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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
c90bfff7-3748-4b80-bc51-8480771e500f | recurrent-neural-networks-with-stochastic | 1902.0498 | null | http://arxiv.org/abs/1902.04980v1 | http://arxiv.org/pdf/1902.04980v1.pdf | Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection | In this paper, we adapt Recurrent Neural Networks with Stochastic Layers,
which are the state-of-the-art for generating text, music and speech, to the
problem of acoustic novelty detection. By integrating uncertainty into the
hidden states, this type of network is able to learn the distribution of
complex sequences. Because the learned distribution can be calculated
explicitly in terms of probability, we can evaluate how likely an observation
is then detect low-probability events as novel. The model is robust, highly
unsupervised, end-to-end and requires minimum preprocessing, feature
engineering or hyperparameter tuning. An experiment on a benchmark dataset
shows that our model outperforms the state-of-the-art acoustic novelty
detectors. | ['Fábio Frazão', 'Duong Nguyen', 'Stan Matwin', 'Ronan Fablet', 'Oliver S. Kirsebom'] | 2019-02-13 | null | null | null | null | ['acoustic-novelty-detection'] | ['audio'] | [ 8.86067748e-02 -5.96808232e-02 1.53739169e-01 -2.58314848e-01
-8.67591023e-01 -4.85033929e-01 5.77742994e-01 -7.91311264e-04
-4.94210631e-01 7.57452250e-01 8.89391303e-02 -1.83531553e-01
-3.48518267e-02 -7.61107802e-01 -9.34930384e-01 -5.70879698e-01
-2.81516284e-01 5.27775407e-01 4.31497008e-01 6.90920353e-02
1.19956769e-01 4.14647162e-01 -1.89402223e+00 2.42432937e-01
3.79692227e-01 1.11374807e+00 2.23098546e-02 1.16105425e+00
-3.21089923e-01 5.13625205e-01 -8.89906466e-01 -2.02664539e-01
-1.01983987e-01 -6.16075099e-01 -3.73965770e-01 -3.95745069e-01
2.90854514e-01 2.01265700e-02 -2.02546313e-01 1.18852937e+00
8.46835613e-01 4.63260293e-01 7.31017053e-01 -9.22484815e-01
-5.23506343e-01 1.23567605e+00 1.65083110e-01 3.52601111e-01
4.05238479e-01 9.21022147e-02 9.86157477e-01 -1.22246277e+00
5.18219531e-01 1.14839375e+00 8.92396629e-01 6.65604591e-01
-1.30597794e+00 -6.61953568e-01 1.81190163e-01 3.01779896e-01
-1.41566241e+00 -6.63224339e-01 9.77631271e-01 -2.94330448e-01
1.02926207e+00 3.18985730e-01 4.39781159e-01 1.65023398e+00
1.42643541e-01 9.80888724e-01 5.23517907e-01 -6.29872203e-01
8.19052398e-01 2.81958170e-02 -2.38236427e-01 6.59452260e-01
6.54513314e-02 4.15447384e-01 -1.04642534e+00 -3.18476528e-01
3.76677185e-01 -1.33894727e-01 -6.95852842e-03 -9.64701846e-02
-1.24773765e+00 6.95318818e-01 -8.37798044e-02 5.88094592e-01
-4.83076692e-01 5.46848714e-01 5.07976353e-01 2.40388587e-01
4.61042702e-01 3.27528000e-01 -4.11422253e-01 -6.58212781e-01
-1.26933885e+00 4.22105283e-01 9.94658530e-01 6.30182564e-01
4.72958565e-01 7.36188412e-01 -3.75263393e-01 7.70007491e-01
2.03991532e-01 6.66275442e-01 1.01191449e+00 -6.75544262e-01
1.28789525e-02 2.18849815e-02 -6.17204905e-02 -6.74280167e-01
-4.16876167e-01 -6.16820633e-01 -9.00526166e-01 2.53281385e-01
1.00122705e-01 -3.12089205e-01 -1.10375369e+00 1.91368854e+00
3.43390442e-02 7.55557418e-01 2.14376628e-01 3.89581114e-01
6.82299435e-01 8.73177469e-01 -3.58709306e-01 -2.73296356e-01
8.50833654e-01 -7.00530589e-01 -9.89958823e-01 -1.00728795e-01
3.85606512e-02 -6.98551416e-01 1.01795781e+00 6.06869638e-01
-1.07728195e+00 -7.60764241e-01 -1.09104323e+00 5.39208353e-01
-5.53955138e-01 1.05089813e-01 3.53164852e-01 8.68232608e-01
-9.26961601e-01 9.45711195e-01 -8.67770076e-01 -2.13643521e-01
1.94739476e-01 8.90728608e-02 1.61625609e-01 6.04763925e-01
-1.39844871e+00 6.00238562e-01 8.45681071e-01 -8.15444812e-02
-1.18124497e+00 -6.38063371e-01 -6.52907789e-01 3.27893585e-01
4.24190491e-01 -5.99478900e-01 1.57481503e+00 -5.75580239e-01
-2.18047190e+00 2.32522488e-01 -2.01673165e-01 -8.11869383e-01
2.17908546e-01 -2.53488243e-01 -7.78282285e-01 -2.06775870e-02
-2.53964961e-01 3.47726285e-01 1.29836965e+00 -8.62344325e-01
-6.67563140e-01 3.53112996e-01 -7.61866450e-01 -2.87159383e-01
-3.37614030e-01 -8.26859847e-02 -2.66196337e-02 -9.74036872e-01
8.51338804e-02 -8.68575871e-01 -1.73411846e-01 -3.30147922e-01
-5.09838820e-01 -4.46698010e-01 7.07077265e-01 -2.64491349e-01
1.38697040e+00 -2.17483687e+00 1.10971399e-01 4.38005745e-01
-2.31469631e-01 1.36897013e-01 6.51589409e-02 4.92885828e-01
1.43476715e-02 1.03927270e-01 -4.36043382e-01 -6.09545469e-01
4.37975615e-01 2.02876776e-01 -7.99465179e-01 1.77021575e-04
2.63619393e-01 8.45071316e-01 -1.02589166e+00 -1.82873428e-01
1.34444281e-01 3.06127608e-01 -3.88011098e-01 3.51365745e-01
-5.93140066e-01 1.35757670e-01 -2.90173031e-02 4.15562451e-01
1.22670911e-01 1.08543880e-01 -1.97196856e-01 2.29448363e-01
-3.94918174e-02 5.40725052e-01 -1.56773210e+00 1.86867881e+00
-3.70058268e-01 7.17989504e-01 -5.81168234e-01 -8.46225441e-01
1.08151841e+00 5.04390121e-01 9.28731561e-02 6.35878146e-02
7.22312555e-02 3.98005784e-01 -7.59036839e-02 -2.24268898e-01
5.68962038e-01 -1.89378962e-01 -2.01246947e-01 6.41840100e-01
5.39076686e-01 -2.50681996e-01 2.99603730e-01 -1.55207552e-02
1.24837732e+00 -1.19481897e-02 1.09376743e-01 1.67557254e-01
4.36246574e-01 -6.32635057e-01 6.47983074e-01 1.31969285e+00
1.37197480e-01 7.17245579e-01 2.27310181e-01 -4.43925947e-01
-1.00538576e+00 -1.62851667e+00 7.83148855e-02 9.69102025e-01
-5.61320126e-01 -4.73755747e-01 -6.73657715e-01 -4.93989289e-01
-6.13006987e-02 1.03565681e+00 -3.98202389e-01 -4.79285747e-01
-4.48103994e-01 -5.12199700e-01 9.61330891e-01 3.46251935e-01
-3.30990925e-03 -1.56537819e+00 -5.03943145e-01 7.48525858e-01
-1.65009685e-02 -9.61390734e-01 -3.56968343e-01 4.73298728e-01
-6.96790218e-01 -2.88991451e-01 -5.83782852e-01 -5.71905255e-01
1.48628419e-02 -5.82230687e-01 1.15760040e+00 -5.41855395e-01
-3.35671484e-01 4.13848728e-01 -4.62564558e-01 -8.09982657e-01
-8.71931612e-01 1.90287411e-01 5.98736405e-01 1.38866156e-01
2.46330485e-01 -9.55454290e-01 -1.44407749e-01 -1.36563197e-01
-1.18642330e+00 -5.06614745e-01 5.29105127e-01 8.99128675e-01
6.96608961e-01 4.22995269e-01 8.50133240e-01 -6.09936774e-01
8.68652880e-01 -2.63106227e-01 -4.53491986e-01 5.25462441e-02
-5.27196467e-01 4.78538454e-01 6.77193224e-01 -9.27707970e-01
-8.15498590e-01 1.35738567e-01 -2.84992367e-01 -6.54831231e-01
-3.47376168e-01 7.04297364e-01 8.13909918e-02 2.98336118e-01
8.26863289e-01 4.76219893e-01 -3.70025009e-01 -5.01563966e-01
5.72598934e-01 4.29005414e-01 9.02942836e-01 -5.01610398e-01
8.08191180e-01 3.60213280e-01 -1.01809032e-01 -9.21567559e-01
-8.56650889e-01 -3.44962984e-01 -5.21616042e-01 -2.73660153e-01
4.21177745e-01 -5.99498987e-01 -5.91413736e-01 4.82347965e-01
-1.28621209e+00 -2.66117044e-02 -9.92889822e-01 6.71983302e-01
-7.54273534e-01 2.91366607e-01 -3.28287989e-01 -1.36689150e+00
-5.53917944e-01 -6.94871306e-01 8.72803509e-01 2.28663296e-01
-4.25177991e-01 -8.55730414e-01 3.93436790e-01 -6.35867953e-01
6.26897693e-01 1.49349704e-01 7.47779369e-01 -8.63893390e-01
-4.91236031e-01 -3.50913852e-01 6.63872540e-01 5.85497558e-01
7.32790902e-02 2.26130769e-01 -1.32501578e+00 1.08816577e-02
4.92781177e-02 -1.29658848e-01 1.16554534e+00 5.05577683e-01
1.26183617e+00 -4.03889090e-01 -1.71056733e-01 3.80791217e-01
8.45348179e-01 9.94615108e-02 5.72550118e-01 -2.27623619e-02
3.99527848e-01 1.25796974e-01 1.51653528e-01 7.50095010e-01
-2.30343565e-02 4.55740541e-01 2.00545967e-01 3.78558487e-01
-1.66263040e-02 -5.33866048e-01 8.34408522e-01 1.27719510e+00
4.45562601e-01 -4.15820688e-01 -8.10626388e-01 6.94605827e-01
-1.98048329e+00 -1.42098892e+00 5.07522285e-01 2.28146553e+00
1.14900374e+00 6.80580080e-01 2.89782971e-01 5.53639531e-01
6.45082057e-01 1.55338049e-01 -6.29003644e-01 -5.71059287e-01
-2.06855267e-01 5.69840252e-01 1.10129997e-01 3.22165608e-01
-9.31300581e-01 1.05943942e+00 7.04030037e+00 1.10225558e+00
-1.06877553e+00 -5.76517098e-02 2.41859674e-01 -3.82077038e-01
-2.71630734e-01 -2.12520331e-01 -8.45963836e-01 4.94141549e-01
1.58844697e+00 -1.56011671e-01 4.84615475e-01 7.72111893e-01
-3.32814418e-02 2.39617795e-01 -1.32105553e+00 9.49551761e-01
1.48118854e-01 -1.49723756e+00 2.43138745e-01 -4.54928190e-01
6.41116440e-01 9.16804075e-02 2.99276590e-01 4.88555372e-01
4.45771903e-01 -8.85369480e-01 9.34780598e-01 1.07006204e+00
4.71819550e-01 -1.02820516e+00 6.61381245e-01 4.63226974e-01
-9.57298458e-01 2.39062142e-02 -4.73898828e-01 1.22707948e-01
3.88716906e-01 1.09712100e+00 -1.22835720e+00 2.20786512e-01
9.03457344e-01 4.82640207e-01 -2.73459703e-01 1.05452859e+00
-3.80170107e-01 1.09036469e+00 -6.60936952e-01 -5.65695167e-01
2.00891942e-01 2.57788897e-01 1.09532654e+00 1.45419371e+00
6.41591311e-01 -4.44516033e-01 2.23151147e-01 1.24405432e+00
-1.01226091e-01 -1.30249366e-01 -5.16716182e-01 -3.94703031e-01
5.19707739e-01 8.84061694e-01 -5.30910313e-01 -3.49887013e-01
1.92850247e-01 9.51247215e-01 1.80066913e-01 4.01251704e-01
-5.09000421e-01 -7.92311609e-01 5.75970888e-01 -3.34280580e-01
7.72269428e-01 -2.59236932e-01 -1.79654919e-02 -1.00547731e+00
1.76458180e-01 -6.96057439e-01 1.65858552e-01 -5.94645739e-01
-1.52753353e+00 7.81931341e-01 -2.47140110e-01 -1.16562998e+00
-9.36961055e-01 -5.56123793e-01 -8.48277211e-01 6.11173451e-01
-1.25561237e+00 -5.98384857e-01 3.11812699e-01 2.61760533e-01
4.60086465e-01 -5.50468922e-01 1.18128467e+00 -5.64591624e-02
-2.97513753e-01 9.16956365e-01 2.43617579e-01 1.31040521e-03
4.28669930e-01 -1.41851199e+00 8.63511264e-01 9.02580321e-01
8.48761261e-01 4.35549200e-01 9.83158112e-01 -6.22163236e-01
-1.03706419e+00 -9.11024570e-01 9.41409945e-01 -4.09768999e-01
8.37499976e-01 -7.25117385e-01 -1.03829670e+00 2.84667522e-01
-5.85314222e-02 -1.33883012e-02 7.01609850e-01 2.57888913e-01
-4.92541581e-01 -1.67744737e-02 -9.68445241e-01 6.46437645e-01
9.72026229e-01 -7.43187249e-01 -7.68280387e-01 2.04004332e-01
1.21248651e+00 -4.32343692e-01 -5.98937094e-01 3.47310334e-01
5.65070689e-01 -8.83768559e-01 9.39947248e-01 -6.40947282e-01
-4.37631495e-02 -3.85982543e-01 -1.60146266e-01 -1.45185328e+00
-1.55444533e-01 -1.29984558e+00 -7.58120060e-01 1.26931977e+00
8.02046359e-01 -3.48253191e-01 7.43964851e-01 -4.56550606e-02
-2.49949485e-01 -5.46770036e-01 -1.46214974e+00 -1.17539275e+00
-1.93884701e-01 -1.07762027e+00 7.38144159e-01 5.25503516e-01
-1.85714647e-01 2.12749347e-01 -5.80309689e-01 3.46374571e-01
3.97561729e-01 -1.54705718e-01 5.73927760e-01 -1.38845909e+00
-5.65833867e-01 -6.17203712e-01 -6.59591556e-01 -6.82771623e-01
3.45735043e-01 -7.35804737e-01 3.71259272e-01 -1.04416895e+00
-4.18283165e-01 -9.24960058e-03 -7.80720949e-01 4.11807448e-01
-5.23289815e-02 -6.39238432e-02 4.63142991e-02 -2.08991304e-01
-6.54046655e-01 8.35003674e-01 3.05201590e-01 -1.68372616e-01
-4.43093538e-01 3.22195649e-01 -1.60467058e-01 7.90917456e-01
9.84603107e-01 -1.00073528e+00 -1.67355835e-01 1.78532153e-01
5.41670859e-01 -3.90850365e-01 2.57016003e-01 -1.42739975e+00
4.67358142e-01 1.85856491e-01 4.16452408e-01 -1.07395911e+00
4.99065429e-01 -5.69601834e-01 2.25397553e-02 3.84404659e-01
-6.45037055e-01 -1.75852012e-02 3.08447778e-01 8.24689388e-01
-3.16925168e-01 -5.55505216e-01 5.69552720e-01 -7.04039410e-02
-3.81025761e-01 1.38661236e-01 -8.20483863e-01 2.46325612e-01
4.51878279e-01 1.65767670e-02 3.22813094e-01 -6.43414319e-01
-6.94413126e-01 -3.29729021e-01 -1.58448800e-01 6.52983785e-01
8.21169436e-01 -1.48447263e+00 -8.85105848e-01 3.12770098e-01
1.45382196e-01 -1.25425354e-01 2.09788173e-01 1.46757826e-01
3.31285261e-02 1.89899951e-01 3.13061565e-01 -8.33805740e-01
-8.55048478e-01 4.67269361e-01 2.13163823e-01 -2.08947867e-01
-5.37346482e-01 9.68617857e-01 -4.57323551e-01 -6.42417073e-01
7.52945542e-01 -6.22628510e-01 1.43783793e-01 -1.11103140e-01
7.21040070e-01 3.69082093e-01 4.26940173e-02 -1.82254240e-01
-4.17059064e-01 3.24915022e-01 -8.15726966e-02 -6.80021286e-01
1.23297536e+00 9.91249457e-02 1.65685534e-01 1.36513281e+00
9.14190650e-01 -1.73907615e-02 -8.53667080e-01 -6.09762549e-01
3.51157755e-01 -3.45961824e-02 6.56391447e-03 -8.45253766e-01
-5.66669941e-01 8.69830549e-01 8.89437973e-01 3.69393319e-01
8.40602219e-01 -6.34024143e-02 8.78106654e-01 1.01035547e+00
1.71336666e-01 -1.37476361e+00 2.97113419e-01 1.03205812e+00
8.65487635e-01 -9.80721354e-01 -2.64690548e-01 3.00900996e-01
-4.37305838e-01 1.11288559e+00 1.12587862e-01 -2.04829365e-01
8.69797289e-01 3.85429263e-01 -1.70959346e-02 2.13002563e-01
-9.82404113e-01 -1.92234546e-01 6.36263490e-01 5.18921673e-01
1.75121978e-01 7.47478679e-02 2.01176643e-01 6.02332115e-01
-6.85338259e-01 -1.98884219e-01 5.99284589e-01 8.03352892e-01
-7.05228567e-01 -9.41513360e-01 -3.17086458e-01 3.34290743e-01
-5.87668240e-01 -3.07993561e-01 -3.13077331e-01 2.12226167e-01
1.51627269e-02 8.87161553e-01 -1.05741173e-02 -4.67513829e-01
4.71267253e-01 6.86904430e-01 8.08246583e-02 -7.22115993e-01
-6.17589951e-01 1.33631870e-01 -1.11755386e-01 -2.87007868e-01
-1.51017472e-01 -7.67263949e-01 -1.29034078e+00 1.80449218e-01
-4.61268276e-01 2.76694775e-01 1.02232432e+00 7.80997336e-01
5.40930808e-01 8.46187055e-01 6.09052539e-01 -7.87505567e-01
-8.56517732e-01 -1.25552154e+00 -6.66351020e-01 1.15602858e-01
5.16209245e-01 -3.96146119e-01 -7.35932946e-01 6.73518181e-02] | [15.585981369018555, 5.709064483642578] |
bfe9263b-45e6-468c-9a75-25cf2db75b5d | from-point-forecasts-to-multivariate | 2204.10154 | null | https://arxiv.org/abs/2204.10154v1 | https://arxiv.org/pdf/2204.10154v1.pdf | From point forecasts to multivariate probabilistic forecasts: The Schaake shuffle for day-ahead electricity price forecasting | Modeling price risks is crucial for economic decision making in energy markets. Besides the risk of a single price, the dependence structure of multiple prices is often relevant. We therefore propose a generic and easy-to-implement method for creating multivariate probabilistic forecasts based on univariate point forecasts of day-ahead electricity prices. While each univariate point forecast refers to one of the day's 24 hours, the multivariate forecast distribution models dependencies across hours. The proposed method is based on simple copula techniques and an optional time series component. We illustrate the method for five benchmark data sets recently provided by Lago et al. (2020). Furthermore, we demonstrate an example for constructing realistic prediction intervals for the weighted sum of consecutive electricity prices, as, e.g., needed for pricing individual load profiles. | ['Fabian Krüger', 'Fabian Kächele', 'Oliver Grothe'] | 2022-04-21 | null | null | null | null | ['prediction-intervals'] | ['miscellaneous'] | [-2.84448445e-01 -1.73761532e-01 1.09539889e-01 -2.16637850e-01
-7.13954389e-01 -7.87860930e-01 7.80048072e-01 3.94551396e-01
-5.38402051e-03 1.08428192e+00 -1.31158322e-01 -6.50143266e-01
-4.68049526e-01 -1.16954362e+00 -5.35716891e-01 -7.81506777e-01
-4.99078929e-01 6.30511284e-01 -3.17954361e-01 1.31447881e-01
-1.21216308e-02 5.47909498e-01 -1.35331798e+00 -3.21964562e-01
9.84192073e-01 1.22934151e+00 2.88131416e-01 4.59319055e-01
5.82720414e-02 -1.31376414e-02 -6.83117568e-01 -4.60711777e-01
3.18042904e-01 -1.22930601e-01 2.07830071e-01 -2.20100671e-01
-9.30932224e-01 -2.50598609e-01 5.17708123e-01 9.95661557e-01
2.81558216e-01 8.24480131e-02 9.12377000e-01 -1.62604547e+00
-3.63589585e-01 9.63937163e-01 -5.57020068e-01 1.23773932e-01
-2.56861951e-02 -2.79504538e-01 9.26949501e-01 -6.05068564e-01
-2.78247535e-01 6.31523788e-01 5.34678280e-01 -2.87210524e-01
-1.40169549e+00 -3.18365902e-01 -1.47968322e-01 7.78621212e-02
-1.27265298e+00 9.48395133e-02 9.26488876e-01 -5.43523431e-01
9.70349491e-01 4.67265308e-01 6.07890904e-01 7.69705832e-01
7.40987659e-01 3.07682693e-01 1.23128700e+00 -3.53260428e-01
5.98012149e-01 9.28228814e-03 -8.78870636e-02 -5.29471993e-01
5.29028177e-01 3.71143103e-01 2.63093203e-01 -6.74403846e-01
1.26250029e-01 5.21288402e-02 -2.50194315e-02 -1.53938726e-01
-8.50043595e-01 8.42800438e-01 -2.36460969e-01 2.52182305e-01
-9.01045620e-01 -4.70347190e-03 1.41022265e-01 -3.36375833e-02
6.93835795e-01 -3.00170571e-01 -8.54771793e-01 -2.12426856e-01
-1.21340525e+00 2.86088228e-01 9.58345294e-01 8.72454047e-01
3.88791829e-01 2.15527922e-01 -2.15207294e-01 3.36653709e-01
3.05389583e-01 9.50934768e-01 1.91745266e-01 -3.51100594e-01
2.62349457e-01 -2.82614291e-01 8.30488324e-01 -5.52136421e-01
-4.67343628e-01 -5.94597518e-01 -8.72872949e-01 2.41444826e-01
4.12611663e-01 -5.36741555e-01 -3.45063508e-01 1.39252114e+00
-8.76348764e-02 2.77810752e-01 7.35855028e-02 8.21815580e-02
-3.08737606e-01 9.19841886e-01 1.70411795e-01 -8.11516404e-01
1.19921482e+00 -2.99698740e-01 -7.23465025e-01 4.40088183e-01
2.30581850e-01 -8.16424370e-01 9.86792818e-02 5.67592025e-01
-1.16592693e+00 -2.22072408e-01 -5.73867500e-01 7.47626126e-01
-6.03299439e-01 6.05099313e-02 1.42372891e-01 8.23416114e-01
-8.13792408e-01 8.67954731e-01 -7.32475638e-01 3.14500868e-01
-3.49246562e-01 -3.51749882e-02 4.48093593e-01 7.26524711e-01
-1.44118571e+00 1.03235495e+00 5.17227829e-01 3.54636133e-01
-3.21494401e-01 -7.44263232e-01 -6.19492054e-01 7.54405618e-01
2.35006567e-02 -3.72475058e-01 1.36002755e+00 -2.76952773e-01
-1.37723887e+00 -2.71396399e-01 -1.16118394e-01 -8.50837231e-01
6.48468196e-01 1.40424564e-01 -8.18457425e-01 -2.19758227e-01
-8.08371790e-03 -3.42660457e-01 7.90718198e-01 -9.79636073e-01
-7.30608284e-01 -1.29893800e-04 -3.32089722e-01 -1.35827616e-01
2.91903168e-01 1.99196339e-01 4.07456577e-01 -8.84350657e-01
-3.73687208e-01 -6.96223378e-01 -3.41356516e-01 -9.15511012e-01
-3.79980564e-01 -3.17595840e-01 3.56536210e-01 -1.05946612e+00
1.31253695e+00 -1.95338213e+00 -1.97805360e-01 8.50404143e-01
-6.13422513e-01 -5.42436242e-01 4.77298915e-01 8.95099044e-01
-6.25834584e-01 2.52565950e-01 -3.33527297e-01 -3.50508839e-01
7.68676043e-01 2.91298211e-01 -6.56134486e-01 3.77353460e-01
1.99012563e-01 8.55258763e-01 -5.80346882e-01 3.16107899e-01
5.26262105e-01 3.86349112e-01 2.16173500e-01 -2.12575004e-01
-2.49092221e-01 7.30204806e-02 -1.81905150e-01 3.46850455e-01
1.03007436e+00 -6.80142790e-02 1.86019212e-01 7.73820505e-02
-5.72919905e-01 -8.01718682e-02 -1.20498741e+00 9.08959985e-01
-5.35340607e-01 -5.95419928e-02 -2.57652193e-01 -8.88669193e-01
8.93509209e-01 4.74268347e-01 6.84876382e-01 -3.57784212e-01
6.34437948e-02 3.54390532e-01 -8.30743089e-02 3.62144172e-01
3.67888659e-01 -3.95409465e-01 -2.94390142e-01 6.13591850e-01
-3.36431712e-01 -2.38725588e-01 5.52688956e-01 -4.32694614e-01
4.97092247e-01 -5.22752889e-02 5.74095428e-01 -5.83601356e-01
3.33636671e-01 -5.59531927e-01 7.17567325e-01 2.72597194e-01
1.46710366e-01 3.33875835e-01 7.59750724e-01 -1.72093138e-01
-9.95105445e-01 -1.22605467e+00 -5.37581682e-01 3.90011579e-01
-4.88051653e-01 -1.82054117e-01 -3.99354935e-01 -2.10839465e-01
3.56728554e-01 1.81216300e+00 -5.87051094e-01 5.00856400e-01
1.82675958e-01 -1.28787863e+00 -2.93715149e-01 4.09369648e-01
-9.80659500e-02 -7.55461693e-01 -7.30461121e-01 6.56128764e-01
1.18997969e-01 -7.80611157e-01 -1.60561368e-01 4.59005266e-01
-4.93502527e-01 -7.52650082e-01 -1.00554538e+00 7.36648291e-02
4.12559032e-01 -2.56577939e-01 1.28866124e+00 -5.76153398e-01
1.77054614e-01 4.21158016e-01 -2.25181624e-01 -1.00269985e+00
-2.00287580e-01 -4.51379210e-01 -1.88828614e-02 2.19964478e-02
1.67160794e-01 -8.07869732e-01 -5.91088116e-01 1.45546824e-01
-6.98333085e-01 -1.26093328e-01 1.85459226e-01 7.21497536e-01
7.15976596e-01 7.15937912e-01 8.61563206e-01 -3.90568435e-01
8.71423841e-01 -8.34438980e-01 -1.44166982e+00 5.53198516e-01
-9.33467269e-01 -1.26927018e-01 7.02399611e-01 1.59495734e-02
-1.14717996e+00 2.55980194e-02 1.31386593e-01 -6.25910014e-02
-8.95837322e-02 9.84669089e-01 1.08956598e-01 3.92652869e-01
-2.47241467e-01 2.82213509e-01 -3.17499965e-01 -5.70608735e-01
1.32322997e-01 2.79179990e-01 4.31761503e-01 -7.04323590e-01
1.09921455e+00 1.60029195e-02 2.80263364e-01 -3.66181761e-01
-1.90228783e-02 -1.08233273e-01 -4.15545762e-01 -1.52970597e-01
5.98461151e-01 -9.71567273e-01 -8.68620276e-01 4.37534899e-01
-1.10955310e+00 -6.17784373e-02 -4.61131006e-01 5.91363370e-01
-7.94445217e-01 2.62493938e-01 -3.30963910e-01 -1.43429780e+00
-2.25236267e-01 -1.05063403e+00 5.72114289e-01 1.59830958e-01
-6.81252331e-02 -1.29953921e+00 1.98490530e-01 -3.44667614e-01
5.53249657e-01 6.74365282e-01 1.01328874e+00 -8.27918172e-01
-2.12381512e-01 -3.80618393e-01 -1.57677941e-02 5.27504802e-01
2.80465186e-01 5.13498604e-01 -7.12179184e-01 -1.94129050e-01
2.43414134e-01 5.29921532e-01 1.27285138e-01 7.38088667e-01
1.17189252e+00 -2.28386313e-01 -4.66519454e-03 7.29447603e-02
1.59268022e+00 6.30221665e-01 5.19184113e-01 2.65459478e-01
-8.71093720e-02 6.06830895e-01 4.66126829e-01 1.06310105e+00
4.81581926e-01 3.45355779e-01 1.96835592e-01 2.96131074e-01
1.00626683e+00 6.72110692e-02 2.47516349e-01 7.48242974e-01
-2.32734114e-01 -2.01340795e-01 -9.25202549e-01 4.95313168e-01
-1.72775578e+00 -1.14157224e+00 -1.95388332e-01 2.59531641e+00
2.59560764e-01 2.37087101e-01 1.45916864e-01 2.01552868e-01
6.15708947e-01 -8.55365396e-02 -3.43869537e-01 -6.64452255e-01
-3.58928055e-01 4.40543741e-01 8.27420056e-01 4.76129830e-01
-9.28095043e-01 -2.66732246e-01 6.70163870e+00 8.19051147e-01
-6.03141665e-01 -1.51659936e-01 1.04017389e+00 1.98873147e-01
-7.34132290e-01 1.73001252e-02 -5.48606992e-01 1.07792163e+00
1.44633317e+00 -1.24279201e+00 3.65511090e-01 8.01934063e-01
7.35784173e-01 -4.70697701e-01 -8.60359073e-01 6.50140822e-01
-6.07116997e-01 -1.03760636e+00 -4.15243238e-01 3.88855159e-01
8.96467507e-01 -1.69132799e-02 1.02807522e-01 1.24956623e-01
2.41868109e-01 -8.50516438e-01 7.22031415e-01 7.99465299e-01
2.66687483e-01 -1.45499015e+00 1.01758122e+00 5.75654209e-01
-1.31911659e+00 -1.49401635e-01 -2.22861767e-01 -1.06521055e-01
7.12486088e-01 1.32714629e+00 -4.16493177e-01 1.31704307e+00
7.01622486e-01 2.91353643e-01 -8.44916180e-02 1.07901359e+00
-2.08995029e-01 4.62343395e-01 -8.08009505e-01 3.03268488e-02
8.10037702e-02 -8.24325562e-01 3.52009207e-01 9.60499108e-01
1.18194413e+00 -5.54659590e-02 -2.66144037e-01 8.35200429e-01
4.00444210e-01 2.02141795e-02 -4.32461828e-01 1.11924030e-01
4.94015604e-01 1.17352486e+00 -7.28619277e-01 -1.84706315e-01
-8.61251354e-01 3.95911872e-01 -5.45834959e-01 6.18447781e-01
-9.84174907e-01 -2.38703400e-01 4.29422855e-01 -3.09205860e-01
6.46872997e-01 -3.94825131e-01 -5.16054153e-01 -9.28033650e-01
1.97897166e-01 -2.96086133e-01 2.87550151e-01 -7.69459605e-01
-1.66533542e+00 2.20017046e-01 5.27072906e-01 -1.26263201e+00
-1.03054476e+00 -4.73651737e-01 -1.23304868e+00 1.68229353e+00
-1.91682661e+00 -5.38000703e-01 5.48421264e-01 3.98927152e-01
1.41719773e-01 1.16481381e-02 9.40478146e-01 -1.53132915e-01
-5.66694379e-01 -3.52199711e-02 1.00255835e+00 -2.61897087e-01
2.06746627e-02 -1.62502933e+00 7.67270029e-01 8.20627570e-01
-1.05902456e-01 2.69434929e-01 1.08825433e+00 -7.04924285e-01
-8.82490754e-01 -7.55564809e-01 1.12082505e+00 -6.43440634e-02
1.16122627e+00 -1.40892684e-01 -7.61200607e-01 4.83090580e-01
6.45212233e-01 -4.76139158e-01 9.95782375e-01 -1.05306424e-01
2.84081042e-01 -1.28460318e-01 -1.33310735e+00 2.04749748e-01
-1.38534635e-01 -1.95653886e-01 -6.97268844e-01 3.08278888e-01
1.50248840e-01 1.36886593e-02 -1.36101377e+00 5.43694615e-01
5.41209400e-01 -1.09675694e+00 7.04528511e-01 2.53605783e-01
-2.53565401e-01 -3.66254866e-01 -2.76100874e-01 -1.83593249e+00
-3.23207587e-01 -9.10488963e-01 -1.74873590e-01 1.32312119e+00
7.00805962e-01 -1.27770996e+00 3.04801911e-01 1.27751827e+00
3.57814312e-01 -5.29301047e-01 -1.51947999e+00 -1.04754043e+00
3.49847615e-01 -6.40236676e-01 1.39744258e+00 6.73661351e-01
-3.42214741e-02 -4.13250864e-01 -2.97297478e-01 6.08086824e-01
1.09607446e+00 5.40306687e-01 1.41906485e-01 -1.27277339e+00
-3.97679657e-01 -5.55552840e-01 9.36421845e-03 -2.69371986e-01
1.54410422e-01 -3.30384672e-01 -1.70233428e-01 -1.31668878e+00
-4.46370170e-02 -1.17333688e-01 -6.89658105e-01 1.44108951e-01
1.15371250e-01 -4.43207562e-01 3.90135109e-01 -2.50354052e-01
3.43890160e-01 8.26607406e-01 4.70213890e-01 1.32576630e-01
1.60234913e-01 5.88014364e-01 -2.01654121e-01 5.37565351e-01
1.30927205e+00 -2.47094303e-01 -5.39514899e-01 2.48061940e-01
4.82452244e-01 3.63045245e-01 2.46160582e-01 -8.23191285e-01
1.28788948e-01 -2.93987304e-01 4.93837416e-01 -1.35808921e+00
2.45358441e-02 -1.17779803e+00 1.13436759e+00 3.57424438e-01
3.68154824e-01 6.52034104e-01 3.68627161e-01 7.19779432e-01
-2.71490574e-01 -4.09624904e-01 5.77549934e-02 1.32438928e-01
-2.33333200e-01 7.14420378e-02 -5.27503610e-01 -5.64143062e-01
1.35264921e+00 2.48673670e-02 -8.43159333e-02 -4.58670229e-01
-1.04638648e+00 6.82033241e-01 1.85954764e-01 6.19179495e-02
1.92474738e-01 -1.28808784e+00 -7.66602874e-01 -4.60978374e-02
-3.67581189e-01 -3.10043573e-01 3.37482512e-01 5.59506774e-01
-8.87710676e-02 5.65438867e-01 -2.73228548e-02 -3.35715353e-01
-5.45234680e-01 7.46200621e-01 1.71154454e-01 -5.06296754e-01
-1.90367371e-01 -3.46904993e-02 -3.91021091e-03 1.04606770e-01
-2.22063184e-01 -7.59077489e-01 -1.72089770e-01 6.24501407e-01
4.80286509e-01 5.08535087e-01 1.43557787e-01 -4.49220568e-01
-1.70670554e-01 3.88789266e-01 5.11774421e-01 -3.31079185e-01
1.43285930e+00 -4.24811929e-01 -2.36400694e-01 1.05848479e+00
6.30797088e-01 -4.97453623e-02 -1.20445430e+00 3.47266346e-01
4.60127831e-01 1.91468783e-02 -1.49016306e-01 -9.70558941e-01
-7.73574054e-01 5.58764100e-01 1.88270360e-01 9.75549698e-01
1.28226590e+00 -5.98749042e-01 4.44051266e-01 -4.13651988e-02
7.28550851e-01 -1.08791780e+00 -1.37732887e+00 2.21529648e-01
1.00479412e+00 -7.60456860e-01 -7.89663047e-02 -1.53970674e-01
-4.00187254e-01 1.23883390e+00 -6.04715288e-01 7.56822154e-03
1.37583208e+00 3.85465652e-01 -4.12249953e-01 4.87575144e-01
-8.48016441e-01 2.63344403e-02 1.82796076e-01 4.61696476e-01
1.44579262e-01 8.96426320e-01 -7.45114982e-01 1.17458189e+00
-3.04396391e-01 8.28617588e-02 7.08766818e-01 8.03998768e-01
1.21323444e-01 -1.12920225e+00 -5.62980533e-01 5.41677773e-01
-5.92018485e-01 -2.60771930e-01 4.08001423e-01 6.91436410e-01
-2.46773452e-01 9.16810989e-01 2.83665180e-01 1.04178987e-01
3.78810823e-01 5.37648737e-01 -1.28016651e-01 -1.29937738e-01
-3.42244208e-01 4.80999559e-01 1.57475658e-02 -4.99679372e-02
-1.90452084e-01 -1.18648815e+00 -9.08660531e-01 -4.83447641e-01
-1.64639920e-01 5.64308763e-01 8.43315899e-01 8.04026723e-01
8.58870745e-02 3.51304948e-01 1.19629562e+00 -1.04114246e+00
-9.84011471e-01 -8.10135365e-01 -1.29098773e+00 -5.70365414e-02
7.72625953e-02 -5.16875803e-01 -7.74526715e-01 -1.91818282e-01] | [6.072834491729736, 3.0294950008392334] |
37475279-b5f6-4ccf-85c0-94589d000021 | assisting-undergraduate-students-in-writing | null | null | https://aclanthology.org/2020.bea-1.11 | https://aclanthology.org/2020.bea-1.11.pdf | Assisting Undergraduate Students in Writing Spanish Methodology Sections | In undergraduate theses, a good methodology section should describe the series of steps that were followed in performing the research. To assist students in this task, we develop machine-learning models and an app that uses them to provide feedback while students write. We construct an annotated corpus that identifies sentences representing methodological steps and labels when a methodology contains a logical sequence of such steps. We train machine-learning models based on language modeling and lexical features that can identify sentences representing methodological steps with 0.939 f-measure, and identify methodology sections containing a logical sequence of steps with an accuracy of 87{\%}. We incorporate these models into a Microsoft Office Add-in, and show that students who improved their methodologies according to the model feedback received better grades on their methodologies. | ['Aurelio Lopez-Lopez', "Samuel Gonz{\\'a}lez-L{\\'o}pez", 'Steven Bethard'] | 2020-07-01 | null | null | null | ws-2020-7 | ['logical-sequence'] | ['reasoning'] | [ 7.44302617e-03 2.16019168e-01 -4.83251333e-01 -6.76561475e-01
-9.56917226e-01 -8.38689446e-01 2.32534260e-01 8.50044847e-01
-3.40501517e-01 7.10074127e-01 1.32079631e-01 -1.06258547e+00
-2.51434267e-01 -9.94890451e-01 -9.97186542e-01 4.01007026e-01
4.30023581e-01 -9.41672921e-02 7.44002685e-02 -2.29670897e-01
1.12850833e+00 5.81868470e-01 -1.46671808e+00 7.40296662e-01
1.22273314e+00 3.50911766e-01 2.63283700e-01 8.61601233e-01
-7.35382795e-01 1.45970607e+00 -1.00103414e+00 -6.05796993e-01
-4.47581559e-01 -5.02074897e-01 -1.16732574e+00 -2.29208574e-01
5.65338850e-01 4.02484089e-02 1.56952783e-01 8.93498719e-01
-4.19715047e-02 1.58187181e-01 7.04099298e-01 -6.59491599e-01
-9.35285151e-01 1.06109977e+00 -3.05530012e-01 2.31160328e-01
1.20375228e+00 -1.53750584e-01 7.23573506e-01 -8.28432024e-01
6.25646830e-01 7.94819534e-01 9.48622704e-01 6.26412630e-01
-1.04947364e+00 -6.75130486e-01 -1.97493866e-01 8.40159655e-02
-9.07764554e-01 -2.39162982e-01 6.65660083e-01 -1.00601923e+00
1.00047600e+00 3.08862984e-01 9.22744751e-01 8.98374617e-01
6.99582398e-01 6.28333211e-01 1.13779438e+00 -1.01081550e+00
1.45451829e-01 8.95550072e-01 6.50350332e-01 9.35335696e-01
3.95775527e-01 -6.13835692e-01 -7.23578334e-01 3.43265504e-01
3.69006366e-01 -9.34491456e-02 3.44246835e-01 3.46636593e-01
-9.25583959e-01 7.14003563e-01 -2.34028161e-01 5.80853999e-01
8.02990496e-02 -2.90936589e-01 1.00509696e-01 4.73717421e-01
-2.05140010e-01 9.06190634e-01 -5.47858417e-01 -6.94029808e-01
-1.17656517e+00 1.65839583e-01 1.05922782e+00 1.29136324e+00
6.10695601e-01 -1.87632218e-01 1.55916903e-02 7.53171265e-01
5.30956686e-01 2.01073959e-01 6.73508704e-01 -1.03282547e+00
4.90521163e-01 1.31844866e+00 -2.23605871e-01 -6.84745193e-01
-5.96123151e-02 -3.34238082e-01 -1.46959260e-01 1.01841673e-01
3.83236080e-01 -8.93723294e-02 -5.01346290e-01 1.13664651e+00
-2.51993895e-01 -3.35951418e-01 1.27442196e-01 -9.63629484e-02
1.42043936e+00 6.98581219e-01 4.03491646e-01 -2.08639681e-01
1.47067487e+00 -7.09569812e-01 -1.01766872e+00 -1.54637650e-01
1.37535191e+00 -1.22816467e+00 1.45860934e+00 7.70467877e-01
-1.60705900e+00 -1.07563710e+00 -1.14618444e+00 -2.20756769e-01
-6.59615219e-01 2.86593556e-01 3.14869136e-01 1.06252599e+00
-8.37604225e-01 8.14828753e-01 -3.37128699e-01 -2.69823283e-01
1.62691832e-01 -1.01485752e-01 -1.70433074e-01 1.18260041e-01
-8.31949651e-01 8.12480807e-01 3.12293060e-02 -3.73322695e-01
-5.18552363e-01 -1.15728700e+00 -1.14108109e+00 1.08281806e-01
-8.53014216e-02 -2.61970937e-01 1.73697877e+00 -3.37805271e-01
-1.37946510e+00 1.29200876e+00 -1.10233530e-01 1.50175551e-02
3.75474662e-01 -1.96316242e-01 -2.80723602e-01 -4.55822572e-02
4.22164023e-01 3.03807378e-01 -1.40954852e-01 -7.69341648e-01
-7.91782558e-01 -1.50537834e-01 2.29622230e-01 -1.03694415e-02
-8.90640438e-01 2.83661485e-01 -2.75041051e-02 -2.83671916e-01
2.31437515e-02 -3.26172650e-01 1.70862004e-02 -4.45464402e-01
-2.62483001e-01 -7.38438904e-01 1.31673336e-01 -5.84079146e-01
1.92783296e+00 -1.44291937e+00 -6.11190498e-01 1.05670810e-01
-3.88361998e-02 3.62189487e-02 1.46547765e-01 6.05707407e-01
-2.79149804e-02 8.95041764e-01 4.14874941e-01 1.33202314e-01
2.28045017e-01 -4.96804953e-01 -1.26076534e-01 -2.52889484e-01
-1.05342522e-01 4.33590323e-01 -8.95561337e-01 -6.64468586e-01
1.33618250e-01 2.50773244e-02 -3.46039295e-01 4.95366901e-01
-5.74273467e-02 -3.35337408e-03 -3.87391984e-01 7.54740179e-01
1.86433479e-01 -1.09178983e-01 1.17728747e-01 5.36236405e-01
-5.62205851e-01 9.04038966e-01 -1.05125308e+00 1.53140068e+00
-9.40369725e-01 7.70868421e-01 -4.80006695e-01 -1.01054621e+00
1.38361430e+00 3.05546701e-01 6.15600087e-02 -6.82695389e-01
-2.11214304e-01 4.68331695e-01 -1.56779930e-01 -1.17505670e+00
3.93163264e-01 2.00079665e-01 -1.90656245e-01 6.17149591e-01
-1.33801270e-02 -6.08360171e-01 8.51813972e-01 4.12860274e-01
1.05396652e+00 3.60127211e-01 4.31496352e-01 -2.20398903e-01
9.26586270e-01 9.32228751e-03 1.32708505e-01 1.02959228e+00
1.25835791e-01 1.14996664e-01 4.55082148e-01 -3.86572510e-01
-7.88742125e-01 -4.56791520e-01 -3.12482361e-02 1.16980267e+00
-6.80798471e-01 -7.76395142e-01 -8.79374564e-01 -6.40498936e-01
-2.95398027e-01 1.19627988e+00 -3.17623079e-01 -2.39028588e-01
-3.89685124e-01 -4.04673479e-02 3.28092068e-01 6.56130493e-01
1.99213967e-01 -1.31398582e+00 -6.55787230e-01 2.23476753e-01
9.39466432e-02 -9.28904414e-01 -2.63551176e-01 3.10872704e-01
-8.23041201e-01 -6.94246233e-01 -5.15568182e-02 -1.50042701e+00
7.72419751e-01 -6.67409226e-02 1.17118824e+00 4.17901307e-01
-2.32497707e-01 3.90578568e-01 -1.40364811e-01 -6.64799213e-01
-7.14764416e-01 1.14528924e-01 -3.86341184e-01 -1.10881829e+00
1.02938819e+00 -2.53339440e-01 7.13879913e-02 -4.62989599e-01
-5.57861865e-01 1.38392910e-01 6.02205515e-01 3.67374480e-01
1.42705575e-01 1.54343724e-01 4.89154041e-01 -1.20691502e+00
9.87908900e-01 -3.53147328e-01 -5.77035367e-01 5.13522923e-01
-8.16064775e-01 7.33013004e-02 7.80122399e-01 -1.85086131e-01
-7.55653799e-01 -1.43703848e-01 -4.56784964e-01 4.92372602e-01
-5.16701996e-01 1.07433784e+00 -6.68596849e-02 -1.57364801e-01
8.22006404e-01 2.16244191e-01 -2.79515207e-01 -3.21381301e-01
-2.06654131e-01 9.17623103e-01 6.79137632e-02 -9.58620489e-01
6.59642220e-01 -5.85957825e-01 -7.25763515e-02 -8.64654541e-01
-1.24750149e+00 -2.64159977e-01 -6.78148150e-01 -6.43937349e-01
5.57969928e-01 -6.99416637e-01 -1.18247843e+00 -1.56082273e-01
-1.10070062e+00 -4.94361103e-01 -2.04941109e-01 7.79025435e-01
-3.48618478e-01 1.27341509e-01 -6.13955140e-01 -1.01613975e+00
6.27507083e-03 -1.13418174e+00 2.82149732e-01 9.57082093e-01
-9.35655832e-01 -1.15617669e+00 -9.93710533e-02 1.00812411e+00
-1.88405234e-02 -1.25891566e-01 1.34121811e+00 -9.40404236e-01
-1.78656995e-01 -3.37800086e-01 3.96423250e-01 3.99553448e-01
-4.52309139e-02 4.90921021e-01 -6.69610262e-01 3.64838719e-01
-1.09295063e-01 -3.96203846e-01 2.24791259e-01 2.55639404e-01
1.41437352e+00 -3.55275571e-01 -2.46261492e-01 2.09128544e-01
1.24156439e+00 7.43448138e-01 1.47265717e-01 9.26765919e-01
3.29701930e-01 1.03850746e+00 3.69925439e-01 -9.12365783e-03
4.61110860e-01 1.43504202e-01 -4.17933643e-01 5.62876523e-01
-9.65655874e-03 -5.85029721e-01 7.93671131e-01 1.32417274e+00
3.10126215e-01 1.22201711e-01 -1.39460289e+00 8.73841107e-01
-1.24388885e+00 -1.04926550e+00 -6.60829544e-01 2.04144335e+00
1.43995750e+00 8.94211233e-01 1.23371758e-01 4.26247537e-01
7.58436620e-02 -3.48643363e-01 3.10261577e-01 -1.29504728e+00
3.78659844e-01 6.71544373e-01 -9.79223996e-02 8.93939853e-01
-5.31749547e-01 8.20975780e-01 6.90235806e+00 5.20086825e-01
-8.17498326e-01 -4.62279677e-01 4.57369536e-01 4.41690773e-01
-6.70414984e-01 -1.02242835e-01 -1.46619439e+00 3.11932772e-01
1.66005063e+00 -6.26721084e-01 -6.91278800e-02 9.00268674e-01
6.86144471e-01 -3.43897268e-02 -1.30049860e+00 2.74002969e-01
1.77015230e-01 -1.65482461e+00 6.41042516e-02 -2.11167946e-01
9.10550714e-01 -1.26742220e+00 -1.34336706e-02 7.48627961e-01
5.54070592e-01 -1.20905435e+00 7.64889717e-01 8.27518046e-01
3.33813339e-01 -6.98367774e-01 7.81546295e-01 6.40739918e-01
-8.49628270e-01 4.65944149e-02 -2.28340983e-01 -7.15018272e-01
-8.35964501e-01 3.30749989e-01 -1.28791904e+00 2.48439834e-01
5.24134696e-01 6.76354468e-01 -9.51291800e-01 1.12882209e+00
-7.43136525e-01 9.63344097e-01 2.57351995e-01 -9.48129237e-01
-8.43293145e-02 -2.84153879e-01 -1.45022616e-01 1.54459572e+00
6.34696484e-01 1.48325741e-01 2.73324430e-01 1.11633980e+00
-1.38500974e-01 5.30756414e-01 -6.11278653e-01 -3.37002039e-01
9.70973015e-01 1.10678554e+00 -8.26266050e-01 -4.37211603e-01
-6.20044708e-01 2.56253034e-01 1.25870734e-01 1.32246867e-01
-3.21459204e-01 -1.05669868e+00 1.35133624e-01 4.51025456e-01
-9.39198509e-02 -1.09682880e-01 -1.28090179e+00 -8.05951834e-01
1.40851170e-01 -1.00352585e+00 -1.64716709e-02 -9.86746967e-01
-9.52522933e-01 2.01871112e-01 -2.67258972e-01 -1.07118022e+00
3.26167271e-02 -5.79968691e-01 -8.39801908e-01 1.07446623e+00
-1.06641614e+00 -7.35116959e-01 -4.13693249e-01 -1.97783738e-01
8.49981666e-01 -3.70391607e-01 9.28985119e-01 4.51263040e-02
-6.10640585e-01 6.89371467e-01 -3.90985340e-01 2.58475959e-01
7.29287863e-01 -1.67819571e+00 4.35299613e-02 9.49802876e-01
6.72434196e-02 1.35640669e+00 7.11344898e-01 -7.89676011e-01
-1.08923757e+00 -6.98088527e-01 1.99866486e+00 -6.54303849e-01
8.36502075e-01 -1.11732751e-01 -7.56457508e-01 7.58692205e-01
4.54095691e-01 -6.62631214e-01 1.45258558e+00 9.14687589e-02
-1.03658848e-01 -6.09504804e-02 -7.09991455e-01 4.61566627e-01
7.59033084e-01 -5.94485343e-01 -1.36061668e+00 5.63481152e-01
3.38415235e-01 -5.20847976e-01 -1.29062784e+00 -4.95019034e-02
6.88887477e-01 -6.23054385e-01 6.77306175e-01 -9.21611309e-01
1.29888499e+00 1.68784298e-02 3.92633766e-01 -9.47144330e-01
-6.15949742e-02 -6.09889925e-01 8.26060474e-02 1.52002931e+00
8.84465933e-01 1.33788481e-01 9.23126996e-01 6.56155467e-01
-3.78312051e-01 -7.88599372e-01 1.41133806e-02 -6.66394234e-01
5.48483014e-01 -5.43387473e-01 2.65699685e-01 1.12620842e+00
5.41409373e-01 3.46339047e-01 4.74778503e-01 -2.67716765e-01
3.20062578e-01 7.07993889e-03 6.37201965e-01 -1.53150892e+00
2.59480268e-01 -8.92616570e-01 -3.17698717e-02 -1.01771796e+00
3.38793576e-01 -1.04833865e+00 -1.28981188e-01 -1.93297207e+00
2.34951660e-01 -2.99551357e-02 8.03522915e-02 6.56311989e-01
-2.91213214e-01 -1.47998571e-01 -2.26673588e-01 -2.84277320e-01
-4.00058836e-01 -3.35056305e-01 7.65589416e-01 5.69435023e-02
-4.24950242e-01 2.01280981e-01 -1.14263809e+00 9.83050406e-01
5.74348927e-01 -4.05570090e-01 -2.27293700e-01 -2.42449686e-01
7.19787180e-01 -4.00400683e-02 -2.67953724e-01 -1.41169429e+00
4.48924571e-01 -6.69059575e-01 7.54201531e-01 -4.50572878e-01
-4.89476919e-01 -6.59587026e-01 -3.84951472e-01 5.38891494e-01
-1.28582180e+00 3.68427068e-01 5.44554949e-01 -3.37655276e-01
-1.13238774e-01 -1.16113734e+00 5.61903834e-01 -2.41040125e-01
-2.50225335e-01 -6.39783323e-01 -1.00457585e+00 -7.13808462e-02
1.20171165e+00 -4.72438604e-01 -2.27588698e-01 -3.04764241e-01
-6.90813839e-01 1.81695104e-01 1.21025063e-01 2.35044241e-01
6.58669055e-01 -1.09289038e+00 -4.52755362e-01 3.12107891e-01
-4.00593951e-02 -2.56195575e-01 -1.83990210e-01 7.01098144e-01
-1.10455644e+00 7.42835641e-01 -3.85668248e-01 -6.25225678e-02
-1.59007537e+00 4.42266464e-02 3.75721529e-02 -4.81321096e-01
-1.97289392e-01 1.22981215e+00 -8.49048793e-01 -8.69007409e-01
5.72074950e-01 -5.69154441e-01 -9.50978220e-01 3.22119534e-01
8.78585219e-01 3.24606866e-01 5.01526177e-01 -4.34132339e-03
-1.17450155e-01 5.69429517e-01 -2.43837699e-01 -3.85547453e-03
1.41212106e+00 2.09075287e-01 1.10120855e-01 1.06536961e+00
1.01950681e+00 7.35495389e-01 -2.01289386e-01 3.13244686e-02
7.26040781e-01 -1.52564526e-01 -1.46605596e-01 -1.28112280e+00
-2.46393234e-01 9.92521226e-01 1.86963379e-01 5.85096300e-01
5.22193909e-01 -5.46478152e-01 3.29163760e-01 4.76965964e-01
-1.67407826e-01 -1.25975752e+00 1.95796177e-01 6.74489975e-01
5.27825296e-01 -1.08008599e+00 2.60101795e-01 -3.85770559e-01
-2.28443250e-01 1.98601532e+00 1.02485526e+00 2.92107821e-01
4.44416642e-01 4.34631199e-01 1.48722213e-02 -7.55845532e-02
-8.96620750e-01 4.02208298e-01 4.37751532e-01 3.05459261e-01
1.59211004e+00 -5.92339449e-02 -9.76142108e-01 8.80824566e-01
-8.13363492e-01 3.43468636e-01 1.24026525e+00 1.46920347e+00
-8.87816012e-01 -1.22619200e+00 -4.76933271e-01 4.50343877e-01
-5.83560050e-01 -2.84717023e-01 -8.73538494e-01 6.74579620e-01
2.33962834e-02 1.20038700e+00 -2.10620344e-01 -5.00198245e-01
6.02467656e-01 7.02940404e-01 3.76219600e-01 -1.29541469e+00
-1.17067385e+00 -2.59800434e-01 8.27655792e-02 4.77542281e-02
-2.27992699e-01 -7.34411001e-01 -1.48330283e+00 -7.06413388e-01
9.93064642e-02 5.48430741e-01 6.44211769e-01 1.04284644e+00
-1.04246713e-01 8.71571243e-01 2.73463130e-01 2.35494539e-01
-7.52649784e-01 -1.20957625e+00 -1.35991387e-02 3.38766836e-02
1.50499344e-02 2.10710671e-02 -3.71644348e-01 4.37098920e-01] | [11.194523811340332, 9.215643882751465] |
614205f2-6a36-4bb3-b7d2-e1ea7a1881e4 | improving-prediction-confidence-in-learning | 2110.03123 | null | https://arxiv.org/abs/2110.03123v1 | https://arxiv.org/pdf/2110.03123v1.pdf | Improving Prediction Confidence in Learning-Enabled Autonomous Systems | Autonomous systems use extensively learning-enabled components such as deep neural networks (DNNs) for prediction and decision making. In this paper, we utilize a feedback loop between learning-enabled components used for classification and the sensors of an autonomous system in order to improve the confidence of the predictions. We design a classifier using Inductive Conformal Prediction (ICP) based on a triplet network architecture in order to learn representations that can be used to quantify the similarity between test and training examples. The method allows computing confident set predictions with an error rate predefined using a selected significance level. A feedback loop that queries the sensors for a new input is used to further refine the predictions and increase the classification accuracy. The method is computationally efficient, scalable to high-dimensional inputs, and can be executed in a feedback loop with the system in real-time. The approach is evaluated using a traffic sign recognition dataset and the results show that the error rate is reduced. | ['Xenofon Koutsoukos', 'Dimitrios Boursinos'] | 2021-10-07 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 4.10771728e-01 4.98460770e-01 -4.19649146e-02 -7.44778633e-01
-2.53972590e-01 -2.47174188e-01 4.95717406e-01 1.52221069e-01
-4.88500834e-01 6.59943223e-01 -2.06308901e-01 -1.98123783e-01
-4.09571439e-01 -1.03674281e+00 -6.52220428e-01 -5.57878971e-01
-4.20709886e-02 8.26939404e-01 7.00600207e-01 -3.00246160e-02
4.04723108e-01 8.07929695e-01 -2.35609126e+00 6.42319620e-01
8.30407262e-01 1.51660323e+00 4.79424559e-03 6.71397448e-01
5.61503246e-02 6.94331765e-01 -3.01128864e-01 -2.60214619e-02
5.64854443e-01 -3.46731506e-02 -3.68869871e-01 -3.96138370e-01
2.86902994e-01 -8.37396085e-02 2.89257407e-01 7.28453636e-01
1.44159734e-01 6.02388196e-02 8.51709723e-01 -1.33348477e+00
1.21683478e-01 2.23124906e-01 4.42402244e-01 -2.86017537e-01
1.28834561e-01 1.57398820e-01 7.32856154e-01 -9.61669922e-01
4.27582443e-01 1.05391335e+00 8.93829048e-01 4.36501235e-01
-1.18385720e+00 -7.62801051e-01 -2.45899513e-01 6.17997468e-01
-1.11422586e+00 -3.20155054e-01 5.43287039e-01 -5.71906030e-01
7.71122575e-01 2.00834125e-01 7.75941253e-01 5.83000898e-01
5.11223555e-01 3.00244063e-01 1.11127114e+00 -6.35402560e-01
8.08262765e-01 5.23440003e-01 2.95169294e-01 6.36084020e-01
3.95441800e-01 5.03027558e-01 -5.96745431e-01 -2.04035491e-02
3.97135109e-01 1.04449093e-01 2.24656463e-01 -4.12677079e-01
-5.84293067e-01 7.97616124e-01 5.86210847e-01 3.76094133e-01
-7.57266104e-01 1.00190610e-01 2.03225106e-01 3.32897872e-01
1.65616035e-01 4.71608162e-01 -4.97211635e-01 -1.47203371e-01
-7.92800188e-01 9.22547579e-02 1.01075006e+00 4.78018135e-01
9.89110589e-01 -2.07574829e-01 -9.61338207e-02 5.34760356e-01
4.59368587e-01 3.78087819e-01 5.34399211e-01 -8.16354513e-01
4.91154343e-02 1.04009485e+00 -9.40876305e-02 -8.50157440e-01
-5.95926106e-01 -2.33245537e-01 -5.24459302e-01 1.15502787e+00
1.89288214e-01 -1.99171022e-01 -9.61403906e-01 1.18299830e+00
4.81930941e-01 2.67874926e-01 1.88337237e-01 8.02208900e-01
4.37485635e-01 4.07932132e-01 3.44843417e-02 1.77503943e-01
1.05188525e+00 -3.55176806e-01 -3.09672683e-01 9.85969529e-02
6.71239436e-01 -2.33678177e-01 8.03471208e-01 6.09545887e-01
-4.15853381e-01 -9.19318557e-01 -1.29362214e+00 2.64910460e-01
-7.87472665e-01 3.36436808e-01 8.14771429e-02 6.53203487e-01
-9.34071600e-01 7.97685444e-01 -8.50124598e-01 -3.97273988e-01
3.93702716e-01 7.99237609e-01 -2.92058975e-01 2.65087903e-01
-9.26487505e-01 1.16649950e+00 8.12321484e-01 6.63129166e-02
-5.52315474e-01 -6.71575189e-01 -5.07255256e-01 8.76907930e-02
-1.38262317e-01 -3.16182196e-01 8.74874353e-01 -1.15986478e+00
-1.79140496e+00 4.80651945e-01 1.65031910e-01 -9.02779341e-01
5.96625984e-01 -1.11912966e-01 -4.28883344e-01 -1.02287546e-01
-3.02956730e-01 6.88119411e-01 8.36148858e-01 -1.02640712e+00
-9.49323595e-01 -4.14661884e-01 -3.59687388e-01 -1.38705015e-01
-3.00120473e-01 -4.12629873e-01 1.51128158e-01 -7.59910152e-04
8.27645510e-02 -1.04253793e+00 -1.27365440e-01 1.70564741e-01
3.11242994e-02 -3.52884263e-01 1.18116593e+00 -4.43547547e-01
7.66556680e-01 -1.95566297e+00 -3.16819668e-01 9.74280775e-01
-3.06372624e-02 6.23248279e-01 1.07174367e-01 1.93323478e-01
5.81748411e-02 -4.21101213e-01 -2.29663625e-01 -3.82336713e-02
-1.14528842e-01 4.30962056e-01 -2.32208624e-01 1.16868943e-01
4.72644627e-01 7.35891879e-01 -3.92129332e-01 -3.48959237e-01
5.49830854e-01 4.82085764e-01 -4.70944285e-01 4.56953645e-01
-3.94872397e-01 1.64330691e-01 -3.21714401e-01 2.23988429e-01
4.04820949e-01 5.78239150e-02 4.51173186e-02 -3.03254455e-01
-1.67824969e-01 2.52183266e-02 -1.35049069e+00 1.05581045e+00
-5.12069941e-01 8.46918285e-01 -5.16633868e-01 -1.08255959e+00
1.73221362e+00 6.98082969e-02 2.19905019e-01 -7.28809059e-01
2.98594832e-01 4.28076595e-01 -4.36028652e-02 -4.37094212e-01
2.74614602e-01 1.10357746e-01 2.78313249e-01 3.80689621e-01
7.11103156e-02 7.33833611e-02 -8.07795376e-02 -3.02605450e-01
1.14445555e+00 2.63656169e-01 5.83147556e-02 -1.80007920e-01
7.87458122e-01 5.04305027e-02 5.05427718e-01 5.08034110e-01
5.06168902e-02 1.68641448e-01 2.00250477e-01 -8.20831001e-01
-1.15308750e+00 -6.55294478e-01 -2.63702989e-01 8.76211226e-01
3.76706608e-02 -2.40688063e-02 -6.92713201e-01 -6.80865347e-01
1.95644274e-01 8.86598706e-01 -8.54328275e-01 -3.20749253e-01
-4.52470839e-01 -1.79656148e-01 1.83825195e-01 7.62849689e-01
4.15667951e-01 -1.22100186e+00 -1.29116356e+00 3.29250276e-01
4.44486231e-01 -8.46953034e-01 5.27297497e-01 4.86068308e-01
-1.10655606e+00 -1.14654911e+00 3.47834155e-02 -4.59041566e-01
6.70221448e-01 -6.63945556e-01 6.76240623e-01 -6.52355701e-02
-2.79855907e-01 2.84321457e-01 -3.56799811e-01 -7.50454366e-01
-7.37928450e-01 -1.06278069e-01 1.65270254e-01 2.74879903e-01
6.54049039e-01 -4.32934433e-01 -3.98991168e-01 4.21918690e-01
-6.91182375e-01 -6.89045936e-02 6.49879813e-01 8.66731048e-01
5.66769063e-01 -2.37620682e-01 5.98454058e-01 -5.32066822e-01
5.29651701e-01 -2.69057214e-01 -1.06924522e+00 3.90310764e-01
-1.04700744e+00 5.42191029e-01 6.69458628e-01 -4.87443000e-01
-7.99499214e-01 4.69154000e-01 3.32519747e-02 -3.73799860e-01
-4.31872189e-01 3.27846617e-01 2.22557515e-01 -4.93827701e-01
1.06267893e+00 -3.29261422e-02 2.53803968e-01 -4.22309309e-01
7.08744749e-02 9.05726314e-01 3.90734673e-01 -1.83901086e-01
5.22803664e-01 2.09875628e-01 4.79987979e-01 -5.03725350e-01
-3.21404815e-01 -2.05478936e-01 -9.55239832e-01 -8.38158846e-01
5.77801108e-01 -4.04753655e-01 -1.02317166e+00 4.15723063e-02
-8.75080228e-01 -2.20790550e-01 -5.47870755e-01 6.94190502e-01
-6.85366392e-01 -1.89455613e-01 1.68002732e-02 -1.07320452e+00
-4.32813346e-01 -9.31559503e-01 8.15355897e-01 5.78118801e-01
-1.93770483e-01 -6.76767647e-01 1.77764550e-01 -5.72323799e-04
5.12756348e-01 2.22677797e-01 6.07770801e-01 -1.12956202e+00
-5.98058939e-01 -6.91842973e-01 -1.12701215e-01 6.32054746e-01
-2.22292900e-01 9.42188278e-02 -1.25823224e+00 1.12526767e-01
-3.72398823e-01 -2.94183344e-01 7.07111955e-01 1.77132964e-01
1.07710159e+00 -1.64513364e-01 -4.53187376e-01 2.18710348e-01
1.18004751e+00 3.69375020e-01 6.92793429e-01 3.83859575e-01
5.11387289e-02 7.36837924e-01 8.69896889e-01 3.53798926e-01
1.24764904e-01 5.61730027e-01 5.05895376e-01 1.00643747e-01
-4.48035680e-05 -2.22526759e-01 1.96453258e-01 2.39523843e-01
-8.22311938e-02 1.94265231e-01 -1.05780292e+00 2.76000947e-01
-2.07707381e+00 -7.92379439e-01 -1.05324626e-01 2.43769813e+00
4.00624812e-01 3.08139384e-01 9.44411010e-02 6.29577339e-01
5.67949593e-01 -7.37444699e-01 -5.40403008e-01 -7.00146735e-01
2.85542756e-01 3.50191742e-01 5.32151639e-01 5.49143076e-01
-9.25492644e-01 6.95604205e-01 6.17559767e+00 3.14869344e-01
-1.33086932e+00 -2.23887548e-01 6.02005780e-01 1.84686095e-01
1.15919247e-01 -7.67347887e-02 -7.83798456e-01 4.37233835e-01
1.38965952e+00 2.01858357e-01 1.20349035e-01 1.11991227e+00
3.29000130e-02 -2.87107348e-01 -1.32383895e+00 7.80616641e-01
-1.02649406e-01 -1.39442968e+00 -1.23755246e-01 5.18592484e-02
2.99217165e-01 2.40024716e-01 -1.38561502e-01 2.27159873e-01
2.29231745e-01 -9.13394690e-01 4.76815373e-01 1.33540416e+00
5.97063959e-01 -7.56620705e-01 1.05512631e+00 4.65641111e-01
-8.79134834e-01 -5.64177096e-01 -2.65225500e-01 -1.60422295e-01
-1.61796898e-01 5.35187483e-01 -1.52488124e+00 1.69264957e-01
7.17428505e-01 3.53355438e-01 -5.48482180e-01 1.19441783e+00
-4.65065539e-02 5.99399567e-01 -7.50042617e-01 -6.65724337e-01
-3.59014794e-02 -2.65537232e-01 3.82246733e-01 1.05700457e+00
4.79755372e-01 5.41279018e-02 -1.11539625e-01 9.32580411e-01
3.65678728e-01 9.76845473e-02 -6.01273715e-01 3.31485510e-01
3.87707174e-01 9.54273820e-01 -5.95024049e-01 -4.70736831e-01
1.41136482e-01 4.92574364e-01 3.18886459e-01 1.22054398e-01
-6.11133039e-01 -4.10522550e-01 4.32600617e-01 1.32675514e-01
4.67989594e-01 4.96884771e-02 -4.71753627e-01 -5.89214385e-01
2.93090016e-01 -4.03891504e-01 1.90582722e-01 -7.05574214e-01
-1.09540856e+00 6.63871586e-01 -2.73043722e-01 -1.41168296e+00
-6.40725195e-01 -1.06937122e+00 -4.51416582e-01 8.17961574e-01
-1.23487198e+00 -8.19071889e-01 -6.73851609e-01 5.46095669e-01
-1.96374804e-02 -4.30634946e-01 1.00190544e+00 9.72388983e-02
-1.21795923e-01 3.64191711e-01 -9.82988931e-05 2.60076094e-02
4.20909464e-01 -1.05103099e+00 -2.14553103e-01 5.62636733e-01
7.54952431e-02 -1.71876222e-01 6.38490379e-01 -6.55250847e-01
-1.04440296e+00 -1.09412646e+00 8.51874828e-01 -2.74295747e-01
3.50681335e-01 -2.93862134e-01 -8.37175608e-01 3.26047897e-01
-3.82978946e-01 3.36732268e-01 7.48210073e-01 -5.50992638e-02
-4.39717829e-01 -8.78793478e-01 -1.47427976e+00 4.47903089e-02
5.37838638e-01 -4.57817972e-01 -6.56233251e-01 1.65582094e-02
2.08176553e-01 -1.00090720e-01 -9.19458389e-01 7.06639230e-01
9.72387493e-01 -1.12050760e+00 5.07020354e-01 -5.12666166e-01
-4.71429639e-02 -4.41134244e-01 -1.37327671e-01 -1.08425117e+00
-1.25555575e-01 -3.22052501e-02 -7.90985972e-02 7.15964913e-01
7.46084750e-01 -9.51281369e-01 9.14506257e-01 1.10605848e+00
1.46737188e-01 -6.18905067e-01 -1.15810311e+00 -4.88092721e-01
-4.49647397e-01 -7.26765990e-01 6.04004204e-01 3.09802890e-01
5.91929443e-03 2.42132358e-02 4.48039267e-03 3.22588682e-01
5.83428800e-01 -1.73625760e-02 9.05600309e-01 -1.91415036e+00
-1.02183513e-01 -1.65310517e-01 -1.37836480e+00 -3.14472735e-01
-7.31525421e-02 -6.32141948e-01 2.33499214e-01 -1.12276495e+00
-3.42345625e-01 -7.59792864e-01 -4.19462264e-01 6.15871787e-01
3.86350542e-01 2.15057209e-01 1.81266576e-01 6.26835376e-02
-3.66367936e-01 5.08405387e-01 4.64879274e-01 6.16810545e-02
-3.40724677e-01 3.32374036e-01 4.09948789e-02 6.34742856e-01
8.46112669e-01 -5.09945393e-01 -1.60007089e-01 2.67327428e-01
-8.34840909e-03 -3.04845423e-01 6.75638139e-01 -1.72449982e+00
5.71053028e-01 1.52756572e-01 7.79631257e-01 -6.02908850e-01
3.62641841e-01 -1.47582245e+00 2.47145727e-01 8.42431962e-01
-5.67182779e-01 -3.00860077e-01 1.99354604e-01 5.65879047e-01
-2.35400975e-01 -2.47292221e-01 8.92648280e-01 5.82990646e-01
-6.69123590e-01 6.99036047e-02 -3.26218337e-01 -7.25127339e-01
1.31040251e+00 -5.22254348e-01 7.40401149e-02 -2.52601713e-01
-8.23475122e-01 4.50275652e-02 4.21881564e-02 4.40523088e-01
6.85807407e-01 -1.23311877e+00 -4.89551723e-01 6.59224808e-01
3.91480625e-01 -3.18670541e-01 -3.07025701e-01 4.91742134e-01
-4.24427718e-01 4.59466040e-01 -5.51185668e-01 -1.02845335e+00
-1.43036187e+00 1.19440988e-01 7.93960154e-01 8.66727382e-02
-3.66270095e-01 5.45538306e-01 -6.40318274e-01 -5.85186422e-01
4.08794433e-01 -6.48152173e-01 -4.65878665e-01 -1.60680756e-01
5.79510212e-01 3.24016035e-01 3.12296927e-01 -3.25405091e-01
-3.47155333e-01 4.59388882e-01 3.43148679e-01 -1.57720432e-01
1.49290371e+00 4.19628978e-01 1.43453211e-01 5.96817851e-01
1.02736175e+00 -6.11504197e-01 -1.35708058e+00 -3.35080504e-01
4.00435299e-01 -2.98953086e-01 1.08710274e-01 -1.02132225e+00
-8.64075601e-01 7.37445772e-01 1.36710274e+00 1.73274204e-01
1.14335883e+00 -2.70358503e-01 1.92788780e-01 8.05185318e-01
4.48008537e-01 -1.55098259e+00 -1.89707950e-01 3.56875241e-01
8.97670388e-01 -1.32450593e+00 -2.52858043e-01 1.61469895e-02
-4.89530176e-01 1.65589035e+00 6.40000761e-01 -2.96839476e-01
1.03902066e+00 4.06907976e-01 -8.16281699e-03 -8.68819188e-03
-8.71203601e-01 -1.23983048e-01 5.73956549e-01 6.49754226e-01
6.15455620e-02 2.43576705e-01 -2.56085396e-01 3.22979271e-01
-1.45386577e-01 5.49594760e-01 4.16244380e-02 7.30471551e-01
-7.88243413e-01 -1.09614503e+00 -4.71378416e-01 7.03564465e-01
3.57346386e-01 3.70967805e-01 -4.66396391e-01 4.09669101e-01
5.23504078e-01 8.84772301e-01 3.65447760e-01 -9.16134357e-01
4.99831170e-01 3.46548736e-01 1.00314505e-01 -2.38032505e-01
-6.26241446e-01 -6.88319266e-01 2.38060042e-01 -9.21320140e-01
-3.35546076e-01 -6.76114082e-01 -1.32804656e+00 1.82815999e-01
-3.89347345e-01 2.32215032e-01 1.21763933e+00 1.08167934e+00
8.38243365e-01 3.18095624e-01 7.07112908e-01 -7.41711140e-01
-4.87663239e-01 -9.94175196e-01 -1.57151565e-01 1.87294617e-01
2.01927811e-01 -7.44705439e-01 -1.14443384e-01 -3.48153301e-02] | [7.6927924156188965, -0.6782286167144775] |
