553431e-01 9.46351349e-01
5.33518195e-01 -3.49263251e-01 -2.31783047e-01 -5.37477396e-02
4.24644768e-01 7.91218728e-02 -8.28408539e-01 1.98305571e+00
-3.57355475e-01 5.89334488e-01 1.43085914e-02 -4.55798239e-01
6.01076722e-01 1.24369383e-01 3.15121204e-01 -4.73136604e-01
1.64695457e-01 1.52673095e-01 -2.99116135e-01 -2.08693430e-01
4.32005793e-01 -1.25654325e-01 -2.68738329e-01 5.73957562e-01
9.54203829e-02 -7.97244787e-01 3.26283008e-01 2.53036499e-01
9.48144674e-01 6.47199869e-01 -9.81604606e-02 6.72529414e-02
-6.30529597e-02 2.04983547e-01 4.51333404e-01 6.49175048e-01
-2.13884890e-01 1.23175800e+00 4.99502659e-01 -3.80024344e-01
-1.44250524e+00 -1.16901505e+00 1.66977744e-03 9.99675095e-01
2.50076335e-02 8.49801674e-03 -1.13672876e+00 -1.05863190e+00
-2.04657149e-02 6.69001698e-01 -8.60921025e-01 8.83494988e-02
-4.27204490e-01 -5.53653955e-01 8.69882703e-01 6.77404344e-01
6.18490994e-01 -1.40867555e+00 -6.77766085e-01 -1.16000146e-01
-3.99019659e-01 -1.45224392e+00 -5.45795143e-01 -1.26109913e-01
-5.65018296e-01 -1.13473320e+00 -9.04844999e-01 -6.40351653e-01
8.52779329e-01 -1.48328379e-01 1.57381582e+00 -1.79892302e-01
-4.25325394e-01 3.29364002e-01 -2.87355870e-01 -6.37135565e-01
-6.95827365e-01 2.06030160e-01 -4.09310281e-01 -1.42366603e-01
-5.17166182e-02 -3.80400479e-01 -7.62213349e-01 4.00529474e-01
-1.06118584e+00 4.57188129e-01 4.72679973e-01 8.16638947e-01
9.33639348e-01 -4.86024648e-01 7.09443331e-01 -1.35229301e+00
2.85117507e-01 -2.73623556e-01 -6.11254036e-01 8.22969973e-02
-4.10783559e-01 -3.65306437e-02 3.44951153e-01 -4.23517197e-01
-1.02750874e+00 4.13077593e-01 -1.23963043e-01 -5.18513441e-01
-3.12583357e-01 -2.39970297e-01 -3.06320220e-01 -1.01565704e-01
9.13104475e-01 1.74302328e-02 4.01020795e-02 -6.31861761e-02
7.89538920e-01 3.32042158e-01 9.37256396e-01 -7.47681677e-01
9.60178912e-01 6.02839887e-01 -6.66090846e-02 -3.13995123e-01
-9.70894694e-01 3.88139896e-02 -6.47161305e-01 -1.24611460e-01
1.01410973e+00 -1.13217545e+00 -1.95502475e-01 6.73521519e-01
-1.03138900e+00 -9.69938457e-01 -6.60183966e-01 -1.45300254e-02
-9.42016721e-01 -3.85414474e-02 -6.05482519e-01 -3.78206521e-01
-4.11036313e-01 -1.25745308e+00 1.66797400e+00 9.89780650e-02
-4.29360777e-01 -7.44801462e-01 -1.98328793e-01 8.23247194e-01
2.79743105e-01 9.86999750e-01 5.11832178e-01 -2.05543801e-01
-8.66231501e-01 -1.83263496e-01 -1.30888090e-01 4.19440657e-01
-1.10719860e-01 -9.07228515e-02 -1.18469560e+00 -1.72919378e-01
-5.41046917e-01 -9.33524251e-01 6.99505210e-01 2.75537401e-01
1.34422338e+00 -1.21671535e-01 -3.20550472e-01 8.71458769e-01
9.97614324e-01 9.68009681e-02 7.36789465e-01 1.66617468e-01
1.01676965e+00 6.26124382e-01 6.25699222e-01 1.94990888e-01
3.43646288e-01 5.01436651e-01 4.39164788e-01 -4.98919070e-01
-5.14869809e-01 -6.84789836e-01 -4.75272834e-02 -1.16478764e-02
2.05119908e-01 -4.60142314e-01 -9.92331982e-01 7.37862945e-01
-1.44180751e+00 -7.74505675e-01 1.02848314e-01 1.89808857e+00
1.01261973e+00 1.37263328e-01 3.08616519e-01 -1.22593455e-01
7.05690980e-01 5.50966635e-02 -9.30161774e-01 5.17252982e-02
1.62936077e-02 4.18404609e-01 6.44318700e-01 2.80560553e-01
-1.13096583e+00 1.25274444e+00 6.58323383e+00 6.12508297e-01
-9.87659216e-01 2.56694824e-01 1.19484687e+00 -1.68891042e-01
-5.59996128e-01 -1.66394129e-01 -6.67041063e-01 5.66053569e-01
6.42365456e-01 1.75525054e-01 3.77166629e-01 1.07030344e+00
-8.43007788e-02 -6.80858418e-02 -1.21371269e+00 1.04313529e+00
1.49327442e-01 -1.35410595e+00 4.38479371e-02 1.21571906e-01
1.34453976e+00 9.81144458e-02 4.20363069e-01 1.86359107e-01
8.50284994e-01 -1.36718607e+00 9.72222686e-01 2.64637738e-01
1.50691640e+00 -5.73056757e-01 4.05701876e-01 1.37610525e-01
-7.32810140e-01 2.88417250e-01 1.44733995e-01 3.17266673e-01
4.88394797e-01 4.49331313e-01 -1.04257226e+00 1.43412963e-01
5.26927948e-01 4.67561573e-01 -6.79180324e-01 6.10217750e-01
-5.69635451e-01 5.09246349e-01 -2.81735361e-01 4.58452553e-01
2.07001045e-01 2.43529119e-02 6.69307113e-02 9.34054494e-01
1.60248667e-01 -5.71947545e-02 3.10422350e-02 1.17897141e+00
-3.66409898e-01 -3.54297072e-01 -6.92637920e-01 -1.67266145e-01
5.32044947e-01 1.16133082e+00 -9.35339987e-01 -3.88662457e-01
-1.04124196e-01 1.03759384e+00 -2.95873396e-02 3.79401267e-01
-1.02988696e+00 -2.33848080e-01 4.34331119e-01 4.28684503e-01
3.26147705e-01 2.00316850e-02 -6.35162473e-01 -1.00409079e+00
-1.97305623e-03 -1.24452507e+00 1.91624582e-01 -1.03631318e+00
-1.13614702e+00 6.87913895e-01 9.91628394e-02 -1.11190879e+00
-6.72566473e-01 -3.87169480e-01 -4.55720395e-01 7.59676576e-01
-1.21805763e+00 -1.59186304e+00 -7.66395688e-01 5.46208084e-01
7.22566068e-01 -1.17639631e-01 6.30419195e-01 3.27981651e-01
-3.28131288e-01 7.03247488e-01 -3.50946277e-01 2.57778794e-01
7.66503930e-01 -1.23865712e+00 1.09823895e+00 8.67807865e-01
6.06479235e-02 1.34712473e-01 6.14399850e-01 -6.91515267e-01
-8.72759342e-01 -1.36240768e+00 3.66564065e-01 -9.90192473e-01
1.94112241e-01 -6.40212476e-01 -3.32691938e-01 9.18234527e-01
2.31468789e-02 3.20826471e-01 3.31667453e-01 -3.97087276e-01
-3.42943609e-01 1.55858040e-01 -1.32050455e+00 5.98551393e-01
1.30989647e+00 -3.72623593e-01 -1.04662478e-01 5.26660264e-01
7.76062667e-01 -8.44496548e-01 -4.06091690e-01 6.82974398e-01
5.67244232e-01 -8.82912636e-01 1.07020915e+00 -4.74456578e-01
9.29871798e-01 -3.27084631e-01 8.17724764e-02 -1.21648967e+00
2.81167001e-01 -7.16198385e-01 1.24271952e-01 1.21828258e+00
5.09926200e-01 -3.26334178e-01 1.24789476e+00 6.74419463e-01
-1.58848688e-01 -6.00758076e-01 -4.65634376e-01 -5.44800818e-01
1.08617678e-01 -3.08129877e-01 7.96146870e-01 9.20144856e-01
-7.82359004e-01 3.71505022e-01 -4.01559860e-01 -2.94737518e-01
8.46199453e-01 -1.14800110e-02 1.46030164e+00 -8.34101975e-01
-3.59512597e-01 -1.17811389e-01 -2.03875825e-01 -9.57927108e-01
2.86509037e-01 -6.94512784e-01 2.15923816e-01 -1.51293099e+00
-4.97282669e-02 -7.60359406e-01 2.66455650e-01 6.44278407e-01
-1.43701926e-01 8.82778227e-01 1.09294973e-01 1.49537370e-01
-6.50054157e-01 4.20576066e-01 1.64422941e+00 -9.15330425e-02
6.09505512e-02 -1.11605473e-01 -8.04751575e-01 7.44210005e-01
7.37981796e-01 -4.88710880e-01 -7.77783036e-01 -6.18481994e-01
5.41717634e-02 4.52536866e-02 7.22484887e-01 -1.09515774e+00
-2.27320373e-01 -2.64223039e-01 6.16142213e-01 -4.69044536e-01
4.74886358e-01 -3.66745472e-01 4.20475990e-01 1.04094110e-01
-4.12963301e-01 3.51547562e-02 1.90764040e-01 5.05384684e-01
-1.28228515e-01 -1.64910627e-03 1.02497470e+00 -5.09554029e-01
-7.22267866e-01 4.46668297e-01 1.62772149e-01 6.77611113e-01
1.18971682e+00 -2.72220790e-01 -3.33682358e-01 -5.40103316e-01
-5.16617179e-01 2.63838798e-01 8.38601410e-01 5.01133680e-01
3.59808534e-01 -1.18297207e+00 -6.64162397e-01 2.29751587e-01
2.25529730e-01 6.85578048e-01 2.10328817e-01 3.24826658e-01
-7.47250557e-01 1.62584260e-02 -3.42872053e-01 -5.73127508e-01
-8.88710141e-01 3.85423988e-01 3.31539989e-01 -8.03796872e-02
-4.98886615e-01 1.13670421e+00 5.36068082e-01 -6.48107350e-01
1.55629544e-02 -6.79817945e-02 4.51316625e-01 -2.02238113e-01
2.42056981e-01 2.30773404e-01 -7.47868270e-02 -5.85338771e-01
-9.19108018e-02 3.42527509e-01 7.32842311e-02 -4.33594614e-01
1.04716682e+00 2.09106684e-01 3.72225463e-01 5.72744533e-02
9.57998693e-01 -4.01220359e-02 -1.83911216e+00 1.13244601e-01
-4.55773860e-01 -4.90921825e-01 -3.50441098e-01 -1.10308981e+00
-1.19886577e+00 7.48500288e-01 4.30136204e-01 -4.17590618e-01
9.29516137e-01 1.53294981e-01 1.15514493e+00 -5.12938648e-02
5.24041474e-01 -9.73395526e-01 4.56192076e-01 1.77367613e-01
7.45028734e-01 -1.48631036e+00 -2.45801568e-01 -6.39985561e-01
-9.47679520e-01 5.65051079e-01 8.73005271e-01 -4.39335033e-02
4.80162464e-02 4.12471980e-01 3.44974756e-01 -1.37366697e-01
-4.01266515e-01 -1.71933532e-01 1.22185014e-01 9.13736582e-01
1.91220686e-01 1.65270284e-01 1.42107025e-01 2.28965819e-01
-6.37113214e-01 1.28318265e-01 5.09650290e-01 8.78624380e-01
1.10646188e-01 -1.24527657e+00 -2.78654456e-01 3.04924697e-01
-4.51995343e-01 -1.22499086e-01 -5.39714992e-01 9.51388717e-01
3.61071646e-01 7.06740201e-01 8.22473131e-03 -2.59076476e-01
1.75140440e-01 1.46438569e-01 4.74259704e-01 -7.64373600e-01
-2.07789749e-01 -1.51463494e-01 1.46493554e-01 -5.51613808e-01
-4.44147885e-01 -7.03195989e-01 -1.01311481e+00 -8.31320211e-02
-8.03733394e-02 -7.23244175e-02 7.04775989e-01 6.76560521e-01
4.43266302e-01 5.11131346e-01 2.47970372e-01 -9.86489832e-01
-3.09295863e-01 -9.04373825e-01 -1.39720172e-01 8.29907715e-01
2.27823198e-01 -7.26044893e-01 -1.42297015e-01 6.44140065e-01] | [11.407658576965332, -0.298392653465271] |
6c6056fb-d771-4e90-b6f5-28727899e25b | ssd-kd-a-self-supervised-diverse-knowledge | 2203.11490 | null | https://arxiv.org/abs/2203.11490v2 | https://arxiv.org/pdf/2203.11490v2.pdf | SSD-KD: A Self-supervised Diverse Knowledge Distillation Method for Lightweight Skin Lesion Classification Using Dermoscopic Images | Skin cancer is one of the most common types of malignancy, affecting a large population and causing a heavy economic burden worldwide. Over the last few years, computer-aided diagnosis has been rapidly developed and make great progress in healthcare and medical practices due to the advances in artificial intelligence. However, most studies in skin cancer detection keep pursuing high prediction accuracies without considering the limitation of computing resources on portable devices. In this case, knowledge distillation (KD) has been proven as an efficient tool to help improve the adaptability of lightweight models under limited resources, meanwhile keeping a high-level representation capability. To bridge the gap, this study specifically proposes a novel method, termed SSD-KD, that unifies diverse knowledge into a generic KD framework for skin diseases classification. Our method models an intra-instance relational feature representation and integrates it with existing KD research. A dual relational knowledge distillation architecture is self-supervisedly trained while the weighted softened outputs are also exploited to enable the student model to capture richer knowledge from the teacher model. To demonstrate the effectiveness of our method, we conduct experiments on ISIC 2019, a large-scale open-accessed benchmark of skin diseases dermoscopic images. Experiments show that our distilled lightweight model can achieve an accuracy as high as 85% for the classification tasks of 8 different skin diseases with minimal parameters and computing requirements. Ablation studies confirm the effectiveness of our intra- and inter-instance relational knowledge integration strategy. Compared with state-of-the-art knowledge distillation techniques, the proposed method demonstrates improved performances for multi-diseases classification on the large-scale dermoscopy database. | ['Z. Jane Wang', 'Chunyan Miao', 'Tim K. Lee', 'Yuheng Wang', 'Yongwei Wang'] | 2022-03-22 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 2.92936206e-01 3.72980461e-02 -6.98558211e-01 -1.14810802e-01
-6.16774678e-01 -1.14517935e-01 3.31949532e-01 2.96674877e-01
-3.95892560e-01 5.60631990e-01 -1.41729549e-01 -2.32672408e-01
-4.19013172e-01 -1.01202500e+00 -2.58087546e-01 -9.19766128e-01
3.56235474e-01 4.32707332e-02 4.08775330e-01 7.36639416e-03
-1.14616022e-01 4.32611674e-01 -1.61059630e+00 3.96206260e-01
1.43939972e+00 1.19929028e+00 1.27576143e-01 2.57058591e-01
-2.35177591e-01 7.33061314e-01 -3.29334736e-01 -7.24006951e-01
1.83785241e-02 -1.89372450e-01 -7.67247975e-01 -1.52941212e-01
3.43605250e-01 -1.08987138e-01 -3.41080219e-01 7.89363563e-01
7.27812946e-01 -3.59454244e-01 4.84804809e-01 -7.88003743e-01
-6.52085066e-01 2.45595455e-01 -5.92321515e-01 -7.15755597e-02
2.20481873e-01 -2.19790288e-03 6.12489343e-01 -4.77786392e-01
6.53434277e-01 6.79597437e-01 7.97118962e-01 7.98818231e-01
-8.35558832e-01 -5.90373695e-01 6.98800609e-02 4.15867388e-01
-1.44430387e+00 6.45917058e-02 5.26293993e-01 -1.79233074e-01
9.31202829e-01 4.21658993e-01 7.36216843e-01 9.87421632e-01
1.81694597e-01 8.96096647e-01 1.21970558e+00 -5.15379727e-01
8.82981420e-02 5.85057080e-01 1.04276903e-01 9.35705185e-01
4.37258661e-01 -1.65880129e-01 -4.98749256e-01 -1.19298130e-01
6.37698710e-01 3.17960769e-01 -3.12471449e-01 -1.98964834e-01
-7.59503722e-01 5.78693807e-01 7.01599598e-01 3.83541316e-01
-3.62692326e-01 -1.53312624e-01 3.59605819e-01 7.29853064e-02
5.43382287e-01 2.23397419e-01 -3.74416322e-01 -9.96289849e-02
-6.63216174e-01 -1.14932314e-01 8.20621789e-01 4.64492738e-01
3.00485998e-01 -3.00088346e-01 -2.00762600e-01 1.08958328e+00
-5.40023930e-02 4.57611650e-01 7.61759460e-01 -1.02420807e-01
1.80395529e-01 1.24916875e+00 -4.61579859e-01 -7.39905477e-01
-3.77263129e-01 -5.52957356e-01 -1.13293219e+00 -2.01718241e-01
1.13607936e-01 8.67091119e-02 -1.12263858e+00 1.38935483e+00
7.16754735e-01 3.80404741e-01 1.07731543e-01 7.40884006e-01
1.01327837e+00 9.55642108e-03 3.91016334e-01 -8.37891921e-03
1.55482459e+00 -9.19544995e-01 -6.48475766e-01 3.45063061e-01
7.26927519e-01 -4.99843657e-01 7.26211369e-01 5.85782647e-01
-6.57710254e-01 -3.57714057e-01 -9.33666408e-01 1.01202473e-04
-6.68310583e-01 3.97903949e-01 9.81722593e-01 8.92478526e-01
-7.21881866e-01 4.12141472e-01 -9.53725994e-01 -7.16245234e-01
7.35095620e-01 3.48589480e-01 -4.97361153e-01 -2.71843672e-01
-1.24246824e+00 1.09507811e+00 3.96141946e-01 -4.20077406e-02
-4.28264916e-01 -1.03431892e+00 -5.32692254e-01 -2.30088726e-01
6.16965711e-01 -7.81724930e-01 8.37341845e-01 -8.30214262e-01
-1.61347485e+00 8.39174867e-01 1.28110051e-01 -3.12530100e-01
4.06609148e-01 -2.82228708e-01 -7.34333694e-01 3.85112584e-01
-3.31259012e-01 2.98935443e-01 6.18883729e-01 -8.90451431e-01
-8.39472413e-01 -4.61277694e-01 8.72384608e-02 2.99805284e-01
-1.12628806e+00 -3.87541026e-01 -7.46155918e-01 -5.86634636e-01
-1.78162694e-01 -9.25192535e-01 -1.79920718e-01 4.15918887e-01
-4.40504819e-01 -2.72241682e-01 7.13291168e-01 -6.02988064e-01
1.55573916e+00 -2.06308365e+00 1.01942144e-01 4.13690716e-01
3.07054877e-01 8.57854366e-01 -3.81171592e-02 2.87359357e-01
1.18477875e-02 2.44068448e-02 -9.09103528e-02 -6.64070472e-02
-3.86378676e-01 2.49189556e-01 7.92682096e-02 2.11652979e-01
3.23530942e-01 7.82043934e-01 -8.50338697e-01 -5.88435531e-01
2.85631299e-01 7.78679848e-01 -3.16098928e-01 7.05066547e-02
-7.69502446e-02 -4.38833004e-03 -5.81317246e-01 1.03850710e+00
5.64697802e-01 -5.72149336e-01 3.87699246e-01 -4.53022808e-01
3.35681200e-01 -1.85652182e-01 -1.10435736e+00 1.69024849e+00
-5.70893645e-01 -4.42612432e-02 -1.96558565e-01 -7.70208418e-01
6.59816146e-01 3.39276165e-01 6.01902425e-01 -6.79449797e-01
-3.96547802e-02 3.37079704e-01 -7.20762461e-02 -8.25412214e-01
1.30932137e-01 -1.51641266e-02 3.04542840e-01 -1.46163451e-02
-1.30074332e-02 2.78882887e-02 9.19123664e-02 2.64117718e-01
1.22426820e+00 -2.39203889e-02 6.08535886e-01 5.83096668e-02
6.81456327e-01 6.76092356e-02 4.69078213e-01 1.53455272e-01
-7.68247172e-02 1.22186758e-01 1.87443227e-01 -2.86739498e-01
-2.70880252e-01 -1.13702190e+00 -4.91242498e-01 6.86923206e-01
2.50930220e-01 -4.38682795e-01 -5.94528735e-01 -9.84983265e-01
4.13699001e-01 1.81980744e-01 -7.98014641e-01 -2.56571919e-01
-5.14774099e-02 -1.07675588e+00 7.78198123e-01 7.15134919e-01
7.68916011e-01 -7.39241481e-01 -3.99844855e-01 -8.77843052e-02
1.34779006e-01 -1.05019379e+00 3.19194980e-02 2.29236460e-03
-7.57462859e-01 -1.39040983e+00 -8.20243061e-01 -7.92165935e-01
7.75555313e-01 3.06177199e-01 6.69439793e-01 2.07672969e-01
-1.06254125e+00 4.41317976e-01 -4.55401927e-01 -4.30597544e-01
-1.65172279e-01 3.26104909e-01 4.64218715e-03 1.51346281e-01
5.72926760e-01 -2.41875634e-01 -7.18570769e-01 1.44847915e-01
-1.18861079e+00 1.22663148e-01 1.05890477e+00 9.99627590e-01
8.21131766e-01 3.16958487e-01 7.16424227e-01 -1.19988608e+00
6.39572680e-01 -6.21101439e-01 -1.66824847e-01 6.53090179e-01
-8.61057580e-01 -9.33178961e-02 6.09517217e-01 -5.07132828e-01
-1.28859222e+00 1.00741535e-02 -6.38301596e-02 -1.66349515e-01
7.52802566e-02 7.65661299e-01 4.76056188e-02 -4.50897217e-01
6.58673525e-01 2.68631756e-01 2.81087905e-01 -3.94720614e-01
3.39334100e-01 8.30233514e-01 3.08644444e-01 -5.07695973e-01
6.72651887e-01 4.50980246e-01 1.49539068e-01 -8.11859131e-01
-8.41060221e-01 -6.30419552e-01 -3.68165523e-01 3.13722268e-02
5.91724634e-01 -1.05991149e+00 -6.91683829e-01 7.79825807e-01
-4.10588086e-01 -1.00096062e-01 -2.55605817e-01 3.66989464e-01
7.94088244e-02 4.64463681e-01 -6.02360964e-01 -6.20157838e-01
-6.66416407e-01 -9.12071943e-01 8.53235722e-01 5.89020193e-01
9.65437479e-03 -1.25766408e+00 7.83895329e-02 5.06679237e-01
5.38340926e-01 1.95098847e-01 1.02429354e+00 -5.89247823e-01
-2.81528860e-01 -3.59726787e-01 -3.73721480e-01 4.21873778e-01
5.93054831e-01 8.41213390e-02 -1.02873158e+00 -1.76713318e-01
-4.90654707e-01 -5.72022140e-01 1.12756634e+00 -1.39881792e-02
1.49711156e+00 2.74113622e-02 -7.02809274e-01 5.76638937e-01
1.72003627e+00 6.01995960e-02 5.91769159e-01 1.10800162e-01
7.20246196e-01 4.67337102e-01 5.53730130e-01 3.03797752e-01
6.13225341e-01 4.58687991e-01 3.21635902e-01 -4.39098060e-01
-3.30808192e-01 -1.37262121e-01 -3.73042189e-02 7.86351025e-01
-3.82757246e-01 1.12213254e-01 -7.86541343e-01 5.88093102e-01
-1.57916486e+00 -6.74175441e-01 2.12154254e-01 2.12197590e+00
1.33803380e+00 -5.86651750e-02 -7.31182918e-02 2.39753112e-01
3.39223862e-01 -1.80749863e-01 -7.68217564e-01 -1.26243785e-01
-7.36890268e-03 6.18521214e-01 3.48978132e-01 6.34422675e-02
-1.05398977e+00 7.72749245e-01 5.16294527e+00 1.28873861e+00
-1.45039213e+00 -1.07803613e-01 4.10125643e-01 -1.24457680e-01
-1.07871093e-01 -5.89517772e-01 -8.78239334e-01 4.22799617e-01
6.62745535e-01 8.29117298e-02 7.28771836e-02 8.99771392e-01
-3.09518039e-01 -3.10498446e-01 -8.47843289e-01 9.00823295e-01
1.91169769e-01 -1.34355974e+00 2.56667614e-01 8.71958584e-02
7.76059926e-01 -4.05244946e-01 3.81506622e-01 2.54693329e-01
6.42045960e-02 -1.06858385e+00 -2.07173780e-01 7.49998212e-01
1.17176187e+00 -5.49933016e-01 9.27080214e-01 2.22471744e-01
-1.19083190e+00 1.68197732e-02 -1.17379054e-01 3.43975991e-01
-3.48751038e-01 8.02041292e-01 -1.05802596e+00 1.12925398e+00
6.31143272e-01 7.68144906e-01 -8.22679758e-01 9.95933652e-01
-2.27506950e-01 5.84640622e-01 -2.64682889e-01 -1.98704749e-01
-7.11628124e-02 1.25162199e-01 -6.69683218e-02 1.13495767e+00
1.42527461e-01 1.32164359e-01 1.20798208e-01 2.85676986e-01
1.04270186e-02 3.06758076e-01 -3.17263305e-01 -1.91608578e-01
3.84918749e-01 1.57735181e+00 -4.28510934e-01 -2.65897840e-01
-4.38435733e-01 9.14567828e-01 3.14051867e-01 6.17160201e-02
-7.61028647e-01 -4.97719884e-01 5.57122469e-01 7.11134262e-03
1.89115152e-01 3.38824749e-01 -1.61082700e-01 -1.14698601e+00
5.12009598e-02 -8.80930722e-01 6.49805188e-01 -1.93199664e-01
-1.56964552e+00 5.61565757e-01 -3.15477163e-01 -1.22501826e+00
4.90432680e-02 -7.89358139e-01 -3.90505195e-01 8.33543718e-01
-2.04701710e+00 -1.61501491e+00 -7.75592387e-01 8.13442767e-01
1.47082672e-01 -1.75277576e-01 1.34021163e+00 3.07986110e-01
-1.03390968e+00 9.78066683e-01 1.28627703e-01 2.36605480e-02
1.01642227e+00 -1.29204011e+00 -3.29957813e-01 4.04016465e-01
-2.35602349e-01 7.95265436e-01 -1.85341120e-01 -6.30332232e-01
-1.47742963e+00 -1.22910035e+00 5.28812230e-01 -1.80518821e-01
6.06298268e-01 8.37469026e-02 -1.03617716e+00 4.57107574e-02
-8.97169709e-02 3.50213259e-01 1.32143927e+00 2.12461784e-01
-5.88216543e-01 -5.58826685e-01 -1.40747523e+00 5.08403838e-01
9.02337492e-01 -5.32549679e-01 -2.08860159e-01 3.32636207e-01
4.05748606e-01 -4.41594541e-01 -1.45577109e+00 6.88620090e-01
8.52047324e-01 -6.44704700e-01 9.15996909e-01 -4.86058205e-01
5.53136170e-01 -8.96022916e-02 3.49596106e-02 -1.18205547e+00
-1.01554401e-01 -1.24785863e-01 -5.19549787e-01 1.12177718e+00
2.26417795e-01 -8.15429151e-01 9.88266289e-01 5.66210210e-01
1.93714544e-01 -1.61717248e+00 -7.58776784e-01 -5.85016131e-01
-7.67213926e-02 3.34917847e-03 3.11622888e-01 1.00014675e+00
1.80368647e-01 3.07608619e-02 -7.03678699e-03 3.19739059e-02
4.54788476e-01 -5.05228788e-02 5.70808589e-01 -1.12584805e+00
-3.17766845e-01 -3.79597247e-01 -5.65062165e-01 -5.53949535e-01
-1.93732858e-01 -9.94078755e-01 -5.16007364e-01 -1.64270091e+00
4.89448160e-01 -7.15438843e-01 -7.39723265e-01 9.02239203e-01
-5.77080488e-01 3.77832383e-01 -2.28496417e-02 -1.44429924e-02
-4.68127936e-01 2.65764773e-01 1.42827094e+00 -2.38746360e-01
-1.84001923e-01 6.83778059e-03 -8.61157119e-01 6.01461351e-01
5.77660561e-01 -3.22633721e-02 -6.36782706e-01 -1.59000233e-01
7.92395324e-02 -2.34741256e-01 2.07497910e-01 -1.04993594e+00
5.48044205e-01 -2.17995584e-01 3.99529248e-01 -2.17332333e-01
4.36843216e-01 -9.05297279e-01 5.60837276e-02 7.89664805e-01
-7.01285601e-02 -7.57646978e-01 2.67689854e-01 6.81123972e-01
-2.90721953e-01 6.31489456e-02 7.36355543e-01 1.43385381e-01
-9.09077764e-01 3.71637195e-01 2.07191989e-01 -2.97041863e-01
1.53113580e+00 -2.59933084e-01 -5.96405268e-01 2.41947711e-01
-5.40531158e-01 1.88781157e-01 1.92214668e-01 3.23742509e-01
6.08350158e-01 -1.07353354e+00 -5.11448026e-01 1.89952880e-01
4.65562046e-01 6.42552674e-02 7.09623814e-01 1.05392432e+00
-5.17680824e-01 4.20473993e-01 -1.69706643e-01 -4.92776543e-01
-1.50927830e+00 3.42760324e-01 2.63539523e-01 -7.38270044e-01
-4.66272980e-01 9.26510930e-01 -1.47834197e-01 -3.29931527e-01
3.43628675e-01 -3.45706344e-01 -2.79235959e-01 4.57728505e-02
6.31214738e-01 4.07256186e-01 3.79207343e-01 -2.64027668e-03
-4.28683877e-01 7.80955911e-01 -5.56104243e-01 6.38863802e-01
1.15104187e+00 3.19914699e-01 -9.57822485e-04 1.39134184e-01
8.38278770e-01 -7.18668625e-02 -8.07019651e-01 -4.85106707e-01
-3.13530773e-01 -2.59368867e-01 1.29524712e-02 -1.40295303e+00
-1.19399476e+00 7.72543132e-01 7.97929525e-01 -4.90657240e-02
1.56006384e+00 -3.04987997e-01 9.23635662e-01 4.72896367e-01
5.21135926e-01 -1.04772329e+00 1.72980249e-01 -1.22575022e-01
4.71057355e-01 -1.31581020e+00 3.57805520e-01 -9.33163226e-01
-8.59167099e-01 1.08858263e+00 8.58242035e-01 7.45387226e-02
7.82472372e-01 3.59473109e-01 1.83478698e-01 -8.51370953e-03
-7.00957477e-01 -3.40526849e-01 4.10306394e-01 5.59238851e-01
3.21811974e-01 2.27529332e-01 -4.75605696e-01 8.34632754e-01
3.30036134e-01 3.49208444e-01 -1.17252360e-03 9.64277864e-01
-1.29420444e-01 -1.34846580e+00 1.89669020e-02 8.19524646e-01
-5.84742785e-01 -6.91594929e-02 -3.66880625e-01 9.34071243e-01
3.49935532e-01 7.29005992e-01 -3.88545513e-01 -6.19351089e-01
3.12571257e-01 -4.61598895e-02 6.58659935e-01 -6.15835667e-01
-8.27559471e-01 -2.50754386e-01 -3.73328244e-03 -5.17039180e-01
-6.75529361e-01 -5.21460176e-02 -1.25533831e+00 -1.67213231e-02
-5.21907032e-01 -9.66018140e-02 6.98765516e-01 8.01179588e-01
5.37325621e-01 8.13752234e-01 5.63892663e-01 -1.93651430e-02
-8.39232862e-01 -7.67042637e-01 -6.81912422e-01 3.88676286e-01
-1.03469044e-01 -7.19714999e-01 3.97079922e-02 6.00512922e-02] | [15.547673225402832, -2.845900297164917] |
47a5aadd-0a4e-4010-9c9d-9b2a7905f086 | zero-shot-cross-lingual-transfer-is-a-hard | null | null | https://aclanthology.org/2021.insights-1.7 | https://aclanthology.org/2021.insights-1.7.pdf | Zero-Shot Cross-Lingual Transfer is a Hard Baseline to Beat in German Fine-Grained Entity Typing | The training of NLP models often requires large amounts of labelled training data, which makes it difficult to expand existing models to new languages. While zero-shot cross-lingual transfer relies on multilingual word embeddings to apply a model trained on one language to another, Yarowski and Ngai (2001) propose the method of annotation projection to generate training data without manual annotation. This method was successfully used for the tasks of named entity recognition and coarse-grained entity typing, but we show that it is outperformed by zero-shot cross-lingual transfer when applied to the similar task of fine-grained entity typing. In our study of fine-grained entity typing with the FIGER type ontology for German, we show that annotation projection amplifies the English model’s tendency to underpredict level 2 labels and is beaten by zero-shot cross-lingual transfer on three novel test sets. | ['Mark Steedman', 'Sabine Weber'] | null | null | null | null | emnlp-insights-2021-11 | ['multilingual-word-embeddings'] | ['methodology'] | [ 1.77679174e-02 5.06643116e-01 -2.01629266e-01 -6.29836440e-01
-9.75175083e-01 -8.47716391e-01 8.26166391e-01 2.54004091e-01
-1.02619445e+00 1.05609322e+00 3.83522242e-01 -5.35340846e-01
2.01943874e-01 -8.30364466e-01 -7.13846326e-01 -2.24904805e-01
1.71764553e-01 9.09081101e-01 6.43580109e-02 -2.31586844e-01
-3.50045174e-01 8.05413276e-02 -1.17629480e+00 4.89643514e-01
9.45325613e-01 2.13161126e-01 1.00595824e-01 3.44740152e-01
-5.14587879e-01 4.50626463e-01 -4.88243550e-01 -9.36946213e-01
4.19875756e-02 -6.42595738e-02 -1.00295663e+00 -4.64611471e-01
6.47988558e-01 5.25799617e-02 1.06814459e-01 8.42646301e-01
6.17001355e-01 1.98075827e-02 8.76845002e-01 -1.09210193e+00
-1.09381354e+00 9.69693244e-01 -1.62870079e-01 -4.71514799e-02
3.67322624e-01 -1.96325064e-01 1.17961586e+00 -9.85260487e-01
1.10676324e+00 1.35756207e+00 1.18219912e+00 9.48035419e-01
-1.51889563e+00 -6.23131156e-01 5.83372973e-02 -9.65828821e-02
-1.37611067e+00 -3.57381314e-01 1.95258886e-01 -6.14049196e-01
1.25676918e+00 -8.45614448e-02 7.21009076e-02 1.25200069e+00
-2.34227449e-01 5.75841784e-01 1.38044548e+00 -8.24740350e-01
-1.27254486e-01 5.32545269e-01 2.46784434e-01 5.44276834e-01
3.86433482e-01 2.89070010e-01 -4.27619457e-01 -1.84157923e-01
4.36886638e-01 -6.84457660e-01 -1.11577742e-01 -1.72762990e-01
-1.23676431e+00 8.72110069e-01 2.35609561e-01 7.10826516e-01
-1.04492374e-01 -8.75477493e-02 5.62618375e-01 4.59730595e-01
7.07595408e-01 8.88588488e-01 -1.08670235e+00 -1.58168823e-01
-7.24560678e-01 1.82136357e-01 9.44854677e-01 1.07552230e+00
9.32959557e-01 5.24010137e-02 -1.89674333e-01 1.22893381e+00
2.00036801e-02 2.60547698e-01 8.52060556e-01 -5.06902874e-01
5.45718193e-01 4.91960287e-01 3.04095924e-01 -1.97587729e-01
-4.19994503e-01 -2.09334493e-01 -2.43815541e-01 1.40696809e-01
8.12612355e-01 -6.24738336e-01 -7.73592710e-01 2.01982164e+00
3.24528098e-01 -1.13847233e-01 4.23092037e-01 3.67403805e-01
5.69131553e-01 4.81078953e-01 7.53356457e-01 8.98866504e-02
1.57399642e+00 -8.54820848e-01 -5.32607436e-01 -3.14825386e-01
1.36592221e+00 -7.12549567e-01 1.27019894e+00 -2.35862389e-01
-7.76437998e-01 -6.42643511e-01 -8.06621432e-01 -3.89399797e-01
-9.96375084e-01 -5.83881065e-02 6.12702250e-01 8.63948047e-01
-8.23866904e-01 4.93163317e-01 -5.82730711e-01 -7.28471398e-01
5.53783067e-02 6.72608614e-02 -7.48943210e-01 -3.93421441e-01
-1.70925248e+00 1.52588868e+00 7.53670990e-01 -2.61516511e-01
-2.86059827e-01 -1.02021146e+00 -1.04002106e+00 5.24163386e-03
1.09405652e-01 -5.59228837e-01 1.30968344e+00 -8.02638769e-01
-1.24131310e+00 1.29486477e+00 -4.77402024e-02 -3.00692111e-01
3.18930030e-01 -2.68105984e-01 -6.05129600e-01 -5.88059902e-01
4.94440675e-01 7.51322687e-01 2.34930962e-01 -9.60659027e-01
-7.85508096e-01 -3.01600784e-01 -3.58203123e-03 1.66980013e-01
-3.49816442e-01 2.54657328e-01 7.07418518e-03 -7.48011470e-01
-5.72463155e-01 -9.61048901e-01 -1.04573652e-01 -5.88638365e-01
2.47948766e-02 -7.89388716e-01 2.76074201e-01 -7.64006674e-01
9.90997612e-01 -1.96166813e+00 1.48429619e-02 -1.11757331e-01
-2.99380869e-01 3.08891594e-01 -4.00043428e-01 5.17275631e-01
-2.86548644e-01 4.45658118e-01 -2.54469186e-01 -2.53156334e-01
4.99969065e-01 3.24305564e-01 -1.33208826e-01 5.10933250e-02
4.38164443e-01 1.02464414e+00 -1.06189835e+00 -3.52458537e-01
-5.94481826e-02 4.60438907e-01 -6.25159204e-01 1.07927352e-01
-4.25785035e-02 1.25526011e-01 2.78975926e-02 1.90106645e-01
3.39198261e-01 1.03349186e-01 5.06162465e-01 -9.73601118e-02
-3.20018142e-01 5.88077128e-01 -9.25329208e-01 1.59055686e+00
-1.00018120e+00 3.72887760e-01 -3.36959213e-01 -3.85807544e-01
6.80390596e-01 7.73969531e-01 -1.12532293e-02 -5.04559994e-01
-2.02408075e-01 6.17129564e-01 1.38695225e-01 -3.20649624e-01
5.95174551e-01 -6.48880363e-01 -7.12992430e-01 6.40916288e-01
5.96752465e-01 2.59834100e-02 1.99922472e-01 2.80355103e-02
8.07689011e-01 4.84790415e-01 3.84499699e-01 -3.83248568e-01
2.17537627e-01 1.68737784e-01 6.73951924e-01 5.89743793e-01
3.66683817e-03 4.79550995e-02 2.04106912e-01 -4.21932131e-01
-1.12828970e+00 -1.08453953e+00 -5.95098257e-01 1.81016469e+00
-4.19865668e-01 -4.62130427e-01 -8.38958740e-01 -1.07380307e+00
1.88904807e-01 1.23587835e+00 -8.42468262e-01 1.34105071e-01
-6.22249186e-01 -7.72018373e-01 1.06301665e+00 5.89897692e-01
-4.64174943e-03 -1.18104386e+00 -1.63020521e-01 4.92236584e-01
-1.26921773e-01 -1.23178625e+00 -3.96151304e-01 2.91373730e-01
-4.29782718e-01 -8.35414886e-01 -8.48131299e-01 -1.13498950e+00
6.22067511e-01 -6.47624850e-01 1.33074224e+00 -2.66096622e-01
-1.47281796e-01 2.83097565e-01 -4.08493519e-01 -5.05490780e-01
-7.17651367e-01 6.68885052e-01 2.35526830e-01 -2.89330781e-01
8.30027223e-01 -1.36590824e-01 2.65699893e-01 2.02169403e-01
-7.42494702e-01 -1.66256502e-01 4.69438404e-01 1.11645174e+00
2.28007928e-01 -4.57041562e-01 8.41342509e-01 -1.63313389e+00
7.65358329e-01 -4.33891594e-01 -4.26800847e-01 4.38501567e-01
-5.66880345e-01 3.96277726e-01 4.85921770e-01 -4.34271336e-01
-1.45295823e+00 -1.21578664e-01 -3.36618662e-01 1.93977356e-01
-4.35561150e-01 6.11263812e-01 -1.71182230e-01 1.33695185e-01
8.07483196e-01 -2.45463148e-01 -6.34078681e-01 -7.93685973e-01
8.57119799e-01 7.53400087e-01 4.15189236e-01 -8.60221088e-01
8.30737591e-01 -1.99223056e-01 -7.33401716e-01 -7.13490963e-01
-9.27722812e-01 -2.28687957e-01 -1.06076324e+00 2.82174855e-01
1.26716399e+00 -9.84889627e-01 -1.07631400e-01 2.24130437e-01
-1.27508593e+00 -6.84508026e-01 -4.71331030e-01 5.19958377e-01
-3.78772676e-01 -1.07247785e-01 -7.07229018e-01 -3.61488968e-01
-9.41508114e-02 -6.36202574e-01 9.80162084e-01 -2.73641378e-01
-7.22945750e-01 -1.54356802e+00 5.60423017e-01 6.08340986e-02
3.17797154e-01 -8.63284990e-02 1.39009333e+00 -1.17471182e+00
3.88690308e-02 -2.35783294e-01 -1.28861293e-01 2.75011033e-01
9.61080045e-02 -2.50134975e-01 -1.06415641e+00 -1.58650219e-01
-4.79156017e-01 -5.48108160e-01 4.91065860e-01 -2.03187317e-01
3.99534553e-01 -2.32686341e-01 -3.94700080e-01 4.86936569e-01
1.39263105e+00 -1.28854677e-01 4.89895254e-01 3.54280531e-01
8.33503783e-01 7.90992618e-01 4.62615520e-01 -2.89834529e-01
6.74890161e-01 8.45680356e-01 -5.84603727e-01 6.47937532e-06
-2.58917898e-01 -5.67938924e-01 3.45756978e-01 1.10239041e+00
-1.24937840e-01 -1.22861952e-01 -1.08419490e+00 8.90690446e-01
-1.49601495e+00 -8.00755739e-01 7.04930443e-03 2.21292233e+00
1.29331005e+00 -5.75238056e-02 -1.66825607e-01 -4.52270299e-01
8.00174832e-01 -1.75014719e-01 2.93587614e-02 -7.23100722e-01
-7.27504641e-02 6.44327760e-01 5.30737579e-01 8.58405054e-01
-9.29341376e-01 1.62590742e+00 6.43186474e+00 7.40696847e-01
-8.35388362e-01 5.84266841e-01 1.53320044e-01 4.74451900e-01
-4.06146854e-01 -1.72997341e-02 -1.18068349e+00 4.41678882e-01
1.33186555e+00 -2.93315619e-01 5.55994064e-02 7.15241134e-01
-4.23417002e-01 1.71635017e-01 -1.46634185e+00 4.76641417e-01
-1.58964153e-02 -9.65307891e-01 -7.68480599e-02 -7.69623145e-02
9.16965902e-01 2.69141227e-01 -3.86552304e-01 9.24698055e-01
1.07397580e+00 -1.00438273e+00 4.00902003e-01 4.49316837e-02
1.21614015e+00 -5.86780906e-01 8.17429960e-01 2.68867850e-01
-9.70686615e-01 2.39883423e-01 -3.94080371e-01 2.99196318e-03
4.23617959e-01 1.42073303e-01 -1.15133989e+00 3.43380690e-01
2.82880485e-01 1.58767581e-01 -5.47333717e-01 7.31342256e-01
-5.71060658e-01 5.25374353e-01 -6.34959489e-02 3.60008068e-02
1.95156485e-01 1.43687636e-01 2.11174518e-01 1.53973114e+00
3.61717582e-01 -1.96917757e-01 1.41638786e-01 4.87486333e-01
-3.65340203e-01 4.50362176e-01 -8.20835888e-01 -2.08162561e-01
5.64295053e-01 1.20482922e+00 -2.22272024e-01 -7.56846189e-01
-6.13598645e-01 1.05811441e+00 7.90099263e-01 3.51281971e-01
-5.04980743e-01 -6.87864244e-01 5.27441978e-01 8.42804164e-02
4.46433902e-01 -2.37181351e-01 -1.92635506e-01 -1.26723528e+00
-3.62373233e-01 -6.41464889e-01 4.51125026e-01 -7.17483461e-01
-1.72624016e+00 8.00822675e-01 1.45358695e-02 -7.69003630e-01
-5.37234545e-01 -8.66210282e-01 -3.09561819e-01 1.31575799e+00
-1.41338098e+00 -1.53804219e+00 1.69162750e-01 4.60305035e-01
4.48402196e-01 -1.15991846e-01 1.58183563e+00 6.57140791e-01
-2.59432465e-01 9.74174142e-01 4.48969714e-02 6.47052586e-01
1.33783877e+00 -1.47964251e+00 7.70213425e-01 5.15341222e-01
3.31820428e-01 7.05146968e-01 4.88385379e-01 -6.40759826e-01
-7.04511166e-01 -1.31422746e+00 1.81747878e+00 -8.68421733e-01
1.06618500e+00 -5.81726670e-01 -1.07538581e+00 1.31425226e+00
4.52305079e-01 -1.48986936e-01 1.13098919e+00 7.79499054e-01
-8.08371127e-01 1.83370665e-01 -1.07326198e+00 5.36296606e-01
9.12302196e-01 -8.23454916e-01 -1.22272682e+00 3.19685042e-01
7.44625628e-01 -5.30062132e-02 -1.28404748e+00 1.46184325e-01
6.04742587e-01 -2.36174703e-01 4.88369405e-01 -1.14237738e+00
2.16538697e-01 -2.05016825e-02 -1.03961520e-01 -1.90619683e+00
-4.44563836e-01 -3.05301368e-01 5.40082097e-01 1.51477182e+00
9.67693746e-01 -9.70266402e-01 2.33918384e-01 6.30727768e-01
-2.57057816e-01 -1.59614682e-01 -9.25484240e-01 -9.48522389e-01
7.21388042e-01 -2.72849351e-01 5.44654727e-01 1.59517801e+00
2.46650770e-01 6.99369550e-01 -1.26365926e-02 4.56507690e-02
4.72153574e-01 -2.82395184e-01 7.04975247e-01 -1.37531054e+00
-4.05951589e-01 -9.95504111e-02 -4.05255765e-01 -4.55447108e-01
6.12866759e-01 -1.31944907e+00 2.75302261e-01 -1.38092077e+00
-1.14728376e-01 -9.58912075e-01 -3.16273808e-01 7.22142458e-01
-2.72696227e-01 4.35492605e-01 2.41420150e-01 -1.41674489e-01
-2.12280199e-01 9.70888734e-02 6.50122643e-01 5.86392730e-02
1.11350566e-01 -3.52343708e-01 -6.31129980e-01 7.38212645e-01
4.91953641e-01 -6.74364984e-01 1.28368009e-02 -8.56897950e-01
3.63087833e-01 -2.27571666e-01 -5.95489293e-02 -7.43407249e-01
-1.83869034e-01 8.28004107e-02 3.64795983e-01 1.44135743e-01
7.79820010e-02 -5.45556366e-01 -6.99247345e-02 1.39056772e-01
-4.41558599e-01 6.78219497e-02 4.17577147e-01 2.00584978e-01
-1.89613566e-01 -4.81580675e-01 6.44356489e-01 -2.11446241e-01
-9.25574660e-01 1.34161204e-01 -5.08049726e-01 6.00620210e-01
8.02531600e-01 -1.72720873e-03 -3.72730970e-01 2.39332184e-01
-1.10784543e+00 1.72208007e-02 6.39984608e-01 6.42074645e-01
-1.88490599e-01 -1.62108183e+00 -8.98281515e-01 3.19130361e-01
3.42352808e-01 -4.07982796e-01 1.27771497e-01 5.55119634e-01
-2.12680116e-01 5.57704687e-01 -3.53007406e-01 -1.54319614e-01
-1.06572509e+00 4.00492370e-01 1.24099672e-01 -5.77105343e-01
-1.90900564e-01 1.05190945e+00 4.05704111e-01 -1.22852361e+00
-2.29008257e-01 -9.43415090e-02 -1.89853106e-02 2.60140389e-01
3.12680423e-01 1.79678395e-01 1.63177475e-01 -7.82033026e-01
-2.80591667e-01 4.67624754e-01 -2.54804194e-01 -3.65870118e-01
1.25923800e+00 4.34040688e-02 3.08879707e-02 9.30169642e-01
1.22372031e+00 4.07124639e-01 -7.44019687e-01 -3.19517910e-01
5.17673075e-01 -7.04433843e-02 -5.11422753e-01 -1.04039085e+00
-2.62253314e-01 1.00956750e+00 3.35322469e-01 3.82997133e-02
3.63955051e-01 2.22455859e-02 7.27361262e-01 3.25065076e-01
6.23247862e-01 -1.18687713e+00 -6.11031950e-01 9.65972245e-01
5.94335318e-01 -1.11947179e+00 -4.35214281e-01 -2.97740847e-01
-6.52326226e-01 7.57764339e-01 6.71965718e-01 2.51801275e-02
5.23845434e-01 3.07312518e-01 3.04344803e-01 5.12497425e-02
-7.69587576e-01 -2.53654242e-01 2.19186664e-01 8.00420761e-01
9.59023714e-01 4.23624545e-01 -3.42368186e-01 6.97770894e-01
-4.38951939e-01 5.49078546e-02 3.86256695e-01 7.30666637e-01
-7.87779689e-02 -1.55364311e+00 -8.65858793e-02 2.89624542e-01
-5.53686321e-01 -6.45851076e-01 -3.33985716e-01 1.16446519e+00
5.19797981e-01 3.33433121e-01 3.87523174e-01 -5.47386222e-02
2.91153580e-01 1.02583551e+00 7.14033008e-01 -1.27745128e+00
-7.74659455e-01 -2.85493344e-01 7.48298705e-01 -4.54528928e-02
-3.35726231e-01 -6.21155977e-01 -8.91159356e-01 2.61746366e-02
-2.50016123e-01 4.32964742e-01 5.79831362e-01 1.00628698e+00
2.48301953e-01 4.17160451e-01 -4.09102887e-02 -5.32517552e-01
-3.34884226e-01 -1.37643385e+00 -5.69673002e-01 6.38605058e-01
-1.09117195e-01 -6.81464970e-01 -2.15920776e-01 2.72867233e-01] | [10.597139358520508, 9.809503555297852] |
0f3d54ac-e0b3-438b-8bd1-b403159106ab | 190600672 | 1906.00672 | null | https://arxiv.org/abs/1906.00672v3 | https://arxiv.org/pdf/1906.00672v3.pdf | Robust Sequence-to-Sequence Acoustic Modeling with Stepwise Monotonic Attention for Neural TTS | Neural TTS has demonstrated strong capabilities to generate human-like speech with high quality and naturalness, while its generalization to out-of-domain texts is still a challenging task, with regard to the design of attention-based sequence-to-sequence acoustic modeling. Various errors occur in those inputs with unseen context, including attention collapse, skipping, repeating, etc., which limits the broader applications. In this paper, we propose a novel stepwise monotonic attention method in sequence-to-sequence acoustic modeling to improve the robustness on out-of-domain inputs. The method utilizes the strict monotonic property in TTS with constraints on monotonic hard attention that the alignments between inputs and outputs sequence must be not only monotonic but allowing no skipping on inputs. Soft attention could be used to evade mismatch between training and inference. The experimental results show that the proposed method could achieve significant improvements in robustness on out-of-domain scenarios for phoneme-based models, without any regression on the in-domain naturalness test. | ['Yan Deng', 'Mutian He', 'Lei He'] | 2019-06-03 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 4.25623983e-01 3.33646238e-02 4.65186566e-01 -4.00088876e-01
-7.98537374e-01 -4.56316292e-01 2.72610784e-01 -4.13944483e-01
-3.95423383e-01 5.51022828e-01 1.13232605e-01 -5.87215185e-01
2.19158813e-01 -3.23074460e-01 -6.85797334e-01 -6.88564062e-01
3.47582310e-01 2.89326549e-01 4.98302191e-01 -4.11199898e-01
7.35430345e-02 1.60964593e-01 -1.50041485e+00 3.62749279e-01
7.91743636e-01 8.07443023e-01 8.24178517e-01 9.89049733e-01
-1.60157442e-01 5.19794643e-01 -8.98190320e-01 -4.01492894e-01
6.95984438e-02 -5.92427373e-01 -4.39150900e-01 -3.12674910e-01
1.31121963e-01 -2.23158091e-01 -2.74080634e-01 1.12609506e+00
1.02917910e+00 3.32692474e-01 6.03536963e-01 -8.83324683e-01
-7.82883346e-01 7.07306623e-01 -1.36884883e-01 3.59849930e-01
3.87884825e-01 4.21210170e-01 9.60493684e-01 -9.52843606e-01
-3.98368686e-02 1.26877820e+00 4.94787097e-01 8.46041203e-01
-7.82693446e-01 -7.46750832e-01 8.29460844e-02 5.03111482e-01
-1.23761165e+00 -7.92754233e-01 7.40362942e-01 -2.90232211e-01
1.33028948e+00 6.38487160e-01 1.52510136e-01 1.49370980e+00
-1.53033823e-01 6.05837762e-01 8.10505807e-01 -7.26895273e-01
1.75995722e-01 1.04852632e-01 -5.44219837e-02 1.45513013e-01
-3.10162455e-01 3.40231866e-01 -5.65028548e-01 8.75668898e-02
4.71371502e-01 -6.83562934e-01 -3.83455276e-01 4.86331969e-01
-1.03314352e+00 6.01824462e-01 -6.77940175e-02 4.18126285e-01
-1.69485062e-01 -8.70303735e-02 5.02359092e-01 2.20165193e-01
1.87764198e-01 2.67246783e-01 -5.87741673e-01 -5.11237800e-01
-8.75519633e-01 -1.59937534e-02 3.93624246e-01 1.14273810e+00
1.14954144e-01 8.04365754e-01 -2.53282994e-01 1.16801071e+00
1.70997798e-01 8.11899602e-01 1.01156747e+00 -3.97180617e-01
7.82294333e-01 -1.07612185e-01 1.00772921e-02 -7.43452132e-01
-2.09105253e-01 -5.23794830e-01 -7.05291450e-01 -6.10091276e-02
2.32488975e-01 -1.12924352e-01 -1.02459919e+00 1.98508191e+00
1.46088034e-01 2.51715362e-01 1.85985908e-01 8.95604670e-01
8.58327687e-01 1.06035995e+00 -1.04325470e-02 -4.09340501e-01
1.14879978e+00 -9.08610165e-01 -1.20420253e+00 -3.04526418e-01
3.63857985e-01 -1.03651166e+00 1.68872261e+00 4.55251932e-01
-1.16368544e+00 -1.13056719e+00 -1.04130268e+00 5.97539246e-02
-1.48114681e-01 7.38730654e-02 -3.69616002e-01 8.01138818e-01
-8.17704797e-01 3.56199324e-01 -6.14821553e-01 -1.20831288e-01
-2.85949528e-01 2.68890411e-01 -6.11210801e-02 2.73280203e-01
-1.60983908e+00 9.46826160e-01 6.52061701e-01 3.28188062e-01
-8.69678497e-01 -4.06252205e-01 -7.76650310e-01 1.73522711e-01
2.38146171e-01 -3.50399047e-01 1.51570642e+00 -7.95752406e-01
-1.98167217e+00 3.21974009e-01 -2.92490751e-01 -3.51440877e-01
4.17638659e-01 -4.08048242e-01 -9.15826380e-01 -3.17735642e-01
-2.33584881e-01 3.07947606e-01 9.37880337e-01 -9.51901257e-01
-5.68234622e-01 -1.09106556e-01 -4.27437097e-01 3.11073989e-01
-4.60311621e-01 1.68914422e-01 -2.77746409e-01 -8.33094656e-01
-3.52300406e-02 -9.24477994e-01 5.37965186e-02 -7.24347532e-01
-7.06072092e-01 -3.06943476e-01 9.70608890e-01 -9.62641180e-01
1.58608639e+00 -2.16880584e+00 -5.54790050e-02 -1.36815803e-02
-6.37414157e-01 7.33645499e-01 -2.07649201e-01 4.40354824e-01
1.70638878e-02 1.23810180e-01 -3.62669706e-01 -4.10483897e-01
-1.69357974e-02 2.32042953e-01 -5.27054369e-01 6.84006661e-02
4.19144928e-01 6.92146122e-01 -7.52485216e-01 -3.94834816e-01
3.00295264e-01 3.78073126e-01 -4.27453011e-01 7.42316425e-01
-1.87507033e-01 6.85297191e-01 -1.36855870e-01 2.91889697e-01
4.87724483e-01 4.57221299e-01 2.60656383e-02 4.17315811e-02
-7.09017292e-02 6.11371160e-01 -9.63680744e-01 1.37936091e+00
-7.77490139e-01 6.65301204e-01 -7.00751022e-02 -8.56614172e-01
1.10019660e+00 6.53673828e-01 -2.50407785e-01 -8.59226406e-01
9.66054946e-02 3.49752814e-01 7.25571334e-01 -8.72936308e-01
6.21893466e-01 -4.20833051e-01 2.41506496e-03 1.06468774e-01
-1.68596908e-01 -2.70814836e-01 -1.80582240e-01 -2.57668138e-01
7.65280843e-01 1.31430075e-01 2.08000854e-01 -1.61606506e-01
6.92435682e-01 -4.96523321e-01 5.25700271e-01 7.46559680e-01
-1.63178012e-01 9.45648193e-01 9.08381715e-02 8.88428688e-02
-1.16341650e+00 -1.03408468e+00 6.36854321e-02 1.26070118e+00
-1.37323543e-01 -1.45816701e-02 -8.99973452e-01 -4.29273814e-01
-4.08398628e-01 1.16855359e+00 -2.00963452e-01 -2.87649393e-01
-8.34896982e-01 -4.08152372e-01 1.04904711e+00 5.32092869e-01
1.92455441e-01 -1.55684030e+00 -2.29425862e-01 5.00019014e-01
-4.40760851e-01 -1.36010635e+00 -7.15469897e-01 2.51256555e-01
-5.22624075e-01 -5.83412468e-01 -8.13415170e-01 -9.08931136e-01
2.75273472e-01 -6.65105805e-02 7.34565854e-01 -4.45352644e-02
2.63403296e-01 -2.60957509e-01 -5.41231811e-01 -4.71880257e-01
-9.52426195e-01 1.15717866e-01 3.20873618e-01 1.27491340e-01
1.74071342e-01 -3.95954669e-01 -2.87334323e-01 5.41981399e-01
-8.82582605e-01 -8.94848034e-02 4.63209152e-01 1.06145203e+00
2.35426530e-01 -7.81750157e-02 1.13590205e+00 -5.56418419e-01
8.19000006e-01 -3.64891648e-01 -3.75758529e-01 1.81242213e-01
-1.80366680e-01 -3.07672881e-02 1.19687092e+00 -7.24590421e-01
-1.12927544e+00 -1.83129594e-01 -8.09865713e-01 -4.24218357e-01
-3.62189978e-01 2.94524938e-01 -6.87198043e-01 5.20919502e-01
5.41663885e-01 6.51854455e-01 -1.82014793e-01 -6.18465543e-01
1.05806008e-01 1.17000198e+00 6.47067964e-01 -4.42484915e-01
7.43358254e-01 -2.57579207e-01 -4.86720294e-01 -1.17848349e+00
-3.66014034e-01 -3.18544209e-01 -4.37037438e-01 -8.64890367e-02
7.36935973e-01 -6.15297794e-01 -4.80054915e-01 5.45453727e-01
-1.42327380e+00 -3.81375879e-01 1.74579173e-01 5.97726941e-01
-6.05090201e-01 7.00415432e-01 -6.13832533e-01 -1.16953826e+00
-3.48713756e-01 -1.31955791e+00 7.84709573e-01 -1.27067678e-02
-4.54903126e-01 -6.88515484e-01 -7.75717646e-02 3.65274280e-01
5.00473738e-01 -4.15223271e-01 9.77244794e-01 -9.78232682e-01
-2.55347520e-01 3.01120356e-02 3.30490470e-01 9.46270883e-01
1.51199192e-01 1.28133446e-01 -1.35279584e+00 -1.67359471e-01
2.13817254e-01 -1.54916570e-01 3.84822488e-01 2.48230129e-01
1.16430902e+00 -5.26615381e-01 1.57712072e-01 4.17780519e-01
8.76940906e-01 8.17315221e-01 9.18617427e-01 -3.26786004e-02
6.36607051e-01 5.91252506e-01 7.54742742e-01 2.51827240e-01
-1.73633203e-01 1.00473642e+00 2.67485052e-01 -7.02985376e-02
-3.08410943e-01 -3.35444063e-01 6.86724246e-01 1.64144111e+00
2.39263937e-01 -7.54320681e-01 -6.79198384e-01 7.40925014e-01
-1.41823208e+00 -9.16850686e-01 -2.56185889e-01 2.45730877e+00
1.02112436e+00 5.00996292e-01 -1.13538660e-01 4.63426888e-01
1.19795942e+00 2.91456562e-02 -3.26872498e-01 -9.66697156e-01
-1.14162445e-01 2.18417138e-01 1.36595592e-03 6.51446819e-01
-6.99014783e-01 1.14972007e+00 5.65548468e+00 1.37808859e+00
-1.21214426e+00 1.90672651e-01 4.59238678e-01 -3.73267122e-02
-3.93763334e-01 -3.23366553e-01 -9.31342959e-01 9.50278878e-01
1.26484716e+00 3.45408499e-01 3.34072620e-01 5.72245657e-01
6.29060805e-01 4.30763096e-01 -1.03564155e+00 7.30099618e-01
-1.34040311e-01 -6.55606329e-01 5.74246645e-02 -3.09542894e-01
6.40817463e-01 -2.29071498e-01 2.58938879e-01 4.77290839e-01
-1.05645128e-01 -1.08000696e+00 9.69015062e-01 1.75134823e-01
1.01394582e+00 -8.21341157e-01 7.79205263e-01 6.75531805e-01
-1.00532699e+00 8.05766359e-02 -3.47563952e-01 -1.07431255e-01
3.82716924e-01 3.64505827e-01 -1.41997564e+00 4.11771178e-01
3.98960233e-01 -4.76565817e-03 -7.33616650e-02 8.19595456e-01
-3.06073070e-01 1.15239334e+00 -3.37914884e-01 -4.20155555e-01
1.79577082e-01 1.88204348e-01 7.71426380e-01 1.57134593e+00
7.20467985e-01 -3.38862166e-02 -1.53641313e-01 5.99259675e-01
3.05221617e-01 3.21683019e-01 -6.36063278e-01 -9.21615108e-04
6.93824351e-01 6.55917346e-01 -1.16566725e-01 -2.35364392e-01
-3.04871827e-01 9.70887125e-01 1.35335401e-01 5.64465642e-01
-1.09691036e+00 -6.06552601e-01 5.39772928e-01 1.30622581e-01
4.65129524e-01 -2.53689557e-01 -3.89922321e-01 -6.60676360e-01
9.76793617e-02 -1.05788994e+00 1.81441531e-02 -8.86754870e-01
-1.22743833e+00 1.05136836e+00 -2.39090815e-01 -1.24326289e+00
-4.31251049e-01 -5.13209403e-01 -8.09851170e-01 1.13162422e+00
-1.41377044e+00 -9.63623047e-01 1.48482561e-01 4.45611179e-01
1.03932512e+00 -2.74079740e-01 7.98760653e-01 4.40388888e-01
-6.16694510e-01 1.02008271e+00 7.31244534e-02 1.29346410e-02
6.51445329e-01 -9.38720047e-01 7.01385796e-01 1.12447238e+00
1.43187314e-01 4.36476052e-01 8.84311259e-01 -6.13893151e-01
-7.65136003e-01 -9.96099055e-01 1.02841568e+00 -1.85317963e-01
4.27466869e-01 -4.88586009e-01 -1.11823881e+00 4.04887468e-01
2.59358943e-01 -2.68409133e-01 5.46434641e-01 -5.25597483e-02
-1.11434810e-01 -8.91294777e-02 -8.43787611e-01 8.28908503e-01
9.21181262e-01 -6.62812591e-01 -8.45165312e-01 -2.70191953e-02
1.06273293e+00 -4.34186935e-01 -4.00333136e-01 5.90870738e-01
4.44169700e-01 -9.31455731e-01 7.52482712e-01 -6.23470306e-01
1.99219063e-01 -4.47219789e-01 -4.58259016e-01 -1.36985350e+00
-3.79777074e-01 -8.63507628e-01 1.44872174e-03 1.48585343e+00
7.14537561e-01 -4.23668087e-01 3.56430262e-01 3.07676703e-01
-8.55640769e-01 -4.45623934e-01 -1.23536456e+00 -1.09093475e+00
2.72364193e-03 -7.34299541e-01 6.44843996e-01 5.97075760e-01
-1.71600822e-02 3.16166908e-01 -7.79335201e-01 5.34023583e-01
-1.81314535e-02 -4.95453328e-01 5.64531982e-01 -5.70982337e-01
-5.48347056e-01 -3.03912222e-01 -2.43466124e-01 -1.26456642e+00
6.40129074e-02 -5.42915583e-01 6.32692337e-01 -1.04041827e+00
-4.32450145e-01 -3.31336796e-01 -4.19673264e-01 2.12523535e-01
-4.93566424e-01 4.35706899e-02 3.30914706e-01 -2.21341819e-01
-2.10250527e-01 8.82591426e-01 1.19137657e+00 -2.85404734e-02
-1.53512090e-01 3.38677287e-01 -2.39782438e-01 5.52261829e-01
9.99440074e-01 -5.77341557e-01 -6.36291146e-01 -5.37582338e-01
-4.11721259e-01 3.34879547e-01 -1.05302975e-01 -1.23034692e+00
1.50475159e-01 -2.06306428e-01 -9.53211933e-02 -6.06964648e-01
5.86525500e-01 -8.31426203e-01 1.41017020e-01 3.93766165e-01
-5.10936260e-01 9.43370983e-02 4.02515352e-01 4.14258361e-01
-2.51289308e-01 -5.86314261e-01 7.94321120e-01 -1.92475542e-02
-6.11108959e-01 -1.29482031e-01 -4.32707638e-01 2.97592163e-01
7.11018562e-01 -3.06583494e-01 -8.36587399e-02 -5.16142190e-01
-6.87834740e-01 -2.32814088e-01 3.63050844e-03 6.30386293e-01
6.35554433e-01 -1.28149521e+00 -1.00739682e+00 5.23491681e-01
-3.76023576e-02 -1.34819970e-01 5.32433391e-01 3.52313608e-01
-2.03426927e-01 6.17656350e-01 1.02315098e-01 -6.14499390e-01
-1.36881077e+00 5.06098628e-01 3.06610733e-01 -4.27647941e-02
-2.11152703e-01 1.00594699e+00 4.93281811e-01 -5.22128940e-01
5.20258248e-01 -4.75276917e-01 7.29669118e-04 -4.81979311e-01
4.01522577e-01 2.52617300e-01 3.73788595e-01 -6.98453605e-01
-3.65227938e-01 4.63634759e-01 1.83966104e-02 -3.45012009e-01
6.94424689e-01 -3.69138718e-01 5.06562114e-01 6.68771923e-01
1.19055188e+00 3.29340160e-01 -1.15806913e+00 -3.43429856e-02
-3.33417952e-02 -3.78794342e-01 -2.61606008e-01 -1.02119803e+00
-6.54639781e-01 1.32790351e+00 3.82357538e-01 3.62401038e-01
8.87980163e-01 -3.23805511e-01 1.25119710e+00 1.69755876e-01
3.84167172e-02 -1.27243459e+00 2.05349043e-01 8.25892806e-01
1.15596735e+00 -1.13960433e+00 -6.54159784e-01 -2.45301142e-01
-8.60893011e-01 9.65568960e-01 8.70647609e-01 3.62442940e-01
3.16023380e-01 3.12458783e-01 2.48994350e-01 5.00123918e-01
-7.90669322e-01 -9.09041911e-02 2.57042140e-01 9.17281151e-01
4.46045130e-01 6.00148439e-02 -7.69627988e-02 6.97313070e-01
-5.24424732e-01 -4.73565668e-01 4.21855330e-01 4.90131348e-01
-7.06170142e-01 -1.02476490e+00 -6.14540756e-01 8.94987881e-02
-5.92137456e-01 -5.40930212e-01 -1.62135512e-01 5.39058805e-01
1.30678952e-01 1.33572304e+00 3.46916057e-02 -5.54407477e-01
5.19369721e-01 4.26620841e-01 2.95605287e-02 -5.34457207e-01
-6.09014452e-01 3.21664125e-01 2.42961243e-01 1.24235183e-01
2.56943703e-01 -4.48464513e-01 -1.27393532e+00 1.13201492e-01
-6.51911199e-01 1.46170616e-01 6.55232728e-01 9.16254759e-01
2.17884943e-01 7.55647004e-01 6.81622624e-01 -5.07089317e-01
-9.36406851e-01 -1.48622954e+00 -3.84603441e-01 5.48271716e-01
5.82752109e-01 -3.61102045e-01 -5.31832516e-01 8.48496109e-02] | [14.683048248291016, 6.7006306648254395] |
fe6e9ba1-ac4c-4e4f-babf-6d126d278165 | real-time-scene-text-detection-with-1 | 2202.10304 | null | https://arxiv.org/abs/2202.10304v1 | https://arxiv.org/pdf/2202.10304v1.pdf | Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion | Recently, segmentation-based scene text detection methods have drawn extensive attention in the scene text detection field, because of their superiority in detecting the text instances of arbitrary shapes and extreme aspect ratios, profiting from the pixel-level descriptions. However, the vast majority of the existing segmentation-based approaches are limited to their complex post-processing algorithms and the scale robustness of their segmentation models, where the post-processing algorithms are not only isolated to the model optimization but also time-consuming and the scale robustness is usually strengthened by fusing multi-scale feature maps directly. In this paper, we propose a Differentiable Binarization (DB) module that integrates the binarization process, one of the most important steps in the post-processing procedure, into a segmentation network. Optimized along with the proposed DB module, the segmentation network can produce more accurate results, which enhances the accuracy of text detection with a simple pipeline. Furthermore, an efficient Adaptive Scale Fusion (ASF) module is proposed to improve the scale robustness by fusing features of different scales adaptively. By incorporating the proposed DB and ASF with the segmentation network, our proposed scene text detector consistently achieves state-of-the-art results, in terms of both detection accuracy and speed, on five standard benchmarks. | ['Xiang Bai', 'Cong Yao', 'Zhaoyi Wan', 'Zhisheng Zou', 'Minghui Liao'] | 2022-02-21 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 3.98994863e-01 -6.09154165e-01 2.33481318e-01 -2.12664440e-01
-4.73545462e-01 -4.23240155e-01 6.09489501e-01 4.83746171e-01
-6.30424857e-01 2.20506545e-02 -2.23813921e-01 -7.36186206e-02
1.64020061e-01 -8.77856612e-01 -2.66697437e-01 -7.43427396e-01
5.23758829e-01 2.71376252e-01 1.03655016e+00 -1.23983614e-01
4.92129087e-01 5.50823987e-01 -1.48838866e+00 -2.73828730e-02
1.28061533e+00 1.14088559e+00 2.90842503e-01 6.38133049e-01
-4.81561214e-01 1.92080617e-01 -4.15291697e-01 -3.35313022e-01
1.93128511e-01 -2.65082002e-01 -1.14519067e-01 3.63515019e-01
2.09822223e-01 -3.32618505e-01 -3.36537570e-01 1.30682158e+00
5.56873322e-01 1.77185666e-02 4.93383616e-01 -9.55044031e-01
-6.75160587e-02 4.36547905e-01 -9.46067274e-01 8.41793567e-02
2.02204928e-01 1.08070254e-01 8.37928891e-01 -7.74249315e-01
1.20397143e-01 1.24943399e+00 5.45436263e-01 -9.90867466e-02
-9.71079171e-01 -5.61234772e-01 1.74644545e-01 1.68041542e-01
-1.57254410e+00 -2.64930129e-01 9.03381348e-01 -3.58941048e-01
6.06936932e-01 3.59211564e-01 6.75885856e-01 3.20047349e-01
7.42049366e-02 1.06078315e+00 7.91359901e-01 -3.34699124e-01
1.33645445e-01 1.42805561e-01 1.94006309e-01 6.18201256e-01
5.03367841e-01 -4.62145567e-01 -2.87165314e-01 2.59143531e-01
8.09405029e-01 2.57027537e-01 -1.73483431e-01 -3.67185295e-01
-1.19296670e+00 5.43988466e-01 6.55601084e-01 5.40473998e-01
-2.68433601e-01 -1.47444010e-01 4.16851848e-01 -3.04366350e-01
5.10316610e-01 1.11810349e-01 -1.05555551e-02 1.32170081e-01
-1.36963701e+00 2.46078484e-02 4.41556215e-01 8.33655238e-01
6.93523645e-01 -3.48292701e-02 -3.45906645e-01 7.83428729e-01
3.56583595e-01 6.47602618e-01 6.58151448e-01 -8.49535167e-02
5.72900414e-01 1.25368321e+00 2.54464559e-02 -1.22683489e+00
-6.24731779e-01 -3.92753989e-01 -8.91809583e-01 -1.56733673e-02
4.23974246e-01 -6.59004739e-03 -1.01628447e+00 1.05435336e+00
7.05600917e-01 -9.64942798e-02 -9.58390161e-02 9.33256805e-01
8.25085580e-01 7.02188909e-01 2.92577548e-03 -1.60544395e-01
1.57850862e+00 -1.00222254e+00 -6.63456798e-01 -2.86256731e-01
4.76348400e-01 -1.12557578e+00 1.09943211e+00 2.71150470e-01
-8.48816097e-01 -5.63778341e-01 -1.23993945e+00 -2.16589838e-01
-5.21081448e-01 5.22085607e-01 3.83320600e-01 6.29777133e-01
-7.36985028e-01 2.54864335e-01 -8.55718732e-01 -7.60408878e-01
3.61641735e-01 3.12882036e-01 -1.46212904e-02 1.55282812e-02
-7.36704528e-01 4.66718197e-01 5.56021512e-01 1.81753337e-01
-1.85356945e-01 -1.08532481e-01 -7.39879549e-01 3.51650149e-01
5.80709696e-01 -3.56188327e-01 8.11662734e-01 -7.59392500e-01
-1.44516563e+00 5.12583435e-01 -1.90332919e-01 -1.74144074e-01
8.79368842e-01 -7.18104094e-02 -3.55600625e-01 4.11065072e-01
1.31427422e-01 6.10036612e-01 1.06779647e+00 -8.60435545e-01
-1.05236053e+00 -4.31656808e-01 -2.12914690e-01 4.45206940e-01
-6.71773732e-01 1.42520458e-01 -1.02316928e+00 -7.98071980e-01
5.11966228e-01 -5.24505019e-01 3.00948322e-02 1.20904759e-01
-6.07448101e-01 -2.75605381e-01 1.29267311e+00 -7.18648136e-01
1.43929970e+00 -2.32069778e+00 -6.80985749e-02 7.06686750e-02
7.22993687e-02 4.23068404e-01 1.30236551e-01 1.66287363e-01
4.09645587e-01 -4.80891503e-02 -1.98634222e-01 -5.36710858e-01
5.92743978e-02 -3.55723411e-01 -8.26223940e-03 6.53124988e-01
1.48742363e-01 7.84387112e-01 -4.79002744e-01 -1.09447694e+00
8.24040651e-01 3.92219484e-01 -2.41007045e-01 1.91880882e-01
-1.71568796e-01 1.76073387e-01 -6.18335903e-01 8.06079984e-01
9.84551489e-01 -2.12051317e-01 -1.65031493e-01 -3.11997026e-01
-4.64514136e-01 -1.56541899e-01 -1.58309758e+00 1.32746029e+00
4.78603924e-03 4.90527272e-01 3.69169533e-01 -9.02860582e-01
9.22639251e-01 4.56569903e-02 4.65870649e-01 -5.70040345e-01
4.87113565e-01 2.60281563e-01 -1.70500726e-01 -3.50178808e-01
6.57547653e-01 1.94768354e-01 4.15250100e-03 1.38470754e-01
-5.13686240e-01 -3.44670504e-01 4.87932354e-01 2.07556859e-01
8.35369527e-01 1.95781849e-02 3.30406517e-01 -1.50383905e-01
9.85590756e-01 1.07834592e-01 5.12256682e-01 5.12844324e-01
-3.36589307e-01 5.81547737e-01 2.87900984e-01 -1.06969059e-01
-9.93659258e-01 -7.22371876e-01 -4.56292361e-01 8.47609460e-01
6.93380177e-01 -3.27293724e-01 -9.63244379e-01 -6.81876242e-01
-9.42379236e-02 3.92451614e-01 -3.86034191e-01 1.30997464e-01
-4.92482454e-01 -8.99978220e-01 4.31750268e-01 5.24711013e-01
1.10396636e+00 -8.58439445e-01 -6.33275151e-01 8.26808736e-02
1.15927625e-02 -1.33120000e+00 -7.14903474e-01 1.16363831e-01
-8.73205662e-01 -9.31704044e-01 -8.42290878e-01 -7.50293434e-01
7.61963129e-01 6.86287165e-01 3.55385929e-01 2.57048368e-01
-3.36414754e-01 -9.47310701e-02 -5.00295103e-01 -2.33563170e-01
-2.30046496e-01 1.71954542e-01 -3.14593136e-01 3.29291224e-01
1.45677581e-01 -1.67405248e-01 -7.61617184e-01 6.44928336e-01
-1.18510890e+00 4.51135844e-01 8.18619072e-01 5.53647578e-01
5.18313587e-01 4.46879059e-01 2.96709657e-01 -6.11553907e-01
2.79314220e-01 1.55915663e-01 -9.42169249e-01 2.69373327e-01
-6.05777860e-01 -2.94110358e-01 7.64728963e-01 -3.83967996e-01
-1.12470376e+00 3.47352564e-01 -3.55323181e-02 -1.39496714e-01
-2.50061750e-01 2.56110460e-01 -4.54388261e-01 -2.62744665e-01
3.99160504e-01 5.88603020e-01 -1.74975365e-01 -3.41121912e-01
2.99095333e-01 9.22007740e-01 3.74551833e-01 -2.10687555e-02
9.92951989e-01 6.91589117e-01 -6.61981013e-03 -1.04959476e+00
-4.29013968e-01 -8.77822936e-01 -8.97650242e-01 -1.33037552e-01
9.40085709e-01 -7.44627833e-01 -4.79423821e-01 1.03504002e+00
-1.00530422e+00 2.97035463e-02 9.28758606e-02 3.07494521e-01
-1.16470188e-01 9.28554177e-01 -5.95448554e-01 -7.91586101e-01
-5.61669707e-01 -1.34695852e+00 1.37665033e+00 7.34135330e-01
3.72368544e-01 -6.91385806e-01 -5.03790617e-01 3.52642238e-01
4.11273539e-01 -4.20311056e-02 7.68484235e-01 -6.54650867e-01
-7.56469786e-01 -5.94813585e-01 -9.01942134e-01 5.66415489e-02
1.86581716e-01 1.75388560e-01 -9.25184548e-01 -2.76256531e-01
-1.40538469e-01 1.15871012e-01 8.90629113e-01 5.22472918e-01
8.76350760e-01 1.50759533e-01 -4.33529556e-01 6.83528543e-01
1.52264535e+00 1.59178793e-01 3.26458603e-01 4.95671660e-01
8.47316384e-01 3.96589994e-01 8.42863679e-01 4.88759130e-01
1.45658150e-01 6.73446715e-01 3.90297920e-01 -4.44755822e-01
-1.63643956e-01 -2.88457479e-02 1.55445889e-01 6.34677112e-01
3.47936243e-01 -2.96022326e-01 -7.55535066e-01 3.07486206e-01
-2.00560522e+00 -5.00247002e-01 -3.12402964e-01 2.16194749e+00
5.83058655e-01 4.10388589e-01 2.42626145e-01 3.47677261e-01
1.13562214e+00 1.54365703e-01 -7.45842636e-01 6.19270205e-02
-2.44900674e-01 -1.72565058e-01 5.80575764e-01 2.68545479e-01
-1.42568004e+00 1.22997487e+00 5.12330151e+00 1.19389749e+00
-1.34128392e+00 -1.64679438e-01 3.56152087e-01 1.97043508e-01
2.46024594e-01 -1.13742463e-01 -1.02953136e+00 4.27287430e-01
1.83613136e-01 -5.98517321e-02 3.28073084e-01 7.66513169e-01
4.58159864e-01 -4.50509608e-01 -7.51291931e-01 9.46254730e-01
1.31515712e-01 -8.47212732e-01 -4.81644422e-02 -2.22755939e-01
5.23449600e-01 -1.83199063e-01 -1.43012553e-01 1.02710560e-01
-2.20583543e-01 -6.04540646e-01 8.60076070e-01 1.47233427e-01
5.17494380e-01 -7.07127452e-01 8.05665910e-01 4.76606071e-01
-1.67235935e+00 -2.16236077e-02 -4.03011948e-01 3.33824396e-01
1.75771013e-01 9.01255012e-01 -8.35274577e-01 5.17097354e-01
6.26613498e-01 7.51619101e-01 -8.97689819e-01 1.38333833e+00
-2.29790196e-01 3.87743175e-01 -6.55068696e-01 -3.56538862e-01
1.42470405e-01 -3.18554282e-01 5.13201952e-01 1.50991380e+00
2.20587939e-01 -1.49170041e-01 4.07527775e-01 8.34504128e-01
1.26132425e-02 5.80933273e-01 -6.82161599e-02 -2.71148030e-02
4.30296838e-01 1.72051513e+00 -1.56878638e+00 -5.30056238e-01
-3.46367359e-01 9.05781806e-01 -1.49985597e-01 1.25315696e-01
-8.17712307e-01 -9.14831281e-01 4.66790125e-02 1.55491695e-01
4.50006276e-01 -2.74667054e-01 -6.90796971e-01 -1.18469739e+00
2.18575284e-01 -6.83338106e-01 2.88402826e-01 -6.43760800e-01
-8.38282347e-01 3.25039506e-01 -2.87539780e-01 -1.14303029e+00
2.87739843e-01 -6.21226430e-01 -7.79839396e-01 8.71961772e-01
-1.61249876e+00 -1.47212481e+00 -7.11011589e-01 5.73044837e-01
8.72807384e-01 2.36592546e-01 1.27412558e-01 3.33389640e-01
-1.13865578e+00 5.79453051e-01 2.27546513e-01 2.03483865e-01
7.55453408e-01 -1.16888869e+00 3.14900190e-01 1.31012869e+00
-1.90070719e-01 2.66006589e-01 5.42086840e-01 -7.90533543e-01
-1.35018098e+00 -1.07695246e+00 4.37035620e-01 1.29528597e-01
6.53863132e-01 -5.00241458e-01 -1.00166440e+00 1.45802543e-01
-1.26249284e-01 -1.62850514e-01 -1.39065823e-02 -3.16258222e-01
1.12565264e-01 -1.87159076e-01 -1.08219719e+00 6.36182547e-01
5.74723005e-01 -1.81869000e-01 -2.85029083e-01 2.44438156e-01
5.01680315e-01 -4.00375456e-01 -4.98064578e-01 4.02433932e-01
4.27053779e-01 -9.47719216e-01 8.02609205e-01 3.50000411e-01
4.50430103e-02 -7.08697140e-01 8.66807848e-02 -6.25625312e-01
-1.92802966e-01 -2.10369572e-01 1.85288414e-01 1.53338945e+00
1.18244022e-01 -6.49921179e-01 8.36493850e-01 2.77568102e-01
-4.82051335e-02 -5.23212135e-01 -9.20168877e-01 -4.51584488e-01
-3.39075893e-01 -4.32560205e-01 4.92681652e-01 5.72402239e-01
-1.34781510e-01 3.55485320e-01 -5.35522662e-02 2.77472526e-01
5.79024494e-01 3.00608426e-01 8.87505352e-01 -1.21246791e+00
-1.22109473e-01 -8.16236854e-01 -4.90010023e-01 -1.31456137e+00
-4.44244415e-01 -5.16611040e-01 4.67123687e-01 -1.73991120e+00
2.06961378e-01 -3.19979370e-01 -7.66173378e-02 2.17977330e-01
-7.01741874e-01 1.88781291e-01 3.00352812e-01 4.12857234e-01
-7.84605742e-01 4.48050231e-01 1.22962737e+00 -1.43712252e-01
-4.32291716e-01 5.49887307e-02 -4.35244739e-01 9.32724118e-01
5.97875118e-01 -2.01682270e-01 -1.21673003e-01 -2.41223663e-01
-4.89844084e-02 -1.63387328e-01 1.37190044e-01 -1.21362817e+00
5.24690866e-01 5.72558027e-04 5.54772615e-01 -1.10625207e+00
2.09929168e-01 -8.26772273e-01 -4.69704479e-01 4.83880848e-01
4.08180691e-02 -2.39794955e-01 2.41010472e-01 5.37438035e-01
-1.45539314e-01 -3.19236785e-01 1.19169915e+00 3.17658514e-01
-6.81744635e-01 1.21834069e-01 -3.00818175e-01 -2.17772752e-01
1.12002003e+00 -5.01704276e-01 -3.31139177e-01 -1.59403551e-02
-2.99452484e-01 3.69311810e-01 7.24291980e-01 1.63880855e-01
5.10645986e-01 -8.71963680e-01 -5.43210626e-01 3.13793600e-01
7.04666302e-02 4.46433157e-01 2.21275374e-01 1.19992101e+00
-7.15329289e-01 4.11255330e-01 1.17743626e-01 -9.29276645e-01
-1.34691584e+00 5.46441138e-01 1.80103019e-01 -4.00501579e-01
-6.52962983e-01 5.19542277e-01 3.15151125e-01 -4.01971191e-02
3.41966242e-01 -5.48722982e-01 -2.84113199e-01 8.28716680e-02
2.80081540e-01 5.47798216e-01 1.09208666e-01 -6.83895469e-01
-2.67778754e-01 9.63274300e-01 -1.27954945e-01 9.64604691e-02
8.27386618e-01 -2.71299303e-01 -3.52086760e-02 1.65476918e-01
7.68962264e-01 1.24849789e-01 -1.18389678e+00 -3.27066809e-01
-1.29207522e-01 -4.08846468e-01 3.93576711e-01 -6.15347683e-01
-9.61405933e-01 1.02319360e+00 6.50907516e-01 3.47159803e-01
1.29031932e+00 -2.33578667e-01 9.03791785e-01 1.94967479e-01
1.93405263e-02 -1.30463719e+00 1.73861668e-01 2.95537859e-01
4.53358740e-01 -1.24255192e+00 2.87171423e-01 -8.04489434e-01
-2.58496732e-01 1.23726976e+00 6.71106100e-01 -3.53416875e-02
3.63101482e-01 2.49356493e-01 -1.27201937e-02 1.42031178e-01
5.41883446e-02 -2.95500100e-01 3.67210299e-01 5.63539937e-02
2.88113862e-01 -4.97760065e-02 -6.10982120e-01 6.41295731e-01
9.86879468e-02 -2.15401798e-01 1.92069918e-01 8.44890475e-01
-9.49392378e-01 -7.94608772e-01 -7.79973745e-01 5.13468504e-01
-2.50067234e-01 -1.11110643e-01 -3.84203076e-01 7.07196534e-01
1.45591766e-01 1.03880382e+00 4.20441553e-02 -3.27064157e-01
4.03151274e-01 -1.84638113e-01 5.02335057e-02 -4.69227642e-01
-5.33315957e-01 7.14492738e-01 -3.80796999e-01 -2.68596888e-01
-1.82034433e-01 -7.19070137e-01 -1.53836906e+00 -2.84077018e-01
-9.64036286e-01 -5.99049069e-02 8.91490936e-01 9.35521603e-01
1.16489194e-01 5.61775148e-01 6.93985224e-01 -9.37428534e-01
-3.92240494e-01 -1.11714196e+00 -6.46443188e-01 3.78712833e-01
3.28569417e-03 -4.36291158e-01 -3.33893299e-01 -3.30794826e-02] | [12.061545372009277, 2.254253387451172] |
99a3600f-4032-41bb-9d29-a4ef731d736b | dynamic-advisor-based-ensemble-dynabe-case | 1805.12111 | null | http://arxiv.org/abs/1805.12111v4 | http://arxiv.org/pdf/1805.12111v4.pdf | Dynamic Advisor-Based Ensemble (dynABE): Case study in stock trend prediction of critical metal companies | Stock trend prediction is a challenging task due to the market's noise, and
machine learning techniques have recently been successful in coping with this
challenge. In this research, we create a novel framework for stock prediction,
Dynamic Advisor-Based Ensemble (dynABE). dynABE explores domain-specific areas
based on the companies of interest, diversifies the feature set by creating
different "advisors" that each handles a different area, follows an effective
model ensemble procedure for each advisor, and combines the advisors together
in a second-level ensemble through an online update strategy we developed.
