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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
69243588-f55e-4cfd-8c89-dd0db34a4ab8 | unpaired-deep-image-dehazing-using | 2203.07677 | null | https://arxiv.org/abs/2203.07677v2 | https://arxiv.org/pdf/2203.07677v2.pdf | Unpaired Deep Image Dehazing Using Contrastive Disentanglement Learning | We offer a practical unpaired learning based image dehazing network from an unpaired set of clear and hazy images. This paper provides a new perspective to treat image dehazing as a two-class separated factor disentanglement task, i.e, the task-relevant factor of clear image reconstruction and the task-irrelevant factor of haze-relevant distribution. To achieve the disentanglement of these two-class factors in deep feature space, contrastive learning is introduced into a CycleGAN framework to learn disentangled representations by guiding the generated images to be associated with latent factors. With such formulation, the proposed contrastive disentangled dehazing method (CDD-GAN) employs negative generators to cooperate with the encoder network to update alternately, so as to produce a queue of challenging negative adversaries. Then these negative adversaries are trained end-to-end together with the backbone representation network to enhance the discriminative information and promote factor disentanglement performance by maximizing the adversarial contrastive loss. During the training, we further show that hard negative examples can suppress the task-irrelevant factors and unpaired clear exemples can enhance the task-relevant factors, in order to better facilitate haze removal and help image restoration. Extensive experiments on both synthetic and real-world datasets demonstrate that our method performs favorably against existing unpaired dehazing baselines. | ['Longgang Dai', 'Pengpeng Li', 'Caihua Kong', 'Yufeng Huang', 'Yufeng Li', 'Zhuoran Zheng', 'Zhentao Fan', 'Xiang Chen'] | 2022-03-15 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 5.96846104e-01 2.26393670e-01 1.50598213e-01 -8.24348852e-02
-8.59320283e-01 -4.56338912e-01 7.11042821e-01 -1.00155294e+00
-1.83900431e-01 7.89831698e-01 4.06788200e-01 -6.54368624e-02
-9.50277224e-02 -7.83977032e-01 -7.64296889e-01 -1.44039953e+00
2.05196202e-01 7.36931935e-02 -3.67976308e-01 -3.52752745e-01
-1.00395657e-01 1.93355769e-01 -1.36836600e+00 1.33370206e-01
1.12871063e+00 6.69052184e-01 8.01986977e-02 8.06484401e-01
5.77322900e-01 1.01821697e+00 -8.64837289e-01 -4.30883616e-01
6.92314029e-01 -8.73245895e-01 -1.07827783e-01 1.49121359e-01
6.44539237e-01 -7.07142353e-01 -8.11987340e-01 1.18188119e+00
5.99047720e-01 5.37084825e-02 7.22731054e-01 -1.53092766e+00
-1.32745433e+00 3.22030783e-01 -8.28206062e-01 9.69346538e-02
-2.02541083e-01 6.66180074e-01 7.28171110e-01 -9.29328084e-01
1.63564518e-01 1.30935812e+00 1.62046403e-01 9.67876911e-01
-1.35124505e+00 -1.14576328e+00 5.95875122e-02 1.97727699e-02
-1.19383717e+00 -5.08499563e-01 1.12427986e+00 -2.89590806e-01
2.59230435e-01 4.25373852e-01 5.78163803e-01 1.20764923e+00
1.93768218e-01 7.91751802e-01 1.38939047e+00 -2.78078765e-01
-5.91894239e-02 1.93499982e-01 -4.57501113e-01 7.13451922e-01
3.77036512e-01 6.11464083e-01 -5.67876875e-01 3.38563360e-02
8.83435249e-01 1.54373854e-01 -7.26416171e-01 -3.61688912e-01
-1.23039901e+00 9.29993451e-01 5.89138031e-01 -9.99662355e-02
-4.05352652e-01 2.57697821e-01 -1.67608514e-01 3.68607432e-01
6.47931814e-01 6.22381926e-01 -1.54384300e-01 6.65130079e-01
-6.49166882e-01 5.10905087e-01 3.46315354e-01 7.45266020e-01
9.96128559e-01 5.89936316e-01 -4.53213602e-01 7.13888526e-01
3.77499163e-01 7.84049749e-01 2.31934339e-01 -9.61503565e-01
5.29496312e-01 1.37117952e-01 9.79505759e-03 -9.18652415e-01
4.41049069e-01 -6.33846700e-01 -1.34106290e+00 8.98089409e-01
1.06165223e-01 -2.35139549e-01 -1.41172671e+00 2.07157230e+00
4.90259007e-02 5.71501195e-01 2.19190404e-01 1.14861691e+00
5.25919259e-01 1.06570470e+00 -1.53874576e-01 -1.69888243e-01
1.34698856e+00 -1.24865341e+00 -1.08286035e+00 -5.12425244e-01
-1.03829488e-01 -7.43380070e-01 1.03270006e+00 4.81120616e-01
-1.19369709e+00 -4.79873389e-01 -1.53235126e+00 -2.17443660e-01
-1.06862955e-01 -1.65355310e-01 3.58815819e-01 5.55999756e-01
-9.73800600e-01 1.14686929e-01 -6.23209000e-01 6.37219965e-01
8.28042805e-01 2.80466259e-01 -3.84897947e-01 -3.61285567e-01
-1.47056746e+00 7.58631885e-01 1.68529332e-01 3.49298149e-01
-1.57720864e+00 -9.96770382e-01 -1.05404150e+00 -2.20243577e-02
2.39158928e-01 -1.26718473e+00 6.48643494e-01 -1.17219424e+00
-1.42704773e+00 8.19478154e-01 3.19946289e-01 -1.51148021e-01
5.62641263e-01 -3.01414132e-01 -4.03243423e-01 1.74566612e-01
1.19040653e-01 6.30923271e-01 1.67865825e+00 -1.81096148e+00
-2.88128734e-01 -3.27910483e-01 1.25597805e-01 4.33030069e-01
-3.19418520e-01 -3.47121984e-01 -1.71816975e-01 -1.25126028e+00
-3.10459852e-01 -7.67831206e-01 -2.67396867e-01 2.50195354e-01
-5.68745732e-01 5.00236928e-01 1.00029492e+00 -9.90629435e-01
7.52696216e-01 -2.29909015e+00 6.55121744e-01 -6.47414522e-03
9.45134223e-01 2.65120417e-01 -4.89629269e-01 8.83890390e-02
-4.60628331e-01 1.34997562e-01 -5.42326570e-01 -4.42226231e-01
-7.20840096e-02 2.49975875e-01 -5.87498248e-01 7.80415654e-01
5.73729157e-01 1.01039588e+00 -1.03005481e+00 -6.93429485e-02
2.82083541e-01 9.28810894e-01 -4.95681584e-01 7.32676744e-01
-4.80794087e-02 7.61418343e-01 -1.83098495e-01 2.00811177e-01
1.11541975e+00 1.20078683e-01 -3.47435355e-01 -3.11712772e-01
2.77168483e-01 -1.27782151e-01 -7.82419801e-01 1.43838429e+00
-4.98159498e-01 5.82471490e-01 2.78696150e-01 -8.01616073e-01
5.15793860e-01 2.34645471e-01 -1.03878394e-01 -7.60838687e-01
3.49634443e-04 3.90399471e-02 -4.73218709e-02 -3.75860363e-01
2.76536286e-01 -5.83265126e-01 4.85630110e-02 3.59774292e-01
3.55267495e-01 -4.33450341e-01 -4.89425719e-01 3.16591203e-01
8.07648540e-01 -4.22747247e-03 1.09548286e-01 -2.51740992e-01
5.66871464e-01 -6.66642785e-01 6.20904446e-01 4.02848035e-01
-2.02342376e-01 8.96777272e-01 3.14279824e-01 -2.80948102e-01
-1.16567433e+00 -1.36661232e+00 3.45791668e-01 8.16051066e-01
3.23980093e-01 1.68360509e-02 -7.25623846e-01 -6.91426575e-01
-2.71538615e-01 7.31635451e-01 -1.05389261e+00 -6.86514497e-01
-7.34418094e-01 -7.74215102e-01 4.93076175e-01 5.37669621e-02
7.23886907e-01 -6.11915410e-01 5.40632494e-02 -3.23139429e-01
-1.48748845e-01 -8.04189980e-01 -7.57222593e-01 2.17565000e-01
-2.57122189e-01 -8.82514477e-01 -1.11511517e+00 -8.64873827e-01
8.49070549e-01 7.46266425e-01 9.63525236e-01 1.16685160e-01
-5.77981584e-02 -3.24246995e-02 -3.30301374e-02 -5.10059774e-01
-5.59269667e-01 -3.73709142e-01 -1.93247005e-01 4.21375960e-01
-2.22395509e-01 -1.08798862e+00 -1.09543371e+00 2.23768786e-01
-1.50520742e+00 3.39438558e-01 6.88756943e-01 1.27277637e+00
2.92613655e-01 1.58096209e-01 4.29420263e-01 -8.72973263e-01
4.16581839e-01 -5.27884126e-01 -5.07362783e-01 8.72390494e-02
-4.64166552e-01 3.54933858e-01 7.49677598e-01 -6.52695358e-01
-1.21587074e+00 -3.19799513e-01 4.60301079e-02 -9.25531447e-01
1.38273612e-01 -4.16754140e-03 -8.80206347e-01 -1.22214042e-01
7.50761271e-01 5.27166247e-01 6.57709315e-02 -1.84916090e-02
8.84957314e-01 2.69585162e-01 9.37650502e-01 -2.83812374e-01
1.85685611e+00 5.99299192e-01 -1.29938886e-01 -3.24482560e-01
-8.95952225e-01 2.20633168e-02 -1.39195636e-01 -8.08778554e-02
1.04525459e+00 -1.48552287e+00 -2.06408978e-01 8.94493639e-01
-1.07604349e+00 -3.53945255e-01 -4.07363474e-01 2.76353657e-01
-4.56317246e-01 1.66300654e-01 -4.84513819e-01 -7.48502076e-01
-3.19523901e-01 -1.14000773e+00 1.16990435e+00 2.73117751e-01
4.05498892e-01 -9.15216804e-01 4.65247929e-02 5.55420041e-01
3.86257708e-01 6.48660839e-01 1.02170479e+00 -1.51081324e-01
-8.40737045e-01 9.28628594e-02 -1.18774824e-01 9.43201005e-01
1.94447488e-01 -4.74881619e-01 -1.23497498e+00 -4.84752864e-01
4.34550136e-01 -3.37016106e-01 1.23377085e+00 -2.46602986e-02
1.00722706e+00 -7.02776194e-01 1.00641474e-01 1.10415447e+00
1.36563838e+00 2.39713863e-02 1.14338732e+00 4.76088077e-02
1.16611588e+00 4.95063037e-01 2.03864858e-01 9.74913463e-02
1.61503002e-01 3.12338620e-01 7.67311037e-01 -5.11647403e-01
-4.90219563e-01 -4.45362151e-01 4.62603927e-01 7.72834659e-01
-1.95896160e-03 -5.59569001e-01 -2.93169677e-01 3.88606101e-01
-1.54710209e+00 -8.92189026e-01 2.46399358e-01 1.98885775e+00
8.97611558e-01 -9.23495442e-02 -2.58470535e-01 1.50086850e-01
8.15908134e-01 8.49296451e-01 -5.46109200e-01 2.44980738e-01
-4.77923334e-01 3.91076028e-01 3.51069361e-01 7.20491827e-01
-9.70164418e-01 7.74339557e-01 5.08895302e+00 9.51035142e-01
-1.03561699e+00 2.60070801e-01 8.04264247e-01 -2.97920585e-01
-9.60862815e-01 -6.23652227e-02 -7.50485510e-02 5.55314481e-01
4.88964498e-01 -5.93953952e-02 9.53115463e-01 2.97135562e-01
4.06285375e-02 2.84414530e-01 -8.66615355e-01 1.12011969e+00
3.90180379e-01 -1.17575359e+00 4.00321096e-01 2.05693066e-01
1.05231214e+00 -4.28977162e-01 7.06442356e-01 3.18019867e-01
3.96917373e-01 -1.26271987e+00 8.86081815e-01 4.44555581e-01
1.01446700e+00 -9.43771005e-01 4.30773079e-01 2.06833273e-01
-6.51665449e-01 3.06205545e-02 -1.33784607e-01 2.89309788e-02
1.33599997e-01 6.29652560e-01 -2.56260931e-01 5.98781288e-01
3.85721177e-01 6.36998773e-01 -3.37043405e-01 3.26995522e-01
-7.73745060e-01 5.55286467e-01 2.98366696e-01 6.72041118e-01
1.34608284e-01 -3.95527929e-01 7.59390116e-01 6.28261447e-01
1.63849756e-01 3.41379672e-01 -3.51374805e-01 1.21528363e+00
-3.18175107e-01 -6.26547098e-01 -7.27886140e-01 1.15480877e-01
1.94892228e-01 1.27346766e+00 -2.51871437e-01 -2.81026870e-01
-2.04174772e-01 1.54099333e+00 2.62848675e-01 7.69697905e-01
-1.06140482e+00 -5.43746710e-01 1.07325017e+00 -1.72932863e-01
2.89693803e-01 -4.60220464e-02 -1.55458763e-01 -1.56410086e+00
7.40112066e-02 -1.17571366e+00 -1.85412113e-02 -1.05971575e+00
-1.60381567e+00 6.70619845e-01 -2.52761304e-01 -1.35070622e+00
2.81004719e-02 -3.90655845e-01 -8.40116322e-01 1.42366230e+00
-1.99146378e+00 -1.51866162e+00 -5.22363603e-01 8.26777279e-01
4.72188741e-01 -6.38253167e-02 5.93636870e-01 3.18797708e-01
-6.18800581e-01 7.16447532e-01 -1.88677516e-02 1.43388093e-01
8.37930024e-01 -1.24635780e+00 3.43838304e-01 1.33176494e+00
-2.81927418e-02 5.80971658e-01 7.75142431e-01 -4.42779720e-01
-1.33294952e+00 -1.39782977e+00 2.00560063e-01 -5.74401975e-01
3.92674327e-01 -7.48969972e-01 -8.77817452e-01 5.98987162e-01
7.98731387e-01 3.52467120e-01 6.82823181e-01 -6.62850678e-01
-9.44747090e-01 -2.33550861e-01 -1.03735948e+00 8.17462564e-01
1.02022779e+00 -7.90435135e-01 -6.29588842e-01 2.13592336e-01
1.29054308e+00 -3.18636447e-01 -3.95513237e-01 4.08041388e-01
3.59600395e-01 -9.26863015e-01 1.12649357e+00 -5.02936900e-01
8.60188246e-01 -4.94870543e-01 -4.89531979e-02 -1.84756505e+00
-5.40437460e-01 -1.15289283e+00 -3.16130877e-01 1.11520636e+00
2.58817345e-01 -7.44365215e-01 6.07912838e-01 1.46764070e-01
-3.36632520e-01 -5.68332374e-01 -6.51112318e-01 -5.38548946e-01
3.34373385e-01 -8.00008047e-03 7.93651700e-01 1.19450915e+00
-7.25087941e-01 4.89234835e-01 -1.04652786e+00 6.26021206e-01
9.23466504e-01 -1.99318647e-01 7.75351107e-01 -5.56236029e-01
-5.52908123e-01 -1.43136680e-01 -2.45260254e-01 -9.47508752e-01
3.47492874e-01 -6.13132060e-01 1.76544517e-01 -9.95525956e-01
3.68314564e-01 -4.89721745e-02 -5.08773863e-01 2.82991141e-01
-8.50803614e-01 6.65479004e-01 3.19087654e-01 1.49210334e-01
-1.30177110e-01 1.10244703e+00 1.94116521e+00 -5.36516845e-01
1.66905075e-01 -3.49787563e-01 -1.32520175e+00 2.89867312e-01
4.78794992e-01 -5.40089607e-01 -8.36751938e-01 -6.61932647e-01
2.45184764e-01 -1.89751461e-02 6.91396177e-01 -7.21779108e-01
-1.70076013e-01 -2.14311272e-01 6.22859478e-01 -8.33218545e-03
4.01404530e-01 -7.60723293e-01 1.38425753e-01 3.46755475e-01
-3.79032910e-01 -1.65077657e-01 -1.26078233e-01 8.88168454e-01
-3.14594448e-01 3.42160195e-01 8.82665813e-01 -6.34704083e-02
-2.98581809e-01 6.39124572e-01 -1.75264910e-01 9.67536420e-02
9.63188350e-01 -9.09746438e-02 -6.37588322e-01 -6.00966632e-01
-4.86486942e-01 1.18207216e-01 4.43787158e-01 3.98398191e-01
8.76837492e-01 -1.54903030e+00 -1.06381214e+00 6.21598005e-01
1.52642906e-01 2.96149611e-01 5.80950439e-01 3.92767638e-01
-4.66410428e-01 -2.87345409e-01 -3.18803877e-01 -1.90552086e-01
-8.57322693e-01 8.19664657e-01 3.70809853e-01 -3.23775142e-01
-3.60939205e-01 1.06359005e+00 1.09015095e+00 -3.69957715e-01
-5.05134501e-02 1.42983198e-01 -5.25081381e-02 -2.47616723e-01
8.79852057e-01 2.18615413e-01 -2.28079304e-01 -5.50623715e-01
7.28183016e-02 3.11013997e-01 -4.48111221e-02 -3.43497723e-01
1.37462306e+00 -2.60959238e-01 -3.40629488e-01 -4.35827859e-02
1.44674754e+00 2.81704277e-01 -1.75549161e+00 -1.89313412e-01
-7.78318822e-01 -7.60150790e-01 2.26452649e-01 -8.00497890e-01
-1.45604312e+00 9.05689657e-01 7.56018639e-01 -1.12213159e-03
1.69675398e+00 -2.25040257e-01 8.48078847e-01 -2.71251023e-01
-2.33892258e-02 -2.22514763e-01 4.96473134e-01 5.73612340e-02
1.11100256e+00 -1.03791249e+00 -6.95790276e-02 -5.23376882e-01
-5.65053284e-01 6.15276814e-01 7.59183586e-01 -4.92387891e-01
4.64704067e-01 2.08163291e-01 1.81276381e-01 -2.06137061e-01
-6.37812853e-01 1.16526939e-01 5.68657458e-01 9.18200195e-01
-2.18923479e-01 9.62934345e-02 1.41296357e-01 6.14657164e-01
-2.21497461e-01 -4.36308146e-01 4.73032415e-01 5.88719189e-01
1.51469698e-02 -8.79442036e-01 -4.81780410e-01 1.25918761e-01
-4.14445519e-01 -3.64405096e-01 -2.33547196e-01 7.61433482e-01
4.24788028e-01 7.89553046e-01 -6.65215254e-02 -5.68194866e-01
7.88243581e-03 -3.43286604e-01 6.44535542e-01 -5.74637473e-01
-3.98319423e-01 7.31128082e-02 -2.10754141e-01 -4.27809656e-01
-3.38665783e-01 -4.77939770e-02 -5.96283376e-01 -2.72934943e-01
-4.18298930e-01 1.78304791e-01 1.31137922e-01 8.20568919e-01
2.63052315e-01 8.08732569e-01 1.20685637e+00 -1.12886262e+00
-5.37128866e-01 -7.82828391e-01 -6.98547959e-01 5.36201179e-01
1.08421063e+00 -4.56521869e-01 -8.51137578e-01 2.09361121e-01] | [10.9655122756958, -3.1074671745300293] |
1c176413-2249-4092-bafc-be1bf4f80663 | e-3-pose-energy-efficient-edge-assisted-multi | 2301.09015 | null | https://arxiv.org/abs/2301.09015v1 | https://arxiv.org/pdf/2301.09015v1.pdf | E$^3$Pose: Energy-Efficient Edge-assisted Multi-camera System for Multi-human 3D Pose Estimation | Multi-human 3D pose estimation plays a key role in establishing a seamless connection between the real world and the virtual world. Recent efforts adopted a two-stage framework that first builds 2D pose estimations in multiple camera views from different perspectives and then synthesizes them into 3D poses. However, the focus has largely been on developing new computer vision algorithms on the offline video datasets without much consideration on the energy constraints in real-world systems with flexibly-deployed and battery-powered cameras. In this paper, we propose an energy-efficient edge-assisted multiple-camera system, dubbed E$^3$Pose, for real-time multi-human 3D pose estimation, based on the key idea of adaptive camera selection. Instead of always employing all available cameras to perform 2D pose estimations as in the existing works, E$^3$Pose selects only a subset of cameras depending on their camera view qualities in terms of occlusion and energy states in an adaptive manner, thereby reducing the energy consumption (which translates to extended battery lifetime) and improving the estimation accuracy. To achieve this goal, E$^3$Pose incorporates an attention-based LSTM to predict the occlusion information of each camera view and guide camera selection before cameras are selected to process the images of a scene, and runs a camera selection algorithm based on the Lyapunov optimization framework to make long-term adaptive selection decisions. We build a prototype of E$^3$Pose on a 5-camera testbed, demonstrate its feasibility and evaluate its performance. Our results show that a significant energy saving (up to 31.21%) can be achieved while maintaining a high 3D pose estimation accuracy comparable to state-of-the-art methods. | ['Jie Xu', 'Letian Zhang'] | 2023-01-21 | null | null | null | null | ['3d-pose-estimation'] | ['computer-vision'] | [ 1.62884295e-02 -3.77006412e-01 -9.99380350e-02 1.93773080e-02
-5.44707596e-01 -4.90258723e-01 -6.57982156e-02 -3.08007389e-01
-6.30802214e-01 4.18517500e-01 -2.17688367e-01 1.31788999e-02
-6.68481588e-02 -5.38952053e-01 -8.24051321e-01 -5.64023256e-01
-8.23922157e-02 2.51179457e-01 1.48982808e-01 1.32290512e-01
-1.39273092e-01 5.54947138e-01 -1.44106984e+00 -3.95761997e-01
4.95240271e-01 1.11141455e+00 4.70923811e-01 7.79210508e-01
6.23605311e-01 3.02018940e-01 -3.55755746e-01 -3.69750440e-01
4.51560169e-01 -2.04159468e-01 -2.04379842e-01 5.66535115e-01
4.64560568e-01 -7.39734828e-01 -5.29564142e-01 8.17862093e-01
7.55328774e-01 2.95003414e-01 5.13073653e-02 -1.55066276e+00
2.88912328e-04 -1.00627899e-01 -6.73825085e-01 9.39733163e-02
5.84493518e-01 2.93063223e-01 6.64984822e-01 -7.29075193e-01
5.21739602e-01 8.29321444e-01 5.85185468e-01 6.82939112e-01
-9.06347394e-01 -7.06434727e-01 5.61580539e-01 4.34389293e-01
-1.55657601e+00 -5.99794328e-01 8.57375383e-01 -9.13741514e-02
9.98233318e-01 1.89058051e-01 1.22817302e+00 1.07654905e+00
3.35493892e-01 7.13930190e-01 7.89800465e-01 -3.71484876e-01
1.56903774e-01 2.50851568e-02 -2.42876261e-01 9.66369569e-01
2.11660698e-01 2.42989771e-02 -7.20809996e-01 4.26737219e-02
8.45067620e-01 2.21737236e-01 -5.08093417e-01 -5.30771792e-01
-1.36485231e+00 4.83083338e-01 3.90089869e-01 -1.50358632e-01
-4.61793274e-01 4.99586344e-01 3.04471970e-01 -1.26468122e-01
1.98321432e-01 9.67368186e-02 -3.89654696e-01 -1.26078606e-01
-8.17209482e-01 -6.35655597e-02 4.44228113e-01 1.15747988e+00
4.77432221e-01 7.21953213e-02 7.18800575e-02 5.48838079e-01
3.13229352e-01 7.73459017e-01 1.59985840e-01 -1.23931396e+00
5.27853072e-01 6.89312160e-01 2.46187791e-01 -1.12371528e+00
-4.00075644e-01 -3.90923738e-01 -8.04080486e-01 1.97558805e-01
-8.15084949e-03 -3.95771623e-01 -5.61557353e-01 1.96706390e+00
5.54147780e-01 2.97575772e-01 -2.11032137e-01 1.16957521e+00
3.14653248e-01 5.54760396e-01 -1.69391766e-01 -5.61752737e-01
1.28953779e+00 -9.67613459e-01 -5.36857724e-01 -5.22403538e-01
-7.34202191e-02 -3.09447855e-01 7.83039927e-01 4.96347338e-01
-1.13614964e+00 -5.72847128e-01 -1.32912993e+00 2.42286593e-01
4.93360534e-02 4.35016572e-01 3.52368802e-01 6.32161021e-01
-1.17599010e+00 2.15232432e-01 -1.06917870e+00 -4.75136817e-01
2.14171082e-01 7.01413929e-01 -3.61831248e-01 -1.45640731e-01
-8.85088265e-01 1.00352693e+00 1.73784450e-01 3.08834106e-01
-1.09837186e+00 -1.27476886e-01 -7.47874379e-01 4.93016429e-02
8.36345255e-01 -1.11677504e+00 9.23447967e-01 -7.83351898e-01
-1.64157867e+00 5.84603190e-01 -1.40588045e-01 -2.93504208e-01
6.69264317e-01 -3.93743336e-01 -1.92986593e-01 2.89291769e-01
-1.33743435e-01 7.65210569e-01 8.32143903e-01 -1.22881961e+00
-7.45105088e-01 -5.21123648e-01 3.24547768e-01 8.57755303e-01
-6.62556231e-01 -1.38278738e-01 -1.18255544e+00 -3.00482482e-01
1.04831703e-01 -1.41359699e+00 -4.37118500e-01 2.95639515e-01
-3.50580007e-01 1.98270857e-01 6.95592165e-01 -4.42975879e-01
1.19209409e+00 -1.73658872e+00 5.17350733e-01 -1.12703800e-01
4.65478629e-01 8.33973587e-02 1.57357007e-01 1.08447850e-01
4.55417782e-01 -1.87369376e-01 2.00307369e-01 -7.83090472e-01
-1.41869277e-01 4.31332439e-02 3.06562126e-01 7.34014094e-01
-2.06921026e-01 5.24922073e-01 -6.68733299e-01 -5.47073126e-01
6.56813622e-01 6.44838870e-01 -4.81531352e-01 3.41512501e-01
1.84106007e-01 1.92753702e-01 -5.26593924e-01 7.34960377e-01
5.33789515e-01 -3.40406746e-01 2.92051405e-01 -4.41552043e-01
4.95898314e-02 -5.66590071e-01 -1.38329315e+00 1.78324246e+00
-6.55254006e-01 7.09985971e-01 3.21767032e-01 -6.56735837e-01
4.81857926e-01 3.10916096e-01 7.98701882e-01 -4.66309309e-01
4.47568774e-01 -1.07725814e-01 -4.01118547e-01 -3.72213632e-01
4.93490487e-01 2.40178362e-01 -2.78485000e-01 1.24357179e-01
-6.91300184e-02 1.92829177e-01 7.79521465e-02 8.92188847e-02
1.03541887e+00 4.09345537e-01 4.18964535e-01 1.21973239e-01
4.88433212e-01 -1.74642980e-01 7.49676466e-01 2.93709874e-01
-5.96417010e-01 5.08868814e-01 -1.35385886e-01 -3.42546880e-01
-1.01806939e+00 -1.05518091e+00 3.49151909e-01 7.30860829e-01
7.60107994e-01 -3.42124939e-01 -8.45377505e-01 -5.25987327e-01
-2.82175422e-01 4.74292159e-01 -2.36760616e-01 -3.05100203e-01
-7.92707741e-01 -6.19154632e-01 1.86296254e-01 5.88225722e-01
9.00315285e-01 -5.57466865e-01 -1.48005140e+00 1.85374722e-01
-3.95088315e-01 -1.48224938e+00 -9.13436532e-01 1.28602847e-01
-6.85213387e-01 -1.22037590e+00 -5.45881510e-01 -4.25343603e-01
8.51687729e-01 5.24463654e-01 8.86336505e-01 2.98099760e-02
-2.47138619e-01 7.98597097e-01 -1.75195381e-01 -2.11729035e-01
9.01553556e-02 -7.89885446e-02 4.52837586e-01 -5.46606584e-03
6.59477338e-02 -4.44390893e-01 -1.03493214e+00 5.22987008e-01
-3.61514419e-01 2.83861876e-01 7.26054907e-01 7.53726602e-01
7.07116663e-01 1.63922697e-01 2.90890634e-01 -1.04536213e-01
9.76619273e-02 -1.44239217e-01 -7.13060319e-01 3.92144144e-01
-6.31023407e-01 -2.94933468e-01 7.48294711e-01 -6.05512083e-01
-8.56360435e-01 5.70118129e-01 5.85595295e-02 -9.98312533e-01
2.02831209e-01 -1.02651119e-02 -4.80857462e-01 -2.04748601e-01
1.01959407e-01 3.19511801e-01 -1.75907373e-01 2.46015992e-02
2.29274094e-01 4.75163460e-01 4.87470329e-01 -1.54185086e-01
7.20212996e-01 4.27345932e-01 8.60283226e-02 -6.26209676e-01
-5.30747533e-01 -3.51932287e-01 -7.14372039e-01 -8.97284687e-01
9.30174828e-01 -1.26668227e+00 -1.08688343e+00 4.84990895e-01
-1.00864041e+00 -1.15748994e-01 1.63090363e-01 5.90835214e-01
-5.30440748e-01 3.17968935e-01 -3.17844182e-01 -1.00222290e+00
-5.99611163e-01 -1.49279678e+00 1.19774783e+00 2.98471242e-01
9.20045450e-02 -7.62995720e-01 -2.04853460e-01 5.58283448e-01
1.25036865e-01 4.46953475e-01 3.27511072e-01 1.15794830e-01
-1.01416278e+00 -3.01251888e-01 1.49904087e-01 1.27401203e-01
-1.63762435e-01 -2.23267615e-01 -6.84737563e-01 -7.97660351e-01
1.35176778e-02 -4.60985824e-02 4.24580097e-01 6.08623266e-01
8.81642044e-01 -1.70291215e-01 -7.30776191e-01 8.29704165e-01
1.77660429e+00 4.62880701e-01 1.85057446e-01 3.42382848e-01
7.38210022e-01 6.73363404e-03 6.43717885e-01 6.45088613e-01
6.78149879e-01 1.06307364e+00 7.68355906e-01 3.19774039e-02
8.56326967e-02 -2.73919761e-01 6.50266528e-01 7.48912215e-01
-6.81954548e-02 -7.87104070e-01 -5.01800358e-01 3.23831856e-01
-1.81048393e+00 -6.99884236e-01 3.61137927e-01 2.45960593e+00
2.57680118e-01 2.39801168e-01 1.14417866e-01 -1.97765641e-02
1.01974761e+00 2.14263275e-01 -1.03388929e+00 3.81391384e-02
2.14897558e-01 -4.50980663e-01 8.10723960e-01 1.82537675e-01
-1.18291199e+00 5.78414321e-01 5.76955175e+00 5.75344920e-01
-9.27670062e-01 1.46210417e-01 6.09130144e-01 -8.08291197e-01
2.37671435e-01 -1.83687270e-01 -8.65692317e-01 6.02077127e-01
8.46844256e-01 -1.22877285e-01 6.87967122e-01 9.76202965e-01
3.17488700e-01 -2.58351415e-01 -1.14308250e+00 1.54833829e+00
3.96768987e-01 -1.08102608e+00 -2.74219751e-01 1.83306739e-01
3.54845613e-01 -9.10504162e-02 -1.32571965e-01 7.00485408e-02
-1.68488070e-03 -5.33353448e-01 1.09887898e+00 2.96883374e-01
1.02229178e+00 -8.74078810e-01 5.30253112e-01 5.66721201e-01
-1.64433122e+00 -3.69834602e-01 -2.77190864e-01 2.96792965e-02
6.58278346e-01 3.56707036e-01 -4.05583113e-01 5.62121511e-01
1.00033450e+00 5.34887075e-01 -4.39282179e-01 9.14961457e-01
-7.15079485e-03 4.89567593e-03 -4.91997778e-01 -7.36215711e-02
-6.31145537e-02 -2.15214327e-01 7.83415437e-01 6.85038924e-01
6.14039183e-01 2.82471955e-01 5.20040870e-01 4.23828930e-01
-1.98750913e-01 -3.28820914e-01 -4.67414826e-01 3.67334127e-01
7.87882924e-01 1.35671473e+00 -8.42725217e-01 -2.74519801e-01
-4.05828059e-01 1.34625053e+00 2.11208537e-01 1.05636373e-01
-1.40312850e+00 -9.38535482e-02 5.79291046e-01 -5.08704670e-02
1.97754785e-01 -4.84292448e-01 1.35162354e-01 -1.34804523e+00
3.61617953e-01 -5.66871226e-01 3.44565839e-01 -1.00902426e+00
-6.59560084e-01 8.37652564e-01 6.60434589e-02 -1.56845748e+00
-6.34343773e-02 -3.41096342e-01 -3.62291574e-01 3.53147030e-01
-1.10051417e+00 -1.14006948e+00 -7.43712306e-01 6.05223417e-01
8.52769971e-01 6.01270087e-02 4.81467068e-01 4.07250285e-01
-1.09797430e+00 6.55314863e-01 -1.06091816e-02 -1.52656019e-01
5.77161908e-01 -9.84158933e-01 2.03173071e-01 1.08989227e+00
1.14826821e-02 1.87055558e-01 5.58523476e-01 -4.16710377e-01
-1.98030150e+00 -9.80028093e-01 2.41963521e-01 -4.75838900e-01
6.67458549e-02 -4.80451256e-01 -2.04164367e-02 6.78004026e-01
3.15573245e-01 1.09919198e-01 4.29805607e-01 -4.14953113e-01
3.92687768e-01 -3.75789046e-01 -1.16070235e+00 7.71536589e-01
1.52789521e+00 -1.41204566e-01 -7.80326650e-02 9.37390774e-02
7.32488394e-01 -6.00890338e-01 -8.07876348e-01 3.29440176e-01
6.83522105e-01 -9.40048635e-01 1.12673867e+00 7.31488168e-02
-1.27147928e-01 -4.09791410e-01 -2.71645129e-01 -1.23010731e+00
-2.58215219e-01 -5.90068638e-01 -4.23627019e-01 8.56434524e-01
-1.77995265e-02 -3.91115278e-01 1.01308751e+00 7.15074420e-01
-1.46198675e-01 -1.00855160e+00 -1.23876524e+00 -5.66273689e-01
-6.57752454e-01 -3.54528457e-01 2.10617051e-01 3.54541063e-01
-2.80667573e-01 2.97748923e-01 -7.26429820e-01 4.52616513e-01
7.76109278e-01 6.93351105e-02 8.31078351e-01 -8.26971531e-01
-2.70623565e-01 -1.09193325e-01 -3.77738923e-01 -1.28783476e+00
-9.44110565e-03 -2.75741369e-01 1.41628280e-01 -1.46135437e+00
4.05194491e-01 -1.00884587e-01 -2.13053405e-01 2.01286972e-01
-2.21815690e-01 2.84357101e-01 4.58730608e-01 9.99481902e-02
-1.06448877e+00 6.39464557e-01 1.04937220e+00 -1.00590229e-01
-2.40197897e-01 -6.18439401e-03 -3.83224398e-01 8.01332831e-01
4.34724480e-01 -1.79944232e-01 -6.26200259e-01 -6.94276273e-01
5.79993501e-02 4.85924870e-01 4.73378032e-01 -1.46928465e+00
5.25192022e-01 -2.08162427e-01 7.25518525e-01 -5.76662779e-01
8.86389911e-01 -1.27486479e+00 5.04523814e-01 5.83662033e-01
4.47607003e-02 5.23712158e-01 6.30526990e-02 8.34488332e-01
1.43653959e-01 2.31401607e-01 7.47924030e-01 -2.76754856e-01
-1.05353963e+00 6.77100897e-01 -3.13690990e-01 -2.90525109e-01
1.58934224e+00 -5.85244000e-01 -2.45635002e-03 -4.42424715e-01
-5.78415811e-01 4.97195810e-01 7.74527848e-01 3.71873498e-01
8.95108104e-01 -1.40417778e+00 -3.19759995e-01 1.04847498e-01
-9.32262391e-02 -1.02737926e-01 5.36036670e-01 7.66395688e-01
-5.07283270e-01 2.84755230e-01 -1.79783478e-01 -7.85865843e-01
-1.57384694e+00 5.90458751e-01 5.22293270e-01 -2.52030045e-01
-4.53703165e-01 6.77193522e-01 -2.12092161e-01 1.51099727e-01
2.71836072e-01 -5.98839074e-02 -1.29697040e-01 -2.28506215e-02
1.76801369e-01 6.06644273e-01 -1.89066514e-01 -8.07366967e-01
-6.91303909e-01 1.00370812e+00 3.54054809e-01 -9.92882326e-02
1.27579129e+00 -7.61153936e-01 3.60141248e-01 9.32107419e-02
1.25901544e+00 -1.63101003e-01 -1.75761366e+00 8.46584961e-02
-5.63665688e-01 -7.12967098e-01 3.21931511e-01 -6.36414289e-01
-1.47979581e+00 4.94560540e-01 9.54686403e-01 -3.42042953e-01
1.55698180e+00 -1.67342141e-01 1.07590067e+00 4.03130770e-01
1.02801633e+00 -1.19707489e+00 2.97915697e-01 -1.17636457e-01
5.39105296e-01 -1.25721657e+00 2.84668833e-01 -2.85537452e-01
-7.03441620e-01 9.99749005e-01 1.04880798e+00 -5.46376333e-02
3.38225096e-01 1.44268230e-01 -2.51802355e-01 -1.26376003e-01
-8.11829388e-01 5.10794576e-04 1.13979183e-01 3.05298477e-01
-1.58938050e-01 -1.25668228e-01 9.52336937e-02 4.45750147e-01
2.04816625e-01 -8.31015259e-02 5.47803521e-01 8.75983417e-01
-3.04837763e-01 -8.05518329e-01 -4.37773556e-01 3.68160725e-01
-1.65614024e-01 3.13491315e-01 2.33270824e-02 7.93918133e-01
2.95223951e-01 1.11032152e+00 -9.19102877e-02 -8.35552216e-01
5.31233847e-01 -2.90981621e-01 5.23977578e-01 -2.64301270e-01
-3.52831811e-01 2.57418424e-01 2.04264998e-01 -8.24258804e-01
-6.10852182e-01 -6.34998024e-01 -9.37862515e-01 -3.71937394e-01
-7.01258004e-01 -1.80453897e-01 6.57106757e-01 8.39763761e-01
5.87020993e-01 6.93598568e-01 8.44286263e-01 -1.42143106e+00
-3.99933934e-01 -4.43523079e-01 -3.44341993e-01 1.11873604e-01
2.52314538e-01 -8.92562449e-01 -3.07965308e-01 1.83890611e-02] | [7.203241348266602, -1.0757694244384766] |
df0cd4c0-a3b0-4ad8-b5e9-92d04b038102 | a-learning-based-method-for-online-adjustment | 2008.06262 | null | https://arxiv.org/abs/2008.06262v1 | https://arxiv.org/pdf/2008.06262v1.pdf | A Learning-based Method for Online Adjustment of C-arm Cone-Beam CT Source Trajectories for Artifact Avoidance | During spinal fusion surgery, screws are placed close to critical nerves suggesting the need for highly accurate screw placement. Verifying screw placement on high-quality tomographic imaging is essential. C-arm Cone-beam CT (CBCT) provides intraoperative 3D tomographic imaging which would allow for immediate verification and, if needed, revision. However, the reconstruction quality attainable with commercial CBCT devices is insufficient, predominantly due to severe metal artifacts in the presence of pedicle screws. These artifacts arise from a mismatch between the true physics of image formation and an idealized model thereof assumed during reconstruction. Prospectively acquiring views onto anatomy that are least affected by this mismatch can, therefore, improve reconstruction quality. We propose to adjust the C-arm CBCT source trajectory during the scan to optimize reconstruction quality with respect to a certain task, i.e. verification of screw placement. Adjustments are performed on-the-fly using a convolutional neural network that regresses a quality index for possible next views given the current x-ray image. Adjusting the CBCT trajectory to acquire the recommended views results in non-circular source orbits that avoid poor images, and thus, data inconsistencies. We demonstrate that convolutional neural networks trained on realistically simulated data are capable of predicting quality metrics that enable scene-specific adjustments of the CBCT source trajectory. Using both realistically simulated data and real CBCT acquisitions of a semi-anthropomorphic phantom, we show that tomographic reconstructions of the resulting scene-specific CBCT acquisitions exhibit improved image quality particularly in terms of metal artifacts. Since the optimization objective is implicitly encoded in a neural network, the proposed approach overcomes the need for 3D information at run-time. | ['Jan-Nico Zäch', 'Russell Taylor', 'Cong Gao', 'Mathias Unberath', 'Nassir Navab', 'Mareike Thies', 'Andreas Maier'] | 2020-08-14 | null | null | null | null | ['tomographic-reconstructions'] | ['medical'] | [ 1.43230513e-01 1.64668068e-01 -1.22378498e-01 -2.62551606e-01
-7.40865469e-01 -3.97996485e-01 2.53192514e-01 4.96480435e-01
-7.23355711e-01 4.37493622e-01 9.10026431e-02 -7.63050675e-01
-3.54583025e-01 -7.07073271e-01 -1.08647394e+00 -3.65108460e-01
-9.67312828e-02 1.03796220e+00 1.92862377e-01 -2.59050969e-02
1.67415693e-01 1.00443232e+00 -1.25852394e+00 1.22828245e-01
6.31818235e-01 1.00969899e+00 5.82694948e-01 5.72746694e-01
1.77601725e-01 7.19314635e-01 -3.14753532e-01 -7.78599903e-02
4.42050844e-01 -5.32551587e-01 -6.58874989e-01 4.46150452e-02
2.36243904e-01 -4.04813766e-01 -1.54359505e-01 9.73115385e-01
3.57588977e-01 -3.16450484e-02 4.41367358e-01 -6.25512719e-01
5.81113771e-02 5.21640241e-01 -9.80230495e-02 1.79934576e-01
1.57925934e-01 3.23689848e-01 2.21169695e-01 -7.34580278e-01
8.00160050e-01 4.36039507e-01 7.22052157e-01 3.57209623e-01
-1.04692793e+00 -4.91531104e-01 -4.24124688e-01 2.44477659e-01
-8.40494871e-01 -7.83737376e-02 5.25752664e-01 -6.07552171e-01
6.70144618e-01 4.69873458e-01 1.18242407e+00 5.54795325e-01
1.05912852e+00 7.16987774e-02 1.24652898e+00 -4.82828885e-01
2.76994199e-01 -1.03971206e-01 -3.60753894e-01 9.60181713e-01
3.25548083e-01 4.01502013e-01 -1.56889364e-01 1.95293009e-01
1.23442912e+00 1.50356695e-01 -8.16730976e-01 -9.01243329e-01
-1.23491240e+00 4.40815181e-01 8.93512607e-01 4.37769622e-01
-8.30609500e-01 1.11015849e-01 2.67920822e-01 1.53632928e-02
-2.58949697e-01 8.29660654e-01 -1.25917658e-01 -1.63149267e-01
-8.41860294e-01 -7.54361078e-02 3.01312804e-01 6.38314545e-01
2.11080581e-01 -1.48156613e-01 2.73427460e-02 4.59114492e-01
1.21516854e-01 2.95306981e-01 8.00923228e-01 -9.34544861e-01
1.73037931e-01 4.62023079e-01 3.28676850e-02 -5.76083481e-01
-8.61329198e-01 -7.05474019e-01 -6.76231861e-01 6.19794369e-01
5.25221050e-01 2.31336594e-01 -1.16772676e+00 1.29141283e+00
3.42019707e-01 -9.87635106e-02 -2.62165248e-01 1.27103484e+00
4.59242314e-01 1.43003345e-01 -2.98396498e-01 -5.36346197e-01
1.33588529e+00 -7.96072841e-01 -6.77634537e-01 -2.57039875e-01
5.71595490e-01 -7.14970887e-01 9.91311431e-01 5.28565288e-01
-1.48922336e+00 -3.19535494e-01 -1.24632454e+00 2.76003480e-01
1.78464487e-01 1.17694356e-01 2.79108822e-01 5.12990355e-01
-7.96032488e-01 7.73300469e-01 -1.41038477e+00 -3.74618685e-03
3.14153284e-01 5.78523517e-01 -5.86783350e-01 6.82114437e-02
-7.89376378e-01 1.48282182e+00 1.37628004e-01 2.48234048e-01
-7.40812242e-01 -8.76189470e-01 -7.26692975e-01 3.15317847e-02
5.68364143e-01 -6.94609821e-01 1.44229972e+00 -7.72536457e-01
-1.64149761e+00 8.11542928e-01 1.74972937e-01 -4.72471744e-01
8.07525158e-01 -7.08031580e-02 -1.45820439e-01 3.90162259e-01
7.25806504e-02 8.89826268e-02 5.75997829e-01 -1.30604327e+00
-2.18615517e-01 -6.16740286e-01 -1.61637291e-02 2.21223369e-01
3.77414316e-01 -2.84006506e-01 -6.08477950e-01 -3.83730650e-01
7.83059061e-01 -1.12611449e+00 -5.97404420e-01 3.03645819e-01
-2.56460577e-01 5.95767736e-01 1.05425552e-01 -6.91983163e-01
7.64711559e-01 -1.59811497e+00 3.59552577e-02 3.81129414e-01
7.39988536e-02 -1.18935086e-01 5.63254416e-01 -6.09578714e-02
-4.79305446e-01 -3.49521756e-01 -2.78044373e-01 4.08694567e-03
-7.52207100e-01 9.74256247e-02 2.92114317e-01 8.93534243e-01
-4.85785246e-01 7.46357083e-01 -7.10053682e-01 -3.48310441e-01
5.99530160e-01 4.43145126e-01 -5.31279266e-01 3.03849339e-01
3.27481627e-02 9.78963852e-01 -2.15937898e-01 3.62003505e-01
5.88230908e-01 -4.46004987e-01 2.69096702e-01 -5.41672587e-01
-1.86538264e-01 2.66693711e-01 -6.97699487e-01 2.17203808e+00
-9.53527093e-01 3.68398905e-01 1.83177531e-01 -5.40565014e-01
5.50840139e-01 3.28022957e-01 8.64523172e-01 -1.05812883e+00
5.92551172e-01 6.44891143e-01 4.23035204e-01 -4.10241455e-01
1.42879456e-01 -6.49924099e-01 3.65206271e-01 4.03751731e-01
-1.49814799e-01 -6.03887200e-01 -4.37540263e-01 -6.15982898e-02
9.02215719e-01 2.38388218e-02 3.68476152e-01 -3.32868725e-01
3.70523751e-01 2.42969722e-01 7.00732619e-02 6.78790808e-01
4.94656563e-02 1.00937319e+00 1.96412489e-01 -7.26612329e-01
-1.20080543e+00 -1.13118005e+00 -4.44683492e-01 -1.43438041e-01
4.21754211e-01 1.32671416e-01 -6.97067618e-01 -4.13154095e-01
-3.02668720e-01 9.76233780e-01 -7.17059553e-01 -2.44900808e-01
-9.80313241e-01 -4.05477017e-01 2.00064257e-02 4.52024549e-01
-5.22218421e-02 -9.81203914e-01 -1.57737231e+00 4.85882849e-01
-1.03452824e-01 -8.59317780e-01 -7.56908581e-02 7.40805089e-01
-1.30514300e+00 -1.41020536e+00 -6.20333195e-01 -2.16645062e-01
9.03567374e-01 -8.50362554e-02 1.08136475e+00 2.23878950e-01
-3.31021875e-01 2.17482030e-01 -3.86476815e-02 -1.51065871e-01
-8.94798040e-01 -2.87947357e-01 -9.42790061e-02 -4.76426661e-01
-4.40159112e-01 -3.45143408e-01 -8.83700311e-01 2.88830578e-01
-8.28560591e-01 4.37498003e-01 5.71805596e-01 9.71648991e-01
9.61497605e-01 -1.78370938e-01 -1.25920594e-01 -7.91074574e-01
3.61390144e-01 -1.50890499e-01 -9.51088190e-01 1.44886717e-01
-8.48434508e-01 1.19257465e-01 7.89671302e-01 -9.19736996e-02
-1.01386440e+00 -4.80059162e-02 -4.03768033e-01 -5.89564562e-01
-1.83488578e-01 4.75853711e-01 2.69390613e-01 -2.20497012e-01
8.27987611e-01 3.80657986e-02 3.47978443e-01 -1.37718499e-01
-2.01702148e-01 2.96454225e-02 7.51654565e-01 -4.28990841e-01
2.75014043e-01 6.31574452e-01 3.37519765e-01 -2.64905006e-01
-4.93336946e-01 -2.05234513e-01 -7.18849540e-01 -6.42879784e-01
9.35302198e-01 -4.57385689e-01 -8.98566067e-01 1.01658940e-01
-9.75104809e-01 -2.21352026e-01 -4.36086863e-01 1.03908646e+00
-8.70753050e-01 8.06713030e-02 -5.23838282e-01 -3.29791874e-01
-2.73359895e-01 -2.07517576e+00 9.24383581e-01 -3.52886356e-02
-6.99514374e-02 -8.53455245e-01 -4.69236299e-02 1.78595439e-01
3.83874714e-01 3.74497503e-01 1.19185877e+00 -2.76646376e-01
-8.60147715e-01 -2.98230231e-01 2.65256464e-01 2.09344197e-02
1.20788619e-01 -3.67974311e-01 -6.87449932e-01 -1.69526771e-01
5.77306390e-01 -5.96064217e-02 2.13422582e-01 9.95049715e-01
1.32089067e+00 1.31554157e-01 -2.79184908e-01 1.16374707e+00
1.72611916e+00 5.93458414e-01 4.54568028e-01 7.64019370e-01
5.72368026e-01 2.47599527e-01 5.18946826e-01 1.40743926e-01
9.02381167e-02 9.87430274e-01 8.68995368e-01 -5.21967299e-02
-1.00364439e-01 -3.00855488e-02 -3.05897474e-01 6.34853780e-01
-3.38619709e-01 9.71180573e-02 -1.14078963e+00 1.99115202e-01
-1.22154295e+00 -4.12050635e-01 -3.98724914e-01 2.62122488e+00
7.87195206e-01 3.16182613e-01 -4.37232316e-01 -3.74001749e-02
2.99738646e-01 -3.03701162e-01 -5.97445488e-01 -2.52921432e-01
2.82792628e-01 4.11948651e-01 9.97357547e-01 6.70333207e-01
-4.65150476e-01 2.28056341e-01 5.50321245e+00 2.10464075e-01
-1.61939514e+00 3.08576673e-01 3.46402675e-01 -2.59755462e-01
-3.09431493e-01 -1.16560973e-01 -4.41729091e-02 2.79364079e-01
8.14969480e-01 6.29659817e-02 2.82118112e-01 6.28011167e-01
3.28802586e-01 -5.47379196e-01 -1.12854552e+00 8.88334811e-01
-1.96000114e-02 -1.73759496e+00 -2.85712451e-01 -1.90373249e-02
4.33356762e-01 1.02335297e-01 -4.68110852e-02 -2.25930139e-01
-9.99822170e-02 -1.06215799e+00 9.85149622e-01 5.31673789e-01
9.64875102e-01 -7.16895044e-01 7.76006937e-01 3.31890166e-01
-5.77754378e-01 1.03367262e-01 -1.18477866e-01 5.54635167e-01
4.77146089e-01 3.53654742e-01 -1.22843552e+00 7.21355319e-01
7.53100812e-01 1.18744723e-01 -3.36409867e-01 1.18918979e+00
-1.07147515e-01 1.17354847e-01 -3.62189591e-01 4.65069473e-01
1.72697157e-01 -2.00795293e-01 4.29194868e-01 5.52689075e-01
5.98904610e-01 3.04022402e-01 -3.94415528e-01 7.42542863e-01
1.68755606e-01 2.22276956e-01 -4.78858232e-01 7.90155172e-01
5.41929975e-02 7.78874159e-01 -1.14414787e+00 -1.28901660e-01
-1.45205647e-01 7.24951863e-01 1.45725086e-01 9.44460649e-03
-7.23419309e-01 5.18945992e-01 1.34152666e-01 5.49884558e-01
-4.97896113e-02 -3.34778032e-03 -7.48098314e-01 -8.43557119e-01
1.16559736e-01 -7.87549198e-01 7.03520924e-02 -1.06823039e+00
-4.35086697e-01 9.96763051e-01 1.23899728e-02 -1.33539855e+00
-5.11911392e-01 -5.85796833e-01 -3.36780578e-01 9.03145373e-01
-1.23558128e+00 -6.78480029e-01 -5.40191829e-01 4.24942970e-01
3.97380263e-01 2.77860910e-01 8.16906869e-01 1.24127353e-02
8.66709426e-02 1.95195496e-01 -6.20568125e-03 -3.36313993e-01
5.70998311e-01 -1.24339068e+00 -7.77764767e-02 6.61994696e-01
-9.46091935e-02 4.96610194e-01 8.92325938e-01 -8.07058454e-01
-1.41971827e+00 -5.99797130e-01 2.78763890e-01 -4.14158493e-01
3.26097578e-01 1.35474324e-01 -7.89767921e-01 8.20781648e-01
-7.65578309e-03 1.57480553e-01 3.93598288e-01 -3.41693223e-01
3.41952026e-01 -4.48013330e-03 -1.11753905e+00 4.00971085e-01
6.31478965e-01 -2.39758000e-01 -5.07801354e-01 3.03287208e-01
1.06155351e-01 -1.29202855e+00 -1.07694304e+00 5.95875978e-01
8.03357065e-01 -1.43810260e+00 1.04460025e+00 -3.61309797e-01
6.70512915e-01 -7.52331391e-02 2.24293157e-01 -1.50896764e+00
7.20282421e-02 -2.93757748e-02 4.29922283e-01 -6.54616952e-02
2.88606763e-01 -3.74376416e-01 9.73977566e-01 6.32793605e-01
-5.54135442e-01 -9.79285002e-01 -1.31114686e+00 -4.66235250e-01
1.10369278e-02 -7.37790823e-01 3.45609367e-01 7.35830963e-01
-1.29241303e-01 -3.49013954e-01 1.28056109e-01 5.33595860e-01
4.90026087e-01 2.10071921e-01 4.63980466e-01 -1.01340115e+00
-4.30898458e-01 -4.31843549e-01 -3.80667120e-01 -7.41182506e-01
-2.18924478e-01 -9.82534707e-01 2.02344239e-01 -1.66181815e+00
-2.35106982e-02 -8.47689867e-01 -3.33296843e-02 2.42410544e-02
-6.11142404e-02 -6.76634610e-02 1.55417040e-01 4.65336204e-01
3.05059522e-01 3.35284323e-01 1.84299695e+00 2.35630885e-01
-2.01935261e-01 2.70181090e-01 3.91937345e-02 9.59617972e-01
5.14881968e-01 -6.06036782e-01 -4.00567830e-01 -5.59150875e-01
3.01401258e-01 9.21615124e-01 3.76623869e-01 -1.14016402e+00
4.56586719e-01 5.80056310e-02 5.62730849e-01 -7.33852923e-01
4.32250082e-01 -1.59761202e+00 6.62622213e-01 9.24882233e-01
-2.51858711e-01 3.96484137e-01 2.64937907e-01 1.52782902e-01
-2.76571155e-01 -4.45202351e-01 1.02546453e+00 -4.50154841e-01
-7.75098130e-02 2.78165728e-01 -2.92928785e-01 -3.81365508e-01
9.32220161e-01 -3.52057278e-01 2.91828364e-01 -3.00515860e-01
-1.06631839e+00 -4.49617028e-01 8.14127922e-01 8.27768221e-02
8.97488296e-01 -1.03198981e+00 -3.75660747e-01 4.99878705e-01
9.62045160e-04 4.00853604e-01 4.46712971e-01 1.27029836e+00
-1.37631595e+00 4.64665115e-01 -5.17229140e-01 -1.00592422e+00
-8.85814607e-01 6.43109798e-01 1.22389054e+00 -4.31139618e-01
-8.67366791e-01 8.03888261e-01 2.58110255e-01 -3.62298429e-01
-7.42825419e-02 -7.13812888e-01 1.76402554e-02 -6.30786002e-01
1.53693661e-01 -9.76497307e-02 7.41772592e-01 -5.03255546e-01
-2.81860709e-01 5.64076483e-01 1.32090420e-01 -1.16088979e-01
1.28441310e+00 -2.00402848e-02 1.09364018e-01 2.66526222e-01
7.28430033e-01 1.16937794e-01 -1.27077639e+00 1.12437814e-01
-3.67551863e-01 -6.33650303e-01 3.74554127e-01 -1.09611702e+00
-1.37737620e+00 7.60691822e-01 9.02807951e-01 -3.10934246e-01
1.22428322e+00 -1.15328886e-01 4.55144227e-01 -1.02880031e-01
6.53151512e-01 -5.89926898e-01 -2.50920951e-01 -2.56971736e-02
9.12558258e-01 -1.06873882e+00 1.47155985e-01 -4.13402885e-01
-4.95783865e-01 1.48995471e+00 6.29590690e-01 -4.64835539e-02
7.57991493e-01 5.13142049e-01 3.08795154e-01 -6.16073906e-01
-2.67545193e-01 4.68902230e-01 2.50143021e-01 3.22165132e-01
2.77075827e-01 2.33759850e-01 -1.36126131e-01 5.89037016e-02
-5.07540941e-01 1.16299026e-01 7.74897635e-01 1.00702345e+00
-1.68980777e-01 -7.71249831e-01 -7.08039224e-01 4.83607173e-01
-5.49840927e-01 2.41170004e-02 3.45986694e-01 9.41651106e-01
-1.21931843e-02 3.40263695e-01 1.63420454e-01 1.89678878e-01
6.95831478e-01 -3.17176789e-01 1.00437474e+00 -4.63945478e-01
-9.18065310e-01 2.47388750e-01 -1.31747276e-01 -7.88880110e-01
-2.29870573e-01 -5.07177770e-01 -1.56933987e+00 1.35088027e-01
-2.99680173e-01 8.55517536e-02 1.15273130e+00 9.79869783e-01
-2.12337554e-01 9.81347501e-01 2.67835498e-01 -8.03381860e-01
-3.45482290e-01 -6.40008390e-01 -4.92597073e-01 2.78925896e-01
4.68566984e-01 -8.03294182e-01 5.90431644e-03 -1.14940643e-01] | [13.689618110656738, -2.779433488845825] |
5c787bae-95fd-4f84-9bb5-f17e56b9056f | invariance-to-quantile-selection-in | 2212.14262 | null | https://arxiv.org/abs/2212.14262v1 | https://arxiv.org/pdf/2212.14262v1.pdf | Invariance to Quantile Selection in Distributional Continuous Control | In recent years distributional reinforcement learning has produced many state of the art results. Increasingly sample efficient Distributional algorithms for the discrete action domain have been developed over time that vary primarily in the way they parameterize their approximations of value distributions, and how they quantify the differences between those distributions. In this work we transfer three of the most well-known and successful of those algorithms (QR-DQN, IQN and FQF) to the continuous action domain by extending two powerful actor-critic algorithms (TD3 and SAC) with distributional critics. We investigate whether the relative performance of the methods for the discrete action space translates to the continuous case. To that end we compare them empirically on the pybullet implementations of a set of continuous control tasks. Our results indicate qualitative invariance regarding the number and placement of distributional atoms in the deterministic, continuous action setting. | ['Ioannis Iossifidis', 'Tobias Glasmachers', 'Muhammad Saif-ur-Rehman', 'Felix Grün'] | 2022-12-29 | null | null | null | null | ['distributional-reinforcement-learning', 'continuous-control'] | ['methodology', 'playing-games'] | [-1.49886802e-01 4.51802909e-02 -3.43794107e-01 -1.12601966e-01
-8.36065114e-01 -8.04392159e-01 1.12551379e+00 1.47073880e-01
-8.92155707e-01 1.15600538e+00 6.29323781e-01 -3.94307196e-01
-4.51354086e-01 -7.51453519e-01 -3.99036258e-01 -8.22683394e-01
-2.21806094e-02 9.14514303e-01 2.06986368e-01 -4.54039186e-01
2.96550095e-01 1.76643059e-01 -1.37349463e+00 -1.97858170e-01
6.29398465e-01 5.49271405e-01 -2.52648950e-01 6.43068075e-01
-1.27166405e-01 7.79038966e-01 -6.84247136e-01 5.28795691e-03
3.32855880e-01 -5.34771562e-01 -6.69045568e-01 -3.34265202e-01
4.16379981e-02 -3.36278558e-01 -5.30892089e-02 1.05825734e+00
7.23801613e-01 4.77533787e-01 1.01390135e+00 -1.43821299e+00
-7.96103239e-01 9.29087698e-01 -6.27986550e-01 1.06977560e-01
2.31648609e-01 5.48784435e-01 8.93012941e-01 -3.47268611e-01
5.10873377e-01 1.66303575e+00 5.49650192e-01 5.23430824e-01
-1.55641365e+00 -3.60860616e-01 1.36347070e-01 -1.30529568e-01
-9.62089658e-01 -2.94714212e-01 4.56049323e-01 -6.17324054e-01
1.02344286e+00 -4.38666731e-01 5.57488263e-01 1.25631702e+00
2.07545027e-01 5.35611451e-01 1.56856823e+00 -6.05716407e-01
7.53055215e-01 -1.00496113e-01 -1.80825695e-01 2.28447810e-01
2.01815978e-01 5.93692958e-01 -5.93207292e-02 -5.16539752e-01
8.27723742e-01 -3.37110341e-01 4.80493903e-02 -7.84933329e-01
-1.11004651e+00 1.23329782e+00 -8.11052173e-02 1.92811131e-01
-4.69411582e-01 6.79294407e-01 8.77063572e-01 3.07386607e-01
4.71246541e-01 4.93218899e-01 -3.95602852e-01 -7.70117223e-01
-5.43036401e-01 9.18369949e-01 8.76801133e-01 8.06603670e-01
3.67493421e-01 4.05915111e-01 -4.86811578e-01 5.84884107e-01
2.60677040e-01 5.79180956e-01 5.73666871e-01 -1.36050022e+00
4.18106258e-01 -8.14657211e-02 6.71469331e-01 -4.49925303e-01
-4.14781719e-01 -4.91121635e-02 -2.64373481e-01 6.38043344e-01
1.06743371e+00 -6.62761629e-01 -5.62988043e-01 2.07930422e+00
3.82199377e-01 -2.60394484e-01 1.91948384e-01 7.62566984e-01
-3.29835147e-01 4.17621285e-01 5.65870285e-01 -9.21229720e-02
8.19692612e-01 -6.11306429e-01 -5.52979052e-01 9.45489481e-02
6.31622076e-01 -5.11823237e-01 1.43333769e+00 4.30401534e-01
-1.01674294e+00 -4.23211515e-01 -7.62218833e-01 1.85769543e-01
-1.97260514e-01 -1.91698626e-01 3.55604708e-01 5.85685849e-01
-1.19292998e+00 9.07850325e-01 -8.37721825e-01 -2.23298013e-01
3.68121356e-01 3.11334342e-01 2.36363456e-01 3.46892208e-01
-1.33044350e+00 1.12848747e+00 4.28086579e-01 -6.05120182e-01
-1.14969718e+00 -6.00831389e-01 -5.62118292e-01 -3.87265719e-02
4.59361106e-01 -5.85973620e-01 1.92820895e+00 -1.22535229e+00
-2.01578975e+00 3.91339928e-01 6.19187534e-01 -7.98162997e-01
7.68369436e-01 -2.30030641e-01 5.28103560e-02 -1.19523190e-01
6.30581304e-02 5.65565705e-01 8.42108309e-01 -8.45599055e-01
-6.22675419e-01 -2.69862384e-01 3.40209246e-01 3.60619426e-01
1.25077283e-02 1.78838428e-02 6.12035394e-01 -6.30640149e-01
-8.92620504e-01 -8.01025391e-01 -3.14897627e-01 -3.08756772e-02
-2.21603978e-02 -6.70745850e-01 3.80971223e-01 -4.41988148e-02
9.80692208e-01 -2.20536089e+00 4.84705687e-01 5.53115644e-03
-1.06180705e-01 1.09080389e-01 -1.51136354e-01 8.82380426e-01
-1.56415358e-01 3.17303017e-02 -2.06759274e-01 -2.31623754e-01
7.49359846e-01 4.81796980e-01 -6.28163457e-01 9.06142831e-01
3.96378376e-02 6.15137458e-01 -1.14427710e+00 -4.83921349e-01
1.16787814e-01 2.66958565e-01 -7.93686688e-01 9.41635817e-02
-6.90100610e-01 2.71475405e-01 -7.14211524e-01 6.60320967e-02
2.27486685e-01 3.09195340e-01 2.77216196e-01 4.56961036e-01
-4.57389951e-01 1.91991061e-01 -1.03387821e+00 1.59138966e+00
-1.51330382e-01 2.55298585e-01 8.27370435e-02 -9.91783798e-01
6.90129757e-01 1.81477621e-01 6.02635086e-01 -6.80376649e-01
1.77878246e-01 1.65987819e-01 3.47516447e-01 -2.32662156e-01
5.79256535e-01 -7.97194302e-01 -2.56809026e-01 7.03275979e-01
-4.40950971e-03 -3.04262936e-01 2.81114548e-01 -4.18309011e-02
9.43004787e-01 8.41152191e-01 5.17833233e-01 -7.83478320e-01
1.10851765e-01 8.27743635e-02 2.32745260e-01 8.31314027e-01
-5.92277825e-01 3.79489958e-01 1.08693504e+00 -1.63654879e-01
-1.26451910e+00 -1.25507343e+00 -1.21789478e-01 1.43074369e+00
-3.68716627e-01 -1.85676694e-01 -8.32877874e-01 -5.54095030e-01
5.27703345e-01 1.09487820e+00 -7.95698822e-01 -2.29817912e-01
-5.83271384e-01 -2.73617089e-01 7.50854850e-01 7.16922522e-01
4.07319628e-02 -1.32659185e+00 -1.16008258e+00 4.78291333e-01
5.04606485e-01 -4.33323115e-01 -3.81224751e-01 4.65755850e-01
-6.16405010e-01 -9.00366366e-01 -7.78900921e-01 -2.59304971e-01
3.54704559e-02 -5.71888387e-01 1.20340514e+00 -5.71789682e-01
1.05842389e-01 8.00046802e-01 -2.50733793e-01 -7.98895121e-01
-6.12295389e-01 -3.18597369e-02 3.48251343e-01 -3.89017195e-01
2.67980009e-01 -5.61867416e-01 -4.25463766e-01 2.60867714e-03
-1.08411229e+00 -6.15956306e-01 2.07796663e-01 7.64907777e-01
4.44373429e-01 -1.99654073e-01 7.64350235e-01 -1.00625491e+00
1.18744838e+00 -6.47241652e-01 -7.80149043e-01 -1.02787197e-01
-6.74595058e-01 4.96390730e-01 9.78026450e-01 -6.25141442e-01
-1.00719142e+00 -2.04075709e-01 -7.19772652e-02 -4.35097665e-01
-9.60674286e-02 4.69177216e-01 1.57138839e-01 6.09733939e-01
9.44527447e-01 -6.61464632e-02 3.28308851e-01 -3.09855938e-01
8.28200340e-01 4.11455601e-01 2.67173022e-01 -1.39044940e+00
5.58060050e-01 2.92569071e-01 -7.86835402e-02 -6.15187466e-01
-5.50432086e-01 5.74679524e-02 -3.57473463e-01 7.19825029e-02
8.81053209e-01 -5.49596786e-01 -6.22971356e-01 3.75020027e-01
-8.37021410e-01 -1.20356703e+00 -1.07367420e+00 4.05427814e-01
-1.33374500e+00 2.81563818e-01 -4.68169361e-01 -1.15728986e+00
1.04061462e-01 -1.24064434e+00 9.98239458e-01 7.58689493e-02
-3.15041959e-01 -1.35944605e+00 8.22078586e-01 -6.70413673e-01
7.01380670e-01 4.06547636e-01 1.30569708e+00 -6.41249716e-01
2.05813289e-01 5.29575706e-01 1.35934904e-01 3.41509253e-01
-3.27096060e-02 -6.84224814e-02 -6.14302516e-01 -3.24448436e-01
-3.99080152e-03 -6.75977349e-01 7.08949745e-01 7.23345041e-01
8.56577039e-01 -1.72916219e-01 3.64958905e-02 2.19375253e-01
1.43156278e+00 4.01853412e-01 4.80081379e-01 5.86990416e-01
3.08950663e-01 4.19160277e-01 8.53318870e-01 7.65553951e-01
9.62365791e-02 6.20076239e-01 3.63286972e-01 2.38927141e-01
9.54829231e-02 -4.52888280e-01 7.32439101e-01 2.35617474e-01
4.12800722e-02 -2.22041532e-01 -8.79157245e-01 5.92776299e-01
-2.17942095e+00 -1.16927230e+00 5.08603454e-01 2.06125617e+00
1.18123746e+00 2.08784834e-01 9.90543902e-01 -1.18991427e-01
5.55527449e-01 3.11699629e-01 -8.16063046e-01 -8.32974613e-01
2.30602890e-01 4.51327354e-01 5.61062276e-01 5.43589115e-01
-8.80939484e-01 7.32586563e-01 7.06862068e+00 8.45434248e-01
-6.98366761e-01 8.15402493e-02 4.11444247e-01 -1.33027300e-01
-2.38285244e-01 -2.76095383e-02 -4.42900449e-01 4.59863484e-01
1.23332798e+00 -3.85249078e-01 7.75246620e-01 1.08606374e+00
2.67100632e-01 -1.91103444e-01 -1.38161469e+00 6.28292859e-01
-4.39954787e-01 -8.25431943e-01 -1.93628013e-01 1.45682603e-01
5.78963280e-01 1.33777618e-01 3.66490692e-01 7.22401500e-01
1.14915061e+00 -1.15046573e+00 1.19824290e+00 5.96226156e-01
8.11669052e-01 -9.97100115e-01 3.97929579e-01 3.27165663e-01
-6.87561691e-01 -3.00271869e-01 -4.70079333e-01 -4.30295616e-01
3.25398594e-02 7.58301392e-02 -5.20418584e-01 1.17335089e-01
3.18540484e-01 5.45118272e-01 -5.04311696e-02 6.86076045e-01
-1.16103433e-01 7.59190440e-01 -3.48345786e-01 -2.97853917e-01
7.02823043e-01 -4.79018986e-01 5.64569056e-01 9.76687431e-01
8.50359872e-02 -1.77722558e-01 5.00352144e-01 1.01995313e+00
3.32337439e-01 -8.22216421e-02 -7.06580758e-01 -2.81222165e-01
3.27656180e-01 7.51520514e-01 -4.68748719e-01 -2.07400069e-01
-2.27557510e-01 1.43458515e-01 5.14984190e-01 4.24973965e-01
-1.19256186e+00 -3.36308479e-01 1.03800523e+00 9.31199417e-02
3.99330586e-01 -4.66401160e-01 1.47191659e-01 -7.68902183e-01
-3.35642129e-01 -1.24058402e+00 4.24830556e-01 -4.03551519e-01
-1.62466764e+00 1.27024695e-01 5.89296639e-01 -8.04744184e-01
-6.55811667e-01 -6.31163657e-01 -3.32579732e-01 7.65334010e-01
-1.26449752e+00 -4.52118099e-01 6.14571035e-01 7.61403024e-01
4.39678669e-01 -1.60354763e-01 8.62731755e-01 -6.99520335e-02
-3.31882149e-01 2.13916883e-01 5.02644479e-01 2.68571638e-02
7.96121061e-01 -1.86465335e+00 2.09159493e-01 3.66747409e-01
-1.28094135e-02 5.08041978e-01 1.03790665e+00 -4.19191152e-01
-1.38511777e+00 -8.44855070e-01 -4.93025966e-02 -6.51084840e-01
1.01735127e+00 -1.37280360e-01 -4.60443556e-01 7.53949642e-01
5.02494752e-01 -1.70092896e-01 1.70524746e-01 -2.31656190e-02
-3.49248141e-01 1.58387139e-01 -1.20034420e+00 6.23117805e-01
8.77776563e-01 -4.98535961e-01 -8.00891459e-01 1.16962880e-01
7.10818529e-01 -3.55139107e-01 -7.76384950e-01 -1.77844360e-01
3.98685575e-01 -1.08134985e+00 7.90169418e-01 -1.02555442e+00
5.48233390e-01 -1.93982437e-01 -3.60679418e-01 -1.86743355e+00
-2.53579557e-01 -7.02959478e-01 7.50778839e-02 1.19783044e+00
5.73956184e-02 -9.35116708e-01 2.82208711e-01 3.17173302e-01
-1.59949049e-01 -8.86637926e-01 -1.01396775e+00 -8.58187854e-01
1.06610632e+00 -2.46896043e-01 6.00547016e-01 5.24649739e-01
1.62874386e-01 2.26289853e-01 -5.21831736e-02 -6.48927987e-01
5.61279356e-01 4.64030653e-02 6.44121051e-01 -9.99494255e-01
-7.72334039e-01 -7.29771495e-01 -1.45186201e-01 -7.88417995e-01
4.31605816e-01 -7.67832994e-01 6.43068701e-02 -1.17300010e+00
-1.21834418e-02 -5.09267867e-01 -3.41494322e-01 5.52261710e-01
1.58046350e-01 -3.20050448e-01 1.96701497e-01 -3.50059867e-02
-7.37906218e-01 8.87322664e-01 1.04441357e+00 9.59984660e-02
-2.47382119e-01 -2.50448883e-01 -7.04553187e-01 8.05293322e-01
9.30994153e-01 -5.36259353e-01 -7.88335681e-01 -3.42516214e-01
3.81203145e-01 2.49837860e-02 1.77059442e-01 -7.38718688e-01
-1.64207667e-01 -5.51339924e-01 1.67105854e-01 2.10281108e-02
-7.98073485e-02 -5.04534721e-01 -2.20630974e-01 4.72351700e-01
-7.60601759e-01 5.77061594e-01 3.12808663e-01 7.59831309e-01
4.81761061e-02 -1.20387912e-01 9.96380925e-01 -1.93693757e-01
-4.01374340e-01 5.80180138e-02 -5.82319319e-01 9.63559330e-01
1.13696432e+00 1.82804734e-01 -3.48528445e-01 -5.25412560e-01
-5.98204434e-01 8.30647573e-02 5.11099637e-01 1.09203652e-01
1.15513034e-01 -1.24405015e+00 -6.71878040e-01 -3.13240916e-01
-2.33479574e-01 -2.34446481e-01 -2.97218233e-01 8.76286745e-01
-3.16013455e-01 2.76730001e-01 -4.98535246e-01 -2.83154190e-01
-5.87835133e-01 7.82802999e-01 5.95154941e-01 -4.30475861e-01
-4.40287530e-01 2.02321589e-01 8.81814957e-02 -3.68715823e-01
2.59898394e-01 -5.81560075e-01 5.15713431e-02 1.88620880e-01
2.79143304e-01 5.58431685e-01 -6.11011982e-01 -3.28473240e-01
-1.70471877e-01 2.79440135e-01 1.72378883e-01 -5.79140186e-01
1.57522845e+00 1.95208281e-01 1.74688533e-01 7.05537736e-01
8.59148145e-01 -1.12109065e-01 -1.79384363e+00 1.22515507e-01
1.92620661e-02 -2.39968076e-01 -1.85110778e-01 -7.64275551e-01
-6.05890274e-01 8.22295785e-01 4.51609999e-01 4.40054864e-01
7.78393149e-01 -1.71034828e-01 1.99099779e-01 9.29957181e-02
6.24682128e-01 -1.33218050e+00 -1.30550656e-02 6.82112038e-01
7.96041310e-01 -5.12251675e-01 1.17910527e-01 5.80726922e-01
-8.92149448e-01 1.03957307e+00 3.83468539e-01 -7.12246597e-01
4.67575788e-01 4.38543618e-01 -1.31349415e-01 -8.85096043e-02
-7.21263826e-01 -2.77038127e-01 -4.50928271e-01 1.01656711e+00
4.12714809e-01 1.53396606e-01 -5.07990241e-01 3.04065555e-01
-1.91380322e-01 -1.17161006e-01 5.70097506e-01 9.31943476e-01
-3.87041420e-01 -1.38447642e+00 -1.65346250e-01 1.34180397e-01
-5.79471767e-01 9.06864107e-02 -2.77184874e-01 1.30733168e+00
-1.18301280e-01 7.02787638e-01 -3.86653445e-03 2.45760798e-01
3.97058636e-01 2.67478287e-01 7.58725166e-01 -5.99129558e-01
-7.58521557e-01 7.70168379e-02 7.84773603e-02 -5.03860652e-01
-5.27151883e-01 -1.04298878e+00 -1.50897145e+00 -2.36560285e-01
4.13887836e-02 2.47678414e-01 3.54588360e-01 8.93632472e-01
1.26422942e-01 3.87348205e-01 2.34535068e-01 -9.94400918e-01
-1.87588251e+00 -8.69230866e-01 -8.74326587e-01 4.19493079e-01
3.45799863e-01 -1.00079322e+00 -3.27085823e-01 -4.23353463e-01] | [4.057559013366699, 2.5656254291534424] |
24971cd6-71d0-4f6c-918c-f778e387f236 | federated-non-negative-matrix-factorization | 2205.13300 | null | https://arxiv.org/abs/2205.13300v1 | https://arxiv.org/pdf/2205.13300v1.pdf | Federated Non-negative Matrix Factorization for Short Texts Topic Modeling with Mutual Information | Non-negative matrix factorization (NMF) based topic modeling is widely used in natural language processing (NLP) to uncover hidden topics of short text documents. Usually, training a high-quality topic model requires large amount of textual data. In many real-world scenarios, customer textual data should be private and sensitive, precluding uploading to data centers. This paper proposes a Federated NMF (FedNMF) framework, which allows multiple clients to collaboratively train a high-quality NMF based topic model with locally stored data. However, standard federated learning will significantly undermine the performance of topic models in downstream tasks (e.g., text classification) when the data distribution over clients is heterogeneous. To alleviate this issue, we further propose FedNMF+MI, which simultaneously maximizes the mutual information (MI) between the count features of local texts and their topic weight vectors to mitigate the performance degradation. Experimental results show that our FedNMF+MI methods outperform Federated Latent Dirichlet Allocation (FedLDA) and the FedNMF without MI methods for short texts by a significant margin on both coherence score and classification F1 score. | ['Jing Xiao', 'Qinliang Su', 'Ruiyi Zhang', 'Jianzong Wang', 'Shijing Si'] | 2022-05-26 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-3.33238095e-01 8.54370371e-03 -5.18965065e-01 -6.56316876e-01
-8.92889977e-01 -2.63871878e-01 6.66144192e-01 -5.83633445e-02
-4.18201312e-02 4.62991506e-01 4.86129731e-01 -2.20234901e-01
-2.00611740e-01 -8.44393253e-01 -3.88871223e-01 -8.90082479e-01
1.35685191e-01 8.35016131e-01 -1.43254012e-01 1.85899615e-01
1.43544227e-01 -2.71907181e-01 -1.23916852e+00 4.81159300e-01
1.08142722e+00 7.10237801e-01 5.90650201e-01 2.61267722e-01
-7.11290121e-01 3.00237417e-01 -6.13521755e-01 -4.17278260e-01
2.79083624e-02 1.61657676e-01 -6.39971852e-01 2.09754214e-01
6.38595670e-02 -3.77653271e-01 -3.53637367e-01 8.23325872e-01
1.91776454e-01 3.94483268e-01 7.25265265e-01 -1.60156584e+00
-4.35858965e-01 8.57657433e-01 -7.42968202e-01 3.61508504e-02
-2.38796785e-01 -2.51439303e-01 1.20094109e+00 -1.16182935e+00
5.85190356e-01 1.37184405e+00 2.11021289e-01 1.89537331e-01
-1.07468247e+00 -9.92165446e-01 6.21929348e-01 4.38077711e-02
-1.13089716e+00 -8.77140909e-02 6.51853621e-01 -3.84988517e-01
7.87079513e-01 1.34424672e-01 3.39308113e-01 9.65982676e-01
3.50715399e-01 1.23218822e+00 5.49061179e-01 -2.11023986e-01
3.58964413e-01 4.61221486e-01 3.98838013e-01 2.29653008e-02
1.48686811e-01 -7.11423993e-01 -1.09838343e+00 -8.50295603e-01
3.56551409e-01 6.05040848e-01 -1.26446828e-01 -3.28094602e-01
-1.16629839e+00 1.35395277e+00 -2.52636909e-01 2.47437075e-01
-5.44469774e-01 -2.86242545e-01 4.68440056e-01 2.92386919e-01
1.24142849e+00 4.56631370e-02 -7.47622371e-01 -2.11063311e-01
-1.17297864e+00 2.17881247e-01 7.38976240e-01 9.16962981e-01
9.78131294e-01 -3.39465410e-01 -2.65746862e-01 1.06878233e+00
6.47243917e-01 6.43407166e-01 8.67756486e-01 -7.36474395e-01
7.57048130e-01 7.54743516e-01 -2.25762725e-02 -1.09975541e+00
-1.03551075e-01 -2.63188690e-01 -7.80616164e-01 -7.11319745e-01
-7.18018115e-02 -3.84149939e-01 -4.28467125e-01 1.51728439e+00
5.84523439e-01 1.82650194e-01 -3.56884673e-02 8.44571888e-01
4.41619605e-01 1.26984072e+00 -9.32991505e-02 -5.73340774e-01
1.19698083e+00 -8.94123435e-01 -9.48488235e-01 1.24394486e-03
8.03538620e-01 -8.97564590e-01 1.07045233e+00 6.59193516e-01
-5.75007379e-01 -8.45841318e-02 -5.26764870e-01 1.33411288e-01
-2.80602332e-02 1.25358284e-01 1.05516315e+00 6.14263058e-01
-7.60023832e-01 1.22600503e-01 -9.83114362e-01 -2.54671395e-01
5.03172398e-01 4.65404093e-01 -3.35625052e-01 -3.50790590e-01
-1.15938187e+00 -1.66053623e-02 8.04033577e-02 -2.69473046e-01
-8.39269578e-01 -1.01827395e+00 -3.92408133e-01 3.51983070e-01
4.24524099e-01 -5.69777191e-01 1.35906649e+00 -3.48766893e-01
-1.25771117e+00 2.65297979e-01 -6.79347813e-01 -4.56568718e-01
2.24269331e-01 -4.88937914e-01 -3.40818673e-01 -1.21393561e-01
3.69523197e-01 5.21050334e-01 1.02843392e+00 -9.32632923e-01
-8.72749031e-01 -6.29837513e-01 -5.98881006e-01 1.29442781e-01
-1.25554740e+00 -7.05575421e-02 -4.34928179e-01 -6.71340287e-01
3.28635335e-01 -6.95252836e-01 -1.71859026e-01 -4.16246265e-01
-3.95584941e-01 -1.00261140e+00 1.62593019e+00 -4.52377319e-01
1.38880861e+00 -2.24699306e+00 -2.86777616e-01 2.55084038e-01
4.20513719e-01 1.48800472e-02 3.61849479e-02 6.60932124e-01
3.29641223e-01 -5.45088341e-03 3.19733381e-01 -7.04965353e-01
6.72622696e-02 2.14737102e-01 -7.34313130e-01 4.32590842e-01
-3.08982462e-01 6.28931820e-01 -5.69928467e-01 -6.01212859e-01
3.36868800e-02 5.28005302e-01 -6.45042896e-01 2.01366559e-01
-5.20409226e-01 5.66675775e-02 -5.82512736e-01 2.59295344e-01
6.83292568e-01 -5.51172376e-01 4.08900440e-01 8.16343576e-02
8.61366466e-02 4.41235006e-01 -1.20524895e+00 1.96360552e+00
-7.50378191e-01 6.68551207e-01 2.34223679e-01 -8.90605032e-01
9.97317016e-01 5.90956330e-01 9.61069286e-01 -2.18499169e-01
-7.12761804e-02 -1.93842590e-01 -5.26390612e-01 -2.39424825e-01
7.23934233e-01 2.54044663e-02 7.19951559e-03 1.40923285e+00
2.14656353e-01 4.58722949e-01 4.84338924e-02 9.39464748e-01
7.47191072e-01 -6.68262839e-01 -2.04358026e-01 -3.63017946e-01
-8.70970264e-03 -1.78851292e-01 7.17754483e-01 5.00092208e-01
1.32031798e-01 3.82949829e-01 3.61169994e-01 -1.65575057e-01
-8.53642166e-01 -8.07261825e-01 -2.27978170e-01 1.65099418e+00
-1.78950056e-01 -8.90127480e-01 -4.32133526e-01 -6.59295976e-01
1.79171070e-01 6.74591601e-01 -3.69153202e-01 -2.23264500e-01
-1.66711286e-01 -1.14931619e+00 8.95216502e-03 2.39780262e-01
1.32989705e-01 -4.07621801e-01 4.84457090e-02 3.00386816e-01
-6.34776413e-01 -8.32393289e-01 -7.29640484e-01 9.42466632e-02
-1.15251958e+00 -5.96415222e-01 -6.17190838e-01 -4.93328989e-01
4.72971022e-01 1.10722589e+00 7.36842990e-01 -5.48836052e-01
1.00215953e-02 1.08991079e-01 -3.66924584e-01 -4.69431758e-01
-2.48695202e-02 3.56493026e-01 2.73260951e-01 2.90192604e-01
8.52776825e-01 -7.06990421e-01 -5.65596223e-01 6.06177986e-01
-1.08490133e+00 1.84258800e-02 2.66158789e-01 9.71577466e-01
2.01018363e-01 5.75981259e-01 8.27183664e-01 -1.16391170e+00
8.53212774e-01 -9.48854744e-01 -2.32918084e-01 1.93799138e-01
-1.12852025e+00 -1.63321212e-01 2.56928891e-01 -6.38776839e-01
-1.39026177e+00 -3.60579520e-01 3.35477442e-01 -5.35897076e-01
1.16449177e-01 5.78442574e-01 -4.32460487e-01 7.58638859e-01
3.99016768e-01 2.64952689e-01 -1.27137274e-01 -7.07587719e-01
4.79742229e-01 1.35370505e+00 1.53436616e-01 -4.52362835e-01
3.58019322e-01 6.74550831e-01 -6.68384671e-01 -7.53915250e-01
-8.10092986e-01 -1.21600914e+00 -2.08193853e-01 4.56955321e-02
3.38098973e-01 -1.35621309e+00 -4.78046954e-01 2.80504405e-01
-1.08262885e+00 -1.11266570e-02 -9.01892260e-02 8.29159260e-01
-3.52582723e-01 2.10369617e-01 -5.43031216e-01 -8.00440848e-01
-7.40371048e-01 -7.72736013e-01 1.31715071e+00 -1.14194885e-01
-2.69120723e-01 -9.93312240e-01 2.15626568e-01 5.68004370e-01
3.44514012e-01 -6.08628392e-01 1.19405615e+00 -1.00416124e+00
-2.69361347e-01 -3.55805248e-01 -1.45037353e-01 9.65200737e-02
3.88439655e-01 -1.83967233e-01 -1.06161320e+00 -4.02243257e-01
2.85276562e-01 -1.86969593e-01 8.13394010e-01 6.42732561e-01
8.50085974e-01 -5.28473318e-01 -6.99904323e-01 1.79901019e-01
1.03175437e+00 -4.04608734e-02 1.46121562e-01 1.98923677e-01
5.84186912e-01 6.81163132e-01 7.51803458e-01 1.02637899e+00
4.99345183e-01 5.29242158e-01 1.46040499e-01 2.62365818e-01
4.23137784e-01 -2.95419604e-01 4.26947653e-01 9.79764998e-01
7.13553488e-01 -5.43892562e-01 -8.14111054e-01 5.78698456e-01
-2.17281675e+00 -7.29297042e-01 -1.20363854e-01 2.04459333e+00
7.62991071e-01 -7.81886280e-02 1.05450429e-01 1.28578901e-01
9.48338687e-01 7.59682432e-02 -6.32250667e-01 -5.21601662e-02
7.52397580e-03 -4.29866701e-01 1.39109362e-02 2.33990885e-02
-1.03764868e+00 8.01171720e-01 5.04766798e+00 1.18997192e+00
-1.03924000e+00 6.70665383e-01 6.60043597e-01 -5.49025893e-01
-5.71815431e-01 -1.11315832e-01 -1.08739519e+00 7.40633011e-01
1.20305943e+00 -5.98488927e-01 -1.22065388e-01 1.26456892e+00
5.80291808e-01 -2.92287115e-02 -7.94288635e-01 9.79517877e-01
3.05008888e-03 -1.29920125e+00 2.73997277e-01 5.41020632e-01
1.18774998e+00 2.69068360e-01 2.37909958e-01 4.20481890e-01
4.99344587e-01 -5.87238133e-01 6.56931102e-02 1.95222020e-01
2.87168950e-01 -9.47001278e-01 5.01213014e-01 8.54019940e-01
-7.24032998e-01 -1.62033305e-01 -7.51120090e-01 2.56466925e-01
2.47144133e-01 1.53238916e+00 -1.06846714e+00 2.92096466e-01
6.98689520e-01 6.78932667e-01 5.54531962e-02 7.31712520e-01
2.79040962e-01 1.13126338e+00 -4.99374419e-01 4.94536459e-02
2.91930944e-01 -1.14329636e-01 4.88947749e-01 9.97290671e-01
3.35176796e-01 -2.38000363e-01 3.38119954e-01 5.42642355e-01
-3.22344989e-01 6.62851036e-01 -2.88025886e-01 -1.70573562e-01
5.07720470e-01 1.42958224e+00 -7.32335031e-01 -3.35166484e-01
-4.07563418e-01 7.43681967e-01 2.65876688e-02 2.27966905e-01
-3.33748460e-01 -3.34516838e-02 8.49462152e-01 2.52686322e-01
3.07292223e-01 -2.54822195e-01 -1.88000411e-01 -1.55981338e+00
1.81733482e-02 -6.17828190e-01 5.41766167e-01 -3.41388315e-01
-1.51515377e+00 3.38409036e-01 -1.16563879e-01 -9.60384429e-01
-4.46612537e-01 6.54688850e-02 -8.02839577e-01 7.92643726e-01
-1.03621018e+00 -1.11550200e+00 7.57707134e-02 8.95924747e-01
9.62335765e-01 -4.46137875e-01 7.10874617e-01 3.10094178e-01
-6.79056287e-01 4.65490103e-01 8.85409534e-01 -1.81009308e-01
1.03274608e+00 -9.59055424e-01 4.12662178e-01 5.54901540e-01
3.89077067e-01 9.95077968e-01 4.20839638e-01 -8.18418086e-01
-1.41657805e+00 -1.38347518e+00 1.02028227e+00 -1.26438648e-01
5.30279636e-01 -7.80819356e-01 -9.88753796e-01 5.31594992e-01
9.40722823e-02 -3.48780960e-01 1.27568924e+00 8.24186802e-01
-3.98112357e-01 -1.05431765e-01 -8.77721906e-01 1.21686690e-01
3.51975143e-01 -3.98972988e-01 -2.30460018e-01 9.83146727e-01
1.06479108e+00 3.31788152e-01 -9.05621886e-01 -8.82083178e-02
3.49916965e-01 -4.65128571e-01 4.48871315e-01 -8.40745270e-01
1.94339469e-01 1.14714593e-01 -3.18798035e-01 -1.40125620e+00
-2.56221622e-01 -8.15747559e-01 -5.45388639e-01 1.62025225e+00
3.77925634e-01 -5.51480532e-01 1.30048943e+00 7.73408532e-01
3.07515144e-01 -6.39550507e-01 -8.45767498e-01 -6.06905162e-01
1.30843138e-02 -8.84038687e-01 6.76397383e-01 1.02504814e+00
5.10228395e-01 5.85812271e-01 -4.94333655e-01 -2.64338590e-02
4.00348872e-01 5.57346523e-01 9.64623094e-01 -1.50449193e+00
-4.96978201e-02 -1.34804165e-02 -7.14513883e-02 -1.14547646e+00
3.63966912e-01 -7.32550204e-01 -2.18584076e-01 -1.38784897e+00
6.34186089e-01 -6.58916533e-01 -6.52790591e-02 3.38242590e-01
-1.83499873e-01 -3.65747511e-01 3.02753933e-02 8.48184884e-01
-8.80684972e-01 8.65817785e-01 8.38537216e-01 -9.61811692e-02
-3.71725410e-01 4.12118793e-01 -8.48445475e-01 4.20484453e-01
7.60299563e-01 -7.40790844e-01 -7.12857723e-01 -3.62253964e-01
2.78054386e-01 -4.93508801e-02 -2.90502578e-01 -3.56360435e-01
5.32429814e-01 -2.20927894e-01 2.19963342e-01 -9.76185620e-01
2.73399830e-01 -6.65785670e-01 -2.32481480e-01 -7.16298372e-02
-4.46736664e-01 -2.37221330e-01 -2.91135490e-01 8.77951443e-01
-3.71226698e-01 6.11429960e-02 1.45725638e-01 5.52014932e-02
-1.55695334e-01 5.42010963e-01 -6.60242736e-01 -1.90950468e-01
8.51360381e-01 3.30193400e-01 -3.62755954e-01 -8.00721049e-01
-5.04783750e-01 5.58646858e-01 7.13911802e-02 4.93713111e-01
4.99358416e-01 -1.16665649e+00 -7.58424163e-01 3.76729339e-01
-1.97200067e-02 4.42239881e-01 6.63132191e-01 7.51796126e-01
4.93502825e-01 1.08686423e+00 5.00708818e-01 -6.92512214e-01
-1.09687483e+00 4.53333139e-01 -4.97321278e-01 -4.48593736e-01
-5.87090492e-01 8.37064564e-01 5.55303216e-01 -5.97791731e-01
3.14150244e-01 6.51153773e-02 9.64780971e-02 3.00742298e-01
7.72435486e-01 4.78167892e-01 5.05376220e-01 -2.41989851e-01
-1.05798589e-02 -1.94159254e-01 -8.07469666e-01 -2.87356585e-01
1.49701786e+00 -4.67487574e-01 -2.23422810e-01 5.51277578e-01
1.52684379e+00 -2.56097555e-01 -1.07295942e+00 -7.92698264e-01
-1.49931222e-01 -5.52842855e-01 5.68188310e-01 -2.96693921e-01
-1.19085145e+00 8.62553656e-01 2.81538337e-01 9.80269462e-02
8.67143929e-01 1.29219353e-01 1.16299176e+00 5.04487753e-01
2.68823415e-01 -1.02738702e+00 1.35921806e-01 2.66446441e-01
3.25908273e-01 -1.19889963e+00 -8.71572550e-03 -4.10006911e-01
-6.72823310e-01 9.07664597e-01 5.97442150e-01 2.89521843e-01
1.16196966e+00 1.61750481e-01 9.93824825e-02 -2.43941933e-01
-1.52677143e+00 5.84804237e-01 1.95362881e-01 3.74003947e-01
3.89625907e-01 3.71322818e-02 -1.87403992e-01 9.72115815e-01
-2.70972192e-01 -2.14915320e-01 1.64729759e-01 8.73658597e-01
-6.47296667e-01 -1.25317538e+00 -5.61483264e-01 1.03703845e+00
-5.99032164e-01 -9.89485830e-02 -7.77622610e-02 6.68758973e-02
-2.11569026e-01 1.17248428e+00 2.26227388e-01 -4.28987831e-01
-4.11758035e-01 2.54913986e-01 -4.70120668e-01 -9.11627114e-01
-2.95038849e-01 7.17675507e-01 -4.62240756e-01 -3.20529014e-01
-2.23084241e-01 -8.86013091e-01 -1.11564243e+00 -4.39079404e-01
-8.95702243e-01 7.33594596e-01 1.13199210e+00 9.57070768e-01
6.53995574e-01 7.42200762e-02 1.08627272e+00 -4.16998774e-01
-5.64975977e-01 -1.25470972e+00 -9.22862172e-01 5.15013747e-02
-1.16933905e-01 -4.43652689e-01 -2.10780978e-01 4.79666516e-02] | [10.388615608215332, 6.929931163787842] |
44d20490-5c12-45f4-9d54-2c331db6bdc4 | the-effects-of-data-size-on-automated-essay | 2108.13275 | null | https://arxiv.org/abs/2108.13275v1 | https://arxiv.org/pdf/2108.13275v1.pdf | The effects of data size on Automated Essay Scoring engines | We study the effects of data size and quality on the performance on Automated Essay Scoring (AES) engines that are designed in accordance with three different paradigms; A frequency and hand-crafted feature-based model, a recurrent neural network model, and a pretrained transformer-based language model that is fine-tuned for classification. We expect that each type of model benefits from the size and the quality of the training data in very different ways. Standard practices for developing training data for AES engines were established with feature-based methods in mind, however, since neural networks are increasingly being considered in a production setting, this work seeks to inform us as to how to establish better training data for neural networks that will be used in production. | ['Paul van Wamelen', 'Amy Harris', 'Milan Patel', 'Susan Lottridge', 'Amir Jafari', 'Christopher Ormerod'] | 2021-08-30 | null | null | null | null | ['automated-essay-scoring'] | ['natural-language-processing'] | [ 3.79843600e-02 -2.81908125e-01 -1.51107639e-01 -3.61102402e-01
-4.29016113e-01 -6.52729332e-01 5.80365956e-01 2.34145969e-01
-6.90786123e-01 4.63421017e-01 4.07122344e-01 -7.87775099e-01
-3.91089916e-01 -8.97437632e-01 -3.16620439e-01 -1.24330834e-01
4.18278515e-01 2.95397222e-01 -1.69743113e-02 -5.53103805e-01
5.75775564e-01 4.98847455e-01 -1.73681891e+00 2.60559708e-01
8.72263908e-01 8.07904005e-01 1.75547540e-01 8.41878712e-01
-2.32079640e-01 1.22835791e+00 -8.85000706e-01 -7.41337180e-01
1.89044073e-01 -4.32870388e-01 -7.05148757e-01 -4.15909678e-01
4.51057732e-01 2.54611094e-02 -4.21362489e-01 5.62539220e-01
3.61234725e-01 1.20964408e-01 7.45919287e-01 -7.69982278e-01
-7.54789174e-01 7.25418687e-01 1.88045114e-01 5.48317373e-01
3.65329266e-01 2.70933747e-01 1.07782185e+00 -7.50685751e-01
4.79462534e-01 5.91326177e-01 8.93504083e-01 4.93865877e-01
-1.07253230e+00 -6.89391553e-01 -1.86958209e-01 2.36443475e-01
-1.03399718e+00 -6.19583368e-01 6.43549562e-01 -4.45455700e-01
1.09988558e+00 2.03656748e-01 8.22880030e-01 1.22399426e+00
3.93840730e-01 5.03036499e-01 1.34728336e+00 -8.86056006e-01
2.17400603e-02 5.11737049e-01 6.37127995e-01 6.44728005e-01
4.11491901e-01 2.87195027e-01 -1.00820100e+00 -1.38057664e-01
6.06576979e-01 -4.56206709e-01 -1.63156837e-01 -1.33063821e-02
-7.68115461e-01 1.09755099e+00 -9.13355779e-03 6.45763934e-01
-4.43488598e-01 -1.34236187e-01 5.84962785e-01 7.81033218e-01
2.02171966e-01 1.12128818e+00 -7.84519315e-01 -6.24484539e-01
-1.35825944e+00 2.60007650e-01 1.04477525e+00 6.37887836e-01
2.90877610e-01 4.11717802e-01 -5.11769056e-01 9.78369236e-01
2.46063098e-01 2.22189829e-01 1.17160308e+00 -7.09308386e-01
4.80752289e-01 8.64801764e-01 -1.39738411e-01 -9.03266311e-01
-3.89681309e-01 -5.87832093e-01 -1.93332002e-01 1.23258233e-01
6.29858613e-01 -1.77603006e-01 -7.42636681e-01 1.64713788e+00
-3.85379076e-01 -2.68940598e-01 4.72189412e-02 6.20145619e-01
8.12103868e-01 4.01906192e-01 2.82593705e-02 1.59867126e-02
1.28554380e+00 -9.25710320e-01 -7.07165241e-01 -1.40518427e-01
9.95644629e-01 -8.29210401e-01 1.53723943e+00 7.77897000e-01
-1.38771617e+00 -7.13149428e-01 -1.37824357e+00 -1.02111407e-01
-6.22173429e-01 3.25735956e-01 5.73428392e-01 1.44502962e+00
-1.18461084e+00 8.13586235e-01 -4.94003594e-01 -8.50605741e-02
-1.37074850e-02 4.14273381e-01 -1.36953786e-01 2.16016963e-01
-1.35713029e+00 1.47471499e+00 1.95875764e-01 -2.76686568e-02
-3.26958835e-01 -5.45328557e-01 -6.39967322e-01 1.81011066e-01
-1.61472365e-01 -5.45251906e-01 1.49876714e+00 -1.11118996e+00
-1.84993851e+00 6.50036037e-01 3.99902314e-02 -5.22908628e-01
1.45775095e-01 2.90305428e-02 -5.06324470e-01 -1.97793514e-01
-3.47293288e-01 -3.13402759e-03 7.52222955e-01 -5.42904258e-01
-5.41953504e-01 -2.63603866e-01 -2.02146228e-02 1.20750554e-01
-9.28507805e-01 1.37573987e-01 7.68683702e-02 -6.44931912e-01
-3.90723258e-01 -8.65405917e-01 1.94546804e-01 -6.75351918e-01
1.48317769e-01 -4.49023068e-01 4.12111014e-01 -8.69575918e-01
1.78998625e+00 -1.85391581e+00 -6.88133687e-02 4.24648404e-01
4.98989597e-02 4.92412925e-01 -2.95018941e-01 3.95322800e-01
-1.69727765e-02 1.01225957e-01 3.56191725e-01 -6.67292476e-02
1.51382342e-01 -2.11473092e-01 -1.31265551e-01 -4.68233004e-02
4.41188812e-02 8.11709046e-01 -7.21152008e-01 -1.98229656e-01
-1.32336080e-01 2.32320741e-01 -3.75436425e-01 4.09385830e-01
1.40007809e-01 -1.66318417e-01 -2.20913187e-01 5.38660944e-01
-4.42042425e-02 -2.16077909e-01 2.39044130e-02 2.53694326e-01
-3.00268948e-01 9.45290029e-01 -9.45895195e-01 1.41976476e+00
-7.74062037e-01 8.32305491e-01 -3.54639173e-01 -5.76733947e-01
1.17081416e+00 4.96000350e-01 -3.05301305e-02 -8.53419304e-01
1.65621713e-01 4.73659307e-01 4.98206168e-01 -2.52352268e-01
1.15012705e+00 -6.42164052e-02 7.95459598e-02 7.28651524e-01
3.22623909e-01 -2.37159267e-01 5.16520917e-01 1.70956895e-01
1.25775313e+00 -1.43901510e-02 -9.60518885e-03 -3.29623312e-01
4.33179796e-01 1.59800500e-01 2.12275997e-01 7.97719061e-01
-5.07917926e-02 3.49547833e-01 3.22441816e-01 -3.63923520e-01
-1.07447314e+00 -5.84264100e-01 -3.17632169e-01 1.30827641e+00
-7.64836133e-01 -7.38839447e-01 -7.55407095e-01 -5.71746707e-01
-2.62608200e-01 1.05795085e+00 -5.47244430e-01 -6.31671727e-01
-3.56367379e-01 -5.61131656e-01 8.06497157e-01 4.11189526e-01
1.31931111e-01 -1.13868272e+00 -8.13108623e-01 4.04358953e-01
3.71841155e-02 -5.14667332e-01 -5.31292975e-01 5.31848907e-01
-1.00711739e+00 -9.41268086e-01 -5.52772343e-01 -4.49875355e-01
3.08607161e-01 5.19319996e-02 1.40875077e+00 3.02403927e-01
3.06575418e-01 5.94077528e-01 -3.80420208e-01 -8.24961603e-01
-5.39532185e-01 6.01158679e-01 4.63491566e-02 -3.51756036e-01
8.43397498e-01 -2.49034882e-01 -2.03917116e-01 1.43106699e-01
-8.37913811e-01 -2.76789337e-01 7.51844943e-01 9.93787587e-01
-2.20148936e-01 5.77805378e-02 5.57480037e-01 -8.56499910e-01
1.48608959e+00 -1.71208188e-01 -3.76920193e-01 3.98852468e-01
-1.20258749e+00 1.98960230e-01 6.47451818e-01 -6.04349315e-01
-9.26163435e-01 -2.17707947e-01 -3.46884131e-02 -2.03398466e-01
1.06094375e-01 9.45972264e-01 4.08755451e-01 -4.77578610e-01
1.04879451e+00 2.64397532e-01 3.92315313e-02 -1.84892833e-01
-1.19864106e-01 8.78688991e-01 2.51055032e-01 -7.82007039e-01
7.54493356e-01 -5.66238225e-01 -3.86992484e-01 -4.15217727e-01
-6.29917145e-01 -1.28697604e-01 -5.59259892e-01 -3.13755006e-01
3.63818973e-01 -7.60460496e-01 -4.52673614e-01 4.47551638e-01
-8.62286270e-01 -5.11070728e-01 -3.61066043e-01 5.65556586e-01
-2.28204519e-01 -1.00015616e-02 -6.81972027e-01 -6.73383117e-01
-5.86020648e-01 -1.24119532e+00 5.25697052e-01 4.26612228e-01
-6.96053505e-01 -1.12233400e+00 4.69369620e-01 5.43024302e-01
9.24269676e-01 -5.07869720e-01 1.08067298e+00 -1.06224573e+00
5.11408411e-03 -5.03346980e-01 1.85433954e-01 6.02766752e-01
-1.17420390e-01 1.79279581e-01 -9.75435019e-01 -5.53331040e-02
2.29793593e-01 -5.65991163e-01 6.08021319e-01 2.14332506e-01
1.03105843e+00 -2.18404919e-01 3.59285265e-01 2.20637247e-01
1.09777367e+00 1.16909981e-01 6.11452043e-01 6.99073255e-01
3.65606159e-01 6.23684764e-01 1.86172634e-01 1.31849393e-01
3.60743701e-01 6.57421887e-01 -3.86749417e-01 3.69686037e-01
1.04946621e-01 -3.31090003e-01 6.17953658e-01 1.34081733e+00
-2.90879756e-01 -8.02062154e-02 -9.73276377e-01 3.47008795e-01
-1.43756402e+00 -8.12947631e-01 -1.59361407e-01 2.35594940e+00
1.22366130e+00 3.65258753e-01 3.55926007e-01 5.78581095e-01
2.44681254e-01 -1.69745777e-02 -7.63108209e-02 -1.12391889e+00
9.87452716e-02 8.61063957e-01 4.58733082e-01 7.91822523e-02
-5.39179981e-01 6.07600629e-01 7.18588829e+00 7.24472344e-01
-1.35393655e+00 -2.87896935e-02 4.44394141e-01 1.23517662e-02
-3.34669590e-01 -1.77026793e-01 -9.72632647e-01 5.44201910e-01
1.85557044e+00 -2.45959237e-01 4.69431758e-01 7.74157822e-01
1.69671685e-01 1.00785941e-01 -9.43469048e-01 6.07153594e-01
-8.98577273e-03 -1.24891889e+00 6.84801266e-02 7.43342862e-02
4.67240006e-01 -2.37659216e-01 2.01765865e-01 7.51284301e-01
3.56338233e-01 -1.20401931e+00 9.38862264e-01 6.71595931e-01
5.51576734e-01 -7.05017388e-01 8.33397746e-01 1.80183738e-01
-7.51303077e-01 -2.27151901e-01 -3.51985663e-01 -5.68690479e-01
-5.62117815e-01 5.65557003e-01 -7.16745198e-01 2.83149153e-01
5.02711713e-01 2.82456338e-01 -1.03780925e+00 9.34225023e-01
-3.19312751e-01 1.06022692e+00 4.41347919e-02 -5.24389625e-01
2.10558265e-01 -1.23218559e-01 1.04556747e-01 1.13274443e+00
3.21481407e-01 -1.74284160e-01 -4.65717673e-01 6.83400571e-01
-3.06598488e-02 2.18247309e-01 -5.33503473e-01 -3.43640238e-01
6.72234058e-01 1.37490046e+00 -4.21813458e-01 -1.15000620e-01
-4.84522223e-01 6.25569701e-01 5.12612700e-01 6.90339580e-02
-4.85573798e-01 -6.38665974e-01 1.48488760e-01 2.22822934e-01
1.88783988e-01 -1.58541054e-01 -7.99098790e-01 -1.12574935e+00
-2.07996666e-01 -1.35564840e+00 3.96038920e-01 -7.93579221e-01
-1.29018998e+00 7.42661834e-01 -3.02279055e-01 -9.09813643e-01
-5.60779572e-01 -9.60946679e-01 -8.33883286e-01 1.10700929e+00
-1.27546346e+00 -8.01658630e-01 -1.54742613e-01 4.20506567e-01
3.88805330e-01 -7.24145472e-01 1.19795263e+00 5.68483844e-02
-5.57685435e-01 1.09784579e+00 1.48597866e-01 3.06449234e-01
6.91881716e-01 -1.33269429e+00 1.19291797e-01 6.85923100e-01
5.19962490e-01 9.73843217e-01 4.95466173e-01 -3.21844369e-01
-1.46874690e+00 -7.23623097e-01 1.30803275e+00 -7.75922835e-01
7.04692841e-01 -2.08928987e-01 -9.27439511e-01 5.53716898e-01
3.01969379e-01 -5.22041857e-01 9.73836124e-01 6.77368164e-01
-3.87772560e-01 -7.31696114e-02 -9.35697317e-01 4.90051985e-01
4.85302687e-01 -9.22776401e-01 -9.73700225e-01 -1.23682537e-03
1.94984198e-01 -4.77397412e-01 -1.13355875e+00 1.64853856e-01
7.26801872e-01 -8.46874714e-01 5.89377761e-01 -5.65227807e-01
7.05605984e-01 3.42132151e-01 2.04010144e-01 -1.72884059e+00
-6.82396054e-01 -4.87936884e-01 -1.91811949e-01 1.12215924e+00
8.40821624e-01 -7.19684184e-01 8.52223992e-01 9.67526197e-01
-1.80865526e-01 -8.64674985e-01 -7.42662907e-01 -6.95145190e-01
4.43880856e-01 -2.95023382e-01 6.39644623e-01 9.42857683e-01
3.09269160e-01 6.65843308e-01 9.99813080e-02 -4.65602040e-01
-6.39338568e-02 -3.53514850e-01 6.29028320e-01 -1.46263075e+00
-1.94768310e-01 -9.74365830e-01 -3.30632150e-01 -5.35234272e-01
3.66181806e-02 -6.59913778e-01 -3.04237425e-01 -9.31097090e-01
8.53299201e-02 -4.44418609e-01 -5.11041999e-01 4.33141440e-01
-1.98482156e-01 1.73245016e-02 2.98094422e-01 1.54549479e-01
-1.36423916e-01 2.67710268e-01 7.98656881e-01 7.17627928e-02
-3.92929941e-01 2.26341665e-01 -1.11462164e+00 3.62667769e-01
8.13410521e-01 -4.68509734e-01 -4.44766074e-01 -2.11556301e-01
4.30111438e-01 -7.87372366e-02 -9.09444690e-02 -1.13651228e+00
4.80470300e-01 2.23333891e-02 5.34456193e-01 -2.30210230e-01
7.90023580e-02 -5.86265564e-01 -1.24008022e-01 3.35190177e-01
-7.06565022e-01 5.95853209e-01 3.57387632e-01 -1.03711672e-01
-1.44183859e-01 -8.81072819e-01 4.60649550e-01 -1.29430175e-01
-3.32509369e-01 -7.17880130e-02 -6.18709326e-01 -4.73309010e-02
4.46410835e-01 -5.00026584e-01 -1.78081796e-01 -4.31202143e-01
-1.38220474e-01 -1.69193164e-01 2.53614008e-01 5.44463873e-01
3.80266488e-01 -1.26012337e+00 -7.74526834e-01 2.90120631e-01
2.68963445e-02 -6.22001052e-01 -1.23258039e-01 7.84087002e-01
-4.11340147e-01 9.20755327e-01 -3.82374495e-01 1.38189085e-02
-1.18321562e+00 5.41428104e-02 4.07781333e-01 -8.70737314e-01
-2.26234868e-01 5.97355366e-01 -5.96215606e-01 -6.70947611e-01
3.44376534e-01 -1.79205105e-01 -5.79764903e-01 -1.63681444e-03
5.85408390e-01 3.67847174e-01 6.81460857e-01 -3.24745089e-01
1.87393218e-01 -1.00798890e-01 -3.69931310e-01 -2.58824795e-01
1.19190860e+00 4.85118359e-01 1.40791401e-01 8.05236518e-01
6.19680047e-01 1.17677897e-01 -5.86307645e-01 -1.87039837e-01
2.51848012e-01 -2.42792085e-01 5.64285398e-01 -1.17754102e+00
-8.80853593e-01 5.39310992e-01 4.19396907e-01 2.74964660e-01
9.76096928e-01 -8.36807787e-01 5.84428668e-01 6.01131916e-01
3.26462269e-01 -1.48284781e+00 8.81976634e-02 8.56784344e-01
7.20003903e-01 -7.07598627e-01 -6.86658770e-02 1.87502474e-01
-5.20094812e-01 1.49607193e+00 6.28293335e-01 -2.16141455e-02
5.04367650e-01 3.45618397e-01 5.19265831e-02 -2.06041053e-01
-1.00374520e+00 2.60061353e-01 6.44130945e-01 3.83766949e-01
8.86963725e-01 -9.51271281e-02 -5.77842534e-01 1.07853365e+00
-8.69676530e-01 3.44112694e-01 8.17976594e-01 9.02754724e-01
-2.82668710e-01 -1.34681189e+00 -6.00241363e-01 1.20482635e+00
-5.38616776e-01 -3.39560688e-01 -6.18482292e-01 6.12880826e-01
-1.22738674e-01 9.35327888e-01 -1.16075411e-01 -7.30007112e-01
5.40689707e-01 7.25196660e-01 5.40327966e-01 -7.32907355e-01
-1.37141132e+00 -6.41326308e-01 3.98554206e-01 -6.55254871e-02
-8.23916644e-02 -7.80995309e-01 -6.53554082e-01 -4.68431890e-01
-7.73986280e-01 4.75240618e-01 8.36211205e-01 1.12623763e+00
2.26891339e-01 6.32210732e-01 4.74760324e-01 -3.90355617e-01
-1.03441739e+00 -1.41334713e+00 -5.34460127e-01 2.19254084e-02
-1.64168462e-01 -5.29001832e-01 -3.56415302e-01 -2.99763948e-01] | [11.348877906799316, 9.306035041809082] |
4cdf3af8-c8fd-4c71-ae38-0153b3d4b870 | a-study-of-the-effect-of-resolving-negation | 1907.03871 | null | https://arxiv.org/abs/1907.03871v1 | https://arxiv.org/pdf/1907.03871v1.pdf | A Study of the Effect of Resolving Negation and Sentiment Analysis in Recognizing Text Entailment for Arabic | Recognizing the entailment relation showed that its influence to extract the semantic inferences in wide-ranging natural language processing domains (text summarization, question answering, etc.) and enhanced the results of their output. For Arabic language, few attempts concerns with Arabic entailment problem. This paper aims to increase the entailment accuracy for Arabic texts by resolving negation of the text-hypothesis pair and determining the polarity of the text-hypothesis pair whether it is Positive, Negative or Neutral. It is noticed that the absence of negation detection feature gives inaccurate results when detecting the entailment relation since the negation revers the truth. The negation words are considered stop words and removed from the text-hypothesis pair which may lead wrong entailment decision. Another case not solved previously, it is impossible that the positive text entails negative text and vice versa. In this paper, in order to classify the text-hypothesis pair polarity, a sentiment analysis tool is used. We show that analyzing the polarity of the text-hypothesis pair increases the entailment accuracy. to evaluate our approach we used a dataset for Arabic textual entailment (ArbTEDS) consisted of 618 text-hypothesis pairs and showed that the Arabic entailment accuracy is increased by resolving negation for entailment relation and analyzing the polarity of the text-hypothesis pair. | ['Fatima T. AL-Khawaldeh'] | 2019-07-05 | null | null | null | null | ['negation-detection'] | ['natural-language-processing'] | [ 5.07474124e-01 5.11471093e-01 1.72811523e-01 -6.14307284e-01
-1.47990555e-01 -1.01678669e+00 8.72839630e-01 8.84124637e-01
-3.23890686e-01 1.12170577e+00 3.44191402e-01 -7.58217812e-01
-2.06822082e-01 -9.74009693e-01 -6.09523118e-01 -2.29817659e-01
2.26162240e-01 5.40494442e-01 1.04363821e-01 -9.62375104e-01
8.19502056e-01 2.04808444e-01 -1.84433949e+00 9.96877670e-01
8.42719734e-01 7.87025750e-01 -2.31737942e-01 7.08400607e-01
-4.72362936e-01 1.28914690e+00 -8.92253995e-01 -4.89948809e-01
6.60099834e-02 -9.34253573e-01 -1.27658474e+00 1.15813851e-01
1.99196488e-01 -3.44684422e-02 7.62831748e-01 1.25834715e+00
9.84413102e-02 -3.19505304e-01 1.12998807e+00 -1.44892168e+00
-1.57421917e-01 8.20864677e-01 -4.24195349e-01 1.74721673e-01
1.02519476e+00 -5.30337453e-01 1.11924481e+00 -8.70012939e-01
6.72583759e-01 1.17136908e+00 4.48196113e-01 -1.18514057e-02
-4.72341895e-01 -8.67940411e-02 7.24029988e-02 5.30156553e-01
-8.62616539e-01 -1.82829723e-01 5.67131817e-01 -2.36773893e-01
1.33132052e+00 6.72141552e-01 7.43003488e-01 2.03208983e-01
5.78975737e-01 3.74848187e-01 1.31966066e+00 -1.01523781e+00
6.58553764e-02 3.20310473e-01 5.82916379e-01 6.20881915e-01
6.16117001e-01 -9.11237836e-01 -1.28686324e-01 9.87868011e-02
-4.50501710e-01 -3.90695095e-01 -1.98997736e-01 5.55083454e-01
-8.65072608e-01 7.57383227e-01 -1.11343220e-01 9.02457476e-01
-5.12223184e-01 -6.58407152e-01 5.92768490e-01 7.78435111e-01
1.13830514e-01 4.25818980e-01 -5.02210438e-01 1.87748298e-02
-8.70916784e-01 3.19302768e-01 1.31058645e+00 7.51750052e-01
6.27320170e-01 -2.98308700e-01 2.51048952e-01 6.30089417e-02
2.17248797e-01 5.59990346e-01 7.98629165e-01 -3.70179564e-01
4.33588773e-01 1.45512700e+00 2.71685511e-01 -1.48779619e+00
-3.01394433e-01 -2.44295578e-02 -3.60715985e-01 2.43327558e-01
5.25594115e-01 -4.29783493e-01 -5.12684822e-01 1.14276111e+00
4.53322202e-01 -6.04935467e-01 9.74351525e-01 8.12874496e-01
1.00619304e+00 8.81166101e-01 -4.20094877e-01 -5.16854763e-01
2.11475825e+00 -4.34630215e-01 -1.26126432e+00 -2.55575657e-01
8.53538573e-01 -1.61979771e+00 7.70481288e-01 6.71738803e-01
-8.79960239e-01 9.38288309e-03 -1.58892536e+00 9.70600098e-02
-8.29716027e-01 2.91938603e-01 9.46848541e-02 3.67019176e-01
-4.64485437e-01 4.27662253e-01 -4.81983088e-02 -4.85030830e-01
-2.44147584e-01 3.18181872e-01 -5.62518239e-01 -8.50850940e-02
-1.47140944e+00 1.35907018e+00 6.00027084e-01 2.44018734e-01
2.26388425e-01 1.44983158e-01 -7.85458446e-01 -1.57701015e-01
5.36914647e-01 -2.42599040e-01 9.86779034e-01 -1.69897652e+00
-1.09124136e+00 9.56197441e-01 -4.10847634e-01 -6.40385389e-01
4.08692986e-01 -1.00795582e-01 -5.00053287e-01 3.03591222e-01
2.70281971e-01 -3.63317579e-02 8.48246515e-01 -8.20596516e-01
-5.98959088e-01 -4.94718879e-01 1.68217048e-01 4.31457460e-01
-3.05786747e-02 1.81551009e-01 4.16987836e-01 -2.23262340e-01
5.69981217e-01 -6.41690433e-01 4.27429050e-01 -9.46065485e-01
-4.57489431e-01 -3.40348572e-01 9.15003002e-01 -7.76862204e-01
1.33462512e+00 -1.67675924e+00 -3.82756442e-01 3.41325492e-01
-1.20891497e-01 2.24322692e-01 3.43712330e-01 7.24328876e-01
-4.29655761e-01 2.95220971e-01 -3.46737355e-01 6.28969789e-01
-1.32323790e-03 4.82171535e-01 -4.20807838e-01 2.72231877e-01
6.18929684e-01 2.98540503e-01 -4.54200715e-01 -8.07959616e-01
-4.53337505e-02 -3.89672294e-02 -2.41848782e-01 -4.34833504e-02
-3.81667137e-01 -2.99583554e-01 -5.12950532e-02 5.40578902e-01
8.09325218e-01 3.39418143e-01 5.07151484e-01 -7.49970913e-01
-2.15113282e-01 1.49578705e-01 -1.50875652e+00 4.21784103e-01
-1.98488891e-01 6.74491405e-01 -2.31853351e-01 -1.28393531e+00
1.17494392e+00 4.25332516e-01 6.52485788e-02 -6.36950195e-01
6.74228728e-01 5.74310541e-01 2.07446367e-01 -9.48306799e-01
8.46159697e-01 -5.77122271e-01 1.40903778e-02 4.75264758e-01
-1.37345240e-01 -4.88786668e-01 8.24405968e-01 3.76162052e-01
4.26279277e-01 5.52062578e-02 9.36297238e-01 -3.26098651e-01
1.28181398e+00 4.72903669e-01 2.05192678e-02 8.12221020e-02
-5.52489469e-03 2.72166312e-01 1.37678444e+00 -3.61825645e-01
-7.10718751e-01 -5.14292955e-01 -5.38809039e-02 5.92373312e-01
-1.73224532e-03 -4.95391190e-01 -7.60244727e-01 -7.55082250e-01
-4.24145788e-01 8.27030301e-01 -5.59443235e-01 9.05733854e-02
-6.60532355e-01 -1.05320382e+00 5.26480794e-01 -2.83323020e-01
4.67309266e-01 -8.01790893e-01 -8.61982405e-01 4.83322795e-03
-8.10005307e-01 -9.80744243e-01 4.24406320e-01 3.47321451e-01
-7.37628818e-01 -1.60146391e+00 -1.61221936e-01 -8.91151905e-01
8.82244587e-01 -4.15166110e-01 1.02299130e+00 1.08696170e-01
1.14617907e-01 -1.89847216e-01 -7.67884493e-01 -9.26734507e-01
-9.17966783e-01 -2.17798382e-01 -3.49362493e-01 -1.10630430e-01
7.08332300e-01 6.32746443e-02 -1.64034724e-01 8.18268135e-02
-1.09748304e+00 -4.22949463e-01 3.74531478e-01 3.55416358e-01
1.98019311e-01 2.38024637e-01 8.94206643e-01 -8.20900977e-01
1.24123478e+00 -5.54162562e-01 -1.12920456e-01 3.76758307e-01
-6.51783228e-01 3.55989963e-01 7.09881544e-01 -1.17742382e-01
-9.40329313e-01 -2.99349546e-01 -4.28427637e-01 8.94713998e-01
-1.78582415e-01 7.95090854e-01 -1.21934518e-01 3.56824487e-01
8.12485099e-01 6.53384179e-02 3.54395628e-01 2.80214638e-01
1.26530841e-01 7.92770207e-01 2.40150526e-01 -1.46246418e-01
2.27430463e-01 2.32939571e-01 5.12258649e-01 -9.21914339e-01
-7.50769556e-01 -3.13889712e-01 -3.51381809e-01 -4.41142678e-01
6.95220113e-01 -4.78088081e-01 -6.06544316e-01 7.16657788e-02
-1.54789102e+00 7.46010959e-01 -1.12866890e-02 3.78875852e-01
-7.17318729e-02 8.17878425e-01 -2.73259968e-01 -9.49067593e-01
-6.94694579e-01 -9.92343962e-01 3.97166133e-01 1.16854012e-01
-9.24848378e-01 -6.95828795e-01 -1.45925552e-01 3.70684534e-01
1.07922249e-01 3.34465921e-01 1.38776648e+00 -1.23163676e+00
4.89087969e-01 -7.28077829e-01 -7.54203554e-03 5.05572021e-01
2.62545377e-01 4.04541105e-01 -4.31942016e-01 4.87878919e-01
4.19209182e-01 -1.64461911e-01 3.41170430e-01 5.98358884e-02
-4.85572480e-02 -7.81858921e-01 3.59611422e-01 -6.88506126e-01
1.46453154e+00 1.42573625e-01 8.00849140e-01 5.89763165e-01
-4.70552556e-02 1.09908295e+00 1.13436127e+00 2.56831765e-01
2.31345221e-01 3.47905122e-02 5.63004076e-01 2.59210825e-01
3.03936094e-01 3.47761571e-01 5.87248921e-01 8.19148421e-01
1.92275614e-01 -4.28728789e-01 -7.82818019e-01 3.82858396e-01
-1.63332450e+00 -1.09447074e+00 -1.16221905e+00 1.87161529e+00
9.62049782e-01 4.36474144e-01 -1.04323924e-02 1.22502947e+00
6.51133716e-01 -1.74978986e-01 3.06274980e-01 -1.18587172e+00
-5.22628844e-01 1.62238866e-01 -1.58378065e-01 1.15404248e+00
-7.99900889e-01 8.48065019e-01 4.53909540e+00 5.95246494e-01
-1.17896330e+00 -1.65391639e-01 2.68125266e-01 2.57088035e-01
-2.90023744e-01 5.28245047e-02 -6.44437313e-01 8.27436522e-02
5.43553770e-01 -1.81975871e-01 -3.12724531e-01 4.25334275e-01
1.75981030e-01 -9.70492065e-01 -8.42767000e-01 5.09507239e-01
7.62987375e-01 -7.77580321e-01 5.29347658e-01 -4.83902454e-01
4.02909160e-01 -6.34062231e-01 -4.16728586e-01 3.87898646e-03
-6.46182120e-01 -8.19109499e-01 8.06512773e-01 3.55558097e-01
-1.63924813e-01 -1.08667231e+00 1.83551705e+00 2.77456403e-01
-8.48555982e-01 2.71721512e-01 -1.21728808e-01 -8.45215738e-01
1.38431013e-01 7.24426389e-01 -1.22064507e+00 7.80763268e-01
2.72593021e-01 3.64566118e-01 -6.56963229e-01 3.39986682e-01
-4.43419188e-01 3.32097203e-01 -5.19477904e-01 -6.35504603e-01
4.85072851e-01 -5.49815893e-01 6.02932870e-01 1.31806695e+00
1.89633369e-01 1.89586714e-01 -5.34877181e-01 3.19373637e-01
2.13674679e-01 9.44672346e-01 -8.45407009e-01 -1.21528514e-01
-2.06104405e-02 9.79669750e-01 -1.21941435e+00 -6.66126311e-01
7.10151717e-02 1.05654514e+00 -3.14485788e-01 -1.18839666e-02
-4.82017070e-01 -1.01904094e+00 -8.20572898e-02 8.14666152e-02
1.98531020e-02 2.94220239e-01 -4.08343613e-01 -7.20981300e-01
4.37600493e-01 -1.13267207e+00 4.16551083e-01 -7.94669867e-01
-8.98376405e-01 9.51227188e-01 -3.95844085e-03 -1.11564159e+00
-4.62884277e-01 -8.70429397e-01 -6.53700113e-01 7.33838618e-01
-1.23185790e+00 -6.96798980e-01 1.51189566e-01 4.25780773e-01
4.04594779e-01 1.77916393e-01 7.88637757e-01 -9.53202415e-03
-1.34340063e-01 -1.26827598e-01 -7.16049016e-01 -9.17236581e-02
8.03928196e-01 -1.23062706e+00 -6.93868041e-01 1.22086620e+00
-2.22500831e-01 6.03546977e-01 1.42959726e+00 -6.82064712e-01
-9.67442572e-01 -3.23182553e-01 1.84708440e+00 -3.10629576e-01
6.41585171e-01 2.34566733e-01 -6.53581560e-01 3.21375132e-01
9.23936963e-01 -8.05342078e-01 8.60843003e-01 -4.28259790e-01
-1.24141753e-01 3.93363647e-02 -1.44994605e+00 7.30661631e-01
-1.31333411e-01 -2.66313225e-01 -1.46059275e+00 3.95978332e-01
3.04100037e-01 4.93757240e-02 -5.00518382e-01 2.58518398e-01
4.65665787e-01 -1.13753164e+00 4.05905694e-01 -6.81735933e-01
8.47788215e-01 -6.65490448e-01 -2.61513323e-01 -8.94043922e-01
7.87715375e-01 3.37488614e-02 3.44519526e-01 1.13349128e+00
1.00044322e+00 -5.41595697e-01 3.40601653e-01 1.12280883e-01
1.27331868e-01 -7.89869130e-01 -5.91535747e-01 2.04404257e-02
-2.54499335e-02 -3.12935352e-01 3.90307814e-01 1.14082515e+00
7.20815539e-01 9.55499053e-01 7.48729631e-02 -1.52706131e-02
6.71806186e-02 4.00459498e-01 2.56095439e-01 -1.10618019e+00
3.04747820e-01 -5.42449951e-01 -5.66969037e-01 -3.09472322e-01
-4.69375998e-02 -7.52751946e-01 -1.34625643e-01 -1.92194736e+00
-4.96611148e-01 4.05338615e-01 5.46838105e-01 4.27460104e-01
4.61463332e-02 3.54138821e-01 8.88903290e-02 -2.73086756e-01
-3.98086488e-01 3.20993066e-02 9.53702807e-01 -1.39251113e-01
-5.41089512e-02 -1.72870323e-01 -6.57981634e-01 1.18854308e+00
1.02074575e+00 -5.09795010e-01 -4.42991436e-01 2.35293761e-01
1.33568752e+00 -2.53391325e-01 -2.20926732e-01 -7.70105660e-01
2.15871539e-02 -1.40188858e-01 1.13557041e-01 -9.38182950e-01
-1.85382321e-01 -1.07723427e+00 -3.91066998e-01 8.52496386e-01
-1.46111194e-03 6.97640538e-01 1.11720160e-01 -3.97825502e-02
-5.47099233e-01 -1.09907317e+00 3.31917912e-01 -3.02738100e-01
-3.83992612e-01 -9.22934175e-01 -1.26747155e+00 2.01859593e-01
1.11991608e+00 -4.14552033e-01 -2.61965096e-01 -4.76859629e-01
-5.29679954e-01 1.26832709e-01 1.71153061e-02 -3.88748348e-02
7.42927253e-01 -7.02168226e-01 -9.31159437e-01 -1.28971934e-01
7.13218078e-02 -3.17499787e-01 -1.16439357e-01 1.24656546e+00
-1.23270392e+00 1.38194934e-01 -4.21056569e-01 -1.44989431e-01
-1.94920468e+00 2.32103780e-01 3.63098443e-01 -1.89712107e-01
2.68292069e-01 3.77464116e-01 -9.03961301e-01 -2.71728396e-01
-2.59827077e-01 -7.22931266e-01 -1.14965665e+00 8.78673911e-01
5.37689209e-01 4.03980255e-01 4.93201524e-01 -1.05894113e+00
-5.52063465e-01 6.84975982e-01 5.08299321e-02 -3.66133720e-01
9.68273282e-01 -1.25414729e-01 -1.08339334e+00 5.68005562e-01
1.02049458e+00 5.40135920e-01 3.32330734e-01 4.78568286e-01
2.72909284e-01 3.30870934e-02 -3.21141303e-01 -1.04419088e+00
-1.25662908e-01 5.06652951e-01 1.52152061e-01 7.39529073e-01
9.91250277e-01 -2.16901749e-01 4.29617345e-01 8.61694813e-01
-1.87824532e-01 -1.36008036e+00 -2.23644435e-01 9.75980639e-01
1.15961242e+00 -1.41798544e+00 3.20542604e-01 -5.98831117e-01
-9.18980539e-01 1.75077474e+00 2.71702886e-01 -7.71752447e-02
5.73086202e-01 3.65198076e-01 2.02622637e-01 -5.49726188e-01
-4.32960510e-01 -1.13554433e-01 1.98841751e-01 2.26837397e-01
1.02253914e+00 1.11329257e-01 -1.61468410e+00 8.56532753e-01
-8.03133547e-01 -2.75262475e-01 1.12603068e+00 1.32250106e+00
-7.76286006e-01 -9.64800715e-01 -9.38224912e-01 4.26898450e-01
-7.69291639e-01 -2.19242796e-01 -1.22417581e+00 7.40508735e-01
3.64353240e-01 1.59614897e+00 -1.37636796e-01 -2.06380412e-01
3.38427693e-01 4.03241068e-01 6.43573284e-01 -1.47336990e-01
-1.01904607e+00 -2.70121872e-01 7.26164222e-01 4.45419885e-02
-7.45708108e-01 -3.64034474e-01 -1.77579832e+00 -4.11446601e-01
-4.64610457e-01 7.00181425e-01 9.32634115e-01 1.61009872e+00
-1.03293337e-01 3.44723731e-01 2.04340547e-01 -4.47436161e-02
-3.53427827e-01 -1.07136846e+00 -7.32378438e-02 5.82929075e-01
3.80404919e-01 -1.42705396e-01 -6.01617455e-01 2.73838073e-01] | [11.05753231048584, 6.929227828979492] |
7eb86337-92ff-4fb2-afce-e210f94b92e5 | unsupervised-learning-for-computational | 1612.08425 | null | http://arxiv.org/abs/1612.08425v2 | http://arxiv.org/pdf/1612.08425v2.pdf | Unsupervised Learning for Computational Phenotyping | With large volumes of health care data comes the research area of
computational phenotyping, making use of techniques such as machine learning to
describe illnesses and other clinical concepts from the data itself. The
"traditional" approach of using supervised learning relies on a domain expert,
and has two main limitations: requiring skilled humans to supply correct labels
limits its scalability and accuracy, and relying on existing clinical
descriptions limits the sorts of patterns that can be found. For instance, it
may fail to acknowledge that a disease treated as a single condition may really
have several subtypes with different phenotypes, as seems to be the case with
asthma and heart disease. Some recent papers cite successes instead using
unsupervised learning. This shows great potential for finding patterns in
Electronic Health Records that would otherwise be hidden and that can lead to
greater understanding of conditions and treatments. This work implements a
method derived strongly from Lasko et al., but implements it in Apache Spark
and Python and generalizes it to laboratory time-series data in MIMIC-III. It
is released as an open-source tool for exploration, analysis, and
visualization, available at https://github.com/Hodapp87/mimic3_phenotyping | ['Chris Hodapp'] | 2016-12-26 | null | null | null | null | ['computational-phenotyping'] | ['medical'] | [-1.51034847e-01 5.92141994e-04 -2.47425497e-01 -4.49807197e-01
-1.96266577e-01 -6.49090230e-01 7.62162879e-02 7.91775763e-01
4.14665043e-02 6.85098708e-01 2.13100448e-01 -7.81436682e-01
-5.04528880e-01 -7.03681111e-01 -1.60665169e-01 -7.54290283e-01
-3.53859782e-01 8.66054416e-01 -1.47282392e-01 1.13266118e-01
1.05617533e-03 4.41998929e-01 -1.48735833e+00 7.44938433e-01
5.82934022e-01 6.55942380e-01 -4.72618155e-02 5.78502297e-01
-3.78565490e-01 1.08616149e+00 -5.72379649e-01 7.72203207e-02
3.15112434e-02 -2.68729746e-01 -6.92317665e-01 -7.91501328e-02
-2.45070919e-01 1.74665153e-02 1.89259797e-01 4.12714452e-01
5.15743136e-01 -4.15719330e-01 4.27680224e-01 -1.33744752e+00
-3.03282946e-01 4.20817614e-01 -1.46630362e-01 1.36620626e-01
6.29515469e-01 4.40514207e-01 6.15709484e-01 -2.32513055e-01
8.00743818e-01 9.83420372e-01 9.69586611e-01 3.65962207e-01
-1.60518968e+00 -5.06659448e-01 -3.38136375e-01 3.13516349e-01
-1.41169059e+00 -9.06040445e-02 2.11199045e-01 -1.05527556e+00
1.09674406e+00 5.45629680e-01 6.87900960e-01 8.19561124e-01
3.87127586e-02 3.15580368e-01 1.28723514e+00 -3.28922540e-01
4.19572532e-01 4.81469005e-01 1.00811787e-01 3.96844655e-01
3.00832748e-01 1.39327005e-01 -2.88761824e-01 -7.93851316e-01
2.63377011e-01 6.59257412e-01 -1.67710900e-01 -2.15694502e-01
-1.23321199e+00 6.46067202e-01 -1.38807476e-01 4.19562101e-01
-5.54844260e-01 -5.64380705e-01 6.16502047e-01 4.93165463e-01
4.02371943e-01 5.36242962e-01 -8.17239702e-01 -3.50596160e-01
-9.92932856e-01 1.55263245e-01 1.13175285e+00 6.48455679e-01
6.79850459e-01 -5.01440108e-01 1.67804018e-01 5.92359543e-01
1.05354950e-01 1.48818582e-01 5.60693443e-01 -7.93985426e-01
-3.49624068e-01 1.15063667e+00 4.96083498e-02 -8.42843652e-01
-7.95840442e-01 -2.03340203e-02 -6.82445049e-01 2.77075291e-01
5.76327085e-01 -3.10150146e-01 -7.33855963e-01 1.26504886e+00
3.89972717e-01 -2.75932495e-02 -1.69754978e-02 5.89111209e-01
8.25101078e-01 2.42978439e-01 2.19316453e-01 -2.18261242e-01
1.27363288e+00 -3.10101360e-01 -6.61936224e-01 2.71566421e-01
1.00334477e+00 -7.07220078e-01 7.68853664e-01 7.93083966e-01
-8.41455698e-01 -9.49410424e-02 -5.58623612e-01 2.88411736e-01
-8.54669273e-01 -1.98077127e-01 7.41002679e-01 6.80475533e-01
-1.11506975e+00 6.47705317e-01 -1.02135372e+00 -9.15273488e-01
3.82454902e-01 2.10921064e-01 -1.87431306e-01 -1.74420074e-01
-9.48705673e-01 7.07708001e-01 4.35559571e-01 -3.63086194e-01
-2.44018167e-01 -1.06099737e+00 -4.26596344e-01 -1.24369450e-01
4.00085330e-01 -7.63854206e-01 9.25952375e-01 -5.96778631e-01
-7.63039887e-01 9.60585952e-01 -3.52455795e-01 -1.55358762e-01
3.93319130e-01 2.99179792e-01 -5.82394540e-01 3.64540815e-02
2.12427229e-02 1.37930334e-01 1.28249247e-02 -7.89576471e-01
-6.29273176e-01 -6.02519035e-01 -2.88439840e-01 -3.90763640e-01
6.99701253e-03 3.28504503e-01 1.85003772e-01 -3.70582044e-01
-8.77547041e-02 -7.34364927e-01 -2.55248755e-01 6.21586256e-02
-4.19520289e-01 -3.64487261e-01 5.88986039e-01 -7.52505362e-01
1.30370402e+00 -2.04323292e+00 -4.07060117e-01 2.86051989e-01
5.20681679e-01 3.03886145e-01 4.75020051e-01 1.03841245e+00
-3.70330930e-01 4.09085959e-01 -9.59866717e-02 2.51072139e-01
-1.87945276e-01 3.34149659e-01 -4.92877513e-02 3.36820662e-01
3.25541764e-01 5.41304410e-01 -8.81064057e-01 -4.28965747e-01
2.52568722e-01 5.97529948e-01 -4.39085841e-01 9.86644551e-02
-2.80338556e-01 6.60845518e-01 -3.23635012e-01 8.07845056e-01
5.55036604e-01 -8.70497584e-01 6.73221886e-01 3.23391944e-01
-3.04864824e-01 4.31640893e-01 -1.18365157e+00 1.06830037e+00
-6.30100742e-02 2.52747178e-01 -5.14719151e-02 -1.27240598e+00
9.03275549e-01 6.91825092e-01 1.02915668e+00 -1.25585243e-01
-2.41169661e-01 1.70529708e-01 1.27420247e-01 -9.89329576e-01
-5.55267572e-01 -5.38106747e-02 3.78066480e-01 7.71398425e-01
-4.21336919e-01 2.63304114e-01 2.35771120e-01 7.76119158e-02
1.68031800e+00 1.22855155e-04 5.74964345e-01 -3.07659894e-01
2.42576569e-01 7.18328357e-01 6.53895676e-01 4.58856463e-01
2.67894901e-02 2.78892100e-01 7.57767022e-01 -8.76612902e-01
-1.01372254e+00 -9.22492266e-01 -5.62955439e-01 8.42501223e-01
-6.39614224e-01 -6.83962762e-01 -1.83222860e-01 -3.11512023e-01
4.06233072e-01 4.16519731e-01 -6.26004398e-01 2.58869201e-01
-1.08770631e-01 -9.35055971e-01 4.32009965e-01 2.02202037e-01
-1.61625758e-01 -1.16716337e+00 -9.91135955e-01 3.44974935e-01
5.80182187e-02 -6.62935197e-01 3.74216765e-01 2.16284111e-01
-8.46751392e-01 -1.48152983e+00 -3.02509904e-01 -4.29748535e-01
4.61234212e-01 -2.86878437e-01 1.32944679e+00 4.24801856e-01
-1.10268104e+00 3.91145408e-01 -4.89307731e-01 -8.07247698e-01
-6.12296879e-01 -2.27801457e-01 -5.35779968e-02 -1.81355551e-01
1.05572534e+00 -6.82934225e-01 -6.79521561e-01 2.00399682e-01
-8.95362020e-01 -1.28067181e-01 4.57867295e-01 6.81207776e-01
5.69480836e-01 7.29401112e-02 6.87102139e-01 -1.17703080e+00
4.45390075e-01 -1.14493883e+00 -2.02630907e-01 3.17962080e-01
-1.09547555e+00 -2.48782888e-01 6.81343138e-01 -2.56145626e-01
-5.15240133e-01 1.17796995e-01 2.07985222e-01 -2.22527981e-01
-8.92306864e-01 5.95498562e-01 2.61979759e-01 6.12394989e-01
7.48215854e-01 2.16061324e-01 4.87527937e-01 -8.31286430e-01
-1.10947326e-01 9.71480608e-01 -6.89088330e-02 -2.35150546e-01
1.50133297e-01 6.71599030e-01 -1.48401350e-01 -7.95346200e-01
-3.10093403e-01 -8.15287292e-01 -4.74996835e-01 1.24536768e-01
6.51105762e-01 -6.44982994e-01 -1.03156769e+00 4.07907888e-02
-7.48646259e-01 -3.71164292e-01 -3.82777095e-01 4.81767446e-01
-2.86167026e-01 6.94666877e-02 -4.55277711e-01 -6.80230439e-01
-1.83121443e-01 -8.02335203e-01 6.95943534e-01 -1.08758695e-01
-8.40555489e-01 -1.27247190e+00 4.95115221e-01 1.62663534e-01
6.27353966e-01 5.87841749e-01 1.17464018e+00 -1.10856545e+00
-6.74507767e-02 -2.62701750e-01 -4.31423523e-02 6.98682293e-02
5.07211268e-01 3.30182701e-01 -8.49060833e-01 -1.07303798e-01
5.95195182e-02 -2.35919759e-01 2.94453442e-01 3.45726818e-01
1.18367517e+00 -4.43184376e-01 -6.48981214e-01 4.28280234e-01
1.44817805e+00 5.66694558e-01 4.87892002e-01 2.19117522e-01
3.30523133e-01 1.00500262e+00 2.83951133e-01 8.22384119e-01
3.71611923e-01 4.40404952e-01 -9.15308222e-02 -3.50571394e-01
1.98376015e-01 2.15639085e-01 -1.76228479e-01 4.21209037e-01
-1.23740546e-01 2.41907388e-01 -1.60629869e+00 4.49596077e-01
-1.80717492e+00 -9.94849801e-01 -4.47887719e-01 2.23199701e+00
1.08701873e+00 -3.78874898e-01 4.77692813e-01 1.39171928e-01
5.40335536e-01 -5.80510974e-01 -6.25863492e-01 -7.24669635e-01
-1.91277533e-04 2.58617133e-01 1.54986903e-01 -3.32015082e-02
-6.83809817e-01 2.02550814e-01 6.72226191e+00 1.82604268e-01
-1.22048998e+00 -2.06006039e-02 7.49061346e-01 -2.36722648e-01
2.99625024e-02 -6.85214624e-02 -2.48634920e-01 6.38890266e-01
1.24336863e+00 -3.31323862e-01 3.88603359e-01 6.74120605e-01
6.60311878e-01 -1.90851793e-01 -1.25864005e+00 7.44882643e-01
-3.89619529e-01 -1.34867060e+00 -5.07397950e-01 1.78882524e-01
3.78271192e-01 2.45752051e-01 -3.06350648e-01 -1.14211209e-01
4.62153703e-01 -1.25424492e+00 -4.91422080e-02 7.57817924e-01
6.87948287e-01 -3.32862198e-01 6.89727664e-01 2.66805619e-01
-7.38683641e-01 -1.60307676e-01 -1.22334987e-01 -3.95054966e-01
-1.52084827e-01 8.93010616e-01 -1.48234248e+00 3.72934222e-01
9.99916375e-01 7.01093435e-01 -5.09089470e-01 1.22763610e+00
2.14840367e-01 8.18428338e-01 -3.79984826e-01 4.15043198e-02
-2.42652923e-01 -5.43843806e-02 2.82894552e-01 1.41992438e+00
1.11220218e-01 1.69883758e-01 4.45007920e-01 7.38358259e-01
7.52222121e-01 4.46352988e-01 -8.46620619e-01 -3.98344785e-01
4.49461609e-01 1.08213508e+00 -6.20177567e-01 -6.80661142e-01
-5.79166174e-01 2.73834676e-01 -1.41234666e-01 2.56829351e-01
-3.77453953e-01 -2.08769575e-01 6.10818624e-01 7.02687323e-01
1.83007762e-01 1.78157285e-01 -4.96901542e-01 -8.14770281e-01
-1.97180942e-01 -1.23564446e+00 9.31285858e-01 -5.99579573e-01
-1.44204342e+00 4.12253410e-01 -6.79395497e-02 -1.09949195e+00
-5.11149883e-01 -5.40562332e-01 -4.64278400e-01 8.57575119e-01
-9.79344964e-01 -6.68082237e-01 -2.39267737e-01 6.60007834e-01
-1.47961909e-02 -7.33737135e-03 1.27133977e+00 3.19537818e-01
-4.57441807e-01 1.19954929e-01 3.27274710e-01 -6.44947365e-02
9.26262021e-01 -1.26422727e+00 -1.25312060e-01 6.96065873e-02
-3.49509239e-01 7.17019737e-01 6.88455820e-01 -7.78406501e-01
-1.18494689e+00 -8.88095856e-01 1.12968099e+00 -4.58478272e-01
7.53679216e-01 -6.88077882e-02 -1.22278619e+00 6.00760758e-01
1.03993952e-01 -2.35563993e-01 1.28508675e+00 2.84574419e-01
-4.28943545e-01 2.20380705e-02 -1.36623657e+00 2.35482424e-01
6.68834209e-01 -3.18955243e-01 -3.78701746e-01 7.65849590e-01
1.06814235e-01 2.32730463e-01 -1.29600632e+00 3.10880750e-01
5.29378951e-01 -1.11549211e+00 6.46244824e-01 -1.03019106e+00
2.59543151e-01 -1.91858485e-01 2.35895202e-01 -1.08207238e+00
-1.40010759e-01 -6.77905679e-01 1.96939766e-01 8.95993471e-01
6.06504321e-01 -1.03863943e+00 6.37480855e-01 9.43419814e-01
2.96860576e-01 -9.93710577e-01 -3.88287872e-01 -7.73720562e-01
-1.51637811e-02 -3.32182318e-01 8.43326271e-01 1.60172355e+00
6.54520094e-01 -8.93067420e-02 1.82504624e-01 -4.77245003e-02
4.06814188e-01 2.45452732e-01 6.75856888e-01 -1.54971349e+00
-4.58795905e-01 -3.32003087e-01 -5.70906818e-01 1.27615752e-02
-5.02490997e-01 -9.22091544e-01 -3.79750401e-01 -1.61858594e+00
2.93892294e-01 -9.14484084e-01 -2.49838427e-01 1.03307152e+00
1.62537009e-01 1.83906022e-03 -2.71867067e-01 4.67574894e-01
-2.32293576e-01 -5.95300853e-01 5.90443611e-01 1.11820482e-01
-4.39561725e-01 3.64847332e-02 -6.25129223e-01 6.12838686e-01
1.28376830e+00 -7.64688551e-01 -1.76219702e-01 1.13409251e-01
3.67530107e-01 9.80227813e-02 4.72072840e-01 -6.84308171e-01
2.39438534e-01 -3.65774751e-01 2.76593894e-01 -3.48236233e-01
-1.58130422e-01 -7.72948802e-01 9.06963825e-01 9.25611675e-01
-2.30139688e-01 2.64336467e-01 2.07669973e-01 2.75570720e-01
1.74720921e-02 1.06694691e-01 3.49106103e-01 -5.49410403e-01
-4.12208140e-01 -9.88461152e-02 -5.55307806e-01 -2.86879018e-02
1.13927329e+00 -1.75606638e-01 -4.88037020e-01 -2.95827419e-01
-1.20809102e+00 3.15816581e-01 7.31881261e-01 1.75980717e-01
2.19035611e-01 -7.82838345e-01 -9.51332152e-01 2.44756758e-01
3.04997712e-01 -1.55570850e-01 1.91417247e-01 1.37302923e+00
-7.71248281e-01 5.66505551e-01 -1.06021509e-01 -7.89400220e-01
-1.32779348e+00 1.03006673e+00 1.19799092e-01 -1.81761593e-01
-9.68742549e-01 1.65112451e-01 -1.04686087e-02 -3.59221458e-01
1.58164874e-01 -1.75428256e-01 -6.72077909e-02 6.25473335e-02
7.76356339e-01 5.22223115e-01 2.05250293e-01 -1.41105324e-01
-7.45703399e-01 1.39352962e-01 -2.96093361e-03 2.75778651e-01
1.58064568e+00 8.81056190e-02 -4.85802889e-01 7.17296004e-01
9.99942422e-01 -9.73540172e-02 -6.04663074e-01 -1.13647714e-01
3.17896515e-01 -2.17704579e-01 -3.76359552e-01 -1.36147320e+00
-7.31720865e-01 8.29584122e-01 6.52392626e-01 8.00297558e-01
1.16247892e+00 3.51777971e-01 2.44375885e-01 2.90875196e-01
1.97472200e-01 -7.54148185e-01 -5.91234386e-01 -1.33269295e-01
5.34426153e-01 -1.07896054e+00 3.64913717e-02 -2.51972347e-01
-6.16503000e-01 1.21927106e+00 5.53479306e-02 1.59123495e-01
8.63010705e-01 5.94531178e-01 4.82387155e-01 -5.23546040e-01
-1.22414637e+00 -6.48804680e-02 -1.40997514e-01 7.53426313e-01
7.18446016e-01 4.83576775e-01 -5.38748562e-01 4.29623604e-01
2.17547826e-02 6.46172285e-01 4.55563158e-01 1.26747930e+00
-8.63709673e-02 -1.40609074e+00 -5.61194122e-01 8.70501459e-01
-7.36963093e-01 -1.19736716e-01 -5.44282556e-01 7.00102627e-01
5.51378548e-01 9.94826317e-01 2.66263455e-01 -2.23290354e-01
1.85974553e-01 6.20195508e-01 2.16113515e-02 -7.73504853e-01
-7.26456761e-01 1.07273705e-01 7.01676235e-02 -6.70599699e-01
-4.75769103e-01 -8.83465827e-01 -1.24643183e+00 -3.74089003e-01
3.49499047e-01 2.64396787e-01 6.46501303e-01 6.18339479e-01
9.38747287e-01 3.85152578e-01 4.31837320e-01 -5.32425679e-02
-3.06265563e-01 -7.62846172e-01 -6.46499693e-01 5.76046705e-01
2.64964223e-01 -4.51573402e-01 -4.28176612e-01 3.46715748e-01] | [7.956182956695557, 6.127477645874023] |
e2ddaec7-a974-4eff-9709-07c1a48b53ea | danzero-mastering-guandan-game-with | 2210.17087 | null | https://arxiv.org/abs/2210.17087v1 | https://arxiv.org/pdf/2210.17087v1.pdf | DanZero: Mastering GuanDan Game with Reinforcement Learning | Card game AI has always been a hot topic in the research of artificial intelligence. In recent years, complex card games such as Mahjong, DouDizhu and Texas Hold'em have been solved and the corresponding AI programs have reached the level of human experts. In this paper, we are devoted to developing an AI program for a more complex card game, GuanDan, whose rules are similar to DouDizhu but much more complicated. To be specific, the characteristics of large state and action space, long length of one episode and the unsure number of players in the GuanDan pose great challenges for the development of the AI program. To address these issues, we propose the first AI program DanZero for GuanDan using reinforcement learning technique. Specifically, we utilize a distributed framework to train our AI system. In the actor processes, we carefully design the state features and agents generate samples by self-play. In the learner process, the model is updated by Deep Monte-Carlo Method. After training for 30 days using 160 CPUs and 1 GPU, we get our DanZero bot. We compare it with 8 baseline AI programs which are based on heuristic rules and the results reveal the outstanding performance of DanZero. We also test DanZero with human players and demonstrate its human-level performance. | ['Houqiang Li', 'Wengang Zhou', 'Youpeng Zhao', 'Jian Zhao', 'Yudong Lu'] | 2022-10-31 | null | null | null | null | ['card-games'] | ['playing-games'] | [-5.14950037e-01 -2.46170074e-01 2.73845568e-02 1.81284919e-01
-2.14994073e-01 -5.85611761e-01 5.69198072e-01 -4.55276668e-01
-4.79145229e-01 9.49545026e-01 -2.43672878e-01 -2.78397292e-01
4.45021354e-02 -1.19309330e+00 -4.04502153e-01 -6.75890625e-01
-1.26815215e-01 1.00665677e+00 5.88644266e-01 -6.53248250e-01
4.19984102e-01 -4.93001044e-02 -1.49301934e+00 6.15899973e-02
1.04535353e+00 5.76232612e-01 -1.40810292e-02 8.83534491e-01
3.95984128e-02 1.38000286e+00 -1.02622378e+00 -3.69723022e-01
4.11503941e-01 -9.47438478e-01 -7.33139873e-01 -1.65133134e-01
-4.66297776e-01 -5.78443289e-01 -6.29917085e-01 1.04076707e+00
6.08485818e-01 1.98120937e-01 4.66812402e-01 -1.51440060e+00
-2.94568598e-01 1.09923971e+00 -6.31241918e-01 3.80498357e-02
2.31165871e-01 7.52783477e-01 7.61721015e-01 -1.74229085e-01
5.68413973e-01 1.19207382e+00 4.51096207e-01 7.32148945e-01
-5.18489122e-01 -8.11824918e-01 1.75720587e-01 3.53420019e-01
-1.12331402e+00 1.88027043e-02 4.22740519e-01 -2.08141357e-01
9.65589404e-01 -7.73490444e-02 1.09730971e+00 8.75476301e-01
4.48650658e-01 1.09337664e+00 1.02045763e+00 -1.73463821e-01
6.91197157e-01 -3.67010772e-01 -1.16515614e-01 7.68657386e-01
1.05773665e-01 4.75217700e-01 -2.42980700e-02 -3.37596387e-01
1.13129175e+00 2.29857527e-02 4.70378399e-01 -2.79592238e-02
-9.97520745e-01 1.05277193e+00 2.33570993e-01 1.62355304e-01
-4.28914636e-01 6.25545859e-01 6.58916891e-01 2.29139954e-01
-4.85424548e-02 5.05761564e-01 -1.87746048e-01 -6.55869901e-01
-4.93554771e-01 8.81064117e-01 1.12090373e+00 8.44596446e-01
3.34451139e-01 4.14076746e-01 -2.07894862e-01 4.28173870e-01
-1.46371461e-02 4.07715827e-01 5.19147635e-01 -1.23450887e+00
2.60720015e-01 5.39370477e-01 2.23726541e-01 -7.27015197e-01
-2.95926154e-01 -2.27392986e-01 -9.99820828e-01 4.92936045e-01
3.55898499e-01 -7.04857051e-01 -6.14241481e-01 1.40729952e+00
3.28675479e-01 5.59821308e-01 2.59065628e-01 9.41655397e-01
5.22965670e-01 1.19734895e+00 -2.75311759e-03 -1.29486516e-01
1.33788431e+00 -1.37463081e+00 -5.04864454e-01 -1.74957201e-01
5.58240891e-01 -2.09793091e-01 9.42919731e-01 8.20912242e-01
-1.27166808e+00 -3.68710607e-01 -9.18776453e-01 5.94972432e-01
-7.30991438e-02 9.19879507e-03 1.23533607e+00 6.63390696e-01
-6.63957000e-01 5.63261867e-01 -9.94266391e-01 -9.53525528e-02
1.77703694e-01 4.70461428e-01 2.36179903e-01 8.73264447e-02
-1.33854699e+00 7.71291792e-01 8.42654943e-01 -6.10542600e-04
-1.15352130e+00 -7.95550738e-03 -4.31662560e-01 2.38338366e-01
8.68409455e-01 -5.15423596e-01 1.64856303e+00 -9.56058741e-01
-2.33765721e+00 3.70485216e-01 5.75843632e-01 -3.80396813e-01
6.96817935e-01 1.01571724e-01 -1.83380306e-01 -8.62081721e-02
-2.58801915e-02 3.15013736e-01 5.12805343e-01 -7.55762398e-01
-1.00638819e+00 1.00017555e-01 4.76779133e-01 3.96825999e-01
2.65734017e-01 2.75322229e-01 -4.00726885e-01 -4.75587249e-01
-4.34481114e-01 -1.08281696e+00 -6.15832746e-01 -7.00890481e-01
-3.99939232e-02 -5.94322383e-01 4.15416926e-01 -7.94109851e-02
1.27136159e+00 -1.92383254e+00 1.41185522e-01 1.35932639e-01
2.35288799e-01 4.05634075e-01 -1.05931520e-01 6.50040209e-01
4.40139502e-01 -7.65916333e-02 -3.65211070e-02 3.13913375e-01
2.22843185e-01 4.16587651e-01 -1.26303896e-01 1.47941545e-01
-1.25676379e-01 9.31772709e-01 -1.22457385e+00 -5.70776761e-01
-6.44509196e-02 -4.28195864e-01 -1.01000309e+00 4.14474368e-01
-4.69383568e-01 2.52902418e-01 -7.31465876e-01 4.40563828e-01
5.13267875e-01 -1.31740272e-01 1.10690027e-01 5.45976460e-01
-2.67183661e-01 -1.00413680e-01 -1.24027753e+00 1.42894983e+00
-1.90750793e-01 2.92847790e-02 -1.42653093e-01 -9.29349720e-01
8.34491014e-01 3.10605228e-01 3.64926249e-01 -7.31170058e-01
5.30634522e-01 2.50848681e-01 5.45059144e-01 -5.01508415e-01
4.87613916e-01 -7.35001266e-02 -5.43981254e-01 8.35359752e-01
-2.66822666e-01 -6.78041279e-01 7.22509027e-01 1.85941190e-01
1.26414621e+00 2.75187224e-01 3.66664708e-01 -7.01262578e-02
4.05183882e-01 5.32411635e-01 8.44920456e-01 1.19950068e+00
-2.96343625e-01 1.57469690e-01 9.69668865e-01 -6.46777093e-01
-1.14983284e+00 -7.28544235e-01 5.58053493e-01 1.17099571e+00
4.33493614e-01 -3.96719337e-01 -9.14298356e-01 -4.91177469e-01
-3.59227687e-01 6.55260324e-01 -4.71571237e-01 -3.50117326e-01
-8.85070682e-01 -7.99541891e-01 7.91844487e-01 5.51110566e-01
1.11025131e+00 -1.73988366e+00 -9.76876736e-01 6.54656231e-01
2.50373006e-01 -6.93005145e-01 -3.74591142e-01 -1.17787495e-01
-5.17959356e-01 -1.00059986e+00 -3.89962524e-01 -8.31909060e-01
2.02631146e-01 1.37828356e-02 1.14136803e+00 2.62276411e-01
-2.04655662e-01 -2.17841536e-01 -5.34782231e-01 -4.95509863e-01
-3.72412503e-01 3.31046820e-01 -1.57990903e-02 -5.83309531e-01
1.39628038e-01 -5.54332435e-01 -5.86217582e-01 4.31961894e-01
-8.53463173e-01 3.29414278e-01 6.56889975e-01 1.11014295e+00
8.53883326e-02 2.88009018e-01 5.55926979e-01 -1.03887296e+00
1.01823866e+00 -5.82843840e-01 -1.05326235e+00 2.10829321e-02
-1.78643212e-01 -3.42277810e-02 1.04721963e+00 -7.78790653e-01
-9.61007833e-01 -1.06689177e-01 -8.43159631e-02 -9.22634676e-02
9.80238095e-02 5.07809818e-01 2.84002852e-02 2.65683502e-01
5.61703980e-01 2.58441985e-01 -1.00733386e-02 2.00302407e-01
1.09511361e-01 6.09050095e-01 4.10129100e-01 -1.08936155e+00
5.44035792e-01 -3.62162553e-02 -3.02502483e-01 -1.82423547e-01
-3.91488016e-01 2.01091051e-01 2.05700368e-01 -2.18778715e-01
6.63920462e-01 -5.93503058e-01 -1.67147970e+00 1.00278080e+00
-1.10860217e+00 -7.28856206e-01 -2.51592785e-01 3.46836060e-01
-7.01763570e-01 -1.21540174e-01 -1.11345339e+00 -7.74003386e-01
-2.91897953e-01 -1.26680350e+00 5.09811938e-01 6.92098677e-01
8.06451514e-02 -5.76697409e-01 6.32805645e-01 6.93043247e-02
4.45485324e-01 1.13246329e-01 7.12243259e-01 -5.00993967e-01
-6.19223833e-01 -1.88312769e-01 1.39636537e-02 -9.23350379e-02
-1.86687976e-01 2.41930053e-01 -3.50188524e-01 -1.88106716e-01
-1.37957379e-01 -4.85753626e-01 1.82366252e-01 2.92436212e-01
9.96743798e-01 -3.41105461e-01 -3.29644009e-02 4.19043243e-01
1.31040442e+00 1.06419551e+00 7.57658422e-01 6.03788972e-01
3.36648583e-01 1.76164210e-01 7.84618378e-01 7.81150460e-01
4.50215846e-01 4.13254052e-01 4.41392928e-01 2.52614021e-01
6.28311932e-01 -3.34420770e-01 4.33013320e-01 8.91392469e-01
-5.82301974e-01 -3.82714689e-01 -9.02985692e-01 2.22442970e-01
-2.30511880e+00 -1.31604445e+00 1.37247771e-01 1.61204112e+00
7.10612237e-01 4.84831959e-01 5.35953701e-01 -1.01038858e-01
5.83256304e-01 5.15934825e-02 -8.30980420e-01 -6.84554517e-01
6.76286519e-02 2.86036730e-01 3.68884683e-01 3.56238931e-01
-7.81242728e-01 1.59125185e+00 6.73849344e+00 1.09480250e+00
-8.86079848e-01 -5.76923974e-02 6.23034000e-01 -1.33215450e-02
1.17250696e-01 3.30164522e-01 -5.75544417e-01 7.58557498e-01
7.76237428e-01 -4.60399508e-01 8.42556477e-01 9.42070007e-01
2.38242522e-01 -4.44996119e-01 -7.04815507e-01 1.01756918e+00
-2.50812501e-01 -1.35351872e+00 -2.37540185e-01 -2.44954359e-02
9.35498118e-01 -2.41520390e-01 -1.82165712e-01 1.02363682e+00
1.27169812e+00 -1.19013536e+00 6.77528739e-01 2.89611191e-01
2.46782914e-01 -1.21331406e+00 9.24547970e-01 7.67830431e-01
-1.14479661e+00 -2.71976471e-01 -6.19558811e-01 -6.89445674e-01
1.41585050e-02 -1.83523551e-01 -4.74779874e-01 3.63521963e-01
5.28655410e-01 4.17467475e-01 -7.72450194e-02 1.05709755e+00
-2.91731626e-01 6.81154549e-01 -4.08108056e-01 -6.55491233e-01
7.38196194e-01 -7.26948798e-01 2.00391397e-01 5.24716675e-01
2.69044727e-01 8.41818511e-01 4.67700660e-01 8.34048271e-01
1.73885375e-01 -4.29411121e-02 -5.11494994e-01 -4.56047095e-02
4.94403332e-01 1.04468858e+00 -7.44869530e-01 -5.68656266e-01
-1.68522730e-01 8.62850726e-01 2.45922193e-01 5.64722121e-02
-1.29785526e+00 -5.60751200e-01 4.40991104e-01 -2.19574615e-01
6.13572523e-02 -2.59270787e-01 -5.11387922e-02 -1.28627396e+00
-4.13948029e-01 -1.44840288e+00 2.49813303e-01 -6.81567729e-01
-9.76577818e-01 7.43325591e-01 -4.37316447e-02 -1.06480134e+00
-6.71539068e-01 -3.06871474e-01 -1.27107382e+00 4.55462158e-01
-7.81019211e-01 -6.80679977e-01 -2.58178949e-01 6.16966009e-01
6.61902726e-01 -6.07550740e-01 5.65155029e-01 1.00515477e-01
-9.83450294e-01 3.58183295e-01 3.25094432e-01 2.32650653e-01
1.74970597e-01 -1.03417301e+00 6.60845935e-01 6.37493253e-01
-3.81733239e-01 2.48222724e-01 7.54072487e-01 -6.23179972e-01
-1.61633813e+00 -5.62201440e-01 1.51259163e-02 -9.76196676e-02
8.13628376e-01 -2.12922707e-01 -5.37709892e-01 6.01942301e-01
3.58560860e-01 -2.20479831e-01 8.66727754e-02 -2.51404047e-01
3.31144094e-01 3.81094925e-02 -9.38582420e-01 8.82930815e-01
1.16169727e+00 1.01835810e-01 -4.66861725e-01 1.55654728e-01
4.58091587e-01 -6.58707142e-01 -3.53888929e-01 -2.03336045e-01
2.87542850e-01 -1.05754685e+00 5.28450310e-01 -7.72773445e-01
5.76259255e-01 -3.72069150e-01 4.01574016e-01 -1.44546103e+00
-3.75857800e-01 -9.37366307e-01 6.90328777e-02 9.76555407e-01
-1.20792389e-01 -8.22438657e-01 1.19684947e+00 4.28307831e-01
2.31338441e-02 -6.75370097e-01 -5.72001576e-01 -7.25555182e-01
2.92283267e-01 -3.39132100e-02 9.81962860e-01 6.32074237e-01
3.33810806e-01 3.30352545e-01 -8.01999807e-01 -2.03287214e-01
3.98977250e-01 3.10416222e-01 1.04881060e+00 -9.29581583e-01
-8.06609511e-01 -5.29747546e-01 -1.41921729e-01 -1.12280059e+00
1.55053750e-01 -3.73669028e-01 1.55002713e-01 -1.29244006e+00
2.39093617e-01 -6.79572880e-01 1.42396823e-01 5.11912942e-01
-1.99421152e-01 -2.23020285e-01 3.43460828e-01 2.14333415e-01
-9.01686192e-01 6.71335340e-01 1.45204592e+00 -1.84867401e-02
-3.57013285e-01 1.42387852e-01 -5.01198649e-01 6.56394958e-01
9.90500450e-01 -5.12113869e-01 -3.74728858e-01 -4.33674157e-01
5.43367267e-01 6.13745093e-01 -1.06787995e-01 -1.17462993e+00
4.40938503e-01 -7.78770506e-01 -2.23303363e-01 -3.55977833e-01
1.18505247e-01 -5.60237288e-01 2.75790811e-01 1.03270257e+00
8.95423889e-02 4.84640539e-01 -6.56480417e-02 8.45663697e-02
-3.36414874e-01 -4.13787842e-01 5.83705127e-01 -4.59984243e-01
-7.93977678e-01 5.26160419e-01 -1.01349390e+00 2.75899112e-01
1.44952846e+00 -1.87571764e-01 -3.87273103e-01 -5.79151392e-01
-2.70238876e-01 7.69668162e-01 3.38460207e-01 1.44470036e-02
4.18064773e-01 -1.19866347e+00 -8.83656383e-01 -1.42182745e-02
-3.25313598e-01 1.31264627e-01 3.37676704e-01 3.75626743e-01
-1.33928430e+00 -4.97592799e-02 -7.83035457e-01 -4.35469449e-02
-8.98608506e-01 5.68299711e-01 5.14892042e-01 -7.09482014e-01
-5.85030437e-01 3.90694916e-01 1.72622070e-01 -5.41478276e-01
-7.77503699e-02 -1.70194641e-01 -2.63071865e-01 -2.86360651e-01
4.23729092e-01 4.25878793e-01 -6.25897050e-01 7.23498538e-02
-2.01760501e-01 2.37617776e-01 -1.60319790e-01 -2.15275005e-01
1.53900468e+00 4.93199557e-01 -8.76004621e-02 5.42102121e-02
3.02078664e-01 -2.19717428e-01 -1.34436154e+00 1.01624556e-01
-1.59214213e-01 -3.06296170e-01 -4.17209387e-01 -4.66295093e-01
-1.32030594e+00 6.82270050e-01 3.27796280e-01 3.17263693e-01
8.40994835e-01 -3.48529249e-01 8.58643055e-01 5.64554632e-01
7.30032444e-01 -1.24906528e+00 1.91072583e-01 1.17968965e+00
3.81159157e-01 -9.76643503e-01 -1.62603289e-01 1.54838124e-02
-1.08022594e+00 9.07370687e-01 1.16095603e+00 -9.22288477e-01
1.38660789e-01 4.84618872e-01 -5.48131391e-02 -9.38724577e-02
-1.09417844e+00 -1.12777621e-01 -7.06438303e-01 4.59360033e-01
-8.54853094e-02 1.55164212e-01 -5.48754215e-01 9.10431564e-01
-4.15491909e-01 2.68271625e-01 8.24666679e-01 1.25078487e+00
-4.51856852e-01 -1.16960967e+00 -4.21799004e-01 2.44013742e-01
-3.43095273e-01 9.81294438e-02 -1.26727134e-01 8.80891025e-01
1.10666744e-01 6.52078807e-01 1.81612313e-01 -2.14361310e-01
3.08436066e-01 -4.57854331e-01 4.79142666e-01 -3.58605385e-01
-1.10980892e+00 -1.02244869e-01 1.11061119e-01 -4.25157964e-01
8.44554156e-02 -3.77452075e-01 -1.46859026e+00 -1.04240036e+00
-1.00132309e-01 6.08935118e-01 6.55720308e-02 8.72725487e-01
3.68838534e-02 5.71676373e-01 5.52488029e-01 -9.04189765e-01
-6.70305789e-01 -7.17740953e-01 -7.58770823e-01 1.74258411e-01
-4.99545485e-01 -5.61946750e-01 4.25389335e-02 -4.21618432e-01] | [3.5773870944976807, 1.5423684120178223] |
2e568929-62c6-44cc-91f0-08baa9696f4b | protein-design-with-guided-discrete-diffusion | 2305.20009 | null | https://arxiv.org/abs/2305.20009v1 | https://arxiv.org/pdf/2305.20009v1.pdf | Protein Design with Guided Discrete Diffusion | A popular approach to protein design is to combine a generative model with a discriminative model for conditional sampling. The generative model samples plausible sequences while the discriminative model guides a search for sequences with high fitness. Given its broad success in conditional sampling, classifier-guided diffusion modeling is a promising foundation for protein design, leading many to develop guided diffusion models for structure with inverse folding to recover sequences. In this work, we propose diffusioN Optimized Sampling (NOS), a guidance method for discrete diffusion models that follows gradients in the hidden states of the denoising network. NOS makes it possible to perform design directly in sequence space, circumventing significant limitations of structure-based methods, including scarce data and challenging inverse design. Moreover, we use NOS to generalize LaMBO, a Bayesian optimization procedure for sequence design that facilitates multiple objectives and edit-based constraints. The resulting method, LaMBO-2, enables discrete diffusions and stronger performance with limited edits through a novel application of saliency maps. We apply LaMBO-2 to a real-world protein design task, optimizing antibodies for higher expression yield and binding affinity to a therapeutic target under locality and liability constraints, with 97% expression rate and 25% binding rate in exploratory in vitro experiments. | ['Andrew Gordon Wilson', 'Kyunghyun Cho', 'Arvind Rajpal', 'Julien Lafrance-Vanasse', 'Isidro Hotzel', 'Tim G. J. Rudner', 'Nathan C. Frey', 'Samuel Stanton', 'Nate Gruver'] | 2023-05-31 | null | null | null | null | ['protein-design', 'bayesian-optimization'] | ['medical', 'methodology'] | [ 5.84533453e-01 2.02667713e-01 -2.87679315e-01 -3.21824670e-01
-5.34714997e-01 -4.26378280e-01 3.42705369e-01 -7.46714249e-02
-3.59606504e-01 9.93968010e-01 3.19148362e-01 -2.67306447e-01
-2.31232911e-01 -4.64384556e-01 -9.16002989e-01 -8.69751453e-01
1.68891564e-01 4.49312121e-01 1.80449076e-02 -2.34785318e-01
4.25468534e-01 6.61542833e-01 -1.05807221e+00 3.42594743e-01
1.15557313e+00 4.69645917e-01 9.07476783e-01 3.67719203e-01
1.02746427e-01 6.75297439e-01 -3.71026486e-01 -2.53718138e-01
-1.44704178e-01 -8.86031866e-01 -6.73836827e-01 -1.38726130e-01
8.23330060e-02 1.75249728e-03 -7.00215101e-02 1.07594943e+00
8.70466948e-01 2.47712731e-01 8.87443721e-01 -6.12240791e-01
-1.30250740e+00 4.39137936e-01 -4.46729541e-01 1.25951171e-01
2.54328161e-01 5.38247883e-01 1.24068701e+00 -1.09784603e+00
1.09447789e+00 1.27642322e+00 5.12604773e-01 8.77815187e-01
-2.00623059e+00 -1.70517847e-01 2.17298701e-01 2.86728293e-01
-1.24286711e+00 -3.81239265e-01 6.89546406e-01 -5.14143825e-01
1.33716106e+00 9.30547267e-02 7.42915154e-01 1.41374683e+00
4.73662704e-01 9.79795992e-01 8.93490136e-01 -1.73132598e-01
8.28875005e-01 -6.64983168e-02 -9.81181934e-02 8.24167788e-01
-2.01541439e-01 3.20571512e-01 -9.91411746e-01 -4.65618402e-01
6.56194210e-01 1.31760150e-01 -3.96154851e-01 -6.94009602e-01
-1.02235186e+00 1.15106428e+00 5.11732459e-01 -1.04409114e-01
-8.40092778e-01 9.77268070e-02 -2.39161730e-01 -7.38582686e-02
3.92395407e-01 8.55342984e-01 -4.04458106e-01 -2.42722780e-02
-1.10609794e+00 5.35306334e-01 5.75693250e-01 8.89699996e-01
6.61813498e-01 6.74052164e-02 -3.98328245e-01 7.84838438e-01
4.56007153e-01 2.20399454e-01 1.14506580e-01 -9.54852641e-01
-2.38651007e-01 2.35122383e-01 1.16638742e-01 -7.23686576e-01
-4.06061381e-01 -6.68286264e-01 -6.13536000e-01 3.17920804e-01
1.57575995e-01 -8.84120632e-03 -9.38128650e-01 2.12312055e+00
3.42682958e-01 -2.68092990e-01 -1.44043058e-01 1.10004699e+00
2.64659286e-01 6.24231339e-01 1.61852702e-01 -3.82117748e-01
1.16581511e+00 -7.76785731e-01 -6.52868629e-01 -1.15504161e-01
4.99840885e-01 -4.97663468e-01 1.18357551e+00 6.89719498e-01
-9.56906319e-01 -1.43494621e-01 -1.11861229e+00 -2.85703652e-02
-8.07032734e-02 3.38169262e-02 7.09094405e-01 4.63830203e-01
-9.47774470e-01 9.69782233e-01 -7.99318790e-01 -2.71963477e-01
7.68966973e-01 4.55967814e-01 8.84001628e-02 -8.72892141e-02
-1.13573742e+00 1.13822043e+00 1.37482941e-01 4.64650616e-02
-1.38967109e+00 -1.00777233e+00 -4.78938669e-01 -1.34732649e-01
3.46763194e-01 -9.87846851e-01 7.40253985e-01 -7.70966828e-01
-1.71769691e+00 4.24461633e-01 -4.82096016e-01 -6.23855770e-01
3.70662063e-01 -2.99995661e-01 1.56568199e-01 -9.71761793e-02
-5.22521399e-02 1.03788459e+00 9.98855710e-01 -7.77739644e-01
-5.90306334e-02 -3.79449338e-01 -3.26865435e-01 1.23930313e-01
-1.10318467e-01 -1.82448536e-01 -1.57116607e-01 -7.99061954e-01
-5.16300872e-02 -8.25567603e-01 -6.76057756e-01 1.44035369e-01
-4.32956666e-01 -1.97934180e-01 5.61206818e-01 -7.25077093e-01
1.11577427e+00 -1.76039016e+00 1.11730850e+00 2.77651101e-01
4.50439006e-01 -2.95547675e-02 -3.17397833e-01 4.66563165e-01
-3.65756452e-02 -1.06646448e-01 -5.68521261e-01 9.21781808e-02
-2.03014128e-02 3.10360896e-03 -1.68492988e-01 4.55092520e-01
5.30189455e-01 1.13991284e+00 -9.78870571e-01 -3.84099782e-02
-1.12181030e-01 7.50995755e-01 -1.13996780e+00 1.52688578e-01
-9.08143103e-01 6.51651502e-01 -5.20151496e-01 6.68502688e-01
4.92389441e-01 -4.21858758e-01 4.93923098e-01 -2.79291719e-01
1.65616293e-02 2.97010299e-02 -6.80410564e-01 2.04967093e+00
-4.24633957e-02 3.62113476e-01 -5.03063835e-02 -8.04444194e-01
1.10790563e+00 -8.34843442e-02 4.48898137e-01 -5.16963661e-01
-2.02669010e-01 1.42708078e-01 4.66195717e-02 -3.45344394e-01
2.55495638e-01 -1.72324821e-01 4.49212968e-01 3.77466500e-01
9.78649855e-02 -1.82731390e-01 6.60436973e-02 4.14201647e-01
1.16385651e+00 8.43029261e-01 1.20750340e-02 -6.92980945e-01
7.12185726e-02 1.08845599e-01 7.69142389e-01 6.33553147e-01
-3.48614454e-02 5.76287866e-01 5.89617431e-01 -4.26849306e-01
-1.31031656e+00 -1.03995049e+00 1.25076901e-02 1.15598917e+00
-1.08187407e-01 -3.72846335e-01 -9.11660194e-01 -6.35049462e-01
2.28099264e-02 8.94083023e-01 -6.98264837e-01 -5.43967724e-01
-4.34110314e-01 -1.14836025e+00 2.26586506e-01 1.16312593e-01
-2.13959739e-01 -1.16809344e+00 -4.88766521e-01 5.53998947e-01
-3.08895353e-02 -3.20001692e-01 -7.83709645e-01 6.03142798e-01
-8.38127017e-01 -8.48003030e-01 -1.24075377e+00 -8.36039245e-01
6.24171138e-01 -1.62361398e-01 1.06336343e+00 -2.78164119e-01
-7.71922886e-01 -1.63964659e-01 -1.06694371e-01 -2.13554665e-01
-4.07913983e-01 9.88178179e-02 2.07775801e-01 -1.25469759e-01
6.70476794e-01 -6.88108504e-01 -8.75797212e-01 4.19657439e-01
-8.98775876e-01 2.64571667e-01 6.71436369e-01 1.26777315e+00
9.25126791e-01 -5.80043256e-01 8.59660387e-01 -6.70996845e-01
9.77799952e-01 -3.34209949e-01 -6.33654535e-01 3.00366789e-01
-7.80676126e-01 5.82821846e-01 2.76391566e-01 -6.18925393e-01
-8.99171174e-01 3.30646187e-01 -2.90573925e-01 -2.52149969e-01
3.05730075e-01 5.78040779e-01 -2.30224043e-01 -2.75014713e-03
1.11431658e+00 4.23341811e-01 4.56673533e-01 -5.44382155e-01
6.21113122e-01 1.94339499e-01 3.95421200e-02 -6.83275104e-01
6.53631389e-02 2.05091059e-01 -3.28110904e-02 -9.62521315e-01
-6.55639768e-01 1.04255334e-01 -5.07162333e-01 -5.56309195e-03
9.05588210e-01 -6.82384372e-01 -7.89249539e-01 8.76398161e-02
-1.10588610e+00 -4.31055933e-01 -2.92488545e-01 4.51694518e-01
-8.98467541e-01 4.47015852e-01 -5.24367452e-01 -5.68438828e-01
-2.35910863e-01 -1.65281284e+00 8.77997696e-01 -4.55180183e-03
-6.43273652e-01 -8.13226700e-01 1.03237823e-01 1.63550332e-01
6.01092398e-01 1.90549865e-01 1.14432752e+00 -2.74126828e-01
-8.97877395e-01 3.83209050e-01 1.20081469e-01 2.53589690e-01
1.31969005e-01 -1.12997860e-01 -6.06755674e-01 -2.92212725e-01
3.47519144e-02 -4.73788887e-01 1.03298068e+00 8.04346800e-01
9.50234056e-01 -1.55948341e-01 -3.91768008e-01 6.71788692e-01
1.20187247e+00 2.99941272e-01 6.84537768e-01 1.79468304e-01
6.19687438e-01 6.36436880e-01 5.12501597e-01 5.45287967e-01
-2.00351864e-01 8.48443210e-01 3.34824860e-01 -1.37665197e-01
-5.41533567e-02 -4.48718607e-01 4.56036389e-01 4.44642127e-01
1.29179746e-01 -1.86552301e-01 -5.12930632e-01 3.45756948e-01
-1.83889747e+00 -9.59818840e-01 2.63537467e-02 1.86516690e+00
1.32550526e+00 -1.35028988e-01 2.00262383e-01 -6.25137866e-01
6.42446041e-01 -3.55176032e-02 -1.27582514e+00 -2.30280280e-01
-4.80836213e-01 2.45570898e-01 3.50401640e-01 6.65872514e-01
-6.20369196e-01 1.02305436e+00 7.20404387e+00 1.18144345e+00
-7.25465477e-01 -6.28096536e-02 8.14606607e-01 -4.47548628e-01
-8.64048839e-01 1.42805010e-01 -9.56473827e-01 4.51981485e-01
7.31175542e-01 1.19258471e-01 7.36673057e-01 7.05718219e-01
6.98841691e-01 -3.93478908e-02 -1.16299772e+00 6.77484035e-01
-1.97311088e-01 -2.06279659e+00 8.35744143e-02 2.74448574e-01
7.59315252e-01 -7.42946565e-02 3.26106787e-01 -2.85506904e-01
4.81739640e-01 -1.21246243e+00 7.55064487e-01 8.34077060e-01
5.07065117e-01 -8.72560203e-01 4.62518632e-02 3.34102213e-01
-5.06648958e-01 1.97967559e-01 -5.79516947e-01 4.79070872e-01
4.75402892e-01 8.23494494e-01 -6.41057551e-01 -2.74407421e-03
4.03078288e-01 8.24104130e-01 -8.13609362e-02 7.08834171e-01
-2.07335621e-01 4.99637187e-01 -2.75897831e-01 -5.04470646e-01
1.80124894e-01 -6.41275227e-01 7.12041974e-01 1.06321490e+00
6.92788437e-02 -1.52736500e-01 -1.48802353e-02 1.66893065e+00
7.95617998e-02 7.01966286e-02 -5.47327280e-01 -2.29423836e-01
2.92764544e-01 7.75699914e-01 -5.44063866e-01 -7.78822452e-02
1.29677206e-01 1.26536191e+00 6.24252975e-01 7.71948636e-01
-1.02028298e+00 -4.32993978e-01 8.13509583e-01 -8.23503211e-02
6.13421857e-01 -3.09813529e-01 -2.31791750e-01 -7.67491341e-01
-2.34863386e-01 -1.28533089e+00 -5.98356910e-02 -6.51496828e-01
-1.22208428e+00 4.11476284e-01 -3.60490769e-01 -5.85427105e-01
1.59418434e-01 -5.48044860e-01 -3.28584760e-02 1.32478356e+00
-1.14331222e+00 -7.51607537e-01 4.05915231e-01 7.36565888e-02
7.94483364e-01 -2.90204436e-01 6.02359712e-01 1.46146372e-01
-6.06861472e-01 4.37581658e-01 4.97601748e-01 -8.21983755e-01
5.17398238e-01 -1.18096948e+00 4.87599701e-01 5.87092042e-01
-1.37624070e-01 1.06890750e+00 1.07319379e+00 -1.04313838e+00
-1.62004375e+00 -9.30833757e-01 5.83258867e-01 -4.19497937e-01
4.20092046e-01 -4.27444041e-01 -9.91182566e-01 7.24295452e-02
3.91353220e-02 -6.43601894e-01 7.04325140e-01 -6.98848115e-03
-4.27763462e-02 4.24859613e-01 -1.22550392e+00 7.70474076e-01
1.31943047e+00 -5.17538667e-01 -2.31618926e-01 6.58915162e-01
9.38216388e-01 -1.08189210e-01 -9.60692525e-01 1.77083671e-01
5.55812299e-01 -7.92855203e-01 1.00615251e+00 -7.95050502e-01
3.05653542e-01 -5.81523657e-01 -2.00491399e-01 -1.36546385e+00
-7.71952868e-01 -1.02454400e+00 -3.41671139e-01 7.55108535e-01
9.03460383e-01 -2.74036497e-01 1.16985929e+00 4.39430177e-01
-1.07084066e-01 -1.13106668e+00 -7.05332875e-01 -6.85977399e-01
1.74734201e-02 -4.44348305e-02 3.95442903e-01 5.60849905e-01
-6.16885908e-02 4.12291408e-01 -6.75695896e-01 -2.32657447e-01
6.90628350e-01 -1.13909990e-01 2.84422576e-01 -8.65770400e-01
-7.49189496e-01 -5.45102835e-01 5.66175394e-02 -1.72090960e+00
-5.25276288e-02 -1.00511646e+00 2.06223339e-01 -1.40014744e+00
3.17294419e-01 -2.26521358e-01 4.67070863e-02 1.98306054e-01
-1.62187412e-01 -9.96185020e-02 -1.38235182e-01 2.61078715e-01
-3.65102470e-01 1.22959006e+00 1.26745296e+00 -2.95017481e-01
-4.21356738e-01 -2.02332869e-01 -8.57686460e-01 2.36377686e-01
5.23429751e-01 -6.07171297e-01 -5.34662127e-01 -2.36348674e-01
3.69664013e-01 -2.82514960e-01 1.14338428e-01 -4.37948525e-01
4.02029864e-02 -3.67219508e-01 3.62589628e-01 -5.64455032e-01
3.67567182e-01 -3.06704313e-01 4.05610025e-01 6.61914527e-01
-6.70979798e-01 -1.13508403e-01 -2.33937979e-01 9.77317274e-01
1.50473610e-01 -3.11203808e-01 8.92262816e-01 -1.14783496e-01
-5.15216768e-01 2.62598664e-01 -6.97460294e-01 -7.33271912e-02
7.89135754e-01 -3.19696903e-01 3.51286791e-02 -1.75511867e-01
-1.17375827e+00 1.31507501e-01 6.33721411e-01 2.94019371e-01
9.89907861e-01 -1.12798142e+00 -7.24030018e-01 1.58075064e-01
-7.17593879e-02 -2.46974781e-01 1.66095272e-01 9.00885999e-01
-6.86341405e-01 3.27902943e-01 -3.18441167e-02 -7.98516452e-01
-1.00667262e+00 6.99556589e-01 4.31226432e-01 4.40803543e-02
-5.09938240e-01 1.16117287e+00 4.38374490e-01 -3.31319153e-01
1.75824299e-01 9.04403627e-02 1.36976868e-01 -1.42331064e-01
5.53769052e-01 2.65580118e-01 -7.45786503e-02 -1.71763048e-01
-3.68316203e-01 2.56960660e-01 -2.34979406e-01 -8.50757733e-02
1.58396697e+00 2.60461383e-02 -7.62617663e-02 3.38587351e-02
9.39674258e-01 -4.25428182e-01 -1.85424650e+00 -1.92648128e-01
2.55909503e-01 -2.85133302e-01 1.87178552e-01 -1.04310811e+00
-6.48231745e-01 6.97671950e-01 7.37493217e-01 -2.88822025e-01
7.96568930e-01 8.38384847e-04 5.25959194e-01 4.27973568e-01
4.37689364e-01 -1.17771506e+00 3.10807765e-01 4.98948812e-01
1.03832114e+00 -8.31562281e-01 1.46659240e-01 -4.56399955e-02
-8.82344782e-01 8.09905231e-01 2.99533576e-01 7.41625950e-02
4.30198520e-01 2.22248882e-02 -2.48342052e-01 -5.25074005e-01
-9.25839186e-01 2.47705233e-04 1.94400549e-01 1.04233027e+00
6.43723726e-01 -1.74441814e-01 -3.94498289e-01 3.35149139e-01
3.41331273e-01 5.54417558e-02 1.15644537e-01 1.01794541e+00
-7.41854906e-01 -1.31366944e+00 -1.36573091e-01 2.97618717e-01
-2.68771231e-01 -4.76920992e-01 -5.22582889e-01 2.55785342e-02
-2.20346004e-01 6.85721576e-01 -4.00734484e-01 -1.54926375e-01
1.80085778e-01 6.61926791e-02 5.97692132e-01 -4.87642258e-01
-3.23805004e-01 4.04175788e-01 -3.04094516e-02 -5.85639536e-01
-1.96555316e-01 -6.71373188e-01 -1.12357974e+00 -1.68695688e-01
-5.33163488e-01 2.16205716e-01 7.21561491e-01 7.39683330e-01
1.02679479e+00 5.83741963e-01 4.16021675e-01 -8.62408221e-01
-6.87506080e-01 -6.91031873e-01 -6.52764916e-01 1.56011894e-01
1.24911204e-01 -6.73350155e-01 1.70423448e-01 1.17727086e-01] | [4.857175350189209, 5.490725040435791] |
eb96f737-7494-4656-8193-a5b886c259df | reinforcement-learning-with-latent-flow-1 | 2101.01857 | null | https://arxiv.org/abs/2101.01857v1 | https://arxiv.org/pdf/2101.01857v1.pdf | Reinforcement Learning with Latent Flow | Temporal information is essential to learning effective policies with Reinforcement Learning (RL). However, current state-of-the-art RL algorithms either assume that such information is given as part of the state space or, when learning from pixels, use the simple heuristic of frame-stacking to implicitly capture temporal information present in the image observations. This heuristic is in contrast to the current paradigm in video classification architectures, which utilize explicit encodings of temporal information through methods such as optical flow and two-stream architectures to achieve state-of-the-art performance. Inspired by leading video classification architectures, we introduce the Flow of Latents for Reinforcement Learning (Flare), a network architecture for RL that explicitly encodes temporal information through latent vector differences. We show that Flare (i) recovers optimal performance in state-based RL without explicit access to the state velocity, solely with positional state information, (ii) achieves state-of-the-art performance on pixel-based challenging continuous control tasks within the DeepMind control benchmark suite, namely quadruped walk, hopper hop, finger turn hard, pendulum swing, and walker run, and is the most sample efficient model-free pixel-based RL algorithm, outperforming the prior model-free state-of-the-art by 1.9X and 1.5X on the 500k and 1M step benchmarks, respectively, and (iv), when augmented over rainbow DQN, outperforms this state-of-the-art level baseline on 5 of 8 challenging Atari games at 100M time step benchmark. | ['Michael Laskin', 'Pieter Abbeel', 'Yang Gao', 'Aravind Rajeswaran', 'Aravind Srinivas', 'Xiaofei Wang', 'Wenling Shang'] | 2021-01-06 | reinforcement-learning-with-latent-flow | http://proceedings.neurips.cc/paper/2021/hash/ba3c5fe1d6d6708b5bffaeb6942b7e04-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/ba3c5fe1d6d6708b5bffaeb6942b7e04-Paper.pdf | neurips-2021-12 | ['montezumas-revenge'] | ['playing-games'] | [ 1.79575637e-01 -3.67130518e-01 -6.59902513e-01 1.68566436e-01
-7.00225413e-01 -4.79990751e-01 1.08049011e+00 -3.95608991e-01
-7.08285630e-01 9.24143851e-01 2.48981580e-01 -4.55686271e-01
-2.62352854e-01 -5.33325434e-01 -9.83362257e-01 -1.00393116e+00
-5.57973564e-01 4.69976246e-01 2.96856582e-01 -5.95820069e-01
3.77310097e-01 1.65817901e-01 -1.51964080e+00 2.53245682e-01
4.66442525e-01 1.18276966e+00 1.54638682e-02 1.12270522e+00
1.58527419e-01 1.64737487e+00 -2.60595530e-01 2.04262450e-01
5.63766122e-01 -5.47297359e-01 -8.26173484e-01 1.53147161e-01
7.16339588e-01 -7.86130011e-01 -9.96553123e-01 4.92556870e-01
4.20486510e-01 4.40508604e-01 3.90218586e-01 -1.14330888e+00
-1.83236599e-01 1.78377464e-01 -5.09105742e-01 4.52925235e-01
3.40379518e-03 1.14381850e+00 1.09663606e+00 -3.51663232e-01
8.34128737e-01 1.34582257e+00 4.91684973e-01 7.30320156e-01
-1.35315144e+00 -5.49915493e-01 4.70477581e-01 5.03953815e-01
-6.65106952e-01 -5.72287381e-01 5.28453648e-01 -3.99731100e-01
1.36812639e+00 -2.01850861e-01 9.88257408e-01 1.48444080e+00
3.38951707e-01 1.25780427e+00 1.26333725e+00 -8.84065628e-02
4.21844006e-01 -1.08774638e+00 -6.90569030e-03 1.05184102e+00
-2.09523126e-01 7.80679464e-01 -8.31503570e-01 2.18516767e-01
1.03950429e+00 -2.45687842e-01 -2.20109522e-01 -6.90949738e-01
-1.41160417e+00 8.58235776e-01 3.79973173e-01 -2.38776252e-01
-4.22586560e-01 1.09046018e+00 6.31974399e-01 2.74465263e-01
2.70217240e-01 4.05285984e-01 -5.88243902e-01 -9.19712484e-01
-1.06752276e+00 6.46850228e-01 6.63721859e-01 6.38019383e-01
7.10128069e-01 7.22854853e-01 -4.58771706e-01 2.25585073e-01
2.01449931e-01 5.06341994e-01 5.64559579e-01 -1.88562477e+00
4.05250371e-01 7.09778455e-04 2.57854968e-01 -4.45546985e-01
-3.75008911e-01 -4.49630558e-01 -6.05262995e-01 7.74965703e-01
5.99306822e-01 -2.08980232e-01 -1.28692985e+00 1.81495750e+00
1.84011057e-01 9.97657120e-01 1.94682032e-01 9.25345898e-01
3.10873389e-01 7.37679005e-01 -2.35684678e-01 -2.85169005e-01
8.40706825e-01 -1.46553397e+00 -5.78669608e-01 -4.26202953e-01
5.12619555e-01 -2.95660555e-01 9.92340922e-01 6.29047275e-01
-1.12578452e+00 -6.54291928e-01 -1.22667849e+00 9.40575451e-02
-2.57339664e-02 -3.87740225e-01 8.50730181e-01 2.68456429e-01
-1.16822302e+00 9.95327294e-01 -1.48529184e+00 -1.44869551e-01
5.46325207e-01 4.14219260e-01 -2.20813885e-01 -2.55600303e-01
-1.07506216e+00 9.00010526e-01 1.05095379e-01 -1.94082148e-02
-1.66826928e+00 -1.01975191e+00 -8.48991573e-01 -3.08549225e-01
8.11828434e-01 -8.72533262e-01 1.74314535e+00 -8.82424772e-01
-2.06650448e+00 2.07419321e-01 -8.21641460e-02 -1.10039711e+00
7.73124337e-01 -6.08670056e-01 7.41331130e-02 4.59642380e-01
-1.17785938e-01 1.01817763e+00 1.12862182e+00 -9.93217230e-01
-7.23749101e-01 1.04928128e-02 3.27606946e-01 4.81783003e-01
2.87659883e-01 -5.04684389e-01 -4.98233438e-01 -4.18366909e-01
-4.20487612e-01 -1.21416092e+00 -3.81216437e-01 1.29203409e-01
-3.40992212e-02 2.82574091e-02 1.25009871e+00 -4.16305602e-01
1.01954412e+00 -1.61896312e+00 4.53870237e-01 -3.92650008e-01
3.13997746e-01 4.56445187e-01 -2.85642147e-01 3.31421822e-01
1.16443716e-01 -2.30039880e-01 -2.08953142e-01 -4.52048272e-01
1.66532099e-02 8.58949959e-01 -3.11079651e-01 5.81132591e-01
1.78881228e-01 1.22237360e+00 -1.31503999e+00 -2.66174614e-01
7.05770850e-01 3.35024804e-01 -7.21342385e-01 1.01779900e-01
-6.51531994e-01 6.39625192e-01 -4.06692266e-01 2.41275564e-01
8.20930302e-02 -2.37120628e-01 1.11734129e-01 -8.11387673e-02
-3.38194221e-01 3.61760587e-01 -9.22084630e-01 2.14184117e+00
-2.92681128e-01 8.49764526e-01 -4.32190858e-02 -1.05308604e+00
2.19791666e-01 2.78214395e-01 9.87788200e-01 -1.00540137e+00
-1.70538247e-01 -1.07252195e-01 1.31624684e-01 -4.09990072e-01
4.53918159e-01 3.06000292e-01 2.14501649e-01 2.55286872e-01
2.29307577e-01 -1.49904072e-01 5.74689686e-01 2.65274793e-01
1.66133451e+00 1.02404761e+00 -2.92042433e-03 -1.06342539e-01
2.31000543e-01 8.30871463e-02 5.73170424e-01 1.05746245e+00
-5.31034589e-01 3.15666825e-01 7.32620120e-01 -6.40269101e-01
-1.15271187e+00 -1.02677071e+00 3.27437550e-01 9.68847752e-01
-1.21809237e-01 -5.56060970e-01 -5.13298869e-01 -5.81770360e-01
2.92317830e-02 5.00544429e-01 -7.83170998e-01 -1.56212851e-01
-9.81066644e-01 -5.94359517e-01 5.49363732e-01 5.23460686e-01
9.14278209e-01 -1.14178061e+00 -1.11707115e+00 5.32630563e-01
-1.20965868e-01 -1.25319374e+00 -3.37758541e-01 3.43777746e-01
-7.88198650e-01 -1.24409986e+00 -5.19139826e-01 -4.05909633e-03
-1.36696249e-01 7.32520521e-02 1.30084038e+00 -2.37820297e-02
-2.22188607e-01 6.75410926e-01 -3.26016366e-01 2.43293926e-01
-1.72551185e-01 -7.20427111e-02 9.45401788e-02 -2.71751672e-01
-3.10231805e-01 -5.05494714e-01 -8.76228154e-01 3.04006208e-02
-6.55292511e-01 1.92857742e-01 5.04226446e-01 1.19895768e+00
4.98311967e-01 -2.64143914e-01 7.41354823e-02 -5.81736743e-01
1.61741093e-01 -1.57537222e-01 -8.38351369e-01 -1.98682189e-01
-5.77448666e-01 5.94749689e-01 5.54839075e-01 -5.04539549e-01
-8.59320641e-01 3.85480225e-02 1.70611404e-02 -1.00064826e+00
4.37424742e-02 2.20657930e-01 4.61099595e-01 -1.32903621e-01
5.07923067e-01 3.42220247e-01 3.71098042e-01 3.55251669e-03
6.40181959e-01 -1.86836049e-01 7.04838455e-01 -9.20314670e-01
4.73881751e-01 7.10954547e-01 5.91264784e-01 -7.06637800e-01
-8.49583149e-01 -4.33487743e-01 -5.27623534e-01 -4.24494594e-01
1.16453660e+00 -1.03003323e+00 -1.12821209e+00 8.12139928e-01
-7.66820908e-01 -1.23100638e+00 -4.03501987e-01 4.95626599e-01
-1.34707606e+00 1.59778625e-01 -9.83861148e-01 -7.67907500e-01
-1.12657040e-01 -1.40830553e+00 1.00629056e+00 -1.30528929e-02
1.65185452e-01 -8.75997007e-01 1.94723174e-01 1.75455034e-01
4.99693632e-01 4.53112662e-01 6.87582254e-01 -1.59496348e-02
-9.21593368e-01 4.24374372e-01 1.61143616e-01 2.72798866e-01
-1.62882552e-01 7.15466887e-02 -8.50197136e-01 -4.22249615e-01
-3.81230742e-01 -6.82546318e-01 1.38629305e+00 8.36107016e-01
9.23506200e-01 -3.03388864e-01 2.32504625e-02 8.35844219e-01
1.32214379e+00 2.56564468e-01 7.78408289e-01 4.04986829e-01
8.40388834e-01 7.21594691e-02 5.63612580e-01 5.95505059e-01
4.62667018e-01 8.49341512e-01 9.41181421e-01 5.22247665e-02
-2.79387772e-01 -2.69600362e-01 7.00200915e-01 4.61676955e-01
-2.73928374e-01 -1.99130833e-01 -7.25175440e-01 3.12572151e-01
-2.33275056e+00 -1.40317631e+00 1.20435491e-01 1.87914562e+00
7.70761430e-01 3.84204298e-01 2.23278508e-01 1.16096377e-01
6.14020675e-02 8.45825851e-01 -9.88439500e-01 -1.34878978e-01
2.19171401e-02 3.98196280e-01 9.27164793e-01 6.55750036e-01
-1.26330924e+00 1.42501092e+00 6.14116240e+00 7.26630807e-01
-1.14083815e+00 -4.83191386e-02 6.48149431e-01 -5.00700772e-01
3.01262885e-01 5.62615767e-02 -6.96963966e-01 2.60458738e-01
1.14201558e+00 3.42678666e-01 1.04285145e+00 5.02713263e-01
5.78078151e-01 -2.77836144e-01 -1.16203785e+00 9.41078424e-01
-1.65875494e-01 -1.74786544e+00 -2.50131875e-01 2.39255831e-01
9.50132728e-01 5.41319191e-01 3.49748373e-01 7.29463756e-01
8.36438715e-01 -1.03036857e+00 6.83563411e-01 6.38279736e-01
6.16948128e-01 -6.94432616e-01 1.74903110e-01 2.93086767e-01
-1.20105970e+00 -3.38894457e-01 -1.08465880e-01 -3.71275097e-01
2.06441969e-01 -3.34721394e-02 -3.57073247e-01 2.94691831e-01
7.43437767e-01 1.26477063e+00 -2.95976788e-01 7.54821777e-01
-2.97902375e-01 9.13847566e-01 -2.36621335e-01 1.87700495e-01
1.10813344e+00 -9.20650214e-02 6.53253853e-01 8.08303893e-01
-2.33577952e-01 1.02587475e-03 5.36085188e-01 3.22538227e-01
8.33848193e-02 -6.16263390e-01 -5.41759908e-01 -5.36991842e-02
8.48045573e-02 8.61097693e-01 -3.85600895e-01 -4.84340668e-01
-4.83493716e-01 9.30223048e-01 2.39537880e-01 6.70444310e-01
-9.31740582e-01 2.30811432e-01 1.22821176e+00 -9.02319029e-02
7.57864833e-01 -6.53008521e-01 3.05648178e-01 -1.09654009e+00
-4.33008522e-01 -1.11618340e+00 2.28563532e-01 -8.27430844e-01
-7.69106030e-01 9.20424238e-02 3.58574428e-02 -1.25463164e+00
-8.77058148e-01 -7.21337676e-01 -3.46739411e-01 3.76612037e-01
-1.75704372e+00 -1.00910568e+00 -1.47889912e-01 7.86735296e-01
9.19026434e-01 -2.55551040e-01 5.81701994e-01 3.44238281e-02
-4.47726220e-01 8.01447630e-02 1.53084069e-01 1.30130678e-01
5.82938373e-01 -1.36864352e+00 5.57385564e-01 8.60558212e-01
1.73455894e-01 1.60187289e-01 7.02545941e-01 -3.67210209e-01
-1.88463700e+00 -9.79799271e-01 -1.50164934e-02 -4.58208084e-01
9.74129677e-01 -7.08050951e-02 -4.29488480e-01 7.60995924e-01
4.11552399e-01 5.72559774e-01 -1.06621265e-01 -1.52845576e-01
-2.71941155e-01 -5.73466495e-02 -6.00077748e-01 8.19270611e-01
1.28559899e+00 -3.64641666e-01 -2.87108570e-01 2.47360393e-01
8.52106631e-01 -7.67147839e-01 -5.40684402e-01 2.47959390e-01
7.73308516e-01 -1.01733041e+00 1.13368535e+00 -1.03273821e+00
6.73286259e-01 -2.98974484e-01 -2.97613114e-01 -1.36186290e+00
-3.11056405e-01 -1.17758262e+00 -8.31972539e-01 3.81287158e-01
3.82769182e-02 -9.24969241e-02 1.07335854e+00 8.13603848e-02
-3.79744172e-01 -1.03923619e+00 -9.98865008e-01 -7.82991230e-01
2.42228687e-01 -6.59048617e-01 1.37411818e-01 5.27490914e-01
-5.02155781e-01 2.63925344e-01 -8.10590029e-01 -1.25697464e-01
6.79624379e-01 -2.08756402e-01 9.10053313e-01 -6.67235672e-01
-6.48740947e-01 -7.08412170e-01 -4.20597374e-01 -1.47863948e+00
4.53647524e-01 -4.17797416e-01 6.21553808e-02 -1.57327318e+00
-2.42585421e-01 -2.33557299e-02 -3.34759563e-01 5.23585618e-01
1.21604763e-02 -6.13971055e-03 6.13729954e-01 8.76830071e-02
-9.73156214e-01 7.76466250e-01 1.43723214e+00 -2.31437802e-01
-1.24988571e-01 -2.76741326e-01 -1.22814834e-01 5.95604658e-01
5.28391838e-01 -1.36678711e-01 -6.75185323e-01 -4.94661421e-01
-1.00844920e-01 5.43066323e-01 6.92175329e-01 -1.28339589e+00
2.22948849e-01 -4.67061579e-01 3.82563055e-01 -5.31497180e-01
7.89335847e-01 -4.93011057e-01 -3.75454366e-01 9.34325039e-01
-4.53044921e-01 3.09407473e-01 2.32725933e-01 7.35680461e-01
5.48968054e-02 3.65804791e-01 8.25563252e-01 -3.15716833e-01
-1.25507891e+00 6.71073198e-01 -4.83339131e-01 4.69430089e-01
7.83456087e-01 -1.71906166e-02 -6.28525615e-01 -5.09197533e-01
-5.08652568e-01 4.36314940e-01 1.03224464e-01 3.81202251e-01
4.87900585e-01 -1.22107148e+00 -6.29953623e-01 -6.15589647e-03
-2.49435946e-01 -9.05277431e-02 2.39454642e-01 8.35098326e-01
-6.67928159e-01 4.82756615e-01 -3.40030432e-01 -9.79677379e-01
-7.54877567e-01 4.58506048e-01 6.47617996e-01 -7.05330610e-01
-7.93489754e-01 6.52571142e-01 1.42425090e-01 -1.47533566e-01
1.27880916e-01 -4.06518906e-01 3.04169077e-02 -1.46415204e-01
4.04035598e-01 4.78144050e-01 -1.81517944e-01 -3.92973155e-01
-7.54522607e-02 6.25923514e-01 -4.60246764e-02 -5.90147853e-01
1.19479036e+00 1.68497875e-01 5.08400261e-01 5.01810789e-01
1.06454539e+00 -5.66428900e-01 -2.55131388e+00 -1.26734510e-01
-2.21385852e-01 -2.75413454e-01 4.49423432e-01 -8.02209496e-01
-1.27448750e+00 9.02617157e-01 5.73276043e-01 -3.61982763e-01
9.02181685e-01 -5.77903152e-01 8.73656094e-01 7.20029831e-01
4.38759774e-01 -1.14239907e+00 5.22549689e-01 9.91048396e-01
6.02878928e-01 -1.29003346e+00 1.40721962e-01 3.64060223e-01
-7.94927657e-01 9.35590386e-01 7.33856857e-01 -5.37267625e-01
5.81835032e-01 3.77684802e-01 -1.26117945e-01 6.76362291e-02
-1.66582274e+00 -5.12105465e-01 1.23320863e-01 5.60218990e-01
7.53410012e-02 -2.72259027e-01 1.28783822e-01 -2.87255913e-01
5.09097762e-02 8.63365605e-02 4.01330411e-01 1.27095997e+00
-4.82939512e-01 -9.12086785e-01 1.35848999e-01 4.44952309e-01
-2.96428531e-01 -1.43764988e-01 2.76837975e-01 1.18757343e+00
-2.20642611e-01 8.47449780e-01 2.62910604e-01 -2.47777507e-01
1.28539219e-01 -7.53124207e-02 8.34219038e-01 -1.08219624e-01
-5.27053714e-01 1.03254661e-01 9.92201790e-02 -1.33119345e+00
-5.60330093e-01 -7.50321627e-01 -1.37581944e+00 -5.99506617e-01
2.89173484e-01 -3.02647740e-01 4.24236417e-01 1.13806629e+00
2.68787026e-01 8.12890291e-01 3.62598509e-01 -1.28556669e+00
-7.30419934e-01 -7.15921938e-01 -1.24653377e-01 2.63583988e-01
8.87606561e-01 -9.88119364e-01 -2.12297082e-01 2.16879658e-02] | [4.2708024978637695, 1.5357977151870728] |
783905cc-305a-4da5-a3d9-ed7bef34851d | yolo-pose-enhancing-yolo-for-multi-person | 2204.06806 | null | https://arxiv.org/abs/2204.06806v1 | https://arxiv.org/pdf/2204.06806v1.pdf | YOLO-Pose: Enhancing YOLO for Multi Person Pose Estimation Using Object Keypoint Similarity Loss | We introduce YOLO-pose, a novel heatmap-free approach for joint detection, and 2D multi-person pose estimation in an image based on the popular YOLO object detection framework. Existing heatmap based two-stage approaches are sub-optimal as they are not end-to-end trainable and training relies on a surrogate L1 loss that is not equivalent to maximizing the evaluation metric, i.e. Object Keypoint Similarity (OKS). Our framework allows us to train the model end-to-end and optimize the OKS metric itself. The proposed model learns to jointly detect bounding boxes for multiple persons and their corresponding 2D poses in a single forward pass and thus bringing in the best of both top-down and bottom-up approaches. Proposed approach doesn't require the postprocessing of bottom-up approaches to group detected keypoints into a skeleton as each bounding box has an associated pose, resulting in an inherent grouping of the keypoints. Unlike top-down approaches, multiple forward passes are done away with since all persons are localized along with their pose in a single inference. YOLO-pose achieves new state-of-the-art results on COCO validation (90.2% AP50) and test-dev set (90.3% AP50), surpassing all existing bottom-up approaches in a single forward pass without flip test, multi-scale testing, or any other test time augmentation. All experiments and results reported in this paper are without any test time augmentation, unlike traditional approaches that use flip-test and multi-scale testing to boost performance. Our training codes will be made publicly available at https://github.com/TexasInstruments/edgeai-yolov5 and https://github.com/TexasInstruments/edgeai-yolox | ['Deepak Poddar', 'Manu Mathew', 'Soyeb Nagori', 'Debapriya Maji'] | 2022-04-14 | null | null | null | null | ['multi-person-pose-estimation'] | ['computer-vision'] | [-4.36491191e-01 -1.65004600e-02 1.02928340e-01 -3.50881994e-01
-1.20295191e+00 -5.26827157e-01 3.37133050e-01 2.24893540e-02
-6.53548002e-01 5.63913405e-01 -2.01112822e-01 3.22358966e-01
6.18430227e-02 -4.28124726e-01 -9.98805225e-01 -5.33595562e-01
-2.17243582e-01 8.56887817e-01 6.02904379e-01 7.89528340e-02
-1.78356245e-01 1.11442134e-01 -1.42134857e+00 7.74441585e-02
4.91247505e-01 8.83682072e-01 -3.97563763e-02 1.11602998e+00
5.61799586e-01 2.67481238e-01 -5.37428021e-01 -6.06460512e-01
4.00252819e-01 -3.56817544e-01 -4.10804093e-01 -7.55834132e-02
1.27273333e+00 -2.40701154e-01 -1.56825572e-01 8.94347608e-01
9.31629479e-01 1.10589206e-01 4.71599966e-01 -1.43498302e+00
-2.60422856e-01 -3.55420797e-03 -9.75557506e-01 1.79257914e-01
5.68746388e-01 2.11893529e-01 8.44255090e-01 -1.32764065e+00
4.42327678e-01 1.32112038e+00 1.13551962e+00 4.67991173e-01
-1.02807999e+00 -6.05576038e-01 6.29342124e-02 7.38313347e-02
-1.44533193e+00 -1.68366060e-01 5.35394490e-01 -4.67197567e-01
6.97220147e-01 2.46279359e-01 6.78016543e-01 9.49037254e-01
3.63983214e-02 1.03928077e+00 1.04385793e+00 -4.01374370e-01
-9.58581567e-02 1.66671455e-01 1.89326599e-01 1.13026106e+00
4.32307005e-01 3.75821926e-02 -5.17076552e-01 -7.04421625e-02
7.51737058e-01 2.27115098e-02 1.39202327e-01 -7.71224082e-01
-1.07147491e+00 6.35171592e-01 6.54686809e-01 -2.37658784e-01
-8.99295583e-02 5.80828905e-01 3.99843901e-01 -5.18462025e-02
2.98490226e-01 6.40260577e-02 -4.86305475e-01 5.39742224e-02
-1.11108828e+00 6.70848548e-01 4.27257627e-01 8.83057833e-01
6.53433442e-01 -3.51494431e-01 -3.51516068e-01 7.35113263e-01
5.87842107e-01 5.88565588e-01 3.04047674e-01 -6.82117760e-01
6.48555040e-01 5.62653065e-01 1.94172919e-01 -6.42128408e-01
-4.99577880e-01 -7.15769351e-01 -3.35109681e-01 3.52072090e-01
7.38575935e-01 -3.20820779e-01 -1.26006591e+00 1.74168682e+00
8.49825501e-01 2.01731965e-01 -3.52567226e-01 1.07378340e+00
8.40931892e-01 4.71328825e-01 1.65600106e-01 3.34790885e-01
1.90708280e+00 -1.44956136e+00 -1.51757658e-01 -5.16666055e-01
5.96815705e-01 -6.82286203e-01 1.07681620e+00 4.70115870e-01
-1.28371871e+00 -8.30641568e-01 -1.05155158e+00 -2.51264781e-01
-3.75272483e-01 6.16876841e-01 5.13032436e-01 8.89275312e-01
-8.13189805e-01 2.87195116e-01 -9.47675765e-01 -4.79286790e-01
5.89177907e-01 3.76485109e-01 -5.78501284e-01 -3.76903974e-02
-8.37783873e-01 7.96052456e-01 2.47376502e-01 2.65091479e-01
-1.05036461e+00 -7.05516815e-01 -9.82486129e-01 -2.70059377e-01
5.97170055e-01 -9.50511813e-01 1.15060198e+00 -3.14706504e-01
-1.07442629e+00 8.66179466e-01 -1.03289537e-01 -3.26157868e-01
1.15472305e+00 -9.38922107e-01 -9.69947726e-02 2.40822241e-01
5.18962026e-01 9.48629737e-01 6.63017929e-01 -1.07218349e+00
-7.93521643e-01 -4.96141464e-01 -8.47065300e-02 3.42874199e-01
-2.75351349e-02 7.75286630e-02 -1.14356554e+00 -6.66039228e-01
4.08758484e-02 -1.19429672e+00 -9.62695479e-02 2.85711914e-01
-7.19810665e-01 -3.59517932e-01 5.77005565e-01 -5.99533796e-01
8.24696302e-01 -1.92061245e+00 5.50529994e-02 -7.36052589e-03
7.82970265e-02 1.20399155e-01 1.41917884e-01 2.39360124e-01
4.07825224e-02 -1.54518679e-01 -2.10145459e-01 -8.16029370e-01
1.07461654e-01 -4.25387979e-01 3.78245592e-01 9.01263535e-01
2.04020083e-01 9.13211048e-01 -6.93361461e-01 -6.61897838e-01
3.87899488e-01 6.82887495e-01 -6.41325355e-01 -1.12070367e-02
6.79922551e-02 2.94005126e-01 -2.08172932e-01 8.47976983e-01
6.03891790e-01 -1.38159499e-01 -2.30830252e-01 -1.25383914e-01
1.56081989e-01 -1.20610774e-01 -1.58499885e+00 1.83620429e+00
-1.45243973e-01 3.84920388e-01 -2.43787453e-01 -6.05376482e-01
5.76866508e-01 2.80987382e-01 3.10302168e-01 -2.63823390e-01
7.55652562e-02 8.48071277e-02 -2.48930469e-01 -2.24823922e-01
2.18670920e-01 5.03365397e-02 -4.53672379e-01 -5.00599183e-02
2.48252138e-01 2.82939821e-01 4.70229417e-01 1.90729171e-01
7.55959988e-01 6.37376428e-01 1.95569381e-01 1.41930757e-02
3.99913877e-01 6.16326220e-02 5.42792559e-01 8.72374594e-01
-4.30313021e-01 1.00013316e+00 2.96817988e-01 -3.87092799e-01
-1.07008350e+00 -1.29675877e+00 -1.87999457e-01 1.27848256e+00
2.27103829e-01 -4.68901485e-01 -8.37157547e-01 -8.44703615e-01
1.91979349e-01 2.04606369e-01 -8.74055266e-01 1.44531861e-01
-6.24071181e-01 -7.97216177e-01 7.95094967e-01 6.73144221e-01
5.23305833e-01 -7.82773137e-01 -7.49811709e-01 2.89332755e-02
2.85759196e-03 -1.15114403e+00 -6.43782139e-01 -5.40226847e-02
-7.13595450e-01 -1.27306843e+00 -1.17929733e+00 -6.94527090e-01
8.21932256e-01 -1.17114913e-02 9.38990057e-01 -2.31836841e-01
-7.05580592e-01 3.52941215e-01 -1.89016491e-01 -4.72067386e-01
2.31429085e-01 -6.11742325e-02 2.21824005e-01 -1.48336753e-01
3.12329382e-01 -2.03041315e-01 -9.91011858e-01 5.04937410e-01
-1.99345201e-01 -9.87249017e-02 7.74377406e-01 7.17977703e-01
6.77514553e-01 -3.19681734e-01 4.45198983e-01 -7.84073293e-01
6.58380054e-03 -2.09262058e-01 -5.23316205e-01 3.55857283e-01
-3.89001757e-01 -2.73421317e-01 1.71205193e-01 -3.87886316e-01
-7.09944963e-01 4.47186261e-01 8.29537138e-02 -4.80547428e-01
-2.56991506e-01 4.11268361e-02 -1.02445140e-01 1.90373845e-02
8.06332231e-01 4.91829179e-02 -2.66500205e-01 -6.89152122e-01
3.90135020e-01 2.57167339e-01 8.12339365e-01 -5.11997461e-01
1.09355903e+00 5.19279540e-01 -1.62790164e-01 -4.78170723e-01
-8.89788449e-01 -8.01631391e-01 -7.67422616e-01 -3.67688626e-01
1.12061608e+00 -1.25192511e+00 -7.01018631e-01 5.05863070e-01
-7.42113531e-01 -9.77176875e-02 -5.40334769e-02 6.16713941e-01
-4.05407697e-01 3.19402933e-01 -4.78176266e-01 -8.02581966e-01
-5.11060417e-01 -8.90365124e-01 1.36182976e+00 4.28213358e-01
-1.58555150e-01 -5.16982615e-01 1.38656884e-01 6.96093440e-01
-1.43173471e-01 5.47462404e-01 1.15642725e-02 -5.23086548e-01
-4.07991678e-01 -9.92375314e-01 -1.71320885e-01 9.89823639e-02
-3.46135736e-01 -2.53292918e-01 -1.02674782e+00 -7.89927900e-01
-5.19189477e-01 -6.40439212e-01 8.10382485e-01 5.40390968e-01
7.16975272e-01 3.81518230e-02 -4.62428391e-01 6.25347495e-01
1.34661591e+00 -3.53803188e-01 2.88493067e-01 5.28934956e-01
9.01202798e-01 4.53868091e-01 8.08571160e-01 2.07579151e-01
5.71985662e-01 1.00423920e+00 1.83410034e-01 -2.85845816e-01
-3.42314869e-01 -4.15351331e-01 4.35320407e-01 1.76908392e-02
-1.40885577e-01 -1.89037099e-02 -9.55084205e-01 6.14163935e-01
-1.93961668e+00 -9.25861478e-01 -2.00361118e-01 2.29575920e+00
5.91535389e-01 3.96827668e-01 6.69370472e-01 6.26408122e-03
8.23522329e-01 -1.81276187e-01 -4.28076684e-01 2.37474754e-01
2.05366030e-01 -2.82495562e-02 5.95310748e-01 5.23331106e-01
-1.56748354e+00 8.97928536e-01 5.05381632e+00 6.96876585e-01
-7.08173811e-01 4.07940000e-01 6.38712347e-01 -6.41796827e-01
6.30290389e-01 -1.15163825e-01 -1.25636756e+00 3.88669223e-01
5.62782109e-01 3.49200547e-01 -2.02397808e-01 1.20373571e+00
1.01439856e-01 -4.01067287e-01 -1.14638174e+00 1.06586361e+00
1.96019739e-01 -9.45645154e-01 -5.22797763e-01 -1.83915764e-01
5.45739889e-01 -1.64807960e-02 -4.81722988e-02 5.37427485e-01
1.60967529e-01 -8.71885121e-01 1.02564943e+00 3.28576386e-01
7.74291039e-01 -7.39511132e-01 7.83922076e-01 1.25831425e-01
-1.41389263e+00 4.57454249e-02 -2.06316531e-01 1.68827847e-01
2.70933896e-01 2.28971854e-01 -7.90484130e-01 5.82290828e-01
1.20701587e+00 2.21063972e-01 -1.01963639e+00 1.46415782e+00
-4.15308446e-01 3.87213498e-01 -6.33256018e-01 -1.06039327e-02
2.19761387e-01 4.16592121e-01 6.80814683e-01 1.40159214e+00
1.59387767e-01 -4.30327445e-01 5.61884820e-01 4.04223979e-01
7.02641085e-02 -1.07928470e-01 -1.08692735e-01 6.09313071e-01
3.86110187e-01 1.29471421e+00 -8.23096395e-01 -4.60777968e-01
-3.15790534e-01 1.14555109e+00 2.59164631e-01 7.18109831e-02
-1.16334927e+00 -3.79602253e-01 2.71059662e-01 3.95488471e-01
4.92510438e-01 -1.55173898e-01 -1.41565114e-01 -9.74164426e-01
4.24074411e-01 -6.01551473e-01 6.65836692e-01 -5.11537313e-01
-1.16167104e+00 4.27085429e-01 4.43674684e-01 -1.26851332e+00
-2.90427893e-01 -5.42397738e-01 -5.37710369e-01 7.91373610e-01
-1.16845655e+00 -1.66039491e+00 -4.89811808e-01 4.74738628e-01
5.87283075e-01 4.35318984e-03 5.08814991e-01 5.40992558e-01
-7.53074527e-01 1.04430008e+00 -2.51212269e-01 5.54779887e-01
1.05478644e+00 -1.53506422e+00 4.79909748e-01 1.11711955e+00
1.97230458e-01 6.12097025e-01 8.90347898e-01 -6.70583069e-01
-1.13397777e+00 -9.82199192e-01 6.25254571e-01 -9.48234379e-01
3.46929103e-01 -7.55987883e-01 -4.09028322e-01 7.48473942e-01
-1.54307485e-01 5.58393896e-01 6.28719330e-01 2.62159735e-01
-2.98422098e-01 -1.01854980e-01 -1.16603899e+00 4.34674144e-01
1.02893376e+00 -1.24822780e-01 -5.47928274e-01 4.82485890e-01
2.38082245e-01 -7.31975257e-01 -6.38324320e-01 4.09285337e-01
7.28262782e-01 -9.21338141e-01 1.16317666e+00 -4.26726878e-01
1.04083613e-01 -7.61714458e-01 1.69796661e-01 -7.07821786e-01
-2.77929068e-01 -3.94566655e-01 -2.75178373e-01 1.06861734e+00
4.96053874e-01 -2.53045619e-01 1.19386923e+00 3.95173728e-01
-2.92383023e-02 -9.35092807e-01 -1.08426666e+00 -9.08413351e-01
-2.84131885e-01 -5.07345378e-01 -5.62131181e-02 4.93589520e-01
-2.49632612e-01 2.89346308e-01 -5.86981177e-01 5.34517765e-01
1.11256146e+00 -1.38062432e-01 1.28777170e+00 -9.83019233e-01
-6.74127698e-01 -1.38719887e-01 -7.56719112e-01 -8.30399096e-01
-6.09905779e-01 -6.57137513e-01 1.61088765e-01 -1.55884647e+00
3.14724058e-01 -3.42955828e-01 -2.11231008e-01 7.24784970e-01
-4.78789151e-01 8.89694691e-01 4.57210720e-01 2.94660985e-01
-8.66897762e-01 3.42740566e-01 8.62096071e-01 1.08855970e-01
-2.76169777e-01 6.37141168e-02 -4.37553138e-01 7.03322530e-01
5.23028553e-01 -7.24948466e-01 -3.14186603e-01 -2.34105423e-01
5.90809919e-02 -1.93119884e-01 9.31484163e-01 -1.53706551e+00
3.05213392e-01 4.39623386e-01 9.56464469e-01 -8.20093453e-01
7.43852496e-01 -4.45042491e-01 -7.16403313e-03 6.96379423e-01
5.50094582e-02 1.49561927e-01 2.31389686e-01 7.00411201e-01
1.07009262e-01 -1.19098604e-01 8.62910032e-01 -1.62440151e-01
-6.81376815e-01 3.78174484e-01 2.85510242e-01 -2.63525099e-02
1.39994574e+00 -5.79901218e-01 -1.29686669e-01 -7.76160434e-02
-1.03437281e+00 5.04513800e-01 3.06821316e-01 4.96572733e-01
3.83862704e-01 -1.19574380e+00 -9.72477257e-01 -1.38371915e-01
1.82765573e-01 2.78347611e-01 3.39680582e-01 9.44440782e-01
-7.55540133e-01 2.77364343e-01 -1.43909276e-01 -8.57579648e-01
-1.55363107e+00 2.87802964e-01 5.19426942e-01 -3.31621051e-01
-7.73517132e-01 1.29375041e+00 2.69284636e-01 -4.90433842e-01
3.41175705e-01 7.42393136e-02 9.69048142e-02 2.85614431e-02
5.16211748e-01 3.70967716e-01 -3.95222530e-02 -5.92849910e-01
-7.82908261e-01 8.04634869e-01 -2.77798116e-01 -2.07121000e-01
1.19192314e+00 1.33323491e-01 4.32137072e-01 3.46532255e-01
9.95285094e-01 4.75316569e-02 -1.53633535e+00 9.09933150e-02
-1.38617426e-01 -4.69284594e-01 -1.84942856e-01 -9.28512573e-01
-7.75854528e-01 6.86832309e-01 1.04569972e+00 -2.37199977e-01
5.73040783e-01 2.82483071e-01 5.86101890e-01 1.69107869e-01
1.69442683e-01 -1.15796483e+00 4.06689167e-01 3.70873436e-02
7.75661886e-01 -1.39807129e+00 3.75794619e-01 -4.91243184e-01
-5.44027507e-01 8.14412534e-01 9.50342059e-01 -4.08859789e-01
4.00365204e-01 1.07039206e-01 6.05879687e-02 -3.02802116e-01
-3.24435085e-01 -3.12210798e-01 6.21275187e-01 4.25990671e-01
4.90122199e-01 2.63316371e-02 -1.22907259e-01 4.56530303e-01
-3.11774671e-01 -6.73744455e-02 -1.38245255e-01 1.03241682e+00
-4.22094911e-01 -8.14809501e-01 -7.83368528e-01 2.89764285e-01
-5.52871704e-01 1.58610120e-01 -2.85792023e-01 1.11300170e+00
3.09697092e-01 5.36769390e-01 -8.53668675e-02 -2.58096993e-01
5.00650227e-01 1.68511018e-01 5.66404819e-01 -7.07245052e-01
-4.47340578e-01 2.29617909e-01 1.33908540e-01 -4.55911547e-01
-1.33875817e-01 -8.64775181e-01 -1.24227095e+00 -1.33084550e-01
-4.55524415e-01 -3.07098497e-02 4.89329100e-01 7.97041774e-01
2.27912575e-01 4.80826795e-01 1.46088168e-01 -1.12749803e+00
-3.45685810e-01 -1.17145562e+00 -1.68007448e-01 3.24200213e-01
2.41930768e-01 -9.10098255e-01 1.19977891e-02 -8.41725394e-02] | [7.358706951141357, -0.706452488899231] |
e18534f8-8edc-4811-b611-a988a0c95e6f | writeahead2-mining-lexical-grammar-patterns | null | null | https://aclanthology.org/N15-3022 | https://aclanthology.org/N15-3022.pdf | WriteAhead2: Mining Lexical Grammar Patterns for Assisted Writing | null | ['Jason Chang', 'Jim Chang'] | 2015-06-01 | null | null | null | naacl-2015-6 | ['grammatical-error-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.393991470336914, 3.7183563709259033] |