e0f27df0-b1af-493a-9d1f-67ea07295a17 | ranking-with-fairness-constraints | 1704.0684 | null | https://arxiv.org/abs/1704.06840v4 | https://arxiv.org/pdf/1704.06840v4.pdf | Ranking with Fairness Constraints | Ranking algorithms are deployed widely to order a set of items in applications such as search engines, news feeds, and recommendation systems. Recent studies, however, have shown that, left unchecked, the output of ranking algorithms can result in decreased diversity in the type of content presented, promote stereotypes, and polarize opinions. In order to address such issues, we study the following variant of the traditional ranking problem when, in addition, there are fairness or diversity constraints. Given a collection of items along with 1) the value of placing an item in a particular position in the ranking, 2) the collection of sensitive attributes (such as gender, race, political opinion) of each item and 3) a collection of constraints that, for each k, bound the number of items with each attribute that are allowed to appear in the top k positions of the ranking, the goal is to output a ranking that maximizes the value with respect to the original rank quality metric while respecting the constraints. This problem encapsulates various well-studied problems related to bipartite and hypergraph matching as special cases and turns out to be hard to approximate even with simple constraints. Our main technical contributions are fast exact and approximation algorithms along with complementary hardness results that, together, come close to settling the approximability of this constrained ranking maximization problem. Unlike prior work on the constrained matching problems, our algorithm runs in linear time, even when the number of constraints is large, its approximation ratio does not depend on the number of constraints, and it produces solutions with small constraint violations. Our results rely on insights about the constrained matching problem when the objective satisfies properties that appear in common ranking metrics such as Discounted Cumulative Gain, Spearman's rho or Bradley-Terry. | ['Nisheeth K. Vishnoi', 'Damian Straszak', 'L. Elisa Celis'] | 2017-04-22 | null | null | null | null | ['hypergraph-matching'] | ['graphs'] | [ 2.53334790e-01 3.45940322e-01 -5.65161884e-01 -3.96363318e-01
-4.25832629e-01 -9.56395447e-01 2.12099090e-01 7.03881383e-01
-4.74309504e-01 6.51537120e-01 2.81462193e-01 -3.47546190e-01
-8.84514093e-01 -9.65642214e-01 -6.25593066e-01 -4.59481418e-01
-1.78564340e-01 8.79304409e-01 1.68862656e-01 -3.90308589e-01
4.58550543e-01 3.24067056e-01 -1.65402925e+00 2.03012198e-01
7.99378514e-01 1.01906514e+00 -3.95632178e-01 4.06192869e-01
9.92604420e-02 4.60122198e-01 -3.18517715e-01 -9.07403350e-01
6.16068006e-01 -2.27500439e-01 -1.01029754e+00 -6.98961243e-02
5.39587557e-01 -2.26525113e-01 -1.60722971e-01 1.28035617e+00
3.25941026e-01 2.69377083e-01 6.18352115e-01 -1.43120837e+00
-4.99984533e-01 8.90459776e-01 -6.41762912e-01 7.99646322e-03
5.51542699e-01 -4.93000865e-01 1.78992522e+00 -3.21681350e-01
6.26382470e-01 1.12570703e+00 1.92006692e-01 2.65860319e-01
-1.35497499e+00 -3.23957831e-01 4.34897751e-01 -7.19323978e-02
-8.44026208e-01 -2.04221636e-01 3.82002473e-01 -3.02881211e-01
4.19160604e-01 8.87774408e-01 5.58300674e-01 3.15485448e-01
-3.09101820e-01 3.62414181e-01 1.19120860e+00 -4.54172194e-01
3.66928726e-01 2.28767499e-01 5.87395787e-01 3.57335597e-01
7.43925154e-01 2.45317593e-02 -5.91155469e-01 -7.84473121e-01
4.08437736e-02 -4.97058630e-02 -2.64052212e-01 -6.00509584e-01
-9.46067572e-01 9.60653901e-01 3.25164497e-01 4.94349860e-02
-9.71962810e-02 -4.61437963e-02 1.43173784e-01 6.01157963e-01
3.87338758e-01 7.05710053e-01 -4.94396031e-01 2.75509328e-01
-7.41366327e-01 4.70433116e-01 1.15086269e+00 7.77710676e-01
8.46896708e-01 -7.21009672e-01 -3.74252141e-01 7.14861095e-01
1.55678436e-01 5.11711359e-01 -1.21182188e-01 -1.03844869e+00
5.28423309e-01 7.61600554e-01 4.56488520e-01 -1.30546737e+00
-3.44019055e-01 -2.99000055e-01 -3.79239917e-01 -1.17569670e-01
7.05535173e-01 -8.56507476e-03 -4.16047543e-01 2.13059330e+00
3.12530637e-01 -3.15001845e-01 -3.56263220e-01 1.02682745e+00
4.30281550e-01 4.51607198e-01 -1.54290676e-01 -5.15090823e-01
1.40440738e+00 -4.72658575e-01 -3.01238090e-01 -1.69317931e-01
4.24235523e-01 -7.73771644e-01 7.88730145e-01 2.14984372e-01
-1.27016664e+00 1.20168924e-01 -8.44106913e-01 7.42362365e-02
-3.59636366e-01 -6.06769681e-01 7.35931754e-01 8.66248012e-01
-1.09146845e+00 6.67640507e-01 -1.48678897e-02 -3.37285906e-01
-7.36391246e-02 7.21044302e-01 -2.14930385e-01 -1.54549032e-01
-1.35654879e+00 8.18743110e-01 -4.06018905e-02 -7.97865838e-02
-6.51774630e-02 -5.76530457e-01 -3.51171076e-01 3.89839351e-01
7.83397436e-01 -8.23753417e-01 1.08157802e+00 -1.17539871e+00
-1.08435988e+00 1.08955824e+00 1.37738660e-01 1.15435578e-01
5.40761769e-01 2.34004661e-01 -1.95635781e-01 -1.71951100e-01
1.78491861e-01 2.15652883e-01 1.47258833e-01 -9.86314416e-01
-8.50424647e-01 -6.63522840e-01 7.84853160e-01 5.39662778e-01
-5.79421937e-01 1.61984637e-01 -5.99719584e-01 -1.91325706e-03
1.68157220e-01 -9.69730616e-01 -5.42265296e-01 -2.41085991e-01
-3.30987304e-01 -2.21385449e-01 -2.14877888e-03 -2.51526415e-01
1.24826598e+00 -1.89007390e+00 2.07368255e-01 1.03541040e+00
1.87935308e-01 -2.04765186e-01 -1.83699250e-01 5.34723520e-01
1.97348103e-01 4.01525646e-01 5.39053120e-02 2.64621168e-01
4.50918704e-01 8.79271030e-02 -2.02728629e-01 5.63260674e-01
-5.67634046e-01 6.67109728e-01 -1.06052279e+00 -2.21262529e-01
-3.54884773e-01 -1.41444087e-01 -8.70116770e-01 -1.70585945e-01
-1.05606362e-01 -2.19355732e-01 -4.33979899e-01 3.98611933e-01
6.49721503e-01 -3.38126689e-01 7.26317406e-01 -4.38797623e-02
8.18271413e-02 3.49279433e-01 -1.46287155e+00 8.59330475e-01
-1.15868896e-01 1.63446784e-01 2.83319145e-01 -7.51251280e-01
5.27167141e-01 -8.21774974e-02 5.91026306e-01 -7.52332389e-01
2.02643685e-02 3.82076740e-01 -6.86590448e-02 -1.13098942e-01
8.14962626e-01 -9.76394862e-02 -4.15738523e-01 5.93754113e-01
-5.73882997e-01 2.38224551e-01 4.90735888e-01 6.80309892e-01
1.16975844e+00 -6.67539418e-01 3.73445638e-02 -4.53529894e-01
2.81967968e-01 6.44184500e-02 5.80505311e-01 1.08450294e+00
2.64748305e-01 1.96846575e-01 9.70589995e-01 -2.06074372e-01
-1.05896044e+00 -7.55008280e-01 -1.09145157e-01 1.40162241e+00
6.98661268e-01 -2.51225084e-01 -3.69525701e-01 -5.23911834e-01
4.87696230e-01 1.59628659e-01 -5.89318573e-01 -8.41181260e-03
-1.67297632e-01 -8.20480347e-01 -6.34399876e-02 1.72489628e-01
-7.18498677e-02 -5.46180487e-01 -2.60909945e-01 1.17743537e-02
-2.69507259e-01 -7.30228603e-01 -7.47708023e-01 5.69820292e-02
-7.27458537e-01 -1.37070298e+00 -4.74677593e-01 -5.46465456e-01
9.91920412e-01 3.83921206e-01 1.32103717e+00 4.08387750e-01
1.25309005e-01 3.93630505e-01 -3.26697379e-01 -1.33262053e-02
1.18096322e-01 5.96801843e-03 -1.81418005e-02 3.27173807e-02
2.67781615e-01 -2.99686760e-01 -7.02766895e-01 6.02975249e-01
-1.08532858e+00 -2.78558701e-01 2.97324866e-01 7.42454767e-01
4.23872590e-01 5.86029962e-02 4.38412488e-01 -1.43974662e+00
8.43476593e-01 -6.53944492e-01 -9.28434789e-01 5.04592955e-01
-1.02850902e+00 1.16788439e-01 3.44199002e-01 -3.51754069e-01
-4.97223258e-01 -1.99646607e-01 4.56791580e-01 3.65889937e-01
5.75528622e-01 8.08756411e-01 -2.34569788e-01 -4.05454403e-03
4.88627017e-01 -4.33409899e-01 -6.01201989e-02 -3.28954309e-01
4.70130146e-01 6.80552185e-01 3.68802488e-01 -7.09682047e-01
6.80082500e-01 4.34599876e-01 3.57187241e-01 -7.98966065e-02
-1.04407477e+00 -6.99529648e-01 -1.92622393e-01 -1.59695148e-01
2.63328254e-01 -5.28609991e-01 -1.16275942e+00 -9.15690437e-02
-5.88273823e-01 4.20132168e-02 -6.50335699e-02 2.05648407e-01
-2.83466905e-01 5.18306434e-01 -3.02068204e-01 -9.03957129e-01
-1.98431700e-01 -7.73582876e-01 3.94159257e-01 1.29696965e-01
-2.12471321e-01 -7.77243972e-01 8.87104422e-02 5.98897219e-01
4.54169571e-01 2.27590457e-01 1.40009725e+00 -8.83320272e-01
-6.87138498e-01 -5.08317053e-01 -2.11136088e-01 -7.04699084e-02
-2.40915820e-01 -7.16162175e-02 -2.35017538e-01 -4.30196732e-01
-5.29948831e-01 -1.44823059e-01 6.94928169e-01 2.24136531e-01
8.58332157e-01 -7.75744617e-01 -2.99830347e-01 1.84678927e-01
1.37503338e+00 -1.17233083e-01 3.67948979e-01 4.59467918e-01
1.96289629e-01 9.39071238e-01 6.32777393e-01 5.97930074e-01
4.15262163e-01 9.41358805e-01 5.03189802e-01 -2.21812595e-02
5.43674767e-01 -5.61546125e-02 1.47053659e-01 4.80710924e-01
-2.59610917e-03 -4.17242736e-01 -4.32952523e-01 5.00409901e-01
-2.09177923e+00 -1.04265773e+00 -2.99950212e-01 3.00393081e+00
7.71916509e-01 1.42087772e-01 4.27093506e-01 7.23340809e-02
9.13513005e-01 -6.28238022e-02 -5.50054312e-01 -8.54627788e-01
-1.70880422e-01 9.41685662e-02 9.73486900e-01 7.51662791e-01
-6.75856471e-01 4.01567221e-01 6.45305777e+00 4.63292301e-01
-6.26098692e-01 -9.02750343e-02 6.55609131e-01 -5.75186133e-01
-1.00602877e+00 3.48306030e-01 -5.06369650e-01 5.13230801e-01
6.46648526e-01 -6.33513153e-01 8.42363656e-01 6.41266942e-01
-1.50264353e-01 -1.39157966e-01 -1.34515011e+00 4.92510647e-01
-3.26281786e-02 -9.14410710e-01 -1.26487374e-01 4.52231586e-01
1.08531630e+00 -3.28022480e-01 3.32856596e-01 1.04577579e-01
6.88750505e-01 -8.46967161e-01 5.90797961e-01 1.89004838e-01
7.14481771e-01 -1.01103127e+00 5.82138121e-01 1.95547611e-01
-7.26989329e-01 -3.76168340e-01 -5.57992578e-01 -1.47486538e-01
-7.29127973e-02 1.00869441e+00 -2.01028302e-01 5.57229459e-01
4.30739731e-01 -7.60265067e-03 -1.15030199e-01 1.30166364e+00
-9.58171040e-02 2.56923646e-01 -4.64183599e-01 -1.84497461e-01
1.23651832e-01 -2.55966395e-01 4.71122444e-01 6.21046722e-01
1.67879209e-01 4.03018117e-01 2.03543350e-01 3.27493429e-01
-5.61671793e-01 5.02986073e-01 -3.36652786e-01 1.05820470e-01
6.54402673e-01 1.32132292e+00 -6.53124392e-01 -1.98413938e-01
-3.67635995e-01 4.34871465e-01 2.93949485e-01 2.80718267e-01
-5.79745591e-01 -3.13023150e-01 6.26224279e-01 4.73404318e-01
6.10566000e-04 3.19054008e-01 -2.90351689e-01 -9.18263495e-01
1.06641725e-01 -8.99634719e-01 9.79569256e-01 -4.06045377e-01
-1.46767044e+00 2.66755164e-01 9.44588333e-02 -7.07018375e-01
-1.43929273e-01 -5.22058308e-01 -3.48490506e-01 8.22873890e-01
-1.51801872e+00 -4.31451827e-01 -8.91939923e-03 4.27170068e-01
-2.76598483e-01 2.39641696e-01 5.34957170e-01 4.70248312e-01
-2.82722026e-01 6.35330975e-01 4.13626105e-01 -3.47168803e-01
7.11232960e-01 -1.28518152e+00 -1.99898124e-01 6.74235284e-01
-1.12708434e-02 5.63853979e-01 8.06460261e-01 -4.27201509e-01
-1.32925236e+00 -5.63388884e-01 1.30888665e+00 -4.19233382e-01
5.00799000e-01 -2.30273888e-01 -4.00334060e-01 4.12984550e-01
-2.24660978e-01 -1.85916156e-01 7.99560010e-01 8.35072339e-01
-5.90291500e-01 -3.83096904e-01 -1.23655546e+00 6.07759237e-01
1.05807924e+00 -3.30037355e-01 -2.45955676e-01 4.79087800e-01
3.47843468e-01 -2.76995689e-01 -6.69292033e-01 4.04645711e-01
9.31036949e-01 -8.13758433e-01 7.98880816e-01 -1.00457549e+00
3.41201156e-01 -1.96040884e-01 -2.84920126e-01 -1.02137470e+00
-7.63056695e-01 -6.33456945e-01 9.24447253e-02 9.72466946e-01
7.61663139e-01 -6.14212215e-01 9.40968931e-01 1.40484607e+00
3.23491395e-01 -9.55090940e-01 -8.97372901e-01 -7.34571517e-01
-2.93308683e-02 3.28811765e-01 7.63890505e-01 8.87963831e-01
4.16015357e-01 2.59887815e-01 -5.27452469e-01 1.66984648e-01
6.29850268e-01 6.46782696e-01 5.36674917e-01 -1.71181333e+00
-4.43121731e-01 -6.51496828e-01 -1.48002818e-01 -8.51936877e-01
-9.73572116e-03 -1.11002111e+00 -2.33937845e-01 -1.65190458e+00
8.40062261e-01 -8.77346516e-01 -6.45081818e-01 5.65155864e-01
-1.62427053e-01 1.47647128e-01 2.25664645e-01 1.79295018e-01
-8.78161192e-01 -2.69918114e-01 9.03822601e-01 -1.51644960e-01
-1.86189175e-01 2.60199875e-01 -1.33185458e+00 3.07253063e-01
4.86505002e-01 -6.44552946e-01 -3.94097775e-01 -2.54892886e-01
1.26740849e+00 3.22193533e-01 9.59334988e-03 -1.04308009e-01
4.86180931e-01 -6.40592813e-01 1.30075244e-02 -2.99614877e-01
2.38431245e-02 -9.19901729e-01 4.03980225e-01 3.93683195e-01
-8.66906345e-01 1.86457470e-01 -3.60072583e-01 5.02188385e-01
3.51747386e-02 -4.06192124e-01 4.99049664e-01 1.28416553e-01
-2.96028405e-01 3.53749782e-01 -8.79531279e-02 2.36784726e-01
8.52535903e-01 -1.71340615e-01 -6.91028655e-01 -6.69153214e-01
-4.82694596e-01 6.05695367e-01 7.38734782e-01 2.93925524e-01
1.38902217e-01 -1.29196584e+00 -7.64497936e-01 -3.79533470e-01
2.63975829e-01 -4.07801569e-01 1.04426168e-01 8.27434778e-01
1.61213917e-03 3.48937452e-01 -8.31887275e-02 5.14661297e-02
-1.39324200e+00 7.20670223e-01 5.49390865e-03 -5.92966259e-01
1.28208727e-01 7.27074742e-01 7.75287226e-02 -2.45867193e-01
2.89671719e-01 2.70878166e-01 -1.33588150e-01 3.41963530e-01
2.16978446e-01 5.27895689e-01 1.75986394e-01 -4.09973681e-01
-6.02056682e-01 3.01459163e-01 -1.56414568e-01 -2.69624323e-01
1.28601456e+00 -1.39490351e-01 -5.14694095e-01 -9.86464918e-02
8.37263584e-01 6.08737350e-01 -4.66240764e-01 -3.18639487e-01
1.80645883e-01 -8.04695070e-01 -1.46820411e-01 -1.10418332e+00
-1.15957558e+00 6.58815131e-02 1.38242945e-01 5.35986781e-01
1.14896035e+00 3.24805826e-02 4.24818277e-01 3.79433602e-01
4.56004828e-01 -1.20864058e+00 -2.03537926e-01 3.14612240e-01
3.37746501e-01 -7.94547260e-01 3.15292120e-01 -3.78960103e-01
-4.69375223e-01 7.39013314e-01 3.68599564e-01 1.73215881e-01
4.50279862e-01 -4.89113592e-02 -1.48327038e-01 -1.47514939e-01
-8.99034500e-01 -2.38703519e-01 4.59944963e-01 5.37958927e-02
3.17701101e-01 3.16085994e-01 -9.65518355e-01 5.72415292e-01
-2.10123450e-01 -4.76168215e-01 5.36444366e-01 6.43473864e-01
-6.91821873e-01 -1.45297372e+00 -2.52464443e-01 9.25081015e-01
-7.66399145e-01 -1.80297330e-01 -7.86284149e-01 2.32993990e-01
-7.12608323e-02 1.22119999e+00 2.20927075e-02 -2.44269818e-01
4.87706512e-01 -3.20522517e-01 3.88719529e-01 -5.23696423e-01
-7.74579883e-01 -2.78785616e-01 2.82726139e-01 -4.07515645e-01
-2.26187482e-01 -5.13359606e-01 -9.09604132e-01 -7.90265858e-01
-7.57595003e-01 6.34129941e-01 4.59572405e-01 6.93164885e-01
3.13338101e-01 -8.64576921e-02 9.66558099e-01 -6.05849251e-02
-8.60376656e-01 -3.70966494e-01 -9.36514497e-01 7.50128746e-01
-2.69815996e-02 -4.10826147e-01 -4.96303856e-01 -5.99905550e-01] | [9.337355613708496, 5.5633745193481445] |
65ed2fb0-a294-4db2-843e-32ecda13271b | efficient-reinforcement-learning-for | 2204.07696 | null | https://arxiv.org/abs/2204.07696v1 | https://arxiv.org/pdf/2204.07696v1.pdf | Efficient Reinforcement Learning for Unsupervised Controlled Text Generation | Controlled text generation tasks such as unsupervised text style transfer have increasingly adopted the use of Reinforcement Learning (RL). A major challenge in applying RL to such tasks is the sparse reward, which is available only after the full text is generated. Sparse rewards, combined with a large action space make RL training sample-inefficient and difficult to converge. Recently proposed reward-shaping strategies to address this issue have shown only negligible gains. In contrast, this work proposes a novel approach that provides dense rewards to each generated token. We evaluate our approach by its usage in unsupervised text style transfer. Averaged across datasets, our style transfer system improves upon current state-of-art systems by 21\% on human evaluation and 12\% on automatic evaluation. Upon ablated comparison with the current reward shaping approach (the `roll-out strategy'), using dense rewards improves the overall style transfer quality by 22\% based on human evaluation. Further the RL training is 2.5 times as sample efficient, and 7 times faster. | ['Arjun Maheswaran', 'Akhilesh Sudhakar', 'Bhargav Upadhyay'] | 2022-04-16 | null | null | null | null | ['text-style-transfoer'] | ['natural-language-processing'] | [ 5.55254400e-01 4.16562021e-01 -1.75105378e-01 -1.62036002e-01
-1.21181822e+00 -5.87852716e-01 9.57755864e-01 -2.88838521e-02
-8.00102055e-01 1.33882177e+00 2.98756182e-01 -1.32830560e-01
3.60675365e-01 -6.47554338e-01 -5.74282646e-01 -4.28543180e-01
3.14346224e-01 7.87857711e-01 1.12870179e-01 -4.71303076e-01
4.51246113e-01 4.95518520e-02 -1.38687563e+00 1.75162628e-01
9.90720689e-01 6.60786211e-01 4.11305815e-01 8.39746058e-01
-4.89163607e-01 9.50335681e-01 -8.26158702e-01 -2.10673541e-01
5.36459573e-02 -7.38755643e-01 -1.03481340e+00 -1.07393734e-01
3.52814108e-01 -6.97650969e-01 3.53047997e-02 6.26608908e-01
7.67886460e-01 4.37410474e-01 6.84136033e-01 -1.00530481e+00
-7.51333654e-01 8.51093829e-01 -4.33502495e-01 -1.85830131e-01
4.54441369e-01 1.16129704e-01 1.07655454e+00 -7.42600560e-01
8.90295446e-01 1.26509559e+00 4.14335221e-01 1.10868061e+00
-1.46351564e+00 -6.28303528e-01 2.34223772e-02 -2.25820974e-01
-6.91525459e-01 -5.60346842e-01 5.04765987e-01 -1.21676803e-01
1.21237695e+00 8.57169032e-02 2.94345945e-01 1.17770290e+00
-7.80759305e-02 1.31703782e+00 1.44779849e+00 -7.33136177e-01
4.08753812e-01 2.90886164e-01 -4.31066811e-01 4.61018234e-01
-7.53694773e-02 2.81739086e-01 -6.20541751e-01 9.42080747e-03
8.23890448e-01 -4.43028122e-01 1.60901576e-01 -3.96707617e-02
-1.32897842e+00 9.86984432e-01 3.64706218e-01 2.32584938e-01
-2.55586714e-01 5.69445133e-01 5.37184179e-01 5.37174165e-01
5.90808094e-01 8.80549848e-01 -3.24763060e-01 -7.72742689e-01
-1.07219052e+00 5.71092546e-01 8.13470125e-01 1.20315456e+00
8.29559922e-01 3.97303700e-01 -6.25195086e-01 1.20008671e+00
9.69181024e-03 6.05972111e-01 6.95870399e-01 -1.10022306e+00
5.95490158e-01 3.62590492e-01 4.53712463e-01 -2.06061304e-01
-7.49859735e-02 -9.91681963e-02 -6.63507342e-01 5.91488123e-01
6.54304266e-01 -5.76172531e-01 -8.44307244e-01 1.76112962e+00
1.14826813e-01 -3.56539547e-01 2.60151252e-02 6.00728691e-01
4.03760284e-01 6.12125099e-01 2.80191600e-01 -2.05135778e-01
9.90515172e-01 -1.00501621e+00 -7.05150962e-01 -1.44098431e-01
7.98289061e-01 -1.05326450e+00 1.40721738e+00 4.54660833e-01
-1.32524621e+00 -4.90694821e-01 -7.87248552e-01 1.66411959e-02
-2.03862756e-01 1.78000987e-01 5.77446163e-01 7.45938540e-01
-1.25996363e+00 9.63463008e-01 -4.65522528e-01 -3.81196499e-01
6.16966903e-01 5.38973927e-01 -1.22342326e-01 6.48576021e-03
-1.02189040e+00 8.89716804e-01 2.13274047e-01 -3.87026101e-01
-7.91525245e-01 -7.06565559e-01 -6.52495146e-01 -5.96439233e-03
3.19997668e-01 -5.39654613e-01 1.84009922e+00 -1.17983115e+00
-2.39876580e+00 5.68697512e-01 -1.57347262e-01 -5.00158787e-01
8.56384456e-01 -6.28149509e-01 -1.13296183e-02 6.86155558e-02
2.45044842e-01 1.27899754e+00 8.70739639e-01 -1.11097789e+00
-6.94224834e-01 2.25508839e-01 -1.40368463e-02 5.26496410e-01
-4.11516190e-01 1.29663706e-01 2.52168206e-03 -8.97744298e-01
-7.36475348e-01 -1.07383013e+00 -3.01121235e-01 -1.20384336e-01
-7.88963288e-02 -5.00671029e-01 6.27335846e-01 -4.17348891e-01
9.17010725e-01 -1.78364730e+00 2.01346174e-01 -1.46030441e-01
-3.62136289e-02 3.89636219e-01 -2.65880316e-01 5.47409892e-01
2.92618155e-01 1.02599241e-01 -2.67128080e-01 -6.75706685e-01
2.11557195e-01 2.72907495e-01 -3.57166171e-01 -1.27314508e-01
4.36578602e-01 1.01557529e+00 -1.24478674e+00 -4.14626002e-01
7.60545442e-03 5.27669489e-02 -7.87860334e-01 4.27024096e-01
-5.56699336e-01 3.06069195e-01 -3.03825915e-01 2.96054959e-01
2.33476952e-01 -1.42028049e-01 1.33938283e-01 5.12350440e-01
-3.53231467e-02 4.21814650e-01 -1.07006001e+00 2.04560113e+00
-6.99598253e-01 7.04718947e-01 -2.20355242e-01 -4.81395245e-01
1.13651168e+00 5.30014694e-01 2.49249384e-01 -7.71158874e-01
-1.68397964e-03 3.01991701e-01 -1.31470010e-01 7.44003952e-02
1.00490034e+00 -3.05503935e-01 -4.60232720e-02 9.43101406e-01
2.58508384e-01 -4.60280389e-01 2.86623567e-01 3.57949138e-01
1.24969327e+00 7.52749264e-01 2.48717498e-02 -3.32545638e-01
3.96898896e-01 2.95071979e-03 1.72170192e-01 9.58519578e-01
-1.70007616e-01 5.75590134e-01 5.67267418e-01 -1.31538168e-01
-1.19358075e+00 -8.66070390e-01 3.35645944e-01 1.55834997e+00
-3.55486989e-01 -3.43904525e-01 -9.23473895e-01 -1.02047241e+00
-5.47336116e-02 9.31943834e-01 -6.34857178e-01 4.64946814e-02
-7.45504737e-01 -3.91298890e-01 6.81855321e-01 6.43434584e-01
5.09212315e-01 -1.86686671e+00 -6.30665720e-01 6.00162685e-01
-1.57999307e-01 -8.01487029e-01 -5.81349432e-01 1.59716487e-01
-9.78806019e-01 -4.34976339e-01 -1.29686391e+00 -5.94593644e-01
6.07315838e-01 -1.49647772e-01 1.31736147e+00 -1.09541394e-01
-1.71411932e-01 3.21803600e-01 -4.99314100e-01 -5.66877723e-01
-7.16289818e-01 3.90984923e-01 -5.48079349e-02 -3.94025087e-01
2.91777831e-02 -3.78744543e-01 -6.03401005e-01 1.57857284e-01
-7.68616199e-01 -5.25490381e-02 7.36835420e-01 1.17684352e+00
2.54513115e-01 -5.77820063e-01 1.18773592e+00 -1.24557805e+00
1.09265649e+00 -1.61822096e-01 -4.05579895e-01 -3.49862198e-03
-9.50975358e-01 4.98219401e-01 8.52419317e-01 -4.77489024e-01
-1.37516820e+00 2.48741098e-02 -6.26454577e-02 -2.04203218e-01
-2.68864304e-01 -7.23825693e-02 4.16225553e-01 2.31304362e-01
9.92499828e-01 1.82028010e-01 1.86657101e-01 -2.82613993e-01
6.42559409e-01 5.75095773e-01 6.72783628e-02 -8.50522876e-01
6.21012270e-01 1.14714891e-01 -2.09533751e-01 -6.42747760e-01
-5.93129039e-01 -2.59903282e-01 -3.97401243e-01 -1.19103171e-01
6.88267410e-01 -6.89041376e-01 -5.09497702e-01 3.34395498e-01
-1.00798726e+00 -1.10108113e+00 -7.22115755e-01 3.09970677e-01
-1.09889698e+00 3.03905576e-01 -6.91773236e-01 -9.69319522e-01
-7.01064467e-01 -9.69205081e-01 1.14129221e+00 1.12983808e-01
-6.52576983e-01 -9.28692222e-01 2.97587961e-01 2.06448436e-01
7.47563839e-01 -6.70756251e-02 6.55162156e-01 -3.34859014e-01
-2.56210744e-01 1.46045849e-01 -2.73019373e-01 4.01963025e-01
2.16232494e-01 -3.51859808e-01 -1.09393680e+00 -3.53266627e-01
-5.82120717e-01 -9.03694630e-01 6.40670478e-01 1.29680470e-01
7.79454768e-01 -1.73866689e-01 1.17598198e-01 -2.42689289e-02
1.11019135e+00 6.12382777e-02 6.29711390e-01 3.00011098e-01
6.04808569e-01 6.43977821e-01 8.72458279e-01 4.36108917e-01
8.76837149e-02 6.65883362e-01 1.40142790e-03 -1.18082307e-01
-3.50025803e-01 -4.12238389e-01 7.07426012e-01 5.78128874e-01
-2.99114704e-01 -1.73715591e-01 -4.31993097e-01 5.12409270e-01
-1.87156320e+00 -9.90202487e-01 2.75296330e-01 2.09118128e+00
1.25785351e+00 3.34891528e-01 4.48328584e-01 2.42338441e-02
3.86104971e-01 -1.15949407e-01 -4.92953807e-01 -9.00182545e-01
2.32443139e-01 6.53523624e-01 3.86745006e-01 6.57793224e-01
-7.05235362e-01 1.37426805e+00 6.92998314e+00 9.89239991e-01
-9.75149512e-01 -1.51331291e-01 5.82919717e-01 -2.18035907e-01
-3.41524839e-01 -8.57596621e-02 -7.93838739e-01 3.93700480e-01
9.69290912e-01 -1.54600218e-01 8.29335511e-01 8.68063092e-01
1.69671863e-01 -1.83809742e-01 -1.07680178e+00 7.63046503e-01
-8.37890208e-02 -1.16243052e+00 8.12627971e-02 5.08570075e-02
9.57552135e-01 -6.15236983e-02 1.78574696e-02 8.02948773e-01
8.26934576e-01 -9.78590667e-01 5.86160541e-01 3.19948465e-01
1.14025867e+00 -1.01014102e+00 4.95404780e-01 2.51296461e-01
-9.01700258e-01 2.20133767e-01 -3.87778252e-01 -2.34835014e-01
8.84332284e-02 1.90561593e-01 -1.25775516e+00 3.52129698e-01
3.64435017e-01 6.36359274e-01 -4.48629707e-01 4.74746883e-01
-4.79723245e-01 7.87033975e-01 -7.59159327e-02 -5.19530773e-01
4.72518355e-01 -3.41871828e-02 4.11891073e-01 1.36324298e+00
1.66070253e-01 -1.80660844e-01 1.03387855e-01 9.41102743e-01
-2.40840241e-01 2.95601696e-01 -7.71220386e-01 -7.78922858e-03
2.18483791e-01 1.30574989e+00 -6.98643327e-01 -6.03928447e-01
-1.16414847e-02 1.40024865e+00 6.50901318e-01 2.39963800e-01
-5.88975966e-01 -7.32774258e-01 2.96640724e-01 -1.56928599e-01
4.22063500e-01 -1.10955834e-01 -3.17416728e-01 -8.56528103e-01
-1.64899483e-01 -9.94108617e-01 1.12236343e-01 -7.16973424e-01
-1.21800232e+00 7.06868708e-01 -2.00118199e-02 -1.13886225e+00
-8.88726652e-01 -4.92306739e-01 -5.43120861e-01 1.00656021e+00
-1.48108375e+00 -8.92980814e-01 -1.02980584e-02 4.90497977e-01
9.12024558e-01 -4.24349129e-01 1.12916851e+00 1.26447082e-02
-1.48426920e-01 9.35067356e-01 1.79704502e-01 -1.21179700e-01
1.04643822e+00 -1.87176144e+00 6.04315162e-01 4.40639973e-01
3.69278044e-02 3.88912350e-01 5.89003861e-01 -6.23877048e-01
-1.14117372e+00 -9.19985294e-01 9.00946796e-01 -5.26424468e-01
4.92263228e-01 -5.38547337e-01 -6.96567893e-01 3.65436792e-01
6.62533879e-01 -4.28357333e-01 5.01994431e-01 2.82087117e-01
-1.47324100e-01 1.20997215e-02 -1.27600634e+00 8.55908215e-01
9.09514785e-01 -3.07102978e-01 -5.49915910e-01 2.25886062e-01
6.53411090e-01 -2.53767639e-01 -8.43389392e-01 -1.03164189e-01
4.26906705e-01 -8.10526252e-01 5.84268749e-01 -4.61285681e-01
7.63779581e-01 9.72314999e-02 1.75188601e-01 -1.65030181e+00
-3.05216849e-01 -1.15255511e+00 -2.78371684e-02 1.31706786e+00
4.72724676e-01 -4.65126932e-01 9.88227487e-01 4.86680269e-01
-4.14043628e-02 -6.28235459e-01 -6.57269895e-01 -7.88071275e-01
4.31292295e-01 -1.57170877e-01 3.25497091e-01 7.10809827e-01
2.88816720e-01 6.51666820e-01 -5.79597116e-01 -7.60296106e-01
5.62421322e-01 -3.09522059e-02 9.12248075e-01 -8.96780252e-01
-6.57441020e-01 -6.79386199e-01 2.30793491e-01 -1.15980244e+00
1.46920845e-01 -8.73929024e-01 3.37153852e-01 -1.49871540e+00
-1.60905674e-01 -7.25588918e-01 -2.32911795e-01 5.40942192e-01
-3.83807003e-01 2.59754270e-01 4.72903073e-01 9.63213369e-02
-7.32232809e-01 7.39516735e-01 1.48754299e+00 4.59675975e-02
-3.87902349e-01 -6.36322051e-02 -6.77760184e-01 3.45940322e-01
1.10674953e+00 -4.70405012e-01 -4.76268649e-01 -3.70418668e-01
1.47169992e-01 8.59808102e-02 -2.02158719e-01 -8.32371593e-01
-1.04838617e-01 -1.38282329e-01 4.07315075e-01 -3.27658564e-01
2.60611057e-01 -4.04521912e-01 -4.94954228e-01 4.75055844e-01
-7.70129204e-01 1.02534138e-01 4.01108980e-01 5.19960463e-01
1.01641141e-01 -3.76115799e-01 7.46566951e-01 -3.28910530e-01
-4.04984564e-01 -3.80108841e-02 -6.62390292e-01 3.39399219e-01
7.75484264e-01 -3.78647819e-02 -3.14811245e-02 -6.74052238e-01
-5.49205303e-01 1.10110641e-01 2.52287745e-01 3.84569257e-01
4.65817541e-01 -1.40932834e+00 -8.50521028e-01 3.53079811e-02
-9.76790190e-02 -5.32356538e-02 -1.77230269e-01 2.97769934e-01
-3.15049499e-01 3.77129793e-01 -4.33757812e-01 -3.69088948e-01
-1.06910288e+00 2.22098514e-01 -3.51625420e-02 -7.45634615e-01
-6.16072059e-01 6.81204557e-01 -1.48636568e-02 -4.97957647e-01
2.32175693e-01 -3.45318586e-01 -1.15432896e-01 -4.58459966e-02
4.98456389e-01 5.85902750e-01 -8.40636641e-02 -3.55237424e-02
1.64547309e-01 1.97683781e-01 -4.83568937e-01 -7.83825099e-01
1.19806337e+00 2.02263385e-01 2.73925513e-01 4.53043848e-01
8.17867637e-01 1.54811861e-02 -1.59612155e+00 -1.71264648e-01
-2.42940206e-02 -3.90027285e-01 -2.04659179e-01 -1.11215198e+00
-5.25600612e-01 8.04190814e-01 4.42628026e-01 2.06331015e-01
7.64757931e-01 -3.98665100e-01 7.32434273e-01 6.71439648e-01
3.82329613e-01 -1.42622244e+00 5.24567902e-01 6.71762109e-01
9.12675142e-01 -1.27053499e+00 -6.34340271e-02 1.66933257e-02
-1.06182492e+00 1.11025107e+00 7.06882834e-01 -4.40092653e-01
7.10702017e-02 1.04334429e-02 8.52923095e-02 3.23267758e-01
-8.06154013e-01 -6.76352829e-02 2.95223854e-02 6.03349805e-01
8.14549744e-01 7.03347698e-02 -2.97686338e-01 -4.80319522e-02
-3.06895554e-01 2.12534457e-01 5.56674957e-01 1.07925224e+00
-4.03979421e-01 -1.75215542e+00 -1.25872970e-01 5.54352462e-01
-5.02853572e-01 -1.97321758e-01 -3.58133227e-01 6.84892297e-01
-3.39244723e-01 1.01343560e+00 -6.91800714e-02 7.18252510e-02
3.24095100e-01 2.25266397e-01 7.06494629e-01 -7.76998460e-01
-1.00291765e+00 1.47628024e-01 2.68594354e-01 -4.55973387e-01
-2.91186243e-01 -6.65804863e-01 -1.34009266e+00 -2.79164582e-01
-3.36195409e-01 1.45923197e-01 5.88326693e-01 8.15803766e-01
2.95906603e-01 5.58106303e-01 7.13005126e-01 -9.84054267e-01
-9.55815017e-01 -1.15871239e+00 -4.33688700e-01 4.83478069e-01
1.10766165e-01 -4.76212174e-01 -4.10452634e-02 1.07093677e-01] | [11.766336441040039, 9.081097602844238] |
dfd771c3-8be7-41af-972a-4cf181415df1 | learning-a-discriminative-prior-for-blind | 1803.03363 | null | http://arxiv.org/abs/1803.03363v2 | http://arxiv.org/pdf/1803.03363v2.pdf | Learning a Discriminative Prior for Blind Image Deblurring | We present an effective blind image deblurring method based on a data-driven
discriminative prior.Our work is motivated by the fact that a good image prior
should favor clear images over blurred images.In this work, we formulate the
image prior as a binary classifier which can be achieved by a deep
convolutional neural network (CNN).The learned prior is able to distinguish
whether an input image is clear or not.Embedded into the maximum a posterior
(MAP) framework, it helps blind deblurring in various scenarios, including
natural, face, text, and low-illumination images.However, it is difficult to
optimize the deblurring method with the learned image prior as it involves a
non-linear CNN.Therefore, we develop an efficient numerical approach based on
the half-quadratic splitting method and gradient decent algorithm to solve the
proposed model.Furthermore, the proposed model can be easily extended to
non-uniform deblurring.Both qualitative and quantitative experimental results
show that our method performs favorably against state-of-the-art algorithms as
well as domain-specific image deblurring approaches. | ['Ming-Hsuan Yang', 'Wei-Sheng Lai', 'Lerenhan Li', 'Nong Sang', 'Jinshan Pan', 'Changxin Gao'] | 2018-03-09 | learning-a-discriminative-prior-for-blind-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Li_Learning_a_Discriminative_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Li_Learning_a_Discriminative_CVPR_2018_paper.pdf | cvpr-2018-6 | ['blind-image-deblurring'] | ['computer-vision'] | [ 3.16565007e-01 -3.69030297e-01 -1.43058309e-02 -3.11869979e-01
-6.05644584e-01 -3.07688296e-01 5.28963447e-01 -6.56970501e-01
-2.89451480e-01 8.20508003e-01 2.61119336e-01 -1.06232494e-01
-2.74088569e-02 -3.41854990e-01 -6.68732345e-01 -1.11399448e+00
5.17985225e-01 6.99782930e-03 -1.08984850e-01 6.59308881e-02
5.10435760e-01 4.83871728e-01 -1.25815129e+00 -9.17496085e-02
1.21364844e+00 9.44956243e-01 3.20335776e-01 5.53428411e-01
2.97131300e-01 8.33772898e-01 -3.96386445e-01 -6.95885420e-02
1.69882998e-01 -5.76588690e-01 -8.55681300e-01 2.12461650e-01
6.84757113e-01 -7.51569808e-01 -5.13036311e-01 1.49309611e+00
4.39852208e-01 1.18985735e-01 8.41405690e-01 -9.07831490e-01
-1.19701171e+00 7.79380649e-02 -8.91425073e-01 2.64221847e-01
3.61050628e-02 2.17878088e-01 5.18406272e-01 -9.48896050e-01
3.34485143e-01 1.14618325e+00 5.46534896e-01 4.09303159e-01
-1.04265392e+00 -5.85756004e-01 -1.26199186e-01 5.04504561e-01
-1.28967583e+00 -6.64253473e-01 9.40962434e-01 -5.39300740e-01
2.00966120e-01 3.02317172e-01 3.02813292e-01 8.52397919e-01
1.80592358e-01 8.31465065e-01 1.43357635e+00 -3.76213342e-01
1.90180808e-01 -8.59624296e-02 2.01642558e-01 5.70430517e-01
3.50925714e-01 2.51868993e-01 -3.69027376e-01 -4.56220135e-02
8.57548237e-01 1.26936793e-01 -1.00679374e+00 -2.62998134e-01
-1.07099867e+00 5.34101069e-01 5.67260981e-01 2.43878320e-01
-3.67502391e-01 8.10023844e-02 -8.81967247e-02 3.18869650e-02
7.80659735e-01 2.90197372e-01 -1.57918155e-01 1.92621246e-01
-1.48537457e+00 6.10619970e-02 5.45109451e-01 3.71240854e-01
9.38959777e-01 6.15219250e-02 -2.52108693e-01 1.04911733e+00
5.25922716e-01 6.48970664e-01 5.27877331e-01 -9.68126953e-01
1.02373317e-01 -2.62040738e-02 5.29614091e-01 -1.22786784e+00
8.07659104e-02 -4.47920859e-01 -1.32967305e+00 3.03939372e-01
3.91077548e-01 -1.91648737e-01 -9.41853464e-01 1.55447495e+00
1.64185375e-01 4.14253980e-01 1.50386408e-01 1.54160857e+00
6.41242087e-01 7.51959562e-01 -4.69988972e-01 -4.55555975e-01
1.21272039e+00 -1.23552382e+00 -1.05628550e+00 -4.10807222e-01
-4.36126068e-02 -8.94847214e-01 6.58134222e-01 4.09123629e-01
-1.10680211e+00 -5.54487407e-01 -1.22223008e+00 -2.22438172e-01
-3.98706906e-02 5.58801472e-01 2.54996181e-01 6.11914933e-01
-1.14008749e+00 4.61125880e-01 -7.55665958e-01 -3.37862670e-01
4.25883442e-01 4.03635241e-02 -2.60930687e-01 -4.70874250e-01
-1.14812803e+00 1.05800021e+00 2.60020137e-01 5.73015332e-01
-1.11337817e+00 -4.18584317e-01 -8.06332350e-01 1.13149963e-01
4.09341678e-02 -7.33510077e-01 9.54356015e-01 -1.18243146e+00
-1.83674049e+00 7.56659806e-01 -3.82239491e-01 -1.18077092e-01
6.26789093e-01 -4.77570206e-01 -1.91970170e-01 3.16074610e-01
-1.25847057e-01 4.55670655e-01 1.46813428e+00 -1.47260034e+00
-2.06194296e-01 -2.38923594e-01 -1.17302090e-01 2.55630314e-01
-3.64321530e-01 2.58988962e-02 -3.87123495e-01 -8.95211756e-01
1.75820842e-01 -7.56292105e-01 8.80639628e-02 2.15741724e-01
-4.82693881e-01 3.36618066e-01 1.06351185e+00 -1.23474526e+00
9.58095431e-01 -1.99472284e+00 3.11833709e-01 -2.57328451e-01
2.04598919e-01 3.24992001e-01 -4.45915423e-02 8.56115967e-02
-2.56615072e-01 -1.63238496e-01 -6.86692238e-01 -3.43942374e-01
-1.95559651e-01 -1.05747238e-01 -3.27290624e-01 1.10676312e+00
5.00043482e-02 8.55012476e-01 -7.46554136e-01 -2.02398330e-01
2.52338976e-01 6.84228301e-01 -3.32574755e-01 5.97248673e-01
8.13624561e-02 6.60640240e-01 -1.32523075e-01 4.79694098e-01
1.39504552e+00 -4.30190980e-01 -1.10214800e-01 -5.75949371e-01
-2.21862078e-01 -2.50390947e-01 -9.68699813e-01 1.71940076e+00
-4.26748455e-01 1.01308620e+00 6.31571054e-01 -1.20028365e+00
7.12872267e-01 1.13901705e-01 1.54694811e-01 -2.19815433e-01
2.03474060e-01 3.19929570e-01 -4.17521238e-01 -8.08023751e-01
5.81235588e-01 -1.18132524e-01 6.38571620e-01 5.56984246e-01
-2.52457619e-01 -1.95079848e-01 -1.31969258e-01 -4.55172844e-02
6.50428355e-01 5.88775799e-02 5.60729280e-02 -4.65335637e-01
8.09681833e-01 -2.75572866e-01 3.10543776e-01 6.27976239e-01
-1.55100524e-01 1.09077382e+00 1.05518224e-02 -2.49344274e-01
-7.89131582e-01 -6.09671593e-01 -3.20002168e-01 5.32772422e-01
6.40088379e-01 2.97477931e-01 -1.00296760e+00 -4.94675964e-01
-4.18440104e-01 4.99187201e-01 -4.77310300e-01 -8.03896636e-02
-4.39196676e-01 -1.19282901e+00 2.82462269e-01 1.08880252e-01
1.09459341e+00 -6.37189507e-01 -1.48758590e-01 -1.35610878e-01
-5.93319893e-01 -1.04286814e+00 -9.04285252e-01 -2.17632428e-01
-6.60089374e-01 -1.02735567e+00 -1.60045123e+00 -9.55151081e-01
9.21302199e-01 7.95802414e-01 6.52888238e-01 -2.18344815e-02
-2.85289064e-03 2.75581926e-01 -2.62412816e-01 3.85172665e-01
-4.16645885e-01 -2.08685279e-01 -1.29604474e-01 5.00061274e-01
2.40146071e-02 -5.70107460e-01 -9.71854925e-01 4.77380186e-01
-1.16037321e+00 3.28461796e-01 7.39143193e-01 1.05602288e+00
1.45535097e-01 1.69181332e-01 3.23610008e-01 -3.02012652e-01
7.70237684e-01 -4.43323016e-01 -5.38296103e-01 4.27854717e-01
-8.40624809e-01 1.62535354e-01 3.77361119e-01 -4.84641463e-01
-1.35332298e+00 -7.80472681e-02 1.97779357e-01 -5.98211169e-01
-2.21282899e-01 4.74148452e-01 -1.53934181e-01 -3.84873271e-01
5.83809555e-01 6.50104046e-01 1.58237264e-01 -6.48359120e-01
5.29010952e-01 1.08070230e+00 7.97470868e-01 -4.84251857e-01
9.53018606e-01 5.81972778e-01 -1.99477419e-01 -8.73026848e-01
-6.79618061e-01 -3.57320160e-01 -3.99314493e-01 -2.24059895e-01
1.01176715e+00 -9.57844138e-01 -5.20850718e-01 1.16845489e+00
-1.55693376e+00 -2.48981059e-01 4.61164296e-01 5.44642687e-01
-2.96961397e-01 1.04072475e+00 -8.25541139e-01 -6.69406712e-01
-4.51016754e-01 -1.36779630e+00 9.59659696e-01 3.93308878e-01
4.00613815e-01 -9.79612648e-01 1.13228880e-01 5.94540238e-01
7.68703043e-01 -1.38194650e-01 5.64544559e-01 -8.85524526e-02
-7.18799412e-01 -6.59277812e-02 -8.19861233e-01 6.96672261e-01
4.20984507e-01 -2.49538705e-01 -1.12023640e+00 -6.02760851e-01
3.84539634e-01 -3.43944490e-01 1.08579671e+00 6.03716612e-01
1.23174930e+00 -6.15227520e-01 -1.84822440e-01 9.03797925e-01
1.35702825e+00 -2.10294083e-01 8.10621440e-01 2.55780518e-01
8.32448483e-01 3.62925082e-01 3.04153174e-01 1.60986960e-01
2.19477698e-01 7.62268782e-01 5.59087992e-01 -3.59754473e-01
-3.60023886e-01 1.22918010e-01 4.55773264e-01 6.76334143e-01
3.75266820e-02 -3.69348735e-01 -5.94579101e-01 5.29950917e-01
-1.86547339e+00 -8.00032973e-01 -1.16794720e-01 2.07139587e+00
1.16791010e+00 -5.34823358e-01 -6.31223202e-01 -1.41733423e-01
1.08177948e+00 4.28118855e-01 -6.15115106e-01 1.16670318e-01
-2.32956305e-01 -1.54861584e-01 4.84215587e-01 8.12104583e-01
-1.21509004e+00 8.41974974e-01 5.91395235e+00 9.61159170e-01
-1.43822622e+00 1.57987997e-01 9.06838775e-01 3.68045807e-01
-2.01941743e-01 -5.47921881e-02 -3.41511518e-01 6.43851817e-01
4.04174358e-01 -1.19009335e-02 8.29696655e-01 5.24626434e-01
4.21636462e-01 -2.52883583e-01 -8.29704165e-01 1.48566389e+00
4.69806880e-01 -1.15239298e+00 -1.61629215e-01 -8.47614259e-02
1.06421793e+00 -1.33666366e-01 2.92809099e-01 -2.58611888e-01
1.05071291e-01 -1.11799645e+00 6.74062252e-01 6.26790106e-01
8.45311463e-01 -1.65863052e-01 6.94736004e-01 4.08277601e-01
-6.31748199e-01 2.09039629e-01 -3.29343170e-01 8.77416655e-02
1.55222580e-01 9.99187827e-01 -2.37076640e-01 6.88614249e-01
6.87201679e-01 1.17308450e+00 -2.31774211e-01 1.12707889e+00
-3.93521070e-01 6.07390344e-01 2.75195893e-02 4.61083114e-01
1.11851124e-02 -5.23558497e-01 6.68022394e-01 1.11892962e+00
5.33895195e-01 1.79031610e-01 -2.97253519e-01 1.11172032e+00
-1.48700356e-01 -1.76874220e-01 -1.70843586e-01 1.03021219e-01
1.24343559e-01 1.22090828e+00 -4.20957357e-01 -2.75015533e-01
-2.03410730e-01 1.53230906e+00 2.41380418e-03 9.32954609e-01
-6.94331110e-01 -3.48562270e-01 5.95949292e-01 -1.60343677e-01
3.46116424e-01 -2.25313738e-01 -3.30241561e-01 -1.72931719e+00
-1.96190719e-02 -9.26302612e-01 -1.61722779e-01 -1.29513896e+00
-1.41622841e+00 6.25819683e-01 -9.87286121e-02 -1.29418433e+00
1.28642172e-01 -6.38515890e-01 -5.71626306e-01 1.26085186e+00
-2.16168833e+00 -9.47124422e-01 -8.24179590e-01 5.00678360e-01
4.54959005e-01 7.00796545e-02 3.11391950e-01 3.13893914e-01
-7.79135108e-01 3.38404953e-01 5.66927195e-01 4.21487838e-02
1.10359979e+00 -8.47904265e-01 -2.80480552e-02 1.30242908e+00
-4.38872546e-01 6.00948930e-01 8.71324062e-01 -4.58334863e-01
-1.27827418e+00 -9.71040845e-01 3.21026862e-01 1.17325395e-01
4.21052605e-01 1.94675073e-01 -1.13419449e+00 3.04388165e-01
7.27644145e-01 2.47388408e-01 1.38900638e-01 -4.86059248e-01
-2.90609986e-01 -1.44472882e-01 -1.22607851e+00 3.04698437e-01
5.20245612e-01 -4.51126873e-01 -5.24680912e-01 2.40054712e-01
3.80870908e-01 -4.63545293e-01 -4.50263113e-01 3.45845371e-01
3.53320211e-01 -9.81455147e-01 9.26689923e-01 -1.26033843e-01
8.23851883e-01 -5.24645329e-01 6.65224120e-02 -1.60413349e+00
-3.92919838e-01 -7.32608140e-01 -2.88973451e-01 1.04447520e+00
-5.23417890e-02 -8.03587139e-01 4.40218329e-01 4.50774550e-01
-2.31581982e-02 -3.52602839e-01 -7.23939180e-01 -5.70946634e-01
5.62794097e-02 7.44437724e-02 3.56150210e-01 1.12766898e+00
-3.24563116e-01 1.74542107e-02 -9.31907237e-01 5.21808565e-01
9.72658098e-01 2.40114659e-01 4.46977794e-01 -8.89807522e-01
-2.08485305e-01 -5.02447784e-01 -1.27662381e-03 -1.87154388e+00
2.63085574e-01 -6.29406154e-01 3.89771372e-01 -1.54384577e+00
5.04540861e-01 -2.42504925e-01 -1.49188444e-01 3.94111127e-01
-4.86201704e-01 3.23311925e-01 -1.85727254e-01 6.82004154e-01
-1.98551074e-01 9.38526392e-01 1.49892986e+00 -5.63654244e-01
1.02937244e-01 -1.53009221e-01 -6.68938994e-01 5.24624407e-01
6.77029073e-01 -3.36336881e-01 -1.78647831e-01 -7.84428775e-01
1.48515683e-02 1.12515613e-01 5.88629484e-01 -7.93329060e-01
3.84177148e-01 -2.87005574e-01 3.22028905e-01 -3.09789479e-01
3.24158639e-01 -7.13659883e-01 -3.66257280e-02 2.78967291e-01
-2.72703558e-01 -5.54496050e-01 -7.97642767e-03 5.82536519e-01
-4.46762562e-01 -3.59311044e-01 1.16646433e+00 4.85749356e-02
-3.56521517e-01 2.55903602e-01 -1.98849797e-01 -1.14288539e-01
5.47410488e-01 -1.51939824e-01 -6.42692983e-01 -6.58549786e-01
-2.78616130e-01 -7.32847825e-02 6.37859166e-01 2.19024271e-01
6.46191418e-01 -1.23980165e+00 -8.39162171e-01 1.93467185e-01
-2.12750614e-01 -1.18718080e-01 3.81779194e-01 1.04149389e+00
-7.06931353e-01 2.09003955e-01 -2.22961292e-01 -5.18862128e-01
-1.08591390e+00 4.98588383e-01 6.36887848e-01 1.73022524e-01
-2.88467973e-01 6.22954130e-01 4.10201699e-01 -6.63406327e-02
1.27815127e-01 -1.62710860e-01 -2.45801494e-01 -3.01406384e-01
7.71914244e-01 5.83148479e-01 -4.06099036e-02 -7.55869031e-01
-1.31304070e-01 8.83011818e-01 -7.55899325e-02 -1.64885268e-01
1.16400456e+00 -4.85660344e-01 -6.89504206e-01 -2.05455616e-01
1.39647436e+00 -2.77603343e-02 -1.59372890e+00 -2.77896941e-01
-5.53267121e-01 -6.98840737e-01 6.99342430e-01 -7.16388822e-01
-1.49699187e+00 9.58608150e-01 8.96675885e-01 9.15853903e-02
1.30630791e+00 -2.73541391e-01 8.40544581e-01 1.15897670e-01
2.14160234e-02 -6.22060537e-01 1.81296766e-01 2.38006338e-01
1.18190300e+00 -1.46889997e+00 1.17491581e-01 -2.58673728e-01
-2.58638263e-01 1.27062309e+00 4.39315826e-01 -6.39606267e-03
7.47543931e-01 -1.52469352e-01 2.05282763e-01 1.02293782e-01
-6.24426529e-02 4.71874252e-02 4.45017517e-01 3.57731462e-01
1.76031008e-01 -2.27285296e-01 -4.22713161e-01 3.65215600e-01
3.50663245e-01 3.58041763e-01 6.52454913e-01 4.29291427e-01
-4.14252073e-01 -7.29972303e-01 -7.97762036e-01 -4.19176929e-02
-4.52075154e-01 -3.03393573e-01 -1.00072317e-01 1.44171774e-01
-5.79875801e-03 1.19603825e+00 -2.09304482e-01 -1.05442666e-01
-3.84972513e-01 -3.38427842e-01 4.23790634e-01 -1.47379085e-01
1.04440331e-01 5.46908453e-02 -4.94499087e-01 -2.46021375e-01
-7.60888994e-01 -3.35175484e-01 -6.72034204e-01 -2.70139784e-01
-5.48950434e-01 1.00316018e-01 5.82067549e-01 1.01923895e+00
2.57891476e-01 2.29411811e-01 6.71174824e-01 -1.26792145e+00
-4.86567587e-01 -1.27481484e+00 -4.75779623e-01 1.66204646e-01
8.23969066e-01 -5.46780348e-01 -8.43879819e-01 4.20701891e-01] | [11.595486640930176, -2.666778564453125] |
d06f90c4-643c-4901-b240-6837cc2bf24a | learning-to-generalize-one-sample-at-a-time | 1910.03915 | null | https://arxiv.org/abs/1910.03915v3 | https://arxiv.org/pdf/1910.03915v3.pdf | Learning to Generalize One Sample at a Time with Self-Supervision | Although deep networks have significantly increased the performance of visual recognition methods, it is still challenging to achieve the robustness across visual domains that is necessary for real-world applications. To tackle this issue, research on domain adaptation and generalization has flourished over the last decade. An important aspect to consider when assessing the work done in the literature so far is the amount of data annotation necessary for training each approach, both at the source and target level. In this paper we argue that the data annotation overload should be minimal, as it is costly. Hence, we propose to use self-supervised learning to achieve domain generalization and adaptation. We consider learning regularities from non annotated data as an auxiliary task, and cast the problem within an Auxiliary Learning principled framework. Moreover, we suggest to further exploit the ability to learn about visual domains from non annotated images by learning from target data while testing, as data are presented to the algorithm one sample at a time. Results on three different scenarios confirm the value of our approach. | ['Tatiana Tommasi', "Antonio D'Innocente", 'Barbara Caputo', 'Silvia Bucci'] | 2019-10-09 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 6.32914245e-01 2.34697670e-01 -1.32084265e-01 -5.61789572e-01
-2.73530334e-01 -7.54051447e-01 7.23982453e-01 2.09565639e-01
-6.68520510e-01 8.71518433e-01 -1.83721960e-01 -1.33204699e-01
-1.25686571e-01 -5.94609320e-01 -6.23273849e-01 -7.53721118e-01
1.60211205e-01 4.32280362e-01 4.13554013e-01 -4.07276824e-02
1.54059812e-01 8.53758037e-01 -1.84881747e+00 2.97101170e-01
8.53211641e-01 1.01406825e+00 1.70194700e-01 4.55487460e-01
-2.16557719e-02 7.81224191e-01 -6.55087829e-01 -3.77909571e-01
4.30427223e-01 -5.02132475e-01 -8.90923500e-01 6.04501545e-01
4.31099206e-01 -1.81193873e-01 2.28559271e-01 1.04090440e+00
4.17391896e-01 1.12920702e-01 8.01373482e-01 -1.34901392e+00
-5.61314583e-01 9.86713096e-02 -2.58638561e-01 2.83537865e-01
-6.55248202e-03 2.19292209e-01 7.64635444e-01 -7.16674864e-01
7.93442667e-01 7.27079391e-01 4.40855920e-01 7.76929677e-01
-1.41634965e+00 -3.47875953e-01 2.42324531e-01 2.50061572e-01
-1.03602302e+00 -4.28552687e-01 9.64223981e-01 -5.98202944e-01
5.75667799e-01 1.20207474e-01 4.73709196e-01 1.29880702e+00
-3.71584535e-01 8.61642122e-01 1.45496905e+00 -7.68881798e-01
4.75929946e-01 7.20950603e-01 2.08082478e-02 2.75403082e-01
3.78026843e-01 1.11713879e-01 -2.29940534e-01 1.61748171e-01
6.65116310e-01 -3.06772619e-01 -1.26430660e-01 -1.01625562e+00
-6.85010731e-01 7.38308311e-01 3.30357134e-01 5.95874429e-01
-2.70915389e-01 -4.55965161e-01 4.34683084e-01 4.46200520e-01
5.87559342e-01 4.36232388e-01 -4.61017162e-01 1.17535278e-01
-8.47680271e-01 5.17944023e-02 7.17505097e-01 7.19188929e-01
7.58153021e-01 9.17186365e-02 1.96753666e-01 9.38825548e-01
3.71487960e-02 1.57053560e-01 4.89148766e-01 -6.61948085e-01
1.77664608e-01 8.96815598e-01 -3.74901928e-02 -7.70270944e-01
-3.65717113e-01 -6.96853638e-01 -8.68897855e-01 6.69317424e-01
7.65492558e-01 -8.57326612e-02 -9.75499809e-01 1.69745278e+00
3.39251816e-01 -1.06096268e-01 2.84531951e-01 8.10313761e-01
4.18323487e-01 2.54131198e-01 2.91433156e-01 -2.78910428e-01
9.22122598e-01 -7.75390685e-01 -4.13505942e-01 -4.81173813e-01
6.52445853e-01 -5.85318565e-01 1.10223567e+00 5.98551750e-01
-8.75499487e-01 -7.39918351e-01 -1.11915660e+00 1.28179714e-01
-5.82380116e-01 2.69850641e-01 3.58312398e-01 6.63859725e-01
-8.63277495e-01 4.99795735e-01 -6.28771603e-01 -7.03482389e-01
4.82105285e-01 3.96457523e-01 -6.23331964e-01 -5.25989830e-02
-9.31339741e-01 1.07043529e+00 8.01441312e-01 -9.08906236e-02
-7.08071649e-01 -3.25407803e-01 -6.31781042e-01 -1.30882096e-02
4.87902641e-01 -3.45594436e-01 1.17472911e+00 -1.83303559e+00
-1.20686829e+00 1.23628020e+00 1.77448630e-01 -5.91876149e-01
6.75994277e-01 3.31298709e-02 -3.97540301e-01 1.69321984e-01
-3.37614901e-02 5.93892753e-01 9.12154853e-01 -1.60096705e+00
-7.67922580e-01 -4.43991721e-01 3.20451081e-01 3.91906872e-03
-6.10264242e-01 -2.27522030e-01 -2.69473195e-01 -5.95712483e-01
-1.15251057e-01 -9.96459424e-01 -3.10292877e-02 -9.05710086e-02
1.06877089e-02 -3.27127397e-01 8.37817788e-01 -5.00781178e-01
9.32053804e-01 -2.27142310e+00 1.48357943e-01 2.19114661e-01
-8.30030218e-02 7.24399745e-01 -7.63497129e-02 2.40626603e-01
-3.00261796e-01 -8.25815126e-02 -5.23243427e-01 -2.23346442e-01
-2.04178989e-01 5.11332154e-01 -2.21044064e-01 3.49519521e-01
4.63165373e-01 6.49919093e-01 -6.79361045e-01 -4.00646806e-01
2.78271526e-01 4.06291008e-01 -4.96818393e-01 4.34986651e-01
-2.63755918e-01 6.62258506e-01 -2.69667298e-01 3.33769351e-01
5.43078482e-01 -1.70461938e-01 3.07719290e-01 -1.05855368e-01
-1.70742832e-02 -1.08586267e-01 -1.20337248e+00 1.50329983e+00
-4.00500983e-01 6.76753402e-01 8.74713063e-03 -1.67062044e+00
1.10692120e+00 2.88384736e-01 4.44933057e-01 -9.27552223e-01
6.83628470e-02 2.99135417e-01 2.35195681e-01 -5.97554684e-01
2.07552433e-01 -3.77970994e-01 2.73419023e-01 3.19591045e-01
3.00328463e-01 9.97359380e-02 3.05837542e-01 -1.74791172e-01
7.04923809e-01 3.66654009e-01 4.72325385e-01 -2.35141978e-01
8.12366009e-01 2.92191982e-01 4.30745333e-01 5.53644061e-01
-2.09714383e-01 4.95108724e-01 5.01237988e-01 -6.16650760e-01
-1.28068638e+00 -8.51339817e-01 -2.20592692e-01 1.17867529e+00
-5.35694733e-02 1.08979866e-01 -8.12640965e-01 -9.85634387e-01
-1.86357155e-01 6.64374471e-01 -7.97360539e-01 -1.99186310e-01
-4.95405734e-01 -5.94334364e-01 3.00975889e-01 6.72390163e-01
5.05623102e-01 -1.16238356e+00 -9.71782804e-01 1.92760170e-01
1.90581992e-01 -1.12415338e+00 1.49214461e-01 4.44510758e-01
-9.41508651e-01 -1.08292806e+00 -7.71981120e-01 -6.69783235e-01
8.10934365e-01 3.98661308e-02 1.13248229e+00 -1.35956015e-02
-3.41444254e-01 6.49992168e-01 -5.36004186e-01 -5.29676676e-01
-5.60052693e-01 1.17227048e-01 -5.07677197e-02 3.15731615e-01
5.21565974e-01 -6.75730288e-01 -3.12144846e-01 3.29577923e-01
-1.29730165e+00 -9.34466720e-02 7.91031718e-01 9.10838723e-01
4.70217258e-01 9.91299562e-03 6.37674928e-01 -8.94686759e-01
5.76740324e-01 -2.70013154e-01 -6.86345935e-01 2.91426480e-01
-6.42481446e-01 1.76578820e-01 7.49529719e-01 -5.77697456e-01
-1.11106074e+00 4.47481066e-01 -5.24465777e-02 -2.47015908e-01
-6.61366105e-01 4.13961440e-01 -3.52326125e-01 -1.90712839e-01
1.02584350e+00 3.17391664e-01 2.25056112e-01 -3.61746222e-01
2.88524777e-01 7.00729072e-01 3.61685574e-01 -5.15436471e-01
8.10829401e-01 5.12799680e-01 7.97739848e-02 -1.00970960e+00
-8.84626031e-01 -3.82399082e-01 -1.06809342e+00 -2.75077671e-01
7.56454945e-01 -5.26579738e-01 -4.03375447e-01 1.19883865e-01
-9.51778769e-01 -4.15419549e-01 -5.17621398e-01 1.81029409e-01
-6.76373720e-01 3.64316732e-01 2.02816844e-01 -9.27527726e-01
1.13376521e-01 -1.05069041e+00 6.14476860e-01 1.85204759e-01
-1.54970631e-01 -1.26521087e+00 2.47793384e-02 2.14205980e-01
3.84335428e-01 3.87135267e-01 8.13082159e-01 -1.20653224e+00
-3.84149164e-01 -1.53896421e-01 -2.66919404e-01 9.00162637e-01
1.92162976e-01 -3.96703213e-01 -1.21166408e+00 -2.78051466e-01
1.71440959e-01 -6.28935397e-01 6.90609932e-01 1.95850525e-02
1.09378350e+00 -1.01065524e-01 -1.17821030e-01 3.39524776e-01
1.44474280e+00 2.25633994e-01 5.62957942e-01 6.31529987e-01
3.87717873e-01 1.13668501e+00 6.49544954e-01 1.79175690e-01
9.40099657e-02 8.07833135e-01 5.38152993e-01 -3.38647097e-01
-1.84613764e-01 -3.60815264e-02 2.05937624e-01 1.81500033e-01
-2.04436317e-01 -1.52309075e-01 -1.18989527e+00 7.04111934e-01
-1.81022596e+00 -8.80382717e-01 1.24893554e-01 2.30460119e+00
5.82250237e-01 2.69301325e-01 3.87857556e-01 4.16370302e-01
5.00614643e-01 -1.71717390e-01 -5.65557063e-01 -3.32392782e-01
-2.00256988e-01 -3.06255743e-03 1.49077043e-01 2.65010625e-01
-1.28392577e+00 7.26119697e-01 5.67578363e+00 5.93329489e-01
-1.32758546e+00 -9.83490124e-02 5.57076693e-01 2.56614387e-01
8.07577074e-02 -2.84275580e-02 -5.61380446e-01 2.11592138e-01
6.97043240e-01 1.87900867e-02 1.28814116e-01 1.13112819e+00
-3.79132926e-02 -2.28209957e-01 -1.32063520e+00 8.39639366e-01
3.53044450e-01 -8.40616941e-01 2.15077307e-02 5.90092503e-02
5.71039736e-01 -2.69411862e-01 9.32584330e-02 2.55659550e-01
1.13044336e-01 -9.91298795e-01 3.99648607e-01 2.57543594e-01
4.61070001e-01 -6.34122193e-01 7.77421832e-01 6.03437424e-01
-8.05266798e-01 -2.58168817e-01 -4.12075579e-01 -1.43271506e-01
-1.72309592e-01 3.61407369e-01 -1.12422848e+00 6.16556525e-01
5.21172345e-01 6.04174614e-01 -9.26968932e-01 1.06977344e+00
-2.90701002e-01 5.33142686e-01 -3.13035816e-01 1.28637195e-01
2.57989228e-01 -1.68862864e-01 4.12837952e-01 1.15842533e+00
1.09471217e-01 -2.72497147e-01 2.38917962e-01 6.18609190e-01
9.16535854e-02 2.10054442e-01 -8.99356902e-01 8.10809433e-02
1.54021561e-01 1.19564378e+00 -9.03201997e-01 -2.87205607e-01
-4.00628418e-01 1.06556368e+00 4.46080506e-01 4.12278950e-01
-5.05706906e-01 -1.34745434e-01 2.47652188e-01 1.48946330e-01
2.02468544e-01 -1.34778753e-01 -2.58228123e-01 -1.00980687e+00
3.46574008e-01 -9.46857870e-01 5.25551140e-01 -5.56669652e-01
-1.37091255e+00 6.82998478e-01 1.89302236e-01 -1.40332210e+00
-4.54474956e-01 -9.29203212e-01 -1.96579129e-01 7.44665682e-01
-1.59374285e+00 -1.17836440e+00 -3.86576504e-01 6.25874937e-01
5.12677133e-01 -1.83280364e-01 7.15668619e-01 1.73200905e-01
-2.82037079e-01 5.39986014e-01 4.15503234e-02 2.15646133e-01
8.25640440e-01 -1.19129336e+00 -4.60204519e-02 9.76389647e-01
3.22034299e-01 4.44164157e-01 8.23290884e-01 -3.28159004e-01
-9.50254381e-01 -9.09911036e-01 6.79797351e-01 -5.58004320e-01
5.38871765e-01 -4.00699437e-01 -1.24787056e+00 5.25248826e-01
2.49139279e-01 1.54609606e-01 7.42630661e-01 6.77080452e-03
-3.94389689e-01 -1.74744979e-01 -1.12041628e+00 4.37932104e-01
8.34229112e-01 -4.76671517e-01 -9.09345746e-01 1.23390362e-01
1.82943553e-01 -6.85684904e-02 -6.24072433e-01 4.77557331e-01
3.36622834e-01 -1.20385611e+00 8.66293430e-01 -7.67015219e-01
2.24639729e-01 -2.84491599e-01 -1.47708520e-01 -1.32906651e+00
-3.21719795e-02 -9.05537512e-03 1.80579245e-01 1.42359018e+00
4.37811553e-01 -5.44091046e-01 1.00149965e+00 5.73159695e-01
1.52575850e-01 -4.67042565e-01 -8.77500951e-01 -1.01964366e+00
2.36350730e-01 -3.83618444e-01 1.89035535e-01 1.03066432e+00
-2.51110733e-01 4.60687429e-01 -3.96296769e-01 1.25944361e-01
5.19805431e-01 3.64659019e-02 9.19851601e-01 -1.61347699e+00
-2.77292222e-01 -2.88359553e-01 -5.68018675e-01 -8.61419976e-01
1.96801409e-01 -7.99547374e-01 -4.36169654e-02 -1.38122237e+00
6.73997402e-02 -3.78721327e-01 -2.82631934e-01 5.92915058e-01
-6.38099015e-02 3.77157986e-01 2.94515789e-01 2.35044762e-01
-4.67213154e-01 3.39208424e-01 1.00873137e+00 -1.27694160e-01
-2.10043862e-01 9.98989120e-02 -5.46299100e-01 8.32060695e-01
8.93933892e-01 -3.02997261e-01 -6.62508011e-01 -2.97200203e-01
-1.52054112e-02 -4.02230561e-01 5.33627808e-01 -1.12848246e+00
1.91634431e-01 -6.92774728e-02 4.92454916e-01 -7.35780671e-02
3.60315233e-01 -1.34527147e+00 -1.50859162e-01 3.68646741e-01
-4.11500752e-01 -2.04307333e-01 3.16197872e-01 3.64860147e-01
-3.47891301e-01 -5.23500860e-01 1.01760149e+00 1.59777645e-02
-1.03635728e+00 5.70359081e-02 -1.86417684e-01 4.98695076e-02
1.24927473e+00 -4.88640517e-01 6.85278252e-02 -2.20950499e-01
-1.08609426e+00 -1.26566634e-01 6.57858431e-01 4.26772535e-01
3.21348280e-01 -1.06439817e+00 -5.32153666e-01 2.17939779e-01
5.22026718e-01 -1.97505906e-01 1.61676347e-01 6.47814751e-01
-2.51073092e-01 3.18975836e-01 -5.64564645e-01 -7.87226021e-01
-1.43121910e+00 9.27220702e-01 2.50122398e-01 -2.29098991e-01
-3.45974445e-01 3.81219238e-01 1.49538606e-01 -3.28597844e-01
4.23381001e-01 1.12118766e-01 -5.08038819e-01 2.55150080e-01
4.91421551e-01 1.78280815e-01 1.85504332e-01 -5.76613247e-01
-2.55112112e-01 5.64200401e-01 -6.52065948e-02 -1.10493205e-01
1.37701416e+00 -9.11308900e-02 1.48508340e-01 5.06154597e-01
9.76965368e-01 -1.81493670e-01 -1.55285645e+00 -3.18793595e-01
5.02598941e-01 -3.16234082e-01 -2.97196180e-01 -9.03578222e-01
-7.57450521e-01 1.18813848e+00 8.50017130e-01 6.21537983e-01
1.42532194e+00 -4.93866429e-02 -7.99078792e-02 4.07489896e-01
2.25133538e-01 -1.30864072e+00 2.02293739e-01 2.78721213e-01
8.58271062e-01 -1.54259467e+00 -6.66722804e-02 -2.54151523e-01
-8.44338119e-01 1.18863666e+00 6.71894133e-01 -1.46816894e-01
2.50139922e-01 -1.89456698e-02 2.92672873e-01 1.72625482e-01
-3.62150669e-01 -4.75096136e-01 4.17055607e-01 1.02855027e+00
4.50878263e-01 -3.61745179e-01 -2.91474640e-01 2.21307471e-01
2.16263786e-01 1.42813772e-01 4.02111620e-01 1.05654609e+00
-5.03457785e-01 -1.56014240e+00 -3.10724437e-01 1.81336641e-01
-4.31750238e-01 2.32660010e-01 -7.74712563e-01 1.15779042e+00
3.59296829e-01 8.86046588e-01 -1.02461822e-01 -3.42749543e-02
4.35287863e-01 3.06249142e-01 4.92340326e-01 -7.36385405e-01
-4.56042022e-01 -1.25251576e-01 -3.75725143e-02 -2.09817484e-01
-8.89418185e-01 -5.35453141e-01 -7.28264034e-01 3.10075551e-01
-3.71923037e-02 2.26162598e-01 5.98698437e-01 1.04065895e+00
2.10115761e-01 4.70885128e-01 4.10730273e-01 -7.68210292e-01
-5.88573515e-01 -9.18148816e-01 -3.21626574e-01 7.76920140e-01
3.33102286e-01 -8.04106355e-01 -3.39790225e-01 3.76190782e-01] | [9.739410400390625, 2.5066866874694824] |