dynABE is able to adapt to price pattern changes of the market during the
active trading period robustly, without needing to retrain the entire model. We
test dynABE on three cobalt-related companies, and it achieves the best-case
misclassification error of 31.12% and an annualized absolute return of 359.55%
with zero maximum drawdown. dynABE also consistently outperforms the baseline
models of support vector machine, neural network, and random forest in all case
studies. | ['Zhengyang Dong'] | 2018-05-24 | null | null | null | null | ['stock-trend-prediction', 'stock-prediction'] | ['time-series', 'time-series'] | [-5.79581439e-01 -3.34328651e-01 -7.70886764e-02 -3.32253724e-01
-2.68369943e-01 -7.92490065e-01 6.78861976e-01 -1.57370120e-01
-2.75658578e-01 9.77249563e-01 -1.54965287e-02 -5.62701881e-01
-1.02845831e-02 -1.04071641e+00 -3.97284716e-01 -6.41359746e-01
-1.22566260e-01 7.96127737e-01 4.49380785e-01 -4.54999298e-01
6.97561979e-01 4.60961074e-01 -1.39991415e+00 3.28878909e-01
8.67088079e-01 1.49748969e+00 -4.29467052e-01 2.58416206e-01
-3.50635648e-01 1.00343502e+00 -7.09357858e-01 -7.33758867e-01
8.54906261e-01 -1.82192847e-01 -1.31207809e-01 -3.01835030e-01
3.35374296e-01 -4.54982907e-01 4.49272394e-02 7.25748122e-01
2.10426494e-01 -2.76975483e-02 5.75943232e-01 -1.19754338e+00
-1.31348118e-01 1.00837982e+00 -6.02716565e-01 6.55045807e-01
-3.64749104e-01 1.09418526e-01 1.10411537e+00 -9.18450594e-01
4.36420202e-01 7.35067129e-01 9.52503860e-01 4.92865629e-02
-1.33494508e+00 -1.06403983e+00 5.20898581e-01 2.10977085e-02
-7.42030919e-01 3.14113162e-02 5.48860967e-01 -7.56478548e-01
1.11696184e+00 3.59567344e-01 9.61306572e-01 6.30644262e-01
7.46569872e-01 5.06665826e-01 1.40046430e+00 -1.22016966e-02
4.04541999e-01 3.50478113e-01 3.24567676e-01 5.96279930e-03
6.51690722e-01 3.07473212e-01 -5.78892112e-01 -4.76158708e-01
4.97876853e-01 3.63660127e-01 7.30905905e-02 -3.18026617e-02
-1.06703615e+00 1.08569229e+00 2.00257525e-01 1.73751831e-01
-8.08201969e-01 -1.42990738e-01 3.14356208e-01 7.88186014e-01
8.89162302e-01 6.89871252e-01 -9.13754582e-01 -7.63486102e-02
-1.40290046e+00 7.61135936e-01 1.34336698e+00 4.68022853e-01
4.04539078e-01 7.11999178e-01 -4.67461757e-02 5.08924663e-01
3.72966155e-02 4.94828820e-01 8.50671589e-01 -5.28977394e-01
4.74382818e-01 7.45151699e-01 2.78291970e-01 -8.06257784e-01
-5.64846635e-01 -1.14664447e+00 -7.59523094e-01 7.44881749e-01
4.19545263e-01 -3.94145519e-01 -8.56537104e-01 1.06611371e+00
1.18730232e-01 1.31372586e-01 1.37579158e-01 2.70406008e-01
8.83109272e-02 6.83512986e-01 -1.32159919e-01 -4.39482272e-01
9.48096871e-01 -1.09100688e+00 -5.26325405e-01 -2.11532027e-01
2.33966365e-01 -5.44982970e-01 1.51287585e-01 8.16617489e-01
-9.17208076e-01 -3.57498437e-01 -1.12847519e+00 1.02606702e+00
-6.31361783e-01 -2.23581210e-01 5.02724528e-01 5.72064698e-01
-7.46180415e-01 1.08152473e+00 -7.37900674e-01 6.01278067e-01
7.70831928e-02 3.81020337e-01 2.39304110e-01 5.88744283e-01
-1.20804775e+00 1.30640483e+00 4.71093595e-01 2.22693667e-01
-2.69760221e-01 -9.16242659e-01 -2.39153638e-01 8.28127265e-02
3.22564006e-01 -6.28130853e-01 1.33659506e+00 -1.03425837e+00
-1.40813398e+00 1.10867299e-01 4.45241451e-01 -1.21894324e+00
9.82702255e-01 -3.06214213e-01 -8.36249173e-01 -5.31555235e-01
-2.07914770e-01 -4.21258956e-02 7.97105610e-01 -8.90204966e-01
-1.13186562e+00 -5.95011830e-01 -5.11202097e-01 -2.01957673e-01
6.17273897e-02 -2.11178996e-02 3.71901125e-01 -1.30912268e+00
1.83572188e-01 -9.41994131e-01 -4.10466582e-01 -8.98170412e-01
-4.55701724e-02 -3.60911596e-03 5.65612376e-01 -1.00681508e+00
1.57842517e+00 -1.60092950e+00 -3.55941564e-01 6.41248345e-01
-1.66301444e-01 1.54772699e-01 4.43423897e-01 5.99578559e-01
-2.08737612e-01 -7.60633126e-02 -4.08171922e-01 3.51442210e-02
1.17786704e-02 1.02117524e-01 -9.16647136e-01 -7.76745752e-02
-9.35918465e-02 7.50424922e-01 -4.17892277e-01 2.06689194e-01
3.15126143e-02 -1.83036417e-01 -2.06741199e-01 -9.12271366e-02
-3.10115159e-01 -3.85909118e-02 -1.13838568e-01 8.10355067e-01
8.28458190e-01 -1.64510030e-02 1.34926960e-01 2.99491316e-01
-4.86006647e-01 1.88816592e-01 -1.58430719e+00 5.53698421e-01
-2.67005265e-01 2.53942311e-01 -2.79414505e-01 -9.14487839e-01
1.38558376e+00 1.04096886e-02 3.88923854e-01 -4.74140525e-01
-3.50232810e-01 8.77448916e-01 1.22833267e-01 1.38067082e-01
4.49467629e-01 -2.35980242e-01 -3.28786038e-02 7.09937513e-01
-1.79826856e-01 1.67367071e-01 1.63831741e-01 -2.79843658e-01
8.98533821e-01 8.49027038e-02 5.78545451e-01 -3.17258745e-01
2.91792601e-01 1.22081690e-01 9.24334586e-01 6.29954755e-01
7.79498741e-02 2.12138012e-01 4.04830873e-01 -1.04959822e+00
-7.89672494e-01 -8.95517349e-01 -2.20729828e-01 1.00363338e+00
-3.74420404e-01 1.89158648e-01 -3.10532451e-01 -8.83220375e-01
9.56486225e-01 1.28227782e+00 -6.87294066e-01 2.26381078e-01
-5.63993037e-01 -1.25130475e+00 5.71753122e-02 6.36478782e-01
6.38616860e-01 -1.16189194e+00 -6.79993093e-01 8.56923103e-01
5.32723427e-01 -4.36675102e-01 -1.77890360e-01 4.65883434e-01
-1.19118643e+00 -8.57946575e-01 -7.33505785e-01 -1.03167243e-01
2.35762313e-01 -2.93178678e-01 1.37924278e+00 -3.45059633e-01
2.97703415e-01 -1.27010122e-01 -2.02721909e-01 -1.08312559e+00
-3.23983788e-01 1.97880358e-01 -2.67688334e-02 3.31318885e-01
3.59331459e-01 -5.90512991e-01 -4.57935452e-01 5.54641962e-01
-4.60734963e-01 -3.54170203e-01 5.45549989e-01 1.04679036e+00
4.75098401e-01 2.57337362e-01 1.21636558e+00 -1.07487655e+00
4.74432766e-01 -7.30324447e-01 -1.19812644e+00 2.78609872e-01
-1.49085867e+00 1.06738530e-01 2.44302660e-01 -3.28517765e-01
-1.16407049e+00 -2.48201042e-02 1.70775756e-01 -8.56798887e-02
2.83320457e-01 8.90418530e-01 4.09991622e-01 2.19906867e-01
3.92512918e-01 5.34064174e-01 1.38098001e-01 -7.15848267e-01
-1.90680161e-01 4.55956280e-01 5.99960387e-01 5.00637628e-02
9.34389055e-01 1.23960748e-01 -1.12458795e-01 -2.18548656e-01
-5.94536304e-01 -2.11638436e-01 -5.68373799e-01 -2.90723026e-01
1.14722513e-01 -9.63482797e-01 -3.03198159e-01 1.10635841e+00
-6.22694790e-01 -2.20014259e-01 -3.09544832e-01 5.75065792e-01
-2.11097822e-01 -2.12622195e-01 -4.82956290e-01 -1.19397879e+00
-8.31977904e-01 -7.27369487e-01 1.89536512e-01 1.78572446e-01
-3.64798754e-01 -1.03853595e+00 2.75019497e-01 -2.50007138e-02
8.16201568e-01 6.56001687e-01 5.52340508e-01 -1.58906209e+00
-5.14413059e-01 -5.18847048e-01 2.42126852e-01 6.75600529e-01
-2.69042067e-02 9.69733745e-02 -7.23676741e-01 -2.40982026e-01
1.67256892e-01 2.84172624e-01 1.21235895e+00 4.39041972e-01
5.81049621e-01 -4.58038121e-01 -1.97041072e-02 3.42254847e-01
1.21892357e+00 7.09454417e-01 2.92239785e-01 1.08930814e+00
-1.15570733e-02 6.48601174e-01 6.73310161e-01 6.74650133e-01
3.96578163e-01 5.80262959e-01 4.14168626e-01 3.33248794e-01
6.86479747e-01 -4.31821235e-02 5.33240974e-01 5.49874246e-01
-4.92203891e-01 2.16822937e-01 -8.00852478e-01 1.50623739e-01
-1.98684156e+00 -1.24588072e+00 -1.11237697e-01 2.17239571e+00
6.83757305e-01 6.85432673e-01 6.09189093e-01 1.35794461e-01
4.27516758e-01 1.23189539e-01 -8.15562844e-01 -3.42340529e-01
-1.42187610e-01 5.42697981e-02 9.98471856e-01 1.06529042e-01
-1.16462278e+00 4.79667664e-01 6.37737322e+00 4.53815132e-01
-1.16784716e+00 -2.68794596e-01 7.91417241e-01 -9.59097147e-02
-2.72675127e-01 -3.75855118e-02 -1.20100832e+00 9.06147182e-01
1.12059534e+00 -5.97948432e-01 1.47802323e-01 1.20012140e+00
-1.51651904e-01 -6.05984516e-02 -7.30068684e-01 4.96734560e-01
-2.02771932e-01 -1.77709258e+00 -8.53167996e-02 3.01306605e-01
9.64624941e-01 1.56114995e-01 4.74206060e-02 5.67114651e-01
4.69520152e-01 -5.72520018e-01 1.01658702e+00 1.01202703e+00
2.78397487e-03 -1.07751298e+00 1.28185129e+00 5.72970450e-01
-1.02798140e+00 -5.71701884e-01 -3.13717276e-02 -1.69347733e-01
1.43103316e-01 7.86678791e-01 -6.46285713e-01 7.09705055e-01
8.81185949e-01 4.93351400e-01 -4.28337932e-01 1.08444667e+00
2.20862776e-01 6.89598799e-01 -4.49783593e-01 -2.47071199e-02
2.07792446e-01 -4.26112711e-01 5.46947241e-01 9.40623403e-01
6.90228999e-01 -8.98406208e-02 2.11716127e-02 5.50762713e-01
2.62765914e-01 9.64711457e-02 -3.16813260e-01 2.81738460e-01
4.18968350e-01 8.46558690e-01 -6.07859612e-01 -5.01315355e-01
-4.01591301e-01 4.16813433e-01 -3.24646205e-01 1.83779478e-01
-4.97278750e-01 -3.87658626e-01 3.80989224e-01 2.09372953e-01
6.67523086e-01 1.01086497e-01 -6.55159950e-01 -1.09807658e+00
9.32588205e-02 -1.01760304e+00 6.70425475e-01 -3.72230887e-01
-1.46277595e+00 7.80227125e-01 1.59212127e-02 -1.53287601e+00
-7.28463709e-01 -7.64541745e-01 -1.06793559e+00 1.00923574e+00
-1.43791580e+00 -7.23032117e-01 1.57360420e-01 2.57099420e-01
5.68754077e-01 -9.93669450e-01 6.02151573e-01 -2.19246894e-01
-5.64710617e-01 4.73634899e-01 6.47480965e-01 -1.17713384e-01
4.24976408e-01 -1.71454895e+00 9.02406335e-01 6.64792895e-01
1.34096265e-01 3.64365280e-01 6.50437474e-01 -1.09271586e+00
-6.85866892e-01 -1.08788371e+00 1.03300226e+00 -4.32150841e-01
1.14053798e+00 7.28007331e-02 -1.18601501e+00 7.96756566e-01
2.07477033e-01 -2.54692137e-01 5.96123934e-01 9.51054767e-02
-4.29193407e-01 -6.14545047e-01 -1.20514667e+00 1.98394254e-01
3.31622928e-01 2.08194494e-01 -8.27964723e-01 -1.07403435e-01
2.80251414e-01 -4.19063091e-01 -1.08666706e+00 3.78832340e-01
8.35417747e-01 -1.30545104e+00 6.96854413e-01 -4.36917156e-01
1.52045339e-02 -5.31986393e-02 1.33382501e-02 -1.69583154e+00
-3.88067663e-01 -6.10975921e-01 -5.69729447e-01 1.23079348e+00
9.34868753e-01 -1.51643550e+00 6.82394326e-01 7.46434093e-01
-2.83976644e-02 -9.32022631e-01 -8.39680314e-01 -9.80359137e-01
1.50176600e-01 -2.14129657e-01 1.22382903e+00 8.03731680e-01
-2.26282462e-01 -6.44947737e-02 -4.93336141e-01 -3.19954418e-02
6.31526709e-01 7.30085969e-01 8.64850819e-01 -1.62837923e+00
-4.93446976e-01 -8.39504421e-01 -1.28255367e-01 -3.96886647e-01
-6.01212196e-02 -6.09755099e-01 -5.31010151e-01 -1.00757933e+00
-1.76130623e-01 -3.50516021e-01 -8.18821669e-01 4.49866980e-01
-1.70551807e-01 -2.23857805e-01 3.35298389e-01 5.14157414e-01
1.00206487e-01 2.29125097e-01 7.51502752e-01 -1.77416936e-01
-4.24087942e-01 1.01325190e+00 -6.80463016e-01 7.70580411e-01
9.99256790e-01 -3.30315799e-01 4.15072031e-02 3.56079757e-01
2.80432731e-01 6.47030696e-02 1.46154299e-01 -8.48364294e-01
1.43810675e-01 -3.14879805e-01 7.91388512e-01 -1.31828654e+00
-1.44500747e-01 -7.20826566e-01 8.21263433e-01 9.51845109e-01
-7.12007210e-02 7.02843606e-01 2.92822331e-01 6.20613217e-01
-4.60482508e-01 -3.18552136e-01 6.65154159e-01 -3.05159017e-02
-5.88223398e-01 1.92694724e-01 -2.92077333e-01 -3.58485103e-01
1.17288327e+00 -2.34845847e-01 -2.51544446e-01 -2.72254914e-01
-8.58860373e-01 5.53773880e-01 1.00577854e-01 2.58226514e-01
1.36914685e-01 -1.20292115e+00 -1.06999910e+00 3.45906943e-01
-1.28309578e-01 -4.98760343e-01 9.52495933e-02 6.60688162e-01
-4.66956437e-01 4.08106267e-01 -2.72949070e-01 -2.62155496e-02
-8.29254985e-01 3.30602229e-01 6.39288723e-01 -9.59383488e-01
-4.78996009e-01 5.66574812e-01 -3.66606891e-01 -2.31583059e-01
8.14087689e-02 -4.42757905e-01 -4.58691478e-01 8.09613347e-01
5.43639719e-01 8.44767869e-01 4.19993699e-01 -2.95477629e-01
-1.20455474e-01 3.51900369e-01 -3.09714019e-01 -1.64516672e-01
1.80110407e+00 2.52780318e-01 -2.94845905e-02 7.44442999e-01
3.81237298e-01 9.36812442e-03 -1.30375981e+00 -3.12768459e-01
7.82367170e-01 -2.57740796e-01 -6.34075180e-02 -1.24865973e+00
-1.31649792e+00 3.45289379e-01 3.82864088e-01 7.63402104e-01
9.99973953e-01 -6.90373421e-01 6.96914792e-01 3.05035084e-01
3.45452428e-01 -1.40750778e+00 -4.50008005e-01 4.84084308e-01
1.18003130e+00 -1.03439188e+00 2.41695970e-01 3.00640553e-01
-1.01589894e+00 1.19864869e+00 4.68546808e-01 -3.30161899e-01
1.09038436e+00 2.56890744e-01 2.26600766e-01 8.09793770e-02
-1.25169778e+00 3.91610444e-01 2.54563034e-01 2.44896397e-01
-6.89013675e-02 2.03059733e-01 -1.67309433e-01 1.07899904e+00
-6.06946886e-01 -6.66397139e-02 3.64503592e-01 8.95589054e-01
-5.67032635e-01 -9.80417013e-01 -5.18527627e-01 1.09987831e+00
-7.87948549e-01 5.82444062e-03 -2.51548022e-01 1.05145216e+00
-2.28276640e-01 3.92493635e-01 4.58185852e-01 -4.31728035e-01
7.49177873e-01 5.96623957e-01 -1.56650677e-01 -2.81005800e-01
-1.17164814e+00 2.91333973e-01 2.12079525e-01 -2.47375354e-01
-7.94396251e-02 -1.37664461e+00 -8.88137400e-01 -4.32729423e-01
-3.37897837e-01 2.58946508e-01 5.48649788e-01 7.66522348e-01
2.40118891e-01 4.08736616e-01 1.22049689e+00 -7.35970736e-01
-1.48145556e+00 -1.15050399e+00 -1.01890743e+00 -6.92871260e-03
3.78494412e-01 -7.38958359e-01 -7.35439181e-01 -1.14602149e-01] | [4.5749382972717285, 4.160758972167969] |
4e4de728-1f55-43b3-9f97-5b2f804ec91f | motif-based-graph-self-supervised-learning | 2110.00987 | null | https://arxiv.org/abs/2110.00987v2 | https://arxiv.org/pdf/2110.00987v2.pdf | Motif-based Graph Self-Supervised Learning for Molecular Property Prediction | Predicting molecular properties with data-driven methods has drawn much attention in recent years. Particularly, Graph Neural Networks (GNNs) have demonstrated remarkable success in various molecular generation and prediction tasks. In cases where labeled data is scarce, GNNs can be pre-trained on unlabeled molecular data to first learn the general semantic and structural information before being fine-tuned for specific tasks. However, most existing self-supervised pre-training frameworks for GNNs only focus on node-level or graph-level tasks. These approaches cannot capture the rich information in subgraphs or graph motifs. For example, functional groups (frequently-occurred subgraphs in molecular graphs) often carry indicative information about the molecular properties. To bridge this gap, we propose Motif-based Graph Self-supervised Learning (MGSSL) by introducing a novel self-supervised motif generation framework for GNNs. First, for motif extraction from molecular graphs, we design a molecule fragmentation method that leverages a retrosynthesis-based algorithm BRICS and additional rules for controlling the size of motif vocabulary. Second, we design a general motif-based generative pre-training framework in which GNNs are asked to make topological and label predictions. This generative framework can be implemented in two different ways, i.e., breadth-first or depth-first. Finally, to take the multi-scale information in molecular graphs into consideration, we introduce a multi-level self-supervised pre-training. Extensive experiments on various downstream benchmark tasks show that our methods outperform all state-of-the-art baselines. | ['Chee-Kong Lee', 'Chengqiang Lu', 'Hao Wang', 'Qi Liu', 'Zaixi Zhang'] | 2021-10-03 | null | http://proceedings.neurips.cc/paper/2021/hash/85267d349a5e647ff0a9edcb5ffd1e02-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/85267d349a5e647ff0a9edcb5ffd1e02-Paper.pdf | neurips-2021-12 | ['retrosynthesis'] | ['medical'] | [ 6.20032489e-01 5.36637120e-02 -6.55830204e-01 -3.38364929e-01
-5.05785108e-01 -7.33646333e-01 5.80149293e-01 4.47322786e-01
7.54285082e-02 1.01023877e+00 7.11450279e-02 -6.34601533e-01
-4.62502465e-02 -1.30366051e+00 -1.15461397e+00 -8.29398990e-01
-1.60468131e-01 3.03931713e-01 2.15483934e-01 -3.57217163e-01
4.39650744e-01 4.29809660e-01 -1.04701579e+00 4.24941927e-01
1.17991710e+00 5.88332176e-01 3.49806756e-01 3.74150515e-01
-2.66777098e-01 7.31557190e-01 -5.44167049e-02 -2.81977624e-01
-2.92500108e-02 -6.73543453e-01 -7.47011721e-01 7.51686702e-03
2.20375031e-01 1.12987064e-01 -4.30264980e-01 8.77879441e-01
5.00973523e-01 1.96947083e-01 8.98114681e-01 -9.81907487e-01
-5.47473192e-01 7.41715014e-01 -4.67455000e-01 2.28432957e-02
4.77820992e-01 3.99316221e-01 1.19989681e+00 -9.59554076e-01
9.25765753e-01 1.13645434e+00 5.35838068e-01 4.91134107e-01
-1.58726811e+00 -6.46833479e-01 3.37963611e-01 1.67382449e-01
-1.01836777e+00 -2.34234944e-01 1.04080975e+00 -4.65359986e-01
1.03309166e+00 -2.54053064e-02 6.32003546e-01 1.37865973e+00
2.58242577e-01 4.69716460e-01 9.23735857e-01 -2.45682001e-01
3.33399713e-01 -6.10244632e-01 -1.28823727e-01 1.07013798e+00
2.59080142e-01 5.83721846e-02 -5.96103013e-01 -2.48873547e-01
7.45762944e-01 2.71776527e-01 -1.04458652e-01 -5.18493295e-01
-1.23131502e+00 1.04481661e+00 8.06374371e-01 2.17342630e-01
-1.89828917e-01 9.56022963e-02 3.70733678e-01 2.20692828e-01
4.81379658e-01 6.79260373e-01 -3.38142335e-01 2.73584336e-01
-9.16730702e-01 1.00525022e-01 6.95490241e-01 9.61030900e-01
1.08240354e+00 -8.73402506e-02 -9.77715030e-02 6.50088489e-01
3.81647706e-01 -1.03897266e-01 3.53229642e-01 -9.00509432e-02
5.23066640e-01 8.56421947e-01 -3.89867663e-01 -1.02500892e+00
-5.52823067e-01 -3.36704761e-01 -1.00118053e+00 -3.72435480e-01
1.95983574e-01 1.41654745e-01 -1.26715720e+00 1.77201271e+00
4.42916185e-01 2.07443878e-01 -5.48772030e-02 7.22571552e-01
1.19622052e+00 8.12928736e-01 3.60925376e-01 -1.51165932e-01
1.09047449e+00 -1.05304098e+00 -2.10454494e-01 -6.63766861e-02
8.60016286e-01 -4.96037334e-01 9.04788613e-01 2.14349434e-01
-6.16064548e-01 -5.13457060e-01 -1.16507804e+00 4.86729890e-02
-7.00175941e-01 -2.15747133e-01 1.07289362e+00 3.39105844e-01
-6.39602423e-01 1.03851509e+00 -7.86572754e-01 -2.95399278e-01
4.90570158e-01 4.94236678e-01 -4.31100279e-01 -1.94508344e-01
-1.20718372e+00 3.55671793e-01 7.64395058e-01 -4.38046679e-02
-1.19481814e+00 -7.38638997e-01 -9.86868501e-01 -1.47696197e-01
5.72733164e-01 -9.25289214e-01 7.72845149e-01 -8.06405306e-01
-1.50133693e+00 7.43928790e-01 -1.15931287e-01 -2.75326341e-01
1.51135072e-01 3.72251213e-01 -3.57120007e-01 1.49121851e-01
2.07202584e-01 7.37959564e-01 6.83938205e-01 -9.86073375e-01
-1.73055932e-01 -2.02430978e-01 9.00491551e-02 5.88066550e-03
-5.59961125e-02 -4.79016125e-01 -2.10122108e-01 -8.53090346e-01
1.35505036e-01 -9.10571992e-01 -5.74239850e-01 -5.20249248e-01
-9.21723604e-01 -2.48847127e-01 4.33169633e-01 -9.85839218e-03
1.26685047e+00 -1.72691691e+00 3.94359708e-01 5.29104531e-01
4.81734991e-01 1.18618585e-01 -3.59692246e-01 9.48447943e-01
-4.38184828e-01 2.76745409e-01 -2.03078568e-01 7.80362412e-02
-6.79370835e-02 8.26607496e-02 -1.89010561e-01 1.39777705e-01
5.11709452e-01 1.14784229e+00 -1.25255215e+00 -4.23159778e-01
7.68606439e-02 2.09147483e-01 -7.86951661e-01 2.63476640e-01
-9.11289334e-01 7.03043342e-01 -7.11140513e-01 9.67450202e-01
4.57549423e-01 -7.05472469e-01 6.00638747e-01 -3.90663415e-01
-2.78029572e-02 5.19809067e-01 -6.81497335e-01 1.99046886e+00
-1.87771022e-01 4.56389152e-02 -5.74482322e-01 -1.35708523e+00
8.39294612e-01 3.70238982e-02 4.60893184e-01 -4.79083478e-01
-1.92923814e-01 2.91877300e-01 1.11933500e-01 -2.51007676e-01
2.28011593e-01 -2.27936968e-01 6.55000657e-02 4.48548138e-01
2.88032055e-01 1.54970244e-01 5.10269880e-01 4.37563032e-01
1.25721467e+00 4.50904071e-01 4.72156137e-01 -9.94275287e-02
5.53586245e-01 1.29999951e-01 6.36215627e-01 5.34709692e-01
3.30199420e-01 4.53718662e-01 7.24944472e-01 -5.33266008e-01
-1.01039565e+00 -8.66271853e-01 1.25628769e-01 1.41029656e+00
6.49536476e-02 -9.29620385e-01 -6.54060423e-01 -9.60302413e-01
-3.40845138e-02 7.50879943e-02 -6.77614808e-01 -4.34243351e-01
-4.15256709e-01 -1.07570517e+00 3.65222245e-01 4.14576918e-01
2.40080506e-01 -1.16419911e+00 2.46332169e-01 5.06325781e-01
1.56871140e-01 -9.03166473e-01 -5.86518705e-01 5.38836241e-01
-1.03850710e+00 -1.27918279e+00 -4.12524641e-01 -9.04339254e-01
1.01088560e+00 3.09223235e-01 1.19259155e+00 1.29695788e-01
-2.64924735e-01 -1.62600905e-01 -3.95996690e-01 -6.71486408e-02
-3.27865869e-01 5.97546220e-01 -1.09598339e-01 -7.39434659e-02
1.90199018e-01 -1.00598395e+00 -9.04960811e-01 2.84827650e-01
-8.78385782e-01 3.65233421e-01 9.45186913e-01 9.40049767e-01
1.02367401e+00 -2.03410506e-01 8.65735233e-01 -1.26707494e+00
5.44646800e-01 -6.17944777e-01 -4.40668672e-01 2.73154736e-01
-6.49374068e-01 4.24583912e-01 1.02870607e+00 -4.89486337e-01
-5.50602674e-01 2.81441450e-01 -1.45608053e-01 -2.15059087e-01
-1.23152733e-01 9.33162510e-01 -5.63708365e-01 -2.33748287e-01
6.11755848e-01 4.48346585e-01 -1.18516549e-01 -2.38718316e-01
5.70551217e-01 2.95922488e-01 2.09685326e-01 -9.51472819e-01
7.67904997e-01 1.71911478e-01 4.91923630e-01 -6.73520684e-01
-8.84690881e-01 -4.05676812e-01 -7.13572264e-01 2.00963765e-01
7.20535636e-01 -8.17289412e-01 -7.00284719e-01 2.10720003e-01
-8.81057203e-01 -4.72187996e-01 1.31144494e-01 1.57627851e-01
-6.61113858e-01 5.44998825e-01 -6.39046431e-01 -2.99974024e-01
-4.18917924e-01 -1.31697083e+00 1.12387526e+00 7.42269084e-02
-6.51285499e-02 -1.17962122e+00 3.49130064e-01 2.29152471e-01
2.01265160e-02 6.12953186e-01 1.35449529e+00 -9.17354167e-01
-8.09199154e-01 1.12080432e-01 -1.75213993e-01 -2.34178111e-01
4.96071070e-01 -7.69427493e-02 -5.62307358e-01 -2.41123721e-01
-8.57903898e-01 -5.52224278e-01 1.19617248e+00 1.77964330e-01
1.30893028e+00 -3.70550513e-01 -7.75022924e-01 8.32018673e-01
1.21564639e+00 2.33506393e-02 5.12396932e-01 2.37081379e-01
1.18943059e+00 3.21369171e-01 4.12868589e-01 2.31441885e-01
3.48811418e-01 5.51679254e-01 4.18286890e-01 -2.36352146e-01
8.54926184e-02 -9.61422145e-01 3.33872467e-01 6.05065525e-01
-2.77908504e-01 -3.38407606e-01 -7.22538173e-01 4.44242731e-02
-2.05920720e+00 -9.90146816e-01 -2.79317070e-02 2.00955582e+00
1.20625579e+00 1.79335952e-01 4.50036705e-01 -1.04068317e-01
6.46786749e-01 3.04272532e-01 -8.63796830e-01 -5.72848991e-02
-3.11530679e-02 3.21307093e-01 3.18390906e-01 2.69670844e-01
-1.19644892e+00 1.35751355e+00 5.76081562e+00 1.03325617e+00
-1.16750443e+00 -3.23717088e-01 6.66281044e-01 3.52026433e-01
-5.33494890e-01 3.18742067e-01 -8.11415672e-01 5.81966400e-01
8.80352736e-01 6.06184043e-02 3.27329218e-01 8.28122258e-01
2.51914799e-01 3.11922073e-01 -1.34119225e+00 8.81273270e-01
-1.97525933e-01 -1.96572888e+00 5.36423862e-01 3.27811837e-01
7.40566552e-01 -4.82347235e-02 -1.27856120e-01 2.07969859e-01
4.57443237e-01 -1.26505077e+00 1.80163443e-01 5.29786110e-01
8.48959506e-01 -9.20978904e-01 2.47565120e-01 2.10075110e-01
-1.46208227e+00 3.49593669e-01 -4.97191787e-01 1.71928644e-01
-8.29945207e-02 8.06593835e-01 -8.76646399e-01 8.85361850e-01
1.02761686e-02 1.17485487e+00 -5.04558086e-01 7.28268743e-01
-3.63281816e-01 6.93019688e-01 -7.16092661e-02 -1.20008759e-01
4.35038537e-01 -5.57206452e-01 2.57140338e-01 1.31826329e+00
1.15506025e-02 7.91648701e-02 5.05713344e-01 1.12943363e+00
-4.19268548e-01 3.14807057e-01 -8.29139233e-01 -5.43797612e-01
2.91948289e-01 1.43314219e+00 -9.13804531e-01 -1.87613621e-01
-3.38270485e-01 9.18689191e-01 5.46172738e-01 3.71051490e-01
-8.09661925e-01 -4.95902985e-01 3.96100998e-01 3.19862992e-01
3.05850506e-01 -2.38412738e-01 2.26019070e-01 -1.26721025e+00
-3.90024334e-01 -1.01481545e+00 4.70167637e-01 -3.88473094e-01
-1.43414009e+00 3.49980831e-01 -3.10447723e-01 -1.08892918e+00
-7.71012008e-02 -8.82559717e-01 -7.83629417e-01 4.69782829e-01
-1.45361388e+00 -1.35119104e+00 -1.56718686e-01 3.34205776e-01
3.95316571e-01 -2.71070361e-01 7.16746390e-01 1.39014259e-01
-9.21161592e-01 5.23760736e-01 2.67932862e-02 1.26487896e-01
7.04604983e-01 -1.31793797e+00 4.99107927e-01 5.46221077e-01
2.58079112e-01 1.04671633e+00 3.29130590e-01 -9.24937427e-01
-1.72156310e+00 -1.45390904e+00 4.59980458e-01 -2.87992924e-01
7.92014241e-01 -6.73598945e-01 -8.29054177e-01 4.83378679e-01
-3.04943651e-01 1.31763220e-01 1.14518178e+00 2.47593477e-01
-5.54864347e-01 1.03688449e-01 -6.92441463e-01 5.90747714e-01
1.57325733e+00 -6.40444577e-01 -2.00126231e-01 8.80714536e-01
7.61954725e-01 -3.39367330e-01 -9.62302625e-01 5.37319303e-01
3.36804509e-01 -8.20038974e-01 1.02377796e+00 -1.15447414e+00
7.07985759e-01 -3.80113572e-01 1.56854279e-02 -1.31151676e+00
-4.50490773e-01 -9.88062322e-01 -2.35734701e-01 1.06094313e+00
7.44403780e-01 -5.35922647e-01 9.91199374e-01 -2.90898457e-02
-3.11931312e-01 -9.66040611e-01 -3.88168067e-01 -6.90306544e-01
-2.73931194e-02 -1.18201301e-01 5.70806623e-01 1.04606402e+00
2.74084985e-01 9.17558789e-01 -2.86074579e-01 -2.09610060e-01
4.80258852e-01 5.53053141e-01 8.89801443e-01 -1.23465443e+00
-4.19577211e-01 -4.61997181e-01 -4.98778522e-01 -1.14865029e+00
3.96523505e-01 -1.46589780e+00 -1.84542269e-01 -1.48649704e+00
5.63965619e-01 -2.84962118e-01 -3.37564707e-01 5.64065695e-01
-3.60355645e-01 1.95481703e-02 -3.59203279e-01 3.27350438e-01
-8.40370774e-01 7.04889297e-01 1.45321727e+00 -2.38735870e-01
-2.77869403e-01 -2.39828005e-01 -7.98308611e-01 4.66619521e-01
7.24281609e-01 -2.77830184e-01 -5.22237480e-01 2.54828215e-01
5.54441035e-01 -1.03933074e-01 3.04663122e-01 -7.41509259e-01
1.03338972e-01 -4.93880391e-01 4.11606222e-01 -6.09727502e-01
-1.20609418e-01 -2.95848429e-01 7.59280622e-02 4.59746927e-01
-2.67511487e-01 -2.35180721e-01 -8.23304132e-02 1.06447446e+00
-1.36720240e-01 8.79817978e-02 5.91729045e-01 -2.78717071e-01
-4.95809168e-01 8.66411924e-01 -1.76016346e-01 -1.51474729e-01
8.37813497e-01 -3.07821691e-01 -3.85952413e-01 -1.69041231e-01
-6.89933717e-01 1.31015524e-01 6.20979071e-01 9.11935493e-02
6.15564644e-01 -1.31772316e+00 -2.21350074e-01 1.19123653e-01
4.92297053e-01 2.37753958e-01 -4.16411050e-02 7.69718647e-01
-4.81189996e-01 6.26787484e-01 -6.22896776e-02 -4.52333421e-01
-8.94016683e-01 8.48219693e-01 1.87633723e-01 -5.38372815e-01
-2.92985797e-01 6.41379476e-01 4.89750326e-01 -5.95575452e-01
-2.41993383e-01 -3.49359483e-01 -2.38044262e-02 2.28706514e-03
1.54064208e-01 -8.77156407e-02 6.17389679e-02 -3.17978799e-01
-3.03758323e-01 4.96986568e-01 -3.64722550e-01 5.26640296e-01
1.42632604e+00 4.59623098e-01 -2.37289667e-01 1.37991294e-01
1.09819949e+00 -7.94290230e-02 -1.16133201e+00 -2.13712931e-01
1.46052793e-01 6.35275543e-02 -2.26834163e-01 -5.32941997e-01
-8.53001058e-01 8.12933505e-01 -4.53063510e-02 -5.36834002e-02
8.68113816e-01 2.80788913e-02 7.95383394e-01 6.10534072e-01
3.10527980e-01 -8.63648415e-01 5.73738396e-01 3.93236011e-01
5.18428862e-01 -1.27127254e+00 1.28118008e-01 -7.40646362e-01
-3.04023653e-01 1.32078302e+00 6.27065957e-01 -7.58628845e-02
4.07703310e-01 -2.58427769e-01 -5.12873709e-01 -5.66708803e-01
-7.35092342e-01 -3.34776223e-01 6.14965200e-01 5.90575993e-01
8.76457572e-01 2.88625509e-02 -2.14038372e-01 6.14835203e-01
-9.57720876e-02 -1.42748699e-01 7.17568696e-02 9.69004929e-01
-5.89788795e-01 -1.62964725e+00 1.56470120e-01 5.29411852e-01
-2.64312387e-01 -5.25098145e-01 -8.63255739e-01 5.10914326e-01
4.65520024e-02 7.01193392e-01 -4.30209339e-01 -6.25973761e-01
4.06833105e-02 7.46085346e-02 6.30689919e-01 -9.71304476e-01
-2.89345652e-01 1.59621730e-01 5.05803265e-02 -5.32253206e-01
-4.57266808e-01 -2.15374321e-01 -1.38304865e+00 -2.68935084e-01
-3.88830900e-01 3.71634960e-01 2.07727015e-01 8.29138935e-01
6.73358381e-01 5.62080562e-01 7.10561574e-01 -9.19607937e-01
-1.46100685e-01 -8.37225318e-01 -3.70915532e-01 3.65513772e-01
1.17575619e-02 -6.84869468e-01 -9.31758955e-02 6.22021593e-02] | [5.117594242095947, 5.925605773925781] |
5aa6e6c0-addd-415c-ad2a-66acf51092e2 | short-text-conversation-based-on-deep-neural | 1907.03070 | null | https://arxiv.org/abs/1907.03070v1 | https://arxiv.org/pdf/1907.03070v1.pdf | Short Text Conversation Based on Deep Neural Network and Analysis on Evaluation Measures | With the development of Natural Language Processing, Automatic question-answering system such as Waston, Siri, Alexa, has become one of the most important NLP applications. Nowadays, enterprises try to build automatic custom service chatbots to save human resources and provide a 24-hour customer service. Evaluation of chatbots currently relied greatly on human annotation which cost a plenty of time. Thus, has initiated a new Short Text Conversation subtask called Dialogue Quality (DQ) and Nugget Detection (ND) which aim to automatically evaluate dialogues generated by chatbots. In this paper, we solve the DQ and ND subtasks by deep neural network. We proposed two models for both DQ and ND subtasks which is constructed by hierarchical structure: embedding layer, utterance layer, context layer and memory layer, to hierarchical learn dialogue representation from word level, sentence level, context level to long range context level. Furthermore, we apply gating and attention mechanism at utterance layer and context layer to improve the performance. We also tried BERT to replace embedding layer and utterance layer as sentence representation. The result shows that BERT produced a better utterance representation than multi-stack CNN for both DQ and ND subtasks and outperform other models proposed by other researches. The evaluation measures are proposed by , that is, NMD, RSNOD for DQ and JSD, RNSS for ND, which is not traditional evaluation measures such as accuracy, precision, recall and f1-score. Thus, we have done a series of experiments by using traditional evaluation measures and analyze the performance and error. | ['Chia-Hui Chang', 'Hsiang-En Cherng'] | 2019-07-06 | null | null | null | null | ['short-text-conversation'] | ['natural-language-processing'] | [-3.35922718e-01 1.73950478e-01 3.67438644e-01 -4.62631375e-01
-5.17423928e-01 -3.55924845e-01 6.24997973e-01 5.24289943e-02
-5.83352149e-01 9.15688396e-01 5.13639212e-01 -1.66547775e-01
1.97586119e-01 -7.91802645e-01 -5.89457690e-04 -2.88664132e-01
3.74398410e-01 5.67253828e-01 5.79479575e-01 -7.14694798e-01
5.92039526e-01 -8.51652920e-02 -1.13054526e+00 5.90223372e-01
8.76771569e-01 9.00809050e-01 4.58673716e-01 9.04219151e-01
-7.46384263e-01 1.14490938e+00 -1.19378853e+00 -3.88683140e-01
-2.79832691e-01 -6.78460062e-01 -1.40867150e+00 -2.43524704e-02
-5.50282001e-01 -6.19340420e-01 -1.64603665e-01 8.68292868e-01
7.46324241e-01 4.73747492e-01 2.73456752e-01 -1.16254294e+00
-7.83317685e-01 5.77476323e-01 -2.26732627e-01 2.77400970e-01
5.26307344e-01 2.48077363e-01 9.61294293e-01 -5.84779084e-01
3.32714647e-01 1.60299671e+00 4.13551658e-01 7.30255067e-01
-6.35595322e-01 -3.38257372e-01 8.02992359e-02 3.74120682e-01
-6.81294799e-01 -1.77717045e-01 6.00204825e-01 -1.63727149e-01
1.43797421e+00 2.27893859e-01 2.84885526e-01 8.58208835e-01
2.15650663e-01 9.09672797e-01 1.01591504e+00 -4.10078406e-01
-9.39829573e-02 4.45124418e-01 6.23592377e-01 6.12322152e-01
-5.09040713e-01 -4.83930916e-01 -1.38764560e-01 -6.09550960e-02
7.14116633e-01 -4.02359441e-02 -6.14556596e-02 6.60038114e-01
-8.68853986e-01 1.15022445e+00 2.83466607e-01 6.99134886e-01
-2.11127087e-01 -1.96475223e-01 7.56981492e-01 6.20736361e-01
3.77512813e-01 5.23280799e-01 -6.63587868e-01 -5.66913843e-01
-1.31193072e-01 2.12865576e-01 1.12452447e+00 8.03137004e-01
5.80666482e-01 -1.21491812e-01 -5.70738912e-01 1.31898224e+00
3.89733136e-01 -1.01586366e-02 9.60044026e-01 -9.67042148e-01
6.65451348e-01 1.04353678e+00 1.19007252e-01 -7.77719915e-01
-7.07220733e-01 3.58783789e-02 -8.12684476e-01 -4.03372288e-01
8.26052204e-02 -5.58443844e-01 -4.71176624e-01 1.54477656e+00
2.68030047e-01 -1.26358554e-01 2.53588229e-01 9.82988954e-01
1.47052634e+00 1.01945710e+00 6.04375564e-02 -2.26715431e-01
1.64559841e+00 -1.50483382e+00 -1.19367993e+00 8.78248364e-02
9.94826436e-01 -8.72986615e-01 1.47867680e+00 2.13572547e-01
-9.98496294e-01 -6.75834537e-01 -9.02373910e-01 -3.23035389e-01
-3.97230834e-01 5.22548221e-02 4.62467879e-01 4.35067475e-01
-8.79040599e-01 4.45212454e-01 -3.14175010e-01 -4.19684827e-01
-1.13171294e-01 3.30173671e-01 -2.46558800e-01 3.16402167e-01
-1.85199976e+00 1.21366298e+00 3.49410743e-01 2.88194418e-01
-6.77086473e-01 6.64031133e-02 -6.43837690e-01 2.00131327e-01
2.72372872e-01 -2.09410414e-01 1.63664806e+00 -7.27052867e-01
-2.04848075e+00 6.02778971e-01 -2.83240583e-02 -3.25652987e-01
-7.97524974e-02 -1.02388278e-01 -4.25777614e-01 -2.89520957e-02
1.36856465e-02 3.13000232e-01 1.09266870e-01 -6.73190415e-01
-6.53911829e-01 -3.49073887e-01 7.59448230e-01 5.52117825e-01
-2.02878028e-01 4.16386247e-01 -1.82807535e-01 6.41429126e-02
-1.57264680e-01 -4.97546464e-01 -1.60953104e-01 -8.01619411e-01
-3.40975195e-01 -9.40771401e-01 8.71342421e-01 -9.81454551e-01
1.44042778e+00 -1.88116622e+00 -1.24306694e-01 -3.18016231e-01
1.43115357e-01 6.14094198e-01 -1.02701917e-01 6.69342816e-01
3.66209209e-01 2.60887533e-01 1.26648583e-02 -1.42119765e-01
5.72788790e-02 3.31423610e-01 9.14416462e-02 -2.36062631e-01
3.44170332e-01 8.35164368e-01 -8.21886182e-01 -6.10641897e-01
2.72993594e-01 -1.00217350e-02 -4.95141357e-01 7.89082944e-01
-4.19198751e-01 4.72636133e-01 -4.78020251e-01 3.17364872e-01
3.20899904e-01 -3.00634295e-01 -9.31409225e-02 1.10433660e-01
-1.30715281e-01 7.11390555e-01 -7.87276387e-01 1.63165128e+00
-8.02834570e-01 5.14532208e-01 3.42873245e-01 -9.00307119e-01
1.23390460e+00 7.62809873e-01 1.73219119e-03 -8.51907492e-01
5.48331797e-01 8.94741938e-02 1.50209129e-01 -1.14491200e+00
6.83801889e-01 -2.32068986e-01 -2.15228721e-01 4.72844630e-01
3.97814572e-01 7.07638562e-02 7.39584342e-02 1.46612018e-01
1.11221492e+00 -1.72007993e-01 1.24442680e-02 8.65776092e-02
9.75098908e-01 -3.52118947e-02 3.29671741e-01 3.42748880e-01
-4.84357089e-01 3.28502834e-01 9.53022122e-01 -1.46396682e-01
-6.81251824e-01 -4.63360101e-01 1.54492080e-01 1.42158389e+00
1.66887820e-01 -2.02241689e-01 -9.84579623e-01 -8.42554033e-01
-4.91979301e-01 6.79370403e-01 -2.38500878e-01 -1.49833038e-01
-6.73668325e-01 -6.34699941e-01 7.05250561e-01 2.78614223e-01
1.20312715e+00 -1.62811971e+00 -1.87382638e-01 4.80002075e-01
-5.28047025e-01 -1.02225637e+00 -4.37124282e-01 -3.54678966e-02
-6.56673253e-01 -9.11387801e-01 -6.30045712e-01 -9.19305384e-01
-3.75991804e-03 6.23032302e-02 9.58771825e-01 3.33258480e-01
6.61626756e-02 -1.99137002e-01 -7.67081499e-01 -2.57923245e-01
-3.65757287e-01 2.91304201e-01 -3.07969421e-01 -1.38439342e-01
4.89191890e-01 -4.36389893e-01 -4.84657347e-01 4.36481416e-01
-7.95656383e-01 -7.40274489e-02 5.18914819e-01 1.03497005e+00
-1.46663517e-01 -3.22175503e-01 1.10473096e+00 -8.68129730e-01
1.43601882e+00 -5.17109096e-01 -1.79508790e-01 2.46488556e-01
-3.90325725e-01 -1.66014675e-02 5.89338481e-01 -2.64008921e-02
-1.36812615e+00 -5.53442299e-01 -8.20806146e-01 2.60644913e-01
-2.39187613e-01 6.47062957e-01 -3.64869833e-01 4.00360793e-01
4.35124338e-01 1.16355285e-01 2.20864877e-01 -6.13919616e-01
1.54522523e-01 1.50237572e+00 7.88923949e-02 -3.63264084e-01
-1.35654509e-01 -3.00552100e-01 -7.58145869e-01 -8.85957718e-01
-7.36583769e-01 -6.49780273e-01 -3.59261811e-01 -2.49950573e-01
1.24681902e+00 -5.84477842e-01 -1.15050483e+00 3.45896095e-01
-1.76689827e+00 -1.96946964e-01 2.08288640e-01 4.21970546e-01
-2.37802789e-01 4.49224323e-01 -1.02952707e+00 -1.03945994e+00
-6.51472449e-01 -1.13215077e+00 7.81406999e-01 5.00509560e-01
-2.46089384e-01 -9.35594678e-01 6.67298511e-02 7.44934559e-01
5.68777144e-01 -1.98901236e-01 9.40308213e-01 -1.13227379e+00
-2.18214691e-01 -1.68015823e-01 -4.20484036e-01 7.37349629e-01
-2.15750141e-03 -2.59777635e-01 -9.59400475e-01 2.31484354e-01
3.83381307e-01 -6.37610853e-01 5.32545865e-01 9.91655588e-02
8.03207576e-01 -5.95120788e-01 1.85312063e-01 -9.84581038e-02
1.04505956e+00 6.01660728e-01 8.06635559e-01 2.76969194e-01
4.23268020e-01 9.04650331e-01 6.98753655e-01 2.68427491e-01
5.92430174e-01 5.24312139e-01 3.00808579e-01 2.61446744e-01
2.10887983e-01 -3.37201282e-02 3.41921598e-01 1.37115693e+00
-6.45771250e-02 -2.64681399e-01 -6.55607700e-01 4.44956660e-01
-2.02767015e+00 -9.03042197e-01 -2.49450922e-01 1.72328293e+00
9.78680909e-01 2.39937529e-01 1.78716049e-01 -9.61557627e-02
7.53984749e-01 1.68546766e-01 -2.28665695e-01 -9.56136644e-01
1.60048395e-01 -3.75145823e-02 -9.89310071e-02 7.20193148e-01
-7.16239870e-01 9.87630010e-01 5.29755211e+00 7.21018374e-01
-8.61131787e-01 5.32682002e-01 6.40077770e-01 2.31388763e-01
-1.09677970e-01 -1.40216008e-01 -7.93867469e-01 6.06520772e-01
1.24630713e+00 1.21541005e-02 3.00779790e-01 7.81427324e-01
2.49011829e-01 -1.15008660e-01 -6.76671803e-01 7.02874482e-01
-9.50998366e-02 -1.13853872e+00 -1.32213011e-01 -4.07260835e-01
3.30980778e-01 -1.43724546e-01 -5.72691917e-01 1.04933679e+00
3.47763360e-01 -1.02347386e+00 -1.22048542e-01 4.06108022e-01
1.55162647e-01 -6.22890830e-01 1.44333041e+00 7.06061423e-01
-9.19434547e-01 -1.51650548e-01 -5.41715682e-01 -4.53022510e-01
2.59940177e-01 1.06443584e-01 -9.02174652e-01 5.89104056e-01
4.15303737e-01 1.54286385e-01 -1.92544445e-01 6.35127127e-01
-1.92711711e-01 4.98265684e-01 -3.31665426e-02 -8.36576462e-01
6.69458985e-01 -2.21817106e-01 1.62953407e-01 1.24898624e+00
-1.47374615e-01 4.08888668e-01 2.06224352e-01 6.72086835e-01
-1.70131892e-01 3.58828992e-01 -2.71911770e-01 -8.11240077e-02
5.95685303e-01 1.32399499e+00 -3.61073226e-01 -3.51972580e-01
-3.94219607e-01 9.14915860e-01 2.62266010e-01 1.16330899e-01
-8.51297438e-01 -9.11433756e-01 3.18765521e-01 -3.33595186e-01
-1.78127557e-01 -2.39699557e-02 -1.70121446e-01 -9.51646149e-01
-5.49568003e-03 -8.44991446e-01 1.25427574e-01 -6.73600793e-01
-1.08865082e+00 7.78963506e-01 -2.92333156e-01 -7.50595689e-01
-2.56676376e-01 -6.24547899e-01 -9.87060189e-01 1.18371260e+00
-1.27413476e+00 -8.43982995e-01 -3.15927625e-01 5.48164427e-01
1.03794026e+00 -1.81458876e-01 1.10362661e+00 4.01241928e-01
-8.56704056e-01 3.95838618e-01 1.72700007e-02 5.41293561e-01
6.66000605e-01 -1.21406567e+00 5.63057475e-02 1.03099488e-01
-3.27170640e-01 5.57089567e-01 4.35483307e-01 -2.34884635e-01
-8.54856670e-01 -7.56907701e-01 1.17288291e+00 -2.28869215e-01
5.71964502e-01 -2.50659376e-01 -9.34275806e-01 4.88925129e-01
5.74061751e-01 -6.93238020e-01 4.99739885e-01 2.58789986e-01
2.22348884e-01 -3.75860259e-02 -1.20983934e+00 3.57373267e-01
4.93915379e-01 -6.26572490e-01 -9.66391146e-01 5.21390617e-01
1.32345116e+00 -3.40669602e-01 -8.95667613e-01 1.04079172e-01
2.27136150e-01 -1.08993828e+00 4.84556288e-01 -6.71199322e-01
5.94827831e-01 5.76606952e-02 -5.84548004e-02 -1.11893082e+00
-1.18225805e-01 -4.48090553e-01 5.66252507e-02 1.48682153e+00
5.48675001e-01 -6.69317424e-01 6.03358984e-01 6.37474954e-01
-4.33321387e-01 -1.06004882e+00 -8.41832101e-01 -4.02783126e-01
-7.15294629e-02 4.64721732e-02 7.96167433e-01 8.14691007e-01
3.00637752e-01 1.33936357e+00 -2.80815750e-01 -3.23585391e-01
-2.57969081e-01 -8.01223069e-02 6.67121947e-01 -1.05785310e+00
-2.43040368e-01 -4.75758880e-01 -1.27599567e-01 -1.40147543e+00
1.92664620e-02 -4.00239825e-01 1.49345636e-01 -2.00893140e+00
-1.08819701e-01 -1.25831783e-01 -2.46544823e-01 2.16370523e-01
-3.21494669e-01 -2.83059418e-01 3.28662321e-02 1.40430927e-01
-9.25499022e-01 7.80709028e-01 1.50533235e+00 -1.34671241e-01
-3.78930509e-01 1.04081042e-01 -6.14973545e-01 6.14630818e-01
1.00965762e+00 -5.22512868e-02 -4.51215208e-01 -5.40254772e-01
-2.51181386e-02 5.76073408e-01 4.02706973e-02 -6.25152946e-01
2.41697878e-01 -1.70869641e-02 -7.60628060e-02 -5.76552451e-01
4.95632142e-01 -4.40841973e-01 -7.68640876e-01 3.29588115e-01
-6.45832896e-01 7.79797062e-02 -7.60025084e-02 2.17492491e-01
-5.84127247e-01 -6.76960588e-01 5.79280436e-01 -5.50346136e-01
-6.18905902e-01 -4.52241786e-02 -5.87979794e-01 1.51998878e-01
7.86570132e-01 5.37243858e-02 -5.86481750e-01 -6.15878046e-01
-7.21824050e-01 7.72263050e-01 -4.13316369e-01 4.68560278e-01
6.47735476e-01 -1.04572439e+00 -7.15484023e-01 -9.26954597e-02
-1.92284897e-01 8.88103023e-02 6.04004204e-01 8.33389580e-01
-6.68084025e-01 6.68475926e-01 -2.46439949e-01 -2.42228568e-01
-1.23956263e+00 8.26870874e-02 3.11460197e-01 -8.29051077e-01
-1.48621127e-01 1.12508929e+00 -7.01483712e-02 -1.10347092e+00
3.67940277e-01 -2.67331839e-01 -8.71975303e-01 3.99507433e-02
5.34289241e-01 3.78188759e-01 8.28532130e-02 -4.37825471e-01
-7.31784552e-02 4.60579582e-02 -1.58659324e-01 -2.16937155e-01
1.03154409e+00 -3.01636755e-01 -4.22479391e-01 5.40939689e-01
1.29310334e+00 -4.89416540e-01 -6.30890906e-01 -9.22478288e-02
1.99436650e-01 -1.22971609e-02 -2.11042732e-01 -9.95459199e-01
-5.53029597e-01 1.21699071e+00 4.56676751e-01 8.32539856e-01
7.65045047e-01 -1.71934381e-01 1.32359242e+00 7.47940838e-01
1.92641571e-01 -1.39339638e+00 3.24261993e-01 1.13394201e+00
1.04621744e+00 -1.35197425e+00 -5.26334882e-01 1.85238849e-02
-9.46055651e-01 9.34268773e-01 1.12411129e+00 -2.25159317e-01
6.04736805e-01 -1.13200806e-01 2.09651753e-01 -2.24714220e-01
-1.00960839e+00 -1.18109673e-01 -2.06705317e-01 3.12913328e-01
8.30786169e-01 -8.22474435e-02 -7.28301525e-01 1.10176599e+00
-2.37424746e-01 -7.21770078e-02 3.92826527e-01 8.10896873e-01
-9.59856689e-01 -1.13900447e+00 -1.37645960e-01 5.97421050e-01
-5.46690524e-01 -6.90693706e-02 -5.94259560e-01 7.22640336e-01
3.35349254e-02 1.46393430e+00 -8.37862268e-02 -8.44021559e-01
5.67252338e-01 3.90419006e-01 2.29507424e-02 -9.27522838e-01
-1.14071524e+00 -2.53214151e-01 6.07093573e-01 -3.13411027e-01
-4.08130944e-01 -3.83681767e-02 -1.49724770e+00 -3.58141989e-01
-8.54366839e-01 7.30575085e-01 6.10424757e-01 1.26429725e+00
1.61987096e-01 7.58071780e-01 8.42624903e-01 -4.03267443e-01
-7.61119425e-01 -1.87712169e+00 -4.71456319e-01 2.87786126e-01
5.31341741e-03 -3.91778827e-01 -9.77255180e-02 -3.98602724e-01] | [12.755526542663574, 7.726679801940918] |
efe93516-74ae-4cd3-8af2-7381d2076fa8 | materobot-material-recognition-in-wearable | 2302.14595 | null | https://arxiv.org/abs/2302.14595v1 | https://arxiv.org/pdf/2302.14595v1.pdf | MateRobot: Material Recognition in Wearable Robotics for People with Visual Impairments | Wearable robotics can improve the lives of People with Visual Impairments (PVI) by providing additional sensory information. Blind people typically recognize objects through haptic perception. However, knowing materials before touching is under-explored in the field of assistive technology. To fill this gap, in this work, a wearable robotic system, MateRobot, is established for PVI to recognize materials before hand. Specially, the human-centric system can perform pixel-wise semantic segmentation of objects and materials. Considering both general object segmentation and material segmentation, an efficient MateViT architecture with Learnable Importance Sampling (LIS) and Multi-gate Mixture-of-Experts (MMoE) is proposed to wearable robots to achieve complementary gains from different target domains. Our methods achieve respective 40.2% and 51.1% of mIoU on COCOStuff and DMS datasets, surpassing previous method with +5.7% and +7.0% gains. Moreover, on the field test with participants, our wearable system obtains a score of 28 in NASA-Task Load Index, indicating low cognitive demands and ease of use. Our MateRobot demonstrates the feasibility of recognizing material properties through visual cues, and offers a promising step towards improving the functionality of wearable robots for PVI. Code will be available at: https://github.com/JunweiZheng93/MATERobot. | ['Rainer Stiefelhagen', 'Kunyu Peng', 'Kailun Yang', 'Jiaming Zhang', 'Junwei Zheng'] | 2023-02-28 | null | null | null | null | ['material-recognition'] | ['computer-vision'] | [-9.77056548e-02 1.31106928e-01 -1.47834972e-01 3.35328653e-02
-6.16494477e-01 -3.19263488e-01 -3.21409076e-01 -2.45286867e-01
-3.17346245e-01 6.41713738e-01 2.00884104e-01 5.87960556e-02
-3.10608387e-01 -5.31617939e-01 -7.53627598e-01 -5.22410631e-01
1.79639563e-01 5.19072227e-02 6.88114464e-02 -1.93783596e-01
6.31100386e-02 3.77710611e-02 -1.89239240e+00 4.92852181e-01
1.38999903e+00 1.43329155e+00 8.47666204e-01 4.90452439e-01
3.87536287e-01 2.71306396e-01 -3.05297166e-01 -2.00828284e-01
4.17964101e-01 3.17871422e-01 -5.25868952e-01 -1.63335636e-01
3.20996612e-01 -7.56508529e-01 -1.68332830e-01 8.62157106e-01
1.14828086e+00 -3.45404327e-01 8.12084377e-01 -1.06433535e+00
-8.03175926e-01 2.72570282e-01 -3.52697670e-01 -4.00214374e-01
6.95614457e-01 4.16095287e-01 7.70081222e-01 -9.80741382e-01
7.34494776e-02 1.09128952e+00 8.51492345e-01 6.67478204e-01
-6.96502328e-01 -5.18891454e-01 7.03104064e-02 6.24850154e-01
-9.95595753e-01 -3.03659827e-01 6.92138374e-01 -6.15898848e-01
9.42189753e-01 4.34570998e-01 1.20198154e+00 1.15343893e+00
9.37879384e-02 1.24776292e+00 1.05173492e+00 -3.10974658e-01
2.50965834e-01 1.02486365e-01 5.31841516e-02 4.27528232e-01
8.34404826e-01 -2.47566365e-02 -5.77354848e-01 1.24087423e-01
7.78035045e-01 1.40434027e-01 -5.52641571e-01 -1.79808348e-01
-1.17070532e+00 6.97245747e-02 8.28689873e-01 -8.14444348e-02
-6.57830834e-01 2.79397875e-01 2.38718539e-02 -1.35418713e-01
2.08732277e-01 3.93175215e-01 -4.26918954e-01 -2.11619228e-01
-1.33214101e-01 1.06372684e-01 5.73476613e-01 1.33116496e+00
1.98491141e-01 -2.53487766e-01 -6.21831775e-01 1.07899678e+00
6.30324423e-01 1.05710101e+00 2.01673612e-01 -1.07896459e+00
4.26148444e-01 8.05019557e-01 8.05529773e-01 -4.04924840e-01
-5.89655161e-01 -2.05433294e-01 -5.60066640e-01 5.27955770e-01
5.12733400e-01 -3.18991870e-01 -1.09534514e+00 1.11981559e+00
2.23375246e-01 -3.87601674e-01 -2.42619127e-01 1.35940254e+00
9.90712821e-01 1.05468392e-01 4.25340533e-01 1.28284127e-01
1.63636148e+00 -7.41817594e-01 -5.30983984e-01 -4.91731137e-01
2.55526423e-01 -5.37874579e-01 1.44594491e+00 7.85507321e-01
-1.00454652e+00 -6.34134173e-01 -9.13467109e-01 -1.68594688e-01
-4.12812755e-02 1.01409221e+00 9.13127840e-01 8.30510080e-01
-1.05601454e+00 1.61814690e-01 -6.71751916e-01 -4.62464035e-01
7.77208030e-01 5.91981292e-01 1.28264204e-01 -2.94756979e-01
-5.82792699e-01 6.32242739e-01 2.97012739e-02 9.80269089e-02
-6.88758790e-01 -8.47236812e-01 -2.37058163e-01 -3.32610488e-01
-4.69431356e-02 -1.03741205e+00 1.27715766e+00 -4.26254064e-01
-1.76688266e+00 6.26023412e-01 -2.37706397e-02 -2.60136247e-01
7.42674708e-01 -9.04200017e-01 -2.27285251e-02 3.65511119e-01
9.87164304e-02 8.74759555e-01 1.11064672e+00 -1.23532987e+00
-7.54113138e-01 -4.08047557e-01 1.10126823e-01 3.54314297e-01
-6.33895278e-01 -2.06852317e-01 -2.32611686e-01 -4.94710922e-01
2.04652786e-01 -9.38151717e-01 5.58816753e-02 7.46057212e-01
-4.45447326e-01 -4.12522763e-01 4.01536107e-01 -1.19701672e+00
6.23064458e-01 -2.13700318e+00 -3.92910577e-02 -1.64763615e-01
5.17302036e-01 4.17435884e-01 9.35321376e-02 -3.49332355e-02
4.94446814e-01 -2.28301674e-01 -2.17754790e-03 -1.61297396e-01
-6.77373377e-04 -6.70179501e-02 1.18067190e-01 3.41466606e-01
-4.56067361e-02 1.01401842e+00 -8.94566596e-01 -3.53405848e-02
4.58951145e-01 5.74400127e-01 -5.96713901e-01 6.74521029e-02
-2.00827166e-01 5.46790779e-01 -3.37510079e-01 1.16460097e+00
7.75460243e-01 4.04719263e-02 -1.02405049e-01 -5.23471057e-01
4.59131263e-02 -6.95580095e-02 -1.02380335e+00 1.82258701e+00
-4.15511072e-01 2.97158867e-01 2.78288782e-01 -6.63580954e-01
6.41291559e-01 4.36277896e-01 8.18725169e-01 -7.55208671e-01
3.45112413e-01 5.16724586e-01 -2.36525923e-01 -1.04084408e+00
2.04852656e-01 5.43860376e-01 2.73411661e-01 -3.34891409e-01
-8.62174854e-02 1.20040961e-02 -2.71808207e-01 -3.37589413e-01
9.59568620e-01 4.17683572e-01 4.25085388e-02 -1.82315782e-01
-1.24297030e-01 -6.35023862e-02 2.65052050e-01 7.54858136e-01
-5.12176275e-01 5.05051374e-01 -2.96493292e-01 -6.80726320e-02
-8.65399301e-01 -1.46977735e+00 -1.46761924e-01 7.68433511e-01
6.42107487e-01 1.30277276e-01 -7.44425118e-01 -7.56952763e-02
3.65489393e-01 2.26956904e-01 -2.42143616e-01 3.74102630e-02
-6.62410408e-02 -4.61047351e-01 1.00306191e-01 8.86705995e-01
8.90731514e-01 -8.49255562e-01 -1.00178778e+00 1.06691331e-01
-5.33586919e-01 -1.01692045e+00 -1.20478392e-01 -3.80663335e-01
-8.65909159e-01 -9.78533506e-01 -1.32092941e+00 -1.16954648e+00
7.88261354e-01 6.85250640e-01 4.74822462e-01 -3.55149597e-01
-6.66859627e-01 1.00436032e+00 -5.17799079e-01 -9.68840480e-01
3.19923848e-01 -5.19585572e-02 4.37758207e-01 -1.72809467e-01
2.06172526e-01 -6.35687053e-01 -1.48840177e+00 5.62741458e-01
-1.14053171e-02 1.62669942e-01 1.01079452e+00 5.47896206e-01
5.79436064e-01 -3.19076687e-01 5.83514273e-01 2.57210761e-01
6.37969494e-01 -2.46166959e-01 -1.36629015e-01 2.18849644e-01
-3.39874536e-01 -3.31453681e-01 1.32156843e-02 -8.00792396e-01
-1.08033264e+00 2.27318734e-01 2.21234243e-02 -1.74088612e-01
-1.28009528e-01 -1.86056662e-02 -3.31630945e-01 -4.88830715e-01
1.07442009e+00 -2.92964578e-01 2.61778384e-02 -5.42910516e-01
3.81458789e-01 1.47281229e+00 5.71388304e-01 -6.21947229e-01
2.69806623e-01 5.16366959e-01 -5.17399251e-01 -8.86490285e-01
-3.63266468e-01 -7.65936911e-01 -1.45141333e-01 -8.64934504e-01
6.59915090e-01 -1.16822088e+00 -1.37583017e+00 9.47350264e-01
-1.18302739e+00 -6.36528552e-01 -3.49193215e-01 9.65973318e-01
-4.95250732e-01 2.19569996e-01 -4.57976222e-01 -1.22982061e+00
-7.40625024e-01 -9.83330250e-01 1.09910309e+00 3.70889604e-01
-3.18972588e-01 -4.87190038e-02 -5.57313740e-01 8.24450612e-01
3.34680915e-01 -1.16917349e-01 4.70709890e-01 2.60499567e-01
-5.83545625e-01 -2.95464665e-01 -6.44990444e-01 4.66883719e-01
3.57387245e-01 -7.05158055e-01 -1.25651371e+00 -3.79410893e-01
-2.37717509e-01 -2.15119362e-01 6.54768646e-01 7.61167228e-01
1.01323438e+00 -2.75894314e-01 -5.11297643e-01 5.92086390e-02
1.19065154e+00 3.48181337e-01 9.50675130e-01 2.48559825e-02
8.59026313e-01 6.16719127e-01 5.70055246e-01 6.82415783e-01
6.00590348e-01 5.26238143e-01 4.92283016e-01 1.48527235e-01
-6.63957298e-01 -7.77657032e-02 4.22415078e-01 5.62211573e-01
-7.64881849e-01 -1.04250843e-02 -8.75723481e-01 8.52187753e-01
-1.72814143e+00 -4.16325539e-01 -2.11563632e-01 2.13667774e+00
5.22965372e-01 -1.36805385e-01 3.86395514e-01 3.90755564e-01
6.98976278e-01 -7.94587791e-01 -9.99671042e-01 3.21412534e-01
1.59560308e-01 1.31450817e-01 8.11095476e-01 1.02636814e-01
-8.69852960e-01 6.24417782e-01 5.25077772e+00 6.97088063e-01
-7.78728187e-01 1.05992921e-01 8.61111209e-02 -2.13976815e-01
-6.53355271e-02 -5.36106527e-01 -4.70368296e-01 5.13858795e-01
2.44818285e-01 6.55492127e-01 8.82264674e-01 1.00296175e+00
2.19585136e-01 -3.81568253e-01 -6.76352978e-01 1.64359975e+00
-1.66039273e-01 -5.95295727e-01 -3.14181864e-01 -9.60249826e-02
3.69113535e-01 1.07918873e-01 2.20415086e-01 -1.38752103e-01
-1.04273841e-01 -5.79901159e-01 1.03422320e+00 1.03417337e+00
9.99139249e-01 -1.89876065e-01 5.08954763e-01 -6.74873143e-02
-1.16623104e+00 -5.53439796e-01 -2.42500275e-01 -3.58936608e-01
-1.30769223e-01 7.63120532e-01 -1.02807820e+00 4.06566679e-01
1.36891353e+00 6.39298618e-01 -3.02325547e-01 1.67758560e+00
-5.90568244e-01 6.42046154e-01 -5.06600440e-01 -4.93200809e-01
-6.29974365e-01 2.96256561e-02 6.87746942e-01 6.23327911e-01
7.40526557e-01 1.19006097e-01 -2.07021739e-02 5.53837419e-01
1.99251086e-01 1.49940059e-01 -2.78603584e-01 2.47060567e-01
4.26555902e-01 9.33419764e-01 -3.81010056e-01 7.13411942e-02
-1.81789324e-01 1.16829491e+00 -1.95092380e-01 6.06088519e-01
-4.97039258e-01 -7.22567320e-01 6.11666560e-01 3.42170656e-01
1.20398976e-01 -5.04205704e-01 -1.16504669e+00 -8.07278812e-01
8.66732359e-01 -4.50517863e-01 -6.77382872e-02 -1.27089739e+00
-1.30912936e+00 1.30324468e-01 -3.07720959e-01 -1.76915061e+00
6.44808829e-01 -1.31127083e+00 -1.27663165e-01 7.27247834e-01
-1.16250670e+00 -1.42872369e+00 -1.00102305e+00 4.64157432e-01
4.35279310e-01 -7.37153590e-02 5.43027043e-01 3.72269124e-01
-2.56403267e-01 7.95493484e-01 5.30764461e-02 -2.89389014e-01
7.18298852e-01 -9.26622748e-01 4.84202504e-02 4.82888013e-01
-4.52733964e-01 4.00863707e-01 4.25045997e-01 -9.75697756e-01
-1.72170842e+00 -8.75100553e-01 1.54179275e-01 -4.84001428e-01
1.66059196e-01 -2.77044177e-01 -4.81823862e-01 4.21242006e-02
-1.50476068e-01 -2.70438552e-01 4.98293132e-01 -7.87808374e-02
-9.50712115e-02 -3.47402543e-01 -1.63873696e+00 7.81921089e-01
1.92137444e+00 -3.69905472e-01 -2.08393887e-01 5.68152726e-01
6.69336796e-01 -4.60675299e-01 -1.00017118e+00 6.92347944e-01
1.05605733e+00 -3.21667165e-01 1.17079473e+00 1.09787844e-01
3.46758701e-02 -4.00429487e-01 -2.47679412e-01 -1.12036765e+00
-3.29988152e-01 -6.72024429e-01 -3.91190380e-01 8.95456254e-01
1.54430151e-01 -1.05789459e+00 7.04769135e-01 8.75845969e-01
-5.70549071e-01 -7.45946884e-01 -8.69686186e-01 -8.47639382e-01
-5.31124890e-01 -9.57688570e-01 3.24578881e-01 3.73417735e-02
3.75623673e-01 -1.97199017e-01 -3.77721250e-01 4.55533713e-01
6.05378032e-01 -4.08187717e-01 4.42292243e-01 -1.54594636e+00
-2.84234613e-01 -2.71250188e-01 -5.45126081e-01 -1.12533915e+00
-5.60092032e-01 -6.76052988e-01 3.25117707e-01 -2.26416254e+00
-2.76188254e-01 -6.16813600e-01 -1.60825163e-01 6.97055221e-01
-7.48077929e-02 3.31059813e-01 -1.65771276e-01 3.91337335e-01
-4.78799522e-01 5.22511899e-01 1.45228577e+00 -6.55533254e-01
-5.20182192e-01 2.00355157e-01 -8.88374686e-01 5.02604127e-01
1.25908208e+00 7.46116042e-02 -4.45805818e-01 -6.13718927e-01
1.59779325e-01 -3.24381620e-01 9.01705801e-01 -1.58295739e+00
1.26985803e-01 1.14307687e-01 4.95689720e-01 -2.78642684e-01
5.47241986e-01 -7.38624096e-01 3.70199233e-02 6.94960117e-01
4.28721935e-01 -9.29533720e-01 2.63144702e-01 5.65008581e-01
7.29809105e-01 -1.82296708e-03 3.83324176e-01 1.06037423e-01
-8.42823565e-01 1.44015610e-01 -5.03634870e-01 -3.41352165e-01
9.45043266e-01 -6.41061962e-01 -4.51980740e-01 4.57386710e-02
-5.86466014e-01 2.32769474e-01 1.54136524e-01 4.02290463e-01
9.04566824e-01 -1.24115443e+00 -4.13262486e-01 -1.79581009e-02
2.92332679e-01 1.38190493e-01 6.47368789e-01 9.63198841e-01
-3.77506703e-01 1.47478372e-01 -5.50888419e-01 -5.90610266e-01
-9.67937171e-01 2.00164571e-01 1.23934671e-01 6.13395333e-01
-8.03736567e-01 7.88231969e-01 -3.02122831e-01 -1.13061026e-01
5.64940691e-01 -7.78506994e-01 -2.01170653e-01 -8.48832875e-02
3.30977589e-01 1.02408338e+00 1.00207515e-01 3.04213651e-02
-3.14725459e-01 7.32087135e-01 2.87010401e-01 -5.60493954e-02
1.13928330e+00 -7.35948980e-02 1.33878961e-01 6.27315342e-02
4.19137180e-01 -1.81678757e-01 -1.56300259e+00 1.37529954e-01
-4.33856934e-01 -3.92917186e-01 -2.15409905e-01 -1.30667996e+00
-6.94212973e-01 6.43048227e-01 1.47947967e+00 -9.99620035e-02
1.40514362e+00 1.66828662e-01 1.02700710e+00 5.89807391e-01
1.06412375e+00 -1.30093408e+00 2.62037277e-01 -2.24260241e-01
1.35015309e+00 -1.22339082e+00 -1.42618775e-01 -8.66605282e-01
-4.77165431e-01 7.13243067e-01 6.43346369e-01 -4.13794965e-02
2.53041267e-01 -4.47849669e-02 -1.65331334e-01 -1.53182428e-02
3.77980500e-01 -4.61096853e-01 6.14743471e-01 1.23101366e+00
-1.17403045e-01 7.03877866e-01 -3.50883752e-01 1.06438386e+00
-1.07940644e-01 3.87592584e-01 -5.52480780e-02 1.03096366e+00
-6.54062390e-01 -6.04169250e-01 -4.29695129e-01 9.97702062e-01
8.84346142e-02 2.49325633e-01 -6.97551221e-02 1.96758285e-02
4.34435666e-01 1.23057449e+00 -4.26001698e-01 -7.10586011e-01
5.85816324e-01 -2.81863719e-01 7.69615889e-01 -3.34091365e-01
-4.08236869e-02 -1.77858859e-01 2.82104522e-01 -5.55447638e-01
-3.19469482e-01 -4.86137390e-01 -9.72958624e-01 3.55454028e-01
-1.76086068e-01 -4.50832635e-01 9.95603740e-01 5.89103341e-01
7.15664327e-01 5.80510139e-01 1.83903798e-01 -1.29894555e+00
-6.27829209e-02 -1.18229151e+00 -4.42702770e-01 -1.10759577e-02
8.75263102e-03 -1.15094399e+00 -3.24460305e-02 -5.90868331e-02] | [7.741673946380615, -1.4363559484481812] |
8d6cee78-e241-41d7-9cdf-28cef974e31a | spts-single-point-text-spotting | 2112.07917 | null | https://arxiv.org/abs/2112.07917v6 | https://arxiv.org/pdf/2112.07917v6.pdf | SPTS: Single-Point Text Spotting | Existing scene text spotting (i.e., end-to-end text detection and recognition) methods rely on costly bounding box annotations (e.g., text-line, word-level, or character-level bounding boxes). For the first time, we demonstrate that training scene text spotting models can be achieved with an extremely low-cost annotation of a single-point for each instance. We propose an end-to-end scene text spotting method that tackles scene text spotting as a sequence prediction task. Given an image as input, we formulate the desired detection and recognition results as a sequence of discrete tokens and use an auto-regressive Transformer to predict the sequence. The proposed method is simple yet effective, which can achieve state-of-the-art results on widely used benchmarks. Most significantly, we show that the performance is not very sensitive to the positions of the point annotation, meaning that it can be much easier to be annotated or even be automatically generated than the bounding box that requires precise positions. We believe that such a pioneer attempt indicates a significant opportunity for scene text spotting applications of a much larger scale than previously possible. The code is available at https://github.com/shannanyinxiang/SPTS. | ['Lianwen Jin', 'Xiang Bai', 'Chunhua Shen', 'Dahua Lin', 'Jing Li', 'Shenggao Zhu', 'Songxuan Lai', 'Mingxin Huang', 'Jiaxin Zhang', 'Yuliang Liu', 'Xinyu Wang', 'Dezhi Peng'] | 2021-12-15 | null | null | null | null | ['text-spotting'] | ['computer-vision'] | [ 6.95126414e-01 -2.94100434e-01 -5.48767447e-02 -3.08811188e-01
-9.33603704e-01 -5.41710734e-01 7.51114964e-01 3.38422805e-01
-5.11712849e-01 2.37369537e-01 -1.20378047e-01 -3.62297416e-01
3.57092470e-01 -5.49733758e-01 -9.03710306e-01 -5.47697783e-01
4.05809343e-01 7.12959588e-01 5.48972130e-01 -1.11787423e-01
5.80478787e-01 2.69322246e-01 -1.49076116e+00 4.51655984e-01
6.75373435e-01 9.09263074e-01 6.07472599e-01 1.01905704e+00
-3.23740065e-01 6.56791508e-01 -5.61049461e-01 -2.21634731e-01
1.95995882e-01 -2.86576062e-01 -6.74778879e-01 3.48594248e-01
6.29530966e-01 -2.11469159e-01 -3.02707732e-01 8.60427797e-01
2.78917879e-01 -1.43120810e-02 6.00267470e-01 -8.72519314e-01
-1.72317028e-01 3.69082481e-01 -6.33468211e-01 -1.96599871e-01
3.60061467e-01 1.78652063e-01 1.05487823e+00 -1.13672781e+00
5.76430619e-01 7.77197123e-01 5.85939467e-01 3.67346853e-01
-1.18932545e+00 -3.47045034e-01 1.18960150e-01 1.53137982e-01
-1.44844055e+00 -4.68191713e-01 6.35730624e-01 -6.85121357e-01
8.42814863e-01 4.36836272e-01 3.98573428e-01 9.63647246e-01
-5.49029708e-02 1.06381357e+00 7.29989827e-01 -8.29350650e-01
1.20155886e-01 3.97343710e-02 5.94479032e-02 9.18184102e-01
1.34442179e-02 -4.26096827e-01 -6.04401588e-01 1.36037022e-01
5.14767289e-01 1.38532132e-01 -1.94073364e-01 -3.82839501e-01
-1.57714713e+00 6.22839034e-01 1.47805423e-01 3.37089449e-01
-1.00393908e-03 2.76940286e-01 7.19322801e-01 1.18197929e-02
6.24598503e-01 1.97977006e-01 -2.70981073e-01 -3.05387318e-01
-1.40882003e+00 6.44118413e-02 5.99155784e-01 9.27819550e-01
7.06460714e-01 -1.52076781e-01 -2.09433392e-01 8.69843006e-01
-7.33723044e-02 5.22388339e-01 6.13564909e-01 -9.64979976e-02
6.79617822e-01 5.37787318e-01 1.05307065e-01 -7.64360666e-01
-2.46129006e-01 9.34530720e-02 -7.06583738e-01 1.34497330e-01
5.80067158e-01 3.01880296e-02 -1.02602100e+00 1.03414130e+00
3.43021601e-01 2.92915076e-01 -3.66989821e-01 8.50121617e-01
4.05621916e-01 8.01230729e-01 -2.53822714e-01 1.75840124e-01
1.54474854e+00 -1.21546447e+00 -4.93046314e-01 -5.30056357e-01
9.94214594e-01 -9.88610268e-01 1.53748274e+00 4.99713093e-01
-6.97207391e-01 -3.81607234e-01 -9.27853227e-01 -2.68603712e-01
-4.17048603e-01 6.56065464e-01 2.73780823e-01 5.55452406e-01
-7.91636586e-01 4.16775584e-01 -9.25743461e-01 -5.83453178e-01
2.46681452e-01 1.16705671e-01 -3.36886883e-01 5.95275871e-02
-6.18386269e-01 6.10774040e-01 5.10450542e-01 1.19385356e-02
-7.78464913e-01 -3.89593601e-01 -8.37467551e-01 -3.62194739e-02
6.35148048e-01 -3.27498794e-01 1.28713524e+00 -1.01656437e+00
-1.47956014e+00 9.41861629e-01 -5.20517766e-01 -5.39400399e-01
8.50058019e-01 -4.76861298e-01 -3.69975083e-02 2.40907222e-01
1.26645401e-01 4.93418962e-01 1.13894260e+00 -1.01455176e+00
-7.63759196e-01 -1.91531241e-01 -4.98761386e-01 2.46279746e-01
-5.70420980e-01 2.15259552e-01 -6.74807131e-01 -7.80817747e-01
-5.56680523e-02 -8.68431687e-01 -1.91065848e-01 3.16317827e-01
-8.21715593e-01 -2.37517491e-01 1.03262866e+00 -8.27572465e-01
1.11721778e+00 -2.19701314e+00 -1.25071570e-01 2.67926808e-02
1.04332693e-01 3.18981797e-01 7.83449318e-03 5.61748266e-01
-1.85029004e-02 2.07723632e-01 -3.53063077e-01 -7.76610613e-01
1.07468804e-02 -2.44998232e-01 -5.32605350e-01 7.37462461e-01
9.42415074e-02 8.41027021e-01 -6.39823198e-01 -5.02526462e-01
6.74336433e-01 3.02719682e-01 -3.04784060e-01 1.15726054e-01
-5.98826349e-01 1.72073945e-01 -3.23641181e-01 3.72656614e-01
3.19913000e-01 -3.27967972e-01 -8.37215409e-02 7.67170042e-02
-4.06975925e-01 3.48301411e-01 -1.03404558e+00 1.67201090e+00
-3.77452910e-01 1.26077533e+00 -3.56243730e-01 -1.03419292e+00
1.03558648e+00 1.30908996e-01 3.81010741e-01 -5.09022593e-01