2496f5db-9831-4d63-889b-aa8158aee312 | pedestrian-alignment-network-for-large-scale | 1707.00408 | null | http://arxiv.org/abs/1707.00408v1 | http://arxiv.org/pdf/1707.00408v1.pdf | Pedestrian Alignment Network for Large-scale Person Re-identification | Person re-identification (person re-ID) is mostly viewed as an image
retrieval problem. This task aims to search a query person in a large image
pool. In practice, person re-ID usually adopts automatic detectors to obtain
cropped pedestrian images. However, this process suffers from two types of
detector errors: excessive background and part missing. Both errors deteriorate
the quality of pedestrian alignment and may compromise pedestrian matching due
to the position and scale variances. To address the misalignment problem, we
propose that alignment can be learned from an identification procedure. We
introduce the pedestrian alignment network (PAN) which allows discriminative
embedding learning and pedestrian alignment without extra annotations. Our key
observation is that when the convolutional neural network (CNN) learns to
discriminate between different identities, the learned feature maps usually
exhibit strong activations on the human body rather than the background. The
proposed network thus takes advantage of this attention mechanism to adaptively
locate and align pedestrians within a bounding box. Visual examples show that
pedestrians are better aligned with PAN. Experiments on three large-scale re-ID
datasets confirm that PAN improves the discriminative ability of the feature
embeddings and yields competitive accuracy with the state-of-the-art methods. | ['Yi Yang', 'Zhedong Zheng', 'Liang Zheng'] | 2017-07-03 | null | null | null | null | ['large-scale-person-re-identification'] | ['computer-vision'] | [-7.20716342e-02 -3.79308164e-01 3.53633389e-02 -2.54932612e-01
-4.23025191e-01 -3.99921685e-01 6.47062719e-01 1.24722384e-02
-9.72436488e-01 4.86338228e-01 2.57085502e-01 4.41402286e-01
5.47780871e-01 -6.17005646e-01 -6.70947015e-01 -6.52907789e-01
2.89619058e-01 3.26818496e-01 2.63404757e-01 1.72410011e-01
4.19225171e-02 4.70847070e-01 -1.74706078e+00 9.52586532e-04
4.92539376e-01 7.09300816e-01 8.05215687e-02 5.86148679e-01
2.50805408e-01 -1.98585130e-02 -7.77688980e-01 -8.37818742e-01
5.79858601e-01 -2.78401434e-01 -4.76354599e-01 1.59104899e-01
1.12545502e+00 -7.00291932e-01 -6.98057473e-01 1.36915648e+00
8.31289589e-01 2.52009064e-01 5.89529634e-01 -1.55200040e+00
-9.32418466e-01 1.69154778e-01 -7.62790442e-01 4.87749875e-01
3.19412917e-01 2.50155091e-01 8.71126115e-01 -1.19475985e+00
3.54896039e-01 1.37082648e+00 7.78761446e-01 8.96809161e-01
-1.14749599e+00 -7.55152643e-01 1.89162672e-01 5.02667665e-01
-1.77835774e+00 -4.75555420e-01 6.10588312e-01 -4.91816252e-01
7.17065156e-01 1.76716045e-01 8.59186172e-01 1.30906165e+00
-2.53956497e-01 9.01435852e-01 6.44801319e-01 -4.28303152e-01
-1.36293232e-01 3.49331439e-01 2.43922606e-01 4.84055191e-01
8.16554129e-01 2.92718887e-01 -5.38869321e-01 5.02599739e-02
8.77994657e-01 1.99836031e-01 -1.10239886e-01 -5.84483266e-01
-1.29742324e+00 4.73032683e-01 6.60993814e-01 1.95215061e-01
-1.66802377e-01 2.82292515e-01 4.76189315e-01 -1.50993347e-01
-1.12101786e-01 1.93597600e-01 1.29162028e-01 2.62592345e-01
-6.95833445e-01 3.49997044e-01 1.89210907e-01 1.00721848e+00
6.54185653e-01 -1.51388168e-01 -4.02422965e-01 8.86746466e-01
2.00845540e-01 7.99494207e-01 4.87912774e-01 -4.87955213e-01
4.80325639e-01 5.34139872e-01 3.91911119e-01 -1.21261060e+00
-2.68192887e-01 -3.59784245e-01 -9.91766214e-01 1.54168800e-01
9.93143857e-01 1.72042940e-02 -6.43245935e-01 1.84267187e+00
3.97742778e-01 2.11039960e-01 -1.98959962e-01 1.28151584e+00
8.37846696e-01 1.25470132e-01 4.93110180e-01 5.09613514e-01
1.77121925e+00 -1.19265187e+00 -2.93621242e-01 -5.99178493e-01
2.65584111e-01 -6.44486248e-01 7.43373334e-01 -1.67939752e-01
-9.15663779e-01 -1.22221732e+00 -1.09123564e+00 -6.26592040e-02
-4.64757591e-01 6.12691522e-01 1.07408844e-01 9.10677850e-01
-9.62632179e-01 4.17809367e-01 -4.31582630e-01 -8.76418650e-01
3.71411145e-01 3.74775320e-01 -7.90937543e-01 7.20944116e-03
-1.02479970e+00 1.01894295e+00 2.01628640e-01 2.77231067e-01
-6.36364520e-01 -4.12264645e-01 -9.28793788e-01 1.70699611e-01
-3.64603847e-02 -1.07670283e+00 9.04951453e-01 -1.04207969e+00
-1.02336478e+00 1.32648420e+00 -4.53693658e-01 -5.70580661e-01
6.43789828e-01 -4.76506621e-01 -4.76825416e-01 1.36824220e-01
4.01235223e-01 9.52594936e-01 9.92548108e-01 -1.19195676e+00
-9.75990951e-01 -6.42033041e-01 -2.33133167e-01 1.88443288e-01
-4.92392361e-01 1.59943998e-01 -6.56842828e-01 -6.44552946e-01
-3.12297642e-01 -9.96843576e-01 -2.46086463e-01 2.62172073e-01
-3.54010701e-01 -3.91410649e-01 4.60486382e-01 -8.91652346e-01
1.04661882e+00 -2.07047153e+00 -3.10784113e-02 1.54367685e-01
2.46327892e-01 3.74708086e-01 -2.04721779e-01 -6.97973138e-03
-2.84660786e-01 -7.00153112e-02 3.75529498e-01 -4.94563282e-01
6.44399673e-02 -1.48866490e-01 6.13634735e-02 7.81722784e-01
7.57491440e-02 1.13901317e+00 -7.85555005e-01 -5.79132497e-01
4.11882102e-01 4.57352549e-01 -3.84595484e-01 2.55932122e-01
6.38190031e-01 3.38235855e-01 -1.03761755e-01 6.37202799e-01
7.74751246e-01 -1.46757290e-01 -7.60617554e-02 -7.29557395e-01
-8.32781121e-02 -2.08923370e-01 -1.30771029e+00 1.42868197e+00
9.47205052e-02 6.77439630e-01 -1.77336201e-01 -1.01235831e+00
7.52373874e-01 -2.03131866e-02 3.01237613e-01 -8.05985093e-01
9.94308293e-02 4.75087799e-02 9.72677022e-02 -2.98311025e-01
6.92583501e-01 4.35307026e-01 -1.03510752e-01 2.93312013e-01
1.96553990e-02 9.22477484e-01 4.86108884e-02 2.13173851e-02
5.45212388e-01 -9.76241902e-02 4.70903188e-01 -2.44132370e-01
7.58475065e-01 -3.25654984e-01 5.35038590e-01 9.57617581e-01
-8.93897295e-01 4.85019743e-01 -1.18225306e-01 -6.72695935e-01
-1.63432622e+00 -1.10640335e+00 -1.44447032e-02 1.12410676e+00
6.39381409e-01 -3.36717755e-01 -8.99166465e-01 -6.74577892e-01
2.22900957e-01 2.27786571e-01 -7.67755151e-01 -2.29059413e-01
-9.38359857e-01 -7.46588528e-01 6.36795580e-01 9.54893053e-01
8.96245182e-01 -7.60469735e-01 -7.62320161e-01 2.73805717e-03
-4.11427975e-01 -1.28320420e+00 -1.05286169e+00 -4.01351482e-01
-2.02650815e-01 -1.16673732e+00 -1.28598142e+00 -9.87990677e-01
9.86943305e-01 7.57487476e-01 8.53639483e-01 1.72807693e-01
-5.59539557e-01 7.89403021e-01 -2.57000048e-02 -5.17290123e-02
-4.80608344e-02 -1.65381029e-01 5.23306072e-01 2.69975930e-01
9.50858951e-01 -1.13496974e-01 -1.16571879e+00 6.85899675e-01
-2.38369808e-01 -5.67458123e-02 4.51403230e-01 1.06178570e+00
4.27783012e-01 -1.83172360e-01 3.14710885e-01 -1.81916840e-02
2.23340914e-01 8.29007551e-02 -5.40175378e-01 4.53379124e-01
-3.13360304e-01 -1.45450085e-01 2.50849068e-01 -6.10112488e-01
-6.68570936e-01 2.86368996e-01 1.58733744e-02 -2.96313584e-01
-4.16800946e-01 -5.25091827e-01 -4.65248615e-01 -2.26156935e-01
4.63942021e-01 3.09110343e-01 -2.21532375e-01 -4.15019602e-01
2.94702679e-01 7.34023392e-01 9.56124842e-01 -4.20649201e-01
9.55438912e-01 5.81003308e-01 -2.31675372e-01 -8.10981035e-01
-5.54435670e-01 -8.09833169e-01 -1.08756030e+00 -4.15339679e-01
1.15834808e+00 -1.18280900e+00 -9.03403699e-01 5.92902243e-01
-1.26439917e+00 1.14699267e-01 -1.37671173e-01 2.93063611e-01
-1.07438132e-01 7.14047015e-01 -4.16354299e-01 -7.35518277e-01
-4.98550594e-01 -9.87126648e-01 1.17495918e+00 6.30622268e-01
-3.24793965e-01 -6.02367759e-01 -1.45918936e-01 1.15596265e-01
2.04132140e-01 1.91815738e-02 3.62048060e-01 -7.02777624e-01
-4.29095894e-01 -4.60506320e-01 -5.76016366e-01 1.65923819e-01
8.55553299e-02 -2.73235708e-01 -1.12933791e+00 -5.64819336e-01
-6.27741814e-01 1.49930194e-01 9.16221917e-01 2.23658532e-01
9.44743991e-01 -2.24286377e-01 -7.58795381e-01 4.91675943e-01
1.13528669e+00 -1.21461578e-01 5.54896057e-01 7.25510240e-01
8.38704646e-01 6.28561914e-01 3.67706865e-01 3.28048080e-01
4.49174851e-01 1.15802360e+00 1.30273521e-01 -3.04000288e-01
-4.35418814e-01 -4.83929694e-01 3.51730257e-01 1.79202810e-01
-1.37236327e-01 5.26837669e-02 -7.30707884e-01 7.34759092e-01
-1.93719125e+00 -1.22129595e+00 6.09683283e-02 2.50962210e+00
4.19065922e-01 -1.84301287e-01 6.52436376e-01 -8.69914424e-03
1.31778407e+00 -2.62052119e-01 -5.42022288e-01 2.08443388e-01
-3.55012953e-01 -2.33772814e-01 7.23304749e-01 2.63158262e-01
-1.55999708e+00 9.56740677e-01 6.01126909e+00 5.37930667e-01
-8.51979554e-01 4.21801805e-02 4.69132751e-01 7.81812370e-02
4.62966591e-01 -4.74644333e-01 -1.44089735e+00 6.85211062e-01
4.67495769e-01 -1.04919575e-01 1.28466621e-01 1.07975256e+00
-1.43746287e-01 5.04697487e-02 -1.43179917e+00 1.67469049e+00
3.52661699e-01 -1.01543081e+00 3.98085043e-02 2.64069773e-02
4.64991033e-01 -3.91401649e-01 1.89051762e-01 9.97047871e-02
1.61846697e-01 -1.02232683e+00 8.68432879e-01 4.12374794e-01
7.71995604e-01 -6.96999967e-01 7.70792127e-01 -4.63977344e-02
-1.64439416e+00 -2.17990175e-01 -7.38060057e-01 1.89688027e-01
2.50801474e-01 2.88038496e-02 -6.77956939e-01 1.95699394e-01
1.13133955e+00 7.66098559e-01 -1.06297576e+00 1.43718708e+00
-1.86999664e-01 -4.87317182e-02 -2.39257500e-01 8.21275339e-02
-1.61853954e-01 3.45798805e-02 5.01448750e-01 1.48813629e+00
3.71020496e-01 -2.06240237e-01 2.09584504e-01 8.70867610e-01
1.04824334e-01 -7.77295604e-02 -4.49040592e-01 3.95555139e-01
5.19695044e-01 1.25910974e+00 -4.76279616e-01 -5.61372817e-01
-5.14752388e-01 1.55466259e+00 3.85421246e-01 4.66231316e-01
-8.50426853e-01 -3.73588093e-02 1.03175211e+00 1.58486739e-01
5.39583862e-01 -2.51142830e-01 -4.82133068e-02 -1.25458002e+00
4.37548719e-02 -7.38350034e-01 5.34098804e-01 -5.29167235e-01
-1.65622818e+00 4.95582163e-01 -1.16482385e-01 -1.23808336e+00
4.74731997e-02 -6.87736392e-01 -3.60081106e-01 8.97336245e-01
-1.26262641e+00 -1.42165399e+00 -7.58890808e-01 7.33213544e-01
4.52060163e-01 -2.87753046e-01 6.76633060e-01 7.30794907e-01
-9.79168594e-01 1.30264127e+00 -2.28086665e-01 8.07773590e-01
1.06464732e+00 -1.12391663e+00 7.68224716e-01 1.20786357e+00
9.92364138e-02 8.24706376e-01 5.14554858e-01 -6.06416821e-01
-9.09177005e-01 -1.21250594e+00 9.12100077e-01 -7.13970244e-01
3.73118281e-01 -4.63893056e-01 -7.24579692e-01 6.09498560e-01
1.24268703e-01 1.29674822e-01 7.61671722e-01 -1.72314078e-01
-5.69837332e-01 -2.22882465e-01 -9.68991876e-01 7.32478440e-01
1.31137037e+00 -4.79392678e-01 -4.65624332e-01 1.11010382e-02
8.95537585e-02 -6.35853186e-02 -5.58004737e-01 1.21518217e-01
8.34007561e-01 -8.13362777e-01 1.66037619e+00 -7.12920070e-01
-2.01172635e-01 -6.47121131e-01 -1.62370667e-01 -9.99749780e-01
-6.78547323e-01 -1.32700846e-01 1.08832426e-01 1.39833724e+00
-2.41493076e-01 -5.19593358e-01 6.90188825e-01 9.38944221e-01
5.01248479e-01 3.17374915e-02 -1.04490697e+00 -8.60093296e-01
-6.81489632e-02 2.27109566e-01 6.84358597e-01 5.79559445e-01
-2.35759705e-01 3.22535455e-01 -6.29696548e-01 5.03366590e-01
1.09548116e+00 -3.18230242e-02 1.26111197e+00 -1.19115424e+00
-8.47926661e-02 -4.78554338e-01 -7.55619943e-01 -1.37464976e+00
7.63257667e-02 -7.49783158e-01 -8.71282369e-02 -1.13207626e+00
6.11815393e-01 -3.94330710e-01 -2.99040943e-01 3.25120658e-01
-5.18533945e-01 4.77308720e-01 3.91249388e-01 3.21206510e-01
-7.23835886e-01 4.52433109e-01 8.87020290e-01 -3.55652153e-01
2.13089570e-01 -4.89034429e-02 -6.21930242e-01 7.51807213e-01
6.55631006e-01 -1.78157136e-01 2.93497562e-01 -4.62782085e-01
-1.70701504e-01 -6.18605435e-01 1.20499671e+00 -1.27936780e+00
5.00932276e-01 3.26036751e-01 1.23045075e+00 -6.05134308e-01
2.95557201e-01 -8.71760666e-01 1.84645429e-01 5.36408782e-01
-1.84515983e-01 3.71878624e-01 9.22719389e-02 5.83717883e-01
-4.18609306e-02 -1.61563352e-01 1.04134715e+00 -2.32755318e-01
-1.06991374e+00 4.43411201e-01 -1.28433555e-01 -2.08142519e-01
1.03530228e+00 -4.62490231e-01 -2.37952963e-01 -1.60017028e-01
-5.31343102e-01 1.46417707e-01 6.75107956e-01 5.11758447e-01
6.22555435e-01 -1.70678079e+00 -6.60506666e-01 3.70339781e-01
2.86175430e-01 -5.70820391e-01 3.20597202e-01 6.85825467e-01
-1.54816955e-01 4.07866925e-01 -4.66636628e-01 -8.00909758e-01
-1.64472473e+00 7.81916797e-01 6.76764190e-01 -3.23555581e-02
-7.95700967e-01 9.52333212e-01 5.03242791e-01 3.31067643e-03
4.34070051e-01 1.66779831e-01 -3.23743880e-01 -9.01395641e-03
9.65471029e-01 4.86269385e-01 -4.57530439e-01 -1.20265806e+00
-6.01468384e-01 9.41755891e-01 -8.74747559e-02 5.75175844e-02
7.03553557e-01 -4.45095211e-01 1.81032866e-01 -1.73883095e-01
1.14821088e+00 -1.05226472e-01 -1.25703859e+00 -5.19148648e-01
-1.60822477e-02 -8.40262473e-01 -5.04287422e-01 -3.30933422e-01
-8.51990044e-01 9.56114054e-01 1.13053000e+00 -2.39720851e-01
6.34887457e-01 -5.90259917e-02 1.00197542e+00 2.49061137e-01
3.65092397e-01 -1.24699450e+00 3.14562798e-01 1.04527868e-01
5.90751886e-01 -1.51885486e+00 -8.47228393e-02 -1.80602625e-01
-6.81397140e-01 1.03631258e+00 1.02898431e+00 -2.18734458e-01
3.46176922e-01 -2.55704056e-02 -3.48570421e-02 3.29753548e-01
1.18539408e-01 -5.74053884e-01 4.95733649e-01 1.02555037e+00
2.64774170e-02 1.64170912e-03 1.35540545e-01 7.85441697e-01
-2.06741199e-01 -5.45300782e-01 -3.35501656e-02 3.20903331e-01
-4.66924280e-01 -1.11339211e+00 -6.76366806e-01 -1.14263808e-02
-1.84447467e-01 5.58012240e-02 -3.98420602e-01 7.67022669e-01
4.04974937e-01 7.13857353e-01 2.15930939e-01 -2.59626985e-01
4.46881831e-01 -7.14279562e-02 6.28106773e-01 -1.56548336e-01
-5.53396881e-01 -1.05191983e-01 -1.31515980e-01 -5.76230884e-01
-4.70312953e-01 -8.23091447e-01 -7.57254422e-01 -3.11298639e-01
-2.94156015e-01 -1.79177448e-01 2.28364632e-01 5.59447229e-01
2.52263606e-01 2.79433757e-01 1.70384094e-01 -9.48145330e-01
-5.09331346e-01 -7.71277249e-01 -3.80747646e-01 8.53433192e-01
2.94939369e-01 -9.36516643e-01 -1.71094183e-02 2.08224818e-01] | [14.753361701965332, 0.853361964225769] |
e3198613-194d-47f0-83e3-6e148c8279ce | variational-deep-logic-network-for-joint | null | null | https://aclanthology.org/2021.cl-4.26 | https://aclanthology.org/2021.cl-4.26.pdf | Variational Deep Logic Network for Joint Inference of Entities and Relations | Abstract Currently, deep learning models have been widely adopted and achieved promising results on various application domains. Despite their intriguing performance, most deep learning models function as black boxes, lacking explicit reasoning capabilities and explanations, which are usually essential for complex problems. Take joint inference in information extraction as an example. This task requires the identification of multiple structured knowledge from texts, which is inter-correlated, including entities, events, and the relationships between them. Various deep neural networks have been proposed to jointly perform entity extraction and relation prediction, which only propagate information implicitly via representation learning. However, they fail to encode the intensive correlations between entity types and relations to enforce their coexistence. On the other hand, some approaches adopt rules to explicitly constrain certain relational facts, although the separation of rules with representation learning usually restrains the approaches with error propagation. Moreover, the predefined rules are inflexible and might result in negative effects when data is noisy. To address these limitations, we propose a variational deep logic network that incorporates both representation learning and relational reasoning via the variational EM algorithm. The model consists of a deep neural network to learn high-level features with implicit interactions via the self-attention mechanism and a relational logic network to explicitly exploit target interactions. These two components are trained interactively to bring the best of both worlds. We conduct extensive experiments ranging from fine-grained sentiment terms extraction, end-to-end relation prediction, to end-to-end event extraction to demonstrate the effectiveness of our proposed method. | ['Sinno Jialin Pan', 'Wenya Wang'] | null | null | null | null | cl-acl-2021-12 | ['relational-reasoning'] | ['natural-language-processing'] | [-1.04544401e-01 4.88342285e-01 -4.36091721e-01 -6.45434141e-01
-3.18322957e-01 -2.67428547e-01 7.69708753e-01 2.14042410e-01
-2.26562604e-01 6.97879732e-01 2.12953001e-01 -2.12113336e-01
-3.62484485e-01 -1.16620219e+00 -8.40731204e-01 -5.52229166e-01
4.98126745e-02 5.55934012e-01 1.07442982e-01 -2.10839987e-01
-9.36448053e-02 6.56254292e-02 -1.38711655e+00 4.36958522e-01
1.09744978e+00 1.10555673e+00 -1.37797743e-01 2.02437360e-02
-4.42691296e-01 1.21593118e+00 -3.17183942e-01 -9.59402084e-01
-1.74245328e-01 -1.88528016e-01 -7.48487949e-01 -2.21824184e-01
-5.01193941e-01 -2.27989167e-01 -5.18106461e-01 1.18218017e+00
1.58458933e-01 1.77385628e-01 7.18413651e-01 -1.19960618e+00
-1.01906395e+00 1.14397240e+00 -5.01216948e-01 -1.74109973e-02
1.66285247e-01 1.24006681e-01 1.45165551e+00 -7.97659338e-01
2.48849660e-01 1.24033678e+00 4.66615587e-01 2.99919069e-01
-1.04464889e+00 -8.23776364e-01 5.14065683e-01 4.47672367e-01
-1.29982138e+00 -1.93295106e-01 9.92857456e-01 -1.81855768e-01
1.22932327e+00 1.06934100e-01 5.21034777e-01 1.23194015e+00
5.66944443e-02 9.31076586e-01 5.37387192e-01 -7.49035403e-02
-5.43365581e-03 4.59236741e-01 2.84263492e-01 6.20141685e-01
3.17472875e-01 3.08546964e-02 -4.22006518e-01 -5.50324423e-03
5.05904555e-01 2.42238462e-01 -2.90086508e-01 -1.04384989e-01
-9.20378089e-01 8.31852913e-01 7.75370836e-01 2.42418185e-01
-4.49151218e-01 -4.18175645e-02 4.85288143e-01 -3.00249103e-02
3.77804309e-01 2.97351331e-01 -8.33706021e-01 2.19871908e-01
-5.72903037e-01 1.98212698e-01 8.06038320e-01 7.53582537e-01
7.22514927e-01 -2.29017660e-01 -2.74675637e-01 7.68154442e-01
7.18783796e-01 1.12338230e-01 4.21287924e-01 -4.96322423e-01
5.86992681e-01 1.14936078e+00 -9.73702874e-03 -1.39001095e+00
-3.36345881e-01 -5.89709520e-01 -1.09793210e+00 -3.09711307e-01
-3.02018318e-02 -2.93233424e-01 -6.55876756e-01 1.94932902e+00
5.14132082e-01 3.97667378e-01 1.78323612e-01 8.66662800e-01
1.21870208e+00 8.09865236e-01 3.22236687e-01 -3.58609378e-01
1.34051192e+00 -8.44655275e-01 -1.16408145e+00 -6.92827627e-02
5.91583908e-01 -2.81475663e-01 8.40184510e-01 3.41861159e-01
-9.58558440e-01 -3.49237084e-01 -1.05503917e+00 -3.35115612e-01
-4.74414408e-01 -1.43400999e-02 9.86170650e-01 1.99379966e-01
-2.82245308e-01 5.23504555e-01 -8.96601737e-01 1.51837483e-01
5.55584729e-01 5.84826887e-01 -1.23968378e-01 3.20949584e-01
-1.87858868e+00 8.82624328e-01 6.21959388e-01 5.92422247e-01
-4.38619077e-01 -4.33208019e-01 -8.74761581e-01 5.17272949e-01
6.38945401e-01 -7.66269207e-01 1.07073116e+00 -9.11184669e-01
-1.52715611e+00 5.96144199e-01 -2.75268741e-02 -4.84103382e-01
4.60299313e-01 -4.81991082e-01 -5.85839331e-01 -3.07791650e-01
-5.68785407e-02 2.36495838e-01 4.22981977e-01 -1.10721040e+00
-6.16573811e-01 -1.40983000e-01 3.51838231e-01 2.03677222e-01
-4.42833543e-01 -8.26830715e-02 -5.51041305e-01 -5.02476335e-01
6.57434687e-02 -4.35362190e-01 -1.53446957e-01 -3.83202165e-01
-7.38463044e-01 -7.32127488e-01 6.53254271e-01 -3.65695506e-01
1.37162030e+00 -1.95661843e+00 2.74035901e-01 1.43085822e-01
2.06268474e-01 1.96318030e-01 3.02412063e-01 6.82736114e-02
-1.29451707e-01 2.55849957e-01 -1.20920599e-01 -3.16543996e-01
2.77126044e-01 3.56112927e-01 -6.94957018e-01 1.56225964e-01
4.87444252e-01 1.11706567e+00 -8.84358227e-01 -6.56752229e-01
8.84539783e-02 5.84471226e-01 -6.55385196e-01 3.34199637e-01
-6.67929709e-01 2.10972011e-01 -7.08536267e-01 4.29034621e-01
6.02603137e-01 -5.35006642e-01 4.29367155e-01 -4.61105645e-01
2.93249846e-01 6.30428672e-01 -1.21626830e+00 1.34155762e+00
-3.65291774e-01 2.31322706e-01 -1.07759297e-01 -1.30816913e+00
7.83661902e-01 4.54827785e-01 3.25776696e-01 -4.39371735e-01
4.62371409e-01 -2.58578993e-02 2.24031135e-02 -7.73131251e-01
2.00208127e-01 -3.16367328e-01 7.17911404e-03 1.03818983e-01
6.00729808e-02 2.14251816e-01 -8.42801109e-03 2.47725368e-01
8.42073143e-01 2.24075362e-01 2.81280071e-01 1.47143617e-01
6.18549168e-01 -1.47040665e-01 9.84973371e-01 3.59627247e-01
1.37573957e-01 2.31884405e-01 8.51948798e-01 -3.20228457e-01
-3.52894485e-01 -9.56320703e-01 -1.36741355e-01 8.80558968e-01
3.96643460e-01 -5.17415166e-01 -2.22031251e-01 -1.03202999e+00
-1.21982843e-01 9.11905110e-01 -7.74916053e-01 -4.39186305e-01
-2.72776186e-01 -1.05846679e+00 3.32241416e-01 4.75242317e-01
6.28642142e-01 -1.18315995e+00 -2.10182548e-01 1.82444349e-01
-4.20178115e-01 -1.18963158e+00 3.80220674e-02 4.01932359e-01
-5.58635592e-01 -1.11018562e+00 2.87845731e-02 -6.47213221e-01
5.60757756e-01 -3.58746707e-01 1.14300621e+00 1.51573703e-01
2.52257288e-01 -1.86515272e-01 -1.75765857e-01 -4.19215649e-01
-5.07964715e-02 1.90660387e-01 1.80429555e-02 3.31164271e-01
6.39246285e-01 -7.41771162e-01 -5.66897154e-01 1.33821949e-01
-9.43913162e-01 1.45843640e-01 7.89368868e-01 1.01268852e+00
5.71027935e-01 3.49214703e-01 6.08590186e-01 -1.26694632e+00
5.60935974e-01 -8.61995876e-01 -4.62652683e-01 3.72200876e-01
-6.41622961e-01 2.20000491e-01 6.86189771e-01 -4.50655580e-01
-1.40694129e+00 -2.15734258e-01 -9.00538489e-02 -4.67147559e-01
-7.95948654e-02 1.02007329e+00 -6.72372818e-01 6.95145130e-01
3.68952274e-01 1.11912534e-01 -4.05241251e-01 -1.55525774e-01
4.82130647e-01 5.45458376e-01 3.88267845e-01 -7.28739440e-01
6.56681597e-01 2.83809721e-01 -2.75859803e-01 -1.96430683e-01
-1.46042538e+00 4.74377424e-02 -3.81297201e-01 2.30745494e-01
8.86901855e-01 -9.21540499e-01 -8.88316631e-01 5.20566888e-02
-1.38937140e+00 -2.80202162e-02 -2.32203171e-01 5.10426521e-01
-1.19300477e-01 1.49432495e-01 -5.76461971e-01 -8.52567852e-01
-3.37948203e-01 -9.84920561e-01 7.14704573e-01 4.88399416e-01
-2.19533950e-01 -1.04992723e+00 -1.12880290e-01 2.25965112e-01
1.18954316e-01 1.37883842e-01 1.17197835e+00 -9.10862446e-01
-7.55705595e-01 -1.32315204e-01 -4.31567073e-01 1.23374097e-01
2.05096766e-01 2.37355351e-01 -9.22282517e-01 3.33496392e-01
-4.79209656e-03 -4.01743114e-01 1.01062250e+00 2.52495944e-01
1.27022624e+00 -5.51442146e-01 -5.12017608e-01 5.75513899e-01
1.12653041e+00 1.32373154e-01 5.92766941e-01 1.21993385e-01
8.08004022e-01 8.15720916e-01 4.96252596e-01 4.68075067e-01
7.51933932e-01 3.89185846e-01 6.71914041e-01 7.45658651e-02
3.36296499e-01 -3.51422787e-01 2.70832211e-01 8.46187234e-01
-2.12335035e-01 -2.75784016e-01 -6.86931252e-01 4.01716888e-01
-2.21253347e+00 -1.05280888e+00 -9.30683613e-02 1.80013311e+00
1.24283469e+00 5.49182177e-01 -3.48929971e-01 4.80705239e-02
5.01923919e-01 1.58325121e-01 -6.94387019e-01 -1.78910911e-01
-1.71080679e-01 7.60753676e-02 -3.08476854e-02 2.95501709e-01
-1.16189885e+00 1.00297499e+00 4.98780584e+00 7.54514337e-01
-1.02554357e+00 2.95811500e-02 6.05945647e-01 -4.41366211e-02
-7.15794802e-01 1.04849279e-01 -7.16445208e-01 4.01701212e-01
7.18160510e-01 6.55101240e-02 2.25169733e-01 6.29580557e-01
-3.48066539e-02 1.45774975e-01 -1.14727271e+00 8.70551229e-01
-4.44543600e-01 -1.36821926e+00 1.66653529e-01 -1.26958534e-01
5.95705390e-01 -1.14618681e-01 6.96045160e-02 8.21141839e-01
6.56429350e-01 -1.20726824e+00 5.59136331e-01 6.67728364e-01
3.14443976e-01 -8.28187704e-01 8.52145612e-01 4.94033098e-01
-1.12562335e+00 -4.57726605e-02 -2.89225698e-01 -2.20199078e-01
3.39899808e-02 9.45905566e-01 -5.23529589e-01 9.23846424e-01
6.38291478e-01 9.33535159e-01 -1.61186799e-01 4.43766266e-01
-6.50091827e-01 6.03490829e-01 -3.73856246e-01 -3.15170497e-01
8.77028257e-02 -2.51839787e-01 2.20452473e-01 1.07255757e+00
-3.79042737e-02 3.74968261e-01 3.63909733e-03 1.34351909e+00
-3.42624903e-01 7.43038431e-02 -4.49620515e-01 -1.47783428e-01
4.27215636e-01 1.30203557e+00 -4.59245563e-01 -4.26339000e-01
-6.07438266e-01 5.98757505e-01 7.33992875e-01 5.09265542e-01
-1.27070534e+00 -4.23357457e-01 5.98470688e-01 -1.61099747e-01
3.07075858e-01 1.16312601e-01 -5.74473798e-01 -1.45501745e+00
-1.72370523e-02 -6.25097692e-01 4.91081387e-01 -4.87721622e-01
-1.51895726e+00 4.56379861e-01 -1.03277445e-01 -9.43294764e-01
-1.35993510e-01 -4.98300284e-01 -6.90720201e-01 7.49003649e-01
-1.56608665e+00 -1.12038219e+00 -8.42871219e-02 5.64758480e-01
2.42755175e-01 1.57658514e-02 7.64043689e-01 4.67077851e-01
-1.10170710e+00 6.36453450e-01 -3.03418338e-01 4.37833518e-01
4.64199215e-01 -1.17262793e+00 -2.07567930e-01 6.11599505e-01
2.65581161e-01 8.68544936e-01 6.90357566e-01 -6.59276128e-01
-1.15660965e+00 -1.16844440e+00 8.85304987e-01 -5.06872535e-01
6.99029863e-01 -2.90515810e-01 -1.14580894e+00 1.01431310e+00
3.08617830e-01 2.01335683e-01 7.41352141e-01 5.87141216e-01
-4.14986640e-01 -1.06479533e-01 -7.61741698e-01 6.66936219e-01
9.13694263e-01 -4.79694217e-01 -9.23980534e-01 3.32168162e-01
9.10897851e-01 -4.35517877e-01 -7.60260880e-01 8.21633041e-01
3.40484023e-01 -8.73500347e-01 8.46392691e-01 -8.08501542e-01
9.22845662e-01 -3.46510053e-01 -6.57078326e-02 -1.05569422e+00
-2.12510660e-01 -3.58144492e-01 -6.99729323e-01 1.61101639e+00
7.86306739e-01 -5.56119084e-01 5.90891242e-01 9.09722805e-01
9.89403278e-02 -1.19310081e+00 -5.79049766e-01 -2.74480045e-01
-9.57928747e-02 -6.14555836e-01 7.88052559e-01 1.26132691e+00
3.02542269e-01 9.65353310e-01 -3.02463979e-01 5.28102815e-01
2.83603311e-01 4.35798198e-01 5.07024467e-01 -1.26670313e+00
-4.43922222e-01 -6.07748508e-01 -3.26829739e-02 -1.05576968e+00
5.20658374e-01 -9.71480727e-01 -1.63775340e-01 -1.70948303e+00
2.56789625e-01 -5.28264642e-01 -6.70227230e-01 5.63187659e-01
-4.99124765e-01 -4.12569255e-01 -4.01790261e-01 5.91359958e-02
-7.19897270e-01 1.01939392e+00 1.04607916e+00 -2.88453668e-01
-2.55726099e-01 3.84673066e-02 -8.43043089e-01 9.22735989e-01
6.18292749e-01 -4.27048534e-01 -6.53629005e-01 -5.18635750e-01
7.52881765e-01 -1.92487407e-02 2.73496240e-01 -4.63758588e-01
3.71898025e-01 -2.90766478e-01 4.03970689e-01 -6.09041333e-01
2.97586620e-01 -9.90730822e-01 4.59054261e-02 5.97806349e-02
-4.23564315e-01 -3.01550776e-01 8.18559527e-03 7.87383199e-01
-4.90722746e-01 -2.23965924e-02 3.99010658e-01 -9.94259417e-02
-5.01822293e-01 4.88770485e-01 1.31095275e-01 1.57824248e-01
9.58561838e-01 2.39058986e-01 -1.42598778e-01 -2.21634924e-01
-7.29147971e-01 6.16102397e-01 -1.67507738e-01 3.72669727e-01
5.53346455e-01 -1.25760925e+00 -5.31428337e-01 1.17008843e-01
-1.40345082e-01 6.12287521e-01 3.38942051e-01 8.80333066e-01
-4.03498076e-02 2.09317699e-01 2.77365088e-01 -4.40878034e-01
-6.92761183e-01 8.10592055e-01 5.11910081e-01 -7.23911703e-01
-3.99658531e-01 9.94903147e-01 3.86482984e-01 -6.21047080e-01
3.87897462e-01 -5.61366379e-01 -6.02979779e-01 1.77328244e-01
1.84523195e-01 -1.99722409e-01 -9.09805149e-02 -2.04167023e-01
-4.00734037e-01 1.79323003e-01 -2.60203898e-01 1.41426355e-01
1.38080931e+00 -7.81294629e-02 -2.63726443e-01 4.24867362e-01
8.43697309e-01 -2.62280181e-02 -1.08232009e+00 -5.03934920e-01
5.62803410e-02 2.85254121e-02 1.55125931e-01 -7.37835705e-01
-1.21134496e+00 1.01284742e+00 -6.23276643e-02 4.48230743e-01
1.04176712e+00 5.41221760e-02 8.32547784e-01 4.02665466e-01
3.95122878e-02 -8.52404296e-01 -1.25254750e-01 5.56007564e-01
6.68823779e-01 -1.38284421e+00 -4.19121310e-02 -5.68866670e-01
-6.52243555e-01 9.52869594e-01 9.10671413e-01 -6.85069785e-02
1.01933920e+00 2.79098570e-01 -2.38266245e-01 -2.89447725e-01
-1.12718391e+00 -1.84258550e-01 3.97494286e-01 1.69283435e-01
5.65367401e-01 1.63628662e-03 -2.17693120e-01 1.54265487e+00
-4.86215577e-02 -2.89501324e-02 -8.47328678e-02 4.47822869e-01
-6.21350408e-02 -8.39234591e-01 5.19027822e-02 4.98757273e-01
-6.52397871e-01 -1.41699746e-01 -2.89738476e-01 7.27506042e-01
4.18525100e-01 8.12419355e-01 -6.27484992e-02 -4.59862769e-01
3.63104045e-01 4.64710444e-02 9.05846134e-02 -5.09554029e-01
-6.05926991e-01 -1.59700751e-01 2.34113019e-02 -3.11599791e-01
-5.61448455e-01 -4.39671159e-01 -1.69395185e+00 -2.42539018e-01
-5.42975545e-01 2.87992001e-01 1.14975691e-01 1.35898864e+00
3.47939670e-01 1.03725123e+00 3.92486125e-01 -2.27679551e-01
-4.88460630e-01 -8.94608557e-01 -4.34440672e-01 3.95176262e-01
1.94451109e-01 -8.45712364e-01 -2.02753812e-01 -1.14549063e-01] | [9.3324613571167, 8.458518028259277] |
bad925de-4f43-4e05-a874-dd780bf9af79 | histogram-layers-for-texture-analysis | 2001.00215 | null | https://arxiv.org/abs/2001.00215v9 | https://arxiv.org/pdf/2001.00215v9.pdf | Histogram Layers for Texture Analysis | We present a histogram layer for artificial neural networks (ANNs). An essential aspect of texture analysis is the extraction of features that describe the distribution of values in local spatial regions. The proposed histogram layer directly computes the spatial distribution of features for texture analysis and parameters for the layer are estimated during backpropagation. We compare our method with state-of-the-art texture encoding methods such as the Deep Encoding Network Pooling (DEP), Deep Texture Encoding Network (DeepTEN), Fisher Vector convolutional neural network (FV-CNN), and Multi-level Texture Encoding and Representation (MuLTER) on three material/texture datasets: (1) the Describable Texture Dataset (DTD); (2) an extension of the ground terrain in outdoor scenes (GTOS-mobile); (3) and a subset of the Materials in Context (MINC-2500) dataset. Results indicate that the inclusion of the proposed histogram layer improves performance. The source code for the histogram layer is publicly available. | ['Weihuang Xu', 'Joshua Peeples', 'Alina Zare'] | 2020-01-01 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [ 2.69016236e-01 -1.37021571e-01 1.19201116e-01 -6.43471718e-01
-6.74763143e-01 2.98269279e-02 5.89678168e-01 3.51248235e-01
-2.99216509e-01 5.69921732e-01 -3.18273567e-02 -1.74949199e-01
-3.66838962e-01 -1.61436903e+00 -7.87484765e-01 -1.07463789e+00
-5.23640811e-01 4.38309759e-01 3.39030385e-01 -5.00434339e-01
3.59232068e-01 7.36492217e-01 -1.90352452e+00 1.09037077e+00
1.85852543e-01 2.24472046e+00 1.44179717e-01 5.87777853e-01
-5.45354187e-02 9.72219408e-01 -1.37573391e-01 -1.33168427e-02
-5.49661778e-02 2.38936648e-01 -9.60856616e-01 -2.05744490e-01
6.59287572e-01 -3.37633133e-01 -5.19204259e-01 8.21369648e-01
3.30121279e-01 3.64709228e-01 9.27315533e-01 -8.23917031e-01
-7.60980010e-01 2.33957067e-01 -3.07618111e-01 7.56195933e-02
-1.02506854e-01 -2.99110770e-01 5.84497333e-01 -8.18028927e-01
7.33985126e-01 1.37227952e+00 1.12694430e+00 -2.29411960e-01
-9.38430727e-01 3.21951415e-03 -4.46646333e-01 3.82133752e-01
-1.64411390e+00 -1.58243343e-01 5.69910824e-01 -4.95218933e-01
1.36931717e+00 1.81256562e-01 6.72735035e-01 1.20763624e+00
8.95072758e-01 8.16145301e-01 1.50071907e+00 -3.80376816e-01
3.62898886e-01 -4.33643371e-01 4.42874320e-02 9.82395709e-01
-5.82130216e-02 2.87020147e-01 -6.52690053e-01 1.65726483e-01
1.10433245e+00 -3.10618758e-01 2.88113415e-01 6.75435588e-02
-8.72720778e-01 7.41606176e-01 5.94666123e-01 -1.60312485e-02
-6.34797931e-01 3.64503443e-01 4.96855587e-01 6.39005303e-01
1.06409335e+00 -1.84204262e-02 -4.77034897e-01 -3.15550774e-01
-1.04234076e+00 5.43331444e-01 8.86452377e-01 7.33302593e-01
1.12088013e+00 1.66403040e-01 -4.42471802e-01 1.18138194e+00
-5.21053448e-02 8.78627360e-01 2.94871241e-01 -5.99708021e-01
4.27587360e-01 3.31388921e-01 -1.95364535e-01 -1.50939643e+00
-4.09394652e-01 1.23785309e-01 -9.62943137e-01 3.44080091e-01
3.33728194e-01 2.04628468e-01 -1.25279272e+00 1.00986469e+00
-1.21935472e-01 -3.86970103e-01 -1.82963401e-01 8.20935309e-01
1.21246135e+00 7.20166385e-01 -2.45472286e-02 8.41549933e-01
1.27660632e+00 -7.76588202e-01 -4.63182032e-01 3.57482620e-02
2.97374099e-01 -6.71539783e-01 8.06948900e-01 4.64999944e-01
-6.44395292e-01 -8.68836522e-01 -1.10952461e+00 -2.85681725e-01
-1.21029794e+00 3.19465399e-01 1.05332637e+00 4.37750906e-01
-1.11908710e+00 1.09805739e+00 -9.18149173e-01 -5.70684671e-01
7.22936273e-01 4.64199752e-01 -7.31026769e-01 -3.46501507e-02
-1.08455682e+00 7.79766738e-01 5.55413783e-01 5.47819793e-01
-7.74982989e-01 -5.09690903e-02 -1.24238086e+00 2.80263543e-01
-8.20219293e-02 -6.14137165e-02 6.67631328e-01 -1.10542119e+00
-1.63866055e+00 8.51026654e-01 1.14126734e-01 -3.62138689e-01
2.18473271e-01 -3.15927923e-01 -1.40520051e-01 2.45479107e-01
-7.08319321e-02 1.09256721e+00 7.13001370e-01 -8.65858078e-01
-5.77068150e-01 -1.31423727e-01 -1.62924692e-01 -6.31259680e-02
-1.36043027e-01 -4.90665913e-01 -3.23696405e-01 -8.34581017e-01
2.26734817e-01 -7.80295014e-01 1.35553703e-01 9.93896052e-02
-4.91851211e-01 4.20208042e-03 1.05601716e+00 -9.33530033e-01
9.91147041e-01 -2.09622574e+00 -6.64394200e-02 5.47585070e-01
-2.19651833e-01 -2.47122034e-01 -3.02996099e-01 4.32806045e-01
8.08952302e-02 9.70746018e-03 -2.32304603e-01 -3.05244718e-02
1.94234088e-01 4.42713469e-01 -6.59890175e-02 4.59779710e-01
4.27638829e-01 9.21127260e-01 -2.94327646e-01 -4.28229004e-01
6.29676163e-01 6.63848877e-01 -5.61512887e-01 2.68013980e-02
-2.57226586e-01 -9.13382694e-02 -4.85814959e-01 1.30176234e+00
9.68759477e-01 -1.05366735e-02 -1.26176745e-01 -6.13963902e-01
-2.19766796e-01 -1.40261725e-01 -9.27555203e-01 1.31993437e+00
-4.20120239e-01 1.00913382e+00 -4.46093410e-01 -9.33317006e-01
1.26154554e+00 5.71117960e-02 3.50451112e-01 -1.20339406e+00
2.95269698e-01 2.34291300e-01 -3.55626106e-01 -7.40860641e-01
1.02625334e+00 3.44101936e-01 -3.87251765e-01 1.17245659e-01
5.57500601e-01 -4.91477996e-02 7.30269924e-02 -8.17667365e-01
8.72269034e-01 5.96027613e-01 6.55262824e-03 -6.66279912e-01
1.43017203e-01 8.75381529e-02 1.94578338e-02 1.02017021e+00
-3.87731451e-03 9.28720772e-01 7.57221818e-01 -1.24864018e+00
-1.39546001e+00 -1.02121794e+00 -5.03709912e-01 1.02742875e+00
-1.40820131e-01 -1.05218202e-01 -7.24068105e-01 -1.38626277e-01
4.49814230e-01 -4.52585928e-02 -1.56191838e+00 2.21990183e-01
-3.70598167e-01 -8.92170846e-01 6.75572753e-01 7.54061639e-01
1.24426520e+00 -1.47806382e+00 -5.64410031e-01 1.95521995e-01
-1.80747539e-01 -1.03232574e+00 4.12585765e-01 8.20382655e-01
-6.68927848e-01 -7.56072104e-01 -4.78136778e-01 -6.10063195e-01
3.27565700e-01 -4.09715772e-01 9.59688008e-01 -2.69787431e-01
-2.39738271e-01 1.02681331e-02 -4.17421937e-01 -2.37326950e-01
1.03559986e-01 3.56981546e-01 -5.38701892e-01 8.35548118e-02
5.52591085e-01 -3.40521485e-01 -3.23963284e-01 1.86299592e-01
-9.31406796e-01 2.59582430e-01 5.01355350e-01 1.17075610e+00
1.05162466e+00 4.61658955e-01 -1.03643127e-01 -8.05347681e-01
3.55451196e-01 -5.60303926e-01 -5.17591178e-01 1.91872314e-01
4.38212492e-02 1.44199252e-01 2.78715402e-01 -5.39620481e-02
-1.17842710e+00 -5.61153069e-02 -4.39296097e-01 -1.97444543e-01
-5.07439017e-01 9.58288908e-01 1.70348480e-01 -5.09252012e-01
5.83673537e-01 1.28883526e-01 -1.00216605e-01 -4.84477639e-01
-5.15732020e-02 6.98801041e-01 5.52753091e-01 -1.04440618e+00
8.36762860e-02 5.10446727e-01 2.29439527e-01 -1.06079638e+00
-6.64535224e-01 -3.65696847e-02 -9.55899119e-01 -3.33040893e-01
9.63991642e-01 -7.53241539e-01 -5.83489478e-01 1.11359513e+00
-6.51476562e-01 -1.13659310e+00 -2.26520255e-01 1.97240889e-01
-9.43136334e-01 -2.35216901e-01 -8.93741548e-01 -5.40812612e-01
-2.10117832e-01 -1.25444591e+00 1.79127824e+00 1.97164163e-01
2.83113774e-02 -1.26312876e+00 -2.01682955e-01 -1.61283866e-01
9.87629950e-01 1.03059494e+00 1.18742585e+00 8.29502344e-02
-3.58286142e-01 -2.07231477e-01 -5.94988227e-01 2.21663371e-01
1.30315730e-02 1.88618436e-01 -1.38637650e+00 8.44708607e-02
-6.80117607e-01 -6.40043676e-01 1.34671998e+00 8.93787086e-01
1.83874905e+00 -2.00328916e-01 6.08613603e-02 9.87543285e-01
1.79251623e+00 1.75480396e-01 1.16262484e+00 8.94461513e-01
6.10848010e-01 5.53697467e-01 4.84300822e-01 6.27643406e-01
3.72385085e-01 5.76457977e-01 4.07489836e-01 -4.90480512e-01
-8.96153748e-02 1.32799178e-01 -5.21266125e-02 4.41903621e-01
-6.77267194e-01 -3.82496685e-01 -1.02440238e+00 2.73180127e-01
-1.90882194e+00 -8.27014744e-01 6.00766987e-02 1.86092496e+00
4.63910997e-01 1.61965445e-01 -2.20530212e-01 3.00118268e-01
2.94529349e-01 2.55626500e-01 1.78461969e-02 -1.02030468e+00
-6.27992988e-01 9.40880358e-01 7.20166683e-01 2.70900756e-01
-2.05352116e+00 1.33815312e+00 6.68357849e+00 1.31034720e+00
-1.52219939e+00 -2.50744700e-01 1.14668727e+00 4.52929795e-01
2.91845322e-01 -5.28811216e-01 -3.80204201e-01 2.17510283e-01
1.11808932e+00 7.03574181e-01 4.45911050e-01 1.00260663e+00
-2.23519295e-01 -6.05365872e-01 -5.50797462e-01 7.24248111e-01
-1.40499339e-01 -1.49315023e+00 8.34182873e-02 8.55827108e-02
5.00416160e-01 3.29066604e-01 3.40399146e-01 4.17194605e-01
-9.04093124e-03 -1.34617639e+00 1.05630183e+00 8.92822742e-01
1.26602912e+00 -8.21328878e-01 1.37439823e+00 -6.56298101e-01
-1.40321505e+00 5.67847267e-02 -8.85941625e-01 -6.06355891e-02
-3.88447195e-01 6.38825893e-01 -2.99917430e-01 6.40600145e-01
1.39878297e+00 8.66367280e-01 -1.09843433e+00 7.47434735e-01
7.25549385e-02 6.87176943e-01 -4.06275630e-01 3.11856829e-02
6.99076772e-01 1.11570723e-01 -9.54365507e-02 1.61465049e+00
3.53637308e-01 -3.50109071e-01 -4.48218687e-03 6.94906592e-01
3.71366978e-01 -1.26713604e-01 -7.05898643e-01 -2.03452334e-01
7.37265050e-02 1.35938346e+00 -1.16617584e+00 -2.93858945e-01
-1.17603615e-01 8.41162980e-01 3.20234716e-01 5.32182157e-01
-5.38003683e-01 -7.78953671e-01 3.53807688e-01 -7.17683136e-02
6.12910211e-01 -1.39649138e-01 -4.72843945e-01 -6.37333632e-01
-2.14526966e-01 -6.44616127e-01 1.25417948e-01 -1.10332525e+00
-1.29031622e+00 9.08525765e-01 5.53812943e-02 -1.02960455e+00
7.32217059e-02 -1.29700232e+00 -1.60844415e-01 1.00980961e+00
-1.49338531e+00 -1.65664029e+00 -7.18517065e-01 5.57283461e-01
1.96688011e-01 -3.37316453e-01 1.17170227e+00 3.98344129e-01
-2.81172007e-01 4.15729254e-01 1.43000945e-01 2.95260429e-01
3.40415776e-01 -1.32964802e+00 4.31081563e-01 1.91466928e-01
-5.07647514e-01 8.67222026e-02 1.29700691e-01 -5.49490571e-01
-1.42421520e+00 -1.30859911e+00 8.08985293e-01 4.59061600e-02
4.59679216e-01 -6.75754964e-01 -8.66817236e-01 5.76598465e-01
5.86243346e-02 2.04826653e-01 3.53318900e-01 7.25686476e-02
-2.13424534e-01 -1.84684083e-01 -1.25088346e+00 1.96781546e-01
5.58743238e-01 -8.16235483e-01 -2.13376172e-02 1.08131878e-01
-5.50080426e-02 -8.43151987e-01 -1.26632214e+00 5.79906344e-01
1.26753628e+00 -9.87781405e-01 7.76197314e-01 -1.57739446e-01
1.12555552e+00 6.47382215e-02 -8.87777388e-01 -1.02761149e+00
-6.35725439e-01 1.12294041e-01 1.64844826e-01 6.02943361e-01
1.34972677e-01 -5.01960397e-01 7.68760026e-01 -6.31921217e-02
-3.18507373e-01 -9.42488670e-01 -9.72451508e-01 -4.89720762e-01
6.53854087e-02 -4.62551713e-01 6.04276240e-01 8.15766156e-01
-4.37559396e-01 -5.82133651e-01 -2.68217891e-01 -2.43112352e-02
3.41353834e-01 6.22668415e-02 4.39819247e-01 -1.19583571e+00
2.16304865e-02 -4.93199915e-01 -9.55477536e-01 -4.05320793e-01
-1.07490219e-01 -6.62518561e-01 1.79439634e-01 -1.50256288e+00
5.32869548e-02 -6.24576569e-01 -4.21979785e-01 1.08309197e+00
3.56520653e-01 5.37259340e-01 -4.35903877e-01 -1.59287930e-01
-3.45279515e-01 6.08176351e-01 1.23598456e+00 -3.88895333e-01
3.02185893e-01 -5.75730443e-01 -1.32134646e-01 5.02504349e-01
6.29105151e-01 -3.99706393e-01 2.30073761e-02 -4.58732635e-01
1.70772195e-01 -9.67990160e-02 7.08298087e-01 -1.37299240e+00
-8.37908015e-02 -1.51663512e-01 1.07427192e+00 -8.51647973e-01
3.44384789e-01 -4.87338245e-01 2.12320276e-02 1.01067081e-01
-7.50766844e-02 2.11864546e-01 8.11251104e-01 -1.26682715e-02
-6.93631411e-01 2.44284064e-01 5.99819541e-01 1.36481980e-02
-1.26435471e+00 4.30617839e-01 -9.04818356e-01 -5.36696494e-01
2.77930975e-01 -4.88482326e-01 -7.67810404e-01 -1.35093987e-01
-6.65157914e-01 -3.78281027e-01 2.01752186e-01 4.54681486e-01
6.43949747e-01 -1.66863680e+00 -4.83094841e-01 5.35968840e-01
1.89424112e-01 -1.07923865e-01 4.98331934e-01 6.05864942e-01
-1.40185034e+00 4.02890205e-01 -1.12670243e+00 -7.57106245e-01
-5.69820583e-01 -2.19225109e-01 6.26118124e-01 -1.97694480e-01
-6.86184466e-01 8.36790860e-01 1.43324614e-01 -5.85041583e-01
1.86489634e-02 -6.25730753e-01 -3.14478219e-01 1.32236106e-03
3.10798377e-01 2.56626725e-01 5.34413457e-01 -6.56856596e-01
-1.94817379e-01 6.40702248e-01 2.26324216e-01 -1.07688997e-02
1.76359630e+00 2.43245244e-01 -3.21974367e-01 4.35868144e-01
1.39147663e+00 -7.46495008e-01 -1.27215815e+00 -1.75089017e-02
-3.52199748e-02 -2.18639970e-01 5.95541954e-01 -8.69118214e-01
-1.16351426e+00 1.08793259e+00 1.19985819e+00 7.43334144e-02
1.22355151e+00 -6.44233167e-01 6.10426784e-01 7.71525562e-01
5.24144053e-01 -1.50533903e+00 -8.01085755e-02 1.11528635e+00
1.00202584e+00 -1.12920070e+00 -1.06118470e-01 -2.26089045e-01
-3.66519660e-01 1.70729220e+00 4.21741277e-01 -3.75455052e-01
1.20391977e+00 7.35144079e-01 8.26118700e-03 -6.72990978e-01
-6.62145734e-01 -2.68668562e-01 4.17500466e-01 6.08644247e-01
5.99985659e-01 3.70350659e-01 4.57795352e-01 3.69685203e-01
-3.82828087e-01 2.08261125e-02 1.64064422e-01 1.26667881e+00
-4.35550630e-01 -6.16852462e-01 -2.55103379e-01 6.01742625e-01
-4.25342590e-01 -3.94522935e-01 -1.43888593e-01 1.11338234e+00
3.54860365e-01 4.73787665e-01 5.97347081e-01 -7.25044906e-01
1.75729424e-01 -9.88290310e-02 7.34420121e-01 -1.18763812e-01
-4.83538926e-01 7.41702020e-02 3.59064221e-01 -8.75136495e-01
-3.70291024e-01 -3.63705486e-01 -6.03176534e-01 -6.34007573e-01
8.06891322e-02 -3.59626055e-01 7.46271729e-01 7.02736855e-01
7.60472938e-02 8.43041778e-01 1.70180246e-01 -1.42537189e+00
4.01577204e-01 -1.23944163e+00 -1.12151277e+00 2.32523456e-01
5.59967086e-02 -1.18939555e+00 1.15968920e-01 2.14054525e-01] | [10.18730640411377, -0.1943131685256958] |
4fc3a0a3-c499-4633-b54d-aa196cbf351a | the-creative-frontier-of-generative-ai | 2306.03601 | null | https://arxiv.org/abs/2306.03601v1 | https://arxiv.org/pdf/2306.03601v1.pdf | The Creative Frontier of Generative AI: Managing the Novelty-Usefulness Tradeoff | In this paper, drawing inspiration from the human creativity literature, we explore the optimal balance between novelty and usefulness in generative Artificial Intelligence (AI) systems. We posit that overemphasizing either aspect can lead to limitations such as hallucinations and memorization. Hallucinations, characterized by AI responses containing random inaccuracies or falsehoods, emerge when models prioritize novelty over usefulness. Memorization, where AI models reproduce content from their training data, results from an excessive focus on usefulness, potentially limiting creativity. To address these challenges, we propose a framework that includes domain-specific analysis, data and transfer learning, user preferences and customization, custom evaluation metrics, and collaboration mechanisms. Our approach aims to generate content that is both novel and useful within specific domains, while considering the unique requirements of various contexts. | ['Hannah Chang', 'Anirban Mukherjee'] | 2023-06-06 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 3.43059123e-01 2.02381492e-01 -8.15302804e-02 -1.18173763e-01
-6.45636916e-02 -6.13096714e-01 8.05751324e-01 1.30629718e-01
-2.52279073e-01 7.98347890e-01 3.79120797e-01 6.51438609e-02
-2.59628892e-01 -8.35811675e-01 -1.74840346e-01 -1.61490604e-01
4.49575603e-01 4.19127047e-01 -4.90199357e-01 -1.40877336e-01
8.10715079e-01 3.49475235e-01 -1.66510010e+00 4.10280377e-01
1.36594188e+00 9.10326064e-01 5.51353879e-02 2.15890899e-01
-3.77813280e-01 8.73194695e-01 -1.12237549e+00 -5.27052760e-01
-1.04791492e-01 -7.12874591e-01 -7.47835517e-01 1.16357371e-01
8.73171464e-02 -1.42667666e-01 3.75412047e-01 7.84844339e-01
5.75196862e-01 -4.85284887e-02 7.45372951e-01 -1.38517916e+00
-1.54576218e+00 7.50524700e-01 1.05484359e-01 5.86538725e-02
5.73365569e-01 6.23120010e-01 6.95419431e-01 -9.29056525e-01
7.12627947e-01 8.82360160e-01 6.44289553e-01 6.79701924e-01
-1.53154659e+00 -6.39785528e-01 -6.67257756e-02 2.07324810e-02
-1.06378329e+00 -5.10817468e-01 1.00114202e+00 -5.89029670e-01
1.10843027e+00 4.44859743e-01 1.22281957e+00 1.54242015e+00
2.77000934e-01 6.71649873e-01 1.12423635e+00 -3.51679444e-01
7.76712120e-01 6.45189464e-01 -4.29160178e-01 -1.35494754e-01
5.02093017e-01 -1.00757927e-01 -7.75431931e-01 -9.29506123e-02
6.79348707e-01 -1.57882035e-01 -2.17928275e-01 -3.83310318e-01
-1.11528015e+00 7.78175771e-01 9.02002156e-02 5.79836786e-01
-6.52930498e-01 -1.73729122e-01 1.48232222e-01 6.03696227e-01
1.57571599e-01 1.65767574e+00 -1.56769708e-01 -4.51547295e-01
-1.04884267e+00 2.87370890e-01 6.99714124e-01 9.65053082e-01
6.13657832e-01 3.15388113e-01 -4.08708304e-01 8.19056511e-01
2.47972589e-02 2.56425172e-01 1.06764805e+00 -9.43292081e-01
-2.41835311e-01 8.83507073e-01 2.23237559e-01 -8.73273015e-01
-2.02722937e-01 -8.03113580e-01 -4.98531282e-01 2.21888304e-01
-2.09555268e-01 -1.12758286e-01 -7.96714842e-01 1.89481020e+00
-1.71340480e-01 -4.17980045e-01 2.43721530e-01 7.21118748e-01
7.75339007e-01 3.38891864e-01 3.83326471e-01 -4.18796927e-01
8.67401779e-01 -7.03680634e-01 -8.44818592e-01 -4.43074971e-01
5.02104282e-01 -6.47801340e-01 1.72559905e+00 6.31193459e-01
-1.59914017e+00 -3.81553054e-01 -1.10812175e+00 1.67130917e-01
-4.97332245e-01 -1.59866899e-01 6.31626904e-01 8.48525703e-01
-1.15520167e+00 6.13181174e-01 -5.68470806e-02 -5.25502861e-01
2.69981802e-01 1.28820194e-02 -2.34204275e-03 3.54874194e-01
-9.97110367e-01 1.17145574e+00 5.36637843e-01 -2.03704149e-01
-4.34261590e-01 -8.22737873e-01 -3.97640586e-01 1.22827142e-01
1.96898997e-01 -1.10485995e+00 1.18910432e+00 -1.63856232e+00
-1.51509416e+00 6.55793190e-01 4.73337978e-01 -2.71061629e-01
5.37636876e-01 -3.36816087e-02 -6.41274929e-01 -1.39509887e-01
-5.34708761e-02 7.18091369e-01 8.35322022e-01 -1.40627265e+00
-2.25144461e-01 -8.97468850e-02 -8.73241052e-02 2.04342633e-01
-7.31151760e-01 -3.22290629e-01 3.68225485e-01 -8.19892108e-01
3.50941718e-02 -6.64743423e-01 2.77025420e-02 -2.94078261e-01
-1.56591058e-01 -6.86027948e-03 5.08079648e-01 -3.63495529e-01
1.55471516e+00 -1.90108538e+00 1.64723709e-01 2.91843742e-01
2.94712484e-01 1.99587822e-01 -1.19641766e-01 5.77167809e-01
2.44568393e-01 5.77908874e-01 1.30373146e-02 6.01992868e-02
4.45285551e-02 1.29275128e-01 -2.22337455e-01 -3.73058349e-01
5.23470998e-01 1.08223021e+00 -1.04843569e+00 -4.24565703e-01
-2.16557086e-01 2.50973463e-01 -8.45359087e-01 2.18393102e-01
-3.17229629e-01 3.29917997e-01 -2.10472882e-01 9.82982993e-01
5.61074257e-01 -3.64923120e-01 1.73518494e-01 1.52688563e-01
-2.56945133e-01 2.43527561e-01 -8.51444840e-01 1.50617611e+00
-6.36186123e-01 4.85337675e-01 -4.95717615e-01 -4.45911795e-01
1.23009908e+00 2.39224598e-01 1.75016433e-01 -1.04590559e+00
4.85861376e-02 3.70009571e-01 2.76082810e-02 -7.27661252e-01
7.04428017e-01 -3.82936448e-01 1.00745268e-01 7.16758668e-01
7.22769722e-02 -5.81589758e-01 2.24837974e-01 4.25877944e-02
9.18558300e-01 1.78194597e-01 4.92107451e-01 -1.82552651e-01
2.91765891e-02 7.65395835e-02 3.72934729e-01 8.15118492e-01
-1.75622338e-03 4.40813869e-01 4.46596593e-01 -2.78691530e-01
-1.19463646e+00 -1.18462944e+00 1.45691872e-01 7.72600055e-01
-1.61902204e-01 -4.11673337e-01 -5.34585297e-01 -3.43899131e-01
-1.05864562e-01 1.45645535e+00 -6.22277200e-01 -6.59855127e-01
-1.29985362e-01 -2.66696215e-01 3.97894770e-01 4.99341428e-01
4.62365925e-01 -1.58339238e+00 -1.31993783e+00 2.25481376e-01
-1.43990740e-01 -3.31839144e-01 -1.80406272e-01 -1.73619427e-02
-1.09127045e+00 -4.36157614e-01 -6.23122990e-01 -4.25102741e-01
5.40026605e-01 1.26171917e-01 1.48660779e+00 8.06346759e-02
-6.29092008e-02 5.14148474e-01 -4.82138366e-01 -5.65883517e-01
-6.20339036e-01 -7.44938701e-02 -1.99191645e-01 -5.45735657e-01
3.84565353e-01 -8.70362461e-01 -5.54017067e-01 1.67459071e-01
-1.30279982e+00 2.23998129e-01 1.06831288e+00 8.88051033e-01
2.36023024e-01 -1.25423163e-01 1.11025202e+00 -7.24409759e-01
1.48576546e+00 -7.61464894e-01 7.54697025e-02 5.22134364e-01
-1.34051907e+00 -1.59013784e-03 4.84225869e-01 -9.34355021e-01
-1.23947120e+00 -4.10466284e-01 4.80147421e-01 -3.12451005e-01
-5.76747246e-02 8.44605327e-01 -1.61675856e-01 6.13410287e-02
1.13886571e+00 4.08399016e-01 1.86704054e-01 -9.37833935e-02
5.35030246e-01 6.86704576e-01 3.13496560e-01 -6.99158311e-01
1.57608971e-01 -1.93104297e-02 -6.21496499e-01 -6.55427873e-01
-3.18997800e-01 1.87894434e-01 -3.14985365e-01 -3.79438072e-01
3.18915218e-01 -5.15864015e-01 -4.46845859e-01 -9.17562936e-03
-1.00810468e+00 -1.27288252e-01 -8.16737533e-01 2.67506897e-01
-8.23564589e-01 3.06085460e-02 7.12512620e-03 -1.05447888e+00
-4.44872528e-01 -9.05108213e-01 3.65146697e-01 4.59273636e-01
-1.13178205e+00 -7.25743055e-01 7.93100148e-02 2.49548465e-01
9.49279964e-01 1.94523260e-01 1.17922544e+00 -7.48476446e-01
-4.44162160e-01 -6.64928630e-02 6.45888522e-02 2.15051919e-01
7.29528666e-02 -9.85190719e-02 -7.74621367e-01 3.25503424e-02
1.64825350e-01 -6.30579591e-01 4.28943276e-01 -1.59675121e-01
8.70281279e-01 -9.47426379e-01 1.84916869e-01 2.84083575e-01
1.23626995e+00 6.49341941e-01 9.19291556e-01 5.56184053e-01
5.98198250e-02 6.82321250e-01 3.76141936e-01 9.32791710e-01
1.11919634e-01 2.79007018e-01 7.18050897e-02 3.32766116e-01
4.19222265e-02 -5.20321846e-01 1.81329861e-01 2.85980314e-01
8.19922949e-04 -3.97487164e-01 -9.38511312e-01 6.45129561e-01
-1.61280441e+00 -1.02529883e+00 4.96456027e-01 2.08037734e+00
1.13243294e+00 1.34546474e-01 2.45091259e-01 -2.05894038e-02
2.78230071e-01 -2.12530211e-01 -7.51990736e-01 -8.77642095e-01
-2.15349764e-01 2.95939803e-01 -1.51871026e-01 -1.18051119e-01
-2.34367281e-01 9.32318926e-01 7.29561138e+00 3.74151230e-01
-1.33152556e+00 -9.10018757e-02 5.77999592e-01 -4.53027517e-01
-1.23326135e+00 6.84301108e-02 4.42142785e-02 5.11998355e-01
6.79272473e-01 -6.13785326e-01 5.15646517e-01 9.27524686e-01
1.90981790e-01 3.04177031e-02 -1.15461874e+00 8.44005287e-01
3.48018706e-01 -1.31256163e+00 5.40413082e-01 -1.49517849e-01
7.32768357e-01 -8.37890625e-01 6.40957236e-01 3.51751566e-01
7.10730180e-02 -1.28812408e+00 1.01200104e+00 7.24516571e-01
5.92090428e-01 -7.42991984e-01 4.15152520e-01 2.54244089e-01
-1.48112953e-01 -2.41238043e-01 -8.93520489e-02 -4.13709193e-01
-9.01269540e-02 3.80922854e-01 -1.17109716e+00 1.47423483e-02
4.14944798e-01 2.93682307e-01 -6.55003548e-01 9.54755366e-01
-2.52265275e-01 3.19021016e-01 1.12712182e-01 -4.04991746e-01
-1.07960008e-01 -1.00333817e-01 5.09757221e-01 1.09409380e+00
7.71788716e-01 -5.48404269e-02 -2.89212286e-01 1.72084236e+00
1.97521865e-01 3.16400766e-01 -7.50567198e-01 -6.39312029e-01
7.79300988e-01 1.03397036e+00 -5.78796327e-01 -2.44381800e-01
2.69557373e-03 9.10966396e-01 8.04592744e-02 3.96189868e-01
-5.42697251e-01 -1.29058808e-01 3.13469827e-01 3.50974321e-01
-1.82504848e-01 -1.18807867e-01 -1.09846914e+00 -9.42551851e-01
-4.92100269e-02 -1.12529552e+00 1.69345081e-01 -1.11209083e+00
-1.20317328e+00 5.02417922e-01 1.55663928e-02 -1.09416568e+00
-3.56146783e-01 -9.88741443e-02 -7.94394374e-01 7.13170409e-01
-9.44628954e-01 -1.27798831e+00 -3.34419638e-01 4.39518578e-02
3.99993241e-01 -3.68316442e-01 7.64663577e-01 -1.23234652e-01
-1.44462466e-01 6.82274461e-01 -3.84129614e-01 -7.90756762e-01
6.58464670e-01 -9.58474278e-01 1.39212146e-01 3.76380324e-01
-9.96862277e-02 9.21158671e-01 1.04639494e+00 -9.85283792e-01
-1.11978185e+00 -5.74962199e-01 1.13852596e+00 -4.24559802e-01
3.16940546e-01 7.81035945e-02 -8.53721201e-01 3.43084127e-01
3.35375190e-01 -9.69225764e-01 1.22208858e+00 2.10825235e-01
-3.86603892e-01 3.51399004e-01 -1.54310596e+00 1.02006340e+00
1.09626329e+00 -1.92222416e-01 -7.05662072e-01 -4.30665985e-02
5.59653699e-01 1.92316130e-01 -7.84124553e-01 2.61902452e-01
7.60795832e-01 -1.18458402e+00 8.24971974e-01 -6.18568778e-01
7.65653193e-01 8.30660909e-02 7.70569891e-02 -1.60228539e+00
-6.69105470e-01 -7.38521993e-01 -7.22789317e-02 1.19469714e+00
6.01603329e-01 -5.10554075e-01 6.03132188e-01 1.02227819e+00
-1.36277735e-01 -5.32375276e-01 -6.41718805e-01 -8.89338017e-01
8.26291069e-02 -2.65059263e-01 1.03128684e+00 1.22148287e+00
5.88262260e-01 3.16493958e-01 -3.54957312e-01 -4.99884307e-01
1.15539387e-01 -1.05875399e-04 6.73460364e-01 -1.36998165e+00
-2.88257807e-01 -9.45309103e-01 -2.03734815e-01 -4.48207468e-01
-1.50460139e-01 -6.92232668e-01 -1.75381288e-01 -1.46329916e+00
1.40542254e-01 -2.71190047e-01 -2.50929415e-01 4.06496286e-01
1.39536530e-01 1.24631897e-01 3.28409255e-01 4.15829062e-01
-4.25464004e-01 5.02645254e-01 1.21187246e+00 -5.77348061e-02
-7.32774019e-01 -3.33186835e-01 -1.39272177e+00 6.23908699e-01
1.05464590e+00 -6.25946969e-02 -7.55469263e-01 -3.54062617e-01
8.46206784e-01 -8.04363564e-02 2.46087030e-01 -1.15149736e+00
1.54453278e-01 -6.81479990e-01 6.88067853e-01 -6.95767105e-02
3.57332289e-01 -6.01273060e-01 6.70398057e-01 4.79184300e-01
-6.71656013e-01 1.94060430e-01 1.52198046e-01 1.30963385e-01
4.91528772e-03 -4.24631774e-01 5.87797403e-01 -4.27845925e-01
-5.42266488e-01 -3.41585368e-01 -5.59473634e-01 3.44347470e-02
1.14377689e+00 -8.06413531e-01 -2.40922615e-01 -5.84242821e-01
-6.55297995e-01 5.25332941e-03 8.69878411e-01 6.84350967e-01
1.04717588e+00 -1.39092946e+00 -6.22111082e-01 4.48182940e-01
5.46147823e-01 -4.96684343e-01 2.85988659e-01 3.87223244e-01
-2.17664525e-01 1.89818233e-01 -7.74619699e-01 -6.14520572e-02
-7.45757580e-01 6.45460725e-01 3.80497649e-02 2.44280040e-01
-2.11158141e-01 5.74742794e-01 -6.40828684e-02 8.48615635e-03
-6.34623915e-02 -2.43144587e-01 -2.88617343e-01 2.49182180e-01
2.38213331e-01 5.40955365e-01 -1.09426953e-01 -1.89211413e-01
-3.89757976e-02 -8.10843706e-02 -9.70302969e-02 -5.34020722e-01
9.16918993e-01 2.52443850e-02 8.72845948e-02 4.98603761e-01
3.77152473e-01 -7.86684901e-02 -7.82460809e-01 5.80625348e-02
2.04261079e-01 -8.15916121e-01 -8.09673145e-02 -1.53388989e+00
-3.93326789e-01 6.44722521e-01 4.14080948e-01 3.97977620e-01
1.20843863e+00 -4.77944426e-02 5.65588057e-01 4.09418583e-01
2.24805772e-01 -1.58701777e+00 7.51848757e-01 2.32747763e-01
1.35465014e+00 -8.24680626e-01 -1.50235251e-01 1.27242997e-01
-1.23898160e+00 9.77236986e-01 1.02679813e+00 2.64365554e-01
1.19597338e-01 -7.05939857e-03 2.73217913e-02 -1.11284032e-01
-1.03252292e+00 1.02212086e-01 1.09100670e-01 8.71540308e-01
7.65231550e-01 8.94148275e-02 -8.57789218e-01 9.30823028e-01
-4.34809685e-01 4.42546815e-01 7.34248817e-01 1.05096471e+00
-5.62133968e-01 -9.93216693e-01 -3.13968688e-01 4.56799895e-01
-8.47285241e-02 -1.25907242e-01 -1.05460453e+00 5.51834762e-01
4.64997917e-01 8.71750355e-01 4.16091569e-02 -4.20243621e-01
3.29899967e-01 2.82643557e-01 4.84709948e-01 -5.22087455e-01
-8.57089341e-01 -7.16006011e-02 6.71471730e-02 -9.44727361e-02
-5.60798533e-02 -5.03252745e-01 -1.03706026e+00 -2.40525141e-01
-1.03954375e-01 1.03723891e-01 7.46087193e-01 5.53921402e-01
7.38550365e-01 4.29558963e-01 2.77027011e-01 -4.24542397e-01
-7.65877545e-01 -1.03504300e+00 -4.02318299e-01 5.64925492e-01
-1.85291097e-01 -6.05417430e-01 -1.97176322e-01 -5.53727336e-02] | [11.61330795288086, 8.773098945617676] |
d2ec65a1-5848-4e7c-83f4-334548873816 | exploring-the-robustness-of-distributional | 2109.08776 | null | https://arxiv.org/abs/2109.08776v5 | https://arxiv.org/pdf/2109.08776v5.pdf | Exploring the Training Robustness of Distributional Reinforcement Learning against Noisy State Observations | In real scenarios, state observations that an agent observes may contain measurement errors or adversarial noises, misleading the agent to take suboptimal actions or even collapse while training. In this paper, we study the training robustness of distributional Reinforcement Learning (RL), a class of state-of-the-art methods that estimate the whole distribution, as opposed to only the expectation, of the total return. Firstly, we validate the contraction of distributional Bellman operators in the State-Noisy Markov Decision Process (SN-MDP), a typical tabular case that incorporates both random and adversarial state observation noises. In the noisy setting with function approximation, we then analyze the vulnerability of least squared loss in expectation-based RL with either linear or nonlinear function approximation. By contrast, we theoretically characterize the bounded gradient norm of distributional RL loss based on the categorical parameterization equipped with the KL divergence. The resulting stable gradients while the optimization in distributional RL accounts for its better training robustness against state observation noises. Finally, extensive experiments on the suite of environments verified that distributional RL is less vulnerable against both random and adversarial noisy state observations compared with its expectation-based counterpart. | ['Linglong Kong', 'Shangling Jui', 'Yingnan Zhao', 'Ke Sun'] | 2021-09-17 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-1.13799244e-01 1.98324054e-01 4.99991849e-02 -1.72151521e-01
-1.12356341e+00 -8.29790056e-01 6.95321977e-01 2.61836872e-02
-7.36723959e-01 9.03546214e-01 1.80815496e-02 -6.54373169e-01
-2.72805214e-01 -5.17065585e-01 -8.74357045e-01 -1.31539547e+00
-3.50354910e-01 3.20689499e-01 -1.79062948e-01 -6.57113269e-02
-9.04645212e-03 6.11836970e-01 -1.23018062e+00 -4.63047802e-01
5.86404204e-01 1.17675686e+00 -2.18509734e-01 7.40823388e-01
3.95153910e-01 1.27107060e+00 -9.02035534e-01 -4.40414876e-01
6.52978539e-01 -2.45785549e-01 -4.51293707e-01 -8.77845585e-02
3.94878447e-01 -5.26687145e-01 -6.48114800e-01 1.62800419e+00
7.01826692e-01 4.10404593e-01 6.32068634e-01 -1.50613260e+00
-4.93344486e-01 8.71152818e-01 -6.34943917e-02 -8.09682067e-03
1.73483253e-01 6.91386700e-01 8.02814066e-01 -3.20489824e-01
3.55234593e-01 1.40574408e+00 6.25272512e-01 7.79122472e-01
-1.51983798e+00 -4.50286627e-01 1.98549762e-01 -9.00296047e-02
-1.07772398e+00 -2.75008023e-01 4.50708747e-01 -5.57982206e-01
7.33111501e-01 7.68990964e-02 3.26979272e-02 1.49198377e+00
3.40405554e-01 8.16916347e-01 1.46434677e+00 6.09428436e-02
6.84209466e-01 3.45585942e-01 -1.88263529e-03 5.66244781e-01
2.62663305e-01 8.96253347e-01 9.18671489e-02 -4.17710096e-01
3.82435143e-01 1.79728493e-02 -2.85423070e-01 -6.38767779e-01
-9.28149283e-01 6.92296505e-01 5.26727736e-02 -1.07017338e-01
-4.96060669e-01 5.11619329e-01 5.86569190e-01 9.18375909e-01
1.15165941e-01 4.45262998e-01 -4.34829593e-01 -3.58838141e-01
-5.67508817e-01 3.79328996e-01 1.02515090e+00 9.34694469e-01
5.84261954e-01 6.37153506e-01 -5.28993666e-01 1.16113022e-01
1.90286487e-01 1.08991337e+00 3.50281745e-01 -1.04242074e+00
6.53329909e-01 -8.81079119e-03 5.33854663e-01 -5.15103102e-01
-4.15382713e-01 -6.19827151e-01 -9.14571404e-01 6.03680670e-01
6.85480654e-01 -7.62974560e-01 -5.24545789e-01 2.31627917e+00
2.65421987e-01 2.47127891e-01 5.05361617e-01 6.18413150e-01
-1.09410152e-01 2.84474194e-01 -5.57504185e-02 -4.12924260e-01
6.53162420e-01 -5.37227750e-01 -7.79570103e-01 -3.10833436e-02
6.14701033e-01 -2.77640522e-01 1.18288302e+00 3.90203923e-01
-9.88168359e-01 -4.52707827e-01 -9.88758504e-01 6.09433651e-01
-1.87672973e-01 -1.68308094e-01 3.48386057e-02 9.30690408e-01
-9.37806010e-01 1.08185899e+00 -1.02784956e+00 1.90154582e-01
2.29696900e-01 2.59806961e-01 -2.15055302e-01 2.32995525e-01
-1.26372373e+00 1.20151436e+00 2.52842844e-01 4.75963987e-02
-1.76708603e+00 -6.70451343e-01 -1.03806663e+00 -1.71047207e-02
7.33354449e-01 -4.63934213e-01 1.49419045e+00 -5.87835193e-01
-1.85466790e+00 1.68480724e-01 4.03955162e-01 -9.67173100e-01
1.13773704e+00 -3.39079827e-01 -2.63600916e-01 -7.49837458e-02
-2.55336583e-01 -1.11026257e-01 1.14609814e+00 -1.15813994e+00
-4.08400297e-01 -4.47920620e-01 1.60872146e-01 -2.17215437e-02
5.49546618e-04 -3.26424688e-01 5.44414580e-01 -3.90886396e-01
-4.73949075e-01 -9.44090128e-01 -4.76658046e-01 -4.05495286e-01
-5.50550938e-01 -1.14608303e-01 7.93248713e-01 -3.79232317e-01
1.18445110e+00 -2.28381300e+00 5.65601289e-02 3.37604105e-01
-1.86035842e-01 1.85060427e-02 -1.55197829e-01 6.26435637e-01
-1.42410249e-01 5.82265705e-02 -3.85747552e-01 -7.69095063e-01
7.18089640e-01 4.63081479e-01 -8.35758984e-01 1.14010835e+00
2.30700690e-02 6.38637364e-01 -9.66413021e-01 -5.31435162e-02
1.68065593e-01 -1.79004055e-02 -2.97199249e-01 3.47548217e-01
-2.65535951e-01 5.87056398e-01 -4.51067954e-01 2.51933157e-01
5.25859237e-01 2.83658206e-01 -1.12499520e-01 2.53100455e-01
-5.33300489e-02 1.77839603e-02 -1.44231343e+00 1.34362292e+00
-5.40453494e-01 2.82363027e-01 2.37496406e-01 -1.05557287e+00
6.77367806e-01 3.38237166e-01 4.35572863e-01 -2.47076064e-01
4.16003525e-01 5.78179248e-02 -1.66760713e-01 -4.14046615e-01
2.88717657e-01 -4.91296232e-01 -4.82860357e-01 6.52077079e-01
1.31337553e-01 -1.58844888e-01 -7.13770613e-02 9.11501274e-02
1.11271667e+00 2.11304098e-01 1.50481105e-01 -2.80066311e-01
4.48790312e-01 -5.26652932e-01 4.63507235e-01 1.23903906e+00
-5.47810316e-01 7.74386451e-02 8.16765070e-01 3.97859849e-02
-9.95165050e-01 -1.40483725e+00 6.33231737e-03 8.94315898e-01
-1.08168863e-01 -5.98376729e-02 -7.38282800e-01 -1.05133414e+00
2.28123993e-01 1.27314079e+00 -7.42417932e-01 -7.71274030e-01
-2.25324512e-01 -5.20228088e-01 9.18600678e-01 4.66720790e-01
4.04321432e-01 -8.51115584e-01 -5.18568635e-01 8.66625607e-02
2.27221102e-01 -9.19219732e-01 -6.05228841e-01 4.68192428e-01
-4.44909602e-01 -9.02418077e-01 -4.50073570e-01 -2.04508975e-01
3.97207379e-01 -4.95927542e-01 9.89749074e-01 -7.02134907e-01
2.67609686e-01 9.41067517e-01 2.38489732e-02 -6.49688184e-01
-8.58930349e-01 -3.47944319e-01 6.73716962e-01 1.53390914e-01
2.26328038e-02 -4.69804764e-01 -3.38100195e-01 1.76823944e-01
-1.00294268e+00 -1.01917362e+00 4.27081615e-01 8.70208204e-01
5.81730545e-01 2.43895859e-01 4.72061217e-01 -7.82031775e-01
8.76219153e-01 -4.81951147e-01 -1.13441312e+00 2.24034309e-01
-6.14684045e-01 5.23784757e-01 9.46587801e-01 -6.33333623e-01
-9.52847600e-01 -1.31289279e-02 1.30025959e-02 -9.27213073e-01
-2.20153347e-01 2.20708653e-01 -1.19922243e-01 2.03995436e-01
7.60857761e-01 5.26426733e-01 1.57240525e-01 -1.71494097e-01
4.00940865e-01 3.96867633e-01 5.27762830e-01 -8.31986368e-01
1.05817103e+00 4.26768869e-01 3.52504462e-01 -6.40855253e-01
-1.00299549e+00 -2.20508635e-01 -1.89813271e-01 -1.97261851e-02
7.13000536e-01 -6.28295660e-01 -1.19457507e+00 5.48142850e-01
-7.58592963e-01 -7.94662893e-01 -1.10450482e+00 6.46701396e-01
-1.21666896e+00 5.24714053e-01 -5.41741669e-01 -1.44509411e+00
-5.79710715e-02 -1.30290651e+00 9.07892585e-01 -1.75248444e-01
3.06478232e-01 -1.20799041e+00 2.94881701e-01 -2.88236946e-01
3.74054253e-01 5.13802350e-01 7.31824458e-01 -1.00614369e+00
-1.48052707e-01 -2.84206301e-01 4.87182319e-01 1.14630294e+00
-1.85786992e-01 -2.21344218e-01 -9.28150177e-01 -6.88593507e-01
5.88633657e-01 -5.10139108e-01 6.98226869e-01 3.68833721e-01
1.01575875e+00 -6.87652707e-01 3.75983626e-01 3.91652942e-01
1.50389612e+00 7.28787705e-02 4.04408962e-01 1.89332992e-01
3.29001665e-01 4.31590438e-01 6.63879395e-01 8.41244102e-01
-7.38656223e-02 1.30584463e-01 7.88131475e-01 3.16704482e-01
4.97600228e-01 -5.96322834e-01 1.19471955e+00 4.40513104e-01
3.50749701e-01 -5.80560751e-02 -5.07045329e-01 2.58240759e-01
-1.77993691e+00 -1.18299067e+00 2.55103678e-01 2.73171473e+00
7.71056235e-01 2.34285295e-01 2.88238168e-01 3.93372029e-02
5.88007748e-01 1.05365038e-01 -9.40568089e-01 -4.65894252e-01
-2.09050953e-01 5.83277568e-02 1.09370053e+00 6.04278564e-01
-1.17764366e+00 6.03641391e-01 6.56530714e+00 8.72273028e-01
-1.00071907e+00 1.80553183e-01 4.12377328e-01 -1.46803260e-01
-9.79756098e-03 -2.40381181e-01 -7.51887023e-01 5.70498884e-01
1.37920976e+00 -2.04477742e-01 6.74934745e-01 1.21841776e+00
3.36124629e-01 1.74655750e-01 -1.35876000e+00 8.20736527e-01
-2.67044753e-01 -6.78797603e-01 -2.49370575e-01 2.79517263e-01
9.11700964e-01 2.32771009e-01 5.79104841e-01 7.90951431e-01
1.03010452e+00 -9.83034313e-01 1.09090495e+00 7.40003109e-01
6.11072481e-01 -9.63214934e-01 1.07693505e+00 6.17689490e-01
-7.03330338e-01 -5.28793752e-01 -4.42191631e-01 7.99267143e-02
1.49902189e-02 2.61487693e-01 -5.78728259e-01 4.08366084e-01
2.24576280e-01 4.08412725e-01 -1.78051189e-01 6.15508378e-01
-4.10344809e-01 8.38880181e-01 -3.54969054e-01 2.48596631e-02
6.57709599e-01 -3.91009986e-01 9.84407961e-01 9.43291783e-01
2.19328448e-01 -4.62983638e-01 6.16035521e-01 8.26987147e-01
3.48133184e-02 -3.48595917e-01 -8.74359131e-01 -3.75275984e-02
2.83535272e-01 8.32684636e-01 8.15263614e-02 -2.91790873e-01
-1.43786043e-01 7.21573651e-01 2.93706447e-01 6.35704637e-01
-1.00342464e+00 -2.35244170e-01 1.08311129e+00 -3.81745040e-01
3.55417788e-01 -2.20357701e-01 3.77724022e-02 -1.00747383e+00
5.11520654e-02 -9.35851395e-01 4.22484130e-01 -1.34209290e-01
-1.64919996e+00 4.04759824e-01 -6.46614879e-02 -1.34230626e+00
-4.38034832e-01 -4.84531105e-01 -5.27724385e-01 7.30505764e-01
-1.34555840e+00 -4.99200553e-01 4.56231296e-01 1.00714648e+00
1.25522554e-01 -2.02878758e-01 6.21901572e-01 -5.54964282e-02
-8.64193082e-01 8.57325613e-01 6.44692361e-01 5.96143380e-02
5.78849912e-01 -1.89802659e+00 -1.77139323e-02 8.80141556e-01
-1.67659923e-01 5.33123434e-01 1.17523110e+00 -3.33866984e-01
-1.51113081e+00 -1.40216351e+00 3.43979639e-03 -5.96934378e-01
1.12846768e+00 -9.44976509e-02 -5.80351233e-01 9.95149434e-01
-1.46040499e-01 4.61481027e-02 3.60286713e-01 -3.33222300e-01
-4.44604218e-01 -4.95326258e-02 -1.57841635e+00 6.84004962e-01
7.16216266e-01 -7.58906782e-01 -5.33787310e-01 2.89419711e-01
8.67046237e-01 -1.60441890e-01 -1.08522332e+00 3.22963335e-02
1.74695939e-01 -8.35813701e-01 6.85797095e-01 -1.10209298e+00
-6.15215711e-02 -3.64688009e-01 -4.51918423e-01 -1.69074988e+00
-1.99538469e-02 -1.21687615e+00 -3.08203012e-01 9.13035154e-01
3.35171521e-01 -1.00404429e+00 4.65130180e-01 5.15661895e-01
-1.91346601e-01 -5.81823289e-01 -1.11333573e+00 -1.39193439e+00
5.28053641e-01 -6.77861750e-01 4.72267270e-01 4.13327754e-01
-5.99521399e-02 -3.19105424e-02 -4.88039553e-01 4.75920498e-01
8.71617436e-01 -1.09058164e-01 8.16219449e-01 -7.17956543e-01
-7.16205180e-01 -5.00928700e-01 -4.57269937e-01 -1.00126922e+00
5.06408811e-01 -7.44075656e-01 4.05771375e-01 -8.00729811e-01
-3.06004107e-01 -2.18319774e-01 -6.29183173e-01 1.82577878e-01
-9.21165943e-02 -4.58089173e-01 2.91781127e-01 -1.22523502e-01
-8.58230829e-01 8.79024744e-01 1.21670568e+00 -1.58454269e-01
-8.29709023e-02 5.75534582e-01 -4.78292584e-01 7.39268899e-01
9.33093905e-01 -6.46673799e-01 -4.79246408e-01 9.59838480e-02
1.09656937e-01 1.82002664e-01 3.81448030e-01 -1.03245902e+00
1.16804257e-01 -2.88007528e-01 -1.77400708e-01 -3.63477647e-01
2.23723367e-01 -9.87158656e-01 -1.32926643e-01 7.56926715e-01
-7.24017859e-01 5.49295805e-02 -2.20762119e-02 1.06066597e+00
-5.04561216e-02 -3.32691729e-01 9.74313140e-01 -5.90819120e-03
-1.76285446e-01 5.15901208e-01 -4.96156693e-01 3.76159728e-01
1.09815633e+00 2.53821254e-01 -1.90294728e-01 -7.49219358e-01
-8.80328715e-01 3.05378407e-01 1.76742911e-01 -1.37701616e-01
4.05208528e-01 -1.22313583e+00 -7.67928183e-01 -2.89156903e-02
-8.37890655e-02 -1.34876162e-01 3.31986845e-01 1.10143518e+00
1.07423700e-01 1.52045071e-01 1.12357542e-01 -3.02943528e-01
-8.64067912e-01 1.03214729e+00 6.19842172e-01 -6.11401141e-01
-3.16659272e-01 6.35245621e-01 3.01753301e-02 -5.46502233e-01
6.77647531e-01 -7.02284873e-01 2.43111476e-01 -1.59346864e-01
5.88355601e-01 7.30655789e-01 -1.03257358e-01 -3.44674289e-01
3.00780181e-02 -7.24300835e-03 2.30557770e-01 -3.78504843e-01
1.01266754e+00 -1.39814362e-01 2.94293851e-01 7.64874935e-01
1.27043819e+00 -1.06694534e-01 -1.86689973e+00 -2.93372571e-01
-3.82552072e-02 -1.54071182e-01 -4.45742384e-02 -7.60375798e-01
-9.35343623e-01 8.37149084e-01 7.52989650e-01 5.78785777e-01
7.84157395e-01 -2.74559915e-01 5.04170477e-01 7.92425394e-01
4.79732394e-01 -1.21355987e+00 -1.28236607e-01 7.81242311e-01
8.23856831e-01 -1.04705918e+00 -3.62730801e-01 3.68644714e-01
-9.30212855e-01 6.86226308e-01 1.80743724e-01 -4.10845727e-01
8.34641218e-01 4.14388329e-01 -2.64646616e-02 2.25217193e-01
-8.45048249e-01 -3.32569391e-01 -2.94169784e-01 8.42273295e-01
-4.31692898e-01 2.78199613e-01 1.73700586e-01 5.20427763e-01
-4.16078120e-01 -4.91748452e-01 6.59029007e-01 9.71630573e-01
-2.21548110e-01 -7.80369937e-01 -3.66984636e-01 2.76500911e-01
-6.75642490e-01 4.25932445e-02 -3.15503329e-02 5.90636551e-01
-2.54894555e-01 1.19581294e+00 -1.06256835e-01 -3.50040108e-01
6.86808527e-01 2.00449005e-01 3.24084252e-01 -2.39915475e-01
-7.06663847e-01 -2.78990239e-01 -1.53814733e-01 -7.55306304e-01
-4.74526249e-02 -9.62984145e-01 -9.54155028e-01 -3.94358814e-01
-2.43450582e-01 1.35293156e-01 7.26510644e-01 9.09864783e-01
4.76228930e-02 5.29622197e-01 1.01658702e+00 -6.07711256e-01
-2.30010247e+00 -1.12436724e+00 -9.82969344e-01 6.79838896e-01
8.10103595e-01 -6.24298692e-01 -9.25888181e-01 -4.29203480e-01] | [4.264456748962402, 2.47283935546875] |
cec9bc2c-21c2-4693-98db-dd5215946059 | morphosyntactic-tagging-with-a-meta-bilstm | 1805.08237 | null | http://arxiv.org/abs/1805.08237v1 | http://arxiv.org/pdf/1805.08237v1.pdf | Morphosyntactic Tagging with a Meta-BiLSTM Model over Context Sensitive Token Encodings | The rise of neural networks, and particularly recurrent neural networks, has
produced significant advances in part-of-speech tagging accuracy. One
characteristic common among these models is the presence of rich initial word
encodings. These encodings typically are composed of a recurrent
character-based representation with learned and pre-trained word embeddings.
However, these encodings do not consider a context wider than a single word and
it is only through subsequent recurrent layers that word or sub-word
information interacts. In this paper, we investigate models that use recurrent
neural networks with sentence-level context for initial character and
word-based representations. In particular we show that optimal results are
obtained by integrating these context sensitive representations through
synchronized training with a meta-model that learns to combine their states. We
present results on part-of-speech and morphological tagging with
state-of-the-art performance on a number of languages. | ['Ryan Mcdonald', 'Joshua Maynez', 'Goncalo Simoes', 'Emily Pitler', 'Daniel Andor', 'Bernd Bohnet'] | 2018-05-21 | morphosyntactic-tagging-with-a-meta-bilstm-1 | https://aclanthology.org/P18-1246 | https://aclanthology.org/P18-1246.pdf | acl-2018-7 | ['morphological-tagging'] | ['natural-language-processing'] | [ 2.80513495e-01 6.85590580e-02 -5.43794692e-01 -4.63098139e-01
-6.05174541e-01 -5.25140643e-01 4.90289271e-01 5.28795481e-01
-7.26588786e-01 5.56646287e-01 6.00830138e-01 -4.91094321e-01
2.21625865e-01 -7.52161145e-01 -3.77480477e-01 -5.53869545e-01
-2.72655308e-01 3.74173403e-01 3.69053662e-01 -4.66537178e-01
-1.25881508e-01 3.84716809e-01 -1.29333794e+00 4.68073875e-01
4.63474870e-01 4.21914518e-01 3.37900788e-01 7.54935205e-01
-7.49136806e-01 6.91292167e-01 -6.96152687e-01 -3.80767494e-01
-3.06918979e-01 -3.90166312e-01 -8.59025896e-01 -2.09532633e-01
1.63441107e-01 3.69354159e-01 -1.78564891e-01 1.06233037e+00
4.84094650e-01 3.18796784e-01 3.10095757e-01 -4.65420544e-01
-8.85894358e-01 1.29053819e+00 5.29875979e-02 5.67190826e-01
2.17975751e-01 -4.06374514e-01 1.49260700e+00 -7.22024202e-01
7.61419833e-01 1.10805261e+00 8.96646261e-01 8.01309347e-01
-1.17943060e+00 -2.58727193e-01 4.93298769e-01 -7.60961697e-02
-1.11852860e+00 -4.20043230e-01 6.63022757e-01 -1.00603484e-01
2.01938009e+00 -1.17328010e-01 6.74500942e-01 1.03732193e+00
4.53873098e-01 7.15958834e-01 6.32721126e-01 -1.06667745e+00
2.80838329e-02 -3.49906692e-03 5.71657717e-01 7.56667078e-01
1.60877824e-01 -8.50944594e-02 -4.00174975e-01 -8.27972218e-02
4.92677480e-01 2.38104373e-01 2.31398687e-01 -1.79031380e-02
-8.33397269e-01 1.06857586e+00 1.98420376e-01 1.13989019e+00
-3.84684145e-01 3.04879934e-01 6.43289506e-01 1.82135820e-01
5.98301351e-01 4.70507383e-01 -7.63210058e-01 -2.08743840e-01
-9.16330636e-01 -2.09755704e-01 8.12590361e-01 7.39032984e-01
6.95084810e-01 2.78194189e-01 -2.09346369e-01 1.37362564e+00
2.03734130e-01 2.36165494e-01 1.09563804e+00 -4.29989070e-01
3.21244270e-01 4.24438149e-01 -4.03219342e-01 -4.62280452e-01
-5.02875686e-01 -7.61859655e-01 -5.55411637e-01 -3.41379970e-01
-7.32626021e-02 -2.74974793e-01 -1.35808253e+00 2.06850243e+00
-1.43010005e-01 1.94085911e-01 2.92393118e-01 2.56398737e-01
6.89136565e-01 8.58263731e-01 5.88421226e-01 -1.51598334e-01
1.58708191e+00 -1.08101416e+00 -1.00814497e+00 -7.02058017e-01
9.59787011e-01 -7.17432082e-01 5.86477697e-01 -9.18954834e-02
-1.20255387e+00 -6.96626365e-01 -1.04089987e+00 -2.49099713e-02
-9.39601481e-01 4.12183963e-02 6.40233636e-01 7.82323003e-01
-1.25073624e+00 6.95172131e-01 -7.44118273e-01 -6.09420121e-01
-1.74020782e-01 4.39477444e-01 -2.26040348e-01 1.93821952e-01
-1.76618242e+00 1.12978804e+00 5.66350996e-01 -2.39113756e-02
-3.55930448e-01 -3.34115714e-01 -1.14992952e+00 1.16482906e-01
-1.21797740e-01 -4.20858711e-01 1.49694753e+00 -9.44873273e-01
-1.34906065e+00 9.28543031e-01 -3.70313019e-01 -6.83466554e-01
-5.00712514e-01 -2.25968748e-01 -8.21298242e-01 -1.26952425e-01
-1.29995763e-01 6.78137422e-01 5.38493514e-01 -9.22228158e-01
-6.42114282e-01 5.83688437e-04 -8.92088562e-02 7.30575696e-02
-5.77287376e-01 4.29685801e-01 -1.53971419e-01 -7.36071944e-01
-2.91358978e-02 -8.20178390e-01 -5.21086633e-01 -7.67977655e-01
-1.81291637e-03 -5.88227570e-01 5.41857123e-01 -5.43787301e-01
1.62716985e+00 -2.06879807e+00 1.39367422e-02 -2.00302288e-01
-2.32019827e-01 7.42697299e-01 -5.80974340e-01 7.93417633e-01
-5.36970258e-01 3.40479434e-01 -1.27143800e-01 -6.96380913e-01
-9.47535187e-02 7.71614432e-01 -3.18078399e-01 5.82006127e-02
7.14256287e-01 9.88521814e-01 -1.06392682e+00 -3.53158176e-01
5.25541157e-02 8.04579020e-01 -3.74811709e-01 1.78740859e-01
-2.08188117e-01 -2.42595464e-01 -8.01471174e-02 4.44875419e-01
1.69065177e-01 5.98522983e-02 6.24103904e-01 2.48635203e-01
-3.18026900e-01 1.17125916e+00 -6.98901474e-01 1.92664230e+00
-8.04050207e-01 4.40806419e-01 -2.79353470e-01 -1.02332342e+00
1.04531419e+00 7.09469020e-01 3.55712593e-01 -6.23958588e-01
1.30747452e-01 3.46877217e-01 2.03289449e-01 -2.16594860e-01
1.03416872e+00 -4.38241273e-01 -2.29191840e-01 3.35344613e-01
6.79266512e-01 2.18698293e-01 3.65588963e-01 -1.35059459e-02
1.30142009e+00 1.47093177e-01 4.88521606e-01 1.41383067e-01
3.86071444e-01 -1.92847162e-01 7.69244075e-01 5.27464986e-01
-5.03964424e-02 6.48827851e-01 2.96316892e-01 -3.53157431e-01
-1.04596663e+00 -6.96686506e-01 -2.27855667e-01 1.62762737e+00
-5.19164503e-01 -6.19928181e-01 -4.92219597e-01 -6.54231369e-01
-2.20713466e-01 7.42465019e-01 -7.71784127e-01 -2.46103942e-01
-1.05077469e+00 -6.60113215e-01 7.04513133e-01 7.92110384e-01
-1.38865411e-01 -1.61077476e+00 -4.90278870e-01 8.53489339e-01
1.26547053e-01 -9.99871314e-01 -2.59306163e-01 1.31446838e+00
-1.06835496e+00 -3.60414118e-01 -7.17195809e-01 -1.19667149e+00
4.49047506e-01 8.22975338e-02 1.46404541e+00 2.95098186e-01
-2.16441378e-02 1.13046691e-01 -9.50361490e-01 -1.96876638e-02
-6.37992203e-01 8.00782919e-01 -1.49345338e-01 -2.76658595e-01
5.53064942e-01 -5.98058164e-01 9.65316221e-02 -2.61467278e-01
-1.03300524e+00 -4.48181957e-01 7.94363081e-01 9.87558007e-01
3.84128243e-01 -3.28230411e-01 4.14317161e-01 -1.12484908e+00
6.21993005e-01 -5.06719112e-01 -1.23023503e-01 3.32408935e-01
-4.25775856e-01 4.81295615e-01 4.77552027e-01 -6.69245660e-01
-9.05933619e-01 1.43650696e-01 -6.58272445e-01 -1.99465305e-02
-3.29443991e-01 8.75308633e-01 2.50215620e-01 4.10961330e-01
4.23971653e-01 1.42134026e-01 -3.32655638e-01 -7.44861364e-01
5.60726523e-01 7.15088248e-01 3.18247139e-01 -6.46504700e-01
3.73675972e-01 -2.30027139e-01 -4.13630724e-01 -8.95087540e-01
-9.74700749e-01 -7.82626510e-01 -9.08485830e-01 1.72030956e-01
9.78197455e-01 -7.88985670e-01 3.39706168e-02 1.47138014e-01
-1.49113131e+00 -3.22714984e-01 -6.25133753e-01 3.85155052e-01
-1.39763936e-01 2.24630013e-01 -1.03616607e+00 -9.78380919e-01
-2.97862828e-01 -7.71347225e-01 8.83123457e-01 3.05149347e-01
-4.13544714e-01 -1.36560881e+00 7.21076965e-01 -3.61496150e-01
6.05275869e-01 -4.37409461e-01 9.81053472e-01 -1.05555260e+00
6.28598034e-02 -3.08199912e-01 1.85893297e-01 3.70570302e-01
2.04658762e-01 -2.60716509e-02 -9.77451265e-01 -2.08243102e-01
-4.88167912e-01 -2.14023754e-01 1.19733858e+00 2.89545774e-01
5.30579567e-01 -2.06526332e-02 -3.15358579e-01 4.04642016e-01
1.35618520e+00 4.12119806e-01 6.27713263e-01 2.79190361e-01
3.93965840e-01 4.27973479e-01 2.32851967e-01 8.25233012e-02
1.53561935e-01 5.58488190e-01 2.85018712e-01 -1.69034265e-02
-4.54373837e-01 -3.21238548e-01 5.93577981e-01 1.74812043e+00
2.81091630e-02 -4.82416540e-01 -9.56624031e-01 9.19779122e-01
-1.77073920e+00 -9.90350187e-01 1.57947302e-01 1.84054077e+00
1.04261827e+00 2.64570504e-01 -2.57272329e-02 5.42680249e-02
8.16552997e-01 6.09012544e-01 -4.26996611e-02 -1.11757100e+00
-2.37151280e-01 8.25589001e-01 4.35472190e-01 5.77459991e-01
-1.11201048e+00 1.44114411e+00 7.12005186e+00 5.31896651e-01
-1.02439129e+00 4.45992708e-01 2.22499043e-01 3.66685987e-01
-4.77971613e-01 1.07743017e-01 -1.25668192e+00 1.37142733e-01
1.62337410e+00 3.44802409e-01 -2.31581666e-02 7.87741184e-01
-3.42559427e-01 1.05055727e-01 -6.94237947e-01 2.99949974e-01
2.52921164e-01 -1.25835228e+00 9.52419490e-02 -7.88027868e-02
6.05347693e-01 3.99358720e-01 -6.43233135e-02 6.46241546e-01
7.12525249e-01 -9.87806380e-01 6.10488653e-01 3.84809703e-01
6.22221649e-01 -7.98707426e-01 1.10271645e+00 -5.16859218e-02
-1.57306957e+00 4.85349558e-02 -5.75039566e-01 -9.85015035e-02
2.90028900e-01 4.42646831e-01 -7.16233134e-01 5.36086857e-01
2.37325877e-01 9.06107187e-01 -5.44786811e-01 9.31430042e-01
-5.03988683e-01 8.21512878e-01 -5.79606369e-03 -4.30131704e-01
5.98703504e-01 1.13091819e-01 2.93690503e-01 2.05692077e+00
1.69029847e-01 -5.79833090e-02 2.74846524e-01 2.35464364e-01
-8.93390998e-02 1.05590738e-01 -5.60631394e-01 -4.30005312e-01
5.40000260e-01 1.28765237e+00 -7.95729041e-01 -3.96562457e-01
-5.26386499e-01 8.47921133e-01 7.63445973e-01 2.46327043e-01
-4.85134453e-01 -6.22993290e-01 1.02050817e+00 -1.90481350e-01
8.09068382e-01 -5.91355145e-01 6.03634045e-02 -1.06268239e+00
-3.55692804e-01 -4.33860362e-01 4.68487620e-01 -5.40846407e-01
-1.19983244e+00 9.82201159e-01 -1.96879283e-01 -7.64691353e-01
-6.75926149e-01 -8.05392981e-01 -5.83154976e-01 8.57797503e-01
-1.77415597e+00 -1.15941048e+00 5.33933282e-01 4.58358377e-02
6.13098800e-01 -3.17091674e-01 1.43724608e+00 2.49802649e-01
-5.77822804e-01 5.15999675e-01 -6.12036772e-02 5.69031596e-01
5.00421166e-01 -1.21682048e+00 8.22406411e-01 9.04497981e-01
8.05501640e-01 9.26910341e-01 4.99000251e-01 -5.99581420e-01
-9.69435811e-01 -9.68685091e-01 1.80885184e+00 -2.80204475e-01
8.75383139e-01 -5.39517701e-01 -1.09362829e+00 7.46167779e-01
6.50808096e-01 3.46415639e-01 8.48940730e-01 8.25353682e-01
-6.54319704e-01 1.10925317e-01 -5.21478415e-01 3.83541733e-01
1.11204886e+00 -9.79691625e-01 -1.24151087e+00 -3.33541036e-02
1.12038934e+00 -1.67800158e-01 -7.12013781e-01 1.50842279e-01
3.36045235e-01 -6.72567964e-01 8.44645083e-01 -1.01093423e+00
5.05730882e-02 1.20380558e-02 -2.66317189e-01 -1.50780928e+00
-6.92604661e-01 -5.26603878e-01 5.88277802e-02 1.39545774e+00
8.06269825e-01 -5.27417600e-01 5.48614264e-01 -2.08281353e-02
-5.57321668e-01 -6.00752056e-01 -1.02725828e+00 -9.05684769e-01
3.85534793e-01 -5.14314532e-01 3.69478017e-01 9.89243567e-01
2.38710612e-01 6.67304993e-01 -7.86728561e-02 -1.33268356e-01
-7.16292262e-02 -2.09062636e-01 -2.09226340e-01 -1.27252746e+00
-1.39295429e-01 -4.52686608e-01 -6.10031664e-01 -8.68113399e-01
5.80624998e-01 -1.07631087e+00 4.55402464e-01 -1.58889532e+00
-4.66767512e-02 -5.02476156e-01 -8.81369352e-01 9.99418139e-01
-5.39233647e-02 1.38100043e-01 -2.20621563e-03 -2.18075991e-01
-6.32630050e-01 2.86511004e-01 4.25787240e-01 -4.15722691e-02
-1.88418746e-01 -3.49907428e-01 -2.17797533e-01 4.40681875e-01
8.94422174e-01 -8.95463705e-01 2.38388702e-02 -6.47590578e-01
6.76555991e-01 -1.06503844e-01 -1.27600133e-01 -1.03873062e+00
2.09057018e-01 2.33640105e-01 1.07054353e-01 -5.34562349e-01
4.86155152e-01 -4.56593841e-01 -1.63918912e-01 5.26464343e-01
-5.56450665e-01 4.34485853e-01 2.38110781e-01 3.27353060e-01
-3.89793992e-01 -7.80784786e-01 6.38000190e-01 -3.23623002e-01
-7.83362687e-01 -3.85149941e-02 -8.27438891e-01 9.58700180e-02
3.40144932e-01 -1.24044724e-01 -9.99267399e-02 6.56889146e-03
-8.88747633e-01 -4.17330921e-01 1.06349073e-01 7.79390395e-01
3.42829585e-01 -1.38965607e+00 -5.21302700e-01 2.95169353e-01
1.03345171e-01 -5.31367898e-01 -2.05539510e-01 2.90370762e-01
-5.77569976e-02 7.50135481e-01 -7.07514063e-02 -2.39966542e-01
-1.24651575e+00 4.01428849e-01 2.43307278e-01 -6.82432532e-01
-4.42898721e-01 9.88627374e-01 -5.32394528e-01 -6.14079654e-01
2.32748926e-01 -6.74729943e-01 -5.84709287e-01 4.31306958e-01
3.68956804e-01 -2.36386061e-01 3.20841312e-01 -8.60440075e-01
-6.32656932e-01 5.29224217e-01 -1.94257468e-01 -3.07614684e-01
1.59639823e+00 1.50750726e-01 -1.56174049e-01 7.94394672e-01
1.11736608e+00 1.15665480e-01 -7.58413792e-01 -4.75568235e-01
4.38431114e-01 2.71482289e-01 -4.16498780e-02 -6.66141093e-01
-1.00371563e+00 1.12344325e+00 2.99080670e-01 4.75933760e-01
6.71090543e-01 1.76168546e-01 9.80222046e-01 4.46865082e-01
1.81409448e-01 -1.23140347e+00 -7.82050018e-04 1.41968608e+00
2.75842249e-01 -7.90117383e-01 -3.01589757e-01 -2.31935605e-02
-3.41602921e-01 1.34360433e+00 2.00736508e-01 -3.38469028e-01
7.66950309e-01 5.54642558e-01 2.76914835e-01 3.65486741e-02
-1.17150998e+00 -8.19887698e-01 2.46344924e-01 5.13034999e-01
1.25489008e+00 1.02710791e-01 -5.09835541e-01 5.31922340e-01
-2.68735643e-02 -4.08490151e-01 2.45656475e-01 1.17704380e+00
-7.92042732e-01 -1.86074114e+00 -5.54692894e-02 8.43986124e-02
-6.22430861e-01 -5.87401986e-01 -2.82460839e-01 5.25317311e-01
2.88906932e-01 8.60377014e-01 3.61425281e-01 -3.43890935e-01
1.72229007e-01 7.15528846e-01 3.49506080e-01 -1.38093901e+00
-1.22614622e+00 7.38196373e-02 4.90857601e-01 -3.15790147e-01
-5.58704734e-01 -7.71202624e-01 -1.38677394e+00 2.69705206e-01
-4.46451217e-01 5.36413014e-01 8.91024292e-01 1.06704414e+00
1.99854150e-01 8.60920489e-01 3.38448793e-01 -6.33288741e-01
-3.43862802e-01 -1.17477930e+00 -4.41567272e-01 -2.90234893e-04
3.38046700e-01 -4.12608415e-01 -1.62209719e-01 -1.74288854e-01] | [10.389246940612793, 9.792105674743652] |
3dc8779c-efe7-4581-b51f-55d2517ab150 | language-is-not-all-you-need-aligning | 2302.14045 | null | https://arxiv.org/abs/2302.14045v2 | https://arxiv.org/pdf/2302.14045v2.pdf | Language Is Not All You Need: Aligning Perception with Language Models | A big convergence of language, multimodal perception, action, and world modeling is a key step toward artificial general intelligence. In this work, we introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale multimodal corpora, including arbitrarily interleaved text and images, image-caption pairs, and text data. We evaluate various settings, including zero-shot, few-shot, and multimodal chain-of-thought prompting, on a wide range of tasks without any gradient updates or finetuning. Experimental results show that Kosmos-1 achieves impressive performance on (i) language understanding, generation, and even OCR-free NLP (directly fed with document images), (ii) perception-language tasks, including multimodal dialogue, image captioning, visual question answering, and (iii) vision tasks, such as image recognition with descriptions (specifying classification via text instructions). We also show that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge from language to multimodal, and from multimodal to language. In addition, we introduce a dataset of Raven IQ test, which diagnoses the nonverbal reasoning capability of MLLMs. | ['Barun Patra', 'Furu Wei', 'Xia Song', 'Subhojit Som', 'Vishrav Chaudhary', 'Johan Bjorck', 'Zewen Chi', 'Kriti Aggarwal', 'Qiang Liu', 'Owais Khan Mohammed', 'Lei Cui', 'Tengchao Lv', 'Shuming Ma', 'Saksham Singhal', 'Yaru Hao', 'Wenhui Wang', 'Li Dong', 'Shaohan Huang'] | 2023-02-27 | null | null | null | null | ['optical-character-recognition'] | ['computer-vision'] | [ 2.78568953e-01 4.41486314e-02 -4.28969860e-02 -4.55615193e-01
-9.44582164e-01 -6.98065817e-01 8.75155747e-01 1.85288116e-03
-4.95555103e-01 3.72796863e-01 3.68210644e-01 -3.91413063e-01
2.69750893e-01 -3.39663446e-01 -9.86699402e-01 -2.99341470e-01
2.86347717e-01 6.26768351e-01 -1.49718374e-01 -3.35640788e-01
2.06030115e-01 -7.43051544e-02 -1.24675643e+00 7.89361596e-01
9.12919104e-01 1.02175283e+00 4.48463529e-01 1.09708548e+00
-2.96231002e-01 1.19000077e+00 -2.52031058e-01 -8.43784928e-01
-4.93804306e-01 -3.02019417e-01 -8.53343368e-01 2.78644174e-01
6.50251567e-01 -6.75262213e-01 -3.72498393e-01 8.53570402e-01
4.18177068e-01 2.47230947e-01 7.10693955e-01 -1.36191583e+00
-1.40779114e+00 3.43944758e-01 -3.38226140e-01 -2.94263393e-01
8.41729522e-01 7.05432713e-01 6.93800092e-01 -1.24305081e+00
4.82670933e-01 2.05751610e+00 5.81365414e-02 9.19125021e-01
-1.27611756e+00 -2.65655369e-01 1.55673683e-01 2.78786778e-01
-1.08884752e+00 -5.73726535e-01 3.71047497e-01 -4.49229836e-01
9.99335766e-01 2.07615033e-01 1.11233801e-01 1.60372424e+00
1.26542479e-01 1.19492483e+00 1.14973068e+00 -5.02748489e-01
1.25688076e-01 1.24763407e-01 1.01628102e-01 9.44330752e-01
-5.59562564e-01 -2.74855085e-02 -9.88202870e-01 4.83181514e-02
5.23628712e-01 -1.15309946e-01 -2.21728295e-01 -1.39153734e-01
-1.69899094e+00 6.21164739e-01 1.99810550e-01 -1.50693923e-01
-1.91969290e-01 1.99649185e-01 3.63076001e-01 4.47446197e-01
3.23801935e-02 2.56985664e-01 -2.46696338e-01 -2.14287251e-01
-1.14377834e-01 -1.60735801e-01 7.01799572e-01 1.03704154e+00
8.08851242e-01 -4.25380133e-02 -5.51582396e-01 1.18059051e+00
4.15784419e-01 1.16095793e+00 6.39422536e-01 -1.03680182e+00
8.34481835e-01 3.97765219e-01 2.22449955e-02 -6.56243563e-01
-3.40933502e-01 4.97350365e-01 -8.41979563e-01 -3.85601580e-01
1.92467198e-01 -2.03970075e-01 -1.12768233e+00 2.05625105e+00
-2.61171535e-02 -1.70326516e-01 6.24220371e-01 8.94302726e-01
1.26977444e+00 1.03268659e+00 3.95344198e-01 -3.78509983e-02
1.53781569e+00 -1.22259653e+00 -6.98075354e-01 -5.66549003e-01
5.16279817e-01 -4.24762547e-01 1.76831234e+00 3.40504587e-01
-1.14576840e+00 -7.45778739e-01 -5.58496356e-01 -5.32376587e-01
-6.58432484e-01 3.53611074e-02 5.46094239e-01 2.14171946e-01
-1.24371111e+00 -2.92988122e-01 -4.84942466e-01 -7.36913323e-01
1.14554986e-01 9.84273180e-02 -6.96155250e-01 -5.27936280e-01
-1.16117251e+00 8.43349636e-01 4.17248935e-01 -1.64173618e-02
-1.26249373e+00 -2.97851920e-01 -1.33684933e+00 9.65481102e-02
5.73273242e-01 -1.00799704e+00 1.41894543e+00 -1.02104092e+00
-1.57031417e+00 1.07312608e+00 -4.36659545e-01 -2.90854990e-01
8.92776698e-02 -2.33562663e-01 -3.56582791e-01 4.22047347e-01
-1.71870813e-01 1.36811018e+00 9.53079998e-01 -1.41188002e+00
-1.89453527e-01 -4.17465329e-01 2.39605069e-01 5.98503113e-01
-4.16743040e-01 3.85120846e-02 -6.46873653e-01 -1.42763421e-01
-2.97058254e-01 -7.44339764e-01 1.69951618e-01 -2.18099892e-01
-5.51783204e-01 -4.47971553e-01 4.94388580e-01 -7.62507141e-01
5.95140696e-01 -2.32169414e+00 4.64392811e-01 -2.77499706e-01
5.04396781e-02 -6.75064474e-02 -9.20635283e-01 7.10633457e-01
3.48160267e-01 1.86886638e-01 -9.19365957e-02 -5.98770797e-01
5.11435211e-01 2.99824178e-01 -5.53371966e-01 -1.35796651e-01
2.19302744e-01 1.64812136e+00 -8.64835024e-01 -4.45519775e-01
2.81829238e-01 3.10881108e-01 -3.47098947e-01 6.30488932e-01
-6.68842554e-01 3.13528627e-01 4.22274321e-02 8.32895100e-01
2.99614757e-01 -4.83084291e-01 -1.05980240e-01 -3.46799910e-01
2.22542837e-01 -1.98966727e-01 -5.03167152e-01 2.15264583e+00
-6.87354803e-01 6.92718744e-01 -6.20222911e-02 -5.80650032e-01
5.07415771e-01 3.69044334e-01 -3.30128878e-01 -1.14447546e+00
-5.73713370e-02 -1.76230118e-01 -2.46298611e-01 -1.02354646e+00
4.35126156e-01 -1.06104100e-02 -2.37691432e-01 4.65821147e-01
5.19407094e-01 -2.04871863e-01 3.60273540e-01 6.24113619e-01
5.96111894e-01 -1.17960416e-01 1.71137527e-01 4.41674024e-01
4.16811198e-01 -2.73268014e-01 -2.65779853e-01 1.03057396e+00
2.21522059e-02 5.23105383e-01 4.89020526e-01 -9.89854243e-03
-7.43673980e-01 -1.46970582e+00 4.51231956e-01 1.77936399e+00
1.56710774e-01 -2.05981493e-01 -7.60683060e-01 -3.65291506e-01
-9.26907212e-02 9.76347506e-01 -4.37334359e-01 -2.71421343e-01
-2.14614184e-03 -3.65422875e-01 5.65829575e-01 4.44054723e-01
6.39401555e-01 -1.36437178e+00 -3.54797989e-01 -1.32654205e-01
-4.37738717e-01 -1.58891845e+00 -7.07378268e-01 -3.41805637e-01
-5.58281362e-01 -7.64464259e-01 -8.12830865e-01 -9.21389163e-01
6.98238850e-01 2.49172434e-01 1.15398049e+00 -2.53236860e-01
-8.97725374e-02 1.43691874e+00 -3.07796001e-01 -1.78972214e-01
-3.22734147e-01 -3.84362817e-01 -6.26398027e-02 2.79968590e-01
1.90349936e-01 -1.05055280e-01 -3.48745435e-01 2.03115821e-01
-1.12093902e+00 5.54923296e-01 9.20110464e-01 9.30737019e-01
4.26353961e-01 -8.48800182e-01 6.57995403e-01 -3.28247428e-01
1.09183407e+00 -3.88105094e-01 -2.38938987e-01 9.75819111e-01
6.21788111e-03 -7.80400913e-03 4.10976112e-01 -8.23766470e-01
-1.17612946e+00 -2.24619403e-01 9.76089016e-02 -5.15011966e-01
-5.61515629e-01 6.00232065e-01 -3.81364703e-01 5.82996011e-02
4.02249813e-01 8.25999200e-01 8.04267526e-02 -1.79801285e-01
1.13043702e+00 9.33255315e-01 1.02133167e+00 -7.83875823e-01
3.99926573e-01 3.24049592e-01 -4.68036056e-01 -1.01281011e+00
-6.31055832e-01 -1.78707987e-01 -3.99352968e-01 -2.54595637e-01
1.25704670e+00 -9.24704492e-01 -1.24680364e+00 4.42254305e-01
-1.48851192e+00 -6.62212133e-01 8.01108703e-02 3.06105405e-01
-8.35273802e-01 3.27405721e-01 -8.51788342e-01 -8.34779501e-01
-3.96370858e-01 -1.12448168e+00 1.34035087e+00 4.82843727e-01
-7.65542910e-02 -9.30285573e-01 -2.04996556e-01 8.66361856e-01
2.13574946e-01 -2.07569495e-01 1.23175263e+00 -5.83321750e-01
-7.31506884e-01 1.19485714e-01 -5.50176620e-01 3.49319100e-01
-2.24819958e-01 -3.33389074e-01 -9.93353188e-01 -1.97542787e-01
-3.38152915e-01 -1.40584469e+00 7.61975229e-01 5.50045669e-02
1.25046682e+00 -4.61118072e-01 3.38190161e-02 4.54132348e-01
1.00387180e+00 3.61767501e-01 6.02267981e-01 -1.87997997e-01
7.09178805e-01 7.14987278e-01 3.87012988e-01 9.98202637e-02
1.05657840e+00 2.46567830e-01 3.91556531e-01 -1.10801063e-01
-1.20721087e-01 -6.19248748e-01 6.13588691e-01 9.20641005e-01
3.53578389e-01 -5.99531710e-01 -1.06532025e+00 3.87898237e-01
-1.86012173e+00 -7.28031695e-01 4.55639124e-01 1.83131361e+00
8.65031481e-01 -2.96286255e-01 -3.26708674e-01 -8.57974648e-01
3.72509062e-01 1.23005293e-01 -8.74875367e-01 -5.72163105e-01
-3.73168141e-01 -3.42560172e-01 -2.54744142e-01 5.89128673e-01
-8.31009567e-01 1.28789401e+00 5.84107208e+00 5.97838223e-01
-9.95105147e-01 2.06468195e-01 6.42391920e-01 2.03231745e-03
-4.15661454e-01 -2.30974585e-01 -4.43378806e-01 3.15067768e-01
1.01341963e+00 -6.59095347e-02 9.27087724e-01 3.51496249e-01
-1.25751585e-01 -2.65803665e-01 -1.47370756e+00 1.55976439e+00
7.98528433e-01 -1.12178624e+00 7.15388238e-01 -3.43839258e-01
5.52907228e-01 6.63939714e-02 3.16882968e-01 7.04488277e-01
1.71548337e-01 -1.28954935e+00 6.48324609e-01 8.44669461e-01
9.82811332e-01 -3.86970878e-01 4.17300701e-01 6.30889118e-01
-7.37055063e-01 -9.24485847e-02 -3.50582376e-02 2.33584791e-01
3.23900998e-01 -3.14532280e-01 -6.07923329e-01 3.31181586e-01
3.46872866e-01 3.87275130e-01 -5.91215968e-01 5.01406670e-01
-2.92229235e-01 2.31664419e-01 6.39992356e-02 -2.94373959e-01
2.15340137e-01 -7.04902709e-02 2.67743707e-01 1.25478661e+00
1.83732420e-01 3.33352774e-01 2.44199023e-01 9.02396441e-01
-4.30148631e-01 1.41177207e-01 -6.90922320e-01 -4.82568115e-01
2.40548208e-01 9.87695396e-01 -1.81579083e-01 -4.95996803e-01
-7.49553025e-01 1.30880976e+00 4.26872611e-01 1.01804876e+00
-4.92498636e-01 -2.49187529e-01 3.22618544e-01 -4.93032277e-01
-9.25469846e-02 -4.32439834e-01 1.13047116e-01 -1.37741971e+00
-2.46496394e-01 -1.15230632e+00 3.85587424e-01 -1.66830170e+00
-1.46655548e+00 4.85474795e-01 1.49515018e-01 -6.34097636e-01
-5.29351354e-01 -9.36986566e-01 -3.64216000e-01 6.97665572e-01
-1.25067794e+00 -1.36578679e+00 -4.25104886e-01 8.46703649e-01
9.76502538e-01 -2.70563483e-01 7.61562169e-01 -5.97477220e-02
-3.86358440e-01 4.99364465e-01 -2.92046726e-01 1.93430752e-01
9.38818276e-01 -1.05056214e+00 8.39400664e-02 3.48634511e-01
3.19120675e-01 6.76129222e-01 2.62832373e-01 -4.80637610e-01
-2.06307840e+00 -7.35622108e-01 6.32291615e-01 -5.01159668e-01
8.27698171e-01 -6.74137533e-01 -8.31881821e-01 8.35568845e-01
8.58394325e-01 -4.46214288e-01 8.77936780e-01 -8.90434682e-02
-4.98067886e-01 1.44813480e-02 -8.41085017e-01 9.70238626e-01
7.80723095e-01 -1.10887980e+00 -7.97524452e-01 5.80628037e-01
1.21074700e+00 -4.07546312e-01 -6.41412795e-01 3.04525018e-01
5.31191647e-01 -6.17655873e-01 1.13826263e+00 -9.86287355e-01
7.28958130e-01 1.04764692e-01 -5.66749036e-01 -1.32748735e+00
1.66743115e-01 -6.07987463e-01 -2.58315355e-01 1.16865683e+00
4.66122329e-01 -5.34577787e-01 -3.19048837e-02 6.93533897e-01
-1.39370218e-01 -4.52807724e-01 -8.17495525e-01 -5.72390676e-01
-1.09879184e-03 -5.00283539e-01 4.03135002e-01 6.78943872e-01
2.42848173e-01 8.57355773e-01 -4.92298603e-01 9.56830084e-02
4.95880753e-01 4.48943861e-02 9.57507491e-01 -6.62958622e-01
-2.43799075e-01 -3.46010476e-01 5.19938692e-02 -1.48156524e+00
4.47743565e-01 -8.30862999e-01 1.81939647e-01 -1.77300370e+00
5.95439494e-01 4.89017159e-01 -4.50483747e-02 6.26505554e-01
-2.09226668e-01 -3.29126976e-02 5.54052234e-01 8.23533460e-02
-1.23517621e+00 8.49602997e-01 1.48360097e+00 -5.32583892e-01
1.24256752e-01 -7.31888533e-01 -6.21035159e-01 6.90346718e-01
4.84498024e-01 2.82855034e-01 -6.31133616e-01 -8.83438170e-01
3.50695163e-01 4.54223305e-01 7.29944050e-01 -5.64522147e-01
3.65317494e-01 -3.77284586e-01 2.80578136e-01 -5.66247344e-01
7.64671504e-01 -3.57088596e-01 -6.42378449e-01 9.29490849e-02
-7.73177564e-01 8.61873925e-02 3.90734017e-01 5.33065319e-01
-2.93959290e-01 3.20357108e-03 2.69070894e-01 -2.20035836e-01
-1.18611383e+00 2.75718302e-01 -2.84888834e-01 2.02159926e-01
7.17671752e-01 2.51136541e-01 -1.05889547e+00 -8.50468397e-01
-8.13578010e-01 8.16793919e-01 2.78923064e-01 7.82301962e-01
1.12889838e+00 -1.41548359e+00 -5.98620594e-01 8.77986252e-02
6.35131299e-01 -3.09127361e-01 5.45499146e-01 7.62654841e-01
-5.98084740e-02 7.32486725e-01 -4.35343245e-03 -8.75441611e-01
-1.07025731e+00 7.23419964e-01 1.90528154e-01 2.52265573e-01
-1.12737440e-01 7.91357338e-01 7.71446764e-01 -7.55829275e-01
5.18777788e-01 -2.30545461e-01 -9.60637406e-02 -6.00740239e-02
8.50228727e-01 -8.05369467e-02 -5.62221110e-01 -4.97374743e-01
-1.48975089e-01 5.95709980e-01 -9.46047716e-03 -5.69252074e-01
5.72294593e-01 -5.40533602e-01 -2.75367826e-01 9.56264079e-01
1.06514215e+00 -5.61930299e-01 -1.24697936e+00 -3.47934216e-01
-2.27265567e-01 -6.03669435e-02 -3.22251141e-01 -1.24947655e+00
-3.94164443e-01 1.39309621e+00 5.41966319e-01 -3.64726894e-02
1.09550178e+00 5.20894587e-01 7.86740422e-01 9.67185557e-01
2.51824766e-01 -8.93559337e-01 7.55312860e-01 7.74301946e-01
1.34108913e+00 -1.72236109e+00 -7.52525389e-01 1.65921807e-01
-1.21704221e+00 9.29901838e-01 9.91680861e-01 7.05876648e-01
-4.14573438e-02 -1.48286998e-01 1.43800378e-01 3.26755568e-02
-1.33330083e+00 -3.89552176e-01 5.37551403e-01 7.58578002e-01
1.41599193e-01 1.98198184e-01 2.89006889e-01 6.98938906e-01
1.08108081e-01 -5.65420166e-02 2.23233894e-01 6.68083847e-01
-4.69696105e-01 -5.07390618e-01 -4.56540108e-01 1.47426620e-01
2.16430694e-01 -4.79442179e-01 -7.36882508e-01 5.77159166e-01
-1.26850381e-01 1.15976524e+00 1.53508529e-01 -3.57339680e-01
3.04360151e-01 4.84633446e-01 6.05265081e-01 -5.32111883e-01
-7.46086538e-02 -2.04208061e-01 -4.52218018e-03 -6.06534958e-01
-1.79556236e-01 -2.40392268e-01 -1.33917999e+00 3.73479328e-03
3.23090643e-01 -1.75747588e-01 6.76632941e-01 1.11800075e+00
3.79931450e-01 2.44232386e-01 1.88876778e-01 -6.78397655e-01
-5.37682533e-01 -1.23606539e+00 -1.08294569e-01 5.11150241e-01
4.72285599e-01 -2.32210472e-01 -2.48567015e-01 3.69348347e-01] | [10.939318656921387, 1.4934486150741577] |
ca595ac3-98e5-48e9-b026-101ea5708dbf | fh-gan-face-hallucination-and-recognition | 1905.06537 | null | https://arxiv.org/abs/1905.06537v1 | https://arxiv.org/pdf/1905.06537v1.pdf | FH-GAN: Face Hallucination and Recognition using Generative Adversarial Network | There are many factors affecting visual face recognition, such as low resolution images, aging, illumination and pose variance, etc. One of the most important problem is low resolution face images which can result in bad performance on face recognition. Most of the general face recognition algorithms usually assume a sufficient resolution for the face images. However, in practice many applications often do not have sufficient image resolutions. The modern face hallucination models demonstrate reasonable performance to reconstruct high-resolution images from its corresponding low resolution images. However, they do not consider identity level information during hallucination which directly affects results of the recognition of low resolution faces. To address this issue, we propose a Face Hallucination Generative Adversarial Network (FH-GAN) which improves the quality of low resolution face images and accurately recognize those low quality images. Concretely, we make the following contributions: 1) we propose FH-GAN network, an end-to-end system, that improves both face hallucination and face recognition simultaneously. The novelty of this proposed network depends on incorporating identity information in a GAN-based face hallucination algorithm via combining a face recognition network for identity preserving. 2) We also propose a new face hallucination network, namely Dense Sparse Network (DSNet), which improves upon the state-of-art in face hallucination. 3) We demonstrate benefits of training the face recognition and GAN-based DSNet jointly by reporting good result on face hallucination and recognition. | ['Usman Ali', 'Te Qi', 'Hongtao Lu', 'Bayram Bayramli'] | 2019-05-16 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 2.77487218e-01 -8.42992887e-02 2.19763502e-01 -3.15536171e-01
-5.28813422e-01 -5.80764338e-02 5.33686459e-01 -1.26455200e+00
2.07111970e-01 8.26418042e-01 3.12590390e-01 3.96699250e-01
1.41587481e-01 -9.75349545e-01 -7.43521333e-01 -6.78379655e-01
4.27346349e-01 1.77767023e-01 -4.09299582e-01 -1.77128986e-01
-1.21358849e-01 8.73398066e-01 -1.83825898e+00 4.10213709e-01
7.69095898e-01 9.31886137e-01 -5.93340099e-02 3.30382317e-01
5.10626808e-02 8.50314558e-01 -5.81659377e-01 -4.52068686e-01
4.09272820e-01 -7.06741214e-01 -3.24219644e-01 3.34267765e-01
6.24007285e-01 -5.96343219e-01 -6.21564388e-01 1.19095969e+00
8.09919417e-01 -2.40840707e-02 6.02943003e-01 -1.24197602e+00
-1.35313666e+00 2.03469113e-01 -8.37701976e-01 -3.97575041e-03
4.13900733e-01 1.96428195e-01 6.92778677e-02 -1.34205651e+00
4.35108155e-01 1.67543733e+00 7.01269686e-01 9.87924278e-01
-1.06220305e+00 -1.06972897e+00 -3.72524559e-01 2.02860177e-01
-1.69124365e+00 -1.04408860e+00 7.73463726e-01 -6.64712042e-02
5.26564419e-01 1.27173677e-01 3.02917451e-01 1.17284822e+00
5.14175184e-02 2.74798214e-01 1.34148693e+00 -1.56112626e-01
-1.33796588e-01 4.26117033e-02 -6.47097468e-01 6.09482586e-01
1.92214862e-01 4.21469480e-01 -6.20113015e-01 6.72231838e-02
1.35112560e+00 2.49334067e-01 -5.07045567e-01 6.43443242e-02
-6.55298591e-01 5.25692225e-01 3.84529531e-01 1.46042600e-01
-5.57331502e-01 -1.60297483e-01 -3.88973542e-02 4.19369996e-01
3.31767768e-01 -9.37211514e-03 2.26699412e-01 2.98916817e-01
-9.48252559e-01 -9.00447369e-02 5.42382956e-01 7.55837858e-01
7.18942285e-01 7.33944058e-01 -2.73799896e-01 1.30228436e+00
3.09763104e-01 6.97774768e-01 5.15060186e-01 -8.60952497e-01
1.75744258e-02 2.21897244e-01 -9.14488509e-02 -1.09392977e+00
1.08873233e-01 -4.33266461e-01 -1.26749837e+00 5.23161352e-01
-1.02139473e-01 3.38524170e-02 -1.15854561e+00 1.86296666e+00
1.12207606e-01 7.22833037e-01 4.90789086e-01 1.05100739e+00
1.27949512e+00 5.81597984e-01 -3.87176313e-02 -5.14721990e-01
1.29720592e+00 -9.45571244e-01 -1.02652597e+00 -1.85622647e-02
-4.78660613e-01 -9.64530528e-01 9.11267579e-01 1.91654250e-01
-1.30684423e+00 -9.59218979e-01 -1.05004668e+00 4.37289178e-02
-1.55861035e-01 1.53699294e-01 3.35514873e-01 8.01243126e-01
-1.16565573e+00 3.22438359e-01 -3.87749851e-01 -2.53587335e-01
7.97621250e-01 5.26430309e-01 -7.75401831e-01 -5.08105636e-01
-1.12779152e+00 8.41442525e-01 -3.70851681e-02 3.24720800e-01
-1.24922693e+00 -6.21372283e-01 -8.58203292e-01 1.39009446e-01
2.15639055e-01 -7.29830921e-01 7.56408095e-01 -1.24550748e+00
-1.58146417e+00 7.57644892e-01 -2.76729316e-01 -7.54606277e-02
3.68591517e-01 -6.96961954e-02 -9.69756424e-01 1.34627029e-01
-1.92131996e-02 5.27250707e-01 1.26231754e+00 -1.50306535e+00
5.46388999e-02 -6.98525548e-01 -4.03030425e-01 2.41176218e-01
-3.14422071e-01 8.12050700e-02 -3.63949239e-01 -5.80294549e-01
-1.70893729e-01 -4.91201699e-01 2.39263505e-01 2.58541889e-02
-2.29144961e-01 3.20963472e-01 1.33684111e+00 -8.47748280e-01
6.78663313e-01 -2.29397750e+00 -1.18658006e-01 -8.89969170e-02
2.18496248e-01 6.18555427e-01 -4.75280195e-01 6.46309406e-02
-2.76130199e-01 9.67549980e-02 -1.18298560e-01 -3.11861187e-01
-2.61050731e-01 2.58665890e-01 -4.82073396e-01 5.00657141e-01
3.99420977e-01 1.07251894e+00 -4.09634680e-01 -2.24702179e-01
2.59590179e-01 1.29236174e+00 -3.89907807e-01 6.33950353e-01
1.77107647e-01 6.34114683e-01 -7.02717751e-02 7.88914800e-01
1.35580111e+00 -4.91412133e-02 -5.77286780e-02 -5.34246087e-01
9.98046398e-02 -4.15893316e-01 -1.06144750e+00 1.23069358e+00
-4.32996809e-01 1.58209994e-01 3.06408286e-01 -6.17475390e-01
1.15610015e+00 6.62405193e-01 3.03076297e-01 -7.62817323e-01
4.75520007e-02 1.22353271e-01 -2.45381728e-01 -4.73702252e-01
1.27339214e-01 -5.79406619e-01 5.08179426e-01 2.47386158e-01
1.52194336e-01 1.24654502e-01 -3.92297149e-01 -2.71211237e-01
7.25613117e-01 3.36230337e-03 2.00667962e-01 1.75233379e-01
7.94646502e-01 -7.27925718e-01 5.95668256e-01 2.65251875e-01
5.46037145e-02 1.09706545e+00 2.85091251e-01 -2.97750741e-01
-1.11120331e+00 -1.18575728e+00 -1.78577825e-01 5.33427417e-01
4.94060703e-02 1.14335313e-01 -7.82772005e-01 -3.55779409e-01
-2.19926715e-01 1.20976977e-01 -7.27082789e-01 -4.06026214e-01
-4.07639652e-01 -7.30854154e-01 7.84031510e-01 4.91762817e-01
1.16391039e+00 -1.33026946e+00 1.05922632e-01 -2.49036968e-01
-3.80868115e-03 -1.18859637e+00 -6.53541028e-01 -5.40904939e-01
-5.37016749e-01 -9.12356019e-01 -8.14338505e-01 -8.44581008e-01
8.93337905e-01 2.94280887e-01 8.44839513e-01 1.30732477e-01
-3.00322980e-01 2.04000622e-01 -7.74645284e-02 -7.49307871e-02
-3.75490576e-01 -3.98350865e-01 3.39266360e-01 4.73266482e-01
2.68500417e-01 -7.94690192e-01 -5.29606164e-01 4.57210481e-01
-1.03455508e+00 -1.53787032e-01 8.32872033e-01 1.12474585e+00
6.76315904e-01 2.63787389e-01 9.31048095e-01 -1.01888943e+00
6.81719840e-01 -4.53856319e-01 -2.72614956e-01 1.95274368e-01
-5.20274580e-01 -1.17414728e-01 1.04004622e+00 -3.89938772e-01
-1.25990212e+00 -4.86466661e-02 -4.55170214e-01 -1.04722989e+00
-2.24037096e-01 4.66074012e-02 -7.63444424e-01 -4.27039683e-01
5.36825359e-01 5.25315940e-01 5.43019116e-01 -4.31103915e-01
1.17443673e-01 7.20673740e-01 9.17451859e-01 -2.35714212e-01
1.03028476e+00 4.58756596e-01 7.05488175e-02 -8.42025876e-01
-5.97337186e-01 9.52829719e-02 -3.01772445e-01 -1.42656982e-01
7.00750709e-01 -1.17805016e+00 -8.69368196e-01 6.01349592e-01
-9.21949863e-01 1.05638765e-01 -1.29721016e-01 2.43730426e-01
-5.45575440e-01 2.60696530e-01 -5.41561544e-01 -8.99240792e-01
-4.59431857e-01 -1.12363327e+00 9.72667158e-01 6.17918015e-01
3.93024772e-01 -6.27971709e-01 -2.21699089e-01 4.88780767e-01
8.67408812e-01 4.58383560e-01 5.58892548e-01 2.81909071e-02
-5.97390413e-01 2.18647808e-01 -4.81569886e-01 5.19507527e-01
4.98427361e-01 -2.39299729e-01 -1.25083292e+00 -5.42348742e-01
7.78472945e-02 -4.15467203e-01 7.72115409e-01 1.71393037e-01
1.00431633e+00 -5.10756433e-01 -1.47611611e-02 1.01487827e+00
1.60867882e+00 1.85638160e-01 1.28725314e+00 -3.05990279e-01
9.01154935e-01 4.53794807e-01 2.64159948e-01 3.51949275e-01
-4.05845493e-02 8.25645208e-01 3.79589707e-01 -5.26716471e-01
-5.39236367e-01 -5.88423073e-01 6.17339492e-01 5.30027151e-01
-2.76979595e-01 3.63566726e-02 -3.01655233e-01 2.78544933e-01
-1.54055774e+00 -1.27417910e+00 3.84095460e-01 2.19866967e+00
6.44696832e-01 -6.29347920e-01 -2.80793700e-02 -1.93673559e-03
8.69704545e-01 8.17843303e-02 -7.54468739e-01 -1.30361080e-01
-3.22741061e-01 6.08965218e-01 8.00336450e-02 4.27385837e-01
-5.51360130e-01 9.42842245e-01 6.19345713e+00 6.91255212e-01
-1.35946071e+00 2.79059708e-01 7.06517875e-01 -6.85025975e-02
-1.83962405e-01 -4.76695210e-01 -7.41849184e-01 4.55891937e-01
7.59302318e-01 -2.25547582e-01 8.96438420e-01 8.00361216e-01
1.39927030e-01 5.37717223e-01 -8.82311404e-01 1.41742158e+00
6.75019324e-01 -1.04438949e+00 5.52564025e-01 8.86683688e-02
7.82660544e-01 -5.89469016e-01 6.14258707e-01 3.44818383e-01
2.26627104e-02 -2.00233579e+00 1.98082566e-01 7.40346372e-01
1.43516624e+00 -1.13554037e+00 6.70885444e-01 2.58144531e-02
-1.12845957e+00 1.88540407e-02 -7.44548142e-01 2.68342674e-01
2.42870636e-02 3.68264675e-01 -5.80031574e-01 5.20577669e-01
5.43668091e-01 6.58893526e-01 -4.14673299e-01 5.35021722e-01
-1.53173089e-01 2.65226752e-01 1.03258416e-01 7.21403062e-01
-3.20475221e-01 -1.45724759e-01 3.30362171e-01 6.32067621e-01
5.49310982e-01 3.95091623e-01 -1.23304829e-01 1.29930317e+00
-6.25151157e-01 -4.37591784e-02 -9.26302433e-01 8.40407982e-02
5.74964285e-01 1.23470008e+00 -1.22507654e-01 -3.72456685e-02
-5.27752876e-01 1.05984318e+00 1.39079541e-01 4.51593518e-01
-8.00984681e-01 -1.92702711e-02 8.85627866e-01 3.57583165e-01
3.23124141e-01 1.10217534e-01 -3.07435468e-02 -1.16526580e+00
4.44125980e-02 -1.28756404e+00 1.36388376e-01 -8.49447787e-01
-1.39142096e+00 9.94525909e-01 -5.18784344e-01 -1.12055969e+00
-2.25336447e-01 -4.01035070e-01 -4.64631647e-01 1.20190156e+00
-1.73069203e+00 -1.42986310e+00 -7.22466886e-01 1.05945218e+00
5.31332076e-01 -6.93629920e-01 8.72914374e-01 6.71669900e-01
-5.97112298e-01 9.71272051e-01 -1.50905222e-01 2.07627252e-01
8.12012017e-01 -5.83418489e-01 1.46620527e-01 8.92968774e-01
-1.20683974e-02 6.68523610e-01 3.45727503e-01 -5.89237571e-01
-1.64203238e+00 -1.47427523e+00 3.88101697e-01 -2.00920731e-01
-2.52881140e-01 -1.26955852e-01 -1.06394613e+00 5.74767649e-01
1.49772063e-01 3.76318038e-01 6.02138221e-01 -3.24679226e-01
-4.51998621e-01 -3.11620861e-01 -1.73804367e+00 4.66649115e-01
1.08596849e+00 -7.53405571e-01 -2.54346550e-01 -1.48008680e-02
4.02397186e-01 -2.40784273e-01 -1.00348091e+00 6.96468055e-01
6.76306665e-01 -1.10743105e+00 1.21115875e+00 -2.43767604e-01
5.95577836e-01 -5.68107486e-01 -1.78509116e-01 -1.19279253e+00
-3.75981241e-01 -4.95888948e-01 -1.15358219e-01 1.41680515e+00
3.82286459e-02 -7.14524627e-01 7.85153866e-01 3.30130011e-01
2.43815079e-01 -5.71536660e-01 -8.20711672e-01 -7.46290028e-01
-1.68254495e-01 3.16657513e-01 9.88316059e-01 9.86453116e-01
-6.22952938e-01 3.95513147e-01 -1.03439009e+00 3.43284309e-01
9.02473032e-01 3.67543288e-03 7.65133440e-01 -9.80289817e-01
-2.78159648e-01 1.33531773e-03 -4.25388783e-01 -6.87738478e-01
3.13781112e-01 -6.21486664e-01 -1.41222164e-01 -1.31031406e+00
4.70097691e-01 -2.88965944e-02 -1.35190964e-01 4.40405041e-01
4.88347113e-02 9.91876245e-01 2.55234931e-02 2.76569545e-01
-1.40097409e-01 8.22414637e-01 1.64434099e+00 5.14751635e-02
4.30141948e-02 -2.33802319e-01 -9.64539707e-01 3.61268282e-01
5.72899282e-01 -1.12769634e-01 -5.00552952e-01 -3.54773343e-01
-4.47154880e-01 4.28977579e-01 4.24712360e-01 -1.16978717e+00
1.61449641e-01 -6.46456853e-02 1.01426566e+00 -4.91908900e-02
5.85529625e-01 -9.12443399e-01 8.30103695e-01 2.08246186e-01
-9.16572884e-02 -6.61776066e-02 2.07821995e-01 3.13888818e-01
-5.08446038e-01 2.05112562e-01 1.35273361e+00 -2.02908978e-01
-5.96131742e-01 7.57394910e-01 1.57110970e-02 -2.70974547e-01
1.05881214e+00 -4.09392923e-01 -3.75860065e-01 -6.01329684e-01
-7.20410943e-01 -2.60672718e-01 6.47290707e-01 5.80653369e-01
1.04776847e+00 -1.80305254e+00 -1.03037274e+00 8.50141287e-01
-2.19790950e-01 -2.22615823e-01 4.37436253e-01 5.72186887e-01
-2.88954049e-01 9.32942554e-02 -8.71906519e-01 -1.38009235e-01
-1.47281706e+00 6.89535558e-01 4.48686481e-01 6.28896281e-02
-4.05139923e-01 5.60520113e-01 4.57064480e-01 -2.66058028e-01
-3.67914587e-02 5.77070713e-01 -4.81524765e-01 -3.81760955e-01
9.35846388e-01 2.24246994e-01 -1.48696929e-01 -1.05392051e+00
-2.31476545e-01 8.60035360e-01 -1.75400630e-01 -4.60740440e-02
1.35574806e+00 -4.97202389e-02 -3.59782547e-01 -1.52926177e-01
1.15245390e+00 7.73855904e-03 -1.18362880e+00 -1.09255344e-01
-8.76728296e-01 -9.72186863e-01 -4.16090190e-02 -8.48623037e-01
-1.62762833e+00 7.83168852e-01 7.74736524e-01 -2.11091474e-01
1.63934600e+00 -2.67462045e-01 7.95574725e-01 -1.49599344e-01
4.46335733e-01 -6.24165654e-01 3.60168934e-01 2.69229591e-01
1.16815138e+00 -1.21260929e+00 -1.48893565e-01 -6.37909412e-01
-5.47217131e-01 9.80326653e-01 1.00975764e+00 -9.85119715e-02
4.88495648e-01 5.43487191e-01 1.48380503e-01 -8.42156783e-02
-5.88441312e-01 -3.41870822e-02 1.98023289e-01 8.83036494e-01
4.04548138e-01 -2.44204283e-01 1.61960479e-02 6.23688638e-01
-1.81608170e-01 2.13018119e-01 2.99660474e-01 2.64158785e-01
-1.90961644e-01 -1.10392499e+00 -6.97851717e-01 3.19914907e-01
-6.25066280e-01 -1.13787226e-01 -3.56158674e-01 4.29410547e-01
3.35228682e-01 8.92123640e-01 -5.73238544e-02 -5.67449927e-01
3.74930382e-01 -1.58238947e-01 7.04492867e-01 -5.78459620e-01
-3.06802452e-01 -1.72906980e-01 -3.59125197e-01 -4.36426342e-01
-3.16483587e-01 -2.25541309e-01 -8.91826630e-01 -7.89470673e-01
-6.34089485e-02 2.67206430e-02 3.63091022e-01 6.00460112e-01
5.50752044e-01 5.34792125e-01 8.86511445e-01 -7.28699267e-01
-4.34287339e-01 -9.61858034e-01 -8.62773359e-01 5.40716946e-01
3.80185366e-01 -6.64243937e-01 -3.00549299e-01 2.03583032e-01] | [12.826318740844727, -0.022560568526387215] |
3b16cf67-4243-4b67-83b0-e8154cb71ca7 | mitoem-dataset-large-scale-3d-mitochondria | null | null | https://donglaiw.github.io/paper/2020_miccai_mitoEM.pdf | https://donglaiw.github.io/paper/2020_miccai_mitoEM.pdf | MitoEM Dataset: Large-scale 3D Mitochondria Instance Segmentation from EM Images | Electron microscopy (EM) allows the identification of intracellular organelles such as mitochondria, providing insights for clinical and scientific studies. However, public mitochondria segmentation datasets only contain hundreds of instances with simple shapes. It is unclear if existing methods achieving human-level accuracy on these small datasets are robust in practice. To this end, we introduce the MitoEM dataset, a 3D mitochondria instance segmentation dataset with two (30um)^3 volumes from human and rat cortices respectively, 3,600x larger than previous benchmarks. With around 40K instances, we find a great diversity of mitochondria in terms of shape and density. For evaluation, we tailor the implementation of the average precision (AP) metric for 3D data with a 45x speedup. On MitoEM, we find existing instance segmentation methods often fail to correctly segment mitochondria with complex shapes or close contacts with other instances. Thus, our MitoEM dataset poses new challenges to the field. We release our code and data: https://donglaiw.github.io/page/mitoEM/index.html. | ['Hanspeter Pfister', 'Jeff Lichtman', 'Ignacio Arganda-Carreras', 'Xueying Wang', 'Won-Dong Jang', 'Aarush Gupta', 'Xin Huang', 'Wenjie Yin', 'Xingyu Liu', 'Nils Wendt', 'Daniel Franco-Barranco', 'Zudi Lin', 'Donglai Wei'] | 2020-10-04 | null | null | null | medical-image-computing-and-computer-assisted-2 | ['3d-instance-segmentation-1'] | ['computer-vision'] | [-1.28767326e-01 -1.32839799e-01 -9.76710096e-02 -1.19962774e-01
-8.49227726e-01 -8.77035499e-01 2.22281277e-01 3.77203822e-01
-3.75895351e-01 1.20703149e+00 -2.44768247e-01 -2.99495757e-01
1.43646896e-01 -3.57845902e-01 -6.93873644e-01 -8.24733019e-01
-9.03753862e-02 9.25805449e-01 2.44439676e-01 4.79066759e-01
5.80182910e-01 5.72893083e-01 -9.21043634e-01 1.83648989e-01
9.39330399e-01 5.94184995e-01 5.20414472e-01 1.04780173e+00
8.46776925e-03 2.34907120e-01 -5.07215142e-01 -2.39954516e-01
1.62861988e-01 -1.24914385e-01 -1.18234241e+00 -5.60524091e-02
5.80000997e-01 -1.42375469e-01 -1.79730847e-01 9.17348266e-01
8.01451802e-01 -4.86680597e-01 8.57825816e-01 -1.51751626e+00
-6.25435948e-01 6.79111063e-01 -5.05754232e-01 9.42671001e-01
1.29501536e-01 4.19352591e-01 5.88506579e-01 -8.18658113e-01
8.99062216e-01 9.95246649e-01 8.68706763e-01 5.30753016e-01
-1.49442756e+00 -4.02559489e-01 -3.85602057e-01 2.97676563e-01
-1.33506846e+00 -5.28098404e-01 1.06488131e-01 -5.69328368e-01
9.00049031e-01 3.55755657e-01 8.83436739e-01 8.97578895e-01
4.13146228e-01 7.93058157e-01 1.32552397e+00 2.71453619e-01
2.43807435e-01 -2.70305544e-01 1.18842959e-01 4.31385070e-01
4.93061483e-01 -7.07966506e-01 -1.59503013e-01 -2.55202502e-01
9.72832382e-01 4.32218276e-02 -2.85246521e-01 -3.19524378e-01
-2.10355377e+00 3.15351397e-01 1.52575091e-01 1.82755858e-01
9.62950382e-03 4.05829996e-01 6.51329815e-01 7.50469835e-03
3.28291729e-02 3.15131038e-01 -6.59894466e-01 -5.08584380e-01
-6.87195420e-01 3.09545338e-01 6.07695282e-01 1.04169202e+00
4.81119394e-01 -4.96247768e-01 2.72971779e-01 7.16943502e-01
-1.15577159e-02 6.07456923e-01 2.25302756e-01 -1.43927991e+00
3.20354015e-01 6.50636494e-01 4.51267771e-02 -6.68287754e-01
-7.78489947e-01 -7.56284297e-02 -1.01944053e+00 -2.14717329e-01
9.55509484e-01 1.18929267e-01 -6.24266803e-01 1.32854772e+00
6.82753861e-01 1.38329923e-01 -1.35525540e-01 7.84157395e-01
1.02765596e+00 2.68410802e-01 -1.27045229e-01 -3.91420573e-02
1.49989629e+00 -7.12257624e-01 -3.56740385e-01 1.05781496e-01
9.06548679e-01 -9.13612247e-01 1.15809226e+00 2.01985776e-01
-1.08985853e+00 -2.12464705e-02 -6.47155762e-01 -3.69753927e-01
-3.55094403e-01 1.10796548e-01 5.70294917e-01 2.19735131e-01
-1.09071290e+00 7.38455594e-01 -1.08813882e+00 -8.71386707e-01
1.03744936e+00 4.44601625e-01 -4.87015605e-01 8.32036361e-02
-2.91902691e-01 7.68268287e-01 1.93638891e-01 -4.57889974e-01
-8.41964424e-01 -1.23502541e+00 -4.95305002e-01 -1.88426882e-01
-3.50320898e-02 -8.15037429e-01 9.35630441e-01 1.55404851e-01
-9.20302451e-01 1.27649605e+00 -2.69442201e-02 -5.60449362e-01
5.48694372e-01 3.69906604e-01 2.10673347e-01 5.86758554e-01
2.38351762e-01 1.00525403e+00 6.24164566e-02 -1.11509979e+00
-3.80335778e-01 -5.11016488e-01 -1.84680969e-01 -1.00934923e-01
1.35928854e-01 -5.34547679e-02 -1.77019536e-01 -3.15688014e-01
1.54881611e-01 -9.42208350e-01 -3.57564203e-02 3.14039230e-01
-6.55948997e-01 -1.87695786e-01 6.74058914e-01 -5.15009463e-01
5.85052013e-01 -1.88492715e+00 5.67035601e-02 -3.29652727e-01
6.95667922e-01 -3.54541391e-02 7.46108741e-02 2.71232307e-01
2.22640201e-01 4.78665411e-01 -3.64742815e-01 -4.67468083e-01
1.68393347e-02 2.10524291e-01 2.09069937e-01 1.10277736e+00
1.19716683e-02 1.08333659e+00 -1.12534714e+00 -9.69966352e-01
1.74753174e-01 5.63791454e-01 -2.77056545e-01 -2.46410564e-01
1.12666026e-01 7.59918749e-01 -2.91692168e-01 1.17771518e+00
7.24659920e-01 -9.99436855e-01 2.48681590e-01 -2.49505222e-01
8.01606253e-02 2.73332059e-01 -8.02587628e-01 1.73687744e+00
2.94518173e-01 6.49629831e-01 2.83697724e-01 -8.64574909e-01
4.01812553e-01 7.34932870e-02 9.53143358e-01 -2.87362844e-01
2.97543913e-01 3.80881280e-01 7.11615235e-02 -2.04432473e-01
-1.09896706e-02 6.53345883e-02 9.69480127e-02 4.58815634e-01
9.42749716e-03 -2.77933329e-01 5.32482684e-01 4.27665561e-01
1.47306657e+00 -8.02996457e-02 2.15467647e-01 -7.18548417e-01
1.26410231e-01 1.21637702e-01 7.73906410e-01 2.99917251e-01
-8.51356745e-01 8.31131816e-01 6.83499575e-01 -4.82701927e-01
-1.43775547e+00 -1.24883020e+00 -7.85590529e-01 3.58383715e-01
2.62332529e-01 -4.72115725e-01 -9.24771011e-01 -5.11182189e-01
2.77127504e-01 -1.61843777e-01 -5.76169014e-01 4.41008329e-01
-3.48820925e-01 -1.26350451e+00 9.34908390e-01 5.37968814e-01
1.96140975e-01 -1.01533568e+00 -7.73299992e-01 6.19013309e-02
-5.16618133e-01 -1.47741699e+00 -5.17084181e-01 2.68220663e-01
-1.22729957e+00 -1.51006889e+00 -7.42175400e-01 -6.79101229e-01
7.86421776e-01 1.71810403e-01 1.31710422e+00 5.91356695e-01
-7.41415501e-01 1.55917898e-01 -9.15121213e-02 -2.76284158e-01
-2.16807589e-01 2.23328695e-01 3.04911852e-01 -8.45845580e-01
1.13382809e-01 -9.77566302e-01 -9.73461688e-01 4.29275513e-01
-4.89486158e-01 1.53738931e-01 2.33617187e-01 5.06243467e-01
1.15900099e+00 -3.85952592e-01 7.92764783e-01 -6.90368652e-01
6.80845827e-02 -6.53848410e-01 -4.10180151e-01 -1.20189972e-01
-3.79624844e-01 -4.56100315e-01 6.19090199e-01 -4.41863388e-01
1.97728425e-02 6.00466095e-02 -1.68622002e-01 -3.26950461e-01
-2.88906306e-01 -2.21182868e-01 -1.05007328e-01 -2.29533553e-01
4.12564814e-01 2.51792729e-01 2.50951082e-01 -4.53135073e-01
1.02851698e-02 4.48040783e-01 6.43386185e-01 -7.27404952e-01
2.93693006e-01 8.53888571e-01 1.68262377e-01 -6.65309608e-01
-4.60960001e-01 -6.44827962e-01 -8.67866695e-01 -4.22002608e-03
7.77536154e-01 -7.67865241e-01 -1.13214469e+00 5.28386891e-01
-8.29397500e-01 -9.55303431e-01 -3.49637300e-01 3.67134720e-01
-1.02993619e+00 7.08573937e-01 -1.20847476e+00 -3.75175774e-01
-4.82635409e-01 -1.34125793e+00 1.19871902e+00 2.26049960e-01
-1.95960060e-01 -9.25270140e-01 2.74656843e-02 6.55837834e-01
2.17954139e-03 3.17146122e-01 1.01410401e+00 -6.05096877e-01
-5.93310416e-01 4.44133192e-01 -4.55108166e-01 -2.19727293e-01
2.64580138e-02 2.93989301e-01 -7.29591191e-01 -3.19875538e-01
-4.40271914e-01 -3.46756309e-01 6.04912043e-01 4.83036667e-01
1.39994669e+00 -3.00786514e-02 -7.33723521e-01 8.16329539e-01
1.40819442e+00 -2.86909252e-01 7.76876271e-01 3.22279304e-01
5.98817587e-01 6.75461665e-02 5.01721203e-01 4.93533075e-01
6.88580871e-01 3.39071840e-01 6.93854272e-01 -9.53151099e-03
-2.36587361e-01 4.49171811e-01 -1.91728100e-02 1.00246882e+00
-1.85003746e-02 -1.10405423e-01 -1.09206152e+00 8.74159455e-01
-1.57604587e+00 -8.91706109e-01 -5.61726987e-01 1.94011974e+00
1.03215730e+00 -1.08652912e-01 5.74865639e-01 1.25574976e-01
7.82157660e-01 -5.21869600e-01 -9.15559113e-01 -1.59470998e-02
-1.09075550e-02 -2.04376310e-01 5.07216811e-01 1.20295689e-01
-9.97189343e-01 4.89646167e-01 6.73171997e+00 8.60268354e-01
-6.62220061e-01 1.20982587e-01 9.96357679e-01 -2.72673786e-01
-4.28961217e-02 -1.29889235e-01 -8.48348081e-01 9.93824303e-01
8.73222113e-01 -5.60878292e-02 5.68646014e-01 4.33538944e-01
2.51516253e-01 -2.36741886e-01 -1.08069968e+00 1.18063200e+00
-3.20958376e-01 -1.65227556e+00 -3.01197946e-01 6.23703539e-01
5.81223786e-01 6.66410029e-01 -1.65093154e-01 -1.90954506e-02
5.86832576e-02 -1.16976714e+00 3.26928377e-01 5.93655825e-01
8.04996431e-01 -6.80794477e-01 6.86065018e-01 4.62433547e-01
-1.15373528e+00 3.28242987e-01 -8.94931376e-01 2.80262947e-01
1.01519592e-01 9.46220458e-01 -9.36404884e-01 1.90003052e-01
1.01434910e+00 7.67638862e-01 -7.58793235e-01 1.51673901e+00
6.08940005e-01 4.89917099e-01 -7.48939574e-01 1.86103895e-01
-2.47971669e-01 -4.46084946e-01 5.02244711e-01 1.44889879e+00
3.47835600e-01 5.32182381e-02 1.21614642e-01 1.13686013e+00
-5.30937433e-01 1.53279126e-01 -4.52122867e-01 -2.10448846e-01
7.41810739e-01 1.84280789e+00 -1.58334041e+00 -5.14572501e-01
1.77203864e-02 7.63214409e-01 5.12126625e-01 -4.98604812e-02
-1.05521739e+00 -2.08202284e-02 9.86499190e-01 2.64218926e-01
2.00200811e-01 -1.85402855e-01 -6.14337325e-01 -1.01650512e+00
-2.17025191e-01 -6.88893676e-01 1.40352502e-01 -8.62679958e-01
-1.61494923e+00 -3.09220850e-02 -4.60616440e-01 -8.12916994e-01
5.89398742e-01 -5.10688245e-01 -5.02310634e-01 2.99365580e-01
-1.16460204e+00 -9.71548975e-01 -2.60476381e-01 3.18133771e-01
3.96836758e-01 3.86803508e-01 6.11044645e-01 5.79172134e-01
-6.65557921e-01 2.82797843e-01 3.46932113e-01 1.92635000e-01
6.56445265e-01 -1.71616709e+00 5.32440722e-01 1.35830805e-01
-8.01889747e-02 6.67070150e-01 5.74308932e-01 -6.30588353e-01
-1.64437091e+00 -1.24127328e+00 6.73851848e-01 -1.06816041e+00
6.73495471e-01 -3.30538690e-01 -7.61879086e-01 7.30689287e-01
7.09306747e-02 6.18141770e-01 7.82389939e-01 -3.51238847e-01
9.25569609e-02 3.55148256e-01 -1.66123796e+00 5.27829170e-01
1.26786745e+00 -2.83371061e-01 -1.37741819e-01 7.54995704e-01
4.07245100e-01 -4.42213953e-01 -1.81189430e+00 4.46556121e-01
4.90538806e-01 -9.15209055e-01 1.19004512e+00 -2.90869921e-01
3.35172921e-01 -6.18721724e-01 -1.66099310e-01 -9.66313958e-01
-2.04200074e-01 -3.31591755e-01 -4.89088118e-01 1.06551814e+00
1.87310845e-01 -5.43271959e-01 8.61649454e-01 -3.98554392e-02
-3.76232445e-01 -1.38583851e+00 -1.06352353e+00 -7.24826992e-01
4.80184406e-01 -2.19223164e-02 6.97690070e-01 1.03056252e+00
3.44966084e-01 2.17280209e-01 3.24792892e-01 -1.31521314e-01
9.26719546e-01 2.81976938e-01 8.77821684e-01 -1.28797579e+00
1.20092168e-01 -4.25240248e-01 -6.68172181e-01 -9.27603185e-01
-4.25959006e-02 -1.30965996e+00 -1.12705797e-01 -1.91200423e+00
1.04371536e+00 -5.65901935e-01 3.45676090e-03 5.54197311e-01
-8.35024565e-02 9.32185888e-01 1.42228916e-01 4.04814303e-01
-1.04064214e+00 1.16210833e-01 1.69207025e+00 1.58216245e-02
4.18747902e-01 -4.89321649e-01 -5.74138165e-01 7.10563958e-01
1.15994561e+00 -4.25166577e-01 -1.33559525e-01 -2.14519382e-01
1.29869789e-01 -2.02119336e-01 7.19592333e-01 -1.21648324e+00
-3.55665199e-02 -1.96633235e-01 7.70173192e-01 -1.01391721e+00
3.60070258e-01 -3.56436521e-01 2.53096998e-01 5.15515447e-01
1.55870825e-01 2.51333892e-01 -1.78535357e-02 3.22656989e-01
6.69725478e-01 7.75957480e-02 1.06991124e+00 -3.50748509e-01
-1.68997914e-01 5.63307822e-01 -6.94816709e-01 6.06172085e-01
9.99804854e-01 -4.52926487e-01 -9.50934350e-01 2.69793361e-01
-7.85790741e-01 4.20758009e-01 1.33057666e+00 -3.83970797e-01
4.59424406e-01 -1.28683341e+00 -4.46004689e-01 -3.55204642e-01
1.13633454e-01 2.77432442e-01 1.69167802e-01 1.69353449e+00
-8.25263619e-01 2.97206163e-01 -4.96328145e-01 -1.11688197e+00
-1.32235348e+00 4.63039488e-01 4.03651774e-01 -6.14800490e-03
-1.01111341e+00 5.44112146e-01 -1.41650990e-01 -8.11183155e-01
-4.42681424e-02 -8.61431837e-01 1.20305799e-01 -2.23798200e-01
4.31056619e-01 8.15967381e-01 -1.11324772e-01 -6.36326253e-01
-6.10809445e-01 4.98507857e-01 3.03904116e-01 5.97658515e-01
1.39065480e+00 -4.92155343e-01 -4.97821718e-01 5.67587137e-01
1.03668010e+00 -9.22671482e-02 -1.21952677e+00 4.59494799e-01
-3.39930654e-02 -3.18785429e-01 -6.00117624e-01 -6.89137936e-01
-1.12004769e+00 8.74951720e-01 5.01416385e-01 1.18850864e-01
6.56554341e-01 4.36856657e-01 1.19535983e+00 3.23406875e-01
6.59502923e-01 -1.00475645e+00 -1.07227378e-01 4.53307092e-01
3.72598588e-01 -1.10655260e+00 1.84312209e-01 -4.23909426e-01
-2.37188786e-01 8.91255319e-01 6.12514257e-01 -9.61766839e-02
3.69895190e-01 6.91791713e-01 -1.49331288e-03 -3.75699908e-01
-9.34767485e-01 2.13778708e-02 -3.63656342e-01 8.33780468e-01
4.10530388e-01 1.97087646e-01 -2.44835138e-01 4.23553407e-01
-2.89946854e-01 -3.08903456e-02 8.85414183e-01 7.35875547e-01
-6.52079761e-01 -8.36892843e-01 -3.24236810e-01 9.67160881e-01
-7.24870324e-01 -3.49769997e-03 -3.55807930e-01 6.40659630e-01
7.26040900e-02 5.16254783e-01 1.33152843e-01 -2.92447694e-02
-3.11458498e-01 -6.24585338e-02 9.66253877e-01 -4.63815898e-01
-2.68634111e-01 -8.77757370e-02 -8.96768793e-02 -6.18614316e-01
-2.70465285e-01 -8.19764912e-01 -2.15345693e+00 -8.82213771e-01
-9.44928732e-03 -5.99965686e-03 5.33468604e-01 8.46150517e-01
6.19471550e-01 2.66568780e-01 -3.85170206e-02 -9.92242336e-01
-2.33622909e-01 -8.57128680e-01 -7.79788971e-01 3.66567373e-01
9.13039520e-02 -6.13996089e-01 -3.56684834e-01 3.72865617e-01] | [14.301351547241211, -3.132519245147705] |
542a6ae6-e0f5-4251-8219-8448e7ca2eee | firerisk-a-remote-sensing-dataset-for-fire | 2303.07035 | null | https://arxiv.org/abs/2303.07035v1 | https://arxiv.org/pdf/2303.07035v1.pdf | FireRisk: A Remote Sensing Dataset for Fire Risk Assessment with Benchmarks Using Supervised and Self-supervised Learning | In recent decades, wildfires, as widespread and extremely destructive natural disasters, have caused tremendous property losses and fatalities, as well as extensive damage to forest ecosystems. Many fire risk assessment projects have been proposed to prevent wildfires, but GIS-based methods are inherently challenging to scale to different geographic areas due to variations in data collection and local conditions. Inspired by the abundance of publicly available remote sensing projects and the burgeoning development of deep learning in computer vision, our research focuses on assessing fire risk using remote sensing imagery. In this work, we propose a novel remote sensing dataset, FireRisk, consisting of 7 fire risk classes with a total of 91872 labelled images for fire risk assessment. This remote sensing dataset is labelled with the fire risk classes supplied by the Wildfire Hazard Potential (WHP) raster dataset, and remote sensing images are collected using the National Agriculture Imagery Program (NAIP), a high-resolution remote sensing imagery program. On FireRisk, we present benchmark performance for supervised and self-supervised representations, with Masked Autoencoders (MAE) pre-trained on ImageNet1k achieving the highest classification accuracy, 65.29%. This remote sensing dataset, FireRisk, provides a new direction for fire risk assessment, and we make it publicly available on https://github.com/CharmonyShen/FireRisk. | ['Michael Kirley', 'Xinye Wanyan', 'Sachith Seneviratne', 'Shuchang Shen'] | 2023-03-13 | null | null | null | null | ['remote-sensing-image-classification'] | ['miscellaneous'] | [ 5.10122359e-01 -2.54964381e-01 -1.55120715e-01 -1.57138258e-01
-2.35235959e-01 -3.34121823e-01 7.82701492e-01 5.39402403e-02
-4.98004317e-01 9.00133908e-01 5.31675696e-01 -7.68105626e-01
-3.01770478e-01 -1.76931703e+00 -3.20283204e-01 -6.87013149e-01
-5.39209545e-01 5.74929686e-03 -2.54415840e-01 -6.37792051e-01
-2.07477465e-01 7.65709698e-01 -1.68276429e+00 1.03259332e-01
8.23866546e-01 7.94098854e-01 4.42540884e-01 5.57356298e-01
2.67595947e-01 2.38546565e-01 -4.02642369e-01 2.49354646e-01
4.89775956e-01 -1.31900236e-01 -8.18032622e-01 -2.51027614e-01
1.16195336e-01 -8.23787689e-01 -3.23893040e-01 1.21766293e+00
4.68645751e-01 1.80001646e-01 8.30439627e-01 -1.11773133e+00
-8.33287477e-01 8.98312449e-01 -6.77146018e-01 4.19445574e-01
-3.73759478e-01 3.18527162e-01 7.92114258e-01 -8.80526721e-01
2.73838758e-01 1.15163386e+00 6.72451079e-01 1.57562122e-01
-1.12780297e+00 -1.08671248e+00 1.54191092e-01 1.38556287e-01
-1.70303297e+00 -2.10055232e-01 3.79670918e-01 -6.77649140e-01
1.18852258e+00 3.03468645e-01 5.54929256e-01 9.02537405e-01
2.87053853e-01 1.95906907e-01 1.11853600e+00 -8.19017217e-02
2.05626100e-01 -6.78750932e-01 -3.13458502e-01 2.32724145e-01
2.35659152e-01 1.02520716e+00 2.15737984e-01 -1.05987906e-01
1.09757245e+00 7.74970651e-01 -3.13987643e-01 4.49199975e-01
-1.18696916e+00 1.09238923e+00 1.36570680e+00 4.07004029e-01
-9.06400442e-01 8.45964719e-03 1.73025221e-01 2.75622517e-01
6.95678771e-01 1.01244278e-01 -4.67053503e-01 3.70394826e-01
-8.87021244e-01 2.52975762e-01 2.09410071e-01 2.96212018e-01
9.42518353e-01 3.46265763e-01 3.37782145e-01 9.97016251e-01
3.80523473e-01 1.05303502e+00 2.58025050e-01 -6.99735820e-01
1.22515343e-01 5.51080644e-01 1.09124102e-01 -1.47900844e+00
-4.58701462e-01 -1.75047711e-01 -1.73392820e+00 7.60605991e-01
2.79141869e-02 -2.38786444e-01 -1.18513250e+00 1.52875257e+00
-7.57232308e-02 -1.71250388e-01 6.49724081e-02 1.15778875e+00
5.41707516e-01 1.13161755e+00 7.16136098e-01 -5.56303561e-02
1.05285776e+00 -4.35953528e-01 -4.26588804e-01 -1.80977449e-01
5.98800965e-02 -2.01745525e-01 8.72034490e-01 2.27515429e-01
-2.11068228e-01 -4.53000426e-01 -9.72176015e-01 5.45974076e-01
-1.02707708e+00 3.90704684e-02 8.57229173e-01 2.59188861e-01
-9.86426115e-01 7.01072395e-01 -7.35648155e-01 -7.21977711e-01
7.13550031e-01 -1.84599578e-01 -3.11398774e-01 8.22424963e-02
-1.58603549e+00 9.95856583e-01 6.72724247e-01 6.98780537e-01
-1.23507583e+00 -6.18134499e-01 -7.77448118e-01 3.87593240e-01
1.89980060e-01 -3.43556963e-02 6.04009569e-01 -7.04704285e-01
-7.28217423e-01 6.98731184e-01 6.27402484e-01 -3.98610771e-01
6.76734149e-02 -2.84228683e-01 -8.89388204e-01 -1.47036791e-01
4.30496544e-01 8.65372300e-01 4.68307018e-01 -1.12302864e+00
-8.07406366e-01 -5.19330978e-01 -1.32868513e-01 8.52454640e-03
-3.03109407e-01 2.59983599e-01 9.10600245e-01 -1.15879941e+00
1.84962079e-02 -5.87510228e-01 -6.91999197e-01 9.46821738e-03
-9.13041905e-02 1.48567781e-01 6.58976316e-01 -7.48467267e-01
1.22320580e+00 -2.07310677e+00 -2.92312771e-01 1.25914529e-01
3.08764894e-02 6.60958290e-01 -4.15396959e-01 5.06263375e-01
-4.34152305e-01 7.30008185e-01 -1.02907991e+00 8.87133956e-01
-2.22256690e-01 5.04855573e-01 -7.67616570e-01 2.37898752e-01
5.43819129e-01 5.51767230e-01 -8.87536645e-01 1.01414777e-01
4.86221462e-01 5.59275806e-01 9.61514562e-03 1.10809363e-01
-2.10700661e-01 1.80533364e-01 -4.04776633e-01 7.64983475e-01
1.15390313e+00 2.32575580e-01 -5.40930331e-02 1.64700434e-01
-4.78463829e-01 4.83514369e-02 -1.06520927e+00 1.13151956e+00
-2.68754423e-01 2.68657148e-01 -3.39071080e-02 -8.58232915e-01
1.26232851e+00 2.55702823e-01 4.88493562e-01 -3.65546882e-01
-1.78270102e-01 1.65523350e-01 -3.04079443e-01 -2.08686829e-01
3.17254573e-01 -2.86724120e-01 -7.70578440e-03 6.73114717e-01
-1.25046834e-01 1.25808194e-01 -3.26435380e-02 -3.18715274e-01
9.81777012e-01 -6.79154471e-02 6.34661019e-01 -3.27455312e-01
1.48128077e-01 4.60928172e-01 6.26581073e-01 5.22620678e-01
-5.85572541e-01 4.51063216e-01 -2.18347490e-01 -1.23685265e+00
-8.95599067e-01 -1.38189530e+00 -5.70036411e-01 1.17471564e+00
-2.60567427e-01 -4.63133268e-02 -2.48056114e-01 -4.84506905e-01
2.01049849e-01 5.68298876e-01 -4.97596830e-01 -1.26398280e-01
-2.32376039e-01 -1.33057427e+00 1.03532159e+00 7.56450295e-01
1.22343540e+00 -1.68810773e+00 -7.19720483e-01 4.18943316e-01
-2.45229587e-01 -3.35110277e-01 1.07392609e-01 2.95585722e-01
-8.15714121e-01 -1.06042171e+00 -7.91709602e-01 -3.92755896e-01
1.73838094e-01 5.85154474e-01 9.81121004e-01 -6.10740222e-02
-4.97689635e-01 -1.87551185e-01 -4.58927035e-01 -4.02275652e-01
-2.43605748e-01 1.46811709e-01 1.41437620e-01 -3.55480194e-01
5.92068076e-01 -1.03499258e+00 -6.89241230e-01 3.71117324e-01
-1.15919983e+00 -1.33810133e-01 7.66138434e-01 7.66098559e-01
4.87189800e-01 6.49117529e-01 7.38167465e-01 -3.16593975e-01
2.28905320e-01 -8.43972087e-01 -4.69434917e-01 7.95084462e-02
-7.33958125e-01 -4.76208240e-01 3.83958876e-01 2.78885663e-02
-1.21811831e+00 2.03257486e-01 -1.38390303e-01 5.74919134e-02
-1.07140315e+00 1.12716794e+00 -8.19385424e-02 3.62063378e-01
1.15816200e+00 2.13974062e-02 -2.79718161e-01 -4.23154294e-01
3.24213713e-01 1.05947757e+00 7.05757320e-01 -1.28803000e-01
9.03068781e-01 3.76800209e-01 -2.91885644e-01 -1.15886700e+00
-5.91135502e-01 -3.78648967e-01 -6.93784297e-01 -4.02897120e-01
9.02112961e-01 -1.14612055e+00 -2.30841503e-01 7.39637077e-01
-9.44493532e-01 -4.44142938e-01 -1.78413257e-01 5.74299455e-01
-2.30990827e-01 -1.23823904e-01 -2.29018778e-01 -7.93516874e-01
-6.38642311e-01 -5.82209289e-01 5.98718584e-01 1.18094750e-01
1.30756855e-01 -5.55507302e-01 3.42183024e-01 -6.75906613e-02
9.22070682e-01 7.01621890e-01 7.51353085e-01 4.56906930e-02
-3.58277440e-01 2.05038771e-01 -6.62102401e-01 5.46670020e-01
7.97505796e-01 8.43723044e-02 -9.95771408e-01 -8.39990973e-02
-3.04942995e-01 -2.18529969e-01 1.58648419e+00 5.95897973e-01
9.44519818e-01 -4.60631549e-01 -1.49694026e-01 8.06921780e-01
1.75739181e+00 3.03037375e-01 8.56914282e-01 5.08605182e-01
5.32805443e-01 4.95612085e-01 4.08118397e-01 5.98101974e-01
7.49599934e-02 3.49986739e-02 8.89842391e-01 -4.12609726e-01
1.15738809e-01 -1.99805126e-01 3.14490676e-01 8.17989632e-02
-7.01405823e-01 2.90720221e-02 -1.52412474e+00 6.21127963e-01
-1.73824894e+00 -1.65399313e+00 -6.03103377e-02 1.94585752e+00
7.88024604e-01 -5.51251709e-01 9.69109759e-02 2.91938335e-01
9.41587210e-01 7.76048660e-01 -5.32774687e-01 9.03682783e-03
-4.13295209e-01 3.87692750e-01 9.41929519e-01 3.36381048e-01
-1.85456479e+00 1.30913532e+00 6.20643091e+00 4.32051927e-01
-1.30060613e+00 2.93992739e-02 5.19871294e-01 3.49022180e-01
-8.46159160e-02 -6.34313151e-02 -4.34756607e-01 9.16909203e-02
9.24153924e-01 -2.82687485e-01 5.78094244e-01 7.32197285e-01
4.80780452e-01 -1.66465759e-01 -9.74666178e-02 3.59992206e-01
-4.92814571e-01 -1.26714432e+00 2.59604186e-01 8.40392858e-02
5.51808059e-01 6.09521329e-01 -7.47218803e-02 3.15775454e-01
1.07459450e+00 -1.30834007e+00 7.10004345e-02 3.41163576e-01
1.06239200e+00 -8.05569530e-01 5.87496281e-01 3.04403514e-01
-1.42740726e+00 -3.01617920e-01 -7.29978621e-01 -5.33663571e-01
-1.13295782e-02 8.46017480e-01 -3.80618274e-01 3.60953599e-01
1.33749104e+00 1.13254774e+00 -3.96905363e-01 6.28787339e-01
-4.42741334e-01 9.69837904e-01 -4.75311488e-01 4.84192103e-01
2.77997583e-01 -3.45714748e-01 3.54437768e-01 9.05477762e-01
3.08727652e-01 4.88463581e-01 3.87212485e-01 1.12840211e+00
1.51304573e-01 -2.47035772e-01 -1.13425100e+00 -1.06245235e-01
3.90250951e-01 1.16822505e+00 -4.53185260e-01 9.67253745e-02
9.47832763e-02 5.92909396e-01 -8.55470300e-02 5.27398467e-01
-6.05067492e-01 -4.98966515e-01 1.08225071e+00 -7.34011456e-02
-7.61930719e-02 -4.15072501e-01 -7.00953081e-02 -1.01776779e+00
-3.20559353e-01 -6.56089962e-01 5.74590385e-01 -8.69954944e-01
-1.45319140e+00 5.52813768e-01 1.80019572e-01 -1.17587292e+00
-5.63115403e-02 -4.38362032e-01 -7.48401582e-01 1.15923977e+00
-1.99945378e+00 -1.31768024e+00 -6.57624900e-01 4.30200905e-01
1.58888772e-01 -2.01565370e-01 1.46941710e+00 1.68767795e-01
-3.66127312e-01 -1.23929903e-01 2.27139965e-01 3.72292936e-01
4.76912379e-01 -9.59004581e-01 6.32218897e-01 8.76178801e-01
-2.60948747e-01 2.18948662e-01 -4.22182269e-02 -7.69880295e-01
-6.83375895e-01 -1.99890900e+00 9.89099026e-01 3.40527624e-01
7.15071082e-01 2.15530783e-01 -9.08131719e-01 7.15815723e-01
-1.30355924e-01 -9.28926766e-02 5.50815821e-01 1.71733387e-02
-4.77799714e-01 -3.97599936e-01 -1.44962192e+00 4.99156266e-01
1.11594748e+00 -4.79090244e-01 -3.75158221e-01 4.30254132e-01
4.76769269e-01 4.69677687e-01 -9.48334634e-01 6.43110514e-01
5.69313467e-01 -8.16006780e-01 9.70389605e-01 -5.52644491e-01
9.15026903e-01 -4.27404821e-01 -6.58567786e-01 -1.61193156e+00
-1.16292071e+00 1.43015936e-01 6.61873281e-01 1.16043270e+00
2.53624707e-01 -6.62110746e-01 3.33258748e-01 3.47193033e-02
1.11988567e-01 1.06735267e-02 -7.34808445e-01 -9.26302850e-01
5.38340211e-01 -3.91855299e-01 9.47372079e-01 1.19989479e+00
-3.31729293e-01 -9.84863639e-02 -2.51247376e-01 6.83836818e-01
5.24868667e-01 -1.29382201e-02 3.31138790e-01 -1.48695219e+00
3.27520341e-01 -7.34488189e-01 -4.67614561e-01 -2.09928662e-01
1.24879844e-01 -8.78510952e-01 7.74992108e-02 -1.43601108e+00
1.23181324e-02 -5.71635306e-01 -4.85783815e-01 1.49312496e+00
-1.92370936e-01 3.18882525e-01 -1.01192407e-01 3.76620501e-01
6.37588680e-01 7.85380304e-01 6.46739185e-01 -7.34812438e-01
-2.32411012e-01 -1.75636914e-02 -5.37588954e-01 7.46128142e-01
1.23314643e+00 -4.07111377e-01 2.76517086e-02 -6.98261380e-01
1.00324145e-02 -1.41881123e-01 7.84478247e-01 -1.31298506e+00
-3.08893532e-01 -9.30604100e-01 4.80982870e-01 -7.90962636e-01
-3.59531939e-01 -9.88926589e-01 5.01779437e-01 6.89217210e-01
-3.83454084e-01 -8.15960467e-02 3.89659911e-01 3.09802711e-01
-2.09020153e-01 1.63773969e-01 7.28043139e-01 -2.79736787e-01
-1.11225295e+00 7.46249557e-01 -8.52814019e-01 -4.93528634e-01
8.19516242e-01 1.85599521e-01 -6.37001276e-01 -8.55455026e-02
-6.72272742e-01 2.56580502e-01 1.93052381e-01 5.56475222e-01
6.85510516e-01 -1.51787412e+00 -1.32030046e+00 6.44922137e-01
3.00638497e-01 -8.81750509e-03 2.88384497e-01 9.12289098e-02
-7.37957001e-01 9.29813012e-02 -8.93595874e-01 -2.83693016e-01
-6.69150233e-01 4.57563907e-01 3.40069622e-01 -5.40753566e-02
-7.49849737e-01 5.58480144e-01 -9.48900580e-02 -8.34651530e-01
-1.39579847e-01 -3.08939248e-01 -4.84227747e-01 1.74634829e-01
8.80117834e-01 3.37344795e-01 -7.24001154e-02 -7.74615109e-01
-3.97314429e-01 1.80052415e-01 2.74005383e-01 -1.36813611e-01
1.83446527e+00 1.32361665e-01 -2.19304293e-01 2.56748468e-01
6.11071408e-01 -7.76399672e-01 -1.04208708e+00 -1.63332626e-01
3.91161293e-02 -4.64996010e-01 7.45218635e-01 -1.21281517e+00
-1.51573384e+00 8.98602545e-01 1.12513506e+00 1.78458095e-01
1.49036145e+00 -3.86638612e-01 5.80821335e-01 5.89482844e-01
4.26338911e-01 -7.61204720e-01 -7.66323864e-01 7.62655020e-01
1.33474696e+00 -1.56840134e+00 -6.56512901e-02 -7.22273141e-02
-2.56642252e-01 8.72477591e-01 3.96979541e-01 -1.64079934e-01
1.15608680e+00 2.33090743e-01 3.16079468e-01 2.07851212e-02
-3.82269442e-01 -7.23536015e-01 -4.70393836e-01 1.20037258e+00
3.36869925e-01 8.04523170e-01 -6.77076727e-02 5.82928181e-01
2.65127998e-02 4.32195544e-01 2.70255327e-01 9.55989003e-01
-8.13742757e-01 -6.80430591e-01 -6.68999910e-01 7.28614867e-01
7.06989914e-02 -4.77671653e-01 -3.03321391e-01 4.63596106e-01
1.08954892e-01 1.07970536e+00 1.24551266e-01 -6.17531657e-01
1.97570696e-01 -2.25567237e-01 -3.05603147e-01 -5.55548728e-01
-6.71478033e-01 -2.58259296e-01 -2.03102559e-01 -4.42255467e-01
-6.47248209e-01 -4.13171738e-01 -1.00815809e+00 -8.72419000e-01
-9.79326107e-03 -1.77658156e-01 6.40214443e-01 6.26334667e-01
3.45360905e-01 1.88128278e-01 9.60503399e-01 -1.29893219e+00
-5.50448298e-01 -1.34915221e+00 -1.18444431e+00 7.91097730e-02
1.78284377e-01 -7.30821848e-01 -4.85878289e-01 -3.78082693e-02] | [9.421952247619629, -1.4454344511032104] |
6a5ae693-5498-46e6-ac87-6a62384861b9 | anifacegan-animatable-3d-aware-face-image | 2210.06465 | null | https://arxiv.org/abs/2210.06465v1 | https://arxiv.org/pdf/2210.06465v1.pdf | AniFaceGAN: Animatable 3D-Aware Face Image Generation for Video Avatars | Although 2D generative models have made great progress in face image generation and animation, they often suffer from undesirable artifacts such as 3D inconsistency when rendering images from different camera viewpoints. This prevents them from synthesizing video animations indistinguishable from real ones. Recently, 3D-aware GANs extend 2D GANs for explicit disentanglement of camera pose by leveraging 3D scene representations. These methods can well preserve the 3D consistency of the generated images across different views, yet they cannot achieve fine-grained control over other attributes, among which facial expression control is arguably the most useful and desirable for face animation. In this paper, we propose an animatable 3D-aware GAN for multiview consistent face animation generation. The key idea is to decompose the 3D representation of the 3D-aware GAN into a template field and a deformation field, where the former represents different identities with a canonical expression, and the latter characterizes expression variations of each identity. To achieve meaningful control over facial expressions via deformation, we propose a 3D-level imitative learning scheme between the generator and a parametric 3D face model during adversarial training of the 3D-aware GAN. This helps our method achieve high-quality animatable face image generation with strong visual 3D consistency, even though trained with only unstructured 2D images. Extensive experiments demonstrate our superior performance over prior works. Project page: https://yuewuhkust.github.io/AniFaceGAN | ['Xin Tong', 'Qifeng Chen', 'Fangyun Wei', 'Jiaolong Yang', 'Yu Deng', 'Yue Wu'] | 2022-10-12 | null | null | null | null | ['face-model'] | ['computer-vision'] | [ 4.22178954e-02 2.76446998e-01 -1.21162966e-01 -2.84840852e-01
-6.06844246e-01 -8.61805320e-01 6.78889513e-01 -1.10265577e+00
4.06322062e-01 5.72780371e-01 2.22183838e-01 -7.52272876e-03
4.43101436e-01 -7.48300612e-01 -8.45295727e-01 -9.18436170e-01
3.32045972e-01 3.86561185e-01 -4.32684869e-01 -2.20825627e-01
-3.88858199e-01 7.89351046e-01 -1.38925326e+00 -9.30636935e-03
6.30054057e-01 9.10723388e-01 -1.87244087e-01 6.49534702e-01
6.31189719e-02 5.97165525e-01 -7.05363333e-01 -5.30143201e-01
6.33960068e-01 -9.04333532e-01 -2.78459072e-01 4.95545089e-01
7.47367680e-01 -6.46876097e-01 -3.94032508e-01 9.78428602e-01
4.77197021e-01 -2.18784064e-01 6.91340566e-01 -1.83306181e+00
-9.82198179e-01 -1.09193228e-01 -8.49996150e-01 -4.66322452e-01
5.53238213e-01 5.11010885e-01 5.80006421e-01 -6.80101454e-01
8.50286603e-01 1.55369961e+00 3.67946804e-01 1.18288910e+00
-1.40076649e+00 -1.02229631e+00 1.04224704e-01 -2.94449717e-01
-1.25425696e+00 -5.67120969e-01 1.17615426e+00 -4.14723665e-01
2.71191001e-01 4.14670378e-01 9.40032065e-01 1.49827933e+00
4.06245515e-02 5.87996960e-01 1.27086067e+00 -9.45407525e-02
1.33674573e-02 -1.60778955e-01 -7.67981946e-01 9.69232321e-01
-1.67200662e-04 1.86026752e-01 -4.30852801e-01 -1.97068170e-01
1.45315182e+00 -6.46169484e-02 -5.16841054e-01 -5.57923973e-01
-1.00225425e+00 8.29206526e-01 2.51066506e-01 -1.22349903e-01
-3.05489153e-01 2.61896163e-01 -4.13770527e-02 2.90216327e-01
4.96265471e-01 3.16403300e-01 -2.12628737e-01 1.01510525e-01
-6.02559388e-01 5.23732901e-01 4.73462999e-01 1.14607930e+00
7.92866230e-01 6.40023291e-01 -2.06395119e-01 5.83387256e-01
3.67901325e-01 9.42904353e-01 1.76237673e-01 -1.44590819e+00
1.33597344e-01 4.18231666e-01 -7.13910675e-03 -1.15885782e+00
2.90290594e-01 -2.06918828e-02 -1.06153631e+00 6.48359954e-01
3.61374766e-01 -1.89770073e-01 -9.66944218e-01 2.30435610e+00
6.22477233e-01 2.50772357e-01 -1.36415944e-01 8.72553468e-01
7.41727114e-01 6.86374664e-01 -1.27732083e-01 -3.10630709e-01
1.18966734e+00 -7.21064627e-01 -8.01643372e-01 -3.99903618e-02
5.84553368e-02 -7.33600199e-01 9.98545051e-01 3.89699335e-03
-1.55498528e+00 -5.59347212e-01 -8.72944057e-01 -1.13536775e-01
2.20350206e-01 -2.17266500e-01 4.71093357e-01 6.36718214e-01
-1.24990547e+00 1.46416470e-01 -7.99180150e-01 7.20089599e-02
6.07855022e-01 3.26571196e-01 -7.28330135e-01 5.61966710e-02
-1.06895304e+00 6.64615929e-01 -2.96017826e-01 -1.14364382e-02
-1.12436831e+00 -7.44909227e-01 -1.01317155e+00 -1.49138525e-01
2.25832701e-01 -1.05245233e+00 1.05773687e+00 -1.31393445e+00
-1.93148839e+00 1.25069475e+00 -2.25472182e-01 9.45335254e-02
7.20484436e-01 6.56849444e-02 -2.22369209e-01 1.13124959e-01
4.11615446e-02 8.81824255e-01 1.39142251e+00 -1.63515937e+00
-1.17759104e-04 -4.55888897e-01 6.59404248e-02 2.46945769e-01
-1.36826649e-01 -1.31818518e-01 -7.32783675e-01 -8.80318940e-01
-1.47103727e-01 -1.14063144e+00 9.43694555e-04 5.63374162e-01
-5.27424097e-01 2.01811999e-01 1.14227784e+00 -5.80762088e-01
5.78971267e-01 -2.00976300e+00 4.86791372e-01 -2.10167114e-02
3.49388868e-01 1.96733311e-01 -3.33336622e-01 1.39587045e-01
-1.95411682e-01 2.12054193e-01 -1.64206922e-01 -5.39200604e-01
1.44007448e-02 3.16933990e-01 -4.48321462e-01 4.78383899e-01
4.53362048e-01 1.08767855e+00 -7.47923732e-01 -4.54203248e-01
1.25123218e-01 1.01445436e+00 -7.02460825e-01 6.75879538e-01
-3.06821048e-01 1.11989677e+00 -6.72450125e-01 6.17181897e-01
8.85509729e-01 -1.42403126e-01 1.95363432e-01 -3.04202229e-01
3.31474811e-01 -2.01228440e-01 -7.93141186e-01 1.69730389e+00
-4.08028066e-01 4.52255815e-01 3.98711920e-01 -5.86054564e-01
9.83425438e-01 4.30158257e-01 4.30996031e-01 -5.03150582e-01
2.18080282e-01 -3.36271673e-02 -2.10841641e-01 -2.18682349e-01
5.88354096e-02 -3.62967670e-01 -1.19351327e-01 4.66161340e-01
-2.23117732e-02 -7.96804428e-01 -3.48076850e-01 1.75922439e-01
6.52811468e-01 3.83178860e-01 1.17777810e-01 -4.05315422e-02
5.01243889e-01 -6.99810982e-01 7.28970468e-01 1.54110521e-01
-6.16669431e-02 9.36919272e-01 6.86509550e-01 -3.85754675e-01
-1.30083287e+00 -1.18797004e+00 2.11241767e-01 3.81682903e-01
1.48326665e-01 -2.58413345e-01 -9.17490363e-01 -7.65857577e-01
-3.98755409e-02 4.10752296e-01 -8.64690065e-01 -2.45979935e-01
-7.45263815e-01 -3.30512911e-01 5.63236475e-01 3.32914889e-01
6.26759529e-01 -7.23972559e-01 -3.38081986e-01 -2.84597427e-01
-1.29076689e-01 -1.11805499e+00 -8.57100070e-01 -5.55252552e-01
-5.77185392e-01 -9.66249168e-01 -8.88010740e-01 -6.07539892e-01
9.88477230e-01 1.58994436e-01 1.19743347e+00 1.41645119e-01
-4.65204082e-02 4.33034450e-01 1.48050915e-02 -2.66374201e-01
-7.91041851e-01 -5.04453242e-01 2.30744570e-01 2.44493842e-01
-2.49817923e-01 -9.71962988e-01 -6.77680612e-01 3.73344094e-01
-8.57825339e-01 4.02567357e-01 1.73941612e-01 9.38894928e-01
6.53221667e-01 -3.15870494e-01 3.79722834e-01 -8.84261012e-01
2.86718071e-01 -1.28726751e-01 -6.06095910e-01 8.18450898e-02
-2.30996847e-01 -3.87180075e-02 8.31089139e-01 -5.19344747e-01
-1.04010594e+00 3.35439760e-03 -2.83760041e-01 -1.13201690e+00
-1.06586546e-01 -4.24110234e-01 -8.26261401e-01 -1.75058559e-01
3.20304900e-01 2.36987621e-01 4.58713502e-01 -1.71394780e-01
5.48660338e-01 1.52518705e-01 5.32086492e-01 -7.98657477e-01
1.38395953e+00 6.02019787e-01 1.74710259e-01 -5.67622423e-01
-5.70615888e-01 5.14998734e-01 -4.18836623e-01 -3.22316349e-01
1.00393116e+00 -1.08290327e+00 -7.87963688e-01 7.71085441e-01
-1.12293875e+00 -4.83308107e-01 -4.60620642e-01 -3.45123257e-03
-7.95788586e-01 2.42088139e-01 -5.52850306e-01 -5.29273093e-01
-2.00367719e-01 -1.43542874e+00 1.40104270e+00 2.70262450e-01
-2.13984922e-01 -8.67795944e-01 2.54399166e-03 3.54795367e-01
2.81539083e-01 9.25591469e-01 9.26574290e-01 2.09929690e-01
-6.49912775e-01 2.63435189e-02 9.35528427e-02 4.22868580e-01
5.15041769e-01 2.55643278e-01 -9.64384675e-01 -4.00753826e-01
1.19686089e-01 -4.65988308e-01 3.10809284e-01 2.27600142e-01
1.22221088e+00 -5.89045703e-01 -1.49323836e-01 1.11834884e+00
1.07844603e+00 1.64009169e-01 6.56577826e-01 -2.55755365e-01
1.10847640e+00 4.65437591e-01 2.31202126e-01 2.76073039e-01
2.42712945e-01 1.03961730e+00 6.58398330e-01 -2.62077242e-01
-5.27711391e-01 -7.16420114e-01 4.95924085e-01 6.40972614e-01
-2.51305848e-01 -3.15393627e-01 -3.30173761e-01 7.09632635e-02
-1.35964334e+00 -1.02915800e+00 2.95495719e-01 2.04280877e+00
9.22681630e-01 -3.17818314e-01 1.27055779e-01 -1.59273386e-01
6.98141158e-01 4.52491790e-01 -8.17705393e-01 -1.94642633e-01
-2.37119973e-01 1.68475986e-01 -3.77900223e-03 5.29420972e-01
-7.55148351e-01 8.77898276e-01 5.61006117e+00 7.77684808e-01
-1.25397635e+00 9.09117311e-02 9.71279740e-01 -2.79746979e-01
-7.75194705e-01 -9.91051644e-02 -5.31002104e-01 6.40702128e-01
3.50573540e-01 -1.27626270e-01 2.71172374e-01 7.71639049e-01
2.27675587e-01 4.46724504e-01 -1.00308275e+00 1.09910679e+00
1.95459962e-01 -1.54185963e+00 4.51323956e-01 3.18490922e-01
1.02960408e+00 -6.78048611e-01 4.94741321e-01 -9.35968235e-02
3.84733051e-01 -1.26884198e+00 8.49525690e-01 3.67519617e-01
1.44103944e+00 -8.02894235e-01 -6.48762845e-03 1.27418026e-01
-9.51700628e-01 3.36178482e-01 -1.22091249e-02 2.69840509e-01
3.53388309e-01 2.07084268e-01 -3.13280970e-01 4.14891005e-01
4.86967236e-01 6.87250078e-01 -1.77628264e-01 6.35759830e-02
-5.39520264e-01 2.14388654e-01 -6.63583949e-02 4.42157149e-01
-4.10800576e-02 -4.94934976e-01 7.71716893e-01 5.41129529e-01
5.39397478e-01 3.40923190e-01 -7.06247538e-02 1.14587319e+00
-4.95195985e-01 -2.85511166e-01 -1.05351973e+00 1.46522477e-01
4.83061880e-01 1.21734679e+00 -2.68762916e-01 -5.99986613e-02
-2.36508280e-01 1.23063076e+00 2.08744854e-01 3.94647419e-01
-1.06834853e+00 2.26920724e-01 1.32857335e+00 2.60991693e-01
1.62010297e-01 -2.51916677e-01 4.59497720e-02 -1.39182901e+00
3.64351505e-03 -1.26245642e+00 -1.23492956e-01 -9.55103219e-01
-1.25540590e+00 8.03889632e-01 -7.03001395e-02 -1.33195329e+00
-5.89639962e-01 -4.39997703e-01 -6.80119157e-01 9.83418584e-01
-1.14414597e+00 -1.58022380e+00 -4.23086226e-01 8.47309589e-01
3.44696015e-01 -2.13285774e-01 9.40690577e-01 4.71699089e-02
-4.88635421e-01 9.37329590e-01 -2.63610750e-01 1.68129697e-01
7.84980953e-01 -9.45973635e-01 5.74569881e-01 7.55490780e-01
1.28103212e-01 3.37123245e-01 5.77167332e-01 -4.50701416e-01
-1.83091366e+00 -1.12770426e+00 2.38597885e-01 -7.18004465e-01
1.76986575e-01 -4.86424357e-01 -7.82817900e-01 7.29477406e-01
3.35649163e-01 2.61272728e-01 4.75123495e-01 -5.64370394e-01
-5.96788049e-01 -1.12237014e-01 -1.29417169e+00 9.14011121e-01
1.36358786e+00 -5.38229465e-01 -3.45861865e-03 1.47421241e-01
7.58644402e-01 -6.01635098e-01 -8.87979627e-01 2.96465099e-01
6.07157469e-01 -1.15452433e+00 1.05581439e+00 -5.34251094e-01
6.51230693e-01 -4.64657128e-01 -1.00447796e-01 -1.29747236e+00
-1.82895865e-02 -1.18350327e+00 -1.73754290e-01 1.35067749e+00
-5.54426908e-02 -5.30871034e-01 9.08599973e-01 6.14823163e-01
3.93637158e-02 -6.79785788e-01 -7.37360775e-01 -7.53764451e-01
3.14381987e-01 -6.39856830e-02 9.73723054e-01 1.07240021e+00
-5.82483113e-01 3.14427853e-01 -7.78719783e-01 1.20859884e-01
5.94674528e-01 3.27678889e-01 1.26926923e+00 -7.39617348e-01
-4.49406028e-01 -4.43944037e-01 -5.79791844e-01 -1.08823228e+00
6.66405916e-01 -7.05581486e-01 -2.14553133e-01 -9.34565902e-01
1.44014180e-01 -3.90220553e-01 5.02983153e-01 4.04221386e-01
-1.00409128e-01 4.90014106e-01 3.54311287e-01 1.70043901e-01
7.50954151e-02 9.06883836e-01 2.00225544e+00 8.79498944e-03
9.05155391e-02 -1.99328423e-01 -7.96659350e-01 6.92042410e-01
5.76674163e-01 -1.67274341e-01 -7.62478113e-01 -6.77698314e-01
-1.27320617e-01 3.62071514e-01 6.77550137e-01 -5.68428338e-01
-2.50384688e-01 -3.52221817e-01 6.62649930e-01 -7.57289827e-02
7.74561286e-01 -7.44048893e-01 7.78526723e-01 2.35991672e-01
-2.12582737e-01 1.30010441e-01 7.11575747e-02 4.62076068e-01
-2.12199092e-01 4.73241985e-01 1.10728061e+00 -2.18274280e-01
-2.20755354e-01 9.53695595e-01 1.06622919e-01 2.49787807e-01
1.09787107e+00 -1.56175599e-01 -2.36220360e-01 -8.72000635e-01
-4.68764305e-01 -1.63503349e-01 1.20964921e+00 4.35992390e-01
6.57022834e-01 -1.73450553e+00 -8.18902254e-01 6.75158918e-01
-1.67508617e-01 3.27670217e-01 3.79566312e-01 4.04959649e-01
-5.12016416e-01 -1.14560343e-01 -4.95419025e-01 -6.04922771e-01
-1.33866739e+00 4.64787751e-01 4.99707818e-01 -8.67025405e-02
-5.73225975e-01 9.17564690e-01 8.04989934e-01 -3.94760549e-01
-1.54904708e-01 2.25169212e-01 1.59578666e-01 -3.44647557e-01
4.50933933e-01 -1.55966356e-01 -3.29102516e-01 -1.01334620e+00
-1.70890182e-01 1.01080847e+00 1.57113239e-01 -1.93593040e-01
1.14582407e+00 -4.55445983e-02 -4.52142544e-02 4.89794128e-02
1.39314115e+00 3.26131791e-01 -1.90353036e+00 1.98681772e-01
-9.47186232e-01 -8.77333820e-01 -2.57725239e-01 -4.03015733e-01
-1.73404872e+00 8.13623846e-01 2.50886261e-01 -3.61108892e-02
1.34427190e+00 5.93835302e-02 8.88276577e-01 -2.95777738e-01
4.04995650e-01 -3.12457174e-01 3.24344039e-01 1.08948462e-01
1.12378526e+00 -1.08683062e+00 -1.61573023e-01 -5.43221474e-01
-7.88952053e-01 8.75497341e-01 8.23049009e-01 -2.38249227e-01
4.66213405e-01 3.97374511e-01 1.61356539e-01 -5.52595668e-02
-6.80284977e-01 3.26427311e-01 3.13805193e-01 9.67466831e-01
2.55586028e-01 6.42081052e-02 2.33412087e-01 2.83618271e-01
-2.94259369e-01 -3.55208606e-01 4.16699290e-01 5.11266708e-01
3.74244362e-01 -1.33306825e+00 -3.02838981e-01 -1.07873142e-01
-3.31934512e-01 2.00042218e-01 -4.48083013e-01 1.05879915e+00
9.17525142e-02 5.84292769e-01 1.44342199e-01 -3.82819444e-01
2.80517727e-01 -1.30891800e-01 8.74227464e-01 -3.81555349e-01
-2.21809104e-01 2.10656315e-01 -2.14604199e-01 -7.03305900e-01
-3.68437618e-01 -7.36762345e-01 -9.35556948e-01 -6.78186119e-01
1.72826946e-01 5.42139485e-02 3.16603869e-01 4.90023375e-01
5.96250236e-01 2.47694969e-01 1.18338692e+00 -1.09496212e+00
-3.58733177e-01 -3.60225469e-01 -5.17908156e-01 6.41888320e-01
4.34923530e-01 -6.85412645e-01 -4.23227429e-01 4.07049656e-01] | [12.612902641296387, -0.29772523045539856] |
bbedec63-2f48-44c6-862b-0c6ecbe219f6 | improving-multi-task-generalization-ability | 2204.02725 | null | https://arxiv.org/abs/2204.02725v2 | https://arxiv.org/pdf/2204.02725v2.pdf | Match-Prompt: Improving Multi-task Generalization Ability for Neural Text Matching via Prompt Learning | Text matching is a fundamental technique in both information retrieval and natural language processing. Text matching tasks share the same paradigm that determines the relationship between two given texts. The relationships vary from task to task, e.g.~relevance in document retrieval, semantic alignment in paraphrase identification and answerable judgment in question answering. However, the essential signals for text matching remain in a finite scope, i.e.~exact matching, semantic matching, and inference matching. Ideally, a good text matching model can learn to capture and aggregate these signals for different matching tasks to achieve competitive performance, while recent state-of-the-art text matching models, e.g.~Pre-trained Language Models (PLMs), are hard to generalize. It is because the end-to-end supervised learning on task-specific dataset makes model overemphasize the data sample bias and task-specific signals instead of the essential matching signals. To overcome this problem, we adopt a specialization-generalization training strategy and refer to it as Match-Prompt. In specialization stage, descriptions of different matching tasks are mapped to a few prompt tokens. In generalization stage, matching model explores the essential matching signals by being trained on diverse matching tasks. High diverse matching tasks avoid model fitting the data bias on a specific task, so that model can focus on learning the essential matching signals. Meanwhile, the prompt tokens obtained in the first step help the model distinguish different task-specific matching signals. Experimental results on public datasets show that Match-Prompt can improve multi-task generalization capability of PLMs in text matching and yield better in-domain multi-task, out-of-domain multi-task and new task adaptation performance than multi-task and task-specific models trained by previous fine-tuning paradigm. | ['Xueqi Cheng', 'HuaWei Shen', 'Liang Pang', 'Shicheng Xu'] | 2022-04-06 | null | null | null | null | ['paraphrase-identification'] | ['natural-language-processing'] | [ 5.35444677e-01 -6.01331949e-01 -4.36898917e-01 -7.45028079e-01
-9.56340730e-01 -5.81529319e-01 7.19112039e-01 1.65537655e-01
-5.34100711e-01 1.46822125e-01 1.71113551e-01 -1.48926437e-01
-3.78467917e-01 -5.17262638e-01 -5.71020961e-01 -1.12932086e-01
6.46289170e-01 6.52904034e-01 3.99912983e-01 -5.07870734e-01
3.84712160e-01 -9.09668207e-02 -1.28165424e+00 1.04409242e+00
1.06036973e+00 1.09430265e+00 7.55194128e-01 4.37398285e-01
-7.23380029e-01 2.44558051e-01 -6.61933541e-01 -4.04235125e-01
1.13114871e-01 -4.38948184e-01 -9.79121804e-01 -5.42622209e-01
6.44773245e-01 1.82645261e-01 -5.23883939e-01 1.12589169e+00
7.09231853e-01 1.63471356e-01 8.61134112e-01 -1.27953029e+00
-9.02996004e-01 8.03527892e-01 -5.06587207e-01 5.84136367e-01
8.01393151e-01 -4.09872234e-02 1.32986975e+00 -9.70647097e-01
3.15690070e-01 1.41802073e+00 6.05893672e-01 6.95359230e-01
-1.26563072e+00 -1.28247857e+00 1.73111200e-01 3.07075948e-01
-1.20863318e+00 -4.44125086e-01 8.21901858e-01 -3.77015024e-01
1.01224780e+00 4.34541762e-01 -1.57383025e-01 1.39162946e+00
4.25467402e-01 1.22784650e+00 1.21914291e+00 -4.38116133e-01
-3.79677355e-01 2.18408912e-01 5.33416867e-01 2.36836568e-01
-1.25495344e-01 -5.13669066e-02 -8.00342679e-01 -2.57690012e-01
2.22360745e-01 2.59423941e-01 -4.64430541e-01 1.94594994e-01
-1.46094120e+00 5.00592411e-01 4.33767676e-01 6.28060997e-01
4.72779907e-02 -1.88678369e-01 6.59659564e-01 1.00562358e+00
1.93340510e-01 4.82901067e-01 -7.43901730e-01 2.67460555e-01
-8.31023037e-01 4.73603785e-01 5.34678400e-01 1.25476217e+00
1.09256971e+00 -2.94205844e-01 -8.89961541e-01 1.29043329e+00
1.90453216e-01 8.85622859e-01 1.21194673e+00 -1.97128698e-01
1.22528195e+00 8.79609883e-01 -4.45212662e-01 -8.63915980e-01
-4.23203200e-01 -3.72857511e-01 -1.01482081e+00 -6.35285199e-01
4.04316306e-01 1.19927280e-01 -7.14195669e-01 1.91840684e+00
-1.75131068e-01 -4.13938127e-02 6.01767413e-02 6.17260754e-01
1.02051413e+00 4.94489998e-01 2.82794058e-01 5.07239215e-02
1.59790707e+00 -8.84734690e-01 -6.91474497e-01 -8.74419510e-01
7.23520279e-01 -9.95565355e-01 1.65389085e+00 -5.95102236e-02
-6.93631828e-01 -1.02290940e+00 -7.54151762e-01 -1.37684181e-01
-4.91056353e-01 5.86111508e-02 2.86814034e-01 3.58638674e-01
-3.43636811e-01 5.09580553e-01 1.00200623e-01 -4.48464274e-01
3.86209739e-03 2.62664020e-01 -1.81830749e-01 -3.07681262e-01
-1.80705261e+00 8.35802734e-01 7.67007828e-01 -2.92279303e-01
-5.29259384e-01 -8.29378545e-01 -5.33275425e-01 1.88870057e-01
2.28579804e-01 -1.06095707e+00 1.46965623e+00 -8.85924041e-01
-1.11712956e+00 1.25424707e+00 -2.90673345e-01 -1.48037419e-01
4.13301259e-01 -1.37723967e-01 -8.36514771e-01 -1.97546273e-01
4.06908631e-01 4.85169828e-01 1.11333477e+00 -7.83816874e-01
-9.24051762e-01 -2.56981820e-01 -2.90590018e-01 2.77913302e-01
-5.35544455e-01 3.12687933e-01 -3.73077720e-01 -7.01355159e-01
1.69320732e-01 -6.68534577e-01 2.22979993e-01 -3.10683906e-01
-2.96640903e-01 -7.28970945e-01 6.32541299e-01 -4.45367068e-01
1.55266428e+00 -2.15697789e+00 -2.34342963e-02 1.03664771e-01
8.01715925e-02 1.93581924e-01 -5.10547400e-01 5.18174231e-01
-2.35376805e-01 1.81997102e-02 -1.91792753e-02 -1.33960158e-01
1.59540072e-01 -1.29133444e-02 -4.55608189e-01 3.32185626e-02
5.21377623e-02 1.33337629e+00 -9.70740378e-01 -6.11959279e-01
-1.40064299e-01 -5.08870184e-01 -2.30792329e-01 4.51089233e-01
-2.63469994e-01 2.07545325e-01 -8.71252358e-01 5.96789420e-01
4.45402026e-01 -2.89565176e-01 -1.74524486e-01 -3.54665488e-01
4.85231280e-01 4.81211931e-01 -1.14254618e+00 2.11222172e+00
-7.19745278e-01 5.08977413e-01 -5.09232320e-02 -1.28960288e+00
1.10490179e+00 2.68359482e-01 1.07111938e-01 -1.16699028e+00
5.79255782e-02 5.00744402e-01 1.17100984e-01 -8.41696084e-01
4.11772251e-01 -2.34816223e-01 -3.77382159e-01 5.12247086e-01
5.49632274e-02 -1.24483600e-01 -2.43480946e-03 4.81901132e-02
9.65309024e-01 -2.28314802e-01 2.57647067e-01 -4.10955071e-01
7.76691079e-01 -8.14409740e-03 2.97888488e-01 1.11659718e+00
-1.35850906e-01 4.93177235e-01 -1.33576632e-01 -1.24755688e-01
-7.52485454e-01 -8.38068724e-01 -3.68210137e-01 1.90363705e+00
2.60292590e-01 -1.91469580e-01 -3.46021980e-01 -5.73196471e-01
3.31106007e-01 5.80442905e-01 -4.91079181e-01 -5.65004587e-01
-7.35326886e-01 -5.86875737e-01 7.40946889e-01 4.37036186e-01
5.92829585e-01 -1.33019888e+00 -4.65265177e-02 4.15340722e-01
-5.57837188e-01 -1.07577598e+00 -9.79171515e-01 2.65260607e-01
-8.85221541e-01 -1.01249135e+00 -7.16152847e-01 -1.00060630e+00
3.66830170e-01 7.00587511e-01 1.42130578e+00 7.26910606e-02
2.90154535e-02 2.10769668e-01 -4.08615112e-01 -3.23375702e-01
-5.92153668e-01 4.63132441e-01 3.35697010e-02 4.28888723e-02
8.01954031e-01 -2.67254472e-01 -5.16035318e-01 6.74023926e-01
-9.14424717e-01 -4.34375592e-02 8.78706634e-01 1.14048243e+00
3.55716258e-01 -3.31754476e-01 9.62167144e-01 -9.71207559e-01
1.40960968e+00 -7.25616813e-01 4.68857959e-02 8.46025586e-01
-8.22705030e-01 3.65366548e-01 8.21609795e-01 -9.19015706e-01
-8.33786488e-01 -3.35607529e-01 7.73108155e-02 -4.46438849e-01
-6.37444183e-02 6.89853311e-01 -1.60297602e-01 8.93935710e-02
1.10492790e+00 5.79334676e-01 -2.51990229e-01 -5.59737444e-01
1.36050016e-01 1.14132392e+00 6.12405539e-01 -1.02998793e+00
8.02553654e-01 -1.86604917e-01 -3.62853825e-01 -2.07530409e-01
-9.93789017e-01 -1.10285091e+00 -4.34678048e-01 5.69013432e-02
5.67250133e-01 -7.62519836e-01 -6.21924818e-01 4.28073049e-01
-1.21703529e+00 -5.54300584e-02 6.01462461e-02 2.75783867e-01
-2.61480510e-01 3.90377969e-01 -2.56604612e-01 -2.66771287e-01
-7.27576554e-01 -9.82925534e-01 1.22996950e+00 1.50802478e-01
-3.95238459e-01 -9.88660455e-01 1.12137767e-02 4.43489492e-01
5.49387455e-01 -5.69641352e-01 1.29399478e+00 -1.29807138e+00
-2.43154585e-01 -4.01916414e-01 -4.27588493e-01 2.20948577e-01
1.35258451e-01 -5.30366838e-01 -1.13100147e+00 -3.50230366e-01
2.04799294e-01 -4.96073335e-01 9.40614760e-01 7.58204237e-02
1.19561887e+00 -3.96735705e-02 -5.93266249e-01 4.82862294e-01
1.07263517e+00 9.64101404e-02 2.45550424e-01 4.04141396e-01
5.37817419e-01 7.39214599e-01 7.05847859e-01 -1.34575829e-01
2.73839861e-01 9.07428861e-01 -2.18589798e-01 2.26139158e-01
-1.81514204e-01 -4.55380112e-01 3.71389419e-01 8.85251522e-01
6.54655337e-01 -1.30429596e-01 -1.09274805e+00 4.13874537e-01
-1.97747040e+00 -1.05301881e+00 -2.11979002e-01 2.18387961e+00
1.22906864e+00 2.32962117e-01 -2.44065687e-01 -2.47617476e-02
8.18092108e-01 5.53609021e-02 -6.33110166e-01 -1.94065332e-01
-2.13651136e-01 1.81380883e-01 1.65617615e-01 2.76937842e-01
-8.17132473e-01 9.75980639e-01 5.71351767e+00 1.54632854e+00
-1.00363708e+00 2.36169159e-01 1.46706879e-01 3.62556994e-01
-6.98872209e-01 -5.07886037e-02 -1.14323294e+00 5.53960323e-01
6.95143163e-01 -7.09859669e-01 3.92741859e-01 7.34977424e-01
-1.45063430e-01 2.01679453e-01 -1.63786530e+00 1.29212689e+00
1.22163735e-01 -1.06862593e+00 5.00215948e-01 -5.28210461e-01
4.73614991e-01 1.00971140e-01 -9.64355916e-02 8.65150750e-01
-1.18515141e-01 -9.97607052e-01 5.72924078e-01 5.36804795e-01
7.59057581e-01 -1.83510274e-01 5.33852756e-01 8.93048108e-01
-1.41961467e+00 -2.06563249e-01 -6.65616214e-01 1.24892898e-01
-1.53917566e-01 2.86305755e-01 -6.66770458e-01 7.59864569e-01
5.38879633e-01 6.39234424e-01 -7.98052430e-01 8.18378925e-01
-3.35887074e-02 3.17937523e-01 -4.61651534e-02 -3.45738351e-01
7.49542192e-02 9.05446429e-03 3.37409705e-01 1.48744619e+00
1.19916081e-01 -1.26458049e-01 3.28909397e-01 9.02000725e-01
-1.87186792e-01 3.75155866e-01 -5.50664604e-01 2.25886360e-01
7.27220476e-01 9.61012244e-01 6.94060698e-02 -4.05727118e-01
-5.97181499e-01 1.03481984e+00 3.89015675e-01 4.72053200e-01
-3.91655833e-01 -4.40623134e-01 3.85473728e-01 -7.08239898e-02
-3.87809426e-01 3.03574085e-01 -3.83534282e-01 -1.29631805e+00
2.09380344e-01 -1.21071362e+00 7.23559201e-01 -6.02190137e-01
-1.91272783e+00 5.16535580e-01 1.80177689e-01 -1.40980470e+00
-3.90427530e-01 -5.75729489e-01 -8.02870035e-01 1.45691288e+00
-1.61713171e+00 -1.31053543e+00 -4.33726460e-01 1.01041818e+00
9.67301309e-01 -6.07720017e-01 7.55683661e-01 4.78019238e-01
-4.42170411e-01 1.11701107e+00 2.67980069e-01 1.46966472e-01
1.36307430e+00 -1.06728160e+00 4.61980283e-01 6.05741560e-01
3.28343928e-01 9.05961573e-01 3.52610379e-01 -5.26361346e-01
-1.37621307e+00 -8.02555323e-01 1.31414807e+00 -6.41100168e-01
7.91862488e-01 -4.23357606e-01 -1.41858816e+00 5.07341027e-01
5.01270443e-02 -1.83279663e-01 5.99604249e-01 4.25632983e-01
-6.84028566e-01 -3.65235299e-01 -7.49389470e-01 5.28435230e-01
1.12690604e+00 -1.19697952e+00 -1.14926076e+00 5.35784900e-01
6.44865215e-01 -3.52832049e-01 -6.86010659e-01 4.44500685e-01
5.22669137e-01 -5.36676109e-01 1.05755115e+00 -9.85748649e-01
3.06035548e-01 -4.47067544e-02 -2.16798037e-01 -1.30780876e+00
-5.76618612e-01 -3.57047021e-01 2.96476543e-01 1.35787833e+00
5.45099616e-01 -6.15286887e-01 3.36132944e-01 5.43736339e-01
-3.11223269e-01 -4.38299328e-01 -9.00531054e-01 -9.30152476e-01
3.98249120e-01 -5.60086548e-01 6.38894856e-01 1.33345616e+00
1.55011237e-01 8.49245727e-01 -1.96501389e-01 -1.06665127e-01
3.41571331e-01 4.60599005e-01 7.91567504e-01 -1.37648666e+00
-3.53340596e-01 -8.63587856e-01 2.23016933e-01 -1.65406561e+00
6.44850731e-01 -1.57836103e+00 -1.60356820e-01 -1.13696778e+00
5.73760927e-01 -6.78594053e-01 -5.29816329e-01 4.53991652e-01
-8.33203316e-01 -4.06201959e-01 1.31937265e-01 7.91917384e-01
-4.32823658e-01 3.43128860e-01 1.43276715e+00 -5.92431426e-01
-1.12909406e-01 3.43689352e-01 -6.49321854e-01 2.43038535e-01
6.28782213e-01 -7.95737624e-01 -5.92612803e-01 -6.74833000e-01
2.57964522e-01 3.10977101e-01 1.92489922e-01 -7.51901448e-01
6.58920407e-01 -3.23929846e-01 2.16443881e-01 -4.11098301e-01
-1.67452261e-01 -8.47031772e-01 -3.60827632e-02 4.58813667e-01
-8.85859251e-01 1.45398691e-01 2.87292600e-01 6.19988918e-01
-4.04548824e-01 -7.42007434e-01 4.80756253e-01 -1.46180183e-01
-7.58576691e-01 2.06334695e-01 6.38130084e-02 6.49002433e-01
3.75445664e-01 -3.60233992e-01 -5.16975164e-01 -9.08945650e-02
-2.62081772e-01 5.79375386e-01 -1.10541575e-01 1.01162529e+00
5.63196063e-01 -1.40657675e+00 -1.04920304e+00 2.09103957e-01
7.70099878e-01 -1.85295105e-01 1.54118374e-01 4.95056063e-01
2.77001709e-01 5.39799750e-01 -6.94097579e-02 -7.47627437e-01
-1.12529075e+00 4.90684062e-01 4.74901408e-01 -5.00859261e-01
-2.70464808e-01 7.01129317e-01 5.08466482e-01 -6.22814417e-01
1.55585930e-01 -4.61032540e-01 -1.94408402e-01 5.55191562e-02
4.94144201e-01 3.70171815e-02 7.69793168e-02 -3.21358353e-01
-3.46501648e-01 8.10726345e-01 -4.15547401e-01 1.40665323e-01
5.93855143e-01 -1.37924984e-01 4.68800366e-02 5.11528492e-01
1.29304171e+00 -2.18412802e-01 -5.28726459e-01 -1.00234985e+00
3.47809553e-01 -4.67602402e-01 -2.52421468e-01 -8.90911996e-01
-5.46837270e-01 1.00518930e+00 4.61915851e-01 1.00716330e-01
1.03639376e+00 2.25399863e-02 9.07821953e-01 6.63664699e-01
2.56167918e-01 -1.12599242e+00 4.30421531e-01 6.73625827e-01
1.06291091e+00 -1.35552204e+00 -3.22758526e-01 -2.50731051e-01
-4.92652565e-01 1.11565471e+00 8.76284897e-01 1.80176631e-01
4.13564056e-01 -2.78646499e-01 1.28188282e-02 -2.35284731e-01
-7.64513671e-01 -9.05175060e-02 8.39949667e-01 4.41984117e-01
5.82913935e-01 -6.86018243e-02 -2.43358552e-01 8.85157824e-01
-7.19228610e-02 -2.12945759e-01 -2.96028078e-01 6.51488900e-01
-5.61883867e-01 -1.35262537e+00 -2.38839298e-01 7.11907446e-01
-2.45861426e-01 -4.31673408e-01 -4.33726519e-01 5.88914275e-01
-1.33517504e-01 9.85938609e-01 -1.64359257e-01 -5.69579422e-01
8.56366456e-01 4.97942537e-01 2.78309226e-01 -7.60079145e-01
-1.08852732e+00 -3.66843909e-01 2.31401529e-02 -2.51712143e-01
-1.24258623e-01 -3.57837141e-01 -9.99558151e-01 -2.28900492e-01
-5.24608672e-01 1.40561730e-01 2.19726607e-01 1.21755695e+00
2.24362388e-01 3.83169025e-01 7.47490764e-01 -1.99639514e-01
-1.25342715e+00 -1.27622962e+00 -2.80205727e-01 1.09653831e+00
1.12176515e-01 -4.16503519e-01 -3.34569693e-01 -1.61453709e-01] | [11.104310989379883, 8.188011169433594] |