67c99aba-61e1-4306-b2ed-18ea84870da7 | inferring-dynamic-regulatory-interaction | 2306.07803 | null | https://arxiv.org/abs/2306.07803v1 | https://arxiv.org/pdf/2306.07803v1.pdf | Inferring dynamic regulatory interaction graphs from time series data with perturbations | Complex systems are characterized by intricate interactions between entities that evolve dynamically over time. Accurate inference of these dynamic relationships is crucial for understanding and predicting system behavior. In this paper, we propose Regulatory Temporal Interaction Network Inference (RiTINI) for inferring time-varying interaction graphs in complex systems using a novel combination of space-and-time graph attentions and graph neural ordinary differential equations (ODEs). RiTINI leverages time-lapse signals on a graph prior, as well as perturbations of signals at various nodes in order to effectively capture the dynamics of the underlying system. This approach is distinct from traditional causal inference networks, which are limited to inferring acyclic and static graphs. In contrast, RiTINI can infer cyclic, directed, and time-varying graphs, providing a more comprehensive and accurate representation of complex systems. The graph attention mechanism in RiTINI allows the model to adaptively focus on the most relevant interactions in time and space, while the graph neural ODEs enable continuous-time modeling of the system's dynamics. We evaluate RiTINI's performance on various simulated and real-world datasets, demonstrating its state-of-the-art capability in inferring interaction graphs compared to previous methods. | ['Smita Krishnaswamy', 'Guy Wolf', 'Frederik Wenkel', 'Aarthi Venkat', 'Edward De Brouwer', 'Sumner Magruder', 'Dhananjay Bhaskar'] | 2023-06-13 | null | null | null | null | ['graph-attention', 'causal-inference', 'causal-inference'] | ['graphs', 'knowledge-base', 'miscellaneous'] | [ 1.39420167e-01 1.48789018e-01 -2.16936599e-02 1.33966118e-01
1.48457944e-01 -7.74469137e-01 7.85436630e-01 3.10296029e-01
3.51569086e-01 8.19698632e-01 1.33975878e-01 -6.19909525e-01
-6.26844704e-01 -8.66802275e-01 -7.31021166e-01 -5.74898183e-01
-1.00358737e+00 5.80664217e-01 5.10252714e-02 -3.10996264e-01
-1.51729256e-01 6.66233778e-01 -8.27688873e-01 -3.16900998e-01
7.86066115e-01 4.90916729e-01 -2.43277013e-01 1.08635283e+00
3.02563578e-01 9.72356975e-01 -4.49096560e-01 2.72364169e-02
-9.80107784e-02 -4.54505175e-01 -5.68993330e-01 -2.06894606e-01
-1.38301879e-01 -1.53052956e-02 -1.06134450e+00 7.40412235e-01
1.85690090e-01 3.11786443e-01 8.09940636e-01 -1.36104500e+00
-9.76404071e-01 7.15509295e-01 -3.91427010e-01 5.59209645e-01
3.39071721e-01 5.80649018e-01 1.12446332e+00 -2.19084486e-01
4.38780755e-01 1.54008520e+00 6.36407852e-01 2.06971988e-01
-1.77827561e+00 -6.05629086e-01 9.14196193e-01 2.77501973e-03
-1.39568293e+00 -9.34014916e-02 8.50934565e-01 -6.23248100e-01
1.30191708e+00 2.26322293e-01 9.08390403e-01 1.08491874e+00
7.63036847e-01 3.96256417e-01 4.53638941e-01 1.65288895e-01
2.25009948e-01 -8.22181821e-01 2.92223722e-01 7.32295394e-01
3.05709451e-01 2.53275603e-01 -1.86139613e-01 -4.72410262e-01
1.01936150e+00 4.60576594e-01 -2.93363631e-01 -5.36826439e-03
-9.52888072e-01 4.69409853e-01 6.58167064e-01 1.90514207e-01
-5.84782720e-01 7.72790432e-01 1.13335654e-01 2.30293527e-01
4.80234116e-01 6.55874670e-01 -4.24949408e-01 1.43479323e-02
-1.29751399e-01 1.68808475e-01 1.17557788e+00 7.77310789e-01
5.39101601e-01 2.28687301e-01 -3.39834183e-01 2.51232237e-01
3.98775220e-01 4.05932307e-01 -2.44048223e-01 -6.58441961e-01
1.04785264e-01 9.14953411e-01 8.11683666e-03 -1.31060982e+00
-4.74654943e-01 -5.57666183e-01 -1.17522228e+00 -3.02621484e-01
4.32173669e-01 -4.23617095e-01 -1.08848953e+00 2.00082040e+00
2.73984939e-01 8.58243048e-01 -2.63159692e-01 5.24997532e-01
3.95866483e-01 9.00489748e-01 2.35779658e-02 -2.84400344e-01
9.68601704e-01 -4.32080537e-01 -9.15433764e-01 -1.53957993e-01
1.42364085e-01 1.21695198e-01 6.49118304e-01 -1.18728029e-02
-9.18363810e-01 -1.22765839e-01 -7.64226258e-01 2.39853755e-01
-6.03357971e-01 -6.12182140e-01 8.41755271e-01 5.43829650e-02
-1.14024901e+00 7.24197626e-01 -1.29579794e+00 -2.30002314e-01
3.85231823e-01 6.10464811e-01 6.43723384e-02 1.04983911e-01
-1.56896043e+00 5.41201115e-01 -6.39615878e-02 5.50836742e-01
-1.34144831e+00 -9.98951614e-01 -9.24124062e-01 3.85415643e-01
6.50382578e-01 -7.96626687e-01 1.01309323e+00 -3.50941688e-01
-1.35358953e+00 -1.07988819e-01 -1.03267647e-01 -5.87720037e-01
2.13702708e-01 2.02118367e-01 -5.46449900e-01 -1.20355994e-01
-2.46430248e-01 -3.17936428e-02 5.39162040e-01 -9.29717243e-01
3.44674774e-02 -1.67828932e-01 3.42987657e-01 7.87950903e-02
3.23841870e-02 -5.18691182e-01 -5.68675458e-01 -4.27442431e-01
-2.44893730e-01 -1.01037729e+00 -5.62961459e-01 -1.72302470e-01
-7.17012644e-01 -3.00911009e-01 9.96289611e-01 -6.22913778e-01
1.66872239e+00 -1.69804418e+00 6.13292575e-01 3.17288160e-01
8.31136465e-01 7.43463635e-02 3.37425321e-02 1.11213040e+00
-1.31733119e-01 3.00561488e-01 -3.25243682e-01 -1.18241489e-01
2.71533746e-02 4.52040225e-01 -3.12017769e-01 3.88485521e-01
5.51909804e-01 1.40068650e+00 -1.16959369e+00 -1.49468288e-01
2.33589649e-01 6.75257146e-01 -3.80672544e-01 3.02008957e-01
-6.25229836e-01 6.39372170e-01 -6.54080153e-01 3.86129349e-01
1.16091929e-01 -8.52929711e-01 6.20556414e-01 1.26901358e-01
1.89313844e-01 8.46244767e-02 -7.44192541e-01 8.19795966e-01
-2.74418116e-01 7.65552700e-01 -1.06339313e-01 -1.10962856e+00
5.38453758e-01 4.15106624e-01 7.06031680e-01 -5.33684433e-01
2.02395350e-01 -5.45371890e-01 3.13758492e-01 -3.21537465e-01
-1.40004516e-01 2.93037504e-01 -3.01350862e-01 4.27837759e-01
-9.45890844e-02 -2.33116169e-02 3.29249650e-01 5.33144474e-01
1.63164902e+00 -1.81376919e-01 2.98556566e-01 -3.57573122e-01
3.43643576e-01 -3.61093372e-01 5.54157972e-01 7.36588418e-01
-4.13778089e-02 -1.81471169e-01 1.18385959e+00 -4.64171857e-01
-8.37872148e-01 -1.18252957e+00 1.58360884e-01 6.18966758e-01
1.91403717e-01 -3.37680995e-01 -5.55893779e-01 -4.21768695e-01
4.75464165e-01 5.74198961e-01 -1.15019512e+00 -5.73040962e-01
-4.65499163e-01 -8.44422758e-01 2.45862007e-01 4.69560742e-01
8.80482793e-03 -9.04862642e-01 -1.80519912e-02 5.89731753e-01
1.44527450e-01 -8.87100399e-01 -7.68384993e-01 4.24409620e-02
-5.85742235e-01 -1.47067833e+00 -5.13404831e-02 -3.01974088e-01
8.80956292e-01 -5.91741391e-02 1.14356267e+00 1.46576673e-01
-5.30593693e-01 6.05565608e-01 2.26470873e-01 -1.66523382e-01
-4.16966796e-01 -1.07140839e-01 4.58678156e-02 1.54698208e-01
-1.39893308e-01 -1.00734079e+00 -5.17411530e-01 2.63660491e-01
-9.84464407e-01 -5.02267256e-02 3.40227246e-01 8.75056326e-01
4.58311468e-01 4.03651148e-01 5.57893634e-01 -1.11013341e+00
9.58525658e-01 -9.01869953e-01 -9.92658436e-01 3.33800554e-01
-7.80265450e-01 2.74629623e-01 1.02514434e+00 -8.95312190e-01
-8.16310644e-01 -2.39086553e-01 4.90388095e-01 -5.33907831e-01
2.08945855e-01 1.03162074e+00 -5.17232716e-02 1.69584945e-01
4.17018980e-01 1.04623459e-01 -7.26225367e-03 -1.81817133e-02
3.98697913e-01 -8.16637427e-02 4.74117160e-01 -7.15097487e-01
7.19701350e-01 3.20196718e-01 4.26933646e-01 -7.05126762e-01
-4.13363367e-01 -3.85864712e-02 -5.74591696e-01 -4.61263388e-01
5.01613319e-01 -5.57523489e-01 -1.38157868e+00 5.33747017e-01
-8.46414804e-01 -8.60179961e-01 -3.87160666e-02 8.87177959e-02
-2.32912555e-01 -3.13643292e-02 -1.04552758e+00 -1.11241949e+00
-1.11565605e-01 -6.98689163e-01 1.03090405e+00 1.54581860e-01
-2.35354811e-01 -1.72192907e+00 2.95164347e-01 -4.50801671e-01
3.74633580e-01 9.05560911e-01 9.94943440e-01 -2.99168020e-01
-8.59390676e-01 -4.43264037e-01 -1.89134657e-01 -3.76253575e-01
4.70225364e-01 4.65929151e-01 -3.57787192e-01 -2.88557231e-01
-5.24545670e-01 4.59300071e-01 5.17495930e-01 7.59756565e-01
9.01633799e-01 -7.36474633e-01 -7.74962008e-01 4.53556508e-01
1.19210374e+00 4.67916012e-01 3.19150805e-01 -6.34479463e-01
1.07010663e+00 3.19546819e-01 8.95723552e-02 4.71470833e-01
6.30365551e-01 3.14374298e-01 6.29270375e-01 -1.68702185e-01
3.02025676e-01 -4.94335383e-01 3.55683833e-01 6.33602560e-01
3.72675024e-02 -6.30611718e-01 -1.19460106e+00 4.98128176e-01
-2.15908384e+00 -9.78137076e-01 -2.08785072e-01 1.98229337e+00
6.08088970e-01 2.77345449e-01 1.94037467e-01 -4.16490614e-01
8.63595963e-01 1.46219954e-01 -1.39956105e+00 -1.01092696e-01
1.22207649e-01 -1.46401137e-01 4.88597065e-01 8.55240285e-01
-7.69666135e-01 8.18070054e-01 7.62761211e+00 1.06516890e-01
-1.04277420e+00 -3.64127815e-01 7.65332460e-01 3.57731506e-02
-3.22443336e-01 1.91663876e-01 -4.55814093e-01 4.33775514e-01
1.36567044e+00 -7.91144013e-01 1.05586398e+00 1.41103759e-01
5.17523408e-01 2.22725883e-01 -1.20324326e+00 3.56198341e-01
-4.62310910e-01 -1.38794315e+00 -1.30599126e-01 4.34175521e-01
6.81864023e-01 3.72329284e-03 1.56007232e-02 3.78840029e-01
1.35913181e+00 -1.16602397e+00 6.88497946e-02 9.73393619e-01
5.05254269e-01 -5.35084426e-01 3.52860749e-01 2.86471874e-01
-1.61788726e+00 -8.92553553e-02 1.99722484e-01 -4.82209057e-01
2.26645574e-01 7.66384780e-01 -8.28840137e-01 5.40028930e-01
3.68032992e-01 1.31994295e+00 -2.90232509e-01 4.48872685e-01
-2.36592680e-01 9.61786330e-01 -4.74989444e-01 -1.97737291e-01
2.86945086e-02 -4.49010581e-01 6.32416129e-01 8.69601786e-01
-1.63564995e-01 3.47397864e-01 2.44677782e-01 1.30940270e+00
2.41509154e-02 -7.89111733e-01 -9.10805345e-01 -8.94325197e-01
5.94900131e-01 8.88177812e-01 -6.88067496e-01 -4.98354018e-01
-1.92058146e-01 4.98677969e-01 5.67333043e-01 9.95388448e-01
-1.13054323e+00 -1.39666542e-01 9.44511831e-01 9.14244130e-02
1.90239146e-01 -7.17553556e-01 2.19764248e-01 -9.82796550e-01
-3.27856153e-01 -5.88027775e-01 7.95789897e-01 -5.99326909e-01
-1.44291294e+00 3.94232154e-01 1.49200022e-01 -7.29975641e-01
-3.91591012e-01 -4.40013766e-01 -8.33503664e-01 8.42920244e-01
-1.07455993e+00 -9.63058114e-01 -8.96020159e-02 4.87963974e-01
6.56722933e-02 4.06441092e-01 4.62361217e-01 -3.72404233e-02
-1.23642945e+00 8.59887376e-02 6.06705248e-02 9.14401412e-02
7.44693503e-02 -1.51851630e+00 9.43324327e-01 9.18161094e-01
-2.21638039e-01 8.59885573e-01 9.08570349e-01 -1.03246784e+00
-2.05468941e+00 -1.30493665e+00 2.99035341e-01 -6.60381794e-01
1.24983394e+00 -7.22154796e-01 -1.07224345e+00 1.08178079e+00
-6.88343914e-03 2.27770045e-01 3.58156383e-01 2.27243319e-01
-3.21891755e-01 1.09951101e-01 -7.14788437e-01 9.44896758e-01
1.28975713e+00 -7.20368803e-01 -1.01346865e-01 4.19750601e-01
1.02919686e+00 -3.86927694e-01 -1.18934572e+00 2.69694686e-01
5.01262486e-01 -2.20769659e-01 1.05088949e+00 -9.61105764e-01
2.48171270e-01 -3.67321968e-01 3.98980260e-01 -1.46338570e+00
-7.67242193e-01 -1.05924916e+00 -8.79916549e-01 9.20568705e-01
5.11357605e-01 -1.08225524e+00 3.08888674e-01 8.14842403e-01
7.93612003e-02 -7.54719973e-01 -3.87638509e-01 -5.89125276e-01
-2.30956808e-01 -1.63541690e-01 7.56583691e-01 1.20097280e+00
2.41537303e-01 5.96719980e-01 -2.34364465e-01 7.42317438e-01
5.42828202e-01 1.40832067e-01 6.17316484e-01 -1.26926696e+00
-5.83454430e-01 -6.62773609e-01 -4.04607803e-01 -7.94201672e-01
2.94333875e-01 -6.62700295e-01 -3.15835476e-02 -1.59695721e+00
-2.52773743e-02 -1.85023453e-02 -4.25191671e-01 2.97994286e-01
-5.79539001e-01 -4.14159089e-01 -2.34791920e-01 -4.15964723e-02
-5.68164229e-01 6.64855301e-01 1.27760077e+00 -3.35477352e-01
-4.07590419e-01 -5.08951545e-02 -7.02637613e-01 1.91552550e-01
5.18771172e-01 -2.15298250e-01 -6.63129389e-01 -4.66941111e-02
2.22820595e-01 5.29453754e-01 3.93728346e-01 -5.75094759e-01
4.66630518e-01 -4.66338784e-01 1.21358626e-01 -4.55166161e-01
2.75526315e-01 -6.84421778e-01 6.76922381e-01 6.80985928e-01
-3.85593385e-01 2.10406974e-01 4.30713326e-01 1.33977532e+00
1.98515300e-02 8.69855821e-01 3.08399916e-01 5.17355390e-02
-3.11349690e-01 1.01747441e+00 -5.93454957e-01 -1.10522390e-03
9.41477597e-01 1.69498026e-01 -4.69891995e-01 -8.85045588e-01
-9.26157832e-01 9.18198347e-01 1.81265861e-01 4.04007256e-01
3.09747130e-01 -9.86199856e-01 -4.48449939e-01 1.41016627e-02
-3.83357774e-03 1.85590938e-01 3.82952511e-01 7.69138038e-01
-2.34581411e-01 3.68599653e-01 1.19275741e-01 -5.97553194e-01
-7.83719003e-01 8.88579071e-01 7.31658578e-01 -5.53996563e-01
-5.59861541e-01 3.85408372e-01 5.16805470e-01 -4.57394689e-01
-1.30806044e-01 -5.57984412e-01 -1.02971427e-01 -3.45157981e-01
1.62197083e-01 2.12481707e-01 -5.82173049e-01 -2.60902643e-01
-1.76005661e-01 2.57774770e-01 1.32234871e-01 1.26790002e-01
1.27060497e+00 -1.25113532e-01 -4.30927873e-01 8.80661607e-01
8.24445963e-01 -4.82152402e-01 -1.75010359e+00 -2.51223564e-01
-5.23539484e-02 -6.58104196e-02 1.02778740e-01 -6.72372401e-01
-9.15480912e-01 5.63234866e-01 -1.10285260e-01 9.77317452e-01
9.96970832e-01 2.41080914e-02 4.74695027e-01 3.26567590e-01
1.11500263e-01 -5.66722393e-01 -3.80651583e-03 5.89691401e-01
8.99012446e-01 -6.56377912e-01 3.16535681e-02 -5.86048126e-01
-5.49190156e-02 8.19517374e-01 6.04413927e-01 -3.78813505e-01
1.03899002e+00 4.56650317e-01 -6.53613150e-01 -5.54615915e-01
-1.54527259e+00 -1.72664002e-02 3.87329608e-01 4.85871315e-01
2.65139908e-01 2.71199346e-01 3.07116687e-01 3.35231870e-01
3.18171650e-01 -3.37814271e-01 4.69390690e-01 7.14972377e-01
9.71508548e-02 -6.68813705e-01 -9.14482102e-02 6.99690163e-01
-5.15598282e-02 2.00622138e-02 -7.54115164e-01 8.13925326e-01
-7.44738460e-01 9.19985592e-01 3.09524238e-01 -3.12589169e-01
3.63345981e-01 -1.75578624e-01 1.89876273e-01 -4.46905613e-01
-3.36099416e-01 -5.61557934e-02 7.94945359e-02 -6.54035330e-01
1.68735906e-01 -7.51033306e-01 -1.19089293e+00 -6.78974569e-01
-3.80476266e-01 2.40545496e-02 1.92416236e-01 8.94635439e-01
9.04794335e-01 1.36055624e+00 6.39188051e-01 -7.89636254e-01
-2.69274563e-02 -8.22924078e-01 -5.02204716e-01 3.10774744e-01
7.67243803e-01 -7.23873079e-01 -4.22145486e-01 1.23855257e-02] | [7.098731994628906, 5.836288928985596] |
e9c241fd-34a6-438d-bbb6-1745cf8eff09 | on-the-efficacy-and-noise-robustness-of | 2305.1254 | null | https://arxiv.org/abs/2305.12540v2 | https://arxiv.org/pdf/2305.12540v2.pdf | On the Efficacy and Noise-Robustness of Jointly Learned Speech Emotion and Automatic Speech Recognition | New-age conversational agent systems perform both speech emotion recognition (SER) and automatic speech recognition (ASR) using two separate and often independent approaches for real-world application in noisy environments. In this paper, we investigate a joint ASR-SER multitask learning approach in a low-resource setting and show that improvements are observed not only in SER, but also in ASR. We also investigate the robustness of such jointly trained models to the presence of background noise, babble, and music. Experimental results on the IEMOCAP dataset show that joint learning can improve ASR word error rate (WER) and SER classification accuracy by 10.7% and 2.3% respectively in clean scenarios. In noisy scenarios, results on data augmented with MUSAN show that the joint approach outperforms the independent ASR and SER approaches across many noisy conditions. Overall, the joint ASR-SER approach yielded more noise-resistant models than the independent ASR and SER approaches. | ['Aravind Ganapathiraju', 'Pushpak Jagtap', 'Malolan Chetlur', 'S. Pavankumar Dubagunta', 'Lokesh Bansal'] | 2023-05-21 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [-4.38454039e-02 -2.28871346e-01 5.37700534e-01 -3.64167720e-01
-1.71294761e+00 -3.81698251e-01 7.44458497e-01 -2.54751772e-01
-7.42610037e-01 9.08020496e-01 4.65745449e-01 -1.73035786e-02
2.87295878e-01 1.65173933e-01 -3.65375817e-01 -8.28798711e-01
5.87247871e-02 2.83552885e-01 -1.75436899e-01 -4.91953999e-01
-3.57668996e-01 -5.67633146e-03 -1.51556516e+00 3.93013090e-01
6.45189166e-01 1.05888689e+00 3.55829537e-01 1.17378080e+00
1.77994028e-01 9.17043924e-01 -1.45620251e+00 -3.44885260e-01
-2.06947662e-02 -3.55092511e-02 -5.13664305e-01 3.39247100e-02
-1.18081182e-01 -1.57483611e-02 -4.13750082e-01 6.42448783e-01
1.15907741e+00 6.81634188e-01 2.34267235e-01 -9.68515456e-01
-2.85883993e-01 6.55047417e-01 -3.69329691e-01 2.73187399e-01
5.11913180e-01 1.00087663e-02 9.28300738e-01 -8.58601391e-01
8.59189406e-02 1.42409182e+00 6.04080439e-01 6.89461887e-01
-9.91235197e-01 -8.12484205e-01 -7.66543970e-02 1.73784748e-01
-1.18430996e+00 -1.16812551e+00 4.90378171e-01 1.49508640e-01
1.55126679e+00 3.43984991e-01 -3.70412618e-02 1.78218198e+00
-3.69556248e-01 1.05905354e+00 1.43160033e+00 -4.35788065e-01
3.70968282e-01 3.67687017e-01 8.31331164e-02 7.73383602e-02
-5.28617740e-01 1.28976226e-01 -1.22183394e+00 -3.71112972e-01
-2.38506887e-02 -7.50392079e-01 -3.89259607e-01 7.84719348e-01
-1.02142966e+00 6.45944655e-01 -4.88382638e-01 4.00558978e-01
-5.61236441e-01 2.02144772e-01 8.38174403e-01 7.13444650e-01
1.13658810e+00 2.54513830e-01 -1.02836668e+00 -9.70412672e-01
-6.19992733e-01 -2.39019245e-01 1.04707897e+00 7.31263578e-01
1.19150423e-01 7.26938307e-01 -2.24649027e-01 1.77891481e+00
1.81614906e-01 8.02093744e-01 5.54192424e-01 -7.05575407e-01
5.87791145e-01 -3.92203003e-01 1.95137218e-01 -2.34642684e-01
-3.64212006e-01 -6.25423074e-01 -5.54554701e-01 -2.51810580e-01
1.50361255e-01 -6.72298670e-01 -8.07371795e-01 1.88906872e+00
-1.19718779e-02 3.17595780e-01 7.01527238e-01 6.13290668e-01
9.64020073e-01 9.65838194e-01 2.74870872e-01 -5.83211005e-01
1.41158366e+00 -9.97927070e-01 -1.46684921e+00 -5.49348354e-01
6.40056312e-01 -1.03613961e+00 9.22697902e-01 7.58577824e-01
-1.11494565e+00 -3.21682751e-01 -9.31736350e-01 3.51696730e-01
-1.81595162e-01 3.11227798e-01 4.14563030e-01 1.32253206e+00
-1.21123755e+00 -2.01162249e-01 -5.74116409e-01 -2.99778134e-01
2.72295363e-02 4.10688221e-02 -4.18609321e-01 2.14867413e-01
-1.50390899e+00 1.16975880e+00 8.25721025e-02 9.81451571e-02
-8.42912078e-01 -4.41443503e-01 -9.73287523e-01 7.32544959e-02
5.47838986e-01 -1.61148445e-03 1.76768422e+00 -8.26488972e-01
-2.24917936e+00 6.39545441e-01 -4.32884961e-01 -7.30241179e-01
3.33847523e-01 -5.29670596e-01 -9.14986134e-01 -1.37200654e-01
-4.99207318e-01 -5.00309616e-02 6.88473225e-01 -1.27575231e+00
-6.21209562e-01 -2.19047308e-01 -4.50085610e-01 3.95606220e-01
-3.13325226e-01 7.39079356e-01 -1.81470856e-01 -6.41969323e-01
-3.18665653e-01 -7.55009532e-01 5.18039949e-02 -8.14599991e-01
-1.42918965e-02 -2.82792479e-01 7.99702048e-01 -1.18467903e+00
1.03609371e+00 -2.33336926e+00 -1.29705384e-01 -3.05187590e-02
-3.58790040e-01 6.31855369e-01 -2.34991118e-01 3.06111127e-01
3.22570354e-02 -2.53613899e-03 1.22698709e-01 -1.06373012e+00
1.06269978e-01 2.63986766e-01 -3.29320505e-02 3.47216427e-01
1.07124127e-01 4.80122238e-01 -5.42692244e-01 2.66038049e-02
3.59369725e-01 8.41840863e-01 -1.76944192e-02 4.97188032e-01
1.64643645e-01 2.93769151e-01 9.57841799e-02 4.96156037e-01
3.54683071e-01 3.77699137e-01 6.98914602e-02 1.29570276e-01
1.28590450e-01 5.97227395e-01 -1.12372077e+00 1.41360331e+00
-1.19713783e+00 8.92906606e-01 8.56487691e-01 -9.50347424e-01
1.25808215e+00 1.09757125e+00 1.62232473e-01 -9.16276515e-01
2.81902462e-01 3.79794240e-01 4.48065996e-02 -3.10046285e-01
5.47038078e-01 -3.53007138e-01 -1.42595291e-01 2.52897084e-01
3.12321842e-01 -3.44073400e-02 -3.11750144e-01 8.58451277e-02
1.12430954e+00 -5.14936328e-01 4.00083721e-01 -3.93797196e-02
5.67871690e-01 -7.05668867e-01 5.28845906e-01 9.67679739e-01
-6.54021144e-01 2.74799109e-01 6.64777495e-03 1.27451226e-01
-8.06398809e-01 -9.38687027e-01 -1.39879957e-01 1.47569132e+00
-4.58300680e-01 -2.29427710e-01 -6.14232123e-01 -3.44219416e-01
-4.13373381e-01 1.07550097e+00 -2.55765617e-01 -4.88642640e-02
-3.88353139e-01 -9.91300344e-01 9.81847703e-01 3.38030040e-01
4.37380046e-01 -1.07351935e+00 1.37918051e-02 4.06915665e-01
-5.85442245e-01 -1.97067869e+00 -3.44971180e-01 6.17839038e-01
-6.74499497e-02 -2.27772325e-01 -7.46889055e-01 -3.92748564e-01
-3.84353548e-01 1.62593365e-01 7.99607456e-01 -1.79544568e-01
5.88283725e-02 5.92410505e-01 -7.48322606e-01 -6.22331560e-01
-9.82027292e-01 -2.03718036e-01 4.46487069e-01 2.36753002e-01
4.15177673e-01 -4.16375935e-01 1.57618210e-01 4.40370172e-01
-4.49673712e-01 -3.87839615e-01 3.58598500e-01 1.10874856e+00
-2.50896215e-01 -1.46492869e-01 1.31863344e+00 -2.81853437e-01
9.40383434e-01 -3.94958109e-01 -2.09507763e-01 2.98950285e-01
-2.95856833e-01 -1.00357540e-01 3.09323072e-01 -4.85865533e-01
-1.62865841e+00 -2.70088613e-01 -5.07675946e-01 -1.04464732e-01
-4.79890555e-01 4.35949326e-01 -3.53645742e-01 2.93242604e-01
6.24584854e-01 3.14338148e-01 -9.27610025e-02 -3.77435416e-01
1.19374767e-01 1.64107180e+00 2.79404759e-01 -5.13366520e-01
2.59022206e-01 -6.08923435e-02 -6.68414891e-01 -1.42353320e+00
-5.98650694e-01 -8.82058859e-01 1.49649680e-01 -3.12932163e-01
7.29999423e-01 -1.47037923e+00 -6.94356859e-01 7.98182905e-01
-1.21035099e+00 -5.60047388e-01 2.04438537e-01 7.34588325e-01
-5.80622435e-01 3.39089870e-01 -7.92759955e-01 -1.78386915e+00
-6.73117518e-01 -1.27100110e+00 1.18729293e+00 -6.23344779e-02
-2.35968456e-01 -8.82925093e-01 -1.23482563e-01 9.09713209e-01
6.54002547e-01 -4.11410153e-01 2.30722800e-01 -1.28168094e+00
2.21763924e-01 -1.43603176e-01 3.22072096e-02 7.64764607e-01
1.75856113e-01 -4.25643235e-01 -1.66274357e+00 -2.47906938e-01
1.21242240e-01 -6.92946672e-01 7.30420887e-01 2.73316532e-01
6.70070410e-01 -1.78433254e-01 2.59004116e-01 1.16505906e-01
8.18102658e-01 5.41189611e-01 6.51244521e-01 1.71245977e-01
4.26519476e-02 4.00030762e-01 5.00061154e-01 6.78821623e-01
1.23374917e-01 9.17946577e-01 -1.76199436e-01 -5.88556267e-02
-1.59161150e-01 4.70573127e-01 9.46526051e-01 1.07793581e+00
6.28769025e-02 -6.88995183e-01 -7.99933314e-01 7.06019521e-01
-1.73464119e+00 -9.73510444e-01 -1.51157044e-02 2.02331996e+00
8.77603889e-01 -2.07206219e-01 2.54286468e-01 2.51953244e-01
7.86743879e-01 2.77033269e-01 1.23107685e-02 -8.22623014e-01
-7.38622189e-01 3.88186932e-01 3.81409615e-01 7.29977548e-01
-9.42885876e-01 9.53644097e-01 6.12390327e+00 1.05378890e+00
-6.87897325e-01 6.58542931e-01 8.44202638e-01 -4.37859774e-01
2.50291616e-01 -7.95656562e-01 -4.80926424e-01 2.34913573e-01
1.57395542e+00 8.82926956e-03 6.86375499e-01 7.43698180e-01
5.05285561e-01 -3.16654027e-01 -6.05146408e-01 1.44895947e+00
2.78201908e-01 -6.61617637e-01 -5.89464724e-01 -1.83215335e-01
5.60654521e-01 4.14628983e-01 1.20765138e-02 7.29824722e-01
5.71355104e-01 -9.48306322e-01 6.41603410e-01 3.98997329e-02
5.63376129e-01 -8.11255455e-01 1.12280726e+00 2.89863557e-01
-7.99054503e-01 -2.44528651e-02 4.29936089e-02 8.33924860e-02
3.90426755e-01 4.70561653e-01 -9.95508671e-01 5.04453182e-01
7.70644784e-01 4.26730998e-02 1.57761410e-01 6.50719523e-01
-5.27927093e-02 1.03551197e+00 -4.61600900e-01 -2.50285000e-01
1.58792287e-01 8.17841440e-02 8.82748604e-01 1.83907163e+00
3.30500424e-01 2.46396333e-01 -4.51735742e-02 1.87087312e-01
-2.25271150e-01 2.24553943e-01 -1.39490962e-01 1.01108789e-01
5.87069929e-01 1.19676554e+00 -1.73850104e-01 -4.03967619e-01
-3.53481472e-01 1.24405622e+00 2.11990610e-01 7.17425823e-01
-8.15176070e-01 -1.03340447e-01 1.03568077e+00 -7.59526610e-01
3.38704914e-01 -3.68548483e-01 -2.67061353e-01 -9.50289249e-01
5.10728806e-02 -1.29940593e+00 1.90719068e-01 -6.93787098e-01
-1.12773669e+00 8.08776498e-01 -4.73174542e-01 -5.38037837e-01
-2.46652782e-01 -5.03979146e-01 -5.37571490e-01 9.61401284e-01
-1.49096334e+00 -8.84702742e-01 7.25080967e-02 4.32498187e-01
1.13090837e+00 -6.51132643e-01 1.04286158e+00 5.38148701e-01
-7.23092377e-01 8.62669826e-01 4.70178515e-01 2.22399384e-02
8.27689469e-01 -1.28738487e+00 2.90408403e-01 5.92437923e-01
2.60647774e-01 1.68011829e-01 1.06082463e+00 -3.21578622e-01
-1.25139916e+00 -7.49144256e-01 6.42966330e-01 -2.63961554e-01
7.41569579e-01 -7.77427197e-01 -7.90248811e-01 4.27514642e-01
5.49969494e-01 -3.12286943e-01 8.10510933e-01 6.02830470e-01
-4.43748504e-01 4.66881916e-02 -1.16828406e+00 3.86375844e-01
5.85513532e-01 -8.11146438e-01 -3.47137839e-01 3.09394151e-01
7.58948445e-01 -1.38341531e-01 -9.08536315e-01 1.51784047e-01
3.30294073e-01 -6.43284798e-01 6.53289020e-01 -5.71918190e-01
-3.43060106e-01 1.48698598e-01 -7.18969226e-01 -1.93999946e+00
2.92410016e-01 -1.24949729e+00 -4.15273644e-02 1.38142145e+00
8.48087072e-01 -6.60077453e-01 2.17663094e-01 4.46314007e-01
-3.94027710e-01 -1.29985422e-01 -1.37436903e+00 -9.85874414e-01
-3.20711255e-01 -1.11055136e+00 6.85696602e-02 7.80797660e-01
1.51690036e-01 5.34779727e-01 -9.07115638e-01 3.65173638e-01
2.53017217e-01 -8.44830215e-01 6.24982476e-01 -6.35265827e-01
-3.73987406e-01 -2.39321634e-01 -2.11919487e-01 -7.14852095e-01
5.28293252e-01 -3.41741025e-01 4.38483804e-01 -1.20270824e+00
-2.04580426e-01 -4.81091551e-02 -3.38382095e-01 2.82058150e-01
-4.05410528e-01 -5.85809425e-02 1.21602304e-01 -3.33378911e-01
-7.53120303e-01 9.77256060e-01 3.28496397e-01 3.12321726e-02
-2.99566358e-01 9.88283306e-02 -2.65145302e-01 4.26763594e-01
9.68636632e-01 -3.26559454e-01 -5.19842356e-02 -1.62459671e-01
-2.85223037e-01 3.17269951e-01 -4.41794321e-02 -8.32346857e-01
8.62712935e-02 2.32002318e-01 -1.06753103e-01 -9.49361399e-02
1.09680378e+00 -4.85633433e-01 -1.95905462e-01 -1.30968004e-01
-3.65914822e-01 -3.36787075e-01 6.47875071e-01 5.45223117e-01
-2.57095307e-01 1.72829106e-01 8.61021340e-01 5.28031886e-02
-4.09127682e-01 -3.54348034e-01 -9.29208100e-01 -1.30505413e-02
7.05578983e-01 7.97778368e-02 -3.40376377e-01 -1.13312817e+00
-8.25928807e-01 9.72606093e-02 -5.16558826e-01 6.68646514e-01
4.38898534e-01 -8.19991827e-01 -1.17248034e+00 -6.37743101e-02
1.20385572e-01 -6.81549251e-01 3.97905082e-01 9.61450815e-01
3.39869112e-01 4.49088871e-01 4.67178732e-01 -2.67536193e-01
-1.84876931e+00 -6.95608482e-02 6.29942000e-01 -1.46553561e-01
-4.43269759e-02 9.20480490e-01 -1.30863532e-01 -5.61067820e-01
7.28219926e-01 1.15378916e-01 8.73615518e-02 2.88947579e-02
7.02690482e-01 4.49590802e-01 5.17917156e-01 -8.30799162e-01
-3.27502370e-01 -2.20868781e-01 -1.54689386e-01 -8.02766144e-01
1.38461077e+00 -4.57089871e-01 3.04746717e-01 7.97226727e-01
1.14402390e+00 8.35782439e-02 -7.75342703e-01 -6.19957685e-01
1.76825553e-01 -1.65142149e-01 5.66924751e-01 -1.40971971e+00
-7.02342093e-01 6.09166801e-01 6.10658109e-01 3.17110509e-01
1.01999772e+00 2.23044325e-02 7.73852170e-01 5.33889592e-01
2.51395881e-01 -1.57819736e+00 4.75521460e-02 7.66078651e-01
1.01081252e+00 -1.38952363e+00 -5.28341711e-01 -1.94535464e-01
-1.08336663e+00 6.95054054e-01 4.28367466e-01 6.42283559e-01
4.37354535e-01 5.96688151e-01 5.93012571e-01 1.02949165e-01
-1.11864638e+00 -4.55139697e-01 5.17586842e-02 6.76475883e-01
5.62213778e-01 2.66851276e-01 7.85910431e-03 7.69145727e-01
-3.37949306e-01 -5.94995201e-01 4.83582050e-01 5.56416214e-01
-4.15051997e-01 -8.36189687e-01 -5.67930996e-01 7.73931444e-02
-8.58696461e-01 -3.15860301e-01 -4.48449969e-01 4.94355291e-01
-6.94368184e-01 1.81590092e+00 -1.32851422e-01 -3.64191353e-01
3.60985219e-01 5.83377421e-01 2.78911000e-04 -3.62993360e-01
-9.63588357e-01 5.13008535e-01 1.12481320e+00 -3.15132439e-01
-3.82660031e-01 -6.95085227e-01 -9.74036694e-01 -4.12241668e-02
-8.01173627e-01 4.01283115e-01 1.24296916e+00 1.24723470e+00
1.35532141e-01 8.29374909e-01 8.09689105e-01 -3.88070494e-01
-7.07481861e-01 -1.64388895e+00 -6.27366185e-01 3.16643953e-01
5.03081918e-01 -3.47923905e-01 -8.01585615e-01 -3.63917083e-01] | [14.515018463134766, 6.301003456115723] |
1ab5ca29-91d5-44e0-9875-7cab38581cce | face-transformer-towards-high-fidelity-and | 2304.0253 | null | https://arxiv.org/abs/2304.02530v1 | https://arxiv.org/pdf/2304.02530v1.pdf | Face Transformer: Towards High Fidelity and Accurate Face Swapping | Face swapping aims to generate swapped images that fuse the identity of source faces and the attributes of target faces. Most existing works address this challenging task through 3D modelling or generation using generative adversarial networks (GANs), but 3D modelling suffers from limited reconstruction accuracy and GANs often struggle in preserving subtle yet important identity details of source faces (e.g., skin colors, face features) and structural attributes of target faces (e.g., face shapes, facial expressions). This paper presents Face Transformer, a novel face swapping network that can accurately preserve source identities and target attributes simultaneously in the swapped face images. We introduce a transformer network for the face swapping task, which learns high-quality semantic-aware correspondence between source and target faces and maps identity features of source faces to the corresponding region in target faces. The high-quality semantic-aware correspondence enables smooth and accurate transfer of source identity information with minimal modification of target shapes and expressions. In addition, our Face Transformer incorporates a multi-scale transformation mechanism for preserving the rich fine facial details. Extensive experiments show that our Face Transformer achieves superior face swapping performance qualitatively and quantitatively. | ['Shijian Lu', 'Fangneng Zhan', 'Rongliang Wu', 'Kaiwen Cui'] | 2023-04-05 | null | null | null | null | ['face-swapping'] | ['computer-vision'] | [ 2.60850787e-01 3.49209338e-01 2.72635221e-01 -5.93873143e-01
-6.58988357e-01 -8.11611533e-01 4.61040586e-01 -9.88731205e-01
4.03077036e-01 7.34471679e-01 2.41085485e-01 4.78427708e-01
3.59551400e-01 -8.70272219e-01 -8.53064835e-01 -8.35357547e-01
2.98335880e-01 3.62414092e-01 -3.13164055e-01 -3.41386348e-01
-2.55729765e-01 9.21714246e-01 -1.59960592e+00 3.15706998e-01
7.04731524e-01 1.10010040e+00 -3.99591774e-01 3.79380554e-01
-1.38174102e-01 4.56946969e-01 -6.20307624e-01 -8.25900972e-01
7.21588492e-01 -8.80051136e-01 -3.83210450e-01 6.45164691e-04
9.98052299e-01 -3.31460267e-01 -4.15360063e-01 1.22247171e+00
4.72983003e-01 -3.12045842e-01 7.31472492e-01 -1.96040082e+00
-1.43264008e+00 -6.29518330e-02 -7.32974291e-01 -4.96101350e-01
4.19858545e-01 1.87001824e-01 3.49043250e-01 -9.40696061e-01
6.14778340e-01 1.70718586e+00 8.38543594e-01 1.03267741e+00
-1.30962563e+00 -1.39980376e+00 -2.30686113e-01 -1.19737163e-01
-1.44667900e+00 -7.30318248e-01 1.11018550e+00 -3.44815731e-01
2.72051655e-02 3.71130913e-01 6.16392910e-01 1.20946634e+00
2.32523143e-01 2.64808893e-01 1.40664566e+00 -7.90237486e-02
-2.26355150e-01 2.10250884e-01 -8.99770856e-01 9.57031965e-01
-9.14972350e-02 3.40446860e-01 -6.22630060e-01 -6.19634241e-02
1.34932446e+00 1.08194046e-01 -3.19816560e-01 -3.88163000e-01
-9.57654476e-01 7.02467263e-01 5.56582212e-01 5.61443046e-02
-3.75102490e-01 1.16852045e-01 -4.51016314e-02 5.26781499e-01
4.12606984e-01 1.08720697e-01 -2.34798476e-01 3.35863799e-01
-5.36881626e-01 2.07673743e-01 3.66412342e-01 1.23798084e+00
9.55634892e-01 4.20737088e-01 -4.57440853e-01 8.10416937e-01
2.88647622e-01 1.07797670e+00 2.28540972e-01 -1.14571965e+00
9.57395285e-02 6.13451779e-01 -1.04425587e-01 -1.25098336e+00
1.64923728e-01 -4.65926155e-02 -1.13756752e+00 6.19050801e-01
3.10790315e-02 -4.04695719e-02 -1.05418599e+00 2.29387999e+00
5.11345208e-01 2.46803090e-01 -3.07234451e-02 5.58292568e-01
1.14923632e+00 4.47538406e-01 -4.46330048e-02 -3.26954909e-02
1.23425579e+00 -7.35950172e-01 -7.78838217e-01 -3.68906222e-02
-4.21563864e-01 -9.92361009e-01 9.24787462e-01 -4.49312210e-01
-1.41780365e+00 -6.54805124e-01 -7.05190539e-01 -1.04941204e-01
-2.18717843e-01 -4.62454569e-04 2.60112435e-01 7.60822654e-01
-1.43587804e+00 4.05827880e-01 -3.78282338e-01 -6.74502878e-03
9.57298875e-01 7.49337375e-01 -1.03820992e+00 -4.34794836e-02
-1.08394063e+00 6.36695445e-01 -2.14231849e-01 -2.34919414e-02
-1.05330968e+00 -1.15288401e+00 -1.15243626e+00 9.15201195e-03
-1.63806319e-01 -1.02029526e+00 9.25271869e-01 -1.55975688e+00
-1.65832365e+00 1.33183014e+00 -2.45895833e-01 1.99273735e-01
6.17703140e-01 2.17436939e-01 -4.23557520e-01 1.35393851e-02
2.68841296e-01 1.01479888e+00 1.47318840e+00 -1.54938078e+00
-2.74976730e-01 -6.64019763e-01 -2.46767297e-01 1.07809670e-01
-3.32623899e-01 7.36290812e-02 -3.17136019e-01 -1.06999648e+00
-1.70426503e-01 -8.34182560e-01 2.97190040e-01 6.93951309e-01
-3.43061894e-01 4.22344178e-01 1.22269809e+00 -9.73907113e-01
3.90471458e-01 -2.17844534e+00 2.73640782e-01 1.61826268e-01
2.68007964e-01 2.31686488e-01 -6.46668792e-01 1.84139743e-01
-4.27702069e-01 1.26044005e-01 -3.23305547e-01 -4.00563955e-01
8.31729919e-03 2.02455744e-01 -3.46202374e-01 4.00263458e-01
3.79829645e-01 1.37090468e+00 -7.09450245e-01 -3.45354468e-01
1.30998254e-01 1.12182224e+00 -5.10205448e-01 4.88650680e-01
1.19070984e-01 9.89127219e-01 -3.12441796e-01 8.39763999e-01
1.14126778e+00 2.31015071e-01 -2.83048809e-01 -3.96502733e-01
5.80750227e-01 -3.90089393e-01 -7.00433373e-01 1.35880315e+00
-4.59119260e-01 3.85195524e-01 5.32545507e-01 -5.36467612e-01
1.23229206e+00 3.43706667e-01 5.78331113e-01 -7.34757960e-01
2.08914667e-01 8.66155997e-02 -3.54836434e-01 4.35278937e-02
-1.08171664e-01 -5.51466286e-01 -1.03653856e-02 4.30393398e-01
9.62988138e-02 -3.74482483e-01 -6.42773926e-01 -1.25547379e-01
5.91749310e-01 7.21586570e-02 1.40449042e-02 -2.51856893e-01
6.44365370e-01 -6.45411789e-01 6.87527955e-01 -1.02299221e-01
-4.30315733e-02 8.86091769e-01 3.48702669e-01 -4.11822736e-01
-1.14491677e+00 -1.41847587e+00 1.66961402e-01 7.00165033e-01
5.97850718e-02 2.61416454e-02 -1.06998837e+00 -7.75463045e-01
2.64309466e-01 3.43698561e-01 -1.10873878e+00 -7.50197411e-01
-5.67125201e-01 -2.27450371e-01 6.68469667e-01 5.71108937e-01
8.01936567e-01 -1.13371694e+00 1.44774944e-01 -1.36917189e-01
-8.18414316e-02 -9.00550663e-01 -1.14687037e+00 -6.82292700e-01
-5.94345212e-01 -1.16213238e+00 -7.96817362e-01 -1.21830630e+00
1.18113327e+00 1.24954641e-01 1.01696682e+00 1.72844026e-02
-1.57354563e-01 3.58279079e-01 -2.25942079e-02 -4.50133443e-01
-7.65183568e-01 -5.68146408e-01 2.31136501e-01 4.69122142e-01
-7.79711753e-02 -9.57393289e-01 -4.84374881e-01 6.19029343e-01
-9.42237198e-01 1.91205397e-01 3.76666725e-01 9.17296588e-01
6.55989826e-01 5.38978800e-02 6.85086131e-01 -8.64183903e-01
4.52352077e-01 -1.50556117e-01 -4.00368661e-01 4.27763611e-01
-2.46557176e-01 -2.00159341e-01 7.29879737e-01 -3.91121596e-01
-1.15424716e+00 1.97924729e-02 -2.54522383e-01 -9.93365109e-01
-1.01152487e-01 -5.35537779e-01 -9.15104866e-01 -6.17992043e-01
3.80157858e-01 3.87569875e-01 5.79463542e-01 -3.35117698e-01
4.81725454e-01 1.85054049e-01 9.38540280e-01 -5.73474824e-01
1.49091709e+00 6.51293933e-01 1.11838713e-01 -3.47068250e-01
-6.39142513e-01 4.45782185e-01 -6.93235755e-01 -6.61446825e-02
6.82539046e-01 -9.45771813e-01 -6.68302119e-01 9.97254252e-01
-1.05119181e+00 -1.45120591e-01 -6.23423636e-01 -2.98286676e-01
-7.68644333e-01 -4.06637527e-02 -5.08927405e-01 -2.65918106e-01
-4.61865455e-01 -9.24084008e-01 1.50848436e+00 4.42086488e-01
-1.58794280e-02 -9.60241377e-01 -1.46549076e-01 2.98603684e-01
5.87524533e-01 9.46976542e-01 8.70729744e-01 1.12357505e-01
-5.18206894e-01 -1.73748508e-01 -2.41695583e-01 4.81659442e-01
9.91048455e-01 -4.60485555e-03 -1.03250778e+00 -4.36987907e-01
3.74487936e-02 -4.74922992e-02 3.78519714e-01 2.70860106e-01
1.18284619e+00 -7.31967568e-01 -2.21377373e-01 1.10463822e+00
1.04819512e+00 3.13060462e-01 9.09222066e-01 -3.99830610e-01
9.64130104e-01 7.88288057e-01 1.86883211e-01 1.36704862e-01
1.74411342e-01 6.64989233e-01 4.10323441e-01 -4.52537715e-01
-6.30231023e-01 -7.54313052e-01 5.25719225e-01 3.53211731e-01
1.57681983e-02 1.51766598e-01 -4.14369941e-01 4.35343117e-01
-1.34236062e+00 -9.61233675e-01 5.37491798e-01 1.99003553e+00
9.70054269e-01 -5.50490856e-01 1.95306109e-03 -2.21986026e-01
1.11866152e+00 -1.73841953e-01 -8.67448151e-01 -8.76032338e-02
-2.87910610e-01 5.46838164e-01 1.53280422e-01 4.73687410e-01
-8.08837175e-01 1.01139140e+00 6.17726898e+00 7.71901369e-01
-1.12250447e+00 2.90472418e-01 8.49858403e-01 -2.31350347e-01
-6.49679840e-01 -4.07318681e-01 -4.05732542e-01 6.69092059e-01
2.95971513e-01 -3.71847808e-01 6.43924832e-01 6.58501148e-01
-2.13336810e-01 8.21348310e-01 -1.05245745e+00 1.29072201e+00
3.83942425e-01 -1.42909873e+00 4.28637475e-01 1.73807353e-01
1.13453877e+00 -7.30611026e-01 5.99828303e-01 -2.98026130e-02
5.22670090e-01 -1.60474777e+00 7.88421571e-01 7.16832280e-01
1.76076698e+00 -1.01058209e+00 5.01084507e-01 -1.98223293e-01
-1.25969470e+00 1.71605393e-01 1.41900191e-02 3.81798208e-01
8.61328542e-02 1.39389798e-01 -4.31685597e-01 3.44301492e-01
8.46063614e-01 7.12225020e-01 -4.70844835e-01 2.51607358e-01
-4.18677360e-01 -3.46301869e-02 -9.63612944e-02 6.37370765e-01
-3.36947024e-01 -4.32037622e-01 5.50698817e-01 4.38726902e-01
6.04096949e-01 1.15058720e-01 -2.84871489e-01 1.34096754e+00
-6.41721368e-01 -1.41062111e-01 -9.36342239e-01 6.38111606e-02
7.12645710e-01 1.21967733e+00 -2.60429353e-01 -1.00388989e-01
-7.29320869e-02 1.24138963e+00 8.64609927e-02 1.85748652e-01
-8.90051484e-01 -9.11561474e-02 1.47340572e+00 2.75000751e-01
2.87656814e-01 3.32087457e-01 -2.15196699e-01 -9.64484692e-01
2.50386208e-01 -1.01616895e+00 -1.47198569e-02 -8.93443644e-01
-1.69450843e+00 8.23549092e-01 -3.63723427e-01 -1.03923237e+00
-8.77285004e-02 -3.44478160e-01 -8.36156428e-01 1.16283107e+00
-1.30949485e+00 -1.95052731e+00 -6.23423159e-01 1.15564787e+00
9.97202620e-02 -4.87880647e-01 1.01077676e+00 2.83506185e-01
-2.64952779e-01 1.16237879e+00 7.22885691e-03 3.99094850e-01
8.42714548e-01 -8.97660434e-01 6.99215710e-01 4.31548029e-01
-6.64623082e-02 5.11447906e-01 2.97830284e-01 -6.10510468e-01
-1.27056885e+00 -1.43625391e+00 4.82806116e-01 -6.66386604e-01
-2.86395475e-02 -4.74036247e-01 -7.16554403e-01 8.27245891e-01
2.12264001e-01 2.76088774e-01 6.50171876e-01 -5.12617290e-01
-6.17786527e-01 -2.76619941e-01 -1.86521733e+00 5.22246361e-01
1.33445513e+00 -8.20448458e-01 -2.60992825e-01 1.47000656e-01
6.44517243e-01 -4.84116137e-01 -9.98331070e-01 4.40930337e-01
6.87816143e-01 -1.06208217e+00 1.22140408e+00 -5.73884368e-01
4.35001969e-01 -3.99445802e-01 -1.12990603e-01 -1.57905316e+00
-2.85376638e-01 -8.25242937e-01 1.07901141e-01 1.76783323e+00
-1.67063907e-01 -9.02623296e-01 7.29609728e-01 5.24894536e-01
1.21192530e-01 -5.21096587e-01 -1.03544474e+00 -6.99929953e-01
3.40419650e-01 3.99012089e-01 1.42768908e+00 1.08275676e+00
-7.56622672e-01 5.68658933e-02 -5.84436357e-01 -3.42090172e-03
7.75815189e-01 3.24634612e-01 8.80523801e-01 -1.05800235e+00
2.10886255e-01 -4.41575974e-01 -5.74046969e-01 -2.99683273e-01
6.66038275e-01 -8.82264853e-01 -2.71548748e-01 -1.05624032e+00
1.51176780e-01 -4.66715515e-01 4.95084934e-02 8.31152380e-01
5.12604974e-03 8.97274435e-01 2.48742849e-01 -1.65693667e-02
2.78253943e-01 1.12608278e+00 1.84522891e+00 -2.08305955e-01
1.49311408e-01 -1.54795140e-01 -1.11394715e+00 6.87097967e-01
6.92237258e-01 -3.45474631e-01 -4.63725269e-01 -4.15980756e-01
-2.82650590e-01 8.79904404e-02 7.71731913e-01 -7.54867375e-01
-1.53525352e-01 -2.25805342e-01 8.97485375e-01 2.81312745e-02
5.52094638e-01 -1.02370918e+00 8.09726477e-01 2.55975723e-01
-4.31283377e-02 7.90517963e-03 4.36348379e-01 3.45275134e-01
-3.55274260e-01 5.67152202e-01 1.34148669e+00 -1.32044747e-01
-4.62159485e-01 9.31430221e-01 4.44495738e-01 2.62432415e-02
1.38820422e+00 -4.49780166e-01 -1.58352673e-01 -6.36460841e-01
-6.74833596e-01 -6.90226108e-02 1.06868148e+00 8.49356711e-01
7.57788122e-01 -2.09634209e+00 -9.79802012e-01 9.66866255e-01
-4.09934521e-02 9.79807973e-02 3.86802316e-01 3.46485913e-01
-3.45653653e-01 -8.70733038e-02 -1.08199418e+00 -2.61564851e-01
-1.56811857e+00 5.20516872e-01 6.96137309e-01 3.01679581e-01
-3.90835017e-01 1.18354011e+00 8.72775078e-01 -7.35975027e-01
-2.93803900e-01 3.88042122e-01 1.14381365e-01 -2.24055454e-01
6.20639741e-01 1.07580319e-01 -2.73474723e-01 -1.24603462e+00
-3.61576825e-01 1.11753881e+00 1.80643514e-01 2.06859991e-01
1.32234418e+00 -3.24630104e-02 -4.93556082e-01 -3.21092159e-01
1.40608144e+00 1.94516644e-01 -1.69071758e+00 -1.95443481e-01
-9.75344360e-01 -1.07692313e+00 -2.44837672e-01 -7.53526330e-01
-1.92733085e+00 6.94601834e-01 5.90153813e-01 -4.47016597e-01
1.54692531e+00 1.74944907e-01 1.00307965e+00 -4.92156625e-01
5.80048501e-01 -4.91490871e-01 1.39339626e-01 3.41571681e-02
1.41125774e+00 -1.04570699e+00 -4.69408244e-01 -7.54668117e-01
-5.34060717e-01 7.83160686e-01 8.66088748e-01 -1.45248994e-01
6.49403811e-01 3.84011775e-01 2.13907719e-01 -2.28284106e-01
-3.04334700e-01 3.53723794e-01 4.92311060e-01 1.23557663e+00
-1.42659005e-02 8.13610256e-02 6.08053446e-01 4.39457774e-01
-6.57876551e-01 -2.26507917e-01 -2.83191819e-02 5.72542071e-01
2.12301314e-01 -1.36838806e+00 -6.23475671e-01 3.14209461e-01
-1.81053430e-01 -3.23071592e-02 -8.26364815e-01 6.81634068e-01
4.77434814e-01 4.59192395e-01 2.37648100e-01 -5.06620586e-01
3.52012604e-01 -1.73866488e-02 9.21403468e-01 -5.02521932e-01
-4.45856094e-01 -2.45262146e-01 -5.51284850e-01 -6.99690044e-01
-4.02816027e-01 -6.23590529e-01 -1.01690960e+00 -8.87308896e-01
2.60840535e-01 -1.44347861e-01 3.29826832e-01 4.79678392e-01
7.31126785e-01 3.53059202e-01 9.41839218e-01 -7.70407438e-01
-2.76733339e-01 -5.36217451e-01 -8.57258499e-01 8.20529461e-01
5.73856592e-01 -8.89441133e-01 -1.66448802e-01 4.61948693e-01] | [12.665267944335938, -0.11663150787353516] |
bdb092ef-6d75-439f-8f3e-a1190b149b4a | learning-semantic-neural-tree-for-human | 1912.09622 | null | https://arxiv.org/abs/1912.09622v1 | https://arxiv.org/pdf/1912.09622v1.pdf | Learning Semantic Neural Tree for Human Parsing | The majority of existing human parsing methods formulate the task as semantic segmentation, which regard each semantic category equally and fail to exploit the intrinsic physiological structure of human body, resulting in inaccurate results. In this paper, we design a novel semantic neural tree for human parsing, which uses a tree architecture to encode physiological structure of human body, and designs a coarse to fine process in a cascade manner to generate accurate results. Specifically, the semantic neural tree is designed to segment human regions into multiple semantic subregions (e.g., face, arms, and legs) in a hierarchical way using a new designed attention routing module. Meanwhile, we introduce the semantic aggregation module to combine multiple hierarchical features to exploit more context information for better performance. Our semantic neural tree can be trained in an end-to-end fashion by standard stochastic gradient descent (SGD) with back-propagation. Several experiments conducted on four challenging datasets for both single and multiple human parsing, i.e., LIP, PASCAL-Person-Part, CIHP and MHP-v2, demonstrate the effectiveness of the proposed method. Code can be found at https://isrc.iscas.ac.cn/gitlab/research/sematree. | ['Siwei Lyu', 'Longyin Wen', 'Libo Zhang', 'Ruyi Ji', 'Yanjun Wu', 'Dawei Du', 'Chen Zhao', 'Feiyue Huang'] | 2019-12-20 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1829_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123580205.pdf | eccv-2020-8 | ['human-parsing'] | ['computer-vision'] | [ 1.20119788e-01 4.33000684e-01 -8.09819847e-02 -6.65681422e-01
-4.51522648e-01 -1.91917270e-01 -2.29369551e-02 -1.38445422e-01
-2.45275602e-01 4.64247197e-01 3.78233463e-01 5.18417135e-02
4.76248026e-01 -7.04356432e-01 -6.80845320e-01 -5.28757572e-01
4.17985409e-01 1.80194959e-01 2.65097618e-01 -4.43241000e-02
1.08064935e-02 -1.88742042e-01 -1.09576941e+00 3.42875808e-01
8.51100922e-01 1.21263433e+00 2.07551807e-01 4.32356238e-01
-1.95739105e-01 5.98961353e-01 -4.90741998e-01 -5.51918924e-01
-2.53425818e-02 -7.56274819e-01 -1.00019777e+00 -1.31642353e-02
6.19681329e-02 -2.59231478e-01 -1.35043353e-01 1.06254041e+00
7.50893414e-01 2.59509794e-02 2.06080183e-01 -9.86723125e-01
-6.91777110e-01 5.80733001e-01 -7.23673046e-01 -4.61837165e-02
2.63628572e-01 1.18868761e-01 8.61264884e-01 -5.39424062e-01
1.63980588e-01 1.65834856e+00 6.97400749e-01 9.87596035e-01
-9.28857565e-01 -8.28927577e-01 3.16243082e-01 2.59310752e-02
-9.81937587e-01 -2.99328357e-01 9.75191534e-01 -2.41188020e-01
4.43657309e-01 -9.57185626e-02 7.52692163e-01 1.20050120e+00
1.26476616e-01 1.17235744e+00 1.09567618e+00 4.70600203e-02
2.00769916e-01 -2.81650603e-01 4.57911491e-01 1.01003313e+00
1.40008941e-01 -2.20657110e-01 -5.08845210e-01 1.40562117e-01
8.41099501e-01 -1.62353851e-02 -1.25667617e-01 -4.36600372e-02
-9.49310184e-01 9.13977206e-01 7.85134733e-01 1.05146617e-01
-4.38926399e-01 2.57775605e-01 8.28247309e-01 -3.19067299e-01
2.97741175e-01 -2.75770016e-02 -5.97792923e-01 2.12704808e-01
-6.54978216e-01 1.73662812e-01 6.14433765e-01 6.93060756e-01
4.60629433e-01 1.98403616e-02 -2.44736150e-01 1.08355832e+00
6.46390021e-01 3.69609624e-01 8.38079453e-01 -1.12821007e+00
5.38411796e-01 5.28825641e-01 -2.25463659e-01 -9.98932958e-01
-7.96154380e-01 -3.35796297e-01 -9.15183127e-01 -2.13961393e-01
1.32202178e-01 -4.25367177e-01 -1.18197119e+00 2.09931350e+00
6.55447841e-01 1.27181113e-01 2.42301911e-01 1.24879575e+00
1.36057508e+00 6.94822431e-01 7.86218703e-01 9.00213197e-02
1.81809759e+00 -1.27785754e+00 -6.65290594e-01 -4.97614145e-01
3.13517243e-01 -2.69676536e-01 1.13883984e+00 2.71976143e-01
-1.15058124e+00 -9.17598963e-01 -9.42336857e-01 -4.03626472e-01
-1.98424712e-01 2.32265756e-01 8.27808261e-01 3.51209551e-01
-8.00681889e-01 4.71563101e-01 -9.65518296e-01 -3.23154539e-01
8.74637127e-01 3.29422265e-01 -1.39164597e-01 1.95426688e-01
-1.53155136e+00 4.08239484e-01 6.74429595e-01 4.82933432e-01
-6.67174757e-01 -3.29773366e-01 -1.09117091e+00 4.51148003e-02
2.63299048e-01 -1.07260537e+00 1.14568686e+00 -1.17393267e+00
-1.69643712e+00 1.03105736e+00 -1.23990878e-01 -3.98836434e-01
2.81309158e-01 -3.75537157e-01 -2.85573881e-02 3.60938728e-01
2.77654856e-01 1.16339326e+00 6.71072245e-01 -1.11320889e+00
-4.07103539e-01 -7.51072347e-01 -1.01029277e-01 3.89371514e-01
-2.27241572e-02 1.39851227e-01 -5.43333888e-01 -7.82402337e-01
2.13701919e-01 -7.38398552e-01 -2.19166368e-01 -2.82186508e-01
-6.55029058e-01 -3.38497519e-01 5.16935527e-01 -1.19913912e+00
1.10683060e+00 -1.95564353e+00 3.38106632e-01 -1.16317675e-01
2.53752947e-01 2.11200118e-01 4.40702140e-02 -1.69671461e-01
5.46178818e-02 2.36819327e-01 -4.96248484e-01 -4.07154262e-01
-3.37263756e-02 1.57566607e-01 2.06940934e-01 2.09263891e-01
1.96067482e-01 1.13138998e+00 -7.64468253e-01 -8.13923538e-01
1.10204257e-01 4.67292637e-01 -5.11002600e-01 2.81000882e-01
-1.56832412e-01 6.48146212e-01 -9.69668329e-01 7.56234229e-01
5.77627957e-01 -2.48707280e-01 -5.19181564e-02 -3.39496136e-01
3.06804895e-01 1.26575738e-01 -7.07763791e-01 2.05970454e+00
-4.89443362e-01 -1.29867569e-01 2.59490162e-01 -1.15400064e+00
8.62572312e-01 2.20222712e-01 3.63530934e-01 -7.93890655e-01
5.33172667e-01 -4.29618731e-03 -1.16021238e-01 -4.99306172e-01
-6.53987005e-02 -3.23373348e-01 -4.27844167e-01 9.65389796e-03
1.84514925e-01 3.20197970e-01 -4.56764810e-02 -5.74779660e-02
7.74560213e-01 4.69609171e-01 8.41980726e-02 -1.87433943e-01
7.13155866e-01 -1.49528325e-01 1.06174791e+00 1.63710609e-01
-5.50064147e-01 7.07060039e-01 5.48055232e-01 -2.21392423e-01
-6.52110159e-01 -1.01190543e+00 7.04149529e-02 1.34123623e+00
5.11031926e-01 -1.70029789e-01 -1.40806532e+00 -6.95703387e-01
-9.57282633e-02 4.64318395e-01 -7.17127919e-01 -4.58259284e-01
-5.45410275e-01 -9.46321011e-01 5.58804452e-01 8.60446274e-01
1.11379170e+00 -1.57897663e+00 -8.42281699e-01 3.14209640e-01
-4.51307207e-01 -1.19929099e+00 -4.92050648e-01 -4.40462716e-02
-9.06574130e-01 -8.67722452e-01 -8.64890516e-01 -1.01851892e+00
3.57290536e-01 -2.95933336e-01 9.14306283e-01 4.44936864e-02
-2.59835333e-01 -1.48549927e-02 -3.76018703e-01 -3.17286730e-01
-4.76364605e-02 1.70266598e-01 -3.45749706e-01 1.07314751e-01
3.31649661e-01 -3.03381234e-01 -9.18589175e-01 2.81664670e-01
-4.75969136e-01 5.16875625e-01 5.94251752e-01 8.85627985e-01
7.62744367e-01 -1.64990470e-01 7.32866049e-01 -9.35477436e-01
5.23727655e-01 -5.37035227e-01 -1.56229973e-01 6.69378340e-02
-1.91380918e-01 -5.41451275e-02 6.95584297e-01 -1.68320909e-01
-1.29336679e+00 1.81176230e-01 -6.41185045e-01 -2.12065995e-01
-3.66193861e-01 3.26875955e-01 -6.49830282e-01 2.17466325e-01
1.40115932e-01 2.29992092e-01 -7.06824958e-02 -6.64324403e-01
2.53095686e-01 6.53980613e-01 8.14961731e-01 -8.04451048e-01
2.52232879e-01 2.74519980e-01 -2.62648493e-01 -4.33476657e-01
-1.32089388e+00 -2.43675381e-01 -6.00043833e-01 -1.05168469e-01
1.57077277e+00 -1.04705775e+00 -6.97402835e-01 8.47683311e-01
-1.07277024e+00 -5.40586054e-01 1.01571590e-01 2.85911649e-01
-6.72571719e-01 3.22248161e-01 -1.01016402e+00 -3.83004248e-01
-8.83136392e-01 -9.99255776e-01 1.41012585e+00 7.91235387e-01
-1.85288310e-01 -8.02371562e-01 -2.22585946e-01 8.98359418e-01
1.19741209e-01 5.09036124e-01 6.96142077e-01 -6.42547488e-01
9.23284739e-02 1.35433629e-01 -3.89865130e-01 4.48937058e-01
-1.00012504e-01 -4.32101160e-01 -9.26876605e-01 4.43132930e-02
3.49954665e-02 -5.95058084e-01 9.70534086e-01 5.61686158e-01
1.54910731e+00 -2.35935688e-01 -2.28522167e-01 7.65523732e-01
1.17313492e+00 1.01154968e-01 5.76476336e-01 -2.86160335e-02
1.08785152e+00 8.29660177e-01 5.87659359e-01 2.79012561e-01
6.49277270e-01 3.47223163e-01 2.20598295e-01 -4.02492493e-01
-1.81551322e-01 -4.54505295e-01 2.54299998e-01 8.08128655e-01
4.38281409e-02 -1.09951785e-02 -7.72093654e-01 3.73936564e-01
-1.75839722e+00 -5.75113416e-01 1.08487289e-02 1.69179940e+00
9.49959576e-01 2.44958729e-01 2.68284887e-01 -1.25916719e-01
1.04193377e+00 1.91317737e-01 -8.69945407e-01 -5.06130815e-01
1.74683183e-01 3.24952155e-01 3.19387585e-01 2.86365062e-01
-1.21990192e+00 1.43476486e+00 4.60602713e+00 7.22072303e-01
-1.07399631e+00 3.17770600e-01 1.04450679e+00 1.36679158e-01
1.52112216e-01 -2.44530082e-01 -8.40915561e-01 8.00319612e-01
9.23443675e-01 3.03611606e-01 2.24375769e-01 7.41575718e-01
4.02479112e-01 9.82009321e-02 -6.66010261e-01 1.04051864e+00
-9.32256132e-02 -7.02251852e-01 -2.04167087e-02 -3.44364673e-01
1.84149966e-01 -2.30573982e-01 -2.24290401e-01 5.17512858e-01
1.36656642e-01 -1.02011180e+00 8.01664472e-01 2.86650807e-01
4.48947251e-01 -6.03894770e-01 6.93468571e-01 1.95454940e-01
-1.30948818e+00 -1.36167884e-01 -3.34571391e-01 1.92464352e-01
2.68791586e-01 4.51123893e-01 6.65660482e-03 4.90778774e-01
9.46273088e-01 7.04821587e-01 -5.14237463e-01 6.16852403e-01
-4.50313389e-01 7.95246065e-01 -2.15301946e-01 1.85447335e-02
2.85311520e-01 -7.82084092e-02 1.24051504e-01 1.21722972e+00
-1.06910020e-01 3.50081235e-01 3.91966552e-01 1.01287067e+00
-2.27640316e-01 3.30439240e-01 7.03755692e-02 3.26900110e-02
2.00472444e-01 1.40107489e+00 -8.95371854e-01 -4.25024331e-01
-3.04543555e-01 1.28931522e+00 2.86870658e-01 3.20645630e-01
-1.21389544e+00 -5.07037401e-01 3.10809374e-01 -1.63149461e-01
2.27265805e-01 1.43285021e-01 -4.79538321e-01 -1.03887427e+00
4.95340861e-02 -6.54657960e-01 5.09113014e-01 -8.13103914e-01
-1.20736217e+00 4.55596596e-01 -2.16808438e-01 -4.20768529e-01
2.01577172e-01 -4.86137956e-01 -7.17847764e-01 9.49532092e-01
-1.18363309e+00 -1.40029073e+00 -5.56731224e-01 6.30305231e-01
6.97201431e-01 3.38715285e-01 4.92995769e-01 2.71594644e-01
-9.40824330e-01 6.05766654e-01 -5.55020869e-01 5.13937414e-01
3.75896186e-01 -1.23977721e+00 4.71728086e-01 4.46181864e-01
-4.70159382e-01 3.46440405e-01 3.60261768e-01 -7.68372297e-01
-9.74834681e-01 -1.25451291e+00 5.33074856e-01 1.59040857e-02
2.27071941e-01 -5.43048799e-01 -9.43062723e-01 5.83955109e-01
9.61585194e-02 1.45878484e-02 5.25519133e-01 8.80729593e-03
-8.48872587e-02 -5.06933443e-02 -1.28354216e+00 3.99605066e-01
1.20218349e+00 -8.40515867e-02 -7.30504632e-01 1.91217020e-01
1.03356135e+00 -5.51225662e-01 -9.83871281e-01 5.61726749e-01
5.79515696e-01 -8.32359135e-01 8.52859318e-01 -5.64732969e-01
6.59867585e-01 -2.55305946e-01 7.51303881e-02 -1.05335617e+00
-2.81229615e-01 -3.64659101e-01 -5.17217629e-02 1.40536880e+00
1.87806457e-01 -7.03139901e-01 8.52526665e-01 4.04571861e-01
-3.08037490e-01 -1.09883976e+00 -6.37166321e-01 -3.69002074e-01
2.09553361e-01 -5.06341532e-02 4.76586908e-01 6.03788376e-01
-2.45198920e-01 7.00300395e-01 -3.19833010e-01 7.59036047e-03
6.97908700e-01 2.18737632e-01 4.12828714e-01 -1.09535301e+00
-2.75672853e-01 -4.50150609e-01 -2.63290256e-01 -1.00431430e+00
4.33442801e-01 -7.56656170e-01 2.49649808e-01 -1.63787174e+00
3.09655130e-01 -2.80365229e-01 -4.55776811e-01 7.37828135e-01
-5.23705959e-01 1.37690991e-01 1.24603808e-01 -2.94584148e-02
-7.17053533e-01 7.35720992e-01 1.43265533e+00 2.14586258e-02
-1.62417322e-01 -6.56870827e-02 -9.96773541e-01 8.89238358e-01
1.12326717e+00 -3.05205017e-01 -3.31931859e-01 -3.82273167e-01
-4.31713164e-01 1.95588946e-01 5.80951035e-01 -1.12360227e+00
-1.24463566e-01 1.47729099e-01 7.82097280e-01 -3.41168761e-01
2.40364134e-01 -3.24854612e-01 -2.77735263e-01 6.49706244e-01
-4.86395299e-01 -1.94687903e-01 1.13790415e-01 4.42483842e-01
-2.91161478e-01 -1.18840542e-02 9.56237137e-01 -4.67883170e-01
-8.21387053e-01 4.57403392e-01 9.32467058e-02 4.03459132e-01
8.00592303e-01 -1.81881879e-02 2.06160732e-02 3.35807949e-02
-9.71673012e-01 7.41962194e-01 1.05734125e-01 4.71536279e-01
4.28690761e-01 -1.20660496e+00 -6.30020976e-01 -5.83742149e-02
-1.83684573e-01 3.08163226e-01 6.23695552e-01 7.01201200e-01
-5.04214644e-01 2.03723803e-01 -3.73753786e-01 -5.79163492e-01
-1.18696547e+00 4.73022640e-01 5.79260409e-01 -1.96112216e-01
-7.53209114e-01 1.05492413e+00 6.06250048e-01 -4.58452284e-01
2.24340156e-01 -3.13296735e-01 -3.11138511e-01 -8.34903643e-02
2.72881836e-01 5.61275966e-02 -3.89479131e-01 -8.47161293e-01
-4.28029269e-01 8.48866224e-01 1.36598870e-01 1.83567211e-01
1.06276309e+00 -2.53566056e-01 -4.15076539e-02 1.57441348e-01
1.36752462e+00 -3.79804671e-01 -1.30392194e+00 9.83929783e-02
-2.27016583e-01 7.75615126e-02 -4.44363244e-02 -9.24454510e-01
-1.53044808e+00 1.19997430e+00 7.01899648e-01 -2.43200749e-01
1.24994755e+00 1.63188070e-01 1.56184971e+00 -1.78059042e-01
1.30841598e-01 -1.15405035e+00 -6.89833909e-02 2.22553581e-01
7.34609842e-01 -1.29202211e+00 -2.71513909e-01 -5.27068675e-01
-9.48649824e-01 8.96640360e-01 9.57291901e-01 -1.92091718e-01
5.44620335e-01 -9.41002145e-02 -1.87549554e-02 -2.49730676e-01
-3.44476521e-01 -3.20208132e-01 8.87437165e-02 3.33787411e-01
4.82827544e-01 1.94741175e-01 -5.55145621e-01 1.36640596e+00
-2.48676255e-01 1.44992173e-01 -7.67778456e-02 7.98996389e-01
-5.68801641e-01 -7.31409848e-01 -2.85328895e-01 3.27736229e-01
-7.92159557e-01 -1.13423336e-02 -3.96599084e-01 5.42940438e-01
4.04473245e-01 7.99407482e-01 -8.83319750e-02 -3.87090385e-01
3.39205205e-01 2.14278877e-01 3.44608933e-01 -6.59057140e-01
-7.01045454e-01 1.48511872e-01 2.97620688e-02 -8.26863825e-01