5.20578139e-02 1.93512946e-01 -3.99781078e-01 -3.76622677e-01
8.68764937e-01 1.51794761e-01 -1.99448138e-01 3.65700275e-01
-3.25733036e-01 -2.49254107e-01 2.82657713e-01 6.60676137e-02
1.02480435e+00 1.52090952e-01 3.36065173e-01 -4.97864224e-02
4.56714034e-01 2.36911386e-01 6.59264848e-02 8.80727351e-01
1.83278754e-01 9.02267098e-01 5.76573551e-01 -4.07700062e-01
-1.67201650e+00 -5.27150989e-01 1.58579703e-02 1.20865047e+00
1.33443668e-01 -7.52387762e-01 -8.92183840e-01 -6.08845472e-01
-2.03792334e-01 7.02193499e-01 -6.15118146e-01 2.42843285e-01
-6.87755346e-01 -2.97413349e-01 6.60258353e-01 4.56028283e-01
3.41616929e-01 -9.55156505e-01 -8.26191306e-01 -6.97869286e-02
-9.39225182e-02 -1.40597415e+00 -6.87553763e-01 2.93267727e-01
-7.18296766e-01 -7.65748322e-01 -9.79414165e-01 -9.03198838e-01
7.98535347e-01 2.95472264e-01 7.34998524e-01 1.38970196e-01
-4.95692909e-01 1.87462971e-01 -5.63056469e-01 -2.20680237e-01
-2.34036759e-01 1.33980900e-01 -2.36748621e-01 3.42273891e-01
9.04378667e-02 -1.65449325e-02 -5.49243748e-01 4.37667936e-01
-9.00914252e-01 6.06242716e-01 3.65744412e-01 7.97700822e-01
7.37467349e-01 -3.16349487e-03 -6.76389933e-02 -9.15914655e-01
1.64258376e-01 8.42622891e-02 -8.05568993e-01 3.16104800e-01
-3.28742504e-01 -8.14593956e-03 8.56526554e-01 -3.81994754e-01
-7.62057126e-01 6.07177556e-01 -8.02755803e-02 -4.34385717e-01
-5.08281469e-01 3.19166660e-01 1.82199791e-01 -2.18128003e-02
7.19880760e-01 6.80081248e-01 -2.95899868e-01 -5.40621698e-01
4.77794707e-01 7.97122717e-01 4.41080660e-01 -4.08436388e-01
8.97594273e-01 6.66911125e-01 -2.45371670e-01 -1.28497148e+00
-6.26802742e-01 -7.98865914e-01 -9.33303475e-01 -2.26890892e-01
8.92368317e-01 -7.13760316e-01 -5.03937244e-01 5.06293714e-01
-1.15132976e+00 -8.77456546e-01 -8.19985271e-02 8.09408277e-02
-6.65934145e-01 8.39086354e-01 -4.65587646e-01 -9.16778803e-01
-3.73044997e-01 -8.91087413e-01 1.70490468e+00 -1.05166435e-01
-1.65701896e-01 -7.93359578e-01 -1.36674240e-01 3.65555733e-01
4.69764546e-02 8.50305557e-02 6.90738678e-01 -6.44839585e-01
-6.01459563e-01 -5.53483427e-01 -3.28560084e-01 4.06056382e-02
-1.00934565e-01 2.85000890e-01 -8.34249377e-01 -1.66165814e-01
-2.47517437e-01 -2.41268933e-01 9.82535839e-01 1.35097042e-01
1.36621547e+00 -2.97030956e-01 -3.79825056e-01 4.68131959e-01
1.47490180e+00 5.54286316e-02 6.94602549e-01 5.02248168e-01
9.63468254e-01 3.70244890e-01 7.08304405e-01 6.26930833e-01
1.06458068e-01 1.04018307e+00 2.25258589e-01 -2.78980732e-01
-2.17630714e-01 -4.25095379e-01 4.39607561e-01 3.53897363e-01
4.11700279e-01 -5.36916912e-01 -1.31588233e+00 5.63701212e-01
-1.98888326e+00 -7.78734028e-01 -4.55066592e-01 2.41952729e+00
6.87873960e-01 2.28318825e-01 2.61449337e-01 4.62086588e-01
1.00453365e+00 1.85628578e-01 -3.87675732e-01 -2.32384995e-01
-6.98954388e-02 -5.27139902e-02 6.72692955e-01 6.08211398e-01
-1.21938396e+00 1.36326456e+00 5.38255262e+00 1.18066812e+00
-1.29165578e+00 -3.27592604e-02 6.75704360e-01 -2.03878619e-02
1.38396934e-01 9.56003144e-02 -9.10523355e-01 5.97975969e-01
6.39026523e-01 1.71275176e-02 4.01106060e-01 9.37053323e-01
4.01498884e-01 -2.45839491e-01 -1.15904582e+00 1.19998288e+00
3.17673296e-01 -1.29158139e+00 1.56899150e-02 -1.64880022e-01
5.04453540e-01 1.03709914e-01 -1.46006450e-01 -1.03957087e-01
-1.83745205e-01 -9.25478637e-01 1.14478314e+00 2.19131559e-01
1.16503537e+00 -3.85029107e-01 3.46849620e-01 5.51150799e-01
-1.33749318e+00 1.20000318e-01 -3.06798458e-01 -2.27188952e-02
2.37506449e-01 7.12879419e-01 -1.34951305e+00 2.76814491e-01
3.33256006e-01 7.21267939e-01 -7.43908346e-01 1.31749737e+00
-2.93808669e-01 8.29351723e-01 -5.39651453e-01 -4.58761126e-01
2.16943532e-01 2.69659534e-02 4.42570359e-01 1.71422410e+00
4.15205300e-01 -4.08081472e-01 3.46275806e-01 6.03121281e-01
-2.95419898e-02 5.12494087e-01 -6.41922951e-01 -5.17576598e-02
2.13273913e-02 1.21423686e+00 -1.26389587e+00 -5.55203557e-01
-2.95489401e-01 1.46673918e+00 2.41748735e-01 1.67896762e-01
-9.02771354e-01 -6.30912662e-01 1.01668790e-01 2.11620182e-01
6.04319334e-01 -4.55910593e-01 -7.21008420e-01 -1.29003942e+00
2.99486756e-01 -6.59268260e-01 2.33326823e-01 -9.02012825e-01
-8.01124811e-01 4.86801714e-01 -3.82067710e-01 -1.23987865e+00
1.59157231e-03 -8.49311113e-01 -6.13372207e-01 5.00252247e-01
-1.25218260e+00 -1.04047823e+00 -5.34220338e-01 5.51539540e-01
1.04248393e+00 1.56101942e-01 6.00287795e-01 2.24934757e-01
-7.23421693e-01 6.03410125e-01 3.59643251e-01 5.66669762e-01
7.17401683e-01 -1.17314494e+00 7.73004770e-01 1.04960036e+00
6.54383004e-01 1.02554962e-01 8.09140861e-01 -7.06239343e-01
-1.23637891e+00 -1.10402870e+00 9.37837362e-01 -6.03542089e-01
7.96179354e-01 -1.05038118e+00 -7.04160035e-01 6.00783229e-01
-1.38556845e-02 -1.67123169e-01 1.81756198e-01 -4.87478115e-02
-3.70690882e-01 1.12426504e-01 -6.47859633e-01 8.25953782e-01
8.40807259e-01 -5.51079154e-01 -2.91078150e-01 8.50438356e-01
6.03767514e-01 -6.05560243e-01 -2.08794042e-01 6.66008145e-02
3.71729255e-01 -7.82905936e-01 6.64446890e-01 -2.47371420e-01
6.15058541e-01 -4.36825573e-01 -1.89171396e-02 -7.34743774e-01
1.06677227e-01 -6.51316881e-01 1.38586476e-01 1.06469536e+00
6.28319740e-01 -4.06417400e-01 1.02579546e+00 4.14127588e-01
-1.88705847e-01 -6.83627367e-01 -8.95344257e-01 -7.62809992e-01
-2.08259568e-01 -6.57224715e-01 1.14681169e-01 7.75743067e-01
1.84858128e-01 3.32646161e-01 -6.25517249e-01 -4.62301485e-02
3.67059052e-01 2.53892303e-01 1.03506005e+00 -8.64662111e-01
-3.19115669e-01 -6.01717472e-01 -4.50153023e-01 -1.57148516e+00
-9.65704955e-03 -7.99580157e-01 5.48787534e-01 -1.46791077e+00
9.98238325e-02 -4.22510326e-01 1.51984394e-01 6.83942735e-01
-1.94669560e-01 3.42208743e-01 2.80903429e-01 2.01989591e-01
-7.91721880e-01 4.35618311e-01 8.89979959e-01 -9.39091891e-02
-5.32445982e-02 -1.31913312e-02 -9.44820121e-02 6.50399089e-01
9.16623652e-01 -4.86999929e-01 -4.13955450e-02 -4.20664191e-01
1.91005990e-01 -3.66373956e-02 4.99730378e-01 -9.44245338e-01
4.82198030e-01 -5.59299141e-02 2.00069636e-01 -7.30257273e-01
3.91710520e-01 -6.31129920e-01 -2.97332734e-01 3.58657777e-01
-4.47400868e-01 -2.19518974e-01 1.33961543e-01 5.22098958e-01
8.81804824e-02 -5.11009276e-01 6.50109291e-01 1.15162358e-01
-7.09849060e-01 1.36750668e-01 -4.67821360e-01 -2.59452999e-01
1.02827632e+00 -4.24205571e-01 -2.92853594e-01 -2.88982958e-01
-2.25386962e-01 -3.31897140e-02 7.21584558e-01 3.08097064e-01
5.93316972e-01 -8.32020164e-01 -7.80386746e-01 5.40231951e-02
2.62996614e-01 1.84101909e-01 -1.08740982e-02 7.45405853e-01
-7.92994082e-01 5.67570090e-01 2.43822441e-01 -9.25662756e-01
-1.57463861e+00 5.96166909e-01 1.35077551e-01 -2.09089592e-01
-1.05891955e+00 7.23647952e-01 2.89031833e-01 -1.10943737e-02
3.13179016e-01 -3.77072781e-01 2.05577210e-01 -1.28367662e-01
4.98134255e-01 1.53103754e-01 1.17480420e-01 -5.38181782e-01
-2.23726258e-01 7.56833315e-01 -1.71276793e-01 -2.15077579e-01
1.07466328e+00 -1.26359556e-02 1.02318153e-01 5.30501187e-01
9.23180580e-01 3.50908525e-02 -1.37610304e+00 -1.25191286e-01
1.65391684e-01 -5.95814407e-01 -1.14233799e-01 -6.54409945e-01
-5.68939805e-01 1.13396299e+00 4.01581556e-01 2.92815059e-01
9.13401961e-01 -1.01261146e-01 7.94591188e-01 6.07934415e-01
1.60851792e-01 -1.02335012e+00 2.43106067e-01 5.60371041e-01
7.07760453e-01 -1.12531447e+00 -9.35288519e-02 -5.66895664e-01
-5.34230232e-01 1.17647505e+00 3.49808455e-01 -1.52231559e-01
1.74808770e-01 4.17133510e-01 -1.57005146e-01 5.18660471e-02
-6.06103122e-01 -2.25990683e-01 2.35435665e-01 2.95302689e-01
4.66300994e-01 2.92645898e-02 6.98488727e-02 -4.89109494e-02
-2.15577126e-01 -2.09832281e-01 5.21488845e-01 7.06246376e-01
-7.27280200e-01 -9.65642154e-01 -5.05886018e-01 5.06614566e-01
-3.82796615e-01 -4.01805520e-01 -5.20122647e-01 5.44767678e-01
-2.61193544e-01 7.15638161e-01 -8.82470533e-02 -2.88659811e-01
2.16593921e-01 1.93794951e-01 2.50891656e-01 -6.98103964e-01
-2.97747284e-01 2.69679099e-01 -7.83062726e-03 -3.97950500e-01
1.23956017e-01 -6.82925224e-01 -1.21840096e+00 -1.24895215e-01
-5.56590796e-01 -1.00049764e-01 8.30166340e-01 8.26424837e-01
2.61954993e-01 1.98074192e-01 5.11179268e-01 -1.02810347e+00
-1.30823657e-01 -8.62145483e-01 -2.32416824e-01 4.09927547e-01
4.04413402e-01 -2.19387308e-01 -2.22811058e-01 5.21611154e-01] | [12.001815795898438, 2.254969835281372] |
6e6cbf5e-98e5-494b-a1b9-386050075e65 | integrating-physiological-time-series-and | 2003.11059 | null | https://arxiv.org/abs/2003.11059v2 | https://arxiv.org/pdf/2003.11059v2.pdf | Integrating Physiological Time Series and Clinical Notes with Deep Learning for Improved ICU Mortality Prediction | Intensive Care Unit Electronic Health Records (ICU EHRs) store multimodal data about patients including clinical notes, sparse and irregularly sampled physiological time series, lab results, and more. To date, most methods designed to learn predictive models from ICU EHR data have focused on a single modality. In this paper, we leverage the recently proposed interpolation-prediction deep learning architecture(Shukla and Marlin 2019) as a basis for exploring how physiological time series data and clinical notes can be integrated into a unified mortality prediction model. We study both early and late fusion approaches and demonstrate how the relative predictive value of clinical text and physiological data change over time. Our results show that a late fusion approach can provide a statistically significant improvement in mortality prediction performance over using individual modalities in isolation. | ['Satya Narayan Shukla', 'Benjamin M. Marlin'] | 2020-03-24 | null | null | null | null | ['icu-mortality'] | ['medical'] | [ 3.49740423e-02 -3.42320740e-01 -9.71578583e-02 -3.54527116e-01
-8.49081993e-01 -2.44614363e-01 2.51337886e-01 1.00683033e+00
-8.01236331e-02 9.31983531e-01 6.14902020e-01 -5.24427831e-01
-2.79182315e-01 -5.44938684e-01 -4.13476467e-01 -4.41457510e-01
-3.52319807e-01 4.76134777e-01 -5.41013777e-01 2.55941510e-01
-1.15007378e-01 4.69627470e-01 -8.38384748e-01 6.76891625e-01
6.77593827e-01 1.32434964e+00 -5.31497478e-01 8.40906680e-01
2.25063026e-01 1.31712520e+00 -3.96543354e-01 2.05831766e-01
2.74622887e-01 -5.81274748e-01 -4.27388012e-01 -5.66062093e-01
5.38357571e-02 -6.81350529e-01 -6.30117357e-01 1.72034264e-01
9.08367634e-01 -1.16043948e-01 6.14759982e-01 -1.00828457e+00
-3.94322932e-01 6.02763474e-01 1.49407923e-01 2.53272384e-01
3.29277724e-01 6.23124659e-01 5.37552357e-01 -5.55208325e-01
1.70016944e-01 5.84582984e-01 1.47942328e+00 4.33629245e-01
-1.55078197e+00 -2.99941272e-01 -3.11628938e-01 7.50494972e-02
-1.14680672e+00 -4.22164381e-01 3.93101752e-01 -7.31165409e-01
1.12513185e+00 1.61929548e-01 8.24088097e-01 1.25732207e+00
7.20631063e-01 4.90150750e-01 8.50802898e-01 -3.37224193e-02
9.12225097e-02 -1.40368983e-01 2.74034262e-01 5.65103531e-01
2.18532518e-01 1.73508614e-01 -5.95328331e-01 -5.84020257e-01
6.52445674e-01 1.08081353e+00 -5.53036511e-01 1.01874910e-01
-1.61167514e+00 3.63951713e-01 1.68122396e-01 9.40260291e-02
-9.44487572e-01 2.27081686e-01 7.07886517e-01 4.33800668e-01
2.78437436e-01 5.59431791e-01 -1.16160417e+00 -4.33887005e-01
-9.32187200e-01 1.14170156e-01 9.39154029e-01 6.95450604e-01
1.93514764e-01 -7.35849589e-02 -4.78443772e-01 6.08351409e-01
5.59680760e-02 2.29108617e-01 7.09107399e-01 -1.00408518e+00
8.52500200e-02 5.45239985e-01 2.75706798e-01 -4.83901590e-01
-8.47415030e-01 -7.76586980e-02 -1.05109000e+00 -3.53539735e-01
3.64196688e-01 -7.22723722e-01 -7.02264547e-01 1.47419381e+00
-7.28925690e-02 6.69128001e-01 2.51238316e-01 3.53073716e-01
8.89488101e-01 3.24029773e-01 4.17924494e-01 -5.49460232e-01
1.04739606e+00 -5.77753484e-01 -8.80894661e-01 3.53335112e-01
9.91370976e-01 -2.56008685e-01 5.91110349e-01 3.06486666e-01
-1.25546491e+00 -3.69814456e-01 -5.83553970e-01 -1.29915953e-01
-3.03350508e-01 4.47931373e-03 3.13492835e-01 -5.16084097e-02
-9.51575577e-01 1.12910438e+00 -1.36204505e+00 -3.21922094e-01
6.02913380e-01 3.25326055e-01 -2.07057342e-01 -5.29105105e-02
-1.11729026e+00 8.99058640e-01 2.49660105e-01 -9.10225213e-02
-5.74485660e-01 -1.60395169e+00 -6.47235632e-01 3.02122027e-01
-2.56158262e-01 -1.45240843e+00 1.23139226e+00 -1.63389623e-01
-1.19026196e+00 4.22999173e-01 -2.88790613e-01 -8.83874536e-01
5.73381782e-01 -4.82763916e-01 -3.69416356e-01 1.76299080e-01
-4.86198097e-01 1.76400200e-01 2.89338738e-01 -6.96818054e-01
-3.82727653e-01 -5.09999394e-01 -6.77499771e-01 -1.28122985e-01
-1.72709018e-01 -2.38454670e-01 2.09947482e-01 -5.18773079e-01
-1.91493765e-01 -6.03580594e-01 -3.16238701e-01 1.05180457e-01
-1.21969834e-01 1.59522481e-02 6.31466687e-01 -9.14629042e-01
1.45191717e+00 -1.97602808e+00 -1.53112739e-01 -2.07371518e-01
7.59840906e-01 1.31486863e-01 2.62023479e-01 6.52972937e-01
-1.47644937e-01 1.18467942e-01 -1.62003934e-01 -5.44250309e-01
-2.33857274e-01 1.60556644e-01 -3.56379688e-01 3.77572924e-01
3.06348324e-01 1.34755349e+00 -7.93130934e-01 -3.84037912e-01
4.20635313e-01 8.74304354e-01 -5.31612813e-01 7.18128204e-01
4.15418148e-02 1.01279473e+00 -2.41823092e-01 7.35775709e-01
1.98476389e-02 -7.67989159e-01 1.31020173e-01 -1.53438359e-01
1.00181632e-01 3.23975623e-01 -2.09170267e-01 1.55932951e+00
-2.67279804e-01 4.95684266e-01 -3.28033298e-01 -9.21819746e-01
7.57500052e-01 9.76524472e-01 1.40668643e+00 -3.41981471e-01
3.18580091e-01 2.93071270e-02 -1.39213249e-01 -8.03537548e-01
-2.62454420e-01 -5.58922529e-01 2.17667133e-01 2.74637371e-01
-8.13054442e-02 2.60572612e-01 -3.63819182e-01 -1.62608594e-01
1.58592904e+00 1.22597339e-02 5.44453144e-01 2.55097332e-03
2.43008569e-01 1.15155466e-01 6.97126567e-01 8.46797943e-01
-4.91442800e-01 8.30568373e-01 4.11521852e-01 -8.67161274e-01
-9.91989255e-01 -1.19541740e+00 -4.65787888e-01 6.89220309e-01
-4.39915866e-01 -4.33476955e-01 -1.97300717e-01 -2.72559881e-01
5.13161361e-01 5.56773067e-01 -7.99141169e-01 -3.26684207e-01
-6.98199034e-01 -8.66615295e-01 6.41819239e-01 9.80075181e-01
-1.72472060e-01 -1.03852594e+00 -1.10549986e+00 6.77426696e-01
-1.20888047e-01 -9.34962153e-01 -1.85093150e-01 5.22613645e-01
-1.57214606e+00 -1.07612634e+00 -5.57613075e-01 -3.05617899e-01
6.75454959e-02 -3.64647925e-01 1.22778869e+00 5.13571955e-04
-4.85105455e-01 5.71525693e-01 -1.51396871e-01 -8.15812886e-01
-5.15105486e-01 -1.80972144e-01 2.51576453e-01 -1.18129596e-01
5.63629091e-01 -7.36751318e-01 -1.01387811e+00 -3.05319041e-01
-6.34759247e-01 3.57984453e-02 2.31339037e-01 9.30351079e-01
7.38404572e-01 -6.68960452e-01 1.04730630e+00 -8.29274178e-01
7.74504125e-01 -1.03634512e+00 -8.56466442e-02 1.50513843e-01
-1.22238195e+00 -1.15517974e-01 1.02696133e+00 -3.67307574e-01
-5.14334142e-01 1.25107018e-03 5.10360561e-02 -9.73546326e-01
-5.09796262e-01 8.15522671e-01 5.25461137e-01 4.83492732e-01
5.05782008e-01 3.05710077e-01 2.87129402e-01 -6.31256104e-01
-1.24477945e-01 5.38642764e-01 6.89063430e-01 -4.36721355e-01
6.50687423e-03 1.98864445e-01 2.93003082e-01 -2.25461274e-01
-5.94304085e-01 -5.36092281e-01 -8.79455090e-01 6.66757375e-02
9.47920680e-01 -1.14961088e+00 -1.14704978e+00 3.15179616e-01
-9.21822071e-01 -5.88815749e-01 -6.50922060e-01 6.81411266e-01
-8.41408730e-01 2.21825719e-01 -1.05594397e+00 -7.34018207e-01
-7.52452850e-01 -6.10331833e-01 1.08250618e+00 -9.71721709e-02
-5.75819850e-01 -1.47540069e+00 3.52736980e-01 -9.74898189e-02
7.05355525e-01 9.75117922e-01 9.99420285e-01 -1.04238498e+00
-2.35463440e-01 -5.39327204e-01 -4.96966280e-02 1.75850287e-01
4.33323294e-01 -1.22566104e-01 -8.22135031e-01 -2.63893664e-01
2.47063473e-01 -2.53605038e-01 6.93973958e-01 6.79805934e-01
1.47656167e+00 -3.58367175e-01 -3.18420529e-01 1.07576966e+00
1.45230186e+00 4.69247878e-01 2.89146572e-01 -1.07417107e-01
7.63818800e-01 8.61827135e-02 5.55157987e-03 1.12047970e+00
4.90461677e-01 -7.24894926e-02 4.01191488e-02 -1.74596608e-01
1.83276713e-01 -4.10874896e-02 6.27604723e-02 9.69225764e-01
-1.74456924e-01 -9.04127955e-02 -1.27020836e+00 4.78921056e-01
-1.95818818e+00 -8.25288415e-01 -2.50286609e-01 2.08287215e+00
1.02827048e+00 -4.14427102e-01 -4.22244752e-03 -5.80466725e-03
2.19711408e-01 -3.17941576e-01 -1.03509402e+00 -2.39008337e-01
5.59114367e-02 3.08269769e-01 2.95978069e-01 -2.85088159e-02
-9.79372859e-01 6.50114194e-02 7.75087166e+00 -3.79144460e-01
-1.17760718e+00 -1.03258211e-02 8.40559721e-01 -4.18122321e-01
3.04267675e-01 -5.28674424e-01 -1.63771972e-01 5.38793147e-01
1.82381928e+00 -3.35158676e-01 3.39901745e-01 5.46299577e-01
5.58993101e-01 3.53306025e-01 -1.93914211e+00 1.27441823e+00
-4.09386098e-01 -1.58715367e+00 -3.34477484e-01 3.70045453e-02
5.82662821e-01 4.24703658e-01 8.89146794e-03 2.75203615e-01
3.02288651e-01 -1.37583363e+00 -2.20372574e-03 1.21817911e+00
9.10906851e-01 -1.53184846e-01 9.60090637e-01 2.99697638e-01
-8.48892212e-01 -3.72465193e-01 2.31563419e-01 -1.59805611e-01
1.96148142e-01 4.45336163e-01 -1.11435032e+00 4.14127946e-01
6.19239211e-01 1.32009804e+00 -3.08701575e-01 1.09703970e+00
5.24366140e-01 9.38787341e-01 -2.80910015e-01 6.15710258e-01
-2.57470161e-01 1.15443304e-01 1.62873909e-01 1.20179284e+00
5.19843876e-01 5.86667717e-01 2.33585387e-01 8.17419171e-01
5.89978881e-02 9.22731869e-03 -8.48647475e-01 -1.83878899e-01
2.82711625e-01 7.81586289e-01 3.32759693e-02 -8.67942333e-01
-6.29390061e-01 4.89609212e-01 3.90764587e-02 4.45305943e-01
-6.75965011e-01 1.26740068e-01 8.88860643e-01 2.35082284e-01
7.08096325e-02 -1.06474876e-01 -1.13646674e+00 -1.45940435e+00
-4.55904096e-01 -7.05736697e-01 7.59260297e-01 -7.26997674e-01
-1.74918771e+00 3.30113709e-01 -3.46159041e-01 -1.39388168e+00
-4.97226417e-01 -4.05188769e-01 -6.25558674e-01 1.24207091e+00
-1.44474685e+00 -6.06811941e-01 -5.28473616e-01 7.78369248e-01
3.16707015e-01 -6.05188757e-02 1.27397275e+00 2.11446598e-01
-6.73928678e-01 5.04815280e-01 4.92151111e-01 2.33640701e-01
8.75694156e-01 -1.22736549e+00 -5.42057119e-02 2.28139341e-01
-7.18508184e-01 9.08622861e-01 3.82309467e-01 -7.20005035e-01
-1.68411064e+00 -1.39030945e+00 7.80728042e-01 -1.10692894e+00
6.43380165e-01 3.58961612e-01 -1.42251730e+00 8.96851361e-01
1.06317557e-01 3.75501394e-01 1.30976057e+00 -6.08717166e-02
-1.43157005e-01 -1.27698362e-01 -1.24441397e+00 1.68316841e-01
5.19756436e-01 -5.48197329e-01 -6.93257868e-01 1.27538517e-01
7.80198514e-01 -2.84074813e-01 -2.05144095e+00 9.08841729e-01
7.41052032e-01 -6.66953385e-01 9.98032391e-01 -1.14025915e+00
7.81187534e-01 1.56483039e-01 -1.13947084e-02 -1.14924252e+00
-3.51112276e-01 -6.74189866e-01 -8.92814219e-01 5.84552884e-01
1.53985051e-02 -7.73736835e-01 3.47097576e-01 1.13024676e+00
-1.98310584e-01 -1.04488897e+00 -5.74095726e-01 -2.73337752e-01
2.22816318e-01 -1.93847954e-01 7.96298087e-01 1.24113548e+00
4.48789835e-01 2.29997486e-01 -4.05383855e-01 -4.92638946e-02
3.89059633e-01 1.64078057e-01 3.69653255e-01 -1.62743795e+00
-3.12661529e-01 -3.98784518e-01 -1.97968692e-01 -3.66805345e-01
-2.78954864e-01 -8.88079941e-01 -6.11517318e-02 -1.72801602e+00
1.52765840e-01 -3.98248732e-01 -1.25595486e+00 6.80792093e-01
-4.83908474e-01 -1.89758129e-02 7.40979761e-02 4.18916404e-01
-3.67058933e-01 4.14393097e-01 9.63098764e-01 1.61689535e-01
-5.58098793e-01 -1.88703611e-01 -5.23333251e-01 3.46649617e-01
9.07890320e-01 -2.88083345e-01 -9.80409384e-02 -3.31035465e-01
-3.04997176e-01 1.04267430e+00 3.63937318e-01 -9.91616786e-01
2.84081191e-01 -2.30647162e-01 8.80890191e-01 -4.79460448e-01
1.43045023e-01 -9.16979492e-01 2.96676368e-01 7.31409788e-01
-6.80019855e-01 4.35731560e-01 4.21817005e-01 8.02549779e-01
-1.63173869e-01 7.05173790e-01 6.90399826e-01 -6.48397058e-02
6.27646595e-02 6.59990370e-01 -4.86422181e-01 2.13475693e-02
7.58541524e-01 -6.21186756e-02 -2.46976614e-01 -2.27794483e-01
-1.14141881e+00 3.47584784e-01 2.41683722e-01 2.67912120e-01
6.28939986e-01 -1.13871181e+00 -8.85416985e-01 3.11249137e-01
1.28923386e-01 -3.51650976e-02 2.46040240e-01 1.42478812e+00
-4.07649070e-01 5.29097855e-01 -2.62328982e-01 -6.50547445e-01
-9.52570975e-01 7.93778241e-01 4.77497697e-01 -1.35994375e-01
-1.02152908e+00 3.14801961e-01 7.45106637e-02 -2.25788951e-01
3.29379112e-01 -7.48741806e-01 -8.44393596e-02 -1.38127223e-01
5.07869303e-01 3.84911478e-01 4.63870801e-02 -1.01806708e-02
-2.82947153e-01 5.22573851e-02 3.02167833e-01 3.80644858e-01
1.74779272e+00 -2.18624864e-02 -2.38727421e-01 1.04402077e+00
1.31155515e+00 -6.25047863e-01 -1.22496557e+00 -3.19985062e-01
-1.00726500e-01 -6.43595681e-02 -1.68072328e-01 -1.07993388e+00
-7.04297602e-01 1.02064717e+00 7.20092654e-01 2.39229470e-01
1.32694578e+00 -2.15652689e-01 1.12205505e+00 4.48234171e-01
3.71705485e-03 -5.61747074e-01 -2.53658056e-01 5.31629086e-01
4.95567322e-01 -1.30888665e+00 -2.42661282e-01 2.73892522e-01
-5.80939770e-01 1.35854423e+00 3.33804905e-01 -3.29870954e-02
9.87959623e-01 3.79278660e-01 1.85799032e-01 -3.92130390e-03
-1.57344031e+00 3.62116337e-01 8.34124163e-02 5.06602764e-01
8.79440725e-01 2.07317963e-01 -1.01203330e-01 9.92121398e-01
1.32038549e-01 7.62208104e-01 4.74062294e-01 1.01370800e+00
-7.40091950e-02 -7.20181346e-01 -2.46003270e-01 1.13801229e+00
-7.85563469e-01 -3.54906470e-01 -6.17864691e-02 1.86719030e-01
4.66335425e-03 7.26329148e-01 1.10612392e-01 -3.70729834e-01
3.01968366e-01 9.55824435e-01 2.28155628e-01 -5.16138434e-01
-1.03260994e+00 2.20742431e-02 -2.07050771e-01 -6.08514845e-01
-3.13219689e-02 -7.65185714e-01 -1.44851935e+00 -2.70783365e-01
4.96543825e-01 -1.91969469e-01 3.65678310e-01 8.56440544e-01
9.81486380e-01 1.00620949e+00 4.06619310e-01 -5.82673669e-01
-7.14045167e-01 -9.60129201e-01 -3.49627048e-01 5.08501470e-01
1.05775130e+00 -1.30795285e-01 -2.23742604e-01 5.36543310e-01] | [7.971783638000488, 6.209127902984619] |
9d7068df-fe91-4abb-a4aa-cec50af73c4e | cross-domain-neural-pitch-and-periodicity | 2301.12258 | null | https://arxiv.org/abs/2301.12258v2 | https://arxiv.org/pdf/2301.12258v2.pdf | Cross-domain Neural Pitch and Periodicity Estimation | Pitch is a foundational aspect of our perception of audio signals. Pitch contours are commonly used to analyze speech and music signals and as input features for many audio tasks, including music transcription, singing voice synthesis, and prosody editing. In this paper, we describe a set of techniques for improving the accuracy of widely-used neural pitch and periodicity estimators to achieve state-of-the-art performance on both speech and music. We also introduce a novel entropy-based method for extracting periodicity and per-frame voiced-unvoiced classifications from statistical inference-based pitch estimators (e.g., neural networks), and show how to train a neural pitch estimator to simultaneously handle both speech and music data (i.e., cross-domain estimation) without performance degradation. While neural pitch trackers have historically been significantly slower than signal processing based pitch trackers, our estimator implementations approach the speed of state-of-the-art DSP-based pitch estimators on a standard CPU, but with significantly more accurate pitch and periodicity estimation. Our experiments show that an accurate, cross-domain pitch and periodicity estimator written in PyTorch with a hopsize of ten milliseconds can run 11.2x faster than real-time on a Intel i9-9820X 10-core 3.30 GHz CPU or 408x faster than real-time on a NVIDIA GeForce RTX 3090 GPU, without hardware optimization. We release all of our code and models as Pitch-Estimating Neural Networks (penn), an open-source, pip-installable Python module for training, evaluating, and performing inference with pitch- and periodicity-estimating neural networks. The code for penn is available at https://github.com/interactiveaudiolab/penn. | ['Bryan Pardo', 'Nathan Pruyne', 'Caedon Hsieh', 'Max Morrison'] | 2023-01-28 | null | null | null | null | ['music-transcription', 'singing-voice-synthesis'] | ['music', 'speech'] | [-1.84897751e-01 -5.22046983e-01 -1.66715294e-01 -1.26715645e-01
-9.23319817e-01 -6.70806885e-01 -1.94226220e-01 -1.81904390e-01
-3.96597028e-01 4.10956681e-01 6.39253110e-02 -3.50613207e-01
1.67428911e-01 -5.31517029e-01 -4.46985573e-01 -6.04523659e-01
-3.07217479e-01 1.29170701e-01 8.91629457e-02 8.62190053e-02
-7.99335986e-02 3.31665009e-01 -2.13821149e+00 7.67135248e-03
2.70799756e-01 1.05459142e+00 9.49034914e-02 1.43899810e+00
3.15656841e-01 3.54876131e-01 -1.02735019e+00 2.02564433e-01
6.30374551e-02 -3.74318272e-01 -3.75992030e-01 -5.86621404e-01
4.24046963e-01 -3.36062729e-01 -1.06451049e-01 1.06722176e+00
1.08947372e+00 3.22143316e-01 1.00116491e-01 -1.09722602e+00
9.25153717e-02 8.66994619e-01 -3.08145463e-01 3.96373242e-01
2.44958341e-01 1.94078892e-01 8.99524391e-01 -7.05800474e-01
-9.54967290e-02 1.00700116e+00 1.07502210e+00 4.30367559e-01
-9.90665138e-01 -1.17639863e+00 -5.83423793e-01 -7.31397495e-02
-1.51893330e+00 -9.41234767e-01 7.22431600e-01 -1.57870382e-01
1.24619830e+00 5.27217031e-01 7.93119133e-01 9.87175047e-01
2.78568012e-03 4.57411766e-01 6.49887145e-01 -4.54647779e-01
2.10146651e-01 -2.71274537e-01 3.71174030e-02 3.75104666e-01
-3.72483522e-01 6.04349792e-01 -1.09495199e+00 -5.12880921e-01
1.07333720e+00 -8.37531447e-01 -3.17517042e-01 6.56816125e-01
-1.28109717e+00 5.53906798e-01 -2.55598545e-01 1.13925226e-01
-5.08736551e-01 5.54528773e-01 8.91567111e-01 2.87866503e-01
4.75595891e-01 4.77044404e-01 -7.80338764e-01 -1.11400807e+00
-1.43777561e+00 4.35434550e-01 1.17557979e+00 5.12696981e-01
2.63153404e-01 9.04117644e-01 8.20782259e-02 9.84007716e-01
6.96327835e-02 7.03308463e-01 8.59489143e-01 -1.36402881e+00
-1.03860147e-01 -6.06880903e-01 -2.19976217e-01 -7.00297356e-01
-5.99616408e-01 -4.97201979e-01 -6.36373639e-01 1.67477503e-01
5.35122871e-01 -5.04722536e-01 -3.68493021e-01 1.92037356e+00
4.81990516e-01 4.05982643e-01 -1.21060908e-01 9.84452486e-01
7.39791155e-01 9.32226241e-01 -1.63808614e-01 -3.96765143e-01
1.65217900e+00 -9.30117965e-01 -7.74870038e-01 2.41661519e-02
6.40648454e-02 -1.24452496e+00 1.32211971e+00 9.27927732e-01
-1.32951581e+00 -9.69817042e-01 -1.14511979e+00 -4.32294980e-02
1.77731559e-01 2.57621884e-01 6.69789374e-01 8.37414980e-01
-1.05920279e+00 9.92453694e-01 -1.18742645e+00 1.44816607e-01
-2.86809921e-01 4.75324482e-01 2.49331981e-01 1.14130163e+00
-1.24348259e+00 1.87082261e-01 5.95337927e-01 -3.32994789e-01
-5.43441653e-01 -1.13391376e+00 -5.37240803e-01 2.63395786e-01
1.57272909e-02 -3.42475474e-01 2.19648433e+00 -9.81024146e-01
-2.19559741e+00 3.50849658e-01 -7.12751150e-02 -5.72487414e-01
-6.86630234e-02 -3.40444982e-01 -7.42791116e-01 1.66582823e-01
-2.54445970e-01 8.23846996e-01 1.23968410e+00 -1.94382519e-01
-6.94762051e-01 1.28226280e-01 -6.32763147e-01 2.53233135e-01
-3.22226167e-01 2.55671114e-01 -2.33334213e-01 -8.70570183e-01
-1.34346142e-01 -8.85219097e-01 3.05812240e-01 -1.17787912e-01
-2.69875675e-01 -1.80369675e-01 6.75658941e-01 -8.98639083e-01
1.48523104e+00 -2.34589720e+00 -4.23693746e-01 -6.15146421e-02
-2.21312031e-01 3.96886408e-01 1.59166437e-02 1.44638330e-01
-2.18450114e-01 -3.00784409e-01 1.11625597e-01 -3.36974084e-01
1.77380264e-01 -8.97554755e-02 -6.60112262e-01 2.76552171e-01
-2.36791223e-01 2.73923278e-01 -6.96189344e-01 -2.11585298e-01
2.92722374e-01 9.14414346e-01 -7.33477890e-01 1.88905299e-01
-4.88283671e-02 3.46689433e-01 2.41845265e-01 6.03359640e-01
3.98646712e-01 2.90237039e-01 9.24920142e-02 -2.65739769e-01
-4.99153674e-01 1.06316245e+00 -1.41045213e+00 1.68722320e+00
-7.89298713e-01 1.09708071e+00 5.02390862e-01 -3.07930321e-01
8.83085251e-01 7.58009374e-01 2.59467542e-01 -2.47972101e-01
1.79392919e-01 4.69940007e-01 3.25005114e-01 -2.59227436e-02
9.05352712e-01 -6.05451502e-02 9.27793011e-02 6.20749414e-01
2.16542900e-01 -3.30659539e-01 8.56648907e-02 -4.41597939e-01
8.52655590e-01 -1.53215388e-02 4.57732350e-01 -2.50323385e-01
9.70977917e-02 -3.62674505e-01 6.43613577e-01 4.52640504e-01
-2.97252446e-01 4.54231232e-01 2.38146648e-01 -2.97633260e-01
-1.07831848e+00 -1.14907277e+00 -4.62379724e-01 1.62266016e+00
-7.41713941e-01 -6.56508863e-01 -1.03799093e+00 3.94085914e-01
-6.76951483e-02 6.60095870e-01 1.84718788e-01 1.56051740e-01
-6.51872754e-01 -4.35855240e-01 1.34298027e+00 6.50002897e-01
2.61666656e-01 -1.57241488e+00 -9.11782324e-01 6.08174860e-01
3.72679867e-02 -7.81847298e-01 -8.96882355e-01 4.97872442e-01
-7.47120261e-01 -4.03488070e-01 -3.70268047e-01 -5.85761726e-01
-3.38829786e-01 -1.35304064e-01 1.18636954e+00 -2.06353053e-01
-2.67953575e-01 1.68367147e-01 5.81318475e-02 -6.39133275e-01
-3.98570836e-01 3.30332190e-01 9.06685293e-01 -5.60728669e-01
2.77598232e-01 -1.20668948e+00 -4.91712660e-01 1.07538015e-01
-4.72202361e-01 -7.21330643e-02 4.17720675e-02 6.81044459e-01
7.03619957e-01 1.02338530e-01 7.43063271e-01 -3.58539037e-02
9.44456816e-01 -4.81727012e-02 -9.76656079e-01 -5.45616925e-01
-3.03219825e-01 -2.35634580e-01 7.89994657e-01 -9.20488238e-01
-5.33229649e-01 1.57223538e-01 -9.21697617e-01 -8.06108713e-01
-1.69844069e-02 3.13590676e-01 2.47037336e-01 2.18529582e-01
6.99468255e-01 1.17067657e-01 1.05399549e-01 -4.18576837e-01
1.74769267e-01 1.12979472e+00 1.09536600e+00 -6.60814643e-01
3.74866754e-01 -1.15800403e-01 -3.07622850e-01 -1.30376709e+00
-6.78817809e-01 -5.68458676e-01 -5.93363009e-02 -8.09041560e-02
6.02590621e-01 -1.10156894e+00 -1.43018794e+00 6.95714295e-01
-1.10235846e+00 -5.20332456e-01 -4.25726682e-01 8.33733916e-01
-9.30829585e-01 1.65700346e-01 -9.61863458e-01 -1.08638811e+00
-1.09287441e+00 -9.04295862e-01 1.02635133e+00 5.60861111e-01
-8.39410722e-01 -7.54920900e-01 4.73065495e-01 -1.90559477e-01
4.41467792e-01 -2.38786086e-01 3.85084331e-01 -3.08789253e-01
1.91543281e-01 1.61215514e-01 2.40565345e-01 5.99406183e-01
2.37903800e-02 4.95239794e-01 -1.57658780e+00 -2.02424228e-01
2.90812731e-01 -2.84497887e-01 3.40968996e-01 8.78434896e-01
1.34983146e+00 -3.00123185e-01 2.50704318e-01 8.87972057e-01
7.67325640e-01 2.50200331e-01 3.87042254e-01 -1.31265685e-01
2.94213861e-01 2.87036031e-01 5.51566482e-01 7.34836459e-01
-7.74316192e-02 7.38884687e-01 -8.27375650e-02 9.81307626e-02
-1.90581650e-01 -2.84027129e-01 6.39835775e-01 1.58488429e+00
8.19489211e-02 1.01217560e-01 -7.46954560e-01 3.90203565e-01
-1.33340466e+00 -1.09221661e+00 -1.78342536e-02 2.37716699e+00
1.43449771e+00 1.00750506e-01 5.49621344e-01 5.62661827e-01
7.28589177e-01 2.65170306e-01 -6.67383492e-01 -1.11394691e+00
3.91335905e-01 9.43183184e-01 3.27831984e-01 6.42771125e-01
-1.18237948e+00 8.59196424e-01 6.14876699e+00 1.20547569e+00
-1.54233289e+00 1.12471662e-01 -1.24933738e-02 -5.76482773e-01
3.20952624e-01 -3.76559258e-01 -9.18283761e-01 4.01715815e-01
1.73575437e+00 -3.01430196e-01 9.65157270e-01 1.23251748e+00
3.62877250e-01 5.15644699e-02 -1.04667604e+00 1.32528222e+00
-2.16902137e-01 -1.20279801e+00 -6.11902177e-01 2.22570486e-02
2.61345595e-01 1.71674013e-01 2.36240044e-01 4.87606168e-01
-1.66537791e-01 -8.46823633e-01 8.91988397e-01 1.01259775e-01
1.16232574e+00 -1.16842580e+00 3.17139804e-01 3.09163541e-01
-1.62554038e+00 2.79151559e-01 -3.89615268e-01 -4.44053620e-01
2.52187133e-01 7.52962053e-01 -9.31636095e-01 -1.51157510e-02
8.63776863e-01 3.65758270e-01 1.61008924e-01 9.01731372e-01
-2.35511377e-01 1.14383495e+00 -7.59055912e-01 -3.74102481e-02
-5.68216145e-02 3.23737711e-01 7.65338123e-01 1.31566131e+00
5.51396251e-01 1.25495225e-01 -7.08340481e-02 7.81732380e-01
-2.40694322e-02 -1.46737888e-01 -4.22886051e-02 -2.66264260e-01
1.08461213e+00 1.41397655e+00 -4.44992185e-01 -2.89396048e-01
5.72127756e-04 4.58072871e-01 -1.57417417e-01 -9.40201581e-02
-1.10738289e+00 -9.56366420e-01 1.35893786e+00 -2.45924428e-01
2.10747838e-01 -4.25233245e-01 -1.37491941e-01 -7.48551190e-01
-1.51115701e-01 -1.18275154e+00 -1.13153651e-01 -6.09736085e-01
-1.09995401e+00 6.16564333e-01 -3.08173001e-01 -1.28735018e+00
-8.50616395e-01 -5.58549166e-01 -8.65152478e-01 1.09795642e+00
-1.07109070e+00 -1.80954948e-01 -6.19750544e-02 3.51912796e-01
6.38873637e-01 4.00292315e-03 1.23067820e+00 3.95292670e-01
-3.68161112e-01 6.08228147e-01 -2.33966131e-02 2.37065442e-02
7.47412741e-01 -1.23976660e+00 1.00312531e+00 4.59826589e-01
4.29873258e-01 5.22928536e-01 8.36240828e-01 -3.57372522e-01
-1.36395669e+00 -8.78707826e-01 6.75371766e-01 -1.44755021e-02
8.04317176e-01 -3.10214490e-01 -1.02793598e+00 3.56448889e-01
2.38732457e-01 -1.63375944e-01 9.25833046e-01 4.10500765e-01
-1.84559926e-01 -1.87378034e-01 -5.73895514e-01 4.86635566e-01
6.00974023e-01 -8.55141699e-01 -6.20899022e-01 1.58613265e-01
8.26476097e-01 -9.26153123e-01 -9.84556019e-01 2.06378639e-01
1.00773585e+00 -1.06083596e+00 8.76566708e-01 2.28868097e-01
-1.18855625e-01 -3.91229391e-01 -1.12224497e-01 -1.17497742e+00
3.17890793e-02 -1.17574036e+00 -3.67996693e-01 1.20204282e+00
1.66061491e-01 -5.34470081e-01 6.78307891e-01 -1.68054819e-01
-3.38116676e-01 -4.26250786e-01 -1.21208906e+00 -9.25984621e-01
-1.08451054e-01 -9.93325710e-01 4.91253465e-01 6.89350367e-01
3.05665851e-01 2.73581922e-01 -2.76678175e-01 2.74581432e-01
4.58467245e-01 -1.34151444e-01 6.73967779e-01 -1.03322887e+00
-8.02506387e-01 -5.96249521e-01 -1.94480658e-01 -8.97994220e-01
1.00521497e-01 -5.26494205e-01 2.81113923e-01 -5.33908665e-01
-4.94196057e-01 -1.60682485e-01 -1.65058613e-01 5.20311117e-01
1.68782905e-01 3.06705326e-01 2.89642662e-01 -1.99172776e-02
7.50104859e-02 3.95458758e-01 7.06188023e-01 3.48320603e-01
-6.15387321e-01 3.93581808e-01 -2.95885317e-02 1.06218588e+00
1.01085448e+00 -6.24554396e-01 -4.07380581e-01 -1.33901745e-01
1.05404958e-01 5.27439535e-01 3.00169438e-01 -1.78245890e+00
3.59515488e-01 3.67205083e-01 2.89541185e-01 -8.72179031e-01
7.92503417e-01 -6.39065057e-02 2.08019838e-01 4.81568396e-01
-2.09300190e-01 2.00083792e-01 8.16735208e-01 -1.55130044e-01
-3.89162511e-01 -1.67451546e-01 7.39624202e-01 2.15557784e-01
-2.71569878e-01 -4.07873988e-02 -5.51088274e-01 1.51791915e-01
3.43546301e-01 -1.37645723e-02 -2.70198733e-01 -5.29264808e-01
-7.22369373e-01 -3.83366644e-01 6.68483526e-02 2.33510792e-01
3.81522775e-01 -1.24618173e+00 -5.96739948e-01 4.30659264e-01
-5.74234247e-01 -3.02335978e-01 3.80227208e-01 6.61340892e-01
-5.84386051e-01 5.47583401e-01 -2.32346848e-01 -7.66352355e-01
-1.57417655e+00 6.99848235e-02 4.87909406e-01 4.20877768e-05
-5.86129606e-01 9.26179409e-01 -1.47352442e-01 -4.44085419e-01
5.84408402e-01 -7.50214577e-01 1.68072298e-01 5.12488596e-02
8.32536519e-01 4.81528282e-01 1.31084055e-01 -1.68975294e-01
-3.90561283e-01 4.41006780e-01 4.19022352e-01 -6.33556902e-01
7.35770822e-01 3.79020095e-01 -5.81388436e-02 8.55090916e-01
1.03041422e+00 3.66606236e-01 -1.12886965e+00 2.46079251e-01
-3.81747246e-01 -1.75833534e-02 4.08891737e-01 -5.64121723e-01
-8.86758745e-01 1.08228970e+00 6.64038122e-01 2.00190648e-01
1.40604138e+00 -4.15167004e-01 1.16743159e+00 4.30029839e-01
2.12400228e-01 -1.16956520e+00 -1.54301494e-01 8.07976842e-01
6.35677040e-01 -5.13229191e-01 -7.92244226e-02 -4.73443978e-02
-4.74330842e-01 1.43195891e+00 3.97931993e-01 -1.91377327e-01
7.27402210e-01 8.43155861e-01 1.49322942e-01 3.37979704e-01
-1.03816855e+00 7.87937418e-02 3.02606493e-01 4.42464978e-01
8.92529368e-01 4.84100848e-01 3.97582315e-02 7.63183117e-01
-1.26185286e+00 -6.00580052e-02 3.63110095e-01 5.55957377e-01
-4.98748183e-01 -9.43336606e-01 -6.12095594e-01 1.18339382e-01
-8.36434782e-01 -5.23987889e-01 1.30267411e-01 6.82742774e-01
-1.69650912e-01 8.60503733e-01 6.59055352e-01 -6.38265550e-01
9.53802541e-02 2.66826957e-01 3.22955370e-01 -5.30736566e-01
-1.20615005e+00 5.16480565e-01 3.36339384e-01 -5.88355005e-01
-2.73022186e-02 -7.03866482e-01 -1.32938719e+00 -6.02893472e-01
-1.77373841e-01 2.56955683e-01 1.10577643e+00 4.02301371e-01
3.34120095e-01 8.37619305e-01 5.89518070e-01 -1.15376961e+00
-7.36736059e-01 -1.04031718e+00 -8.24761450e-01 -6.36113644e-01
4.02984589e-01 -2.27002099e-01 -5.66672564e-01 -3.79954427e-02] | [15.516680717468262, 5.812486171722412] |
bbdee48d-6cd3-4e3b-bf5f-801372abaf1b | modifying-optimal-sat-based-approach-to-multi | 1707.00228 | null | http://arxiv.org/abs/1707.00228v1 | http://arxiv.org/pdf/1707.00228v1.pdf | Modifying Optimal SAT-based Approach to Multi-agent Path-finding Problem to Suboptimal Variants | In multi-agent path finding (MAPF) the task is to find non-conflicting paths
for multiple agents. In this paper we focus on finding suboptimal solutions for
MAPF for the sum-of-costs variant. Recently, a SAT-based approached was
developed to solve this problem and proved beneficial in many cases when
compared to other search-based solvers. In this paper, we present SAT-based
unbounded- and bounded-suboptimal algorithms and compare them to relevant
algorithms. Experimental results show that in many case the SAT-based solver
significantly outperforms the search-based solvers. | ['Eli Boyarski', 'Pavel Surynek', 'Ariel Felner', 'Roni Stern'] | 2017-07-02 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [-9.07919109e-02 2.78972656e-01 -3.29646587e-01 -7.82308728e-02
-5.30278802e-01 -8.72412801e-01 2.11841747e-01 2.62438834e-01
-1.09431349e-01 1.61159599e+00 -2.33472452e-01 -3.51403654e-01
-9.12957191e-01 -1.01660323e+00 -2.83749372e-01 -4.06025022e-01
-7.79065013e-01 1.35860598e+00 6.16781473e-01 -5.06775200e-01
2.82116294e-01 4.01154459e-01 -1.01131535e+00 4.29877639e-02
7.83760488e-01 4.39958543e-01 1.57398000e-01 4.18560952e-01
-2.57378906e-01 3.98273408e-01 -6.24260545e-01 -4.03907329e-01
6.44701660e-01 -5.94271183e-01 -1.24478662e+00 -5.27373590e-02
-1.14789344e-01 -3.38361189e-02 -6.27995655e-02 9.66475904e-01
7.80593231e-03 6.89664930e-02 2.06135035e-01 -2.18508649e+00
3.12999338e-01 7.90255427e-01 -7.71838248e-01 2.81115592e-01
9.87015188e-01 -1.62939966e-01 1.06820345e+00 -1.14071585e-01
1.02187335e+00 1.29840350e+00 4.16091144e-01 3.96587133e-01
-1.07723606e+00 -2.33586684e-01 6.13503397e-01 6.10026300e-01
-1.25630462e+00 -8.07750896e-02 2.54357666e-01 1.56556636e-01
1.67762458e+00 7.29617715e-01 8.63325715e-01 2.92796016e-01
3.23738813e-01 5.13289094e-01 1.27888894e+00 -4.72188711e-01
3.36339861e-01 -7.69085437e-02 9.37638655e-02 5.57099342e-01
6.98767781e-01 3.50501269e-01 -3.10077518e-01 -5.43713868e-01
6.17616415e-01 -6.45995021e-01 -2.55844951e-01 -4.19368178e-01
-1.15808475e+00 1.18679893e+00 3.11948657e-01 5.10422997e-02
-4.41481471e-01 3.31900775e-01 2.41483480e-01 4.53921676e-01
6.42029271e-02 8.24213505e-01 -5.84701121e-01 -5.31630851e-02
-5.95598996e-01 8.18343341e-01 1.23021936e+00 1.12591457e+00
4.06894952e-01 -1.98019415e-01 2.64013618e-01 1.48741305e-01
2.10601121e-01 2.21624836e-01 -5.52129388e-01 -1.33772719e+00
6.93697214e-01 5.51954269e-01 5.39878547e-01 -1.00318193e+00
-8.80014300e-01 -2.40110800e-01 -3.37861866e-01 6.40866339e-01
4.75768387e-01 -1.68097109e-01 -5.42504907e-01 1.57336211e+00
7.21174598e-01 1.04172014e-01 1.96836695e-01 9.74044025e-01
5.01086116e-01 1.01311696e+00 -5.56491971e-01 -1.02043498e+00
9.59592819e-01 -1.56422162e+00 -7.96991467e-01 -3.42757136e-01
5.48850954e-01 -5.75867593e-01 3.26326489e-02 6.84772313e-01
-1.65390646e+00 5.14949739e-01 -9.48035657e-01 5.99178731e-01
-4.30204898e-01 -9.46078181e-01 8.06526780e-01 5.82075477e-01
-1.44714046e+00 1.98193938e-01 -4.62774605e-01 -5.93338430e-01
2.51643155e-02 7.33003020e-01 -3.44667941e-01 -4.83356178e-01
-9.70850706e-01 1.31896913e+00 5.41763604e-01 -4.68501002e-02
-5.57435274e-01 -2.12186664e-01 -5.63314676e-01 9.64758247e-02
1.23293042e+00 -9.30768251e-01 1.35205293e+00 -5.02759159e-01
-1.48287308e+00 4.33674902e-01 -3.59402776e-01 -5.67822039e-01
4.72242892e-01 6.20207250e-01 -2.61559635e-01 1.44387051e-01
2.79445887e-01 4.04343307e-01 1.24680541e-01 -1.30562568e+00
-9.23886359e-01 2.85720788e-02 7.85976768e-01 2.11250991e-01
3.88393551e-01 4.20526475e-01 -2.43562341e-01 -5.32575250e-02
3.79220508e-02 -1.03363681e+00 -9.57021236e-01 -3.15880805e-01
-3.88409734e-01 -1.50878847e-01 4.40095663e-01 -2.80555384e-03
1.30702353e+00 -1.29449093e+00 6.10691667e-01 7.74802566e-01
5.10815457e-02 -1.05892412e-01 -6.41752481e-01 1.08827114e+00
8.72360915e-02 -1.12962954e-01 -6.20862767e-02 7.43868724e-02
1.93361208e-01 6.82048678e-01 4.86422628e-02 5.65356135e-01
-8.20686892e-02 7.23230600e-01 -1.07856846e+00 -6.19756997e-01
1.41134728e-02 -3.80171090e-01 -7.54315495e-01 -2.55774349e-01
-5.94980955e-01 -7.61202676e-03 -3.93436551e-01 6.24318957e-01
9.28632557e-01 1.70525555e-02 7.48418927e-01 6.28972232e-01
-5.56561112e-01 2.38050595e-01 -1.54367197e+00 1.52482224e+00
-1.52904615e-01 3.14270347e-01 6.73477769e-01 -8.26332688e-01
3.67704064e-01 9.65599194e-02 5.83568633e-01 -8.49313200e-01
8.36197808e-02 4.25422013e-01 2.74717808e-01 -1.71547234e-01
4.29964721e-01 -5.46170212e-03 -2.55604744e-01 7.39544868e-01
-6.21579409e-01 -8.38913471e-02 1.01300836e+00 4.11719888e-01
1.57364357e+00 -3.16563606e-01 5.65351248e-01 -4.38374192e-01
7.90713608e-01 9.36577141e-01 7.46308684e-01 6.12571657e-01
-1.46510348e-01 -1.55681625e-01 6.94781840e-01 -5.74408770e-01
-5.33073664e-01 -6.99376941e-01 2.49068707e-01 6.25301242e-01
6.87995076e-01 -8.95504236e-01 -5.88465631e-01 -5.28830826e-01
8.12485516e-02 7.75897682e-01 -3.66799563e-01 4.10884321e-01
-9.54348743e-01 -7.21791983e-01 -1.66226193e-01 1.17037520e-01
-1.85582368e-03 -7.07395434e-01 -9.56928015e-01 7.84746110e-01
-2.38771707e-01 -1.06260788e+00 -1.79080650e-01 1.95478231e-01
-4.23455536e-01 -1.50306106e+00 -3.19280058e-01 -6.50364041e-01
7.64002860e-01 6.76465213e-01 9.32484567e-01 3.07324439e-01
-1.54046819e-01 2.35230863e-01 -5.57739139e-01 -2.68285215e-01
-1.71821669e-01 -7.73065304e-03 -1.16399601e-01 -8.56178880e-01
-1.63957551e-01 -3.14423770e-01 -9.67565030e-02 7.05179155e-01
-3.15650165e-01 -1.99668601e-01 2.23382279e-01 5.42958140e-01
4.50306028e-01 7.69644320e-01 6.19228542e-01 -5.87385118e-01
9.27449048e-01 -4.88447279e-01 -1.18159878e+00 3.79767537e-01
-6.10481560e-01 -2.43521526e-01 5.61420798e-01 2.98827495e-02
-5.83334923e-01 4.59789671e-02 4.77124959e-01 5.33328578e-02
1.86495468e-01 8.99622083e-01 2.78902445e-02 -6.26011491e-01
6.31754175e-02 -1.55514926e-01 -1.10720843e-01 8.19240212e-02
1.39046028e-01 -8.25653374e-02 9.57018510e-02 -6.11996293e-01
7.42637217e-01 3.02623540e-01 8.12032282e-01 -1.76086798e-01
-1.97785631e-01 -3.16662848e-01 9.17521939e-02 -4.19996679e-01
4.19386238e-01 -3.05147111e-01 -9.87461925e-01 3.04184239e-02
-1.41231382e+00 -4.77260917e-01 -1.12002976e-01 1.86258718e-01
-8.21194589e-01 2.05999330e-01 -3.07991147e-01 -1.10827792e+00
1.25574291e-01 -1.19890499e+00 3.57648551e-01 3.01476628e-01
-3.45452815e-01 -8.92547011e-01 6.57520652e-01 2.35605747e-01
4.68427151e-01 5.96409261e-01 6.64919496e-01 -5.95880091e-01
-8.38754535e-01 4.83197309e-02 -2.02728659e-01 -1.01583266e+00
-1.72836185e-01 -5.02992831e-02 1.65497705e-01 -5.91569602e-01
-4.79694277e-01 1.86679453e-01 1.17980942e-01 6.69005156e-01
1.92498818e-01 -6.27476752e-01 -1.01215935e+00 1.61704004e-01
1.73681128e+00 5.39953232e-01 3.51869255e-01 1.11165321e+00
-3.12145919e-01 9.15240169e-01 1.16674614e+00 6.30249143e-01
6.32888556e-01 1.00083029e+00 8.13645840e-01 3.13564062e-01
4.16924715e-01 5.01784742e-01 1.54089287e-01 9.26690027e-02
-5.46550043e-02 -8.90547276e-01 -1.14232326e+00 6.27572179e-01
-2.50196743e+00 -9.85502303e-01 -7.79478908e-01 1.61829495e+00
4.06449944e-01 2.34968141e-01 7.31221199e-01 1.16992950e-01
7.88166404e-01 -2.32405826e-01 -2.78442383e-01 -1.02950549e+00
-1.62580192e-01 7.31878951e-02 6.17570817e-01 9.98930335e-01
-7.58176506e-01 8.59958887e-01 7.64785099e+00 3.09119523e-01
-2.95025080e-01 5.51616475e-02 -1.60254687e-01 -6.62271678e-01
-2.24146917e-01 1.94768831e-01 -3.29345703e-01 7.33242556e-02
9.97662187e-01 -7.94717014e-01 9.32597339e-01 5.92428923e-01
3.37056994e-01 -5.94297767e-01 -1.03308356e+00 5.88765085e-01
-1.15374684e-01 -1.35817409e+00 -4.19808686e-01 4.25746650e-01
1.02312851e+00 -3.10823292e-01 -2.75941253e-01 -1.53566346e-01
7.04655707e-01 -9.66894567e-01 8.04495633e-01 -2.37757012e-01
1.76733032e-01 -1.18369210e+00 8.93079996e-01 3.07443917e-01
-1.36530292e+00 -4.38364655e-01 -1.74247190e-01 -7.04266131e-01
9.70339775e-01 2.54553169e-01 -7.95933664e-01 1.14224160e+00
7.42493331e-01 2.94587612e-01 3.64562035e-01 1.82480597e+00
-1.25150695e-01 -2.82733858e-01 -6.49277329e-01 -1.57638282e-01
8.26402843e-01 -4.36016977e-01 9.18923140e-01 9.34435129e-01
3.35045516e-01 6.53375506e-01 4.89028692e-01 6.75097167e-01
7.12930799e-01 -1.33364126e-01 -3.34408104e-01 2.42901091e-02
5.50539792e-01 1.05641484e+00 -1.19063938e+00 -1.21252127e-02
-3.36217910e-01 5.07041097e-01 1.70117632e-01 2.03007340e-01
-1.08410323e+00 -2.10907489e-01 9.07857478e-01 -1.13469213e-01
3.57298076e-01 -3.30915600e-01 -1.84640080e-01 -4.57528085e-01
-8.63297954e-02 -8.72758508e-01 7.76085734e-01 -6.04614735e-01
-9.82933342e-01 7.10495889e-01 5.27147591e-01 -8.97324920e-01
-3.97404075e-01 -3.77038896e-01 -9.10016954e-01 6.21308327e-01
-1.74070799e+00 -6.63530767e-01 -4.12357561e-02 6.16221249e-01
3.95577580e-01 -6.88481927e-02 7.49448061e-01 8.44549909e-02
-6.86853290e-01 8.20006132e-02 -5.73260263e-02 -7.96481133e-01
1.20625198e-01 -1.07352901e+00 5.04049053e-03 9.84152257e-01
-4.23991919e-01 3.41516882e-01 1.23432946e+00 -6.63269699e-01
-1.82008541e+00 -5.63771605e-01 8.36528659e-01 9.99180898e-02
8.22165728e-01 2.70973682e-01 -2.37781599e-01 9.05683815e-01
7.05931842e-01 -3.98914754e-01 4.34172928e-01 -4.91337776e-02
1.93680882e-01 1.03091359e-01 -1.48588300e+00 4.16444242e-01
1.16257834e+00 5.68087697e-01 -3.58379215e-01 6.00661576e-01
5.28976321e-01 -6.56912267e-01 -3.67733657e-01 7.16724247e-02
1.39421046e-01 -1.00537288e+00 8.85992825e-01 -6.16531312e-01
-3.09488531e-02 -6.80878162e-01 -1.19130664e-01 -1.58130538e+00
-7.13087559e-01 -9.27005231e-01 5.99307828e-02 7.28681624e-01
6.12045228e-01 -1.21412885e+00 9.25813973e-01 5.11755645e-01
-3.84074926e-01 -7.38431871e-01 -1.56501412e+00 -1.26499701e+00
-1.06163092e-01 1.10040963e-01 8.37722182e-01 8.93729389e-01
8.53709936e-01 4.51647192e-02 -2.05671296e-01 5.57814777e-01
8.12688351e-01 6.71505868e-01 6.40455961e-01 -1.11121023e+00
-2.29442403e-01 -7.28575528e-01 -1.24555930e-01 -4.29098785e-01
2.65957117e-01 -7.14384317e-01 -9.63926092e-02 -2.21132469e+00
1.74404204e-01 -5.66254139e-01 1.45833895e-01 6.92324400e-01
4.83566612e-01 -2.15950340e-01 4.01553422e-01 -1.98701262e-01
-1.03655255e+00 2.38768961e-02 1.28701138e+00 -2.80054212e-01
-2.36320004e-01 -5.83631322e-02 -4.07278895e-01 3.72051656e-01
1.02000535e+00 -7.13493645e-01 -4.13684398e-01 -2.28081346e-01
8.48601639e-01 6.90388083e-01 4.76384871e-02 -5.90327442e-01
5.59895396e-01 -1.16480422e+00 -6.75432980e-01 -7.15651274e-01
4.36559469e-01 -1.21890342e+00 8.24982285e-01 9.87596154e-01
1.01435728e-01 4.19590592e-01 3.28981549e-01 3.40496361e-01
-2.34500617e-01 -6.38618529e-01 2.92835027e-01 -4.19773757e-01
-6.59968972e-01 -6.59121126e-02 -6.27603412e-01 -2.38735393e-01
1.94783986e+00 -3.82996976e-01 -7.68552721e-01 -5.31171262e-01
-6.20431364e-01 9.36287761e-01 3.96079957e-01 -2.04109088e-01
6.90407395e-01 -1.02912843e+00 -8.05591047e-01 -4.43188041e-01
-1.60249263e-01 -7.14564174e-02 6.20918274e-02 1.20899320e+00
-8.52830529e-01 9.05309677e-01 -5.33050120e-01 1.99605711e-02
-1.51416874e+00 1.00238907e+00 3.27219963e-01 -6.89276695e-01
-3.65147948e-01 7.51818419e-01 -3.12519372e-01 -7.90653080e-02
1.17191419e-01 -1.30649164e-01 1.40949115e-01 -5.99141978e-02
5.77928305e-01 9.40686405e-01 -3.60997856e-01 -2.02137485e-01
-1.08288932e+00 5.07119477e-01 3.67322541e-03 -2.77335107e-01
1.69028986e+00 -3.03887665e-01 -6.99814916e-01 -3.87836874e-01
4.38844979e-01 -2.05298662e-01 -5.01166523e-01 9.97207537e-02
8.91750157e-02 -7.18670249e-01 -1.55669730e-02 -1.17477429e+00
-1.18571138e+00 -6.89446703e-02 -3.30323040e-01 7.42752135e-01
1.23837197e+00 8.65960568e-02 6.60319984e-01 4.80961651e-01
1.26149213e+00 -9.59841847e-01 -4.14579362e-01 6.37770712e-01
8.72894526e-01 -7.58469284e-01 3.91697854e-01 -1.12262249e+00
-3.25462550e-01 1.25164008e+00 8.19346070e-01 -6.50295615e-02
5.05948104e-02 7.84363687e-01 -4.11721110e-01 -3.68363500e-01
-1.21017063e+00 -3.45430851e-01 -5.42743862e-01 7.38320112e-01
-4.94581759e-01 5.29008638e-03 -7.98280478e-01 1.95844084e-01
-2.75257707e-01 -9.04818997e-02 1.03279579e+00 1.39079559e+00
-6.78382874e-01 -1.55147791e+00 -8.40490460e-01 -4.08150256e-02
1.75161764e-01 2.69289732e-01 -8.80709410e-01 1.05518603e+00
-5.83794825e-02 1.46521461e+00 -9.66516510e-02 2.61653718e-02
4.34714913e-01 -5.90403497e-01 1.09192669e+00 -1.21499300e-01
-8.57472897e-01 -5.95285976e-03 1.06214178e+00 -9.69797254e-01
-6.46812320e-01 -8.38769138e-01 -1.54494143e+00 -8.29895079e-01
-3.25480223e-01 4.86401290e-01 3.02598059e-01 6.65372133e-01
-9.05951485e-02 4.45581615e-01 4.06367242e-01 -6.92918539e-01
-2.27074891e-01 2.38876827e-02 -5.10260284e-01 -5.36207438e-01
3.73813510e-02 -8.68466914e-01 -3.24077874e-01 -8.93142045e-01] | [4.988466262817383, 1.8631861209869385] |
267b4e0e-cde4-4123-9cf1-978c1e2481fc | symmetric-dense-inception-network-for | null | null | https://openreview.net/forum?id=91RHZVOKj1M | https://openreview.net/pdf?id=91RHZVOKj1M | Symmetric Dense Inception Network for Simultaneous Cell Detection and Classification in Multiplex Immunohistochemistry Images | Deep-learning-based automatic analysis of the multiplex immunohistochemistry (mIHC) enables distinct cell populations to be localized on a large scale, providing insights into disease biology and therapeutic targets. However, standard deep-learning pipelines performed cell detection and classification as two-stage tasks, which is computationally inefficient and faces challenges to incorporate neighbouring tissue context for determining the cell identity. To overcome these limitations and to obtain a more accurate mapping of cell phenotypes, we presented a symmetric dense inception neural network for detecting and classifying cells in mIHC slides simultaneously. The model was applied with a novel stop-gradient strategy and a loss function accounted for class imbalance. When evaluated on an ovarian cancer dataset containing 6 cell types, the model achieved an F1 score of 0.835 in cell detection, and a weighted F1-score of 0.867 in cell classification, which outperformed separate models trained on individual tasks by 1.9% and 3.8% respectively. Taken together, the proposed method boosts the learning efficiency and prediction accuracy of cell detection and classification by jointly learning from both tasks. | ['Yinyin Yuan', 'Jonathan A. Ledermann', 'Teresa Marafioti', 'Ayse U. Akarca', 'Tami Grunewald', 'Hanyun Zhang'] | 2021-07-20 | null | null | null | miccai-workshop-compay-2021-9 | ['cell-detection'] | ['computer-vision'] | [ 2.11895183e-01 -1.11939706e-01 -2.45017275e-01 -1.47394225e-01
-1.14388585e+00 -5.26743174e-01 3.50556046e-01 7.53883064e-01
-7.40251005e-01 9.72978830e-01 -3.83531332e-01 -2.37760842e-01
1.75596967e-01 -6.28482699e-01 -3.85759085e-01 -1.35845923e+00
2.34168861e-02 5.40891349e-01 1.07542679e-01 3.06510419e-01
2.68003225e-01 7.68634617e-01 -1.35851645e+00 4.73663479e-01
6.41792059e-01 1.15992248e+00 3.17542367e-02 9.36961889e-01
-4.35787775e-02 7.07035124e-01 -4.61982965e-01 -2.30390042e-01
-1.71794593e-01 -1.31034106e-01 -4.39690709e-01 -1.73162684e-01
4.89658654e-01 -1.75958991e-01 5.88958412e-02 7.57047117e-01
7.62580991e-01 -5.03352344e-01 8.93455625e-01 -1.02824175e+00
-1.01008946e-02 1.19493559e-01 -8.04230332e-01 2.50161797e-01
-1.08140379e-01 2.11495459e-02 8.31941485e-01 -8.24836910e-01
7.21692562e-01 4.69996989e-01 8.83714914e-01 5.49565017e-01
-1.62319911e+00 -8.27186763e-01 -3.54347140e-01 -1.01112038e-01
-1.59924650e+00 -4.13352579e-01 2.18827665e-01 -6.82631850e-01
8.80403399e-01 2.61620224e-01 7.42905438e-01 6.99343920e-01
5.14876723e-01 8.37312698e-01 1.02336919e+00 -2.08635882e-01
3.16512853e-01 1.94809571e-01 1.54455658e-02 6.58084691e-01
3.43224525e-01 -2.50995904e-01 -3.33617210e-01 -1.65867552e-01
7.08568394e-01 2.57934928e-01 -2.90127993e-02 -1.35720998e-01
-1.29704046e+00 7.93744206e-01 2.29918823e-01 5.13344824e-01
-1.66654319e-01 -1.99163370e-02 6.41827106e-01 -2.15858575e-02
4.83395845e-01 3.81102294e-01 -4.94128644e-01 1.03304394e-01
-9.61033344e-01 1.02699371e-02 5.08147657e-01 2.79418290e-01
4.88908559e-01 -2.55853802e-01 -3.72883022e-01 7.88618982e-01
-4.59725186e-02 4.50842261e-01 6.28770709e-01 -7.03356624e-01
-2.31685176e-01 8.65215242e-01 -5.98196909e-02 -7.38502145e-01
-9.26209807e-01 -9.36331451e-01 -1.05419910e+00 2.22695351e-01
7.57606745e-01 -5.75378835e-02 -8.66701007e-01 1.50396633e+00
4.53572154e-01 4.00241166e-02 -9.44661051e-02 7.60944605e-01
7.19173670e-01 2.97902167e-01 2.25919113e-01 -2.02236086e-01
1.45170939e+00 -4.79360610e-01 -6.87168896e-01 1.30865827e-01
1.43673730e+00 -5.98964393e-01 7.41729200e-01 2.20548928e-01
-7.72493124e-01 -1.81719959e-01 -1.10431886e+00 -8.43866095e-02
-4.97968912e-01 4.94007409e-01 8.37732673e-01 4.76784259e-01
-1.08602822e+00 2.38732174e-01 -8.88684690e-01 -4.30624098e-01
1.11805475e+00 7.20314145e-01 -6.78308010e-01 5.37297726e-02
-6.68242574e-01 6.03670835e-01 4.85635251e-02 7.44097531e-02
-7.76985943e-01 -1.09796095e+00 -5.97322226e-01 1.08808920e-01
-2.75156468e-01 -5.28230011e-01 1.02616799e+00 -7.12274134e-01
-1.50974917e+00 1.43230915e+00 -2.18000889e-01 -5.40805340e-01
4.31101173e-01 3.77936602e-01 -1.01056673e-01 2.59349644e-01
1.91080719e-01 7.93000579e-01 8.79458264e-02 -7.54492581e-01
-1.11405778e+00 -4.99563873e-01 -4.48007196e-01 -8.29018727e-02
-4.03350174e-01 -1.24892257e-01 -2.31185377e-01 -2.91362315e-01
-8.60624015e-02 -8.06592584e-01 -2.26372823e-01 2.60753125e-01
-1.23715356e-01 9.35688093e-02 5.48482418e-01 -5.34438789e-01
9.38449919e-01 -2.13127756e+00 5.67948706e-02 1.60185233e-01
4.93234187e-01 3.17048699e-01 -2.45974604e-02 6.30940571e-02
2.65146405e-01 6.86441436e-02 -1.54021960e-02 -3.53759497e-01
-2.53104538e-01 -3.78417671e-01 2.27527231e-01 9.53677952e-01
3.60489964e-01 9.83650863e-01 -9.46521163e-01 -5.78960061e-01
2.55085379e-02 6.03619277e-01 -3.55288088e-01 1.49436057e-01
1.87230304e-01 2.69105792e-01 -5.81700467e-02 1.11255443e+00
6.16418004e-01 -4.67835635e-01 4.05560166e-01 -1.40217006e-01
1.40344158e-01 -1.82636067e-01 -6.21301949e-01 1.33419311e+00
-3.50358129e-01 9.32542205e-01 4.17907327e-01 -9.00894821e-01
8.02067459e-01 2.44856253e-01 6.67744517e-01 -6.99023902e-01
3.38612854e-01 5.16666234e-01 1.00164644e-01 -2.43134737e-01
-1.68281481e-01 -5.61247706e-01 -1.01251334e-01 1.69990495e-01
9.29803029e-02 3.34147304e-01 3.03766906e-01 -9.56782624e-02
1.29921854e+00 -3.78993094e-01 2.55529642e-01 -4.71339077e-01
7.56428003e-01 1.28279701e-01 7.44787812e-01 4.62479860e-01
-4.53715086e-01 6.13521814e-01 7.73679137e-01 -6.44243658e-01
-6.77207887e-01 -8.22056115e-01 -5.14825344e-01 1.02659178e+00
-8.94162953e-02 1.14781074e-01 -4.90009487e-01 -8.17808032e-01
3.17254364e-01 8.49718004e-02 -8.93606424e-01 -9.35363099e-02
-3.08995068e-01 -1.25743866e+00 7.58911788e-01 4.50651020e-01
1.26233041e-01 -7.27580667e-01 -3.32100421e-01 2.20991030e-01
-1.67119764e-02 -1.03515494e+00 -8.86796564e-02 4.49125171e-01
-7.84125149e-01 -1.30964124e+00 -7.82919347e-01 -1.09067833e+00
9.51489925e-01 6.11843131e-02 8.50601137e-01 2.31645942e-01
-7.96289802e-01 -3.00811827e-01 4.39040139e-02 -4.35856402e-01
-2.92271316e-01 4.76506114e-01 -1.07779071e-01 2.10989833e-01
6.38221502e-01 -3.00006539e-01 -7.54900455e-01 8.13444480e-02
-6.77810609e-01 8.73477906e-02 9.65750813e-01 1.22537398e+00
8.82891595e-01 -1.42298490e-01 9.78073716e-01 -1.18096542e+00
9.05743614e-02 -4.09578055e-01 -5.85899472e-01 -3.06028314e-02
-4.83380556e-01 -3.58470559e-01 6.64525330e-01 -2.93359071e-01
-8.49058390e-01 3.47113788e-01 -2.16954395e-01 4.26469110e-02
-1.37998179e-01 4.40251768e-01 5.62482402e-02 -2.77658701e-01
4.89120007e-01 2.29273364e-01 5.55543184e-01 -4.29264642e-03
-3.13065886e-01 7.12701857e-01 4.12367761e-01 7.89510980e-02
1.46249369e-01 7.17970788e-01 3.94554883e-01 -5.70310533e-01
-7.46270299e-01 -7.34145761e-01 -5.53140402e-01 -6.38131127e-02
7.85369515e-01 -1.18753290e+00 -8.95108879e-01 8.36540043e-01
-7.54422009e-01 -5.56842387e-01 1.16074227e-01 5.33979297e-01
-2.73731470e-01 4.47835885e-02 -1.12896943e+00 -4.28153664e-01
-5.80538809e-01 -8.49083364e-01 1.29519618e+00 2.10214868e-01
-3.16279650e-01 -1.14710355e+00 2.76993006e-01 2.83720821e-01
4.06990558e-01 5.39202869e-01 1.00106835e+00 -8.20101619e-01
-2.67375678e-01 -8.67997408e-01 -4.43564385e-01 -7.38916695e-02
2.52986401e-01 1.86589882e-01 -1.22121549e+00 -4.51053292e-01
-4.59895462e-01 -4.31255132e-01 8.69491220e-01 5.57067573e-01
1.18379998e+00 1.41120061e-01 -9.71813798e-01 7.71377504e-01
1.57104659e+00 7.77242780e-02 5.92951000e-01 3.13921779e-01
5.18303514e-01 5.32299757e-01 4.63564485e-01 3.42424572e-01
1.36915371e-01 3.55519444e-01 3.18644494e-01 -6.89413249e-01
-9.83558893e-02 1.83079123e-01 -1.14627719e-01 3.21711451e-01
3.06065410e-01 -1.26962155e-01 -1.05950344e+00 7.19143331e-01
-1.67324054e+00 -6.46998942e-01 -2.25828364e-01 2.07069063e+00
9.06395614e-01 1.23594619e-01 1.02988169e-01 1.11937076e-01
7.43461907e-01 -3.84350836e-01 -5.54878771e-01 -8.11395422e-02
-2.79550433e-01 2.01612106e-03 4.92991447e-01 3.83713514e-01
-1.19802856e+00 6.62722766e-01 6.79522610e+00 1.02310741e+00
-1.48826814e+00 -2.03484192e-01 1.24507689e+00 -4.20100749e-01
1.34315416e-01 -4.53535467e-01 -9.65100884e-01 4.95548338e-01
7.51943946e-01 1.28816813e-01 -2.17359021e-01 3.58580232e-01
3.69679779e-02 -1.87273785e-01 -1.14390242e+00 9.07568157e-01
-1.61412776e-01 -1.66351104e+00 -1.25260994e-01 4.12102282e-01
6.82304084e-01 2.01106612e-02 9.86532867e-02 4.15737659e-01
-1.46911368e-02 -1.13865507e+00 2.84321625e-02 3.99210513e-01
1.21615052e+00 -8.01031888e-01 1.50327110e+00 3.28994453e-01
-7.71763206e-01 -2.19768379e-02 -1.52626455e-01 -1.08921342e-01
-3.67904752e-01 8.67610514e-01 -1.27059805e+00 1.15856282e-01
3.92431408e-01 5.88711083e-01 -5.33171117e-01 9.67286825e-01
4.72933143e-01 3.99740726e-01 -2.65182674e-01 -3.40526432e-01
-8.10195655e-02 2.90750891e-01 -2.42988728e-02 1.70408833e+00
3.46508473e-01 -1.38187677e-01 -7.18813464e-02 5.16472638e-01
-2.23389357e-01 1.63871288e-01 -1.42576873e-01 -1.21688791e-01
4.63873595e-01 1.89180815e+00 -1.17327476e+00 -1.06739506e-01
-2.74407774e-01 5.72316170e-01 6.45890594e-01 1.83886886e-01
-8.41088891e-01 -6.10449731e-01 5.83021522e-01 3.33279103e-01
1.36210293e-01 3.82660419e-01 -5.29277146e-01 -7.35030353e-01
-3.95112842e-01 -4.36069429e-01 5.27119815e-01 -1.14019282e-01
-1.39470339e+00 2.65662313e-01 -6.68344975e-01 -1.14148986e+00
4.55451943e-02 -9.28975463e-01 -4.80426937e-01 7.11698771e-01
-1.62680137e+00 -1.14773619e+00 -2.98485190e-01 -1.04235217e-01
9.37210470e-02 -1.71673581e-01 1.05935788e+00 4.15478587e-01
-8.72670770e-01 8.62823665e-01 3.49927127e-01 1.75924048e-01
7.98445344e-01 -1.37267387e+00 -3.82052869e-01 3.30368072e-01
-4.08513755e-01 3.37548286e-01 2.83659101e-01 -2.67129362e-01
-1.19352090e+00 -1.30590045e+00 1.08458257e+00 -2.83268005e-01
5.39321601e-01 -5.98177314e-01 -8.25004160e-01 3.34306508e-01
-1.71795934e-01 4.32945043e-01 1.41288674e+00 9.64998677e-02
-3.85067761e-02 -4.16825771e-01 -1.32613325e+00 4.77297723e-01
5.01159489e-01 -4.18640196e-01 3.44268113e-01 3.89309466e-01
1.39695536e-02 -4.63291615e-01 -1.20228076e+00 4.82268453e-01
9.09556687e-01 -8.86677265e-01 6.89714551e-01 -3.55420589e-01
3.12467426e-01 -3.67003560e-01 4.95268628e-02 -7.31639087e-01
-6.59598291e-01 -1.83078423e-01 -1.54934898e-01 1.03165770e+00
7.11874902e-01 -6.29692614e-01 1.10314369e+00 2.43394762e-01
-1.67594522e-01 -1.39881599e+00 -1.08331120e+00 -2.42757007e-01
2.67370313e-01 2.34311566e-01 2.40031689e-01 7.96843112e-01
1.95842862e-01 2.63933897e-01 2.67492294e-01 -7.20766932e-03
4.88057673e-01 1.54905036e-01 7.14117646e-01 -1.31081700e+00
7.29915798e-02 -9.00710881e-01 -6.48033798e-01 -4.34200794e-01
3.00664425e-01 -1.09869027e+00 -1.33616388e-01 -1.31643236e+00
6.90270185e-01 -3.98283541e-01 -8.05834413e-01 5.20618439e-01
-3.62798333e-01 8.43896747e-01 -3.60958546e-01 2.37785652e-01
-7.35300243e-01 7.31049255e-02 1.00273061e+00 -2.34290704e-01
6.51512742e-02 -2.69021630e-01 -8.92955959e-01 3.55532408e-01
8.27459037e-01 -3.46681833e-01 -9.34939738e-03 -7.92101771e-02