6d51e4a7-b29e-4406-ac46-5d9ada777b20 | graph-tensor-networks-an-intuitive-framework | 2303.13565 | null | https://arxiv.org/abs/2303.13565v1 | https://arxiv.org/pdf/2303.13565v1.pdf | Graph Tensor Networks: An Intuitive Framework for Designing Large-Scale Neural Learning Systems on Multiple Domains | Despite the omnipresence of tensors and tensor operations in modern deep learning, the use of tensor mathematics to formally design and describe neural networks is still under-explored within the deep learning community. To this end, we introduce the Graph Tensor Network (GTN) framework, an intuitive yet rigorous graphical framework for systematically designing and implementing large-scale neural learning systems on both regular and irregular domains. The proposed framework is shown to be general enough to include many popular architectures as special cases, and flexible enough to handle data on any and many data domains. The power and flexibility of the proposed framework is demonstrated through real-data experiments, resulting in improved performance at a drastically lower complexity costs, by virtue of tensor algebra. | ['Danilo P. Mandic', 'Kriton Konstantinidis', 'Yao Lei Xu'] | 2023-03-23 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [-2.71045178e-01 -2.12995317e-02 -3.46946530e-02 -4.24754888e-01
2.83026189e-01 -5.09544611e-01 8.12631130e-01 -8.21205005e-02
-1.53009564e-01 4.13480401e-01 -3.14486288e-02 -4.90280509e-01
-5.34250140e-01 -7.72610128e-01 -4.17286366e-01 -7.72755742e-01
-8.16370904e-01 1.28668517e-01 7.46523356e-03 -3.92655909e-01
1.40844630e-02 9.13294315e-01 -1.41147697e+00 6.18130388e-03
4.16103095e-01 1.15457046e+00 -3.21776509e-01 3.57521892e-01
-3.84607017e-02 9.00190592e-01 -1.59511745e-01 -7.96946704e-01
3.31765532e-01 1.19924150e-01 -8.92472565e-01 3.56228709e-01
8.77453506e-01 -1.06197372e-01 -7.15340674e-01 1.10432577e+00
1.29765391e-01 7.22271502e-02 3.12711209e-01 -1.29471052e+00
-9.34305251e-01 6.15453959e-01 -2.10303918e-01 3.90691847e-01
-1.70611948e-01 -6.05013408e-02 1.57977104e+00 -1.13020670e+00
3.28513533e-01 1.23841083e+00 8.10111880e-01 1.70852929e-01
-1.33238518e+00 -2.68081099e-01 2.34478787e-01 6.91480786e-02
-1.15954316e+00 -3.36666614e-01 9.11806524e-01 -5.48359394e-01
8.46159399e-01 1.04704566e-01 8.66066217e-01 7.49123454e-01
1.48533076e-01 9.05314386e-01 1.09065557e+00 -3.44626494e-02
2.31497437e-01 -3.82502347e-01 5.99686801e-01 1.18911910e+00
5.33721030e-01 1.53002376e-02 -5.76818764e-01 -1.24789134e-01
9.60445642e-01 1.41726062e-01 -8.15364644e-02 -6.68513596e-01
-1.46808839e+00 8.32300603e-01 7.83911705e-01 5.54636776e-01
-4.46951449e-01 4.62754190e-01 6.20551229e-01 4.93803054e-01
4.29313987e-01 2.40604088e-01 -8.82285386e-02 -6.42422065e-02
-6.11440659e-01 4.82024878e-01 8.30423236e-01 9.19417620e-01
5.43951035e-01 6.21610224e-01 2.63700545e-01 7.69493341e-01
3.34582597e-01 7.50306845e-02 3.21424574e-01 -8.20105851e-01
3.71392071e-01 6.94768846e-01 -4.11486328e-01 -1.42558110e+00
-5.77890873e-01 -9.37284410e-01 -1.35809231e+00 4.26115021e-02
3.27133805e-01 7.92154968e-02 -6.30792618e-01 1.61600268e+00
2.29540318e-01 1.36526793e-01 -6.34345859e-02 9.49836433e-01
6.46404505e-01 2.11410686e-01 -1.52373567e-01 2.71482825e-01
1.24256957e+00 -7.88244188e-01 -6.15743041e-01 8.40712921e-04
8.12103748e-01 -3.20726097e-01 8.69318008e-01 5.84287643e-01
-1.05562294e+00 -3.79579008e-01 -1.15711010e+00 -4.46598738e-01
-4.32709813e-01 6.50117919e-02 1.51912248e+00 6.33702695e-01
-1.30791545e+00 6.45144820e-01 -1.04420686e+00 -2.47598335e-01
7.19150305e-01 4.42172587e-01 -5.53324819e-01 -1.75561845e-01
-1.01566994e+00 7.09675729e-01 5.35830021e-01 6.74900234e-01
-7.86499858e-01 -7.17176795e-01 -7.59366393e-01 6.85049966e-02
2.44091060e-02 -9.31257665e-01 1.14683366e+00 -7.67815113e-01
-1.39642358e+00 6.04552984e-01 3.44610512e-01 -5.43225646e-01
3.26355875e-01 -5.33951148e-02 -2.29261503e-01 -5.13797514e-02
-4.05850857e-01 2.20991865e-01 7.04513669e-01 -8.20227265e-01
-3.69756460e-01 -6.03467345e-01 4.69903916e-01 -5.31821102e-02
-7.11976171e-01 -1.93152785e-01 -1.91241160e-01 -7.70330846e-01
2.99822420e-01 -8.68927240e-01 -2.92908281e-01 2.51663357e-01
-5.49668729e-01 -4.50046211e-01 6.93027139e-01 -2.11155355e-01
8.64586174e-01 -2.00983024e+00 3.69734526e-01 2.97997266e-01
9.09478605e-01 3.48396868e-01 -2.44061530e-01 4.65114176e-01
-2.95023054e-01 -4.47270907e-02 -7.77826980e-02 -3.34535480e-01
2.28965685e-01 5.18840134e-01 -2.86275119e-01 7.50097275e-01
2.84441859e-01 8.12207580e-01 -8.80435705e-01 -1.40053287e-01
1.15832083e-01 5.99979639e-01 -5.91622293e-01 2.93934508e-03
4.80604544e-02 -1.21531680e-01 -6.48094237e-01 6.13256335e-01
5.14046788e-01 -3.61147881e-01 1.78641990e-01 -4.18165296e-01
-1.73071638e-01 4.13004249e-01 -1.26040423e+00 1.62873495e+00
-2.68484056e-01 6.92370713e-01 3.20089549e-01 -1.37965012e+00
9.32600021e-01 1.81793556e-01 5.45156598e-01 -3.34692627e-01
1.28774002e-01 2.79143602e-01 2.15370685e-01 -3.62119943e-01
4.40111965e-01 -1.61102563e-01 1.92593247e-01 5.82826495e-01
4.17326361e-01 2.50919849e-01 4.97409046e-01 1.49785519e-01
1.11195397e+00 -2.74342954e-01 8.91746953e-02 -6.24164641e-01
4.71001208e-01 -1.74913168e-01 1.70017824e-01 4.11270410e-01
-7.81690776e-02 -1.54949538e-02 7.14372635e-01 -8.74879003e-01
-1.16021061e+00 -1.02205408e+00 -3.25978488e-01 1.19342208e+00
-2.63947487e-01 -4.19743955e-01 -5.34194231e-01 -4.17006373e-01
7.62431249e-02 1.70950487e-01 -5.47387362e-01 1.26778170e-01
-4.69774276e-01 -7.41648436e-01 7.88831830e-01 4.20553952e-01
5.24320543e-01 -7.51506150e-01 -1.61947548e-01 2.65452955e-02
4.28460151e-01 -1.23483932e+00 -2.03754723e-01 6.62832782e-02
-1.22526407e+00 -9.83223617e-01 -6.29316866e-01 -8.71514201e-01
5.42805135e-01 3.64673287e-01 1.09280181e+00 2.79758692e-01
-1.14691392e-01 3.47377151e-01 -3.06674521e-02 7.75634730e-03
-1.73801318e-01 1.34853035e-01 3.08365166e-01 4.01787043e-01
2.11424351e-01 -9.16335821e-01 -4.13387656e-01 5.53689338e-02
-1.30070293e+00 2.52385825e-01 4.86367643e-01 8.83573949e-01
2.70083427e-01 1.71666518e-01 2.66996354e-01 -7.25986421e-01
1.00825560e+00 -4.70666140e-01 -7.43628502e-01 8.92890766e-02
-6.62318289e-01 2.09421411e-01 8.73099267e-01 -3.21511567e-01
-5.49973190e-01 -3.05069506e-01 -1.89013295e-02 -3.01543534e-01
7.86935836e-02 9.78587329e-01 8.50892067e-02 -5.14747381e-01
4.43820834e-01 2.00719461e-01 1.65756494e-01 -6.29618526e-01
4.50235546e-01 2.78562725e-01 4.94151324e-01 -8.68955672e-01
8.93506169e-01 4.37939644e-01 6.08591080e-01 -1.06803155e+00
-6.09795988e-01 -9.48366895e-02 -7.46498466e-01 -1.33354187e-01
3.98702145e-01 -8.27079892e-01 -9.20059860e-01 4.69234884e-01
-1.18131983e+00 -3.37828025e-02 3.08882371e-02 4.24522251e-01
-3.23668897e-01 3.91442269e-01 -7.87093461e-01 -5.21449149e-01
-2.56648064e-01 -1.18144500e+00 6.37093127e-01 -2.02381253e-01
2.91421145e-01 -1.51101041e+00 9.22198314e-03 2.12355003e-01
4.47571576e-01 3.00715387e-01 1.06805027e+00 -5.84803522e-01
-7.56250799e-01 -3.27513069e-01 -5.75808406e-01 7.10961103e-01
-1.69298932e-01 1.60744682e-01 -8.73967946e-01 -3.03822547e-01
-1.95515513e-01 -2.57913202e-01 6.78876698e-01 -1.81020468e-01
1.24215519e+00 -3.99974495e-01 1.36083677e-01 7.16343582e-01
1.51096797e+00 -3.28240365e-01 4.06722039e-01 4.76932824e-02
1.13378263e+00 5.77961922e-01 -2.31694236e-01 4.04381543e-01
5.44216216e-01 6.32794321e-01 7.30308771e-01 -1.08364753e-01
-1.80550560e-01 -8.18347558e-03 2.04522848e-01 1.39213002e+00
-5.23565412e-01 1.51364610e-01 -1.13079000e+00 5.68451107e-01
-1.71781027e+00 -8.33287477e-01 -3.82427990e-01 1.79273129e+00
5.97557604e-01 1.38690203e-01 2.48062924e-01 2.92037666e-01
3.46618861e-01 3.04621279e-01 -2.92129040e-01 -6.63481593e-01
-8.85966420e-02 4.23098594e-01 4.14210737e-01 2.00432464e-01
-1.04928815e+00 7.88572192e-01 7.00396633e+00 5.39631128e-01
-1.17889893e+00 1.99736631e-03 2.69120425e-01 2.54695535e-01
-2.52137601e-01 -1.75213546e-01 -4.31151032e-01 -8.39136615e-02
8.37231338e-01 -2.67858237e-01 8.27984929e-01 1.00528169e+00
2.13268809e-02 6.51283920e-01 -1.42380309e+00 1.03014493e+00
-1.94841757e-01 -1.71301115e+00 1.31514356e-01 1.73294380e-01
5.61574519e-01 3.60441327e-01 1.67070270e-01 2.08899766e-01
2.31892630e-01 -9.73919988e-01 6.97417974e-01 4.09232289e-01
4.02182609e-01 -5.90394557e-01 5.75274944e-01 6.76190779e-02
-1.06276524e+00 -1.27106622e-01 -5.44563830e-01 -4.40109134e-01
-3.64993960e-01 8.73136282e-01 -6.67696536e-01 6.33518159e-01
5.63885748e-01 1.08317435e+00 -6.59269452e-01 1.09119546e+00
1.20452531e-01 5.51982760e-01 -2.74429917e-01 -3.17347348e-01
6.63682878e-01 -4.05860186e-01 3.25852633e-01 1.33525145e+00
1.82540998e-01 -1.98448703e-01 6.23011850e-02 8.17314804e-01
-2.39475936e-01 6.14578873e-02 -9.25176322e-01 -3.86805207e-01
1.19154163e-01 1.38078701e+00 -4.59621578e-01 -1.24672651e-01
-5.63955545e-01 3.74882013e-01 6.78077042e-01 6.59435928e-01
-5.58132350e-01 -4.91081119e-01 1.01189137e+00 -3.77543196e-02
2.35263720e-01 -8.86993647e-01 -6.42832458e-01 -1.35298836e+00
3.18671614e-01 -9.32200372e-01 1.96402505e-01 -4.59843159e-01
-1.57507908e+00 7.33487785e-01 -5.92958108e-02 -1.00189006e+00
-7.25352019e-02 -1.29550195e+00 -3.16849291e-01 6.99801624e-01
-1.40677798e+00 -1.11435103e+00 -1.95363238e-01 8.15602005e-01
-7.87151381e-02 -3.71144444e-01 8.30978692e-01 5.05569577e-01
-7.93313146e-01 3.62138510e-01 1.37187332e-01 4.31224614e-01
-2.05924511e-01 -1.35115004e+00 3.83012056e-01 7.75047123e-01
3.97964865e-01 1.11333740e+00 7.14153767e-01 8.79357755e-02
-2.13846326e+00 -1.03954041e+00 8.15266967e-01 -1.40423611e-01
1.42508328e+00 -6.46674573e-01 -1.01495349e+00 7.42301226e-01
1.30504355e-01 4.82449621e-01 5.51809609e-01 4.67129380e-01
-8.66345882e-01 -5.01938343e-01 -8.07961941e-01 8.65625858e-01
1.20861423e+00 -7.15363324e-01 -5.26052058e-01 5.56074977e-01
4.68599349e-01 -1.16702549e-01 -1.44836640e+00 2.92222738e-01
5.18686473e-01 -1.08035135e+00 9.85026121e-01 -8.41729283e-01
3.30406070e-01 -1.86261028e-01 -3.68868649e-01 -1.20901155e+00
-4.41514105e-01 -8.28332782e-01 -4.05254990e-01 9.01601911e-01
2.21242145e-01 -7.77589142e-01 7.61827111e-01 5.43682218e-01
-4.75668043e-01 -1.00115836e+00 -1.09319639e+00 -8.11709404e-01
1.78495273e-01 -8.26395392e-01 6.59658372e-01 1.08491623e+00
1.38923138e-01 2.99450696e-01 -1.21325664e-01 8.63520280e-02
7.59387314e-01 3.35922688e-02 6.10896409e-01 -1.47102952e+00
-2.39362538e-01 -9.04754877e-01 -9.05542850e-01 -1.04120564e+00
1.67590231e-01 -1.19211793e+00 -3.98138106e-01 -1.37836528e+00
-1.81069687e-01 -6.59733713e-01 -3.29892784e-01 6.86626494e-01
4.32208747e-01 4.14459407e-01 1.48068205e-01 2.61435539e-01
-5.08922160e-01 6.61722839e-01 1.43625295e+00 -4.71361354e-02
2.79631615e-01 -6.95652068e-02 -6.84493363e-01 6.89467967e-01
6.00431800e-01 -1.01616390e-01 -5.96373618e-01 -8.09047461e-01
4.34431046e-01 -1.13297306e-01 6.11426592e-01 -9.85249758e-01
1.52461112e-01 2.38941051e-03 -1.24713220e-02 -5.37560023e-02
3.39348525e-01 -9.06430066e-01 -3.47800180e-02 3.02683771e-01
-4.74721462e-01 5.28568506e-01 4.13251892e-02 4.18973237e-01
-3.30845743e-01 8.55739862e-02 5.33736348e-01 2.33437791e-01
-6.64946437e-01 8.60073268e-01 3.28490883e-02 -6.96114823e-02
7.48123169e-01 -4.41849530e-02 -3.89826983e-01 5.75074218e-02
-6.64758503e-01 -1.31348476e-01 1.07425980e-01 3.93576920e-01
6.75794065e-01 -1.78090012e+00 -4.42871451e-01 3.15626472e-01
1.52592614e-01 -7.57037103e-02 1.16573855e-01 1.08939993e+00
-7.38755703e-01 7.15089798e-01 -3.71130645e-01 -6.54971540e-01
-8.03043008e-01 5.43518484e-01 4.42817688e-01 -2.55555809e-01
-8.08709800e-01 6.12991512e-01 9.38391462e-02 -4.67111260e-01
5.05117655e-01 -6.70726120e-01 -1.03363162e-02 -2.58785814e-01
3.47788990e-01 3.34286392e-01 3.95010054e-01 -5.75490355e-01
-2.50989437e-01 2.86483496e-01 9.52550322e-02 2.17673033e-01
1.43844688e+00 2.73100764e-01 -6.11907780e-01 5.15658259e-01
1.40762794e+00 -6.08857453e-01 -1.00033069e+00 -5.31184196e-01
2.90620506e-01 -9.62307155e-02 4.32581872e-01 -2.56712079e-01
-1.42207515e+00 1.06657326e+00 2.78912485e-01 8.43686402e-01
9.64859426e-01 -2.04579756e-01 5.85771263e-01 9.29084539e-01
5.16295314e-01 -6.82990909e-01 4.41664718e-02 7.97016501e-01
1.03563702e+00 -8.23848903e-01 1.16604073e-02 -2.27536291e-01
-3.45573843e-01 1.49106002e+00 3.22724521e-01 -4.90044892e-01
1.09581196e+00 6.87756464e-02 -3.15764844e-01 -5.73299646e-01
-8.62302363e-01 -5.68190683e-03 5.04611552e-01 5.54495990e-01
8.10288310e-01 1.84672922e-01 -1.62969530e-01 1.59649178e-01
-3.05995286e-01 3.58853824e-02 5.28802574e-01 6.32655442e-01
-1.66242436e-01 -9.55955803e-01 -1.50621533e-01 4.75109786e-01
-5.11814058e-01 -1.38532557e-02 -1.65638670e-01 8.99815083e-01
6.71850592e-02 5.75991154e-01 3.67174596e-02 -4.62592542e-01
1.33651450e-01 -2.23942921e-01 5.15969038e-01 -5.10042071e-01
-5.36823094e-01 -3.66921008e-01 1.56492397e-01 -6.12760127e-01
-6.06868446e-01 -5.61753213e-01 -1.00591314e+00 -7.24584222e-01
2.36297816e-01 -9.12819952e-02 9.59076583e-01 9.44185436e-01
3.31528783e-01 3.38403851e-01 5.86120844e-01 -7.82057047e-01
-7.90636480e-01 -8.10403049e-01 -9.26377594e-01 2.81144023e-01
5.68334162e-01 -8.46299052e-01 -1.95205852e-01 -1.52917266e-01] | [6.249281883239746, 5.105773448944092] |
2172bd73-a050-4ccd-bcc9-f65c8ec9c07a | huber-additive-models-for-non-stationary-time | null | null | https://openreview.net/forum?id=9kpuB2bgnim | https://openreview.net/pdf?id=9kpuB2bgnim | Huber Additive Models for Non-stationary Time Series Analysis | Sparse additive models have shown promising flexibility and interpretability in processing time series data. However, existing methods usually assume the time series data to be stationary and the innovation is sampled from a Gaussian distribution. Both assumptions are too stringent for heavy-tailed and non-stationary time series data that frequently arise in practice, such as finance and medical fields. To address these problems, we propose an adaptive sparse Huber additive model for robust forecasting and inference (e.g., Granger causal discovery) in both non-Gaussian data and (non)stationary data. In theory, the generalization bounds of our estimator are established for both stationary and non-stationary time series data, which are independent of the widely used mixing conditions in learning theory of dependent observations. Moreover, the error bound for non-stationary time series contains a discrepancy measure for the shifts of the data distributions over time. Such a discrepancy measure can be estimated empirically and used as a penalty in our method. Experimental results on both synthetic and real-world benchmark datasets validate the effectiveness of the proposed method. | ['DaCheng Tao', 'Hong Chen', 'Fengxiang He', 'Xianrui Zhong', 'Yingjie Wang'] | 2021-09-29 | null | null | null | iclr-2022-4 | ['additive-models'] | ['methodology'] | [ 1.47760227e-01 -2.90585130e-01 -9.93027911e-02 -3.41066897e-01
-6.97308302e-01 -4.60362166e-01 5.44422865e-01 1.38583541e-01
-2.53030807e-01 8.45541060e-01 9.38496217e-02 -3.26005876e-01
-5.74936330e-01 -5.83629251e-01 -7.01000035e-01 -1.14607036e+00
-3.05371493e-01 3.71954560e-01 4.24680579e-03 2.10512683e-01
1.15243204e-01 1.96329638e-01 -1.14916205e+00 -4.34787482e-01
1.15356112e+00 1.08828640e+00 1.04649626e-01 4.21100110e-01
1.13708168e-01 6.33869827e-01 -4.78228837e-01 -1.93711445e-01
3.84672791e-01 -4.08947736e-01 -4.80678640e-02 1.03307385e-02
-1.26107171e-01 1.22105859e-01 -2.33045265e-01 1.26650143e+00
3.54045242e-01 3.08652282e-01 8.78487110e-01 -1.43437231e+00
-6.03534520e-01 6.18698478e-01 -8.79570484e-01 4.14744347e-01
-1.43076628e-01 -2.60229558e-01 6.53666019e-01 -7.39172876e-01
-9.97639000e-02 1.14128220e+00 7.93839514e-01 -6.97187753e-03
-9.04476881e-01 -7.92799711e-01 1.43974215e-01 2.53293663e-01
-1.28743494e+00 -2.99516231e-01 9.16673660e-01 -6.31639659e-01
1.60091281e-01 2.47084171e-01 2.61891782e-01 1.21815169e+00
5.17487943e-01 6.05142772e-01 1.12738252e+00 -2.19888031e-01
6.80313468e-01 -5.79277724e-02 4.43702452e-02 1.24948278e-01
5.66630960e-01 1.30618572e-01 -2.46908158e-01 -5.56021035e-01
8.53268504e-01 4.15508837e-01 -1.29206493e-01 -1.28142923e-01
-1.19764173e+00 8.97665620e-01 -9.16583315e-02 4.76621687e-01
-8.07521641e-01 9.30806696e-02 3.77322346e-01 3.11229825e-01
9.21284437e-01 -9.86423865e-02 -5.17307401e-01 -1.73377454e-01
-8.65836680e-01 4.29819599e-02 6.20239735e-01 8.96719992e-01
2.56596237e-01 4.98360634e-01 -5.77470213e-02 5.13307869e-01
1.56590670e-01 9.26140010e-01 5.37851095e-01 -7.12606788e-01
3.28967959e-01 -3.63691300e-02 3.80065590e-01 -1.20530045e+00
-3.19428056e-01 -8.07247758e-01 -1.46370268e+00 -3.81448597e-01
7.29614854e-01 -2.61741668e-01 -6.23898804e-01 1.83994329e+00
3.99977237e-01 1.06972039e+00 -2.69142929e-02 7.58840501e-01
2.96579599e-01 6.90319240e-01 -5.87101504e-02 -9.24082875e-01
1.00986111e+00 -3.40286672e-01 -9.93333995e-01 1.00162499e-01
2.93023378e-01 -8.02850842e-01 7.80699670e-01 3.71146768e-01
-9.27937448e-01 -3.20108920e-01 -5.18052518e-01 4.17022705e-01
1.25461400e-01 -1.24873526e-01 5.59490860e-01 3.42632592e-01
-4.49995756e-01 3.64805460e-01 -1.02957141e+00 -4.95122895e-02
1.77715749e-01 4.27212566e-02 -2.70760804e-01 -4.24384773e-02
-1.21890390e+00 3.07586014e-01 3.70902359e-01 1.62145555e-01
-5.77759743e-01 -9.49459195e-01 -4.69698548e-01 6.87345639e-02
3.59469712e-01 -8.17514479e-01 1.05545723e+00 -1.07674301e+00
-1.21069717e+00 1.45610899e-01 -3.04180950e-01 -6.65226042e-01
5.02672672e-01 -1.03887551e-01 -7.56627619e-01 -1.04881965e-01
1.32426292e-01 -5.43335915e-01 1.37634397e+00 -8.31854582e-01
-5.83027363e-01 -4.63935524e-01 -3.63629192e-01 -2.86677539e-01
-2.60203451e-01 -1.32985786e-01 1.39244020e-01 -1.10155249e+00
2.99787730e-01 -9.90865529e-01 -4.71025109e-01 -3.21823537e-01
-2.32758284e-01 -3.95318478e-01 6.84447169e-01 -7.36390114e-01
1.18822610e+00 -2.44384885e+00 -1.72676474e-01 3.68560374e-01
-3.82484309e-02 -2.82017708e-01 9.24345776e-02 4.04321551e-01
-3.12220037e-01 -2.82601655e-01 -2.71532655e-01 -1.64551780e-01
-1.52057096e-01 2.34681308e-01 -5.30601382e-01 1.03003013e+00
-2.11841211e-01 2.80645430e-01 -1.15629518e+00 -2.85739213e-01
-5.24536856e-02 3.86574745e-01 -2.35069960e-01 7.35929608e-02
1.12223722e-01 7.05956161e-01 -4.87913400e-01 3.49092305e-01
6.52350962e-01 -4.96726036e-01 -2.30794936e-01 -4.74424995e-02
2.33013574e-02 -7.59409890e-02 -1.38630855e+00 1.21834588e+00
-4.28514689e-01 4.28675622e-01 -2.15146601e-01 -1.38854921e+00
8.94659042e-01 6.01717114e-01 7.53406942e-01 -3.70177448e-01
2.15565324e-01 2.87802517e-01 1.89713240e-01 -5.13664722e-01
1.92253992e-01 -5.65975606e-01 8.77853334e-02 2.50878483e-01
-2.17056930e-01 2.49884084e-01 6.45488501e-02 -2.33729899e-01
1.00902402e+00 -4.22016472e-01 5.23378789e-01 -4.10766721e-01
2.91473716e-01 -4.71774787e-01 9.26365554e-01 8.61535728e-01
5.81970513e-02 3.62753719e-01 3.33937526e-01 -1.91151917e-01
-1.08073485e+00 -1.05281150e+00 -1.28137797e-01 9.07741904e-01
-1.28868490e-01 -1.60987731e-02 -2.17859074e-01 -3.53948683e-01
-1.81652568e-02 1.01391733e+00 -6.44143462e-01 -1.65114090e-01
-2.94139981e-01 -1.00490737e+00 1.22300431e-01 6.13514662e-01
-1.27534634e-02 -4.47261751e-01 -3.14455360e-01 3.36603165e-01
-1.29656360e-01 -1.06985617e+00 -5.80147624e-01 8.50925744e-02
-1.02053499e+00 -9.56135511e-01 -8.63912463e-01 -3.30430686e-01
7.32018173e-01 4.45129007e-01 8.04493308e-01 -5.32198071e-01
8.42005610e-02 4.69078660e-01 -2.39473745e-01 -7.06862450e-01
-1.88776568e-01 -4.25268263e-01 2.84419358e-01 4.75643247e-01
1.83055043e-01 -7.01824009e-01 -7.02239394e-01 3.57141942e-01
-1.01665223e+00 -2.67837197e-01 4.25062418e-01 9.29347813e-01
6.56978309e-01 6.97757363e-01 8.62831295e-01 -5.59675217e-01
5.97237945e-01 -1.03666210e+00 -8.18429351e-01 2.01944038e-01
-6.25805974e-01 -1.14702269e-01 7.74774969e-01 -1.08805728e+00
-1.01611972e+00 -1.13910571e-01 3.60946715e-01 -6.39584422e-01
-2.89584529e-02 9.52079654e-01 9.26379859e-02 2.79600948e-01
3.72649014e-01 4.19272989e-01 2.48792372e-03 -3.35445076e-01
2.06760138e-01 5.19873798e-01 5.99255621e-01 -5.06003380e-01
9.56610024e-01 6.96491897e-01 4.14845139e-01 -9.80944157e-01
-7.27342784e-01 -6.32786214e-01 -2.88854957e-01 7.56596774e-02
4.43546593e-01 -9.94518876e-01 -5.65176606e-01 3.46946836e-01
-8.81755114e-01 3.58625315e-02 -3.08751613e-01 1.05466223e+00
-6.08970404e-01 4.27974373e-01 -4.45003003e-01 -1.42726612e+00
-2.23606199e-01 -6.49380565e-01 8.66869450e-01 1.90374777e-01
-2.35195071e-01 -1.59116304e+00 1.01317793e-01 -1.21734098e-01
4.19935822e-01 4.22112256e-01 9.67430949e-01 -9.14132833e-01
-8.13594759e-02 -6.49916053e-01 6.78846687e-02 2.68410027e-01
3.70870024e-01 5.35547324e-02 -5.18999755e-01 -1.70967102e-01
4.54179704e-01 3.50357980e-01 5.00233233e-01 8.53027105e-01
1.20262837e+00 -6.13870323e-01 -2.55869091e-01 2.63034910e-01
1.13907313e+00 3.61449122e-01 2.88011909e-01 -4.04799096e-02
4.05027002e-01 4.47459906e-01 6.83602750e-01 1.03497577e+00
2.12843269e-01 3.12065244e-01 3.38830650e-01 1.40231743e-01
7.19157696e-01 -1.95227191e-01 3.13245833e-01 1.22667468e+00
-6.89582825e-02 -3.02025139e-01 -8.26353490e-01 7.55104721e-01
-2.07158399e+00 -1.34507561e+00 -6.74769640e-01 2.73172736e+00
8.32837045e-01 5.42205572e-02 1.98426157e-01 1.28779113e-01
8.36956322e-01 -1.50055841e-01 -7.30734229e-01 2.14762211e-01
-3.09213132e-01 -1.21383574e-02 7.59539604e-01 7.81979412e-02
-1.12435567e+00 2.87619680e-01 5.94199610e+00 9.16850269e-01
-1.22737288e+00 2.67598063e-01 4.67075735e-01 8.93194154e-02
-3.34440887e-01 -6.53485116e-03 -2.44320169e-01 8.93318474e-01
1.11829591e+00 -7.97390640e-01 1.32042035e-01 7.28583992e-01
6.02315485e-01 1.91596866e-01 -8.44994485e-01 1.34509730e+00
-2.72294283e-01 -8.50048542e-01 -5.15175104e-01 2.10869219e-02
9.37883198e-01 -9.97701064e-02 3.25837940e-01 1.48071364e-01
2.94699311e-01 -7.19501376e-01 5.95330358e-01 7.37728655e-01
2.72954196e-01 -6.81660771e-01 8.05648744e-01 6.31733775e-01
-9.08744574e-01 -2.09010705e-01 -3.92879069e-01 -8.71612281e-02
2.51760274e-01 1.33878160e+00 -5.47269881e-01 5.72231114e-01
4.73506272e-01 8.88419926e-01 -1.40532434e-01 1.39356267e+00
-6.92928443e-03 1.11127448e+00 -6.31208837e-01 1.97070390e-01
1.63376719e-01 -4.51170295e-01 7.43627727e-01 7.80404747e-01
9.62778151e-01 1.14194244e-01 1.07245825e-01 5.12951732e-01
3.43225032e-01 1.75955951e-01 -4.52182353e-01 -3.28599125e-01
6.14635050e-01 9.62538600e-01 -6.82768822e-01 -3.48496765e-01
-5.93636870e-01 6.01113319e-01 -3.75421435e-01 6.99194670e-01
-1.13092768e+00 -6.29020780e-02 4.85943705e-01 1.93498492e-01
3.79411727e-01 -4.51080978e-01 -2.58281648e-01 -1.19546425e+00
6.06292151e-02 -7.24101305e-01 5.25334418e-01 -3.50420147e-01
-1.90278506e+00 3.39413196e-01 4.06215005e-02 -1.51102328e+00
-4.04691249e-01 -1.22563697e-01 -7.06234515e-01 6.17410719e-01
-1.11923635e+00 -7.96930432e-01 2.12672018e-02 8.95866871e-01
4.57492203e-01 -9.84376743e-02 4.85193163e-01 3.53696972e-01
-4.65788782e-01 3.28499734e-01 7.60491550e-01 -1.48891926e-01
7.51612425e-01 -1.17860174e+00 4.10340987e-02 8.52531552e-01
2.00382974e-02 7.91238427e-01 1.19107139e+00 -6.82890892e-01
-1.40031099e+00 -1.18997467e+00 4.52016771e-01 -5.24466671e-02
1.12906897e+00 -5.06466851e-02 -1.16002715e+00 6.01011097e-01
-3.55716869e-02 1.60344794e-01 9.79192734e-01 2.01430276e-01
-2.51049548e-01 -1.52948588e-01 -9.09581721e-01 2.64768451e-01
5.50747931e-01 -1.94120705e-01 -4.65322435e-01 6.80629551e-01
5.13058066e-01 -1.28662899e-01 -9.71351027e-01 4.86202627e-01
4.11359757e-01 -5.11739075e-01 7.85276294e-01 -9.11589146e-01
2.38860801e-01 -3.16822171e-01 -2.42148563e-01 -1.43604994e+00
-6.10198319e-01 -8.60367417e-01 -4.38142598e-01 1.16701853e+00
1.30217612e-01 -9.64848220e-01 2.09219411e-01 4.70819980e-01
9.90033001e-02 -3.88801187e-01 -1.15457070e+00 -1.28768218e+00
-4.35296968e-02 -5.40596426e-01 5.31185389e-01 1.17691517e+00
-2.63412684e-01 1.94311395e-01 -6.72985673e-01 4.97910321e-01
1.02317238e+00 1.83989912e-01 5.63368499e-01 -1.57849240e+00
-3.53231728e-01 -3.77868474e-01 -3.96987557e-01 -8.43967319e-01
4.00328040e-01 -4.23863053e-01 1.33989975e-01 -1.09671545e+00
1.09880790e-02 -5.80760598e-01 -6.34848952e-01 -4.53721806e-02
-4.07060504e-01 -3.26837927e-01 -2.34055400e-01 3.56903434e-01
-4.12504613e-01 8.08402479e-01 8.25847208e-01 1.85495868e-01
-6.76204041e-02 6.16007984e-01 -5.39540946e-01 9.35282409e-01
7.72644341e-01 -6.50111496e-01 -7.78038442e-01 -1.24420978e-01
2.46735603e-01 3.90437305e-01 3.82885635e-01 -6.67575657e-01
2.51399517e-01 -5.34533322e-01 4.53014821e-02 -6.37808025e-01
1.78636014e-02 -1.29141653e+00 5.19535601e-01 2.94892371e-01
-1.90493464e-01 4.80543785e-02 -1.91195652e-01 1.17333353e+00
-2.71688610e-01 2.55268253e-02 6.22837484e-01 2.93018758e-01
-2.44642481e-01 4.65383530e-01 -1.22135304e-01 1.50038362e-01
1.10183728e+00 9.57610831e-02 -1.86935261e-01 -9.27953362e-01
-4.65396136e-01 3.22165787e-02 -5.16659021e-02 4.27433968e-01
4.66896027e-01 -1.43299520e+00 -9.03491974e-01 1.23720996e-01
-1.51269078e-01 -1.20629609e-01 5.64129770e-01 1.56591535e+00
1.71338782e-01 3.54558080e-01 2.55382299e-01 -6.82203650e-01
-9.30806696e-01 1.00826478e+00 -1.35473832e-01 9.94660109e-02
-4.19160664e-01 4.58405256e-01 3.56728017e-01 2.15496793e-01
1.29886523e-01 -5.30737162e-01 1.08009927e-01 5.79939932e-02
7.07687140e-01 6.13229811e-01 -1.40036911e-01 -4.24578249e-01
-3.45239997e-01 2.01434299e-01 1.04418509e-01 3.25611383e-02
1.47914040e+00 -2.30808169e-01 -2.40247220e-01 1.22487879e+00
1.09962010e+00 1.05974481e-01 -1.13291669e+00 -5.81440628e-01
1.18605874e-01 -4.02020603e-01 -8.75020623e-02 -2.24087283e-01
-8.13534439e-01 6.33346200e-01 5.52711785e-01 7.67406166e-01
1.23643160e+00 -1.71472922e-01 7.83935785e-01 -1.31166160e-01
4.54163969e-01 -8.54334235e-01 -1.78877473e-01 3.74712586e-01
7.88181663e-01 -1.10786712e+00 -1.22505717e-01 -2.66166151e-01
-4.08175498e-01 9.70184624e-01 8.22848082e-03 -2.78126210e-01
1.03695512e+00 1.93806723e-01 -2.47426778e-01 2.96950608e-01
-8.77168536e-01 4.53794003e-02 5.41935384e-01 4.10121083e-01
5.53220391e-01 2.80219346e-01 -4.56114054e-01 8.35640609e-01
-1.41846061e-01 1.27243951e-01 3.87744486e-01 6.19850814e-01
-2.41452575e-01 -5.20863116e-01 -6.54914916e-01 4.06703740e-01
-6.27459049e-01 3.51663530e-02 3.22216541e-01 5.55463493e-01
-2.12869599e-01 9.60919142e-01 5.56432791e-02 8.14351737e-02
6.27836287e-02 -8.07877406e-02 1.77706584e-01 -1.54043674e-01
2.77836137e-02 6.38396621e-01 -4.50404882e-01 -1.47561744e-01
-4.82184052e-01 -8.17113876e-01 -9.76894021e-01 -3.11248273e-01
-4.59856838e-01 2.94886947e-01 5.28898895e-01 1.03070211e+00
3.57443571e-01 5.26211441e-01 8.04948986e-01 -2.35671759e-01
-9.42656696e-01 -1.04858780e+00 -8.47476721e-01 3.52039307e-01
3.79645437e-01 -7.93290734e-01 -6.61032975e-01 2.87298173e-01] | [6.9487409591674805, 3.7449095249176025] |
1fb288b4-fe6f-438d-af0e-4d504592b003 | forest-parameter-prediction-by-multiobjective | 2306.11103 | null | https://arxiv.org/abs/2306.11103v1 | https://arxiv.org/pdf/2306.11103v1.pdf | Forest Parameter Prediction by Multiobjective Deep Learning of Regression Models Trained with Pseudo-Target Imputation | In prediction of forest parameters with data from remote sensing (RS), regression models have traditionally been trained on a small sample of ground reference data. This paper proposes to impute this sample of true prediction targets with data from an existing RS-based prediction map that we consider as pseudo-targets. This substantially increases the amount of target training data and leverages the use of deep learning (DL) for semi-supervised regression modelling. We use prediction maps constructed from airborne laser scanning (ALS) data to provide accurate pseudo-targets and free data from Sentinel-1's C-band synthetic aperture radar (SAR) as regressors. A modified U-Net architecture is adapted with a selection of different training objectives. We demonstrate that when a judicious combination of loss functions is used, the semi-supervised imputation strategy produces results that surpass traditional ALS-based regression models, even though \sen data are considered as inferior for forest monitoring. These results are consistent for experiments on above-ground biomass prediction in Tanzania and stem volume prediction in Norway, representing a diversity in parameters and forest types that emphasises the robustness of the approach. | ['Lennart Noordermeer', 'Terje Gobakken', 'Erik Næsset', 'Michael Kampffmeyer', 'Stian N. Anfinsen', 'Sara Björk'] | 2023-06-19 | null | null | null | null | ['imputation', 'parameter-prediction', 'imputation', 'imputation'] | ['computer-vision', 'miscellaneous', 'miscellaneous', 'time-series'] | [ 9.33659554e-01 1.30383343e-01 -5.16826153e-01 -6.78099394e-01
-9.66686010e-01 -4.56501007e-01 7.69474089e-01 -1.00789264e-01
-4.77303594e-01 1.48444438e+00 1.96858287e-01 -6.45240128e-01
-3.45915586e-01 -1.11738229e+00 -7.78028905e-01 -7.52127528e-01
-3.78975987e-01 6.31227314e-01 -2.83573478e-01 -4.30663675e-01
-5.43093216e-03 6.95810556e-01 -1.56615734e+00 1.66800261e-01
1.04215360e+00 1.00792694e+00 4.96311724e-01 5.90999782e-01
9.51194167e-02 8.51796985e-01 -1.86993241e-01 -8.17923099e-02
6.69734061e-01 -1.58210725e-01 -5.44433475e-01 -2.27176905e-01
7.49801457e-01 -5.81491947e-01 -3.88029851e-02 6.20805144e-01
3.51054311e-01 -3.88036445e-02 7.57172048e-01 -9.00189757e-01
-3.34194571e-01 9.45449054e-01 -6.35842800e-01 -6.66306093e-02
-3.74842674e-01 1.90177754e-01 1.10819781e+00 -3.72143060e-01
4.58522469e-01 9.76641715e-01 1.48334622e+00 2.16762677e-01
-2.08008194e+00 -9.15883183e-01 -5.90028502e-02 -2.39544645e-01
-1.46652353e+00 -7.73449004e-01 3.54445606e-01 -5.37830472e-01
8.32483709e-01 4.92348373e-01 6.10185921e-01 9.85276520e-01
3.77770871e-01 2.98395246e-01 1.32027066e+00 -3.22851360e-01
3.51202600e-02 -3.62502709e-02 -2.66902030e-01 4.92038392e-02
2.88943768e-01 9.34082508e-01 -1.86608449e-01 -1.28597721e-01
7.82908678e-01 8.67629275e-02 -2.51492172e-01 -4.78311449e-01
-1.20433033e+00 9.93380785e-01 7.84325421e-01 -1.72670037e-01
-7.28752255e-01 2.22061560e-01 1.81040943e-01 3.79824609e-01
8.80344748e-01 4.42610383e-01 -8.24624658e-01 3.72656256e-01
-1.63554013e+00 3.42414737e-01 6.63428187e-01 5.82312286e-01
7.36733019e-01 6.21140480e-01 2.51317751e-02 1.11864531e+00
1.51713088e-01 9.23648655e-01 3.12909037e-02 -9.60381687e-01
3.66630167e-01 3.91348451e-01 4.08339709e-01 -7.70320177e-01
-3.25185031e-01 -9.92152512e-01 -1.04711330e+00 5.74357152e-01
3.91868055e-01 -2.86497682e-01 -1.10580635e+00 1.67418110e+00
-9.03435722e-02 -8.98990333e-02 1.15301684e-01 6.72911942e-01
3.88941258e-01 5.85262775e-01 4.28465873e-01 -4.17697579e-01
4.05503869e-01 -3.74925435e-01 -3.28217745e-01 -6.45091236e-01
5.75567365e-01 -3.70792717e-01 6.45196319e-01 2.02023193e-01
-5.70575535e-01 -5.71524262e-01 -9.22041714e-01 4.91429657e-01
-4.78915155e-01 3.47048789e-01 8.69661570e-01 5.96172690e-01
-8.92077267e-01 1.03160679e+00 -7.85702169e-01 -3.68450969e-01
7.67074168e-01 4.05635029e-01 -3.43673110e-01 5.64972945e-02
-8.88997674e-01 1.26950872e+00 3.36833477e-01 6.19140744e-01
-7.82628834e-01 -9.25273478e-01 -5.46239018e-01 -1.47384703e-01
4.37558480e-02 -2.34173253e-01 9.12191153e-01 -1.49564922e+00
-9.03594434e-01 8.07794154e-01 2.11002097e-01 -8.65234554e-01
5.94516397e-01 -1.28033310e-01 -5.36277115e-01 -5.53068936e-01
4.20164466e-01 8.52700412e-01 7.07614481e-01 -1.53974104e+00
-7.08623528e-01 -6.95385993e-01 -3.34784657e-01 1.76262483e-02
4.09040689e-01 -1.48164049e-01 9.16799963e-01 -6.89111412e-01
3.76882434e-01 -8.25104296e-01 -5.71181357e-01 -1.97904572e-01
-2.68649191e-01 8.24303091e-01 4.26953495e-01 -9.47616041e-01
9.67844486e-01 -1.69477546e+00 -3.66112813e-02 2.49630347e-01
-1.39535308e-01 3.05337906e-02 -3.76823068e-01 4.75171596e-01
-3.94765943e-01 1.98261231e-01 -8.13033283e-01 1.73997045e-01
-2.99512684e-01 5.65539122e-01 -7.46236384e-01 5.23735881e-01
2.80315518e-01 7.63404608e-01 -5.53428829e-01 -1.72520772e-01
1.99837863e-01 2.68365890e-01 -1.63759276e-01 1.28837377e-01
-3.55013132e-01 5.29452860e-01 -9.24859568e-02 8.38941395e-01
9.94562089e-01 4.13636893e-01 2.28333041e-01 -6.64555328e-03
-4.26444441e-01 2.77379155e-01 -7.87459254e-01 1.13336051e+00
-6.02501869e-01 5.37448168e-01 3.41819555e-01 -1.04065740e+00
1.28386927e+00 -1.90987974e-01 3.94209355e-01 -4.65618998e-01
-4.55143720e-01 3.04986894e-01 3.60168129e-01 9.41848233e-02
3.59528244e-01 -5.82531631e-01 3.25817168e-01 1.92569748e-01
1.21995108e-02 -2.80623466e-01 -2.48474553e-01 -4.75860804e-01
9.53242779e-01 8.57683122e-01 3.03769082e-01 -4.54435110e-01
2.49894455e-01 8.45471978e-01 7.65885472e-01 8.14635038e-01
1.00735895e-01 4.07096326e-01 2.30592832e-01 -6.34017885e-01
-1.16234040e+00 -1.14566875e+00 -8.67727101e-01 1.24732947e+00
-7.26421952e-01 1.67672038e-01 6.78519383e-02 -3.67127270e-01
5.82139790e-01 1.17271900e+00 -7.20220506e-01 1.67914882e-01
-1.50081277e-01 -1.36360490e+00 6.96595967e-01 5.61469257e-01
2.66811192e-01 -1.05242455e+00 -5.86467981e-01 4.03087288e-01
4.29801233e-02 -4.89723861e-01 5.27399480e-01 1.02104437e+00
-1.33387935e+00 -9.47723508e-01 -4.85533267e-01 -1.25550702e-01
3.68560016e-01 2.84150336e-02 1.35942495e+00 -4.19103235e-01
-9.17049944e-02 -1.46140516e-01 -3.21479261e-01 -5.30557573e-01
-5.49094975e-01 3.75727266e-01 4.78187837e-02 -4.29021090e-01
5.42325318e-01 -1.05836141e+00 -1.70960546e-01 1.64911568e-01
-4.94027376e-01 5.99115603e-02 1.06894994e+00 1.06075478e+00
6.10890031e-01 -1.56245112e-01 9.27612841e-01 -1.01714242e+00
1.03880607e-01 -5.29283762e-01 -7.54156232e-01 8.52039233e-02
-9.13254857e-01 -4.94755767e-02 4.88828778e-01 -1.62282884e-01
-9.34142530e-01 3.52459818e-01 1.67606547e-01 -1.67586580e-01
-4.08817470e-01 9.62241411e-01 9.26881433e-02 -1.54747829e-01
9.43362772e-01 3.36209983e-01 4.49753255e-01 -4.44940209e-01
1.06955141e-01 8.37173104e-01 5.32414556e-01 -2.71993846e-01
8.94823909e-01 9.39576849e-02 3.62273782e-01 -8.41756523e-01
-9.55376565e-01 -5.73672168e-02 -1.00297964e+00 -2.17242196e-01
2.62086481e-01 -1.17596948e+00 -1.88911110e-01 2.35827699e-01
-5.44981480e-01 -8.52739573e-01 -6.25718474e-01 7.11820006e-01
-8.32972348e-01 -1.33099943e-01 9.80541483e-02 -1.19512689e+00
-3.65372628e-01 -6.04213774e-01 8.57083261e-01 -4.16931719e-01
-6.42311852e-03 -7.56018937e-01 1.52421087e-01 1.19772077e-01
6.83227003e-01 7.84498394e-01 6.18695796e-01 -4.42365348e-01
-5.05993605e-01 -2.83406913e-01 -4.49869633e-01 6.13230526e-01
2.71124929e-01 3.97852547e-02 -9.11270797e-01 -1.77999869e-01
-3.96547079e-01 -4.57254320e-01 1.36204350e+00 9.22587454e-01
9.18400288e-01 -3.31373006e-01 -3.07711542e-01 9.29761171e-01
1.78144407e+00 -7.92564154e-02 6.06980681e-01 5.30203521e-01
4.43396389e-01 7.51181483e-01 6.78892016e-01 2.91017234e-01
8.89125094e-02 4.06143576e-01 7.79537201e-01 -2.34423146e-01
2.52262652e-01 -2.27300107e-01 3.89718771e-01 -1.33133158e-01
-4.27894861e-01 1.77170664e-01 -1.13822925e+00 4.14503396e-01
-1.84105718e+00 -1.45341909e+00 -4.30059612e-01 2.66003418e+00
8.18243325e-01 -3.53969559e-02 2.72349566e-02 -5.23766913e-02
5.13884306e-01 3.90542120e-01 -7.60631502e-01 -1.68933809e-01
-4.36853588e-01 4.87490505e-01 1.58426332e+00 5.66561937e-01
-1.30858171e+00 9.84767556e-01 7.02889729e+00 5.26933432e-01
-1.10596359e+00 -1.64284393e-01 6.53857172e-01 -7.75317475e-02
-1.18001506e-01 1.22000478e-01 -6.78096712e-01 1.28446192e-01
1.37233400e+00 2.27878541e-01 6.82826936e-01 7.59974897e-01
4.86239761e-01 -3.01748455e-01 -7.54340708e-01 3.04211378e-01
-2.54691482e-01 -1.24112880e+00 -1.90787569e-01 2.44655803e-01
4.97432500e-01 7.35617578e-01 -2.04632670e-01 4.81701642e-01
8.40967119e-01 -1.39294970e+00 4.69545037e-01 7.66085684e-01
1.14590323e+00 -7.18601763e-01 8.53594720e-01 4.40262914e-01
-8.32486749e-01 -3.34124327e-01 -6.77412450e-01 -4.09044057e-01
-2.13847741e-01 5.38264394e-01 -8.54353547e-01 5.34741461e-01
6.78438663e-01 8.66251230e-01 -6.30289137e-01 7.02615857e-01
8.53894502e-02 1.04020798e+00 -6.55025482e-01 5.16869485e-01
9.76728275e-02 -5.80575883e-01 2.77385622e-01 8.82116437e-01
4.11612123e-01 -8.72885883e-02 -3.29070445e-03 1.02695394e+00
3.26071888e-01 -4.32461873e-02 -8.79817784e-01 6.02726126e-03
3.55452120e-01 1.15105045e+00 -8.82276744e-02 3.12046975e-01
-7.59797022e-02 2.84407675e-01 2.27683485e-01 2.04671144e-01
-3.78111601e-01 1.86870601e-02 5.25470257e-01 5.10937572e-01
2.09601149e-01 -2.94445038e-01 -5.31354845e-01 -7.40253687e-01
-4.62383151e-01 -6.83978975e-01 3.01830973e-02 -1.02764797e+00
-1.37707126e+00 4.57980067e-01 1.82756260e-02 -1.36392665e+00
-4.00549710e-01 -5.40257096e-01 -6.38108075e-01 1.52685678e+00
-1.72452343e+00 -1.76667237e+00 -3.96275312e-01 -1.82641000e-02
2.91844960e-02 -3.17722023e-01 1.12282038e+00 -3.58756423e-01
-2.64648676e-01 -1.04961954e-01 7.03056335e-01 -2.51444429e-01
5.58163762e-01 -1.36593783e+00 3.18898976e-01 7.23559439e-01
-2.31614470e-01 1.74505621e-01 8.30979347e-01 -8.66466284e-01
-1.03337049e+00 -1.67467749e+00 7.71498144e-01 -3.21189135e-01
6.02367997e-01 5.33083044e-02 -6.27046227e-01 1.08799624e+00
-3.80460739e-01 -3.77633721e-02 8.91220212e-01 4.39779997e-01
-2.18089167e-02 -7.22709119e-01 -1.23726308e+00 1.23022057e-01
9.06979382e-01 -1.48613781e-01 -3.58144164e-01 4.19554114e-01
8.84720981e-02 1.19212478e-01 -9.89224136e-01 8.14410388e-01
1.00974452e+00 -8.76866698e-01 1.01991069e+00 -8.27805996e-01
8.76535356e-01 -1.36708856e-01 -9.20262098e-01 -1.50006068e+00
-5.09719133e-01 1.65412843e-01 4.27196741e-01 1.15361822e+00
5.53483129e-01 -5.66137075e-01 8.34798038e-01 2.48053282e-01
-7.34212697e-02 -2.43016884e-01 -9.67607975e-01 -8.18132639e-01
4.47590947e-01 -3.83607566e-01 4.78833050e-01 8.52727413e-01
-7.73532867e-01 1.24081776e-01 -5.06396830e-01 2.62786835e-01
9.83968437e-01 2.87553698e-01 1.10404480e+00 -1.99753046e+00
-7.46482164e-02 -2.17419505e-01 -1.75980926e-01 -4.31284785e-01
4.05476779e-01 -9.01592791e-01 1.31048322e-01 -1.34793091e+00
-1.46512181e-01 -1.05350101e+00 -1.39711201e-01 9.24907088e-01
2.23826021e-01 3.12618494e-01 -1.47987276e-01 4.97432739e-01
6.68810248e-01 6.95507169e-01 8.33447635e-01 -1.58775389e-01
-2.98189163e-01 5.14268100e-01 -6.40210629e-01 4.69881415e-01
8.98512483e-01 -5.23345470e-01 3.92904431e-02 -1.57809615e-01
-1.26434481e-02 3.66285563e-01 7.81732678e-01 -1.04426527e+00
-5.47054648e-01 -7.80690789e-01 1.01447451e+00 -9.50384557e-01
4.45735604e-01 -1.10837376e+00 6.51616156e-01 4.68505114e-01
-4.52158034e-01 -3.56196016e-01 2.79913336e-01 5.07795811e-01
8.94134417e-02 -9.38960761e-02 7.89003015e-01 -2.08984613e-01
-4.78631258e-01 2.11489126e-01 -2.31358722e-01 -5.84237874e-01
5.35804868e-01 -4.73130196e-01 -2.55396932e-01 -3.49769771e-01
-8.90687883e-01 1.73236445e-01 4.95589286e-01 -5.88940531e-02
7.86954984e-02 -9.80954766e-01 -1.37090552e+00 4.04150009e-01
1.08691074e-01 -5.76217473e-02 -4.00852561e-02 8.05843413e-01
-6.40739322e-01 3.41913253e-01 -7.18018532e-01 -5.79377353e-01
-7.26357698e-01 4.85454546e-03 5.83546340e-01 -1.58786133e-01
-3.51619929e-01 6.08062208e-01 -4.25564945e-01 -1.00455570e+00
-4.44266945e-01 -9.72429588e-02 -1.25726596e-01 2.64452696e-01
-4.54796404e-02 9.77145433e-02 -6.50679320e-02 -9.41144049e-01
-1.69557467e-01 -7.00642914e-02 3.41818631e-01 -2.31069326e-02
1.99586952e+00 4.50677797e-02 -1.30927980e-01 6.62598968e-01
3.46514314e-01 -5.72347082e-02 -1.38143814e+00 -8.11818466e-02
-9.75784496e-04 -6.09414876e-01 7.92797983e-01 -1.33275759e+00
-9.90071058e-01 6.51125729e-01 8.47892225e-01 -8.39967430e-02
1.01652110e+00 -5.49660921e-01 -7.45394602e-02 5.60049236e-01
2.56657481e-01 -8.59496713e-01 -1.17002094e+00 1.52463362e-01
1.23487186e+00 -1.34680200e+00 4.02865738e-01 2.53619980e-02
-4.70185935e-01 1.12374926e+00 4.85734731e-01 -2.36957729e-01
6.20447099e-01 2.70388812e-01 1.24755902e-02 2.18154714e-01
-5.92108786e-01 -6.17617846e-01 -2.42531061e-01 1.16583991e+00
5.36620498e-01 5.42504549e-01 -1.33296639e-01 4.79715288e-01
-5.48176706e-01 2.55743265e-01 5.65666735e-01 7.05745757e-01
-5.70436180e-01 -1.11358345e+00 -4.98157769e-01 1.37457430e+00
-1.47684112e-01 -4.74362701e-01 -4.32259917e-01 9.43337917e-01
-7.57414922e-02 7.28863478e-01 1.03593841e-01 -2.59437770e-01
1.03608221e-02 2.15714499e-01 2.50155360e-01 -6.32880449e-01
-4.95811194e-01 -2.06669122e-01 3.51522654e-01 -2.26565242e-01
-7.39481091e-01 -8.72858822e-01 -2.56342798e-01 -5.96856594e-01
-5.62758327e-01 -1.32978275e-01 7.25376308e-01 6.73723996e-01
4.67667878e-02 4.15027857e-01 7.13326156e-01 -1.15265739e+00
-9.72009242e-01 -1.53506279e+00 -1.03898656e+00 -2.90871203e-01
3.86454284e-01 -7.43641078e-01 -3.63093287e-01 -1.05617277e-01] | [9.474952697753906, -1.5252399444580078] |
617502e6-1e62-4838-91c8-10b1264d9947 | generative-zero-shot-network-quantization | 2101.08430 | null | https://arxiv.org/abs/2101.08430v1 | https://arxiv.org/pdf/2101.08430v1.pdf | Generative Zero-shot Network Quantization | Convolutional neural networks are able to learn realistic image priors from numerous training samples in low-level image generation and restoration. We show that, for high-level image recognition tasks, we can further reconstruct "realistic" images of each category by leveraging intrinsic Batch Normalization (BN) statistics without any training data. Inspired by the popular VAE/GAN methods, we regard the zero-shot optimization process of synthetic images as generative modeling to match the distribution of BN statistics. The generated images serve as a calibration set for the following zero-shot network quantizations. Our method meets the needs for quantizing models based on sensitive information, \textit{e.g.,} due to privacy concerns, no data is available. Extensive experiments on benchmark datasets show that, with the help of generated data, our approach consistently outperforms existing data-free quantization methods. | ['Jian Cheng', 'Peisong Wang', 'Qinghao Hu', 'Xiangyu He'] | 2021-01-21 | null | null | null | null | ['data-free-quantization', 'data-free-quantization'] | ['computer-vision', 'methodology'] | [ 6.91923976e-01 4.22637314e-01 -2.33481199e-01 -5.26390016e-01
-1.14267159e+00 -2.24468082e-01 8.06950986e-01 -3.15554738e-01
-4.89570260e-01 8.22659552e-01 2.81882256e-01 4.41938788e-02
2.68899411e-01 -9.05853271e-01 -1.14803481e+00 -8.95494759e-01
5.38288295e-01 2.04498947e-01 -3.68049324e-01 2.57553142e-02
2.13133484e-01 3.12584907e-01 -1.63624620e+00 1.72696799e-01
6.94836199e-01 1.23380077e+00 1.20216906e-02 5.75113356e-01
2.80918609e-02 9.45251644e-01 -8.25259507e-01 -8.04914415e-01
6.34829819e-01 -6.56382918e-01 -4.56037343e-01 5.86244762e-01
6.38649940e-01 -8.08236718e-01 -6.62860692e-01 1.50921774e+00
5.08735120e-01 2.15523854e-01 7.02946961e-01 -1.51430428e+00
-1.07302868e+00 5.13515949e-01 -3.54345381e-01 -2.17640609e-01
-1.16503492e-01 5.03475666e-01 7.71647274e-01 -9.26162124e-01
7.65582025e-01 1.11120486e+00 3.46042365e-01 9.96108770e-01
-1.74027419e+00 -6.06062770e-01 -3.11375707e-01 -1.08924873e-01
-1.57394254e+00 -9.03565705e-01 7.58741856e-01 -3.29201818e-01
2.50121683e-01 2.10003003e-01 1.42769605e-01 1.59405005e+00
-3.81442579e-03 6.39661074e-01 1.14816499e+00 -3.83351594e-01
7.48618960e-01 3.21622521e-01 -3.31782758e-01 4.67393905e-01
3.16775382e-01 1.42547756e-01 -7.38530338e-01 -1.48801580e-01
9.47741747e-01 3.93550918e-02 -5.11478424e-01 -5.73516309e-01
-1.06904399e+00 9.38741505e-01 2.66617894e-01 -2.23705754e-01
-4.61114705e-01 4.88320321e-01 1.75989464e-01 1.94672376e-01
3.98613691e-01 1.70102149e-01 -1.68771408e-02 2.65145242e-01
-1.16709340e+00 1.02891140e-01 5.56404591e-01 1.48190856e+00
1.00272298e+00 5.44425130e-01 -6.81001306e-01 7.89019704e-01
5.67982532e-02 5.72773635e-01 6.02161109e-01 -1.28998435e+00
3.50010633e-01 -2.75926232e-01 2.92421550e-01 -6.67796552e-01
5.86747110e-01 -2.77774900e-01 -1.41132259e+00 3.21094662e-01
2.02006087e-01 -2.78887283e-02 -1.44262683e+00 1.92904365e+00
-1.20066702e-01 1.91073552e-01 2.68248707e-01 5.93057394e-01
7.69585907e-01 5.72108209e-01 -8.59752074e-02 -1.89134851e-01
1.06810486e+00 -6.14678383e-01 -9.07696128e-01 -1.45158544e-01
4.29419149e-03 -3.11068863e-01 1.32150793e+00 4.87660527e-01
-1.08563757e+00 -5.29450595e-01 -1.22880840e+00 -7.15106130e-02
-2.14764893e-01 9.37847942e-02 3.77079725e-01 1.05467284e+00
-1.21371102e+00 5.21892965e-01 -7.27848470e-01 -4.01326567e-02
1.07417393e+00 1.29888415e-01 -3.75751525e-01 -3.70380908e-01
-1.03354347e+00 3.86160612e-01 4.16734606e-01 -2.05799326e-01
-1.53351939e+00 -6.69604421e-01 -1.03466606e+00 6.85764477e-02
2.55948603e-01 -7.19275832e-01 1.02144587e+00 -1.11053860e+00
-1.50404632e+00 9.59830165e-01 -6.41191527e-02 -6.61955118e-01
7.52429426e-01 1.24711990e-01 -7.64260292e-02 2.26605564e-01
1.98940918e-01 1.14372015e+00 1.39583838e+00 -1.65495646e+00
-2.55756825e-01 -5.99635020e-02 -1.38697237e-01 -2.00678915e-01
-3.99104685e-01 -1.91058844e-01 -6.42359495e-01 -9.37789023e-01
-1.71539783e-01 -5.74516118e-01 -5.22966743e-01 2.93948829e-01
-7.13595092e-01 3.83073121e-01 4.84744728e-01 -5.78029990e-01
6.73013926e-01 -2.18967247e+00 -2.87857085e-01 2.24650562e-01
2.40794078e-01 -1.36430785e-02 -1.05053827e-01 -1.52118327e-02
-1.36753723e-01 3.19768816e-01 -4.34037149e-01 -8.36472332e-01
2.17070431e-01 4.84589726e-01 -6.18335903e-01 5.57666540e-01
3.57204467e-01 1.02435732e+00 -7.76887953e-01 -4.63804513e-01
2.79680282e-01 6.09229803e-01 -5.43223917e-01 4.45522010e-01
-2.76932389e-01 4.81592178e-01 7.05136657e-02 4.49752688e-01
7.96603322e-01 -4.79851067e-01 -1.12322181e-01 -3.01309139e-01
6.51653647e-01 -1.54357001e-01 -1.06226790e+00 1.87801361e+00
-3.07126105e-01 5.67554772e-01 5.14224824e-03 -8.98749292e-01
7.77394652e-01 3.39967728e-01 1.34347260e-01 -5.13532102e-01
4.08913374e-01 -2.08843127e-01 -4.53971297e-01 1.80408340e-02
3.22261512e-01 -1.63187757e-01 -4.35998961e-02 3.64924878e-01
5.93791127e-01 -3.16870183e-01 6.62216619e-02 4.32468295e-01
9.36903775e-01 -1.38385534e-01 9.70464423e-02 7.10139722e-02
1.63070373e-02 -3.89164388e-01 7.22403586e-01 1.21550357e+00
-6.73423409e-02 1.17278802e+00 2.96016425e-01 -1.43124044e-01
-1.51012850e+00 -1.17463326e+00 -8.53877291e-02 6.63227141e-01
-9.99465659e-02 -3.71913671e-01 -1.25423217e+00 -3.17082077e-01
-4.97926205e-01 9.96646762e-01 -7.66877115e-01 -2.29815140e-01
5.88417649e-02 -8.54738951e-01 5.90919554e-01 3.43492091e-01
6.74504042e-01 -9.02888477e-01 -3.65817189e-01 6.04139678e-02
-2.19492853e-01 -1.32815838e+00 -4.93940800e-01 2.33499289e-01
-5.53263128e-01 -4.48302001e-01 -9.33039844e-01 -5.42073965e-01
1.02124202e+00 4.29562554e-02 1.24421942e+00 -9.20315459e-02
-4.25040096e-01 4.78100896e-01 -1.06904939e-01 -3.98882568e-01
-6.32507384e-01 -4.28648174e-01 1.01304911e-02 4.51265007e-01
3.08551956e-02 -8.17960680e-01 -6.29064202e-01 -1.19593954e-02
-1.33931589e+00 1.28334299e-01 5.31881869e-01 1.18411231e+00
9.93240774e-01 2.73390889e-01 2.90786266e-01 -9.70341742e-01
4.48940128e-01 -3.97717386e-01 -7.00103819e-01 1.91836894e-01
-7.30837882e-01 2.29981810e-01 6.13086700e-01 -5.44125319e-01
-1.12191629e+00 1.42225781e-02 7.67707527e-02 -1.03014576e+00
-3.77398849e-01 -2.20537946e-01 -7.44780958e-01 -1.66607931e-01
8.66011798e-01 5.45296371e-01 -3.01517919e-02 -2.05257177e-01
8.64922822e-01 6.76503241e-01 1.04173362e+00 -6.13332391e-01
9.55678165e-01 6.36213422e-01 -3.60831432e-02 -6.11105621e-01
-8.21511149e-01 5.24114221e-02 -4.62920696e-01 8.83210003e-02
9.92599547e-01 -1.13592684e+00 -5.23963749e-01 6.51346087e-01
-9.74006951e-01 -2.76251256e-01 -7.81859994e-01 1.78296238e-01
-1.02010274e+00 2.95165122e-01 -6.01890683e-01 -7.90867448e-01
-3.67710382e-01 -1.22771585e+00 1.09080458e+00 4.22948264e-02
2.64811069e-01 -7.55971193e-01 -4.35793519e-01 4.16179687e-01
3.93445700e-01 4.24731910e-01 5.85234642e-01 -4.23918784e-01
-1.03041196e+00 -1.37702629e-01 -9.45689231e-02 9.79849935e-01
1.53407812e-01 -2.95009404e-01 -1.36156428e+00 -3.66648406e-01
3.69971335e-01 -7.27306426e-01 8.44279885e-01 3.63027900e-01
1.79077256e+00 -8.29587400e-01 2.04587817e-01 1.12953842e+00
1.53102505e+00 -1.57604620e-01 1.00766075e+00 -1.55704081e-01
8.34918976e-01 2.62033761e-01 3.81649733e-01 7.18171775e-01
7.93702304e-02 2.00724140e-01 6.14723027e-01 -1.12769946e-01
-2.56342769e-01 -4.89901423e-01 2.53954917e-01 2.68570960e-01
3.69002998e-01 -4.75531071e-01 -5.11778474e-01 6.01851344e-01
-1.55946624e+00 -9.71185207e-01 3.70920032e-01 2.29774928e+00
1.12838018e+00 1.25472829e-01 -3.97519410e-01 4.14526686e-02
7.83781886e-01 2.20117211e-01 -7.55780101e-01 2.84341406e-02
-3.39734286e-01 4.66703475e-01 1.06013155e+00 1.63405001e-01
-1.11626291e+00 8.54980707e-01 6.76648951e+00 1.24667203e+00
-8.83019447e-01 2.53794104e-01 1.30510104e+00 -2.42699772e-01
-5.81218064e-01 -1.17410988e-01 -7.35284328e-01 7.01468706e-01
1.08670604e+00 -1.74505278e-01 7.26096988e-01 8.84378135e-01
-1.10003345e-01 1.71502739e-01 -1.25068033e+00 1.31018388e+00
3.47449541e-01 -1.76074183e+00 4.64318782e-01 1.81701705e-01
1.09811556e+00 2.61481642e-03 5.33280492e-01 -6.35629566e-03
8.67705941e-01 -1.36086774e+00 7.91937351e-01 6.24402642e-01
1.41325629e+00 -6.61739528e-01 4.43172157e-01 4.63925749e-01
-5.00861406e-01 1.54290646e-01 -6.01908445e-01 4.97428894e-01
5.37200496e-02 6.83405995e-01 -7.64503777e-01 4.37142521e-01
5.99574625e-01 3.29218775e-01 -4.82146472e-01 6.63405240e-01
-3.13012153e-01 6.58132195e-01 -2.37723514e-01 4.52811658e-01
1.00873321e-01 -2.12880388e-01 1.25952333e-01 6.37817323e-01
3.56582314e-01 1.33678289e-02 -1.42502291e-02 1.41012967e+00
-6.98007047e-01 -2.87175365e-02 -6.43104315e-01 -4.19637598e-02
5.14079213e-01 1.02554977e+00 -5.60013771e-01 -5.40016770e-01
-2.74718404e-01 1.33226645e+00 6.79147094e-02 5.59857786e-01
-6.40638292e-01 -2.64794320e-01 6.10132694e-01 -1.09213017e-01
4.39456999e-01 8.97956491e-02 -4.72998679e-01 -1.34013093e+00
-7.51430839e-02 -1.10799301e+00 8.28432292e-02 -1.03066146e+00
-1.45409572e+00 4.84739274e-01 2.04408239e-03 -1.09732926e+00
-5.03574312e-01 -3.67814660e-01 -5.24436891e-01 9.85428989e-01
-1.27277792e+00 -1.21608329e+00 -3.32971871e-01 8.60878110e-01
4.48332340e-01 -3.66772175e-01 9.00238276e-01 4.13712077e-02
-3.68723810e-01 1.00186610e+00 4.00851309e-01 4.09446150e-01
5.33280253e-01 -1.14062190e+00 5.79616547e-01 1.12709260e+00
3.51283461e-01 5.10299623e-01 7.44572639e-01 -4.42467839e-01
-1.14304757e+00 -1.51609111e+00 1.48277536e-01 -3.87046635e-01
2.85208255e-01 -6.73102498e-01 -7.58103669e-01 8.22151601e-01
3.48110229e-01 4.65101242e-01 7.74466276e-01 -6.52753949e-01
-4.99503791e-01 -1.90825135e-01 -1.63588548e+00 5.91733992e-01
8.56674254e-01 -7.14776158e-01 -2.56624490e-01 4.89776075e-01
8.67390633e-01 -2.12918386e-01 -9.84687567e-01 3.15242447e-02
1.15114413e-02 -8.65350962e-01 1.04713631e+00 -4.60428715e-01
5.08639097e-01 -2.03954965e-01 -4.96145040e-01 -1.21575975e+00
-8.85865092e-02 -9.00373101e-01 -5.50308563e-02 1.47515070e+00
-5.75641580e-02 -2.73639381e-01 1.11075139e+00 1.01291263e+00
2.54590392e-01 -1.96907029e-01 -8.74036491e-01 -7.95689583e-01
-4.38832566e-02 -3.86894494e-01 8.71655166e-01 8.31382632e-01
-8.56028140e-01 2.31526382e-02 -9.09870327e-01 1.57407388e-01
1.37546849e+00 -2.21974850e-01 1.04907405e+00 -7.94536352e-01
-3.99987102e-01 -7.79632628e-02 -5.16934633e-01 -1.08772421e+00
3.17384392e-01 -6.70977294e-01 3.54735404e-01 -1.07787895e+00
1.95403308e-01 -2.25780621e-01 -3.91481161e-01 4.33776021e-01
-9.16091260e-03 7.04955697e-01 1.12135939e-01 6.87664524e-02
-3.07534784e-01 9.19837892e-01 1.11117256e+00 -4.09451336e-01
3.62497807e-01 -1.97876543e-01 -9.55884635e-01 5.00135541e-01
7.17556536e-01 -5.25012493e-01 -7.81983674e-01 -2.33199701e-01
-1.37605965e-01 -5.80465384e-02 4.86570954e-01 -9.74428892e-01
2.42969364e-01 -3.39677513e-01 6.41516507e-01 -4.65106070e-01
5.76861322e-01 -7.28372991e-01 3.76630008e-01 4.95697856e-02
-6.24131322e-01 -6.95351779e-01 -2.82712668e-01 7.50925839e-01
-1.79801658e-01 -3.40845406e-01 9.78837550e-01 -3.56275827e-01
-5.18578827e-01 8.23621392e-01 -1.74071655e-01 1.99706033e-01
8.02711248e-01 -2.15523824e-01 -1.52092665e-01 -8.54014695e-01
-5.97209752e-01 -9.54074413e-02 8.20867181e-01 1.68842316e-01
7.98101187e-01 -1.57195282e+00 -6.13429427e-01 4.21616077e-01
2.69959450e-01 3.02750051e-01 3.95122208e-02 6.04936741e-02
-1.62311897e-01 -3.77109945e-02 -3.44920337e-01 -4.18081194e-01
-7.73586929e-01 8.45806122e-01 2.48257980e-01 7.06704557e-02
-5.48151433e-01 9.91041958e-01 5.50333917e-01 3.90703281e-05
5.03064632e-01 -1.12935238e-01 4.19387370e-01 -2.90435493e-01
6.78932548e-01 -8.49138126e-02 1.14675537e-02 -4.70633626e-01
1.05435774e-01 -1.33149996e-01 -1.17295019e-01 -4.66586947e-01
1.32443285e+00 -2.34814987e-01 5.13831563e-02 3.82058680e-01
1.29251909e+00 -2.95294017e-01 -1.72105825e+00 -4.15263325e-01
-3.53902042e-01 -8.45186949e-01 1.84701532e-01 -4.82710660e-01
-1.36733878e+00 1.01612794e+00 6.56242430e-01 -1.83049425e-01
1.23001337e+00 -2.55385280e-01 6.82776570e-01 3.54424059e-01
5.29605985e-01 -1.09167910e+00 1.49042487e-01 2.68545039e-02
7.90862679e-01 -1.37344265e+00 -1.04453854e-01 -2.04896033e-01
-6.04654729e-01 5.78971207e-01 3.74162078e-01 -1.62456751e-01
5.16141951e-01 3.54390830e-01 3.29327844e-02 1.64199814e-01
-7.76661754e-01 4.28316109e-02 2.47026258e-03 1.11661088e+00
-2.23368317e-01 -1.01790920e-01 2.99165756e-01 5.12489736e-01
-2.19307557e-01 4.36642580e-02 7.42185354e-01 6.99391067e-01
-1.75729781e-01 -1.05532539e+00 -5.01302421e-01 6.42726123e-01
-5.56562483e-01 -4.68240201e-01 1.46943284e-02 2.74679005e-01
1.25761017e-01 8.79013419e-01 1.71206981e-01 -2.09934130e-01
-1.20060958e-01 -4.15361784e-02 3.85461271e-01 -4.84867871e-01
-9.09226835e-02 -6.86789304e-02 -2.63503551e-01 -5.85154831e-01
-1.89935744e-01 -5.37933409e-01 -7.95518458e-01 -3.62854034e-01
-3.41305099e-02 -1.93210065e-01 6.21513605e-01 5.42236507e-01
3.58837157e-01 4.47303683e-01 8.34548712e-01 -8.99448633e-01
-9.46442604e-01 -7.51171768e-01 -6.91679478e-01 8.03667367e-01
3.46039295e-01 -4.29514974e-01 -4.98980075e-01 6.47453964e-01] | [11.6868896484375, -0.31729620695114136] |
1764ea06-ee09-488d-90ba-67ecbfcbaa31 | unifying-semi-supervised-and-robust-learning | null | null | https://openreview.net/forum?id=r1gp1jRN_4 | https://openreview.net/pdf?id=r1gp1jRN_4 | Unifying semi-supervised and robust learning by mixup | Supervised deep learning methods require cleanly labeled large-scale datasets, but collecting such data is difficult and sometimes impossible. There exist two popular frameworks to alleviate this problem: semi-supervised learning and robust learning to label noise. Although these frameworks relax the restriction of supervised learning, they are studied independently. Hence, the training scheme that is suitable when only small cleanly-labeled data are available remains unknown. In this study, we consider learning from bi-quality data as a generalization of these studies, in which a small portion of data is cleanly labeled, and the rest is corrupt. Under this framework, we compare recent algorithms for semi-supervised and robust learning. The results suggest that semi-supervised learning outperforms robust learning with noisy labels. We also propose a training strategy for mixing mixup techniques to learn from such bi-quality data effectively. | ['Hideki Nakayama', 'Ryuichiro Hataya'] | 2019-03-24 | null | null | null | iclr-workshop-lld-2019 | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 1.14393331e-01 1.84905544e-01 -3.81443232e-01 -6.64234936e-01
-1.33922100e+00 -7.13066697e-01 3.56033027e-01 8.70280564e-02
-4.26658422e-01 9.71511662e-01 2.13592261e-01 -4.92247082e-02
-1.58738092e-01 -8.17593753e-01 -9.93760943e-01 -1.08342075e+00
4.28322583e-01 3.13162774e-01 -3.78736466e-01 1.95941925e-01
-2.30045140e-01 2.45580658e-01 -1.54927790e+00 3.29468310e-01
1.12573028e+00 8.33512306e-01 -1.07056551e-01 3.46787512e-01
-3.62535089e-01 1.01700497e+00 -6.96764827e-01 -4.29450870e-01
4.95307207e-01 -4.39048856e-01 -9.37749267e-01 2.32665151e-01
6.61683798e-01 -1.38813749e-01 -8.22984874e-02 1.44759250e+00
6.76511467e-01 -7.50484467e-02 4.60894287e-01 -1.45265853e+00
-6.64303362e-01 1.04084504e+00 -2.96174437e-01 -1.29041463e-01
-9.75310504e-02 -5.01654744e-02 8.82881701e-01 -7.61198878e-01
5.24653792e-01 1.03462553e+00 7.63549805e-01 7.78843701e-01
-1.24835002e+00 -9.94100332e-01 1.31916046e-01 3.64023075e-02
-1.20592892e+00 -7.43582785e-01 9.22286808e-01 -4.21635747e-01
2.09610984e-01 5.42679727e-02 1.54723361e-01 1.55907667e+00
-3.13866019e-01 9.71689641e-01 1.68155742e+00 -5.99727273e-01
6.41754806e-01 2.54156262e-01 4.88895684e-01 2.93115199e-01
6.74267888e-01 2.91449934e-01 -5.03582239e-01 -1.64395377e-01
1.95420325e-01 1.31894469e-01 -1.96621999e-01 -4.29829180e-01
-1.03813088e+00 6.35992050e-01 2.58729756e-01 3.18246365e-01
-1.44857392e-01 -1.84689537e-01 4.32765871e-01 5.58515787e-01
8.44871461e-01 5.67468464e-01 -7.15681016e-01 -2.18652356e-02
-1.22668147e+00 1.29026212e-02 8.17315936e-01 1.13194096e+00
9.50908422e-01 1.40070006e-01 -5.09102494e-02 8.16343963e-01
2.36359462e-01 7.52420008e-01 3.69282871e-01 -9.82196212e-01
6.55811846e-01 3.35278004e-01 2.92811185e-01 -6.54029310e-01
-4.28163558e-01 -4.64522153e-01 -1.16196358e+00 2.15889648e-01
5.92783570e-01 -3.81535232e-01 -1.30338550e+00 1.78720486e+00
3.50212395e-01 1.64504424e-01 1.39395773e-01 6.53364420e-01
9.62161958e-01 4.17587996e-01 -4.78838943e-02 -5.01299739e-01
7.88981974e-01 -1.20744324e+00 -1.21618724e+00 -2.22191945e-01
6.43305004e-01 -5.57860076e-01 1.01591003e+00 8.02983820e-01
-7.78698504e-01 -5.93967021e-01 -9.94225502e-01 2.16700342e-02
-5.81540465e-01 -6.85734749e-02 3.02156240e-01 9.03276145e-01
-9.04189348e-01 7.64315963e-01 -7.60080099e-01 -7.61344358e-02
6.70836806e-01 -1.38520887e-02 -5.29822350e-01 -4.39887643e-01
-1.13163400e+00 6.42217219e-01 4.17886823e-01 2.31791481e-01
-1.11517739e+00 -4.56702709e-01 -9.30077553e-01 -2.30350807e-01
5.80168068e-01 -2.35570773e-01 1.30312502e+00 -1.01831448e+00
-1.18364978e+00 9.14550602e-01 -1.77087814e-01 -3.47467959e-01
6.51981354e-01 -5.19957066e-01 -4.87059742e-01 2.45389808e-02
2.31237680e-01 1.91869631e-01 1.22262144e+00 -1.95816183e+00
-5.69975793e-01 -3.67334932e-01 -1.39000759e-01 -1.11999869e-01
-5.08267581e-01 -4.84911799e-02 -2.18548268e-01 -8.24921370e-01
1.15221754e-01 -8.81640255e-01 -1.83261797e-01 -2.63771027e-01
-7.01482058e-01 -9.16103423e-02 7.47370005e-01 -4.58021045e-01
1.00634432e+00 -2.13533044e+00 -4.01960537e-02 9.17784199e-02
5.05392015e-01 5.36334634e-01 -4.39575672e-01 2.87446231e-01
-2.45563731e-01 5.80122471e-01 -4.03673738e-01 -8.93952012e-01
6.32250011e-02 5.02962530e-01 -5.13185978e-01 5.82022607e-01
1.24114148e-01 8.04892063e-01 -1.28937113e+00 -4.83563811e-01
4.07680310e-02 2.23270431e-01 -1.09272391e-01 6.58640921e-01
8.44166204e-02 6.81302190e-01 -2.78564095e-01 8.83128881e-01
1.01602030e+00 -1.83699310e-01 7.17935562e-02 -2.72458196e-01
1.31789744e-01 2.60194749e-01 -1.56179488e+00 1.61222029e+00
-3.31847221e-01 1.99962705e-01 4.20338303e-01 -1.14074171e+00
8.17364454e-01 4.70522940e-01 4.37657118e-01 -5.22137642e-01
1.07064776e-01 1.79856434e-01 -4.74676251e-01 -7.79166639e-01
1.16590969e-01 -3.70527834e-01 3.29948217e-02 6.85896695e-01
2.54194975e-01 -2.15090752e-01 2.85299093e-01 8.99762735e-02
1.13916934e+00 1.70055494e-01 1.34361356e-01 -5.20858429e-02
5.09138008e-05 -2.15001643e-01 1.04293251e+00 1.22568178e+00
-5.70664942e-01 9.82800066e-01 2.34137177e-01 -3.87841910e-01
-9.60177422e-01 -9.46830750e-01 -2.85049736e-01 1.08249521e+00
1.53586403e-01 -5.21394789e-01 -8.66295040e-01 -1.26352012e+00
-5.00447303e-02 3.81083429e-01 -8.11401725e-01 -1.68890625e-01
-3.92272860e-01 -8.61259818e-01 5.93004942e-01 4.88637120e-01
3.32108617e-01 -1.08246255e+00 1.76935762e-01 -9.20426250e-02
-2.97176361e-01 -9.84957874e-01 -4.78644595e-02 7.55764782e-01
-8.43063891e-01 -1.21993172e+00 -2.69076765e-01 -6.92727566e-01
8.57834935e-01 4.65747178e-01 1.23121405e+00 5.99530526e-02
2.60543734e-01 1.19215481e-01 -8.67636561e-01 -4.13371116e-01
-6.54877663e-01 1.44799529e-02 4.69078958e-01 2.07612544e-01
3.47538859e-01 -4.84832764e-01 -1.02823533e-01 2.23939091e-01
-1.25653529e+00 -2.11954609e-01 5.35878003e-01 1.01025844e+00
8.98457348e-01 4.51507568e-01 7.81261206e-01 -1.43774569e+00
2.99485832e-01 -6.55895472e-01 -2.58765996e-01 2.94231623e-01
-8.49032104e-01 2.18524873e-01 1.08895195e+00 -4.41684902e-01
-1.05441391e+00 1.11912459e-01 -1.62170038e-01 -5.42161465e-01
-6.82641506e-01 2.69928098e-01 -7.40781963e-01 -2.53095329e-02
8.13417912e-01 -1.09379314e-01 -4.01343286e-01 -9.78045285e-01
6.48422778e-01 9.41875219e-01 3.79994243e-01 -7.19624460e-01
1.33144474e+00 6.46531641e-01 -3.45650971e-01 -4.65453476e-01
-1.59136748e+00 -5.69009840e-01 -9.01060045e-01 1.21906966e-01
6.31109715e-01 -1.01934671e+00 2.12418646e-01 7.45372534e-01
-6.60432220e-01 -4.23977643e-01 -7.43782818e-01 3.60691816e-01
-3.10433775e-01 4.36950147e-01 -5.62638342e-01 -7.07141638e-01
-3.24962646e-01 -1.00483692e+00 1.16085792e+00 1.82089612e-01
1.68698639e-01 -9.86623704e-01 2.26206630e-01 6.57917619e-01
3.41961980e-01 3.03284615e-01 3.19613904e-01 -1.04045153e+00
-2.02520192e-01 -2.87942499e-01 -3.28172371e-02 1.00171793e+00
6.25314593e-01 -1.27569437e-01 -1.40274894e+00 -5.66314936e-01
3.56780797e-01 -1.02841270e+00 8.88190329e-01 1.41706109e-01
1.24688935e+00 -4.05203432e-01 -3.84902731e-02 8.55383873e-01
1.25078249e+00 -2.48410508e-01 3.73402178e-01 3.61153215e-01
1.10479116e+00 5.88117719e-01 7.12307870e-01 1.44357026e-01
2.56306112e-01 1.59311920e-01 4.33692932e-01 -3.18185002e-01
-2.18543589e-01 -3.86112183e-01 2.31137037e-01 1.12836528e+00
2.34261066e-01 -6.86323345e-02 -7.60382831e-01 6.17525220e-01
-1.86184645e+00 -9.13480341e-01 -1.11151382e-01 2.12916493e+00
1.41383326e+00 1.31072953e-01 -1.01654187e-01 5.69883883e-01
6.73883319e-01 2.53836840e-01 -5.38534760e-01 2.02301696e-01
-3.56315166e-01 2.06468537e-01 4.23159003e-01 3.03317487e-01
-1.61888564e+00 6.90474927e-01 6.49232817e+00 9.74588454e-01
-9.73519623e-01 4.24319506e-01 7.37622082e-01 -1.43091783e-01
-3.09848577e-01 -1.00219719e-01 -7.70480037e-01 5.74657798e-01
9.27682400e-01 2.99459904e-01 3.05734277e-01 9.41852748e-01
2.65263051e-01 8.04906785e-02 -1.19865787e+00 1.06843960e+00
2.31572509e-01 -8.35206211e-01 -2.63486803e-01 -1.25777662e-01
1.36593914e+00 6.54441938e-02 -4.57915254e-02 3.79553944e-01
7.40797818e-01 -1.06055951e+00 7.14857161e-01 5.15087128e-01
7.61009574e-01 -5.71555018e-01 9.59764540e-01 7.21762896e-01
-7.83727586e-01 -1.77321792e-01 -2.35405147e-01 4.46435921e-02
6.46333769e-02 1.29026258e+00 -1.42608926e-01 7.05600321e-01
9.55271959e-01 9.25274134e-01 -8.94266546e-01 1.21739614e+00
-7.64106929e-01 1.20813203e+00 -2.02027649e-01 5.59659064e-01
-8.36798623e-02 -2.70193994e-01 2.44855747e-01 1.20815301e+00
1.29730478e-02 -2.00327814e-01 4.71973121e-01 6.41589999e-01
-4.02670741e-01 -1.40177891e-01 -6.52514696e-01 1.68361858e-01
6.92305326e-01 1.34521270e+00 -5.97727895e-01 -5.11172652e-01
-4.42260474e-01 7.09795654e-01 6.46456361e-01 4.32489008e-01
-5.14628768e-01 -1.69152752e-01 4.24408972e-01 -1.82452098e-01
2.81921495e-02 4.92543355e-02 -7.56992519e-01 -1.62202811e+00
1.32752836e-01 -1.25239193e+00 3.55554819e-01 -5.51553845e-01
-1.78931689e+00 4.15902525e-01 -1.21999212e-01 -1.32608783e+00
7.78085515e-02 -3.38905454e-01 -1.57995507e-01 5.15587270e-01
-1.88216805e+00 -1.30007422e+00 -3.36319178e-01 5.67389488e-01
4.07137036e-01 3.29216234e-02 7.53650904e-01 4.13373828e-01
-8.36655021e-01 7.07557857e-01 6.14038289e-01 3.96416515e-01
1.24550664e+00 -1.71987522e+00 1.21763252e-01 1.27669108e+00
5.08528352e-01 6.88739359e-01 5.46383500e-01 -7.21361399e-01
-1.23966372e+00 -1.51252520e+00 8.67928624e-01 -7.62569487e-01
5.70840418e-01 -6.14360213e-01 -1.21475542e+00 5.70118845e-01
1.30295470e-01 6.21250510e-01 9.17972326e-01 1.14881307e-01
-7.80441284e-01 -4.42079246e-01 -1.23271823e+00 1.24873020e-01
1.04060066e+00 -6.73169255e-01 -8.24638128e-01 6.27578378e-01
7.71752536e-01 -3.10933590e-01 -8.38121831e-01 4.23375458e-01
-4.00474183e-02 -8.48274231e-01 6.85344756e-01 -6.09518826e-01
2.70780236e-01 -5.16179085e-01 -2.32049953e-02 -1.70397520e+00
-2.48037338e-01 -6.54804826e-01 -4.21803236e-01 1.56116712e+00
2.98228592e-01 -2.55498230e-01 7.33847320e-01 6.04977906e-01
-1.63806587e-01 -2.26934254e-01 -8.32406223e-01 -9.93190348e-01
3.53491426e-01 -4.61381286e-01 5.55630565e-01 1.45003664e+00
-1.91164836e-01 1.10430785e-01 -7.48439193e-01 1.99898660e-01
1.01720977e+00 8.28431398e-02 7.81944036e-01 -1.23043787e+00
-8.56297910e-02 4.45605591e-02 2.91148961e-01 -5.88441014e-01
4.30658162e-01 -8.27923954e-01 4.20115352e-01 -1.52253795e+00
2.45426536e-01 -7.26746380e-01 -4.40420687e-01 1.00347459e+00
-5.32761335e-01 3.87979537e-01 -2.69189328e-02 3.83849651e-01
-8.87029052e-01 4.68635648e-01 1.03069985e+00 -3.24712545e-01
-6.80548102e-02 1.38991848e-01 -9.22239244e-01 6.94102287e-01
9.06006098e-01 -8.88668954e-01 -3.36061627e-01 -5.24262607e-01
2.51999021e-01 -6.80947602e-01 -4.14142897e-03 -9.24281299e-01
1.08904719e-01 -3.03801745e-01 1.24911666e-01 -3.08295637e-01
-1.65371656e-01 -1.05901027e+00 -8.18612874e-02 -2.30608396e-02
-6.44042552e-01 -4.82482821e-01 -3.07759643e-01 6.87671065e-01
-3.38101357e-01 -4.31525886e-01 9.84665096e-01 -1.67979971e-01
-4.06292528e-01 3.68478477e-01 -7.02247098e-02 4.97233182e-01
7.38114655e-01 2.23476902e-01 -5.21515548e-01 -2.76581168e-01
-8.89054954e-01 1.27934039e-01 5.10896742e-01 4.12949860e-01
3.77591163e-01 -1.50466490e+00 -7.11045802e-01 1.59533501e-01
2.34062150e-01 4.43991452e-01 1.05838224e-01 3.78164411e-01
-4.12316658e-02 -3.02296858e-02 1.63719937e-01 -2.37619489e-01
-9.62570190e-01 1.13129008e+00 1.44422397e-01 -3.18100184e-01
-5.16957402e-01 7.47127593e-01 -2.10818231e-01 -8.97251189e-01
6.15742743e-01 -2.92826712e-01 -2.46898830e-01 3.54140162e-01
7.94423282e-01 3.86901677e-01 4.82367754e-01 -7.42886722e-01
-2.02896576e-02 4.69984502e-01 7.00591132e-02 2.60114402e-01
1.34790337e+00 -2.96579659e-01 -3.05944443e-01 7.15200841e-01
1.14712787e+00 3.95112693e-01 -1.36600292e+00 -7.73838818e-01
1.00736216e-01 -4.41471756e-01 2.16648635e-02 -8.05514872e-01
-1.31234515e+00 8.17650378e-01 4.60540384e-01 4.39339578e-01
1.14097607e+00 -1.08034752e-01 6.08237267e-01 5.32363951e-01
4.00430292e-01 -1.39819479e+00 6.74683005e-02 4.96703714e-01
4.67986345e-01 -1.81048048e+00 1.73115321e-02 -2.99800396e-01
-4.66465712e-01 7.91323841e-01 7.28751540e-01 5.19931279e-02
7.72003233e-01 5.00552237e-01 5.44737458e-01 -3.34894806e-02
-4.49077994e-01 -4.15557742e-01 7.08449483e-02 9.62594151e-01
1.22757619e-02 3.53831910e-02 6.61514467e-03 8.83778036e-01
-1.31782293e-01 -1.05747625e-01 5.55077672e-01 1.20459676e+00
-3.85106504e-01 -1.23136544e+00 -7.55710363e-01 4.70005155e-01
-4.90454733e-01 -1.31337330e-01 -3.18672359e-01 3.24697912e-01
6.53049648e-01 1.52745020e+00 -4.92767721e-01 -4.70556289e-01
3.65152448e-01 2.50817060e-01 9.47147049e-03 -9.13142622e-01
-5.30502915e-01 1.05129220e-01 2.89550405e-02 -5.38857937e-01
-8.92342746e-01 -4.77769881e-01 -7.92742908e-01 -3.35539728e-02
-5.10332048e-01 2.88637519e-01 4.62717712e-01 1.15347362e+00
1.22953594e-01 2.91696668e-01 9.32624280e-01 -7.97641158e-01
-8.07973087e-01 -1.02237141e+00 -7.87037969e-01 8.97398055e-01
8.20700884e-01 -4.57322240e-01 -7.30665624e-01 4.08590317e-01] | [9.41451358795166, 3.92673397064209] |
e65ae6d9-cbcb-4faf-aff8-55a74e3e8351 | geometric-multi-model-fitting-with-a-convex | 1706.01553 | null | http://arxiv.org/abs/1706.01553v1 | http://arxiv.org/pdf/1706.01553v1.pdf | Geometric Multi-Model Fitting with a Convex Relaxation Algorithm | We propose a novel method to fit and segment multi-structural data via convex
relaxation. Unlike greedy methods --which maximise the number of inliers-- this
approach efficiently searches for a soft assignment of points to models by
minimising the energy of the overall classification. Our approach is similar to
state-of-the-art energy minimisation techniques which use a global energy.
However, we deal with the scaling factor (as the number of models increases) of
the original combinatorial problem by relaxing the solution. This relaxation
brings two advantages: first, by operating in the continuous domain we can
parallelize the calculations. Second, it allows for the use of different
metrics which results in a more general formulation.
We demonstrate the versatility of our technique on two different problems of
estimating structure from images: plane extraction from RGB-D data and
homography estimation from pairs of images. In both cases, we report accurate
results on publicly available datasets, in most of the cases outperforming the
state-of-the-art. | ['Pedro Pinies', 'Lina M. Paz', 'Paul Amayo', 'Paul Newman'] | 2017-06-05 | geometric-multi-model-fitting-with-a-convex-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Amayo_Geometric_Multi-Model_Fitting_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Amayo_Geometric_Multi-Model_Fitting_CVPR_2018_paper.pdf | cvpr-2018-6 | ['homography-estimation'] | ['computer-vision'] | [ 3.83536726e-01 1.30873412e-01 2.06633657e-02 -2.63831735e-01
-9.69045758e-01 -6.36670470e-01 3.08268607e-01 6.64491504e-02
-3.39698374e-01 4.46905822e-01 -1.30981624e-01 -5.28385229e-02
-3.52811754e-01 -6.67148590e-01 -7.35819221e-01 -6.49387479e-01
1.23786584e-01 7.19443619e-01 2.96462387e-01 -3.74025591e-02
5.67473352e-01 8.16203237e-01 -1.53062904e+00 -1.33772403e-01
5.96437991e-01 1.12794673e+00 3.26897986e-02 4.53510821e-01
1.89731702e-01 3.09154633e-02 -1.75865903e-01 -4.03878301e-01
6.36516869e-01 -2.84841388e-01 -9.10647035e-01 5.22845089e-01
6.66552842e-01 1.11877145e-02 4.68123294e-02 8.42862487e-01
3.29628766e-01 1.62797570e-01 4.82476205e-01 -1.11561489e+00
-7.75050418e-03 -2.13829875e-01 -8.80862236e-01 -3.26960832e-01
6.24522328e-01 -1.60146296e-01 1.10940695e+00 -7.73824632e-01
8.44919920e-01 1.01753008e+00 9.21259701e-01 6.31207302e-02
-1.61467946e+00 -4.26557511e-02 -4.11993377e-02 1.35967769e-02
-1.42894030e+00 -5.81725597e-01 8.82271588e-01 -3.58527571e-01
9.89985943e-01 6.06473446e-01 7.64485896e-01 3.23186606e-01
-1.49907872e-01 6.45562828e-01 1.16392708e+00 -6.40475094e-01
8.81630927e-02 -1.40166864e-01 -9.26999822e-02 7.72528172e-01
1.73216850e-01 1.35731936e-01 -3.40463102e-01 -3.48857701e-01
9.32344198e-01 -2.16108963e-01 -2.58061856e-01 -1.02410519e+00
-1.41537285e+00 7.59642065e-01 2.38067955e-01 1.58233851e-01
-3.92429292e-01 2.36305833e-01 1.09643362e-01 4.30324003e-02
3.49836558e-01 5.79785049e-01 -3.45063865e-01 -1.68529376e-01
-1.35711181e+00 2.76842237e-01 8.31574082e-01 8.07516634e-01
9.17395115e-01 -4.75742549e-01 6.04012132e-01 7.58305967e-01
4.31371242e-01 2.21515626e-01 1.34049669e-01 -1.12825632e+00
3.61551225e-01 5.96589267e-01 -6.07977435e-02 -1.32561672e+00
-5.36617696e-01 -2.52913028e-01 -5.93784630e-01 3.23956996e-01
3.77272338e-01 1.88724831e-01 -6.48440421e-01 1.40481985e+00
6.67530298e-01 -5.84631506e-03 -2.85002887e-01 8.33272457e-01
2.80674219e-01 3.32658947e-01 -6.25856102e-01 -3.99245113e-01
1.11857677e+00 -9.15759683e-01 -4.05866921e-01 -1.25266731e-01
3.95655364e-01 -1.08897114e+00 6.67484760e-01 6.96296513e-01
-1.32826579e+00 -3.80338192e-01 -1.12093544e+00 -2.48021558e-01
-8.49631131e-02 -3.74158472e-02 6.61456883e-01 4.71819550e-01
-1.12927198e+00 7.75606453e-01 -9.85645950e-01 -2.97961593e-01
7.84808248e-02 7.21480072e-01 -5.43109715e-01 9.96021107e-02
-3.21193874e-01 8.86521041e-01 3.10354292e-01 2.77943090e-02
-2.12310806e-01 -6.37650073e-01 -7.89754093e-01 -1.07644014e-01
4.93809253e-01 -6.47800922e-01 9.40344870e-01 -8.26177895e-01
-1.54145491e+00 1.17854166e+00 -2.87598014e-01 -3.90698947e-02
6.57875121e-01 -9.67895016e-02 1.98927134e-01 2.43033871e-01
-1.24333523e-01 6.57190382e-01 7.26232827e-01 -1.38041818e+00
-3.36331785e-01 -2.82558501e-01 -2.94103939e-02 2.04656631e-01
3.36680599e-02 -2.08523557e-01 -7.82816887e-01 -5.48987210e-01
7.11220980e-01 -1.14566410e+00 -4.33097959e-01 2.93149948e-01
-4.17487025e-01 5.76087199e-02 6.78226650e-01 -5.25071442e-01
1.11599350e+00 -2.21642590e+00 7.05082595e-01 6.05529845e-01
2.17825741e-01 -2.02806275e-02 1.16002865e-01 3.86778086e-01
-1.37947798e-01 -1.00380167e-01 -5.10523140e-01 -6.70146644e-01
3.57812201e-03 3.95995289e-01 6.37979573e-03 8.63561928e-01
-3.16341184e-02 6.29274070e-01 -6.14913523e-01 -6.21544898e-01
4.86956477e-01 4.21083421e-01 -6.56624377e-01 -1.56308878e-02
5.48422113e-02 3.05792183e-01 -2.25689217e-01 4.65367228e-01
8.50412428e-01 -1.40961125e-01 1.92232803e-01 -4.52581644e-01
-2.62521207e-01 1.25262290e-01 -1.80425322e+00 1.94970965e+00
-2.63702452e-01 5.38841903e-01 2.36131385e-01 -1.07544565e+00
7.95101285e-01 1.13898866e-01 1.01688695e+00 -3.66221488e-01
-8.11932515e-03 4.36983734e-01 -3.65144968e-01 -2.41668567e-01
5.05763412e-01 -2.13986397e-01 6.08221218e-02 4.11787599e-01
-1.48582786e-01 -5.05064428e-01 1.62060574e-01 -1.78965911e-01
8.21244836e-01 4.63780493e-01 5.12957573e-01 -3.27975690e-01
4.38766688e-01 6.04048967e-02 2.93323129e-01 3.01550776e-01
1.78041831e-01 7.68020868e-01 4.66772377e-01 -4.49530363e-01
-1.11475420e+00 -8.42274547e-01 -3.45385700e-01 4.38553959e-01
3.18086326e-01 -5.55437148e-01 -7.85747349e-01 -3.95999819e-01
1.46006614e-01 2.81610668e-01 -4.96222854e-01 2.91001439e-01
-8.17955792e-01 -7.14445055e-01 2.23646034e-02 2.75116563e-01
2.83180594e-01 -5.46571016e-01 -1.02499390e+00 1.93289295e-01
-2.40351796e-01 -1.14827514e+00 -3.64115685e-01 1.44478217e-01
-1.29121876e+00 -1.15069723e+00 -5.64641297e-01 -6.32707834e-01
8.47430170e-01 1.89005107e-01 1.19869351e+00 2.37938389e-01
-4.27330762e-01 5.23739159e-01 -3.20006758e-02 2.61937566e-02
-6.16344586e-02 1.57106072e-01 -1.50903493e-01 -8.54067504e-02
-8.70676991e-03 -5.27816057e-01 -4.95339513e-01 6.59179688e-01
-7.92217314e-01 1.50691465e-01 4.30826575e-01 5.42890489e-01
1.10911155e+00 3.63415815e-02 -2.27976162e-02 -5.66857457e-01
2.93460101e-01 -6.47464171e-02 -9.56132233e-01 3.04529935e-01
-5.32916248e-01 3.11057776e-01 2.55720317e-01 -2.88633764e-01
-5.45296848e-01 6.27238631e-01 -6.65487768e-03 -3.45253021e-01
-2.54258234e-02 2.87755042e-01 3.10273245e-02 -7.00619161e-01
3.51801485e-01 -4.42743078e-02 8.37930143e-02 -5.72473288e-01
3.92019004e-01 3.36884499e-01 5.91954708e-01 -5.60612857e-01
7.76995242e-01 6.96712911e-01 7.73867965e-01 -9.82999802e-01
-5.63253403e-01 -7.63863921e-01 -9.56098258e-01 -1.62910357e-01
7.73875654e-01 -2.93483466e-01 -7.67453611e-01 3.20119679e-01
-1.03363848e+00 9.01024789e-02 -3.01603675e-01 4.57244515e-01
-1.02803588e+00 7.44929492e-01 -1.81139424e-01 -7.35477924e-01
7.95074701e-02 -1.37818205e+00 1.28518164e+00 -1.70000121e-01
-2.04528138e-01 -1.09734106e+00 3.65742475e-01 2.83498347e-01
5.43551482e-02 7.57100701e-01 5.57124019e-01 -1.01851352e-01
-6.40235424e-01 -2.96462357e-01 -3.36435623e-02 1.81090787e-01
-6.23515248e-02 1.43308699e-01 -7.49916911e-01 -3.46839011e-01
2.36823082e-01 -8.70512202e-02 5.65885782e-01 5.10408342e-01
1.00076437e+00 -1.88436836e-01 -4.32019264e-01 8.83314371e-01
1.74949253e+00 -2.29456544e-01 8.19143116e-01 5.82790494e-01
7.90110648e-01 7.03090429e-01 7.28080690e-01 2.33342782e-01
3.29274744e-01 1.17423034e+00 5.98836124e-01 -4.63062882e-01
1.45504385e-01 1.17924392e-01 -8.82046372e-02 8.02116334e-01
-2.74583906e-01 1.65463090e-02 -8.67803037e-01 4.88680750e-01
-2.08400321e+00 -6.22665286e-01 -3.71535510e-01 2.59356594e+00
5.68271399e-01 -1.27332211e-01 4.03852612e-01 3.23924273e-01
3.83946121e-01 4.62168083e-03 -9.42329019e-02 -4.02085811e-01
1.31779209e-01 2.23668084e-01 8.24276090e-01 7.10489035e-01
-1.04927409e+00 5.15305459e-01 6.88987350e+00 5.97079933e-01
-9.33915138e-01 -1.46171957e-01 2.56846905e-01 -3.10148865e-01
-5.97653799e-02 2.42573634e-01 -6.19892240e-01 1.84871227e-01
6.03402734e-01 1.87514409e-01 6.02050543e-01 6.48785770e-01
-5.02606668e-02 -3.47439408e-01 -1.25570929e+00 1.15080786e+00
2.46251538e-01 -1.18301868e+00 -3.90252531e-01 4.72755849e-01
6.85544193e-01 1.11969712e-03 -1.78586990e-01 -5.23610413e-01
-2.53685296e-01 -8.06027710e-01 9.28696513e-01 6.88683093e-01
6.34157658e-01 -8.13718319e-01 4.59661365e-01 2.28942618e-01
-1.13416481e+00 2.79795915e-01 -2.91938692e-01 1.96767837e-01
4.92887348e-01 6.58425093e-01 -4.27645266e-01 8.34482253e-01
4.36089247e-01 4.83578265e-01 -4.22240555e-01 1.31327987e+00
4.93903123e-02 1.52335670e-02 -8.80678058e-01 3.27152818e-01
2.94090450e-01 -5.54092646e-01 6.39295280e-01 1.08566439e+00
2.32295007e-01 4.59530912e-02 3.26157600e-01 6.02545261e-01
1.53840840e-01 8.59575868e-02 -4.05447483e-01 2.97029555e-01
7.81024322e-02 1.30528998e+00 -9.82008934e-01 -6.04955480e-02
-3.29600155e-01 8.48465800e-01 2.77142644e-01 4.15007994e-02
-6.05468214e-01 -3.17837626e-01 2.94153154e-01 2.98197627e-01
5.23776412e-01 -6.05868399e-01 -4.74812090e-01 -9.58494186e-01
3.81049454e-01 -8.19024801e-01 2.81583875e-01 -5.97883284e-01
-9.90265906e-01 4.38345194e-01 3.06468606e-01 -1.19805682e+00
-3.60020518e-01 -6.83264911e-01 -1.94531828e-01 6.63310170e-01
-1.40548003e+00 -1.11625195e+00 -2.82616735e-01 3.64097327e-01
2.73621440e-01 4.46056575e-01 7.69703805e-01 2.58942544e-01
-2.65578091e-01 3.00094187e-01 1.56859830e-01 -3.70560944e-01
4.86220092e-01 -1.19953036e+00 3.80673945e-01 8.07228863e-01
2.45832965e-01 5.82266033e-01 7.70131588e-01 -3.93623769e-01
-1.38229036e+00 -3.28306079e-01 7.70778060e-01 -4.56472546e-01
5.24804056e-01 -2.93262362e-01 -7.35205591e-01 6.55719936e-01
-7.26078078e-02 -2.07093880e-02 5.76328397e-01 4.94461618e-02
1.04073919e-01 8.32958668e-02 -1.34620464e+00 2.36969054e-01
9.39616263e-01 -2.52729177e-01 -4.39870536e-01 4.34646755e-01
1.32916167e-01 -7.21499681e-01 -1.06046247e+00 3.11801046e-01
7.64526069e-01 -1.14349663e+00 1.19202077e+00 -7.80816749e-02
1.36774242e-01 -4.62863952e-01 -3.22335273e-01 -1.07508504e+00
-3.07827741e-01 -9.07412767e-01 -2.11633176e-01 9.38447297e-01
1.64782241e-01 -6.00478888e-01 7.57817507e-01 5.10708988e-01
-8.98676440e-02 -1.08698189e+00 -1.19358730e+00 -7.62542009e-01
-2.94504493e-01 -2.87912846e-01 6.02520764e-01 8.17733765e-01
-1.31272748e-01 4.91081476e-02 -2.84351617e-01 5.59702702e-02
9.83772337e-01 1.36360675e-01 8.35443616e-01 -1.40217555e+00
-5.98881245e-01 -4.08317834e-01 -7.78134286e-01 -1.22976077e+00
-1.12544261e-01 -6.16739392e-01 1.17601957e-02 -1.43136466e+00
-5.70004666e-03 -6.38674378e-01 1.82079330e-01 4.90246832e-01
1.63058922e-01 4.56139028e-01 2.23123372e-01 3.90588313e-01
-4.74107891e-01 1.05342254e-01 1.05048680e+00 2.89500356e-01
-2.38346040e-01 -1.27345979e-01 -3.56037229e-01 8.85759711e-01
4.38689232e-01 -5.54901242e-01 -1.58172950e-01 -4.91092324e-01
3.95108938e-01 -5.25657833e-02 4.43109989e-01 -1.00128996e+00
1.33035570e-01 -1.97148353e-01 2.15206161e-01 -8.19891870e-01
5.08832216e-01 -1.16610670e+00 5.87184370e-01 3.60907584e-01
3.09455348e-03 3.08813035e-01 1.33658409e-01 3.08231473e-01
5.89435687e-03 -5.89044631e-01 9.72287536e-01 -2.02027351e-01
-3.56334239e-01 1.06219590e-01 1.57562494e-01 -2.06999183e-01
1.17541122e+00 -7.16447830e-01 1.27432999e-04 -1.82590693e-01
-6.35203123e-01 7.18932375e-02 1.05696118e+00 4.98044901e-02
5.97569466e-01 -1.31931424e+00 -4.94938016e-01 3.57891887e-01
-7.76880309e-02 2.80261874e-01 -2.14500815e-01 1.18701434e+00
-8.74831200e-01 2.96696126e-01 -1.40146716e-02 -9.39079702e-01
-1.38995612e+00 5.63429892e-01 3.96729320e-01 -2.89661974e-01
-5.40792465e-01 5.53017676e-01 -1.17975706e-02 -3.74295264e-01
2.24663690e-02 -3.49682927e-01 2.70008951e-01 -2.62027280e-03
4.73618247e-02 7.84512639e-01 3.42617989e-01 -8.68083417e-01
-5.30617535e-01 1.13720596e+00 2.28934139e-01 -2.85016894e-01
1.59837306e+00 -7.10701868e-02 -4.29981023e-01 1.97284177e-01
1.36884725e+00 1.59609050e-01 -1.24280667e+00 1.08117744e-01
6.75034290e-03 -8.10042977e-01 2.59365916e-01 -4.52130377e-01
-1.00728011e+00 5.87178886e-01 5.06773949e-01 3.53388876e-01
1.12446511e+00 5.22078946e-02 6.60996437e-01 2.77126282e-01
4.95024264e-01 -1.14805305e+00 -1.48561925e-01 1.58462733e-01
9.13728595e-01 -9.56759989e-01 6.22115493e-01 -6.76341176e-01
-1.87323198e-01 1.33544242e+00 2.17435397e-02 -3.10345709e-01
4.23318237e-01 1.58612177e-01 -1.32029057e-01 -4.75923806e-01
-1.25537694e-01 1.33857746e-02 6.47686720e-01 4.50846493e-01
2.62160212e-01 -1.00680746e-01 -3.40595573e-01 -2.03991041e-01
-3.62689763e-01 -1.69258669e-01 3.99213046e-01 1.14303517e+00
-2.86822766e-01 -1.46466553e+00 -6.71341002e-01 3.04267436e-01
-2.54511178e-01 1.98409900e-01 -5.23825347e-01 8.71657968e-01
4.67006899e-02 8.74422371e-01 -1.40407807e-04 -7.04814419e-02
5.68123341e-01 -2.26725683e-01 8.74391854e-01 -3.97835612e-01
-4.56690460e-01 3.89333278e-01 1.30228214e-02 -8.52116823e-01
-8.63940537e-01 -1.01966405e+00 -9.56385672e-01 -3.20397466e-01
-7.67453849e-01 -1.36146061e-02 9.30862665e-01 8.50663722e-01
2.90764183e-01 1.63119107e-01 6.74733877e-01 -1.24332607e+00
-5.25894463e-01 -4.42485899e-01 -4.39485878e-01 3.56981277e-01
4.17777389e-01 -8.31092715e-01 -2.84396619e-01 2.93334182e-02] | [7.974825859069824, -2.4483113288879395] |
98b3f410-9319-4d9e-94f0-293657aa8981 | align-deep-features-for-oriented-object | 2008.09397 | null | https://arxiv.org/abs/2008.09397v3 | https://arxiv.org/pdf/2008.09397v3.pdf | Align Deep Features for Oriented Object Detection | The past decade has witnessed significant progress on detecting objects in aerial images that are often distributed with large scale variations and arbitrary orientations. However most of existing methods rely on heuristically defined anchors with different scales, angles and aspect ratios and usually suffer from severe misalignment between anchor boxes and axis-aligned convolutional features, which leads to the common inconsistency between the classification score and localization accuracy. To address this issue, we propose a Single-shot Alignment Network (S$^2$A-Net) consisting of two modules: a Feature Alignment Module (FAM) and an Oriented Detection Module (ODM). The FAM can generate high-quality anchors with an Anchor Refinement Network and adaptively align the convolutional features according to the anchor boxes with a novel Alignment Convolution. The ODM first adopts active rotating filters to encode the orientation information and then produces orientation-sensitive and orientation-invariant features to alleviate the inconsistency between classification score and localization accuracy. Besides, we further explore the approach to detect objects in large-size images, which leads to a better trade-off between speed and accuracy. Extensive experiments demonstrate that our method can achieve state-of-the-art performance on two commonly used aerial objects datasets (i.e., DOTA and HRSC2016) while keeping high efficiency. The code is available at https://github.com/csuhan/s2anet. | ['Gui-Song Xia', 'Jiaming Han', 'Jian Ding', 'Jie Li'] | 2020-08-21 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 8.85572284e-02 -4.95997965e-01 1.23971507e-01 -4.25111145e-01
-3.70553285e-01 -6.62967622e-01 2.94681549e-01 -1.52119562e-01
-3.50917548e-01 1.64089128e-01 -1.75856695e-01 -2.08372399e-02
-2.05440685e-01 -8.94802868e-01 -5.21633863e-01 -7.76493251e-01
-1.16397776e-01 -1.04253300e-01 7.14408159e-01 -3.04965019e-01
1.88847438e-01 5.54689646e-01 -1.52453315e+00 -4.66110231e-03
9.91049349e-01 1.28749323e+00 3.56548399e-01 4.88549083e-01
1.73803300e-01 2.42072359e-01 -5.56545675e-01 -2.72767693e-01
7.65216470e-01 -1.86704904e-01 -6.08181238e-01 1.83214396e-01
5.83832562e-01 -4.00935918e-01 -3.39104384e-01 1.24067891e+00
6.17805719e-01 -4.32956479e-02 4.18035150e-01 -1.31971872e+00
-7.41487265e-01 2.37485215e-01 -9.21761513e-01 3.43585312e-01
5.15540428e-02 3.51853430e-01 8.76955330e-01 -8.97012413e-01
2.36149773e-01 9.55506325e-01 8.95676494e-01 2.99860518e-02
-1.08525872e+00 -8.46084774e-01 3.48944485e-01 1.54385090e-01
-1.78879333e+00 -2.35864431e-01 6.38089359e-01 -4.14753765e-01
5.10066986e-01 2.30191141e-01 6.77782357e-01 6.47604585e-01
1.35479882e-01 5.31713903e-01 7.86535442e-01 -2.28270814e-01
4.68392260e-02 -3.05764169e-01 -1.84073284e-01 8.52753818e-01
4.45255756e-01 -3.51770222e-02 -2.15946093e-01 6.73069432e-02
1.01425505e+00 3.28504890e-01 -4.58419472e-01 -5.16344428e-01
-1.46832681e+00 6.48553789e-01 1.03097105e+00 8.82422999e-02
-3.88408929e-01 -7.15415031e-02 2.05265045e-01 6.14209361e-02
2.99215466e-01 5.92837989e-01 -5.77652812e-01 4.77613300e-01
-7.51938224e-01 2.30953187e-01 2.25750878e-01 1.10774851e+00
7.26105034e-01 -1.38405874e-01 -2.05695137e-01 7.77486145e-01
3.04203302e-01 7.18885064e-01 4.86820221e-01 -6.07559323e-01
4.08117682e-01 8.82509232e-01 3.64396274e-01 -1.44346023e+00
-5.65923154e-01 -6.14009559e-01 -8.71812701e-01 1.86300918e-01
3.68695706e-01 -1.74193501e-01 -8.93702567e-01 1.56452858e+00
6.44102514e-01 1.73535839e-01 -2.04898864e-01 1.25785506e+00
8.18073988e-01 4.75777954e-01 -1.57798901e-01 3.17192286e-01
1.58476233e+00 -1.03854382e+00 -3.73035699e-01 -4.70839709e-01
5.99371195e-01 -9.95325744e-01 9.43510354e-01 -1.21563645e-02
-6.77643955e-01 -8.17802072e-01 -1.33986521e+00 1.06281817e-01
-3.97019684e-01 8.62465739e-01 6.21312976e-01 2.28399947e-01
-8.15074205e-01 4.32632476e-01 -8.15634847e-01 -4.08569336e-01
5.26184320e-01 4.81285632e-01 -3.80107403e-01 1.23040564e-01
-9.26696241e-01 4.91048247e-01 4.40192580e-01 4.84599262e-01
-4.86413002e-01 -4.10983980e-01 -8.11461091e-01 -1.34684011e-01
4.10197735e-01 -5.73674381e-01 1.07073152e+00 -1.06618190e+00
-1.42290521e+00 5.62447429e-01 1.92083731e-01 -2.32465401e-01
3.17046970e-01 -3.56338739e-01 -4.50342208e-01 9.76867452e-02
3.90225232e-01 8.49627316e-01 8.83453548e-01 -7.82587409e-01
-1.13002205e+00 -5.00989795e-01 2.38624364e-01 3.19451183e-01
-4.61558878e-01 8.76159295e-02 -4.55166727e-01 -8.22414160e-01
5.28752506e-01 -9.42857325e-01 -1.94393948e-01 3.64272028e-01
-2.56023824e-01 -8.12477693e-02 8.88101161e-01 -4.06469464e-01
1.17712462e+00 -2.16895771e+00 1.25014916e-01 -7.18190894e-02
1.68486252e-01 4.44419712e-01 -3.31313938e-01 4.14386280e-02
-1.46921262e-01 -7.15161413e-02 -1.06145196e-01 8.85741040e-02
-3.40890974e-01 -2.87562191e-01 -3.55458528e-01 6.98415339e-01
5.19093037e-01 7.48479068e-01 -9.55700517e-01 -3.48184228e-01
2.73068100e-01 3.65309596e-01 -3.64007294e-01 3.32145721e-01
7.23747984e-02 3.30542952e-01 -6.13833010e-01 1.01335847e+00
8.87351036e-01 -3.79677951e-01 -9.01686847e-02 -6.33430660e-01
-5.04924119e-01 -5.65331094e-02 -1.47488701e+00 1.56575716e+00
-1.67612717e-01 5.43672025e-01 -9.19304937e-02 -7.62485445e-01
1.09913576e+00 2.41470952e-02 3.83749098e-01 -3.85824621e-01
2.62016058e-01 2.09392339e-01 7.10170940e-02 -3.48869860e-01
5.66273034e-01 4.35826302e-01 6.67859018e-02 -1.68359786e-01
1.04116246e-01 1.17841744e-04 1.31816432e-01 -2.55222857e-01
9.13423479e-01 2.70829707e-01 5.31942427e-01 -3.35270047e-01
3.59979063e-01 3.51850758e-03 7.18896449e-01 4.96561348e-01
-2.47078136e-01 7.80952275e-01 2.27484882e-01 -8.02124023e-01
-8.99523795e-01 -7.06852853e-01 -3.04697216e-01 7.84126461e-01
6.16076589e-01 -4.63819414e-01 -5.58252335e-01 -7.99139321e-01
-1.12207383e-01 -5.95365316e-02 -6.89504266e-01 -1.29428446e-01
-4.30597156e-01 -9.75698471e-01 4.22887504e-01 7.92499244e-01
1.06253815e+00 -6.90215230e-01 -9.88938034e-01 9.68078058e-03
-2.13726461e-01 -1.07799518e+00 -5.04667878e-01 1.95120558e-01
-7.70884812e-01 -1.06217515e+00 -7.00927436e-01 -6.99397564e-01
9.49432373e-01 8.88087749e-01 8.82652938e-01 1.63426265e-01
-4.29317087e-01 -1.45631224e-01 -5.51229656e-01 -4.52273726e-01
5.29046476e-01 3.67875755e-01 1.02634326e-01 1.05657965e-01
2.95667291e-01 -4.18407708e-01 -1.03689873e+00 8.55088651e-01
-9.38068867e-01 2.08292335e-01 7.84221053e-01 8.46762002e-01
5.69991529e-01 -4.37958762e-02 2.43034303e-01 -2.48770803e-01
-4.73328158e-02 -3.17711294e-01 -1.14818537e+00 1.41560301e-01
-4.30880457e-01 -1.28365397e-01 6.75047457e-01 -4.90490347e-01
-4.74489987e-01 4.67947423e-01 9.21770781e-02 -2.81624973e-01
-1.92653090e-01 2.17451632e-01 -3.02853912e-01 -4.02257651e-01
6.88465834e-01 5.98133425e-04 -3.03850979e-01 -1.16345041e-01
7.16427192e-02 6.91742897e-01 5.17759204e-01 -2.61340410e-01
1.18407154e+00 5.53897381e-01 -7.27112219e-02 -6.09204829e-01
-1.15357673e+00 -5.04143119e-01 -9.65060055e-01 -1.34549886e-01
6.93864942e-01 -1.18992674e+00 -3.17471504e-01 7.81466126e-01
-1.04480886e+00 -2.22139955e-01 -1.36849478e-01 5.64590096e-01
-1.81297496e-01 1.74585000e-01 -2.12911785e-01 -4.26768661e-01
-4.08664763e-01 -1.22166121e+00 1.30484939e+00 7.92442918e-01
1.08637482e-01 -4.33167160e-01 -1.59606591e-01 -1.43879414e-01
4.44131702e-01 3.22667778e-01 3.82767797e-01 -3.49663168e-01
-8.03259015e-01 -3.71006012e-01 -5.56576610e-01 2.00364500e-01
2.32322529e-01 2.23824129e-01 -8.33151698e-01 -4.26098019e-01
-4.46098059e-01 -2.38265023e-01 7.16231346e-01 4.05127436e-01
1.18449736e+00 -5.81535958e-02 -4.22421247e-01 1.11682129e+00
1.37425709e+00 1.13147855e-01 6.01865947e-01 5.67523122e-01
6.09381795e-01 1.71126708e-01 1.06926310e+00 5.17807961e-01
2.59590477e-01 1.00733387e+00 6.67015731e-01 -4.19378251e-01
6.47764653e-03 -7.11058378e-02 1.97690055e-01 4.31991905e-01
-2.92579979e-01 -1.24939181e-01 -7.91298449e-01 4.99449879e-01
-1.93720603e+00 -6.71895444e-01 -5.38428649e-02 2.20172882e+00
6.23746932e-01 5.24903201e-02 3.63031104e-02 -7.22078532e-02
8.20073962e-01 3.56172740e-01 -5.12447059e-01 5.03990173e-01
-1.21213064e-01 -1.37122542e-01 9.37173247e-01 4.56212424e-02
-1.70055664e+00 9.79886115e-01 5.37341547e+00 7.83727705e-01
-1.35317945e+00 -2.68064737e-01 4.21361744e-01 1.73433542e-01
5.26343405e-01 2.81771850e-02 -1.15662098e+00 6.44257426e-01
3.68271083e-01 2.28733435e-01 1.49708867e-01 1.14484143e+00
-3.87934744e-02 5.32613471e-02 -5.19906938e-01 8.81481349e-01
1.73377637e-02 -1.11582279e+00 -1.29948303e-01 -7.50282854e-02
6.13265336e-01 1.47891924e-01 3.03078108e-02 6.26533926e-02
1.30922750e-01 -7.21935570e-01 9.09055352e-01 3.12644809e-01
9.09808218e-01 -6.26524806e-01 1.06733441e+00 9.55009833e-02
-1.76623130e+00 -3.37586403e-01 -7.62856603e-01 -6.24907650e-02
-2.39507407e-01 4.21626329e-01 -4.91855890e-01 6.68877959e-01
1.13823020e+00 9.16992605e-01 -9.34249222e-01 1.38481653e+00
-3.08896005e-01 2.96876997e-01 -4.26356673e-01 1.00929715e-01
2.68605351e-01 -2.51271158e-01 4.25389290e-01 9.42151368e-01
6.18755043e-01 1.01915672e-01 3.75780731e-01 5.60271919e-01
-1.27375843e-02 2.66174730e-02 -4.89935011e-01 1.77306220e-01
7.52097666e-01 1.77483189e+00 -1.15209544e+00 -1.20293781e-01
-3.83365035e-01 1.09730327e+00 3.03564966e-01 1.17140189e-01
-1.12571776e+00 -7.18475223e-01 6.80220783e-01 1.88730791e-01
6.64059818e-01 -3.13772410e-01 6.65980726e-02 -1.27882385e+00
1.71657458e-01 -8.59049141e-01 2.15469703e-01 -7.40190506e-01
-1.14514101e+00 7.72935748e-01 -6.05109520e-02 -1.75697815e+00
3.92678350e-01 -8.74336243e-01 -6.15480602e-01 7.46981502e-01
-1.55359757e+00 -1.23829281e+00 -1.05737257e+00 3.27151805e-01
5.71506321e-01 -1.62156731e-01 7.42954075e-01 3.37602705e-01
-7.92465866e-01 4.67586666e-01 -1.91436838e-02 5.27909696e-01
9.01963353e-01 -1.06309617e+00 5.51053762e-01 1.04940379e+00
9.34429467e-02 5.13930857e-01 3.70469481e-01 -3.93815458e-01
-1.16987383e+00 -1.46188962e+00 3.70748937e-01 -4.29472655e-01
4.66111451e-01 -2.83459872e-01 -6.79580033e-01 4.97456729e-01
-1.33281142e-01 5.81317842e-01 4.49722081e-01 -2.18548432e-01
-3.14935446e-01 -4.13120747e-01 -7.29369521e-01 4.95205402e-01
1.16816425e+00 -5.92396297e-02 -2.12990284e-01 3.63038808e-01
5.96213102e-01 -6.70638084e-01 -7.07204998e-01 7.45272040e-01
6.01219177e-01 -9.92547274e-01 1.07779229e+00 -2.34259397e-01
3.66440803e-01 -1.03089976e+00 -2.23927572e-01 -1.25696790e+00
-7.71444857e-01 -2.63176084e-01 1.97611108e-01 1.08512294e+00
1.95320994e-01 -6.44702852e-01 3.37401509e-01 1.30646983e-02
-8.63052309e-02 -1.01852357e+00 -6.52693570e-01 -7.02619076e-01
-5.12235284e-01 5.43920808e-02 9.22850907e-01 7.98863471e-01
-5.78585207e-01 3.03909510e-01 -1.57746464e-01 7.65066981e-01
3.26109886e-01 4.75638062e-01 8.78611565e-01 -1.25337219e+00
-3.63886394e-02 -3.04852754e-01 -7.43444502e-01 -1.33975923e+00
-3.28885227e-01 -5.05761504e-01 2.39772856e-01 -1.26359737e+00
-6.17007688e-02 -5.57557881e-01 -3.15855801e-01 6.82068110e-01
-3.99118304e-01 6.00913167e-01 6.67061210e-02 3.65538150e-01
-8.18717122e-01 5.52985072e-01 9.56364274e-01 3.87209430e-02
-1.95222110e-01 3.76711078e-02 -6.64307475e-01 1.03856349e+00
9.31671143e-01 -4.22744095e-01 -1.75699502e-01 -7.19307065e-01
9.79870707e-02 -4.49244499e-01 6.94382012e-01 -1.42417741e+00
2.21623167e-01 -1.51233882e-01 9.39804852e-01 -6.17423713e-01
-2.59415004e-02 -8.55133295e-01 -5.44967465e-02 3.94046754e-01
-3.87961678e-02 4.02396232e-01 1.10220976e-01 4.66393471e-01
-1.82577953e-01 -9.50108394e-02 9.78966951e-01 -2.44209561e-02
-8.24544311e-01 5.19477010e-01 1.19879827e-01 -2.65964568e-01
1.11690915e+00 -2.11066723e-01 -4.99516994e-01 1.45425722e-01
-3.36020589e-02 3.48015696e-01 6.91140711e-01 4.58732486e-01
6.02728426e-01 -1.34299755e+00 -6.55443490e-01 4.91889745e-01
3.42083901e-01 4.54099506e-01 1.86774462e-01 1.02576327e+00
-6.91229939e-01 2.59865224e-01 -3.38252485e-01 -8.27993691e-01
-1.17012084e+00 1.98609114e-01 5.34400582e-01 -6.13723211e-02
-5.70046663e-01 1.00055766e+00 1.81023523e-01 -3.90601158e-01
1.39810532e-01 -4.26211387e-01 -2.27434024e-01 1.32051542e-01
6.25467837e-01 1.13035470e-01 -1.76696815e-02 -6.39012516e-01
-4.65325058e-01 9.01059031e-01 1.11396685e-02 4.85565662e-01
1.27073562e+00 -1.01971678e-01 8.15301016e-02 7.35847726e-02
7.79994428e-01 -5.95201515e-02 -1.61185944e+00 -2.74248213e-01
-2.18709558e-01 -9.01036143e-01 1.03070639e-01 -4.78006542e-01
-1.33301604e+00 6.09418869e-01 9.95460093e-01 1.87677443e-01
1.25922120e+00 -2.89170474e-01 6.66299105e-01 4.13851023e-01
3.06197554e-01 -9.29670751e-01 1.40270442e-01 3.18412691e-01
8.49831402e-01 -1.37763441e+00 2.78197467e-01 -4.41911310e-01
-4.55435425e-01 1.25495207e+00 9.99319375e-01 -3.78654897e-01
4.23173726e-01 3.06246221e-01 1.97315052e-01 -1.29783556e-01
-1.98895290e-01 -3.43133837e-01 4.39720184e-01 5.81145525e-01
3.31462830e-01 4.55567893e-03 -6.65792525e-02 6.50434434e-01
-8.50200951e-02 -1.63223967e-01 1.05582438e-01 9.04468179e-01
-4.89768833e-01 -7.36926317e-01 -4.51432288e-01 2.38415033e-01
-2.69021571e-01 -9.35545936e-02 -2.52466738e-01 8.61245513e-01
5.25550187e-01 6.51717305e-01 1.56112000e-01 -7.07808495e-01
4.25242692e-01 -3.82601231e-01 1.68749630e-01 -4.92576063e-01
-3.42943877e-01 1.06730992e-02 -4.12667245e-01 -6.85849130e-01
-5.74955285e-01 -4.56995726e-01 -9.73143339e-01 4.29887511e-02
-8.92725110e-01 -1.29212305e-01 3.63329351e-01 5.90108871e-01
6.36301458e-01 6.11156285e-01 7.66995728e-01 -9.52624440e-01
-6.06947124e-01 -1.01331162e+00 -4.51718062e-01 1.47533238e-01
2.10422322e-01 -8.53359878e-01 -2.50851721e-01 2.91359965e-02] | [8.728741645812988, -0.8277407884597778] |
bb572170-b1a1-4c93-ad4d-2bc0db4c6469 | supporting-medical-relation-extraction-via | 2208.13472 | null | https://arxiv.org/abs/2208.13472v1 | https://arxiv.org/pdf/2208.13472v1.pdf | Supporting Medical Relation Extraction via Causality-Pruned Semantic Dependency Forest | Medical Relation Extraction (MRE) task aims to extract relations between entities in medical texts. Traditional relation extraction methods achieve impressive success by exploring the syntactic information, e.g., dependency tree. However, the quality of the 1-best dependency tree for medical texts produced by an out-of-domain parser is relatively limited so that the performance of medical relation extraction method may degenerate. To this end, we propose a method to jointly model semantic and syntactic information from medical texts based on causal explanation theory. We generate dependency forests consisting of the semantic-embedded 1-best dependency tree. Then, a task-specific causal explainer is adopted to prune the dependency forests, which are further fed into a designed graph convolutional network to learn the corresponding representation for downstream task. Empirically, the various comparisons on benchmark medical datasets demonstrate the effectiveness of our model. | ['Xiaohui Hu', 'Chengbo Jiao', 'Zheng Lian', 'Jiangmeng Li', 'Yifan Jin'] | 2022-08-29 | null | https://aclanthology.org/2022.coling-1.216 | https://aclanthology.org/2022.coling-1.216.pdf | coling-2022-10 | ['medical-relation-extraction'] | ['medical'] | [ 3.98500413e-01 7.91806936e-01 -4.94768590e-01 -5.18838286e-01
-4.94384497e-01 -4.00381684e-02 2.29088545e-01 3.28477234e-01
-7.06448480e-02 8.00391555e-01 6.63449645e-01 -6.73755348e-01
-1.51455864e-01 -8.54054272e-01 -4.72789496e-01 -4.65197891e-01
-1.58265859e-01 3.20869565e-01 4.73380461e-02 -1.44832149e-01
-1.69687226e-01 1.21472634e-01 -6.72942579e-01 4.80468839e-01
1.07492948e+00 4.93247718e-01 1.37894034e-01 4.72518563e-01
-4.83344495e-01 1.02514803e+00 -4.01834458e-01 -5.07943153e-01
-4.04591233e-01 -8.59111369e-01 -1.19400525e+00 -8.50100294e-02
-7.15840518e-01 -1.50185898e-01 -4.69025850e-01 9.53228235e-01
1.23624802e-01 -3.68683130e-01 5.89442313e-01 -9.85743761e-01
-6.45469368e-01 1.18449438e+00 -3.94699633e-01 3.79837573e-01
3.14021975e-01 -1.12659983e-01 1.33463132e+00 -4.33425635e-01
8.66818964e-01 1.15133965e+00 3.33287299e-01 7.17637777e-01
-9.61970031e-01 -6.82414174e-01 3.95815551e-01 1.63250372e-01
-1.03977048e+00 -1.58369750e-01 6.92165434e-01 -1.64956346e-01
1.02079475e+00 1.03455685e-01 5.81077874e-01 1.10958099e+00
6.31634712e-01 7.22972453e-01 8.51138830e-01 -2.51714259e-01
-2.38031492e-01 -2.45861828e-01 3.05126607e-01 1.08173776e+00
3.55346203e-01 -1.49031267e-01 -4.62597549e-01 -2.52823651e-01
5.92991352e-01 -3.53430249e-02 -2.75043756e-01 2.30999619e-01
-1.02474642e+00 9.81209159e-01 7.88722456e-01 2.85165042e-01
-5.51634669e-01 -6.84029534e-02 3.59911144e-01 -2.40650959e-02
5.39779663e-01 2.98064262e-01 -6.89033687e-01 4.32128102e-01
-5.34845650e-01 -1.72628284e-01 6.93417013e-01 9.00054395e-01
3.42550844e-01 -4.75884020e-01 -1.96835414e-01 5.90839207e-01
6.83624625e-01 -1.76958703e-02 3.27965260e-01 -1.44856542e-01
5.84497333e-01 1.01878548e+00 -4.62196529e-01 -9.50099707e-01
-7.58734941e-01 -4.94281381e-01 -9.71945405e-01 -5.66500902e-01
2.14327276e-02 -4.02293622e-01 -1.08914936e+00 1.67009819e+00
5.95902026e-01 1.81151435e-01 5.68985581e-01 8.94253016e-01
1.32659042e+00 3.76613051e-01 7.60318816e-01 -4.24460351e-01
1.76796222e+00 -9.55259204e-01 -1.13209581e+00 -3.92777920e-01
8.66249442e-01 -6.21759772e-01 4.98723358e-01 2.86738519e-02
-5.10128498e-01 6.96734861e-02 -9.11923230e-01 -8.96246359e-02
-3.25006321e-02 -1.03343219e-01 9.97023284e-01 6.50985315e-02
-3.59333336e-01 5.02792478e-01 -1.04884255e+00 -1.41950369e-01
7.91238844e-01 3.45097601e-01 -5.17747641e-01 -1.18072741e-01
-1.63046443e+00 8.96511853e-01 6.59010887e-01 2.89101303e-01
-5.40656626e-01 -4.74671662e-01 -1.04118395e+00 1.02191523e-01
4.00612235e-01 -1.19703889e+00 1.02696729e+00 -5.99676430e-01
-1.01951170e+00 7.52748370e-01 -3.00076455e-01 -3.99138421e-01
1.25965849e-02 -2.42751718e-01 -5.28327525e-01 2.21392959e-01
3.23495299e-01 2.98038542e-01 3.94561231e-01 -9.11423326e-01
-6.66503906e-01 -4.60104376e-01 -3.18674035e-02 1.91830441e-01
-5.63873760e-02 9.00870413e-02 -4.56225425e-01 -5.35112679e-01
3.10378850e-01 -6.79714441e-01 -8.57808471e-01 -7.70129740e-01
-1.14759564e+00 -4.43618804e-01 4.41484660e-01 -8.29384625e-01
1.53712058e+00 -1.89237833e+00 6.29245266e-02 9.68352705e-02
5.96640050e-01 -1.24494441e-01 7.83453509e-02 3.03419143e-01
-4.49726641e-01 3.89230371e-01 -2.80525923e-01 7.08400458e-02
-5.85013568e-01 4.62094247e-01 -1.42232329e-01 1.52674213e-01
7.30914772e-01 1.11116827e+00 -1.15765572e+00 -1.08433390e+00
-2.43083045e-01 3.38174075e-01 -4.83335197e-01 3.56129318e-01
-3.17228258e-01 7.48972952e-01 -1.20781636e+00 5.64420581e-01
3.62656474e-01 -6.33858263e-01 6.74726784e-01 -1.80492327e-01
2.91256756e-01 8.91193807e-01 -4.83849406e-01 1.69770777e+00
-3.48529071e-01 4.20839936e-02 -3.40880096e-01 -8.99284005e-01
8.67473364e-01 5.17016768e-01 5.70875227e-01 -2.73005038e-01
1.42372623e-01 2.52465010e-02 3.99768949e-01 -8.15667987e-01
-1.23700544e-01 -6.20386302e-01 -1.45507708e-01 1.71763867e-01
-7.28751905e-03 3.59491915e-01 -9.46124122e-02 3.83730590e-01
1.54623699e+00 -3.16606695e-03 7.62027740e-01 -1.26468167e-01
5.02309024e-01 3.12209874e-01 9.09547865e-01 1.56793833e-01
3.68471928e-02 3.72907192e-01 1.03991115e+00 -4.47952092e-01
-4.98974115e-01 -7.69742489e-01 -2.71246791e-01 6.47428930e-01
4.80799451e-02 -6.96721435e-01 -7.22017527e-01 -1.40236652e+00
-3.14384013e-01 7.21866608e-01 -7.99343348e-01 -2.14658722e-01
-7.09353924e-01 -1.15111673e+00 4.38720077e-01 5.15502751e-01
2.37605378e-01 -1.29464746e+00 -4.69818115e-01 4.04389143e-01
-5.30568898e-01 -1.28582513e+00 -2.79487729e-01 3.02587569e-01
-1.15601587e+00 -1.41214633e+00 -1.75699778e-02 -6.26680970e-01
8.75098288e-01 -1.57187045e-01 1.15539336e+00 4.68163013e-01
-1.01691857e-01 -4.20787096e-01 -4.43079203e-01 -3.99008006e-01
-5.11090815e-01 2.90929317e-01 -4.69130665e-01 -3.02641571e-01
6.85227275e-01 -4.49955106e-01 -8.01517725e-01 1.59670711e-02
-6.90967321e-01 5.06769598e-01 8.82795274e-01 1.00155568e+00
7.39631414e-01 1.77920222e-01 5.91960549e-01 -1.55575275e+00
5.14872670e-01 -7.07023561e-01 1.66816078e-02 1.89527601e-01
-8.07279289e-01 4.02831048e-01 5.58218837e-01 -8.83720964e-02
-1.30370593e+00 2.86823660e-01 -4.42729920e-01 1.22252256e-01
-1.39079005e-01 9.29862797e-01 -3.43585759e-01 7.18642592e-01
4.33926821e-01 1.83942076e-02 -3.99235219e-01 -2.61761636e-01
4.10190046e-01 5.46923161e-01 4.21920329e-01 -1.79965004e-01
5.07547259e-01 1.41345069e-01 1.62213787e-01 -1.81215256e-01
-1.51055372e+00 -1.87226892e-01 -5.48955619e-01 3.19216043e-01
1.32586896e+00 -7.11132646e-01 -4.59481716e-01 -1.52346462e-01
-1.36321080e+00 5.59113175e-02 4.89685684e-02 6.66818202e-01
-2.08807752e-01 1.76451672e-02 -8.45441759e-01 -5.88867545e-01
-5.99047840e-01 -1.02104521e+00 1.05759561e+00 2.72972703e-01
-4.24406171e-01 -9.08600390e-01 6.61426857e-02 4.11981612e-01
-3.36239934e-01 5.07332683e-01 1.54405391e+00 -9.01360631e-01
-4.40843374e-01 -2.84285359e-02 -3.80930573e-01 -3.34304780e-01
3.84216130e-01 -1.02960609e-01 -7.40307570e-01 4.10371989e-01
-1.88638285e-01 1.24271415e-01 1.01371539e+00 4.34482753e-01
1.06794965e+00 -5.82003415e-01 -1.00617552e+00 5.55651128e-01
1.14310038e+00 1.03812143e-01 5.46168506e-01 1.63986653e-01
9.94548202e-01 8.33991408e-01 7.05981970e-01 1.03602186e-01
6.32456839e-01 2.47951642e-01 4.20959204e-01 -3.10347021e-01
-1.00837156e-01 -6.12519622e-01 -2.23041758e-01 8.56470883e-01
-1.73454419e-01 -1.03296369e-01 -1.04892027e+00 4.19545293e-01
-1.89985478e+00 -4.40410912e-01 -5.80376446e-01 1.48077917e+00
1.22800732e+00 3.70442420e-01 -1.85029775e-01 -1.00907922e-01
5.69301665e-01 -7.11579770e-02 -4.20797139e-01 -1.89310387e-01
1.48794264e-01 2.29659095e-01 2.35752240e-01 3.14244568e-01
-1.09513569e+00 1.18366194e+00 5.77039003e+00 4.33735967e-01
-7.04492986e-01 1.48696661e-01 6.40038490e-01 4.31280255e-01
-5.25313556e-01 2.91126490e-01 -7.43187070e-01 2.61263102e-01
8.47120762e-01 -1.57170206e-01 -2.22876534e-01 7.48716950e-01
3.18611681e-01 1.29798353e-01 -1.04644525e+00 4.87087876e-01
-4.07652259e-01 -1.34209585e+00 7.12286681e-02 1.91417113e-01
3.80901963e-01 -2.39381105e-01 -4.16050583e-01 7.50519782e-02
5.86132824e-01 -1.28819358e+00 2.51042340e-02 3.85942429e-01
7.40946293e-01 -7.73912847e-01 1.01452029e+00 3.03220510e-01
-1.23695743e+00 -4.43451405e-02 -3.18467706e-01 1.30304441e-01
3.06758970e-01 9.03802752e-01 -1.18041217e+00 9.57547665e-01
5.15015781e-01 7.37663090e-01 -3.87548834e-01 4.28595215e-01
-1.06500256e+00 9.40182686e-01 8.55350792e-02 5.84203377e-02
8.99087861e-02 8.95441622e-02 2.10958257e-01 1.41011024e+00
-3.98560911e-02 7.10965693e-01 -1.98826455e-02 7.40239441e-01
-2.87462115e-01 3.56122226e-01 -6.01054072e-01 -8.63133371e-02
2.29849175e-01 1.55138445e+00 -8.88771653e-01 -2.26514056e-01
-4.53313708e-01 8.89603734e-01 6.67113304e-01 4.98549007e-02
-9.43697751e-01 -1.74066469e-01 3.92737865e-01 -2.48044059e-02
7.14044794e-02 2.33127147e-01 -6.22400880e-01 -1.09521294e+00
-2.25487247e-01 -6.99738503e-01 9.41158831e-01 -5.07465780e-01
-1.29677844e+00 9.25296545e-01 -1.51661500e-01 -8.51962328e-01
-2.93101639e-01 -2.90114999e-01 -6.74785078e-01 8.88194442e-01
-1.48816180e+00 -1.29579091e+00 -1.12426557e-01 3.89408231e-01
5.25644720e-01 1.26287699e-01 1.14002907e+00 3.70891243e-02
-9.37392592e-01 3.66113037e-01 -7.39143193e-01 5.06353438e-01
3.62260163e-01 -1.31374729e+00 4.74342346e-01 7.40570784e-01
2.49439478e-02 8.74457598e-01 5.52563608e-01 -1.16203249e+00
-9.90953147e-01 -1.23719680e+00 1.54108465e+00 -4.02883232e-01
6.95051551e-01 -3.21041718e-02 -1.10725117e+00 8.37290049e-01
1.43302396e-01 4.12422381e-02 9.74155188e-01 4.97887671e-01
-3.85889620e-01 3.35173666e-01 -8.35947990e-01 5.87161541e-01
1.35075986e+00 -2.29539230e-01 -9.82214630e-01 4.00619775e-01
1.10605538e+00 -4.30786490e-01 -1.06753254e+00 6.24799132e-01
3.17868918e-01 -5.20150065e-01 8.43095124e-01 -1.14634824e+00
1.29954886e+00 -1.29968911e-01 3.56880575e-01 -1.22231245e+00
-3.21462721e-01 -3.67083341e-01 -1.30176365e-01 1.14418221e+00
9.83889759e-01 -3.90696853e-01 6.73057258e-01 4.86347198e-01
-1.30844519e-01 -1.17855942e+00 -6.36480033e-01 -9.44081321e-02
-4.91427258e-02 -2.22336590e-01 6.27533674e-01 1.18432117e+00
3.93933713e-01 1.15570867e+00 -1.41396597e-01 4.64804977e-01
4.62261021e-01 4.21821952e-01 2.26207182e-01 -1.22961724e+00
-3.44938457e-01 -7.75991455e-02 -1.68541130e-02 -7.35609770e-01
3.83160263e-01 -1.05938971e+00 1.83240592e-01 -2.05034542e+00
6.24323785e-01 -5.52589476e-01 -3.51808459e-01 7.93804586e-01
-9.54201460e-01 -2.34489202e-01 -2.96160579e-01 2.66948909e-01
-3.52720350e-01 4.18400824e-01 1.59453082e+00 1.35669047e-02
-3.03399920e-01 6.14528880e-02 -1.08027077e+00 9.18002307e-01
7.90626526e-01 -1.02901340e+00 -4.15957421e-01 -1.89789966e-01
2.03228757e-01 5.13433874e-01 1.72531500e-01 -2.15924904e-01
3.30548175e-02 -3.98503184e-01 4.68622863e-01 -4.47474360e-01
-3.10037047e-01 -5.57175338e-01 1.27838895e-01 7.58248508e-01
-4.97007489e-01 -1.77288115e-01 -1.48157999e-01 7.47134566e-01
-2.09087729e-01 -2.21601978e-01 3.19691509e-01 -1.05674654e-01
-3.26103985e-01 3.24972719e-01 -3.86231765e-02 -3.17793190e-02
7.53537893e-01 3.35542172e-01 -4.52720821e-01 1.14920348e-01
-9.84295249e-01 3.66162837e-01 -2.23340496e-01 2.97046065e-01
6.24749482e-01 -1.12010205e+00 -8.94959748e-01 -1.76034123e-01
1.25199303e-01 4.25400704e-01 2.34994575e-01 9.62106049e-01
-2.58934736e-01 3.57987553e-01 1.42009825e-01 -3.08355361e-01
-1.30529189e+00 8.07849884e-01 2.19463661e-01 -9.93005872e-01
-1.01925719e+00 8.00673902e-01 5.63035727e-01 -1.46335050e-01
-3.07663947e-01 -2.42288724e-01 -6.64874732e-01 -2.28673801e-01
2.88002163e-01 -3.72119248e-01 -1.16079196e-01 -3.46387029e-01
-7.87627280e-01 8.46102461e-02 -9.86685231e-02 1.37446418e-01
1.50469363e+00 1.05079427e-01 -3.57055843e-01 -1.69900674e-02
8.97773802e-01 -3.99324857e-02 -7.96320558e-01 -1.49826780e-01
3.90994936e-01 -3.96271870e-02 -7.49821588e-02 -6.46092355e-01
-1.11453164e+00 6.52731240e-01 -8.58853981e-02 8.12494457e-02
1.25696445e+00 4.57308561e-01 8.09155464e-01 1.52069807e-01
-2.99096331e-02 -5.38001835e-01 -1.65478200e-01 1.71614379e-01
7.44666517e-01 -1.04689097e+00 1.27211764e-01 -1.22998643e+00
-7.32153535e-01 1.06257510e+00 6.54074848e-01 5.74372262e-02
7.29691327e-01 5.76748669e-01 5.57638034e-02 -6.53439164e-01
-8.57892931e-01 -4.56958979e-01 3.68350923e-01 5.40779531e-01
8.94891798e-01 2.56832927e-01 -6.43969357e-01 1.20070934e+00
-3.82039100e-01 5.29864095e-02 2.50716478e-01 5.04389644e-01
-1.49748996e-01 -1.19361222e+00 3.96662354e-02 5.63317120e-01
-9.56000209e-01 -5.69724977e-01 -4.56440896e-01 6.95463240e-01
1.39819458e-01 1.06559157e+00 -3.70255977e-01 -3.94222140e-01
3.89943182e-01 8.09389353e-02 2.32271954e-01 -8.62340868e-01
-6.04784727e-01 2.79035389e-01 5.03588557e-01 -4.67331111e-01
-4.97397840e-01 -4.08628762e-01 -1.98912191e+00 1.54663175e-01
-3.92095447e-01 2.70543993e-01 2.13543788e-01 1.23155069e+00
2.02455074e-01 1.12604737e+00 4.45717752e-01 2.92625874e-01
-4.65769134e-03 -9.17045414e-01 -1.83355674e-01 3.81373823e-01
1.52028665e-01 -5.14809191e-01 9.11360532e-02 2.15272278e-01] | [8.7108736038208, 8.796212196350098] |
34deccb2-d67f-4902-a86d-7d566db2001c | vus-at-iwslt-2021-a-finetuned-pipeline-for | null | null | https://aclanthology.org/2021.iwslt-1.12 | https://aclanthology.org/2021.iwslt-1.12.pdf | VUS at IWSLT 2021: A Finetuned Pipeline for Offline Speech Translation | In this technical report, we describe the fine-tuned ASR-MT pipeline used for the IWSLT shared task. We remove less useful speech samples by checking WER with an ASR model, and further train a wav2vec and Transformers-based ASR module based on the filtered data. In addition, we cleanse the errata that can interfere with the machine translation process and use it for Transformer-based MT module training. Finally, in the actual inference phase, we use a sentence boundary detection model trained with constrained data to properly merge fragment ASR outputs into full sentences. The merged sentences are post-processed using part of speech. The final result is yielded by the trained MT module. The performance using the dev set displays BLEU 20.37, and this model records the performance of BLEU 20.9 with the test set. | ['Won Ik Cho', 'Jihyung Moon', 'Jungyoon Choi', 'Minji Jung', 'Youngki Moon', 'Yong Rae Jo'] | null | null | null | null | acl-iwslt-2021-8 | ['boundary-detection'] | ['computer-vision'] | [ 4.09184724e-01 4.15218323e-01 1.95779294e-01 -5.04088104e-01
-1.46866918e+00 -7.28894055e-01 4.40772653e-01 -1.07069127e-01
-6.14327192e-01 8.21854711e-01 5.75543523e-01 -9.34730172e-01
4.42919850e-01 -3.71724963e-01 -6.41389668e-01 -3.69870156e-01
4.99862492e-01 5.17205596e-01 1.51971459e-01 -4.48704392e-01
2.44545192e-02 -2.15388350e-02 -8.93469095e-01 6.12352729e-01
9.79760349e-01 5.46811879e-01 2.36082271e-01 1.05521667e+00
-1.17593326e-01 6.19512618e-01 -1.07857275e+00 -5.93491435e-01
3.74496549e-01 -6.45081699e-01 -8.82156730e-01 -1.02776058e-01
3.88825744e-01 -2.04916283e-01 -5.79616308e-01 7.87521183e-01
6.83342934e-01 2.04143807e-01 2.32359067e-01 -4.59108651e-01
-5.20607531e-01 1.41330683e+00 -1.57080609e-02 4.24308389e-01
3.69449168e-01 3.00012290e-01 9.99050975e-01 -1.23952210e+00
6.36571884e-01 1.23234117e+00 2.73305923e-01 8.69950056e-01
-1.09893715e+00 -6.71495438e-01 -1.33419588e-01 8.97517353e-02
-1.13206041e+00 -1.15209019e+00 2.89124072e-01 -1.28965512e-01
1.65486550e+00 5.38528979e-01 2.35033587e-01 1.23910558e+00
2.30374917e-01 6.76773310e-01 8.62104595e-01 -7.03959525e-01
5.76747442e-03 -6.65569827e-02 4.39829767e-01 4.24670637e-01
-1.41132995e-01 -1.64127648e-01 -5.63018024e-01 3.94171894e-01
1.77963525e-01 -6.26945317e-01 -3.59592497e-01 8.96233559e-01
-1.28087986e+00 5.40610909e-01 -1.22496650e-01 4.92872357e-01
-2.53503442e-01 -6.02926165e-02 4.02427226e-01 7.74447381e-01
8.31444740e-01 4.51661915e-01 -7.78273642e-01 -4.48955446e-01
-1.45044529e+00 -2.48040780e-01 6.36068463e-01 1.11191642e+00
5.45287371e-01 2.55641162e-01 -9.41210687e-01 1.22322881e+00
2.96585977e-01 6.69019639e-01 6.94075644e-01 -7.29872108e-01
1.12145078e+00 2.84152776e-01 -1.78093702e-01 1.00389592e-01
1.31285012e-01 -5.01563549e-01 -3.75493973e-01 -2.54785240e-01
3.04238766e-01 -4.18821454e-01 -1.58273387e+00 1.33624923e+00
-6.85792565e-02 -2.14606628e-01 3.42228949e-01 6.11507297e-01
8.67132783e-01 1.04717422e+00 -1.43078744e-01 -3.50336373e-01
1.04259610e+00 -1.26192129e+00 -1.22856069e+00 -4.01802808e-01
9.40213621e-01 -1.16913664e+00 1.13163304e+00 1.60380214e-01
-1.48660839e+00 -4.94911283e-01 -1.19150531e+00 -3.59391987e-01
-1.89302951e-01 3.30384791e-01 -1.81505695e-01 5.77662408e-01
-1.32218146e+00 6.19213283e-01 -8.21901977e-01 -3.34584892e-01
-7.10720718e-02 1.96302474e-01 -1.67926475e-01 -6.78807124e-02
-1.40661323e+00 1.28938973e+00 4.51270193e-01 2.21594408e-01
-8.35912943e-01 -4.61315811e-01 -1.14071560e+00 6.64591417e-02
-8.95820092e-03 -3.49184573e-01 1.83890820e+00 -6.45337522e-01
-2.00708771e+00 7.43465662e-01 -6.89389169e-01 -5.48010647e-01
4.32442665e-01 -2.19190776e-01 -5.97681761e-01 -3.14486712e-01
-2.39177756e-02 3.16179723e-01 5.81759572e-01 -6.97160840e-01
-6.62935257e-01 -1.67398468e-01 -4.96878177e-01 2.83801705e-01
-2.06516800e-03 4.11966175e-01 -5.43588281e-01 -5.37796974e-01
1.04584880e-01 -6.83604360e-01 -3.19716148e-02 -1.17488348e+00
-6.39970541e-01 -2.32477739e-01 4.44521904e-01 -1.58667588e+00
1.82804465e+00 -2.04415178e+00 1.43608347e-01 1.59199804e-01
-2.68038988e-01 4.90141302e-01 -5.62890530e-01 4.39672440e-01
2.69966433e-03 4.19366777e-01 -2.13363841e-01 -8.08183849e-01
-1.99168324e-02 7.42258504e-02 -3.50006998e-01 9.03045610e-02
3.90899748e-01 1.03180945e+00 -6.50498748e-01 -2.83888280e-01
2.94824034e-01 1.52109459e-01 -6.78544566e-02 4.42479104e-01
-1.41038701e-01 3.43862176e-01 3.74469832e-02 3.07378918e-01
4.46599573e-01 3.97775620e-01 1.27372980e-01 2.24786162e-01
-3.29787642e-01 1.42597651e+00 -5.94883204e-01 1.66627097e+00
-7.63608813e-01 8.84570599e-01 -5.98365366e-02 -4.22710866e-01
8.90564144e-01 6.86166763e-01 -1.07987896e-01 -8.23329329e-01
1.48778170e-01 4.02216762e-01 2.36360222e-01 -2.84675419e-01
7.52731144e-01 2.20005825e-01 8.89622271e-02 2.86836743e-01
4.26514983e-01 -4.16526586e-01 2.83218503e-01 1.44000500e-01
1.36514759e+00 1.18987858e-01 -3.70939486e-02 1.11802772e-01
4.90956068e-01 1.23243593e-01 6.07360482e-01 5.59922397e-01
-4.36103670e-03 8.79951417e-01 6.30059689e-02 1.58590451e-01
-1.29325616e+00 -1.07554626e+00 5.85047826e-02 9.38692987e-01
-5.99387884e-01 -5.84731340e-01 -1.10886919e+00 -8.39065313e-01
-5.89729249e-01 1.48721600e+00 -2.01444671e-01 -9.90767404e-02
-1.01049793e+00 -3.37761551e-01 8.32647204e-01 1.70022622e-01
1.95708588e-01 -1.10465455e+00 2.44455233e-01 5.25708139e-01
-6.56751990e-01 -1.03728926e+00 -8.73757005e-01 2.72174120e-01
-5.72415590e-01 -5.32136142e-01 -6.94231689e-01 -8.69702816e-01
3.00640196e-01 6.84064999e-02 9.89813089e-01 -6.73923939e-02
3.59555304e-01 -2.25936800e-01 -6.15852952e-01 -2.36414313e-01
-1.22585821e+00 3.64216626e-01 -1.58218026e-01 -2.41888285e-01
4.12530750e-01 -9.46127325e-02 -9.22837853e-02 5.94514385e-02
-4.00963217e-01 8.30926746e-02 4.39330310e-01 8.56959701e-01
5.70259094e-01 -2.79176414e-01 4.83500898e-01 -5.50731599e-01
6.55179322e-01 -1.84388146e-01 -5.01806557e-01 4.98567164e-01
-4.45096552e-01 2.55710274e-01 5.04710257e-01 -3.75650167e-01
-1.32981896e+00 -1.50039777e-01 -7.15110362e-01 -1.23894386e-01
1.00945592e-01 3.94249737e-01 -6.22128129e-01 6.22264564e-01
4.51081932e-01 2.96107531e-01 -3.96994054e-01 -6.48917913e-01
3.16559106e-01 1.43769288e+00 3.84471118e-01 2.24338844e-02
8.34396601e-01 -5.68437159e-01 -8.39025199e-01 -7.82606602e-01
-9.47656631e-01 -3.01964462e-01 -4.19317961e-01 5.24198152e-02
8.83699536e-01 -8.02416027e-01 7.44210109e-02 3.33871871e-01
-1.69400740e+00 -7.02321887e-01 -3.91907394e-01 6.79090679e-01
-2.85546809e-01 2.82068282e-01 -9.95900095e-01 -8.66116941e-01
-8.86352003e-01 -1.30991530e+00 7.74803102e-01 1.80018432e-02
-5.22153378e-01 -6.37664378e-01 1.99370921e-01 6.85310721e-01
3.56446654e-01 -6.99710429e-01 6.49105132e-01 -1.04577219e+00
-2.91582555e-01 3.66125181e-02 1.55864060e-01 7.76943445e-01
1.10943742e-01 8.17339867e-02 -1.29568958e+00 -1.55109182e-01
-2.96029355e-02 1.13862358e-01 8.68938327e-01 3.62926275e-01
7.01847434e-01 -3.87869656e-01 -3.38975415e-02 5.06984651e-01
8.08406234e-01 3.10108989e-01 8.75118375e-01 2.61173218e-01
5.80351532e-01 4.28153276e-01 6.54324293e-01 -1.01641335e-01
1.95358321e-01 5.12260258e-01 -1.33662656e-01 8.97833630e-02
-6.05587602e-01 -4.95475948e-01 9.62925375e-01 1.65468800e+00
4.08536196e-01 -6.80501282e-01 -7.40117192e-01 5.11802793e-01
-1.57864332e+00 -8.21209431e-01 -4.39471751e-01 2.29931355e+00
1.23561096e+00 3.36866051e-01 -4.14858460e-01 1.35135770e-01
8.78712595e-01 -3.95637751e-03 -8.99768248e-03 -9.89843190e-01
-2.69389451e-01 8.12601924e-01 7.01574028e-01 1.04325616e+00
-7.41708517e-01 1.75046587e+00 6.47724104e+00 9.04556811e-01
-9.71768498e-01 5.15731633e-01 4.01805222e-01 -1.82927959e-02
-6.68679535e-01 8.73110890e-02 -1.01477313e+00 4.41164970e-01
1.78881061e+00 -8.49610865e-02 5.80457151e-01 2.79960901e-01
7.01394677e-01 1.14342712e-01 -1.12732434e+00 7.02360451e-01
-6.84973374e-02 -9.64150429e-01 -1.02344774e-01 -1.18422255e-01
4.52450544e-01 3.80838841e-01 -1.67543739e-01 6.33775592e-01
5.50461769e-01 -1.02549529e+00 9.22283113e-01 1.67382479e-01
1.02303958e+00 -6.05370820e-01 9.09455895e-01 2.64039814e-01
-7.70496249e-01 3.35474223e-01 -3.43879700e-01 4.03153114e-02
2.70203054e-01 8.02951217e-01 -1.31712556e+00 5.68702698e-01
2.54994810e-01 4.70470041e-01 -4.62293863e-01 9.14609730e-01
-7.52049387e-01 1.42450559e+00 -2.38118574e-01 5.18977828e-02
-3.90178780e-03 -6.94365427e-02 7.61526346e-01 1.75516343e+00
5.42733729e-01 -3.25561643e-01 -3.85531276e-01 5.84771633e-01
-4.14611459e-01 1.86649680e-01 -3.59474212e-01 -2.21771281e-02
7.38522768e-01 1.12307930e+00 -1.61536694e-01 -5.73357403e-01
-1.92616478e-01 1.53311908e+00 3.56885582e-01 5.31233549e-01
-6.58224344e-01 -5.26375532e-01 7.13489413e-01 -2.53392249e-01
3.54970694e-01 -2.65148997e-01 -3.71151984e-01 -1.37491131e+00
2.02854514e-01 -1.06547511e+00 8.33848193e-02 -7.63642907e-01
-8.31733763e-01 8.08971643e-01 -5.52836597e-01 -8.65707397e-01
-4.83919531e-01 -2.27606013e-01 -7.49395192e-01 1.41920066e+00
-1.41819656e+00 -8.34470570e-01 3.00245494e-01 2.27590099e-01
1.19910944e+00 -1.18431091e-01 8.28643382e-01 3.00022572e-01
-7.51420140e-01 1.00882351e+00 1.35467336e-01 5.24501979e-01
9.76935923e-01 -1.27183270e+00 1.15199327e+00 1.43453276e+00
8.12686533e-02 5.84463358e-01 6.45055413e-01 -9.12070870e-01
-1.30364704e+00 -1.45636499e+00 1.68411410e+00 -6.70237541e-01
6.16417885e-01 -5.84865153e-01 -8.57832909e-01 9.58156168e-01
6.73428416e-01 -4.82438326e-01 4.05640572e-01 1.89225152e-02
-1.83971882e-01 -7.07772048e-03 -9.31257129e-01 6.42346680e-01
9.40683484e-01 -5.99311054e-01 -1.12202346e+00 6.65033609e-02
1.40475214e+00 -6.14463627e-01 -8.18520010e-01 1.03555150e-01
1.80478111e-01 -1.46846041e-01 2.64232665e-01 -4.51534569e-01
1.67839944e-01 -2.44277522e-01 -2.39471436e-01 -2.02052855e+00
-3.28888834e-01 -1.02857780e+00 4.25358601e-02 1.64544928e+00
1.21866834e+00 -3.75163764e-01 2.21590459e-01 3.73090029e-01
-7.76398242e-01 -8.76269564e-02 -1.18132579e+00 -8.36190760e-01
2.62839258e-01 -5.28305590e-01 5.45307517e-01 5.35386384e-01
1.82903826e-01 6.99675381e-01 -2.59253532e-02 7.61578009e-02
3.54223669e-01 -6.06035769e-01 3.94360542e-01 -6.41396880e-01
-2.62742430e-01 -1.78898498e-01 1.60243928e-01 -9.99717951e-01
1.91605628e-01 -1.06877005e+00 5.65808773e-01 -1.70041275e+00
-7.45404586e-02 -8.05071667e-02 -2.39235014e-01 6.27032042e-01
-3.48407298e-01 2.88503245e-02 2.92470962e-01 -5.97917959e-02
-2.91791528e-01 4.00858164e-01 1.01269019e+00 -3.25721949e-01
-4.48395520e-01 2.56717838e-02 -3.92752379e-01 1.05009429e-01
1.24102235e+00 -5.72491348e-01 -7.89884571e-03 -8.87426794e-01
-1.82171181e-01 2.09575608e-01 -3.23919237e-01 -8.65014255e-01
-6.72960281e-02 5.90492934e-02 2.83861369e-01 -7.89014101e-01
2.00469583e-01 -4.94379669e-01 -2.70930856e-01 2.45457962e-01
-5.64041555e-01 1.91565081e-02 1.90300405e-01 -3.10889482e-01
-1.41146243e-01 -3.63850385e-01 6.80464566e-01 1.01579435e-01
-6.85794204e-02 1.04315476e-02 -8.82826030e-01 1.69474140e-01
4.14946288e-01 -2.21541569e-01 -2.87740409e-01 -3.63611996e-01
-6.64107859e-01 2.95289218e-01 2.28770599e-01 5.39595962e-01
5.13912320e-01 -9.44771588e-01 -1.19430280e+00 2.63627917e-01
-2.76820194e-02 -1.08878210e-01 -1.88072577e-01 6.83878541e-01
-1.06216982e-01 3.84182513e-01 2.92246670e-01 -1.86811671e-01
-1.35319865e+00 1.52742237e-01 3.50409269e-01 -3.09169561e-01
-4.16777700e-01 6.07608736e-01 -4.13124949e-01 -5.69642007e-01
1.85069576e-01 -5.96945107e-01 6.81743994e-02 -1.90187410e-01
6.19349658e-01 4.62039024e-01 7.48021007e-01 -5.14147162e-01
-3.50005746e-01 -5.04862145e-02 -1.84794679e-01 -9.13837850e-01
1.22452843e+00 -4.96646225e-01 -3.49352881e-02 4.47003067e-01
1.19470310e+00 4.29837286e-01 -8.04184377e-01 -1.41377762e-01
1.97964177e-01 2.63263047e-01 1.35982946e-01 -1.23656654e+00
-5.88235736e-01 1.01548350e+00 3.06228250e-01 -1.18382290e-01
8.16925585e-01 -1.65207952e-01 1.07549524e+00 3.57742637e-01
-1.28237635e-01 -1.54311979e+00 -5.41587770e-01 1.45180833e+00
7.11768210e-01 -9.85823631e-01 -6.15830898e-01 -1.51773676e-01
-4.07313377e-01 1.03899670e+00 4.30992037e-01 2.54031211e-01
1.83572665e-01 4.21230346e-01 3.95531088e-01 4.58436579e-01
-9.49200332e-01 -2.41666600e-01 2.29752153e-01 4.37746167e-01
8.60593796e-01 2.24078566e-01 -6.86693370e-01 5.49754083e-01
-6.78711891e-01 -1.07917897e-01 7.46017992e-01 5.64284682e-01
-7.12423325e-01 -1.31746984e+00 -2.65398234e-01 2.32216164e-01
-7.51488507e-01 -7.61101365e-01 -5.87736785e-01 2.76624888e-01
-2.65883625e-01 1.64660037e+00 1.62029371e-01 -6.83738291e-01
4.75306064e-01 5.56061447e-01 3.85741055e-01 -9.79734838e-01
-9.80293334e-01 3.08499604e-01 7.65830219e-01 -5.25176108e-01
4.02668297e-01 -4.77469683e-01 -1.26382530e+00 -1.28180504e-01
-4.36831772e-01 3.69895160e-01 9.37847078e-01 1.02345872e+00
3.09977531e-01 6.58333778e-01 7.43372083e-01 -3.54973584e-01
-7.31915116e-01 -1.76514125e+00 5.79743758e-02 1.31082430e-01
4.36834097e-01 3.01058352e-01 -5.11960089e-01 -1.24766333e-02] | [14.403006553649902, 7.163525581359863] |
4dc56ef9-aeae-4f12-aa4e-321e0250580d | pareto-secure-machine-learning-psml | 2307.01292 | null | https://arxiv.org/abs/2307.01292v1 | https://arxiv.org/pdf/2307.01292v1.pdf | Pareto-Secure Machine Learning (PSML): Fingerprinting and Securing Inference Serving Systems | With the emergence of large foundational models, model-serving systems are becoming popular. In such a system, users send the queries to the server and specify the desired performance metrics (e.g., accuracy, latency, etc.). The server maintains a set of models (model zoo) in the back-end and serves the queries based on the specified metrics. This paper examines the security, specifically robustness against model extraction attacks, of such systems. Existing black-box attacks cannot be directly applied to extract a victim model, as models hide among the model zoo behind the inference serving interface, and attackers cannot identify which model is being used. An intermediate step is required to ensure that every input query gets the output from the victim model. To this end, we propose a query-efficient fingerprinting algorithm to enable the attacker to trigger any desired model consistently. We show that by using our fingerprinting algorithm, model extraction can have fidelity and accuracy scores within $1\%$ of the scores obtained if attacking in a single-model setting and up to $14.6\%$ gain in accuracy and up to $7.7\%$ gain in fidelity compared to the naive attack. Finally, we counter the proposed attack with a noise-based defense mechanism that thwarts fingerprinting by adding noise to the specified performance metrics. Our defense strategy reduces the attack's accuracy and fidelity by up to $9.8\%$ and $4.8\%$, respectively (on medium-sized model extraction). We show that the proposed defense induces a fundamental trade-off between the level of protection and system goodput, achieving configurable and significant victim model extraction protection while maintaining acceptable goodput ($>80\%$). We provide anonymous access to our code. | ['Alexey Tumanov', 'Sibin Mohan', 'Tianhao Wang', 'Somesh Jha', 'Shahab Nikkhoo', 'Prahlad Jasti', 'Manav Agrawal', 'Jui-Tse Hung', 'Debopam Sanyal'] | 2023-07-03 | null | null | null | null | ['model-extraction', 'model-extraction'] | ['adversarial', 'methodology'] | [ 6.96242750e-02 -4.55491364e-01 -1.74467042e-01 -1.35244682e-01
-9.58500087e-01 -1.07433915e+00 -2.23238934e-02 6.03196137e-02
-4.76335287e-01 6.81492537e-02 -6.66000247e-01 -1.00763929e+00
2.60390760e-03 -1.01528656e+00 -7.25955904e-01 -2.76764840e-01
-4.45931494e-01 3.48980814e-01 5.47542214e-01 1.03579260e-01
3.77269596e-01 7.60889232e-01 -1.18233860e+00 1.57304734e-01
4.78011966e-01 1.25111401e+00 -4.01290894e-01 1.09198618e+00
-2.18077805e-02 4.16674137e-01 -9.35780525e-01 -3.21967393e-01
4.78760928e-01 -1.41731456e-01 -5.71341634e-01 -5.11295438e-01
2.93934882e-01 -8.38642180e-01 -3.45052183e-01 9.01723564e-01
3.42471451e-01 -3.92655283e-01 1.56862661e-01 -1.82298911e+00
1.94058955e-01 6.54784024e-01 -5.88795066e-01 2.35013559e-01
2.93625206e-01 2.69407809e-01 9.64807332e-01 -4.32248652e-01
1.55028403e-01 8.84398937e-01 2.78737724e-01 2.98341423e-01
-1.57484090e+00 -1.35425043e+00 -2.14454591e-01 -1.57960236e-01
-1.66168332e+00 -6.89199984e-01 3.72073978e-01 -9.13930759e-02
8.60646009e-01 8.76911104e-01 -1.67213008e-01 7.23987222e-01
-2.00327523e-02 2.54701108e-01 9.47750926e-01 -2.64401227e-01
4.86728400e-01 3.36568654e-01 5.09292066e-01 5.72158515e-01
6.23338640e-01 3.48035187e-01 -3.28349710e-01 -1.17784309e+00
5.79303443e-01 -1.02649160e-01 -2.20395371e-01 -1.55082926e-01
-6.98264778e-01 5.35715520e-01 -1.01058464e-02 -1.39910415e-01
-1.14178963e-01 6.33637011e-01 3.64313036e-01 4.90180910e-01
-3.11945558e-01 3.39947641e-01 -4.82573330e-01 -2.75629431e-01
-1.13949442e+00 2.08011214e-02 1.22888386e+00 1.13571835e+00
6.20086074e-01 2.29489636e-02 -1.98895950e-02 1.44234583e-01
2.94818193e-01 1.15183651e+00 -2.59474367e-01 -1.09891570e+00
4.45462257e-01 3.66121382e-01 4.43642020e-01 -9.45189238e-01
-7.22820172e-03 -4.81985360e-01 -5.24482191e-01 2.89544404e-01
4.85306501e-01 -1.88057542e-01 -2.28261262e-01 1.90228188e+00
2.61019766e-01 2.82966793e-01 -1.90450639e-01 3.21966529e-01
-2.82124192e-01 5.42252839e-01 -1.89038098e-01 -3.36705983e-01
1.54754317e+00 -4.59849685e-01 -1.89567551e-01 1.78640023e-01
5.22517800e-01 -9.96977031e-01 1.14950252e+00 2.89924592e-01
-1.03525245e+00 -3.25985014e-01 -1.13134730e+00 7.43830979e-01
-8.28831084e-03 -2.96404004e-01 1.47683352e-01 1.29341900e+00
-9.26068485e-01 4.53329831e-02 -9.91719067e-01 -1.42667904e-01
-3.20221633e-02 7.82684922e-01 1.23533837e-01 2.87242383e-01
-8.07993889e-01 2.69884825e-01 -2.01114088e-01 -5.27988851e-01
-9.96879220e-01 -7.86880672e-01 -2.81061918e-01 5.42503834e-01
5.12541294e-01 -5.06262362e-01 1.28319728e+00 -2.30819166e-01
-1.12955153e+00 3.29760611e-01 -2.10210130e-01 -5.21279097e-01
3.81697208e-01 -6.30550832e-02 -6.36106551e-01 2.82129109e-01
-1.56296998e-01 -2.19634339e-01 8.08614552e-01 -1.49273801e+00
-7.08952188e-01 -1.70813188e-01 4.19185221e-01 -5.93570828e-01
-5.66298902e-01 5.37392259e-01 -3.35293919e-01 -3.47224921e-01
-6.10972829e-02 -1.10991442e+00 -3.57768275e-02 4.87256981e-02
-5.61552227e-01 5.59605241e-01 9.83793914e-01 -3.07836711e-01
1.76930618e+00 -2.24318957e+00 -8.57606828e-01 1.04085851e+00
4.63672519e-01 4.19359475e-01 8.76004472e-02 6.92757726e-01
5.06069541e-01 5.90393245e-01 -1.70975879e-01 -3.68723199e-02
1.34607941e-01 -5.98664545e-02 -7.68835902e-01 4.54697996e-01
-3.99395615e-01 3.45147729e-01 -3.79814118e-01 -2.48519644e-01
-1.10849284e-01 1.94745675e-01 -9.37696397e-01 5.25795698e-01
-2.86305044e-02 -2.17565611e-01 -5.74783266e-01 6.27129316e-01
9.96600509e-01 -4.98201072e-01 5.04352510e-01 -1.30475521e-01
-8.55574682e-02 3.12422723e-01 -1.43986905e+00 6.64850891e-01
-7.61054993e-01 3.44825760e-02 6.61106884e-01 -3.77966017e-01
7.86631346e-01 2.78310478e-01 4.22839552e-01 -3.79443169e-01
1.77634563e-02 4.22628790e-01 1.41039908e-01 -2.27016266e-02
2.43248686e-01 3.47649187e-01 -4.02545571e-01 1.27870321e+00
-6.39325857e-01 3.02425206e-01 -2.37699419e-01 2.94694990e-01
1.52099216e+00 -5.45715988e-01 -1.10961692e-02 -2.70269960e-01
4.85370189e-01 -1.93560362e-01 3.07013333e-01 1.28887475e+00
-2.49336019e-01 -2.54726559e-02 5.59913337e-01 -6.62819520e-02
-7.18490541e-01 -1.18661773e+00 2.37087756e-02 1.07920814e+00
3.53901207e-01 -7.92689323e-01 -9.13905680e-01 -7.02132463e-01
1.33509740e-01 7.29080677e-01 -8.81851986e-02 -3.67094189e-01
-8.55205953e-01 -3.15448493e-01 1.37953877e+00 4.02937680e-01
7.21460819e-01 -4.56680894e-01 -9.03545320e-01 2.88621664e-01
-1.14443839e-01 -1.06455159e+00 -7.89529800e-01 -4.62657399e-02
-5.73363900e-01 -1.25863659e+00 2.14259252e-01 -3.04210901e-01
5.49460113e-01 5.89276910e-01 8.90358090e-01 5.94355822e-01
-3.00703719e-02 3.44787359e-01 8.35955739e-02 4.26012417e-03
-6.92924857e-01 2.36036435e-01 1.95689008e-01 2.01684251e-01
2.92497009e-01 -8.30724239e-01 -6.60113454e-01 6.21291280e-01
-9.41197932e-01 -4.12291408e-01 2.35050395e-01 5.75538993e-01
2.33568817e-01 5.82509153e-02 4.69334960e-01 -8.87242019e-01
5.45387030e-01 -4.06714916e-01 -1.14690053e+00 3.67831856e-01
-8.95342529e-01 1.54145449e-01 1.09820640e+00 -4.09834534e-01
-4.96791154e-01 -2.19397068e-01 -1.46158591e-01 -2.48103797e-01
7.45788291e-02 -1.29429530e-02 -3.89779836e-01 -3.78877640e-01
8.22399199e-01 3.21714222e-01 1.11007459e-01 -3.12488765e-01
1.65535450e-01 8.43907475e-01 3.02197874e-01 -8.73719573e-01
1.28094339e+00 5.19370079e-01 -5.14347479e-03 -4.75555003e-01
-9.20518860e-02 -2.50785738e-01 3.66212092e-02 1.45274714e-01
2.40692659e-03 -4.05675709e-01 -1.55271375e+00 3.21621299e-01
-8.77346516e-01 -2.08959967e-01 2.49952421e-01 1.34459659e-01
-3.43102843e-01 5.90871096e-01 -5.80270946e-01 -1.07845080e+00
-7.60275960e-01 -1.28803766e+00 6.74580276e-01 -5.45468517e-02
-2.21380159e-01 -2.80982792e-01 -3.40942919e-01 1.39950514e-01
1.01946020e+00 7.49073476e-02 1.07492280e+00 -9.11942840e-01
-8.79477024e-01 -4.10927534e-01 -5.38226664e-01 1.22301981e-01
-8.44349489e-02 1.88136801e-01 -9.23389733e-01 -6.12559021e-01
8.56029913e-02 2.23206848e-01 1.69609606e-01 -1.50395036e-01
1.29729211e+00 -8.04370999e-01 -3.75256866e-01 7.12949157e-01
1.52620983e+00 5.07159531e-01 2.94375628e-01 8.73575434e-02
2.71849722e-01 -3.28859910e-02 3.05312067e-01 6.91873431e-01
-1.08531006e-01 8.87235880e-01 4.36418951e-01 1.22458383e-01
4.00706083e-01 -1.29075885e-01 4.98084545e-01 2.84202158e-01
5.38994193e-01 -2.29061320e-01 -9.44502056e-01 5.51009960e-02
-1.33908868e+00 -8.92316461e-01 4.57467558e-03 2.79916239e+00
8.13358724e-01 8.15120518e-01 3.43287319e-01 2.88138002e-01
6.03041649e-01 -1.32864907e-01 -6.03349626e-01 -6.13920808e-01
4.31029439e-01 5.37020922e-01 1.03579426e+00 5.86251080e-01
-5.54770172e-01 6.56906605e-01 6.07906675e+00 7.64310062e-01
-1.28734195e+00 -4.07874919e-02 3.44403952e-01 -1.18918873e-01
-3.50382835e-01 3.68958890e-01 -8.65155697e-01 7.56106496e-01
1.58553600e+00 -5.84605932e-01 7.18070805e-01 9.84666526e-01
4.30856973e-01 1.60623845e-02 -1.06843317e+00 7.51526475e-01
-3.93738776e-01 -1.08300352e+00 -1.46182582e-01 4.90590572e-01
-7.56269544e-02 -3.47318053e-01 8.87602419e-02 3.04340214e-01
4.10789639e-01 -6.84567809e-01 4.61690992e-01 1.45384043e-01
9.80249822e-01 -1.02496302e+00 4.61670130e-01 4.88899589e-01
-1.20166075e+00 -2.93529570e-01 2.87930062e-03 1.80471733e-01
8.97531509e-02 2.90258259e-01 -8.59564900e-01 2.64813572e-01
5.12287855e-01 -7.44705141e-01 -4.16945606e-01 7.79635668e-01
3.54919523e-01 1.11090171e+00 -9.38434660e-01 -1.03478553e-02
-1.30120888e-01 1.97737426e-01 4.29304481e-01 1.05711365e+00
3.50267947e-01 -3.97270992e-02 2.87624419e-01 8.39845240e-01
-2.28935555e-01 -6.40351474e-02 -3.06542039e-01 2.68134624e-01
1.29292727e+00 8.58597338e-01 -3.74010503e-01 -5.47634542e-01
-2.35775843e-01 7.96697736e-01 -1.69224143e-01 3.35949153e-01
-1.28388155e+00 -7.32640564e-01 9.78749871e-01 5.94739020e-01
-3.30370106e-02 -3.80103081e-01 -3.63439530e-01 -9.43124473e-01
1.04231603e-01 -1.24101102e+00 4.59080160e-01 -1.64751448e-02
-8.83889079e-01 7.38880098e-01 1.13854138e-02 -9.47158337e-01
-2.58686513e-01 1.04935875e-03 -5.64194620e-01 9.95861471e-01
-1.03886044e+00 -8.17780614e-01 -1.96763858e-01 8.20148051e-01
-2.24079683e-01 -2.99184415e-02 1.04467475e+00 5.57263672e-01
-5.29623806e-01 1.49162841e+00 2.03556672e-01 1.03460595e-01
6.28786445e-01 -8.56612384e-01 7.25392342e-01 1.13049686e+00
6.75076246e-02 1.27966273e+00 6.66190565e-01 -3.45213562e-01
-1.70933831e+00 -8.38912904e-01 6.70642734e-01 -3.18044484e-01
5.43007135e-01 -6.79361701e-01 -7.69635379e-01 4.27757770e-01
-3.50986481e-01 1.92565143e-01 8.91090155e-01 -2.10675359e-01
-9.56839144e-01 -6.13070428e-01 -1.53223550e+00 7.03757465e-01
6.74704432e-01 -8.43476355e-01 2.98792273e-01 -1.47458270e-01
6.75319016e-01 1.79586317e-02 -8.77176821e-01 2.32448041e-01
9.02226806e-01 -1.14566493e+00 9.34278667e-01 -6.42935157e-01
-6.12854600e-01 -4.38574612e-01 -5.70670009e-01 -3.16082627e-01
-4.58617732e-02 -9.91709471e-01 -6.27594650e-01 1.18789434e+00
4.83038723e-01 -8.90378177e-01 8.50015163e-01 8.88390243e-01
3.77192438e-01 -4.93571937e-01 -9.12545919e-01 -9.43449199e-01
-1.28333613e-01 -6.17188632e-01 1.14150143e+00 5.57752132e-01
-1.82643801e-01 -4.17367555e-02 -3.36760312e-01 8.87649238e-01
9.87734437e-01 3.37710738e-01 1.26358974e+00 -8.57739449e-01
-7.52050400e-01 -3.89414519e-01 7.03637209e-03 -1.19400728e+00
-3.26034933e-01 -4.69076037e-01 -5.10915220e-01 -7.68537700e-01
-7.97827616e-02 -8.68655920e-01 -5.09230018e-01 5.09371698e-01
2.12513775e-01 8.87743682e-02 4.48528975e-01 4.22840327e-01
-2.22636536e-01 -2.81836331e-01 3.56445819e-01 1.85608700e-01
-3.19936246e-01 4.99895424e-01 -6.88992977e-01 3.13072711e-01
1.12838089e+00 -7.23410964e-01 -4.91700768e-01 -2.50688959e-02
-7.63388872e-02 5.75008988e-01 4.38590944e-01 -1.10721910e+00
3.35048258e-01 -3.15720409e-01 -1.54606476e-01 -2.62302130e-01
2.86065906e-01 -1.22356725e+00 4.41462725e-01 9.07719493e-01
-2.61302233e-01 2.62014419e-01 3.15203853e-02 4.00558472e-01
2.37210810e-01 2.45139092e-01 8.56151640e-01 3.87399614e-01
2.27936711e-02 3.50568891e-01 -3.39176387e-01 -6.14549480e-02
9.96511102e-01 -2.63062239e-01 -5.68899870e-01 -4.26641226e-01
-4.48812544e-01 1.14978164e-01 7.93749213e-01 5.41213453e-02
2.78141290e-01 -9.40891027e-01 -3.76203477e-01 6.19975567e-01
7.93146417e-02 -8.13305080e-01 -6.87387399e-03 7.04064786e-01
-5.97104609e-01 2.33063862e-01 1.18167326e-01 -3.77530456e-01
-1.44399202e+00 7.84036756e-01 4.81822997e-01 -3.32340986e-01
-8.00160393e-02 4.43819374e-01 1.55561483e-02 -9.09621343e-02
4.01479453e-01 -1.57316998e-01 7.26152539e-01 -6.34094238e-01
7.07395494e-01 5.46239257e-01 3.96899227e-03 -2.11479649e-01
-4.86971170e-01 7.07930252e-02 -6.31767437e-02 -3.35479230e-01
6.42989159e-01 -1.51000276e-01 -5.22940122e-02 -2.34407008e-01
1.32068217e+00 6.16841853e-01 -8.27424526e-01 -2.90764362e-01
5.53672537e-02 -7.36268520e-01 -2.15430811e-01 -7.46189296e-01
-1.02869534e+00 6.85221970e-01 5.98985076e-01 7.46706665e-01
1.07071900e+00 -4.12028581e-01 1.22253513e+00 2.49236792e-01
8.63516271e-01 -5.39965153e-01 -2.74375737e-01 1.40958443e-01
1.69811830e-01 -6.12605631e-01 -4.07833368e-01 -5.84273219e-01
8.95967707e-02 9.06866670e-01 7.38849044e-01 -6.56789094e-02
9.35422003e-01 8.48509252e-01 -7.89957196e-02 1.85171664e-02
-8.14257145e-01 3.42814803e-01 -3.50761384e-01 4.44579244e-01
-1.43900543e-01 2.06759959e-01 -1.68631539e-01 9.11300480e-01
-3.41476016e-02 -2.07487851e-01 6.57767177e-01 1.10537052e+00
-6.89113200e-01 -1.57963920e+00 -7.04233944e-01 5.79534531e-01
-7.73526490e-01 -1.51037529e-01 -5.83329163e-02 6.04771018e-01
-4.64470774e-01 1.37873900e+00 -8.32046866e-02 -7.09673405e-01
3.25070113e-01 -4.83121781e-04 -2.37120301e-01 -3.22919458e-01
-9.41723704e-01 1.33626118e-01 2.63628930e-01 -8.61056268e-01
4.54183519e-01 -2.94198450e-02 -1.30579758e+00 -1.06488681e+00
-4.27637070e-01 5.31779230e-01 4.99241889e-01 2.95161456e-01
7.61849165e-01 -2.79771350e-02 1.37272394e+00 8.83275792e-02
-1.09979999e+00 -4.00977939e-01 -5.73312163e-01 -1.00827336e-01
1.72549382e-01 -3.74874949e-01 -6.62493348e-01 -2.55798221e-01] | [5.683661460876465, 7.309680938720703] |
18865bc2-e009-4e47-95bc-1e9498e10710 | on-the-development-of-a-large-scale-corpus | null | null | http://www.ep.liu.se/ecp/article.asp?issue=155&article=012&volume= | http://www.ep.liu.se/ecp/155/012/ecp18155012.pdf | On the Development of a Large Scale Corpus for Native Language Identification | Native Language Identification (NLI) is the task of identifying an author’s native language from their writings in a second language. In this paper, we introduce a new corpus (italki), which is larger than the current corpora. It can be used for training machine learning based systems for classifying and identifying the native language of authors of English text. To examine the usefulness of italki, we evaluate it by using it to train and test some of the well performing NLI systems presented in the 2017 NLI shared task. In this paper, we present some aspects of italki. We show the impact of the variation of italki’s training dataset size of some languages on systems performance. From our empirical finding, we highlight the potential of italki as a large scale corpus for training machine learning classifiers for classifying the native language of authors from their written English text. We obtained promising results that show the potential of italki to improve the performance of current NLI systems. More importantly, we found that training the current NLI systems on italki generalize better than training them on the current corpora. | ['Sardar Jaf', 'Thomas Hudson'] | 2018-12-10 | null | null | null | tlt17-2018-12 | ['native-language-identification'] | ['natural-language-processing'] | [-1.79406390e-01 1.48680145e-02 -2.46486798e-01 -8.59851465e-02
-8.96091342e-01 -1.16406500e+00 1.05123103e+00 -6.97766244e-02
-6.79078758e-01 6.44079506e-01 3.14324111e-01 -6.71932697e-01
5.75047061e-02 -2.58637220e-01 -4.81041402e-01 -1.20940380e-01
3.12720567e-01 6.85544133e-01 -9.12510008e-02 -1.69570353e-02
5.45762241e-01 6.40226722e-01 -1.23462522e+00 2.89271623e-01
8.56474459e-01 5.01724601e-01 3.62857915e-02 6.56190276e-01
-2.66783327e-01 7.47194707e-01 -1.04833508e+00 -3.56309861e-01
2.97994941e-01 -2.00148061e-01 -1.02006173e+00 -3.92165512e-01
9.49011743e-01 2.28282046e-02 -2.61396617e-01 6.53424919e-01
3.06430489e-01 -1.68197393e-01 8.71857464e-01 -1.12219465e+00
-6.96837366e-01 9.76530612e-01 2.62905229e-02 3.67208272e-01
6.84914052e-01 1.05760828e-01 1.05298328e+00 -9.38051224e-01
1.19372463e+00 1.30570769e+00 7.92142034e-01 5.26803136e-01
-1.02846086e+00 -1.09653795e+00 -1.22282691e-01 -1.11509159e-01
-1.69385350e+00 -7.61451125e-01 4.24645454e-01 -9.34872150e-01
1.03909206e+00 1.67142168e-01 3.51576917e-02 1.29073489e+00
2.36539934e-02 8.79270077e-01 1.34675324e+00 -8.98979127e-01
-1.13583691e-01 7.16798365e-01 5.42738497e-01 4.33098942e-01
1.63978726e-01 -7.93979242e-02 -7.68327534e-01 -2.88938463e-01
1.37680724e-01 -6.36640668e-01 -2.63530821e-01 2.53151953e-01
-1.68234921e+00 7.20751405e-01 -2.83155978e-01 9.68965650e-01
2.68977702e-01 -3.24870586e-01 6.22021258e-01 3.86089176e-01
4.89533514e-01 1.24104857e+00 -6.95385635e-01 -4.73278791e-01
-1.32225621e+00 3.91868241e-02 1.38352430e+00 1.04162419e+00
4.67754662e-01 -2.49945968e-01 -6.57671094e-02 9.79144394e-01
9.93812308e-02 5.33910871e-01 1.11730361e+00 -8.07969928e-01
6.15851283e-01 6.98044121e-01 2.30715852e-02 -5.97055554e-01
-2.99480826e-01 -3.88862759e-01 -4.03481811e-01 -4.71689701e-02
9.61542010e-01 -1.40198782e-01 -4.33009773e-01 1.51415789e+00
-5.97137988e-01 -2.35785171e-01 3.79738003e-01 3.23202282e-01
9.30218160e-01 6.71405137e-01 -6.66112527e-02 5.65246865e-02
1.11603475e+00 -7.27090001e-01 -5.05362809e-01 -2.44683102e-01
1.05717158e+00 -1.00728190e+00 1.17806053e+00 5.24505138e-01
-8.90056729e-01 -7.28835642e-01 -8.59994054e-01 -3.02016735e-04
-8.81647110e-01 9.41174507e-01 4.41713125e-01 8.38431478e-01
-1.14401686e+00 4.02050912e-01 -2.66195506e-01 -8.05883825e-01
-3.55669647e-03 1.94922805e-01 -5.53359866e-01 2.16225356e-01
-1.24547112e+00 8.42606843e-01 3.07531089e-01 -3.52324486e-01
-3.52702081e-01 -8.39641869e-01 -6.14453614e-01 -4.20949273e-02
8.06170329e-02 1.54171675e-01 1.21638358e+00 -1.02493608e+00
-1.35741746e+00 1.53740299e+00 -2.44967565e-01 -2.96645939e-01
4.97706264e-01 3.39033119e-02 -6.70278013e-01 -1.57241702e-01
5.79183698e-01 3.81314129e-01 3.43959242e-01 -8.74998033e-01
-8.28617871e-01 -2.85996765e-01 -3.97086471e-01 -4.83477652e-01
-7.04757154e-01 3.56611669e-01 -2.66430199e-01 -6.37482464e-01
-3.81337494e-01 -1.09417593e+00 4.80743796e-01 -5.55635512e-01
-3.36773932e-01 -8.80674899e-01 4.19183671e-01 -1.00264573e+00
1.50639033e+00 -2.29228735e+00 -1.92805529e-01 1.95251629e-01
1.68806478e-01 4.57992226e-01 -1.85763523e-01 5.93306124e-01
1.11349583e-01 7.42260814e-01 2.40992397e-01 -1.60327926e-01
3.20876278e-02 2.64427606e-02 -3.26004654e-01 2.82348424e-01
-6.78028911e-02 8.96051705e-01 -7.48170972e-01 -3.71326506e-01
-2.83508331e-01 1.45302853e-02 -3.92999314e-02 3.28714073e-01
1.48820430e-01 3.29935610e-01 -7.94130713e-02 4.84283119e-01
3.86897415e-01 1.77764166e-02 5.69507182e-02 1.19810082e-01
-5.66609263e-01 6.09128535e-01 -9.75479722e-01 1.35714877e+00
-6.91689789e-01 1.27181447e+00 7.20368773e-02 -6.30988955e-01
1.04434311e+00 3.43064308e-01 1.29476190e-04 -3.92661333e-01
-2.09184319e-01 7.45395124e-01 4.72663969e-01 -2.19599158e-01
2.44205967e-01 3.61103684e-01 -3.83042663e-01 6.00438297e-01
2.87079960e-01 1.41536146e-01 6.89163327e-01 3.29745799e-01
8.79349887e-01 -3.58267426e-02 4.48168814e-01 -8.51357579e-01
1.14651978e+00 8.44243094e-02 1.76263466e-01 1.16168904e+00
-2.43979126e-01 1.95818186e-01 2.62045860e-01 -2.80158728e-01
-1.08892226e+00 -9.79783177e-01 -4.01080579e-01 1.20122230e+00
-6.76558971e-01 -5.42509675e-01 -5.87326586e-01 -6.83251917e-01
1.45520210e-01 8.07848871e-01 -4.09391791e-01 1.20715387e-01
-6.35127366e-01 -2.56707519e-01 1.17282104e+00 3.37130129e-01
1.86801031e-01 -1.14032686e+00 -1.15146078e-01 -4.23346087e-02
-5.18505946e-02 -1.22323036e+00 -6.33053899e-01 1.92747727e-01
-2.27791891e-01 -9.72562373e-01 -8.28035355e-01 -1.11042011e+00
3.79434168e-01 -3.14159542e-02 1.26753783e+00 -3.58431339e-02
-2.07484931e-01 7.75404513e-01 -3.66938382e-01 -5.62823772e-01
-1.29355752e+00 1.09440243e+00 2.69420207e-01 -4.91576612e-01
6.74112737e-01 1.69467051e-02 2.54655570e-01 6.36757091e-02
-5.65817893e-01 -2.81230986e-01 5.01852870e-01 5.84982216e-01
-2.71425396e-01 -1.81497231e-01 6.21743202e-01 -1.14751375e+00
9.52830374e-01 -1.90825298e-01 -4.40903217e-01 5.60408235e-01
-7.99560010e-01 1.19483590e-01 9.14331555e-01 -6.42220497e-01
-9.53836024e-01 -3.06125104e-01 -4.82872538e-02 2.46023893e-01
-3.36461365e-01 6.38198078e-01 -1.59734055e-01 -3.24762166e-01
5.51380694e-01 2.58212090e-01 -1.66151058e-02 -7.74694920e-01
2.05291808e-02 1.39443135e+00 6.13297999e-01 -6.09377027e-01
7.67927468e-01 -1.80552036e-01 -3.55918705e-01 -9.68823373e-01
-7.77259648e-01 -7.47401059e-01 -1.23688722e+00 5.28108981e-03
4.17684257e-01 -1.03921485e+00 -6.50633395e-01 6.37414098e-01
-1.03344369e+00 -3.95116329e-01 1.96629553e-03 3.44725221e-01
-1.02335729e-01 3.74720782e-01 -5.19142032e-01 -5.80276608e-01
-3.55077207e-01 -1.18587887e+00 8.22911263e-01 1.37407959e-01
-8.14646363e-01 -1.39907575e+00 1.95028290e-01 4.33320940e-01
3.77280653e-01 -3.53355795e-01 7.50798404e-01 -1.49557471e+00
8.91604051e-02 -4.48130786e-01 -6.47345036e-02 4.52530384e-01
4.72884476e-02 2.71836311e-01 -8.03633571e-01 -4.46000427e-01
-3.32982123e-01 -3.59811515e-01 6.19720995e-01 -2.69475490e-01
7.45070815e-01 -4.01185572e-01 -3.87574345e-01 5.86945355e-01
1.22864914e+00 1.50530443e-01 1.10339321e-01 9.48356926e-01
7.14910388e-01 6.44202769e-01 2.66101956e-01 1.62902087e-01
3.82711917e-01 5.66936374e-01 -6.43040001e-01 1.32244751e-01
-3.11089844e-01 -2.59028524e-01 6.72464192e-01 1.01430726e+00
3.65806431e-01 -2.54372329e-01 -1.57763588e+00 6.74908519e-01
-1.20989442e+00 -7.31434464e-01 -9.11523253e-02 2.17676878e+00
1.09922600e+00 1.03091188e-01 1.90149814e-01 -1.48712397e-01
4.28573072e-01 -3.34128648e-01 -1.77283093e-01 -7.58891046e-01
-3.72960031e-01 2.64344454e-01 7.74156511e-01 7.41136670e-01
-1.06943595e+00 1.29963112e+00 6.76131010e+00 8.40691149e-01
-1.25253844e+00 -4.10133675e-02 5.63948750e-01 2.15995654e-01
4.37632464e-02 2.52962280e-02 -1.54727876e+00 6.15812957e-01
1.46819019e+00 -5.81355691e-01 5.51155627e-01 6.48020267e-01
2.18490791e-02 -1.31385133e-01 -1.47388208e+00 8.78560901e-01
3.79771709e-01 -1.22026825e+00 5.37807345e-02 2.19480649e-01
7.33983636e-01 9.26740021e-02 -1.12541653e-01 4.97091442e-01
2.51633584e-01 -1.07230842e+00 7.83410609e-01 4.23752457e-01
9.15492415e-01 -5.04616141e-01 9.14370179e-01 6.19135261e-01
-6.16810024e-01 -4.41324972e-02 -2.54747033e-01 -3.16453397e-01
-4.98134226e-01 2.95629561e-01 -1.18526459e+00 1.89931050e-01
5.24478734e-01 8.99617195e-01 -1.29864800e+00 5.19346654e-01
-1.52600661e-01 1.12943172e+00 -2.72855729e-01 -2.29581192e-01
3.60243767e-01 -8.43949243e-02 4.98662204e-01 1.70483100e+00
2.75573045e-01 -5.72319150e-01 5.10499179e-01 6.54985607e-01
-3.27226102e-01 4.26242560e-01 -7.89870203e-01 -5.95113099e-01
7.36433327e-01 1.17483735e+00 -3.76157761e-01 -6.63318813e-01
-5.55432379e-01 1.01659322e+00 4.83495891e-01 2.83757359e-01
-1.25083894e-01 -6.50460184e-01 4.13490802e-01 1.76973820e-01
-1.94286443e-02 -3.80821437e-01 -4.66361046e-01 -1.32352126e+00
-5.68059497e-02 -1.11208224e+00 4.19474095e-01 -6.75848126e-01
-1.47132266e+00 5.74869812e-01 -1.42963484e-01 -8.04606438e-01
-5.54361999e-01 -1.09687972e+00 -4.42624778e-01 1.26239419e+00
-1.17582953e+00 -1.27912688e+00 -2.01009512e-02 6.85468763e-02
3.82773370e-01 -1.01430011e+00 7.98277020e-01 -9.10358280e-02
-6.93508148e-01 1.14102757e+00 8.29604983e-01 7.05281198e-01
1.40647697e+00 -1.23590577e+00 4.93212044e-01 8.26113284e-01
5.25862694e-01 1.05807376e+00 4.98643875e-01 -6.74129426e-01
-1.11939776e+00 -7.00427890e-01 1.61587811e+00 -8.01736474e-01
1.18680632e+00 -6.68960452e-01 -7.41524518e-01 9.63052452e-01
4.93315905e-01 -6.18027031e-01 8.09514940e-01 2.50707060e-01
-4.14659917e-01 8.51058811e-02 -9.16672051e-01 5.56655943e-01
8.52721512e-01 -7.52500415e-01 -6.64385974e-01 4.93860185e-01
2.98065394e-01 -3.22082378e-02 -1.36115754e+00 -6.72384948e-02
7.06048429e-01 -6.29330337e-01 8.68694246e-01 -5.21282792e-01
4.35137779e-01 1.08934350e-01 3.08501393e-01 -1.26334071e+00
-5.18610060e-01 -5.90133667e-01 4.34895128e-01 1.91535890e+00
7.79679954e-01 -9.33720887e-01 4.29284245e-01 6.72987521e-01
2.84135610e-01 -3.47778171e-01 -9.19026256e-01 -1.17582238e+00
8.33274305e-01 -1.73119575e-01 3.11891407e-01 1.08976758e+00
1.94449946e-01 5.04001319e-01 3.46346274e-02 -2.47446850e-01
2.21292481e-01 -1.70614555e-01 9.72023904e-01 -1.55022418e+00
5.00430651e-02 -6.99839652e-01 -3.29505861e-01 -5.60538232e-01
9.31357265e-01 -1.47936904e+00 -2.45178625e-01 -1.04573762e+00
3.66926610e-01 -3.47072780e-01 1.07994355e-01 5.71831703e-01
-2.18564674e-01 1.54694453e-01 4.13342506e-01 8.71272385e-01
-3.23175430e-01 -2.32727632e-01 6.81082666e-01 -3.00812244e-01
-1.41428262e-01 -1.38969973e-01 -7.72249818e-01 5.68514466e-01
9.16659474e-01 -2.34899685e-01 2.41665598e-02 -2.42509723e-01
9.73534733e-02 -6.84434772e-01 -7.54781142e-02 -1.18326950e+00
3.01217735e-01 1.28523380e-01 3.83233815e-01 -2.85015523e-01
-4.32702750e-01 -4.08028275e-01 -3.90447319e-01 3.72984439e-01
-6.77382469e-01 1.34862050e-01 3.82286221e-01 -2.06100121e-01
-1.63635686e-01 -7.31974125e-01 6.44502044e-01 -4.23424274e-01
-7.71272838e-01 -1.53730467e-01 -9.97601509e-01 5.31095326e-01
6.81731105e-01 -2.29741588e-01 -1.87758073e-01 -1.46414548e-01
-3.03198516e-01 -6.98381439e-02 7.90412068e-01 6.45836949e-01
-2.44719721e-02 -8.45230758e-01 -8.38721454e-01 2.65147507e-01
4.35023516e-01 -8.70716929e-01 -5.26381552e-01 4.72434461e-01
-5.36488116e-01 1.16631472e+00 -2.23887563e-01 -2.05871314e-01
-1.32745206e+00 5.79971611e-01 9.22516808e-02 -6.65161788e-01
-4.44108546e-01 5.65364242e-01 3.46982256e-02 -8.78158271e-01
3.01411688e-01 -1.61493961e-02 -3.21246296e-01 4.31683026e-02
5.00470459e-01 2.70098299e-01 -6.86914474e-02 -8.06904852e-01
-6.30007207e-01 1.69204846e-01 -3.14307183e-01 -4.34510231e-01
1.06457925e+00 -1.27037242e-01 -4.31821167e-01 1.02741635e+00
1.32367349e+00 7.67652690e-01 -1.62665710e-01 -2.50546575e-01
5.17360866e-01 -1.22513175e-01 -1.17771305e-01 -1.27873290e+00
-3.54265422e-01 7.57857919e-01 4.74312514e-01 -1.44651681e-02
5.41480660e-01 7.96676204e-02 6.59077644e-01 4.62541848e-01
1.79131538e-01 -1.28016675e+00 -4.23633099e-01 8.15978050e-01
7.19540477e-01 -1.32427573e+00 -1.25867024e-01 -1.67779997e-01
-4.94952530e-01 1.48752511e+00 4.51575249e-01 2.73585707e-01
5.51554799e-01 3.23593318e-01 5.10052741e-01 1.08714513e-01
-4.05421674e-01 7.24435672e-02 4.88992006e-01 5.31753242e-01
1.11076903e+00 8.28647614e-02 -7.08553314e-01 8.79173696e-01
-6.66418195e-01 -7.26658851e-02 8.88721764e-01 5.44238389e-01
-1.01210892e-01 -1.44081140e+00 -6.21030509e-01 5.66721976e-01
-6.19652629e-01 -4.23086703e-01 -1.02721906e+00 1.02250659e+00
1.08371951e-01 9.66760218e-01 -4.14357968e-02 -1.55540943e-01
-1.89489336e-03 6.46626890e-01 1.29765734e-01 -7.21754909e-01
-1.08577061e+00 -4.95407015e-01 1.89114243e-01 -3.01107429e-02
-8.39333162e-02 -1.09247708e+00 -5.99314213e-01 -1.36651248e-01
9.94206294e-02 5.10521770e-01 5.69758534e-01 1.10301936e+00
2.91950315e-01 1.33133933e-01 3.85282785e-01 -5.20728767e-01
-7.08915830e-01 -1.22973526e+00 -5.10651410e-01 5.53367913e-01
-1.00849733e-01 -2.52029687e-01 -6.84871972e-01 2.27465793e-01] | [10.409684181213379, 10.505586624145508] |
6e56d016-4664-4329-add3-94f8cd2af680 | learning-object-centric-neural-scattering | 2303.06138 | null | https://arxiv.org/abs/2303.06138v3 | https://arxiv.org/pdf/2303.06138v3.pdf | Learning Object-Centric Neural Scattering Functions for Free-Viewpoint Relighting and Scene Composition | Photorealistic object appearance modeling from 2D images is a constant topic in vision and graphics. While neural implicit methods (such as Neural Radiance Fields) have shown high-fidelity view synthesis results, they cannot relight the captured objects. More recent neural inverse rendering approaches have enabled object relighting, but they represent surface properties as simple BRDFs, and therefore cannot handle translucent objects. We propose Object-Centric Neural Scattering Functions (OSFs) for learning to reconstruct object appearance from only images. OSFs not only support free-viewpoint object relighting, but also can model both opaque and translucent objects. While accurately modeling subsurface light transport for translucent objects can be highly complex and even intractable for neural methods, OSFs learn to approximate the radiance transfer from a distant light to an outgoing direction at any spatial location. This approximation avoids explicitly modeling complex subsurface scattering, making learning a neural implicit model tractable. Experiments on real and synthetic data show that OSFs accurately reconstruct appearances for both opaque and translucent objects, allowing faithful free-viewpoint relighting as well as scene composition. | ['Jiajun Wu', 'Thomas Funkhouser', 'Ruohan Gao', 'Eric Ryan Chan', 'Yen-Yu Chang', 'Alireza Fathi', 'Michelle Guo', 'Hong-Xing Yu'] | 2023-03-10 | null | null | null | null | ['inverse-rendering'] | ['computer-vision'] | [ 5.30815244e-01 1.55173510e-01 4.03744072e-01 -3.91613722e-01
-1.37457564e-01 -4.46131110e-01 7.10214794e-01 -5.81913471e-01
2.95416653e-01 5.94328105e-01 2.81607267e-03 -2.27506474e-01
2.99244493e-01 -9.23825622e-01 -1.04219508e+00 -7.21284866e-01
3.87656510e-01 6.73302412e-01 1.55043483e-01 -4.24869098e-02
1.98081303e-02 1.01689577e+00 -1.60696304e+00 5.08566916e-01
6.98629916e-01 9.33295548e-01 2.13820621e-01 9.75047112e-01
-3.00806016e-01 1.21159494e+00 -4.16656315e-01 -2.15017021e-01
3.57038736e-01 -1.33806050e-01 -4.75305498e-01 9.66413468e-02
1.13475931e+00 -9.78863776e-01 -5.15500128e-01 7.27723539e-01
2.75374223e-02 1.59345180e-01 1.16511881e+00 -1.08509731e+00
-1.53931391e+00 -2.23934844e-01 -6.15761161e-01 -5.64611435e-01
1.69837758e-01 1.41610026e-01 7.22104430e-01 -9.39319253e-01
5.05410373e-01 1.60385203e+00 7.68404901e-01 9.25975740e-01
-1.49651146e+00 -5.65173626e-01 2.57290930e-01 -3.46202612e-01
-9.00242805e-01 -4.10314918e-01 8.65861714e-01 -4.18477595e-01
8.14700961e-01 7.50290036e-01 9.62359011e-01 8.47208261e-01
3.82548630e-01 7.04275906e-01 1.25107849e+00 -1.97590277e-01
1.91587031e-01 1.09092698e-01 8.14214423e-02 8.52363110e-01
2.22478345e-01 4.56060082e-01 -4.23085332e-01 -2.61674732e-01
1.53221643e+00 2.64125645e-01 -6.92212462e-01 -4.99706537e-01
-1.05320990e+00 5.66382349e-01 7.77104378e-01 -4.95653600e-01
-4.87881042e-02 8.38942885e-01 -3.30124706e-01 1.85896993e-01
7.90759921e-01 4.18312460e-01 -2.99614072e-01 4.07474190e-01
-3.72219652e-01 2.46590674e-01 8.72804046e-01 1.12069559e+00
1.05295908e+00 5.03064811e-01 2.82314956e-01 8.10644925e-01
6.05784118e-01 1.08745849e+00 -4.43506420e-01 -1.63614249e+00
-1.67934179e-01 1.79945305e-01 5.62195897e-01 -6.16586387e-01
-5.03343418e-02 -2.08373755e-01 -7.39402950e-01 1.18948340e+00
3.38333398e-01 2.81382084e-01 -1.38277268e+00 1.43731260e+00
4.53213006e-01 3.56148809e-01 -1.29682153e-01 1.14476883e+00
1.06101286e+00 1.06828165e+00 -3.35757077e-01 1.10986277e-01
1.15940475e+00 -8.85112345e-01 -5.78890979e-01 -1.10350259e-01
3.22828859e-01 -7.50733614e-01 1.12508762e+00 5.20819545e-01
-1.37141395e+00 -1.24705933e-01 -8.15320909e-01 -6.94720089e-01
1.79524627e-02 -3.07216555e-01 9.83294308e-01 3.58582586e-01
-1.09408367e+00 3.47061753e-01 -9.91037548e-01 2.71428883e-01
3.68785083e-01 2.28959292e-01 -2.82698989e-01 -3.88948590e-01
-5.86662292e-01 7.67135203e-01 -5.27105451e-01 2.75160640e-01
-1.21972978e+00 -1.35458004e+00 -6.79916382e-01 -9.92109850e-02
-1.68969065e-01 -1.11990106e+00 1.38369465e+00 -1.47878897e+00
-1.71456349e+00 8.83592010e-01 -1.89136207e-01 6.60119876e-02
2.68809140e-01 -1.38487324e-01 8.36343914e-02 1.21966869e-01
-5.37555814e-01 7.04370499e-01 9.50565040e-01 -2.16784453e+00
-1.40772294e-02 -3.06252182e-01 4.98015434e-01 4.90308851e-01
1.18836850e-01 -3.44273269e-01 1.84350640e-01 -4.78303373e-01
3.93127978e-01 -7.18277693e-01 -5.70421927e-02 1.40826488e+00
-2.79849321e-01 3.92269820e-01 1.12251878e+00 -6.57848597e-01
-6.83358312e-02 -1.68162966e+00 -3.96086201e-02 -1.98221534e-01
4.24214125e-01 -5.34656122e-02 -3.00790697e-01 1.05573952e-01
2.84324568e-02 -1.88773319e-01 -2.89665312e-01 -6.67314887e-01
-1.44577086e-01 4.18477237e-01 -5.68060637e-01 5.97159743e-01
7.28219822e-02 1.02628410e+00 -8.78763437e-01 -3.04527134e-02
3.58157009e-01 1.27086532e+00 -7.24100053e-01 2.99480081e-01
-6.45720065e-01 5.90622485e-01 -3.67359102e-01 7.02199399e-01
1.13482225e+00 -1.73105285e-01 -4.01102334e-01 -3.39619458e-01
-2.17645437e-01 2.63592213e-01 -6.52337313e-01 1.37120092e+00
-1.16643620e+00 9.44122314e-01 3.33245575e-01 -3.37836027e-01
8.59747708e-01 1.52692556e-01 2.91481704e-01 -8.45118940e-01
-1.03451293e-02 1.09331943e-01 -3.40454906e-01 -2.35415205e-01
4.48801070e-01 -5.81617594e-01 7.66078711e-01 5.76661408e-01
-4.76686835e-01 -8.59506369e-01 -8.63042951e-01 -3.29118669e-02
7.72791386e-01 6.21633291e-01 -3.85880768e-01 -1.15414456e-01
-9.34910625e-02 -1.04870290e-01 2.19971165e-01 3.89024943e-01
6.22642398e-01 1.29644096e+00 -9.98581201e-02 -8.98489237e-01
-1.23035622e+00 -1.76231980e+00 -3.08358431e-01 6.93075001e-01
4.79935229e-01 6.01451457e-01 -5.82671344e-01 5.66282496e-02
1.68661296e-01 1.00702786e+00 -7.92732835e-01 -1.94643602e-01
-5.93839109e-01 -4.90603715e-01 -4.63296883e-02 2.73903400e-01
2.76007503e-01 -9.47697759e-01 -8.20021272e-01 -1.11977667e-01
-1.31896809e-01 -8.82217526e-01 -3.58438849e-01 -8.65060389e-02
-9.32509899e-01 -8.57157290e-01 -1.04168952e+00 -5.73268175e-01
9.04855072e-01 7.73641407e-01 1.47080338e+00 3.44716489e-01
-6.66182816e-01 6.71159804e-01 6.95518553e-02 -4.61184293e-01
-6.12181425e-01 -7.73033857e-01 -3.20867658e-01 3.05701226e-01
-2.71482527e-01 -6.55054629e-01 -8.03976417e-01 3.89794171e-01
-9.04169321e-01 7.25493312e-01 1.49936387e-02 5.11341512e-01
6.67497814e-01 -6.48328006e-01 2.10089106e-02 -9.13945258e-01
-7.70817623e-02 -1.52691469e-01 -8.17977548e-01 2.45649219e-01
-1.35112852e-01 -6.81628585e-02 4.33293343e-01 -7.77702153e-01
-1.62911880e+00 -2.31354207e-01 7.64449239e-02 -9.13464129e-01
-6.43483773e-02 -2.76964009e-01 2.69427061e-01 -4.46940541e-01
7.76798666e-01 1.79789111e-01 -1.33560123e-02 -4.78604048e-01
2.56914020e-01 2.62268990e-01 3.69551092e-01 -7.54939675e-01
1.01058674e+00 1.29062581e+00 2.28350028e-01 -1.27962220e+00
-7.60939419e-01 1.22621104e-01 -1.93996191e-01 -3.45440537e-01
8.39829564e-01 -8.35223138e-01 -1.11464584e+00 5.80080926e-01
-1.53237796e+00 -9.82468069e-01 -7.12127268e-01 3.10423672e-01
-8.20803285e-01 4.54970114e-02 -6.60837471e-01 -1.09611082e+00
-1.23590551e-01 -1.00497901e+00 1.46047127e+00 1.48247525e-01
1.40775181e-02 -1.14847493e+00 -6.37498358e-03 2.81509310e-01
4.33531582e-01 4.21790689e-01 1.26280010e+00 8.98071826e-01
-1.38225842e+00 1.95662156e-01 -5.48179805e-01 2.76891321e-01
3.83355647e-01 2.85978407e-01 -1.44566119e+00 -9.39942077e-02
1.61726594e-01 -4.65486199e-01 7.70453811e-01 6.88107610e-01
1.17215824e+00 -3.34313363e-01 -2.94348150e-01 1.15893924e+00
1.37266827e+00 1.67049691e-02 7.10098684e-01 -1.32800460e-01
1.21727562e+00 8.07193637e-01 2.40534738e-01 1.20862365e-01
3.95934105e-01 5.40715337e-01 1.15155268e+00 -5.95048308e-01
-8.21117461e-01 -1.34743467e-01 2.62313336e-01 4.51364666e-01
-3.55215579e-01 -5.10370910e-01 -4.70412731e-01 1.03995167e-01
-1.23423147e+00 -7.43215263e-01 -6.79557562e-01 2.32137036e+00
6.64868414e-01 -4.43645716e-01 -5.34164488e-01 -2.50087827e-01
-3.29871327e-02 -1.22545362e-01 -8.77601027e-01 -6.52004898e-01
-2.83873707e-01 1.46889731e-01 3.55984569e-01 1.14067066e+00
-1.92160890e-01 7.33154058e-01 6.26263809e+00 1.98259011e-01
-1.30697322e+00 1.18802585e-01 6.51880503e-01 -3.34677726e-01
-1.51507723e+00 3.78929973e-02 -4.26897615e-01 -1.82200968e-02
3.80605429e-01 4.65378582e-01 8.61993968e-01 3.30769807e-01
2.56419599e-01 -1.26881346e-01 -1.08037543e+00 9.34597731e-01
5.10218553e-02 -1.51748145e+00 5.95710576e-01 6.09168485e-02
1.03813040e+00 2.19231963e-01 3.46809328e-01 -2.89903402e-01
4.24761653e-01 -1.09808242e+00 1.13760519e+00 8.13649058e-01
1.07180893e+00 -3.89959604e-01 -2.12350965e-01 3.92512888e-01
-8.15454900e-01 2.47896641e-01 -6.61354721e-01 1.68840975e-01
2.09415764e-01 5.85556567e-01 -5.57258964e-01 -1.51108699e-02
6.84238911e-01 5.75817227e-01 -1.63083375e-02 7.79761910e-01
-3.66928242e-02 2.82047242e-01 -4.73844260e-01 3.50446627e-02
1.61012206e-02 -5.41796565e-01 5.10485530e-01 6.35540485e-01
2.75863260e-01 3.16082358e-01 -2.79377043e-01 1.49022520e+00
-1.59133561e-02 -3.91120702e-01 -6.87769771e-01 2.67316699e-01
-1.47274256e-01 1.00138390e+00 -4.88050848e-01 -1.83312036e-02
-3.49704683e-01 1.05554259e+00 3.80272955e-01 1.11699593e+00
-7.63579667e-01 1.03164442e-01 1.14186645e+00 7.43359268e-01
-2.10390285e-01 -4.28770304e-01 -4.36243564e-01 -1.24326217e+00
-3.00040364e-01 -3.44295591e-01 -5.71772993e-01 -1.87241650e+00
-1.14559257e+00 5.45393288e-01 6.73754420e-03 -1.20121014e+00
3.08589458e-01 -1.14090157e+00 -4.06692386e-01 1.06272531e+00
-1.92721641e+00 -1.58192194e+00 -5.02377212e-01 5.17560601e-01
5.16566873e-01 6.11223042e-01 7.81099617e-01 -1.28394902e-01
3.46394807e-01 -2.01483890e-02 1.51054516e-01 -3.65000457e-01
3.53789151e-01 -1.08620739e+00 4.76861179e-01 2.98944861e-01
1.61393002e-01 5.87812483e-01 8.24870825e-01 -4.21617448e-01
-1.95186341e+00 -8.96774709e-01 5.35524599e-02 -6.60074115e-01
1.21605918e-01 -5.90034366e-01 -1.16745031e+00 7.46115088e-01
1.84177801e-01 4.59584534e-01 1.71631455e-01 -3.10093492e-01
-8.64018798e-01 8.89137611e-02 -1.13201118e+00 9.69465315e-01
1.22901368e+00 -6.66880250e-01 -3.05810012e-02 5.13422906e-01
8.44911933e-01 -6.25024319e-01 -5.00781536e-01 3.03594500e-01
8.90626311e-01 -1.13314450e+00 1.39279962e+00 -4.18449163e-01
5.71660161e-01 -3.89834553e-01 -2.24182323e-01 -1.34143531e+00
-2.70053178e-01 -4.20707315e-01 -4.93858993e-01 5.89446425e-01
2.07965061e-01 -8.74490201e-01 8.74793231e-01 8.18088293e-01
-5.05366981e-01 -6.73087597e-01 -5.77721477e-01 -6.06499970e-01
2.94195384e-01 -3.20231497e-01 5.15459538e-01 9.19027984e-01
-1.02846968e+00 -1.31229591e-02 -3.42609435e-01 5.06370783e-01
1.15425622e+00 5.61868191e-01 5.48256636e-01 -1.43563831e+00
-4.27790970e-01 -2.14788780e-01 6.11652024e-02 -1.40663373e+00
2.19320625e-01 -8.97471011e-01 2.85598040e-01 -1.83764935e+00
1.07608862e-01 -1.06615829e+00 3.37434262e-01 2.07002878e-01
2.20626742e-01 6.50379837e-01 -1.52804166e-01 3.40439342e-02
3.03469330e-01 8.23662639e-01 2.12403512e+00 -3.48928124e-01
7.03977942e-02 6.12233859e-03 -4.05497462e-01 9.28287685e-01
3.98660123e-01 -3.69720727e-01 -7.86309958e-01 -1.30971229e+00
5.01743495e-01 2.77472436e-02 9.23554838e-01 -4.81088310e-01
-2.87337154e-01 -5.72930455e-01 5.74835300e-01 -5.01761377e-01
1.25498068e+00 -1.04657972e+00 6.27431929e-01 4.12926525e-01
-1.83622181e-01 -4.69721526e-01 1.20216012e-01 6.33930564e-01
3.93717170e-01 -1.10303730e-01 1.07106173e+00 -3.30694824e-01
-1.10841721e-01 7.02355564e-01 -2.65632510e-01 -1.37711808e-01
5.37631631e-01 -7.18304694e-01 -5.21589041e-01 -5.73280394e-01
-4.49192405e-01 -2.87528187e-01 1.05660045e+00 2.81386852e-01
1.05076110e+00 -1.17485189e+00 -6.13272250e-01 3.12562823e-01
-3.30795608e-02 4.49092954e-01 3.35352719e-01 2.12826297e-01
-1.18143022e+00 -1.91209733e-01 -8.74574929e-02 -8.12082708e-01
-1.09992099e+00 2.81112254e-01 9.18679357e-01 5.97675562e-01
-9.81663585e-01 1.09169519e+00 1.50672591e+00 -7.74353564e-01
-3.94883640e-02 -7.01244831e-01 5.17444074e-01 -7.29735136e-01
5.52410066e-01 1.87582284e-01 -1.44711405e-01 -5.06011903e-01
2.76180685e-01 1.06022203e+00 2.27964744e-01 -4.79837097e-02
1.32448161e+00 -5.26398011e-02 -1.47160947e-01 7.36932635e-01
1.16002929e+00 -1.86348885e-01 -1.93909836e+00 -6.11299239e-02
-1.04117692e+00 -7.65035748e-01 3.46076906e-01 -7.76948094e-01
-1.21436930e+00 1.42966366e+00 2.16704443e-01 1.60989799e-02
7.95913517e-01 -1.68375999e-01 7.22338200e-01 4.82819200e-01
6.36330247e-01 -2.99095094e-01 2.53387332e-01 4.61604446e-01
1.24279892e+00 -9.31107640e-01 -2.29068119e-02 -9.36466515e-01
-1.18869573e-01 1.10385716e+00 7.07372129e-01 -2.47958869e-01
8.66363108e-01 6.61540210e-01 1.86181203e-01 -2.29093686e-01
-8.63018274e-01 5.34966528e-01 3.45736533e-01 8.81983578e-01
3.99707496e-01 -8.24829563e-02 8.38524580e-01 -5.59926033e-01
5.88109791e-02 -1.95888028e-01 7.26768851e-01 5.98849356e-01
-3.75320375e-01 -5.71467459e-01 -4.42440987e-01 3.91763270e-01
1.33361280e-01 -1.00442767e-01 -1.90341741e-01 4.83575404e-01
-1.35299683e-01 3.22965771e-01 3.98848593e-01 3.31081659e-01
2.12318555e-01 -6.06833220e-01 1.13856864e+00 -7.42976964e-01
-2.46032104e-01 -7.13854060e-02 -7.58683756e-02 -4.62558657e-01
-3.22873622e-01 -4.22650278e-01 -1.40499091e+00 -4.16391551e-01
-3.81295174e-01 -3.56133908e-01 8.16147923e-01 5.29685736e-01
-2.76120193e-02 4.89524275e-01 5.40780544e-01 -1.42177010e+00
-1.26123548e-01 -3.33991379e-01 -6.98194027e-01 7.81944618e-02
1.01779461e+00 -7.61799514e-01 -5.10041952e-01 2.59614229e-01] | [9.60857105255127, -3.1497154235839844] |
3cb049bb-c1a1-4e6f-83e7-565b6ad8ce1e | hierarchical-automatic-power-plane-generation | 2210.16314 | null | https://arxiv.org/abs/2210.16314v2 | https://arxiv.org/pdf/2210.16314v2.pdf | Hierarchical Automatic Power Plane Generation with Genetic Optimization and Multilayer Perceptron | We present an automatic multilayer power plane generation method to accelerate the design of printed circuit boards (PCB). In PCB design, while automatic solvers have been developed to predict important indicators such as the IR-drop, power integrity, and signal integrity, the generation of the power plane itself still largely relies on laborious manual methods. Our automatic power plane generation approach is based on genetic optimization combined with a multilayer perceptron and is able to automatically generate power planes across a diverse set of problems with varying levels of difficulty. Our method GOMLP consists of an outer loop genetic optimizer (GO) and an inner loop multi-layer perceptron (MLP) that generate power planes automatically. The critical elements of our approach include contour detection, feature expansion, and a distance measure to enable island-minimizing complex power plane generation. We compare our approach to a baseline solution based on A*. The A* method consisting of a sequential island generation and merging process which can produce less than ideal solutions. Our experimental results show that on single layer power plane problems, our method outperforms A* in 71% of the problems with varying levels of board layout difficulty. We further describe H-GOMLP, which extends GOMLP to multilayer power plane problems using hierarchical clustering and net similarities based on the Hausdorff distance. | ['Levent Burak Kara', 'Mirko Spasojevic', 'Taylor Hogan', 'Elias Fallon', 'Devika Shanbhag', 'Xuliang Dong', 'Vinay Patil', 'Haiguang Liao'] | 2022-10-28 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [ 4.40455317e-01 1.30042121e-01 3.09781302e-02 -6.81430846e-03
-7.60488987e-01 -6.99748755e-01 -8.47864449e-02 2.81536549e-01
2.92487472e-01 6.59603536e-01 -4.53602701e-01 -3.45845848e-01
-9.42949533e-01 -8.64650130e-01 -6.32769525e-01 -6.55947983e-01
-3.96681577e-01 5.42353213e-01 3.14518601e-01 -1.74611986e-01
6.36417389e-01 7.85028934e-01 -1.27763891e+00 4.14365202e-01
8.48371208e-01 1.23178518e+00 -9.30665806e-02 9.54546273e-01
2.51063675e-01 3.87183577e-01 -1.03926313e+00 -1.91225320e-01
2.49742836e-01 -4.22251433e-01 -3.28882188e-01 1.04201861e-01
1.98296994e-01 1.58007741e-01 3.50621402e-01 9.36801910e-01
6.80735290e-01 -2.30452120e-01 7.05759764e-01 -1.51323926e+00
8.87596607e-02 6.60226524e-01 -6.23998523e-01 -2.34954879e-01
5.11363745e-01 -5.00165932e-02 1.07751024e+00 -2.61852622e-01
3.83810371e-01 9.38144207e-01 1.01220286e+00 -3.58640477e-02
-1.74919796e+00 -4.63410020e-01 -2.95258671e-01 -1.07361287e-01
-1.45701289e+00 -1.84310842e-02 1.06505263e+00 -3.16492081e-01
1.38897860e+00 5.76339841e-01 8.57114017e-01 5.66502325e-02
5.50397336e-01 4.26952809e-01 6.44096076e-01 -7.41128087e-01
6.61365688e-01 2.32627288e-01 -6.55154288e-02 7.02159286e-01
2.60407001e-01 -2.66154110e-01 -1.09052084e-01 -4.64760453e-01
8.82013917e-01 -8.45124185e-01 -1.01166978e-01 -4.88519579e-01
-8.93218160e-01 5.06407976e-01 1.72444403e-01 1.33592024e-01
-1.45493522e-01 4.41117398e-02 1.20368879e-02 2.96765000e-01
8.37658569e-02 1.46460819e+00 -5.01038313e-01 4.88867983e-02
-1.18173909e+00 3.21781993e-01 1.28505671e+00 1.00589228e+00
5.68569660e-01 3.99753936e-02 -1.00677609e-02 1.07180488e+00
2.63899595e-01 2.52107412e-01 -1.67279452e-01 -9.21171606e-01