-2.96724200e-01 -5.16705155e-01 -2.01735115e+00 7.45297000e-02
1.36599392e-01 1.21125087e-01 5.24305344e-01 9.10035491e-01
2.69746244e-01 7.94039786e-01 3.57316226e-01 -7.82863855e-01
-3.63589257e-01 -9.56392229e-01 -3.86648655e-01 7.17548966e-01
-8.84864926e-02 -7.90088058e-01 -4.13013920e-02 1.72248200e-01] | [8.79185962677002, 0.13781601190567017] |
0a94be85-7523-429a-8fa9-19e61fae4ab1 | dual-progressive-transformations-for-weakly | 2209.15211 | null | https://arxiv.org/abs/2209.15211v1 | https://arxiv.org/pdf/2209.15211v1.pdf | Dual Progressive Transformations for Weakly Supervised Semantic Segmentation | Weakly supervised semantic segmentation (WSSS), which aims to mine the object regions by merely using class-level labels, is a challenging task in computer vision. The current state-of-the-art CNN-based methods usually adopt Class-Activation-Maps (CAMs) to highlight the potential areas of the object, however, they may suffer from the part-activated issues. To this end, we try an early attempt to explore the global feature attention mechanism of vision transformer in WSSS task. However, since the transformer lacks the inductive bias as in CNN models, it can not boost the performance directly and may yield the over-activated problems. To tackle these drawbacks, we propose a Convolutional Neural Networks Refined Transformer (CRT) to mine a globally complete and locally accurate class activation maps in this paper. To validate the effectiveness of our proposed method, extensive experiments are conducted on PASCAL VOC 2012 and CUB-200-2011 datasets. Experimental evaluations show that our proposed CRT achieves the new state-of-the-art performance on both the weakly supervised semantic segmentation task the weakly supervised object localization task, which outperform others by a large margin. | ['Qingyao Wu', 'Yukun Su', 'Dongjian Huo'] | 2022-09-30 | null | null | null | null | ['weakly-supervised-object-localization'] | ['computer-vision'] | [ 2.66093850e-01 2.39309952e-01 -1.95196316e-01 -5.49968004e-01
-7.49876320e-01 -3.19302976e-01 5.31432569e-01 -1.98348016e-01
-4.72566932e-01 5.64546943e-01 -2.02191964e-01 4.49016783e-03
1.32581234e-01 -6.77515328e-01 -8.90130222e-01 -8.18801820e-01
3.68555009e-01 2.07786605e-01 1.03316140e+00 -1.49886057e-01
2.25101948e-01 1.67104781e-01 -1.55580676e+00 3.84278804e-01
1.05434537e+00 1.37541413e+00 4.43306446e-01 -1.25044445e-02
-4.30053473e-01 9.14699256e-01 -4.25157249e-01 -1.73101306e-01
1.82315454e-01 -2.51662493e-01 -1.05128980e+00 6.05846383e-02
4.25049454e-01 7.43927583e-02 2.38788024e-01 1.46378136e+00
1.52642474e-01 -1.57187402e-01 3.97523791e-01 -1.02773368e+00
-4.27696139e-01 5.57692885e-01 -7.88598001e-01 1.55042827e-01
-1.40518814e-01 1.19609788e-01 7.98217118e-01 -1.20642877e+00
3.88376147e-01 1.17955601e+00 7.33468473e-01 5.32490194e-01
-8.23062003e-01 -6.44029498e-01 6.20733082e-01 1.27434641e-01
-1.48926020e+00 -1.11656733e-01 1.05851698e+00 -1.53695479e-01
4.23848331e-01 1.02970660e-01 5.19944489e-01 8.65362287e-01
-4.85929906e-01 1.29627860e+00 1.45847201e+00 -1.73345327e-01
1.75948113e-01 2.73717403e-01 3.24578553e-01 9.33443844e-01
1.36126459e-01 -1.85819879e-01 -3.74207288e-01 1.51910618e-01
6.96445823e-01 5.51739633e-02 -1.57927901e-01 -5.27276874e-01
-1.18373168e+00 5.77808857e-01 1.09360993e+00 4.84807074e-01
-1.60553485e-01 1.75761245e-02 3.12523842e-01 -3.18928331e-01
7.39895403e-01 3.32677215e-01 -6.91898227e-01 2.31934682e-01
-9.79128063e-01 5.99008501e-02 3.70375007e-01 9.69762623e-01
8.20844829e-01 -7.56055266e-02 -3.31362784e-01 8.66741598e-01
2.92131275e-01 -9.22529586e-03 2.94984460e-01 -5.45496166e-01
5.47084093e-01 1.13619590e+00 -2.17851121e-02 -6.91333711e-01
-3.06089908e-01 -9.37468529e-01 -6.92526400e-01 1.04381658e-01
4.90121692e-01 4.34809178e-02 -1.38552237e+00 1.56400025e+00
5.11330366e-01 3.27730536e-01 3.31128947e-02 1.16226399e+00
1.11648047e+00 6.93148375e-01 4.57904607e-01 6.53769001e-02
1.27139866e+00 -1.50388193e+00 -5.34979045e-01 -6.18940592e-01
3.97631824e-01 -5.77322185e-01 1.15039420e+00 3.57153863e-02
-1.01448870e+00 -9.02343154e-01 -9.48783159e-01 7.24643543e-02
-5.55040240e-01 4.59469020e-01 6.69982791e-01 3.08327287e-01
-8.54110658e-01 4.40173388e-01 -8.43465567e-01 -4.31645691e-01
1.15057719e+00 2.27786899e-01 -1.79092005e-01 -6.90428987e-02
-9.85623717e-01 6.29386604e-01 6.46533668e-01 5.21607459e-01
-1.08778489e+00 -5.95542669e-01 -7.40181804e-01 1.82830334e-01
7.26907194e-01 -3.21097642e-01 1.02500224e+00 -1.35223413e+00
-1.23053122e+00 9.53428328e-01 3.09442263e-02 -2.63430387e-01
5.39295495e-01 -2.81599253e-01 8.45505446e-02 1.23194307e-01
2.69952416e-01 9.98022318e-01 7.77995586e-01 -1.47415364e+00
-9.08203542e-01 -6.05932057e-01 1.11191325e-01 3.36784601e-01
-4.26273435e-01 -1.18698321e-01 -8.48159730e-01 -6.84131920e-01
4.26327080e-01 -7.69808173e-01 -4.29580212e-01 1.20992951e-01
-6.46789789e-01 -6.46764338e-01 1.03052628e+00 -2.68171430e-01
7.81805634e-01 -2.05462003e+00 -8.31678808e-02 -6.44887015e-02
-6.24636449e-02 5.94395339e-01 1.91032253e-02 -2.80350745e-01
1.46004513e-01 6.19250163e-02 -7.61493683e-01 -3.29604238e-01
-3.36642712e-01 1.62305057e-01 -5.29229306e-02 4.57089782e-01
5.83248675e-01 1.26608181e+00 -7.99231350e-01 -7.07810223e-01
2.42191076e-01 2.40961343e-01 -2.38698095e-01 4.42091316e-01
-4.60862607e-01 4.51698482e-01 -8.01293433e-01 1.11673665e+00
7.83781648e-01 -3.15632820e-01 -2.38799915e-01 -4.42082822e-01
-1.72648251e-01 -8.50633234e-02 -9.56308067e-01 1.97677255e+00
-2.67941202e-03 3.06115180e-01 1.35991871e-01 -1.50786757e+00
8.58741105e-01 -9.12543610e-02 3.40469539e-01 -7.74139404e-01
1.52553454e-01 3.85889620e-01 -1.63147181e-01 -5.22909999e-01
1.00965835e-01 3.51058356e-02 1.42192449e-02 -1.00209683e-01
2.09709167e-01 6.11852109e-02 -5.95073625e-02 -4.62348312e-02
5.69604099e-01 5.86122453e-01 -6.45448193e-02 -7.85383821e-01
7.37013400e-01 2.54429281e-01 7.63520241e-01 7.22066641e-01
-4.08121556e-01 7.44130790e-01 4.19881850e-01 -3.26515138e-01
-6.67734504e-01 -9.13429022e-01 -3.10341775e-01 1.12323844e+00
7.28986084e-01 -5.22984006e-03 -1.14677584e+00 -1.25778949e+00
-2.77911872e-01 2.16524035e-01 -7.17227399e-01 -1.10495016e-01
-5.79907179e-01 -8.84201109e-01 5.30880094e-01 9.39607680e-01
1.21956062e+00 -1.29316008e+00 -5.38745821e-01 3.19030546e-02
-2.23043069e-01 -1.38613474e+00 -2.46638328e-01 2.59316713e-01
-8.38480651e-01 -1.18608618e+00 -1.14401162e+00 -1.40424955e+00
1.04178631e+00 2.63945580e-01 8.82977247e-01 9.85523537e-02
-2.29001567e-01 -1.27453551e-01 -3.90268505e-01 -4.29887593e-01
2.36560538e-01 2.88657188e-01 -4.27117825e-01 2.46166527e-01
3.48229975e-01 -3.37204710e-02 -8.81817460e-01 4.99964863e-01
-7.80629754e-01 2.28681400e-01 8.45898092e-01 8.85580361e-01
8.97235811e-01 -1.39628798e-02 7.60156691e-01 -1.19329071e+00
1.19362734e-01 -3.60184431e-01 -6.47364736e-01 2.54670352e-01
-6.29163861e-01 -1.34599641e-01 4.44207728e-01 -2.67879426e-01
-1.35059202e+00 4.31363791e-01 -3.01463336e-01 -1.94853321e-01
-4.00005668e-01 4.42751527e-01 -2.93788224e-01 -2.33057454e-01
4.00258243e-01 4.04315859e-01 -4.44062591e-01 -6.71893418e-01
2.35104486e-01 6.23746812e-01 6.58285141e-01 -5.89043915e-01
6.60890102e-01 6.45561457e-01 -2.25265071e-01 -4.17502612e-01
-1.50883913e+00 -8.04605246e-01 -6.36737764e-01 -9.66668949e-02
1.10941815e+00 -1.07135773e+00 -3.38292450e-01 6.20535672e-01
-1.03589189e+00 -3.53111386e-01 -2.09118098e-01 6.21823668e-02
-4.14016783e-01 1.19089440e-01 -3.65673125e-01 -6.74511671e-01
-3.98188919e-01 -1.31730115e+00 1.34248543e+00 5.48802674e-01
3.67536813e-01 -5.43837547e-01 -4.51185346e-01 5.80436289e-01
4.72997695e-01 3.11599314e-01 6.57552719e-01 -6.54610932e-01
-7.24268377e-01 -6.58025444e-02 -8.20563793e-01 4.83835399e-01
-7.76903257e-02 -3.82057607e-01 -1.28520775e+00 -6.20788634e-02
-4.87313531e-02 -5.52567840e-01 1.25202549e+00 3.10591906e-01
1.86171997e+00 -1.04355000e-01 -4.64575112e-01 7.25913107e-01
1.53968012e+00 2.25282516e-02 4.58946139e-01 2.88679004e-01
1.04575086e+00 7.74904251e-01 1.00339961e+00 -5.90849854e-02
2.88232744e-01 4.63220716e-01 7.98941731e-01 -6.47173643e-01
-2.23187268e-01 -3.52365345e-01 2.27188542e-02 3.87337297e-01
-5.98563850e-02 -4.90416251e-02 -7.12828934e-01 9.22349393e-01
-2.01773262e+00 -3.31332266e-01 -2.48338625e-01 1.73970127e+00
8.25363398e-01 5.00670850e-01 -3.58729772e-02 1.13093190e-01
7.26552665e-01 2.23434031e-01 -6.11905038e-01 1.17539927e-01
-1.43749326e-01 2.15582758e-01 5.55531442e-01 3.73249389e-02
-1.38432860e+00 1.38940394e+00 5.07016706e+00 1.08104885e+00
-1.00924301e+00 3.05335820e-01 9.47687328e-01 2.83307612e-01
1.28703982e-01 -7.32944533e-02 -8.82716179e-01 4.53608304e-01
2.86400259e-01 6.02740824e-01 -1.06550604e-01 1.15608275e+00
-1.58092484e-01 -1.75714269e-01 -9.01226640e-01 8.76305103e-01
-8.58344324e-03 -1.18848372e+00 3.97670269e-02 -3.46029490e-01
9.84520018e-01 1.73116416e-01 2.11250801e-02 4.75598365e-01
-1.67284414e-01 -1.02455032e+00 9.65110660e-01 2.61541724e-01
6.47159040e-01 -6.12630427e-01 1.06548095e+00 5.53728223e-01
-1.31496465e+00 -2.45737523e-01 -4.96551573e-01 2.34804690e-01
-1.14202634e-01 5.52842140e-01 -4.46726859e-01 2.85363674e-01
1.09699762e+00 7.41014481e-01 -8.07260871e-01 1.11818874e+00
-3.60985786e-01 9.18440700e-01 -3.14764351e-01 -1.24766260e-01
7.39843845e-01 5.21899387e-03 2.90404588e-01 1.26930487e+00
2.68418528e-02 1.79253608e-01 3.41293126e-01 1.32302475e+00
-1.78550214e-01 7.14519471e-02 -1.73973143e-01 1.61191270e-01
6.81812838e-02 1.43999052e+00 -1.20487154e+00 -3.54042560e-01
-2.17089072e-01 9.05004859e-01 3.61452311e-01 5.43616116e-01
-7.03378558e-01 -3.45102668e-01 2.41471812e-01 1.90939188e-01
3.61024827e-01 6.85099512e-02 -5.12149811e-01 -9.53959525e-01
1.85967982e-01 -4.83725339e-01 3.44960600e-01 -7.63353467e-01
-1.12294519e+00 6.20665550e-01 -4.84994017e-02 -1.02027225e+00
3.48367870e-01 -6.28566623e-01 -6.92552030e-01 6.06224537e-01
-1.97109652e+00 -1.54623508e+00 -6.16602123e-01 6.56563044e-01
1.06208599e+00 -2.61736545e-03 3.30089003e-01 2.15969861e-01
-5.88391721e-01 4.17905331e-01 -3.42604071e-01 2.13460714e-01
3.04603517e-01 -1.30924106e+00 1.47725850e-01 8.06381285e-01
6.99546933e-02 4.32634085e-01 3.29938233e-01 -5.28996885e-01
-9.33297515e-01 -1.46230710e+00 4.23960537e-01 -2.65310496e-01
2.45552912e-01 -5.56151390e-01 -1.02474582e+00 2.99285650e-01
7.84296170e-02 6.32431626e-01 -4.64159697e-02 -1.78668424e-01
-9.97097790e-02 -2.85692662e-01 -1.21732843e+00 3.97192895e-01
1.22189498e+00 -3.23825568e-01 -6.18461430e-01 3.81652564e-01
8.52130771e-01 -3.02271515e-01 -5.44763565e-01 9.58519757e-01
1.25636235e-01 -9.17863190e-01 9.98149395e-01 -3.24811518e-01
3.13128382e-01 -6.15749538e-01 1.02049589e-01 -1.04803395e+00
-4.89869639e-02 -7.62954308e-03 8.97362232e-02 1.58382201e+00
3.90719265e-01 -4.50581759e-01 1.04324472e+00 1.45555213e-01
-3.75423342e-01 -1.09188616e+00 -7.83032417e-01 -5.76333880e-01
-6.32455647e-02 -2.86348462e-01 5.10929406e-01 8.69915843e-01
-4.05650467e-01 2.25489914e-01 1.03431687e-01 1.92530021e-01
7.62056291e-01 2.12614119e-01 3.69397372e-01 -1.24459994e+00
1.79371625e-01 -4.65134859e-01 -1.87624842e-01 -1.20236194e+00
1.41834810e-01 -7.79876292e-01 6.28963709e-01 -1.66072667e+00
3.33669692e-01 -7.43866742e-01 -7.23773539e-01 7.16112137e-01
-4.85990942e-01 5.64820945e-01 -2.81274486e-02 1.95725903e-01
-1.09569013e+00 8.25320065e-01 1.38489974e+00 -2.86787033e-01
7.49470666e-02 1.07583717e-01 -7.52751708e-01 1.01853716e+00
7.09316254e-01 -4.24510747e-01 -4.77763027e-01 -3.97549778e-01
-2.30989292e-01 -3.99660736e-01 6.30470693e-01 -9.57127035e-01
2.73301154e-01 -1.01500697e-01 4.68102604e-01 -8.25053692e-01
2.11761802e-01 -8.07124078e-01 -5.32087028e-01 4.37409580e-01
-3.01222861e-01 -4.93825734e-01 1.51241705e-01 6.22488678e-01
-5.32279670e-01 -1.49030790e-01 1.05481923e+00 -4.37739372e-01
-1.09430504e+00 5.03412902e-01 7.65092224e-02 1.18465640e-01
1.06658292e+00 -2.78359205e-01 -2.53387362e-01 2.90031403e-01
-3.25549513e-01 3.96817356e-01 1.73650429e-01 5.63371241e-01
6.66569591e-01 -1.01484931e+00 -4.60911840e-01 2.61692196e-01
5.25056243e-01 8.23888302e-01 9.19762179e-02 7.56181598e-01
-4.11330849e-01 3.93350750e-01 7.34716700e-03 -1.03678322e+00
-9.47955787e-01 4.58059311e-01 3.83147329e-01 -1.69781502e-02
-4.60438132e-01 1.24080539e+00 7.29415536e-01 -4.91855890e-01
4.35117394e-01 -3.26260746e-01 -5.30869663e-01 -1.70587212e-01
2.10623801e-01 -4.47623767e-02 -3.99060547e-03 -6.50544167e-01
-5.43885529e-01 6.88252509e-01 -7.09620165e-03 3.87283266e-01
1.28298330e+00 -1.55312240e-01 -9.28684026e-02 1.15612626e-01
1.10012281e+00 -5.38033664e-01 -1.55985236e+00 -3.97772521e-01
2.82617092e-01 -3.00855041e-01 1.92492962e-01 -9.25575018e-01
-1.52370191e+00 9.44829166e-01 8.94641697e-01 -2.21162904e-02
1.03530455e+00 2.91660339e-01 8.90030384e-01 1.81663275e-01
3.91063273e-01 -1.22450125e+00 1.73564970e-01 2.53846198e-01
6.15097761e-01 -1.61119342e+00 -3.14976752e-01 -7.25544453e-01
-5.46020925e-01 8.10640872e-01 1.22654688e+00 -2.04342872e-01
6.61869586e-01 1.78753659e-02 3.71890701e-02 -3.73111278e-01
-1.61298469e-01 -6.24888897e-01 4.27083939e-01 5.14763355e-01
2.66444385e-01 -1.04931258e-01 -3.34386855e-01 7.79448092e-01
3.63250136e-01 -9.95657817e-02 3.99955735e-02 8.53791714e-01
-6.94800079e-01 -6.47793472e-01 -2.58090138e-01 3.58962208e-01
-6.27733707e-01 -4.58501168e-02 -5.32054007e-01 6.48391128e-01
5.23686469e-01 8.72911513e-01 -1.10225648e-01 -4.66589667e-02
3.06432188e-01 -1.46557704e-01 2.33057588e-01 -5.91417074e-01
-5.76188266e-01 1.77487701e-01 -1.64752647e-01 -6.85034811e-01
-8.23530436e-01 -3.32845747e-01 -1.47767043e+00 4.74258900e-01
-5.74727654e-01 9.49411094e-02 6.71835542e-01 1.15033937e+00
1.95889082e-02 6.76847994e-01 3.89969915e-01 -7.69730628e-01
-4.33923692e-01 -1.04785025e+00 -2.94071466e-01 5.17486036e-01
2.37042278e-01 -7.33991683e-01 -6.60779253e-02 -1.10326625e-01] | [9.564949989318848, 0.6073523163795471] |
b05b2089-6995-43c4-a3ca-50b49399a9b1 | n2sky-neural-networks-as-services-in-the | 1401.2468 | null | http://arxiv.org/abs/1401.2468v1 | http://arxiv.org/pdf/1401.2468v1.pdf | N2Sky - Neural Networks as Services in the Clouds | We present the N2Sky system, which provides a framework for the exchange of
neural network specific knowledge, as neural network paradigms and objects, by
a virtual organization environment. It follows the sky computing paradigm
delivering ample resources by the usage of federated Clouds. N2Sky is a novel
Cloud-based neural network simulation environment, which follows a pure service
oriented approach. The system implements a transparent environment aiming to
enable both novice and experienced users to do neural network research easily
and comfortably. N2Sky is built using the RAVO reference architecture of
virtual organizations which allows itself naturally integrating into the Cloud
service stack (SaaS, PaaS, and IaaS) of service oriented architectures. | ['Erwin Mann', 'Erich Schikuta'] | 2014-01-10 | null | null | null | null | ['neural-network-simulation'] | ['computer-code'] | [-9.09303784e-01 -2.62377918e-01 2.38499045e-01 -2.44017079e-01
5.02270162e-01 -6.41859114e-01 7.74662197e-01 -2.90153861e-01
-2.75076866e-01 4.10673201e-01 -3.32266659e-01 -3.85174513e-01
-3.88360590e-01 -9.37590003e-01 -2.26338953e-01 -8.18205357e-01
7.40317479e-02 5.07976532e-01 2.19952151e-01 -4.94659215e-01
6.54044226e-02 1.13861275e+00 -2.21079755e+00 1.11630373e-01
4.96042401e-01 1.22928929e+00 2.13680491e-01 6.51841521e-01
-5.30139923e-01 9.56593335e-01 -5.14203012e-01 -3.77176136e-01
2.54644185e-01 2.75265515e-01 -6.79502964e-01 -5.99681735e-01
1.74128786e-01 9.92102772e-02 -4.59309630e-02 9.88056600e-01
3.84195179e-01 1.44442990e-01 -8.00713822e-02 -1.74204469e+00
-6.55435264e-01 2.85561740e-01 6.44885242e-01 4.27898645e-01
1.58569142e-01 1.22359343e-01 2.97537059e-01 -6.66470110e-01
1.11309302e+00 7.43568301e-01 3.88910621e-01 7.12773979e-01
-9.07183051e-01 -5.49593151e-01 1.15151703e-01 5.84222913e-01
-1.42075765e+00 -3.34127992e-01 2.72725582e-01 -2.44159043e-01
1.28485262e+00 8.53018403e-01 1.27766490e+00 8.35412979e-01
3.35622519e-01 2.29712263e-01 1.16243696e+00 -2.21278057e-01
8.82503510e-01 6.76510274e-01 5.93091011e-01 2.49279812e-01
3.88762832e-01 7.57576674e-02 -6.66534126e-01 -2.01111555e-01
8.38638067e-01 3.00875872e-01 -6.48275763e-02 -4.25323337e-01
-8.33026707e-01 1.12001806e-01 5.71068764e-01 6.58421874e-01
-9.76777017e-01 3.50483745e-01 5.14509439e-01 5.99104583e-01
1.89807624e-01 3.72461230e-02 -6.12359822e-01 -3.81780893e-01
-5.81970751e-01 5.45985639e-01 1.04168189e+00 1.01283336e+00
2.51443118e-01 6.18507147e-01 3.94788265e-01 3.23868036e-01
5.47167838e-01 1.98858827e-01 9.82968807e-01 -1.32229209e+00
-6.41291738e-01 6.85663939e-01 2.27649342e-02 -5.99183738e-01
1.51459901e-02 -9.62966383e-01 -7.58001283e-02 6.96841955e-01
-3.20147097e-01 8.75024050e-02 -5.45383751e-01 1.27750945e+00
4.86671716e-01 3.82091463e-01 5.84531724e-01 8.50222588e-01
1.19988751e+00 1.59787774e-01 -5.61654754e-02 -1.02033727e-01
1.14550459e+00 -1.20291460e+00 -8.33548069e-01 3.87382656e-01
2.02852055e-01 -7.87774250e-02 8.62574816e-01 6.23159766e-01
-1.34963393e+00 -1.65226042e-01 -7.48378456e-01 2.58210152e-01
-1.19936895e+00 -7.13001251e-01 1.06259573e+00 1.10946035e+00
-1.73662424e+00 5.28083920e-01 -7.34202147e-01 -4.88872647e-01
2.12962538e-01 4.52982008e-01 -5.84610939e-01 1.87376454e-01
-8.10191572e-01 1.01340437e+00 2.29577720e-01 7.19428062e-02
-1.19397426e+00 -5.07510841e-01 -3.59295428e-01 5.90048671e-01
-3.87370251e-02 -1.01929164e+00 1.11757028e+00 -1.26286983e+00
-1.58449399e+00 7.42773354e-01 2.05305785e-01 -5.99270761e-01
3.42479229e-01 1.63232401e-01 -8.47223580e-01 1.92197368e-01
-4.06587154e-01 4.58068438e-02 2.33465433e-01 -1.03302324e+00
-3.85247529e-01 -8.68579209e-01 1.15298882e-01 1.69916987e-01
-3.32459927e-01 4.81894255e-01 -9.99326929e-02 -8.35252274e-03
2.76077036e-02 -6.05571151e-01 -4.58994627e-01 -9.70755368e-02
3.72219533e-01 -1.45517707e-01 7.50190973e-01 -8.62084255e-02
5.35981119e-01 -1.94910848e+00 -1.74375385e-01 3.77921134e-01
5.39657831e-01 3.02618086e-01 9.55694318e-02 5.48649788e-01
-1.52984515e-01 3.28549580e-03 6.20097458e-01 1.46799117e-01
1.20336808e-01 3.27242941e-01 1.05020538e-01 -6.92925081e-02
-7.34045148e-01 4.92415935e-01 -6.54152870e-01 -7.47155100e-02
-1.64088026e-01 6.23790324e-01 -3.47117692e-01 2.40727261e-01
1.03524901e-01 -9.54944491e-02 -2.42575452e-01 8.21363747e-01
7.90310085e-01 -3.62770915e-01 3.18037271e-01 3.84315729e-01
-5.25761604e-01 -1.01528615e-01 -1.10782564e+00 1.68893993e+00
-5.75044811e-01 3.75826389e-01 4.29670185e-01 -6.75753891e-01
1.01306891e+00 9.59160089e-01 3.27813953e-01 -4.48666692e-01
2.16086969e-01 3.77611428e-01 -2.00382307e-01 -6.21438563e-01
2.16806307e-01 1.01469375e-01 6.20410800e-01 2.47222871e-01
6.26399755e-01 1.70277864e-01 -1.70847073e-01 2.81115234e-01
1.20224833e+00 3.67631286e-01 2.47974526e-02 -4.56241906e-01
3.92373919e-01 -6.04476854e-02 4.72352058e-01 4.44870472e-01
-5.76005459e-01 -1.20889552e-01 1.96429387e-01 -6.68547213e-01
-7.63017833e-01 -1.39911175e+00 -1.17609620e-01 9.27862227e-01
-1.52150080e-01 -2.52450407e-01 -1.09063697e+00 -2.36029983e-01
-1.86529055e-01 9.03143883e-01 -3.85037571e-01 1.21830180e-01
4.82858717e-01 -1.18714221e-01 1.20237842e-01 1.94058239e-01
4.75501686e-01 -1.41192496e+00 -9.17085588e-01 3.15628238e-02
4.45563555e-01 -9.77871716e-01 4.33907449e-01 6.06706627e-02
-1.03926790e+00 -6.07318759e-01 -3.85767192e-01 -4.25890207e-01
2.02686578e-01 4.45324272e-01 1.34432542e+00 2.82163620e-01
-2.16535702e-01 7.82757461e-01 -1.01225890e-01 -7.02498317e-01
2.02831663e-02 -2.78461933e-01 3.32402259e-01 1.46324784e-01
5.79363167e-01 -1.28845346e+00 -4.83437955e-01 -1.64363131e-01
-7.93586195e-01 -1.27231693e-02 -1.38881132e-01 2.35929087e-01
3.31433773e-01 -2.27857396e-01 7.96132207e-01 -9.73256350e-01
5.94600856e-01 -8.13968599e-01 -9.11426544e-01 2.04950511e-01
-1.39031649e+00 -7.01377213e-01 5.06438076e-01 4.80449498e-02
-8.63733709e-01 -3.29150796e-01 -1.40062436e-01 -7.57759392e-01
-6.28385246e-01 6.13669038e-01 -2.16331318e-01 -5.99855721e-01
6.90802574e-01 5.54949403e-01 2.50479400e-01 -7.29112983e-01
1.81639835e-01 1.07077956e+00 5.68973422e-01 -3.53731923e-02
1.43308535e-01 5.41819572e-01 -3.02572399e-01 -4.21263188e-01
1.44841805e-01 -5.64083755e-01 -1.84095249e-01 -6.72403395e-01
5.10500014e-01 -7.86232293e-01 -1.31247163e+00 4.82565045e-01
-9.96586382e-01 -1.07463151e-01 -3.94503027e-01 5.20492315e-01
-5.44252217e-01 -4.78015184e-01 -5.29994845e-01 -9.91479516e-01
-8.62462521e-01 -8.05646062e-01 1.11503191e-02 7.80658960e-01
3.06536645e-01 -8.60838652e-01 2.52123922e-01 4.74908292e-01
1.17345309e+00 -1.08270980e-01 5.29576421e-01 -9.70501482e-01
-7.45975375e-01 -3.26882571e-01 -3.49184088e-02 3.47171724e-01
-5.16454101e-01 3.51721913e-01 -1.29939747e+00 8.72573853e-02
3.03059340e-01 4.74588089e-02 -8.65577534e-02 5.83844744e-02
9.59619999e-01 -2.45566756e-01 -7.38340914e-02 7.73098111e-01
1.96772850e+00 3.22900891e-01 7.29008377e-01 8.89531136e-01
8.76893029e-02 6.70348167e-01 1.27318025e-01 2.95510262e-01
6.11753404e-01 5.37265599e-01 1.11567378e+00 1.36989132e-01
3.96608800e-01 4.69165474e-01 -1.13740573e-02 1.09331822e+00
-6.46489382e-01 3.63236576e-01 -1.11300981e+00 5.44068575e-01
-1.89422035e+00 -1.05152321e+00 -3.75469863e-01 2.05867815e+00
2.95790464e-01 -1.48045108e-01 2.06350386e-02 -2.00097397e-01
3.91887039e-01 -4.14735019e-01 -4.02913898e-01 -1.03800344e+00
1.50951684e-01 3.23778778e-01 1.60981238e-01 -2.03233048e-01
-7.90634844e-03 7.69365311e-01 7.31366396e+00 4.10676956e-01
-1.36738944e+00 7.85784602e-01 1.40792876e-02 -4.42783177e-01
-4.63556230e-01 9.58893970e-02 -4.54350024e-01 4.03845191e-01
1.89658380e+00 -9.00160193e-01 9.86792445e-01 1.54065967e+00
2.45672002e-01 3.45928699e-01 -3.04998457e-01 8.43349159e-01
-2.44165257e-01 -2.10456014e+00 -1.93444535e-01 5.36390170e-02
3.53950977e-01 9.32213604e-01 -1.36770859e-01 1.91521838e-01
1.62349224e-01 -7.86409855e-01 1.00132620e+00 8.77085328e-01
3.49105984e-01 -8.35078239e-01 8.42277765e-01 3.75455856e-01
-7.86588311e-01 -2.28971869e-01 -2.31789559e-01 -1.46866769e-01
-2.94275999e-01 1.08829588e-01 -1.67290956e-01 6.59805954e-01
1.26812375e+00 5.07525392e-02 -3.41779530e-01 1.34334624e+00
4.39007074e-01 3.18838596e-01 7.96997324e-02 -1.48149952e-01
4.58579324e-02 -5.10052860e-01 4.04386401e-01 8.67288411e-01
1.63553730e-01 -4.92252642e-04 -4.69923407e-01 8.65082443e-01
2.57781833e-01 4.22267705e-01 -8.82566750e-01 1.37119606e-01
5.46230257e-01 1.85748470e+00 -6.34853423e-01 -5.08163691e-01
-3.53345901e-01 6.06124043e-01 2.94558462e-02 5.20180523e-01
-4.76607680e-01 -5.65715969e-01 9.40327108e-01 7.63111785e-02
6.98312512e-03 3.15353349e-02 -1.65471524e-01 -1.15856481e+00
-2.77130574e-01 -1.03973758e+00 -7.83709288e-02 -1.47153521e+00
-8.78962815e-01 1.44831288e+00 -3.51144344e-01 -7.62828946e-01
-2.59459645e-01 -6.93508625e-01 -6.63845301e-01 9.21109200e-01
-1.21567464e+00 -1.25750756e+00 -7.54786968e-01 8.84966075e-01
-1.76997073e-02 -7.96833456e-01 1.31838095e+00 4.28172261e-01
-7.10546792e-01 -2.58414727e-02 2.89076000e-01 -4.36311543e-01
-7.77866095e-02 -1.18921113e+00 1.67326376e-01 6.70735359e-01
2.33052075e-02 8.79870474e-01 7.35211670e-01 -3.09001684e-01
-1.55200303e+00 -5.73401093e-01 8.50803077e-01 -1.27753526e-01
5.34683049e-01 -2.82207698e-01 -9.00452733e-01 7.28717208e-01
4.98088390e-01 5.52964211e-01 9.72169816e-01 2.18102619e-01
-2.14704126e-01 -7.05117166e-01 -1.53299963e+00 6.39317691e-01
8.69642794e-01 -5.58310032e-01 -7.34008923e-02 5.63658118e-01
7.23199725e-01 1.21998772e-01 -1.19013703e+00 -2.76812702e-01
5.17877281e-01 -1.33790636e+00 7.07965314e-01 -9.06004846e-01
-1.38104437e-02 -1.69398382e-01 -2.55752444e-01 -1.06609952e+00
-1.61628619e-01 -8.16453874e-01 -3.71862680e-01 9.82958436e-01
6.61549121e-02 -1.29030299e+00 9.63150442e-01 1.10667133e+00
-2.52594560e-01 -4.08893496e-01 -1.24171734e+00 -9.40798104e-01
-3.94837469e-01 -3.99056166e-01 1.27603650e+00 1.09876060e+00
1.59280390e-01 -2.21892059e-01 5.51549196e-01 -5.94226159e-02
5.06014049e-01 -1.16892911e-01 5.41352868e-01 -1.46589923e+00
-2.66678482e-01 -5.34388542e-01 -8.88278127e-01 2.62058258e-01
-8.09264705e-02 -9.93169963e-01 -5.78205109e-01 -1.49005151e+00
-1.15054008e-02 -4.91692722e-01 -7.26778984e-01 3.58389795e-01
6.87556922e-01 3.16169173e-01 3.93123150e-01 2.96421707e-01
-5.12790918e-01 2.35980630e-01 6.87110841e-01 5.91392636e-01
1.20697699e-01 1.55975804e-01 -5.61598241e-01 7.06058800e-01
6.80839360e-01 -5.14550686e-01 -4.55189288e-01 -3.19231093e-01
3.03916454e-01 2.42706195e-01 6.28509104e-01 -1.39741123e+00
5.01835823e-01 -2.83090144e-01 -1.51630312e-01 -1.60695300e-01
2.61269122e-01 -1.36679649e+00 1.12647128e+00 6.91159070e-01
-6.22176453e-02 6.04360878e-01 -7.89761022e-02 1.56514406e-01
-1.40607998e-01 -6.23479784e-01 2.82798439e-01 -4.18587625e-01
-7.01058090e-01 3.40821892e-01 -6.91497445e-01 -7.84093678e-01
1.41536880e+00 -3.30073059e-01 -5.93986452e-01 -1.42787099e-01
-1.13564539e+00 -1.20687678e-01 9.21517074e-01 2.06820846e-01
5.94629765e-01 -1.12231565e+00 -3.27916116e-01 4.85085458e-01
5.33799641e-04 -6.61059439e-01 4.53270882e-01 6.16706908e-01
-1.15368593e+00 5.26124597e-01 -9.59405541e-01 -1.36906728e-01
-1.20287776e+00 8.41262579e-01 8.39062452e-01 1.07098445e-01
-6.50443077e-01 7.14467824e-01 -4.27276850e-01 -7.44493246e-01
2.11821467e-01 3.89614254e-01 -7.41953731e-01 -2.18306929e-01
8.86727273e-01 5.19934952e-01 5.36342919e-01 -1.73183993e-01
-4.51016933e-01 -3.51356179e-01 2.25480780e-01 -4.28822339e-01
1.65032959e+00 -1.30835459e-01 -5.95223844e-01 6.57677650e-01
6.00805521e-01 -1.72773451e-01 -6.82073236e-01 2.93068141e-01
1.40060812e-01 -4.88535315e-01 5.45787156e-01 -1.11218834e+00
-1.54181349e+00 6.92861259e-01 1.18264890e+00 5.73397517e-01
1.10802436e+00 -3.76040488e-01 6.91193044e-02 2.89868027e-01
7.98639596e-01 -9.79348004e-01 -6.65467381e-01 1.48353130e-01
6.45965159e-01 -4.29741114e-01 -4.37132239e-01 -1.17663577e-01
-4.74216938e-01 1.18874514e+00 7.61876583e-01 -2.72437125e-01
9.89773989e-01 5.07290602e-01 6.52391493e-01 -6.82385743e-01
-1.66212153e+00 -3.33547778e-02 -4.24837351e-01 1.03096092e+00
2.38045886e-01 2.33338118e-01 -4.49044079e-01 9.82406080e-01
5.92209809e-02 7.09364235e-01 7.89361477e-01 8.15662920e-01
-5.31137995e-02 -8.63022804e-01 -1.25083134e-01 2.95754433e-01
-7.93033481e-01 -1.96271285e-01 2.75939256e-02 3.57925504e-01
1.68088660e-01 5.37367582e-01 2.15342760e-01 -2.33164385e-01
3.03248554e-01 6.78894147e-02 7.30050355e-02 -2.58549243e-01
-1.61640143e+00 -3.76442581e-01 2.21165437e-02 -8.67961407e-01
-7.63312057e-02 -4.89934713e-01 -1.28294945e+00 -7.61679649e-01
-1.83910146e-01 4.98954296e-01 1.73187947e+00 5.59065700e-01
8.24683785e-01 7.63548076e-01 5.22644937e-01 -7.06962287e-01
-3.87213469e-01 -4.96646762e-01 -1.19858611e+00 -1.04411751e-01
-4.09302652e-01 -3.87487471e-01 -8.99861753e-02 -2.95740932e-01] | [8.173344612121582, 2.9691081047058105] |
3ed1e3b0-eb0d-48b3-af4a-540487b8a7ab | texrel-a-green-family-of-datasets-for | 2105.12804 | null | https://arxiv.org/abs/2105.12804v1 | https://arxiv.org/pdf/2105.12804v1.pdf | TexRel: a Green Family of Datasets for Emergent Communications on Relations | We propose a new dataset TexRel as a playground for the study of emergent communications, in particular for relations. By comparison with other relations datasets, TexRel provides rapid training and experimentation, whilst being sufficiently large to avoid overfitting in the context of emergent communications. By comparison with using symbolic inputs, TexRel provides a more realistic alternative whilst remaining efficient and fast to learn. We compare the performance of TexRel with a related relations dataset Shapeworld. We provide baseline performance results on TexRel for sender architectures, receiver architectures and end-to-end architectures. We examine the effect of multitask learning in the context of shapes, colors and relations on accuracy, topological similarity and clustering precision. We investigate whether increasing the size of the latent meaning space improves metrics of compositionality. We carry out a case-study on using TexRel to reproduce the results of an experiment in a recent paper that used symbolic inputs, but using our own non-symbolic inputs, from TexRel, instead. | ['Hugh Perkins'] | 2021-05-26 | null | null | null | null | ['emergent-communications-on-relations'] | ['natural-language-processing'] | [ 1.54484034e-01 1.58337027e-01 4.69330281e-01 -1.89974383e-01
-3.19443017e-01 -6.43215716e-01 1.09902692e+00 3.25517058e-01
-3.01101565e-01 6.70184255e-01 5.00592887e-01 -4.76179391e-01
-5.99626720e-01 -8.53718221e-01 -6.58374965e-01 -4.98233885e-01
-7.14673996e-01 7.70102084e-01 1.39432430e-01 -4.75104779e-01
1.85013428e-01 3.08511585e-01 -1.50998127e+00 6.67534769e-01
7.34832287e-01 8.74597609e-01 -1.29046887e-01 8.63365293e-01
2.08750814e-01 1.01436114e+00 -8.60422850e-01 -1.51741490e-01
2.19065726e-01 -4.17320907e-01 -1.25194883e+00 -1.14213370e-01
5.08289099e-01 2.53799617e-01 -1.78342462e-01 1.39484599e-01
6.20709598e-01 2.91896850e-01 7.09728897e-01 -1.35299575e+00
-8.18890750e-01 1.02055192e+00 -1.07120000e-01 1.88465774e-01
4.18107837e-01 1.63299635e-01 1.20807266e+00 -4.09708619e-01
6.15373254e-01 1.41928196e+00 8.49711835e-01 -9.06319171e-03
-1.83820367e+00 -5.40928483e-01 -1.47683751e-02 -1.38815045e-01
-1.25524867e+00 -6.72109962e-01 5.26822567e-01 -5.32577813e-01
1.35606563e+00 3.73348325e-01 7.37302065e-01 1.27745330e+00
-2.11495589e-02 2.69470215e-01 1.29037905e+00 -3.08245629e-01
2.93111861e-01 -9.56827775e-02 -2.39411652e-01 5.70745051e-01
-1.63176283e-01 2.40280956e-01 -6.33848667e-01 -4.17016149e-01
9.59731698e-01 -5.98036706e-01 -1.94395296e-02 -4.00907516e-01
-1.67746544e+00 4.22905385e-01 7.35550284e-01 5.99158406e-01
-9.92414951e-02 4.95365709e-01 3.79981339e-01 9.45669115e-01
6.00107729e-01 9.90042686e-01 -6.53307021e-01 -3.98422748e-01
-5.71253538e-01 2.78300703e-01 1.08331227e+00 8.63998175e-01
5.47849894e-01 -2.01623887e-01 -2.61119246e-01 7.76820719e-01
-1.92746222e-01 9.36408155e-03 2.25002542e-01 -1.07539248e+00
3.37745100e-01 6.16556227e-01 -6.72058389e-02 -9.79664266e-01
-8.02851856e-01 -7.12844014e-01 -7.09545493e-01 8.98451582e-02
7.72167504e-01 -2.39503682e-01 -4.40102816e-01 1.68507695e+00
-5.67266084e-02 2.56657582e-02 1.83349609e-01 7.26819396e-01
9.29556251e-01 4.55470562e-01 -9.93108302e-02 1.07946098e-01
1.13207912e+00 -8.89831066e-01 -8.45428631e-02 2.07618132e-01
9.57415700e-01 -8.25855374e-01 1.43710530e+00 5.10772347e-01
-8.23279381e-01 -4.86406803e-01 -6.31090760e-01 -2.25824267e-01
-5.38742065e-01 2.02236593e-01 8.53362858e-01 2.63417304e-01
-1.31102133e+00 7.56134212e-01 -5.36994576e-01 -7.79523075e-01
1.83794498e-01 3.19505036e-01 -1.64992407e-01 6.77632213e-01
-9.33527052e-01 1.02313936e+00 8.94111395e-02 -3.63425702e-01
-6.12950981e-01 -5.39911568e-01 -3.21464747e-01 1.89014375e-02
2.07295626e-01 -5.88542521e-01 1.13721275e+00 -9.42126155e-01
-1.39209080e+00 7.93677866e-01 3.00679117e-01 -3.66802424e-01
4.78363961e-01 1.92542300e-02 -1.64148360e-01 -1.69815555e-01
-6.15785867e-02 8.38754952e-01 4.36421275e-01 -1.10123289e+00
-3.30607407e-02 7.38341287e-02 4.33815181e-01 -1.34890214e-01
-1.61369517e-01 -7.78349712e-02 1.74150184e-01 -8.45562816e-01
-2.35328138e-01 -1.13840675e+00 2.12916136e-02 -1.04051597e-01
-3.21652353e-01 -4.73706245e-01 6.53241158e-01 -1.80723608e-01
8.62155616e-01 -2.09437728e+00 5.31255305e-01 2.59022981e-01
3.25184882e-01 -2.41874501e-01 -5.06418645e-01 8.67160797e-01
-1.06448516e-01 2.89724499e-01 -4.14463222e-01 -2.77045131e-01
-1.20728239e-01 2.59933442e-01 -1.98006630e-01 2.10381284e-01
2.95058250e-01 9.98943210e-01 -8.60384822e-01 -2.89439648e-01
-1.04700521e-01 4.45836067e-01 -5.86724043e-01 -5.14173210e-02
-5.75704157e-01 6.70216858e-01 -8.07341784e-02 3.04530084e-01
5.36933728e-02 -6.93961203e-01 1.13963619e-01 1.95744842e-01
-1.90448746e-01 5.67803979e-01 -6.07036412e-01 1.86546350e+00
-8.67525816e-01 1.20635414e+00 -2.50425845e-01 -6.92126572e-01
1.16136801e+00 2.92933643e-01 3.87858093e-01 -5.76801956e-01
1.25931635e-01 3.91521268e-02 5.37337422e-01 -4.82667863e-01
3.66253495e-01 -2.27307132e-03 -1.59089472e-02 9.70337212e-01
-8.97273794e-02 -3.18624526e-01 9.01343003e-02 4.03771013e-01
1.38743830e+00 3.63007754e-01 7.66637921e-02 -7.28583455e-01
6.10152185e-02 -1.73199531e-02 -3.68795335e-01 6.59228623e-01
3.61306489e-01 3.42449218e-01 6.64087355e-01 -8.46484542e-01
-1.07425964e+00 -1.05671120e+00 -2.27090865e-01 1.36702514e+00
-3.37029286e-02 -8.66680503e-01 -3.15837175e-01 -5.20024776e-01
9.79442522e-02 5.89046657e-01 -8.36447239e-01 1.42592728e-01
-4.14180249e-01 -6.10116482e-01 7.60422826e-01 3.21551651e-01
3.17133904e-01 -1.30387139e+00 -1.06884587e+00 1.40679032e-01
-1.32694587e-01 -1.03391778e+00 1.26902461e-01 3.14420700e-01
-7.04659700e-01 -9.39305604e-01 -3.58849823e-01 -5.08709490e-01
3.75556827e-01 -6.90200627e-02 1.62110996e+00 4.33754534e-01
-2.28382394e-01 2.45775431e-01 -5.83386421e-01 -8.97453353e-02
-4.46715236e-01 6.58705890e-01 -1.55186281e-01 -4.58981842e-01
-1.48848938e-02 -1.14347422e+00 -2.87393272e-01 3.26810330e-01
-8.26868594e-01 3.12898159e-01 6.34906650e-01 6.51981533e-01
-3.60111862e-01 -1.60998568e-01 5.66025853e-01 -8.68668318e-01
1.00881064e+00 -3.82474959e-01 -3.23201776e-01 2.63262659e-01
-3.33317131e-01 3.36003363e-01 8.92792225e-01 -5.10542393e-01
-3.69067699e-01 -4.90730196e-01 2.83821285e-01 5.96136674e-02
-1.79034457e-01 4.81344789e-01 6.15655422e-01 -1.49428874e-01
1.13503230e+00 -1.96464583e-01 1.60057008e-01 -4.11525488e-01
4.16650623e-01 5.21189332e-01 2.25612119e-01 -1.13168395e+00
5.54372549e-01 2.71123558e-01 1.70899436e-01 -1.02174234e+00
-2.68041492e-01 1.18180588e-01 -7.20294595e-01 8.01842734e-02
3.50826323e-01 -7.95887411e-01 -9.38075185e-01 1.45631880e-01
-1.08439410e+00 -9.29104805e-01 -3.97944778e-01 2.71503270e-01
-6.99160159e-01 -6.54681474e-02 -5.29274404e-01 -5.88322639e-01
4.36456762e-02 -9.56417382e-01 1.04448831e+00 -5.55220008e-01
-6.84126914e-01 -1.16693473e+00 3.47728878e-02 -4.46774773e-02
8.34055603e-01 6.30653799e-01 1.19886315e+00 -6.98015213e-01
-7.07721174e-01 4.99552280e-01 -3.45016718e-01 -2.55880296e-01
1.07654177e-01 2.51827389e-01 -8.59916985e-01 -1.56652570e-01
-6.05084121e-01 -7.15020895e-01 1.05300677e+00 -9.26156715e-02
8.27161789e-01 -5.64391971e-01 -2.30304897e-01 3.64433259e-01
9.30422544e-01 -8.55272859e-02 4.08830881e-01 4.18971241e-01
5.25241196e-01 9.50851083e-01 7.09448615e-03 4.61445957e-01
6.99347854e-01 8.16500008e-01 -8.40851571e-03 -2.37034097e-01
-1.44923612e-01 -2.35224105e-02 1.92581788e-01 7.18331873e-01
-1.94914714e-01 -2.76900351e-01 -1.28309083e+00 4.08572048e-01
-1.87758017e+00 -6.96388602e-01 7.22849788e-03 1.81097507e+00
8.42348874e-01 7.44769797e-02 6.71560645e-01 1.38081219e-02
1.79982439e-01 2.08516344e-02 -1.14628062e-01 -5.39328396e-01
-1.90543443e-01 2.35628948e-01 9.65322778e-02 6.30019546e-01
-8.64850700e-01 9.41912830e-01 7.29009771e+00 5.23855329e-01
-1.03009045e+00 -1.50481062e-02 6.77676082e-01 -3.26380193e-01
-4.60990489e-01 3.14925581e-01 1.03352092e-01 2.14269474e-01
8.97180617e-01 3.65848571e-01 1.03934371e+00 1.92803502e-01
-3.64906602e-02 -1.87323570e-01 -1.54951072e+00 6.24733388e-01
-2.27066189e-01 -1.29729033e+00 -6.22368716e-02 3.17273848e-02
6.12393618e-01 2.44238138e-01 -7.27180243e-02 2.36458674e-01
6.72960877e-01 -1.37704992e+00 7.52150774e-01 5.54172754e-01
8.57792974e-01 -4.59532201e-01 2.36131757e-01 3.02662045e-01
-8.95160973e-01 -8.79583135e-02 -2.13200748e-02 -8.33075225e-01
-3.63276482e-01 2.47577220e-01 -9.01162922e-01 3.21211576e-01
6.96836054e-01 7.61444628e-01 -9.25357699e-01 6.63221180e-01
-2.48683095e-02 4.77182508e-01 -6.21887922e-01 -3.82742226e-01
1.98661923e-01 -1.81619659e-01 4.86239672e-01 1.28867805e+00
1.01072341e-01 -4.38450761e-02 2.58965522e-01 9.01498139e-01
-2.02088449e-02 9.08242688e-02 -9.95331705e-01 -1.40392900e-01
5.00503361e-01 1.07653880e+00 -9.62106347e-01 -3.17548603e-01
-1.95582598e-01 4.37277019e-01 5.67058206e-01 4.67482716e-01
-5.40012002e-01 -2.76526004e-01 2.96763897e-01 3.38612199e-01
1.40819997e-01 -5.45236707e-01 -4.50056672e-01 -8.78721714e-01
-1.45731926e-01 -9.49392796e-01 -1.61374267e-02 -8.43562126e-01
-1.38968277e+00 9.86553550e-01 3.76242220e-01 -8.88759732e-01
-2.94184208e-01 -3.52965236e-01 -5.42184830e-01 5.16795993e-01
-8.42143774e-01 -1.26248670e+00 -3.48049939e-01 1.06169485e-01
1.96994781e-01 -2.75794268e-01 1.02968454e+00 -2.89028883e-01
-1.63660467e-01 4.42452461e-01 -4.88896556e-02 1.23948924e-01
5.17021775e-01 -1.29246724e+00 8.37969840e-01 3.39162618e-01
4.93248880e-01 6.59851372e-01 7.12333798e-01 -3.69040459e-01
-9.98742044e-01 -7.29506373e-01 6.49547935e-01 -7.30822206e-01
8.57334256e-01 -7.29447424e-01 -5.60577929e-01 6.82829559e-01
5.51106155e-01 -2.93202996e-01 5.93165696e-01 7.64532328e-01
-5.20825982e-01 2.44122632e-02 -7.88631439e-01 6.76624775e-01
1.58923280e+00 -6.93211973e-01 -3.17757696e-01 2.88972765e-01
9.35090959e-01 -8.03638473e-02 -1.10815954e+00 4.80405569e-01
8.00648153e-01 -1.16611564e+00 7.41167367e-01 -4.88738716e-01
6.62788033e-01 -2.27832511e-01 -1.10720456e-01 -1.51681376e+00
-5.90171635e-01 -9.63883877e-01 1.02481671e-01 1.03606141e+00
6.23999000e-01 -7.72725761e-01 3.79033357e-01 1.25565723e-01
9.88701880e-02 -9.78708565e-01 -9.20624375e-01 -7.58749783e-01
1.48711532e-01 -2.19925210e-01 7.01345146e-01 1.00009751e+00
1.79109395e-01 9.04523075e-01 -1.54006377e-01 -3.23266953e-01
2.43745774e-01 4.58916008e-01 1.07092333e+00 -1.49564898e+00
-5.66185415e-01 -6.80342555e-01 -3.24535817e-01 -9.26699519e-01
1.61973193e-01 -1.07373536e+00 -2.29960874e-01 -1.46998847e+00
-1.54392168e-01 -9.93436396e-01 5.58937415e-02 6.14025533e-01
3.81895810e-01 3.06321174e-01 2.35251054e-01 3.71383429e-01
-5.73095441e-01 5.54611683e-01 1.23958158e+00 5.23582054e-03
-2.29959577e-01 -2.74505585e-01 -6.92465305e-01 2.81839371e-01
8.91432941e-01 -2.66873002e-01 -5.69368899e-01 -6.10130012e-01
6.02386117e-01 -1.71539158e-01 5.86008847e-01 -1.14492846e+00
1.34054109e-01 1.53415740e-01 3.86168599e-01 -1.27955765e-01
4.26052690e-01 -7.82882452e-01 2.96866149e-01 1.78940609e-01
-7.53904879e-01 4.87253398e-01 2.95787752e-01 2.29465544e-01
1.94575980e-01 3.91003817e-01 3.49088401e-01 -2.43586496e-01
-5.23232892e-02 -1.71723366e-01 -5.72985947e-01 2.07427174e-01
8.38063121e-01 -1.27609581e-01 -6.85725868e-01 -5.33353508e-01
-8.14450800e-01 1.69633105e-01 8.50137353e-01 5.43195188e-01
3.91827762e-01 -9.94899571e-01 -8.77083957e-01 3.66234511e-01
1.23565346e-01 -8.67994279e-02 -5.14334202e-01 9.34552193e-01
-5.46232641e-01 2.46316358e-01 -3.64362627e-01 -6.80271924e-01
-1.16945410e+00 5.38380504e-01 2.23089263e-01 -1.88928489e-02
-6.23427868e-01 7.11463034e-01 -3.37128751e-02 -5.74021399e-01
2.16629297e-01 -8.01806271e-01 1.26530185e-01 -6.04879260e-02
-5.91151379e-02 2.97280818e-01 -4.40293476e-02 -4.15924609e-01
-3.08529586e-01 4.97181892e-01 3.48113030e-01 -3.66192311e-01
1.45231700e+00 2.03728989e-01 -4.82429951e-01 8.23549211e-01
9.26602304e-01 9.59543288e-02 -1.02328992e+00 -2.70800442e-01
1.74765691e-01 -3.04279894e-01 -4.04253811e-01 -1.07006192e+00
-5.59346914e-01 7.06164479e-01 2.42965013e-01 7.95580328e-01
1.05755770e+00 3.24488193e-01 2.83074915e-01 7.41339147e-01
4.10134524e-01 -5.33783436e-01 4.62107301e-01 7.98408210e-01
1.16423190e+00 -8.38387668e-01 8.01281780e-02 -1.55744895e-01
-4.97978330e-01 1.04119325e+00 4.87528592e-01 -3.02254021e-01
6.33849919e-01 4.83665586e-01 9.49898064e-02 -4.22018528e-01
-1.36653209e+00 -1.70781970e-01 2.71503627e-01 5.59689760e-01
8.39648128e-01 1.72491595e-01 1.34422436e-01 -1.80257589e-01
-8.01567376e-01 -3.01223516e-01 2.62329668e-01 8.58376026e-01
-1.76084623e-01 -1.12958121e+00 -1.19097568e-01 4.58139628e-01
-1.78649440e-01 -2.95047581e-01 -1.19396961e+00 8.92855704e-01
2.73589790e-01 9.97447848e-01 4.25047636e-01 -7.56070912e-01
8.33018720e-02 -1.43361419e-01 8.32564235e-01 -5.58711827e-01
-1.08169520e+00 -1.54895425e-01 5.55375397e-01 -1.89816982e-01
-6.16801679e-01 -3.98579895e-01 -9.87708330e-01 -5.32279611e-01
-1.55723155e-01 3.18733752e-01 2.34661058e-01 7.95171142e-01
5.87644517e-01 5.37085474e-01 3.00362080e-01 -7.82767415e-01
-7.89629221e-02 -1.44927728e+00 -7.50097632e-02 5.13790131e-01
4.82735395e-01 -6.67090356e-01 -4.87197600e-02 -1.02217615e-01] | [9.49374008178711, 7.023814678192139] |
2bbf259e-8bf4-4e51-acd0-d297a3bbec15 | an-entity-driven-recursive-neural-network | 1704.04336 | null | http://arxiv.org/abs/1704.04336v1 | http://arxiv.org/pdf/1704.04336v1.pdf | An entity-driven recursive neural network model for chinese discourse coherence modeling | Chinese discourse coherence modeling remains a challenge taskin Natural
Language Processing field.Existing approaches mostlyfocus on the need for
feature engineering, whichadoptthe sophisticated features to capture the logic
or syntactic or semantic relationships acrosssentences within a text.In this
paper, we present an entity-drivenrecursive deep modelfor the Chinese discourse
coherence evaluation based on current English discourse coherenceneural network
model. Specifically, to overcome the shortage of identifying the entity(nouns)
overlap across sentences in the currentmodel, Our combined modelsuccessfully
investigatesthe entities information into the recursive neural network
freamework.Evaluation results on both sentence ordering and machine translation
coherence rating task show the effectiveness of the proposed model, which
significantly outperforms the existing strong baseline. | ['Shujing Du', 'Mingwen Wang', 'Maoxi Li', 'Fan Xu'] | 2017-04-14 | null | null | null | null | ['sentence-ordering', 'coherence-evaluation'] | ['natural-language-processing', 'natural-language-processing'] | [-3.48962754e-01 3.17360431e-01 -2.41007939e-01 -3.69677961e-01
-5.40364087e-01 -2.01663494e-01 9.24269199e-01 -4.42238599e-02
-4.60938662e-01 9.77426469e-01 1.17818213e+00 -3.69856320e-02
-4.17879261e-02 -7.54673958e-01 -3.40191334e-01 -5.26397824e-01
2.61223763e-01 2.79482782e-01 -9.17510390e-02 -5.65942883e-01
3.31219167e-01 -2.84856021e-01 -8.89277458e-01 6.02071881e-01
1.37121320e+00 4.70139980e-01 2.94855088e-01 4.93152559e-01
-2.76697516e-01 1.46446526e+00 -7.53246963e-01 -4.72520232e-01
-4.79948312e-01 -7.16406047e-01 -1.13940370e+00 -2.49790952e-01
6.57059029e-02 -3.20426464e-01 -4.95001584e-01 1.09759116e+00
3.09369445e-01 4.06873971e-02 4.66510177e-01 -7.24260092e-01
-8.45016181e-01 1.38306904e+00 -1.00037210e-01 4.66265708e-01
3.58139038e-01 -2.27248684e-01 1.34395015e+00 -1.20663571e+00
8.27536047e-01 1.22715390e+00 5.64689875e-01 2.94020474e-01
-8.60978603e-01 -5.04752517e-01 1.99699402e-01 6.45089924e-01
-1.02185702e+00 -5.00870466e-01 9.12290633e-01 -3.19692492e-01
1.55184174e+00 3.64792138e-01 5.20987630e-01 9.63315427e-01
4.36391741e-01 9.17274892e-01 8.97862136e-01 -3.44592690e-01
-1.65497437e-01 -2.80727684e-01 8.46243262e-01 3.38517159e-01
1.27567187e-01 -8.12123194e-02 -4.79607344e-01 2.25551859e-01
2.83143222e-01 -3.71834517e-01 -2.36713618e-01 5.10353327e-01
-1.03295910e+00 7.66595125e-01 6.26559079e-01 8.57943118e-01
-4.50000316e-01 8.77440050e-02 4.66934502e-01 4.02071565e-01
5.04932582e-01 8.53152871e-01 1.64091792e-02 -1.01308949e-01
-6.70663118e-01 2.87361950e-01 9.89874482e-01 9.74202096e-01
6.39868140e-01 1.18256673e-01 -3.33272278e-01 6.33476257e-01
2.67869622e-01 2.73221284e-01 6.73007488e-01 -8.03651094e-01
8.30963671e-01 6.34203911e-01 -9.23834294e-02 -1.64326918e+00
-6.06796503e-01 -3.14115256e-01 -1.08706009e+00 -5.72045982e-01
-2.10549489e-01 -5.61586738e-01 -6.87200278e-02 1.58312261e+00
-8.59289169e-02 -8.95822495e-02 4.88340348e-01 8.84933650e-01
1.54816139e+00 9.17991221e-01 2.05808237e-01 -4.75930393e-01
1.07310605e+00 -1.30245733e+00 -1.23403668e+00 7.84197226e-02
6.52353823e-01 -6.97583318e-01 7.99889922e-01 9.50037967e-03
-1.37937200e+00 -4.99024093e-01 -1.11304927e+00 -3.11492860e-01
3.56741324e-02 1.24866273e-02 7.84881890e-01 1.35627016e-01
-1.04157305e+00 5.75957596e-01 -7.69418716e-01 -1.60562769e-01
2.50442356e-01 2.45886311e-01 -2.40589857e-01 3.33852291e-01
-1.54093325e+00 1.31093287e+00 1.09041131e+00 4.43807065e-01
-3.44378203e-01 -2.22072169e-01 -7.37014949e-01 3.56292516e-01
3.40697348e-01 -7.04949975e-01 1.24526024e+00 -1.06685209e+00
-1.63160849e+00 5.44198513e-01 -7.95408860e-02 -6.15367651e-01
-1.42463043e-01 -5.06267369e-01 -3.16394359e-01 -5.80333248e-02
-1.12823665e-01 4.26321983e-01 8.85944255e-03 -7.25675762e-01
-4.88137007e-01 1.17701389e-01 2.30870917e-01 7.87475049e-01
-3.80293638e-01 2.29152218e-01 -7.88308978e-02 -6.84100330e-01
3.58227715e-02 -6.19987547e-01 1.39855985e-02 -1.10184205e+00
-8.14640820e-01 -9.96532917e-01 3.18650603e-01 -8.16985965e-01
1.69586933e+00 -1.89132893e+00 6.17485881e-01 -4.36471313e-01
3.38096976e-01 2.29627654e-01 -5.73357791e-02 7.03061342e-01
6.77303821e-02 2.42219239e-01 1.17734581e-01 -9.88396853e-02
1.12607881e-01 -1.03262924e-01 -5.16527891e-01 1.68986589e-01
4.26293969e-01 1.35598612e+00 -8.12270582e-01 -5.84553123e-01
-1.37710959e-01 -1.27680555e-01 -5.70743978e-01 6.17231786e-01
-4.12275642e-01 2.05217615e-01 -5.54928243e-01 2.98860162e-01
4.45154876e-01 -4.02563781e-01 6.71502352e-01 -1.46236107e-01
-1.61662698e-01 1.08246827e+00 -5.32637358e-01 1.74200559e+00
-2.67024219e-01 7.24361598e-01 -2.69210577e-01 -8.21651816e-01
6.37781799e-01 6.00093305e-01 1.85773402e-01 -6.87430441e-01
4.47044164e-01 2.40020469e-01 4.70506996e-01 -7.31236875e-01
9.80565786e-01 2.14317050e-02 -4.63557750e-01 4.02988374e-01
-7.21274992e-04 3.79390195e-02 2.04100847e-01 1.11442439e-01
9.68315959e-01 1.57260686e-01 5.67106545e-01 -4.91479486e-01
6.15296125e-01 2.44914964e-01 4.94485378e-01 4.01585549e-01
-4.45311181e-02 4.31835979e-01 5.55997133e-01 -2.06492007e-01
-8.23247194e-01 -7.54937351e-01 -5.89737818e-02 9.15841699e-01
2.35511601e-01 -5.81909418e-01 -7.15702593e-01 -4.58590508e-01
-5.71780741e-01 6.38533235e-01 -3.08925956e-01 1.43708065e-01
-1.17445159e+00 -9.17837143e-01 7.64108300e-01 5.58576941e-01
1.04870415e+00 -1.20652878e+00 -4.06507671e-01 3.80454242e-01
-6.98928654e-01 -1.14028966e+00 -2.32872337e-01 -1.31821603e-01
-5.95181942e-01 -8.58226538e-01 -2.95125525e-02 -1.23219633e+00
5.40140979e-02 6.40576333e-02 1.29984808e+00 3.41277987e-01
4.64657754e-01 -3.35565686e-01 -3.39227229e-01 -6.14867033e-03
-3.11022133e-01 4.33475196e-01 -2.54085124e-01 -6.85566604e-01
7.31737792e-01 -4.12030727e-01 -3.47780824e-01 -3.92735481e-01
-3.45037788e-01 4.64845389e-01 3.04119438e-01 1.13449478e+00
2.12560613e-02 -1.44984974e-02 9.99063909e-01 -8.87588084e-01
1.32988656e+00 -6.79027200e-01 -2.03398123e-01 4.60392356e-01
-4.67459500e-01 -9.79752764e-02 5.86918056e-01 -3.65333110e-02
-1.27538633e+00 -6.41866863e-01 -2.89307803e-01 2.19655022e-01
4.32152934e-02 1.10849452e+00 -2.77742296e-01 8.78569663e-01
5.45314312e-01 -6.91476390e-02 -5.05846500e-01 -1.67871006e-02
7.17611313e-01 6.52100742e-01 5.49226880e-01 -4.69346732e-01
2.31219232e-01 -2.78128367e-02 -5.43127477e-01 -6.23998940e-01
-9.47913945e-01 -1.70498714e-01 -5.94321787e-01 8.78038853e-02
1.16236758e+00 -1.20061088e+00 -7.01007664e-01 4.13435489e-01
-2.00601530e+00 1.42421976e-01 -2.55120266e-02 5.64931512e-01
-1.34871781e-01 2.80977994e-01 -1.00316775e+00 -3.79662991e-01
-7.43356347e-01 -7.15036929e-01 4.55127925e-01 4.57587838e-01
-5.76170564e-01 -1.22855532e+00 2.28121534e-01 5.14179915e-02
4.59758461e-01 2.20066279e-01 1.09370482e+00 -9.16034460e-01
-8.63517880e-01 1.79965705e-01 -2.45169804e-01 1.44213691e-01
9.33633074e-02 -1.69181213e-01 -7.13397980e-01 1.57505542e-01
2.74370372e-01 -5.61751306e-01 8.26599836e-01 3.08626622e-01
6.11905813e-01 -7.26229072e-01 -5.00015877e-02 2.47101039e-01
1.26004744e+00 3.27851363e-02 7.16253459e-01 2.79279500e-01
7.13946819e-01 7.47280896e-01 3.94788921e-01 8.96161348e-02
1.00452459e+00 3.80616248e-01 4.31302227e-02 8.96030590e-02
-2.71065682e-01 -1.06361277e-01 2.88238794e-01 1.71615624e+00
-1.15848556e-01 -4.42543119e-01 -1.01555216e+00 3.93204689e-01
-2.16120267e+00 -1.17452073e+00 -3.12427610e-01 1.25006163e+00
1.39987063e+00 1.91660181e-01 -3.35717559e-01 -5.23462951e-01
5.65185964e-01 4.29698437e-01 -7.32279271e-02 -4.76779491e-01
-4.99665231e-01 4.03097980e-02 -4.06675935e-01 7.88271964e-01
-1.12322938e+00 1.39850044e+00 6.01066256e+00 6.09009326e-01
-1.07243240e+00 2.71690369e-01 5.19472957e-01 1.38302356e-01
-5.64010799e-01 3.27643976e-02 -8.26403856e-01 3.17076027e-01
9.26026106e-01 -4.73650306e-01 2.89851576e-01 4.68878180e-01
1.07014365e-01 -2.50104457e-01 -1.15738893e+00 6.81030273e-01
3.00121635e-01 -1.75037622e+00 -1.47874877e-01 -4.35320139e-01
9.04412746e-01 1.25747964e-01 -1.55892402e-01 6.33321106e-01
5.58344901e-01 -1.37938106e+00 2.30150580e-01 7.16536462e-01
3.23995769e-01 -5.86141527e-01 1.04947269e+00 5.43129683e-01
-7.36004949e-01 1.38314232e-01 -4.70472008e-01 -4.94706362e-01
1.18321918e-01 2.59605139e-01 -9.66248155e-01 6.94806635e-01
4.06640619e-01 1.16263592e+00 -4.24542725e-01 4.80006605e-01
-4.97588187e-01 6.60051584e-01 1.34106055e-01 -5.63934863e-01
2.75245935e-01 -1.76669136e-01 4.89518940e-01 1.45608270e+00
4.84455004e-02 2.67936766e-01 3.60923558e-01 1.12314963e+00
-3.52776021e-01 2.76327133e-01 -5.02790093e-01 -3.05138826e-01
7.36267924e-01 1.10625958e+00 -2.98056483e-01 -3.16696018e-01
-2.30836585e-01 5.89733899e-01 7.06487417e-01 2.48986974e-01
-7.22764790e-01 -3.66393566e-01 1.86379552e-01 -4.40858394e-01
5.64820878e-02 -4.00231838e-01 -8.97266626e-01 -1.43493569e+00
-5.07459277e-03 -8.65867198e-01 7.71778375e-02 -5.54837167e-01
-1.25665605e+00 1.02529073e+00 1.97540447e-01 -9.62923408e-01
-6.46951199e-01 -1.71387196e-02 -1.02906251e+00 1.03262067e+00
-1.54174328e+00 -1.36338544e+00 -1.25157759e-01 1.55025080e-01
9.37839985e-01 -4.29783314e-01 9.69328523e-01 -4.01400663e-02
-7.32684195e-01 3.05563420e-01 -1.69477701e-01 2.78887838e-01
4.06451970e-01 -1.14845788e+00 3.71002793e-01 1.00794005e+00
-2.43433639e-01 1.17596400e+00 7.00679541e-01 -8.85922670e-01
-8.96493316e-01 -8.19791138e-01 1.73732746e+00 -3.81009251e-01
9.17584658e-01 -1.44032180e-01 -1.04771769e+00 5.26337147e-01
1.47212040e+00 -8.14657331e-01 7.33627141e-01 6.59050405e-01
-7.99887534e-03 2.20005080e-01 -2.95110941e-01 8.75706375e-01
8.02113593e-01 -5.61253726e-01 -1.36208665e+00 4.69822079e-01
1.05143714e+00 -5.60236812e-01 -8.22769284e-01 5.41025519e-01
2.43213788e-01 -7.45763779e-01 4.27156359e-01 -9.86892879e-01
1.26234376e+00 -1.34926423e-01 -1.64595470e-01 -1.25756085e+00
-4.44017857e-01 -5.10596335e-01 -2.87032068e-01 1.22048771e+00
4.27424937e-01 -1.12352826e-01 4.88013208e-01 6.07771277e-01
-5.48674703e-01 -8.26475024e-01 -9.32990909e-01 -3.41688275e-01
6.94451988e-01 1.37219757e-01 5.29931009e-01 1.09427738e+00
7.84161508e-01 1.26619971e+00 -2.11640492e-01 -2.20554352e-01
-2.94419050e-01 2.56907374e-01 5.02801180e-01 -8.57826352e-01
-1.96390331e-01 -6.92395449e-01 -3.03419176e-02 -1.18615603e+00
7.30165601e-01 -1.15625775e+00 1.90957427e-01 -1.76001668e+00
5.23167193e-01 -1.72755957e-01 -5.56349568e-02 2.30648339e-01
-5.56261182e-01 -3.81570220e-01 1.50575995e-01 3.06786209e-01
-9.34146345e-01 9.47122991e-01 1.12791872e+00 -1.10318810e-01
-1.97181016e-01 -4.98690426e-01 -7.70070314e-01 7.02460706e-01
9.16624427e-01 -2.02695310e-01 -4.43975627e-01 -8.72159243e-01
5.00091016e-01 2.32766211e-01 2.08398119e-01 -5.81720412e-01
6.46069586e-01 -3.23357433e-01 1.97754502e-02 -9.45737600e-01
1.08317234e-01 -5.09293675e-01 -1.44406687e-02 3.74650329e-01
-6.26870632e-01 5.76498449e-01 -1.83171764e-01 3.03997666e-01
-5.61635971e-01 -3.42784435e-01 3.21136564e-01 -2.08929718e-01
-5.14615834e-01 -3.84240657e-01 -4.02887106e-01 3.21997643e-01
6.70497894e-01 2.49506682e-01 -1.01911724e+00 -3.84429842e-01
-6.00943625e-01 3.04165781e-01 1.14719898e-01 5.71432114e-01
7.09026694e-01 -1.43379176e+00 -1.06182694e+00 -3.13262701e-01
-3.02026153e-01 8.77951756e-02 1.82653576e-01 9.15732801e-01
-5.34852624e-01 8.01525474e-01 -9.39924717e-02 -4.18623477e-01
-1.07609951e+00 1.81885943e-01 4.51973468e-01 -6.01164877e-01
-5.26062191e-01 8.22544932e-01 2.44292557e-01 -3.44654620e-01
3.21568251e-02 -4.47927803e-01 -9.79316711e-01 6.63793981e-02
3.89388621e-01 1.56784087e-01 -1.08834999e-02 -6.65103614e-01
-1.77268371e-01 9.29871202e-03 -4.55862939e-01 8.58872384e-02
1.37054181e+00 -2.95225054e-01 -6.07559979e-01 4.20802832e-01
1.00502753e+00 -3.92441303e-02 -7.10064113e-01 -2.77712107e-01
5.97486556e-01 2.27359701e-02 -3.00899714e-01 -5.61573267e-01
-4.19233531e-01 7.95736074e-01 -1.80133861e-02 2.30412781e-01
8.06455970e-01 -1.43727779e-01 8.42091978e-01 8.12880993e-01
6.13002405e-02 -1.33028769e+00 3.58628690e-01 1.36002159e+00
1.27837265e+00 -1.29859114e+00 9.66825411e-02 -3.59402150e-01
-8.78453255e-01 1.35850585e+00 8.94464433e-01 -6.24556541e-01
6.73783779e-01 4.18462940e-02 -1.22347564e-01 -3.05738360e-01
-1.08475733e+00 8.81529320e-03 3.37977350e-01 1.09074950e-01
1.01401246e+00 1.44265994e-01 -1.09572029e+00 1.10588825e+00
-5.77444136e-01 -2.86537945e-01 5.87241530e-01 5.27778447e-01
-5.97784877e-01 -8.54399025e-01 2.74053305e-01 1.73613578e-01
-3.37267786e-01 -6.65324926e-01 -5.27123809e-01 6.43141091e-01
7.41423294e-02 1.20053029e+00 1.30273432e-01 -5.43719471e-01
2.31354520e-01 -2.67223656e-01 1.46698043e-01 -8.56202900e-01
-1.03104126e+00 2.72758335e-01 7.27572501e-01 -9.74197909e-02
-1.00822830e+00 -6.93903983e-01 -1.58421671e+00 -3.24011743e-01
-6.58455968e-01 1.96391597e-01 1.48654059e-01 1.30012202e+00
2.61988401e-01 7.72442400e-01 5.57126701e-01 -2.27014184e-01
-5.85074067e-01 -1.49592924e+00 8.53603985e-03 2.19903737e-01
1.33164585e-01 -1.67469203e-01 2.04769209e-01 -7.18089938e-02] | [11.373621940612793, 9.304925918579102] |
add9b772-08b2-4d15-8fa8-3b8a04459d2d | ccg-supertagging-as-top-down-tree-generation | null | null | https://aclanthology.org/2021.scil-1.34 | https://aclanthology.org/2021.scil-1.34.pdf | CCG Supertagging as Top-down Tree Generation | null | ['Vivek Srikumar', 'Nathan Schneider', 'Jakob Prange'] | null | null | null | null | scil-2021-2 | ['ccg-supertagging'] | ['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] | [-7.467878818511963, 3.631009578704834] |