1.37942955e-01 -1.02194376e-01 2.85353154e-01 -1.09650195e+00
3.90822470e-01 1.28446937e-01 1.13484502e+00 -7.08149314e-01
1.99575171e-01 -5.07496655e-01 1.12847708e-01 7.94511080e-01
-3.80141050e-01 -6.53875589e-01 4.98287916e-01 4.80495304e-01
-2.53389716e-01 1.91259950e-01 9.05492544e-01 1.05033420e-01
-2.56124943e-01 3.18602830e-01 -7.95571804e-01 -3.89409900e-01
1.26011252e+00 -5.68503082e-01 -3.78954142e-01 1.53583989e-01
-4.79478568e-01 4.13677186e-01 5.94258606e-01 -2.00348049e-01
3.26114804e-01 -1.29511344e+00 -8.44265521e-01 3.84342015e-01
4.39393193e-01 6.97844699e-02 4.83518720e-01 1.20750606e+00
-8.88332486e-01 4.96869355e-01 -2.74934053e-01 -1.01880372e+00
-1.28161371e+00 1.51989982e-01 6.35450304e-01 -6.73302591e-01
-1.18250318e-01 1.11639094e+00 3.39545071e-01 -4.83590096e-01
1.83115453e-01 -5.21028303e-02 -3.99019241e-01 1.86319292e-01
5.37561953e-01 3.07656407e-01 2.83219576e-01 -5.20352900e-01
-5.95515549e-01 3.62909764e-01 -5.94122946e-01 3.28944743e-01
1.18712533e+00 4.61872928e-02 -3.19436580e-01 4.14883226e-01
1.47142565e+00 -1.50193959e-01 -1.30240202e+00 5.37681617e-02
-5.60701303e-02 -1.09560400e-01 2.92977303e-01 -9.74201143e-01
-1.12432265e+00 6.45202518e-01 7.36283541e-01 7.07015470e-02
1.09786141e+00 -1.42221794e-01 7.46826589e-01 1.15703985e-01
1.55274376e-01 -9.31263328e-01 -2.19405860e-01 2.42881015e-01
3.06688070e-01 -1.39363539e+00 7.13970438e-02 -4.22205985e-01
-1.85505748e-01 1.16143227e+00 6.90402329e-01 1.38456449e-01
4.30343181e-01 5.08557141e-01 3.15065533e-01 -2.53066301e-01
-9.16391909e-01 9.20067579e-02 -5.53340390e-02 6.03030443e-01
8.87782991e-01 1.14875413e-01 -2.82277554e-01 8.36450160e-01
2.14741796e-01 2.78559685e-01 2.90571958e-01 8.60327661e-01
-3.71041417e-01 -7.51374662e-01 2.20850687e-02 7.41048634e-01
-9.92187560e-01 2.12820306e-01 -3.79155427e-01 7.91470528e-01
-1.11536831e-02 6.14814639e-01 3.76933634e-01 -2.62258977e-01
-1.22134618e-01 -1.72640845e-01 3.96159500e-01 -4.72909927e-01
-4.87150729e-01 2.30844527e-01 -7.22077563e-02 -3.47092360e-01
-2.61875212e-01 -5.31240821e-01 -1.36588049e+00 -2.23029062e-01
-4.31498468e-01 -4.44879048e-02 3.59646708e-01 6.54307604e-01
6.45261347e-01 5.26814282e-01 7.05621541e-01 -7.16591716e-01
-1.54204667e-01 -9.41588521e-01 -9.38312054e-01 2.02624157e-01
5.18834889e-01 -3.75783473e-01 -5.11939406e-01 9.99463052e-02] | [15.058904647827148, -3.0920567512512207] |
9665d908-2363-42eb-8ade-4d8a23c17fc9 | robust-sound-guided-image-manipulation | 2208.14114 | null | https://arxiv.org/abs/2208.14114v3 | https://arxiv.org/pdf/2208.14114v3.pdf | Robust Sound-Guided Image Manipulation | Recent successes suggest that an image can be manipulated by a text prompt, e.g., a landscape scene on a sunny day is manipulated into the same scene on a rainy day driven by a text input "raining". These approaches often utilize a StyleCLIP-based image generator, which leverages multi-modal (text and image) embedding space. However, we observe that such text inputs are often bottlenecked in providing and synthesizing rich semantic cues, e.g., differentiating heavy rain from rain with thunderstorms. To address this issue, we advocate leveraging an additional modality, sound, which has notable advantages in image manipulation as it can convey more diverse semantic cues (vivid emotions or dynamic expressions of the natural world) than texts. In this paper, we propose a novel approach that first extends the image-text joint embedding space with sound and applies a direct latent optimization method to manipulate a given image based on audio input, e.g., the sound of rain. Our extensive experiments show that our sound-guided image manipulation approach produces semantically and visually more plausible manipulation results than the state-of-the-art text and sound-guided image manipulation methods, which are further confirmed by our human evaluations. Our downstream task evaluations also show that our learned image-text-sound joint embedding space effectively encodes sound inputs. | ['Sang Ho Yoon', 'Sangpil Kim', 'Jinkyu Kim', 'Gyeongrok Oh', 'Wonmin Byeon', 'Seung Hyun Lee'] | 2022-08-30 | null | null | null | null | ['image-manipulation'] | ['computer-vision'] | [ 5.03611803e-01 -1.66828245e-01 2.09888890e-01 -3.35343033e-01
-5.83308458e-01 -7.90446162e-01 9.15161669e-01 -3.16714942e-01
-1.53369695e-01 3.61828685e-01 4.02465880e-01 -2.36635968e-01
3.87561023e-01 -8.72743785e-01 -9.09308791e-01 -7.12949991e-01
3.29143316e-01 -1.98258385e-01 -2.15201408e-01 -3.37104797e-01
2.94441760e-01 1.08112246e-01 -1.88098323e+00 2.14244127e-01
5.38061798e-01 5.42381644e-01 5.67717612e-01 1.03955853e+00
-4.68352109e-01 7.97585189e-01 -8.41261566e-01 -8.71403739e-02
9.30142850e-02 -5.75581133e-01 -4.59947169e-01 2.38651291e-01
8.24059844e-01 -6.49344385e-01 -2.84709841e-01 9.73185956e-01
5.03758132e-01 3.16630185e-01 4.25982207e-01 -1.51733422e+00
-1.05974376e+00 5.44836938e-01 -4.94426787e-01 -5.76872975e-02
6.62999332e-01 6.60993218e-01 1.05880094e+00 -9.36446786e-01
5.87803483e-01 1.67913115e+00 2.42431089e-02 5.28261781e-01
-1.44313443e+00 -7.68289626e-01 3.50817800e-01 -1.98354319e-01
-1.04104161e+00 -5.09674788e-01 8.03998947e-01 -3.73542637e-01
6.74555063e-01 6.80356145e-01 6.83974504e-01 1.51723015e+00
5.02976663e-02 7.59379327e-01 1.11612725e+00 -4.70737815e-01
1.76778972e-01 1.96596414e-01 -4.05969620e-01 6.89071894e-01
-4.18738760e-02 2.22310036e-01 -8.97345781e-01 2.05275300e-03
9.57372367e-01 -2.50175446e-02 -5.88715196e-01 -1.27864823e-01
-1.48480737e+00 8.02085340e-01 4.29200143e-01 1.63700972e-02
-2.49156356e-01 6.26544416e-01 1.62126869e-01 2.54346281e-01
5.08706629e-01 5.81599236e-01 -1.68578494e-02 -2.50898987e-01
-9.51369286e-01 4.43497926e-01 6.53950632e-01 8.42944562e-01
7.28193343e-01 3.16712081e-01 -4.11734909e-01 7.13043392e-01
4.91687775e-01 1.09332085e+00 -2.09294502e-02 -1.08673739e+00
3.57150108e-01 2.51625348e-02 4.36552763e-01 -9.89664912e-01
-8.04097727e-02 4.15994003e-02 -5.56727469e-01 5.96351266e-01
1.09023400e-01 -5.04294448e-02 -1.08310401e+00 1.91355085e+00
2.85369605e-01 2.18502253e-01 5.43378443e-02 1.12219810e+00
1.04380631e+00 1.00733447e+00 2.77379394e-01 9.81460363e-02
1.52818346e+00 -1.10600591e+00 -1.00751793e+00 -2.75007665e-01
1.15578562e-01 -8.21690440e-01 1.98349357e+00 2.68880904e-01
-9.72040832e-01 -6.00294292e-01 -8.21859062e-01 -7.51376897e-02
-2.89355874e-01 -9.26018581e-02 5.42233706e-01 4.29094553e-01
-1.17324138e+00 9.42499563e-02 -6.31637096e-01 -4.74595398e-01
8.81073847e-02 -3.30097198e-01 -3.66064072e-01 1.65051594e-02
-1.13354576e+00 5.66925049e-01 -1.22681148e-02 3.23342159e-02
-1.10666716e+00 -6.89099431e-01 -1.24815571e+00 -4.75416780e-02
3.06300789e-01 -1.03567326e+00 1.23969162e+00 -1.26017261e+00
-1.90735877e+00 8.08995008e-01 -2.33688161e-01 6.36359304e-02
3.53932023e-01 -4.42335635e-01 -1.56689480e-01 5.73502302e-01
1.89788014e-01 1.14043677e+00 1.60825861e+00 -1.82615197e+00
-2.81257123e-01 -3.44738970e-03 6.11299396e-01 5.11553943e-01
-3.44265670e-01 1.74531087e-01 -4.46503460e-01 -9.57530320e-01
-2.20695794e-01 -8.99077356e-01 -1.65231511e-01 6.28708243e-01
-6.10839367e-01 3.91894519e-01 9.61939156e-01 -3.75917614e-01
1.12973380e+00 -2.36924958e+00 1.95503533e-01 -3.53189781e-02
1.72890127e-01 -1.42819151e-01 -4.02083576e-01 5.20634711e-01
-5.40655032e-02 4.53012466e-01 -3.14691544e-01 -6.55393302e-01
1.87621549e-01 3.20486963e-01 -7.54355371e-01 1.31681800e-01
2.91807592e-01 9.99724805e-01 -1.07112801e+00 -4.36324388e-01
3.65677655e-01 6.89879179e-01 -5.48459828e-01 3.95947665e-01
-5.06319463e-01 6.87430501e-01 -3.20160121e-01 6.28947437e-01
5.41933477e-01 -1.04785919e-01 -2.37041399e-01 -2.95816749e-01
-1.18025437e-01 1.17066465e-02 -7.98691690e-01 1.81386673e+00
-9.29665148e-01 8.24586928e-01 1.90881774e-01 -3.56168389e-01
7.51841903e-01 3.18802983e-01 -1.40416750e-03 -5.40322185e-01
-1.35110512e-01 -2.47385740e-01 -5.83938301e-01 -9.34164882e-01
6.70261860e-01 -2.90503263e-01 -4.10424545e-02 7.05452263e-01
-3.63423318e-01 -8.53904724e-01 -2.47389928e-01 4.60293204e-01
9.01419759e-01 5.00801861e-01 -8.74586254e-02 7.02149719e-02
-9.64895412e-02 -1.96555555e-01 -3.43469493e-02 9.55332637e-01
-5.74734807e-02 9.07870054e-01 3.68950725e-01 -6.11076206e-02
-7.99056292e-01 -1.31815958e+00 4.03657369e-02 1.35941589e+00
3.29177588e-01 -4.28426951e-01 -5.74705243e-01 -4.34620202e-01
-4.79296893e-02 9.68205094e-01 -7.29234815e-01 -5.78680336e-02
-2.81269372e-01 -3.32756281e-01 6.28791451e-01 4.16695833e-01
4.09289122e-01 -1.38670778e+00 -1.03995359e+00 -7.42271021e-02
-4.80736405e-01 -1.15869021e+00 -5.52950561e-01 -1.50539994e-01
-4.92399693e-01 -5.31273007e-01 -7.61568606e-01 -5.24058163e-01
7.25279033e-01 6.22482479e-01 1.30473912e+00 6.31828308e-02
-3.93865079e-01 8.16109002e-01 -6.44812763e-01 -3.94496560e-01
-3.01703095e-01 -6.07692242e-01 -2.07694352e-01 1.71457008e-02
-2.10493833e-01 -6.17719293e-01 -7.11040020e-01 -2.64445809e-03
-1.60656440e+00 5.41526198e-01 3.27371269e-01 7.46124268e-01
3.00371706e-01 -1.96005985e-01 2.61285365e-01 -5.92713177e-01
6.54435277e-01 -4.18871433e-01 -2.16234252e-01 6.77303448e-02
4.64361832e-02 9.10436064e-02 4.63292539e-01 -6.59118414e-01
-1.20583963e+00 -3.31792310e-02 4.15430181e-02 -5.71084976e-01
-3.49328756e-01 4.53959554e-01 -1.67018041e-01 9.18154195e-02
7.05079556e-01 2.01759920e-01 -2.04116981e-02 -7.21134767e-02
9.68476534e-01 6.48402333e-01 6.68127239e-01 -6.82518661e-01
1.14607680e+00 9.13000882e-01 -2.37952828e-01 -9.40792918e-01
-6.92733049e-01 -1.89281642e-01 -8.71102959e-02 -4.10155177e-01
1.12082946e+00 -9.69340861e-01 -6.08061552e-01 5.61772108e-01
-1.20712054e+00 -8.19863617e-01 -8.85062665e-02 2.20384389e-01
-6.14822447e-01 2.44350940e-01 -5.52834749e-01 -7.63559401e-01
-2.15566859e-01 -1.23919368e+00 1.70928669e+00 9.57874805e-02
-3.66699606e-01 -7.99904585e-01 -7.69360662e-02 2.18320385e-01
6.44535363e-01 4.86083210e-01 9.18818057e-01 2.59311378e-01
-6.89878583e-01 2.66886920e-01 -2.68015385e-01 8.88924897e-02
4.58347589e-01 3.62412930e-01 -1.12614620e+00 -1.39862701e-01
-9.87174660e-02 -5.65837681e-01 8.42392623e-01 1.82970017e-01
1.09763956e+00 -3.32356125e-01 2.40364689e-02 6.93350792e-01
1.43273270e+00 -9.15064588e-02 6.77088022e-01 3.42818826e-01
9.00282562e-01 5.25329709e-01 6.71434045e-01 5.80125809e-01
4.19213086e-01 7.38994837e-01 5.32591224e-01 -4.61178333e-01
-1.79579407e-01 -3.48726064e-01 6.93504572e-01 2.93938667e-01
3.23579699e-01 -4.85747188e-01 -6.61015809e-01 3.17183077e-01
-1.60990906e+00 -9.71659660e-01 1.48041591e-01 1.90877116e+00
1.02357280e+00 -7.61191025e-02 -2.56996304e-01 -8.74800831e-02
3.82274479e-01 6.39169157e-01 -4.00318474e-01 -3.57599705e-01
-1.54278621e-01 3.16199362e-01 1.11624584e-01 7.14708924e-01
-1.01997805e+00 1.20061219e+00 6.21142864e+00 4.48939949e-01
-1.39537823e+00 -1.04319327e-01 2.84516484e-01 -2.83507764e-01
-9.56738651e-01 -7.63819292e-02 -2.95647234e-01 3.68818611e-01
5.49611568e-01 3.90383750e-02 5.76635182e-01 4.66232687e-01
8.21771502e-01 -4.50533688e-01 -1.03249168e+00 1.13213527e+00
2.72229731e-01 -1.04499817e+00 5.38149059e-01 -3.34578723e-01
3.97413075e-01 -4.74937201e-01 4.76876318e-01 2.61857361e-01
2.81117827e-01 -1.07856607e+00 1.07518780e+00 5.96640527e-01
1.06483018e+00 -4.32200044e-01 -1.18094489e-01 -6.56129047e-03
-1.13449502e+00 1.98050141e-01 7.88642094e-02 -9.00950655e-02
4.37683553e-01 4.65147942e-01 -4.74997401e-01 2.33587444e-01
8.60114336e-01 4.47487503e-01 -4.27054167e-01 5.07762790e-01
-6.16284966e-01 7.20758200e-01 -2.20689699e-01 1.34076327e-01
1.69257551e-01 -1.81495339e-01 8.73580575e-01 1.35641527e+00
2.29807854e-01 7.76233822e-02 1.38393536e-01 1.22165477e+00
5.71898222e-02 -5.23516685e-02 -1.10861313e+00 -1.12535536e-01
3.86402130e-01 1.28767228e+00 -5.84351003e-01 -4.02707428e-01
-2.46508002e-01 1.31870878e+00 -1.19588099e-01 8.25917542e-01
-1.16396499e+00 -4.96844023e-01 7.87438333e-01 -2.29425013e-01
3.77655514e-02 -3.79897535e-01 -2.62320995e-01 -1.16374493e+00
2.38169096e-02 -1.01829469e+00 -2.89577872e-01 -1.57232642e+00
-1.00287855e+00 5.11269152e-01 1.30195722e-01 -1.13669121e+00
-1.13333240e-01 -2.62065798e-01 -8.75125110e-01 7.74885237e-01
-1.56624985e+00 -1.35311544e+00 -7.34703064e-01 5.57417512e-01
7.47134089e-01 3.92358929e-01 8.38676631e-01 -2.08654366e-02
-3.60127211e-01 5.00899732e-01 -3.44678462e-01 -1.03945166e-01
9.39898431e-01 -1.44035637e+00 4.51395512e-01 7.71730363e-01
3.50153416e-01 5.96645772e-01 1.01469862e+00 -4.31719691e-01
-1.71864426e+00 -1.21118236e+00 4.25178289e-01 -4.72605467e-01
6.00407302e-01 -7.07397461e-01 -8.70387137e-01 5.84207058e-01
3.86193216e-01 -1.42139941e-01 4.66598451e-01 -1.79040968e-01
-5.80199361e-01 1.85192019e-01 -9.22427833e-01 1.24802196e+00
8.83865237e-01 -8.93191576e-01 -4.49691772e-01 2.88514584e-01
1.27166700e+00 -3.79771262e-01 -3.01143318e-01 1.39280543e-01
5.04676223e-01 -8.70584488e-01 1.06102812e+00 -4.39512253e-01
8.67251515e-01 -4.86520827e-01 -2.92058021e-01 -1.56695998e+00
-3.60678323e-02 -7.47067571e-01 1.40181363e-01 1.27880406e+00
1.62618801e-01 -3.69595557e-01 2.88088590e-01 5.37587762e-01
4.62881289e-02 -3.98769945e-01 -3.62160534e-01 -5.08080959e-01
-1.99748859e-01 -6.27410710e-01 4.37277436e-01 9.14276659e-01
-8.48875940e-02 2.45206654e-01 -5.28385341e-01 3.53243172e-01
3.52571636e-01 1.86171889e-01 1.10692203e+00 -5.07619381e-01
-4.54077810e-01 -3.34658533e-01 -1.43018916e-01 -1.10840535e+00
3.00004035e-01 -6.72552288e-01 5.90628088e-01 -1.51133955e+00
1.77943408e-01 -2.14050621e-01 -6.24090852e-03 6.07081890e-01
-5.84772229e-01 3.34924668e-01 5.21041691e-01 1.12080656e-01
-4.48107034e-01 8.27412307e-01 1.49020064e+00 -1.33370668e-01
-2.96375573e-01 -6.11307859e-01 -7.20419109e-01 5.87832808e-01
6.97366238e-01 -1.90461367e-01 -5.59588790e-01 -7.25031614e-01
3.99743199e-01 2.13344306e-01 8.47449064e-01 -6.70043766e-01
-1.70534030e-01 -4.32402283e-01 3.35322618e-02 -7.57894740e-02
6.69932663e-01 -5.60793221e-01 -7.61175994e-03 1.47792265e-01
-5.83882332e-01 -5.07874340e-02 4.48226094e-01 7.14630485e-01
-2.58387178e-01 1.72999516e-01 5.95988691e-01 -6.00684248e-02
-4.71941113e-01 5.71862422e-02 -6.63967669e-01 -9.77427214e-02
6.19514167e-01 -1.12847708e-01 -5.18489003e-01 -9.73625779e-01
-5.30473948e-01 1.22525968e-01 5.13208687e-01 8.56629014e-01
8.45167279e-01 -1.24646211e+00 -8.04980695e-01 1.30344301e-01
3.28041881e-01 -1.03878148e-01 2.40844071e-01 6.49506450e-01
-4.16423589e-01 -1.75774083e-01 -1.05534494e-01 -7.95674026e-01
-1.33118582e+00 2.71440089e-01 5.05590551e-02 3.10320735e-01
-7.83767700e-01 8.48736942e-01 8.78240407e-01 -2.82738775e-01
2.34487131e-01 -5.34507692e-01 1.12377234e-01 -1.42138809e-01
5.81034184e-01 -1.89528525e-01 -4.90859210e-01 -3.33769739e-01
5.15529774e-02 4.78965282e-01 2.93733776e-01 -7.81978428e-01
9.85838532e-01 -2.96649665e-01 -9.62844566e-02 6.99556530e-01
1.08491397e+00 1.51056319e-01 -1.42312479e+00 -1.12648390e-01
-5.81706405e-01 -8.17587018e-01 1.81957617e-01 -8.62753332e-01
-9.55098808e-01 9.42608654e-01 5.55300057e-01 1.11527517e-01
1.27570462e+00 -2.35930085e-02 7.77500451e-01 4.33342427e-01
2.23864421e-01 -6.63382947e-01 6.22427583e-01 4.06208038e-01
1.39545989e+00 -1.19870329e+00 -4.25179213e-01 -3.01402479e-01
-9.23925400e-01 1.06167090e+00 5.63146830e-01 2.08242819e-01
4.27095532e-01 4.79454935e-01 5.73905528e-01 -3.49672347e-01
-7.71126330e-01 -2.78373003e-01 2.23170640e-03 5.04005671e-01
3.29656482e-01 1.70796826e-01 2.29087293e-01 1.56545237e-01
-3.47299457e-01 -1.40271068e-01 6.31952822e-01 1.12721086e+00
-4.84348118e-01 -7.77541280e-01 -7.31633127e-01 1.17710516e-01
-1.08660877e-01 -5.17457247e-01 -2.77119875e-01 4.83442277e-01
-4.29082178e-02 1.09392536e+00 4.33620140e-02 -2.34132737e-01
2.06731483e-01 -1.14492394e-01 5.58976769e-01 -7.57173777e-01
-4.22822028e-01 1.05374284e-01 -8.24583173e-02 -7.13071764e-01
-5.44046104e-01 -4.18662488e-01 -1.22577810e+00 6.12784028e-02
-1.05034828e-01 -2.39909291e-01 8.36134851e-01 7.29106963e-01
1.93979859e-01 5.82980394e-01 7.47176766e-01 -1.36038196e+00
-2.25313783e-01 -7.42583990e-01 -4.47600007e-01 7.68085539e-01
4.98881966e-01 -6.02295637e-01 -6.43975735e-01 5.45143545e-01] | [11.556941032409668, -0.20375262200832367] |
37f6e94f-36b0-4ec2-93cb-9280bdfff8be | recurrent-ladder-networks | 1707.09219 | null | http://arxiv.org/abs/1707.09219v4 | http://arxiv.org/pdf/1707.09219v4.pdf | Recurrent Ladder Networks | We propose a recurrent extension of the Ladder networks whose structure is
motivated by the inference required in hierarchical latent variable models. We
demonstrate that the recurrent Ladder is able to handle a wide variety of
complex learning tasks that benefit from iterative inference and temporal
modeling. The architecture shows close-to-optimal results on temporal modeling
of video data, competitive results on music modeling, and improved perceptual
grouping based on higher order abstractions, such as stochastic textures and
motion cues. We present results for fully supervised, semi-supervised, and
unsupervised tasks. The results suggest that the proposed architecture and
principles are powerful tools for learning a hierarchy of abstractions,
learning iterative inference and handling temporal information. | ['Tele Hotloo Hao', 'Isabeau Prémont-Schwarz', 'Alexander Ilin', 'Rinu Boney', 'Harri Valpola', 'Antti Rasmus'] | 2017-07-28 | recurrent-ladder-networks-1 | http://papers.nips.cc/paper/7182-recurrent-ladder-networks | http://papers.nips.cc/paper/7182-recurrent-ladder-networks.pdf | neurips-2017-12 | ['music-modeling'] | ['music'] | [ 1.63866282e-01 -1.29778206e-01 -5.65507948e-01 -2.58019000e-01
-4.43891257e-01 -3.68552715e-01 7.20941901e-01 -2.58456856e-01
4.08184305e-02 5.05794346e-01 3.41942698e-01 -9.66870710e-02
-5.73324084e-01 -3.79064530e-01 -4.16983783e-01 -8.54668081e-01
-8.70752633e-01 4.91465747e-01 3.01915050e-01 -3.90938632e-02
2.77184874e-01 6.18834853e-01 -1.77837062e+00 7.63651013e-01
1.99867576e-01 8.18527758e-01 1.61025882e-01 9.30326104e-01
-6.80378377e-02 1.23398352e+00 -1.41303897e-01 -4.60931808e-02
-1.09846786e-01 -4.45636094e-01 -7.05606282e-01 1.40084982e-01
4.57947582e-01 -1.42751172e-01 -2.88376898e-01 4.46537733e-01
2.11490810e-01 3.92026395e-01 8.85652542e-01 -1.15719628e+00
-4.22801524e-01 9.40439641e-01 -4.10554856e-01 1.92122191e-01
2.32169956e-01 -2.33615756e-01 1.32071614e+00 -7.82224059e-01
7.08050549e-01 1.62867630e+00 1.01415551e+00 5.73697448e-01
-1.72064245e+00 -4.72240657e-01 5.40448427e-01 5.04852355e-01
-9.47687387e-01 -4.23835814e-01 8.47578764e-01 -6.97613955e-01
8.31181228e-01 2.29152486e-01 6.16557956e-01 1.57588351e+00
2.69618809e-01 1.09275985e+00 1.01273322e+00 -5.33398449e-01
2.36837119e-01 -2.15755507e-01 3.35279882e-01 9.13176477e-01
-4.67333227e-01 3.92558515e-01 -1.26725423e+00 -3.08839381e-01
1.08445179e+00 -1.72043741e-02 2.06022829e-01 -5.08910120e-01
-1.18811536e+00 7.33513951e-01 9.89160016e-02 2.58297354e-01
-2.88346857e-01 5.97172022e-01 4.77254033e-01 3.97047490e-01
4.42759544e-01 1.09210320e-01 -4.54532832e-01 -2.13229924e-01
-1.37564409e+00 7.05632195e-02 6.66472137e-01 9.69131947e-01
5.69782317e-01 4.54372525e-01 -1.53077513e-01 7.76051104e-01
3.82990509e-01 1.79644674e-01 5.01688123e-01 -1.64728725e+00
-1.77623518e-02 -1.25029549e-01 5.64478412e-02 -7.75404990e-01
-5.48010826e-01 -6.19131505e-01 -1.04831791e+00 4.31412548e-01
2.54979223e-01 2.79868662e-01 -9.71256554e-01 1.93264294e+00
-2.48337358e-01 4.27170306e-01 -2.25229710e-01 3.98399174e-01
5.13290048e-01 6.16350472e-01 7.57927969e-02 -7.81576455e-01
1.08386111e+00 -1.10048831e+00 -1.00616729e+00 3.04931015e-01
1.88552126e-01 -4.03680444e-01 8.54032576e-01 8.68906081e-01
-1.29721367e+00 -1.04044819e+00 -9.44491088e-01 -8.50305110e-02
8.53537116e-03 8.09541643e-02 9.98714387e-01 2.89295405e-01
-1.16248393e+00 1.10333860e+00 -1.37330222e+00 -3.42021942e-01
-6.47594333e-02 3.19382519e-01 1.67689502e-01 4.32337910e-01
-9.22789812e-01 2.92178541e-01 2.16990247e-01 1.60604209e-01
-1.31516695e+00 -4.95873421e-01 -7.30261445e-01 2.67093163e-03
-3.02711744e-02 -8.98065209e-01 1.41735065e+00 -7.86152363e-01
-1.71146131e+00 4.71440017e-01 -5.31575561e-01 -7.54253387e-01
6.67734563e-01 -4.68481988e-01 -3.18524361e-01 3.89520615e-01
-9.93088037e-02 7.40738153e-01 1.26187229e+00 -1.19597447e+00
-4.89608467e-01 -1.33400053e-01 -2.02246949e-01 -1.85399596e-02
-3.90282661e-01 -1.51039390e-02 -4.01608437e-01 -1.04640901e+00
3.29671860e-01 -1.11881232e+00 -4.69926029e-01 -1.10055104e-01
7.36431044e-04 -2.43739992e-01 9.64996338e-01 -5.46204209e-01
1.44250071e+00 -2.34743595e+00 7.01679826e-01 8.47726390e-02
2.13059157e-01 -5.12756586e-01 -1.11229211e-01 6.17161393e-01
-2.74676144e-01 -5.73880412e-03 -4.47870838e-03 -7.10285187e-01
1.33135736e-01 4.57874954e-01 -7.97404945e-01 2.68755127e-02
-3.17480654e-01 6.61079586e-01 -8.55103135e-01 -5.37942886e-01
2.46516749e-01 3.01700115e-01 -6.27582967e-01 1.60901487e-01
-4.99907762e-01 4.68550056e-01 -9.18347612e-02 4.68987375e-01
-7.09402338e-02 -4.61072385e-01 2.19313204e-01 -1.84294969e-01
-1.10723674e-01 2.76256710e-01 -1.36315989e+00 1.83554506e+00
-3.49745542e-01 9.86238956e-01 -2.43381068e-01 -6.14343405e-01
6.68497145e-01 8.07517350e-01 5.88050902e-01 -9.49022919e-02
-4.35064316e-01 -9.51258466e-02 -3.55172336e-01 -4.57100272e-01
3.10871273e-01 -2.07059726e-01 2.04583686e-02 5.63640177e-01
3.30052733e-01 -7.00232163e-02 4.35957372e-01 3.15161467e-01
7.79567838e-01 5.43834686e-01 4.94815670e-02 -2.19593570e-01
2.66195655e-01 -3.55538607e-01 6.73289120e-01 1.12466681e+00
1.72820389e-01 3.59603077e-01 4.54290241e-01 -7.84486711e-01
-9.71138656e-01 -1.47552860e+00 -2.27743927e-02 1.60840809e+00
-2.21474752e-01 -7.93580711e-01 -2.75580972e-01 -1.22589096e-01
-2.56257951e-01 5.13924599e-01 -7.25573301e-01 1.23126782e-01
-4.77679580e-01 -4.69637781e-01 4.40459013e-01 9.34404373e-01
1.67079732e-01 -1.15738332e+00 -4.68254596e-01 4.12767440e-01
-1.96494624e-01 -1.03913689e+00 4.91541177e-02 5.45711160e-01
-1.42617571e+00 -5.42391002e-01 -4.29967880e-01 -6.86201036e-01
1.57727957e-01 7.04383031e-02 1.13905394e+00 -1.92807540e-01
-3.07865113e-01 5.15339673e-01 -1.60590261e-01 2.23164633e-02
-3.56031269e-01 -3.20274830e-02 4.03126359e-01 -9.32303444e-02
-5.78638799e-02 -1.19090950e+00 -1.93540022e-01 3.35527360e-01
-8.00911725e-01 3.44350070e-01 3.39290053e-01 9.15893734e-01
4.52286661e-01 2.48989418e-01 1.51285946e-01 -6.57820702e-01
4.00387198e-01 -8.13762769e-02 -4.36029643e-01 1.86599165e-01
-4.04578298e-01 4.09491986e-01 4.90526199e-01 -8.20788383e-01
-1.18881416e+00 1.01028562e-01 1.87964022e-01 -6.36714637e-01
-2.25210696e-01 5.77507794e-01 4.08715427e-01 1.64554849e-01
6.10708535e-01 1.77654445e-01 -3.56008530e-01 -7.04041898e-01
5.08507311e-01 5.09383082e-02 6.40248954e-01 -9.85130608e-01
7.15315402e-01 8.50466013e-01 2.09916115e-01 -1.06165612e+00
-1.02939069e+00 -4.13056105e-01 -1.09519327e+00 -2.53099710e-01
9.48286116e-01 -9.06983674e-01 -6.81776702e-01 3.85523856e-01
-1.04065371e+00 -5.88274062e-01 -3.69378835e-01 6.52060270e-01
-1.16415250e+00 4.00815606e-01 -1.28605771e+00 -9.73408401e-01
1.76568225e-01 -7.98347354e-01 9.96329844e-01 -1.48412004e-01
-6.05034053e-01 -1.28674543e+00 2.67633736e-01 1.05297752e-01
-4.40451596e-03 8.96426141e-02 1.19663620e+00 -1.71996914e-02
-9.96785760e-01 3.15753728e-01 2.81606495e-01 1.47389501e-01
-4.26525734e-02 4.33955193e-01 -9.31484222e-01 -2.33369261e-01
-4.09297720e-02 -4.76332039e-01 1.28033328e+00 5.77455103e-01
1.15309274e+00 -3.10937334e-02 -1.53950140e-01 6.79122746e-01
1.11006176e+00 4.54036444e-02 2.85078406e-01 1.21668512e-02
6.10182464e-01 5.14508247e-01 1.37705460e-01 5.99029422e-01
-6.91973558e-03 8.21864188e-01 1.19949095e-01 1.36473477e-01
-4.55946401e-02 -2.93894887e-01 6.14415765e-01 1.17802250e+00
-4.55619991e-01 1.83748826e-01 -8.57318282e-01 5.58750629e-01
-2.27823043e+00 -1.52845430e+00 -9.61902887e-02 1.62368679e+00
8.07825863e-01 4.11930174e-01 3.50758702e-01 2.86601186e-01
4.00086939e-01 2.80367136e-01 -3.92262161e-01 -4.50917006e-01
-1.45802319e-01 7.33798817e-02 -3.50041091e-02 7.01837540e-01
-1.06149161e+00 9.69707072e-01 8.73353291e+00 7.44293630e-01
-7.28957295e-01 4.54685017e-02 2.72472173e-01 -1.32230744e-01
-2.86575228e-01 3.19885239e-02 -8.97936821e-01 -4.91392426e-03
1.08909476e+00 6.58271760e-02 5.12406170e-01 5.30383587e-01
5.94702601e-01 4.08148527e-01 -1.39117730e+00 9.14454639e-01
-1.89489976e-01 -1.54484177e+00 5.43139219e-01 -1.13932259e-01
1.00440073e+00 -1.68522030e-01 3.12473238e-01 2.56599814e-01
6.43962324e-01 -9.69620764e-01 7.97042191e-01 9.01388168e-01
3.44718844e-01 -5.40860415e-01 1.84710905e-01 4.19084013e-01
-1.59582067e+00 -4.05223846e-01 3.94410603e-02 -4.38605636e-01
2.82861292e-01 2.68656760e-01 -1.48420215e-01 2.76899189e-01
7.85844266e-01 1.16147149e+00 -4.39770907e-01 8.96550119e-01
-3.89658600e-01 6.53882444e-01 -2.15768278e-01 2.53402501e-01
3.57934535e-01 -1.94089990e-02 6.11721098e-01 1.55114508e+00
-1.48860095e-02 -1.15356259e-01 3.07095647e-01 5.99040568e-01
3.92679244e-01 -2.90773064e-01 -5.13083875e-01 1.32569028e-02
1.57902807e-01 8.42090666e-01 -8.55858922e-01 -4.94834602e-01
-1.56229466e-01 7.94008553e-01 2.77969182e-01 8.14803958e-01
-8.14733148e-01 1.98251486e-01 4.41927522e-01 -2.50089675e-01
5.08242249e-01 -8.68264019e-01 -2.59273469e-01 -1.28686070e+00
-2.89618671e-01 -7.80892849e-01 6.92265809e-01 -1.01956689e+00
-1.09492946e+00 3.29870343e-01 3.97349298e-01 -1.09285164e+00
-9.48005855e-01 -7.34246492e-01 -4.37315494e-01 4.04770136e-01
-1.01373446e+00 -1.06411302e+00 -6.92944229e-02 7.69902468e-01
8.81831467e-01 -2.83329248e-01 9.64716315e-01 4.02386263e-02
-1.90972090e-01 2.58663267e-01 2.21164241e-01 -1.42668590e-01
4.62053508e-01 -1.43538690e+00 1.43168747e-01 6.23771846e-01
8.25676620e-01 8.33905876e-01 9.41935778e-01 -3.85293782e-01
-1.00396371e+00 -5.56907713e-01 7.04582036e-01 -4.02799159e-01
9.23354447e-01 -4.57828373e-01 -7.16694474e-01 1.01185369e+00
1.07066430e-01 -4.81211036e-01 8.16752255e-01 8.02784145e-01
-6.69722080e-01 -1.55675318e-02 -4.24720377e-01 8.35987389e-01
1.23342812e+00 -1.01017213e+00 -6.99842155e-01 2.74059236e-01
8.08413684e-01 -4.06934582e-02 -7.28579462e-01 5.14934003e-01
9.63176191e-01 -1.06444132e+00 1.09841859e+00 -9.06658888e-01
5.84147155e-01 -1.08917058e-01 -2.27540255e-01 -7.78001487e-01
-8.80177677e-01 -1.10190296e+00 -8.15786302e-01 8.41962516e-01
3.02941412e-01 2.19586208e-01 9.08876777e-01 1.49098098e-01
1.11027256e-01 -3.64112020e-01 -7.99305141e-01 -8.70353103e-01
-3.74899030e-01 -7.37948418e-01 -2.66728222e-01 9.90236461e-01
-9.85081047e-02 4.65598971e-01 -7.64577925e-01 1.93802625e-01
1.06328487e+00 4.06699896e-01 5.43497920e-01 -1.39376724e+00
-7.38219023e-01 -5.75630188e-01 -3.18233520e-01 -1.41551661e+00
2.87717134e-01 -5.09429932e-01 -2.38973677e-01 -1.25167322e+00
2.35308558e-01 -4.62413728e-02 -6.14045322e-01 2.11032450e-01
3.86188924e-01 3.18826616e-01 1.82152182e-01 4.87631768e-01
-7.33312249e-01 4.88039851e-01 8.78275275e-01 -5.78847900e-02
-3.87407452e-01 1.73991814e-01 -4.32982855e-02 1.22131765e+00
4.88669306e-01 -3.83151948e-01 -5.67848623e-01 -3.82121384e-01
4.26719368e-01 5.57072461e-01 5.78001261e-01 -1.16793382e+00
4.34109569e-01 -2.13398486e-01 4.42132533e-01 -1.02339160e+00
7.63705015e-01 -6.84894800e-01 3.24833602e-01 4.82167989e-01
-8.36257339e-01 1.45431668e-01 1.53696790e-01 7.57434130e-01
-3.45753431e-01 3.02629080e-02 6.93040788e-01 -1.57930151e-01
-8.17982197e-01 1.69137850e-01 -6.97803438e-01 -2.72545964e-01
4.63090181e-01 -2.59174526e-01 2.44338334e-01 -6.95741475e-01
-1.93751371e+00 -3.50309610e-02 -3.27040665e-02 4.76215214e-01
4.87587959e-01 -1.52604735e+00 -4.60185796e-01 1.23938009e-01
-1.49909139e-01 -7.14289546e-01 7.22029284e-02 6.53966248e-01
-1.57624260e-01 4.89645243e-01 -3.59729052e-01 -1.11262131e+00
-1.44398093e+00 6.24297798e-01 1.25020802e-01 -4.97199953e-01
-7.99933612e-01 6.41002715e-01 3.20804060e-01 4.70232479e-02
7.52555549e-01 -4.66915518e-01 -1.16286084e-01 4.08285767e-01
4.44560707e-01 4.17664260e-01 -5.10511458e-01 -1.51506692e-01
1.00755535e-01 5.93293130e-01 -2.35322230e-02 -6.01119101e-01
1.39587879e+00 -3.48532021e-01 -1.70194894e-01 1.53326690e+00
8.03956628e-01 -2.22101614e-01 -1.53418171e+00 -4.99187261e-01
2.98664689e-01 -4.16825572e-03 -1.28911763e-01 -3.66912156e-01
-4.62200046e-01 1.11419678e+00 4.92243826e-01 3.63139778e-01
1.07787609e+00 4.79914583e-02 2.61674523e-01 7.73405850e-01
5.17912209e-01 -1.11946118e+00 5.80890715e-01 8.65717709e-01
7.50276864e-01 -5.67317247e-01 7.21388161e-02 -2.05535367e-01
-1.57065660e-01 1.37013233e+00 1.96692571e-01 -1.08580694e-01
9.38378215e-01 2.67612070e-01 -1.05143569e-01 -1.00069523e-01
-1.40507960e+00 8.96501541e-03 4.14419591e-01 3.42984587e-01
5.52722514e-01 -9.03654173e-02 9.70389172e-02 2.61367738e-01
-1.78276718e-01 -6.78868368e-02 2.34224111e-01 7.68453956e-01
-4.49410349e-01 -1.04624069e+00 -3.04598480e-01 -2.61744354e-02
-3.86150956e-01 -1.54506102e-01 -3.59073617e-02 6.35941565e-01
-1.09117121e-01 5.70367455e-01 -1.80282090e-02 -1.29047617e-01
-2.81825006e-01 3.28902006e-01 7.47945249e-01 -5.12588263e-01
-2.87197024e-01 6.44751608e-01 8.65059718e-02 -6.46222055e-01
-8.30182731e-01 -8.28502595e-01 -1.01427317e+00 -1.09009765e-01
1.52451560e-01 2.33400658e-01 2.63275951e-01 8.89343798e-01
-6.96203858e-02 7.05754578e-01 5.09752452e-01 -1.20649409e+00
-2.62526065e-01 -8.24326575e-01 -6.07620776e-01 3.98694783e-01
6.02418959e-01 -7.15505362e-01 -4.74664718e-01 7.63763368e-01] | [8.579971313476562, 0.9350833296775818] |
100b566a-4a87-4ef5-b405-96a4c4584fcf | predicting-electricity-infrastructure-induced | 2206.02930 | null | https://arxiv.org/abs/2206.02930v2 | https://arxiv.org/pdf/2206.02930v2.pdf | Predicting Electricity Infrastructure Induced Wildfire Risk in California | This paper examines the use of risk models to predict the timing and location of wildfires caused by electricity infrastructure. Our data include historical ignition and wire-down points triggered by grid infrastructure collected between 2015 to 2019 in Pacific Gas & Electricity territory along with various weather, vegetation, and very high resolution data on grid infrastructure including location, age, materials. With these data we explore a range of machine learning methods and strategies to manage training data imbalance. The best area under the receiver operating characteristic we obtain is 0.776 for distribution feeder ignitions and 0.824 for transmission line wire-down events, both using the histogram-based gradient boosting tree algorithm (HGB) with under-sampling. We then use these models to identify which information provides the most predictive value. After line length, we find that weather and vegetation features dominate the list of top important features for ignition or wire-down risk. Distribution ignition models show more dependence on slow-varying vegetation variables such as burn index, energy release content, and tree height, whereas transmission wire-down models rely more on primary weather variables such as wind speed and precipitation. These results point to the importance of improved vegetation modeling for feeder ignition risk models, and improved weather forecasting for transmission wire-down models. We observe that infrastructure features make small but meaningful improvements to risk model predictive power. | ['Duncan S. Callaway', 'Baihong Jin', 'Zheng Zhang', 'Meghana Bharadwaj', 'Mengqi Yao'] | 2022-06-06 | null | null | null | null | ['weather-forecasting'] | ['miscellaneous'] | [-1.08446017e-01 -4.72439587e-01 -4.95539725e-01 6.80610240e-02
-4.14865643e-01 -6.09850526e-01 8.67129028e-01 5.37797272e-01
-9.19600353e-02 9.37052608e-01 4.56217438e-01 -1.10361278e+00
-5.68335950e-01 -1.39182734e+00 -8.01703557e-02 -9.42349553e-01
-6.87031806e-01 2.34483808e-01 -1.93490699e-01 -3.90049577e-01
1.37559056e-01 7.74944007e-01 -1.19313431e+00 -2.67309368e-01
8.92900467e-01 8.07607472e-01 4.65912446e-02 5.86875856e-01
3.14639956e-01 5.94907880e-01 -5.27149439e-01 5.16366482e-01
4.11226988e-01 -1.54645681e-01 -4.07634079e-01 -5.12481451e-01
-3.18762511e-01 -5.31511188e-01 -5.35788760e-02 5.08053958e-01
5.43239951e-01 1.93215013e-02 1.01686931e+00 -1.48640180e+00
-1.25973076e-01 8.69249105e-01 -9.69813704e-01 9.84361649e-01
1.24139935e-01 2.01847196e-01 9.95437086e-01 -4.02222753e-01
1.84387952e-01 7.62714982e-01 1.04219031e+00 -4.45805490e-01
-1.53917086e+00 -9.02688622e-01 5.04247192e-03 3.38980734e-01
-1.56809652e+00 -9.39626619e-02 4.84009683e-01 -7.17784524e-01
1.57931399e+00 5.12249768e-01 1.02567601e+00 4.25511897e-01
3.58306646e-01 4.73590866e-02 1.15492070e+00 -5.47890246e-01
3.25841993e-01 -7.85582885e-02 -3.66306528e-02 2.59640962e-01
3.20690304e-01 6.26608670e-01 -1.81125492e-01 -4.67993349e-01
5.10859489e-01 -2.23960042e-01 -1.78856477e-01 4.84397233e-01
-6.71641707e-01 7.76591420e-01 5.52488446e-01 4.20707196e-01
-5.99997044e-01 -1.76950216e-01 2.04024240e-01 3.09897333e-01
9.79131162e-01 2.93121606e-01 -6.79791749e-01 -7.63483495e-02
-1.44931424e+00 3.32927525e-01 7.49641895e-01 2.80624449e-01
7.38845050e-01 2.71468282e-01 3.78238447e-02 6.75668299e-01
1.29057318e-01 9.01788294e-01 -9.75943729e-02 -5.30231178e-01
3.26782435e-01 -1.72053017e-02 2.93990195e-01 -1.05264270e+00
-9.35489953e-01 -6.73541427e-01 -1.06648195e+00 4.32207704e-01
3.03193510e-01 -3.87556612e-01 -7.30880558e-01 1.34099245e+00
-5.63781932e-02 7.44403079e-02 -3.21128428e-01 4.73648727e-01
-3.30343209e-02 9.93000388e-01 4.32867467e-01 -5.98629892e-01
1.26124954e+00 -4.14634138e-01 -4.77206916e-01 -3.94068770e-02
6.11538768e-01 -3.36317927e-01 4.21957433e-01 2.79574633e-01
-6.43740356e-01 -3.48946974e-02 -8.22441638e-01 8.30938339e-01
-6.44261360e-01 -3.81736696e-01 6.55704021e-01 6.68192267e-01
-8.97148013e-01 8.70095551e-01 -6.72834933e-01 -5.55481195e-01
3.59583527e-01 -1.82633832e-01 2.75627702e-01 2.43655801e-01
-1.25118184e+00 1.39490056e+00 3.69384661e-02 2.21091866e-01
-7.34868944e-01 -1.02028644e+00 -5.07240593e-01 3.73108745e-01
-1.11066557e-01 -2.25056425e-01 6.16251528e-01 -3.37975889e-01
-4.80312407e-01 2.60381550e-01 1.04342438e-01 -6.67857826e-01
2.65742034e-01 4.27536994e-01 -7.68362045e-01 -9.02977660e-02
3.51063043e-01 7.09878700e-03 3.07522774e-01 -1.28288913e+00
-1.07624757e+00 -2.85619110e-01 -3.14562440e-01 2.24749953e-01
-1.71938688e-01 2.53099620e-01 6.15204275e-01 -8.03662598e-01
-1.99713245e-01 -6.15471542e-01 -3.54941040e-01 -5.66224158e-01
-2.97240913e-01 -2.03287572e-01 4.60140318e-01 -1.24744248e+00
1.58176112e+00 -1.70728374e+00 -4.32402521e-01 7.70952642e-01
-2.77284086e-01 -2.58476079e-01 1.84025064e-01 6.24237299e-01
-2.13001356e-01 7.50597000e-01 -3.61193240e-01 4.14137065e-01
-1.01040982e-01 3.07549119e-01 -3.64847481e-01 4.44841325e-01
1.48857981e-01 3.37153584e-01 -7.52329171e-01 -6.05914555e-02
7.14873254e-01 1.72484770e-01 -7.25541711e-02 -1.77727327e-01
-1.10910283e-02 2.39286035e-01 -1.00641489e-01 7.91452169e-01
8.91888857e-01 2.57956624e-01 -9.46273282e-02 1.38321787e-01
-8.54160786e-01 2.80619144e-01 -7.14027166e-01 8.35863173e-01
-5.27015328e-01 4.46091175e-01 1.63008392e-01 -9.91637409e-01
8.81068289e-01 3.40138197e-01 9.19137836e-01 -7.93706536e-01
-2.29861304e-01 -1.12311542e-02 -3.67509536e-02 -4.32373792e-01
4.42835921e-03 -3.70134205e-01 1.33904487e-01 6.83366954e-01
-5.01137853e-01 -2.68515080e-01 1.56908318e-01 -9.85349715e-02
1.05848885e+00 -2.07919717e-01 3.44460636e-01 -8.19835305e-01
-1.32198408e-01 4.23779637e-01 6.71292424e-01 5.35744846e-01
-1.51659101e-01 -6.26422465e-02 4.18639123e-01 -5.26741385e-01
-1.09016323e+00 -1.08400095e+00 -7.68580496e-01 9.87737119e-01
-2.46794224e-01 -2.08489433e-01 -1.44250974e-01 -1.12397142e-01
3.89599830e-01 1.35631549e+00 -4.66035992e-01 -7.20254937e-03
-1.51669979e-01 -1.73566818e+00 5.21331012e-01 5.81696451e-01
3.66507620e-01 -7.43554473e-01 -4.17281419e-01 4.53635961e-01
-2.99975455e-01 -4.52454120e-01 3.70022506e-01 7.92911172e-01
-8.35409939e-01 -8.98999333e-01 -4.64166999e-01 -2.48391613e-01
2.65352070e-01 5.13025746e-02 1.38732088e+00 2.61497676e-01
-5.35812855e-01 7.99860358e-02 -3.65989119e-01 -5.54677784e-01
-3.16104516e-02 1.79919451e-01 1.87262855e-02 -7.75814652e-01
2.54041106e-01 -1.20819533e+00 -6.97904050e-01 3.26877594e-01
-3.69986057e-01 -2.34571975e-02 2.17522174e-01 5.22682607e-01
2.79852897e-01 1.00332320e+00 7.53417313e-01 -1.06034361e-01
2.83953726e-01 -1.05791378e+00 -5.20107508e-01 1.37486339e-01
-1.32147348e+00 -2.90921092e-01 2.50409335e-01 2.44560957e-01
-9.46663022e-01 -3.47577274e-01 -1.18152052e-01 2.88290381e-01
-3.32406580e-01 8.63275051e-01 1.82259440e-01 1.45270362e-01
6.86378717e-01 -2.39341766e-01 -4.70397234e-01 -5.70054233e-01
-8.28661472e-02 5.99649251e-01 3.92672479e-01 -5.72568655e-01
1.25253582e+00 4.65109825e-01 -8.60281475e-03 -1.02626085e+00
-3.67969334e-01 -4.71124083e-01 -4.44453299e-01 -4.48019564e-01
6.80328548e-01 -1.07004535e+00 -4.78093028e-01 5.85025549e-01
-4.85468775e-01 -6.98460102e-01 -1.63902313e-01 4.89415199e-01
-1.80635318e-01 -1.70871496e-01 -3.52917045e-01 -1.35912418e+00
-4.54844445e-01 -5.03803492e-01 4.57660884e-01 6.16885461e-02
-2.02560455e-01 -1.08514333e+00 4.38544780e-01 5.99157102e-02
7.79011786e-01 7.08043516e-01 1.48189068e+00 -1.34019047e-01
-2.23351955e-01 6.56513721e-02 -2.80016601e-01 -4.73146848e-02
3.65802199e-01 4.19871360e-01 -8.59784424e-01 -3.15212071e-01
-3.26067090e-01 2.79147208e-01 9.28641617e-01 8.23400438e-01
9.81009543e-01 -2.42432415e-01 -5.99078238e-01 5.78329682e-01
1.85457218e+00 2.20214546e-01 4.38240051e-01 9.28255141e-01
3.43173504e-01 5.61373055e-01 3.97454083e-01 8.86858225e-01
5.12752533e-01 5.26163876e-01 7.28331506e-01 -5.76026380e-01
4.03398216e-01 -6.04293123e-02 1.38048068e-01 1.04181826e-01
-5.92088223e-01 -7.34864324e-02 -1.43735433e+00 8.50073874e-01
-1.40772808e+00 -1.47689295e+00 -4.29955870e-01 2.21268582e+00
6.25085592e-01 7.18392357e-02 3.55845124e-01 4.10256088e-01
4.52423304e-01 3.29829097e-01 -3.51008803e-01 -4.69801307e-01
-2.43270457e-01 2.90814519e-01 1.25355661e+00 5.70085466e-01
-9.85553384e-01 3.74524564e-01 6.96217585e+00 5.81218004e-01
-7.82682061e-01 -9.74580124e-02 8.29905987e-01 -1.04337215e-01
-6.56324387e-01 1.18114553e-01 -4.02468681e-01 3.32174599e-01
1.25014651e+00 -4.93648142e-01 8.33372414e-01 2.41820678e-01
1.13354957e+00 -6.33756638e-01 -4.71481532e-01 1.62829310e-01
-5.05045116e-01 -1.04926050e+00 -5.32684326e-01 1.86577752e-01
6.91541910e-01 4.55387384e-01 -5.42597950e-01 2.37012997e-01
6.02614999e-01 -1.05156171e+00 5.20402431e-01 4.39811140e-01
6.56151414e-01 -1.03895187e+00 5.46568513e-01 2.75974452e-01
-1.54492509e+00 -3.52232933e-01 -4.67331527e-04 -4.59079832e-01
3.71262670e-01 1.20794725e+00 -6.16508245e-01 7.16361284e-01
1.26015770e+00 7.89142668e-01 -2.66097009e-01 8.89563739e-01
-1.96366996e-01 1.11440814e+00 -9.69018042e-01 6.13969088e-01
6.38887063e-02 -3.57403666e-01 4.46708441e-01 1.04677045e+00
5.17070651e-01 1.72350958e-01 1.26425266e-01 6.20215714e-01
5.76675773e-01 -6.23459592e-02 -7.52105951e-01 3.55329067e-01
8.37162375e-01 1.15456975e+00 -6.79461956e-01 -4.20437194e-02
-3.94418359e-01 2.50467986e-01 -3.77424449e-01 4.59811002e-01
-7.59801447e-01 -1.59908563e-01 9.64969337e-01 4.16667759e-01
1.88869052e-02 -2.72863209e-01 -7.04498887e-01 -8.32121730e-01
-3.89837563e-01 -3.15766662e-01 5.50929010e-01 -1.02287519e+00
-1.27399540e+00 8.50095227e-02 3.08894396e-01 -8.71647835e-01
-2.55491823e-01 -1.86188608e-01 -1.27230048e+00 1.39501858e+00
-2.04218721e+00 -8.91152978e-01 -1.85018629e-01 4.93059576e-01
2.40490958e-01 1.87535986e-01 1.14434457e+00 1.19899251e-01
-6.42928541e-01 -5.02088107e-02 6.48497760e-01 4.28304449e-02
1.45862937e-01 -1.45849109e+00 4.94736046e-01 7.18306541e-01
-2.01993391e-01 3.89269032e-02 6.11280680e-01 -9.07608032e-01
-8.16157699e-01 -1.13535380e+00 9.10655081e-01 1.06895030e-01
8.21241975e-01 7.48652816e-02 -5.91041028e-01 4.54997122e-01
3.98407668e-01 -5.49114883e-01 7.75300205e-01 5.27554452e-01
-5.55044301e-02 -2.84796774e-01 -1.49778891e+00 4.04506922e-01
6.53437436e-01 -3.49730015e-01 -1.67311504e-01 5.78675389e-01
-1.28515899e-01 3.68729532e-01 -1.34663486e+00 5.94663143e-01
6.90311134e-01 -7.33920991e-01 7.90931940e-01 -2.57825106e-01
5.80366924e-02 -2.62721628e-01 -1.25249073e-01 -1.67761517e+00
-8.45942557e-01 -4.25767750e-01 4.44091946e-01 1.39330184e+00
7.68067181e-01 -6.13706827e-01 4.13866639e-01 5.49238861e-01
2.23161310e-01 -4.28091973e-01 -1.07519948e+00 -8.51086259e-01
5.48337579e-01 -6.15584433e-01 8.38462174e-01 1.13973689e+00
2.18862861e-01 -1.23646885e-01 -1.50421411e-01 6.26867890e-01
8.13618839e-01 1.77700683e-01 -9.72620770e-03 -1.43338084e+00
2.61053652e-01 -5.34821391e-01 -1.64190996e-02 6.11170486e-04
2.53239423e-02 -8.69616866e-01 -2.11211860e-01 -1.93348360e+00
7.96691105e-02 -9.80327547e-01 -4.51700538e-01 1.04110861e+00
2.67516393e-02 4.37350571e-02 -1.69372082e-01 3.19502771e-01
8.15156400e-01 4.39579695e-01 1.82821780e-01 -3.57646435e-01
-3.62094909e-01 1.88218325e-01 -2.98022270e-01 6.79450989e-01
1.36755085e+00 -5.86939335e-01 -9.51724648e-02 -4.11947846e-01
3.03811729e-01 2.32384607e-01 4.85324472e-01 -9.78518307e-01
-3.64065111e-01 -6.14807904e-01 5.91981292e-01 -7.10464060e-01
-2.93483943e-01 -1.11905706e+00 3.89074951e-01 6.29083514e-01
3.14449482e-02 3.05812836e-01 5.26094735e-01 4.49819446e-01
4.02327538e-01 6.26251251e-02 4.46007937e-01 4.33336496e-02
-3.77633214e-01 2.76869535e-01 -1.16910934e+00 -3.47028643e-01
9.26834106e-01 -7.71655366e-02 -4.64472771e-01 -2.94801295e-01
-7.73617566e-01 6.70597315e-01 4.97559369e-01 3.15105945e-01
-2.44652376e-01 -1.10328293e+00 -1.05407524e+00 2.31837407e-01
-1.77534133e-01 -4.27566856e-01 2.76693571e-02 8.05882514e-01
-4.56433088e-01 2.26176113e-01 -2.97409147e-01 -3.75534385e-01
-9.76725698e-01 1.70162767e-01 5.50267637e-01 -3.11154187e-01
-6.48676217e-01 5.52073300e-01 -6.30718708e-01 -4.79086712e-02
-1.37461973e-02 -4.80560303e-01 -2.08443135e-01 5.61157465e-01
3.87049675e-01 7.90878773e-01 2.69290119e-01 -3.67321521e-01
-5.14068067e-01 3.43598038e-01 5.42969108e-01 -2.35255420e-01
1.51814651e+00 -2.60235548e-01 -2.45857403e-01 4.33910638e-01
5.29693663e-01 -3.28146875e-01 -1.05570877e+00 2.32430726e-01
8.36534500e-02 -2.09037751e-01 7.45946407e-01 -1.28546715e+00
-1.22667646e+00 4.23549324e-01 7.38725305e-01 7.72079349e-01
1.41383934e+00 -3.27118695e-01 3.93886417e-01 2.96660326e-02
4.42058802e-01 -1.11716521e+00 -1.21839297e+00 7.06746057e-02
8.17426860e-01 -8.75990033e-01 4.28471208e-01 -3.45381498e-02
-4.62342501e-02 8.36824119e-01 1.54727027e-01 2.30749413e-01
1.22261357e+00 8.56411576e-01 -1.49264187e-01 -2.07980424e-02
-7.48577237e-01 -3.77138942e-01 -6.50409222e-01 7.89228022e-01
2.91386824e-02 6.88260555e-01 -3.07971358e-01 2.84006268e-01
-2.49557391e-01 -2.90807746e-02 2.57991612e-01 7.74226844e-01
-5.62330365e-01 -7.10184813e-01 -5.68517268e-01 1.11224174e+00
-3.22766483e-01 -5.71795523e-01 1.71577912e-02 5.52254260e-01
2.04218850e-01 1.35881770e+00 4.07801598e-01 -1.97354943e-01
1.89892933e-01 7.83024803e-02 -9.75847803e-03 -1.97417349e-01
-7.66310513e-01 -6.82868659e-02 3.17987800e-01 -1.09372333e-01
-3.37059230e-01 -1.04528940e+00 -1.04426861e+00 -8.98362219e-01
-6.36986375e-01 3.17914009e-01 8.77037823e-01 1.03259087e+00
-3.59055102e-02 4.84438241e-01 1.14443433e+00 -1.11946750e+00
-2.35252246e-01 -1.26310503e+00 -1.15258658e+00 -2.26606771e-01
3.44990671e-01 -6.36560977e-01 -9.57599998e-01 -2.20297307e-01] | [6.202621936798096, 2.809185028076172] |
fa997e03-8176-4ee3-a632-9ee4fbc3e9d9 | facial-expression-recognition-at-the-edge-cpu | 2305.15422 | null | https://arxiv.org/abs/2305.15422v1 | https://arxiv.org/pdf/2305.15422v1.pdf | Facial Expression Recognition at the Edge: CPU vs GPU vs VPU vs TPU | Facial Expression Recognition (FER) plays an important role in human-computer interactions and is used in a wide range of applications. Convolutional Neural Networks (CNN) have shown promise in their ability to classify human facial expressions, however, large CNNs are not well-suited to be implemented on resource- and energy-constrained IoT devices. In this work, we present a hierarchical framework for developing and optimizing hardware-aware CNNs tuned for deployment at the edge. We perform a comprehensive analysis across various edge AI accelerators including NVIDIA Jetson Nano, Intel Neural Compute Stick, and Coral TPU. Using the proposed strategy, we achieved a peak accuracy of 99.49% when testing on the CK+ facial expression recognition dataset. Additionally, we achieved a minimum inference latency of 0.39 milliseconds and a minimum power consumption of 0.52 Watts. | ['Ramtin Zand', 'Lareb Khan', 'Heath Smith', 'Mohammadreza Mohammadi'] | 2023-05-17 | null | null | null | null | ['facial-expression-recognition'] | ['computer-vision'] | [-8.12350735e-02 -3.00418407e-01 -1.76462993e-01 -6.38072550e-01
1.04999661e-01 -2.22051933e-01 1.83641925e-01 -2.31116965e-01
-6.32439733e-01 4.11731005e-01 -3.16063315e-01 -3.89869779e-01
2.91871339e-01 -8.29851210e-01 -3.35320234e-01 -6.25144780e-01
-7.60751292e-02 -7.33698457e-02 -1.31016061e-01 -1.48650691e-01
-3.25476766e-01 9.66986895e-01 -1.81608844e+00 2.51107097e-01
2.39171982e-01 1.88953412e+00 -3.84218872e-01 7.88926661e-01
2.24352702e-01 1.13989401e+00 -7.74899840e-01 -6.10218227e-01
1.02146700e-01 -5.79837635e-02 -3.20100367e-01 -5.31277478e-01
2.57591337e-01 -4.89276499e-01 -3.49715263e-01 8.74383271e-01
7.58093834e-01 -7.93443695e-02 8.25337619e-02 -1.79908288e+00
4.71629724e-02 1.34169042e-01 -4.60863143e-01 4.25827026e-01
-5.27925976e-03 9.16102156e-03 4.44119126e-01 -6.24915302e-01
2.72672087e-01 9.43788648e-01 6.88181341e-01 7.67650187e-01
-4.54857230e-01 -1.20605969e+00 -3.57441843e-01 2.54181862e-01
-1.37990129e+00 -8.79730999e-01 6.54440701e-01 1.09306112e-01
1.73726285e+00 2.71764845e-01 1.03244543e+00 1.18744874e+00
5.46499848e-01 6.31706655e-01 8.35140646e-01 -2.00330719e-01
4.75121617e-01 -2.11267427e-01 -1.51689639e-02 8.67723882e-01
3.81080434e-02 -6.69587106e-02 -9.64388311e-01 -2.48003066e-01
4.70018476e-01 -1.72954455e-01 4.01506633e-01 4.95496184e-01
-5.43166280e-01 6.67908847e-01 5.57532787e-01 1.47241756e-01
-5.23249447e-01 9.28704262e-01 1.10741341e+00 5.61302342e-03
5.42712808e-01 1.08659081e-01 -4.30216759e-01 -7.65635550e-01
-5.86760521e-01 -6.08253144e-02 7.43830502e-01 8.97571564e-01
2.36814588e-01 6.48573697e-01 1.05121292e-01 5.45018196e-01
3.35268117e-02 5.38911343e-01 2.68835247e-01 -9.14121568e-01
-7.59666935e-02 5.79703033e-01 -2.68039227e-01 -9.37434733e-01
-7.51324654e-01 -1.67373747e-01 -1.15537274e+00 2.94305623e-01
-1.12486081e-02 -5.06562769e-01 -7.60756195e-01 1.57310808e+00
3.96067291e-01 2.74443984e-01 -8.25804472e-03 9.99378741e-01
9.16307569e-01 5.28329611e-01 5.18852472e-01 9.93147790e-02
1.70167208e+00 -9.60081935e-01 -5.91685832e-01 -2.17833251e-01
6.42537117e-01 -4.82850730e-01 7.39651382e-01 5.26722491e-01
-9.87886548e-01 -5.69541156e-01 -1.33497226e+00 -2.37420902e-01
-2.72383630e-01 4.67433572e-01 1.51658010e+00 1.12821293e+00
-1.21704185e+00 3.31364632e-01 -1.30259562e+00 -2.68291563e-01
7.87681043e-01 9.17167664e-01 -1.74386665e-01 3.71934861e-01
-8.27589333e-01 5.05028188e-01 5.56073152e-02 3.74307334e-01
-4.21626061e-01 -5.57326794e-01 -5.84460080e-01 3.54438394e-01
-1.71346903e-01 -6.84557974e-01 1.39466321e+00 -1.22548401e+00
-1.99880230e+00 7.51611948e-01 -1.03087932e-01 -5.35385370e-01
2.44846120e-02 -1.89899132e-01 -7.92583525e-01 8.48361030e-02
-4.58815277e-01 9.11700964e-01 5.68486035e-01 -8.30718726e-02
-6.14782751e-01 -4.41180825e-01 2.67121959e-02 -9.56474990e-02
-6.42893434e-01 6.15175188e-01 -1.57857373e-01 -1.27356976e-01
-2.52939105e-01 -1.14501452e+00 -1.01088315e-01 3.05563599e-01
3.99772525e-02 -3.16695064e-01 1.06041396e+00 -1.47911668e-01
8.91824424e-01 -2.18551064e+00 -4.78418648e-01 1.49861202e-01
1.95366278e-01 5.58563292e-01 3.53355706e-01 -2.53450215e-01
1.71499595e-01 -8.59585777e-02 4.97481942e-01 -3.74063194e-01
7.77838379e-02 2.44279012e-01 -8.21447000e-03 4.04724687e-01
1.62156865e-01 9.58305180e-01 -5.61827183e-01 -2.30125710e-01
2.13231549e-01 8.50107789e-01 -6.22941732e-01 7.32600763e-02
4.48392518e-02 1.45378426e-01 -4.88254905e-01 1.21519387e+00
5.54802895e-01 -1.77200750e-01 2.67900497e-01 -3.39551181e-01
-5.64681999e-02 1.98200811e-02 -6.29932106e-01 1.43087995e+00
-6.78799808e-01 1.03495133e+00 2.98385620e-01 -5.74849844e-01
9.37328935e-01 2.87636876e-01 4.59531128e-01 -1.00434935e+00
9.06017363e-01 2.06644773e-01 1.51002079e-01 -2.58499444e-01
7.11887538e-01 9.16764233e-03 -1.76685646e-01 2.69822180e-01
2.50570565e-01 4.04960990e-01 1.26429023e-02 -2.83530980e-01
1.34097469e+00 -1.69647932e-01 1.62387311e-01 -3.71439964e-01
1.36995405e-01 -7.12403134e-02 5.33001065e-01 3.61904144e-01
-4.84574109e-01 -1.70691147e-01 4.53873307e-01 -1.10677540e+00
-8.83654237e-01 -6.04625881e-01 -2.33783901e-01 1.24246073e+00
-2.85352707e-01 -5.74424446e-01 -8.93516183e-01 -1.08260661e-01
-4.45421755e-01 3.33403975e-01 -3.33365679e-01 -2.87907898e-01
-4.94867295e-01 -8.83505881e-01 1.13968778e+00 1.02829874e+00
1.00505280e+00 -1.00910044e+00 -1.51609838e+00 1.20068796e-01
3.86912227e-01 -1.57452536e+00 1.52923644e-01 3.51725608e-01
-5.52979648e-01 -6.57239974e-01 -9.15404409e-02 -5.39229274e-01
4.66650277e-01 -1.83337569e-01 1.08634019e+00 1.38579100e-01
-6.51931763e-01 1.05967388e-01 -1.12943597e-01 -7.99597800e-01
8.65718573e-02 1.16216347e-01 4.40333664e-01 -5.83106689e-02
6.74976885e-01 -4.89583760e-01 -7.50483036e-01 6.41498044e-02
-4.30402458e-01 6.52110875e-02 4.59881812e-01 6.31128371e-01
4.12884682e-01 -8.60938281e-02 2.73005366e-01 -5.14606893e-01
4.82883096e-01 -4.17459369e-01 -7.50961304e-01 -6.31467104e-02
-3.43138844e-01 -4.37203199e-01 8.11859846e-01 -5.11462033e-01
-9.24437761e-01 4.07446027e-01 -6.41070366e-01 -5.67846715e-01
-1.50182873e-01 1.44215137e-01 -1.01833232e-01 -5.02514482e-01
4.74215090e-01 -3.67116511e-01 -9.34734568e-02 2.16729119e-01
-1.48513779e-01 8.32045019e-01 5.82397878e-01 -6.25542939e-01
-2.28460088e-01 4.21419442e-01 4.60380822e-01 -1.05038643e+00
-4.59197998e-01 -2.73074992e-02 -1.53017983e-01 -5.15391767e-01
9.85344648e-01 -1.26102197e+00 -1.67610705e+00 6.96717620e-01
-1.03940916e+00 -4.04947758e-01 2.31854543e-01 3.47561508e-01
-2.27174342e-01 -3.01599711e-01 -8.70661616e-01 -9.00243521e-01
-1.15059018e+00 -1.21035922e+00 1.23890293e+00 8.42479944e-01
-4.43999976e-01 -7.11629391e-01 -3.60170752e-01 1.31138474e-01
8.75762582e-01 4.40849215e-01 3.54034871e-01 -1.89390957e-01
-2.92573333e-01 -3.07770908e-01 -4.10512894e-01 1.58414260e-01
-2.87548244e-01 2.91961461e-01 -1.41415954e+00 -3.32308739e-01
-1.97384685e-01 -7.10901916e-01 2.59713501e-01 2.38061994e-01
1.66029143e+00 -2.49547623e-02 -3.26315403e-01 1.01253545e+00
1.21657372e+00 4.87157166e-01 7.01746225e-01 3.56387906e-02
5.33658803e-01 -6.74106479e-02 3.18211526e-01 7.52533734e-01
-1.83674484e-03 7.71727800e-01 5.16567290e-01 -1.44074425e-01
2.16180801e-01 5.82390279e-02 2.65177041e-01 5.91405571e-01
-2.38336071e-01 -4.33697015e-01 -1.00943112e+00 3.95117290e-02
-1.55623674e+00 -6.16313696e-01 -1.54222235e-01 1.61690998e+00
2.04431310e-01 2.63888221e-02 -8.00816808e-03 -5.32638468e-02
2.24365592e-01 -4.76840995e-02 -7.95087218e-01 -1.06561923e+00
3.79506424e-02 1.01855767e+00 4.01157588e-01 -1.71501935e-01
-1.00606120e+00 1.00417149e+00 6.39662552e+00 8.00575256e-01
-1.66358101e+00 2.34168857e-01 8.99191856e-01 -5.17532349e-01
4.96033788e-01 -4.74394858e-01 -8.61277580e-01 4.39338654e-01
1.70291781e+00 -2.60791536e-02 1.35888949e-01 1.50522292e+00
-1.12359405e-01 -2.70336270e-01 -8.21351171e-01 1.40373230e+00
-9.69298929e-02 -1.33644831e+00 -5.12034595e-01 5.77506013e-02
2.76973456e-01 2.38355324e-01 -6.45739725e-03 3.26718450e-01
5.32523766e-02 -1.25412941e+00 5.82003474e-01 -1.97736118e-02
1.13417590e+00 -1.32407391e+00 1.05564749e+00 -4.89153415e-02
-1.13574064e+00 -8.55844095e-02 -5.70391178e-01 -5.90296090e-01
6.47830889e-02 4.62157935e-01 -6.36741340e-01 -3.58447880e-02
1.19248426e+00 2.95902908e-01 -1.92306608e-01 2.56135851e-01
-6.72919899e-02 5.37739038e-01 -6.22653365e-01 -6.93395853e-01
1.59229591e-01 4.52651046e-02 -1.64229155e-01 1.27456272e+00
4.65654284e-01 5.17443120e-01 -1.90204799e-01 3.87879670e-01
-3.56273204e-01 -1.94685072e-01 -3.05344075e-01 -4.80203256e-02
3.45450252e-01 1.84111941e+00 -9.10589695e-01 -3.12684715e-01
-2.82595634e-01 1.11285806e+00 2.98034549e-01 -2.65573263e-01
-1.33207405e+00 -3.53844404e-01 1.38292933e+00 -3.21524531e-01
-2.55204886e-01 -3.30865055e-01 -2.67677218e-01 -8.54958832e-01
6.53571635e-02 -5.79419255e-01 5.68439402e-02 -9.45093811e-01
-6.95176244e-01 1.00554466e+00 -4.46160883e-01 -7.32977211e-01
-2.77427554e-01 -1.05596626e+00 -6.05408132e-01 5.54092824e-01
-1.04740715e+00 -1.19984043e+00 -9.25809622e-01 5.06884038e-01
1.28609985e-01 -1.72288731e-01 1.26844633e+00 6.72369957e-01
-9.28859115e-01 9.31148589e-01 -3.98688912e-01 2.57260114e-01
2.44722068e-02 -5.97795248e-01 5.60324132e-01 6.13222241e-01
-1.57674253e-01 5.61828971e-01 5.18906534e-01 -1.87463745e-01
-1.98639667e+00 -1.05554461e+00 4.53566968e-01 6.58126399e-02
5.59509814e-01 -6.69844866e-01 -3.70205939e-01 6.07299328e-01
2.34662876e-01 6.05237663e-01 1.01710248e+00 3.48168314e-02
-2.20040604e-01 -3.30344319e-01 -1.34185421e+00 6.70313656e-01
1.13572145e+00 -5.26376963e-01 5.91799736e-01 3.10413867e-01
1.60928220e-01 -9.61543500e-01 -9.38760877e-01 3.92404258e-01
9.62242842e-01 -1.01501286e+00 6.41144753e-01 -5.94109178e-01
4.05480266e-01 3.97268534e-02 -2.64140457e-01 -8.87382329e-01
-9.32069793e-02 -5.82552195e-01 -3.14210922e-01 9.83110905e-01
-3.50972600e-02 -5.81189990e-01 1.23575270e+00 9.00662422e-01
-2.64437228e-01 -1.05884910e+00 -1.26601040e+00 -5.21557450e-01
-5.00892878e-01 -6.99256837e-01 5.33283353e-01 6.10636115e-01
1.33995637e-01 3.49193722e-01 -2.63332754e-01 1.00446567e-01
2.92768925e-01 -2.37385467e-01 6.48162067e-01 -7.12843120e-01
-8.28620698e-03 -3.10524046e-01 -1.06345010e+00 -6.40218854e-01
5.06213963e-01 -6.15435302e-01 -3.79004508e-01 -7.02653050e-01
5.57127893e-02 -4.03751910e-01 -5.88648878e-02 7.80934870e-01
5.03103495e-01 6.35657787e-01 4.79671136e-02 -4.52753991e-01
-6.48843348e-01 4.37916517e-01 4.11327988e-01 1.08616874e-01
2.54598767e-01 -2.19245896e-01 -3.37765962e-01 8.37592542e-01
1.18522751e+00 -2.38645688e-01 -3.34383577e-01 -5.11566460e-01
3.94870400e-01 -3.19184177e-02 2.90924907e-01 -1.40244341e+00
4.95908171e-01 1.54717162e-01 8.34598601e-01 -9.26228985e-02
7.89709568e-01 -8.02560568e-01 4.83760566e-01 5.23596644e-01
2.58548528e-01 4.68454331e-01 6.99861944e-01 -1.28746361e-01
-6.60352260e-02 2.77990997e-01 8.54027152e-01 2.27879673e-01
-1.04053867e+00 4.01501805e-01 -6.71262264e-01 -5.41074991e-01
1.20340443e+00 -9.00305212e-02 -4.88956094e-01 -1.69927880e-01
-3.25512916e-01 -1.94614716e-02 2.44656458e-01 4.66771990e-01
6.08155787e-01 -1.33084464e+00 -1.72551170e-01 4.32696193e-01
5.73841203e-03 -2.85417229e-01 4.55378503e-01 6.18254364e-01
-9.68498468e-01 4.33584660e-01 -7.87525594e-01 -6.77569568e-01
-1.70549166e+00 -4.01501032e-03 5.23122013e-01 1.15065694e-01
-4.13562149e-01 9.68199968e-01 -3.68373185e-01 -3.83302942e-02
1.44085348e-01 -2.73074389e-01 1.83217272e-01 -2.80808508e-01
8.68020535e-01 4.25347835e-01 5.42323589e-01 -5.59177577e-01
-6.28155410e-01 2.93705553e-01 2.66050905e-01 2.58049041e-01
1.21731234e+00 4.32392031e-01 -1.72614425e-01 -6.04709797e-02
1.30713224e+00 -4.53651786e-01 -8.77916574e-01 5.05942404e-01
-2.60169089e-01 -1.72986493e-01 4.98253971e-01 -5.51510453e-01
-1.47605741e+00 8.30272615e-01 9.81785953e-01 -2.36503020e-01
1.59094775e+00 -2.69008696e-01 1.00989306e+00 4.91239667e-01
7.57843196e-01 -1.06600761e+00 -9.35318321e-02 5.50661564e-01
2.30856910e-01 -1.03338671e+00 4.99567538e-02 -4.84837592e-01
-4.22771364e-01 1.35505784e+00 1.01743472e+00 -3.06448769e-02
7.65088797e-01 1.11116385e+00 -2.43022367e-02 -4.25788492e-01
-9.33713257e-01 4.54685129e-02 -1.95504427e-01 3.72593462e-01
6.07698262e-01 3.98342490e-01 8.70621726e-02 6.71250522e-01
-4.28307861e-01 3.33042026e-01 2.44789526e-01 1.15312910e+00
-9.54755023e-03 -6.40813410e-01 1.56632230e-01 5.21572709e-01
-8.33643377e-01 1.20442368e-01 -2.57239848e-01 6.00296557e-01
2.25322351e-01 8.78401995e-01 6.91755176e-01 -9.19153869e-01
2.18678892e-01 -6.70988262e-02 4.64732945e-01 -3.37187722e-02
-9.88097250e-01 -3.33971053e-01 4.43827033e-01 -1.06812572e+00
-3.65090251e-01 -2.25326955e-01 -1.36619473e+00 -1.01712942e+00
1.56255811e-01 -2.34056473e-01 1.28014171e+00 6.48032546e-01
9.13181067e-01 6.37410522e-01 3.40553194e-01 -1.03438759e+00
6.02663159e-02 -6.75518513e-01 -4.54760522e-01 -1.64327964e-01
-1.43291384e-01 -6.49135113e-01 2.54782964e-03 -3.68471533e-01] | [8.427695274353027, 2.7749664783477783] |
17babf5c-1044-402f-8ba4-9069b9aa7883 | a-probabilistic-rotation-representation-for | 2305.18947 | null | https://arxiv.org/abs/2305.18947v1 | https://arxiv.org/pdf/2305.18947v1.pdf | A Probabilistic Rotation Representation for Symmetric Shapes With an Efficiently Computable Bingham Loss Function | In recent years, a deep learning framework has been widely used for object pose estimation. While quaternion is a common choice for rotation representation, it cannot represent the ambiguity of the observation. In order to handle the ambiguity, the Bingham distribution is one promising solution. However, it requires complicated calculation when yielding the negative log-likelihood (NLL) loss. An alternative easy-to-implement loss function has been proposed to avoid complex computations but has difficulty expressing symmetric distribution. In this paper, we introduce a fast-computable and easy-to-implement NLL loss function for Bingham distribution. We also create the inference network and show that our loss function can capture the symmetric property of target objects from their point clouds. | ['Koichi Nishiwaki', 'Takuya Ikeda', 'Hiroya Sato'] | 2023-05-30 | null | null | null | null | ['pose-estimation'] | ['computer-vision'] | [-4.53729302e-01 -2.25399837e-01 -3.30263793e-01 -4.96646643e-01
-6.09702587e-01 -2.72157103e-01 3.75829607e-01 -8.06264207e-02
-4.57488745e-01 8.39352608e-01 -4.44206089e-01 -1.39319301e-01
-1.81011707e-01 -9.41950440e-01 -8.81105065e-01 -7.57511914e-01
1.18451342e-01 5.14751673e-01 1.00541957e-01 5.90406470e-02
5.49124703e-02 8.20840955e-01 -1.20796454e+00 -5.25114715e-01
7.39029944e-01 1.11409497e+00 1.89390212e-01 6.32001534e-02
8.84661153e-02 3.80885750e-01 -6.38075531e-01 -4.35965925e-01
3.62726688e-01 -2.45985195e-01 -4.89713997e-01 -3.27362604e-02
2.76305139e-01 -4.84337717e-01 -2.77154803e-01 1.23973632e+00
5.69851279e-01 -7.29892626e-02 8.79470944e-01 -1.45812654e+00
-3.23518127e-01 1.59699172e-01 -7.08793283e-01 -1.88640997e-01
2.38314375e-01 -2.90020794e-01 9.44298267e-01 -8.10271442e-01
5.83156943e-01 1.32944131e+00 7.29557931e-01 -6.17715195e-02
-9.33151364e-01 -8.16946924e-01 -1.33343875e-01 2.25866720e-01
-1.66426611e+00 3.59497011e-01 8.41439188e-01 -1.68285176e-01
4.79143590e-01 1.61167681e-01 6.96761906e-01 9.02419150e-01
3.58229190e-01 1.01574886e+00 9.37754691e-01 -1.84750587e-01
-3.79113555e-02 7.24075213e-02 -3.02617639e-01 6.35645628e-01
5.51767945e-01 -2.04805896e-01 -1.09088115e-01 -5.36852404e-02
1.18723702e+00 1.52022168e-01 -1.72001213e-01 -7.53105402e-01
-1.04417205e+00 8.63315642e-01 7.51581669e-01 -1.90374464e-01
-2.61146635e-01 3.21462750e-01 1.78634435e-01 -2.37460108e-03
3.37146670e-01 2.24608243e-01 -3.40095609e-01 -1.96739703e-01
-4.69083041e-01 6.24834418e-01 7.55139947e-01 8.87785733e-01
8.70723844e-01 -2.00585872e-02 2.09118679e-01 6.78482175e-01
7.57701516e-01 7.47328997e-01 2.07544789e-01 -6.23994827e-01
2.97735721e-01 3.69418055e-01 2.33874068e-01 -1.27674031e+00
-5.56079745e-01 -6.54280663e-01 -1.20255744e+00 3.40745568e-01
4.89458799e-01 -5.72721520e-03 -6.39558315e-01 1.63280582e+00
4.47459161e-01 1.33201405e-01 -2.75176197e-01 1.20263529e+00
3.00908595e-01 6.73490882e-01 -2.38021389e-01 -6.77439347e-02
1.12339914e+00 -4.57259983e-01 -6.81025386e-01 1.39606625e-01