5.69010854e-01 8.51867557e-01 -1.53653696e-01 -1.24936748e+00
-7.12916017e-01 -4.83275145e-01 -5.65278351e-01 5.95678210e-01
2.35646348e-02 -6.79217696e-01 -5.50559580e-01 8.93393338e-01
1.25268400e-01 -3.32466632e-01 -1.23548292e-01 4.03887272e-01
2.61423081e-01 1.03390670e+00 -4.85839188e-01 -1.32911772e-01
8.72137725e-01 -9.21223104e-01 -5.71413219e-01 8.21959227e-02
7.06125617e-01 -7.84273148e-01 4.40700263e-01 1.01185071e+00
-1.29369879e+00 -2.11155266e-01 -2.03252006e+00 6.27889276e-01
-2.75319308e-01 3.12603146e-01 5.94330788e-01 1.25546539e+00
-1.28499711e+00 1.13621342e+00 -4.35746998e-01 -9.07847378e-03
4.20183152e-01 9.02445495e-01 1.10388502e-01 3.01122338e-01
-4.18031991e-01 5.84362209e-01 3.44954938e-01 1.03353120e-01
-1.50684908e-01 -1.20407510e+00 -6.68234527e-01 1.14961468e-01
2.67984748e-01 -5.20844340e-01 7.70217180e-01 -9.27146614e-01
-2.07577419e+00 1.66685879e-01 5.79977989e-01 6.95797522e-03
2.71603018e-01 6.19885400e-02 -3.50093424e-01 2.61831377e-02
-2.85646766e-01 7.12624192e-01 7.46064126e-01 -1.33970523e+00
-7.76180267e-01 1.90048486e-01 -2.55452991e-01 -3.95163409e-02
-1.61770850e-01 -1.83172479e-01 -3.44792962e-01 -4.78041768e-01
3.69693749e-02 -6.85763180e-01 -3.68670285e-01 -2.32131615e-01
-8.33008051e-01 1.67129964e-01 6.63757980e-01 -7.79676735e-01
1.37389636e+00 -1.79309714e+00 9.10054669e-02 1.14069533e+00
-1.40769348e-01 -1.74174279e-01 2.68107262e-02 5.07141352e-01
-3.92997772e-01 1.69633880e-01 -2.74958938e-01 4.10326153e-01
2.56100029e-01 -3.77910703e-01 7.55582824e-02 5.26444137e-01
6.41729951e-01 5.39043546e-01 -4.07106161e-01 -2.63101637e-01
3.06372136e-01 8.56057778e-02 -1.19572270e+00 -1.26494080e-01
-3.70296210e-01 -2.88670421e-01 -1.14119463e-01 8.15046430e-01
5.59803009e-01 -2.30129734e-01 4.39224124e-01 -3.77577692e-01
-2.22872227e-01 -5.98943755e-02 -1.50276506e+00 1.24296606e+00
-4.09641773e-01 8.35813344e-01 2.05363587e-01 -7.13760138e-01
1.28739750e+00 1.08206449e-02 6.70491934e-01 -5.60881317e-01
3.55929881e-01 1.38824314e-01 1.29468903e-01 -1.16709121e-01
3.61409783e-01 1.05048977e-01 -4.43194568e-01 3.44426304e-01
6.77509308e-02 -7.05156803e-01 4.50759768e-01 -1.13492720e-01
1.73395360e+00 -3.45351025e-02 -2.03430966e-01 -7.40528107e-01
5.17701685e-01 2.14761958e-01 6.89687371e-01 3.59892935e-01
4.55956727e-01 8.12173069e-01 1.27959239e+00 -7.02904761e-02
-1.33189917e+00 -1.11383641e+00 -2.33965397e-01 5.43796420e-01
2.09456891e-01 -3.33607465e-01 -7.32132554e-01 -4.50804532e-01
-1.70323357e-01 8.95821810e-01 6.20351471e-02 1.11510724e-01
-7.06484020e-01 -1.26641083e+00 4.81395423e-01 3.39721143e-01
1.11925393e-01 -7.25620627e-01 -3.77336025e-01 4.02204543e-01
5.44744432e-01 -1.06205368e+00 1.02127962e-01 6.05388403e-01
-9.74914551e-01 -8.45309079e-01 -3.96286964e-01 -1.11356425e+00
1.14722157e+00 -7.18216658e-01 1.22058094e+00 1.04941651e-01
-9.72903550e-01 1.57096595e-01 -1.63600698e-01 -2.54811347e-01
-5.10897458e-01 8.64895806e-02 -4.90512282e-01 -5.21085203e-01
-3.98547381e-01 -7.42270589e-01 -3.19373041e-01 5.34520984e-01
-6.92016363e-01 1.57627106e-01 9.59212899e-01 5.66813707e-01
3.91293585e-01 9.93111789e-01 4.12609488e-01 -4.14954841e-01
7.13124812e-01 -2.81835586e-01 -1.21271813e+00 2.39224762e-01
-2.26848304e-01 3.87636453e-01 7.21351922e-01 -3.07463169e-01
-6.29879653e-01 6.42207190e-02 -2.28757158e-01 6.54933602e-03
9.08485875e-02 2.06179228e-02 -7.18560576e-01 -5.67018569e-01
3.72381181e-01 -4.88003165e-01 -2.49441981e-01 -2.79961377e-01
6.11328967e-02 2.61726022e-01 7.90139958e-02 -5.69467187e-01
8.81363273e-01 6.28906190e-02 3.26896846e-01 -1.19448495e+00
2.61027187e-01 1.50440082e-01 -5.38058698e-01 -4.15365875e-01
4.97849435e-01 -1.20018542e-01 -8.42125475e-01 3.50134224e-01
-1.18831003e+00 -5.53087294e-01 -2.13756472e-01 7.22999647e-02
-4.97903883e-01 8.49900171e-02 -6.18649781e-01 -6.29154801e-01
-4.66014892e-01 -1.06669950e+00 1.08027434e+00 1.41379520e-01
-6.30865037e-01 -1.03501248e+00 1.19979247e-01 1.04544565e-01
1.25043929e-01 5.10507643e-01 1.72914922e+00 -4.66021866e-01
-5.77506125e-01 -4.45076853e-01 -2.46657822e-02 4.25870776e-01
-8.51798281e-02 4.20966834e-01 -4.08000290e-01 -2.17430636e-01
-3.12141269e-01 2.52279580e-01 2.47573927e-02 5.67821503e-01
1.01290548e+00 -3.61694694e-01 -5.84968507e-01 6.98036134e-01
1.82413208e+00 6.44349158e-01 9.37502563e-01 2.70371109e-01
3.61515939e-01 5.73099554e-01 2.38901436e-01 4.72002327e-01
-6.90615997e-02 7.21602798e-01 2.64633238e-01 -3.23354751e-01
4.34975140e-02 1.20509706e-01 1.93461120e-01 9.10192847e-01
3.76151919e-01 -3.15691769e-01 -9.01646435e-01 4.22376424e-01
-1.38011110e+00 -4.40281689e-01 3.11427303e-02 1.90732861e+00
6.35390043e-01 6.06333613e-01 1.18382335e-01 7.48069584e-01
8.61503124e-01 -3.13509643e-01 -3.20954382e-01 -9.13711429e-01
1.99515283e-01 4.68919665e-01 9.47350919e-01 5.38728833e-01
-9.84209120e-01 2.30491921e-01 6.80572414e+00 9.59440649e-01
-9.84453022e-01 -7.20808327e-01 7.33861864e-01 -4.63279784e-01
-3.97445679e-01 -2.96642125e-01 -7.96618402e-01 3.77017379e-01
5.31011343e-01 1.22904889e-02 3.06701303e-01 6.59001470e-01
-6.49736077e-02 -2.08785191e-01 -1.40375614e+00 9.37919676e-01
9.96650606e-02 -1.30791652e+00 -3.60518992e-01 1.98679231e-02
1.04459071e+00 -7.83599734e-01 -7.44067803e-02 -4.13812138e-02
3.39022756e-01 -9.73407209e-01 6.17773533e-01 3.20500314e-01
4.38724309e-01 -1.03445601e+00 5.84432662e-01 -3.32507677e-02
-1.06103826e+00 -2.05323771e-01 2.58218888e-02 2.45454744e-01
2.07323357e-01 1.09098971e+00 -1.04017723e+00 6.57202899e-01
6.25154674e-01 2.08567068e-01 -3.83570343e-01 1.51785553e+00
-1.65590700e-02 2.82182634e-01 -7.61542261e-01 -6.00968361e-01
-8.39680508e-02 -1.85819328e-01 7.31519341e-01 9.45189893e-01
5.27984560e-01 -2.03131646e-01 -1.08229093e-01 9.57034647e-01
2.64593005e-01 2.06749007e-01 -8.41355994e-02 1.50065497e-01
4.97511953e-01 1.31211984e+00 -1.06842554e+00 3.24494988e-01
3.94199975e-02 5.79217434e-01 -3.26693475e-01 2.29172379e-01
-1.00296926e+00 -1.19152820e+00 6.00831568e-01 4.41273123e-01
4.50206131e-01 -1.15227997e-01 -9.68319654e-01 -1.70722857e-01
2.66572446e-01 -7.91271687e-01 7.31203705e-02 -7.32610285e-01
-1.10088098e+00 4.84391868e-01 3.18953544e-02 -1.11585736e+00
-3.69671822e-01 -9.89458978e-01 -8.34544063e-01 4.79629129e-01
-1.07264042e+00 -5.46982288e-01 9.16646570e-02 -5.98980002e-02
4.26345766e-01 -2.74245322e-01 3.78139794e-01 4.06180352e-01
-6.57800555e-01 6.08322740e-01 4.09940779e-02 -1.26369104e-01
2.24723235e-01 -1.37562096e+00 2.29014650e-01 5.33650935e-01
-4.14813876e-01 -2.78542303e-02 7.99177945e-01 -7.59374738e-01
-1.90142047e+00 -8.25475633e-01 2.38081768e-01 -5.24063855e-02
7.67448127e-01 -5.35088360e-01 -3.62530947e-01 9.71704274e-02
2.16671929e-01 -6.10797584e-01 6.94419801e-01 -3.02766338e-02
4.14245963e-01 -6.53041542e-01 -1.26268303e+00 8.30490470e-01
6.65534079e-01 2.79037833e-01 -1.59823194e-01 1.89770117e-01
3.10083330e-01 -1.09420098e-01 -1.19567001e+00 7.81422555e-01
5.39674580e-01 -8.49621236e-01 8.88504505e-01 3.35988760e-01
4.24176604e-01 -4.60882485e-01 1.83874562e-01 -1.22507179e+00
-4.53254730e-01 -9.50842261e-01 3.73933852e-01 1.19817841e+00
9.67516959e-01 -5.62273622e-01 1.09698081e+00 4.15751755e-01
-2.87453532e-01 -8.47071409e-01 -8.46488535e-01 -6.71091080e-01
-2.61018038e-01 -4.43718731e-01 6.18916273e-01 5.43525517e-01
4.02728200e-01 1.74119040e-01 2.44293362e-01 4.77905899e-01
7.23871112e-01 -6.79860264e-02 5.12672424e-01 -1.29249442e+00
-7.47702718e-01 -9.51001108e-01 -7.50451386e-01 -7.36051917e-01
-2.25688502e-01 -5.74578106e-01 2.41246089e-01 -1.39842498e+00
-3.07769150e-01 -8.01721215e-01 9.42435935e-02 2.64501393e-01
4.15116012e-01 1.68927878e-01 2.81105959e-03 -5.22515595e-01
-5.77239931e-01 1.23041749e-01 7.93009937e-01 -2.65144467e-01
-7.43981659e-01 -1.92690015e-01 -4.76869255e-01 7.82326639e-01
8.09213519e-01 -4.12656039e-01 -1.48973688e-01 -1.31986767e-01
8.86329472e-01 1.57572046e-01 1.18339742e-02 -1.71248972e+00
4.44269627e-01 3.44941765e-01 7.04315603e-01 -6.61807656e-01
1.94022313e-01 -9.02068496e-01 3.48378867e-01 4.56509531e-01
6.83545768e-02 2.19226152e-01 7.78389752e-01 -9.63358358e-02
2.10945215e-02 -2.83350021e-01 6.61979139e-01 3.50453079e-01
-2.65085280e-01 -4.65434879e-01 -6.76857531e-01 -5.13618052e-01
1.17962158e+00 -4.67388988e-01 -2.83333182e-01 -5.49294129e-02
-3.68119359e-01 4.04203206e-01 4.11990583e-01 2.75928676e-01
5.03245473e-01 -1.32190824e+00 -5.30640781e-01 6.06852412e-01
-4.98327523e-01 -3.33650231e-01 -1.46037107e-03 5.68467021e-01
-1.21420836e+00 2.77529955e-01 -3.14638585e-01 -8.32730114e-01
-1.33702803e+00 1.46565467e-01 4.24304754e-01 -5.05131721e-01
-3.79111975e-01 7.84334779e-01 -2.45421454e-01 -2.33361557e-01
1.11658834e-01 -4.51641172e-01 1.27701955e-02 -7.78641328e-02
1.56998724e-01 7.61447251e-01 4.56996828e-01 1.02955952e-01
-3.89474928e-01 8.71146858e-01 2.53376722e-01 -3.83551925e-01
1.48584878e+00 4.91383672e-01 -4.45970446e-01 3.37709188e-02
1.21457767e+00 4.85974224e-03 -9.27911162e-01 8.51496816e-01
1.47911444e-01 -1.84503630e-01 2.90704906e-01 -9.22081649e-01
-1.09593773e+00 2.54462749e-01 4.75092351e-01 5.03123343e-01
1.58420491e+00 -3.20776761e-01 5.93818963e-01 4.04886872e-01
4.61135268e-01 -1.43551421e+00 1.65164426e-01 1.00809321e-01
8.15783083e-01 -2.73427993e-01 5.29698849e-01 -8.20281684e-01
-6.96603656e-02 1.28934133e+00 4.03684467e-01 -3.49648625e-01
7.03888178e-01 1.29918313e+00 -2.20334843e-01 -6.68875426e-02
-8.29415262e-01 4.89251465e-01 4.08823818e-01 4.53696877e-01
6.42808825e-02 -2.12190241e-01 -2.70450991e-02 7.59348989e-01
-5.43933570e-01 -3.79553974e-01 2.85294443e-01 1.16821563e+00
-5.13003707e-01 -1.22472799e+00 -8.88321042e-01 6.22125506e-01
-2.70506233e-01 1.93497837e-01 -4.47990477e-01 7.17644334e-01
5.60166314e-02 9.38304901e-01 3.12326670e-01 -5.32620490e-01
6.42489672e-01 -3.15667123e-01 8.43485117e-01 -2.49882936e-01
-8.39451432e-01 1.67700142e-01 4.89121526e-01 -6.09471619e-01
5.16212821e-01 -4.44153458e-01 -9.68942046e-01 -1.49514228e-01
-5.07135391e-01 -2.80663371e-02 9.71030235e-01 5.67310274e-01
1.98917046e-01 1.15973663e+00 7.87257075e-01 -1.01315892e+00
1.93107948e-01 -2.80654609e-01 -7.81881332e-01 -3.91841382e-01
-8.08110982e-02 -3.53874356e-01 -2.21327782e-01 -2.36606762e-01] | [5.87297248840332, 3.409100294113159] |
77d59747-c91e-4a7d-8339-f9984267c35e | few-shot-camouflaged-animal-detection-and | 2304.07444 | null | https://arxiv.org/abs/2304.07444v1 | https://arxiv.org/pdf/2304.07444v1.pdf | Few-shot Camouflaged Animal Detection and Segmentation | Camouflaged object detection and segmentation is a new and challenging research topic in computer vision. There is a serious issue of lacking data of camouflaged objects such as camouflaged animals in natural scenes. In this paper, we address the problem of few-shot learning for camouflaged object detection and segmentation. To this end, we first collect a new dataset, CAMO-FS, for the benchmark. We then propose a novel method to efficiently detect and segment the camouflaged objects in the images. In particular, we introduce the instance triplet loss and the instance memory storage. The extensive experiments demonstrated that our proposed method achieves state-of-the-art performance on the newly collected dataset. | ['Tam V. Nguyen', 'Minh-Triet Tran', 'Thanh-Toan Do', 'Thanh Duc Ngo', 'Vinh-Tiep Nguyen', 'Nhat-Duy Nguyen', 'Anh-Khoa Nguyen Vu', 'Thanh-Danh Nguyen'] | 2023-04-15 | null | null | null | null | ['camouflaged-object-segmentation'] | ['computer-vision'] | [ 3.56210321e-01 -4.60475832e-01 -1.69961333e-01 -3.94979212e-03
-5.58212876e-01 -4.31297392e-01 2.65335172e-01 -2.72041142e-01
-5.21216393e-01 7.48875201e-01 -4.36935484e-01 -1.79923289e-02
2.47900784e-01 -4.90954995e-01 -7.62348652e-01 -8.30349267e-01
2.47796580e-01 1.01045616e-01 8.22684765e-01 1.12962209e-01
3.43394727e-01 3.71657372e-01 -1.40480971e+00 -6.39420897e-02
9.24420118e-01 1.15103483e+00 2.91193306e-01 3.87185574e-01
2.68021859e-02 6.09412074e-01 -8.10548127e-01 -3.35225403e-01
6.03276789e-01 -3.34318578e-01 -4.80596453e-01 5.15998840e-01
7.84737051e-01 -4.18953091e-01 -3.24720234e-01 1.49530447e+00
2.00262949e-01 2.74723053e-01 4.64638740e-01 -1.43384922e+00
-7.45064437e-01 9.75905135e-02 -8.75309169e-01 4.61103588e-01
-2.02893555e-01 5.06826222e-01 3.77949446e-01 -8.90954494e-01
6.51685655e-01 1.22652864e+00 2.93794543e-01 8.09682488e-01
-9.60081458e-01 -9.19231415e-01 2.31393769e-01 5.34105301e-01
-1.27269042e+00 -2.26777568e-01 7.28859305e-01 -4.68123555e-01
1.66199997e-01 1.75691247e-01 7.75917292e-01 6.22319520e-01
1.62920237e-01 1.00997043e+00 1.16695178e+00 -1.97245881e-01
3.71427089e-01 -8.30297768e-02 3.49995911e-01 7.69360542e-01
6.65086687e-01 4.04606491e-01 4.07917686e-02 -1.25961319e-01
5.31523049e-01 3.94606829e-01 -3.92738044e-01 -3.93134713e-01
-9.62932467e-01 8.94992948e-01 5.02224386e-01 -1.08253799e-01
1.83397651e-01 1.27075627e-01 6.58795655e-01 1.11447088e-01
5.61297834e-01 4.20146853e-01 -6.07805610e-01 8.30858722e-02
-7.27875590e-01 2.85069078e-01 6.08143747e-01 9.23339725e-01
6.69191957e-01 1.04389399e-01 -3.56696874e-01 8.54335606e-01
-1.09031975e-01 6.72687531e-01 1.48612037e-01 -7.19313025e-01
7.67049938e-02 3.44028771e-01 2.19261348e-01 -1.04953718e+00
1.49716198e-01 1.43725649e-01 -7.62801230e-01 2.68792152e-01
2.13702261e-01 -2.46231005e-01 -1.46031809e+00 1.31933224e+00
7.11343527e-01 1.15579391e+00 -2.24518806e-01 1.25542641e+00
9.05937791e-01 7.72151649e-01 4.03128229e-02 -4.52249259e-01
1.35454440e+00 -1.43379915e+00 -8.31242919e-01 -3.83675426e-01
3.70662659e-01 -8.87881219e-01 9.61981893e-01 3.14896673e-01
-7.04934478e-01 -5.90304732e-01 -1.03528631e+00 -4.25642394e-02
-4.55317408e-01 -7.91545734e-02 8.73177588e-01 6.77624702e-01
-2.50922114e-01 3.11212063e-01 -7.61298954e-01 4.10200022e-02
1.12942994e+00 1.28777459e-01 -1.50413573e-01 -2.89929301e-01
-8.59841704e-01 7.72105217e-01 6.15602553e-01 8.76100957e-02
-1.09466517e+00 -7.61165202e-01 -8.95991325e-01 1.07120094e-03
9.78564501e-01 -1.98782071e-01 1.15353048e+00 -1.06163847e+00
-1.04331720e+00 9.33953643e-01 1.03085645e-01 -5.61429024e-01
4.09190983e-01 -3.08611333e-01 -4.26862746e-01 1.78746656e-01
2.57398598e-02 6.60011590e-01 1.20362794e+00 -1.33903098e+00
-1.01168060e+00 -2.16576442e-01 -9.31717157e-02 -2.19712645e-01
-1.02282554e-01 -1.42933810e-02 -4.54996675e-01 -8.45291138e-01
-3.43249649e-01 -9.35054660e-01 -4.00912017e-01 4.34894770e-01
-3.96637380e-01 -2.56330401e-01 1.62785232e+00 -4.08108890e-01
8.54223728e-01 -2.38414502e+00 -1.70057312e-01 -3.57413381e-01
2.57593095e-01 1.14126861e+00 -1.66912273e-01 -3.23696673e-01
2.33707473e-01 -1.18522555e-01 -4.02551115e-01 -1.31972462e-01
-4.00180072e-01 2.67693341e-01 -5.24542511e-01 8.32546413e-01
2.92175472e-01 8.12362432e-01 -9.69467878e-01 -7.50646949e-01
5.39477348e-01 2.37842456e-01 -4.57759440e-01 2.10491598e-01
-5.37063718e-01 6.05249763e-01 -3.87035251e-01 9.37645495e-01
1.32025027e+00 -6.43132478e-02 -4.57027227e-01 9.85448435e-03
3.08846198e-02 -4.83340383e-01 -9.57358062e-01 1.34472966e+00
4.78573218e-02 5.84021449e-01 -7.49456063e-02 -8.93594801e-01
6.04579151e-01 -1.83483317e-01 2.43363768e-01 -5.36738753e-01
7.04121947e-01 2.84414530e-01 -2.83499658e-02 -6.04539514e-01
5.01720130e-01 -7.32170716e-02 -1.49802063e-02 4.36172858e-02
5.31439818e-02 -2.65545785e-01 4.01185900e-01 -4.91018146e-02
8.40462863e-01 -2.15485245e-01 3.77750963e-01 -2.54523754e-01
5.49535573e-01 2.71238983e-01 8.04814756e-01 5.91762960e-01
-6.28272474e-01 4.17063534e-01 2.18183100e-01 -5.00073075e-01
-9.66525912e-01 -9.38352346e-01 -1.61354348e-01 7.99284697e-01
8.46314013e-01 1.31522700e-01 -8.75039816e-01 -8.51972520e-01
1.23829968e-01 4.29804981e-01 -5.65408349e-01 -4.48480308e-01
-6.45016968e-01 -4.56968397e-01 2.73072630e-01 2.42289051e-01
8.58909726e-01 -1.11028218e+00 -7.62803078e-01 -1.71697795e-01
2.07620859e-02 -1.07597554e+00 -7.95080781e-01 -2.33355001e-01
-7.40442216e-01 -1.32532334e+00 -8.46738994e-01 -1.23582828e+00
7.84229338e-01 9.67186809e-01 5.83346009e-01 3.49043608e-01
-7.24278092e-01 -2.38576993e-01 -3.70683193e-01 -5.23231804e-01
2.52418593e-02 -2.60179043e-01 -1.37602121e-01 1.24794982e-01
3.00287187e-01 -4.50601690e-02 -7.68738210e-01 3.86061281e-01
-1.02741575e+00 -1.69616431e-01 4.92451698e-01 1.05723739e+00
9.36060429e-01 3.72453183e-02 3.58880162e-01 -1.08217824e+00
2.46500835e-01 -3.21394682e-01 -1.12376666e+00 1.39529899e-01
-3.58039528e-01 -4.70678180e-01 6.51624739e-01 -9.60765719e-01
-8.40193272e-01 2.83643752e-01 1.29458368e-01 -9.63732362e-01
-2.56110370e-01 -4.01366465e-02 -2.05511525e-01 -5.58387339e-01
1.99845985e-01 5.78338504e-02 -1.53211638e-01 -6.18292332e-01
5.75586259e-01 6.25586152e-01 7.64149308e-01 -1.42314479e-01
9.19491768e-01 8.58224988e-01 -9.08801239e-03 -7.65887618e-01
-1.11901796e+00 -6.60358608e-01 -2.76061922e-01 -8.94846693e-02
8.56357694e-01 -6.80411339e-01 -7.31931031e-01 7.99729884e-01
-1.14389217e+00 -1.24315284e-01 -3.70836884e-01 2.51854599e-01
-3.24395865e-01 5.91316044e-01 -4.70145464e-01 -7.42532670e-01
-4.53513056e-01 -1.10087466e+00 1.05016172e+00 5.49922109e-01
5.78520775e-01 -7.11181819e-01 5.42277843e-02 5.06717503e-01
4.40925434e-02 3.33556563e-01 5.99921227e-01 -5.88923872e-01
-9.28358853e-01 -1.91536307e-01 -6.92055047e-01 5.14461100e-01
2.21017554e-01 -2.02506036e-01 -7.09967136e-01 -4.74826723e-01
2.81662643e-01 -3.28186989e-01 1.17930973e+00 3.99757117e-01
1.57543063e+00 -3.29262942e-01 -4.49765801e-01 7.28866339e-01
1.57502937e+00 5.10231614e-01 7.41099477e-01 -1.55017331e-01
9.57340539e-01 1.66386276e-01 1.30930746e+00 2.36602738e-01
-2.37693578e-01 3.62849146e-01 4.36672121e-01 -2.22579390e-01
-2.15949461e-01 -1.54125527e-01 -1.49667695e-01 6.01651549e-01
4.39717591e-01 -3.03363323e-01 -4.98204619e-01 7.02257752e-01
-1.92977643e+00 -9.01280463e-01 -2.57150292e-01 1.94359326e+00
6.05074704e-01 2.31538489e-01 -1.45310014e-01 6.11929260e-02
1.17839527e+00 3.29904258e-01 -8.20087016e-01 -8.28414708e-02
3.44440304e-02 3.09482455e-01 8.24588060e-01 1.92014650e-01
-1.60113728e+00 1.37713313e+00 6.21337318e+00 1.29981661e+00
-1.20174527e+00 4.84913588e-01 7.23904908e-01 5.65906093e-02
4.96062517e-01 -4.11470383e-02 -7.80692756e-01 9.80292022e-01
1.61869004e-01 -1.45630926e-01 3.82765085e-01 1.18143988e+00
-3.19173783e-02 -4.28385228e-01 -7.66386390e-01 1.14200795e+00
3.88194978e-01 -1.38585615e+00 2.44643248e-04 -9.91207361e-02
1.19629395e+00 -2.49766678e-01 1.69002756e-01 1.59309104e-01
7.45710954e-02 -7.72526026e-01 3.89603376e-01 2.13633478e-01
9.39416111e-01 -7.39998400e-01 6.81713820e-01 2.09538221e-01
-1.09908485e+00 -1.89209014e-01 -6.88928723e-01 -2.28447840e-01
2.34344229e-01 6.74691975e-01 -4.47389454e-01 4.60402146e-02
4.69143957e-01 6.98804796e-01 -5.95231593e-01 1.70505011e+00
-8.33042637e-02 5.90432644e-01 -5.44663295e-02 -1.75653070e-01
2.37965047e-01 -1.50261924e-01 4.90145445e-01 7.97985673e-01
2.38592234e-02 2.20681056e-01 5.99291444e-01 6.78438008e-01
-2.75440514e-01 -6.59233779e-02 -5.67946553e-01 -5.77194877e-02
4.60182458e-01 1.26549327e+00 -9.83891487e-01 -5.96773207e-01
-4.02175844e-01 9.90733147e-01 5.89835718e-02 7.84381181e-02
-1.11263192e+00 -6.07164562e-01 7.44280756e-01 -1.41997769e-01
6.60124719e-01 -4.86092940e-02 -2.06432134e-01 -1.27598512e+00
-2.48986319e-01 -6.79417670e-01 3.71110350e-01 -5.89909315e-01
-1.47630572e+00 4.54532579e-02 -6.38329536e-02 -1.27951956e+00
4.52319622e-01 -7.56056786e-01 -7.07614064e-01 2.45849088e-01
-1.48907304e+00 -1.12058675e+00 -4.57166970e-01 4.67233717e-01
9.16110396e-01 -5.40326722e-02 2.62572467e-01 5.58990479e-01
-4.51141357e-01 3.29144150e-01 1.73455387e-01 2.11970747e-01
6.49757683e-01 -7.84999251e-01 4.93237376e-01 1.19496727e+00
-3.13948020e-02 3.05854142e-01 7.60882497e-01 -9.41253662e-01
-1.38832116e+00 -1.35861969e+00 1.48105338e-01 -3.68343890e-02
5.48343182e-01 -2.60543466e-01 -9.37110305e-01 4.84246552e-01
-6.34168908e-02 4.83346105e-01 1.27280980e-01 -5.93558669e-01
-2.16017112e-01 -1.28620779e-02 -1.58511162e+00 5.82037389e-01
1.19102144e+00 -2.90460587e-02 -7.37558007e-01 5.24530530e-01
1.15604007e+00 -5.62224209e-01 -6.27663583e-02 4.51308787e-01
3.04247946e-01 -4.79301035e-01 9.21029150e-01 -6.66555643e-01
2.94061571e-01 -4.13047075e-01 4.40227389e-02 -1.20170248e+00
-2.24557705e-02 -4.80447114e-01 -2.59003311e-01 9.89594340e-01
-2.94419378e-01 -6.63106680e-01 8.31604958e-01 2.18482584e-01
-1.84499353e-01 -7.15530217e-01 -9.57692742e-01 -8.81793559e-01
-2.79945135e-02 1.67999715e-02 3.88602853e-01 8.86495829e-01
-5.65922081e-01 5.57581009e-03 -7.29966402e-01 -3.24907154e-02
8.57531667e-01 3.04871678e-01 7.38276422e-01 -1.20005691e+00
2.07985006e-02 2.46117241e-04 -7.07560956e-01 -1.10057807e+00
1.56649008e-01 -5.47899902e-01 4.20099258e-01 -1.13040900e+00
4.97692674e-01 -2.65293270e-01 -2.51609027e-01 2.26596013e-01
-4.78924066e-01 6.79363549e-01 3.62206399e-01 2.59894542e-02
-8.64830017e-01 4.67397958e-01 1.64921963e+00 -4.28419262e-01
-2.48599220e-02 -1.22570090e-01 -5.13617575e-01 8.54558170e-01
8.17345440e-01 -6.90467179e-01 -2.59038836e-01 -3.26450229e-01
-6.44266963e-01 -2.11865708e-01 6.40925229e-01 -8.66190016e-01
2.35746950e-02 -5.70763052e-01 1.28387809e-01 -8.16915035e-01
4.52644914e-01 -7.71458924e-01 -2.97433197e-01 9.18684244e-01
1.59379408e-01 -5.55467248e-01 1.55455261e-01 8.87883961e-01
1.37875546e-02 -2.27804825e-01 1.44605672e+00 -8.03901851e-02
-1.22377729e+00 6.42985821e-01 -2.06325278e-01 4.08586353e-01
1.45935392e+00 -7.84741491e-02 -7.99021721e-01 3.36623818e-01
-1.57500207e-01 4.12244499e-01 5.56079149e-01 5.54456890e-01
8.79458070e-01 -1.22418904e+00 -3.12836945e-01 2.45566368e-01
1.52386293e-01 4.82667014e-02 3.99639338e-01 5.54984450e-01
-6.17758214e-01 1.98634222e-01 -5.08034885e-01 -4.33121860e-01
-1.60422683e+00 1.16457343e+00 1.06788434e-01 3.07089984e-01
-6.83358133e-01 9.17999864e-01 3.89043748e-01 -4.93065976e-02
1.97008416e-01 -3.66725713e-01 3.34005803e-02 -2.68669784e-01
7.05333531e-01 5.92449725e-01 -3.51287782e-01 -5.71091473e-01
-3.13017190e-01 5.39418280e-01 -2.18636945e-01 4.04513091e-01
9.71667945e-01 8.03500861e-02 -1.80170268e-01 3.17569524e-01
1.29347646e+00 -2.76747376e-01 -1.39487076e+00 -1.70470104e-01
-1.51793867e-01 -1.27781641e+00 -4.55932021e-02 -6.90492868e-01
-1.43176734e+00 7.70103037e-01 9.86901581e-01 -6.69924635e-03
9.84316826e-01 5.11004440e-02 1.53644562e+00 4.57308590e-01
5.20430684e-01 -1.10968816e+00 1.79162443e-01 2.70904869e-01
3.85873586e-01 -1.33377588e+00 5.24868891e-02 -8.75319839e-01
-3.85237932e-01 3.28060001e-01 8.79024804e-01 -7.22771287e-01
6.72019243e-01 1.83570638e-01 3.99035327e-02 -1.20123737e-01
-4.68619525e-01 -4.42680866e-01 8.36725682e-02 6.39525294e-01
-3.92139226e-01 1.92319900e-01 -7.36382961e-01 4.31063682e-01
3.09154510e-01 2.30316579e-01 4.70448732e-01 9.93265510e-01
-8.33570659e-01 -7.01143503e-01 -4.64062452e-01 7.53554344e-01
-5.67902744e-01 1.56084329e-01 -4.47707981e-01 5.63318491e-01
6.29215479e-01 9.91583824e-01 5.42711578e-02 -1.80181861e-01
1.45881981e-01 -4.98617530e-01 5.73460639e-01 -6.64770424e-01
-1.76709041e-01 7.92276412e-02 -3.50535423e-01 -3.10157269e-01
-2.99925417e-01 -3.26580316e-01 -1.15194964e+00 -2.17931882e-01
-7.70358562e-01 4.15476784e-02 5.10694146e-01 9.27360117e-01
1.24897696e-01 3.49629313e-01 6.58985555e-01 -8.98735046e-01
-5.29506564e-01 -7.45020390e-01 -6.90157712e-01 5.06017923e-01
4.03007418e-01 -9.50439870e-01 -4.48094249e-01 1.97706327e-01] | [9.663321495056152, -0.18220816552639008] |
608be0ca-19a2-4ae8-a9c8-b8f141b1942b | relight-my-nerf-a-dataset-for-novel-view | 2304.10448 | null | https://arxiv.org/abs/2304.10448v1 | https://arxiv.org/pdf/2304.10448v1.pdf | ReLight My NeRF: A Dataset for Novel View Synthesis and Relighting of Real World Objects | In this paper, we focus on the problem of rendering novel views from a Neural Radiance Field (NeRF) under unobserved light conditions. To this end, we introduce a novel dataset, dubbed ReNe (Relighting NeRF), framing real world objects under one-light-at-time (OLAT) conditions, annotated with accurate ground-truth camera and light poses. Our acquisition pipeline leverages two robotic arms holding, respectively, a camera and an omni-directional point-wise light source. We release a total of 20 scenes depicting a variety of objects with complex geometry and challenging materials. Each scene includes 2000 images, acquired from 50 different points of views under 40 different OLAT conditions. By leveraging the dataset, we perform an ablation study on the relighting capability of variants of the vanilla NeRF architecture and identify a lightweight architecture that can render novel views of an object under novel light conditions, which we use to establish a non-trivial baseline for the dataset. Dataset and benchmark are available at https://eyecan-ai.github.io/rene. | ['Samuele Salti', 'Luigi Di Stefano', 'Daniele De Gregorio', 'Riccardo Spezialetti', 'Riccardo De Matteo', 'Marco Toschi'] | 2023-04-20 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Toschi_ReLight_My_NeRF_A_Dataset_for_Novel_View_Synthesis_and_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Toschi_ReLight_My_NeRF_A_Dataset_for_Novel_View_Synthesis_and_CVPR_2023_paper.pdf | cvpr-2023-1 | ['image-relighting'] | ['computer-vision'] | [ 2.45858133e-01 -2.10758656e-01 4.53823566e-01 -3.63018721e-01
-4.48564798e-01 -1.01866817e+00 7.52369046e-01 -7.23789811e-01
-7.58528337e-02 3.67783964e-01 2.22552996e-02 -1.32988229e-01
1.55245051e-01 -6.04940891e-01 -1.25855148e+00 -5.91290355e-01
1.67233273e-01 1.72076225e-01 -1.09620295e-01 -2.69656122e-01
9.94698405e-02 6.77340031e-01 -1.74055195e+00 2.78172761e-01
2.86186993e-01 8.82454336e-01 4.43583101e-01 9.90430653e-01
6.63882077e-01 8.89598310e-01 -2.99695045e-01 -4.74342912e-01
1.03760374e+00 1.19454160e-01 -4.39160913e-01 1.65920779e-01
1.39880908e+00 -7.95431316e-01 -6.21957600e-01 6.65396631e-01
4.53644007e-01 3.28183472e-01 3.06314081e-01 -1.40082443e+00
-9.71438468e-01 7.47547427e-04 -6.47674143e-01 7.12803230e-02
6.11810803e-01 7.48432398e-01 1.05829322e+00 -1.08440423e+00
9.89986837e-01 1.15820718e+00 6.50962234e-01 5.49842358e-01
-1.36486626e+00 -5.08882701e-01 1.35444090e-01 -2.84906998e-02
-1.16946983e+00 -7.39969969e-01 8.04605305e-01 -3.82494956e-01
9.24257636e-01 2.26563543e-01 8.20692778e-01 1.61256278e+00
3.16570193e-01 4.58055317e-01 1.24327707e+00 -7.41522312e-02
2.59950191e-01 -5.80412038e-02 -2.03839183e-01 6.73555136e-01
2.14892462e-01 5.10004759e-01 -7.36663461e-01 1.34204432e-01
7.71467328e-01 1.20565161e-01 -7.42648304e-01 -6.78047419e-01
-1.30411172e+00 4.01405573e-01 7.83980310e-01 -3.45762223e-01
-2.26430446e-01 4.61976439e-01 -2.31964335e-01 1.61053315e-01
2.45461956e-01 4.88790393e-01 -3.98408234e-01 2.93900281e-01
-2.22959965e-01 4.57934260e-01 5.52923918e-01 1.25620961e+00
8.95162463e-01 1.48228437e-01 2.59649791e-02 3.78275722e-01
2.17121437e-01 7.81462550e-01 -2.24043295e-01 -1.40543365e+00
4.49650347e-01 1.96223110e-01 4.29606736e-01 -8.26521933e-01
-5.01371264e-01 -4.08162266e-01 -4.91648912e-01 5.63517153e-01
2.76669651e-01 -1.68932259e-01 -9.56131041e-01 1.72099125e+00
4.87875134e-01 2.44533390e-01 3.93419303e-02 1.28811097e+00
8.45642090e-01 7.20744431e-01 -5.26189685e-01 1.31155536e-01
1.30149472e+00 -1.13656068e+00 -3.61594647e-01 -4.12070602e-01
1.96554646e-01 -7.84026742e-01 1.43120396e+00 7.80560613e-01
-1.06987941e+00 -3.58708382e-01 -8.95606279e-01 -6.53818369e-01
-2.93932050e-01 1.96343794e-01 7.23723352e-01 3.46089721e-01
-1.11954176e+00 7.92571157e-02 -6.16905630e-01 -2.84172416e-01
3.83830637e-01 -2.37607092e-01 -4.32169884e-01 -5.80522537e-01
-6.40489221e-01 7.49535680e-01 7.24059530e-03 4.59644020e-01
-1.42289603e+00 -1.10477352e+00 -7.59087324e-01 -2.01374859e-01
5.91925859e-01 -1.00768459e+00 1.24157441e+00 -6.81180179e-01
-1.54692471e+00 8.32578421e-01 1.67799830e-01 -1.39451876e-01
5.83574533e-01 -5.44621170e-01 -2.01603353e-01 2.53613085e-01
-1.58334643e-01 7.24781275e-01 8.46916378e-01 -1.81501925e+00
-3.45904708e-01 -2.48964980e-01 6.58752680e-01 2.93715417e-01
3.17928225e-01 -1.57679945e-01 -3.63393068e-01 -4.21470582e-01
-9.86347720e-02 -1.08987916e+00 4.71151024e-02 3.97245944e-01
-6.62955880e-01 2.27528334e-01 7.77375281e-01 -3.06877226e-01
3.34636778e-01 -2.01875544e+00 1.37457341e-01 -1.74140036e-01
3.41558278e-01 -1.65277973e-01 -3.61477703e-01 4.69912618e-01
-2.15532511e-01 -1.73739702e-01 -1.65138975e-01 -5.78259647e-01
2.19842419e-02 2.45858908e-01 -6.62382185e-01 6.03308260e-01
4.52603512e-02 9.82827544e-01 -8.77640724e-01 9.73979663e-03
3.87000591e-01 8.24895442e-01 -7.65728176e-01 3.28918815e-01
-5.31122088e-01 7.83250093e-01 -1.54743016e-01 7.70248950e-01
9.37967956e-01 -7.73149431e-02 -2.22676411e-01 -5.86374283e-01
-2.99102038e-01 -1.26957029e-01 -1.08293569e+00 2.13963866e+00
-6.57804549e-01 9.59560394e-01 5.36892489e-02 -1.25544682e-01
6.55564547e-01 -2.75533378e-01 2.46198729e-01 -7.60216236e-01
2.65986174e-01 -1.26319453e-01 -4.65600550e-01 -6.91851437e-01
7.13742435e-01 1.09753087e-01 1.24669820e-01 4.18328702e-01
-1.98050532e-02 -3.66562366e-01 1.18104175e-01 2.42592290e-01
1.17929208e+00 5.25442839e-01 -8.82187188e-02 -7.29738455e-03
-7.93281011e-03 -2.84642041e-01 4.54659671e-01 7.35449910e-01
9.09743160e-02 1.04707742e+00 9.88292620e-02 -7.49816477e-01
-1.15852189e+00 -1.55032265e+00 -2.36676410e-01 9.61736739e-01
2.45801941e-01 -3.37833256e-01 -4.24123108e-01 -3.30148965e-01
-9.97288972e-02 8.82401645e-01 -6.88095152e-01 1.80571884e-01
-4.99283582e-01 -5.18714786e-01 1.83767334e-01 1.04069494e-01
6.73449397e-01 -7.78706193e-01 -1.33697140e+00 -4.00242537e-01
-3.32432777e-01 -1.51667774e+00 -2.77084619e-01 -3.07938270e-02
-2.61945695e-01 -1.21339357e+00 -5.01421571e-01 -1.51062831e-01
4.60299551e-01 9.93514538e-01 1.54828918e+00 -2.19474599e-01
-6.53358400e-01 8.67402196e-01 -3.40226591e-01 -4.29056644e-01
-6.59703389e-02 -2.50033557e-01 -2.15899460e-02 1.85176343e-01
-2.06809014e-01 -6.28557503e-01 -9.68916953e-01 3.18842918e-01
-1.04227102e+00 3.94595027e-01 4.76070642e-01 4.11692798e-01
5.38486362e-01 -2.29316384e-01 -2.97741681e-01 -5.35824597e-01
-3.12347189e-02 -4.12584126e-01 -1.03325474e+00 2.38554910e-01
-5.15827574e-02 -2.53731161e-01 7.34551132e-01 -1.55261874e-01
-1.30254662e+00 1.12795554e-01 4.24875706e-01 -9.49872017e-01
-2.55359173e-01 -2.04039276e-01 -2.30835766e-01 -1.65308431e-01
1.05233371e+00 -1.31558239e-01 -6.05786681e-01 -4.24588293e-01
8.53271723e-01 2.34569833e-01 8.61002505e-01 -7.38754690e-01
1.13590050e+00 9.56211448e-01 1.69083893e-01 -6.51020408e-01
-1.18062377e+00 -3.01090311e-02 -5.34634829e-01 -4.09804821e-01
9.02103901e-01 -1.02593076e+00 -9.60261703e-01 4.18460518e-01
-1.13064897e+00 -8.30976725e-01 -3.79758090e-01 4.27708536e-01
-7.29145229e-01 -7.29888603e-02 -2.48421624e-01 -6.27892435e-01
-1.40746370e-01 -1.11262560e+00 1.42945826e+00 3.44306827e-01
4.99359250e-01 -5.17280519e-01 1.86369658e-01 5.44164956e-01
1.89384237e-01 5.50460577e-01 3.96412611e-01 4.21390712e-01
-1.37173724e+00 2.24863246e-01 -3.47684890e-01 2.80743659e-01
-2.10581124e-01 1.79652348e-01 -1.47238839e+00 -4.82656956e-01
-1.54926172e-02 -6.97465539e-01 8.23718488e-01 1.69752985e-01
1.22630286e+00 -1.14671029e-01 4.36488353e-02 1.28193092e+00
1.78525567e+00 -2.20752195e-01 6.18501902e-01 3.51690441e-01
1.04341030e+00 6.38746262e-01 4.02917981e-01 5.03485978e-01
5.80742836e-01 8.10208917e-01 1.08841157e+00 -1.30468041e-01
-3.00517082e-01 -1.82878613e-01 3.80035967e-01 1.20707467e-01
-3.02509218e-01 -7.28289545e-01 -9.19430315e-01 3.32122266e-01
-1.36090910e+00 -8.31042469e-01 -1.47544399e-01 2.09524035e+00
5.07158518e-01 -3.11153769e-01 -2.62293726e-01 -5.17240167e-01
3.03496569e-01 4.82242495e-01 -9.84731674e-01 -7.57206902e-02
-3.54419291e-01 -1.25675052e-01 6.04294717e-01 4.12465602e-01
-7.46595800e-01 7.75760055e-01 5.70177317e+00 -1.99427884e-02
-1.19815183e+00 3.81892012e-03 3.34786326e-01 -6.06706321e-01
-5.43437898e-01 7.22942576e-02 -6.00005865e-01 1.26760393e-01
7.07937241e-01 2.71238983e-01 1.07806635e+00 6.16661727e-01
2.29488641e-01 -2.60339409e-01 -1.28807187e+00 1.15475619e+00
5.16594529e-01 -1.12546349e+00 -8.23603757e-03 9.94190797e-02
9.40549254e-01 5.89648902e-01 4.89545584e-01 -1.51226461e-01
6.38443589e-01 -8.15480411e-01 9.09801126e-01 8.92413139e-01
1.00405765e+00 -3.08255553e-01 -6.85004741e-02 9.93619040e-02
-8.62501264e-01 -1.97902963e-01 -4.48850811e-01 8.07759911e-02
3.43182951e-01 4.72003043e-01 -6.64838195e-01 7.84786582e-01
1.18990624e+00 9.50633466e-01 -7.45876372e-01 7.59925365e-01
-3.92031699e-01 1.11136980e-01 -5.43242753e-01 6.26040101e-01
2.98336484e-02 -3.84115040e-01 6.48904622e-01 6.41311407e-01
3.48667473e-01 1.85131952e-01 -1.45717636e-01 1.15666950e+00
-3.44145060e-01 -4.80797440e-01 -9.65761662e-01 5.41343927e-01
8.71046782e-02 1.58560610e+00 -3.75353605e-01 1.56091705e-01
-4.47685331e-01 9.84139383e-01 3.74715507e-01 9.15681005e-01
-1.12271869e+00 -2.20667734e-03 7.60221541e-01 8.44260380e-02
1.39631316e-01 -4.37065780e-01 5.79444021e-02 -1.58529222e+00
3.36204827e-01 -6.51434600e-01 7.14175217e-03 -1.95166266e+00
-1.21947026e+00 5.76372445e-01 1.97387323e-01 -1.29252446e+00
2.10467830e-01 -8.36465299e-01 -3.57219726e-01 6.42421007e-01
-1.92508984e+00 -1.48021257e+00 -1.12516069e+00 7.68437564e-01
5.06243885e-01 8.82450417e-02 4.26472664e-01 9.26258489e-02
-5.81419289e-01 1.53704599e-01 1.14069909e-01 -2.65901089e-01
9.00681198e-01 -1.10588050e+00 5.15317142e-01 1.04432714e+00
3.29570711e-01 5.58891356e-01 7.77875960e-01 -2.06433713e-01
-2.02472448e+00 -1.30932975e+00 1.14385612e-01 -1.07490873e+00
4.36915100e-01 -9.44779754e-01 -5.14612854e-01 1.22222555e+00
3.48214954e-01 4.37572598e-01 1.74676672e-01 -2.49455154e-01
-6.30348682e-01 -2.77237624e-01 -9.39394414e-01 9.87559915e-01
1.54014456e+00 -3.91651928e-01 -2.72732377e-01 5.76871753e-01
8.15715551e-01 -7.85662770e-01 -6.51983142e-01 2.33812451e-01
7.61437654e-01 -1.44564331e+00 1.30823088e+00 -3.55801433e-01
6.55002356e-01 -5.00394821e-01 -6.00130975e-01 -1.39608943e+00
-1.49667889e-01 -7.97681630e-01 -3.21239382e-02 8.34328055e-01
-5.16970195e-02 -7.84420311e-01 3.15551877e-01 5.54455578e-01
-3.11382264e-01 -5.76438129e-01 -5.91250300e-01 -6.18502796e-01
-1.48294136e-01 -5.19468129e-01 6.69193983e-01 7.28086770e-01
-1.04381335e+00 4.02452230e-01 -5.69509625e-01 4.40043747e-01
8.20683002e-01 2.40202338e-01 1.36971605e+00 -9.52562749e-01
-3.16493094e-01 2.92413831e-02 -3.46779488e-02 -1.09936547e+00
9.13464502e-02 -7.75293648e-01 1.35650024e-01 -1.33806598e+00
1.04894161e-01 -3.19684118e-01 1.57099366e-01 3.99181217e-01
1.81873068e-01 5.74124873e-01 3.93494517e-01 3.36683244e-01
-5.80871046e-01 7.07802594e-01 1.35233915e+00 2.01496836e-02
9.79280919e-02 -4.81209636e-01 -5.28945684e-01 7.13096261e-01
7.21910059e-01 -2.08496094e-01 -7.04831481e-01 -1.19674003e+00
7.99503088e-01 -9.85931754e-02 9.46277201e-01 -9.62233603e-01
7.62014836e-02 -2.97080308e-01 4.39957798e-01 -5.89266837e-01
7.17392981e-01 -1.04017496e+00 5.24147928e-01 -6.05509803e-02
-2.49342680e-01 2.22509265e-01 1.50765805e-02 8.04483354e-01
5.02693772e-01 1.37813881e-01 6.84664309e-01 -2.22290665e-01
-8.57882857e-01 4.79365915e-01 2.89937556e-01 1.45414203e-01
1.17273486e+00 -1.44374102e-01 -1.12583351e+00 -1.73177555e-01
-3.01324010e-01 1.12714432e-01 9.86064732e-01 5.80917776e-01
6.23415828e-01 -1.19153035e+00 -6.67812765e-01 5.37253082e-01
5.59992611e-01 4.29803461e-01 5.64966083e-01 6.01013720e-01
-8.55289817e-01 1.97441369e-01 -3.41554075e-01 -7.95749724e-01
-8.69874537e-01 7.06909060e-01 5.03487229e-01 2.81096905e-01
-1.01560366e+00 7.52170265e-01 7.80184269e-01 -5.95861435e-01
-7.80523568e-02 -4.97885227e-01 3.02861452e-01 -3.86759132e-01
5.13489962e-01 3.81296337e-01 7.77921155e-02 -6.94738507e-01
-7.98755810e-02 8.27542841e-01 2.12278306e-01 -1.06280051e-01
1.33275700e+00 -4.90300685e-01 8.88339356e-02 5.51960945e-01
1.11838305e+00 1.34737849e-01 -1.75806868e+00 -2.06500486e-01
-9.14466262e-01 -1.01863170e+00 2.72925019e-01 -1.06160152e+00
-1.23692238e+00 7.31887698e-01 5.21453142e-01 -2.43366703e-01
1.21217859e+00 -1.91403949e-03 6.50041819e-01 7.44950294e-01
6.28551722e-01 -7.28519022e-01 9.50393528e-02 3.97882909e-01
1.26262963e+00 -1.29496157e+00 -3.27336937e-02 -2.72392809e-01
-5.41010320e-01 1.01704252e+00 6.77837133e-01 -1.88704550e-01
3.29711109e-01 2.11095840e-01 1.36651903e-01 -4.12917703e-01
-1.03040636e+00 -9.33552906e-02 5.14962897e-02 7.21332133e-01
-1.62542775e-01 -1.45164758e-01 7.37104416e-01 -1.28837779e-01
-6.03438377e-01 -8.68328661e-02 9.62852538e-01 8.61856103e-01
3.12455427e-02 -3.11394960e-01 -3.96994323e-01 1.39709175e-01
3.30289192e-02 -2.66717076e-02 -2.49907181e-01 7.23214626e-01
1.35933384e-01 8.86115432e-01 2.09903434e-01 -1.65657327e-01
5.85251272e-01 -5.29504657e-01 8.55182230e-01 -4.30549651e-01
-1.92117542e-01 -1.88552439e-01 1.12314932e-02 -1.18241227e+00
-4.94597346e-01 -5.85146844e-01 -7.00235307e-01 -3.63240808e-01
9.37346090e-03 -7.04667985e-01 8.33203912e-01 4.46316659e-01
4.42668945e-01 4.92844820e-01 8.05953085e-01 -1.37839508e+00
-2.10130677e-01 -5.87935269e-01 -5.63347876e-01 4.27976251e-01
7.07038820e-01 -8.95692647e-01 -4.73837882e-01 3.01661760e-01] | [9.423449516296387, -3.0414936542510986] |
f61e8958-4cc7-4f18-ac38-823766399f48 | measuring-and-reducing-non-multifact | 2005.00789 | null | https://arxiv.org/abs/2005.00789v3 | https://arxiv.org/pdf/2005.00789v3.pdf | Is Multihop QA in DiRe Condition? Measuring and Reducing Disconnected Reasoning | Has there been real progress in multi-hop question-answering? Models often exploit dataset artifacts to produce correct answers, without connecting information across multiple supporting facts. This limits our ability to measure true progress and defeats the purpose of building multi-hop QA datasets. We make three contributions towards addressing this. First, we formalize such undesirable behavior as disconnected reasoning across subsets of supporting facts. This allows developing a model-agnostic probe for measuring how much any model can cheat via disconnected reasoning. Second, using a notion of \emph{contrastive support sufficiency}, we introduce an automatic transformation of existing datasets that reduces the amount of disconnected reasoning. Third, our experiments suggest that there hasn't been much progress in multi-hop QA in the reading comprehension setting. For a recent large-scale model (XLNet), we show that only 18 points out of its answer F1 score of 72 on HotpotQA are obtained through multifact reasoning, roughly the same as that of a simpler RNN baseline. Our transformation substantially reduces disconnected reasoning (19 points in answer F1). It is complementary to adversarial approaches, yielding further reductions in conjunction. | ['Tushar Khot', 'Ashish Sabharwal', 'Niranjan Balasubramanian', 'Harsh Trivedi'] | 2020-05-02 | null | https://aclanthology.org/2020.emnlp-main.712 | https://aclanthology.org/2020.emnlp-main.712.pdf | emnlp-2020-11 | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 1.70371860e-01 9.95478094e-01 -5.37929386e-02 -3.15634638e-01
-1.61815727e+00 -1.08419168e+00 5.21899402e-01 2.86955267e-01
-4.35586393e-01 9.84475434e-01 7.30844557e-01 -8.68486345e-01
-2.36434132e-01 -9.82410908e-01 -1.11076224e+00 -1.67072222e-01
5.52523851e-01 8.19956899e-01 3.59623283e-01 -8.19251955e-01
2.86635011e-01 -8.47606584e-02 -9.66105759e-01 6.35916710e-01
1.25826931e+00 3.49714428e-01 -6.19333506e-01 9.21743274e-01
-2.10881963e-01 1.58136559e+00 -9.93661225e-01 -1.07845557e+00
1.96660116e-01 -5.27929008e-01 -1.86021924e+00 -6.63725019e-01
1.16107130e+00 -5.10524213e-01 -3.66208166e-01 9.56430078e-01
4.69558448e-01 7.66503513e-02 3.23608816e-01 -1.17205191e+00
-8.96067679e-01 1.23123026e+00 -2.04466000e-01 4.46601868e-01
7.13193536e-01 5.15610218e-01 1.51544297e+00 -6.81413487e-02
7.81547725e-01 1.33970308e+00 8.74822736e-01 7.19373643e-01
-1.35737371e+00 -3.18832725e-01 5.65130860e-02 3.44074607e-01
-7.47266293e-01 -4.88824397e-01 5.59739947e-01 -1.60871819e-01
1.10560274e+00 6.82037830e-01 2.53942728e-01 1.25073349e+00
-9.91784260e-02 7.06573725e-01 1.30984879e+00 -4.71901059e-01
-1.24217406e-01 -1.46640733e-01 4.51901644e-01 8.22612882e-01
7.98411295e-02 -2.52676278e-01 -4.88546938e-01 -3.32531571e-01
6.09269366e-02 -7.70664632e-01 -4.87263352e-01 -7.25183412e-02
-1.24163616e+00 9.93177474e-01 5.25128663e-01 2.09004939e-01
9.82848331e-02 2.77775109e-01 4.10865247e-01 9.00876224e-01
1.53553531e-01 1.20291102e+00 -7.74958432e-01 -5.44277012e-01
-7.29991972e-01 6.68692946e-01 1.16106558e+00 7.58679748e-01
3.21280599e-01 -3.07490170e-01 -3.30850542e-01 4.23077762e-01
-3.01936597e-01 5.60122788e-01 3.54749799e-01 -1.84505880e+00
1.03936720e+00 6.24272108e-01 2.67490953e-01 -9.67799842e-01
-4.29907382e-01 -2.34997034e-01 -3.23669046e-01 -1.02707036e-01
1.05474305e+00 -2.12707266e-01 -3.92325222e-01 2.16443300e+00
1.27540261e-01 -4.00335848e-01 2.30200306e-01 6.66969001e-01
7.73620486e-01 3.86988759e-01 -2.84019522e-02 1.68975279e-01
1.28533864e+00 -1.10638261e+00 -4.82488453e-01 -2.49874182e-02
9.76562381e-01 -5.63703716e-01 1.77932262e+00 4.97277439e-01
-1.55396867e+00 -2.23226935e-01 -8.32764566e-01 -6.49150908e-01
-2.38225684e-01 -6.87833309e-01 5.72795093e-01 6.53388500e-01
-1.14696467e+00 4.19446290e-01 -5.16079485e-01 -1.91949129e-01
3.09064507e-01 1.07763886e-01 -2.71024823e-01 -1.67486861e-01
-1.46853650e+00 1.29401922e+00 6.75080866e-02 -4.19521362e-01
-5.67513227e-01 -8.93138230e-01 -6.34440899e-01 8.74538124e-02
7.86404073e-01 -1.05646360e+00 1.64592612e+00 -6.60739779e-01
-1.31774116e+00 7.53870428e-01 -2.37001970e-01 -7.94876277e-01
8.95732641e-01 -4.74691302e-01 -2.50424474e-01 2.59477317e-01
2.70919323e-01 7.69690394e-01 3.19363058e-01 -1.08161485e+00
-3.24829996e-01 -2.15119958e-01 8.23956907e-01 2.62358814e-01
-9.82085615e-02 -1.92367703e-01 6.06773794e-02 -3.86970490e-01
-1.26332372e-01 -7.67172754e-01 1.26076117e-01 -2.27809682e-01
-6.67283416e-01 -4.43117142e-01 3.11984986e-01 -8.30922782e-01
9.83197689e-01 -1.58320498e+00 1.72561482e-01 -8.75410438e-03
5.47827244e-01 1.95297912e-01 -5.34973919e-01 5.05982041e-01
1.68410063e-01 4.68432963e-01 -3.98031563e-01 9.27160233e-02
2.52884835e-01 2.08763450e-01 -8.60662818e-01 -1.12583404e-02
2.37902552e-01 1.51420414e+00 -1.00174820e+00 -5.15529156e-01
-2.97302723e-01 -6.83387518e-02 -8.22050869e-01 -1.62066985e-02
-7.60891020e-01 2.03388587e-01 -2.57649451e-01 5.13213634e-01
4.42703128e-01 -3.20094615e-01 -3.44189741e-02 2.71473099e-02
3.51399094e-01 8.77906561e-01 -5.08833170e-01 1.78811181e+00
-4.18626189e-01 4.84808296e-01 -9.84653458e-02 -7.19873071e-01
4.18067545e-01 1.54133111e-01 2.33414210e-02 -8.49530935e-01
-3.86164635e-02 1.91226557e-01 3.59802723e-01 -8.09419632e-01
5.51253676e-01 -1.72751740e-01 -1.18992046e-01 7.67031372e-01
-1.81207493e-01 -5.21893322e-01 1.59925133e-01 6.56054020e-01
1.73103297e+00 -1.19740963e-01 -3.12550932e-01 -2.00349882e-01
4.84055221e-01 5.28453887e-01 2.52004474e-01 1.05539429e+00
-3.50605130e-01 4.64753985e-01 9.01010752e-01 -2.12505102e-01
-9.70216930e-01 -1.20495808e+00 2.11114943e-01 1.06194556e+00
-3.88641022e-02 -4.46970224e-01 -9.78741348e-01 -1.20522904e+00
2.17464361e-02 1.17594409e+00 -5.90727329e-01 -2.74926454e-01
-8.62951159e-01 -3.06944162e-01 1.49641144e+00 5.11849105e-01
6.41276658e-01 -9.37132120e-01 -4.27090943e-01 9.04690400e-02
-9.97694433e-01 -8.61568570e-01 -2.23857388e-01 6.91907015e-03
-9.15185988e-01 -1.31891930e+00 -2.06929341e-01 -3.30125242e-01
2.79049337e-01 9.59439203e-02 1.74591100e+00 3.13485652e-01
2.33083293e-01 4.55718726e-01 -2.75687397e-01 -2.19335333e-01
-7.46556938e-01 5.10095179e-01 -3.61376554e-01 -9.18991268e-01
6.40911043e-01 -3.99087906e-01 -4.96227026e-01 2.12691769e-01
-7.20191896e-01 -2.73564816e-01 3.63490164e-01 8.09900224e-01
1.98985174e-01 -1.40635997e-01 7.85316765e-01 -1.40581608e+00
1.07247972e+00 -6.11627877e-01 -2.13639289e-01 6.02710187e-01
-5.70375025e-01 3.05125952e-01 9.72666681e-01 -1.86275527e-01
-9.78981674e-01 -6.19817019e-01 -3.85930777e-01 1.41534463e-01
-1.85764357e-01 3.10505986e-01 -5.43998405e-02 1.64004326e-01
1.21428311e+00 -1.55383602e-01 4.15036418e-02 -2.62298316e-01
9.84174013e-01 3.56507719e-01 8.52203906e-01 -9.81393039e-01
9.76270854e-01 3.41919124e-01 -3.11894983e-01 -2.11736977e-01
-1.37606132e+00 -1.11963697e-01 -1.89022556e-01 4.11811769e-01
7.62262642e-01 -5.86419702e-01 -1.11720932e+00 -2.34358292e-02
-1.14063263e+00 -6.32130682e-01 -7.05530047e-01 -6.06238469e-03
-7.40042627e-01 4.81237382e-01 -8.82160485e-01 -3.50720346e-01
-4.63218182e-01 -9.10588264e-01 6.50002539e-01 -7.12997764e-02
-8.33386302e-01 -9.72265065e-01 3.29079717e-01 1.54618454e+00
4.66267318e-01 1.07988983e-01 1.36248207e+00 -8.96374941e-01
-5.17423093e-01 6.27855286e-02 -8.86763409e-02 2.62807310e-01
-2.33214572e-01 -2.27948532e-01 -1.00187290e+00 -1.27552748e-01
2.02061385e-01 -1.07661021e+00 7.94709980e-01 -1.54800877e-01
1.13903475e+00 -7.31803954e-01 2.58039445e-01 2.02656642e-01
1.07248461e+00 -2.35780135e-01 8.19036663e-01 5.28086483e-01
6.35734618e-01 7.10852265e-01 3.32222611e-01 -4.22181606e-01
7.27448523e-01 3.16669524e-01 3.35815102e-01 2.30475023e-01
-2.58150548e-01 -6.35436773e-01 2.96992302e-01 7.55013764e-01
1.02322765e-01 -4.34345752e-01 -9.70477939e-01 6.65645421e-01
-1.64616501e+00 -1.21726143e+00 -4.92120266e-01 1.79091227e+00
1.21528816e+00 3.70501310e-01 1.43422291e-01 1.13305122e-01
2.39421964e-01 1.16477318e-01 -6.55340791e-01 -7.55721986e-01
-4.05532569e-01 4.78057384e-01 3.51733685e-01 1.03418982e+00
-5.52396536e-01 9.82559800e-01 6.97273540e+00 6.75933123e-01
-4.61887836e-01 2.23976046e-01 5.32127261e-01 -1.63984001e-01
-1.01114714e+00 1.25995040e-01 -3.13655943e-01 2.85263628e-01
1.15194666e+00 2.96427161e-02 6.58919394e-01 3.67701024e-01
-2.23038331e-01 -4.27940451e-02 -1.21547055e+00 1.73628956e-01
1.44189417e-01 -1.41441298e+00 1.63786396e-01 -1.70352191e-01
7.18945920e-01 2.68327910e-02 1.92563504e-01 6.48197591e-01
9.52220678e-01 -1.31095755e+00 6.16465688e-01 4.15744305e-01
5.39740086e-01 -6.60235524e-01 7.43892610e-01 6.37189567e-01
-2.92203158e-01 -2.18351081e-01 -2.58758813e-01 -2.38882974e-01
1.66460484e-01 2.54270613e-01 -7.97158182e-01 4.89424765e-01
4.69883204e-01 -1.01999260e-01 -7.95168161e-01 4.00322616e-01
-6.14303827e-01 1.02798533e+00 -2.78494269e-01 1.46527067e-01
3.42687398e-01 1.79978892e-01 5.35348356e-01 7.37264514e-01
-1.94903925e-01 2.27456793e-01 -3.20637047e-01 9.77292180e-01
-7.15230525e-01 -1.47169888e-01 -5.71415484e-01 1.08014748e-01
6.69709623e-01 8.09068024e-01 9.76414606e-03 -3.20815891e-01
-3.95058542e-01 8.08092356e-01 7.64945030e-01 8.70596841e-02
-8.19816470e-01 -3.36188078e-01 1.49166435e-01 6.77314252e-02
-1.45179465e-01 5.75795956e-02 -5.67436874e-01 -1.17204189e+00
3.42081755e-01 -1.72066486e+00 6.89313650e-01 -8.77215922e-01
-1.44299924e+00 3.68180096e-01 -2.10262835e-01 -4.88772720e-01
-1.85440242e-01 -2.14151949e-01 -6.92376614e-01 8.32611084e-01
-1.44085431e+00 -1.09983468e+00 -1.81073964e-01 6.40152216e-01
3.87289882e-01 1.61033217e-02 9.47749257e-01 9.13692545e-03
-2.02954948e-01 1.06224465e+00 -2.06783414e-01 1.06022663e-01
9.72413242e-01 -1.72401237e+00 6.37252331e-01 8.95246029e-01
3.21840227e-01 8.32332611e-01 8.36782992e-01 -4.97108698e-01
-1.28125155e+00 -6.51743948e-01 1.09908831e+00 -1.46560836e+00
1.00625181e+00 8.70786235e-02 -1.21158671e+00 1.01571536e+00
3.99152726e-01 -5.16440511e-01 5.50570309e-01 4.16985244e-01
-9.38923717e-01 1.12816274e-01 -1.33897412e+00 7.98358858e-01
1.15118027e+00 -7.34283745e-01 -1.34323072e+00 4.50473577e-01
1.30137563e+00 -3.97859216e-01 -9.00145829e-01 2.76708871e-01
2.08371550e-01 -1.09940982e+00 8.54299605e-01 -1.14517653e+00
9.06525671e-01 -7.47831017e-02 -2.25485638e-01 -1.20666683e+00
-1.27528831e-01 -7.42062569e-01 -3.60356778e-01 1.16396236e+00
7.79768825e-01 -7.58651614e-01 9.64422584e-01 9.97906446e-01
-4.56396341e-02 -7.80995071e-01 -9.01600420e-01 -6.07037127e-01
1.01692891e+00 -2.98673213e-01 8.27751219e-01 1.06774879e+00
1.53517604e-01 6.22096539e-01 8.36089998e-02 2.35560741e-02
3.79549444e-01 4.07894701e-02 9.31898236e-01 -9.17959750e-01
-6.84207022e-01 -4.85515296e-01 1.37914509e-01 -1.18355834e+00
2.83474296e-01 -9.97496665e-01 -1.89074397e-01 -1.61842144e+00
1.67845875e-01 -5.62136114e-01 -1.58514362e-02 6.67006314e-01
-3.37590337e-01 7.84907117e-02 1.70890838e-01 2.46496752e-01
-7.42937028e-01 2.43962765e-01 1.41970193e+00 -3.31264377e-01
2.07986727e-01 -2.18510151e-01 -1.30660176e+00 6.05333626e-01
9.44302499e-01 -4.24323708e-01 -7.57151604e-01 -9.33531284e-01
7.15285242e-01 1.36034057e-01 5.07933915e-01 -7.06866562e-01
2.04926565e-01 -3.81900370e-02 -1.15518980e-01 -3.00469190e-01
2.09086299e-01 -3.94063771e-01 -3.56815070e-01 5.66273987e-01
-9.78486359e-01 2.61335999e-01 1.59181282e-01 4.19998974e-01
-1.21726863e-01 -3.46601397e-01 4.75232720e-01 -3.04590791e-01
-9.78132933e-02 -4.05602247e-01 -1.12122208e-01 1.02865434e+00
5.73433936e-01 1.52736843e-01 -1.39322197e+00 -5.96592367e-01
-4.94311333e-01 4.60144311e-01 4.57304686e-01 1.53091371e-01
1.84633225e-01 -7.88933814e-01 -8.96160185e-01 -3.72189939e-01
-2.72936970e-02 1.88987389e-01 2.43376598e-01 8.02513421e-01
-7.03994572e-01 5.08002341e-01 -3.46931480e-02 -1.03708774e-01
-1.06343973e+00 5.97211897e-01 5.16777217e-01 -5.89487195e-01
-6.30062938e-01 1.06332803e+00 -2.08175600e-01 -9.51128602e-01
3.16151488e-03 -3.44303191e-01 9.50685665e-02 -2.36016050e-01
3.01335335e-01 6.35814309e-01 2.54327476e-01 -1.56475425e-01
-2.80409516e-03 2.86347598e-01 -3.31482023e-01 -3.77297819e-01
1.10318577e+00 1.57349892e-02 -2.00836211e-01 2.47075304e-01
9.51777697e-01 6.42792881e-01 -9.41624761e-01 -6.85907304e-02
-9.75676253e-02 -2.35150516e-01 -4.71318811e-01 -1.56182551e+00
-6.15109444e-01 9.11976993e-01 -5.34732826e-02 4.94470358e-01
8.02351952e-01 3.67346331e-02 1.35787129e+00 1.01072729e+00
2.81689674e-01 -8.88898909e-01 1.53132588e-01 6.59979045e-01
8.92446637e-01 -1.24434173e+00 -1.81006193e-01 -1.75511122e-01
-6.81430995e-01 6.44827545e-01 8.93732488e-01 1.14237346e-01
-1.35799408e-01 6.69020787e-02 3.63895863e-01 -4.53395545e-01
-9.57993805e-01 1.53462544e-01 -4.87290360e-02 5.53239167e-01
4.04607892e-01 5.39707951e-02 -1.21794380e-01 4.21923190e-01
-1.13274372e+00 -3.69292349e-01 7.68095553e-01 8.64756763e-01
-4.44268525e-01 -1.00855088e+00 -3.31320733e-01 4.88930434e-01
-6.67713284e-01 -2.92315394e-01 -8.39853883e-01 9.52447295e-01
-1.71985865e-01 1.34156704e+00 -2.36105293e-01 -3.34564447e-01
5.49242496e-01 3.50372225e-01 6.93109453e-01 -4.44990963e-01
-9.32249010e-01 -9.79670346e-01 5.12168884e-01 -6.21575236e-01
-1.46819623e-02 -2.53022999e-01 -1.25194204e+00 -1.03438270e+00
-2.84769624e-01 3.28664780e-01 1.01011880e-01 9.73136187e-01
3.39732468e-01 3.76832515e-01 1.41486108e-01 3.15878063e-01
-1.10969567e+00 -8.83142054e-01 1.75482735e-01 6.61086440e-01
3.56524825e-01 -1.49789840e-01 -4.87847537e-01 -2.29350910e-01] | [11.003938674926758, 7.969555377960205] |
ffb5d188-9bab-4e12-8659-d356f79f545b | on-robust-inference-in-time-series-regression | 2203.04080 | null | https://arxiv.org/abs/2203.04080v1 | https://arxiv.org/pdf/2203.04080v1.pdf | On Robust Inference in Time Series Regression | Least squares regression with heteroskedasticity and autocorrelation consistent (HAC) standard errors has proved very useful in cross section environments. However, several major difficulties, which are generally overlooked, must be confronted when transferring the HAC estimation technology to time series environments. First, most economic time series have strong autocorrelation, which renders HAC regression parameter estimates highly inefficient. Second, strong autocorrelation similarly renders HAC conditional predictions highly inefficient. Finally, the structure of most popular HAC estimators is ill-suited to capture the autoregressive autocorrelation typically present in economic time series, which produces large size distortions and reduced power in hypothesis testing, in all but the largest sample sizes. We show that all three problems are largely avoided by the use of a simple dynamic regression (DynReg), which is easily implemented and also avoids possible problems concerning strong exogeneity. We demonstrate the advantages of DynReg with detailed simulations covering a range of practical issues. | ['Kun Ho Kim', 'George Kapetanios', 'Francis X. Diebold', 'Richard T. Baillie'] | 2022-03-08 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [-3.55818152e-01 -4.16468889e-01 -5.09540081e-01 -2.15500921e-01
-7.30068624e-01 -4.88688588e-01 5.11094511e-01 -3.15021098e-01
-4.31466252e-01 1.03291357e+00 1.29747331e-01 -8.53446901e-01
-4.86717671e-01 -4.90406781e-01 -4.60596204e-01 -5.77661335e-01
-1.32166892e-01 -1.37586966e-01 -3.07109237e-01 2.31051311e-01
3.09885591e-01 4.16883022e-01 -1.02862608e+00 -7.58063674e-01
1.07997179e+00 8.52929235e-01 2.52220511e-01 3.77546757e-01
2.69721836e-01 1.00504422e+00 -5.00292540e-01 -5.05179763e-01
3.33594114e-01 -3.66639167e-01 -2.04779729e-01 5.03304414e-02
-2.56706685e-01 -2.45618552e-01 6.77715763e-02 7.25424051e-01
2.34850913e-01 2.59485990e-01 1.01983595e+00 -1.18875027e+00
-6.05069578e-01 5.35842180e-01 -6.91429496e-01 3.17636997e-01
7.83335343e-02 1.56218991e-01 7.16994822e-01 -9.30591583e-01
2.25006536e-01 1.26628232e+00 8.46498728e-01 -9.07739326e-02
-1.30518508e+00 -7.48297930e-01 -1.93372220e-01 -4.59514439e-01
-1.40953612e+00 -5.52214086e-01 4.91023779e-01 -5.35151064e-01
9.61682081e-01 2.18610525e-01 3.72576624e-01 9.28905964e-01
6.70001984e-01 2.30799139e-01 1.47750676e+00 -5.53153872e-01
2.74900775e-02 3.13059002e-01 2.18692794e-01 1.53156653e-01
6.73869848e-01 3.76424253e-01 1.20418958e-01 -3.48478824e-01
1.17591536e+00 1.73824415e-01 1.34945005e-01 -8.30916166e-02
-1.18622088e+00 1.13241053e+00 -1.52099848e-01 4.42012757e-01
-7.96357512e-01 3.13898027e-01 3.06799144e-01 6.60878241e-01
3.97881895e-01 3.29431295e-01 -6.97009087e-01 -1.96593016e-01
-7.06034958e-01 3.54837269e-01 6.75830066e-01 9.76948440e-01
5.03335476e-01 6.81446850e-01 1.57297984e-01 7.49013066e-01
-1.51322931e-02 1.12433791e+00 7.07909584e-01 -1.28672290e+00
4.90361512e-01 4.27657273e-03 2.30429128e-01 -1.24965203e+00
-5.05831003e-01 -5.91925204e-01 -1.03842795e+00 4.17209417e-02
6.02979541e-01 -7.84461021e-01 -2.94743508e-01 1.73179424e+00
-1.23262301e-01 -1.24389529e-01 1.28665537e-01 3.51637989e-01
-2.75742650e-01 7.06971943e-01 1.82893902e-01 -8.76550913e-01
1.14092958e+00 -3.94894332e-01 -8.29740226e-01 -1.67922541e-01
8.44699323e-01 -7.30425358e-01 8.24590504e-01 6.17735125e-02
-1.22435713e+00 -4.58234996e-01 -3.81934106e-01 2.85379708e-01
-2.30376437e-01 -1.00549772e-01 7.87318766e-01 8.37484419e-01
-7.18420506e-01 4.44590598e-01 -8.48428547e-01 -3.88874002e-02
-1.97492838e-01 3.82807374e-01 -1.77989900e-01 3.74767274e-01
-9.90848005e-01 8.30144703e-01 4.67937887e-02 -8.03615451e-02
-1.15920350e-01 -8.61858010e-01 -1.06779444e+00 4.61017996e-01
3.53726178e-01 -4.25174475e-01 1.27881086e+00 -1.32678783e+00
-1.38202834e+00 2.74905027e-03 -3.13219160e-01 -3.55989337e-01
6.05989575e-01 -1.42660493e-03 -8.52327883e-01 -3.61102253e-01
6.03416383e-01 -1.15380034e-01 5.63441992e-01 -8.20330620e-01
-6.40859187e-01 -3.67427081e-01 -7.42106199e-01 -1.18955530e-01
-3.70913669e-02 5.28398514e-01 -1.19534358e-01 -1.18135190e+00
-2.53822118e-01 -9.50720966e-01 -5.31218171e-01 -1.01187956e+00
1.75625384e-02 -1.73582777e-01 1.52766258e-01 -6.60311699e-01
1.80148041e+00 -2.06791401e+00 -5.03477156e-01 7.16759324e-01
-4.03549880e-01 -3.99256825e-01 7.35245645e-02 4.65012670e-01
-5.41439295e-01 2.17203394e-01 7.53092254e-03 -4.53182757e-02
8.19322765e-02 5.71350381e-03 -3.11842293e-01 6.42421305e-01
5.03618419e-02 8.96018386e-01 -4.74859625e-01 -2.81306684e-01
2.69114435e-01 -1.00274622e-01 -3.97111654e-01 -2.89360762e-01
6.73174024e-01 6.69362620e-02 -3.68860960e-01 4.54558432e-01
6.29598439e-01 -2.37492487e-01 3.19006026e-01 6.15173936e-01
-8.68890941e-01 2.38930732e-01 -1.17396379e+00 6.45888448e-01
-5.52677274e-01 5.26834786e-01 -3.69067341e-01 -1.50874293e+00
7.30012596e-01 4.41914916e-01 4.71911579e-01 -9.42756832e-01
-8.07770193e-02 5.04027009e-01 1.16631024e-01 -4.28194582e-01
4.43425864e-01 -4.22447085e-01 -3.26310873e-01 3.22552890e-01
-1.78623140e-01 6.77054282e-03 7.04125091e-02 -3.71176749e-01
6.65204406e-01 -3.52178365e-01 7.83715069e-01 -6.84907496e-01
1.49172574e-01 1.19693624e-02 1.19507062e+00 9.69731271e-01
1.45322561e-01 1.61883220e-01 9.40120339e-01 1.34107415e-02
-1.06913543e+00 -9.04795587e-01 -3.12841862e-01 6.91437244e-01
-5.73911786e-01 1.00471407e-01 -2.11256072e-01 -3.41854811e-01
4.55700636e-01 7.88454294e-01 -4.93638188e-01 2.01956481e-01
-4.50289398e-01 -8.04687679e-01 2.13655040e-01 7.44826317e-01
2.20542759e-01 -7.92857289e-01 -5.68157732e-01 4.34963405e-01
2.67010897e-01 -8.45250666e-01 -2.20982984e-01 2.85043735e-02
-1.00492382e+00 -8.20881546e-01 -1.09323108e+00 -5.54663002e-01
4.83182073e-01 6.46790802e-01 1.03585470e+00 -2.00188562e-01
3.02859724e-01 4.13333744e-01 -1.11076161e-01 -8.28014195e-01
-3.21005374e-01 -1.17673844e-01 1.59807190e-01 -2.99986869e-01
5.30182183e-01 -3.44902456e-01 -4.71001863e-01 3.33405226e-01
-4.96118218e-01 -4.25322413e-01 5.99722445e-01 1.05359185e+00
1.22723155e-01 4.18119490e-01 1.29126954e+00 -8.73495162e-01
5.09963870e-01 -6.81398511e-01 -1.20167291e+00 2.93783337e-01
-1.06678021e+00 -3.63844126e-01 6.87693119e-01 -5.63658357e-01
-1.41801441e+00 -4.93857414e-01 2.85918713e-01 -2.41168682e-03
1.28659219e-01 1.16042662e+00 4.36847925e-01 3.76719594e-01
4.10164326e-01 -8.19395557e-02 4.18068260e-01 -4.42816287e-01
-4.77858067e-01 6.39408231e-01 2.55861253e-01 -2.40003824e-01
6.81316495e-01 1.63114429e-01 2.15220779e-01 -9.02099371e-01
-2.08245963e-01 -5.92729867e-01 -2.31448784e-01 1.83399186e-01
6.44350827e-01 -1.35642600e+00 -6.48174345e-01 4.64995027e-01
-5.65639913e-01 -6.45839334e-01 -1.46703377e-01 1.32081270e+00
-7.00576723e-01 1.35488018e-01 -5.88004470e-01 -1.27519977e+00
7.65041038e-02 -1.05248654e+00 4.30112451e-01 1.58122942e-01
-4.55825061e-01 -1.33169878e+00 -2.99704298e-02 -7.61372149e-02
4.15233850e-01 3.32163155e-01 1.05869484e+00 -3.83457452e-01
-6.83111027e-02 -3.65338385e-01 -1.27299428e-01 3.07844162e-01
1.77518636e-01 4.45500225e-01 -3.59026998e-01 -3.42285156e-01
-2.35740338e-02 9.30815935e-02 2.95566320e-01 1.14311981e+00
9.98181581e-01 -3.54035407e-01 -1.53530449e-01 3.26782227e-01
1.45821023e+00 6.42520487e-01 4.92891043e-01 2.66234279e-01
6.47369698e-02 6.88172221e-01 6.33367181e-01 8.88321757e-01
4.89967138e-01 4.67399269e-01 -5.28216541e-01 -3.84882420e-01
8.05467725e-01 -2.48812929e-01 4.82613266e-01 9.38999414e-01
-3.34364384e-01 -6.55994043e-02 -8.18686426e-01 6.59171581e-01
-1.83689511e+00 -1.25453532e+00 -5.45687854e-01 2.33954453e+00
6.12647712e-01 2.78132018e-02 4.93075937e-01 1.55682433e-02
4.38999355e-01 -1.04722351e-01 -6.72403648e-02 -6.07803524e-01
-2.90797442e-01 2.18836635e-01 1.06726956e+00 2.44522825e-01
-8.48327935e-01 5.80348194e-01 7.81393766e+00 4.93437171e-01
-7.94229984e-01 -1.37209594e-01 6.56617582e-01 3.59614164e-01
-3.03830892e-01 4.28192616e-01 -4.46666688e-01 6.71476960e-01
1.31468320e+00 -7.63752639e-01 7.48396963e-02 9.03048933e-01
9.19630468e-01 -4.46357965e-01 -4.79939222e-01 8.15644681e-01
-5.55480123e-01 -8.66385221e-01 -5.59781313e-01 4.64654356e-01
1.10788500e+00 -2.08521649e-01 4.39733863e-01 6.58940852e-01
5.65995634e-01 -8.73065829e-01 5.77317417e-01 4.20300722e-01
7.07194090e-01 -1.15912008e+00 8.42141688e-01 3.14526796e-01
-8.21183145e-01 -5.95116317e-01 -8.05301130e-01 -5.42038679e-01
5.40587083e-02 7.89673984e-01 -3.10445189e-01 4.21704382e-01
6.48997366e-01 5.83865702e-01 -2.05752373e-01 8.37257087e-01
1.87499955e-01 7.58105576e-01 -2.15280846e-01 1.58290103e-01
3.71431142e-01 -6.39271438e-01 4.30606082e-02 1.13668513e+00
1.05241716e+00 1.40642345e-01 -4.67407882e-01 5.87430537e-01
2.82232672e-01 2.58051693e-01 -1.03865242e+00 -3.35458070e-02
5.28435349e-01 6.93807006e-01 -3.17439407e-01 -3.80889565e-01
-1.27071786e+00 3.26317042e-01 -6.85388595e-02 7.24543631e-01
-5.99605024e-01 -2.11883426e-01 4.24117684e-01 4.69390377e-02
5.34999147e-02 -3.68149281e-01 -2.85361946e-01 -1.30009222e+00
-2.30762467e-01 -1.06717408e+00 6.16901278e-01 -5.13015687e-01
-1.07305908e+00 -3.83197308e-01 2.67378122e-01 -1.06280041e+00
-8.82518768e-01 -4.88970131e-01 -6.15253985e-01 8.99683475e-01
-1.27065468e+00 -5.12616456e-01 3.30910653e-01 6.77215517e-01
5.70825994e-01 -1.45673007e-01 5.45237303e-01 1.79462969e-01
-9.24210727e-01 4.86242622e-01 7.92321861e-01 1.40531817e-02
7.38497674e-01 -1.37708104e+00 2.58184224e-01 9.79940653e-01
-3.85697633e-01 9.56337512e-01 5.87013602e-01 -7.75687158e-01
-1.13668704e+00 -8.25219393e-01 1.30254996e+00 -2.05154255e-01
9.26826179e-01 -3.05026565e-02 -7.78846145e-01 1.06532562e+00
-3.03964429e-02 -4.51452821e-01 6.24574482e-01 4.31088179e-01
8.07169229e-02 -1.99057192e-01 -6.35846615e-01 8.02997470e-01
3.65457654e-01 -4.32026982e-01 -4.20527011e-01 9.27761942e-02
3.30266118e-01 3.50773782e-01 -1.22030556e+00 4.96964723e-01
9.54852939e-01 -9.59890842e-01 6.62798762e-01 -5.37566185e-01
9.53500122e-02 2.76225001e-01 4.49968129e-02 -1.22384679e+00
-6.34290636e-01 -9.03012514e-01 6.17770910e-01 1.44617844e+00
7.13092804e-01 -1.12591422e+00 1.70703486e-01 9.57091570e-01
2.16066122e-01 1.67642776e-02 -7.54333019e-01 -1.32284009e+00
5.52971601e-01 -5.10073066e-01 7.02685654e-01 1.35693228e+00
1.23099737e-01 -2.03741360e-02 -6.85934782e-01 -1.58703640e-01
4.79605615e-01 7.56809041e-02 9.61732924e-01 -1.27065146e+00
-2.45277733e-01 -3.64645243e-01 7.82188028e-02 -7.67323434e-01
4.41713303e-01 -5.46780452e-02 -1.98008120e-01 -5.60288966e-01
2.52421647e-01 -4.46400166e-01 -2.15578377e-01 -1.48807079e-01
-1.56068355e-01 -3.49326640e-01 -1.94043405e-02 1.97884828e-01
1.52314618e-01 1.32377401e-01 8.08154583e-01 4.39529657e-01
-4.51883167e-01 5.72641969e-01 -6.29484117e-01 6.93661034e-01
8.81280422e-01 -4.91591156e-01 -3.64593595e-01 -1.46783531e-01
2.55189329e-01 8.98110807e-01 3.88424844e-01 -4.56664175e-01
-1.55618321e-02 -7.79164016e-01 4.94112879e-01 -5.12459159e-01
-3.08683574e-01 -1.00193667e+00 6.97598815e-01 4.10912991e-01
-1.50381565e-01 1.07519364e+00 -7.15669850e-03 4.15033340e-01
-2.45491520e-01 -1.56297639e-01 6.17908597e-01 -8.78480524e-02
-2.03961924e-01 -4.10351679e-02 -9.28848684e-01 -2.07355663e-01
8.52710247e-01 -4.66632470e-02 5.20736128e-02 -8.47814560e-01
-2.32032776e-01 2.42817640e-01 4.78654981e-01 1.00663006e-02
1.10206269e-01 -1.32920873e+00 -5.49722493e-01 2.99644768e-01
-2.12536916e-01 -5.67913473e-01 3.12106282e-01 1.48671913e+00
-2.43049502e-01 9.74912763e-01 1.35638013e-01 6.96184486e-02
-8.58059347e-01 9.34009016e-01 8.88850540e-02 -2.19852716e-01
-6.02775037e-01 -7.50384331e-02 4.85939562e-01 1.34008750e-01
-2.71070153e-01 -4.55516636e-01 9.15734395e-02 1.29334629e-01
2.74816543e-01 8.98945391e-01 -3.59650582e-01 -6.11984015e-01
-1.73203304e-01 6.01526976e-01 4.09099847e-01 -3.54609966e-01
1.19913769e+00 -6.38885081e-01 1.11783750e-01 6.63145185e-01
1.15154386e+00 2.81748652e-01 -1.12140656e+00 1.38304038e-02
4.91843611e-01 -3.49057853e-01 -2.06068084e-02 -2.10052654e-01
-7.13764191e-01 4.18715566e-01 2.83517372e-02 5.48819244e-01
1.01638484e+00 -5.66306710e-01 2.45688662e-01 2.20089510e-01
1.49793491e-01 -1.48976552e+00 -5.23018420e-01 3.06110084e-01
4.99376088e-01 -1.32493329e+00 3.62141468e-02 -7.13085607e-02
-9.04139876e-01 9.08494532e-01 -2.87957545e-02 -2.82015890e-01
8.98871303e-01 5.73776923e-02 5.58282062e-02 8.77886266e-02
-9.30828094e-01 -1.35350838e-01 9.15533975e-02 3.22855890e-01
4.80281025e-01 2.42361829e-01 -9.90787208e-01 5.00998914e-01
-1.81827381e-01 -2.45552793e-01 6.94714248e-01 6.64214849e-01
-1.00697810e-02 -7.80832112e-01 -5.43130577e-01 4.38755006e-01
-1.30929637e+00 6.20495202e-03 1.94570974e-01 1.69858038e+00
-5.01308680e-01 1.18313444e+00 4.84927505e-01 3.18722904e-01
2.96404392e-01 1.29007474e-01 -3.40900362e-01 -1.29140124e-01
-3.46108347e-01 9.59381521e-01 4.77216989e-02 -2.53065765e-01
-4.81680572e-01 -1.05391824e+00 -6.97495580e-01 -7.34728634e-01
-6.78535402e-01 2.00312361e-01 3.41847420e-01 8.86386812e-01
3.76191214e-02 3.20375292e-03 1.32445323e+00 -2.62915432e-01
-9.49142277e-01 -9.26938474e-01 -1.17039335e+00 2.04248607e-01
3.45324397e-01 -6.79468513e-01 -8.69135082e-01 -2.98748016e-01] | [6.435882568359375, 4.207743167877197] |
ab62031a-49ff-4bb6-ad74-89866d6cb797 | label-flipping-data-poisoning-attack-against | 2208.08433 | null | https://arxiv.org/abs/2208.08433v1 | https://arxiv.org/pdf/2208.08433v1.pdf | Label Flipping Data Poisoning Attack Against Wearable Human Activity Recognition System | Human Activity Recognition (HAR) is a problem of interpreting sensor data to human movement using an efficient machine learning (ML) approach. The HAR systems rely on data from untrusted users, making them susceptible to data poisoning attacks. In a poisoning attack, attackers manipulate the sensor readings to contaminate the training set, misleading the HAR to produce erroneous outcomes. This paper presents the design of a label flipping data poisoning attack for a HAR system, where the label of a sensor reading is maliciously changed in the data collection phase. Due to high noise and uncertainty in the sensing environment, such an attack poses a severe threat to the recognition system. Besides, vulnerability to label flipping attacks is dangerous when activity recognition models are deployed in safety-critical applications. This paper shades light on how to carry out the attack in practice through smartphone-based sensor data collection applications. This is an earlier research work, to our knowledge, that explores attacking the HAR models via label flipping poisoning. We implement the proposed attack and test it on activity recognition models based on the following machine learning algorithms: multi-layer perceptron, decision tree, random forest, and XGBoost. Finally, we evaluate the effectiveness of K-nearest neighbors (KNN)-based defense mechanism against the proposed attack. | ['Tauhidul Alam', 'Diane A. Igoche', 'Peter Y. Wu', 'Ahmed Imteaj', 'Abdur R. Shahid'] | 2022-08-17 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 8.20889294e-01 3.76228020e-02 -2.39022061e-01 -9.29075927e-02
-2.89344072e-01 -8.75925899e-01 4.47687447e-01 3.61435562e-01
-5.58165312e-01 8.13414931e-01 -9.67097282e-02 -6.73384428e-01
-1.23176403e-01 -9.25605953e-01 -8.19569468e-01 -8.70689213e-01
-2.11775810e-01 -7.61416703e-02 1.41180545e-01 2.26240873e-01
3.88875067e-01 6.38704777e-01 -1.21409893e+00 3.34392548e-01
3.62114370e-01 1.08717835e+00 -8.71277630e-01 7.82743752e-01
4.22738761e-01 1.03105760e+00 -1.05608642e+00 -2.53377557e-01
5.50296068e-01 -3.14675957e-01 -7.27702677e-01 -1.68573186e-01
-5.01538701e-02 -3.03984433e-01 7.53485933e-02 1.06694925e+00
4.02768701e-01 -3.01339716e-01 4.49015796e-01 -2.11759663e+00
-3.15828145e-01 5.32552063e-01 -5.76580465e-01 1.89227015e-01
7.61134267e-01 1.94043830e-01 1.88569382e-01 -1.26446545e-01
9.39791501e-02 8.73354614e-01 7.86042690e-01 6.30707741e-01
-1.01183641e+00 -1.18815839e+00 -3.82696867e-01 7.18600824e-02
-1.51896918e+00 -1.49762273e-01 8.22205961e-01 -2.87814528e-01
5.18371940e-01 7.75295436e-01 5.67081988e-01 1.69876349e+00
4.90921766e-01 4.72410947e-01 1.58745921e+00 -3.06727916e-01
9.12600100e-01 1.87495664e-01 4.36357111e-01 4.01891798e-01
9.08659637e-01 5.69260836e-01 -6.30802453e-01 -1.12320971e+00
7.82012865e-02 2.97824711e-01 -4.78237383e-02 -1.01865388e-01
-9.10161078e-01 7.25820065e-01 3.22622597e-01 -7.49937892e-02
-5.14531255e-01 7.12527335e-02 4.19121504e-01 1.99228182e-01
-1.72616959e-01 3.11208010e-01 -5.04827321e-01 -5.85378781e-02
-5.73786855e-01 -4.39338498e-02 1.03341508e+00 4.58956659e-01
2.96279192e-01 -9.71468538e-02 1.01395905e-01 -1.67739183e-01
6.63694084e-01 7.92804599e-01 6.66079104e-01 -5.18720210e-01
2.31699124e-01 6.21320009e-01 3.21066022e-01 -1.30514967e+00
-3.57679397e-01 -4.36478145e-02 -6.85251832e-01 3.82631212e-01
4.06538516e-01 -6.10677361e-01 -5.87015033e-01 1.35452390e+00
6.20574355e-01 5.57752311e-01 2.86425412e-01 6.12773478e-01
2.26440951e-01 3.56469125e-01 6.40795469e-01 -3.00663441e-01
1.39258134e+00 -4.49837029e-01 -7.43567765e-01 7.73804560e-02
7.72520721e-01 -1.69700027e-01 7.95659184e-01 7.62417197e-01
-2.99308568e-01 -1.23552538e-01 -1.45478749e+00 7.39406586e-01
-7.78387070e-01 -6.28507674e-01 6.64002180e-01 1.74757946e+00
-2.17037544e-01 2.75912404e-01 -1.02509177e+00 -5.75198710e-01
7.56184459e-01 6.48620725e-01 -3.30171704e-01 1.66537955e-01
-1.31959641e+00 6.76998615e-01 2.76087910e-01 -1.40840188e-01
-8.69414270e-01 -2.85003930e-01 -6.06196404e-01 -5.14062226e-01
3.09844345e-01 -2.38414615e-01 8.79156351e-01 -6.46105886e-01
-1.14979970e+00 5.25959194e-01 2.77188182e-01 -8.32087040e-01
5.38456917e-01 -1.93131432e-01 -8.78311634e-01 -8.65647495e-02
-1.50291711e-01 -4.20218967e-02 9.42002118e-01 -1.05008984e+00
-4.01283145e-01 -7.85238743e-01 -2.76807606e-01 -2.59210646e-01
-2.65729785e-01 2.46779602e-02 9.43245173e-01 -4.82472956e-01
-2.31239960e-01 -1.12918627e+00 -1.90205768e-01 -4.80467111e-01
-7.59697974e-01 2.36885861e-01 1.42378056e+00 -3.34894717e-01
1.33068943e+00 -2.17328668e+00 -8.46099555e-01 5.97654045e-01
-7.50603527e-02 5.89784205e-01 4.92573619e-01 6.18046165e-01
1.09670162e-01 4.31181401e-01 -2.81531334e-01 1.93811461e-01
-8.70761275e-02 4.59130734e-01 -5.34069061e-01 8.70406210e-01
-4.47295457e-01 7.02696741e-01 -7.21078157e-01 -2.17987433e-01
2.04016656e-01 3.05127770e-01 4.73364480e-02 7.34380335e-02
1.08414397e-01 4.62765515e-01 -7.53115475e-01 9.56153274e-01
5.81222475e-01 6.68336675e-02 1.48620605e-01 -3.93603705e-02
3.02407682e-01 -5.27541786e-02 -1.19940865e+00 8.45970273e-01
3.91920842e-02 -1.28758438e-02 -3.13568771e-01 -8.25123072e-01
9.27472651e-01 3.91103327e-01 4.54458177e-01 -2.30033055e-01
4.01761055e-01 4.79557440e-02 -1.78783104e-01 -7.02128172e-01
-7.94420913e-02 2.63642877e-01 -5.75588644e-01 9.17348325e-01
-5.93885899e-01 7.99848437e-01 -6.57893956e-01 -2.03994051e-01
1.84487545e+00 -1.29119813e-01 7.39021838e-01 1.74326777e-01
5.87309837e-01 1.41014427e-01 3.99704337e-01 1.11555779e+00
-7.26128042e-01 -1.34314150e-01 -1.12357944e-01 -4.70574379e-01
-3.99908811e-01 -1.16272736e+00 5.99188693e-02 9.89166021e-01
2.13476360e-01 -2.26812690e-01 -9.79591191e-01 -1.24950266e+00
2.30038583e-01 7.81397343e-01 -6.18917644e-01 -5.58101535e-01
-2.84953415e-01 -9.21626091e-01 1.66219389e+00 3.76079082e-01
1.13181746e+00 -1.14046478e+00 -1.24169922e+00 8.64884853e-02
7.23978877e-02 -8.96882474e-01 -8.17117805e-04 4.02584612e-01
-6.40912712e-01 -1.62842715e+00 3.49341720e-01 -1.96424007e-01
4.78953063e-01 1.30443737e-01 3.24114293e-01 1.47064731e-01
-2.22298682e-01 3.98411512e-01 -5.71967542e-01 -1.03099370e+00
-5.76873660e-01 -8.53883922e-02 3.83248597e-01 3.14214945e-01
1.28336287e+00 -5.10497808e-01 -4.02577132e-01 5.12519300e-01
-1.07542706e+00 -7.97955692e-01 3.98923904e-01 2.44233042e-01
1.42663643e-01 6.12882197e-01 5.55094779e-01 -1.07416892e+00
8.79473984e-01 -7.23272562e-01 -3.10588121e-01 2.94318795e-01
-7.88768291e-01 -2.36321837e-01 5.12983739e-01 -8.90871465e-01
-5.26842654e-01 5.04521310e-01 8.66795331e-02 -6.16747374e-03
-7.76096642e-01 1.41946182e-01 -6.10446393e-01 -3.90881956e-01
1.10176313e+00 7.72565454e-02 -1.22757688e-01 -3.39950383e-01
9.35743377e-03 1.19257593e+00 4.09288764e-01 -4.45144512e-02
8.93355608e-01 7.21979856e-01 1.61015958e-01 -8.07607830e-01
-4.94395077e-01 -3.56275141e-01 -1.72414437e-01 -1.66941673e-01
8.57446432e-01 -4.69560862e-01 -9.82049704e-01 1.00777268e+00
-8.45891178e-01 1.02233432e-01 1.15943357e-01 3.27545851e-01
-6.31423518e-02 4.32659864e-01 -3.20239782e-01 -1.18594837e+00
-5.48194408e-01 -7.34636903e-01 7.46383190e-01 1.19530298e-01
-6.62647367e-01 -7.80819476e-01 1.77901059e-01 8.63257706e-01
2.31923610e-01 9.41824853e-01 3.90184939e-01 -1.41582000e+00
-5.23453057e-02 -7.09845543e-01 5.04015386e-01 8.38775039e-02
3.71791750e-01 -1.97429806e-01 -1.15145755e+00 -1.58800110e-01
4.79207397e-01 -2.63909489e-01 1.10680044e-01 -8.70930329e-02
1.01427996e+00 -9.00865018e-01 -5.23022175e-01 2.15492412e-01
1.06168985e+00 6.11847043e-01 7.70464540e-01 5.54846466e-01
7.32262433e-01 3.97436500e-01 6.57697380e-01 5.73025763e-01
-2.58427039e-02 3.33277524e-01 6.21311963e-01 1.39106646e-01
6.97841406e-01 -6.10628545e-01 6.36889517e-01 -3.02611709e-01
3.81125659e-01 -2.06180945e-01 -8.64646554e-01 -1.56938568e-01
-1.88048315e+00 -1.02113891e+00 -3.16652268e-01 2.36158800e+00
6.41206086e-01 2.31813774e-01 4.24151361e-01 9.32021022e-01
7.29266346e-01 1.39774410e-02 -6.01072967e-01 -6.45255387e-01
2.21107364e-01 1.85167581e-01 1.21682954e+00 3.62563908e-01
-1.49697161e+00 6.33433461e-01 5.84886599e+00 4.08581793e-01
-9.83844101e-01 3.00826758e-01 3.05271298e-01 1.40242293e-01
3.82556260e-01 -1.47602595e-02 -6.85390174e-01 7.87557423e-01
1.18942583e+00 2.81910330e-01 2.97417283e-01 1.08753610e+00
3.71830553e-01 -3.01714242e-01 -8.62252116e-01 9.68950808e-01
1.14089116e-01 -7.95930803e-01 -2.39959031e-01 4.59255010e-01
1.38201073e-01 -2.27986008e-01 -2.21213788e-01 5.15001528e-02
7.12986231e-01 -1.21849155e+00 2.17054680e-01 1.62696838e-01
9.18515027e-02 -9.36449945e-01 7.82819211e-01 7.64812887e-01
-6.81648672e-01 -4.19454783e-01 -1.69058703e-02 -4.99937743e-01
4.98295501e-02 4.40891266e-01 -9.28270519e-01 2.38970034e-02
8.17641616e-01 -2.81645562e-02 -8.72646987e-01 5.99329710e-01
-4.14794147e-01 1.19797528e+00 -7.00890005e-01 -2.92814970e-01
-4.52130064e-02 3.25769395e-01 4.81946826e-01 9.07525420e-01
-1.47003159e-02 8.14744756e-02 1.98662922e-01 3.78161132e-01
1.68323219e-01 -1.93050221e-01 -1.17166948e+00 6.39330670e-02
8.08106840e-01 9.45200443e-01 -5.94887912e-01 7.20138475e-02
-6.00374304e-03 9.56798136e-01 -3.51106644e-01 1.43263027e-01
-9.16142344e-01 -1.83142334e-01 3.64416987e-01 3.75095457e-01
-1.28824458e-01 -1.70231052e-03 -5.74540615e-01 -6.14564240e-01
-1.93571016e-01 -1.09051466e+00 7.30600834e-01 -4.37426448e-01
-1.23453832e+00 1.14224575e-01 4.29053754e-02 -9.88393307e-01
-6.73702210e-02 -2.32584193e-01 -4.71071213e-01 2.66758621e-01
-7.54616559e-01 -1.19942713e+00 -2.35019743e-01 1.00102055e+00
-2.22569197e-01 -3.49628478e-01 1.01614213e+00 4.38997746e-02
-5.68087041e-01 7.79320776e-01 -3.70952636e-01 4.50197816e-01
4.04273182e-01 -7.77654052e-01 1.06888980e-01 1.01671565e+00
3.07330608e-01 6.89257801e-01 6.38416052e-01 -1.12016559e+00
-1.40978944e+00 -1.22310925e+00 6.25507176e-01 -7.71819830e-01
4.03529644e-01 -5.07406592e-01 -7.48453140e-01 7.95001745e-01
5.46081038e-03 1.48903370e-01 1.55646288e+00 -3.84771675e-01
-4.53620434e-01 -9.83677506e-02 -1.96920240e+00 3.23602438e-01
6.78775370e-01 -4.77757603e-01 -6.24760628e-01 1.09334230e-01
2.40597978e-01 1.44265309e-01 -7.18896747e-01 3.82993460e-01
7.67480850e-01 -7.49759912e-01 7.60572016e-01 -8.71443331e-01
-4.66646045e-01 -5.83795846e-01 -2.97754973e-01 -7.62360811e-01
7.72522092e-02 -8.81082237e-01 -6.00420415e-01 1.23573923e+00
2.02913910e-01 -7.91225195e-01 9.44803476e-01 7.93071151e-01
8.11272144e-01 -1.51105866e-01 -9.41584229e-01 -6.28759503e-01
-3.07030082e-01 -5.67065001e-01 9.30910289e-01 1.20920479e+00
2.67725050e-01 2.44376928e-01 -6.58714354e-01 8.05220544e-01
9.56999421e-01 -7.32645214e-01 9.40259874e-01 -1.13407290e+00
-2.92314202e-01 3.99501950e-01 -7.04678953e-01 -2.57171392e-01
-3.33938479e-01 -3.07152629e-01 -3.06443930e-01 -5.36827326e-01
-2.91530579e-01 -2.41157711e-01 -4.41325933e-01 8.33661497e-01
1.55391231e-01 5.36016047e-01 7.34640956e-02 3.13568294e-01
-5.54948509e-01 -1.35307208e-01 2.25700319e-01 -1.19167604e-01
-3.38593006e-01 5.07305026e-01 -7.10649431e-01 8.64043891e-01
1.41017854e+00 -1.17535150e+00 -5.79166353e-01 3.40052247e-01
7.33319074e-02 -2.65248835e-01 6.42679811e-01 -1.28719044e+00
4.01450455e-01 -3.21617007e-01 4.70032185e-01 -3.39247853e-01
-1.38960406e-01 -1.61666477e+00 5.92913926e-01 9.33572590e-01
-2.78970152e-01 1.17331877e-01 -1.32332981e-01 8.61953497e-01
3.88622910e-01 -1.34441376e-01 6.04127049e-01 -9.75075066e-02
-3.41404110e-01 -1.07457899e-01 -8.96872342e-01 -5.47345459e-01
1.71975386e+00 -7.10792124e-01 -4.95298326e-01 -9.55164060e-02
-5.24892807e-01 -8.07445198e-02 4.91250664e-01 3.45019341e-01
5.23513615e-01 -1.19852901e+00 -3.32289608e-03 4.76552308e-01
2.86409378e-01 -4.91519034e-01 -4.68961805e-01 6.16698623e-01
-4.49571609e-01 -7.86436647e-02 -2.31934935e-01 -2.71821439e-01
-1.72003496e+00 8.24496746e-01 2.59187996e-01 -1.63270712e-01
-2.05341652e-01 1.85234800e-01 -7.77733624e-01 -2.67414063e-01
6.29027247e-01 8.14038217e-02 -1.43514395e-01 -3.11829597e-01
7.78482080e-01 9.42790747e-01 1.40218716e-03 -5.61016500e-01
-8.77362430e-01 3.30464132e-02 1.73982456e-01 -2.09897548e-01
5.02286375e-01 -2.64476091e-02 1.13338605e-01 3.16830873e-01
9.69926417e-01 1.72647774e-01 -6.97957397e-01 1.95434585e-01
3.75919372e-01 -4.69310433e-01 -7.75276348e-02 -1.15456390e+00
-6.89897299e-01 3.12675953e-01 1.02522504e+00 6.11237645e-01
8.85423481e-01 -5.34642041e-01 8.75357151e-01 5.74233770e-01
6.91647351e-01 -8.64930153e-01 -1.03596710e-01 -2.65934318e-01
3.51238139e-02 -1.25555348e+00 5.54223880e-02 -2.51939207e-01
-7.32745767e-01 5.59322536e-01 4.56261903e-01 -1.39089540e-01
1.10247433e+00 6.35196984e-01 3.56931001e-01 -1.54835388e-01
-1.48096621e-01 5.04082739e-01 -4.70188230e-01 1.27971005e+00
-3.69045585e-01 3.25143397e-01 -2.70220280e-01 7.39335358e-01
-7.57495761e-02 3.40393215e-01 6.17398918e-01 1.60390306e+00
-5.49889565e-01 -1.16871440e+00 -9.90379632e-01 3.65130603e-01
-8.31473053e-01 4.35494393e-01 -1.10017800e+00 5.34250617e-01
4.20725226e-01 1.59838450e+00 -4.25687194e-01 -9.09211695e-01
2.49211058e-01 3.14288378e-01 -1.69680938e-01 -1.87423557e-01
-9.91936624e-01 -5.58733642e-01 1.45887077e-01 -6.44152343e-01
-5.46268821e-01 -6.18043721e-01 -1.18358231e+00 -5.91905773e-01
-1.61036596e-01 2.44329765e-01 7.02355683e-01 1.01842666e+00
3.45377892e-01 -7.65621066e-02 8.12350929e-01 -1.23350233e-01
-8.70107889e-01 -8.24945331e-01 -7.37620711e-01 5.47064722e-01
2.32407629e-01 -7.01712072e-01 -5.80740213e-01 -8.28897282e-02] | [5.586164474487305, 7.1304402351379395] |
1e7c2be8-ff60-473d-8c71-d765895ab143 | correlated-time-series-self-supervised | 2306.06994 | null | https://arxiv.org/abs/2306.06994v2 | https://arxiv.org/pdf/2306.06994v2.pdf | Correlated Time Series Self-Supervised Representation Learning via Spatiotemporal Bootstrapping | Correlated time series analysis plays an important role in many real-world industries. Learning an efficient representation of this large-scale data for further downstream tasks is necessary but challenging. In this paper, we propose a time-step-level representation learning framework for individual instances via bootstrapped spatiotemporal representation prediction. We evaluated the effectiveness and flexibility of our representation learning framework on correlated time series forecasting and cold-start transferring the forecasting model to new instances with limited data. A linear regression model trained on top of the learned representations demonstrates our model performs best in most cases. Especially compared to representation learning models, we reduce the RMSE, MAE, and MAPE by 37%, 49%, and 48% on the PeMS-BAY dataset, respectively. Furthermore, in real-world metro passenger flow data, our framework demonstrates the ability to transfer to infer future information of new cold-start instances, with gains of 15%, 19%, and 18%. The source code will be released under the GitHub https://github.com/bonaldli/Spatiotemporal-TS-Representation-Learning | ['Fugee Tsung', 'Rui Zhao', 'Ziyue Li', 'Lei Bai', 'Luxuan Wang'] | 2023-06-12 | null | null | null | null | ['correlated-time-series-forecasting', 'time-series'] | ['time-series', 'time-series'] | [-1.10569783e-01 -2.13216990e-01 -1.91179439e-01 -4.86845434e-01
-1.09437037e+00 -4.86464798e-01 5.33422053e-01 9.98360217e-02
1.22919768e-01 7.64333248e-01 2.26654395e-01 -4.43992525e-01
-4.26973671e-01 -8.51504564e-01 -6.36945367e-01 -7.07137048e-01
-4.47848499e-01 1.87462851e-01 -1.88027665e-01 -3.02697182e-01
3.86225176e-03 5.49030364e-01 -1.44307494e+00 4.18179870e-01
8.70841801e-01 1.25186825e+00 1.77063107e-01 4.18803602e-01
8.69673267e-02 1.05850482e+00 -5.17658532e-01 3.70416827e-02
3.47060084e-01 -8.55859071e-02 -2.22132877e-01 -2.82668680e-01
-1.49260089e-01 3.65277492e-02 -6.22207999e-01 3.45320672e-01
2.20910639e-01 6.55808508e-01 5.41164935e-01 -1.27587080e+00
-4.00950283e-01 5.24114728e-01 -5.07137060e-01 4.36905086e-01
-5.91736771e-02 6.16140887e-02 5.87746441e-01 -7.98480749e-01
3.72581393e-01 7.41057754e-01 8.26538920e-01 5.38156368e-02
-9.72559810e-01 -1.01240110e+00 3.64353001e-01 3.27670425e-01
-1.32260549e+00 -2.46165857e-01 7.27656186e-01 -5.06425738e-01
1.17626536e+00 3.07275981e-01 4.34470147e-01 1.07886243e+00
3.83438110e-01 5.78952968e-01 9.43595350e-01 2.07829718e-02
5.23983836e-02 -7.23484382e-02 2.59086698e-01 2.16380596e-01
-3.26649621e-02 4.88644809e-01 -4.15392250e-01 1.65537536e-01
4.30491745e-01 8.12057793e-01 9.94041786e-02 2.94111788e-01
-1.14486456e+00 6.07474625e-01 5.95399380e-01 3.45905364e-01
-8.05079460e-01 4.14026499e-01 4.31236833e-01 2.65264750e-01
1.17012525e+00 4.50442314e-01 -7.68187940e-01 -6.03664279e-01
-9.73976672e-01 7.67914653e-02 4.13706869e-01 9.48939860e-01
6.72066271e-01 3.48574817e-01 -1.60778552e-01 7.24786878e-01
-4.11670864e-01 7.11495876e-01 5.77937901e-01 -7.26153672e-01
8.18495154e-01 3.48584145e-01 1.35920331e-01 -9.58514750e-01
-6.39425874e-01 -8.31539154e-01 -1.02184224e+00 -4.33648050e-01
4.36068177e-02 -5.83954155e-01 -8.63715231e-01 1.44202507e+00
-5.57725038e-03 1.08469057e+00 -1.01637179e-02 6.91138864e-01
4.21925306e-01 1.35648704e+00 -6.33619130e-02 -4.19291973e-01
8.61026466e-01 -1.02968740e+00 -8.10103834e-01 -3.42525393e-02
7.41434395e-01 -5.63588917e-01 6.23637557e-01 3.19220573e-01
-8.02269578e-01 -6.72946393e-01 -5.74141800e-01 3.31974894e-01
-6.71248615e-01 1.69869155e-01 8.62441003e-01 1.53973565e-01
-5.15632272e-01 6.89543247e-01 -1.02102697e+00 -9.94154438e-02
2.54689187e-01 4.79727425e-02 -6.61532134e-02 -2.92628348e-01
-1.16936707e+00 7.01326251e-01 1.65240452e-01 3.07635099e-01
-8.25949907e-01 -1.38559723e+00 -6.26442611e-01 3.69625986e-01
3.01993519e-01 -2.03304857e-01 1.06413114e+00 -4.74204212e-01
-1.23187149e+00 -6.56407177e-02 -4.06180024e-01 -5.77269077e-01
3.43474060e-01 -4.96915698e-01 -1.10137749e+00 -1.44515708e-01
3.22502069e-02 -2.59073973e-02 4.83517110e-01 -8.79592955e-01
-7.49313295e-01 -1.92655355e-01 -4.14306790e-01 -2.63396710e-01
-2.19362482e-01 -3.07041377e-01 -1.54313520e-01 -9.50535476e-01
-6.43353583e-03 -1.15830874e+00 -4.91383195e-01 -5.18728077e-01
6.86286986e-02 -1.25508234e-01 7.82554686e-01 -9.58169639e-01
1.35991454e+00 -2.36514759e+00 -2.35109359e-01 3.02719861e-01
-1.78957656e-01 1.10465288e-01 -3.06105077e-01 8.38806808e-01
-2.79013008e-01 -8.17561001e-02 1.19249024e-04 -2.05092475e-01
-2.22474366e-01 2.34634653e-01 -9.30286586e-01 2.42051765e-01
2.46473730e-01 8.61596346e-01 -9.34893608e-01 3.10953707e-01
5.32794714e-01 5.90088964e-01 -1.71122789e-01 1.86023563e-01
5.95052652e-02 7.36013889e-01 -4.15908068e-01 5.49021006e-01
6.39522314e-01 -3.58900338e-01 1.04377441e-01 1.05151489e-01
-2.24020854e-01 2.03788131e-01 -9.06146526e-01 1.69937193e+00
-1.04919457e+00 8.14137280e-01 -6.59580052e-01 -1.13225257e+00
1.26432884e+00 3.76199841e-01 9.02467728e-01 -1.15988183e+00
-2.82824695e-01 9.69926938e-02 -2.57588983e-01 -3.88332814e-01
5.40980697e-01 -1.40665164e-02 -1.93410337e-01 4.46693271e-01
-3.25773865e-01 2.06525505e-01 6.96782917e-02 -3.27042043e-02
1.02329993e+00 4.74971533e-02 5.33648441e-03 1.27884179e-01
2.47696057e-01 -1.05072819e-01 6.98328614e-01 3.55860740e-01
1.38251022e-01 4.73028570e-01 3.44469607e-01 -7.97753632e-01
-6.71150982e-01 -9.17266309e-01 -7.48661831e-02 1.19290686e+00
-1.71736732e-01 -5.84068477e-01 9.93027240e-02 -3.95179212e-01
4.03633833e-01 1.28334820e+00 -6.32474899e-01 -3.68917793e-01
-6.13242805e-01 -7.00436711e-01 1.44488633e-01 8.93657029e-01
9.90259945e-02 -7.83850014e-01 -3.62410009e-01 3.56379926e-01
-2.71620691e-01 -1.06473243e+00 -5.35613596e-02 2.50729233e-01
-9.70477104e-01 -7.09551871e-01 -7.23394752e-01 -2.23157376e-01
5.02040446e-01 5.30703008e-01 8.48285735e-01 -2.96808749e-01
-1.53226197e-01 1.48724660e-01 -4.20888126e-01 -5.85991323e-01
-9.64906503e-05 3.13065588e-01 -1.18711419e-01 6.61997125e-02
1.85768127e-01 -6.57552779e-01 -6.60227001e-01 4.55639541e-01
-4.83576179e-01 -7.32280388e-02 3.40451390e-01 7.62451708e-01
6.74692512e-01 4.69943658e-02 1.01176119e+00 -8.36886287e-01
4.26564693e-01 -1.21640491e+00 -7.60361254e-01 1.12731770e-01
-8.32395673e-01 -2.42116630e-01 8.17753911e-01 -3.74975801e-01
-1.17103529e+00 -5.62339686e-02 9.55909193e-02 -8.13418388e-01
6.09237850e-02 9.08541739e-01 7.33257234e-01 5.32269359e-01
4.88244265e-01 3.18784446e-01 -2.12079436e-01 -7.41394818e-01
4.05811489e-01 6.05306804e-01 2.86774158e-01 -5.09085298e-01
8.49089742e-01 3.67853463e-01 3.56091931e-02 -5.42852283e-01
-8.12745154e-01 -7.20712066e-01 -4.53947484e-01 -2.98517704e-01
3.82197887e-01 -1.39208663e+00 -5.56840360e-01 8.82648826e-02
-7.29093790e-01 -4.82745290e-01 -4.74680394e-01 6.80141032e-01
-4.85712171e-01 -1.67316154e-01 -3.44177544e-01 -9.31435764e-01
-2.72912055e-01 -7.50922978e-01 1.01072478e+00 -9.59149152e-02
-8.64720624e-03 -9.71764565e-01 1.46981984e-01 3.13023329e-01
8.54585648e-01 5.87663591e-01 7.61027873e-01 -4.91588205e-01
-4.58891034e-01 -6.36980534e-01 -1.49089932e-01 1.74643233e-01
3.75797182e-01 -1.34204283e-01 -1.07904553e+00 -4.21403468e-01
-4.81596112e-01 -1.95398019e-03 9.44666505e-01 3.55157405e-01
1.86481738e+00 -1.33975372e-01 -3.97753507e-01 7.86776543e-01
1.23224425e+00 5.12349784e-01 4.73158568e-01 2.70875078e-02
5.04467249e-01 5.41945994e-01 9.57862437e-01 9.02502358e-01
2.74967521e-01 5.80040038e-01 4.29140240e-01 5.21469628e-03
6.58584014e-02 -2.16679558e-01 2.91204721e-01 1.06065667e+00
-4.71134990e-01 -2.29303703e-01 -1.12511826e+00 6.18193448e-01
-2.23432946e+00 -1.23475778e+00 -9.80739743e-02 2.13764715e+00
1.42474502e-01 -1.04506075e-01 -1.37733966e-02 2.94268299e-02
4.15808380e-01 3.16143185e-01 -7.35754132e-01 -2.19287634e-01
2.49702111e-01 1.19669646e-01 5.18830955e-01 4.16772813e-02
-9.24922347e-01 7.54989266e-01 5.68022394e+00 5.78780591e-01
-1.41713274e+00 2.28524655e-01 8.17903399e-01 -4.68046278e-01
-2.09984824e-01 -1.94244996e-01 -5.08925498e-01 7.43316531e-01
1.70597267e+00 -5.44196904e-01 7.01086581e-01 8.90884459e-01
6.53096199e-01 4.29855019e-01 -9.19074059e-01 1.04129982e+00
-2.63960123e-01 -1.69214880e+00 -2.50431329e-01 3.13837267e-02
1.00107896e+00 5.55963159e-01 3.49137604e-01 8.97922516e-01
9.87353548e-02 -1.07626283e+00 3.73827964e-01 8.51455867e-01
8.36894393e-01 -1.08433998e+00 8.98288369e-01 3.26123476e-01
-1.59107161e+00 -5.31607926e-01 -3.32276672e-01 -1.45780802e-01
3.43256086e-01 8.37266445e-01 -7.72167981e-01 1.00046933e+00
6.75133705e-01 1.36341882e+00 -3.98101389e-01 8.55437219e-01
2.61262283e-02 1.08145607e+00 -2.29831576e-01 2.99468786e-01
2.02237815e-01 -2.90499061e-01 2.35027120e-01 1.20221210e+00
9.87930894e-01 3.83009054e-02 1.00447558e-01 5.24627388e-01
-2.30120867e-02 -6.52158335e-02 -8.09735596e-01 -1.17740713e-01
5.00097156e-01 1.19466043e+00 -1.31832823e-01 -3.68185967e-01
-3.77944350e-01 5.85324407e-01 2.22916707e-01 7.75630295e-01
-1.19113839e+00 -3.65466237e-01 9.05822098e-01 1.21491715e-01
4.01517838e-01 -3.65893632e-01 -1.58644363e-01 -1.16715598e+00
6.76731020e-02 -5.85398853e-01 4.44949061e-01 -8.27807367e-01
-1.35499716e+00 6.02417946e-01 9.13125053e-02 -1.74249864e+00
-7.68632770e-01 -4.04085904e-01 -8.22528124e-01 9.60609972e-01
-1.72111630e+00 -1.15398741e+00 -4.78400886e-01 5.53831398e-01
7.05764711e-01 -4.21618491e-01 7.35691309e-01 5.00125170e-01
-7.73920178e-01 2.90273100e-01 6.86502278e-01 -7.53462985e-02
5.15860140e-01 -8.18579972e-01 5.38140535e-01 7.90617526e-01
1.04139253e-01 5.43844461e-01 3.90420407e-01 -2.54065692e-01
-1.51350915e+00 -1.88084280e+00 9.01364446e-01 -2.82764137e-01
8.56737554e-01 -3.26511949e-01 -1.01661944e+00 9.87159431e-01
3.92265506e-02 3.46736938e-01 8.47495496e-01 5.46000242e-01
-3.52980018e-01 -7.62283266e-01 -8.33712101e-01 2.95666039e-01
8.66362631e-01 -4.71177369e-01 -2.10119188e-01 7.36820877e-01
8.72344553e-01 -3.09447765e-01 -1.41971159e+00 3.85226071e-01
5.15120685e-01 -5.00964940e-01 8.75796854e-01 -7.16225505e-01
4.24816966e-01 -1.94313094e-01 -2.27111921e-01 -1.55167377e+00
-6.08824492e-01 -4.68184233e-01 -4.21126157e-01 1.09308326e+00
5.78455210e-01 -1.00842738e+00 3.71868759e-01 3.83641839e-01
-4.25029337e-01 -7.67814279e-01 -1.10835052e+00 -1.06236184e+00
1.71129182e-01 -8.62457871e-01 1.10165250e+00 1.07255042e+00
1.33272512e-02 -7.08812550e-02 -5.86854458e-01 3.32119852e-01
1.57548204e-01 5.19317746e-01 7.85966396e-01 -1.06962097e+00
-1.59558713e-01 -2.07710654e-01 -2.12830842e-01 -7.25908875e-01
2.77091444e-01 -9.74898815e-01 -2.75192350e-01 -1.39855862e+00
-2.72252291e-01 -5.16339362e-01 -1.16037571e+00 6.32810533e-01
1.12953454e-01 -1.12812601e-01 3.90127331e-01 2.34562576e-01
-4.27474767e-01 6.75973535e-01 9.18316126e-01 -2.39553839e-01
-2.26376042e-01 3.64237487e-01 -3.23145479e-01 1.12499490e-01
1.08196104e+00 -5.17757416e-01 -7.42755413e-01 -4.08920974e-01
1.04458548e-01 4.35047269e-01 7.92275518e-02 -1.13852346e+00
6.13552295e-02 -3.94024193e-01 4.93653148e-01 -7.87720501e-01
5.67565858e-01 -9.57539856e-01 5.77722847e-01 5.07381201e-01
-3.86491030e-01 4.00884002e-01 5.51426589e-01 8.25134277e-01
-3.52439612e-01 5.28063536e-01 1.19422674e-01 1.64764702e-01
-9.71312881e-01 4.74778652e-01 -3.68916005e-01 -2.79733747e-01
1.29224229e+00 1.11567453e-01 -4.90988225e-01 -5.85038543e-01
-7.25812376e-01 4.63808119e-01 -2.21839026e-01 7.95268714e-01
5.95933557e-01 -1.32766175e+00 -6.87583685e-01 2.50456303e-01
3.99796814e-01 -1.47291332e-01 5.78869641e-01 9.45539296e-01
-1.60640061e-01 5.88647962e-01 -1.20129632e-02 -4.74468529e-01
-6.84728444e-01 7.24570334e-01 1.24490805e-01 -4.03279126e-01
-7.04452932e-01 2.78275430e-01 -2.62775011e-02 -4.21010137e-01
-1.25547558e-01 -8.23686600e-01 -3.51448208e-02 1.30702674e-01
5.14941871e-01 7.44205475e-01 2.64259547e-01 -3.80344450e-01
-3.69100332e-01 4.39939976e-01 7.60658607e-02 2.74054199e-01
1.79155767e+00 9.30879414e-02 3.80883038e-01 8.64726305e-01
1.33557439e+00 -3.20184946e-01 -1.28835940e+00 -1.96346477e-01
2.22211368e-02 -4.52078849e-01 1.56456232e-01 -9.47236717e-01
-1.44641304e+00 9.07423794e-01 7.15011060e-01 1.43397689e-01
1.15507734e+00 -2.64638066e-01 9.37167048e-01 4.87661749e-01
4.44131821e-01 -7.94442594e-01 -2.64720023e-01 5.10737598e-01
1.03353798e+00 -1.24519217e+00 -3.31432931e-02 -1.33807436e-01
-8.59122157e-01 1.01015162e+00 3.33563507e-01 -2.01434463e-01
1.00843668e+00 1.12110145e-01 8.52974132e-02 -3.86063457e-02
-1.34900677e+00 3.09737753e-02 3.61725539e-01 4.61721867e-01
4.33640748e-01 4.06181514e-01 2.03077123e-01 6.05978251e-01
-1.11445144e-01 2.14030251e-01 1.90329626e-01 6.10890746e-01
9.41034630e-02 -6.48488224e-01 2.65102293e-02 6.09872341e-01
-1.29681438e-01 3.69962603e-02 4.22171831e-01 7.04426110e-01
-1.30160034e-01 1.02926695e+00 5.27505338e-01 -5.69975019e-01
4.61913854e-01 1.30936205e-01 -3.10762703e-01 -4.82877612e-01
-6.93918169e-01 -1.98510699e-02 8.86950493e-02 -8.15363884e-01
-2.16280907e-01 -6.50768578e-01 -1.29510820e+00 -4.17172402e-01
1.07752107e-01 1.58996716e-01 7.97712922e-01 7.34362006e-01
9.00978863e-01 9.09841776e-01 1.08764851e+00 -8.31476867e-01
-4.84994054e-01 -9.89539921e-01 -5.63822389e-01 4.49227750e-01
4.07077342e-01 -5.49237728e-01 -2.91356504e-01 -9.26463306e-02] | [6.9184675216674805, 2.8826940059661865] |
905f1a6f-cd3f-4d21-abc1-2371bd8e5631 | 3d-pick-mix-object-part-blending-in-joint | 1811.01068 | null | http://arxiv.org/abs/1811.01068v1 | http://arxiv.org/pdf/1811.01068v1.pdf | 3D Pick & Mix: Object Part Blending in Joint Shape and Image Manifolds | We present 3D Pick & Mix, a new 3D shape retrieval system that provides users
with a new level of freedom to explore 3D shape and Internet image collections
by introducing the ability to reason about objects at the level of their
constituent parts. While classic retrieval systems can only formulate simple
searches such as "find the 3D model that is most similar to the input image"
our new approach can formulate advanced and semantically meaningful search
queries such as: "find me the 3D model that best combines the design of the
legs of the chair in image 1 but with no armrests, like the chair in image 2".
Many applications could benefit from such rich queries, users could browse
through catalogues of furniture and pick and mix parts, combining for example
the legs of a chair from one shop and the armrests from another shop. | ['Adrian Penate-Sanchez', 'Lourdes Agapito'] | 2018-11-02 | null | null | null | null | ['3d-shape-retrieval'] | ['computer-vision'] | [-1.93911850e-01 -4.55758758e-02 7.01445192e-02 -5.71288802e-02
-6.98947728e-01 -9.61626291e-01 4.46495175e-01 3.57202739e-01
3.07360172e-01 -4.12992500e-02 -6.60550641e-03 -5.31536818e-01
-7.64737010e-01 -6.61832750e-01 -1.78013548e-01 -4.35135543e-01
4.35101315e-02 8.89183700e-01 4.93182987e-01 -5.69029808e-01
3.94236177e-01 1.22816396e+00 -1.67662799e+00 2.00400218e-01
4.00088951e-02 1.17105317e+00 6.43224716e-01 3.89929086e-01
-3.67022842e-01 -1.50355354e-01 -4.56163794e-01 -5.09248853e-01
3.34720492e-01 2.07710266e-03 -9.20195937e-01 4.58724886e-01
3.49060088e-01 -2.44962677e-01 -1.32542357e-01 8.08687925e-01
4.20541883e-01 1.06492966e-01 5.99361539e-01 -9.81986880e-01
-5.60304582e-01 -2.42446452e-01 -5.65980256e-01 -1.94872603e-01
1.10256040e+00 -1.63320854e-01 9.97085512e-01 -8.64175081e-01
1.08848786e+00 1.40889347e+00 1.78086415e-01 2.53833979e-01
-1.06140709e+00 -6.00345340e-03 1.38068780e-01 -3.96177806e-02
-1.59045589e+00 -1.78660631e-01 8.77414763e-01 -5.80504872e-02
8.65905762e-01 8.74901414e-01 6.42546177e-01 2.77409077e-01
2.45282575e-02 9.70019817e-01 6.17828369e-01 -5.43480158e-01
1.09683834e-01 3.32408220e-01 -8.42502117e-02 8.98088694e-01
4.28235345e-02 -4.11897153e-01 -3.04783463e-01 -4.30919766e-01
1.11462474e+00 2.35083103e-01 1.59709573e-01 -9.01852310e-01
-1.07172883e+00 4.95999187e-01 6.43027782e-01 3.95986736e-01
-3.49010438e-01 -2.81108022e-02 -8.64119306e-02 1.86388075e-01
1.22679338e-01 6.40856147e-01 -5.98649859e-01 2.41805106e-01
-6.70868635e-01 6.39880598e-01 8.87945116e-01 1.45234358e+00
9.17412877e-01 -6.39424026e-01 2.93453902e-01 7.37333357e-01
5.54321349e-01 4.85637814e-01 -6.41685128e-02 -1.12786877e+00
2.17215300e-01 1.05563760e+00 2.26632133e-01 -9.68010008e-01
-2.80821353e-01 -8.81112814e-02 -2.75570899e-01 3.53512824e-01
-2.20813304e-02 3.92772794e-01 -7.66431510e-01 9.53344107e-01
5.46140611e-01 -7.80811846e-01 -2.50561208e-01 9.45466161e-01
1.21935165e+00 2.12699756e-01 -1.93869531e-01 1.82621390e-01
1.98023057e+00 -6.13881826e-01 -1.63380548e-01 8.83764122e-03
4.23615038e-01 -1.19762671e+00 8.51661563e-01 9.02846903e-02
-1.42285895e+00 -4.83822286e-01 -1.05012763e+00 -3.01689953e-01
-8.06552708e-01 2.75365915e-02 6.84758842e-01 2.86962688e-01
-1.09632623e+00 5.04469514e-01 -4.20583159e-01 -6.32609487e-01
1.24521397e-01 3.72411817e-01 -4.81569767e-01 -3.02118838e-01
-5.38726807e-01 9.80593443e-01 1.14896253e-01 -2.50121504e-01
-5.02525270e-01 -3.72658879e-01 -7.60354996e-01 1.05263054e-01
8.29386473e-01 -1.11297905e+00 1.22910511e+00 -3.02909195e-01
-7.81475782e-01 1.42612922e+00 -3.11347991e-01 1.86402068e-01
2.11272582e-01 -2.45509502e-02 -1.73684046e-01 2.84930646e-01
3.51394922e-01 7.64280200e-01 6.37319148e-01 -1.49349749e+00
-3.15790862e-01 -8.66748333e-01 6.70099318e-01 4.94382679e-01
5.07925689e-01 1.24791779e-01 -9.24215674e-01 -4.54561174e-01
8.90362680e-01 -7.94241846e-01 -2.59769231e-01 6.03576422e-01
-3.93586695e-01 -3.03452730e-01 1.33451152e+00 -4.20529306e-01
8.98950100e-01 -2.16477680e+00 -1.21553577e-01 5.76317787e-01
1.06150797e-02 -1.09476242e-02 9.85582694e-02 7.01783895e-01
-5.14034703e-02 4.62770045e-01 2.99408585e-01 -1.00713633e-01
3.73349935e-01 2.17420638e-01 1.42124638e-01 -1.83447331e-01
-5.56379650e-03 7.83345997e-01 -9.62117791e-01 -6.58822119e-01
2.82407880e-01 1.89256981e-01 -1.77085325e-01 -1.33741975e-01
-4.42417026e-01 -3.03944051e-02 -9.08088624e-01 1.08168221e+00
6.83059633e-01 -4.67757344e-01 1.09934866e-01 -2.17416421e-01
9.40591991e-02 1.50589541e-01 -1.39869964e+00 1.87140143e+00
-1.20108172e-01 1.68812752e-01 3.91835988e-01 -4.37555909e-01
1.12936795e+00 3.75377059e-01 6.37327194e-01 -6.57295287e-01
-2.19384387e-01 2.75055587e-01 -7.88473964e-01 -4.40699935e-01
6.11228168e-01 -3.68590988e-02 -2.98625529e-01 6.24341428e-01
-5.41907072e-01 -1.04775989e+00 -1.94566511e-03 3.38737011e-01
7.99386978e-01 3.50036651e-01 2.74548918e-01 -2.20420614e-01
5.11463106e-01 9.10870135e-02 -3.72785717e-01 6.24294102e-01
3.06569099e-01 6.79730177e-01 1.69279113e-01 -5.95882952e-01
-9.93201971e-01 -1.40515983e+00 1.38935205e-02 7.95541286e-01
5.84136903e-01 -4.94488358e-01 -1.90616935e-01 -4.97821122e-01
1.41647846e-01 2.45251104e-01 -1.30390406e-01 4.53587621e-01
-1.61350310e-01 1.54757529e-01 -1.60461888e-01 2.22921908e-01
5.12943625e-01 -8.30046654e-01 -7.34153271e-01 -1.46167189e-01
-9.24191773e-02 -7.21785426e-01 -4.29886878e-01 1.77631825e-01
-1.03849399e+00 -9.78323936e-01 -7.03817248e-01 -1.05828381e+00
9.11549985e-01 6.02935374e-01 1.36587918e+00 6.61390543e-01
-5.67778468e-01 8.00685823e-01 -3.29867363e-01 -3.08215499e-01
-4.86947969e-02 -2.00441424e-02 -2.52013445e-01 -4.89407569e-01
6.97943717e-02 -5.35488307e-01 -5.78435659e-01 6.01492107e-01
-1.05084038e+00 1.28375351e-01 3.22090417e-01 2.35666886e-01
8.75934780e-01 3.18439096e-01 1.40301660e-01 -3.56053919e-01
6.08366251e-01 -4.02021468e-01 -3.64516377e-01 5.17937660e-01
-2.55575269e-01 2.51955688e-01 -7.13961720e-02 -3.92478965e-02
-7.72055387e-01 2.82869458e-01 -4.21275757e-02 -3.41285497e-01
-3.84406924e-01 2.71496564e-01 -5.29465377e-01 3.31752701e-03
3.87489200e-01 1.48144007e-01 -2.46920399e-02 -9.82750714e-01
5.32883465e-01 3.49515259e-01 4.89844531e-01 -7.41963565e-01
6.03926539e-01 5.18621087e-01 3.34126562e-01 -7.84095883e-01
-3.49834830e-01 -8.28814447e-01 -9.50369239e-01 -1.76669672e-01
6.06618643e-01 -4.84470129e-01 -8.86559784e-01 -1.45078510e-01
-1.14130497e+00 4.68480289e-01 -4.66812104e-01 6.09899238e-02
-8.34358275e-01 3.52183431e-01 -9.29076225e-02 -7.84902036e-01
-7.76294470e-02 -8.84661794e-01 1.61272633e+00 2.66952723e-01
-4.51228052e-01 -7.46739209e-01 -5.90960383e-01 3.93204451e-01
2.15216175e-01 3.42039376e-01 1.48459828e+00 -6.10645652e-01
-1.21429706e+00 -4.89393473e-01 -1.55170038e-01 -3.15919787e-01
2.76933014e-01 -1.28675073e-01 -5.06345272e-01 -1.18211061e-01
-3.01465720e-01 1.39959455e-01 1.18164614e-01 1.32093221e-01
7.96561182e-01 -1.37988880e-01 -7.39813030e-01 1.03377797e-01
1.56038868e+00 5.34131289e-01 6.43236935e-01 2.16392606e-01
1.35396421e-01 7.49542058e-01 4.71439421e-01 2.77482808e-01
2.20287278e-01 1.01554096e+00 5.84389508e-01 -3.01233195e-02
-1.21288329e-01 -2.52729088e-01 -3.79745096e-01 1.94654256e-01
4.23474796e-02 -1.29678413e-01 -6.22200251e-01 4.91185755e-01
-1.61432481e+00 -6.80239379e-01 7.01369718e-02 2.23502731e+00
1.65440887e-01 5.05940914e-02 1.31995291e-01 6.92468137e-02
3.86725366e-01 -2.68190410e-02 -6.38671637e-01 -4.65956658e-01
1.08973883e-01 1.17608286e-01 6.83488473e-02 3.37580591e-01
-8.01584125e-01 5.93635678e-01 7.09461355e+00 9.14992929e-01
-3.98383081e-01 -4.59606051e-01 4.53827083e-01 3.26362588e-02
-7.43116081e-01 3.34379107e-01 -8.55529785e-01 -7.63858408e-02
5.66508397e-02 -1.01871215e-01 2.96974629e-01 7.75887489e-01
9.18462351e-02 -6.15374386e-01 -1.28285062e+00 1.11750102e+00
1.76554006e-02 -1.28444946e+00 2.76093066e-01 2.26547107e-01
4.24893230e-01 -5.84535301e-01 -1.22262523e-01 -1.50961876e-01
1.37320176e-01 -9.34795141e-01 7.06645310e-01 7.40597606e-01
5.40569961e-01 -5.26415408e-01 3.11286807e-01 6.12385094e-01
-1.41275311e+00 2.87452161e-01 -6.96102753e-02 1.10212058e-01
1.05764247e-01 2.90972646e-02 -1.02943134e+00 9.21820998e-01
7.53881633e-01 7.95618147e-02 -6.71434641e-01 1.28159976e+00
1.77540734e-01 -6.56163454e-01 -7.32490063e-01 -3.37513477e-01
8.37770775e-02 -2.10356608e-01 7.98317015e-01 7.52225995e-01
5.36194861e-01 3.91547501e-01 7.94507191e-02 9.59551156e-01
9.88051593e-02 1.63769409e-01 -1.01303387e+00 1.74353153e-01
3.85043979e-01 1.05434787e+00 -1.03913438e+00 -3.18338066e-01
-1.71136916e-01 8.51695657e-01 -3.80761564e-01 2.38283858e-01
-1.40407279e-01 -3.06779385e-01 3.97519648e-01 7.06643224e-01
4.48082000e-01 -4.94454265e-01 -7.42781088e-02 -5.53804457e-01
3.16035807e-01 -7.34270573e-01 3.72328520e-01 -1.62806070e+00
-1.06786633e+00 3.02242219e-01 5.69225967e-01 -1.19515002e+00
-3.84619117e-01 -4.80715215e-01 -2.31191069e-01 9.83337402e-01
-6.87306643e-01 -1.31956339e+00 -2.10648552e-02 6.58731997e-01
6.64333284e-01 9.20169950e-02 1.03256810e+00 4.13898239e-03
3.36975455e-01 -1.23097390e-01 -1.18181236e-01 -3.33606780e-01
1.06807739e-01 -9.61868346e-01 2.99416840e-01 4.11588065e-02
3.54397357e-01 6.10966921e-01 4.71384466e-01 -5.43736517e-01
-1.49150324e+00 -1.22927157e-02 1.20036316e+00 -6.93863451e-01
1.98269725e-01 -9.77179483e-02 -4.87719148e-01 2.80661881e-01
1.77660864e-02 -5.62531412e-01 3.25339943e-01 -2.92296298e-02
-1.39892638e-01 -5.59385121e-02 -1.26998043e+00 7.91713834e-01
1.11081123e+00 -5.29013753e-01 -8.53467107e-01 5.81397891e-01
5.35518348e-01 -4.93633986e-01 -7.59779632e-01 2.10116506e-01
8.17697167e-01 -9.15026724e-01 1.69851804e+00 -5.04712164e-01
-1.16952598e-01 -4.27040219e-01 -3.94240528e-01 -8.27225447e-01
-1.69498354e-01 -6.13024414e-01 2.69273072e-01 7.23546386e-01
3.38995337e-01 -2.25405648e-01 8.67419839e-01 1.08930981e+00
5.01120575e-02 -7.22388268e-01 -8.16918194e-01 -6.74411416e-01
-5.14870346e-01 -2.71407902e-01 1.04048073e+00 3.07864338e-01
1.02368640e-02 2.21377596e-01 1.99393585e-01 4.18817662e-02
4.86910909e-01 6.43112063e-01 6.92049563e-01 -1.47527075e+00
-6.75238520e-02 -5.30156732e-01 -4.03190881e-01 -1.57785726e+00
-5.65866470e-01 -8.99736702e-01 -2.64947981e-01 -1.85238838e+00
9.32137594e-02 -6.02624238e-01 1.84275210e-01 4.27005917e-01
6.03296876e-01 2.35892497e-02 5.85402489e-01 2.81369805e-01
-7.20767796e-01 -1.11272842e-01 1.65163195e+00 -2.31915995e-01
-1.22572154e-01 3.00553620e-01 -8.15762460e-01 8.32769752e-01
1.68023765e-01 -2.85824269e-01 -5.16584635e-01 -2.98747063e-01
5.06142378e-01 5.84897578e-01 3.56579810e-01 -5.66265583e-01
3.21593612e-01 -8.50659907e-02 6.13612592e-01 -1.15461004e+00
8.10413539e-01 -1.39275825e+00 7.60773122e-01 2.14923263e-01
2.25974917e-02 4.23203230e-01 2.02318490e-01 3.35095346e-01
8.18197131e-02 -6.57144427e-01 1.02787197e-01 -9.38975334e-01
-6.93862319e-01 1.84531987e-01 -3.43017161e-01 -5.65958202e-01
8.66635919e-01 -6.73930585e-01 -1.13004781e-01 -6.01823151e-01
-1.33914113e+00 2.44685844e-01 7.63125658e-01 6.12098455e-01
8.94077718e-01 -1.56437194e+00 -1.27718493e-01 4.69339639e-01
2.26787906e-02 2.15309545e-01 -6.71940595e-02 2.43535191e-01
-7.23577440e-01 5.77891111e-01 -1.46062583e-01 -4.38443929e-01
-1.53974664e+00 7.38051832e-01 1.43284202e-01 -2.17793629e-01
-5.03527224e-01 5.00119269e-01 1.26793995e-01 -4.00910318e-01
1.50215089e-01 -3.03000927e-01 -2.66591907e-02 4.49090265e-02
1.93745524e-01 2.81477034e-01 2.62205958e-01 -5.62442124e-01
-5.73940635e-01 9.38282609e-01 2.25274235e-01 -7.69742131e-02
1.11819100e+00 -4.25658911e-01 -3.54656875e-02 1.25886276e-01
1.19731820e+00 -7.06704333e-02 -7.08001733e-01 -2.61906028e-01
1.48473913e-02 -8.17687094e-01 -4.26536441e-01 -9.30854917e-01
-5.66577733e-01 5.12969792e-01 4.24266756e-01 5.64495444e-01
1.12294555e+00 7.52245426e-01 6.20622933e-01 6.19844317e-01
8.61048281e-01 -8.00004363e-01 3.35943639e-01 2.60813057e-01
1.24547160e+00 -7.72314787e-01 3.35068971e-01 -4.58731592e-01
-4.09425765e-01 1.15528929e+00 2.64601707e-02 -1.01214368e-02
9.14759099e-01 9.00453329e-02 -1.94826245e-01 -9.43834841e-01
-5.38262367e-01 -4.30799454e-01 7.12172866e-01 5.49164236e-01
1.03665898e-02 9.51603130e-02 -1.39119059e-01 1.73249230e-01
-9.96113271e-02 -2.81681925e-01 -1.37687325e-01 1.29312563e+00
-6.23593509e-01 -1.31466591e+00 -5.21994531e-01 3.19258332e-01
-1.30447954e-01 1.27170518e-01 -7.43336260e-01 8.98442864e-01
1.77597359e-01 7.40448534e-01 2.97749005e-02 -1.02327988e-01
7.61692584e-01 2.23230615e-01 6.64107859e-01 -6.78687453e-01
-4.20907527e-01 4.11691934e-01 7.51695558e-02 -2.83811152e-01
-4.79456872e-01 -5.76894879e-01 -1.26860368e+00 -1.51985839e-01
-3.57843161e-01 -1.27236894e-03 9.90532160e-01 7.62640476e-01
4.97659504e-01 -2.62299955e-01 3.87716293e-01 -8.88069212e-01
-2.10161701e-01 -3.86411726e-01 -9.43826854e-01 6.19248390e-01
1.57982290e-01 -6.49028599e-01 -3.83728519e-02 1.79562777e-01] | [8.459378242492676, -3.505995273590088] |
95db263a-4975-4c97-843e-57ebdb82da6f | spatial-temporal-transformer-for-3d-point | 2110.09783 | null | https://arxiv.org/abs/2110.09783v1 | https://arxiv.org/pdf/2110.09783v1.pdf | Spatial-Temporal Transformer for 3D Point Cloud Sequences | Effective learning of spatial-temporal information within a point cloud sequence is highly important for many down-stream tasks such as 4D semantic segmentation and 3D action recognition. In this paper, we propose a novel framework named Point Spatial-Temporal Transformer (PST2) to learn spatial-temporal representations from dynamic 3D point cloud sequences. Our PST2 consists of two major modules: a Spatio-Temporal Self-Attention (STSA) module and a Resolution Embedding (RE) module. Our STSA module is introduced to capture the spatial-temporal context information across adjacent frames, while the RE module is proposed to aggregate features across neighbors to enhance the resolution of feature maps. We test the effectiveness our PST2 with two different tasks on point cloud sequences, i.e., 4D semantic segmentation and 3D action recognition. Extensive experiments on three benchmarks show that our PST2 outperforms existing methods on all datasets. The effectiveness of our STSA and RE modules have also been justified with ablation experiments. | ['Yulan Guo', 'Qiuhong Ke', 'TingTing Xie', 'Hao liu', 'Yimin Wei'] | 2021-10-19 | null | null | null | null | ['3d-human-action-recognition'] | ['computer-vision'] | [ 2.71814913e-01 -3.13378602e-01 -1.19868621e-01 -3.42575938e-01
-7.27830529e-01 -2.90293157e-01 6.66810751e-01 2.26905923e-02
-2.09444731e-01 1.50788561e-01 2.88817972e-01 -6.18683826e-03
-7.46882409e-02 -6.54568911e-01 -8.78615022e-01 -5.74226618e-01
-1.51237562e-01 1.19715165e-02 8.87511075e-01 -9.62119102e-02
3.38190228e-01 9.23263550e-01 -1.56027186e+00 4.38919842e-01
6.88640296e-01 1.24249279e+00 3.11117411e-01 4.90832210e-01
-3.48097384e-01 6.85750484e-01 -3.58837008e-01 1.38650946e-02
4.98088568e-01 -1.04216188e-01 -7.25570858e-01 2.32988894e-01
1.65533945e-01 -3.45499545e-01 -4.30533528e-01 7.09784031e-01
3.96759272e-01 3.39731663e-01 5.39627254e-01 -1.42252529e+00
-3.18673700e-01 8.23468268e-02 -9.38413858e-01 3.92460912e-01
3.87186140e-01 3.96116078e-01 7.76025057e-01 -1.06548905e+00
5.91604829e-01 1.57911193e+00 8.10249269e-01 2.37883508e-01
-8.96138251e-01 -6.12350106e-01 5.09426296e-01 3.64629745e-01
-1.31018889e+00 -1.19937152e-01 9.76414323e-01 -4.86954302e-01
1.27082241e+00 2.82318834e-02 7.40770340e-01 9.43270087e-01
2.25906521e-01 1.19516313e+00 9.05135214e-01 -1.07586859e-02
1.72219202e-01 -3.46777409e-01 -3.03445626e-02 4.97727692e-01
-3.27070683e-01 9.06698033e-02 -5.77122152e-01 6.02397881e-03
1.43698061e+00 3.58445346e-01 2.52884869e-02 -5.77357471e-01
-1.34400761e+00 6.13675177e-01 6.97425842e-01 3.24838281e-01
-5.10969520e-01 3.29100758e-01 4.69338328e-01 2.30948836e-01
7.16055453e-01 9.79717448e-02 -5.67496121e-01 -2.85003215e-01
-7.30188131e-01 2.74317890e-01 2.22829416e-01 1.02692449e+00
7.17550218e-01 -2.65570194e-01 -3.26368421e-01 8.53293896e-01
9.52511951e-02 4.13287282e-01 4.45802003e-01 -1.20533288e+00
6.48893416e-01 9.40454841e-01 5.14174365e-02 -1.09361148e+00
-2.54240811e-01 -1.34195000e-01 -7.19808519e-01 2.57049531e-01
-2.44485624e-02 3.49398047e-01 -1.03812528e+00 1.46067107e+00
5.72694600e-01 8.55554104e-01 -1.26013666e-01 1.06462431e+00
8.73659968e-01 8.35044682e-01 1.44751027e-01 5.04887551e-02
1.05369115e+00 -1.11425710e+00 -4.11258012e-01 -8.90892893e-02
5.04872441e-01 -5.06352663e-01 9.91002321e-01 -1.54513270e-01
-1.16409540e+00 -8.08062911e-01 -7.57042110e-01 -3.46104592e-01
-3.13993543e-01 -7.65478760e-02 4.45244700e-01 -1.08666345e-01
-8.21130395e-01 6.88073516e-01 -1.15420651e+00 -2.92796224e-01
7.30165839e-01 2.62064755e-01 -5.70654690e-01 -2.96595618e-02
-9.72256422e-01 4.21265155e-01 1.18387565e-01 -1.08576491e-01
-7.95726717e-01 -7.74060845e-01 -9.53687966e-01 9.52196133e-05
2.64945418e-01 -7.15371072e-01 1.05288780e+00 -6.38370037e-01
-1.47947431e+00 9.71293926e-01 -3.95275474e-01 -3.56283277e-01
5.04244924e-01 -3.87317747e-01 -1.11853011e-01 3.90157700e-01
4.28303182e-01 8.25104892e-01 9.19215858e-01 -1.13459074e+00
-9.25809085e-01 -6.10076427e-01 4.69611995e-02 4.33266521e-01
1.22211307e-01 -1.46335542e-01 -7.32167661e-01 -8.33500385e-01
4.78691697e-01 -8.50697339e-01 -2.53626168e-01 1.95332065e-01
-1.97461784e-01 -4.75366414e-01 1.16860688e+00 -4.07189012e-01
9.16180789e-01 -2.31383562e+00 2.87346572e-01 -1.39795160e-02
4.63034473e-02 1.41421556e-01 -1.88091621e-01 2.86246300e-01
-2.19100490e-01 7.37846596e-03 -3.52329254e-01 -3.92804265e-01
-2.56731175e-02 2.63499290e-01 -4.72740889e-01 2.54860699e-01
6.83978677e-01 1.17963398e+00 -9.71379161e-01 -4.67204124e-01
6.66990280e-01 5.91503620e-01 -4.25119847e-01 2.47189954e-01
-2.15600505e-01 5.89026809e-01 -8.09240103e-01 7.88896918e-01
6.47443831e-01 -4.16749656e-01 -5.18205881e-01 -3.36750716e-01
-2.32016832e-01 3.22328150e-01 -1.07883680e+00 2.18902493e+00
-3.21525276e-01 2.84006894e-01 -3.30195397e-01 -8.19609582e-01
9.05382991e-01 7.71264806e-02 9.34114099e-01 -9.39056575e-01
-5.29589169e-02 1.35658076e-02 -6.07169986e-01 -5.59665382e-01
3.67901921e-01 1.78657591e-01 -1.37709141e-01 2.21421152e-01
-3.99774164e-02 -2.07688078e-01 -1.52596295e-01 1.71282649e-01
1.40966916e+00 4.35468823e-01 5.00967167e-02 5.38242161e-02
6.11639619e-01 9.59770009e-02 6.99349880e-01 3.51081282e-01
-4.06439602e-01 7.51997232e-01 5.03354192e-01 -5.67561746e-01
-9.40719783e-01 -1.22375822e+00 8.53257254e-02 6.54063284e-01
7.33696759e-01 -4.22007978e-01 -3.13973159e-01 -9.86510098e-01
1.91427961e-01 3.13819051e-01 -7.37167716e-01 -1.32324442e-01
-6.68812573e-01 -6.49016947e-02 1.13600165e-01 1.04871428e+00
8.38908255e-01 -1.05476880e+00 -9.00287807e-01 1.03272274e-01
-1.59811631e-01 -1.34180796e+00 -6.68968260e-01 -3.09416596e-02
-1.14084208e+00 -9.72467899e-01 -7.93259561e-01 -5.92216790e-01
3.25636148e-01 7.19533026e-01 8.75986636e-01 -3.52958590e-01
-1.18738152e-01 5.66686571e-01 -6.76060081e-01 -2.23316416e-01
1.59174860e-01 1.23900222e-02 -1.84003443e-01 1.03651196e-01
3.84969115e-01 -9.44168925e-01 -7.86448658e-01 6.29596055e-01
-8.43584299e-01 3.12935829e-01 7.48387218e-01 5.68627834e-01
9.67766941e-01 -1.79472379e-02 1.96149692e-01 -4.73159075e-01
2.29591832e-01 -4.43853259e-01 -3.72760475e-01 -2.71380562e-02
2.27543116e-02 -1.00213148e-01 2.01885045e-01 -2.24212646e-01
-7.78135478e-01 2.57997334e-01 -1.87111244e-01 -1.16704571e+00
-2.33543932e-01 1.61440715e-01 -2.55052835e-01 4.31697369e-02
2.08303779e-01 3.44442636e-01 -1.69847101e-01 -6.30872011e-01
1.44443125e-01 3.50149572e-01 5.16685426e-01 -3.34339708e-01
8.47562790e-01 7.29084492e-01 1.01775313e-02 -6.87777877e-01
-8.36941242e-01 -6.78387165e-01 -1.00113225e+00 -1.79458231e-01
1.12217677e+00 -1.02569139e+00 -5.25885165e-01 5.46391666e-01
-1.16097462e+00 -4.36939687e-01 -6.15714610e-01 3.92512202e-01
-8.95971060e-01 3.03108454e-01 -4.42200303e-01 -5.30160308e-01
-2.07386196e-01 -1.20361173e+00 1.84761906e+00 1.80823043e-01
4.23940606e-02 -6.73715234e-01 5.45048602e-02 3.81871879e-01
-3.91165763e-02 6.04827404e-01 6.97438896e-01 -1.59372106e-01
-9.42518294e-01 6.60377890e-02 -5.16587853e-01 2.91771770e-01
2.31665194e-01 -2.35349432e-01 -8.71843338e-01 -8.09577256e-02
-2.52807289e-02 -1.78394884e-01 1.05356479e+00 4.74027634e-01
1.44510710e+00 4.02386375e-02 -4.28743660e-01 8.90693724e-01
1.17168140e+00 1.79873511e-01 6.58979475e-01 2.56985247e-01
9.18396592e-01 3.68266135e-01 1.03162372e+00 4.27219957e-01
3.84419441e-01 9.38957512e-01 5.33616304e-01 -1.10917218e-01
-1.15963414e-01 -4.28779781e-01 4.12237853e-01 6.43405378e-01
-2.37109646e-01 1.36530429e-01 -8.99384141e-01 5.79816639e-01
-2.14654326e+00 -9.58046019e-01 1.41952723e-01 1.87139654e+00
4.19500113e-01 1.86455444e-01 2.16152042e-01 1.85169876e-01
6.19074106e-01 4.68879938e-01 -8.82189929e-01 -5.99705167e-02
7.80026838e-02 1.74736828e-01 3.66591066e-01 2.03653455e-01
-1.34932005e+00 1.04035532e+00 5.26523972e+00 8.67450118e-01
-1.02360570e+00 1.47927076e-01 3.99072826e-01 -1.15071356e-01
-1.45498052e-01 -2.27240443e-01 -5.56880593e-01 4.28404659e-01
3.21012795e-01 -2.98828608e-03 -1.34026006e-01 7.84743488e-01
3.23188603e-01 7.63882473e-02 -1.04833269e+00 1.27420723e+00
-1.27748311e-01 -1.45202303e+00 1.89541861e-01 -4.86880504e-02
6.06031775e-01 1.91341594e-01 3.67758758e-02 2.73791909e-01
6.17248267e-02 -8.84322405e-01 6.15376115e-01 5.44452906e-01
8.64078999e-01 -8.70507598e-01 4.40909207e-01 2.85923094e-01
-1.58367598e+00 -1.77838176e-01 -1.95866629e-01 1.61196962e-01
2.82838404e-01 5.05464554e-01 -3.11467946e-01 6.85529172e-01
9.98748541e-01 1.53431427e+00 -4.59339678e-01 9.52212274e-01
-2.28045851e-01 1.90085873e-01 -3.88784945e-01 3.52343976e-01
6.27701581e-01 -9.99211147e-02 7.76989222e-01 9.53365505e-01
2.49064222e-01 4.81041789e-01 1.10372603e-01 8.52317870e-01
1.55985922e-01 -1.18517436e-01 -5.64031839e-01 5.39481975e-02
4.80366588e-01 9.93299842e-01 -7.13617265e-01 -2.82070398e-01
-4.01539773e-01 1.26616883e+00 2.07161412e-01 2.48226121e-01
-9.84245121e-01 -2.06297860e-01 1.05812991e+00 2.41915450e-01
6.70683622e-01 -5.34155488e-01 -2.03329697e-01 -1.19822407e+00
2.14465439e-01 -5.02839327e-01 3.84182096e-01 -1.11013019e+00
-1.20718157e+00 4.38258857e-01 1.03845082e-01 -1.65971744e+00
1.87361777e-01 -3.57931286e-01 -6.15396440e-01 7.18452156e-01
-1.62164640e+00 -1.18699872e+00 -6.32556081e-01 9.83151674e-01
9.88891840e-01 8.43364447e-02 4.22436744e-01 1.39577016e-01
-5.44118345e-01 2.27532804e-01 -2.98009872e-01 7.82162696e-02
3.36687773e-01 -1.02678180e+00 8.16888571e-01 6.18011415e-01
1.72336653e-01 3.19832295e-01 8.79677013e-02 -6.81206346e-01
-1.24957228e+00 -1.40407825e+00 6.02929473e-01 -5.55755675e-01
3.68822664e-01 -3.63753974e-01 -1.04477000e+00 6.82966292e-01
-3.27224344e-01 2.72178292e-01 2.08727986e-01 -2.62935460e-01
-3.19324136e-01 -2.01416686e-01 -9.38984752e-01 4.55845416e-01
1.59995115e+00 -5.00405252e-01 -7.20610440e-01 2.89487600e-01
1.08921063e+00 -5.79372466e-01 -9.87594783e-01 7.84894943e-01
3.83530527e-01 -1.06344748e+00 1.32142925e+00 -2.97901034e-01
6.69722974e-01 -4.35028195e-01 -2.63854057e-01 -1.09025967e+00
-4.79383022e-01 -3.90659094e-01 -2.84814060e-01 9.09274876e-01
-1.01907561e-02 -3.50019008e-01 8.20434868e-01 2.31986821e-01
-3.32990915e-01 -9.25057888e-01 -1.19693649e+00 -8.73019338e-01
-8.73141736e-02 -6.50448680e-01 8.11070800e-01 9.33715761e-01
-4.44104761e-01 3.01198125e-01 -7.24075437e-02 2.16571763e-01
4.10071343e-01 4.21998978e-01 9.39835429e-01 -1.04882503e+00
2.65736096e-02 -4.59684968e-01 -9.34758842e-01 -1.52740705e+00
6.27761409e-02 -7.44452894e-01 -1.43502891e-01 -1.53244936e+00
8.82351771e-02 -3.42665255e-01 -4.44312066e-01 4.23380882e-01
-1.37948453e-01 6.73046038e-02 2.41813332e-01 3.72289002e-01
-8.25557709e-01 9.74982679e-01 1.41690564e+00 -1.92505419e-01
-4.29148912e-01 6.17572442e-02 -2.09614530e-01 7.02705622e-01
5.65974712e-01 -4.14024651e-01 -6.19740963e-01 -5.50109863e-01
-3.14031303e-01 1.29247963e-01 7.82430887e-01 -1.28355932e+00
1.54634446e-01 -3.21765482e-01 4.83052909e-01 -1.10988545e+00
6.89159572e-01 -9.09449756e-01 -9.52332467e-02 2.57904053e-01
-1.81995541e-01 8.98867249e-02 2.65991449e-01 7.74429619e-01
-3.31970066e-01 4.93889809e-01 7.33303368e-01 -1.53198197e-01
-1.16836131e+00 7.57552028e-01 1.31581604e-01 -9.45005342e-02
1.32391906e+00 -5.25309443e-01 6.51286766e-02 -1.81581274e-01
-7.66105056e-01 4.73078340e-01 6.02241755e-01 8.57980490e-01
1.14154804e+00 -1.58666241e+00 -4.95111436e-01 4.95160908e-01
2.79159904e-01 3.83293450e-01 5.50992191e-01 9.64546800e-01
-4.55214381e-01 4.21929240e-01 -3.42097640e-01 -1.23535502e+00
-1.21382678e+00 5.12454510e-01 2.52444297e-01 -1.94692343e-01
-1.17320704e+00 8.88161182e-01 4.38485712e-01 -2.86158264e-01
2.51939684e-01 -6.95707262e-01 -1.71228260e-01 -2.70241469e-01
4.65221703e-01 3.11645120e-01 -2.27874801e-01 -8.34797978e-01
-4.81673479e-01 1.17633832e+00 3.63629013e-02 4.06766403e-03
1.45207119e+00 -1.62317872e-01 2.41667792e-01 7.02920556e-01
1.32358062e+00 -5.68761587e-01 -1.70930088e+00 -4.13535088e-01
-9.08363760e-02 -9.74459767e-01 4.23511257e-03 -4.23916668e-01
-1.14157987e+00 1.01945782e+00 6.30108178e-01 -3.61083169e-03
1.38658381e+00 1.10174328e-01 1.19090331e+00 -1.12543795e-02
4.92065072e-01 -8.48043263e-01 3.98229092e-01 6.43883348e-01
1.07819951e+00 -1.24358106e+00 -1.44865125e-01 -5.89624345e-01
-7.91439950e-01 7.62633204e-01 6.25474155e-01 -3.90661895e-01
7.26144493e-01 5.95492981e-02 -2.32637927e-01 -3.20435971e-01
-8.49962294e-01 -3.72908771e-01 3.33926052e-01 5.15134215e-01
-2.12687515e-02 -2.46635899e-01 -4.11770493e-02 4.53623891e-01
1.26196250e-01 4.90852930e-02 -1.06209375e-01 1.11890733e+00
-2.47532353e-01 -9.49638963e-01 -3.91358696e-02 1.48908347e-01
8.67699981e-02 3.15890431e-01 -4.28436279e-01 8.98211718e-01
2.77284294e-01 5.84002018e-01 4.16177660e-01 -7.92867422e-01
6.74913943e-01 -8.50671679e-02 3.70152026e-01 -5.49219310e-01
-4.25245970e-01 1.06888160e-01 -3.91043663e-01 -1.33129323e+00
-6.81950510e-01 -7.91253865e-01 -1.39424932e+00 -7.32526705e-02
1.86831608e-01 -8.98531154e-02 5.44480562e-01 7.33561277e-01
7.45359778e-01 6.50929093e-01 8.66606951e-01 -1.08817029e+00
-9.82851759e-02 -6.58313513e-01 -4.66322601e-01 6.68854952e-01
3.89261901e-01 -9.24701512e-01 -1.58533946e-01 1.25368033e-02] | [8.36065673828125, 0.11850260943174362] |
8cb09920-a68e-4b03-ae05-947415cbe9e2 | zero-shot-audio-source-separation-through | 2112.07891 | null | https://arxiv.org/abs/2112.07891v4 | https://arxiv.org/pdf/2112.07891v4.pdf | Zero-shot Audio Source Separation through Query-based Learning from Weakly-labeled Data | Deep learning techniques for separating audio into different sound sources face several challenges. Standard architectures require training separate models for different types of audio sources. Although some universal separators employ a single model to target multiple sources, they have difficulty generalizing to unseen sources. In this paper, we propose a three-component pipeline to train a universal audio source separator from a large, but weakly-labeled dataset: AudioSet. First, we propose a transformer-based sound event detection system for processing weakly-labeled training data. Second, we devise a query-based audio separation model that leverages this data for model training. Third, we design a latent embedding processor to encode queries that specify audio targets for separation, allowing for zero-shot generalization. Our approach uses a single model for source separation of multiple sound types, and relies solely on weakly-labeled data for training. In addition, the proposed audio separator can be used in a zero-shot setting, learning to separate types of audio sources that were never seen in training. To evaluate the separation performance, we test our model on MUSDB18, while training on the disjoint AudioSet. We further verify the zero-shot performance by conducting another experiment on audio source types that are held-out from training. The model achieves comparable Source-to-Distortion Ratio (SDR) performance to current supervised models in both cases. | ['Taylor Berg-Kirkpatrick', 'Shlomo Dubnov', 'Zejun Ma', 'Bilei Zhu', 'Xingjian Du', 'Ke Chen'] | 2021-12-15 | null | null | null | null | ['audio-tagging', 'audio-source-separation'] | ['audio', 'audio'] | [ 2.53462821e-01 -2.55333722e-01 -1.41432017e-01 -2.47255608e-01
-1.67728031e+00 -8.02611411e-01 1.88258499e-01 1.93927661e-01
-2.32568562e-01 1.50452957e-01 4.33466852e-01 -6.19398020e-02
3.35064456e-02 -3.86198312e-01 -5.69656551e-01 -5.74218392e-01
-1.82428971e-01 2.47501120e-01 5.20881295e-01 7.54333735e-02
-2.90836632e-01 2.27059856e-01 -1.65028298e+00 5.72510421e-01
5.12956083e-01 1.00692856e+00 9.53861773e-02 1.05659831e+00
9.66708660e-02 6.92935646e-01 -9.22265053e-01 -1.33828307e-02
3.71294320e-01 -6.25175714e-01 -5.27646899e-01 -8.64407420e-02
5.47646701e-01 -4.01071489e-01 -2.95706213e-01 7.06869423e-01
9.78339851e-01 5.85286729e-02 6.06896818e-01 -1.35917675e+00
-2.76054680e-01 9.97190416e-01 -2.45225757e-01 4.18227971e-01
3.78646344e-01 -4.34983782e-02 1.30479848e+00 -1.10308635e+00
-7.26275817e-02 1.16342688e+00 6.46778047e-01 4.76820201e-01
-1.37455606e+00 -1.04543459e+00 7.20201507e-02 1.57681987e-01
-1.44752967e+00 -1.02013981e+00 9.16673601e-01 -3.68218005e-01
8.36004913e-01 3.97834986e-01 2.35473737e-01 1.19462466e+00
-3.93672436e-01 9.74034905e-01 5.92626870e-01 -5.11271417e-01
3.55150133e-01 1.56386182e-01 3.65786821e-01 7.29628876e-02
-1.71301425e-01 5.49545176e-02 -9.70474124e-01 -3.22394162e-01
3.62988293e-01 -2.35104024e-01 -4.67605919e-01 5.86483479e-02
-9.66004848e-01 5.82136512e-01 9.85281691e-02 1.55616060e-01
6.30493239e-02 1.86971098e-01 4.01856631e-01 3.87291461e-01
5.09288192e-01 2.20040455e-01 -4.44946945e-01 -7.06007108e-02
-1.30481565e+00 1.15276337e-01 9.19200122e-01 9.01467860e-01
5.52177131e-01 4.83218968e-01 -3.79261076e-01 1.16824663e+00
3.37818056e-01 4.98124599e-01 6.19970679e-01 -7.91896045e-01
4.37500417e-01 -2.97422544e-03 -7.91920051e-02 -4.85605359e-01
-1.68491244e-01 -8.37015569e-01 -4.56521064e-01 4.30994928e-02
1.69443667e-01 -2.46628687e-01 -1.10409987e+00 1.91720688e+00
2.43971050e-01 7.07922518e-01 2.35958099e-01 9.32172596e-01
8.84355903e-01 8.31890821e-01 -1.29897937e-01 -1.23504838e-02
1.32485461e+00 -1.02652919e+00 -4.97734338e-01 -2.48195291e-01
3.41558993e-01 -8.15252066e-01 1.08955872e+00 6.62746072e-01
-1.10227633e+00 -6.73462987e-01 -1.06589651e+00 -1.04392201e-01
-6.82557821e-02 3.16607565e-01 7.71457627e-02 7.42290676e-01
-8.42769623e-01 2.11671099e-01 -8.68148565e-01 -1.17336549e-01
1.57506868e-01 1.23143055e-01 -2.58668661e-02 2.17911378e-01
-1.27480221e+00 1.67661190e-01 2.61388928e-01 -3.65983665e-01
-1.65503132e+00 -9.52464581e-01 -9.02359843e-01 6.25018895e-01
4.04830873e-01 -3.49329621e-01 1.60995829e+00 -9.22014654e-01
-1.42027903e+00 5.11765897e-01 -4.02588874e-01 -7.04706311e-01
6.40600473e-02 -3.85945767e-01 -7.39496350e-01 4.07207668e-01
2.28637531e-01 4.98424381e-01 1.17901421e+00 -1.35379779e+00
-8.25155914e-01 2.86528468e-02 -3.17402221e-02 1.02121003e-01
-6.72417939e-01 3.36550772e-01 -5.18909454e-01 -1.11736667e+00
2.50556804e-02 -7.53513694e-01 2.44296253e-01 -2.09368616e-01
-4.52479988e-01 -8.33284259e-02 6.51127458e-01 -4.65518445e-01
1.38496053e+00 -2.65986872e+00 -4.93417010e-02 -2.08178572e-02
1.09086759e-01 1.96989179e-01 -4.42911685e-01 2.93349266e-01
-2.24417865e-01 -6.84335083e-03 -2.44880185e-01 -8.96945596e-01
2.49438167e-01 -1.89857800e-02 -9.94210899e-01 7.53582362e-03
4.00199950e-01 3.19697529e-01 -8.88766289e-01 -4.19097990e-01
-3.12456399e-01 5.26888788e-01 -8.69974077e-01 5.07784784e-01
-3.42719890e-02 8.68533850e-02 1.05661787e-01 7.00231671e-01
4.89741713e-01 9.20122564e-02 -8.85127857e-02 -1.66311905e-01
2.03030139e-01 7.69803047e-01 -1.48591375e+00 1.70626104e+00
-5.57529688e-01 6.24816656e-01 4.01369691e-01 -7.96100259e-01
6.86185598e-01 8.28144014e-01 4.85491365e-01 -1.94906831e-01
2.79500298e-02 3.89694273e-01 -1.75938420e-02 -1.54986292e-01
2.92256266e-01 -3.65261376e-01 -2.17732459e-01 6.93732858e-01
6.41236424e-01 -6.64593279e-02 1.04456082e-01 1.52378753e-01
1.26010847e+00 -2.57906675e-01 -2.84363747e-01 1.84417069e-01
2.25858688e-01 -4.74649608e-01 6.93975389e-01 7.85435915e-01
-1.60897404e-01 1.10822642e+00 1.37105063e-01 2.14770287e-01
-4.25609767e-01 -1.52876043e+00 -8.04647952e-02 1.63333941e+00
1.55697763e-02 -8.22964728e-01 -5.87150216e-01 -5.76263011e-01
-1.37365654e-01 6.74606562e-01 -4.93953526e-02 -4.85095471e-01
-4.19833243e-01 -5.76763034e-01 1.21167076e+00 5.78433216e-01
-5.41207902e-02 -7.20159829e-01 -4.99222577e-01 3.05153877e-01
-3.13772261e-01 -1.15861559e+00 -6.40981615e-01 7.75915325e-01
-4.32801604e-01 -7.86544681e-01 -6.49215400e-01 -8.58034074e-01
3.03215859e-03 5.33201337e-01 1.05694795e+00 -3.16185266e-01
3.17026190e-02 5.78255057e-01 -3.69865626e-01 -7.39207685e-01
-4.00741965e-01 1.89357325e-01 4.54136193e-01 2.60682672e-01
2.60326952e-01 -8.22178185e-01 -2.94905752e-01 4.03501183e-01
-1.02682936e+00 -4.76196289e-01 3.11731130e-01 5.83977342e-01
5.35924792e-01 3.77388120e-01 1.04075551e+00 -1.88689679e-01
6.80578649e-01 -7.21769989e-01 -2.09400967e-01 -3.47387642e-02
-1.69088900e-01 -1.04746170e-01 7.88415015e-01 -8.49194288e-01
-9.67890263e-01 6.49807043e-03 -3.15332592e-01 -7.02953637e-01
-3.49328697e-01 3.09958965e-01 -5.07860601e-01 4.81057465e-01
7.27648616e-01 1.88239053e-01 -5.16614854e-01 -1.00712466e+00
2.62912929e-01 1.11745250e+00 6.90574169e-01 -5.79780459e-01
7.96135008e-01 3.36514831e-01 -6.47710502e-01 -8.53005290e-01
-9.65824485e-01 -7.58599460e-01 -2.37780154e-01 1.75855294e-01
5.96199214e-01 -1.41043675e+00 -1.22108966e-01 4.34968442e-01
-1.01079357e+00 -3.63130182e-01 -5.20648241e-01 7.39183843e-01
-3.83044511e-01 5.70212491e-02 -5.98353744e-01 -9.13488209e-01
-2.41272405e-01 -9.32760239e-01 1.35211718e+00 -1.05365597e-01
-3.23025852e-01 -7.06097841e-01 3.80819291e-01 9.20326486e-02
1.97900429e-01 -3.43471885e-01 5.57821393e-01 -1.24713707e+00
-3.42764646e-01 -1.31377339e-01 2.05138877e-01 6.79936349e-01
3.66267651e-01 -1.92870468e-01 -1.68561137e+00 -3.19808781e-01
3.12802196e-01 -6.31094158e-01 1.17805028e+00 1.86451614e-01
1.05987740e+00 -2.32521549e-01 -2.66746581e-01 6.63890064e-01
8.86545718e-01 1.03230178e-01 1.28044352e-01 -5.15229814e-02
6.06530130e-01 2.65962273e-01 3.27114791e-01 3.27297032e-01
2.03136206e-01 6.78368986e-01 1.08549915e-01 -2.24015668e-01
-5.47663748e-01 -3.78316462e-01 8.94108295e-01 1.01885068e+00
5.95791042e-01 -4.26184922e-01 -7.47372746e-01 9.06916201e-01
-1.40691507e+00 -9.46312785e-01 2.28552982e-01 2.26679254e+00
1.31556165e+00 4.22577083e-01 3.74113947e-01 7.49999404e-01
4.92898732e-01 3.71722765e-02 -3.95441562e-01 -7.43989199e-02
-3.28328088e-02 5.61256349e-01 3.20947170e-02 5.41831851e-01
-1.16994488e+00 6.38759673e-01 6.37281752e+00 1.04312921e+00
-1.42627883e+00 3.56785625e-01 1.58390969e-01 -6.91767514e-01
-3.67038012e-01 2.08219104e-02 -9.00885105e-01 5.31025350e-01
1.34316182e+00 -7.75768608e-02 2.63840616e-01 6.14328682e-01
1.21474124e-01 1.36328325e-01 -1.51145315e+00 1.04544556e+00
2.14792937e-01 -7.97352254e-01 3.49276885e-02 -3.72747213e-01
2.30064198e-01 1.27323702e-01 9.42813531e-02 5.80785513e-01
2.26668432e-01 -7.40191758e-01 1.16473675e+00 1.31775374e-02
7.95141160e-01 -6.67643607e-01 2.51553297e-01 4.14636999e-01
-1.28370965e+00 -2.50018924e-01 -1.20666340e-01 7.93029740e-02
2.15818271e-01 6.09206140e-01 -1.00266194e+00 5.52913129e-01
8.65536630e-01 4.53422725e-01 -4.97123003e-01 1.21879590e+00
-2.21978888e-01 1.39564168e+00 -6.60886347e-01 6.11249328e-01
-2.14522362e-01 4.52030331e-01 8.72017205e-01 1.54832101e+00
4.42873776e-01 -3.00051630e-01 2.88141102e-01 7.50787139e-01
-2.33886808e-01 -1.85847446e-01 -2.94884950e-01 1.08823158e-01
7.77458012e-01 1.07195699e+00 -5.34996450e-01 -3.37872744e-01
-2.08367839e-01 8.27074587e-01 2.32797954e-02 5.39245844e-01
-1.11837924e+00 -7.03554690e-01 8.27785432e-01 1.09990746e-01
4.21982080e-01 -3.53127755e-02 -7.08815157e-02 -1.33629119e+00
-9.83033795e-03 -9.92283642e-01 4.95133221e-01 -6.95905387e-01
-1.24365342e+00 6.97023988e-01 1.56595275e-01 -1.47507596e+00
-2.90671766e-01 -2.66005248e-01 -8.08793426e-01 8.63730013e-01
-1.65756834e+00 -9.38017368e-01 -7.37828319e-04 6.72046244e-01
7.05984831e-01 -2.80066669e-01 9.11052942e-01 6.91831052e-01
-5.95447719e-01 9.58195746e-01 -1.00787491e-01 3.95814538e-01
1.25375462e+00 -1.25348616e+00 2.18847468e-01 9.82064247e-01
8.03725421e-01 5.00974715e-01 5.76646864e-01 -1.69123337e-01
-1.01695633e+00 -1.35302579e+00 6.19464219e-01 -5.05147815e-01
6.19352520e-01 -8.58503222e-01 -1.10890865e+00 6.94592178e-01
1.82178065e-01 5.46071194e-02 1.34910905e+00 1.96719140e-01
-7.31750548e-01 -4.87812728e-01 -6.93414390e-01 2.84866214e-01
7.53199399e-01 -9.82416391e-01 -8.93531442e-01 1.13657452e-01
1.11363447e+00 -1.28571957e-01 -5.20320892e-01 2.17507944e-01
4.01092887e-01 -8.32520604e-01 1.15693808e+00 -6.09737277e-01
1.97436154e-01 -4.77226853e-01 -3.41146350e-01 -1.46663702e+00
-1.99054629e-01 -7.73312032e-01 -4.08050507e-01 1.78729749e+00
4.52769458e-01 -3.68511051e-01 3.60644877e-01 1.37644798e-01
-3.89733493e-01 -3.33935529e-01 -9.59928453e-01 -1.14646196e+00
-5.59527352e-02 -9.46481407e-01 6.12373531e-01 7.41033316e-01
-2.25599147e-02 7.59919941e-01 -3.51625562e-01 5.69399118e-01
5.11564016e-01 1.08155996e-01 7.61841178e-01 -1.09880078e+00
-8.55868220e-01 -2.43513748e-01 -1.17582306e-01 -1.22174263e+00
2.58427948e-01 -1.04266393e+00 3.26889247e-01 -1.22237122e+00
-1.50170803e-01 -4.63539630e-01 -7.73800254e-01 7.08220840e-01
-2.02105016e-01 3.86624068e-01 2.02841237e-01 2.92887181e-01
-5.40413916e-01 6.31592989e-01 4.13761109e-01 -3.22929919e-01
-4.35993999e-01 1.63574621e-01 -7.39035904e-01 8.12106371e-01
7.70142615e-01 -8.55936110e-01 -7.20006287e-01 -7.32862473e-01
-3.32190879e-02 -7.96027035e-02 4.28314716e-01 -1.36604369e+00
2.63072819e-01 2.01736376e-01 1.08849421e-01 -4.75226313e-01
5.44424713e-01 -7.63471663e-01 -1.39151901e-01 -4.10487577e-02
-5.56540191e-01 -6.11819208e-01 4.66838986e-01 5.49996734e-01
-5.69323838e-01 -2.42117763e-01 5.90538085e-01 3.39270085e-01
-1.78341523e-01 1.32350191e-01 -4.63743567e-01 3.30698669e-01
5.81204414e-01 -4.85770125e-03 8.52361321e-02 -4.47105795e-01
-8.14514875e-01 1.24241754e-01 5.77209666e-02 5.42830527e-01
6.29032373e-01 -1.45968580e+00 -8.60751212e-01 4.17765826e-01
1.57119319e-01 8.95572230e-02 -7.77051300e-02 5.52655578e-01
1.81901127e-01 5.64804263e-02 2.36103579e-01 -7.44171739e-01
-1.30117404e+00 5.60369670e-01 2.89598674e-01 6.25725463e-02
-2.98750848e-01 1.05152726e+00 4.70261067e-01 -2.47918323e-01
6.88542902e-01 -6.22078538e-01 9.40022692e-02 3.06068599e-01
7.07494974e-01 3.13288301e-01 1.88594133e-01 -4.54187751e-01
-3.97958696e-01 2.77181953e-01 2.22364381e-01 -5.59263170e-01
1.30540121e+00 -9.75903869e-02 3.19505036e-01 9.53962147e-01
1.30153668e+00 6.10982299e-01 -1.14209330e+00 -4.03994769e-01
-1.42295256e-01 -2.93970823e-01 9.46808308e-02 -6.82357013e-01
-9.54883158e-01 1.23810863e+00 6.74888253e-01 5.19899547e-01
1.45307922e+00 6.82655498e-02 9.62167203e-01 1.25112742e-01
2.74234507e-02 -7.83220470e-01 3.05229753e-01 5.46405315e-01
7.37962067e-01 -9.94067073e-01 -5.34853160e-01 -2.15667665e-01
-4.42156881e-01 9.53808486e-01 3.52355480e-01 5.87882251e-02
8.46955717e-01 7.55709231e-01 2.90012687e-01 9.59556252e-02
-8.87795091e-01 -3.26767743e-01 5.24984241e-01 5.03300905e-01
4.00595695e-01 -2.02756494e-01 5.28863490e-01 1.25727046e+00
-3.04178149e-01 -1.71026304e-01 3.94931287e-01 8.88552368e-01
-5.47967494e-01 -1.15737545e+00 -6.39503121e-01 8.99183899e-02
-6.24360681e-01 -3.79397094e-01 -5.36109567e-01 2.79596969e-02
1.83253855e-01 1.34819472e+00 9.69333574e-02 -5.10712326e-01
4.34580803e-01 3.97293091e-01 2.01166302e-01 -1.09040558e+00
-5.94871342e-01 6.60854220e-01 -9.17720497e-02 -2.17822447e-01
-1.51041657e-01 -3.63453209e-01 -1.28377426e+00 4.50724989e-01
-4.10344183e-01 3.98575991e-01 2.22583339e-01 6.32627189e-01
4.95803386e-01 8.86227190e-01 7.76069283e-01 -9.19814169e-01
-8.07657421e-01 -9.47514176e-01 -6.27030432e-01 2.66157359e-01
9.52299476e-01 -5.14754653e-01 -6.61509573e-01 3.59441757e-01] | [15.254385948181152, 5.393296718597412] |
e582c57b-148b-4592-8399-7744a36f4e8e | robust-cross-modal-knowledge-distillation-for | 2304.07775 | null | https://arxiv.org/abs/2304.07775v2 | https://arxiv.org/pdf/2304.07775v2.pdf | Robust Cross-Modal Knowledge Distillation for Unconstrained Videos | Cross-modal distillation has been widely used to transfer knowledge across different modalities, enriching the representation of the target unimodal one. Recent studies highly relate the temporal synchronization between vision and sound to the semantic consistency for cross-modal distillation. However, such semantic consistency from the synchronization is hard to guarantee in unconstrained videos, due to the irrelevant modality noise and differentiated semantic correlation. To this end, we first propose a \textit{Modality Noise Filter} (MNF) module to erase the irrelevant noise in teacher modality with cross-modal context. After this purification, we then design a \textit{Contrastive Semantic Calibration} (CSC) module to adaptively distill useful knowledge for target modality, by referring to the differentiated sample-wise semantic correlation in a contrastive fashion. Extensive experiments show that our method could bring a performance boost compared with other distillation methods in both visual action recognition and video retrieval task. We also extend to the audio tagging task to prove the generalization of our method. The source code is available at \href{https://github.com/GeWu-Lab/cross-modal-distillation}{https://github.com/GeWu-Lab/cross-modal-distillation}. | ['Di Hu', 'Dejing Dou', 'Haoyi Xiong', 'Andong Deng', 'Xingjian Li', 'Wenke Xia'] | 2023-04-16 | null | null | null | null | ['audio-tagging', 'video-retrieval', 'action-recognition-in-videos'] | ['audio', 'computer-vision', 'computer-vision'] | [ 5.75138740e-02 -1.40928850e-01 1.18688624e-02 -3.01840812e-01
-9.81073797e-01 -6.52354300e-01 7.24814236e-01 -4.62414883e-03
-4.03862655e-01 5.34669936e-01 2.71025687e-01 7.53334016e-02
-2.33377382e-01 -3.95771980e-01 -7.53320754e-01 -8.34054589e-01
7.31497258e-02 -5.02088666e-03 4.22723502e-01 -8.56731739e-03
2.91649010e-02 -5.99403791e-02 -1.76335335e+00 4.50366855e-01
8.26448143e-01 8.44267666e-01 5.04926026e-01 4.77115721e-01
-6.52271882e-02 5.86358190e-01 -3.32993299e-01 -2.90806323e-01
8.92507881e-02 -6.87507927e-01 -8.54279280e-01 -3.64347190e-01
5.15183985e-01 -7.83936009e-02 -4.85565066e-01 1.47667623e+00
8.68663371e-01 3.52479815e-01 4.59350646e-01 -1.45553279e+00
-7.16542900e-01 7.54510820e-01 -5.38696706e-01 2.87389755e-01
5.66973865e-01 1.24656260e-01 9.10873473e-01 -8.83763850e-01
5.33068776e-01 1.22885871e+00 4.43746060e-01 7.29719400e-01
-1.12827015e+00 -1.06270552e+00 3.21609735e-01 6.39029860e-01
-1.49803352e+00 -4.16114151e-01 8.84912252e-01 -3.62347901e-01
1.39687255e-01 3.75297636e-01 6.47723317e-01 1.43548381e+00
-4.91818249e-01 1.09281194e+00 1.28973043e+00 -3.70041877e-01
1.20180566e-02 8.11233744e-02 6.36473000e-02 5.51378250e-01
-9.88922492e-02 8.64942968e-02 -7.97420740e-01 3.56889158e-01
6.05407119e-01 -3.32229622e-02 -7.34618843e-01 -1.79608837e-01
-1.44675779e+00 4.63843435e-01 4.08611357e-01 6.47380412e-01
4.91578355e-02 1.89302325e-01 5.18039107e-01 3.41227055e-01
2.14926124e-01 2.85745353e-01 -3.34030360e-01 -4.10693616e-01
-6.87630594e-01 -1.35094941e-01 4.58710402e-01 9.62421536e-01
6.64665520e-01 -3.89055945e-02 -4.51336443e-01 1.09940934e+00
2.18508452e-01 6.50943756e-01 7.18134046e-01 -1.17419457e+00
2.13702247e-01 1.36115998e-01 -2.22880155e-01 -8.19260538e-01
-2.13924885e-01 -1.44958287e-01 -8.47418666e-01 -7.09235966e-02
3.65527183e-01 -1.41119793e-01 -6.31004572e-01 2.32246923e+00
5.44561207e-01 6.63222790e-01 -3.47669795e-02 1.07917058e+00
1.14306986e+00 4.92262930e-01 2.34827861e-01 -2.38916099e-01
1.66924870e+00 -7.79387057e-01 -7.15141535e-01 1.86459832e-02
5.53454876e-01 -9.06945825e-01 1.29224300e+00 3.61786634e-01
-7.26548076e-01 -7.10598588e-01 -8.61676514e-01 -9.67212319e-02
-2.69356281e-01 1.02487631e-01 3.42770517e-01 1.74637094e-01
-8.54779303e-01 3.88948530e-01 -8.12614918e-01 -5.09573758e-01
2.81520694e-01 -7.96565972e-03 -5.36704600e-01 -1.74927041e-01
-1.49932837e+00 6.24410033e-01 7.37791181e-01 -1.13628181e-02
-8.02993357e-01 -7.09458709e-01 -8.22331011e-01 -1.39098495e-01
5.50046027e-01 -6.79761648e-01 1.26463437e+00 -1.05897653e+00
-1.46800792e+00 7.42950797e-01 -7.55102634e-02 -1.65589098e-02
2.70599008e-01 -2.76403755e-01 -5.27193367e-01 4.95380908e-01
-1.59573108e-02 9.49459910e-01 8.98891687e-01 -1.30364490e+00
-6.39321625e-01 -2.70642012e-01 2.54809499e-01 5.09822190e-01
-5.67476928e-01 -1.97367564e-01 -8.51238251e-01 -9.45573986e-01
1.63456962e-01 -9.73393261e-01 4.34135914e-01 -7.88852870e-02
-2.32551247e-01 -2.70723671e-01 7.66090333e-01 -6.14402592e-01
1.25494885e+00 -2.40672541e+00 2.45778844e-01 -1.41946897e-01
4.37713005e-02 1.14579938e-01 -4.32385594e-01 3.06283712e-01
-2.27271140e-01 -2.35730752e-01 -1.08926244e-01 -3.31544757e-01
1.61818117e-01 1.89180076e-01 -1.71778962e-01 3.10135901e-01
-1.87432244e-02 7.47989893e-01 -1.13048708e+00 -6.79231644e-01
2.81884998e-01 6.52480245e-01 -4.67168212e-01 2.10953653e-01
-1.37889102e-01 8.54763865e-01 -4.08709377e-01 5.50042391e-01
6.33154631e-01 -2.53589332e-01 1.82308838e-01 -6.42820895e-01
1.73866063e-01 3.25258300e-02 -1.42412412e+00 2.18193269e+00
-2.01647401e-01 5.18095434e-01 1.09861076e-01 -1.22607172e+00
5.29198647e-01 4.11281675e-01 6.00907683e-01 -8.68778765e-01
2.49353424e-01 2.25901157e-01 -1.43850848e-01 -7.68347800e-01
3.00974429e-01 -2.28702888e-01 -4.59472910e-02 2.06625134e-01
4.83043969e-01 -2.38636091e-01 1.97402030e-01 2.63231456e-01
7.64452636e-01 3.28625768e-01 1.36211356e-02 -1.45803168e-01
6.53363347e-01 -3.70625108e-01 5.28340459e-01 5.64246893e-01
-3.60517949e-01 6.79030180e-01 2.86950916e-01 2.96624273e-01
-4.80110258e-01 -1.03851700e+00 -2.36221775e-01 1.26152551e+00
5.24572670e-01 -4.15592700e-01 -6.86593771e-01 -6.64048076e-01
-2.16111958e-01 5.48004091e-01 -6.03136957e-01 -3.36220205e-01
-3.04852635e-01 -3.08393598e-01 7.26751685e-01 4.36685711e-01
6.94222033e-01 -8.41697454e-01 -2.25831956e-01 -3.89284819e-01
-6.48292840e-01 -1.17632067e+00 -5.94888270e-01 8.69812593e-02
-5.66178262e-01 -1.11530995e+00 -8.56086791e-01 -7.23045051e-01
4.15117502e-01 4.16569024e-01 7.66512573e-01 -1.18476264e-01
-7.26231113e-02 8.82687807e-01 -6.45525873e-01 -9.75111797e-02
-1.40800506e-01 -2.05969214e-01 1.54303640e-01 2.90931407e-02
4.44432378e-01 -8.28152537e-01 -8.05255592e-01 2.68711776e-01
-1.22348130e+00 1.97558045e-01 3.91218275e-01 9.07392263e-01
6.70225561e-01 -3.33734117e-02 2.73216397e-01 -4.43465382e-01
4.38184768e-01 -4.74332303e-01 -4.31623966e-01 2.38483131e-01
-2.83230484e-01 2.31897831e-01 3.96381766e-01 -9.19228017e-01
-9.93934929e-01 -9.84190330e-02 -1.78417414e-01 -8.45137715e-01
-4.39426482e-01 4.65145290e-01 -3.50830138e-01 2.53708631e-01
4.52066094e-01 4.24286783e-01 -2.71084532e-02 -5.23569942e-01
6.67445779e-01 5.04278243e-01 7.77201056e-01 -8.91256213e-01
7.74931967e-01 4.25026417e-01 -1.17178239e-01 -5.58164060e-01
-9.90628660e-01 -3.94235224e-01 -4.64132875e-01 -3.88872862e-01
9.64565396e-01 -1.09616065e+00 -8.42794240e-01 4.99684066e-01
-8.52149010e-01 -4.16600823e-01 -3.46712023e-01 1.00034273e+00
-5.00141084e-01 6.38826370e-01 -4.90083516e-01 -4.80233938e-01
4.66369502e-02 -1.05777478e+00 9.80086863e-01 4.12673444e-01
-2.07321942e-02 -9.32112098e-01 1.94508612e-01 4.95817542e-01
9.86318290e-02 -1.38769820e-01 5.60057104e-01 -7.32692122e-01
-3.31143826e-01 1.20988406e-01 -2.44466588e-01 5.40030897e-01
6.71206266e-02 -2.58045465e-01 -1.25264406e+00 -2.70868331e-01
-1.89238936e-01 -4.05443519e-01 9.28101599e-01 2.57893860e-01
1.30629182e+00 -8.46080631e-02 -9.90588963e-02 5.02776742e-01
1.14562941e+00 6.17605262e-02 5.61405063e-01 2.15883672e-01
6.21100545e-01 5.31263173e-01 7.57006168e-01 3.87699872e-01
4.14709300e-01 8.09635878e-01 3.04779440e-01 5.95571809e-02
-4.66660738e-01 -2.73523539e-01 5.53709090e-01 1.09916437e+00
-4.92056496e-02 -6.33507073e-02 -5.78472376e-01 5.46878219e-01
-1.93046224e+00 -1.17085052e+00 1.05648778e-01 2.23572946e+00
1.28329730e+00 -2.13860303e-01 1.34853363e-01 1.43973410e-01
7.86729932e-01 7.89542943e-02 -2.71895945e-01 2.10686296e-01
-1.79763064e-01 1.03252351e-01 3.75608802e-02 6.60652041e-01
-1.16781688e+00 8.64385545e-01 4.17816257e+00 1.44632196e+00
-1.25025129e+00 5.17618299e-01 1.39740095e-01 -1.37690797e-01
-3.96127075e-01 9.25830286e-03 -5.86797655e-01 6.76632941e-01
5.64696431e-01 -2.44926326e-02 3.60949993e-01 3.63790452e-01
1.00286275e-01 -2.95269489e-01 -1.08634043e+00 1.23424554e+00
1.97445661e-01 -8.23572695e-01 -7.17203505e-03 -4.33287650e-01
4.92782772e-01 -1.36436954e-01 9.98998433e-02 4.51781273e-01
1.05209395e-01 -5.48014760e-01 8.18384886e-01 7.25341797e-01
8.31201375e-01 -3.61920863e-01 4.32707906e-01 1.71355829e-01
-1.34958911e+00 2.82621533e-02 4.09336425e-02 3.30183327e-01
6.65280074e-02 5.29589713e-01 -3.21457028e-01 8.49556029e-01
9.88776445e-01 8.07232738e-01 -6.45365834e-01 9.34753418e-01
-3.83257836e-01 4.85384047e-01 -3.56360435e-01 2.92694479e-01
-1.71753883e-01 -1.48030922e-01 7.69354701e-01 1.28206825e+00
4.85195071e-01 2.06753612e-01 8.76099151e-03 5.05317688e-01
5.38564548e-02 7.13711306e-02 -4.53717321e-01 -5.44063374e-03
7.21807182e-01 1.08143103e+00 -4.96441543e-01 -4.30002064e-01
-3.98178667e-01 1.12506866e+00 1.48966322e-02 5.54777920e-01
-1.14120054e+00 -4.69725460e-01 5.37975252e-01 -1.89446762e-01
3.35677058e-01 -1.56858832e-01 1.91866904e-01 -1.54170990e+00
1.25083461e-01 -8.28482330e-01 7.16037631e-01 -1.12272525e+00
-1.37402856e+00 4.53664631e-01 3.34044665e-01 -1.66335356e+00
5.15691303e-02 -3.40037793e-01 -1.15267597e-01 6.05460346e-01
-1.57855153e+00 -1.16931665e+00 -5.71431160e-01 1.12763703e+00
4.60939676e-01 7.41231441e-02 6.77593231e-01 7.35197246e-01
-4.95559990e-01 7.54499912e-01 -5.48575558e-02 6.12289570e-02
1.31285059e+00 -9.98103321e-01 -6.61844671e-01 8.58609021e-01
2.48047054e-01 5.33426344e-01 7.16146469e-01 -4.06834424e-01
-1.26207685e+00 -9.11573291e-01 4.41372782e-01 -3.19131374e-01
8.30710590e-01 -1.49793148e-01 -1.06179512e+00 5.08122981e-01
3.63604933e-01 6.04168735e-02 6.30727828e-01 -1.49050606e-02
-6.72930121e-01 -4.04497355e-01 -7.47359216e-01 5.69382250e-01
1.00945723e+00 -8.77064288e-01 -6.53605998e-01 1.67679980e-01
9.58220541e-01 -2.76160449e-01 -9.86933410e-01 6.10470951e-01
6.07660830e-01 -8.72680426e-01 9.09339368e-01 -2.05986023e-01
3.27840507e-01 -6.98942006e-01 -4.33922827e-01 -1.09665930e+00
-1.74336076e-01 -4.69205856e-01 -1.92690149e-01 1.63027322e+00
1.37388438e-01 -5.59618592e-01 1.24365188e-01 1.19208477e-01
-1.70786709e-01 -2.34408200e-01 -1.07769728e+00 -9.88809586e-01
3.94397043e-02 -6.93544090e-01 3.01437736e-01 1.27983725e+00
2.85583198e-01 3.61959219e-01 -4.03609693e-01 3.45510125e-01
4.76924002e-01 1.58016771e-01 5.73536515e-01 -9.59560394e-01
-6.62083685e-01 -6.65422201e-01 -3.22496623e-01 -1.29159307e+00
1.12997279e-01 -9.30326164e-01 8.31604749e-02 -1.27890611e+00
3.83436203e-01 -2.77013630e-01 -5.89703560e-01 6.60640419e-01
-2.83125788e-01 4.60435569e-01 3.82998824e-01 2.86402851e-01
-9.70613778e-01 7.64870226e-01 1.37055731e+00 -2.18186840e-01
-6.16817102e-02 -3.01683009e-01 -6.02209270e-01 4.97134984e-01
6.33101046e-01 -3.79989326e-01 -5.53686142e-01 -4.27713096e-01
7.11019114e-02 3.18979658e-02 5.39665461e-01 -1.02094066e+00
2.05135450e-01 -2.23561645e-01 -3.64524238e-02 -3.15366745e-01
5.49143374e-01 -8.47724199e-01 1.61342099e-01 2.63561338e-01
-3.76139849e-01 -3.20896685e-01 4.01781231e-01 5.58849096e-01
-5.56257367e-01 -2.31162414e-01 8.01437080e-01 -2.81513872e-04
-9.68022525e-01 1.90065786e-01 -1.26767620e-01 3.12964529e-01
8.68477643e-01 6.38028793e-03 -5.32744884e-01 -4.25163239e-01
-1.04067194e+00 1.94895893e-01 2.07542077e-01 5.49991369e-01
3.50748062e-01 -1.62305093e+00 -4.81042534e-01 4.79982458e-02
3.06240261e-01 -2.01329678e-01 6.82779074e-01 1.51911390e+00
7.20507503e-02 2.36313075e-01 -2.43239775e-01 -7.20898688e-01
-1.43397951e+00 6.98239684e-01 2.21333459e-01 1.58080161e-01
-4.77223188e-01 1.03992271e+00 4.44589853e-01 -1.17898226e-01
4.75612700e-01 -2.63286412e-01 -9.73267481e-02 3.11051697e-01
4.01504576e-01 2.18310550e-01 -2.00885117e-01 -6.20360315e-01
-4.62933689e-01 7.15532005e-01 3.80780697e-01 -3.63490254e-01
9.89214718e-01 -4.49385792e-01 -6.52702302e-02 6.55517578e-01
1.20561469e+00 2.30293386e-02 -1.09928322e+00 -4.00068402e-01
-2.72382528e-01 -4.30580288e-01 -2.12841392e-01 -7.03029633e-01
-1.15700281e+00 1.00323594e+00 8.91917288e-01 2.20315814e-01
1.56709170e+00 3.24484587e-01 5.30343890e-01 1.93394125e-01
9.72991139e-02 -1.02139568e+00 4.51007187e-01 4.20826197e-01
9.23548043e-01 -1.28713214e+00 -2.07562938e-01 -3.04330826e-01
-7.79957235e-01 8.76040757e-01 7.28270888e-01 2.25818247e-01
8.61916065e-01 1.13906503e-01 3.76621261e-02 -1.52040556e-01
-5.15366018e-01 -6.05202615e-01 4.00438637e-01 5.56473613e-01
4.10327345e-01 -2.07333509e-02 -4.03794110e-01 6.59171641e-01
1.44045651e-02 4.05525379e-02 7.09710494e-02 8.45983326e-01
-1.77752346e-01 -1.15734220e+00 -2.87779391e-01 -1.80794999e-01
-1.25698477e-01 -1.97495997e-01 -1.20556280e-01 7.32530534e-01
4.19774354e-01 9.22348320e-01 -1.10536255e-01 -5.38335204e-01
2.33975708e-01 2.81720370e-01 7.41510868e-01 -3.25908363e-01
-2.67670304e-01 4.76435483e-01 -5.26946448e-02 -6.01805151e-01
-8.78106534e-01 -6.22633278e-01 -1.32528675e+00 -8.98923278e-02
-2.82736212e-01 2.02869952e-01 3.28740478e-01 8.34435225e-01
3.82444620e-01 5.62375605e-01 4.26969886e-01 -6.36623800e-01
-2.70472854e-01 -1.10575783e+00 -5.65717041e-01 7.40911424e-01
1.45598188e-01 -9.33123529e-01 -4.46299136e-01 3.80770564e-01] | [10.118688583374023, 0.8346592783927917] |
66e51a8b-32fd-492d-9d70-0617ce56ac3e | intkb-a-verifiable-interactive-framework-for | null | null | https://aclanthology.org/2020.coling-main.490 | https://aclanthology.org/2020.coling-main.490.pdf | IntKB: A Verifiable Interactive Framework for Knowledge Base Completion | Knowledge bases (KBs) are essential for many downstream NLP tasks, yet their prime shortcoming is that they are often incomplete. State-of-the-art frameworks for KB completion often lack sufficient accuracy to work fully automated without human supervision. As a remedy, we propose : a novel interactive framework for KB completion from text based on a question answering pipeline. Our framework is tailored to the specific needs of a human-in-the-loop paradigm: (i) We generate facts that are aligned with text snippets and are thus immediately verifiable by humans. (ii) Our system is designed such that it continuously learns during the KB completion task and, therefore, significantly improves its performance upon initial zero- and few-shot relations over time. (iii) We only trigger human interactions when there is enough information for a correct prediction. Therefore, we train our system with negative examples and a fold-option if there is no answer. Our framework yields a favorable performance: it achieves a hit@1 ratio of 29.7{\%} for initially unseen relations, upon which it gradually improves to 46.2{\%}. | ['Dennis Diefenbach', 'Stefan Feuerriegel', 'Guo Kunpeng', 'Bernhard Kratzwald'] | 2020-12-01 | null | null | null | coling-2020-8 | ['knowledge-base-completion', 'knowledge-base-completion'] | ['graphs', 'knowledge-base'] | [ 1.30467638e-01 6.65531993e-01 -2.00890437e-01 -1.75434709e-01
-1.00947320e+00 -6.46606445e-01 4.94199693e-01 4.99118596e-01
-4.13102746e-01 1.07602704e+00 -1.39932679e-02 -6.24765158e-01
-6.88102795e-03 -9.72923875e-01 -8.19746673e-01 -1.27334118e-01
2.53337562e-01 1.03616655e+00 6.44381106e-01 -5.30381083e-01
-4.62433510e-02 5.11687770e-02 -1.37791181e+00 4.42881584e-01
1.18570244e+00 7.80065656e-01 1.16247281e-01 6.83002532e-01
-3.37481886e-01 1.08609569e+00 -5.04110694e-01 -8.34078014e-01
1.11387439e-01 -3.43295068e-01 -1.40217853e+00 -3.80335510e-01
1.07528232e-01 -3.26400787e-01 -3.29468995e-01 9.43622887e-01
2.25183591e-01 5.86787350e-02 6.52296603e-01 -1.07531071e+00
-5.85842490e-01 7.32286453e-01 -1.17684402e-01 1.55415058e-01
7.01569438e-01 1.53885439e-01 1.46808362e+00 -9.89564121e-01
8.29238415e-01 7.52395570e-01 4.81599987e-01 4.77476865e-01
-1.26970649e+00 -2.76693463e-01 1.82409182e-01 3.53984505e-01
-1.34261024e+00 -5.85232019e-01 2.93756098e-01 -4.16871876e-01
1.29849303e+00 4.02112663e-01 4.81430888e-01 8.72014582e-01
-1.97239593e-01 7.54049242e-01 5.02860427e-01 -6.48395836e-01
2.65830547e-01 3.39704901e-01 3.85063261e-01 6.28666341e-01
3.01036030e-01 -4.15713042e-01 -6.73737943e-01 -1.97547913e-01
4.48850125e-01 -3.13585550e-01 -4.24808472e-01 -9.40514356e-02
-1.24150598e+00 6.35299623e-01 1.75522402e-01 2.40986794e-01
-4.31964189e-01 -1.53331801e-01 1.77323237e-01 4.49943900e-01
2.98937738e-01 6.64882362e-01 -5.83593249e-01 -4.85847555e-02
-8.13875377e-01 4.90161449e-01 1.33739924e+00 1.13957179e+00
9.87834811e-01 -6.36853993e-01 -2.63300806e-01 7.28224814e-01
-1.04294069e-01 2.47046962e-01 2.76232600e-01 -5.87634563e-01
7.54767776e-01 7.97471821e-01 6.07370973e-01 -9.03565228e-01
-2.15265110e-01 -3.05560172e-01 -5.26603103e-01 -3.49418610e-01
5.70442498e-01 -1.30468175e-01 -7.89606631e-01 1.45756030e+00
5.54928124e-01 -3.73201892e-02 3.61680686e-01 7.41764069e-01
8.97309780e-01 6.57705843e-01 1.45548761e-01 -4.06584978e-01
1.48953557e+00 -9.69648540e-01 -7.91961908e-01 -5.54097950e-01
9.42888677e-01 -4.63687211e-01 1.23215616e+00 3.81724268e-01
-7.90500879e-01 -2.46677905e-01 -8.74515831e-01 -1.93580955e-01
-3.56754839e-01 7.75578544e-02 3.73190314e-01 3.25442076e-01
-9.24797475e-01 4.80258137e-01 -6.20470047e-01 -3.15202236e-01
4.69769537e-02 2.29829356e-01 -4.60188508e-01 -1.86186120e-01
-1.43356442e+00 1.00700343e+00 7.27851093e-01 1.29613489e-01
-7.30984449e-01 -6.69871926e-01 -8.12611699e-01 1.85020611e-01
1.03638113e+00 -9.58257079e-01 1.54308534e+00 -6.56514347e-01
-1.34688830e+00 6.03516042e-01 -3.30237925e-01 -4.76594478e-01
5.69141924e-01 -6.56001568e-01 -2.98689663e-01 3.32847029e-01
3.17142963e-01 4.21432525e-01 5.59107065e-01 -1.16183698e+00
-7.76251495e-01 -1.71095550e-01 2.52763420e-01 2.29893595e-01
-3.74313146e-01 -8.22671428e-02 -6.85008049e-01 -3.95262122e-01
2.70252656e-02 -7.77988791e-01 -1.63053885e-01 -2.80371636e-01
-7.49224722e-01 -3.43149245e-01 4.53971177e-01 -8.89415860e-01
1.50030231e+00 -1.72160935e+00 -4.27327007e-02 1.56842142e-01
3.16981792e-01 5.71816981e-01 -4.95992973e-02 7.01605380e-01
5.47987446e-02 3.20168674e-01 -1.52078539e-01 -2.63493434e-02
-1.26062736e-01 1.20721750e-01 -5.18539965e-01 -1.27624989e-01
4.56967205e-01 1.00725389e+00 -1.18767715e+00 -5.34223020e-01
-1.69132277e-01 -6.12067766e-02 -5.85042715e-01 5.15958965e-01
-7.60085583e-01 2.25093186e-01 -6.14377677e-01 4.70078826e-01
2.27277815e-01 -6.77274644e-01 3.68346810e-01 -1.38968006e-01
1.08393721e-01 5.84736347e-01 -1.04906964e+00 1.34673798e+00
-1.79573312e-01 3.50244105e-01 -1.36501327e-01 -8.88915539e-01
8.80009770e-01 5.14255404e-01 2.46575288e-02 -2.93450058e-01
-1.74199268e-01 2.26421416e-01 -1.10451572e-01 -7.60842860e-01
6.08180642e-01 -1.17552042e-01 7.98470601e-02 4.65286106e-01
7.58785233e-02 -9.70690474e-02 4.95347917e-01 6.37431443e-01
1.57919586e+00 1.88726857e-02 5.54234803e-01 4.26504537e-02
4.69232440e-01 3.03223372e-01 6.09689891e-01 9.23523486e-01
3.49361300e-02 3.78755897e-01 8.20716918e-01 -4.45611954e-01
-9.35923338e-01 -7.91205943e-01 3.02571774e-01 1.02436554e+00
1.03230104e-01 -6.80794835e-01 -6.71944201e-01 -9.25670564e-01
-5.84946759e-02 9.04577434e-01 -4.94554520e-01 -5.62403984e-02
-4.24779624e-01 -2.74908960e-01 5.25152385e-01 3.97250652e-01
3.10033530e-01 -1.21867239e+00 -3.91291142e-01 3.97575766e-01
-6.91963077e-01 -1.24946213e+00 -1.53089747e-01 6.06971420e-02
-5.99821687e-01 -1.30641735e+00 -3.92526120e-01 -6.68022275e-01
6.53327346e-01 1.64663166e-01 1.22720563e+00 1.86954245e-01
-3.81501988e-02 1.45455837e-01 -7.13734150e-01 -3.04804772e-01
-5.79434395e-01 2.81443655e-01 -1.46126926e-01 -2.15143695e-01
4.43447292e-01 -4.58509147e-01 -5.25414228e-01 2.93886870e-01
-7.57764339e-01 1.59637079e-01 6.07708454e-01 8.48816097e-01
5.31970918e-01 3.57316919e-02 8.76295865e-01 -1.32083881e+00
7.15258002e-01 -4.33526158e-01 -3.47693354e-01 6.94406986e-01
-6.66019320e-01 3.59745733e-02 6.79857492e-01 -1.67601109e-01
-1.19843924e+00 -2.01211171e-03 -2.62360036e-01 -1.55417904e-01
-2.30736718e-01 8.01958203e-01 -1.17854655e-01 3.47615451e-01
1.02369916e+00 1.70341209e-01 -3.34302306e-01 -4.63889331e-01
5.27436018e-01 6.72670901e-01 5.64137936e-01 -5.64101696e-01
8.80894780e-01 1.89872250e-01 -5.14991105e-01 -7.00863898e-01
-1.17925823e+00 -5.79177141e-01 -6.02785885e-01 -5.74537627e-02
5.81997931e-01 -9.08831239e-01 -8.56055081e-01 1.02055542e-01
-1.21749711e+00 -3.90813470e-01 -2.31755883e-01 1.88028872e-01
-3.51010770e-01 3.58870476e-01 -6.40843511e-01 -9.25877571e-01
-6.82396412e-01 -7.03896105e-01 8.02093446e-01 2.85429031e-01
-5.45822084e-01 -6.05682135e-01 7.21128359e-02 6.48682237e-01
9.30171311e-02 -2.22377181e-01 1.10684931e+00 -1.14692235e+00
-4.88613814e-01 -3.06631267e-01 -1.75144985e-01 2.25860491e-01
4.98139150e-02 -1.79009929e-01 -9.89794254e-01 4.18302007e-02
-4.56964821e-01 -5.39500117e-01 6.85814381e-01 -3.14753443e-01
9.05263007e-01 -6.04230046e-01 -5.50146163e-01 -8.80658701e-02
1.02125084e+00 8.40225145e-02 5.61124742e-01 3.54501665e-01
4.82307911e-01 7.81349957e-01 8.02332759e-01 3.21659178e-01
6.37146592e-01 5.44822156e-01 2.92008996e-01 1.90997019e-01
2.54707098e-01 -6.11852407e-01 4.72645573e-02 6.46040916e-01
9.93683785e-02 -4.07817632e-01 -1.19241226e+00 8.22104037e-01
-2.16447902e+00 -8.13172817e-01 -1.19337238e-01 2.16909695e+00
1.37480390e+00 4.09233004e-01 -4.56139594e-02 1.25659630e-01
6.51229262e-01 -1.98252305e-01 -4.42668319e-01 1.15117356e-02
1.97566584e-01 1.63526818e-01 4.66413237e-02 5.58518589e-01
-9.68701899e-01 1.35178638e+00 5.59075499e+00 7.36752033e-01
-6.53504312e-01 -1.95635885e-01 5.60990334e-01 7.94123709e-02
-4.36465263e-01 2.50413805e-01 -1.01330328e+00 2.18888015e-01
7.83881366e-01 -3.24105680e-01 2.15976641e-01 9.38634038e-01
-7.34874886e-03 -3.05337131e-01 -1.25465894e+00 6.70886993e-01
-9.40626785e-02 -1.40644586e+00 1.18829153e-01 -2.79665709e-01
3.97851169e-01 -2.93327957e-01 -4.40936357e-01 6.57807112e-01
5.25340974e-01 -8.65260541e-01 5.99514842e-01 4.40339088e-01
6.67098522e-01 -6.01079106e-01 8.20961058e-01 9.41711485e-01
-8.79065990e-01 1.04724495e-02 -2.87571758e-01 -1.73930258e-01
4.16557074e-01 9.35579240e-01 -1.38893342e+00 7.64623046e-01
5.33605754e-01 2.43914962e-01 -4.26682204e-01 8.35745811e-01
-7.60400593e-01 6.89832866e-01 -3.62113863e-01 -2.67086387e-01
-9.01377127e-02 6.94794953e-02 3.99397254e-01 1.28801787e+00
1.05195962e-01 4.91789997e-01 2.17717439e-01 6.28380954e-01
-2.56834567e-01 3.02192926e-01 -6.18663907e-01 -1.19733259e-01
7.60459065e-01 1.19632208e+00 -5.88146091e-01 -6.02193236e-01
-2.74739116e-01 9.23273981e-01 8.68251145e-01 3.99416268e-01
-5.28572202e-01 -6.49340570e-01 3.16922843e-01 8.12759716e-03
5.16914248e-01 1.85168415e-01 -1.74036816e-01 -1.34981716e+00
3.41273457e-01 -8.17660451e-01 6.29448414e-01 -7.96239495e-01
-1.31669712e+00 5.84081113e-01 -2.11305991e-01 -6.63709819e-01
-5.27433395e-01 -4.04884338e-01 -3.32632631e-01 6.81626678e-01
-1.45119417e+00 -9.67947662e-01 -3.19575131e-01 4.71020013e-01
3.01320463e-01 2.75303751e-01 1.00998366e+00 2.03493014e-01
-5.78524470e-01 4.79940981e-01 -3.96601975e-01 3.39061081e-01
7.43987143e-01 -1.23968446e+00 5.89282334e-01 8.55418384e-01
2.87587136e-01 8.68331254e-01 8.30333233e-01 -9.55829144e-01
-1.30792260e+00 -1.09492624e+00 1.39433300e+00 -6.23596907e-01
7.86478579e-01 -3.76391858e-01 -1.23392880e+00 8.45115900e-01
-2.07967266e-01 -9.36955363e-02 7.61369169e-01 5.19541860e-01
-4.34625059e-01 -1.03505291e-01 -8.83364975e-01 7.75082707e-01
8.59805703e-01 -5.94062567e-01 -9.23778176e-01 5.61074138e-01
1.00370312e+00 -4.26721841e-01 -6.64887488e-01 4.52958226e-01
3.39418977e-01 -7.51903117e-01 6.47788703e-01 -7.95047522e-01
4.19097811e-01 -4.77760464e-01 1.37732714e-01 -1.00459599e+00
-1.09525107e-01 -9.07718360e-01 -4.77391213e-01 1.13422096e+00
9.84672070e-01 -5.31391621e-01 8.41722310e-01 9.53012645e-01
6.17194502e-03 -8.30174208e-01 -7.60908484e-01 -6.40345037e-01
-3.69620591e-01 -4.50344592e-01 3.38298738e-01 9.83106315e-01
5.89043438e-01 8.62421989e-01 -3.69917512e-01 2.17939451e-01
1.75665870e-01 1.36702910e-01 1.04740369e+00 -1.16238987e+00
-4.36776698e-01 6.53197896e-03 1.14342809e-01 -1.30471921e+00
-1.01628825e-01 -6.47784710e-01 3.45110953e-01 -1.72029173e+00
3.57869059e-01 -4.43070620e-01 -5.47102140e-03 8.66302788e-01
-7.01068640e-01 -7.50869364e-02 -7.13178739e-02 2.71591127e-01
-8.31192076e-01 4.47865933e-01 8.81992579e-01 1.78857431e-01
-4.40974981e-01 1.30813057e-02 -8.12738836e-01 5.82181633e-01
6.97344124e-01 -3.89013797e-01 -4.68330830e-01 -2.26902992e-01
7.01204479e-01 3.10963660e-01 2.11213201e-01 -5.55350482e-01
5.79249084e-01 -1.83990404e-01 5.04647531e-02 -5.08666992e-01
1.43044457e-01 -5.89622855e-01 7.41069242e-02 2.19342113e-01
-3.61898780e-01 -2.68925428e-01 6.96975887e-02 7.37107992e-01
-1.39871359e-01 -4.49845046e-01 2.62977809e-01 -2.09041417e-01
-5.78731716e-01 1.81095794e-01 -2.78023332e-01 3.43714923e-01
9.73287642e-01 1.09470941e-01 -4.86195862e-01 -5.74299932e-01
-7.45587409e-01 4.65249956e-01 1.72400489e-01 8.63331705e-02
4.30103987e-01 -8.44260216e-01 -6.04769170e-01 -9.30129066e-02
4.16927189e-01 5.09894431e-01 5.06855734e-02 5.68015873e-01
-4.66598421e-01 5.57125688e-01 3.67416650e-01 -2.63611227e-01
-1.11844218e+00 5.99920869e-01 7.81360269e-02 -6.36416972e-01
-7.13567734e-01 9.28073049e-01 1.47027269e-01 -3.86870563e-01
2.02799171e-01 -2.38517463e-01 -2.37078145e-01 1.17363833e-01
7.58940816e-01 1.32861927e-01 3.48958760e-01 -2.02151313e-02
-2.39597395e-01 -1.03228532e-01 -5.50970614e-01 -2.64369756e-01
1.37505269e+00 1.19491823e-01 -9.00911763e-02 3.40105236e-01
5.23307920e-01 1.57668442e-02 -9.93499577e-01 -4.92347747e-01
3.51301998e-01 -3.43123466e-01 -3.87975425e-01 -1.08289695e+00
-3.48334432e-01 6.26425683e-01 -4.10435379e-01 4.20549512e-01
8.32495034e-01 2.30339736e-01 8.97368491e-01 9.59548116e-01
3.02010208e-01 -1.04169977e+00 2.22915798e-01 7.33737886e-01
7.23152936e-01 -1.19741905e+00 -8.77504125e-02 -7.26150990e-01
-7.02100933e-01 9.04807866e-01 7.22105682e-01 3.77430052e-01
2.30744377e-01 1.11706719e-01 -2.90497113e-02 -2.60977834e-01
-1.18088686e+00 -2.63588756e-01 2.37937018e-01 5.49285352e-01
3.16763490e-01 -1.84651405e-01 -1.96292013e-01 8.73334527e-01
-2.31638417e-01 1.40354887e-01 3.70424956e-01 9.86733973e-01
-7.57081628e-01 -9.29255664e-01 -4.93375398e-02 5.99190950e-01
-3.76532078e-01 -3.78084689e-01 -5.66807449e-01 6.75563574e-01
-3.08319718e-01 1.07236052e+00 -3.50506127e-01 -3.23025495e-01
5.59594750e-01 3.84979665e-01 1.48969904e-01 -9.68481898e-01
-5.14436781e-01 -1.95436805e-01 7.10684061e-01 -3.58261138e-01
1.35809720e-01 -3.87222260e-01 -1.28145909e+00 -1.84903130e-01
-5.50023139e-01 4.99370813e-01 2.03920275e-01 1.21207440e+00
4.88956720e-01 2.48472497e-01 2.01493710e-01 -2.30013356e-01
-5.90461254e-01 -1.02829790e+00 -2.28713334e-01 3.19780380e-01
2.10892677e-01 -4.05952424e-01 -1.53116718e-01 3.88110340e-01] | [9.718048095703125, 8.410362243652344] |
25b7dd26-61cd-4829-8f13-da927f571c39 | a-comprehensive-review-of-3d-convolutional | 2306.09418 | null | https://arxiv.org/abs/2306.09418v1 | https://arxiv.org/pdf/2306.09418v1.pdf | A comprehensive review of 3D convolutional neural network-based classification techniques of diseased and defective crops using non-UAV-based hyperspectral images | Hyperspectral imaging (HSI) is a non-destructive and contactless technology that provides valuable information about the structure and composition of an object. It can capture detailed information about the chemical and physical properties of agricultural crops. Due to its wide spectral range, compared with multispectral- or RGB-based imaging methods, HSI can be a more effective tool for monitoring crop health and productivity. With the advent of this imaging tool in agrotechnology, researchers can more accurately address issues related to the detection of diseased and defective crops in the agriculture industry. This allows to implement the most suitable and accurate farming solutions, such as irrigation and fertilization before crops enter a damaged and difficult-to-recover phase of growth in the field. While HSI provides valuable insights into the object under investigation, the limited number of HSI datasets for crop evaluation presently poses a bottleneck. Dealing with the curse of dimensionality presents another challenge due to the abundance of spectral and spatial information in each hyperspectral cube. State-of-the-art methods based on 1D- and 2D-CNNs struggle to efficiently extract spectral and spatial information. On the other hand, 3D-CNN-based models have shown significant promise in achieving better classification and detection results by leveraging spectral and spatial features simultaneously. Despite the apparent benefits of 3D-CNN-based models, their usage for classification purposes in this area of research has remained limited. This paper seeks to address this gap by reviewing 3D-CNN-based architectures and the typical deep learning pipeline, including preprocessing and visualization of results, for the classification of hyperspectral images of diseased and defective crops. Furthermore, we discuss open research areas and challenges when utilizing 3D-CNNs with HSI data. | ['Christopher J. Henry', 'Christopher P. Bidinosti', 'Michael A. Beck', 'Nooshin Noshiri'] | 2023-06-15 | null | null | null | null | ['classification-of-hyperspectral-images'] | ['computer-vision'] | [ 5.35503328e-01 -5.57013571e-01 -1.13356441e-01 3.45107280e-02
-2.99997211e-01 -8.62510443e-01 9.48462728e-03 5.29093802e-01
-1.67970080e-02 3.98008794e-01 -4.16464984e-01 -6.45566046e-01
-3.48419368e-01 -1.14966893e+00 -2.52767295e-01 -1.13014317e+00
-2.38518000e-01 -1.50761204e-02 -1.70186296e-01 -5.35624027e-01
-1.55970141e-01 1.08349168e+00 -1.89880800e+00 1.74905524e-01
8.59844983e-01 1.34753978e+00 5.64299822e-01 4.43129480e-01
-5.72580583e-02 -7.84042105e-02 -3.23928535e-01 4.07490194e-01
3.44705850e-01 -3.97726029e-01 -2.31349051e-01 2.46988818e-01
3.22390556e-01 -5.95025122e-01 -1.03483304e-01 1.31438982e+00
5.29790580e-01 -2.04616845e-01 5.74166596e-01 -9.73975658e-01
-7.41237044e-01 6.80979043e-02 -8.96843255e-01 -8.15791264e-02
-5.85494824e-02 1.77725777e-01 6.45665348e-01 -6.42045021e-01
1.77052677e-01 7.38007963e-01 5.89057624e-01 6.78651258e-02
-1.17264068e+00 -5.68988681e-01 -1.79135635e-01 1.74527124e-01
-1.07949436e+00 -8.16419125e-02 5.95301807e-01 -4.36563760e-01
8.37065578e-01 2.83042669e-01 1.15022075e+00 5.69423318e-01
1.27327591e-01 6.20175838e-01 1.08392251e+00 -3.42164010e-01
4.35566455e-02 -3.94082546e-01 2.07944866e-02 4.94340330e-01
6.01895392e-01 4.60437506e-01 -1.73778877e-01 3.30207758e-02
8.67639065e-01 2.75465101e-01 -4.42766130e-01 -5.26469588e-01
-9.15836513e-01 6.55135274e-01 9.40768301e-01 5.09248614e-01
-7.94743180e-01 -3.06580752e-01 2.22383931e-01 9.74462628e-02
6.77082181e-01 6.53683901e-01 -6.88163400e-01 3.15196723e-01
-1.01987708e+00 4.45609316e-02 4.65113729e-01 6.27783298e-01
9.28582251e-01 2.90062726e-01 1.58263475e-01 7.90355563e-01
5.19405790e-02 8.56878996e-01 -5.84831787e-03 -7.01237082e-01
-1.10654183e-01 9.20693636e-01 6.28639534e-02 -9.70217764e-01
-5.07834792e-01 -5.36299527e-01 -1.14160144e+00 6.58690274e-01
2.26407826e-01 -9.05418843e-02 -1.20475733e+00 1.16793180e+00
1.21771529e-01 -2.55817711e-01 6.69588745e-02 1.10856974e+00
7.86509812e-01 6.79225385e-01 7.76037276e-02 -2.10823044e-01
1.37162483e+00 -2.85834551e-01 -6.36945486e-01 -3.71123284e-01
5.87685227e-01 -7.36290514e-01 7.85809219e-01 2.87478715e-01
-6.44307554e-01 -1.24942191e-01 -1.28238928e+00 1.66555807e-01
-9.88333106e-01 4.61010844e-01 9.52560663e-01 5.49654365e-01
-8.57289732e-01 6.15868568e-01 -8.65432680e-01 -7.01685727e-01
7.83299029e-01 2.78731585e-01 -6.09640539e-01 -4.92085338e-01
-7.80905187e-01 9.10573244e-01 5.39270580e-01 7.14119315e-01
-6.72293842e-01 -8.78582835e-01 -7.91287005e-01 2.89509326e-01
2.88132429e-01 6.64393082e-02 8.60334933e-01 -6.64996862e-01
-1.23708892e+00 9.09554482e-01 7.22370371e-02 -3.24531123e-02
-1.22512244e-01 -1.10939600e-01 -1.41151056e-01 2.10530221e-01
9.20519698e-03 4.42789614e-01 4.48964804e-01 -1.29022849e+00
-6.75594509e-01 -8.92440677e-01 -8.78547356e-02 1.30015567e-01
-5.01767635e-01 -1.07341588e-01 1.24244601e-01 -2.23457232e-01
7.67607689e-01 -7.72852361e-01 -1.21092856e-01 4.25895572e-01
-2.81724393e-01 4.77926224e-01 1.32999158e+00 -7.52035737e-01
5.27480841e-01 -2.20637393e+00 -1.04672089e-01 -5.59477881e-03
2.21738800e-01 9.74962831e-01 -3.20135772e-01 4.37380791e-01
-2.30012134e-01 1.80351943e-01 -4.71882612e-01 3.95144075e-01
-4.01529938e-01 5.77229932e-02 1.69171765e-01 6.32680714e-01
5.42547047e-01 9.26024258e-01 -8.68382454e-01 3.46979015e-02
5.37340522e-01 6.50110722e-01 1.54980630e-01 1.55650705e-01
-1.38687953e-01 2.69338399e-01 -4.13285226e-01 1.36460590e+00
1.15027833e+00 -9.90098044e-02 2.13043660e-01 -5.56840658e-01
-6.74703300e-01 -2.84760058e-01 -6.99544549e-01 1.31347835e+00
-2.63825983e-01 7.33911693e-01 6.54458225e-01 -1.44561505e+00
9.64268446e-01 4.39056098e-01 5.83359778e-01 -7.09741592e-01
7.12877885e-02 4.55156624e-01 -5.62681556e-02 -6.26611352e-01
2.06589624e-01 -3.56821828e-02 5.18743217e-01 1.13176294e-01
1.86118893e-02 -3.32518518e-01 1.47118300e-01 -4.80693609e-01
7.31552601e-01 2.51580626e-01 2.67168075e-01 -3.75798196e-01
8.83207470e-02 5.26480794e-01 3.89442503e-01 2.58803368e-01
-3.15984309e-01 4.34811980e-01 3.14220876e-01 -4.57730949e-01
-9.54070628e-01 -6.49307311e-01 -2.21104652e-01 9.55874324e-01