eef77d1a-4516-4172-b0aa-8116c903eddf | visual-prompting-modifying-pixel-space-to | 2203.17274 | null | https://arxiv.org/abs/2203.17274v2 | https://arxiv.org/pdf/2203.17274v2.pdf | Exploring Visual Prompts for Adapting Large-Scale Models | We investigate the efficacy of visual prompting to adapt large-scale models in vision. Following the recent approach from prompt tuning and adversarial reprogramming, we learn a single image perturbation such that a frozen model prompted with this perturbation performs a new task. Through comprehensive experiments, we demonstrate that visual prompting is particularly effective for CLIP and robust to distribution shift, achieving performance competitive with standard linear probes. We further analyze properties of the downstream dataset, prompt design, and output transformation in regard to adaptation performance. The surprising effectiveness of visual prompting provides a new perspective on adapting pre-trained models in vision. Code is available at http://hjbahng.github.io/visual_prompting . | ['Phillip Isola', 'Swami Sankaranarayanan', 'Ali Jahanian', 'Hyojin Bahng'] | 2022-03-31 | null | null | null | null | ['visual-prompting'] | ['computer-vision'] | [ 2.43313998e-01 2.21893042e-01 2.87243407e-02 -2.83865303e-01
-9.38994646e-01 -1.03682256e+00 6.56246185e-01 -3.92790169e-01
-4.01005656e-01 5.68260968e-01 2.80447960e-01 -3.03110629e-01
2.72350818e-01 -4.80497703e-02 -1.18385303e+00 -7.22166896e-01
3.82797509e-01 2.97423489e-02 1.80616945e-01 2.98517663e-02
5.51806763e-02 4.98171210e-01 -1.10814226e+00 8.32021907e-02
6.51362538e-01 5.82425833e-01 1.99897915e-01 1.22491598e+00
2.50147104e-01 6.14854395e-01 -8.52080286e-01 -3.77805501e-01
6.24691725e-01 -6.29694462e-01 -6.21703923e-01 -1.75307050e-01
1.05822623e+00 -4.41967607e-01 -6.27326190e-01 9.19466734e-01
1.11870050e+00 1.04285128e-01 4.77197409e-01 -1.17412865e+00
-1.15393007e+00 4.75128949e-01 -2.85777539e-01 5.25228620e-01
2.27303207e-01 1.13675070e+00 4.69742656e-01 -6.35822892e-01
6.77668631e-01 1.09367442e+00 6.62062049e-01 1.11358809e+00
-1.58423269e+00 -6.33127332e-01 3.64182293e-01 6.84986562e-02
-1.09626436e+00 -9.05239999e-01 4.12512302e-01 -7.17157006e-01
7.94907749e-01 3.19834203e-01 5.61292231e-01 1.76732898e+00
1.78385943e-01 3.50738734e-01 1.14375532e+00 -4.10474688e-01
3.05525720e-01 2.51114275e-02 3.80030945e-02 4.03982848e-01
-1.86378226e-01 6.40435815e-01 -5.32409668e-01 -2.69816935e-01
7.02307880e-01 -3.89442027e-01 -5.43038905e-01 2.70001367e-02
-1.06270087e+00 3.91974658e-01 5.60011566e-01 -2.44819090e-01
-1.81850448e-01 6.99857533e-01 2.27997169e-01 3.37587088e-01
1.11625813e-01 8.58555794e-01 -5.84082007e-01 -7.99895301e-02
-4.62033927e-01 3.87709796e-01 5.25597930e-01 1.20485342e+00
4.26398665e-01 2.27686226e-01 -9.35312331e-01 5.16162157e-01
-6.77879229e-02 7.54853964e-01 4.12966847e-01 -1.46670043e+00
-5.42496666e-02 -1.66707765e-02 9.20753479e-02 -3.79453033e-01
-4.74099874e-01 -4.44208920e-01 -5.82019567e-01 4.03624833e-01
6.07866585e-01 -5.04275143e-01 -1.37714112e+00 2.16402721e+00
3.66783917e-01 3.87727380e-01 -1.35992467e-01 7.25973725e-01
1.05771983e+00 3.13404649e-01 2.34239742e-01 2.69726664e-02
1.06744385e+00 -9.02641118e-01 -4.16330099e-01 -2.83987224e-01
6.88791126e-02 -8.40263009e-01 1.64250588e+00 7.84657970e-02
-1.08622420e+00 -5.85055649e-01 -5.46324492e-01 -1.97923809e-01
-1.22888274e-02 -5.70279993e-02 3.46856058e-01 3.13919634e-01
-1.48434520e+00 2.07325280e-01 -8.20898533e-01 -7.67285466e-01
8.03472579e-01 3.66460443e-01 -2.26640821e-01 3.22540179e-02
-7.31499612e-01 8.43794227e-01 -8.69604424e-02 -1.29867613e-01
-1.31574821e+00 -1.21822762e+00 -4.61230427e-01 -9.38193724e-02
2.28683576e-01 -1.31821263e+00 1.75885797e+00 -1.00867271e+00
-1.72064316e+00 8.51767659e-01 -1.43136352e-01 -5.52616775e-01
7.37351835e-01 -5.85186593e-02 -1.09927431e-01 1.49492428e-01
-2.75458187e-01 1.15806961e+00 1.29526937e+00 -1.26116359e+00
-1.35057837e-01 -1.03242733e-01 6.15634061e-02 6.79583326e-02
-3.27852935e-01 1.44715965e-01 -6.63905919e-01 -5.14651716e-01
-6.43030465e-01 -1.09948957e+00 -5.13354182e-01 1.52463377e-01
-8.95103514e-01 2.48006627e-01 3.97100002e-01 -4.72202897e-01
1.04734898e+00 -2.14156699e+00 4.28717844e-02 -5.14532179e-02
3.53782296e-01 4.87037927e-01 -6.71230555e-01 3.53165686e-01
-4.00218405e-02 1.72605410e-01 -2.23741308e-01 -2.72958368e-01
9.21722408e-03 2.73956582e-02 -5.76055825e-01 3.79786670e-01
1.99967563e-01 1.55008650e+00 -5.98319530e-01 -1.98331222e-01
4.11306694e-02 5.93095243e-01 -7.59340703e-01 4.38307732e-01
-5.60502052e-01 1.02817714e+00 -2.43284002e-01 7.95161366e-01
5.56253850e-01 -3.88476700e-01 -1.11319683e-01 -1.47712916e-01
1.26043767e-01 -2.46122271e-01 -3.71320099e-01 1.49951243e+00
-1.05914041e-01 8.20909083e-01 -5.60908951e-02 -4.08028960e-01
5.00866175e-01 -5.31290323e-02 4.68975939e-02 -6.10339344e-01
6.54647425e-02 -3.55733484e-01 4.72349627e-03 -7.64272749e-01
1.37998849e-01 1.09610617e-01 3.39202695e-02 -2.57839002e-02
2.03444511e-01 -2.71634728e-01 -2.40548864e-01 3.33987832e-01
1.51463544e+00 1.81904078e-01 2.55813122e-01 -1.39157027e-01
-1.05869628e-01 5.50974943e-02 4.11643207e-01 1.13770235e+00
-5.56975842e-01 8.42780292e-01 4.43629235e-01 -2.07053438e-01
-1.12600946e+00 -1.33410394e+00 -1.08587801e-01 1.09470248e+00
-2.03284055e-01 -2.88784593e-01 -9.85910356e-01 -5.40460527e-01
3.12163740e-01 6.06057644e-01 -9.83914196e-01 -5.82256794e-01
-4.07211572e-01 -4.53277230e-01 8.56096148e-01 5.54056406e-01
1.15242794e-01 -1.07611382e+00 -5.08921325e-01 -2.45729327e-01
7.80536756e-02 -1.01008713e+00 -9.17505741e-01 1.56329721e-01
-5.32872796e-01 -8.91391516e-01 -7.64479339e-01 -6.38793409e-01
8.92965674e-01 2.17936784e-02 1.25426507e+00 6.44030236e-03
-4.13340449e-01 9.12558854e-01 -1.01580188e-01 -6.24037206e-01
-6.08079672e-01 1.76560611e-01 -1.18177598e-02 -3.82717967e-01
2.62667555e-02 -7.01327622e-01 -8.38052630e-01 1.82695895e-01
-9.73647118e-01 -2.13890374e-01 3.51583630e-01 8.05336356e-01
7.24136591e-01 -7.68175662e-01 5.97222269e-01 -9.54846621e-01
7.08123863e-01 -2.89655656e-01 -8.63182724e-01 1.58941671e-01
-5.30873358e-01 -1.24788471e-01 9.06972945e-01 -8.71688724e-01
-9.61501122e-01 2.30167106e-01 -1.15588047e-01 -8.37203562e-01
-3.47744167e-01 1.44632719e-02 -2.25664869e-01 -3.61079603e-01
1.39075553e+00 7.12672919e-02 -6.94567710e-02 -3.35602373e-01
8.74555588e-01 3.29216808e-01 1.08018434e+00 -5.34985662e-01
1.10105526e+00 3.60059977e-01 -1.07218809e-01 -5.11951566e-01
-7.37637460e-01 2.69763209e-02 -3.91924679e-01 -2.09100738e-01
7.56067574e-01 -9.59297359e-01 -9.45672810e-01 5.62413812e-01
-9.67870653e-01 -1.10845339e+00 -7.13717461e-01 8.29934627e-02
-7.53759027e-01 1.04552291e-01 -4.19630438e-01 -2.64583498e-01
-3.12459260e-01 -9.07691061e-01 6.74412131e-01 6.91514552e-01
-2.48558432e-01 -9.22525525e-01 3.53697032e-01 2.47219250e-01
6.40672326e-01 2.81764567e-01 7.04768240e-01 -6.00236118e-01
-7.78046489e-01 7.93216005e-02 4.98781279e-02 4.07787740e-01
-1.12813078e-01 2.64433354e-01 -1.33998477e+00 -5.52356660e-01
-3.46038997e-01 -5.36894083e-01 9.75704670e-01 8.15094411e-01
1.13269591e+00 -4.48072344e-01 -3.52196574e-01 1.34944701e+00
1.24762690e+00 -1.84201542e-03 7.11529195e-01 3.75061929e-01
7.19074249e-01 1.35516286e-01 2.28044525e-01 5.42042255e-01
1.94785863e-01 5.99884510e-01 3.26833040e-01 -2.18146175e-01
-5.52980542e-01 -4.71170038e-01 4.82318699e-01 1.37521580e-01
3.05538952e-01 -3.30286026e-01 -8.10569823e-01 3.65650266e-01
-1.63976443e+00 -8.97904575e-01 3.45995873e-01 2.15507150e+00
1.08437037e+00 -1.41157538e-01 2.34318703e-01 -9.11344111e-01
6.36685193e-01 -1.20210901e-01 -1.31354308e+00 -3.11397374e-01
-3.40657741e-01 2.11806878e-01 7.04044640e-01 6.61754608e-01
-7.57247686e-01 1.26656795e+00 7.65095139e+00 5.36886513e-01
-1.19255030e+00 9.07018036e-02 6.29468501e-01 -6.92072809e-01
-5.81098437e-01 6.11217842e-02 -8.81140769e-01 6.17144346e-01
1.03900015e+00 -4.01831686e-01 8.14071417e-01 5.04571021e-01
4.92270261e-01 1.36917353e-01 -1.14889383e+00 7.64993191e-01
-5.89127168e-02 -1.42026401e+00 9.93407965e-02 -5.28421029e-02
8.49308848e-01 4.23061967e-01 5.80323815e-01 1.32494673e-01
6.85982764e-01 -1.17113507e+00 5.90652168e-01 8.36241543e-01
1.21103489e+00 -2.87711829e-01 1.57573558e-02 -1.16366800e-02
-4.56681132e-01 -1.54006884e-01 -4.81371373e-01 2.31977746e-01
8.84633660e-02 3.82517695e-01 -9.60246146e-01 -4.06539813e-02
8.30213010e-01 3.68741751e-01 -9.66053843e-01 1.16598964e+00
-3.98063719e-01 9.91272628e-01 -2.47526601e-01 4.21460479e-01
-3.30056280e-01 2.02938542e-01 8.02368999e-01 1.26363766e+00
1.97359845e-01 -1.68257337e-02 -7.07349405e-02 9.98980105e-01
-3.43622178e-01 -2.10210815e-01 -7.79937088e-01 2.84282621e-02
7.13649035e-01 1.15381527e+00 -2.19563320e-01 -2.19078898e-01
-1.26939401e-01 1.05445981e+00 5.54912388e-01 9.04686451e-01
-9.81347024e-01 3.65208425e-02 9.09422576e-01 -5.23455860e-03
3.38815302e-01 -1.18026696e-02 -2.57723272e-01 -9.81521070e-01
6.87967688e-02 -1.03928828e+00 2.34071672e-01 -1.09642982e+00
-1.33533406e+00 6.25164211e-01 -4.74622697e-02 -9.85985816e-01
-1.85670286e-01 -2.83679098e-01 -7.89268076e-01 8.64130914e-01
-1.27399814e+00 -1.26908970e+00 -5.49628556e-01 8.13042819e-01
2.47669056e-01 -2.43080348e-01 7.38828123e-01 -1.65569991e-01
-6.89231873e-01 1.21411300e+00 1.43674016e-01 -2.45798677e-01
1.25629866e+00 -1.31526923e+00 8.77915680e-01 1.12101412e+00
-8.28681141e-03 6.59803271e-01 1.03059304e+00 -5.08922040e-01
-1.36260927e+00 -1.18731797e+00 1.57310694e-01 -1.00615609e+00
4.09218580e-01 -3.23260874e-01 -9.00564790e-01 1.08475363e+00
5.54330051e-01 3.43977869e-01 5.53784072e-01 -2.01013058e-01
-5.68322957e-01 -1.57709181e-01 -1.11072695e+00 9.35369551e-01
1.32476318e+00 -5.67084253e-01 -8.98081064e-02 4.39494342e-01
1.11216187e+00 -7.75293469e-01 -7.48353601e-01 1.98667124e-01
3.58758241e-01 -8.33526134e-01 1.03567803e+00 -1.01828718e+00
2.04683095e-01 -2.74984360e-01 -5.05308136e-02 -1.63176489e+00
-6.34856820e-01 -1.40315676e+00 -1.65789634e-01 1.28875566e+00
3.82966310e-01 -8.66215050e-01 5.93339801e-01 7.29375362e-01
-2.04815596e-01 -5.45252919e-01 -7.58013427e-01 -7.99212754e-01
3.83741915e-01 -4.16821390e-02 4.53805119e-01 6.58209741e-01
-3.93009484e-01 3.29677701e-01 -3.75623196e-01 3.26789051e-01
3.99018347e-01 -2.11977884e-01 1.19772696e+00 -5.91247201e-01
-7.04969108e-01 -3.28449994e-01 -1.73409477e-01 -9.45779204e-01
2.42190734e-01 -8.63522708e-01 7.76406378e-02 -1.33675218e+00
3.36182207e-01 -1.15323411e-02 -2.93469191e-01 6.88050389e-01
-5.28847277e-01 2.54080266e-01 4.59282815e-01 2.96257883e-01
-4.15173650e-01 4.47409153e-01 1.43589687e+00 -8.01981837e-02
-2.95109630e-01 2.53888797e-02 -1.22848642e+00 3.40057522e-01
9.98578310e-01 -4.03725207e-01 -4.15357649e-01 -5.93644559e-01
3.58057320e-02 -3.07953656e-01 7.43033767e-01 -8.07313979e-01
3.38362545e-01 -2.70655215e-01 5.73725760e-01 1.32993132e-01
1.19251385e-01 -4.76268589e-01 1.26042590e-01 1.88875943e-01
-6.30522966e-01 1.07637264e-01 7.87231207e-01 5.49629569e-01
2.81002253e-01 1.62949219e-01 1.08407080e+00 -7.18937293e-02
-6.68866456e-01 4.82632339e-01 -3.32897633e-01 5.68454742e-01
9.09482837e-01 1.14122957e-01 -9.08297420e-01 -4.32948828e-01
-8.93607736e-01 2.09453642e-01 9.19041157e-01 2.57152587e-01
4.23983276e-01 -1.02795660e+00 -6.95286036e-01 1.46947473e-01
2.67814606e-01 -9.10721049e-02 4.96897936e-01 7.64283478e-01
-1.63295642e-01 2.90625896e-02 -3.37028056e-01 -6.92178428e-01
-1.22006023e+00 7.66092241e-01 7.28112876e-01 2.88284004e-01
-6.13190889e-01 1.07716107e+00 4.76749718e-01 -4.15969014e-01
2.52576321e-01 -3.64676476e-01 3.13926041e-01 -4.78500098e-01
4.66336370e-01 7.54982010e-02 -2.32399404e-01 -4.85867448e-02
-2.56279200e-01 5.87347448e-01 -4.92191985e-02 -1.29130930e-01
9.09798145e-01 -3.13213676e-01 1.50374159e-01 1.23214804e-01
8.95083904e-01 6.92313612e-02 -1.95561445e+00 -1.96281090e-01
-5.36794841e-01 -4.58766222e-01 -2.30915308e-01 -1.27661407e+00
-8.35422516e-01 5.37498653e-01 8.42453539e-01 -2.18418762e-01
1.36102593e+00 2.92559624e-01 3.59547019e-01 4.46449548e-01
-8.59258622e-02 -7.84108520e-01 1.50408849e-01 2.99752951e-01
1.12571490e+00 -1.19148266e+00 -2.53174186e-01 1.51920915e-02
-7.25432694e-01 5.47816217e-01 1.00917125e+00 -1.15366641e-03
5.51618457e-01 6.94695830e-01 7.08303511e-01 7.01936111e-02
-1.00934005e+00 -1.59452558e-01 1.48941353e-01 1.28673542e+00
-2.74976790e-02 -5.95270731e-02 4.34908152e-01 3.53365183e-01
-3.52963895e-01 2.34239742e-01 7.02255726e-01 5.84322989e-01
-2.10154861e-01 -9.11412299e-01 -3.26019853e-01 4.22243953e-01
-3.18675339e-01 -1.89843819e-01 -6.08358264e-01 5.56819737e-01
-1.22985289e-01 7.80796885e-01 -8.07641670e-02 -4.41665113e-01
5.13496578e-01 -6.16241153e-03 7.43111014e-01 -5.76378047e-01
-7.48070896e-01 -8.66425708e-02 -2.34053329e-01 -8.69520724e-01
8.50135833e-02 -5.89025676e-01 -9.85087216e-01 -2.18740374e-01
8.97002444e-02 -4.24541593e-01 3.41467053e-01 5.62121987e-01
8.92186999e-01 6.50337100e-01 5.11889517e-01 -8.34025741e-01
-9.25401449e-01 -1.01327181e+00 -1.26139984e-01 3.78088892e-01
7.51112342e-01 -2.48903707e-01 -4.81614411e-01 5.94754636e-01] | [10.287446975708008, 2.0543127059936523] |
673d9487-0dd6-48de-9149-e499f96dd660 | calibrated-interpretation-confidence | 2211.07443 | null | https://arxiv.org/abs/2211.07443v6 | https://arxiv.org/pdf/2211.07443v6.pdf | Calibrated Interpretation: Confidence Estimation in Semantic Parsing | Sequence generation models are increasingly being used to translate natural language into programs, i.e. to perform executable semantic parsing. The fact that semantic parsing aims to predict programs that can lead to executed actions in the real world motivates developing safe systems. This in turn makes measuring calibration -- a central component to safety -- particularly important. We investigate the calibration of popular generation models across four popular semantic parsing datasets, finding that it varies across models and datasets. We then analyze factors associated with calibration error and release new confidence-based challenge splits of two parsing datasets. To facilitate the inclusion of calibration in semantic parsing evaluations, we release a library for computing calibration metrics. | ['Benjamin Van Durme', 'Elias Stengel-Eskin'] | 2022-11-14 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 3.23436052e-01 3.35125327e-01 -6.06222987e-01 -6.12432241e-01
-9.84284937e-01 -9.52023685e-01 6.30052686e-01 2.61157662e-01
-1.24928705e-01 4.96970445e-01 5.48990071e-01 -7.44339645e-01
3.26755643e-01 -1.04296482e+00 -1.09707654e+00 1.74265608e-01
2.40499347e-01 1.61902919e-01 4.27787781e-01 -1.34370923e-01
4.29405540e-01 -5.12502305e-02 -1.55124331e+00 3.53085816e-01
9.59459424e-01 4.35266584e-01 -2.57391006e-01 8.56925309e-01
-2.73686320e-01 9.82489944e-01 -6.55728996e-01 -5.41870117e-01
1.00449100e-01 -5.06148994e-01 -1.14360821e+00 -3.83305162e-01
2.02728614e-01 -1.39958218e-01 2.55961657e-01 1.23029160e+00
5.14861867e-02 -1.57179162e-01 3.18849057e-01 -1.60709870e+00
-4.02702957e-01 9.84636247e-01 1.73125621e-02 -1.90142635e-02
6.07013583e-01 5.19519746e-01 1.18777537e+00 5.78439683e-02
8.79969537e-01 1.31650209e+00 7.52787650e-01 7.08293617e-01
-1.35399282e+00 -4.29610968e-01 7.50099719e-02 -3.12332213e-01
-8.35677445e-01 -6.17674112e-01 3.18049878e-01 -6.37706935e-01
1.36313379e+00 1.83867380e-01 3.01170716e-04 1.63218009e+00
2.98050672e-01 3.94022197e-01 1.16735888e+00 -3.74465495e-01
4.82281297e-01 2.23555155e-02 6.97183013e-01 6.55100167e-01
6.55091822e-01 4.78667974e-01 -4.19575065e-01 -4.53753531e-01
7.90641531e-02 -6.47273481e-01 1.37828723e-01 -5.07263727e-02
-9.49971259e-01 9.50171828e-01 -1.83109507e-01 -5.11009304e-04
1.54469788e-01 7.15762317e-01 8.84762466e-01 2.55842656e-01
1.33771926e-01 8.51197660e-01 -4.86815602e-01 -7.98427403e-01
-4.55383033e-01 5.67368567e-01 1.16581678e+00 1.12314665e+00
6.59941494e-01 -6.11131154e-02 -9.37236920e-02 3.75711024e-01
1.91520289e-01 4.10637796e-01 2.41242647e-01 -1.10128820e+00
6.19460464e-01 8.00849438e-01 2.94327494e-02 -5.31353354e-01
-4.15646613e-01 -1.34092690e-02 2.62538373e-01 1.98585853e-01
3.95729780e-01 -2.66973644e-01 -6.51505530e-01 2.12839341e+00
-7.18048587e-03 8.58202130e-02 3.09436083e-01 4.56174910e-01
3.41787755e-01 3.68037969e-01 6.61050797e-01 -1.10016670e-02
1.32318962e+00 -5.81614375e-01 -2.14550674e-01 -6.52179062e-01
1.09274423e+00 -4.75031555e-01 1.48462248e+00 1.86953932e-01
-7.38823593e-01 -3.54656577e-01 -1.20615005e+00 1.09481663e-01
-2.26371244e-01 -4.03831303e-01 9.58047092e-01 1.03585589e+00
-9.55803931e-01 6.89317703e-01 -1.03361404e+00 -4.03012753e-01
1.66305229e-01 4.20848746e-03 -2.32563213e-01 2.90771872e-01
-1.04497147e+00 9.81494725e-01 6.61306143e-01 -7.68370569e-01
-9.94853437e-01 -6.79853141e-01 -1.11882257e+00 -1.72252819e-01
4.40036267e-01 -7.80483782e-01 1.65808082e+00 -9.35910940e-01
-1.19453132e+00 6.61760926e-01 -6.09777384e-02 -5.33817232e-01
4.39018190e-01 -4.31330621e-01 -4.47365016e-01 -2.91963845e-01
3.49166363e-01 5.39453983e-01 6.34818912e-01 -8.65072370e-01
-6.02134705e-01 -3.05494219e-01 4.81868446e-01 -2.84750044e-01
1.27967060e-01 4.74788606e-01 -1.27920419e-01 -1.59556314e-01
-3.72879416e-01 -1.04768574e+00 -1.88110575e-01 -6.61283016e-01
-3.86616290e-01 -1.95807129e-01 4.47057009e-01 -7.42883980e-01
1.17694342e+00 -1.99625862e+00 -3.23650129e-02 -1.97878350e-02
-1.51241779e-01 -1.16915777e-01 -2.65186012e-01 3.54599684e-01
-8.90113339e-02 6.39108598e-01 -2.76551396e-01 7.97146037e-02
3.40877712e-01 1.21648945e-01 -5.73482215e-01 6.34862706e-02
5.68816483e-01 1.11382878e+00 -9.03066635e-01 -4.39465880e-01
3.69595997e-02 5.83291687e-02 -9.37517762e-01 2.64465630e-01
-9.63286459e-01 2.44685695e-01 -6.06990874e-01 6.30257428e-01
1.64308488e-01 -1.72160074e-01 1.52261332e-01 3.95996004e-01
-8.33521709e-02 8.17526937e-01 -5.83359599e-01 1.72296309e+00
-4.98646170e-01 3.38015586e-01 -4.36066628e-01 -6.33742511e-01
7.66055346e-01 1.34056941e-01 -9.30686817e-02 -8.35260391e-01
1.34686381e-02 1.75395325e-01 -5.58961928e-02 -4.75847065e-01
6.72076702e-01 -1.76900983e-01 -7.29758859e-01 5.07952809e-01
-3.09981346e-01 -3.81176621e-01 2.27025107e-01 1.84369355e-01
1.39036584e+00 5.87000489e-01 3.17046732e-01 -5.31555295e-01
2.15065971e-01 5.75903952e-01 7.02686250e-01 6.89447939e-01
-1.98677585e-01 3.24584812e-01 1.31565654e+00 -4.57534671e-01
-1.07347465e+00 -8.20557535e-01 1.36161610e-01 1.06407630e+00
3.70313339e-02 -5.70743382e-01 -1.23986495e+00 -1.02950907e+00
-4.45918180e-02 1.58126605e+00 -5.27769864e-01 -6.72127664e-01
-6.35560930e-01 -5.46394110e-01 9.90720391e-01 7.98911214e-01
1.33358493e-01 -9.50676858e-01 -1.08310223e+00 8.12868178e-02
-3.41103882e-01 -1.20803404e+00 -3.17422748e-01 2.46995762e-01
-8.68802726e-01 -1.43638086e+00 4.61415023e-01 -3.18317086e-01
3.17501873e-01 -1.87493280e-01 1.75752902e+00 3.23179156e-01
-4.62990440e-02 6.22011542e-01 -5.50680637e-01 -4.51933444e-01
-1.33742869e+00 2.61074811e-01 -2.45863557e-01 -1.00123668e+00
5.39868534e-01 -2.38611132e-01 -3.70200649e-02 2.01142147e-01
-7.84060657e-01 1.24746054e-01 1.55153796e-01 4.82115328e-01
1.92412138e-01 -2.05516309e-01 4.55987364e-01 -1.26290154e+00
8.44738960e-01 -4.44755554e-01 -1.17104983e+00 4.42079455e-01
-7.85205543e-01 5.03990531e-01 8.23443532e-01 3.44791822e-02
-8.93202960e-01 -8.21058750e-02 -3.40283841e-01 6.95203543e-02
-2.94880658e-01 3.79815847e-01 -1.95275024e-01 2.64865190e-01
1.07606316e+00 -1.75239310e-01 -3.42115127e-02 -1.28424149e-02
3.91624331e-01 3.19404423e-01 5.99669456e-01 -1.27635467e+00
6.54074669e-01 -9.42383558e-02 -2.14737505e-01 -3.41878921e-01
-7.14209497e-01 8.54013935e-02 -4.07638140e-02 2.74133205e-01
9.98666108e-01 -8.06535065e-01 -6.65870428e-01 1.57825693e-01
-1.15232623e+00 -7.33456314e-01 -2.10809827e-01 6.82822689e-02
-8.41103673e-01 3.01713347e-01 -4.06810224e-01 -6.95005715e-01
-2.45899662e-01 -1.71880531e+00 1.27378964e+00 9.17825326e-02
-9.05507445e-01 -8.36727321e-01 4.06813651e-01 3.38726491e-01
3.31344038e-01 4.72826868e-01 1.23760784e+00 -7.41365612e-01
-4.99306291e-01 -4.43398766e-02 -1.58947334e-02 3.86370569e-01
-2.33142674e-01 3.38099808e-01 -1.02427018e+00 7.53258914e-02
-2.72270804e-03 -4.14241016e-01 4.73517179e-01 1.67737901e-01
9.93369937e-01 -2.88872927e-01 -3.38730991e-01 5.21594346e-01
1.44545937e+00 1.50595590e-01 7.13280618e-01 4.12806541e-01
3.85894537e-01 5.45068085e-01 8.48718345e-01 1.33368984e-01
4.41950887e-01 5.50476670e-01 2.73366421e-01 6.62664413e-01
1.15319528e-01 -6.75540507e-01 6.82522774e-01 2.09223509e-01
3.38459074e-01 -1.26215532e-01 -1.36020184e+00 4.05428171e-01
-1.54190433e+00 -7.10795283e-01 -1.38174236e-01 2.10457373e+00
8.50995719e-01 4.92346227e-01 2.19419137e-01 -1.82461694e-01
4.98430520e-01 3.72910649e-02 -2.82022208e-01 -7.58292496e-01
2.65801430e-01 4.87318754e-01 7.31420934e-01 7.40474343e-01
-9.37535882e-01 1.19944763e+00 7.28591490e+00 2.95435607e-01
-1.13902760e+00 6.89030392e-03 6.49199963e-01 3.65841597e-01
-7.18428195e-01 6.02244914e-01 -9.05927658e-01 5.87241471e-01
1.70017636e+00 -5.94859540e-01 3.68024290e-01 1.25899839e+00
9.24113989e-02 -3.50083768e-01 -1.49626315e+00 2.53746748e-01
-4.42290962e-01 -1.12847519e+00 -1.19709738e-01 -4.06437337e-01
5.45056462e-01 -8.59548431e-03 -1.34001762e-01 4.93424207e-01
9.78326499e-01 -1.23682654e+00 1.00353384e+00 1.46750174e-02
6.63841009e-01 -6.76038802e-01 5.55405319e-01 1.89082071e-01
-8.00065100e-01 -5.12290560e-02 -9.55521762e-02 -1.20937094e-01
7.33755305e-02 3.54321480e-01 -9.98386204e-01 3.21308017e-01
4.69925106e-01 4.13798124e-01 -7.40258574e-01 4.35575515e-01
-5.28398514e-01 8.45416367e-01 7.34996796e-02 8.31893533e-02
2.22443789e-02 -7.07195795e-05 2.41877139e-01 1.18331778e+00
2.22287133e-01 -1.40511647e-01 -1.20077292e-02 1.32261527e+00
1.20823219e-01 -2.90674597e-01 -6.32472932e-01 -7.96169102e-01
6.29288316e-01 8.24093282e-01 -6.86457574e-01 -3.11772138e-01
-5.31033516e-01 5.83044767e-01 2.09877118e-01 3.20155211e-02
-1.31052291e+00 -8.96415412e-02 1.04186356e+00 -3.86507921e-02
-2.54332930e-01 -3.55335802e-01 -7.60103106e-01 -9.18441892e-01
1.52990580e-01 -1.38100982e+00 4.17753875e-01 -5.50290346e-01
-7.27556169e-01 4.85370636e-01 2.85953075e-01 -7.10747063e-01
-5.27386904e-01 -4.65078562e-01 -6.60311103e-01 7.40643024e-01
-1.27210283e+00 -8.70481491e-01 -3.62193555e-01 -1.92260032e-03
3.70337844e-01 2.13697195e-01 8.71750653e-01 -8.51861760e-02
-5.73352158e-01 6.39187574e-01 -7.80014694e-01 -1.02267250e-01
5.85468352e-01 -1.14089918e+00 1.47641945e+00 1.27941871e+00
-3.17664817e-02 9.63220716e-01 9.45163786e-01 -9.49578822e-01
-1.54969084e+00 -1.24966478e+00 6.14851713e-01 -1.01096654e+00
8.01625431e-01 -2.16957092e-01 -7.99162865e-01 9.33894813e-01
-8.03178623e-02 -4.14455801e-01 4.50851738e-01 1.58851892e-01
-8.55571330e-01 3.34626734e-01 -1.01249444e+00 4.35778797e-01
1.49976122e+00 -8.11301589e-01 -6.63080573e-01 2.85870098e-02
1.32721472e+00 -5.42309344e-01 -8.06853771e-01 4.50225174e-01
3.06637436e-01 -1.26421881e+00 6.74561560e-01 -9.10543501e-01
8.15441966e-01 -2.01316714e-01 -2.06518263e-01 -1.20354080e+00
1.36183351e-01 -6.94590867e-01 2.70693183e-01 9.71241832e-01
8.15256715e-01 -9.52229738e-01 7.98139453e-01 1.26269937e+00
-3.33818227e-01 -1.04256384e-01 -4.96347576e-01 -8.62249494e-01
3.32947105e-01 -8.75346720e-01 9.20781255e-01 7.92309463e-01
2.47640461e-01 2.63720840e-01 1.49660081e-01 2.76335239e-01
3.87408346e-01 2.26132981e-02 9.06774402e-01 -8.61807406e-01
-6.88100040e-01 -5.36318123e-01 -3.80906016e-01 -4.23441350e-01
6.00241899e-01 -9.50319827e-01 1.13661475e-01 -1.16572952e+00
2.76071012e-01 -4.39820677e-01 1.93291843e-01 6.37498915e-01
-4.40177292e-01 -4.17151392e-01 -2.22016592e-02 -1.56600565e-01
-6.86948419e-01 2.18753353e-01 4.37873244e-01 1.64442122e-01
-4.73388918e-02 -1.72695100e-01 -1.11910200e+00 5.28438568e-01
9.80648458e-01 -4.97026116e-01 -7.01063573e-01 -5.60057700e-01
5.49659669e-01 1.53528735e-01 4.66924608e-01 -1.23652279e+00
-2.15807438e-01 -5.69647431e-01 -3.38375717e-01 -1.07049257e-01
-3.22106272e-01 -5.32653272e-01 4.98601496e-01 6.02237344e-01
-7.28958488e-01 3.88307810e-01 4.46247607e-01 2.91125923e-01
-1.53522298e-01 -4.33354616e-01 7.03443766e-01 -1.79795444e-01
-9.60371792e-01 -1.57922864e-01 -2.36488171e-02 7.43230641e-01
1.11310410e+00 -1.42822742e-01 -7.27138817e-01 -6.49001449e-02
-1.78945452e-01 1.27656937e-01 1.19950974e+00 8.56264472e-01
6.46090135e-02 -7.68717527e-01 -3.92384529e-01 2.01735511e-01
4.81616169e-01 -1.69856250e-01 -2.19264418e-01 5.38025498e-01
-7.34213412e-01 4.54074919e-01 -1.43874645e-01 -4.13690060e-01
-1.14043236e+00 5.43542087e-01 2.31423482e-01 -1.44901261e-01
-5.39956570e-01 6.85041249e-01 1.15046099e-01 -5.09022057e-01
-1.49016365e-01 -6.24951184e-01 3.51501226e-01 -6.70461774e-01
4.69318390e-01 2.24253669e-01 1.78590640e-01 -2.15507939e-01
-5.05940557e-01 3.94932665e-02 1.62036359e-01 -1.29467085e-01
1.11611760e+00 5.97752035e-01 -3.37489456e-01 1.55057818e-01
1.09817421e+00 -1.97049547e-02 -1.08588076e+00 3.24188471e-01
6.57337904e-01 -4.58157957e-01 -3.71693343e-01 -7.79990375e-01
-5.35382152e-01 4.88211900e-01 1.63953573e-01 1.04174592e-01
7.01799154e-01 -2.09550951e-02 6.22902691e-01 3.04666102e-01
9.82511044e-01 -1.12847102e+00 -1.31891683e-01 5.46989620e-01
4.39841300e-01 -1.08928239e+00 -2.42229402e-01 -5.76850235e-01
-6.97548926e-01 7.58950472e-01 8.87214839e-01 1.05638362e-01
-3.06742229e-02 7.64525235e-01 -9.55838263e-02 -6.78802282e-02
-1.00728643e+00 3.14661950e-01 -2.78420985e-01 6.65815115e-01
6.47190034e-01 3.88324797e-01 -3.89753014e-01 7.51360655e-01
-4.53957826e-01 1.33412406e-01 7.57032573e-01 1.09579039e+00
-3.13073635e-01 -1.55557573e+00 -2.97998786e-01 4.31098551e-01
-4.49117005e-01 -8.13103467e-02 -6.70673192e-01 7.21235037e-01
-2.73731261e-01 1.09604883e+00 -1.74772620e-01 -6.77411675e-01
2.23905757e-01 5.46763837e-01 3.61512095e-01 -8.22833836e-01
-7.64325678e-01 -7.35471368e-01 8.51299524e-01 -1.17850363e+00
1.62697546e-02 -6.46180689e-01 -1.43463767e+00 -4.28758979e-01
-9.59667116e-02 2.29222894e-01 7.13343143e-01 7.27802753e-01
5.94071686e-01 4.79176760e-01 1.37605205e-01 -1.23502957e-02
-1.03359497e+00 -4.75902051e-01 1.69283882e-01 6.24028981e-01
-9.37879309e-02 -4.22627181e-01 -2.19625100e-01 6.84131831e-02] | [8.053108215332031, 7.717579364776611] |
aa21c836-0395-4425-ae11-498ccecaa58c | multi-task-semantic-dependency-parsing-with | 1906.01239 | null | https://arxiv.org/abs/1906.01239v1 | https://arxiv.org/pdf/1906.01239v1.pdf | Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies | In Semantic Dependency Parsing (SDP), semantic relations form directed acyclic graphs, rather than trees. We propose a new iterative predicate selection (IPS) algorithm for SDP. Our IPS algorithm combines the graph-based and transition-based parsing approaches in order to handle multiple semantic head words. We train the IPS model using a combination of multi-task learning and task-specific policy gradient training. Trained this way, IPS achieves a new state of the art on the SemEval 2015 Task 18 datasets. Furthermore, we observe that policy gradient training learns an easy-first strategy. | ['Anders Søgaard', 'Shuhei Kurita'] | 2019-06-04 | multi-task-semantic-dependency-parsing-with-1 | https://aclanthology.org/P19-1232 | https://aclanthology.org/P19-1232.pdf | acl-2019-7 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [ 3.47752780e-01 6.91509724e-01 -7.05459774e-01 -6.71361089e-01
-8.64014208e-01 -7.41116047e-01 5.92920065e-01 3.19320083e-01
-4.53825742e-01 8.00524592e-01 2.67592639e-01 -7.14575648e-01
4.61595505e-02 -7.41716623e-01 -7.79813349e-01 -1.88928545e-01
-7.32639953e-02 1.10900080e+00 7.10981548e-01 -1.37969777e-01
-1.86335028e-03 1.62078917e-01 -1.02061439e+00 4.78678137e-01
8.14324141e-01 8.50972950e-01 2.95101315e-01 6.57225966e-01
-6.77167356e-01 1.04987013e+00 -3.41638327e-01 -7.51619875e-01
-1.61238492e-01 -2.82900870e-01 -1.53143132e+00 -3.49645346e-01
2.19446123e-01 1.86421618e-01 8.01135898e-02 9.39718366e-01
2.70667702e-01 2.23775640e-01 2.80544698e-01 -1.09536874e+00
-3.65712047e-01 1.30733705e+00 -1.61403269e-01 2.64915287e-01
4.19458479e-01 -3.30856085e-01 1.66829765e+00 -6.23503089e-01
9.44303393e-01 1.81295180e+00 3.50746989e-01 8.97585154e-01
-1.12760687e+00 -4.37357426e-01 6.13104761e-01 4.27087545e-01
-4.43099618e-01 -1.21512197e-01 7.72309065e-01 -3.58795226e-02
1.72762108e+00 -1.71516776e-01 3.80569398e-01 1.50734770e+00
-4.04627323e-02 1.22069883e+00 1.30166340e+00 -7.28458285e-01
2.99846262e-01 -3.81327420e-01 6.95763052e-01 9.38883960e-01
3.02644465e-02 8.09417367e-02 -8.50546360e-01 -2.35241145e-01
2.58105993e-01 -7.71690965e-01 2.54511476e-01 -3.38899523e-01
-8.53178859e-01 1.06165397e+00 1.11027554e-01 1.91819593e-01
-4.23073843e-02 2.73970723e-01 7.88224697e-01 4.36411828e-01
5.47527134e-01 5.53172231e-01 -1.42903602e+00 -1.76787317e-01
-2.87763238e-01 2.09575221e-02 1.32062626e+00 8.49099994e-01
6.75689936e-01 -1.78850010e-01 -1.13544844e-01 1.10364103e+00
3.82379174e-01 3.83296847e-01 3.48557115e-01 -9.82443392e-01
8.70790899e-01 3.49670798e-01 -2.88588345e-01 -3.18282157e-01
-7.64010906e-01 -3.28396000e-02 -2.35484660e-01 -1.51213914e-01
4.30743933e-01 -3.29118371e-01 -1.21938288e+00 2.07172847e+00
4.42099541e-01 1.32785395e-01 5.20939410e-01 3.55704814e-01
9.12875533e-01 5.90867043e-01 8.87715101e-01 -1.46042407e-01
1.52942407e+00 -1.25580132e+00 -6.21347606e-01 -8.76689732e-01
9.36597407e-01 -1.96053371e-01 1.04276776e+00 1.69960380e-01
-9.68199670e-01 -8.03896710e-02 -8.19760859e-01 -1.38433009e-01
-4.36910599e-01 -9.15198177e-02 9.43429708e-01 3.89708310e-01
-1.01866806e+00 8.18836451e-01 -1.16037250e+00 -5.00329792e-01
4.56727743e-01 1.90456256e-01 -4.01017010e-01 1.83634572e-02
-1.45909011e+00 1.19284904e+00 9.85029578e-01 -2.38929555e-01
-7.99073994e-01 -6.31022155e-01 -1.24461603e+00 4.72056568e-02
7.79081106e-01 -7.94005334e-01 1.50439036e+00 -6.26325428e-01
-1.97316396e+00 1.22472346e+00 -3.76054853e-01 -6.84039891e-01
1.07104994e-01 -5.41105211e-01 -1.19916379e-01 2.87579805e-01
2.50571907e-01 5.07362664e-01 6.42886341e-01 -1.20080400e+00
-9.81658936e-01 -4.24536943e-01 1.10958733e-01 2.33722001e-01
2.05001533e-01 4.65848237e-01 -4.07030165e-01 -1.09550379e-01
1.17095567e-01 -8.43646467e-01 -3.58293176e-01 -7.92639792e-01
-7.96082497e-01 -8.49152923e-01 6.01619899e-01 -6.66701257e-01
1.02884936e+00 -1.83901036e+00 4.24373418e-01 8.54579285e-02
3.71461883e-02 2.90269017e-01 -2.67243683e-01 3.05316120e-01
-1.41172083e-02 2.21141458e-01 -4.08983231e-01 -5.42320848e-01
5.87058887e-02 7.84008086e-01 -1.42189085e-01 -1.64416775e-01
5.48303962e-01 9.64897454e-01 -1.30197763e+00 -7.84714818e-01
4.12839837e-03 -4.85289507e-02 -4.35226321e-01 1.83468774e-01
-8.80097091e-01 5.47606766e-01 -7.33137250e-01 7.50514507e-01
4.32096034e-01 -3.52935523e-01 9.03070688e-01 2.27211848e-01
1.10335916e-01 9.36529219e-01 -5.74153423e-01 1.94427490e+00
-6.93041027e-01 9.70124230e-02 -9.03236121e-02 -1.40735614e+00
6.93902373e-01 2.21865088e-01 2.12484792e-01 -7.02254772e-01
-2.78483722e-02 3.44590932e-01 -1.25063956e-01 -3.14317286e-01
1.14897147e-01 -1.12394705e-01 -4.35765803e-01 1.13839716e-01
7.06140161e-01 -1.18120819e-01 4.98398185e-01 2.99591601e-01
1.33977449e+00 6.75400019e-01 5.47457576e-01 -4.04562026e-01
5.67337275e-01 1.31070569e-01 9.45959747e-01 8.06553721e-01
-7.68991634e-02 -5.60900643e-02 1.24165750e+00 -4.01462406e-01
-6.62759721e-01 -1.17237377e+00 -4.87873368e-02 1.53864014e+00
-1.86098710e-01 -6.01227283e-01 -5.71208119e-01 -1.55629385e+00
-3.31253484e-02 1.11360967e+00 -4.71497655e-01 -1.48704976e-01
-1.13750052e+00 -6.70233488e-01 3.76050651e-01 6.77827895e-01
4.59289521e-01 -1.52700353e+00 -4.35356200e-01 5.67831337e-01
-6.92647994e-02 -1.92132103e+00 2.71566421e-01 8.12142968e-01
-9.34386909e-01 -1.33180320e+00 9.12869796e-02 -1.24567449e+00
3.62814039e-01 -5.35465240e-01 1.75988483e+00 -3.54920067e-02
4.27483082e-01 2.15116009e-01 -7.69227147e-01 -3.25235099e-01
-6.12545848e-01 5.13410687e-01 -4.11747605e-01 -6.60124302e-01
3.87724340e-01 -5.14700055e-01 -1.33231487e-02 -2.15344414e-01
-1.75500184e-01 7.78618827e-03 4.97058928e-01 1.14176345e+00
6.34770215e-01 -4.34830636e-01 5.34076393e-01 -1.56649494e+00
5.23547649e-01 -3.98517191e-01 -7.17006862e-01 6.59699023e-01
-6.85149372e-01 5.89021266e-01 7.43808687e-01 -7.12256357e-02
-1.37804043e+00 1.67580396e-01 -5.07233977e-01 7.21470192e-02
-1.70350119e-01 5.30471921e-01 -2.82143921e-01 2.35211581e-01
3.21003765e-01 -9.88512766e-04 -5.45823157e-01 -6.38181090e-01
6.31818712e-01 -5.93625801e-03 4.38285440e-01 -1.24391341e+00
3.57534707e-01 1.02380358e-01 1.26895592e-01 -3.73871744e-01
-1.66687632e+00 -9.36508924e-02 -7.32180059e-01 3.37515175e-01
1.30588937e+00 -6.31180704e-01 -6.00596488e-01 3.59431773e-01
-1.42025077e+00 -9.82151866e-01 -2.27975011e-01 2.76583225e-01
-5.89638531e-01 2.65465587e-01 -9.52968597e-01 -5.12582183e-01
-4.87370521e-01 -7.98336267e-01 1.18051004e+00 4.21369560e-02
-9.08691138e-02 -1.27896702e+00 2.17943639e-01 2.04895750e-01
-4.25251462e-02 3.86332124e-02 1.41549504e+00 -1.37098277e+00
-1.52548805e-01 3.24236214e-01 -2.97588438e-01 2.98481971e-01
-9.54319537e-02 -1.70732275e-01 -7.95043826e-01 1.09550230e-01
-2.53065139e-01 -5.34062266e-01 1.12445092e+00 3.36897761e-01
9.40550327e-01 -1.44908726e-01 -5.62772214e-01 6.87196255e-01
1.35059655e+00 8.81239027e-02 1.22082077e-01 6.79904878e-01
9.78936553e-01 7.27122426e-01 7.35700846e-01 1.29908370e-03
7.92600274e-01 4.34382588e-01 3.77417028e-01 1.66742161e-01
-1.67771533e-01 -5.15932620e-01 2.95956403e-01 5.05462050e-01
1.44641057e-01 -1.69126093e-01 -1.30799818e+00 4.85635340e-01
-1.94558096e+00 -4.81174022e-01 -1.15211487e-01 1.77522159e+00
1.04921985e+00 5.20191371e-01 -1.92796677e-01 -1.66865081e-01
6.37429714e-01 3.38809758e-01 -4.15489435e-01 -9.26507533e-01
-2.53910422e-01 8.52470934e-01 5.65620542e-01 7.53072441e-01
-1.38629889e+00 2.08897948e+00 6.64112902e+00 4.83684152e-01
-7.26468205e-01 4.47576553e-01 3.65835547e-01 5.40454686e-01
-1.72869787e-01 4.41673070e-01 -1.15368009e+00 2.53806710e-01
1.18094075e+00 1.36509794e-03 1.76964447e-01 1.06094837e+00
-4.62874442e-01 -3.01879626e-02 -1.04831815e+00 4.91230011e-01
-3.98980767e-01 -1.11595547e+00 -1.07191270e-03 -6.23486340e-01
3.84552389e-01 3.96594971e-01 -3.99340481e-01 7.42189944e-01
1.18432486e+00 -5.84343970e-01 5.41199803e-01 -3.06032866e-01
4.10569847e-01 -4.74243373e-01 5.52028954e-01 1.38307303e-01
-1.28217316e+00 -3.13615471e-01 -1.48678020e-01 1.69526219e-01
6.54038787e-01 6.21794343e-01 -7.43966937e-01 7.95274496e-01
7.62995064e-01 7.53329933e-01 -2.72460341e-01 3.50957155e-01
-1.43878663e+00 8.68661344e-01 -2.50569463e-01 -7.44943693e-02
3.54875356e-01 -1.63295910e-01 5.68041503e-01 1.30906188e+00
-3.17504376e-01 -1.07154340e-01 4.36372221e-01 3.48907918e-01
-2.19100937e-01 8.04999694e-02 -4.61922348e-01 -4.82491404e-02
5.05016804e-01 1.15252209e+00 -8.24136496e-01 -6.15565419e-01
-6.39551699e-01 1.00810862e+00 1.02063262e+00 2.16557741e-01
-5.59057057e-01 1.30998284e-01 5.69226801e-01 -6.61422491e-01
5.47078252e-01 -3.29337865e-01 -3.53681803e-01 -1.12735224e+00
-1.35477647e-01 -6.72774255e-01 1.08923733e+00 -4.11866814e-01
-1.27591872e+00 6.02936029e-01 4.10620660e-01 -2.26812974e-01
-4.04569924e-01 -9.57638264e-01 -5.88909626e-01 3.68633986e-01
-2.04635835e+00 -1.32861507e+00 3.81852478e-01 4.65000123e-01
7.10479677e-01 7.11671039e-02 1.15860868e+00 -1.26002273e-02
-7.20977545e-01 3.03335875e-01 -4.05921608e-01 2.64033020e-01
3.95614177e-01 -1.87539005e+00 9.19551373e-01 9.18052971e-01
1.78934678e-01 4.98140045e-02 5.09571791e-01 -7.53947377e-01
-1.11507678e+00 -8.92695904e-01 1.40230441e+00 -4.43558842e-01
1.00676763e+00 -3.44875216e-01 -7.80534506e-01 1.17485821e+00
2.42604181e-01 2.54212599e-02 5.12937248e-01 8.52159619e-01
-5.94534814e-01 2.46352598e-01 -8.53867888e-01 2.84783512e-01
1.39833522e+00 -3.96176666e-01 -9.94849265e-01 6.70621753e-01
1.15005684e+00 -7.22238600e-01 -7.18634903e-01 6.04096889e-01
4.06068377e-02 -5.31239927e-01 8.94724846e-01 -1.15722024e+00
4.53406662e-01 4.24510211e-01 -6.64987415e-02 -1.56802940e+00
-1.63617149e-01 -6.33941770e-01 -3.92826021e-01 1.21875858e+00
1.06115353e+00 -8.65222394e-01 7.72472203e-01 3.92154664e-01
-5.15859544e-01 -7.40541577e-01 -1.00435209e+00 -9.25263882e-01
3.58383626e-01 -4.03802723e-01 3.68624359e-01 8.77099395e-01
1.42577291e-01 9.88970220e-01 -3.42856981e-02 1.88813969e-01
6.11846507e-01 2.11630896e-01 1.52549684e-01 -1.39492488e+00
-3.89945000e-01 -2.78743148e-01 -3.99665115e-03 -8.28400254e-01
1.10496843e+00 -1.13740873e+00 1.30669773e-01 -1.78213525e+00
-8.78237933e-02 -4.56529766e-01 -5.14023602e-01 9.64287877e-01
-5.31601191e-01 -7.09600925e-01 1.56827763e-01 -1.88681334e-01
-8.22459459e-01 3.46197784e-01 1.06332278e+00 1.86208084e-01
-2.57076889e-01 -8.31040263e-04 -5.21486044e-01 8.60819936e-01
1.08113086e+00 -8.49635303e-01 -3.36510867e-01 -6.29102409e-01
5.02174199e-01 7.17205629e-02 -5.43708280e-02 -4.63321924e-01
-1.92914382e-02 -8.76408294e-02 -2.14387149e-01 -3.65216732e-01
1.26026422e-01 -2.71193475e-01 -7.36048102e-01 5.41269183e-01
-3.28906596e-01 -5.39078861e-02 9.54794288e-02 4.04572994e-01
-2.81354219e-01 -2.07884640e-01 7.08374858e-01 -5.05792677e-01
-1.18934011e+00 2.20929965e-01 -1.07809700e-01 6.01534903e-01
6.89751565e-01 4.95131671e-01 -4.55181777e-01 3.24901074e-01
-1.03352928e+00 6.35105968e-01 -1.54311612e-01 4.26482260e-01
3.16465288e-01 -8.38300943e-01 -7.00782120e-01 -1.70336470e-01
-5.35167847e-03 3.28320384e-01 -3.17418426e-01 5.92650473e-01
-3.28623414e-01 2.44138300e-01 -7.45179802e-02 -3.95573378e-01
-1.11357164e+00 4.31747317e-01 1.49467856e-01 -9.94843185e-01
-8.66304159e-01 1.21431148e+00 -1.20824650e-01 -8.02202523e-01
-1.30330985e-02 -3.21522176e-01 -4.84216720e-01 1.57713741e-02
7.13416189e-02 -6.46451637e-02 6.75111488e-02 -1.16167195e-01
-7.13412404e-01 3.97085220e-01 -2.26723507e-01 -2.52068013e-01
1.41326153e+00 1.41705051e-01 -2.06822172e-01 2.75786728e-01
1.00073993e+00 -2.66334474e-01 -1.11719275e+00 -4.79970425e-01
9.53019977e-01 -5.55713475e-02 -2.67786324e-01 -1.03617954e+00
-8.36258829e-01 6.42975867e-01 -8.08101892e-02 2.79412359e-01
9.16061521e-01 6.14572167e-01 7.82085478e-01 6.57326639e-01
3.66590589e-01 -1.40929425e+00 -5.87036572e-02 1.17450106e+00
3.86052847e-01 -1.38244438e+00 -2.10264593e-01 -9.32089508e-01
-8.67610097e-01 9.60364759e-01 7.45415151e-01 -2.39034772e-01
6.48246169e-01 3.56770694e-01 -7.58751854e-02 -3.01889330e-01
-1.12072957e+00 -6.01908684e-01 -1.41107067e-01 7.33457685e-01
4.36004430e-01 1.91258997e-01 -6.69322908e-01 6.79530621e-01
-3.35517108e-01 -3.83855730e-01 -6.57618493e-02 1.25620210e+00
-4.00623858e-01 -1.74376702e+00 3.92002165e-01 1.43584624e-01
-6.78686857e-01 -4.72330540e-01 -4.87666994e-01 8.89308989e-01
-4.34609830e-01 9.03063238e-01 -2.34007701e-01 -1.32297456e-01
3.44982862e-01 6.04171753e-01 7.49405801e-01 -9.84717607e-01
-4.85278070e-01 -4.53856438e-01 9.90587652e-01 -1.03289044e+00
-5.13439357e-01 -5.75357676e-01 -1.73068523e+00 3.03886414e-01
-6.91489875e-02 2.61298716e-01 8.83198857e-01 1.39322340e+00
1.42095730e-01 5.31480968e-01 5.34213424e-01 -3.63893509e-01
-5.48089087e-01 -8.07673573e-01 -1.79873630e-01 3.83800209e-01
-1.35947615e-01 -6.60022736e-01 -2.39790514e-01 -2.73987979e-01] | [10.334651947021484, 9.544744491577148] |
5ab176e7-7867-4d9f-82bd-ac3a8c77d1bb | towards-part-based-understanding-of-rgb-d | 2012.02094 | null | https://arxiv.org/abs/2012.02094v1 | https://arxiv.org/pdf/2012.02094v1.pdf | Towards Part-Based Understanding of RGB-D Scans | Recent advances in 3D semantic scene understanding have shown impressive progress in 3D instance segmentation, enabling object-level reasoning about 3D scenes; however, a finer-grained understanding is required to enable interactions with objects and their functional understanding. Thus, we propose the task of part-based scene understanding of real-world 3D environments: from an RGB-D scan of a scene, we detect objects, and for each object predict its decomposition into geometric part masks, which composed together form the complete geometry of the observed object. We leverage an intermediary part graph representation to enable robust completion as well as building of part priors, which we use to construct the final part mask predictions. Our experiments demonstrate that guiding part understanding through part graph to part prior-based predictions significantly outperforms alternative approaches to the task of semantic part completion. | ['Angela Dai', 'Evgeny Burnaev', 'Alexey Artemov', 'Denis Zorin', 'Emil Bogomolov', 'Vladislav Ishimtsev', 'Alexey Bokhovkin'] | 2020-12-03 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Bokhovkin_Towards_Part-Based_Understanding_of_RGB-D_Scans_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Bokhovkin_Towards_Part-Based_Understanding_of_RGB-D_Scans_CVPR_2021_paper.pdf | cvpr-2021-1 | ['3d-instance-segmentation-1'] | ['computer-vision'] | [ 4.88268107e-01 6.55337393e-01 -3.17658260e-02 -6.33202493e-01
-3.48543316e-01 -8.13357830e-01 6.21579349e-01 1.98396608e-01
3.66705328e-01 -4.61553149e-02 2.43625805e-01 -1.48523197e-01
8.29547793e-02 -7.52782345e-01 -1.02884257e+00 1.93216689e-02
1.05009459e-01 1.01268172e+00 5.16294897e-01 5.30198440e-02
3.06504160e-01 9.67677355e-01 -1.75893009e+00 4.52708423e-01
6.51048779e-01 9.88452435e-01 4.40629810e-01 6.43958807e-01
-4.48067814e-01 4.53265965e-01 -7.59817334e-03 -4.09294665e-02
4.84545410e-01 5.12774475e-02 -1.04776216e+00 1.09574151e+00
3.28404188e-01 -4.98827100e-01 -1.54617593e-01 6.86244547e-01
-3.88974011e-01 2.44768187e-01 6.97142363e-01 -1.25828969e+00
-1.29617035e-01 2.58386314e-01 -4.24674630e-01 -5.27220368e-01
8.05689275e-01 2.37976998e-01 1.10042453e+00 -8.00137162e-01
7.98375666e-01 1.56287062e+00 4.48911607e-01 4.57715958e-01
-1.45770037e+00 -1.64451316e-01 6.98696136e-01 -1.73568517e-01
-1.12563574e+00 -4.27004904e-01 9.42421317e-01 -6.03332818e-01
1.27723145e+00 6.84992224e-02 1.22361612e+00 4.30180967e-01
-1.33482009e-01 1.01432395e+00 8.88224423e-01 -1.00264020e-01
3.49102795e-01 -1.79349303e-01 2.97393650e-01 7.87629724e-01
2.78346121e-01 -3.51400942e-01 -4.39506620e-01 -1.18618630e-01
1.22921550e+00 3.75377119e-01 -2.77213193e-02 -1.11200118e+00
-1.00385571e+00 3.59773457e-01 6.34244502e-01 -2.66595662e-01
-6.52812839e-01 4.14279282e-01 -2.79664427e-01 -2.73317456e-01
5.47019839e-01 2.60395825e-01 -8.00219774e-01 2.54254311e-01
-5.44573843e-01 5.79428852e-01 9.00097013e-01 1.34453249e+00
1.23142779e+00 -4.04695481e-01 2.24486291e-01 3.23722988e-01
6.69545829e-01 4.50819075e-01 -5.81077754e-01 -1.56103992e+00
2.77158618e-01 1.27966344e+00 2.83275664e-01 -3.56848687e-01
-4.85778630e-01 -1.66135967e-01 -9.14549232e-02 3.96237195e-01
2.86878675e-01 5.56832910e-01 -1.61652899e+00 1.39563942e+00
8.67887914e-01 1.82930037e-01 -2.32274920e-01 8.44350398e-01
5.09179950e-01 2.21218020e-01 1.98708773e-01 5.06669998e-01
1.43328226e+00 -6.99158967e-01 3.51206884e-02 -6.01386547e-01
3.72082174e-01 -4.38695967e-01 7.83712626e-01 2.56093681e-01
-1.23220384e+00 -5.24897814e-01 -9.00141776e-01 -5.15981019e-01
-1.97338820e-01 -4.22944218e-01 1.12755406e+00 3.79348218e-01
-9.01167512e-01 5.15977859e-01 -1.27174866e+00 -4.14897144e-01
9.21249211e-01 3.64253342e-01 -6.36834502e-01 -3.80602032e-01
-1.01339832e-01 7.96313763e-01 5.40368021e-01 -1.93080649e-01
-1.28723717e+00 -8.97369206e-01 -1.22559977e+00 6.21485077e-02
6.58567131e-01 -1.22908604e+00 1.40253019e+00 -5.74617147e-01
-9.73682284e-01 1.24370992e+00 -2.61445373e-01 -2.31251538e-01
1.46542728e-01 -5.34423828e-01 4.67893302e-01 2.19763562e-01
6.54865876e-02 1.08313227e+00 7.74460971e-01 -1.78519094e+00
-5.09637058e-01 -8.00262272e-01 4.70661938e-01 5.47689140e-01
6.84629977e-01 -5.45953155e-01 -7.90124714e-01 -6.51365519e-02
1.09769356e+00 -7.32667506e-01 -6.18183970e-01 5.67469180e-01
-7.43965030e-01 -1.14542216e-01 7.70421863e-01 -5.78443646e-01
1.11469209e-01 -2.03522944e+00 1.22032613e-01 2.82779217e-01
4.29458767e-01 -4.24075723e-01 -6.13918416e-02 2.97583282e-01
1.97705794e-02 1.18616425e-01 -4.60122824e-01 -7.23753572e-01
2.72677809e-01 5.22497594e-01 -4.20776010e-01 3.10564578e-01
4.53999251e-01 9.91162598e-01 -7.34992087e-01 -2.57376075e-01
6.66911185e-01 4.82220560e-01 -9.58205998e-01 4.37678516e-01
-1.09076035e+00 6.52661622e-01 -1.20131731e+00 8.21681559e-01
8.36223722e-01 -4.46183592e-01 8.63796659e-03 -2.39288196e-01
-2.32699029e-02 5.04644454e-01 -1.07537162e+00 2.37015438e+00
-2.30615035e-01 2.37234100e-03 3.79026711e-01 -8.22355688e-01
6.04080319e-01 9.05264691e-02 7.02392638e-01 -3.14409062e-02
2.14488097e-02 -1.50327414e-01 -4.84951019e-01 -2.33217448e-01
3.34722161e-01 -2.74602145e-01 -8.31500366e-02 5.98096728e-01
1.65812783e-02 -1.34100103e+00 -2.42179617e-01 5.09679317e-01
1.13037074e+00 1.00481021e+00 3.00927907e-01 -2.70486772e-02
1.47054285e-01 6.20016277e-01 2.17144236e-01 6.30779326e-01
1.69423312e-01 7.32219577e-01 2.85414487e-01 -3.78108323e-01
-1.06937814e+00 -1.47864377e+00 1.40645709e-02 5.43871224e-01
5.96667945e-01 -2.66533762e-01 -9.16870296e-01 -7.12420106e-01
3.40574622e-01 8.90970886e-01 -5.61882257e-01 1.81702137e-01
-4.55684334e-01 -1.72256857e-01 -3.69157225e-01 5.94240606e-01
2.91113943e-01 -7.96329975e-01 -8.85867000e-01 9.61499363e-02
-3.29066142e-02 -1.41593468e+00 -1.44011870e-01 3.95144999e-01
-1.19525075e+00 -1.43157768e+00 -1.14410385e-01 -4.35889691e-01
1.06171739e+00 6.73426688e-01 1.30576909e+00 3.27019453e-01
-5.98448038e-01 1.09280980e+00 -2.85669476e-01 -4.68040824e-01
-3.52010757e-01 -4.64539498e-01 -2.09312558e-01 -1.37619406e-01
-9.34616663e-03 -8.05011153e-01 -6.38814509e-01 1.70856953e-01
-7.82876432e-01 6.63519621e-01 7.14730024e-01 5.01206741e-02
1.14227653e+00 -5.49787171e-02 -4.04464275e-01 -8.33453357e-01
-2.83160806e-01 -2.59086341e-01 -5.49811006e-01 2.44283602e-01
-4.33319956e-02 2.60054350e-01 1.15223101e-03 -8.05856809e-02
-1.32495379e+00 6.35714769e-01 9.00118649e-02 -5.15067577e-01
-7.22516894e-01 -5.40247783e-02 -6.14341617e-01 1.38540104e-01
2.03767300e-01 3.49129550e-02 -2.64080405e-01 -8.30006719e-01
8.99946451e-01 8.07231218e-02 5.69935918e-01 -8.88913810e-01
8.41001034e-01 9.60647583e-01 2.15631917e-01 -6.52707040e-01
-1.15404677e+00 -6.14574850e-01 -1.26780534e+00 1.30315721e-02
1.32149065e+00 -1.12711215e+00 -6.49328947e-01 1.85809076e-01
-1.46599352e+00 -5.18659055e-01 -6.32661164e-01 1.65775299e-01
-1.08299041e+00 4.54704054e-02 -2.22719148e-01 -7.60814965e-01
3.22410077e-01 -1.02635288e+00 1.83997464e+00 1.44985868e-02
-2.92936683e-01 -7.14842379e-01 -2.12575689e-01 6.63442791e-01
-3.81585360e-01 4.38571066e-01 1.12359321e+00 -4.57442671e-01
-1.42502165e+00 -6.78538233e-02 -2.71204144e-01 -8.95914063e-02
3.25930268e-02 -3.08429182e-01 -1.18108749e+00 3.22546244e-01
1.61545098e-01 -1.64260641e-01 4.76583183e-01 5.26468277e-01
1.32739890e+00 2.13512838e-01 -8.03757071e-01 5.93139172e-01
1.36289430e+00 -8.69015157e-02 2.43157759e-01 -1.69581592e-01
9.52080965e-01 8.78747761e-01 6.15438879e-01 5.38669467e-01
6.15998805e-01 3.80392402e-01 7.94970393e-01 -8.02670270e-02
-4.48770374e-01 -6.96571171e-01 -3.26298833e-01 8.05785507e-02
1.92678329e-02 -1.02881745e-01 -1.02426839e+00 2.55879790e-01
-1.73913276e+00 -4.02660728e-01 -3.10865283e-01 1.92686224e+00
1.80464491e-01 3.75609368e-01 -9.04969573e-02 -9.26840529e-02
3.36952388e-01 -7.28640854e-02 -1.14978564e+00 1.04330935e-01
2.95050383e-01 2.20761329e-01 3.47016871e-01 5.81004679e-01
-8.69384408e-01 1.22361863e+00 6.48851013e+00 2.52066195e-01
-3.13656211e-01 -4.42183437e-03 5.98617136e-01 2.36368060e-01
-7.77850866e-01 6.26969337e-01 -6.15247071e-01 -5.35302043e-01
1.35510474e-01 1.90543115e-01 5.04637778e-01 9.73781109e-01
2.08932996e-01 -4.40822542e-01 -1.66784334e+00 8.48938823e-01
-1.03677638e-01 -1.08857632e+00 3.13718826e-01 3.03896308e-01
5.88089585e-01 2.85611693e-02 -5.13423860e-01 1.83771774e-02
6.53394580e-01 -8.23836923e-01 1.12924635e+00 6.40688837e-01
3.64677578e-01 -3.42503428e-01 -6.16106242e-02 6.75516725e-01
-1.41207182e+00 1.69611812e-01 -1.50811449e-01 -3.38168442e-01
4.86946046e-01 5.44877470e-01 -9.98204887e-01 5.22873342e-01
4.32114929e-01 8.27868342e-01 -2.86806494e-01 8.35653186e-01
-4.07292068e-01 1.15064666e-01 -5.42450905e-01 4.28697973e-01
-7.44551197e-02 -2.01424062e-01 5.05195856e-01 4.82428640e-01
-1.46738663e-02 6.98366165e-01 6.35869861e-01 1.40851450e+00
-5.40660620e-02 -4.77428585e-01 -5.56361139e-01 5.02929688e-02
3.29242110e-01 1.10957801e+00 -1.37802637e+00 -4.80942518e-01
-3.01820695e-01 1.11331213e+00 3.79477620e-01 3.65354180e-01
-5.24154186e-01 4.03238505e-01 1.07359838e+00 2.87094980e-01
4.07706648e-01 -6.43954754e-01 -8.79047871e-01 -1.10969102e+00
2.35715769e-02 -2.45433852e-01 2.91522071e-02 -1.37655127e+00
-8.55065584e-01 2.74802689e-02 4.26245093e-01 -8.66630673e-01
1.25221759e-01 -7.81044126e-01 -3.32164347e-01 6.79266810e-01
-1.25962770e+00 -1.57838583e+00 -4.50127512e-01 4.85161990e-01
9.33954239e-01 7.44723797e-01 7.11041391e-01 -5.78738749e-01
4.62208427e-02 -6.90797091e-01 -7.72017300e-01 -3.08707088e-01
-2.04813853e-01 -1.21366417e+00 8.45031023e-01 6.65215313e-01
3.27741146e-01 5.52340209e-01 6.26292229e-01 -1.15035331e+00
-1.87339103e+00 -8.38228881e-01 2.29592517e-01 -1.12639594e+00
2.98286498e-01 -7.89190471e-01 -6.46171272e-01 9.82502103e-01
-5.20413697e-01 1.39827266e-01 2.07577720e-01 5.84064983e-02
-6.20919406e-01 3.48242462e-01 -1.28890347e+00 5.52765429e-01
1.82089233e+00 -6.81989849e-01 -7.90679038e-01 4.02366459e-01
1.12804592e+00 -6.04244351e-01 -8.16514969e-01 5.31504035e-01
5.89593887e-01 -1.00204825e+00 1.39248800e+00 -6.01260602e-01
1.35561332e-01 -4.66198832e-01 -4.43493217e-01 -8.27084303e-01
-2.02875763e-01 -2.74001718e-01 -1.57784224e-01 6.62528157e-01
1.12410583e-01 -1.40700400e-01 1.25859571e+00 1.19762957e+00
-6.08592927e-01 -5.64472973e-01 -4.99132961e-01 -3.31609607e-01
-4.38440353e-01 -1.18352759e+00 8.02650452e-01 4.98404801e-01
-4.35052007e-01 5.79133965e-02 4.45382357e-01 6.60053372e-01
9.37179208e-01 6.39159560e-01 1.09582329e+00 -1.59058726e+00
-3.21712345e-01 -3.93387288e-01 -5.05051255e-01 -1.72974205e+00
2.23662019e-01 -7.14075565e-01 3.84055406e-01 -2.09000015e+00
3.81557286e-01 -6.30251169e-01 2.05234960e-01 7.27084517e-01
-9.47474167e-02 1.39992014e-01 2.13893726e-01 1.64563254e-01
-7.06938684e-01 4.78388458e-01 1.44633579e+00 -5.78154884e-02
-2.33434096e-01 -7.04239011e-02 -5.27228415e-01 1.03061211e+00
1.43092632e-01 -2.74207830e-01 -4.88817841e-01 -6.04879558e-01
-1.65956020e-01 2.21892387e-01 6.67040944e-01 -7.66078055e-01
-1.21528558e-01 -3.14244896e-01 5.99043131e-01 -1.02687776e+00
9.87443507e-01 -1.04618061e+00 3.47654104e-01 4.34025049e-01
-2.90616620e-02 -5.40095806e-01 1.69977486e-01 6.91632867e-01
2.80832171e-01 1.22664265e-01 2.92798460e-01 -6.57821119e-01
-9.94703174e-01 7.02366412e-01 7.96639696e-02 -1.55016974e-01
9.62145627e-01 -6.16011679e-01 2.52295405e-01 -1.20098628e-01
-1.04614854e+00 2.70771176e-01 1.04334664e+00 4.18442786e-01
8.37542236e-01 -8.34055781e-01 -3.01825285e-01 4.65002865e-01
1.99807823e-01 8.07155550e-01 3.04214537e-01 3.51949722e-01
-4.74957943e-01 3.70250642e-01 -6.99874992e-03 -1.02224231e+00
-9.81621802e-01 4.58248079e-01 2.33941451e-01 2.41033092e-01
-1.01568425e+00 1.04086888e+00 1.08457625e+00 -8.51128995e-01
4.06218991e-02 -8.16363573e-01 3.48801106e-01 -6.87623143e-01
1.41275495e-01 1.25227809e-01 -1.07915007e-01 -6.23228252e-01
-4.33766156e-01 9.53211248e-01 3.27836186e-01 -3.49470198e-01
1.40658832e+00 -2.77815938e-01 -6.00809082e-02 3.72344524e-01
6.33492112e-01 -3.12096149e-01 -1.91829634e+00 -5.50809093e-02
3.66301648e-02 -5.70708871e-01 -1.64814249e-01 -7.29881406e-01
-4.09707218e-01 8.62963676e-01 -8.88594761e-02 4.45160232e-02
7.91624248e-01 7.89459944e-01 4.65414852e-01 3.62068325e-01
8.70206237e-01 -5.45206249e-01 7.30179921e-02 4.97110337e-01
7.69946754e-01 -1.04363430e+00 3.63553762e-01 -1.13878369e+00
-2.48075277e-01 9.45609570e-01 6.13461673e-01 -1.86200589e-01
7.63704002e-01 1.46123603e-01 -2.80043513e-01 -7.57756531e-01
-5.56074739e-01 -4.98419255e-01 5.52275121e-01 7.29203224e-01
-2.61030137e-03 1.64835289e-01 6.70134544e-01 2.97991037e-01
-2.14479566e-01 -4.02286887e-01 -5.43623092e-03 1.01154208e+00
-7.80696809e-01 -1.00914383e+00 -5.48853576e-01 3.00011307e-01
1.26138091e-01 2.40200326e-01 -6.70342267e-01 8.06352198e-01
2.84333557e-01 7.82431901e-01 2.47648269e-01 -8.61257389e-02
4.88718420e-01 2.16022447e-01 1.01602483e+00 -1.31543458e+00
1.21877780e-02 1.70998186e-01 1.54848874e-01 -1.14411080e+00
-5.00696182e-01 -9.12482738e-01 -1.87669599e+00 3.69485587e-01
-1.65559292e-01 -1.86409071e-01 1.20724702e+00 1.38463318e+00
2.77528286e-01 4.19446200e-01 3.61716658e-01 -1.54000831e+00
1.29814729e-01 -5.42800605e-01 -5.02884209e-01 5.82668185e-01
2.90292442e-01 -7.02876449e-01 -6.15759194e-02 2.54378408e-01] | [8.346024513244629, -2.9343550205230713] |
16d2e85f-0b6a-492e-8792-7f1a6623ef75 | deep-alignment-network-a-convolutional-neural | 1706.01789 | null | http://arxiv.org/abs/1706.01789v2 | http://arxiv.org/pdf/1706.01789v2.pdf | Deep Alignment Network: A convolutional neural network for robust face alignment | In this paper, we propose Deep Alignment Network (DAN), a robust face
alignment method based on a deep neural network architecture. DAN consists of
multiple stages, where each stage improves the locations of the facial
landmarks estimated by the previous stage. Our method uses entire face images
at all stages, contrary to the recently proposed face alignment methods that
rely on local patches. This is possible thanks to the use of landmark heatmaps
which provide visual information about landmark locations estimated at the
previous stages of the algorithm. The use of entire face images rather than
patches allows DAN to handle face images with large variation in head pose and
difficult initializations. An extensive evaluation on two publicly available
datasets shows that DAN reduces the state-of-the-art failure rate by up to 70%.