2.57904559e-01 -8.11383486e-01 8.86812747e-01 5.49420834e-01
-9.13822234e-01 -2.87248135e-01 -1.15953493e+00 -1.61641777e-01
-1.03667274e-01 1.29893184e-01 8.96274924e-01 4.45734769e-01
-6.54281914e-01 4.81631607e-01 -9.49904501e-01 -1.12524852e-01
2.96720982e-01 3.87074530e-01 -3.64417702e-01 2.20658273e-01
-9.97790217e-01 1.22301233e+00 4.32132691e-01 4.40439172e-02
-6.45056009e-01 -3.05078864e-01 -7.92160809e-01 8.36411640e-02
3.46445233e-01 -5.85728109e-01 1.16189492e+00 -5.13612986e-01
-1.81211603e+00 6.95877075e-01 1.27041191e-01 -4.63384837e-01
7.91825891e-01 -4.44509000e-01 3.84714417e-02 -1.01767175e-01
-1.77849591e-01 4.88111466e-01 8.97507250e-01 -9.94839072e-01
-3.52802634e-01 -2.74626851e-01 2.17381477e-01 3.33932966e-01
-2.86688898e-02 -1.40496582e-01 -3.11206758e-01 -7.51461983e-01
6.25714362e-01 -9.50219452e-01 -1.45455196e-01 6.26270175e-01
-3.40253234e-01 -3.92119616e-01 7.37510860e-01 -4.32761788e-01
9.83676255e-01 -2.06034589e+00 -1.29826978e-01 1.45972893e-01
-3.27364691e-02 3.27190429e-01 8.85534585e-02 3.54048669e-01
8.77846479e-02 -8.39811265e-02 -2.04606146e-01 -1.37399986e-01
3.75743359e-01 2.24743322e-01 -5.29543161e-01 8.68770599e-01
3.89713526e-01 6.84649646e-01 -8.14728498e-01 -3.64526421e-01
3.06878835e-01 6.90349221e-01 -5.71713150e-01 2.75705040e-01
-9.91233066e-02 2.12139413e-01 -4.78294671e-01 4.88340437e-01
1.18256557e+00 -1.98774204e-01 -2.68818643e-02 -4.21031505e-01
7.79802725e-02 2.61708677e-01 -1.52940989e+00 1.27193022e+00
-3.34912717e-01 5.04078269e-01 -7.53243566e-02 -1.06917143e+00
1.25966716e+00 -8.53236318e-02 5.44649780e-01 -3.74846607e-01
2.61095673e-01 2.45705619e-01 -9.76595357e-02 -2.28961617e-01
3.77053916e-01 -4.74959403e-01 -1.44479290e-01 2.60277778e-01
-7.73958936e-02 -3.81268919e-01 3.51850577e-02 -4.15914744e-01
4.53885496e-01 3.74884725e-01 7.34555900e-01 3.44452821e-02
3.19666386e-01 -4.18186277e-01 7.62083530e-01 3.33228856e-01
-1.41548514e-01 9.20885026e-01 4.68185335e-01 -5.49371421e-01
-8.37294519e-01 -1.26549327e+00 -4.83802944e-01 3.42077851e-01
5.40992856e-01 -8.45408440e-02 -2.97422916e-01 -4.65817034e-01
2.69455194e-01 3.27392817e-01 1.31523991e-02 -2.81499207e-01
-6.96798205e-01 -7.34823644e-01 5.16698778e-01 6.77025080e-01
8.34121525e-01 -5.58493078e-01 -6.11215532e-01 9.45993364e-02
-1.42759860e-01 -7.81669676e-01 -2.43895561e-01 -4.75302301e-02
-7.06260443e-01 -1.05607212e+00 -8.84844780e-01 -5.62034369e-01
5.78357458e-01 5.96189499e-02 8.92488420e-01 -3.29091996e-01
-1.40439287e-01 -4.41647209e-02 -1.31282121e-01 -6.79157019e-01
1.50918067e-01 -5.23549616e-02 1.42478272e-01 -1.45826221e-01
4.75378454e-01 -3.62868071e-01 -5.20497084e-01 3.52556169e-01
-8.58956039e-01 -2.19074920e-01 6.24452531e-01 7.18112528e-01
8.61330807e-01 6.05589449e-02 4.53918666e-01 -2.23898411e-01
4.95786637e-01 -2.88213730e-01 -1.07835329e+00 1.49236366e-01
-3.68216842e-01 2.90245473e-01 4.86467183e-01 -4.83641863e-01
-6.50481343e-01 1.00566909e-01 -4.26770985e-01 -4.95937526e-01
7.72163048e-02 6.21281564e-01 -2.27762446e-01 -3.52091283e-01
8.88808295e-02 1.22691214e-01 9.89572257e-02 -5.63302457e-01
2.48291016e-01 6.08278215e-01 3.30751926e-01 -6.83887780e-01
9.30415630e-01 5.07093728e-01 4.60281789e-01 -7.51163125e-01
-8.80260885e-01 -2.22784743e-01 -3.47005844e-01 -1.60537381e-02
7.49923885e-01 -8.43436658e-01 -1.18021452e+00 6.58270538e-01
-1.37543595e+00 1.86408713e-01 -1.11002363e-01 9.17238176e-01
-6.52240396e-01 5.86790383e-01 -3.85854542e-01 -1.00210142e+00
-9.46606994e-02 -1.14235246e+00 1.15856659e+00 2.33517900e-01
7.85736516e-02 -7.39884377e-01 2.49461066e-02 -3.75186026e-01
2.47396529e-01 4.93349105e-01 5.62094450e-01 -2.95083523e-01
-8.37162673e-01 -4.35186505e-01 -4.38377023e-01 5.14294386e-01
-3.99378240e-02 1.34821817e-01 -6.22983813e-01 -4.01372015e-01
2.86376864e-01 -4.28465784e-01 5.30577302e-01 3.63867402e-01
1.44916284e+00 -7.60971755e-02 -8.47842470e-02 9.83537734e-01
1.44388974e+00 1.10321119e-01 6.53066635e-01 5.14275432e-01
5.88344872e-01 1.62435129e-01 9.08962965e-01 6.61367536e-01
5.77999234e-01 6.04750395e-01 6.84394300e-01 1.00025035e-01
2.51601189e-01 -3.20072085e-01 1.17343247e-01 7.92410851e-01
-4.87995856e-02 -3.29976439e-01 -7.00614989e-01 9.48894024e-02
-1.75194454e+00 -5.08513033e-01 6.27730861e-02 2.30094314e+00
7.79495418e-01 1.40597939e-01 -4.56049060e-03 1.15849771e-01
5.54360211e-01 1.39801800e-01 -4.88110334e-01 -1.03286535e-01
-2.30121464e-01 1.78861037e-01 5.45305908e-01 4.49663877e-01
-1.21660864e+00 6.60456002e-01 6.66173792e+00 9.32800412e-01
-1.41229284e+00 -1.85691908e-01 2.39082485e-01 3.43654782e-01
-2.20289454e-01 -6.12706356e-02 -9.48607087e-01 4.23042625e-01
1.55372679e-01 -2.87171841e-01 -4.15359735e-02 1.14696014e+00
-2.19911441e-01 3.28394547e-02 -9.62214708e-01 1.25364149e+00
3.45420577e-02 -9.23045754e-01 -2.76007149e-02 3.74036655e-02
3.65541786e-01 -6.27972791e-03 -2.05730461e-02 2.74569660e-01
-1.13620937e-01 -8.48470271e-01 7.55187988e-01 5.23336411e-01
7.93787420e-01 -9.65336859e-01 8.28082621e-01 5.46959460e-01
-1.21707916e+00 1.37392864e-01 -9.41072106e-01 -2.88927734e-01
3.42700303e-01 6.58784270e-01 -8.37827981e-01 5.85962415e-01
4.76382941e-01 5.30030906e-01 -2.33739510e-01 1.53197920e+00
-4.37383980e-01 2.81894803e-01 -7.32091486e-01 -2.99645275e-01
2.52442837e-01 -5.64614594e-01 6.86743438e-01 8.54214668e-01
6.97620690e-01 -2.24271178e-01 5.42383790e-01 8.36528540e-01
-7.08564371e-02 -5.91397062e-02 -6.10213697e-01 6.34360239e-02
3.79126936e-01 1.01209128e+00 -5.93881667e-01 3.08765862e-02
-2.25977853e-01 6.49325728e-01 4.13433760e-01 3.88674200e-01
-1.08162057e+00 -6.39766514e-01 8.42614055e-01 1.05277605e-01
2.30554849e-01 -4.87114817e-01 -1.33078188e-01 -1.33106232e+00
1.67196780e-01 -6.47822678e-01 9.28472281e-02 -7.62716711e-01
-1.41546178e+00 3.64971995e-01 3.50509793e-01 -1.64122987e+00
-3.23743105e-01 -1.14688039e+00 -3.66365314e-01 9.86385047e-01
-1.61612773e+00 -1.02016222e+00 -2.00664148e-01 3.35532933e-01
5.74970897e-03 -4.58528139e-02 6.03769243e-01 4.54465419e-01
-4.11003917e-01 5.93811810e-01 1.40509546e-01 6.24007396e-02
7.57212520e-01 -1.36485088e+00 -1.19849276e-02 5.68608522e-01
-3.95462997e-02 7.54029512e-01 8.74948800e-01 -4.77588207e-01
-1.47759569e+00 -7.98497558e-01 6.95806086e-01 -3.46370786e-02
5.75636387e-01 -1.67001069e-01 -9.62217629e-01 7.33360827e-01
-1.02019124e-01 1.19724736e-01 2.55618036e-01 -2.40568250e-01
-3.34728241e-01 -4.14681494e-01 -1.16847587e+00 5.59388757e-01
5.05501866e-01 -3.80773723e-01 -6.15980029e-01 2.88275659e-01
4.57461357e-01 -7.78456748e-01 -8.10573041e-01 8.06564629e-01
5.61035395e-01 -8.41897011e-01 1.10877502e+00 -2.72562534e-01
8.17236677e-02 -5.63023090e-01 -1.29081681e-01 -1.01878655e+00
-3.20391446e-01 -5.68549216e-01 -2.01177508e-01 9.79618251e-01
-1.12922385e-01 -1.06212556e+00 7.94657409e-01 2.13350430e-01
2.15082109e-01 -1.07494950e+00 -9.45708215e-01 -9.51540470e-01
2.40710929e-01 -4.59196448e-01 7.10938811e-01 6.31272733e-01
-3.46431494e-01 8.38167593e-02 -5.48881054e-01 1.34173110e-01
7.03471661e-01 3.05351824e-01 9.49920237e-01 -1.33602440e+00
-2.12946683e-01 -4.41272825e-01 -8.34963500e-01 -1.53759086e+00
2.63482749e-01 -8.55746210e-01 2.83104237e-02 -1.41412163e+00
3.09239421e-02 -4.44266707e-01 -2.34391123e-01 1.69625074e-01
1.08635435e-02 3.09413463e-01 1.75032914e-01 7.87732229e-02
-4.08036292e-01 8.94425988e-01 1.51130724e+00 1.77306812e-02
2.93160915e-01 2.08940268e-01 -2.72240311e-01 1.11439991e+00
8.16923797e-01 -4.79252666e-01 -3.25912386e-01 -3.86644542e-01
5.11347353e-01 -1.79278791e-01 5.32345355e-01 -8.28839719e-01
8.60339180e-02 -1.48687050e-01 4.98703301e-01 -1.21761000e+00
4.44096714e-01 -8.40136111e-01 1.17926694e-01 5.51729679e-01
1.30814001e-01 1.56470641e-01 -8.70667920e-02 4.31206256e-01
-5.27173460e-01 -4.02050287e-01 8.83695066e-01 3.89869958e-02
-3.13793451e-01 6.26950860e-01 -2.30886787e-02 -1.19987436e-01
9.15458322e-01 -7.58783370e-02 -1.00867085e-01 -4.75927204e-01
-2.46404141e-01 1.08669758e-01 5.21048367e-01 3.34896713e-01
6.88820660e-01 -1.58557212e+00 -6.14366889e-01 2.44746774e-01
-1.77629426e-01 4.02989447e-01 -2.93760329e-01 7.60379076e-01
-1.04120636e+00 3.94054443e-01 -1.88318714e-01 -8.15012455e-01
-8.48474979e-01 3.10184389e-01 2.76658714e-01 -3.50349009e-01
-4.81760919e-01 8.08212221e-01 2.36233801e-01 -6.57165647e-01
3.76030833e-01 -5.99246025e-01 9.89655182e-02 -8.45947117e-02
2.57197797e-01 4.73261833e-01 -2.21906975e-01 -6.19570434e-01
-4.94821012e-01 7.81404853e-01 2.29521483e-01 3.68019869e-03
1.19547009e+00 -9.67408344e-02 -2.25268632e-01 4.87885237e-01
1.33627069e+00 -3.17460477e-01 -1.37026608e+00 -1.98832035e-01
-4.94821742e-03 -6.53239012e-01 -2.78849900e-01 -2.08130658e-01
-8.23293984e-01 1.16906488e+00 4.59975570e-01 3.14688295e-01
9.73711848e-01 -2.93063223e-01 8.93656969e-01 7.28982270e-01
7.05947161e-01 -9.25255895e-01 1.06740005e-01 8.03218246e-01
1.09639299e+00 -1.24122000e+00 2.80642509e-01 -6.07943892e-01
-3.74727339e-01 1.26511610e+00 7.41936982e-01 -4.82381821e-01
9.44737971e-01 2.43851930e-01 -2.80425302e-03 1.43354043e-01
-1.59515992e-01 -7.44203106e-02 3.62174898e-01 7.29114175e-01
5.19931853e-01 -6.36901148e-03 -5.45079827e-01 4.53970134e-01
-4.20002043e-01 -8.66846144e-02 1.52426541e-01 7.93286800e-01
-3.50595057e-01 -1.07294917e+00 -3.03344101e-01 9.05763283e-02
-4.83413607e-01 3.60665590e-01 -4.63333167e-02 9.73616242e-01
4.23146077e-02 2.60009766e-01 3.89845371e-02 3.42990234e-02
2.95343935e-01 -4.24684882e-01 7.11506188e-01 -3.17530900e-01
1.82618365e-01 2.38454521e-01 -3.93287808e-01 -5.34271419e-01
-4.03345764e-01 -2.76978314e-01 -1.32647049e+00 -1.34266526e-01
-4.92337227e-01 -1.09918289e-01 7.43371248e-01 8.16910446e-01
1.78854670e-02 1.18384317e-01 6.58521473e-01 -9.85069513e-01
-1.17358387e+00 -8.60709190e-01 -7.63372958e-01 3.88341963e-01
3.57632995e-01 -1.06766033e+00 -3.27723294e-01 -4.44756925e-01] | [8.474921226501465, -1.0333870649337769] |
61bb8260-2afd-4234-b227-0f9fb3cfd8d8 | toist-task-oriented-instance-segmentation | 2210.10775 | null | https://arxiv.org/abs/2210.10775v1 | https://arxiv.org/pdf/2210.10775v1.pdf | TOIST: Task Oriented Instance Segmentation Transformer with Noun-Pronoun Distillation | Current referring expression comprehension algorithms can effectively detect or segment objects indicated by nouns, but how to understand verb reference is still under-explored. As such, we study the challenging problem of task oriented detection, which aims to find objects that best afford an action indicated by verbs like sit comfortably on. Towards a finer localization that better serves downstream applications like robot interaction, we extend the problem into task oriented instance segmentation. A unique requirement of this task is to select preferred candidates among possible alternatives. Thus we resort to the transformer architecture which naturally models pair-wise query relationships with attention, leading to the TOIST method. In order to leverage pre-trained noun referring expression comprehension models and the fact that we can access privileged noun ground truth during training, a novel noun-pronoun distillation framework is proposed. Noun prototypes are generated in an unsupervised manner and contextual pronoun features are trained to select prototypes. As such, the network remains noun-agnostic during inference. We evaluate TOIST on the large-scale task oriented dataset COCO-Tasks and achieve +10.9% higher $\rm{mAP^{box}}$ than the best-reported results. The proposed noun-pronoun distillation can boost $\rm{mAP^{box}}$ and $\rm{mAP^{mask}}$ by +2.8% and +3.8%. Codes and models are publicly available at https://github.com/AIR-DISCOVER/TOIST. | ['Ya-Qin Zhang', 'Guyue Zhou', 'Hao Zhao', 'Xiaoxue Chen', 'Yongliang Shi', 'Beiwen Tian', 'Pengfei Li'] | 2022-10-19 | null | null | null | null | ['referring-expression'] | ['computer-vision'] | [ 5.07488549e-01 4.92884755e-01 -2.29424447e-01 -7.11918890e-01
-1.07330692e+00 -6.68125927e-01 5.48601687e-01 1.12929985e-01
-6.84700072e-01 5.05367994e-01 1.23680115e-01 -3.10962290e-01
-1.66043907e-01 -7.81317592e-01 -9.03449535e-01 -4.69155997e-01
1.68867067e-01 8.52345705e-01 -5.93437534e-03 -2.92220235e-01
2.36693740e-01 4.03833002e-01 -1.55136955e+00 3.03495288e-01
9.70014274e-01 1.04153574e+00 3.66449893e-01 2.42135435e-01
-2.48032268e-02 4.02476788e-01 -3.92331898e-01 -5.24917662e-01
3.34280506e-02 -2.41252676e-01 -1.10005665e+00 -8.19599032e-02
3.06445241e-01 -1.71897009e-01 -1.95138250e-02 1.15271175e+00
3.86568069e-01 3.52261871e-01 7.14926541e-01 -1.00498497e+00
-5.61713696e-01 8.73742521e-01 -3.20705980e-01 1.90866649e-01
4.96731937e-01 1.82407796e-01 1.50238645e+00 -1.06019258e+00
6.11463726e-01 1.32454908e+00 7.67843351e-02 6.51800215e-01
-1.35808969e+00 -5.02660811e-01 4.84819293e-01 2.34104604e-01
-1.58705425e+00 -3.48315686e-01 7.01184094e-01 -1.76148951e-01
1.11244750e+00 5.20812571e-01 3.49042892e-01 1.34345531e+00
-3.26900244e-01 1.24261236e+00 9.67542052e-01 -3.89676601e-01
1.55002818e-01 2.71203313e-02 3.11221153e-01 5.60929418e-01
-3.74440998e-02 -2.14920998e-01 -4.96114582e-01 1.45506561e-01
4.88135099e-01 -1.98390797e-01 -3.60073447e-01 -1.12963036e-01
-1.30248678e+00 9.82061267e-01 7.38155067e-01 2.74930030e-01
-4.94608462e-01 3.43254834e-01 1.99152291e-01 -1.30135179e-01
2.39502400e-01 8.45108509e-01 -6.00163698e-01 -2.09344909e-01
-7.13136852e-01 4.47057396e-01 6.43189490e-01 1.35212696e+00
7.40083516e-01 -2.38655642e-01 -3.40655446e-01 1.00163364e+00
2.93389887e-01 4.02007997e-01 5.16710617e-02 -1.05190158e+00
6.80962265e-01 6.92812145e-01 4.52263579e-02 -7.47506857e-01
-6.23218179e-01 -4.96658713e-01 -4.40574944e-01 -3.93055946e-01
5.52774251e-01 3.25915255e-02 -9.35091913e-01 1.89520943e+00
2.73330003e-01 -4.11804497e-01 7.82231614e-03 1.18488348e+00
7.11757839e-01 4.41714227e-01 4.37495321e-01 4.88108583e-02
1.90222251e+00 -8.49040687e-01 -5.30197859e-01 -6.09432042e-01
7.09792614e-01 -4.35346276e-01 1.49907863e+00 3.31254333e-01
-1.05972993e+00 -3.39908153e-01 -7.65776336e-01 -4.42235410e-01
-3.86078119e-01 1.10193364e-01 9.37538326e-01 2.24304542e-01
-6.30958319e-01 3.99409890e-01 -8.11781526e-01 -4.79309678e-01
7.50742912e-01 6.21738732e-01 -2.32299402e-01 -1.84662387e-01
-1.20032859e+00 8.02474499e-01 4.99849319e-01 3.66288364e-01
-8.12556624e-01 -4.51349884e-01 -9.63090658e-01 1.89042613e-01
7.89105773e-01 -6.01596713e-01 1.32917500e+00 -8.10901701e-01
-1.25380588e+00 1.13541043e+00 -4.42507178e-01 -4.81914341e-01
4.27318096e-01 -5.63427866e-01 4.39729579e-02 2.10199624e-01
4.73062098e-01 1.20040548e+00 5.84179163e-01 -1.28155494e+00
-7.95154154e-01 -4.62706834e-01 4.45032150e-01 3.23143512e-01
-4.34929281e-02 1.43138155e-01 -7.15879917e-01 -5.17319918e-01
6.30237460e-01 -1.03897297e+00 -7.24537671e-02 -1.77799612e-01
-7.72342682e-01 -7.16447651e-01 3.11037779e-01 -4.47155327e-01
9.50404763e-01 -2.01561117e+00 4.13047552e-01 2.15411022e-01
2.18297720e-01 -1.54367864e-01 -1.94198173e-02 6.48359135e-02
-2.60775536e-02 2.36001626e-01 -4.19988096e-01 -3.20371330e-01
3.35257173e-01 3.27297181e-01 -3.67367089e-01 3.11869472e-01
4.52118725e-01 1.08748412e+00 -8.71156573e-01 -4.76398587e-01
1.69493444e-02 2.53455639e-01 -7.07676888e-01 2.62401938e-01
-6.27021313e-01 4.65689480e-01 -7.65177965e-01 9.71989751e-01
4.34436023e-01 -1.04523793e-01 2.75343750e-02 -3.19782883e-01
8.66135731e-02 4.91811156e-01 -1.00135839e+00 1.74098551e+00
-3.97583783e-01 3.92786711e-01 9.14889053e-02 -1.16290808e+00
9.09063578e-01 9.24164653e-02 2.31370121e-01 -8.03469718e-01
3.35615486e-01 4.44671720e-01 1.54045179e-01 -5.65295875e-01
3.93408895e-01 -4.96495329e-02 -5.26226997e-01 6.08807914e-02
-2.33568810e-02 -3.91711861e-01 2.12755889e-01 1.47132292e-01
9.56567883e-01 5.32374144e-01 9.69936550e-02 -4.09099013e-01
3.67620111e-01 2.83376992e-01 4.95588899e-01 7.68694699e-01
-2.92108506e-02 4.39270705e-01 7.41372705e-01 -1.93799928e-01
-5.76873779e-01 -1.04697418e+00 -3.26608121e-01 1.62576020e+00
3.01818520e-01 -1.12972893e-01 -9.11450744e-01 -6.67961359e-01
-1.62560925e-01 1.28205466e+00 -5.32955587e-01 4.79225218e-02
-8.93649042e-01 -7.46195674e-01 4.24828976e-01 7.09304333e-01
4.21549141e-01 -1.41945744e+00 -8.82479906e-01 1.49475008e-01
-3.28274459e-01 -1.19255078e+00 -3.27529907e-01 7.57317841e-01
-6.00187778e-01 -8.60033035e-01 -4.65041310e-01 -8.74086440e-01
8.68186414e-01 -1.70148179e-01 1.20405185e+00 -1.94758344e-02
-7.10079595e-02 2.79975086e-01 -4.64773089e-01 -5.12302101e-01
3.40919546e-03 4.46867406e-01 -1.35361239e-01 -1.47576407e-01
7.97107697e-01 -5.46173692e-01 -5.85097492e-01 2.39913762e-01
-5.51662266e-01 2.54118498e-02 5.97416520e-01 9.00028527e-01
8.95362318e-01 -5.02858698e-01 4.54976797e-01 -1.02578139e+00
5.04343152e-01 -3.99847150e-01 -6.54129326e-01 -1.25428773e-02
-4.16636229e-01 1.61044344e-01 4.39091593e-01 -2.79465258e-01
-9.50090706e-01 2.34922647e-01 -1.91925704e-01 -2.73118407e-01
-4.23039764e-01 4.13958281e-01 -5.81513345e-01 6.12634242e-01
5.55855036e-01 1.54917985e-01 -4.95279938e-01 -4.40023869e-01
6.46109879e-01 5.72001159e-01 6.78457558e-01 -9.69437659e-01
5.38129449e-01 3.27615201e-01 -1.97581336e-01 -4.88843322e-01
-1.06042838e+00 -4.33428735e-01 -6.81647897e-01 1.90552458e-01
1.10777092e+00 -8.44858408e-01 -9.84495103e-01 -4.41813171e-02
-1.48172998e+00 -3.33688289e-01 -1.89495280e-01 2.40900025e-01
-7.06916869e-01 -6.28482774e-02 -4.20016170e-01 -7.87957072e-01
-3.68175745e-01 -1.29304695e+00 1.47319615e+00 2.15653196e-01
-5.20935714e-01 -4.87686068e-01 -7.03658581e-01 6.82829320e-01
2.40004376e-01 9.18210968e-02 1.04957891e+00 -1.17022872e+00
-8.99895847e-01 -8.48264247e-02 -5.37158787e-01 -1.47191314e-02
-1.22102186e-01 -5.16607285e-01 -1.01146519e+00 -4.24361192e-02
-8.41102228e-02 -2.42510721e-01 9.28420603e-01 1.10625587e-01
1.42667127e+00 -3.05126101e-01 -5.59521735e-01 6.46497905e-01
9.96605575e-01 2.73655921e-01 4.92183000e-01 3.69412482e-01
8.13212872e-01 9.45423782e-01 9.38437462e-01 1.99273124e-01
5.01759171e-01 8.20010602e-01 7.51555026e-01 9.29778889e-02
1.82770476e-01 -1.40090376e-01 -2.32854206e-02 1.68067664e-01
1.59226637e-02 -4.36878979e-01 -1.10689306e+00 8.60321462e-01
-1.81127346e+00 -7.35252798e-01 -1.40102923e-01 1.74624300e+00
9.57339227e-01 3.37958246e-01 -2.91810602e-01 -6.10210747e-02
5.60865998e-01 -4.51690964e-02 -6.30753458e-01 -3.62401694e-01
-3.88088226e-02 5.27724743e-01 3.16152722e-01 4.92223501e-01
-1.27538872e+00 1.48099148e+00 3.77314711e+00 8.26302350e-01
-9.38640893e-01 1.55428573e-01 6.33005917e-01 -1.45848216e-02
-2.51739800e-01 1.29433796e-01 -9.07797337e-01 1.49932459e-01
6.30052090e-01 1.83057427e-01 4.55706477e-01 1.00429976e+00
2.58503973e-01 -3.14072222e-01 -1.60611260e+00 8.52220416e-01
-9.10755619e-02 -9.24683988e-01 3.94084565e-02 -1.68938175e-01
2.18099087e-01 -9.32051763e-02 1.27388656e-01 5.81469893e-01
7.01370835e-02 -1.23123085e+00 9.66797590e-01 3.29862148e-01
6.84902132e-01 -4.92354602e-01 6.19869649e-01 2.48295426e-01
-1.07272065e+00 -1.87973142e-01 -1.18756302e-01 3.03148516e-02
2.65839219e-01 3.34078431e-01 -9.14981842e-01 3.44361216e-01
7.58247137e-01 4.32012826e-01 -4.03165638e-01 5.93516707e-01
-6.13743365e-01 5.76912642e-01 -5.42328596e-01 -2.54855782e-01
3.59247148e-01 -3.30262221e-02 6.52617812e-01 1.32099092e+00
1.20417126e-01 3.22825462e-01 1.23024657e-01 1.34800494e+00
-2.04625592e-01 2.90740013e-01 -2.32172593e-01 7.17787892e-02
5.77444613e-01 1.12867260e+00 -8.13196838e-01 -1.71076044e-01
-1.29366651e-01 1.17034268e+00 3.92975420e-01 3.99587095e-01
-1.03551579e+00 -5.25823474e-01 4.40797001e-01 1.52029842e-01
4.46517795e-01 -1.38175905e-01 -3.49224925e-01 -8.40510428e-01
2.60867268e-01 -7.88605213e-01 3.62191647e-01 -7.84911394e-01
-1.04713213e+00 6.60788536e-01 3.15233409e-01 -8.57949138e-01
-2.44405136e-01 -8.17299247e-01 -3.67852211e-01 8.83049250e-01
-1.35925817e+00 -1.37025082e+00 -3.49876046e-01 3.10515374e-01
7.76693285e-01 1.37650162e-01 8.67043614e-01 1.62333205e-01
-5.29482782e-01 6.51248455e-01 -4.19886380e-01 2.89027125e-01
4.08385396e-01 -1.34047186e+00 2.03465596e-01 6.25250936e-01
2.85695106e-01 9.29718435e-01 7.88747072e-01 -4.39103007e-01
-1.40140414e+00 -1.07693124e+00 1.09119523e+00 -7.50702202e-01
4.76877749e-01 -5.85347295e-01 -9.28390503e-01 8.64349186e-01
-3.22269909e-02 6.99568465e-02 3.14340740e-01 2.34170645e-01
-4.05014485e-01 4.72249314e-02 -1.07230806e+00 8.57193470e-01
1.39371991e+00 -4.10911322e-01 -7.43988335e-01 6.21391475e-01
9.03105617e-01 -6.07448280e-01 -7.00867772e-01 4.05008167e-01
1.95369884e-01 -6.60384059e-01 7.92461336e-01 -5.65797448e-01
3.07202995e-01 -1.22189328e-01 -5.83191037e-01 -7.88659394e-01
1.27167016e-01 -5.90791821e-01 1.44491598e-01 1.26031375e+00
8.92179787e-01 -5.46103895e-01 7.27263629e-01 8.22078526e-01
-3.39212924e-01 -1.01425576e+00 -1.18499184e+00 -3.93318087e-01
1.11232743e-01 -6.74869895e-01 5.02452850e-01 6.76689804e-01
4.98458594e-02 6.69248104e-01 2.45615140e-01 2.72803426e-01
2.96937019e-01 1.93323463e-01 4.17258143e-01 -1.05872047e+00
-2.47734219e-01 -5.14656246e-01 -3.12536135e-02 -1.53397739e+00
4.49637532e-01 -1.09329259e+00 3.41503769e-01 -1.44852448e+00
1.84164628e-01 -6.77594960e-01 -1.20177425e-01 8.18016410e-01
-2.70528585e-01 1.73007348e-03 2.28298962e-01 9.55829918e-02
-6.73483133e-01 5.70514023e-01 9.76632714e-01 -2.66735256e-01
-2.68304646e-01 2.17237715e-02 -8.65844131e-01 8.10825706e-01
7.76653290e-01 -4.63221639e-01 -2.76770353e-01 -7.25768447e-01
2.02522159e-01 -9.25335139e-02 5.36548257e-01 -5.24075687e-01
1.23437531e-01 -1.43673584e-01 -1.49830272e-02 -4.78651166e-01
6.67273223e-01 -7.68800914e-01 -4.04174119e-01 7.14215562e-02
-5.02806902e-01 3.64976451e-02 1.64783880e-01 3.73481661e-01
-1.27664395e-02 -3.25467914e-01 5.34861326e-01 -2.36667752e-01
-7.35824525e-01 -4.60659526e-02 -1.53911099e-01 1.68735176e-01
8.50038826e-01 -1.28450856e-01 -2.30526850e-01 -2.35360712e-01
-8.85280132e-01 5.59105217e-01 7.16021135e-02 4.64754850e-01
4.91160363e-01 -8.91937852e-01 -5.62138855e-01 -1.95162088e-01
3.12011689e-01 6.19453073e-01 -7.87465423e-02 8.75669301e-01
-1.82301596e-01 5.85298896e-01 1.90748498e-01 -6.94441855e-01
-8.70054960e-01 4.79171544e-01 4.94403958e-01 -5.59356471e-04
-4.37317282e-01 1.28840864e+00 2.63317794e-01 -4.64038521e-01
3.03250939e-01 -5.46981394e-01 -1.97300225e-01 1.12235248e-01
7.50686154e-02 1.14865461e-02 -3.88757437e-02 -7.80169427e-01
-5.25307834e-01 3.94121319e-01 -1.79133028e-01 -6.89882636e-02
1.27542710e+00 2.04216763e-02 -1.98908433e-01 9.01095420e-02
1.08562613e+00 -1.90848768e-01 -1.08512199e+00 -1.23796798e-01
4.73594874e-01 -1.01811588e-01 -2.22678334e-01 -8.80833924e-01
-8.03514838e-01 9.28332388e-01 3.98017675e-01 -1.88897923e-02
1.01793289e+00 5.70349216e-01 5.33630848e-01 6.66328907e-01
3.58149022e-01 -9.58497286e-01 4.48579714e-02 5.63841462e-01
1.03377712e+00 -1.34480858e+00 -1.87738404e-01 -7.19968796e-01
-6.02951527e-01 8.01837564e-01 9.41176951e-01 3.15589597e-03
1.57053441e-01 -9.79255140e-02 -4.21990007e-02 -5.20641923e-01
-5.20433664e-01 -5.10944247e-01 2.31026098e-01 4.26486254e-01
5.71642756e-01 1.44956574e-01 -5.03574789e-01 1.07576835e+00
-4.70945776e-01 -4.93856251e-01 1.76943168e-01 8.27989697e-01
-3.24732065e-01 -9.51606929e-01 -2.85332680e-01 5.41580439e-01
-4.08492059e-01 -3.41875106e-01 -2.71038294e-01 8.51539552e-01
1.88923329e-01 9.47897315e-01 1.12764284e-01 1.74108148e-02
7.19113350e-01 2.69439936e-01 2.98922509e-01 -8.26496422e-01
-5.61540782e-01 4.14369628e-02 3.38025212e-01 -6.51448488e-01
-3.84953916e-01 -7.92236924e-01 -1.81119668e+00 3.37043673e-01
-3.72317195e-01 8.47849697e-02 4.65489596e-01 1.01930249e+00
2.34547183e-01 6.11470938e-01 1.30843669e-01 -8.83754253e-01
-6.35650456e-01 -8.91604543e-01 -8.59191343e-02 4.99790430e-01
-3.26960795e-02 -7.72290647e-01 -2.28552282e-01 -4.07755971e-02] | [10.379800796508789, 1.3293808698654175] |
3637919b-d153-46f0-acc0-1b11e998be4d | fusionnet-3d-object-classification-using | 1607.05695 | null | http://arxiv.org/abs/1607.05695v4 | http://arxiv.org/pdf/1607.05695v4.pdf | FusionNet: 3D Object Classification Using Multiple Data Representations | High-quality 3D object recognition is an important component of many vision
and robotics systems. We tackle the object recognition problem using two data
representations, to achieve leading results on the Princeton ModelNet
challenge. The two representations: 1. Volumetric representation: the 3D object
is discretized spatially as binary voxels - $1$ if the voxel is occupied and
$0$ otherwise. 2. Pixel representation: the 3D object is represented as a set
of projected 2D pixel images. Current leading submissions to the ModelNet
Challenge use Convolutional Neural Networks (CNNs) on pixel representations.
However, we diverge from this trend and additionally, use Volumetric CNNs to
bridge the gap between the efficiency of the above two representations. We
combine both representations and exploit them to learn new features, which
yield a significantly better classifier than using either of the
representations in isolation. To do this, we introduce new Volumetric CNN
(V-CNN) architectures. | ['Vishakh Hegde', 'Reza Zadeh'] | 2016-07-19 | null | null | null | null | ['3d-object-classification', '3d-object-recognition'] | ['computer-vision', 'computer-vision'] | [ 3.32236469e-01 1.67898148e-01 -6.79393560e-02 -2.74727136e-01
-3.88279170e-01 -4.70793545e-01 8.68157685e-01 -6.11623898e-02
-3.06427777e-01 3.95817459e-01 5.25836349e-02 -4.49144602e-01
1.52473360e-01 -9.64797497e-01 -8.70293260e-01 -5.19470632e-01
-2.17934698e-01 3.39787275e-01 2.74270982e-01 1.68013766e-01
3.77357513e-01 1.10926104e+00 -1.58077896e+00 3.05851012e-01
6.98005781e-02 1.68705308e+00 -1.57143638e-01 6.03264213e-01
-5.06962299e-01 7.10833907e-01 -3.38691175e-01 2.96153486e-01
7.43719339e-01 3.86128165e-02 -9.04864371e-01 2.18562454e-01
8.02904069e-01 -2.34239087e-01 -5.80074370e-01 9.50207591e-01
1.55082539e-01 8.79057497e-03 9.63242769e-01 -1.18704164e+00
-6.56679809e-01 2.06618413e-01 -7.22148597e-01 3.16455543e-01
1.19145572e-01 3.14009011e-01 9.73973811e-01 -1.17836761e+00
7.65724599e-01 1.40642595e+00 9.08361793e-01 3.77313375e-01
-1.34122455e+00 -6.98868185e-02 2.42231011e-01 5.17856926e-02
-1.31549907e+00 -2.12794229e-01 7.68070281e-01 -6.83309913e-01
1.63048458e+00 8.88194516e-02 1.08324289e+00 7.78573275e-01
2.49295875e-01 8.48616362e-01 1.06611216e+00 -2.16747165e-01
3.26641172e-01 -2.14579746e-01 2.66323358e-01 8.72311056e-01
2.20135286e-01 2.19509527e-01 -2.67612487e-01 -2.80949194e-02
1.18663049e+00 1.75049722e-01 -2.06112534e-01 -8.50868642e-01
-1.24829388e+00 7.38844454e-01 1.04198825e+00 1.80368274e-01
-3.82071197e-01 6.91513777e-01 2.91110843e-01 1.66702405e-01
4.26932812e-01 3.87240410e-01 -2.13388428e-01 5.68116903e-02
-9.88080204e-01 4.88611579e-01 6.02716625e-01 8.57717037e-01
9.08198535e-01 2.69946933e-01 -7.53255635e-02 7.48854935e-01
4.18099284e-01 2.02179402e-01 2.42402196e-01 -1.05476093e+00
2.24432945e-01 9.18187141e-01 -8.41604024e-02 -7.28619277e-01
-4.79207039e-01 -2.68979818e-01 -8.69456053e-01 8.02019656e-01
2.17564821e-01 3.58927399e-01 -1.57048523e+00 1.25205529e+00
2.47394159e-01 2.77157605e-01 -4.12541553e-02 8.70330095e-01
1.23442674e+00 6.69409573e-01 -1.82502180e-01 3.93391758e-01
1.11693394e+00 -9.33978021e-01 -5.65029941e-02 -1.34751052e-01
5.71297765e-01 -4.61270809e-01 6.41544938e-01 1.78191170e-01
-1.21163857e+00 -6.23647273e-01 -1.43071103e+00 -3.36933076e-01
-8.13093960e-01 -2.24238485e-01 6.93835557e-01 5.39810121e-01
-1.29633868e+00 6.58886075e-01 -8.07593763e-01 -1.75657123e-01
9.46812391e-01 4.64261323e-01 -4.37327802e-01 -3.77416611e-01
-7.22562253e-01 1.03410757e+00 6.91246390e-02 -5.14103919e-02
-8.83232832e-01 -8.54709864e-01 -1.06781006e+00 -1.03468060e-01
-2.70607144e-01 -6.89480841e-01 1.23579943e+00 -6.78430617e-01
-1.07473576e+00 1.33379662e+00 6.05590753e-02 -6.19609952e-01
5.69245458e-01 1.93728939e-01 2.81235337e-01 -4.24269810e-02
-1.83005944e-01 1.10967278e+00 7.40225971e-01 -1.35251415e+00
-4.69320863e-01 -4.98667330e-01 3.07601765e-02 1.58647120e-01
2.50354093e-02 -3.70661080e-01 -4.19913650e-01 -3.62765104e-01
5.07905662e-01 -7.78649867e-01 -4.17582750e-01 6.52579665e-01
-4.61381793e-01 -3.80757987e-01 1.19319999e+00 -1.46007523e-01
2.60769665e-01 -2.21549320e+00 1.14095457e-01 1.92911103e-01
6.26508951e-01 2.45627552e-01 -1.88057348e-01 -1.63701236e-01
-4.35181893e-03 2.26095095e-01 -4.72087473e-01 -4.65821832e-01
7.81084225e-02 3.03743958e-01 -3.31874222e-01 5.72142661e-01
6.10473096e-01 1.15808797e+00 -6.07842684e-01 -3.47698241e-01
4.94129151e-01 7.76007771e-01 -4.39383417e-01 -3.45986560e-02
-2.65399754e-01 1.09903127e-01 -3.91512424e-01 8.56970549e-01
9.21088636e-01 -1.64863423e-01 -1.60997629e-01 -2.90620357e-01
-2.28984535e-01 3.63318682e-01 -1.20162606e+00 1.55939412e+00
-1.30089015e-01 8.71262133e-01 2.25189447e-01 -1.40666723e+00
1.05890250e+00 8.55289325e-02 7.84450233e-01 -7.26225972e-01
2.16402650e-01 2.34777138e-01 -2.91845053e-01 -4.25364897e-02
5.31744421e-01 -7.91052580e-02 -4.69196104e-02 2.34222621e-01
-7.68275466e-04 -8.01317334e-01 -1.72366679e-01 -6.13576807e-02
1.25306535e+00 2.90852100e-01 1.76735535e-01 -3.25692564e-01
9.07069147e-02 3.12319100e-01 3.21715206e-01 8.51036251e-01
-5.18692076e-01 8.81119668e-01 5.12801290e-01 -9.05367076e-01
-1.01013696e+00 -1.31153035e+00 -4.07918930e-01 3.72788757e-01
1.40893832e-01 -7.20772892e-02 -4.04030889e-01 -6.42313063e-01
5.62876821e-01 1.84275895e-01 -7.66078711e-01 -7.34081045e-02
-6.44480944e-01 -3.65507096e-01 3.66918325e-01 9.69554305e-01
4.89949375e-01 -1.03181851e+00 -1.06359100e+00 9.83686969e-02
4.53476995e-01 -1.06729770e+00 6.29619434e-02 7.02483952e-01
-9.72131491e-01 -1.13095093e+00 -8.09295356e-01 -9.82717216e-01
5.15063107e-01 3.71029884e-01 1.20839095e+00 7.41850063e-02
-8.08980703e-01 6.71098411e-01 -1.35865316e-01 -6.00038230e-01
8.74333680e-02 -4.37183976e-02 1.88037826e-04 -3.68486345e-01
4.61795449e-01 -5.81793845e-01 -5.03593504e-01 -1.32776514e-01
-6.70464933e-01 5.19591533e-02 4.67018247e-01 6.75544679e-01
9.79969561e-01 -2.42738709e-01 1.06236383e-01 -6.52951658e-01
3.41613919e-01 -3.58905822e-01 -6.11837804e-01 -1.83057457e-01
-1.97524458e-01 -5.22891358e-02 3.10283601e-01 -3.32830846e-01
-3.08979571e-01 4.27313507e-01 -2.45574325e-01 -8.99071455e-01
-3.18082005e-01 4.58290935e-01 2.01334357e-01 -4.31766212e-01
5.47592819e-01 1.26166880e-01 -2.55805440e-02 -6.00015402e-01
5.37028551e-01 4.79457945e-01 5.78723550e-01 -5.22381186e-01
5.62441528e-01 6.09974504e-01 3.97493303e-01 -9.40273762e-01
-4.97579008e-01 -3.57551664e-01 -8.93811584e-01 -1.78706363e-01
9.17655170e-01 -7.09741473e-01 -6.19351447e-01 4.01277691e-01
-1.31466579e+00 -5.65682352e-01 -8.35023999e-01 2.95013577e-01
-7.85131276e-01 9.92097929e-02 -5.53694367e-01 -6.86645806e-01
-1.11114904e-01 -1.42692482e+00 1.10767090e+00 6.87520131e-02
1.82115112e-03 -7.85280824e-01 -1.85621413e-03 -4.44565713e-02
4.93758917e-01 5.34892440e-01 1.08750200e+00 -3.69669348e-01
-7.98575103e-01 -4.56393778e-01 -7.37999320e-01 2.99753785e-01
-1.54086798e-01 -2.70896987e-03 -1.03317845e+00 -2.32114062e-01
-1.71081141e-01 -5.81522822e-01 1.26058400e+00 5.58940768e-01
1.48680997e+00 1.31109893e-01 -4.11047429e-01 8.05986345e-01
1.55523098e+00 1.67139024e-02 6.20794654e-01 1.84828818e-01
6.12572014e-01 3.96319032e-01 -1.52852029e-01 2.82564044e-01
3.03098440e-01 5.94945133e-01 9.12891507e-01 -2.53253222e-01
-5.23179293e-01 -2.64516706e-03 1.59516428e-02 3.51662904e-01
-1.13178372e-01 1.43025860e-01 -1.20189631e+00 6.77821577e-01
-1.53969836e+00 -7.00547934e-01 -1.75921008e-01 1.92159534e+00
4.11085308e-01 3.55594486e-01 -8.54555145e-02 2.84125566e-01
2.98872441e-01 4.24256861e-01 -8.13421011e-01 -5.85257649e-01
-1.22226484e-01 5.58251143e-01 4.81700003e-01 2.77230710e-01
-1.25577319e+00 8.01502228e-01 7.69051123e+00 2.82746822e-01
-1.35285652e+00 -1.12605765e-01 6.27803981e-01 -2.48465121e-01
-8.95405114e-02 -3.01439643e-01 -9.04589534e-01 -4.73849066e-02
5.58719873e-01 2.37283573e-01 2.01982111e-01 1.09063113e+00
-3.83466274e-01 1.69209149e-02 -1.39641321e+00 1.19806492e+00
1.36190370e-01 -1.86345243e+00 2.53183812e-01 3.61243933e-01
7.35132992e-01 5.05403578e-01 3.52492958e-01 3.79986346e-01
5.04930675e-01 -1.57889128e+00 9.43857491e-01 4.92470801e-01
8.18728089e-01 -3.63548130e-01 4.25980419e-01 1.92741677e-01
-1.40433145e+00 6.07577898e-02 -6.30677879e-01 -2.22073987e-01
-4.11659334e-04 6.33945584e-01 -6.33675158e-01 1.57716885e-01
8.97379100e-01 8.93650055e-01 -2.59626478e-01 1.17543924e+00
3.25584203e-01 1.22139126e-01 -5.20860553e-01 1.14845417e-01
5.86399198e-01 2.26164376e-03 4.89859819e-01 1.16129041e+00
2.17867732e-01 1.33251296e-02 3.44828695e-01 1.34910250e+00
-4.52460557e-01 -2.82596737e-01 -1.03200293e+00 6.36294335e-02
2.00007409e-01 1.02563632e+00 -7.20935464e-01 -3.82288128e-01
-6.27098143e-01 7.74354815e-01 6.20256484e-01 2.35581934e-01
-3.80498648e-01 -3.64783645e-01 9.55453873e-01 9.86192562e-03
6.65918350e-01 -6.08210385e-01 -7.15195596e-01 -7.67536879e-01
-7.87116811e-02 -2.61991143e-01 3.77946496e-02 -8.51024866e-01
-1.38998687e+00 4.42920864e-01 -1.66356176e-01 -1.03575885e+00
5.31951673e-02 -1.25993669e+00 -5.26434600e-01 9.44634259e-01
-1.88290417e+00 -1.14407361e+00 -4.60056454e-01 5.90575695e-01
3.83180439e-01 1.01042129e-01 8.04700255e-01 8.48351717e-02
-2.24637270e-01 2.38125607e-01 -7.99698010e-02 2.18782336e-01
8.37940425e-02 -1.31616271e+00 7.22864926e-01 2.91704684e-01
3.12374055e-01 2.66699046e-01 2.01858599e-02 -3.66881192e-01
-1.48926651e+00 -1.27199984e+00 7.33755350e-01 -5.77895939e-01
2.20976889e-01 -4.03410614e-01 -7.07412064e-01 8.31022799e-01
-1.03479676e-01 7.00445652e-01 3.87408853e-01 -7.84762874e-02
-7.59469330e-01 3.44149083e-01 -1.25352037e+00 3.51137519e-01
1.16149890e+00 -5.26690960e-01 -6.66871786e-01 2.75627643e-01
6.38860881e-01 -4.93462592e-01 -9.73883152e-01 5.65428853e-01
4.72001731e-01 -9.45686162e-01 1.42281067e+00 -7.02427924e-01
5.59545696e-01 -1.56780764e-01 -5.97781897e-01 -1.23190546e+00
-3.31799567e-01 1.44240886e-01 -3.11287135e-01 4.84619051e-01
2.25553095e-01 -5.04152238e-01 1.05059206e+00 3.82402301e-01
-4.73017216e-01 -1.19363213e+00 -1.31224811e+00 -8.18061650e-01
5.23946404e-01 -6.79681063e-01 6.36936724e-01 7.85845220e-01
-2.28730217e-01 7.10771680e-02 1.39475852e-01 -1.29332840e-01
6.78053439e-01 2.62427002e-01 5.57158411e-01 -1.39011025e+00
9.16205347e-02 -1.02798450e+00 -1.02726388e+00 -1.46499538e+00
-1.04586624e-01 -1.15529656e+00 1.43193483e-01 -1.89180052e+00
9.17249396e-02 -5.63524187e-01 -4.00985956e-01 6.71097457e-01
4.35948700e-01 7.87597477e-01 2.35976532e-01 2.60125220e-01
-3.30271453e-01 4.78919983e-01 1.19814420e+00 -5.71924508e-01
-1.25118434e-01 -2.81375349e-01 -2.83641398e-01 6.09577179e-01
4.43365037e-01 2.91420612e-02 -2.30242405e-03 -5.96783817e-01
-1.83720902e-01 -2.27483302e-01 6.80040419e-01 -1.13771820e+00
6.18947148e-02 -8.04111138e-02 7.51410842e-01 -8.90290201e-01
7.37088084e-01 -7.44759679e-01 -2.29081690e-01 4.61913705e-01
-1.64948612e-01 -1.92567095e-01 3.66254866e-01 4.95695233e-01
-7.06839710e-02 2.69791465e-02 1.03190005e+00 -6.14187300e-01
-8.25072169e-01 5.82150638e-01 -4.72031683e-01 -1.87596679e-01
1.03851390e+00 -7.57307291e-01 -1.92285493e-01 6.18870333e-02
-8.57752383e-01 -6.97604492e-02 4.60016310e-01 3.95176530e-01
8.25025022e-01 -1.48504829e+00 -4.16259527e-01 4.86024231e-01
1.32253230e-01 4.54433948e-01 -2.60724090e-02 6.64819598e-01
-7.93518305e-01 6.16974056e-01 -3.56828868e-01 -9.92158294e-01
-8.50507319e-01 3.99841279e-01 6.49460256e-01 -9.96768251e-02
-9.06452477e-01 1.09780991e+00 3.41823399e-02 -5.82507849e-01
5.09880304e-01 -6.39646947e-01 -5.72709888e-02 -6.86973333e-02
3.68646234e-01 5.80773838e-02 1.04369208e-01 -6.18455648e-01
-3.82192314e-01 7.62476206e-01 2.00805850e-02 3.12752500e-02
1.58672988e+00 3.44150156e-01 1.08899148e-02 5.46610892e-01
1.46952987e+00 -6.93640351e-01 -1.48402548e+00 -3.99564117e-01
-1.32125273e-01 -4.90935713e-01 4.06331956e-01 -4.50262398e-01
-1.14765918e+00 1.28869772e+00 6.85802817e-01 3.72880518e-01
5.63807726e-01 8.92445073e-02 6.83239162e-01 4.46348637e-01
2.82296568e-01 -8.92554641e-01 1.70671985e-01 9.07709181e-01
9.45845187e-01 -1.12562275e+00 7.50292391e-02 -3.38194817e-01
-1.63923964e-01 1.21380484e+00 5.65716445e-01 -4.94710922e-01
1.03434825e+00 2.20054269e-01 -6.71779290e-02 -6.37314677e-01
-4.16258633e-01 -3.11915040e-01 2.47868255e-01 7.62840331e-01
2.85731256e-01 8.31855237e-02 3.77724946e-01 2.85419554e-01
-1.00299522e-01 -1.21734202e-01 2.22685352e-01 1.24268746e+00
-5.10177851e-01 -6.66785538e-01 -3.25927317e-01 7.36253500e-01
1.31122530e-01 6.79689040e-03 -6.11771047e-01 7.14995682e-01
1.22240037e-01 4.11191314e-01 6.28628135e-01 -3.55316639e-01
4.80747372e-01 1.76169932e-01 7.23839521e-01 -7.22276151e-01
-4.87613887e-01 -2.01699317e-01 -3.16991806e-01 -7.71761298e-01
-3.54028136e-01 -4.67023611e-01 -1.26390743e+00 -2.43219391e-01
8.74831676e-02 -4.49497849e-01 9.91873682e-01 7.30521202e-01
3.59537750e-01 4.69396770e-01 4.98006672e-01 -1.60212326e+00
-5.75257540e-01 -7.11184680e-01 -6.72648907e-01 2.36851528e-01
4.51082170e-01 -9.21931267e-01 -3.94701600e-01 -1.46559961e-02] | [8.132647514343262, -3.6849288940429688] |
3c3d88d3-55ba-42d6-970b-c4d375b20dfa | predictive-modeling-of-hospital-readmission | 2106.08488 | null | https://arxiv.org/abs/2106.08488v1 | https://arxiv.org/pdf/2106.08488v1.pdf | Predictive Modeling of Hospital Readmission: Challenges and Solutions | Hospital readmission prediction is a study to learn models from historical medical data to predict probability of a patient returning to hospital in a certain period, 30 or 90 days, after the discharge. The motivation is to help health providers deliver better treatment and post-discharge strategies, lower the hospital readmission rate, and eventually reduce the medical costs. Due to inherent complexity of diseases and healthcare ecosystems, modeling hospital readmission is facing many challenges. By now, a variety of methods have been developed, but existing literature fails to deliver a complete picture to answer some fundamental questions, such as what are the main challenges and solutions in modeling hospital readmission; what are typical features/models used for readmission prediction; how to achieve meaningful and transparent predictions for decision making; and what are possible conflicts when deploying predictive approaches for real-world usages. In this paper, we systematically review computational models for hospital readmission prediction, and propose a taxonomy of challenges featuring four main categories: (1) data variety and complexity; (2) data imbalance, locality and privacy; (3) model interpretability; and (4) model implementation. The review summarizes methods in each category, and highlights technical solutions proposed to address the challenges. In addition, a review of datasets and resources available for hospital readmission modeling also provides firsthand materials to support researchers and practitioners to design new approaches for effective and efficient hospital readmission prediction. | ['Xingquan Zhu', 'Shuwen Wang'] | 2021-06-16 | null | null | null | null | ['readmission-prediction'] | ['medical'] | [ 1.38671309e-01 -1.00118622e-01 -7.58720517e-01 -3.47823411e-01
-5.41370273e-01 -3.43105011e-02 -5.53409420e-02 1.01914179e+00
-1.76413089e-01 7.61746645e-01 6.24050021e-01 -6.49260938e-01
-8.18535924e-01 -7.09003747e-01 -1.55273721e-01 -6.11284256e-01
-2.33783513e-01 8.40863109e-01 -5.29029012e-01 1.87512130e-01
2.13366002e-01 6.75446451e-01 -1.43458343e+00 6.61580145e-01
8.59611869e-01 9.19700205e-01 8.12815353e-02 5.86105049e-01
1.48343474e-01 9.53389466e-01 -3.74249816e-01 -1.19376816e-01
2.75314182e-01 -6.41183019e-01 -9.67635870e-01 -2.62200683e-01
-4.14506197e-01 -5.75740576e-01 -1.80136889e-01 3.97556722e-01
6.19353950e-01 -3.03432822e-01 7.16162860e-01 -1.12323654e+00
-4.49633002e-01 4.59778041e-01 1.50601670e-01 3.78714561e-01
2.57721424e-01 -1.75500847e-02 5.39328098e-01 -6.80530608e-01
9.58197415e-02 6.71191275e-01 8.47003162e-01 7.77608395e-01
-9.48803127e-01 -4.61408377e-01 -9.97111499e-02 1.67674839e-01
-1.35132110e+00 -4.42253321e-01 1.15486279e-01 -7.04768062e-01
1.05610716e+00 9.26185727e-01 1.17950666e+00 8.76043260e-01
6.12104833e-01 4.87733543e-01 5.73486209e-01 -2.34003946e-01
1.48414612e-01 1.76451504e-01 3.54855895e-01 3.65973502e-01
5.89639425e-01 2.15693861e-01 -3.87261540e-01 -9.10608947e-01
6.28030062e-01 1.13723171e+00 -5.17826438e-01 -3.24375868e-01
-1.46465659e+00 8.08954239e-01 1.45519361e-01 1.18975960e-01
-7.68698275e-01 -3.46602648e-01 5.76131165e-01 6.10149875e-02
2.15824232e-01 5.03740311e-01 -8.40764761e-01 -2.05375239e-01
-9.06951904e-01 2.35755444e-01 8.03172767e-01 8.31752241e-01
2.48785213e-01 -3.67033124e-01 -3.50455821e-01 6.92516148e-01
4.35792617e-02 4.48628068e-01 6.35844886e-01 -5.92113435e-01
5.29771090e-01 6.20895386e-01 1.72907263e-01 -8.39852035e-01
-7.62856305e-01 -4.76774722e-01 -1.36340702e+00 -5.08334577e-01
-5.63097000e-02 -8.47971588e-02 -5.38512230e-01 1.01514161e+00
-1.10799059e-01 2.06561744e-01 3.21611941e-01 5.74878871e-01
9.13630068e-01 5.56598246e-01 8.34067445e-03 -6.74993217e-01
1.33226621e+00 -6.52517200e-01 -6.98151827e-01 -2.53835857e-01
1.11580348e+00 -3.81739080e-01 6.50602579e-01 9.71077234e-02
-1.15052783e+00 2.55148765e-02 -4.27015722e-01 2.56437063e-01
-1.69562683e-01 -7.57320300e-02 3.66987318e-01 4.33937758e-01
-5.66670835e-01 5.01789570e-01 -1.09039652e+00 -6.80535972e-01
4.94277894e-01 5.85035920e-01 2.86269281e-02 -1.71841666e-01
-9.78096306e-01 9.47155058e-01 2.23578095e-01 -9.54263210e-02
-5.06811023e-01 -1.13203609e+00 -6.48893535e-01 4.20821875e-01
-2.90826887e-01 -1.52997601e+00 8.83089066e-01 -4.37376469e-01
-6.95712149e-01 9.70957160e-01 -4.26272571e-01 -5.63258529e-01
4.31919247e-01 -1.73040420e-01 -2.68009245e-01 -2.31381431e-01
-1.81681186e-01 9.35485363e-02 -1.77765545e-02 -1.07891440e+00
-7.25686610e-01 -4.88886297e-01 -5.82190275e-01 2.58044392e-01
-4.05444771e-01 8.28564242e-02 4.55822349e-02 -5.41345716e-01
-1.17099471e-02 -7.65483737e-01 -6.64152145e-01 -2.62823015e-01
-3.61074030e-01 2.75219232e-01 4.05276000e-01 -7.07186818e-01
1.98016310e+00 -1.81043017e+00 -1.83197618e-01 2.21718505e-01
3.62991154e-01 3.66599023e-01 1.86581329e-01 8.56955588e-01
-2.47134700e-01 3.29527050e-01 -2.46405080e-01 -2.38643482e-01
-6.15874708e-01 2.72893906e-01 -4.22697783e-01 2.70383477e-01
-1.72072262e-01 9.50296342e-01 -8.15778375e-01 -2.33946592e-01
5.44930696e-01 4.26256537e-01 -6.82192504e-01 5.58010757e-01
6.46446943e-01 6.20445549e-01 -5.06052732e-01 6.82979882e-01
3.71721715e-01 -5.54085851e-01 1.43580884e-01 2.11443201e-01
6.68871179e-02 4.92821485e-01 -5.74305236e-01 8.22075307e-01
-1.99212685e-01 9.23044309e-02 -3.47315550e-01 -1.01378357e+00
9.67299998e-01 4.68833506e-01 1.00797510e+00 -3.19180846e-01
5.35189062e-02 2.78758198e-01 -1.61867812e-01 -8.59477162e-01
5.67466691e-02 -5.99051192e-02 -1.13607636e-02 3.36390167e-01
-8.74807954e-01 3.86783212e-01 -4.58000064e-01 -2.08758205e-01
1.13697648e+00 -7.38312602e-01 9.62801218e-01 -3.83372784e-01
1.78759798e-01 1.75212249e-01 8.27741444e-01 8.21732998e-01
-3.60785961e-01 8.26015472e-01 2.28868008e-01 -1.37434113e+00
-8.60268593e-01 -1.05381799e+00 -3.29438180e-01 3.45644951e-01
4.68280874e-02 -4.05390292e-01 -4.97110188e-01 -2.87133157e-01
1.28516853e-01 5.77971935e-01 -4.83016580e-01 -4.72182631e-01
-5.32810032e-01 -1.49513650e+00 1.97637811e-01 5.72986901e-01
1.22670211e-01 -8.17072988e-01 -1.02721739e+00 3.91004980e-01
-6.73719168e-01 -5.66296160e-01 -2.16360837e-01 2.93347657e-01
-1.37383974e+00 -1.23396134e+00 -5.12949228e-01 -7.89697111e-01
7.63817906e-01 2.75916278e-01 1.37945819e+00 6.00958943e-01
-5.99617839e-01 2.95747727e-01 -1.24776892e-01 -8.23450983e-01
-4.30554032e-01 1.28486857e-01 1.59383804e-01 -3.81995618e-01
6.88730419e-01 -3.10687274e-01 -1.05987310e+00 8.03950503e-02
-7.63363361e-01 1.19274773e-01 4.40426737e-01 1.12153733e+00
8.68644834e-01 -1.79667816e-01 9.52125549e-01 -9.17962849e-01
8.92239630e-01 -8.95897925e-01 -1.07771136e-01 4.05993760e-01
-1.21216619e+00 -5.03123641e-01 6.09390020e-01 1.70717910e-01
-3.45676810e-01 1.94627151e-01 -5.09068184e-02 -1.02772996e-01
-1.82681173e-01 5.50824881e-01 2.52902657e-01 7.22546697e-01
4.79891062e-01 4.63543147e-01 1.51113585e-01 -4.41289961e-01
-5.78969836e-01 9.81379569e-01 1.24801688e-01 -3.66256610e-02
1.18466347e-01 4.26379412e-01 1.23347297e-01 -4.26734746e-01
-8.77698123e-01 -8.68869722e-01 -6.64489508e-01 1.24319814e-01
1.00388086e+00 -9.65705097e-01 -7.54681230e-01 2.45534822e-01
-9.09546137e-01 -1.94660857e-01 -4.01470274e-01 6.69714093e-01
-6.07712567e-01 1.23476297e-01 -7.10531294e-01 -7.72910416e-01
-8.18948030e-01 -1.12833977e+00 7.00746477e-01 -2.18126506e-01
-7.01913297e-01 -1.24161065e+00 5.76393083e-02 5.52147567e-01
4.30797726e-01 4.06713635e-01 1.41309607e+00 -6.72915041e-01
-4.03946131e-01 -3.80739033e-01 -9.74505469e-02 1.60999730e-01
5.96973717e-01 -1.63780674e-01 -7.55838335e-01 -4.95824575e-01
3.98460589e-02 1.79420695e-01 4.60799634e-01 8.44211638e-01
1.69514751e+00 -7.03772485e-01 -8.31403255e-01 7.94093490e-01
1.10365474e+00 4.35309350e-01 7.59267032e-01 2.81310648e-01
3.50158840e-01 5.92412591e-01 7.05008030e-01 1.07555425e+00
8.25354993e-01 2.13953748e-01 6.81149960e-01 -3.79700094e-01
4.54129994e-01 -9.56188217e-02 -2.43937880e-01 1.14914870e+00
-2.06740126e-01 -2.98972756e-01 -1.37149715e+00 6.04823232e-01
-2.15256453e+00 -1.01432741e+00 -4.92302150e-01 2.59267807e+00
4.42068517e-01 -4.15033370e-01 -5.60423136e-02 1.02491073e-01
3.31769615e-01 -3.16524029e-01 -4.89958972e-01 -4.27005380e-01
8.21842626e-02 -2.20004931e-01 5.24724305e-01 1.05526008e-01
-9.04070973e-01 1.41959772e-01 7.69806004e+00 -1.39226839e-01
-1.10083389e+00 -1.90249085e-01 1.23253810e+00 -2.18374059e-01
-1.13212533e-01 -4.72337492e-02 -6.26602590e-01 5.13459206e-01
1.06593704e+00 -2.28515252e-01 2.95155317e-01 7.87407637e-01
7.01479197e-01 2.07763344e-01 -1.42902064e+00 1.29422939e+00
4.39401157e-02 -1.81498718e+00 2.46736899e-01 2.68644035e-01
6.15790367e-01 1.51085168e-01 -1.34178087e-01 -5.01278080e-02
-7.28772283e-02 -1.34563005e+00 5.02371117e-02 9.45883214e-01
6.09962761e-01 -7.51414597e-01 1.22276998e+00 6.26200378e-01
-8.70250523e-01 -5.07197976e-01 -3.47280502e-01 -3.99720669e-01
8.96957293e-02 7.04201698e-01 -1.04266763e+00 4.73802388e-01
1.07981956e+00 8.97527337e-01 -2.45450407e-01 1.37120998e+00
6.39724970e-01 4.18345451e-01 9.40986946e-02 2.19570801e-01
-3.14166874e-01 -6.62938505e-02 1.09795801e-01 1.12765980e+00
7.15688229e-01 3.06742936e-01 2.40119398e-01 2.36493856e-01
1.05856523e-01 2.40924135e-01 -7.04991579e-01 3.20325553e-01
6.93753421e-01 4.56171602e-01 -2.51000524e-01 -3.63006681e-01
-3.90475422e-01 7.11790383e-01 1.24004804e-01 2.15344429e-01
-5.28768182e-01 6.10872582e-02 1.05129945e+00 7.71180689e-01
-3.72308433e-01 2.61954308e-01 -8.15353274e-01 -1.17714632e+00
-2.18215346e-01 -7.64092684e-01 9.65933263e-01 -5.12431324e-01
-1.22520959e+00 5.17960370e-01 -8.81206803e-03 -1.51767385e+00
-3.50557566e-01 -2.97374517e-01 -4.87072110e-01 8.28921139e-01
-1.58476675e+00 -5.22517562e-01 -5.05617797e-01 4.69266921e-01
3.15591455e-01 -1.98091164e-01 1.54061306e+00 5.02469063e-01
-6.94437742e-01 3.28157723e-01 4.59921032e-01 -6.65681884e-02
6.01785064e-01 -6.49340332e-01 -2.24319965e-01 2.23805636e-01
-7.76526928e-01 8.03291082e-01 3.67486000e-01 -5.62790751e-01
-1.23358452e+00 -1.58037174e+00 1.51332057e+00 -7.49537945e-01
-1.06556341e-01 2.61727087e-02 -8.88636708e-01 9.22828376e-01
-2.94785768e-01 -6.78427294e-02 1.29989088e+00 -3.10997050e-02
2.96715915e-01 -3.66984725e-01 -1.25816894e+00 5.50987542e-01
7.75565803e-01 -6.44321321e-03 -3.19586366e-01 4.96852994e-01
2.40647301e-01 -2.37075746e-01 -1.19630373e+00 6.34360015e-01
5.98016143e-01 -1.08150995e+00 1.10192633e+00 -8.61062407e-01
3.03745389e-01 1.92899078e-01 3.76169980e-02 -1.00647902e+00
-6.56067073e-01 -2.87175298e-01 -9.90456436e-03 7.51971304e-01
3.35866600e-01 -9.88587737e-01 7.43343055e-01 1.07311916e+00
-1.44536078e-01 -1.39463401e+00 -8.99226308e-01 -3.59253466e-01
2.04698816e-02 -2.22448241e-02 1.11444569e+00 1.20603490e+00
2.03080282e-01 -1.40203640e-01 -6.16417229e-01 1.69308946e-01
3.81654084e-01 2.42403194e-01 6.94514930e-01 -1.44091201e+00
8.24683383e-02 -3.22046697e-01 -3.99326742e-01 -7.47100174e-01
-5.22868693e-01 -9.33342099e-01 -3.79858851e-01 -2.25391054e+00
9.25059557e-01 -7.55932689e-01 -5.99559486e-01 5.67940593e-01
-2.60775924e-01 -3.43137056e-01 -1.89546689e-01 7.75094390e-01
-1.18811972e-01 2.53093332e-01 1.02964294e+00 1.54911876e-01
-3.56122255e-01 4.06374693e-01 -8.11366856e-01 5.71685672e-01
1.02327204e+00 -8.04940343e-01 -5.26448667e-01 -2.90850103e-01
8.95264074e-02 6.31371915e-01 4.21724737e-01 -7.37999260e-01
3.65036987e-02 -7.44658232e-01 5.42815924e-01 -7.39986062e-01
1.85307786e-01 -1.19182861e+00 5.91155410e-01 1.21095216e+00
-1.69873655e-01 6.10271394e-01 -4.70806360e-02 3.90613109e-01
-3.07061166e-01 1.31281614e-02 5.89805841e-01 -1.39993802e-01
-8.40155855e-02 5.08992076e-01 -7.81786323e-01 -1.70985669e-01
1.31294310e+00 -2.79246002e-01 -1.40052423e-01 -3.56835365e-01
-7.33833730e-01 4.40164149e-01 5.04396856e-01 3.88303071e-01
8.26805711e-01 -1.12010682e+00 -8.78811896e-01 4.68715817e-01
4.80074346e-01 8.73888731e-02 3.45505476e-01 9.86806750e-01
-6.61953926e-01 6.75422788e-01 1.48868598e-02 -6.15359843e-01
-1.49695444e+00 6.41682088e-01 5.29716253e-01 -5.26978374e-01
-6.65303111e-01 3.04896653e-01 3.17803234e-01 -4.84395355e-01
3.51166189e-01 -6.82308793e-01 -9.15174372e-03 -3.58614743e-01
6.38448715e-01 4.80073512e-01 1.57691926e-01 -2.93440312e-01
-6.27339542e-01 2.87500411e-01 8.35628361e-02 1.02015471e+00
1.52399802e+00 -4.09552783e-01 -2.17246667e-01 3.16899568e-01
1.06432092e+00 -4.72049236e-01 -6.66429520e-01 3.67730632e-02
-4.87040728e-02 -3.96796346e-01 -2.27801308e-01 -7.90700138e-01
-9.01871920e-01 8.39701176e-01 7.22505927e-01 1.41614348e-01
1.28952622e+00 -4.12265863e-03 1.02117133e+00 2.82391518e-01
1.09539896e-01 -7.02791393e-01 -3.49364817e-01 3.80936801e-01
7.91208804e-01 -1.40774155e+00 6.63327947e-02 -3.99141282e-01
-6.24070048e-01 1.18890536e+00 1.45909727e-01 3.64979476e-01
1.13856554e+00 2.53512949e-01 3.18342477e-01 -1.39302522e-01
-9.63089645e-01 6.01310015e-01 5.15953777e-03 5.68338394e-01
5.27959108e-01 4.11941826e-01 -2.29518682e-01 8.43927205e-01
-1.51143819e-01 2.90474683e-01 3.83064657e-01 1.01583683e+00
-3.20237547e-01 -1.25446784e+00 -4.87189829e-01 1.29433465e+00
-6.13309503e-01 -2.51592189e-01 -2.19000861e-01 3.67952853e-01
1.63805515e-01 9.86252248e-01 1.77515224e-01 -2.78564394e-01
4.98921931e-01 5.95106892e-02 -1.31394818e-01 -6.51299715e-01
-8.66777062e-01 -3.59653264e-01 1.41121969e-02 -2.65470803e-01
-9.08489078e-02 -8.49839151e-01 -1.38190520e+00 -5.33491135e-01
7.55332112e-02 2.40846872e-01 4.07891780e-01 5.73360920e-01
1.00405180e+00 6.32424355e-01 4.81858253e-01 -1.38403371e-01
-4.38322634e-01 -4.69683677e-01 -3.79629225e-01 3.14049244e-01
5.94102323e-01 -2.54519787e-02 -3.12906891e-01 1.42725095e-01] | [7.982918739318848, 6.221606731414795] |
00356dd1-9599-48b9-9a3b-a119a046125e | deep-learning-based-generalized-models-for | 2011.06739 | null | https://arxiv.org/abs/2011.06739v3 | https://arxiv.org/pdf/2011.06739v3.pdf | Generalized Dilated CNN Models for Depression Detection Using Inverted Vocal Tract Variables | Depression detection using vocal biomarkers is a highly researched area. Articulatory coordination features (ACFs) are developed based on the changes in neuromotor coordination due to psychomotor slowing, a key feature of Major Depressive Disorder. However findings of existing studies are mostly validated on a single database which limits the generalizability of results. Variability across different depression databases adversely affects the results in cross corpus evaluations (CCEs). We propose to develop a generalized classifier for depression detection using a dilated Convolutional Neural Network which is trained on ACFs extracted from two depression databases. We show that ACFs derived from Vocal Tract Variables (TVs) show promise as a robust set of features for depression detection. Our model achieves relative accuracy improvements of ~10% compared to CCEs performed on models trained on a single database. We extend the study to show that fusing TVs and Mel-Frequency Cepstral Coefficients can further improve the performance of this classifier. | ['Carol Espy-Wilson', 'Nadee Seneviratne'] | 2020-11-13 | null | null | null | null | ['cross-corpus'] | ['computer-vision'] | [-2.59958059e-02 -2.94025332e-01 -2.19964847e-01 -3.74107033e-01
-9.18150604e-01 -3.78932655e-01 5.64539850e-01 2.90439934e-01
-5.70658624e-01 3.87029141e-01 4.23664957e-01 5.70222028e-02
-2.21520752e-01 -4.25199330e-01 1.23183727e-02 -3.34736526e-01
-1.23515785e-01 3.85391936e-02 -2.84012645e-01 -2.17198789e-01
6.26380965e-02 6.68776810e-01 -1.32342541e+00 4.92480695e-01
6.49170637e-01 5.86607516e-01 1.25020623e-01 5.54941952e-01
2.23000512e-01 4.16359976e-02 -9.28893805e-01 -1.38369322e-01
-2.72622675e-01 -4.38904732e-01 -4.65004861e-01 -3.38932574e-01
1.92256048e-01 -4.37601507e-01 -1.48269773e-01 6.20037377e-01
1.15415835e+00 -4.60321754e-02 8.59939158e-01 -1.03538835e+00
-2.07630947e-01 3.23822767e-01 -2.79748470e-01 6.65618539e-01
2.72252589e-01 1.31824892e-02 7.59830415e-01 -6.74361110e-01
5.60479820e-01 1.18173409e+00 1.10485435e+00 8.10186923e-01
-1.19130182e+00 -9.62883592e-01 -2.39599600e-01 1.33929119e-01
-1.08024716e+00 -8.09735000e-01 5.82394123e-01 -5.21745026e-01
1.58460212e+00 -8.70153829e-02 7.91255772e-01 1.03986275e+00
3.64765674e-01 4.45658147e-01 7.45949388e-01 -1.85758218e-01
1.25249512e-02 -1.34862602e-01 -1.02035575e-01 5.27234256e-01
2.58171737e-01 -7.51417596e-04 -6.72919571e-01 -1.83323130e-01
2.57225513e-01 -2.12553382e-01 5.88829666e-02 1.64660811e-01
-9.81292248e-01 1.05059099e+00 -1.45087205e-02 5.34950554e-01
-4.26714987e-01 1.80435747e-01 8.46406281e-01 3.39746565e-01
7.75570035e-01 3.44063252e-01 -4.82222617e-01 -6.30327046e-01
-1.08824956e+00 2.45864481e-01 2.98553497e-01 3.80868390e-02
8.01311992e-03 5.99960797e-02 -8.32386091e-02 1.33949006e+00
3.73167396e-01 7.35843360e-01 9.75405037e-01 -9.07384932e-01
3.57461870e-01 5.51647067e-01 -5.19194663e-01 -9.87420380e-01
-1.24418318e+00 -3.62642258e-01 -2.89032310e-01 6.96506118e-04
5.56300171e-02 -6.40512586e-01 -3.16062838e-01 1.86920106e+00
1.78661525e-01 -2.08322942e-01 -1.00998133e-02 6.60846651e-01
8.77683759e-01 -1.34092823e-01 -8.87965262e-02 -2.75683761e-01
9.18479860e-01 -4.08903003e-01 -7.33589351e-01 -5.76891378e-03
9.29061174e-01 -6.61131620e-01 6.36136234e-01 6.37433290e-01
-8.45364034e-01 -4.37714040e-01 -1.17603183e+00 2.06803419e-02
-4.79954690e-01 4.80665654e-01 5.86908579e-01 1.01647186e+00
-1.00500190e+00 6.68145180e-01 -1.25019348e+00 -5.01212060e-01
5.79053342e-01 9.03004050e-01 -7.42002785e-01 2.27073878e-01
-1.05032051e+00 9.48777139e-01 -3.79581809e-01 1.20175080e-02
-5.42337120e-01 -6.46172822e-01 -7.30535150e-01 -3.54378313e-01
-6.70510590e-01 -2.86724776e-01 1.18703210e+00 -6.95880234e-01
-1.40792894e+00 9.38875914e-01 -1.03707761e-01 -4.43483680e-01
1.74464792e-01 -4.39911596e-02 -5.82342863e-01 5.37245750e-01
2.25846440e-01 7.07688928e-01 7.10334003e-01 -2.06865728e-01
-5.17283738e-01 -6.63900256e-01 -2.36650556e-01 8.47151875e-02
-6.31845236e-01 2.56549150e-01 1.40122667e-01 -5.21384299e-01
-7.90687129e-02 -9.38885033e-01 3.56857657e-01 -6.72960142e-03
-9.50345770e-02 -3.66573870e-01 8.09161484e-01 -6.97077453e-01
1.33595824e+00 -2.06010032e+00 -4.37122993e-02 -2.41512150e-01
4.21283960e-01 5.70228755e-01 -4.73103881e-01 4.28470522e-01
-3.02562565e-01 3.11379939e-01 1.65556222e-01 -2.80013919e-01
5.13279904e-03 -2.24121422e-01 3.15568030e-01 8.89535666e-01
6.70760930e-01 6.77557528e-01 -7.21415818e-01 -5.99422446e-03
2.22572982e-01 5.94165921e-01 -8.48401904e-01 -6.23776652e-02
4.42914188e-01 2.11325847e-02 -1.89262390e-01 4.52171296e-01
6.51868522e-01 4.12364095e-01 2.12412089e-01 -1.46069288e-01
3.88501659e-02 6.21615052e-01 -5.33901155e-01 1.60670710e+00
-4.81193185e-01 1.06248069e+00 2.30113454e-02 -1.26289904e+00
9.25997019e-01 4.67899173e-01 6.19123340e-01 -7.24987805e-01
2.20796853e-01 2.41937682e-01 8.50493908e-01 -6.89370453e-01
-3.43839914e-01 -5.40544391e-01 8.70789401e-03 4.27668035e-01
3.32114309e-01 -1.83923408e-01 2.29734018e-01 -2.58356154e-01
1.47835946e+00 -9.92225483e-02 9.97288749e-02 -1.72927901e-01
3.96129251e-01 -2.45540023e-01 5.06157517e-01 2.22096369e-01
-6.44788921e-01 7.56034672e-01 8.77983212e-01 1.79964185e-01
-6.66516006e-01 -8.73323321e-01 -7.29494810e-01 9.75798726e-01
-7.05077291e-01 -6.27647996e-01 -7.77832985e-01 -5.70033491e-01
3.13696623e-01 2.80948609e-01 -8.70090902e-01 -5.21663249e-01
-2.01556072e-01 -1.15589797e+00 1.27653611e+00 7.20249772e-01
1.44921139e-01 -8.74272168e-01 -6.15569174e-01 3.54596674e-01
5.69088049e-02 -1.01721156e+00 -2.77353883e-01 4.64610785e-01
-8.74427497e-01 -1.20144069e+00 -7.04577506e-01 -6.98438406e-01
-7.28740031e-03 6.26924485e-02 5.84236979e-01 -9.98758003e-02
-5.38285136e-01 5.11160493e-01 -2.83835769e-01 -7.87500203e-01
-4.83129710e-01 1.20435365e-01 5.63377261e-01 -4.38721269e-01
7.50214994e-01 -7.10238039e-01 -6.32875919e-01 -1.13394283e-01
-5.79204798e-01 -5.32624006e-01 3.64618808e-01 7.97585070e-01
3.17342371e-01 -3.66600245e-01 1.61391342e+00 -5.30807912e-01
1.10487545e+00 -6.22182906e-01 2.19091833e-01 -4.15537536e-01
-6.38239324e-01 -3.14342707e-01 4.04951453e-01 -4.79109317e-01
-5.89477301e-01 -2.67316639e-01 -4.71494287e-01 -3.16362023e-01
-2.59275436e-01 6.23017311e-01 2.92049885e-01 1.90393820e-01
6.29855096e-01 -2.04040930e-01 6.18313253e-01 -4.50208366e-01
-9.61043388e-02 1.21376228e+00 2.65299499e-01 -1.35985598e-01
7.42570013e-02 3.23168576e-01 9.81858745e-03 -1.18298066e+00
-4.08427835e-01 -7.06676126e-01 -7.31720269e-01 -1.70425624e-02
8.80245984e-01 -1.03428507e+00 -6.34753108e-01 5.40782273e-01
-8.30126584e-01 -1.83192551e-01 2.81402260e-01 9.18370068e-01
-5.22667646e-01 1.96019620e-01 -5.61723113e-01 -6.67430222e-01
-7.43928313e-01 -1.04371214e+00 1.26849926e+00 -1.50059378e-02
-8.12285900e-01 -1.07179034e+00 7.30378091e-01 3.09421718e-01
3.36141199e-01 3.94405514e-01 1.07822227e+00 -9.92908239e-01
8.13434958e-01 -3.79227489e-01 8.55833068e-02 6.99510634e-01
6.50993288e-01 2.16402367e-01 -1.04458344e+00 -1.93679050e-01
-4.79339287e-02 -6.29347444e-01 7.82650411e-01 6.05657697e-01
6.16916835e-01 3.92485112e-01 -3.72521073e-01 2.75229722e-01
7.96101451e-01 3.75701129e-01 3.09832484e-01 1.72688350e-01
2.40769058e-01 4.72616345e-01 2.89157957e-01 4.24878746e-01
6.84971660e-02 6.32915854e-01 1.52568966e-01 -1.35359734e-01
-5.98065078e-01 2.69311309e-01 7.00821102e-01 9.09459054e-01
-1.76548455e-02 1.75564528e-01 -8.31302822e-01 7.69344211e-01
-1.28840423e+00 -8.33154917e-01 -1.51281178e-01 1.86650956e+00
8.06309164e-01 -2.40061134e-02 4.52466905e-01 3.99555564e-01
6.19526029e-01 -1.23907402e-01 -6.84138417e-01 -8.04443359e-01
-3.31500731e-02 8.26675475e-01 -2.69343913e-01 7.94669539e-02
-8.56754899e-01 4.67539132e-01 7.18895578e+00 5.20869672e-01
-1.51137543e+00 1.55546591e-01 1.32844478e-01 -6.13736749e-01
2.65469790e-01 -9.42815304e-01 -5.63562393e-01 2.70830035e-01
1.50203323e+00 -2.79605128e-02 3.82039189e-01 2.89497226e-01
8.60433161e-01 3.29746492e-02 -9.80567157e-01 8.57248545e-01
1.27446175e-01 -8.98308992e-01 -4.64304477e-01 1.47382692e-01
4.34001178e-01 6.85994208e-01 3.85052681e-01 3.02277148e-01
-4.28367287e-01 -1.27865052e+00 2.00722277e-01 4.37595963e-01
1.19846284e+00 -1.03631580e+00 8.76223207e-01 3.67644504e-02
-8.13238978e-01 -5.18573672e-02 -2.54629701e-01 -1.24799833e-02
-5.13837993e-01 2.98574179e-01 -1.35527897e+00 -2.70021297e-02
5.10672688e-01 9.73511159e-01 -5.46883762e-01 8.72500062e-01
2.38301247e-01 8.89109552e-01 -2.91290313e-01 -1.86780974e-01