8.28600824e-02 5.20836592e-01 -4.09076631e-01 -1.12738036e-01
3.08906347e-01 2.83785313e-01 -7.10253358e-01 -5.85714802e-02
-1.34739131e-01 -1.25874853e+00 3.54569316e-01 4.77350652e-01
8.90283048e-01 -9.72674310e-01 -9.35447156e-01 3.19618851e-01
-1.16095886e-01 -9.66512144e-01 5.02483428e-01 7.97665298e-01
-1.06470680e+00 -1.37336373e+00 -8.56758833e-01 -6.31485105e-01
4.77030784e-01 1.01774883e+00 7.74187267e-01 -1.21881381e-01
-8.30661118e-01 -8.64989832e-02 -6.17423832e-01 -9.12680626e-01
-1.52731419e-01 1.42795414e-01 -5.61557114e-01 -3.81439030e-01
6.38612747e-01 -5.25594950e-01 -6.85017288e-01 4.84999642e-03
-1.05042934e+00 -1.68267831e-01 7.55897164e-01 8.07863355e-01
5.94406307e-01 3.32948357e-01 1.64622709e-01 -4.83807951e-01
2.58595437e-01 -4.62478518e-01 -8.62641096e-01 4.49577034e-01
-3.96777064e-01 -5.03326297e-01 2.89810359e-01 -2.31707677e-01
-7.75634527e-01 2.91165859e-01 2.54013777e-01 -1.01458497e-01
-7.12687194e-01 1.19486117e+00 -2.54333764e-01 -3.74203354e-01
8.02383125e-01 3.55149247e-02 2.93253183e-01 -3.47733408e-01
-7.30000511e-02 6.27219975e-01 5.40788174e-01 1.42668724e-01
8.46114635e-01 4.98794824e-01 4.19447571e-01 -1.43719256e+00
-9.29087400e-01 -7.75076807e-01 -8.76200557e-01 -3.46355915e-01
8.50504935e-01 -8.35596859e-01 -7.82721400e-01 8.97204757e-01
-9.17432427e-01 -3.53616744e-01 1.32880788e-02 5.52492619e-01
-1.82931051e-02 2.07314819e-01 -3.44288975e-01 -9.32566404e-01
-4.04427528e-01 -9.89732146e-01 1.04896748e+00 3.54751796e-01
3.89463216e-01 -8.60057354e-01 -4.15834188e-01 1.24492124e-01
7.17390835e-01 8.21291804e-01 1.18930006e+00 6.54470995e-02
-4.10310954e-01 -6.12787724e-01 -6.88153923e-01 3.73717010e-01
5.86961925e-01 2.65713185e-01 -1.28808641e+00 -2.23050594e-01
-2.23876685e-01 -3.72375578e-01 9.43181932e-01 9.18763280e-01
9.65212941e-01 3.50066751e-01 -2.33493611e-01 9.29520845e-01
1.82237303e+00 4.09342170e-01 4.63087261e-01 3.93307507e-01
5.53572237e-01 9.06765521e-01 8.36160421e-01 3.71913135e-01
-2.41170153e-01 1.18665315e-01 1.31663108e+00 -8.33812594e-01
1.24533512e-01 2.26521164e-01 -2.16146499e-01 1.96639314e-01
-2.30511606e-01 -3.60347897e-01 -1.08056688e+00 5.84288359e-01
-1.51627362e+00 -9.43258703e-01 -3.21344703e-01 1.95734036e+00
4.65731502e-01 -4.27739024e-01 -2.06370726e-01 5.59687614e-01
6.58393145e-01 4.65431549e-02 -7.75888443e-01 2.54230034e-02
-7.37635136e-01 3.20871532e-01 9.55907762e-01 -1.12372734e-01
-1.38476956e+00 1.01564538e+00 6.16807938e+00 1.78968564e-01
-1.67714274e+00 -3.58944178e-01 4.05160069e-01 3.73779744e-01
2.16154411e-01 -1.98953912e-01 -4.19026256e-01 -3.58645409e-01
4.65985954e-01 2.93964028e-01 4.95519131e-01 5.56158423e-01
3.80359143e-01 -6.02276862e-01 -7.47520745e-01 6.76386058e-01
-2.66823798e-01 -1.17931533e+00 -2.28897601e-01 2.24112108e-01
5.97503901e-01 2.19358474e-01 1.42464280e-01 -2.54258186e-01
-1.86761911e-03 -9.61381435e-01 2.70835519e-01 1.33845627e-01
1.10727465e+00 -5.96925080e-01 1.00305760e+00 2.55136847e-01
-1.27449954e+00 -4.09943074e-01 -5.66788137e-01 -1.33122116e-01
-3.65696728e-01 9.06034768e-01 -6.19638383e-01 7.26014197e-01
8.18292797e-01 7.98464477e-01 -4.50983822e-01 1.19826293e+00
-1.43847254e-04 5.88958681e-01 -2.77371973e-01 1.62858441e-01
5.00683308e-01 -5.04429579e-01 5.18698208e-02 9.99499440e-01
8.57237518e-01 3.60858530e-01 6.85788468e-02 1.02319336e+00
3.81516963e-01 -1.04357250e-01 -8.11429501e-01 -6.74578130e-01
3.94641191e-01 1.48009288e+00 -9.99197185e-01 1.85238600e-01
-3.68581384e-01 5.84445834e-01 -2.08303064e-01 4.83854949e-01
-1.51677743e-01 -4.32019800e-01 7.27073789e-01 -4.10534553e-02
3.35826159e-01 -5.97957551e-01 -2.76527166e-01 -6.29725814e-01
-1.94809735e-01 -9.35616553e-01 -6.14371523e-02 -9.53762412e-01
-1.02940333e+00 1.32641241e-01 -9.41420645e-02 -1.00061536e+00
2.15672344e-01 -1.30294704e+00 -3.26646924e-01 1.22018552e+00
-1.99157703e+00 -1.42799032e+00 -1.05918074e+00 1.54755503e-01
2.43493870e-01 -8.40792581e-02 1.51352561e+00 -1.60844217e-03
-6.32463634e-01 -1.84201986e-01 5.07452667e-01 5.17197698e-02
3.92333359e-01 -1.09063530e+00 -3.55235953e-03 6.93276227e-01
-4.16247129e-01 9.93816853e-02 4.76250410e-01 -5.38655043e-01
-1.69826555e+00 -1.15256178e+00 4.43599582e-01 4.29852813e-01
4.47277129e-01 6.12147190e-02 -9.31856096e-01 2.67074853e-01
-1.48131102e-01 1.59923315e-01 8.44760537e-01 -1.60853229e-02
-1.19187795e-01 -2.47699335e-01 -1.07757723e+00 2.24878132e-01
6.92465067e-01 -5.66531062e-01 3.16862345e-01 4.67784047e-01
2.54705310e-01 -3.23214591e-01 -8.35924387e-01 6.33646965e-01
6.86761916e-01 -1.01322973e+00 9.02752578e-01 -1.54027611e-01
5.48632622e-01 -1.98356673e-01 -2.71475434e-01 -1.51232374e+00
-6.63446248e-01 4.68297154e-02 3.94084007e-01 7.57677853e-01
2.66571105e-01 -4.62637246e-01 7.72378385e-01 -1.79104190e-02
-2.22991794e-01 -4.38629419e-01 -3.48445654e-01 -7.15331495e-01
6.68768287e-02 -1.29686564e-01 8.18187058e-01 9.25161719e-01
-3.74784887e-01 -2.84278631e-01 2.31160387e-01 7.39203632e-01
6.03272855e-01 2.80429751e-01 1.68754905e-01 -1.85058522e+00
4.55993652e-01 -4.41149026e-01 -4.19894546e-01 -3.15571576e-01
1.12708956e-01 -5.86115897e-01 1.41349018e-01 -1.80218220e+00
5.28109297e-02 -2.91968584e-01 -9.44576785e-02 7.94401586e-01
-1.24561079e-01 3.83871138e-01 -9.67147411e-04 8.31931606e-02
4.11062747e-01 2.40621388e-01 1.34618676e+00 -5.06645143e-01
-4.04656976e-01 2.76405644e-03 -5.52731395e-01 3.66596133e-01
1.19942701e+00 -7.41742253e-02 -2.82920778e-01 -6.69819713e-01
1.56355888e-01 1.42858371e-01 7.32097328e-01 -8.70784879e-01
-6.87136799e-02 -3.42712879e-01 7.25561798e-01 -9.90874827e-01
3.63494396e-01 -1.01918232e+00 1.88815091e-02 4.98127580e-01
4.24018800e-02 -2.96508253e-01 6.13823295e-01 1.51114449e-01
-3.45854998e-01 -2.12713003e-01 6.30542994e-01 -3.04509789e-01
-8.35454106e-01 4.73759383e-01 -6.15300894e-01 -6.44461930e-01
8.82278144e-01 -4.47380185e-01 -5.19240618e-01 -1.83955878e-01
-5.44775546e-01 1.18213609e-01 3.94511163e-01 1.56465575e-01
6.30348027e-01 -1.05787969e+00 -7.12221026e-01 4.59120452e-01
3.81064951e-01 1.12301141e-01 2.43125901e-01 6.21317089e-01
-1.03284991e+00 5.29819489e-01 -7.94254899e-01 -7.39442468e-01
-1.30414331e+00 3.45151156e-01 5.04788816e-01 5.61634451e-02
-9.12284032e-02 6.84843242e-01 2.91643124e-02 -3.69884223e-01
3.68096195e-02 -3.12719285e-01 -4.42901969e-01 3.62073541e-01
2.86645234e-01 3.94385815e-01 4.01766300e-01 -5.16493917e-01
-1.63425863e-01 3.84045750e-01 4.71775532e-01 2.38029763e-01
1.81589830e+00 1.36097416e-01 -4.05245215e-01 3.78547192e-01
9.95858610e-01 -8.59906077e-01 -1.31268740e+00 -1.39080450e-01
-1.15231767e-01 -3.79732609e-01 7.35454500e-01 -9.02451515e-01
-1.27025092e+00 1.34932530e+00 1.07689524e+00 7.05204844e-01
1.52210689e+00 -4.70005214e-01 2.77282000e-01 3.57213944e-01
1.05601631e-01 -9.19057846e-01 -2.45319352e-01 5.43496132e-01
8.12551141e-01 -1.60584879e+00 9.57606956e-02 -6.30483210e-01
-7.66740218e-02 1.44422710e+00 4.95688587e-01 4.05002594e-01
7.45592535e-01 4.11066949e-01 3.69665861e-01 -4.13068473e-01
-8.25887173e-02 -6.05943024e-01 8.34502056e-02 1.22128808e+00
8.13685298e-01 2.99604118e-01 2.63920754e-01 -2.65329123e-01
1.79410920e-01 6.26273677e-02 1.72995284e-01 1.28542626e+00
-6.77566111e-01 -8.11906099e-01 -5.55476665e-01 5.88820577e-01
-1.97963089e-01 -1.39206707e-01 -4.81561691e-01 7.40527093e-01
1.60312504e-01 9.75901008e-01 -6.31132945e-02 -1.50564715e-01
3.73224437e-01 -8.32079723e-02 5.12535572e-01 -6.78301573e-01
-3.47292393e-01 2.98366606e-01 -2.98381269e-01 -4.09908533e-01
-7.47433186e-01 -7.25923955e-01 -6.66520536e-01 -3.30739826e-01
-7.97584236e-01 -3.59048963e-01 1.21656835e+00 7.70333707e-01
1.12069391e-01 5.71264803e-01 6.89151406e-01 -1.24691069e+00
-4.55902100e-01 -9.18681383e-01 -1.23575795e+00 -1.90894261e-01
5.86433053e-01 -5.68754256e-01 -9.87708941e-02 -6.83744671e-03] | [9.258797645568848, -1.5756549835205078] |
71463e99-06e2-4fb1-bfb4-7d99c3050521 | local-optimization-achieves-global-optimality | 2305.04819 | null | https://arxiv.org/abs/2305.04819v1 | https://arxiv.org/pdf/2305.04819v1.pdf | Local Optimization Achieves Global Optimality in Multi-Agent Reinforcement Learning | Policy optimization methods with function approximation are widely used in multi-agent reinforcement learning. However, it remains elusive how to design such algorithms with statistical guarantees. Leveraging a multi-agent performance difference lemma that characterizes the landscape of multi-agent policy optimization, we find that the localized action value function serves as an ideal descent direction for each local policy. Motivated by the observation, we present a multi-agent PPO algorithm in which the local policy of each agent is updated similarly to vanilla PPO. We prove that with standard regularity conditions on the Markov game and problem-dependent quantities, our algorithm converges to the globally optimal policy at a sublinear rate. We extend our algorithm to the off-policy setting and introduce pessimism to policy evaluation, which aligns with experiments. To our knowledge, this is the first provably convergent multi-agent PPO algorithm in cooperative Markov games. | ['Jason D. Lee', 'Zhaoran Wang', 'Zhuoran Yang', 'Yulai Zhao'] | 2023-05-08 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [-4.34725106e-01 3.31267387e-01 -7.14294016e-01 3.86796057e-01
-9.36508358e-01 -7.66647160e-01 2.77801812e-01 4.07701313e-01
-9.44486320e-01 1.31346142e+00 6.40931800e-02 -4.59710568e-01
-4.68071401e-01 -4.50515717e-01 -7.47362614e-01 -9.13990259e-01
-4.08608675e-01 7.84913957e-01 -2.84634158e-02 -3.66109848e-01
1.88138068e-01 1.03871047e-01 -7.87128210e-01 -4.07102197e-01
1.02901328e+00 8.91857147e-01 -4.49973065e-03 1.18822467e+00
4.98618662e-01 9.93315279e-01 -6.16776049e-01 -7.91475177e-02
5.54002643e-01 -5.00882924e-01 -8.62999499e-01 4.34587970e-02
-9.18829739e-02 -6.77321017e-01 -3.87187153e-01 1.30241287e+00
7.21624017e-01 4.06354994e-01 4.47306901e-01 -1.42397594e+00
-3.94552946e-01 8.38104069e-01 -7.56687760e-01 3.46974909e-01
1.07956715e-01 4.13182616e-01 1.33561397e+00 7.34997392e-02
4.47739720e-01 1.35887158e+00 3.09924841e-01 7.26090491e-01
-1.25553131e+00 -9.80600491e-02 5.04438698e-01 -1.82638586e-01
-8.75311613e-01 5.36320033e-03 2.53529578e-01 -1.35005623e-01
6.20383143e-01 8.90348628e-02 8.64857495e-01 8.61396194e-01
4.75268126e-01 1.02565432e+00 1.44573092e+00 -4.64186490e-01
6.22168481e-01 -7.72680193e-02 -2.83667445e-01 9.56766188e-01
3.83640468e-01 4.48129058e-01 -1.87602594e-01 -5.99138379e-01
7.99447536e-01 -1.78865299e-01 -1.28792927e-01 -6.91299796e-01
-9.08393502e-01 1.10526729e+00 -1.21188685e-01 -1.95862710e-01
-9.35865462e-01 8.30218911e-01 2.41557017e-01 7.38581836e-01
3.32279623e-01 6.07562184e-01 -5.64992785e-01 -4.76830810e-01
-5.03949463e-01 7.93672681e-01 1.13910794e+00 6.51637495e-01
5.78393579e-01 2.10520193e-01 -2.77130425e-01 3.88260096e-01
3.59996796e-01 8.00197005e-01 1.33469626e-01 -1.72094655e+00
4.74655777e-01 -1.34239376e-01 9.75515902e-01 -4.72434103e-01
-3.52005839e-01 -2.68010199e-01 -2.99959719e-01 6.23232424e-01
7.21512675e-01 -1.03714705e+00 -1.78206339e-01 2.02426243e+00
3.89819175e-01 -2.78592687e-02 2.45393887e-01 9.25807953e-01
-4.35442805e-01 5.61707377e-01 -1.03554145e-01 -8.30040157e-01
1.00437856e+00 -8.81117821e-01 -5.72508931e-01 2.25960575e-02
6.53759539e-01 -2.92561948e-01 8.13199282e-01 3.26183468e-01
-1.34238231e+00 3.30336034e-01 -7.89750814e-01 7.59621024e-01
2.97747433e-01 -4.95444745e-01 4.71237063e-01 5.69617867e-01
-1.20218539e+00 6.91868126e-01 -8.44999373e-01 -3.18075597e-01
3.06927592e-01 3.75995129e-01 2.40974814e-01 4.23990130e-01
-9.51794922e-01 9.91385400e-01 3.84946287e-01 -2.59128541e-01
-1.58675063e+00 -6.70465827e-01 -3.37073594e-01 2.80410927e-02
1.17370951e+00 -7.77261674e-01 2.00274944e+00 -1.07894838e+00
-2.19420600e+00 9.82884616e-02 9.27748382e-02 -6.60806775e-01
8.48545730e-01 -4.57779355e-02 2.04150647e-01 3.35810781e-01
3.51544142e-01 1.13448396e-01 8.08394551e-01 -1.28781354e+00
-1.11524272e+00 -1.50942713e-01 7.16122925e-01 8.31252396e-01
-6.68583959e-02 -8.36691707e-02 3.58845860e-01 -1.60123006e-01
-7.44431138e-01 -1.13020730e+00 -9.50477362e-01 -2.54461706e-01
-1.00817956e-01 -3.54483634e-01 4.30991828e-01 -9.37954560e-02
8.66986394e-01 -1.63888621e+00 2.14427203e-01 3.08921635e-01
4.89836276e-01 -1.85872734e-01 -1.50148481e-01 5.94937146e-01
5.05530238e-01 2.01163888e-01 1.39486328e-01 -2.39217043e-01
6.29293978e-01 2.68530309e-01 -3.84847432e-01 9.86118734e-01
-4.56737608e-01 8.95862997e-01 -1.20130157e+00 -3.68653446e-01
-2.29865648e-02 -4.75952983e-01 -9.02455986e-01 1.49675757e-01
-6.58296168e-01 3.54298711e-01 -1.13820279e+00 1.76368505e-01
3.03788751e-01 -1.14437073e-01 4.48169619e-01 7.07981110e-01
-2.67537236e-01 -2.12125093e-01 -1.13961458e+00 1.25290549e+00
-4.01601911e-01 2.08533779e-01 6.61746323e-01 -1.20016813e+00
2.65852243e-01 5.43365240e-01 1.01337063e+00 -4.28594619e-01
2.81620413e-01 1.91977099e-01 -9.69197787e-03 -3.77156734e-01
5.65094709e-01 -4.66941357e-01 -1.96801484e-01 8.58022571e-01
-7.71248639e-02 -8.78360942e-02 2.36233905e-01 1.72784105e-01
1.05702782e+00 7.25949034e-02 5.61817527e-01 -6.98224664e-01
6.59877285e-02 2.55836308e-01 5.66806138e-01 1.39979196e+00
-8.05394828e-01 -3.28235358e-01 1.05425453e+00 -1.52308285e-01
-1.14532661e+00 -8.66212606e-01 4.48608488e-01 1.40372097e+00
2.82195926e-01 -1.77732185e-01 -6.91262782e-01 -6.48534536e-01
3.98668289e-01 5.20372927e-01 -7.97920406e-01 2.62458712e-01
-5.73558390e-01 -7.97435343e-01 2.85707474e-01 6.20226637e-02
3.55561823e-01 -8.41487467e-01 -7.58835375e-01 6.28651619e-01
2.32835159e-01 -9.37790275e-01 -9.40634012e-01 -5.18457741e-02
-5.52296698e-01 -1.20616388e+00 -8.38218868e-01 -3.32025141e-01
4.92860585e-01 7.42868036e-02 7.57935762e-01 -2.47048646e-01
3.89549524e-01 1.09914494e+00 -4.51674834e-02 -5.40822566e-01
-6.64891183e-01 6.61751553e-02 3.99720758e-01 -2.78784912e-02
-2.60564208e-01 -3.79703432e-01 -9.20172632e-01 9.55343246e-02
-5.43911636e-01 -2.35874772e-01 3.02821368e-01 7.32505620e-01
4.54376817e-01 -1.92728918e-02 6.23453200e-01 -4.38195556e-01
1.31575847e+00 -6.49036646e-01 -1.40915322e+00 2.84353435e-01
-6.75182402e-01 3.50958556e-01 1.01873350e+00 -5.07412314e-01
-7.49485612e-01 -2.06388861e-01 4.40316051e-01 -1.92589849e-01
2.57088810e-01 2.98130959e-01 3.43070358e-01 -1.76563919e-01
4.37312901e-01 1.70437425e-01 5.66662788e-01 -6.14508092e-02
5.07315993e-01 2.55230367e-01 -3.13480794e-02 -1.29579031e+00
5.22788763e-01 4.89942372e-01 2.68094033e-01 -6.65850341e-01
-6.81851625e-01 -3.43163460e-02 9.26910490e-02 -3.71452332e-01
7.25630343e-01 -7.02492833e-01 -1.95398259e+00 1.27707496e-01
-9.72186446e-01 -9.44038749e-01 -5.03944814e-01 5.77381134e-01
-1.31585062e+00 3.66515815e-01 -5.21348417e-01 -1.38238740e+00
-2.89998055e-02 -1.12376082e+00 5.51693022e-01 4.14319366e-01
4.39301372e-01 -1.24692583e+00 6.38542473e-01 -3.09962612e-02
4.38495606e-01 1.55291215e-01 3.62545699e-01 -3.10288697e-01
-4.65052783e-01 3.26540321e-01 3.27194273e-01 3.08121871e-02
-2.18612030e-01 -1.93887696e-01 -2.78276056e-01 -7.27006972e-01
2.99731195e-02 -6.68503821e-01 4.64584768e-01 8.70107532e-01
8.04019153e-01 -8.54616106e-01 -2.77278125e-01 2.10097805e-01
1.62354636e+00 3.42215776e-01 -1.00818023e-01 6.34762406e-01
3.09515327e-01 1.12155810e-01 5.83831370e-01 1.08249331e+00
6.09480739e-01 4.81409878e-01 5.35744131e-01 4.41842645e-01
8.04909110e-01 -2.82989979e-01 7.67968595e-01 1.96648955e-01
-2.36182719e-01 -2.93406487e-01 -5.60854316e-01 3.99151355e-01
-2.45993423e+00 -1.23057997e+00 5.49498141e-01 2.26018834e+00
9.30037200e-01 2.32193843e-02 7.46134341e-01 -5.83376586e-01
5.63796163e-01 2.25352377e-01 -9.69782114e-01 -6.79631650e-01
-6.69855103e-02 5.86274676e-02 1.23868036e+00 1.02315331e+00
-9.34758306e-01 9.71951604e-01 7.05588245e+00 7.28214800e-01
-7.85839438e-01 4.08721000e-01 5.42476714e-01 -5.07821620e-01
-1.26093999e-01 4.96573038e-02 -5.69306314e-01 4.20104653e-01
9.22894537e-01 -6.97837055e-01 1.13949108e+00 9.25376058e-01
1.00651228e+00 -3.07199895e-01 -7.18388021e-01 6.12208366e-01
-5.16693592e-01 -1.32210159e+00 -7.27457762e-01 4.78766888e-01
1.17507243e+00 1.47620991e-01 -3.15349363e-02 2.93476224e-01
1.44922352e+00 -9.41888094e-01 8.96233261e-01 1.97425038e-01
2.12882936e-01 -1.02838862e+00 3.41790944e-01 3.33280236e-01
-1.01602662e+00 -5.34394085e-01 -2.10917622e-01 -3.29521686e-01
2.72969723e-01 -1.33889318e-01 -5.23219943e-01 3.03450048e-01
1.87092870e-01 1.39241561e-01 2.20392644e-01 8.80734444e-01
-7.96089992e-02 6.31425977e-01 -4.46106553e-01 -4.97403026e-01
8.54676962e-01 -6.04740560e-01 7.75376737e-01 5.50189137e-01
-3.14662680e-02 9.20316875e-02 9.64110076e-01 6.12056136e-01
-4.37410399e-02 1.62528843e-01 -4.70131338e-01 -2.59052485e-01
4.11898971e-01 1.03828263e+00 -5.00529885e-01 -2.70765632e-01
-3.01036000e-01 8.33525777e-01 6.34608567e-01 6.27672553e-01
-8.50616276e-01 1.26793198e-02 1.17792547e+00 -4.58330363e-01
3.93718302e-01 -4.01031792e-01 4.54015955e-02 -1.06573248e+00
-2.52837967e-02 -9.48506474e-01 4.65130031e-01 -5.63152507e-02
-1.18205512e+00 -7.41823912e-02 -1.31312860e-02 -7.48146892e-01
-6.26962900e-01 -3.22629809e-01 -6.36274695e-01 4.75876272e-01
-1.34958303e+00 -5.20490825e-01 6.69560909e-01 4.31662023e-01
1.84234783e-01 -2.11869389e-01 5.37476599e-01 -3.23475510e-01
-6.65064752e-01 3.64727318e-01 9.32959855e-01 -1.05995156e-01
2.33680218e-01 -1.74197459e+00 -3.75412673e-01 6.80532157e-01
-3.03570569e-01 -2.97931168e-04 1.01427829e+00 -3.34052831e-01
-1.90050638e+00 -6.77797019e-01 -2.46529520e-01 -1.18108742e-01
1.32145345e+00 4.17267680e-02 -2.53593534e-01 8.05885971e-01
5.66750944e-01 3.43801565e-02 1.39650285e-01 6.19005933e-02
2.41792664e-01 -1.18817300e-01 -1.05647063e+00 1.06383324e+00
7.28586495e-01 -9.07710940e-02 -1.13986887e-01 5.37256360e-01
7.29316175e-01 -6.14403844e-01 -9.67967033e-01 -3.43502522e-01
2.87649274e-01 -4.13132310e-01 4.97455925e-01 -1.12344730e+00
4.35042903e-02 -7.30037019e-02 -2.08239660e-01 -1.86377931e+00
-1.33191198e-01 -1.68139958e+00 -3.44563901e-01 4.63761479e-01
3.41023117e-01 -1.03854966e+00 8.90784025e-01 5.05271077e-01
4.85368893e-02 -8.44614089e-01 -1.28520846e+00 -1.08936977e+00
7.80681252e-01 -6.49197400e-02 1.47713691e-01 5.12285352e-01
4.32126105e-01 3.91338170e-02 -5.79506755e-01 4.47871655e-01
1.12265027e+00 1.58769071e-01 6.32647216e-01 -4.83265251e-01
-8.44298780e-01 -9.04202938e-01 2.57324338e-01 -1.32787633e+00
5.17976761e-01 -4.02263165e-01 1.64421141e-01 -1.30497432e+00
3.22128057e-01 -5.85909724e-01 -2.60234118e-01 1.80929482e-01
-2.08312199e-01 -4.01210487e-01 6.29736185e-01 -4.32424359e-02
-1.19619298e+00 8.65641773e-01 1.54011202e+00 -4.70023826e-02
-3.80557686e-01 1.81837574e-01 -9.30741727e-01 4.33309942e-01
1.08112407e+00 -4.68141258e-01 -3.91389310e-01 -1.13519624e-01
7.24657625e-02 6.18594766e-01 7.85059258e-02 -3.26162308e-01
2.13024333e-01 -1.12256885e+00 -4.70112562e-01 1.21078685e-01
5.78320026e-02 -3.57607603e-01 -4.42715377e-01 1.01746964e+00
-6.65833116e-01 3.39861393e-01 -1.90393910e-01 9.12289917e-01
3.88753802e-01 -3.81640196e-01 1.07828474e+00 -1.94343671e-01
-4.02624339e-01 6.77879989e-01 -9.25656736e-01 7.59369075e-01
1.25642478e+00 4.09039050e-01 -3.78316492e-01 -9.54436421e-01
-4.88463759e-01 7.97091484e-01 2.81406939e-01 -1.57453924e-01
2.38111243e-01 -1.12982833e+00 -8.41810465e-01 -5.81787407e-01
-4.92689937e-01 -4.84276325e-01 -6.24070689e-02 9.45013285e-01
-2.86904007e-01 2.70583570e-01 2.29838751e-02 -1.39731616e-01
-8.51114213e-01 5.28712213e-01 9.49922025e-01 -5.08034945e-01
-4.13167328e-01 1.67738304e-01 -1.52813196e-01 -8.70833397e-02
8.09104890e-02 -1.57666877e-01 3.02142262e-01 -2.47282878e-01
5.11183798e-01 5.11456013e-01 -6.55910671e-01 -1.12864651e-01
-4.25778888e-03 1.43295154e-01 -1.30472198e-01 -9.63116109e-01
1.20255828e+00 -3.39181274e-01 9.97396857e-02 2.54639000e-01
8.20747018e-01 -1.29722757e-03 -1.85321724e+00 -2.24220067e-01
1.10490834e-02 -2.54553467e-01 -4.68973890e-02 -4.68477637e-01
-8.88146460e-01 1.16016284e-01 2.90109873e-01 6.45972610e-01
4.87415135e-01 -1.58635706e-01 3.36118490e-01 5.98616421e-01
6.85346365e-01 -1.56372726e+00 2.30530232e-01 6.89954400e-01
6.12938225e-01 -1.25901783e+00 4.04292857e-03 4.56476420e-01
-9.66653585e-01 8.74188423e-01 5.43998957e-01 -5.06452918e-01
5.53133547e-01 2.81808674e-01 -1.66347697e-01 7.34648779e-02
-1.16029644e+00 -4.79852051e-01 -5.67936599e-01 5.63601434e-01
-2.21163258e-01 5.51784515e-01 -4.79508221e-01 6.64623082e-02
-6.18973421e-03 -5.39540835e-02 9.18092370e-01 1.04533195e+00
-7.99115837e-01 -1.10096908e+00 -4.24705088e-01 4.01515186e-01
-6.68664873e-01 1.80559292e-01 8.10574666e-02 8.58010709e-01
-7.26670742e-01 1.14582622e+00 -1.20336398e-01 3.81617486e-01
1.63662247e-02 -2.21110284e-01 7.87215650e-01 -1.81891426e-01
-4.09613132e-01 4.55593802e-02 1.62231736e-02 -6.90129697e-01
-3.35435390e-01 -6.91666007e-01 -1.24432480e+00 -8.12828600e-01
-2.86664013e-02 4.63738590e-01 2.91717798e-01 1.18105650e+00
2.43912205e-01 1.76070452e-01 1.04665315e+00 -5.38816929e-01
-1.62436044e+00 -6.73349977e-01 -6.89606309e-01 5.20910211e-02
7.40936279e-01 -5.84054708e-01 -3.48314881e-01 -6.80542111e-01] | [4.213704586029053, 2.6356866359710693] |
4c8489bb-a9a8-4368-a4c1-02efc6b61c52 | connect-the-dots-in-situ-4d-seismic | 2105.11622 | null | https://arxiv.org/abs/2105.11622v2 | https://arxiv.org/pdf/2105.11622v2.pdf | Connect the Dots: In Situ 4D Seismic Monitoring of CO2 Storage with Spatio-temporal CNNs | 4D seismic imaging has been widely used in CO$_2$ sequestration projects to monitor the fluid flow in the volumetric subsurface region that is not sampled by wells. Ideally, real-time monitoring and near-future forecasting would provide site operators with great insights to understand the dynamics of the subsurface reservoir and assess any potential risks. However, due to obstacles such as high deployment cost, availability of acquisition equipment, exclusion zones around surface structures, only very sparse seismic imaging data can be obtained during monitoring. That leads to an unavoidable and growing knowledge gap over time. The operator needs to understand the fluid flow throughout the project lifetime and the seismic data are only available at a limited number of times. This is insufficient for understanding the reservoir behavior. To overcome those challenges, we have developed spatio-temporal neural-network-based models that can produce high-fidelity interpolated or extrapolated images effectively and efficiently. Specifically, our models are built on an autoencoder, and incorporate the long short-term memory (LSTM) structure with a new loss function regularized by optical flow. We validate the performance of our models using real 4D post-stack seismic imaging data acquired at the Sleipner CO$_2$ sequestration field. We employ two different strategies in evaluating our models. Numerically, we compare our models with different baseline approaches using classic pixel-based metrics. We also conduct a blind survey and collect a total of 20 responses from domain experts to evaluate the quality of data generated by our models. Via both numerical and expert evaluation, we conclude that our models can produce high-quality 2D/3D seismic imaging data at a reasonable cost, offering the possibility of real-time monitoring or even near-future forecasting of the CO$_2$ storage reservoir. | ['Youzuo Lin', 'Neill Symons', 'Brendt Wohlberg', 'Xitong Zhang', 'Shihang Feng'] | 2021-05-25 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [-7.57551193e-02 -2.40344197e-01 3.49896699e-01 -1.35258555e-01
-8.09463382e-01 -3.36593479e-01 2.94084221e-01 3.43808010e-02
-3.87966573e-01 5.58130801e-01 2.94737101e-01 -4.78919476e-01
-2.16510177e-01 -1.22019708e+00 -7.62801051e-01 -7.27762222e-01
-7.06000388e-01 2.89639324e-01 2.03408554e-01 -1.50782049e-01
3.96021545e-01 8.01353693e-01 -1.36407542e+00 1.12464346e-01
6.22795403e-01 9.99553144e-01 5.02781332e-01 3.52455854e-01
-8.55125040e-02 6.56667113e-01 -2.39046246e-01 3.21707100e-01
3.29653203e-01 4.97323908e-02 -7.07837522e-01 1.25630256e-02
1.05725266e-01 -1.09151852e+00 -3.93826336e-01 7.28751183e-01
4.86641586e-01 9.18496773e-02 4.76251215e-01 -4.03718293e-01
-2.57972807e-01 3.58429819e-01 -7.00226009e-01 3.28395754e-01
-1.52129121e-02 4.84279484e-01 6.98111475e-01 -1.18254280e+00
3.62883657e-01 9.90156174e-01 7.85487711e-01 1.33627385e-01
-1.33408475e+00 -4.15926129e-01 4.76146862e-02 -1.05227895e-01
-1.23800755e+00 -4.24285799e-01 7.59553492e-01 -8.19698989e-01
1.04769480e+00 -1.21999711e-01 8.22489142e-01 4.61328268e-01
5.49264066e-02 3.54100406e-01 1.01502371e+00 -1.84260771e-01
3.26012820e-01 -2.57973015e-01 -2.68025607e-01 6.03525698e-01
-2.06842292e-02 2.97592163e-01 -7.29430616e-01 -1.58912644e-01
1.15414560e+00 5.15136728e-03 -5.82225502e-01 1.91801220e-01
-9.22006845e-01 7.64207244e-01 3.88293892e-01 4.32684720e-01
-6.25014007e-01 3.69529754e-01 7.22127780e-02 1.75442696e-02
7.13189721e-01 3.33105952e-01 -1.72288656e-01 -2.33654141e-01
-1.38392758e+00 4.22492027e-01 6.22171760e-01 3.10229093e-01
9.02614117e-01 6.92625403e-01 2.89647341e-01 6.44499898e-01
5.61271608e-01 9.81752217e-01 1.56060979e-01 -1.48861480e+00
5.03028691e-01 1.73364177e-01 4.97488827e-01 -1.05831277e+00
-1.64078936e-01 -1.51014343e-01 -6.55803084e-01 5.35867035e-01
3.57824743e-01 -2.81024903e-01 -7.20049560e-01 1.33168089e+00
-1.40671879e-02 2.28628457e-01 -9.90266502e-02 8.99589598e-01
3.25422078e-01 1.05031753e+00 -9.37857479e-02 -1.12813085e-01
9.56383944e-01 -2.39270896e-01 -5.28458118e-01 -3.08643788e-01
4.47137892e-01 -2.92505324e-01 9.75241661e-01 1.79116502e-01
-1.19375587e+00 -5.27791642e-02 -9.70840693e-01 5.24023950e-01
-8.13706741e-02 -3.02263647e-01 5.70650339e-01 4.08736795e-01
-1.04361725e+00 1.07232761e+00 -1.39668262e+00 4.69874740e-02
4.02179480e-01 4.97297198e-02 -2.90137917e-01 -5.68475015e-02
-1.18861234e+00 8.37935865e-01 -2.27560788e-01 8.95055234e-01
-1.30830920e+00 -8.79014969e-01 -7.66881108e-01 1.69703320e-01
-1.32945791e-01 -1.55471995e-01 1.14125431e+00 -1.49020359e-01
-1.09355485e+00 2.71598160e-01 -2.09765211e-01 -2.83758193e-01
5.68130493e-01 4.24864851e-02 -2.89913099e-02 3.19511831e-01
1.80867627e-01 2.95252830e-01 3.06453586e-01 -1.35065520e+00
-4.92245942e-01 -4.50435758e-01 -1.76616982e-01 1.15120210e-01
-4.39436704e-01 -3.66977975e-02 -7.60359690e-03 -3.78180891e-01
3.60211521e-01 -5.56133747e-01 -4.37161952e-01 1.97835192e-01
9.83067006e-02 5.65110266e-01 8.42915475e-01 -1.22799325e+00
9.41166043e-01 -1.80646014e+00 -3.35862070e-01 2.04814598e-01
-5.73603660e-02 6.81044068e-03 1.00205958e-01 4.65839177e-01
1.18902996e-01 3.22143972e-01 -9.30001080e-01 -1.84873849e-01
-3.16520542e-01 1.66138217e-01 -4.05988485e-01 4.45975602e-01
2.58124799e-01 4.40478504e-01 -7.93665171e-01 -9.72583145e-02
3.40914339e-01 5.17698109e-01 -2.00732127e-01 2.57970929e-01
-3.73834223e-02 6.17284238e-01 -5.21835446e-01 8.56805325e-01
8.39217424e-01 -1.50328457e-01 -2.52650939e-02 1.63793247e-02
-5.10875165e-01 2.02211663e-02 -1.26620340e+00 1.41526580e+00
-6.77943587e-01 6.17069364e-01 4.45605069e-01 -1.12992632e+00
9.92189467e-01 3.41631949e-01 6.97258651e-01 -1.09542751e+00
-5.89709103e-01 6.94813371e-01 -4.43383485e-01 -8.65490735e-01
6.56562030e-01 -4.96060342e-01 3.78945440e-01 6.44426882e-01
-4.76697624e-01 -2.83652723e-01 -1.20453998e-01 -1.59022227e-01
1.18797123e+00 1.91107355e-02 -8.27330351e-01 -4.38555479e-01
1.32877439e-01 1.85024858e-01 4.38992560e-01 6.10940456e-01
-1.44039597e-02 6.93018854e-01 3.00839007e-01 -4.67097312e-01
-1.28767538e+00 -9.91671681e-01 -3.01865727e-01 2.67282456e-01
-6.57878369e-02 3.13118637e-01 -2.72830188e-01 9.05664116e-02
2.42404476e-01 5.09117723e-01 -4.25178349e-01 2.08938256e-01
-9.30250585e-01 -9.89313781e-01 4.73956227e-01 8.62138271e-01
4.37522441e-01 -8.59561980e-01 -9.03626680e-01 6.98008776e-01
-1.53833196e-01 -7.30967164e-01 1.89462930e-01 1.06387004e-01
-1.35856748e+00 -7.37235546e-01 -1.00450182e+00 -4.93308753e-01
5.29812038e-01 1.44318879e-01 8.98597896e-01 -1.11108020e-01
-2.45279633e-02 1.96572080e-01 4.36635688e-02 2.33951975e-02
-1.82300240e-01 -5.44424117e-01 -1.77264780e-01 -2.70271868e-01
-6.40073717e-02 -6.93973899e-01 -9.04249668e-01 3.82529318e-01
-9.75360215e-01 -2.53823042e-01 1.39966920e-01 8.50120187e-01
5.31712770e-01 2.97763318e-01 4.22977209e-01 -3.24238747e-01
5.16191363e-01 -5.69062114e-01 -9.24725533e-01 4.54596840e-02
-5.18771529e-01 2.98914351e-02 2.30806887e-01 -1.29444199e-02
-1.30363989e+00 -1.12748116e-01 -2.12735176e-01 -3.88493448e-01
1.84898585e-01 1.26929319e+00 1.77265361e-01 -1.81389824e-01
4.95207280e-01 1.83392033e-01 1.39955059e-01 -6.14199460e-01
-2.31839478e-01 7.45172262e-01 5.06824613e-01 -7.44110882e-01
6.57514632e-01 9.64550257e-01 -5.59143126e-02 -1.09250104e+00
-1.34928688e-01 -6.84424415e-02 -1.98729873e-01 -6.40889525e-01
4.34137046e-01 -1.10057473e+00 -6.41061723e-01 8.40917826e-01
-1.03916371e+00 -7.31117845e-01 -1.85308412e-01 5.80673456e-01
-1.82017863e-01 2.93159157e-01 -7.09689558e-01 -1.26072967e+00
-2.96847224e-01 -1.26668930e+00 9.75136340e-01 2.94016767e-02
1.50568634e-01 -9.40730751e-01 1.06092215e-01 2.07450420e-01
9.06709492e-01 5.04148304e-01 7.27663279e-01 1.70933917e-01
-8.48196089e-01 -2.14713991e-01 -2.13175938e-01 2.79269397e-01
1.18504331e-01 6.54327795e-02 -1.10927391e+00 -2.77457058e-01
4.10725713e-01 -1.90807849e-01 1.07669008e+00 7.39072800e-01
1.10896003e+00 -1.59379885e-01 2.87717935e-02 5.12046099e-01
1.63425195e+00 4.55949605e-01 7.24715531e-01 4.63779241e-01
3.15348387e-01 9.34617102e-01 3.67517173e-01 7.38566220e-01
2.71934807e-01 1.44549534e-01 5.47488928e-01 -1.46675166e-02
2.98636287e-01 -1.66831791e-01 1.94741905e-01 6.48831844e-01
-3.88027400e-01 -7.66333267e-02 -1.55617714e+00 1.11603880e+00
-1.55812407e+00 -9.84544456e-01 -7.98980519e-02 2.29978037e+00
4.64551479e-01 4.13222983e-02 -3.58702451e-01 6.78195506e-02
6.55118465e-01 3.78306389e-01 -5.64795434e-01 6.20048158e-02
-1.44358352e-01 2.68908620e-01 8.48814189e-01 6.72495961e-01
-7.18499303e-01 3.61511141e-01 6.48399258e+00 4.44883525e-01
-1.42424357e+00 -2.33372557e-03 8.05469155e-01 -8.39651283e-03
-7.67245591e-01 2.18359455e-01 -4.27393705e-01 4.94479597e-01
1.16174364e+00 3.88198420e-02 3.82238388e-01 4.05170292e-01
9.80580091e-01 -5.62768817e-01 -6.91412926e-01 6.27936661e-01
-4.65755671e-01 -1.88486564e+00 -3.79890293e-01 1.70850679e-01
4.62192595e-01 3.05178642e-01 -2.44568363e-01 -1.86905622e-01
2.55530894e-01 -1.03985858e+00 8.33025694e-01 8.74777198e-01
8.59663486e-01 -4.63174045e-01 6.23654127e-01 2.62814075e-01
-1.35405374e+00 -1.74448088e-01 -8.75652730e-02 -2.17225000e-01
7.56473958e-01 8.11640024e-01 -4.44822073e-01 3.03908616e-01
9.07454431e-01 5.10967731e-01 2.31739897e-02 9.98970151e-01
5.89912534e-02 8.00291955e-01 -7.08423555e-01 2.13494614e-01
2.80762792e-01 -1.76147014e-01 3.21128696e-01 6.96125269e-01
7.87430525e-01 2.88082331e-01 -7.26099014e-02 1.21386528e+00
1.99113101e-01 -4.11005318e-01 -7.33709812e-01 -6.90648556e-02
6.29224479e-01 5.84601104e-01 -3.56557429e-01 -2.64492221e-02
-4.91577923e-01 3.27487320e-01 -1.63290262e-01 6.28585398e-01
-3.34541738e-01 -3.47023040e-01 4.50876743e-01 6.48661375e-01
1.35972664e-01 -5.87131560e-01 -4.75217402e-01 -9.19981182e-01
3.82355422e-01 -3.97637248e-01 3.41010168e-02 -8.79219651e-01
-1.10039902e+00 2.37847641e-01 2.99671441e-02 -1.02966666e+00
-3.00529599e-01 -4.63893652e-01 -6.92801654e-01 1.25157082e+00
-1.80489349e+00 -7.78770328e-01 -3.95754546e-01 1.21688239e-01
2.23342687e-01 -1.62130445e-02 6.79931879e-01 4.84742671e-01
-4.45890278e-01 -1.30264401e-01 5.14827609e-01 1.96695253e-01
1.53808594e-01 -8.79218996e-01 4.12212968e-01 8.98919523e-01
-5.45652628e-01 3.63095909e-01 7.30989039e-01 -9.49320078e-01
-1.42702377e+00 -8.22489321e-01 5.46228170e-01 9.92504209e-02
9.41578805e-01 1.22668058e-01 -1.51300752e+00 5.17966449e-01
-1.64105147e-01 2.28106126e-01 2.81322867e-01 -4.23574179e-01
2.02838808e-01 -3.01164418e-01 -1.21572518e+00 2.03995496e-01
5.09495795e-01 -7.82317340e-01 -2.57163286e-01 1.51740357e-01
3.06607008e-01 -3.19290370e-01 -1.21957541e+00 5.78477979e-01
6.60109103e-01 -9.25730526e-01 7.62100160e-01 -9.33107808e-02
7.94057429e-01 -2.28929266e-01 -5.02370775e-01 -9.85790610e-01
1.20139971e-01 -2.12048337e-01 3.94364111e-02 1.01798654e+00
6.57194138e-01 -6.36403978e-01 1.06263149e+00 1.26438332e+00
-1.59342051e-01 -6.81255817e-01 -1.15522921e+00 -7.42788672e-01
4.62473571e-01 -7.29333162e-01 5.66070139e-01 6.90123320e-01
-2.74616003e-01 -5.77995718e-01 -2.85117567e-01 5.41370690e-01
7.87168205e-01 4.78883609e-02 1.94434926e-01 -8.79963100e-01
-1.51111275e-01 -6.97974414e-02 -1.24019809e-01 -8.58702898e-01
-1.34015113e-01 -3.12315613e-01 2.37839416e-01 -1.59675598e+00
4.65339459e-02 -6.94760084e-01 -9.56878215e-02 4.39107835e-01
3.27497751e-01 -1.08315028e-01 -1.63395092e-01 5.31322837e-01
5.86075723e-01 8.44829381e-01 1.10035503e+00 -3.39049697e-01
-2.24242136e-01 -2.90110081e-01 -8.07492137e-02 6.69042170e-01
4.58398789e-01 -3.19352597e-01 6.07523248e-02 -1.28331983e+00
1.81885153e-01 8.04806113e-01 6.78539455e-01 -9.82737362e-01
2.77138412e-01 -3.73691559e-01 3.17262113e-01 -7.38555491e-01
3.92843157e-01 -7.97243774e-01 4.83589023e-01 5.72182000e-01
5.18144853e-02 -1.26832232e-01 2.00803563e-01 5.08639455e-01
-5.84059834e-01 -4.35213000e-01 7.27211475e-01 -5.69939613e-01
-7.57398486e-01 3.83166671e-01 -5.23241520e-01 -3.20749462e-01
5.35748184e-01 -2.89096534e-01 -3.46469849e-01 -4.30683970e-01
-7.21900940e-01 4.10635859e-01 5.72338581e-01 -2.42679566e-01
8.32779288e-01 -1.03565729e+00 -8.21306407e-01 1.86740369e-01
-2.75029182e-01 4.27583158e-01 5.97424924e-01 5.22503555e-01
-1.19203866e+00 -1.75048839e-02 -6.07436374e-02 -8.28071892e-01
-3.18336040e-01 -2.26152927e-01 8.04359734e-01 -2.22538844e-01
-7.43237793e-01 7.62002409e-01 -2.40076289e-01 -2.18116015e-01
-1.29719198e-01 -1.58942237e-01 3.94115411e-02 1.36301816e-01
6.10246301e-01 4.11597013e-01 2.39122316e-01 -4.45175141e-01
-1.58964932e-01 6.04526639e-01 4.25914228e-01 -5.88340938e-01
2.05469775e+00 -6.37481064e-02 -2.70251967e-02 5.09459674e-01
1.00129187e+00 -3.35503191e-01 -1.70866358e+00 -1.36001721e-01
1.20398067e-02 -8.20245504e-01 5.57711601e-01 -3.06368619e-01
-1.60075784e+00 1.14145386e+00 7.93410122e-01 1.37123346e-01
9.64993060e-01 -2.62642920e-01 7.90275037e-01 4.10621822e-01
4.50977683e-01 -1.02979207e+00 -3.92582864e-02 4.31038946e-01
9.93987501e-01 -1.11304152e+00 -4.33609169e-03 2.26328030e-01
-2.27308810e-01 1.11999738e+00 3.63321483e-01 -1.35663310e-02
9.39093351e-01 6.67396545e-01 9.91291627e-02 -4.63387340e-01
-4.30614769e-01 3.33499163e-01 -7.32864261e-01 3.10715616e-01
1.41858429e-01 -1.86781973e-01 1.78121492e-01 2.33427778e-01
2.58480966e-01 1.72703248e-02 7.48436749e-01 1.34814417e+00
-6.47949576e-01 -7.42359579e-01 -7.17371702e-01 5.37675917e-01
-3.60225409e-01 -4.58222292e-02 5.01038790e-01 4.71817702e-01
-4.49314266e-01 7.32890725e-01 2.35284314e-01 -8.09190646e-02
4.28296059e-01 -2.05962792e-01 1.07455542e-02 -2.90093750e-01
-1.11640781e-01 2.65846819e-01 1.01478919e-01 -5.03052533e-01
-5.31707406e-01 -9.33654189e-01 -1.29381549e+00 -4.60445106e-01
-4.66717362e-01 1.24224856e-01 9.32313204e-01 8.21392715e-01
1.73816741e-01 3.14942181e-01 9.21447575e-01 -1.36728930e+00
-6.19903922e-01 -9.58055437e-01 -1.04932451e+00 3.70972045e-02
4.78208512e-01 -6.30203009e-01 -5.20702004e-01 6.09234646e-02] | [6.864328861236572, 2.6064493656158447] |
2ae9b4d1-6ba0-4dd4-b8f8-f982a4f70b00 | model-compression-for-dnn-based-text | 2210.17326 | null | https://arxiv.org/abs/2210.17326v4 | https://arxiv.org/pdf/2210.17326v4.pdf | Model Compression for DNN-Based Text-Independent Speaker Verification Using Weight Quantization | DNN-based models achieve significant performance in the speaker verification (SV) task with substantial computation costs. Model compression can be applied to reduce the model size for lower resource consumption. The present study exploits weight quantization to compress two widely-used SV models, ECAPA-TDNN and ResNet. The experiments on VoxCeleb indicate that quantization is effective for compressing SV models, where the model size can be reduced by multiple times with no noticeable performance decline. ResNet achieves more robust results than ECAPA-TDNN using lower-bitwidth quantization. The analysis of layer weights shows that the smooth distribution of ResNet may contribute to its robust results. The additional experiments on CN-Celeb validate the quantized model's generalization ability in the language mismatch scenario. Furthermore, information probing results demonstrate that the quantized models can preserve most of the learned speaker-relevant knowledge compared to the original models. | ['Tan Lee', 'Jiong Wang', 'Zhaoyang Zhang', 'Wei Liu', 'Jingyu Li'] | 2022-10-31 | null | null | null | null | ['text-independent-speaker-verification', 'speaker-verification'] | ['speech', 'speech'] | [ 5.97367026e-02 2.28300959e-01 -2.95096666e-01 -4.94999886e-01
-7.82089889e-01 -2.28900779e-02 3.09651613e-01 -6.49018586e-02
-6.04488313e-01 3.92998844e-01 5.38369596e-01 -3.54486614e-01
-2.95818243e-02 -4.68129337e-01 -3.43531072e-01 -5.62718213e-01
8.34906623e-02 1.50897458e-01 1.70055643e-01 -3.76200825e-01
1.54528126e-01 5.38274050e-01 -1.86486030e+00 2.65613288e-01
6.74350083e-01 1.04171693e+00 4.48771209e-01 6.88756227e-01
-1.09808795e-01 6.98439717e-01 -7.90205300e-01 -4.73255843e-01
3.54251832e-01 -1.86066300e-01 -7.58463681e-01 -3.53443444e-01
7.52725303e-01 -3.74874085e-01 -7.26002634e-01 1.14640677e+00
8.53529036e-01 3.27637911e-01 1.83738962e-01 -1.10985887e+00
-4.54351753e-01 1.06030893e+00 -7.72378221e-03 4.74255979e-01
3.72374728e-02 -1.20927535e-01 9.58444476e-01 -7.95186162e-01
4.33235258e-01 1.64902341e+00 8.38626385e-01 9.79022503e-01
-1.13135302e+00 -1.15922248e+00 -7.96063468e-02 7.63166547e-01
-1.88587201e+00 -1.02864897e+00 6.38452590e-01 2.80776829e-01
1.37753963e+00 6.26298308e-01 5.85880518e-01 8.49466324e-01
-1.09632999e-01 5.90598106e-01 7.48130143e-01 -4.81727511e-01
3.32447052e-01 3.21426570e-01 3.07833821e-01 5.64830244e-01
6.84354678e-02 3.01315397e-01 -1.13419545e+00 -1.84319362e-01
3.03041846e-01 -4.57726330e-01 -4.77467328e-01 4.60907191e-01
-7.48373687e-01 7.61923134e-01 3.11785996e-01 3.92011523e-01
-3.23237985e-01 1.43182769e-01 6.92442775e-01 5.21829188e-01
4.21610147e-01 1.01935111e-01 -4.35672581e-01 -4.25689399e-01
-1.30306184e+00 2.16937453e-01 6.44305706e-01 1.14106238e+00
4.34959739e-01 8.65604639e-01 -1.69779971e-01 1.19710636e+00
3.63548249e-01 5.40952444e-01 1.13500166e+00 -1.09270239e+00
5.59550524e-01 1.31310493e-01 -4.93611515e-01 -7.45699883e-01
-7.43961409e-02 -5.07814169e-01 -1.00609803e+00 -1.13493994e-01
-8.12771842e-02 1.38201103e-01 -9.63217795e-01 1.90575516e+00
3.26345526e-02 3.07128519e-01 3.03667724e-01 6.51032507e-01
9.99933958e-01 7.68962383e-01 2.50020176e-01 -1.62063733e-01
1.28716910e+00 -7.68227398e-01 -1.01867974e+00 -1.86009333e-01
5.27323723e-01 -6.57249808e-01 9.06640351e-01 2.12348938e-01
-1.22102880e+00 -8.47242594e-01 -1.16598105e+00 -1.26763269e-01
-2.18545437e-01 -1.38165548e-01 3.31172019e-01 1.09742379e+00
-1.53346467e+00 6.12385571e-01 -6.93276644e-01 -2.02745631e-01
4.12674695e-01 4.69616055e-01 -1.43232346e-01 -8.21712166e-02
-1.54182875e+00 9.27032471e-01 5.05306482e-01 1.05772927e-01
-9.28576469e-01 -8.41658175e-01 -9.97068405e-01 5.58301210e-01
-4.22621936e-01 -3.81755948e-01 1.31537020e+00 -6.27344906e-01
-1.60345018e+00 4.88867462e-01 -6.12333417e-01 -1.13399017e+00
2.70585060e-01 3.85806084e-01 -7.62888134e-01 3.47977459e-01
-6.01923168e-01 1.21076787e+00 8.90854239e-01 -7.98637807e-01
-5.38812816e-01 -1.26467973e-01 -2.49940351e-01 1.87092081e-01
-9.89843845e-01 1.31804407e-01 -2.71323919e-01 -8.70810211e-01
3.41917694e-01 -7.33717382e-01 2.73801517e-02 -3.74461412e-02
-1.73006400e-01 -3.35314333e-01 9.33347523e-01 -1.17763746e+00
1.52042603e+00 -2.27033448e+00 -1.59998760e-01 2.10027575e-01
1.27958030e-01 6.55081093e-01 -3.84260535e-01 1.94038033e-01
-3.13070836e-03 4.03298110e-01 -1.61972523e-01 -6.78801537e-01
2.57508159e-02 3.30857694e-01 -3.22449356e-01 3.35906088e-01
-3.19345951e-01 8.64213705e-01 -2.25631759e-01 -7.23652422e-01
8.88558030e-02 8.78337562e-01 -5.67061782e-01 -3.45821492e-02
1.51287168e-01 -2.58720011e-01 9.82053801e-02 5.17271161e-01
9.68532503e-01 3.36138964e-01 1.02777243e-01 -2.32895553e-01
2.35361964e-01 7.04509974e-01 -1.00283659e+00 1.55969489e+00
-3.38548213e-01 1.11033106e+00 3.85666728e-01 -7.08488464e-01
9.58762646e-01 6.62827730e-01 9.13772359e-02 -9.93432164e-01
-9.10470821e-03 -9.74695161e-02 1.51474580e-01 -2.18603581e-01
8.16581786e-01 -3.02446455e-01 4.21446741e-01 2.14687645e-01
1.01348922e-01 -2.44394183e-01 -8.31260085e-02 2.21575856e-01
5.70949912e-01 -7.84916639e-01 -7.64773414e-02 -3.38542134e-01
5.38481474e-01 -4.38301176e-01 9.50987339e-01 5.68699360e-01
-5.55147886e-01 3.71477574e-01 -1.03172302e-01 -3.66078280e-02
-1.07581711e+00 -7.22782791e-01 -4.21213686e-01 9.89187717e-01
-3.73188779e-02 -7.03734577e-01 -1.04590905e+00 -2.08300203e-01
-1.56973794e-01 1.08019471e+00 -1.55224636e-01 -4.44403917e-01
-6.66120589e-01 -4.54022259e-01 1.11395419e+00 5.56664169e-01
7.88563311e-01 -9.56225574e-01 -2.27837682e-01 1.47866413e-01
-4.61948007e-01 -1.16507113e+00 -6.08479261e-01 -1.09611154e-01
-1.22886968e+00 -3.14583987e-01 -5.45113027e-01 -1.09938514e+00
3.79825205e-01 3.20183456e-01 8.43317747e-01 2.40900517e-01
1.37188643e-01 1.08367771e-01 -2.57554114e-01 -5.34763277e-01
-8.73661757e-01 1.18586175e-01 5.10546327e-01 -2.75332689e-01
5.98773241e-01 -4.43304539e-01 -2.60120362e-01 2.46801317e-01
-7.87200153e-01 -1.87369995e-02 1.58869594e-01 9.06816840e-01
3.95194411e-01 3.74085367e-01 7.52756000e-01 -2.72543848e-01
8.50252271e-01 8.49748179e-02 -4.73334938e-01 1.55475512e-01
-9.90889311e-01 1.10667408e-01 3.54826659e-01 -5.06843984e-01
-9.99869704e-01 -3.40410173e-01 -4.22471732e-01 -6.11088812e-01
1.59255207e-01 3.22712690e-01 -1.64267004e-01 -3.50695044e-01
5.25480926e-01 4.28999275e-01 1.61982462e-01 -5.94146490e-01
2.09367767e-01 1.10378230e+00 2.94543177e-01 -5.38968183e-02
4.65321839e-01 1.55881330e-01 -4.91916358e-01 -1.25535309e+00
-4.72687744e-02 -3.21799606e-01 -1.70719147e-01 -2.23513041e-02
5.67936063e-01 -1.08943498e+00 -7.00194061e-01 3.25709194e-01
-1.15128839e+00 -1.61235392e-01 -5.22667885e-01 6.70139194e-01
-2.36856759e-01 5.11537075e-01 -1.02048874e+00 -8.62362146e-01
-6.53715789e-01 -1.04472172e+00 6.22882485e-01 1.04023308e-01
-1.71333954e-01 -8.67184162e-01 -2.49372080e-01 5.66854477e-01
7.21373022e-01 -7.13368535e-01 9.16213751e-01 -8.15946817e-01
-2.92393893e-01 -1.45081639e-01 4.39056475e-03 8.56402874e-01
-1.83218062e-01 -3.22890490e-01 -1.54450500e+00 -5.92756569e-01
3.01088065e-01 -1.16614141e-01 9.27077055e-01 4.35215801e-01
1.38625884e+00 -7.28484452e-01 -1.84934959e-02 6.26198232e-01
9.34376657e-01 1.47521555e-01 6.81225717e-01 -6.14542067e-02
3.48286718e-01 4.90336061e-01 2.44849280e-01 1.56589419e-01
3.66405934e-01 7.36037731e-01 2.32600808e-01 5.33625633e-02
-7.38200247e-01 -3.71180356e-01 4.66566473e-01 1.68342435e+00
5.94316311e-02 -1.21675409e-01 -7.45997012e-01 5.39403796e-01
-1.32977951e+00 -1.17005742e+00 3.88912588e-01 2.07502890e+00
9.34249282e-01 1.21012092e-01 -2.26750001e-01 6.27844095e-01
8.42561305e-01 2.63582915e-01 -4.06543344e-01 -6.87494159e-01
-4.12710756e-01 8.35184008e-02 5.61275542e-01 7.51650155e-01
-4.76228654e-01 8.39966655e-01 7.71110439e+00 1.26866424e+00
-1.19609046e+00 4.55192983e-01 5.23055375e-01 -3.48544061e-01
-3.12373668e-01 -4.14111197e-01 -1.18570411e+00 5.31318069e-01
1.84915340e+00 -5.41952789e-01 4.93177235e-01 7.60372698e-01
3.04768860e-01 5.23554027e-01 -9.87368405e-01 1.32636130e+00
2.26068690e-01 -1.59253061e+00 3.30223441e-01 1.03820466e-01
4.75830525e-01 2.45178118e-01 1.46692067e-01 4.31875795e-01
-1.16264692e-03 -9.59420025e-01 9.92987692e-01 2.75268793e-01
9.45240915e-01 -8.76713037e-01 8.74938607e-01 3.46559405e-01
-1.35310638e+00 -2.71421671e-01 -9.10849452e-01 2.15871543e-01
2.07348913e-01 1.75398201e-01 -1.23367894e+00 2.49856889e-01
7.62139142e-01 4.77395862e-01 -6.66195989e-01 9.17152822e-01
8.27200636e-02 1.04187834e+00 -4.36227649e-01 7.48885721e-02
-1.24209590e-01 2.68327534e-01 6.23912334e-01 1.41406071e+00
2.62189776e-01 -9.66083705e-02 -3.49864513e-01 7.32604980e-01
-3.60119998e-01 -1.04774237e-02 -3.11825961e-01 1.42399579e-01
1.06563914e+00 6.22815371e-01 -4.41273525e-02 -6.28358960e-01
-6.37819096e-02 6.72743142e-01 2.98903678e-02 4.67071325e-01
-6.72019243e-01 -4.05608952e-01 8.00753236e-01 4.01372127e-02
4.69945937e-01 -7.37828091e-02 -3.54098111e-01 -9.41740572e-01
-8.41099210e-03 -1.13253498e+00 2.41535202e-01 -5.96385837e-01
-8.44393075e-01 9.76036131e-01 9.82711390e-02 -1.17494559e+00
-4.42517459e-01 -1.08675934e-01 -2.95965463e-01 9.32294071e-01
-1.73397362e+00 -9.55375433e-01 -1.12005927e-01 6.41821921e-01
9.14204180e-01 -5.35685480e-01 9.89399612e-01 5.66573679e-01
-5.29707968e-01 1.33249617e+00 4.63049337e-02 -2.53443979e-02
2.92102814e-01 -6.77790284e-01 5.52147388e-01 9.71123755e-01
1.78035155e-01 6.23618066e-01 6.29822552e-01 -3.88627201e-01
-1.13773537e+00 -9.93074715e-01 1.35354340e+00 1.76169619e-01
7.09854364e-02 -4.99655157e-01 -1.14751422e+00 4.74901170e-01
3.59698117e-01 -2.34953761e-01 6.22639894e-01 -2.04785168e-01
-6.39224529e-01 -4.90152717e-01 -1.65691721e+00 4.62210447e-01
9.18335855e-01 -1.06952190e+00 -6.56848013e-01 7.33638257e-02
1.23418105e+00 -1.62964165e-01 -9.51168180e-01 1.37920126e-01
3.75660330e-01 -7.73816288e-01 1.07817292e+00 -3.34837139e-01
-3.29948425e-01 7.41919503e-02 -5.91947436e-01 -1.15277624e+00
-2.80587494e-01 -5.62880397e-01 -4.25367922e-01 1.49394226e+00
4.03921008e-01 -8.00691366e-01 8.44977438e-01 8.68021190e-01
-7.94319436e-02 -2.97619462e-01 -1.69027138e+00 -9.36773777e-01
-1.43851321e-02 -7.28669882e-01 9.58790362e-01 7.77259767e-01
-4.10488658e-02 -5.81420846e-02 -1.45990640e-01 9.33765694e-02
7.58696854e-01 -5.59766769e-01 1.57455236e-01 -1.05783701e+00
7.16025159e-02 -4.44247484e-01 -7.50063241e-01 -1.01235306e+00
6.20510995e-01 -1.17080843e+00 -1.58743590e-01 -1.16989100e+00
-9.90522578e-02 -4.07550514e-01 -5.03948212e-01 4.82854158e-01
2.07769021e-01 3.41378719e-01 5.31128168e-01 3.15052867e-01
-4.97434102e-02 6.99785531e-01 9.19755399e-01 -5.29298902e-01
-1.45361170e-01 -1.73046798e-01 -4.94816571e-01 4.95799989e-01
1.00173485e+00 -5.47427237e-01 -5.69718480e-01 -4.38413858e-01
-3.60314459e-01 -7.83133954e-02 2.15528347e-02 -1.21669364e+00
5.92295885e-01 2.53562868e-01 1.21102072e-01 -6.12748206e-01
7.59996355e-01 -7.99970210e-01 9.39965323e-02 7.42842615e-01
-6.46610796e-01 8.21360014e-03 6.27996743e-01 2.89822996e-01
-4.99228120e-01 -3.71255517e-01 1.08080566e+00 1.95061609e-01
-6.45327628e-01 1.66416228e-01 -5.66227257e-01 -1.87764928e-01
4.24900174e-01 -4.39744711e-01 -2.69298349e-02 -5.81512451e-01
-6.20967329e-01 -9.13458392e-02 1.14074267e-01 4.72409844e-01
9.71791685e-01 -1.42871022e+00 -8.51455867e-01 6.89104319e-01
-1.56214014e-01 -3.90368789e-01 4.16385680e-01 3.81209642e-01
-5.38862228e-01 7.89886415e-01 -9.31026936e-02 -5.96159518e-01
-1.78564298e+00 2.65974700e-01 5.63528657e-01 6.92893565e-02
-5.94483078e-01 9.98041928e-01 -2.05828533e-01 -3.52745056e-01
8.55252922e-01 -5.66051722e-01 -7.68188909e-02 -4.03452776e-02
8.62566948e-01 7.54227281e-01 3.42337132e-01 -8.07612181e-01
-3.46056789e-01 2.99142033e-01 -3.20187569e-01 -3.20808828e-01
1.25967681e+00 -4.14043874e-01 2.36704461e-02 4.09051701e-02
1.26047111e+00 -2.62411803e-01 -9.99930680e-01 -3.51845205e-01
-7.80405626e-02 -3.92549992e-01 5.61147809e-01 -5.46128511e-01
-1.34115064e+00 1.12097633e+00 1.05470550e+00 4.18135114e-02
1.22576606e+00 -3.21368426e-01 1.09222114e+00 3.07393968e-01
3.24105412e-01 -1.29024553e+00 -4.56432164e-01 3.76556814e-01
9.26456332e-01 -9.15213466e-01 -1.23465203e-01 -1.01760268e-01
-5.86729765e-01 8.63385737e-01 3.31047744e-01 5.19346893e-01
7.46102571e-01 2.73133487e-01 1.23325869e-01 2.62173593e-01
-9.48340595e-01 2.76521951e-01 2.50186294e-01 8.03515851e-01
6.60298541e-02 1.80381060e-01 -3.11077554e-02 4.14419651e-01
-8.55528116e-01 -2.04814866e-01 3.52131099e-01 5.50548494e-01
-4.92766738e-01 -9.42719340e-01 -3.65842849e-01 3.15709382e-01
-3.97917122e-01 -4.63944733e-01 -8.34353939e-02 4.42810297e-01
-1.09928027e-01 1.24575031e+00 2.00705081e-01 -6.41752303e-01
3.27619582e-01 4.02203530e-01 1.54817119e-01 -1.22807428e-01
-9.82619882e-01 5.17671928e-02 1.24444552e-01 -6.58464313e-01
-2.14326292e-01 -4.37645376e-01 -1.17802060e+00 -9.49377179e-01
-4.85564440e-01 2.75817782e-01 1.06838167e+00 4.31797355e-01
5.61673105e-01 5.13941348e-01 4.32097554e-01 -5.46723008e-01
-9.35845673e-01 -1.27303159e+00 -6.30589247e-01 2.56192267e-01
6.43393874e-01 -6.77311346e-02 -6.57357812e-01 1.29083335e-01] | [14.27401351928711, 6.234599590301514] |
1bf78881-40a1-4841-9b68-c9541ddc77b2 | pacific-towards-proactive-conversational | 2210.08817 | null | https://arxiv.org/abs/2210.08817v2 | https://arxiv.org/pdf/2210.08817v2.pdf | PACIFIC: Towards Proactive Conversational Question Answering over Tabular and Textual Data in Finance | To facilitate conversational question answering (CQA) over hybrid contexts in finance, we present a new dataset, named PACIFIC. Compared with existing CQA datasets, PACIFIC exhibits three key features: (i) proactivity, (ii) numerical reasoning, and (iii) hybrid context of tables and text. A new task is defined accordingly to study Proactive Conversational Question Answering (PCQA), which combines clarification question generation and CQA. In addition, we propose a novel method, namely UniPCQA, to adapt a hybrid format of input and output content in PCQA into the Seq2Seq problem, including the reformulation of the numerical reasoning process as code generation. UniPCQA performs multi-task learning over all sub-tasks in PCQA and incorporates a simple ensemble strategy to alleviate the error propagation issue in the multi-task learning by cross-validating top-$k$ sampled Seq2Seq outputs. We benchmark the PACIFIC dataset with extensive baselines and provide comprehensive evaluations on each sub-task of PCQA. | ['Tat-Seng Chua', 'Wai Lam', 'Wenxuan Zhang', 'Wenqiang Lei', 'Yang Deng'] | 2022-10-17 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [-1.41807329e-02 6.73688948e-02 4.79202658e-01 -5.74685633e-01
-1.74272668e+00 -8.61067712e-01 5.33066630e-01 3.34683396e-02
-1.79740608e-01 8.75510454e-01 7.20869899e-01 -5.98044157e-01
-3.21925543e-02 -9.18113768e-01 -5.19605815e-01 -5.12923658e-01
2.97376066e-01 7.60154784e-01 -6.56486228e-02 -8.94879043e-01
4.13781554e-01 -1.56197757e-01 -1.26530576e+00 1.19609988e+00
1.41441774e+00 1.33476031e+00 -1.42736971e-01 1.16540468e+00
-7.59458721e-01 1.40991724e+00 -9.88250613e-01 -1.22915316e+00
-1.14507824e-01 -5.80296099e-01 -1.54495966e+00 -3.51538837e-01
3.53644043e-01 -2.20267028e-01 2.48683263e-02 5.84613502e-01
8.16770792e-01 3.59739602e-01 5.35616696e-01 -1.48013806e+00
-6.16321623e-01 8.08315277e-01 7.32796788e-02 2.46555313e-01
9.65990007e-01 4.54677761e-01 1.63054502e+00 -8.00139189e-01
1.34276479e-01 1.59388196e+00 6.22358382e-01 6.87543571e-01
-7.56162822e-01 -7.36296058e-01 1.24151528e-01 2.39658684e-01
-7.74682760e-01 -2.83495575e-01 7.70664632e-01 -1.92066833e-01
1.28793061e+00 4.02317345e-01 1.33275926e-01 1.21597075e+00
1.03286214e-01 1.12681317e+00 1.00438881e+00 1.38879688e-02
3.92015606e-01 -4.57957610e-02 7.74620622e-02 6.10944450e-01
-7.10833907e-01 -3.60186130e-01 -5.24271250e-01 -5.68612039e-01
-4.17348519e-02 -3.49728972e-01 -3.40447687e-02 2.26611778e-01
-1.23964989e+00 1.18305147e+00 1.61534220e-01 -1.56609163e-01
-5.83707802e-02 -2.69075520e-02 7.62681127e-01 6.50817811e-01
2.99619675e-01 9.11860764e-01 -9.81757641e-01 -6.72338486e-01
-1.00590803e-01 8.12210023e-01 1.41185498e+00 1.12082350e+00
5.11447906e-01 -2.41416097e-01 -8.69917452e-01 1.25981092e+00
-8.68400559e-02 6.79875135e-01 4.93194312e-01 -1.47372437e+00
1.37340462e+00 7.94082999e-01 3.86647172e-02 -6.68733656e-01
-3.62166077e-01 -7.38701001e-02 -9.36658263e-01 -5.19858718e-01
5.49785614e-01 -5.68030655e-01 -4.22161818e-02 1.66243970e+00
2.29868114e-01 -2.86414266e-01 6.92854285e-01 5.98759234e-01
1.36183953e+00 8.89377415e-01 1.53169915e-01 -5.35728335e-02
1.46903884e+00 -1.47459161e+00 -7.48544812e-01 3.14453281e-02
8.10815096e-01 -7.10075617e-01 1.40575719e+00 4.59271550e-01
-1.17515993e+00 -5.48366547e-01 -5.17562985e-01 -5.04337490e-01
-5.59929252e-01 -2.81917870e-01 6.28042042e-01 5.24614632e-01
-9.01332855e-01 1.19132668e-01 2.69452780e-01 1.29970238e-01
1.03938729e-01 -7.27756545e-02 2.89229341e-02 4.49707806e-02
-1.82905245e+00 5.98919928e-01 6.17593937e-02 3.78682204e-02
-6.47718668e-01 -9.74905431e-01 -8.72574806e-01 3.00727576e-01
4.59286690e-01 -8.09102178e-01 1.99237144e+00 -6.71936274e-01
-1.96058679e+00 6.88169062e-01 -6.73152953e-02 -3.13184321e-01
5.97998679e-01 -3.95551920e-01 -4.49316293e-01 1.78121433e-01
1.62289202e-01 6.46919787e-01 4.60850149e-01 -9.01635230e-01
-8.75392914e-01 -7.76367113e-02 6.14149749e-01 5.04242897e-01
1.84673294e-01 -2.77001023e-01 -2.68157840e-01 -6.27860665e-01
-5.62279940e-01 -5.88766575e-01 -7.96528831e-02 -5.75158179e-01
-3.75036627e-01 -9.56878364e-01 2.48484537e-01 -6.46000206e-01
1.29492879e+00 -1.84491050e+00 -2.49344744e-02 -1.85815990e-01
1.06851161e-01 1.68590054e-01 -6.19794071e-01 6.37514234e-01
1.33663103e-01 9.92654022e-05 -5.24837375e-01 -3.22816104e-01
1.83615476e-01 1.15183271e-01 -6.13051295e-01 -4.69097733e-01
5.86968362e-01 1.08790088e+00 -1.17573166e+00 -6.19207799e-01
-3.75102937e-01 -2.76854843e-01 -9.99201298e-01 7.04820454e-01
-8.41042101e-01 3.69891405e-01 -6.23509109e-01 7.51332700e-01
7.24821746e-01 -2.72994220e-01 2.36238819e-02 2.69031912e-01
1.72279626e-01 8.03766251e-01 -9.13015246e-01 1.77851510e+00
-7.92659402e-01 8.43621120e-02 1.26672238e-01 -9.61726606e-01
1.11068392e+00 3.32921356e-01 4.10578609e-01 -1.16381383e+00
-3.03691804e-01 1.59129858e-01 -9.02267545e-02 -8.75864565e-01
8.56060207e-01 -1.90320425e-02 -6.51573658e-01 6.24559760e-01
5.12592196e-02 -7.09860086e-01 1.72962204e-01 4.60244000e-01
1.16627872e+00 -1.20740145e-01 -1.81200579e-02 -2.47010335e-01
1.06532896e+00 7.66533464e-02 4.48145181e-01 9.38459039e-01
-4.68304783e-01 6.04109347e-01 1.01114106e+00 -1.82851285e-01
-6.78710997e-01 -1.15335202e+00 5.53354137e-02 1.76898801e+00
-3.67231369e-01 -3.63677919e-01 -6.34596527e-01 -1.10703838e+00
2.41094783e-01 8.57735872e-01 -5.40029347e-01 -4.41356609e-03
-7.52004683e-01 -6.77802265e-01 7.83221066e-01 5.41343272e-01
9.52544153e-01 -1.43317842e+00 -1.12393759e-02 4.18938160e-01
-9.66542542e-01 -1.08729696e+00 -8.50844264e-01 -1.07083537e-01
-6.64115965e-01 -1.26022983e+00 -5.60945809e-01 -8.63625050e-01
-1.76231280e-01 5.15826270e-02 1.90506232e+00 -7.90977292e-03
2.12046176e-01 6.79637194e-01 -7.33368933e-01 -2.71900773e-01
-5.68021953e-01 3.70783061e-01 -5.68339705e-01 -1.25004321e-01
3.01773787e-01 -1.41866475e-01 -6.15673363e-01 2.65094727e-01
-6.81042969e-01 -5.94993010e-02 4.65477854e-01 1.00748730e+00
1.42249838e-01 -4.11266744e-01 1.45398426e+00 -1.20970666e+00
1.31484222e+00 -6.68715000e-01 -2.76743650e-01 5.76723337e-01
-3.91250819e-01 1.23577289e-01 7.84139991e-01 -5.42817498e-03
-1.56004953e+00 -3.57611626e-01 -7.00628281e-01 3.58076990e-01
2.07216308e-01 4.36727613e-01 -3.80137473e-01 5.52529633e-01
5.38199663e-01 2.93975830e-01 8.59027579e-02 -3.07626784e-01
7.19491303e-01 8.24635267e-01 5.86037517e-01 -1.03897834e+00
2.30156451e-01 -1.28908053e-01 -5.53172290e-01 -3.77924711e-01
-1.10318947e+00 -5.87366045e-01 -3.19119751e-01 -1.63709030e-01
8.92686188e-01 -9.96362627e-01 -1.43506110e+00 5.52704871e-01
-1.48856425e+00 -5.29677391e-01 -4.73315082e-02 -6.27973974e-02
-7.90418565e-01 3.48138630e-01 -7.70190239e-01 -8.62959146e-01
-9.23616648e-01 -1.06159377e+00 1.05077815e+00 9.84621719e-02
-1.78406194e-01 -1.01743805e+00 2.04136610e-01 1.16150033e+00
4.70453948e-01 -1.64035559e-01 1.27376461e+00 -9.96466398e-01
-3.11233729e-01 3.08382869e-01 -2.97385275e-01 6.07887506e-01
-1.21554978e-01 -1.88871339e-01 -1.03081477e+00 1.16864055e-01
1.73785407e-02 -9.36326265e-01 7.32288241e-01 -2.39453003e-01
1.46962786e+00 -6.13049507e-01 3.87661278e-01 1.77872539e-01
8.02485585e-01 2.92718709e-01 5.68122506e-01 2.18503311e-01
2.72611380e-01 7.27816284e-01 7.47193575e-01 4.65497106e-01
1.17805624e+00 3.12452734e-01 1.96680933e-01 4.06830937e-01
1.84706673e-01 -1.64372623e-01 5.25930882e-01 1.28892291e+00
3.49575073e-01 -2.91642249e-01 -9.72873569e-01 4.38814312e-01
-1.79168808e+00 -9.09228921e-01 -1.00935839e-01 1.51667690e+00
1.42645419e+00 -3.41786183e-02 1.68017954e-01 -1.54045030e-01
3.27711940e-01 2.26948008e-01 -5.77610970e-01 -8.54005635e-01
-2.70856917e-01 2.93297142e-01 -3.45674723e-01 6.52960837e-01
-9.77168858e-01 6.74299181e-01 6.50633383e+00 7.56402910e-01
-4.21309114e-01 8.27079490e-02 8.90737712e-01 1.32314980e-01
-6.45736635e-01 -4.85152602e-01 -7.09875166e-01 5.36294460e-01
9.48018789e-01 -1.10776134e-01 5.11272371e-01 8.36517394e-01
-2.31608525e-01 4.17104736e-02 -1.14060140e+00 8.16600144e-01
1.54643014e-01 -1.29414451e+00 3.63258034e-01 -7.17656553e-01
7.54387259e-01 -2.49023438e-01 1.08585142e-01 1.39439428e+00
7.75785148e-01 -9.39893067e-01 2.96835572e-01 6.63482964e-01
5.53607523e-01 -7.54887164e-01 1.06034017e+00 2.69190848e-01
-9.62159634e-01 -5.13531089e-01 -1.66754395e-01 -1.18025623e-01
3.21904570e-02 2.98731118e-01 -7.93692291e-01 7.69818842e-01
8.42961788e-01 3.96592706e-01 -5.66356182e-01 5.42308509e-01
-7.37859383e-02 4.54872012e-01 1.85527936e-01 -4.99759108e-01
5.04469812e-01 -3.28787029e-01 1.55426428e-01 1.42147708e+00
-1.52302702e-04 5.16922355e-01 -1.09763481e-01 6.90945923e-01
-4.72994655e-01 2.67550379e-01 -9.11946073e-02 8.43015611e-02
5.72355807e-01 1.15241325e+00 2.46217251e-01 -5.68713486e-01
-5.92236340e-01 7.73031473e-01 5.45976102e-01 2.91602343e-01
-6.95282280e-01 -5.14564097e-01 7.31317997e-01 -6.33407891e-01
2.04972222e-01 2.40854472e-01 -2.89114475e-01 -1.12745035e+00
2.91460250e-02 -1.48881626e+00 9.63634014e-01 -8.44978452e-01
-1.82497704e+00 6.33528709e-01 -2.59635478e-01 -8.68573904e-01
-4.88005519e-01 -3.94243240e-01 -8.65727901e-01 9.80308115e-01
-1.75098670e+00 -8.53469610e-01 -3.29879135e-01 7.92722046e-01
9.54299390e-01 -3.71004760e-01 9.02723312e-01 3.19054574e-01
-4.21008229e-01 8.52628350e-01 8.41695964e-02 3.81696284e-01
9.54612792e-01 -1.68969524e+00 4.34733778e-01 7.99813494e-02
-4.42614317e-01 3.67125213e-01 2.64786780e-01 -3.22401017e-01
-1.47742784e+00 -1.20050561e+00 1.12335980e+00 -9.23761666e-01
7.10608065e-01 -4.94638324e-01 -9.83611763e-01 4.68948632e-01
3.89544755e-01 -5.82720876e-01 8.75805795e-01 4.72536653e-01
-5.25365770e-01 -3.97628844e-01 -1.04193401e+00 3.33893925e-01
6.23608649e-01 -9.12601590e-01 -1.02840817e+00 5.18390357e-01
1.38208830e+00 -6.52852058e-01 -9.96825397e-01 3.19270819e-01
2.96340078e-01 -9.62802291e-01 9.74028766e-01 -9.33432281e-01
8.88067901e-01 3.18260491e-01 -1.95391491e-01 -1.33080530e+00
-1.01998851e-01 -7.00244665e-01 -1.48035631e-01 1.47135520e+00
6.95412040e-01 -5.43035507e-01 5.46255946e-01 5.36244452e-01
-3.62020314e-01 -9.04678583e-01 -9.63674426e-01 -3.28894079e-01
7.50618994e-01 -4.33559865e-01 1.06362963e+00 8.13170671e-01
2.91585594e-01 7.93946564e-01 -4.35290001e-02 -2.31540054e-01
6.09675609e-02 4.65778410e-01 8.98332953e-01 -7.75437653e-01
-2.85360515e-01 -6.58321500e-01 6.46904588e-01 -1.27377129e+00
2.67011553e-01 -9.62592661e-01 2.56589532e-01 -1.36345768e+00
4.53141220e-02 -5.61804771e-01 1.47161037e-02 8.11231956e-02
-5.58650494e-01 -4.18307543e-01 2.55377054e-01 -1.14420600e-01
-1.08602595e+00 8.32335711e-01 1.61678839e+00 -2.97256738e-01
-1.07731886e-01 3.11812431e-01 -1.02433753e+00 1.21306263e-01
6.96241915e-01 -1.47654265e-02 -4.12042707e-01 -3.99318784e-01
5.50983369e-01 7.15346932e-01 1.08832128e-01 -5.70902705e-01
2.16150746e-01 -1.21582210e-01 -1.84633240e-01 -8.38402390e-01
2.96255857e-01 -2.69707531e-01 -7.49872804e-01 1.92302287e-01
-1.01559186e+00 3.88329834e-01 1.80758566e-01 6.17379367e-01
-4.92095649e-01 -1.23420015e-01 4.80126560e-01 -3.33931506e-01
-4.54808444e-01 8.09563100e-02 -3.17330867e-01 9.52587306e-01
4.63936031e-01 6.22325122e-01 -8.14410329e-01 -8.46303403e-01
-4.71696913e-01 1.16504896e+00 -4.98972923e-01 6.65610909e-01
4.18245375e-01 -1.27903652e+00 -1.09879732e+00 -2.11360544e-01
2.60070413e-01 3.23301703e-01 6.28941298e-01 5.56909919e-01
-4.37758088e-01 6.99343324e-01 1.37385294e-01 -3.20111215e-01
-7.93994069e-01 4.90911543e-01 5.78618705e-01 -7.88062632e-01
-2.31111437e-01 9.90556002e-01 7.84068108e-02 -1.27269948e+00
3.33651304e-01 -5.38363576e-01 -5.51088750e-01 3.11408669e-01
5.01204550e-01 3.50645304e-01 2.54788488e-01 -9.88855399e-03
-7.43703768e-02 -1.58928912e-02 -7.63486773e-02 -2.76516266e-02
9.56102669e-01 -8.31144080e-02 -2.62260973e-01 3.84831995e-01
1.36547089e+00 -8.41734707e-02 -1.14689696e+00 -4.74477440e-01
1.32211834e-01 9.78586078e-02 -6.20393276e-01 -1.51252186e+00
-7.98713624e-01 9.90637600e-01 5.13144955e-02 3.69804919e-01
1.00635684e+00 -5.63313365e-02 1.16760659e+00 9.08380032e-01
-1.24759309e-01 -1.30223942e+00 6.93188727e-01 1.19754851e+00
1.42854357e+00 -1.35447526e+00 -5.39246559e-01 -8.73459801e-02
-1.29313767e+00 1.05250406e+00 9.94795978e-01 3.15084726e-01
5.27688324e-01 -4.62566949e-02 3.13391656e-01 -6.82826936e-02
-1.38271618e+00 -2.03439724e-02 -1.90640856e-02 4.58653241e-01
4.19705451e-01 4.60425578e-02 -5.01122698e-02 1.20923519e+00
-7.59446442e-01 -2.64499217e-01 3.58367920e-01 6.49709225e-01
-9.76614356e-02 -7.78583229e-01 -1.50787488e-01 4.72277403e-01
-3.65766376e-01 -3.81444752e-01 -4.45610046e-01 5.16299605e-01
-2.41158888e-01 1.43276012e+00 3.62335384e-01 -3.80288571e-01
6.26795590e-01 4.42164540e-01 -7.33004585e-02 -3.60952526e-01
-1.29179335e+00 -5.79841375e-01 5.94400883e-01 -6.16998255e-01
-2.89762288e-01 -7.27860451e-01 -1.12600648e+00 -3.92460138e-01
7.53316507e-02 7.39782453e-01 2.65099168e-01 9.61019754e-01
4.78858948e-01 6.04290009e-01 1.01351559e+00 1.01191945e-01
-9.81066227e-01 -1.30296171e+00 -2.90464312e-01 5.95794559e-01
4.62615192e-01 -1.86359793e-01 -4.22225416e-01 -1.86866790e-01] | [11.793158531188965, 8.013343811035156] |
c8fe341f-f1f9-4c08-99e9-31952348abc1 | unsupervised-paraphrasing-consistency | null | null | https://aclanthology.org/2021.emnlp-main.430 | https://aclanthology.org/2021.emnlp-main.430.pdf | Unsupervised Paraphrasing Consistency Training for Low Resource Named Entity Recognition | Unsupervised consistency training is a way of semi-supervised learning that encourages consistency in model predictions between the original and augmented data. For Named Entity Recognition (NER), existing approaches augment the input sequence with token replacement, assuming annotations on the replaced positions unchanged. In this paper, we explore the use of paraphrasing as a more principled data augmentation scheme for NER unsupervised consistency training. Specifically, we convert Conditional Random Field (CRF) into a multi-label classification module and encourage consistency on the entity appearance between the original and paraphrased sequences. Experiments show that our method is especially effective when annotations are limited. | ['Ricardo Henao', 'Rui Wang'] | null | null | null | null | emnlp-2021-11 | ['low-resource-named-entity-recognition'] | ['natural-language-processing'] | [ 4.09049988e-01 5.05742610e-01 -5.09389639e-01 -9.28973794e-01
-5.99863589e-01 -6.73080087e-01 5.77391863e-01 3.27791274e-01
-5.62049568e-01 1.03608716e+00 3.68265152e-01 -2.33863726e-01
5.09303391e-01 -4.32628274e-01 -8.00787628e-01 -2.79455125e-01
4.15418565e-01 5.02109289e-01 -2.18148921e-02 1.12823077e-01
2.22738665e-02 3.75879377e-01 -8.71437252e-01 5.32455027e-01
8.08144093e-01 2.45843872e-01 2.94358015e-01 4.12730396e-01
-5.73708296e-01 1.16349149e+00 -5.42223454e-01 -6.09806001e-01
-1.02833044e-02 -4.34712321e-01 -1.53152430e+00 -6.01078989e-03
4.33601081e-01 2.92859115e-02 -4.55201045e-02 1.07310307e+00
9.91426408e-02 3.33437920e-01 4.07595545e-01 -1.15621924e+00
-9.75990832e-01 8.92679513e-01 -2.78317779e-01 -4.25500236e-02
4.80694383e-01 -3.26795608e-01 9.43360209e-01 -9.45146501e-01
8.53839636e-01 9.23319161e-01 1.03294218e+00 8.28480482e-01
-1.37194085e+00 -4.40758586e-01 6.30932748e-02 1.51270255e-01
-1.41709220e+00 -3.34501624e-01 5.58755457e-01 -3.15948844e-01
1.08161902e+00 2.09775284e-01 1.64814770e-01 1.13648748e+00
-1.51084736e-02 6.72335863e-01 1.25081754e+00 -9.56420064e-01
3.41607392e-01 3.36735576e-01 3.12551051e-01 6.21329010e-01
-1.43558541e-02 -1.28585190e-01 -4.70979452e-01 -5.54927468e-01
4.63275075e-01 9.45246741e-02 3.00700013e-02 -8.39544907e-02
-1.20722592e+00 6.35597348e-01 1.95309043e-01 4.19099271e-01
-3.05784434e-01 -4.09758240e-02 4.15930212e-01 1.01685636e-01
4.98945743e-01 5.17032981e-01 -8.17559898e-01 1.10544316e-01
-1.04298377e+00 -1.17656782e-01 6.67141020e-01 1.39070964e+00
9.28625524e-01 -2.04177022e-01 -1.47065833e-01 9.58232701e-01
4.67804164e-01 2.02541560e-01 6.51167333e-01 -8.28663528e-01
2.32128695e-01 3.16867828e-01 2.75674194e-01 -4.61936057e-01
-2.26557955e-01 -8.14853138e-06 -7.80166447e-01 -2.63814926e-01
1.47625655e-01 -5.85828982e-02 -9.99585867e-01 1.87384570e+00
3.94822747e-01 3.69010806e-01 3.18345785e-01 3.36437464e-01
6.31968975e-01 5.83192647e-01 8.45231950e-01 -2.24481657e-01
1.18594527e+00 -1.26309574e+00 -1.01452529e+00 -3.83193582e-01
1.08129990e+00 -8.14415276e-01 7.90117919e-01 -1.62764400e-01
-7.06620276e-01 -8.44813466e-01 -6.06860876e-01 -8.51470884e-03
-4.91078615e-01 1.83325842e-01 5.76162279e-01 5.22502184e-01
-8.59029233e-01 7.31983483e-01 -8.14211309e-01 -4.45827454e-01
-7.29526132e-02 9.13970694e-02 -9.20318246e-01 -7.36990646e-02
-1.35056794e+00 1.28527880e+00 8.76989901e-01 3.73780355e-02
-1.84301689e-01 -4.66929257e-01 -1.18646884e+00 -7.87806809e-02
3.53257507e-02 -2.69080669e-01 1.53278661e+00 -9.97805178e-01
-1.24202657e+00 1.00357544e+00 -4.19635087e-01 -2.80577332e-01
8.54294226e-02 -2.36679196e-01 -6.45880818e-01 -2.60029554e-01
2.32150286e-01 8.79259884e-01 5.05357206e-01 -1.31925189e+00
-4.38530087e-01 -2.82554105e-02 -3.38318080e-01 -3.71977426e-02
-1.20448284e-01 3.37513030e-01 -1.15610510e-01 -1.00400770e+00
8.95161927e-02 -1.02977097e+00 -5.75105071e-01 -3.10694963e-01
-6.01437449e-01 -2.24483177e-01 5.22635579e-01 -1.00753105e+00
1.04222965e+00 -2.18513584e+00 -1.93582997e-01 2.60772079e-01
-1.77813649e-01 9.42362621e-02 -2.32730299e-01 5.75957239e-01
-4.58688498e-01 4.72044677e-01 -4.26042527e-01 -6.69964015e-01
-1.65930957e-01 6.74495578e-01 -2.71700770e-01 2.82277882e-01
4.09105718e-01 7.66591311e-01 -1.08527493e+00 -8.27157617e-01
-9.77311730e-02 2.39498883e-01 -2.03346759e-01 5.70823193e-01
-3.06086004e-01 7.66076982e-01 -1.51999101e-01 4.11012352e-01
7.59706736e-01 -2.94948250e-01 7.88579643e-01 -8.84890556e-02
-1.41108721e-01 4.53774005e-01 -1.11177826e+00 1.87356877e+00
-3.89516026e-01 5.30841589e-01 -3.83138627e-01 -6.02297246e-01
1.11653697e+00 5.84064543e-01 2.04434082e-01 -2.40642458e-01
-4.28769082e-01 9.44587123e-03 -5.95771432e-01 -3.40090662e-01
8.33207726e-01 -4.34758574e-01 -1.74439356e-01 4.54673529e-01
2.12122530e-01 1.51708111e-01 -4.26403843e-02 2.36570239e-01
9.09486353e-01 4.75621194e-01 7.99977601e-01 1.40270665e-01
2.62029827e-01 1.73484832e-01 9.04538393e-01 8.38251650e-01
-1.04757965e-01 2.90712297e-01 -4.72764932e-02 -3.20024759e-01
-1.43432665e+00 -9.12740052e-01 -3.17802936e-01 1.04451728e+00
-1.87025908e-02 -5.84023774e-01 -5.87077975e-01 -1.24419987e+00
-2.54457116e-01 1.01307690e+00 -5.25639951e-01 -1.26386806e-02
-6.05898321e-01 -3.36461663e-01 9.43685889e-01 1.09227633e+00
3.38784307e-01 -1.17468703e+00 3.62989865e-02 2.39294991e-01
-5.30276239e-01 -1.25975251e+00 -5.45028150e-01 7.13028669e-01
-9.07086134e-01 -5.49865305e-01 -3.67047578e-01 -1.31303084e+00
1.01992166e+00 -1.84024721e-01 9.34042692e-01 1.02837712e-01
2.95360953e-01 1.36572286e-01 -7.07192183e-01 1.89584692e-03
-9.78595138e-01 -1.90091878e-01 1.36824921e-01 -3.95217597e-01
3.33276749e-01 -2.38966495e-01 1.70821249e-01 2.98727900e-01
-1.07898331e+00 2.55411208e-01 5.09012163e-01 1.10900557e+00
7.57299483e-01 -2.61526406e-01 3.70926470e-01 -1.49551737e+00
1.25403509e-01 -5.26890516e-01 -7.98781365e-02 7.93903232e-01
-6.57882452e-01 5.89587629e-01 6.90585554e-01 -4.37949210e-01
-1.60078442e+00 5.84298909e-01 -4.04826969e-01 -3.38733166e-01
-7.91257679e-01 5.90243578e-01 -1.26652971e-01 1.88206837e-01
6.72155499e-01 2.61433333e-01 -4.76170391e-01 -9.08829272e-01
6.69156611e-01 9.55240667e-01 8.54746580e-01 -6.20319605e-01
5.22969842e-01 1.53186247e-01 -2.87899584e-01 -5.24500072e-01
-1.01298809e+00 -6.29417002e-01 -1.21422970e+00 9.72436965e-02
8.08665752e-01 -1.02364028e+00 2.71725114e-02 1.42261013e-01
-1.46210873e+00 -3.82313967e-01 -4.96536195e-01 5.50339401e-01
-4.28772390e-01 9.36367631e-01 -7.68546820e-01 -6.41399503e-01
-8.27262625e-02 -4.41676766e-01 1.01419079e+00 2.24129304e-01
-5.63188076e-01 -1.11526537e+00 6.73066258e-01 2.77304798e-01
9.72349271e-02 6.55820742e-02 8.97666097e-01 -1.42831886e+00
9.94730666e-02 -2.63779819e-01 5.57385534e-02 2.88576186e-01
3.31804901e-01 -2.28222720e-02 -1.05402863e+00 1.29414108e-02
-2.98158914e-01 -4.57588077e-01 4.78374213e-01 -2.44030520e-01
9.39081073e-01 -4.46583837e-01 -3.65544766e-01 1.85097203e-01
1.20064688e+00 -8.51150379e-02 6.41747117e-01 3.82491887e-01
7.69649744e-01 6.00622296e-01 8.52446377e-01 4.10284460e-01
3.30841988e-01 6.11700714e-01 -9.57975164e-02 -2.73857236e-01
3.43328132e-03 -6.80447519e-01 1.12060301e-01 7.37891316e-01
4.32790041e-01 -2.85770714e-01 -8.68912399e-01 4.20609564e-01
-1.90479445e+00 -1.19002128e+00 -1.36800110e-01 2.02036428e+00
1.40137029e+00 -1.16960309e-01 -2.29059190e-01 -2.63959646e-01
1.26070082e+00 -6.86774552e-02 -1.58779547e-01 -3.81107926e-01
-1.79011777e-01 2.34311253e-01 4.89496350e-01 4.57902491e-01
-1.24583983e+00 1.12670922e+00 7.26013088e+00 5.94209611e-01
-6.70570493e-01 2.85039455e-01 3.61502498e-01 6.42204583e-01
-3.51654857e-01 5.02418160e-01 -8.53981435e-01 4.88840193e-01
1.20454669e+00 3.28759253e-01 -8.94083902e-02 1.16842163e+00
2.01686937e-02 3.94530743e-02 -1.28889525e+00 5.04109442e-01
-7.81798959e-02 -1.26511168e+00 -3.92091908e-02 -2.66619682e-01
8.14215064e-01 -1.50652185e-01 -4.54623520e-01 2.87621289e-01
6.25418067e-01 -7.24144697e-01 8.47046494e-01 7.34247506e-01
8.17240179e-01 -3.85595441e-01 1.01352382e+00 3.69261533e-01
-1.12762225e+00 4.26077992e-01 -3.90950769e-01 1.33828849e-01
5.16063571e-01 5.15917003e-01 -1.04916263e+00 5.63848674e-01
4.37579364e-01 6.15956604e-01 -5.95297039e-01 9.90692854e-01
-5.34303427e-01 7.29426503e-01 -1.81996524e-01 1.48175880e-01
-1.05462305e-01 1.84937060e-01 1.62137896e-01 1.87565148e+00
-1.55906677e-01 -7.01295882e-02 4.71129596e-01 5.85052490e-01
-2.96752006e-01 1.25626579e-01 -5.54178715e-01 -5.20284697e-02
7.87847161e-01 1.10896838e+00 -7.19491661e-01 -6.53820634e-01
-6.05235815e-01 1.43447435e+00 7.92589843e-01 1.45500749e-01
-8.86083543e-01 -1.68269768e-01 2.55760968e-01 -4.88579661e-01
4.02355999e-01 -2.52709925e-01 -3.14425439e-01 -1.31678104e+00
-8.68073404e-02 -5.96387088e-01 5.89432359e-01 -8.59663665e-01
-1.57651961e+00 5.51788270e-01 -1.62928067e-02 -1.05916131e+00
-1.85968578e-01 -2.16502607e-01 -6.27520144e-01 8.13343883e-01
-1.43132544e+00 -1.35097623e+00 -5.30420877e-02 2.81687140e-01
5.71623921e-01 2.71726102e-01 1.25541019e+00 1.64517939e-01
-5.81165314e-01 7.82515883e-01 3.41387391e-01 4.56613362e-01
1.10169601e+00 -1.44486308e+00 7.46830404e-01 9.32732403e-01
4.49840128e-01 9.45384622e-01 6.53521657e-01 -1.26996553e+00
-5.13250291e-01 -1.40285277e+00 1.56343794e+00 -4.96849775e-01
5.04607201e-01 -1.86825573e-01 -1.24982524e+00 1.21407521e+00
3.05783361e-01 7.76522085e-02 1.34991276e+00 1.92831129e-01
-7.29833663e-01 6.90568388e-01 -1.18942201e+00 3.98090422e-01
7.97373116e-01 -9.19366777e-01 -1.04676425e+00 5.29235542e-01
7.63456643e-01 -2.74014026e-01 -1.03673434e+00 2.93067724e-01
2.82023460e-01 -2.67596096e-01 5.37452579e-01 -1.15805495e+00
1.69208929e-01 -4.41011608e-01 -2.37619683e-01 -1.14647567e+00
-5.29690385e-01 -3.88618261e-01 -1.32954597e-01 1.61828470e+00
6.68856859e-01 -1.87812001e-01 7.20105290e-01 1.03922522e+00
-2.27881208e-01 -1.28046066e-01 -6.66147053e-01 -1.02667558e+00
-8.49547833e-02 -1.55496508e-01 3.72512788e-01 1.70578063e+00
6.50335908e-01 3.28807592e-01 -8.24424326e-01 3.86467338e-01
2.84098983e-01 -1.74472600e-01 3.89561921e-01 -1.03938043e+00
-4.55471903e-01 4.16835278e-01 -1.52334556e-01 -1.02683759e+00
6.27531469e-01 -9.99856472e-01 6.38204992e-01 -1.34328759e+00
4.94609565e-01 -5.22397578e-01 -1.77710250e-01 1.17715573e+00
-4.19737846e-01 -5.67982979e-02 -3.49971280e-02 5.60270607e-01
-8.14010382e-01 2.86251128e-01 4.57473576e-01 1.25503451e-01
-2.32675582e-01 -1.22046694e-01 -3.06841224e-01 6.52263761e-01
9.66744602e-01 -1.03545451e+00 1.22583397e-01 -3.03032219e-01
8.69333893e-02 -1.01304084e-01 1.88766733e-01 -7.04484224e-01
2.66054034e-01 -3.95297319e-01 4.88801450e-01 -3.86782229e-01
-5.26134484e-02 -8.12678576e-01 5.49320698e-01 3.49912018e-01
-1.02432466e+00 1.75055519e-01 2.16404706e-01 7.46698916e-01
-2.28852063e-01 -8.51931989e-01 8.45547020e-01 -1.16356127e-01
-9.85136807e-01 -1.28115162e-01 -5.77761352e-01 -9.41008553e-02
7.85015643e-01 -7.95236230e-02 -2.62507051e-01 -2.98443157e-02
-1.10592747e+00 -3.55931260e-02 6.04410589e-01 2.20132619e-01
3.93623471e-01 -1.35522115e+00 -4.05244499e-01 7.41956308e-02
2.96755880e-01 -1.68362468e-01 7.84657598e-02 3.65504533e-01
-3.21391553e-01 3.38125587e-01 -2.59684324e-01 -2.40561113e-01
-1.38033915e+00 6.56528234e-01 3.00512426e-02 -5.52660048e-01
-4.68197078e-01 6.33081615e-01 -2.84895003e-01 -9.69538629e-01
2.60797143e-01 -1.73800420e-02 -2.50629544e-01 -4.72929686e-01
3.80416632e-01 1.58447415e-01 1.42581947e-02 -7.20439732e-01
-4.55762893e-01 4.92837057e-02 -4.63120073e-01 -1.29967183e-01
1.19278741e+00 -3.12578619e-01 -1.26851290e-01 4.38602149e-01
1.00530326e+00 2.05394968e-01 -1.00598121e+00 -5.14159143e-01
6.47538722e-01 -2.74879515e-01 -3.27062100e-01 -6.98926985e-01
-3.68533164e-01 3.29347253e-01 3.10125530e-01 -6.55119941e-02
6.77946150e-01 3.29812378e-01 5.60305774e-01 7.38159478e-01
1.91696838e-01 -1.10936725e+00 -2.24981323e-01 6.98384583e-01
3.19961369e-01 -1.29443121e+00 -9.46724862e-02 -5.13413846e-01
-9.72287834e-01 1.05539882e+00 8.46165240e-01 3.52114081e-01
4.22863305e-01 2.90506512e-01 5.01862988e-02 3.97037305e-02
-5.15564501e-01 -1.19823869e-02 1.49123129e-02 7.97561586e-01
8.05269539e-01 -1.35883763e-01 -3.45209032e-01 5.49804807e-01
1.29820913e-01 -1.31050169e-01 4.65281844e-01 1.38645577e+00
-3.98364037e-01 -1.54311645e+00 -2.38654315e-01 1.62815645e-01
-4.24346298e-01 -4.93729055e-01 -7.02635229e-01 4.96141762e-01
1.62222967e-01 8.73045146e-01 7.71168768e-02 -3.33440542e-01
2.44847104e-01 6.93678975e-01 1.81852475e-01 -1.07925618e+00
-7.42240071e-01 -2.10049227e-01 3.36196721e-01 -1.98901698e-01
-7.01977432e-01 -6.96092546e-01 -1.66890979e+00 7.37081021e-02
-7.25805581e-01 5.37524283e-01 6.47645056e-01 1.36198354e+00
3.23756427e-01 -3.60882841e-02 6.68691099e-01 -3.94158274e-01
-5.94583035e-01 -9.96977448e-01 -5.14019668e-01 6.52903318e-01
1.32515937e-01 -3.06640059e-01 -1.47078827e-01 6.75113738e-01] | [9.780606269836426, 9.48792552947998] |
06f2a90b-c6fc-4960-a408-c4cd0e86cab1 | downstream-task-performance-of-bert-models | null | null | https://aclanthology.org/2022.lrec-1.451 | https://aclanthology.org/2022.lrec-1.451.pdf | Downstream Task Performance of BERT Models Pre-Trained Using Automatically De-Identified Clinical Data | Automatic de-identification is a cost-effective and straightforward way of removing large amounts of personally identifiable information from large and sensitive corpora. However, these systems also introduce errors into datasets due to their imperfect precision. These corruptions of the data may negatively impact the utility of the de-identified dataset. This paper de-identifies a very large clinical corpus in Swedish either by removing entire sentences containing sensitive data or by replacing sensitive words with realistic surrogates. These two datasets are used to perform domain adaptation of a general Swedish BERT model. The impact of the de-identification techniques is assessed by training and evaluating the models using six clinical downstream tasks. The results are then compared to a similar BERT model domain-adapted using an unaltered version of the clinical corpus. The results show that using an automatically de-identified corpus for domain adaptation does not negatively impact downstream performance. We argue that automatic de-identification is an efficient way of reducing the privacy risks of domain-adapted models and that the models created in this paper should be safe to distribute to other academic researchers. | ['Hercules Dalianis', 'Aron Henriksson', 'Anastasios Lamproudis', 'Thomas Vakili'] | null | null | null | null | lrec-2022-6 | ['de-identification'] | ['natural-language-processing'] | [ 5.92965662e-01 6.04745865e-01 7.98621699e-02 -5.62128305e-01
-1.36021161e+00 -8.00186872e-01 3.28159183e-01 7.13197768e-01
-1.05465722e+00 1.17737091e+00 5.98304272e-01 -2.87586629e-01
-2.31928110e-01 -3.25151533e-01 -3.22380066e-01 -6.37878418e-01
4.61292505e-01 7.77830124e-01 7.53465071e-02 -5.23088947e-02
1.59053072e-01 3.06406677e-01 -8.87382746e-01 5.30573785e-01
7.40955412e-01 1.87742323e-01 -2.01275334e-01 4.29979116e-01
2.79722065e-01 3.19800258e-01 -9.43665266e-01 -8.86277556e-01
5.49110174e-01 -3.90207797e-01 -1.03380322e+00 -4.03072149e-01
3.26445512e-02 -4.03023899e-01 2.41173178e-01 9.95348215e-01
1.02959824e+00 -2.66251087e-01 3.19906086e-01 -5.95448673e-01
-2.27364123e-01 7.91658759e-01 -6.09718449e-02 3.56976762e-02
3.82538915e-01 -1.08755380e-01 6.18897915e-01 -1.57724828e-01
9.62396801e-01 8.12482655e-01 9.76491392e-01 8.94074321e-01
-1.47209525e+00 -8.84828866e-01 -5.13167858e-01 -4.35541630e-01
-1.50441647e+00 -8.52519095e-01 3.13069344e-01 -5.37378728e-01
8.18181813e-01 5.46724081e-01 2.26222143e-01 1.23381960e+00
7.45133162e-02 8.30339501e-04 9.24912989e-01 -6.98429704e-01
2.92822927e-01 8.74083102e-01 1.66206434e-01 1.96184963e-01
7.62666762e-01 1.48058668e-01 -2.15859815e-01 -9.80930090e-01
-3.45497206e-02 -4.22398955e-01 -1.51774630e-01 -3.09863359e-01
-9.75922644e-01 8.21278095e-01 -2.86940485e-01 5.36706686e-01
-1.90782636e-01 -5.03467262e-01 8.65369201e-01 2.00380057e-01
5.76959610e-01 8.98437023e-01 -6.12770140e-01 -2.01551348e-01
-9.28693593e-01 5.43340929e-02 7.89629579e-01 9.36579466e-01
1.59782588e-01 -5.96347928e-01 -1.82729483e-01 9.82386172e-01
-5.43033667e-02 1.17371611e-01 9.79093611e-01 -6.76909268e-01
4.14302289e-01 5.62223613e-01 3.35947782e-01 -7.37669110e-01
-4.53508049e-01 -6.84724525e-02 -3.72032672e-01 -1.68766409e-01
7.10109830e-01 -4.41672772e-01 -6.77690864e-01 1.74464500e+00
3.32552612e-01 -4.69350308e-01 4.36963081e-01 4.41075265e-01
7.26720989e-01 3.20611633e-02 8.05596054e-01 -2.60891348e-01
1.37549567e+00 -1.28257051e-01 -8.67026508e-01 -1.38612226e-01
1.23573422e+00 -7.11272061e-01 4.53193605e-01 1.68890476e-01
-8.70732129e-01 -3.82542377e-03 -8.87232184e-01 -1.71033919e-01
-3.09179872e-01 -1.59423545e-01 1.23474926e-01 1.38977790e+00
-9.74685073e-01 4.85315263e-01 -7.09238827e-01 -6.64094746e-01
6.43029392e-01 7.42218733e-01 -7.46398509e-01 8.23467970e-02
-1.57188082e+00 1.01737976e+00 6.83356166e-01 -4.34981227e-01
-8.95183310e-02 -7.38518000e-01 -7.58848369e-01 -1.69910535e-01
4.64032628e-02 -6.50649309e-01 1.11228681e+00 -1.02004361e+00
-9.01705146e-01 1.47041619e+00 -7.24300221e-02 -5.83044887e-01
7.54632235e-01 5.99592105e-02 -5.88208020e-01 8.24211314e-02
4.10062164e-01 1.11854680e-01 3.64271432e-01 -1.09211135e+00
-8.21323991e-01 -4.99036700e-01 -5.51536202e-01 3.96208130e-02
-3.72272432e-01 4.34740365e-01 1.80546753e-02 -6.31046712e-01
-3.68995100e-01 -8.26396704e-01 -3.91138494e-01 -5.42812824e-01
-4.30054426e-01 3.31965566e-01 2.19457433e-01 -1.14560354e+00
1.47115576e+00 -2.37158585e+00 -5.46679258e-01 3.20063144e-01
2.06014335e-01 7.13333130e-01 5.97769581e-02 3.51494610e-01
-3.32457036e-01 6.57405794e-01 -5.28430700e-01 -3.11550915e-01
-3.21734220e-01 7.86286592e-02 -4.24177572e-03 6.06902838e-01
-1.59961417e-01 5.48618078e-01 -9.94412780e-01 -6.62435353e-01
-6.05094768e-02 3.96985859e-01 -5.35803318e-01 -1.02058142e-01
3.45807433e-01 3.61030638e-01 -3.41642290e-01 2.39623293e-01
7.15919852e-01 3.15973371e-01 5.49676299e-01 2.23873168e-01
2.44709373e-01 4.61892396e-01 -6.09329224e-01 1.43712723e+00
-1.25382636e-02 2.16447279e-01 6.15881942e-02 -7.09825575e-01
9.79519725e-01 6.28930926e-01 6.04536831e-01 -5.30425072e-01
1.34140074e-01 3.83643895e-01 1.91578656e-01 -6.36341512e-01
4.40531790e-01 -8.99786532e-01 -4.04400617e-01 4.65252846e-01
-2.77096748e-01 1.32201523e-01 -1.29997283e-01 1.31798759e-01
1.11237085e+00 -3.75851244e-01 8.41140330e-01 -3.59995753e-01
6.36955380e-01 2.28070915e-01 9.99825001e-01 5.55401802e-01
-5.86136341e-01 7.21029997e-01 4.27517176e-01 -8.98293406e-02
-9.35440183e-01 -6.28091276e-01 -6.28102541e-01 5.56054592e-01
-5.58244646e-01 -5.02003908e-01 -1.04110706e+00 -1.01586306e+00
1.45713881e-01 1.22875917e+00 -6.98701501e-01 -5.55041254e-01
-3.52948636e-01 -1.01242781e+00 1.11497521e+00 2.42597282e-01
-3.09011117e-02 -7.45835483e-01 -6.59998655e-01 3.02552134e-01
-3.36277783e-01 -9.92767334e-01 -4.74596977e-01 1.74747720e-01
-7.50267982e-01 -1.03218746e+00 -5.68342626e-01 -6.18354857e-01
8.43346596e-01 -3.37591231e-01 8.56718242e-01 -3.27121979e-03
-2.83361465e-01 2.36452192e-01 -4.23886865e-01 -8.03488493e-01
-1.23822188e+00 2.40114868e-01 -3.68315130e-02 -1.38951233e-02
1.27395165e+00 -1.10811330e-01 -3.04035515e-01 1.41700253e-01
-9.19295132e-01 -6.33086801e-01 3.35234880e-01 8.31385553e-01
3.20734411e-01 -2.10995391e-01 8.27313364e-01 -1.91812050e+00
9.46679115e-01 -3.87906373e-01 -2.32115880e-01 1.98873892e-01
-9.27551985e-01 1.06106997e-01 5.69839597e-01 -1.65068403e-01
-1.20257831e+00 4.34817940e-01 -3.18923295e-01 3.10529262e-01
-4.10114110e-01 1.51279613e-01 -4.13209856e-01 1.21469438e-01
1.12825930e+00 -2.00350747e-01 3.21101755e-01 -5.57654321e-01
-1.29614398e-01 1.28728962e+00 3.34780604e-01 -1.88612685e-01
5.23766220e-01 4.08501238e-01 -3.73447627e-01 -6.29692316e-01
-5.18186212e-01 -7.18275309e-01 -7.77188599e-01 5.32042742e-01
7.49850094e-01 -7.92384505e-01 -2.50096858e-01 2.82706678e-01
-6.53582573e-01 -4.52937782e-02 -5.12854874e-01 4.93179739e-01
-2.93290317e-01 6.09885037e-01 -2.99874395e-01 -7.02196002e-01
-4.64455515e-01 -7.97907829e-01 9.98711288e-01 -2.40399078e-01
-1.17479086e+00 -1.07071459e+00 3.99391234e-01 6.73724592e-01
5.17387018e-02 3.95830750e-01 9.51032400e-01 -1.41416287e+00
4.21730191e-01 -7.19004214e-01 1.87139064e-01 2.96587944e-01
5.41090429e-01 -3.45075518e-01 -1.14096189e+00 -3.65998805e-01
3.70664626e-01 -1.13780282e-01 5.75786531e-01 1.27373397e-01
6.18552804e-01 -4.39041615e-01 -7.38520026e-01 5.44368088e-01
1.28461266e+00 4.73037153e-01 7.50803232e-01 5.95454097e-01
3.60069692e-01 8.28317583e-01 6.94082499e-01 3.62208575e-01
1.62384376e-01 4.73583758e-01 -2.56210864e-01 8.29495676e-03
4.38768178e-01 -2.38712028e-01 1.41334951e-01 2.59905756e-01
2.33436093e-01 2.07585492e-03 -1.10048461e+00 9.51753557e-01
-1.43413877e+00 -9.88261342e-01 -6.58784434e-02 2.52742934e+00
1.26952100e+00 -5.66426478e-02 3.74909520e-01 1.08778983e-01
8.44888151e-01 -4.74811912e-01 -3.11742991e-01 -8.43974411e-01
-8.46429989e-02 3.36319655e-01 9.89573121e-01 4.90105242e-01
-1.04591250e+00 6.62948370e-01 6.70085573e+00 3.51990253e-01
-7.24336207e-01 2.16253176e-01 6.11468971e-01 -2.89518863e-01
-1.51535675e-01 -2.14034021e-02 -7.60806084e-01 6.11288488e-01
1.59289980e+00 -6.89295590e-01 -1.91638485e-01 6.90853655e-01
3.44608873e-01 -1.58516496e-01 -1.20436227e+00 6.69798434e-01
5.77565767e-02 -1.14300311e+00 -1.87035635e-01 4.11160439e-01
3.51153284e-01 -4.17741358e-01 -8.98506418e-02 4.99993563e-03
6.07533097e-01 -9.27746594e-01 4.19877559e-01 4.37731385e-01
1.08233500e+00 -8.20954144e-01 1.32193959e+00 2.88112909e-01
-1.41151398e-01 -1.48407698e-01 -3.74504745e-01 2.83942610e-01
9.25829709e-02 4.43615645e-01 -1.46174455e+00 4.41545874e-01
6.60407186e-01 2.51512289e-01 -6.19321942e-01 8.45583737e-01
2.88210392e-01 5.26989937e-01 -2.18903646e-01 1.76570892e-01
-2.13927597e-01 1.55629829e-01 3.87766331e-01 1.48299301e+00
7.30140433e-02 1.92425102e-01 -5.42744935e-01 4.37115252e-01
-5.46350777e-02 3.30178142e-01 -8.57999146e-01 -3.27340513e-01
4.62984651e-01 9.05980289e-01 -3.08495909e-01 -2.51883566e-01
-3.76630336e-01 9.83350813e-01 -1.13388561e-02 -3.76977623e-02
-2.54686773e-01 -5.00038087e-01 7.55939364e-01 4.50180292e-01
-4.81133088e-02 5.69585204e-01 -4.71733868e-01 -7.39232838e-01
-8.52252245e-02 -1.25150621e+00 1.01185226e+00 -3.16854447e-01
-1.13660777e+00 5.11967003e-01 8.06977600e-03 -1.16448617e+00
-4.98672098e-01 -1.46419942e-01 1.76862553e-01 1.25813198e+00
-9.56982553e-01 -7.21755266e-01 1.15767844e-01 3.46184522e-01
-1.47608772e-01 -1.59685150e-01 1.38676751e+00 2.75343180e-01
-4.47642118e-01 1.28320885e+00 4.78825390e-01 4.11142141e-01
1.40802395e+00 -1.00929487e+00 3.44477624e-01 6.52923584e-01
-4.80928570e-01 1.08481395e+00 6.17603183e-01 -9.38094318e-01
-4.76460755e-01 -1.06477273e+00 1.57256198e+00 -1.05217385e+00
2.25618646e-01 -5.23510993e-01 -1.06829679e+00 8.41134191e-01
-1.49354964e-01 -6.09494209e-01 1.35682547e+00 -1.32517830e-01
-5.46488762e-02 -1.24341184e-02 -2.02502632e+00 3.08435738e-01
7.51245081e-01 -5.53283155e-01 -9.23090160e-01 3.63578945e-01
5.77678382e-01 -7.46293291e-02 -1.03553498e+00 3.06258164e-02
4.91586596e-01 -8.03202331e-01 8.52337837e-01 -9.77426946e-01
6.63469285e-02 1.63932860e-01 5.51853776e-02 -1.23850024e+00
-3.60134631e-01 -7.33012915e-01 8.07270288e-01 1.51952565e+00
7.65966833e-01 -9.71819997e-01 7.27508903e-01 1.50250316e+00
4.25757885e-01 3.33713666e-02 -1.11722112e+00 -6.01674676e-01
4.52830195e-01 9.26913396e-02 7.05964565e-01 1.31899786e+00
5.56036770e-01 2.56822556e-01 -1.75071180e-01 1.04691476e-01
4.80668455e-01 -5.68731546e-01 5.25270760e-01 -1.32170689e+00
-2.65567750e-02 1.83888033e-01 -5.20611107e-01 2.17090249e-01
7.48985261e-02 -9.25081313e-01 -6.89045191e-02 -1.08768618e+00
2.79092312e-01 -5.61863542e-01 -2.02992141e-01 5.01071811e-01
-3.52104366e-01 -1.00959903e-02 -8.30220357e-02 1.84818700e-01
3.42259333e-02 -1.51185691e-01 5.77928007e-01 1.90683886e-01
-4.46112126e-01 6.16498403e-02 -1.40908229e+00 5.54231107e-01
9.40053880e-01 -1.19898653e+00 -3.59857708e-01 -1.75221395e-02
-4.79583889e-02 -2.15908885e-01 5.82342967e-02 -5.52754998e-01
9.32025723e-03 7.86257237e-02 2.49968812e-01 5.60131893e-02
1.72558334e-02 -9.96087968e-01 3.68638664e-01 5.81624031e-01
-6.35101914e-01 -2.05759987e-01 5.75380206e-01 5.43906868e-01
8.11277851e-02 -7.13228583e-01 8.81911099e-01 -4.82451975e-01
-3.17153066e-01 -3.55845004e-01 -5.42318463e-01 2.48885125e-01
1.19175899e+00 -5.49557209e-01 -1.86427340e-01 -2.52375931e-01
-8.99644911e-01 -3.30954678e-02 7.53633261e-01 1.86512824e-02
1.21731579e-01 -7.08011508e-01 -7.70393610e-01 3.39384496e-01
3.66653442e-01 -2.76661783e-01 2.15414703e-01 6.46080315e-01
-5.41166365e-01 6.21767998e-01 -1.99459895e-01 -3.56727168e-02
-1.82094443e+00 7.83168495e-01 3.50312859e-01 -3.04556131e-01
-6.96739137e-01 7.20256925e-01 9.03493688e-02 -5.08211195e-01
4.64316569e-02 -3.31697613e-02 -1.34468138e-01 1.48501515e-01
5.81605196e-01 8.50197747e-02 3.01036149e-01 -8.16914022e-01
-6.95848644e-01 -7.29557732e-03 -5.42879641e-01 -3.58456194e-01
1.32149804e+00 -2.54232526e-01 2.20152233e-02 6.70987666e-02
1.19460785e+00 3.97809684e-01 -5.71154058e-01 -2.44423617e-02
3.37118298e-01 -4.39009517e-01 -1.70405060e-01 -1.12286317e+00
-5.07535875e-01 5.47133803e-01 5.78661501e-01 -1.56145513e-01
1.09353995e+00 -2.29284480e-01 5.55933893e-01 1.16559997e-01
2.12224394e-01 -1.27473545e+00 -9.29488838e-01 -7.13170022e-02
5.40244579e-01 -1.04444671e+00 2.00559616e-01 -3.74681354e-01
-1.00998235e+00 5.75747669e-01 1.60932630e-01 4.86081481e-01
4.60427910e-01 3.39386135e-01 3.90957624e-01 3.13259959e-02
-4.00389642e-01 1.68199688e-01 -1.96263254e-01 1.22987521e+00
4.06817287e-01 2.39885841e-02 -9.77393150e-01 9.73233044e-01
-3.35433245e-01 2.37586305e-01 8.32709074e-01 9.21326816e-01
2.49581501e-01 -1.63767445e+00 -5.08419871e-01 6.40752196e-01
-1.06281924e+00 -3.18282157e-01 -1.10413098e+00 9.23071265e-01
9.78730991e-03 1.01485097e+00 -2.13852257e-01 -3.31527710e-01
4.81497765e-01 6.45088136e-01 -9.00973454e-02 -8.94699931e-01
-1.16524518e+00 -9.09365863e-02 7.82791853e-01 -1.45186648e-01
-2.16017425e-01 -1.21666050e+00 -1.03838992e+00 -1.92800969e-01
-1.81928396e-01 4.05588925e-01 4.34620678e-01 6.79549754e-01
7.78643072e-01 1.84285700e-01 7.84317777e-02 2.89118081e-01
-9.75853205e-01 -7.29412615e-01 -5.68112314e-01 8.56260657e-01
2.84847707e-01 -2.08646119e-01 -3.61076862e-01 2.83771813e-01] | [6.810202121734619, 7.039198875427246] |
42a3dcef-3e6d-4c1f-b247-5a29d12d374b | deformable-filter-convolution-for-point-cloud | 1907.13079 | null | https://arxiv.org/abs/1907.13079v1 | https://arxiv.org/pdf/1907.13079v1.pdf | Deformable Filter Convolution for Point Cloud Reasoning | Point clouds are the native output of many real-world 3D sensors. To borrow the success of 2D convolutional network architectures, a majority of popular 3D perception models voxelize the points, which can result in a loss of local geometric details that cannot be recovered. In this paper, we propose a novel learnable convolution layer for processing 3D point cloud data directly. Instead of discretizing points into fixed voxels, we deform our learnable 3D filters to match with the point cloud shape. We propose to combine voxelized backbone networks with our deformable filter layer at 1) the network input stream and 2) the output prediction layers to enhance point level reasoning. We obtain state-of-the-art results on LiDAR semantic segmentation and producing a significant gain in performance on LiDAR object detection. | ['Yuwen Xiong', 'Raquel Urtasun', 'Mengye Ren', 'Kelvin Wong', 'Renjie Liao'] | 2019-07-30 | null | null | null | null | ['lidar-semantic-segmentation'] | ['computer-vision'] | [ 2.19020978e-01 2.92405009e-01 3.98609862e-02 -5.77310979e-01
-5.15159428e-01 -7.55410612e-01 4.15811151e-01 2.53366619e-01
-2.40905926e-01 -4.69059721e-02 -2.75452733e-01 -3.57393563e-01
2.31258810e-01 -1.33582878e+00 -1.27245581e+00 -7.02642649e-02
7.25466525e-03 8.66263092e-01 7.96885252e-01 -5.15211485e-02
1.54540822e-01 1.32140899e+00 -1.60304356e+00 5.16453326e-01
7.94757128e-01 1.15210903e+00 1.51959360e-01 6.62658155e-01
-6.66856349e-01 1.58600494e-01 -1.47686765e-01 -2.38443896e-01
7.84761667e-01 6.38036489e-01 -6.60041153e-01 1.02120437e-01
1.09854186e+00 -7.96375215e-01 -2.98596829e-01 1.04468560e+00
6.93058446e-02 -6.43998161e-02 5.38205326e-01 -1.19804776e+00
-6.23748422e-01 4.17526007e-01 -4.96484637e-01 -1.28302097e-01
-2.90879887e-02 4.54004735e-01 8.58079791e-01 -1.06572580e+00
6.26127183e-01 1.67083156e+00 1.05659270e+00 4.09123182e-01
-1.24946308e+00 -7.66106486e-01 5.08644998e-01 -2.60259569e-01
-1.28068542e+00 -1.49436817e-01 7.76822031e-01 -5.32807231e-01
1.25021696e+00 5.19193374e-02 7.11589634e-01 6.27546072e-01
6.14471361e-02 5.98205745e-01 7.04888642e-01 8.96567702e-02
1.54804319e-01 -3.11738968e-01 1.03304356e-01 7.26178288e-01
1.96722388e-01 8.72421861e-02 -2.95029014e-01 1.58425327e-02
1.23316956e+00 4.32748914e-01 2.49525145e-01 -6.12795115e-01
-8.72712493e-01 7.35296249e-01 1.12649810e+00 -2.24682182e-01
-3.71962458e-01 6.75940037e-01 -9.80649143e-03 8.92428681e-02
7.17450082e-01 5.81528172e-02 -5.63440561e-01 3.37545455e-01
-1.02007663e+00 5.28235078e-01 4.97151405e-01 1.09661090e+00
1.08444285e+00 -1.29782990e-01 4.09387983e-02 2.63412416e-01
6.10479832e-01 8.49341094e-01 -4.94811773e-01 -1.29165089e+00
4.63932663e-01 1.01412487e+00 -1.86010642e-04 -7.20820725e-01
-2.91108906e-01 -2.04312354e-01 -3.41750026e-01 9.19327140e-01
2.66383946e-01 3.54891002e-01 -1.52917707e+00 1.15045667e+00
4.17251408e-01 4.42049861e-01 -3.62040669e-01 9.59991992e-01
8.81697655e-01 5.51908612e-01 1.63200215e-01 7.74630904e-01
9.63922203e-01 -5.65679550e-01 1.97824284e-01 -3.88218075e-01
2.01036364e-01 -5.28312325e-01 9.30076241e-01 1.21590771e-01
-1.41070032e+00 -7.53660500e-01 -1.01813126e+00 -6.79855883e-01
-4.62367594e-01 -4.73128647e-01 7.16613650e-01 3.34918380e-01
-1.05936289e+00 9.81762946e-01 -1.31566834e+00 -2.03012064e-01
1.09805167e+00 5.01020193e-01 -8.45184326e-02 -1.97335154e-01
-4.06047970e-01 7.10069060e-01 1.76832035e-01 -1.92304537e-01
-8.86532068e-01 -1.31457746e+00 -6.14382982e-01 2.31757030e-01
3.03395301e-01 -1.04179490e+00 1.34094012e+00 -5.25070667e-01
-1.04803205e+00 1.18057859e+00 -1.33672580e-01 -4.85171556e-01
7.25695312e-01 -4.54051226e-01 2.49403328e-01 2.87315309e-01
4.54110615e-02 1.30171633e+00 7.54899085e-01 -1.57123613e+00
-7.65352011e-01 -6.43827736e-01 1.76095635e-01 8.04154053e-02
3.10704917e-01 -6.14628375e-01 -3.78683388e-01 -2.46392056e-01
6.43908858e-01 -6.96221709e-01 -3.40206802e-01 8.54527950e-01
-3.24836463e-01 -4.78719831e-01 1.09844208e+00 -2.57794797e-01
1.28206864e-01 -2.06220865e+00 -1.37843102e-01 2.76011229e-01
4.37403917e-01 1.60785303e-01 -1.44572645e-01 -2.10075974e-02
7.04929680e-02 3.40140343e-01 -3.99286628e-01 -5.78485012e-01
2.48112142e-01 4.92773592e-01 -8.02933097e-01 2.44463488e-01
7.32483923e-01 1.17170942e+00 -8.58518839e-01 -2.76771247e-01
7.22490072e-01 7.54116476e-01 -8.83257627e-01 8.01315159e-03
-8.15987289e-01 1.75423965e-01 -6.38543844e-01 7.42121756e-01
1.29626572e+00 -9.13057998e-02 -4.85334337e-01 -2.80851185e-01
-2.50964731e-01 5.67597687e-01 -1.03586435e+00 2.06793165e+00
-2.48453960e-01 3.79689664e-01 2.56894350e-01 -6.29813790e-01
8.39996696e-01 -6.23044483e-02 6.34292424e-01 -4.23878133e-01
-6.20655119e-02 7.84452781e-02 -4.09493387e-01 5.02772741e-02
4.81543273e-01 -2.22658128e-01 1.51014611e-01 1.08762898e-01
3.82137075e-02 -9.99642432e-01 -4.99393553e-01 4.58504483e-02
1.04260623e+00 4.33690667e-01 -4.53553319e-01 -1.29032135e-01
3.41228879e-04 4.02643174e-01 2.57597059e-01 8.07730377e-01
3.86466086e-02 8.29865277e-01 2.29695737e-01 -5.80430031e-01
-1.24719441e+00 -1.83865547e+00 -3.05239886e-01 6.55573785e-01
3.89111400e-01 -5.73657453e-02 -7.33243704e-01 -5.91792047e-01
8.58065307e-01 7.12890327e-01 -1.78238407e-01 -4.07976620e-02
-7.87788749e-01 6.09985292e-02 5.50072551e-01 9.08826292e-01
5.55339813e-01 -7.47408807e-01 -7.84481823e-01 1.41244337e-01
4.18223292e-01 -1.31951976e+00 -2.15488635e-02 3.98100734e-01
-1.37269270e+00 -9.47707772e-01 -4.99116518e-02 -5.12699246e-01
5.73897779e-01 5.98589599e-01 1.35457599e+00 1.95912838e-01
-1.93941608e-01 3.61168146e-01 -4.17337343e-02 -7.82011807e-01
-2.03314438e-01 1.81700245e-01 -1.94603115e-01 -5.47037661e-01
6.77434862e-01 -7.76661456e-01 -6.36245906e-01 -5.07040136e-02
-8.38012576e-01 2.75361855e-02 3.71366948e-01 8.58074203e-02
1.07248664e+00 -1.57339767e-01 -1.22210160e-01 -7.62482941e-01
7.28586572e-04 -2.23967209e-01 -8.04048538e-01 -4.12916690e-01
-3.24385583e-01 1.35948509e-02 2.88946599e-01 -2.11724028e-01
-6.98900461e-01 4.48745072e-01 -4.52173620e-01 -1.05778241e+00
-6.09248042e-01 3.26982327e-02 -4.81468216e-02 -3.02769452e-01
4.99804467e-01 -2.33661845e-01 -1.91145748e-01 -8.26821625e-01
7.54115343e-01 3.83057356e-01 6.50781333e-01 -6.13048494e-01
1.31445765e+00 1.02487004e+00 1.49167359e-01 -6.61701918e-01
-8.55870306e-01 -2.81887919e-01 -1.11042130e+00 2.88428459e-02
1.21599102e+00 -1.09300160e+00 -8.10581326e-01 2.17085123e-01
-1.54287958e+00 -5.10251403e-01 -8.24231088e-01 4.47900780e-02
-5.68656325e-01 1.70082524e-02 -6.36159956e-01 -5.42763650e-01
-1.62571594e-01 -1.12291563e+00 1.59164691e+00 1.01419732e-01
2.10981090e-02 -4.94465500e-01 -2.92207658e-01 1.84561133e-01
2.96031758e-02 4.30469334e-01 9.92952824e-01 -2.21330389e-01
-1.49175835e+00 1.15316808e-02 -6.32558584e-01 3.37710381e-01
-2.83640295e-01 8.95613655e-02 -1.27820480e+00 1.30425617e-02
-1.58030495e-01 -1.06770478e-01 1.10072577e+00 3.82480472e-01
1.51415694e+00 1.43275172e-01 -5.85845530e-01 1.01344264e+00
1.38277638e+00 -2.45486647e-01 4.62205529e-01 -1.72487512e-01
1.02713203e+00 3.02621394e-01 2.00145692e-01 1.36108115e-01
4.59275514e-01 2.30314657e-01 1.08105540e+00 -2.45352715e-01
-4.30783510e-01 -6.72082961e-01 -9.06631351e-02 3.24717194e-01
-1.14873581e-01 -1.31144390e-01 -1.14548826e+00 4.77230310e-01
-1.63626778e+00 -7.29732037e-01 -3.14245939e-01 1.91321278e+00
5.05703211e-01 6.24401629e-01 -9.91662815e-02 -1.13049857e-01
3.49006265e-01 8.13649315e-03 -9.34392095e-01 -3.81900132e-01
1.06400244e-01 8.48868489e-01 9.14005041e-01 7.08161831e-01
-1.06083167e+00 1.35063565e+00 6.24061584e+00 4.11626935e-01
-1.27732539e+00 1.04848437e-01 1.78921670e-01 -2.30249003e-01
-5.39299726e-01 1.26410034e-02 -8.16638827e-01 1.30538493e-01
5.21360397e-01 2.54664153e-01 2.57949561e-01 9.69744742e-01
8.63869786e-02 2.31779918e-01 -1.43885052e+00 9.66220975e-01
-5.33287764e-01 -1.55049467e+00 4.23205942e-01 3.08759660e-01
4.42957431e-01 6.38527572e-01 9.38797146e-02 7.87957981e-02
6.16425455e-01 -1.32447064e+00 1.08666635e+00 4.67110336e-01
7.35618591e-01 -7.44772077e-01 1.06628112e-01 4.99505311e-01
-1.24523020e+00 1.72081411e-01 -7.44298875e-01 -2.73798823e-01
1.44189477e-01 6.50267601e-01 -9.63730335e-01 2.45234072e-01
9.15539324e-01 7.26206720e-01 -3.54014337e-01 1.01586843e+00
-2.76998878e-01 5.12034655e-01 -8.92717540e-01 3.39765280e-01
5.08735001e-01 -2.44537108e-02 7.57119596e-01 1.07473505e+00
2.38291711e-01 2.56145537e-01 4.22270447e-01 1.68361318e+00
-3.89374405e-01 -4.74790961e-01 -7.56513357e-01 1.08561926e-01
6.68747604e-01 1.05264938e+00 -8.83582294e-01 -2.99060971e-01
-4.13211763e-01 7.93354452e-01 4.28737581e-01 1.66031972e-01
-5.83864689e-01 -1.46230161e-01 1.21305740e+00 5.66103816e-01
6.75058842e-01 -8.36075187e-01 -9.33652163e-01 -8.10858727e-01
-1.92995109e-02 3.32349353e-02 -8.01831260e-02 -9.28952694e-01
-1.45665205e+00 4.95180786e-02 -5.03747091e-02 -8.52995157e-01
3.06770831e-01 -8.41597021e-01 -4.52016473e-01 1.15390170e+00
-1.81238401e+00 -1.44184315e+00 -4.79451299e-01 4.47233498e-01
5.32514513e-01 6.26656294e-01 5.04503012e-01 1.77607119e-01
2.19601601e-01 9.83421206e-02 -4.56523269e-01 1.35780752e-01
2.75702208e-01 -1.37041688e+00 1.11646700e+00 5.77156723e-01
9.24574062e-02 4.86077487e-01 3.16298246e-01 -8.62531424e-01
-1.38287866e+00 -1.38850069e+00 5.81108212e-01 -8.79364789e-01
3.48092556e-01 -6.25035524e-01 -1.13385618e+00 9.37843025e-01
-3.08619291e-01 2.60085046e-01 8.31535980e-02 -1.01241827e-01
-6.56937122e-01 1.60529260e-02 -1.45230591e+00 4.81282979e-01
1.65047169e+00 -6.23703957e-01 -7.64962375e-01 2.73764342e-01
1.19067860e+00 -8.07238519e-01 -9.32561398e-01 6.41390324e-01
3.49104762e-01 -8.71815443e-01 1.62118018e+00 -6.85299516e-01
4.85845059e-01 -3.96197051e-01 -2.85440624e-01 -9.91156757e-01
-4.25590575e-01 -8.69587734e-02 -1.50622845e-01 7.71796644e-01
8.09354335e-02 -3.04057509e-01 1.27505493e+00 6.89466476e-01
-6.31530821e-01 -5.60460031e-01 -1.02256167e+00 -6.21613741e-01
5.78597903e-01 -8.51377428e-01 1.01709187e+00 6.83115482e-01
-7.01045215e-01 5.37176542e-02 6.21958792e-01 6.46158755e-01
9.18420792e-01 3.27405185e-01 7.07871139e-01 -1.68035173e+00
8.11925158e-02 -6.60693347e-01 -5.43777764e-01 -1.55541480e+00
1.00867115e-01 -1.26289928e+00 1.28405884e-01 -1.68262458e+00
-2.13235453e-01 -9.07101095e-01 -6.13598488e-02 5.99264979e-01
1.21883653e-01 4.36040163e-01 4.19555902e-01 1.72652632e-01
-1.63685203e-01 2.74577707e-01 1.34162521e+00 -2.75051326e-01
-2.85402834e-01 -7.97223896e-02 -3.86380404e-01 9.43789661e-01
6.07448161e-01 -4.67885792e-01 -2.89491355e-01 -1.21808290e+00
6.59777522e-02 -3.11850816e-01 1.03079033e+00 -1.03678024e+00
2.04374060e-01 -2.34421864e-01 8.07550013e-01 -1.33171821e+00
6.46806598e-01 -1.00442255e+00 -1.33676767e-01 4.07156914e-01
-7.34890327e-02 -1.84600756e-01 4.23955560e-01 4.22832578e-01
3.05429071e-01 3.95984165e-02 8.46232653e-01 -4.92372483e-01
-6.84593558e-01 7.22927153e-01 4.03271019e-02 -2.06228256e-01
8.08769882e-01 -4.29952443e-01 -2.09005728e-01 2.86714375e-01
-6.91894889e-01 3.03402126e-01 9.34653223e-01 5.52175462e-01
9.46240246e-01 -1.10246277e+00 -4.55704689e-01 2.68900067e-01
-2.57705241e-01 1.12515688e+00 9.31457505e-02 1.09209605e-01
-8.95884037e-01 3.24590653e-01 -1.67931303e-01 -1.13802862e+00
-8.76971424e-01 4.25638825e-01 5.13668358e-01 3.38070422e-01
-1.07682920e+00 9.89788949e-01 1.44494221e-01 -8.01513672e-01
1.59837365e-01 -1.04304504e+00 4.65538502e-01 -2.99055368e-01
-1.67164300e-02 2.12088555e-01 2.29143828e-01 -2.97518998e-01
-4.80657220e-01 8.67873609e-01 2.20541403e-01 6.94077387e-02
1.51878893e+00 2.23957106e-01 4.34310175e-02 3.92461181e-01
9.88770306e-01 -1.99132994e-01 -1.73140764e+00 -2.33192116e-01
-2.01409236e-01 -6.04020417e-01 2.81002313e-01 -6.03907526e-01
-1.02015209e+00 1.29869235e+00 6.70383096e-01 1.01131849e-01
4.83431399e-01 3.05130839e-01 1.02179480e+00 3.97441715e-01
4.77977753e-01 -6.41562402e-01 -1.78276539e-01 6.55721903e-01
5.68710625e-01 -1.21290684e+00 1.49383498e-02 -9.05020833e-01
1.24314837e-01 1.14447439e+00 6.43672168e-01 -7.29617059e-01
8.24294508e-01 5.50579607e-01 -5.89355826e-02 -5.32488704e-01
-3.69357347e-01 -2.93491751e-01 2.30147228e-01 8.32624674e-01
1.13243483e-01 1.83150128e-01 4.97732759e-01 3.56049329e-01
-3.41829658e-01 3.88063043e-02 7.84539729e-02 1.00806725e+00
-9.99353409e-01 -8.54547858e-01 -3.88938516e-01 3.68345290e-01
1.71379402e-01 -1.47427926e-02 -4.94898528e-01 6.42304718e-01
4.35031742e-01 5.05171478e-01 7.02667177e-01 -2.69251972e-01
6.94427133e-01 1.87575698e-01 7.21309781e-01 -9.32326734e-01
-4.96083677e-01 -2.50290394e-01 -4.71337646e-01 -8.99878323e-01
-2.52703596e-02 -6.50186718e-01 -1.90094411e+00 -5.49450994e-01
1.91340029e-01 -5.01711547e-01 8.31350982e-01 7.88297355e-01
6.34519160e-01 5.84033847e-01 1.64544195e-01 -1.44790328e+00
-5.35926044e-01 -4.67206657e-01 -2.61715680e-01 4.36952949e-01
4.03452039e-01 -5.79667926e-01 -2.88330335e-02 9.51410085e-02] | [8.060582160949707, -3.4418880939483643] |
0abd70e2-399d-475a-bf8c-7df894578250 | models-genesis-generic-autodidactic-models | 1908.06912 | null | https://arxiv.org/abs/1908.06912v1 | https://arxiv.org/pdf/1908.06912v1.pdf | Models Genesis: Generic Autodidactic Models for 3D Medical Image Analysis | Transfer learning from natural image to medical image has established as one of the most practical paradigms in deep learning for medical image analysis. However, to fit this paradigm, 3D imaging tasks in the most prominent imaging modalities (e.g., CT and MRI) have to be reformulated and solved in 2D, losing rich 3D anatomical information and inevitably compromising the performance. To overcome this limitation, we have built a set of models, called Generic Autodidactic Models, nicknamed Models Genesis, because they are created ex nihilo (with no manual labeling), self-taught (learned by self-supervision), and generic (served as source models for generating application-specific target models). Our extensive experiments demonstrate that our Models Genesis significantly outperform learning from scratch in all five target 3D applications covering both segmentation and classification. More importantly, learning a model from scratch simply in 3D may not necessarily yield performance better than transfer learning from ImageNet in 2D, but our Models Genesis consistently top any 2D approaches including fine-tuning the models pre-trained from ImageNet as well as fine-tuning the 2D versions of our Models Genesis, confirming the importance of 3D anatomical information and significance of our Models Genesis for 3D medical imaging. This performance is attributed to our unified self-supervised learning framework, built on a simple yet powerful observation: the sophisticated yet recurrent anatomy in medical images can serve as strong supervision signals for deep models to learn common anatomical representation automatically via self-supervision. As open science, all pre-trained Models Genesis are available at https://github.com/MrGiovanni/ModelsGenesis. | ['Jianming Liang', 'Michael B. Gotway', 'Zongwei Zhou', 'Ruibin Feng', 'Nima Tajbakhsh', 'Vatsal Sodha', 'Md Mahfuzur Rahman Siddiquee'] | 2019-08-19 | null | null | null | null | ['pulmonary-embolism-detection', 'lung-nodule-detection', 'lung-nodule-segmentation', 'liver-segmentation'] | ['medical', 'medical', 'medical', 'medical'] | [ 2.44251683e-01 5.92648804e-01 -2.22161382e-01 -5.43655038e-01
-9.20903206e-01 -4.21108633e-01 3.80557179e-01 -1.09530970e-01
-4.99514729e-01 5.93593121e-01 1.54084742e-01 -3.42666179e-01
-8.57265741e-02 -6.16788089e-01 -8.59376371e-01 -7.03901470e-01
-2.77635425e-01 6.78588271e-01 2.33954847e-01 -1.74431562e-01
-2.89187878e-01 4.92377996e-01 -1.01762927e+00 2.80122101e-01
9.66169715e-01 8.79573047e-01 4.28670347e-01 5.21038294e-01
-1.50241447e-03 4.77839500e-01 -3.56601596e-01 -2.42350087e-01
3.19649071e-01 -4.64475900e-01 -1.00318420e+00 6.16687126e-02
4.41608906e-01 -4.28158104e-01 -3.02346438e-01 1.01243448e+00
5.96671939e-01 -3.33074450e-01 6.69497252e-01 -7.81694114e-01
-6.33038878e-01 6.40217960e-01 -5.78390896e-01 2.81011015e-01
-6.13059439e-02 1.76921368e-01 7.02658176e-01 -5.45164049e-01
8.21954668e-01 9.79247510e-01 7.46095121e-01 9.22290146e-01
-1.31350195e+00 -5.29591918e-01 -5.67755066e-02 -1.26783714e-01
-1.05359805e+00 -1.37241796e-01 6.73330605e-01 -5.42766690e-01
5.63038230e-01 1.84797511e-01 7.13161469e-01 1.35378337e+00
4.02103037e-01 9.78855133e-01 1.32606637e+00 -1.94029406e-01
-5.14935665e-02 7.73373321e-02 4.35200743e-02 8.70269060e-01
2.82468740e-02 1.11113653e-01 5.86359277e-02 5.40862838e-03
1.07920504e+00 1.05229236e-01 -3.74810725e-01 -5.37722826e-01
-1.30794370e+00 7.69709468e-01 9.46580946e-01 4.90371555e-01
-3.42684776e-01 2.69782636e-02 4.71826315e-01 1.24145545e-01
4.92697686e-01 4.69364583e-01 -6.08115494e-01 2.07369626e-01
-8.24584067e-01 1.24612013e-02 4.20573682e-01 6.54690504e-01
6.99892461e-01 -5.52676432e-02 -8.39360356e-02 8.54073226e-01
5.02117537e-02 2.71409810e-01 9.81607020e-01 -6.90614402e-01
1.12646654e-01 6.13277256e-01 -5.50426126e-01 -4.53648597e-01
-8.06193709e-01 -9.56483543e-01 -1.11712635e+00 1.13054797e-01
4.64387327e-01 -2.25976795e-01 -1.27296650e+00 1.80586958e+00
3.74359697e-01 4.31369059e-02 -6.48645014e-02 9.25928593e-01
1.00188184e+00 2.49849603e-01 1.36702165e-01 9.74464566e-02
1.27939749e+00 -9.85242605e-01 -3.68320704e-01 -2.42532268e-01
9.01129007e-01 -3.47924441e-01 1.22145808e+00 1.99397549e-01
-1.03822374e+00 -5.65644920e-01 -8.10444951e-01 2.20154244e-02
-1.99843958e-01 8.06979984e-02 8.16273332e-01 3.71887147e-01
-1.14431167e+00 7.15621293e-01 -1.12710178e+00 -2.68171012e-01
8.96867752e-01 4.15373772e-01 -6.22655630e-01 1.98131483e-02
-1.13705325e+00 1.05676270e+00 2.93263137e-01 2.10273247e-02
-1.27941763e+00 -9.61683095e-01 -8.46165061e-01 -2.02316821e-01
3.17081541e-01 -9.63072836e-01 1.13731408e+00 -9.48098898e-01
-1.30448067e+00 1.57232952e+00 2.68574953e-01 -6.76695585e-01
6.66688085e-01 -1.08667500e-01 -3.49602439e-02 4.48604435e-01
3.09449553e-01 9.05105770e-01 7.85564780e-01 -1.28946483e+00
-1.55816823e-01 -6.65224910e-01 1.16989270e-01 5.66404499e-02
-1.81619063e-01 -3.48899186e-01 -3.64715666e-01 -7.27220893e-01
1.05264224e-01 -9.52787876e-01 -5.47006667e-01 -3.97080975e-03
-5.63370764e-01 6.98302910e-02 4.73552287e-01 -7.02530801e-01
6.97840929e-01 -2.08821964e+00 1.91659629e-01 1.09800212e-01
5.20995319e-01 3.08769673e-01 -9.24399048e-02 -5.73064946e-02
-4.57605749e-01 1.66534632e-01 -5.44227660e-01 -1.90190449e-01
-3.28169674e-01 3.00568610e-01 1.17191583e-01 4.27013159e-01
1.73558608e-01 1.24408531e+00 -1.01363611e+00 -6.06229186e-01
2.40718141e-01 4.53443438e-01 -5.43384671e-01 1.76839426e-01
-4.75882962e-02 1.14119709e+00 -6.30779624e-01 3.71972263e-01
4.61344421e-01 -6.32719219e-01 8.90167430e-02 -4.14340585e-01
3.83628160e-01 2.25783810e-01 -4.90729958e-01 2.22949696e+00
-6.30394280e-01 -5.18688448e-02 9.51959491e-02 -1.42702281e+00
6.29901767e-01 4.69315112e-01 8.43208730e-01 -7.39200950e-01
1.44682452e-01 3.37142736e-01 2.85071641e-01 -6.05801582e-01
-3.46883565e-01 -4.30025458e-01 5.11915497e-02 4.67502922e-01
5.20077586e-01 -4.09613401e-01 -9.52284113e-02 1.59545660e-01
1.00672340e+00 2.19255656e-01 1.94124803e-01 -2.44555995e-01
4.65187490e-01 4.95382585e-02 4.55063194e-01 7.23232627e-01
-3.45559061e-01 7.29125321e-01 5.01619220e-01 -3.64542514e-01
-9.29697096e-01 -1.18891573e+00 -4.84395862e-01 6.33750260e-01
-3.65498960e-02 -1.80556253e-01 -9.78990018e-01 -1.12633657e+00
-1.84175909e-01 2.42704585e-01 -1.03772759e+00 -3.34527284e-01
-5.30852735e-01 -9.73581731e-01 6.45454645e-01 4.23891425e-01
4.37369227e-01 -9.86415505e-01 -6.17006719e-01 9.07588229e-02
2.65017115e-02 -1.05901015e+00 -3.94849479e-01 4.99603629e-01
-1.21260810e+00 -1.13390541e+00 -1.15643215e+00 -7.71060586e-01
8.99511755e-01 -7.65218679e-03 1.19106328e+00 1.97426841e-01
-5.74437141e-01 4.25212353e-01 -2.10529938e-01 -2.36570522e-01
-6.00953758e-01 3.85451823e-01 -1.55522287e-01 -9.03964192e-02
-1.94038168e-01 -8.55514526e-01 -7.57794261e-01 1.20743744e-01
-9.45418298e-01 2.87626415e-01 9.52932656e-01 1.09139371e+00
7.52703488e-01 -3.01392049e-01 7.16463983e-01 -1.36693048e+00
3.91098648e-01 -4.94291276e-01 -1.95503145e-01 1.80349931e-01
-3.92861009e-01 6.77565187e-02 5.05369663e-01 -2.00914636e-01
-8.39792132e-01 2.03020245e-01 -6.49328113e-01 -4.81437624e-01
-4.96086389e-01 5.19790351e-01 1.18273333e-01 -7.10875914e-02
8.28421533e-01 2.98933983e-01 3.88345838e-01 -5.60643375e-01
2.75029033e-01 2.04134017e-01 5.66658080e-01 -5.94627261e-01
7.67399371e-01 7.05878556e-01 -3.12856375e-03 -3.88394684e-01
-1.20082080e+00 -2.02778637e-01 -9.61647689e-01 -3.01887188e-02
1.16117549e+00 -8.33122134e-01 -2.80478269e-01 4.35140073e-01
-7.58049607e-01 -6.29744768e-01 -4.59951490e-01 4.77087259e-01
-5.93906581e-01 3.42722625e-01 -8.07115734e-01 -1.57433957e-01
-5.36590755e-01 -1.38645041e+00 1.17130435e+00 3.37097682e-02
-1.68917865e-01 -1.37942779e+00 4.55782413e-02 4.03564036e-01
4.24919367e-01 6.09934151e-01 1.09834063e+00 -7.07007587e-01
-1.52468622e-01 -1.45345613e-01 -1.49091557e-01 5.81548631e-01
4.30201024e-01 -4.43020880e-01 -1.04203212e+00 -2.33355194e-01
1.77266896e-01 -5.80169678e-01 8.51927817e-01 5.83304524e-01
1.52379358e+00 1.50285950e-02 -3.55565250e-01 9.24396574e-01
1.21854126e+00 -9.93125811e-02 3.31714302e-01 2.38687366e-01
6.94801986e-01 5.09295285e-01 1.46141022e-01 -1.54878683e-02
4.04995292e-01 6.35565996e-01 5.34458339e-01 -7.67607749e-01
-4.09794450e-01 -2.07282797e-01 5.39669134e-02 8.20235193e-01
-3.38279083e-03 4.72410202e-01 -1.07852662e+00 4.58271354e-01
-1.57454455e+00 -5.41927159e-01 -3.50223184e-02 2.00472832e+00
1.21009922e+00 2.21922815e-01 1.04688689e-01 -1.68015838e-01
4.00105417e-01 -6.64720461e-02 -7.64989376e-01 9.81158316e-02
-2.18263604e-02 5.90754747e-01 4.04931754e-01 3.43013972e-01
-1.15189099e+00 8.30276787e-01 6.21502876e+00 7.43386686e-01
-1.55745721e+00 4.81781632e-01 1.01632547e+00 8.26880410e-02
-2.40459651e-01 -2.53678530e-01 -3.98376644e-01 3.45130235e-01
6.76611483e-01 -3.38783190e-02 -1.10811414e-02 9.65956807e-01
1.61628515e-01 1.14138335e-01 -1.20019853e+00 8.78365040e-01
-5.20191528e-02 -1.34793580e+00 9.37565416e-02 9.21510309e-02
7.40291357e-01 3.37222844e-01 8.25519413e-02 3.63535643e-01
1.85125843e-01 -1.22634721e+00 3.43786508e-01 1.55041978e-01
9.65189755e-01 -2.60849237e-01 6.57954216e-01 3.99570167e-01
-6.39953315e-01 2.68010825e-01 -1.46973655e-01 3.42526078e-01
1.99551821e-01 7.35356152e-01 -8.94607902e-01 7.50884771e-01
7.50314713e-01 8.94096792e-01 -6.55899882e-01 9.62379396e-01
-3.44871014e-01 5.87257743e-01 -1.78151712e-01 5.98181069e-01
4.55691099e-01 4.37883064e-02 4.15753931e-01 1.04029572e+00
-1.54734561e-02 -1.49007887e-01 1.60351545e-01 1.03170264e+00
-3.55523638e-02 4.24896143e-02 -5.37446856e-01 3.39289516e-01
-1.62707269e-01 1.42836690e+00 -8.29304874e-01 -5.10224581e-01
-2.26923659e-01 9.39938724e-01 2.06056714e-01 2.41037175e-01
-9.28602040e-01 -1.48226414e-02 1.83761463e-01 2.81410128e-01
7.39282444e-02 -5.67494668e-02 -4.14927542e-01 -1.36172903e+00
-1.90310821e-01 -7.64490247e-01 4.97349590e-01 -5.68761706e-01
-1.50306368e+00 9.42911923e-01 6.88593909e-02 -1.20036352e+00
-1.62624732e-01 -7.64789999e-01 -5.92830181e-01 7.47661054e-01
-1.67307913e+00 -1.15670514e+00 -3.98059458e-01 8.62399876e-01
2.93696463e-01 9.38051194e-02 9.01262820e-01 4.98903543e-01
-4.58721995e-01 5.66537738e-01 -1.36640728e-01 3.85432303e-01
8.49990964e-01 -1.42929637e+00 2.07556352e-01 4.49504197e-01
2.10679412e-01 6.26045406e-01 2.34610811e-01 -4.36544895e-01
-1.05329466e+00 -1.04552865e+00 3.85403335e-01 -4.19543386e-01
6.65169716e-01 -3.44935805e-01 -1.02536750e+00 7.82345474e-01
1.66300964e-02 2.94794619e-01 7.68085122e-01 -1.40664559e-02
-1.29013538e-01 -4.40652780e-02 -1.09758306e+00 4.42960918e-01
1.11879361e+00 -3.72762769e-01 -7.61663854e-01 6.82581544e-01
6.93341792e-01 -7.48984456e-01 -1.23194909e+00 6.09779358e-01
2.56661117e-01 -8.83219361e-01 1.02816844e+00 -6.06785357e-01
6.12773120e-01 2.31795162e-02 3.13583434e-01 -1.38222063e+00
-6.04272746e-02 -3.25551867e-01 2.80817579e-02 6.41654611e-01
5.02712965e-01 -7.21450269e-01 8.65849793e-01 2.86677897e-01
-4.71924722e-01 -1.14772332e+00 -9.31816161e-01 -6.91852391e-01
5.58281064e-01 -2.66524941e-01 3.10324818e-01 1.22583401e+00
-1.50510207e-01 3.43635738e-01 -1.15674354e-01 -1.11499019e-01
7.37056911e-01 9.61572528e-02 6.06636703e-01 -1.09134042e+00
-5.09633064e-01 -5.42350471e-01 -2.78557271e-01 -1.15605664e+00
1.75032705e-01 -1.51166475e+00 -2.77453005e-01 -1.45773244e+00
3.33684772e-01 -8.05271149e-01 -3.41102183e-01 7.61690021e-01
-2.22444609e-01 4.45042580e-01 -2.08427440e-02 3.57289046e-01
-2.77028382e-01 3.97780508e-01 1.94449663e+00 -1.53222114e-01
7.25818351e-02 1.28236070e-01 -8.30037951e-01 7.26589620e-01
6.74644470e-01 -4.43315417e-01 -4.51053113e-01 -5.71901560e-01
-1.86921403e-01 1.46472126e-01 6.29365087e-01 -8.19798768e-01
-9.77492034e-02 2.65251040e-01 4.14996952e-01 -2.20847994e-01
1.10530198e-01 -7.55586982e-01 -1.23580672e-01 7.26066411e-01
-3.15209210e-01 -2.40613729e-01 2.99090743e-01 1.20406926e-01
-3.17929327e-01 -2.23234102e-01 9.49286997e-01 -6.29631698e-01
-4.93401110e-01 6.76166594e-01 -7.73209482e-02 2.27217317e-01
9.03138638e-01 -2.25989208e-01 -6.45548925e-02 -2.39481568e-01
-1.33723307e+00 2.14109749e-01 2.59043217e-01 2.30248168e-01
4.25215513e-01 -1.03397071e+00 -7.23276734e-01 1.48051396e-01
-4.36408557e-02 4.73273814e-01 5.53705812e-01 1.34977400e+00
-4.45751399e-01 4.37960654e-01 -5.40004909e-01 -1.03111064e+00
-6.95992291e-01 3.92085969e-01 7.62274086e-01 -7.13785768e-01
-8.91381621e-01 8.64399374e-01 7.45277107e-01 -9.06755984e-01
1.27365828e-01 -4.53699768e-01 -1.60806291e-02 -3.49020362e-01
2.17059255e-01 -4.32072282e-01 2.44904831e-01 -3.41188461e-01
-2.88335264e-01 5.82410038e-01 -3.14040542e-01 1.96482331e-01
1.60983133e+00 1.23744443e-01 2.61802729e-02 3.10723126e-01
1.19122064e+00 -3.45604926e-01 -1.18330109e+00 -2.45591357e-01
-1.80289317e-02 -2.50921398e-02 2.46945694e-02 -1.02814758e+00
-1.49026883e+00 1.14470708e+00 5.95291257e-01 -1.43822715e-01
1.08566940e+00 2.18847558e-01 8.54289770e-01 1.56834983e-04
3.64546478e-01 -5.57332575e-01 1.70399249e-01 2.75070637e-01
9.00308311e-01 -1.33694291e+00 -2.31299996e-01 -3.69891793e-01
-7.10788667e-01 9.21748161e-01 7.17182934e-01 -2.70655036e-01
7.84040630e-01 2.97465861e-01 2.86211580e-01 -3.81424546e-01
-4.53404278e-01 -1.76929086e-01 3.51198763e-01 7.26510048e-01
5.92500925e-01 -5.12443632e-02 -1.67194843e-01 7.89800167e-01
-2.12005749e-01 1.49433343e-02 2.63809383e-01 7.40433574e-01
5.11250645e-02 -1.26883113e+00 -8.20167437e-02 5.81723273e-01
-5.96896529e-01 -1.51940614e-01 -1.15492716e-01 8.82337034e-01
1.63499802e-01 1.28193676e-01 -2.19338819e-01 -1.48832768e-01
1.72666669e-01 -3.22739370e-02 6.38656497e-01 -9.72234130e-01
-7.88984478e-01 9.16769132e-02 -3.69363725e-01 -4.88425434e-01
-4.29625213e-01 -3.83550793e-01 -1.40772736e+00 2.02191100e-01
-9.64566991e-02 1.36879221e-01 5.87168336e-01 9.93685722e-01
3.29589039e-01 7.33435333e-01 4.90233302e-01 -9.16937470e-01
-5.67655146e-01 -9.69071448e-01 -3.84579688e-01 6.37723744e-01
2.50844508e-01 -6.80974364e-01 -1.99798226e-01 1.68382809e-01] | [14.636885643005371, -2.189363718032837] |
54d4abf7-6d32-435e-b44a-20f89b2263c0 | deep-learning-techniques-for-humor-detection | null | null | https://aclanthology.org/W19-1307 | https://aclanthology.org/W19-1307.pdf | Deep Learning Techniques for Humor Detection in Hindi-English Code-Mixed Tweets | We propose bilingual word embeddings based on word2vec and fastText models (CBOW and Skip-gram) to address the problem of Humor detection in Hindi-English code-mixed tweets in combination with deep learning architectures. We focus on deep learning approaches which are not widely used on code-mixed data and analyzed their performance by experimenting with three different neural network models. We propose convolution neural network (CNN) and bidirectional long-short term memory (biLSTM) (with and without Attention) models which take the generated bilingual embeddings as input. We make use of Twitter data to create bilingual word embeddings. All our proposed architectures outperform the state-of-the-art results, and Attention-based bidirectional LSTM model achieved an accuracy of 73.6{\%} which is an increment of more than 4{\%} compared to the current state-of-the-art results. | ['Suraj Tripathi', 'Radhika Mamidi', 'Koushik Reddy Sane', 'Sushmitha Reddy Sane'] | 2019-06-01 | null | null | null | ws-2019-6 | ['humor-detection'] | ['natural-language-processing'] | [-6.71017408e-01 -2.87008822e-01 -1.37820810e-01 -7.54044801e-02
-5.68721592e-01 -7.26318965e-03 6.93678796e-01 8.15507397e-02
-7.61819899e-01 6.13449752e-01 7.27787197e-01 -8.79657388e-01
5.04060268e-01 -9.00460720e-01 -4.64737356e-01 -1.33746132e-01
-1.38866846e-02 1.95197850e-01 -6.85308054e-02 -8.81647587e-01
4.20241743e-01 -6.63301200e-02 -9.18776155e-01 5.66338897e-01
9.10841882e-01 3.37560475e-01 -2.57404029e-01 7.49768972e-01
-5.04003465e-01 1.51643670e+00 -3.37635070e-01 -9.10187244e-01
-8.91879648e-02 -3.70736331e-01 -9.08005655e-01 -6.38200939e-01
1.15802214e-01 -4.14380580e-01 -1.04325783e+00 1.13184404e+00
5.81215262e-01 5.83013557e-02 4.04437780e-01 -9.62107718e-01
-1.59628975e+00 7.65972614e-01 -5.42091846e-01 4.79804993e-01
2.86808640e-01 -1.77774224e-02 1.15381753e+00 -1.47509122e+00
4.72647041e-01 8.47116888e-01 1.05111432e+00 4.06534940e-01
-9.06662583e-01 -1.00314486e+00 -3.18354756e-01 7.30126441e-01
-1.44343019e+00 -5.43112233e-02 7.68223166e-01 -8.55237722e-01
1.49764752e+00 -2.30696484e-01 5.18526912e-01 1.44417620e+00
5.97538888e-01 7.75423884e-01 9.67237234e-01 -4.80847269e-01
-4.19520497e-01 4.69141930e-01 4.40905958e-01 9.66605365e-01
3.71896382e-03 -1.50627807e-01 -5.22662938e-01 -1.45791754e-01
4.06788290e-01 1.69887975e-01 -4.72319908e-02 3.96203250e-01
-1.13795674e+00 1.55078030e+00 8.34466577e-01 7.58749485e-01
-1.67024761e-01 2.34343782e-01 7.32155263e-01 4.75016683e-01
5.10893226e-01 4.36740547e-01 -7.18389153e-02 -3.49850267e-01
-9.48173046e-01 2.79042900e-01 7.53925920e-01 1.08533251e+00
6.91212714e-01 4.29939717e-01 -2.15030178e-01 1.17064583e+00
2.10979655e-01 4.07937676e-01 1.06017077e+00 -1.72299534e-01
6.10428989e-01 5.82173645e-01 -1.49500087e-01 -1.45599866e+00
-4.27909106e-01 -6.80680811e-01 -7.65783072e-01 -1.85242131e-01
-7.96553865e-02 -2.00975791e-01 -5.17864823e-01 1.35879040e+00
-6.04463816e-01 1.17428444e-01 -5.49681075e-02 7.36591160e-01
1.34848011e+00 8.25587869e-01 -2.64234431e-02 2.36020833e-01
1.15700948e+00 -1.62201011e+00 -8.22681606e-01 -2.98713982e-01
1.02979743e+00 -1.01647437e+00 1.20592880e+00 4.59997542e-02
-8.39651942e-01 -6.28148139e-01 -1.29214907e+00 -5.08320212e-01
-8.25241268e-01 -1.98895872e-01 2.93305695e-01 6.84076369e-01
-1.04110301e+00 4.32219207e-01 -4.27212805e-01 -4.14266825e-01
2.46249110e-01 1.04763262e-01 -3.87138128e-01 4.87501100e-02
-1.52593184e+00 1.26430881e+00 2.43230276e-02 -1.87240392e-01
-1.05546629e+00 -4.17488813e-01 -1.04318035e+00 3.74174826e-02
-5.18942773e-01 -3.56808454e-01 1.24983573e+00 -8.26525331e-01
-1.28157651e+00 8.38243961e-01 -1.47146925e-01 -6.64710104e-01
2.12630689e-01 -2.49252483e-01 -8.20162237e-01 -3.13792139e-01
1.52004108e-01 3.28497857e-01 3.04331392e-01 -8.77928197e-01
-2.51895934e-01 1.27730042e-01 1.97695836e-01 -1.82349041e-01
-1.23358870e+00 3.24010670e-01 -1.38505772e-01 -6.23789012e-01
-4.73455995e-01 -9.77064073e-01 3.69076766e-02 -5.22695124e-01
-2.46022135e-01 -1.09390862e-01 7.09530234e-01 -1.18500388e+00
1.95901656e+00 -1.85962975e+00 -1.07971199e-01 -3.07670504e-01
3.60471487e-01 5.35034120e-01 -4.18154001e-01 8.98191690e-01
-1.30542323e-01 1.48656800e-01 -1.21883556e-01 -5.09653568e-01
5.32517135e-02 1.46214902e-01 -1.86121628e-01 5.94121218e-01
3.50142978e-02 1.05210412e+00 -9.70657825e-01 -2.30328068e-01
1.15662180e-01 5.98475814e-01 -8.11544478e-01 3.09877902e-01
3.93961787e-01 5.89146838e-03 6.24215566e-02 3.33526641e-01
6.32249057e-01 -1.26258180e-01 1.49400294e-01 3.19915533e-01
-2.55511194e-01 5.57614803e-01 -4.40451533e-01 1.64609218e+00
-1.11449611e+00 1.26322055e+00 -5.37158728e-01 -6.69333279e-01
1.13437688e+00 3.52129161e-01 8.16201568e-02 -1.04069912e+00
3.55473220e-01 5.00268102e-01 2.30109006e-01 -9.08536732e-01
8.78708363e-01 -4.46168721e-01 -9.03690830e-02 4.20773387e-01
4.57786471e-01 4.13707227e-01 -3.90347727e-02 2.26697430e-01
1.20663428e+00 -2.91494399e-01 1.22600652e-01 -5.93553543e-01
8.85631561e-01 -1.72266543e-01 1.15879431e-01 3.91627759e-01
-1.96941659e-01 5.39139628e-01 5.28513372e-01 -8.39325547e-01
-1.42886734e+00 -5.76314569e-01 -8.44345689e-02 1.37652540e+00
-2.49222845e-01 -6.37535512e-01 -5.95975518e-01 -5.52665055e-01
-1.44268155e-01 1.07193708e+00 -8.09167266e-01 -1.64366871e-01
-6.49276376e-01 -8.40063155e-01 8.40737164e-01 6.25330150e-01
4.16346371e-01 -1.10909450e+00 -3.65428746e-01 4.74857002e-01
-4.55024302e-01 -1.05506408e+00 -3.94671321e-01 1.59729913e-01
-3.94402832e-01 -9.06948447e-01 -6.20439351e-01 -9.14653838e-01
2.64506727e-01 2.49387309e-01 1.22193646e+00 3.31963420e-01
8.82488713e-02 -3.02246779e-01 -6.29850328e-01 -4.71432507e-02
-4.14718539e-01 3.74016047e-01 -1.84003934e-02 -1.70877576e-01
1.02803457e+00 -6.58335745e-01 -4.74299222e-01 -9.61875468e-02
-8.36083114e-01 6.87145144e-02 3.16429585e-01 1.29161203e+00
-5.74469745e-01 -6.90955460e-01 6.33375347e-01 -8.67373168e-01
1.24118638e+00 -1.08064783e+00 -3.81546989e-02 -1.72211662e-01
-6.99475467e-01 6.90971389e-02 7.98873305e-01 -2.45128348e-01
-7.12154746e-01 -6.45281971e-01 -6.22562766e-01 -2.72513717e-01
3.62450898e-01 8.70712042e-01 8.04666102e-01 -1.55488700e-01
8.96490574e-01 2.52140969e-01 -2.13186964e-01 -4.27236646e-01
4.03803438e-01 1.40141523e+00 2.14000374e-01 -2.17921481e-01
5.55804312e-01 1.61432579e-01 -5.39909422e-01 -5.61229825e-01
-6.56332076e-01 -5.60425520e-01 -7.55837142e-01 -5.23166284e-02
1.14177620e+00 -1.14476895e+00 -4.58630025e-01 3.54053080e-01
-1.54414582e+00 -1.71105862e-01 5.25906622e-01 4.47177231e-01
-1.91863552e-01 3.00124347e-01 -1.30388725e+00 -6.20181143e-01
-5.38694203e-01 -1.28429663e+00 3.60289395e-01 -1.26817152e-01
-2.63462096e-01 -1.38524246e+00 5.22438765e-01 5.18595815e-01
8.48201692e-01 1.47327140e-01 1.06299555e+00 -1.02113569e+00
4.75546382e-02 -3.79768878e-01 -5.45537651e-01 5.13815939e-01
-3.69830161e-01 -9.41174552e-02 -1.09444809e+00 -2.48090804e-01
-1.65303946e-01 -5.94414413e-01 9.03080165e-01 -1.27477497e-01
8.56663764e-01 -4.66228276e-01 9.90644097e-02 3.21071446e-01
1.65783525e+00 -2.83417348e-02 7.25064576e-01 7.55749106e-01
8.16578507e-01 1.30716607e-01 -2.26676852e-01 6.50962412e-01
8.22171926e-01 4.38625634e-01 5.09638608e-01 -4.57199290e-02
-8.30258578e-02 -3.95319492e-01 4.87140000e-01 1.77990556e+00
1.25447717e-02 1.21752471e-01 -1.36509871e+00 1.09501612e+00
-1.91805565e+00 -1.02938151e+00 -9.02690828e-01 1.86316419e+00
7.89945662e-01 1.44930243e-01 1.76105760e-02 -1.17648371e-01
4.52159554e-01 3.77781987e-01 2.03512177e-01 -1.21236241e+00
4.76916060e-02 4.92459059e-01 5.12219608e-01 8.44260335e-01
-9.15418684e-01 9.99340415e-01 6.11757088e+00 7.20776737e-01
-1.25428545e+00 8.47508252e-01 2.78968900e-01 6.57137707e-02
-5.47975898e-01 -1.59439191e-01 -3.83670628e-01 4.57261741e-01
1.45466948e+00 -1.02816008e-01 5.68512440e-01 9.35576081e-01
-6.94393143e-02 3.74022782e-01 -8.11217606e-01 1.06542742e+00
5.77628076e-01 -1.34286630e+00 -1.35117546e-01 3.88589688e-02
1.28089571e+00 6.48317814e-01 6.76956400e-02 9.59112227e-01
4.63043451e-01 -1.41832900e+00 6.45051777e-01 3.25083494e-01
6.57422543e-01 -9.19799387e-01 1.34090269e+00 2.91546434e-01
-8.63651514e-01 -4.39248055e-01 -5.53796709e-01 -6.92195833e-01
-4.34489036e-03 4.11515027e-01 -6.67929471e-01 3.57362628e-01
5.52934229e-01 9.98770356e-01 -7.73128331e-01 8.39103878e-01
-3.30220878e-01 5.73192000e-01 4.20432806e-01 -5.04309118e-01
9.79985118e-01 1.70841232e-01 1.47822332e-02 1.94461429e+00
4.29482490e-01 -3.11321974e-01 -4.74043526e-02 9.92664814e-01
-5.47524929e-01 3.19710225e-01 -7.19355524e-01 -3.73423360e-02
2.91290373e-01 1.32901800e+00 2.78451643e-03 -4.49441284e-01
-1.04393268e+00 1.07025027e+00 7.07648814e-01 2.49752313e-01
-1.11916578e+00 -9.15109277e-01 5.87705672e-01 2.28829272e-02
2.15028301e-01 -5.11060834e-01 -3.74948323e-01 -1.33059239e+00
-3.23183179e-01 -7.04867899e-01 1.72680363e-01 -6.11862600e-01
-1.46815062e+00 1.03444660e+00 -4.13556814e-01 -1.23992670e+00
-8.07575062e-02 -7.72910714e-01 -7.81028152e-01 8.61889243e-01
-1.70988369e+00 -1.35021675e+00 -3.42487335e-01 5.40132880e-01
6.45876586e-01 -4.66517240e-01 9.61296320e-01 9.93491948e-01
-4.89493459e-01 7.77113557e-01 5.17520666e-01 6.88941479e-01
6.44887507e-01 -9.17161584e-01 4.48090672e-01 6.88562810e-01
3.60952578e-02 8.33381176e-01 5.83631217e-01 -2.22738639e-01
-1.11693907e+00 -1.01720834e+00 1.78523648e+00 -6.54088736e-01
1.34647822e+00 -5.08308470e-01 -7.23678946e-01 7.64020860e-01
1.20372629e+00 -1.83456704e-01 1.00944638e+00 1.07786499e-01
-8.06226730e-01 2.03340784e-01 -7.97413349e-01 4.63912189e-01
7.24171042e-01 -9.56438839e-01 -7.26002872e-01 2.38829613e-01
7.22480774e-01 -2.78862119e-02 -7.27223456e-01 6.26206845e-02
4.23696607e-01 -1.23863685e+00 5.23903668e-01 -9.45734799e-01
1.44287479e+00 3.66946161e-02 -4.15056050e-01 -1.31730449e+00
-7.80387640e-01 -1.35510340e-01 8.87085348e-02 9.04769301e-01
5.30417800e-01 -4.35920477e-01 9.84574184e-02 1.57247767e-01
-2.54442006e-01 -8.38594854e-01 -9.51946318e-01 -4.99581248e-01
9.02137756e-01 -4.05389369e-01 2.40490213e-01 1.54359722e+00
5.82780361e-01 5.89039505e-01 -8.15954626e-01 -4.39040959e-01
1.31665766e-01 -1.54651329e-01 5.61830401e-01 -5.53023934e-01
3.40792984e-02 -6.89797759e-01 -5.03562868e-01 -8.44822764e-01
6.22682989e-01 -1.27247226e+00 -3.85736883e-01 -1.55948520e+00
7.20903456e-01 2.39655122e-01 -4.75462168e-01 5.16544759e-01
1.21482508e-02 7.16562092e-01 2.85367191e-01 1.58035811e-02
-4.09323722e-01 7.69122660e-01 7.08405197e-01 -3.98372531e-01
3.14011395e-01 -6.41252577e-01 -6.61126375e-01 4.64533776e-01
6.72011554e-01 -5.24632931e-01 7.33935237e-02 -8.38989615e-01
5.44688582e-01 -1.18956856e-01 2.78558671e-01 -1.00589716e+00
2.68068403e-01 3.23029488e-01 8.81803781e-02 -2.46003047e-01
-7.79311964e-03 -6.24519706e-01 -3.74682456e-01 7.71967709e-01
-4.67068374e-01 6.38690948e-01 2.50484675e-01 1.48639843e-01
-4.03768122e-01 -4.67207670e-01 7.39170671e-01 -3.04352880e-01
-4.80367005e-01 -8.22718218e-02 -8.44658375e-01 -6.29066303e-02
6.40666783e-01 2.46090621e-01 -5.73146760e-01 -6.10495925e-01
-3.57698560e-01 6.14651442e-02 1.03865944e-01 8.55932713e-01
6.43119097e-01 -1.88773441e+00 -1.02847993e+00 1.41737044e-01
4.51778024e-01 -1.06811094e+00 1.17303267e-01 1.07259536e+00
-8.90574098e-01 6.83493137e-01 -4.60005462e-01 7.08019659e-02
-9.59909856e-01 7.48253763e-01 2.09581748e-01 -3.14410388e-01
-2.78538495e-01 8.80571425e-01 -3.53214681e-01 -9.82703805e-01
-2.04185262e-01 3.77617106e-02 -4.30491954e-01 1.31907666e-06
4.28846598e-01 5.90456486e-01 1.76318705e-01 -9.64966953e-01
-3.05943310e-01 3.45318019e-01 2.10198052e-02 4.60029356e-02
1.47124219e+00 1.83420867e-01 -6.05206132e-01 5.98157346e-01
1.89333546e+00 5.83836474e-02 -1.52233258e-01 -3.25416058e-01
-5.85916750e-02 -6.58288598e-01 1.06022835e-01 -5.05950987e-01
-9.90382791e-01 1.51832807e+00 5.54793954e-01 1.64033934e-01
4.66593534e-01 -4.63498414e-01 1.25228584e+00 2.36336514e-01
2.44723588e-01 -1.13138247e+00 4.74085838e-01 1.19576490e+00
7.69769430e-01 -1.25333738e+00 -3.16929251e-01 5.40968955e-01
-6.70366883e-01 1.67361784e+00 8.49063694e-01 -5.10331035e-01
8.02831590e-01 5.77039942e-02 1.89865023e-01 -1.44707933e-01
-7.25929201e-01 -2.83445064e-02 2.61226207e-01 6.03532791e-03
1.21584821e+00 5.54594323e-02 -7.60716558e-01 7.54228115e-01
-3.15034539e-01 -5.81682138e-02 9.03037429e-01 5.69272459e-01
-5.57810426e-01 -9.31402028e-01 -1.00306615e-01 3.28015178e-01
-6.20299816e-01 -8.41282070e-01 -2.01848343e-01 6.08714700e-01
2.07791343e-01 1.03508711e+00 -5.84229231e-02 -1.18174374e+00
1.65747926e-01 2.97001302e-01 -1.25720829e-01 -5.29282749e-01
-1.03698039e+00 -3.45638931e-01 2.50238955e-01 -2.38417089e-01
-2.15721339e-01 -3.08589991e-02 -1.04184163e+00 -9.94060040e-01
-3.23533028e-01 3.27435881e-02 5.44918418e-01 8.42265487e-01
2.02026740e-01 5.78783393e-01 6.70668542e-01 -5.72381079e-01
-4.44464147e-01 -1.51389396e+00 -3.43434036e-01 6.20227814e-01
3.83035570e-01 -3.91253173e-01 -2.34404147e-01 -1.54457003e-01] | [8.920578956604004, 10.898447036743164] |
3f9719c4-a0d3-42ee-a848-5c70075506f8 | talecrafter-interactive-story-visualization | 2305.18247 | null | https://arxiv.org/abs/2305.18247v2 | https://arxiv.org/pdf/2305.18247v2.pdf | TaleCrafter: Interactive Story Visualization with Multiple Characters | Accurate Story visualization requires several necessary elements, such as identity consistency across frames, the alignment between plain text and visual content, and a reasonable layout of objects in images. Most previous works endeavor to meet these requirements by fitting a text-to-image (T2I) model on a set of videos in the same style and with the same characters, e.g., the FlintstonesSV dataset. However, the learned T2I models typically struggle to adapt to new characters, scenes, and styles, and often lack the flexibility to revise the layout of the synthesized images. This paper proposes a system for generic interactive story visualization, capable of handling multiple novel characters and supporting the editing of layout and local structure. It is developed by leveraging the prior knowledge of large language and T2I models, trained on massive corpora. The system comprises four interconnected components: story-to-prompt generation (S2P), text-to-layout generation (T2L), controllable text-to-image generation (C-T2I), and image-to-video animation (I2V). First, the S2P module converts concise story information into detailed prompts required for subsequent stages. Next, T2L generates diverse and reasonable layouts based on the prompts, offering users the ability to adjust and refine the layout to their preference. The core component, C-T2I, enables the creation of images guided by layouts, sketches, and actor-specific identifiers to maintain consistency and detail across visualizations. Finally, I2V enriches the visualization process by animating the generated images. Extensive experiments and a user study are conducted to validate the effectiveness and flexibility of interactive editing of the proposed system. | ['Yingqing He', 'Yujiu Yang', 'Ying Shan', 'Xintao Wang', 'Yong Zhang', 'Longyue Wang', 'Haoxin Chen', 'Menghan Xia', 'Xiaodong Cun', 'Youxin Pang', 'Yuan Gong'] | 2023-05-29 | null | null | null | null | ['story-visualization'] | ['computer-vision'] | [ 1.89905301e-01 -1.58247367e-01 2.02842221e-01 -3.12748432e-01
-2.58192897e-01 -9.23542917e-01 7.94679701e-01 -2.38247234e-02
1.39536679e-01 3.46243858e-01 2.12529331e-01 -2.38015264e-01
7.49032646e-02 -6.38403118e-01 -7.02138066e-01 -1.57839134e-01
1.56214118e-01 2.10338652e-01 5.09739041e-01 -2.30280697e-01
4.76428062e-01 5.10799050e-01 -1.62132668e+00 5.59580386e-01
9.29345608e-01 6.36904776e-01 5.16891360e-01 7.75926113e-01
-6.27102911e-01 6.41517520e-01 -6.38577104e-01 -2.89427131e-01
1.99682653e-01 -4.14671957e-01 -3.71527702e-01 3.47382158e-01
2.86212444e-01 -4.97517467e-01 -1.23723432e-01 8.67582023e-01
2.90603817e-01 1.40751332e-01 5.38179338e-01 -1.41113651e+00
-6.68114841e-01 6.00928426e-01 -7.31526315e-01 -2.11129040e-01
7.37500727e-01 5.05671740e-01 4.11996216e-01 -1.00592256e+00
1.12767315e+00 1.41219282e+00 2.60911673e-01 4.06821936e-01
-1.21824968e+00 -6.94036365e-01 3.68588239e-01 6.34944141e-02
-1.38537169e+00 -3.67382616e-01 9.36846852e-01 -7.87407815e-01
4.77951914e-01 4.97144252e-01 1.04231036e+00 1.01272476e+00
-1.48555681e-01 6.57199979e-01 9.47310209e-01 -5.40910363e-01
1.44343585e-01 3.62129420e-01 -1.57457069e-01 5.87612450e-01
-6.06880188e-02 -1.22122504e-01 -6.60384595e-01 1.02633297e-01
1.30197525e+00 -1.95248142e-01 -3.14582676e-01 -6.17749155e-01
-1.43568707e+00 3.44682515e-01 -1.96575634e-02 1.22651853e-01
-1.90873638e-01 -1.30382739e-02 4.75040376e-01 1.32996857e-01
2.25255445e-01 5.23329914e-01 -1.07731715e-01 -1.86959073e-01
-9.99058247e-01 4.63378817e-01 3.64221632e-01 1.47887659e+00
5.02599537e-01 1.60957426e-01 -5.20327270e-01 7.00378478e-01
1.79840848e-01 5.54703116e-01 1.91602618e-01 -9.96327579e-01
5.92195392e-01 6.62585497e-01 3.25523853e-01 -1.17029166e+00
-1.92613915e-01 -1.01044446e-01 -7.69905984e-01 4.68005061e-01
2.46768013e-01 -1.02586970e-01 -7.09492505e-01 1.69807744e+00
5.18793941e-01 -1.68106362e-01 -2.90981412e-01 7.21778512e-01
9.39019561e-01 8.98061574e-01 8.08617547e-02 -1.63142577e-01
1.27002716e+00 -1.01283967e+00 -8.49309325e-01 -3.56368273e-02
3.72769654e-01 -9.08081889e-01 1.62819326e+00 2.72010505e-01
-1.28774357e+00 -8.17920029e-01 -9.48265314e-01 -1.19450301e-01
-1.66307285e-01 3.98993284e-01 2.94281721e-01 3.08569551e-01
-9.86347020e-01 2.55431563e-01 -3.36614460e-01 -4.88362730e-01
1.08566649e-01 -3.68037000e-02 -3.28415841e-01 1.19857296e-01
-9.32164729e-01 4.69359577e-01 5.89339554e-01 8.03962275e-02
-7.09584355e-01 -8.61512005e-01 -7.61205435e-01 9.64843780e-02
3.83874565e-01 -7.47429729e-01 1.17552006e+00 -1.20479679e+00
-1.75478601e+00 6.29581809e-01 1.02939010e-01 1.57652706e-01
9.76452172e-01 -2.01347798e-01 -2.17193350e-01 2.95940876e-01
-1.11443345e-02 9.68709111e-01 9.53672469e-01 -1.53949165e+00
-5.29363990e-01 1.13644607e-01 1.93382099e-01 2.39989728e-01
-3.10292006e-01 1.61602467e-01 -1.06684983e+00 -1.06042755e+00
-2.31809482e-01 -7.70743966e-01 6.37597889e-02 3.68529290e-01
-6.36824310e-01 1.56851172e-01 1.23278379e+00 -7.24182963e-01
1.51530623e+00 -2.33564091e+00 2.35996306e-01 3.09090555e-01
5.73709272e-02 2.76872188e-01 -3.15207273e-01 5.98569572e-01
-2.00347736e-01 1.52545810e-01 1.32643759e-01 -5.20912051e-01
-1.00630417e-01 -9.96269584e-02 -1.59915835e-01 -1.49111867e-01
4.47680391e-02 7.68587947e-01 -9.35920238e-01 -8.10524702e-01
5.82180023e-01 2.94540137e-01 -5.76829135e-01 4.68775690e-01
-4.66183543e-01 7.03772604e-01 -3.33366245e-01 3.83592486e-01
5.80594838e-01 -2.66582072e-01 3.27015400e-01 -5.10922372e-01
-5.23518503e-01 -1.58191517e-01 -1.53459477e+00 1.66789269e+00
-3.29539627e-01 7.01882541e-01 4.90330951e-03 -3.93064529e-01
1.07012630e+00 1.88582674e-01 2.96848476e-01 -4.95267004e-01
4.86817434e-02 -1.34485334e-01 -3.90742242e-01 -8.38064551e-01
7.94581890e-01 5.25270104e-01 -5.61849922e-02 6.09701395e-01
-3.46330106e-01 -3.25222075e-01 6.37150943e-01 5.58729351e-01
4.07620549e-01 5.87312400e-01 1.70951307e-01 -7.21105412e-02
3.47639710e-01 -1.93624813e-02 1.78599805e-01 6.38271868e-01
4.41160500e-01 7.27390409e-01 5.32882929e-01 -4.07631099e-01
-1.46659636e+00 -8.43198657e-01 2.49881759e-01 9.62442935e-01
3.79610449e-01 -6.08047187e-01 -9.32650447e-01 -3.06413889e-01
-3.88109863e-01 9.59135771e-01 -6.08421803e-01 2.12252691e-01
-6.27718270e-01 -1.08619593e-01 1.06456578e-01 4.67301339e-01
5.45223534e-01 -1.24085534e+00 -1.16156995e+00 -8.13901518e-03
-2.29274794e-01 -8.44087243e-01 -9.18145716e-01 -3.87749285e-01
-6.63049281e-01 -9.49782908e-01 -7.23033547e-01 -7.87437558e-01
9.73582625e-01 4.92763162e-01 8.33963931e-01 1.94189399e-01
-1.91851959e-01 4.03355718e-01 -5.92485309e-01 -7.61843026e-02
-5.94084740e-01 -5.11063561e-02 -3.27487707e-01 7.63221309e-02
-7.32987046e-01 -6.30462885e-01 -5.76619446e-01 3.55487645e-01
-1.28747344e+00 1.23085999e+00 3.05725127e-01 5.89602232e-01
5.79886258e-01 -9.05254930e-02 6.67317808e-02 -1.04046142e+00
6.01539850e-01 -8.66873339e-02 -5.57657003e-01 5.81342459e-01
-3.56670439e-01 2.27688882e-03 7.91945159e-01 -8.36387575e-01
-1.35685480e+00 6.90651014e-02 1.80398852e-01 -6.29442692e-01
-6.48154467e-02 3.69707584e-01 -4.64879215e-01 1.69060066e-01
2.98105896e-01 9.64675322e-02 -2.05596328e-01 -4.17089969e-01
8.09782803e-01 4.74413961e-01 7.70207167e-01 -6.27602875e-01
1.07912147e+00 7.02597871e-02 -4.49779779e-01 -6.16675019e-01
-2.45473057e-01 1.49028346e-01 -9.29103732e-01 -6.78889215e-01
7.40687251e-01 -5.12323618e-01 -4.18186814e-01 6.20539308e-01
-1.31369233e+00 -6.48995936e-01 -3.90826464e-01 7.23635852e-02
-5.01560986e-01 3.12875003e-01 -3.68887782e-01 -5.99130154e-01
-3.27677459e-01 -1.26685238e+00 9.76490557e-01 4.83254373e-01
-4.38801825e-01 -7.15232313e-01 -1.94276184e-01 9.83000472e-02
3.94367933e-01 5.33499122e-01 1.29027390e+00 -3.93361077e-02
-7.73193777e-01 -7.41179511e-02 -3.83995265e-01 -1.70981988e-01
2.38940954e-01 7.35932648e-01 -5.41032016e-01 -1.22492880e-01
-5.99983096e-01 -1.59062684e-01 1.48192167e-01 3.41850400e-01
1.36552346e+00 -4.92632955e-01 -2.75196105e-01 8.09609771e-01
1.08010972e+00 6.42835379e-01 7.97469139e-01 5.69056392e-01
8.93582463e-01 5.12755334e-01 7.89474905e-01 5.88277221e-01
4.65811074e-01 9.15324450e-01 2.38508433e-01 -3.93084466e-01
-2.73733884e-01 -7.88504004e-01 1.69673428e-01 7.65943050e-01
-5.20364121e-02 -2.98661560e-01 -5.58158338e-01 2.92696446e-01
-1.78589511e+00 -1.02450430e+00 -1.54852137e-01 2.13516808e+00
7.93399394e-01 1.30038679e-01 1.18801534e-01 -9.05336160e-03
8.47852826e-01 2.12152869e-01 -6.70025408e-01 -3.76189083e-01
-5.08617833e-02 -3.43653321e-01 3.52668166e-02 4.46467996e-01
-7.29544342e-01 1.03803253e+00 5.94253635e+00 8.53029370e-01
-1.33560014e+00 -1.03741899e-01 7.77939856e-01 -2.11390197e-01
-7.68035054e-01 1.85772136e-01 -4.27364767e-01 4.30021524e-01
3.61913741e-02 -2.43194625e-01 5.37602246e-01 6.67469800e-01
6.85741186e-01 -6.60341978e-02 -9.30964291e-01 1.11901021e+00
-5.44264652e-02 -1.67203665e+00 5.57582915e-01 -5.16566336e-01
7.02731848e-01 -8.41428459e-01 1.47321224e-01 1.05743602e-01
6.43325299e-02 -6.72256351e-01 1.52572274e+00 6.81023121e-01
1.53528273e+00 -4.97429788e-01 5.18077649e-02 2.51743466e-01
-1.42168200e+00 7.40897730e-02 5.70931882e-02 2.27635771e-01
6.08782530e-01 2.55645476e-02 -5.98255813e-01 4.90536302e-01
6.71317995e-01 5.76337278e-01 -7.50763953e-01 7.87843525e-01
-3.62870455e-01 1.65997818e-01 -8.46486166e-02 -6.57740682e-02
-1.54375091e-01 -4.04637694e-01 5.54274499e-01 1.39840710e+00
3.92675221e-01 2.48338372e-01 1.96918800e-01 9.40184176e-01
1.69685647e-01 4.03597713e-01 -4.14967805e-01 -1.16182446e-01
7.24241376e-01 1.32640207e+00 -8.27432454e-01 -5.74127316e-01
-1.10949747e-01 9.57032263e-01 9.30317715e-02 5.72942376e-01
-8.71051967e-01 -3.79340231e-01 1.49014771e-01 5.90543032e-01
1.80827752e-01 -3.82419705e-01 -2.91181415e-01 -1.03457415e+00
-6.34251609e-02 -1.10542941e+00 5.06217629e-02 -1.42686343e+00
-6.96120977e-01 8.21815729e-01 5.04732311e-01 -1.30655646e+00
-1.44930169e-01 -1.68358997e-01 -8.74368608e-01 6.35569572e-01
-9.46307480e-01 -1.46710360e+00 -6.92796528e-01 5.95889866e-01
9.27893341e-01 4.96447682e-02 3.82625908e-01 2.68604368e-01
-6.67337239e-01 6.00809216e-01 -6.08349256e-02 -1.01690374e-01
7.11903393e-01 -9.32681441e-01 5.52278817e-01 8.47001255e-01
5.25862426e-02 5.43481290e-01 7.67066836e-01 -8.38675201e-01
-1.32871544e+00 -1.11642957e+00 4.48850065e-01 1.56041458e-02
4.04848784e-01 -5.67595601e-01 -8.24214578e-01 5.17540038e-01
2.75944918e-01 -4.15705323e-01 2.63163924e-01 -4.22521621e-01
-2.18626499e-01 -4.67176326e-02 -8.27702880e-01 1.02282131e+00
1.14243400e+00 -2.94677109e-01 -1.53273091e-01 2.33845249e-01
7.15444565e-01 -7.43773043e-01 -5.08967578e-01 6.84013143e-02
7.96029389e-01 -1.01386940e+00 9.98283327e-01 -3.47329944e-01
6.26104951e-01 -6.68997884e-01 1.45391196e-01 -1.08766747e+00
-4.50633168e-01 -9.46899176e-01 2.10703760e-01 1.59182346e+00
2.53313303e-01 -6.72827065e-02 2.23976642e-01 7.71372378e-01
-1.31506965e-01 -4.96374875e-01 -4.47884679e-01 -4.07779217e-01
-4.61759686e-01 -3.47761869e-01 8.78249705e-01 8.50045800e-01
8.02264810e-02 1.53936595e-01 -6.78960085e-01 1.00164019e-01
2.55630702e-01 3.65432918e-01 1.39917243e+00 -8.62163842e-01
-3.04317981e-01 -5.21502674e-01 1.22478180e-01 -1.18310261e+00
-2.83379018e-01 -4.96844828e-01 -1.39827833e-01 -1.71624351e+00
3.26400727e-01 -5.48388541e-01 4.45989341e-01 3.47023398e-01
-2.94217110e-01 3.89188863e-02 7.73937702e-01 4.61593211e-01
-5.63822508e-01 5.24606645e-01 1.58556461e+00 -1.11978613e-02
-7.18194783e-01 -3.15508276e-01 -3.20146829e-01 7.91447461e-01
6.01505756e-01 -6.62839487e-02 -7.66020358e-01 -5.69370210e-01
1.88651726e-01 2.92968184e-01 3.41757745e-01 -7.81948149e-01
1.01653934e-01 -5.22497952e-01 4.95695859e-01 -8.70094955e-01
2.85719961e-01 -5.49678922e-01 8.07879925e-01 2.16592968e-01
-4.39321250e-01 4.78534639e-01 2.84643084e-01 3.02587390e-01
3.49342711e-02 -1.54085293e-01 6.22575223e-01 -9.28303823e-02
-7.22422600e-01 2.08413616e-01 -3.81546617e-01 -1.03604831e-01
1.10226321e+00 -5.77805340e-01 -3.41777921e-01 -6.45115256e-01
-5.69233835e-01 3.02146018e-01 8.99076581e-01 6.11996830e-01
7.83926487e-01 -1.44107223e+00 -4.49828923e-01 3.73851448e-01
1.14591695e-01 1.07399456e-01 5.44404089e-01 4.21799630e-01
-8.47517550e-01 5.37058637e-02 -3.58552545e-01 -6.34010375e-01
-1.43249583e+00 7.37038016e-01 4.46596295e-02 -1.37610883e-01
-9.66032207e-01 5.70499897e-01 6.91900611e-01 -5.46343625e-02
1.87977448e-01 -3.83332968e-01 -3.17245871e-01 3.47667821e-02
6.49020135e-01 2.30535120e-01 -5.38828433e-01 -5.26015699e-01
6.28915206e-02 7.29241848e-01 -6.57712156e-03 -4.00627255e-01
1.19220710e+00 -3.78537536e-01 2.30988100e-01 4.32927549e-01
5.48576355e-01 1.27600804e-01 -1.66923141e+00 2.12946013e-02
-3.39664936e-01 -7.54320562e-01 -4.57165003e-01 -7.76942909e-01
-1.02936888e+00 7.91211426e-01 3.03929895e-01 5.24725020e-02
1.20269334e+00 -3.03002864e-01 6.99020684e-01 -3.73196006e-02
3.86311531e-01 -9.59554851e-01 4.09170389e-01 1.60505638e-01
1.28812909e+00 -6.22900188e-01 -9.90413278e-02 -5.22474110e-01
-9.57974911e-01 1.28426075e+00 8.27294052e-01 2.60868609e-01
2.81753838e-01 3.37918103e-01 1.67897612e-01 -1.18427493e-01
-6.94762766e-01 3.22859198e-01 3.24771315e-01 6.32348895e-01
1.46011740e-01 -3.80752459e-02 -1.63448229e-01 4.07730967e-01
-1.52322203e-01 4.46386002e-02 5.36195278e-01 6.36645079e-01
-2.31736138e-01 -1.08561385e+00 -5.56106567e-01 1.66307941e-01
3.05310398e-01 5.65624647e-02 -2.82922208e-01 9.04672980e-01
3.32053751e-01 7.63633847e-01 -3.05712111e-02 -3.15619648e-01
4.84485358e-01 -2.52403498e-01 3.63716692e-01 -5.76770961e-01
-6.13655269e-01 3.22337180e-01 -6.82294965e-02 -4.46045101e-01
-2.39029616e-01 -6.10277593e-01 -1.07252073e+00 -2.87702590e-01
-1.79783896e-01 -2.97493730e-02 5.48970759e-01 6.21248722e-01
4.57345635e-01 6.13911867e-01 5.61572492e-01 -1.19033110e+00
3.11118700e-02 -9.52043593e-01 -3.48110169e-01 6.91708386e-01
-5.75749502e-02 -6.20011628e-01 1.59357578e-01 6.77072346e-01] | [11.234031677246094, 0.0020187068730592728] |
794d4877-51b0-449f-80ce-715486efe9f5 | a-specifically-designed-machine-learning | 2006.09067 | null | https://arxiv.org/abs/2006.09067v1 | https://arxiv.org/pdf/2006.09067v1.pdf | A specifically designed machine learning algorithm for GNSS position time series prediction and its applications in outlier and anomaly detection and earthquake prediction | We present a simple yet efficient supervised machine learning algorithm that is designed for the GNSS position time series prediction. This algorithm has four steps. First, the mean value of the time series is subtracted from it. Second, the trends in the time series are removed. Third, wavelets are used to separate the high and low frequencies. And fourth, a number of frequencies are derived and used for finding the weights between the hidden and the output layers, using the product of the identity and sine and cosine functions. The role of the observation precision is taken into account in this algorithm. A large-scale study of three thousand position times series of GNSS stations across the globe is presented. Seventeen different machine learning algorithms are examined. The accuracy levels of these algorithms are checked against the rigorous statistical method of Theta. It is shown that the most accurate machine learning algorithm is the method we present, in addition to being faster. Two applications of the algorithm are presented. In the first application, it is shown that the outliers and anomalies in a time series can be detected and removed by the proposed algorithm. In a large scale study, ten other methods of time series outlier detection are compared with the proposed algorithm. The study reveals that the proposed algorithm is approximately 3.22 percent more accurate in detecting outliers. In the second application, the suitability of the algorithm for earthquake prediction is investigated. A case study is presented for the Tohoku 2011 earthquake. It is shown that this earthquake could have been predicted approximately 2 hours before its happening, solely based on each of the 845 GEONET station time series. Comparison with four different studies show the improvement in prediction of the time of the earthquake. | ['M. Kiani'] | 2020-06-16 | null | null | null | null | ['earthquake-prediction'] | ['computer-vision'] | [-2.04227850e-01 -2.28918165e-01 2.33320400e-01 -6.76762983e-02
-4.20918077e-01 -1.99474975e-01 2.77950972e-01 3.79001617e-01
-4.47536081e-01 5.92321813e-01 -5.55760190e-02 -4.68629122e-01
-3.77641261e-01 -7.36138701e-01 -3.98915082e-01 -1.01581097e+00
-6.92797065e-01 5.12419567e-02 2.17619449e-01 -4.83328223e-01
4.08390582e-01 6.62830353e-01 -1.47505653e+00 -6.33718446e-02
6.15625381e-01 1.10934258e+00 -2.91741610e-01 6.34281337e-01
3.42132837e-01 5.24263024e-01 -8.99669588e-01 2.86901504e-01
5.04975498e-01 -4.61378455e-01 -3.00437272e-01 -3.77230465e-01
-2.17386469e-01 -1.82015777e-01 -2.36647651e-01 6.80462122e-01
6.47372007e-01 2.47520700e-01 6.33817971e-01 -1.34999418e+00
1.37616575e-01 3.72414827e-01 -4.51510161e-01 5.46941996e-01
2.22182274e-01 -2.93396741e-01 5.62787473e-01 -7.75745451e-01
8.25968236e-02 5.16543925e-01 1.30499864e+00 -1.90304041e-01
-1.04045939e+00 -7.01065898e-01 -4.41533208e-01 1.36958018e-01
-1.46153331e+00 1.29792439e-02 9.68759775e-01 -4.82311219e-01
1.25197399e+00 4.29145098e-01 7.26497233e-01 3.24987888e-01
6.84671581e-01 5.78799136e-02 8.30460370e-01 -6.69841766e-01
2.89337844e-01 -3.52107167e-01 2.48310626e-01 4.43179905e-01
6.87914610e-01 2.31614381e-01 -4.64689434e-02 -7.10471392e-01
5.39542913e-01 2.78141826e-01 -8.06121901e-02 3.59765410e-01
-9.17946815e-01 8.35058272e-01 1.39253423e-01 8.90625119e-01
-6.82962358e-01 -1.91708971e-02 5.43019891e-01 7.64793694e-01
7.48479009e-01 4.02262837e-01 -7.66906619e-01 -5.55832461e-02
-1.41060781e+00 4.33767512e-02 9.54674244e-01 3.98598194e-01
5.37605643e-01 7.96151102e-01 4.99401778e-01 3.32613498e-01
3.04509550e-01 5.85707784e-01 8.54082167e-01 -4.64608252e-01
1.96061000e-01 3.44519407e-01 8.31976309e-02 -1.51935780e+00
-7.93351054e-01 -4.97620672e-01 -9.29996848e-01 3.52657109e-01
5.30986726e-01 -4.62120980e-01 -9.12760317e-01 1.08776009e+00
2.97681708e-02 5.33266068e-01 3.84207405e-02 3.49833399e-01
3.96777004e-01 9.23788309e-01 -1.52870625e-01 -5.37172258e-01
7.98202753e-01 -2.71238059e-01 -7.44330704e-01 1.41412258e-01
8.76237452e-01 -8.02845597e-01 2.19654202e-01 4.48055357e-01
-7.29285896e-01 -5.08664191e-01 -1.04482663e+00 5.87390304e-01
-5.64779282e-01 1.28864095e-01 3.42619449e-01 5.34018219e-01
-6.70037448e-01 1.09509635e+00 -1.17912245e+00 -3.78538966e-01
-4.05011415e-01 4.43483979e-01 -1.85792759e-01 6.72576964e-01
-1.43364370e+00 1.00380659e+00 4.96204942e-01 3.42475176e-01
8.27087164e-02 -3.94567549e-01 -7.32233286e-01 1.60361707e-01
-5.53834736e-01 -1.62765816e-01 8.22047651e-01 -1.15757418e+00
-1.04277980e+00 4.08431679e-01 -8.58905464e-02 -7.01343417e-01
4.55951631e-01 2.90363487e-02 -1.15935242e+00 1.30240679e-01
8.07485357e-02 -5.91585398e-01 8.95389438e-01 -7.77341783e-01
-8.08428645e-01 -1.22942857e-01 -8.66278291e-01 -3.81851614e-01
4.43452336e-02 -2.21723225e-02 1.77042857e-01 -9.62527871e-01
7.67619312e-01 -8.82517278e-01 -3.03870529e-01 -8.55388522e-01
-2.41214279e-02 7.73259019e-03 8.39750588e-01 -1.11514258e+00
1.75283408e+00 -2.37743640e+00 -6.30496562e-01 1.17553687e+00
-1.99663475e-01 -1.22645916e-03 4.78426874e-01 9.92578328e-01
-7.29225039e-01 1.67655740e-02 -2.05515429e-01 9.38784331e-02
-2.64201880e-01 2.52316087e-01 -6.77325428e-01 9.63871717e-01
-1.39105886e-01 1.67750582e-01 -5.85651040e-01 3.13460864e-02
3.01807523e-01 1.36466980e-01 -2.29953364e-01 -1.65297851e-01
5.64118505e-01 2.97865629e-01 -1.81024164e-01 4.52279776e-01
5.32918513e-01 2.26993501e-01 -2.09421396e-01 -1.09513573e-01
-4.23450232e-01 3.24856192e-02 -1.26847100e+00 6.99876130e-01
-2.63376143e-02 7.09231853e-01 -6.97818458e-01 -1.17329216e+00
1.36206710e+00 8.25685024e-01 8.63733470e-01 -6.59914672e-01
1.43253431e-01 6.88761711e-01 1.24945529e-01 -5.89989543e-01
3.19526851e-01 -8.52123946e-02 -9.35662389e-02 3.86456817e-01
-1.71168029e-01 -3.54703218e-02 2.45351315e-01 -2.46324241e-01
9.90431070e-01 -1.43778145e-01 7.59344041e-01 -3.06345373e-01
3.29959273e-01 1.46782249e-01 6.22037888e-01 6.38557494e-01
-8.90929475e-02 5.34736097e-01 5.02970934e-01 -8.77728581e-01
-1.12611604e+00 -7.68412530e-01 -2.46049501e-02 5.71976304e-01
-3.61833602e-01 -1.60838053e-01 -3.20556372e-01 -1.19452976e-01
2.59475857e-01 7.93412685e-01 -5.05769074e-01 -1.07300349e-01
-7.68893838e-01 -1.12069964e+00 7.18406677e-01 3.08161855e-01
2.18532488e-01 -8.66289198e-01 -7.08543301e-01 4.06405658e-01
-2.06000105e-01 -6.01096332e-01 2.27401659e-01 3.58304709e-01
-1.53854084e+00 -9.67017651e-01 -4.03713912e-01 -7.38941014e-01
7.68043041e-01 1.34627253e-01 7.76301801e-01 3.91338676e-01
7.65625667e-03 1.51961848e-01 -3.78104687e-01 -6.72495067e-01
-3.47146451e-01 -1.28727973e-01 3.09226513e-01 -7.22032413e-02
4.91588026e-01 -8.79835367e-01 -3.78177553e-01 2.14943275e-01
-7.17484534e-01 -6.12086475e-01 4.31220889e-01 7.03578055e-01
3.50387156e-01 6.30657315e-01 8.62789512e-01 -5.05910754e-01
4.47821766e-01 -9.15162742e-01 -7.08660424e-01 -2.81752735e-01
-7.12609649e-01 -2.67185390e-01 8.16370368e-01 -2.82530248e-01
-5.82256854e-01 1.23255663e-01 -1.60268724e-01 5.82816720e-04
-1.22437634e-01 8.52401257e-01 5.37707210e-01 -1.42777026e-01
8.10984075e-01 2.65078634e-01 -6.44096360e-02 -4.59894955e-01
-4.13168162e-01 5.72216749e-01 5.74509144e-01 -9.28199813e-02
1.04372942e+00 6.11412823e-01 2.10424047e-02 -1.34176064e+00
-2.37922892e-01 -8.23609591e-01 -6.64608479e-01 -3.69786024e-01
3.14007640e-01 -7.92212427e-01 -3.35844487e-01 6.33366644e-01
-9.05177295e-01 1.79388091e-01 -2.37319499e-01 9.98773873e-01
-4.69592363e-01 2.84154713e-01 -4.45165902e-01 -1.07663023e+00
-3.15034688e-01 -3.47437680e-01 3.82455289e-01 9.38304588e-02
-5.45096755e-01 -1.12154377e+00 3.93999130e-01 -4.21788514e-01
4.21655029e-01 8.25244009e-01 7.90220380e-01 -1.30891383e+00
8.98621082e-02 -8.82539034e-01 3.09894174e-01 3.01605731e-01
3.36519271e-01 4.01743948e-01 -8.69805217e-01 -5.50940394e-01
5.70021927e-01 6.40385866e-01 5.62576711e-01 5.73686957e-01
5.58005273e-01 -3.19927663e-01 -1.77800849e-01 5.89010060e-01
1.68383467e+00 7.09350109e-01 7.54981518e-01 8.26915860e-01
2.44336411e-01 3.19737613e-01 5.13733029e-01 5.93241692e-01
-1.47071749e-01 1.23884425e-01 2.94830739e-01 -3.17209154e-01
6.22928143e-01 1.42511263e-01 4.14004534e-01 1.31426764e+00
-6.95471525e-01 1.24063894e-01 -1.15408683e+00 6.65690303e-01
-1.81919813e+00 -1.43472874e+00 -7.33494878e-01 2.36681485e+00
2.34296396e-01 3.43411416e-01 1.44406289e-01 9.77874458e-01
6.17737770e-01 5.82406260e-02 3.89184058e-02 -6.34868145e-01
-2.37610750e-02 2.97092706e-01 8.82822096e-01 3.30497593e-01
-1.34240580e+00 1.99636877e-01 6.77315807e+00 3.35365415e-01
-1.38865745e+00 -1.90466732e-01 2.25318521e-01 2.67178178e-01
2.74337709e-01 -9.65695232e-02 -2.46110544e-01 7.63050318e-01
1.33874345e+00 -3.55414599e-01 -7.85709172e-02 8.68775308e-01
6.96356177e-01 -3.04585099e-01 -5.00866771e-01 7.91347325e-01
-3.90987918e-02 -8.93418610e-01 -3.06448907e-01 -1.33027092e-01
7.96622515e-01 1.21191800e-01 -3.66111159e-01 1.05721019e-01
-3.52633744e-01 -6.69711709e-01 6.28474474e-01 8.64271581e-01
6.38862997e-02 -1.15979207e+00 1.34991515e+00 1.74535736e-01
-1.15504706e+00 -2.59607553e-01 -2.09085047e-01 -6.22477710e-01
1.24363452e-01 9.12233412e-01 -6.60417318e-01 8.11982214e-01
8.86149347e-01 3.95015359e-01 -2.97098666e-01 1.58093536e+00
8.42252001e-02 1.11627877e+00 -7.34949291e-01 2.37582833e-01
2.35871717e-01 -2.12363809e-01 5.88071942e-01 1.30244207e+00
9.69249070e-01 1.55790791e-01 -1.35650516e-01 -6.75751828e-03
7.28464842e-01 1.82836205e-01 -8.01556408e-01 2.51735657e-01
3.91982079e-01 6.74267232e-01 -7.94652045e-01 -3.03367078e-01
-5.56608200e-01 5.31523228e-01 -5.19985676e-01 6.33586109e-01
-8.18139791e-01 -9.13901687e-01 2.33937174e-01 1.75416350e-01
2.31260508e-01 -4.16341305e-01 -5.04937172e-01 -7.94053078e-01
3.98966391e-03 -6.25771701e-01 6.81873500e-01 -5.55637360e-01
-1.10962427e+00 5.48779905e-01 7.71508589e-02 -1.87595522e+00
-6.37895286e-01 -4.21819538e-01 -1.02546358e+00 9.91702855e-01
-1.02384889e+00 -4.17209983e-01 3.52479368e-02 5.25007784e-01
6.30775020e-02 -3.98339033e-01 9.30487037e-01 5.28155744e-01
-4.25124407e-01 1.38872012e-01 7.31807292e-01 3.37826371e-01
4.72153217e-01 -9.94015098e-01 4.67485726e-01 1.18429470e+00
-2.42812663e-01 5.05127311e-01 1.11077881e+00 -9.28818882e-01
-6.46795273e-01 -8.82157147e-01 1.52957308e+00 5.29286638e-02
8.00665677e-01 1.73879579e-01 -1.12099254e+00 8.41506183e-01
-1.81995034e-01 -5.99495396e-02 6.86072767e-01 -3.32339972e-01
1.94504142e-01 -1.55836433e-01 -1.21398365e+00 3.56439412e-01
-4.63231653e-02 -2.75860310e-01 -1.06098557e+00 7.76637718e-02
1.17959470e-01 -2.92228848e-01 -8.36699605e-01 3.88956845e-01
6.57518327e-01 -1.02236021e+00 9.15996432e-01 -4.47313607e-01
-8.26112553e-02 -6.05469823e-01 -5.95616288e-02 -1.26262391e+00
-4.15393770e-01 -7.87000179e-01 -8.07786807e-02 7.91652799e-01
4.25180137e-01 -1.11928105e+00 4.42951411e-01 2.28495523e-01
-8.26996937e-02 -2.95119643e-01 -1.14475644e+00 -1.05154872e+00
-2.68365800e-01 -5.48586369e-01 3.80864322e-01 1.24181640e+00
5.70775475e-04 -2.77401358e-01 -4.65172470e-01 5.74822605e-01
4.87602293e-01 -7.10949451e-02 5.55358171e-01 -1.56518233e+00
6.46892488e-02 -1.00257240e-01 -7.59081483e-01 -1.31409660e-01
-5.13035834e-01 -4.31681931e-01 -1.62137792e-01 -1.25164866e+00
-8.28089952e-01 -1.43932134e-01 -5.49123883e-01 3.41687709e-01
1.96595192e-01 2.57368624e-01 -1.52026594e-01 4.68736857e-01
4.22109723e-01 -1.89084727e-02 3.24899226e-01 3.30465674e-01
-4.32321370e-01 5.62254667e-01 7.03412890e-02 1.23236740e+00
1.22128081e+00 -9.18616474e-01 -1.08650006e-01 3.94536406e-02
3.26833427e-01 1.46011591e-01 2.79335856e-01 -1.32295537e+00
2.91381210e-01 8.86164680e-02 6.46475017e-01 -9.07776952e-01
-1.37695059e-01 -1.24482429e+00 5.70836842e-01 9.46142256e-01
4.29731160e-01 7.73477793e-01 3.28845143e-01 5.07279038e-01
-4.63096201e-01 -3.71380091e-01 6.80348217e-01 2.05216721e-01
-7.09496200e-01 -2.24213898e-01 -8.05802047e-01 -3.62072051e-01
9.63395655e-01 -4.54556286e-01 6.86265901e-02 -5.41070938e-01
-9.00145769e-01 -1.70349196e-01 5.62484302e-02 -2.57801153e-02
4.39488709e-01 -1.39504862e+00 -6.61622345e-01 3.95398229e-01
-9.47567970e-02 -6.31403327e-01 1.78846195e-01 1.24206865e+00
-1.06139326e+00 2.75567263e-01 -2.97505319e-01 -3.19284707e-01
-1.04017711e+00 2.35519797e-01 3.15348178e-01 -1.04524225e-01
-6.58111572e-01 1.25682712e-01 -7.12502241e-01 -9.37524438e-02
5.69925494e-02 -6.70815587e-01 -2.70623654e-01 3.46922517e-01
4.96102035e-01 7.94509113e-01 3.67801368e-01 -7.48119712e-01
-3.40911001e-01 7.22517610e-01 4.62130308e-01 2.46373601e-02
1.58448374e+00 -5.48658781e-02 -3.97559017e-01 8.57244492e-01
1.11400294e+00 2.91333288e-01 -5.42873204e-01 7.45844841e-02
6.05238378e-01 -2.17869207e-01 -3.65603149e-01 -4.05459613e-01
-8.35611761e-01 3.17415446e-01 6.06390119e-01 8.90668929e-01
1.44611251e+00 -8.29258204e-01 7.57912993e-01 4.32068914e-01
2.77894288e-01 -1.27259398e+00 -6.34786427e-01 7.97684729e-01
6.52340591e-01 -9.84286964e-01 1.65782392e-01 6.21874705e-02
-1.61628708e-01 1.57798505e+00 -1.20689847e-01 -7.12052524e-01
9.15452302e-01 2.27743670e-01 2.34584883e-01 2.72512180e-03
-2.67616272e-01 7.45093822e-02 2.21026510e-01 3.17424387e-01
4.31983292e-01 -1.12770081e-01 -7.65641272e-01 6.04040325e-01
-4.41329002e-01 -1.32701680e-01 6.25366688e-01 1.10292542e+00
-7.18827069e-01 -5.20525157e-01 -8.65034401e-01 4.52509940e-01
-8.62334371e-01 -4.97750416e-02 4.16369215e-02 1.26372552e+00
7.33425468e-02 1.02664649e+00 2.46853068e-01 -3.50033820e-01
4.94734347e-01 3.89070630e-01 -2.96112031e-01 -6.16132654e-02
-7.89720297e-01 1.91535518e-01 3.70340310e-02 -4.57224041e-01
-2.87192404e-01 -5.70734143e-01 -1.53828645e+00 -4.03993905e-01
-3.53239477e-01 7.10279346e-01 7.64569402e-01 9.07040179e-01
-1.17681690e-01 4.40924168e-01 9.12356436e-01 -8.96610975e-01
-2.89552659e-01 -7.31700957e-01 -8.09480488e-01 3.25503409e-01
6.35763466e-01 -2.56069571e-01 -1.10648394e+00 7.58595169e-02] | [6.6946940422058105, 2.974480152130127] |
7f20e1cf-50bc-4407-9ffb-fbc01fdefb48 | an-empirical-study-on-leveraging-position | 2109.01238 | null | https://arxiv.org/abs/2109.01238v1 | https://arxiv.org/pdf/2109.01238v1.pdf | An Empirical Study on Leveraging Position Embeddings for Target-oriented Opinion Words Extraction | Target-oriented opinion words extraction (TOWE) (Fan et al., 2019b) is a new subtask of target-oriented sentiment analysis that aims to extract opinion words for a given aspect in text. Current state-of-the-art methods leverage position embeddings to capture the relative position of a word to the target. However, the performance of these methods depends on the ability to incorporate this information into word representations. In this paper, we explore a variety of text encoders based on pretrained word embeddings or language models that leverage part-of-speech and position embeddings, aiming to examine the actual contribution of each component in TOWE. We also adapt a graph convolutional network (GCN) to enhance word representations by incorporating syntactic information. Our experimental results demonstrate that BiLSTM-based models can effectively encode position information into word representations while using a GCN only achieves marginal gains. Interestingly, our simple methods outperform several state-of-the-art complex neural structures. | ['Nikolaos Aletras', 'Kai Sun', 'Samuel Mensah'] | 2021-09-02 | null | https://aclanthology.org/2021.emnlp-main.722 | https://aclanthology.org/2021.emnlp-main.722.pdf | emnlp-2021-11 | ['target-oriented-opinion-words-extraction'] | ['natural-language-processing'] | [ 3.12010676e-01 2.69050181e-01 -3.13557863e-01 -4.47474062e-01
-7.41965890e-01 -5.64091206e-01 8.02681446e-01 5.00864863e-01
-5.38455784e-01 1.80302888e-01 6.92052841e-01 -6.42155588e-01
3.42752695e-01 -9.37215567e-01 -5.22894800e-01 -3.78927737e-01
-7.76802450e-02 1.34784445e-01 8.11150074e-02 -8.00257444e-01
5.32025814e-01 2.26746783e-01 -1.09526718e+00 4.79028672e-01
4.53250915e-01 9.28533316e-01 -2.38665923e-01 8.00346851e-01
-7.16196895e-01 7.03519881e-01 -8.41218054e-01 -7.03795612e-01
-1.56846121e-02 -3.36521447e-01 -5.25658488e-01 -1.05279842e-02
3.50103647e-01 -2.41201311e-01 -2.50765145e-01 1.05563319e+00
4.57279503e-01 -4.52234447e-02 8.06202292e-01 -9.93989885e-01
-1.27391541e+00 8.60741675e-01 -5.79912782e-01 3.56578976e-01
1.02101736e-01 -9.55139548e-02 1.73037386e+00 -1.18882191e+00
5.16155779e-01 1.23048496e+00 6.54468477e-01 3.95204991e-01
-9.19625640e-01 -2.62658775e-01 7.20664799e-01 2.51717512e-02
-6.99977338e-01 -3.79358530e-01 9.73842680e-01 -2.16794685e-01
1.65604734e+00 -1.62060782e-01 7.21006572e-01 1.18986857e+00
6.43031359e-01 8.96055579e-01 7.84595668e-01 -5.19647419e-01
1.91108346e-01 -3.41255590e-02 5.89680254e-01 7.78126121e-01
4.63994205e-01 -1.07138082e-01 -8.55291605e-01 9.77052189e-03
2.11673990e-01 3.12354807e-02 -6.33700937e-02 -3.15450370e-01
-9.14056420e-01 1.21495390e+00 6.71873331e-01 4.83652264e-01
-4.31639493e-01 4.33995932e-01 4.92973596e-01 2.46459365e-01
1.06669915e+00 7.44170904e-01 -8.25540781e-01 -1.21891730e-01
-7.47821450e-01 5.73424995e-02 8.94726574e-01 8.25835109e-01
7.51401663e-01 3.11130226e-01 -4.85819846e-01 5.77128053e-01
6.14501953e-01 4.45083708e-01 7.39058077e-01 -1.74942628e-01
4.66713667e-01 8.00981581e-01 -3.06642234e-01 -9.18284893e-01
-3.45347583e-01 -5.32525718e-01 -2.72036344e-01 1.35067791e-01
-5.60981734e-03 -5.07142782e-01 -1.42216766e+00 1.41115344e+00
1.58510152e-02 -3.70993644e-01 1.24836102e-01 6.46929920e-01
8.84354651e-01 7.71313548e-01 6.48415163e-02 2.39157751e-01
1.56445003e+00 -1.30317783e+00 -7.31776893e-01 -1.05481136e+00
9.56869841e-01 -8.58589649e-01 8.24602067e-01 5.82378879e-02
-9.78714049e-01 -3.87496352e-01 -1.30702245e+00 -3.18911940e-01
-1.02673006e+00 1.85528286e-02 5.29298007e-01 7.46270418e-01
-1.23120105e+00 5.05490184e-01 -6.63574815e-01 -2.56972671e-01
4.46078956e-01 3.66096646e-01 -2.63119400e-01 3.34480070e-02
-1.44071686e+00 1.18921888e+00 1.68975472e-01 4.31509316e-02
-3.39351803e-01 -6.83934331e-01 -1.42080235e+00 1.46967560e-01
1.32682413e-01 -6.47297084e-01 1.26989388e+00 -1.22002935e+00
-1.42844319e+00 6.01573110e-01 -5.08655906e-01 -4.38663483e-01
-4.27658319e-01 -3.57517362e-01 -5.05782783e-01 9.87199619e-02
1.16820104e-01 6.74475014e-01 1.09247744e+00 -9.56808329e-01
-4.32942212e-01 -5.49780428e-01 2.73587257e-01 2.05620721e-01
-7.99670577e-01 1.30188257e-01 -3.89660954e-01 -8.63356948e-01
-2.92837113e-01 -7.03361034e-01 -2.79477179e-01 3.15489760e-03
-3.68494451e-01 -4.33364630e-01 8.36530745e-01 -5.63293874e-01
1.32127869e+00 -2.06629372e+00 1.44766748e-01 3.59611288e-02
3.46174270e-01 4.92110133e-01 -7.37509072e-01 7.57037401e-01
-1.01885073e-01 4.16330367e-01 -1.65307775e-01 -6.23044074e-01
2.21418872e-01 1.32528082e-01 -3.88693988e-01 1.45838231e-01
9.02984917e-01 1.48868632e+00 -9.52697814e-01 -1.86053559e-01
5.88753894e-02 8.48384082e-01 -5.62748790e-01 1.08785154e-02
-2.98967510e-01 -5.04025400e-01 -5.03439248e-01 5.54706097e-01
2.55206704e-01 -1.34183004e-01 1.56773791e-01 -3.41965497e-01
7.72113577e-02 9.07897592e-01 -5.09391725e-01 1.47755432e+00
-7.28244781e-01 9.51714337e-01 -1.30445868e-01 -8.85757625e-01
9.17219400e-01 1.27632678e-01 2.73466688e-02 -7.46625125e-01
3.86344671e-01 2.32439473e-01 3.04353446e-01 -1.75588861e-01
1.02257335e+00 -2.27570817e-01 -1.98320955e-01 4.51983213e-01
4.44903731e-01 -1.40727580e-01 3.09272945e-01 3.29207391e-01
1.24812460e+00 1.38669135e-02 4.73149210e-01 -2.38051936e-01
2.42965654e-01 -1.41334692e-02 2.57628318e-02 5.18411875e-01
-1.76067293e-01 6.50202632e-01 7.45841980e-01 -3.19452792e-01
-8.36215913e-01 -7.34931469e-01 2.78518766e-01 1.38874757e+00
-2.78277159e-01 -7.75316596e-01 -4.45269883e-01 -9.50053632e-01
4.61704005e-03 1.03678310e+00 -1.06771624e+00 -2.81205446e-01
-5.53013206e-01 -7.41581261e-01 2.23737448e-01 7.68990993e-01
-1.82901509e-02 -1.17075145e+00 -4.88705993e-01 4.02855694e-01
2.52975315e-01 -9.99471605e-01 -8.19385886e-01 6.00108504e-01
-6.78259194e-01 -8.82314146e-01 -7.12767661e-01 -7.92636693e-01
7.57353365e-01 2.86708146e-01 1.36318052e+00 -1.28987119e-01
5.38007058e-02 4.27031934e-01 -8.09655130e-01 -8.04128885e-01
-1.79363474e-01 4.83513802e-01 -3.99586797e-01 -1.30759895e-01
1.02130532e+00 -2.12378800e-01 -5.23527741e-01 -3.16224843e-01
-9.67480063e-01 -4.33926761e-01 8.08957577e-01 7.42724895e-01
3.49753141e-01 -3.63856703e-01 4.30500269e-01 -1.18381882e+00
1.16293752e+00 -3.18774253e-01 -1.93812639e-01 -2.22582594e-02
-7.33076572e-01 4.62049663e-01 6.27436280e-01 -3.76240909e-01
-6.27593637e-01 -4.55789864e-01 -4.83953625e-01 -3.09764259e-02
2.44596019e-01 9.35191214e-01 1.20878793e-01 1.91290557e-01
3.52557629e-01 1.78510666e-01 -1.18036054e-01 -5.77541552e-02
7.07870185e-01 7.15360463e-01 -1.17373168e-01 -1.69917226e-01
5.66000581e-01 2.11576536e-01 -2.64856070e-01 -8.24225307e-01
-1.28812289e+00 -5.39712071e-01 -6.72335386e-01 1.65977433e-01
8.89415145e-01 -8.58840466e-01 -1.05870381e-01 2.00376466e-01
-1.43274653e+00 -1.16942778e-01 -3.57384712e-01 7.73517191e-02
-4.24703807e-02 1.80904984e-01 -5.62553167e-01 -7.37358630e-01
-7.60662973e-01 -1.07786608e+00 1.47614074e+00 2.67257411e-02
-4.60631192e-01 -1.36271322e+00 4.63353060e-02 1.69405323e-02
7.36590862e-01 -6.24843985e-02 8.60119343e-01 -1.08553374e+00
3.26143019e-03 -4.85358953e-01 -1.09387204e-01 4.44007427e-01
2.95735478e-01 8.39683879e-03 -8.55178833e-01 -2.49243259e-01
-1.06407717e-01 -1.48833066e-01 1.31576598e+00 3.02841842e-01
7.25854278e-01 -3.37961912e-01 -2.30963781e-01 2.70834416e-01
1.26770341e+00 -2.46802513e-02 5.41239142e-01 4.20451045e-01
8.80821168e-01 6.73792958e-01 2.51878530e-01 -1.07916817e-02