Our method has also been submitted for evaluation as part of the Menpo
challenge. | ['Tomasz Trzcinski', 'Marek Kowalski', 'Jacek Naruniec'] | 2017-06-06 | null | null | null | null | ['robust-face-alignment'] | ['computer-vision'] | [-1.93060473e-01 2.82183617e-01 1.49916753e-01 -6.67648137e-01
-5.65712392e-01 -2.42304817e-01 7.04337001e-01 -5.12404554e-02
-5.80946803e-01 3.36054415e-01 1.90090805e-01 2.12768465e-01
2.52318531e-01 -3.83380353e-01 -7.14514375e-01 -4.83667731e-01
-7.75538236e-02 7.82182753e-01 4.65860330e-02 -2.58255690e-01
1.76580161e-01 8.66995156e-01 -1.48466563e+00 3.23368348e-02
9.77065489e-02 8.24888706e-01 -3.40465754e-01 3.97150069e-01
1.81704924e-01 2.66738273e-02 -5.61006188e-01 -5.87056279e-01
5.70905149e-01 -3.29863042e-01 -7.58451700e-01 -1.15377912e-02
1.33335173e+00 -4.77571100e-01 -8.80668536e-02 1.05974388e+00
8.86227429e-01 -1.52159510e-02 4.23744291e-01 -1.38541746e+00
-2.49114722e-01 3.85297686e-01 -1.03182948e+00 7.74828941e-02
3.96599263e-01 -9.13293958e-02 7.05947161e-01 -1.18086100e+00
6.03081346e-01 1.42439938e+00 1.17458916e+00 7.28749812e-01
-1.29756606e+00 -6.77674294e-01 -4.95632738e-02 1.90254226e-01
-1.79504502e+00 -1.03746939e+00 7.48249650e-01 -2.16955900e-01
9.37887847e-01 -9.07924473e-02 6.61065817e-01 7.60059118e-01
4.91139926e-02 4.55555350e-01 8.12919974e-01 -6.30961895e-01
-8.77621025e-03 -2.40804538e-01 -1.70722589e-01 1.07857478e+00
2.16887578e-01 5.95193403e-03 -5.31755209e-01 -2.06221476e-01
7.78206408e-01 -2.28277877e-01 -4.27651815e-02 -5.31928182e-01
-9.24522758e-01 6.59836233e-01 6.85320497e-01 3.00146043e-01
-4.05798316e-01 1.24916807e-01 2.58978963e-01 2.76030712e-02
5.12380481e-01 1.67857960e-01 -2.16908827e-01 2.27946222e-01
-1.37244809e+00 2.34983981e-01 7.14967251e-01 6.98570669e-01
7.95347691e-01 -1.50166273e-01 -8.20716023e-02 8.14952493e-01
4.18344617e-01 2.31107935e-01 2.75439382e-01 -9.22908723e-01
3.54497790e-01 4.28486854e-01 -1.84481978e-01 -1.18150568e+00
-8.56802642e-01 -1.80483274e-02 -7.81944096e-01 4.27646816e-01
7.13999629e-01 -1.52863130e-01 -1.27078199e+00 1.78627658e+00
5.95156908e-01 4.32661086e-01 -3.78288090e-01 6.74227715e-01
8.97643328e-01 1.31611884e-01 -3.21716107e-02 9.05919671e-02
1.29324126e+00 -1.06439984e+00 -6.80472374e-01 -2.66231567e-01
2.45191082e-01 -9.53311861e-01 6.04223907e-01 1.69217601e-01
-1.29607689e+00 -7.14483023e-01 -1.10493279e+00 -1.96265802e-01
-3.28488827e-01 2.32516438e-01 3.65905732e-01 9.13651049e-01
-1.88580108e+00 4.34205472e-01 -9.90538657e-01 -5.19231498e-01
7.61847734e-01 9.77610767e-01 -9.14340913e-01 4.08846177e-02
-5.62527180e-01 9.38763082e-01 -6.83131739e-02 5.81261098e-01
-7.14382410e-01 -6.74815714e-01 -9.95661795e-01 -7.28721321e-02
8.90540257e-02 -5.40746093e-01 1.26257586e+00 -9.22166944e-01
-1.60618591e+00 1.13456035e+00 -5.41201651e-01 -2.48761237e-01
6.54640675e-01 -3.53160560e-01 -7.23141879e-02 5.51661551e-02
-2.61825025e-02 1.25733149e+00 7.77285039e-01 -9.85419631e-01
-2.48038888e-01 -5.73993444e-01 -1.80566028e-01 -1.16150580e-01
-1.03697158e-01 3.21517020e-01 -9.67700183e-01 -4.19134617e-01
1.11189894e-01 -1.17676330e+00 -1.49441078e-01 1.80891767e-01
-4.55072969e-01 -2.75103062e-01 6.72512054e-01 -8.43719661e-01
7.23125517e-01 -1.87139940e+00 2.05842972e-01 5.07104576e-01
1.87756255e-01 3.65075171e-01 -4.28170860e-01 1.39305010e-01
-4.05333221e-01 -5.25409169e-02 -1.92779690e-01 -9.58801806e-01
-4.35055979e-03 -1.06275372e-01 1.11660689e-01 7.89681017e-01
1.82933792e-01 9.26433325e-01 -3.60626489e-01 -4.59506691e-01
1.06988676e-01 8.74297380e-01 -5.83981633e-01 1.17534637e-01
1.14896975e-01 4.41217303e-01 2.19956592e-01 7.26952195e-01
9.92676377e-01 1.46570161e-01 1.36321872e-01 -2.67468601e-01
4.05856334e-02 1.18716508e-01 -1.11597276e+00 1.81834674e+00
-2.63336301e-01 5.77672899e-01 1.88760027e-01 -6.24444008e-01
7.51707196e-01 4.92955714e-01 5.51671386e-01 -6.55078232e-01
2.69875824e-01 4.78302054e-02 -5.02254963e-02 -3.29671614e-02
3.15310597e-01 1.48509473e-01 3.97015154e-01 4.50987607e-01
2.51390576e-01 1.65932551e-01 1.10185422e-01 -9.60983709e-02
8.17871988e-01 2.18800351e-01 2.86439836e-01 -2.99951345e-01
6.18419349e-01 -4.54391211e-01 5.45633972e-01 3.22608083e-01
-3.75151217e-01 9.96431410e-01 4.86167014e-01 -9.09535050e-01
-1.03736877e+00 -6.82611763e-01 -3.19876164e-01 9.02970791e-01
-2.73247331e-01 -6.06367588e-01 -1.22313035e+00 -8.69539320e-01
-6.92437217e-02 7.80081078e-02 -1.04726815e+00 1.33579388e-01
-8.97049487e-01 -7.66008615e-01 5.89003325e-01 6.84111834e-01
5.59359372e-01 -1.16500843e+00 -3.54182661e-01 -3.41012306e-03
8.14534873e-02 -1.06799400e+00 -5.84092021e-01 -1.69694379e-01
-6.24411702e-01 -1.13702130e+00 -7.60234237e-01 -9.04301822e-01
1.19073129e+00 -1.31884620e-01 1.21942687e+00 4.80979651e-01
-5.30128419e-01 2.49186188e-01 -1.80353709e-02 -3.80852640e-01
-8.90982896e-02 2.97904342e-01 1.92672580e-01 5.16999699e-02
5.11531472e-01 -4.14566666e-01 -6.62789345e-01 4.46403772e-01
-5.73712111e-01 -2.44436488e-01 3.11772496e-01 7.43857741e-01
5.10608137e-01 -3.96408737e-01 2.93660462e-01 -7.49784768e-01
2.01502562e-01 -2.42712259e-01 -8.76650095e-01 2.15033546e-01
-2.85741061e-01 -2.50687357e-02 1.77126974e-01 1.86461341e-02
-7.44251192e-01 5.71384847e-01 -5.40339470e-01 -1.99140906e-01
-3.09642494e-01 -6.80647120e-02 -2.29198784e-01 -4.82655197e-01
4.73432243e-01 -3.12709272e-01 3.14996839e-01 -4.90172088e-01
2.73957670e-01 2.50468791e-01 6.38950706e-01 -3.70007277e-01
8.70466888e-01 4.50606108e-01 2.65126437e-01 -6.83336675e-01
-5.47165930e-01 -1.98465839e-01 -1.31801450e+00 -7.00638369e-02
8.43907118e-01 -7.41057098e-01 -8.78199399e-01 5.62111318e-01
-1.27182400e+00 -3.38354677e-01 1.55832827e-01 6.05456047e-02
-2.60645658e-01 1.85465500e-01 -4.21099544e-01 -3.35507095e-01
-4.74698812e-01 -1.33406377e+00 1.33403480e+00 4.89031672e-01
-3.02776992e-01 -9.20241177e-01 2.08199367e-01 7.69720152e-02
3.82602870e-01 3.11375827e-01 4.33649600e-01 -6.13074005e-01
-3.48826140e-01 -2.99778819e-01 -1.43058216e-02 4.18932177e-02
1.09201163e-01 2.01602250e-01 -1.17215264e+00 -6.18457079e-01
-4.20863658e-01 -2.57719576e-01 7.01662302e-01 4.70631987e-01
1.10983503e+00 -2.08978832e-01 -3.74860674e-01 8.31345141e-01
1.21746397e+00 -6.34522438e-02 8.79769683e-01 4.97670740e-01
7.51960576e-01 6.44829869e-01 1.70022428e-01 2.07739338e-01
4.64857936e-01 1.19124722e+00 4.40688640e-01 -4.64752734e-01
-3.39907765e-01 4.83807474e-02 2.10115746e-01 2.59187669e-01
-1.89748585e-01 2.04753920e-01 -1.12574995e+00 5.14820457e-01
-1.63129616e+00 -9.29321468e-01 2.02473506e-01 2.27651024e+00
6.73738003e-01 -2.56884784e-01 4.55647141e-01 -1.11226529e-01
6.62610888e-01 6.23076335e-02 -1.88517928e-01 -4.52154458e-01
1.98383749e-01 7.61571050e-01 1.65927738e-01 6.80251718e-01
-1.20636618e+00 1.09135652e+00 7.39669752e+00 2.64510721e-01
-1.11150002e+00 8.92785564e-02 8.53121102e-01 -2.97864079e-01
2.54344791e-01 -3.43720019e-01 -1.17704213e+00 2.75005043e-01
7.27630794e-01 3.28723371e-01 3.36491555e-01 8.01799715e-01
-2.10176200e-01 -1.44706354e-01 -1.34271741e+00 9.71743405e-01
4.29441482e-01 -1.16553104e+00 -2.42111742e-01 1.68449044e-01
6.68187976e-01 2.02907547e-01 1.96761757e-01 -2.50078440e-02
2.07902685e-01 -1.34068978e+00 5.17199516e-01 3.19691777e-01
5.73687553e-01 -1.07697904e+00 8.37077141e-01 -2.73671240e-01
-1.23088658e+00 2.58326143e-01 -4.96925354e-01 3.21209520e-01
4.61837314e-02 3.13013285e-01 -1.14012730e+00 2.36051604e-01
8.45932066e-01 3.34215552e-01 -1.00920498e+00 1.23481309e+00
-2.73247331e-01 2.86634922e-01 -7.03508139e-01 6.35804415e-01
6.27202764e-02 1.21577093e-02 2.55554497e-01 1.13177645e+00
2.66325444e-01 -2.61813551e-01 3.10239904e-02 5.02250135e-01
-3.36016268e-01 1.89585626e-01 -4.93597507e-01 5.49218535e-01
3.82218748e-01 1.58026135e+00 -7.84535408e-01 1.17717208e-02
-3.93877685e-01 1.03583241e+00 5.95655739e-01 8.96207541e-02
-7.71151006e-01 -2.90764153e-01 5.51067770e-01 1.15605056e-01
3.50676358e-01 -3.09226632e-01 -1.09595433e-01 -6.18349314e-01
1.32295653e-01 -8.59763682e-01 3.88076812e-01 -3.67041171e-01
-9.40684319e-01 1.03505826e+00 3.77151258e-02 -7.41003573e-01
-4.50720400e-01 -5.02871037e-01 -7.88775265e-01 1.03545332e+00
-1.43375397e+00 -1.45676160e+00 -3.56667489e-01 6.03316188e-01
2.53413677e-01 -2.77655631e-01 1.00989032e+00 3.72668564e-01
-9.29224610e-01 1.21905565e+00 -3.14672530e-01 4.38353151e-01
9.87154067e-01 -1.12474740e+00 9.41520691e-01 8.83581638e-01
4.09480065e-01 7.89196610e-01 4.80841756e-01 -5.12937427e-01
-9.15456653e-01 -8.62686694e-01 1.09683657e+00 -5.90626061e-01
8.14804137e-02 -4.38497752e-01 -7.73533821e-01 9.31076586e-01
5.93498230e-01 3.22238028e-01 5.57617605e-01 3.23660880e-01
-4.61364925e-01 -3.16743225e-01 -1.36740446e+00 4.97313380e-01
9.26648974e-01 -3.44254285e-01 -4.22158808e-01 2.99504548e-01
-8.74798894e-02 -6.67351067e-01 -6.59395039e-01 4.48542029e-01
7.14944065e-01 -1.04957676e+00 1.10054159e+00 -4.39170390e-01
9.12670121e-02 -3.74093354e-01 1.81081802e-01 -1.36436987e+00
-3.19273800e-01 -6.63654745e-01 1.26985371e-01 1.29713213e+00
4.05492812e-01 -4.13271248e-01 8.75562966e-01 6.89761162e-01
8.81560892e-02 -5.80548823e-01 -1.18032551e+00 -3.82313043e-01
-1.07582204e-01 1.17601864e-01 8.71664703e-01 8.35528553e-01
-3.38970542e-01 7.24887326e-02 -1.63908005e-01 2.98818767e-01
6.10691965e-01 -2.61603624e-01 1.00682390e+00 -1.29161024e+00
1.49057120e-01 -6.14529610e-01 -6.81951761e-01 -4.87481058e-01
4.53420758e-01 -9.09312248e-01 1.33198962e-01 -1.28814292e+00
1.35765970e-01 -4.11100268e-01 -1.06401451e-01 1.14677644e+00
-1.68884039e-01 1.05082786e+00 2.09270939e-01 -7.41980001e-02
-3.58728200e-01 3.89780134e-01 7.12105513e-01 -4.07052524e-02
-1.19867280e-01 -1.85028180e-01 -4.94772613e-01 8.05805027e-01
8.15065861e-01 -4.10213739e-01 9.98417363e-02 -5.76509893e-01
-1.25145525e-01 -5.28145730e-01 3.30576539e-01 -1.27323425e+00
2.66601771e-01 4.39142764e-01 1.06318069e+00 -5.86345017e-01
5.15755892e-01 -9.07007337e-01 1.54111698e-01 3.72869611e-01
-1.40884309e-03 7.09851861e-01 6.64442480e-01 4.48284522e-02
-1.51635274e-01 -4.94232820e-03 1.08028638e+00 1.44499853e-01
-5.02291679e-01 4.35245246e-01 1.75687879e-01 -3.46590191e-01
9.65648890e-01 -2.25788221e-01 -7.94438124e-02 -2.48353407e-01
-7.30857790e-01 1.43879637e-01 5.21529377e-01 5.66009104e-01
4.12405372e-01 -1.46427763e+00 -7.92298973e-01 6.17592454e-01
-1.41137138e-01 -8.17721337e-02 -3.51950340e-02 1.09746516e+00
-7.48326838e-01 3.05590451e-01 -6.11631453e-01 -8.10245574e-01
-1.73608506e+00 3.02717507e-01 5.64031303e-01 -5.16255014e-02
-5.21509588e-01 1.18578005e+00 1.45956874e-02 -4.55169231e-01
3.43965650e-01 9.45152268e-02 -3.28028888e-01 1.65453702e-01
9.10196722e-01 2.84493476e-01 5.27135551e-01 -1.25538743e+00
-7.63724208e-01 1.00282478e+00 -3.38195652e-01 -2.31791675e-01
1.46385670e+00 1.23262450e-01 -5.11462033e-01 -3.19232285e-01
1.27882206e+00 1.40951246e-01 -1.15913916e+00 -9.89769325e-02
7.49664605e-02 -5.25669158e-01 5.29993623e-02 -7.73903787e-01
-1.56695473e+00 8.08404326e-01 8.75127912e-01 -4.48729664e-01
1.09306729e+00 -2.14495108e-01 6.06488287e-01 1.39306396e-01
2.54645318e-01 -8.48049581e-01 -5.61123155e-02 5.65874696e-01
1.13099480e+00 -1.19555807e+00 7.77850226e-02 -2.24369988e-01
-3.12667042e-01 1.24049413e+00 8.38164032e-01 4.58615571e-02
6.97340369e-01 3.26340288e-01 3.84483427e-01 -2.87083983e-01
-2.25994378e-01 -1.49498492e-01 5.92948139e-01 5.68997562e-01
5.72448492e-01 -2.65120715e-01 -1.59042664e-02 2.12352220e-02
-3.52270871e-01 -1.71404272e-01 1.37139193e-03 8.11690748e-01
-1.38980061e-01 -1.58350527e+00 -6.13674879e-01 -2.67912317e-02
-5.10983288e-01 8.55478719e-02 -6.00977838e-01 9.08644378e-01
1.85745686e-01 6.90125585e-01 3.63885760e-01 -2.43290842e-01
3.01393121e-01 3.10661614e-01 8.53911281e-01 -4.75318015e-01
-7.14008570e-01 5.27125560e-02 -1.50285244e-01 -9.28063333e-01
-4.39318001e-01 -7.97852278e-01 -1.17269838e+00 -4.17801350e-01
-2.16214471e-02 -7.13764876e-02 8.14492345e-01 7.93542624e-01
5.16374230e-01 1.47778690e-01 4.53757882e-01 -1.52977037e+00
-1.36238739e-01 -1.00583875e+00 -1.22554980e-01 2.74345636e-01
3.91684502e-01 -7.78242528e-01 -1.70883797e-02 2.15828642e-02] | [13.485325813293457, 0.3320131301879883] |
66671a43-daa8-43ee-bc4e-09dfbf8c03c7 | language-expansion-in-text-based-games | 1805.07274 | null | http://arxiv.org/abs/1805.07274v1 | http://arxiv.org/pdf/1805.07274v1.pdf | Language Expansion In Text-Based Games | Text-based games are suitable test-beds for designing agents that can learn
by interaction with the environment in the form of natural language text. Very
recently, deep reinforcement learning based agents have been successfully
applied for playing text-based games. In this paper, we explore the possibility
of designing a single agent to play several text-based games and of expanding
the agent's vocabulary using the vocabulary of agents trained for multiple
games. To this extent, we explore the application of recently proposed policy
distillation method for video games to the text-based game setting. We also use
text-based games as a test-bed to analyze and hence understand policy
distillation approach in detail. | ['Balaraman Ravindran', 'Ghulam Ahmed Ansari', 'Sagar J P', 'Sarath Chandar'] | 2018-05-17 | null | null | null | null | ['text-based-games'] | ['playing-games'] | [-8.20470303e-02 2.13523850e-01 2.28330772e-02 9.42106619e-02
-9.95347276e-02 -6.73528135e-01 1.06704223e+00 -1.81014597e-01
-1.07140458e+00 9.77251112e-01 1.13764964e-01 -7.82872140e-01
-5.61625510e-02 -1.19917047e+00 -4.51690793e-01 -4.86338228e-01
8.32541753e-03 9.77573991e-01 4.20234352e-01 -1.15877581e+00
9.94293615e-02 2.31067047e-01 -1.55395091e+00 1.78566396e-01
6.90416217e-01 2.89925814e-01 6.98003054e-01 9.83485460e-01
-1.57765284e-01 1.33557189e+00 -8.58780861e-01 -2.33652458e-01
5.53349257e-01 -5.43983102e-01 -7.40159094e-01 -6.36696070e-02
-1.40424579e-01 -5.89092016e-01 -6.69204056e-01 1.09578300e+00
3.03197026e-01 6.63614333e-01 3.90432656e-01 -9.86759961e-01
-1.97314039e-01 9.15984869e-01 1.44233540e-01 1.41119599e-01
4.48890179e-01 4.51398253e-01 1.05315173e+00 -1.49501443e-01
6.60248995e-01 1.26759899e+00 -7.84015134e-02 1.10042310e+00
-7.58686066e-01 -5.85030973e-01 2.68884838e-01 2.82854140e-01
-9.30626512e-01 -1.87904835e-02 5.70603132e-01 -3.57676834e-01
1.43317020e+00 1.00344308e-02 1.13107848e+00 1.19419765e+00
1.78238168e-01 9.44103777e-01 1.17366850e+00 -6.41226292e-01
5.73963284e-01 -1.23530954e-01 -4.89475608e-01 7.72925079e-01
9.88217965e-02 8.42973232e-01 -1.50850803e-01 -2.86337547e-02
1.00091791e+00 -3.71198863e-01 2.97682405e-01 -1.14962779e-01
-9.53943908e-01 1.42379189e+00 -1.72926560e-02 3.93466204e-01
-3.86386871e-01 4.69547361e-01 6.49334669e-01 6.37817860e-01
2.97111392e-01 1.07698011e+00 -5.36869586e-01 -6.39744520e-01
-5.20731091e-01 8.86340797e-01 1.11115122e+00 7.95127153e-01
5.31318605e-01 4.26290512e-01 -1.15094431e-01 4.81091619e-01
3.90378684e-01 5.20181239e-01 7.48654902e-01 -8.41167450e-01
4.07977313e-01 3.76189768e-01 1.40696704e-01 -3.98384601e-01
-5.74844778e-01 1.07735664e-01 -4.34590727e-01 6.14688873e-01
3.78734350e-01 -7.86341250e-01 -8.99221599e-01 1.68351924e+00
1.62497848e-01 2.84851819e-01 4.71682817e-01 5.82462788e-01
6.26993418e-01 7.65056908e-01 1.92692265e-01 -2.15983585e-01
1.40943611e+00 -9.82725859e-01 -4.80478197e-01 -4.35856640e-01
8.85132253e-01 -2.26875961e-01 1.02036548e+00 5.63899755e-01
-9.93267357e-01 -2.88588732e-01 -7.61428654e-01 4.58177626e-01
-6.30873621e-01 -2.80035049e-01 7.59208977e-01 7.84004807e-01
-1.23978519e+00 4.19561952e-01 -8.20275605e-01 -5.62210441e-01
1.25228122e-01 7.19438434e-01 -2.86382474e-02 2.75050431e-01
-1.61997223e+00 1.17321253e+00 9.93740380e-01 -4.43985045e-01
-1.18436062e+00 -8.44301060e-02 -9.99908686e-01 4.07886244e-02
9.38652635e-01 -8.10421050e-01 1.72964633e+00 -7.83856392e-01
-2.44904137e+00 5.41612148e-01 4.19885933e-01 -6.66464388e-01
4.69728529e-01 1.71181500e-01 8.36756006e-02 1.24450631e-01
-1.59609482e-01 5.96390784e-01 8.21914554e-01 -7.53078222e-01
-1.10630167e+00 -5.90506271e-02 1.13038516e+00 6.41839683e-01
-1.53430641e-01 7.89938644e-02 -8.25786591e-02 -6.15057409e-01
-8.50358367e-01 -9.18504775e-01 -6.35695398e-01 -5.59259892e-01
9.46067944e-02 -5.79810500e-01 5.29300630e-01 -1.17686577e-01
1.04058087e+00 -1.80281520e+00 4.61386710e-01 3.41111086e-02
3.59624296e-01 5.17978609e-01 -3.30719858e-01 7.34426618e-01
2.30854884e-01 2.08090156e-01 2.97837079e-01 -6.39292924e-03
2.41989598e-01 8.44344437e-01 -1.07681841e-01 -6.81435466e-02
7.54593536e-02 1.08666229e+00 -1.20616055e+00 -3.03858966e-01
6.76512420e-01 -2.64604360e-01 -1.10180736e+00 2.45700866e-01
-9.43841755e-01 4.47621077e-01 -9.99955416e-01 -2.55989611e-01
1.91035848e-02 2.05402702e-01 3.19675952e-01 6.08767629e-01
-1.87677816e-01 2.68541962e-01 -9.15042162e-01 1.73928881e+00
-5.36056280e-01 6.44801855e-01 -1.90347090e-01 -1.23162031e+00
6.13366425e-01 4.58758891e-01 5.20079911e-01 -9.10159051e-01
4.48980749e-01 -1.05749749e-01 4.99031663e-01 -5.97661436e-01
7.33417153e-01 -6.38388872e-01 -3.20463777e-01 6.79516733e-01
1.56644374e-01 -6.98676705e-01 8.20058823e-01 1.03915319e-01
1.11637402e+00 1.92616150e-01 6.32092655e-01 -1.43786743e-01
5.07036030e-01 2.89553136e-01 -1.02518864e-01 1.42183530e+00
-1.50074705e-01 -1.58713996e-01 4.04381782e-01 -5.84623396e-01
-1.26513624e+00 -5.85068405e-01 4.49640512e-01 1.61557555e+00
-4.38604206e-02 -8.26516688e-01 -7.54141271e-01 -6.61784291e-01
-3.18675578e-01 7.99485385e-01 -7.07063913e-01 -2.15454027e-01
-6.75074816e-01 -5.66820860e-01 7.32208014e-01 4.12210107e-01
5.96920431e-01 -1.54574823e+00 -1.13155091e+00 4.90437567e-01
7.64885396e-02 -1.10806751e+00 -1.09804794e-01 3.63857895e-01
-5.53441346e-01 -9.35819805e-01 -4.92380589e-01 -7.99557865e-01
-4.54948982e-03 -1.09459229e-01 9.90456104e-01 1.23051882e-01
1.73244551e-01 6.91819847e-01 -8.02990675e-01 -6.50678575e-01
-9.41493213e-01 2.32338443e-01 1.48747459e-01 -6.70922816e-01
2.81338811e-01 -3.74497116e-01 -1.96853638e-01 1.32338524e-01
-1.21963346e+00 4.83899176e-01 2.54769444e-01 1.05203319e+00
-1.09067187e-01 2.81630367e-01 3.64883542e-01 -6.42905533e-01
1.33217108e+00 -1.50906876e-01 -8.92771244e-01 8.65003541e-02
-1.23198226e-01 4.47716802e-01 1.01612866e+00 -8.12679887e-01
-9.00694728e-01 -2.73851454e-01 -4.23142463e-01 2.33016331e-02
-2.74163157e-01 8.08334708e-01 3.28422040e-01 3.41171585e-02
7.52823114e-01 4.65192407e-01 5.71368448e-02 2.11230278e-01
5.27273178e-01 6.48626089e-01 -1.55500054e-01 -9.87574160e-01
7.94778287e-01 8.53153914e-02 -1.61374673e-01 -9.66175556e-01
-3.08575660e-01 -2.06869140e-01 -1.42784119e-01 -1.78047791e-01
9.78459835e-01 -6.72763526e-01 -1.15556455e+00 5.12364447e-01
-9.99180853e-01 -1.14760494e+00 -6.02853417e-01 4.70028281e-01
-1.04788697e+00 1.44619942e-01 -5.30043125e-01 -6.48350060e-01
-6.95748180e-02 -1.45433986e+00 8.48171771e-01 4.48035955e-01
1.52673656e-05 -1.26023245e+00 5.36234319e-01 2.91255973e-02
3.12930912e-01 -3.16544950e-01 7.04270422e-01 -1.06922162e+00
-1.00856118e-01 8.84019658e-02 2.26775408e-01 5.91937676e-02
1.37533337e-01 -4.25121874e-01 -5.97766876e-01 -4.62551534e-01
-4.15296964e-02 -5.56940556e-01 3.23878944e-01 4.35233384e-01
7.29236484e-01 -1.99089900e-01 7.07505271e-02 3.36607665e-01
1.29144156e+00 7.24190056e-01 7.76350260e-01 8.77795100e-01
3.44532639e-01 2.24732861e-01 4.90175754e-01 7.42068529e-01
3.23198557e-01 6.91340744e-01 4.66535270e-01 2.53280342e-01
4.73171860e-01 -2.30840340e-01 4.23180640e-01 4.73790079e-01
-4.00557369e-01 -5.90958953e-01 -8.55701983e-01 9.27563235e-02
-1.97583449e+00 -1.18959653e+00 5.40064871e-01 1.58939636e+00
8.57286334e-01 3.01071346e-01 4.31724876e-01 -3.05789798e-01
3.96596789e-01 1.37577429e-01 -4.49553162e-01 -6.49028122e-01
1.65200487e-01 8.20723712e-01 3.00863713e-01 7.23969817e-01
-1.05296445e+00 1.68739986e+00 6.42131901e+00 1.03019309e+00
-1.04403698e+00 -1.15324557e-01 5.25425039e-02 -1.93032846e-02
-2.11341097e-03 -8.18350315e-02 -6.03935122e-01 1.41217530e-01
9.27940309e-01 -5.11973441e-01 9.78420317e-01 7.08882153e-01
6.65708721e-01 -2.81990230e-01 -9.02927876e-01 7.18540668e-01
-3.53295952e-01 -1.38175249e+00 3.42107058e-01 2.83890635e-01
5.86522460e-01 1.97872654e-01 -4.40373607e-02 1.03016412e+00
1.11165679e+00 -1.02956736e+00 6.82731330e-01 -2.11053967e-01
5.42814016e-01 -5.88642359e-01 5.59082150e-01 7.27192521e-01
-1.10979712e+00 -4.41810936e-01 -4.33659941e-01 -7.84077227e-01
-1.44763082e-01 -7.66933441e-01 -1.06509757e+00 5.70993781e-01
3.44210088e-01 5.75147986e-01 -2.27610230e-01 7.80779481e-01
-4.22097236e-01 5.19166887e-01 -4.13312554e-01 -7.13994265e-01
9.90628898e-01 -3.77229989e-01 5.72327733e-01 7.87267268e-01
1.23447433e-01 4.89162862e-01 5.47046185e-01 4.70032722e-01
2.96766609e-01 2.09734425e-01 -9.24366713e-01 -4.62904245e-01
-2.45238487e-02 7.37138987e-01 -6.24611259e-01 -5.26091516e-01
-5.21781623e-01 8.11298609e-01 3.52290273e-01 4.68971848e-01
-7.01169908e-01 -2.45660827e-01 7.45259404e-01 -2.52923310e-01
4.01592284e-01 -4.21953380e-01 3.62594664e-01 -1.42217147e+00
-5.25331795e-01 -1.45610666e+00 1.44668117e-01 -8.85049999e-01
-9.21440840e-01 6.77260220e-01 4.52247798e-01 -8.82591128e-01
-8.31945300e-01 -1.00027549e+00 -6.75464213e-01 5.80826461e-01
-1.04762375e+00 -7.97393262e-01 -1.65990591e-02 9.86058295e-01
8.92359555e-01 -8.39772701e-01 9.31150675e-01 -1.86071172e-01
-3.79944861e-01 1.80176720e-01 3.35036784e-01 2.75965363e-01
1.75218821e-01 -1.35037887e+00 5.07665813e-01 6.11049950e-01
2.67039806e-01 4.71341372e-01 1.05598783e+00 -5.32702744e-01
-1.34295619e+00 -7.65499413e-01 -4.50762548e-02 -1.71617836e-01
1.14064288e+00 -3.86116624e-01 -2.63707340e-01 6.95118248e-01
5.68237066e-01 -7.00165033e-01 3.71692687e-01 -8.93568248e-02
2.52081513e-01 3.18979889e-01 -8.82189751e-01 1.24204814e+00
9.02324557e-01 -3.61568689e-01 -8.88321042e-01 1.22784361e-01
8.83206725e-01 -6.03083670e-01 -4.76255685e-01 -8.26286711e-03
2.89112121e-01 -5.91532052e-01 7.24915624e-01 -1.15698290e+00
4.90558594e-01 7.87853543e-03 -9.55156535e-02 -1.90752470e+00
-1.88811004e-01 -1.00301492e+00 1.99777797e-01 2.88772792e-01
1.96373254e-01 -7.09921062e-01 1.04114211e+00 2.77143836e-01
3.19001228e-02 -4.44807410e-01 -8.88711393e-01 -5.64736784e-01
7.28505075e-01 -6.69596493e-01 5.53683162e-01 5.78966320e-01
5.21600842e-01 4.31398273e-01 -3.93066525e-01 -2.44738162e-01
4.26955260e-02 -4.06661421e-01 8.70512724e-01 -1.02824140e+00
-8.10289681e-01 -6.58141673e-01 -3.54602039e-01 -1.35749805e+00
3.10246378e-01 -7.31129944e-01 1.08351789e-01 -1.35693431e+00
-2.47921813e-02 -2.10057318e-01 1.90200090e-01 3.55453104e-01
-1.11475021e-01 -2.08625883e-01 6.00118160e-01 -2.42169499e-01
-7.41549075e-01 7.06034958e-01 1.66645801e+00 -3.48376870e-01
-4.16070908e-01 7.09652677e-02 -5.92190325e-01 5.72627187e-01
1.16633344e+00 -3.34586233e-01 -8.71638775e-01 -2.85779953e-01
6.99932694e-01 1.91087633e-01 -5.05257621e-02 -7.18033075e-01
2.54144073e-01 -8.11489820e-01 -4.06541169e-01 9.37137455e-02
1.74619004e-01 -7.90409029e-01 -4.74463314e-01 7.15877056e-01
-5.09358525e-01 1.18033104e-01 5.62831283e-01 4.71199751e-01
-1.06136642e-01 -6.05654180e-01 4.02365446e-01 -6.36906147e-01
-9.09287632e-01 3.68782640e-01 -1.52089572e+00 3.70826930e-01
1.24301219e+00 -2.05083534e-01 -1.47244245e-01 -7.68480539e-01
-8.84603620e-01 4.36107159e-01 2.50196010e-01 3.38714719e-01
6.09978020e-01 -8.97965133e-01 -7.27360010e-01 6.06613010e-02
2.72541586e-02 -2.31909886e-01 -1.46692649e-01 8.03736001e-02
-8.67528498e-01 5.04641294e-01 -5.36288738e-01 -1.14313021e-01
-1.04922795e+00 5.33454537e-01 7.88505375e-01 -8.38431716e-01
-5.60763180e-01 3.61492723e-01 5.49329817e-01 -5.87369382e-01
3.98311988e-02 -5.16682863e-01 -5.37672222e-01 -3.35176080e-01
5.55405974e-01 -2.25371584e-01 -3.65927190e-01 -7.01805428e-02
4.03480604e-02 1.94504589e-01 -1.99299112e-01 -7.07430840e-01
1.42046416e+00 1.47916123e-01 2.10732922e-01 -2.97066420e-02
4.76048261e-01 -4.89877880e-01 -1.03270304e+00 -1.60071880e-01
-1.52230293e-01 -4.76226732e-02 -4.22656536e-02 -5.32564282e-01
-7.76238739e-01 5.74631095e-01 3.46270502e-01 5.41216493e-01
9.02879536e-01 -1.08674414e-01 1.83477059e-01 8.97912621e-01
8.04544151e-01 -1.09605753e+00 3.76238555e-01 1.32276881e+00
5.35562515e-01 -1.12154651e+00 -3.52563649e-01 2.83487171e-01
-6.64712667e-01 1.19981956e+00 7.31864154e-01 -4.02096182e-01
4.63619411e-01 4.67150778e-01 -1.49249941e-01 -2.20665231e-01
-1.05630958e+00 -6.92415476e-01 -2.16626883e-01 9.50729251e-01
7.26409480e-02 6.01749346e-02 -4.10308331e-01 1.40575513e-01
-4.54947323e-01 1.61870956e-01 9.33809340e-01 1.20495808e+00
-6.11035228e-01 -1.58503139e+00 -2.11204648e-01 4.58128393e-01
-2.37016201e-01 -3.00211936e-01 -4.25246477e-01 9.85378683e-01
-6.89406991e-02 1.14964354e+00 -8.76982287e-02 -5.46686649e-01
2.15982482e-01 -6.57146946e-02 9.27857757e-01 -1.02048385e+00
-9.17305589e-01 -1.49094790e-01 3.19039971e-01 -2.85657376e-01
-4.17626143e-01 -4.51634318e-01 -1.16672301e+00 -4.28877562e-01
-2.08240315e-01 4.06981677e-01 3.43313158e-01 1.48467362e+00
-2.68257588e-01 6.54277980e-01 2.17782378e-01 -9.02359664e-01
-6.46652102e-01 -1.06382191e+00 -6.25917614e-01 2.29444668e-01
1.12844802e-01 -7.09317625e-01 5.29324412e-02 -2.51080871e-01] | [3.803438663482666, 1.473789095878601] |
ffd7927f-507b-4eba-92ac-78eaecbe4b9a | continual-training-of-language-models-for-few | 2210.05549 | null | https://arxiv.org/abs/2210.05549v1 | https://arxiv.org/pdf/2210.05549v1.pdf | Continual Training of Language Models for Few-Shot Learning | Recent work on applying large language models (LMs) achieves impressive performance in many NLP applications. Adapting or posttraining an LM using an unlabeled domain corpus can produce even better performance for end-tasks in the domain. This paper proposes the problem of continually extending an LM by incrementally post-train the LM with a sequence of unlabeled domain corpora to expand its knowledge without forgetting its previous skills. The goal is to improve the few-shot end-task learning in these domains. The resulting system is called CPT (Continual PostTraining), which to our knowledge, is the first continual post-training system. Experimental results verify its effectiveness. | ['Bing Liu', 'Lei Shu', 'Hu Xu', 'Yijia Shao', 'Haowei Lin', 'Zixuan Ke'] | 2022-10-11 | null | null | null | null | ['continual-pretraining'] | ['methodology'] | [ 2.27812052e-01 3.14299196e-01 -2.59676814e-01 -3.36604387e-01
-1.02560437e+00 -7.73628235e-01 6.29022837e-01 -3.13852429e-01
-9.08716023e-01 1.27652407e+00 -1.49110392e-01 -6.06969059e-01
4.04080123e-01 -3.65305990e-01 -7.17241824e-01 -2.55593061e-01
6.78204447e-02 1.04981303e+00 6.46917284e-01 -9.79186371e-02
8.08236524e-02 2.34378293e-01 -1.11438215e+00 5.08913457e-01
8.78405452e-01 3.69404644e-01 6.96824551e-01 6.62926376e-01
-6.29591644e-01 9.08198595e-01 -6.61599159e-01 -4.33637321e-01
1.18996665e-01 -5.24092793e-01 -1.18992674e+00 2.51157403e-01
3.86119157e-01 -2.43424878e-01 -1.16253994e-01 8.14582467e-01
5.35216808e-01 7.90576696e-01 4.23645258e-01 -7.55900323e-01
-8.14026117e-01 1.00885808e+00 -4.10638332e-01 5.42484581e-01
1.64346755e-01 -9.91018042e-02 4.88717735e-01 -1.28394949e+00
7.36382306e-01 1.28731918e+00 4.87523288e-01 9.21259820e-01
-9.90824163e-01 -4.92089987e-01 4.77742225e-01 5.57263717e-02
-1.21402264e+00 -6.20801926e-01 5.08662701e-01 -1.45895720e-01
1.48729050e+00 -4.81548637e-01 -1.98255345e-01 1.07080281e+00
-1.36945754e-01 1.11204183e+00 9.38566208e-01 -9.84771371e-01
2.66822189e-01 4.32637542e-01 -6.84520826e-02 4.35047865e-01
-1.47506446e-01 -5.87562658e-02 -3.72964293e-01 -4.17302996e-02
3.96999240e-01 -3.54646683e-01 1.82096735e-01 -9.25859809e-02
-9.08993542e-01 5.59723914e-01 -2.27349177e-01 5.91910303e-01
-1.80526420e-01 -1.74796075e-01 5.11917710e-01 7.82803297e-01
8.34201097e-01 7.11319983e-01 -1.05965614e+00 -3.03922236e-01
-9.45962369e-01 -1.10888980e-01 9.76262748e-01 1.13919723e+00
6.73036456e-01 2.45606333e-01 1.37969047e-01 1.21222997e+00
-2.42911190e-01 3.68274629e-01 1.23130870e+00 -8.24958920e-01
6.20514810e-01 2.86447585e-01 1.67375684e-01 2.48635665e-01
-2.26658002e-01 -1.98402494e-01 -2.39782646e-01 1.23924939e-02
2.55397946e-01 -7.90170193e-01 -1.03975320e+00 1.75385702e+00
2.67001897e-01 6.15081608e-01 4.55544949e-01 1.05361633e-01
5.71486831e-01 1.03043115e+00 6.48291290e-01 -5.77461243e-01
5.86888552e-01 -1.42543375e+00 -6.15269661e-01 -8.17478597e-01
1.00457537e+00 -7.17981756e-01 1.25513732e+00 6.47891939e-01
-1.29153001e+00 -1.04419088e+00 -7.83554077e-01 -1.15445759e-02
-5.89253426e-01 1.27880976e-01 2.48674601e-01 2.86788136e-01
-1.09309351e+00 5.59405744e-01 -7.03737080e-01 -4.05588984e-01
-2.51919609e-02 5.15034974e-01 -1.30564436e-01 -2.19515979e-01
-1.41351080e+00 1.30652511e+00 9.35623288e-01 -5.77267408e-01
-8.89658093e-01 -8.05806696e-01 -8.13400269e-01 -1.51105240e-01
7.08106995e-01 -2.34589010e-01 2.15023780e+00 -1.33126998e+00
-1.94571590e+00 8.06046069e-01 -2.92654574e-01 -4.77847964e-01
3.45694900e-01 -3.37423712e-01 -7.07976699e-01 -1.38131693e-01
3.49732526e-02 9.29530084e-01 7.46953428e-01 -1.04893327e+00
-9.43341076e-01 -2.44782306e-02 -4.45133224e-02 5.31564415e-01
-8.43784451e-01 3.12162172e-02 -5.97704887e-01 -5.79320312e-01
-4.11744595e-01 -8.98162365e-01 -4.00246382e-01 -8.48673940e-01
2.75064081e-01 -7.29104042e-01 9.07041073e-01 -5.63668549e-01
1.43052542e+00 -1.92935181e+00 -9.20888186e-02 -3.29834163e-01
-2.27017820e-01 8.31789434e-01 -7.19794095e-01 5.62107027e-01
2.36647520e-02 7.88017921e-03 -1.36390880e-01 -6.80129528e-01
-3.05447668e-01 2.72981226e-01 -3.70179355e-01 4.22503091e-02
2.88377643e-01 1.08850169e+00 -1.10739207e+00 -6.50221527e-01
6.43785819e-02 -1.78695153e-02 -3.31618756e-01 2.52043992e-01
-7.42980123e-01 5.35526812e-01 -5.06577551e-01 2.15952754e-01
3.27286512e-01 -2.95426399e-01 3.56832206e-01 6.74760699e-01
-3.84512693e-02 2.96622813e-01 -8.40459287e-01 2.21905017e+00
-7.78966069e-01 7.13458061e-01 -3.65520775e-01 -1.18434453e+00
1.08485854e+00 4.99502927e-01 4.18881804e-01 -7.14665115e-01
-5.47571182e-02 2.00177133e-01 -2.15404499e-02 -5.61866224e-01
7.10621893e-01 -5.86033762e-01 -2.06352785e-01 5.64577878e-01
4.90936279e-01 -8.60333666e-02 4.69590038e-01 2.65601933e-01
1.01050425e+00 4.13909703e-01 4.63327646e-01 -4.02257368e-02
6.54481232e-01 3.40935320e-01 2.44326904e-01 7.66965926e-01
-3.05383563e-01 1.33113369e-01 -1.47394612e-01 -1.77613348e-01
-1.10552859e+00 -1.10554707e+00 2.53209341e-02 1.86165416e+00
-2.54893571e-01 -7.58234113e-02 -5.52809954e-01 -1.02736175e+00
-1.95627898e-01 1.07020688e+00 -2.25267395e-01 -2.06610963e-01
-8.77599776e-01 -1.91910610e-01 5.64554513e-01 7.68287599e-01
3.41970682e-01 -1.51722467e+00 -2.39677340e-01 6.55574322e-01
-1.91621885e-01 -1.24306107e+00 -4.21614200e-01 4.27605897e-01
-1.11184156e+00 -5.67423522e-01 -8.88556004e-01 -1.47630167e+00
7.85380661e-01 2.20335070e-02 1.37897134e+00 -4.16483283e-01
1.33214980e-01 3.50087374e-01 -4.52048630e-01 -5.59909105e-01
-8.31225872e-01 3.78660828e-01 3.19641709e-01 -6.28524780e-01
9.05882239e-01 -4.37044442e-01 1.99504137e-01 8.19515139e-02
-8.61971617e-01 -2.45561883e-01 8.20163488e-01 8.36120844e-01
6.50973916e-01 2.30191663e-01 1.21216440e+00 -1.44674277e+00
9.99751210e-01 -6.53585255e-01 -4.89742219e-01 5.99938750e-01
-5.27287483e-01 1.45799622e-01 1.05522990e+00 -1.17319953e+00
-1.87783241e+00 2.51467198e-01 9.09343082e-03 -3.41044515e-01
-3.36636066e-01 8.08655024e-01 1.15743510e-01 1.84243247e-01
1.03595352e+00 1.59041613e-01 -6.59357727e-01 -7.80792713e-01
6.71862185e-01 7.47443795e-01 8.19815755e-01 -9.47581291e-01
4.55964923e-01 -4.89312224e-02 -5.61987817e-01 -7.27392614e-01
-1.24558651e+00 -8.58379126e-01 -1.08952105e+00 4.75585163e-02
2.83502758e-01 -9.44681227e-01 -7.31028095e-02 1.00291550e-01
-1.07081807e+00 -1.06966031e+00 -5.65412462e-01 4.45053190e-01
-5.31149268e-01 5.06024599e-01 -6.24927104e-01 -8.07869315e-01
-1.90897882e-01 -5.41740179e-01 6.96268022e-01 2.77006119e-01
-1.89702556e-01 -1.49993062e+00 5.53610802e-01 -1.79181248e-01
1.82017118e-01 -3.83621782e-01 6.01289451e-01 -1.19310880e+00
2.72912979e-01 -3.32566798e-01 3.07005644e-01 6.86337829e-01
4.66968445e-03 -5.96145809e-01 -9.70431685e-01 -5.22996366e-01
1.84715930e-02 -9.80539262e-01 6.61388636e-01 2.80506968e-01
1.02777064e+00 -7.70398900e-02 -5.47128379e-01 1.76637471e-02
1.39777112e+00 5.68441510e-01 2.74564564e-01 3.94190371e-01
1.16102539e-01 4.73190308e-01 8.83242786e-01 4.20372367e-01
-9.55234468e-02 1.83430046e-01 -4.71616507e-01 5.67442738e-03
-3.82464617e-01 -5.50774217e-01 7.78636158e-01 9.58510995e-01
3.91280830e-01 -5.74801803e-01 -1.15302813e+00 7.66048014e-01
-1.72148144e+00 -7.45593429e-01 5.00017285e-01 1.98334253e+00
1.25562036e+00 4.83462006e-01 -8.77536833e-02 -5.09663343e-01
6.99968040e-01 -2.09721267e-01 -9.96670902e-01 -5.26089430e-01
6.52328655e-02 5.88489532e-01 2.86978185e-01 7.25205064e-01
-1.06983280e+00 1.68472683e+00 7.13354588e+00 8.90892863e-01
-9.40717041e-01 4.97763097e-01 4.98900563e-01 -9.05738473e-02
2.19461709e-01 -1.81298293e-02 -1.37877750e+00 9.84673649e-02
1.55910349e+00 -6.03084683e-01 2.29440674e-01 1.28359854e+00
8.96515474e-02 -8.24886113e-02 -1.26853573e+00 4.29566056e-01
3.19322526e-01 -8.48805785e-01 -1.90690570e-02 -3.36513251e-01
1.26895344e+00 4.67509806e-01 2.27984805e-02 1.18869591e+00
9.68767226e-01 -5.84581673e-01 2.70872563e-01 1.20431006e-01
1.10736275e+00 -7.78476357e-01 5.70739031e-01 9.72965479e-01
-8.87288868e-01 -1.76201448e-01 -7.45364785e-01 -9.25299153e-02
4.80888784e-01 1.33856788e-01 -1.35656357e+00 1.38891518e-01
1.99340269e-01 5.97814262e-01 -3.77248406e-01 8.51820946e-01
-3.08637023e-01 8.09137404e-01 -1.44263789e-01 4.86755185e-03
4.86991197e-01 2.07846925e-01 2.32322782e-01 1.54044616e+00
2.98166513e-01 2.79515684e-01 7.49102294e-01 3.13474089e-01
-4.10219878e-01 1.96550503e-01 -7.08313763e-01 -3.15757513e-01
5.79554737e-01 8.52675021e-01 -4.56866711e-01 -9.98841047e-01
-6.98096216e-01 1.18967509e+00 6.99977636e-01 6.14826798e-01
-5.32425880e-01 -3.17840904e-01 -2.48332638e-02 -1.41243905e-01
2.72446126e-01 -2.19014257e-01 3.37994844e-02 -1.18414843e+00
-2.29833171e-01 -8.04852247e-01 6.51309133e-01 -7.96178401e-01
-1.54909313e+00 7.98505843e-01 1.57995194e-01 -1.12081516e+00
-4.93946314e-01 -6.49800003e-01 -3.51991355e-01 7.83093333e-01
-1.66611791e+00 -1.00914168e+00 2.80202180e-01 6.91987455e-01
1.51678193e+00 -4.99209017e-01 8.94616306e-01 2.18572646e-01
-2.73488313e-01 5.37794173e-01 5.22021949e-01 -4.40429896e-02
1.01504290e+00 -1.18431389e+00 4.87266958e-01 9.94708002e-01
3.42607915e-01 3.88758302e-01 5.96978068e-01 -9.09169197e-01
-7.85497487e-01 -1.17882800e+00 1.29236555e+00 -7.02063680e-01
7.55012155e-01 -1.38244405e-01 -9.08779621e-01 1.16691780e+00
3.58465105e-01 -2.79141366e-01 6.48453236e-01 2.25764245e-01
-3.96439992e-02 2.56115079e-01 -8.82142723e-01 4.24826354e-01
9.21812654e-01 -4.15127099e-01 -1.23299098e+00 6.47486866e-01
9.28465962e-01 -4.78524238e-01 -5.28457046e-01 3.79693508e-01
-2.48863287e-02 -1.33989856e-01 9.02113080e-01 -1.09886754e+00
3.07311893e-01 1.87376082e-01 2.68090218e-01 -1.50654209e+00
-3.07681978e-01 -6.05983078e-01 -3.13165277e-01 1.22186399e+00
7.09714651e-01 -3.07953566e-01 8.84497762e-01 5.86790621e-01
-4.66317117e-01 -3.22663963e-01 -5.00142097e-01 -1.33058262e+00
7.28773892e-01 -5.15340686e-01 1.34912103e-01 9.70759988e-01
4.78224933e-01 7.62789369e-01 -4.43858325e-01 -2.10233182e-02
9.38173905e-02 -1.41919792e-01 6.11693859e-01 -1.19473243e+00
-3.70631009e-01 -5.09623699e-02 4.20822948e-01 -1.58452368e+00
6.99543834e-01 -1.04794586e+00 4.60044205e-01 -1.38090599e+00
1.00999117e-01 -4.24782902e-01 -5.39402723e-01 7.26071894e-01
-1.71656221e-01 -1.52282834e-01 1.16894543e-02 1.44789025e-01
-1.27297580e+00 3.13384175e-01 1.28820813e+00 1.34464623e-02
-7.83619940e-01 4.38012443e-02 -4.88885403e-01 7.99571037e-01
8.94107938e-01 -6.29857004e-01 -9.87189293e-01 -7.34781802e-01
-1.60504222e-01 -9.74329337e-02 -6.21836662e-01 -9.32215035e-01
5.87507069e-01 -2.74173737e-01 3.52203190e-01 -6.27442300e-01
3.16573769e-01 -6.60243869e-01 -3.96353632e-01 3.39581072e-01
-7.22729445e-01 1.18157648e-01 8.20009708e-01 4.43563879e-01
-3.49657118e-01 -7.23281085e-01 8.95922363e-01 -5.86284757e-01
-1.52330136e+00 1.48520708e-01 -6.28525496e-01 5.40216923e-01
1.18483937e+00 1.94419338e-03 2.29120161e-02 -2.25963488e-01
-1.09669578e+00 3.50848049e-01 8.26746896e-02 5.25716782e-01
4.65251625e-01 -1.03606164e+00 -7.53203809e-01 -6.08934201e-02
-5.02229622e-03 3.70848272e-03 6.13352470e-02 2.28003219e-01
1.42094996e-02 6.96448505e-01 7.64838308e-02 -3.04386616e-01
-1.12058318e+00 9.49087083e-01 1.28652811e-01 -8.94817889e-01
-4.42560345e-01 1.17749298e+00 2.42799059e-01 -6.38176024e-01
4.32451129e-01 2.48946305e-02 -2.00815380e-01 -9.67280045e-02
7.59032965e-01 2.52605885e-01 8.15085787e-03 -5.24490140e-04
5.91172241e-02 1.94777206e-01 -4.52906013e-01 -5.19016683e-01
1.41607630e+00 -4.74440783e-01 3.93380433e-01 6.63767695e-01
1.05719340e+00 -6.24575764e-02 -1.25055134e+00 -9.09129858e-01
4.42363262e-01 -1.34932458e-01 -1.87894925e-01 -1.20950043e+00
-3.32492322e-01 8.97598267e-01 2.28075743e-01 -2.55065620e-01
1.05092824e+00 2.26937473e-01 8.81264329e-01 1.08609998e+00
4.92097348e-01 -1.58369279e+00 5.31229138e-01 9.54486012e-01
5.18803775e-01 -1.32561815e+00 -1.47404522e-01 1.29417181e-01
-9.86378193e-01 1.08098412e+00 1.26138592e+00 -2.35633925e-02
5.94259143e-01 3.84054244e-01 9.18379202e-02 2.52741843e-01
-1.02349961e+00 -2.23757461e-01 8.09228867e-02 6.28864229e-01
6.14424884e-01 -1.53221592e-01 -3.67818594e-01 7.35562205e-01
9.93329957e-02 5.36552787e-01 2.14772210e-01 1.11638975e+00
-9.86731231e-01 -1.55136418e+00 -4.81813997e-02 1.99130669e-01
-3.91710281e-01 -3.64203393e-01 -4.98621553e-01 6.67529464e-01
1.97485238e-01 9.38066065e-01 -4.17537764e-02 1.88091025e-02
3.44012469e-01 6.93732083e-01 3.69083464e-01 -1.29579997e+00
-3.85812908e-01 2.32778370e-01 2.28572786e-01 1.08781997e-02
-2.95096934e-01 -5.42960763e-01 -1.57256091e+00 3.28820124e-02
-3.65772307e-01 4.49324667e-01 4.07665908e-01 1.02535295e+00
1.65702939e-01 2.98668921e-01 3.71695787e-01 -5.01007497e-01
-8.23143125e-01 -1.32979560e+00 -3.71386677e-01 3.47905487e-01
7.16123432e-02 -4.66766685e-01 -1.32407531e-01 6.27532721e-01] | [10.664831161499023, 8.225811004638672] |
c8ca8d48-e1b4-4474-9440-67e73c1fc9da | strengths-and-weaknesses-of-3d-pose | 2306.06117 | null | https://arxiv.org/abs/2306.06117v1 | https://arxiv.org/pdf/2306.06117v1.pdf | Strengths and Weaknesses of 3D Pose Estimation and Inertial Motion Capture System for Movement Therapy | 3D pose estimation offers the opportunity for fast, non-invasive, and accurate motion analysis. This is of special interest also for clinical use. Currently, motion capture systems are used, as they offer robust and precise data acquisition, which is essential in the case of clinical applications. In this study, we investigate the accuracy of the state-of-the-art 3D position estimation approach MeTrabs, compared to the established inertial sensor system MTw Awinda for specific motion exercises. The study uses and provides an evaluation dataset of parallel recordings from 10 subjects during various movement therapy exercises. The information from the Awinda system and the frames for monocular pose estimation are synchronized. For the comparison, clinically relevant parameters for joint angles of ankle, knee, back, and elbow flexion-extension were estimated and evaluated using mean, median, and maximum deviation between the calculated joint angles for the different exercises, camera positions, and clothing items. The results of the analysis indicate that the mean and median deviations can be kept below 5{\deg} for some of the studied angles. These joints could be considered for medical applications even considering the maximum deviations of 15{\deg}. However, caution should be applied to certain particularly problematic joints. In particular, elbow flexions, which showed high maximum deviations of up to 50{\deg} in our analysis. Furthermore, the type of exercise plays a crucial role in the reliable and safe application of the 3D position estimation method. For example, all joint angles showed a significant deterioration in performance during exercises near the ground. | ['Ted Preuß', 'Hannah Siebers', 'Shawan Mohammed'] | 2023-06-01 | null | null | null | null | ['pose-estimation', '3d-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-1.03740081e-01 -6.24096580e-02 -2.59704530e-01 2.93967783e-01
-4.43277746e-01 -3.58205378e-01 1.61060542e-01 1.80633277e-01
-9.01096761e-01 7.48512208e-01 1.09280668e-01 2.37939179e-01
-5.93492389e-01 -2.69438535e-01 -2.85468221e-01 -6.21025980e-01
-4.97868866e-01 5.39603055e-01 3.53567153e-01 -2.50334859e-01
4.39667329e-03 7.88956404e-01 -1.47526896e+00 -1.82486653e-01
5.27026892e-01 6.16037548e-01 4.61807013e-01 4.41563547e-01
5.55954039e-01 1.78153560e-01 -8.07395518e-01 -2.40998909e-01
4.49305587e-02 -1.57256365e-01 -2.12825835e-01 2.86840171e-01
1.47383630e-01 -6.17196381e-01 7.18583465e-02 5.08131504e-01
9.48738933e-01 2.47730434e-01 3.30634773e-01 -8.07414353e-01
2.91511744e-01 1.15015328e-01 -4.05596316e-01 1.06416851e-01
7.51890004e-01 1.58162683e-01 3.10025752e-01 -6.87040865e-01
1.09217155e+00 5.15918612e-01 6.30111158e-01 3.78336102e-01
-1.02189004e+00 -3.84943604e-01 -3.89824420e-01 3.72676492e-01
-1.50675309e+00 -2.37488076e-01 9.91580784e-01 -6.56578362e-01
7.10197568e-01 4.64174926e-01 1.17331803e+00 9.52441394e-01
7.77220845e-01 3.35498989e-01 1.05116034e+00 -4.74937171e-01
3.18405867e-01 -1.25382207e-02 -9.76750404e-02 -1.53376702e-02
7.87808776e-01 -2.51415104e-01 -3.88956845e-01 1.62634239e-01
8.58773708e-01 -8.82884935e-02 -5.73037326e-01 -4.89495695e-01
-1.27902043e+00 3.19658935e-01 5.18196672e-02 6.80889070e-01
-6.53774321e-01 7.23928362e-02 4.53141183e-01 -1.76845387e-01
2.80193955e-01 4.11539644e-01 -3.93736809e-01 -9.01861608e-01
-6.96683407e-01 3.32946777e-01 4.94220614e-01 4.13113624e-01
-9.46364552e-02 -6.53881282e-02 7.23506883e-02 5.68213522e-01
1.66421473e-01 2.91914105e-01 5.41469514e-01 -7.81470358e-01
4.86997575e-01 5.76404214e-01 1.93946198e-01 -1.01967514e+00
-8.90154243e-01 -4.87093478e-01 -3.76696527e-01 2.11328551e-01
7.36676335e-01 -2.26308882e-01 -4.22541469e-01 1.38809526e+00
6.33588910e-01 -3.81585777e-01 -1.80788755e-01 1.27328265e+00
2.83743650e-01 -3.00407354e-02 2.17386112e-02 -5.75778782e-01
1.58810663e+00 -1.52307570e-01 -9.61528301e-01 -2.07131684e-01
6.53847396e-01 -1.14710999e+00 8.46052587e-01 5.81254244e-01
-1.16638505e+00 -5.83658695e-01 -9.88521516e-01 3.98013204e-01
1.57384872e-01 5.60330331e-01 1.33773580e-01 6.80050731e-01
-3.36282760e-01 9.14415896e-01 -1.16319203e+00 -4.54482168e-01
-4.86392617e-01 4.68645781e-01 -6.75191045e-01 2.74678618e-01
-1.06826234e+00 1.30725884e+00 2.40755051e-01 8.20876658e-01
8.84861350e-02 -2.82862097e-01 -5.95326900e-01 -3.23004693e-01
4.43742573e-01 -4.56012428e-01 8.12967658e-01 -2.08132297e-01
-1.75030959e+00 9.30819452e-01 2.90028036e-01 -6.77723158e-03
9.36807632e-01 -9.85450506e-01 -3.55711192e-01 4.61306751e-01
-2.17992738e-01 -1.37215838e-01 4.24595505e-01 -6.39806807e-01
2.36631468e-01 -6.23598516e-01 -1.78315371e-01 3.07875782e-01
-1.46427721e-01 -1.32908836e-01 -4.06988502e-01 -6.48086965e-01
4.83343989e-01 -1.26449215e+00 -1.39577851e-01 2.15769455e-01
-2.40865275e-01 1.38597965e-01 4.22791302e-01 -8.54821026e-01
1.10770428e+00 -2.09708428e+00 3.51002246e-01 3.24465722e-01
-7.72099569e-02 2.27757499e-01 6.49690628e-01 5.37505329e-01
1.44832144e-02 -4.55940276e-01 2.16176301e-01 2.05271676e-01
-3.93021554e-01 1.89573690e-01 4.69623893e-01 9.16823328e-01
-4.26128596e-01 1.95751220e-01 -6.09271407e-01 -4.02697414e-01
6.63854003e-01 5.28849721e-01 -2.22706065e-01 2.07830861e-01
2.90118486e-01 6.13098860e-01 -2.95353502e-01 4.12751585e-01
2.57060260e-01 3.51206213e-01 5.66008866e-01 -6.34824991e-01
-1.58251956e-01 -1.38783902e-01 -1.53796554e+00 1.60582924e+00
-2.61937439e-01 5.14399767e-01 -5.04945703e-02 -6.02533698e-01
1.06110013e+00 8.29309523e-01 1.04001689e+00 -3.75643432e-01
3.78481776e-01 6.84250355e-01 3.54024351e-01 -8.91231716e-01
4.48829591e-01 -3.70791219e-02 1.89190462e-01 4.03068662e-02
-2.75055140e-01 -1.92388982e-01 2.43842140e-01 -3.72862697e-01
4.04085457e-01 4.09130394e-01 6.25373363e-01 -2.70487875e-01
5.20413697e-01 -2.07239121e-01 2.46288642e-01 -5.06446175e-02
-6.33442448e-03 6.93375945e-01 4.35493290e-01 -1.75445333e-01
-7.26585388e-01 -9.99462724e-01 -1.85908422e-01 6.07137159e-02
1.96764860e-02 -4.56152946e-01 -7.47672081e-01 2.44744480e-01
4.61943597e-02 2.74773449e-01 -1.26411602e-01 -3.22481811e-01
-7.85365105e-01 -6.05161488e-01 1.79239601e-01 5.95560372e-01
1.46434739e-01 -7.50292659e-01 -1.24159408e+00 3.68507445e-01
-3.11827451e-01 -1.23429525e+00 5.30995615e-02 -1.95032477e-01
-1.30524719e+00 -1.32307410e+00 -1.05760026e+00 -1.82689995e-01
3.43008190e-01 -8.52073058e-02 6.62142098e-01 -8.08904320e-02
-2.84338206e-01 5.56040943e-01 -3.49267006e-01 -9.12272185e-02
-2.33998492e-01 6.81414502e-03 3.81811738e-01 -4.07331407e-01
4.42730216e-03 -4.71568167e-01 -8.29570949e-01 6.06724977e-01
-5.17008901e-01 -1.87239662e-01 4.79516268e-01 2.76492208e-01
5.38439631e-01 -3.64364684e-01 2.16884673e-01 -2.45943755e-01
7.17814565e-01 1.46390861e-02 -2.80274004e-01 -1.11757167e-01
-2.71438330e-01 -4.01350379e-01 1.87022313e-01 -4.64823812e-01
-7.71519899e-01 4.60940637e-02 -3.49114358e-01 -1.22007295e-01
-2.64459074e-01 4.61982906e-01 -1.34587243e-01 1.13223992e-01
7.48830616e-01 -2.57597297e-01 4.53617275e-01 -5.64726889e-01
-1.22255705e-01 5.32618880e-01 6.08055055e-01 -5.58918536e-01
3.23691547e-01 2.35093102e-01 4.80923146e-01 -1.15799940e+00
2.78977323e-02 -5.42200923e-01 -8.54533672e-01 -8.93386245e-01
8.31970394e-01 -5.95610678e-01 -1.07541919e+00 6.33725584e-01
-8.68511617e-01 -4.32567038e-02 -3.16342890e-01 1.52764356e+00
-8.73329699e-01 7.62156665e-01 -5.40734351e-01 -9.15423393e-01
-4.38003093e-01 -1.31140757e+00 9.58652794e-01 1.27323042e-03
-9.78337348e-01 -7.39411473e-01 -5.02528474e-02 4.22463387e-01
8.00534040e-02 8.73012900e-01 4.59121525e-01 -1.71103105e-01
1.59914158e-02 -6.80401802e-01 6.14412546e-01 1.24554597e-01
3.18438709e-01 4.45915833e-02 -4.31161791e-01 -4.25501198e-01
1.49983570e-01 -7.54331648e-02 -1.01342402e-01 5.91010511e-01
6.00181222e-01 2.01749906e-01 -1.97833374e-01 3.39167789e-02
1.32131910e+00 3.44898105e-02 7.54691184e-01 5.49902320e-01
3.94177109e-01 8.91878605e-01 1.04927707e+00 5.75461805e-01
-3.02070916e-01 1.36853182e+00 5.08093715e-01 2.20897391e-01
-1.56465713e-02 2.91056484e-01 3.83557349e-01 7.97562182e-01
-9.35160518e-01 2.51194090e-01 -6.84355557e-01 3.87257457e-01
-1.43817806e+00 -5.63398182e-01 -5.67355692e-01 2.69886565e+00
5.99690378e-01 2.74344295e-01 2.00133890e-01 7.28883743e-01
5.59614956e-01 3.94610167e-02 -1.89874619e-01 -3.57047200e-01
3.26326787e-01 2.54804462e-01 4.04459804e-01 1.78577438e-01
-4.07826513e-01 4.63768803e-02 5.53077555e+00 3.11005503e-01
-1.33173978e+00 -1.51698723e-01 -4.34742063e-01 -4.07869160e-01
1.46118328e-01 -1.78843990e-01 -4.32878941e-01 5.87347031e-01
9.66996133e-01 -4.07608524e-02 -3.74730915e-01 7.80327082e-01
6.58337057e-01 -7.49235928e-01 -7.35595286e-01 6.95228815e-01
-1.06144466e-01 -8.43703508e-01 -7.35982716e-01 1.75461069e-01
2.44461536e-01 -2.49737397e-01 -3.52397919e-01 -4.31482166e-01
-8.61930609e-01 -2.20349789e-01 6.11243606e-01 6.77977920e-01
6.28290832e-01 -6.34960353e-01 9.14263129e-01 3.82643968e-01
-1.07998061e+00 2.44164735e-01 -1.41044885e-01 -2.13100687e-01
6.83838248e-01 8.92107368e-01 -4.66709942e-01 8.36883068e-01
2.82835215e-01 3.65085930e-01 -2.07580209e-01 1.05048895e+00
-5.68722308e-01 2.41201162e-01 -5.91019809e-01 -2.69049034e-02
-1.79616347e-01 -2.73142487e-01 8.48936617e-01 5.23324668e-01
5.02344728e-01 -4.99636717e-02 -7.83632472e-02 1.84091970e-01
5.76482415e-01 3.47069561e-01 -4.75050539e-01 2.63922513e-01
2.04791531e-01 1.13112891e+00 -7.85321474e-01 9.20777470e-02
-1.27641901e-01 7.47282207e-01 -3.63232881e-01 2.18212098e-01
-7.17584729e-01 -1.93662465e-01 5.79083204e-01 6.35624349e-01
-3.09855461e-01 -6.69006288e-01 -8.06934014e-02 -9.02782917e-01
6.57355249e-01 -5.30896246e-01 3.56051534e-01 -8.42376530e-01
-4.85292733e-01 1.53591692e-01 6.65755630e-01 -1.69150710e+00