8.90455171e-02 7.24386945e-02 3.47747058e-01 1.46350861e+00
3.20422024e-01 -3.14977139e-01 -5.57513893e-01 6.54001474e-01
-7.62103079e-03 2.85070181e-01 -7.43172228e-01 -7.33539283e-01
1.51182681e-01 1.49843681e+00 -4.87560183e-01 2.35742256e-01
-7.24686265e-01 5.89187801e-01 3.36912483e-01 -1.26200154e-01
-5.60810328e-01 -6.63233459e-01 1.08153892e+00 -5.90757467e-02
1.87948093e-01 8.35933462e-02 -2.01609675e-02 -8.09126198e-01
-2.79159904e-01 -1.15513301e+00 2.16255337e-01 -4.23921019e-01
-1.18914700e+00 3.64418805e-01 -2.87742764e-01 -1.13855922e+00
-4.77003425e-01 -7.03623950e-01 -6.36174440e-01 9.75973189e-01
-1.35389984e+00 -8.90721917e-01 6.07332587e-02 3.80300850e-01
5.77333987e-01 -3.40155333e-01 1.23961127e+00 3.47515732e-01
-8.00367653e-01 8.07352006e-01 3.84951919e-01 2.09625036e-01
9.15974021e-01 -1.13210940e+00 -5.70214838e-02 3.61828983e-01
-2.51797251e-02 6.78023279e-01 1.67037025e-01 -5.55913210e-01
-1.05927837e+00 -1.03370178e+00 8.59800756e-01 -4.01503384e-01
5.87717354e-01 -1.56835377e-01 -5.40140867e-01 2.02808097e-01
1.53213814e-02 -3.40497017e-01 1.24517787e+00 2.99181491e-01
2.54438035e-02 -1.42739564e-01 -1.32506239e+00 1.65209025e-01
8.74166071e-01 -7.11004913e-01 -6.48049414e-01 2.78012633e-01
1.54959098e-01 -9.54916030e-02 -1.14732242e+00 4.47625071e-01
9.54651177e-01 -8.83400977e-01 7.31251121e-01 -7.50559449e-01
6.87106550e-01 2.22074106e-01 -1.22864932e-01 -1.55737138e+00
-1.50190398e-01 -2.18050748e-01 3.63641083e-02 1.12725651e+00
3.88853669e-01 -4.99838442e-01 4.39655721e-01 -4.28879214e-03
-1.84741259e-01 -7.45480299e-01 -9.85025167e-01 -5.65330684e-01
6.17411137e-01 -6.02949679e-01 2.07608283e-01 7.01276183e-01
7.38244534e-01 4.76981193e-01 2.91748166e-01 -2.31150955e-01
-1.58246621e-01 -2.76601404e-01 1.60726368e-01 -1.36193860e+00
-5.41286319e-02 -6.34891272e-01 -6.86806321e-01 -4.31786142e-02
4.25692886e-01 -1.38186288e+00 -1.93578839e-01 -1.62184703e+00
2.14992076e-01 2.45957393e-02 -3.31972897e-01 5.90092957e-01
4.99479324e-02 6.10079944e-01 -3.17801833e-02 -3.28203380e-01
-3.01067084e-02 4.39701170e-01 1.04178798e+00 -1.33435845e-01
-4.44570869e-01 -3.14548984e-02 -9.18798864e-01 8.53433847e-01
1.02639198e+00 -5.39819539e-01 -7.30078578e-01 -3.78019027e-02
-1.40087530e-01 1.13227159e-01 -8.94355401e-02 -1.05352890e+00
-2.21686617e-01 2.78739899e-01 7.49459147e-01 -3.94839317e-01
5.86331844e-01 -6.60277754e-02 -3.94576043e-01 8.12446654e-01
-3.86453718e-01 2.03744695e-01 6.63949728e-01 3.48672152e-01
-7.32552260e-02 -3.04889619e-01 6.24833882e-01 2.55899310e-01
-1.71497539e-02 -6.86736405e-02 -1.01793301e+00 -1.27078698e-03
4.29215699e-01 6.98161647e-02 -2.46398643e-01 -4.06141520e-01
-8.48851979e-01 -2.46439442e-01 -1.57938316e-01 4.35441524e-01
4.20326442e-01 -1.20765674e+00 -6.61984921e-01 1.38055727e-01
-3.05030327e-02 -8.04284751e-01 7.82754458e-03 1.26274288e+00
-4.07204002e-01 8.72570038e-01 -3.54276448e-01 -5.05783916e-01
-1.60524046e+00 2.03948971e-02 7.40563810e-01 3.44413221e-02
-4.52609539e-01 7.00770080e-01 -2.02226445e-01 -5.30208170e-01
4.96414937e-02 -5.35103381e-01 -6.67521119e-01 6.84626937e-01
5.45343935e-01 6.10623717e-01 7.63106406e-01 -6.78848982e-01
-7.59675384e-01 3.54528457e-01 -9.71822143e-02 -2.21534565e-01
1.56516075e+00 -4.00042571e-02 5.97084984e-02 4.35850263e-01
1.50224721e+00 2.98327178e-01 -7.49953270e-01 3.85036856e-01
-3.00745934e-01 2.26584837e-01 4.20557141e-01 -9.26745474e-01
-1.03297830e+00 1.13178158e+00 1.28422427e+00 3.84435318e-02
1.06406665e+00 -2.12064087e-01 5.90317130e-01 4.56620485e-01
-2.02743977e-01 -1.23894942e+00 1.73548281e-01 2.65338868e-01
9.57363665e-01 -8.57413471e-01 -1.57867119e-01 1.33133128e-01
-4.86863196e-01 1.52124155e+00 3.89641285e-01 -2.57599354e-01
7.79912114e-01 4.22853470e-01 2.81832069e-01 -2.42147192e-01
-8.66580546e-01 -5.09985447e-01 2.87633032e-01 1.17527664e+00
9.83439863e-01 4.17465977e-02 -9.60813224e-01 1.17722487e+00
-5.08049190e-01 1.11385416e-02 6.18215919e-01 5.50339997e-01
-2.34392926e-01 -1.21820748e+00 8.74825940e-02 1.07104325e+00
-1.01112211e+00 -9.38468128e-02 -6.08083248e-01 8.87182891e-01
5.16086280e-01 1.43772125e+00 2.49147266e-01 -5.23430169e-01
2.67290086e-01 3.29960465e-01 5.49682498e-01 -9.78483856e-01
-7.51045048e-01 2.89488107e-01 4.44014102e-01 -3.74554962e-01
-7.36645699e-01 -1.26081789e+00 -1.51379001e+00 -1.25832809e-02
-2.91723579e-01 -2.55561739e-01 6.58804595e-01 1.08629298e+00
4.45288301e-01 6.83160245e-01 6.04252458e-01 -4.72404033e-01
-4.04590130e-01 -1.36060357e+00 -7.48086035e-01 9.80297402e-02
6.17269039e-01 -8.20140064e-01 -2.29097128e-01 -5.44096045e-02] | [13.799832344055176, 5.067220211029053] |
0479f8a3-0b2d-43a7-9f39-20184704d6e4 | analysis-of-impact-of-emotions-on-target | 2208.07091 | null | https://arxiv.org/abs/2208.07091v1 | https://arxiv.org/pdf/2208.07091v1.pdf | Analysis of impact of emotions on target speech extraction and speech separation | Recently, the performance of blind speech separation (BSS) and target speech extraction (TSE) has greatly progressed. Most works, however, focus on relatively well-controlled conditions using, e.g., read speech. The performance may degrade in more realistic situations. One of the factors causing such degradation may be intrinsic speaker variability, such as emotions, occurring commonly in realistic speech. In this paper, we investigate the influence of emotions on TSE and BSS. We create a new test dataset of emotional mixtures for the evaluation of TSE and BSS. This dataset combines LibriSpeech and Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS). Through controlled experiments, we can analyze the impact of different emotions on the performance of BSS and TSE. We observe that BSS is relatively robust to emotions, while TSE, which requires identifying and extracting the speech of a target speaker, is much more sensitive to emotions. On comparative speaker verification experiments we show that identifying the target speaker may be particularly challenging when dealing with emotional speech. Using our findings, we outline potential future directions that could improve the robustness of BSS and TSE systems toward emotional speech. | ['Jan Černocký', 'Ladislav Mošner', 'Tsubasa Ochiai', 'Marc Delcroix', 'Martin Kocour', 'Kateřina Žmolíková', 'Ján Švec'] | 2022-08-15 | null | null | null | null | ['speech-separation', 'speech-extraction'] | ['speech', 'speech'] | [-3.13609272e-01 -2.98384219e-01 3.84011418e-01 -3.30845296e-01
-8.61952960e-01 -6.45244360e-01 3.32067221e-01 -2.79040188e-01
-1.74705788e-01 4.98406351e-01 3.72783363e-01 -1.51925668e-01
1.38325244e-01 9.99368131e-02 -4.32714760e-01 -9.86076057e-01
-4.99685071e-02 -1.91163458e-02 -2.10076839e-01 -6.35827407e-02
-7.05914758e-03 6.39247239e-01 -1.85748315e+00 6.57171533e-02
7.52721906e-01 8.87962282e-01 1.66406274e-01 8.75649631e-01
3.19232233e-02 5.24620771e-01 -1.26610589e+00 -2.51938134e-01
1.05783097e-01 -4.54052567e-01 -2.73571968e-01 1.56323597e-01
3.46093476e-02 3.12335771e-02 -9.39444155e-02 1.17368805e+00
1.11319077e+00 2.09591001e-01 6.47562504e-01 -1.37598801e+00
-4.13536310e-01 4.94001001e-01 -2.05031797e-01 1.75600827e-01
4.10968065e-01 3.86924669e-02 5.83753169e-01 -1.07003593e+00
8.90986323e-02 1.28603923e+00 4.37974125e-01 6.57605112e-01
-1.01941788e+00 -7.77580619e-01 -6.58202842e-02 2.89647520e-01
-1.51337564e+00 -1.42319858e+00 7.98004031e-01 -3.21974397e-01
8.39888871e-01 6.49872839e-01 3.80136579e-01 1.43741715e+00
-3.91732067e-01 7.11284101e-01 1.31312346e+00 -5.63331604e-01
4.81792390e-01 6.87908947e-01 2.28158996e-01 -2.66683754e-02
-1.04503572e-01 3.28695208e-01 -5.60325921e-01 -1.77793741e-01
9.39560086e-02 -7.14561522e-01 -8.84166300e-01 1.47177577e-01
-1.01721454e+00 4.24476296e-01 -1.24839462e-01 5.04719317e-01
-3.98172259e-01 -3.66848081e-01 2.53857762e-01 4.09152240e-01
4.48553890e-01 3.61633033e-01 -2.98673362e-01 -4.41401064e-01
-1.19586349e+00 -2.27863371e-01 9.95019913e-01 6.35035038e-01
1.02273367e-01 3.81780535e-01 -2.78925300e-01 1.42139244e+00
4.21952397e-01 8.02398801e-01 6.15938842e-01 -6.92974806e-01
-3.60372551e-02 -2.51414895e-01 3.36793423e-01 -7.98615098e-01
-2.54838794e-01 -6.29202604e-01 -5.27771831e-01 2.67779738e-01
2.67835855e-01 -4.11038667e-01 -9.09458220e-01 1.74424350e+00
1.51815012e-01 2.73274392e-01 5.74916661e-01 1.11262548e+00
9.81193841e-01 7.30806470e-01 -2.13028342e-01 -7.01445937e-01
1.19691706e+00 -6.51588142e-01 -1.08763778e+00 -3.73782218e-01
6.52276576e-02 -1.11297679e+00 1.00879514e+00 5.51643550e-01
-9.89579916e-01 -2.29741320e-01 -9.74240482e-01 4.93517101e-01
-1.98996574e-01 3.86400551e-01 5.11613339e-02 1.24277067e+00
-1.20601618e+00 -5.24592958e-02 -4.87289816e-01 -5.38008869e-01
-8.14191028e-02 3.08841854e-01 -2.02594921e-01 2.72734374e-01
-1.26852918e+00 1.06846857e+00 -1.31104633e-01 2.39428386e-01
-8.26610684e-01 -1.24906451e-01 -6.29249275e-01 1.50099292e-01
2.85908543e-02 -1.89318523e-01 1.26358855e+00 -1.27288771e+00
-1.79344881e+00 7.21215665e-01 -6.78535640e-01 -2.84392953e-01
1.63126290e-01 8.89446866e-03 -1.04518306e+00 9.79609787e-03
-2.33664319e-01 3.20600688e-01 1.26764297e+00 -1.68644261e+00
3.30032408e-02 -2.41308153e-01 -5.27638674e-01 1.92289352e-01
-3.44829202e-01 7.67575681e-01 -4.01253581e-01 -8.47859740e-01
-5.74167147e-02 -1.06085300e+00 3.56556416e-01 -5.92717290e-01
-3.65668029e-01 8.47497657e-02 7.01894403e-01 -1.07202554e+00
1.08384430e+00 -2.76807714e+00 1.23968907e-01 2.78145403e-01
-2.07409188e-01 4.73114580e-01 -6.33690953e-02 8.90640244e-02
-4.65343088e-01 1.13333501e-01 -4.56265286e-02 -5.45537114e-01
2.28532508e-01 -3.55737880e-02 -3.61906350e-01 4.82980818e-01
1.21933147e-02 4.08666790e-01 -4.16347653e-01 -3.72165382e-01
4.55657579e-02 6.22212946e-01 -2.17796609e-01 4.78107631e-01
2.69330770e-01 3.59768152e-01 1.91102684e-01 7.54449785e-01
7.22325623e-01 4.23027635e-01 -1.94833785e-01 -7.97444209e-02
-4.66087312e-02 2.68698722e-01 -1.16530836e+00 9.26826477e-01
-3.99164498e-01 1.09326065e+00 7.61418462e-01 -6.78530276e-01
1.00421250e+00 8.04978311e-01 -4.29816879e-02 -4.51198906e-01
2.17864633e-01 2.56194860e-01 3.04668248e-01 -6.58314943e-01
4.04382616e-01 -5.50357223e-01 2.17294097e-01 3.03303182e-01
-6.37626722e-02 -2.41148874e-01 -2.35627025e-01 5.81953079e-02
6.48834348e-01 -7.67965376e-01 1.24608144e-01 -2.15870272e-02
4.93081599e-01 -6.68909669e-01 4.42650735e-01 4.18354928e-01
-6.74325109e-01 5.51602781e-01 4.21903491e-01 7.80016959e-01
-5.64828932e-01 -1.08395767e+00 -1.42323554e-01 9.05878246e-01
-3.31092090e-03 -1.79968178e-01 -9.03548062e-01 -1.60473987e-01
-2.83041239e-01 1.15642560e+00 -2.36415640e-01 -2.83858418e-01
1.17802754e-01 -8.61045659e-01 8.36477101e-01 2.32027709e-01
1.99089214e-01 -9.37931359e-01 -1.61392450e-01 -2.29193002e-01
-4.61010247e-01 -1.23644996e+00 -4.62475061e-01 1.40860006e-01
-3.03357780e-01 -4.02482927e-01 -8.95883262e-01 -7.28234410e-01
4.50917602e-01 2.72249609e-01 8.02031159e-01 -2.88062871e-01
1.89798623e-01 5.73380411e-01 -4.58573073e-01 -6.36945248e-01
-7.76875496e-01 -5.09932876e-01 5.24637282e-01 3.63083631e-01
3.85441512e-01 -4.80805784e-01 -1.33732915e-01 6.48147762e-01
-6.06501400e-01 -3.37965459e-01 4.10427958e-01 6.87300742e-01
-3.91560979e-02 4.53752816e-01 8.64186764e-01 3.23773138e-02
1.04160273e+00 -3.98758620e-01 -1.57918811e-01 2.13005736e-01
-2.91529030e-01 -1.51163310e-01 2.65678287e-01 -7.00734556e-01
-1.27307260e+00 -3.46353263e-01 -2.55255789e-01 -6.18566036e-01
-5.47957897e-01 3.74716163e-01 -6.15905046e-01 -1.07555911e-01
6.07706666e-01 1.73571467e-01 -1.14007197e-01 -3.56199652e-01
-3.49671505e-02 1.66021442e+00 4.35912967e-01 -3.66519183e-01
5.97427547e-01 3.48669402e-02 -6.83613122e-01 -1.45644760e+00
-1.88651428e-01 -5.04023552e-01 9.86212790e-02 -5.56099713e-01
5.14928162e-01 -1.04408169e+00 -4.79094476e-01 8.26061845e-01
-1.02085185e+00 -2.44047821e-01 2.00877264e-01 7.90188789e-01
-1.90245047e-01 4.51976359e-01 -4.37845170e-01 -1.47267199e+00
-1.23007044e-01 -1.30639780e+00 9.20196712e-01 1.93947226e-01
-6.12848461e-01 -4.26473230e-01 -7.36912489e-02 6.15511239e-01
5.21436572e-01 -3.06046128e-01 6.35618210e-01 -8.65857661e-01
-7.39509705e-03 4.10213880e-02 1.41601652e-01 8.53458881e-01
2.54273891e-01 2.39786178e-01 -1.56809282e+00 -3.42969805e-01
5.26535928e-01 -1.71976224e-01 4.87362951e-01 5.35670519e-01
6.13206625e-01 -1.37810484e-01 -2.51941122e-02 3.86923462e-01
6.79651141e-01 6.39103830e-01 6.97952926e-01 4.78334464e-02
2.13553682e-01 9.38516915e-01 6.11669719e-01 2.61491507e-01
-2.50518061e-02 7.05845177e-01 7.94326589e-02 -1.77169740e-01
-1.49635196e-01 3.57226282e-01 9.69883740e-01 1.13029063e+00
1.05276354e-01 -4.30761307e-01 -9.32683110e-01 5.76331794e-01
-1.22920752e+00 -8.15726519e-01 2.14768633e-01 2.27795386e+00
7.25763977e-01 -1.65592849e-01 2.35139057e-01 5.65256596e-01
1.06373596e+00 9.83803049e-02 -4.42706108e-01 -4.18158084e-01
-5.68938076e-01 -4.97687198e-02 8.32813308e-02 5.77845633e-01
-7.50791132e-01 6.12070680e-01 6.23598433e+00 6.84198737e-01
-1.51174402e+00 1.81130953e-02 4.91715044e-01 -6.30696595e-01
-9.11065340e-02 -5.50552547e-01 -1.57429263e-01 6.61206305e-01
1.15667403e+00 -1.53046519e-01 7.77879775e-01 4.57260132e-01
6.63213551e-01 -1.02045693e-01 -8.90668035e-01 1.51165450e+00
4.62014854e-01 -2.67802626e-01 -5.52900910e-01 -1.24097437e-01
2.44605511e-01 2.77600363e-02 2.66679496e-01 2.83979714e-01
-1.70697957e-01 -1.06507468e+00 8.95231962e-01 2.24528179e-01
5.82734346e-01 -7.30135322e-01 7.03710973e-01 1.93502501e-01
-5.59642255e-01 -6.95115626e-02 3.15206721e-02 3.17923397e-01
-1.65055059e-02 6.33853793e-01 -1.00942993e+00 1.27205729e-01
8.05478573e-01 2.41609976e-01 -4.25455064e-01 1.02698565e+00
-2.91822940e-01 9.98763263e-01 -2.96639323e-01 -1.72630787e-01
-4.07661349e-01 -1.63153559e-01 1.08459663e+00 1.39404428e+00
6.42592609e-01 6.36375844e-02 -4.97141153e-01 6.25374854e-01
4.73256968e-02 2.50759870e-01 -3.61221910e-01 -4.70099449e-01
6.65989280e-01 1.07886207e+00 -5.70371628e-01 -2.50882089e-01
-2.48951182e-01 9.85155821e-01 -1.73325673e-01 8.58999133e-01
-1.00578094e+00 -3.13711405e-01 1.00549591e+00 -3.93099219e-01
2.98229486e-01 -1.41586632e-01 -1.56673461e-01 -1.32205021e+00
2.07353711e-01 -1.47736931e+00 -7.84679875e-02 -1.24304640e+00
-9.98686790e-01 7.45944619e-01 -3.44917446e-01 -1.11865771e+00
-1.16266586e-01 -4.28094953e-01 -5.44919312e-01 1.00587606e+00
-1.18940389e+00 -4.36971188e-01 -1.37601882e-01 6.20046914e-01
4.84278589e-01 -3.12694609e-01 7.14311361e-01 3.46442342e-01
-9.39900696e-01 7.37859905e-01 1.45088166e-01 2.68275253e-02
1.12746501e+00 -1.00177121e+00 1.61611270e-02 9.92333055e-01
2.20342785e-01 6.99270666e-01 1.24843097e+00 -2.93300062e-01
-1.15278244e+00 -6.52248144e-01 7.02820182e-01 -1.33365601e-01
3.88502419e-01 -5.39719164e-01 -9.31411028e-01 3.33967447e-01
3.04679811e-01 -5.27388692e-01 9.92180467e-01 1.88732013e-01
-2.45089829e-01 -2.59971470e-02 -1.18584704e+00 7.23959506e-01
5.18568456e-01 -9.12383020e-01 -7.67417729e-01 -3.68253551e-02
5.29455900e-01 -1.24221891e-01 -5.60266972e-01 1.97884500e-01
5.30052185e-01 -1.23497522e+00 9.09751713e-01 -2.12412700e-01
-2.49881148e-01 -4.77244526e-01 -3.09545994e-01 -1.93971860e+00
-4.52273991e-03 -7.23150551e-01 4.53410968e-02 1.62536323e+00
5.07945120e-01 -8.31185758e-01 2.19959527e-01 7.55859613e-01
5.03459536e-02 -1.31880358e-01 -8.92038345e-01 -9.44477737e-01
-3.50281507e-01 -6.72785521e-01 6.18461370e-01 1.09909880e+00
2.51360357e-01 3.51305723e-01 -4.60625470e-01 6.98451281e-01
2.34127700e-01 -1.57582536e-01 6.72093749e-01 -8.27143431e-01
-4.22170401e-01 -5.21811903e-01 -3.34297001e-01 -6.45095766e-01
3.84264261e-01 -4.98173326e-01 3.83079499e-01 -1.17726851e+00
-2.62062341e-01 7.08713457e-02 -3.21823567e-01 3.93696874e-02
-2.06055731e-01 -1.63715854e-01 2.40364224e-01 -9.06888992e-02
-1.63499370e-01 6.90992773e-01 7.27766812e-01 -1.54449582e-01
-2.88318306e-01 6.33798763e-02 -6.38204873e-01 5.10556042e-01
9.03102458e-01 -3.37139130e-01 -3.66423905e-01 -3.97860892e-02
-1.63812429e-01 1.53378755e-01 2.48695284e-01 -8.18260908e-01
2.88433939e-01 1.93030030e-01 7.39011839e-02 -2.29497954e-01
7.56400049e-01 -7.82767057e-01 2.63360769e-01 -1.68461073e-02
-6.45589456e-02 -4.77419168e-01 5.58442891e-01 3.24568361e-01
-5.49993575e-01 -1.98379219e-01 8.06594729e-01 3.56046885e-01
-2.44923785e-01 -4.00369346e-01 -8.22625875e-01 -1.60181731e-01
8.70406210e-01 -2.59066284e-01 -1.99945018e-01 -1.06326020e+00
-8.52504432e-01 -4.66411794e-03 4.32924807e-01 4.06761914e-01
6.06979132e-01 -1.04039025e+00 -7.46265233e-01 3.75389844e-01
7.74879521e-03 -5.80382645e-01 3.45614910e-01 1.11318636e+00
4.78135608e-02 2.08415851e-01 1.56180918e-01 -4.30860490e-01
-1.89163566e+00 5.57624936e-01 6.22440875e-01 6.71686888e-01
1.68977156e-01 9.65896428e-01 1.01272285e-01 -7.51762316e-02
5.95354736e-01 -1.58779025e-01 -3.11638206e-01 3.74116272e-01
3.95231992e-01 5.18332779e-01 2.19199017e-01 -9.54200745e-01
-6.31839812e-01 2.30769008e-01 2.33243987e-01 -5.97313404e-01
9.70273256e-01 -2.83230335e-01 -2.69102067e-01 6.50063753e-01
9.96703565e-01 6.26815438e-01 -6.02908611e-01 1.29362896e-01
-3.84487808e-02 -4.41733599e-01 4.05234426e-01 -9.58516896e-01
-9.88022625e-01 8.18407595e-01 7.79908121e-01 4.80829924e-01
1.51542246e+00 5.21325991e-02 5.36319792e-01 1.24615334e-01
1.47091597e-01 -1.03862631e+00 -2.65337914e-01 2.70741403e-01
1.14451277e+00 -1.21194351e+00 -6.53132975e-01 -3.62828285e-01
-8.88576329e-01 7.88849115e-01 2.93266326e-01 5.83851993e-01
7.01759100e-01 4.70993102e-01 6.14033163e-01 3.34511288e-02
-6.31266654e-01 -2.10032552e-01 3.67412627e-01 5.66541016e-01
4.78475213e-01 3.49138290e-01 2.25598827e-01 8.07471335e-01
-5.82184076e-01 -2.84344316e-01 4.81632143e-01 4.20828164e-01
-2.17794687e-01 -8.35201204e-01 -1.19884741e+00 1.54670775e-01
-5.41648984e-01 -2.53947377e-01 -8.88466358e-01 2.35810891e-01
-2.31643498e-01 1.73134220e+00 -6.04207627e-02 -4.51647341e-01
3.85087669e-01 5.49986005e-01 1.87603891e-01 -3.07864249e-01
-4.35193747e-01 5.38513064e-01 5.28644085e-01 -1.41022816e-01
-6.16617680e-01 -8.60186160e-01 -9.29094970e-01 -7.54777417e-02
-5.29162943e-01 4.61756676e-01 9.24913824e-01 7.81112373e-01
3.94233704e-01 5.43614626e-01 5.93307495e-01 -6.71826780e-01
-2.95213073e-01 -1.21197224e+00 -9.21181679e-01 2.93783426e-01
6.87342525e-01 -6.53129280e-01 -1.16425824e+00 -4.16536927e-02] | [14.53093433380127, 6.003451347351074] |
6397d4a3-68f6-49b6-b1a2-c6831586cb03 | contrastive-weighted-learning-for-near | 2211.03073 | null | https://arxiv.org/abs/2211.03073v2 | https://arxiv.org/pdf/2211.03073v2.pdf | Contrastive Weighted Learning for Near-Infrared Gaze Estimation | Appearance-based gaze estimation has been very successful with the use of deep learning. Many following works improved domain generalization for gaze estimation. However, even though there has been much progress in domain generalization for gaze estimation, most of the recent work have been focused on cross-dataset performance -- accounting for different distributions in illuminations, head pose, and lighting. Although improving gaze estimation in different distributions of RGB images is important, near-infrared image based gaze estimation is also critical for gaze estimation in dark settings. Also there are inherent limitations relying solely on supervised learning for regression tasks. This paper contributes to solving these problems and proposes GazeCWL, a novel framework for gaze estimation with near-infrared images using contrastive learning. This leverages adversarial attack techniques for data augmentation and a novel contrastive loss function specifically for regression tasks that effectively clusters the features of different samples in the latent space. Our model outperforms previous domain generalization models in infrared image based gaze estimation and outperforms the baseline by 45.6\% while improving the state-of-the-art by 8.6\%, we demonstrate the efficacy of our method. | ['Adam Lee'] | 2022-11-06 | null | null | null | null | ['gaze-estimation'] | ['computer-vision'] | [ 1.98352247e-01 -2.85589248e-02 -9.62765068e-02 -6.60486519e-01
-6.20418668e-01 -4.03546154e-01 2.57508099e-01 -5.62692761e-01
-4.10512835e-01 6.32286727e-01 1.88826043e-02 -1.29377797e-01
-3.57248858e-02 5.89003637e-02 -7.15498447e-01 -9.12248790e-01
2.56178677e-01 -1.64435640e-01 -1.12926938e-01 -1.48915306e-01
5.83808899e-01 1.51992843e-01 -1.93508673e+00 -5.03083430e-02
7.44379759e-01 1.09134161e+00 -2.08691239e-01 7.65428960e-01
3.40911746e-01 7.21740067e-01 -6.56763494e-01 -1.12787612e-01
2.69456983e-01 -5.24070799e-01 -6.81620955e-01 -2.26079106e-01
1.49451125e+00 -4.55877125e-01 1.33134738e-01 9.03338134e-01
7.57191300e-01 2.77856261e-01 7.79521585e-01 -1.80624151e+00
-8.71170282e-01 -2.16324940e-01 -1.12653172e+00 2.83892304e-01
5.41662633e-01 1.40180722e-01 6.18184984e-01 -5.39728522e-01
4.36752081e-01 1.13486004e+00 6.10127091e-01 1.22057760e+00
-1.15897298e+00 -1.50893497e+00 2.75111049e-01 3.05228919e-01
-1.18605089e+00 -6.15848720e-01 7.11547196e-01 -3.66856426e-01
7.43581116e-01 1.84983499e-02 1.90081954e-01 1.51636195e+00
9.18495953e-02 6.97253227e-01 1.71510828e+00 -6.01599753e-01
1.60095375e-02 2.79080600e-01 2.46361047e-02 4.99919981e-01
-1.28556579e-01 1.01459935e-01 -1.11584222e+00 2.56647587e-01
3.07425052e-01 -4.76843938e-02 -4.21492875e-01 -4.67386633e-01
-7.48445272e-01 6.11678064e-01 7.68913686e-01 -2.59805650e-01
1.12485602e-01 2.21393853e-01 4.91857529e-02 3.05992603e-01
1.08169591e+00 3.74326080e-01 -4.26870465e-01 -9.88908932e-02
-1.05196345e+00 3.22604626e-01 4.11194831e-01 1.01179039e+00
7.92494476e-01 -1.81247979e-01 -1.03541628e-01 5.31766355e-01
6.12189829e-01 7.94920862e-01 2.13324681e-01 -1.02229548e+00
4.50830430e-01 4.17728990e-01 8.40652213e-02 -5.95045924e-01
-4.76313889e-01 8.84399097e-03 -2.28335842e-01 9.90102708e-01
6.96712017e-01 -3.94143939e-01 -1.05840206e+00 2.06895041e+00
4.09268260e-01 2.50685811e-01 -2.65480667e-01 9.75016534e-01
6.23424351e-01 1.15775257e-01 4.08382207e-01 -1.73045360e-02
1.10154426e+00 -7.57876337e-01 -7.53649414e-01 -2.15352848e-01
5.27394295e-01 -7.53937006e-01 1.30631340e+00 5.22124708e-01
-9.86926496e-01 -4.08047259e-01 -9.20890927e-01 -3.58193338e-01
-4.21283633e-01 3.01934425e-02 4.46162462e-01 1.24098301e+00
-1.42004693e+00 1.42880529e-01 -5.91304064e-01 -8.20535600e-01
8.34469080e-01 7.83449888e-01 -4.24209327e-01 -5.92187000e-03
-5.75785577e-01 8.23326111e-01 -2.60685802e-01 -1.58481047e-01
-4.58969176e-01 -1.05750847e+00 -9.04235065e-01 -2.95574456e-01
1.36049017e-01 -6.36518359e-01 1.12457788e+00 -1.29947317e+00
-1.58827102e+00 1.11663532e+00 -6.52920127e-01 -2.51183420e-01
2.99824774e-01 -5.83371878e-01 -3.01222116e-01 -2.50900481e-02
-7.01538473e-02 1.18936610e+00 1.38229477e+00 -1.10380566e+00
-5.90088487e-01 -8.14965546e-01 1.70751005e-01 3.78824830e-01
-3.64154816e-01 4.02338564e-01 -7.44583309e-02 -1.67212799e-01
-2.02884421e-01 -1.24647117e+00 5.38029552e-01 3.88870776e-01
-2.15530232e-01 -4.35724884e-01 1.27262211e+00 -5.29040694e-01
9.10950124e-01 -2.18877244e+00 -7.40255415e-02 -2.01009482e-01
5.14469802e-01 1.15145281e-01 1.05738923e-01 -1.44775137e-01
-3.57471466e-01 1.33229345e-02 1.03107288e-01 -9.34037209e-01
4.61208448e-02 -3.23234856e-01 -4.12592232e-01 8.17247748e-01
2.88109332e-01 7.56597281e-01 -7.32629001e-01 -3.58024895e-01
2.47552961e-01 7.46174872e-01 -5.32993674e-01 1.10431187e-01
-5.03363833e-03 8.06866825e-01 5.85765876e-02 7.21054316e-01
9.27373290e-01 -7.18263835e-02 -3.33908468e-01 -5.17942803e-03
-1.12272866e-01 -1.74255930e-02 -4.79926556e-01 1.91365528e+00
-3.85496318e-01 1.21715939e+00 -7.98496306e-02 -4.25508171e-01
6.35413945e-01 -1.15289157e-02 3.12541902e-01 -9.68311071e-01
2.91072637e-01 -8.93915743e-02 -1.51613533e-01 -5.64276040e-01
5.47871709e-01 4.77242749e-03 2.01180071e-01 6.63145661e-01
1.40483946e-01 1.09155752e-01 -3.20215732e-01 -1.06376037e-02
6.22402549e-01 9.70150173e-01 -1.82715461e-01 -3.80905010e-02
3.89978945e-01 -3.29187155e-01 1.55899227e-01 3.18111479e-01
-7.21090972e-01 8.95447910e-01 3.29366744e-01 -1.33227140e-01
-8.01755667e-01 -9.23101366e-01 -1.83324397e-01 1.81699693e+00
2.07459703e-01 -3.56706493e-02 -1.14523613e+00 -9.71893251e-01
-7.61556923e-02 6.73804820e-01 -1.22311378e+00 -1.44416511e-01
-2.05432355e-01 -5.52904606e-01 5.39467275e-01 5.03442287e-01
4.39806849e-01 -8.68966162e-01 -6.52915776e-01 -7.62749612e-01
4.10736911e-02 -9.53452945e-01 -4.71748114e-01 -1.21525072e-01
-6.41167462e-01 -1.26749980e+00 -8.72228384e-01 -4.72922146e-01
7.04042137e-01 4.85449076e-01 1.08060467e+00 -3.04141194e-01
-2.94550240e-01 8.17321837e-01 -3.46060634e-01 -1.07025325e+00
4.22760434e-02 2.13449374e-01 2.68154651e-01 7.98376650e-02
1.02988923e+00 -2.04618871e-01 -9.82424259e-01 2.72055000e-01
-5.05358458e-01 -4.57679987e-01 2.12454647e-01 6.04844391e-01
3.55247781e-02 -6.63623095e-01 2.73555517e-01 -7.55317569e-01
5.58802783e-01 -4.78495777e-01 -5.33229947e-01 8.96692872e-02
-8.65030110e-01 5.31910583e-02 -4.10003997e-02 -4.03724462e-01
-1.23633671e+00 -2.72492468e-01 3.24234068e-01 -7.41980851e-01
-5.05209029e-01 -4.04171646e-01 1.23602062e-01 -5.24237990e-01
1.12323344e+00 -2.76311845e-01 3.11238468e-01 -2.51857966e-01
1.73630953e-01 8.22690189e-01 1.15362994e-01 -2.83989608e-01
6.78762078e-01 6.56491578e-01 2.32076272e-01 -7.02905178e-01
-1.05439317e+00 -5.93069017e-01 -6.75598383e-01 -2.50510514e-01
1.06781626e+00 -1.07680762e+00 -1.26258516e+00 6.47701085e-01
-6.81011915e-01 -3.68616879e-01 1.21616930e-01 3.68797123e-01
-5.02295315e-01 6.49136901e-02 7.97861144e-02 -1.14408362e+00
-3.60685706e-01 -1.01933372e+00 1.36961007e+00 5.34798741e-01
-3.55227709e-01 -9.77321029e-01 2.21327692e-01 5.27917743e-01
4.60231751e-01 2.80926257e-01 2.65942961e-01 -1.29256666e-01
-4.72987533e-01 7.81834498e-02 -4.29742575e-01 3.48967165e-01
-8.03647470e-03 5.56911603e-02 -1.80428624e+00 -4.23781276e-01
-1.99451447e-01 -7.79938400e-01 6.83287382e-01 6.90505564e-01
1.11411297e+00 1.40643343e-01 -5.03730416e-01 1.06784248e+00
1.19487035e+00 -2.34167203e-01 7.45843053e-01 5.48931062e-01
8.87553632e-01 9.14939344e-01 8.03464711e-01 1.89136758e-01
5.46970487e-01 6.03330374e-01 6.85821533e-01 -1.81023344e-01
-2.71959424e-01 5.50576821e-02 4.06143934e-01 -2.75018990e-01
-3.13118637e-01 -1.92358956e-01 -8.40045691e-01 4.26228851e-01
-1.48988640e+00 -1.05897343e+00 -6.31106496e-02 2.28653908e+00
7.44303226e-01 -3.81570011e-01 5.10754108e-01 -2.31773835e-02
4.63831425e-01 2.89441999e-02 -8.72870445e-01 -5.04599035e-01
1.41825184e-01 5.19955337e-01 5.79412639e-01 1.42735794e-01
-1.17361474e+00 8.29222977e-01 6.63711834e+00 2.27668121e-01
-1.50476336e+00 1.01546124e-01 4.92975891e-01 -7.41128743e-01
1.87429681e-01 -2.99616933e-01 -1.01929724e+00 6.17714286e-01
8.79681230e-01 4.76427525e-01 3.79524261e-01 7.40241587e-01
3.40692848e-02 -5.27888894e-01 -9.94150102e-01 1.29330683e+00
6.82978988e-01 -5.07897377e-01 -6.58887088e-01 2.47158110e-01
7.10343122e-01 2.84048021e-01 1.01981020e+00 3.03386360e-01
-1.41969457e-01 -1.24753785e+00 4.88457322e-01 5.81748247e-01
1.26868618e+00 -7.23859549e-01 3.50168735e-01 -5.26919626e-02
-5.72068572e-01 -3.04666460e-01 -1.71260640e-01 -1.24949262e-01
-5.16657531e-01 -3.30676138e-01 -9.00200248e-01 5.50765246e-02
1.23830521e+00 8.74173164e-01 -8.95357251e-01 7.65052676e-01
-2.74492919e-01 5.92885613e-01 -2.01286048e-01 2.77310144e-02
-2.95159351e-02 4.31148261e-02 3.17380697e-01 8.61971021e-01
3.96184102e-02 -3.14245492e-01 -4.88503695e-01 9.18472767e-01
-1.48389131e-01 -3.33452940e-01 -5.94634652e-01 5.10576904e-01
2.20131189e-01 1.10010934e+00 7.04786181e-02 3.74098450e-01
-4.38661933e-01 1.03573191e+00 3.58069241e-01 7.38153577e-01
-9.07320678e-01 -4.13003564e-01 1.14895785e+00 2.90574849e-01
1.62142694e-01 -6.40015304e-02 -3.18189383e-01 -8.10376346e-01
-3.82400565e-02 -7.25061655e-01 1.98415384e-01 -1.25001097e+00
-1.10528231e+00 3.09322149e-01 1.40829757e-01 -1.28422320e+00
-4.01328981e-01 -8.19333673e-01 -3.89254332e-01 1.42158604e+00
-1.96086657e+00 -1.63097143e+00 -8.19762051e-01 9.64935541e-01
2.86839008e-01 -1.95342287e-01 8.63653481e-01 -1.12398081e-01
-5.74680388e-01 1.12982607e+00 1.83255613e-01 -8.00334662e-02
1.62139201e+00 -1.52113426e+00 2.47848809e-01 7.81163514e-01
-2.70827353e-01 8.88166010e-01 6.65572047e-01 -1.19457625e-01
-1.12826693e+00 -7.98342407e-01 6.01201355e-01 -1.21215737e+00
3.42011243e-01 -5.71812332e-01 -6.61316335e-01 8.39065671e-01
6.30901933e-01 1.08637623e-01 9.81332600e-01 4.72635180e-01
-7.30971217e-01 -3.86637807e-01 -1.34349346e+00 5.38537502e-01
9.33951855e-01 -8.02289546e-01 -2.90586680e-01 2.88728159e-02
2.90752530e-01 -6.47177577e-01 -5.50096929e-01 1.41939268e-01
8.95455837e-01 -1.23741412e+00 8.61930430e-01 -5.47175944e-01
2.74522871e-01 -2.33714148e-01 2.19635412e-01 -1.04464436e+00
2.38265693e-02 -7.88743675e-01 -2.59539992e-01 1.07943606e+00
2.14366671e-02 -6.98744416e-01 1.00455725e+00 9.39480305e-01
3.11144382e-01 -4.16288882e-01 -5.01823843e-01 -3.94121200e-01
1.48602501e-01 -3.00384432e-01 4.40513343e-01 7.86367178e-01
-2.02691928e-01 4.02185410e-01 -4.13053155e-01 1.56899691e-01
1.03026223e+00 -2.60412067e-01 1.23083282e+00 -1.24194145e+00
2.70012438e-01 -3.15730304e-01 -4.99026805e-01 -7.85777628e-01
5.19974113e-01 -2.70024091e-01 -4.14899029e-02 -9.84561861e-01
2.68189944e-02 -3.47564340e-01 -4.01011288e-01 4.76256520e-01
-4.30979073e-01 8.28586638e-01 2.15821728e-01 1.94172472e-01
-7.46840239e-01 2.11891815e-01 1.06654489e+00 -1.29330978e-02
-2.04666004e-01 1.70356616e-01 -8.74866426e-01 3.58343154e-01
7.72950709e-01 -4.16428894e-01 -6.44878447e-01 -4.72667575e-01
4.25208926e-01 -7.85686135e-01 5.00139296e-01 -1.19874024e+00
3.65066469e-01 1.16025195e-01 6.08767569e-01 -4.92087930e-01
6.44154489e-01 -8.06006849e-01 -5.95717072e-01 -1.44200921e-01
-2.71479040e-01 -1.72901452e-01 5.78969121e-01 7.19922304e-01
1.34122878e-01 1.84480235e-01 8.11521947e-01 2.56937414e-01
-7.57855654e-01 3.65538687e-01 2.71085948e-01 2.39065528e-01
1.14128745e+00 -7.01456845e-01 -6.03392243e-01 -4.05903846e-01
-5.03050923e-01 -4.31656204e-02 9.06268537e-01 5.56181431e-01
3.48545462e-01 -7.14649856e-01 -5.02716422e-01 4.50429648e-01
5.45066595e-01 -5.23126870e-02 1.13615289e-01 1.10891247e+00
-1.97894722e-01 2.68318534e-01 -4.94739532e-01 -1.00341177e+00
-1.68060219e+00 4.84287143e-01 4.12066638e-01 5.70893347e-01
7.04837963e-02 1.35339761e+00 4.47819024e-01 -2.35124990e-01
3.79175991e-01 -1.91702530e-01 -4.13168848e-01 1.58907808e-02
7.18955100e-01 5.51957428e-01 -4.53514829e-02 -8.03496957e-01
-4.42839801e-01 9.86587405e-01 -2.05142066e-01 1.67871863e-02
1.23154569e+00 -5.95575035e-01 2.53073256e-02 4.83578026e-01
1.27343524e+00 7.74898082e-02 -1.68739164e+00 -2.56079305e-02
-5.05444527e-01 -5.75126827e-01 2.35211596e-01 -1.03000867e+00
-7.80990839e-01 1.20030785e+00 1.17744339e+00 -1.65495932e-01
1.43205845e+00 -1.40318424e-01 3.36533606e-01 6.75019473e-02
1.45098403e-01 -9.58598554e-01 1.89924255e-01 4.21216995e-01
5.16524076e-01 -1.84753466e+00 1.90007240e-01 -8.47834423e-02
-8.15871835e-01 8.48645508e-01 7.72929490e-01 -4.54164110e-02
7.06719041e-01 -9.56307650e-02 5.43884456e-01 -1.81726784e-01
-2.62115538e-01 -2.85315305e-01 7.12407708e-01 1.21901858e+00
6.14296198e-01 -3.50106329e-01 4.74560142e-01 -2.15464637e-01
-1.10204853e-01 5.79616725e-02 1.84751913e-01 8.10170710e-01
2.43951268e-02 -1.04432142e+00 -5.87132692e-01 2.29303554e-01
-8.35027993e-01 -1.41466081e-01 -4.72008824e-01 9.01463270e-01
1.27908006e-01 1.08757865e+00 2.29365796e-01 -2.67028511e-01
-1.23542175e-02 1.92675829e-01 9.90647495e-01 -5.68896055e-01
-6.11540675e-01 -2.86259592e-01 -3.64860833e-01 -7.84047604e-01
-9.53113735e-01 -8.45070839e-01 -5.70677698e-01 -4.47815776e-01
-4.79413480e-01 -5.52389562e-01 9.95682955e-01 7.35774338e-01
5.08827329e-01 3.55060548e-01 5.23038745e-01 -1.11058176e+00
-3.30541641e-01 -1.17444384e+00 -7.24470556e-01 4.28707689e-01
8.82809341e-01 -9.44105327e-01 -5.28300226e-01 1.97837144e-01] | [14.141717910766602, 0.05177021026611328] |
ee0f51a1-9a01-48ac-8f69-047bcb713ba8 | person-in-wifi-fine-grained-person-perception | 1904.00276 | null | http://arxiv.org/abs/1904.00276v1 | http://arxiv.org/pdf/1904.00276v1.pdf | Person-in-WiFi: Fine-grained Person Perception using WiFi | Fine-grained person perception such as body segmentation and pose estimation
has been achieved with many 2D and 3D sensors such as RGB/depth cameras, radars
(e.g., RF-Pose) and LiDARs. These sensors capture 2D pixels or 3D point clouds
of person bodies with high spatial resolution, such that the existing
Convolutional Neural Networks can be directly applied for perception. In this
paper, we take one step forward to show that fine-grained person perception is
possible even with 1D sensors: WiFi antennas. To our knowledge, this is the
first work to perceive persons with pervasive WiFi devices, which is cheaper
and power efficient than radars and LiDARs, invariant to illumination, and has
little privacy concern comparing to cameras. We used two sets of off-the-shelf
WiFi antennas to acquire signals, i.e., one transmitter set and one receiver
set. Each set contains three antennas lined-up as a regular household WiFi
router. The WiFi signal generated by a transmitter antenna, penetrates through
and reflects on human bodies, furniture and walls, and then superposes at a
receiver antenna as a 1D signal sample (instead of 2D pixels or 3D point
clouds). We developed a deep learning approach that uses annotations on 2D
images, takes the received 1D WiFi signals as inputs, and performs body
segmentation and pose estimation in an end-to-end manner. Experimental results
on over 100000 frames under 16 indoor scenes demonstrate that Person-in-WiFi
achieved person perception comparable to approaches using 2D images. | ['Stanislav Panev', 'Fei Wang', 'Sanping Zhou', 'Dong Huang', 'Jinsong Han'] | 2019-03-30 | person-in-wifi-fine-grained-person-perception-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Wang_Person-in-WiFi_Fine-Grained_Person_Perception_Using_WiFi_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Wang_Person-in-WiFi_Fine-Grained_Person_Perception_Using_WiFi_ICCV_2019_paper.pdf | iccv-2019-10 | ['rf-based-pose-estimation'] | ['computer-vision'] | [ 4.97652829e-01 1.93168595e-01 5.12519181e-01 -2.85653859e-01
-4.51049894e-01 -6.15453243e-01 1.66624248e-01 -3.22886556e-01
-4.45130408e-01 4.70352322e-01 -5.85153466e-03 1.30383894e-01
1.14021957e-01 -1.21799994e+00 -1.06329381e+00 -4.54267740e-01
-2.86589712e-01 6.35593593e-01 9.48344991e-02 1.34934619e-01
-6.00924015e-01 4.31907535e-01 -1.60484755e+00 1.53061226e-01
4.40753907e-01 1.40949905e+00 -1.75011739e-01 8.38357389e-01
3.98993760e-01 -1.99838936e-01 -6.51099205e-01 -6.41361237e-01
5.98365307e-01 3.06803793e-01 -1.64159730e-01 1.91025510e-01
9.45146978e-01 -7.35784352e-01 -3.56110901e-01 6.39803112e-01
8.10118794e-01 -2.84088135e-01 3.20497394e-01 -1.15853715e+00
-2.44365081e-01 2.61695921e-01 -6.34556651e-01 -3.58114451e-01
1.20069718e+00 2.30826735e-01 1.53347790e-01 -6.61908031e-01
-2.78779622e-02 1.51978672e+00 1.24275291e+00 6.64331436e-01
-9.60145354e-01 -6.83372796e-01 2.81448513e-02 -5.34773469e-01
-1.54559100e+00 -2.49968484e-01 5.54764807e-01 -3.47928286e-01
5.64941108e-01 2.71256834e-01 1.00687897e+00 1.71712482e+00
2.22916491e-02 4.96745288e-01 1.19676840e+00 -1.54089183e-01
9.74445567e-02 -3.05578653e-02 4.97647636e-02 8.29510450e-01
6.57619715e-01 2.68419743e-01 -4.68215287e-01 -2.18248859e-01
1.16947353e+00 5.53191960e-01 -4.44605440e-01 -2.65861452e-01
-1.23713970e+00 3.09322029e-01 6.85959697e-01 1.95534714e-02
-5.07951140e-01 5.46220303e-01 -1.27494052e-01 -3.90688516e-02
1.94792882e-01 -2.30280235e-01 -3.41844529e-01 -1.03307001e-01
-8.15188706e-01 3.03453058e-01 6.70536697e-01 1.12158978e+00
5.71418285e-01 -3.26913685e-01 4.49191444e-02 4.40977573e-01
7.64472008e-01 1.45635927e+00 -1.92077123e-02 -8.93693089e-01
7.09657550e-01 2.43165493e-01 3.96417290e-01 -1.07466722e+00
-6.61922932e-01 -4.85649765e-01 -9.57904279e-01 2.11255640e-01
6.24406934e-01 -6.28358245e-01 -1.00716066e+00 1.52810335e+00
4.76576239e-01 3.90066504e-01 -2.84578800e-01 1.46367395e+00
9.62536156e-01 2.46320724e-01 -3.10121268e-01 2.95491815e-01
1.77853596e+00 -4.21309143e-01 -2.56494731e-01 -4.49781328e-01
-1.35969609e-01 -3.08940530e-01 1.00348437e+00 8.03598940e-01
-1.01595712e+00 -9.77977097e-01 -9.39412534e-01 2.25509837e-01
-1.89341411e-01 -2.06995457e-02 7.27843404e-01 1.47680628e+00
-7.92385757e-01 2.35397264e-01 -9.97218609e-01 -6.25527859e-01
6.05456650e-01 7.36749828e-01 -3.38593185e-01 -3.76070738e-01
-1.05195570e+00 1.73016474e-01 -2.46892869e-01 1.97813943e-01
-8.01166177e-01 -5.70148051e-01 -7.65368044e-01 -1.39182180e-01
3.31021905e-01 -1.34809268e+00 1.05767846e+00 -6.79646909e-01
-1.56800115e+00 7.98255682e-01 -5.04965223e-02 -4.62767124e-01
5.14403462e-01 -7.88348794e-01 -4.54233676e-01 2.03617901e-01
9.22230706e-02 4.89246994e-01 9.68887687e-01 -1.37190557e+00
-5.90378940e-01 -9.58650291e-01 3.37368578e-01 2.61809910e-03
-3.31728943e-02 -3.07045728e-01 -6.83314860e-01 -2.10805938e-01
2.29679599e-01 -1.13981295e+00 -1.84871688e-01 1.73719898e-01
-6.45899296e-01 3.07607740e-01 4.84710693e-01 -4.41306770e-01
4.26019937e-01 -2.10253477e+00 -4.00702029e-01 6.25773251e-01
2.35998407e-01 -2.61562783e-02 2.55925238e-01 -4.27888371e-02
4.37860996e-01 -1.46552548e-01 -4.49413918e-02 -6.32450819e-01
3.05349946e-01 4.59343418e-02 4.16090824e-02 6.78445101e-01
-4.59677368e-01 6.45903707e-01 -6.74047410e-01 -3.00086409e-01
4.36090082e-01 9.93470192e-01 -3.63397390e-01 -1.53441802e-02
3.48723859e-01 4.91249591e-01 -6.14155829e-01 1.02249026e+00
1.03289723e+00 2.89091561e-03 -1.18035682e-01 -3.69327486e-01
5.54257445e-02 7.16322586e-02 -1.56473839e+00 1.93457162e+00
-3.91806424e-01 6.71514198e-02 5.05328596e-01 -6.70242012e-01
7.64896929e-01 5.31659424e-01 4.29434836e-01 -6.67577088e-01
2.16233134e-01 -4.05978672e-02 -6.80030346e-01 -4.90779012e-01
3.13021652e-02 -1.12370826e-01 -5.67081749e-01 8.70582014e-02
-1.48445055e-01 2.49493808e-01 -4.20848727e-01 -1.55455261e-01
1.16775846e+00 3.19111079e-01 -1.11568198e-01 1.76344052e-01
3.36096793e-01 -2.55563319e-01 4.19289678e-01 1.10275829e+00
-3.68932001e-02 8.37899327e-01 -5.05236983e-01 -5.09269416e-01
-4.19440866e-01 -1.69313896e+00 1.35439942e-02 8.52607965e-01
4.60252434e-01 -2.64502823e-01 -9.34981763e-01 -5.48288047e-01
4.85669225e-01 6.53693303e-02 -4.54533309e-01 2.59013206e-01
-6.21433198e-01 -4.67164606e-01 9.03582811e-01 7.51549721e-01
9.97184753e-01 -5.09011507e-01 -9.83216465e-01 1.36380449e-01
-1.74503714e-01 -1.26785254e+00 -1.94876194e-01 -1.12180166e-01
-6.02602899e-01 -9.55019295e-01 -8.87206674e-01 -3.23128045e-01
4.77824867e-01 4.09705341e-01 1.04684067e+00 -3.77326310e-01
-2.86257952e-01 1.14757645e+00 -1.67946175e-01 -6.57536983e-01
6.86231971e-01 -3.34047645e-01 5.34915328e-01 2.77423531e-01
4.81617510e-01 -8.48239064e-01 -1.01028943e+00 4.06365484e-01
-1.46572605e-01 -3.27193260e-01 6.10598683e-01 1.28899828e-01
6.54703200e-01 1.63766414e-01 -1.58398673e-01 -4.78750497e-01
2.08154589e-01 -2.14615315e-01 -3.84580523e-01 -2.38281444e-01
9.13679898e-02 -6.62961662e-01 3.89696091e-01 -1.34413123e-01
-9.63613987e-01 4.57592428e-01 -2.60435581e-01 -4.16973919e-01
-9.36749101e-01 -6.45159334e-02 -5.68125665e-01 -4.38085496e-02
6.70863330e-01 -1.58181768e-02 -2.25827441e-01 -6.61126614e-01
3.45870107e-01 8.97648752e-01 9.14638400e-01 -5.60778856e-01
1.00714064e+00 9.59263504e-01 -1.12889288e-03 -9.49084818e-01
-6.07097685e-01 -4.93895292e-01 -7.33297408e-01 -4.75236267e-01
1.10146523e+00 -1.40349460e+00 -1.25241935e+00 6.13019228e-01
-9.66182113e-01 -8.10563341e-02 -6.68743774e-02 6.30288303e-01
-3.46662283e-01 1.93919718e-01 -4.79934543e-01 -1.12326288e+00
-3.69638324e-01 -6.47396326e-01 1.66641533e+00 4.10503894e-01
-3.21679384e-01 -4.82143730e-01 -2.40942448e-01 8.60290647e-01
9.62797403e-02 6.48843527e-01 -2.35110312e-03 1.76765412e-01
-6.72072947e-01 -3.34441721e-01 9.70283672e-02 -2.53962904e-01
4.91886251e-02 -7.58555353e-01 -1.39258325e+00 -4.74700063e-01
-1.09315567e-01 -1.68354690e-01 5.91164649e-01 7.91156769e-01
1.01885080e+00 -1.56821340e-01 -9.22972500e-01 9.33655560e-01
1.19750965e+00 -1.46404907e-01 6.72295928e-01 1.19925253e-01
9.80506241e-01 3.03294897e-01 5.07939517e-01 4.82231408e-01
5.56103528e-01 8.85521591e-01 6.90314591e-01 -3.89675945e-01
-1.15782566e-01 -2.47399837e-01 4.25272077e-01 -1.92409962e-01
-6.85418189e-01 -3.40517730e-01 -6.51660144e-01 1.12952016e-01
-1.53143084e+00 -8.78150702e-01 -4.40848768e-01 2.36030555e+00
1.27492756e-01 2.61797518e-01 5.37233651e-01 2.45416790e-01
6.19632363e-01 -2.79337436e-01 -3.45969528e-01 7.11008161e-02
1.52161703e-01 4.21182901e-01 8.43257785e-01 1.71856299e-01
-1.41746426e+00 3.00590307e-01 5.22348642e+00 4.53574099e-02
-8.77105236e-01 1.50629044e-01 4.40600030e-02 -4.68466192e-01
4.87460122e-02 -6.09624982e-01 -8.59249473e-01 5.70894897e-01
6.82249069e-01 9.97068763e-01 3.57280910e-01 8.83986712e-01
1.58898428e-01 7.53303692e-02 -1.13892496e+00 1.55277169e+00
-5.45300208e-02 -8.06109130e-01 -4.95480150e-01 4.57255304e-01
2.08036512e-01 3.70293446e-02 -1.96914207e-02 1.13203451e-01
4.54239808e-02 -1.09962356e+00 9.05921400e-01 7.52988398e-01
8.72147799e-01 -7.14361429e-01 6.99350893e-01 4.51816201e-01
-1.43953419e+00 -8.62962455e-02 -2.87896842e-01 -3.28233927e-01
2.96077192e-01 8.95636976e-01 -7.56249189e-01 4.92876440e-01
1.45168769e+00 2.50479877e-01 -2.25337207e-01 8.41302574e-01
-1.34739801e-01 5.76742589e-01 -9.63549793e-01 1.48354784e-01
-1.70726329e-01 4.55136299e-02 4.90945071e-01 1.30118072e+00
6.76229358e-01 2.62647778e-01 4.11917418e-01 6.92124903e-01
-1.58652160e-02 -5.84505498e-01 -6.94697380e-01 6.38999462e-01
5.01219392e-01 1.15129745e+00 -5.91879427e-01 -8.94451961e-02
-5.86551309e-01 1.17608905e+00 -3.91026020e-01 2.76541501e-01
-9.28740621e-01 -2.94162035e-01 7.94609845e-01 7.79713750e-01
5.46955347e-01 -4.09109890e-01 -1.16481423e-01 -9.81739223e-01
2.24350095e-01 -6.66215658e-01 1.90788060e-01 -7.03131914e-01
-1.36481130e+00 1.97541028e-01 -2.69448847e-01 -1.11674213e+00
-9.60562564e-03 -7.19605327e-01 -3.23507756e-01 8.50791454e-01
-1.01050699e+00 -1.45878720e+00 -8.79357517e-01 1.20768929e+00
6.45736000e-03 2.47328520e-01 9.79982078e-01 5.54751217e-01
-2.31965303e-01 6.79232061e-01 -2.29736850e-01 4.51650918e-01
3.52200985e-01 -1.22874451e+00 5.17124355e-01 6.76777899e-01
7.87369460e-02 6.67104006e-01 4.70318645e-01 -7.04889655e-01
-1.87517428e+00 -1.02682221e+00 3.80439728e-01 -6.42805219e-01
-3.83822411e-01 -8.65991116e-01 -9.21168625e-02 7.30670750e-01
-1.97721481e-01 1.09859526e-01 8.73663843e-01 2.80724078e-01
-1.57332778e-01 -5.87885439e-01 -1.60280907e+00 2.11872309e-01
1.61592126e+00 -2.46404469e-01 -6.80863798e-01 1.41843617e-01
4.47688282e-01 -5.92436492e-01 -8.58079195e-01 3.34407091e-01
1.14575052e+00 -1.10928953e+00 1.63374221e+00 2.76050806e-01
-4.01856810e-01 -5.40230572e-01 -3.86981606e-01 -9.82776046e-01
-4.75082576e-01 -4.88537878e-01 -2.10531831e-01 1.14282644e+00
-3.62699963e-02 -6.98854089e-01 9.89325762e-01 5.69245398e-01
-3.77077758e-02 -3.05866688e-01 -9.65971649e-01 -6.48535669e-01
-6.10672832e-01 -8.51664603e-01 9.85179722e-01 3.35962504e-01
-4.05214310e-01 1.26295194e-01 -3.97684366e-01 8.84118557e-01
1.20945108e+00 1.67597845e-01 1.15873635e+00 -1.56102216e+00
-4.69321609e-01 2.28071779e-01 -6.65723026e-01 -1.65212965e+00
-5.74672401e-01 -3.01884055e-01 -1.01001943e-02 -1.67167234e+00
-3.92184496e-01 -4.41293746e-01 7.97903985e-02 3.80580068e-01
3.40695709e-01 7.16992795e-01 5.38113154e-02 -1.45816877e-01
-4.29921657e-01 4.55002859e-02 8.32627833e-01 -2.64381111e-01
-8.32163990e-02 5.67853272e-01 -7.07085848e-01 1.04856443e+00
5.38733125e-01 -1.17716871e-01 -1.98185623e-01 -6.71262622e-01
1.38325244e-01 -1.90084777e-03 1.11032295e+00 -1.72336757e+00
1.71546176e-01 3.82707506e-01 1.31707180e+00 -6.76234007e-01
9.72920775e-01 -1.20698190e+00 4.27766293e-01 4.83658075e-01
4.05891150e-01 -4.72391635e-01 1.40481461e-02 6.23261631e-01
4.20996487e-01 4.94200140e-01 3.75598967e-01 -5.48983037e-01
-3.64122987e-01 4.69508350e-01 -3.31083238e-01 -3.69312912e-01
7.38924623e-01 -8.03962171e-01 -7.53527656e-02 -4.41070378e-01
-7.96091735e-01 3.07013206e-02 3.96029800e-01 3.34742814e-01
8.57573152e-01 -1.36219704e+00 -5.07169425e-01 5.46289444e-01
-2.60608643e-01 2.76250720e-01 1.60844833e-01 5.47798455e-01
-1.91026688e-01 5.92835069e-01 -9.96403620e-02 -1.11334348e+00
-1.38317883e+00 3.01368535e-01 3.90393823e-01 2.24674433e-01
-8.56302381e-01 1.02115357e+00 1.54946312e-01 -5.97462893e-01
4.66156483e-01 -5.56990623e-01 4.86436747e-02 -2.62243867e-01
6.62643552e-01 3.19855899e-01 -6.45467937e-02 -6.67174637e-01
-8.37096870e-01 1.29689455e+00 5.77285349e-01 -1.80788144e-01
1.09708512e+00 -4.16822582e-01 5.16206086e-01 7.06943646e-02
7.24260867e-01 3.29645574e-01 -1.41842544e+00 -3.70807834e-02
-6.71665490e-01 -6.34691000e-01 -1.41637817e-01 -9.54398692e-01
-1.05174553e+00 8.32209110e-01 1.15032864e+00 4.07426089e-01
1.18852127e+00 2.32399538e-01 1.13247979e+00 2.15197951e-01
1.19030404e+00 -7.35132694e-01 -2.93683946e-01 1.59842372e-01
3.50876302e-01 -9.17235017e-01 -2.00920030e-01 -6.24509394e-01
-2.16943659e-02 8.57082665e-01 5.10083497e-01 -9.50983316e-02
5.38966715e-01 4.79964107e-01 4.25721407e-02 -4.03998286e-01
2.08053872e-01 -4.33600456e-01 1.78986475e-01 1.42129350e+00
1.26572475e-01 4.54525828e-01 5.93626678e-01 1.03178787e+00
-6.28004193e-01 2.49686375e-01 -1.66144744e-01 8.78286541e-01
-5.37108421e-01 -6.69245005e-01 -1.17353773e+00 4.24675435e-01
-5.07041216e-01 3.08640659e-01 -2.36500755e-01 6.57402396e-01
1.05436814e+00 1.17814267e+00 6.62790611e-02 -4.75596070e-01
8.10244977e-01 -2.05973208e-01 9.36058104e-01 -3.82170975e-01
-5.72174072e-01 1.30196854e-01 1.92960650e-01 -9.32260573e-01
-4.66687083e-01 -7.77068913e-01 -1.16747034e+00 -2.35541105e-01
1.64523020e-01 -2.34136090e-01 6.99233651e-01 9.02541757e-01
3.72306406e-01 4.25345957e-01 3.18884075e-01 -1.28706801e+00
-5.09232655e-02 -6.69890225e-01 -7.46480405e-01 3.18483561e-01
4.36432302e-01 -7.37487972e-01 -4.50950339e-02 8.52204189e-02] | [6.836877346038818, 0.3930681347846985] |
c86d4ee3-275e-4f17-b771-66971aaee6ff | open-set-recognition-of-breast-cancer | 2201.02923 | null | https://arxiv.org/abs/2201.02923v1 | https://arxiv.org/pdf/2201.02923v1.pdf | Open-Set Recognition of Breast Cancer Treatments | Open-set recognition generalizes a classification task by classifying test samples as one of the known classes from training or "unknown." As novel cancer drug cocktails with improved treatment are continually discovered, predicting cancer treatments can naturally be formulated in terms of an open-set recognition problem. Drawbacks, due to modeling unknown samples during training, arise from straightforward implementations of prior work in healthcare open-set learning. Accordingly, we reframe the problem methodology and apply a recent existing Gaussian mixture variational autoencoder model, which achieves state-of-the-art results for image datasets, to breast cancer patient data. Not only do we obtain more accurate and robust classification results, with a 24.5% average F1 increase compared to a recent method, but we also reexamine open-set recognition in terms of deployability to a clinical setting. | ['Yuan Luo', 'Diego Klabjan', 'Alexander Cao'] | 2022-01-09 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [ 6.09770358e-01 3.04672033e-01 -4.39977199e-01 -2.43802696e-01
-1.23479903e+00 -4.70280528e-01 5.55273294e-01 1.50302708e-01
-2.51609176e-01 1.00725281e+00 4.81699891e-02 -4.54417080e-01
-2.04269215e-01 -6.66323364e-01 -8.63859832e-01 -9.64594722e-01
3.17114264e-01 8.79005134e-01 -5.05951345e-01 1.91601038e-01
-1.76990420e-01 2.32423395e-01 -1.53204024e+00 5.87730467e-01
7.41235316e-01 1.00290358e+00 -1.27725257e-02 5.69739878e-01
-1.27283841e-01 5.36115885e-01 -4.78994191e-01 -2.17801452e-01
-7.04460442e-02 -2.81645447e-01 -9.74193394e-01 4.20523494e-01
4.37425941e-01 -1.85312614e-01 -3.06697100e-01 9.20578241e-01
4.58007783e-01 -7.88099319e-02 1.17514729e+00 -9.89978254e-01
-9.42857683e-01 1.22195661e-01 -1.60208732e-01 -9.30650309e-02
1.23485662e-01 7.27749988e-02 7.29574263e-01 -7.38409281e-01
7.13929176e-01 5.93293190e-01 7.27230072e-01 1.22296000e+00
-1.46653807e+00 -4.15519536e-01 3.72325815e-02 7.81752244e-02
-1.35567379e+00 -4.61164415e-01 2.11108223e-01 -7.93063700e-01
8.07970703e-01 6.45119131e-01 6.15592122e-01 1.77788019e+00
4.65639740e-01 9.98639584e-01 1.09479702e+00 -2.63337016e-01
4.70205784e-01 4.11221385e-01 3.33945304e-01 4.43146735e-01
2.56446749e-01 3.12251657e-01 2.86305770e-02 -4.27650243e-01
6.26314282e-01 5.25477171e-01 -7.51821578e-01 -4.37882006e-01
-1.05032885e+00 1.05591118e+00 3.28720540e-01 4.26303715e-01
-4.38131332e-01 -9.43444222e-02 2.39433601e-01 3.57378311e-02
7.11760521e-01 3.56997609e-01 -6.58559561e-01 -6.41202480e-02
-1.01746321e+00 -5.07149436e-02 9.58201468e-01 6.91633224e-01
4.09660906e-01 -6.48466572e-02 -4.36269224e-01 7.86957145e-01
4.14388120e-01 4.15405244e-01 8.31172228e-01 -6.02743506e-01
-1.85157701e-01 5.34916937e-01 -1.41362175e-01 -5.11425555e-01
-4.32536423e-01 -7.88020968e-01 -1.01141131e+00 -1.15577787e-01
3.08301449e-01 -5.95159978e-02 -1.24437547e+00 1.29257059e+00
3.97593856e-01 4.65864807e-01 5.96960962e-01 4.39554036e-01
1.14480364e+00 4.54406172e-01 -7.55826607e-02 -5.40434480e-01
1.33403146e+00 -7.72381186e-01 -9.25591350e-01 7.12123513e-03
8.14662814e-01 -2.16518730e-01 5.13272524e-01 5.96829891e-01
-4.95017439e-01 -2.27225408e-01 -9.55145657e-01 1.72946766e-01
-5.01021326e-01 -1.07280277e-01 8.66743684e-01 8.80800843e-01
-7.88031340e-01 8.05789888e-01 -1.01955223e+00 -3.25027168e-01
1.07589388e+00 4.98946697e-01 -3.69001746e-01 -3.26002568e-01
-6.94495559e-01 7.65422106e-01 1.65624723e-01 -1.31723478e-01
-1.14976466e+00 -1.13231778e+00 -1.01838410e+00 5.25066480e-02
3.21157098e-01 -1.02193415e+00 1.14069653e+00 -8.93949807e-01
-1.44303703e+00 8.62705946e-01 -2.13441193e-01 -3.83208841e-01
3.64587039e-01 1.19868584e-01 -5.63060641e-01 -8.79378542e-02
-1.82292804e-01 5.30567050e-01 7.59597003e-01 -1.24764109e+00
-3.14845562e-01 -5.49781203e-01 -4.38625097e-01 -2.45932311e-01
-5.65259039e-01 -5.34561157e-01 4.41806093e-02 -4.11372989e-01
-2.33593285e-02 -1.04871368e+00 -5.51880538e-01 1.96869913e-02
-4.19174224e-01 -2.45147392e-01 6.59526825e-01 -4.86785799e-01
1.07633698e+00 -1.93268216e+00 2.34001309e-01 2.11650077e-02
6.28655791e-01 1.28150076e-01 6.36931881e-02 3.07160541e-02
-3.59025896e-01 2.00141132e-01 -5.21775544e-01 -2.24710986e-01
-1.69072419e-01 5.57303011e-01 -7.55784363e-02 6.40153110e-01
2.85312533e-01 9.44284379e-01 -8.94811630e-01 -2.73352623e-01
1.40426740e-01 5.63885808e-01 -6.31562471e-01 1.07218197e-03
-3.31816763e-01 5.62167406e-01 -3.95350665e-01 9.62787092e-01
5.84614575e-01 -7.47680068e-01 1.44988954e-01 2.87437551e-02
3.09776753e-01 -3.40678662e-01 -7.77112126e-01 1.57307804e+00
-1.31944194e-01 3.49305242e-01 -9.26572382e-02 -1.31146884e+00
5.46651185e-01 5.93182504e-01 9.19418573e-01 -2.37921894e-01
4.44182694e-01 1.17900513e-01 3.71712958e-03 -5.86748004e-01
5.44395763e-03 -3.12096477e-01 1.98588774e-01 -8.63969177e-02
3.52057755e-01 1.81610867e-01 -1.57255158e-01 -2.37443715e-01
1.20907521e+00 -3.64661179e-02 3.48417491e-01 -4.34769303e-01
2.88573444e-01 2.55710483e-01 3.97989243e-01 7.52370179e-01
-3.77745152e-01 6.27996564e-01 1.74821541e-01 -4.81909513e-01
-6.23290360e-01 -1.10278821e+00 -9.84690726e-01 4.29672867e-01
-5.52559853e-01 -1.30849555e-01 -6.54753864e-01 -9.24140692e-01
2.58490771e-01 4.50491220e-01 -9.97989595e-01 -1.38814867e-01
4.30742688e-02 -1.10832584e+00 4.20999080e-01 5.93622029e-01
-2.79997230e-01 -6.06257200e-01 -2.23076001e-01 2.24872351e-01
1.06650546e-01 -8.89472663e-01 -1.69665292e-01 4.23652947e-01
-8.87985766e-01 -1.45796764e+00 -1.15473914e+00 -7.75353253e-01
6.13611162e-01 -1.61226034e-01 1.06437027e+00 -1.55661345e-01
-5.75417399e-01 5.84734619e-01 -2.63047785e-01 -8.43783617e-01
-7.84018338e-01 -2.51359195e-01 4.11446631e-01 2.24065080e-01
5.75431943e-01 -9.11399350e-02 -5.78283727e-01 -1.64388586e-02
-8.19834769e-01 -2.51652449e-01 6.02582574e-01 1.39856100e+00
1.06357646e+00 -1.73414618e-01 7.19965339e-01 -1.17067182e+00
2.64516532e-01 -8.91990423e-01 -1.57421485e-01 1.88110799e-01
-8.96771967e-01 -5.78086413e-02 4.82164204e-01 -6.69223309e-01
-6.51309371e-01 4.55247536e-02 -2.48642057e-01 -8.84981751e-01
-5.91398537e-01 6.40498042e-01 4.82489057e-02 1.06224574e-01
8.27178180e-01 4.25924838e-01 5.05408823e-01 -2.31092930e-01
2.38382265e-01 1.02123165e+00 3.12016487e-01 -2.91925251e-01
2.80339807e-01 7.72489667e-01 -8.76767039e-02 -8.92912328e-01
-9.59415793e-01 -5.60689390e-01 -4.18826848e-01 1.04565516e-01
1.05491340e+00 -1.00991631e+00 -9.43527281e-01 1.80421710e-01
-6.88447475e-01 -2.90748984e-01 -4.44477379e-01 5.52981377e-01
-6.88752353e-01 2.75149465e-01 -6.17302537e-01 -5.90978622e-01
-4.87640560e-01 -1.37540710e+00 1.09903073e+00 8.87864009e-02
-3.09401751e-01 -1.23958230e+00 2.55187213e-01 6.40787303e-01
2.30476737e-01 3.60557556e-01 7.65277088e-01 -1.18360555e+00
-3.54570955e-01 -3.87052625e-01 1.97414592e-01 4.06212330e-01
3.07198048e-01 -2.85791188e-01 -1.34961343e+00 -5.20950258e-01
1.31937012e-01 -4.36086863e-01 9.31819975e-01 7.09549725e-01
1.47033226e+00 -3.32718194e-01 -8.18818390e-01 7.78369546e-01
1.41311860e+00 1.97073057e-01 6.80457056e-01 -5.90903778e-03
6.34768844e-01 2.16173679e-01 3.68292853e-02 4.58997846e-01
2.98786610e-02 2.40203589e-01 2.63296843e-01 -1.64808795e-01
1.63145423e-01 3.04913335e-02 1.22043073e-01 3.91489267e-01
4.19540972e-01 -1.65721819e-01 -7.88961649e-01 4.46455300e-01
-1.70603466e+00 -1.05911064e+00 -2.36952871e-01 2.15472126e+00
9.04018998e-01 -1.61745757e-01 -1.18146844e-01 4.25608158e-02
3.74129802e-01 -2.67066896e-01 -7.42965758e-01 -3.82252969e-02
-4.37161922e-02 5.62219203e-01 3.24987978e-01 1.47129238e-01
-1.17566204e+00 4.30918276e-01 7.01312780e+00 9.67217147e-01
-1.15725350e+00 2.17648104e-01 1.00121677e+00 -2.01558068e-01
-3.09349030e-01 -3.37616354e-01 -7.06878126e-01 2.72680551e-01
1.32505095e+00 3.11173081e-01 9.29540470e-02 8.70337963e-01
-1.13514557e-01 2.48094812e-01 -1.71178865e+00 1.20940518e+00
3.72136712e-01 -1.71409023e+00 4.16936614e-02 5.49633980e-01
8.13838840e-01 1.49811327e-01 2.14409336e-01 6.29378319e-01
7.98468888e-02 -1.41373169e+00 6.96818158e-02 6.82894170e-01
8.48010242e-01 -3.85131031e-01 7.41684735e-01 5.18435419e-01
-4.12920982e-01 -2.02306017e-01 -1.89891592e-01 5.07282130e-02
-3.75904649e-01 5.28139234e-01 -9.98329461e-01 5.98792613e-01
5.85456908e-01 9.80966330e-01 -3.10076743e-01 8.92769873e-01
4.64866936e-01 6.90868735e-01 -9.31534693e-02 -5.27137741e-02
-2.15387642e-02 -1.34400725e-01 2.45001972e-01 7.94022202e-01
2.06566483e-01 9.88456756e-02 4.70770687e-01 1.01073647e+00
-3.20163593e-02 6.75598858e-03 -7.68646955e-01 -3.89868587e-01
1.01268361e-03 1.10542989e+00 -4.07183647e-01 -5.02930582e-01
-4.76638138e-01 1.09475398e+00 1.05104566e-01 2.76843250e-01
-6.27389848e-01 9.74023417e-02 7.12970555e-01 3.58307082e-03
3.16870093e-01 5.11287391e-01 -1.95433319e-01 -1.31534958e+00
-2.24181712e-01 -9.73971248e-01 7.28519201e-01 -3.81624490e-01
-1.53750622e+00 5.18687189e-01 -3.13922495e-01 -1.35401487e+00
-1.44955993e-01 -1.14122415e+00 -1.26431823e-01 5.56898952e-01
-1.32354474e+00 -9.87581670e-01 5.76405844e-04 4.22122627e-01
6.17256820e-01 -3.33335757e-01 1.60358536e+00 1.67429119e-01
-8.15962970e-01 8.15008700e-01 9.35045183e-01 1.40794322e-01
4.19171929e-01 -1.22271180e+00 -3.30995530e-01 3.29655290e-01
2.56026298e-01 4.16132778e-01 5.56851149e-01 -3.99864644e-01
-1.46168780e+00 -1.13681889e+00 3.41288328e-01 -1.11911869e+00
6.12374723e-01 -2.65883416e-01 -9.47580278e-01 7.60154843e-01
2.56896000e-02 3.90537739e-01 1.42937720e+00 3.94703329e-01
-2.36243650e-01 2.39686489e-01 -1.30552256e+00 3.69379222e-01
6.46663427e-01 -3.80846798e-01 -4.60016489e-01 6.62364066e-01
5.75563073e-01 -5.37658632e-01 -1.50412560e+00 4.94476497e-01
4.22601849e-01 -3.61063480e-01 1.00277531e+00 -1.13909411e+00
6.39235735e-01 1.45489290e-01 -4.48524773e-01 -1.30269945e+00
-4.33678478e-01 -3.38476211e-01 -4.46481168e-01 6.78830147e-01
5.39045036e-01 -6.21083140e-01 1.19304776e+00 6.61510110e-01
-2.49128953e-01 -1.24516749e+00 -1.18437076e+00 -5.94910741e-01
4.82732505e-01 -3.59780014e-01 2.00770333e-01 1.32738781e+00
3.07434320e-01 3.13153505e-01 -2.07417831e-01 1.15776874e-01
5.86476684e-01 8.80206898e-02 4.05354381e-01 -1.54750156e+00
-6.96040571e-01 -4.20512646e-01 -4.36608285e-01 -6.08426869e-01
3.50722015e-01 -1.19109952e+00 -8.15881342e-02 -1.40650272e+00
6.87244534e-01 -2.71402746e-01 -4.71242666e-01 4.66021806e-01
-2.19585925e-01 1.74303919e-01 -2.60040075e-01 1.78300917e-01
-3.89075428e-01 5.87503016e-01 1.02430427e+00 -7.45362997e-01
-1.87542200e-01 2.16553405e-01 -9.95042324e-01 5.04984975e-01
3.59180331e-01 -4.58579838e-01 -2.27258563e-01 -6.26708344e-02
-3.58961552e-01 4.40107912e-01 3.53751451e-01 -9.59910512e-01
-1.51741371e-01 -2.26467237e-01 7.30159342e-01 -3.66886288e-01
4.25852001e-01 -8.09401751e-01 2.95747906e-01 7.18229890e-01
-1.68910161e-01 -6.60250664e-01 3.86026025e-01 1.06057167e+00
1.46589488e-01 -2.64166415e-01 5.83490074e-01 -1.27192050e-01
-5.16816854e-01 5.91774762e-01 -5.65405011e-01 -2.06550598e-01
1.19140553e+00 -5.04741549e-01 -2.65879780e-01 -2.75666872e-03
-1.17027020e+00 1.29029855e-01 2.29206994e-01 3.89742762e-01
6.10256732e-01 -1.08822024e+00 -8.00171256e-01 2.56229699e-01
5.55976927e-01 1.79959148e-01 6.50223553e-01 9.29219365e-01
-1.43423289e-01 6.01483822e-01 2.79017270e-01 -9.25060630e-01
-1.09000182e+00 9.34862256e-01 5.77807128e-01 -3.25861424e-01
-5.70307136e-01 8.10942709e-01 3.62991810e-01 -6.83445156e-01
2.66690552e-01 -2.70242095e-01 -2.55920053e-01 -1.09783716e-01
6.32717669e-01 -4.81259748e-02 2.82156378e-01 -4.85331655e-01
-3.61798555e-01 6.92993477e-02 -3.46984863e-01 4.94340062e-01
1.43170846e+00 4.48118776e-01 2.94482857e-01 5.25985241e-01
1.55910671e+00 -5.20504415e-01 -8.42308283e-01 -8.82228762e-02
-4.12702233e-01 7.53929652e-03 3.66608143e-01 -8.80217433e-01
-7.69146323e-01 5.58003843e-01 1.16807961e+00 -2.01869428e-01
6.81951523e-01 3.06293994e-01 4.88547981e-01 5.52205384e-01
1.68110237e-01 -7.94908643e-01 -7.24844187e-02 2.32672974e-01
6.13163173e-01 -1.72521293e+00 -2.92388257e-02 -2.74545431e-01
-4.20364648e-01 9.42182600e-01 3.70457381e-01 6.65716901e-02
9.61833119e-01 1.84743732e-01 2.50624329e-01 -4.47609127e-01
-8.35442781e-01 8.11820384e-03 3.03543091e-01 8.13117862e-01
5.37537932e-01 3.85241419e-01 6.35189712e-02 1.05985224e+00
1.22704245e-01 3.75063717e-01 5.61796129e-01 8.74059856e-01
-2.16281995e-01 -9.68681633e-01 -4.40471143e-01 1.08935618e+00
-5.44170916e-01 1.33438930e-02 -4.68696244e-02 6.45219862e-01
-7.42494911e-02 8.22278500e-01 2.61286318e-01 -3.49658459e-01
1.13550305e-01 3.06347698e-01 4.73341644e-01 -7.81877279e-01
-3.01596791e-01 5.81151713e-03 -2.20794141e-01 -3.69363666e-01
-2.80543894e-01 -6.98281705e-01 -1.03427839e+00 1.52832538e-01
-6.62410319e-01 1.70800060e-01 5.09961605e-01 1.10710013e+00
6.38851345e-01 8.95079255e-01 3.41861457e-01 -6.51313901e-01
-1.03148842e+00 -9.75577354e-01 -5.98997772e-01 4.63008016e-01
4.87037092e-01 -6.67720795e-01 -1.86765030e-01 3.88960615e-02] | [6.125430107116699, 5.77226448059082] |
6e0ff0ad-551f-44a9-a02b-028bf7867584 | enabling-my-robot-to-play-pictionary | 1608.03369 | null | http://arxiv.org/abs/1608.03369v1 | http://arxiv.org/pdf/1608.03369v1.pdf | Enabling My Robot To Play Pictionary : Recurrent Neural Networks For Sketch Recognition | Freehand sketching is an inherently sequential process. Yet, most approaches
for hand-drawn sketch recognition either ignore this sequential aspect or
exploit it in an ad-hoc manner. In our work, we propose a recurrent neural
network architecture for sketch object recognition which exploits the long-term
sequential and structural regularities in stroke data in a scalable manner.