5.72187364e-01 1.81469113e-01 4.14947629e-01 -4.80221808e-02
-2.56520361e-01 -3.20462644e-01 7.96702862e-01 1.16391611e+00
3.86623621e-01 -6.22357726e-01 -6.56615078e-01 9.93498564e-01
-1.52215755e+00 -5.60355127e-01 -1.25521451e-01 1.50216615e+00
6.52911365e-01 6.34280443e-01 -3.38443309e-01 1.38727091e-02
3.79016876e-01 9.31049526e-01 -4.47361916e-01 -1.04069698e+00
-3.86454165e-02 6.46889806e-01 6.56019270e-01 5.81340611e-01
-1.09393179e+00 1.15471447e+00 6.51303768e+00 6.38212204e-01
-1.15068996e+00 3.85788865e-02 4.32721704e-01 2.40178872e-02
-7.08543420e-01 -1.02945894e-01 -9.13680255e-01 6.59434050e-02
1.36334932e+00 -1.47370184e-02 -7.00794682e-02 8.39699030e-01
-8.01174045e-02 2.69100219e-01 -1.02074277e+00 3.46316695e-01
6.70013428e-01 -1.15826154e+00 4.15870368e-01 1.66978836e-01
5.90750575e-01 3.09498847e-01 2.40608156e-01 4.69141841e-01
4.08886194e-01 -9.98192132e-01 5.98621547e-01 1.72725961e-01
3.94574732e-01 -8.16498101e-01 9.75421607e-01 -2.44690165e-01
-1.28651142e+00 8.46663564e-02 -3.75492960e-01 -2.49539375e-01
3.92908484e-01 7.06261456e-01 -5.76247454e-01 3.25570017e-01
3.36250484e-01 9.46459532e-01 -7.26951301e-01 2.45196953e-01
-7.30733514e-01 6.74285710e-01 4.01721969e-02 -7.69845426e-01
7.78019845e-01 1.23962037e-01 4.22649682e-01 1.32506716e+00
3.08140665e-01 -1.34463564e-01 -1.23678267e-01 6.83976114e-01
-2.99537420e-01 1.68142915e-01 -7.50310659e-01 -6.45997226e-01
8.77914503e-02 1.45420408e+00 -5.39249361e-01 -4.38086241e-01
-8.18444729e-01 1.09060788e+00 6.28302872e-01 3.61395180e-01
-4.78581607e-01 -7.86597371e-01 9.15911019e-01 -5.14240935e-02
1.07144928e+00 -4.56995010e-01 -2.81721860e-01 -1.21711636e+00
1.68367047e-02 -8.67968142e-01 7.22652003e-02 -8.03335011e-01
-1.49285185e+00 8.55582178e-01 -4.01020020e-01 -9.62915659e-01
-3.51179093e-01 -1.34519100e+00 -7.91323900e-01 9.26945031e-01
-1.87024462e+00 -1.33046448e+00 2.80063927e-01 1.45123050e-01
7.86909580e-01 -1.11708924e-01 9.72732186e-01 -1.85327947e-01
-3.81316900e-01 4.89801407e-01 -2.06537127e-01 5.22356629e-01
5.52368879e-01 -1.47123158e+00 1.14629900e+00 9.27283049e-01
5.38876593e-01 9.97481763e-01 5.13487339e-01 -5.83584547e-01
-1.58511472e+00 -1.20173407e+00 1.46848941e+00 -7.62007654e-01
1.19916260e+00 -5.87963939e-01 -5.85677326e-01 8.82152498e-01
6.91137791e-01 4.60963883e-02 9.46718991e-01 4.14968729e-01
-7.42646098e-01 1.25237674e-01 -5.63512743e-01 8.02247822e-01
7.37998366e-01 -7.48832107e-01 -1.00496387e+00 1.43782143e-03
1.11197269e+00 -5.80813214e-02 -5.68844855e-01 1.69674471e-01
3.56584579e-01 -8.85397971e-01 7.67682731e-01 -7.19116509e-01
8.40580046e-01 -5.68154156e-02 -1.42823786e-01 -1.85797000e+00
-2.03039870e-01 -1.51784226e-01 -3.88959348e-01 1.02129376e+00
9.60889995e-01 -7.18407273e-01 5.98005474e-01 3.57567817e-01
-2.52547234e-01 -1.06143177e+00 -5.45327067e-01 -3.32225114e-01
4.56712902e-01 -4.84853029e-01 3.99419248e-01 4.79840130e-01
2.45457351e-01 9.99682009e-01 -8.62438697e-03 4.74996455e-02
2.02586800e-01 -1.71495471e-02 2.52383620e-01 -9.15047586e-01
-1.24524690e-01 -6.65839851e-01 -4.26537275e-01 -1.26024663e+00
4.05805737e-01 -8.62166584e-01 1.42057329e-01 -2.14935446e+00
-1.77962095e-01 4.11239356e-01 -4.89100963e-01 5.00689745e-01
-5.64071953e-01 3.06731582e-01 1.96684837e-01 -4.10828173e-01
-5.93977451e-01 8.72537374e-01 1.02463877e+00 -4.98970330e-01
8.83741006e-02 -3.88069391e-01 -1.21132648e+00 5.84019601e-01
8.53543758e-01 -3.98597836e-01 -3.90697330e-01 -9.00919974e-01
5.05620241e-01 -5.68237007e-01 -3.87790911e-02 -3.69924307e-01
3.75530332e-01 3.78500372e-01 3.57324511e-01 -5.17825961e-01
3.54609936e-01 -5.11914432e-01 -1.04626799e+00 3.34654570e-01
-4.50818479e-01 5.87653399e-01 3.97488624e-01 5.63736439e-01
-5.17304003e-01 -3.26144934e-01 2.57508576e-01 -1.42315537e-01
-7.71698773e-01 1.49538264e-01 -7.23879397e-01 2.32473508e-01
6.20978594e-01 7.97107741e-02 -3.60911340e-01 -4.83971000e-01
-2.49133095e-01 3.42148058e-02 3.14105600e-02 7.79009938e-01
7.85734892e-01 -1.20107555e+00 -5.71241856e-01 2.70018458e-01
4.19576228e-01 -3.12138796e-01 -3.27313036e-01 4.03663248e-01
-2.89202541e-01 5.58454096e-01 2.59113163e-01 -4.78589423e-02
-9.92108703e-01 5.52446008e-01 2.83816606e-01 -5.95420480e-01
-2.44543001e-01 9.76042330e-01 -7.82106258e-03 -6.41351640e-01
-1.38843551e-01 -5.78567743e-01 -4.24218774e-01 4.87264186e-01
5.42283893e-01 -1.59981370e-01 2.83229530e-01 -7.05785215e-01
-5.11917055e-01 5.96843183e-01 -3.53702009e-01 -2.12914199e-01
1.27721453e+00 8.70663226e-02 -2.78966039e-01 5.10962069e-01
1.54384005e+00 1.15736008e-01 -8.01279902e-01 -2.05865934e-01
1.54372424e-01 3.57504934e-02 4.39718395e-01 -5.66849768e-01
-1.12652719e+00 1.11603665e+00 7.75062814e-02 2.54113197e-01
8.59236121e-01 -9.99400616e-02 9.39738631e-01 4.18433726e-01
1.36342952e-02 -9.99113262e-01 2.49996141e-01 8.73430729e-01
6.01130128e-01 -1.38228762e+00 -6.80330992e-02 -1.74580052e-01
-6.68382406e-01 1.40289211e+00 4.37192678e-01 -3.24691862e-01
9.71597612e-01 3.07284236e-01 3.79638672e-01 -5.29911816e-01
-8.60198379e-01 -6.40665293e-01 6.48080885e-01 4.77848023e-01
8.79983366e-01 -1.53169662e-01 -3.96349549e-01 5.45473397e-01
-3.23461086e-01 -4.78941798e-01 5.91787636e-01 1.12608397e+00
-5.23886681e-01 -1.24826419e+00 1.24521025e-01 6.98789954e-01
-7.03061163e-01 -9.39587772e-01 -6.82913721e-01 7.07499683e-01
-2.24362597e-01 1.33074737e+00 2.48060182e-01 -3.45736474e-01
5.76630533e-01 2.42256701e-01 1.69033781e-01 -1.00215423e+00
-8.29532325e-01 -1.39412414e-02 2.62622386e-01 -4.49179143e-01
-4.64551002e-01 -2.84345567e-01 -1.15363550e+00 -2.29123384e-02
-3.26624215e-01 5.95437139e-02 8.06690753e-01 1.02608335e+00
5.08012235e-01 9.94123757e-01 3.32356095e-01 -8.10230076e-01
-4.44916338e-01 -1.10594523e+00 -4.89337772e-01 2.06484869e-01
6.66757822e-01 -1.60849914e-01 -4.31408972e-01 -2.68153548e-01] | [11.41247844696045, 6.7529616355896] |
02ac4345-4cbc-4c0d-b824-f5b47e6dc8fa | streaming-classification-of-variable-stars | 1912.02235 | null | https://arxiv.org/abs/1912.02235v1 | https://arxiv.org/pdf/1912.02235v1.pdf | Streaming Classification of Variable Stars | In the last years, automatic classification of variable stars has received substantial attention. Using machine learning techniques for this task has proven to be quite useful. Typically, machine learning classifiers used for this task require to have a fixed training set, and the training process is performed offline. Upcoming surveys such as the Large Synoptic Survey Telescope (LSST) will generate new observations daily, where an automatic classification system able to create alerts online will be mandatory. A system with those characteristics must be able to update itself incrementally. Unfortunately, after training, most machine learning classifiers do not support the inclusion of new observations in light curves, they need to re-train from scratch. Naively re-training from scratch is not an option in streaming settings, mainly because of the expensive pre-processing routines required to obtain a vector representation of light curves (features) each time we include new observations. In this work, we propose a streaming probabilistic classification model; it uses a set of newly designed features that work incrementally. With this model, we can have a machine learning classifier that updates itself in real time with new observations. To test our approach, we simulate a streaming scenario with light curves from CoRot, OGLE and MACHO catalogs. Results show that our model achieves high classification performance, staying an order of magnitude faster than traditional classification approaches. | ['Lukas Zorich', 'Karim Pichara', 'Pavlos Protopapas'] | 2019-12-04 | null | null | null | null | ['classification-of-variable-stars'] | ['miscellaneous'] | [-8.08837786e-02 -5.70223629e-01 -9.27091092e-02 -5.24660349e-01
-3.16010058e-01 -8.22968066e-01 7.72195280e-01 3.44774812e-01
-4.49873000e-01 6.64589167e-01 -6.77581429e-01 -5.06613553e-01
-2.16511741e-01 -9.73710299e-01 -4.47406679e-01 -7.89297760e-01
-1.54261336e-01 9.72558975e-01 8.52914929e-01 -1.23224914e-01
2.63358504e-01 8.39961946e-01 -2.28266764e+00 6.96532950e-02
7.76910305e-01 9.23875451e-01 3.70264679e-01 1.08430934e+00
-4.00700033e-01 4.84054506e-01 -7.58947372e-01 -1.63331226e-01
5.12825847e-01 -2.49582782e-01 -6.77298486e-01 3.13028723e-01
5.38386762e-01 -9.81699079e-02 -1.59775317e-01 8.12388778e-01
1.50149792e-01 1.88301116e-01 5.11425018e-01 -1.39903057e+00
2.88223386e-01 4.46828067e-01 -1.99189082e-01 2.77885646e-01
-4.05845977e-03 2.82987922e-01 1.07350218e+00 -4.98682857e-01
5.14241219e-01 6.21767282e-01 5.93598723e-01 1.55658290e-01
-1.25566351e+00 -5.34456074e-01 3.95168997e-02 7.38001883e-01
-1.08339560e+00 5.90685047e-02 6.58205092e-01 -6.46688223e-01
7.04402626e-01 4.77021545e-01 9.64152634e-01 5.53301513e-01
-1.04022048e-01 3.20386589e-01 1.12231851e+00 -5.26260674e-01
6.54840946e-01 6.55438900e-01 3.99934679e-01 2.30339631e-01
3.42703789e-01 2.22282946e-01 -4.12266761e-01 -3.70952189e-01
2.54744232e-01 1.47728443e-01 -9.09768939e-02 -5.09681821e-01
-7.97697067e-01 1.07502544e+00 3.88919637e-02 1.81261718e-01
-3.75820309e-01 -3.24175432e-02 4.21791285e-01 6.72919691e-01
3.95235956e-01 2.91587055e-01 -8.09984326e-01 -3.79323542e-01
-1.06239212e+00 2.94217438e-01 1.10018814e+00 6.73695803e-01
1.00370693e+00 -1.33620664e-01 4.14245933e-01 7.50806272e-01
-3.24182399e-02 5.13584852e-01 8.08651030e-01 -6.89626992e-01
-1.51956573e-01 5.88198483e-01 1.09939180e-01 -4.01046634e-01
-3.94070178e-01 -5.06987393e-01 -4.67006773e-01 6.80485368e-01
5.44594646e-01 -3.60251367e-02 -5.93521059e-01 1.10335302e+00
5.90458810e-01 3.25726897e-01 1.17017396e-01 6.83227897e-01
3.68924648e-01 6.55184448e-01 -4.30171251e-01 -4.08927202e-01
9.96289670e-01 -6.29941046e-01 -3.42399299e-01 1.01321135e-02
5.48318624e-01 -7.51453817e-01 8.44070494e-01 8.38427424e-01
-5.21221519e-01 -5.73536217e-01 -9.93808210e-01 5.50083637e-01
-3.50151509e-01 -2.81569418e-02 9.42765236e-01 7.01225698e-01
-6.84788644e-01 7.39888012e-01 -7.80565798e-01 -4.44588184e-01
5.15591092e-02 2.56651551e-01 -2.04642072e-01 2.16192245e-01
-6.11483097e-01 6.03160977e-01 3.23941886e-01 -2.33180776e-01
-8.03246856e-01 -5.23223758e-01 -4.14520711e-01 2.94358075e-01
3.81003052e-01 -3.22509885e-01 1.68938470e+00 -1.09692156e+00
-1.60904944e+00 4.43198234e-01 -2.27592006e-01 -6.45902991e-01
3.53611231e-01 8.63204002e-02 -2.69000500e-01 2.24247407e-02
-4.81036097e-01 1.24948956e-01 1.12856483e+00 -1.03208482e+00
-1.19806790e+00 -2.31112853e-01 1.03360191e-02 -1.45468771e-01
-4.77724373e-01 2.10545123e-01 -2.48143658e-01 -3.55480194e-01
2.15503037e-01 -1.22968900e+00 -1.21739388e-01 -2.50240773e-01
2.92901754e-01 -6.72958076e-01 9.65295792e-01 -1.91340148e-01
8.53580713e-01 -2.05396008e+00 -3.42913419e-02 3.57331932e-01
-1.78181939e-02 3.86453897e-01 3.86768252e-01 4.66053873e-01
-1.15361229e-01 -3.09634954e-01 -3.02832961e-01 -2.73693621e-01
-1.93982631e-01 4.41069931e-01 -4.63065058e-01 3.46814722e-01
-9.58081931e-02 5.56875840e-02 -6.61407590e-01 -1.73163712e-01
3.02926600e-01 5.03063900e-03 -6.54562950e-01 3.48428726e-01
-6.27573907e-01 3.07733446e-01 -2.60201525e-02 1.27910852e-01
7.58750677e-01 -3.27441365e-01 2.40975812e-01 3.93790245e-01
-5.23691118e-01 1.56078950e-01 -1.51204646e+00 1.27031004e+00
-6.61124945e-01 9.79008853e-01 -3.27724636e-01 -1.25928092e+00
1.14033449e+00 1.54319823e-01 5.86533248e-01 2.47612875e-02
-2.84729488e-02 1.87692389e-01 1.41688049e-01 -5.09235978e-01
3.63800704e-01 -7.65273422e-02 3.51906896e-01 5.62895119e-01
1.05622746e-01 -4.58777040e-01 3.82959753e-01 1.36653408e-01
1.22722912e+00 8.19517225e-02 3.71920504e-02 2.41526902e-01
5.94651878e-01 2.59599924e-01 5.15218794e-01 7.25191534e-01
1.58239171e-01 3.50396693e-01 2.13867053e-01 -7.11599171e-01
-1.06388235e+00 -7.25756407e-01 -3.69539738e-01 1.13135278e+00
-2.78066218e-01 -5.68442345e-01 -2.22288206e-01 -7.29799569e-01
1.93599716e-01 6.43132508e-01 -2.45365158e-01 5.72531335e-02
-1.83014050e-01 -8.00995767e-01 -1.14119343e-01 4.14353125e-02
1.81614086e-01 -9.41757083e-01 -9.69141781e-01 4.46692318e-01
3.38081241e-01 -9.57343161e-01 2.73826480e-01 4.75549102e-01
-1.09210646e+00 -1.08145809e+00 -2.61611372e-01 -5.43121457e-01
4.04642224e-01 4.93956327e-01 9.20404673e-01 2.98462987e-01
-4.85287815e-01 2.60165006e-01 -7.04554796e-01 -7.51581252e-01
-4.85360116e-01 2.07554549e-01 1.28140107e-01 2.31652901e-01
5.43941915e-01 -7.23280907e-01 -1.60250247e-01 2.99099922e-01
-9.66936409e-01 -1.17527105e-01 3.73639882e-01 8.27137053e-01
2.66510785e-01 5.23413777e-01 5.40295839e-01 -8.38313460e-01
-4.47011404e-02 -5.29754877e-01 -1.47972143e+00 1.28969714e-01
-9.16996658e-01 1.88298121e-01 8.50971162e-01 -6.30983055e-01
-1.02343225e+00 3.97439510e-01 -1.29731148e-01 -3.74790579e-01
-4.03993726e-01 3.99993330e-01 3.69403601e-01 -2.03704610e-01
7.64469385e-01 2.13924259e-01 6.20825440e-02 -8.16102147e-01
2.76821226e-01 1.07810295e+00 4.65023637e-01 -4.40027624e-01
1.25843835e+00 5.57222843e-01 5.28122447e-02 -9.27103281e-01
-7.79623032e-01 -9.93373752e-01 -7.02876270e-01 -4.63370740e-01
1.93128899e-01 -6.09909475e-01 -7.11083531e-01 5.49672365e-01
-9.98167396e-01 -1.76516742e-01 -5.10668516e-01 8.46379817e-01
-3.51934940e-01 4.60167587e-01 2.72333212e-02 -8.22462440e-01
-9.09159780e-02 -8.78814816e-01 5.66685975e-01 5.52879870e-01
2.25225106e-01 -8.64031971e-01 4.23348397e-01 5.34945838e-02
4.91760373e-01 -1.26134872e-01 5.38992345e-01 -8.92071605e-01
-6.56462073e-01 -4.85361010e-01 3.86946276e-02 5.03572881e-01
3.18488181e-02 3.73781443e-01 -1.14133012e+00 -4.56302762e-01
8.09552222e-02 -1.71095625e-01 8.10826898e-01 5.90694323e-02
1.44902587e+00 4.34825383e-02 -2.17705548e-01 6.68025196e-01
1.30713093e+00 2.25307345e-01 2.43949190e-01 4.38036770e-01
1.50361583e-01 7.04004049e-01 6.26460016e-01 9.40301299e-01
2.77232021e-01 7.23223567e-01 3.31201315e-01 2.91539997e-01
6.29033744e-02 1.87594905e-01 3.59676510e-01 7.68506467e-01
-9.77656394e-02 1.87429458e-01 -8.18806946e-01 4.54525679e-01
-1.73681617e+00 -1.05528879e+00 -5.14479399e-01 2.74286580e+00
7.27036476e-01 3.27784896e-01 1.10859491e-01 5.05424559e-01
3.56191784e-01 -3.73424023e-01 -5.57648301e-01 -2.93751359e-01
1.40769228e-01 4.83263969e-01 5.79937756e-01 4.10987526e-01
-8.89068365e-01 6.37914181e-01 5.79052782e+00 4.07799304e-01
-1.46339619e+00 1.13649026e-01 -1.18487366e-01 -1.38374537e-01
-1.30283415e-01 6.94444478e-01 -1.10680425e+00 5.97748339e-01
1.29243088e+00 -2.90009886e-01 5.78113616e-01 1.06247771e+00
1.28139064e-01 -4.81825590e-01 -9.45561945e-01 1.13437510e+00
1.56906322e-01 -1.19259441e+00 -3.90198588e-01 2.20535118e-02
5.16366065e-01 4.31032479e-01 -4.12063748e-01 6.01395667e-01
4.26781088e-01 -3.87437046e-01 3.92714322e-01 6.05063975e-01
3.97017181e-01 -5.96724212e-01 8.06057334e-01 6.69174194e-01
-1.10453928e+00 -3.08817804e-01 -6.72888696e-01 -3.05782884e-01
-6.81086108e-02 9.23542798e-01 -1.19226253e+00 7.10316896e-01
9.62029159e-01 4.57591325e-01 -1.02465463e+00 1.66762650e+00
-9.70248431e-02 1.12283087e+00 -9.17512238e-01 -2.50003427e-01
-8.86948779e-02 -2.28629842e-01 5.39543390e-01 8.72775197e-01
4.43845600e-01 -2.42571592e-01 2.61492968e-01 2.01081529e-01
2.90865690e-01 1.83446646e-01 -5.69153726e-01 2.75786221e-01
3.12566876e-01 1.52942538e+00 -6.11504495e-01 -6.86692834e-01
-7.10961401e-01 7.05779850e-01 2.63243645e-01 -3.08868643e-02
-6.72841966e-01 -4.96618509e-01 4.87327427e-01 1.20524839e-01
5.54205298e-01 -3.96488369e-01 -5.05177267e-02 -1.19744396e+00
-1.09501712e-01 -6.36567295e-01 5.50942838e-01 -6.50229990e-01
-1.27338731e+00 5.28421819e-01 6.00003041e-02 -1.36665881e+00
-5.60143292e-01 -5.61851799e-01 -7.19492614e-01 5.87030888e-01
-1.43387914e+00 -6.24916494e-01 -6.77854121e-01 2.98404813e-01
6.47050023e-01 -5.79973459e-01 6.05701745e-01 2.67611802e-01
-4.29992616e-01 1.42950252e-01 4.53461051e-01 -3.64422947e-01
9.05290782e-01 -1.54334331e+00 5.03395833e-02 7.42478132e-01
5.79210460e-01 4.45688106e-02 1.09833074e+00 -5.28329670e-01
-1.31537676e+00 -9.41334426e-01 7.63702750e-01 -3.53659868e-01
8.95440638e-01 -2.79811144e-01 -1.28420115e+00 5.57802856e-01
4.49861325e-02 1.87560737e-01 6.48831844e-01 3.16291809e-01
-2.50734270e-01 -5.42948961e-01 -7.68692851e-01 -2.96415705e-02
5.07848144e-01 -2.25658372e-01 -6.05272710e-01 6.47377729e-01
3.96173656e-01 1.14075638e-01 -8.82186174e-01 2.05994219e-01
4.55576181e-01 -9.35972750e-01 4.63912785e-01 -5.38704574e-01
-3.86991888e-01 -4.79764253e-01 1.15391001e-01 -1.49358225e+00
-5.17640114e-02 -7.05376804e-01 -1.18867010e-01 1.13781929e+00
3.12894523e-01 -9.47930932e-01 8.19999993e-01 2.84762055e-01
9.01883319e-02 -1.11568123e-01 -8.71170461e-01 -1.10848010e+00
-1.33344740e-01 -6.65874064e-01 6.67889118e-01 8.20750713e-01
-1.02465801e-01 3.58597398e-01 -1.56994313e-01 4.20823395e-01
5.49717069e-01 5.01825750e-01 1.48848760e+00 -1.98400545e+00
-8.50219965e-01 -2.51374096e-01 -5.85520387e-01 -7.46547580e-01
3.24613862e-02 -9.90847647e-01 -5.52146211e-02 -1.10066462e+00
1.43657029e-01 -8.95569384e-01 5.94168678e-02 4.29065466e-01
-6.94963560e-02 -1.56621933e-02 1.27920285e-01 3.75211030e-01
-2.69935817e-01 3.64242464e-01 4.66286272e-01 1.45588025e-01
-3.33736241e-01 5.38729966e-01 1.05308205e-01 7.05264509e-01
9.47538793e-01 -6.68688893e-01 -2.06724167e-01 3.70238721e-02
3.20692211e-01 -1.32476255e-01 2.29802921e-01 -1.31505561e+00
4.10755724e-01 -2.50087500e-01 1.03003696e-01 -7.85392106e-01
3.64214957e-01 -1.00211120e+00 2.68350989e-01 4.50084388e-01
1.87245548e-01 -5.26864529e-02 3.98285538e-02 5.02853513e-01
-2.38374233e-01 -9.95255470e-01 8.74822080e-01 1.06832072e-01
-6.29516602e-01 3.22472960e-01 -6.89777672e-01 -4.55833405e-01
1.32593668e+00 2.97855616e-01 -1.12115048e-01 -3.50693762e-01
-6.71709359e-01 6.59257993e-02 5.52124500e-01 3.73906612e-01
1.79981530e-01 -6.94932461e-01 -7.39645660e-01 4.36192572e-01
4.07930613e-01 -1.02830958e-02 4.59318347e-02 5.18357217e-01
-6.09391868e-01 1.22146666e-01 7.32999146e-02 -9.36576605e-01
-1.45344388e+00 7.48789608e-01 2.16282219e-01 -1.15094475e-01
-8.70875359e-01 4.37725723e-01 -3.21027905e-01 -4.18304443e-01
2.10306421e-01 -3.12163413e-01 -4.25987095e-01 2.46327981e-01
6.38480008e-01 1.02321878e-01 4.61741954e-01 -7.68440291e-02
2.05077007e-01 3.21567535e-01 -1.13236077e-01 -2.48408780e-01
1.54479218e+00 1.53484821e-01 -8.96214470e-02 7.53298521e-01
6.16411746e-01 1.14845641e-01 -1.16441739e+00 -5.39593458e-01
3.28708947e-01 -7.79773176e-01 1.60320655e-01 -4.91686165e-01
-8.08553100e-01 8.86089027e-01 8.55769336e-01 6.29052401e-01
1.09686267e+00 2.01616943e-01 5.14086783e-01 8.48209560e-01
6.32253528e-01 -1.07115960e+00 -2.74450332e-01 5.06617486e-01
5.85845292e-01 -1.29272115e+00 -7.18035642e-03 -2.12199122e-01
-2.72700161e-01 1.50738084e+00 3.92371386e-01 7.47673912e-03
8.20762098e-01 2.00217977e-01 1.93910804e-02 -2.79816017e-02
-1.37336862e+00 -5.56450665e-01 4.02924493e-02 3.72581512e-01
-7.56559446e-02 6.88743144e-02 -3.50956261e-01 1.20069973e-01
-4.77374643e-01 -9.61854160e-02 9.73992705e-01 9.69312727e-01
-8.82657290e-01 -1.51659131e+00 -7.36974716e-01 7.60386467e-01
-4.67694849e-02 3.47177088e-01 -1.33056238e-01 4.28160936e-01
-2.69165821e-02 9.94083881e-01 3.58708948e-01 -2.71774560e-01
1.69581041e-01 3.78281832e-01 2.36582518e-01 -7.21287251e-01
-5.04554808e-01 -8.53871033e-02 1.88602949e-04 -9.40303430e-02
-3.34105045e-01 -1.03503549e+00 -1.14965606e+00 -3.28433156e-01
-4.91611093e-01 5.51514208e-01 1.31630874e+00 8.84274840e-01
-2.42988374e-02 4.03531998e-01 1.35643482e+00 -7.86661208e-01
-8.14122617e-01 -1.26505172e+00 -4.72251296e-01 3.40851098e-01
3.86991538e-02 -9.38299298e-01 -7.71337748e-01 2.94300735e-01] | [7.6394500732421875, 3.1035611629486084] |
8b66468b-85b6-4c45-b9dd-d27c70fb17d3 | segmentation-guided-domain-adaptation-for | 2210.09213 | null | https://arxiv.org/abs/2210.09213v2 | https://arxiv.org/pdf/2210.09213v2.pdf | Segmentation-guided Domain Adaptation for Efficient Depth Completion | Complete depth information and efficient estimators have become vital ingredients in scene understanding for automated driving tasks. A major problem for LiDAR-based depth completion is the inefficient utilization of convolutions due to the lack of coherent information as provided by the sparse nature of uncorrelated LiDAR point clouds, which often leads to complex and resource-demanding networks. The problem is reinforced by the expensive aquisition of depth data for supervised training. In this work, we propose an efficient depth completion model based on a vgg05-like CNN architecture and propose a semi-supervised domain adaptation approach to transfer knowledge from synthetic to real world data to improve data-efficiency and reduce the need for a large database. In order to boost spatial coherence, we guide the learning process using segmentations as additional source of information. The efficiency and accuracy of our approach is evaluated on the KITTI dataset. Our approach improves on previous efficient and low parameter state of the art approaches while having a noticeably lower computational footprint. | ['Stefan Rudolph', 'Anselm Haselhoff', 'Martin Sunkel', 'Fabian Märkert'] | 2022-10-14 | null | null | null | null | ['depth-completion'] | ['computer-vision'] | [ 1.50508881e-01 -1.09609552e-01 1.32748768e-01 -5.86107790e-01
-6.53437912e-01 -3.93745780e-01 4.77233142e-01 1.09827369e-01
-8.43992829e-01 9.18723643e-01 -1.42973185e-01 -2.42238835e-01
-2.57960320e-01 -9.83450055e-01 -6.87749028e-01 -4.85929281e-01
1.74760282e-01 7.42072046e-01 4.13451999e-01 -1.14247486e-01
2.10358754e-01 9.13405895e-01 -1.75812304e+00 -1.06453747e-01
1.03898776e+00 1.20221782e+00 6.27731740e-01 4.97261524e-01
-2.68599719e-01 4.90695357e-01 -1.83580890e-01 -9.38691720e-02
4.33491200e-01 3.53285149e-02 -5.49663484e-01 3.07046659e-02
7.56286621e-01 -5.47174990e-01 -3.80468309e-01 8.12368810e-01
4.77243513e-01 1.73984721e-01 4.10646260e-01 -1.06355345e+00
2.58290946e-01 -1.28294295e-02 -4.34642345e-01 4.15120237e-02
-4.35831517e-01 1.44312441e-01 6.24487936e-01 -9.73895133e-01
6.06476784e-01 8.36909652e-01 8.60914588e-01 2.84862399e-01
-1.24409223e+00 -6.81372583e-01 -1.79279521e-02 3.30195010e-01
-1.44616890e+00 -5.36463141e-01 8.78676295e-01 -3.79920691e-01
1.09821951e+00 -2.17795894e-01 7.71419883e-01 6.66556060e-01
-1.72020227e-01 3.60992104e-01 8.82141113e-01 -2.27226093e-01
5.30449212e-01 1.80154935e-01 -1.21109068e-01 6.25438750e-01
5.22425413e-01 3.11334819e-01 -6.77457035e-01 2.82862872e-01
8.72703552e-01 -3.89498472e-02 3.98874469e-02 -7.78663695e-01
-8.49894583e-01 9.64756787e-01 7.72462249e-01 -1.42121151e-01
-4.42280263e-01 4.04635549e-01 1.18958510e-01 9.63023007e-02
6.69233024e-01 3.10710430e-01 -5.06364942e-01 -2.01067194e-01
-1.21404505e+00 2.91042149e-01 6.17628813e-01 8.87585461e-01
1.40249527e+00 1.08090900e-01 5.35500526e-01 6.87256873e-01
3.53965104e-01 5.79345882e-01 -7.29093328e-02 -1.37495315e+00
6.36748910e-01 5.79186916e-01 4.54017706e-02 -6.20819390e-01
-4.40482765e-01 -6.73083425e-01 -7.99209297e-01 5.97578943e-01
5.72013855e-01 1.50438622e-02 -9.52741325e-01 1.43042350e+00
2.98047751e-01 2.94623315e-01 7.56179448e-03 7.77252138e-01
5.26887357e-01 3.14613134e-01 -1.40300676e-01 2.22211108e-01
8.75552535e-01 -6.54938400e-01 -2.20375776e-01 -6.48024440e-01
5.80517113e-01 -4.70234126e-01 7.85832226e-01 3.69279921e-01
-6.05085492e-01 -5.77372253e-01 -1.20843446e+00 -3.86801362e-01
-3.43791068e-01 1.57618880e-01 9.19015646e-01 5.12359679e-01
-1.04950738e+00 7.88534641e-01 -9.55619752e-01 -3.94080490e-01
9.26052630e-01 4.61089224e-01 -4.92244601e-01 -5.26602924e-01
-7.22780347e-01 9.15642202e-01 4.55521882e-01 5.57195917e-02
-7.29477406e-01 -1.02028561e+00 -1.03253388e+00 -1.19250722e-01
2.26265132e-01 -6.68299079e-01 1.13740444e+00 -5.21742344e-01
-1.36690462e+00 5.83770037e-01 -6.43370003e-02 -7.85440683e-01
6.77940905e-01 -2.81316310e-01 2.24939674e-01 2.75987774e-01
1.77262500e-01 1.15507054e+00 7.33027756e-01 -1.08301604e+00
-6.59783483e-01 -5.58664680e-01 -1.44304307e-02 2.43387595e-01
-3.43941599e-02 -6.84307754e-01 -4.64677155e-01 -7.19638914e-02
3.32518578e-01 -8.76621425e-01 -5.55481732e-01 3.75171632e-01
5.95513955e-02 1.66434735e-01 8.72941613e-01 -3.78058761e-01
4.54579920e-01 -2.12458587e+00 -1.78087845e-01 1.31283566e-01
4.06815290e-01 1.66454434e-01 -6.75464496e-02 2.98495770e-01
1.06805645e-01 -4.20327455e-01 -5.62675953e-01 -7.07693994e-01
-2.69963533e-01 6.66389763e-01 -1.91549003e-01 5.63434660e-01
4.74009871e-01 7.75990307e-01 -7.66318023e-01 -4.79783237e-01
6.30364895e-01 7.71925867e-01 -7.41760075e-01 1.99760988e-01
-3.83614093e-01 7.23846614e-01 -2.13868335e-01 3.44450593e-01
1.04450953e+00 -3.30484100e-02 -3.05327654e-01 -1.84388518e-01
-4.84201729e-01 5.85017264e-01 -1.17838347e+00 2.23635840e+00
-7.48044491e-01 7.88818538e-01 1.45971507e-01 -1.14221406e+00
8.77798975e-01 -9.40455571e-02 4.24956560e-01 -8.84160161e-01
9.26243588e-02 3.68517697e-01 -1.88013092e-01 -1.72909930e-01
5.90841293e-01 -1.94660917e-01 2.69529909e-01 1.93617195e-01
1.64809555e-01 -8.60851824e-01 1.09436736e-01 1.44735321e-01
1.02349865e+00 4.98806745e-01 -8.42738077e-02 -2.48713821e-01
2.30876878e-01 3.36360335e-01 4.54793006e-01 5.07803619e-01
5.19093201e-02 6.83906794e-01 1.96207434e-01 -5.21396339e-01
-1.19635510e+00 -8.67517054e-01 -2.20897079e-01 3.43355030e-01
-3.54135782e-02 -9.52764750e-02 -5.33638537e-01 -2.63496548e-01
1.26844645e-01 6.13663495e-01 -3.92661780e-01 9.36674699e-02
-4.25215960e-01 -3.70436311e-01 4.37681168e-01 7.10703731e-01
8.09774637e-01 -6.94490910e-01 -9.79928851e-01 3.21283728e-01
-5.50569259e-02 -1.46070850e+00 2.67837703e-01 5.98322153e-01
-1.32738209e+00 -8.89859676e-01 -4.32062536e-01 -3.11061144e-01
4.86093253e-01 3.84287804e-01 1.18087697e+00 -2.81714261e-01
-4.40134585e-01 1.50110021e-01 7.41816387e-02 -6.89457774e-01
7.59052038e-02 2.87511230e-01 -5.82009703e-02 -2.16377750e-01
3.59498054e-01 -9.95458245e-01 -7.09836423e-01 -5.44324964e-02
-8.42396736e-01 6.46734238e-02 6.54478788e-01 7.27783024e-01
7.55533278e-01 -5.93005754e-02 2.46419966e-01 -8.59141767e-01
4.61297035e-02 -1.29255399e-01 -1.17087567e+00 -5.65061331e-01
-7.63366878e-01 2.73119569e-01 3.17610115e-01 -2.02090442e-02
-1.06750178e+00 6.48829281e-01 -2.35905752e-01 -5.63961923e-01
-2.51344025e-01 4.50096458e-01 -5.52909896e-02 -4.56872135e-01
6.88853621e-01 1.70081835e-02 1.81147814e-01 -4.25395221e-01
4.66326207e-01 3.45353186e-01 5.67596138e-01 -3.28450233e-01
9.54404056e-01 9.23186004e-01 4.42465395e-01 -1.22797954e+00
-7.74533272e-01 -7.48654366e-01 -9.75823760e-01 -4.00509536e-02
6.65625274e-01 -1.29303050e+00 -3.18355083e-01 4.63464648e-01
-1.29065621e+00 -5.41560829e-01 -5.87739706e-01 6.80126786e-01
-6.05343938e-01 3.18555683e-01 -1.16241701e-01 -7.01849580e-01
-1.82627812e-02 -8.36712420e-01 1.08066201e+00 2.74250042e-02
1.15236491e-01 -7.92887509e-01 1.54921994e-01 2.91585773e-01
5.54875135e-01 1.83495671e-01 6.10898018e-01 6.29887804e-02
-1.20195067e+00 -1.92889541e-01 -6.78923607e-01 4.62410003e-01
-1.67587653e-01 -2.97574401e-01 -1.26332438e+00 -2.15813518e-02
-3.96473818e-02 -3.65003675e-01 1.13372099e+00 3.89345556e-01
1.01962543e+00 2.82751620e-01 -1.92206308e-01 9.18379664e-01
1.74548280e+00 -2.62635618e-01 6.57830775e-01 2.82194048e-01
7.63867557e-01 8.36437821e-01 6.55061603e-01 4.54203337e-01
5.29413402e-01 3.76239270e-01 8.29793692e-01 -3.15496475e-01
-3.22630227e-01 -2.19347998e-01 -2.60199249e-01 5.22998631e-01
-3.37795094e-02 1.93189755e-02 -1.08359683e+00 8.32541764e-01
-1.79948330e+00 -7.07667708e-01 -2.63909131e-01 2.30600381e+00
4.48555261e-01 2.33636975e-01 -2.15661556e-01 2.30403900e-01
7.82693177e-02 8.78403187e-02 -5.58332801e-01 -3.45814265e-02
6.69126511e-02 7.05077589e-01 9.21574175e-01 5.97871065e-01
-9.46578205e-01 1.03327382e+00 5.77092266e+00 5.46702921e-01
-1.15966582e+00 1.26524076e-01 2.45019928e-01 -2.93356508e-01
-1.32834136e-01 2.76870397e-03 -6.97919369e-01 5.13434596e-02
1.01102030e+00 1.49026737e-01 2.85623759e-01 9.86610830e-01
2.76247054e-01 -6.86192334e-01 -1.06443632e+00 1.34421253e+00
-1.60873830e-01 -1.55361640e+00 -3.77626210e-01 2.13755459e-01
6.15331650e-01 8.20657492e-01 -1.56617090e-01 5.60706016e-03
2.34294564e-01 -9.15587902e-01 5.98137558e-01 3.97730261e-01
8.91892552e-01 -9.74433184e-01 6.78839743e-01 4.96332854e-01
-1.17528307e+00 1.21006742e-01 -7.01555550e-01 -5.04303396e-01
6.23877458e-02 1.00463116e+00 -8.58138025e-01 5.05011559e-01
7.40979671e-01 7.41036475e-01 -4.25112307e-01 1.10805690e+00
-3.15109134e-01 3.82635087e-01 -9.04297352e-01 2.31067345e-01
3.50982875e-01 -2.98476815e-01 1.97171167e-01 9.37937617e-01
2.55763918e-01 -2.41279844e-02 -1.78696886e-02 9.92208242e-01
-4.14973386e-02 -2.38073274e-01 -9.74664092e-01 1.75362021e-01
4.27589208e-01 1.25994813e+00 -7.19077349e-01 -3.83596681e-02
-3.73480111e-01 7.92764485e-01 4.80558693e-01 1.57044008e-01
-4.41946983e-01 -3.18447083e-01 8.14101160e-01 3.86863321e-01
5.35552263e-01 -7.83126593e-01 -5.88579059e-01 -8.30240488e-01
9.88267139e-02 -2.51239270e-01 -1.76023498e-01 -6.20167196e-01
-8.69882107e-01 4.68406349e-01 -6.89926371e-02 -1.26692104e+00
-3.72284442e-01 -6.05489194e-01 -2.84707308e-01 1.13703144e+00
-2.20478296e+00 -1.08457661e+00 -9.37709391e-01 6.24337792e-01
4.03752238e-01 9.10263062e-02 6.42258525e-01 5.65639079e-01
-5.64798191e-02 4.42563184e-02 -9.01938155e-02 -1.24175102e-01
4.70624864e-01 -1.09565604e+00 5.61502695e-01 8.78694534e-01
8.63749534e-02 2.15752631e-01 5.17727852e-01 -4.40262973e-01
-1.28902376e+00 -1.29554403e+00 8.02657604e-01 -3.79278898e-01
4.35858935e-01 -4.11587030e-01 -8.18250358e-01 4.18712586e-01
-2.18829602e-01 2.58630186e-01 3.54800820e-01 -6.87969178e-02
-2.86020100e-01 -4.29485112e-01 -1.01267314e+00 2.76299059e-01
1.07562494e+00 -7.04308033e-01 -2.82884061e-01 8.44179913e-02
6.01932347e-01 -4.00114954e-01 -3.43936920e-01 5.01532197e-01
3.66091818e-01 -1.20753348e+00 8.50341618e-01 1.03080064e-01
5.00721514e-01 -3.33991677e-01 -2.22592309e-01 -1.12710238e+00
-2.82209013e-02 -1.56041995e-01 1.51289985e-01 8.21218967e-01
2.65899330e-01 -6.19138420e-01 1.41017830e+00 4.62080032e-01
-2.29510665e-01 -4.06781554e-01 -1.22493923e+00 -7.49790788e-01
-5.75799542e-03 -9.75832582e-01 3.34353924e-01 6.81432426e-01
-6.09789252e-01 4.16779160e-01 -1.11190733e-02 2.79124439e-01
9.73088384e-01 6.90293536e-02 1.06793058e+00 -1.60283422e+00
-1.75134704e-01 -7.95083866e-02 -6.26407444e-01 -1.21120524e+00
1.25674754e-02 -6.27797604e-01 3.42308372e-01 -1.62716877e+00
-2.82151163e-01 -8.17697585e-01 1.76854700e-01 2.96662271e-01
3.46535832e-01 6.00318372e-01 -8.60277265e-02 7.70304948e-02
-2.92190999e-01 7.21438706e-01 8.03629160e-01 -1.31461015e-02
-1.44766212e-01 -1.21139467e-01 -1.71336934e-01 6.79080367e-01
7.56346226e-01 -5.38426638e-01 -7.30429173e-01 -7.83256590e-01
2.77133465e-01 -5.79502247e-02 5.68687618e-01 -1.50132143e+00
3.58494937e-01 1.20148331e-01 3.33931118e-01 -1.08958340e+00
8.43967080e-01 -1.05345702e+00 -8.99447594e-04 4.35394913e-01
2.72060603e-01 -9.32935178e-02 4.04802293e-01 4.93627876e-01
-3.84203434e-01 -1.08131282e-01 8.55533242e-01 -1.80067867e-01
-9.02564704e-01 5.93942523e-01 -5.41238785e-02 -2.21298024e-01
7.60618567e-01 -4.85090047e-01 7.79034048e-02 -4.08132881e-01
-3.07368279e-01 1.67778432e-01 6.07125759e-01 1.42833725e-01
7.71398246e-01 -9.98131692e-01 -6.54023945e-01 4.82242852e-01
2.97160894e-01 8.04410338e-01 1.96827531e-01 7.40918636e-01
-8.75144541e-01 5.41217387e-01 -3.45982194e-01 -9.00090575e-01
-9.79361594e-01 8.26015547e-02 3.89687777e-01 -1.46737888e-01
-7.54246294e-01 9.62219954e-01 1.14285737e-01 -4.73919749e-01
9.56771597e-02 -4.72172529e-01 7.84913003e-02 -3.91647266e-03
2.11771443e-01 3.14807117e-01 4.89749074e-01 -3.67660642e-01
-3.10972780e-01 6.15642190e-01 9.52939615e-02 -2.97280788e-01
1.67833900e+00 -1.31278053e-01 -4.11797315e-02 3.55228543e-01
1.02138066e+00 -2.22353354e-01 -1.74905157e+00 -4.07094657e-01
-7.75681138e-02 -5.74230015e-01 5.10373950e-01 -4.62632984e-01
-1.14631629e+00 1.32833457e+00 7.05844998e-01 -3.88604969e-01
8.50781024e-01 -1.44177243e-01 6.27179086e-01 6.48840785e-01
6.27445877e-01 -1.11137056e+00 -7.65490010e-02 7.64881074e-01
5.91385186e-01 -1.51398206e+00 2.12754056e-01 -5.86167812e-01
-2.42990822e-01 1.04160190e+00 5.03577948e-01 -3.19529980e-01
6.61169469e-01 3.16368073e-01 1.69171259e-01 -3.67976010e-01
-4.61078197e-01 -5.80640852e-01 -1.63527969e-02 8.83652031e-01
5.56500442e-02 -2.48972937e-01 8.39038938e-02 -1.04361415e-01
-2.91815400e-01 2.03746915e-01 3.40961784e-01 8.84682417e-01
-4.22913700e-01 -1.10383451e+00 -4.88301404e-02 3.45354676e-01
4.61030975e-02 -1.91246584e-01 -1.67997688e-01 8.10077488e-01
2.74502456e-01 7.38492668e-01 3.96058768e-01 -1.72725245e-01
1.60577700e-01 -1.10985391e-01 6.48591816e-01 -7.72394896e-01
6.24389164e-02 -1.52799100e-01 7.17364028e-02 -8.76718760e-01
-4.38682646e-01 -6.22300863e-01 -1.11158550e+00 -3.14253181e-01
-2.03081861e-01 -1.33250609e-01 1.20548463e+00 1.06046700e+00
4.78239328e-01 5.57433844e-01 3.40306640e-01 -1.19541299e+00
-3.70175064e-01 -9.33123112e-01 -4.64732289e-01 5.27377166e-02
4.69418228e-01 -7.64882028e-01 -9.48420390e-02 -2.54785866e-01] | [8.366994857788086, -2.5291032791137695] |
3d3d5c20-bc04-40c3-b94d-4e0594d45fe3 | imprecise-label-learning-a-unified-framework | 2305.12715 | null | https://arxiv.org/abs/2305.12715v2 | https://arxiv.org/pdf/2305.12715v2.pdf | Imprecise Label Learning: A Unified Framework for Learning with Various Imprecise Label Configurations | In this paper, we introduce the imprecise label learning (ILL) framework, a unified approach to handle various imprecise label configurations, which are commonplace challenges in machine learning tasks. ILL leverages an expectation-maximization (EM) algorithm for the maximum likelihood estimation (MLE) of the imprecise label information, treating the precise labels as latent variables. Compared to previous versatile methods attempting to infer correct labels from the imprecise label information, our ILL framework considers all possible labeling imposed by the imprecise label information, allowing a unified solution to deal with any imprecise labels. With comprehensive experimental results, we demonstrate that ILL can seamlessly adapt to various situations, including partial label learning, semi-supervised learning, noisy label learning, and a mixture of these settings. Notably, our simple method surpasses the existing techniques for handling imprecise labels, marking the first unified framework with robust and effective performance across various imprecise labels. We believe that our approach has the potential to significantly enhance the performance of machine learning models on tasks where obtaining precise labels is expensive and complicated. We hope our work will inspire further research on this topic with an open-source codebase release. | ['Bhiksha Raj', 'Rita Singh', 'Masashi Sugiyama', 'Xing Xie', 'Yidong Wang', 'Ran Tao', 'Jindong Wang', 'Ankit Shah', 'Hao Chen'] | 2023-05-22 | null | null | null | null | ['learning-with-noisy-labels', 'partial-label-learning', 'learning-with-noisy-labels'] | ['computer-vision', 'methodology', 'natural-language-processing'] | [ 3.05458248e-01 1.86349779e-01 -5.72003126e-01 -1.05802655e+00
-1.33192849e+00 -9.52579260e-01 6.35780215e-01 2.52326012e-01
-3.83475959e-01 7.75706172e-01 -3.72121274e-01 -2.92568892e-01
-3.59405249e-01 -3.99399936e-01 -7.47521520e-01 -4.90788549e-01
4.64277267e-02 6.95752740e-01 -1.73955351e-01 2.34973267e-01
2.30322748e-01 2.55955964e-01 -1.53624821e+00 3.60179991e-01
5.89577258e-01 1.13617158e+00 -4.52314734e-01 1.91512898e-01
6.75400794e-02 8.32455337e-01 -4.43721354e-01 -6.47612214e-01
1.60029233e-01 3.06678385e-01 -1.12996459e+00 -2.96803880e-02
3.86817575e-01 -1.91828862e-01 4.34928626e-01 1.08838785e+00
2.56046265e-01 6.26875386e-02 9.92334366e-01 -1.55583417e+00
-4.99189109e-01 7.01897204e-01 -5.21020949e-01 -2.01204926e-01
6.74363613e-01 -3.77952158e-01 1.05265677e+00 -1.03129172e+00
4.34530318e-01 1.42020631e+00 1.14291584e+00 3.14609140e-01
-1.27141976e+00 -6.49075210e-01 2.44114920e-01 2.92697877e-01
-1.71846282e+00 -6.10814333e-01 5.29058695e-01 -4.27562535e-01
7.17030525e-01 2.90671378e-01 -2.94812977e-01 1.05089986e+00
8.24336782e-02 9.19465661e-01 1.27543759e+00 -8.27093720e-01
5.71769953e-01 9.71519426e-02 3.24243158e-01 5.35820305e-01
1.78407714e-01 1.90780729e-01 -4.27555561e-01 -6.43847167e-01
4.35661003e-02 4.43542860e-02 -1.59343705e-01 -4.71304744e-01
-1.05564666e+00 6.19425952e-01 -2.67664832e-03 -1.26506999e-01
-6.12106323e-02 2.74078667e-01 4.49928045e-01 3.69194478e-01
8.19111824e-01 3.15804392e-01 -9.37805116e-01 -4.91262898e-02
-1.11282897e+00 2.32975364e-01 9.18369472e-01 1.20831943e+00
7.16156662e-01 -5.62422276e-01 -4.37045790e-04 8.91631186e-01
8.22134078e-01 2.05259055e-01 6.59985393e-02 -1.33477604e+00
2.90555447e-01 3.46104026e-01 5.58910847e-01 -5.69997311e-01
-6.67178750e-01 -4.73779477e-02 -4.39658791e-01 1.53684825e-01
6.11422241e-01 -6.17437139e-02 -1.02709067e+00 1.72328520e+00
4.52157199e-01 3.23468536e-01 -2.02939004e-01 5.71238518e-01
4.50878888e-01 2.95954973e-01 4.30150121e-01 -5.11354327e-01
1.41928947e+00 -6.16142571e-01 -9.68602419e-01 -2.29881361e-01
8.72006893e-01 -7.61556923e-01 6.13778949e-01 6.36023819e-01
-8.26514184e-01 -3.03892381e-02 -1.02287626e+00 -2.55158305e-01
-4.61129874e-01 -6.76561072e-02 9.01600003e-01 8.56141031e-01
-9.60906267e-01 5.82975090e-01 -8.30547690e-01 3.10721528e-03
3.63377601e-01 4.46108133e-01 -3.51464719e-01 -5.96396446e-01
-1.32606578e+00 8.83097947e-01 7.44022667e-01 1.42136812e-01
-6.55592144e-01 -5.82020760e-01 -1.13540590e+00 -2.28384286e-01
8.36048603e-01 -4.90374058e-01 1.80555153e+00 -4.66888368e-01
-1.30282748e+00 1.02166426e+00 -2.44247466e-01 -4.98103797e-02
4.43859220e-01 -6.83093742e-02 -2.75997698e-01 -3.43838334e-01
1.31901950e-01 3.96848202e-01 7.10171938e-01 -1.35534179e+00
-6.64503574e-01 -4.24930841e-01 7.39517109e-03 2.15812460e-01
2.54486740e-01 3.13912481e-01 -1.28666550e-01 -5.15057921e-01
6.65513158e-01 -1.06179976e+00 3.49022895e-02 -4.65487549e-03
-2.79299974e-01 -7.17944682e-01 4.42847431e-01 -1.18090458e-01
1.06474745e+00 -2.07656980e+00 -1.21295199e-01 3.06080908e-01
1.41884044e-01 2.86908951e-02 3.56426209e-01 3.66310447e-01
-1.46142617e-01 3.68476689e-01 -5.24407744e-01 -7.32364058e-01
5.34209967e-01 5.09312630e-01 -2.99202234e-01 6.46569967e-01
-6.43678457e-02 8.75640690e-01 -1.24419034e+00 -6.59846723e-01
5.34654967e-03 2.73238629e-01 1.63789034e-01 -3.73651274e-02
-3.65839213e-01 5.19377850e-02 -4.04333025e-01 1.02561593e+00
8.83168221e-01 -1.15689173e-01 2.65155286e-01 -9.75173153e-03
3.12014937e-01 1.15846187e-01 -1.53085649e+00 1.68339908e+00
-4.36720997e-01 -3.24497372e-02 2.06062064e-01 -8.50381672e-01
6.96805179e-01 7.63128817e-01 4.41651851e-01 -1.14690393e-01
5.83329163e-02 3.51472288e-01 -9.25266743e-01 -5.31710505e-01
1.44218847e-01 -7.27719128e-01 -5.44313550e-01 7.99986362e-01
1.23852186e-01 -2.07684070e-01 -2.25903820e-02 6.28283620e-02
9.00037348e-01 4.15662915e-01 6.33142948e-01 -9.79733989e-02
2.00091854e-01 -5.34956276e-01 7.80570388e-01 9.09409523e-01
-5.59833348e-01 6.12926245e-01 3.26510638e-01 -5.13600528e-01
-5.22147894e-01 -1.10691082e+00 -7.42606163e-01 1.52810991e+00
1.04636904e-02 -4.33301061e-01 -3.99185836e-01 -1.06198263e+00
1.35094628e-01 8.70709836e-01 -4.27836001e-01 -6.95931986e-02
-8.18351191e-03 -9.48326945e-01 7.89277077e-01 3.90025854e-01
2.30395854e-01 -7.00629354e-01 -6.11469187e-02 1.51825279e-01
-5.05813479e-01 -9.92822707e-01 -2.41641492e-01 4.72858787e-01
-7.15758085e-01 -1.02325904e+00 -8.88584182e-02 -7.33908117e-01
4.71580774e-01 -1.44496784e-01 1.33184433e+00 -2.91576944e-02
1.86049029e-01 3.36357057e-01 -4.69060361e-01 -4.23452437e-01
-3.75146747e-01 -1.44553125e-01 3.03462446e-01 -8.55173543e-02
7.80883670e-01 -3.27741176e-01 -8.56380686e-02 5.03880203e-01
-1.22088313e+00 -3.77453864e-01 1.60496876e-01 7.19702542e-01
7.42124438e-01 4.69725966e-01 7.72355795e-01 -1.44699740e+00
6.48441613e-01 -1.02995646e+00 -3.90515029e-01 6.77469850e-01
-1.10842741e+00 4.67079610e-01 2.85654396e-01 -1.31120235e-01
-1.36269832e+00 3.93631309e-01 -1.80298984e-01 5.63083589e-02
-5.74963033e-01 4.88918066e-01 -3.32844585e-01 6.47735000e-02
8.27831149e-01 -4.25358295e-01 -5.03101826e-01 -6.58113480e-01
6.62377834e-01 1.04491782e+00 6.64116502e-01 -1.12580001e+00
3.32968563e-01 4.82616305e-01 2.05093056e-01 2.09062144e-01
-1.28919458e+00 -7.36785531e-01 -7.47153938e-01 -2.32339483e-02
2.49861449e-01 -8.96600902e-01 -5.76758564e-01 5.65357447e-01
-1.02239168e+00 6.55264407e-02 -6.26551080e-03 2.74624139e-01
-7.68186450e-01 6.81938946e-01 -7.96540856e-01 -9.72116411e-01
-7.57803544e-02 -1.11453366e+00 1.45269668e+00 -7.51565546e-02
-4.05927807e-01 -1.23911524e+00 9.03664604e-02 5.17815948e-01
-2.18232665e-02 3.88648212e-01 6.93886757e-01 -8.75781357e-01
-3.14420849e-01 -5.24979591e-01 -6.55396804e-02 2.01783121e-01
1.08332299e-01 -7.40145147e-02 -1.14448404e+00 -2.88661093e-01
2.43948072e-01 -8.34058225e-01 8.38146687e-01 9.11447257e-02
7.86840260e-01 -2.40035444e-01 -4.07526970e-01 3.91537696e-01
1.41867125e+00 -1.50884509e-01 2.62753099e-01 3.54376107e-01
4.95501816e-01 5.54964483e-01 1.00853586e+00 6.17182136e-01
7.33866930e-01 7.35498488e-01 4.53519911e-01 3.34890008e-01
3.89567554e-01 -6.60832524e-02 5.65618351e-02 2.32756928e-01
3.11388791e-01 -1.45629466e-01 -1.10284245e+00 3.44419271e-01
-2.10550332e+00 -6.83820605e-01 -8.84237364e-02 2.14981008e+00
1.37957096e+00 -1.43535480e-01 -3.26881468e-01 3.46744388e-01
6.96924269e-01 -5.75454384e-02 -4.30626780e-01 -5.53543031e-01
3.16101581e-01 -5.96511178e-02 4.00310874e-01 6.48138642e-01
-1.43208730e+00 6.96777046e-01 7.49153805e+00 9.56596673e-01
-3.68041724e-01 2.87460804e-01 4.44462419e-01 -6.98330179e-02
-4.44738925e-01 1.66034505e-01 -7.02954590e-01 6.37574375e-01
1.00046515e+00 2.56126672e-01 5.38836062e-01 8.75101447e-01
-1.93831429e-01 -2.44897604e-01 -1.66646302e+00 8.58783424e-01
2.35277750e-02 -5.22345483e-01 -6.02161169e-01 -1.25200599e-01
9.43340421e-01 -5.08614704e-02 1.41425282e-01 5.15245616e-01
8.31220031e-01 -1.00857246e+00 8.15307736e-01 6.63001955e-01
9.41725492e-01 -8.40799093e-01 1.09120297e+00 8.41648161e-01
-7.60935724e-01 -1.80939406e-01 -2.72707909e-01 -2.18207747e-01
1.98004201e-01 1.38752282e+00 -9.00047600e-01 6.08338058e-01
6.34962082e-01 4.58823383e-01 -5.25447547e-01 8.85981798e-01
-4.64614868e-01 4.47766393e-01 -7.11201966e-01 6.42385721e-01
-1.64608568e-01 1.76615730e-01 1.35648280e-01 1.25916564e+00
3.05030327e-02 4.10484038e-02 3.93834084e-01 6.81323826e-01
-2.97745734e-01 -1.15796372e-01 -5.20137727e-01 4.15540904e-01
9.12788332e-01 1.09811556e+00 -7.36311316e-01 -3.09462607e-01
-2.11340144e-01 7.57304966e-01 5.70294678e-01 4.07556474e-01
-6.38927579e-01 -1.17570207e-01 2.92769551e-01 -4.88756150e-01
-2.07760006e-01 -1.41107114e-02 -6.58199966e-01 -1.18051016e+00
1.36620238e-01 -7.10065305e-01 8.50899339e-01 -6.91511154e-01
-1.55167961e+00 9.13052931e-02 4.23208386e-01 -8.68667960e-01
-6.78983688e-01 -4.74376231e-01 -4.50417176e-02 6.24386907e-01
-1.58237541e+00 -1.22568619e+00 2.14214996e-01 2.20793739e-01
2.15503618e-01 4.10391301e-01 9.73305345e-01 1.68724701e-01
-4.25227642e-01 7.72475660e-01 3.56108338e-01 -3.20739478e-01
1.24757290e+00 -1.54425538e+00 1.23307355e-01 6.55742466e-01
2.39735559e-01 7.13497460e-01 6.72972560e-01 -6.72937870e-01
-7.98998415e-01 -1.05675387e+00 1.32153058e+00 -1.19557357e+00
3.96801412e-01 -3.34332436e-01 -8.59203935e-01 1.31803250e+00
-3.25054765e-01 2.68128723e-01 1.11546743e+00 4.75493580e-01
-9.65503573e-01 2.02643856e-01 -1.70076597e+00 9.01464596e-02
9.57090497e-01 -6.36990845e-01 -7.45339751e-01 5.47048569e-01
5.43379605e-01 -5.67270935e-01 -1.27030563e+00 7.29509234e-01
4.49741125e-01 -8.07759821e-01 9.05159175e-01 -3.90689671e-01
-1.09752566e-01 -5.01889467e-01 -3.19533795e-01 -1.30088735e+00
-3.94863605e-01 -2.78559387e-01 -4.10051823e-01 1.51286328e+00
5.06360948e-01 -5.12872219e-01 4.44582075e-01 1.36005819e+00
-7.35004200e-03 -6.96242154e-01 -1.11937845e+00 -6.63236320e-01
7.72318542e-02 -1.05996299e+00 8.38884473e-01 1.18579435e+00
3.79144698e-01 -1.32487267e-01 -3.76838654e-01 2.73876339e-01
9.40035284e-01 -1.38944358e-01 -8.41417164e-02 -1.44599724e+00
-1.24596171e-01 -6.15516352e-03 -3.08111697e-01 -7.08414912e-01
6.36240840e-01 -1.02850723e+00 4.23703045e-01 -1.29398286e+00
2.32670307e-01 -8.13065946e-01 -5.19593120e-01 1.01260746e+00
-3.18006486e-01 5.20997167e-01 -2.14745253e-02 3.30716789e-01
-1.31947184e+00 1.41323730e-01 7.07158089e-01 -4.18333858e-02
1.51926741e-01 3.60045373e-01 -8.36324692e-01 9.23623621e-01
7.35695839e-01 -8.80844355e-01 -4.94417757e-01 -3.49059552e-01
7.67181933e-01 -1.78934246e-01 8.99331719e-02 -5.33105850e-01
3.93050820e-01 -2.95727313e-01 1.37767226e-01 -1.68184504e-01
1.19112402e-01 -1.09566951e+00 2.27401450e-01 -3.98205370e-01
-5.70510864e-01 -3.74995321e-01 -8.37269276e-02 7.71949172e-01
-4.36493233e-02 -8.17770600e-01 5.27824759e-01 -2.09075242e-01
-7.63610184e-01 -3.19846384e-02 -2.91684121e-01 1.15800679e-01
1.14916551e+00 1.19360626e-01 -1.58066049e-01 -2.13272467e-01
-9.37344193e-01 5.16402006e-01 6.21967971e-01 2.65260547e-01
3.64226341e-01 -1.46738851e+00 -6.14455044e-01 1.15706317e-01
3.70898932e-01 4.11874980e-01 -4.94524874e-02 5.05496681e-01
-1.08442597e-01 3.85709047e-01 4.47105944e-01 -6.56719148e-01
-9.17057991e-01 8.61023962e-01 2.16031760e-01 -3.38543534e-01
-2.69455463e-02 8.28667521e-01 -2.77842611e-01 -9.59289193e-01
4.45396066e-01 -1.99288502e-01 1.26510449e-02 4.59223203e-02
5.62783420e-01 7.14861691e-01 4.52022880e-01 -6.79763675e-01
-5.00341535e-01 4.64118958e-01 -2.67200947e-01 -6.56917542e-02
8.61642480e-01 -5.12301505e-01 -4.21524853e-01 7.93637037e-01
1.14401817e+00 -3.34350318e-01 -1.30693042e+00 -6.80205405e-01
6.95431948e-01 -3.11250538e-01 1.13601513e-01 -1.23337662e+00
-5.03415048e-01 3.47081840e-01 2.86735356e-01 7.72791877e-02
1.06210029e+00 2.33768955e-01 4.31352288e-01 4.67832297e-01
9.32414949e-01 -1.21765161e+00 -5.45091629e-01 4.27804470e-01
3.77913892e-01 -1.45536256e+00 2.50009507e-01 -5.40973842e-01
-5.18829942e-01 8.91084254e-01 1.29368082e-01 5.77089965e-01
8.11114550e-01 5.36442161e-01 2.53128856e-01 -1.28677920e-01
-8.74632657e-01 1.22258395e-01 1.96411341e-01 7.10416496e-01
3.73026371e-01 2.80387908e-01 -2.25283384e-01 7.52215803e-01
1.94601923e-01 3.14052939e-01 4.17124778e-01 1.20734644e+00
-4.51215595e-01 -1.40753365e+00 -7.72646010e-01 3.46787453e-01
-6.98541582e-01 9.58463624e-02 -2.51307100e-01 2.60576427e-01
5.06762445e-01 1.51436663e+00 -3.55554968e-01 -1.16702959e-01
2.06743076e-01 9.11296129e-01 3.96463335e-01 -9.10723150e-01
-1.40807405e-01 -3.75955969e-01 6.64740354e-02 -5.26441157e-01
-4.94221330e-01 -6.62767410e-01 -1.41789341e+00 -2.03832835e-01
-6.06361449e-01 3.40661071e-02 8.14170241e-01 1.35444701e+00
1.56940669e-01 -1.01294115e-01 5.52989841e-01 -8.36803555e-01
-1.12261236e+00 -7.90445387e-01 -8.80163252e-01 6.49584115e-01
3.63706708e-01 -9.78398919e-01 -6.50442600e-01 1.65902913e-01] | [9.329264640808105, 4.193579196929932] |
7989e37a-0875-406e-8ed5-e57a5241c2ce | class-conditional-embeddings-for-music-source | 1811.03076 | null | http://arxiv.org/abs/1811.03076v1 | http://arxiv.org/pdf/1811.03076v1.pdf | Class-conditional embeddings for music source separation | Isolating individual instruments in a musical mixture has a myriad of
potential applications, and seems imminently achievable given the levels of
performance reached by recent deep learning methods. While most musical source
separation techniques learn an independent model for each instrument, we
propose using a common embedding space for the time-frequency bins of all
instruments in a mixture inspired by deep clustering and deep attractor
networks. Additionally, an auxiliary network is used to generate parameters of
a Gaussian mixture model (GMM) where the posterior distribution over GMM
components in the embedding space can be used to create a mask that separates
individual sources from a mixture. In addition to outperforming a
mask-inference baseline on the MUSDB-18 dataset, our embedding space is easily
interpretable and can be used for query-based separation. | ['Shrikant Venkataramani', 'Prem Seetharaman', 'Jonathan Le Roux', 'Gordon Wichern'] | 2018-11-07 | null | null | null | null | ['music-source-separation'] | ['music'] | [ 6.07177727e-02 8.65306109e-02 -7.20615685e-03 -2.85421684e-02
-1.10233271e+00 -9.67332304e-01 5.67511082e-01 -5.15087619e-02
-1.80143267e-01 1.47443205e-01 4.78316873e-01 -4.40558866e-02
-4.69275892e-01 -2.76545525e-01 -4.51445073e-01 -6.99623466e-01
-1.36679292e-01 6.21112823e-01 -2.03928664e-01 -7.27668330e-02
-7.87399709e-02 4.16807532e-01 -1.51338744e+00 2.77836084e-01
7.33105659e-01 9.02512491e-01 6.63899556e-02 8.27450514e-01
-4.65335473e-02 4.36032832e-01 -1.05719244e+00 -3.85933638e-01
5.70446551e-01 -7.67590404e-01 -5.54473102e-01 6.92445189e-02
5.90558112e-01 -1.08934581e-01 -3.06390405e-01 1.00235176e+00
5.84303916e-01 2.83192664e-01 9.24721658e-01 -1.03685224e+00
-3.88746321e-01 9.91154194e-01 -2.60726124e-01 6.11587204e-02
1.17183551e-01 3.76184322e-02 1.14362442e+00 -5.64359725e-01
1.87078118e-01 1.09374762e+00 5.29941261e-01 3.55562091e-01
-1.75420332e+00 -7.05414414e-01 -7.68370330e-02 -3.10138408e-02
-1.25203276e+00 -6.29996777e-01 9.83341217e-01 -4.97562468e-01
4.49814707e-01 4.76842225e-01 8.53097141e-01 1.10948825e+00
-2.08392397e-01 8.80415499e-01 6.35844111e-01 -2.86549568e-01
2.37933099e-01 1.31143168e-01 -2.20088184e-01 3.48329097e-01
3.03624687e-03 -2.20397562e-01 -6.25480592e-01 -2.94678420e-01
1.02031410e+00 -1.04125492e-01 -2.43194938e-01 -4.85059112e-01
-1.34625673e+00 7.78695047e-01 9.43477452e-02 2.84184784e-01
-1.98665619e-01 2.29037464e-01 2.14685157e-01 1.69302776e-01
1.84868470e-01 9.40532327e-01 -9.74703133e-02 -3.71039286e-03
-1.49524641e+00 3.45368803e-01 8.99542332e-01 5.85332453e-01
5.33207715e-01 3.37687075e-01 -2.56442249e-01 8.54316831e-01
3.76734525e-01 1.40573800e-01 6.99170232e-01 -1.47867072e+00
2.35943452e-01 4.41372812e-01 3.33366506e-02 -8.86769891e-01
-8.27410370e-02 -7.96724677e-01 -5.88251472e-01 3.87812287e-01
7.14636564e-01 -1.89292818e-01 -8.38976800e-01 1.91986513e+00
1.52613789e-01 6.19083166e-01 3.14291567e-02 8.85411739e-01
4.86167967e-01 4.51854885e-01 -3.55139315e-01 1.22249141e-01
1.25958633e+00 -9.23879445e-01 -5.01292109e-01 -3.39764357e-01
1.45482689e-01 -7.85829365e-01 8.64091814e-01 6.27776682e-01
-1.19258296e+00 -8.55334818e-01 -1.15991151e+00 -1.03634454e-01
-1.99100897e-01 2.25166425e-01 6.08012617e-01 6.60605669e-01
-8.47898245e-01 7.53766298e-01 -8.25386405e-01 2.21675292e-01
3.49243611e-01 3.94627422e-01 -2.18944967e-01 4.49112922e-01
-9.07652617e-01 4.80446458e-01 4.27569598e-01 5.69921546e-02
-1.03292751e+00 -8.18231642e-01 -7.81091094e-01 3.06269705e-01
1.72533080e-01 -7.00327218e-01 1.03899801e+00 -1.05095255e+00
-1.63520145e+00 7.49689877e-01 4.85131145e-02 -5.27611136e-01
3.68227094e-01 -2.84794986e-01 -4.68595415e-01 4.23301935e-01
-4.64177728e-02 7.55680621e-01 1.36904681e+00 -1.14316034e+00
-5.12733400e-01 -1.49066851e-01 -1.94735408e-01 2.63045132e-01
-5.15987039e-01 -1.58249706e-01 -5.11553645e-01 -9.83146906e-01
2.43877873e-01 -1.06515253e+00 -1.61199480e-01 -2.38413140e-01
-7.31236279e-01 -6.95402771e-02 5.07850707e-01 -9.45640922e-01
1.27816284e+00 -2.63265467e+00 6.40300393e-01 2.35795349e-01
2.82123238e-01 2.00333837e-02 -2.54829735e-01 1.79866225e-01
-1.88769862e-01 -5.98399229e-02 -2.49270201e-01 -7.64484942e-01
3.76711488e-01 -1.78782597e-01 -6.74426854e-01 3.78852129e-01
1.52186275e-01 5.17623186e-01 -7.26636410e-01 -2.77537495e-01
9.89903286e-02 5.76970100e-01 -6.26587808e-01 2.58312672e-01
-3.59726489e-01 5.79631031e-01 7.57185444e-02 3.09322983e-01
3.91701669e-01 3.01188026e-02 2.51699328e-01 -1.46568984e-01
3.86333019e-01 5.89586914e-01 -1.60314631e+00 2.15679932e+00
-2.04125404e-01 6.61440015e-01 4.04260248e-01 -8.71800303e-01
9.35919881e-01 4.91725057e-01 7.31871009e-01 1.43120348e-01
1.24013036e-01 1.19400471e-01 3.56455773e-01 -3.29554975e-02
6.22027099e-01 -1.64567471e-01 -2.95816988e-01 6.24328196e-01
5.87777615e-01 -2.56241828e-01 8.42481777e-02 1.07887775e-01
8.96714509e-01 6.72595501e-02 -4.67596054e-01 -1.61939725e-01
1.23248227e-01 -4.72691536e-01 4.21022207e-01 5.53380787e-01
-5.25913723e-02 1.01237035e+00 4.10932422e-01 4.67603654e-02
-8.64710152e-01 -1.57347167e+00 -6.37021065e-02 9.49662685e-01
-9.31625888e-02 -6.73515379e-01 -6.78086996e-01 -4.61034298e-01
1.02966286e-01 6.95004761e-01 -3.54164451e-01 -3.97492439e-01
-3.78308833e-01 -4.43024218e-01 8.72095585e-01 3.53121579e-01
3.18724872e-03 -8.05642366e-01 -3.35382700e-01 2.70753324e-01
-3.70593667e-01 -8.49330425e-01 -5.96832156e-01 6.13072515e-01
-7.92456686e-01 -8.60021532e-01 -5.80658138e-01 -6.89553678e-01
7.87040032e-03 1.24508562e-02 1.17423260e+00 -4.49987024e-01
-1.31102815e-01 4.41715509e-01 2.77491003e-01 -5.06366968e-01
-5.08743942e-01 2.46021077e-01 1.89727694e-01 2.71961093e-01
3.94928962e-01 -9.71944511e-01 -6.33293211e-01 5.05473791e-03
-1.02586913e+00 -3.08552772e-01 3.87371868e-01 3.74651134e-01
4.05766815e-01 3.69010538e-01 5.22523522e-01 -2.52298683e-01
6.37806296e-01 -5.96179962e-01 -2.20465928e-01 -1.79203004e-01
-2.82504439e-01 1.51108414e-01 5.54832160e-01 -8.40699434e-01
-7.35278249e-01 7.48837739e-02 -1.39086228e-02 -8.20147097e-01
-3.79652917e-01 2.46145219e-01 -4.27418321e-01 6.71693981e-01
5.32975495e-01 4.57203202e-02 1.85852364e-01 -8.51522148e-01
7.31519520e-01 6.21888459e-01 1.06531322e+00 -5.35657346e-01
7.86539674e-01 4.61090714e-01 -3.23263556e-01 -6.88519835e-01
-7.25468993e-01 -4.53989267e-01 -6.96728587e-01 1.29356757e-01
1.04041839e+00 -1.12907481e+00 -6.05011225e-01 1.17615536e-01
-8.47844481e-01 -2.13181362e-01 -6.75893962e-01 6.14601016e-01
-6.70700967e-01 2.21103206e-01 -5.80131769e-01 -7.72165298e-01
2.14429498e-02 -1.01726556e+00 1.03632009e+00 2.62176275e-01
-8.52401078e-01 -1.00248039e+00 2.59059280e-01 4.52820450e-01
-1.02797961e-02 8.41015428e-02 8.95040214e-01 -1.01645744e+00
-6.06326759e-01 -3.58374000e-01 5.20000041e-01 5.70422232e-01
3.61519307e-01 6.64634854e-02 -1.25565100e+00 -2.13459820e-01
1.79588735e-01 -1.64825574e-01 1.12601912e+00 4.80542362e-01
1.10833895e+00 -2.10240290e-01 -3.39841306e-01 7.56520629e-01
9.31587934e-01 7.25610852e-02 5.68097830e-01 1.00169636e-01
8.38887036e-01 5.51280797e-01 -1.79125041e-01 1.06721856e-01
4.46877070e-02 6.70760989e-01 1.94470614e-01 -1.60565645e-01
-2.16482058e-01 -3.58427614e-01 5.31699419e-01 8.70151520e-01
2.35699743e-01 -9.34246406e-02 -4.47359025e-01 6.69391572e-01
-1.60193121e+00 -1.31107557e+00 4.26475674e-01 2.36746955e+00
8.86052787e-01 2.40381882e-01 6.79524362e-01 3.69156092e-01
6.00664556e-01 3.81836385e-01 -6.32172883e-01 -2.14437306e-01
-3.28003645e-01 4.45954353e-01 -7.01682642e-02 3.30840379e-01
-1.27411890e+00 5.77810585e-01 7.13457298e+00 9.25513387e-01
-1.01181269e+00 -2.01434359e-01 4.31967854e-01 -5.64167261e-01
-3.78165483e-01 6.65880889e-02 -6.32377088e-01 5.69004297e-01
1.09239483e+00 1.45325869e-01 5.34042239e-01 4.39204603e-01
-1.19433992e-01 1.47869080e-01 -1.40178692e+00 1.08330011e+00
1.28556341e-01 -1.13996935e+00 5.10806637e-03 3.13963085e-01
4.81244385e-01 -9.70607847e-02 4.46098000e-01 3.50581467e-01
4.76688981e-01 -1.35894680e+00 9.09999609e-01 5.32457113e-01
5.48021376e-01 -8.44826102e-01 7.57218152e-02 4.90527928e-01
-8.12898040e-01 -2.72764474e-01 -2.82422870e-01 -5.13244085e-02
-1.09913394e-01 5.19753277e-01 -7.34220982e-01 4.65322882e-01
5.27244925e-01 4.57590878e-01 -5.37019968e-01 8.98972332e-01
-9.47167948e-02 8.74769270e-01 -4.27837819e-01 5.14087081e-01
-6.08291440e-02 -5.93251467e-01 8.53738189e-01 1.18637586e+00
2.82362968e-01 -5.95588565e-01 1.58153966e-01 1.23459578e+00
-1.39933079e-01 -1.68227851e-01 -3.77173454e-01 -3.82057905e-01
5.38633406e-01 1.43933821e+00 -8.42639983e-01 -3.07326257e-01
7.42208958e-02 1.09227753e+00 2.41053656e-01 3.79384279e-01
-7.13800311e-01 -3.45396578e-01 1.08969748e+00 -9.97273400e-02
6.45182252e-01 -3.24831635e-01 -3.35451782e-01 -1.22613919e+00
-7.32688084e-02 -1.12453938e+00 3.67706388e-01 -5.01950681e-01
-1.25651538e+00 5.19820213e-01 -1.70773774e-01 -1.45731437e+00
-4.14538234e-01 -3.74005735e-01 -7.48547792e-01 1.15767360e+00
-7.37352848e-01 -8.08702469e-01 2.06496343e-01 4.04717773e-01
3.06109309e-01 -4.59916234e-01 1.04290938e+00 1.64500177e-01
-5.22351861e-01 5.54248810e-01 2.40204737e-01 4.13279593e-01
8.99977744e-01 -1.50685728e+00 2.22173020e-01 8.18244100e-01
1.17826402e+00 7.94109344e-01 7.19297409e-01 -3.09847802e-01
-1.15242863e+00 -9.24393177e-01 3.32200944e-01 -8.75272334e-01
5.32508969e-01 -7.11211503e-01 -8.60617876e-01 6.72388673e-01
3.70003998e-01 -3.69779050e-01 1.15806866e+00 2.87283719e-01
-5.65023422e-01 -2.29739189e-01 -6.36955142e-01 7.24653184e-01
7.39987731e-01 -8.65824103e-01 -6.46521389e-01 6.48166239e-02
5.06592453e-01 -2.83567786e-01 -8.69214118e-01 7.63048083e-02
6.06584132e-01 -1.17630613e+00 1.32348382e+00 -6.72139406e-01
4.29111362e-01 -4.24964458e-01 -2.09101364e-01 -1.63199043e+00
-5.24706721e-01 -9.20178533e-01 -4.06885147e-01 1.55533862e+00
3.15617561e-01 -1.17043346e-01 7.53186822e-01 4.79059994e-01
-2.11572535e-02 -3.86847198e-01 -8.63178909e-01 -9.04532075e-01
2.35579222e-01 -5.99104285e-01 7.53434777e-01 8.89248431e-01
-5.58548346e-02 6.62421167e-01 -2.75118500e-01 2.92458802e-01
5.98772585e-01 3.85258228e-01 8.04111958e-01 -1.32081687e+00
-1.01124024e+00 -9.39359307e-01 -3.12729031e-01 -1.22467363e+00
3.59473288e-01 -1.22739363e+00 -1.08732171e-01 -1.34213912e+00
-4.01317403e-02 -4.00114387e-01 -5.97721219e-01 1.23602197e-01
-2.18287960e-01 3.11863035e-01 4.22010183e-01 3.17544162e-01
-3.01858366e-01 5.81231773e-01 7.35349059e-01 -4.39470321e-01
-4.59157288e-01 1.58818841e-01 -9.30139601e-01 8.58540893e-01
6.70835435e-01 -4.33922440e-01 -4.57362264e-01 -3.23464394e-01
-2.16313358e-02 3.53071727e-02 3.56422335e-01 -1.22014546e+00
-3.36977653e-02 3.48669797e-01 5.85926592e-01 -4.72419798e-01
7.44744837e-01 -6.74225748e-01 4.08829063e-01 1.31201699e-01
-4.76417542e-01 -4.76458520e-01 2.58888930e-01 6.57318234e-01
-2.96194226e-01 -2.62623668e-01 5.97295463e-01 1.29012063e-01
-3.25216912e-02 -2.96837203e-02 -3.88618082e-01 5.12680002e-02
5.53806782e-01 -3.67092520e-01 2.23918572e-01 -5.91523051e-01
-8.03711891e-01 -1.44685417e-01 2.67782182e-01 5.22624731e-01
2.12125033e-01 -1.51912665e+00 -6.25521839e-01 2.09497318e-01
-1.58737019e-01 -8.93353764e-03 1.53952673e-01 6.03696764e-01
-2.71641742e-02 -7.30711520e-02 -6.25519827e-02 -6.35080934e-01
-9.06762600e-01 6.47909284e-01 3.72065783e-01 -3.53409424e-02
-5.35721242e-01 9.65396285e-01 2.71546185e-01 -3.91962588e-01
4.23992217e-01 -3.02125812e-01 -1.70956198e-02 2.42538065e-01
4.50896353e-01 3.05430621e-01 -1.45773172e-01 -4.99497682e-01
-1.79045558e-01 2.07689717e-01 3.47195983e-01 -6.26480103e-01
1.28213644e+00 1.59167483e-01 6.60393015e-02 9.85727489e-01
1.15697002e+00 6.65668488e-01 -1.50290310e+00 -3.18458937e-02
8.73470455e-02 -1.04658768e-01 -1.66361868e-01 -6.67552829e-01
-8.12460184e-01 9.21713173e-01 3.98936898e-01 5.34165978e-01
9.74772930e-01 1.07845940e-01 7.52415955e-01 -2.83475895e-03
-3.30216080e-01 -1.02675307e+00 2.05410957e-01 3.13836277e-01
7.52556145e-01 -9.01623368e-01 -3.10664862e-01 -3.03710531e-02
-4.16330218e-01 1.04259527e+00 2.07413405e-01 -4.20041621e-01
6.43408179e-01 5.63115835e-01 3.23567867e-01 -6.37751967e-02
-3.61062020e-01 -2.39703301e-02 7.77334929e-01 5.37916541e-01
4.02957976e-01 -3.35471295e-02 4.65673357e-01 1.00095189e+00
-6.98091030e-01 -6.30797446e-01 3.43954235e-01 3.34630907e-01
-3.04499090e-01 -1.20790553e+00 -5.84435225e-01 3.31081122e-01
-5.71566522e-01 -9.74873677e-02 -5.13757110e-01 3.35059673e-01
3.75514746e-01 1.02789986e+00 3.49163383e-01 -4.74316061e-01
-7.95697197e-02 5.20319045e-01 5.83656967e-01 -7.31727481e-01
-5.54240823e-01 8.73465776e-01 -1.92870945e-01 -2.69308150e-01
-2.43218690e-01 -7.11109996e-01 -1.11900473e+00 1.71654463e-01
-2.12648496e-01 3.95035863e-01 5.76112390e-01 7.53830194e-01
4.60516274e-01 5.75328290e-01 4.83648032e-01 -9.30889666e-01
-5.10398269e-01 -1.12369788e+00 -9.24945414e-01 5.08437455e-01
3.85603637e-01 -3.83976340e-01 -4.56784248e-01 3.03029090e-01] | [15.469719886779785, 5.577300548553467] |
22d0963f-a997-4ab6-86e1-e87d44aff133 | using-connectome-features-to-constrain-echo | 2206.02094 | null | https://arxiv.org/abs/2206.02094v2 | https://arxiv.org/pdf/2206.02094v2.pdf | Using Connectome Features to Constrain Echo State Networks | We report an improvement to the conventional Echo State Network (ESN) across three benchmark chaotic time-series prediction tasks using fruit fly connectome data alone. We also investigate the impact of key connectome-derived structural features on prediction performance -- uniquely bridging neurobiological structure and machine learning function; and find that both increasing the global average clustering coefficient and modifying the position of weights -- by permuting their synapse-synapse partners -- can lead to increased model variance and (in some cases) degraded performance. In all we consider four topological point modifications to a connectome-derived ESN reservoir (null model): namely, we alter the network sparsity, re-draw nonzero weights from a uniform distribution, permute nonzero weight positions, and increase the network global average clustering coefficient. We compare the four resulting ESN model classes -- and the null model -- with a conventional ESN by conducting time-series prediction experiments on size-variants of the Mackey-Glass 17 (MG-17), Lorenz, and Rossler chaotic time series; denoting each model's performance and variance across train-validate trials. | ['Mark Daley', 'Jacob Morra'] | 2022-06-05 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 1.65009111e-01 8.75730533e-03 1.14400804e-01 6.05557784e-02
4.00593907e-01 -6.44949555e-01 6.50737941e-01 -6.84386236e-04
-4.55724955e-01 7.94461966e-01 3.28922458e-03 -4.31939662e-01
-6.19822025e-01 -5.76715708e-01 -4.60109919e-01 -7.81475246e-01
-1.18676150e+00 6.62578583e-01 4.31446046e-01 -4.10349607e-01
5.71849108e-01 4.02903199e-01 -8.03345501e-01 -1.33007437e-01
7.55688548e-01 8.85136724e-01 -6.17908128e-02 6.33011997e-01
3.25766027e-01 4.39201981e-01 -4.92680460e-01 -2.28383690e-01
6.57261252e-01 -4.19694662e-01 -5.11992395e-01 -6.31201327e-01
1.37532353e-01 3.36690843e-01 -5.22459865e-01 8.83268654e-01
4.04029936e-01 -6.87388480e-02 7.45512307e-01 -1.23020411e+00
-3.14902812e-01 7.81928003e-01 -3.21163327e-01 7.40559578e-01
-2.29290947e-01 4.28716868e-01 6.83097184e-01 -3.44682373e-02
8.29030752e-01 9.35343862e-01 1.42176318e+00 5.61553001e-01
-2.09863043e+00 -8.29172313e-01 -1.08826652e-01 -9.25047919e-02
-9.83660460e-01 -2.72083104e-01 7.72021294e-01 -6.00409091e-01
1.16542792e+00 -1.02579676e-01 1.30962992e+00 1.17032182e+00
9.68735754e-01 -2.99045205e-01 1.25315475e+00 2.33654469e-01
4.25918192e-01 -4.39129502e-01 1.66353077e-01 6.07310712e-01
4.73733664e-01 3.43175262e-01 -2.94538915e-01 -4.83937114e-01
9.98588085e-01 -1.86117053e-01 -4.35328126e-01 -4.54542607e-01
-1.65323663e+00 4.59018081e-01 5.71179390e-01 4.77633893e-01
-3.28088075e-01 4.98419821e-01 3.25477570e-01 8.63974929e-01
2.18382001e-01 9.83706534e-01 -4.85029966e-01 -1.14135630e-01
-9.94150341e-01 1.53131545e-01 1.22582793e+00 3.38956594e-01
3.48062843e-01 3.65299553e-01 2.54594296e-01 3.60104591e-01
-3.83848026e-02 5.62920272e-01 8.04470837e-01 -8.19653332e-01
3.21463972e-01 3.97675425e-01 -2.39796236e-01 -1.24193835e+00
-9.68988538e-01 -7.79035985e-01 -1.38554776e+00 -6.45286515e-02
4.60668027e-01 -4.90938902e-01 -6.34232819e-01 1.89744890e+00
-1.39210194e-01 5.36124468e-01 -8.58225301e-03 6.80487394e-01
3.20894361e-01 5.72314918e-01 7.70459920e-02 -1.22112125e-01
6.75456285e-01 -3.73600096e-01 -9.57077742e-02 -5.75816184e-02
4.95188206e-01 -2.18362197e-01 6.89990163e-01 1.93647608e-01
-1.04839790e+00 -2.73909956e-01 -1.23534095e+00 7.23983884e-01
-4.07211840e-01 -3.00921381e-01 3.05166721e-01 4.07097489e-01
-1.33563530e+00 1.51359260e+00 -9.68735933e-01 -4.32433128e-01
3.15597490e-03 7.56410420e-01 -3.62268478e-01 6.78448856e-01
-1.14756382e+00 9.90545690e-01 1.59654662e-01 -4.21209335e-02
-8.33299875e-01 -9.00373042e-01 -2.41171882e-01 3.49862456e-01
-4.91465479e-01 -7.35461116e-01 4.52547133e-01 -7.92720318e-01
-1.40244615e+00 2.78316557e-01 3.77589077e-01 -1.02742481e+00
3.78632069e-01 6.36266410e-01 -5.19035697e-01 1.04743026e-01
-2.25174233e-01 1.00068307e+00 7.81416178e-01 -9.70425427e-01
3.41077924e-01 -8.80376324e-02 -5.26563168e-01 -2.53412068e-01
-3.57784569e-01 -5.84713459e-01 6.46330655e-01 -7.62194097e-01
3.87545109e-01 -1.38511300e+00 -4.21463668e-01 -9.96650681e-02
-6.05839491e-01 3.14680547e-01 5.92366576e-01 -3.33341718e-01
1.03962767e+00 -1.88613307e+00 3.21551800e-01 7.64498174e-01
5.73787093e-01 2.16669321e-01 -6.38752639e-01 6.55158162e-01
-5.17071307e-01 3.38620305e-01 -3.66940975e-01 5.07930256e-02
-3.15326810e-01 8.95092711e-02 -3.15280616e-01 5.93044519e-01
1.88236341e-01 9.92494762e-01 -8.47718298e-01 -1.17582716e-01
-1.10981703e-01 4.63831872e-01 -8.16172063e-01 -4.73447770e-01
-8.71390570e-03 6.28284812e-01 -1.47738189e-01 -1.85246598e-02
3.99822384e-01 -6.27473950e-01 4.54334706e-01 -1.82297453e-03
7.39181116e-02 2.38346636e-01 -7.07147717e-01 9.91802454e-01
-6.47346452e-02 8.69811773e-01 -3.51563752e-01 -1.08325720e+00
1.17500484e+00 4.88293692e-02 8.15757990e-01 -7.66048610e-01
-1.14130042e-01 1.12363160e-01 1.03098226e+00 -5.30506782e-02
1.02606587e-01 -1.01533815e-01 5.66922687e-02 5.39241552e-01
1.79971650e-01 1.31625265e-01 1.61186025e-01 2.00362340e-01
1.68926072e+00 -4.04349953e-01 -2.07763612e-01 -9.82015669e-01
2.46695578e-01 -6.46632090e-02 5.14305413e-01 6.48274660e-01
-3.72250497e-01 2.82951087e-01 1.23656964e+00 -4.97707039e-01
-1.58547795e+00 -1.24910212e+00 -2.21088216e-01 4.06643063e-01
3.75453755e-02 -3.96440744e-01 -5.51130474e-01 -2.21357703e-01
2.43290827e-01 2.65463859e-01 -9.32265639e-01 -5.72752714e-01
-7.82765090e-01 -1.05420911e+00 1.01001084e+00 1.22743756e-01
6.00031972e-01 -1.10650456e+00 -6.82810366e-01 2.62610763e-01
1.40450656e-01 -5.90601385e-01 -3.10925752e-01 4.89746064e-01
-1.28250480e+00 -1.05457318e+00 -6.72822893e-01 -8.14471006e-01
7.06131995e-01 -3.46065462e-01 1.06370199e+00 2.50031710e-01
-1.47465333e-01 1.01677656e-01 4.04976159e-02 3.24563384e-01
-3.78216654e-01 3.46488953e-01 3.60184491e-01 -4.15560067e-01
-2.67191261e-01 -1.32949841e+00 -8.86606097e-01 6.25398338e-01
-6.44418001e-01 -1.61714077e-01 3.08329672e-01 9.27097797e-01
2.93298453e-01 -1.27200350e-01 7.95484543e-01 -4.57418740e-01
9.65007484e-01 -5.89258075e-01 -4.82066751e-01 1.10316545e-01
-8.39438736e-01 4.74797636e-01 8.64830196e-01 -8.97263527e-01
-1.01131350e-01 -2.45040163e-01 3.82100195e-01 -4.74144727e-01
4.86479014e-01 5.61568677e-01 7.93043673e-01 -5.33360183e-01
6.51112378e-01 5.31073630e-01 3.72057766e-01 -1.94606379e-01
-8.41357484e-02 -1.25784993e-01 3.31822485e-01 -4.75792289e-01
6.25429869e-01 2.06826910e-01 6.43665671e-01 -6.53459847e-01
1.36760086e-01 2.07992792e-01 -5.93982279e-01 -1.82174921e-01
5.39151371e-01 -4.58726794e-01 -9.82482135e-01 4.14608717e-01
-9.53938842e-01 -8.47701371e-01 -1.67052269e-01 3.76576722e-01
-4.15318161e-01 1.06342927e-01 -9.63611245e-01 -2.25827381e-01
-2.98783958e-01 -8.31293643e-01 2.98364431e-01 1.88970212e-02
-2.66803622e-01 -1.33803904e+00 6.98173702e-01 -6.31645322e-01
1.13149691e+00 4.23133880e-01 1.36031890e+00 -8.42266083e-01
5.87039888e-02 3.33341435e-02 -1.42518818e-01 -3.36593352e-02
-3.85585099e-01 5.00896834e-02 -1.90115854e-01 -4.72516149e-01
-7.40479678e-02 1.12199791e-01 9.43044126e-01 4.89181638e-01
6.12735152e-01 -4.66848046e-01 -5.52484512e-01 4.81156528e-01
1.31030226e+00 2.09042709e-02 4.99814391e-01 2.73932904e-01
-2.41400278e-03 5.47449172e-01 -5.79812288e-01 3.32625121e-01
1.08603366e-01 3.72670949e-01 2.97874004e-01 3.46856326e-01
5.94923832e-02 -3.09902400e-01 5.35434127e-01 1.17120695e+00
-1.41600952e-01 8.50466173e-03 -1.15645492e+00 3.05177987e-01
-1.61501861e+00 -1.21973908e+00 -9.09110457e-02 2.03063989e+00
6.57984316e-01 3.09982240e-01 3.32717061e-01 -8.81507918e-02
8.76167774e-01 1.39838472e-01 -8.90806496e-01 -9.80912223e-02
-2.77795136e-01 1.42355934e-01 6.99396372e-01 2.70279109e-01
-5.50871611e-01 4.88628715e-01 7.15108490e+00 2.80483931e-01
-1.53026760e+00 -7.05096796e-02 6.52550399e-01 -5.02234757e-01
-1.33193523e-01 1.14235751e-01 -3.75283450e-01 8.73482585e-01
1.48771513e+00 -2.54620790e-01 8.40237141e-01 3.67042810e-01
1.10074915e-01 1.86349332e-01 -1.05747032e+00 7.88999796e-01
-3.01785946e-01 -1.51225770e+00 -1.53774038e-01 4.65389431e-01
7.48895168e-01 6.52230859e-01 2.37898052e-01 1.96155086e-01
2.91993231e-01 -8.75080824e-01 4.02787775e-01 1.02753913e+00
6.20573759e-01 -3.49934995e-01 4.30067241e-01 1.67975768e-01
-1.17876947e+00 -2.99792320e-01 -4.64690268e-01 3.15839835e-02
7.82657489e-02 5.85778892e-01 -4.34535652e-01 -2.79661596e-01
5.88498294e-01 7.74278522e-01 -7.77989686e-01 1.30928159e+00
5.00898659e-01 8.43872726e-01 -7.27600574e-01 -3.36874157e-01
1.77538246e-01 -2.51375437e-01 9.32948589e-01 9.38901126e-01
2.73842156e-01 -4.74717095e-02 -2.21011728e-01 1.05796695e+00
2.51729283e-02 -5.15686154e-01 -5.90918899e-01 -6.48552328e-02
6.20749950e-01 1.10813749e+00 -1.18635905e+00 -1.21256381e-01
3.63440305e-01 4.95957822e-01 1.50642723e-01 3.69622767e-01
-7.01875985e-01 -6.35939181e-01 5.39496481e-01 4.42796320e-01
3.66631776e-01 -5.05171359e-01 -1.56550363e-01 -1.17428434e+00
-2.77003616e-01 -4.87784952e-01 -1.89897373e-01 -6.19823396e-01
-1.54661274e+00 8.29246819e-01 -1.20956354e-01 -1.20601332e+00
-2.90371895e-01 -4.12311167e-01 -1.13701797e+00 4.26285267e-01
-1.08156359e+00 -4.03320611e-01 1.03300288e-01 4.76773888e-01
-3.02377403e-01 -3.48370224e-01 5.79946458e-01 1.93869784e-01
-5.21049857e-01 4.13305491e-01 3.82821262e-01 5.62708080e-02
2.63509035e-01 -9.78808343e-01 7.15280056e-01 3.02377850e-01
-4.93774302e-02 1.00302863e+00 7.85087109e-01 -9.83701110e-01
-1.22568452e+00 -9.69068646e-01 5.81066132e-01 -2.43892699e-01
1.11936927e+00 -4.95657831e-01 -8.20954442e-01 5.34019291e-01
1.03863187e-01 1.39817610e-01 3.67071152e-01 -8.20670202e-02
-2.21720114e-01 -7.51361623e-02 -1.44538808e+00 7.53184080e-01
1.16957188e+00 -2.89354712e-01 -3.49436134e-01 9.75837111e-02
4.58485365e-01 3.75830680e-01 -1.15186596e+00 3.25482100e-01
7.82647431e-01 -9.41808164e-01 8.78227174e-01 -7.46738195e-01
4.46996868e-01 -1.37864351e-01 2.31879681e-01 -1.67152500e+00
-5.71369350e-01 -8.80567729e-01 -1.06386794e-02 5.74717164e-01
5.81949055e-01 -1.15278542e+00 8.87847662e-01 2.78999954e-01
7.63862878e-02 -8.70416343e-01 -1.37583077e+00 -9.43851113e-01
4.02411491e-01 1.91900000e-01 4.17109936e-01 1.04471874e+00
6.27820790e-01 3.82199943e-01 -1.76794633e-01 -3.01597625e-01
5.34673512e-01 -2.61140406e-01 2.73559511e-01 -1.56976533e+00
-2.93751210e-01 -9.40110028e-01 -7.52920866e-01 -7.34206796e-01
4.60282788e-02 -9.54693437e-01 -3.40290576e-01 -7.39146650e-01
-9.57218260e-02 -7.82535911e-01 -3.16721797e-01 3.48742962e-01
4.69142795e-01 2.53821075e-01 4.39667672e-01 6.22984767e-01
-4.56587076e-01 4.56024110e-01 1.11670232e+00 8.90166909e-02
-1.57255054e-01 -1.60549775e-01 -9.44363922e-02 3.40221435e-01
9.35211837e-01 -7.09421337e-01 -3.36491287e-01 -1.42603423e-02
7.39678085e-01 4.04590219e-01 6.06889546e-01 -1.77370167e+00
5.15569329e-01 3.14734280e-01 6.43480480e-01 -1.00821719e-01
2.65706629e-01 -6.52240396e-01 3.61988544e-01 1.20928502e+00
-6.67410791e-01 7.88938224e-01 2.78417617e-01 8.23822379e-01
3.10889363e-01 8.74857008e-02 8.45331728e-01 2.32870989e-02
1.93513677e-01 1.68553352e-01 -6.58652902e-01 1.81861855e-02
8.23507488e-01 -1.88786134e-01 -7.65562952e-01 -2.93984503e-01
-9.26959455e-01 2.49351129e-01 4.97542262e-01 6.73920959e-02
3.13093483e-01 -1.07056975e+00 -5.31466961e-01 4.93431687e-01
-4.78664011e-01 -8.80637467e-01 -1.87672917e-02 1.24393404e+00
-5.01674473e-01 5.04006863e-01 -8.05002034e-01 -6.41183376e-01
-6.96878612e-01 1.82295889e-01 8.71361852e-01 -4.58471626e-01
-5.80999911e-01 4.37935531e-01 -4.36880857e-01 -6.14018619e-01
-1.52446300e-01 -6.01839781e-01 -9.05456692e-02 -1.83163166e-01
-1.77506492e-01 3.95379812e-01 -2.29238614e-01 -3.74305993e-01
-6.63966313e-02 6.16385639e-01 2.48244077e-01 1.42046548e-02
1.61870754e+00 5.77030852e-02 -3.31325203e-01 3.75476629e-01
9.86549854e-01 -5.13490915e-01 -1.23602331e+00 2.18839832e-02
1.40871301e-01 1.76466703e-01 -2.66821235e-01 -8.11553419e-01
-1.14292383e+00 8.30917418e-01 8.36399853e-01 7.73608685e-01
5.79809070e-01 -2.74706364e-01 7.14049339e-01 8.60067964e-01
4.28810567e-01 -6.18981361e-01 1.75553441e-01 7.26673901e-01
6.76959157e-01 -5.82584918e-01 -3.50039601e-02 4.45830703e-01
-3.84757102e-01 1.22313499e+00 6.34570599e-01 -9.53539789e-01
1.22596335e+00 1.94780678e-01 -6.22128546e-01 -3.67980510e-01
-1.18787050e+00 5.27294338e-01 -1.14407390e-02 5.40431380e-01
1.03158504e-01 -1.46279648e-01 -7.13747442e-02 4.38502669e-01
-5.51519454e-01 -1.95988968e-01 5.95895588e-01 4.80683833e-01
-4.35391426e-01 -7.66101360e-01 9.22434926e-02 1.03917933e+00
-7.57403374e-02 -3.20150524e-01 -4.30129945e-01 7.84728467e-01
-2.36028165e-01 3.78888339e-01 5.05534828e-01 -6.78506017e-01
-1.02640994e-01 9.11679640e-02 3.24128658e-01 -4.21901755e-02
-1.00506639e+00 -3.32373083e-01 -1.39158681e-01 -3.34892035e-01
-1.16430506e-01 -6.38020635e-01 -1.08722901e+00 -9.82544601e-01
-3.43053192e-01 2.58144766e-01 6.89828575e-01 7.57561505e-01
8.93878579e-01 4.02592719e-01 4.46678936e-01 -1.13482082e+00
-6.75505280e-01 -1.16033292e+00 -5.66927254e-01 1.22549953e-02
4.48579222e-01 -4.52174276e-01 -7.45645583e-01 -2.53257900e-01] | [6.784858703613281, 3.595660448074341] |
9b7d4c76-2d3d-4c4e-b819-a7432191603b | acoustic-non-line-of-sight-imaging | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Lindell_Acoustic_Non-Line-Of-Sight_Imaging_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Lindell_Acoustic_Non-Line-Of-Sight_Imaging_CVPR_2019_paper.pdf | Acoustic Non-Line-Of-Sight Imaging | Non-line-of-sight (NLOS) imaging enables unprecedented capabilities in a wide range of applications, including robotic and machine vision, remote sensing, autonomous vehicle navigation, and medical imaging. Recent approaches to solving this challenging problem employ optical time-of-flight imaging systems with highly sensitive time-resolved photodetectors and ultra-fast pulsed lasers. However, despite recent successes in NLOS imaging using these systems, widespread implementation and adoption of the technology remains a challenge because of the requirement for specialized, expensive hardware. We introduce acoustic NLOS imaging, which is orders of magnitude less expensive than most optical systems and captures hidden 3D geometry at longer ranges with shorter acquisition times compared to state-of-the-art optical methods. Inspired by hardware setups used in radar and algorithmic approaches to model and invert wave-based image formation models developed in the seismic imaging community, we demonstrate a new approach to seeing around corners.
| [' Vladlen Koltun', ' Gordon Wetzstein', 'David B. Lindell'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['seismic-imaging'] | ['miscellaneous'] | [ 5.69972754e-01 -2.39468992e-01 4.26148355e-01 -1.76359683e-01
-5.89938462e-01 -5.79808891e-01 2.93096632e-01 -4.23449099e-01
-8.44341218e-01 5.41395187e-01 -1.59262389e-01 -2.21154600e-01
-2.69045621e-01 -6.36574447e-01 -2.89269239e-01 -8.68988514e-01
7.47283250e-02 7.20017910e-01 4.38632280e-01 1.55231088e-01
3.78388971e-01 1.02639461e+00 -1.37769806e+00 -3.03336024e-01
2.24011049e-01 8.46837521e-01 9.66421291e-02 5.70542753e-01
-5.42916730e-03 2.14450046e-01 -8.89442861e-02 2.31894180e-01
1.77084789e-01 -4.67858076e-01 -1.46781448e-02 5.87909706e-02
6.49566591e-01 -5.34299433e-01 -6.21266663e-01 8.56613636e-01
5.03830135e-01 7.67277852e-02 5.00679731e-01 -6.17007554e-01
-3.62456203e-01 -1.86956540e-01 -7.41724908e-01 2.58725613e-01
6.84322476e-01 2.99191535e-01 6.30729914e-01 -8.78460348e-01
4.80420172e-01 7.60437489e-01 7.82516003e-01 3.87779146e-01
-1.03613341e+00 -4.81575251e-01 -8.58692586e-01 1.39299437e-01
-1.27215123e+00 -6.73537433e-01 5.68322301e-01 -4.37119484e-01
1.03872132e+00 -6.47653565e-02 5.35680830e-01 7.66240954e-01
4.26691949e-01 5.09576499e-02 1.21387100e+00 -7.43260086e-01
4.76679765e-02 1.51908025e-03 -1.02158926e-01 9.76848722e-01
6.33434415e-01 4.92087185e-01 -1.08336174e+00 -1.97023124e-01
9.29789722e-01 2.88262367e-01 -5.68956137e-01 -2.63273567e-01
-1.26949024e+00 5.07487416e-01 2.65889224e-02 3.48128378e-01
-4.85102087e-01 3.50501895e-01 -4.15859103e-01 4.65256833e-02
3.75037491e-01 7.74717450e-01 1.50550812e-01 6.89027191e-04
-1.11256182e+00 -5.66003583e-02 5.17063320e-01 3.37225974e-01
8.91974986e-01 2.30215177e-01 5.82044840e-01 3.63197476e-01
5.69100201e-01 1.25552595e+00 1.52815253e-01 -1.08469915e+00
-1.34741798e-01 6.87353387e-02 1.85133517e-01 -5.62619030e-01
-6.45028710e-01 -5.67063689e-01 -2.25444987e-01 6.17231250e-01
3.58200997e-01 -1.23119719e-01 -9.74889159e-01 1.01642835e+00
3.39761704e-01 4.53201771e-01 4.95749712e-01 9.21801269e-01
5.73111951e-01 6.36265635e-01 -7.46971786e-01 -5.96927941e-01
1.22299373e+00 -4.83514875e-01 -5.32201469e-01 -7.07254648e-01
3.04873168e-01 -8.89741898e-01 5.32107651e-01 6.58818722e-01
-8.91813099e-01 5.95693700e-02 -1.02016366e+00 2.14426458e-01
3.94599319e-01 -2.06997648e-01 4.06053990e-01 7.31445789e-01
-7.63161361e-01 2.23862082e-01 -8.76995862e-01 -5.63795686e-01
1.50560677e-01 -4.48502749e-02 -3.14435691e-01 -5.39611638e-01
-4.51671451e-01 5.74442625e-01 -2.35878184e-01 7.22659826e-02
-6.25052333e-01 -7.08535671e-01 -5.95369697e-01 -2.72391915e-01
5.99420667e-01 -6.87823355e-01 7.91311860e-01 2.25721970e-02
-1.70969450e+00 7.40588844e-01 -2.73149759e-01 -4.21791673e-01
1.08478680e-01 -4.46464986e-01 -2.77189285e-01 9.90117669e-01
1.39276370e-01 1.15547270e-01 8.70726764e-01 -9.63634312e-01
-5.24632744e-02 -6.04832888e-01 -4.84620780e-01 3.58867832e-02
-2.25279406e-02 4.85618934e-02 1.32465199e-01 3.08236986e-01
8.71913373e-01 -9.06500340e-01 -3.49142700e-01 2.71707088e-01
-3.89779836e-01 4.84064937e-01 1.08685362e+00 1.17688872e-01
2.66446918e-01 -2.04766512e+00 -3.88199747e-01 -9.57588628e-02
1.95844710e-01 3.31335843e-01 -1.22412130e-01 9.45451498e-01
5.09028554e-01 -3.83389920e-01 -4.55040693e-01 -2.33737990e-01
-4.77917761e-01 1.01454534e-01 -3.55548590e-01 7.80443132e-01
-3.94105852e-01 5.03365934e-01 -8.23568404e-01 -3.60498756e-01
3.99420559e-01 5.11471331e-01 -2.22772524e-01 -1.06411621e-01
-7.73063526e-02 8.15522373e-01 -6.46226346e-01 9.30266380e-01
7.65316904e-01 4.63353805e-02 -1.15250617e-01 -2.07766175e-01
-8.16821039e-01 -8.07288736e-02 -9.30940568e-01 1.51831377e+00
-2.98688173e-01 8.57735574e-01 2.76582509e-01 -7.38124967e-01
8.95961225e-01 4.67170745e-01 5.11179090e-01 -9.03651655e-01
-2.36097232e-01 6.07484519e-01 -4.79118794e-01 -1.15691996e+00
3.89548898e-01 -9.34719503e-01 2.33288974e-01 6.39606416e-01
-2.64998198e-01 -6.26990736e-01 -2.26205125e-01 6.69406503e-02
1.44444489e+00 3.12535502e-02 1.56972140e-01 5.36719784e-02
7.90027380e-02 3.02597761e-01 2.51645565e-01 9.95691001e-01
1.05544962e-01 6.36356533e-01 -3.47981811e-01 -5.15615702e-01
-7.83838034e-01 -1.18072319e+00 -3.14169139e-01 1.70408025e-01
2.17064202e-01 1.08365312e-01 -3.10449839e-01 1.86603099e-01
-8.87903348e-02 4.13476497e-01 6.63062930e-02 3.62964988e-01
-4.51924622e-01 -8.09193671e-01 8.13362122e-01 5.00592887e-02
5.34059525e-01 -8.76775146e-01 -1.37978578e+00 2.92430431e-01
3.69837768e-02 -1.73325002e+00 3.62984240e-01 -3.19987535e-01
-8.07676136e-01 -1.12615418e+00 -6.53339684e-01 -1.88748971e-01
4.29991364e-01 8.46756160e-01 7.28681684e-01 -4.42300513e-02
-8.89120102e-01 1.03846157e+00 -2.10871845e-01 -4.99853045e-01
-7.05186129e-02 -6.41507447e-01 4.92343456e-01 6.84685037e-02
3.60580772e-01 -8.60103548e-01 -7.91024446e-01 9.77022871e-02
-7.68825948e-01 -3.19761306e-01 8.76875460e-01 5.01348138e-01
5.23544788e-01 -1.66765243e-01 2.72857815e-01 -7.28114724e-01
2.24538401e-01 -7.62722939e-02 -8.94949675e-01 -1.41211540e-01
-4.96994406e-01 -7.89624825e-02 1.75380051e-01 5.08083776e-03
-9.62522328e-01 -3.61451134e-02 -1.58299223e-01 -2.38348800e-03
-5.58493853e-01 7.42414117e-01 5.19606709e-01 -7.20710516e-01
9.11114335e-01 3.49909812e-01 3.27479661e-01 -1.81790516e-01
-2.31234599e-02 6.51204050e-01 6.08709216e-01 7.94546958e-03
9.89839971e-01 1.33004677e+00 6.32046402e-01 -2.03035188e+00
-1.05197823e+00 -8.41114581e-01 -4.52760071e-01 -2.70588070e-01
9.66710150e-01 -7.51252890e-01 -7.20739186e-01 4.86184239e-01
-1.02593017e+00 -3.02217174e-02 -2.18733132e-01 1.20082152e+00
-3.38384360e-01 6.01712584e-01 -3.16690028e-01 -1.18059254e+00
-9.94303226e-02 -7.38867521e-01 1.09305418e+00 5.32386661e-01
1.65174261e-01 -8.49411666e-01 4.10396546e-01 6.54488385e-01
4.35509175e-01 4.02749509e-01 4.66082364e-01 3.40850264e-01
-1.26110303e+00 -3.51953238e-01 -8.83529186e-02 -5.46450615e-02
-2.08482623e-01 -1.17064908e-01 -1.03915572e+00 -5.74878305e-02
3.49876881e-01 -3.84880722e-01 1.06025171e+00 8.57189000e-01
1.80901691e-01 4.01077598e-01 -6.32436633e-01 9.46238816e-01
1.71715307e+00 1.85391635e-01 6.41887367e-01 2.70510912e-01
5.70244610e-01 6.10027909e-01 4.49707925e-01 3.31025809e-01
2.67777201e-02 6.06047630e-01 4.07077074e-01 -9.68640950e-03
5.77947358e-03 3.45523715e-01 2.26855814e-01 3.76056314e-01
-2.10345224e-01 -1.37762144e-01 -8.44348371e-01 3.49159390e-01
-1.34387743e+00 -1.20469534e+00 -5.30603647e-01 2.34205508e+00
1.29545376e-01 -8.87289122e-02 -5.70606828e-01 -5.93634695e-02
-7.18957633e-02 1.60211042e-01 -4.33806241e-01 1.67417526e-03
-1.51763901e-01 4.52798069e-01 5.96347630e-01 7.74074912e-01
-4.68183666e-01 6.19885266e-01 6.55084038e+00 7.82479644e-02
-1.39078689e+00 1.32905692e-01 -3.98313105e-01 -5.84123656e-02
-5.61984599e-01 1.68000847e-01 -7.89201260e-01 -1.38102338e-01
6.75849259e-01 2.70561218e-01 2.25187868e-01 -4.83893380e-02
4.20392841e-01 -9.55360651e-01 -6.70782447e-01 1.09135866e+00
2.22266227e-01 -1.37255347e+00 -2.51109242e-01 2.11535841e-01
4.29050565e-01 4.60972428e-01 1.03894345e-01 -5.90261936e-01
-4.85936850e-01 -7.10681379e-01 3.62225533e-01 6.99961126e-01
9.39846098e-01 -3.76906902e-01 2.67483801e-01 5.88497221e-01
-7.25877285e-01 -1.95910037e-02 -5.75974941e-01 -2.71155119e-01
7.09256649e-01 1.27314889e+00 -8.10236216e-01 2.79859483e-01
8.27255249e-01 3.03888053e-01 1.38329551e-01 1.05211031e+00
-3.31255913e-01 7.09435701e-01 -6.86365068e-01 3.15834908e-03
3.51502657e-01 -5.35788059e-01 9.71571922e-01 7.90557325e-01
7.85187006e-01 4.29081678e-01 1.14657551e-01 8.37708831e-01
3.30196589e-01 -5.96854746e-01 -1.02455616e+00 -1.51554123e-01
4.10881609e-01 1.28761065e+00 -7.75725663e-01 1.99005172e-01
-4.80508834e-01 5.11751831e-01 -4.40711021e-01 6.50532365e-01
-1.98515281e-01 -5.60369611e-01 4.50676352e-01 5.59622824e-01
1.49017468e-01 -1.11643207e+00 -1.33348666e-02 -9.36508179e-01
-1.03598006e-01 -1.68454312e-02 -1.80295438e-01 -1.09568751e+00
-7.78171599e-01 4.43056405e-01 6.83580115e-02 -9.40402865e-01
-1.52913645e-01 -6.66510761e-01 -5.32284200e-01 7.27725863e-01
-2.14190483e+00 -1.04318595e+00 -5.85402608e-01 3.75691950e-01
6.46355301e-02 9.50575992e-02 9.08394575e-01 -6.80771172e-02
1.05878627e-02 -4.47282642e-01 4.38061774e-01 -1.54999956e-01
8.58657181e-01 -5.94121933e-01 1.96081504e-01 1.07064629e+00
4.47660536e-01 3.61838579e-01 9.75165427e-01 -5.36584020e-01
-2.05414629e+00 -5.05454063e-01 5.90474308e-01 2.54705679e-02
6.51524007e-01 5.11014322e-03 -5.44535041e-01 5.83807766e-01
1.61761507e-01 4.49813575e-01 8.90040696e-01 -2.75569320e-01
-2.40759119e-01 -4.73584272e-02 -1.15565979e+00 6.42842129e-02
8.57522726e-01 -7.08864212e-01 -5.27205884e-01 5.31896174e-01
2.28443190e-01 -2.81381428e-01 -1.73464909e-01 3.95645410e-01
9.27451193e-01 -1.39810705e+00 1.08963048e+00 2.98094094e-01
1.64312571e-02 -3.15459937e-01 -6.75142407e-02 -1.08966625e+00
-1.33179650e-01 -7.57186592e-01 4.37594056e-01 5.02648115e-01
1.81019306e-01 -1.23563135e+00 1.10125315e+00 -2.23052859e-01
-5.12657344e-01 -2.62109846e-01 -1.21886349e+00 -7.78530061e-01
-5.45977950e-01 -2.83700168e-01 -1.96825951e-01 5.91594458e-01
-3.30888808e-01 3.85431379e-01 -1.33636937e-01 9.30554032e-01
1.52415454e+00 3.95762950e-01 4.38689232e-01 -1.53974450e+00
-3.65729630e-01 3.53365153e-01 -4.88838583e-01 -1.15637076e+00
-1.26667783e-01 -4.06041890e-01 2.43528545e-01 -1.66326439e+00
-8.41675028e-02 -4.65486228e-01 4.75128740e-01 1.32635320e-02
6.33253098e-01 9.90620315e-01 -1.74167365e-01 5.32329261e-01
-2.97214955e-01 2.07023948e-01 8.96757245e-01 1.85149878e-01
-4.90423813e-02 -6.16245754e-02 -5.87325394e-02 9.57653821e-01
3.74514282e-01 -7.17161298e-01 -2.84593552e-01 -6.59803271e-01
6.83063865e-01 3.22338253e-01 7.50260532e-01 -1.48666024e+00
6.39665842e-01 -2.43666172e-01 5.57747297e-02 -6.37951136e-01
8.16753030e-01 -5.76539755e-01 2.22772017e-01 6.58769131e-01
3.22453052e-01 -4.94211853e-01 -2.15984985e-01 7.74368823e-01
-2.40303636e-01 -8.61680627e-01 9.16597545e-01 -5.94422519e-01
-6.86992288e-01 1.02744892e-01 -7.08364129e-01 -5.02047949e-02
8.25309932e-01 -4.77591395e-01 -8.22566807e-01 -4.47115570e-01
-4.17073548e-01 -3.58961821e-01 5.58576941e-01 -6.28919065e-01
1.13290024e+00 -6.13838494e-01 -5.13473213e-01 2.39741415e-01
8.87961537e-02 -1.04174525e-01 4.03887540e-01 1.18061781e+00
-1.29517353e+00 4.73745883e-01 1.82040110e-02 -9.17113543e-01
-1.20144010e+00 -1.01821803e-01 4.12106335e-01 2.55793519e-02
-8.59717250e-01 9.48032916e-01 -9.26954672e-02 -2.28345573e-01
-4.01398182e-01 1.66876808e-01 1.96498647e-01 -3.14421952e-01
6.46994770e-01 4.50920939e-01 7.88957402e-02 -3.50547552e-01
-4.01122332e-01 1.26449919e+00 2.76105821e-01 -4.98657554e-01
1.49169052e+00 -2.43136615e-01 -1.24990176e-02 6.09881341e-01
6.40973389e-01 3.50602448e-01 -1.22438848e+00 -3.70284885e-01
-2.85017043e-01 -4.19115394e-01 3.62537295e-01 -3.24387312e-01
-4.70596075e-01 1.28228402e+00 3.45589966e-01 2.79993743e-01
7.51784205e-01 2.68568665e-01 9.22754288e-01 8.75118792e-01
6.80272818e-01 -7.34605670e-01 3.48691911e-01 5.02006471e-01
3.33808303e-01 -9.99807715e-01 5.21926641e-01 -4.81749564e-01
1.83885489e-02 1.42038810e+00 -1.83616757e-01 -5.76964812e-03
7.66068876e-01 7.20688999e-01 3.64635259e-01 -6.75968945e-01
-3.05940747e-01 -2.07969785e-01 -2.51039386e-01 7.88658202e-01
2.20810398e-01 -1.30696714e-01 2.35767942e-02 -4.45128322e-01
2.14843690e-01 4.64976244e-02 1.26242912e+00 1.21204305e+00
-1.03105080e+00 -9.12470698e-01 -6.64545357e-01 1.98245838e-01
-4.31129843e-01 -1.48275882e-01 2.24616736e-01 4.58336771e-01
-3.68680567e-01 9.55363393e-01 6.22799061e-02 7.38399178e-02
2.78064162e-01 -3.01346648e-02 8.61385882e-01 -9.22677934e-01
1.78110301e-01 3.49655479e-01 1.63763806e-01 -4.66797262e-01
-9.98597085e-01 -9.93476331e-01 -1.51403332e+00 3.88881326e-01
-4.38807189e-01 6.16808981e-02 1.03254414e+00 1.21938252e+00
3.08936954e-01 -1.34243011e-01 3.24276954e-01 -1.03289330e+00
-4.53827858e-01 -6.48431897e-01 -1.04185128e+00 -3.43336314e-01
5.80635846e-01 -8.08732927e-01 -6.75286531e-01 -2.19106093e-01] | [9.846494674682617, -2.7764647006988525] |
9d39155f-4474-4270-8c67-e6478292bd44 | less-is-more-rethinking-state-of-the-art | 2209.00243 | null | https://arxiv.org/abs/2209.00243v1 | https://arxiv.org/pdf/2209.00243v1.pdf | Less is More: Rethinking State-of-the-art Continual Relation Extraction Models with a Frustratingly Easy but Effective Approach | Continual relation extraction (CRE) requires the model to continually learn new relations from class-incremental data streams. In this paper, we propose a Frustratingly easy but Effective Approach (FEA) method with two learning stages for CRE: 1) Fast Adaption (FA) warms up the model with only new data. 2) Balanced Tuning (BT) finetunes the model on the balanced memory data. Despite its simplicity, FEA achieves comparable (on TACRED or superior (on FewRel) performance compared with the state-of-the-art baselines. With careful examinations, we find that the data imbalance between new and old relations leads to a skewed decision boundary in the head classifiers over the pretrained encoders, thus hurting the overall performance. In FEA, the FA stage unleashes the potential of memory data for the subsequent finetuning, while the BT stage helps establish a more balanced decision boundary. With a unified view, we find that two strong CRE baselines can be subsumed into the proposed training pipeline. The success of FEA also provides actionable insights and suggestions for future model designing in CRE. | ['Zhifang Sui', 'Yunbo Cao', 'Binghuai Lin', 'Rundong Gao', 'Tianyu Liu', 'YiFan Song', 'Peiyi Wang'] | 2022-09-01 | null | null | null | null | ['continual-relation-extraction'] | ['natural-language-processing'] | [ 1.52742773e-01 5.57956517e-01 -3.66574883e-01 -6.40383184e-01
-5.18110931e-01 -3.63173157e-01 7.13241339e-01 2.07217231e-01
-4.17384267e-01 6.95070863e-01 2.25373805e-01 -5.52441120e-01
-2.86299456e-02 -9.55095351e-01 -7.08218455e-01 -3.36489141e-01
-9.88758728e-02 5.75014055e-01 4.80807483e-01 -3.40463340e-01
-1.77214786e-01 2.68796891e-01 -1.38009202e+00 5.58617234e-01
5.89790881e-01 1.12723041e+00 -9.58361253e-02 5.82930624e-01
3.56127769e-02 1.19705951e+00 -5.64844310e-01 -9.54503834e-01
1.86472356e-01 -2.76964992e-01 -1.26258767e+00 -3.25334847e-01
3.47317010e-01 -3.61794233e-01 -4.95151758e-01 4.24851269e-01
4.26609427e-01 3.07788216e-02 3.25738281e-01 -1.17006278e+00
-7.18068361e-01 1.22158599e+00 -4.61314023e-01 3.70254636e-01
1.99866295e-02 -5.94416521e-02 1.24797416e+00 -1.13892972e+00
7.89288998e-01 1.11495376e+00 8.76887679e-01 5.56137145e-01
-1.21361208e+00 -8.25528443e-01 4.91032958e-01 2.19427884e-01
-1.24680245e+00 -8.94177854e-01 5.20779788e-01 -3.15183520e-01
1.50952423e+00 2.54323035e-01 6.41260028e-01 1.04085612e+00
-4.72476594e-02 9.73917842e-01 6.59342229e-01 -4.69640672e-01
2.16327198e-02 7.22700432e-02 5.68467855e-01 4.87085193e-01
2.41893187e-01 2.26266861e-01 -9.01911139e-01 1.26860484e-01
4.21429664e-01 -3.03179920e-01 -1.21421941e-01 -1.54421195e-01
-8.60032678e-01 5.05096853e-01 5.35095394e-01 4.59641606e-01
-1.87507555e-01 -3.66312414e-02 5.13045013e-01 4.59905505e-01
3.53194207e-01 7.28907585e-01 -8.48769128e-01 -2.62383044e-01
-8.94068182e-01 2.55866081e-01 6.93816125e-01 1.08730686e+00
7.44031489e-01 -2.31249630e-01 -3.81035358e-01 8.82856548e-01
6.25429377e-02 -1.44189775e-01 4.33678150e-01 -6.89218163e-01
8.19882810e-01 7.32919991e-01 -2.73186952e-01 -7.99793184e-01
-4.64803785e-01 -7.70395815e-01 -7.18515337e-01 -1.71408296e-01
4.96629238e-01 -1.90504625e-01 -9.42798436e-01 1.86653185e+00
1.57956824e-01 -1.49001062e-01 1.73868313e-01 4.12847191e-01
8.66854846e-01 4.14818674e-01 3.16900969e-01 -3.41696590e-01
1.14252579e+00 -1.18159282e+00 -7.28673398e-01 -6.84650838e-01
6.33644700e-01 -6.52341247e-01 8.58350217e-01 3.27784747e-01
-1.19464874e+00 -7.93793619e-01 -1.32226515e+00 -3.14865768e-01
-3.98750216e-01 -8.59683752e-02 1.10054886e+00 4.59260434e-01
-8.96978199e-01 7.13471532e-01 -9.56364214e-01 -3.78549665e-01
6.25227869e-01 5.14719129e-01 -3.22145104e-01 -3.36753465e-02
-1.46418238e+00 1.12756872e+00 5.60119212e-01 3.43384922e-01
-5.07108867e-01 -6.95835054e-01 -7.69972146e-01 6.24967963e-02
6.50843799e-01 -6.79350615e-01 1.48686552e+00 -6.26717627e-01
-1.32294321e+00 7.48815119e-01 -4.00253236e-01 -6.15032911e-01
4.42075372e-01 -5.92719913e-01 -2.72576928e-01 -3.22427660e-01
-1.91062629e-01 7.34124243e-01 6.64275706e-01 -1.28877211e+00
-8.56439114e-01 -2.56901145e-01 1.19174726e-01 7.13653862e-02
-4.25227225e-01 -1.98737122e-02 -4.13404047e-01 -6.85724735e-01
1.01556666e-01 -5.52700102e-01 1.19386353e-01 -5.47467649e-01
-2.98074484e-01 -4.31525826e-01 6.33658946e-01 -4.10383463e-01
2.03837276e+00 -1.97050714e+00 1.42435104e-01 -7.96205550e-03
4.49386001e-01 4.22292650e-01 -1.31169081e-01 2.53508210e-01
-3.93046260e-01 1.07611656e-01 1.17379259e-02 -3.34875315e-01
-2.01654553e-01 2.67226100e-01 -2.53511935e-01 -2.09914088e-01
5.46087205e-01 1.26565886e+00 -1.01929164e+00 -3.54677796e-01
-4.17291641e-01 9.13887694e-02 -4.00984287e-01 4.58700269e-01
-2.24338789e-02 -2.30509583e-02 -8.23638886e-02 6.12123787e-01
5.60289800e-01 -2.90309787e-01 3.71521771e-01 -2.68615991e-01
6.83295503e-02 8.91109407e-01 -1.00202107e+00 1.42987859e+00
-2.12590739e-01 6.44613206e-01 -2.86104530e-01 -8.70361984e-01
1.02557325e+00 2.01796323e-01 2.25259781e-01 -7.97433257e-01
-1.44386366e-02 1.79462597e-01 4.41458255e-01 -4.34075773e-01
6.44439518e-01 -1.73265100e-01 5.93585223e-02 3.89956921e-01
3.42921168e-01 5.15944436e-02 3.22007746e-01 2.24314108e-01
1.28618634e+00 2.12952808e-01 3.35193694e-01 1.20215550e-01
2.35969648e-01 -1.69567272e-01 8.03719640e-01 7.20923841e-01
-1.82266042e-01 3.30349326e-01 4.68414575e-01 -5.76171994e-01
-8.86498511e-01 -8.27780128e-01 -1.26221642e-01 1.38660347e+00
-1.81830645e-01 -8.58105719e-01 -4.78215158e-01 -1.06420159e+00
6.53672442e-02 5.93022346e-01 -7.81030178e-01 -4.27607864e-01
-1.01985538e+00 -1.09466445e+00 5.82154632e-01 1.00340390e+00
6.81228042e-01 -1.07831860e+00 -5.51853895e-01 4.08282399e-01
-1.89902544e-01 -9.57246482e-01 -2.17180938e-01 8.94574046e-01
-8.96120369e-01 -8.70865643e-01 -3.07592750e-01 -7.58476377e-01
2.94762105e-01 1.36878356e-01 1.44232011e+00 9.18869749e-02
6.67596906e-02 -1.46173701e-01 -5.20865619e-01 -4.80743170e-01
-4.17076379e-01 6.86258852e-01 -1.59520194e-01 -3.17070931e-01
7.11715817e-01 -5.83407104e-01 -2.23966852e-01 2.50990242e-01
-4.65496004e-01 3.01605821e-01 6.95071518e-01 1.00944769e+00
1.87048599e-01 1.74000040e-01 8.57109964e-01 -1.35344326e+00
6.75125718e-01 -5.40024877e-01 -1.67141676e-01 4.92855370e-01
-1.15428555e+00 2.09344536e-01 2.67499626e-01 -5.08324265e-01
-1.24505019e+00 -1.36003420e-01 -1.54759765e-01 3.68854217e-02
1.71003863e-01 5.42515814e-01 -1.86168671e-01 3.74814600e-01
9.23697591e-01 -2.95709521e-01 -4.24798936e-01 -4.45672959e-01
5.11255443e-01 6.21287107e-01 7.09843755e-01 -6.70952499e-01
6.88041270e-01 -4.80132513e-02 -4.39421445e-01 -1.63653910e-01
-1.18978214e+00 -1.69448555e-01 -9.25127685e-01 9.44713410e-03
5.44345558e-01 -8.37550938e-01 -3.98024023e-01 5.24218678e-01
-1.20308495e+00 -8.44876647e-01 -4.38297153e-01 1.13545403e-01
-1.35013863e-01 -2.97517180e-01 -8.96910906e-01 -7.40670681e-01
-3.51710677e-01 -7.99042523e-01 7.08756685e-01 3.20503950e-01
-8.20863128e-01 -8.20932567e-01 -5.80510199e-02 4.01552111e-01
4.52597022e-01 3.04012373e-02 1.06673062e+00 -6.95473909e-01
-4.36402470e-01 -5.45724388e-03 -3.31858575e-01 2.84893751e-01
1.48963854e-01 2.15301439e-02 -1.41693306e+00 -1.31636918e-01
-1.95223153e-01 -5.86397827e-01 1.14741230e+00 -1.24603748e-01
8.54000092e-01 -2.11494863e-01 -4.66172397e-01 5.65700710e-01
8.28082085e-01 5.29539406e-01 4.68874454e-01 4.74207461e-01
7.87847638e-01 5.60965955e-01 8.34376395e-01 1.32151395e-01
7.98810780e-01 4.39469039e-01 1.43868579e-02 -1.80267274e-01
-4.99168605e-01 -4.67486233e-01 3.28336507e-01 8.66438925e-01
-1.82049498e-01 -1.56543553e-01 -9.79345620e-01 5.40931582e-01
-1.99756765e+00 -7.58590698e-01 1.11777380e-01 2.01214981e+00
1.50139499e+00 7.81394720e-01 1.14333434e-02 5.15234947e-01
3.69249433e-01 2.14485794e-01 -6.20085835e-01 -5.77465355e-01
-1.16483778e-01 5.55898607e-01 1.88807309e-01 5.61524332e-01
-1.14602721e+00 1.26500261e+00 7.05654812e+00 6.88869596e-01
-1.07317340e+00 1.03746898e-01 8.18362713e-01 -9.42099169e-02
-1.83207124e-01 1.31695569e-01 -1.19787753e+00 3.08526363e-02
1.04225397e+00 -1.86007433e-02 3.74899596e-01 7.29052246e-01
-4.42114800e-01 -2.51673639e-01 -1.71177232e+00 7.18898833e-01
-9.81409103e-02 -1.34400630e+00 -1.21018007e-01 -2.46625170e-01
4.93709981e-01 -6.21295907e-02 -1.06354691e-01 7.65232265e-01
4.52535152e-01 -9.67253864e-01 7.97683895e-01 4.12312835e-01
8.12822998e-01 -5.93574286e-01 8.32374394e-01 1.91921592e-01
-1.16445374e+00 -3.11146379e-01 -1.01266734e-01 -4.21644837e-01
2.69016605e-02 6.21586442e-01 -8.87483776e-01 5.99447370e-01
6.03947997e-01 6.79622173e-01 -9.52910006e-01 5.16975999e-01
-4.82023269e-01 7.56368041e-01 -1.44178003e-01 3.41739148e-01
-1.41329035e-01 4.58400339e-01 1.21267684e-01 1.41175222e+00
-1.48752242e-01 1.13696955e-01 -1.35614038e-01 5.82428873e-01
-2.92632401e-01 -1.12716332e-01 -2.33754158e-01 -7.71846846e-02
8.47418547e-01 1.09273124e+00 -5.25229573e-01 -4.02921379e-01
-3.98282081e-01 6.55560374e-01 9.30825531e-01 3.39123666e-01
-4.87034172e-01 -3.42612088e-01 3.36383730e-01 1.86385304e-01
2.40354776e-01 3.24613415e-02 -6.50887430e-01 -1.08578396e+00
-2.29962673e-02 -8.92063618e-01 7.19344914e-01 -4.56546754e-01
-1.15270138e+00 7.91759431e-01 -4.50395755e-02 -5.30564368e-01
-4.69304711e-01 -3.85294437e-01 -3.40520352e-01 6.72513425e-01
-1.44992995e+00 -1.29521155e+00 -2.76002347e-01 1.64600074e-01
5.57714880e-01 -2.83610094e-02 8.47935021e-01 5.65120220e-01
-9.81843352e-01 1.11216283e+00 -5.96490204e-01 2.71360934e-01
1.03251302e+00 -1.24938035e+00 7.89651215e-01 9.05728757e-01
2.19991431e-01 8.19454789e-01 4.16005015e-01 -7.00013220e-01
-9.28483248e-01 -9.36856389e-01 1.44308472e+00 -7.69760251e-01
6.55793786e-01 -5.78382969e-01 -1.16346037e+00 9.66610909e-01
1.02877535e-01 -8.69947076e-02 6.95454717e-01 8.37185323e-01
-6.68688416e-01 -4.51623440e-01 -7.01034963e-01 2.62261689e-01
1.19288719e+00 -3.78853083e-01 -7.77874470e-01 -2.86637098e-01
8.59350383e-01 -3.86120826e-01 -1.08353770e+00 7.33750701e-01
8.30838859e-01 -9.19052124e-01 7.44934618e-01 -7.42815554e-01
4.92254496e-01 4.98213395e-02 2.61030216e-02 -1.11847663e+00
-6.14141643e-01 -8.66352260e-01 -7.29568183e-01 1.71007526e+00
7.87753582e-01 -4.77164060e-01 5.33702135e-01 8.96038651e-01
4.20889072e-02 -1.28896344e+00 -4.87801611e-01 -6.24216676e-01
2.63028145e-01 -6.75277472e-01 8.00400019e-01 1.00021255e+00
3.11176330e-01 1.04906547e+00 -1.99551016e-01 -1.77296802e-01
2.18969490e-02 -5.35657518e-02 8.84366274e-01 -1.17598546e+00
-5.22405863e-01 -4.28249449e-01 9.86139178e-02 -1.13976789e+00
-1.86156034e-01 -8.42427671e-01 -5.18282428e-02 -1.24729455e+00
3.46724242e-01 -8.66415739e-01 -4.37156469e-01 1.00448608e+00
-6.20781064e-01 6.20562993e-02 3.21281016e-01 1.86976448e-01
-5.86034477e-01 2.35737398e-01 1.00263667e+00 3.72819975e-02
-4.23821002e-01 7.02822283e-02 -1.08640325e+00 5.34133971e-01
6.49080515e-01 -4.20171976e-01 -5.31979084e-01 -5.90131044e-01
6.32563174e-01 -3.45130920e-01 -1.06013827e-01 -6.99402392e-01
2.26893440e-01 3.64108719e-02 5.30332625e-01 -4.47227925e-01
7.34633133e-02 -3.92818689e-01 -1.27345115e-01 1.75654754e-01
-4.72493142e-01 -2.39063893e-02 2.46125400e-01 1.22188732e-01
-2.29242161e-01 -1.22521453e-01 7.05710828e-01 1.90949545e-03
-4.70532507e-01 1.21872447e-01 -1.12763122e-02 2.68087864e-01
6.11109316e-01 -8.75465572e-02 -5.17869830e-01 -2.29377493e-01
-8.27010930e-01 4.56682682e-01 -6.82784757e-03 6.36211634e-01
2.44389325e-01 -1.30770528e+00 -5.57791114e-01 3.64770651e-01
1.29135400e-01 4.72337633e-01 -1.35059536e-01 6.19387269e-01
3.36870141e-02 3.70014250e-01 1.69499427e-01 -1.90617710e-01
-1.10031462e+00 6.43765748e-01 2.14642510e-01 -7.60936439e-01
-3.12925130e-01 1.37551165e+00 -2.68185381e-02 -1.55572534e-01
5.22631168e-01 -5.20012081e-01 -2.14690998e-01 5.57656050e-01
5.10278881e-01 4.72868413e-01 3.51954937e-01 -2.95949399e-01
-3.08139771e-01 1.22452751e-01 -6.08551562e-01 -6.83761463e-02
1.53216600e+00 -9.01231840e-02 -8.00205208e-03 7.47550428e-01
7.66414702e-01 -1.22239522e-04 -1.28023648e+00 -4.56914216e-01
4.02030706e-01 -1.08405791e-01 -1.49123117e-01 -1.17575252e+00
-9.05972183e-01 8.63681197e-01 1.52324229e-01 2.52793282e-01
1.20083070e+00 1.47697866e-01 9.60549414e-01 2.81031013e-01
2.81023562e-01 -1.14584172e+00 8.93994421e-03 6.56166375e-01
7.54358232e-01 -1.19327331e+00 1.31372273e-01 -5.18248439e-01
-5.57471395e-01 1.04539382e+00 9.92050946e-01 1.81665123e-01
5.99144042e-01 7.40232348e-01 7.19958544e-02 -1.07534654e-01
-1.28565848e+00 -2.92804688e-01 2.30351403e-01 5.46598613e-01
6.95740998e-01 -1.01914003e-01 -2.51003176e-01 1.14411104e+00
-5.63123822e-01 8.07217509e-03 1.04318708e-01 9.21068072e-01
-2.42188156e-01 -1.37077880e+00 -3.96588966e-02 4.70087856e-01
-2.31323272e-01 -2.36589924e-01 -4.90668237e-01 8.71853173e-01
6.42559171e-01 1.01659751e+00 1.56600997e-02 -6.63345516e-01
5.75321019e-01 5.26546478e-01 5.64337254e-01 -7.60222793e-01
-8.72816920e-01 4.38836366e-02 5.10573149e-01 -4.82180089e-01
-2.10642219e-01 -6.94962740e-01 -1.17917156e+00 -9.29415450e-02
-6.34017944e-01 -5.93789890e-02 1.04799710e-01 9.81047332e-01
3.28830600e-01 7.67413378e-01 3.85357618e-01 -4.55431283e-01
-2.95084804e-01 -1.29626703e+00 -4.04662937e-02 2.20840618e-01
2.71046966e-01 -7.94925511e-01 6.40171859e-03 8.89101326e-02] | [9.226712226867676, 8.579191207885742] |
80c73f44-e17c-45b5-b914-98753ac58b30 | leveraging-language-foundation-models-for | 2209.05479 | null | https://arxiv.org/abs/2209.05479v2 | https://arxiv.org/pdf/2209.05479v2.pdf | Leveraging Language Foundation Models for Human Mobility Forecasting | In this paper, we propose a novel pipeline that leverages language foundation models for temporal sequential pattern mining, such as for human mobility forecasting tasks. For example, in the task of predicting Place-of-Interest (POI) customer flows, typically the number of visits is extracted from historical logs, and only the numerical data are used to predict visitor flows. In this research, we perform the forecasting task directly on the natural language input that includes all kinds of information such as numerical values and contextual semantic information. Specific prompts are introduced to transform numerical temporal sequences into sentences so that existing language models can be directly applied. We design an AuxMobLCast pipeline for predicting the number of visitors in each POI, integrating an auxiliary POI category classification task with the encoder-decoder architecture. This research provides empirical evidence of the effectiveness of the proposed AuxMobLCast pipeline to discover sequential patterns in mobility forecasting tasks. The results, evaluated on three real-world datasets, demonstrate that pre-trained language foundation models also have good performance in forecasting temporal sequences. This study could provide visionary insights and lead to new research directions for predicting human mobility. | ['Bhanu Prakash Voutharoja', 'Flora D. Salim', 'Hao Xue'] | 2022-09-11 | null | null | null | null | ['sequential-pattern-mining', 'temporal-sequences'] | ['natural-language-processing', 'reasoning'] | [ 1.60598144e-01 -2.42622510e-01 -7.30481982e-01 -7.70030081e-01
-3.67167234e-01 -2.56758898e-01 5.89832723e-01 2.54560500e-01
-5.26222169e-01 6.36357367e-01 7.73629606e-01 -7.42940068e-01
7.73041174e-02 -1.02828074e+00 -6.62570119e-01 -1.64203331e-01
-4.68025416e-01 4.33920443e-01 2.20381111e-01 -5.56582868e-01
1.04872055e-01 1.18642621e-01 -1.71356869e+00 8.01256835e-01
6.98087752e-01 1.07244754e+00 1.19264066e-01 5.35999894e-01
-5.11267364e-01 1.37994695e+00 -3.16400796e-01 -3.80520850e-01
-2.49762069e-02 -3.50000322e-01 -8.88839185e-01 -3.18562150e-01
-2.03482002e-01 -4.84125257e-01 -6.29240215e-01 3.75285387e-01
8.77161250e-02 3.92625898e-01 5.28790951e-01 -1.35481799e+00
-5.36261201e-01 1.01411653e+00 -2.35510886e-01 7.25589931e-01
6.00831628e-01 1.22365147e-01 1.26814270e+00 -4.12552238e-01
5.54810762e-01 1.23229587e+00 7.05801249e-01 2.62725890e-01
-6.53234065e-01 -4.83637482e-01 5.91377616e-01 5.14933705e-01
-1.05348492e+00 -3.00135702e-01 5.80872774e-01 -4.07883376e-01
1.45678341e+00 3.47241193e-01 4.78704780e-01 1.36253178e+00
1.73646331e-01 1.32439148e+00 3.32346618e-01 -7.49222413e-02
3.32990475e-02 7.08285393e-03 3.71848345e-01 4.69921440e-01
-2.16082901e-01 -3.84596623e-02 -7.17123449e-01 1.68467108e-02
2.05358818e-01 4.51394379e-01 3.04964602e-01 5.86252153e-01
-1.25099027e+00 4.91493315e-01 3.46156567e-01 4.88833785e-01
-6.30814195e-01 5.79131916e-02 6.52047515e-01 3.96618903e-01
6.05701447e-01 -8.93945545e-02 -6.85625732e-01 -7.69839168e-01
-9.32150662e-01 4.19745803e-01 7.64333367e-01 1.17438483e+00
6.41463399e-01 -2.44271800e-01 -4.91279542e-01 5.62224746e-01
3.11977714e-01 6.68643832e-01 7.62256622e-01 -4.81027216e-01
1.04795861e+00 7.16830552e-01 2.81130433e-01 -1.23990548e+00
-6.26072407e-01 -2.72072196e-01 -6.25700414e-01 -9.05973315e-01
4.61630404e-01 -1.52018115e-01 -6.83457971e-01 1.63519132e+00
1.03624836e-01 5.19804418e-01 -5.57750575e-02 8.07398200e-01
7.48720884e-01 1.06457722e+00 3.30713242e-01 7.78465867e-02
1.36343801e+00 -9.51858521e-01 -5.50999701e-01 -3.52636546e-01
1.16275811e+00 -2.76851028e-01 1.45339072e+00 -4.88872863e-02
-7.90556669e-01 -5.75037181e-01 -5.78636885e-01 -3.04244339e-01
-4.72221255e-01 2.02592567e-01 7.80364990e-01 3.86620343e-01
-6.97580695e-01 4.16604757e-01 -1.15026987e+00 -6.69192016e-01
2.56557971e-01 1.17960714e-01 1.29627377e-01 1.90926403e-01
-1.51169002e+00 4.63319182e-01 3.21440518e-01 1.28182679e-01
-4.54793602e-01 -5.85650861e-01 -8.98389161e-01 2.67388850e-01
4.09619398e-02 -4.57496107e-01 1.46115768e+00 -7.97520578e-01
-1.31952238e+00 5.86993754e-01 -9.40075099e-01 -8.56203496e-01
2.81434506e-01 6.35176301e-02 -9.48300064e-01 -2.94616997e-01
4.84331667e-01 3.49407405e-01 2.38815159e-01 -4.42361176e-01
-1.37955070e+00 -7.74717182e-02 4.16765846e-02 -1.19445935e-01
-6.50100708e-01 1.34833694e-01 -5.45030773e-01 -5.91067195e-01
-2.23206624e-01 -8.62934053e-01 -3.12446922e-01 -6.41406000e-01
-2.26661384e-01 -4.49854553e-01 4.77991939e-01 -8.43387783e-01
1.91552079e+00 -2.16378880e+00 -4.87214744e-01 4.61350083e-01
-2.14362025e-01 -7.51262307e-02 -1.13708206e-01 6.13209665e-01
5.32804430e-01 -4.38387535e-04 -1.69553950e-01 -3.84245902e-01
1.70459583e-01 3.33965957e-01 -9.66564596e-01 -1.03398003e-01
5.10851573e-03 1.42755294e+00 -1.04336262e+00 -3.66570175e-01
1.94341131e-02 1.56357601e-01 -6.86732829e-01 5.17880656e-02
-3.76155436e-01 6.36619985e-01 -5.63725591e-01 4.17754531e-01
3.40521455e-01 -5.61321259e-01 1.64972916e-01 2.57043749e-01
-2.87175685e-01 1.21053684e+00 -5.76604068e-01 1.47470272e+00
-8.02235901e-01 6.01880550e-01 -7.07089126e-01 -9.45661306e-01
7.41933107e-01 2.00065896e-01 7.06192493e-01 -1.12182236e+00
-1.84527427e-01 9.54451114e-02 -2.62689680e-01 -7.99499691e-01
7.80020177e-01 1.72326982e-01 -6.01376474e-01 5.39979696e-01
-3.68281066e-01 8.07519078e-01 5.43208361e-01 3.33279781e-02
1.23942220e+00 -3.07436194e-02 1.04521178e-01 1.58892602e-01
6.83224142e-01 4.18133050e-01 7.40376234e-01 7.06529558e-01
-8.52878168e-02 5.69585450e-02 4.89157766e-01 -8.50283802e-01
-9.14475322e-01 -8.88346851e-01 2.05882043e-01 1.53216457e+00
6.71930909e-02 -7.23780274e-01 -2.42620632e-01 -5.72089374e-01
4.06732224e-02 7.87198186e-01 -5.17902374e-01 9.70704481e-02
-9.10759866e-01 -6.27702296e-01 6.40070438e-01 9.07894790e-01
4.10255849e-01 -1.21062124e+00 -5.47692835e-01 4.73420858e-01
-7.12016225e-01 -1.32635081e+00 -6.39896512e-01 -3.05019796e-01
-8.47134531e-01 -5.68071902e-01 -3.74911487e-01 -9.06237662e-01
3.86205763e-01 2.94400275e-01 1.01738596e+00 -6.44396693e-02
2.40677640e-01 -5.15180677e-02 -6.06437147e-01 -2.04499871e-01
-6.00737184e-02 6.52867913e-01 -2.59280205e-02 2.27271840e-01
1.08744085e+00 -8.36619258e-01 -5.72540283e-01 5.10581195e-01
-6.27723694e-01 1.68102428e-01 2.75521249e-01 3.39931428e-01
3.75903755e-01 1.66409135e-01 7.59682477e-01 -9.62172925e-01
6.14116192e-01 -1.09323883e+00 -2.26332918e-01 9.10019353e-02
-2.62461126e-01 -3.40563804e-02 1.07527530e+00 -4.31300610e-01
-1.20547307e+00 -1.91836938e-01 -4.90865648e-01 2.54284114e-01
-2.58936852e-01 1.12671113e+00 -1.07341424e-01 8.48200321e-01
3.99462879e-01 5.41734695e-01 -5.28266370e-01 -5.78745604e-01
1.74885690e-01 1.06735814e+00 4.19882178e-01 -2.38383144e-01
3.17307234e-01 7.83425033e-01 -3.79948795e-01 -1.03611910e+00
-8.74115765e-01 -8.75256240e-01 -4.47131574e-01 1.96743775e-02
6.76616669e-01 -9.41319168e-01 -1.00475550e+00 2.44434372e-01
-1.05786669e+00 -5.10358214e-01 -1.82736233e-01 5.88841021e-01
-4.49495375e-01 8.11768323e-02 -8.58929217e-01 -8.49635065e-01
-1.72184527e-01 -7.25392520e-01 1.01111853e+00 -1.05550410e-02
-6.06676936e-01 -1.14450860e+00 -1.17382213e-01 3.51699054e-01
5.12845993e-01 -2.03503773e-01 1.01362967e+00 -6.60225630e-01
-5.01356840e-01 -2.28921592e-01 1.16413562e-02 -2.55463779e-01
1.86806619e-01 -3.65387619e-01 -7.09279835e-01 1.80673406e-01
-3.81501108e-01 1.11891858e-01 8.99137378e-01 3.17634046e-01
1.22308087e+00 -8.11798751e-01 -6.71933532e-01 6.15454316e-01
6.94550991e-01 4.13321257e-01 5.85190415e-01 3.93874675e-01
7.18460500e-01 7.05505013e-01 7.39224315e-01 7.73659885e-01
1.17521906e+00 6.86487257e-01 -2.02483699e-01 3.01813364e-01
2.23535255e-01 -9.88356352e-01 5.49941838e-01 9.40591455e-01
1.25667110e-01 -5.53908885e-01 -1.07730341e+00 8.57790232e-01
-2.16803002e+00 -1.27998877e+00 -2.40501210e-01 1.82346594e+00
5.13825476e-01 2.79671758e-01 5.48593700e-01 1.91607580e-01
3.29304606e-01 2.54825503e-01 -5.00233591e-01 -2.98729688e-01
1.28388137e-01 -2.69465655e-01 6.67594492e-01 5.02043784e-01
-1.10473347e+00 1.17387807e+00 5.75214481e+00 5.52829206e-01
-1.24735498e+00 1.27926573e-01 5.58982611e-01 -1.84142604e-01
-5.78452706e-01 -1.07272558e-01 -1.15412867e+00 1.00533807e+00
1.61164701e+00 -2.77827829e-01 3.58926356e-01 7.57760823e-01
8.48901987e-01 2.06002802e-01 -1.04329336e+00 9.87498999e-01
-2.52575755e-01 -1.38006639e+00 2.61632979e-01 5.43233939e-02
4.59590137e-01 1.95978165e-01 1.55094147e-01 6.18252575e-01
2.23224983e-01 -8.21169436e-01 7.38075674e-01 7.33175755e-01
3.81333053e-01 -6.51954412e-01 6.20329320e-01 6.49500132e-01
-1.58328032e+00 -3.82990539e-01 -1.12883896e-01 -5.66598177e-01
9.15066898e-01 5.32812655e-01 -1.02611792e+00 3.78805280e-01
6.96990848e-01 1.35146415e+00 -2.23025709e-01 8.11304450e-01
-6.98006377e-02 1.22925699e+00 -5.46001613e-01 -2.47493848e-01
4.51497912e-01 6.19141497e-02 1.82602242e-01 1.48435056e+00
4.56273764e-01 1.19169153e-01 2.44245738e-01 5.83978176e-01
-2.28559840e-02 3.06270540e-01 -6.43290222e-01 -1.72355756e-01
4.15313274e-01 5.86830795e-01 -7.90434539e-01 -5.45202076e-01
-6.46044493e-01 1.02256668e+00 6.08552285e-02 4.27931100e-01
-9.77275074e-01 -3.28482538e-01 8.11766803e-01 6.29459918e-01
3.93961132e-01 -4.85693634e-01 -4.89638239e-01 -1.28269148e+00
3.05720776e-01 -4.51208442e-01 6.78749144e-01 -3.75716388e-01
-1.21218979e+00 5.05751193e-01 1.04833819e-01 -1.29241228e+00
-6.86366498e-01 -4.74580497e-01 -7.04583526e-01 6.58603668e-01
-1.53726375e+00 -1.06659317e+00 1.92982741e-02 4.70324874e-01
6.53185904e-01 -1.43273011e-01 4.52903628e-01 8.50128233e-01
-4.74824786e-01 7.94850826e-01 3.22977640e-03 3.90666366e-01
1.41211122e-01 -7.08709598e-01 1.30583918e+00 8.06734920e-01
2.71872193e-01 6.63723111e-01 4.51731592e-01 -8.91167223e-01
-1.15814555e+00 -1.34827828e+00 1.95559275e+00 -4.99434710e-01
7.22609043e-01 -3.67088258e-01 -7.69165277e-01 1.08205533e+00
-4.59297150e-01 -2.30373576e-01 9.03148234e-01 2.24020958e-01
4.23922911e-02 -1.55304655e-01 -7.44168401e-01 6.21931434e-01
1.54173374e+00 -6.23727620e-01 -4.18135673e-01 1.69871628e-01
8.47371697e-01 -2.12832585e-01 -5.94970465e-01 4.27998938e-02
6.37946308e-01 -4.86744553e-01 8.44186425e-01 -1.03966427e+00
4.92007345e-01 -2.03192383e-01 -8.16489980e-02 -9.40688431e-01
-2.50665754e-01 -5.62892199e-01 -3.59846115e-01 1.06209970e+00
9.51247096e-01 -5.49176395e-01 8.60577047e-01 7.38498807e-01
-2.57351808e-02 -4.63456661e-01 -8.55114222e-01 -7.93375194e-01
-4.64656532e-01 -1.11889565e+00 1.12173235e+00 7.11942077e-01
4.13026601e-01 4.03453767e-01 -5.88013947e-01 1.58434704e-01
-3.97587121e-02 1.79627255e-01 8.28619480e-01 -1.05220032e+00
-1.86561614e-01 -2.70358801e-01 -4.29445863e-01 -1.83762717e+00
3.81167710e-01 -1.27615309e+00 -2.11679459e-01 -1.59004390e+00
-2.52670139e-01 -5.71354806e-01 -3.10803682e-01 6.14000738e-01
1.62932500e-01 -9.46935639e-02 -3.55836563e-02 4.10105914e-01
-8.40416193e-01 5.77359974e-01 8.08272898e-01 -2.70005047e-01
-7.32139111e-01 5.45725524e-01 -6.06478095e-01 5.98735690e-01
8.11293304e-01 -4.24380720e-01 -6.53815567e-01 -6.26229703e-01
6.00839794e-01 9.81449112e-02 9.95043367e-02 -8.82780313e-01
3.52485865e-01 -3.78385484e-01 -1.00186460e-01 -8.19511294e-01
1.12843543e-01 -7.82637119e-01 -2.71976292e-01 1.53006077e-01
-6.70924366e-01 3.03727657e-01 1.68909863e-01 6.35147691e-01
-2.66769290e-01 2.95013964e-01 -1.95265964e-01 1.08645290e-01
-9.55160081e-01 4.89581853e-01 -8.58951986e-01 1.42870918e-01
7.41017282e-01 -2.41317585e-01 1.38955086e-01 -6.80658400e-01
-7.35685766e-01 5.13391018e-01 8.41237046e-03 7.63138592e-01
4.88414496e-01 -1.42718375e+00 -5.90909421e-01 4.18719858e-01
3.15878242e-01 -3.40587169e-01 2.93635130e-01 9.16768849e-01
-2.90448874e-01 8.54897320e-01 -4.93835611e-03 -4.63237971e-01
-7.62138426e-01 6.45244062e-01 -3.06151509e-02 -4.23657238e-01
-8.02053392e-01 5.59419036e-01 -1.57023489e-01 -4.49599862e-01
3.22589815e-01 -1.16199577e+00 -3.38131458e-01 2.74813343e-02
8.22100937e-01 5.01383781e-01 -1.14010528e-01 -7.45001972e-01
-5.67997396e-01 5.81907816e-02 -7.81774521e-02 -2.49025866e-01
1.41050136e+00 -5.25602639e-01 7.78274983e-02 6.59481168e-01
1.35753298e+00 -1.38011143e-01 -1.03081846e+00 -5.57661891e-01
5.97283900e-01 -3.37848097e-01 -3.99358213e-01 -6.31519973e-01
-8.12975705e-01 8.39007854e-01 1.46382362e-01 1.70068830e-01
1.00566292e+00 -1.65874120e-02 1.71646988e+00 5.40779531e-01
4.61409867e-01 -1.01676285e+00 -4.54574674e-01 9.74995971e-01
2.99180597e-01 -1.19413757e+00 -7.24916399e-01 -8.18998367e-02
-1.02762783e+00 7.90082216e-01 3.13057214e-01 2.55873650e-01
1.04188895e+00 2.02176064e-01 -3.90739273e-03 -1.16465066e-03
-1.05318153e+00 -4.85267341e-01 2.76745558e-01 1.92370325e-01
6.51170254e-01 2.67646044e-01 -3.52712423e-01 1.09339404e+00
-7.11217105e-01 4.34128523e-01 2.11153597e-01 8.90086293e-01
-3.60839963e-01 -1.06240714e+00 1.33562520e-01 7.19788432e-01
-3.85668069e-01 -2.04312682e-01 6.89279661e-02 1.67611450e-01
7.60741904e-02 1.13477886e+00 4.95873988e-01 -6.72156572e-01
4.95148033e-01 2.29116499e-01 -2.15260670e-01 -4.51732576e-01
-5.26705801e-01 -6.37678087e-01 3.20596069e-01 -6.84935689e-01
-2.49682605e-01 -8.28669369e-01 -1.61828005e+00 -5.18457413e-01
6.09783053e-01 3.08998048e-01 3.23084652e-01 1.16432405e+00
6.60621524e-01 2.57631093e-01 5.72691441e-01 -4.75917459e-01
-7.10974913e-03 -8.04495633e-01 -3.36922258e-01 5.57205856e-01
4.90908325e-01 -3.39785159e-01 -1.13641657e-01 1.76919088e-01] | [6.581500053405762, 2.0792582035064697] |
2cf4a2aa-346c-413a-83a6-7e4d83369dae | going-for-goal-a-resource-for-grounded | 2211.04534 | null | https://arxiv.org/abs/2211.04534v1 | https://arxiv.org/pdf/2211.04534v1.pdf | Going for GOAL: A Resource for Grounded Football Commentaries | Recent video+language datasets cover domains where the interaction is highly structured, such as instructional videos, or where the interaction is scripted, such as TV shows. Both of these properties can lead to spurious cues to be exploited by models rather than learning to ground language. In this paper, we present GrOunded footbAlL commentaries (GOAL), a novel dataset of football (or `soccer') highlights videos with transcribed live commentaries in English. As the course of a game is unpredictable, so are commentaries, which makes them a unique resource to investigate dynamic language grounding. We also provide state-of-the-art baselines for the following tasks: frame reordering, moment retrieval, live commentary retrieval and play-by-play live commentary generation. Results show that SOTA models perform reasonably well in most tasks. We discuss the implications of these results and suggest new tasks for which GOAL can be used. Our codebase is available at: https://gitlab.com/grounded-sport-convai/goal-baselines. | ['Verena Rieser', 'Ioannis Konstas', 'Lu Yu', 'Malvina Nikandrou', 'Shubham Agarwal', 'Andrea Vanzo', 'Emanuele Bastianelli', 'José Lopes', 'Alessandro Suglia'] | 2022-11-08 | null | null | null | null | ['moment-retrieval'] | ['computer-vision'] | [ 5.28248474e-02 1.48914903e-02 -4.77521241e-01 -2.47296527e-01
-1.36459005e+00 -8.31129432e-01 8.38041127e-01 1.08557984e-01
-4.60107803e-01 7.70855784e-01 7.94588923e-01 -3.86308789e-01
5.24731040e-01 -3.67028922e-01 -1.05969179e+00 -3.89795691e-01
5.47911157e-04 1.31836012e-01 4.83114839e-01 -3.87098670e-01
4.15903270e-01 -1.26810551e-01 -1.49672806e+00 1.16508365e+00
4.19073135e-01 4.55943584e-01 3.07625353e-01 1.13157785e+00
5.63655086e-02 1.62982833e+00 -7.80854285e-01 -5.92596531e-01
-5.42120151e-02 -7.53184915e-01 -8.51332307e-01 -8.37155506e-02
8.10897708e-01 -4.18570161e-01 -6.33012176e-01 7.75411010e-01
5.98831892e-01 5.99283218e-01 3.14547569e-01 -1.34786332e+00
-4.87704575e-01 1.07924843e+00 -3.67356539e-01 6.27922773e-01
1.05599475e+00 2.00653180e-01 1.12532961e+00 -9.86917377e-01
1.14962339e+00 1.21113491e+00 6.49098873e-01 5.89287341e-01
-8.99909794e-01 -8.44284415e-01 3.64810646e-01 3.94561738e-01
-1.27186632e+00 -8.98909450e-01 5.57495773e-01 -7.44644403e-01
9.40406322e-01 2.49979720e-01 6.54699504e-01 1.77819371e+00
2.01011941e-01 1.18941700e+00 7.03162491e-01 -4.53713953e-01
-1.39541566e-01 -2.11777642e-01 -3.65617126e-02 9.06859517e-01
-1.21198751e-01 -2.14458942e-01 -1.36273956e+00 4.66023944e-02
3.66900951e-01 -3.14472049e-01 -3.75943512e-01 1.25670135e-01
-1.54925883e+00 7.31206775e-01 2.24241745e-02 1.30150974e-01
7.96826482e-02 3.43411714e-01 7.57176995e-01 3.61582875e-01
7.18384683e-01 3.80932003e-01 6.35190755e-02 -9.72518682e-01
-1.12604535e+00 8.64153862e-01 8.15151930e-01 1.22166324e+00
3.31932604e-01 6.10859059e-02 -3.81592810e-01 6.81735158e-01
1.27616867e-01 3.52508694e-01 5.30593991e-01 -9.79662776e-01
9.48102474e-01 -3.43952104e-02 2.49182004e-02 -1.09188116e+00
-1.21719487e-01 1.49712652e-01 -9.09366310e-02 -2.90860564e-01
6.58035159e-01 -2.81366318e-01 -5.39186954e-01 1.56247783e+00
-4.61145975e-02 5.21434426e-01 -2.31129482e-01 7.98920751e-01
1.49873710e+00 8.04479063e-01 2.40915101e-02 -1.31562948e-01
1.22735882e+00 -1.25071979e+00 -8.15854073e-01 -1.42477393e-01
9.08899248e-01 -1.13888216e+00 1.41768944e+00 4.08988953e-01
-1.36997020e+00 -4.00216192e-01 -8.33314538e-01 -4.42095578e-01
-1.65220276e-02 -2.43593365e-01 2.63861775e-01 1.42671347e-01
-8.55783224e-01 5.71595907e-01 -1.12792754e+00 -3.59949619e-01
2.25474656e-01 -2.48102814e-01 -2.54327804e-01 -6.34004772e-02
-1.13355517e+00 5.50539732e-01 1.65556952e-01 -1.18776366e-01
-1.06011355e+00 -6.73855186e-01 -1.02933109e+00 -4.26687628e-01
5.99596977e-01 -1.32572293e-01 1.89932299e+00 -1.13513494e+00
-1.40400553e+00 1.34140253e+00 -3.26388568e-01 -4.36363459e-01
7.93787956e-01 -5.51264822e-01 -2.61554658e-01 3.46100599e-01
3.17673475e-01 5.91118157e-01 5.49385130e-01 -8.41168046e-01
-6.75337851e-01 2.28249654e-01 4.80454117e-01 2.85404295e-01
-9.10063684e-02 4.33708429e-01 -6.71395183e-01 -8.39901626e-01
-3.01235557e-01 -1.23823977e+00 2.19275802e-01 -4.31454182e-01
-5.28682172e-01 -1.55718476e-01 5.30146599e-01 -7.25492835e-01
1.68066394e+00 -2.28664398e+00 1.59952804e-01 -3.49709183e-01
1.50362834e-01 -1.78877905e-01 -7.04531223e-02 8.12046885e-01
4.54222076e-02 3.33630294e-01 3.14611942e-01 -4.53686565e-01
-9.20181349e-02 4.02720086e-02 -4.31592047e-01 5.80144286e-01
-1.18325420e-01 8.82114649e-01 -1.10444653e+00 -7.06460178e-01
-1.51537374e-01 2.08982721e-01 -7.72277117e-01 2.35356256e-01
-4.81290400e-01 5.31592488e-01 -3.65319043e-01 4.98683244e-01
-1.60662532e-01 -1.54016420e-01 3.60993557e-02 6.76035807e-02
-3.52201819e-01 8.13539922e-01 -8.56253624e-01 2.02455783e+00
-3.39181721e-01 1.36448932e+00 -1.48813277e-01 -6.64174795e-01
3.95711929e-01 4.70500529e-01 3.04445237e-01 -6.04689658e-01
-2.25726026e-03 -1.12889931e-01 -9.71383974e-02 -8.29480588e-01
9.07439053e-01 4.40690666e-02 -2.74996281e-01 5.34431934e-01
1.21183209e-01 -3.08228731e-01 7.58073628e-01 6.44793808e-01
1.12472236e+00 4.13537115e-01 2.45882630e-01 -1.81723312e-01
9.41318572e-02 3.74195427e-01 3.75374943e-01 9.76079285e-01
-2.25615963e-01 9.66227055e-01 6.38759553e-01 -4.93165463e-01
-9.28452015e-01 -7.92171478e-01 2.07212105e-01 1.57429922e+00
1.53232254e-02 -1.04506516e+00 -5.88765860e-01 -4.07070786e-01
-3.40182036e-01 6.72652841e-01 -3.69847953e-01 1.14635505e-01
-8.77994120e-01 -1.81487679e-01 5.71391165e-01 4.71520305e-01
2.07192630e-01 -1.06383479e+00 -3.77286017e-01 2.35419512e-01
-8.21452081e-01 -1.44490719e+00 -8.75958025e-01 -2.66947269e-01
-5.22786319e-01 -1.17255485e+00 -5.48340499e-01 -7.71199226e-01
1.84194937e-01 4.26637769e-01 1.57378614e+00 3.96080136e-01
1.80689514e-01 5.58206558e-01 -7.62045503e-01 -5.66672623e-01
-5.63810885e-01 2.35436648e-01 -1.11475863e-01 -3.30839097e-01
4.00718957e-01 -1.79563001e-01 -2.78330564e-01 1.50880367e-01
-8.22787166e-01 4.68430012e-01 -1.38502926e-01 7.79054046e-01
4.74538684e-01 -4.72678751e-01 1.33593440e-01 -1.23796487e+00
6.75211430e-01 -6.01153255e-01 -2.92765170e-01 -1.18511647e-01
2.36357093e-01 -4.55066144e-01 4.88829523e-01 -5.20732284e-01
-8.94530535e-01 -3.19567949e-01 -1.04174092e-01 -3.14238578e-01
-3.15468833e-02 8.10681224e-01 2.20023721e-01 4.04881537e-01
9.79615271e-01 -1.50285125e-01 -2.52109140e-01 -1.95780531e-01
1.54840127e-01 5.67093909e-01 5.66479087e-01 -9.51329410e-01
5.16318142e-01 3.25751215e-01 -5.30039668e-01 -1.14437068e+00
-1.08566129e+00 -5.10100067e-01 -4.17167842e-01 -6.56346679e-01
9.20169830e-01 -1.54438853e+00 -5.95977902e-01 2.90794611e-01
-1.11496139e+00 -9.94495034e-01 -1.54748023e-01 5.62211990e-01
-7.43578255e-01 1.24635078e-01 -1.03588843e+00 -4.97972608e-01
1.76313639e-01 -1.24227679e+00 1.02185047e+00 1.32265314e-01
-7.71255255e-01 -9.27721739e-01 -5.27766440e-03 6.53332353e-01
-2.06958279e-01 5.17427444e-01 4.02009100e-01 -6.29067183e-01
-5.78580797e-01 -1.58932190e-02 3.69423926e-01 -2.31127199e-02
-2.18665794e-01 5.59385657e-01 -8.57573152e-01 -3.60689610e-01
-4.04716462e-01 -7.63533354e-01 6.37160420e-01 2.45264128e-01
1.06292284e+00 -4.77880597e-01 -1.00080997e-01 5.45885563e-01
9.14198220e-01 1.12696819e-01 5.01778483e-01 4.12818700e-01
6.56725883e-01 3.14032078e-01 8.26576889e-01 5.17554104e-01
6.03226244e-01 6.48045123e-01 1.45590575e-02 3.23451161e-01
-1.99868605e-01 -6.90958321e-01 7.95034051e-01 1.38923037e+00
-2.57589042e-01 -6.94991767e-01 -1.21415210e+00 7.68061042e-01
-2.00973868e+00 -1.47347224e+00 -2.12439969e-01 1.87172294e+00
9.39616680e-01 3.99169505e-01 4.07214880e-01 -2.36115754e-01
7.35861659e-01 5.83120584e-01 -3.09771616e-02 -4.78919595e-01
2.76660360e-03 5.12524545e-02 1.39624372e-01 6.99359238e-01
-1.02938437e+00 1.03171909e+00 5.87939548e+00 7.06008673e-01
-1.22561884e+00 2.85843313e-01 5.94632924e-01 -6.56655073e-01
-3.37782860e-01 -1.06658466e-01 -9.65926647e-01 6.55801773e-01
1.21131635e+00 -3.29742014e-01 2.54027396e-01 7.56499231e-01
5.35779178e-01 -1.38248369e-01 -1.29443777e+00 9.27819371e-01
3.50173533e-01 -1.58439922e+00 -1.23045243e-01 -3.47708076e-01
9.69834983e-01 2.94481009e-01 -5.74733056e-02 4.83020633e-01
5.85379601e-01 -8.99088085e-01 1.31360698e+00 1.92384750e-01
9.18268859e-01 -4.53754485e-01 3.30495268e-01 2.65914112e-01
-1.04824209e+00 2.49535024e-01 -7.65301064e-02 -3.05259436e-01
2.91278541e-01 6.04745820e-02 -7.24532604e-01 1.49122074e-01
8.63803864e-01 1.21741068e+00 -4.82305288e-01 7.78478503e-01
-3.93372744e-01 9.55018759e-01 -1.03727825e-01 -1.61574036e-01
4.49001253e-01 1.15190357e-01 7.17908263e-01 1.45807230e+00
2.94139385e-01 4.23901677e-01 5.83817005e-01 3.78551453e-01
-1.82988033e-01 4.32163291e-02 -9.14061606e-01 -4.32193935e-01
4.66479242e-01 9.76815879e-01 -7.74802327e-01 -4.20378417e-01
-7.68682778e-01 6.82632744e-01 2.00334594e-01 4.36409354e-01
-9.75716472e-01 -1.24167226e-01 6.38917983e-01 5.48929393e-01
2.11085733e-02 -4.98992294e-01 2.65597671e-01 -1.72788882e+00
-1.53666094e-01 -1.36281037e+00 4.71797258e-01 -9.98360574e-01
-9.82622862e-01 4.64992195e-01 2.64114141e-01 -1.35993981e+00
-5.42063951e-01 -2.85306215e-01 -5.72970629e-01 1.89063802e-01
-1.09298396e+00 -9.07537341e-01 -3.40055853e-01 7.61532247e-01
1.14548171e+00 -8.79297480e-02 3.01844835e-01 4.47833657e-01
-4.46669102e-01 4.72722173e-01 -1.66934192e-01 3.58834445e-01
1.18278062e+00 -9.87808347e-01 3.96777302e-01 1.06897771e+00
6.05492115e-01 4.51272160e-01 1.04330218e+00 -6.58206165e-01
-1.32541025e+00 -7.57108510e-01 9.31521595e-01 -7.43444502e-01
1.01708686e+00 -5.69062054e-01 -7.10132062e-01 1.11103892e+00
5.02390027e-01 -7.83702731e-02 9.93413568e-01 1.98824272e-01
-1.45784006e-01 2.27478027e-01 -3.97214860e-01 9.84850347e-01
1.31354880e+00 -8.34717214e-01 -5.50596952e-01 7.23515093e-01
7.97733307e-01 -1.23448861e+00 -5.42601347e-01 -9.03518796e-02
6.39833987e-01 -9.60723460e-01 7.27253854e-01 -9.07552540e-01
1.19047558e+00 -9.06067714e-03 -1.98080286e-01 -1.20298433e+00
-8.90214071e-02 -1.01451886e+00 -9.99905318e-02 1.32874405e+00
4.61974353e-01 6.86591417e-02 7.72059560e-01 5.62763393e-01
-5.64308643e-01 -3.42934012e-01 -7.34396398e-01 -4.86965537e-01
1.05184697e-01 -6.76261067e-01 1.56035140e-01 1.08176708e+00
2.66795963e-01 3.46621662e-01 -3.59438837e-01 -2.22822428e-01
5.56175373e-02 -1.23293504e-01 1.07315898e+00 -6.96187377e-01
-2.72429794e-01 -2.94068247e-01 -3.01311344e-01 -1.06666708e+00
5.67509055e-01 -8.15527260e-01 2.85201728e-01 -1.16171563e+00
1.03966638e-01 -4.42346409e-02 4.46909815e-02 3.88967663e-01
-1.18950129e-01 4.50782955e-01 5.34102321e-01 2.11887673e-01
-1.04457426e+00 9.43642259e-02 1.32115853e+00 -3.73088345e-02
-1.87765211e-01 -8.05375576e-02 -4.62870806e-01 7.59579420e-01
4.92908239e-01 -3.64326268e-01 -4.26825672e-01 -7.02484965e-01
5.54482520e-01 3.46093625e-01 1.91179156e-01 -9.96044099e-01
3.16944599e-01 -3.47373635e-01 -8.50338042e-02 -3.96019489e-01
2.26387680e-01 -2.79369980e-01 7.18685687e-02 1.05002478e-01
-8.03632379e-01 5.41743696e-01 3.66004318e-01 3.50948781e-01
-5.98771393e-01 -2.20080450e-01 4.60386425e-01 -3.87873143e-01
-8.47580433e-01 1.09065950e-01 -8.80827725e-01 8.25842559e-01
8.75869215e-01 -2.32517347e-01 -5.44021308e-01 -9.58156884e-01
-6.44112468e-01 1.00794286e-01 5.64040601e-01 7.70406246e-01
3.94115746e-01 -1.28057730e+00 -9.30848002e-01 -1.91459835e-01
2.60013282e-01 4.86375131e-02 1.94274619e-01 7.45398045e-01
-9.49015677e-01 3.41602385e-01 5.80364466e-02 -7.49914050e-01
-1.56779623e+00 1.58621475e-01 -1.40282229e-01 7.26687675e-03
-7.43351400e-01 1.09921730e+00 1.32544339e-01 9.82254520e-02
2.01415271e-01 -4.67340112e-01 -5.72982468e-02 3.88692468e-01
7.40770042e-01 1.76937133e-01 3.50992684e-03 -7.84300983e-01
-2.28865519e-01 3.93520929e-02 -2.12048471e-01 -2.47690454e-01
1.33200598e+00 -2.31596276e-01 2.16397837e-01 1.01755774e+00
1.05782914e+00 3.65717947e-01 -1.16746306e+00 -2.12274596e-01
-9.49809551e-02 -5.83835065e-01 -1.15970477e-01 -3.09477806e-01
-7.53307700e-01 7.72616684e-01 -1.52309015e-01 1.32580131e-01
6.60221040e-01 -1.13532588e-01 8.17193031e-01 2.66781688e-01
3.21774036e-01 -1.01437235e+00 4.54585552e-01 7.92288899e-01
9.43205774e-01 -1.33568168e+00 9.76984054e-02 -1.40438005e-01
-9.85326052e-01 1.06786692e+00 7.03462243e-01 -1.05775297e-01
4.35020566e-01 4.41228241e-01 1.48604497e-01 -5.38974367e-02
-1.27359641e+00 5.62201366e-02 3.48988682e-01 2.73478597e-01
1.05227947e+00 -8.37677121e-02 -3.09647113e-01 4.96128708e-01
-5.86475313e-01 3.18764225e-02 1.10951984e+00 1.21305656e+00
-1.76803410e-01 -7.64631510e-01 -2.91995466e-01 2.80502051e-01
-9.14947808e-01 -3.32555056e-01 -5.52708089e-01 9.35098529e-01
-2.11615220e-01 1.03234267e+00 2.06363931e-01 -3.24255943e-01
2.21339881e-01 1.71188071e-01 4.14160848e-01 -1.10672462e+00
-8.34686399e-01 1.27728283e-01 5.86336076e-01 -7.90679216e-01
-6.03670597e-01 -8.04472864e-01 -1.04128182e+00 -8.38956475e-01
-7.28550926e-02 1.57284796e-01 3.22692066e-01 8.19167078e-01
1.11251533e-01 4.38246757e-01 -4.34748642e-03 -9.12875891e-01
1.18243277e-01 -8.78617167e-01 -2.80005932e-01 5.88410735e-01
4.16734755e-01 -5.73937953e-01 -2.76423544e-01 6.95736885e-01] | [10.447897911071777, 0.7943729162216187] |
80edd125-e491-4817-97d7-be415cab028d | encore-pre-training-entity-encoders-using | 2305.12924 | null | https://arxiv.org/abs/2305.12924v1 | https://arxiv.org/pdf/2305.12924v1.pdf | EnCore: Pre-Training Entity Encoders using Coreference Chains | Entity typing is the task of assigning semantic types to the entities that are mentioned in a text. Since obtaining sufficient amounts of manual annotations is expensive, current state-of-the-art methods are typically trained on automatically labelled datasets, e.g. by exploiting links between Wikipedia pages. In this paper, we propose to use coreference chains as an additional supervision signal. Specifically, we pre-train an entity encoder using a contrastive loss, such that entity embeddings of coreferring entities are more similar to each other than to the embeddings of other entities. Since this strategy is not tied to Wikipedia, we can pre-train our entity encoder on other genres than encyclopedic text and on larger amounts of data. Our experimental results show that the proposed pre-training strategy allows us to improve the state-of-the-art in fine-grained entity typing, provided that only high-quality coreference links are exploited. | ['Steven Schockaert', 'Frank Mtumbuka'] | 2023-05-22 | null | null | null | null | ['entity-embeddings', 'entity-typing'] | ['methodology', 'natural-language-processing'] | [-1.30549103e-01 5.09713113e-01 -3.57422113e-01 -3.38827193e-01
-6.13981366e-01 -7.66982496e-01 6.81106806e-01 6.32548809e-01
-1.04504597e+00 8.99195790e-01 2.64091730e-01 4.64984775e-02
3.04092020e-02 -9.29636478e-01 -1.15823793e+00 -1.81272522e-01
-5.87441958e-02 7.45949209e-01 4.56838727e-01 -2.40992501e-01
9.22825467e-03 9.37461555e-02 -1.59062397e+00 1.08802855e-01
1.16637623e+00 6.88933671e-01 3.05969357e-01 2.97787309e-01
-3.35617781e-01 4.82070684e-01 -4.67841357e-01 -9.96335983e-01
-1.59800261e-01 -7.60121504e-04 -1.08700037e+00 -3.26148123e-01
4.78610575e-01 4.63916622e-02 -3.44146848e-01 1.21511388e+00
2.87070394e-01 4.78331186e-02 6.88443959e-01 -9.33532059e-01
-6.36379421e-01 8.19370687e-01 -2.99549878e-01 -1.75525565e-02
2.61589766e-01 -4.74759698e-01 1.52365255e+00 -8.60682964e-01
9.83619809e-01 1.02200365e+00 8.01069677e-01 7.02394664e-01
-1.27212560e+00 -4.83858436e-01 1.96671158e-01 3.34483743e-01
-1.24663031e+00 -3.26486111e-01 5.35095572e-01 -5.42839229e-01
9.13231194e-01 6.25878060e-03 9.27599445e-02 1.15620208e+00
-4.56948608e-01 6.51246130e-01 5.86185277e-01 -6.71673179e-01
-3.75296399e-02 5.19148707e-01 4.22854275e-01 6.74471736e-01
6.15940332e-01 -2.26168767e-01 -3.99395525e-01 -1.13132186e-02
2.93695867e-01 -3.12899292e-01 -5.21071434e-01 -7.01898515e-01
-1.07823169e+00 7.51576066e-01 5.69223762e-01 4.52092439e-01
-3.09432685e-01 -8.27271417e-02 5.54561257e-01 1.23424558e-02
4.66796547e-01 8.37434947e-01 -6.69841766e-01 -7.29897097e-02
-7.27091312e-01 9.25423950e-02 1.01002634e+00 1.05170739e+00
1.02644706e+00 -7.40274727e-01 2.19482798e-02 9.78818715e-01
2.23520294e-01 1.51642457e-01 4.09496933e-01 -7.03201950e-01
8.65348458e-01 8.76086831e-01 4.14653450e-01 -6.94574177e-01
-3.58341128e-01 -4.46687162e-01 -4.96747524e-01 -2.36874521e-01
5.60727477e-01 -3.56672525e-01 -6.25098288e-01 1.98737669e+00
4.16695207e-01 1.12813003e-01 1.62374824e-01 7.29910791e-01
6.71708584e-01 1.75166786e-01 1.71825305e-01 1.43206373e-01
1.68534970e+00 -9.21107590e-01 -7.71252692e-01 -2.53189653e-01
9.02351737e-01 -4.80871975e-01 1.02656829e+00 -1.41899005e-01
-7.08331823e-01 -3.67188185e-01 -9.56538975e-01 -2.47258410e-01
-6.74315751e-01 3.78969789e-01 4.17418271e-01 2.38493040e-01
-4.45940435e-01 8.40739489e-01 -9.74537373e-01 -4.48858976e-01
2.74410158e-01 2.46572837e-01 -7.24942982e-01 -4.17202748e-02
-1.67599535e+00 1.09459460e+00 6.30993783e-01 -7.16295466e-02
-4.31999564e-01 -8.25134814e-01 -1.14317060e+00 2.95387447e-01
5.40091872e-01 -7.34748423e-01 1.12710226e+00 -8.46275985e-01
-9.97286201e-01 1.08358276e+00 -2.03025252e-01 -4.02589202e-01
3.77862275e-01 -4.88857329e-01 -3.95952553e-01 1.01071723e-01
3.38259220e-01 4.28014904e-01 2.38309845e-01 -1.22778940e+00
-8.62256765e-01 -4.31330651e-01 5.53723156e-01 1.61723718e-01
-6.31531119e-01 1.69359371e-02 -6.79200351e-01 -4.32965696e-01
-4.63563532e-01 -9.48351622e-01 -3.24285142e-02 -1.80185720e-01
-5.23781240e-01 -6.19075418e-01 4.50808823e-01 -6.31029129e-01
1.39973247e+00 -2.08446360e+00 4.71295655e-01 -6.45594597e-02
1.35877803e-01 3.72928947e-01 1.26915544e-01 4.55008417e-01
9.48466584e-02 1.33756623e-01 -2.20028698e-01 -5.27315617e-01
3.97044867e-01 2.17843577e-01 2.86416244e-03 1.62623554e-01
1.52217925e-01 5.78810990e-01 -1.24627864e+00 -4.66277272e-01
-2.13255212e-01 4.70115513e-01 -6.53608084e-01 3.73994589e-01
-3.25880885e-01 1.44066170e-01 -4.72682267e-01 -2.58325413e-02
3.81934434e-01 -4.50804293e-01 3.94077718e-01 -4.84565943e-01
-1.25929773e-01 8.81915808e-01 -1.12797606e+00 1.87655628e+00
-8.12728465e-01 5.49762368e-01 -4.75461669e-02 -9.28179860e-01
5.59691012e-01 3.67500663e-01 1.37418315e-01 -2.41350785e-01
-1.45402893e-01 4.95738566e-01 -1.01452574e-01 -4.87535357e-01
6.70608759e-01 2.40694713e-02 -2.70525575e-01 2.45191187e-01
6.37744367e-01 5.67779183e-01 5.91942966e-01 2.60298759e-01
9.47057605e-01 2.13463232e-01 2.74178982e-01 -1.72154725e-01
5.59714854e-01 1.10288681e-02 7.69594669e-01 4.97396231e-01
2.95220286e-01 2.11930200e-01 5.08169055e-01 1.09471105e-01
-1.10205841e+00 -8.26723337e-01 -3.93685758e-01 1.14741683e+00
3.20126593e-01 -7.08965182e-01 -9.98238504e-01 -1.29212642e+00
1.60026357e-01 6.39450431e-01 -7.98948109e-01 5.27573824e-02
-6.33610785e-01 -5.05762041e-01 5.88727295e-01 6.93118572e-01
3.32171410e-01 -9.33352232e-01 -2.76360512e-01 3.40644389e-01
-2.95304120e-01 -1.33313143e+00 -4.63712394e-01 3.33742768e-01
-4.62228119e-01 -1.30787647e+00 -6.33760870e-01 -8.36191714e-01
7.54538655e-01 -2.76483446e-01 1.30207646e+00 4.10184823e-02
6.36164621e-02 7.34171346e-02 -4.59263057e-01 -1.38924599e-01
-1.77958906e-01 6.71014607e-01 8.38426724e-02 3.52884680e-02
6.85337782e-01 -3.78560215e-01 -3.95543516e-01 9.83377993e-02
-5.92996478e-01 -1.82135463e-01 4.98888493e-01 1.17418540e+00
4.20979619e-01 -6.62881732e-02 4.80765700e-01 -1.57548583e+00
2.50485152e-01 -4.92007881e-01 -5.79222381e-01 5.02816498e-01
-4.74688619e-01 5.57252884e-01 8.89313877e-01 -3.06681603e-01
-1.26475441e+00 -3.50146666e-02 -2.61689544e-01 -2.24061772e-01
-2.58530766e-01 7.58001506e-01 -4.31000143e-01 2.19516739e-01
5.22613227e-01 -1.70776471e-01 -6.72277808e-01 -9.33866084e-01
4.82944578e-01 8.37454617e-01 8.50707531e-01 -8.22893560e-01
7.77168334e-01 1.11809775e-01 -3.05683464e-01 -5.40337622e-01
-1.17749643e+00 -6.39801502e-01 -9.25179183e-01 3.11497539e-01
9.43876266e-01 -9.42562342e-01 -4.96178627e-01 2.52388209e-01
-1.15624988e+00 -1.63315117e-01 -1.14317439e-01 6.35179758e-01
-3.04779798e-01 2.97220230e-01 -6.68407738e-01 -3.44802588e-01
-1.42091498e-01 -7.94711590e-01 9.15406883e-01 2.41726026e-01
-1.56734556e-01 -1.14597583e+00 3.61457378e-01 2.41597697e-01
1.14762150e-02 -4.73866984e-02 8.90704513e-01 -1.02713907e+00
-3.98688346e-01 -1.23534925e-01 -2.91271150e-01 1.65575281e-01
2.26994425e-01 -1.23126663e-01 -9.30701077e-01 -1.08969942e-01
-5.79706132e-01 -2.93146431e-01 9.01887596e-01 -2.66618133e-01
7.63421714e-01 -5.02865553e-01 -8.00731301e-01 5.70256710e-01
1.45834553e+00 -3.33912104e-01 4.49352562e-01 6.77676857e-01
9.94275153e-01 7.41033733e-01 7.57871270e-01 2.19335392e-01
8.23668957e-01 9.16518450e-01 1.88996658e-01 9.39984918e-02
-8.73515829e-02 -5.80009758e-01 6.63623437e-02 7.56858766e-01
-8.81792456e-02 -2.85791457e-01 -8.00361514e-01 9.78505731e-01
-1.93735361e+00 -9.17912602e-01 -2.10756324e-02 2.29108262e+00
1.44179320e+00 2.12854460e-01 -1.30436465e-01 -1.74975425e-01
9.12066221e-01 -8.30464959e-02 -3.68801445e-01 9.86551791e-02
3.74252237e-02 1.44849241e-01 6.47127151e-01 6.18643641e-01
-1.42680490e+00 9.78848755e-01 4.60656977e+00 5.24183989e-01
-8.68766904e-01 2.14019969e-01 -2.14105889e-01 2.49174148e-01
-3.84481132e-01 4.77053858e-02 -1.24944031e+00 7.51786172e-01
8.40252757e-01 -1.81656718e-01 1.79064170e-01 7.42261767e-01
-3.45350206e-01 1.59989059e-01 -1.43621123e+00 4.80666995e-01
-6.71292022e-02 -1.16048682e+00 -3.76524687e-01 -4.61854339e-02
6.12796009e-01 2.05418095e-01 -3.91466856e-01 4.78785902e-01
6.58538342e-01 -6.36117399e-01 6.25465512e-01 3.19921076e-01
7.96161413e-01 -6.12835646e-01 1.04067349e+00 3.64642411e-01
-1.13138163e+00 1.82871521e-01 -3.74289304e-01 4.42823648e-01
1.62629813e-01 6.17170215e-01 -5.50725102e-01 8.48538280e-01
5.98799646e-01 8.35283518e-01 -4.95744646e-01 1.13126981e+00
-7.57188082e-01 4.04083014e-01 -4.09674793e-01 8.50067735e-02
-3.43139842e-02 2.08824515e-01 4.40076888e-01 1.56195128e+00
2.32678100e-01 -1.30604073e-01 3.14484425e-02 7.22555816e-01
-6.74578071e-01 -4.09962004e-03 -4.18159515e-01 -1.58952996e-01
9.00766492e-01 1.37637472e+00 -1.43344596e-01 -5.07535815e-01
-8.12830746e-01 1.09175432e+00 1.00620198e+00 1.87663808e-01
-6.78668916e-01 -1.00993347e+00 7.96398878e-01 -1.15562761e-02
7.78414905e-01 -1.22383852e-02 1.68883204e-01 -1.55464447e+00
9.12345648e-02 -5.54825962e-01 6.14922583e-01 -4.38575059e-01
-1.44120657e+00 7.00047493e-01 -2.26916686e-01 -1.00910187e+00
-3.18353593e-01 -5.91803670e-01 -2.60945946e-01 8.09014976e-01
-1.90592849e+00 -1.17970204e+00 -1.81295246e-01 3.18965673e-01
3.23579758e-02 1.42306060e-01 1.03524685e+00 7.12589622e-01
-6.57572031e-01 8.85584712e-01 1.57755524e-01 6.83395147e-01
1.10814703e+00 -1.70555830e+00 2.42515281e-01 8.38153780e-01
3.41327012e-01 9.70737100e-01 7.30408490e-01 -4.67373580e-01
-8.57679248e-01 -1.15973699e+00 1.41797519e+00 -5.07623553e-01
8.63079786e-01 -5.04200637e-01 -1.20410526e+00 9.24593925e-01
3.01871449e-01 2.27843434e-01 7.12572038e-01 8.05747628e-01
-7.02293694e-01 -5.90615487e-03 -8.17444742e-01 3.09907079e-01
1.10489655e+00 -8.15118134e-01 -1.03388870e+00 -6.35822192e-02
6.03250325e-01 -5.07003725e-01 -1.22146547e+00 2.53390372e-01
3.50254118e-01 -3.46957177e-01 8.10634375e-01 -7.78788924e-01
3.79943460e-01 -4.18242604e-01 6.65074307e-03 -1.66453862e+00
-1.32095963e-01 -2.25143790e-01 -3.83248001e-01 1.70068038e+00
7.32384145e-01 -4.88856137e-01 5.61163545e-01 5.54766059e-01
-2.33465984e-01 -3.46748024e-01 -6.60883784e-01 -8.12173903e-01
-3.98147292e-02 2.23716199e-01 4.75865752e-01 1.34575915e+00
4.49450135e-01 7.76269376e-01 -9.23913419e-02 4.54079986e-01
7.58366048e-01 1.76302880e-01 5.71623385e-01 -1.60725808e+00
-4.29741949e-01 -1.75875753e-01 -3.07801843e-01 -1.05758250e+00
7.66388416e-01 -9.96469378e-01 1.57996729e-01 -1.41584885e+00
2.94150978e-01 -7.73045421e-01 -4.88851339e-01 5.85016549e-01
-5.91455400e-01 1.08694114e-01 5.23466710e-03 1.04296699e-01
-8.04550886e-01 4.76945817e-01 8.78667712e-01 -1.60342649e-01
1.90850005e-01 -3.31597924e-01 -6.74373031e-01 6.86138451e-01
4.13882434e-01 -6.24774456e-01 -6.24767505e-02 -6.83628440e-01
3.32089782e-01 -2.66423583e-01 1.91114277e-01 -6.84640706e-01
4.40960586e-01 2.03505486e-01 5.57676889e-03 -1.40257210e-01
2.10685089e-01 -7.28474736e-01 -1.52653813e-01 -8.20891783e-02
-4.32639390e-01 -3.07236046e-01 4.20517139e-02 5.41079640e-01
-4.88165885e-01 -6.80009365e-01 5.55747867e-01 -2.56160963e-02
-8.09520006e-01 1.10567145e-01 1.09014012e-01 6.60142124e-01
6.24862790e-01 3.89525652e-01 -5.16574621e-01 1.77464589e-01
-7.77367055e-01 2.72491544e-01 8.08965743e-01 3.15592527e-01
-9.88098681e-02 -1.37096167e+00 -4.41295207e-01 -1.19784497e-01
4.43399251e-01 1.65438071e-01 -2.20585577e-02 7.17585921e-01
-9.57261994e-02 6.18042231e-01 -1.90399855e-01 -2.33136207e-01
-1.27202237e+00 6.25782967e-01 2.82832801e-01 -6.33429110e-01
-5.23074567e-01 7.92609096e-01 2.36199185e-01 -6.76798105e-01
2.61462152e-01 -1.60987135e-02 -5.32369435e-01 3.81562948e-01
3.58012021e-01 3.36337797e-02 2.67771542e-01 -6.32939517e-01
-5.65538108e-01 5.01981616e-01 -3.73041928e-01 4.28317487e-02
1.48579133e+00 -1.45640209e-01 4.96097319e-02 3.75502706e-01
1.32379854e+00 3.13282698e-01 -1.12725902e+00 -4.20955718e-01
5.76201260e-01 -2.94330239e-01 -1.09339729e-01 -7.03763127e-01
-9.33111966e-01 8.82685304e-01 1.23756140e-01 1.72351584e-01
6.88274264e-01 2.99336672e-01 7.32697189e-01 5.92471600e-01
5.19005835e-01 -1.12289464e+00 -3.77764195e-01 6.24542892e-01
4.06009585e-01 -1.23747659e+00 -3.80591512e-01 -5.72524071e-01
-5.02935886e-01 8.41636300e-01 6.56328797e-01 -2.95062084e-02
4.64695424e-01 1.81491598e-01 -1.54328227e-01 -1.43745437e-01
-7.52034664e-01 -5.25119781e-01 4.21175480e-01 5.05373955e-01
7.93183148e-01 -9.33754165e-03 -4.91501689e-01 9.61819053e-01
-4.91118506e-02 4.12091315e-02 4.32876974e-01 6.19813263e-01
-1.66780397e-01 -1.44479382e+00 2.06456602e-01 2.87248731e-01
-7.63013840e-01 -2.85925299e-01 -2.55478054e-01 7.44510531e-01
1.13290198e-01 5.52224100e-01 1.48026869e-01 -9.00777951e-02
4.46807295e-01 1.06225401e-01 6.88976347e-01 -8.94136846e-01
-4.75622594e-01 -4.35263723e-01 7.98371315e-01 -3.27121317e-01
-4.63857681e-01 -6.26146793e-01 -1.26704919e+00 -4.46592607e-02
-6.72852337e-01 5.81951439e-01 4.29347575e-01 9.44609821e-01
5.24701595e-01 4.41144168e-01 3.49824131e-01 -5.73818088e-01
-4.08579797e-01 -9.68730807e-01 -4.54643637e-01 9.47550833e-01
3.21684033e-01 -1.13288569e+00 -3.23495239e-01 1.43870696e-01] | [9.507608413696289, 8.883102416992188] |
2e4d0590-b34d-4303-b777-d06efe920dae | clustered-object-detection-in-aerial-images | 1904.08008 | null | https://arxiv.org/abs/1904.08008v3 | https://arxiv.org/pdf/1904.08008v3.pdf | Clustered Object Detection in Aerial Images | Detecting objects in aerial images is challenging for at least two reasons: (1) target objects like pedestrians are very small in pixels, making them hardly distinguished from surrounding background; and (2) targets are in general sparsely and non-uniformly distributed, making the detection very inefficient. In this paper, we address both issues inspired by observing that these targets are often clustered. In particular, we propose a Clustered Detection (ClusDet) network that unifies object clustering and detection in an end-to-end framework. The key components in ClusDet include a cluster proposal sub-network (CPNet), a scale estimation sub-network (ScaleNet), and a dedicated detection network (DetecNet). Given an input image, CPNet produces object cluster regions and ScaleNet estimates object scales for these regions. Then, each scale-normalized cluster region is fed into DetecNet for object detection. ClusDet has several advantages over previous solutions: (1) it greatly reduces the number of chips for final object detection and hence achieves high running time efficiency, (2) the cluster-based scale estimation is more accurate than previously used single-object based ones, hence effectively improves the detection for small objects, and (3) the final DetecNet is dedicated for clustered regions and implicitly models the prior context information so as to boost detection accuracy. The proposed method is tested on three popular aerial image datasets including VisDrone, UAVDT and DOTA. In all experiments, ClusDet achieves promising performance in comparison with state-of-the-art detectors. Code will be available in \url{https://github.com/fyangneil}. | ['Heng Fan', 'Haibin Ling', 'Fan Yang', 'Peng Chu', 'Erik Blasch'] | 2019-04-16 | clustered-object-detection-in-aerial-images-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Yang_Clustered_Object_Detection_in_Aerial_Images_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Yang_Clustered_Object_Detection_in_Aerial_Images_ICCV_2019_paper.pdf | iccv-2019-10 | ['object-detection-in-aerial-images'] | ['computer-vision'] | [-3.24830301e-02 -3.10818285e-01 2.47658789e-02 -1.04701832e-01
-3.60943526e-01 -6.65352881e-01 2.82759815e-01 2.20075861e-01
-5.47588348e-01 2.99839497e-01 -4.04856801e-01 2.68345326e-02
1.50022343e-01 -9.28370953e-01 -4.51540500e-01 -8.47903609e-01
-1.75496235e-01 2.84711689e-01 1.12450242e+00 1.90792143e-01
-1.17446631e-01 7.22270131e-01 -1.49830639e+00 -1.27172410e-01
8.44515502e-01 1.21482813e+00 5.38095832e-01 7.37626553e-01
2.37924159e-01 6.22400105e-01 -5.64417183e-01 -1.95038378e-01
5.22517800e-01 -3.38417262e-01 -2.20067158e-01 3.11750621e-01
4.48967606e-01 -4.44319248e-01 -2.10392192e-01 1.47688210e+00
4.47972357e-01 1.03917070e-01 5.86175203e-01 -1.25251114e+00
4.04375046e-03 4.93158877e-01 -1.21639037e+00 3.70031565e-01
-3.84439886e-01 2.24681705e-01 7.88540483e-01 -7.20353484e-01
2.39814520e-01 1.27796876e+00 6.73760235e-01 2.69599617e-01
-1.14734030e+00 -9.33380485e-01 2.88572609e-01 -7.14863017e-02
-1.79668999e+00 -1.81591019e-01 4.04771507e-01 -4.81000990e-01
3.90942305e-01 3.73152494e-01 7.60606885e-01 5.10493994e-01
-3.96734998e-02 7.67426789e-01 9.21839118e-01 -2.04530329e-01
4.51358765e-01 1.47262990e-01 -1.46404933e-02 6.49672329e-01
6.28819227e-01 2.13535309e-01 9.78332758e-02 -7.68044963e-02
9.74016428e-01 1.57960966e-01 -1.95313394e-01 -5.88118494e-01
-1.07394218e+00 7.41099417e-01 9.76964474e-01 2.12193087e-01
-4.32111293e-01 2.58471876e-01 3.49657863e-01 -1.49023399e-01
3.69102389e-01 7.15399012e-02 -2.66876578e-01 4.53330785e-01
-1.05885541e+00 1.84242532e-01 4.20230806e-01 8.99711311e-01
7.19544411e-01 -8.06948915e-02 1.22751412e-03 7.68070042e-01
3.87575030e-01 7.96115994e-01 1.52109161e-01 -6.31279290e-01
2.77196646e-01 7.32783735e-01 3.26071680e-01 -1.19583356e+00
-6.42140985e-01 -6.27117872e-01 -1.00275564e+00 4.72835690e-01
4.42530572e-01 -3.22615504e-01 -7.98502445e-01 1.53610742e+00
6.46171987e-01 1.27305523e-01 -1.48816586e-01 1.19743896e+00
9.19340968e-01 6.99408472e-01 1.84392557e-01 -3.40358056e-02
1.60940039e+00 -8.66343617e-01 -2.26820067e-01 -5.63769460e-01
2.25969866e-01 -8.37172389e-01 4.79602933e-01 2.50729471e-01
-8.17560196e-01 -8.65050733e-01 -1.01136112e+00 4.23685044e-01
-2.81101346e-01 9.94258761e-01 4.99790490e-01 6.05881512e-01
-9.07912970e-01 1.12453274e-01 -9.96949136e-01 -8.40326130e-01
4.81132299e-01 4.40992296e-01 1.41131103e-01 4.43722010e-02
-4.79735494e-01 4.63197052e-01 8.26457322e-01 4.77359183e-02
-8.98095191e-01 -1.79304093e-01 -6.84277177e-01 2.26844158e-02
6.86848700e-01 -4.95396793e-01 9.93183494e-01 -9.48732436e-01
-9.70752299e-01 7.90709734e-01 2.16039401e-02 -5.07249713e-01
4.93893892e-01 2.23023500e-02 -3.26415360e-01 2.54282147e-01
3.42858344e-01 9.15603042e-01 9.17664289e-01 -1.03606594e+00
-1.25055361e+00 -4.12297308e-01 4.34753969e-02 1.11627601e-01
-3.61064553e-01 1.91303805e-01 -8.54558408e-01 -6.87620640e-01
1.30098000e-01 -8.66638720e-01 -2.64724612e-01 2.97077239e-01
-4.84296918e-01 -1.35485977e-01 1.17698944e+00 -3.44791204e-01
1.10639369e+00 -2.30418563e+00 -2.25674473e-02 8.05369616e-02
3.79116923e-01 5.40707350e-01 -1.52843431e-01 1.48503765e-01
1.10791974e-01 -3.79503965e-01 -9.33991671e-02 -7.76677951e-02
-3.07729572e-01 -2.21471414e-01 -1.92590971e-02 7.19485164e-01
1.88403398e-01 5.06567597e-01 -7.72170305e-01 -6.88228607e-01
5.65515280e-01 2.44701624e-01 -2.65345216e-01 1.32374808e-01
-8.30958858e-02 1.96633831e-01 -5.45287967e-01 1.06702650e+00
1.12840068e+00 -1.62164673e-01 7.62652904e-02 -4.58898455e-01
-6.77177370e-01 -4.46085066e-01 -1.63991940e+00 9.16346967e-01
3.54939736e-02 5.61995208e-01 5.10472894e-01 -8.89868677e-01
1.03000343e+00 -6.10174648e-02 2.45484918e-01 -1.19948603e-01
4.26539272e-01 6.77158535e-02 3.84623967e-02 -1.53801456e-01
4.32840496e-01 2.07725301e-01 -8.73295870e-03 -1.68344080e-02
-1.79870054e-02 -1.43576443e-01 6.57048285e-01 2.21912369e-01
9.48769212e-01 -2.01078355e-01 5.80548465e-01 -3.80039155e-01
5.34575999e-01 1.40021592e-01 8.99792612e-01 7.47252464e-01
-5.49531996e-01 3.09532613e-01 2.43921563e-01 -3.83742809e-01
-8.20985138e-01 -1.06753361e+00 -1.53331012e-01 9.44887459e-01
6.76775396e-01 -2.62957752e-01 -7.85245419e-01 -5.30575037e-01
3.37957256e-02 2.32976675e-01 -4.82338667e-01 1.77954122e-01
-2.41977096e-01 -8.74632120e-01 4.69103634e-01 7.06174731e-01
8.13282728e-01 -7.93110549e-01 -9.21654403e-01 1.92998648e-01
-6.61073029e-02 -1.21389735e+00 -4.79023755e-01 3.39060366e-01
-9.05169308e-01 -1.25306189e+00 -6.73536956e-01 -7.87120283e-01
6.90898180e-01 9.76830006e-01 7.99361765e-01 1.15091182e-01
-6.41707480e-01 6.39390200e-02 -3.69123399e-01 -6.45056605e-01
1.59107596e-02 -6.53808191e-02 9.93747637e-02 1.18910089e-01
5.35697460e-01 -2.51514465e-01 -7.60945797e-01 6.79194927e-01
-8.13787997e-01 -6.99827373e-02 9.20043707e-01 5.76872647e-01
5.97963810e-01 5.83172023e-01 2.10584447e-01 -4.59140539e-01
-7.03478605e-02 -1.70414954e-01 -1.42411602e+00 7.31187686e-02
-5.39838783e-02 -6.99821115e-01 5.64536035e-01 -4.44125712e-01
-8.36959600e-01 4.99961317e-01 2.43700951e-01 -2.88835347e-01
-4.63191956e-01 -2.81586424e-02 -2.28261560e-01 -2.30290398e-01
7.58050561e-01 7.80640021e-02 -2.25736424e-01 -2.69256979e-01
8.47851038e-02 6.04200125e-01 5.75847805e-01 -9.29154158e-02
1.22079480e+00 6.78386033e-01 -6.03700392e-02 -1.06704533e+00
-6.60227358e-01 -9.14298773e-01 -8.91226411e-01 -4.67402995e-01
9.38504755e-01 -1.33332741e+00 -6.76473856e-01 4.86708164e-01
-9.99383390e-01 -3.84788036e-01 -1.92959324e-01 6.50069118e-01
-8.42807889e-02 3.21962714e-01 -5.42269945e-01 -1.05038607e+00
-2.33075082e-01 -9.78735685e-01 1.04224133e+00 8.73765707e-01
1.83633596e-01 -6.65927708e-01 -4.25739050e-01 1.59485992e-02
3.16176206e-01 3.83945048e-01 3.64958584e-01 -2.81751454e-01
-8.18763793e-01 -4.05810356e-01 -8.57309878e-01 3.05894494e-01
8.78101885e-02 2.82972068e-01 -7.71591783e-01 -5.61520040e-01
-3.60594332e-01 -2.32405260e-01 9.57600892e-01 8.35860431e-01
7.84866333e-01 3.03050220e-01 -7.54191339e-01 4.49144512e-01
1.62042224e+00 2.56213039e-01 2.22254217e-01 1.22365318e-01
4.67361152e-01 3.68335664e-01 1.01955628e+00 5.90930223e-01
5.71497157e-02 6.83715045e-01 7.73328543e-01 -5.77447534e-01
-2.67641842e-01 1.80202127e-01 3.83632779e-01 1.99361563e-01
-1.84259251e-01 -2.61896044e-01 -6.14767492e-01 5.98397970e-01
-1.97503150e+00 -8.91036749e-01 -4.57990766e-01 2.07419777e+00
3.10393572e-01 1.08232133e-01 5.66312551e-01 -1.93028212e-01
1.14027953e+00 1.08829275e-01 -7.29727507e-01 3.64280313e-01
-1.20244257e-01 -1.31968915e-01 9.36277866e-01 2.31098626e-02
-1.68803644e+00 1.13207507e+00 4.83361769e+00 8.63886416e-01
-9.48419750e-01 1.75552696e-01 3.44525754e-01 -1.43432617e-03
9.51985478e-01 -5.98969422e-02 -1.14360809e+00 3.77247959e-01
1.33138359e-01 2.86042336e-02 1.72041953e-01 1.23786867e+00
2.16413170e-01 -3.79396409e-01 -6.03056610e-01 8.26227486e-01
-1.45232797e-01 -8.03679168e-01 -3.20021123e-01 1.22451643e-03
5.72531521e-01 2.17226416e-01 -1.99580237e-01 1.03355065e-01
5.83748281e-01 -3.56783628e-01 9.26335454e-01 1.57875996e-02
5.88787317e-01 -9.15375710e-01 9.48628724e-01 4.91889298e-01
-1.83284688e+00 -3.91131848e-01 -1.12226498e+00 1.52915657e-01
-1.29847357e-03 8.02157581e-01 -5.68785667e-01 4.01793092e-01
1.01764035e+00 6.27915919e-01 -7.68557966e-01 1.59355175e+00
-2.80718923e-01 6.07454121e-01 -5.68302631e-01 -7.47857615e-02
4.20477837e-01 -2.52625406e-01 6.78450406e-01 1.43339586e+00
3.83265167e-01 2.51099080e-01 6.64082944e-01 9.41365182e-01
9.12273452e-02 -3.77071053e-02 -2.92894989e-01 1.65295690e-01
7.30252385e-01 1.80485058e+00 -1.49407434e+00 -3.87376308e-01
-3.71025711e-01 8.71606469e-01 9.43522081e-02 1.68785274e-01
-8.89055550e-01 -5.30704796e-01 6.09487236e-01 3.65363173e-02
8.55660439e-01 -1.83141276e-01 2.01058552e-01 -9.87032890e-01
-1.78818062e-01 -5.86725175e-01 4.35436010e-01 -5.36597252e-01
-1.07859147e+00 7.14819014e-01 5.21463156e-03 -1.44474709e+00
3.08012694e-01 -6.61159456e-01 -7.97887981e-01 5.58693051e-01
-1.25318468e+00 -1.28723097e+00 -9.28799629e-01 4.70536530e-01
6.25012577e-01 -1.30729124e-01 5.35234213e-01 4.03117746e-01
-8.28183353e-01 2.68528908e-01 -2.95986775e-02 4.21315342e-01
6.19310439e-01 -1.12238264e+00 2.52813607e-01 1.35452092e+00
-5.68618551e-02 2.71207541e-01 5.29250503e-01 -6.94917500e-01
-1.13887930e+00 -1.58218575e+00 1.56091005e-01 -9.77262408e-02
5.47125340e-01 -3.95123035e-01 -5.18501878e-01 2.85345823e-01
-1.05649158e-01 2.81224936e-01 3.47118437e-01 -2.65479952e-01
-4.59548309e-02 -2.70061523e-01 -1.03944230e+00 2.57699639e-01
7.24451721e-01 1.63185611e-01 -1.26098106e-02 5.33746183e-01
3.84704977e-01 -3.39718759e-01 -4.08174247e-01 3.30466837e-01
3.03555280e-01 -1.15370810e+00 9.67076242e-01 1.47797376e-01
-9.36504602e-02 -1.04480648e+00 -1.79103836e-01 -1.09960353e+00
-7.15549409e-01 -8.36720914e-02 1.53642669e-01 1.36043596e+00
8.83003622e-02 -4.97202635e-01 7.13956058e-01 -3.74656431e-02
1.13251247e-02 -3.71554643e-01 -7.17231154e-01 -1.06082022e+00
-4.67462242e-01 -7.42013529e-02 3.93677145e-01 5.76822639e-01
-6.31536186e-01 2.78381139e-01 -2.07891747e-01 7.24926710e-01
1.07629657e+00 2.00048879e-01 9.92579579e-01 -1.42375934e+00
-2.73791581e-01 -3.51756662e-01 -6.50086403e-01 -1.14699519e+00
-4.42446291e-01 -4.86288190e-01 2.73438960e-01 -1.42549253e+00
3.23938817e-01 -6.27103388e-01 -2.73574535e-02 3.93891990e-01
-1.87813118e-01 6.64379478e-01 3.21960628e-01 3.45508516e-01
-7.98299253e-01 1.87952831e-01 7.39086688e-01 -8.38115737e-02
-2.97607213e-01 3.10203701e-01 -4.67281133e-01 9.80895638e-01
8.72051835e-01 -4.61866915e-01 -8.41273516e-02 -1.64601147e-01
-4.89184201e-01 -1.38746515e-01 7.34898329e-01 -1.60094571e+00
3.68905216e-01 3.31818708e-03 7.37122893e-01 -1.10511518e+00
3.11345547e-01 -9.23566759e-01 -1.32687360e-01 7.35810816e-01
2.73461193e-01 -2.27869272e-01 3.18820357e-01 6.14025652e-01
-1.20153792e-01 -2.42949501e-01 1.36459589e+00 -7.18013272e-02
-8.49699795e-01 2.41453961e-01 -5.48347175e-01 -3.54207128e-01
1.39154172e+00 -2.34433934e-01 -1.78269237e-01 -3.33627197e-03
-4.92645293e-01 3.12805384e-01 6.02034926e-01 1.79696783e-01
3.73526365e-01 -1.11375546e+00 -6.71895325e-01 9.71427411e-02
2.05019459e-01 2.11323544e-01 2.06222981e-01 1.12204552e+00
-5.94670951e-01 2.95075178e-01 -8.49946588e-02 -8.83541346e-01
-1.69036293e+00 5.74338675e-01 2.04165474e-01 -9.65272170e-03
-5.18513143e-01 1.01568365e+00 5.58920503e-01 -2.37040401e-01
3.30013156e-01 -4.45639104e-01 -1.11683607e-01 7.18636066e-02
7.66887009e-01 4.55682665e-01 -2.86453456e-01 -6.82598114e-01
-4.15387422e-01 7.31455564e-01 -1.06964335e-01 3.17854315e-01
1.18471277e+00 -4.06174734e-02 -1.81968406e-01 -4.32798974e-02
6.93886638e-01 -2.17755482e-01 -1.34565854e+00 -3.55348438e-01
-2.18653873e-01 -5.77093542e-01 2.58025050e-01 -5.38901627e-01
-1.24955368e+00 6.42983854e-01 9.95996833e-01 2.87052393e-01
1.38860261e+00 1.69455528e-01 5.54656088e-01 2.50964195e-01
4.00179148e-01 -1.08577752e+00 1.55530676e-01 2.23586500e-01
4.75094527e-01 -1.14824283e+00 3.47695678e-01 -7.71679997e-01
-4.52500463e-01 1.14748549e+00 8.77190828e-01 -2.39153370e-01
4.45410579e-01 3.89403939e-01 -4.14886251e-02 -2.37323180e-01
-2.55045116e-01 -6.49851382e-01 1.38129637e-01 6.30974591e-01
6.06324673e-02 2.83788711e-01 -7.65307024e-02 3.76952857e-01
1.85868472e-01 -1.83748096e-01 2.61570007e-01 8.47925901e-01
-9.75486338e-01 -6.71113729e-01 -8.49203706e-01 3.43504876e-01
-1.87922895e-01 1.18821956e-01 -5.18363357e-01 9.94313180e-01
5.43618619e-01 1.08140981e+00 1.41244873e-01 -3.44513357e-01
2.65886754e-01 -6.31211221e-01 1.65683597e-01 -6.07613802e-01
-3.69295567e-01 4.11851943e-01 -2.27132201e-01 -5.64570546e-01
-4.47639763e-01 -6.61035836e-01 -1.08630586e+00 -6.71721324e-02
-9.28319931e-01 1.60408206e-02 6.35687828e-01 4.10764545e-01
1.40403241e-01 5.52303791e-01 5.22535980e-01 -1.07198548e+00
-3.56972456e-01 -1.01825452e+00 -8.84926975e-01 -1.82666972e-01
-6.72131032e-02 -7.75952637e-01 -2.47623384e-01 9.29597169e-02] | [8.682943344116211, -0.7771471738815308] |
47aaa4c5-8bba-4a99-9334-a1a88578a7f3 | pruning-pre-trained-language-models-with | 2305.12394 | null | https://arxiv.org/abs/2305.12394v1 | https://arxiv.org/pdf/2305.12394v1.pdf | Pruning Pre-trained Language Models with Principled Importance and Self-regularization | Iterative pruning is one of the most effective compression methods for pre-trained language models. We discovered that finding the optimal pruning decision is an equality-constrained 0-1 Integer Linear Programming problem. The solution to this optimization problem leads to a principled importance criterion which we use to rank parameters during iterative model pruning. To mitigate the poor generalization at high sparsity levels, we propose a self-regularization scheme where model prediction is regularized by the latest checkpoint with increasing sparsity throughout pruning. Our experiments on natural language understanding, question-answering, named entity recognition, and data-to-text generation with various Transformer-based PLMs show the effectiveness of the approach at various sparsity levels. | ['Kenny Q. Zhu', 'Siyu Ren'] | 2023-05-21 | null | null | null | null | ['data-to-text-generation'] | ['natural-language-processing'] | [ 4.34971690e-01 3.20014834e-01 -8.78338277e-01 -4.46855485e-01
-8.58606160e-01 -1.16278157e-01 2.08807111e-01 2.88314253e-01
-3.17748070e-01 6.39618039e-01 2.51261175e-01 -7.30153978e-01
-3.18013996e-01 -6.88736022e-01 -6.76899433e-01 -4.11342867e-02
-4.98188548e-02 8.41175258e-01 1.63760394e-01 -1.89600974e-01
6.00479007e-01 3.98820311e-01 -1.31991863e+00 7.07195103e-01
1.07846189e+00 9.48813975e-01 4.54099208e-01 5.57519078e-01
-6.73856497e-01 9.90924895e-01 -5.16905487e-01 -3.94874573e-01
3.79058957e-01 -2.44418904e-01 -1.21023202e+00 6.32999018e-02
2.68882304e-01 5.20041175e-02 -5.32954410e-02 8.76997828e-01
1.88903138e-02 1.42840877e-01 3.61695200e-01 -1.07937872e+00
-1.97510511e-01 8.60301435e-01 -4.73249018e-01 2.80874848e-01
2.50967324e-01 -3.73266488e-01 1.29051912e+00 -7.37044513e-01
6.21249199e-01 1.27181625e+00 4.67862874e-01 5.65073371e-01
-1.32632554e+00 -4.11977947e-01 2.63400406e-01 1.99310362e-01
-1.34565139e+00 -5.31060040e-01 6.10858560e-01 -1.27013743e-01
1.28566992e+00 3.20494533e-01 5.28030515e-01 6.13166869e-01
1.45972535e-01 7.99295306e-01 7.09883928e-01 -7.47978628e-01
2.63994992e-01 3.65185857e-01 4.74092484e-01 1.00001645e+00
2.93237567e-01 -2.19622299e-01 -7.49673784e-01 -8.67054701e-01
3.74395102e-01 -1.79770470e-01 -1.00996513e-02 -1.38856143e-01
-4.79836136e-01 1.06237805e+00 -1.91199839e-01 9.62391570e-02
-2.18442678e-01 8.59462172e-02 3.67303997e-01 5.32332182e-01
3.95217389e-01 7.40734398e-01 -9.57331061e-01 -2.17534170e-01
-1.26114333e+00 1.37558728e-01 1.22258854e+00 9.25652862e-01
5.57740211e-01 7.76684061e-02 -1.39861733e-01 1.17684150e+00
3.59602153e-01 1.22222289e-01 6.73413336e-01 -9.54203367e-01
7.87883222e-01 7.99117625e-01 -1.06522955e-01 -7.53598630e-01
-1.22203715e-01 -5.54000318e-01 -6.89422011e-01 -3.62869322e-01
-7.82807171e-02 -3.19270939e-02 -8.88767302e-01 1.72786462e+00
1.73603952e-01 7.49107869e-03 5.81991971e-02 2.28127331e-01
2.34863445e-01 7.48269081e-01 2.33003184e-01 -7.11266100e-01
1.09584987e+00 -1.06290805e+00 -6.09543979e-01 -3.64593834e-01
9.58952725e-01 -6.32878125e-01 9.81518388e-01 5.30777156e-01
-1.10806334e+00 -2.37141237e-01 -7.55147636e-01 5.85181685e-03
5.93430251e-02 1.22992814e-01 7.13116169e-01 5.57273448e-01
-8.15819979e-01 6.36373997e-01 -8.68541539e-01 -2.50299685e-02
2.96533614e-01 5.37056148e-01 -1.71432942e-01 -3.43253613e-02
-8.46430779e-01 9.28760409e-01 7.12196827e-01 -3.88248682e-01
-5.73662519e-01 -7.94934571e-01 -5.91509581e-01 2.45364785e-01
4.53592092e-01 -6.00013554e-01 1.27751303e+00 -6.50384724e-01
-1.19763541e+00 7.57325888e-01 -8.21419179e-01 -1.13336527e+00
-9.28516313e-02 -3.57788622e-01 -3.27404708e-01 5.43705747e-02
-9.00112763e-02 6.44054651e-01 9.43346441e-01 -1.11516893e+00
-6.73033059e-01 -3.40807773e-02 5.24024898e-03 1.37226969e-01
-6.02302849e-01 1.89124286e-01 -4.86862749e-01 -5.36935210e-01
3.94856632e-01 -6.67584956e-01 -5.89364886e-01 -5.01948297e-01
-3.49021137e-01 -4.20609981e-01 7.84384847e-01 -9.57118690e-01
1.80266654e+00 -1.64722812e+00 5.27586564e-02 4.28009808e-01
1.40728444e-01 1.49442464e-01 -1.75404966e-01 4.76792932e-01
-3.49458284e-03 4.51164603e-01 -1.70447439e-01 -2.74056435e-01
-3.26969564e-01 3.84233028e-01 -6.97801113e-01 -2.06690729e-01
8.42024013e-02 5.50306916e-01 -6.59089804e-01 -9.28065479e-01
-1.49682090e-01 7.50909140e-03 -8.87579978e-01 3.30118597e-01
-6.46910965e-01 -1.36728153e-01 -5.24061799e-01 6.36043489e-01
3.59902531e-01 -4.46190089e-01 3.40708435e-01 -8.84636790e-02
1.26995832e-01 7.23155975e-01 -1.07317352e+00 1.46244383e+00
-4.98119444e-01 4.11049515e-01 6.02496192e-02 -1.20158958e+00
1.05493832e+00 1.61409765e-01 5.82326710e-01 -4.62376833e-01
-2.18009725e-01 6.94211870e-02 -8.92694965e-02 -4.08180296e-01
6.15301609e-01 -1.17105611e-01 8.30744803e-02 5.30404627e-01
1.21421285e-01 -3.26029927e-01 5.71229696e-01 3.86813670e-01
1.30005515e+00 -3.77689838e-01 3.98712933e-01 -3.00816119e-01
4.89985377e-01 2.57323563e-01 7.94052720e-01 9.64902639e-01
3.16879481e-01 3.89938176e-01 6.58889294e-01 -5.59583664e-01
-9.52066660e-01 -5.69744349e-01 -1.41590327e-01 1.30965626e+00
-3.13538522e-01 -1.00597441e+00 -6.42427742e-01 -7.22605824e-01
-1.22338332e-01 1.10679114e+00 -7.76980072e-02 -1.94302991e-01
-8.30621004e-01 -6.70735657e-01 4.86598492e-01 2.74568826e-01
2.34518453e-01 -9.80519533e-01 -4.57942784e-01 4.90664393e-01
-4.24245179e-01 -1.02835059e+00 -4.16956544e-01 7.07588494e-01
-1.57821941e+00 -9.25007343e-01 -1.47757113e-01 -9.55982745e-01
9.05476272e-01 3.34823132e-02 1.37142634e+00 4.14727032e-01
-7.99959823e-02 2.57080346e-01 -2.65712112e-01 -3.82348090e-01
-6.86424792e-01 5.01325846e-01 7.32559115e-02 -4.20549899e-01
2.52814293e-01 -4.96876508e-01 -6.68516457e-02 1.39855236e-01
-8.86568546e-01 1.94522381e-01 5.27485967e-01 1.06846881e+00
9.14251029e-01 1.68817475e-01 3.22152346e-01 -1.13181424e+00
1.08097744e+00 -3.69273245e-01 -6.30228519e-01 7.90613890e-01
-9.00512695e-01 6.55829310e-01 8.26165974e-01 -5.03843069e-01
-1.00551367e+00 2.21040770e-01 -1.57718763e-01 -2.88575500e-01
8.43960717e-02 7.35213220e-01 2.25996841e-02 1.71495695e-02
6.39427185e-01 3.27679992e-01 -4.55022335e-01 -6.53574884e-01
1.85487568e-01 4.06049490e-01 2.82057375e-01 -8.73539567e-01
4.91294086e-01 5.98497540e-02 -4.27603424e-02 -9.59537685e-01
-1.04577959e+00 -5.00439048e-01 -2.77392089e-01 2.76776165e-01
3.40807378e-01 -7.41767347e-01 -3.24446023e-01 -8.73367488e-02
-1.32471991e+00 -2.71448851e-01 -5.80445468e-01 1.52348235e-01
-4.35718834e-01 5.93555510e-01 -7.02871084e-01 -9.27021682e-01
-9.05056894e-01 -6.78711891e-01 7.78800786e-01 -7.91761354e-02
-4.24505860e-01 -7.31449068e-01 2.14884877e-01 5.70336163e-01
2.48225302e-01 -3.70802224e-01 1.59411335e+00 -9.53268766e-01
-5.59615910e-01 -1.61222503e-01 -1.43204322e-02 5.30251265e-01
-2.89748758e-01 -2.01564148e-01 -3.63312542e-01 -1.84097677e-01
3.94099146e-01 -4.92319196e-01 9.52189624e-01 2.64739811e-01
1.51645398e+00 -7.03022540e-01 -4.52733248e-01 6.67135894e-01
1.41959929e+00 1.75582945e-01 3.41751158e-01 7.23514557e-02
5.01095116e-01 3.70228410e-01 6.55270875e-01 5.83188951e-01
1.34939075e-01 4.24308926e-01 2.07139432e-01 2.47094348e-01
3.05825830e-01 -5.16190708e-01 2.60126233e-01 1.16443288e+00
2.71568418e-01 -3.05609107e-01 -1.02974391e+00 4.00519758e-01
-1.72296095e+00 -8.67500484e-01 1.54535219e-01 2.28555846e+00
1.12227905e+00 4.90639448e-01 -3.80108356e-01 8.51080418e-02
3.50288600e-01 4.25309800e-02 -4.19973254e-01 -7.67850220e-01
-6.30676076e-02 5.47848403e-01 5.26795149e-01 5.69172919e-01
-8.58020365e-01 1.15894651e+00 7.44991970e+00 1.11299825e+00
-8.42562735e-01 4.92626652e-02 7.73811042e-01 -1.44219533e-01
-3.37635577e-01 3.53367060e-01 -1.32457936e+00 2.92182922e-01
1.18898177e+00 -4.74483848e-01 4.51181561e-01 1.07694924e+00
1.79625958e-01 -1.00377217e-01 -9.75568771e-01 7.63592124e-01
-2.31980141e-02 -1.58334792e+00 4.10360336e-01 -1.49905592e-01
8.40223312e-01 1.10160269e-01 -3.07867914e-01 3.49370748e-01
4.10193831e-01 -9.91672516e-01 3.44072789e-01 4.79108781e-01
3.32663000e-01 -6.25315845e-01 2.34410197e-01 8.09369326e-01
-1.14769459e+00 -3.76427770e-01 -7.18726814e-01 1.46435261e-01
3.53942364e-01 7.51712739e-01 -1.05083466e+00 8.47482122e-03
3.77742797e-01 2.46462181e-01 -3.45640242e-01 1.08357966e+00
-1.20385356e-01 9.99002516e-01 -8.10871124e-01 1.81762248e-01
-4.67324816e-02 -1.17626481e-01 5.27170360e-01 1.33689022e+00
2.36356914e-01 2.50072300e-01 3.13211590e-01 6.71545804e-01
-2.31313810e-01 3.31043094e-01 -4.71295446e-01 -2.19552800e-01
5.24291098e-01 7.93441236e-01 -5.64719558e-01 -3.78715247e-01
-1.77079499e-01 6.55912220e-01 6.25396073e-01 1.38814017e-01
-4.32081521e-01 -2.47753155e-03 3.05803746e-01 2.98406184e-01
3.29630613e-01 -3.60041797e-01 -5.67504525e-01 -1.28152370e+00
7.81473294e-02 -1.17429292e+00 6.25618339e-01 -2.74404258e-01
-7.87786305e-01 5.47502160e-01 2.67315000e-01 -9.00626063e-01
-4.80119526e-01 -2.65757322e-01 -6.73700035e-01 6.38661921e-01
-1.27746129e+00 -6.55042529e-01 3.15892905e-01 3.54385406e-01
8.44980717e-01 -3.68464321e-01 1.02784646e+00 6.33143261e-02
-6.00094557e-01 7.31237173e-01 6.61764666e-02 -2.79627651e-01
2.02389121e-01 -8.95145595e-01 3.06655854e-01 8.70396316e-01
4.18206990e-01 8.20759356e-01 5.82431853e-01 -8.64068866e-01
-1.38087475e+00 -8.55594635e-01 1.46390665e+00 -4.30056453e-02
4.66505855e-01 -1.42873749e-01 -1.13000572e+00 4.77886736e-01
2.45032795e-02 -3.87254357e-01 8.16811025e-01 2.16364145e-01
-2.28459999e-01 -3.79364379e-02 -1.10550547e+00 4.85139608e-01
9.75470245e-01 -3.68671894e-01 -7.44037569e-01 7.38313675e-01
1.08558333e+00 -4.04271007e-01 -8.29937994e-01 3.86398822e-01
2.26382136e-01 -4.88880754e-01 8.78445685e-01 -1.04546404e+00
5.26206195e-01 2.18241096e-01 -1.57398582e-01 -8.32804441e-01
-1.12454452e-01 -7.87524939e-01 -8.12662661e-01 1.03160059e+00
7.34227300e-01 -2.18769073e-01 1.19847023e+00 8.42098355e-01
1.54927149e-01 -1.17686093e+00 -9.41983879e-01 -6.64101362e-01
-2.38766834e-01 -6.21974349e-01 4.22664434e-01 7.55106807e-01
7.61345327e-02 4.64102566e-01 -5.64807832e-01 -1.19832411e-01
5.75930774e-01 1.68588236e-01 5.72358251e-01 -1.19496250e+00