-9.06812906e-01 -8.72070789e-01 -2.54501075e-01 5.75060248e-01
-5.79207778e-01 -5.23425996e-01 -3.21186185e-01 4.62577850e-01
4.02273178e-01 4.61862892e-01 6.98133051e-01 3.25849146e-01
-3.62139523e-01 1.91193089e-01 -4.61517908e-02 -2.24198923e-01
7.38682389e-01 -7.74994552e-01 -2.34406054e-01 5.36794066e-01
-1.59662560e-01 7.58661449e-01 1.07009518e+00 -6.53258622e-01
-1.44015825e+00 -1.61863789e-01 7.81896830e-01 -1.21114932e-01
3.86744976e-01 8.06430448e-03 -6.04135334e-01 4.84105587e-01
-2.71054238e-01 -8.47088993e-02 6.01282239e-01 -1.67321548e-01
5.92815042e-01 -1.28487781e-01 -1.02270889e+00 6.38838112e-01
6.58293724e-01 -1.57553330e-01 -7.11708546e-01 4.17443603e-01
-5.22211902e-02 -9.14015353e-01 -1.49550748e+00 4.71641570e-01
1.40249264e+00 -8.92062485e-01 1.05143797e+00 -1.08803049e-01
1.76907748e-01 -2.49412939e-01 1.82261616e-01 -1.07704544e+00
1.80613950e-01 -3.81610125e-01 -5.78808710e-02 8.53827655e-01
-1.00295626e-01 -5.67476392e-01 8.75391364e-01 5.70844293e-01
8.02252814e-02 -7.29833782e-01 -1.34908080e+00 -8.34266245e-01
-3.24463636e-01 -4.98606890e-01 -1.16028339e-01 5.02312779e-01
4.37618136e-01 -3.00692260e-01 -6.03942215e-01 -3.45156603e-02
5.01927018e-01 -1.17629112e-04 9.90025938e-01 -1.11710739e+00
-3.89593273e-01 3.57032120e-02 -1.02749527e+00 -3.85741115e-01
-4.25806314e-01 -1.17783941e-01 -3.17633897e-01 -1.49504590e+00
-3.88121158e-01 1.56396523e-01 6.57628551e-02 -1.97846472e-01
-9.81714483e-03 2.30373248e-01 3.38535696e-01 3.42515409e-01
4.27281469e-01 1.29515797e-01 1.38754272e+00 4.79183882e-01
-4.22957838e-01 4.99923646e-01 2.18672454e-01 7.18674302e-01
6.45164788e-01 -3.98857147e-01 -4.64366525e-01 -1.09546386e-01
2.64589339e-01 1.52090833e-01 2.34623715e-01 -1.29165232e+00
-1.83968563e-02 1.69115886e-02 4.43583161e-01 -6.78781807e-01
5.13905525e-01 -1.27858078e+00 1.05649781e+00 1.10301220e+00
2.35352904e-01 7.70731792e-02 9.31980982e-02 -6.05295822e-02
-1.17343917e-01 -2.32204765e-01 4.22661334e-01 -1.42322257e-01
-5.36726058e-01 -3.70556504e-01 -6.65428460e-01 -3.76295835e-01
1.18084359e+00 -9.59965169e-01 -6.13942929e-02 -3.63142848e-01
-1.34507430e+00 -3.07179540e-01 3.92011762e-01 3.62461090e-01
4.09528524e-01 -1.16184735e+00 -2.86419451e-01 1.19074777e-01
1.59558684e-01 -5.29406928e-02 6.39292419e-01 1.61891615e+00
-7.38340676e-01 3.28398257e-01 -6.71637535e-01 -8.70431244e-01
-1.44020903e+00 1.00748695e-01 1.75670058e-01 -2.17293009e-01
-6.72114253e-01 -9.41821933e-02 -6.01089835e-01 2.49449074e-01
-1.47425070e-01 -5.41139424e-01 -3.92395467e-01 3.01619619e-01
8.58143568e-02 1.09077084e+00 4.87781614e-01 -6.49667799e-01
-5.40913939e-01 1.14079237e+00 3.95403683e-01 -1.53051943e-01
1.02328968e+00 -2.94491202e-01 2.17932478e-01 4.69471067e-01
7.70453513e-01 3.81581515e-01 -1.06468439e+00 4.06574488e-01
-1.10679217e-01 -6.27511382e-01 -4.89235103e-01 -4.00801033e-01
-8.47473145e-01 7.27156460e-01 8.39870036e-01 -2.05970019e-01
1.16928101e+00 -3.55833173e-02 4.24024791e-01 -2.71317270e-02
6.19294584e-01 -1.22024500e+00 -2.61661381e-01 -3.14173728e-01
1.07836342e+00 -4.49682921e-01 6.21138632e-01 -8.05279613e-01
-3.03700358e-01 1.34769940e+00 4.00176674e-01 -1.50254086e-01
3.91446799e-01 1.54478386e-01 1.96198091e-01 -5.81987128e-02
-4.62718308e-02 -3.47328559e-02 4.98692453e-01 4.91648734e-01
8.70388985e-01 2.01421365e-01 -1.45640993e+00 1.27027959e-01
-1.32424846e-01 4.74917382e-01 6.17860258e-01 1.30173337e+00
-2.78279692e-01 -9.85012770e-01 -8.03644419e-01 1.90364234e-02
-6.25183403e-01 7.91973293e-01 -6.30713776e-02 1.67799425e+00
-3.67875732e-02 7.13335395e-01 -9.06678960e-02 -2.73524582e-01
9.02645648e-01 -2.34542750e-02 8.19075525e-01 -2.93129712e-01
-7.08754957e-01 3.91192049e-01 3.64501506e-01 -7.95246303e-01
-7.22729683e-01 -8.26382399e-01 -1.24195445e+00 -1.36468410e-01
-4.93664563e-01 6.33427873e-02 1.04268634e+00 9.20029104e-01
-8.16689953e-02 3.76459032e-01 2.13352218e-01 -9.40622211e-01
-5.51293910e-01 -1.09386253e+00 -8.68638694e-01 5.78966677e-01
-2.13723868e-01 -1.14101219e+00 -1.65345341e-01 -2.01257393e-01] | [7.010500431060791, 0.1575242429971695] |
a78a5fd9-9bc6-4bee-b790-260176b259d3 | a-unified-prompt-guided-in-context-inpainting | 2305.11577 | null | https://arxiv.org/abs/2305.11577v1 | https://arxiv.org/pdf/2305.11577v1.pdf | A Unified Prompt-Guided In-Context Inpainting Framework for Reference-based Image Manipulations | Recent advancements in Text-to-Image (T2I) generative models have yielded impressive results in generating high-fidelity images based on consistent text prompts. However, there is a growing interest in exploring the potential of these models for more diverse reference-based image manipulation tasks that require spatial understanding and visual context. Previous approaches have achieved this by incorporating additional control modules or fine-tuning the generative models specifically for each task until convergence. In this paper, we propose a different perspective. We conjecture that current large-scale T2I generative models already possess the capability to perform these tasks but are not fully activated within the standard generation process. To unlock these capabilities, we introduce a unified Prompt-Guided In-Context inpainting (PGIC) framework, which leverages large-scale T2I models to re-formulate and solve reference-guided image manipulations. In the PGIC framework, the reference and masked target are stitched together as a new input for the generative models, enabling the filling of masked regions as producing final results. Furthermore, we demonstrate that the self-attention modules in T2I models are well-suited for establishing spatial correlations and efficiently addressing challenging reference-guided manipulations. These large T2I models can be effectively driven by task-specific prompts with minimal training cost or even with frozen backbones. We synthetically evaluate the effectiveness of the proposed PGIC framework across various tasks, including reference-guided image inpainting, faithful inpainting, outpainting, local super-resolution, and novel view synthesis. Our results show that PGIC achieves significantly better performance while requiring less computation compared to other fine-tuning based approaches. | ['Yanwei Fu', 'Yunuo Cai', 'Yikai Wang', 'Qiaole Dong', 'Chenjie Cao'] | 2023-05-19 | null | null | null | null | ['image-manipulation', 'image-inpainting', 'novel-view-synthesis'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 5.36378264e-01 7.00151399e-02 5.25985435e-02 -1.64308727e-01
-7.99061358e-01 -4.79206979e-01 9.49311137e-01 -6.22786403e-01
1.98289216e-01 7.01364934e-01 4.47949678e-01 1.69484578e-02
-5.40973954e-02 -6.29324913e-01 -9.01220977e-01 -6.10686421e-01
4.24089283e-01 3.31026137e-01 2.95501165e-02 -2.89764255e-01
3.59690189e-01 5.17962158e-01 -1.47894156e+00 4.86184657e-01
1.12359965e+00 7.53285348e-01 6.62388861e-01 7.33265340e-01
-2.16908809e-02 7.32013166e-01 -7.51108408e-01 -2.81647772e-01
2.66352564e-01 -5.61950684e-01 -5.76207101e-01 3.51108164e-01
6.47759497e-01 -5.60783327e-01 -3.04134369e-01 6.32232487e-01
5.68796873e-01 2.09745482e-01 5.14144480e-01 -1.10413873e+00
-9.46518540e-01 2.96299011e-01 -8.55054855e-01 -7.45615214e-02
4.19661969e-01 3.64733934e-01 6.95638418e-01 -1.10940301e+00
9.37721074e-01 1.42684507e+00 3.97406310e-01 5.94119728e-01
-1.40704298e+00 -5.98672211e-01 2.94052124e-01 5.13864495e-03
-1.23243463e+00 -5.40380776e-01 9.30254161e-01 -3.12718242e-01
8.64881158e-01 3.59943926e-01 5.40720046e-01 1.25637412e+00
2.87534416e-01 7.91984320e-01 1.16441560e+00 -4.47126001e-01
-2.48489473e-02 -1.97266370e-01 -4.70249295e-01 4.99694377e-01
-1.72649011e-01 4.12091643e-01 -9.02467966e-01 1.77968487e-01
1.31884170e+00 1.50322750e-01 -4.19707000e-01 -1.86863780e-01
-1.45852828e+00 6.54804587e-01 3.73366684e-01 8.56128410e-02
-4.78214651e-01 2.60785639e-01 8.71560797e-02 1.46151185e-01
6.96315348e-01 6.30858898e-01 -1.66145712e-01 1.44826829e-01
-1.44438279e+00 5.97701073e-01 3.03643376e-01 1.27868950e+00
7.53430486e-01 3.20630670e-01 -7.54005969e-01 8.14363837e-01
6.58572838e-02 4.44647729e-01 2.72420764e-01 -1.22417688e+00
4.53088611e-01 2.78075665e-01 2.39330441e-01 -9.79459584e-01
-3.18144006e-03 -6.18398070e-01 -9.20116246e-01 2.94237047e-01
-6.51369393e-02 1.07423171e-01 -1.23187530e+00 1.89198887e+00
3.34514707e-01 2.66523778e-01 -1.97843418e-01 8.39969516e-01
7.01536894e-01 7.74863660e-01 -1.37241393e-01 -9.27838311e-02
1.10172176e+00 -1.34217417e+00 -7.59285390e-01 -2.89093286e-01
1.60309419e-01 -9.72532749e-01 1.19455051e+00 2.45645657e-01
-1.50329685e+00 -8.17566574e-01 -8.71061981e-01 -4.03332740e-01
6.06923588e-02 8.20977464e-02 5.29462159e-01 1.78836361e-01
-1.33543921e+00 4.92593169e-01 -6.24880016e-01 -1.92729831e-01
4.06624883e-01 1.05193695e-02 -3.32378179e-01 -3.02502096e-01
-8.57822478e-01 7.28556275e-01 1.93233103e-01 2.36042038e-01
-1.26278269e+00 -9.31221187e-01 -9.44980919e-01 2.09720284e-02
4.62624997e-01 -1.13833570e+00 9.73594487e-01 -7.95265555e-01
-1.50580144e+00 7.24177122e-01 -3.94178629e-01 -2.60669380e-01
4.86891270e-01 -2.99000859e-01 1.18911285e-02 8.33861381e-02
3.73531252e-01 1.23899043e+00 1.44896901e+00 -1.62459052e+00
-3.07955831e-01 -8.49789232e-02 4.54325452e-02 3.49461168e-01
5.14569320e-02 -1.70040131e-02 -6.49441421e-01 -1.18158984e+00
5.96394204e-02 -8.80108058e-01 -1.78350195e-01 3.34413275e-02
-5.64695835e-01 2.16676548e-01 1.20597601e+00 -6.14396334e-01
1.22106898e+00 -1.98806930e+00 5.03807247e-01 -2.69418269e-01
3.10285300e-01 3.50934505e-01 -4.27215129e-01 5.37720382e-01
-8.93657953e-02 1.94680914e-01 -1.58989877e-01 -7.76480317e-01
-1.16471365e-01 4.78761971e-01 -6.89381063e-01 -1.45558879e-01
4.27546561e-01 1.36466575e+00 -7.28135884e-01 -4.68912154e-01
4.48302954e-01 6.37897551e-01 -9.32414412e-01 5.06875515e-01
-4.45356250e-01 9.80009198e-01 -2.45180324e-01 5.91437638e-01
6.70252562e-01 -3.35064203e-01 -6.11437075e-02 -5.65576553e-01
-7.44393915e-02 5.91376349e-02 -9.21116173e-01 2.07848787e+00
-6.54070437e-01 4.57316548e-01 3.24201614e-01 -8.53574693e-01
7.70884752e-01 7.01968595e-02 3.13004076e-01 -7.16700137e-01
-2.00735316e-01 -3.72793712e-02 -3.47022086e-01 -2.50983149e-01
7.31870830e-01 -3.76575775e-02 7.65545741e-02 4.08350289e-01
8.32592770e-02 -4.97983903e-01 2.04491869e-01 3.74802530e-01
7.24453986e-01 7.96307445e-01 2.96966936e-02 -1.02448254e-03
2.65615791e-01 -1.17722943e-01 3.64174932e-01 8.40436935e-01
2.14209497e-01 1.19159317e+00 1.48818731e-01 -2.77554333e-01
-1.22956455e+00 -1.10995567e+00 2.13278115e-01 9.45569336e-01
1.36784345e-01 -4.37535137e-01 -7.57151067e-01 -2.65819162e-01
-2.79638261e-01 7.79006541e-01 -6.93876088e-01 -4.56109717e-02
-6.73467815e-01 -4.10828978e-01 2.34659165e-01 5.78096271e-01
7.34553754e-01 -1.20705581e+00 -5.46392918e-01 2.75232792e-01
-4.27157551e-01 -1.09122145e+00 -8.69112253e-01 3.75925377e-02
-9.03477848e-01 -6.55012846e-01 -9.49245036e-01 -5.79517066e-01
7.22882330e-01 6.76673472e-01 1.29868877e+00 8.47710073e-02
-2.77830750e-01 4.18879986e-01 -2.47507274e-01 -1.00123227e-01
-4.35858399e-01 -2.22857688e-02 -3.94044995e-01 -5.23943752e-02
-5.20065427e-01 -8.65613997e-01 -8.96926403e-01 4.24149156e-01
-1.33764660e+00 7.17893362e-01 6.34217322e-01 1.12594521e+00
7.65928566e-01 -2.89669991e-01 4.73178267e-01 -7.15906560e-01
5.79494953e-01 -2.26559013e-01 -2.91974694e-01 2.69543797e-01
-4.44236785e-01 1.48159370e-01 5.21336377e-01 -4.94736969e-01
-1.43063533e+00 -2.43004948e-01 -7.94403777e-02 -9.00706708e-01
1.19493101e-02 2.70572871e-01 -2.68245876e-01 -8.09971690e-02
4.26411867e-01 4.32974368e-01 -5.42894118e-02 -3.76526028e-01
7.09395528e-01 1.64204478e-01 6.65250719e-01 -9.54708755e-01
9.48150516e-01 5.13914049e-01 -5.62905967e-02 -4.76185322e-01
-9.53290761e-01 5.21870963e-02 -5.56093395e-01 -8.32252651e-02
8.81322622e-01 -9.46639180e-01 -3.49414021e-01 4.67885971e-01
-1.24571407e+00 -5.79225123e-01 -4.01101559e-01 -3.69049124e-02
-7.72244513e-01 3.28015178e-01 -4.78663057e-01 -4.38806742e-01
-5.05861819e-01 -1.38704669e+00 1.71194589e+00 1.94624424e-01
-2.23985732e-01 -9.05449033e-01 -1.02866530e-01 5.69246411e-01
7.68592000e-01 2.76330709e-01 8.24991643e-01 6.55118003e-02
-1.01283407e+00 2.77706772e-01 -3.81897688e-01 2.17512980e-01
2.05542013e-01 -7.26394430e-02 -8.56317878e-01 -3.40464830e-01
4.13895063e-02 -3.18326652e-01 8.35565388e-01 4.37052518e-01
1.19910312e+00 -3.46028745e-01 -2.47452676e-01 1.13473856e+00
1.29410815e+00 2.33252063e-01 9.13111985e-01 2.55168736e-01
9.82058346e-01 4.36133683e-01 6.83136642e-01 3.98530036e-01
3.32311869e-01 1.02733779e+00 3.94314766e-01 -3.63202363e-01
-5.96439362e-01 -6.00373685e-01 1.74216613e-01 5.57523668e-01
-1.31011933e-01 -3.59833807e-01 -5.03549457e-01 5.99140048e-01
-1.84418857e+00 -1.17737877e+00 2.90893614e-01 1.85838175e+00
9.85440314e-01 -2.17028916e-01 -2.56350040e-01 -1.74222142e-01
5.54702103e-01 4.36335415e-01 -6.42543197e-01 -2.19519228e-01
-3.04885767e-02 4.51376915e-01 7.69049972e-02 6.17164552e-01
-8.17434609e-01 1.25833797e+00 6.26580048e+00 1.05021417e+00
-1.25796068e+00 1.35844037e-01 8.47456396e-01 -2.63867587e-01
-6.34915233e-01 1.35642037e-01 -8.08513403e-01 3.70901018e-01
4.95194376e-01 -4.31546420e-02 6.03268504e-01 6.08870983e-01
6.18039012e-01 -1.73299834e-01 -9.79452372e-01 1.01017892e+00
1.93976030e-01 -1.66688192e+00 4.44714189e-01 -3.94000448e-02
1.03046477e+00 -4.47480649e-01 3.70614946e-01 2.40576193e-01
1.69597566e-01 -1.06528509e+00 9.09236670e-01 5.26889443e-01
1.14020514e+00 -5.54914713e-01 7.27710798e-02 2.08191976e-01
-1.14030886e+00 1.28671855e-01 -1.58995286e-01 1.11141913e-01
5.09925842e-01 4.74929124e-01 -6.02622271e-01 7.87622511e-01
5.24431229e-01 6.52556658e-01 -4.65988874e-01 5.72578251e-01
-2.49697328e-01 3.39452326e-01 -1.02751017e-01 6.86528921e-01
2.48237997e-01 -2.57488430e-01 4.73555624e-01 9.16556776e-01
5.06813049e-01 5.40410168e-02 2.38087893e-01 1.37359250e+00
6.82861209e-02 -2.60200083e-01 -6.24720216e-01 7.52002820e-02
3.61149758e-01 1.20704830e+00 -5.66799939e-01 -3.90542746e-01
-1.78586960e-01 1.40482461e+00 2.64779150e-01 5.87889314e-01
-1.07794619e+00 -5.48774302e-02 6.18603408e-01 3.07164967e-01
4.28030610e-01 -4.00749326e-01 -3.74795079e-01 -1.33806777e+00
-1.80838723e-02 -1.02902818e+00 -2.15450078e-01 -1.41394925e+00
-8.86594057e-01 7.13169932e-01 2.61783808e-01 -1.08354211e+00
-5.41862488e-01 -2.33559728e-01 -5.90871871e-01 1.06363046e+00
-1.61244285e+00 -1.75411797e+00 -6.37901247e-01 8.12813401e-01
8.47600877e-01 1.31113112e-01 5.77367425e-01 1.98168039e-01
-4.71803606e-01 5.43604612e-01 -1.37210727e-01 -3.16529483e-01
8.93573821e-01 -9.13112581e-01 6.12966120e-01 1.06558907e+00
1.41663536e-01 9.27401185e-01 5.56735516e-01 -8.71348560e-01
-1.34926379e+00 -1.36561787e+00 4.44562346e-01 -3.24150205e-01
1.70936495e-01 -3.89658689e-01 -7.49142885e-01 7.12666035e-01
4.14015412e-01 4.66816761e-02 8.46374854e-02 -3.70759457e-01
-2.77370185e-01 -1.81131903e-02 -1.01132214e+00 8.74814034e-01
1.41480124e+00 -5.19998908e-01 -2.75456399e-01 2.94686049e-01
8.27522874e-01 -6.48231685e-01 -7.47690916e-01 3.44508499e-01
3.77090037e-01 -1.19811308e+00 1.33633554e+00 -2.45931089e-01
9.30830538e-01 -4.46338654e-01 -5.75489365e-02 -1.35075390e+00
-3.78695846e-01 -1.12553906e+00 2.02737395e-02 1.25111532e+00
2.64105592e-02 -3.16484779e-01 6.50978982e-01 5.32303333e-01
-5.14456749e-01 -7.70800114e-01 -5.65194964e-01 -5.68742633e-01
-3.79329510e-02 -3.11508954e-01 5.37279010e-01 6.98800743e-01
-5.28522909e-01 3.11060131e-01 -8.52183402e-01 -5.32021299e-02
5.82965076e-01 4.41363931e-01 1.09539199e+00 -6.43827379e-01
-4.94379878e-01 -1.97252810e-01 2.81123556e-02 -1.43201435e+00
-2.95571797e-02 -7.20047951e-01 -1.61207747e-02 -1.65701640e+00
2.66855717e-01 -4.16172236e-01 2.69000143e-01 5.35426438e-01
-5.27260959e-01 4.21046376e-01 3.87236327e-01 4.22874212e-01
-3.72102588e-01 7.73154259e-01 1.88037109e+00 7.65305832e-02
-1.49969175e-01 -3.31581265e-01 -8.73455763e-01 4.48391408e-01
4.22150373e-01 -2.89514605e-02 -8.74284744e-01 -7.16212869e-01
7.36474292e-03 3.96417975e-01 6.34041667e-01 -8.71930540e-01
6.93208054e-02 -3.37627292e-01 4.95315224e-01 -7.40125895e-01
4.92657691e-01 -5.22985995e-01 5.88436604e-01 -1.02204934e-01
-3.41399968e-01 1.30131647e-01 1.14600152e-01 5.62303901e-01
-3.50076258e-01 4.33437079e-02 7.87580490e-01 -3.94619316e-01
-6.26508296e-01 4.22553718e-01 4.77671921e-02 1.47645995e-01
8.88308704e-01 -4.22101587e-01 -2.34771639e-01 -6.59038305e-01
-7.41924226e-01 -5.94288446e-02 7.53802955e-01 3.75537336e-01
8.56335342e-01 -1.28190506e+00 -6.80060804e-01 4.29786831e-01
-8.19217563e-02 2.80039489e-01 4.72586930e-01 7.08352089e-01
-2.87701964e-01 3.73696655e-01 -2.67800778e-01 -7.53434896e-01
-1.07888269e+00 7.29660690e-01 1.66713327e-01 -6.03183150e-01
-7.38006949e-01 7.47572839e-01 9.64540541e-01 -1.93615735e-01
-4.71120663e-02 -2.29418531e-01 2.80854940e-01 -3.32788378e-01
4.36308652e-01 1.89615451e-02 -1.26477882e-01 -4.03147578e-01
8.48390162e-02 6.92955792e-01 -3.05382699e-01 -3.39563936e-01
1.34248888e+00 -2.91846454e-01 5.51247299e-02 -2.05043599e-01
8.88058662e-01 -2.79026143e-02 -1.69525599e+00 -1.30676746e-01
-5.47731161e-01 -6.95629656e-01 2.81903017e-02 -8.61981153e-01
-1.19616830e+00 8.42866540e-01 1.83348641e-01 -3.17375988e-01
1.44260597e+00 -2.11985260e-01 8.82617950e-01 -1.90249488e-01
5.03714561e-01 -6.39752090e-01 4.88954961e-01 4.06956464e-01
1.38037682e+00 -9.31361020e-01 -4.25601527e-02 -5.62424719e-01
-7.48654604e-01 7.87872016e-01 7.18312919e-01 -1.15508988e-01
1.74571648e-01 5.00524640e-01 -1.57213077e-01 -8.83055478e-02
-9.01674330e-01 -9.25952271e-02 3.33950579e-01 8.62893581e-01
4.25204009e-01 -3.71697426e-01 -1.83420107e-01 2.04678982e-01
-5.30230440e-02 2.67752148e-02 3.44484001e-01 8.37212980e-01
-1.04801074e-01 -1.24680769e+00 -4.89763409e-01 1.36579156e-01
-1.21541835e-01 -4.23415899e-01 -1.08530015e-01 6.96332276e-01
1.81409828e-02 7.79084802e-01 -1.69971451e-01 -2.13718668e-01
1.70756951e-01 -8.63901600e-02 7.66746640e-01 -6.53619587e-01
-5.09285748e-01 3.94132495e-01 -1.50731757e-01 -9.05803800e-01
-4.74202186e-01 -3.23628277e-01 -7.30736315e-01 -1.00167349e-01
-1.28927022e-01 -2.33685508e-01 3.82574797e-01 8.41790497e-01
9.00023699e-01 7.64929712e-01 5.18840611e-01 -1.54658866e+00
-3.18499535e-01 -9.11307871e-01 -1.17761895e-01 5.46041608e-01
1.03325427e-01 -7.77349949e-01 -1.44238070e-01 3.58007580e-01] | [11.401127815246582, -0.643374502658844] |
e32f48d9-3892-4177-a7fe-a249b9adbcc1 | 6-pack-category-level-6d-pose-tracker-with | 1910.1075 | null | https://arxiv.org/abs/1910.10750v1 | https://arxiv.org/pdf/1910.10750v1.pdf | 6-PACK: Category-level 6D Pose Tracker with Anchor-Based Keypoints | We present 6-PACK, a deep learning approach to category-level 6D object pose tracking on RGB-D data. Our method tracks in real-time novel object instances of known object categories such as bowls, laptops, and mugs. 6-PACK learns to compactly represent an object by a handful of 3D keypoints, based on which the interframe motion of an object instance can be estimated through keypoint matching. These keypoints are learned end-to-end without manual supervision in order to be most effective for tracking. Our experiments show that our method substantially outperforms existing methods on the NOCS category-level 6D pose estimation benchmark and supports a physical robot to perform simple vision-based closed-loop manipulation tasks. Our code and video are available at https://sites.google.com/view/6packtracking. | ['Li Fei-Fei', 'Roberto Martín-Martín', 'Silvio Savarese', 'Jun Lv', 'Danfei Xu', 'Cewu Lu', 'Yuke Zhu', 'Chen Wang'] | 2019-10-23 | null | null | null | null | ['6d-pose-estimation-using-rgbd'] | ['computer-vision'] | [-4.06006455e-01 -1.60638332e-01 -6.27618492e-01 -1.18707679e-01
-7.57227540e-01 -7.14425266e-01 4.49894428e-01 2.73768127e-01
-2.99996585e-01 1.89740703e-01 -2.92414457e-01 -9.13322717e-02
-1.10049598e-01 -3.29923809e-01 -1.27366161e+00 -1.13629051e-01
-5.35191476e-01 9.94721413e-01 5.05688250e-01 1.46058455e-01
2.72390872e-01 1.08229065e+00 -1.41569805e+00 -8.81317481e-02
-1.23236410e-01 1.28523207e+00 4.69047040e-01 8.04746628e-01
2.50848711e-01 3.37954283e-01 -2.83264667e-01 1.91131160e-01
7.48770714e-01 3.91282618e-01 -3.78784180e-01 9.46787745e-02
9.91233051e-01 -7.78316855e-01 -6.02422118e-01 9.21455801e-01
2.25659400e-01 2.14923069e-01 3.75203550e-01 -1.61171675e+00
-2.19885468e-01 2.65300661e-01 -5.93931198e-01 -6.38170987e-02
5.26537955e-01 5.34197748e-01 6.03433967e-01 -9.48310792e-01
7.89649725e-01 1.53943646e+00 7.93999672e-01 5.51184237e-01
-9.90727901e-01 -9.73996639e-01 3.17925751e-01 1.61644444e-03
-1.17170525e+00 -3.12999994e-01 6.15066290e-01 -4.83875692e-01
9.83612239e-01 -1.37476444e-01 9.13006842e-01 9.15280521e-01
6.17446840e-01 7.83864081e-01 6.23284161e-01 -1.03182651e-01
2.03210652e-01 -3.84989351e-01 5.79088181e-02 1.06225145e+00
4.57496226e-01 2.29385376e-01 -5.74090958e-01 -2.23769069e-01
1.19668984e+00 4.80187505e-01 1.17104456e-01 -1.39281809e+00
-1.74251425e+00 4.31585759e-01 8.95421743e-01 -2.43750945e-01
-4.39726591e-01 1.02116680e+00 1.43210083e-01 1.27300516e-01
-3.77610400e-02 2.40079746e-01 -6.57996118e-01 -2.72594720e-01
-2.00836360e-01 6.55387521e-01 5.84877372e-01 1.64781344e+00
6.60676360e-01 -2.55824775e-01 1.36220485e-01 1.01731077e-01
6.28308654e-01 9.35301185e-01 2.30729561e-02 -1.40216780e+00
3.91017705e-01 6.97434187e-01 6.49003506e-01 -5.87916434e-01
-4.53634799e-01 -2.95347393e-01 -1.25313913e-02 7.97051907e-01
3.69204581e-01 5.11322469e-02 -1.13670075e+00 1.32791615e+00
7.74491191e-01 2.24079728e-01 -3.92169744e-01 1.12925375e+00
5.80407739e-01 4.49349940e-01 -1.05722211e-01 4.16518658e-01
1.25984144e+00 -9.70469296e-01 -2.79659778e-01 -3.56984377e-01
3.21267188e-01 -6.12686992e-01 7.53264368e-01 3.17037165e-01
-9.32354689e-01 -6.57321036e-01 -8.87908101e-01 -1.19994365e-01
-3.34435165e-01 1.73069257e-02 8.00053537e-01 2.36410405e-02
-7.27305830e-01 6.22974157e-01 -1.56372583e+00 -3.74260217e-01
5.47658682e-01 6.48963213e-01 -5.65207899e-01 1.49643207e-02
-4.00880635e-01 9.89415705e-01 4.50696200e-01 -7.36019388e-03
-1.45070481e+00 -7.54304528e-01 -9.01316166e-01 -4.02284682e-01
5.06570756e-01 -6.95280671e-01 1.81394410e+00 -8.51763189e-02
-1.27865803e+00 9.80148673e-01 8.07482600e-02 -4.97581005e-01
5.03244221e-01 -8.02525222e-01 1.00724876e-01 9.04191881e-02
6.62287623e-02 8.50949407e-01 9.98946309e-01 -1.35329247e+00
-8.61331463e-01 -6.70135736e-01 2.74404764e-01 2.31429622e-01
2.44956776e-01 -2.93593049e-01 -6.49610817e-01 -2.52109647e-01
6.10135615e-01 -1.32385886e+00 -1.89923227e-01 1.03678823e+00
-2.50318706e-01 -3.74448448e-01 1.36720324e+00 -6.37728646e-02
1.21277839e-01 -1.90084255e+00 1.49564072e-01 -6.50779009e-02
2.99889952e-01 2.93073468e-02 1.99402478e-02 1.91337585e-01
2.52528608e-01 -5.47313333e-01 4.83546108e-01 -4.08449590e-01
2.76608884e-01 -4.64689173e-02 -1.25015706e-01 8.31915855e-01
-1.67160675e-01 9.49637055e-01 -1.17010391e+00 -1.38009757e-01
5.38741827e-01 4.45611507e-01 -4.18281198e-01 2.71105647e-01
-5.64428508e-01 2.86340982e-01 -6.13569856e-01 1.03438020e+00
6.12495959e-01 -1.72876239e-01 -1.95325226e-01 -4.74193662e-01
-1.08532440e-02 3.55203599e-01 -1.26390076e+00 2.31647825e+00
-2.96856314e-01 5.50415218e-01 2.48210043e-01 -3.40618104e-01
7.60076463e-01 -2.10460320e-01 6.15667045e-01 -2.91472137e-01
1.88588515e-01 8.52174759e-02 -2.73704797e-01 -1.68530233e-02
4.67549473e-01 3.48957151e-01 -3.19368780e-01 3.14232945e-01
9.09454674e-02 -4.60224539e-01 -1.57862186e-01 1.33744523e-01
1.22653461e+00 5.57308257e-01 2.41762444e-01 -4.60935980e-02
-1.60942093e-01 4.28064317e-01 2.98738360e-01 8.16719651e-01
-3.13994169e-01 2.95499593e-01 -1.80446953e-01 -4.94193673e-01
-1.00858784e+00 -1.44706416e+00 -6.40726555e-03 9.15831804e-01
7.35025465e-01 -4.00514483e-01 -1.79840297e-01 -6.27319396e-01
9.16537702e-01 2.87361503e-01 -2.82695800e-01 -1.23741254e-01
-6.84497476e-01 3.82621855e-01 1.00387797e-01 8.71096849e-01
1.69287488e-01 -8.09250712e-01 -1.50421584e+00 2.79769868e-01
4.68305647e-01 -9.72690046e-01 -4.01793659e-01 5.82013190e-01
-1.07986605e+00 -1.33731949e+00 -3.68779629e-01 -8.50379169e-01
6.49654329e-01 6.54459178e-01 7.87382662e-01 -2.49234080e-01
-5.67563713e-01 9.76214468e-01 -2.94052392e-01 -4.95655924e-01
-2.92965502e-01 -2.12428689e-01 4.83463585e-01 -8.47942650e-01
3.97389114e-01 -2.29703575e-01 -8.29061985e-01 3.49705487e-01
-1.80601642e-01 -5.07351644e-02 4.45405841e-01 3.41436833e-01
1.00878143e+00 -2.59555995e-01 -2.70418882e-01 5.24488613e-02
1.01869749e-02 -1.38873577e-01 -1.24244869e+00 -6.20735884e-02
-1.33902341e-01 -2.02228367e-01 -5.13350777e-02 -7.76368201e-01
-4.51543361e-01 7.31804073e-01 4.24974978e-01 -1.08349502e+00
-3.59005600e-01 6.93687052e-02 1.14052333e-01 -5.00682354e-01
3.94316435e-01 -1.99149832e-01 1.01270400e-01 -7.29433894e-01
4.30990934e-01 3.12045544e-01 7.43231297e-01 -6.94879711e-01
1.14421833e+00 5.01687646e-01 1.84684262e-01 -2.97010779e-01
-5.43438792e-01 -6.61990285e-01 -8.30395639e-01 -4.62614506e-01
5.31342924e-01 -1.18892419e+00 -1.38808584e+00 5.74328244e-01
-1.08417273e+00 -6.40101552e-01 -1.73530906e-01 7.21260667e-01
-8.74422073e-01 -1.63608491e-01 -6.59233630e-01 -6.61952794e-01
-2.76924521e-01 -1.11220288e+00 1.75580895e+00 2.79346436e-01
-1.39193505e-01 -4.11389261e-01 -8.56456831e-02 1.86282117e-02
5.56778088e-02 4.27727342e-01 4.08717811e-01 -2.34561875e-01
-1.10497797e+00 -4.29944545e-01 5.55101782e-02 -2.98729867e-01
2.38409445e-01 -1.93040848e-01 -4.75176424e-01 -8.15648794e-01
-3.56760472e-01 -4.78128850e-01 4.73691404e-01 2.38134742e-01
1.08618021e+00 -1.27317339e-01 -9.41083908e-01 6.55148804e-01
1.38034129e+00 2.53042489e-01 -1.08217992e-01 4.97730076e-01
8.40657771e-01 -9.84638929e-02 1.28275239e+00 4.37727898e-01
3.80143106e-01 9.39720392e-01 1.21183848e+00 3.32925379e-01
-9.46329758e-02 -4.35541183e-01 1.94494233e-01 2.27056146e-01
2.48489678e-01 3.28544751e-02 -1.04112101e+00 4.44939166e-01
-1.79899991e+00 -5.63048244e-01 4.72705029e-02 2.07843161e+00
4.42171127e-01 5.74063063e-01 2.26785511e-01 -3.50708991e-01
6.24357402e-01 -5.91758788e-02 -1.29730713e+00 1.17585748e-01
6.40765548e-01 1.07961218e-03 1.04816723e+00 2.91741669e-01
-1.31833839e+00 1.00821221e+00 5.43674421e+00 2.04543725e-01
-9.80159283e-01 -1.81415584e-02 -2.37222090e-01 -4.03684378e-01
2.41545424e-01 -6.41106488e-03 -1.23718464e+00 1.66728511e-01
3.97227675e-01 1.01842759e-02 2.13328376e-01 1.38562202e+00
-1.84524149e-01 -1.97662562e-01 -1.56613266e+00 1.06334424e+00
-2.16520563e-01 -1.31069839e+00 -3.27757448e-01 6.35429248e-02
3.34800810e-01 4.26014334e-01 2.61617035e-01 2.71666676e-01
5.99921703e-01 -6.10421956e-01 1.16152918e+00 3.35060626e-01
4.90500718e-01 -4.85298246e-01 1.68968573e-01 4.13007557e-01
-1.34619927e+00 -1.92896664e-01 -3.35653663e-01 5.53383976e-02
1.01545326e-01 3.39076556e-02 -1.13423085e+00 -3.78295854e-02
1.00563252e+00 6.84370577e-01 -3.74339193e-01 1.48765457e+00
-6.27415627e-02 -2.47971993e-02 -6.72040284e-01 -1.65152341e-01
3.90106857e-01 3.88369113e-01 8.70281279e-01 7.42295504e-01
2.98655272e-01 1.39880344e-01 4.49007273e-01 8.32471013e-01
-8.30613598e-02 -6.98138833e-01 -6.19466484e-01 1.10187314e-01
7.88334131e-01 1.08067548e+00 -8.38911355e-01 -9.28898752e-02
-2.04797015e-02 8.48189294e-01 2.44697869e-01 9.05660781e-05
-8.57705891e-01 -3.89487118e-01 1.08457398e+00 1.47197157e-01
6.75455093e-01 -8.99608076e-01 2.71792531e-01 -9.22847271e-01
2.02098563e-01 -6.23118699e-01 1.05616517e-01 -1.12235856e+00
-9.22298670e-01 3.32612768e-02 2.87055612e-01 -1.51641703e+00
-2.11613730e-01 -1.06903398e+00 -1.56286091e-01 3.14827442e-01
-1.03706193e+00 -1.24660122e+00 -6.73707843e-01 6.87941492e-01
6.40745699e-01 7.66302198e-02 6.71007335e-01 -6.59861322e-03
2.22782090e-01 3.13839495e-01 -2.26546265e-02 1.06942281e-01
6.38610482e-01 -1.12611258e+00 9.39410269e-01 3.38527381e-01
1.40493438e-01 6.47271216e-01 7.65054226e-01 -9.23898876e-01
-2.33939052e+00 -1.07158530e+00 6.16574436e-02 -9.02867854e-01
7.30483651e-01 -7.18416631e-01 -5.73745131e-01 1.18214631e+00
-4.56211448e-01 5.44964969e-01 -1.43200994e-01 -1.25776619e-01
-4.35256541e-01 -2.66468644e-01 -1.18068910e+00 4.54915375e-01
1.56149960e+00 -3.64142329e-01 -7.70674646e-01 6.64648056e-01
9.60751414e-01 -1.48990369e+00 -1.02755344e+00 4.70463574e-01
9.09289598e-01 -3.85833263e-01 1.35998881e+00 -5.78569889e-01
-9.36670303e-02 -7.00120449e-01 -1.85490444e-01 -1.08590269e+00
-3.87645364e-01 -4.44397837e-01 -7.68784702e-01 4.07427669e-01
-2.48907194e-01 -3.04014027e-01 1.03596377e+00 3.89895588e-01
-1.50475129e-01 -4.71304059e-01 -9.90538180e-01 -1.15107107e+00
-3.72199595e-01 -3.66360694e-01 6.45497620e-01 3.57060492e-01
-1.46566302e-01 -2.12237149e-01 -1.42693415e-01 4.99335557e-01
1.22232831e+00 6.00473821e-01 1.29486871e+00 -1.24893677e+00
-6.10930845e-02 -3.67699444e-01 -9.59283888e-01 -1.45764053e+00
1.74628466e-01 -6.04626179e-01 3.89568746e-01 -1.42005932e+00
-5.24604134e-02 -6.58076823e-01 -2.59172410e-01 7.47902155e-01
3.58640492e-01 2.03449532e-01 4.24106389e-01 1.76342756e-01
-1.07266378e+00 3.34722698e-01 1.18599832e+00 -3.13471556e-01
-6.87867478e-02 1.33615062e-01 -4.12328616e-02 6.55916214e-01
8.15141141e-01 -6.31348193e-01 -4.91174646e-02 -6.07424378e-01
-1.89644545e-01 2.82895088e-01 8.41581523e-01 -1.30250835e+00
4.92566943e-01 -1.67193830e-01 7.45743752e-01 -1.26841652e+00
7.89503813e-01 -1.30979323e+00 2.64459193e-01 9.11342919e-01
-2.71516770e-01 1.52881458e-01 4.65480149e-01 7.41969347e-01
4.11617279e-01 6.34644628e-02 5.69998085e-01 -3.12943786e-01
-1.20816624e+00 6.56922519e-01 -1.24490976e-01 -4.50451881e-01
1.29669297e+00 -2.87839264e-01 -2.74639696e-01 -6.84157237e-02
-5.78689158e-01 4.02840018e-01 7.62897730e-01 9.33957875e-01
7.01345384e-01 -1.49785984e+00 -1.59859374e-01 5.25363311e-02
3.79661202e-01 4.12688524e-01 -1.81024268e-01 4.20260489e-01
-6.19490445e-01 4.48909968e-01 -4.31555927e-01 -1.22337270e+00
-1.27178836e+00 6.57011092e-01 3.11334878e-01 5.10689139e-01
-1.00181115e+00 8.98680091e-01 -1.84901714e-01 -6.28210545e-01
8.42308819e-01 -7.08380580e-01 5.61014354e-01 -6.08050823e-01
2.97444761e-01 3.48889053e-01 -5.05096875e-02 -5.52371919e-01
-7.17007399e-01 7.48008311e-01 -4.77182940e-02 3.19952965e-01
1.31181073e+00 -7.43753389e-02 2.09780365e-01 5.32641053e-01
1.12611890e+00 -3.55861872e-01 -2.02519870e+00 -3.38249058e-01
-4.22996022e-02 -7.05195665e-01 3.18991505e-02 -6.77564979e-01
-8.05080652e-01 4.45167810e-01 1.05067456e+00 -5.11511445e-01
4.01656330e-01 3.09373021e-01 8.43488276e-01 1.06732953e+00
1.23980403e+00 -7.11572707e-01 3.27210814e-01 4.39991027e-01
8.10880661e-01 -1.29482150e+00 1.96980730e-01 -3.36932987e-02
3.32702212e-02 1.17790782e+00 9.22778845e-01 -4.68072951e-01
5.31058073e-01 6.33940220e-01 -7.30866194e-02 -3.52558017e-01
-5.51876545e-01 2.16465583e-03 2.98354447e-01 8.08524907e-01
-2.70997256e-01 1.40831590e-01 7.30341852e-01 -6.62800521e-02
-3.59213315e-02 -1.06796555e-01 8.95425379e-02 1.52985132e+00
-7.98372746e-01 -5.86982787e-01 -5.84512174e-01 2.35898271e-01
1.41871516e-02 5.37374198e-01 -1.44213423e-01 1.09324014e+00
-4.99833263e-02 3.88660401e-01 2.85088539e-01 -3.47712994e-01
5.14604509e-01 -2.10311756e-01 1.09427977e+00 -8.12215567e-01
-2.44090065e-01 -1.02758653e-01 -2.28935033e-01 -1.17724013e+00
-3.66883785e-01 -7.51543045e-01 -1.49843335e+00 -1.46256775e-01
-3.07953447e-01 -3.39688390e-01 1.05406392e+00 5.25556564e-01
4.47085738e-01 2.25389078e-01 1.63994461e-01 -1.65945828e+00
-7.39196599e-01 -5.46207666e-01 -2.55108386e-01 1.42331660e-01
7.15760112e-01 -1.10382068e+00 -8.43246579e-02 -6.09073378e-02] | [7.080930233001709, -2.383798122406006] |
d986230f-ea15-4c39-b314-59557dc7b2d4 | deterministic-parsing-using-pcfgs | null | null | https://aclanthology.org/E14-1036 | https://aclanthology.org/E14-1036.pdf | Deterministic Parsing using PCFGs | null | ['Mark-Jan Nederhof', 'Martin McCaffery'] | 2014-04-01 | null | null | null | eacl-2014-4 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.414912700653076, 3.640712022781372] |
e4915ec2-59b1-4b86-8a4a-df4c7918f3a7 | u-time-a-fully-convolutional-network-for-time | 1910.11162 | null | https://arxiv.org/abs/1910.11162v1 | https://arxiv.org/pdf/1910.11162v1.pdf | U-Time: A Fully Convolutional Network for Time Series Segmentation Applied to Sleep Staging | Neural networks are becoming more and more popular for the analysis of physiological time-series. The most successful deep learning systems in this domain combine convolutional and recurrent layers to extract useful features to model temporal relations. Unfortunately, these recurrent models are difficult to tune and optimize. In our experience, they often require task-specific modifications, which makes them challenging to use for non-experts. We propose U-Time, a fully feed-forward deep learning approach to physiological time series segmentation developed for the analysis of sleep data. U-Time is a temporal fully convolutional network based on the U-Net architecture that was originally proposed for image segmentation. U-Time maps sequential inputs of arbitrary length to sequences of class labels on a freely chosen temporal scale. This is done by implicitly classifying every individual time-point of the input signal and aggregating these classifications over fixed intervals to form the final predictions. We evaluated U-Time for sleep stage classification on a large collection of sleep electroencephalography (EEG) datasets. In all cases, we found that U-Time reaches or outperforms current state-of-the-art deep learning models while being much more robust in the training process and without requiring architecture or hyperparameter adaptation across tasks. | ['Poul Jørgen Jennum', 'Michael Hejselbak Jensen', 'Mathias Perslev', 'Christian Igel', 'Sune Darkner'] | 2019-10-24 | u-time-a-fully-convolutional-network-for-time-1 | http://papers.nips.cc/paper/8692-u-time-a-fully-convolutional-network-for-time-series-segmentation-applied-to-sleep-staging | http://papers.nips.cc/paper/8692-u-time-a-fully-convolutional-network-for-time-series-segmentation-applied-to-sleep-staging.pdf | neurips-2019-12 | ['sleep-staging'] | ['medical'] | [ 2.55996913e-01 -3.41694981e-01 -9.99280438e-02 -5.61380982e-01
-4.16828096e-01 -4.88571614e-01 2.34606847e-01 7.84748495e-02
-6.14780247e-01 6.27670884e-01 5.18022701e-02 -4.22196060e-01
-3.31399262e-01 -2.92939097e-01 -4.51997876e-01 -6.16360724e-01
-4.85615671e-01 1.70946896e-01 1.13616362e-01 -1.47756889e-01
1.44374028e-01 3.95287693e-01 -1.39198923e+00 3.21231186e-01
6.40607953e-01 1.43399549e+00 1.16503574e-01 6.03024960e-01
1.68212280e-02 4.82963294e-01 -7.45216310e-01 2.69320190e-01
-1.36135370e-02 -6.97551250e-01 -7.94700742e-01 -2.51590401e-01
-2.28432149e-01 2.34705284e-01 -3.56765866e-01 5.46686947e-01
5.82490742e-01 2.80404001e-01 3.64573717e-01 -1.04268110e+00
-3.37486625e-01 6.40256166e-01 2.97850110e-02 9.86525536e-01
1.61443517e-01 -1.78313497e-02 8.53848040e-01 -3.48501980e-01
4.24145252e-01 5.91094077e-01 1.06204188e+00 6.35262728e-01
-1.54369116e+00 -7.31376708e-01 2.96468325e-02 4.62610543e-01
-1.15925527e+00 -3.43374670e-01 8.38916421e-01 -4.01885211e-01
1.59324241e+00 2.45904356e-01 1.12054181e+00 1.39307296e+00
8.78572643e-01 7.17628956e-01 1.15471506e+00 -4.20447551e-02
4.38316464e-01 -5.18627107e-01 4.79166240e-01 2.54175752e-01
-6.36404812e-01 -1.29914179e-01 -6.84338868e-01 1.38127893e-01
5.95369279e-01 3.50624472e-01 -2.68649399e-01 1.92890182e-01
-1.28151345e+00 5.13817310e-01 7.19162941e-01 7.25515008e-01
-4.83019471e-01 1.41207650e-01 7.18999147e-01 6.39306128e-01
9.04553890e-01 7.45073915e-01 -8.79926562e-01 -5.20141244e-01
-1.47530127e+00 -8.52068812e-02 4.97663319e-01 4.52525526e-01
3.85130376e-01 -2.98707876e-02 -2.89945960e-01 7.39926875e-01
-1.86245069e-01 -2.66059518e-01 1.18270946e+00 -8.31641138e-01
-5.45819476e-02 6.47427797e-01 -1.45687148e-01 -3.12302619e-01
-1.20006216e+00 -6.00590050e-01 -9.11123574e-01 -9.53276083e-02
2.77321994e-01 -1.83927312e-01 -1.16776359e+00 1.73256218e+00
-3.85419160e-01 5.01115441e-01 -9.47974771e-02 7.03707814e-01
6.63902521e-01 6.80765986e-01 5.29077882e-03 -4.98762101e-01
1.27992404e+00 -5.64223766e-01 -8.24678957e-01 -3.67692858e-01
4.34386194e-01 -2.47214004e-01 9.32907045e-01 5.38219333e-01
-1.00432956e+00 -4.96118158e-01 -1.21469486e+00 -1.05037019e-01
-6.80586040e-01 3.71723212e-02 7.34824717e-01 3.15227658e-01
-1.33836019e+00 1.20803380e+00 -1.35236096e+00 -4.51232046e-01
4.43400741e-01 9.46619511e-01 -2.27601856e-01 7.71187425e-01
-1.25581193e+00 8.79163206e-01 1.66436329e-01 4.14635718e-01
-8.78708601e-01 -6.79446816e-01 -6.50099099e-01 7.36531094e-02
-1.21786453e-01 -4.33341503e-01 1.39352620e+00 -8.72088552e-01
-1.52408886e+00 8.70864749e-01 -4.05526519e-01 -1.02263224e+00
1.41513556e-01 -1.84450880e-01 -5.80567479e-01 -5.38080707e-02
4.43266258e-02 5.48760414e-01 7.83557832e-01 -3.10137391e-01
-1.92254975e-01 -4.52402890e-01 -2.84999996e-01 -1.69845238e-01
-2.34089166e-01 1.24127835e-01 -2.23287046e-01 -6.09871686e-01
1.96612567e-01 -9.28639650e-01 -2.70819783e-01 -3.36259991e-01
2.77078338e-02 -5.35651147e-01 8.92393470e-01 -4.51384068e-01
1.29613316e+00 -2.25950098e+00 3.44362468e-01 -1.88083678e-01
3.54166716e-01 1.37813166e-01 8.56053680e-02 2.62735188e-01
-4.72495049e-01 -7.69554377e-02 -3.84352893e-01 -5.06477594e-01
-1.52730927e-01 2.63691932e-01 -3.63774657e-01 5.69937348e-01
1.78425297e-01 1.12914336e+00 -9.48141515e-01 -6.17146455e-02
2.51541972e-01 4.31825787e-01 -2.47598946e-01 2.60633260e-01
-3.19413662e-01 6.87093914e-01 -6.57255277e-02 2.58711755e-01
-1.01898082e-01 -3.13280940e-01 -1.12852044e-01 -5.08855423e-03
-3.36573005e-01 8.09284806e-01 -3.16490561e-01 2.02573156e+00
-5.44772565e-01 1.16654623e+00 -5.73335290e-01 -1.24153185e+00
9.42570150e-01 6.21637285e-01 9.66916203e-01 -1.02490771e+00
5.22142708e-01 1.92725867e-01 2.22571760e-01 -4.94112551e-01
1.35181651e-01 -1.65576607e-01 -1.57746136e-01 7.98126042e-01
5.00368655e-01 6.83071241e-02 2.35651866e-01 -3.33004117e-01
1.46800542e+00 1.60511240e-01 1.92584887e-01 -3.09084177e-01
1.49415463e-01 -4.73348826e-01 6.69219971e-01 3.73666704e-01
-3.51635993e-01 6.37768209e-01 5.83324015e-01 -8.97193670e-01
-8.59894693e-01 -9.59382117e-01 -4.06569541e-01 1.21756065e+00
-1.28229141e-01 -5.75269461e-01 -7.43105233e-01 -3.36716950e-01
-4.70574468e-01 4.74303663e-01 -1.09961760e+00 -4.26806867e-01
-4.59614515e-01 -9.40772772e-01 4.91450399e-01 7.56516874e-01
1.88377976e-01 -1.70245457e+00 -1.21561861e+00 6.62397742e-01
-1.34395778e-01 -1.04997694e+00 -2.91374415e-01 1.20380259e+00
-1.04335511e+00 -8.49968731e-01 -6.25865579e-01 -6.63751423e-01
3.56284499e-01 -2.12343276e-01 1.00129890e+00 -2.86490202e-01
-3.68887842e-01 1.15742072e-01 -3.59700292e-01 -4.71729785e-01
1.78340793e-01 3.49499166e-01 1.92704231e-01 9.80525371e-03
6.80356741e-01 -1.10448396e+00 -8.26569736e-01 2.97007412e-01
-7.79295385e-01 -7.65843987e-02 3.06195378e-01 8.12258422e-01
7.03742385e-01 2.48619765e-02 7.04264760e-01 -4.49583143e-01
9.02913630e-01 -4.73467737e-01 -2.52598852e-01 4.44295853e-02
-5.01744449e-01 1.02153808e-01 8.97973716e-01 -6.26231968e-01
-3.08434546e-01 -1.31841719e-01 -2.03856364e-01 -6.64215982e-01
6.11683726e-02 6.64122522e-01 5.55881619e-01 -1.53431688e-02
7.69890070e-01 5.05851388e-01 -2.00733081e-01 -2.81249076e-01
-4.04586121e-02 5.00824511e-01 7.24663854e-01 -1.44332200e-01
-2.77952999e-02 2.95264035e-01 -2.41891474e-01 -9.16819751e-01
-9.95209992e-01 -5.10294437e-01 -9.86628175e-01 -2.90466428e-01
1.15710926e+00 -5.76726556e-01 -6.24643147e-01 4.84612584e-01
-9.76418972e-01 -8.68229866e-01 -3.02205384e-01 4.22819018e-01
-7.87252843e-01 -1.61879048e-01 -6.44506752e-01 -5.84031761e-01
-5.77174962e-01 -1.05483186e+00 1.00922930e+00 2.40490556e-01
-6.35432661e-01 -9.96419072e-01 1.32115319e-01 -1.80117339e-01
4.61128056e-01 2.27529973e-01 8.16724658e-01 -6.01895630e-01
-2.59256940e-02 -1.35728911e-01 1.92445472e-01 1.19652554e-01
1.05163664e-01 -1.69667289e-01 -1.13965189e+00 -1.88855380e-01
4.20863301e-01 -3.21177363e-01 9.48384285e-01 6.81234419e-01
1.81006002e+00 5.62729277e-02 -2.26298749e-01 1.01422656e+00
8.71329069e-01 4.42844898e-01 6.02892160e-01 3.86303335e-01
2.10017875e-01 1.82277426e-01 5.50020076e-02 3.94828975e-01
2.69600928e-01 5.35291791e-01 3.02728862e-01 -1.32623184e-02
3.54476631e-01 2.58285761e-01 3.60172212e-01 8.81769776e-01
-2.13839889e-01 1.56932920e-01 -9.75094318e-01 7.14375854e-01
-1.90352964e+00 -8.57215166e-01 2.64960945e-01 2.05862045e+00
9.80145454e-01 4.19684470e-01 2.10627034e-01 4.72194731e-01
2.29541689e-01 1.30173147e-01 -8.42315316e-01 -4.78451699e-01
8.48271325e-02 6.82426691e-01 2.03150302e-01 -1.48923442e-01
-1.00640881e+00 8.78692746e-01 6.95918322e+00 3.14995557e-01
-1.67842448e+00 2.33262703e-01 6.21001482e-01 -4.47837085e-01
1.39637738e-01 -3.59008163e-01 -3.58336538e-01 6.40561461e-01
1.71474195e+00 -2.47424960e-01 8.01921070e-01 3.16656530e-01
5.34622192e-01 3.26654688e-02 -1.36313522e+00 1.20747972e+00
-2.28934601e-01 -1.32831895e+00 -6.93365991e-01 -2.63102204e-01
5.64765513e-01 5.38224876e-01 4.24065255e-02 4.01810527e-01
-9.12571177e-02 -1.33681345e+00 7.76151061e-01 6.09773993e-01
7.40184963e-01 -5.97096682e-01 6.28925622e-01 2.24155024e-01
-1.12645209e+00 -3.42841476e-01 -1.93180233e-01 -3.50105792e-01
4.91346978e-02 3.52556676e-01 -4.79253113e-01 1.33811265e-01
1.08914530e+00 1.18510914e+00 -6.88592136e-01 1.11840570e+00
-9.71539915e-02 7.55112588e-01 -3.38620692e-01 -8.42154697e-02
4.23721999e-01 -9.71641112e-03 1.35345221e-01 1.18954563e+00
4.28373575e-01 9.04552713e-02 -1.59466818e-01 8.99858177e-01
-9.26208049e-02 -4.29503828e-01 -4.68115419e-01 -3.37070823e-01
1.03032529e-01 1.26329446e+00 -1.18743348e+00 -3.12522650e-02
-1.17982082e-01 1.03258944e+00 4.17661250e-01 3.77611399e-01
-8.06152821e-01 -4.48277205e-01 6.31008446e-01 -2.46900618e-01
7.80075118e-02 -4.97677833e-01 -3.92560989e-01 -1.07430053e+00
-2.24859901e-02 -3.53706211e-01 3.89456898e-01 -9.00999963e-01
-1.16954565e+00 1.13231230e+00 -1.66587368e-01 -1.17234910e+00
-4.87131596e-01 -4.58934039e-01 -8.63124967e-01 6.89476609e-01
-1.18164361e+00 -7.13524997e-01 -1.30510151e-01 8.03237319e-01
6.95157051e-01 6.37410507e-02 9.84153211e-01 1.86180577e-01
-6.82893574e-01 2.66151041e-01 -8.25598743e-03 -3.83376144e-02
4.41048950e-01 -1.42135680e+00 7.42913663e-01 8.22426319e-01
2.71484554e-01 9.08779919e-01 6.37260675e-01 -2.40602866e-01
-1.11721897e+00 -1.04611957e+00 9.00044918e-01 -3.81933630e-01
9.54437017e-01 -7.89153337e-01 -1.05056858e+00 7.41761982e-01
2.73910493e-01 1.56840011e-01 8.70945692e-01 2.41333380e-01
5.36467209e-02 -1.66551352e-01 -7.61346936e-01 5.28526306e-01
9.16573286e-01 -9.49643970e-01 -8.13926339e-01 2.01991782e-01
4.77026850e-01 -4.06720459e-01 -1.06580460e+00 2.99187213e-01
5.66046417e-01 -9.80122805e-01 5.40452361e-01 -3.37843865e-01
2.26950094e-01 -8.10095519e-02 3.61094922e-01 -1.45374882e+00
-3.82262945e-01 -1.14152479e+00 -1.51380897e-01 5.35139501e-01
4.54766124e-01 -5.27255476e-01 6.48071647e-01 3.31241190e-01
-5.74992716e-01 -1.05534124e+00 -1.10181010e+00 -6.06492400e-01
-2.01517478e-01 -7.01550722e-01 4.52895731e-01 6.65233552e-01
3.87354910e-01 5.85847080e-01 -1.77974015e-01 -3.01273823e-01
1.04775473e-01 2.17530042e-01 4.38272469e-02 -1.37626684e+00
1.18049949e-01 -7.50834942e-01 -4.62272704e-01 -7.19513953e-01
4.39250916e-01 -9.65558529e-01 2.27001742e-01 -1.55422199e+00
-2.78541774e-01 -2.41559848e-01 -9.61234450e-01 8.10726702e-01
1.08120635e-01 4.96347845e-01 -2.99436778e-01 2.80047566e-01
-5.81756115e-01 6.50677562e-01 8.96566451e-01 -1.60295442e-01
-5.82417250e-01 3.09459746e-01 -2.99328387e-01 4.70372111e-01
1.04761064e+00 -3.75157952e-01 -6.13150656e-01 -3.25076550e-01
2.46505022e-01 1.98799759e-01 1.69398576e-01 -1.44987237e+00
3.37041020e-01 3.52529228e-01 7.08663821e-01 -6.16585195e-01
5.27726710e-01 -6.35421932e-01 1.01545274e-01 3.62860829e-01
-4.84201878e-01 2.65947133e-01 4.06894207e-01 4.02875155e-01
-1.36472419e-01 1.91234946e-01 7.20023870e-01 -2.07933396e-01
-7.18153059e-01 4.91100341e-01 -5.70475161e-01 -2.41314188e-01
8.75461578e-01 -3.79425615e-01 -1.03010438e-01 -1.69437632e-01
-1.04734623e+00 2.68469036e-01 3.83195281e-02 5.65265596e-01
5.21879554e-01 -1.06175649e+00 -1.96726605e-01 5.98726869e-01
3.58408429e-02 -2.62793303e-01 1.32428885e-01 1.15219212e+00
-1.74903482e-01 6.04133666e-01 -6.88125372e-01 -8.43307674e-01
-8.97541761e-01 3.98393124e-01 7.06075549e-01 -1.55296728e-01
-9.92884517e-01 8.26010406e-01 -7.95791745e-02 -6.53165355e-02
2.58663863e-01 -8.90639782e-01 -5.96646249e-01 3.17272961e-01
4.71533418e-01 -1.05934143e-01 4.55437452e-01 -3.37984443e-01
-3.33231330e-01 3.30461532e-01 -1.73725579e-02 -1.13780625e-01
1.75593591e+00 1.52496248e-01 -3.51236433e-01 1.19776499e+00
1.20771050e+00 -8.63784015e-01 -1.40070128e+00 7.96629637e-02
1.73166871e-01 3.85522932e-01 1.61980212e-01 -6.73081219e-01
-1.11217487e+00 1.01640761e+00 6.19911253e-01 5.14093339e-01
1.46092010e+00 -1.85846999e-01 9.61000741e-01 3.27137768e-01
3.73138577e-01 -8.50577831e-01 -1.10596744e-02 6.30344808e-01
6.72383904e-01 -8.75120699e-01 -2.11715549e-01 3.86875600e-01
-2.78401136e-01 1.36009943e+00 3.84424090e-01 -2.91293353e-01
7.82389283e-01 1.43404976e-01 8.67155492e-02 -3.65028739e-01
-1.11970270e+00 -1.02665119e-01 3.81438136e-01 3.32446247e-01
6.98190391e-01 -2.91723683e-02 -3.09939921e-01 7.98848748e-01
-5.74104249e-01 3.07979703e-01 3.34809273e-01 7.71470845e-01
-2.34475344e-01 -9.00615871e-01 3.64176333e-02 9.12524939e-01
-7.81804681e-01 8.73285532e-03 -2.46676177e-01 3.90976578e-01
3.69314887e-02 7.34389722e-01 2.35321864e-01 -7.15807557e-01
-9.03848838e-03 3.35260123e-01 5.22636116e-01 -6.44568384e-01
-9.42257643e-01 1.42700225e-01 -3.07226479e-01 -7.93275595e-01
-4.56156880e-01 -6.20703340e-01 -1.32502949e+00 2.20657378e-01
3.91872600e-02 2.69144416e-01 6.59125686e-01 1.29854167e+00
3.04490685e-01 1.06143248e+00 3.75750989e-01 -1.31281793e+00
3.46480682e-02 -1.15278602e+00 -4.81489211e-01 1.52075052e-01
7.50794470e-01 -6.03820980e-01 -1.78714961e-01 8.48442167e-02] | [13.32819652557373, 3.508892774581909] |
6d28bbec-cc7a-4a96-aaef-fff2331c5754 | uwaterloo-at-semeval-2017-task-8-detecting | null | null | https://aclanthology.org/S17-2080 | https://aclanthology.org/S17-2080.pdf | UWaterloo at SemEval-2017 Task 8: Detecting Stance towards Rumours with Topic Independent Features | This paper describes our system for subtask-A: SDQC for RumourEval, task-8 of SemEval 2017. Identifying rumours, especially for breaking news events as they unfold, is a challenging task due to the absence of sufficient information about the exact rumour stories circulating on social media. Determining the stance of Twitter users towards rumourous messages could provide an indirect way of identifying potential rumours. The proposed approach makes use of topic independent features from two categories, namely cue features and message specific features to fit a gradient boosting classifier. With an accuracy of 0.78, our system achieved the second best performance on subtask-A of RumourEval. | ['Olga Vechtomova', 'Hareesh Bahuleyan'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['rumour-detection'] | ['natural-language-processing'] | [-3.55628371e-01 4.57774214e-02 -3.61891031e-01 -3.01928163e-01
-7.57040322e-01 -2.57466972e-01 1.24928892e+00 5.83539128e-01
-2.57478595e-01 8.71069431e-01 5.67780972e-01 -3.32878143e-01
3.19998384e-01 -5.97369790e-01 -5.40785313e-01 -3.03822339e-01
-1.05299674e-01 3.66789252e-01 4.62884635e-01 -6.94390774e-01
8.81115913e-01 1.70964822e-01 -1.34422410e+00 7.81912565e-01
5.82921028e-01 9.82588172e-01 -9.46638733e-02 5.14080703e-01
-4.21740621e-01 1.52564013e+00 -9.12141442e-01 -2.97498554e-01
-3.28010917e-01 -6.04947150e-01 -1.20077324e+00 -2.01429147e-02
5.10337174e-01 -2.66629040e-01 -1.45973280e-01 8.44640255e-01
8.01225230e-02 -1.04493927e-02 6.83974028e-01 -1.16025710e+00
-2.91269451e-01 9.23523188e-01 -5.61241210e-01 9.84008253e-01
5.41903377e-01 -3.04130077e-01 9.19703484e-01 -1.10491598e+00
8.11996341e-01 1.19103920e+00 6.46593332e-01 2.22017586e-01
-1.05655241e+00 -7.69390404e-01 -1.90931395e-01 6.21847749e-01
-9.21142995e-01 -4.61596668e-01 8.21529806e-01 -9.36511695e-01
5.23266554e-01 3.57803583e-01 3.77458602e-01 1.48659551e+00