Specifically, we introduce a Gated Recurrent Unit based framework which
leverages deep sketch features and weighted per-timestep loss to achieve
state-of-the-art results on a large database of freehand object sketches across
a large number of object categories. The inherently online nature of our
framework is especially suited for on-the-fly recognition of objects as they
are being drawn. Thus, our framework can enable interesting applications such
as camera-equipped robots playing the popular party game Pictionary with human
players and generating sparsified yet recognizable sketches of objects. | ['Babu R. Venkatesh', 'Ravi Kiran Sarvadevabhatla', 'Jogendra Kundu'] | 2016-08-11 | null | null | null | null | ['sketch-recognition'] | ['computer-vision'] | [ 1.40827954e-01 -5.20457923e-01 -2.90724903e-01 -3.40803117e-01
-3.32640201e-01 -6.88005924e-01 7.45529652e-01 -3.02170753e-01
-2.97751874e-01 1.72802210e-01 9.70445946e-02 -2.60151207e-01
-1.66291386e-01 -8.36552978e-01 -5.61558902e-01 -3.11839819e-01
-2.02506594e-02 5.88702142e-01 2.17626587e-01 -2.43440211e-01
5.86064160e-01 1.11785877e+00 -1.58468151e+00 4.21131581e-01
2.13356614e-01 8.80921423e-01 3.54253888e-01 1.04238236e+00
-4.15253878e-01 1.00921583e+00 -3.89869571e-01 -4.69574898e-01
5.06892025e-01 -1.62094876e-01 -2.38977417e-01 -4.57501709e-02
8.12788785e-01 -7.44727910e-01 -1.00915837e+00 6.49165273e-01
4.88660306e-01 2.99612880e-01 6.64420068e-01 -1.07872570e+00
-6.69666648e-01 5.22457480e-01 -2.45108485e-01 -1.27069175e-01
1.81774676e-01 3.53024513e-01 1.20179737e+00 -1.24099827e+00
8.46448839e-01 1.09981084e+00 7.40656257e-01 6.66948795e-01
-1.17507303e+00 -8.68291259e-01 5.36886513e-01 -3.64561044e-02
-1.05612481e+00 -4.07792062e-01 1.09904003e+00 -2.04576880e-01
9.98100638e-01 1.14002682e-01 8.92234564e-01 1.33127713e+00
-1.96602941e-01 1.43734920e+00 6.68925822e-01 -8.27364773e-02
2.19129816e-01 -4.62563187e-01 -1.09245010e-01 1.05791879e+00
-1.12221405e-01 1.55089766e-01 -7.36307561e-01 -9.59027559e-02
1.62018919e+00 7.31947541e-01 2.88148373e-01 -6.33632779e-01
-1.39049923e+00 6.80374384e-01 4.25677598e-01 1.89246297e-01
-5.03708780e-01 8.23648036e-01 6.00880444e-01 4.42463905e-01
4.73864786e-02 4.45097178e-01 -1.36359394e-01 -3.63013148e-01
-1.40771437e+00 5.60128331e-01 8.35132420e-01 1.10733068e+00
2.97405332e-01 3.45286787e-01 -2.72832990e-01 8.80756319e-01
-4.89102025e-03 5.54995835e-01 3.38668168e-01 -8.35987866e-01
4.56728637e-01 4.92623091e-01 -3.01073901e-02 -9.67942536e-01
1.08319131e-04 6.12724088e-02 -9.14732397e-01 5.42694092e-01
2.78940052e-01 2.11127430e-01 -1.16732252e+00 1.45649409e+00
-9.07743201e-02 1.43960565e-01 -2.83087611e-01 9.61836934e-01
8.55321646e-01 5.19290030e-01 8.61336011e-03 5.33912838e-01
1.02249062e+00 -1.11799812e+00 -5.17500699e-01 7.37342015e-02
-2.43627176e-01 -5.10447323e-01 1.31830370e+00 5.84383845e-01
-1.12015271e+00 -6.46978378e-01 -1.14745069e+00 -5.02046824e-01
-3.92172635e-01 5.37406325e-01 9.90685523e-01 3.02355707e-01
-5.11289358e-01 1.10187316e+00 -9.58196580e-01 -2.44575784e-01
8.47208619e-01 2.63710231e-01 -1.84684157e-01 7.29926280e-04
-4.71548587e-01 5.40403605e-01 -1.58269987e-01 3.67940396e-01
-1.06247163e+00 -7.14941859e-01 -6.22093976e-01 2.29048371e-01
3.77826989e-01 -6.14926934e-01 1.25603151e+00 -7.02786922e-01
-1.98314798e+00 6.38265431e-01 3.85495834e-02 -2.84463882e-01
9.82113481e-01 -3.10841113e-01 -1.88738424e-02 5.21220639e-02
-3.72922331e-01 7.02238441e-01 1.41418755e+00 -1.01730776e+00
-3.72177571e-01 -3.20923090e-01 5.68570569e-02 -1.34374397e-02
-2.01906905e-01 -1.11136384e-01 -4.33965504e-01 -1.27552187e+00
1.67508066e-01 -8.04160655e-01 -3.15791279e-01 7.64353573e-01
-3.22044045e-01 -3.67266268e-01 1.02936780e+00 -3.63540649e-01
1.04476774e+00 -1.91364551e+00 3.99831861e-01 2.05843985e-01
3.34254444e-01 3.73446643e-01 -4.29026693e-01 8.04841578e-01
3.02716702e-01 -2.29657426e-01 -1.23366728e-01 -4.94962662e-01
4.00927246e-01 2.25827366e-01 -1.00980008e+00 1.13053799e-01
3.51159841e-01 1.37116694e+00 -9.91402507e-01 -1.01488747e-01
3.21478009e-01 4.29350764e-01 -4.42692965e-01 5.80723584e-01
-4.21054721e-01 -9.79648456e-02 -5.47598660e-01 8.86157751e-01
2.37387449e-01 -1.81126148e-02 2.79478997e-01 2.70795953e-02
-1.29247293e-01 1.73406422e-01 -1.13742459e+00 2.10955524e+00
-5.95216632e-01 8.37800980e-01 3.31729837e-03 -6.46577656e-01
1.10727906e+00 1.21522367e-01 1.77720264e-01 -5.37849247e-01
-2.08413377e-01 2.76542425e-01 -3.61985922e-01 -2.06039816e-01
9.01904285e-01 -1.32939219e-01 -1.62897427e-02 8.68358612e-01
7.57012144e-02 -3.99996728e-01 -3.63410749e-02 1.97970346e-01
1.22629261e+00 4.75264162e-01 2.17076465e-01 8.49459618e-02
1.80943962e-02 -4.20557231e-01 -2.90055219e-02 1.08894765e+00
3.93630937e-02 9.02640224e-01 4.28402185e-01 -1.00133014e+00
-1.46628404e+00 -1.16357553e+00 3.44306380e-01 1.32956278e+00
-1.59501657e-01 -2.06410185e-01 -1.07845636e-02 -6.90862536e-01
4.84738737e-01 1.98812529e-01 -7.09129751e-01 1.57621101e-01
-8.61792028e-01 2.40941301e-01 7.57428467e-01 9.91028190e-01
2.69617975e-01 -1.75225568e+00 -9.94206965e-01 5.22515595e-01
7.18857586e-01 -8.73276234e-01 -6.95713520e-01 7.13185966e-02
-8.65733683e-01 -8.84321630e-01 -1.29828894e+00 -7.49290109e-01
6.02167547e-01 3.66325498e-01 9.35745060e-01 1.83293805e-01
-5.65425217e-01 5.83313227e-01 -3.00835259e-02 -4.41184580e-01
-7.75193423e-02 2.06459746e-01 -1.67924364e-03 8.92172456e-02
-5.92655726e-02 -9.17483926e-01 -5.94951153e-01 1.35901244e-02
-7.83320248e-01 3.18903960e-02 8.16234410e-01 1.06458533e+00
4.54004288e-01 -5.31489491e-01 5.77336431e-01 -8.77063751e-01
8.93410206e-01 8.43264442e-03 -5.99343181e-01 4.09919500e-01
-2.23733738e-01 2.15222672e-01 8.21271837e-01 -1.00433195e+00
-7.96322048e-01 3.21390808e-01 -5.05750068e-03 -1.09368050e+00
1.11261143e-04 9.43475813e-02 2.35563070e-01 -2.09601775e-01
2.56667137e-01 4.13986772e-01 -1.04040675e-01 -5.95100880e-01
5.89076340e-01 3.34930122e-01 8.16261649e-01 -9.36659336e-01
8.63664031e-01 5.67922056e-01 -3.76262097e-03 -9.55646098e-01
-3.27231616e-01 -2.95019478e-01 -7.61117339e-01 -2.27497935e-01
3.20698529e-01 -6.14571750e-01 -1.16914284e+00 5.90784132e-01
-1.21472013e+00 -7.74651885e-01 -5.39779186e-01 -1.35298893e-01
-7.21715510e-01 3.22780788e-01 -6.92389131e-01 -1.12204099e+00
-7.72967935e-01 -8.29339087e-01 1.37242877e+00 8.11329857e-02
-2.68979281e-01 -5.55155635e-01 5.48054930e-03 -3.30099106e-01
5.92274070e-01 5.01609780e-02 9.13077533e-01 -3.67553890e-01
-9.65088308e-01 -5.36907017e-01 -5.18533885e-01 3.57806720e-02
2.47868225e-02 2.63326555e-01 -8.63432705e-01 -2.19280511e-01
-6.68283641e-01 -6.44735396e-01 1.05699384e+00 -5.91716822e-03
1.64573896e+00 -3.63260359e-01 2.52457932e-02 6.15052640e-01
1.22155333e+00 -8.76263157e-02 5.66511154e-01 -2.50290632e-01
9.01036322e-01 3.15565228e-01 2.01769367e-01 6.70350313e-01
1.28828570e-01 6.13874495e-01 7.52286389e-02 1.10461339e-01
-1.59352452e-01 -9.48747456e-01 8.85722935e-02 7.47409821e-01
-3.10385287e-01 -6.29242882e-02 -6.73104107e-01 5.89382887e-01
-1.90316403e+00 -1.25973582e+00 4.53671783e-01 2.09851432e+00
5.27061641e-01 2.30283529e-01 3.07113737e-01 4.36973311e-02
4.23503876e-01 5.82340658e-01 -7.59020627e-01 -4.93011475e-01
9.32970867e-02 6.13490641e-01 1.47687286e-01 4.77757826e-02
-9.33338344e-01 1.19472027e+00 6.37053108e+00 7.00959742e-01
-1.34670293e+00 -4.37397301e-01 1.07055314e-01 -2.46847942e-01
-4.18741792e-01 -1.82147190e-01 -4.90100950e-01 2.73196459e-01
2.35688627e-01 4.66629118e-02 8.59796286e-01 9.23083425e-01
-3.38340312e-01 2.98561960e-01 -1.40423357e+00 1.28866363e+00
-3.73235904e-02 -1.74152827e+00 3.85983288e-01 -7.33582824e-02
5.26666462e-01 -1.11487463e-01 3.03505510e-01 4.16486830e-01
3.51551771e-01 -1.16758454e+00 9.73366141e-01 7.70652950e-01
1.17892420e+00 -5.20595968e-01 1.22854952e-03 2.05480039e-01
-1.38720834e+00 -2.27000803e-01 -2.83295423e-01 -2.37566799e-01
3.50747220e-02 1.54671252e-01 -5.52071393e-01 1.07174389e-01
3.12624723e-01 8.95079255e-01 -3.71844262e-01 8.67688715e-01
-1.73342779e-01 2.21289590e-01 -3.99181694e-01 -4.36892897e-01
5.51767826e-01 -1.36916533e-01 4.64701444e-01 1.43274438e+00
1.70904696e-01 3.13537717e-01 8.78199115e-02 1.10391939e+00
-4.18147206e-01 -2.97233254e-01 -9.95215893e-01 -6.53690279e-01
5.41448057e-01 1.18284512e+00 -9.40313399e-01 -4.91592735e-01
-1.91451818e-01 1.28034472e+00 6.57359958e-01 4.13552254e-01
-4.48299319e-01 -6.98445499e-01 6.68668568e-01 -2.97543913e-01
8.83510590e-01 -7.30073869e-01 -4.12521541e-01 -1.24550593e+00
2.76292354e-01 -7.92063177e-01 5.57804219e-02 -5.81344426e-01
-1.58644903e+00 4.21633065e-01 -5.52383661e-01 -1.07191372e+00
-1.46422520e-01 -7.60062814e-01 -1.05555356e+00 4.94238079e-01
-1.19135284e+00 -1.60718632e+00 -3.67737204e-01 3.75891060e-01
1.05216956e+00 -4.84374613e-01 9.66300786e-01 1.18200861e-01
-3.47964853e-01 6.79133296e-01 -5.30235358e-02 4.19235021e-01
4.80283618e-01 -1.20257044e+00 1.12987399e+00 4.06653851e-01
8.69717658e-01 7.97261775e-01 2.09339157e-01 -6.71530545e-01
-1.87819338e+00 -7.06775725e-01 5.64534903e-01 -5.18463552e-01
7.24374712e-01 -8.77393305e-01 -8.22447598e-01 5.48437357e-01
-2.61658490e-01 2.55237341e-01 1.29184723e-01 1.85063034e-01
-9.32924509e-01 4.94047925e-02 -7.54718304e-01 9.12420511e-01
1.49449372e+00 -1.08384526e+00 -5.76655567e-01 3.39316241e-02
1.88406199e-01 -4.25551385e-01 -5.60962498e-01 1.48216292e-01
1.56215334e+00 -6.30069137e-01 1.14566433e+00 -1.12597501e+00
6.92358494e-01 1.00808911e-01 -9.21859294e-02 -7.34471381e-01
-2.68782347e-01 -9.27176356e-01 -5.05690038e-01 8.30679178e-01
2.79679764e-02 -1.35642424e-01 1.16104674e+00 5.90774298e-01
1.66575864e-01 -1.02743018e+00 -6.51360154e-01 -6.68713033e-01
-1.67670816e-01 -4.27030593e-01 5.55683494e-01 7.27110565e-01
-2.19355747e-01 5.34112714e-02 -6.52818859e-01 -2.71913528e-01
5.95635235e-01 6.82061255e-01 1.06747365e+00 -1.13503516e+00
-1.97517723e-01 -8.63461673e-01 -4.79006886e-01 -1.47988987e+00
3.66607994e-01 -7.97341049e-01 -4.64841276e-02 -1.10521865e+00
5.64675853e-02 -7.86298990e-01 -1.96159169e-01 5.30278027e-01
4.98812422e-02 2.98505396e-01 7.01462805e-01 2.77560622e-01
-6.91474259e-01 6.90279841e-01 1.06271052e+00 -3.95047635e-01
-1.78229824e-01 5.37251718e-02 -1.16266645e-01 5.81766605e-01
3.16895247e-01 -3.22769940e-01 -1.87768996e-01 -7.18043864e-01
4.79042269e-02 1.16654843e-01 6.93310142e-01 -8.51774395e-01
6.27833009e-01 -1.82968348e-01 5.55261731e-01 -7.67876685e-01
6.41478598e-01 -7.15138316e-01 -2.98454948e-02 3.11717540e-01
-6.93536282e-01 2.33342081e-01 1.26977324e-01 8.13940048e-01
6.48012683e-02 -1.69315971e-02 5.35497427e-01 -3.59926224e-01
-7.91379690e-01 6.69104517e-01 -2.05063239e-01 -1.66654468e-01
5.48422039e-01 -1.91124007e-01 1.26983836e-01 -3.97922307e-01
-4.07722712e-01 -3.97403874e-02 2.97426134e-01 6.80110991e-01
1.10936296e+00 -1.54329717e+00 -5.35208523e-01 4.05246735e-01
8.86663571e-02 -1.54198289e-01 2.57641643e-01 -4.50185081e-03
-5.28304458e-01 3.20734829e-01 -5.93451202e-01 -2.50180006e-01
-1.05346167e+00 4.94941622e-01 5.37055451e-03 -1.28837973e-01
-1.32614315e+00 1.00147700e+00 -2.10116222e-03 -5.55380344e-01
5.65050602e-01 -5.34563959e-01 3.08816165e-01 -1.13641154e-02
5.61267018e-01 4.33944523e-01 -2.63621062e-01 -1.18467040e-01
6.65203705e-02 5.56634426e-01 -1.89037323e-01 -2.04815581e-01
1.62791145e+00 5.94724238e-01 1.82007328e-01 7.41357207e-01
9.78871167e-01 -9.61674601e-02 -1.77191091e+00 -2.54557073e-01
5.79960123e-02 -6.41125560e-01 -2.48180598e-01 -6.60882473e-01
-1.07859147e+00 1.11050868e+00 9.51666236e-02 -2.61023819e-01
5.91463685e-01 -3.29598844e-01 9.95728016e-01 1.18953288e+00
5.93377948e-01 -9.42129612e-01 4.41620201e-01 5.20025849e-01
1.31750369e+00 -1.02628529e+00 1.18401133e-01 1.68391228e-01
-6.09974682e-01 1.57620156e+00 2.56241977e-01 -8.08175802e-01
4.02590960e-01 3.41271162e-01 -2.46690840e-01 -3.25805396e-01
-8.90548229e-01 3.48341763e-02 3.27529222e-01 2.24768415e-01
8.79017487e-02 1.63488060e-01 2.80896276e-01 5.99611938e-01
-1.64995138e-02 2.34383687e-01 3.04418385e-01 1.02787352e+00
1.07374908e-02 -1.05217052e+00 2.50668973e-01 6.21114492e-01
2.37578359e-02 1.57831591e-02 -6.71511471e-01 6.25334382e-01
-5.40660203e-01 -4.18338142e-02 8.16596486e-03 -2.62960821e-01
5.74676096e-01 1.70644045e-01 8.19255471e-01 -4.40316439e-01
-6.25755608e-01 -3.49800825e-01 -1.22846439e-01 -7.22543955e-01
-3.80227864e-02 -5.20931721e-01 -9.30734873e-01 -3.23993891e-01
4.43628617e-02 -4.47387934e-01 7.43229091e-01 5.69752157e-01
3.72534215e-01 3.15930903e-01 5.29211044e-01 -1.64415479e+00
-7.82385349e-01 -7.18504190e-01 -5.74322581e-01 3.64583343e-01
5.15727997e-01 -6.20080411e-01 9.24688503e-02 -2.11335510e-01] | [11.734039306640625, 0.430096834897995] |
4b225d67-24ae-4fe2-b94b-351eb6834615 | bootstrapping-ternary-relation-extractors | 1511.08952 | null | https://arxiv.org/abs/1511.08952v2 | https://arxiv.org/pdf/1511.08952v2.pdf | Bootstrapping Ternary Relation Extractors | Binary relation extraction methods have been widely studied in recent years. However, few methods have been developed for higher n-ary relation extraction. One limiting factor is the effort required to generate training data. For binary relations, one only has to provide a few dozen pairs of entities per relation, as training data. For ternary relations (n=3), each training instance is a triplet of entities, placing a greater cognitive load on people. For example, many people know that Google acquired Youtube but not the dollar amount or the date of the acquisition and many people know that Hillary Clinton is married to Bill Clinton by not the location or date of their wedding. This makes higher n-nary training data generation a time consuming exercise in searching the Web. We present a resource for training ternary relation extractors. This was generated using a minimally supervised yet effective approach. We present statistics on the size and the quality of the dataset. | ['Ndapandula Nakashole'] | 2015-11-29 | null | null | null | null | ['binary-relation-extraction'] | ['natural-language-processing'] | [ 1.73553348e-01 5.04055142e-01 -5.63960254e-01 -3.20897162e-01
-5.89392304e-01 -6.36936784e-01 5.95812559e-01 8.03091109e-01
-4.40548837e-01 1.36833930e+00 -1.71846300e-01 -5.21228194e-01
-2.94604599e-01 -1.22795045e+00 -4.28409398e-01 -5.92105500e-02
-7.22040609e-02 9.22810137e-01 -6.78572208e-02 -2.39371151e-01
-8.05416107e-02 4.10282165e-01 -1.33463407e+00 4.02762601e-03
7.46957123e-01 9.89275336e-01 -2.16520399e-01 5.42879701e-01
1.61554525e-03 7.85220623e-01 -8.73203337e-01 -1.03341544e+00
2.19206810e-01 -4.72425073e-01 -1.04775429e+00 -9.48112682e-02
3.03517401e-01 -2.53538817e-01 -4.47798342e-01 8.63457382e-01
1.67405337e-01 -3.69813368e-02 8.22987378e-01 -1.44612002e+00
-3.34942520e-01 9.77084517e-01 -2.37516686e-01 3.94464642e-01
7.82782853e-01 -4.99369681e-01 1.22617781e+00 -6.36196971e-01
9.70661640e-01 7.21058190e-01 5.18440962e-01 7.77305141e-02
-9.09357071e-01 -8.33822072e-01 -3.09430718e-01 2.18011096e-01
-1.47876167e+00 -4.97191757e-01 3.09386760e-01 -2.80651242e-01
1.09651065e+00 6.95485830e-01 7.03087151e-01 7.86268890e-01
-4.15181220e-02 3.21835577e-01 9.45935905e-01 -5.76263249e-01
4.99840230e-02 4.26726371e-01 1.29915267e-01 6.07297182e-01
9.66085732e-01 -2.79036462e-01 -6.29805326e-01 -3.05800408e-01
7.93238282e-01 -2.66088128e-01 1.25610754e-01 -1.40241668e-01
-9.95428205e-01 7.03872144e-01 2.13231578e-01 3.81814897e-01
-2.05085769e-01 -3.36989552e-01 1.01374134e-01 3.90286058e-01
1.65813848e-01 8.60580742e-01 -7.42663682e-01 -3.86153847e-01
-9.74683046e-01 4.23738867e-01 1.51986051e+00 1.40570796e+00
8.13879550e-01 -5.15674651e-01 3.73282075e-01 3.38698775e-01
-1.06715068e-01 1.43370479e-01 1.72396779e-01 -6.29082978e-01
9.00903940e-01 1.04631972e+00 2.13291779e-01 -1.15043378e+00
-2.62329519e-01 -1.08723961e-01 -9.52206790e-01 -1.33478492e-01
7.91300774e-01 -5.82176268e-01 -6.24872923e-01 1.05390477e+00
2.89483964e-01 -2.37202376e-01 2.42424205e-01 4.42100972e-01
1.03482735e+00 2.71400422e-01 3.18991169e-02 -5.82746923e-01
1.77610886e+00 -3.41978967e-01 -9.47093487e-01 -3.83664370e-01
6.56163216e-01 -8.03126752e-01 3.88603538e-01 3.18639934e-01
-1.28703535e+00 -5.47664911e-02 -1.16768861e+00 -1.14531383e-01
-8.50680947e-01 1.63396254e-01 1.29334283e+00 8.17545533e-01
-2.37357259e-01 7.26698339e-01 -6.88878059e-01 -6.23941362e-01
4.98838007e-01 7.09033251e-01 -7.86292017e-01 -2.44109496e-03
-1.55423760e+00 1.21348548e+00 6.06494427e-01 -2.01911420e-01
3.73024829e-02 -3.43452066e-01 -1.01235485e+00 1.30014241e-01
8.74531507e-01 -5.32910287e-01 1.06628501e+00 -3.56097966e-01
-8.56965244e-01 8.14730465e-01 -2.28824601e-01 -5.61802924e-01
3.68182063e-01 -1.50654688e-01 -6.76197827e-01 -5.02425432e-02
3.83930624e-01 1.27215415e-01 2.81535536e-01 -7.54170775e-01
-8.54794025e-01 -3.69763792e-01 3.38999182e-01 4.03337240e-01
-3.50609660e-01 3.30894619e-01 -4.16376382e-01 -4.22463387e-01
2.70197988e-01 -7.23097146e-01 -1.13828316e-01 -5.60028851e-01
-7.43084311e-01 -4.27448601e-01 7.26566195e-01 -5.45178056e-01
1.54346633e+00 -1.65771520e+00 -2.55841613e-01 5.74116766e-01
4.09951597e-01 2.34644815e-01 5.27553022e-01 5.63993216e-01
-3.49055171e-01 5.87240398e-01 1.74261749e-01 2.00372592e-01
-1.05785921e-01 1.85740769e-01 1.02815561e-01 -5.35444431e-02
1.58723667e-01 8.05416048e-01 -8.91570866e-01 -7.79227138e-01
-2.86353141e-01 1.42172083e-01 -6.33271039e-02 1.51552316e-02
-1.12831034e-02 -3.45629781e-01 -4.40368980e-01 6.88017547e-01
2.06329852e-01 -4.15301979e-01 6.18482292e-01 -2.52947003e-01
2.03663521e-02 7.94842601e-01 -1.37782013e+00 1.09779453e+00
-6.31792843e-02 6.40060127e-01 -4.12795871e-01 -6.28845155e-01
8.04014862e-01 6.85118914e-01 5.75641751e-01 -3.63000453e-01
3.21690381e-01 2.69176900e-01 4.21253555e-02 -4.46909696e-01
6.45147622e-01 5.34040630e-02 -1.31414220e-01 4.76788074e-01
-6.51973635e-02 -5.01159251e-01 1.04600418e+00 5.01215100e-01
1.35934293e+00 -1.07183330e-01 9.22912300e-01 2.58640796e-01
8.62492174e-02 3.55054468e-01 6.22366250e-01 3.75451386e-01
3.12208533e-01 1.74611211e-01 9.43056703e-01 -4.75778610e-01
-1.11317027e+00 -6.40603364e-01 -2.26792302e-02 6.21615887e-01
-1.10186227e-01 -1.10589325e+00 -4.90221083e-01 -8.54943871e-01
-9.97724161e-02 5.00152051e-01 -4.02846038e-01 6.21257350e-02
-5.01870573e-01 -6.37119651e-01 4.77318019e-01 2.38931760e-01
5.37622809e-01 -8.22936535e-01 -4.87191737e-01 1.76720217e-01
-5.11089802e-01 -1.38422561e+00 -1.28001630e-01 5.25611103e-01
-7.86700487e-01 -1.20141602e+00 -1.60159186e-01 -7.65825212e-01
7.24526048e-01 -2.62360811e-01 1.41125679e+00 2.09500805e-01
-2.17153087e-01 -3.09677660e-01 -3.53232652e-01 -5.50342679e-01
-9.05677751e-02 5.85955024e-01 4.70237024e-02 -5.48347652e-01
8.36824238e-01 -6.39047563e-01 -2.71750037e-02 2.99889505e-01
-5.22577763e-01 -7.58286193e-02 7.17504621e-01 5.46271384e-01
2.99962163e-01 7.75848448e-01 2.49266565e-01 -1.46771181e+00
7.74526477e-01 -6.06833816e-01 -3.55315059e-01 2.39674076e-01
-8.73645544e-01 -2.79861242e-02 2.91354597e-01 -4.24744904e-01
-6.49509311e-01 3.03901553e-01 2.54670411e-01 4.11773056e-01
-3.03560495e-01 8.09019864e-01 -2.73325533e-01 2.80926675e-01
7.38901019e-01 -2.84726977e-01 -5.11868834e-01 -1.96892582e-02
9.33976620e-02 6.97514832e-01 5.09782314e-01 -4.72658873e-01
9.56582844e-01 4.35903892e-02 1.68340579e-01 -7.58100152e-01
-7.86667585e-01 -3.13378215e-01 -9.91049469e-01 1.32727429e-01
6.90231740e-01 -7.43139088e-01 -9.09418225e-01 8.55497345e-02
-1.02525103e+00 -6.92671165e-02 -2.23298550e-01 4.32189196e-01
-1.53789356e-01 6.34587333e-02 -5.48332334e-01 -7.42821395e-01
-1.27890259e-01 -5.25960326e-01 3.61199975e-01 3.38583410e-01
-8.76357019e-01 -7.83203781e-01 -3.34055126e-01 5.38155138e-01
-1.10236116e-01 3.57280463e-01 1.03702474e+00 -8.59166265e-01
-6.09956324e-01 -7.59662688e-01 -4.55550402e-01 -1.68940723e-01
5.35882950e-01 -9.48465429e-03 -3.29902828e-01 1.73849866e-01
-3.63671809e-01 -2.30853960e-01 4.02225032e-02 -1.20661996e-01
5.10838389e-01 -5.70870697e-01 -6.60156369e-01 7.82386139e-02
1.07290709e+00 4.34993207e-01 6.48623109e-01 3.71890217e-01
7.73882270e-01 4.75084543e-01 8.98599565e-01 3.26797038e-01
6.54286861e-01 7.89595127e-01 -5.11418283e-02 -8.05012435e-02
1.10858776e-01 -2.63999492e-01 -2.21962050e-01 2.48796195e-01
-7.20210195e-01 -1.16346724e-01 -1.01637828e+00 5.23862600e-01
-1.52616477e+00 -1.08776891e+00 -3.59364450e-01 2.11379004e+00
1.38577414e+00 5.37334263e-01 1.37424126e-01 5.84431946e-01
3.78707319e-01 -3.33311439e-01 -2.96922982e-01 -5.08467387e-03
-1.08636826e-01 4.75720078e-01 5.52858770e-01 4.69737321e-01
-1.03683412e+00 9.66306210e-01 6.15829277e+00 5.97508132e-01
-4.45865601e-01 -3.00418258e-01 5.99775255e-01 -6.23720400e-02
-8.42356905e-02 3.41057688e-01 -1.16688597e+00 2.12510854e-01
8.80134881e-01 -4.04876322e-01 1.16822399e-01 7.95155287e-01
-1.42939150e-01 -6.23457670e-01 -1.45075953e+00 1.10748816e+00
-1.01056509e-01 -1.14696002e+00 -2.07804367e-01 4.40007120e-01
6.60428882e-01 -6.98610604e-01 -3.47208679e-01 2.21968457e-01
5.71467280e-01 -1.17915487e+00 3.44312757e-01 3.36390823e-01
8.49640846e-01 -8.71894896e-01 9.57307935e-01 3.03856105e-01
-1.28242266e+00 1.88756913e-01 -8.26888531e-02 -5.35512686e-01
2.41177324e-02 9.11152244e-01 -1.10252798e+00 6.78758144e-01
7.00625241e-01 3.86990786e-01 -7.76801050e-01 6.77049339e-01
-4.00223374e-01 2.62204468e-01 -6.89083457e-01 -3.06648940e-01
-2.08212852e-01 -1.80174857e-01 1.44201010e-01 8.75328720e-01
2.55967379e-01 6.17858887e-01 -8.93083662e-02 3.29292297e-01
-2.99651116e-01 8.56478661e-02 -7.13308454e-01 -4.17083412e-01
7.53798783e-01 1.35324466e+00 -9.03705478e-01 -5.33286691e-01
-3.68742347e-01 5.62500119e-01 4.96442884e-01 1.61313727e-01
-4.78572160e-01 -7.01652169e-01 3.63143921e-01 4.63076890e-01
7.22551793e-02 -3.15526992e-01 -4.90795434e-01 -9.46897805e-01
8.44949186e-02 -1.07449341e+00 5.81265748e-01 -6.47612154e-01
-8.74565482e-01 7.60185301e-01 2.05747768e-01 -8.97704542e-01
-6.77861333e-01 -3.83717626e-01 -9.84866843e-02 8.05762470e-01
-7.69008040e-01 -8.38397384e-01 -2.27370143e-01 4.56223220e-01
7.78200710e-03 -3.04075509e-01 1.11167026e+00 5.42283893e-01
-6.16921842e-01 5.47054529e-01 -6.80784762e-01 5.63387513e-01
5.71883857e-01 -1.28800416e+00 4.49033737e-01 7.30255783e-01
4.23397958e-01 8.95245492e-01 8.68299127e-01 -9.35916901e-01
-9.82014537e-01 -5.15715659e-01 1.91467094e+00 -7.63144672e-01
7.47394800e-01 -3.57759774e-01 -5.39098620e-01 8.50878000e-01
8.23937431e-02 -2.80947477e-01 8.96166384e-01 5.90496421e-01
-8.98954794e-02 -5.47006838e-02 -9.53737497e-01 6.97618723e-01
8.80883813e-01 -5.02896249e-01 -5.87522626e-01 5.18083096e-01
2.24827543e-01 -6.11774564e-01 -1.14307761e+00 3.68449748e-01
4.55191314e-01 -7.93699324e-01 7.94142485e-01 -7.02460885e-01
5.02417505e-01 -1.51346982e-01 1.58894092e-01 -8.77512872e-01
2.91714352e-02 -7.96260476e-01 -5.08170962e-01 1.43341792e+00
1.05713212e+00 -4.56138521e-01 1.22036219e+00 1.33298802e+00
5.74117005e-01 -9.49419081e-01 -5.86476445e-01 -5.85379422e-01
-3.60522896e-01 -1.69857576e-01 8.40991557e-01 1.31893241e+00
5.10348201e-01 8.24170709e-01 -2.22900316e-01 -3.84295620e-02
3.26605171e-01 6.65558502e-02 8.34699810e-01 -1.52531505e+00
-2.99622342e-02 -1.04531221e-01 -4.11753058e-01 -7.95508921e-01
-2.97123432e-01 -6.17334485e-01 -3.65075231e-01 -1.69394410e+00
5.01106754e-02 -6.83985114e-01 2.33819336e-01 8.14878643e-01
-1.31172583e-01 2.34730065e-01 -8.78460854e-02 1.32463634e-01
-2.97680676e-01 -2.73172200e-01 9.94851232e-01 1.34214535e-01
-2.34286636e-01 4.22716051e-01 -9.22011554e-01 7.12637305e-01
8.29338849e-01 -6.59883976e-01 -5.02588689e-01 6.10977039e-02
9.58606958e-01 2.47639731e-01 6.17508451e-03 -8.67288709e-01
5.39539337e-01 -2.21152753e-01 6.07554674e-01 -5.85980058e-01
5.00019372e-01 -1.11910152e+00 4.17575061e-01 7.78752193e-02
-5.86801730e-02 1.38649464e-01 -1.91984162e-01 9.20422971e-02
-3.66494566e-01 -2.95724541e-01 1.78956851e-01 -2.91329056e-01
-2.07327485e-01 3.56665015e-01 -2.29870096e-01 2.21705865e-02
1.07147825e+00 -2.49057591e-01 -2.49288246e-01 -4.42060590e-01
-8.65782976e-01 1.43640349e-02 1.81748942e-01 2.47206137e-01
4.10123110e-01 -1.24352360e+00 -4.50258434e-01 3.77887525e-02
-2.04912379e-01 8.80910456e-02 -4.04689521e-01 5.64862251e-01
-3.40223670e-01 4.29804832e-01 -1.99157476e-01 2.96588302e-01
-1.61367881e+00 3.59188884e-01 -5.97233437e-02 -6.51247442e-01
-3.06321740e-01 8.15717638e-01 -8.01563501e-01 -2.06213426e-02
1.36786759e-01 -3.50473434e-01 -5.69737554e-01 5.58344841e-01
3.21506321e-01 3.74547213e-01 2.62871712e-01 -3.75921369e-01
-2.71267831e-01 8.54839087e-02 -2.44391009e-01 -7.59909675e-02
1.33757997e+00 1.31259620e-01 -3.03331792e-01 3.43451351e-01
6.87427402e-01 1.54066905e-01 -4.88542229e-01 -6.97405934e-02
2.77065247e-01 -3.87646377e-01 -4.54189241e-01 -6.76984072e-01
-8.22406650e-01 1.26004502e-01 -1.85448885e-01 9.00529921e-01
8.80722582e-01 2.04668254e-01 6.82512820e-01 6.53070807e-01
6.01134598e-01 -9.78210449e-01 -3.66994023e-01 5.81777334e-01
4.51174647e-01 -1.34804344e+00 7.86905766e-01 -1.13359153e+00
-5.86992383e-01 9.53973651e-01 4.84680325e-01 1.65927157e-01
8.21738780e-01 5.16346574e-01 -8.44012424e-02 -4.55589920e-01
-6.17864490e-01 -2.53225803e-01 3.35767239e-01 5.30942202e-01
7.28254974e-01 2.13314682e-01 -6.22801244e-01 5.77663779e-01
-9.61494505e-01 -1.49534509e-01 5.66504538e-01 9.91068959e-01
1.25299441e-02 -1.64334607e+00 -1.60495445e-01 9.91880715e-01
-6.62932456e-01 -4.41746026e-01 -7.90224671e-01 9.47225451e-01
3.23636055e-01 1.18268645e+00 -1.85041994e-01 -3.84170145e-01
5.44556558e-01 1.71595961e-01 4.88630772e-01 -9.43439543e-01
-4.49575633e-01 -3.47262055e-01 8.94941568e-01 -1.46624476e-01
-4.56562102e-01 -8.93295944e-01 -1.27083564e+00 -6.53721333e-01
-5.50786555e-01 5.50107002e-01 4.14877921e-01 1.05864477e+00
1.22058932e-02 2.52169311e-01 1.48631975e-01 -4.09695417e-01
2.22841293e-01 -1.04625797e+00 -7.74399281e-01 3.76765549e-01
-1.28959030e-01 -6.86990976e-01 -1.85442597e-01 3.26117337e-01] | [9.301706314086914, 8.653947830200195] |
d3c8f6b0-66eb-4e3e-adb4-9bde1d70d80d | natural-response-generation-for-chinese | 2302.08817 | null | https://arxiv.org/abs/2302.08817v1 | https://arxiv.org/pdf/2302.08817v1.pdf | Natural Response Generation for Chinese Reading Comprehension | Machine reading comprehension (MRC) is an important area of conversation agents and draws a lot of attention. However, there is a notable limitation to current MRC benchmarks: The labeled answers are mostly either spans extracted from the target corpus or the choices of the given candidates, ignoring the natural aspect of high-quality responses. As a result, MRC models trained on these datasets can not generate human-like responses in real QA scenarios. To this end, we construct a new dataset called Penguin to promote the research of MRC, providing a training and test bed for natural response generation to real scenarios. Concretely, Penguin consists of 200k training data with high-quality fluent, and well-informed responses. Penguin is the first benchmark towards natural response generation in Chinese MRC on a relatively large scale. To address the challenges in Penguin, we develop two strong baselines: end-to-end and two-stage frameworks. Following that, we further design Prompt-BART: fine-tuning the pre-trained generative language models with a mixture of prefix prompts in Penguin. Extensive experiments validated the effectiveness of this design. | ['Jia Li', 'Baoyuan Wang', 'Yinan Bao', 'Hongguang Li', 'Nuo Chen'] | 2023-02-17 | null | null | null | null | ['response-generation', 'reading-comprehension', 'machine-reading-comprehension'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 2.5977042e-01 3.1626269e-01 2.5378585e-01 -6.7994153e-01
-1.4370737e+00 -6.2675822e-01 7.2492415e-01 -4.1475713e-01
-2.4167152e-01 1.0712588e+00 9.2025304e-01 -2.3048125e-01
3.5834795e-01 -7.7718896e-01 -5.7797575e-01 -3.2326326e-01
4.8501903e-01 8.3791494e-01 1.4402555e-01 -7.4025983e-01
1.7603071e-01 -2.6086038e-01 -1.1455547e+00 8.5402209e-01
1.2642531e+00 6.3078052e-01 3.5335359e-01 1.0664269e+00
2.0230892e-01 1.2827549e+00 -1.0000318e+00 -7.5310886e-01
-6.7931913e-02 -9.5310187e-01 -1.3972458e+00 -2.0041621e-01
1.9890565e-01 -6.5628493e-01 -3.6702067e-01 4.8449928e-01
9.7960573e-01 4.8695239e-01 4.9307731e-01 -9.9781001e-01
-1.1539298e+00 8.9729804e-01 1.0959507e-01 1.0637337e-01
9.0093607e-01 6.7614156e-01 1.1187941e+00 -8.8679743e-01
4.9665734e-01 1.5632465e+00 2.4417168e-01 1.0612712e+00
-8.8512021e-01 -3.0462909e-01 -2.3883291e-02 2.0738859e-01
-8.0987942e-01 -7.4272132e-01 5.1975882e-01 1.6429359e-02
1.0200883e+00 4.6600571e-01 8.3739974e-02 1.8144267e+00
-2.4725141e-01 1.1621070e+00 1.3412899e+00 -5.0834817e-01
-1.3516341e-02 -1.3355059e-02 2.2732830e-01 2.8365627e-01
-4.9400160e-01 -6.9014765e-02 -7.4795711e-01 -1.4086998e-01
3.0824035e-01 -4.8815653e-01 -8.0969745e-01 3.5982212e-01
-1.4408511e+00 9.8278654e-01 3.2871085e-01 -1.4672121e-01
-2.9705548e-01 -1.8266118e-01 1.7879409e-01 4.9714279e-01
3.2289448e-01 8.0433834e-01 -4.5100221e-01 -5.1692122e-01
-4.2636362e-01 7.6151544e-01 1.2141099e+00 1.2252223e+00
4.4240102e-01 -2.0838301e-01 -8.5338110e-01 1.0945690e+00
9.5050633e-02 6.4441603e-01 5.3504241e-01 -9.4261599e-01
1.0310553e+00 5.1777822e-01 2.9670462e-01 -5.9717840e-01
-1.6170463e-01 -2.1148224e-01 -6.4896488e-01 -5.3158039e-01
7.1974033e-01 -4.5242700e-01 -4.2968693e-01 1.8573043e+00
2.3974313e-01 -3.0836979e-01 4.4786027e-01 1.1039529e+00
1.2336783e+00 9.3384856e-01 -1.5115939e-01 -1.4534067e-01
1.3534265e+00 -1.6106167e+00 -7.5577909e-01 -4.3214869e-01
3.8945913e-01 -7.6825982e-01 1.8476925e+00 2.8771067e-01
-1.3256792e+00 -5.0247568e-01 -5.7792449e-01 -3.9229885e-01
1.9258951e-01 9.7932316e-02 3.2169202e-01 2.2119325e-01
-9.4586772e-01 2.2698222e-03 -3.5671905e-01 -1.4921993e-01
-2.7942549e-02 -1.7993723e-01 6.6320390e-02 -4.7541466e-01
-1.6657205e+00 9.8351276e-01 5.5533402e-02 1.4176178e-01
-1.0153486e+00 -6.2484270e-01 -7.4838430e-01 5.6500003e-02
6.1755592e-01 -8.4571236e-01 2.1994720e+00 -8.5266042e-01
-2.0040381e+00 6.5096432e-01 -2.2426172e-01 -3.6766428e-01
7.8319341e-01 -5.1823771e-01 -3.2424033e-01 3.2367185e-02
1.7228812e-01 7.4171406e-01 4.7037685e-01 -1.0650399e+00
-5.3131819e-01 1.4410116e-01 5.0498176e-01 3.3597013e-01
1.3934420e-01 1.5270534e-01 -1.9938590e-01 -4.7855905e-01
-5.0148076e-01 -8.5989475e-01 -1.6027530e-01 -9.4222677e-01
-4.0755248e-01 -8.2053429e-01 2.6009133e-02 -7.1393096e-01
1.2322959e+00 -1.6772317e+00 8.2584620e-03 -4.0561348e-01
1.1185894e-01 1.7174587e-01 -5.6161934e-01 9.5867109e-01
4.1460612e-01 -3.1080814e-02 -2.4293143e-01 -1.8827420e-01
2.8839001e-01 5.1809318e-02 -7.8490871e-01 -2.9384771e-01
5.6613857e-01 1.3438864e+00 -1.2709765e+00 -3.1203729e-01
-3.4717357e-01 -3.5731543e-02 -6.2108827e-01 9.9647403e-01
-7.1608561e-01 6.3840771e-01 -5.4744178e-01 4.8189145e-01
4.1382498e-01 -4.3087763e-01 -4.9116369e-02 2.4181944e-01
1.6107459e-01 1.1314759e+00 -5.3009796e-01 1.7197615e+00
-5.4920161e-01 3.7727243e-01 -2.9045507e-01 -2.8982610e-01
8.9316905e-01 3.2672247e-01 -2.2115505e-01 -1.0156755e+00
6.9168000e-04 2.4443540e-01 2.3085210e-01 -7.4671012e-01
9.2992878e-01 6.0132254e-02 -4.2761472e-01 8.9327645e-01
-1.4263283e-02 -3.4247255e-01 4.1332084e-01 6.1226040e-01
1.2088475e+00 -3.3190299e-02 1.9430007e-01 1.4867643e-02
5.3954917e-01 1.6979158e-01 4.3682271e-01 1.0098323e+00
-5.3479895e-02 9.6736270e-01 6.1160862e-01 -9.6781760e-02
-5.8179432e-01 -8.7340754e-01 3.3994490e-01 1.4229410e+00
-6.9038138e-02 -4.9105105e-01 -1.0569311e+00 -9.1051322e-01
-4.9994582e-01 9.5482308e-01 -4.7461930e-01 -8.6028084e-02
-8.9700800e-01 -7.3067260e-01 7.6675123e-01 5.0721848e-01
6.9435894e-01 -1.6631807e+00 -6.9263273e-01 3.6421832e-01
-1.1199570e+00 -1.2037748e+00 -6.6463017e-01 -3.6880046e-01
-1.7090228e-01 -1.0689601e+00 -7.6049113e-01 -7.0484358e-01
3.3391976e-01 3.6861309e-01 1.9113247e+00 1.9516279e-01
5.2634364e-01 2.3683013e-01 -9.6429777e-01 -3.5433561e-01
-8.8146627e-01 4.0400857e-01 -3.6544544e-01 -2.0944467e-01
2.7508429e-01 -1.8753125e-01 -6.7585582e-01 5.5808735e-01
-6.1298394e-01 4.7889644e-01 6.7031246e-01 1.1024318e+00
7.4775353e-02 -7.0860511e-01 1.1601629e+00 -1.0473959e+00
1.3980427e+00 -6.7221224e-01 -1.4231461e-01 6.1943638e-01
-1.2981209e-01 -6.8966128e-02 9.4915098e-01 -5.0904351e-01
-1.3681980e+00 -5.0791711e-01 -3.2659450e-01 3.9135107e-01
-1.5084407e-01 5.4917467e-01 -4.3613777e-01 6.5269041e-01
8.8249218e-01 3.5726777e-01 -2.1138473e-01 -4.1119504e-01
5.9430534e-01 1.0660058e+00 8.0254591e-01 -1.0099947e+00
3.2072803e-01 -3.0568904e-01 -8.4446192e-01 -3.6330119e-01
-1.1839271e+00 -1.8415551e-01 -1.5392366e-01 -1.0856797e-01
5.8163613e-01 -9.9514270e-01 -7.0742953e-01 6.3207579e-01
-1.4718341e+00 -9.6327657e-01 -1.0343535e-01 2.0011693e-01
-6.5174395e-01 1.0919243e-01 -8.5233372e-01 -6.5747559e-01
-7.1399558e-01 -1.1466316e+00 8.8868678e-01 3.5478839e-01
-5.3196049e-01 -6.3769269e-01 2.9830554e-01 9.4823349e-01
6.6220546e-01 -1.3580202e-01 9.0997392e-01 -7.7878451e-01
-6.0389531e-01 1.8473761e-01 -1.7289554e-01 3.4109631e-01
-7.4042894e-02 -1.8640487e-01 -1.0326066e+00 -1.3356325e-01
1.4952262e-01 -1.2230698e+00 6.0348403e-01 -2.6462430e-01
8.5806054e-01 -6.1132288e-01 2.2731304e-01 8.3962873e-02
6.9380093e-01 4.4761058e-03 6.2345386e-01 3.9979868e-02
4.8192891e-01 9.6762717e-01 8.9398414e-01 2.7866566e-01
1.0693032e+00 6.5223759e-01 1.6591430e-01 1.7063968e-01
-9.1623493e-02 -7.1076143e-01 5.6166065e-01 1.3668920e+00
7.8670561e-02 -8.3453476e-01 -9.6020567e-01 6.8584776e-01
-1.8862120e+00 -8.3836472e-01 -4.2447227e-01 1.8820812e+00
1.4144484e+00 -9.9733651e-02 1.7391728e-01 -3.6388385e-01
3.5765043e-01 1.6847186e-01 -3.7342179e-01 -3.8970220e-01
-2.2648624e-01 3.0728412e-01 -3.1096563e-01 5.2764261e-01
-6.0868013e-01 1.1109269e+00 5.7718515e+00 7.6473957e-01
-8.3009833e-01 1.3668756e-01 6.3865650e-01 -6.5240688e-03
-5.5453008e-01 -3.4221392e-02 -7.9311550e-01 5.5897611e-01
1.1616575e+00 -2.1508598e-01 5.3130686e-01 6.1515027e-01
2.5280133e-01 6.7008816e-02 -1.1771804e+00 6.0309541e-01
2.2136740e-01 -1.0441449e+00 3.7433002e-02 -2.6784372e-01
7.8688014e-01 1.5817722e-02 -1.3247195e-01 8.9386243e-01
9.0771002e-01 -1.2454034e+00 6.9535279e-01 4.2161283e-01
6.7612994e-01 -5.5000347e-01 7.6262987e-01 8.1496650e-01
-5.0837165e-01 1.8371915e-02 -4.6207997e-01 -3.6605680e-01
5.2569962e-01 1.4084648e-01 -1.0821851e+00 5.4059434e-01
4.0226784e-01 2.8719169e-01 -5.6596452e-01 5.5759078e-01
-9.2340744e-01 1.0532323e+00 4.1870210e-02 -3.8503617e-01
3.9504302e-01 2.9281314e-02 2.1419452e-01 1.2257032e+00
4.7821399e-02 5.7427180e-01 1.2517443e-01 9.1581130e-01
-4.2171660e-01 6.9776036e-02 -1.8371351e-01 -5.9465799e-03
7.8565991e-01 1.2810887e+00 2.8212912e-02 -3.1756574e-01
-4.4409135e-01 8.8648450e-01 6.9158590e-01 3.9651817e-01
-7.6908368e-01 -1.5609935e-01 2.9157373e-01 -4.1554444e-02
-1.8950135e-01 1.1418885e-01 4.5490935e-02 -1.2411522e+00
1.6501348e-01 -1.9222463e+00 3.4429905e-01 -9.3584269e-01
-1.5643824e+00 8.9354539e-01 -3.4450538e-02 -9.5754606e-01
-8.1369823e-01 -2.5638124e-01 -7.4986118e-01 1.0666585e+00
-1.4628348e+00 -1.2380459e+00 -5.9193403e-01 4.8388940e-01
1.1223038e+00 1.5701858e-02 8.3288896e-01 2.1432832e-02
-4.5613727e-01 7.6119995e-01 -3.0891097e-01 1.8133099e-01
1.0938350e+00 -1.4397056e+00 9.8605710e-01 9.5068896e-01
2.0038363e-02 7.7492189e-01 5.7577109e-01 -3.7394005e-01
-1.2123246e+00 -9.1833919e-01 1.1943630e+00 -9.2481595e-01
4.8694041e-01 -5.7091027e-01 -1.1852452e+00 5.7484996e-01
7.5588429e-01 -5.3613400e-01 6.0809314e-01 -3.7183013e-02
-2.5876332e-02 2.6226240e-01 -7.1757632e-01 8.4853542e-01
1.1162403e+00 -4.7503740e-01 -8.8741648e-01 5.8300406e-01
1.0830307e+00 -8.6281258e-01 -6.1567086e-01 4.1407010e-01
3.1767318e-01 -9.2545545e-01 6.6055471e-01 -8.6485791e-01
9.4129771e-01 2.6444126e-02 4.6048228e-02 -1.7224723e+00
-1.1610869e-01 -1.1295187e+00 -1.5581846e-02 1.6229557e+00
5.4488063e-01 -4.2657289e-01 3.9639845e-01 8.4864348e-01
-4.6663645e-01 -9.0152353e-01 -5.1234961e-01 -2.7330813e-01
3.2795927e-01 -1.4213713e-01 1.0218248e+00 6.6246337e-01
6.6704690e-02 9.9615824e-01 -5.9109533e-01 -2.4353796e-01
-3.0385377e-02 2.6903197e-01 1.4400284e+00 -5.9091038e-01
-5.7906443e-01 -2.5883794e-01 6.1608237e-01 -1.8904902e+00
1.6323380e-01 -6.0084194e-01 5.0758487e-01 -1.4556700e+00
1.3677153e-01 -4.6082819e-01 2.4513094e-01 3.1739807e-01
-8.7466764e-01 -2.3199175e-01 1.4084268e-01 2.4980032e-01
-7.2911930e-01 8.4294206e-01 1.6783112e+00 1.7875645e-01
-1.3088159e-01 1.7973636e-01 -1.1325727e+00 2.9488143e-01
8.4277904e-01 -3.0289301e-01 -7.7379167e-01 -8.8820487e-01
3.3901960e-01 5.1133817e-01 2.8921998e-01 -3.4684402e-01
1.6033150e-01 -4.0700153e-01 -2.7948523e-01 -5.5730897e-01
1.1959691e-01 2.3974497e-02 -3.1046578e-01 -1.0515533e-01
-8.9198667e-01 3.4478131e-01 -3.2537681e-01 2.8150859e-01
-3.1109244e-01 -1.4518638e-01 3.4024492e-01 -2.6374060e-01
-5.0487351e-01 -9.2543505e-02 -3.5839215e-01 8.7546736e-01
4.7101453e-01 1.5834273e-01 -1.0386109e+00 -8.9323848e-01
1.0350704e-01 7.8018111e-01 2.7600595e-01 6.9754940e-01
5.9617293e-01 -1.0740889e+00 -1.3536309e+00 -6.7022100e-02
3.4671500e-01 5.9540653e-01 3.1612015e-01 7.1047777e-01
-3.7006244e-01 4.4702744e-01 -3.9528579e-02 -2.9772300e-01
-9.6945953e-01 2.6023969e-01 2.5386313e-01 -6.1414605e-01
-4.7605753e-01 1.0606235e+00 3.1302291e-01 -8.9088178e-01
-3.7067037e-03 -3.9600518e-01 -3.3962110e-01 -2.4270055e-01
8.0019456e-01 2.8970858e-01 4.5824453e-02 -3.8828051e-01
1.2817633e-02 -3.0239442e-01 -2.8367287e-01 -3.7511176e-01
9.4635713e-01 -1.9208112e-01 -7.5573154e-02 2.4679652e-01
6.3844419e-01 3.1763491e-01 -1.2943454e+00 -4.3547922e-01
-2.7173787e-02 -4.5422053e-01 -7.4930322e-01 -1.3655533e+00
-5.4852408e-01 8.8887918e-01 -3.8448247e-01 1.6166863e-01
9.4981760e-01 -3.0173607e-02 1.2581547e+00 5.6216282e-01
3.5027480e-01 -9.5677370e-01 4.9038061e-01 8.9746898e-01
1.4345647e+00 -1.2500668e+00 -6.3146007e-01 -3.4416881e-01
-1.1686064e+00 8.1524634e-01 1.0941713e+00 6.4353697e-02
-3.5481772e-01 -4.2113680e-02 6.4178455e-01 4.7838304e-02
-1.4987379e+00 -7.1367875e-02 2.5031182e-01 5.8744007e-01
7.0368338e-01 1.4637692e-01 -3.2164297e-01 1.0755872e+00
-9.0992779e-01 -1.2946519e-01 7.0255464e-01 7.2479397e-01
-1.8941434e-01 -1.1645423e+00 -8.2643129e-02 1.5965545e-01
-3.2320350e-01 -3.1165078e-01 -8.9716578e-01 6.2533998e-01
-4.1902512e-01 1.5713570e+00 -4.3517664e-01 -3.4885943e-01
5.9573781e-01 -4.9552601e-02 3.4753644e-01 -7.3846096e-01
-9.6479750e-01 -3.5558015e-01 5.7011777e-01 -3.7238020e-01
-8.2874157e-02 -4.6630129e-01 -9.9991298e-01 -3.1448644e-01
-3.2111385e-01 4.5103160e-01 -5.5496633e-02 1.0590184e+00
3.3208719e-01 4.8511741e-01 6.7238086e-01 -3.1385329e-01
-1.2021090e+00 -1.5944500e+00 2.1909563e-01 6.4662391e-01
1.1991649e-01 -1.6623065e-01 -1.6902183e-01 -6.5505035e-02] | [12.299242973327637, 8.310821533203125] |
0c22b5ee-7690-4253-8dde-4b24fcc22d35 | billion-scale-similarity-search-with-gpus | 1702.08734 | null | http://arxiv.org/abs/1702.08734v1 | http://arxiv.org/pdf/1702.08734v1.pdf | Billion-scale similarity search with GPUs | Similarity search finds application in specialized database systems handling
complex data such as images or videos, which are typically represented by
high-dimensional features and require specific indexing structures. This paper
tackles the problem of better utilizing GPUs for this task. While GPUs excel at
data-parallel tasks, prior approaches are bottlenecked by algorithms that
expose less parallelism, such as k-min selection, or make poor use of the
memory hierarchy.
We propose a design for k-selection that operates at up to 55% of theoretical
peak performance, enabling a nearest neighbor implementation that is 8.5x
faster than prior GPU state of the art. We apply it in different similarity
search scenarios, by proposing optimized design for brute-force, approximate
and compressed-domain search based on product quantization. In all these
setups, we outperform the state of the art by large margins. Our implementation
enables the construction of a high accuracy k-NN graph on 95 million images
from the Yfcc100M dataset in 35 minutes, and of a graph connecting 1 billion
vectors in less than 12 hours on 4 Maxwell Titan X GPUs. We have open-sourced
our approach for the sake of comparison and reproducibility. | ['Hervé Jégou', 'Matthijs Douze', 'Jeff Johnson'] | 2017-02-28 | null | null | null | null | ['image-similarity-search'] | ['computer-vision'] | [ 2.35276371e-02 -4.86995667e-01 -2.63635188e-01 -1.94574088e-01
-8.74336064e-01 -6.05030119e-01 6.09693348e-01 5.16097784e-01
-7.35364079e-01 2.05892444e-01 1.32577941e-01 -4.05616820e-01
-4.16140825e-01 -9.40274894e-01 -6.12570405e-01 -4.73821700e-01
-6.89063817e-02 6.55839980e-01 5.50076485e-01 -2.79073119e-01
8.04826736e-01 6.09689534e-01 -2.11832166e+00 2.35601813e-01
4.67659891e-01 1.34337902e+00 3.29102576e-01 7.71176100e-01
-1.32063672e-01 2.25178063e-01 -1.92016929e-01 -3.58138889e-01
5.25246620e-01 5.57700209e-02 -8.62078071e-01 -3.13218564e-01
9.00388360e-01 -3.31114203e-01 -4.35011804e-01 7.92101383e-01
7.56152332e-01 1.34805202e-01 5.61984003e-01 -1.10340166e+00
-2.56234765e-01 9.92096737e-02 -5.52362442e-01 3.66349101e-01
5.89845598e-01 -2.41395030e-02 1.11492336e+00 -9.64736223e-01
7.46895254e-01 9.42181349e-01 8.02665651e-01 -1.26810884e-03
-1.12867367e+00 -4.14335787e-01 -5.17410457e-01 6.44925714e-01
-1.69792342e+00 -4.45294470e-01 8.74921605e-02 -1.41694576e-01
1.49375188e+00 5.81316411e-01 7.20841110e-01 7.17764676e-01
2.03385875e-01 4.49869961e-01 8.61515284e-01 -4.19331402e-01
4.12266701e-01 -6.33729622e-02 4.92975831e-01 7.18755186e-01
4.88918751e-01 7.41705447e-02 -8.48432302e-01 -8.38153481e-01
5.25422096e-01 8.53909403e-02 -1.62571654e-01 -4.20285553e-01
-1.33299160e+00 1.07697272e+00 2.74761468e-01 1.86935410e-01
-3.03417474e-01 2.51372278e-01 7.48506844e-01 4.35366243e-01
1.79914385e-01 4.28181529e-01 -3.46054763e-01 -4.09784794e-01
-1.27604198e+00 3.66666943e-01 1.10141337e+00 1.21993804e+00
8.79823148e-01 -6.48464322e-01 -1.01662099e-01 6.68029010e-01
-1.58931371e-02 5.21612465e-01 6.97380722e-01 -1.06190491e+00
3.42916757e-01 3.91462803e-01 -2.11427972e-01 -1.25624132e+00
-5.54640770e-01 -3.98027033e-01 -9.62771833e-01 -4.47501354e-02
2.56991684e-01 4.39399421e-01 -5.03199995e-01 9.79499280e-01
7.03787625e-01 1.28289223e-01 -2.29542688e-01 1.01291704e+00
6.47316396e-01 6.57971323e-01 -4.63968933e-01 -3.45306396e-02
1.84175897e+00 -1.17529964e+00 -2.16123715e-01 2.42792442e-01
1.10221946e+00 -1.06908143e+00 1.05386090e+00 4.88824904e-01
-1.00201571e+00 -4.81862754e-01 -1.12397218e+00 -4.57592905e-01
-3.58125925e-01 -1.35681063e-01 8.43811512e-01 7.00601041e-01
-1.26485884e+00 1.00144768e+00 -8.63981783e-01 -7.29942918e-01
2.63532877e-01 4.37246501e-01 -4.64610636e-01 -2.11346194e-01
-7.75968432e-01 6.58754349e-01 2.87879527e-01 -4.64217633e-01
-5.85604794e-02 -1.02066517e+00 -5.15509367e-01 2.07387581e-01
3.53735059e-01 -9.26332057e-01 7.95216680e-01 -1.75512180e-01
-1.21086872e+00 9.86592948e-01 -1.00370154e-01 -7.20716536e-01
2.89403051e-01 -2.86156654e-01 -9.25717130e-02 2.57183313e-01
-1.54700223e-02 4.78907287e-01 7.20946133e-01 -2.66503364e-01
-6.89483643e-01 -5.47185361e-01 -2.81327695e-01 2.35505342e-01
-5.87807238e-01 2.62002517e-02 -9.30163741e-01 -5.85120797e-01
2.16666907e-01 -1.22556317e+00 -4.17646348e-01 3.12869936e-01
-1.24220438e-01 -2.94502854e-01 5.99376500e-01 -1.81696683e-01
1.32285011e+00 -2.14674616e+00 -1.94908790e-02 5.69020271e-01
3.13356668e-01 3.55719030e-01 7.81560689e-02 7.72815764e-01
3.22556794e-01 -2.90410012e-01 1.09937757e-01 -2.87136346e-01
1.81276605e-01 2.21444398e-01 -2.58554488e-01 8.69305849e-01
-5.35113096e-01 5.86798310e-01 -5.96980214e-01 -6.72153592e-01
1.20956607e-01 4.51872677e-01 -8.13544333e-01 -7.95479640e-02
2.33707756e-01 -2.07917243e-01 -4.07229424e-01 4.89943385e-01
7.70818233e-01 -6.26681983e-01 3.09343054e-03 -5.36238313e-01
-2.51397043e-01 1.89429387e-01 -1.40120566e+00 2.20919108e+00
-4.53743875e-01 5.33613503e-01 -1.76503032e-01 -1.04557049e+00
8.09819996e-01 -1.56419173e-01 3.05792958e-01 -8.96409810e-01
-1.01337455e-01 4.93522644e-01 -4.25763339e-01 -7.03663155e-02
8.10665488e-01 6.68494701e-01 3.14797498e-02 5.59906840e-01
-1.79702491e-01 -3.24892491e-01 3.24389756e-01 2.14309916e-01
1.43822229e+00 -3.05628985e-01 3.42005551e-01 -7.21803606e-01
4.87789422e-01 3.03836673e-01 -1.15852527e-01 1.03089821e+00
1.63559869e-01 6.44302607e-01 1.99167565e-01 -7.67539859e-01
-1.15350211e+00 -7.05261588e-01 -3.79041940e-01 1.34104431e+00
2.97367990e-01 -1.06564212e+00 -6.36856854e-01 -1.10793918e-01
1.66682169e-01 3.10542792e-01 -1.71493307e-01 -4.34813872e-02
-5.51354051e-01 -7.29468942e-01 4.50583071e-01 1.35760605e-01
3.61035466e-01 -5.56592762e-01 -1.05788541e+00 2.23449260e-01
4.09580588e-01 -1.08302832e+00 -6.24844611e-01 2.55566806e-01
-8.78604770e-01 -9.49958622e-01 -6.79752469e-01 -6.54196560e-01
1.65734202e-01 4.50223058e-01 1.30466235e+00 1.56542122e-01
-7.82555044e-01 3.67953360e-01 -4.01709795e-01 1.69759154e-01
8.83736387e-02 5.34602404e-01 1.86087295e-01 -4.21570808e-01
4.68660265e-01 -7.49653697e-01 -1.06056964e+00 4.02417421e-01
-7.20973969e-01 -7.99199790e-02 5.51747262e-01 9.45472598e-01
9.80833471e-01 -1.02330849e-01 -1.49831176e-01 -6.85763240e-01
5.36935508e-01 -4.76799816e-01 -8.65025938e-01 6.42605796e-02
-1.03362310e+00 3.99240255e-01 5.75160086e-01 -3.16854894e-01
-2.70036012e-01 3.60340923e-01 -2.20617011e-01 -4.23943102e-01
1.89084187e-01 1.57426476e-01 5.35743535e-01 -5.36258578e-01
8.14275801e-01 3.64572734e-01 1.51392445e-01 -4.19968814e-01
4.52963054e-01 6.22794330e-01 6.34892225e-01 -6.54061735e-01
4.20825332e-01 6.84845090e-01 4.63572562e-01 -8.99424434e-01
-3.86802316e-01 -9.89171803e-01 -3.21756005e-01 3.42327833e-01
5.78499079e-01 -8.27483952e-01 -9.17041719e-01 1.46591604e-01
-1.04086363e+00 8.12427774e-02 -1.83397442e-01 4.92485762e-01
-6.82176411e-01 5.77860892e-01 -6.58700109e-01 -1.88712209e-01
-8.30102384e-01 -1.19957709e+00 1.59363234e+00 -8.22002813e-02
-3.27235341e-01 -6.21123791e-01 2.21821859e-01 1.70543149e-01
6.45237088e-01 -1.25380337e-01 6.28154457e-01 -7.15826869e-01
-6.93245649e-01 -2.64953494e-01 -4.55637068e-01 -2.41164073e-01
-3.21642995e-01 -3.11466604e-01 -6.44848049e-01 -5.17863572e-01
-1.84156686e-01 -2.30571136e-01 8.48968863e-01 7.97167271e-02
1.35031271e+00 -1.44995064e-01 -6.17090821e-01 1.09631777e+00
1.66106439e+00 -2.41139516e-01 5.93892455e-01 5.12194455e-01
5.36360621e-01 4.21091288e-01 7.59477675e-01 8.19096863e-01
3.04631054e-01 1.27692938e+00 3.10385853e-01 6.72491733e-04
-9.88442674e-02 2.02654570e-01 -2.02975258e-01 8.93989503e-01
1.53208017e-01 -1.12140231e-01 -9.99440134e-01 4.82647151e-01
-1.78937304e+00 -6.91067278e-01 -4.24561679e-01 2.36023951e+00
6.96754098e-01 1.46511476e-03 9.16422233e-02 2.30927721e-01
3.77800703e-01 3.00845895e-02 -4.37572241e-01 -5.18214822e-01
1.94473155e-02 6.07088387e-01 9.32644725e-01 3.83737594e-01
-9.79159236e-01 7.58938134e-01 6.02671671e+00 1.48537552e+00
-1.02381194e+00 9.52009112e-02 4.06889349e-01 -4.84564483e-01
-6.46103844e-02 1.58035532e-02 -1.11864614e+00 6.29488885e-01
1.07635796e+00 -1.90290526e-01 4.33550000e-01 1.14535582e+00
-2.60881960e-01 -2.58471102e-01 -1.25032496e+00 1.78651929e+00
1.79059915e-02 -1.73338604e+00 -1.25737071e-01 3.84726316e-01
5.23783982e-01 3.54589134e-01 5.45568019e-02 -7.60672800e-03
-1.48858249e-01 -8.74719501e-01 3.68414909e-01 3.81548852e-01
8.34481001e-01 -6.29789233e-01 5.70358455e-01 1.92626864e-01
-1.19414675e+00 2.71882594e-01 -6.82462692e-01 4.69018593e-02
1.57608334e-02 1.00751555e+00 -5.03490627e-01 3.67579222e-01
1.00546646e+00 3.47394824e-01 -6.44065678e-01 1.06573582e+00
8.07455897e-01 3.39012086e-01 -9.94638145e-01 -3.36982518e-01
2.98551470e-01 -1.92746088e-01 2.57941514e-01 1.27996147e+00
7.20423281e-01 1.26583561e-01 -2.63419487e-02 1.64470211e-01
1.88374624e-01 5.47779560e-01 -6.06793165e-01 3.80453944e-01
6.36220634e-01 1.18299401e+00 -7.92254925e-01 -5.36472678e-01
-4.38608527e-01 1.41513181e+00 4.38074231e-01 -2.24102169e-01
-7.11255312e-01 -7.38171935e-01 7.75517642e-01 1.30967528e-01
6.42543554e-01 -3.45192552e-01 -2.11009711e-01 -8.33558917e-01
2.56339490e-01 -8.11706483e-01 4.98883486e-01 -4.40275908e-01
-1.24713719e+00 6.32894814e-01 -1.65238306e-01 -1.15215564e+00
-4.74161714e-01 -6.25307977e-01 -2.04562306e-01 7.50184000e-01
-1.18809569e+00 -5.95882177e-01 -4.90415663e-01 6.46528423e-01
2.62804687e-01 -2.18265682e-01 9.11377072e-01 6.04994893e-01
-6.74675927e-02 8.64621222e-01 6.17302239e-01 -6.61742866e-01
6.97277665e-01 -8.29819858e-01 8.42591465e-01 1.94726169e-01
4.58373398e-01 6.03841066e-01 6.82138145e-01 -3.27202320e-01
-2.13630652e+00 -7.08387434e-01 9.27689195e-01 -2.75056511e-01
8.47889006e-01 -3.87017190e-01 -8.45461607e-01 -1.03830770e-02
-1.11227088e-01 6.09052598e-01 7.20727503e-01 1.87203735e-02
-4.60222721e-01 -1.84385970e-01 -1.10567069e+00 4.36372608e-01
1.53141999e+00 -5.87426007e-01 -2.14667737e-01 8.36650431e-01
6.47207558e-01 -6.35422051e-01 -1.16623247e+00 8.54425356e-02
6.29796445e-01 -1.26680243e+00 1.36758137e+00 -2.04750150e-01
2.56672371e-02 -1.95996955e-01 -5.19896567e-01 -7.75879622e-01
-3.68624806e-01 -7.13026226e-01 -2.11052105e-01 7.19028592e-01
9.41422582e-03 -6.28692031e-01 1.03913450e+00 2.48926312e-01
5.58854677e-02 -1.16606331e+00 -1.08255041e+00 -9.46074069e-01
-3.29635143e-01 -4.54120249e-01 7.21275747e-01 7.74135709e-01
-8.95764679e-02 2.44139299e-01 -1.70793727e-01 3.62333395e-02
7.32113004e-01 3.85321140e-01 8.92319798e-01 -1.01605356e+00
-4.95197922e-01 -5.69122672e-01 -9.28375125e-01 -1.43082869e+00
-8.82363543e-02 -9.02019799e-01 -5.88028669e-01 -1.14832449e+00
1.60277098e-01 -6.73923731e-01 2.14518949e-01 1.89494655e-01
1.17450856e-01 6.23519719e-01 -1.18995821e-02 4.69422013e-01
-7.57310390e-01 3.31564188e-01 6.59612656e-01 7.52292424e-02
6.91017658e-02 -2.61374086e-01 -2.67311245e-01 3.40168327e-01
4.69402462e-01 -4.07750368e-01 -2.62905061e-01 -4.79423225e-01
4.11502272e-01 8.12341794e-02 2.42387831e-01 -1.22139978e+00
8.24705899e-01 2.73833394e-01 1.00938678e-01 -8.32166314e-01
5.45826435e-01 -8.44729722e-01 3.39926690e-01 6.25750363e-01
-8.34942088e-02 4.71130520e-01 8.80155638e-02 4.27415013e-01
-1.60643145e-01 -3.30973715e-01 4.67500627e-01 1.24315739e-01
-7.03474343e-01 3.84587973e-01 8.31446499e-02 -9.52754840e-02
8.83582950e-01 -3.84103417e-01 -3.39805126e-01 -1.49966076e-01
-3.15224767e-01 5.59888110e-02 6.42091095e-01 3.97445261e-02
4.15705472e-01 -1.46352875e+00 -7.41272151e-01 1.83602899e-01
1.18046016e-01 -1.81304455e-01 2.51339674e-01 7.95301199e-01
-1.05991685e+00 7.52250493e-01 -8.56694505e-02 -9.86003578e-01
-1.50672174e+00 7.76391327e-01 -1.57164842e-01 -3.14247966e-01
-8.50116372e-01 7.78398812e-01 -4.49827686e-02 -5.55028804e-02
1.80483773e-01 -1.38210610e-01 1.67592093e-01 7.31664822e-02
6.42103136e-01 7.77889013e-01 6.56184912e-01 -4.50510561e-01
-5.20361900e-01 1.01291656e+00 -2.06494913e-01 1.59694597e-01
1.13497579e+00 -3.51634249e-02 -1.61557212e-01 -1.33834302e-01
1.89063561e+00 -8.99924263e-02 -7.07107842e-01 -2.34339163e-01
3.22332233e-02 -6.78048670e-01 2.14716807e-01 -6.61307247e-03
-8.98736596e-01 4.85439688e-01 9.82070327e-01 1.94187894e-01
1.10606337e+00 1.64771482e-01 1.16562974e+00 8.98508728e-01
8.22009921e-01 -1.02684188e+00 -1.77095354e-01 4.85427111e-01
6.03123367e-01 -1.23273361e+00 5.06470382e-01 -4.73659217e-01
-1.11845113e-01 1.26618338e+00 9.03808028e-02 -4.04813290e-01
6.35424793e-01 3.26541483e-01 -4.22254384e-01 -3.78746450e-01
-8.25007200e-01 -1.14819594e-01 4.33081657e-01 3.82317275e-01
1.53008521e-01 -9.97073874e-02 -6.56113386e-01 -2.30321418e-02
-5.98251939e-01 -2.51921207e-01 -3.39269266e-02 9.64358211e-01
-3.68059993e-01 -1.16171110e+00 -4.36895221e-01 7.30868757e-01
-3.57438356e-01 -4.69242156e-01 1.90880999e-01 6.07883155e-01
-2.08114788e-01 6.06071770e-01 4.88558978e-01 -3.35203886e-01
2.64497727e-01 -3.81518364e-01 5.48334837e-01 -1.99940503e-01
-7.10482836e-01 -3.03053737e-01 4.53988239e-02 -1.16622543e+00
-8.43966305e-02 -7.32932210e-01 -1.01625526e+00 -7.18017638e-01
-2.52044320e-01 9.66387242e-02 9.64735687e-01 4.93781686e-01
9.29615140e-01 -2.34285265e-01 4.61607993e-01 -1.00887883e+00
-8.06837916e-01 -4.38081414e-01 -5.27520299e-01 4.34360147e-01
-3.48525308e-02 -5.08935332e-01 -3.02930027e-01 -3.45220625e-01] | [8.565876960754395, 3.425962448120117] |
7cd0ae01-5954-49da-ba54-28233ddc3ee2 | pdsum-prototype-driven-continuous | 2302.05550 | null | https://arxiv.org/abs/2302.05550v1 | https://arxiv.org/pdf/2302.05550v1.pdf | PDSum: Prototype-driven Continuous Summarization of Evolving Multi-document Sets Stream | Summarizing text-rich documents has been long studied in the literature, but most of the existing efforts have been made to summarize a static and predefined multi-document set. With the rapid development of online platforms for generating and distributing text-rich documents, there arises an urgent need for continuously summarizing dynamically evolving multi-document sets where the composition of documents and sets is changing over time. This is especially challenging as the summarization should be not only effective in incorporating relevant, novel, and distinctive information from each concurrent multi-document set, but also efficient in serving online applications. In this work, we propose a new summarization problem, Evolving Multi-Document sets stream Summarization (EMDS), and introduce a novel unsupervised algorithm PDSum with the idea of prototype-driven continuous summarization. PDSum builds a lightweight prototype of each multi-document set and exploits it to adapt to new documents while preserving accumulated knowledge from previous documents. To update new summaries, the most representative sentences for each multi-document set are extracted by measuring their similarities to the prototypes. A thorough evaluation with real multi-document sets streams demonstrates that PDSum outperforms state-of-the-art unsupervised multi-document summarization algorithms in EMDS in terms of relevance, novelty, and distinctiveness and is also robust to various evaluation settings. | ['Jiawei Han', 'Hou Pong Chan', 'Susik Yoon'] | 2023-02-10 | null | null | null | null | ['multi-document-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.76423478e-01 -4.11702275e-01 -5.55776954e-02 -8.93981382e-02
-9.27488804e-01 -7.14773715e-01 5.98752320e-01 1.17361069e+00
-1.69150963e-01 7.90572762e-01 6.54344380e-01 5.39270997e-01
-4.25012529e-01 -6.16508007e-01 -2.24497870e-01 -5.73137820e-01
-2.04962820e-01 6.49436414e-01 5.75046778e-01 -3.36411893e-01
8.26565921e-01 3.32565993e-01 -1.81574905e+00 4.17727649e-01
1.31001997e+00 4.93759781e-01 4.20022666e-01 1.24121308e+00
-4.05431300e-01 2.19736785e-01 -1.30652356e+00 -1.60228312e-01
-8.38802084e-02 -7.98787415e-01 -7.07468569e-01 2.93245733e-01
4.04802918e-01 -1.17415361e-01 -6.32290691e-02 8.68109167e-01
7.94449627e-01 4.87014234e-01 5.57090640e-01 -8.78315032e-01
-3.44883382e-01 9.19915259e-01 -7.01732099e-01 5.90165257e-01
5.51996291e-01 -3.80438149e-01 1.01403236e+00 -5.57890117e-01
6.56030655e-01 1.07604027e+00 3.87504667e-01 2.55811930e-01
-7.29045451e-01 -5.86881191e-02 3.22735846e-01 1.56351388e-01
-1.08286500e+00 -5.34355700e-01 7.47605383e-01 -4.07640561e-02
9.67108905e-01 7.24387348e-01 7.32341349e-01 5.61998069e-01
3.50673944e-01 1.04855609e+00 1.69125631e-01 -5.13637662e-01
3.65874082e-01 2.42162198e-02 4.93458629e-01 3.02997649e-01
7.24171817e-01 -1.00334001e+00 -6.57647908e-01 -4.39857662e-01
-1.25465281e-02 4.83248606e-02 -3.49204242e-01 1.08993120e-01
-1.21908188e+00 5.60373783e-01 -4.53255534e-01 7.19560027e-01
-7.17306376e-01 -2.81916797e-01 9.51508939e-01 2.12971762e-01
6.69360816e-01 4.93319273e-01 -7.51777217e-02 -4.19301480e-01
-1.44767833e+00 4.98477489e-01 8.30711901e-01 9.82575715e-01
4.99601364e-01 7.20870718e-02 -4.74680960e-01 1.00455391e+00
-1.41071409e-01 3.88192296e-01 1.18846989e+00 -6.73811197e-01
6.78195715e-01 8.95680964e-01 -1.02251448e-01 -1.20061469e+00
-2.57132471e-01 -4.91602331e-01 -1.13863480e+00 -7.69266427e-01
-5.41394651e-01 -4.68775220e-02 -3.89689088e-01 1.16188681e+00
3.77815425e-01 -1.87826470e-01 3.69750649e-01 1.45745158e-01
9.11049783e-01 1.04827380e+00 -5.41024804e-01 -1.00863290e+00
1.00847042e+00 -9.34003472e-01 -7.37189293e-01 1.17430156e-02
2.15480119e-01 -8.13434362e-01 6.66798890e-01 3.91987920e-01
-1.45929623e+00 -6.12246692e-01 -1.21466720e+00 3.65190804e-01
-6.89070523e-02 -7.82267153e-02 1.12258293e-01 5.23652494e-01
-9.83531415e-01 6.05132520e-01 -4.85048592e-01 -7.83225536e-01
2.11987212e-01 8.92860517e-02 -5.02319708e-02 2.27757152e-02
-8.74060273e-01 4.45391923e-01 9.27040756e-01 -2.93441594e-01
-3.80091459e-01 -4.71777797e-01 -5.22546947e-01 3.17475230e-01
2.13440746e-01 -8.67904365e-01 1.34131718e+00 -6.99414253e-01
-1.40676320e+00 1.96274847e-01 -3.82484764e-01 -3.43555063e-01
4.36276615e-01 -1.61823675e-01 -6.45805001e-01 7.29491889e-01
3.22120726e-01 7.55292699e-02 8.67586672e-01 -1.28275836e+00
-1.19902837e+00 -3.51486742e-01 -6.04326010e-01 4.88199323e-01
-9.22444761e-01 8.79285187e-02 -4.79542822e-01 -7.50000179e-01
-1.82850193e-02 -5.24977088e-01 -4.48556989e-02 -9.89180744e-01
-5.93912125e-01 -3.02570850e-01 1.10783803e+00 -4.93729502e-01
2.02906227e+00 -1.94745636e+00 1.93742126e-01 3.79576311e-02
1.48131028e-01 4.99910295e-01 -1.69557452e-01 1.13422263e+00
3.86766583e-01 1.27140060e-01 -4.06756788e-01 -5.49311042e-01
-2.61582166e-01 1.31312668e-01 -5.39281011e-01 3.16017196e-02
-1.04331560e-01 5.33826947e-01 -1.17231083e+00 -8.86721969e-01
-7.49894306e-02 -1.20737232e-01 -2.38099501e-01 2.09928807e-02
-2.21313655e-01 1.64160818e-01 -5.29791951e-01 5.54519773e-01
5.35094976e-01 -6.10260852e-02 8.45536068e-02 1.27708867e-01
-3.98072749e-01 -3.28303695e-01 -1.24909461e+00 1.63715625e+00
-6.66748732e-02 5.88363588e-01 -3.82799715e-01 -1.07640207e+00
9.95784402e-01 3.41866851e-01 7.87478566e-01 -3.69143814e-01
-1.51202418e-02 2.53290266e-01 -3.52278620e-01 -6.39484763e-01
1.68046331e+00 2.21650630e-01 -2.95881838e-01 7.98857570e-01
-6.02625795e-02 -3.62274647e-01 1.22867787e+00 8.04564476e-01
1.19086826e+00 -6.52038693e-01 6.01598918e-01 -2.18850877e-02
7.23872006e-01 1.62005574e-01 4.12256449e-01 8.63794982e-01
1.83784276e-01 7.91810572e-01 2.60390013e-01 7.00102672e-02
-9.36990619e-01 -7.82046378e-01 1.37449399e-01 9.34742570e-01
1.45156711e-01 -8.90236735e-01 -6.37781978e-01 -4.39842761e-01
5.27041778e-02 8.51453304e-01 -2.53051102e-01 -3.37629974e-01
-6.51675344e-01 -7.23488271e-01 5.41103125e-01 2.54689366e-01
5.47491133e-01 -9.76073503e-01 -8.59510183e-01 6.71739876e-01
-3.54602635e-01 -6.99885666e-01 -7.67912388e-01 -3.36767465e-01
-1.05705893e+00 -8.46396327e-01 -9.40283358e-01 -7.33915329e-01
6.98016524e-01 7.44580507e-01 8.27038646e-01 -1.57230213e-01
-7.56491944e-02 8.04117620e-01 -8.10989559e-01 -5.86257994e-01
-9.83072162e-01 5.20701766e-01 8.47878084e-02 1.14351176e-01
-2.08824590e-01 -5.40536404e-01 -3.04653794e-01 -2.16791383e-03
-1.52158952e+00 -9.43154022e-02 6.34674966e-01 5.47802150e-01
6.31869614e-01 5.89219928e-01 1.25416708e+00 -8.03026438e-01
1.53851831e+00 -6.13919497e-01 6.35726154e-02 6.63817644e-01
-6.53110802e-01 -2.64563803e-02 1.03331184e+00 -5.33897817e-01
-1.16992223e+00 -4.96116430e-01 2.14525327e-01 4.24247719e-02
1.46954849e-01 8.71857226e-01 1.77798346e-01 5.86934865e-01
4.80732024e-01 8.04579616e-01 -2.15417176e-01 -3.68537128e-01
3.52379292e-01 1.02144790e+00 8.44035387e-01 -3.52258474e-01
4.74766880e-01 4.45331305e-01 -1.99138194e-01 -1.33836246e+00
-5.83241165e-01 -1.02130795e+00 -6.69636726e-01 -3.32148165e-01
2.25889534e-01 -4.55688536e-01 -2.24779546e-01 5.55836380e-01
-1.27206624e+00 4.72125381e-01 -7.43651211e-01 1.21298566e-01
-2.74758220e-01 1.17196858e+00 -1.29856020e-01 -6.07623398e-01
-1.23127031e+00 -4.57668751e-01 9.93275225e-01 5.53299844e-01
-6.23479187e-01 -9.56817389e-01 5.98403156e-01 -9.60833952e-02
4.31827039e-01 3.06897610e-01 8.42810810e-01 -1.04077590e+00
-1.70598086e-02 -7.41678655e-01 3.01964521e-01 4.43428308e-01
6.01384282e-01 4.07554358e-01 -3.64239305e-01 -6.03726506e-01
1.07822299e-01 7.70727098e-02 1.02552760e+00 4.81080770e-01
7.66485393e-01 -6.34272337e-01 -3.86249363e-01 -4.95496355e-02
1.13383567e+00 2.08907112e-01 1.32661343e-01 2.50180066e-01
3.36398780e-01 3.43906075e-01 7.19840884e-01 1.23330677e+00
4.38465446e-01 8.05135071e-02 -4.88134399e-02 5.44581771e-01