-5.67059398e-01 -2.25756451e-01 1.07031614e-01 -1.65784073e+00
3.30253214e-01 -1.07536566e+00 -1.06781274e-01 -1.43911040e+00
1.73399910e-01 -6.24743581e-01 -2.46245444e-01 6.68871164e-01
1.14977425e-02 -6.00431383e-01 1.83869570e-01 3.41433167e-01
-6.08546495e-01 4.14988488e-01 7.40211427e-01 -1.53826550e-01
-5.41047633e-01 2.71590263e-01 -6.93259001e-01 6.73456311e-01
8.43821824e-01 -1.09664512e+00 -4.85483557e-01 -5.83051622e-01
4.48298782e-01 3.69071782e-01 -2.07975477e-01 -8.68032336e-01
4.49389696e-01 -4.23672348e-01 -9.73879844e-02 -9.34210658e-01
3.65982622e-01 -8.04936230e-01 -1.88963711e-02 6.75716758e-01
-8.42236876e-01 1.53678566e-01 9.84572917e-02 4.09653664e-01
-4.12226349e-01 -8.04779530e-01 7.13885248e-01 -2.02279598e-01
-5.37402272e-01 1.95414752e-01 -2.22866163e-01 4.46569443e-01
4.44601297e-01 -4.29344885e-02 9.90419462e-02 -3.07014465e-01
-4.82943296e-01 3.83536547e-01 1.39875310e-02 2.77334958e-01
7.52222776e-01 -8.82954061e-01 -7.27174044e-01 2.50138432e-01
-2.68960834e-01 -6.71302378e-02 -1.97163910e-01 4.01854396e-01
-3.61656129e-01 6.64046407e-01 2.03358978e-01 -4.15429831e-01
-1.44323635e+00 1.99902996e-01 1.22814484e-01 -1.21999967e+00
-2.92981267e-01 9.78677094e-01 -3.55745345e-01 -3.71486753e-01
4.67695802e-01 -3.24162781e-01 -1.61390558e-01 -2.68156469e-01
3.49487305e-01 4.40652400e-01 2.22716555e-01 -1.52561530e-01
-5.00530191e-02 1.45312414e-01 -5.22752345e-01 -3.79037857e-02
1.34276116e+00 2.05037519e-01 -3.68289113e-01 9.24474839e-03
1.15868628e+00 -1.33828819e-01 -6.44282460e-01 -3.66045117e-01
5.82685113e-01 -3.15800339e-01 9.02418569e-02 -4.74604636e-01
-9.48424637e-01 6.91641986e-01 1.65623948e-01 2.41983101e-01
1.13350725e+00 -2.59569108e-01 9.90469396e-01 1.07953095e+00
4.33499068e-01 -1.23380446e+00 -1.29156768e-01 1.13383520e+00
7.62798667e-01 -9.77154672e-01 3.46238136e-01 -4.16147768e-01
-4.79641318e-01 1.11647463e+00 6.70982540e-01 1.60600305e-01
6.87821150e-01 4.86906171e-01 -3.90219897e-01 1.15306422e-01
-1.39130855e+00 1.42235696e-01 3.14144373e-01 2.80384690e-01
4.33357060e-01 -1.37791902e-01 -7.07980573e-01 6.63830578e-01
-4.14700806e-01 7.41887689e-02 1.06868343e-02 1.05268431e+00
-7.27222860e-01 -1.52648818e+00 -2.23160073e-01 1.09434056e+00
-6.52457595e-01 -4.75443810e-01 -3.22348267e-01 3.44409347e-01
-6.56348988e-02 9.62285399e-01 -1.88776553e-01 -3.53327155e-01
2.24517763e-01 5.26853323e-01 2.37212181e-01 -8.83538127e-01
-8.60090613e-01 -1.25957578e-01 2.44435385e-01 -5.38257718e-01
-3.51141281e-02 -3.11523974e-01 -1.27072608e+00 -6.05445653e-02
-6.96660638e-01 5.39834142e-01 5.33146977e-01 8.21423590e-01
5.92327118e-01 5.63826337e-02 6.44601643e-01 -1.21613689e-01
-9.52665329e-01 -7.89596558e-01 -1.71711326e-01 1.73696369e-01
-1.21612884e-01 -1.46192864e-01 -3.76979202e-01 8.64378288e-02] | [8.735090255737305, 3.6623895168304443] |
b1e7af04-dbea-4c8e-979c-b1d4949ab547 | editing-implicit-assumptions-in-text-to-image | 2303.08084 | null | https://arxiv.org/abs/2303.08084v1 | https://arxiv.org/pdf/2303.08084v1.pdf | Editing Implicit Assumptions in Text-to-Image Diffusion Models | Text-to-image diffusion models often make implicit assumptions about the world when generating images. While some assumptions are useful (e.g., the sky is blue), they can also be outdated, incorrect, or reflective of social biases present in the training data. Thus, there is a need to control these assumptions without requiring explicit user input or costly re-training. In this work, we aim to edit a given implicit assumption in a pre-trained diffusion model. Our Text-to-Image Model Editing method, TIME for short, receives a pair of inputs: a "source" under-specified prompt for which the model makes an implicit assumption (e.g., "a pack of roses"), and a "destination" prompt that describes the same setting, but with a specified desired attribute (e.g., "a pack of blue roses"). TIME then updates the model's cross-attention layers, as these layers assign visual meaning to textual tokens. We edit the projection matrices in these layers such that the source prompt is projected close to the destination prompt. Our method is highly efficient, as it modifies a mere 2.2% of the model's parameters in under one second. To evaluate model editing approaches, we introduce TIMED (TIME Dataset), containing 147 source and destination prompt pairs from various domains. Our experiments (using Stable Diffusion) show that TIME is successful in model editing, generalizes well for related prompts unseen during editing, and imposes minimal effect on unrelated generations. | ['Yonatan Belinkov', 'Bahjat Kawar', 'Hadas Orgad'] | 2023-03-14 | null | null | null | null | ['model-editing'] | ['natural-language-processing'] | [ 5.65496624e-01 3.75506401e-01 2.66758204e-02 -6.49940014e-01
-3.29205960e-01 -7.37865746e-01 1.12241793e+00 3.09093148e-01
-4.96450871e-01 6.00577295e-01 1.45616889e-01 -2.95004457e-01
2.91134208e-01 -7.92563081e-01 -1.03119552e+00 -4.69663918e-01
4.47448939e-01 7.67791867e-01 2.03929003e-02 -2.70123422e-01
2.96774536e-01 3.25897753e-01 -1.24154043e+00 3.58163685e-01
8.35069478e-01 5.91862559e-01 5.04720509e-01 8.09569657e-01
-3.41243297e-01 9.53948796e-01 -1.15932488e+00 -7.85473526e-01
2.35791840e-02 -6.07656300e-01 -6.89309478e-01 1.78108945e-01
5.08612871e-01 -5.55569947e-01 -3.94099593e-01 9.91431296e-01
4.30678338e-01 8.28060806e-02 8.05247068e-01 -1.25192416e+00
-1.43945444e+00 8.30746591e-01 -7.64909685e-01 1.12663656e-01
2.00439230e-01 5.35653889e-01 6.33542359e-01 -7.03704834e-01
1.00586939e+00 1.15219831e+00 4.25962210e-01 8.48415315e-01
-1.45139110e+00 -6.31939471e-01 4.19846177e-01 -1.15141690e-01
-1.24806392e+00 -4.56352770e-01 6.71576023e-01 -5.94196022e-01
8.58323395e-01 3.90914589e-01 7.85081863e-01 1.66061890e+00
1.88666731e-01 4.75808710e-01 9.70629752e-01 -4.37283427e-01
2.72227079e-01 5.63754559e-01 -2.10550219e-01 5.08183062e-01
-2.52686907e-02 4.35531279e-03 -4.98090625e-01 -1.72588304e-01
6.58640206e-01 -7.22720176e-02 -2.23218128e-02 -1.19146265e-01
-1.18850994e+00 8.09988499e-01 3.35399270e-01 1.23942308e-01
-4.12459195e-01 3.66726488e-01 2.00095996e-02 2.84754187e-01
6.92705750e-01 5.80022216e-01 -2.05981329e-01 -2.04272151e-01
-8.79777431e-01 3.42868984e-01 5.78248739e-01 1.21919680e+00
7.51108229e-01 5.30132502e-02 -4.62119222e-01 9.24776495e-01
-2.22953483e-02 5.86022854e-01 4.05522764e-01 -1.05030751e+00
2.95065135e-01 3.57636005e-01 3.34326327e-01 -9.75809693e-01
1.67869683e-02 -2.13833809e-01 -8.41916800e-01 2.62439787e-01
3.25636715e-01 -4.19413656e-01 -1.27113712e+00 2.08545589e+00
2.14538112e-01 -1.24327421e-01 -6.59343749e-02 7.01931596e-01
5.61358333e-01 8.86855900e-01 2.50499010e-01 -2.44191475e-02
1.17324924e+00 -1.01141298e+00 -6.48687243e-01 -4.76467341e-01
5.15944958e-01 -9.04058397e-01 1.45420849e+00 9.05964449e-02
-1.20289242e+00 -5.06065309e-01 -7.02871442e-01 -1.79493845e-01
-5.07043123e-01 9.48479548e-02 4.63456273e-01 4.71898556e-01
-1.22130227e+00 6.29337072e-01 -4.17648226e-01 -7.56901860e-01
3.41808051e-01 7.92269856e-02 -2.27217942e-01 -1.51298940e-01
-1.09955907e+00 9.39669192e-01 1.05890386e-01 -2.04396576e-01
-1.11216915e+00 -8.90039265e-01 -6.52303219e-01 3.73294391e-02
2.13491529e-01 -9.75187778e-01 1.53395295e+00 -1.75282633e+00
-1.38999283e+00 9.96650040e-01 -3.08831573e-01 -3.40141088e-01
9.72227335e-01 -1.26244813e-01 -4.51539874e-01 -2.60762535e-02
2.33848527e-01 1.21624577e+00 1.27941597e+00 -1.45388877e+00
-5.03233075e-01 -3.52835804e-02 4.39770341e-01 2.56829053e-01
-5.78149855e-01 1.73673779e-01 -7.40269780e-01 -1.04884791e+00
-4.03497100e-01 -1.08415198e+00 -1.38905570e-01 3.13646108e-01
-6.35813057e-01 6.96960166e-02 8.22090805e-01 -6.88890994e-01
1.23046589e+00 -2.24481225e+00 1.47787258e-01 1.02413103e-01
2.71985203e-01 -4.67282869e-02 -4.06849742e-01 6.15153611e-01
-1.46345124e-01 5.03069282e-01 -3.99534762e-01 -5.49791694e-01
-1.45808351e-03 1.77914321e-01 -4.26427037e-01 1.51127398e-01
2.98959374e-01 7.60607779e-01 -9.68314886e-01 -4.31921393e-01
-1.82736386e-02 6.47762120e-01 -6.03051305e-01 2.36632079e-01
-6.18056595e-01 3.74508858e-01 -8.77431482e-02 9.01487693e-02
5.19670308e-01 -3.33360940e-01 -5.57695739e-02 -1.40091524e-01
-1.44808233e-01 7.66477780e-03 -7.36763835e-01 1.64751089e+00
-3.99588197e-01 8.90654325e-01 -3.23465526e-01 -2.19137058e-01
7.18536794e-01 2.94274949e-02 1.14580370e-01 -6.79955542e-01
-1.31329507e-01 -1.45222038e-01 -1.02432825e-01 -5.16912580e-01
7.76230037e-01 1.16302287e-02 7.85321072e-02 9.24666882e-01
-3.13718230e-01 -4.27456021e-01 3.21589381e-01 7.08091497e-01
9.70641792e-01 1.31658226e-01 -1.26921296e-01 -2.83209197e-02
-8.44896734e-02 1.05951063e-01 3.78424585e-01 9.03431296e-01
2.04903662e-01 8.55274737e-01 6.76209271e-01 -3.43870610e-01
-1.40082896e+00 -1.00558126e+00 3.49772185e-01 1.05774987e+00
-2.03287825e-02 -5.20595193e-01 -9.85311031e-01 -6.78452194e-01
-6.29664287e-02 1.41754091e+00 -1.08157063e+00 -3.87270689e-01
-3.95751089e-01 -4.71757203e-01 5.90225041e-01 4.12114978e-01
3.01857203e-01 -9.85444367e-01 -5.98891437e-01 1.88353986e-01
2.82310732e-02 -7.02032089e-01 -1.00524628e+00 -2.00969204e-01
-4.62796152e-01 -6.48829103e-01 -1.05556130e+00 -6.03389502e-01
1.31099963e+00 1.96009442e-01 1.20462012e+00 3.03163409e-01
-6.24940544e-02 4.80759442e-01 -1.39363065e-01 -5.84394395e-01
-6.94022417e-01 -1.14698652e-02 -1.77128643e-01 1.63214356e-01
1.70155808e-01 -2.50365138e-01 -4.28983688e-01 1.59502044e-01
-1.15466881e+00 5.96699774e-01 4.37425643e-01 7.79377639e-01
3.60037357e-01 -5.73732890e-02 2.55680025e-01 -1.38320708e+00
9.16973650e-01 -4.92402256e-01 -4.63240802e-01 4.05340075e-01
-6.05688632e-01 2.27487907e-02 5.23550212e-01 -8.46146703e-01
-1.38381565e+00 1.10729528e-03 2.41377234e-01 -5.70542514e-01
-1.12181962e-01 4.48413789e-01 4.38157395e-02 3.08224082e-01
9.14338350e-01 2.10109442e-01 -1.51747704e-01 -3.43756318e-01
5.01796365e-01 4.92122829e-01 4.02914792e-01 -7.08526194e-01
9.05437708e-01 4.96391356e-01 -6.90285623e-01 -7.51412153e-01
-6.73232555e-01 3.04443806e-01 -3.83581817e-01 -2.00575575e-01
7.53402829e-01 -6.84733510e-01 -1.27425894e-01 8.57459605e-01
-1.40528524e+00 -8.92363310e-01 -5.80977142e-01 4.00280394e-02
-2.55481482e-01 -2.03939583e-02 -6.08075619e-01 -5.77953875e-01
-2.05870420e-02 -9.30794418e-01 7.21965492e-01 3.94516766e-01
-7.20173657e-01 -1.05254257e+00 -4.54744548e-02 9.30635035e-02
7.29179323e-01 1.15059607e-01 1.25105894e+00 -3.49207968e-01
-4.37146425e-01 7.50187486e-02 -2.40485221e-01 2.10699975e-01
3.68896633e-01 7.52142251e-01 -9.46500957e-01 -1.78033799e-01
-2.56767511e-01 -1.84405103e-01 6.25857234e-01 1.11041985e-01
1.19115245e+00 -5.69667399e-01 -3.99302006e-01 4.46612209e-01
1.00192297e+00 3.91799957e-01 5.80026269e-01 1.22135378e-01
5.99239469e-01 7.66063213e-01 3.99618626e-01 6.17156744e-01
4.54970479e-01 3.25502813e-01 1.94971144e-01 -4.09373105e-01
-2.07727522e-01 -6.06712341e-01 3.39950949e-01 4.27340180e-01
1.70562133e-01 -6.45183563e-01 -8.20039690e-01 7.04139888e-01
-1.59238660e+00 -9.59429979e-01 1.11210309e-01 2.21361876e+00
1.15311992e+00 2.00671345e-01 -2.21879929e-01 -4.80612338e-01
8.52755189e-01 2.92720169e-01 -9.38924789e-01 -6.11253798e-01
-3.16714644e-01 -1.21083595e-01 2.56841540e-01 6.83150232e-01
-5.93415678e-01 1.19828725e+00 5.98832607e+00 5.63392758e-01
-1.36174524e+00 1.08794123e-01 1.02571499e+00 -4.34714675e-01
-1.03323734e+00 1.76121756e-01 -4.80982929e-01 6.51155114e-01
6.49365187e-01 -5.33554971e-01 8.02072704e-01 5.12661517e-01
2.90899545e-01 -1.72246322e-01 -1.31399572e+00 7.20088720e-01
2.50015944e-01 -1.40743983e+00 4.80458200e-01 -1.00979522e-01
9.38202679e-01 -2.53033668e-01 4.89902854e-01 2.04145864e-01
7.20236301e-01 -1.12306392e+00 1.26600075e+00 5.89182734e-01
1.11917818e+00 -5.47616780e-01 1.11541413e-02 4.13279206e-01
-5.25377333e-01 2.64079839e-01 -4.02662069e-01 -2.98330523e-02
1.61528647e-01 4.53288853e-01 -8.29196036e-01 -5.07556759e-02
7.46109068e-01 5.90272188e-01 -7.11745024e-01 6.03108883e-01
-4.75416332e-01 4.62167591e-01 -2.76327252e-01 -1.21036530e-01
2.54435450e-01 -1.42653152e-01 4.09934372e-01 1.19188249e+00
5.28465033e-01 1.58256069e-01 -4.18235540e-01 1.16367126e+00
-4.16456670e-01 -4.28544283e-02 -8.56767118e-01 -3.60279828e-01
5.41992664e-01 1.09311187e+00 -4.65051204e-01 -7.27747023e-01
-1.71855927e-01 1.52760220e+00 3.86163592e-01 8.15807760e-01
-1.13137937e+00 -4.25960362e-01 6.54327273e-01 3.31506997e-01
2.00052664e-01 5.13290241e-02 -1.44679546e-01 -9.37798738e-01
-1.24489792e-01 -9.67579424e-01 9.98288617e-02 -1.45972812e+00
-1.14382005e+00 6.08407080e-01 -2.87115779e-02 -6.53861642e-01
-2.51693487e-01 -9.03250426e-02 -7.79147923e-01 1.13305199e+00
-1.21783018e+00 -1.10294557e+00 -3.55217040e-01 5.41038096e-01
6.61038339e-01 5.38276508e-02 6.90967798e-01 2.62921721e-01
-5.24266481e-01 6.27128720e-01 -6.29534423e-02 -5.10725155e-02
1.10372126e+00 -1.28722870e+00 8.91917527e-01 7.47782290e-01
2.09683850e-02 8.22528243e-01 1.03171837e+00 -9.29565430e-01
-1.04455853e+00 -1.13095486e+00 1.21626508e+00 -6.48296356e-01
4.85113591e-01 -3.53742361e-01 -9.51775312e-01 1.07716978e+00
6.42300248e-01 -4.35615391e-01 4.95929122e-01 -1.87137768e-01
-4.53592181e-01 -4.20451537e-02 -1.04510260e+00 1.16612101e+00
1.04391563e+00 -5.17106295e-01 -3.19684952e-01 6.72529459e-01
8.88639927e-01 -5.78575850e-01 -4.71216857e-01 -3.20709169e-01
3.31331730e-01 -8.03886592e-01 6.87018633e-01 -8.71971846e-01
6.69773638e-01 -1.08761340e-01 3.19053113e-01 -1.79570878e+00
-4.70906049e-01 -8.59372258e-01 1.62467808e-01 1.52878487e+00
7.53621817e-01 -6.02983713e-01 6.35235965e-01 1.15190029e+00
3.85057251e-03 -4.07607555e-01 -4.89922971e-01 -4.61918861e-01
1.43060848e-01 -1.98811769e-01 8.29477668e-01 1.34861517e+00
-4.11642671e-01 4.10566777e-01 -5.93447268e-01 -1.71071626e-02
3.22263420e-01 -1.37051031e-01 8.36411953e-01 -8.80863070e-01
-2.69081414e-01 -5.00437856e-01 4.16014075e-01 -1.09287477e+00
-6.35073287e-03 -5.70293248e-01 -1.40225142e-01 -1.59584713e+00
1.93577081e-01 -7.03924179e-01 -3.45014632e-02 6.10913694e-01
-2.18924314e-01 -4.37911600e-02 2.73446888e-01 1.64578393e-01
-1.25381961e-01 5.45089364e-01 1.30008233e+00 -5.81759512e-01
-9.15100276e-02 -2.91655779e-01 -9.85039592e-01 6.99188113e-01
8.45906079e-01 -7.32962906e-01 -7.20224440e-01 -1.19387603e+00
5.03422737e-01 -2.01947972e-01 2.97777981e-01 -6.79421484e-01
4.23657596e-01 -4.86528516e-01 4.98661786e-01 -1.36804864e-01
5.21920085e-01 -6.80756450e-01 5.67319453e-01 4.24919069e-01
-6.46337450e-01 5.07656872e-01 2.13059634e-01 4.57358986e-01
1.30200014e-01 -3.40354264e-01 5.69027185e-01 -2.30925739e-01
-5.84555268e-01 2.19784737e-01 -3.58243823e-01 1.33212224e-01
1.14732230e+00 -2.97885150e-01 -6.17047191e-01 -7.99490809e-01
-4.34527934e-01 1.34906933e-01 1.00191033e+00 6.86088502e-01
5.18706322e-01 -1.10706437e+00 -7.71491349e-01 2.16026738e-01
1.77990228e-01 -4.49164882e-02 3.75017792e-01 3.80695105e-01
-5.37909269e-01 -1.06605120e-01 -1.12823851e-01 -3.50703627e-01
-1.10865712e+00 6.54100776e-01 2.04570770e-01 1.26326621e-01
-4.95682389e-01 1.04075444e+00 5.88207483e-01 -3.16557109e-01
1.43090904e-01 -2.77261496e-01 3.60973924e-02 8.98643360e-02
5.34727156e-01 -9.36765522e-02 -3.93483847e-01 -4.99889016e-01
-5.45969792e-02 3.99877161e-01 -4.43318933e-01 -5.17626286e-01
1.10856307e+00 -1.18260860e-01 -1.07352072e-02 5.50900698e-01
7.94942200e-01 1.56684086e-01 -1.61262381e+00 -2.32258528e-01
-4.61018562e-01 -5.56465864e-01 -2.16196492e-01 -1.15391552e+00
-1.06449389e+00 9.80018258e-01 3.73426795e-01 1.25334352e-01
8.98094118e-01 -1.04401119e-01 5.40832758e-01 2.63866842e-01
5.06309792e-03 -1.26440156e+00 2.47653887e-01 3.19305032e-01
1.26027942e+00 -1.03604579e+00 -1.11370027e-01 -1.01298384e-01
-1.05608737e+00 7.16764092e-01 9.37833965e-01 3.43773186e-01
1.99158534e-01 1.61406651e-01 3.03580791e-01 -5.80038950e-02
-1.16958928e+00 4.66487139e-01 -1.29881814e-01 7.63128281e-01
3.46153975e-01 4.35279906e-02 8.92703384e-02 2.17036232e-01
-2.87073106e-01 -9.71919000e-02 7.26509452e-01 8.76216054e-01
-1.18087262e-01 -8.54093432e-01 -2.66496569e-01 5.05874217e-01
-1.65732801e-01 -4.16346252e-01 -7.77069271e-01 7.49569178e-01
-2.73324512e-02 6.53238893e-01 4.04445201e-01 -2.12420538e-01
3.00246894e-01 -1.09730922e-01 3.42117697e-01 -7.17881501e-01
-8.10969174e-01 -1.66820288e-01 1.28432602e-01 -1.86532825e-01
1.67307314e-02 -6.52204156e-01 -1.09940028e+00 -6.24809384e-01
2.69633252e-02 -2.34052110e-02 6.37536764e-01 5.95542252e-01
5.82629383e-01 5.38873255e-01 3.09364766e-01 -5.78657925e-01
-3.87352854e-01 -9.16892648e-01 -2.99240440e-01 6.96757019e-01
9.04625356e-02 -2.71651626e-01 -1.75737679e-01 5.20065427e-01] | [11.409403800964355, -0.22310779988765717] |
d5d42635-4621-48b4-b5e0-3b6017b1b538 | mutually-guided-few-shot-learning-for | 2306.13310 | null | https://arxiv.org/abs/2306.13310v1 | https://arxiv.org/pdf/2306.13310v1.pdf | Mutually Guided Few-shot Learning for Relational Triple Extraction | Knowledge graphs (KGs), containing many entity-relation-entity triples, provide rich information for downstream applications. Although extracting triples from unstructured texts has been widely explored, most of them require a large number of labeled instances. The performance will drop dramatically when only few labeled data are available. To tackle this problem, we propose the Mutually Guided Few-shot learning framework for Relational Triple Extraction (MG-FTE). Specifically, our method consists of an entity-guided relation proto-decoder to classify the relations firstly and a relation-guided entity proto-decoder to extract entities based on the classified relations. To draw the connection between entity and relation, we design a proto-level fusion module to boost the performance of both entity extraction and relation classification. Moreover, a new cross-domain few-shot triple extraction task is introduced. Extensive experiments show that our method outperforms many state-of-the-art methods by 12.6 F1 score on FewRel 1.0 (single-domain) and 20.5 F1 score on FewRel 2.0 (cross-domain). | ['Lianghua He', 'Chen Ma', 'Bowei He', 'Shuai Jiang', 'Chengmei Yang'] | 2023-06-23 | null | null | null | null | ['cross-domain-few-shot', 'knowledge-graphs', 'few-shot-learning', 'relation-classification'] | ['computer-vision', 'knowledge-base', 'methodology', 'natural-language-processing'] | [-2.17579335e-01 5.88099718e-01 -6.18991673e-01 -4.22419697e-01
-1.17617309e+00 -2.14790717e-01 6.16059721e-01 5.49455464e-01
-1.90304533e-01 9.40634906e-01 1.73647225e-01 -4.64733131e-02
8.63745958e-02 -1.21789300e+00 -7.95380473e-01 -3.56586516e-01
1.09676741e-01 6.95504069e-01 6.28023148e-01 -3.94374251e-01
-5.01371503e-01 -1.19218938e-01 -1.36306310e+00 3.95786852e-01
1.14759147e+00 9.91272569e-01 -1.58879369e-01 5.70693351e-02
-3.14842820e-01 1.14458179e+00 -1.39482692e-01 -9.64932203e-01
-1.10339247e-01 -2.45971978e-01 -1.12001884e+00 -1.54002577e-01
-2.67948002e-01 1.39804736e-01 -4.15078074e-01 1.00597990e+00
5.44239402e-01 1.33900926e-01 6.24037802e-01 -1.22290266e+00
-5.55940390e-01 1.26184428e+00 -7.68200517e-01 4.29246724e-02
3.47960800e-01 -2.58095235e-01 1.41927743e+00 -1.20572388e+00
9.78474796e-01 1.08294308e+00 5.23407042e-01 9.12900567e-02
-9.96228933e-01 -6.49347901e-01 -1.28051221e-01 4.82293546e-01
-1.63250089e+00 -7.14893520e-01 4.21585053e-01 -2.65425533e-01
1.22186553e+00 -8.82468969e-02 3.37299466e-01 8.29402030e-01
-2.16162905e-01 9.12200212e-01 4.77197617e-01 -5.54736197e-01
1.40226573e-01 9.51384455e-02 4.67664063e-01 6.64269090e-01
5.92248142e-01 -1.22264385e-01 -6.87634408e-01 2.86435820e-02
6.61699772e-02 -3.03026855e-01 -1.07997470e-01 -3.13377291e-01
-1.00731623e+00 5.85506260e-01 5.57667196e-01 3.99650872e-01
-3.16710621e-01 -2.80200779e-01 6.35814965e-01 2.09000364e-01
6.23600364e-01 1.08743779e-01 -7.52075791e-01 1.13135897e-01
-6.30209506e-01 1.79793656e-01 9.28088129e-01 1.48516703e+00
9.35943663e-01 -4.51884419e-01 -4.22653675e-01 9.86945748e-01
2.75103509e-01 1.64740175e-01 2.02626884e-01 -1.67987943e-01
9.48944926e-01 1.09385967e+00 -1.52514845e-01 -9.34356153e-01
-5.47562122e-01 -2.29695305e-01 -8.35637450e-01 -5.91106296e-01
8.15052018e-02 -3.71191323e-01 -8.80051553e-01 1.38747382e+00
8.88628125e-01 2.93702632e-01 5.25243163e-01 6.99750245e-01
1.53879392e+00 5.13478637e-01 2.96010286e-01 -3.98757041e-01
1.74484742e+00 -1.01410794e+00 -9.57111895e-01 -2.28912458e-01
1.21733320e+00 -4.53705668e-01 5.66855013e-01 -3.34757715e-01
-7.85322726e-01 -3.62593591e-01 -1.03194213e+00 -4.47553396e-01
-5.95764935e-01 6.88957945e-02 6.18035376e-01 2.74846762e-01
-2.65849590e-01 3.94086301e-01 -7.28443384e-01 -3.66996169e-01
4.37861890e-01 2.37771869e-01 -5.58591902e-01 -4.14396599e-02
-1.87360859e+00 9.47378695e-01 1.10329425e+00 -9.30865034e-02
-4.16039646e-01 -8.89345169e-01 -1.36481404e+00 3.58390778e-01
1.06010067e+00 -5.92050254e-01 1.16831088e+00 1.83524251e-01
-9.31910753e-01 8.76466870e-01 -2.28335679e-01 -4.34395224e-01
9.35764760e-02 -2.81259418e-01 -8.08492899e-01 -9.49963927e-02
4.03613955e-01 2.40882248e-01 1.09801650e-01 -1.02335644e+00
-1.01272082e+00 -3.64351809e-01 2.99917185e-03 2.59639531e-01
-1.56037763e-01 -3.96872833e-02 -8.21230471e-01 -3.39252144e-01
5.69374533e-03 -5.96468568e-01 -5.24931140e-02 -7.99642026e-01
-8.48003805e-01 -6.34398162e-01 7.97903061e-01 -4.75239575e-01
1.46460116e+00 -2.04508615e+00 4.70462851e-02 4.83967327e-02
5.06163061e-01 5.09616613e-01 1.39073119e-01 5.22919953e-01
-2.57670701e-01 1.29630968e-01 1.02474399e-01 -8.05032998e-02
2.28372477e-02 -3.59974839e-02 -5.75918965e-02 1.09519914e-01
4.73821074e-01 1.33411837e+00 -1.18309915e+00 -9.20175254e-01
-1.72044143e-01 2.26399377e-01 -6.01851270e-02 1.53918475e-01
-2.30706885e-01 -3.93853150e-02 -6.24499261e-01 8.72360170e-01
5.88625371e-01 -6.01392567e-01 6.09471083e-01 -7.36993670e-01
2.04903156e-01 3.80330294e-01 -1.02895451e+00 1.43096340e+00
-1.61310509e-01 1.90642998e-01 -4.05039698e-01 -9.40285265e-01
9.19235289e-01 5.32296658e-01 5.09527266e-01 -5.73319972e-01
3.08568120e-01 3.15454483e-01 -9.52766836e-02 -6.84109747e-01
5.11836529e-01 -3.39456260e-01 -3.93598855e-01 9.81232971e-02
5.54218233e-01 3.22648883e-01 7.45571375e-01 6.36968970e-01
1.42066145e+00 8.51097256e-02 7.51744330e-01 1.26959816e-01
4.88935441e-01 3.80109400e-02 1.08141351e+00 1.59395203e-01
-4.73638810e-03 2.01084450e-01 8.51638496e-01 -2.82282174e-01
-8.04090917e-01 -6.76340044e-01 4.46594954e-02 9.54137444e-01
3.80789012e-01 -9.09208953e-01 -3.77456695e-01 -1.14883912e+00
1.37220547e-01 6.28745496e-01 -4.89755750e-01 -2.86671162e-01
-3.07593822e-01 -9.27552521e-01 5.97992301e-01 5.50579667e-01
5.96738577e-01 -8.66625428e-01 -4.86447401e-02 2.43627310e-01
-5.88026166e-01 -1.70642197e+00 -9.81913954e-02 2.95751035e-01
-3.55573624e-01 -1.35724068e+00 -2.76016384e-01 -8.52498829e-01
2.28947446e-01 5.64984269e-02 1.41326296e+00 -1.40270278e-01
6.23231474e-03 -3.67073953e-01 -8.06262136e-01 -1.95368946e-01
-3.25899012e-02 5.45427024e-01 -6.44028112e-02 -5.77891767e-02
9.55950677e-01 -3.66713554e-01 -7.88497552e-02 1.26484260e-01
-5.22623658e-01 2.36280486e-01 6.81999683e-01 8.40115488e-01
6.94316268e-01 3.64312917e-01 8.49991739e-01 -1.57865918e+00
4.67669636e-01 -7.47416317e-01 -3.60686868e-01 7.61206567e-01
-6.53217077e-01 2.01563567e-01 3.89194131e-01 -6.00921288e-02
-1.32825196e+00 1.73498631e-01 1.37538388e-01 -2.14644477e-01
1.29557401e-01 9.91755545e-01 -6.37747228e-01 4.58475202e-01
6.72037542e-01 -1.50834516e-01 -7.06356227e-01 -4.78507936e-01
7.13113248e-01 7.98171282e-01 5.21387637e-01 -5.23745716e-01
8.97262633e-01 4.80708629e-02 -6.58482239e-02 -4.75566268e-01
-1.46494365e+00 -6.46203756e-01 -8.57338905e-01 4.66180071e-02
9.19999003e-01 -1.26881683e+00 -7.23801553e-01 1.70322537e-01
-1.05217564e+00 2.50085555e-02 -3.91691983e-01 4.10196155e-01
-1.95568442e-01 9.95396003e-02 -7.18267858e-01 -6.79649293e-01
-4.47696924e-01 -8.15683484e-01 1.16855955e+00 3.51616979e-01
-6.55783117e-02 -6.71759307e-01 2.06956685e-01 5.87054014e-01
-3.24053675e-01 2.83583224e-01 1.00031447e+00 -1.12424052e+00
-4.57836956e-01 -1.76259935e-01 -6.27202570e-01 -2.74090409e-01
2.60156959e-01 -2.96970963e-01 -7.51301348e-01 1.12768546e-01
-6.01221502e-01 -6.57039225e-01 9.35015261e-01 -2.34518975e-01
4.21789348e-01 -3.80663574e-02 -9.52126622e-01 2.29642138e-01
1.22746336e+00 1.57150365e-02 5.37916064e-01 1.02741388e-03
1.13151312e+00 5.67608237e-01 1.08836305e+00 3.94653171e-01
9.58292544e-01 6.86209381e-01 -2.76493151e-02 2.91967429e-02
-1.31512970e-01 -5.04028440e-01 5.29551692e-02 8.83278847e-01
-5.45004345e-02 -2.65928060e-01 -1.16036999e+00 6.94942772e-01
-2.23646855e+00 -8.89853895e-01 -3.75993371e-01 1.78683817e+00
1.37893438e+00 2.35041380e-01 1.92539260e-01 1.82890251e-01
9.32085276e-01 5.35111763e-02 -4.48749036e-01 2.81196356e-01
-8.96794423e-02 3.58396620e-02 4.06176895e-01 2.61388540e-01
-1.32788014e+00 1.35689712e+00 4.92819738e+00 1.11747634e+00
-4.61261272e-01 1.98786527e-01 3.79258245e-01 1.65211305e-01
-2.46899158e-01 2.91569144e-01 -1.30636060e+00 3.61888766e-01
1.08642817e+00 -5.63845575e-01 -2.94927329e-01 7.94251502e-01
-2.18862608e-01 -1.08267255e-01 -1.18520010e+00 8.40419292e-01
-1.59004748e-01 -1.40728486e+00 -1.57906085e-01 -9.30758286e-03
6.34269893e-01 -4.06829193e-02 -6.64029419e-01 9.01616812e-01
5.87843716e-01 -6.71780705e-01 2.52868891e-01 4.11340922e-01
9.32256103e-01 -9.08091843e-01 1.03051329e+00 3.53438944e-01
-1.76986182e+00 1.80189773e-01 -3.67337018e-01 2.62177169e-01
4.80967790e-01 1.26176345e+00 -8.34182382e-01 1.27253425e+00
5.95984399e-01 9.56348956e-01 -3.18483382e-01 7.30901539e-01
-4.39198703e-01 3.03663164e-01 -1.13707989e-01 1.06797308e-01
-8.90053883e-02 4.59680855e-02 2.48672619e-01 1.26770127e+00
4.74014841e-02 5.61152756e-01 3.18160355e-01 5.29693246e-01
-5.48131883e-01 2.69668221e-01 -5.90341747e-01 -2.42598861e-01
7.96040654e-01 1.50600791e+00 -5.71104825e-01 -5.96227050e-01
-6.81984723e-01 6.05780065e-01 9.27403569e-01 1.69641659e-01
-8.63046885e-01 -7.96727777e-01 3.57297182e-01 1.19236065e-02
5.94015956e-01 2.70925313e-01 1.06449556e-02 -1.47238827e+00
-3.85759003e-03 -5.81212163e-01 7.43299544e-01 -5.62163413e-01
-1.37489676e+00 5.61490834e-01 2.57247891e-02 -1.15092850e+00
-2.21142039e-01 -2.21728772e-01 -1.99159741e-01 5.22887588e-01
-1.47555339e+00 -1.41232383e+00 -2.18663350e-01 3.00425231e-01
2.42826834e-01 -1.83331653e-01 7.09989250e-01 7.36679792e-01
-1.06836164e+00 7.83010960e-01 -2.90204048e-01 6.95314407e-01
7.53132284e-01 -1.18202066e+00 4.82393444e-01 7.93840945e-01
2.04309002e-01 5.18230200e-01 3.61134768e-01 -1.07873738e+00
-1.29228437e+00 -1.38006973e+00 1.51260519e+00 -3.46582979e-01
7.68607855e-01 -2.92226791e-01 -9.75391507e-01 9.23087955e-01
-2.10907888e-02 4.54625517e-01 8.89309168e-01 7.56701171e-01
-6.15574241e-01 -1.77874744e-01 -1.08935869e+00 4.70256776e-01
1.14370322e+00 -5.38641632e-01 -6.95041180e-01 4.34344620e-01
1.10932326e+00 -5.86408377e-01 -1.33890629e+00 8.07427466e-01
2.18762666e-01 -5.93243420e-01 8.09083700e-01 -7.75073290e-01
6.05557144e-01 -3.72102499e-01 -1.75363332e-01 -1.22794998e+00
-2.67387807e-01 -4.44820076e-01 -9.40737307e-01 1.76864552e+00
8.56211543e-01 -2.36950219e-01 6.47474408e-01 6.02584898e-01
-7.37034809e-03 -8.75922024e-01 -6.98304355e-01 -8.56023431e-01
-3.49712670e-01 -2.04381153e-01 7.00106800e-01 1.13589609e+00
8.09760571e-01 1.38094246e+00 -3.31270933e-01 2.82652706e-01
5.62529743e-01 3.94492865e-01 6.84134901e-01 -1.41713929e+00
-1.00396134e-01 1.11842923e-01 -4.39867109e-01 -6.07464671e-01
3.36540580e-01 -1.11803222e+00 1.11018382e-01 -1.64597261e+00
5.90006709e-01 -5.31710505e-01 -2.36391231e-01 9.30818558e-01
-5.90774596e-01 -7.11406842e-02 -8.94495100e-02 5.02207242e-02
-1.09910870e+00 7.26087451e-01 9.03797328e-01 -8.54576826e-02
-1.84379861e-01 -2.87123650e-01 -8.91054153e-01 5.13020217e-01
4.94718105e-01 -5.61681747e-01 -5.19314229e-01 -3.45153827e-03
3.97463292e-01 1.85037047e-01 -2.10877538e-01 -8.00957203e-01
5.30918956e-01 -1.41304964e-02 1.14288114e-01 -8.57758403e-01
1.54849857e-01 -6.05048537e-01 1.00052282e-01 3.30552720e-02
-1.52087152e-01 -4.60173219e-01 -1.20512865e-01 7.33843923e-01
-3.91883552e-01 -1.16797082e-01 4.96449113e-01 7.06162006e-02
-9.18807089e-01 4.90768999e-01 3.20410103e-01 5.64720154e-01
1.37982702e+00 3.23666096e-01 -6.25552654e-01 1.41639514e-02
-8.61046970e-01 6.44797742e-01 -1.56695545e-02 2.94827938e-01
3.90476704e-01 -1.62620485e+00 -7.57002175e-01 -1.27682373e-01
6.25948846e-01 3.15882951e-01 2.08883911e-01 8.97293687e-01
1.14739221e-02 3.94150406e-01 2.68620402e-01 -9.49151218e-02
-1.29808426e+00 7.70008504e-01 1.49191301e-02 -7.08906233e-01
-7.13772893e-01 9.39418256e-01 -1.04822442e-01 -4.30053294e-01
-2.45532598e-02 -8.40424597e-02 -4.54195857e-01 5.28519273e-01
5.22025466e-01 4.50492710e-01 2.39109680e-01 -7.24160314e-01
-6.26954734e-01 1.01346947e-01 -4.27009284e-01 1.06523953e-01
1.34937406e+00 -8.32386762e-02 -1.71351060e-01 4.81714934e-01
1.01877058e+00 -7.51517192e-02 -6.64318919e-01 -7.50989139e-01
5.13777077e-01 -2.35981360e-01 -1.24727689e-01 -7.10306704e-01
-1.16233695e+00 5.07228851e-01 -1.82214230e-01 2.48342663e-01
8.27529132e-01 4.42570955e-01 1.05949044e+00 4.76084650e-01
5.05972028e-01 -9.83138025e-01 -3.37772787e-01 6.63133323e-01
3.68129700e-01 -1.46463633e+00 1.95554048e-01 -1.17964935e+00
-8.61280322e-01 5.90004086e-01 7.98890531e-01 3.23637336e-01
8.34157884e-01 4.34878349e-01 -2.16285050e-01 -5.27389944e-01
-9.99893963e-01 -8.03267956e-01 4.45724398e-01 4.83193040e-01
7.07592905e-01 8.23607445e-02 -4.08798188e-01 1.13100696e+00
5.82577400e-02 1.78912237e-01 1.22497275e-01 8.58635783e-01
-4.73333359e-01 -1.27915823e+00 3.12021464e-01 6.03396893e-01
-3.10227603e-01 -3.29487890e-01 -2.93256670e-01 6.20837748e-01
2.37892464e-01 1.07574701e+00 -3.15981179e-01 -8.11240911e-01
5.19830048e-01 2.04464316e-01 1.71774864e-01 -1.04179418e+00
-3.91642064e-01 -1.51445806e-01 7.18846023e-01 -3.58896762e-01
-4.48707074e-01 -4.81351137e-01 -1.55901849e+00 -4.14710760e-01
-7.50373244e-01 4.29852724e-01 -1.22171842e-01 1.27998626e+00
6.36862934e-01 6.40402555e-01 3.26612800e-01 -1.92488551e-01
-1.92634717e-01 -1.00631440e+00 -7.16402948e-01 2.82874256e-01
-1.90632120e-01 -9.44974303e-01 2.01471522e-01 4.14780490e-02] | [9.289665222167969, 8.601180076599121] |
53b97897-4f97-478b-9d5d-32b14e2212b5 | neural-temporality-adaptation-for-document | null | null | https://aclanthology.org/P19-1403 | https://aclanthology.org/P19-1403.pdf | Neural Temporality Adaptation for Document Classification: Diachronic Word Embeddings and Domain Adaptation Models | Language usage can change across periods of time, but document classifiers models are usually trained and tested on corpora spanning multiple years without considering temporal variations. This paper describes two complementary ways to adapt classifiers to shifts across time. First, we show that diachronic word embeddings, which were originally developed to study language change, can also improve document classification, and we show a simple method for constructing this type of embedding. Second, we propose a time-driven neural classification model inspired by methods for domain adaptation. Experiments on six corpora show how these methods can make classifiers more robust over time. | ['Xiaolei Huang', 'Michael J. Paul'] | 2019-07-01 | null | null | null | acl-2019-7 | ['diachronic-word-embeddings'] | ['natural-language-processing'] | [-3.47436070e-02 -5.06069601e-01 -7.96253562e-01 -5.74060857e-01
-1.08227804e-01 -6.99108899e-01 1.10876620e+00 3.14140081e-01
-9.07558978e-01 8.91421676e-01 4.61262167e-01 -4.88776326e-01
-1.40776336e-02 -9.13312316e-01 -2.37030819e-01 -3.18556756e-01
-1.99158132e-01 1.96210384e-01 2.33628988e-01 -3.91939998e-01
3.49045455e-01 5.87546766e-01 -1.32452703e+00 2.27260552e-02
6.76629364e-01 3.21310967e-01 -1.72812894e-01 7.94639587e-01
-5.12725294e-01 2.09586024e-01 -8.00574958e-01 -2.79098421e-01
-5.93362190e-02 -1.71759740e-01 -1.02177823e+00 -1.27277434e-01
1.97795764e-01 -4.08249348e-02 -6.53090119e-01 8.36846292e-01
2.68835008e-01 3.77829880e-01 1.04561031e+00 -9.61459458e-01
-1.44202244e+00 9.22610104e-01 -2.32590765e-01 7.65854001e-01
3.29664826e-01 -2.15557232e-01 8.31870437e-01 -5.99379957e-01
9.33354199e-01 1.15979981e+00 1.09454739e+00 8.56295228e-01
-1.19471824e+00 -5.14019191e-01 6.12473965e-01 5.62703669e-01
-1.05170202e+00 -2.12170526e-01 1.00733984e+00 -6.83903217e-01
1.28025150e+00 1.80105582e-01 7.61512041e-01 1.75720370e+00
4.29772556e-01 7.60556459e-01 1.04756129e+00 -8.93614829e-01
1.04826433e-03 3.66469502e-01 6.52152181e-01 2.27643400e-01
2.45760828e-01 -1.04377288e-02 -3.93146425e-01 -1.56426340e-01
2.03118935e-01 3.08538586e-01 -3.41385454e-01 -1.96869746e-01
-1.14226079e+00 9.96266425e-01 1.91408500e-01 1.32483041e+00
4.23560627e-02 -5.85509278e-02 7.68642306e-01 5.16767561e-01
1.04622054e+00 7.17477918e-01 -9.18659866e-01 -4.26786989e-01
-7.51519442e-01 4.26926911e-02 6.85205162e-01 6.68240905e-01
5.58171332e-01 2.70377666e-01 -1.89777017e-01 1.22451544e+00
-7.10880756e-02 3.52598101e-01 1.59433126e+00 -2.51207858e-01
1.54802546e-01 2.74604440e-01 -1.08666122e-01 -7.53097534e-01
-6.08762443e-01 -1.50758594e-01 -7.04658985e-01 -4.01673794e-01
3.55307847e-01 2.90580397e-03 -9.58779871e-01 1.82892787e+00
-4.43996340e-02 1.67190563e-02 8.15960616e-02 4.49138470e-02
2.58590460e-01 7.16236711e-01 1.69700846e-01 -3.61374438e-01
1.07467055e+00 -8.75947654e-01 -9.81176376e-01 -8.19602832e-02
9.85525906e-01 -2.92095244e-01 1.07618749e+00 2.13461772e-01
-5.94792008e-01 -5.06250978e-01 -1.00769508e+00 -5.43730333e-03
-1.41281486e+00 -4.78237629e-01 6.93688035e-01 8.29730153e-01
-8.26836228e-01 8.79038990e-01 -7.15355635e-01 -1.01668453e+00
1.10296682e-01 9.15697664e-02 -2.80081213e-01 2.17321739e-01
-1.65053618e+00 1.19981444e+00 5.59775054e-01 -2.29988962e-01
-2.39839762e-01 -4.44117963e-01 -7.96498954e-01 -1.88025311e-01
-3.46110880e-01 -4.06879216e-01 1.46731591e+00 -1.09544313e+00
-1.63337958e+00 8.98929060e-01 -2.80697703e-01 -5.25585115e-01
1.38305813e-01 2.38395315e-02 -1.09399438e+00 -3.94890875e-01
3.50024663e-02 2.94924349e-01 6.87721193e-01 -8.70353878e-01
-6.27497375e-01 -4.62059855e-01 -1.80732220e-01 -2.34591633e-01
-1.43735158e+00 -8.99385288e-02 -2.09749758e-01 -1.06873620e+00
-4.11013752e-01 -7.52077997e-01 -2.01777175e-01 -4.33932066e-01
1.41147658e-01 -8.88241768e-01 7.81189024e-01 -7.70945728e-01
1.60661566e+00 -2.03792191e+00 2.40151823e-01 3.74412327e-03
-2.64040768e-01 2.17298225e-01 -3.10707152e-01 5.55089593e-01
-1.46244377e-01 5.07946432e-01 -3.89112025e-01 -1.58818722e-01
9.84201208e-02 5.58179557e-01 -4.18972611e-01 3.39386255e-01
1.13113590e-01 8.85832846e-01 -1.12160921e+00 -3.59761715e-03
3.57737206e-02 3.01448375e-01 -3.00920159e-01 -1.90860555e-01
9.11106635e-03 1.06127359e-01 -1.95468843e-01 5.16412675e-01
4.61105913e-01 3.20684195e-01 4.51983958e-01 4.12339568e-01
-2.84931809e-01 3.53050947e-01 -5.67127585e-01 1.62131834e+00
-8.61135423e-01 1.10773230e+00 -6.30553126e-01 -1.29939151e+00
8.75455201e-01 1.11565568e-01 4.27430212e-01 -8.26523721e-01
8.11312422e-02 1.19415656e-01 -5.89484125e-02 -5.97470105e-01
6.65665746e-01 -1.32423699e-01 -4.95089114e-01 5.00475168e-01
1.33131549e-01 -2.33601496e-01 3.64476770e-01 -3.11204165e-01
1.17538369e+00 -1.71003208e-01 6.00640595e-01 -3.47480118e-01
4.83474284e-01 9.91666913e-02 6.06974423e-01 5.97583115e-01
-4.71764863e-01 1.08551010e-01 1.63875103e-01 -8.54823828e-01
-1.09095395e+00 -8.78691971e-01 -6.02711320e-01 1.64495564e+00
-3.18479806e-01 -4.16072994e-01 -2.94446558e-01 -9.36315298e-01
3.00492972e-01 9.77570474e-01 -1.01549101e+00 -4.38009858e-01
-6.25117242e-01 -8.23501348e-01 4.77846593e-01 8.55099797e-01
5.60080893e-02 -1.08129704e+00 -2.12183967e-01 4.48356479e-01
2.57482797e-01 -6.36408091e-01 -5.47055364e-01 4.84615266e-01
-1.07159579e+00 -8.33141387e-01 -9.41550493e-01 -9.04941022e-01
3.21989924e-01 2.06193194e-01 1.13323438e+00 -8.53351429e-02
-1.16987176e-01 8.21439505e-01 -6.92297518e-01 -6.90725744e-01
-4.69843507e-01 6.16424859e-01 6.88908041e-01 -2.85201967e-01
1.01130641e+00 -6.15899742e-01 -1.09091699e-01 -1.70875594e-01
-1.04977095e+00 -7.11819291e-01 1.36846855e-01 1.05287957e+00
4.59551299e-03 5.18864766e-02 7.48854518e-01 -1.13317990e+00
1.22574246e+00 -7.26832926e-01 -2.40928322e-01 3.63923639e-01
-1.15844166e+00 1.43287465e-01 6.73826635e-01 -1.12331128e+00
-8.85987639e-01 -3.64959776e-01 -6.77271485e-02 -7.30360597e-02
-1.30207554e-01 5.71375251e-01 4.34105515e-01 4.33744311e-01
8.58267784e-01 3.87467027e-01 -3.63251030e-01 -6.79189444e-01
6.31387234e-01 1.11181509e+00 3.59721512e-01 -5.17773390e-01
8.46741557e-01 2.56782681e-01 -7.22116828e-01 -9.10387039e-01
-5.29612005e-01 -6.45818353e-01 -1.26455784e+00 8.57715309e-02
6.38477087e-01 -5.80516994e-01 6.24587163e-02 4.05078799e-01
-1.18891072e+00 -4.49891299e-01 -4.23181295e-01 5.39029300e-01
-1.82091683e-01 2.25515604e-01 -4.27743703e-01 -6.58505261e-01
-5.37064299e-02 -3.73105079e-01 5.07631660e-01 2.08059952e-01
-6.80901051e-01 -1.84260678e+00 7.95494318e-01 -4.22978044e-01
5.90811074e-01 5.14422432e-02 9.98029292e-01 -8.78208041e-01
3.99643213e-01 -2.28245139e-01 2.50688344e-01 3.89576077e-01
6.19310856e-01 2.86601067e-01 -1.01246083e+00 -4.44343030e-01
-2.87961841e-01 1.61355123e-01 1.06846738e+00 2.12413579e-01
1.50774288e+00 -3.91481936e-01 -7.58195937e-01 4.73256648e-01
1.12793648e+00 6.34541154e-01 2.58690625e-01 8.45117450e-01
5.53262055e-01 4.43935424e-01 1.68278843e-01 2.30716512e-01
3.92208070e-01 6.91793740e-01 -3.51774812e-01 1.73066035e-01
-9.52494964e-02 -1.21174529e-01 5.82167268e-01 1.47603667e+00
-2.19591260e-02 -2.03445047e-01 -1.19154215e+00 1.23879361e+00
-1.73990417e+00 -1.11189854e+00 4.55365218e-02 1.79077339e+00
1.15023077e+00 2.19118163e-01 1.58965603e-01 3.00026327e-01
4.12889212e-01 3.17794383e-01 -6.08822048e-01 -1.14173830e+00
-2.40148142e-01 6.05456233e-01 6.17761016e-01 4.23704237e-01
-1.06411636e+00 1.06762159e+00 8.11731815e+00 4.38707262e-01
-1.21888328e+00 4.59954739e-01 1.03040248e-01 7.47552589e-02
-5.05196214e-01 -2.57548332e-01 -6.26509964e-01 5.76503456e-01
1.32100761e+00 -6.81993783e-01 1.28927082e-01 7.79571772e-01
-1.82532445e-01 3.81540030e-01 -1.31001949e+00 7.24518716e-01
3.36872816e-01 -1.24201918e+00 1.14755630e-01 -1.92677587e-01
1.01763475e+00 8.11733976e-02 1.24831788e-01 8.69562268e-01
4.22801197e-01 -7.29278564e-01 4.05868322e-01 6.41543806e-01
7.57234752e-01 -6.78093553e-01 6.97159886e-01 1.56492963e-01
-9.81694341e-01 -4.34903949e-01 -4.00427610e-01 -2.65889615e-01
-1.45972148e-01 4.66151386e-01 -7.45148242e-01 3.68388176e-01
9.10609722e-01 1.19265723e+00 -9.01058435e-01 6.87570453e-01
-1.02572806e-01 6.45056009e-01 5.28764054e-02 -4.56523657e-01
1.24823794e-01 6.46523908e-02 4.34279770e-01 1.79948783e+00
4.51520294e-01 -3.74445915e-01 -1.64535508e-01 2.57619500e-01
-1.13932557e-01 1.80016905e-01 -1.31385505e+00 -2.77880728e-01
5.30100524e-01 7.27110624e-01 -4.17878121e-01 -4.16203469e-01
-6.61799312e-01 1.29122949e+00 4.60919768e-01 4.92596298e-01
-5.29291928e-01 -7.81715095e-01 9.12267566e-01 -6.84089810e-02
2.93811828e-01 -4.97272283e-01 -2.15666935e-01 -1.35355222e+00
-9.92779583e-02 -5.68080068e-01 6.02321744e-01 -3.08345139e-01
-1.82735610e+00 4.55543637e-01 2.09258139e-01 -1.12241900e+00
-4.37352747e-01 -1.05023968e+00 -7.00368166e-01 5.05523086e-01
-1.60926235e+00 -9.07602012e-01 2.18316466e-01 3.35147768e-01
9.36423838e-01 -5.36137402e-01 1.12535775e+00 6.06657751e-02
-6.84309244e-01 7.16253102e-01 8.91214609e-01 1.99475303e-01
9.41811979e-01 -1.49310899e+00 4.89930898e-01 6.90190673e-01
4.94195521e-01 8.68406534e-01 4.43764448e-01 -2.66340703e-01
-9.49805081e-01 -1.06053019e+00 1.35960770e+00 -8.11133027e-01
1.05156338e+00 -4.66192663e-01 -1.20202720e+00 8.63727152e-01
5.99313021e-01 -4.90215302e-01 9.78229702e-01 8.25835466e-01
-7.43545055e-01 -2.09492579e-01 -8.70935321e-01 7.54559100e-01
1.17402077e+00 -8.55793238e-01 -1.24245620e+00 5.69403410e-01
8.17132950e-01 1.83944538e-01 -9.77196872e-01 2.17523798e-01
7.29328752e-01 -3.57108146e-01 7.26782918e-01 -1.05105805e+00
3.33128497e-02 1.70134947e-01 -1.49053127e-01 -1.84432828e+00
-7.22145259e-01 -4.72511351e-01 -2.77399808e-01 1.40271652e+00
4.25015986e-01 -1.01104045e+00 1.51352257e-01 1.68557286e-01
2.50197556e-02 -2.98972130e-01 -1.16674578e+00 -1.37963474e+00
1.03025508e+00 -5.09548545e-01 7.58916497e-01 1.72153759e+00
3.05309445e-01 2.84267008e-01 -4.08386365e-02 -3.36328804e-01
-2.16207981e-01 -1.56031936e-01 5.06186008e-01 -1.53994572e+00
1.35068446e-01 -1.01502657e+00 -5.07093370e-01 -7.27520287e-01
7.39772320e-01 -1.17815650e+00 -1.25817016e-01 -1.37522352e+00
9.81764495e-02 -1.43214330e-01 -7.73899853e-01 5.08065462e-01
-1.02128118e-01 -1.88185293e-02 -1.45551965e-01 1.96276337e-01
-9.44920480e-02 5.66688597e-01 7.33914733e-01 -7.28057206e-01
-3.60668987e-01 -2.59693623e-01 -6.28325641e-01 6.09863341e-01
9.36123192e-01 -5.64562798e-01 -1.70066044e-01 -5.95618963e-01
1.46651283e-01 -8.15940797e-01 -1.83733791e-01 -8.93200219e-01
-8.46292451e-02 -4.88227427e-01 5.72918892e-01 -2.94401586e-01
-3.12200915e-02 -7.77579308e-01 -2.96007395e-01 5.37669599e-01
-2.33394518e-01 4.02263403e-01 5.24180114e-01 7.85343349e-01
-2.25509420e-01 -4.66955304e-01 6.18743300e-01 4.48822454e-02
-9.88760293e-01 1.06284052e-01 -9.11557198e-01 -2.03265578e-01
1.04151022e+00 -2.22299904e-01 -1.24039508e-01 -7.02342689e-02
-6.89326227e-01 1.61902070e-01 2.95543134e-01 1.13591802e+00
1.09757982e-01 -1.78885353e+00 -4.37952429e-01 1.16579287e-01
5.51348865e-01 -6.83693469e-01 -1.37386158e-01 4.13613319e-01
-1.08328193e-01 6.45261645e-01 -2.17419580e-01 -4.08932745e-01
-1.11413050e+00 8.41395080e-01 3.08263123e-01 -3.34765583e-01
-4.33416516e-01 7.27702618e-01 -4.42443222e-01 -8.80395472e-01
1.83983043e-01 -7.82592118e-01 -5.61557829e-01 8.11872482e-01
3.86812657e-01 1.60670385e-01 -6.35202378e-02 -4.21898186e-01
-5.08483529e-01 8.88771653e-01 -2.67648995e-01 -1.66327849e-01
1.57842386e+00 -1.16065636e-01 7.17999190e-02 1.36979592e+00
1.39890659e+00 2.62347665e-02 -5.03159523e-01 -4.59065467e-01
5.48696458e-01 -3.38728368e-01 -6.20134510e-02 -5.77558160e-01
-4.97586191e-01 7.87464738e-01 8.64375055e-01 7.99771547e-01
9.81549382e-01 -1.79947808e-01 7.20551133e-01 6.32972002e-01
1.26677766e-01 -1.63133085e+00 4.13992219e-02 9.35023487e-01
9.13391769e-01 -1.18703473e+00 -1.27453998e-01 4.12840754e-01
-2.53182411e-01 1.57911325e+00 4.26840901e-01 1.09705441e-01
1.06099045e+00 -1.72571778e-01 6.60040379e-02 1.62923574e-01
-8.49450946e-01 -2.74302155e-01 2.31862813e-01 1.01068926e+00
6.89792752e-01 1.67921379e-01 -6.71274960e-01 4.86613214e-01
-2.84540027e-01 -1.29591882e-01 5.62694788e-01 1.02851152e+00
-2.25185707e-01 -1.58330715e+00 -2.28392333e-01 3.80524427e-01
-2.81263828e-01 1.87730670e-01 -6.72810078e-01 1.00584543e+00
2.07351908e-01 6.71098888e-01 3.85515988e-01 -4.84358281e-01
4.60282743e-01 8.91933203e-01 4.41832840e-01 -8.36435676e-01
-4.40091372e-01 -5.38529217e-01 -6.52853847e-02 -1.13365091e-01
-6.23756588e-01 -8.99120986e-01 -7.49230385e-01 -3.20823401e-01
-1.28252596e-01 4.39323597e-02 6.65650070e-01 7.59089947e-01
8.36819857e-02 6.67388976e-01 9.33153749e-01 -6.02559626e-01
-3.95098031e-01 -1.54970741e+00 -5.96777439e-01 7.70767272e-01
4.40926850e-01 -6.52925193e-01 -5.34910321e-01 2.61671990e-01] | [10.22545337677002, 8.871541976928711] |
85bc89a9-98d5-43ce-8571-5b62822335f7 | asymmetric-co-teaching-for-unsupervised-cross | 1912.01349 | null | https://arxiv.org/abs/1912.01349v1 | https://arxiv.org/pdf/1912.01349v1.pdf | Asymmetric Co-Teaching for Unsupervised Cross Domain Person Re-Identification | Person re-identification (re-ID), is a challenging task due to the high variance within identity samples and imaging conditions. Although recent advances in deep learning have achieved remarkable accuracy in settled scenes, i.e., source domain, few works can generalize well on the unseen target domain. One popular solution is assigning unlabeled target images with pseudo labels by clustering, and then retraining the model. However, clustering methods tend to introduce noisy labels and discard low confidence samples as outliers, which may hinder the retraining process and thus limit the generalization ability. In this study, we argue that by explicitly adding a sample filtering procedure after the clustering, the mined examples can be much more efficiently used. To this end, we design an asymmetric co-teaching framework, which resists noisy labels by cooperating two models to select data with possibly clean labels for each other. Meanwhile, one of the models receives samples as pure as possible, while the other takes in samples as diverse as possible. This procedure encourages that the selected training samples can be both clean and miscellaneous, and that the two models can promote each other iteratively. Extensive experiments show that the proposed framework can consistently benefit most clustering-based methods, and boost the state-of-the-art adaptation accuracy. Our code is available at https://github.com/FlyingRoastDuck/ACT_AAAI20. | ['Fengxiang Yang', 'Zhun Zhong', 'Xiaowei Guo', 'Shaozi Li', 'Ke Li', 'Hao Cheng', 'Zhiming Luo', 'Xing Sun', 'Rongrong Ji', 'Feiyue Huang'] | 2019-12-03 | null | null | null | null | ['miscellaneous'] | ['miscellaneous'] | [-8.58121272e-03 -2.86172986e-01 -7.59637579e-02 -5.49987257e-01
-4.68881994e-01 -3.57103080e-01 3.85368824e-01 -1.81872621e-01
-6.02211237e-01 8.27651560e-01 2.76957117e-02 2.98835009e-01
2.69073565e-02 -6.04107082e-01 -5.11509836e-01 -1.01417100e+00
3.29829663e-01 6.70095980e-01 -9.53596532e-02 2.79214591e-01
-1.43802762e-01 1.75062284e-01 -1.63970602e+00 1.16248868e-01
1.25700665e+00 7.37405479e-01 2.32412055e-01 3.09348665e-02
-2.09912378e-02 4.90761071e-01 -6.66571021e-01 -7.33335257e-01
3.63663137e-01 -5.04583955e-01 -5.23476005e-01 2.81754613e-01
3.96860510e-01 -2.54542381e-01 -1.52022958e-01 1.43101323e+00
7.00342059e-01 4.46130812e-01 5.48321307e-01 -1.25182700e+00
-9.55151141e-01 6.71651065e-01 -7.45818138e-01 -4.55601104e-02
-6.91720545e-02 1.90478235e-01 5.69382608e-01 -9.76029277e-01
3.51416856e-01 1.17230070e+00 6.75029814e-01 1.16761434e+00
-1.35151720e+00 -1.20374477e+00 4.77843493e-01 3.37515295e-01
-1.68696141e+00 -5.73990762e-01 8.53211999e-01 -2.68319547e-01
-9.91806984e-02 2.82490045e-01 4.72008020e-01 1.58678281e+00
-5.98894954e-01 9.61433649e-01 1.12311649e+00 -1.93182215e-01
2.71160901e-01 5.72672129e-01 2.15268567e-01 1.93615943e-01
4.36467350e-01 3.77064990e-03 -3.50777477e-01 1.24610206e-02
4.99873340e-01 3.43330353e-01 -4.64100242e-01 -2.46429816e-01
-1.15010524e+00 5.81898332e-01 4.86965567e-01 2.40012035e-01
-2.84226149e-01 -2.16780141e-01 2.77184576e-01 6.73626438e-02
4.64951575e-01 1.93797275e-01 -2.17163876e-01 1.72013864e-01
-9.20442164e-01 1.48818403e-01 3.72822911e-01 9.25728559e-01
7.94082940e-01 -2.71650344e-01 -2.16668114e-01 1.36344409e+00
1.44277260e-01 4.72182363e-01 6.79384530e-01 -7.99003005e-01
3.10487717e-01 4.55001831e-01 2.53631115e-01 -8.90524626e-01
-1.26079932e-01 -6.92475140e-01 -1.27221107e+00 1.36731848e-01
4.80392784e-01 -2.87389576e-01 -9.12904322e-01 1.93805099e+00
4.45476621e-01 4.99893457e-01 -1.38319075e-01 1.14477062e+00
6.88158393e-01 3.64275336e-01 2.87089318e-01 -2.05426857e-01
1.08880115e+00 -1.07533705e+00 -6.11657679e-01 -3.35407078e-01
3.67700756e-01 -6.24479830e-01 9.83286381e-01 5.27340472e-01
-6.87440574e-01 -9.69937921e-01 -9.28142011e-01 2.20624462e-01
-1.52859256e-01 4.67609763e-01 3.24068069e-01 5.76483786e-01
-7.88876295e-01 5.80545723e-01 -5.62970757e-01 -3.09906125e-01
6.94004714e-01 2.17158005e-01 -3.57419014e-01 -4.84729111e-01
-1.05561984e+00 5.24131954e-01 5.05031526e-01 3.19792658e-01
-7.48612940e-01 -4.59812462e-01 -4.84471560e-01 -1.15977399e-01
4.71036136e-01 -5.73292911e-01 1.01689184e+00 -1.31253099e+00
-1.25434959e+00 9.05863941e-01 -3.69571716e-01 -1.49077028e-01
7.89690971e-01 -3.15739781e-01 -7.45529115e-01 -1.15506470e-01
3.92817527e-01 8.22986245e-01 9.22309518e-01 -1.71869493e+00
-8.64305139e-01 -4.23227757e-01 -2.68722028e-01 3.10699075e-01
-5.99336326e-01 -2.20871732e-01 -7.44158626e-01 -7.08084345e-01
1.12221047e-01 -9.56461668e-01 -2.31678441e-01 -3.91633175e-02
-4.48220462e-01 -2.90733367e-01 4.35996264e-01 -6.21575356e-01
1.05286646e+00 -2.46131229e+00 1.77229494e-02 2.95332104e-01
2.90601283e-01 4.54444408e-01 -1.53279826e-01 -1.60774559e-01
-1.57384858e-01 2.54994571e-01 -1.76607177e-01 -7.30598390e-01
-1.82181492e-01 1.06371552e-01 -9.23167691e-02 4.96820360e-01
-3.78915071e-02 5.91054380e-01 -1.03704095e+00 -4.91973400e-01
1.63845673e-01 3.74498099e-01 -3.11828405e-01 2.95892149e-01
7.02123344e-02 7.72512913e-01 -3.78328711e-01 5.99078119e-01
1.04782867e+00 -3.25217962e-01 1.17456906e-01 -2.75533587e-01
1.71709716e-01 -2.01201439e-01 -1.52085197e+00 1.39013565e+00
-1.16204493e-01 2.58387893e-01 2.01766819e-01 -1.11756444e+00
9.53954101e-01 1.88082442e-01 3.48985732e-01 -5.27594090e-01
6.24652617e-02 1.88812077e-01 -4.86101583e-02 -4.93476242e-01
2.25982606e-01 -1.14786536e-01 2.84763604e-01 3.78819197e-01
-8.90249014e-02 5.98434329e-01 2.89833229e-02 1.00288242e-02
3.98733407e-01 8.43347162e-02 -3.44002619e-02 -2.99850143e-02
5.36326051e-01 -1.60327166e-01 1.03241825e+00 8.66542399e-01
-5.34340739e-01 6.96240425e-01 -1.17821902e-01 -4.05514121e-01
-1.02077973e+00 -9.93983209e-01 -1.72700182e-01 1.07744610e+00
5.20354569e-01 -6.16902895e-02 -8.26621294e-01 -7.95779109e-01
-1.60217453e-02 5.97667754e-01 -7.06723750e-01 -3.45055848e-01
-4.78544444e-01 -8.94044876e-01 5.02880514e-01 3.95617753e-01
7.59804785e-01 -9.96351779e-01 1.49018556e-01 6.02543168e-02
-5.64019561e-01 -8.48629355e-01 -5.74793756e-01 -4.34368802e-03
-5.80451429e-01 -1.00540721e+00 -8.78790259e-01 -9.38320160e-01
9.96596098e-01 6.41361535e-01 8.98294091e-01 3.55390340e-01
6.00645877e-02 1.92121007e-02 -6.38582289e-01 -2.55841434e-01
-2.26294085e-01 4.01763916e-02 4.61169362e-01 5.52397847e-01
8.25134754e-01 -4.89001334e-01 -6.54837251e-01 6.35246515e-01
-6.77445412e-01 7.00272471e-02 4.44820702e-01 1.00436532e+00
7.25124419e-01 2.60058165e-01 7.96763241e-01 -9.62901175e-01
4.57228124e-01 -5.67193925e-01 -2.22480565e-01 3.16066384e-01
-6.02707744e-01 -2.77076870e-01 9.13067997e-01 -8.44296873e-01
-1.16043913e+00 3.78119797e-02 -6.32196069e-02 -6.20399415e-01
-6.07867479e-01 1.35498360e-01 -5.98993659e-01 9.69621912e-02
7.10746169e-01 3.01693261e-01 3.65103669e-02 -5.96506476e-01
2.08733976e-01 8.84751916e-01 6.52122974e-01 -6.76069975e-01
9.61728156e-01 6.12938941e-01 -5.89998007e-01 -5.30706763e-01
-8.88415575e-01 -5.85271001e-01 -6.54522419e-01 -3.54472190e-01
7.09199369e-01 -1.04674339e+00 -4.93714124e-01 6.18606627e-01
-7.45087743e-01 -1.77628368e-01 -2.28649542e-01 5.91624737e-01
5.30751422e-04 5.26866734e-01 -4.05165076e-01 -8.94587457e-01
-1.34656250e-01 -1.03616095e+00 7.89657831e-01 7.63136566e-01
-1.19885474e-01 -6.29334867e-01 -2.75614500e-01 4.56346005e-01
1.65471554e-01 -3.48645262e-02 3.35283667e-01 -9.47528005e-01
-4.02229995e-01 -2.55498409e-01 -2.62235999e-01 5.36896050e-01
3.08013171e-01 -2.51468837e-01 -1.26804447e+00 -4.53883231e-01
1.04950503e-01 -4.03965831e-01 9.45170164e-01 2.83590287e-01
1.46788704e+00 -1.78889379e-01 -6.47417843e-01 7.86944091e-01
1.14038646e+00 1.18788876e-01 4.82438773e-01 4.29497629e-01
8.41802001e-01 7.05019355e-01 6.08420789e-01 3.49153638e-01
2.96229213e-01 4.91566360e-01 2.14072570e-01 -1.34733558e-01
-1.71173036e-01 -2.44075105e-01 1.46566257e-01 6.26596928e-01
-1.86343238e-01 -1.69004321e-01 -6.43648326e-01 5.18543422e-01
-1.96722162e+00 -1.15863895e+00 5.79177476e-02 2.38617849e+00
9.61665750e-01 -4.75291237e-02 2.20017761e-01 -2.02612802e-02
1.34016156e+00 -1.57814920e-01 -8.79485786e-01 6.10134423e-01
-2.32792094e-01 -9.07709897e-02 3.63367081e-01 2.62106180e-01
-1.26672101e+00 8.34334135e-01 4.78050756e+00 1.13268638e+00
-9.45922434e-01 1.75227270e-01 9.34737623e-01 -3.16552341e-01
-1.85417049e-02 -2.20197827e-01 -8.94961774e-01 9.95414674e-01
5.16276240e-01 -9.09807310e-02 6.25023603e-01 9.71564472e-01
1.85551703e-01 1.11929543e-01 -1.04840326e+00 1.30424702e+00
9.27209929e-02 -8.96455467e-01 -6.71601389e-04 -3.14383358e-02
6.97455645e-01 -2.27118328e-01 7.60593712e-02 3.43102902e-01
4.36143368e-01 -8.87351811e-01 7.03352511e-01 6.21577263e-01
6.49807930e-01 -7.81001270e-01 7.97125220e-01 5.62406540e-01
-1.07123077e+00 -2.69395024e-01 -6.54468894e-01 2.66187012e-01
2.14195624e-02 7.72919476e-01 -3.46184671e-01 6.39019191e-01
1.09732747e+00 9.01342750e-01 -7.39889026e-01 1.27466631e+00
-2.54253328e-01 6.98825121e-01 -2.57309973e-01 1.46743670e-01
-1.97497472e-01 -5.12221217e-01 3.22911352e-01 1.05010033e+00
2.87548155e-01 8.82322043e-02 5.22930741e-01 8.24816644e-01
-2.29300708e-01 1.04489056e-02 -2.03071699e-01 4.42976505e-01
8.41724217e-01 1.31179631e+00 -6.03989303e-01 -5.55841327e-01
-3.43906522e-01 1.28279829e+00 5.41330278e-01 6.52945161e-01
-9.84143257e-01 -1.42722517e-01 6.19897962e-01 -1.15326438e-02
1.49708778e-01 1.41646415e-01 -2.67882973e-01 -1.44584394e+00
2.87825406e-01 -9.49002147e-01 6.10307395e-01 -5.74769795e-01
-2.04809785e+00 5.35717428e-01 -2.79584110e-01 -1.54014874e+00
2.84694850e-01 -3.19988012e-01 -4.25112158e-01 9.86806273e-01
-1.37915969e+00 -1.08392262e+00 -6.60294890e-01 7.14709818e-01
4.62029189e-01 -2.32095867e-01 5.58467269e-01 7.40195930e-01
-1.00025582e+00 9.59657669e-01 4.05497253e-01 4.34035003e-01
1.18953335e+00 -1.04200089e+00 2.84374673e-02 9.83605742e-01
1.27306534e-02 8.22912872e-01 3.90781850e-01 -7.05428839e-01
-6.75108254e-01 -1.43738413e+00 4.29664105e-01 -2.44828820e-01
2.07084224e-01 -3.99943918e-01 -1.20740533e+00 5.50533116e-01
-1.44397259e-01 3.39555405e-02 7.79806912e-01 2.27574751e-01
-3.08287710e-01 -3.40294302e-01 -1.22472847e+00 8.03073943e-01
1.26893115e+00 -2.33200714e-01 -3.45879823e-01 4.43280667e-01
4.55065429e-01 -1.32961258e-01 -5.75768113e-01 2.74166316e-01
3.19478542e-01 -9.70790207e-01 1.01511371e+00 -4.38126415e-01
7.90123940e-02 -6.31889164e-01 1.47255585e-02 -1.49410176e+00
-8.08646381e-01 -2.30820656e-01 7.07113137e-03 1.73173618e+00
2.81333297e-01 -6.76941693e-01 8.83457780e-01 8.70194316e-01
2.09604520e-02 -3.53079945e-01 -8.00016046e-01 -9.24759209e-01
-2.90567297e-02 -9.93842781e-02 7.24648952e-01 1.30515599e+00
-3.21418136e-01 3.03449094e-01 -7.42436767e-01 3.83334219e-01
8.37780952e-01 -1.98589098e-02 9.23238277e-01 -1.42874849e+00
-1.31905943e-01 -4.71763551e-01 -1.20434090e-02 -1.12872219e+00
1.60222515e-01 -9.09756541e-01 1.63814902e-01 -1.17722237e+00
4.24415201e-01 -9.08156812e-01 -3.47508907e-01 5.00314057e-01
-7.56597161e-01 2.97657996e-01 6.09791577e-02 7.27555275e-01
-6.24814749e-01 6.30041957e-01 1.15500736e+00 -1.83101520e-01
-1.90412149e-01 1.53141633e-01 -1.02621591e+00 6.67148948e-01
9.99655366e-01 -4.53887552e-01 -3.12527776e-01 -5.31381428e-01
-3.05206388e-01 -7.20484495e-01 4.27386105e-01 -1.20081186e+00
3.65780443e-01 -1.66430384e-01 9.65357661e-01 -2.64889479e-01
2.02561870e-01 -9.41841066e-01 3.06023866e-01 2.36255243e-01
-3.74416202e-01 -3.64626706e-01 -1.50929585e-01 6.66080832e-01
-9.28137973e-02 -3.57548147e-01 9.95984912e-01 -2.60946363e-01
-7.55945444e-01 6.46236002e-01 2.56740768e-02 2.45125853e-02
9.40261960e-01 -3.28007072e-01 -2.51067936e-01 -2.14847639e-01
-8.41705739e-01 4.52719033e-01 7.03523993e-01 4.42454398e-01
3.37992013e-01 -1.50423872e+00 -8.85586321e-01 1.52860269e-01
1.30179912e-01 1.65856495e-01 5.82925737e-01 6.60044432e-01
2.32932176e-02 -2.23149419e-01 -1.25529543e-01 -6.71071649e-01
-1.09040344e+00 9.66217458e-01 4.17471975e-01 3.54991369e-02
-5.58100164e-01 9.52195942e-01 3.12902182e-01 -6.03966236e-01
4.18209344e-01 2.93519974e-01 -2.96468943e-01 6.39674067e-02
8.06496143e-01 3.11136156e-01 -2.79321253e-01 -6.18518114e-01
-3.15437198e-01 4.42782223e-01 -4.39099997e-01 1.79564714e-01
1.06558061e+00 -3.97331208e-01 -6.40194640e-02 3.01041514e-01
9.16159272e-01 -3.99051001e-03 -1.27267098e+00 -4.94967788e-01
-1.52359650e-01 -6.67783499e-01 -3.89310032e-01 -7.87402511e-01
-1.19396698e+00 6.43046975e-01 8.09346974e-01 1.75067022e-01
1.22324395e+00 -5.12556732e-02 7.30453253e-01 2.16573402e-01
3.05084586e-01 -1.33066964e+00 1.06712151e-03 7.36380517e-02
4.75310534e-01 -1.53294253e+00 -8.36377069e-02 -4.80513841e-01
-7.90768921e-01 6.38813317e-01 1.12063491e+00 -2.16112006e-02
4.37853277e-01 -1.91523418e-01 2.30265632e-01 2.36901715e-01
-1.21013895e-01 -2.51397252e-01 8.97607505e-02 9.23834443e-01
1.42492771e-01 2.67809927e-01 -8.42639208e-02 9.83838618e-01
-1.33770267e-02 -1.70137000e-03 2.41656810e-01 5.13319612e-01
-3.24037701e-01 -1.13318110e+00 -6.88847125e-01 5.96343219e-01
-3.13177675e-01 -3.79931927e-02 -3.45290989e-01 5.69653332e-01
7.18915343e-01 1.12734497e+00 -1.77883618e-02 -5.26925385e-01
2.70625770e-01 5.45920245e-02 1.09524556e-01 -4.85017151e-01
-4.36733484e-01 4.12838534e-02 -1.02344655e-01 -2.27837831e-01
-5.75997949e-01 -7.79897988e-01 -8.80997121e-01 -3.76344472e-01
-3.85440022e-01 3.41502577e-01 1.41492784e-01 7.75262892e-01
4.60764527e-01 2.53462046e-01 7.09897637e-01 -7.33633399e-01
-5.26861429e-01 -1.00809634e+00 -6.11560822e-01 8.86119843e-01
2.29617164e-01 -8.24380517e-01 -4.43118423e-01 2.76484489e-01] | [14.825742721557617, 1.0947003364562988] |
2ada8d9f-c685-40fe-9a00-49ff90d6341d | emergent-a-novel-data-set-for-stance | null | null | https://aclanthology.org/N16-1138 | https://aclanthology.org/N16-1138.pdf | Emergent: a novel data-set for stance classification | null | ['William Ferreira', 'Andreas Vlachos'] | 2016-06-01 | null | null | null | naacl-2016-6 | ['rumour-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.380704879760742, 3.715984344482422] |
089f4fda-295e-4eac-82da-e9ad2cb857d5 | predicting-future-shanghai-stock-market-price | 1609.05394 | null | http://arxiv.org/abs/1609.05394v1 | http://arxiv.org/pdf/1609.05394v1.pdf | Predicting Future Shanghai Stock Market Price using ANN in the Period 21-Sep-2016 to 11-Oct-2016 | Predicting the prices of stocks at any stock market remains a quest for many
investors and researchers. Those who trade at the stock market tend to use
technical, fundamental or time series analysis in their predictions. These
methods usually guide on trends and not the exact likely prices. It is for this
reason that Artificial Intelligence systems, such as Artificial Neural Network,
that is feedforward multi-layer perceptron with error backpropagation, can be
used for such predictions. A difficulty in neural network application is the
determination of suitable network parameters. A previous research by the author
already determined the network parameters as 5:21:21:1 with 80% training data
or 4-year of training data as a good enough model for stock prediction. This
model has been put to the test in predicting selected Shanghai Stock Exchange
stocks in the future period of 21-Sep-2016 to 11-Oct-2016, about one week after
the publication of these predictions. The research aims at confirming that
simple neural network systems can be quite powerful in typical stock market
predictions. | ['Barack Wamkaya Wanjawa'] | 2016-09-17 | null | null | null | null | ['stock-prediction'] | ['time-series'] | [-8.28666031e-01 -1.97872102e-01 -3.52864116e-01 -2.64197350e-01
1.38049975e-01 -4.74032789e-01 4.82748657e-01 1.27577305e-01
-4.73658592e-01 1.03714907e+00 9.65775996e-02 -8.61545801e-01
-1.09654471e-01 -9.65600491e-01 -3.14082623e-01 -2.80641407e-01
-1.52145445e-01 3.56928855e-01 2.61971891e-01 -6.52319252e-01
8.47848833e-01 6.08996928e-01 -1.39902902e+00 7.11808056e-02
2.69311428e-01 1.33903182e+00 6.20298758e-02 4.32350904e-01
-5.26060879e-01 1.20840597e+00 -4.92611259e-01 -5.49499273e-01
8.25302780e-01 -4.02773805e-02 -2.16220453e-01 -5.02472520e-01
-3.73810440e-01 -6.11046076e-01 -1.90463766e-01 1.10754716e+00
9.08269063e-02 -2.52911270e-01 4.00262624e-01 -9.45042193e-01
-7.72627473e-01 1.17708433e+00 -2.22447902e-01 6.57694757e-01
-2.46202841e-01 -1.14523899e-03 1.15199137e+00 -7.94376433e-01
1.69275180e-01 5.92830300e-01 8.67677987e-01 8.72736722e-02
-7.35771239e-01 -9.39907849e-01 3.29410397e-02 2.59161353e-01
-9.38716173e-01 6.82147071e-02 1.02199972e+00 -7.41677880e-01
9.43812668e-01 3.07927221e-01 1.18474853e+00 3.23580831e-01
6.27807617e-01 4.15683925e-01 1.35737753e+00 -3.62353384e-01
3.51791680e-01 6.88613176e-01 2.30339125e-01 -1.46137133e-01
7.26792574e-01 3.02772164e-01 -2.09490657e-01 -3.78207117e-02
8.63423645e-01 2.26472348e-01 -1.87095162e-02 6.41807437e-01
-8.57916832e-01 1.11837614e+00 3.95769507e-01 8.77333164e-01
-9.09215689e-01 -1.91431358e-01 2.17288136e-01 7.31504858e-01
6.64932370e-01 7.79646635e-01 -9.69313979e-01 -1.83468789e-01
-1.29736567e+00 6.14213288e-01 1.19451034e+00 3.60757172e-01
3.15552562e-01 6.30757153e-01 4.59578842e-01 3.26429576e-01
4.09256160e-01 3.02359700e-01 1.16151118e+00 -4.87734467e-01
2.32997641e-01 7.38864660e-01 3.51964444e-01 -1.00784552e+00
-5.54615378e-01 -6.39757752e-01 -6.58969402e-01 7.37005115e-01
5.85604012e-01 -6.77661180e-01 -6.50708318e-01 9.10456121e-01
-3.05801570e-01 1.85249612e-01 3.24518412e-01 6.46119118e-01
3.34238470e-01 1.12518561e+00 -2.12876141e-01 -4.46560472e-01
8.49488318e-01 -6.92915976e-01 -6.22968793e-01 9.92105994e-03
2.38052219e-01 -8.81663918e-01 2.84260958e-01 4.20226097e-01
-1.05874896e+00 -5.37641346e-01 -9.63908374e-01 7.45634735e-01
-8.45903397e-01 7.79748932e-02 5.46103835e-01 4.65954542e-01
-9.21615362e-01 1.11457467e+00 -5.01025021e-01 2.59950459e-01
-4.57499981e-01 3.67957026e-01 2.16588870e-01 1.04410934e+00
-1.49588215e+00 1.46636784e+00 1.01440215e+00 4.15828139e-01
1.25466138e-01 -7.54910886e-01 -1.93809882e-01 1.77428588e-01
-3.63557547e-01 5.77193685e-02 1.33379233e+00 -1.16083562e+00
-1.48196399e+00 2.88522005e-01 4.47288543e-01 -1.01836061e+00
5.34938276e-01 1.51384575e-02 -9.47717309e-01 -3.22196960e-01
-2.76906312e-01 2.44893283e-01 3.34999144e-01 -5.86342871e-01
-1.03537214e+00 -5.45895882e-02 -1.05353773e-01 -1.50157496e-01
-1.69092864e-01 7.01069012e-02 4.87236530e-01 -9.66632426e-01
3.56105030e-01 -8.09838772e-01 -2.99087793e-01 -5.86866081e-01
-8.74504521e-02 -2.73089975e-01 5.08224666e-01 -1.10376942e+00
1.23973215e+00 -1.54625142e+00 -8.26839864e-01 5.60389817e-01
-2.85543084e-01 3.96731377e-01 5.70724428e-01 5.80304980e-01
-5.69142759e-01 1.19739823e-01 2.38479868e-01 3.56305897e-01
1.41693085e-01 -9.38568413e-02 -8.22012901e-01 7.99572766e-02
1.15791596e-01 7.64232278e-01 -2.51575947e-01 6.64780512e-02
1.40786245e-01 1.73892751e-02 -1.21364854e-01 6.19235449e-02
-1.96598276e-01 -5.66647276e-02 -4.48192477e-01 4.78814960e-01
6.35376334e-01 -1.78080752e-01 -1.57621890e-01 7.16458249e-04
-8.85282338e-01 3.23976308e-01 -1.19122791e+00 3.97164643e-01
5.25243804e-02 9.12207007e-01 -5.91248035e-01 -9.42044735e-01
1.52200854e+00 6.76422834e-01 5.56969047e-01 -6.68278813e-01
1.40589073e-01 9.31235552e-01 3.69050354e-01 -2.76124567e-01
6.63371980e-01 -5.19285321e-01 4.17268544e-01 4.76107299e-01
-3.62609506e-01 1.75881803e-01 3.65708500e-01 -5.69067419e-01
4.13446605e-01 3.04828621e-02 8.79285708e-02 -4.40208495e-01
3.69700164e-01 1.97513968e-01 6.50659502e-01 1.19012237e-01
-3.65480967e-02 1.99305221e-01 5.02831459e-01 -1.04726887e+00
-1.37634575e+00 -4.72169608e-01 -3.26760560e-01 4.95870978e-01
-4.85274345e-01 3.68111849e-01 -3.68192405e-01 1.71192527e-01
2.10658714e-01 1.00315177e+00 -4.09337878e-01 2.39227667e-01
-3.10891867e-01 -9.18603599e-01 1.74068242e-01 4.32521015e-01
5.07934690e-01 -1.35722840e+00 -5.92417777e-01 7.19756484e-01
6.27946258e-01 -5.69389880e-01 2.43635818e-01 3.52998912e-01
-1.18681264e+00 -8.07151854e-01 -1.21408284e+00 -5.38660228e-01
3.88417572e-01 -4.08337921e-01 8.95007670e-01 3.08061000e-02
4.36645299e-01 -4.33404565e-01 -1.21601410e-01 -1.12170529e+00
-3.86616558e-01 1.81753874e-01 1.50276527e-01 -8.02799910e-02
6.80915475e-01 -7.31266379e-01 -4.48891848e-01 -5.56272939e-02
-5.15842855e-01 -5.36316894e-02 7.02142656e-01 4.67705131e-01
6.76884800e-02 3.98167372e-01 1.03699994e+00 -5.67806482e-01
8.95835042e-01 -7.16075778e-01 -1.35936284e+00 2.71800369e-01
-1.33296752e+00 2.27480471e-01 5.80249310e-01 -2.66708374e-01
-7.98966944e-01 -1.91420212e-01 -2.48208866e-01 -1.83439225e-01
2.69730508e-01 1.01222479e+00 7.02110052e-01 -3.39699052e-02
3.51064295e-01 5.42235553e-01 8.83858502e-02 -6.73080802e-01
-4.76031691e-01 5.88148952e-01 2.86570132e-01 8.82107541e-02
9.27440286e-01 -1.67068228e-01 -1.65449575e-01 -5.06202161e-01
-4.53088641e-01 -2.83132285e-01 -6.36920154e-01 -3.73272240e-01
4.39676553e-01 -9.76381600e-01 -5.77608585e-01 8.77603829e-01
-1.01145959e+00 -6.86729848e-02 -2.20481977e-01 9.78499353e-01
-8.64522606e-02 -3.22638810e-01 -9.30364132e-01 -1.35894704e+00
-5.50380826e-01 -7.45832741e-01 -6.92374930e-02 5.26206911e-01
-2.30136365e-01 -1.40584779e+00 2.48000309e-01 -1.85551062e-01
8.23243678e-01 1.55321807e-01 6.18504405e-01 -1.16454637e+00
-5.08308530e-01 -6.04759097e-01 -4.66372371e-02 6.47565961e-01
7.96034560e-02 2.93752849e-01 -7.26087630e-01 8.27843621e-02
2.43173212e-01 1.27435595e-01 5.76906443e-01 6.04404449e-01
2.58399934e-01 -5.73633432e-01 2.30360776e-01 1.98030308e-01
1.72929573e+00 9.75306630e-01 5.92706144e-01 9.38505769e-01
-4.44472954e-02 4.50562865e-01 3.77554804e-01 6.13525271e-01
2.67350078e-01 -3.71116214e-02 1.63696274e-01 8.32747146e-02
6.91082120e-01 -1.50219440e-01 3.59499663e-01 1.10280776e+00
-4.51360524e-01 3.57665062e-01 -8.20050061e-01 3.86673629e-01
-1.61638451e+00 -1.10314000e+00 -1.53352335e-01 1.90207231e+00
7.44702339e-01 9.28335547e-01 3.11986536e-01 6.80032551e-01
5.65878332e-01 8.06302205e-02 -4.94806975e-01 -5.82369328e-01
-1.07532836e-01 -3.95312645e-02 1.16399980e+00 1.91121370e-01
-8.70343089e-01 3.82250369e-01 6.80909157e+00 3.10185224e-01
-1.73582864e+00 -5.40079415e-01 9.51691687e-01 1.69765562e-01
-3.62775296e-01 -1.92323942e-02 -1.07720804e+00 1.01474118e+00
1.32478464e+00 -6.36796057e-01 1.69693157e-01 1.05671227e+00
4.55201775e-01 -8.62535760e-02 -4.36796159e-01 6.56392753e-01
-4.93848532e-01 -1.81764472e+00 9.75540653e-02 2.99348086e-02
7.19974995e-01 -6.49335608e-03 2.12803856e-03 2.14518026e-01
5.94401695e-02 -7.77689517e-01 9.97550130e-01 1.06618023e+00
-2.85907179e-01 -8.63441825e-01 1.30342126e+00 5.13077021e-01
-1.00106823e+00 -2.60808080e-01 -6.14147544e-01 -8.35903108e-01
1.24310762e-01 6.30084455e-01 -8.11546922e-01 2.11236015e-01
5.84674776e-01 5.49594164e-01 -1.99818894e-01 1.27384639e+00
6.14828095e-02 8.04530621e-01 -4.86475497e-01 -7.65568197e-01
5.47770798e-01 -5.16316712e-01 9.90384296e-02 7.61247218e-01
7.63424635e-01 2.72919685e-01 -3.81463319e-01 8.18927586e-01
4.25351083e-01 2.18333155e-01 -4.58229810e-01 -1.64767921e-01
3.57910216e-01 7.19769955e-01 -6.90783441e-01 -2.61642188e-01
-5.90017617e-01 1.77372828e-01 -2.03876778e-01 2.91000187e-01
-3.74685556e-01 -4.16790128e-01 4.45290953e-01 4.25323278e-01
2.98421890e-01 -1.91092864e-01 -6.47149026e-01 -8.71872663e-01
7.89209828e-02 -3.78643036e-01 5.41200675e-02 -8.38489294e-01
-1.24094677e+00 6.37815058e-01 -6.37038378e-03 -1.41103268e+00
-7.04112470e-01 -1.21551752e+00 -1.09049571e+00 1.20771778e+00
-1.62204468e+00 -5.34917831e-01 5.70271552e-01 1.76615849e-01
2.91899234e-01 -9.96723354e-01 4.95407164e-01 1.87066704e-01
-4.53123122e-01 -4.61905412e-02 6.29623711e-01 4.73401040e-01
-6.39260635e-02 -1.38666117e+00 5.69345057e-01 7.20474303e-01
-1.15281902e-01 2.06420973e-01 9.67905700e-01 -7.93743551e-01
-7.73056746e-01 -6.69619083e-01 1.43096244e+00 1.77959740e-01
1.31636488e+00 3.89101058e-01 -9.36843514e-01 8.22929680e-01
3.87834281e-01 -4.21485543e-01 3.83608133e-01 -3.44451100e-01
2.84671754e-01 -3.48756880e-01 -9.54603612e-01 4.61193860e-01
-2.14066148e-01 -1.58326596e-01 -1.02181661e+00 -4.72073928e-02
2.59194463e-01 -1.45630360e-01 -1.14431620e+00 9.51777399e-02
6.80615962e-01 -1.17971849e+00 6.30654275e-01 -4.78370249e-01
2.39640325e-01 1.31206668e-03 2.02599689e-01 -1.45021403e+00
-3.85395676e-01 -5.34484506e-01 2.67306924e-01 9.32031751e-01
1.03157258e+00 -1.21101737e+00 1.09866083e+00 1.24254417e+00
2.14301422e-01 -6.67009652e-01 -7.82371104e-01 -7.59648442e-01
3.52541536e-01 -2.96803206e-01 9.48839426e-01 9.97414470e-01
1.05077244e-01 -1.19347341e-01 -4.14881438e-01 -1.24406330e-01
3.90636563e-01 2.06270382e-01 2.84527451e-01 -1.57127929e+00
-1.00522757e-01 -9.58491504e-01 -4.23459768e-01 -4.65541959e-01
-4.92971651e-02 -3.18507254e-01 -5.68671167e-01 -1.21074116e+00
-5.82379282e-01 -4.17048246e-01 -6.52387977e-01 3.08998883e-01
1.73704162e-01 -8.71821493e-02 3.93430293e-01 5.99897325e-01
6.01764917e-01 1.24484487e-01 1.05922091e+00 4.25572619e-02
-1.92405954e-01 6.79964840e-01 -4.50473458e-01 8.29830587e-01
1.15385294e+00 -1.68441623e-01 -2.45536909e-01 1.41800091e-01
6.17695570e-01 2.01567903e-01 -9.40807238e-02 -1.12562549e+00
3.89627695e-01 -4.27089214e-01 6.87444091e-01 -9.91793871e-01
1.74892426e-01 -1.01922858e+00 6.83761477e-01 9.20630574e-01
-2.50343472e-01 7.44922519e-01 2.04485580e-01 -9.15908366e-02
-7.19686866e-01 -7.87136197e-01 3.47048134e-01 -4.21152800e-01
-9.73538578e-01 9.85289440e-02 -4.15378422e-01 -4.63864237e-01
1.14751041e+00 -5.54232717e-01 -2.93893591e-02 -5.57500184e-01
-6.22081697e-01 1.46530256e-01 -8.33418295e-02 2.13888079e-01
3.45160395e-01 -1.38659608e+00 -8.88698757e-01 1.47060022e-01
-5.31295359e-01 -6.63730145e-01 -2.26159677e-01 5.05198121e-01
-1.19885528e+00 8.80127847e-01 -6.38419986e-01 1.06247135e-01
-5.73487639e-01 3.64664972e-01 6.16546631e-01 -4.20781553e-01
-5.07820547e-01 4.67377603e-01 -5.39366245e-01 1.73822314e-01
2.39055604e-03 -7.55529165e-01 -6.51621640e-01 4.52531904e-01
7.27529705e-01 3.05242181e-01 7.67024755e-02 -4.53968376e-01
7.74730965e-02 5.90064108e-01 3.15241627e-02 -2.47035578e-01
1.82793140e+00 2.74223447e-01 -2.80359872e-02 1.10566175e+00
9.95546341e-01 -2.99162567e-01 -1.09877515e+00 3.60461771e-02
7.07489133e-01 -1.07695989e-01 1.94607422e-01 -7.73768604e-01
-1.35484362e+00 4.28380787e-01 5.52016497e-01 9.93519485e-01
9.34453547e-01 -6.18234456e-01 9.23421919e-01 5.43260276e-01
2.03550592e-01 -1.42475343e+00 -7.66145408e-01 6.07496083e-01
1.03738296e+00 -1.20017231e+00 4.04765410e-03 2.66293436e-01
-3.50670636e-01 1.74435723e+00 9.32631344e-02 -6.06147826e-01
1.40288246e+00 3.59802604e-01 3.98830920e-01 1.67360246e-01
-9.83062744e-01 1.24281764e-01 2.07281664e-01 -2.71897037e-02
5.12695372e-01 1.82949618e-01 -6.39442503e-01 7.55725920e-01
-7.29742825e-01 4.75618690e-01 6.79576397e-01 9.39385593e-01
-7.80265868e-01 -8.72702539e-01 -5.53536773e-01 7.52311826e-01
-9.85374093e-01 -1.92041934e-01 -1.41821668e-01 1.07659531e+00
-1.92729652e-01 4.73693907e-01 6.21205509e-01 -3.88166994e-01
3.14580888e-01 2.17785135e-01 -3.29898804e-01 3.25433202e-02
-9.09101307e-01 -1.15633398e-01 -6.41181273e-03 1.79131612e-01
-3.36235464e-01 -7.65292883e-01 -1.20244193e+00 -7.54274845e-01
-2.48451218e-01 6.66951180e-01 8.65840793e-01 9.12443757e-01
-2.56069303e-01 1.44778579e-01 1.01319885e+00 -7.70924747e-01
-1.02811444e+00 -1.30507612e+00 -1.27342224e+00 -2.35309117e-02
3.80317122e-01 -3.03139776e-01 -6.15706265e-01 9.77848545e-02] | [4.506277084350586, 4.207151412963867] |
b3e58a20-9c1e-4474-9cfb-a63a0918dd1d | semi-supervised-graph-embedding-approach-to | 1610.04351 | null | http://arxiv.org/abs/1610.04351v1 | http://arxiv.org/pdf/1610.04351v1.pdf | Semi-supervised Graph Embedding Approach to Dynamic Link Prediction | We propose a simple discrete time semi-supervised graph embedding approach to
link prediction in dynamic networks. The learned embedding reflects information
from both the temporal and cross-sectional network structures, which is
performed by defining the loss function as a weighted sum of the supervised
loss from past dynamics and the unsupervised loss of predicting the
neighborhood context in the current network. Our model is also capable of
learning different embeddings for both formation and dissolution dynamics.
These key aspects contributes to the predictive performance of our model and we
provide experiments with three real--world dynamic networks showing that our
method is comparable to state of the art methods in link formation prediction
and outperforms state of the art baseline methods in link dissolution
prediction. | ['Ryohei Hisano'] | 2016-10-14 | null | null | null | null | ['dynamic-link-prediction'] | ['graphs'] | [-3.36149037e-01 4.98435259e-01 -9.26980317e-01 -7.47012943e-02
2.41298839e-01 -4.35164809e-01 9.26780999e-01 4.70616549e-01
1.09851979e-01 8.56382728e-01 4.09312189e-01 -1.52636573e-01
-8.80290866e-01 -1.42830133e+00 -5.62306821e-01 -4.83898997e-01
-1.24561894e+00 1.07188559e+00 5.95548511e-01 -4.52698588e-01
-1.77302420e-01 6.69858277e-01 -1.04243457e+00 -1.27558231e-01
6.82790518e-01 7.42199063e-01 -3.32617939e-01 6.30398750e-01
6.46968000e-03 9.57296193e-01 -7.51677295e-03 -7.45672643e-01
2.02694684e-01 -7.28928447e-02 -9.37177777e-01 -1.21356092e-01
2.95147717e-01 -1.45641595e-01 -1.29667974e+00 6.53279841e-01
2.75125086e-01 7.20883720e-03 9.50407505e-01 -1.40607512e+00
-5.74968755e-01 8.73502970e-01 -9.50925350e-02 7.85147071e-01
4.23856169e-01 -1.35310352e-01 1.86297429e+00 -5.57532728e-01
1.40372765e+00 1.39785337e+00 9.14767265e-01 1.42249063e-01
-1.93181670e+00 -3.29094112e-01 3.25531632e-01 5.07883132e-01
-1.04529154e+00 -2.62921214e-01 1.04055405e+00 -7.81735718e-01
1.04343879e+00 -2.07103819e-01 9.72688973e-01 1.15307236e+00
5.44723928e-01 4.33697283e-01 5.53315938e-01 6.22563437e-02
-2.25591376e-01 -1.41101331e-02 1.08044371e-01 8.02939832e-01
1.32099181e-01 5.14500380e-01 -4.24037367e-01 -3.85949284e-01
8.73975456e-01 -8.65556151e-02 -1.67797148e-01 -8.35293770e-01
-9.49179053e-01 9.12413180e-01 7.80851066e-01 1.38391718e-01
-1.33378401e-01 2.30500549e-01 3.96418035e-01 8.64310861e-01
9.36841965e-01 6.53827935e-02 -4.05059248e-01 1.72808375e-02
-7.54496694e-01 2.56038457e-01 1.30077815e+00 6.20731831e-01
6.53464496e-01 -2.57226169e-01 8.84746090e-02 6.53452992e-01
5.77881634e-01 1.06078684e-02 2.00071204e-02 -7.58120537e-01
5.96253097e-01 7.99738586e-01 -2.53253371e-01 -1.34799588e+00
-3.53994608e-01 -4.75612044e-01 -6.80646896e-01 1.44082576e-01
3.10225934e-01 1.09533489e-01 -2.86362141e-01 1.66682708e+00
2.99701571e-01 5.86442709e-01 -2.00402930e-01 2.65617311e-01
7.33321607e-01 6.38029039e-01 -1.74867600e-01 -3.70085776e-01
3.17586333e-01 -1.12323534e+00 -7.92812407e-01 2.56853521e-01
4.93078917e-01 -2.26995721e-01 1.97048157e-01 -3.15720230e-01
-1.22618067e+00 -3.33387226e-01 -1.13352311e+00 1.70803547e-01
-5.76482117e-01 -6.26078904e-01 8.97767246e-01 1.57587767e-01
-1.25397062e+00 1.53187788e+00 -7.82250524e-01 -6.53905869e-01
3.87811840e-01 5.04304767e-01 -4.82691050e-01 1.19806468e-01
-1.53206813e+00 8.05710196e-01 2.72361934e-01 -1.72243983e-01
-8.79219294e-01 -1.00257123e+00 -8.40738177e-01 2.73585618e-01
1.96842417e-01 -7.39110708e-01 4.00551498e-01 -3.81002277e-01
-1.23429251e+00 6.89151585e-01 7.26147415e-03 -8.89704287e-01
7.14293540e-01 2.14437857e-01 -7.32641160e-01 2.40643352e-01
2.68091932e-02 2.78762132e-01 6.22548759e-01 -1.17265081e+00
-2.21200138e-01 -1.04378462e-01 1.79112226e-01 -5.47133572e-03
-5.84468126e-01 -6.25007927e-01 -3.88129622e-01 -6.35381341e-01
6.77546998e-03 -8.54884505e-01 -1.64480507e-01 5.12587368e-01
-3.97430032e-01 -4.06477839e-01 1.06604421e+00 -5.43543518e-01
1.64908719e+00 -1.67732215e+00 6.07169032e-01 6.33905172e-01
6.18990541e-01 7.73260044e-03 -2.97554255e-01 1.11721504e+00
-3.70137036e-01 3.90169680e-01 9.89066213e-02 -3.77643019e-01
-6.80153444e-02 4.14090425e-01 -3.34481388e-01 3.79916370e-01
1.47713915e-01 1.16484523e+00 -1.31747842e+00 -6.80847347e-01
1.21290497e-01 4.62643057e-01 -3.67005527e-01 2.43616357e-01
-2.80370414e-01 2.19211027e-01 -3.72455657e-01 4.51694995e-01
1.45369187e-01 -4.73774135e-01 8.34940016e-01 -1.56273752e-01
3.42025787e-01 5.51702678e-01 -1.08185303e+00 1.30639744e+00
-1.33585762e-02 7.12454557e-01 -5.67308590e-02 -1.19965172e+00
7.88942099e-01 3.71023893e-01 1.02017283e+00 -2.72630453e-01
-3.10428351e-01 -1.21661201e-01 1.91298593e-02 -3.29275161e-01
-1.12453766e-01 3.09577227e-01 4.84404474e-01 7.87352622e-01
4.31587130e-01 1.34465843e-01 5.98404229e-01 8.40697944e-01
1.44302630e+00 -4.88049723e-02 1.68745872e-02 -3.16896021e-01
4.41526115e-01 -3.28351408e-01 6.08152270e-01 3.83678883e-01
-3.55937570e-01 -1.31893679e-01 1.14791334e+00 -5.65492392e-01
-1.17994714e+00 -1.64254868e+00 -1.98272645e-01 7.95721054e-01
4.92424183e-02 -8.18652928e-01 -7.73101598e-02 -9.45544600e-01
7.31094241e-01 -1.73010945e-01 -9.80413914e-01 -2.89530754e-01
-5.91400981e-01 -5.25902331e-01 2.50683099e-01 5.20872593e-01
-6.68749064e-02 -7.90571809e-01 9.76321101e-01 6.25533521e-01
1.34650886e-01 -1.01147413e+00 -5.46674728e-01 -5.70543893e-02
-1.39025724e+00 -1.53397000e+00 1.20626107e-01 -1.01618361e+00
4.09017980e-01 -1.18972294e-01 1.52886713e+00 5.22872031e-01
-3.69769871e-01 5.38364112e-01 -6.82120100e-02 3.90700936e-01
-4.21845436e-01 3.20378572e-01 3.88145715e-01 -5.39929084e-02
-9.01202336e-02 -1.41903293e+00 -5.57278395e-01 2.20853582e-01
-4.12578702e-01 -4.71715391e-01 1.21884994e-01 9.74009335e-01
4.62723494e-01 4.17217821e-01 6.11391485e-01 -1.01122844e+00
9.55553889e-01 -8.58772814e-01 -4.09913003e-01 2.18521446e-01
-1.24895525e+00 1.21138528e-01 4.25786227e-01 -4.98161703e-01
-5.44051707e-01 -3.15304637e-01 3.35715145e-01 -1.65835291e-01
6.15243673e-01 6.43649518e-01 2.09069595e-01 -2.78787941e-01
3.53653401e-01 -7.42738545e-02 4.38205332e-01 -4.15931851e-01
5.57509422e-01 4.10673805e-02 1.71437740e-01 -4.39118862e-01
1.30153632e+00 6.12544715e-01 3.90703380e-01 -5.09651244e-01
-5.84854424e-01 -3.15157652e-01 -1.17512512e+00 -3.94755542e-01
3.26342762e-01 -7.68458009e-01 -6.83122337e-01 -2.85573117e-02
-8.10289264e-01 -4.29145634e-01 -4.04807091e-01 1.82652980e-01
-5.33094883e-01 4.11785126e-01 -1.21417522e+00 -5.15945196e-01
-2.83993930e-01 -5.79094172e-01 2.42542550e-01 -2.60844201e-01
-1.69467688e-01 -2.05386496e+00 6.88357711e-01 1.06336750e-01
3.65465134e-01 5.92110753e-01 1.15878844e+00 -4.47099149e-01
-7.75952518e-01 -2.36440748e-01 -1.88745335e-01 5.56502938e-02
1.59897402e-01 4.42829698e-01 -3.11895400e-01 -4.27280426e-01
-8.77157629e-01 -1.59576178e-01 1.25919521e+00 2.81537622e-01
6.45383835e-01 -3.61026973e-01 -1.00628889e+00 5.41151106e-01
1.49656832e+00 -3.12072933e-01 6.54944539e-01 2.54965574e-01
7.40496874e-01 6.08559072e-01 3.32359135e-01 3.27699095e-01
6.35264277e-01 7.87588894e-01 7.38004267e-01 1.40395150e-01
-3.40205967e-01 -7.38527954e-01 3.67789567e-01 9.64081228e-01
-2.39118218e-01 -2.41053671e-01 -7.83178449e-01 8.99921894e-01
-2.24615335e+00 -1.58092177e+00 -2.21663222e-01 1.73645854e+00
9.38015282e-01 6.16591156e-01 4.16424185e-01 2.25642249e-01
6.47126317e-01 6.58610463e-01 -5.29995799e-01 -2.69468784e-01
-1.24400347e-01 5.25500961e-02 3.90977710e-01 9.62872386e-01
-1.27987206e+00 9.69745040e-01 8.01738548e+00 5.92265725e-01
-5.79499662e-01 -6.86041359e-03 2.27290496e-01 -8.31752494e-02
-4.26584065e-01 2.54963636e-01 -6.00895405e-01 4.24372494e-01
9.64275956e-01 -5.16911626e-01 6.46164477e-01 5.19043148e-01
3.07093501e-01 6.04054630e-01 -1.37720537e+00 3.11493933e-01
-2.55942404e-01 -1.89262462e+00 1.78328380e-01 4.53533292e-01
9.26654637e-01 2.26016581e-01 1.27440225e-02 1.18975788e-01
6.48858130e-01 -1.01373374e+00 1.21577732e-01 7.85406291e-01
4.02457982e-01 -6.04657829e-01 5.12714148e-01 -7.40057509e-03
-1.69950223e+00 -2.63443440e-01 -1.96411237e-01 -2.61308849e-01
4.73332167e-01 9.01503861e-01 -7.90025353e-01 6.11531436e-01
4.59930956e-01 1.74583554e+00 -5.14350414e-01 1.00580668e+00
-3.43029857e-01 5.56759715e-01 -2.67580152e-01 2.17151463e-01
-3.07239406e-02 -6.39065146e-01 9.25878823e-01 8.95141363e-01
-3.07176560e-01 -2.86916971e-01 3.66938263e-01 7.34392166e-01
-4.01520908e-01 -1.66465610e-01 -9.03504431e-01 -4.24603045e-01
6.98140323e-01 1.03081739e+00 -3.32521945e-01 -3.31188627e-02
-3.14294308e-01 6.61933482e-01 9.28435266e-01 3.50143433e-01
-5.19593954e-01 -2.14646742e-01 1.13471913e+00 7.29612231e-01
4.10290331e-01 -4.42901999e-01 9.84475240e-02 -9.60912526e-01
-1.15526527e-01 -4.34977226e-02 6.88135982e-01 -6.86488748e-02
-1.79242277e+00 3.31078440e-01 -1.66733399e-01 -1.16441727e+00
-1.48430705e-01 -5.98328173e-01 -1.13700199e+00 4.59013551e-01
-1.67497420e+00 -1.22189462e+00 2.36713499e-01 4.28875178e-01
-4.97161336e-02 -3.41974407e-01 8.43141556e-01 5.13351858e-01
-7.27655947e-01 5.44376552e-01 4.42031294e-01 3.73261631e-01
5.29007375e-01 -1.52694786e+00 4.56347078e-01 4.24702168e-01
2.69636571e-01 4.77441162e-01 6.05066359e-01 -8.68391514e-01
-1.18411589e+00 -1.19765782e+00 1.20953166e+00 -4.71634090e-01
1.59265399e+00 -3.70119423e-01 -6.80554926e-01 1.04867506e+00
-1.07347824e-01 4.00414914e-01 5.87305427e-01 5.66753745e-01
-4.54496562e-01 -3.65656525e-01 -1.04815805e+00 5.93540788e-01
1.66002560e+00 -8.05998623e-01 -2.56389469e-01 6.78613424e-01
7.76119411e-01 9.61134881e-02 -1.65641630e+00 5.00639081e-01
6.59296811e-01 -6.78160965e-01 1.51907122e+00 -8.92791688e-01
6.04557872e-01 1.40321255e-01 3.31992686e-01 -1.32415617e+00
-8.47243488e-01 -6.85808301e-01 -1.10582674e+00 1.12395179e+00
6.41366243e-01 -7.48957217e-01 1.15738642e+00 6.80981996e-03
4.36664075e-01 -1.32597578e+00 -9.69902813e-01 -9.55956399e-01
1.67249963e-01 1.05822146e-01 3.31755042e-01 1.23208201e+00
2.96824694e-01 3.39329123e-01 -4.02876854e-01 1.33738562e-01
8.64772499e-01 5.90411462e-02 5.00001967e-01 -2.02317882e+00
-3.20778579e-01 -5.14467061e-01 -9.64787364e-01 -9.48694885e-01
6.62149072e-01 -1.27067471e+00 -9.44782972e-01 -1.68856120e+00
1.78671032e-01 -6.17585480e-01 -4.64354813e-01 1.87114179e-02
2.05848604e-01 8.01508278e-02 -1.89962342e-01 3.69086921e-01
-5.31469822e-01 8.27333093e-01 1.26630354e+00 -5.70429921e-01
-2.66023785e-01 -4.96916249e-02 -6.85503110e-02 3.86742502e-01
4.85447526e-01 -5.28991282e-01 -5.57495475e-01 -5.65116219e-02
4.48192418e-01 1.70217156e-01 2.19528049e-01 -7.12051630e-01
3.38200897e-01 -1.84383746e-02 1.32445306e-01 -6.10839307e-01
4.37541991e-01 -7.27039814e-01 1.32230341e-01 6.38126969e-01
-2.81766921e-01 -1.30203992e-01 -3.48358899e-01 1.41353989e+00
-8.32922235e-02 2.01821715e-01 4.04940844e-01 1.49236962e-01
-5.35148501e-01 9.98520136e-01 1.10663235e-01 2.54484396e-02
1.02828753e+00 -2.26450250e-01 -4.28669870e-01 -6.61126733e-01
-1.40888739e+00 7.33175635e-01 3.63183409e-01 4.87909675e-01
5.71367145e-01 -1.80012834e+00 -6.23969138e-01 -1.51236102e-01
9.29337461e-03 -6.62731707e-01 -3.61969084e-01 8.34623337e-01
-4.22634512e-01 2.31566519e-01 -5.09009585e-02 -2.22541586e-01
-1.13396978e+00 6.11026168e-01 5.12029469e-01 -9.90011215e-01
-6.89354599e-01 6.42127395e-01 -6.29940987e-01 -6.65085077e-01
3.67643923e-01 2.60039568e-01 -5.10650456e-01 3.40140939e-01
4.48050648e-02 6.54382706e-01 -3.78537774e-01 -6.05607331e-01
-2.69455492e-01 5.46828032e-01 -2.19767198e-01 1.47138789e-01
1.69684029e+00 -1.76837385e-01 -5.06099999e-01 5.79179168e-01
1.48906434e+00 -1.84396252e-01 -1.22876334e+00 -6.32903159e-01
1.23059131e-01 -2.40267277e-01 -1.83913466e-02 -4.40707117e-01
-1.19040120e+00 3.47635180e-01 3.93221587e-01 5.93661904e-01
3.38975936e-01 2.95351833e-01 8.21470618e-01 3.02605510e-01
3.16974074e-01 -1.04250014e+00 4.27695960e-01 6.88514888e-01
7.24862099e-01 -1.20819116e+00 4.60100412e-01 -8.80782723e-01
-1.47631345e-02 1.20640516e+00 3.98614198e-01 -6.55178785e-01
1.58398008e+00 -3.91821489e-02 -4.56084162e-01 -5.27695239e-01
-1.14627302e+00 -1.27454638e-01 3.72326672e-01 7.33132541e-01
3.83954734e-01 -6.81679845e-02 -4.60429192e-01 1.06393658e-01
-2.53339093e-02 -4.73596990e-01 3.10850561e-01 7.19020188e-01
-1.30255938e-01 -1.61889470e+00 4.49963003e-01 8.17138672e-01
-1.53519511e-01 1.22166589e-01 -6.28862560e-01 7.62383342e-01
-1.84398457e-01 7.91633606e-01 1.94711819e-01 -6.09033048e-01
9.42899212e-02 -2.02472066e-03 3.82082909e-01 -6.21594310e-01
-2.60153770e-01 -4.28874344e-01 3.95003051e-01 -7.37823486e-01
-4.13278401e-01 -5.18514335e-01 -8.97151649e-01 -8.68609250e-01
-4.27157074e-01 9.58515704e-03 1.11674415e-02 5.95688283e-01
3.45997036e-01 5.18855512e-01 9.14965868e-01 -9.41358507e-01
-2.43938163e-01 -7.31472731e-01 -9.30154443e-01 7.02324688e-01
3.12584370e-01 -1.01248658e+00 -6.88415945e-01 -1.72819078e-01] | [7.274959087371826, 6.250614166259766] |
53dd6957-94ad-468b-9671-2704c00da9c6 | dialogue-act-recognition-via-crf-attentive | 1711.05568 | null | http://arxiv.org/abs/1711.05568v1 | http://arxiv.org/pdf/1711.05568v1.pdf | Dialogue Act Recognition via CRF-Attentive Structured Network | Dialogue Act Recognition (DAR) is a challenging problem in dialogue
interpretation, which aims to attach semantic labels to utterances and
characterize the speaker's intention. Currently, many existing approaches
formulate the DAR problem ranging from multi-classification to structured
prediction, which suffer from handcrafted feature extensions and attentive
contextual structural dependencies. In this paper, we consider the problem of
DAR from the viewpoint of extending richer Conditional Random Field (CRF)
structural dependencies without abandoning end-to-end training. We incorporate
hierarchical semantic inference with memory mechanism on the utterance
modeling. We then extend structured attention network to the linear-chain
conditional random field layer which takes into account both contextual
utterances and corresponding dialogue acts. The extensive experiments on two
major benchmark datasets Switchboard Dialogue Act (SWDA) and Meeting Recorder
Dialogue Act (MRDA) datasets show that our method achieves better performance
than other state-of-the-art solutions to the problem. It is a remarkable fact
that our method is nearly close to the human annotator's performance on SWDA
within 2% gap. | ['Rongqin Yang', 'Zhou Zhao', 'Zheqian Chen', 'Deng Cai', 'Xiaofei He'] | 2017-11-15 | dialogue-act-recognition-via-crf-attentive-1 | https://dl.acm.org/doi/10.1145/3209978.3209997 | https://dl.acm.org/doi/pdf/10.1145/3209978.3209997 | sigir-2018-7 | ['dialogue-act-classification', 'dialogue-interpretation'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.20396298e-01 7.65046358e-01 -1.81070156e-02 -1.07550657e+00
-7.06288159e-01 -2.68397927e-01 8.60761404e-01 -1.33078814e-01
-4.32327330e-01 1.00225449e+00 8.96081686e-01 -2.08381489e-01
3.87996286e-01 -2.40379900e-01 1.51131619e-02 -4.65160757e-01
2.45924264e-01 1.00238919e+00 1.50283903e-01 -5.32777071e-01
3.06110501e-01 -1.20888025e-01 -9.21244621e-01 5.65488458e-01
8.40584874e-01 9.64698195e-01 1.81468382e-01 8.67032111e-01
-3.84826481e-01 1.57216978e+00 -8.02053690e-01 -5.07398188e-01
-3.87690753e-01 -5.35371244e-01 -1.71128964e+00 4.19651210e-01
-4.59522456e-02 -5.58919966e-01 -1.44375578e-01 7.62954712e-01
4.69779193e-01 3.73976797e-01 6.32477820e-01 -1.06968176e+00
-4.85496938e-01 6.72357023e-01 -1.69245124e-01 8.03277642e-02
7.45182931e-01 2.62749419e-02 1.24625492e+00 -6.64246082e-01
2.55099386e-01 1.84439433e+00 4.05710876e-01 9.30138946e-01
-9.63900864e-01 -3.49573381e-02 4.43661273e-01 1.52975291e-01
-8.57579827e-01 -6.75040007e-01 8.18688512e-01 -3.29444110e-01
1.35979712e+00 2.37084791e-01 -1.25443101e-01 1.28389502e+00
-2.50218421e-01 1.23638177e+00 1.19979775e+00 -6.17291451e-01
2.81987667e-01 2.83152722e-02 7.75202453e-01 7.03934193e-01
-8.41245830e-01 -5.24443448e-01 -3.72113734e-01 -3.65398794e-01
4.66390461e-01 -2.18588233e-01 -1.64583012e-01 4.29068655e-01
-9.70151246e-01 1.05825067e+00 -1.33006826e-01 2.56992847e-01
-1.49233788e-01 -3.53435516e-01 6.68895781e-01 2.45531738e-01
5.59835374e-01 2.46038139e-01 -8.17734480e-01 -3.70935947e-01
-1.16722800e-01 2.43743747e-01 1.38722217e+00 8.80078018e-01
6.82833970e-01 -7.61264712e-02 -4.99219239e-01 1.26388347e+00
5.29943407e-01 1.83873937e-01 5.25786638e-01 -1.00270939e+00
6.75568342e-01 6.24346793e-01 1.27988070e-01 -5.59683383e-01
-5.44327974e-01 2.65415758e-01 -8.52895021e-01 -2.97773749e-01
3.93319041e-01 -6.10305369e-01 -5.08865476e-01 1.61935508e+00
4.22177613e-01 1.54458627e-01 5.20753086e-01 8.39956820e-01
1.16119444e+00 8.41370702e-01 5.56204379e-01 -5.10159552e-01
1.62088060e+00 -1.44757998e+00 -1.26822841e+00 -4.74789053e-01
6.92433298e-01 -5.84802270e-01 9.78532195e-01 2.09100440e-01
-9.56898153e-01 -5.56765139e-01 -7.26358831e-01 -2.83171654e-01
-6.05804771e-02 1.66607186e-01 6.78241551e-01 3.46777767e-01
-7.12447405e-01 1.20727956e-01 -6.21790349e-01 -1.45587206e-01
-3.25833783e-02 4.52872068e-01 -2.82301962e-01 2.03649536e-01
-1.43505156e+00 1.06416607e+00 3.94160628e-01 2.47917637e-01
-8.16134393e-01 5.13522960e-02 -1.08601797e+00 2.95699984e-02
5.52483439e-01 -3.54314655e-01 2.01655698e+00 -1.11500609e+00
-2.36475825e+00 8.90654206e-01 -4.97730225e-01 -5.49161136e-01
4.63793986e-03 -3.97399366e-01 -1.60337538e-01 -4.68181893e-02
-1.27503321e-01 4.70858663e-01 5.57263374e-01 -9.45033967e-01
-6.79780006e-01 -4.39601839e-01 4.10779715e-01 6.93834186e-01
-9.39287320e-02 4.10347402e-01 -8.12117159e-02 -2.91769445e-01
-1.87377259e-01 -8.21176708e-01 -4.35982287e-01 -8.56377125e-01
-5.44732988e-01 -1.07519555e+00 8.28517377e-01 -8.35723042e-01
1.28068805e+00 -1.90796340e+00 4.89102125e-01 -5.61729491e-01
8.80014151e-02 3.49376678e-01 1.59983292e-01 4.80085164e-01
2.10597321e-01 -3.05235773e-01 -4.32764322e-01 -8.60595286e-01
2.27239847e-01 5.36096573e-01 -3.79484385e-01 1.68695718e-01
2.52756774e-01 6.79988265e-01 -7.42644846e-01 -6.21319294e-01
2.27915719e-01 3.22792009e-02 -4.15914297e-01 1.06562507e+00
-7.19434798e-01 7.68897772e-01 -7.46208310e-01 3.86156380e-01
4.23227340e-01 -3.07958007e-01 5.79276681e-01 9.89096984e-02
1.69144064e-01 8.72969449e-01 -7.59429812e-01 1.91566110e+00
-6.34380996e-01 4.10673976e-01 3.60112548e-01 -1.07731903e+00
1.09201646e+00 6.88589334e-01 4.20453623e-02 -4.06981021e-01
4.81260829e-02 -2.69671641e-02 2.95404792e-02 -6.44348919e-01
6.18086934e-01 -3.47513944e-01 -5.60053527e-01 4.99204904e-01
4.39010322e-01 7.89068863e-02 -3.44899356e-01 8.07974339e-02
9.36711669e-01 4.43282537e-02 7.26206839e-01 -1.90023571e-01
1.12382972e+00 -2.47350350e-01 6.35212421e-01 3.91888559e-01
-3.63676727e-01 1.49964765e-01 6.51945055e-01 -6.22918248e-01
-5.90894401e-01 -6.91378415e-01 1.68896411e-02 1.74350178e+00
-1.05500007e-02 -3.41861337e-01 -9.44708049e-01 -1.06869304e+00
-5.31196654e-01 9.49309707e-01 -3.09881121e-01 3.40450048e-01
-8.75740409e-01 -7.37016082e-01 7.41264820e-01 4.71662015e-01
1.04624629e+00 -1.36064160e+00 -2.39909098e-01 4.75349277e-01
-5.89374840e-01 -1.44545436e+00 -3.41093898e-01 2.15802073e-01
-4.49679196e-01 -7.07397044e-01 -3.37939173e-01 -1.04430997e+00
4.17895585e-01 -2.04906821e-01 1.13538635e+00 -6.27304390e-02
2.19353214e-01 1.92747697e-01 -5.41723788e-01 -2.54438698e-01
-7.54265130e-01 1.88985839e-01 -1.74765751e-01 2.13142231e-01
7.18266726e-01 -2.84187913e-01 -1.05350941e-01 3.86973202e-01
-4.58404899e-01 3.63615602e-01 2.27395952e-01 1.13883758e+00
-3.68199050e-02 -4.64136362e-01 8.96748900e-01 -1.15163755e+00
9.44620788e-01 -4.18019384e-01 -1.72558933e-01 4.04841423e-01
-1.47798285e-01 3.58938813e-01 6.42850995e-01 8.72761533e-02
-1.93291295e+00 1.77964956e-01 -6.11164868e-01 2.57370412e-01
-8.27020526e-01 3.09064776e-01 -4.85495090e-01 5.99784851e-01
4.33133245e-01 1.60166323e-01 2.55395900e-02 -7.56643474e-01
2.59674191e-01 1.38090050e+00 5.37201285e-01 -7.51160800e-01
-1.30276039e-01 7.43208379e-02 -4.08248335e-01 -8.42334211e-01
-1.36728549e+00 -6.32071495e-01 -8.59556556e-01 -3.96999456e-02
1.35193074e+00 -9.10296559e-01 -1.01230836e+00 5.71019232e-01
-1.65038264e+00 -5.44451058e-01 3.16288322e-01 1.86546654e-01
-8.38615119e-01 5.87867677e-01 -1.10926354e+00 -1.26681149e+00
-3.55511308e-01 -1.12624061e+00 1.19500339e+00 2.81388015e-01
-3.69875401e-01 -1.29174030e+00 1.31940424e-01 8.99965107e-01
1.03015773e-01 -7.56828263e-02 9.11449134e-01 -1.38313973e+00
-6.33124486e-02 1.56111106e-01 -8.36397894e-03 4.56566364e-01
1.06640197e-01 -6.05975926e-01 -1.26992202e+00 7.95736760e-02
5.26146054e-01 -8.95238638e-01 6.35641098e-01 8.11879188e-02
8.13197911e-01 -6.85921907e-01 -1.36356473e-01 -1.49942681e-01
7.63956964e-01 5.01384079e-01 6.71315730e-01 -2.07007900e-01
6.39815688e-01 9.42723215e-01 8.86124790e-01 4.64189112e-01
8.96621406e-01 8.76303136e-01 2.60261685e-01 -1.18137607e-02
3.20623845e-01 -2.56849080e-01 4.60444361e-01 1.07250416e+00
7.58647174e-02 -4.35631305e-01 -9.59121048e-01 2.46447816e-01
-2.30875635e+00 -8.70754480e-01 -6.64143860e-02 1.72136939e+00
1.19526637e+00 9.67964306e-02 1.38451234e-01 -3.06318134e-01
8.28145802e-01 5.10563850e-01 -1.39793739e-01 -8.34982336e-01
1.34052366e-01 -1.15638144e-01 -1.39238507e-01 1.08955824e+00
-1.43836761e+00 1.21150160e+00 5.64137888e+00 6.69357896e-01
-5.80815613e-01 3.64138693e-01 9.16356504e-01 6.56387627e-01
1.41953260e-01 -1.49579728e-02 -1.17284298e+00 2.22148716e-01
1.31527293e+00 2.40530938e-01 2.58690745e-01 9.50035095e-01
3.99651080e-02 -1.44572109e-01 -1.26511717e+00 7.52580285e-01
2.94211984e-01 -1.07286298e+00 -2.88194329e-01 -1.99811876e-01
4.72752512e-01 -2.40583792e-01 -5.67011833e-01 7.08838999e-01
7.48480976e-01 -1.32816017e+00 1.06986731e-01 3.99764538e-01
4.06649232e-01 -4.54705566e-01 1.11505985e+00 8.26225281e-01
-9.56444919e-01 1.79219004e-02 -2.26781964e-01 -5.21071970e-01
4.10271049e-01 -5.20782210e-02 -1.29904437e+00 3.32696348e-01
1.66527584e-01 6.67437375e-01 -9.81009826e-02 2.35445172e-01
-4.14419472e-01 8.91521394e-01 1.56371333e-02 -3.26258838e-01
5.23328900e-01 -1.73681498e-01 3.57707500e-01 1.36507177e+00
-5.56689203e-01 5.05625188e-01 6.78649187e-01 3.40980560e-01
-9.05640870e-02 1.14440314e-01 -4.14960414e-01 1.64787546e-01
3.13795507e-01 9.74908829e-01 -1.89095184e-01 -5.31150877e-01
-6.96160674e-01 1.15205312e+00 5.76790571e-01 1.56075984e-01
-6.57459319e-01 -5.95618449e-02 5.07206917e-01 -4.77742285e-01
9.53569710e-02 -9.79493633e-02 -8.01936015e-02 -1.07449138e+00
-1.18216954e-01 -9.87717688e-01 5.38690448e-01 -3.92441869e-01
-1.35032558e+00 8.51780415e-01 -8.22794661e-02 -7.83594012e-01
-7.23747909e-01 -6.04300499e-01 -7.35183835e-01 8.83077860e-01
-1.34880495e+00 -1.23665500e+00 -4.72268313e-02 5.76870084e-01
1.46239436e+00 -3.52700740e-01 1.47359490e+00 3.03479508e-02
-4.99316782e-01 3.14463705e-01 -2.29137167e-01 5.15030921e-01
6.85466468e-01 -1.43508112e+00 2.59049386e-01 2.53380418e-01
-1.98014572e-01 1.82651371e-01 5.70308506e-01 -4.43376660e-01
-8.71117055e-01 -7.67318606e-01 1.29108489e+00 -5.63164830e-01
5.09897947e-01 -4.40464139e-01 -1.00025022e+00 1.05455375e+00
8.73019755e-01 -4.69012111e-01 9.77809131e-01 4.77957428e-01
-1.36120366e-02 4.05243754e-01 -9.55911219e-01 3.86490464e-01
7.86589801e-01 -5.37057638e-01 -1.24732900e+00 5.24816096e-01
9.93942320e-01 -5.45808792e-01 -9.06610727e-01 1.98604867e-01
1.47559093e-02 -8.65178466e-01 5.96097708e-01 -1.00353968e+00
4.93448704e-01 -2.06908328e-04 -1.89946830e-01 -1.13656414e+00
-2.58914772e-02 -9.28314626e-01 3.88175528e-03 1.42673957e+00
4.13265109e-01 -3.14433008e-01 4.70555097e-01 9.68892038e-01
-5.92659056e-01 -5.77368557e-01 -1.10481012e+00 -1.39500558e-01
4.63845097e-02 -1.17338575e-01 3.55676681e-01 1.00090253e+00
5.94895184e-01 1.33981383e+00 -8.16817939e-01 -3.44566554e-02
1.61971405e-01 5.55012338e-02 8.15821826e-01 -1.18256664e+00
-6.36264265e-01 2.58613396e-02 -1.25958458e-01 -1.87712169e+00
8.19043577e-01 -5.11322320e-01 5.62407374e-01 -1.36027133e+00
4.78865094e-02 -1.88578546e-01 2.16596365e-01 4.73060668e-01
-4.79369581e-01 -5.68072081e-01 -1.85096469e-02 2.09502503e-01
-1.22675419e+00 9.18502688e-01 9.92279768e-01 -1.88646480e-01
-3.54347646e-01 3.75388354e-01 -4.15496260e-01 1.03752398e+00
7.66768098e-01 -2.26124093e-01 -4.62918013e-01 -3.86786193e-01
-5.26723921e-01 8.41091812e-01 1.89439598e-02 -5.43837130e-01
3.33104938e-01 -3.72224182e-01 -2.41911843e-01 -4.66762215e-01
6.37100637e-01 -4.31335300e-01 -5.62665403e-01 1.41518846e-01
-9.20975566e-01 -1.95584074e-01 -1.19476043e-01 5.92105091e-01
-5.75244427e-01 -5.37292659e-01 6.76425040e-01 -3.61561924e-01
-8.78500223e-01 -1.23166725e-01 -7.35680640e-01 2.97466993e-01
7.97035992e-01 3.80650491e-01 -4.36983049e-01 -5.75064898e-01
-9.44345236e-01 6.06089771e-01 -3.32300693e-01 3.59502405e-01
4.37034994e-01 -9.65326548e-01 -8.35964799e-01 -4.01477814e-02
-5.61823919e-02 3.39705199e-01 3.73278439e-01 6.11442327e-01
-1.85659051e-01 7.55701721e-01 9.41488296e-02 -5.62282085e-01
-1.45978284e+00 1.60176337e-01 3.38955313e-01 -7.91926742e-01
-3.84362340e-01 9.97523010e-01 3.35863948e-01 -8.52954745e-01
5.52574575e-01 -4.20638733e-02 -6.21247530e-01 -1.52724937e-01
5.90640247e-01 1.03380322e-01 -2.97612846e-01 -9.05935526e-01
-3.11409146e-01 -6.68983981e-02 -4.19585973e-01 -1.07689507e-01
1.14891100e+00 -4.31826890e-01 -1.35756731e-01 6.46575451e-01
1.03812325e+00 -2.87005842e-01 -1.29601026e+00 -6.48583531e-01
3.73006552e-01 -1.15692370e-01 -2.62532949e-01 -9.07960832e-01
-3.61779183e-01 1.00026000e+00 1.02810085e-01 4.29056048e-01
8.07045817e-01 2.06714466e-01 7.74015546e-01 7.69439101e-01
2.94780552e-01 -1.32538974e+00 3.21616471e-01 1.04245496e+00
1.03486764e+00 -1.61513913e+00 -2.97976792e-01 -5.82790136e-01
-1.42648196e+00 1.11421382e+00 1.10320997e+00 1.87438820e-02
5.54806352e-01 6.83214143e-02 2.33213261e-01 -6.43695369e-02
-1.15677977e+00 -7.67003149e-02 -1.08779661e-01 4.55233485e-01
7.21364677e-01 2.59543628e-01 -3.42230231e-01 9.07422483e-01
3.90150696e-02 -2.73910642e-01 3.80889028e-01 8.26529980e-01
-6.53299630e-01 -1.26419199e+00 -6.01007901e-02 1.90129519e-01
-5.14001250e-01 -1.73560575e-01 -6.73799455e-01 3.76125813e-01
-3.72724384e-01 1.37348092e+00 3.13305631e-02 -2.05279201e-01
1.47801474e-01 6.38760149e-01 -7.98721835e-02 -1.03448629e+00
-8.42745960e-01 1.69090912e-01 8.78657043e-01 -2.35485241e-01
-7.59750366e-01 -6.70115173e-01 -1.58290315e+00 6.29201308e-02
-3.44413638e-01 4.45450693e-01 3.07884753e-01 1.53013313e+00
2.30473187e-02 3.74887615e-01 8.05888355e-01 -5.57415843e-01
-8.29353392e-01 -1.61188316e+00 -3.33057016e-01 4.75116134e-01
2.84573942e-01 -4.79222953e-01 -1.70811117e-01 1.90091878e-01] | [12.67773723602295, 7.6604485511779785] |
41bc46ce-ee73-4ab8-bccc-d18eddbba1b6 | fairness-constraint-in-structural | 2202.08977 | null | https://arxiv.org/abs/2202.08977v1 | https://arxiv.org/pdf/2202.08977v1.pdf | Fairness constraint in Structural Econometrics and Application to fair estimation using Instrumental Variables | A supervised machine learning algorithm determines a model from a learning sample that will be used to predict new observations. To this end, it aggregates individual characteristics of the observations of the learning sample. But this information aggregation does not consider any potential selection on unobservables and any status-quo biases which may be contained in the training sample. The latter bias has raised concerns around the so-called \textit{fairness} of machine learning algorithms, especially towards disadvantaged groups. In this chapter, we review the issue of fairness in machine learning through the lenses of structural econometrics models in which the unknown index is the solution of a functional equation and issues of endogeneity are explicitly accounted for. We model fairness as a linear operator whose null space contains the set of strictly {\it fair} indexes. A {\it fair} solution is obtained by projecting the unconstrained index into the null space of this operator or by directly finding the closest solution of the functional equation into this null space. We also acknowledge that policymakers may incur a cost when moving away from the status quo. Achieving \textit{approximate fairness} is obtained by introducing a fairness penalty in the learning procedure and balancing more or less heavily the influence between the status quo and a full fair solution. | ['Jean-Michel Loubes', 'Jean-Pierre Florens', 'Samuele Centorrino'] | 2022-02-16 | null | null | null | null | ['econometrics'] | ['miscellaneous'] | [ 9.94748846e-02 7.11871803e-01 -8.03778291e-01 -6.23915255e-01
-5.30781209e-01 -4.40115005e-01 3.68100584e-01 1.62039608e-01
-7.35375345e-01 9.99711573e-01 5.40584326e-01 -7.09800720e-01
-5.64354420e-01 -7.35978484e-01 -4.07318175e-01 -7.14071810e-01
3.20218891e-01 3.76241624e-01 -8.36230695e-01 3.22633445e-01
3.86574656e-01 3.21271062e-01 -1.25170016e+00 -2.06983481e-02
1.38232410e+00 8.38149428e-01 -4.99990851e-01 2.37833470e-01
-1.17792286e-01 1.07687664e+00 -7.03173757e-01 -6.28533959e-01
6.99604332e-01 -7.29436159e-01 -6.08456969e-01 4.55832854e-02
6.50022507e-01 -2.26540476e-01 -1.07843943e-01 1.13410068e+00
2.63905436e-01 1.68302104e-01 1.12195778e+00 -1.56801951e+00
-8.59811723e-01 6.54826045e-01 -3.46842498e-01 -4.61947024e-02
1.01024918e-01 1.92804281e-02 1.22859919e+00 -6.38681710e-01
4.29403394e-01 1.35158575e+00 5.11602879e-01 3.68611187e-01
-1.48631477e+00 -8.62902641e-01 9.53320563e-02 -2.37922698e-01
-1.33300972e+00 -6.64525628e-01 4.21004981e-01 -9.16066468e-01
3.05405140e-01 8.47857952e-01 4.31488216e-01 4.18935984e-01
3.91447097e-01 3.88289690e-01 1.42760038e+00 -5.86786032e-01
4.25342411e-01 5.40127516e-01 2.49247357e-01 4.74685222e-01
5.32532752e-01 5.52187204e-01 -3.56338918e-01 -4.39588517e-01
6.49267316e-01 -5.26673496e-02 -4.51305173e-02 -4.00590777e-01
-8.02675605e-01 1.17196846e+00 3.03658098e-01 -1.03445448e-01
-4.53667611e-01 -6.90803751e-02 4.80982428e-03 6.16128504e-01
6.16638303e-01 7.38037527e-01 -4.07161444e-01 2.86352038e-01
-1.06885302e+00 1.20871149e-01 7.84837842e-01 3.99108529e-01
9.66767728e-01 5.26641794e-02 -3.14757943e-01 5.49912155e-01
1.38820946e-01 4.52574492e-01 3.01891625e-01 -1.53175819e+00
3.85776162e-01 6.89703286e-01 3.06505919e-01 -8.11572194e-01
-4.66405004e-01 -7.75616288e-01 -6.41313732e-01 6.75790370e-01
9.30698991e-01 -5.42486429e-01 -4.07575309e-01 1.84705186e+00
3.47528785e-01 -3.76141727e-01 -4.97329682e-02 7.97383487e-01
2.32017130e-01 4.87600327e-01 3.17288309e-01 -7.25784957e-01
7.59995162e-01 -4.63693351e-01 -5.59502423e-01 -1.69012234e-01
7.51368999e-01 -5.32606542e-01 8.62284482e-01 1.40612870e-01
-1.01099253e+00 -3.04981649e-01 -5.96350431e-01 -5.82122318e-02
1.17084816e-01 -1.58416346e-01 6.20513678e-01 1.08038247e+00
-8.55244696e-01 7.10693717e-01 -3.56457055e-01 1.87273160e-01
4.25503194e-01 5.17256677e-01 -2.20096931e-01 3.08949202e-01
-9.49746907e-01 9.74031568e-01 1.47201166e-01 -2.62823552e-01
-2.63276458e-01 -1.07313800e+00 -8.76348495e-01 4.19918358e-01
6.39439702e-01 -6.22520924e-01 7.36137688e-01 -1.67226064e+00
-1.07134867e+00 7.77476490e-01 -1.18665464e-01 -7.02345446e-02
7.68407106e-01 8.56253654e-02 -2.32716173e-01 -5.03571570e-01
4.65571016e-01 1.71129599e-01 6.87083900e-01 -1.15861094e+00
-7.65042663e-01 -6.52955890e-01 -2.54711471e-02 2.97081292e-01
-3.01366180e-01 8.76556858e-02 4.42262262e-01 -7.45044470e-01
2.30544917e-02 -9.58609104e-01 -1.81177020e-01 4.95137982e-02
-2.31955692e-01 -3.19977403e-01 1.81108236e-01 -6.71464622e-01
1.52743161e+00 -2.06017208e+00 -3.11621815e-01 5.90164423e-01
3.32188487e-01 -1.34405732e-01 -3.40313017e-02 1.89341784e-01
-3.56968611e-01 3.54748905e-01 -9.64107290e-02 -4.11919411e-03
1.68448076e-01 -2.86458340e-02 -6.31690770e-02 8.67158115e-01
-1.23715609e-01 5.75428426e-01 -6.38180494e-01 -3.53367865e-01
1.22263089e-01 -1.28887936e-01 -8.31088066e-01 -3.45868780e-03
3.39143813e-01 2.84089535e-01 -2.45591998e-01 1.79934114e-01
5.97638190e-01 8.68282542e-02 1.60276249e-01 3.35275382e-01
-5.43046892e-01 3.73152018e-01 -1.29812253e+00 7.95843482e-01
-2.07605660e-01 3.76155764e-01 -1.94166191e-02 -1.28614187e+00
8.34511459e-01 1.99985310e-01 6.82415724e-01 -7.17582703e-01
1.00478418e-02 2.27698579e-01 4.13841248e-01 -3.25328827e-01
2.58611619e-01 -6.83334291e-01 1.82378124e-02 7.27547109e-01
-7.29152635e-02 1.27937466e-01 -1.40948579e-01 -1.10444084e-01
5.01929283e-01 -2.29926080e-01 6.39842212e-01 -1.09475756e+00
3.12065750e-01 -1.26673967e-01 9.75919664e-01 9.99737084e-01
-2.09891081e-01 3.60173523e-01 1.03311121e+00 -4.00583267e-01
-9.80317116e-01 -1.03879094e+00 -4.66437787e-01 1.19220412e+00
-4.64509368e-01 1.27980962e-01 -5.83139360e-01 -7.87105322e-01
3.91942501e-01 1.21217549e+00 -1.04178441e+00 -3.88965815e-01
-9.41369832e-02 -8.41631055e-01 1.05468377e-01 6.66375533e-02
6.13374785e-02 -5.12515903e-01 -6.71502054e-01 -1.47108167e-01
3.69798578e-02 -2.70029217e-01 -5.02426744e-01 -2.36683264e-02
-6.74187899e-01 -9.18904483e-01 -5.66268444e-01 -1.98157489e-01
8.33979070e-01 -1.65227994e-01 9.57057834e-01 1.61561027e-01
2.23984733e-01 1.61469907e-01 4.88430448e-02 -9.43853080e-01
-5.72467029e-01 -2.33948141e-01 -2.88641416e-02 2.97070980e-01
5.07984042e-01 -1.74656704e-01 -5.53434193e-01 1.69139758e-01
-4.44513768e-01 -8.21190998e-02 -3.59852202e-02 8.06681335e-01
1.20445542e-01 -2.78316606e-02 9.01062250e-01 -1.07034719e+00
4.64308202e-01 -6.29193425e-01 -7.29438365e-01 1.70788050e-01
-1.08110988e+00 1.19735807e-01 3.46635610e-01 -2.73090392e-01
-1.23414397e+00 -3.25054109e-01 3.28591853e-01 2.10231677e-01
1.64785609e-01 7.73845732e-01 -2.25712180e-01 1.80371493e-01
9.27692115e-01 -4.84472096e-01 3.45514834e-01 -3.59057605e-01
5.03007248e-02 8.25388134e-01 1.48935929e-01 -5.38895965e-01
6.21493459e-01 3.68111283e-01 1.62231728e-01 -4.19436157e-01
-1.14016497e+00 -1.68726966e-01 -4.12395686e-01 -2.38959134e-01
4.63267714e-01 -6.64235055e-01 -8.29808593e-01 -5.18041197e-03
-4.15553361e-01 -3.80034268e-01 -9.18296516e-01 9.72045779e-01
-5.43506384e-01 -1.01589672e-01 -7.06548169e-02 -1.19315779e+00
-2.87851930e-04 -9.51113224e-01 3.21266264e-01 1.61602527e-01
-7.29950905e-01 -1.07990623e+00 -2.90478133e-02 5.71901679e-01
2.15769276e-01 2.90397346e-01 1.23366714e+00 -7.67704964e-01
-1.59397408e-01 -4.05987054e-01 -1.05103962e-01 4.15158093e-01
2.85123527e-01 5.15606590e-02 -8.77392411e-01 -3.18998069e-01
3.54160577e-01 -7.76076391e-02 7.36685455e-01 8.16973507e-01
1.02576518e+00 -9.02369380e-01 1.01599373e-01 4.93547082e-01
1.33796561e+00 4.44278717e-01 2.18409494e-01 2.54756987e-01
5.68871379e-01 1.12799275e+00 3.65369201e-01 6.54147625e-01
5.90893626e-01 2.89566517e-01 1.12243384e-01 -2.29789674e-01
2.96469063e-01 -1.94116786e-01 9.99545380e-02 1.28173083e-01
-5.84406815e-02 9.47658941e-02 -1.01125491e+00 2.70741194e-01
-1.72155130e+00 -1.12899578e+00 -1.56894058e-01 2.92535281e+00
8.12017798e-01 6.81853816e-02 2.61523366e-01 2.01283827e-01
5.79674423e-01 -1.02660097e-01 -5.30418217e-01 -8.80550504e-01
6.78526312e-02 1.15078716e-02 6.20674789e-01 1.11366498e+00
-7.83416510e-01 3.63397419e-01 6.82255840e+00 6.25403762e-01
-1.00334167e+00 -2.36923009e-01 1.14785457e+00 -3.68306547e-01
-7.12888420e-01 2.68918127e-01 -3.01004738e-01 4.91071343e-01
6.83401048e-01 -7.54565060e-01 3.36453259e-01 4.91667897e-01
6.51983619e-01 -4.67397124e-01 -1.20886493e+00 3.79756361e-01
-1.38018936e-01 -8.91547918e-01 -8.16610605e-02 6.94775760e-01
1.14324164e+00 -4.18322057e-01 3.78998220e-01 2.30749220e-01
5.72612107e-01 -1.43149316e+00 1.02444398e+00 5.93241096e-01
9.36013639e-01 -8.69180381e-01 6.31355166e-01 6.28408253e-01
-4.13364857e-01 -4.67898190e-01 -3.96436125e-01 -8.86690557e-01
-3.61962765e-01 8.83115470e-01 -4.44668025e-01 1.12200595e-01
2.50247031e-01 2.37339929e-01 -3.38364691e-01 7.99601376e-01
2.12112099e-01 6.59371316e-01 -3.87185030e-02 3.88075262e-01
8.07039291e-02 -6.44463897e-01 2.94494748e-01 7.26574302e-01
4.15274411e-01 2.92419136e-01 -7.68928155e-02 9.50412989e-01
-2.24664286e-01 3.48377556e-01 -5.34693241e-01 1.23602055e-01
4.62812662e-01 7.40332723e-01 -2.77499348e-01 -3.98505956e-01
-6.56278431e-01 3.08311284e-01 2.28193015e-01 6.24551475e-01
-3.07264805e-01 7.77182132e-02 8.22786450e-01 1.25232235e-01
-5.44142485e-01 2.67147541e-01 -9.13446784e-01 -1.12787402e+00
-2.24876612e-01 -1.05837822e+00 7.67528117e-01 -2.86245912e-01
-1.03577805e+00 -3.31899136e-01 -2.21232437e-02 -6.73816979e-01
-1.94892421e-01 -2.92237699e-01 -4.62568760e-01 1.43623066e+00
-1.02995884e+00 -5.09037495e-01 2.64061213e-01 3.82453471e-01
-3.08910850e-02 -2.19473064e-01 7.12892175e-01 3.06544844e-02
-4.99107629e-01 7.34306812e-01 4.67430383e-01 -5.18745780e-02
6.80380285e-01 -1.33624232e+00 -1.66716069e-01 7.32800186e-01
-1.84704542e-01 7.04814374e-01 7.47024477e-01 -6.86676383e-01
-5.40360332e-01 -8.93105328e-01 1.52196598e+00 -6.70085728e-01
3.58769089e-01 9.03752632e-03 -7.58468091e-01 9.30548549e-01
-2.39183977e-01 -2.70311952e-01 1.07116807e+00 3.02206397e-01
-2.66598880e-01 -1.76288918e-01 -1.43656802e+00 5.48129916e-01
7.74623632e-01 -5.07408440e-01 -3.47934067e-01 -4.88874242e-02
3.67923766e-01 3.05590145e-02 -9.04640317e-01 3.71421605e-01
9.21019673e-01 -1.13939607e+00 7.04446316e-01 -9.69118118e-01
5.22083342e-01 2.31398642e-01 -7.27343634e-02 -1.39711106e+00
-6.70616925e-01 -4.95417327e-01 3.38349640e-01 1.14024770e+00
5.84843040e-01 -8.24177206e-01 9.65128601e-01 1.51446843e+00
2.48401210e-01 -5.38177609e-01 -1.10018790e+00 -5.10850489e-01
7.73317218e-01 -1.60218149e-01 7.11964607e-01 1.46949041e+00
9.32731554e-02 3.19560803e-02 -3.31610084e-01 -2.05694273e-01
9.33602631e-01 4.24927510e-02 5.42429924e-01 -1.46549737e+00
-1.65070504e-01 -6.60407662e-01 -3.11672855e-02 -2.84089386e-01
2.60987639e-01 -9.46725965e-01 -4.07350570e-01 -1.19240308e+00
4.12828147e-01 -7.07061410e-01 -3.16891253e-01 2.60189861e-01
-3.83359671e-01 -2.18216315e-01 5.72914183e-01 2.10897774e-01
1.34326667e-01 3.07530433e-01 1.15367675e+00 1.11553408e-01
-4.49596733e-01 2.85426348e-01 -1.35705626e+00 8.90428960e-01
9.29888487e-01 -8.60169172e-01 -3.34729731e-01 -3.29123586e-01
3.29560697e-01 3.21561545e-01 4.55208242e-01 -4.04659033e-01
-3.52228656e-02 -9.15498495e-01 5.69702268e-01 2.03878522e-01
-1.26383737e-01 -8.65164757e-01 4.76437628e-01 6.23279870e-01
-1.00236356e+00 -1.84876889e-01 -2.77770400e-01 -1.72330849e-02
2.11791158e-01 -4.84740555e-01 8.02026808e-01 -1.81326807e-01
2.64317334e-01 1.95262861e-02 -4.63876426e-01 2.19599485e-01
8.37735534e-01 -1.59262389e-01 -1.97979704e-01 -5.78046620e-01
-5.87019563e-01 2.57861584e-01 6.46801054e-01 1.04573376e-01
-2.05373149e-02 -1.34868884e+00 -1.01011884e+00 2.14577511e-01
-4.89524677e-02 -4.89974320e-01 -2.99130008e-02 6.82505310e-01
-5.70059605e-02 4.88641262e-01 -1.99722081e-01 1.95233628e-01
-9.64060962e-01 4.65915680e-01 6.02480888e-01 2.86410972e-02
-7.25267157e-02 5.71177661e-01 6.17831290e-01 -5.06080508e-01
8.75706822e-02 9.52682719e-02 -6.39539137e-02 4.32860464e-01
3.24607283e-01 1.02818191e+00 -3.03795964e-01 -7.86922395e-01
-2.14857429e-01 1.87835321e-01 3.82849008e-01 -2.99588025e-01
1.01465201e+00 -3.71397108e-01 -3.72578830e-01 5.85451663e-01
1.26342535e+00 2.97616512e-01 -1.35404456e+00 -2.02706859e-01
8.11725855e-02 -8.08732212e-01 7.91940987e-02 -9.90440190e-01
-9.15810645e-01 4.21110392e-01 3.60691994e-01 2.56960183e-01
7.98995435e-01 -2.49160111e-01 -1.26314893e-01 9.49441642e-02
1.68471206e-02 -1.31453896e+00 -3.96825790e-01 3.01386490e-02
8.56907964e-01 -1.32497835e+00 2.03297079e-01 1.50665045e-02
-6.06681347e-01 7.07804561e-01 3.12810987e-01 -1.39760777e-01
7.64701784e-01 -5.61650023e-02 4.16160494e-01 1.19143181e-01
-4.75986004e-01 7.90862739e-02 4.72166985e-01 4.69880462e-01
5.23996174e-01 5.35595715e-01 -9.98635352e-01 5.22536099e-01
-6.52962983e-01 -1.34885916e-03 6.74142063e-01 4.38766897e-01
-4.19422716e-01 -8.88581514e-01 -7.38230050e-01 1.06563735e+00
-7.87437439e-01 -7.61537924e-02 -2.63333738e-01 6.24952853e-01
3.93834442e-01 1.08509922e+00 2.80788153e-01 3.93703207e-02
2.38067761e-01 1.48269057e-01 -3.42260227e-02 -6.01432920e-01
-4.40612793e-01 4.32496741e-02 9.11958739e-02 -3.49212170e-01
-2.73431987e-01 -1.01524687e+00 -8.35315406e-01 -5.77225029e-01
-2.49030918e-01 3.91712397e-01 2.78167009e-01 1.00490117e+00
-1.04276024e-01 1.37796715e-01 9.12635684e-01 -3.42923313e-01
-1.07503426e+00 -4.65732574e-01 -8.38038087e-01 4.86608565e-01
6.02073550e-01 -3.20781946e-01 -6.42632425e-01 -2.65967935e-01] | [8.71043872833252, 5.317661285400391] |
c464e0fe-4d3d-4bdb-921a-3e87d4768726 | cola-coarse-label-pre-training-for-3d | 2202.06884 | null | https://arxiv.org/abs/2202.06884v3 | https://arxiv.org/pdf/2202.06884v3.pdf | COLA: COarse LAbel pre-training for 3D semantic segmentation of sparse LiDAR datasets | Transfer learning is a proven technique in 2D computer vision to leverage the large amount of data available and achieve high performance with datasets limited in size due to the cost of acquisition or annotation. In 3D, annotation is known to be a costly task; nevertheless, pre-training methods have only recently been investigated. Due to this cost, unsupervised pre-training has been heavily favored. In this work, we tackle the case of real-time 3D semantic segmentation of sparse autonomous driving LiDAR scans. Such datasets have been increasingly released, but each has a unique label set. We propose here an intermediate-level label set called coarse labels, which can easily be used on any existing and future autonomous driving datasets, thus allowing all the data available to be leveraged at once without any additional manual labeling. This way, we have access to a larger dataset, alongside a simple task of semantic segmentation. With it, we introduce a new pre-training task: coarse label pre-training, also called COLA. We thoroughly analyze the impact of COLA on various datasets and architectures and show that it yields a noticeable performance improvement, especially when only a small dataset is available for the finetuning task. | ['François Goulette', 'Jean-Emmanuel Deschaud', 'Jules Sanchez'] | 2022-02-14 | null | null | null | null | ['real-time-3d-semantic-segmentation', 'unsupervised-pre-training'] | ['computer-vision', 'methodology'] | [ 4.30129528e-01 3.39043528e-01 -1.50193721e-01 -5.76732814e-01
-8.45695019e-01 -6.29064143e-01 6.68969870e-01 3.22778672e-01
-7.73470283e-01 6.12418056e-01 -3.53159398e-01 -3.53872329e-01
-1.69393476e-02 -6.79457963e-01 -7.77416229e-01 -4.52039361e-01
-6.85351901e-03 8.80654991e-01 5.43102384e-01 6.92670001e-03
-4.08471823e-02 5.51732838e-01 -1.95649207e+00 -2.32170582e-01
8.18358898e-01 1.06463087e+00 4.30807352e-01 2.61594296e-01
-2.25051120e-01 2.66029269e-01 -2.12272525e-01 -1.71807885e-01
6.75718129e-01 -2.99820658e-02 -7.09144413e-01 2.66689122e-01
6.49064600e-01 -2.18007371e-01 1.27229199e-01 8.35068464e-01
5.25575936e-01 1.07660048e-01 6.14969075e-01 -1.21228766e+00
4.02774453e-01 3.73922378e-01 -4.83743966e-01 -1.25386357e-01
-9.25567374e-02 1.61669344e-01 8.88409495e-01 -8.11638415e-01
8.22531223e-01 9.08844888e-01 7.13857651e-01 3.48813087e-01
-1.40444899e+00 -6.52501702e-01 2.66913548e-02 1.73416421e-01
-1.34039712e+00 -3.50593954e-01 8.23884070e-01 -7.52362907e-01
6.85804605e-01 -9.35285836e-02 6.65885091e-01 8.75225663e-01
-4.78114337e-01 7.21690834e-01 1.41324282e+00 -4.36003089e-01
5.22209823e-01 2.17096955e-01 1.81806386e-01 5.82900882e-01
2.05714747e-01 8.21953714e-02 -3.27943236e-01 2.84868807e-01
3.55261385e-01 -2.25366190e-01 1.10187627e-01 -8.68557394e-01
-9.52461064e-01 8.83176565e-01 5.19505262e-01 -5.15747257e-02
-1.97328076e-01 5.46711637e-03 4.25355583e-01 1.76231354e-01
5.03549993e-01 2.77611315e-01 -5.01252055e-01 -2.20587224e-01
-1.07117045e+00 1.83830887e-01 7.25228190e-01 9.36639249e-01
1.30133688e+00 -3.70045543e-01 2.29454085e-01 7.98722267e-01
-5.54424264e-02 5.17279029e-01 8.56388286e-02 -9.13069248e-01
4.51671809e-01 8.04126084e-01 6.70782998e-02 -5.27571738e-01
-6.62479639e-01 -4.38817263e-01 -6.70812726e-01 5.83865404e-01
5.22343278e-01 -4.58655544e-02 -1.09952474e+00 1.75621271e+00
6.17714763e-01 1.17382683e-01 -1.02514252e-01 7.66330302e-01
3.58471304e-01 1.86025783e-01 5.11400960e-02 1.49124884e-04
1.15325904e+00 -8.74655724e-01 -2.73464859e-01 -4.32359874e-01
6.73817337e-01 -5.17469764e-01 1.11375713e+00 4.74092603e-01
-5.53865314e-01 -6.55424654e-01 -1.10259378e+00 -2.20247328e-01
-6.23633444e-01 7.19324574e-02 6.22777224e-01 6.09446049e-01
-9.37952578e-01 5.56490719e-01 -9.14680064e-01 -4.97953176e-01
7.73821592e-01 3.96528572e-01 -5.42425990e-01 -3.17684770e-01
-7.98127711e-01 9.62950945e-01 6.04095757e-01 -9.53701213e-02
-7.83854246e-01 -7.35864043e-01 -7.97703266e-01 -2.53260911e-01
8.73690546e-01 -5.56611001e-01 1.12926686e+00 -4.79835361e-01
-1.37557685e+00 1.22832441e+00 2.14369550e-01 -6.00464702e-01
9.13055599e-01 -2.73057997e-01 1.75564170e-01 -1.14814648e-02
2.64740139e-01 1.16908133e+00 9.98427629e-01 -1.41309822e+00
-7.10446477e-01 -4.91796970e-01 2.60544717e-01 8.58265236e-02
-1.97330117e-01 -5.15856743e-01 -5.63932121e-01 -3.04954618e-01
1.49998050e-02 -1.38428926e+00 -2.88611561e-01 -2.18530148e-02
-3.48213553e-01 -2.05498248e-01 8.96570742e-01 -2.10676670e-01
6.18556142e-01 -2.25169730e+00 1.43311605e-01 2.01910719e-01
1.58871606e-01 3.23099434e-01 2.94297487e-02 1.76783219e-01
1.37123734e-01 -1.10085547e-01 -8.33179653e-01 -6.40004456e-01
1.72404587e-01 4.55596566e-01 3.20073813e-02 3.13423961e-01
3.76231611e-01 7.34928012e-01 -8.74955535e-01 -6.43701017e-01
4.56272870e-01 3.01432639e-01 -6.21575296e-01 1.91444874e-01
-5.10553181e-01 8.26770484e-01 -4.67005312e-01 5.54082513e-01
7.25645900e-01 -1.54807732e-01 -7.72340521e-02 -7.75411651e-02
-1.83412269e-01 1.06028795e-01 -1.26723015e+00 2.08065486e+00
-7.11028337e-01 4.69247311e-01 6.12935126e-02 -1.24694693e+00
9.80756938e-01 -7.78457671e-02 6.35181308e-01 -5.76264441e-01
2.55066037e-01 5.17326236e-01 -6.45032749e-02 -4.33722526e-01
3.94123554e-01 -1.66688353e-01 -1.74830511e-01 6.04932308e-01
-1.99104603e-02 -5.82851768e-01 4.36421901e-01 -1.01902664e-01
1.06301534e+00 5.10896206e-01 1.48322150e-01 -1.46587595e-01
4.48093832e-01 3.69136363e-01 3.86397868e-01 6.34735167e-01
-1.17746532e-01 4.84176755e-01 2.42644861e-01 -3.25993240e-01
-1.07801938e+00 -7.75469184e-01 -3.98847371e-01 1.06379867e+00
2.08163053e-01 -2.31456086e-01 -8.75520468e-01 -8.90114129e-01
2.32496560e-01 4.56138968e-01 -5.82798839e-01 4.55990955e-02
-5.85163295e-01 -4.87217337e-01 3.46327156e-01 4.59902465e-01
6.37503088e-01 -1.01625180e+00 -1.21998537e+00 -3.42174619e-02
2.29701385e-01 -1.51220560e+00 8.68637115e-02 7.32019484e-01
-9.46168780e-01 -9.79331136e-01 -4.57997918e-01 -4.31728631e-01
4.71086353e-01 2.19555184e-01 1.06843889e+00 -1.47104591e-01
-4.81363624e-01 2.99976557e-01 -3.33586127e-01 -5.87176085e-01
-2.83521026e-01 6.19586885e-01 -6.34745508e-02 -1.71321966e-02
2.93909907e-01 -8.01234126e-01 -3.80700886e-01 2.61953205e-01
-9.16249990e-01 2.04989895e-01 9.05798733e-01 6.32499874e-01
7.30096340e-01 -1.96062237e-01 5.98943353e-01 -1.15183651e+00
-1.28538441e-02 -3.35707217e-01 -8.73697758e-01 -3.25541377e-01
-6.43261790e-01 1.70741484e-01 3.59703600e-01 -9.48235989e-02
-7.66951084e-01 4.86474574e-01 -2.09707230e-01 -4.92208958e-01
-6.67468667e-01 6.28133297e-01 -3.02477896e-01 -1.19530320e-01
5.05382955e-01 -2.73835421e-01 2.55618572e-01 -7.30237186e-01
7.52205193e-01 7.07858622e-01 5.88958621e-01 -4.79144007e-01
9.40558612e-01 5.28361261e-01 3.15780163e-01 -8.16756845e-01
-1.16348839e+00 -6.65953577e-01 -1.20268261e+00 -1.08799987e-01
8.60332549e-01 -9.26602840e-01 -1.97857708e-01 3.37787718e-01
-6.91834867e-01 -7.21006930e-01 -7.30494857e-01 3.95258367e-01
-7.67819107e-01 1.67042032e-01 -1.11813927e-02 -5.31531990e-01
5.37600219e-02 -1.21788085e+00 1.17878199e+00 -5.58060743e-02
-1.64640341e-02 -7.63901711e-01 -1.26718327e-01 4.69435304e-01
2.94158280e-01 4.20045912e-01 9.12663400e-01 -7.66434729e-01
-7.30147421e-01 -2.47226492e-01 -3.83344859e-01 4.03119594e-01
4.54232395e-02 -4.59272861e-01 -1.26334512e+00 -2.79953092e-01
-1.00681007e-01 -5.72079599e-01 9.49116588e-01 3.33301984e-02
1.07004607e+00 4.21706051e-01 -3.48347008e-01 4.73701119e-01
1.19464445e+00 -2.68078715e-01 1.47756547e-01 4.58520204e-01
7.64492333e-01 8.57920468e-01 9.35718358e-01 2.69568026e-01
4.75659460e-01 8.66737425e-01 6.84723318e-01 -1.75634399e-01
-3.45490843e-01 -1.46091744e-01 -5.78638352e-02 6.25441194e-01
1.05617773e-02 2.24571768e-02 -1.16017091e+00 4.61566001e-01
-1.87433827e+00 -4.51697916e-01 2.91619338e-02 2.30756402e+00
7.50240028e-01 4.83116001e-01 1.40464872e-01 3.85379344e-01
4.94653016e-01 -6.21921457e-02 -8.01361442e-01 1.45827398e-01
1.19375572e-01 3.41076344e-01 7.68442154e-01 3.62261385e-01
-1.39644945e+00 1.00768137e+00 5.28771019e+00 8.07505250e-01
-1.21917272e+00 2.17697307e-01 3.33284378e-01 -2.48547317e-03
2.26496663e-02 5.37654571e-03 -8.68774891e-01 3.41924131e-01
9.23256755e-01 2.10626557e-01 4.62788045e-02 9.87583518e-01
1.53649613e-01 -4.31655556e-01 -1.28730035e+00 1.10603642e+00
-1.07508093e-01 -9.71843481e-01 -1.57626271e-01 3.72105122e-01
6.35634542e-01 3.83649319e-01 -1.85857564e-01 4.59437400e-01
1.30055159e-01 -8.57526839e-01 7.39280999e-01 1.78900212e-01
9.75760043e-01 -6.90533280e-01 8.28467727e-01 7.56861985e-01
-9.37639058e-01 -1.18427135e-01 -2.58868754e-01 -3.74443159e-02
2.50634879e-01 8.50740433e-01 -8.08169901e-01 6.35881841e-01
5.92425704e-01 8.33576381e-01 -7.54859567e-01 1.06212103e+00
-2.50366241e-01 4.41066533e-01 -7.57582843e-01 3.83649975e-01
3.68885189e-01 -3.46464306e-01 4.00006831e-01 9.55858290e-01
3.52009624e-01 -2.77533770e-01 6.34710550e-01 6.09239280e-01
-1.91370726e-01 7.90579244e-02 -6.89365268e-01 2.56071180e-01
2.60632485e-01 1.48869669e+00 -9.29932296e-01 -2.33422086e-01
-3.32743645e-01 6.27957463e-01 4.20385718e-01 -7.02167302e-02
-6.73303962e-01 -6.64655417e-02 4.43472266e-01 3.43891442e-01
5.45300543e-01 -4.24830586e-01 -3.10684025e-01 -7.48719156e-01
2.54419651e-02 -4.51648653e-01 2.71168023e-01 -4.73274738e-01
-1.06301892e+00 4.34734285e-01 2.38385201e-01 -1.21337259e+00
-3.30142081e-01 -6.11653209e-01 -6.93030432e-02 5.37432194e-01
-1.93352306e+00 -1.33585942e+00 -6.08615875e-01 4.15803581e-01
5.26969492e-01 2.22791620e-02 6.44304633e-01 5.60466170e-01
-4.01582658e-01 2.74885744e-01 -1.04306169e-01 -8.21747780e-02
7.66345978e-01 -1.38992751e+00 1.67491361e-01 5.98919630e-01
2.56628722e-01 6.64170608e-02 5.42353690e-01 -4.37518865e-01
-1.25679266e+00 -1.24524391e+00 4.38326895e-01 -4.62297738e-01
5.73562384e-01 -6.28943443e-01 -8.66023600e-01 5.08394778e-01
-1.36111781e-01 2.97047913e-01 5.33373594e-01 9.83139575e-02
-2.35432848e-01 -1.84942916e-01 -1.00847089e+00 1.49191618e-01
1.22053683e+00 -4.92579609e-01 -3.91294479e-01 1.98344707e-01
6.75123274e-01 -5.12394905e-01 -6.88859344e-01 6.48940444e-01
2.18330920e-01 -1.06811857e+00 7.52816081e-01 -2.45145801e-02
2.05142036e-01 -5.16604066e-01 -9.11985561e-02 -1.15620303e+00
2.16988683e-01 -3.23968321e-01 1.12368591e-01 1.27137160e+00
4.35803235e-01 -4.87392306e-01 1.02678883e+00 4.05281544e-01
-4.28642511e-01 -5.40683985e-01 -1.05635631e+00 -8.45593333e-01
-4.25127596e-02 -7.60785758e-01 5.02840102e-01 8.49034607e-01
-5.21256387e-01 5.37982762e-01 -1.09292388e-01 1.93593018e-02
5.89577436e-01 4.70965028e-01 1.27650774e+00 -1.70421076e+00
-9.08714309e-02 -3.27378154e-01 -5.12501776e-01 -1.03929710e+00
2.66910166e-01 -1.19581294e+00 3.21492106e-01 -1.28672063e+00
-6.71816897e-03 -1.09045768e+00 -4.50619645e-02 6.67518854e-01
3.48384492e-03 6.23169959e-01 2.02312574e-01 2.91454792e-01
-4.71618414e-01 5.48969328e-01 1.01753974e+00 -1.84997782e-01
-2.57231534e-01 2.82060821e-02 -1.94332168e-01 8.45572710e-01
7.74295449e-01 -5.62566221e-01 -5.28091967e-01 -3.32909852e-01
2.85921954e-02 -3.44114155e-01 3.32752794e-01 -1.23355365e+00
2.14569494e-02 1.91668019e-01 -6.47418872e-02 -7.37668991e-01
5.37390471e-01 -1.06651866e+00 2.85580028e-02 1.49771884e-01
-1.49588764e-01 -3.89894515e-01 3.04323137e-01 6.71071947e-01
-2.52524167e-01 -4.27275300e-01 8.09424162e-01 -1.57761034e-02
-1.05435586e+00 4.35309261e-01 2.37513930e-02 1.26768351e-01
1.19127965e+00 -3.05966020e-01 1.13762468e-01 7.94521794e-02
-8.12831879e-01 2.68086582e-01 6.81642413e-01 3.47650677e-01
3.54201272e-02 -9.58350241e-01 -5.02351224e-01 2.53424257e-01
5.19871473e-01 7.48296380e-01 5.77786267e-02 8.05071115e-01
-3.66822034e-01 2.01714590e-01 -2.94785678e-01 -1.14808905e+00
-1.01898861e+00 6.13264263e-01 -7.99462795e-02 -3.03754747e-01
-8.64454091e-01 5.72616160e-01 -7.07822368e-02 -5.50086319e-01
1.02516055e-01 -3.06332409e-01 -2.80693650e-01 4.72529143e-01
2.64825225e-02 1.10375538e-01 3.78865719e-01 -6.23923182e-01
-1.70613363e-01 7.99475312e-01 1.10231519e-01 2.83146109e-02
1.47818077e+00 -2.85080016e-01 6.82124719e-02 6.82406843e-01
9.87274289e-01 -1.31422803e-01 -1.61705470e+00 -2.81239986e-01
4.38120008e-01 -5.15732944e-01 1.35702297e-01 -5.11531830e-01
-1.06237006e+00 1.16737592e+00 7.05962360e-01 2.23909467e-01
9.36093986e-01 1.23930737e-01 7.04488456e-01 4.20013934e-01
7.10160434e-01 -1.14527905e+00 -5.74145615e-02 5.04903674e-01
4.63991731e-01 -1.64215422e+00 1.03147998e-02 -5.78006566e-01
-4.24930364e-01 8.63477170e-01 3.56017679e-01 4.24212068e-02
6.94811702e-01 3.38204980e-01 2.80029148e-01 -1.80473998e-01
-2.85808891e-01 -7.72699773e-01 5.81167266e-02 7.47573853e-01
-4.11096551e-02 -4.42430675e-02 9.62185150e-04 1.63762987e-01
-2.69797862e-01 6.02844059e-02 3.15410674e-01 1.00848246e+00
-4.39311862e-01 -1.38342428e+00 -9.43112373e-02 4.35363442e-01
-3.44827659e-02 2.01050758e-01 -2.00774446e-01 8.94784510e-01
4.73690301e-01 6.67300284e-01 8.26973841e-02 -1.58848867e-01
3.58209193e-01 3.01531643e-01 3.77817750e-01 -7.87787080e-01
-1.54980451e-01 -9.82674733e-02 9.40791219e-02 -5.39439023e-01
-7.47305095e-01 -6.58527911e-01 -1.18569171e+00 1.36291385e-01
-2.05941841e-01 -1.07541913e-02 1.00306940e+00 1.09401107e+00
3.93143803e-01 4.46049690e-01 4.04233992e-01 -1.31148827e+00
-4.20036405e-01 -8.52067649e-01 -4.18425113e-01 4.42081600e-01
2.01263905e-01 -1.14848828e+00 -1.39080510e-01 2.69286007e-01] | [8.177294731140137, -2.6768925189971924] |
6a051859-8357-4592-bdee-fb83b5f0fb0a | hybrid-active-learning-via-deep-clustering | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Rana_Hybrid_Active_Learning_via_Deep_Clustering_for_Video_Action_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Rana_Hybrid_Active_Learning_via_Deep_Clustering_for_Video_Action_Detection_CVPR_2023_paper.pdf | Hybrid Active Learning via Deep Clustering for Video Action Detection | In this work, we focus on reducing the annotation cost for video action detection which requires costly frame-wise dense annotations. We study a novel hybrid active learning (AL) strategy which performs efficient labeling using both intra-sample and inter-sample selection. The intra-sample selection leads to labeling of fewer frames in a video as opposed to inter-sample selection which operates at video level. This hybrid strategy reduces the annotation cost from two different aspects leading to significant labeling cost reduction. The proposed approach utilize Clustering-Aware Uncertainty Scoring (CLAUS), a novel label acquisition strategy which relies on both informativeness and diversity for sample selection. We also propose a novel Spatio-Temporal Weighted (STeW) loss formulation, which helps in model training under limited annotations. The proposed approach is evaluated on UCF-101-24 and J-HMDB-21 datasets demonstrating its effectiveness in significantly reducing the annotation cost where it consistently outperforms other baselines. Project details available at https://sites.google.com/view/activesparselabeling/home | ['Yogesh S. Rawat', 'Aayush J. Rana'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['deep-clustering', 'deep-clustering'] | ['miscellaneous', 'natural-language-processing'] | [ 3.68946552e-01 5.71937561e-02 -4.95324403e-01 -5.22144318e-01
-1.58989120e+00 -4.28845525e-01 3.53645533e-01 3.51435065e-01
-8.26103508e-01 8.12077343e-01 9.82521251e-02 1.78159177e-01
1.14943281e-01 -3.08725923e-01 -6.37068272e-01 -8.62548530e-01
1.46967232e-01 3.32220465e-01 6.94157004e-01 5.51177621e-01
3.00909877e-01 1.66397199e-01 -1.39598310e+00 4.66594130e-01
8.45138431e-01 1.19251275e+00 4.71392483e-01 7.06604838e-01
4.64722477e-02 1.26510286e+00 -3.92058104e-01 -2.21410602e-01
4.34002072e-01 -2.90922493e-01 -8.46872926e-01 3.87031019e-01
5.52601933e-01 -3.21398079e-01 1.75180748e-01 8.41898143e-01
6.69001281e-01 5.71211755e-01 4.57969815e-01 -1.16720176e+00
9.27864835e-02 4.01746064e-01 -8.84546041e-01 4.58306789e-01
2.49700397e-01 4.83854637e-02 1.16603696e+00 -1.03799379e+00
5.86274147e-01 1.01311719e+00 6.29676104e-01 4.24042970e-01
-1.09441495e+00 -5.16101062e-01 2.28615597e-01 5.23981452e-01
-1.41161680e+00 -7.36966550e-01 7.35289156e-01 -4.69747335e-01
7.94553995e-01 2.33261868e-01 4.82028037e-01 9.29398894e-01
-2.44395465e-01 1.25881386e+00 8.90918672e-01 -7.40498424e-01
4.81955558e-01 2.94413790e-02 2.50169098e-01 9.42221284e-01
6.27246648e-02 -2.11429566e-01 -8.35302651e-01 -2.56357878e-01
4.60258633e-01 -6.33075535e-02 -3.64284128e-01 -3.97863269e-01
-1.09742129e+00 6.48908138e-01 -1.68162465e-01 -3.73392925e-02
-4.93905157e-01 2.48920709e-01 6.51181519e-01 -2.00224176e-01
7.25952208e-01 7.01776594e-02 -4.20913815e-01 -4.28402752e-01
-1.35247397e+00 -8.57456923e-02 3.04769397e-01 1.00057137e+00
7.30119526e-01 -8.71374682e-02 -6.80146575e-01 1.02651465e+00
5.18030643e-01 6.31776601e-02 2.58583546e-01 -1.43817270e+00
6.96158051e-01 6.20225310e-01 2.22958639e-01 -7.42343962e-01
-8.91251862e-02 -3.00356537e-01 -2.67373800e-01 1.65058643e-01
1.95226923e-01 -8.66941586e-02 -9.16787803e-01 1.58057654e+00
6.04197443e-01 5.78666151e-01 -3.29102099e-01 6.65697575e-01
4.73741323e-01 5.11080801e-01 3.54154617e-01 -5.96183360e-01
8.96849275e-01 -1.42770994e+00 -8.99040818e-01 1.22240894e-01
7.01667905e-01 -6.51448131e-01 1.01466441e+00 4.98126984e-01
-1.31925797e+00 -6.38831973e-01 -8.83768260e-01 -8.12169388e-02
-8.39449167e-02 5.44222057e-01 4.03802186e-01 4.31435347e-01
-1.01590645e+00 2.75125831e-01 -1.18185806e+00 -2.46568829e-01
7.60124624e-01 3.95478666e-01 -2.66915739e-01 -5.32542542e-02
-7.64694631e-01 3.74908209e-01 5.75071216e-01 -1.32082507e-01
-1.02886283e+00 -5.87247431e-01 -8.48434448e-01 -7.39172706e-03
9.74424601e-01 -1.81406230e-01 1.39309502e+00 -1.01819479e+00
-1.38248503e+00 6.05846345e-01 -3.82853776e-01 -7.23770797e-01
6.38864815e-01 -3.67278337e-01 -1.06580794e-01 5.04363656e-01
1.21367231e-01 1.02282536e+00 7.91805685e-01 -1.06450188e+00
-1.03925967e+00 -1.12439081e-01 9.18223336e-02 6.17406905e-01
-4.25034732e-01 1.13591209e-01 -9.20155704e-01 -8.55394781e-01
-8.10732245e-02 -1.05627644e+00 -4.45019267e-02 1.37892455e-01
-2.66799569e-01 -3.78357857e-01 9.69765306e-01 -6.21093392e-01
1.49976277e+00 -2.16802907e+00 7.28870779e-02 -1.02273889e-01
9.20473859e-02 4.29961234e-01 1.15601674e-01 1.29572511e-01
7.63643906e-02 -1.44858677e-02 -3.57252955e-01 -8.20206106e-01
-1.22987404e-01 1.17212281e-01 1.82262272e-01 4.10843819e-01
1.87718153e-01 6.00966394e-01 -9.51410711e-01 -1.11037409e+00
3.46948266e-01 3.84028882e-01 -5.41243494e-01 3.13337266e-01
-2.75075793e-01 4.30107176e-01 -1.81399792e-01 8.69025350e-01
4.23119336e-01 -2.15912268e-01 8.22679028e-02 -2.44592428e-01
-6.50739297e-02 -2.45295046e-03 -1.27326357e+00 1.92346442e+00
-2.89760888e-01 5.27027249e-01 -1.49752766e-01 -1.06947291e+00
4.05917794e-01 5.56928754e-01 7.33148396e-01 -3.12752932e-01
-2.52382569e-02 -7.41915870e-03 -3.24066222e-01 -4.51176852e-01
4.81336713e-01 4.28679854e-01 1.62926048e-01 4.11557466e-01
2.10136876e-01 6.35112524e-01 6.53687537e-01 4.59315896e-01
1.18749714e+00 6.19260252e-01 5.56443572e-01 -1.82037324e-01
6.89846992e-01 -1.75094679e-01 1.07697868e+00 6.67490184e-01
-7.66579390e-01 4.95514721e-01 2.47888744e-01 -7.13409707e-02
-5.98986030e-01 -7.22515881e-01 9.56996530e-02 1.17550826e+00
1.74158037e-01 -4.74354953e-01 -7.47943759e-01 -1.23358202e+00
-3.67016554e-01 6.85904980e-01 -5.68159878e-01 1.21594988e-01
-5.68591475e-01 -6.31716311e-01 3.50068331e-01 5.33910453e-01
6.72973216e-01 -8.58199418e-01 -7.00552166e-01 -1.42295808e-02
-4.80332792e-01 -1.08019280e+00 -5.61414599e-01 1.89165398e-01
-9.01696324e-01 -1.12180102e+00 -6.52240098e-01 -3.64748657e-01
6.40795827e-01 3.33226621e-01 9.74840939e-01 -4.44982648e-02
-3.12025249e-01 7.20978856e-01 -8.35216522e-01 -3.89668077e-01
-1.71273146e-02 3.58174257e-02 -1.51948094e-01 1.42677516e-01
5.39316237e-01 1.26064047e-02 -6.92353547e-01 4.06427503e-01
-7.22368777e-01 1.53811304e-02 4.76486027e-01 7.03946233e-01
1.12366927e+00 1.47328507e-02 4.21134621e-01 -1.03625882e+00
7.50417560e-02 -5.82258284e-01 -5.10411024e-01 5.68147540e-01
-5.71510375e-01 -2.50669539e-01 5.24423271e-02 -3.36461037e-01
-1.43358052e+00 5.19059956e-01 1.85783744e-01 -5.12156487e-01
-1.74142435e-01 1.16699509e-01 -7.64596835e-02 1.97650436e-02
4.02906358e-01 -9.66914892e-02 -3.45931441e-01 -4.94157314e-01
-6.34675622e-02 6.58021986e-01 2.58874387e-01 -4.53549951e-01
2.79048622e-01 5.76993346e-01 -1.86841249e-01 -7.00557828e-01
-1.08991253e+00 -9.30595696e-01 -8.03343892e-01 -5.38932920e-01
1.02060628e+00 -1.05334878e+00 -2.98655391e-01 3.19631994e-01
-8.48238230e-01 -3.95739496e-01 -4.23213810e-01 7.34888256e-01
-6.85851932e-01 6.62290514e-01 -4.24868971e-01 -1.19024265e+00
-2.89665610e-01 -1.27925777e+00 1.30239010e+00 1.12265341e-01
-2.46379346e-01 -8.88500690e-01 7.21905902e-02 7.82332599e-01
-1.62190031e-02 1.67052954e-01 2.28396326e-01 -7.13366032e-01
-7.90089011e-01 -3.99765708e-02 -3.94172519e-02 5.37198305e-01
2.89187673e-02 -1.57496072e-02 -8.98220122e-01 -4.20087665e-01
-2.40092784e-01 -5.44765711e-01 1.00480223e+00 5.95233381e-01
1.25903583e+00 1.14247845e-02 -2.26633534e-01 3.38992208e-01
1.45377970e+00 4.93242979e-01 6.57999694e-01 2.03594491e-01
7.07953274e-01 4.14364159e-01 1.30849504e+00 6.24001324e-01
1.97331101e-01 1.05950928e+00 2.45456547e-01 -8.86372253e-02
-3.74781698e-01 1.83230311e-01 4.77272093e-01 6.38294578e-01
-2.76459485e-01 -5.52566588e-01 -8.81906688e-01 5.19217670e-01
-2.20135999e+00 -1.12794685e+00 -1.40675381e-02 2.36141253e+00
8.67128551e-01 1.91836044e-01 2.78395593e-01 4.21245515e-01
8.69448602e-01 1.03473768e-01 -5.46965420e-01 6.46284595e-02
6.44055158e-02 8.28501880e-02 5.76731622e-01 5.98743141e-01
-1.45288599e+00 7.05676377e-01 5.20272636e+00 1.18803549e+00
-6.16002500e-01 6.88102484e-01 8.15168679e-01 -3.84943068e-01
4.56122339e-01 -4.82262671e-02 -1.03021216e+00 7.84085095e-01
8.23988378e-01 2.65002877e-01 1.06552895e-02 9.01375949e-01
5.69767416e-01 -6.90389812e-01 -9.92710769e-01 1.09654856e+00
2.06332982e-01 -1.22217023e+00 -1.73168272e-01 -9.98565480e-02
6.53618574e-01 -1.03091896e-01 -3.13559264e-01 1.48719504e-01
5.28492257e-02 -2.98664153e-01 9.01725113e-01 5.77175975e-01
7.00772583e-01 -6.50014102e-01 7.56755114e-01 2.19765320e-01
-1.33001268e+00 -1.86403021e-01 -1.29036784e-01 2.32218906e-01
5.17285526e-01 6.10290110e-01 -7.42962539e-01 5.67811310e-01
8.29433858e-01 8.12698662e-01 -6.27222359e-01 1.19429374e+00
-4.48461249e-02 9.59351063e-01 -2.00579703e-01 4.02501374e-01
5.71944714e-02 -4.02218811e-02 6.05807304e-01 1.39182580e+00
2.08491087e-01 7.66342431e-02 5.25536895e-01 3.77052456e-01
-9.03771743e-02 7.13488534e-02 -1.02618881e-01 6.79083392e-02
7.10693061e-01 1.24628544e+00 -1.05504584e+00 -5.11136353e-01
-4.99310553e-01 1.11715949e+00 3.82704765e-01 2.67858744e-01
-1.12708390e+00 -1.37780130e-01 -1.02243656e-02 2.53059287e-02
4.45778102e-01 -1.06492564e-01 7.62263238e-02 -1.03654385e+00
-4.04742546e-02 -5.14731705e-01 7.44269550e-01 -6.13016009e-01
-6.95874929e-01 3.50569457e-01 3.18860382e-01 -1.37617135e+00
-9.51153189e-02 -1.08531609e-01 -2.47706383e-01 3.96362543e-01
-1.32125485e+00 -1.07401061e+00 -4.07823771e-01 5.01359403e-01
1.14208698e+00 -1.81370974e-01 4.64125544e-01 8.01379383e-01
-9.91969585e-01 6.63017392e-01 -1.55765712e-01 -2.78035104e-02
8.80679309e-01 -1.31013751e+00 -2.65680701e-01 1.07653236e+00
3.31491292e-01 2.38780439e-01 3.24029595e-01 -7.11317182e-01
-7.35767305e-01 -1.23177183e+00 7.33175695e-01 -2.89890289e-01
1.54785156e-01 -2.48558402e-01 -5.61623752e-01 6.61139607e-01
1.87571153e-01 1.54450566e-01 9.72486317e-01 -2.34652579e-01
-7.60280117e-02 -9.46212262e-02 -1.30594933e+00 1.34190649e-01
9.77862000e-01 -2.78649211e-01 -1.71448603e-01 4.13662583e-01
6.02083921e-01 -2.27212444e-01 -7.02853501e-01 5.67805290e-01
4.31513011e-01 -1.08111393e+00 6.99810803e-01 -1.26255125e-01
1.03114821e-01 -5.26090443e-01 -3.22932959e-01 -7.27645576e-01
-3.46367627e-01 -5.05843639e-01 -5.79268754e-01 1.49271941e+00
4.55406070e-01 -1.39912367e-01 8.70345473e-01 5.59080243e-01
-1.89725697e-01 -9.80162561e-01 -8.27649295e-01 -7.30854094e-01
-7.94031262e-01 -4.29058880e-01 -1.02165081e-01 8.59050393e-01
-3.62220913e-01 1.32727474e-01 -4.65952724e-01 -4.18336913e-02
7.66756594e-01 -4.72195476e-01 4.15026456e-01 -9.87759292e-01
-5.33813298e-01 1.30154952e-01 -4.35344189e-01 -8.55233550e-01
1.39269922e-02 -3.98960799e-01 1.27528518e-01 -1.35804486e+00
5.22166967e-01 -4.75834370e-01 -6.21955872e-01 6.95165873e-01
-2.32454985e-01 5.67434728e-01 1.97993293e-01 4.63422418e-01
-1.46523559e+00 3.45072120e-01 6.55842900e-01 9.13990662e-02
-1.71965882e-01 -1.46566918e-02 -1.26239792e-01 7.82216609e-01
8.46196353e-01 -5.11899650e-01 -7.67107427e-01 -3.72181237e-01
5.05452082e-02 -1.38752952e-01 1.40994877e-01 -1.18855286e+00
1.57034352e-01 3.26431245e-02 1.65160120e-01 -7.90897608e-01
5.32293200e-01 -9.35164273e-01 9.60315466e-02 2.38549665e-01
-6.80591524e-01 -1.01806797e-01 -3.48777883e-02 1.05118084e+00
-2.67114103e-01 -4.82389569e-01 1.11476743e+00 -2.39588320e-01
-1.07620966e+00 3.05194139e-01 -6.30787686e-02 -1.49576981e-02
1.53534365e+00 -4.08985943e-01 -4.93700542e-02 -1.74411654e-01
-8.10209870e-01 1.66467994e-01 3.55010331e-01 1.24323264e-01
2.35763371e-01 -1.28916216e+00 -5.52495360e-01 -3.37639958e-01
1.37903720e-01 3.83071229e-02 3.58908236e-01 1.27254272e+00
-6.08651519e-01 3.98060143e-01 1.75481692e-01 -7.68977046e-01
-1.85250318e+00 2.71947443e-01 5.51497564e-02 -4.04739678e-01
-3.00653458e-01 1.25904095e+00 4.07495582e-03 1.18769705e-01
7.15705812e-01 4.85344529e-02 -3.84446084e-01 3.18573564e-01
5.06857872e-01 9.82496798e-01 9.82549042e-02 -8.42387080e-01
-4.12491143e-01 2.83941388e-01 -2.19156876e-01 -1.57674626e-01
1.09714830e+00 -4.28646386e-01 1.46529198e-01 5.84013760e-01
1.13192379e+00 -7.76846632e-02 -1.53035378e+00 -2.86939561e-01
2.88361341e-01 -8.25406611e-01 3.66381705e-01 -8.35218370e-01
-1.05793083e+00 5.60044646e-01 9.05965269e-01 -7.63651952e-02
1.31219649e+00 -1.26595512e-01 6.59673750e-01 2.75354385e-01
3.96311700e-01 -1.57993567e+00 2.59500980e-01 2.61043385e-02
5.17863452e-01 -1.42937469e+00 3.02280396e-01 -6.03699327e-01
-8.21894825e-01 6.05210185e-01 6.16181254e-01 1.68795228e-01
3.78870398e-01 1.60844252e-01 -1.89012453e-01 -3.94164659e-02
-8.60785484e-01 -2.60501385e-01 2.73682475e-01 2.44606391e-01
5.66866875e-01 -1.45209238e-01 -5.54166496e-01 1.16272926e-01
8.11587512e-01 2.78429717e-01 3.04058254e-01 1.32061982e+00
-5.58409333e-01 -1.21487021e+00 -1.71026915e-01 5.18541753e-01
-7.67770886e-01 -1.81008317e-02 -2.05654070e-01 5.04788578e-01
5.27866244e-01 1.01872373e+00 9.00778323e-02 -1.77149951e-01
-1.23005942e-01 2.49247059e-01 4.29659754e-01 -8.94469082e-01
-3.13771784e-01 4.37112570e-01 4.18473184e-01 -7.73656070e-01
-1.28884959e+00 -8.82450700e-01 -1.14619076e+00 1.99269831e-01
-6.73836470e-01 1.06912039e-01 3.98231119e-01 7.12979853e-01
5.64325452e-01 3.16839516e-01 4.60205734e-01 -7.88732648e-01
-1.67448550e-01 -9.40256536e-01 -6.06282115e-01 2.89488643e-01
5.99919148e-02 -9.94437456e-01 -3.60192597e-01 5.51500082e-01] | [8.483945846557617, 0.5685277581214905] |
3e5d3ca3-eff0-4c0d-89a8-cf0a19e08329 | comparing-approaches-for-automatic-question | null | null | https://aclanthology.org/S17-1013 | https://aclanthology.org/S17-1013.pdf | Comparing Approaches for Automatic Question Identification | Collecting spontaneous speech corpora that are open-ended, yet topically constrained, is increasingly popular for research in spoken dialogue systems and speaker state, inter alia. Typically, these corpora are labeled by human annotators, either in the lab or through crowd-sourcing; however, this is cumbersome and time-consuming for large corpora. We present four different approaches to automatically tagging a corpus when general topics of the conversations are known. We develop these approaches on the Columbia X-Cultural Deception corpus and find accuracy that significantly exceeds the baseline. Finally, we conduct a cross-corpus evaluation by testing the best performing approach on the Columbia/SRI/Colorado corpus. | ['Kara Schechtman', 'Sarah Ita Levitan', 'Angel Maredia', 'Julia Hirschberg'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['cross-corpus'] | ['computer-vision'] | [-2.27850080e-01 3.28150541e-01 -9.28088427e-02 -6.43493176e-01
-1.46248996e+00 -1.17227221e+00 6.51658118e-01 -1.60425305e-01
-5.63755274e-01 1.02269971e+00 5.83831131e-01 -1.90130875e-01
5.66440105e-01 1.84779659e-01 -5.67996651e-02 -3.12106252e-01
-6.36140034e-02 6.33473456e-01 1.53922334e-01 -3.55545908e-01
2.89702475e-01 2.74140090e-02 -8.32907498e-01 2.65648276e-01
8.66343617e-01 3.19644362e-01 -1.87061116e-01 7.63086975e-01
-1.70582071e-01 9.34484243e-01 -1.27726817e+00 -6.84487939e-01
-1.39819875e-01 -4.29608285e-01 -1.48490810e+00 2.58319288e-01
5.24013758e-01 -2.93317705e-01 -8.62571523e-02 1.05532742e+00
4.74865466e-01 1.75225034e-01 3.07991326e-01 -1.09131706e+00
-3.91653776e-01 8.25628817e-01 -2.76443809e-01 3.27785313e-01
6.19724631e-01 -8.18340182e-02 8.85522604e-01 -5.81691980e-01
7.04761028e-01 1.30693161e+00 3.62739414e-01 1.04063451e+00
-1.02036870e+00 -6.70603335e-01 1.13771848e-01 -1.89461544e-01
-1.38483810e+00 -1.06436062e+00 8.18345308e-01 -3.79043967e-01
8.08429360e-01 4.04862642e-01 1.02668479e-01 1.82903469e+00
-5.27845144e-01 1.01510262e+00 1.00512266e+00 -4.14437950e-01
3.34984004e-01 4.21408147e-01 3.85854542e-01 3.25418711e-01
-2.22117975e-01 -3.14485610e-01 -9.23085749e-01 -6.41434431e-01
7.35036358e-02 -7.36416578e-01 -6.66908681e-01 2.67014861e-01
-9.33168411e-01 1.07088649e+00 -3.63313258e-01 4.63142246e-01
5.35843447e-02 -4.09221917e-01 7.41374254e-01 4.39376622e-01
8.79382670e-01 8.42482030e-01 -5.59663773e-01 -8.90888095e-01
-9.00333762e-01 2.92916059e-01 1.60527170e+00 9.76763546e-01
1.98107466e-01 8.10958911e-03 1.46757886e-01 1.32962728e+00
2.84378290e-01 4.20926213e-01 5.86926103e-01 -1.12303185e+00
6.38383090e-01 2.46199518e-02 4.62552756e-01 -6.47814035e-01
-2.73794353e-01 1.99407622e-01 -3.14599305e-01 -3.73419851e-01
7.59711385e-01 -7.30509996e-01 -4.43458289e-01 1.70648479e+00
4.01994944e-01 -4.71521951e-02 4.80240047e-01 8.98781717e-01
8.13708127e-01 7.26084530e-01 1.96989551e-01 -4.84335005e-01
1.43482709e+00 -1.11599040e+00 -1.10572207e+00 -4.14241731e-01
8.76794040e-01 -9.22270536e-01 1.17977333e+00 5.60876071e-01
-9.42168295e-01 1.37825802e-01 -7.81348586e-01 -1.47692516e-01
-2.22127751e-01 -9.79200304e-02 2.23481372e-01 1.06062233e+00
-9.34609890e-01 8.15962031e-02 -6.88458264e-01 -6.78679526e-01
7.48771131e-02 -4.43231091e-02 -4.10859436e-01 2.12740123e-01
-1.39200473e+00 1.06424773e+00 -6.31888732e-02 -2.18595371e-01
-8.15519094e-01 -2.48575091e-01 -9.61129606e-01 -1.41262829e-01
3.86384040e-01 3.84300977e-01 1.95255280e+00 -8.00722539e-01
-1.99586284e+00 1.04274857e+00 -3.61411959e-01 -3.18740726e-01
2.26683646e-01 -5.41595936e-01 -7.37596393e-01 2.77868450e-01
1.57239050e-01 3.71069968e-01 5.60355067e-01 -1.24362767e+00
-4.10572469e-01 -2.65379339e-01 -2.15057731e-02 2.48598546e-01
-4.13336694e-01 6.69740081e-01 -1.16972663e-01 -3.07066351e-01
-2.39747971e-01 -9.68149722e-01 2.02906042e-01 -5.97014070e-01
-7.34030902e-01 -5.23045182e-01 1.12773263e+00 -9.98994470e-01
1.22837019e+00 -2.35898781e+00 -2.45257378e-01 -1.51954800e-01
1.45642264e-02 6.21652305e-01 8.65343213e-02 6.95760190e-01
2.12562561e-01 5.38299859e-01 -6.97464347e-02 -6.13933444e-01
2.37516873e-02 2.50997424e-01 -4.62430656e-01 3.66042048e-01
9.92549658e-02 4.16815966e-01 -9.12826478e-01 -5.78321099e-01
-1.73664689e-01 3.13099027e-02 -3.23661983e-01 4.91971642e-01
-7.45748729e-02 7.11405277e-01 -4.50915277e-01 4.79407579e-01
3.21542859e-01 8.64172801e-02 3.30574542e-01 4.95897025e-01
-2.59599507e-01 1.00400102e+00 -5.56152582e-01 1.81362700e+00
-3.95959616e-01 7.90754676e-01 6.94169879e-01 -7.01876819e-01
9.06379998e-01 8.28972280e-01 -1.29294857e-01 -8.97747874e-02
6.77045211e-02 2.53821611e-01 -4.89020087e-02 -7.80145586e-01
8.10500324e-01 -2.27423832e-01 -6.09349608e-01 8.18212032e-01
1.56484857e-01 -5.88275254e-01 4.97582816e-02 4.39178407e-01
1.04051697e+00 -3.68626356e-01 4.08118963e-01 -3.69785815e-01
4.91611660e-01 2.37588495e-01 5.94895065e-01 4.83135164e-01
-7.95555115e-01 5.76628149e-01 7.29252875e-01 -1.61535125e-02
-7.95390666e-01 -5.87951064e-01 -1.46811455e-01 1.22345805e+00
-1.74072295e-01 -3.23794931e-01 -1.20137513e+00 -1.00215602e+00
-4.87034351e-01 1.04238129e+00 -1.90552026e-01 -4.14322875e-02
-5.02042651e-01 -4.23721522e-02 1.22386765e+00 9.50889736e-02
4.57061410e-01 -8.79199505e-01 -1.37553334e-01 8.89227763e-02
-8.44722748e-01 -1.42465127e+00 -8.00636947e-01 6.05316972e-03
-1.76678121e-01 -7.66698599e-01 -4.13784713e-01 -7.72809088e-01
1.62188590e-01 1.04337007e-01 9.89844620e-01 2.49971822e-02
2.53239304e-01 1.57025516e-01 -5.58431506e-01 -3.73315811e-01
-6.70535386e-01 4.27127808e-01 1.06789179e-01 -9.07725617e-02
6.09245956e-01 -2.51797169e-01 -6.71143755e-02 4.60775614e-01
-4.29508477e-01 -5.66394329e-01 -3.94580960e-02 8.44949841e-01
-4.13885176e-01 -6.43903255e-01 8.98702025e-01 -1.09564304e+00
9.91796970e-01 -5.43227553e-01 -4.19036716e-01 1.03049032e-01
1.59305453e-01 -2.89061159e-01 5.12656510e-01 -4.91074383e-01
-1.32676804e+00 -2.17627421e-01 -4.87023532e-01 -8.60742405e-02
-6.01176441e-01 3.46800119e-01 -3.05863917e-01 7.96254873e-02
7.95936942e-01 -9.70913395e-02 -7.77238533e-02 -5.45410395e-01
1.76674157e-01 1.46526301e+00 5.67217648e-01 -6.51845872e-01
5.15459538e-01 1.99305508e-02 -1.07491100e+00 -1.18490458e+00
-1.11324024e+00 -6.93066120e-01 -5.47199905e-01 -1.44679785e-01
8.30481529e-01 -7.89021611e-01 -7.82564938e-01 4.77992058e-01
-1.54752982e+00 -3.79043221e-01 1.00596890e-01 5.72838664e-01
-1.22180015e-01 5.47524333e-01 -8.34260046e-01 -1.09733951e+00
-2.46853098e-01 -9.15558338e-01 1.06869519e+00 5.28196171e-02
-1.03614938e+00 -1.10277724e+00 3.77162486e-01 8.55113447e-01
2.53862262e-01 -1.67765781e-01 6.20420218e-01 -1.59242809e+00
1.24355145e-01 -2.28772134e-01 2.38405317e-01 3.68105054e-01
-1.88890416e-02 5.64245321e-02 -1.34323299e+00 -1.00658476e-01
1.64295956e-01 -1.13200378e+00 1.54617086e-01 -1.93982467e-01
6.51893258e-01 -6.42953396e-01 -1.59683406e-01 -2.00510979e-01
5.82556605e-01 4.12548751e-01 2.58594990e-01 -1.27119139e-01
3.64401907e-01 9.97946322e-01 5.79487920e-01 4.80142713e-01
6.27350092e-01 5.75502217e-01 -1.27212986e-01 3.71497333e-01
3.25727731e-01 -3.02395552e-01 5.45069098e-01 1.32014012e+00
4.65434253e-01 -7.39095986e-01 -1.01276445e+00 8.04398894e-01
-1.64195096e+00 -9.92510200e-01 -3.91738079e-02 2.04380035e+00
1.30508733e+00 -1.13075286e-01 1.88764200e-01 -1.60584509e-01
9.13074970e-01 3.61230582e-01 -6.74876869e-02 -8.33259106e-01
-3.10828127e-02 1.02984728e-02 -8.25224593e-02 8.94613266e-01
-1.02791727e+00 1.29560530e+00 7.17142963e+00 5.42679310e-01
-1.04974115e+00 4.99150455e-01 6.71725988e-01 -2.27372885e-01
-3.72435749e-02 -7.92589784e-02 -8.67693186e-01 4.21329886e-01
1.39530408e+00 -2.33218133e-01 3.82358491e-01 8.49783838e-01
2.70757496e-01 -3.20302308e-01 -1.12005198e+00 7.99765885e-01
4.04721856e-01 -8.20640922e-01 -6.00479305e-01 3.56367044e-03
5.75868130e-01 6.14680387e-02 -3.55771214e-01 4.46698427e-01
7.92688549e-01 -7.96830893e-01 5.83595455e-01 -2.66015232e-01
5.20164013e-01 -6.32349014e-01 8.00135672e-01 7.23005176e-01
-2.75439680e-01 3.00384521e-01 -2.84061003e-02 2.36152601e-03
5.06128788e-01 2.17005178e-01 -1.01931846e+00 8.39267205e-03
6.14705503e-01 7.55376965e-02 -7.53162727e-02 6.16786301e-01
-4.56723243e-01 1.21437430e+00 -3.31402779e-01 -4.73622590e-01
3.16644400e-01 1.30471826e-01 8.76594007e-01 1.24770868e+00
-1.39740586e-01 4.55705583e-01 4.93180275e-01 6.12795591e-01
-4.12521839e-01 2.15831324e-02 -7.33580709e-01 -2.86143184e-01
8.21174562e-01 1.35807979e+00 -1.85116097e-01 -3.62267852e-01
-3.48789394e-01 9.09571826e-01 6.18688166e-01 3.86045992e-01
-4.85426396e-01 -3.52442741e-01 7.50475168e-01 -2.77668118e-01
-6.83898404e-02 -4.09245819e-01 -3.90669033e-02 -1.20954359e+00
-1.66333281e-02 -1.22698402e+00 3.12109739e-01 -5.42856693e-01
-1.41401649e+00 9.13323343e-01 -2.62435526e-02 -7.51239419e-01
-8.97700608e-01 -2.24014148e-01 -5.15396237e-01 7.44570911e-01
-1.23257089e+00 -6.31037295e-01 -2.64525460e-03 5.73275566e-01
8.60318542e-01 -1.30231500e-01 1.24493837e+00 2.89816916e-01
-8.76715124e-01 7.47674942e-01 -1.00042760e-01 4.98067737e-01
1.09704161e+00 -1.10684383e+00 4.03016776e-01 7.19581902e-01
6.04071766e-02 6.93711340e-01 6.97017848e-01 -5.57772338e-01
-8.92591000e-01 -7.18297839e-01 1.44285202e+00 -4.47176218e-01
9.62668240e-01 -8.98967385e-01 -1.36594665e+00 7.95159817e-01
8.58807087e-01 -2.50893354e-01 1.11274052e+00 5.30454099e-01
-4.42485034e-01 4.00050581e-01 -1.32032800e+00 4.16109830e-01
7.62837887e-01 -7.87214279e-01 -9.04170454e-01 5.22136807e-01
7.46676505e-01 -4.95565414e-01 -8.26200962e-01 -2.19939679e-01
2.06560493e-01 -5.36860585e-01 5.32573462e-02 -7.44304836e-01
2.65572309e-01 1.13174632e-01 6.46090880e-03 -1.62197542e+00
2.72849619e-01 -1.35574162e+00 4.19567287e-01 1.85596836e+00
6.67050540e-01 -7.05643594e-01 5.26119471e-01 1.14533460e+00
-3.28872740e-01 3.26515734e-02 -1.16939950e+00 -6.97174847e-01
3.32383335e-01 -2.73974657e-01 3.05257171e-01 1.49572682e+00
9.99247730e-01 8.58637094e-01 -4.97949213e-01 6.49254993e-02
1.15116984e-01 -3.87730002e-01 8.83428931e-01 -1.04709041e+00
1.23786435e-01 -1.06121570e-01 -2.33783945e-02 -1.15242839e+00
7.96808720e-01 -2.96905190e-01 6.23206735e-01 -6.90251112e-01
-2.08944634e-01 -3.43392491e-01 6.24415338e-01 6.02046251e-01
-5.76355122e-02 1.32067902e-02 -2.22026825e-01 2.09722370e-01
-7.20777988e-01 5.94470501e-01 7.79020071e-01 6.46734014e-02
-2.25889653e-01 -2.75940672e-02 -7.47501135e-01 7.64807463e-01
8.50879312e-01 -5.57888091e-01 -2.28498593e-01 -5.10257781e-01
-2.82905579e-01 4.58072454e-01 -3.76702361e-02 -5.39433956e-01
3.03084612e-01 -1.30743608e-01 -4.76787686e-01 -3.85669440e-01
4.80978012e-01 -4.67424303e-01 -6.22862518e-01 -1.58304334e-01
-6.06408417e-01 -2.44995236e-01 2.01763451e-01 2.33284816e-01
-6.03899360e-01 -6.20641589e-01 8.31317246e-01 -1.50085732e-01
-4.17364180e-01 -1.15264170e-01 -8.66983414e-01 8.14917088e-01
8.29888284e-01 1.66497484e-01 -7.72030652e-01 -9.67343748e-01
-4.99521583e-01 2.19212011e-01 2.87185103e-01 4.76499438e-01
1.58648267e-01 -8.67113292e-01 -8.19990993e-01 -1.49512962e-02
2.18078583e-01 -2.91617453e-01 3.79740703e-03 6.53296411e-01
-2.72898436e-01 3.58208328e-01 2.42813617e-01 -4.75140631e-01
-1.43791246e+00 1.56791300e-01 2.28763610e-01 3.28366682e-02
-2.17800707e-01 9.50813949e-01 -5.90038709e-02 -7.10327387e-01
3.11040640e-01 1.29601985e-01 -2.32417122e-01 -4.03533280e-02
7.19850361e-01 1.89621031e-01 -8.96660052e-03 -9.60685968e-01
-4.38034058e-01 -3.07445198e-01 -3.25709641e-01 -7.08048761e-01
8.46141875e-01 -4.92438376e-01 -7.27108121e-02 7.91602850e-01
1.30898321e+00 4.52203363e-01 -8.21512699e-01 -3.01053613e-01
2.58613497e-01 -3.82907331e-01 -1.57063097e-01 -7.46720850e-01
-6.56777859e-01 7.76937962e-01 -1.19682848e-01 7.29802847e-01
4.28110301e-01 1.48320392e-01 9.72890198e-01 5.51495075e-01
2.61483163e-01 -1.34190583e+00 4.50127386e-02 9.43756342e-01
1.00098562e+00 -1.32156527e+00 -4.46876109e-01 -6.53574884e-01
-1.13327718e+00 8.18333507e-01 7.17572927e-01 3.01006973e-01
5.88591933e-01 3.30452651e-01 7.53751218e-01 -3.02911550e-01
-8.72332215e-01 1.89077467e-01 -4.92539480e-02 5.84603012e-01
7.72046268e-01 9.30705145e-02 -4.18285280e-01 5.93005598e-01
-5.94337940e-01 -6.27905190e-01 8.33406925e-01 9.96882915e-01
-2.84017354e-01 -1.09440529e+00 -3.93463373e-01 -3.75693887e-02
-8.67481828e-01 -2.74155676e-01 -1.23478496e+00 7.77369082e-01
-6.41408741e-01 1.81185555e+00 2.82790307e-02 -2.11979479e-01
2.32933491e-01 6.19227827e-01 -1.45521954e-01 -9.95319843e-01
-9.40409601e-01 9.86650661e-02 1.14372981e+00 -2.48838484e-01
-3.62399995e-01 -1.10469913e+00 -1.00089693e+00 -3.66261631e-01
-7.40162611e-01 7.40519285e-01 6.07607901e-01 1.17582011e+00
2.48721957e-01 -3.19431514e-01 9.08306181e-01 -5.53354144e-01
-6.31297648e-01 -1.53319883e+00 -4.96847272e-01 4.69377965e-01
2.84426242e-01 -3.32695216e-01 -6.09285355e-01 5.84817268e-02] | [12.91567325592041, 7.829317569732666] |
9f93db06-6e1d-4054-a5af-8dfb77b383b6 | scaf-skip-connections-in-auto-encoder-for | null | null | https://link.springer.com/chapter/10.1007/978-3-031-06427-2_36 | https://hal.archives-ouvertes.fr/hal-03687091/file/SCAF_submitted%20%282%29.pdf | SCAF: Skip-Connections in Auto-encoder for Face alignment with few annotated data | Supervised face alignment methods need large amounts of training data to achieve good performance in terms of accuracy and generalization. However face alignment datasets rarely exceed a few thousand samples making these methods prone to overfitting on the specific training dataset. Semi-supervised methods like TS3 or 3FabRec have emerged to alleviate this issue by using labeled and unlabeled data during the training. In this paper we propose Skip-Connections in Auto-encoder for Face alignment (SCAF), we build on 3FabRec by adding skip-connections between the encoder and the decoder. These skip-connections lead to better landmark predictions, especially on challenging examples. We also apply for the first time active learning to the face alignment task and introduce a new acquisition function, the Negative Neighborhood Magnitude, specially designed to assess the quality of heatmaps. These two proposals show their effectiveness on several face alignment datasets when training with limited data. | ['Bertrand Coüasnon', 'Yann Ricquebourg', 'Christian Raymond', 'Philippe-Henri Gosselin', 'Martin Dornier'] | 2022-05-15 | null | null | null | iciap-2022-5 | ['face-alignment'] | ['computer-vision'] | [ 2.00269714e-01 4.02507275e-01 -2.46854365e-01 -1.00589395e+00
-6.75761640e-01 -2.11075798e-01 6.91756368e-01 -1.39759108e-01
-3.66991758e-01 6.22739792e-01 6.75319582e-02 2.85225123e-01
-2.14198306e-02 -6.24389291e-01 -7.52417266e-01 -5.60390353e-01
-1.04800031e-01 8.33187103e-01 -1.01064831e-01 -4.49136645e-02
5.44382259e-02 5.46191394e-01 -1.45917058e+00 1.39435023e-01
7.46086538e-01 8.37687790e-01 1.19621746e-01 1.54463008e-01
-2.78820038e-01 5.43912947e-01 -3.51069629e-01 -6.40576303e-01
4.27326918e-01 -3.32484007e-01 -7.02842236e-01 3.73430224e-03
8.81231368e-01 -3.22362483e-01 1.65865853e-01 7.07401693e-01
8.28812480e-01 -1.69682041e-01 5.37040174e-01 -1.34000850e+00
-2.11374745e-01 5.83982527e-01 -5.43090999e-01 8.81278329e-03
3.19434226e-01 6.56914487e-02 9.10500228e-01 -1.11023569e+00
5.74494243e-01 1.15880656e+00 1.09851396e+00 9.06145751e-01
-1.46096802e+00 -7.41030574e-01 -2.49010250e-01 3.61702472e-01
-1.31097889e+00 -1.03716993e+00 7.90138245e-01 -4.53381091e-01
9.70850468e-01 -9.49799269e-03 4.58753467e-01 1.21816659e+00
-1.00587018e-01 5.47669649e-01 9.67293561e-01 -4.89159465e-01
1.95682868e-01 1.16379067e-01 -1.75892472e-01 7.12114871e-01
-4.94014435e-02 2.14564383e-01 -7.13365316e-01 -5.10815531e-02
6.74052477e-01 -3.23702395e-01 7.20429569e-02 -6.45570219e-01
-7.98435032e-01 7.48018146e-01 5.22397161e-01 2.73148626e-01
-1.98290333e-01 -8.26689750e-02 2.68392652e-01 2.77174324e-01
5.49357295e-01 4.56321895e-01 -3.66212219e-01 4.06391807e-02
-1.29186904e+00 -9.98274982e-02 4.26737815e-01 7.30820060e-01
9.84210193e-01 9.86351669e-02 -9.10541341e-02 9.73792493e-01
3.25835913e-01 2.68564492e-01 3.32370549e-01 -4.85680133e-01
5.58792949e-01 8.31235766e-01 -2.34689549e-01 -8.22822392e-01
-5.96287668e-01 -4.52955455e-01 -7.06508636e-01 3.64677519e-01
3.68501127e-01 3.53925489e-02 -8.60481262e-01 1.83657265e+00
3.22903901e-01 3.27050984e-01 -2.59157896e-01 6.69239759e-01
8.26132357e-01 3.41522932e-01 1.30748823e-01 -2.79384941e-01
7.12552726e-01 -1.05585718e+00 -6.43947423e-01 -2.45087013e-01
7.98033834e-01 -8.21584046e-01 1.02690828e+00 1.30006894e-01
-1.04319429e+00 -9.32740510e-01 -9.89577889e-01 9.06704143e-02
-2.99988061e-01 4.75542545e-01 3.69452238e-01 7.57626593e-01
-1.21990752e+00 7.86093473e-01 -8.50430548e-01 -2.99273670e-01
9.02850926e-01 1.03182387e+00 -8.37947667e-01 2.34191507e-01
-7.60277152e-01 1.00118208e+00 8.53464454e-02 2.25607023e-01
-7.99540460e-01 -6.12355232e-01 -6.38258100e-01 5.28766327e-02
1.64596409e-01 -4.61679161e-01 7.32400358e-01 -1.34891582e+00
-1.69198298e+00 1.04919767e+00 -2.77410131e-02 -5.07012427e-01
4.50000048e-01 -3.81574422e-01 -3.00541759e-01 -4.68384624e-02
-1.44422695e-01 1.16355085e+00 1.00023198e+00 -9.87596393e-01
1.00766100e-01 -6.72139347e-01 -1.97685629e-01 -4.64549698e-02
-5.64622164e-01 -6.98447004e-02 -2.04716921e-01 -5.24444938e-01
-1.73334613e-01 -9.47948694e-01 -5.06697856e-02 -5.11423871e-02
-2.69278347e-01 -1.88204959e-01 9.22868490e-01 -6.15918875e-01
8.80050838e-01 -1.94450986e+00 2.38123536e-01 3.02034706e-01
-1.02098338e-01 5.87217212e-01 -5.58844566e-01 3.26751679e-01
-5.28957188e-01 -3.01179975e-01 -2.86176354e-01 -6.99828565e-01
-2.18352363e-01 1.59100816e-01 -1.53234899e-01 4.75890845e-01
6.52924359e-01 9.75004494e-01 -5.02385736e-01 -6.42522097e-01
1.16707116e-01 5.41333437e-01 -7.22177446e-01 4.29721832e-01
-1.83659390e-01 7.30698645e-01 8.86760652e-03 3.83542120e-01
6.98161542e-01 -1.14103295e-01 1.00849234e-01 -1.75643891e-01
5.74903078e-02 1.99710399e-01 -8.26602757e-01 1.97520411e+00
-6.49652600e-01 5.61366439e-01 -3.03097576e-01 -1.05272424e+00
1.22601402e+00 5.18511117e-01 6.84183180e-01 -7.68915832e-01
1.78310454e-01 1.00130379e-01 1.56797260e-01 -3.69959563e-01
1.53944105e-01 1.16185285e-01 4.88223046e-01 4.11219716e-01
6.19873881e-01 2.65980244e-01 1.34822041e-01 -8.11430067e-02
7.03559458e-01 5.46214938e-01 2.68138856e-01 -3.43460470e-01
7.82407105e-01 -3.97872835e-01 5.59522748e-01 1.62897423e-01
1.19594999e-01 7.76900232e-01 3.53357404e-01 -5.38440704e-01
-1.24427760e+00 -6.12728536e-01 -2.81019628e-01 8.92699122e-01
-4.03785110e-01 -7.07252383e-01 -1.08320558e+00 -1.05130351e+00
-1.38364762e-01 3.71240199e-01 -7.44005859e-01 -1.74853116e-01
-6.75148904e-01 -7.88609982e-01 4.95334029e-01 5.96812248e-01
5.04707396e-01 -9.79662240e-01 -2.57299840e-01 9.92006287e-02
1.29911229e-01 -1.01084208e+00 -1.32414877e-01 1.33969977e-01
-1.02288008e+00 -1.06475782e+00 -5.70861936e-01 -6.63005590e-01
1.06985795e+00 -3.83175313e-01 1.10749483e+00 2.91111946e-01
-4.41245139e-01 1.67956099e-01 -2.05312639e-01 -1.76943943e-01
-2.74070203e-01 4.26683962e-01 1.67393684e-01 1.56669259e-01
4.05032128e-01 -6.39268994e-01 -3.44775885e-01 4.82189864e-01
-3.88234794e-01 -4.13465723e-02 6.13286614e-01 9.93107200e-01
3.49865854e-01 -6.67432725e-01 7.43229687e-01 -8.60026658e-01
1.32751808e-01 -1.45602554e-01 -7.59082198e-01 3.37002724e-01
-5.61704516e-01 2.00285167e-01 7.52107561e-01 -2.92555213e-01
-7.91117191e-01 5.47005653e-01 -4.08028781e-01 -3.23113263e-01
-1.50943054e-02 -2.68449411e-02 -3.65047961e-01 -4.73151475e-01
8.34678471e-01 -1.50671020e-01 2.79382646e-01 -5.50042868e-01
2.69562565e-02 4.54564154e-01 2.51491308e-01 -3.28683347e-01
8.58139396e-01 2.28086635e-01 1.99928626e-01 -5.96758962e-01
-7.06968546e-01 -1.49423003e-01 -1.20636678e+00 -3.52338433e-01
5.94072819e-01 -7.54683435e-01 -5.42163074e-01 2.17206612e-01
-8.90306950e-01 -4.21431810e-01 -2.72487640e-01 4.55955416e-01
-7.74815261e-01 2.28096768e-01 -1.24154821e-01 -6.29667401e-01
-4.67179537e-01 -1.22933722e+00 1.11855292e+00 1.76220745e-01
-3.06332320e-01 -9.70169067e-01 3.14738959e-01 3.39554667e-01
6.05219364e-01 -6.26131669e-02 7.55535781e-01 -8.67484868e-01
-3.25796425e-01 -1.47136688e-01 1.36567608e-01 4.45096344e-01
2.55455881e-01 1.60249416e-02 -1.29477835e+00 -4.26757991e-01
-2.17207164e-01 -5.70612907e-01 7.68605053e-01 1.24430001e-01
1.20680642e+00 -2.36761093e-01 -3.71026278e-01 5.58843791e-01
1.13827634e+00 4.83217761e-02 9.31198895e-01 -2.13812534e-02
5.32219291e-01 8.50015104e-01 4.67212856e-01 3.64096090e-02
-7.66083524e-02 1.31580758e+00 3.04634929e-01 -2.33046442e-01
-3.37777287e-01 -2.57707864e-01 4.41792637e-01 7.46204436e-01
-1.63079649e-01 9.16208699e-02 -1.05528820e+00 1.39746532e-01
-1.48223543e+00 -9.80518758e-01 1.74433723e-01 2.47263598e+00
8.90266955e-01 1.04979642e-01 3.07906836e-01 2.79048234e-01
5.22649825e-01 -7.61154369e-02 -3.57531697e-01 -1.96330622e-01
-1.10721998e-01 4.65742081e-01 1.91561803e-01 5.68156004e-01
-1.11786234e+00 1.07521224e+00 6.55840540e+00 7.88728178e-01
-1.46530390e+00 1.90784588e-01 8.03622127e-01 -1.65151395e-02
6.52010832e-03 1.26195163e-01 -9.54376578e-01 4.25242007e-01
8.84831131e-01 5.50877690e-01 1.62225351e-01 9.31618512e-01
-9.41941291e-02 4.41032238e-02 -1.32942533e+00 1.10689414e+00
2.36126766e-01 -1.14756024e+00 4.69815321e-02 2.28319736e-03
6.92652643e-01 -2.40448937e-01 -3.27292411e-03 -5.48813716e-02
-1.61180660e-01 -1.32222879e+00 2.55765080e-01 5.49269080e-01
8.96492362e-01 -6.80195093e-01 7.45095253e-01 4.76039872e-02
-9.37308908e-01 -3.88612747e-02 -4.94214386e-01 2.77641177e-01
1.06243715e-01 4.70323443e-01 -1.23873305e+00 3.28435808e-01
3.34382594e-01 6.82635307e-01 -9.81857955e-01 1.00723994e+00
-5.91386035e-02 4.71847028e-01 -3.75351250e-01 2.33566031e-01
-1.34783685e-01 -3.13682556e-01 3.94166142e-01 7.25510418e-01
2.76381522e-01 -3.52562606e-01 -1.86027735e-01 7.16843843e-01
1.14052957e-02 4.25367981e-01 -5.97137332e-01 4.36731167e-02
2.71651238e-01 1.38043904e+00 -6.13927603e-01 -2.85094995e-02
-3.79812866e-01 9.38603461e-01 5.60452044e-01 -1.25285629e-02
-7.12456167e-01 4.25959229e-02 2.90281206e-01 3.10494393e-01
1.54059500e-01 -1.17819853e-01 -1.80146664e-01 -7.93788493e-01
-4.93259393e-02 -8.22662354e-01 1.38688296e-01 -6.20966971e-01
-1.05738223e+00 9.52601314e-01 -2.06265505e-03 -1.21475780e+00
-4.98337686e-01 -6.89439654e-01 -5.70813358e-01 5.23354948e-01
-1.00781965e+00 -1.32034028e+00 -2.84647942e-01 5.25678635e-01
3.14825237e-01 -8.23296666e-01 1.05563998e+00 5.76659620e-01
-7.12130725e-01 1.03660703e+00 -2.38491505e-01 2.44521901e-01
9.85849321e-01 -1.00314534e+00 5.78824341e-01 5.31098425e-01
5.83713889e-01 5.99111617e-01 4.65829879e-01 -4.94978070e-01
-1.12482071e+00 -7.52785385e-01 8.81818056e-01 -3.94972742e-01
1.27403125e-01 -7.05795228e-01 -7.45992541e-01 6.75194979e-01
2.05825970e-01 1.18307494e-01 8.32368016e-01 4.39137697e-01
-4.38720793e-01 -5.37854314e-01 -1.27913523e+00 3.28515828e-01
1.23152244e+00 -4.27278996e-01 -4.63951439e-01 2.72875845e-01
4.90426377e-04 -9.26893055e-02 -6.77263141e-01 6.94017291e-01
4.91180331e-01 -1.05059636e+00 8.74322653e-01 -6.33921564e-01
2.64128953e-01 -1.45266995e-01 1.74433395e-01 -1.21027923e+00
-1.93512663e-01 -5.29524446e-01 9.83089060e-02 1.43139565e+00
6.39872313e-01 -4.82195407e-01 1.16950953e+00 1.45368978e-01
-8.06825310e-02 -8.34414661e-01 -1.08238089e+00 -5.21432519e-01
-1.55666545e-01 1.97985340e-02 7.57512748e-01 1.01392961e+00
-9.47244093e-02 6.07241750e-01 -4.35464323e-01 -2.22682610e-01
5.04243553e-01 -7.02835694e-02 1.03304768e+00 -1.37707686e+00
6.95062475e-03 -1.02864392e-01 -7.93476105e-01 -7.18258858e-01
3.66746783e-01 -9.73889828e-01 -1.88938901e-01 -1.04656589e+00
1.28422052e-01 -7.52165198e-01 2.33294144e-01 8.91756177e-01
1.73445493e-01 6.53458655e-01 -7.87135512e-02 5.97731322e-02
-2.98183680e-01 7.69088745e-01 9.16015625e-01 -3.96434739e-02
2.01145019e-02 -5.17248176e-03 -5.01936823e-02 5.72909653e-01
8.83687258e-01 -4.31553960e-01 -4.84512508e-01 -4.16840315e-01
3.13948661e-01 -2.11089239e-01 1.82549179e-01 -1.31017578e+00
1.49301752e-01 3.43665987e-01 7.57221162e-01 -5.15924096e-01
5.57761788e-01 -1.09829593e+00 3.28910708e-01 2.75084555e-01
-4.16537970e-01 1.87581256e-01 1.81998521e-01 4.09324132e-02
-3.45798463e-01 -3.20209473e-01 9.76164579e-01 1.51214540e-01
-4.25852448e-01 4.94532526e-01 1.65194079e-01 -2.23467678e-01
8.75958502e-01 -4.10287350e-01 -4.81685847e-02 -1.94615304e-01
-9.92238104e-01 -1.65582374e-02 5.23767352e-01 4.67631519e-01
4.53875482e-01 -1.47323489e+00 -7.36002088e-01 7.19904304e-01
9.34155774e-04 -1.86958700e-01 2.28854671e-01 1.04550803e+00
-5.31762362e-01 3.56624812e-01 -7.88480103e-01 -8.45012426e-01
-1.54564273e+00 4.01185066e-01 4.66389477e-01 -9.32893008e-02
-2.97538340e-01 8.43426943e-01 -1.08810142e-01 -6.37024760e-01
3.31027627e-01 2.70230174e-01 -4.30265397e-01 2.99193859e-01
3.86911273e-01 2.03089908e-01 5.05114436e-01 -9.83694792e-01
-4.16593581e-01 9.16238725e-01 -1.57548919e-01 8.83457735e-02
1.45687866e+00 1.45628288e-01 -2.12063000e-01 1.06050991e-01
1.36112404e+00 1.05431765e-01 -1.37491393e+00 -1.43229067e-01
2.90087432e-01 -4.64222938e-01 -2.53382653e-01 -5.95415950e-01
-1.43901467e+00 1.05632770e+00 1.02449358e+00 -4.59420621e-01
1.08888364e+00 -2.15682715e-01 2.16375098e-01 2.82199889e-01
3.71920139e-01 -1.05799806e+00 3.30068260e-01 2.36223623e-01
9.79583561e-01 -1.26908004e+00 1.85322866e-01 -4.09981757e-01
-4.68900174e-01 1.32203877e+00 9.04910445e-01 -1.02665223e-01
4.91382480e-01 3.42610925e-01 6.52946457e-02 -3.33751366e-02
-6.38104081e-01 -5.58855161e-02 4.55999702e-01 5.47311485e-01
6.27975464e-01 -3.53128791e-01 -1.74464628e-01 1.83910787e-01
-4.72285956e-01 -4.76285517e-02 -8.29504058e-02 6.25189364e-01
-9.39958468e-02 -1.63432348e+00 -1.67533338e-01 3.35379481e-01
-3.41739327e-01 7.37469122e-02 -6.95016623e-01 7.10125625e-01
2.64396489e-01 5.06056905e-01 4.83567715e-01 -5.34431458e-01
1.64305475e-02 3.33923817e-01 9.38706994e-01 -5.22228897e-01
-8.38856339e-01 -7.85460770e-02 4.07823361e-02 -6.02612793e-01
-6.26244664e-01 -5.45273125e-01 -8.49139929e-01 -8.55296105e-02
-5.35719216e-01 2.04670161e-01 5.12468874e-01 8.16819429e-01
5.08765519e-01 7.03833252e-02 8.12647343e-01 -8.02016556e-01
-3.82784992e-01 -1.38497245e+00 -1.87106624e-01 3.56698573e-01
6.67028502e-02 -8.53900492e-01 -2.63516903e-02 3.55412997e-02] | [13.485594749450684, 0.3555440604686737] |
ae26c3cf-e099-4c13-98bc-d73e289c4be5 | modeling-spatio-temporal-human-track | 1806.11008 | null | http://arxiv.org/abs/1806.11008v1 | http://arxiv.org/pdf/1806.11008v1.pdf | Modeling Spatio-Temporal Human Track Structure for Action Localization | This paper addresses spatio-temporal localization of human actions in video.
In order to localize actions in time, we propose a recurrent localization
network (RecLNet) designed to model the temporal structure of actions on the
level of person tracks. Our model is trained to simultaneously recognize and
localize action classes in time and is based on two layer gated recurrent units
(GRU) applied separately to two streams, i.e. appearance and optical flow
streams. When used together with state-of-the-art person detection and
tracking, our model is shown to improve substantially spatio-temporal action
localization in videos. The gain is shown to be mainly due to improved temporal
localization. We evaluate our method on two recent datasets for spatio-temporal
action localization, UCF101-24 and DALY, demonstrating a significant
improvement of the state of the art. | ['Ivan Laptev', 'Guilhem Chéron', 'Anton Osokin', 'Cordelia Schmid'] | 2018-06-28 | null | null | null | null | ['spatio-temporal-action-localization'] | ['computer-vision'] | [ 9.21792090e-02 -4.97302234e-01 -2.78370976e-01 -2.83285026e-02
-5.22329450e-01 -3.58372808e-01 7.73592174e-01 -1.35681719e-01
-6.71476841e-01 5.15460789e-01 8.38236570e-01 3.98717999e-01
1.67114735e-01 -4.16729510e-01 -7.04289317e-01 -3.72526407e-01
-5.03258884e-01 5.24467565e-02 4.86319095e-01 8.03460702e-02
-4.09752689e-02 5.18798172e-01 -1.39251566e+00 6.31236732e-01
6.17738180e-02 5.69463551e-01 -1.94620162e-01 1.13103509e+00
4.25384343e-01 1.63536787e+00 -4.01886970e-01 -1.08674511e-01
1.91635430e-01 -6.46230936e-01 -8.28316033e-01 2.95414776e-01
8.46543968e-01 -5.69407225e-01 -1.09701979e+00 5.75213611e-01
3.08715135e-01 5.01281917e-01 3.76654774e-01 -1.21325409e+00
-4.83968049e-01 2.02130303e-01 -3.18958730e-01 6.24911964e-01
7.43967772e-01 1.63075656e-01 8.58066976e-01 -5.52481949e-01
9.75965798e-01 1.49592388e+00 1.11729455e+00 6.23784661e-01
-9.68924582e-01 -3.26040834e-01 4.29659843e-01 3.49245042e-01
-1.22922564e+00 -5.55470645e-01 2.86327124e-01 -4.68979150e-01
1.42066705e+00 -2.61087805e-01 6.78450525e-01 1.29386926e+00
3.61361414e-01 1.29151464e+00 5.30791521e-01 -2.58155137e-01
-2.97693331e-02 -4.36268628e-01 -2.58015413e-02 1.10805357e+00
-2.33253166e-01 2.53471017e-01 -7.76126266e-01 6.15519397e-02
1.23921013e+00 3.35768938e-01 -1.83605105e-01 -2.42859960e-01
-1.50643241e+00 4.41059887e-01 6.06839657e-01 5.17114699e-01
-6.22323394e-01 8.62285495e-01 6.74563408e-01 1.71581522e-01
5.85453033e-01 2.18674824e-01 -2.73719281e-01 -6.20073676e-01
-1.01068687e+00 1.58918858e-01 4.94825661e-01 6.92716539e-01
1.05402038e-01 -2.63623204e-02 -8.82938266e-01 4.10552800e-01
-1.30721498e-02 5.72412610e-01 3.09131324e-01 -1.22187269e+00
3.43819797e-01 5.33073127e-01 5.28536379e-01 -8.83036077e-01
-4.99495238e-01 -2.02401534e-01 -4.43045586e-01 8.82816464e-02
6.91166580e-01 -1.44787848e-01 -9.17323232e-01 1.88683259e+00
-5.16643412e-02 9.60811496e-01 -8.94158855e-02 8.03890347e-01
4.54054087e-01 5.69753051e-01 4.40052271e-01 2.07896810e-02
1.22093356e+00 -1.24288797e+00 -5.64521492e-01 -4.29050252e-02
8.46550465e-01 -2.90913582e-01 4.33696508e-01 4.28746268e-02
-1.06217253e+00 -9.58317041e-01 -5.45917094e-01 -1.49381831e-02
-3.11870247e-01 6.29281819e-01 4.27387387e-01 3.98746163e-01
-1.24083567e+00 8.08957100e-01 -1.49535358e+00 -8.58905375e-01
5.22497535e-01 2.85334170e-01 -5.68156958e-01 -6.20811284e-02
-9.29866374e-01 5.86467862e-01 9.23550874e-02 1.25509342e-02
-1.17557395e+00 -4.71588492e-01 -9.52800810e-01 -5.91415130e-02
2.07241237e-01 -7.61542439e-01 1.26244020e+00 -9.88928199e-01
-1.32090735e+00 5.59993744e-01 -5.54822505e-01 -1.03704274e+00
7.71927059e-01 -7.63614237e-01 -5.61189830e-01 6.99556351e-01
3.10641438e-01 8.14710915e-01 6.60863101e-01 -5.66768706e-01
-1.01720095e+00 2.73975823e-02 9.97393802e-02 1.80679709e-02
-2.62054116e-01 6.17964685e-01 -7.03582644e-01 -8.66400480e-01
-3.32806081e-01 -9.39832509e-01 -2.50695378e-01 1.22123294e-01
-1.07705534e-01 -3.94031584e-01 9.28518653e-01 -9.47732806e-01
1.19592071e+00 -2.03201938e+00 1.68409944e-01 -2.50954002e-01
-5.38314544e-02 2.19665393e-01 -3.15111220e-01 3.29551846e-01
3.27943116e-02 -2.67694294e-01 1.55604675e-01 -8.39548290e-01
-3.02938260e-02 -1.19804190e-02 -2.35022351e-01 9.52274263e-01
2.22445115e-01 1.20093083e+00 -1.13391054e+00 -3.94099146e-01
5.39719641e-01 8.05712879e-01 -3.68536055e-01 2.03103453e-01
-1.50749549e-01 6.32400036e-01 -3.97985518e-01 6.68289185e-01
-7.69977942e-02 -4.30709660e-01 -4.26667817e-02 -4.18424904e-02
-2.58886486e-01 1.96115240e-01 -1.00964463e+00 1.91068971e+00
-1.90296382e-01 7.91359425e-01 -3.83260071e-01 -5.96848786e-01
2.02642605e-01 7.82310486e-01 8.93691659e-01 -7.36681581e-01
-2.87987925e-02 -3.58835578e-01 -3.12995762e-01 -4.39987451e-01
3.92666757e-01 3.06117147e-01 5.08895554e-02 7.02823162e-01
2.97507703e-01 1.07898068e+00 5.03595889e-01 3.56821537e-01
1.54075265e+00 7.87578106e-01 1.02670893e-01 2.44957604e-03
5.93920946e-01 -8.27897638e-02 4.54801261e-01 1.04278767e+00
-6.53039992e-01 4.44967687e-01 2.44484708e-01 -7.17403114e-01
-9.13934529e-01 -1.03739476e+00 4.97790217e-01 1.53178966e+00
1.79197323e-02 -6.20498955e-01 -5.03210127e-01 -1.06470072e+00
-1.49409901e-02 2.43836120e-01 -9.13624287e-01 -3.50190662e-02
-1.14802063e+00 -2.44519532e-01 7.33920574e-01 1.18586433e+00
7.99758375e-01 -1.38866591e+00 -1.08092451e+00 3.75902414e-01
-9.08363834e-02 -1.48389781e+00 -8.45733106e-01 -3.33123237e-01
-6.61250532e-01 -1.17291093e+00 -9.69610214e-01 -5.92783689e-01
3.89140010e-01 3.94125640e-01 1.24217224e+00 -1.09888434e-01
-3.94513577e-01 1.09915137e+00 -4.46170121e-01 1.34318605e-01
-4.90232185e-02 -1.00580998e-01 2.31282532e-01 2.93150187e-01
4.59798843e-01 -2.76026100e-01 -7.40182877e-01 4.88514632e-01
-3.34040880e-01 -2.52083421e-01 3.72178316e-01 4.99827117e-01
2.84295827e-01 3.47250625e-02 1.47257060e-01 -4.00121838e-01
2.50945450e-03 -1.46424264e-01 -3.88620168e-01 4.14051205e-01
1.07727185e-01 -1.78789183e-01 3.75824958e-01 -5.00886917e-01
-1.05302215e+00 5.18679976e-01 2.32088968e-01 -7.31567740e-01
-4.56089973e-01 -5.09386733e-02 2.90458620e-01 -1.09091096e-01
4.35011923e-01 1.86688498e-01 -2.29643703e-01 -4.07370180e-01
4.72655356e-01 1.35152161e-01 7.89520442e-01 -2.10799679e-01
4.32382733e-01 9.28176343e-01 -6.13129325e-02 -8.24862957e-01
-8.16485345e-01 -9.41678286e-01 -1.03219509e+00 -5.01276910e-01
1.40850222e+00 -1.31961679e+00 -8.03077519e-01 5.03311276e-01
-9.80047762e-01 -7.56077111e-01 -2.19511420e-01 7.53828347e-01
-7.25503862e-01 2.67582566e-01 -1.15915406e+00 -9.34421301e-01
-1.15746655e-01 -6.08524203e-01 1.36010802e+00 1.30623922e-01
-3.83597374e-01 -1.24529207e+00 2.04709783e-01 -8.74059796e-02
2.02450961e-01 4.33739603e-01 -6.39965683e-02 -3.19946200e-01
-8.35064530e-01 -3.97245437e-01 -2.26250663e-01 -8.38238746e-03
1.72355026e-01 -2.91658998e-01 -7.62705564e-01 -5.60277700e-01
-5.31681418e-01 -3.25960040e-01 1.34391201e+00 8.19437563e-01
5.96658528e-01 6.74779639e-02 -7.41647124e-01 4.59543288e-01
1.00803602e+00 3.38169038e-02 6.97171092e-01 1.87080070e-01
9.05740917e-01 3.24461937e-01 4.67494458e-01 5.10098279e-01
5.59264086e-02 9.65183854e-01 -1.84589922e-01 -6.20553866e-02
-5.66182375e-01 -4.74294901e-01 8.51182103e-01 5.66228069e-02
-5.12471795e-01 -3.09277475e-01 -7.97507405e-01 6.46130443e-01
-2.33690691e+00 -1.61781204e+00 8.01211447e-02 1.95383430e+00
1.64642558e-01 1.96364060e-01 8.57771039e-01 -1.98023319e-01
6.80497348e-01 3.56460899e-01 -2.18362778e-01 2.81822264e-01
-8.25017467e-02 -1.74841657e-02 5.87645650e-01 4.30361748e-01
-1.92477071e+00 1.20844066e+00 6.95309639e+00 2.02772945e-01
-7.73059726e-01 1.97353184e-01 2.10229665e-01 -3.72413397e-01
6.62853301e-01 -4.55463409e-01 -8.90324593e-01 2.53617942e-01
1.04788768e+00 3.33414763e-01 2.41656139e-01 4.37479168e-01
5.08019924e-01 4.26939763e-02 -1.26868498e+00 8.79165828e-01
2.86374539e-01 -1.36656213e+00 -1.81237370e-01 -6.79828003e-02
7.15814233e-01 1.04652762e-01 -4.62872833e-02 4.21500921e-01
5.17072082e-01 -1.04141760e+00 7.79702783e-01 9.26885247e-01
6.53668344e-01 -6.60673022e-01 6.58574522e-01 4.01664572e-03
-1.78155518e+00 -3.07463646e-01 -3.35214995e-02 -2.16929644e-01
6.05066776e-01 -1.18325129e-01 -5.51730990e-01 2.62796849e-01
9.08349156e-01 1.79370785e+00 -7.17942297e-01 1.06246161e+00
-3.86107177e-01 6.66340649e-01 -1.17823794e-01 2.80745417e-01
3.94608945e-01 1.75561368e-01 4.83270317e-01 1.68730807e+00
9.94205922e-02 1.39930397e-01 6.73280418e-01 5.74804842e-01
6.50229678e-02 -3.68483692e-01 -5.07401764e-01 -2.57905900e-01
-3.28559019e-02 7.44082272e-01 -8.06198478e-01 -5.68836331e-01
-6.38763309e-01 1.50264108e+00 2.75544256e-01 5.89208603e-01
-1.00835586e+00 -1.91808194e-02 7.61755347e-01 7.96377957e-02
6.49145842e-01 -4.24021006e-01 3.33939195e-01 -1.28454411e+00
-2.17719674e-01 -4.43261623e-01 7.13388681e-01 -6.68013871e-01
-1.12314594e+00 3.68219405e-01 -1.66195065e-01 -1.32368445e+00
-3.80405366e-01 -6.05264306e-01 -4.79701996e-01 6.71742499e-01
-1.02354705e+00 -1.44237185e+00 -1.05761588e-01 8.31509292e-01
6.69292629e-01 -3.09639215e-01 6.56081200e-01 4.46231484e-01
-6.45137548e-01 4.98047680e-01 -2.66989708e-01 7.66834736e-01
7.84317017e-01 -1.23515773e+00 7.86538005e-01 1.27756763e+00
6.00201607e-01 5.41753411e-01 5.03143311e-01 -9.12609160e-01
-1.00586104e+00 -1.43193305e+00 1.08179855e+00 -9.74081397e-01
6.74049079e-01 -4.59540069e-01 -6.01400375e-01 1.25120604e+00
8.83047208e-02 3.23235720e-01 3.49132746e-01 6.75861835e-02
-3.50689828e-01 2.88190365e-01 -7.77363479e-01 5.31830490e-01
1.42498648e+00 -8.49184275e-01 -6.21446669e-01 4.45355147e-01
4.60828006e-01 -2.55244792e-01 -5.63864946e-01 1.49284109e-01
7.34692395e-01 -9.26276624e-01 1.33310890e+00 -1.01055670e+00
2.90991783e-01 -4.17233825e-01 -1.26375956e-02 -7.75003076e-01
-7.52061963e-01 -6.87481880e-01 -4.09907699e-01 7.98547566e-01
-7.20467567e-02 -1.23585127e-01 9.83244240e-01 3.07805628e-01
-7.10329041e-03 -2.69633442e-01 -9.18047309e-01 -8.81696165e-01
-4.72762614e-01 -4.07263905e-01 -1.41764298e-01 5.37444711e-01
-1.87513605e-01 1.25164881e-01 -8.26755941e-01 2.42705598e-01
5.31103194e-01 -8.42778236e-02 7.17782438e-01 -8.79135847e-01
-3.98654431e-01 -3.26265633e-01 -7.96275795e-01 -1.31413162e+00
3.93358916e-01 -5.00377536e-01 1.27217859e-01 -1.54946971e+00
2.24900365e-01 3.27685207e-01 -7.95368671e-01 7.12880254e-01
-1.87423192e-02 5.69494247e-01 3.88679504e-01 2.41898611e-01
-1.44763052e+00 4.61677045e-01 7.18319714e-01 -1.03507973e-01
-2.79426515e-01 1.29513964e-01 1.75226092e-01 6.69203699e-01
3.89035314e-01 -4.12351668e-01 -1.58087343e-01 -2.89916039e-01
-3.20626259e-01 1.53042063e-01 1.04845667e+00 -1.64942968e+00
4.46492225e-01 2.41722986e-01 7.44695127e-01 -5.67037702e-01
5.78994155e-01 -5.69700122e-01 -2.83968210e-01 7.73685575e-01
-6.94561958e-01 2.00410843e-01 7.53819942e-02 1.03906357e+00
-1.45929992e-01 5.33752143e-01 6.57367170e-01 -2.69554049e-01
-1.23669052e+00 3.10562104e-01 -5.11546314e-01 -6.51223660e-02
8.82529020e-01 -3.14616412e-02 -2.29068547e-01 -5.77141523e-01
-1.08820450e+00 1.77886873e-01 2.89337307e-01 4.33224320e-01
4.18339998e-01 -1.46930349e+00 -7.34116316e-01 -2.25829724e-02
5.71642034e-02 -1.01386738e+00 4.07889009e-01 1.07302916e+00
-3.82390648e-01 8.28595221e-01 -2.77188122e-01 -6.45434022e-01
-1.42457342e+00 6.68014705e-01 6.11626685e-01 -4.07815903e-01
-1.13993895e+00 1.04071677e+00 -5.53584397e-02 9.76184979e-02
5.37175000e-01 -3.30134511e-01 -2.45974824e-01 -2.06097767e-01
9.96638954e-01 6.01475179e-01 -5.39220154e-01 -9.14487779e-01
-5.94268680e-01 5.34634590e-01 2.30666488e-01 -2.19468281e-01
1.08658361e+00 -1.22691281e-01 1.84695467e-01 4.54656631e-01
1.10825479e+00 -2.74272174e-01 -1.93366945e+00 -3.07839066e-01
1.08214006e-01 -5.04012525e-01 -2.56712914e-01 -7.90704548e-01
-9.32361722e-01 6.90645397e-01 8.41056168e-01 -7.49039054e-02
8.49902868e-01 1.60587743e-01 9.89631534e-01 4.02545989e-01
4.28562492e-01 -9.60731387e-01 3.17047030e-01 6.36747003e-01
5.57684064e-01 -1.16779387e+00 1.64143667e-02 -1.31616712e-01
-4.94822741e-01 1.02682185e+00 4.27224368e-01 -4.67293233e-01
5.16470492e-01 9.90361348e-02 -1.14560574e-01 -1.34928569e-01
-6.61722481e-01 -5.87478280e-01 5.13440907e-01 4.81592417e-01
5.84628046e-01 -7.72626176e-02 2.61659145e-01 -1.16027240e-02
5.10919750e-01 3.61870170e-01 1.17019810e-01 1.05444670e+00
-2.67389238e-01 -8.01743805e-01 -1.91204458e-01 8.48841891e-02
-4.44587201e-01 2.35458866e-01 -3.04088056e-01 8.09848905e-01
3.79551977e-01 1.05819094e+00 3.40984762e-01 -2.27144420e-01
3.53013754e-01 6.37057126e-02 5.41363716e-01 -4.30541992e-01
-5.89602411e-01 1.15501605e-01 2.49490216e-01 -1.24667680e+00
-9.86625195e-01 -1.05977440e+00 -1.14234459e+00 1.11238062e-02
4.07262117e-01 -8.89967680e-02 -1.73700809e-01 1.00664318e+00
3.88373166e-01 7.95421124e-01 7.79891983e-02 -9.64699984e-01
-7.73387775e-02 -8.61102939e-01 -5.36322534e-01 7.00749815e-01
5.14674664e-01 -7.36876249e-01 -3.55062149e-02 4.23765928e-01] | [8.240006446838379, 0.42355313897132874] |
e7fa143e-fe10-44b6-a78d-afe565039584 | a-machine-learning-approach-to-the-prediction | 2305.18406 | null | https://arxiv.org/abs/2305.18406v1 | https://arxiv.org/pdf/2305.18406v1.pdf | A machine learning approach to the prediction of heat-transfer coefficients in micro-channels | The accurate prediction of the two-phase heat transfer coefficient (HTC) as a function of working fluids, channel geometries and process conditions is key to the optimal design and operation of compact heat exchangers. Advances in artificial intelligence research have recently boosted the application of machine learning (ML) algorithms to obtain data-driven surrogate models for the HTC. For most supervised learning algorithms, the task is that of a nonlinear regression problem. Despite the fact that these models have been proven capable of outperforming traditional empirical correlations, they have key limitations such as overfitting the data, the lack of uncertainty estimation, and interpretability of the results. To address these limitations, in this paper, we use a multi-output Gaussian process regression (GPR) to estimate the HTC in microchannels as a function of the mass flow rate, heat flux, system pressure and channel diameter and length. The model is trained using the Brunel Two-Phase Flow database of high-fidelity experimental data. The advantages of GPR are data efficiency, the small number of hyperparameters to be trained (typically of the same order of the number of input dimensions), and the automatic trade-off between data fit and model complexity guaranteed by the maximization of the marginal likelihood (Bayesian approach). Our paper proposes research directions to improve the performance of the GPR-based model in extrapolation. | ['Omar K. Matar', 'Tassos G. Karayiannis', 'Luca Magri', 'Francesco Coletti', 'Tullio Traverso'] | 2023-05-28 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-5.08341305e-02 -2.52151698e-01 -1.17909983e-02 -3.82706910e-01
-4.80685264e-01 -2.14380085e-01 4.50546831e-01 4.69586968e-01
-2.70795912e-01 8.67844224e-01 -2.32137129e-01 -5.86164355e-01
-3.37565899e-01 -6.50797009e-01 -4.73285139e-01 -8.81322980e-01
5.56637347e-02 5.15960038e-01 3.63532938e-02 2.71756023e-01
6.25362635e-01 5.83561957e-01 -1.24304760e+00 -3.88980389e-01
1.10076272e+00 1.18117583e+00 2.35021278e-01 6.10858917e-01
-1.11346543e-01 4.59316283e-01 -1.18020602e-01 1.53099254e-01
-3.47536542e-02 -5.61722636e-01 -5.95418155e-01 -2.12137401e-01
-4.34039026e-01 -3.37083153e-02 -1.48529455e-01 5.22354841e-01
4.15446043e-01 4.17066693e-01 1.09035003e+00 -1.08220685e+00
-4.23418134e-02 1.15183532e-01 -1.60045639e-01 -2.05843076e-01
-3.66128013e-02 1.79815099e-01 7.13502169e-01 -6.83101118e-01
1.22550823e-01 1.10089076e+00 5.51980495e-01 5.19485734e-02
-1.39129627e+00 -5.74677706e-01 -3.32879066e-01 -3.45012814e-01
-1.34294939e+00 -2.27021068e-01 4.18172538e-01 -7.69791603e-01
8.39384973e-01 1.09014496e-01 7.47265637e-01 4.24146473e-01
7.20484614e-01 2.49670133e-01 1.21434593e+00 -5.16746402e-01
6.52567804e-01 4.17136788e-01 -1.07664160e-01 4.80238259e-01
3.60708117e-01 4.48473811e-01 -2.60128826e-01 -1.81404576e-01
1.26590025e+00 -2.34389126e-01 -1.04438454e-01 -5.36405802e-01
-4.75886613e-01 1.08246768e+00 1.36794522e-01 8.60296637e-02
-2.49154299e-01 7.37542361e-02 3.99732113e-01 7.14646801e-02
5.61887383e-01 1.05803549e+00 -6.81644261e-01 -4.72166568e-01
-8.45228493e-01 3.67720634e-01 1.21912682e+00 7.95114875e-01
6.25182271e-01 -7.03465715e-02 5.06397113e-02 6.10953748e-01
6.34473264e-01 6.31963909e-01 3.31271082e-01 -6.49384856e-01
4.19664025e-01 5.75114429e-01 4.11832452e-01 -6.82491899e-01
-4.85002548e-01 -2.65096545e-01 -6.57237172e-01 1.11724421e-01
4.84956831e-01 -6.31610513e-01 -8.23821604e-01 1.07259846e+00
2.58378237e-01 -4.09671694e-01 2.42893733e-02 7.63111830e-01
2.36796826e-01 8.23243022e-01 3.44801515e-01 -3.05108577e-01
8.61879647e-01 -6.10066056e-01 -7.82064319e-01 -6.73712045e-02
6.45263314e-01 -8.19262445e-01 7.55174756e-01 -8.71691946e-03
-7.98605382e-01 -5.38663447e-01 -9.61937487e-01 1.86987370e-01
-2.38442555e-01 1.84137627e-01 6.93302989e-01 6.98889196e-01
-1.61531821e-01 1.07084823e+00 -1.16317344e+00 -8.56140554e-02
1.43986657e-01 3.01289082e-01 -3.41668129e-02 1.15355127e-01
-8.19771469e-01 1.14746177e+00 5.49274266e-01 4.89984363e-01
-5.00942409e-01 -1.05871558e+00 -8.21860254e-01 1.62733778e-01
4.69901673e-02 -3.69142354e-01 1.10796463e+00 -3.60084802e-01
-2.00291491e+00 7.09971711e-02 -1.23984165e-01 -8.86651576e-02
7.23829925e-01 -5.42959630e-01 -1.91888809e-01 -2.63856854e-02
-2.82369673e-01 9.23395008e-02 5.65717101e-01 -1.17350471e+00
-4.50484782e-01 -1.88769788e-01 -6.19938910e-01 7.81134143e-02
1.67471379e-01 -3.78948212e-01 -1.39660940e-01 -1.38724864e-01
4.75280255e-01 -9.69210088e-01 -5.50095975e-01 -2.27905661e-01
-1.26685560e-01 1.11133426e-01 7.72850811e-01 -6.52454555e-01
1.05054045e+00 -2.07206368e+00 -8.93981829e-02 4.64823961e-01
-3.75952125e-01 1.73730284e-01 4.90694910e-01 6.92872584e-01
1.03163145e-01 1.86513681e-02 -4.31960672e-01 4.00142893e-02
-4.12035771e-02 1.26117140e-01 -7.46063888e-02 6.23704255e-01
4.27478313e-01 8.02918851e-01 -8.87894630e-01 -2.82925963e-01
7.47725725e-01 4.27234411e-01 -2.66346991e-01 6.01698518e-01
9.81791168e-02 6.26717627e-01 -4.99811888e-01 2.63482600e-01
6.65310919e-01 -1.84688911e-01 1.27694145e-01 -1.50556326e-01
-4.87343431e-01 6.56707212e-02 -1.42070639e+00 1.13361526e+00
-6.82073414e-01 4.76272315e-01 1.17499679e-01 -8.87004972e-01
1.44933355e+00 4.62036580e-01 6.69106483e-01 -5.15891373e-01
1.47966757e-01 3.77817154e-01 -1.10716887e-01 -7.45233238e-01
4.81539369e-01 -7.57239997e-01 2.90100306e-01 2.95129448e-01
6.71740472e-02 -7.73487806e-01 -6.93613887e-02 -4.25033778e-01
3.50284129e-01 3.01485151e-01 2.11597368e-01 -7.77156293e-01
4.37836409e-01 -1.26915142e-01 4.13688064e-01 5.10725260e-01
1.99013650e-01 3.69448245e-01 7.22515285e-01 -3.10376614e-01
-1.40557969e+00 -8.01305115e-01 -6.77257419e-01 4.11491990e-01
3.86763602e-01 -1.67035192e-01 -3.78656715e-01 7.61769786e-02
3.71437281e-01 8.15944195e-01 -2.55481392e-01 -2.24329934e-01
-5.04699171e-01 -1.10870767e+00 1.89914361e-01 7.53489017e-01
3.21781456e-01 -5.04760027e-01 -6.96910858e-01 4.99878585e-01
3.49856496e-01 -1.02379394e+00 1.83906123e-01 5.67926466e-01
-1.18204951e+00 -1.31677759e+00 -4.18982267e-01 -3.45684528e-01
7.17834353e-01 -4.82638657e-01 7.55244792e-01 -1.56482503e-01
-4.71419245e-01 2.05208007e-02 -8.93021375e-02 -5.62292576e-01
-7.04849601e-01 -6.30734935e-02 -1.97958976e-01 -2.95252800e-01
7.66638890e-02 -1.15534216e-01 -5.53188026e-01 7.50667810e-01
-7.09802389e-01 -1.24278732e-01 6.62228644e-01 8.75779390e-01
4.20654476e-01 4.09110576e-01 5.85604072e-01 -8.87319088e-01
7.31560171e-01 -4.00054485e-01 -9.08391714e-01 1.46255940e-01
-1.15013814e+00 2.45996207e-01 7.44643092e-01 -2.95132399e-01
-1.41941857e+00 1.17481008e-01 8.92436877e-02 -2.52434880e-01
-1.70208633e-01 6.89163387e-01 6.76172078e-02 -5.08341752e-02
3.70050311e-01 -1.69590767e-02 3.04295719e-01 -4.50624824e-01
2.25121006e-02 4.89870489e-01 5.84047288e-02 -9.40703571e-01
4.29281294e-01 -1.10507920e-01 6.48627818e-01 -1.05780220e+00
-2.24966839e-01 -6.25865340e-01 -7.11125612e-01 -2.70643324e-01
5.88708103e-01 -6.84377611e-01 -9.51649487e-01 4.67282206e-01
-4.09431994e-01 -3.26848775e-01 -1.89295292e-01 9.57606971e-01
-6.36193931e-01 2.74715245e-01 -6.80180907e-01 -1.50147820e+00
-3.80101860e-01 -1.18735123e+00 7.22361624e-01 4.63726759e-01
-2.43453875e-01 -1.32132995e+00 -2.45725259e-01 6.81067631e-02
6.81013525e-01 4.17761683e-01 1.20262051e+00 -3.09584975e-01
-3.53843898e-01 -5.91257691e-01 -5.53929247e-02 5.18444180e-01
2.73792565e-01 3.46515357e-01 -8.21501315e-01 -3.27484220e-01
2.85065889e-01 -1.41568676e-01 4.86434281e-01 8.28741908e-01
8.09647560e-01 -2.36440878e-02 -5.08017778e-01 4.44121987e-01
1.54515052e+00 3.05991590e-01 3.82246375e-01 4.38494422e-02
4.26985204e-01 5.65132380e-01 7.98173189e-01 6.34521723e-01
-2.79785693e-01 2.36984089e-01 3.31141129e-02 -1.48433462e-01
5.85746050e-01 -4.77940798e-01 -4.91486005e-02 5.43925226e-01
-2.53262371e-01 4.97102775e-02 -9.24679399e-01 1.91691503e-01
-1.70505500e+00 -4.84839857e-01 -4.56416130e-01 2.65296817e+00
4.75668222e-01 1.80592284e-01 -1.06585696e-01 1.45037532e-01
6.04252517e-01 -3.48957360e-01 -4.11228776e-01 -4.67692882e-01
2.97723323e-01 1.41085669e-01 9.06216145e-01 7.02763259e-01
-9.43235278e-01 3.65468144e-01 6.90792179e+00 5.17817020e-01
-1.23283696e+00 -6.05540931e-01 8.59981537e-01 3.53949517e-01
3.68360132e-02 2.65598625e-01 -9.05181646e-01 4.43448335e-01
1.12693095e+00 -1.93433583e-01 4.04226393e-01 5.60166061e-01
4.74870443e-01 -6.31864607e-01 -1.19480956e+00 6.67762697e-01
-4.87951100e-01 -9.52982724e-01 -4.11225975e-01 -6.22360632e-02
6.04189038e-01 -4.55081850e-01 -1.17335677e-01 2.44613931e-01
-1.43239841e-01 -1.12454081e+00 4.48223740e-01 6.84234619e-01
6.10220611e-01 -6.36296272e-01 1.15190005e+00 3.00659835e-01
-9.94444788e-01 -2.00831458e-01 -2.93941110e-01 -2.07114980e-01
3.40423375e-01 8.29523563e-01 -1.15958035e+00 7.10727096e-01
4.10903543e-01 2.91145921e-01 -7.35033378e-02 1.22668326e+00
-4.19323519e-02 6.65065169e-01 -7.32804716e-01 -3.12404037e-01
7.30292946e-02 -7.35754728e-01 1.37608245e-01 1.03204918e+00
4.69357461e-01 2.29995977e-02 -9.50326398e-02 1.22397685e+00
5.08034110e-01 2.27521941e-01 -2.46601731e-01 -2.66559601e-01
4.08590198e-01 1.00097418e+00 -7.81055927e-01 1.72611088e-01
-1.77165166e-01 2.19211757e-01 -2.39869058e-01 5.40006280e-01
-4.61724788e-01 -3.95793527e-01 4.89447147e-01 3.50121796e-01
1.70251593e-01 -4.82550830e-01 -5.94196081e-01 -4.02663589e-01
5.43826930e-02 -1.52083948e-01 -2.11660918e-02 -2.52918094e-01
-1.20385170e+00 -5.08954525e-02 3.02943677e-01 -8.92793953e-01
-3.71986598e-01 -9.97214556e-01 -6.00230277e-01 1.30748796e+00
-1.39705348e+00 -6.25394523e-01 -2.21250188e-02 -7.75526762e-02
2.46119991e-01 1.70114726e-01 1.00418234e+00 8.24854895e-02
-6.31728709e-01 1.69925392e-02 7.83199906e-01 -1.13782562e-01
4.78122115e-01 -1.23097837e+00 2.65305806e-02 3.39164793e-01
-9.38001275e-01 5.82259655e-01 1.13589346e+00 -7.16535807e-01
-1.56811607e+00 -8.26517045e-01 4.96019125e-01 -1.11384802e-01
5.72240353e-01 -2.91717917e-01 -9.40183640e-01 1.91127881e-01
-4.52923745e-01 1.28440163e-03 6.78956091e-01 2.84152597e-01
5.30134141e-01 -9.41445902e-02 -1.08681953e+00 1.38638943e-01
2.37944070e-02 -4.10352498e-01 -2.31735349e-01 -1.03571285e-02
4.86441404e-02 -5.80758214e-01 -1.49433017e+00 5.82429767e-01
5.65339386e-01 -5.12229025e-01 4.89019215e-01 -5.63459575e-01
4.91401583e-01 -6.81402981e-02 2.34705284e-01 -1.32066989e+00
-5.64551465e-02 -5.77001035e-01 -1.72596320e-03 1.16556346e+00
6.15311861e-01 -6.29353046e-01 8.42308581e-01 1.55783319e+00
-5.85167073e-02 -1.12715650e+00 -9.87490237e-01 -6.56523407e-01
1.72409534e-01 -1.42168388e-01 3.25266570e-01 4.76341337e-01
2.82420784e-01 3.09927911e-01 -2.37211138e-01 2.09074736e-01
4.87104267e-01 1.04309052e-01 6.20260417e-01 -1.43533611e+00
-1.08191021e-01 -1.10485025e-01 -1.46431267e-01 -8.54025364e-01
-1.25162542e-01 -3.05881202e-01 3.76332402e-01 -1.24710000e+00
-2.72007793e-01 -9.13743615e-01 -2.99311196e-03 -1.89654022e-01
-3.19620252e-01 -6.89343035e-01 -1.09432474e-01 1.56353593e-01
1.68433905e-01 8.77352178e-01 1.43385124e+00 3.11779857e-01
-6.72296941e-01 3.95184487e-01 -3.08716167e-02 3.40154171e-01
8.19988906e-01 -4.22346860e-01 -3.27736288e-01 1.71294257e-01
2.00631060e-02 4.53314602e-01 2.39420254e-02 -9.63439524e-01
1.53457522e-01 -2.58154482e-01 6.67579651e-01 -3.62338752e-01
1.97604775e-01 -1.06316924e+00 3.49452585e-01 5.64893067e-01
-3.08699787e-01 -1.81034118e-01 4.32497799e-01 5.82088947e-01
-3.07313919e-01 -4.83683527e-01 7.35795379e-01 4.09760997e-02
-1.77567005e-01 3.19085494e-02 -4.98829067e-01 -2.01477751e-01
1.03320181e+00 -2.35403717e-01 -6.49250224e-02 -4.09433655e-02
-2.61545748e-01 3.30760509e-01 2.81941980e-01 1.20698988e-01
3.51109773e-01 -7.45842755e-01 -4.97278333e-01 4.78906184e-01
-1.90024495e-01 4.03952897e-01 2.28807315e-01 7.31918693e-01
-9.02914166e-01 6.78630650e-01 -1.11784883e-01 -5.88077068e-01
-5.99160969e-01 3.75423938e-01 5.26842654e-01 -1.71569407e-01
-5.29855192e-01 3.89768302e-01 -1.57363698e-01 -3.46054822e-01
-2.21669704e-01 -4.99892414e-01 1.67946175e-01 -3.50279748e-01
5.49492799e-02 7.77946472e-01 1.89428180e-01 -1.79416254e-01
7.27298856e-02 3.52059335e-01 3.03972751e-01 1.58039749e-01
1.40531850e+00 -4.33863141e-02 3.81284915e-02 6.76775575e-01
1.07376933e+00 -4.90062743e-01 -1.69342268e+00 2.35121071e-01
-5.75881265e-02 -5.15463889e-01 3.94321173e-01 -6.01581931e-01
-7.91315615e-01 5.74148417e-01 5.21096408e-01 2.35843480e-01
6.25963867e-01 -2.89112091e-01 3.16783011e-01 1.76012829e-01
-1.27732843e-01 -1.49907720e+00 -3.00032973e-01 3.43153089e-01
5.75605571e-01 -1.16236687e+00 3.42744589e-01 -6.14660323e-01
-5.16139686e-01 1.46896720e+00 5.11633813e-01 6.43777177e-02
9.15948570e-01 4.68299270e-01 1.78041205e-01 -8.02383199e-02
-3.99756700e-01 4.89694059e-01 8.40145722e-02 1.42553762e-01
6.64236724e-01 2.12172821e-01 -4.14418340e-01 2.91665822e-01
-7.22369254e-02 1.69846386e-01 4.23371457e-02 1.03384793e+00
-3.92311007e-01 -9.45493400e-01 -4.19918329e-01 7.01119483e-01
-2.43369386e-01 2.33640298e-01 3.41948926e-01 8.88594329e-01
-2.87499368e-01 1.02266753e+00 1.62430238e-02 1.10724673e-01
4.58511978e-01 1.50029778e-01 2.96915889e-01 -1.84598535e-01
-1.47301137e-01 5.46922684e-02 1.61142815e-02 -1.62504718e-01
-2.68209755e-01 -5.73111117e-01 -1.16375005e+00 -2.56420672e-01
-7.60673702e-01 6.20663822e-01 1.11850989e+00 1.08992159e+00
-1.71816014e-02 3.84615064e-01 6.50161445e-01 -6.97303951e-01
-6.96259916e-01 -1.19285309e+00 -1.09577787e+00 3.89756739e-01
2.64711678e-02 -8.28313947e-01 -4.51281786e-01 -2.01588988e-01] | [6.262622356414795, 3.3700647354125977] |
eaefa01d-e581-4837-8a0f-dab78304881d | idd-a-dataset-for-exploring-problems-of | 1811.10200 | null | http://arxiv.org/abs/1811.10200v1 | http://arxiv.org/pdf/1811.10200v1.pdf | IDD: A Dataset for Exploring Problems of Autonomous Navigation in Unconstrained Environments | While several datasets for autonomous navigation have become available in
recent years, they tend to focus on structured driving environments. This
usually corresponds to well-delineated infrastructure such as lanes, a small
number of well-defined categories for traffic participants, low variation in
object or background appearance and strict adherence to traffic rules. We
propose IDD, a novel dataset for road scene understanding in unstructured
environments where the above assumptions are largely not satisfied. It consists
of 10,004 images, finely annotated with 34 classes collected from 182 drive
sequences on Indian roads. The label set is expanded in comparison to popular
benchmarks such as Cityscapes, to account for new classes. It also reflects
label distributions of road scenes significantly different from existing
datasets, with most classes displaying greater within-class diversity.
Consistent with real driving behaviours, it also identifies new classes such as
drivable areas besides the road. We propose a new four-level label hierarchy,
which allows varying degrees of complexity and opens up possibilities for new
training methods. Our empirical study provides an in-depth analysis of the
label characteristics. State-of-the-art methods for semantic segmentation
achieve much lower accuracies on our dataset, demonstrating its distinction
compared to Cityscapes. Finally, we propose that our dataset is an ideal
opportunity for new problems such as domain adaptation, few-shot learning and
behaviour prediction in road scenes. | ['C. V. Jawahar', 'Anbumani Subramanian', 'Anoop Namboodiri', 'Manmohan Chandraker', 'Girish Varma'] | 2018-11-26 | null | null | null | null | ['road-scene-understanding'] | ['computer-vision'] | [ 2.61790812e-01 1.38055652e-01 -4.74132389e-01 -7.05562711e-01
-5.66636920e-01 -5.80812454e-01 8.74458134e-01 -9.73640010e-03
-4.94006515e-01 7.24893034e-01 9.89613608e-02 -4.03887451e-01
-2.94951677e-01 -1.00518274e+00 -4.40808058e-01 -5.35856128e-01
-8.63743946e-02 8.38612378e-01 9.30756152e-01 -5.41392744e-01
1.74453929e-01 5.76714039e-01 -2.24035239e+00 2.32613739e-02
1.06535256e+00 7.61669159e-01 3.59473884e-01 3.63510489e-01
-3.61480594e-01 6.80666268e-01 -2.20122963e-01 -4.26445901e-01
4.57446516e-01 6.72023296e-02 -8.14416051e-01 3.31649423e-01
9.07141924e-01 -1.48766607e-01 -5.06801605e-01 1.05418754e+00
2.29054585e-01 6.23441100e-01 9.22470093e-01 -1.18917596e+00
-2.35718802e-01 1.67829111e-01 -4.58811224e-01 3.53667974e-01
-1.13704056e-02 2.74558455e-01 9.84346926e-01 -4.99406189e-01
9.77950394e-01 1.17257440e+00 6.88443124e-01 5.60658097e-01
-1.11103189e+00 -5.30751109e-01 3.88856858e-01 6.36724234e-01
-1.38059521e+00 -3.18347484e-01 6.43374205e-01 -7.46508420e-01
6.09102488e-01 2.17598975e-01 4.57260817e-01 1.18085325e+00
-1.95250943e-01 6.33557796e-01 1.26288939e+00 -6.60956204e-02
4.40421849e-01 3.67302150e-01 6.16307080e-01 3.39062005e-01
2.11891979e-01 1.87478870e-01 -1.19644776e-01 3.70611548e-01
2.19432965e-01 -4.03122865e-02 1.79484367e-01 -8.08135033e-01
-9.46406245e-01 8.19875538e-01 3.74116272e-01 1.26836255e-01
-5.62381409e-02 -2.84791231e-01 5.59184194e-01 2.47718897e-02
4.59624171e-01 -1.32495863e-02 -4.31358069e-01 -3.11976194e-01
-7.41913080e-01 3.52611780e-01 6.69845462e-01 1.04344928e+00
1.30614316e+00 4.72950675e-02 -2.03597918e-02 1.29375124e+00
-2.11114269e-02 5.54987550e-01 2.43107930e-01 -1.24854457e+00
5.20146608e-01 4.41606343e-01 1.51200816e-01 -8.88239741e-01
-6.52023315e-01 -3.77001971e-01 -5.76546133e-01 3.92676920e-01
6.55921340e-01 1.39595374e-01 -1.32222021e+00 1.55545831e+00
3.53227884e-01 3.79112184e-01 2.81673279e-02 8.38249087e-01
9.16734934e-01 4.45025802e-01 1.76896170e-01 3.12514603e-01
1.37298679e+00 -1.06130803e+00 -4.36298549e-01 -4.83126551e-01
7.97499359e-01 -2.45418131e-01 1.23602009e+00 2.36553550e-01
-3.61843675e-01 -8.77863884e-01 -7.09829688e-01 4.73561734e-02
-1.03993511e+00 -3.54722172e-01 6.77837074e-01 1.07266951e+00
-9.52830970e-01 2.01368153e-01 -3.06133300e-01 -8.14937055e-01
7.03149140e-01 -5.96063435e-02 -2.83007652e-01 -2.32078716e-01
-1.26787102e+00 1.08959520e+00 4.30608749e-01 -1.57904595e-01
-1.05238152e+00 -6.14386678e-01 -1.11637676e+00 -3.89754325e-01
6.59081697e-01 -1.89886108e-01 1.08067667e+00 -6.22924685e-01
-1.28497243e+00 1.35400832e+00 -2.00645939e-01 -7.89817393e-01
7.02127934e-01 8.87039304e-02 -8.98164868e-01 1.30707806e-03
5.02789915e-01 1.00841439e+00 4.00975019e-01 -1.59939194e+00
-1.25528467e+00 -2.01612622e-01 2.73562402e-01 1.10679142e-01
1.87110871e-01 -3.26146156e-01 -5.72841346e-01 -3.43193591e-01
-8.14980865e-02 -1.09104872e+00 -6.17147267e-01 -3.45995039e-01
-3.02808225e-01 -1.38523042e-01 1.01679838e+00 -1.96629331e-01
1.04420412e+00 -2.01702929e+00 -4.77785200e-01 3.29861999e-01
2.05223978e-01 2.43842408e-01 -1.09517306e-01 2.22031623e-01
8.84567574e-02 -3.73571850e-02 -4.84764159e-01 2.15427708e-02
1.84279814e-01 7.71203279e-01 -2.15798199e-01 4.31972980e-01
3.97582501e-02 8.73146176e-01 -9.94709969e-01 -5.52254379e-01
7.47901142e-01 5.51834218e-02 -2.58102417e-01 -3.23821515e-01
-9.11005288e-02 5.42086005e-01 -3.57587874e-01 6.62740707e-01
8.96101773e-01 1.81144819e-01 -3.08185279e-01 2.10764974e-01
-3.83835673e-01 1.36215344e-01 -1.29252934e+00 1.53096378e+00
-3.91848892e-01 1.07304478e+00 -1.29860595e-01 -1.19855320e+00
1.06860638e+00 -2.92987585e-01 3.76286894e-01 -1.21343279e+00
7.73655251e-02 1.14538424e-01 -6.01098277e-02 -7.38876164e-01
7.58508861e-01 5.10640740e-02 -3.04032266e-01 -5.80772832e-02
-1.63667619e-01 -3.51735622e-01 6.03284538e-01 4.39667813e-02
8.30539405e-01 -5.83161749e-02 7.78507441e-02 -4.44114417e-01
4.94898677e-01 4.16240484e-01 5.82336187e-01 1.13203394e+00
-7.68832445e-01 6.76976800e-01 3.02015543e-01 -6.37607455e-01
-1.07063699e+00 -1.26020133e+00 -6.25303924e-01 1.10005486e+00
6.58978879e-01 -2.24940982e-02 -7.92572021e-01 -6.41826987e-01
-5.45088165e-02 8.74357939e-01 -8.44634891e-01 7.46292248e-02
-5.49810648e-01 -6.04250968e-01 5.78470886e-01 3.78557622e-01
7.20348954e-01 -8.73852193e-01 -5.41602135e-01 1.23441964e-01
-2.34413892e-01 -1.50306940e+00 1.39120534e-01 2.97988594e-01
-5.49130857e-01 -1.17107260e+00 -4.45518851e-01 -7.96091676e-01
3.57638061e-01 6.59531534e-01 1.38379669e+00 -4.04554993e-01
-4.15618509e-01 3.12226981e-01 -4.12297070e-01 -3.88963282e-01
-2.30580315e-01 1.84903249e-01 -1.47514477e-01 1.61298215e-01
6.99832320e-01 -2.84756839e-01 -5.41705489e-01 9.33185756e-01
-4.81853396e-01 -3.30742039e-02 3.91257524e-01 4.93102759e-01
6.30924284e-01 3.34722817e-01 5.78553498e-01 -1.23864794e+00
4.79413942e-02 -6.90763652e-01 -4.62857932e-01 -1.88945860e-01
-4.20983136e-01 -2.85025716e-01 2.43400633e-01 -1.58903912e-01
-1.34480941e+00 -1.40166292e-02 -4.36526209e-01 9.04392079e-02
-1.09246695e+00 -9.29030553e-02 -2.04258546e-01 -7.38312602e-02
7.43234873e-01 1.22036727e-03 -9.31713507e-02 -3.89139265e-01
5.74007511e-01 7.84526646e-01 7.36711442e-01 -5.59019744e-01
7.91025221e-01 9.02637661e-01 -1.02883734e-01 -1.35174119e+00
-8.75786781e-01 -1.02851212e+00 -9.60842133e-01 -5.47210932e-01
9.82116461e-01 -8.53641808e-01 -2.31550142e-01 5.17063975e-01
-4.66621131e-01 -7.04863727e-01 -5.65592468e-01 4.07225966e-01
-7.75082469e-01 2.68694997e-01 -2.49638796e-01 -7.60958254e-01
4.28801984e-01 -1.03665841e+00 9.74817336e-01 2.19157934e-01
-2.17415079e-01 -1.09484375e+00 7.92266428e-02 5.98990023e-01
2.98051149e-01 3.27978879e-01 6.82269990e-01 -5.53109169e-01
-4.93109703e-01 2.98576672e-02 -3.89880776e-01 2.81410486e-01
3.52840088e-02 -7.14339539e-02 -1.13590240e+00 2.52586156e-01
-4.77467209e-01 -1.62177756e-01 1.28497159e+00 4.71626461e-01
1.13260543e+00 3.99116665e-01 -5.13199627e-01 3.45938355e-01
1.29303741e+00 4.17371720e-01 9.01216626e-01 8.55278969e-01
6.93298399e-01 1.27382135e+00 1.01475072e+00 1.04045913e-01
7.32580006e-01 7.57850051e-01 6.17633104e-01 -2.22710103e-01
-2.79994190e-01 1.84297450e-02 -8.25506449e-02 2.56618589e-01
-1.98579729e-02 -1.47335947e-01 -1.33281469e+00 8.95647466e-01
-1.90774643e+00 -1.12642825e+00 -7.80787170e-01 2.12460804e+00
2.67029375e-01 6.22365713e-01 4.58819509e-01 1.18776694e-01
7.72226334e-01 3.61004323e-01 -5.58992505e-01 -3.53897274e-01
-2.96553135e-01 -9.13772434e-02 1.04370296e+00 5.90237558e-01
-1.39700913e+00 1.18060613e+00 6.73889351e+00 1.23849058e+00
-6.81286633e-01 1.65150702e-01 5.95511913e-01 2.92015016e-01
-1.52368546e-01 -1.31186351e-01 -1.17069590e+00 5.05252361e-01
1.05882573e+00 3.32312435e-01 -2.92525049e-02 9.92651284e-01
3.97956431e-01 -3.49839270e-01 -6.03213608e-01 8.97871912e-01
-6.11985475e-02 -1.16272628e+00 -1.97400734e-01 1.78727508e-01
9.58444834e-01 5.13585448e-01 3.67729478e-02 6.32814884e-01
6.19989455e-01 -8.93890500e-01 8.06625605e-01 3.12898815e-01
6.65071487e-01 -7.84942687e-01 6.29418612e-01 4.07257587e-01
-1.30019784e+00 -2.44224444e-01 -5.20048678e-01 -2.02987999e-01
2.80735970e-01 4.26187724e-01 -5.65791488e-01 4.22425985e-01
9.78323936e-01 1.03018773e+00 -8.36453319e-01 1.16766202e+00
1.20613612e-01 5.61730921e-01 -1.89180866e-01 1.15836576e-01
8.70831668e-01 -5.50556064e-01 4.02517766e-01 1.45639932e+00
4.80252467e-02 -1.03389196e-01 3.22624147e-01 3.45037848e-01
4.89818156e-01 -1.15132295e-02 -1.07200944e+00 6.02318048e-01
3.47706914e-01 1.19884336e+00 -1.07604492e+00 -5.30204952e-01
-7.04020560e-01 5.51707625e-01 9.45431218e-02 5.24430633e-01
-1.02279353e+00 -3.36093366e-01 1.15186059e+00 4.67701584e-01
3.20684075e-01 -3.14420700e-01 -4.98443037e-01 -8.45212519e-01
-1.84228823e-01 -4.88192260e-01 3.00492436e-01 -5.03499568e-01
-1.15120065e+00 3.32551837e-01 2.51076788e-01 -1.43952990e+00
4.28944966e-03 -8.36936474e-01 -3.92951488e-01 4.13974673e-01
-1.92517006e+00 -1.01404417e+00 -6.56775177e-01 5.64770639e-01
1.03373981e+00 -1.29178300e-01 4.75164205e-01 5.96912384e-01
-5.74450970e-01 1.84419632e-01 3.78092527e-01 1.19287157e-02
5.38748860e-01 -1.30230308e+00 5.29851079e-01 6.25972688e-01
1.98321342e-02 -5.31103462e-02 7.89209604e-01 -3.42512757e-01
-5.68138957e-01 -1.43576825e+00 5.91830015e-01 -7.53462851e-01
7.60742068e-01 -5.71447134e-01 -9.55567420e-01 4.54991013e-01
-5.81036396e-02 -6.33965358e-02 4.36046302e-01 1.71166867e-01
-2.43087485e-01 -3.93810302e-01 -1.04806590e+00 6.15918875e-01
1.59324849e+00 -3.25014710e-01 -6.05863869e-01 3.04486960e-01
4.27170783e-01 -3.03926170e-01 -2.80377179e-01 4.46227074e-01
2.99055457e-01 -1.24400759e+00 1.17059195e+00 -2.98911780e-01
1.61789417e-01 -3.18467349e-01 -2.56803334e-01 -1.11219931e+00
-4.59425658e-01 -1.39524341e-02 3.37634563e-01 1.26359558e+00
4.16327447e-01 -7.12730289e-01 1.25432372e+00 5.30549407e-01
-6.78045988e-01 -3.90731752e-01 -9.96957600e-01 -1.08580267e+00
1.01902552e-01 -9.04189050e-01 4.10524607e-01 8.70300651e-01
-6.66964412e-01 1.59443945e-01 -3.34992528e-01 5.02311578e-03
7.68225372e-01 -4.21288460e-02 1.07297182e+00 -1.71609747e+00
3.69472474e-01 -7.11156130e-01 -7.95578837e-01 -1.16252327e+00
4.72324133e-01 -7.20678389e-01 1.25012890e-01 -1.79935455e+00
-7.02407807e-02 -8.94952834e-01 -1.18689530e-01 2.54461586e-01
1.85717046e-01 4.64852393e-01 -4.01609898e-01 7.36769140e-02
-7.80437827e-01 3.99932325e-01 1.15766060e+00 -4.85887736e-01
-7.61040971e-02 9.65279937e-02 -5.07853508e-01 8.67731810e-01
8.82131577e-01 -2.60120600e-01 -6.12830520e-01 -1.38240710e-01
-1.50306255e-01 -5.82753539e-01 3.51442605e-01 -1.31591916e+00
4.88009490e-02 -4.16619301e-01 -1.70194715e-01 -9.85197067e-01
3.84575605e-01 -1.08434296e+00 9.49292779e-02 3.49425018e-01
-7.19029382e-02 -6.23891354e-01 1.41564682e-01 7.78172493e-01
-3.42894107e-01 -2.30994761e-01 1.00344026e+00 -2.39001200e-01
-1.96351349e+00 3.01214635e-01 -7.48447716e-01 4.38544482e-01
1.48181748e+00 -1.12153614e+00 -3.73562276e-01 -1.08380340e-01
-7.53975272e-01 5.90381920e-01 4.85271037e-01 7.42247105e-01
1.70774236e-01 -1.10692716e+00 -6.30241632e-01 2.17166960e-01
6.42066479e-01 9.06785801e-02 6.01985097e-01 6.05928600e-01
-5.72650075e-01 5.43265522e-01 -3.93224508e-01 -9.02849853e-01
-1.03888357e+00 4.27035511e-01 3.63948405e-01 3.03975306e-02
-9.99947429e-01 4.61039156e-01 5.20974159e-01 -8.77380908e-01
1.04294553e-01 -2.14579463e-01 -5.93624771e-01 4.99054611e-01
2.65142113e-01 7.02027559e-01 5.85541837e-02 -1.11959708e+00
-2.93807417e-01 7.68697917e-01 1.03233352e-01 1.11768231e-01
1.10887945e+00 -7.31308579e-01 3.39648843e-01 7.28167593e-01
1.00110459e+00 -2.02792421e-01 -1.51317632e+00 -2.92923599e-01
2.68414885e-01 -6.18294060e-01 -1.90312490e-01 -5.50245941e-01
-9.01544631e-01 8.95147920e-01 8.98178041e-01 3.01960647e-01
6.66647911e-01 2.64871210e-01 6.88103914e-01 4.16476995e-01
6.53554201e-01 -1.58620906e+00 -9.88055840e-02 9.74152029e-01
2.17752650e-01 -1.77189720e+00 -5.97198248e-01 -6.51858032e-01
-8.14873993e-01 7.61907995e-01 7.30562329e-01 8.44786838e-02
7.37085521e-01 8.75587948e-03 2.76255429e-01 -2.31753528e-01
-3.51305425e-01 -1.03504181e+00 1.39816821e-01 1.24742186e+00
-2.55850226e-01 3.11895281e-01 2.16573458e-02 2.13457808e-01
-3.61597598e-01 -4.27097291e-01 5.71280181e-01 6.09496295e-01
-9.37576532e-01 -8.36574256e-01 -2.16577768e-01 6.05113149e-01
2.29896070e-03 2.13773265e-01 -6.90076500e-02 1.19359195e+00
5.53581297e-01 1.11754429e+00 4.15182084e-01 -2.80612320e-01
6.73246145e-01 -1.75979324e-02 1.34182826e-03 -6.90683246e-01
1.18978787e-03 -4.25191939e-01 5.99296093e-01 -5.20064056e-01
-4.60550189e-01 -8.47975791e-01 -1.15586102e+00 -2.98956931e-01
-2.67235395e-02 1.12434365e-02 5.54287016e-01 8.18039179e-01
7.16041476e-02 4.64433491e-01 4.45394337e-01 -9.85248446e-01
2.14614123e-01 -6.94679618e-01 -8.98353517e-01 7.14964092e-01
2.10213780e-01 -1.17049348e+00 -4.02171046e-01 4.92930636e-02] | [8.238816261291504, -1.633687973022461] |
247fc165-f402-41c6-a31b-5c87741354d9 | sgloc-scene-geometry-encoding-for-outdoor | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_SGLoc_Scene_Geometry_Encoding_for_Outdoor_LiDAR_Localization_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_SGLoc_Scene_Geometry_Encoding_for_Outdoor_LiDAR_Localization_CVPR_2023_paper.pdf | SGLoc: Scene Geometry Encoding for Outdoor LiDAR Localization | LiDAR-based absolute pose regression estimates the global pose through a deep network in an end-to-end manner, achieving impressive results in learning-based localization. However, the accuracy of existing methods still has room to improve due to the difficulty of effectively encoding the scene geometry and the unsatisfactory quality of the data. In this work, we propose a novel LiDAR localization framework, SGLoc, which decouples the pose estimation to point cloud correspondence regression and pose estimation via this correspondence. This decoupling effectively encodes the scene geometry because the decoupled correspondence regression step greatly preserves the scene geometry, leading to significant performance improvement. Apart from this decoupling, we also design a tri-scale spatial feature aggregation module and inter-geometric consistency constraint loss to effectively capture scene geometry. Moreover, we empirically find that the ground truth might be noisy due to GPS/INS measuring errors, greatly reducing the pose estimation performance. Thus, we propose a pose quality evaluation and enhancement method to measure and correct the ground truth pose. Extensive experiments on the Oxford Radar RobotCar and NCLT datasets demonstrate the effectiveness of SGLoc, which outperforms state-of-the-art regression-based localization methods by 68.5% and 67.6% on position accuracy, respectively. | ['Chenglu Wen', 'Siqi Shen', 'Guosheng Hu', 'Cheng Wang', 'Shangshu Yu', 'Wen Li'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['outdoor-localization'] | ['robots'] | [-2.22525164e-01 -3.32054019e-01 -2.00232379e-02 -7.77434766e-01
-1.42771220e+00 -5.20992398e-01 3.94179881e-01 1.28045946e-01
-5.37833810e-01 6.16502941e-01 -1.36131287e-01 2.38794778e-02
-1.79624587e-01 -9.03573096e-01 -1.07948649e+00 -6.69800520e-01
6.91999123e-02 6.76219583e-01 5.36777265e-02 -6.42749816e-02
1.51906475e-01 5.82641721e-01 -1.17086613e+00 -5.68181992e-01
1.14427340e+00 1.16502726e+00 1.76998138e-01 2.62048304e-01
1.24612339e-01 3.72665107e-01 -3.61855358e-01 -1.75144017e-01
3.23791146e-01 1.76023975e-01 -9.65734497e-02 -1.99988052e-01
1.12675130e+00 -5.93627453e-01 -3.64917785e-01 1.09899390e+00
6.70414805e-01 -4.33793329e-02 2.32791469e-01 -1.05980992e+00
-3.08845818e-01 2.07655281e-01 -8.13389361e-01 -3.57066154e-01
9.31340829e-02 -1.55522957e-01 9.40998554e-01 -1.01288271e+00
3.33924621e-01 1.15704548e+00 1.14484501e+00 -1.14498988e-01
-1.20213294e+00 -8.92719805e-01 1.01225466e-01 1.85772236e-02
-1.93797326e+00 -3.73453498e-01 9.24547732e-01 -3.41300219e-01
6.04881704e-01 -3.89232710e-02 3.42987180e-01 7.52024472e-01
2.78862625e-01 5.13476253e-01 7.26362646e-01 -6.17557950e-02
1.59637332e-01 -2.99982488e-01 -1.61610276e-01 9.40494895e-01
4.33212101e-01 2.19317228e-01 -6.24050021e-01 -6.44939840e-02
9.06913936e-01 1.65886417e-01 -1.42258987e-01 -1.16780138e+00
-1.30197001e+00 8.45835984e-01 1.09866166e+00 -2.21280545e-01
-2.27750629e-01 5.65414369e-01 2.77345896e-01 -1.85371712e-01
4.49066579e-01 4.81274068e-01 -4.87457216e-01 3.82886338e-03
-9.84880984e-01 3.17334950e-01 3.83425295e-01 1.24465179e+00
9.90304649e-01 9.82268378e-02 2.21608415e-01 6.90824568e-01
4.43449408e-01 1.16140747e+00 -1.44386694e-01 -1.06227267e+00
7.58293271e-01 6.18502975e-01 3.42545360e-01 -1.49029982e+00
-6.37202621e-01 -9.83863235e-01 -8.58639061e-01 -1.24685273e-01
2.75034189e-01 -4.86643240e-02 -7.61418402e-01 1.85148919e+00
3.15621763e-01 2.54737079e-01 -7.60590956e-02 1.23059452e+00
5.00145018e-01 3.58129323e-01 -2.32368618e-01 1.70374200e-01
1.08295238e+00 -8.20367396e-01 -4.95252073e-01 -6.00085199e-01
7.14063227e-01 -6.44681871e-01 7.43651509e-01 1.64936900e-01
-6.30737841e-01 -6.06378734e-01 -1.33607459e+00 -2.15302005e-01
-5.03292941e-02 7.97392309e-01 8.32986534e-01 3.27429503e-01
-7.27898419e-01 2.10306793e-01 -1.07533348e+00 -1.77444413e-01
3.18935424e-01 3.34453076e-01 -2.99067229e-01 -4.53877091e-01
-8.86411250e-01 8.60241652e-01 2.69777447e-01 5.02818882e-01
-4.80981797e-01 -8.96352112e-01 -1.26576161e+00 -1.11801364e-01
3.92595798e-01 -5.62900543e-01 1.00615883e+00 -1.46981940e-01
-1.09373355e+00 3.76379043e-01 -3.06067020e-01 -4.77444440e-01
6.53170288e-01 -5.90546906e-01 -1.34100080e-01 -8.89473110e-02
4.22454566e-01 8.37897658e-01 4.52618420e-01 -1.33909249e+00
-7.26045191e-01 -7.07867742e-01 -1.56859472e-01 1.70279875e-01
2.10056245e-01 -7.20341980e-01 -7.15753615e-01 -4.86257970e-01
9.06546712e-01 -1.02827251e+00 -2.91540504e-01 1.99884817e-01
-1.15419373e-01 1.75464645e-01 7.73796201e-01 -6.04749501e-01
7.50954926e-01 -2.07571959e+00 -1.13243610e-01 1.30349964e-01
1.71211556e-01 -2.23673269e-01 -1.68514460e-01 8.90631899e-02
2.95312375e-01 -2.76767492e-01 -3.25849354e-01 -5.39206743e-01
-7.60339154e-03 2.55937696e-01 -5.63833296e-01 9.04489577e-01
2.55786180e-01 1.08080220e+00 -8.86979818e-01 -2.15973079e-01
4.65402365e-01 7.80555904e-01 -6.91553473e-01 -1.40209198e-01
-9.63309854e-02 5.79600275e-01 -5.81383944e-01 9.40776110e-01
1.24086761e+00 1.13012746e-01 -9.08372998e-02 -5.64356327e-01
-2.10494027e-01 3.44129831e-01 -1.08996356e+00 2.33600307e+00
-6.52987778e-01 5.92106879e-01 1.24524206e-01 -6.09736919e-01
1.24372685e+00 -3.89613867e-01 5.30651450e-01 -9.93260145e-01
1.62616178e-01 3.30007672e-01 -2.63881356e-01 1.61090285e-01
7.62786508e-01 8.54335725e-02 -3.63397777e-01 -2.09711283e-01
-2.17554122e-01 -3.84306520e-01 -3.39169860e-01 -7.70728216e-02
9.28629041e-01 5.23995280e-01 1.10218741e-01 -1.00872889e-01
5.07532477e-01 2.23947674e-01 8.85881364e-01 6.14055634e-01
-3.59261222e-02 7.88747847e-01 -1.40330475e-02 -2.96061218e-01
-8.43334675e-01 -1.16759694e+00 -2.63071805e-01 6.55728936e-01
6.14781618e-01 -4.76300329e-01 -3.04649413e-01 -3.67707849e-01
4.45671618e-01 4.52067107e-01 -2.40866542e-01 -2.28204072e-01
-8.96238565e-01 -6.11177742e-01 5.25114834e-01 7.74139166e-01
7.94688344e-01 -2.35726863e-01 -4.48589742e-01 3.22528481e-01
-3.90500873e-01 -1.32685137e+00 -3.54490101e-01 1.57716244e-01
-9.18740034e-01 -8.00494552e-01 -7.37038553e-02 -4.48103160e-01
7.06711173e-01 6.78027749e-01 6.97367072e-01 -2.17889920e-01
-7.45074376e-02 1.84374303e-02 -1.74487725e-01 -1.60542086e-01
3.16893548e-01 2.86931872e-01 2.01538727e-01 -3.26654017e-01
2.40390986e-01 -5.86795449e-01 -6.03727818e-01 3.49412680e-01
-2.93726802e-01 -1.02711700e-01 8.99955869e-01 7.90742815e-01
8.67537796e-01 -1.17827542e-01 1.01751111e-01 -4.84862715e-01
-4.89564724e-02 -1.83348060e-01 -1.07475734e+00 -9.88643393e-02
-6.17746651e-01 1.60230577e-01 4.35891271e-01 4.80908491e-02
-7.44100571e-01 5.51279843e-01 -1.07594334e-01 -5.97343147e-01
1.55836716e-01 7.77036250e-01 -3.58052999e-01 -5.44995844e-01
4.27705258e-01 2.06925929e-01 -1.25831500e-01 -5.45004070e-01
3.63718003e-01 4.48686182e-01 1.07124245e+00 -6.84762537e-01
1.14161241e+00 5.58569133e-01 2.95880735e-01 -5.38942277e-01
-1.10958827e+00 -5.94247639e-01 -8.44629109e-01 1.52944013e-01
4.95875895e-01 -1.57271743e+00 -6.34752452e-01 2.83576369e-01
-1.21819103e+00 2.77522355e-02 1.41488105e-01 6.83588743e-01
-4.87692982e-01 3.87566805e-01 -2.67065495e-01 -6.38840795e-01
-3.63805205e-01 -1.16266978e+00 1.69287467e+00 -7.98582435e-02
1.79511383e-01 -5.03415704e-01 2.48331297e-02 4.23860788e-01
2.20789880e-01 4.81782407e-01 3.42438161e-01 7.23366737e-02
-1.22773302e+00 -5.50544083e-01 -5.36369205e-01 -1.67679694e-02
-4.83978428e-02 -4.00046915e-01 -8.62771809e-01 -6.18116260e-01
-7.68313110e-02 -1.36510238e-01 8.98563325e-01 3.29118490e-01
8.64615262e-01 2.58485407e-01 -4.37228411e-01 1.11643612e+00
1.55064261e+00 -1.70319542e-01 4.72801059e-01 5.24736166e-01
1.11191535e+00 2.63105392e-01 1.24536228e+00 3.13947350e-01
8.04562211e-01 9.28620934e-01 7.92413890e-01 -1.58838347e-01
1.40796481e-02 -6.92082942e-01 2.94959575e-01 8.40651214e-01
3.21427047e-01 6.31379336e-02 -9.25069273e-01 3.14997584e-01
-2.06595778e+00 -3.92875135e-01 -3.51332337e-01 2.29548454e+00
5.02013505e-01 1.46489307e-01 -5.10559201e-01 -2.83249259e-01
4.96132642e-01 3.06290776e-01 -5.55815756e-01 2.78093636e-01
-4.05209363e-02 -3.71462516e-02 1.23756874e+00 7.04268575e-01
-1.27688468e+00 1.20069826e+00 4.89216089e+00 7.53011167e-01
-1.34628415e+00 2.25218534e-02 7.43945464e-02 1.14940815e-01
5.96883930e-02 1.06520176e-01 -1.07620537e+00 2.63323158e-01
6.84363961e-01 2.71732688e-01 1.82847708e-01 1.12286210e+00
1.60693109e-01 -2.35091075e-01 -9.46269393e-01 1.12843812e+00
1.26858562e-01 -1.20180726e+00 -1.87147900e-01 1.85480401e-01
5.47608256e-01 3.92395943e-01 -1.25422060e-01 4.26940918e-01
5.31281121e-02 -6.66677952e-01 1.11108303e+00 3.80797416e-01
9.60605919e-01 -1.04961860e+00 9.28391516e-01 5.08374035e-01
-1.58949840e+00 1.13329642e-01 -6.66945040e-01 -1.03096776e-01
1.82187721e-01 7.96907246e-01 -8.09715927e-01 8.79047751e-01
5.44793546e-01 7.82604754e-01 -7.57025361e-01 1.21487653e+00
-4.75536615e-01 2.32766643e-01 -6.84067965e-01 3.75483215e-01
3.19150776e-01 -3.03958595e-01 4.68071610e-01 8.96065891e-01
4.71787244e-01 -2.21548855e-01 4.71286178e-01 1.05304885e+00
3.85913923e-02 -2.55119741e-01 -4.85878736e-01 3.97924066e-01
9.04695868e-01 1.39534760e+00 -4.14508581e-01 1.78360656e-01
-1.50350168e-01 6.51835263e-01 3.60862583e-01 6.69215396e-02
-1.12954068e+00 -3.76076430e-01 8.06028903e-01 9.52648092e-03
3.53668272e-01 -8.87782335e-01 -5.09357154e-01 -1.15281582e+00
3.57787967e-01 -4.34538335e-01 -2.02708825e-01 -7.43987441e-01
-9.77754295e-01 4.04284686e-01 -2.42685497e-01 -1.55118978e+00
-1.57272160e-01 -3.30879986e-01 -1.21018119e-01 8.37840021e-01
-1.69929731e+00 -1.56227708e+00 -7.45905161e-01 4.87374291e-02
2.20400840e-01 1.45299032e-01 4.31162059e-01 4.73599464e-01
-4.01098430e-01 6.37027562e-01 1.58562601e-01 1.77800596e-01
1.02657521e+00 -1.03664958e+00 6.06232345e-01 8.89421761e-01
3.48349325e-02 7.96971500e-01 5.71545243e-01 -7.34225988e-01
-1.83758247e+00 -1.48925245e+00 7.88158178e-01 -6.52208507e-01
5.26162624e-01 -7.73373485e-01 -7.15799272e-01 6.18668020e-01
-6.06410563e-01 1.91403881e-01 -9.49629303e-03 1.68960709e-02
-4.49258924e-01 -5.54967225e-01 -9.64478374e-01 5.03598273e-01
1.12875319e+00 -4.70807344e-01 -3.89528453e-01 1.59361288e-01
8.72299969e-01 -9.00126934e-01 -6.86173022e-01 8.69163811e-01
5.09892523e-01 -7.59806573e-01 1.22244966e+00 1.90331295e-01
1.24878645e-01 -8.05875838e-01 -5.25432229e-01 -1.14007449e+00
-4.05768186e-01 -1.55763999e-01 -4.51271385e-02 1.43850482e+00
3.22762690e-02 -6.73909485e-01 9.40715075e-01 2.30091244e-01
-2.54672468e-01 -5.99494159e-01 -1.09629524e+00 -8.41578245e-01
-1.17813408e-01 -6.03927493e-01 6.50266051e-01 7.84149647e-01
-6.62764251e-01 4.71653342e-01 -4.05327380e-01 7.23497331e-01
8.98918390e-01 3.96980971e-01 1.16667771e+00 -1.35918272e+00
1.14110023e-01 -7.49174803e-02 -6.10222995e-01 -1.55646729e+00
2.88507462e-01 -6.84852123e-01 5.97334146e-01 -1.30428231e+00
-1.52279601e-01 -8.85890365e-01 4.25546197e-03 2.45238364e-01
-4.26222458e-02 3.59433591e-01 1.01703927e-02 3.19274336e-01
-6.73192203e-01 9.39836323e-01 9.52893198e-01 -9.69078019e-02
-2.70595886e-02 -1.49335980e-01 -5.74860871e-01 7.33582795e-01
6.35720670e-01 -5.16163051e-01 -2.02825218e-01 -8.98815274e-01
1.44115329e-01 7.72072002e-02 6.39935613e-01 -1.24202955e+00
4.97566164e-01 -4.33209725e-02 5.89951158e-01 -1.26772702e+00
6.09350443e-01 -1.02273452e+00 -2.08070986e-02 2.97343284e-01
1.12455137e-01 -2.05041505e-02 1.59523383e-01 8.18768919e-01
-3.35675001e-01 2.64695972e-01 7.68485844e-01 3.56861681e-01
-7.77145624e-01 4.55339462e-01 3.30610216e-01 -2.11005062e-01
6.95859492e-01 -2.40029141e-01 -2.56608903e-01 -3.01900357e-01
7.55555555e-02 5.46339810e-01 6.64312780e-01 4.20095444e-01
5.55125535e-01 -1.58162522e+00 -5.49908936e-01 3.49985152e-01
3.51337880e-01 6.59675002e-01 6.26997054e-02 1.18319952e+00
-7.54802227e-01 5.79689384e-01 1.17783614e-01 -1.02624023e+00
-9.10014808e-01 1.10036433e-01 2.39178851e-01 -2.40519848e-02
-6.34222567e-01 7.73065627e-01 1.14529438e-01 -9.48861897e-01
3.10046285e-01 -6.18079662e-01 3.95298779e-01 -1.31036058e-01
2.70097822e-01 3.18568826e-01 2.43935302e-01 -8.80177736e-01
-8.00562918e-01 1.07357144e+00 5.74814305e-02 1.45166457e-01
1.33996630e+00 -2.97162682e-01 -8.28319043e-02 4.67227399e-02
1.30391514e+00 3.44922692e-01 -1.41527605e+00 -3.55807960e-01
3.02061774e-02 -6.84770763e-01 4.45068538e-01 -5.82306802e-01
-1.13232052e+00 8.97710860e-01 6.24385297e-01 -5.48314393e-01
7.55593061e-01 -2.31289119e-01 9.64707613e-01 7.07107544e-01
8.68592441e-01 -8.24339092e-01 -2.64148563e-01 8.59236956e-01
8.74635100e-01 -1.45627952e+00 4.19917405e-01 -6.53711557e-01
-3.91431302e-01 1.01224744e+00 6.49355710e-01 -4.42190826e-01
3.24189335e-01 4.28982586e-01 -3.21770944e-02 -8.53755996e-02
-1.02869011e-01 -8.68387520e-02 3.82834554e-01 5.98858237e-01
2.15519458e-01 1.24790408e-01 1.90426670e-02 5.91512859e-01
-5.14810324e-01 -1.93341866e-01 -5.16262576e-02 8.94935727e-01
-6.27362609e-01 -8.79230559e-01 -5.09301841e-01 2.93512009e-02
2.68282712e-01 -1.90738384e-02 -3.81685048e-01 9.09129262e-01
2.64288992e-01 5.99942744e-01 1.76512644e-01 -5.56634009e-01
5.01320422e-01 -4.35014576e-01 5.33339620e-01 -3.69820595e-01
-4.55043465e-02 3.04008752e-01 -5.69092520e-02 -9.02459323e-01
-3.36628570e-03 -6.34910166e-01 -1.46411455e+00 -3.45290422e-01
-5.09094954e-01 -4.76164259e-02 8.66227567e-01 8.11036408e-01
5.54327667e-01 4.17145103e-01 5.21553695e-01 -1.05935037e+00
-7.09168851e-01 -8.41944814e-01 -4.01991308e-01 -2.55197376e-01
4.81599689e-01 -1.07188272e+00 -1.43566206e-01 -5.10508001e-01] | [7.535606861114502, -2.2347075939178467] |
c1752d34-b0b4-4e9c-bbd9-9eb53eb5f2ef | the-future-of-artificial-intelligence-ai-and | 2305.02327 | null | https://arxiv.org/abs/2305.02327v1 | https://arxiv.org/pdf/2305.02327v1.pdf | The Future of Artificial Intelligence (AI) and Machine Learning (ML) in Landscape Design: A Case Study in Coastal Virginia, USA | There have been theory-based endeavours that directly engage with AI and ML in the landscape discipline. By presenting a case that uses machine learning techniques to predict variables in a coastal environment, this paper provides empirical evidence of the forthcoming cybernetic environment, in which designers are conceptualized not as authors but as choreographers, catalyst agents, and conductors among many other intelligent agents. Drawing ideas from posthumanism, this paper argues that, to truly understand the cybernetic environment, we have to take on posthumanist ethics and overcome human exceptionalism. | ['Ben Bowes', 'Zihao Zhang'] | 2023-05-03 | null | null | null | null | ['ethics'] | ['miscellaneous'] | [ 1.11185655e-01 4.06562120e-01 5.29809557e-02 1.81831628e-01
5.38265049e-01 -6.12134635e-01 1.02045631e+00 -1.18915848e-01
-4.65742201e-01 5.57265997e-01 8.10283720e-01 -7.48015523e-01
-4.47129220e-01 -7.55807579e-01 -2.76563138e-01 -3.55507016e-01
-8.23134854e-02 -9.65043083e-02 -3.21699589e-01 -1.11103940e+00
7.14047730e-01 5.44206023e-01 -1.22257388e+00 -1.05962180e-01
6.02434099e-01 1.01757161e-01 -2.22519189e-01 9.18579876e-01
3.54890883e-01 1.48769951e+00 -5.67638099e-01 -4.50378388e-01
3.00335020e-01 -7.81673849e-01 -6.33874536e-01 -3.04221839e-01
-2.17763782e-01 -2.57086039e-01 -9.98012647e-02 6.71829164e-01
4.66773272e-01 -5.56943659e-03 7.30798006e-01 -1.06819880e+00
-5.97484171e-01 8.70453298e-01 -1.90610021e-01 4.43621865e-03
4.43590879e-01 7.34412551e-01 7.99455583e-01 -3.46359640e-01
7.65514314e-01 1.21891618e+00 8.59789193e-01 2.25734133e-02
-1.07804894e+00 -6.36245251e-01 -5.34164086e-02 -5.28328493e-02
-1.29230416e+00 -5.49534023e-01 7.04884589e-01 -9.32498395e-01
9.12190318e-01 4.28584963e-01 1.64242709e+00 8.03271651e-01
7.12816000e-01 1.94852933e-01 9.21342492e-01 -8.06367517e-01
3.79336298e-01 1.32229805e-01 -6.10130847e-01 1.39387965e-01
1.34960964e-01 4.41350967e-01 -3.87017995e-01 -3.78411859e-02
7.98480034e-01 -2.06622228e-01 -3.04250985e-01 2.61471182e-01
-1.23367047e+00 7.90381849e-01 3.59993130e-01 6.28734112e-01
-6.61958218e-01 1.74329102e-01 1.92134038e-01 4.07379150e-01
1.73544750e-01 1.26049531e+00 -4.10562754e-01 -8.35270047e-01
-4.60743427e-01 1.35178566e-01 1.26513565e+00 4.35552686e-01
2.21453637e-01 1.36544272e-01 8.53049755e-01 1.40262499e-01
7.08379090e-01 5.11736691e-01 -4.06467989e-02 -1.26587355e+00
-2.32700169e-01 5.53248227e-01 1.84874460e-01 -1.48853922e+00
-3.87009919e-01 -3.06763083e-01 -1.62609220e-01 6.06703043e-01
4.20807675e-02 -6.94736421e-01 -3.99020404e-01 1.16208017e+00
7.78145865e-02 -1.72072649e-01 1.83927312e-01 8.95415425e-01
1.70093209e-01 6.05000973e-01 3.33147347e-01 -1.52710974e-01
7.09354222e-01 -4.15400833e-01 -6.54144526e-01 -1.66373119e-01
7.02140510e-01 -6.56705916e-01 9.60308611e-01 6.24600112e-01
-8.52340162e-01 2.96599090e-01 -1.15322661e+00 6.40771747e-01
-4.65753466e-01 -7.96340525e-01 7.29288340e-01 8.76625896e-01
-1.02737105e+00 6.39704525e-01 -5.84849179e-01 -9.16291893e-01
2.37591058e-01 -1.59197927e-01 -7.65858367e-02 6.70497417e-01
-9.85913336e-01 1.50839877e+00 1.93353459e-01 5.05940259e-01
-4.58563596e-01 -4.73844022e-01 -5.98231137e-01 -3.58165324e-01
1.52394876e-01 -3.56997430e-01 9.72867250e-01 -1.45253134e+00
-1.33877826e+00 6.52048886e-01 6.63479149e-01 -2.03562349e-01
6.20863557e-01 -1.82774171e-01 -6.13889933e-01 -1.19719967e-01
-2.54883200e-01 -9.73139033e-02 -9.22329277e-02 -1.60815692e+00
-6.87180340e-01 -7.47645944e-02 3.98871273e-01 3.04089725e-01
-2.13138074e-01 3.58816028e-01 7.91249573e-01 -3.63123596e-01
-4.40541375e-03 -7.53656387e-01 -4.25321430e-01 -1.42325446e-01
8.83604772e-03 1.56012163e-01 3.67683798e-01 -4.58651423e-01
1.41937387e+00 -1.96940160e+00 2.78327186e-02 4.89928603e-01
2.96046883e-01 -3.57196666e-02 1.79527909e-01 1.63245559e+00
4.40862060e-01 5.93398929e-01 2.17376292e-01 7.43127525e-01
2.05921888e-01 2.75305390e-01 -1.72626108e-01 6.15252495e-01
-1.60556167e-01 2.80498058e-01 -1.22822678e+00 -3.99175256e-01
1.73277646e-01 3.56765538e-01 -4.37219322e-01 -3.05488735e-01
2.04729363e-01 4.45322871e-01 -5.48307776e-01 5.27986407e-01
1.73072487e-01 1.74239442e-01 7.16672719e-01 4.26616818e-01
-9.24017072e-01 -2.34753966e-01 -7.61586130e-01 1.06612039e+00
-5.35498142e-01 1.27042937e+00 3.36399764e-01 -2.48661160e-01
8.53313625e-01 3.86634290e-01 1.88022301e-01 -1.04476523e+00
4.39903647e-01 4.86007810e-01 6.26328349e-01 -1.02170181e+00
4.57099587e-01 -5.77339292e-01 1.73655860e-02 5.48178911e-01
-6.66332066e-01 -5.85477889e-01 -5.61760783e-01 -1.37589321e-01
1.20634449e+00 6.71816051e-01 3.70109856e-01 -7.26293981e-01
2.86252707e-01 3.02851111e-01 6.54450893e-01 2.44739830e-01
-3.96530300e-01 -1.82460636e-01 3.67583632e-01 -9.14519548e-01
-1.19337976e+00 -6.87296808e-01 -1.75802007e-01 9.61400926e-01
3.11720163e-01 -3.62790912e-01 -4.67567772e-01 3.49892117e-02
-2.36835539e-01 1.20845687e+00 -5.08209050e-01 -5.92833087e-02
-6.17275894e-01 -3.35084319e-01 6.40614152e-01 1.09360196e-01
2.43208762e-02 -1.06390917e+00 -1.64325929e+00 4.12549347e-01
5.34182072e-01 -5.10891438e-01 4.32505161e-01 3.14808547e-01
-4.95209426e-01 -9.49972451e-01 5.63394129e-02 -5.30211508e-01
5.05939126e-01 -1.43624857e-01 9.72760975e-01 4.81215805e-01
-2.51374513e-01 3.83527219e-01 -5.86737037e-01 -7.70932376e-01
-8.54191661e-01 -1.89615563e-01 -8.19643289e-02 -4.70758647e-01
2.36610830e-01 -1.10559916e+00 -7.06652105e-01 1.04284339e-01
-7.60695815e-01 4.19724882e-01 6.24136746e-01 4.01159942e-01
-4.08051342e-01 1.02076970e-01 7.71528542e-01 -5.66699624e-01
3.95075381e-01 -7.32715368e-01 -1.45369053e-01 1.66179836e-01
-7.29131281e-01 -5.20269275e-01 5.71260214e-01 -2.64529407e-01
-1.02223933e+00 -3.63774031e-01 1.07341997e-01 5.65539479e-01
-3.00118804e-01 7.31509984e-01 -1.67984188e-01 -3.03884536e-01
9.67706680e-01 -2.19140083e-01 3.86889100e-01 -2.30070651e-02
4.10200000e-01 9.39435840e-01 2.30240420e-01 -4.13008988e-01
1.07607186e+00 3.24139416e-01 1.04724546e-03 -1.39659369e+00
2.62962818e-01 2.15348125e-01 -8.35856318e-01 -1.17588747e+00
7.14264989e-01 -8.40789914e-01 -6.91981792e-01 1.38286516e-01
-8.10897052e-01 -9.19833362e-01 -3.85950148e-01 7.56732762e-01
-4.99487787e-01 -3.93924087e-01 -6.48828819e-02 -1.23335981e+00
1.38413876e-01 -5.40781915e-01 2.22329333e-01 5.56382835e-01
-6.10119402e-01 -1.39600360e+00 3.89214218e-01 1.79462805e-01
8.91152799e-01 1.15699637e+00 9.63592350e-01 -4.73057419e-01
-2.94705749e-01 -5.54527231e-02 3.63012582e-01 5.14240675e-02
2.37073064e-01 5.51578879e-01 -5.97919047e-01 2.14305192e-01
2.32003942e-01 4.15541008e-02 -4.25605655e-01 -4.41177696e-01
-4.65163141e-01 -9.02991116e-01 -7.64761493e-02 4.04418170e-01
1.93667340e+00 8.78637016e-01 7.55670130e-01 1.21530950e+00
3.71908277e-01 1.08559418e+00 3.95079523e-01 7.41115451e-01
5.75723827e-01 -5.32623194e-02 3.27812046e-01 -2.43740141e-01
3.16863567e-01 -2.06185073e-01 1.27727419e-01 8.30889463e-01
-3.43199253e-01 -2.07334891e-01 -1.62755871e+00 7.61921167e-01
-1.89020085e+00 -1.07835054e+00 -6.10206723e-01 1.52055383e+00
6.05353057e-01 7.40085021e-02 1.59197703e-01 -1.48148075e-01
3.07211220e-01 2.51387864e-01 -6.95019141e-02 -7.42531002e-01
1.19447678e-01 -1.80105746e-01 6.94657147e-01 3.25652868e-01
-6.98461413e-01 8.03567111e-01 7.70797968e+00 -1.05797172e-01
-8.16611409e-01 -1.38286695e-01 1.84707463e-01 -5.78188002e-02
-9.20817733e-01 4.93580490e-01 2.46846363e-01 7.98515901e-02
8.77561808e-01 -4.21122015e-01 6.41514242e-01 3.10337812e-01
8.55884671e-01 -4.76125985e-01 -5.98942935e-01 1.52546808e-01
1.23791918e-01 -1.26452398e+00 -4.59821522e-01 3.46678108e-01
6.61059141e-01 -5.55209182e-02 -2.38324583e-01 -2.63769746e-01
5.91858566e-01 -1.24965405e+00 1.20921338e+00 7.77492821e-01
4.47068512e-02 -7.53060102e-01 5.24082303e-01 4.62400317e-01
-5.34254909e-01 -4.31983888e-01 1.29042417e-01 -1.19787359e+00
2.27568775e-01 -2.28156298e-01 -5.32916188e-01 7.70949125e-02
5.15943289e-01 3.28802943e-01 -3.07634950e-01 1.02753496e+00
-1.85442954e-01 6.64604366e-01 -2.65005976e-01 -3.25751692e-01
5.17171264e-01 -4.90440100e-01 6.54876530e-01 1.00525963e+00
-1.19878247e-01 5.43393850e-01 -4.14709270e-01 5.56865931e-01
7.84569383e-01 1.71077088e-01 -1.05502856e+00 -4.23646212e-01
6.08965397e-01 1.07928860e+00 -7.86709130e-01 4.89233404e-01
-3.65608424e-01 1.99161887e-01 -6.05886459e-01 3.16723615e-01
-6.50326431e-01 -7.38460243e-01 7.01619327e-01 5.45970261e-01
-2.10157931e-01 -4.30167705e-01 -6.27219558e-01 -4.57723022e-01
-6.23991728e-01 -1.14727521e+00 -3.90175253e-01 -4.64729309e-01
-8.50617111e-01 3.63994002e-01 -4.98071574e-02 -8.88732791e-01
-1.35751352e-01 -3.60462099e-01 -9.44519281e-01 3.80396724e-01
-9.22460675e-01 -1.69852805e+00 -2.18163371e-01 -1.94712907e-01
-6.33158162e-02 -2.27201298e-01 6.59260571e-01 -4.83179986e-02
-3.20022255e-01 1.41641852e-02 3.18705618e-01 -1.31235532e-02
1.43794224e-01 -9.25761402e-01 3.60675961e-01 6.97028160e-01
-3.78868490e-01 6.94379926e-01 1.09566164e+00 -7.75756419e-01
-1.92088497e+00 -3.68853003e-01 7.59332299e-01 -4.36614841e-01
1.24460542e+00 -1.08705536e-01 -4.29332048e-01 7.32280135e-01
9.17706728e-01 -8.65772486e-01 1.14981711e+00 -1.11648351e-01
2.81812042e-01 2.55088154e-02 -1.02174985e+00 1.07617593e+00
9.68039989e-01 -3.82768482e-01 -8.16029608e-01 1.43015712e-01
5.90077221e-01 1.56681791e-01 -1.13558650e+00 -1.10118836e-01
1.13171053e+00 -6.97414517e-01 4.55136269e-01 -4.35754359e-01
1.57541037e-01 -4.33827698e-01 -2.29786366e-01 -8.91740024e-01
-3.09993774e-01 -1.36972404e+00 8.57317805e-01 1.30471301e+00
5.16430378e-01 -9.03758407e-01 5.69826543e-01 8.77692461e-01
-2.82980472e-01 -4.19511080e-01 -8.97633731e-01 -5.42811096e-01
4.99310166e-01 -9.49525088e-02 7.95758486e-01 1.18511188e+00
7.09066927e-01 1.15357846e-01 -1.81062624e-01 -4.06777561e-02
4.31508064e-01 -5.60339689e-01 8.47318232e-01 -1.45329642e+00
2.86050290e-02 -6.25629246e-01 -6.01413548e-01 1.99711204e-01
-1.69352382e-01 -3.89819503e-01 4.37260628e-01 -1.70932114e+00
-3.16383690e-01 -2.07493320e-01 2.10140571e-01 2.47660592e-01
5.83258629e-01 -1.67955354e-01 3.21238577e-01 2.67222196e-01
6.39641210e-02 3.57222915e-01 9.75381434e-01 3.37151289e-01
-6.30488455e-01 -7.49842644e-01 -1.23709476e+00 1.10023177e+00
8.22623253e-01 -4.09575731e-01 -4.45350558e-01 -2.55528271e-01
1.20119965e+00 6.21705651e-02 3.95497829e-01 -1.23711216e+00
3.64686430e-01 -1.19941962e+00 1.93550706e-01 1.01926982e-01
-3.02401125e-01 -1.27183306e+00 1.03506494e+00 9.22735929e-01
-2.36964181e-01 1.23572767e-01 1.15820639e-01 -4.74988669e-02
1.66301310e-01 -1.16707444e-01 4.74923939e-01 -5.68541102e-02
-6.47609532e-01 -4.47376400e-01 -1.12310624e+00 2.12976802e-02
1.39024198e+00 -5.76973796e-01 -7.98501432e-01 -3.29580486e-01
-2.45585665e-01 2.90806949e-01 1.38027298e+00 -6.74968138e-02
5.86162984e-01 -8.08855176e-01 -8.13408196e-01 9.37885195e-02
-1.85845017e-01 -4.68017191e-01 1.44842297e-01 6.66670859e-01
-1.63241708e+00 -2.39645764e-01 -5.84778070e-01 2.73388207e-01
-4.58373547e-01 1.37437999e-01 5.24146855e-01 7.27433622e-01
-1.01283991e+00 6.07625067e-01 -2.50337511e-01 -1.62394986e-01
-1.42548755e-01 4.03306544e-01 -3.44687432e-01 3.28917801e-01
2.85786450e-01 5.61893404e-01 -7.73310602e-01 -9.63687003e-01
-4.85470027e-01 5.34009278e-01 6.41980886e-01 -6.18314326e-01
1.72889411e+00 -2.19603509e-01 -3.00477207e-01 6.03937387e-01
7.56765127e-01 -1.25291124e-02 -1.13336205e+00 5.84744990e-01
2.04260930e-01 -7.35458314e-01 2.26587892e-01 -1.15319514e+00
-3.40138078e-01 5.12710631e-01 5.08510232e-01 3.07765186e-01
1.16971505e+00 -4.66399759e-01 3.89753073e-01 4.43837583e-01
4.36750203e-01 -1.42734551e+00 -1.86183095e-01 4.40917537e-02
1.24605381e+00 -3.71414959e-01 2.86973476e-01 1.94712862e-01
-8.33308101e-01 1.20959067e+00 4.77664858e-01 -9.47273448e-02
8.20212603e-01 5.59223413e-01 4.93028700e-01 -2.72191077e-01
-9.38664973e-01 3.31115387e-02 -6.63215339e-01 1.00711417e+00
3.32213730e-01 2.59940416e-01 -6.26370490e-01 4.23752874e-01
-3.61988068e-01 6.34688959e-02 1.04028821e+00 1.32905865e+00
-7.34787464e-01 -9.54824567e-01 -5.59567750e-01 7.05265924e-02
-3.33938360e-01 9.06993002e-02 -8.08164001e-01 1.35668409e+00
4.59180385e-01 1.10139143e+00 9.91110504e-02 -8.13742638e-01
4.31585401e-01 -2.74972647e-01 -8.86978731e-02 -8.64444971e-02
-1.33936548e+00 1.46015555e-01 3.90297890e-01 -1.66349277e-01
-3.42793494e-01 -8.40554237e-01 -1.19513619e+00 -8.80542338e-01
-1.38717845e-01 3.89965951e-01 8.34874451e-01 7.27028131e-01
6.51809871e-02 3.32408637e-01 9.05037403e-01 -7.13352323e-01
-1.03920080e-01 -8.17984700e-01 -8.07684064e-01 -4.12660062e-01
2.26216018e-01 -1.94242358e-01 -6.96297526e-01 2.84808315e-03] | [9.035406112670898, 6.235238075256348] |
2283f811-ee8e-4ef0-88d4-03dfd81faff0 | improving-pixel-level-contrastive-learning-by | 2211.10177 | null | https://arxiv.org/abs/2211.10177v1 | https://arxiv.org/pdf/2211.10177v1.pdf | Improving Pixel-Level Contrastive Learning by Leveraging Exogenous Depth Information | Self-supervised representation learning based on Contrastive Learning (CL) has been the subject of much attention in recent years. This is due to the excellent results obtained on a variety of subsequent tasks (in particular classification), without requiring a large amount of labeled samples. However, most reference CL algorithms (such as SimCLR and MoCo, but also BYOL and Barlow Twins) are not adapted to pixel-level downstream tasks. One existing solution known as PixPro proposes a pixel-level approach that is based on filtering of pairs of positive/negative image crops of the same image using the distance between the crops in the whole image. We argue that this idea can be further enhanced by incorporating semantic information provided by exogenous data as an additional selection filter, which can be used (at training time) to improve the selection of the pixel-level positive/negative samples. In this paper we will focus on the depth information, which can be obtained by using a depth estimation network or measured from available data (stereovision, parallax motion, LiDAR, etc.). Scene depth can provide meaningful cues to distinguish pixels belonging to different objects based on their depth. We show that using this exogenous information in the contrastive loss leads to improved results and that the learned representations better follow the shapes of objects. In addition, we introduce a multi-scale loss that alleviates the issue of finding the training parameters adapted to different object sizes. We demonstrate the effectiveness of our ideas on the Breakout Segmentation on Borehole Images where we achieve an improvement of 1.9\% over PixPro and nearly 5\% over the supervised baseline. We further validate our technique on the indoor scene segmentation tasks with ScanNet and outdoor scenes with CityScapes ( 1.6\% and 1.1\% improvement over PixPro respectively). | ['Gabriele Facciolo', 'Adrien Courtois', 'Axel Davy', 'Josselin Kherroubi', 'Kristina Prokopetc', 'Ahmed Ben Saad'] | 2022-11-18 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 6.52629077e-01 -5.90770058e-02 -3.25545929e-02 -4.58666980e-01
-8.08734655e-01 -6.78953588e-01 4.91801858e-01 3.71391386e-01
-6.29660785e-01 7.01632440e-01 -2.93785810e-01 -9.84200761e-02
-2.73255557e-01 -1.11173093e+00 -9.09744322e-01 -1.02812445e+00
1.01570234e-01 2.58085042e-01 5.36530674e-01 -2.38964438e-01
2.48641968e-01 6.65763676e-01 -1.74463224e+00 2.32125610e-01
1.02990794e+00 1.12507915e+00 6.25684738e-01 4.37101364e-01
-2.07938761e-01 4.04143631e-01 -3.09834629e-01 7.52701610e-02
6.13003731e-01 -2.88442135e-01 -5.97240448e-01 3.44899595e-01
6.56452417e-01 -2.84557510e-03 2.00770676e-01 1.15162551e+00
3.70141387e-01 1.88928470e-01 7.70146191e-01 -1.02747440e+00
-1.96412280e-01 3.43734711e-01 -9.43350315e-01 9.96301770e-02
-1.94574490e-01 2.77516730e-02 9.35595512e-01 -6.83171034e-01
4.06296879e-01 1.16456914e+00 6.37374818e-01 2.51959801e-01
-1.28279555e+00 -6.57279789e-01 3.26222241e-01 1.24732293e-01
-1.17631626e+00 -1.82394952e-01 8.88907254e-01 -4.15596455e-01
5.97915828e-01 1.39266387e-01 2.91345000e-01 5.78041613e-01
-1.44470319e-01 8.59536648e-01 1.38450098e+00 -7.27459192e-01
2.80496508e-01 3.73722225e-01 2.23918870e-01 6.25337780e-01
2.40988612e-01 1.50474593e-01 -4.22014296e-02 2.25012168e-01
7.62385666e-01 1.55850488e-03 -2.85095483e-01 -6.83587253e-01
-7.37394929e-01 1.03229535e+00 9.08255577e-01 3.21011961e-01
-3.28807801e-01 -4.00163792e-03 2.42789790e-01 1.94384769e-01
6.34373188e-01 4.11349088e-01 -7.37222016e-01 5.85141718e-01
-9.68664944e-01 -1.02433003e-03 3.31715077e-01 5.32946348e-01
1.18253469e+00 -1.06213810e-02 5.43639474e-02 9.79176104e-01
3.33641499e-01 5.10216117e-01 3.31246197e-01 -9.39903378e-01
4.86847013e-01 6.65923893e-01 6.65575936e-02 -1.00661385e+00
-4.25514668e-01 -4.51816648e-01 -7.29246736e-01 7.73240030e-01
6.44866049e-01 -1.45229965e-01 -1.17228413e+00 1.62006009e+00
2.28187814e-01 7.80220255e-02 1.42363682e-01 7.46465504e-01
6.63625598e-01 7.45019853e-01 9.49963033e-02 -9.45782140e-02
1.14482355e+00 -9.24646318e-01 -1.73396260e-01 -4.88348544e-01
6.56492651e-01 -5.56031406e-01 1.06894839e+00 5.62872648e-01
-8.46471369e-01 -6.74913466e-01 -1.09959090e+00 1.53592601e-01
-6.95834100e-01 2.85213172e-01 7.32961535e-01 7.07308412e-01
-8.51888478e-01 7.59850323e-01 -7.27747679e-01 -4.49481726e-01
5.97068071e-01 3.60699862e-01 -3.64598960e-01 -1.43553600e-01
-7.74363339e-01 6.12720251e-01 4.00490671e-01 8.35928544e-02
-8.11191201e-01 -6.00581884e-01 -9.24482286e-01 2.03901883e-02
4.43378359e-01 -1.87120631e-01 7.49531865e-01 -1.34709477e+00
-1.23607206e+00 1.04702902e+00 -7.80215673e-03 -7.37465858e-01
5.45617998e-01 -2.37706721e-01 1.03507616e-01 3.25209171e-01
3.28492045e-01 1.07973504e+00 8.08057308e-01 -1.58422172e+00
-9.92384315e-01 -5.73460996e-01 2.32588708e-01 2.55960941e-01
-6.72590733e-02 -2.79158205e-01 -2.84669548e-01 -6.64380312e-01
4.87225175e-01 -9.05092478e-01 -4.33473587e-01 2.43716270e-01
-1.78496823e-01 2.33433433e-02 7.74825871e-01 -4.76524562e-01
4.93421942e-01 -2.26584840e+00 -4.64874022e-02 1.44766316e-01
-2.14493349e-01 3.12080085e-01 -1.83485329e-01 4.54940088e-02
-2.32580766e-01 8.11399743e-02 -7.78564036e-01 -2.63877869e-01
-3.88741285e-01 2.22837493e-01 -1.96557984e-01 5.90601146e-01
4.08036143e-01 5.65959811e-01 -7.52216160e-01 -3.80041182e-01
6.13097548e-01 3.60098720e-01 -3.60888511e-01 3.94332148e-02
-1.99144185e-01 5.82105517e-01 -2.25499555e-01 4.43440378e-01
1.12070441e+00 7.99785703e-02 -1.63141936e-01 -1.94971398e-01
-2.84491807e-01 -4.23055179e-02 -1.28994668e+00 1.45490956e+00
-6.34339869e-01 6.58894539e-01 1.80964455e-01 -1.42895722e+00
1.10414875e+00 -1.39997914e-01 3.59586358e-01 -7.62697458e-01
1.31688174e-02 1.74985111e-01 -6.78663850e-02 -3.10478300e-01
2.70256460e-01 -3.49096805e-02 3.08663696e-02 4.52118963e-02
-2.78680511e-02 -4.27964985e-01 2.24181876e-01 -1.86772913e-01
5.99080741e-01 2.46316940e-01 1.91960603e-01 -3.84880841e-01
7.44086206e-01 2.48727381e-01 5.13928413e-01 7.97270417e-01
-2.01203257e-01 8.78359258e-01 3.34807128e-01 -1.43739462e-01
-8.60794783e-01 -9.55458105e-01 -3.80218714e-01 9.99087930e-01
3.08719128e-01 2.44615704e-01 -5.82795382e-01 -6.72337890e-01
5.74132986e-02 6.08566225e-01 -6.76093996e-01 2.75095142e-02
-6.40134215e-01 -1.12374270e+00 2.51117378e-01 7.03918636e-01
9.77133751e-01 -1.15631175e+00 -8.65371108e-01 1.25842482e-01
9.66516882e-02 -1.10889053e+00 2.71796137e-01 9.02808845e-01
-1.12732470e+00 -9.67534661e-01 -7.00707674e-01 -8.73918056e-01
6.59610331e-01 4.85854238e-01 8.25044274e-01 -1.42921895e-01
-2.13411197e-01 1.78424448e-01 -4.46026623e-01 -2.63398290e-01
-8.46337825e-02 3.93425748e-02 -4.19947594e-01 1.24969959e-01
8.82852897e-02 -4.10260767e-01 -5.08913219e-01 2.70689368e-01
-9.24301744e-01 5.58236986e-02 5.77909768e-01 8.80239606e-01
7.47483373e-01 2.42038876e-01 4.51576263e-01 -1.21197224e+00
-7.58658573e-02 -3.03485662e-01 -8.57033908e-01 1.22024380e-01
-6.14200115e-01 8.85660723e-02 4.60838437e-01 -2.14460984e-01
-1.20471442e+00 4.49197829e-01 -1.72937140e-01 -2.65710115e-01
-4.84890074e-01 2.03587979e-01 -2.54456848e-01 -1.95153207e-01
6.13407195e-01 2.08249744e-02 -2.03671858e-01 -4.08492208e-01
4.38619018e-01 5.21891415e-01 4.07465905e-01 -4.02600050e-01
7.11929560e-01 8.72617722e-01 4.55962792e-02 -8.30689192e-01
-9.19191957e-01 -6.85479760e-01 -6.98892534e-01 5.17002568e-02
9.08229947e-01 -9.78186190e-01 -3.25028628e-01 3.92912030e-01
-7.92753637e-01 -5.12264132e-01 -4.31500524e-01 4.46531624e-01
-4.69380409e-01 3.80077958e-01 -3.72847140e-01 -1.03883004e+00
-2.01271683e-01 -1.08658934e+00 1.04006302e+00 4.71100748e-01
3.80583465e-01 -9.19285595e-01 -2.87061095e-01 2.22573176e-01
1.74983129e-01 3.85047197e-01 9.62763250e-01 -2.75581121e-01
-5.84541202e-01 2.50359606e-02 -6.02957010e-01 7.26272821e-01
1.37996241e-01 -8.03828537e-02 -1.28274548e+00 -8.91882181e-02
6.11274317e-02 -2.52628088e-01 1.43587840e+00 8.35960865e-01
1.18660367e+00 2.24848479e-01 -3.46509933e-01 6.92730606e-01
1.85105228e+00 2.23665848e-01 6.19862795e-01 4.78358924e-01
7.57310390e-01 9.51966465e-01 8.65069389e-01 1.99682996e-01
6.60929158e-02 5.06071866e-01 8.11123252e-01 -4.34198171e-01
-1.74702778e-01 3.40903066e-02 1.28308564e-01 9.71636996e-02
-3.15885246e-02 -2.51113355e-01 -7.16027200e-01 7.80999482e-01
-1.88021970e+00 -7.70135403e-01 -3.20404708e-01 2.34838438e+00
4.89214599e-01 3.01781267e-01 -6.92041144e-02 5.03994524e-01
6.99651301e-01 1.03819132e-01 -5.29291034e-01 -3.01353037e-01
-5.63117504e-01 4.31936592e-01 9.79753554e-01 5.22204220e-01
-1.48257565e+00 1.01300836e+00 5.25069952e+00 8.03061068e-01
-1.25769329e+00 -1.03199035e-01 9.27649260e-01 4.46082413e-01
2.42236722e-02 4.21900004e-02 -8.33566964e-01 2.58594424e-01
3.66628975e-01 3.28222603e-01 1.48228124e-01 8.93637240e-01
2.13931888e-01 -6.32871330e-01 -8.72411668e-01 9.97274041e-01
8.37383792e-02 -1.09997082e+00 -2.11162671e-01 -1.62834451e-01
8.90341699e-01 8.38610679e-02 2.27052718e-02 2.63083935e-01
2.02107728e-01 -8.08954835e-01 5.43243229e-01 1.46729290e-01
3.89095038e-01 -6.66080117e-01 8.20999622e-01 3.68540734e-01
-1.20988822e+00 -2.91225225e-01 -6.49129808e-01 -9.37966853e-02
-1.39292613e-01 7.44735777e-01 -7.53016293e-01 6.94350779e-01
8.94496799e-01 8.31565738e-01 -7.66411364e-01 1.11989832e+00
-2.84562528e-01 6.34154618e-01 -4.97190148e-01 2.30077595e-01
4.68908727e-01 -3.77470493e-01 1.84591055e-01 1.10696101e+00
1.97039157e-01 -1.72398791e-01 2.59511352e-01 8.01801801e-01
1.27441511e-01 1.45322412e-01 -6.65997624e-01 4.33784783e-01
6.85927048e-02 1.39869249e+00 -1.05860054e+00 -1.82841033e-01
-2.71939665e-01 1.00363302e+00 8.99208784e-02 2.98714191e-01
-5.23877800e-01 -3.96501034e-01 3.67386281e-01 1.33144870e-01
6.24924839e-01 -1.89129002e-02 -5.03830016e-01 -7.38681912e-01
-1.18447192e-01 -3.96598101e-01 3.88955414e-01 -7.07435787e-01
-1.36803198e+00 3.83522570e-01 2.22739149e-02 -1.25618696e+00
-8.44957158e-02 -7.99225926e-01 -5.73901296e-01 7.75194347e-01
-1.96537757e+00 -9.93798852e-01 -6.09769642e-01 3.85425419e-01
6.79642856e-01 1.88747138e-01 5.56544483e-01 2.50265390e-01
-3.82595628e-01 2.34008104e-01 2.30021194e-01 7.35527650e-02
5.73409677e-01 -1.39428115e+00 -4.25822325e-02 8.30543399e-01
1.69248834e-01 2.09408533e-02 5.31960607e-01 -2.28733286e-01
-7.64820993e-01 -1.28119671e+00 5.15564561e-01 -2.57090479e-01
2.85512269e-01 -2.85442084e-01 -9.12560940e-01 3.81161243e-01
1.19132493e-02 2.51838654e-01 2.57081896e-01 -3.48947138e-01
-1.02597706e-01 -3.67592096e-01 -1.41866803e+00 2.75014848e-01
8.76689017e-01 -2.36741439e-01 -2.74375826e-01 2.41535172e-01
6.75481975e-01 -9.09952819e-02 -4.00815040e-01 8.08544397e-01
2.68569648e-01 -1.17832577e+00 1.03591585e+00 -1.46784127e-01
3.91077816e-01 -4.56763923e-01 -3.65360558e-01 -1.11820173e+00
-7.89580345e-02 1.56851172e-01 6.04172111e-01 1.34398079e+00
2.84954250e-01 -5.60986221e-01 9.70871449e-01 9.79242921e-02
-7.81954378e-02 -5.75580895e-01 -6.78477108e-01 -7.63943672e-01
2.24146768e-01 -4.94690448e-01 1.24341458e-01 8.25378478e-01
-6.65705740e-01 2.34381035e-01 3.29405181e-02 4.53337342e-01
5.58319986e-01 3.47148597e-01 6.22144818e-01 -1.50690758e+00
-2.46292755e-01 -3.52511108e-01 -3.66964877e-01 -1.03106618e+00
1.65423810e-01 -8.57943177e-01 3.35182428e-01 -1.52974010e+00
-1.37699276e-01 -9.28233206e-01 -3.14240456e-01 4.37448710e-01
-8.26900899e-02 4.84056592e-01 2.26454452e-01 2.38513555e-02
-2.04520866e-01 3.88490051e-01 1.01378119e+00 -3.06547970e-01
-3.55126202e-01 5.18999659e-02 -5.04617989e-01 9.35139239e-01
1.02575433e+00 -4.84189659e-01 -3.75137985e-01 -4.24519151e-01
-5.06697828e-03 -2.36349016e-01 5.46879947e-01 -1.21842670e+00
-9.36041996e-02 -6.76074158e-03 6.15208983e-01 -6.49549186e-01
3.60637695e-01 -8.56505692e-01 -3.14706385e-01 5.27457297e-01
-1.85521528e-01 -3.21901441e-01 2.91793108e-01 5.27170897e-01
-2.08707452e-01 -6.58913374e-01 1.07609785e+00 -3.54917914e-01
-9.26289320e-01 -1.21884216e-02 -9.47509557e-02 -2.14026690e-01
1.01788855e+00 -3.32555830e-01 -1.65576816e-01 -1.07103884e-01
-6.07767582e-01 6.77617788e-02 3.69060487e-01 8.80091712e-02
3.41883391e-01 -8.63454461e-01 -6.00791037e-01 2.68197834e-01
1.38584614e-01 3.52362514e-01 1.26596708e-02 7.02691317e-01
-6.46625996e-01 2.17491314e-01 -3.73019725e-01 -8.07193398e-01
-1.13476753e+00 4.20794100e-01 4.17766660e-01 -2.74479151e-01
-4.32017565e-01 9.50007439e-01 6.39484942e-01 -4.19520020e-01
1.94979280e-01 -5.21552026e-01 -5.08129001e-01 2.20537052e-01
2.29989082e-01 2.50544310e-01 1.83958396e-01 -4.67277735e-01
-3.85957867e-01 8.90510380e-01 4.88104038e-02 -5.09100258e-02
1.59274542e+00 -1.60246372e-01 7.98242539e-02 4.82313633e-01
1.07418215e+00 8.27276483e-02 -1.63081205e+00 -1.14466444e-01
-3.01762708e-02 -4.45694476e-01 2.02164084e-01 -6.89840257e-01
-1.38671017e+00 1.10709620e+00 1.18310452e+00 1.31123811e-01
1.31994641e+00 3.96220051e-02 4.39561009e-01 3.38587970e-01
4.17750329e-01 -9.68065858e-01 4.19838447e-03 2.70485073e-01
6.03684723e-01 -1.61616182e+00 -5.83318882e-02 -7.19708502e-01
-5.00083566e-01 1.03442395e+00 5.41220009e-01 -6.15961075e-01
5.71695328e-01 1.85597688e-01 1.59808591e-01 -2.23364308e-02
-2.30151907e-01 -7.02115774e-01 1.50460554e-02 8.28989983e-01
3.40544552e-01 7.55289942e-03 -3.52146655e-01 2.56420612e-01
2.06909359e-01 -3.11162829e-01 4.10885632e-01 8.38819623e-01
-7.02979684e-01 -1.16144133e+00 -4.50137168e-01 4.71398354e-01
-2.82555282e-01 -4.12266515e-02 -1.48838326e-01 7.33611643e-01
5.32773972e-01 9.54160631e-01 1.76277786e-01 -2.15973929e-02
2.53515095e-01 -2.56327331e-01 5.75570166e-01 -6.74333572e-01
-5.83974481e-01 1.10913992e-01 -2.02771887e-01 -2.32531950e-01
-7.27938294e-01 -7.03513920e-01 -1.33787704e+00 2.01130658e-01
-5.68874776e-01 -1.16598219e-01 6.99128509e-01 6.89161479e-01
-1.37646005e-01 4.12780881e-01 7.83545971e-01 -1.05963874e+00
-1.60957024e-01 -7.75004566e-01 -6.83633089e-01 3.90609115e-01
2.96309590e-01 -7.35517681e-01 -6.07507467e-01 8.33383352e-02] | [9.45419979095459, -0.8937613368034363] |
defc6a22-a018-4499-8f91-d15d25c49dff | comprehensive-event-representations-using | 2303.04794 | null | https://arxiv.org/abs/2303.04794v1 | https://arxiv.org/pdf/2303.04794v1.pdf | Comprehensive Event Representations using Event Knowledge Graphs and Natural Language Processing | Recent work has utilised knowledge-aware approaches to natural language understanding, question answering, recommendation systems, and other tasks. These approaches rely on well-constructed and large-scale knowledge graphs that can be useful for many downstream applications and empower knowledge-aware models with commonsense reasoning. Such knowledge graphs are constructed through knowledge acquisition tasks such as relation extraction and knowledge graph completion. This work seeks to utilise and build on the growing body of work that uses findings from the field of natural language processing (NLP) to extract knowledge from text and build knowledge graphs. The focus of this research project is on how we can use transformer-based approaches to extract and contextualise event information, matching it to existing ontologies, to build a comprehensive knowledge of graph-based event representations. Specifically, sub-event extraction is used as a way of creating sub-event-aware event representations. These event representations are then further enriched through fine-grained location extraction and contextualised through the alignment of historically relevant quotes. | ['Tin Kuculo'] | 2023-03-08 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 3.91546845e-01 7.48007655e-01 -3.00431818e-01 -3.97049993e-01
-4.50664818e-01 -5.89015305e-01 1.15779972e+00 1.20434380e+00
-2.68991530e-01 6.91121638e-01 9.32893634e-01 -4.57603335e-01
-7.34061360e-01 -1.29902363e+00 -4.42015231e-01 2.91021645e-01
-3.09877366e-01 4.58271295e-01 4.20455545e-01 -6.52976453e-01
4.96190518e-01 5.91327488e-01 -1.78074253e+00 6.39801502e-01
5.03619671e-01 7.37819850e-01 -2.05018356e-01 4.72414821e-01
-8.59962523e-01 1.60975301e+00 -4.54742253e-01 -5.82146466e-01
-3.17739159e-01 -4.06072587e-01 -1.53247797e+00 -2.81423360e-01
-2.61762530e-01 3.27621609e-01 -1.20290644e-01 5.81916213e-01
1.26112312e-01 4.99890596e-01 3.34524393e-01 -1.24227977e+00
-6.33964837e-01 1.11159468e+00 1.56640142e-01 5.45213580e-01
1.10060108e+00 -4.32738960e-01 1.03417063e+00 -5.68558097e-01
1.15248096e+00 1.43002284e+00 6.78043962e-01 -5.82015663e-02
-7.69671202e-01 -3.11340213e-01 -1.49307633e-02 7.82757759e-01
-1.15237951e+00 -2.61110634e-01 7.84425557e-01 -3.83287489e-01
1.67616487e+00 3.45789701e-01 9.10672009e-01 8.84344637e-01
-7.75513202e-02 2.69113570e-01 8.53311002e-01 -1.18793142e+00
1.16710581e-01 5.87422103e-02 3.23977709e-01 5.35862625e-01
2.64660925e-01 -2.27207150e-02 -1.01409519e+00 -1.85004607e-01
6.48123622e-01 1.04604095e-01 6.37628213e-02 4.86487038e-02
-1.20567310e+00 8.28524947e-01 2.84045577e-01 9.30342197e-01
-1.04137206e+00 -1.30733252e-02 5.59876025e-01 2.76546508e-01
3.78449708e-01 7.78781116e-01 -5.65930784e-01 -2.46194929e-01
-7.97326744e-01 4.26141411e-01 1.33925247e+00 7.73608208e-01
9.18260098e-01 -2.44635135e-01 -1.44181341e-01 6.23242199e-01
4.71106112e-01 -1.26321226e-01 5.30680537e-01 -8.15099657e-01
4.75005507e-01 1.41095638e+00 5.03201820e-02 -1.22457397e+00
-5.19016862e-01 1.59434080e-01 -7.92712846e-04 -3.15213293e-01
1.92501977e-01 9.98969525e-02 -8.71026039e-01 1.38263583e+00
5.86740851e-01 4.43672717e-01 3.40830654e-01 3.06345314e-01
1.23494601e+00 4.44740981e-01 5.59273481e-01 -1.96387872e-01
1.72617245e+00 -3.41220140e-01 -9.00237560e-01 -1.14771381e-01
7.34756708e-01 -5.36772192e-01 6.96127415e-01 9.75038260e-02
-9.39139009e-01 -3.56925912e-02 -9.34165120e-01 -1.92710713e-01
-1.23981321e+00 -7.58337975e-01 1.08724892e+00 3.77797127e-01
-6.70638442e-01 5.50079167e-01 -6.17578387e-01 -7.71460056e-01
4.35224473e-01 -1.49418443e-01 -3.86979103e-01 -1.69235513e-01
-1.83999300e+00 1.41275585e+00 1.12353861e+00 -3.19433987e-01
-3.12952578e-01 -9.32772040e-01 -1.27765334e+00 -4.87294346e-02
9.24177647e-01 -7.51848102e-01 9.98955965e-01 -6.71578109e-01
-8.91391098e-01 9.32283282e-01 -4.09373008e-02 -8.40024233e-01
-3.64782125e-01 -1.34919599e-01 -1.06355631e+00 3.01752746e-01
3.37064296e-01 -1.84646137e-02 2.69921452e-01 -9.75971818e-01
-8.05620849e-01 -5.32403767e-01 3.85947078e-01 1.44645557e-01
-2.48247623e-01 6.61789179e-01 -8.70981514e-02 -5.64571559e-01
-1.55008823e-01 -3.31299722e-01 -9.25905779e-02 -9.02615607e-01
-2.43195832e-01 -5.43613195e-01 5.67411542e-01 -9.12493110e-01
1.76053703e+00 -1.59116387e+00 -1.43249780e-01 5.79586625e-01
1.81724131e-01 -4.79954258e-02 3.36479366e-01 1.34294307e+00
-1.88206092e-01 2.76667982e-01 1.14980623e-01 5.23318887e-01
4.15357679e-01 6.26598477e-01 -4.95113134e-01 -4.38417733e-01
3.57480794e-01 1.09001672e+00 -1.22177863e+00 -6.32994771e-01
4.10139173e-01 4.42560852e-01 -2.63981968e-01 -1.51985645e-01
-2.96133816e-01 -1.28162950e-01 -6.72182620e-01 4.94654059e-01
-3.85872573e-01 -2.54414994e-02 5.41957617e-01 -4.07903552e-01
-1.87342688e-01 7.74611473e-01 -1.32117307e+00 1.61999965e+00
-5.54940343e-01 5.71012616e-01 -6.45366609e-01 -1.12885857e+00
9.68336582e-01 5.72508097e-01 4.97119099e-01 -6.68685973e-01
6.48850948e-02 -8.72382522e-03 -4.07218814e-01 -7.11621106e-01
7.73300648e-01 -4.05132890e-01 -3.89355659e-01 3.47355753e-01
3.52634311e-01 -1.64941207e-01 5.54637492e-01 6.81558013e-01
1.24923527e+00 3.52745056e-01 1.06096447e+00 -9.99662951e-02
5.55596769e-01 5.49118340e-01 2.31521189e-01 3.52620691e-01
3.42124701e-01 -1.35389939e-01 4.24223483e-01 -5.71225405e-01
-8.05972576e-01 -8.01907241e-01 5.87001368e-02 1.10319030e+00
-1.40690550e-01 -1.09223759e+00 -3.59935373e-01 -4.49006349e-01
-1.73820388e-02 1.27154362e+00 -6.56577528e-01 -4.52969335e-02
-4.49752808e-01 -3.55379552e-01 6.83325946e-01 5.76923370e-01
4.03335318e-02 -1.47895467e+00 -8.59231710e-01 6.87925220e-01
-3.57697964e-01 -1.03755903e+00 5.55585742e-01 1.08999610e-01
-5.22826433e-01 -1.57779109e+00 2.60709137e-01 -5.67544878e-01
8.99218470e-02 -2.65913457e-01 1.56262922e+00 8.56664032e-02
-4.80815023e-01 7.09421873e-01 -8.24877918e-01 -1.18510938e+00
-4.95540142e-01 -1.49340719e-01 -3.46592069e-01 -2.93140322e-01
9.99929965e-01 -8.89983058e-01 -1.92591563e-01 -1.72233313e-01
-1.29165018e+00 -2.31204987e-01 1.63165897e-01 2.41102993e-01
5.66172242e-01 4.53538805e-01 9.65597570e-01 -9.82843816e-01
1.02555835e+00 -9.32181180e-01 3.50597389e-02 5.00877142e-01
-6.76998019e-01 1.81306928e-01 1.08275369e-01 -1.84389234e-01
-1.42135274e+00 -3.49149257e-01 2.24243209e-01 7.01450109e-02
-2.05574647e-01 1.27192652e+00 1.10457577e-02 3.25800478e-01
1.04550993e+00 -2.06982002e-01 -3.96118075e-01 -1.12716362e-01
9.79404092e-01 3.96648288e-01 7.19631612e-01 -8.34322512e-01
5.34840107e-01 4.43105757e-01 1.16047762e-01 -6.81019425e-01
-9.03929591e-01 -6.57046795e-01 -7.58920252e-01 -3.85991305e-01
7.66975760e-01 -6.11708641e-01 -4.70452160e-01 -2.44783878e-01
-7.72490859e-01 -2.02656418e-01 -1.08339620e+00 3.99949402e-01
-3.91126871e-01 2.92746782e-01 -1.27626091e-01 -8.02033365e-01
-3.53558034e-01 2.75812205e-02 7.68154919e-01 3.68865401e-01
-7.65220284e-01 -1.40357184e+00 2.01549605e-01 3.43817860e-01
1.82741731e-01 8.52003336e-01 1.06610036e+00 -1.20588732e+00
-3.28255743e-01 -2.98496276e-01 -2.43480615e-02 -3.47115308e-01
1.49023861e-01 -1.63390875e-01 -5.17905474e-01 5.31217277e-01
-3.12362880e-01 -2.12019041e-01 2.45301276e-01 -4.80317026e-02
4.25176948e-01 -5.05158782e-01 -4.44338769e-01 -1.26257062e-01
1.52090180e+00 3.47640008e-01 9.78380919e-01 8.92899632e-01
4.65585589e-01 9.04403329e-01 6.11057758e-01 4.79757696e-01
9.56674755e-01 4.52412665e-01 6.23917859e-03 4.08353448e-01
-1.98322296e-01 -4.05019015e-01 -1.71795592e-01 5.01148939e-01
-4.58355367e-01 1.61348119e-01 -1.26379776e+00 1.16528046e+00
-2.13805389e+00 -1.51735091e+00 -3.62288624e-01 1.70230222e+00
1.07551539e+00 2.12245181e-01 -4.95695919e-02 4.01951641e-01
4.83009815e-01 5.31339506e-03 7.89047685e-03 -6.11361563e-01
3.76359560e-02 7.91696429e-01 2.14850605e-01 2.89800465e-01
-6.68932796e-01 1.02664959e+00 5.77014399e+00 6.40113890e-01
-3.14724475e-01 4.93125394e-02 -2.85858333e-01 1.89084068e-01
-4.79525417e-01 6.59340739e-01 -5.28321803e-01 -3.82085890e-02
1.43093574e+00 -6.55589640e-01 4.96708184e-01 3.99342149e-01
5.95482551e-02 -3.28315556e-01 -1.08774471e+00 7.24509954e-01
9.89428982e-02 -1.86975217e+00 2.51151592e-01 -1.00903116e-01
5.38700938e-01 -2.45293424e-01 -9.54876900e-01 4.89837646e-01
7.98084199e-01 -9.15945649e-01 5.79509258e-01 1.04295802e+00
3.22188497e-01 -7.99664319e-01 6.86502755e-01 3.83566655e-02
-1.58302546e+00 -9.34346914e-02 4.84115668e-02 -3.89718652e-01
7.02038586e-01 6.29545331e-01 -1.05206192e+00 1.42602146e+00
7.32312977e-01 7.05746651e-01 -5.46987176e-01 7.48231769e-01
-5.48963845e-01 5.85088849e-01 -4.24696654e-01 1.22900650e-01
4.54990827e-02 1.01209641e-01 3.59623790e-01 1.44101620e+00
5.49371205e-02 5.63161075e-01 1.77434400e-01 6.75540745e-01
-1.45260440e-02 2.80771047e-01 -7.51617134e-01 -3.32326531e-01
7.65329063e-01 1.12579453e+00 -8.09327543e-01 -7.07022130e-01
-3.53479862e-01 3.39965641e-01 2.61402041e-01 1.89336851e-01
-4.28803742e-01 -8.33734035e-01 2.02505693e-01 1.72003567e-01
2.37893254e-01 -3.45533155e-02 -4.05862816e-02 -9.05204117e-01
-2.20880181e-01 -6.11743808e-01 1.00087309e+00 -1.14655042e+00
-1.34771740e+00 3.31322014e-01 6.51713610e-01 -3.29306036e-01
-8.46597791e-01 -1.64290473e-01 -6.89091325e-01 6.92634046e-01
-1.28436887e+00 -1.53374279e+00 -2.67347515e-01 8.78823459e-01
8.63672420e-02 2.63810724e-01 1.10512865e+00 3.01220030e-01
-7.73238204e-03 -1.55032992e-01 -7.51673162e-01 1.34032354e-01
2.90347606e-01 -1.30950761e+00 2.56028533e-01 8.57861280e-01
5.00394821e-01 7.28839517e-01 5.14078498e-01 -1.12961996e+00
-1.29495394e+00 -9.74493802e-01 1.54684961e+00 -9.54532266e-01
1.29779935e+00 2.55005360e-01 -1.07280493e+00 1.05473161e+00
1.73400924e-01 -1.74356401e-01 9.80654418e-01 3.08257163e-01
-4.18660462e-01 1.46422222e-01 -1.09858549e+00 4.95800853e-01
1.26309633e+00 -9.35381889e-01 -1.75169897e+00 2.04966277e-01
6.91453397e-01 -4.37659658e-02 -1.57591283e+00 2.21907780e-01
3.63449186e-01 -4.92641211e-01 1.20262873e+00 -1.15884292e+00
1.62761495e-01 -3.23198915e-01 -1.44014552e-01 -1.00023019e+00
1.06258867e-02 -7.47657478e-01 -4.27221507e-01 1.75304389e+00
5.21678805e-01 -5.86452425e-01 2.30569080e-01 7.77950346e-01
-9.88550410e-02 -2.32583761e-01 -7.37399638e-01 -5.32682598e-01
-5.48168123e-01 -9.37763035e-01 7.99123108e-01 1.51314569e+00
7.68622816e-01 6.31845355e-01 1.31846324e-01 -5.50619438e-02
3.05180430e-01 1.17938109e-01 5.00484228e-01 -1.73013520e+00
1.93269253e-01 -2.15270489e-01 -7.20470607e-01 1.85088739e-01
5.73841035e-02 -1.11055553e+00 -4.05942708e-01 -2.41881323e+00
-8.20822194e-02 -1.39795914e-01 -1.49711102e-01 8.36405635e-01
1.23770617e-01 -3.27271633e-02 -1.95270628e-02 -1.67451762e-02
-6.64723158e-01 -2.38769148e-02 6.97316110e-01 3.36014092e-01
-4.88745689e-01 -5.96529007e-01 -1.00294936e+00 8.02134395e-01
7.58128405e-01 -5.22791982e-01 -5.08337021e-01 2.57195592e-01
1.00592196e+00 -9.83841419e-02 5.43542266e-01 -6.43694043e-01
3.03341389e-01 -3.07738572e-01 6.32461309e-02 -4.23454612e-01
-5.26480116e-02 -8.46700490e-01 6.77658975e-01 1.16581142e-01
-2.56788611e-01 -1.96789000e-02 3.45518589e-01 5.77707171e-01
-4.46614474e-01 -3.25081408e-01 -1.17690163e-02 -3.29487175e-01
-1.28287935e+00 -3.00727099e-01 -3.42556715e-01 4.27821785e-01
1.09222937e+00 -3.46464336e-01 -2.80496538e-01 -2.00416237e-01
-1.01999509e+00 5.72408922e-02 5.93934022e-02 5.49345791e-01
4.78092164e-01 -1.36001933e+00 -5.81523001e-01 -3.44561815e-01
3.24193150e-01 -1.45953298e-01 3.86131480e-02 5.42140901e-01
-3.37906271e-01 4.11020935e-01 -2.83465862e-01 1.62164316e-01
-1.01421487e+00 7.67702222e-01 -7.61047155e-02 -6.48646235e-01
-8.83319616e-01 3.08868140e-01 -6.36637568e-01 -2.44745314e-01
-1.71178684e-01 -4.64636922e-01 -8.81698132e-01 5.21299958e-01
6.37117147e-01 4.58997190e-01 1.40192941e-01 -5.10635853e-01
-6.79952085e-01 2.90454119e-01 4.37084883e-01 -2.88854003e-01
1.56893408e+00 -1.61018103e-01 -3.31713468e-01 4.06772941e-01
4.32682991e-01 9.05483291e-02 -4.14762288e-01 -3.56024265e-01
8.37567031e-01 -6.29664063e-02 -1.33007661e-01 -1.03466940e+00
-2.78281599e-01 3.45243335e-01 -1.95688486e-01 9.20998991e-01
9.71121371e-01 4.56585646e-01 6.16009474e-01 3.83112401e-01
3.38058710e-01 -1.18340123e+00 -1.36802047e-01 5.41621149e-01
9.70303476e-01 -6.25565946e-01 2.97185898e-01 -5.98758936e-01
-5.36246896e-01 1.21533644e+00 9.99777541e-02 1.20730951e-01
6.73559189e-01 1.17041864e-01 -3.70909065e-01 -1.05764413e+00
-4.82286096e-01 -8.09441805e-01 6.12468600e-01 7.96322644e-01
2.52451897e-01 7.67432079e-02 -4.87116158e-01 1.01401019e+00
-4.34403449e-01 4.23348069e-01 3.16989571e-01 1.37924409e+00
-5.33077836e-01 -1.30425334e+00 -3.17014366e-01 5.71528971e-01
-7.05530524e-01 -2.93720067e-01 -7.27555156e-01 9.81528759e-01
1.74848914e-01 1.16961265e+00 -9.49738026e-02 -1.83410063e-01
6.78803980e-01 5.84727764e-01 5.55865288e-01 -1.02131867e+00
-9.89854693e-01 -7.05468178e-01 1.02952707e+00 -5.75805008e-01
-1.12892985e+00 -5.18148541e-01 -1.63780618e+00 -1.36705324e-01
-2.08644539e-01 3.05028558e-01 6.00333035e-01 1.41032505e+00
5.54354548e-01 9.38637435e-01 -1.82120979e-01 -5.37225544e-01
2.76271641e-01 -7.43111312e-01 -2.10116088e-01 7.34941661e-01
-5.14260948e-01 -6.81150675e-01 9.03356373e-02 4.76235330e-01] | [9.421257019042969, 8.38774585723877] |
ed6aff3e-ca31-4d4f-b0db-d69c7c499c09 | data-dependent-gaussian-prior-objective-for | null | null | https://openreview.net/forum?id=S1efxTVYDr | https://openreview.net/pdf?id=S1efxTVYDr | Data-dependent Gaussian Prior Objective for Language Generation | For typical sequence prediction problems such as language generation, maximum likelihood estimation (MLE) has commonly been adopted as it encourages the predicted sequence most consistent with the ground-truth sequence to have the highest probability of occurring. However, MLE focuses on once-to-all matching between the predicted sequence and gold-standard, consequently treating all incorrect predictions as being equally incorrect. We refer to this drawback as {\it negative diversity ignorance} in this paper. Treating all incorrect predictions as equal unfairly downplays the nuance of these sequences' detailed token-wise structure. To counteract this, we augment the MLE loss by introducing an extra Kullback--Leibler divergence term derived by comparing a data-dependent Gaussian prior and the detailed training prediction. The proposed data-dependent Gaussian prior objective (D2GPo) is defined over a prior topological order of tokens and is poles apart from the data-independent Gaussian prior (L2 regularization) commonly adopted in smoothing the training of MLE. Experimental results show that the proposed method makes effective use of a more detailed prior in the data and has improved performance in typical language generation tasks, including supervised and unsupervised machine translation, text summarization, storytelling, and image captioning.
| ['Kehai Chen', 'Zuchao Li', 'Rui Wang', 'Masso Utiyama', 'Zhuosheng Zhang', 'Eiichiro Sumita', 'Hai Zhao'] | 2020-05-01 | null | null | null | iclr-2020-1 | ['l2-regularization', 'unsupervised-machine-translation'] | ['methodology', 'natural-language-processing'] | [ 7.08054483e-01 5.25208235e-01 -2.07955867e-01 -3.58538270e-01
-9.49521124e-01 -3.86562824e-01 8.49219441e-01 3.15516710e-01
-4.95676666e-01 1.17287958e+00 4.83346134e-01 -2.36836657e-01
2.94132441e-01 -5.42477489e-01 -8.97730470e-01 -7.20943809e-01
3.53694469e-01 4.94040281e-01 5.44773787e-02 -1.32679194e-01
6.10755980e-01 4.14303318e-02 -1.27022994e+00 2.09251940e-01
1.28281868e+00 7.84722447e-01 6.75978601e-01 4.55115050e-01
-2.83187300e-01 8.16841900e-01 -8.27625573e-01 -7.09805369e-01
2.92255525e-02 -7.72743285e-01 -7.19124854e-01 2.12863117e-01
2.15706170e-01 -1.04743168e-01 1.86576732e-02 1.15724754e+00
4.86569703e-01 1.64772660e-01 9.73071814e-01 -1.10466886e+00
-7.67188787e-01 4.95598108e-01 -6.44766033e-01 -4.16521095e-02
4.53111947e-01 1.01779923e-01 9.75061536e-01 -1.00092852e+00
7.33850718e-01 1.17807269e+00 5.81905842e-01 5.10021865e-01
-1.19194674e+00 -3.44270647e-01 -8.16633925e-02 5.88787906e-02
-1.45837581e+00 -4.24793780e-01 6.13356292e-01 -5.22527158e-01
9.63926256e-01 2.79837251e-01 2.86879003e-01 1.22387493e+00
3.59891683e-01 9.04797137e-01 7.90015817e-01 -7.05459237e-01
2.74323285e-01 2.20449165e-01 -2.20111802e-01 5.00119805e-01
1.95881113e-01 -7.06183910e-02 -5.99178851e-01 -3.50398630e-01
4.76750404e-01 -5.66018164e-01 -4.82160866e-01 -1.74531907e-01
-1.31263566e+00 8.76905024e-01 -1.73573896e-01 1.39848560e-01
-6.16856456e-01 -1.09285526e-02 4.66305137e-01 -2.50109844e-02
6.04450166e-01 3.54918510e-01 -2.00406611e-01 -2.72008389e-01
-1.04764926e+00 3.53148520e-01 6.59909725e-01 1.08789444e+00
7.02624321e-01 1.08967632e-01 -4.84095216e-01 1.03926504e+00
2.97050208e-01 3.56696099e-01 7.83182800e-01 -7.75571704e-01
5.70708692e-01 2.35575125e-01 3.84492755e-01 -9.17079091e-01
2.38975242e-01 -4.64490443e-01 -9.75228727e-01 -1.25297725e-01
2.81064719e-01 -2.06023276e-01 -7.26438880e-01 1.98129165e+00
5.12619056e-02 2.34573573e-01 1.23170726e-01 8.61688375e-01
2.45156705e-01 9.11564350e-01 -1.39300781e-03 -6.27883732e-01
9.52358961e-01 -8.70547771e-01 -8.95095229e-01 -2.77521819e-01
6.61851943e-01 -1.00971460e+00 1.02261508e+00 3.39371502e-01
-1.06802249e+00 -3.34560663e-01 -1.00527215e+00 2.73755309e-03
1.22953765e-01 1.43503681e-01 2.17698127e-01 5.79975665e-01
-1.02650332e+00 6.30301833e-01 -4.85726565e-01 -1.92370206e-01
1.44737959e-01 3.22998278e-02 -3.49685252e-01 1.04065903e-01
-1.28916359e+00 1.09689820e+00 8.16819012e-01 2.50838939e-02
-5.05659580e-01 -5.92516840e-01 -7.86214888e-01 3.48501541e-02
3.77240509e-01 -7.38995612e-01 1.25545907e+00 -1.16092026e+00
-1.66198874e+00 7.23964989e-01 -4.27412152e-01 -6.85866773e-01
7.08773375e-01 -2.72391945e-01 -1.31739512e-01 -3.88151966e-02
1.17117912e-01 7.15071917e-01 8.88332963e-01 -1.30110550e+00
-2.96498865e-01 9.21606049e-02 -5.22684634e-01 4.07934457e-01
-1.33420214e-01 -1.39113039e-01 -1.51730821e-01 -9.89971578e-01
-8.10250044e-02 -7.94595599e-01 -2.01810136e-01 -2.81674027e-01
-4.95113403e-01 -3.12826186e-01 4.45805818e-01 -9.84728873e-01
1.36765695e+00 -1.71167207e+00 5.07560521e-02 1.45777911e-01
-1.97210848e-01 3.78682464e-01 -1.78719372e-01 6.97609365e-01
4.88602854e-02 1.11234672e-01 -5.71027756e-01 -3.45099300e-01
-1.37476448e-03 2.33990982e-01 -4.25179482e-01 3.91892672e-01
3.50486189e-01 6.14902198e-01 -1.11520040e+00 -6.33331835e-01
4.51020189e-02 4.00848925e-01 -4.12906915e-01 2.16577649e-01
-5.40937543e-01 3.97413254e-01 -3.45276833e-01 5.56769036e-02
6.46990597e-01 -3.35157454e-01 2.61365294e-01 7.89434612e-02
-6.84825554e-02 3.39221925e-01 -8.97038341e-01 1.73735845e+00
-1.41074836e-01 5.05676627e-01 -3.79112214e-01 -1.06689882e+00
1.14783144e+00 5.13102710e-01 7.95338973e-02 -3.79081160e-01
-8.24511945e-02 2.93866038e-01 -8.56735036e-02 -3.97023857e-01
7.70393491e-01 -3.99168432e-01 1.28529087e-01 3.74033362e-01
-1.38761364e-02 -1.84020594e-01 7.94613883e-02 3.54229748e-01
6.04331791e-01 2.70819396e-01 7.35030115e-01 -2.75128365e-01
5.08102059e-01 -1.13066547e-01 7.17798591e-01 8.92784655e-01
-3.26111764e-02 7.11817265e-01 5.03495276e-01 1.00054838e-01
-1.48001969e+00 -9.05905664e-01 9.26707685e-02 6.70449376e-01
1.91279668e-02 -4.03771073e-01 -9.06820893e-01 -5.92025816e-01
-3.82377595e-01 1.43252707e+00 -2.99685180e-01 -1.18700065e-01
-4.91031259e-01 -6.88208163e-01 6.29869640e-01 1.95094168e-01
3.26648027e-01 -9.35582161e-01 -3.17187101e-01 3.79754663e-01
-6.61392272e-01 -8.71533811e-01 -7.51482546e-01 -1.21575527e-01
-7.06938386e-01 -4.98732209e-01 -1.00794411e+00 -5.86982429e-01
6.26107275e-01 -5.73844314e-02 1.01113474e+00 -9.88619775e-02
4.30649891e-02 3.38064656e-02 -5.06447971e-01 -4.18221921e-01
-9.86283004e-01 -2.03872696e-01 3.99143100e-02 -2.71235108e-02
2.07802802e-01 -4.32783216e-01 -3.90854955e-01 3.52008529e-02
-9.24011350e-01 4.12734985e-01 6.54482484e-01 1.09770620e+00
5.95181465e-01 -2.26338074e-01 7.99158931e-01 -7.42376447e-01
9.13504899e-01 -4.46135283e-01 -3.42198610e-01 5.04268706e-01
-6.76881671e-01 3.27684492e-01 7.21501529e-01 -4.37696040e-01
-1.31686580e+00 -2.58780211e-01 -2.62298226e-01 -3.35387915e-01
-1.18889585e-01 5.88933647e-01 -1.29313812e-01 4.24486190e-01
5.27351558e-01 7.33264744e-01 1.95141092e-01 -3.13926250e-01
3.49028081e-01 7.10532844e-01 5.34021020e-01 -5.59821010e-01
4.02336895e-01 7.25085363e-02 -4.77388389e-02 -9.56495881e-01
-7.62810826e-01 -3.67465019e-01 -3.72551292e-01 -1.63334683e-01
7.50393987e-01 -6.92285299e-01 -3.42041433e-01 4.81736928e-01
-1.53260636e+00 1.01562634e-01 -2.23959565e-01 5.86996853e-01
-9.13793325e-01 9.81197774e-01 -4.25478160e-01 -1.11289251e+00
-5.00016391e-01 -9.35070634e-01 9.80738401e-01 -6.80945739e-02
-5.32863021e-01 -9.83614922e-01 2.39082962e-01 2.26651102e-01
1.68502867e-01 2.88349658e-01 1.07990849e+00 -9.23131108e-01
-4.22975987e-01 -2.69105256e-01 -5.99953532e-03 7.12994218e-01
5.28183579e-02 -1.81098104e-01 -7.29204834e-01 -1.72052324e-01
1.22597545e-01 -3.06946337e-01 7.41062164e-01 3.16878378e-01
8.17559838e-01 -6.75010920e-01 -7.12453276e-02 -2.53411625e-02
1.34374070e+00 3.55224982e-02 8.25324357e-01 1.39061391e-01
5.44352531e-01 7.37893641e-01 8.04059148e-01 6.91594839e-01
6.28955588e-02 7.34289408e-01 2.00792521e-01 2.35702366e-01
1.16779208e-01 -6.09240532e-01 5.14387906e-01 9.90705669e-01
1.23511560e-01 -7.43700743e-01 -6.97681844e-01 4.97101307e-01
-2.07458091e+00 -1.25417924e+00 -1.70014754e-01 2.51443601e+00
1.12286365e+00 5.03491871e-02 -6.36305436e-02 -6.97149010e-03
1.03442907e+00 1.75462559e-01 -3.38420808e-01 -5.17734110e-01
-4.14134532e-01 -1.25805020e-01 4.24237609e-01 6.51140630e-01
-7.83964574e-01 8.34897697e-01 6.16172791e+00 1.56212223e+00
-9.03857529e-01 -8.33196267e-02 6.89199746e-01 1.73302501e-01
-5.23155928e-01 7.59784207e-02 -5.75397313e-01 7.86999345e-01
8.92965972e-01 -5.24548590e-01 1.62372485e-01 6.23857379e-01
6.45117879e-01 -2.71348059e-01 -1.08054364e+00 8.50481570e-01
1.43052906e-01 -1.42510867e+00 4.06660944e-01 5.03040180e-02
8.25610936e-01 -3.29247802e-01 -2.57281028e-02 3.57949995e-02
1.51762635e-01 -9.01391029e-01 1.05420887e+00 6.15992963e-01
6.16118550e-01 -7.23969221e-01 8.51329744e-01 9.41349030e-01
-6.18192852e-01 3.91105741e-01 -3.04184556e-01 -2.53202375e-02
4.65993196e-01 7.58166492e-01 -1.32701838e+00 7.16326237e-01
-1.35164842e-01 3.44316125e-01 -5.32623678e-02 8.30942214e-01
-1.34158239e-01 6.96019948e-01 -1.10951453e-01 -1.34625584e-01
3.87869388e-01 -4.65060949e-01 8.38611782e-01 1.46530664e+00
6.77055061e-01 -6.66459426e-02 2.08855361e-01 7.77299225e-01
-1.87009558e-01 3.32585216e-01 -4.92437780e-01 -2.02518493e-01
3.84556174e-01 6.38981104e-01 -4.71101701e-01 -5.68770230e-01
-2.91220337e-01 1.34267461e+00 2.27341488e-01 2.64634848e-01
-7.11540639e-01 -1.51463062e-01 4.01744366e-01 3.41923349e-02
2.46304959e-01 -5.28655760e-02 -3.95040482e-01 -9.44397807e-01
1.80757254e-01 -7.55452931e-01 9.51210677e-05 -8.25753570e-01
-1.36900353e+00 4.13126230e-01 2.75774542e-02 -1.27969658e+00
-6.18574321e-01 -2.72695631e-01 -6.24071658e-01 1.26642632e+00
-1.38526952e+00 -7.25548506e-01 3.92196983e-01 9.32606831e-02
7.86561787e-01 -1.48083225e-01 7.40871251e-01 -1.05757423e-01
-3.04369807e-01 6.20018721e-01 3.71844083e-01 -1.98585510e-01
7.62697339e-01 -1.14837527e+00 3.43133360e-01 8.56419861e-01
-2.18184412e-01 6.68955922e-01 1.17905247e+00 -9.56963420e-01
-8.27189982e-01 -1.04112291e+00 1.45362985e+00 -3.06552108e-02
6.15627408e-01 -7.54422918e-02 -1.04663539e+00 3.89074147e-01
3.75898361e-01 -6.89818621e-01 6.67010963e-01 -4.35930669e-01
-1.72932252e-01 5.47318876e-01 -1.10048831e+00 7.01279938e-01
7.31552541e-01 -3.23780149e-01 -7.01401234e-01 4.94475961e-01
7.32748568e-01 -6.53595254e-02 -5.80558896e-01 2.76786953e-01
3.36034745e-01 -8.95958304e-01 6.02970839e-01 -4.60621804e-01
5.94580412e-01 -2.04822034e-01 -1.64788306e-01 -1.29104447e+00
-1.53749377e-01 -9.55061913e-01 -5.82932793e-02 1.20455754e+00
3.57168615e-01 -4.01542038e-01 8.12526762e-01 3.47605139e-01
-3.11520159e-01 -7.32874036e-01 -8.95632267e-01 -9.34088409e-01
1.02903344e-01 -2.47080043e-01 3.06401432e-01 9.19641256e-01
1.68608099e-01 4.10732925e-01 -8.73522639e-01 -6.79235831e-02
6.56783044e-01 -2.78510630e-01 6.03140473e-01 -9.23773110e-01
-4.23154593e-01 -3.94843012e-01 -2.75432199e-01 -1.46655440e+00
2.06788331e-01 -1.02972472e+00 3.92721087e-01 -1.47113717e+00
3.97650689e-01 -3.63590084e-02 8.03596675e-02 -4.47716117e-02
-4.51556057e-01 -1.44663498e-01 1.27202004e-01 3.77379060e-01
-4.06415820e-01 9.34116602e-01 1.27165294e+00 1.43354684e-01
2.17840299e-02 -1.88771989e-02 -5.09188712e-01 7.24672675e-01
7.72593796e-01 -4.37588274e-01 -4.39830542e-01 -1.63267121e-01
1.18207701e-01 3.33089858e-01 1.45201325e-01 -5.54873824e-01
7.47719556e-02 -2.93586612e-01 3.02732959e-02 -6.05691075e-01
3.31137151e-01 -3.60988677e-01 2.34923899e-01 3.54305625e-01
-6.94540203e-01 -1.87929824e-01 -1.53193638e-01 7.88972020e-01
-2.27703914e-01 -6.93851829e-01 6.88224733e-01 -1.90057248e-01
-5.39810598e-01 -6.23227209e-02 -5.13766885e-01 1.06532060e-01
8.18839312e-01 -4.20751572e-01 -1.08368725e-01 -6.64350450e-01
-4.02874649e-01 7.80544654e-02 4.95698094e-01 2.87614048e-01
6.63613260e-01 -1.22259510e+00 -1.13534248e+00 -9.97803956e-02
1.39864162e-01 -2.19657958e-01 1.90153986e-01 6.92824423e-01
-4.33192194e-01 5.46630681e-01 4.97851670e-02 -5.16710281e-01
-1.17732871e+00 4.20519799e-01 -9.49193761e-02 -5.18847764e-01
-4.87202466e-01 8.42203915e-01 3.26427490e-01 -2.11719111e-01
1.40318945e-01 6.50757924e-02 7.97755495e-02 -1.98382363e-01
4.23783481e-01 3.97301257e-01 -2.72564683e-02 -7.79245079e-01
-7.11546987e-02 9.27594304e-03 -3.22273165e-01 -3.81793380e-01
9.27133739e-01 -2.67189831e-01 -7.21465647e-02 5.56553125e-01
1.09661520e+00 -4.19909991e-02 -1.30930293e+00 -1.69925645e-01
2.91638136e-01 -4.58081812e-01 -2.03422070e-01 -8.09140325e-01
-8.58664438e-02 7.59978890e-01 1.02692127e-01 -2.52067167e-02
8.13006699e-01 -2.58828521e-01 7.91894019e-01 3.07211757e-01
3.95080805e-01 -1.26633096e+00 5.60567230e-02 6.00620151e-01
9.45625186e-01 -9.96826410e-01 -7.19218254e-02 -4.36579555e-01
-9.98930335e-01 9.07780647e-01 2.78927892e-01 6.92918524e-02
7.67072737e-02 -2.92001646e-02 -1.79341838e-01 4.42065060e-01
-8.53997469e-01 2.42262185e-01 3.22413832e-01 6.08319938e-01
6.26537979e-01 -1.71140730e-02 -8.29059839e-01 4.88527358e-01
-2.03249618e-01 1.59370840e-01 5.72492182e-01 7.93128252e-01
-6.42385840e-01 -1.16897404e+00 -4.43810195e-01 3.55741680e-01
-4.85937029e-01 -4.15851116e-01 -3.36940557e-01 3.65368605e-01
-1.38964655e-03 9.20785725e-01 6.68304265e-02 -4.58765738e-02
-8.58874246e-02 3.16001207e-01 4.94788736e-01 -4.94060785e-01
-2.94544399e-01 2.53923893e-01 1.81078956e-01 -4.67347838e-02
-2.68601298e-01 -5.43449461e-01 -1.09076786e+00 -2.36115217e-01
-5.41699529e-01 4.09885734e-01 6.64808631e-01 9.31008160e-01
3.77907664e-01 1.60821095e-01 5.31414628e-01 -7.52083480e-01
-1.01515484e+00 -1.09623110e+00 -4.18698519e-01 5.66708148e-01
8.20360631e-02 -3.02833706e-01 -2.70879507e-01 3.54463428e-01] | [11.881521224975586, 9.23363208770752] |
8c61eda7-e05c-4b7e-9581-170dca587bfe | why-deep-surgical-models-fail-revisiting | 2209.08647 | null | https://arxiv.org/abs/2209.08647v2 | https://arxiv.org/pdf/2209.08647v2.pdf | Why Deep Surgical Models Fail?: Revisiting Surgical Action Triplet Recognition through the Lens of Robustness | Surgical action triplet recognition provides a better understanding of the surgical scene. This task is of high relevance as it provides the surgeon with context-aware support and safety. The current go-to strategy for improving performance is the development of new network mechanisms. However, the performance of current state-of-the-art techniques is substantially lower than other surgical tasks. Why is this happening? This is the question that we address in this work. We present the first study to understand the failure of existing deep learning models through the lens of robustness and explainability. Firstly, we study current existing models under weak and strong $\delta-$perturbations via an adversarial optimisation scheme. We then analyse the failure modes via feature based explanations. Our study reveals that the key to improving performance and increasing reliability is in the core and spurious attributes. Our work opens the door to more trustworthy and reliable deep learning models in surgical data science. | ['Angelica I. Aviles-Rivero', 'Carola-Bibiane Schönlieb', 'Yueming Jin', 'Shujun Wang', 'Lihao Liu', 'Yanqi Cheng'] | 2022-09-18 | null | null | null | null | ['action-triplet-recognition'] | ['computer-vision'] | [ 4.64291900e-01 4.49532628e-01 -3.54353368e-01 -2.51801878e-01
-7.82896161e-01 -4.95491564e-01 3.22946370e-01 2.68328935e-01
-4.15262908e-01 6.56102002e-01 5.33863127e-01 -5.86820722e-01
-6.95409536e-01 -3.56920063e-01 -7.72297204e-01 -8.85856211e-01
-2.41125748e-01 8.59109536e-02 -4.79164980e-02 -4.78764981e-01
3.69804412e-01 6.00462079e-01 -1.28314674e+00 4.76436257e-01
4.81041133e-01 9.10434425e-01 -1.99329868e-01 6.72948718e-01
4.37951952e-01 8.78658831e-01 -2.73094475e-01 -4.82908398e-01
5.16514838e-01 -4.27851290e-01 -9.34485018e-01 -2.45277360e-01
2.01306477e-01 -2.04977512e-01 -3.54335010e-01 8.22188616e-01
7.07023203e-01 -2.54188001e-01 1.80770382e-01 -1.13022125e+00
-3.75512153e-01 4.44063246e-01 -1.15869828e-01 4.39843088e-01
4.44704182e-02 4.16396201e-01 8.08908582e-01 -2.98777968e-01
6.05441868e-01 6.18697643e-01 7.67768681e-01 8.43789697e-01
-1.08815229e+00 -6.15040958e-01 7.05780787e-03 1.26232490e-01
-9.74851370e-01 -6.30477965e-01 6.28852844e-01 -3.36660415e-01
7.79583693e-01 4.30312455e-01 6.47261679e-01 1.34444940e+00
7.97903955e-01 5.01480877e-01 1.07589543e+00 -3.28216702e-01
3.03150192e-02 -1.61633734e-02 -5.43438382e-02 9.37745452e-01
4.50910479e-01 7.68772602e-01 -7.36211896e-01 9.47340205e-03
8.16068113e-01 1.59803569e-01 -3.10905010e-01 -4.97896016e-01
-1.45001578e+00 8.19481730e-01 6.93799257e-01 4.10327554e-01
-3.32092941e-01 6.27974808e-01 4.64857459e-01 3.29571009e-01
5.27590103e-02 1.09306502e+00 -5.15138447e-01 -3.37171644e-01
-7.54106581e-01 -9.20040086e-02 7.10678458e-01 4.52871501e-01
2.10070372e-01 -4.43783104e-02 1.04954854e-01 2.80677617e-01
1.79568380e-01 -1.58159971e-01 4.57364142e-01 -1.19573748e+00
1.77104454e-02 5.40401995e-01 -2.56606400e-01 -8.94173741e-01
-8.11131477e-01 -9.60945845e-01 -7.81787395e-01 5.60776889e-01
4.34280634e-01 -7.52583966e-02 -9.00676429e-01 1.76490772e+00
5.87766394e-02 2.63899207e-01 8.15634504e-02 7.28551388e-01
7.30754912e-01 -3.01900476e-01 6.34140074e-02 -2.16455874e-03
1.19166136e+00 -6.83024883e-01 -5.93142509e-01 -3.72106373e-01
9.43315268e-01 -6.68753088e-01 7.44253516e-01 3.86102736e-01
-9.04056370e-01 -1.26359135e-01 -1.05953681e+00 8.82783607e-02
-1.41920626e-01 -1.15961649e-01 1.03106987e+00 7.75840342e-01
-9.97718871e-01 7.41616130e-01 -1.14717650e+00 -4.93501931e-01
6.93098009e-01 9.20838416e-01 -7.63486803e-01 -1.19792387e-01
-1.07357883e+00 1.20756578e+00 9.82709080e-02 2.45277449e-01
-1.12941468e+00 -7.24450409e-01 -9.19691265e-01 -2.20888734e-01
4.76269960e-01 -1.05953407e+00 9.97106135e-01 -9.62235808e-01
-1.11957431e+00 1.00233090e+00 1.19535320e-01 -6.63881123e-01
4.89548475e-01 -1.29419550e-01 -1.46189332e-01 -1.85979996e-02
-3.10232848e-01 5.02776742e-01 5.43437898e-01 -1.45038640e+00
-4.09818679e-01 -4.83053476e-01 1.43166885e-01 1.35409571e-02
-4.40480337e-02 -2.21398652e-01 -4.55931714e-03 -6.46150351e-01
9.48461071e-02 -1.34765720e+00 -6.19745374e-01 2.34205231e-01
-4.66454118e-01 4.07319844e-01 3.94362099e-02 -4.33635175e-01
1.02045393e+00 -2.22715998e+00 2.69863009e-01 1.88371763e-01
5.92236042e-01 1.74544185e-01 6.94718538e-03 6.02163851e-01
-3.93182278e-01 3.10490668e-01 -1.95619781e-02 -2.80391604e-01
-3.26871395e-01 4.95573997e-01 -1.21075414e-01 6.47764802e-01
2.08585888e-01 1.00123966e+00 -8.44362259e-01 -3.22913617e-01
2.21860766e-01 3.35175395e-01 -9.01000082e-01 2.84457523e-02
2.49212742e-01 7.80507267e-01 -3.40360403e-01 7.39395618e-01
-5.00431880e-02 -8.20938945e-02 9.33851600e-02 -1.87401071e-01
4.63298671e-02 9.42464620e-02 -7.20605671e-01 1.99366808e+00
-1.92654371e-01 5.61095238e-01 -4.29822952e-02 -1.06408429e+00
6.07548356e-01 2.17535853e-01 7.84603179e-01 -4.61163640e-01
4.60247129e-01 4.30075437e-01 6.75478697e-01 -7.65215814e-01
1.87016636e-01 -5.89826941e-01 -7.02819750e-02 1.03054695e-01
-1.05037093e-01 8.81296173e-02 -3.54129761e-01 1.09151520e-01
1.45153904e+00 4.83913720e-02 4.42299575e-01 -4.12780881e-01
5.29855490e-02 1.67805955e-01 4.70732212e-01 8.38720024e-01
-4.53916490e-01 8.07130277e-01 7.51998961e-01 -7.01656461e-01
-7.34175503e-01 -8.57370675e-01 -1.47967637e-01 7.24127352e-01
6.75464645e-02 -2.33092338e-01 -5.43893754e-01 -8.50464344e-01
4.59036119e-02 5.61230183e-01 -1.34374619e+00 -8.69853854e-01
-4.72258121e-01 -8.28877330e-01 5.93661785e-01 6.82857871e-01
-1.61152259e-01 -1.00401735e+00 -7.22025573e-01 1.15092516e-01
-1.35905847e-01 -1.01987600e+00 1.26043183e-03 4.45825815e-01
-1.12555051e+00 -1.26555598e+00 -3.50588322e-01 -5.21586835e-01
7.26478636e-01 1.46232136e-02 9.13059592e-01 5.51579177e-01
-5.63413620e-01 3.04551899e-01 -4.23907906e-01 -6.49369717e-01
-6.58590436e-01 1.14067741e-01 2.50414521e-01 -3.13889325e-01
1.63236141e-01 -5.11234224e-01 -1.03661513e+00 2.32238293e-01
-1.06713736e+00 -8.09258968e-02 1.00297248e+00 1.08445299e+00
4.42235708e-01 -1.63481921e-01 4.25877959e-01 -1.00942135e+00
5.29403865e-01 -4.78074193e-01 2.80222539e-02 -1.75758209e-02
-8.80890310e-01 2.78067648e-01 3.98139119e-01 -1.74610153e-01
-5.02150118e-01 2.33767152e-01 -2.43088737e-01 -3.31165880e-01
-2.39465982e-01 5.32817543e-01 2.45983779e-01 -4.38977063e-01
9.20927167e-01 3.54762599e-02 5.22850394e-01 -2.49152422e-01
1.67159483e-01 8.82727578e-02 5.18541336e-01 -3.58115137e-01
6.48782492e-01 7.10829794e-01 3.06988925e-01 -3.88381064e-01
-8.88386250e-01 -2.20841810e-01 -5.10618746e-01 -2.32354030e-01
7.18502164e-01 -6.21001542e-01 -9.30317640e-01 1.84775367e-01
-7.99544454e-01 -2.79276013e-01 -4.95025128e-01 5.20144224e-01
-7.92557895e-01 2.28602618e-01 -4.51496750e-01 -3.91847193e-01
-1.82046905e-01 -1.70383549e+00 1.05187738e+00 -7.33933896e-02
-4.30279642e-01 -9.06371891e-01 1.90634638e-01 6.21475577e-01
4.82968241e-01 5.73372424e-01 8.51004004e-01 -8.31346810e-01
-6.56191528e-01 -4.61101413e-01 8.05985853e-02 3.20635796e-01
1.37965679e-01 -1.52442381e-01 -8.75757217e-01 -3.41574341e-01
1.14498459e-01 -1.04859836e-01 8.57126772e-01 4.79596943e-01
1.36555517e+00 -2.41355523e-01 -4.48755413e-01 9.91001248e-01
1.51273680e+00 -7.03224614e-02 8.49861085e-01 6.61341727e-01
4.49761510e-01 7.07778931e-01 4.36444432e-01 2.08243087e-01
3.26231509e-01 5.24038792e-01 1.03033543e+00 -3.66650492e-01
-1.91692203e-01 -1.25918612e-01 9.38698463e-03 4.24686730e-01
-1.99164599e-01 -3.54295387e-03 -1.21070349e+00 6.05100989e-01
-1.87219465e+00 -8.17906141e-01 -6.40604319e-03 2.21833134e+00
7.34258890e-01 3.87939155e-01 -2.10751325e-01 2.95921952e-01
7.79872537e-02 -2.12687239e-01 -4.64687556e-01 -4.18992192e-01
3.76892269e-01 1.95000902e-01 6.35320783e-01 3.23064864e-01
-7.50848234e-01 6.04362607e-01 6.65813732e+00 4.52765405e-01
-1.17365730e+00 -1.78100795e-01 6.59702837e-01 -3.59453499e-01
-6.51194379e-02 -1.24663590e-02 -4.19716746e-01 2.32520893e-01
9.83606875e-01 4.75957580e-02 2.58300424e-01 6.59208834e-01
2.53248960e-01 -1.50609583e-01 -1.37097979e+00 7.34588206e-01
3.45169634e-01 -1.61435425e+00 -2.81154662e-01 3.62202883e-01
3.91132802e-01 1.77873835e-01 2.61272132e-01 -1.40672565e-01
2.75538653e-01 -1.41111362e+00 3.73570234e-01 7.67185926e-01
6.68306172e-01 -5.70504487e-01 1.10250247e+00 1.88045874e-01
-4.46870863e-01 -2.71033376e-01 1.86730623e-01 1.61619335e-01
-1.29390866e-01 1.24800593e-01 -9.03291345e-01 5.30574322e-01
6.48906350e-01 6.20523095e-01 -6.06300414e-01 1.11072516e+00
-5.76276369e-02 4.60578412e-01 -1.59287244e-01 2.09416971e-01
9.62790847e-02 4.84943748e-01 4.58878726e-01 9.69173551e-01
1.45892188e-01 7.77355060e-02 -1.31654352e-01 2.91296542e-01
7.35768080e-02 -2.96522886e-01 -7.36779571e-01 9.45801660e-02
1.49596697e-02 1.09847498e+00 -6.61142111e-01 1.88084736e-01
-1.49126321e-01 7.92239785e-01 2.90523976e-01 -1.32581100e-01
-5.73123395e-01 -4.62421551e-02 8.50989461e-01 2.11441308e-01
-1.21663526e-01 -2.26573333e-01 -8.04028153e-01 -1.02980828e+00
-1.05284743e-01 -1.31854689e+00 5.52104175e-01 -4.67698246e-01
-9.92402613e-01 5.64820588e-01 -3.19627374e-01 -1.33142304e+00
-2.61494249e-01 -8.20656419e-01 -3.08153272e-01 4.66191292e-01
-1.71998644e+00 -1.18635738e+00 -3.63516867e-01 5.99096417e-01
3.22002649e-01 -8.23103711e-02 1.12289417e+00 1.97609770e-03
-5.01525879e-01 7.73236692e-01 -3.08826584e-02 3.56397927e-02
8.14989746e-01 -1.11708581e+00 9.59733501e-02 7.14238405e-01
8.12978894e-02 8.79148901e-01 1.13795114e+00 -3.07274878e-01
-1.60612869e+00 -5.35848200e-01 5.32572746e-01 -1.14435434e+00
7.30379224e-01 -6.22635335e-02 -6.66106403e-01 6.30230248e-01
2.78491843e-02 9.61249769e-02 1.22278368e+00 2.32137069e-01
-8.19161758e-02 -1.07519068e-01 -1.05664051e+00 5.53192735e-01
1.07001603e+00 -2.77343988e-01 -6.07183218e-01 1.82008520e-01
6.19211972e-01 -5.24714768e-01 -1.06649101e+00 7.74763882e-01
7.10985243e-01 -1.15839708e+00 1.08931935e+00 -9.91428018e-01
8.34361315e-01 2.06432324e-02 -1.92644242e-02 -1.37238753e+00
-1.52790546e-01 -7.07865000e-01 1.97127804e-01 5.16978621e-01
6.35273457e-01 -5.42876422e-01 9.79421973e-01 8.95298123e-01
-3.62296075e-01 -1.26285064e+00 -1.08573437e+00 -5.49620450e-01
2.43292004e-01 -6.61695540e-01 3.86368483e-01 1.02930045e+00
1.77171215e-01 -2.46321112e-01 -3.95588577e-01 8.86122808e-02
5.74549496e-01 -1.95005417e-01 6.03493989e-01 -9.58176792e-01
-2.23913535e-01 -5.33119380e-01 -8.65923643e-01 -1.97076693e-01
-1.94390967e-01 -9.26486611e-01 4.65148911e-02 -1.55548429e+00
1.81051642e-01 -4.02567387e-01 -6.04802191e-01 6.94690764e-01
-2.10441306e-01 3.19165021e-01 2.31703356e-01 2.94095397e-01
-3.42420965e-01 2.80873805e-01 1.14148843e+00 3.05209786e-01
1.25350997e-01 1.00950383e-01 -1.20860708e+00 5.52692473e-01
9.92454112e-01 -7.77808428e-01 -1.25613183e-01 -4.55693275e-01
4.11471963e-01 -1.11500598e-01 6.47705436e-01 -8.41980159e-01
4.03653383e-01 -1.00670904e-01 2.40707934e-01 1.74207211e-01
2.39634827e-01 -1.10591626e+00 3.16289254e-02 1.01984656e+00
-5.96469998e-01 8.39594156e-02 4.46801394e-01 6.60964489e-01
9.37383473e-02 5.95221147e-02 8.00114453e-01 -2.82622784e-01
-5.53243160e-01 1.83200017e-01 -2.54928261e-01 -9.84038264e-02
1.11500347e+00 -2.99598008e-01 -4.34627503e-01 -2.56397814e-01
-1.00063229e+00 5.33691011e-02 5.75331867e-01 6.80112004e-01
7.31311738e-01 -1.08674896e+00 -7.72556186e-01 3.00560892e-01
3.35758150e-01 -2.69554853e-01 2.39942163e-01 1.25711596e+00
-5.56809783e-01 3.56857955e-01 -2.65765578e-01 -4.36820805e-01
-1.28607666e+00 5.84355414e-01 7.24323094e-01 -1.77682832e-01
-5.42868614e-01 8.96773636e-01 5.50598726e-02 -2.57743269e-01
3.88596237e-01 -1.47100791e-01 -2.00596657e-02 -2.98071325e-01
3.05697680e-01 1.20745425e-03 1.48519546e-01 -3.57549608e-01
-5.41017592e-01 2.86132455e-01 -1.56667128e-01 1.63872093e-01
1.56906295e+00 1.16829291e-01 2.94620451e-02 1.21986635e-01
9.69246328e-01 -2.22736806e-01 -1.02193594e+00 4.39247228e-02
-2.27654371e-02 -5.17487288e-01 2.38441885e-01 -1.19446456e+00
-1.14998031e+00 8.55848849e-01 7.51625478e-01 2.81691663e-02
1.08011734e+00 1.18307896e-01 5.69090724e-01 1.03960671e-01
3.51207852e-01 -7.26984322e-01 6.74029142e-02 3.01961917e-02
9.82499301e-01 -1.53002858e+00 8.69055539e-02 -2.25844309e-01
-7.50682712e-01 1.03711414e+00 5.11839747e-01 -6.26922995e-02
6.74568415e-01 4.76631790e-01 3.69120687e-01 -6.02210581e-01
-7.69529283e-01 -2.18992278e-01 4.11251694e-01 4.13367242e-01
4.37924623e-01 -6.19181357e-02 -2.96766996e-01 5.76095641e-01
-3.07912856e-01 2.34957680e-01 5.29966593e-01 1.05791891e+00
-1.69933885e-01 -1.25936270e+00 -1.10637367e-01 6.11647010e-01
-9.32734191e-01 -2.94368356e-01 -3.68199050e-01 8.47807348e-01
2.19062865e-02 8.78014326e-01 -5.13776958e-01 -6.99894786e-01
4.66228157e-01 -6.35032058e-02 3.09738487e-01 -6.35491490e-01
-8.48860681e-01 -2.14566186e-01 1.89437851e-01 -1.00500357e+00
-3.12114656e-01 -6.69298232e-01 -1.06922054e+00 -1.97531343e-01
-3.00483227e-01 -1.99094862e-01 9.35574830e-01 9.55299616e-01
4.98519868e-01 9.30811584e-01 4.82841313e-01 -4.92057472e-01
-5.39249778e-01 -4.11113530e-01 -1.47113889e-01 3.47902447e-01
7.03291953e-01 -5.97269595e-01 -5.82257986e-01 -1.86592519e-01] | [14.145244598388672, -3.2947945594787598] |
6113852f-b008-44c1-a6de-39a00bb749b9 | diversified-patch-based-style-transfer-with | 2101.06381 | null | https://arxiv.org/abs/2101.06381v2 | https://arxiv.org/pdf/2101.06381v2.pdf | DivSwapper: Towards Diversified Patch-based Arbitrary Style Transfer | Gram-based and patch-based approaches are two important research lines of style transfer. Recent diversified Gram-based methods have been able to produce multiple and diverse stylized outputs for the same content and style images. However, as another widespread research interest, the diversity of patch-based methods remains challenging due to the stereotyped style swapping process based on nearest patch matching. To resolve this dilemma, in this paper, we dive into the crux of existing patch-based methods and propose a universal and efficient module, termed DivSwapper, for diversified patch-based arbitrary style transfer. The key insight is to use an essential intuition that neural patches with higher activation values could contribute more to diversity. Our DivSwapper is plug-and-play and can be easily integrated into existing patch-based and Gram-based methods to generate diverse results for arbitrary styles. We conduct theoretical analyses and extensive experiments to demonstrate the effectiveness of our method, and compared with state-of-the-art algorithms, it shows superiority in diversity, quality, and efficiency. | ['Dongming Lu', 'Wei Xing', 'Ailin Li', 'Zhiwen Zuo', 'Haibo Chen', 'Lei Zhao', 'Zhizhong Wang'] | 2021-01-16 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [ 1.10729359e-01 -2.65368581e-01 -3.24709378e-02 -1.87895983e-01
-5.40828466e-01 -7.48845875e-01 3.55385065e-01 -4.51473325e-01
1.67048335e-01 9.50314581e-01 1.70386180e-01 -1.13012329e-01
7.11911470e-02 -8.72320712e-01 -6.68188274e-01 -6.40306294e-01
4.34232652e-01 2.38579839e-01 2.28692934e-01 -5.93805075e-01
3.96314263e-01 4.22634214e-01 -1.39082015e+00 4.69643891e-01
1.19830108e+00 7.86366105e-01 2.62639344e-01 3.99875909e-01
-2.82146931e-01 2.65022159e-01 -6.32751584e-01 -7.49839067e-01
2.84281373e-01 -1.04960942e+00 -5.34511089e-01 3.02561261e-02
4.53394890e-01 -1.97632775e-01 -1.25207916e-01 1.04333448e+00
8.47841024e-01 -8.11392535e-03 8.06422234e-01 -1.20309782e+00
-1.33815587e+00 4.96368349e-01 -7.42929935e-01 -1.41078055e-01
3.04532140e-01 1.38753846e-01 1.02981973e+00 -9.90807831e-01
7.57690549e-01 1.57958186e+00 5.01177490e-01 8.11682582e-01
-1.34131336e+00 -8.06068361e-01 3.09551954e-01 -8.62616077e-02
-1.13333583e+00 -8.98309723e-02 9.78814125e-01 -3.43827784e-01
2.90720105e-01 4.50654030e-01 9.57957447e-01 1.38605893e+00
1.95476964e-01 1.24607968e+00 1.36872351e+00 -3.59679997e-01
2.35858753e-01 2.64474839e-01 -3.87343526e-01 5.24820626e-01
2.44553626e-01 9.79621634e-02 -7.25987613e-01 5.84814288e-02
1.34022510e+00 -9.36067849e-02 -2.73646355e-01 -3.63071501e-01
-1.18096673e+00 8.10006976e-01 5.58863401e-01 1.84987262e-01
-1.03172818e-02 5.93638122e-02 3.00655693e-01 4.68335539e-01
5.07725000e-01 7.59704888e-01 -6.49115145e-02 -2.57374138e-01
-9.35704529e-01 5.45706689e-01 6.61376655e-01 1.11609185e+00
5.98652303e-01 3.68197054e-01 -6.00351274e-01 1.07241285e+00
-6.56361207e-02 4.43179011e-01 5.24198174e-01 -8.92720699e-01
1.72251269e-01 5.87207913e-01 -3.63484323e-02 -9.79952276e-01
9.51560885e-02 -6.20433748e-01 -9.74325001e-01 4.25109088e-01
1.46136180e-01 -2.09569097e-01 -8.64226043e-01 1.60557306e+00
7.69077241e-02 -4.89394106e-02 -4.19678509e-01 8.53262067e-01
7.34506309e-01 6.78035915e-01 -1.17948785e-01 1.86601475e-01
1.04933500e+00 -1.12547791e+00 -5.12109458e-01 4.56415862e-02
2.56732494e-01 -8.82754147e-01 1.61084354e+00 3.83673251e-01
-1.06245482e+00 -5.66412628e-01 -9.42655861e-01 1.34205848e-01
-1.60336852e-01 2.67812572e-02 6.16207957e-01 7.76175976e-01
-1.10934067e+00 7.80704498e-01 -1.87608898e-01 -4.19277519e-01
5.24879634e-01 -2.49595474e-02 -8.14050958e-02 1.58670768e-01
-1.01003194e+00 6.30520165e-01 1.23142548e-01 -1.17604181e-01
-5.40588319e-01 -7.73995697e-01 -5.34430623e-01 -6.65800497e-02
2.51960512e-02 -1.08551967e+00 1.16339338e+00 -1.35985291e+00
-1.98078907e+00 7.66005754e-01 -6.14354797e-02 1.50647983e-01
5.61857283e-01 -1.84284434e-01 -9.42323133e-02 -1.70901254e-01
1.02718756e-01 9.20448244e-01 9.58803356e-01 -1.55498981e+00
-5.08101106e-01 1.07101714e-02 7.00670928e-02 2.91921973e-01
-6.20329082e-01 -9.04700533e-02 -4.35597986e-01 -1.37317073e+00
-1.20433509e-01 -9.34661746e-01 -1.36436105e-01 3.46421778e-01
-3.49728644e-01 9.43124816e-02 6.77416146e-01 -4.89300996e-01
1.35646868e+00 -2.24810815e+00 5.47461033e-01 2.25070654e-03
1.11973308e-01 2.71865845e-01 -3.92637372e-01 4.54185724e-01
2.39175126e-01 2.55457908e-01 -2.83004493e-01 -2.09603071e-01
7.23593161e-02 7.98090920e-02 -4.58336622e-01 -1.21150464e-01
3.95074815e-01 1.14801598e+00 -8.68686259e-01 -3.13871533e-01
3.01076528e-02 3.31460923e-01 -8.48225653e-01 2.85877734e-01
-1.92416921e-01 7.23019600e-01 -4.77100074e-01 8.38052988e-01
5.75364053e-01 6.59733191e-02 -1.21968471e-01 -2.09481835e-01
3.97429764e-02 -2.09789108e-02 -1.02676237e+00 1.79389131e+00
-4.17688876e-01 7.10371673e-01 -1.86639279e-01 -5.89319944e-01
1.20855570e+00 4.84531671e-02 1.94326341e-02 -7.10542858e-01
2.32550368e-01 5.22018552e-01 -2.05031186e-01 -1.54116407e-01
7.91171193e-01 -2.51671851e-01 -1.39467791e-01 4.28882927e-01
-2.92376373e-02 -4.93086278e-01 1.37435943e-01 7.93780982e-02
7.41349280e-01 5.85577309e-01 1.29583567e-01 -4.80548203e-01
2.81487048e-01 -2.19634056e-01 5.00746608e-01 6.54779792e-01
-3.33757214e-02 1.05632007e+00 5.91561854e-01 -2.23149344e-01
-1.19218457e+00 -1.06606913e+00 6.51373342e-02 1.18648887e+00
3.01010102e-01 -2.43662611e-01 -8.96894336e-01 -5.44022322e-01
1.17698926e-02 6.24799192e-01 -7.63389945e-01 -9.41730440e-02
-5.00588000e-01 -5.90485871e-01 5.50547302e-01 6.35182977e-01
6.13245964e-01 -1.26025498e+00 -2.78684855e-01 2.28952244e-01
-9.78084467e-03 -4.56391186e-01 -6.67911232e-01 -1.72642469e-01
-9.32587504e-01 -3.96919817e-01 -1.38442862e+00 -8.60348523e-01
6.82834387e-01 3.66864741e-01 1.23159313e+00 -2.39881836e-02
-3.41039859e-02 7.35669807e-02 -5.89203358e-01 -4.82136607e-01
-3.91956687e-01 2.47132599e-01 -2.13078499e-01 1.47952437e-01
-8.16726610e-02 -8.89265001e-01 -8.24173748e-01 5.61133504e-01
-1.17375934e+00 2.13355005e-01 6.86121762e-01 9.26538408e-01
4.83051419e-01 -4.14741307e-01 6.47587538e-01 -1.00880277e+00
1.13922560e+00 -2.50472873e-01 -1.93635166e-01 2.40368143e-01
-3.58162940e-01 7.14236870e-02 7.05965221e-01 -6.61346793e-01
-1.01444173e+00 -2.14206710e-01 -1.46894798e-01 -4.36454564e-01
1.31407723e-01 2.62951821e-01 -4.21876401e-01 -3.36368859e-01
7.07283676e-01 3.06024134e-01 -9.91015509e-02 -5.39363086e-01
6.70070648e-01 6.13724828e-01 3.43102664e-01 -9.33758855e-01
8.15314174e-01 2.93225765e-01 -2.97423273e-01 -5.17171562e-01
-6.45936191e-01 1.53404266e-01 -4.47773308e-01 -3.25421035e-01
4.82015520e-01 -8.08382750e-01 -2.40328595e-01 7.89867461e-01
-1.01896918e+00 -5.47771096e-01 -4.06306714e-01 -1.70481324e-01
-5.82553208e-01 2.21991748e-01 -5.74817240e-01 -4.90852326e-01
-3.37292939e-01 -1.22923255e+00 1.20607865e+00 4.41254497e-01
-5.07525027e-01 -9.09731209e-01 7.49621093e-02 -1.11950114e-01
6.71353400e-01 2.38350287e-01 9.81292367e-01 -8.39548409e-02
-5.16788304e-01 1.48418534e-03 -3.47209960e-01 2.35907212e-01
3.15006793e-01 1.12446234e-01 -7.63196409e-01 -3.41033548e-01
-3.48081678e-01 -2.40612254e-01 7.91501164e-01 2.06208020e-01
1.19327414e+00 -2.55335510e-01 -1.38010293e-01 8.72263312e-01
1.31177378e+00 2.41730243e-01 8.50064874e-01 4.53019232e-01
8.60978246e-01 4.07223791e-01 4.47951555e-01 3.83743048e-01
8.43774229e-02 7.25452900e-01 1.82034820e-01 -2.56631285e-01
-3.76833379e-01 -4.02337551e-01 3.85061741e-01 8.88377786e-01
-1.61344185e-01 -4.33868080e-01 -3.37265015e-01 5.50547004e-01
-1.76180100e+00 -9.27312076e-01 1.84360340e-01 1.95578635e+00
1.00237834e+00 7.98001513e-03 3.33619088e-01 8.04679617e-02
7.87090540e-01 1.85327172e-01 -6.62814021e-01 -6.54332042e-01
-3.92285198e-01 2.83004463e-01 8.58235583e-02 1.47821411e-01
-6.69208884e-01 1.22012949e+00 6.98931837e+00 1.22594535e+00
-1.35664439e+00 -1.48129269e-01 8.21279943e-01 -2.14967892e-01
-9.42018449e-01 -2.29636550e-01 -7.23296821e-01 5.90186596e-01
2.07821682e-01 -3.50377202e-01 6.43336117e-01 7.76180983e-01
-2.79421732e-02 1.25055417e-01 -1.04393959e+00 1.14801931e+00
2.18393207e-01 -1.44126928e+00 6.48239791e-01 -1.64092794e-01
1.16095102e+00 -6.54714286e-01 4.85815018e-01 2.81203240e-01
3.58612061e-01 -1.00568509e+00 1.10975325e+00 3.07966173e-01
9.97140765e-01 -7.24790931e-01 2.91608691e-01 -1.03888199e-01
-1.15807331e+00 5.74647188e-02 -4.03743654e-01 -7.62531534e-02
2.60764241e-01 6.81427300e-01 -7.44811147e-02 6.28379881e-01
7.02906430e-01 8.35909784e-01 -6.63179040e-01 1.08580732e+00
-2.88119286e-01 6.57660067e-01 5.95423877e-02 -3.54779691e-01
1.95654377e-01 -4.20761317e-01 7.18191385e-01 1.16909480e+00
6.38764977e-01 -1.17320478e-01 -1.66708842e-01 1.22141814e+00
-6.72785491e-02 2.36179963e-01 -7.42792666e-01 4.18737680e-02
4.90911901e-01 1.09589851e+00 -6.86105072e-01 -2.77585864e-01
-1.58386976e-01 1.44114327e+00 3.43033940e-01 4.66466099e-01
-8.26119542e-01 -6.44717991e-01 5.67088366e-01 1.15504175e-01
4.58228469e-01 -2.52964377e-01 -8.03150535e-01 -1.19626486e+00
1.05749816e-01 -1.29206645e+00 6.73397779e-02 -9.19973135e-01
-1.62776506e+00 8.68365705e-01 -8.78950581e-02 -1.42707539e+00
8.09828416e-02 -6.25155866e-01 -7.73213446e-01 1.02600956e+00
-1.07411683e+00 -1.31754017e+00 -3.64383012e-01 3.53905737e-01
7.17047274e-01 -4.11953419e-01 6.26307011e-01 3.05178523e-01
-4.82073784e-01 9.80868697e-01 3.38628799e-01 -2.11569797e-02
9.86339092e-01 -1.07572293e+00 6.52951002e-01 7.99112797e-01
-4.60202480e-03 6.62891805e-01 6.18567824e-01 -6.55551195e-01
-1.33584750e+00 -1.20050275e+00 4.13091272e-01 -4.18942332e-01
5.83282351e-01 -3.84139955e-01 -8.21020663e-01 2.98454434e-01
3.79782021e-01 -4.29471016e-01 6.46363378e-01 -1.02226501e-02
-4.60829765e-01 -1.69200808e-01 -9.64875281e-01 1.18450141e+00
1.46941638e+00 -3.50211233e-01 -3.99993062e-01 -2.32763320e-01
6.77124143e-01 -3.79309833e-01 -5.94763815e-01 3.67309570e-01
8.16571474e-01 -1.18351328e+00 7.36570418e-01 -4.43589777e-01
8.37517798e-01 -2.38949999e-01 -1.65539339e-01 -1.77610874e+00
-7.01921463e-01 -8.78648043e-01 2.47907460e-01 1.49044275e+00
4.63144511e-01 -6.61208451e-01 5.83500981e-01 2.78284043e-01
-9.27776396e-02 -9.62500215e-01 -6.81427300e-01 -8.59601438e-01
4.97309744e-01 6.15601353e-02 7.86013782e-01 7.59135008e-01
-2.49480262e-01 3.76056433e-01 -5.79185963e-01 -4.47042435e-01
3.36209178e-01 4.13993955e-01 9.16297853e-01 -9.45340633e-01
-4.37963963e-01 -8.91389370e-01 -1.99878111e-01 -1.29167783e+00
-1.11151136e-01 -7.98792422e-01 -4.32968400e-02 -1.50127673e+00
2.04602495e-01 -6.36401176e-01 -1.55856714e-01 3.27548057e-01
-5.87552071e-01 6.79520547e-01 4.92080480e-01 2.43093073e-01
-3.14847946e-01 8.80899251e-01 1.83199453e+00 -3.12520027e-01
-3.32731813e-01 -2.06997693e-01 -1.16771042e+00 3.47086817e-01
8.88836324e-01 -2.24197641e-01 -4.79656875e-01 -6.58284128e-01
5.09724200e-01 -2.97660589e-01 2.00725988e-01 -9.77604449e-01
-3.11860472e-01 -2.96226710e-01 2.42259726e-01 -2.96321630e-01
1.49657458e-01 -4.03366029e-01 3.04019570e-01 2.18795806e-01
-3.38484377e-01 1.15183696e-01 3.04416090e-01 4.54368472e-01
-2.91975737e-01 -2.50734780e-02 8.80725265e-01 -2.33734310e-01
-5.22150159e-01 3.01149786e-01 -3.11989158e-01 1.42600417e-01
8.16302121e-01 -5.08685470e-01 -1.41244426e-01 -6.03919923e-01
-2.86012530e-01 -2.02373028e-01 8.06582034e-01 6.51972413e-01
4.99978274e-01 -1.77677691e+00 -8.31051826e-01 2.13807821e-01
1.77692734e-02 -2.18270913e-01 6.93610832e-02 3.84287536e-01
-5.92102349e-01 1.10911401e-02 -7.15485096e-01 -5.21237135e-01
-9.07517314e-01 3.77916753e-01 1.42164856e-01 1.61275752e-02
-3.98488343e-01 1.04366183e+00 5.86636603e-01 -3.92825842e-01
-1.38762835e-02 -8.82373750e-02 5.64558804e-02 -3.23365815e-02
4.23108637e-01 2.38368928e-01 -1.98839307e-01 -2.81978577e-01
5.09655215e-02 8.25901866e-01 7.76886269e-02 -3.13707471e-01
1.16258252e+00 -1.60736609e-02 -1.29746646e-01 5.47880411e-01
7.89307177e-01 1.93939999e-01 -1.32124746e+00 -2.43697930e-02
-5.78473032e-01 -7.88189292e-01 -4.18117762e-01 -7.57563174e-01
-1.20845938e+00 9.33949411e-01 3.64382178e-01 8.63709524e-02
1.27788460e+00 -2.78877616e-01 9.59932268e-01 -6.52038977e-02
6.49340391e-01 -8.51479709e-01 3.32974732e-01 4.58018780e-01
1.23168838e+00 -9.19105887e-01 -2.02730209e-01 -5.44533432e-01
-8.87437999e-01 9.45331991e-01 7.74644852e-01 -3.21298510e-01
2.25481585e-01 1.36739954e-01 2.24460825e-01 2.62375206e-01
-5.00956476e-01 1.11798279e-01 4.67699826e-01 6.87632680e-01
6.08905554e-01 4.15067039e-02 -4.63019133e-01 5.74535251e-01
-4.41553146e-01 -2.61608995e-02 2.22144872e-01 8.28564882e-01
-3.48215073e-01 -1.43185890e+00 -3.63085479e-01 3.96332353e-01
-1.23817042e-01 -2.92811424e-01 -6.06497288e-01 5.48407495e-01
-1.04270484e-02 6.06578887e-01 -1.57919243e-01 -5.99858761e-01
3.40365142e-01 -1.38180494e-01 7.20907509e-01 -3.98559064e-01
-7.36878991e-01 -4.70072888e-02 -1.73606291e-01 -4.69949365e-01
-2.16303781e-01 -3.75434428e-01 -6.18812084e-01 -5.12656987e-01
-1.50514096e-01 -1.33475602e-01 3.23972315e-01 6.02797270e-01
6.14561439e-01 7.23205864e-01 7.90849268e-01 -1.18718052e+00
-2.70393103e-01 -8.33396971e-01 -6.74200833e-01 6.06335521e-01
-1.50522113e-01 -6.79658592e-01 -1.47004038e-01 1.73633561e-01] | [11.62535285949707, -0.45740777254104614] |
3e0ad605-c77d-455a-9048-2f95b500c987 | a-corpus-based-approach-for-spanish-chinese | null | null | https://aclanthology.org/W16-4913 | https://aclanthology.org/W16-4913.pdf | A Corpus-based Approach for Spanish-Chinese Language Learning | Due to the huge population that speaks Spanish and Chinese, these languages occupy an important position in the language learning studies. Although there are some automatic translation systems that benefit the learning of both languages, there is enough space to create resources in order to help language learners. As a quick and effective resource that can give large amount language information, corpus-based learning is becoming more and more popular. In this paper we enrich a Spanish-Chinese parallel corpus automatically with part of-speech (POS) information and manually with discourse segmentation (following the Rhetorical Structure Theory (RST) (Mann and Thompson, 1988)). Two search tools allow the Spanish-Chinese language learners to carry out different queries based on tokens and lemmas. The parallel corpus and the research tools are available to the academic community. We propose some examples to illustrate how learners can use the corpus to learn Spanish and Chinese. | ['Shuyuan Cao', 'Mikel Iruskieta', 'Iria da Cunha'] | 2016-12-01 | null | null | null | ws-2016-12 | ['discourse-segmentation'] | ['natural-language-processing'] | [-1.88460082e-01 1.64490212e-02 -5.72949409e-01 -1.98867798e-01
-8.88660312e-01 -8.32519710e-01 6.53039634e-01 3.55917603e-01
-5.82671344e-01 1.09483695e+00 1.70623735e-01 -1.00880659e+00
2.65467584e-01 -8.03077102e-01 -2.67294109e-01 -4.10668761e-01
2.70986676e-01 6.24459863e-01 5.03831804e-01 -6.14597023e-01
6.22316778e-01 3.49279374e-01 -1.43945813e+00 3.81000042e-01
1.64717841e+00 2.52687559e-02 1.08180439e+00 2.74865299e-01
-1.09241784e+00 9.03175592e-01 -6.95688844e-01 -4.71007705e-01
-3.03624541e-01 -6.76942468e-01 -1.28562212e+00 -2.74903506e-01
-2.21002981e-01 6.66242167e-02 3.96712363e-01 1.19407177e+00
3.15543979e-01 -1.40265718e-01 5.06629109e-01 -7.74278820e-01
-6.59468293e-01 1.02239943e+00 -2.91869760e-01 1.88138261e-01
6.34098589e-01 -4.45885748e-01 7.63669312e-01 -8.01616192e-01
9.63717163e-01 1.36813259e+00 1.86927110e-01 4.75516558e-01
-5.58509588e-01 -5.13101399e-01 1.88237101e-01 5.59974253e-01
-1.04974830e+00 8.96805674e-02 6.65006936e-01 -2.89424896e-01
7.99499333e-01 1.60267502e-01 1.04246759e+00 9.07851994e-01
-2.11833175e-02 9.98367846e-01 1.42313457e+00 -1.24024260e+00
-1.80248860e-02 8.33534062e-01 1.81271911e-01 3.76693040e-01
1.04824379e-01 -2.19147772e-01 -2.79285073e-01 3.60258132e-01
4.51896876e-01 -1.16271749e-01 -6.20506816e-02 5.25738038e-02
-1.10895765e+00 9.45051491e-01 -2.97480244e-02 1.12347710e+00
1.18174501e-01 -4.88540947e-01 4.21799809e-01 6.11505389e-01
-1.21300220e-02 6.79480672e-01 -6.84038401e-01 -6.24895573e-01
-5.43825865e-01 1.38251901e-01 1.14466739e+00 1.22296000e+00
5.41510999e-01 -2.61259019e-01 2.44219497e-01 8.73529255e-01
3.65837783e-01 8.24946523e-01 8.51964533e-01 -7.29193687e-01
5.04909277e-01 7.75723517e-01 -2.59740025e-01 -6.20889306e-01
2.01325059e-01 -1.20780379e-01 -3.55234563e-01 -1.21366404e-01
6.58712983e-01 -2.56100267e-01 -3.06155503e-01 1.29770374e+00
2.15349779e-01 -5.83288491e-01 4.16866422e-01 3.88838351e-01
9.05198872e-01 9.55741644e-01 2.57376581e-01 -6.31022573e-01
1.59073961e+00 -9.57005024e-01 -8.89410377e-01 1.46749904e-02
9.54152882e-01 -1.25312364e+00 1.36204481e+00 3.73031706e-01
-1.08585894e+00 -4.26748633e-01 -7.77341187e-01 -2.44389951e-01
-8.06572199e-01 -5.29259667e-02 4.80815679e-01 6.52541339e-01
-8.20026159e-01 2.74540037e-01 -6.75473928e-01 -6.86638653e-01
-2.55204421e-02 1.86370075e-01 -4.61209230e-02 -3.68334860e-01
-1.27585435e+00 1.13802361e+00 5.24527311e-01 -3.94712389e-01
-4.67508197e-01 -7.13696778e-02 -7.17276037e-01 -1.22579597e-01
4.25024480e-01 -1.01506868e-02 1.51004314e+00 -1.47106814e+00
-1.66174829e+00 9.65545058e-01 -1.12548955e-01 5.61925694e-02
4.29353833e-01 -1.61252066e-01 -2.73219943e-01 2.99951136e-01
4.72937435e-01 5.11060894e-01 1.01982363e-01 -1.03907323e+00
-1.09094775e+00 -2.46775299e-01 8.60541910e-02 4.70444083e-01
-5.21966338e-01 7.94277847e-01 -5.56375206e-01 -5.84972918e-01
-1.91602446e-02 -6.86547339e-01 4.24693935e-02 -5.88601530e-01
3.25602083e-03 -8.40727568e-01 7.52616286e-01 -9.04077053e-01
1.40133786e+00 -1.73179507e+00 -7.32851997e-02 5.83260842e-02
-3.21265519e-01 4.69180405e-01 3.54355648e-02 7.53043890e-01
3.21185887e-01 4.08967108e-01 -1.23357765e-01 3.62747610e-01
-1.78474665e-01 5.37752807e-01 1.77124545e-01 -2.37853184e-01
-1.19476885e-01 6.55644894e-01 -1.20435417e+00 -9.06868994e-01
-5.95836677e-02 2.38541782e-01 -3.30972672e-01 2.30486065e-01
-1.94081634e-01 5.78920424e-01 -8.58102322e-01 7.98173606e-01
1.99660912e-01 5.06630689e-02 5.92095852e-01 8.18054020e-01
-7.75054514e-01 7.14687586e-01 -8.44390810e-01 1.43755782e+00
-6.07130229e-01 6.72481537e-01 1.04107521e-01 -1.16680562e+00
9.56059635e-01 6.46481812e-01 -1.24083154e-01 -7.94808567e-01
7.51385912e-02 9.14718390e-01 3.50072205e-01 -7.66949236e-01
3.68801147e-01 -1.56212211e-01 -1.66779518e-01 5.12516141e-01
-7.37373978e-02 -3.03218186e-01 7.01908529e-01 1.12573192e-01
4.69319969e-01 3.47921401e-01 5.99864244e-01 -5.79335093e-01
1.02013624e+00 5.06181538e-01 4.32244301e-01 1.91900998e-01
-1.49545157e-02 -2.03174893e-02 3.61750990e-01 -2.13789865e-01
-9.40444410e-01 -5.73119760e-01 -1.81144997e-01 1.29202986e+00
-1.13977239e-01 -3.64295542e-01 -9.35067356e-01 -6.24918520e-01
-6.21088326e-01 8.37184966e-01 5.59539208e-03 5.77873468e-01
-1.02478838e+00 -7.45212808e-02 1.06085770e-01 3.55102926e-01
6.13448501e-01 -1.65686452e+00 -4.66662794e-01 2.78463751e-01
-3.67479742e-01 -7.99486756e-01 -2.58454770e-01 2.12531313e-01
-7.44825900e-01 -9.81426954e-01 -9.67576504e-01 -1.72404647e+00
6.93740904e-01 2.99640775e-01 8.83715868e-01 4.89978790e-01
1.41897798e-01 2.44990975e-01 -8.20109308e-01 -8.77031922e-01
-8.33034039e-01 4.71113056e-01 -3.15167129e-01 -9.52118099e-01
7.98778057e-01 -4.30491537e-01 3.77691388e-01 -1.44630387e-01
-6.40196145e-01 2.34288052e-01 6.38823867e-01 5.61996043e-01
2.53465742e-01 -1.02954499e-01 6.48483515e-01 -9.67764676e-01
7.13442206e-01 -2.92327762e-01 -7.44222581e-01 4.58976865e-01
-6.29137516e-01 7.14674667e-02 6.93471909e-01 -3.06277812e-01
-1.29401493e+00 -2.83532470e-01 -2.92292386e-01 3.83780748e-01
-3.94759834e-01 8.71844530e-01 -4.59245414e-01 1.11595601e-01
3.59380513e-01 2.97886848e-01 -6.82437569e-02 -5.16809642e-01
1.91354051e-01 9.23474073e-01 1.36377245e-01 -8.39667082e-01
6.29640758e-01 -3.72663379e-01 -2.80304819e-01 -1.18202055e+00
-5.49995661e-01 -5.63791931e-01 -9.06128764e-01 -2.05938786e-01
1.03387988e+00 -8.00275505e-01 -4.77409154e-01 3.46552767e-02
-1.33459187e+00 -3.02710205e-01 -4.03356105e-02 8.09104860e-01
-1.47076309e-01 3.80230665e-01 -6.80884421e-01 -9.07719135e-01
7.87063316e-02 -1.16039884e+00 3.49546611e-01 6.49707556e-01
-7.55176097e-02 -1.20573282e+00 7.96997640e-03 2.92811334e-01
2.21883684e-01 -2.45667920e-01 1.48625386e+00 -9.36651945e-01
-5.95786214e-01 2.08718702e-01 1.63098037e-01 2.74600387e-01
-1.14577286e-01 1.65982470e-02 -4.18297589e-01 1.07020922e-01
4.65162210e-02 -4.77693617e-01 2.07977340e-01 -5.95176183e-02
6.08726442e-01 -5.58882952e-01 -3.25427443e-01 2.15938408e-02
1.37040532e+00 7.57432759e-01 4.43915933e-01 8.11326087e-01
4.46191669e-01 1.14857709e+00 6.52939439e-01 -1.45371675e-01
5.63051224e-01 3.50039184e-01 -2.91538894e-01 2.37251282e-01
4.54226658e-02 -4.36953336e-01 7.57719755e-01 1.73543286e+00
-2.00195014e-01 1.23850301e-01 -1.32529390e+00 7.10015893e-01
-1.58782613e+00 -8.93665195e-01 -3.50614190e-01 1.98515821e+00
1.40634573e+00 8.45898967e-03 1.38759345e-01 3.36470228e-04
6.40212238e-01 -2.84450114e-01 3.69163901e-01 -6.15721524e-01
-2.47043267e-01 3.02457005e-01 1.53165177e-01 7.86866069e-01
-5.60820103e-01 1.38843071e+00 5.84210491e+00 9.18726742e-01
-1.18649340e+00 -6.46341965e-02 4.02089149e-01 7.36268520e-01
-5.83916545e-01 2.80611753e-01 -1.03959739e+00 2.53510028e-01
9.91573870e-01 -4.64879483e-01 3.05299282e-01 8.78876925e-01
2.11182579e-01 -3.87519181e-01 -7.05622315e-01 8.07135403e-01
-1.75278671e-02 -1.20731378e+00 6.00216798e-02 8.76311585e-02
8.68280470e-01 -2.02347144e-01 -3.47836256e-01 5.63055575e-01
3.57649595e-01 -8.43505025e-01 5.95060050e-01 2.12609600e-02
6.49705589e-01 -7.31425881e-01 9.67124939e-01 8.83992553e-01
-9.64938819e-01 9.15779471e-02 -4.89040285e-01 -3.12846780e-01
-1.27427354e-01 -1.09313205e-01 -8.61831486e-01 4.64900881e-01
5.69012046e-01 4.54134643e-01 -6.53140306e-01 9.04562593e-01
-8.11239183e-01 7.85785615e-01 -6.47788420e-02 -1.02456546e+00
3.30218941e-01 -7.16951847e-01 4.38857734e-01 1.12182021e+00
5.96656442e-01 3.53389323e-01 5.99246860e-01 5.00237942e-01
3.33883345e-01 9.08936381e-01 -6.46861792e-01 -2.42864817e-01
6.72962725e-01 9.99656141e-01 -9.78379846e-01 -5.40400863e-01
-7.16677487e-01 7.81550646e-01 1.33255586e-01 3.08560401e-01
-3.14582065e-02 -6.83082521e-01 -7.03028888e-02 1.91215649e-01
9.31773037e-02 -3.82452548e-01 -2.84263879e-01 -9.19060111e-01
-1.66228026e-01 -1.41804600e+00 1.10910438e-01 -4.01806533e-01
-9.96104538e-01 5.16483843e-01 2.71527231e-01 -1.00163996e+00
-5.27160764e-01 -7.15751588e-01 -6.61434710e-01 8.58560264e-01
-1.58587992e+00 -1.14007676e+00 6.37582690e-02 4.23507184e-01
9.26837206e-01 -6.19057536e-01 9.95603919e-01 2.26148129e-01
-2.87515998e-01 1.75959095e-01 1.74298257e-01 4.00373638e-01
5.83758414e-01 -1.34379041e+00 -2.94604003e-01 6.67384624e-01
1.23063013e-01 6.79594755e-01 4.47382897e-01 -6.67761624e-01
-1.26275742e+00 -3.12765867e-01 2.00484347e+00 -1.40666276e-01
6.58739865e-01 -1.46699578e-01 -8.61490250e-01 4.91186857e-01
9.89424288e-01 -7.27901161e-01 9.60088134e-01 2.39183363e-02
3.11638176e-01 7.07526654e-02 -6.29859388e-01 7.95556545e-01
6.66890740e-01 -3.09904456e-01 -1.28323901e+00 4.17853028e-01
5.87620854e-01 -8.88926610e-02 -4.53262299e-01 -1.36565030e-01
2.36636341e-01 -7.55007625e-01 2.84784079e-01 -4.17205602e-01
3.88021648e-01 -1.86749965e-01 1.47505581e-01 -1.06326687e+00
1.68180853e-01 -8.94430637e-01 6.98447943e-01 1.40184104e+00
7.77078092e-01 -5.11398852e-01 4.77891743e-01 2.67373025e-01
-3.19768012e-01 -4.61219102e-01 -6.48669899e-01 -6.04164898e-01
6.50302827e-01 -2.49852553e-01 4.77570832e-01 1.11800873e+00
7.44521916e-01 8.15423191e-01 2.62414902e-01 -5.19494653e-01
-3.72662838e-03 1.57479376e-01 6.76605999e-01 -1.54636979e+00
-7.64986351e-02 -6.21050060e-01 3.75469118e-01 -1.11415362e+00
2.50331789e-01 -1.07523286e+00 6.46037161e-02 -1.65833509e+00
-1.56102218e-02 -5.25595486e-01 2.71332592e-01 3.27916533e-01
-1.97552323e-01 -4.66997951e-01 2.59306818e-01 2.78423131e-01
-5.49654245e-01 3.79660696e-01 1.54161394e+00 2.55827516e-01
-4.83471215e-01 4.37091216e-02 -6.36207581e-01 8.80295932e-01
7.98727214e-01 -5.17626226e-01 -2.39186347e-01 -4.13969666e-01
2.52407253e-01 1.38382167e-01 -5.23418665e-01 -7.45099723e-01
2.03949600e-01 -6.23259366e-01 1.99636072e-01 -6.68811738e-01
-3.62523705e-01 -7.89331675e-01 -5.79715371e-01 6.49971247e-01
-2.08083078e-01 5.28383255e-01 1.09257966e-01 -2.65544325e-01
-6.57185137e-01 -8.10730696e-01 6.46978915e-01 -6.02451622e-01
-5.74425757e-01 -1.11141920e-01 -8.00201774e-01 2.40026578e-01
1.15990424e+00 -1.22752808e-01 1.31223979e-03 -4.41313237e-01
-3.08757484e-01 3.12855929e-01 4.36523020e-01 3.55788827e-01
2.72242516e-01 -1.20407307e+00 -6.08909547e-01 1.06176227e-01
-7.70793259e-02 -7.34951347e-02 -3.62578541e-01 8.14028978e-01
-9.93686199e-01 9.07897115e-01 -4.39342529e-01 -2.98249394e-01
-1.26774800e+00 4.62069392e-01 -2.25548089e-01 -4.11279827e-01
-2.13513970e-01 6.23963714e-01 -1.32805824e-01 -5.92952847e-01
3.27796727e-01 -2.37434044e-01 -9.38996971e-01 2.52841771e-01
6.27956629e-01 3.18683922e-01 -3.93472224e-01 -6.79057717e-01
-9.47658569e-02 4.52071458e-01 1.77505746e-01 -5.77541828e-01
1.11502314e+00 -2.34445125e-01 -3.54071766e-01 6.62732184e-01
9.56866086e-01 6.69890702e-01 -4.26339269e-01 -1.33518741e-01
6.38014317e-01 -2.06910685e-01 -4.11065429e-01 -7.73032725e-01
-2.24923804e-01 1.23034310e+00 2.36634761e-01 3.22361857e-01
7.90997803e-01 6.48215339e-02 6.78456187e-01 5.98440111e-01
3.69711280e-01 -1.64538157e+00 -4.87008356e-02 1.19286823e+00
5.54180801e-01 -1.25311267e+00 -3.07968944e-01 -5.66700876e-01
-7.04472065e-01 1.43456197e+00 6.28535271e-01 2.00324416e-01
6.05289400e-01 1.42308369e-01 5.03154516e-01 2.72039056e-01
-4.64732647e-01 -3.65603000e-01 4.79173698e-02 8.85711312e-01
1.19851995e+00 3.49765196e-02 -1.46536970e+00 6.45418286e-01
-5.12942135e-01 -2.23472193e-01 3.86900514e-01 1.11487985e+00
-1.03746593e+00 -2.15917349e+00 -5.73164940e-01 -1.19608611e-01
-6.59883440e-01 -3.21807384e-01 -5.62709332e-01 1.26534212e+00
2.08806396e-01 1.01251602e+00 -1.34605229e-01 2.12947324e-01
3.13406810e-02 3.37581635e-01 3.95763695e-01 -9.78445768e-01
-8.06668401e-01 5.94034016e-01 2.13282451e-01 1.12228714e-01
-5.91052175e-01 -8.95873606e-01 -1.64430594e+00 -2.16833323e-01
-1.03518799e-01 9.20905352e-01 6.64363503e-01 1.21022463e+00
-3.66671443e-01 1.31147519e-01 3.56259167e-01 -2.75657773e-01
-4.26062495e-01 -1.22031581e+00 -4.11166251e-01 -1.14741936e-01
-3.35353881e-01 -2.31183663e-01 -3.37857120e-02 1.34673016e-03] | [10.514493942260742, 10.066756248474121] |
72602498-1e4c-4f74-8f38-f4253af80fc9 | understanding-user-resistance-strategies-in | null | null | https://aclanthology.org/2020.findings-emnlp.431 | https://aclanthology.org/2020.findings-emnlp.431.pdf | Understanding User Resistance Strategies in Persuasive Conversations | Persuasive dialog systems have various usages, such as donation persuasion and physical exercise persuasion. Previous persuasive dialog systems research mostly focused on analyzing the persuader{'}s strategies and paid little attention to the persuadee (user). However, understanding and addressing users{'} resistance strategies is an essential job of a persuasive dialog system. So, we adopt a preliminary framework on persuasion resistance in psychology and design a fine-grained resistance strategy annotation scheme. We annotate the PersuasionForGood dataset with the scheme. With the enriched annotations, we build a classifier to predict the resistance strategies. Furthermore, we analyze the relationships between persuasion strategies and persuasion resistance strategies. Our work lays the ground for developing a persuasive dialogue system that can understand and address user resistance strategy appropriately. The code and data will be released. | ['Zhou Yu', 'Chen Li', 'Weiyan Shi', 'Youzhi Tian'] | 2020-11-01 | null | null | null | findings-of-the-association-for-computational | ['persuasion-strategies'] | ['computer-vision'] | [ 4.42493021e-01 8.03294480e-01 -5.53512812e-01 -6.62204862e-01
-2.40262643e-01 -4.33516681e-01 8.00099134e-01 2.39334583e-01
-3.19204986e-01 8.90874267e-01 9.44858134e-01 -7.93220460e-01
-3.54658097e-01 -6.59709275e-01 2.77706087e-01 -1.84770286e-01
8.12098801e-01 1.53197289e-01 4.24359962e-02 -1.08637583e+00
8.14189434e-01 -6.82732165e-02 -1.13440347e+00 1.45604730e-01
1.12794280e+00 2.23570302e-01 1.19803011e-01 8.09681296e-01
-4.16548073e-01 1.24160206e+00 -5.16726851e-01 -2.39554554e-01
-2.58812219e-01 -9.80077267e-01 -1.42523491e+00 -3.51314813e-01
-7.48315006e-02 -3.94988149e-01 1.57461867e-01 7.87439287e-01
4.47396457e-01 1.91428348e-01 6.04818583e-01 -1.16301370e+00
-1.04856658e+00 1.46487737e+00 -8.74977857e-02 3.10558289e-01
8.26837003e-01 -5.76770455e-02 7.96842396e-01 -1.71136633e-01
2.67666727e-01 1.56962407e+00 6.54095292e-01 1.42138040e+00
-1.07184112e+00 -3.52699786e-01 1.43977970e-01 3.23654145e-01
-6.35542631e-01 -5.70362985e-01 1.02537489e+00 -5.36757231e-01
4.02190000e-01 7.98431933e-01 4.99974191e-01 1.61074913e+00
-3.34444970e-01 7.48989940e-01 1.29817975e+00 -5.24232090e-01
1.44388050e-01 5.24521768e-01 1.37193620e+00 6.73600137e-01
7.46994046e-04 -2.31676891e-01 -2.59645462e-01 -3.04290235e-01
4.57117915e-01 -3.20888549e-01 -2.19076708e-01 4.31221187e-01
-8.22613358e-01 1.30647194e+00 3.42538245e-02 4.00520623e-01
-1.89766929e-01 -3.10600847e-01 3.88479233e-01 4.23745960e-01
2.37001300e-01 5.81842542e-01 -4.22940850e-01 -5.08148193e-01
9.09826756e-02 9.72166583e-02 1.27024269e+00 4.11987484e-01
6.54208183e-01 -1.71066269e-01 -3.35588932e-01 1.17392254e+00
5.32762349e-01 2.42543429e-01 4.76673722e-01 -1.15105891e+00
-2.44752541e-01 1.02370751e+00 3.91061306e-01 -1.01719260e+00
-6.41686678e-01 4.96716440e-01 -5.20499706e-01 2.17589699e-02
4.62994605e-01 -4.87108976e-01 4.37005349e-02 1.84913099e+00
5.86443067e-01 -3.71018767e-01 3.28689218e-01 1.10097814e+00
1.26661384e+00 1.49032131e-01 7.12159932e-01 -3.83081764e-01
1.75817406e+00 -7.70321190e-01 -1.12341642e+00 -1.05747312e-01
7.55310416e-01 -6.63519263e-01 1.65963733e+00 -1.37038723e-01
-6.74337566e-01 -5.78198433e-01 -8.97600770e-01 -3.68834108e-01
-1.87971622e-01 -6.96329623e-02 9.07154918e-01 1.15792692e+00
-6.63680315e-01 3.54781359e-01 -3.11258227e-01 -6.60974920e-01
2.26380274e-01 -1.32562369e-01 2.18912959e-01 7.06639647e-01
-1.71890211e+00 1.24712074e+00 3.54336798e-02 -2.35003263e-01
-3.14016134e-01 -6.77508950e-01 -7.80939519e-01 -8.75426754e-02
5.72810411e-01 -8.42464089e-01 1.51773417e+00 -1.06346452e+00
-1.90243161e+00 7.22245634e-01 3.28066975e-01 -2.94436544e-01
4.47434425e-01 -2.99346000e-01 -3.66637409e-01 -1.33291006e-01
1.73784956e-01 3.56548369e-01 5.18918812e-01 -8.95763874e-01
-6.28157496e-01 -2.72988468e-01 3.78344864e-01 5.77689528e-01
-4.60844785e-01 5.21187901e-01 5.73983788e-01 -2.06739768e-01
-5.17928064e-01 -4.20959383e-01 -2.39375129e-01 -4.24930274e-01
-4.42271173e-01 -1.07426953e+00 7.84754038e-01 -3.90922666e-01
1.49907124e+00 -1.89535224e+00 -4.06696051e-02 -3.51951689e-01
6.16126239e-01 4.19676512e-01 1.22239597e-01 4.25765991e-01
2.48475209e-01 4.40234452e-01 4.75968540e-01 2.12779805e-01
4.46396053e-01 5.46804480e-02 -2.69395530e-01 1.08790502e-01
-3.21513891e-01 6.66693389e-01 -8.48283172e-01 -5.14888823e-01
3.23203683e-01 2.39674494e-01 -3.86446536e-01 3.90531451e-01
-2.20806822e-01 4.33461040e-01 -1.27783525e+00 1.20678864e-01
3.76219481e-01 -2.70455688e-01 4.48248804e-01 6.68297857e-02
-4.55072016e-01 6.92643523e-01 -5.80035806e-01 1.18335152e+00
-2.05538958e-01 3.36051732e-01 1.47393420e-01 -6.98295832e-01
1.18690956e+00 1.09088287e-01 2.02169880e-01 -4.20968771e-01
6.43137574e-01 -8.94312486e-02 3.07042867e-01 -7.33471513e-01
6.87099814e-01 -3.31761874e-02 -3.54565680e-01 9.26726699e-01
-7.39561319e-01 3.24048579e-01 -1.12420544e-01 5.38655102e-01
9.84255373e-01 7.09924549e-02 6.65487826e-01 -5.78630447e-01
1.03830552e+00 2.90745974e-01 2.93092161e-01 9.20200169e-01
-5.72063804e-01 -3.24647576e-01 4.27740753e-01 -5.25875092e-01
-4.96068209e-01 -4.90642309e-01 1.24353155e-01 1.58202946e+00
4.22593206e-01 -5.19758284e-01 -1.15478361e+00 -8.98197711e-01
-3.09609979e-01 1.13967240e+00 -6.72497451e-01 -3.35319966e-01
-4.88952100e-01 -6.80384099e-01 7.10155368e-01 2.08800480e-01
9.48602498e-01 -1.13277853e+00 -7.85917819e-01 2.51030236e-01
-5.11365354e-01 -5.73865712e-01 -4.11353469e-01 -3.41424614e-01
-2.76741952e-01 -1.43397796e+00 3.60227704e-01 -5.42634428e-01
1.40096515e-01 4.81188923e-01 7.89405942e-01 6.42300785e-01
-3.77827100e-02 6.72117651e-01 -5.30068398e-01 -5.48731387e-01
-9.77028549e-01 -2.14331481e-03 2.24027053e-01 -4.29972112e-01
9.28872406e-01 -2.58068144e-01 -7.59198368e-01 6.70726955e-01
-2.18154401e-01 3.86268795e-01 -6.66881949e-02 4.42009807e-01
-2.33365387e-01 -5.19594014e-01 1.08185470e+00 -1.14775002e+00
1.62413228e+00 -4.18399513e-01 -8.09402205e-03 2.14939013e-01
-1.07621157e+00 -1.61479458e-01 5.95671892e-01 -7.04598904e-01
-1.53976774e+00 -4.80070442e-01 -6.29680932e-01 7.68850625e-01
-4.76961851e-01 1.09930858e-02 -2.36886472e-01 2.72652030e-01
1.15194106e+00 -2.20766813e-01 2.68637180e-01 -7.20877051e-01
7.36346245e-01 1.05225921e+00 1.42689899e-01 -8.43354583e-01
1.51799887e-01 3.72765660e-02 -6.50548160e-01 -1.03134489e+00
-1.16910040e+00 -1.84115335e-01 -7.09359646e-02 -5.68883061e-01
1.12022424e+00 -1.92035168e-01 -1.58366227e+00 1.35180559e-02
-9.13964927e-01 -4.38211411e-01 -2.03089640e-01 2.05915749e-01
-6.07379735e-01 8.61192822e-01 -5.50036073e-01 -1.23336625e+00
-7.82006741e-01 -7.36189842e-01 6.19759083e-01 7.24118114e-01
-1.12683523e+00 -1.30499339e+00 2.23300040e-01 1.00616586e+00
5.20526528e-01 -2.44959220e-01 1.04533648e+00 -1.10414088e+00
2.56298363e-01 3.02322358e-01 -8.17498043e-02 9.15054418e-03
3.12216401e-01 -3.18643987e-01 -6.23183846e-01 4.98390406e-01
2.07263112e-01 -4.83881235e-01 2.24045172e-01 1.49592355e-01
8.00101340e-01 -5.72585285e-01 -2.87782371e-01 -2.08607748e-01
6.12157464e-01 1.62060052e-01 6.72537684e-01 3.51867020e-01
4.12410527e-01 8.93325567e-01 9.92650867e-01 3.25362593e-01
9.82723534e-01 4.59686279e-01 -1.84056610e-01 -4.25350331e-02
-1.22269318e-01 -5.28203666e-01 4.31899577e-01 3.92929822e-01
-2.06601918e-01 -7.85775855e-02 -6.57377481e-01 5.36568835e-03
-2.05782437e+00 -1.12410367e+00 -7.48714626e-01 1.58001602e+00
1.31273985e+00 -2.21764192e-01 2.60781914e-01 2.97042400e-01
6.79660141e-01 3.52962315e-02 -2.65755266e-01 -9.18706954e-01
1.05483398e-01 -3.31973940e-01 -2.05820054e-01 1.22002590e+00
-7.00851619e-01 1.31001568e+00 6.47775936e+00 4.33137029e-01
-6.32882416e-01 1.85540035e-01 4.88551676e-01 7.44040370e-01
-4.60441142e-01 1.76693916e-01 -1.02208674e+00 3.05370450e-01
8.82861674e-01 -3.08338523e-01 2.98781753e-01 7.80856431e-01
7.37069011e-01 -2.45885208e-01 -1.11799407e+00 4.70981002e-01
-1.08303092e-01 -1.18288887e+00 -3.41774881e-01 -1.41607299e-01
-1.35615259e-01 -8.44063401e-01 -3.80885571e-01 6.55748367e-01
8.53109300e-01 -7.75625527e-01 8.59314054e-02 5.95829964e-01
2.39750281e-01 -1.15910873e-01 4.40122545e-01 6.66778624e-01
-3.04772735e-01 1.41245415e-02 -2.74948120e-01 -6.31380320e-01
1.58950761e-01 -7.74791837e-02 -1.16132474e+00 3.26015353e-02
4.25088942e-01 4.10301208e-01 -3.16591233e-01 -3.21196824e-01
-5.20833015e-01 8.64100754e-01 8.11237767e-02 -8.68691325e-01
-3.22343171e-01 -5.44393837e-01 4.81349170e-01 1.05629992e+00
-3.93393546e-01 7.82994986e-01 3.22344482e-01 8.46273899e-01
2.58785486e-01 4.83346283e-01 -3.10455620e-01 -1.70333281e-01
5.84527969e-01 1.34908926e+00 -4.18932617e-01 -4.21149284e-01
-1.71092287e-01 6.55281305e-01 2.89726466e-01 4.78304289e-02
-7.78476059e-01 4.11749259e-02 9.04738665e-01 -5.66581525e-02
-5.48201025e-01 2.30638042e-01 -7.29157865e-01 -7.64779747e-01
-8.40132535e-01 -1.02439332e+00 7.04348385e-01 -6.24565065e-01
-1.47128630e+00 3.40012848e-01 -4.36642580e-02 -2.65056938e-01
-8.80327076e-02 -3.39620173e-01 -6.99615479e-01 8.43789577e-01
-1.17922986e+00 -1.03054738e+00 -5.86846113e-01 4.85637605e-01
8.15411270e-01 -1.18915446e-01 1.10858369e+00 -1.98297203e-01
-6.36701643e-01 5.07229269e-01 -8.42228115e-01 -9.16587040e-02
6.54126227e-01 -1.19528377e+00 -1.64774194e-01 2.87059456e-01
-8.71910810e-01 9.50002789e-01 1.13624847e+00 -6.56363249e-01
-1.42460120e+00 -4.35062736e-01 9.10886943e-01 -8.68470430e-01
1.02270627e+00 -1.02316260e-01 -9.60709989e-01 4.08154786e-01
6.41098082e-01 -1.21817088e+00 1.29935598e+00 6.08429551e-01
-1.60507292e-01 3.90484005e-01 -1.20010102e+00 1.10595000e+00
1.37916815e+00 -2.19628796e-01 -1.21121812e+00 4.44351017e-01
6.46338403e-01 -3.07688080e-02 -8.91963780e-01 -1.44870475e-01
4.21889722e-01 -8.96498740e-01 9.59698319e-01 -1.03674364e+00
5.78427851e-01 -9.94963199e-03 1.41118571e-01 -1.07347882e+00
-3.71330053e-01 -9.24880624e-01 -7.12716766e-03 1.40165973e+00
1.53391942e-01 -6.91237748e-01 4.71378386e-01 1.14503813e+00
3.68445888e-02 -1.28615633e-01 -3.13461065e-01 8.15794021e-02
3.41907628e-02 -2.14821428e-01 6.35083079e-01 1.39736724e+00
6.98246062e-01 1.64536345e+00 -6.77995801e-01 -2.58156598e-01
4.04312581e-01 8.58815238e-02 9.91724312e-01 -1.30374157e+00
3.77663486e-02 -6.85023785e-01 3.88813943e-01 -1.18208849e+00
4.72630322e-01 -7.43177116e-01 -1.31904021e-01 -1.66823614e+00
1.71935335e-01 -5.97115874e-01 3.14889044e-01 8.26371014e-01
-7.80358791e-01 -2.80516833e-01 -2.94085354e-01 -4.62819599e-02
-5.55067420e-01 5.08249581e-01 1.18869543e+00 2.45762989e-01
-5.43894053e-01 -6.06217310e-02 -1.98477614e+00 8.11932027e-01
1.32715082e+00 -1.04024038e-01 -7.02219844e-01 8.57993290e-02
1.65507287e-01 1.29935384e-01 1.07968993e-01 8.81745815e-02
2.37758700e-02 -8.24769378e-01 -2.56819069e-01 -1.03154458e-01
5.84035106e-02 -4.96509522e-01 -7.35059083e-02 4.53481048e-01
-8.99096310e-01 -3.02482039e-01 -3.32102448e-01 4.97132748e-01
6.66401327e-01 -3.31578881e-01 7.47070312e-01 1.24615759e-01
-6.31291866e-01 -3.05536926e-01 -1.10801661e+00 -8.91640708e-02
8.79601419e-01 -2.92839766e-01 -1.00253117e+00 -6.31860316e-01
-6.72365785e-01 5.60493529e-01 3.45784158e-01 7.97458947e-01
4.66812223e-01 -1.01122737e+00 -8.99972200e-01 -2.21838251e-01
-1.03458695e-01 -8.75555694e-01 3.57383579e-01 8.61386716e-01
-4.28017136e-03 -1.20605193e-01 -2.94912636e-01 -9.51945186e-02
-1.67395294e+00 5.80761731e-01 4.70369697e-01 2.46128380e-01
-7.92387187e-01 4.43312287e-01 -6.23724759e-02 -4.69778270e-01
8.04753229e-02 4.31556433e-01 -1.24869239e+00 2.21718140e-02
9.16559339e-01 6.87931359e-01 -7.39488900e-01 -3.07153225e-01
-3.06508660e-01 6.00470416e-02 -1.97675020e-01 9.18434700e-04
8.78513038e-01 -5.17971337e-01 -2.83808470e-01 2.85328031e-01
4.33583289e-01 3.68986279e-01 -6.22282386e-01 -4.68314476e-02
2.05989286e-01 -4.79131222e-01 -2.31516391e-01 -1.16416276e+00
-1.13033345e-02 3.46137792e-01 3.75383377e-01 8.78797889e-01
7.56665826e-01 -2.02989168e-02 4.07744974e-01 4.01514381e-01
-1.20133631e-01 -1.39551556e+00 7.17706755e-02 4.94377345e-01
7.87773013e-01 -1.31662452e+00 9.47321672e-03 -8.48832905e-01
-1.02023172e+00 9.41835225e-01 9.19873595e-01 2.79675364e-01
5.28033137e-01 1.21312901e-01 5.68759263e-01 -4.19375211e-01
-7.75492013e-01 -6.26099994e-03 3.41413505e-02 7.42124200e-01
8.80841792e-01 3.29586625e-01 -1.54366624e+00 1.23403966e+00
-4.81354326e-01 -2.69239042e-02 6.73080742e-01 6.66146517e-01
-8.45534146e-01 -1.26024103e+00 -2.09639758e-01 6.03525378e-02
-2.40834922e-01 9.27853063e-02 -1.34981096e+00 5.89839637e-01
-3.13021153e-01 1.75336027e+00 -4.47769612e-01 -5.46426117e-01
4.71921355e-01 6.92615956e-02 1.91723019e-01 -5.60520530e-01
-1.07166123e+00 -1.82241097e-01 9.86958504e-01 -2.80734122e-01
-7.21807063e-01 -5.15586436e-01 -1.47632480e+00 -8.76530647e-01
-1.31850481e-01 5.89127183e-01 3.78801048e-01 1.05355763e+00
2.13770315e-01 3.52600157e-01 6.98476315e-01 8.42854455e-02
-9.35738623e-01 -1.28131640e+00 3.15732509e-02 5.31132698e-01
-1.30747542e-01 -4.30009037e-01 -2.06745416e-01 -3.20567399e-01] | [12.821542739868164, 7.874197006225586] |
ec4ec31c-666f-42c6-8cf8-d057500b3b36 | lexicon-injected-semantic-parsing-for-task | 2211.14508 | null | https://arxiv.org/abs/2211.14508v1 | https://arxiv.org/pdf/2211.14508v1.pdf | Lexicon-injected Semantic Parsing for Task-Oriented Dialog | Recently, semantic parsing using hierarchical representations for dialog systems has captured substantial attention. Task-Oriented Parse (TOP), a tree representation with intents and slots as labels of nested tree nodes, has been proposed for parsing user utterances. Previous TOP parsing methods are limited on tackling unseen dynamic slot values (e.g., new songs and locations added), which is an urgent matter for real dialog systems. To mitigate this issue, we first propose a novel span-splitting representation for span-based parser that outperforms existing methods. Then we present a novel lexicon-injected semantic parser, which collects slot labels of tree representation as a lexicon, and injects lexical features to the span representation of parser. An additional slot disambiguation technique is involved to remove inappropriate span match occurrences from the lexicon. Our best parser produces a new state-of-the-art result (87.62%) on the TOP dataset, and demonstrates its adaptability to frequently updated slot lexicon entries in real task-oriented dialog, with no need of retraining. | ['Qun Liu', 'Xin Jiang', 'Zhiyong Wu', 'Baojun Wang', 'Yasheng Wang', 'Wenlin Dai', 'Xiaojun Meng'] | 2022-11-26 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 1.86031953e-01 5.32453001e-01 -1.51380524e-01 -7.26481676e-01
-7.54284024e-01 -8.56580853e-01 2.78803200e-01 2.12922439e-01
-3.29885393e-01 7.81762719e-01 5.78766167e-01 -3.65228474e-01
1.74980566e-01 -8.45224023e-01 -1.29263535e-01 -2.26719514e-01
1.64964899e-01 9.03484762e-01 8.62741709e-01 -7.60486126e-01
1.38959184e-01 -4.39343564e-02 -1.34788334e+00 4.89883780e-01
5.66701293e-01 9.00825381e-01 4.81130123e-01 5.03365993e-01
-1.25784099e+00 6.61722779e-01 -1.02118814e+00 -3.76478225e-01
-3.77464062e-03 -3.61064821e-01 -1.23108268e+00 9.33450386e-02
-1.23128243e-01 -1.53834030e-01 -7.42752329e-02 7.83122301e-01
4.12494034e-01 2.32960418e-01 1.79877371e-01 -1.11376798e+00
-2.97092319e-01 1.09297669e+00 -9.01936516e-02 1.60343364e-01
5.69937229e-01 -4.41497505e-01 1.39822686e+00 -6.95091307e-01
6.48696959e-01 1.68239903e+00 4.81663704e-01 8.09142470e-01
-9.01026130e-01 -4.81160641e-01 5.32370329e-01 -2.03806505e-01
-9.04870510e-01 -2.97269583e-01 8.26730251e-01 -1.51933372e-01
1.27609551e+00 3.18950206e-01 1.30396470e-01 1.06843209e+00
-3.70320864e-02 7.47039139e-01 8.27611387e-01 -6.14909232e-01
3.83380532e-01 1.06289342e-01 6.59252703e-01 5.05799294e-01
-1.86767906e-01 -5.43132186e-01 -7.57898748e-01 -3.73502076e-01
4.29933876e-01 -1.55129328e-01 3.12037259e-01 -2.25950673e-01
-8.37472558e-01 1.09781146e+00 5.81259169e-02 2.00807929e-01
-9.11588147e-02 -4.01755333e-01 8.39581847e-01 1.32767588e-01
3.93542290e-01 2.62168676e-01 -8.94873619e-01 -3.06735516e-01
-3.18740666e-01 2.47735113e-01 1.20606899e+00 1.22972810e+00
6.51012719e-01 -1.01691887e-01 -2.70764500e-01 1.46067452e+00
2.03658447e-01 3.32820803e-01 8.60433578e-01 -8.68524194e-01
6.65235341e-01 1.09220219e+00 4.25114073e-02 -5.75266123e-01
-9.59349990e-01 7.43547976e-02 -2.40098208e-01 -4.78602886e-01
3.02411139e-01 -1.59945279e-01 -8.18785906e-01 2.01531005e+00
6.15974486e-01 -3.91847849e-01 4.87779588e-01 5.95340610e-01
1.24511802e+00 5.39116144e-01 7.09986806e-01 -3.07387650e-01
2.18885875e+00 -9.89145875e-01 -9.77198601e-01 -5.76489031e-01
6.47098124e-01 -7.05953419e-01 1.33464420e+00 -1.24622509e-01
-8.26395214e-01 -4.19881880e-01 -6.76229894e-01 -2.32773006e-01
-7.80461490e-01 -1.56054795e-01 8.34941864e-01 8.43195379e-01
-8.64128590e-01 -6.96610939e-03 -3.89801741e-01 -6.75513685e-01
-3.22526187e-01 4.08014804e-01 -2.68525302e-01 4.35774118e-01
-1.70825827e+00 6.21891320e-01 6.52473986e-01 -2.61489242e-01
-3.31573546e-01 -2.49049768e-01 -1.17551696e+00 1.94935009e-01
8.62272143e-01 -3.82307380e-01 1.79085946e+00 -4.01508063e-01
-1.81469274e+00 8.59875858e-01 -5.06386578e-01 -4.50859696e-01
-3.11893895e-02 -2.34644681e-01 -4.43722039e-01 -1.29813822e-02
5.38807809e-01 7.68536627e-01 6.37224317e-01 -1.01833797e+00
-8.17474246e-01 -4.45857465e-01 4.98302191e-01 5.75261295e-01
-3.31066221e-01 1.51794791e-01 -4.60273504e-01 -5.05600095e-01
6.16614223e-01 -8.61868620e-01 -2.87926793e-01 -6.89167261e-01
-5.14851511e-01 -7.07532287e-01 7.94123352e-01 -6.28115773e-01
1.55211794e+00 -2.17438626e+00 -1.19138531e-01 -2.96015948e-01
-3.33380327e-02 5.46147749e-02 6.23753667e-02 6.65076613e-01
1.75096467e-01 -3.32810357e-02 -2.81717896e-01 -5.57758510e-01
1.91457823e-01 6.46444380e-01 -6.93398714e-01 -5.75505316e-01
1.40221938e-01 7.28839219e-01 -6.93550289e-01 -7.45554566e-01
2.35096827e-01 -2.19897091e-01 -6.90662861e-01 4.48289692e-01
-6.45531118e-01 1.78385392e-01 -6.71175122e-01 8.12684000e-01
6.62818491e-01 -7.92181566e-02 7.15905368e-01 -3.86773683e-02
-1.02412803e-02 7.55971968e-01 -1.05284262e+00 1.87160707e+00
-4.95415866e-01 -1.59123708e-02 7.98175037e-02 -8.33022416e-01
1.26783931e+00 3.70153278e-01 2.76722968e-01 -8.25693548e-01
7.17307925e-02 7.89421052e-02 -2.47409478e-01 -3.45375776e-01
9.93417919e-01 -2.61745807e-02 -1.09585071e+00 2.62277246e-01
2.38620669e-01 -7.34709799e-02 2.90651739e-01 2.55162001e-01
1.19403219e+00 3.34496535e-02 5.37690699e-01 -1.89511180e-01
8.24932098e-01 3.85796338e-01 6.89056277e-01 7.05717564e-01
-3.55039269e-01 2.38203183e-01 6.55502379e-01 -5.32086790e-01
-6.51065171e-01 -1.13712227e+00 -2.48541012e-02 2.11175704e+00
3.28054249e-01 -5.86741209e-01 -1.09160197e+00 -9.72434580e-01
-2.07500204e-01 5.64026713e-01 -2.22200185e-01 1.13341399e-01
-8.33791137e-01 -6.73948944e-01 5.47566414e-01 4.69201893e-01
7.67227113e-01 -1.41772056e+00 -6.56594872e-01 7.47159243e-01
-4.97170180e-01 -1.65318692e+00 -2.10807726e-01 5.70512950e-01
-7.95111656e-01 -1.09914529e+00 -2.33117878e-01 -1.14242351e+00
2.65814573e-01 1.35785088e-01 1.28702283e+00 -1.57886654e-01
-7.43054152e-02 3.07443589e-01 -7.29838014e-01 -1.73417345e-01
-4.73993540e-01 5.08360445e-01 -6.13277406e-02 -3.53800833e-01
7.76166916e-01 -2.27289230e-01 -2.92225182e-01 5.99116325e-01
-7.15601742e-01 -2.34806887e-03 2.28676200e-01 1.06425273e+00
2.33930111e-01 -1.45378858e-01 1.01572967e+00 -1.37562883e+00
1.03132713e+00 -3.01748693e-01 -5.23170412e-01 3.22944403e-01
-2.09627792e-01 2.87053943e-01 7.60546148e-01 -1.97011113e-01
-1.55881882e+00 2.48537183e-01 -5.06480515e-01 3.44492614e-01
-3.61092776e-01 3.17798227e-01 -3.53788465e-01 4.69277501e-01
5.46812296e-01 2.00542733e-01 -1.33929595e-01 -7.53801048e-01
4.80409384e-01 9.79551077e-01 5.47406375e-01 -9.04126942e-01
2.35685453e-01 1.52598768e-01 -3.96883100e-01 -7.60802507e-01
-9.89724338e-01 -7.84782350e-01 -7.47861326e-01 1.67551592e-01
9.36648846e-01 -8.41203451e-01 -7.13078737e-01 2.36040130e-01
-1.27741718e+00 -3.06110799e-01 -2.31413499e-01 -5.12997657e-02
-5.89724243e-01 4.34082359e-01 -7.78506815e-01 -9.34456110e-01
-4.96391326e-01 -1.10181832e+00 1.35610807e+00 3.64147812e-01
-5.31573594e-01 -9.27177131e-01 -5.22324629e-02 6.34944618e-01
4.25997555e-01 -2.84602612e-01 1.21468842e+00 -1.26484978e+00
-2.29293242e-01 2.05094248e-01 -1.93510041e-01 -1.47971839e-01
2.34929994e-01 -7.88324893e-01 -1.04044509e+00 5.44447191e-02
4.71519902e-02 -3.79618347e-01 4.53555763e-01 -3.73887010e-02
7.59019554e-01 -5.69605008e-02 -3.26488465e-01 2.77388934e-02
8.37650955e-01 7.05318987e-01 1.89561889e-01 4.80002105e-01
2.11624742e-01 9.20155585e-01 9.86167312e-01 6.20483696e-01
6.48407698e-01 9.11307216e-01 1.64400145e-01 2.08680540e-01
2.62295804e-03 -5.12772918e-01 2.18563676e-01 8.52766693e-01
7.49604523e-01 -2.51346976e-01 -7.63694346e-01 4.71977741e-01
-1.97486436e+00 -3.56118083e-01 1.97544187e-01 1.79661179e+00
9.78555679e-01 4.11203474e-01 1.71952769e-01 -1.55655891e-01
8.39489877e-01 3.94260675e-01 -6.59353971e-01 -6.14404917e-01
1.94984544e-02 2.60343432e-01 1.91829577e-01 7.13985562e-01
-1.27616692e+00 1.94421101e+00 6.61977816e+00 6.75857782e-01
-7.19955862e-01 3.17099661e-01 3.06024194e-01 5.13702929e-01
-1.52826875e-01 1.38244882e-01 -1.30549920e+00 3.45257759e-01
9.54872310e-01 -2.03184951e-02 1.44800618e-01 1.28777206e+00
-4.13651094e-02 -1.31520152e-01 -6.43684089e-01 8.38149428e-01
-1.54723898e-01 -9.29005861e-01 1.50580287e-01 -3.86090547e-01
-7.24141002e-02 -3.60284269e-01 -2.55389750e-01 9.43781972e-01
5.86592317e-01 -5.71218073e-01 3.77626747e-01 -2.42129996e-01
5.75287640e-01 -3.32218885e-01 7.45070636e-01 2.85769880e-01
-1.54567873e+00 -1.49395525e-01 -5.52180707e-01 -1.56935453e-01
5.15110314e-01 1.77729711e-01 -1.28326237e+00 3.76513094e-01
5.95012307e-01 1.21045224e-01 -4.75053698e-01 1.75329000e-01
-1.73556626e-01 4.54669893e-01 -3.66845220e-01 -1.90926343e-01
2.07150638e-01 1.14458650e-01 4.13185924e-01 1.31630528e+00
3.02267615e-02 2.48663381e-01 5.83896756e-01 3.92014593e-01
4.37997766e-02 4.04994816e-01 -4.32582259e-01 9.34281722e-02
9.35949147e-01 1.14669168e+00 -1.01701546e+00 -4.36683536e-01
-4.00227875e-01 9.72208738e-01 4.48099941e-01 1.09955989e-01
-6.00096226e-01 -2.57308125e-01 7.43642390e-01 -2.06132859e-01
9.32109579e-02 -3.32661659e-01 -2.74873555e-01 -9.41262126e-01
-9.28281695e-02 -7.06040144e-01 1.00429773e+00 -5.54551423e-01
-1.04904163e+00 9.72113013e-01 2.14392066e-01 -9.38753128e-01
-5.96393883e-01 -7.93629646e-01 -4.48172182e-01 5.38478792e-01
-1.33145297e+00 -9.25184906e-01 -2.60482192e-01 5.50304711e-01
1.38806438e+00 -4.28266257e-01 1.30183101e+00 8.54889899e-02
-5.34590662e-01 6.04815423e-01 -3.89150470e-01 2.03118846e-01
6.43196702e-01 -1.36850917e+00 6.92939162e-01 4.59387243e-01
-8.19280371e-02 6.39996946e-01 6.82598829e-01 -7.57650316e-01
-1.03609467e+00 -7.41186917e-01 1.21357954e+00 -4.47226495e-01
6.31055474e-01 -7.01206744e-01 -8.73992026e-01 7.19421089e-01
1.18277110e-01 -4.24043924e-01 8.59013259e-01 4.59908783e-01
-4.21660364e-01 1.99969798e-01 -1.19922364e+00 5.51287353e-01
1.36572850e+00 -4.17182595e-01 -1.18639028e+00 2.90143311e-01
1.40638030e+00 -8.46868336e-01 -5.97579718e-01 3.31688821e-01
2.69819140e-01 -9.15818155e-01 9.02418554e-01 -6.02635860e-01
-1.64618462e-01 -9.66062471e-02 -3.95418823e-01 -9.24469590e-01
6.45883977e-02 -6.00629032e-01 1.02280281e-01 1.37257278e+00
4.85573053e-01 -6.40570879e-01 1.03475463e+00 7.71538019e-01
-3.75074655e-01 -4.11964953e-01 -1.17598784e+00 -6.56490147e-01
-2.17748180e-01 -3.06238532e-01 7.77059674e-01 6.36059165e-01
4.90606070e-01 8.88948560e-01 -3.25717151e-01 1.12153791e-01
1.72951937e-01 2.46394262e-01 7.65196443e-01 -1.35742819e+00
-4.30158265e-02 -1.48970455e-01 -1.60635874e-01 -1.44969356e+00
4.80287105e-01 -6.23916924e-01 2.45055750e-01 -1.48298621e+00
-1.29166678e-01 -6.86534882e-01 -1.25145596e-02 6.89058006e-01
-2.34008774e-01 -3.12998593e-01 2.22103208e-01 1.90996483e-01
-8.83998513e-01 5.36771178e-01 7.35237658e-01 -5.39068691e-02
-6.47375524e-01 3.88873704e-02 -6.88362956e-01 7.66506076e-01
9.67059672e-01 -4.24888492e-01 -5.82193434e-01 -1.09526068e-01
-1.39959782e-01 3.00049990e-01 -3.17197204e-01 -8.20762932e-01
3.91493797e-01 -5.01483120e-02 -2.92987674e-01 -6.86991930e-01
7.23229647e-01 -8.27128351e-01 -2.23134771e-01 1.99873194e-01
-5.32658935e-01 2.41326243e-01 3.57684821e-01 4.51071829e-01
-4.44707572e-01 -5.71274102e-01 4.25254971e-01 -4.43610072e-01
-1.21686137e+00 -1.70465663e-01 -3.90765250e-01 4.10560250e-01
7.92327881e-01 -4.94314693e-02 -5.17474115e-01 -1.60791159e-01
-1.01705205e+00 3.58834952e-01 -1.79734439e-01 8.91064465e-01
3.44950229e-01 -1.02754569e+00 4.86427872e-03 3.65221798e-01
2.47357950e-01 1.41796514e-01 3.66443068e-01 -6.27572164e-02
-3.40211034e-01 6.69627428e-01 -9.00732651e-02 -5.09166241e-01
-1.30520332e+00 3.86901796e-01 1.13729089e-01 -6.73780084e-01
-6.49632692e-01 7.89079428e-01 3.09694350e-01 -9.43040848e-01
5.02287447e-01 -3.28172535e-01 -5.49537778e-01 3.36178243e-01
3.96248996e-01 -2.05281407e-01 8.98877606e-02 -6.89308643e-01
-4.60568964e-01 3.38241577e-01 -3.01212758e-01 -2.88985610e-01
7.96528578e-01 -4.42388982e-01 3.61238839e-03 5.00084519e-01
6.79897070e-01 2.55492795e-03 -8.89449179e-01 -5.66526353e-01
6.98125303e-01 -1.49409369e-01 -6.64679646e-01 -7.77416050e-01
-3.53619009e-01 6.73298836e-01 3.39218438e-01 6.71129227e-01
7.86547780e-01 2.83345431e-01 1.22927403e+00 6.28469288e-01
7.67461777e-01 -1.28166044e+00 1.80771098e-01 1.13198602e+00
6.60894156e-01 -1.25621474e+00 -5.00304818e-01 -8.67550671e-01
-8.68048668e-01 1.06144059e+00 9.49448645e-01 4.83263642e-01
5.13995647e-01 1.98606178e-01 4.85725522e-01 -2.22039431e-01
-6.86663866e-01 -4.29042727e-01 -3.84961456e-01 5.78717589e-01
4.75722045e-01 1.79821804e-01 -5.43173313e-01 1.07985687e+00
-6.55632436e-01 -5.32476783e-01 1.41169503e-01 1.19294071e+00
-1.02749825e+00 -1.52214158e+00 -2.74032384e-01 8.05850998e-02
-3.98046196e-01 -2.54406661e-01 -5.63614547e-01 7.57314742e-01
-2.47864276e-01 1.16768718e+00 1.21171311e-01 -3.06904376e-01
7.27250814e-01 8.52629483e-01 -2.21655756e-01 -1.11008275e+00
-6.60723746e-01 -6.93879556e-03 4.67181921e-01 -4.54257071e-01
-1.96697250e-01 -3.31536740e-01 -1.66484642e+00 1.99782088e-01
-1.67049095e-01 5.69787920e-01 5.31646013e-01 8.72733176e-01
4.32683885e-01 5.14509380e-01 4.83129710e-01 -4.92078602e-01
-5.59236825e-01 -1.00576806e+00 -5.29318333e-01 5.03505468e-01
-1.38075501e-01 -9.15373862e-01 -1.64923579e-01 -1.64102837e-01] | [12.624560356140137, 7.509890079498291] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.