2.63754100e-01 -3.69985029e-02 4.03759003e-01 1.64242995e+00
1.22614086e-01 -8.60652208e-01 -5.22506058e-01 8.01100135e-01
-2.18087888e+00 -1.43866730e+00 -4.99450378e-02 2.09380436e+00
8.58982980e-01 2.06816509e-01 3.02111268e-01 3.88894975e-01
1.02816164e+00 4.09305155e-01 -6.83384120e-01 -5.90123057e-01
-5.27629554e-01 -1.21126302e-01 -8.83843228e-02 -3.81605290e-02
-7.56721318e-01 4.65505481e-01 5.77435350e+00 7.76850939e-01
-9.04301822e-01 -2.44621187e-01 1.26642063e-01 -4.01381910e-01
-4.36053127e-01 -7.58208930e-02 -7.82317579e-01 6.11891270e-01
1.04785395e+00 -1.25685501e+00 -1.48845345e-01 5.33624232e-01
3.71266723e-01 -5.74903131e-01 -9.23784018e-01 9.24247205e-01
7.44129896e-01 -1.48796940e+00 7.05906987e-01 -4.14930552e-01
1.18320572e+00 -2.37436190e-01 -1.83906436e-01 7.46355206e-02
-2.12744251e-02 -1.69256225e-01 6.45043194e-01 6.21694803e-01
3.79973263e-01 -9.85873938e-01 6.70831203e-01 6.34492695e-01
-1.37048483e+00 -8.36147964e-02 -3.53584677e-01 3.56697470e-01
4.63988870e-01 8.39601457e-01 -8.33774626e-01 1.17684710e+00
4.27178860e-01 9.62332487e-01 -7.47980416e-01 1.30224216e+00
2.97288179e-01 4.69189793e-01 -1.54933989e-01 -2.98707217e-01
2.04937577e-01 -1.66017458e-01 1.09444261e+00 1.61132228e+00
6.39113188e-01 -9.16296020e-02 3.26830268e-01 3.05638593e-02
-7.23670572e-02 4.75550741e-01 -2.74556845e-01 -3.31003815e-01
6.91898465e-01 1.27577519e+00 -7.95191109e-01 -8.69109392e-01
2.28722841e-02 1.10223246e+00 -2.00975716e-01 4.84286807e-03
-3.21858674e-01 -9.15556133e-01 1.36516318e-01 -1.49620712e-01
2.77992189e-01 -2.49189541e-01 -8.57274160e-02 -1.12801301e+00
3.05186749e-01 -7.81224489e-01 7.93770134e-01 -3.60159755e-01
-1.01150465e+00 5.93359172e-01 2.78812468e-01 -1.43703961e+00
-5.57548404e-01 5.18815398e-01 -1.00726438e+00 2.99903214e-01
-1.22564173e+00 -6.75978541e-01 -4.80430305e-01 2.87601203e-01
1.11635041e+00 -4.48060006e-01 7.12974966e-01 -1.72400139e-02
-6.54076397e-01 3.19900751e-01 8.31865668e-01 -5.65069497e-01
7.66287565e-01 -1.18211687e+00 4.40322727e-01 1.07589543e+00
-4.07157280e-02 5.16227245e-01 8.94068241e-01 -8.70837629e-01
-1.41157234e+00 -1.14041269e+00 8.43043447e-01 2.71642068e-03
4.80881065e-01 2.06511706e-01 -1.06308126e+00 2.22278878e-01
4.54251945e-01 -7.20586300e-01 8.34479272e-01 -4.33795482e-01
3.00254852e-01 -2.72721857e-01 -1.05652511e+00 4.77740973e-01
6.14598393e-01 9.87521559e-02 -8.04518461e-01 3.28162819e-01
6.65389776e-01 -1.81009009e-01 -6.53186381e-01 1.29477277e-01
3.84619594e-01 -8.16718876e-01 5.21938562e-01 -3.69643986e-01
3.27196062e-01 -3.07210475e-01 -8.75763688e-03 -1.80175424e+00
-2.60813385e-01 -9.65416133e-01 -4.82869148e-01 1.72620344e+00
-1.05981469e-01 -5.27147770e-01 4.98184025e-01 -6.43381029e-02
-4.85176414e-01 -5.74395001e-01 -6.47356510e-01 -7.48722732e-01
-2.81465471e-01 4.68980968e-02 6.29944205e-01 5.84615827e-01
3.41500282e-01 5.31191766e-01 -2.19367817e-01 -6.01586998e-02
5.33118606e-01 4.96193588e-01 9.66773331e-01 -1.44070172e+00
-8.65942836e-02 -7.93801129e-01 -2.83622146e-01 -9.69328284e-01
-1.25080064e-01 -8.59048605e-01 -3.75081487e-02 -2.21043015e+00
5.06890714e-01 1.40618622e-01 -1.31538212e-01 -1.48511812e-01
-4.09802169e-01 -4.50176686e-01 1.43493474e-01 6.03215337e-01
-1.25594616e+00 7.91528344e-01 9.88685608e-01 -3.34381282e-01
-7.46851802e-01 3.02610666e-01 -9.78339016e-01 4.65639770e-01
8.78073454e-01 -4.42910314e-01 -5.25599599e-01 -2.62483388e-01
1.17516451e-01 -3.39845978e-02 -4.56722975e-01 -1.15975952e+00
7.81749666e-01 -8.57644528e-02 7.32002854e-02 -1.37299657e+00
-1.30727217e-01 -2.21051186e-01 1.54532986e-02 3.86413693e-01
-2.78306901e-01 5.10863423e-01 3.03105921e-01 7.29798436e-01
-5.54700494e-01 -5.97399175e-01 5.12205780e-01 -1.57418400e-01
-7.15632737e-01 2.17489317e-01 -5.74438155e-01 3.22872549e-01
1.03401124e+00 -5.40570378e-01 -4.09911066e-01 -2.67957181e-01
-7.65113086e-02 5.76466143e-01 4.16844845e-01 4.48590428e-01
9.16575611e-01 -9.23519671e-01 -1.23181558e+00 -1.18079633e-01
2.73973078e-01 3.15419197e-01 5.97615421e-01 3.92860025e-01
-4.41286623e-01 3.03550810e-01 -5.49238361e-02 -5.73524058e-01
-1.60533154e+00 2.79268146e-01 -3.46097469e-01 -4.54708099e-01
-7.16167271e-01 3.62219125e-01 -5.38041830e-01 2.84541398e-01
8.40428397e-02 -2.81836629e-01 -4.62478340e-01 7.22723842e-01
8.32759082e-01 9.09275055e-01 1.97176799e-01 -5.00460684e-01
-2.24486422e-02 4.69764382e-01 -6.00998461e-01 -4.26155515e-02
1.58779454e+00 -3.61927927e-01 -4.15334046e-01 7.16609597e-01
9.69957709e-01 2.28270143e-01 -5.95580518e-01 -3.56021404e-01
2.77473956e-01 -2.36953527e-01 -4.21296060e-01 -3.32304686e-01
-5.19581974e-01 2.71292657e-01 -5.50997332e-02 6.86815262e-01
1.28630805e+00 -1.94706060e-02 1.09378266e+00 6.12278819e-01
2.09801644e-02 -1.42401958e+00 6.31277323e-01 6.03002727e-01
1.03295410e+00 -7.84487009e-01 4.61393118e-01 1.50882334e-01
-7.75381386e-01 1.25050938e+00 1.75860748e-01 1.68604627e-01
2.99122036e-01 -4.33512926e-02 -4.01421934e-01 -1.58070549e-02
-8.76299739e-01 2.18368128e-01 9.77327898e-02 2.65464693e-01
5.50450943e-02 -1.03912652e-01 -4.43806589e-01 5.70270479e-01
-2.62338549e-01 -2.60667115e-01 8.53133261e-01 1.31891978e+00
-1.06681955e+00 -1.07762027e+00 -4.78899539e-01 9.21113431e-01
-2.10510015e-01 2.54999936e-01 -5.19683003e-01 3.13659608e-01
-3.74897540e-01 1.13241315e+00 7.18635470e-02 -1.48048192e-01
6.29679263e-01 -4.02139202e-02 3.06121498e-01 -8.82576942e-01
-7.52317607e-01 -1.66193340e-02 -1.14051297e-01 3.03423107e-01
-4.94995058e-01 -1.05092502e+00 -1.41800106e+00 -3.49805176e-01
-4.10054713e-01 6.51561797e-01 6.08354688e-01 7.45442212e-01
6.99692786e-01 7.51656771e-01 1.02483988e+00 -8.47241342e-01
-7.45050430e-01 -1.11087656e+00 -6.03354931e-01 3.62319648e-01
3.54825318e-01 -8.95592757e-03 -1.95002109e-01 1.02073900e-01] | [12.61446762084961, 9.558448791503906] |
f20aee3f-f7ac-4fa8-b795-3f03a0bc78fb | eulerian-phase-based-motion-magnification-for | 2212.04923 | null | https://arxiv.org/abs/2212.04923v1 | https://arxiv.org/pdf/2212.04923v1.pdf | Eulerian Phase-based Motion Magnification for High-Fidelity Vital Sign Estimation with Radar in Clinical Settings | Efficient and accurate detection of subtle motion generated from small objects in noisy environments, as needed for vital sign monitoring, is challenging, but can be substantially improved with magnification. We developed a complex Gabor filter-based decomposition method to amplify phases at different spatial wavelength levels to magnify motion and extract 1D motion signals for fundamental frequency estimation. The phase-based complex Gabor filter outputs are processed and then used to train machine learning models that predict respiration and heart rate with greater accuracy. We show that our proposed technique performs better than the conventional temporal FFT-based method in clinical settings, such as sleep laboratories and emergency departments, as well for a variety of human postures. | ['Tauhidur Rahman', 'Suren Jayasuriya', 'Deepak Ganesan', 'Stephanie Carreiro', 'Toral Surti', 'Md Farhan Tasnim Oshim'] | 2022-12-03 | null | null | null | null | ['motion-magnification'] | ['computer-vision'] | [ 4.49766129e-01 -4.77026761e-01 2.63559788e-01 8.27560127e-02
-4.91068214e-01 -6.18604720e-01 -3.36521447e-01 -5.03355153e-02
-6.82969987e-01 8.49979699e-01 1.48805588e-01 -3.81359607e-01
-1.45257354e-01 -2.25181386e-01 1.14901468e-01 -1.05352664e+00
-4.44367975e-01 -1.40652031e-01 4.43174034e-01 3.02245378e-01
2.90337801e-01 6.82457924e-01 -9.01718199e-01 1.98724583e-01
3.60856742e-01 8.98402333e-01 5.43138273e-02 1.45219290e+00
6.27668440e-01 6.89177215e-01 -1.01686668e+00 2.50367284e-01
8.52763504e-02 -6.72469497e-01 -3.59459341e-01 -3.25823784e-01
1.48435712e-01 -2.62621403e-01 -2.58305967e-01 5.96821785e-01
1.15779543e+00 3.05670470e-01 5.89610875e-01 -2.73718148e-01
-1.12864748e-01 5.41703291e-02 -6.19860351e-01 1.23391056e+00
3.58782768e-01 3.22763503e-01 4.75807600e-02 -6.41826093e-01
3.02781045e-01 6.74570024e-01 1.02091873e+00 4.88830000e-01
-1.03873205e+00 -5.23018599e-01 -7.92683423e-01 4.03123707e-01
-1.01489961e+00 -2.55087644e-01 8.63723099e-01 -1.95988670e-01
1.17113960e+00 5.93060851e-01 8.65412474e-01 6.14184380e-01
8.70689154e-01 -6.22263253e-02 1.19989491e+00 -4.19912994e-01
-3.39644775e-02 -1.76563561e-01 6.95869699e-02 8.30365419e-01
5.07906318e-01 -1.37134148e-02 -5.04057586e-01 -2.52542973e-01
1.01065385e+00 1.77822411e-02 -8.46088409e-01 4.20182437e-01
-1.65206611e+00 3.93670499e-01 7.95518421e-03 5.61723948e-01
-6.24682724e-01 3.51784915e-01 1.33029491e-01 -8.93901363e-02
8.77200291e-02 7.04906642e-01 -1.35816887e-01 -8.95856023e-01
-9.66727972e-01 -2.27710366e-01 7.94824600e-01 3.03322487e-02
-7.14276955e-02 3.04462135e-01 -3.94826859e-01 3.04381371e-01
1.76045090e-01 7.81974018e-01 8.27206850e-01 -1.18824840e+00
1.44604951e-01 -1.90373257e-01 4.33480620e-01 -9.15516019e-01
-1.03321409e+00 -4.64108884e-01 -8.25428128e-01 -7.95846805e-02
3.97289515e-01 -2.76216865e-01 -7.68429458e-01 1.05051613e+00
5.60507417e-01 5.56097329e-01 1.00806430e-01 1.10753655e+00
9.17202652e-01 4.19621438e-01 3.71146612e-02 -8.20325077e-01
1.60207164e+00 -4.39278185e-01 -1.18777144e+00 -6.76307231e-02
4.46179360e-01 -1.03659964e+00 7.46350288e-01 4.81316179e-01
-1.09774458e+00 -5.42001426e-01 -1.03439951e+00 2.93085068e-01
3.92041713e-01 2.30027869e-01 3.06176186e-01 1.13786614e+00
-7.29763985e-01 9.48813856e-01 -1.45916140e+00 -8.85517225e-02
2.32906327e-01 2.16079980e-01 1.10735394e-01 2.93864280e-01
-8.38776767e-01 8.61125290e-01 -1.24507219e-01 4.87121910e-01
7.65252933e-02 -5.82500041e-01 -6.22121871e-01 -1.26762688e-01
-2.52401799e-01 -9.47031975e-01 1.22966993e+00 2.35424832e-01
-1.61644328e+00 4.78699595e-01 -4.87074435e-01 -4.15021628e-01
3.72745335e-01 -8.32428113e-02 -4.80745792e-01 9.58447039e-01
-4.33905333e-01 -1.10058360e-01 1.20186114e+00 -4.79015172e-01
-1.63351431e-01 -1.85529932e-01 -4.46473777e-01 2.14608431e-01
-3.75065468e-02 8.90444741e-02 2.04696909e-01 -6.94872916e-01
6.03145003e-01 -8.74294996e-01 -1.98670462e-01 -1.46206066e-01
1.18061043e-01 3.34773511e-01 1.01931691e+00 -7.47479916e-01
1.43856549e+00 -2.08054519e+00 -3.57271433e-01 2.46763509e-02
4.45530534e-01 5.58481753e-01 3.60591322e-01 1.32116839e-01
1.01819649e-01 -1.28313988e-01 -5.56383245e-02 1.42425165e-01
-6.81585193e-01 -2.29885373e-02 4.94098291e-03 8.29000056e-01
1.37829781e-01 7.94361770e-01 -9.65538800e-01 -5.04756987e-01
4.23652589e-01 7.81459391e-01 -1.48419246e-01 9.18900594e-02
6.59934402e-01 9.74242687e-01 -2.66073763e-01 5.56135416e-01
5.67360222e-01 -3.21442783e-01 1.57106239e-02 -8.18758011e-01
-7.13364109e-02 3.32443416e-01 -1.11172235e+00 1.04308677e+00
-3.83592665e-01 8.23924184e-01 -2.43381411e-01 -9.26932633e-01
7.29413092e-01 6.15461349e-01 8.23108971e-01 -6.67606294e-01
2.16711059e-01 1.93184778e-01 3.35695595e-01 -1.35775828e+00
9.77907777e-02 -7.28813410e-01 2.86662906e-01 5.62797368e-01
-3.76673102e-01 -2.35480353e-01 -6.82249740e-02 -4.70352560e-01
1.22938216e+00 -2.40641519e-01 3.52797955e-01 -1.37585118e-01
3.17406893e-01 -1.95843562e-01 3.28739285e-01 7.82952130e-01
-6.51871502e-01 6.83257580e-01 -8.53308216e-02 -6.56994462e-01
-4.35651690e-01 -9.28071260e-01 -2.93020606e-01 2.11347997e-01
1.54506773e-01 -1.51570976e-01 -5.15410125e-01 -2.64027029e-01
-2.35902458e-01 2.88317621e-01 -1.75003141e-01 -3.21851313e-01
-1.09707487e+00 -1.15722561e+00 6.26058936e-01 6.48738444e-01
2.77599961e-01 -1.13949430e+00 -1.88119042e+00 5.60873568e-01
-6.53674066e-01 -1.29691696e+00 -5.45242667e-01 2.07927227e-01
-1.07297862e+00 -1.10842609e+00 -8.67646515e-01 -5.54887474e-01
3.69260520e-01 2.87606239e-01 8.44628751e-01 -1.17555350e-01
-1.10174990e+00 5.56667626e-01 -3.22887480e-01 -6.95971131e-01
-2.25020140e-01 -5.24673820e-01 2.78305233e-01 -7.36973286e-02
2.06004292e-01 -6.41463399e-01 -1.30817819e+00 2.45801523e-01
-4.87170428e-01 -4.13458467e-01 3.93049181e-01 6.81214631e-01
4.29818839e-01 3.75014655e-02 3.10568213e-01 -4.03826535e-01
8.33563030e-01 1.15510210e-01 -2.76542038e-01 -3.28816324e-01
-1.64702982e-01 -1.61069274e-01 4.86678392e-01 -9.59162235e-01
-7.62769699e-01 -3.28539371e-01 -1.15730964e-01 -3.92984599e-01
-1.15345819e-02 2.57476747e-01 8.70652974e-01 -5.48025846e-01
1.10731387e+00 3.27791005e-01 2.59862300e-02 -8.49366039e-02
-2.97645330e-01 5.93128264e-01 8.53475451e-01 -1.02562964e-01
7.36292303e-01 5.83026290e-01 6.72856033e-01 -1.29504371e+00
-2.90745080e-01 -7.43180096e-01 -4.93093342e-01 -3.84897321e-01
1.09970713e+00 -5.44965923e-01 -9.62422848e-01 5.02600312e-01
-1.14149153e+00 -3.58890057e-01 -2.95453995e-01 1.42868793e+00
-4.21591669e-01 7.16627479e-01 -7.58069098e-01 -1.04416752e+00
-7.47859061e-01 -7.22683549e-01 8.33354414e-01 5.19850314e-01
-5.13072431e-01 -9.37941849e-01 2.04381853e-01 2.84667641e-01
7.04105437e-01 6.91776633e-01 7.46387422e-01 9.75118205e-02
-3.06356788e-01 -3.33975941e-01 3.05527210e-01 9.77543816e-02
5.72033465e-01 1.65542942e-02 -9.48518872e-01 -2.86132097e-01
7.81553030e-01 8.74361321e-02 4.79342848e-01 1.11861801e+00
1.08301437e+00 -2.88399696e-01 -3.53970826e-01 8.48998129e-01
1.18107259e+00 4.86519188e-01 6.80995345e-01 -6.29636347e-02
4.41401958e-01 1.57608941e-01 6.07233524e-01 6.41978979e-01
-2.27311254e-01 3.13524485e-01 -1.20177314e-01 -3.50369990e-01
-2.27248549e-01 5.39954484e-01 -1.04390539e-01 9.56629038e-01
-7.73077786e-01 -1.19839143e-03 -5.39231956e-01 2.59644061e-01
-1.08970368e+00 -1.15707088e+00 -3.44117284e-01 2.17070436e+00
6.98113620e-01 1.91953126e-02 2.61169374e-01 4.99094486e-01
3.75891238e-01 -6.75168186e-02 -3.63277674e-01 -4.00581539e-01
2.57384896e-01 8.60261738e-01 3.90565157e-01 3.15633416e-01
-9.86364245e-01 2.05874722e-02 7.83674002e+00 1.08439207e-01
-1.69622028e+00 8.58478844e-02 2.76735097e-01 -1.78527012e-01
2.01622382e-01 -6.03615165e-01 -4.09758657e-01 5.00922859e-01
9.84942079e-01 3.59543413e-02 -4.19305684e-03 2.71105677e-01
7.75935709e-01 -2.80441076e-01 -6.07579410e-01 1.37078059e+00
-2.71072239e-01 -1.13263595e+00 -6.66608691e-01 -1.96446404e-01
3.24563026e-01 -3.50802481e-01 1.51097372e-01 -4.07721937e-01
-8.61581028e-01 -1.02348399e+00 1.79498613e-01 6.04165316e-01
9.11606848e-01 -4.61930454e-01 7.40442753e-01 3.72042984e-01
-1.21928060e+00 3.53092179e-02 -2.98435837e-01 -5.45431525e-02
2.66396374e-01 6.98108256e-01 -1.06888747e+00 1.68984339e-01
6.37195051e-01 2.33641475e-01 -1.14263058e-01 1.40732884e+00
-4.99083959e-02 1.05503368e+00 -5.98842025e-01 -2.54486322e-01
-1.51050970e-01 -1.93010587e-02 8.13645720e-01 1.38168550e+00
6.42317474e-01 6.73808753e-01 -3.64658535e-01 4.52059150e-01
4.33924049e-01 -2.28304595e-01 -2.05980211e-01 4.05856408e-02
2.40899414e-01 1.33260155e+00 -1.12790990e+00 -1.91816971e-01
-3.44989866e-01 7.67258048e-01 -5.13488889e-01 5.72121322e-01
-9.54431117e-01 -8.01665843e-01 3.21602315e-01 4.94605273e-01
2.13084385e-01 -4.96526748e-01 -1.12126961e-01 -1.04237235e+00
2.38620982e-01 -6.40509188e-01 3.68811131e-01 -7.11109877e-01
-6.11059964e-01 5.52954316e-01 -1.75810400e-02 -1.49851727e+00
-3.41061950e-01 -5.40479422e-01 -9.35130239e-01 1.08307469e+00
-1.47309744e+00 -2.87666321e-01 -5.67380428e-01 6.17681682e-01
1.98145166e-01 5.06191313e-01 1.02317810e+00 3.30117077e-01
-1.63968816e-01 4.04368550e-01 -9.21007991e-02 5.87884337e-03
6.07836604e-01 -1.06461072e+00 1.21219933e-01 9.03976679e-01
-3.98803130e-02 5.84123313e-01 9.08478618e-01 -5.12250006e-01
-1.33486009e+00 -8.43849599e-01 7.45793939e-01 -2.88740277e-01
1.37939662e-01 2.07889870e-01 -8.24146330e-01 -1.75637268e-02
-2.33222753e-01 5.29256523e-01 9.18636680e-01 -8.62527192e-01
3.33281219e-01 -8.29251111e-02 -1.39156628e+00 1.82663679e-01
5.14727116e-01 -3.66767347e-01 -7.04432607e-01 3.51877838e-01
3.16662341e-01 -9.09465134e-01 -1.12335920e+00 5.60845077e-01
9.21197891e-01 -6.95825815e-01 1.11646283e+00 -2.53310174e-01
-1.65778622e-01 -4.69528645e-01 3.68095934e-01 -1.05960143e+00
-4.02719826e-01 -1.26688015e+00 -4.19561625e-01 1.92412660e-01
1.22387195e-02 -9.34133887e-01 7.98892140e-01 1.30460098e-01
5.31072132e-02 -8.05730462e-01 -1.10759354e+00 -5.73725224e-01
-5.12071609e-01 -1.70289457e-01 -1.80283144e-01 6.08318627e-01
1.33943155e-01 1.26352813e-02 -7.65708685e-02 3.06751430e-01
7.36774504e-01 6.54505268e-02 1.24502011e-01 -8.12137663e-01
-5.96838236e-01 -6.55853450e-02 -8.29726934e-01 -8.41697872e-01
-7.97032952e-01 -1.16708063e-01 -4.04360481e-02 -1.47094679e+00
-2.05764696e-01 -1.65153574e-02 -3.70140970e-01 6.42806813e-02
-7.08568752e-01 9.61352646e-01 -2.10357830e-01 2.00266570e-01
1.11157149e-01 -1.50498345e-01 1.57100141e+00 1.56228095e-01
-5.47100306e-01 5.85551918e-01 -2.67829448e-01 6.23376846e-01
6.05726898e-01 -6.28124356e-01 -4.45999175e-01 1.27829060e-01
-3.15652452e-02 4.22137022e-01 3.23932320e-01 -1.47784710e+00
-4.87280414e-02 1.35169029e-02 8.41318429e-01 -5.97612381e-01
2.92801023e-01 -4.81517613e-01 2.27853611e-01 1.11461556e+00
5.61446063e-02 2.21250206e-01 3.59984100e-01 3.03343058e-01
-8.28991234e-02 -6.94113597e-02 1.01778913e+00 -2.98478454e-01
-8.02602693e-02 1.35845551e-02 -7.96526790e-01 6.11510053e-02
7.65497148e-01 -6.91609979e-01 -3.11219811e-01 -4.71062660e-01
-1.00948799e+00 -4.07524288e-01 -1.47450462e-01 -3.81580412e-01
9.21142459e-01 -9.94266093e-01 -5.21795273e-01 3.18360448e-01
-3.77562851e-01 -1.38402447e-01 4.74435836e-01 1.67257011e+00
-1.08442116e+00 2.97536820e-01 -1.12732373e-01 -9.08054531e-01
-1.57552814e+00 1.70551926e-01 6.28048599e-01 -3.31330985e-01
-9.50466752e-01 8.07968616e-01 -4.59771216e-01 8.77496779e-01
3.88106215e-03 -8.85571182e-01 -4.53561008e-01 -2.14553759e-01
8.38092089e-01 5.68640351e-01 3.03827435e-01 -3.69590014e-01
-4.96620566e-01 1.16033483e+00 6.14133477e-01 -2.61367857e-02
9.35854554e-01 -1.91067055e-01 2.37705648e-01 3.29995006e-01
9.04908001e-01 2.63705462e-01 -1.00126493e+00 2.18485713e-01
-5.60578525e-01 -5.35564303e-01 -1.84937447e-01 -5.35574973e-01
-7.62135148e-01 1.19336009e+00 9.84389901e-01 3.60181719e-01
1.58554101e+00 -4.25214678e-01 8.80076706e-01 2.44198784e-01
1.21642098e-01 -6.80445135e-01 5.02574980e-01 -1.79290608e-01
3.20063919e-01 -6.84054911e-01 9.10186619e-02 -5.89424133e-01
-2.83501297e-01 1.61676419e+00 1.28964975e-01 2.43934076e-02
5.87691128e-01 4.49149877e-01 5.65844476e-01 3.58083919e-02
-3.06884199e-01 4.81120497e-02 3.92727196e-01 9.20231938e-01
6.07714295e-01 -2.22091764e-01 -4.89994556e-01 2.98610449e-01
-9.90695134e-02 2.61733860e-01 8.52718651e-01 1.23057473e+00
-6.64030075e-01 -4.83248353e-01 -7.22828329e-01 4.81975049e-01
-1.11347067e+00 -7.00354762e-03 4.11530346e-01 3.87188077e-01
-3.99846882e-02 9.03210223e-01 -5.36129205e-03 -8.31976086e-02
3.98938417e-01 2.49195606e-01 1.12079227e+00 -4.81148899e-01
-3.73735100e-01 4.86629814e-01 -5.04850037e-02 -5.76025903e-01
-7.74010420e-01 -3.34420413e-01 -1.45901728e+00 2.24981710e-01
-2.67640680e-01 1.10428452e-01 6.36035979e-01 7.76148796e-01
1.21885158e-01 8.40713561e-01 4.85077232e-01 -8.29134583e-01
-5.80004692e-01 -1.03210306e+00 -5.28512597e-01 3.95402759e-01
9.72311318e-01 -3.94495785e-01 -7.98267484e-01 3.03188086e-01] | [13.944354057312012, 2.959150791168213] |
8ceb1e13-5e05-4a0f-92f9-15deebe1f421 | fast-distributed-inference-serving-for-large | 2305.05920 | null | https://arxiv.org/abs/2305.05920v1 | https://arxiv.org/pdf/2305.05920v1.pdf | Fast Distributed Inference Serving for Large Language Models | Large language models (LLMs) power a new generation of interactive AI applications exemplified by ChatGPT. The interactive nature of these applications demand low job completion time (JCT) for model inference. Existing LLM serving systems use run-to-completion processing for inference jobs, which suffers from head-of-line blocking and long JCT. We present FastServe, a distributed inference serving system for LLMs. FastServe exploits the autoregressive pattern of LLM inference to enable preemption at the granularity of each output token. FastServe uses preemptive scheduling to minimize JCT with a novel skip-join Multi-Level Feedback Queue scheduler. Based on the new semi information-agnostic setting of LLM inference, the scheduler leverages the input length information to assign an appropriate initial queue for each arrival job to join. The higher priority queues than the joined queue are skipped to reduce demotions. We design an efficient GPU memory management mechanism that proactively offloads and uploads intermediate states between GPU memory and host memory for LLM inference. We build a system prototype of FastServe based on NVIDIA FasterTransformer. Experimental results show that compared to the state-of-the-art solution Orca, FastServe improves the average and tail JCT by up to 5.1$\times$ and 6.4$\times$, respectively. | ['Xin Jin', 'Xuanzhe Liu', 'Gang Huang', 'Zili Zhang', 'Yinmin Zhong', 'Bingyang Wu'] | 2023-05-10 | null | null | null | null | ['blocking'] | ['natural-language-processing'] | [-1.29832551e-01 -1.57269880e-01 -4.98391896e-01 -4.69092160e-01
-6.55591905e-01 -2.99667269e-01 4.66789901e-01 5.31052835e-02
-5.27862012e-01 4.33083385e-01 -1.01625808e-01 -7.46052980e-01
9.45934728e-02 -8.56237471e-01 -5.67958713e-01 -4.50650871e-01
-1.55786827e-01 1.02072060e+00 5.65960050e-01 -3.69569622e-02
3.74435306e-01 4.56058323e-01 -1.83901072e+00 7.96265364e-01
8.84441078e-01 9.53909159e-01 4.39939737e-01 1.18157661e+00
-5.18192172e-01 1.13072944e+00 -6.24565661e-01 3.12146574e-01
1.10412419e-01 -2.43594497e-02 -7.39194393e-01 -2.91175693e-01
4.22497064e-01 -6.71294689e-01 -2.19964057e-01 3.01835775e-01
5.87681532e-01 3.74069214e-01 2.45757490e-01 -1.60247636e+00
1.66737959e-01 5.34897089e-01 -7.57475555e-01 7.09491789e-01
8.22075829e-02 5.25345027e-01 5.85225344e-01 -7.49358475e-01
4.97641206e-01 1.37199283e+00 3.64686072e-01 1.27275318e-01
-1.25773752e+00 -7.36657381e-01 1.63151488e-01 3.56544778e-02
-1.35170400e+00 -8.05403233e-01 -6.92854226e-02 -3.07483673e-01
1.64142013e+00 7.28843093e-01 2.54481375e-01 4.55037266e-01
8.85860324e-01 7.98897982e-01 1.06831968e+00 -5.34882247e-01
5.03010333e-01 -2.36584425e-01 5.63565493e-01 9.00686145e-01
-3.57018530e-01 -1.32197112e-01 -9.93318498e-01 -5.57885468e-01
8.56379271e-01 -2.60532647e-01 3.95346344e-01 5.34263074e-01
-1.02059305e+00 6.44164383e-01 -1.87640395e-02 -3.14686596e-01
-4.62575912e-01 3.77306104e-01 7.92282104e-01 1.16710663e-01
5.93958259e-01 2.06636742e-01 -3.24237764e-01 -6.81817830e-01
-1.32290256e+00 2.91806549e-01 1.07770467e+00 1.08991098e+00
6.92046821e-01 3.59768234e-02 -8.83713722e-01 6.47425294e-01
1.24877729e-01 7.11771071e-01 5.03260791e-02 -1.43204522e+00
1.44821629e-01 2.88300574e-01 -1.48691818e-01 -6.03743494e-01
-6.34019077e-01 -3.74467283e-01 -8.69669676e-01 8.63701552e-02
6.85574710e-02 -1.81412119e-02 -7.21968472e-01 1.34723091e+00
5.50117731e-01 4.51880485e-01 -4.62316513e-01 1.08085513e+00
3.42779696e-01 1.11185992e+00 2.82097638e-01 -1.39551610e-01
1.44296539e+00 -1.18508744e+00 -4.68203962e-01 -3.60796958e-01
6.93212807e-01 -1.04482400e+00 1.24432778e+00 4.52727407e-01
-1.18677521e+00 -6.26298130e-01 -5.79590380e-01 -2.79848278e-01
1.58661842e-01 -3.91888134e-02 1.09688330e+00 2.50761658e-01
-1.07573283e+00 3.61694515e-01 -1.54983103e+00 7.66743794e-02
-5.15930764e-02 4.23375785e-01 3.50232214e-01 1.68435320e-01
-7.82033801e-01 3.12111706e-01 2.01756954e-01 3.37995961e-02
-8.31193805e-01 -1.11718273e+00 -3.52610618e-01 7.57463276e-02
5.46773493e-01 -9.39673483e-01 1.57982719e+00 -5.25337636e-01
-1.68942511e+00 6.02739751e-01 -6.16373241e-01 -4.25954938e-01
2.64186084e-01 -2.27233112e-01 -1.33462980e-01 -5.64822406e-02
2.68201530e-01 4.49532807e-01 8.42840552e-01 -8.29476774e-01
-7.63428867e-01 -8.24621692e-02 -1.00127839e-01 2.62440771e-01
-1.02278277e-01 1.55864254e-01 -8.51884246e-01 -7.40810633e-02
-3.04815084e-01 -1.00171483e+00 -4.28816140e-01 -3.60641062e-01
-5.82713246e-01 -1.80797175e-01 6.86199069e-01 -1.42182380e-01
1.66954577e+00 -1.93844700e+00 -1.61367580e-01 2.52505332e-01
4.44855720e-01 1.27299815e-01 4.40872423e-02 4.28913116e-01
7.02270091e-01 -2.63248444e-01 2.77062386e-01 -6.41023576e-01
2.20985517e-01 6.43087268e-01 -5.95213771e-01 7.80941993e-02
-2.86468387e-01 5.95383584e-01 -8.37298036e-01 -7.93826103e-01
2.86647797e-01 4.29437384e-02 -9.43941653e-01 5.48598528e-01
-4.65151399e-01 1.88913956e-01 -5.56518972e-01 3.83148611e-01
7.31484711e-01 -4.85562712e-01 1.38273224e-01 3.19661885e-01
-6.95743442e-01 6.31673276e-01 -1.08101618e+00 1.51532769e+00
-8.72645140e-01 5.10057390e-01 5.66365898e-01 -3.68763208e-01
5.80615759e-01 -2.87760515e-02 1.84742123e-01 -7.54561484e-01
-1.59834266e-01 8.87122080e-02 -9.40794051e-02 -1.76242337e-01
1.09970987e+00 3.69016051e-01 -3.33430201e-01 6.14916861e-01
-5.05693793e-01 -1.22136727e-01 3.06399524e-01 5.28361857e-01
1.26731145e+00 5.76250851e-02 -1.36223838e-01 -7.63425708e-01
2.64542371e-01 1.98011801e-01 7.11177170e-01 1.35890734e+00
7.70123675e-02 -1.29389361e-01 5.82372725e-01 -4.84059662e-01
-9.05311763e-01 -1.09083748e+00 -7.89325014e-02 2.11726427e+00
-2.49024078e-01 -8.53220105e-01 -9.03352678e-01 3.33869196e-02
1.65033400e-01 9.80318606e-01 -4.41368250e-03 -3.89826968e-02
-6.99616432e-01 -6.08694315e-01 4.47549194e-01 4.50006276e-01
5.50451875e-01 -1.19796169e+00 -1.18527746e+00 5.01648724e-01
1.75353393e-01 -1.12891722e+00 -7.04233825e-01 9.57064107e-02
-8.17791760e-01 -2.82465309e-01 1.84304655e-01 -2.76204437e-01
3.94276500e-01 1.24788545e-01 1.60867262e+00 1.70325816e-01
-4.24851328e-01 8.38763565e-02 2.35721394e-01 -2.34063208e-01
-7.16540396e-01 6.11097693e-01 8.28352198e-02 -2.61496753e-01
1.88681081e-01 -5.13581276e-01 -5.80541551e-01 2.04479024e-01
-6.94854856e-01 7.99361050e-01 3.21101248e-01 6.62443995e-01
7.66844809e-01 -2.69149840e-01 2.56194830e-01 -1.10813630e+00
6.07641101e-01 -5.68308651e-01 -9.00548220e-01 -1.48421645e-01
-6.81814253e-01 1.53781936e-01 1.09224844e+00 -2.66968757e-01
-1.14851546e+00 -3.86955768e-01 -1.49548739e-01 -5.08531511e-01
-7.21731186e-02 3.87762725e-01 3.04578662e-01 4.37394381e-01
3.40341032e-01 1.56284019e-01 -3.80031355e-02 -2.12327123e-01
1.98219433e-01 8.07533443e-01 6.79093719e-01 -1.09517479e+00
-9.22683552e-02 3.05306166e-01 1.17529862e-01 -9.41913366e-01
-5.03400981e-01 -4.29217100e-01 -8.58917013e-02 -3.08822751e-01
5.63264251e-01 -7.64313698e-01 -1.62826228e+00 5.13753593e-01
-1.16408110e+00 -1.04702425e+00 1.62020311e-01 7.78935254e-02
-6.22247219e-01 -1.05987936e-01 -1.27928174e+00 -1.02692938e+00
-8.75264943e-01 -1.39248919e+00 1.51356232e+00 2.07704753e-01
-4.94567007e-01 -7.33268559e-01 -3.14649343e-01 2.79580712e-01
8.33866775e-01 -2.31893539e-01 8.73651147e-01 -2.85785407e-01
-8.99581134e-01 2.08733723e-01 -2.81297415e-01 -3.17044616e-01
-5.58829129e-01 1.84766382e-01 -8.42253447e-01 -3.48598540e-01
-2.78445423e-01 -4.99937870e-02 6.12993002e-01 4.48595583e-01
1.36007452e+00 -1.72244564e-01 -6.37287378e-01 6.93665624e-01
1.21431994e+00 -3.78635824e-02 3.44136089e-01 1.45700663e-01
9.04482961e-01 1.24875205e-02 7.95966148e-01 9.73833501e-01
4.83860373e-01 8.24319422e-01 1.07152380e-01 -8.04909021e-02
1.58018798e-01 1.44012928e-01 5.85773230e-01 1.29944468e+00
2.08716139e-01 -3.04032236e-01 -1.18797243e+00 1.51893035e-01
-2.10429478e+00 -5.81699669e-01 -4.31677848e-01 2.44289565e+00
1.02447557e+00 6.76624835e-01 -1.11822203e-01 -3.56025070e-01
1.42472416e-01 4.25880887e-02 -6.17815018e-01 -1.26991713e+00
4.70014662e-01 4.48833585e-01 6.66176617e-01 7.56571233e-01
-6.40695214e-01 1.30966973e+00 5.39320374e+00 1.39039159e+00
-1.16932261e+00 2.42915124e-01 6.29723310e-01 -4.41511929e-01
-3.92027795e-02 5.29592037e-02 -1.22819209e+00 7.27285624e-01
1.69887745e+00 -1.72933519e-01 6.93479538e-01 8.77900481e-01
6.75717413e-01 -7.21043468e-01 -9.77618933e-01 7.81557083e-01
-4.57506686e-01 -1.39626217e+00 -2.51007199e-01 1.10859677e-01
3.26066285e-01 3.31406325e-01 -1.00103520e-01 6.16314709e-01
3.89249027e-01 -7.04262197e-01 5.13953269e-01 8.55923533e-01
5.98973930e-01 -1.05509949e+00 4.06416893e-01 7.17843235e-01
-1.26213956e+00 3.70489299e-01 -2.05109328e-01 -7.85569727e-01
3.18660110e-01 8.81880105e-01 -9.35408950e-01 2.11889878e-01
5.31169295e-01 1.94290020e-02 -2.35274568e-01 3.50148886e-01
2.15826303e-01 9.80551600e-01 -6.21903777e-01 3.62515114e-02
1.67484209e-01 -6.81140199e-02 5.60691953e-01 1.55735099e+00
-2.91687578e-01 2.44425267e-01 1.01044583e+00 8.93146157e-01
1.47452965e-01 -2.38759205e-01 -3.84954698e-02 2.02229053e-01
9.10158515e-01 1.42831945e+00 -8.53001118e-01 -9.43185270e-01
-1.98025241e-01 1.10752475e+00 3.12731117e-01 1.05486751e-01
-1.05557144e+00 -2.11967409e-01 1.01545846e+00 2.04751581e-01
-4.70008105e-02 -6.03171766e-01 -4.16413516e-01 -7.69940615e-01
-3.33186328e-01 -7.85179794e-01 3.31040442e-01 -6.05471551e-01
-1.10100818e+00 5.23090661e-01 2.43630297e-02 -2.59484261e-01
-6.67581201e-01 -2.70453840e-01 -6.39929712e-01 1.09027994e+00
-1.07565475e+00 -8.04335773e-01 -2.42130324e-01 1.98905915e-01
8.80974054e-01 8.00870508e-02 1.06168687e+00 2.72447884e-01
-8.68012547e-01 5.87344289e-01 1.80704251e-01 -3.74492645e-01
5.98719358e-01 -1.07517159e+00 8.43501151e-01 6.74137354e-01
-4.95751619e-01 7.69340694e-01 7.45389760e-01 -8.15721571e-01
-2.06795216e+00 -8.74763906e-01 7.26800501e-01 -3.17843348e-01
4.81517047e-01 -7.08433390e-01 -1.02330577e+00 6.46621943e-01
2.57221967e-01 1.34237140e-01 4.76964861e-01 3.59947264e-01
7.85543472e-02 -1.06979653e-01 -5.89542747e-01 7.59073257e-01
8.06491971e-01 -4.54086423e-01 1.01431914e-01 4.70886767e-01
6.49106503e-01 -1.08100188e+00 -1.00692129e+00 -7.82340467e-02
4.80513245e-01 -1.01183379e+00 6.87091410e-01 -4.76147562e-01
3.60885203e-01 -6.41406178e-02 2.67626182e-03 -8.02394867e-01
-3.96937639e-01 -8.93579543e-01 -3.00652981e-01 1.10980237e+00
-6.18027262e-02 -5.09846985e-01 7.92361498e-01 9.45288599e-01
-5.29752254e-01 -1.02231848e+00 -8.00923824e-01 -5.88655949e-01
-3.57270360e-01 -8.44267964e-01 3.63261908e-01 3.98210853e-01
-2.33805344e-01 7.57835627e-01 -1.92296773e-01 3.17489892e-01
7.96875656e-01 5.59114993e-01 1.09104395e+00 -6.43428445e-01
-7.50540793e-01 -2.56268591e-01 2.48628125e-01 -1.27986360e+00
3.83600354e-01 -1.00440526e+00 1.26770318e-01 -1.15491176e+00
1.64188191e-01 -8.10916960e-01 7.08316788e-02 5.39833486e-01
-2.16609016e-01 -5.47498874e-02 4.22882855e-01 3.49065244e-01
-1.06016016e+00 1.23908006e-01 9.48886275e-01 3.68761778e-01
-4.20846641e-01 -5.29152751e-02 1.72080055e-01 7.53381014e-01
6.39175892e-01 -3.98146898e-01 -2.81957299e-01 -6.35935783e-01
1.52046010e-01 4.17452186e-01 2.43509024e-01 -9.75749314e-01
5.73050141e-01 -3.59436154e-01 3.11956611e-02 -5.98221600e-01
4.41447318e-01 -2.37417072e-01 1.96557835e-01 4.12078708e-01
-2.90654778e-01 4.08449233e-01 6.73157275e-01 1.07811600e-01
3.78747016e-01 1.88972399e-01 4.84334409e-01 -7.87010044e-02
-9.65251744e-01 2.82316118e-01 -1.00102925e+00 4.13974971e-01
6.05718970e-01 2.64684349e-01 -3.39950085e-01 -1.55005470e-01
-3.38712960e-01 6.40731215e-01 3.21103930e-01 3.32341045e-01
3.33601713e-01 -8.93108308e-01 -5.69582939e-01 3.08138192e-01
-3.46482247e-01 1.99193701e-01 5.34165502e-01 9.41957057e-01
-6.13438308e-01 4.54828888e-01 -6.93249777e-02 -9.11599994e-01
-1.22045636e+00 4.58799064e-01 -6.15268238e-02 -5.59823871e-01
-5.33754289e-01 8.10914040e-01 1.64679304e-01 -1.77704602e-01
-1.86354607e-01 -3.14923584e-01 6.69281244e-01 -4.33872998e-01
6.69339895e-01 7.91823983e-01 3.10302168e-01 -1.97642386e-01
-4.83973145e-01 -2.59062231e-01 -3.33153337e-01 -1.77559093e-01
9.64402139e-01 -1.16528079e-01 -6.19155049e-01 7.04977155e-01
9.48938847e-01 7.54768997e-02 -1.13241839e+00 -1.25666693e-01
-7.12208897e-02 -4.37311262e-01 4.99040455e-01 -6.21455371e-01
-7.40400493e-01 5.87280691e-01 1.55582041e-01 4.38076407e-02
1.03412032e+00 -2.83583671e-01 1.28590155e+00 3.41831893e-02
6.14223778e-01 -1.18991303e+00 -3.90403092e-01 1.01127803e+00
3.04103315e-01 -8.37147832e-01 -3.04913940e-03 -5.06280661e-01
-3.29047531e-01 9.34997976e-01 9.28370535e-01 1.45959884e-01
3.31606746e-01 1.14537370e+00 -1.03788644e-01 -1.36110514e-01
-1.41515088e+00 2.66582668e-01 -7.37243472e-03 1.64962932e-02
5.79923153e-01 4.35472280e-01 -3.03153545e-01 5.23825884e-01
-2.15724617e-01 3.17928821e-01 3.59539390e-01 1.21626961e+00
-4.13249016e-01 -1.07707655e+00 -5.59861958e-01 7.67865777e-01
-2.74314374e-01 -3.55497152e-01 4.45728332e-01 4.43383068e-01
-1.60529569e-01 8.63973320e-01 9.36674297e-01 -1.01749927e-01
1.00522682e-01 -1.05066918e-01 2.81809330e-01 -6.76758647e-01
-9.86204326e-01 1.87964275e-01 2.62672782e-01 -8.44843268e-01
4.54400480e-01 -3.11132729e-01 -1.90551436e+00 -1.07094657e+00
1.46929519e-02 9.13293436e-02 4.92427826e-01 7.19989121e-01
1.18944263e+00 9.53155398e-01 3.29583019e-01 -1.06563187e+00
-4.24719155e-01 -6.87418938e-01 -4.76274252e-01 -1.14010207e-01
-8.96096788e-03 -4.99533147e-01 1.41251668e-01 -2.15113670e-01] | [8.62181282043457, 3.470515727996826] |
0aeb8c8b-1ee2-403c-b587-b68a76e9ce9f | deep-reparametrization-of-multi-frame-super | 2108.08286 | null | https://arxiv.org/abs/2108.08286v1 | https://arxiv.org/pdf/2108.08286v1.pdf | Deep Reparametrization of Multi-Frame Super-Resolution and Denoising | We propose a deep reparametrization of the maximum a posteriori formulation commonly employed in multi-frame image restoration tasks. Our approach is derived by introducing a learned error metric and a latent representation of the target image, which transforms the MAP objective to a deep feature space. The deep reparametrization allows us to directly model the image formation process in the latent space, and to integrate learned image priors into the prediction. Our approach thereby leverages the advantages of deep learning, while also benefiting from the principled multi-frame fusion provided by the classical MAP formulation. We validate our approach through comprehensive experiments on burst denoising and burst super-resolution datasets. Our approach sets a new state-of-the-art for both tasks, demonstrating the generality and effectiveness of the proposed formulation. | ['Radu Timofte', 'Luc van Gool', 'Fisher Yu', 'Martin Danelljan', 'Goutam Bhat'] | 2021-08-18 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Bhat_Deep_Reparametrization_of_Multi-Frame_Super-Resolution_and_Denoising_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Bhat_Deep_Reparametrization_of_Multi-Frame_Super-Resolution_and_Denoising_ICCV_2021_paper.pdf | iccv-2021-1 | ['multi-frame-super-resolution', 'burst-image-super-resolution'] | ['computer-vision', 'computer-vision'] | [ 3.56113076e-01 -1.09169737e-01 1.85067430e-02 -2.88429409e-01
-1.07881308e+00 -2.00061537e-02 8.63018095e-01 -2.96712101e-01
-3.25006366e-01 5.50506473e-01 4.94956166e-01 3.51009935e-01
-4.08052266e-01 -6.88704312e-01 -9.14023995e-01 -1.22503912e+00
1.03418216e-01 1.23898476e-01 -4.90765199e-02 -1.01745032e-01
1.39302239e-01 5.05131483e-01 -1.40484214e+00 2.78086215e-01
6.52333379e-01 1.01154947e+00 4.88296419e-01 4.65332717e-01
4.53028321e-01 9.06239450e-01 -8.31070393e-02 -1.65930256e-01
4.43438828e-01 -3.79359543e-01 -6.59625292e-01 6.24571383e-01
5.12381375e-01 -5.61269701e-01 -6.13159835e-01 9.47221100e-01
2.47335091e-01 2.64245003e-01 4.50895399e-01 -7.48097658e-01
-6.47956491e-01 -7.12529663e-03 -6.66800320e-01 3.55597764e-01
2.92150676e-01 -8.01501945e-02 1.17271352e+00 -9.71688509e-01
6.53684497e-01 1.23760617e+00 7.75672793e-01 1.24690168e-01
-1.66736782e+00 -1.78276747e-01 -1.84478268e-01 2.74457157e-01
-1.08189476e+00 -8.56526792e-01 8.63686204e-01 -5.54776549e-01
6.63213789e-01 -1.94140270e-01 5.46903670e-01 1.09401810e+00
2.59390384e-01 8.32474887e-01 1.31565678e+00 -5.86851656e-01
1.55719295e-01 -4.91601378e-01 1.63327605e-02 6.53391421e-01
-1.36032164e-01 3.77836645e-01 -8.06324005e-01 -4.12172340e-02
1.16651082e+00 9.10067782e-02 -4.96602386e-01 -6.05985045e-01
-1.28548300e+00 8.98719430e-01 1.81212097e-01 2.51983494e-01
-6.99126422e-01 3.32198441e-01 1.44883826e-01 1.48199096e-01
8.48984241e-01 -5.25666066e-02 -1.22595064e-01 -3.32271419e-02
-1.29536331e+00 2.95173705e-01 3.82616907e-01 5.67651808e-01
8.64788532e-01 2.84244437e-02 -1.83622688e-01 7.30749190e-01
5.25709808e-01 4.67135727e-01 2.88487047e-01 -1.67275810e+00
-1.07950410e-02 -4.49472629e-02 3.05074096e-01 -9.07648981e-01
-1.67284504e-01 -4.87736255e-01 -8.33071172e-01 2.82411218e-01
2.94111490e-01 3.34400147e-01 -7.03672707e-01 1.84624088e+00
1.16892569e-01 7.59164214e-01 -6.98561668e-02 8.50772679e-01
8.08458477e-02 7.73509562e-01 -2.88066655e-01 -6.56231582e-01
1.02817237e+00 -7.76546836e-01 -8.28904629e-01 5.09598712e-03
8.03369284e-02 -7.33226657e-01 7.15015352e-01 5.79768419e-01
-1.36438036e+00 -6.63170159e-01 -1.03996038e+00 -2.69763231e-01
4.52063024e-01 -4.69372571e-02 3.72489363e-01 2.97192603e-01
-1.34275019e+00 9.61302757e-01 -1.26938188e+00 -2.08414569e-01
5.00175476e-01 1.24073774e-01 -5.54200292e-01 -2.96060383e-01
-7.58856714e-01 8.82359624e-01 2.07651868e-01 9.47965086e-02
-1.04694855e+00 -7.78161943e-01 -9.50715661e-01 1.88501105e-02
7.08541125e-02 -1.13457441e+00 1.07857752e+00 -8.51834059e-01
-1.79373336e+00 8.46927643e-01 -4.44158912e-01 -6.09055340e-01
5.18394172e-01 -4.12405998e-01 -1.75421499e-02 5.89715421e-01
1.74248576e-01 4.43381965e-01 1.21342349e+00 -1.25929630e+00
-4.97730196e-01 -2.84521133e-01 1.48854434e-01 1.14002638e-01
-2.11272642e-01 -4.73503321e-02 -2.39951745e-01 -9.49109077e-01
3.00795496e-01 -7.59534359e-01 -2.94263482e-01 1.32887065e-01
1.99334044e-02 2.93555140e-01 6.07062340e-01 -9.41516638e-01
8.72512341e-01 -2.10518956e+00 8.80042672e-01 -7.27083907e-02
3.05093288e-01 -2.00622708e-01 -1.94621816e-01 4.19955641e-01
-2.25739464e-01 -3.06294769e-01 -5.06617010e-01 -9.76573229e-01
-8.58323798e-02 5.80683887e-01 -6.47818983e-01 9.76476133e-01
9.26950648e-02 8.75719190e-01 -8.46465886e-01 -2.11988389e-01
5.80250025e-01 8.36611867e-01 -5.96765876e-01 3.73690158e-01
7.00886175e-02 7.87684381e-01 -1.73782215e-01 2.44206578e-01
6.78524733e-01 -3.89279336e-01 -3.89317162e-02 -5.59896708e-01
-1.03211075e-01 1.75566256e-01 -1.07410204e+00 2.34644628e+00
-5.77239692e-01 5.08373141e-01 2.13922352e-01 -1.45343041e+00
5.05902350e-01 3.67097586e-01 9.48814750e-01 -6.31199300e-01
-3.77251916e-02 1.32248446e-01 -5.37256896e-01 -4.57405627e-01
2.53142506e-01 -5.73637784e-01 3.62994164e-01 4.75273460e-01
4.76165384e-01 7.40484670e-02 7.94036537e-02 1.29635110e-01
8.21615934e-01 6.54756725e-01 3.23597014e-01 -4.61414039e-01
5.09187520e-01 -5.62684774e-01 6.05015278e-01 7.27882743e-01
-3.62709910e-02 7.94166148e-01 2.64084518e-01 -4.25865412e-01
-1.22740734e+00 -1.33904898e+00 -3.86300534e-01 8.02845240e-01
-1.43986121e-02 -3.61041456e-01 -5.37518263e-01 -2.95464188e-01
-3.22580069e-01 4.65882778e-01 -6.64693594e-01 5.57560213e-02
-7.91490734e-01 -8.92014980e-01 -1.73285585e-02 2.07747385e-01
4.05796409e-01 -5.54662287e-01 -5.14202297e-01 3.12630683e-01
-6.18578911e-01 -1.45991325e+00 -2.56263584e-01 -2.06221759e-01
-1.05413795e+00 -8.40878189e-01 -8.00446689e-01 -6.00342631e-01
3.43394130e-01 5.06286442e-01 1.01519501e+00 -1.34195462e-01
-1.26025110e-01 7.45562017e-01 -2.04938486e-01 3.07456344e-01
-5.60594678e-01 -4.98675257e-01 1.67670608e-01 5.22806346e-01
-2.28232384e-01 -1.14662886e+00 -6.63152814e-01 1.49654999e-01
-1.11140668e+00 3.18095297e-01 4.32075918e-01 1.07515538e+00
7.59556592e-01 6.24416769e-02 4.78032559e-01 -3.38401973e-01
1.33934364e-01 -4.17749226e-01 -7.14628041e-01 3.15689817e-02
-5.28919041e-01 1.71194717e-01 2.24738687e-01 -3.05931885e-02
-1.15262568e+00 8.62279460e-02 -2.17156172e-01 -5.04644215e-01
1.90652398e-04 4.03741509e-01 -2.19645038e-01 -2.72017121e-01
3.10950190e-01 3.85534853e-01 2.59211034e-01 -7.28965163e-01
6.98668420e-01 2.46562377e-01 9.74510729e-01 -7.90308774e-01
8.02846849e-01 1.15802813e+00 1.77603871e-01 -7.87720442e-01
-9.39208627e-01 -5.65306008e-01 -8.43692422e-01 -1.85660213e-01
9.89223719e-01 -1.20180011e+00 -5.16190231e-01 5.53859890e-01
-1.35090292e+00 -8.99923071e-02 -5.94327033e-01 4.02750522e-01
-1.15218234e+00 9.11181748e-01 -8.26525211e-01 -4.88412350e-01
-2.19434910e-02 -1.24380624e+00 1.42371261e+00 -1.29860744e-01
2.86086798e-01 -1.17312300e+00 3.28140378e-01 4.40484494e-01
3.13043952e-01 2.73893923e-01 6.26918375e-01 2.63487315e-03
-8.97689223e-01 1.12275757e-01 -2.37081662e-01 6.11172378e-01
1.62790176e-02 -4.48779494e-01 -1.21463335e+00 -4.68738765e-01
6.89627290e-01 -2.10857749e-01 1.24968028e+00 7.88651764e-01
9.00711656e-01 -8.87518600e-02 -1.05775532e-03 9.93121743e-01
1.61745048e+00 -4.25096512e-01 9.02249336e-01 5.31821907e-01
6.13177598e-01 5.05036294e-01 3.88626009e-01 5.92584133e-01
4.39862341e-01 9.51480508e-01 6.02223635e-01 -3.05614676e-02
-2.08269924e-01 6.39705360e-02 4.60002214e-01 9.11553025e-01
-2.15968773e-01 5.32500185e-02 -6.48789763e-01 6.20715380e-01
-2.06214833e+00 -1.02954960e+00 5.63320853e-02 2.06109738e+00
8.19984555e-01 -2.40624294e-01 -2.13332817e-01 9.75044891e-02
4.57688779e-01 4.49285299e-01 -1.28366515e-01 2.96327949e-01
-3.95033538e-01 4.72181499e-01 1.24100536e-01 8.57101142e-01
-1.30475140e+00 6.48188829e-01 6.38543224e+00 8.16225052e-01
-8.52450788e-01 4.37165111e-01 4.00299430e-01 8.92504081e-02
-1.67708695e-01 9.10752639e-02 -4.23398286e-01 1.99344143e-01
9.29833949e-01 -1.84209213e-01 5.78469515e-01 2.41377205e-01
4.88553464e-01 -8.65414068e-02 -1.32336712e+00 1.20433068e+00
2.26322576e-01 -1.59522903e+00 6.45283982e-02 3.43950450e-01
8.08307171e-01 9.90185738e-02 1.17952764e-01 -1.58885852e-01
8.79853815e-02 -8.48966002e-01 8.75870883e-01 8.03112805e-01
5.19143045e-01 -4.80065882e-01 3.61837834e-01 1.76290959e-01
-8.74289274e-01 -1.68380514e-01 -3.73961538e-01 -1.57296851e-01
6.97953820e-01 8.14440727e-01 -1.52144536e-01 9.75400507e-01
6.24045074e-01 1.34105659e+00 -1.60290852e-01 9.29895163e-01
-2.76590854e-01 4.30357218e-01 -2.52263516e-01 1.22136843e+00
-8.62252340e-02 -5.05610645e-01 8.90635192e-01 1.07614505e+00
3.69425029e-01 -6.45556394e-03 2.72563040e-01 9.47849393e-01
3.66990976e-02 -1.93308741e-01 -3.21655303e-01 2.98359126e-01
-7.52796943e-04 1.22805905e+00 -3.71077538e-01 -2.12689936e-01
-4.77398604e-01 1.09852707e+00 1.69852287e-01 5.61702490e-01
-7.28493094e-01 3.55403543e-01 7.33505070e-01 1.00922018e-01
5.57510436e-01 -5.32056987e-01 -2.38773555e-01 -1.52763987e+00
1.71395466e-01 -7.93555200e-01 7.19914809e-02 -8.36534441e-01
-1.36613977e+00 3.62622470e-01 2.09630817e-01 -1.28515005e+00
-3.59219342e-01 -5.30683875e-01 -3.77838492e-01 8.58696818e-01
-1.93929648e+00 -1.56792057e+00 -2.23054439e-01 6.59737468e-01
4.41524565e-01 1.16706766e-01 7.32900381e-01 4.02957469e-01
-4.53788906e-01 6.78227097e-02 4.14825052e-01 -2.61587054e-01
6.47719204e-01 -1.11116362e+00 2.17577621e-01 1.31076562e+00
2.53454030e-01 6.11412942e-01 7.55959034e-01 -2.50724435e-01
-1.36471510e+00 -8.74505460e-01 4.43468601e-01 -3.40763837e-01
9.60515261e-01 -1.74990073e-01 -8.86555135e-01 8.65645707e-01
2.54422605e-01 2.89842844e-01 6.15346313e-01 -2.02520430e-01
-2.95364052e-01 -5.86996749e-02 -8.66747379e-01 2.26906523e-01
8.19074512e-01 -7.90393293e-01 -6.92982435e-01 4.22783822e-01
6.93780363e-01 -2.99084991e-01 -1.18867552e+00 3.38253915e-01
5.16863346e-01 -1.12133324e+00 1.53392255e+00 -3.06292772e-01
7.71975875e-01 -1.98629975e-01 -6.56975567e-01 -1.18893063e+00
-6.62020147e-01 -6.72259569e-01 -5.00889957e-01 9.95132327e-01
-2.52914816e-01 -4.85783041e-01 4.27986771e-01 3.01030457e-01
-3.22634995e-01 -6.00972414e-01 -1.34059596e+00 -6.19618952e-01
6.11131899e-02 -3.83171886e-01 1.17603086e-01 8.38971198e-01
-4.40484971e-01 5.97151108e-02 -7.93697357e-01 5.40376544e-01
1.30565679e+00 1.01574808e-01 4.77518380e-01 -1.17985940e+00
-8.05510104e-01 -1.67771325e-01 -4.71832186e-01 -1.47092843e+00
3.84827077e-01 -7.96951592e-01 3.92883122e-02 -1.36823618e+00
4.01203781e-01 3.42448428e-02 -4.82910067e-01 1.96635067e-01
5.19947382e-03 5.32860696e-01 3.63347918e-01 6.33610904e-01
-5.62566996e-01 9.13425088e-01 1.01216722e+00 4.56865653e-02
6.11444972e-02 -2.66316772e-01 -5.13199329e-01 9.31205392e-01
2.12851688e-01 -3.99882585e-01 -2.93677390e-01 -8.11813176e-01
1.61762357e-01 2.77094364e-01 9.51783061e-01 -7.37283349e-01
1.39208063e-01 -8.23624805e-02 2.27209419e-01 -2.35602424e-01
6.98502183e-01 -7.65263200e-01 2.45790705e-02 1.41440943e-01
-1.37309745e-01 -2.97125816e-01 -3.06494553e-02 8.68738353e-01
-2.60880947e-01 -2.56623086e-02 1.18231797e+00 3.41843888e-02
-6.98757052e-01 3.27064842e-01 -9.22731161e-02 -9.15836468e-02
5.93181849e-01 -1.67967811e-01 -2.11803988e-01 -3.59055221e-01
-9.86286938e-01 -1.96527496e-01 7.27353752e-01 1.91247553e-01
6.79320157e-01 -1.39626217e+00 -9.11560655e-01 2.81942606e-01
-9.04595032e-02 -1.46765277e-01 3.42038780e-01 1.36835480e+00
-4.13074136e-01 -1.10965986e-02 -2.74458408e-01 -1.06311440e+00
-7.87483037e-01 3.50642234e-01 5.14466465e-01 -3.24455410e-01
-1.14199948e+00 5.57761908e-01 5.48805654e-01 5.23617454e-02
-1.30149245e-01 -1.90842450e-01 2.55890097e-02 -1.58303216e-01
7.15062141e-01 4.41292614e-01 -1.22921495e-02 -9.57636595e-01
-4.63337265e-02 6.87711596e-01 4.04749624e-02 -4.60376859e-01
1.78022993e+00 -6.15617037e-01 -4.30361897e-01 3.70333731e-01
1.35533810e+00 -2.05457713e-02 -1.70013463e+00 -5.23025870e-01
-3.89445759e-02 -6.85268283e-01 5.96211374e-01 -3.45075727e-01
-9.40119147e-01 7.88626552e-01 6.54894412e-01 -2.24468753e-01
1.35797489e+00 -5.99076301e-02 8.65737140e-01 1.45034209e-01
4.36003208e-01 -6.06888950e-01 2.83431828e-01 3.47532034e-01
1.05052245e+00 -1.13957977e+00 1.97996512e-01 -1.84196904e-01
-6.88929781e-02 1.15642750e+00 -2.17775494e-01 -4.22898054e-01
7.79147387e-01 1.91980258e-01 -1.80438757e-01 -8.65956917e-02
-4.93671060e-01 3.84387188e-02 1.36573300e-01 5.27222276e-01
1.02348462e-01 -3.36760849e-01 -1.04551889e-01 2.99988955e-01
1.69233292e-01 1.44925430e-01 6.78356230e-01 8.19850147e-01
-4.17999446e-01 -1.28201020e+00 -3.52861375e-01 -1.24142461e-01
-4.65396315e-01 -1.22556545e-01 4.20793176e-01 4.17673469e-01
7.94321373e-02 8.98645461e-01 -7.09879547e-02 4.01851386e-02
5.58580793e-02 -1.04431607e-01 8.85261118e-01 -3.32799286e-01
1.23944111e-01 6.08350337e-01 -3.23905140e-01 -8.89210999e-01
-1.00723100e+00 -8.43596637e-01 -9.28906739e-01 -6.26087040e-02
-7.52928555e-02 -2.19995558e-01 5.48901379e-01 1.09218740e+00
2.89024383e-01 4.30243671e-01 6.50452375e-01 -1.46607327e+00
-5.76492488e-01 -8.38888943e-01 -6.23738408e-01 5.57491720e-01
7.59388089e-01 -8.17071140e-01 -3.71014953e-01 5.50836444e-01] | [11.540752410888672, -2.3683090209960938] |
3dbc92bb-28f2-4fe5-af51-7b31431f4d3d | centering-based-neural-coherence-modeling | null | null | https://aclanthology.org/2020.emnlp-main.604 | https://aclanthology.org/2020.emnlp-main.604.pdf | Centering-based Neural Coherence Modeling with Hierarchical Discourse Segments | Previous neural coherence models have focused on identifying semantic relations between adjacent sentences. However, they do not have the means to exploit structural information. In this work, we propose a coherence model which takes discourse structural information into account without relying on human annotations. We approximate a linguistic theory of coherence, Centering theory, which we use to track the changes of focus between discourse segments. Our model first identifies the focus of each sentence, recognized with regards to the context, and constructs the structural relationship for discourse segments by tracking the changes of the focus. The model then incorporates this structural information into a structure-aware transformer. We evaluate our model on two tasks, automated essay scoring and assessing writing quality. Our results demonstrate that our model, built on top of a pretrained language model, achieves state-of-the-art performance on both tasks. We next statistically examine the identified trees of texts assigned to different quality scores. Finally, we investigate what our model learns in terms of theoretical claims. | ['Michael Strube', 'Sungho Jeon'] | null | null | null | null | emnlp-2020-11 | ['automated-essay-scoring'] | ['natural-language-processing'] | [ 9.17691216e-02 5.58241665e-01 -3.09057325e-01 -2.09532842e-01
-7.35768616e-01 -6.82847619e-01 9.74839032e-01 5.08192718e-01
-3.24857384e-01 4.74969029e-01 1.02052295e+00 -4.21578139e-01
3.84565033e-02 -8.25869322e-01 -5.48793852e-01 -1.74240232e-01
2.47417465e-01 3.48537713e-01 3.45995903e-01 -3.52788627e-01
7.09615052e-01 -1.40763864e-01 -1.01313269e+00 7.94564009e-01
1.16532099e+00 5.24200380e-01 1.95004046e-01 5.35556972e-01
-2.55703360e-01 1.48543203e+00 -7.67557919e-01 -6.11360312e-01
-3.29933971e-01 -7.80547023e-01 -1.29809761e+00 -1.15836803e-02
5.87738752e-01 -2.45300844e-01 -4.22265917e-01 1.00482929e+00
-2.24867426e-02 -1.16015948e-01 6.78635120e-01 -5.82759142e-01
-6.67647064e-01 1.29531121e+00 -2.74395466e-01 5.34186065e-01
6.20586872e-01 -2.11112529e-01 1.60003304e+00 -7.15052426e-01
9.62474346e-01 1.15850258e+00 7.92926669e-01 2.38693982e-01
-1.26406598e+00 -1.13839179e-01 2.77999699e-01 5.53015649e-01
-7.04010606e-01 -6.49998546e-01 1.06762493e+00 -7.47413695e-01
9.88264680e-01 1.88054591e-01 7.12363362e-01 9.73763347e-01
1.66317105e-01 8.51345003e-01 1.10307217e+00 -7.49696493e-01
-2.62169703e-03 -2.72349209e-01 8.89627635e-01 9.10405993e-01
-7.19769597e-02 -2.13653147e-01 -8.25934231e-01 9.94139165e-02
1.77253723e-01 -6.13621116e-01 -4.85185713e-01 -1.10107018e-02
-1.24305058e+00 8.74255717e-01 5.65402031e-01 7.71816313e-01
-3.64157617e-01 8.47378150e-02 3.91157299e-01 3.44459832e-01
6.15223587e-01 7.54361093e-01 5.49002998e-02 -2.51481719e-02
-1.20363855e+00 1.64816484e-01 1.00545597e+00 4.91036624e-01
5.78809977e-01 -3.60697091e-01 -6.02297246e-01 5.59735298e-01
2.85484314e-01 -1.04368795e-02 4.91434395e-01 -1.29882884e+00
6.85965121e-01 8.98817301e-01 -1.10598519e-01 -1.38731945e+00
-4.61426020e-01 -3.81420314e-01 -3.93153787e-01 -4.04415548e-01
4.76684868e-01 9.07376409e-02 -1.33055940e-01 1.89054835e+00
-4.97867614e-02 7.97786471e-03 1.17312804e-01 5.16176879e-01
9.58667457e-01 3.87027234e-01 -1.35215193e-01 -4.79977220e-01
1.49865317e+00 -1.16015482e+00 -1.08597648e+00 -1.29978862e-02
7.62477040e-01 -6.82526350e-01 1.14744234e+00 1.47967432e-02
-1.40731072e+00 -3.41469288e-01 -1.12934542e+00 -3.98724258e-01
1.00438423e-01 2.48263896e-01 2.24581808e-01 2.79803723e-01
-1.14496160e+00 8.31030428e-01 -9.06106651e-01 -1.27232075e-01
3.49281400e-01 -2.51799732e-01 3.90579402e-02 4.65087205e-01
-1.23554778e+00 1.27675593e+00 3.81918401e-01 -2.19096631e-01
-3.66258264e-01 -5.63083708e-01 -8.31164181e-01 3.46973121e-01
4.75959003e-01 -9.51502025e-01 1.54584634e+00 -9.38426256e-01
-1.65564609e+00 1.17662525e+00 -5.40592849e-01 -7.08950520e-01
2.68677294e-01 -1.96550831e-01 3.51133309e-02 3.25032234e-01
2.50731468e-01 3.20808023e-01 3.71768773e-01 -9.21488762e-01
-4.14197117e-01 -2.74149299e-01 3.76328915e-01 2.40756676e-01
-5.68165898e-01 2.13338256e-01 -1.35726422e-01 -5.02164006e-01
9.72691327e-02 -6.82429254e-01 3.28691423e-01 -3.82525355e-01
-7.02954948e-01 -6.98056757e-01 3.51767093e-01 -8.89426053e-01
1.99213243e+00 -1.73225546e+00 5.08465290e-01 -1.68808594e-01
5.40511250e-01 -7.24306852e-02 -1.51868230e-02 6.33278310e-01
3.64021063e-01 1.78123519e-01 -3.13192219e-01 -4.32299525e-01
2.17813671e-01 -1.83652341e-01 -4.72842813e-01 2.09872857e-01
3.57831530e-02 1.18637776e+00 -8.41593027e-01 -6.16770327e-01
-3.40060562e-01 -7.64723793e-02 -6.46636248e-01 2.31042847e-01
-5.48676014e-01 2.79395282e-01 -3.08189124e-01 2.02077582e-01
1.01875000e-01 -6.26833916e-01 6.77246451e-01 -3.54177147e-01
-2.79297054e-01 1.29178035e+00 -4.15577441e-01 1.70385110e+00
-1.93311602e-01 1.04065955e+00 -1.72384128e-01 -1.01970875e+00
6.94214940e-01 3.52995515e-01 4.82088104e-02 -7.75619566e-01
1.03149183e-01 1.34557724e-01 4.18938696e-01 -6.49733722e-01
8.12098980e-01 1.04777567e-01 -2.76767969e-01 9.21523869e-01
-6.53398708e-02 6.38170242e-02 4.75516170e-01 5.09076118e-01
1.21638632e+00 -3.03054824e-02 5.76511800e-01 -3.99680793e-01
6.16796255e-01 3.88955064e-02 2.32111022e-01 5.21856904e-01
-4.43155728e-02 2.38295078e-01 9.86873746e-01 -2.04021975e-01
-1.03487408e+00 -8.26134086e-01 3.01078930e-02 1.13903904e+00
-1.79126002e-02 -7.27070510e-01 -1.07230461e+00 -6.35017812e-01
-2.34544158e-01 9.41406608e-01 -8.27435732e-01 6.93030804e-02
-1.08330548e+00 -2.67579138e-01 6.63271368e-01 3.48666251e-01
5.08208752e-01 -1.17250896e+00 -9.29927528e-01 2.24798113e-01
-6.91621125e-01 -9.78973687e-01 -4.50419366e-01 -1.83637768e-01
-8.20349932e-01 -1.32742167e+00 -1.13352478e-01 -7.26420701e-01
3.50143343e-01 1.33246318e-01 1.49080956e+00 3.52456897e-01
4.88112718e-01 2.72988647e-01 -2.01643541e-01 -1.46655306e-01
-8.05018425e-01 5.03758729e-01 -4.74763393e-01 -3.49688053e-01
3.53267789e-01 -5.33203661e-01 -4.20746684e-01 -1.43592432e-01
-5.47830462e-01 4.24050570e-01 1.56071305e-01 8.27884555e-01
1.08484291e-01 -5.35824358e-01 6.16327941e-01 -1.07255387e+00
1.22737575e+00 -3.73123646e-01 -2.68789768e-01 5.62165082e-01
-5.84428489e-01 2.33402982e-01 2.66396374e-01 -3.06044798e-02
-1.08849084e+00 -6.23350143e-01 9.02646855e-02 9.90236625e-02
2.38681793e-01 1.04726887e+00 1.71594709e-01 6.20795906e-01
6.73701644e-01 -7.30793085e-03 -8.57966468e-02 -2.59410530e-01
6.00796223e-01 4.99891192e-01 7.69959211e-01 -7.63287306e-01
3.51716369e-01 3.35037619e-01 -3.00017893e-01 -4.30812478e-01
-1.63060379e+00 -2.48778298e-01 -8.73713970e-01 -1.93536833e-01
8.70356262e-01 -7.02457309e-01 -6.42419994e-01 2.69441277e-01
-1.62985706e+00 -2.38951176e-01 -1.73324242e-01 1.61704868e-01
-5.29839337e-01 6.41641438e-01 -9.34352219e-01 -5.20887196e-01
-5.31138420e-01 -7.55612075e-01 6.76962197e-01 8.08799639e-02
-8.29775274e-01 -1.26603973e+00 4.50083733e-01 6.12275839e-01
3.16065252e-01 1.89172864e-01 1.23330879e+00 -7.10889101e-01
-5.10536611e-01 3.92688274e-01 -8.43653828e-02 -8.08880627e-02
1.32271618e-01 -2.96780802e-02 -8.27593923e-01 5.74450940e-02
2.49859929e-01 -4.46650118e-01 1.14390683e+00 3.47575188e-01
7.77181983e-01 -7.14027047e-01 -1.95196420e-01 1.83205783e-01
8.64303410e-01 -1.60642534e-01 4.66718763e-01 6.28561974e-01
4.93819565e-01 8.45631480e-01 1.57842219e-01 1.05521210e-01
8.70912194e-01 7.10873663e-01 2.98495945e-02 2.59569347e-01
-2.87156671e-01 -3.27144861e-01 3.86345834e-01 1.40240264e+00
-2.22167987e-02 -1.48486167e-01 -1.12350833e+00 6.95935667e-01
-2.00386691e+00 -1.27613497e+00 -3.60278785e-01 1.54160655e+00
1.54515147e+00 3.81332934e-01 -4.22543213e-02 2.01197760e-03
5.54446161e-01 4.75761175e-01 -2.63814121e-01 -3.87553692e-01
-1.26328170e-01 2.17218190e-01 -3.13614547e-01 7.21155345e-01
-1.04225731e+00 1.00357914e+00 6.21693993e+00 3.92377049e-01
-8.98263037e-01 1.64681911e-01 4.63629454e-01 2.06104428e-01
-5.95731080e-01 1.44478947e-01 -4.69036371e-01 4.03499126e-01
9.85503674e-01 -3.39081556e-01 1.77518830e-01 3.06528509e-01
1.66044906e-01 -2.09189907e-01 -1.58751082e+00 2.78802514e-01
2.67910749e-01 -1.85777450e+00 -2.50210203e-02 -1.44732624e-01
8.36874127e-01 -1.99733272e-01 -1.65542796e-01 2.01140121e-01
3.49664450e-01 -1.04289436e+00 9.91414547e-01 9.11363900e-01
3.14438850e-01 -1.88458860e-01 6.70433402e-01 5.28009951e-01
-7.49117136e-01 1.67616948e-01 -8.76841098e-02 -5.19332230e-01
1.26872808e-01 4.04941827e-01 -8.17580819e-01 4.71658915e-01
2.21987352e-01 8.67819905e-01 -6.26345634e-01 4.94902015e-01
-7.34202862e-01 9.66698229e-01 1.83261812e-01 -1.47198707e-01
7.13908672e-02 -1.11837856e-01 5.34115016e-01 1.24000406e+00
1.15546621e-01 4.27808389e-02 3.04831833e-01 1.18544281e+00
-4.96693611e-01 -3.22744213e-02 -3.73200119e-01 -1.02713071e-01
8.92428100e-01 9.18314874e-01 -5.36600053e-01 -4.36357349e-01
-3.56315047e-01 6.56525254e-01 8.83038402e-01 1.26438336e-02
-4.71316397e-01 1.08136669e-01 1.60504967e-01 1.48713663e-02
1.49755523e-01 -3.14434677e-01 -6.95889115e-01 -1.27124488e+00
1.14225067e-01 -8.33097875e-01 2.85794705e-01 -5.36117196e-01
-1.25428104e+00 4.26566392e-01 -9.67876166e-02 -6.46206319e-01
-4.36819017e-01 -2.61991024e-01 -9.37459469e-01 7.93252170e-01
-1.58972168e+00 -1.07001770e+00 -1.45887956e-01 7.79926628e-02
5.82442403e-01 -1.63339734e-01 9.25056934e-01 -2.70869821e-01
-5.71231246e-01 4.03547436e-01 -1.76248237e-01 4.91282225e-01
5.82385898e-01 -1.31823301e+00 4.45456445e-01 8.30661952e-01
3.64139855e-01 9.07302856e-01 6.21167898e-01 -5.48432648e-01
-6.91416264e-01 -6.16750240e-01 1.73159468e+00 -6.84585392e-01
1.10942543e+00 -7.08199441e-02 -1.21677411e+00 6.98131680e-01
8.93274546e-01 -7.26860523e-01 8.42610657e-01 6.90759361e-01
-6.35018945e-01 3.36885512e-01 -4.26852226e-01 4.86457288e-01
1.03125966e+00 -9.20250475e-01 -1.53645182e+00 2.25159287e-01
9.45834398e-01 -5.24665058e-01 -8.59223962e-01 1.47249833e-01
3.14561188e-01 -1.08494687e+00 6.35471821e-01 -7.24402249e-01
1.24854112e+00 -1.92547031e-02 -2.95032617e-02 -1.24016190e+00
-4.66916949e-01 -3.38821024e-01 -4.83087093e-01 1.21212256e+00
5.21283090e-01 -2.19944090e-01 4.55011725e-01 3.19751859e-01
-4.85421926e-01 -7.73246050e-01 -9.40118015e-01 -2.98342317e-01
4.33637470e-01 7.86571670e-03 3.16979796e-01 1.24777281e+00
7.34673619e-01 1.03148818e+00 1.25127465e-01 -2.94301480e-01
1.75959006e-01 5.76696813e-01 4.37213242e-01 -1.37080896e+00
-3.54954183e-01 -1.14413309e+00 1.87210009e-01 -9.57183063e-01
7.07016945e-01 -1.18694174e+00 -7.88306147e-02 -1.74254882e+00
6.57897353e-01 -1.94284692e-02 -5.05239256e-02 2.85885721e-01
-5.52930653e-01 -1.24625675e-01 3.43866199e-01 5.71019650e-01
-9.22705591e-01 6.45181537e-01 1.08094645e+00 -3.38688940e-01
-1.28056943e-01 -3.17187846e-01 -8.67845476e-01 8.83316517e-01
8.21020484e-01 -2.57974952e-01 -2.36858770e-01 -7.12673485e-01
5.06607652e-01 1.14009626e-01 2.80708760e-01 -7.08987594e-01
5.06318390e-01 -8.33560452e-02 -7.23053217e-02 -5.13938487e-01
-9.62887779e-02 -1.49807334e-01 -3.42071682e-01 5.12053490e-01
-1.10658145e+00 1.53231457e-01 -9.25506279e-02 2.90200830e-01
-3.55580389e-01 -5.68961799e-01 4.31480944e-01 -1.44856021e-01
-1.94238335e-01 -5.32063365e-01 -4.02230114e-01 3.61858189e-01
5.61872065e-01 1.37276441e-01 -9.66359913e-01 -3.33740592e-01
-6.53583229e-01 2.01324508e-01 4.57870573e-01 1.47015199e-01
3.71977061e-01 -1.23891056e+00 -9.70472217e-01 -4.82243896e-01
-5.16378284e-02 -2.57328928e-01 -9.11996216e-02 9.79770243e-01
-2.97442079e-01 7.71022737e-01 -6.58169435e-03 -5.00929296e-01
-1.37662542e+00 3.94649625e-01 3.69287282e-01 -9.20729458e-01
-5.90150177e-01 6.03581607e-01 4.66576293e-02 -2.73208588e-01
1.20935693e-01 -6.40976906e-01 -7.63898730e-01 4.64786530e-01
3.29839677e-01 2.68774182e-01 -5.54980151e-03 -5.62045157e-01
-6.69892207e-02 4.56019700e-01 -1.25891581e-01 -2.07773924e-01
1.20255220e+00 -2.43944228e-01 -6.15657330e-01 7.15337038e-01
8.53741527e-01 3.31051797e-01 -8.95605385e-01 -6.75342977e-01
7.79253304e-01 3.24325124e-03 -9.21437144e-02 -8.04307699e-01
-4.37149882e-01 9.48515534e-01 -2.17899367e-01 4.29099768e-01
7.84026623e-01 2.75800258e-01 7.57194161e-01 6.27487898e-01
-6.37695119e-02 -1.12812972e+00 6.18857503e-01 1.16115344e+00
1.03780162e+00 -9.34010804e-01 2.51338398e-03 -3.83376330e-01
-4.07468498e-01 1.20978153e+00 4.40983862e-01 -1.38312191e-01
3.04888636e-01 9.15764794e-02 -1.12471871e-01 -6.18887067e-01
-1.06461394e+00 -1.63518503e-01 6.02094352e-01 1.87471107e-01
1.02299726e+00 1.31000921e-01 -7.94691145e-01 7.04983652e-01
-7.60703981e-01 -2.95852184e-01 7.36307144e-01 6.07535720e-01
-7.96187043e-01 -9.68479037e-01 -1.46492139e-01 2.59841621e-01
-4.19142485e-01 -3.97136003e-01 -9.18011546e-01 4.91251409e-01
-9.10502449e-02 1.02133083e+00 3.60510767e-01 -2.74705499e-01
2.47855440e-01 3.80052626e-01 6.04276180e-01 -8.15221131e-01
-8.75297368e-01 -8.17518979e-02 6.20871127e-01 -2.31272951e-01
-9.46288109e-01 -9.41546917e-01 -1.30226302e+00 -4.36569840e-01
-2.62878060e-01 2.51656532e-01 2.41313517e-01 1.21041071e+00
2.00538620e-01 9.40617561e-01 4.19650376e-01 -2.64844656e-01
-6.67661428e-01 -1.32170260e+00 1.05247058e-01 4.50757653e-01
2.53731996e-01 -3.37990165e-01 2.55908631e-02 2.36869425e-01] | [11.219230651855469, 9.31920337677002] |
e9e686d9-eff3-4dfe-955e-b292600f5b10 | labeling-documents-with-timestamps-learning | null | null | https://aclanthology.org/P12-1011 | https://aclanthology.org/P12-1011.pdf | Labeling Documents with Timestamps: Learning from their Time Expressions | null | ['Nathanael Chambers'] | 2012-07-01 | null | null | null | acl-2012-7 | ['document-dating'] | ['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] | [-7.31355619430542, 3.650650978088379] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.