3.45910519e-01 8.04221690e-01 1.71798968e+00 1.03058800e-01
2.58445621e-01 6.54553533e-01 6.83113873e-01 2.06470355e-01
-1.11532725e-01 -2.03018695e-01 -9.78451073e-01 -6.98226511e-01
1.38435692e-01 -2.26489305e-01 -3.24691832e-01 3.05389255e-01
-9.32240725e-01 1.35173881e+00 5.99962056e-01 2.40390718e-01
-7.78559864e-01 -5.37162721e-01 7.26777494e-01 8.64347756e-01
1.21624601e+00 6.68568850e-01 -1.43255532e-01 -5.07877991e-02
-1.30925953e+00 7.06427336e-01 1.16119230e+00 3.09667766e-01
6.67874098e-01 -1.61178678e-01 -1.68819040e-01 1.11098766e+00
-7.04830959e-02 2.94577152e-01 5.20827591e-01 -1.55249998e-01
4.57216650e-01 2.84676313e-01 5.70066690e-01 -1.32072163e+00
-7.43951797e-01 -5.28431773e-01 -3.77216876e-01 -8.17968175e-02
2.77411968e-01 -4.41978514e-01 -4.52147335e-01 1.10385633e+00
4.18960541e-01 3.21129449e-02 -2.90026695e-01 1.14569354e+00
1.07605553e+00 8.43783796e-01 -9.43728015e-02 -4.15235549e-01
1.60518026e+00 -8.02305341e-01 -7.88904369e-01 -3.38065714e-01
1.98402002e-01 -1.13853502e+00 6.53705239e-01 4.35208172e-01
-9.08960640e-01 -7.86577687e-02 -1.10101604e+00 2.50665516e-01
-3.67051572e-01 -3.78198206e-01 3.41367245e-01 3.57448190e-01
-5.39717734e-01 5.24300039e-01 -1.96260676e-01 -3.81522387e-01
2.81369954e-01 -4.22480702e-01 7.61785880e-02 2.24265270e-02
-1.64834523e+00 1.15808535e+00 1.12643272e-01 -2.57644981e-01
-1.15654695e+00 -5.36510170e-01 -3.30253750e-01 -3.40537637e-01
3.53407830e-01 -1.68024853e-01 1.41516542e+00 -9.12181735e-01
-1.24585283e+00 9.74119127e-01 -2.15239927e-01 -8.47290397e-01
7.71201968e-01 -3.96203935e-01 -5.45171380e-01 5.91856055e-02
2.80855507e-01 -2.63432741e-01 1.25944984e+00 -9.52893496e-01
-6.41398847e-01 -3.38143110e-01 -1.40278071e-01 -1.04722809e-02
-5.19286431e-02 6.67890370e-01 5.63363016e-01 -3.81294698e-01
1.21196352e-01 -6.63753510e-01 1.65466636e-01 -1.08791137e+00
-5.52975774e-01 -2.45827153e-01 9.12648320e-01 -9.65245068e-01
1.18805873e+00 -1.72550428e+00 -3.39682460e-01 -7.12369150e-03
4.23970670e-01 -1.45828063e-02 4.81362909e-01 8.09688509e-01
1.97327271e-01 -1.08262964e-01 3.03864241e-01 5.66333048e-02
-3.54044527e-01 -4.21863824e-01 -9.11924660e-01 7.32970238e-01
-3.66577730e-02 4.23605412e-01 -1.15970492e+00 -2.33432740e-01
-3.01241368e-01 6.41620085e-02 -2.22501382e-02 3.27569962e-01
-3.35091561e-01 3.79829288e-01 -5.36938190e-01 1.95073441e-01
7.78939724e-01 -2.82557815e-01 -1.72433630e-01 3.01114887e-01
-4.46398258e-01 1.17621231e+00 -4.83092487e-01 6.32258654e-01
-4.10441369e-01 1.05703676e+00 1.57886446e-01 -5.75650990e-01
1.20302343e+00 4.22476083e-01 1.91374570e-01 -3.90492439e-01
3.00879240e-01 1.67580307e-01 -3.49400342e-01 -4.55815822e-01
7.88504124e-01 -4.74914670e-01 -2.05295041e-01 8.28329861e-01
-2.90988296e-01 3.50021459e-02 -1.33965835e-01 4.03393567e-01
1.14049888e+00 -5.60320616e-01 5.72515011e-01 -5.19084692e-01
4.27654058e-01 3.11622143e-01 4.84340005e-02 9.01861072e-01
-1.74332723e-01 5.36111951e-01 7.51401246e-01 -6.48907661e-01
-8.02953362e-01 -5.30448139e-01 -4.81850386e-01 1.47419631e+00
-9.15685520e-02 -2.05665827e-01 -5.04713356e-01 -7.04354703e-01
1.47420868e-01 1.12561774e+00 -9.71542060e-01 2.17341185e-01
-3.03835034e-01 -1.11849761e+00 4.31739062e-01 -1.95659801e-01
3.46867532e-01 -1.11732411e+00 -2.83956200e-01 3.85691881e-01
-7.27061391e-01 -1.02975702e+00 -2.77341247e-01 4.45798710e-02
-6.71500027e-01 -1.07909048e+00 -5.93939006e-01 -3.54978025e-01
4.01916616e-02 6.12376869e-01 1.09856224e+00 1.21943310e-01
2.15814918e-01 -3.88484180e-01 -4.82296258e-01 -6.21128738e-01
-8.58403206e-01 9.17672738e-02 -1.54383600e-01 1.92403167e-01
3.60585690e-01 -5.06143570e-01 -3.98139507e-01 5.67504346e-01
-4.01387513e-01 -6.98219314e-02 2.18465075e-01 7.30948865e-01
-2.44734466e-01 -5.53441703e-01 1.04837334e+00 -1.29933929e+00
1.11070251e+00 -1.48477566e+00 -4.71778587e-02 -5.99886000e-01
-6.19125485e-01 -4.61127102e-01 5.56869030e-01 -3.24525714e-01
-1.02601337e+00 -8.60542297e-01 -1.22911207e-01 1.46298617e-01
-2.14595288e-01 6.75165236e-01 6.71902597e-01 3.14799696e-01
1.29741383e+00 2.48096079e-01 1.22082829e-02 -6.21752262e-01
7.00340495e-02 1.17325282e+00 2.16022521e-01 -1.87402964e-02
3.98112088e-01 4.71534729e-01 -5.38895130e-01 -1.12237561e+00
-1.59865296e+00 -1.09654427e+00 -1.40001535e-01 -3.81323963e-01
3.83254677e-01 -9.57858622e-01 -5.01818240e-01 5.95471978e-01
-1.28666186e+00 4.10487950e-02 4.34232444e-01 2.04218730e-01
-3.19460183e-01 4.32034284e-02 -1.01136196e+00 -7.80767679e-01
-5.49286067e-01 -4.21495706e-01 3.19824845e-01 -2.11677894e-01
-6.49706483e-01 -7.05079913e-01 3.83829385e-01 7.06867278e-01
7.23453283e-01 2.99773037e-01 4.56220061e-01 -1.40294433e+00
-1.23281591e-01 -3.25087756e-01 -3.79114330e-01 -3.41035202e-02
-6.54687583e-02 -4.44716811e-01 -1.11062562e+00 -2.42437556e-01
3.96328092e-01 -7.54558563e-01 1.10425019e+00 3.39614935e-02
5.86145699e-01 -8.40488553e-01 -1.50690436e-01 -1.86680734e-01
8.79299462e-01 -5.08512616e-01 2.94620454e-01 7.95385361e-01
2.82775939e-01 6.13244832e-01 7.71514475e-01 9.04869914e-01
4.52539921e-01 7.40847409e-01 6.05166733e-01 3.12755853e-01
-6.20180666e-02 -1.51975870e-01 6.88419461e-01 7.90178776e-01
9.38601345e-02 4.47258763e-02 -9.11891639e-01 8.44749689e-01
-1.75195003e+00 -1.29800236e+00 -8.40422213e-01 2.01546550e+00
8.99159729e-01 3.19041073e-01 5.20786464e-01 2.19787061e-02
7.97585845e-01 7.35851407e-01 -8.61627705e-05 -7.89572299e-01
-1.32394684e-02 -4.94472176e-01 3.44604999e-01 5.70911169e-01
-1.11295962e+00 6.59018457e-01 6.04861784e+00 4.44875509e-01
-1.17388856e+00 6.40468359e-01 3.00969362e-01 -1.52707130e-01
-8.58560391e-03 -7.41894171e-02 -8.83748949e-01 7.14980543e-01
1.33076155e+00 -6.90567493e-02 2.80037969e-01 7.81528354e-01
5.87105215e-01 -3.24373692e-01 -6.47798896e-01 2.08950713e-01
5.35862386e-01 -1.35656297e+00 -1.57165453e-01 -1.16523348e-01
8.56045008e-01 6.58110619e-01 -1.70138121e-01 3.53826463e-01
2.98898220e-01 -6.90980613e-01 9.33314085e-01 3.69387805e-01
-1.02074377e-01 -8.73161197e-01 1.03987277e+00 8.72804701e-01
-1.33342654e-01 3.01393177e-02 -3.86869729e-01 -4.38647836e-01
3.24305296e-01 1.06202614e+00 -1.80706739e+00 8.42056647e-02
5.29548228e-01 6.56967878e-01 -3.70727956e-01 1.05294538e+00
-4.11597073e-01 1.32583690e+00 4.59875911e-02 -2.90052295e-01
5.03704309e-01 4.31321234e-01 1.06145239e+00 1.68599701e+00
-3.10075618e-02 -1.23544253e-01 1.98200822e-01 6.53250754e-01
-2.03339219e-01 3.73654813e-01 -3.86717975e-01 2.49967247e-01
4.38390106e-01 1.18140006e+00 -4.38819319e-01 -3.06057066e-01
-1.31711543e-01 7.81006098e-01 5.01773536e-01 -9.68665555e-02
-7.82160699e-01 -1.73776671e-02 3.98441643e-01 6.89249814e-01
1.73304796e-01 2.45568514e-01 -3.36068660e-01 -9.26704168e-01
-4.23213571e-01 -9.58565831e-01 3.58951688e-01 -4.90605503e-01
-1.53482580e+00 7.23753870e-01 -1.23397388e-01 -7.57533729e-01
-2.77649850e-01 -3.18829156e-03 -1.05636203e+00 8.83073032e-01
-1.59593534e+00 -9.98185754e-01 -4.41250324e-01 3.51475209e-01
6.91804290e-01 -8.89947545e-03 5.61779857e-01 -9.46212001e-03
-2.77479321e-01 1.30906656e-01 2.56298363e-01 -2.58508027e-02
1.01004577e+00 -1.17443585e+00 4.29827958e-01 2.90465355e-01
-1.68753535e-01 3.35773557e-01 1.62221694e+00 -8.10665667e-01
-8.14409733e-01 -9.97955322e-01 1.26431251e+00 -6.24500930e-01
1.28754771e+00 -9.29494053e-02 -1.37306190e+00 5.89642227e-01
4.11272854e-01 -5.31794846e-01 5.81929564e-01 6.72900438e-01
-4.85120386e-01 3.32567424e-01 -9.71741140e-01 1.98955595e-01
2.85458326e-01 -5.67874610e-01 -1.03515518e+00 7.10192680e-01
2.88935483e-01 -3.99047256e-01 -7.60437906e-01 -3.47113669e-01
3.44772667e-01 -1.08343530e+00 6.27386689e-01 -7.81004786e-01
1.01309097e+00 9.79007334e-02 2.02120945e-01 -1.71455753e+00
-2.29636654e-01 -5.69952190e-01 -1.33961618e-01 1.08871078e+00
5.08570135e-01 -7.57780194e-01 5.05409837e-01 -2.48036280e-01
1.81554094e-01 -4.83566880e-01 -7.51303673e-01 -4.95790541e-01
9.29952115e-02 -3.77523899e-01 3.73035550e-01 1.33653772e+00
5.70283592e-01 8.06543648e-01 -8.89313340e-01 6.94737583e-02
5.94603896e-01 4.73719120e-01 9.24954236e-01 -1.30931807e+00
4.87167090e-02 -4.39739227e-01 -9.97939408e-02 -8.92491400e-01
7.22335130e-02 -8.78973186e-01 2.29523957e-01 -1.09011471e+00
5.85207701e-01 -1.73454672e-01 -3.37431610e-01 1.07553214e-01
-3.24350506e-01 2.80004382e-01 1.85564067e-02 6.81173623e-01
-4.82358009e-01 2.95446664e-01 9.59492028e-01 -2.17502993e-02
-1.05429597e-01 5.03702879e-01 -6.41710520e-01 8.53411436e-01
9.25150275e-01 -8.59491110e-01 1.19826756e-01 3.25273454e-01
4.67265218e-01 4.02196646e-01 4.86105114e-01 -4.85815316e-01
-2.20139953e-03 -2.38020375e-01 -1.52211949e-01 -9.65287149e-01
2.54752010e-01 -8.28067511e-02 -2.43607759e-01 1.94739088e-01
-5.11377633e-01 -1.82986781e-01 7.95748383e-02 6.85861647e-01
-3.72704834e-01 -3.29307228e-01 7.95989811e-01 -9.03726444e-02
-4.27255780e-01 -1.76100403e-01 -1.00419354e+00 1.94352403e-01
6.87852502e-01 2.89362937e-01 -7.56479979e-01 -7.86568046e-01
-6.11571193e-01 1.66987389e-01 3.47635001e-01 5.51426589e-01
5.16360939e-01 -8.88144672e-01 -1.56894064e+00 -3.21214199e-01
1.19408511e-01 -7.65619338e-01 1.98473439e-01 1.34133935e+00
-1.14587404e-01 4.15892750e-01 -3.36486518e-01 -3.00016880e-01
-1.32542205e+00 4.33570981e-01 2.02329699e-02 -3.79991323e-01
-6.44070327e-01 6.50478899e-01 -1.05775319e-01 -4.14591916e-02
-3.00213516e-01 3.42039317e-01 -7.37639308e-01 6.74629986e-01
1.15486360e+00 7.92474687e-01 4.13095534e-01 -8.86385977e-01
-2.32826918e-01 -5.33147335e-01 -5.75389802e-01 -6.93742484e-02
1.26270390e+00 -4.58736867e-01 -3.28956932e-01 8.30877602e-01
1.08077168e+00 1.84825107e-01 -6.23561442e-01 -4.00329202e-01
3.74016911e-01 -5.07113218e-01 6.08681023e-01 -9.53036487e-01
-3.34591568e-01 6.96746707e-01 -2.66500175e-01 8.93140078e-01
2.72715867e-01 2.02209905e-01 1.03881896e+00 -2.93576308e-02
2.71091551e-01 -7.49498487e-01 -3.36982571e-02 9.04238760e-01
1.38260663e+00 -1.39695823e+00 2.45411187e-01 -1.88227192e-01
-7.86500990e-01 1.02227533e+00 2.29657695e-01 -2.06625223e-01
5.49572527e-01 -2.49554038e-01 2.20440626e-01 -6.15442991e-01
-9.88334656e-01 2.05234647e-01 4.69577134e-01 -6.13291860e-02
5.76253355e-01 3.15779150e-01 -8.72524142e-01 6.06234491e-01
-4.54453945e-01 -2.39109039e-01 1.05693030e+00 6.03784859e-01
-9.97712910e-01 -2.14539289e-01 -8.43686700e-01 6.55144691e-01
-1.09401226e+00 5.18386774e-02 -7.10022509e-01 5.81033230e-01
-3.99206191e-01 1.18406415e+00 -1.98129579e-01 -3.54352564e-01
2.76274588e-02 2.01221377e-01 -4.74097244e-02 -5.41981816e-01
-9.03181136e-01 -2.94051468e-01 8.75410318e-01 -4.78740901e-01
-1.93451002e-01 -9.89032865e-01 -7.19067156e-01 -7.91853547e-01
-5.09736717e-01 5.51440239e-01 8.35589826e-01 1.19338775e+00
-3.91957574e-02 1.06513858e-01 1.13137865e+00 -8.19111824e-01
-1.03380167e+00 -1.45218301e+00 -6.44070327e-01 5.42913377e-01
8.00361931e-01 -6.81189179e-01 -5.10260761e-01 -3.18768024e-01] | [8.22685718536377, 10.117347717285156] |
6c8c8a07-a249-4f34-bd9b-d3e12ab83f03 | semantic-enhanced-text-to-sql-parsing-via | 2208.03903 | null | https://arxiv.org/abs/2208.03903v1 | https://arxiv.org/pdf/2208.03903v1.pdf | Semantic Enhanced Text-to-SQL Parsing via Iteratively Learning Schema Linking Graph | The generalizability to new databases is of vital importance to Text-to-SQL systems which aim to parse human utterances into SQL statements. Existing works achieve this goal by leveraging the exact matching method to identify the lexical matching between the question words and the schema items. However, these methods fail in other challenging scenarios, such as the synonym substitution in which the surface form differs between the corresponding question words and schema items. In this paper, we propose a framework named ISESL-SQL to iteratively build a semantic enhanced schema-linking graph between question tokens and database schemas. First, we extract a schema linking graph from PLMs through a probing procedure in an unsupervised manner. Then the schema linking graph is further optimized during the training process through a deep graph learning method. Meanwhile, we also design an auxiliary task called graph regularization to improve the schema information mentioned in the schema-linking graph. Extensive experiments on three benchmarks demonstrate that ISESL-SQL could consistently outperform the baselines and further investigations show its generalizability and robustness. | ['Lijie Wen', 'Li Lin', 'Xuming Hu', 'Aiwei Liu'] | 2022-08-08 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [ 2.65088677e-01 2.94123054e-01 -2.37107366e-01 -8.54674101e-01
-8.17923248e-01 -6.06484354e-01 4.13779020e-01 5.42726696e-01
-1.94373041e-01 1.37238964e-01 2.57997423e-01 -4.67091233e-01
1.31409392e-01 -9.26643431e-01 -1.05289364e+00 2.23247018e-02
3.98102552e-01 4.73165274e-01 3.89260232e-01 -3.56942654e-01
2.00147077e-01 2.01294813e-02 -1.05176282e+00 5.63821733e-01
1.00147474e+00 8.70888591e-01 1.28912121e-01 -2.46899091e-02
-9.58123684e-01 7.52836347e-01 -3.81800383e-01 -8.49351466e-01
1.78459838e-01 -3.97044867e-01 -9.41879570e-01 1.22962236e-01
6.26359940e-01 7.91912749e-02 -1.93478152e-01 1.26159906e+00
1.66339502e-01 2.71904003e-02 6.59523904e-02 -1.12852311e+00
-6.72931492e-01 9.45557058e-01 -4.18510586e-01 -1.46404162e-01
6.71200514e-01 -1.43826455e-02 1.53836846e+00 -9.82868910e-01
5.65615356e-01 1.36960256e+00 6.40480876e-01 4.06899720e-01
-9.86500204e-01 -4.90079045e-01 3.75631154e-01 1.14171103e-01
-1.22971344e+00 -3.52171838e-01 9.52092886e-01 -2.02577084e-01
1.02663350e+00 2.86555052e-01 1.13833658e-01 8.90548289e-01
-2.30959401e-01 7.97199130e-01 6.86518550e-01 -2.15815648e-01
-1.79543961e-02 3.97013754e-01 3.90100151e-01 9.59031641e-01
2.42939472e-01 -3.01551610e-01 -4.40216929e-01 -1.74605384e-01
4.51832831e-01 -1.25587985e-01 -5.90276308e-02 -6.84595048e-01
-1.07927620e+00 8.45111668e-01 5.35733283e-01 1.16582483e-01
-2.92484015e-01 -3.19136493e-02 6.65868640e-01 3.68643314e-01
1.42182410e-01 5.52312672e-01 -4.86240536e-01 3.42275858e-01
-6.30586326e-01 1.92829683e-01 9.48450744e-01 1.48580134e+00
1.03840733e+00 -2.72259027e-01 1.22030091e-03 8.63988340e-01
4.98283595e-01 3.75073642e-01 3.52323860e-01 -6.24198079e-01
1.17537844e+00 1.27591205e+00 -1.79423153e-01 -1.05527902e+00
-4.15376246e-01 -3.26463282e-01 -4.41693068e-01 -7.23047078e-01
1.82202831e-01 5.24558946e-02 -5.92500210e-01 1.72545493e+00
5.45977294e-01 -6.06672242e-02 2.73039073e-01 7.45866120e-01
1.00648248e+00 5.91047585e-01 4.02708173e-01 4.36729491e-02
1.53606439e+00 -9.91346359e-01 -7.10193694e-01 -5.81014216e-01
8.78707290e-01 -6.47484183e-01 1.58692706e+00 -1.62563637e-01
-9.28342044e-01 -6.04734302e-01 -8.59721363e-01 -2.08741933e-01
-4.56628770e-01 -6.28842711e-02 7.35743523e-01 6.20340168e-01
-4.84687150e-01 1.12613037e-01 -6.31963134e-01 -6.33969486e-01
2.42050529e-01 2.39368323e-02 -3.19211632e-01 -7.45862350e-02
-1.36772084e+00 5.89307964e-01 9.53177452e-01 1.52487261e-02
-3.41256618e-01 -7.56063342e-01 -1.25092614e+00 1.75163373e-01
1.02512610e+00 -7.51421154e-01 1.13254297e+00 -8.95756721e-01
-1.20823848e+00 1.03345203e+00 -2.44360119e-01 -3.58370095e-01
1.03997946e-01 -1.42885715e-01 -5.11687100e-01 -1.41776055e-01
2.74313182e-01 2.06119061e-01 3.56065065e-01 -1.21000028e+00
-5.46451032e-01 -5.14788628e-01 2.86890835e-01 1.81307226e-01
-1.13119125e-01 1.64139038e-03 -9.47798193e-01 -5.58310151e-01
4.57297534e-01 -6.08337939e-01 -9.85858962e-02 -4.00210470e-01
-7.79739082e-01 -5.07067144e-01 2.86350906e-01 -6.97086751e-01
1.50666273e+00 -2.13377285e+00 -4.27210443e-02 3.91439945e-01
7.78991655e-02 2.00532377e-01 -1.30497187e-01 7.35434175e-01
-1.90465555e-01 9.26996991e-02 -2.84748286e-01 -5.77227287e-02
2.96930134e-01 1.76235870e-01 -5.09182155e-01 -3.74280959e-02
4.09542114e-01 1.17971897e+00 -7.63447762e-01 -5.63355803e-01
-2.82165229e-01 -1.56839952e-01 -7.43277371e-01 6.21151209e-01
-7.36303627e-01 -1.22207642e-01 -6.23192847e-01 5.51891387e-01
6.76505148e-01 -5.62017143e-01 5.67060292e-01 -6.70124233e-01
3.11380446e-01 6.88298523e-01 -1.21015811e+00 2.04317403e+00
-4.34119880e-01 -3.58697586e-02 -1.73392355e-01 -1.06208968e+00
1.23934388e+00 -3.32285650e-02 2.78579116e-01 -9.84274805e-01
-2.76367664e-01 2.49757588e-01 -3.37326795e-01 -9.36484337e-01
5.57466805e-01 -2.98310071e-01 -4.78763580e-01 2.90773183e-01
7.85794482e-02 -4.10018973e-02 2.99882255e-02 5.16130984e-01
8.86523485e-01 1.72103509e-01 2.92739451e-01 -4.72423136e-02
8.82688880e-01 9.13455561e-02 5.72149277e-01 6.39721155e-01
1.92982495e-01 2.00868905e-01 8.19695115e-01 -2.01052174e-01
-8.37467968e-01 -1.16674185e+00 2.54524380e-01 1.18996847e+00
3.66428792e-01 -6.42873824e-01 -8.29562485e-01 -1.25561965e+00
1.40918568e-01 9.76719856e-01 -3.36650789e-01 -3.90424222e-01
-8.03144276e-01 -3.66857588e-01 6.37387037e-01 5.87892830e-01
5.05830169e-01 -9.80975926e-01 -1.47094592e-01 3.41162473e-01
-3.46386105e-01 -1.46551692e+00 -7.92848825e-01 1.32512301e-01
-6.51822388e-01 -1.18925905e+00 -1.52996490e-02 -9.39007759e-01
6.08795941e-01 7.14915544e-02 1.21288061e+00 1.66342691e-01
2.25880891e-01 2.74612248e-01 -1.79658353e-01 -1.66231230e-01
-4.62497294e-01 4.32052970e-01 -7.10376024e-01 1.59198090e-01
7.29929864e-01 -1.89510554e-01 -4.07861173e-01 1.71289995e-01
-1.22591317e+00 1.47936568e-01 4.79292631e-01 6.34506285e-01
6.27550244e-01 -1.56444579e-01 6.23695612e-01 -1.78084433e+00
6.02314889e-01 -6.66689515e-01 -7.85781562e-01 8.09513628e-01
-6.57680631e-01 5.69135368e-01 6.07369721e-01 6.44975947e-03
-1.27320015e+00 5.37820272e-02 -2.79483140e-01 -6.97226003e-02
2.68307235e-02 9.70799327e-01 -8.19463909e-01 3.78603727e-01
4.71902281e-01 1.78479612e-01 -2.49717772e-01 -5.07929385e-01
5.89181125e-01 4.10261691e-01 7.16574490e-01 -7.22413898e-01
8.21418226e-01 2.23706141e-01 -2.32924506e-01 -3.31429273e-01
-1.06993675e+00 -6.12267673e-01 -4.29536313e-01 2.89345354e-01
7.60166049e-01 -8.30883741e-01 -5.94320893e-01 1.85913265e-01
-1.23550403e+00 -1.24388132e-02 -9.22279805e-02 3.60282771e-02
-4.20324743e-01 6.41242683e-01 -3.39896560e-01 -2.87139744e-01
-2.86226243e-01 -9.70365942e-01 1.13699174e+00 8.59376565e-02
-8.07660297e-02 -1.15556443e+00 1.10640720e-01 6.14391387e-01
1.93357155e-01 -5.85719347e-02 1.56730020e+00 -1.22933900e+00
-6.89606309e-01 -8.38031024e-02 -5.01851141e-01 3.06894183e-01
3.30621004e-01 -3.25187653e-01 -4.00256783e-01 -9.21996981e-02
1.31965615e-03 -2.82674044e-01 6.31675661e-01 -4.61063862e-01
1.04351819e+00 -4.59574133e-01 -1.70102000e-01 8.28539371e-01
1.69257677e+00 2.30799153e-01 4.77951765e-01 3.71265084e-01
9.96183872e-01 8.85769725e-01 6.38046205e-01 1.65785074e-01
8.40747118e-01 6.42519593e-01 3.62079442e-01 9.58039761e-02
3.65902372e-02 -8.37274551e-01 1.50660455e-01 1.01878154e+00
8.82732034e-01 -2.76478231e-01 -1.07721210e+00 4.42252785e-01
-1.80320370e+00 -4.99783933e-01 -1.17562279e-01 2.11267948e+00
1.03258896e+00 1.88730419e-01 -2.48745218e-01 -5.65303564e-01
7.01225519e-01 3.10940057e-01 -6.19755149e-01 -2.52838641e-01
-8.70984420e-02 2.27984086e-01 3.21157277e-01 4.80891407e-01
-9.58701491e-01 1.46313500e+00 5.06089783e+00 5.10840774e-01
-9.90510643e-01 -1.59455910e-02 1.56553492e-01 3.23435456e-01
-8.91924500e-01 3.90094399e-01 -1.00398827e+00 3.40334117e-01
7.84931123e-01 -2.54309744e-01 3.13017845e-01 7.51359344e-01
-1.61546841e-01 3.23480844e-01 -1.20814097e+00 9.13675725e-01
1.73427284e-01 -1.22095537e+00 4.40721303e-01 -5.67148745e-01
3.69135141e-01 -1.69471368e-01 -3.29096377e-01 5.33355236e-01
1.43807188e-01 -6.05486512e-01 5.24002612e-01 5.98515034e-01
4.61328894e-01 -5.67212522e-01 5.82689524e-01 1.91748589e-01
-1.24888062e+00 2.59183943e-01 -5.12238204e-01 5.57063937e-01
3.70581597e-02 2.66799420e-01 -8.96292984e-01 8.38174045e-01
4.04799968e-01 5.26479721e-01 -9.05608416e-01 6.36357427e-01
-3.66404474e-01 3.76642108e-01 -1.48817431e-02 -6.85000122e-02
3.32939088e-01 -2.20321506e-01 3.05649132e-01 1.35894859e+00
8.59031603e-02 4.39719968e-02 8.24173465e-02 1.25847733e+00
-3.95036042e-01 4.17822868e-01 -5.12386680e-01 -2.01299056e-01
5.53148866e-01 1.05164063e+00 -2.72446483e-01 -4.54080522e-01
-9.15091455e-01 8.63475263e-01 6.40352786e-01 5.02781928e-01
-8.31523120e-01 -5.41472256e-01 4.88318890e-01 1.37601808e-01
3.89019191e-01 -2.74471492e-02 -1.20988742e-01 -1.45728457e+00
4.82218534e-01 -1.26493263e+00 6.85615957e-01 -7.69637644e-01
-1.33216083e+00 5.36306024e-01 -5.92535548e-02 -6.34396911e-01
-2.73733363e-02 -3.24519098e-01 -5.07918119e-01 7.97468305e-01
-1.62167764e+00 -1.18812644e+00 -3.79131168e-01 8.13670933e-01
4.10728514e-01 -1.19756214e-01 4.71800029e-01 3.47909212e-01
-7.77603269e-01 8.64902675e-01 -3.87853891e-01 3.24016988e-01
7.15546668e-01 -1.25234544e+00 8.82946253e-01 8.92398298e-01
1.83825731e-01 1.06252265e+00 4.77541208e-01 -8.16029906e-01
-1.65948474e+00 -1.36042130e+00 1.19067407e+00 -2.87873864e-01
8.40986669e-01 -5.98062515e-01 -1.60609186e+00 1.07352614e+00
2.20019594e-01 -2.58664846e-01 6.24954224e-01 1.18347585e-01
-8.51834059e-01 -2.45393142e-01 -7.24426687e-01 5.16517341e-01
1.09943497e+00 -8.84971857e-01 -9.99011040e-01 3.80053550e-01
1.11066234e+00 -5.12590051e-01 -7.65032709e-01 4.04606640e-01
7.52192587e-02 -7.00061858e-01 7.62073100e-01 -1.12061656e+00
2.96313494e-01 -2.78342813e-01 -3.93712431e-01 -8.66702199e-01
1.08218499e-01 -5.84642708e-01 5.09051681e-02 1.62142324e+00
5.91022074e-01 -6.61869824e-01 8.71469438e-01 6.70450211e-01
-2.67983437e-01 -5.98095894e-01 -6.48695588e-01 -7.81556964e-01
-2.14116611e-02 -3.67819786e-01 1.08512378e+00 1.01163793e+00
9.24995914e-02 7.45191038e-01 6.26474321e-02 4.01312113e-01
4.91758466e-01 4.29340243e-01 9.38042819e-01 -7.93065190e-01
-2.84318894e-01 -3.09218794e-01 -1.06943592e-01 -1.44729853e+00
4.65030074e-01 -1.22030997e+00 2.68157944e-02 -1.56648266e+00
1.62996083e-01 -5.33133626e-01 -2.83820361e-01 3.87977004e-01
-6.62762165e-01 -7.26898849e-01 7.87578523e-02 1.76613629e-02
-9.01618123e-01 5.59582114e-01 8.78923118e-01 -2.48412699e-01
-1.38357058e-01 -1.22643523e-01 -1.02995718e+00 3.89135122e-01
6.51496887e-01 -6.04974031e-01 -6.33089960e-01 -6.39327586e-01
8.04804444e-01 3.11226547e-01 2.59684473e-01 -6.75099790e-01
4.83200520e-01 -3.11737418e-01 -4.25186783e-01 -4.81875271e-01
-1.27188951e-01 -9.28091705e-01 1.21375591e-01 1.44753993e-01
-7.01617241e-01 3.60093743e-01 1.24197476e-01 5.99294841e-01
-5.11146784e-01 -4.07004029e-01 4.62114543e-01 -2.05151603e-01
-9.50802028e-01 3.70503902e-01 3.06345493e-01 6.82549238e-01
7.22470939e-01 8.69793221e-02 -4.59239006e-01 -8.86861756e-02
-3.99096668e-01 4.41342890e-01 3.04843307e-01 6.77093625e-01
6.06598258e-01 -1.41185927e+00 -5.04902363e-01 3.51849943e-01
6.94850445e-01 2.15108022e-02 1.45943183e-03 6.09989226e-01
-4.64206189e-01 4.97458369e-01 2.27612287e-01 -4.05826092e-01
-9.75815713e-01 9.12661374e-01 4.19631004e-01 -3.45688522e-01
-4.04844254e-01 8.66298616e-01 4.95431542e-01 -9.24450874e-01
3.31442088e-01 -4.85910296e-01 -6.69353902e-02 -1.64705753e-01
3.47425044e-02 -1.00719988e-01 2.94471860e-01 -4.77710843e-01
-4.28008825e-01 4.28624570e-01 -3.29958618e-01 2.56230295e-01
1.06794131e+00 -3.01504046e-01 -4.13490564e-01 3.31366688e-01
1.49658203e+00 8.99964124e-02 -8.86516750e-01 -8.93616915e-01
7.98973143e-01 -3.22757781e-01 -5.14540613e-01 -5.79139769e-01
-1.17595112e+00 7.43348777e-01 1.27835155e-01 2.31829435e-01
1.00860322e+00 2.83749640e-01 1.13258004e+00 6.03706658e-01
1.29211053e-01 -8.81140709e-01 2.46811166e-01 6.86422110e-01
8.03847969e-01 -1.36202765e+00 -2.95650989e-01 -7.19839573e-01
-7.97349274e-01 1.13591301e+00 8.58318448e-01 1.26034155e-01
3.97775888e-01 -1.19040474e-01 2.42670059e-01 -6.58069789e-01
-7.54965186e-01 -2.30941221e-01 4.36395109e-01 2.83890694e-01
4.32298243e-01 -1.27802953e-01 -2.55467892e-01 8.84928048e-01
-1.57895058e-01 -3.52162808e-01 1.27290696e-01 7.99268603e-01
-3.54190677e-01 -1.26764154e+00 2.21710056e-01 3.17542672e-01
-3.89437407e-01 -3.94896954e-01 -4.57269460e-01 8.18345726e-01
-1.96978450e-01 8.07627380e-01 -7.02252164e-02 -3.74023169e-01
9.93853986e-01 3.24905962e-01 1.74697205e-01 -9.72742319e-01
-9.57503855e-01 -1.57785952e-01 1.28115982e-01 -8.18687022e-01
-6.81301877e-02 -5.27347445e-01 -1.53859496e+00 -2.11879276e-02
-3.77574146e-01 2.81497955e-01 5.69293559e-01 8.21416259e-01
4.38884437e-01 4.24717873e-01 5.89513898e-01 5.76625645e-01
-7.04015791e-01 -4.35426056e-01 -2.88954973e-01 9.12822366e-01
4.28456813e-02 -1.28727779e-01 -1.05755225e-01 5.04328273e-02] | [9.900267601013184, 7.922833442687988] |
30249d10-995c-4c00-bc39-6d64a4332e2a | anticipatory-thinking-challenges-in-open | 2306.13157 | null | https://arxiv.org/abs/2306.13157v1 | https://arxiv.org/pdf/2306.13157v1.pdf | Anticipatory Thinking Challenges in Open Worlds: Risk Management | Anticipatory thinking drives our ability to manage risk - identification and mitigation - in everyday life, from bringing an umbrella when it might rain to buying car insurance. As AI systems become part of everyday life, they too have begun to manage risk. Autonomous vehicles log millions of miles, StarCraft and Go agents have similar capabilities to humans, implicitly managing risks presented by their opponents. To further increase performance in these tasks, out-of-distribution evaluation can characterize a model's bias, what we view as a type of risk management. However, learning to identify and mitigate low-frequency, high-impact risks is at odds with the observational bias required to train machine learning models. StarCraft and Go are closed-world domains whose risks are known and mitigations well documented, ideal for learning through repetition. Adversarial filtering datasets provide difficult examples but are laborious to curate and static, both barriers to real-world risk management. Adversarial robustness focuses on model poisoning under the assumption there is an adversary with malicious intent, without considering naturally occurring adversarial examples. These methods are all important steps towards improving risk management but do so without considering open-worlds. We unify these open-world risk management challenges with two contributions. The first is our perception challenges, designed for agents with imperfect perceptions of their environment whose consequences have a high impact. Our second contribution are cognition challenges, designed for agents that must dynamically adjust their risk exposure as they identify new risks and learn new mitigations. Our goal with these challenges is to spur research into solutions that assess and improve the anticipatory thinking required by AI agents to manage risk in open-worlds and ultimately the real-world. | ['Leilani H. Gilpin', 'Dustin Dannenhauer', 'Adam Amos-Binks'] | 2023-06-22 | null | null | null | null | ['adversarial-robustness', 'autonomous-vehicles', 'management', 'starcraft'] | ['adversarial', 'computer-vision', 'miscellaneous', 'playing-games'] | [-5.91709465e-03 5.50664425e-01 -3.33316848e-02 -1.58817589e-01
-5.45854211e-01 -1.03366554e+00 1.02945316e+00 1.51163191e-01
-6.43102825e-01 4.00432765e-01 2.97200292e-01 -6.80497885e-01
-2.29474068e-01 -1.11715233e+00 -7.31142044e-01 -2.25438341e-01
-6.36152983e-01 7.16796041e-01 1.69420332e-01 -8.95244956e-01
2.34524637e-01 2.80016601e-01 -1.42974806e+00 -3.26352455e-02
8.19875240e-01 6.42660260e-01 -4.26141113e-01 7.59479284e-01
4.46356684e-01 1.06722236e+00 -8.27793002e-01 -8.10406983e-01
8.87330890e-01 3.65291893e-01 -5.50336182e-01 -5.68333745e-01
3.05170357e-01 -5.25452912e-01 -1.89266592e-01 1.12736177e+00
3.15301985e-01 1.11750692e-01 6.83649659e-01 -1.73842096e+00
-5.38709819e-01 7.66669810e-01 1.35658663e-02 1.75352141e-01
2.64324933e-01 1.11157620e+00 6.05256140e-01 1.66658357e-01
4.79598761e-01 1.75735092e+00 9.09580529e-01 7.31930017e-01
-1.13600647e+00 -8.32234263e-01 6.73557162e-01 2.32298598e-02
-8.43565464e-01 -1.69245392e-01 3.46449643e-01 -8.22174072e-01
1.16589558e+00 4.11744773e-01 6.51246607e-01 1.40118515e+00
4.40437973e-01 1.43930376e-01 1.16022038e+00 2.18128577e-01
5.41251063e-01 2.41741195e-01 -1.37860164e-01 2.61219293e-01
3.43656331e-01 1.18814397e+00 -2.28948861e-01 -6.64680123e-01
1.80992797e-01 -5.23804761e-02 2.38493517e-01 -4.03953977e-02
-9.31052625e-01 7.99739063e-01 5.50130188e-01 -1.77802205e-01
-4.76634234e-01 4.70593125e-01 4.44575936e-01 5.62561750e-01
2.79075235e-01 1.00899112e+00 -3.83491009e-01 -2.98492223e-01
-1.12555847e-01 8.55388641e-01 1.10862315e+00 5.93100727e-01
4.93668109e-01 1.93506658e-01 1.98916867e-01 1.04864724e-01
2.51364380e-01 6.57485604e-01 -3.64673021e-03 -1.19019341e+00
3.93326789e-01 4.50517148e-01 4.67541754e-01 -1.06381452e+00
-4.65073615e-01 -2.28909567e-01 -3.07481319e-01 1.27997208e+00
6.52228892e-01 -5.46610236e-01 -1.12972617e+00 2.08151388e+00
2.16043368e-01 1.49174809e-01 4.80629981e-01 6.95552528e-01
-2.67720014e-01 5.37564278e-01 7.16277242e-01 5.45158386e-02
1.37420273e+00 -4.49813515e-01 -3.14179212e-01 -8.31296861e-01
3.86771321e-01 -4.28592861e-01 1.06136727e+00 5.84624410e-01
-9.21252966e-01 -4.27600145e-02 -1.11783183e+00 4.38297838e-01
-7.31497526e-01 -1.24946606e+00 9.88086462e-01 9.59810317e-01
-6.99371636e-01 5.62923968e-01 -7.53276348e-01 -7.18770847e-02
5.99684656e-01 1.46339253e-01 -1.00791402e-01 3.54083210e-01
-1.70214200e+00 1.53638721e+00 1.05007678e-01 -3.22833866e-01
-1.76067138e+00 -1.18622625e+00 -6.69849753e-01 -4.86686267e-02
9.13061380e-01 -6.12961411e-01 1.39207315e+00 -1.22641850e+00
-1.01029670e+00 5.55504262e-01 5.99659503e-01 -1.00231028e+00
9.41624939e-01 -2.04186112e-01 -3.98595929e-01 -3.57258767e-01
4.47981730e-02 4.71682400e-01 9.02394176e-01 -1.49112546e+00
-5.49799740e-01 -2.46376157e-01 7.90574074e-01 3.59084368e-01
9.64934658e-03 2.85321534e-01 7.12464631e-01 -7.87635624e-01
-6.65972769e-01 -9.98163879e-01 -9.00187910e-01 -1.47183657e-01
-3.86557691e-02 6.78907111e-02 4.21975285e-01 -4.23450828e-01
1.01536953e+00 -1.87264681e+00 -1.24729402e-01 4.46820557e-01
2.20336884e-01 3.47236633e-01 4.49444950e-02 5.64839363e-01
1.08189382e-01 5.92649519e-01 -1.85000554e-01 9.98449996e-02
5.68451762e-01 7.17767775e-02 -8.46018553e-01 3.94794457e-02
2.06265807e-01 6.25457168e-01 -1.10194659e+00 6.67902529e-02
1.30212784e-01 1.12644896e-01 -6.20910645e-01 3.60969305e-01
-7.18807817e-01 3.13838455e-03 -2.67972857e-01 5.02400875e-01
5.15304863e-01 6.04948282e-01 -5.22258645e-03 3.15667212e-01
-1.04136653e-01 1.78233206e-01 -1.02525270e+00 7.42328942e-01
-4.78406161e-01 6.97843209e-02 2.68840551e-01 -3.77568603e-01
4.12396938e-01 6.32978603e-02 3.70871555e-03 -5.22826493e-01
2.91550010e-01 2.96681118e-03 3.77549917e-01 -3.92239451e-01
3.03988248e-01 -5.97412646e-01 -6.17422521e-01 6.91341937e-01
-4.41422015e-01 -7.11628675e-01 -1.57581806e-01 3.88110131e-01
1.36023033e+00 -3.70053858e-01 1.92178458e-01 -3.06444019e-01
-3.01500335e-02 4.09607559e-01 8.54749382e-01 9.96442854e-01
-7.11030126e-01 -4.14938433e-03 7.45177150e-01 -9.00741696e-01
-8.31457973e-01 -1.51122332e+00 1.79151416e-01 1.03053737e+00
2.17446387e-01 -1.95597395e-01 -9.74192202e-01 -8.99179399e-01
4.02078956e-01 1.28801000e+00 -7.69679785e-01 -7.32394576e-01
-4.38132554e-01 -6.55931115e-01 8.77221942e-01 5.09288073e-01
2.83398569e-01 -8.55361998e-01 -1.05579031e+00 9.89539698e-02
1.34794340e-01 -4.24600661e-01 -3.97331297e-01 5.08027412e-02
-1.74978510e-01 -1.18460631e+00 -5.01218028e-02 -5.51858395e-02
3.05053622e-01 -8.62896591e-02 1.22109830e+00 2.04809278e-01
-2.41571590e-01 6.00854814e-01 1.00133933e-01 -1.35613418e+00
-9.09510434e-01 -7.25530207e-01 4.84626710e-01 -4.02164161e-01
3.92010093e-01 -7.89619625e-01 -6.49658918e-01 5.99606395e-01
-7.76839435e-01 -5.16826987e-01 3.95269841e-01 4.60681528e-01
1.27534851e-01 2.08813474e-01 6.84988558e-01 -8.65689516e-01
9.14624393e-01 -9.16531384e-01 -9.27917302e-01 2.62813300e-01
-7.71323264e-01 -2.05560818e-01 6.52792692e-01 -9.19592202e-01
-1.31016624e+00 -3.12371910e-01 1.78004384e-01 -1.45607263e-01
-2.80569881e-01 1.41124174e-01 -2.73949891e-01 -7.90113360e-02
1.42177129e+00 -6.11901820e-01 1.28708363e-01 1.06389277e-01
6.27431035e-01 3.47798228e-01 3.97233903e-01 -8.04381788e-01
1.37301207e+00 5.60866833e-01 -1.10616408e-01 -1.46795094e-01
-5.83729684e-01 2.90807128e-01 8.48064572e-02 -3.83459389e-01
8.93962145e-01 -8.93262923e-01 -1.09011233e+00 7.04678297e-01
-6.73351467e-01 -9.54522252e-01 -3.96023005e-01 6.17282614e-02
-7.42443740e-01 -4.55438830e-02 -2.85738915e-01 -1.09901690e+00
8.52297843e-02 -1.17254615e+00 4.25161123e-01 2.25086734e-01
-5.33080935e-01 -9.94615972e-01 1.86283007e-01 5.17414987e-01
8.21717143e-01 3.82644892e-01 9.40740705e-01 -9.00656521e-01
-8.31054688e-01 -2.37555444e-01 3.00666571e-01 2.50765234e-01
-1.60387635e-01 -3.44377279e-01 -1.18359125e+00 -8.01131353e-02
6.96496740e-02 -5.25109291e-01 5.28349876e-01 -5.06421328e-02
6.62100434e-01 -7.35798359e-01 -2.48527557e-01 3.17316025e-01
9.79607165e-01 6.50138497e-01 5.85868537e-01 6.87779248e-01
3.49409908e-01 1.09063137e+00 9.81908321e-01 2.36439615e-01
7.10788429e-01 4.23979402e-01 8.99027169e-01 2.02619210e-01
2.93387026e-01 -4.32566822e-01 6.61537826e-01 -5.49302638e-01
-3.42089981e-02 -1.47435293e-01 -1.34386885e+00 3.99239272e-01
-1.70959044e+00 -1.43024850e+00 4.49973047e-01 2.33875585e+00
8.34363520e-01 7.34180808e-01 4.01568830e-01 -3.07765722e-01
2.68780798e-01 1.05104767e-01 -9.22815204e-01 -8.14789653e-01
1.40072092e-01 -4.14712459e-01 7.76917517e-01 9.60340679e-01
-1.24145079e+00 1.02394438e+00 6.89242697e+00 5.82246244e-01
-6.99919939e-01 9.30524021e-02 9.87340868e-01 -2.96396434e-01
-7.39916801e-01 3.34819794e-01 -3.97506267e-01 3.90940160e-01
1.14188004e+00 -6.04695737e-01 7.59126663e-01 9.68145847e-01
1.27423674e-01 -2.64938504e-01 -1.36533821e+00 3.51190627e-01
-2.88000312e-02 -1.13596153e+00 -9.16277990e-02 5.34499921e-02
6.60477698e-01 -7.63733163e-02 6.29442692e-01 5.77287495e-01
1.37743890e+00 -1.50976098e+00 1.10589290e+00 5.99580228e-01
2.07253307e-01 -9.39635217e-01 5.67990780e-01 5.38218319e-01
-6.17857754e-01 -7.57311642e-01 -2.58677285e-02 -5.72034597e-01
3.57243091e-01 8.84013921e-02 -7.88969338e-01 -1.85655788e-01
8.08158159e-01 2.72659361e-02 -3.04996371e-01 4.41585124e-01
-1.15976654e-01 3.84505898e-01 -4.32233602e-01 1.36620879e-01
2.70447284e-01 5.41203469e-03 1.01504314e+00 6.21548116e-01
-1.99527398e-01 3.52967441e-01 3.47664982e-01 9.37310874e-01
3.99308562e-01 -6.89678371e-01 -1.02266383e+00 -8.76908079e-02
6.69427454e-01 7.20709622e-01 -2.69506723e-01 -1.97286040e-01
-8.84650275e-03 2.29346529e-01 6.78554177e-02 2.02088103e-01
-8.68373752e-01 -1.15608849e-01 1.68153262e+00 7.88986236e-02
-6.24775887e-01 -6.10823743e-02 -3.70621413e-01 -7.50648916e-01
-1.73057988e-01 -1.60863554e+00 5.95145226e-01 -3.80876839e-01
-1.58032727e+00 4.23753351e-01 3.82934660e-01 -1.10340750e+00
-4.74968255e-01 -4.59998339e-01 -9.28144157e-01 8.04583251e-01
-1.22757018e+00 -1.40397084e+00 1.97534740e-01 3.17824066e-01
2.79525667e-01 -3.48977685e-01 6.72769189e-01 -1.88832596e-01
-9.81435403e-02 6.43211603e-01 -5.35937369e-01 -2.71937758e-01
7.14595020e-01 -1.28416538e+00 9.18074429e-01 9.87869859e-01
-4.15273935e-01 7.82471001e-01 9.34627295e-01 -1.17137897e+00
-1.26886785e+00 -1.03765166e+00 2.45974228e-01 -1.20424402e+00
1.07351816e+00 -5.32714546e-01 -7.65854120e-01 8.93906772e-01
-7.40538314e-02 -5.48106849e-01 5.02302587e-01 1.89062566e-01
-8.05933833e-01 -6.45298213e-02 -1.62792015e+00 1.25060654e+00
1.23179102e+00 -5.43999672e-01 -8.82467270e-01 3.44802588e-01
1.18196380e+00 -2.04447225e-01 -6.43197596e-01 3.80165428e-01
3.84069741e-01 -1.04673755e+00 1.30294216e+00 -8.62532735e-01
-1.36464378e-02 -3.00397187e-01 -9.24664140e-02 -1.67072058e+00
-7.49931708e-02 -1.14545584e+00 4.16094929e-01 1.05391979e+00
4.92569387e-01 -1.24786782e+00 5.49270928e-01 1.35146928e+00
-8.59594941e-02 -2.99748391e-01 -7.34281480e-01 -8.50227475e-01
5.36118209e-01 -9.61023748e-01 9.35962081e-01 8.51050794e-01
1.25907183e-01 -4.86039788e-01 -2.04857215e-01 7.25165844e-01
1.00805974e+00 -5.47802806e-01 8.27821791e-01 -9.75642562e-01
-3.68233830e-01 -6.20480478e-01 -5.67677438e-01 1.85367558e-02
9.67487469e-02 -4.94627386e-01 1.25814110e-01 -8.87603641e-01
-1.66225463e-01 -9.51005161e-01 -3.79188210e-02 4.61362869e-01
-2.45793611e-01 -2.09741011e-01 5.49995542e-01 -8.06135759e-02
-1.69361874e-01 7.35597089e-02 7.16047943e-01 -4.71259385e-01
-6.80734590e-02 2.48445049e-01 -1.21747124e+00 1.39234817e+00
9.25962210e-01 -3.61486703e-01 -8.53926063e-01 -3.50826085e-01
7.17098892e-01 -1.88441455e-01 9.22352135e-01 -1.13248944e+00
1.33956239e-01 -1.05296552e+00 -1.48896620e-01 1.06899150e-01
4.15331602e-01 -9.71370578e-01 3.57128173e-01 7.35495031e-01
-4.41443741e-01 1.26944512e-01 5.15023172e-01 7.37783432e-01
2.91552514e-01 -3.40267606e-02 7.10166812e-01 -4.11885500e-01
-7.69574642e-01 2.38298401e-01 -8.82484019e-01 3.68976951e-01
1.60048747e+00 1.70973703e-01 -9.53404069e-01 -7.38969684e-01
-7.61004150e-01 7.94768810e-01 7.48981953e-01 5.79686821e-01
4.27306443e-01 -7.93770492e-01 -4.81890649e-01 1.55695409e-01
6.49746731e-02 -5.20303771e-02 4.40932721e-01 6.40627220e-02
-3.73872966e-01 -2.44937390e-01 -4.04056638e-01 1.20481640e-01
-9.88058507e-01 9.78043437e-01 8.97100031e-01 -8.70823041e-02
-2.50671506e-01 1.11223280e+00 5.80576897e-01 -6.24431610e-01
5.31081080e-01 -1.56791613e-01 -8.59443843e-02 -3.23057622e-02
8.24038625e-01 4.30019408e-01 -4.22220320e-01 -3.68026197e-01
-3.26958984e-01 2.19635710e-01 -8.19926336e-02 -3.48023891e-01
1.08546329e+00 1.40731633e-01 2.16698572e-02 1.23542726e-01
2.98960149e-01 5.02059385e-02 -1.58236897e+00 3.17249477e-01
8.89523178e-02 -7.92267144e-01 -2.90904135e-01 -1.47204232e+00
-3.39040488e-01 6.65342629e-01 7.17169344e-01 4.48270738e-01
8.33461046e-01 -1.81165606e-01 6.26689970e-01 5.63993037e-01
7.20345080e-01 -1.28888702e+00 1.70817733e-01 4.73854363e-01
1.07434654e+00 -1.30955338e+00 -2.36521408e-01 -1.81866497e-01
-1.08384442e+00 3.59494627e-01 1.01144540e+00 -1.89413920e-01
7.46017575e-01 6.56466484e-01 6.05605602e-01 -1.54140562e-01
-1.12346435e+00 -1.92913651e-01 -2.36525878e-01 1.33365428e+00
-5.34939826e-01 5.55094242e-01 3.86025995e-01 7.45057046e-01
-4.86904979e-01 -5.07241607e-01 7.19749987e-01 7.43571877e-01
-4.41090912e-01 -1.06118047e+00 -6.51168227e-01 3.28184158e-01
-3.48157883e-01 8.64180326e-02 -5.39496303e-01 6.18224263e-01
4.27628458e-01 1.32297838e+00 -4.80175875e-02 -5.85254252e-01
6.13670051e-01 -8.28507319e-02 -1.61358386e-01 -4.93697703e-01
-1.12438309e+00 -7.36637890e-01 5.06946623e-01 -9.31105793e-01
3.31098586e-01 -6.64774418e-01 -1.10677052e+00 -6.05166256e-01
2.39200264e-01 -4.40249555e-02 4.56249624e-01 5.74725449e-01
2.77083308e-01 2.58709490e-01 6.30926728e-01 -1.02301240e+00
-1.29524314e+00 -3.77700090e-01 5.71200624e-02 5.82611442e-01
4.72659618e-01 -7.38822162e-01 -8.34273160e-01 -3.25710297e-01] | [4.538244247436523, 2.1346216201782227] |
76103265-dcc5-45ad-a5ee-d7f8682c3681 | recurrent-attention-networks-for-long-text | 2306.06843 | null | https://arxiv.org/abs/2306.06843v1 | https://arxiv.org/pdf/2306.06843v1.pdf | Recurrent Attention Networks for Long-text Modeling | Self-attention-based models have achieved remarkable progress in short-text mining. However, the quadratic computational complexities restrict their application in long text processing. Prior works have adopted the chunking strategy to divide long documents into chunks and stack a self-attention backbone with the recurrent structure to extract semantic representation. Such an approach disables parallelization of the attention mechanism, significantly increasing the training cost and raising hardware requirements. Revisiting the self-attention mechanism and the recurrent structure, this paper proposes a novel long-document encoding model, Recurrent Attention Network (RAN), to enable the recurrent operation of self-attention. Combining the advantages from both sides, the well-designed RAN is capable of extracting global semantics in both token-level and document-level representations, making it inherently compatible with both sequential and classification tasks, respectively. Furthermore, RAN is computationally scalable as it supports parallelization on long document processing. Extensive experiments demonstrate the long-text encoding ability of the proposed RAN model on both classification and sequential tasks, showing its potential for a wide range of applications. | ['Qing Li', 'Fu Lee Wang', 'Yingbin Zhao', 'Xing Lee', 'Haoran Xie', 'Xiaotian Luo', 'Zongxi Li', 'Xianming Li'] | 2023-06-12 | null | null | null | null | ['chunking'] | ['natural-language-processing'] | [ 2.89611042e-01 2.18424454e-01 -4.72855270e-01 -2.43426323e-01
-4.63705838e-01 9.54484008e-03 4.68983263e-01 3.51051211e-01
-5.27562141e-01 3.98097247e-01 4.58605409e-01 -3.61755192e-01
-4.56852354e-02 -8.65487397e-01 -3.35220337e-01 -6.52104080e-01
-5.00429608e-03 2.29697019e-01 2.42837772e-01 4.47558910e-02
5.90265870e-01 8.84743929e-02 -1.94618702e+00 5.10699213e-01
1.03628278e+00 1.04994035e+00 7.41837621e-01 7.60954499e-01
-8.86851013e-01 1.06207919e+00 -6.20303810e-01 -2.76050776e-01
-1.31490901e-01 -3.05941552e-01 -1.08678889e+00 -2.57602390e-02
-1.26833871e-01 -1.54811740e-01 -3.41648012e-01 7.99017549e-01
4.39067513e-01 3.28801066e-01 2.55788267e-01 -7.69710362e-01
-8.62638474e-01 9.68123257e-01 -7.26678133e-01 3.98594379e-01
2.41689101e-01 -5.33701420e-01 1.44220495e+00 -8.98176193e-01
1.75107032e-01 9.77085590e-01 7.14959919e-01 3.03573549e-01
-5.33910036e-01 -4.50833768e-01 5.23572683e-01 4.05439913e-01
-1.13413513e+00 -3.27434987e-01 5.36290705e-01 -1.46292657e-01
1.66875625e+00 2.43985355e-01 6.69124544e-01 6.16387427e-01
2.77573049e-01 1.35203469e+00 5.84705949e-01 -7.30415165e-01
8.94135684e-02 -1.81003332e-01 8.32467735e-01 4.93711442e-01
3.36578578e-01 -3.99098217e-01 -5.53253770e-01 2.32241470e-02
4.70405996e-01 3.67164314e-01 -2.00103596e-02 2.39353985e-01
-9.48157072e-01 7.72907734e-01 3.76307666e-01 5.98993897e-01
-5.28583407e-01 2.88254600e-02 8.42743158e-01 1.61097646e-01
6.61128104e-01 2.55967021e-01 -2.43676096e-01 -2.75925905e-01
-1.04151642e+00 -4.07145321e-02 5.81391692e-01 1.14156330e+00
4.87101674e-01 2.71094926e-02 -4.30050462e-01 9.28851187e-01
2.41629817e-02 1.39525384e-01 1.36583817e+00 -1.72186315e-01
5.62361956e-01 7.72816956e-01 -1.61070809e-01 -8.99512470e-01
-5.37207901e-01 -5.82348228e-01 -1.00819981e+00 -5.13535142e-01
-1.64994761e-01 7.15081096e-02 -9.03431952e-01 1.37040579e+00
2.53807396e-01 -1.57626886e-02 1.60993993e-01 6.36818051e-01
7.69675255e-01 9.06700015e-01 2.95006037e-01 -3.38849425e-01
1.61059380e+00 -1.46551275e+00 -8.99773896e-01 -3.21627736e-01
8.79859746e-01 -5.70080638e-01 1.14737809e+00 1.33941546e-01
-1.19141114e+00 -7.05383599e-01 -1.15268302e+00 -4.48197335e-01
-4.58756775e-01 2.15330690e-01 7.57283151e-01 5.66055894e-01
-8.14315438e-01 4.76352602e-01 -9.07417536e-01 -3.75597179e-01
3.47096771e-01 4.94289309e-01 4.42033671e-02 2.31700987e-01
-1.19912028e+00 5.58409512e-01 7.05881894e-01 1.79850698e-01
-8.75783414e-02 -4.64696527e-01 -8.31710458e-01 7.75421858e-01
1.81851417e-01 -6.43481374e-01 1.39820397e+00 -1.05334997e+00
-1.61386037e+00 5.91062307e-01 -5.59021533e-01 -6.37234271e-01
-1.16056278e-01 -6.06484830e-01 -4.81652230e-01 3.22250612e-02
1.75981447e-02 3.16937804e-01 9.82417583e-01 -4.01737273e-01
-1.03082049e+00 -5.33212423e-01 -2.24938840e-01 5.45400321e-01
-1.04854298e+00 2.12173045e-01 -6.04517519e-01 -9.35031772e-01
3.23321186e-02 -5.58999956e-01 -1.54691413e-01 -7.94791579e-01
-3.34043145e-01 -4.75271314e-01 9.56325889e-01 -4.96671647e-01
1.83888495e+00 -2.16609192e+00 -1.97550207e-02 -1.65526986e-01
1.38789132e-01 4.72007424e-01 -1.59877181e-01 6.25183165e-01
-6.45989627e-02 -2.21064668e-02 -7.79049322e-02 -4.72063601e-01
-1.95545390e-01 1.94768697e-01 -5.97516418e-01 5.93543909e-02
1.10442571e-01 1.20368063e+00 -7.23479927e-01 -5.95116735e-01
-9.13746357e-02 1.96129084e-01 -5.07533550e-01 3.64314348e-01
-1.46771640e-01 -2.74237871e-01 -5.89115322e-01 2.95838416e-01
5.42443573e-01 -3.85469258e-01 2.13597298e-01 2.63306648e-01
-2.03899294e-01 6.77616596e-01 -8.73275876e-01 1.53189075e+00
-7.11862981e-01 4.45781291e-01 6.17547110e-02 -1.16201091e+00
8.99249554e-01 1.87521175e-01 3.29821408e-01 -9.44302320e-01
6.19763732e-02 6.23057745e-02 -2.48015389e-01 -5.93869269e-01
1.20274842e+00 -1.80366665e-01 -1.75657526e-01 9.84190345e-01
1.15533926e-01 4.55152869e-01 3.10508043e-01 1.83902130e-01
9.10331428e-01 -1.04500614e-01 5.27512312e-01 -3.18382591e-01
6.42485857e-01 -7.40009174e-02 4.36488837e-01 7.46201932e-01
2.42620140e-01 3.43799144e-01 2.10415497e-01 -5.38798213e-01
-8.62773776e-01 -3.69439244e-01 3.71052437e-02 1.86967909e+00
1.75523430e-01 -7.83754706e-01 -7.86763728e-01 -7.28471041e-01
6.98437244e-02 4.99900222e-01 -6.19261026e-01 -1.60746798e-01
-7.16650963e-01 -8.00062001e-01 5.97828567e-01 9.17317688e-01
4.03081179e-01 -1.32969069e+00 -9.44908977e-01 5.43306649e-01
-1.77122802e-01 -8.03270936e-01 -5.90904295e-01 2.84888983e-01
-1.02055526e+00 -7.05735743e-01 -7.11762786e-01 -1.02399099e+00
4.64887530e-01 7.37141550e-01 9.56326425e-01 2.58765131e-01
-2.48683572e-01 -1.13469973e-01 -7.58526564e-01 -5.51923990e-01
-1.38002500e-01 6.65409267e-01 -2.29707956e-01 2.06010714e-02
6.46872342e-01 -4.26868558e-01 -4.20667499e-01 9.48585346e-02
-9.23086226e-01 2.43509769e-01 8.27688038e-01 1.08037782e+00
4.32781965e-01 2.22314328e-01 9.76868510e-01 -1.06116354e+00
8.25090706e-01 -5.47004521e-01 -3.19718361e-01 4.08628136e-01
-8.79618526e-01 1.66584477e-01 9.13280070e-01 -1.31767422e-01
-1.41300797e+00 -3.96103024e-01 -2.73168683e-01 5.52070774e-02
1.34250551e-01 8.27112615e-01 1.16892211e-01 4.74760592e-01
1.84661135e-01 6.09077394e-01 3.23916636e-02 -6.99535370e-01
2.45175734e-01 1.25190473e+00 3.95660698e-01 -3.97282213e-01
8.67747441e-02 2.75740117e-01 -4.30951834e-01 -1.09723961e+00
-1.21776772e+00 -9.53361690e-01 -4.82741743e-01 3.25116515e-01
7.04924822e-01 -8.18660200e-01 -5.45062482e-01 4.96332079e-01
-1.05456161e+00 -9.47016254e-02 -3.27167720e-01 3.85089546e-01
-4.01617587e-01 6.40989006e-01 -8.26375306e-01 -8.82358432e-01
-1.12172937e+00 -7.10613728e-01 1.18695426e+00 2.34753340e-01
-2.95168638e-01 -9.88910496e-01 -9.45459083e-02 1.24485843e-01
4.92882162e-01 -6.65561318e-01 1.10435188e+00 -8.71180415e-01
-1.89519852e-01 -3.88952613e-01 -4.18808848e-01 8.25451612e-02
3.85420360e-02 -4.07178640e-01 -1.03352880e+00 -3.36811185e-01
1.75550431e-01 -3.26486796e-01 1.06550395e+00 3.11222911e-01
1.53125453e+00 -3.33865166e-01 -3.44466656e-01 5.61238110e-01
1.17662036e+00 1.33356333e-01 6.55336320e-01 4.97282445e-01
7.49698997e-01 5.55695057e-01 7.52356887e-01 5.79729617e-01
3.14477324e-01 5.20035028e-01 3.18878144e-02 -5.62032796e-02
-1.83993559e-02 -2.76147515e-01 3.65533888e-01 1.36775959e+00
-9.88878123e-03 -3.52468640e-01 -7.82774568e-01 5.69286883e-01
-2.05122423e+00 -1.17257893e+00 -5.35368454e-03 2.04005480e+00
6.58930182e-01 7.83253312e-02 -7.56205199e-03 3.16587418e-01
7.68359125e-01 3.36631656e-01 -4.10840362e-01 -8.54591191e-01
1.35968253e-01 2.57699430e-01 2.58191079e-01 1.97530106e-01
-1.03139544e+00 1.02095866e+00 6.58241415e+00 1.08365130e+00
-1.09294438e+00 8.58662128e-02 4.24628526e-01 -1.24128826e-01
-1.87392280e-01 -2.55279005e-01 -1.24891078e+00 3.81361812e-01
1.30954826e+00 -5.31579792e-01 -3.60976197e-02 9.76394415e-01
4.80685122e-02 1.08080627e-02 -7.66781211e-01 7.19286263e-01
1.75790042e-01 -1.41374075e+00 3.53403479e-01 -1.16954066e-01
4.48070228e-01 4.16714735e-02 -1.20259225e-01 5.89103401e-01
-8.59865099e-02 -9.73236799e-01 7.64340401e-01 3.68808880e-02
7.44940817e-01 -8.94733131e-01 1.03298736e+00 5.70587456e-01
-1.38589346e+00 -4.43006635e-01 -5.31887591e-01 -2.72446364e-01
1.60049513e-01 6.52210057e-01 -6.65461481e-01 8.64610076e-01
6.81624413e-01 8.11862350e-01 -4.41152930e-01 6.60693824e-01
8.01037550e-02 5.73648810e-01 -2.07876757e-01 -2.66474277e-01
5.33699274e-01 -8.14454705e-02 2.30228513e-01 1.70643497e+00
3.26560080e-01 -1.30839195e-04 2.17830777e-01 3.03122073e-01
6.44668099e-03 4.59396511e-01 -2.29624212e-01 -2.61120826e-01
4.02568549e-01 1.12082696e+00 -8.83542001e-01 -6.15433097e-01
-5.99820316e-01 9.31505442e-01 6.66892886e-01 8.77384767e-02
-7.23233223e-01 -9.43105280e-01 3.44043583e-01 -4.79876734e-02
7.34832525e-01 -9.48478281e-02 -6.89736426e-01 -1.15002060e+00
2.14820594e-01 -6.56119764e-01 7.03378975e-01 -4.04084384e-01
-8.37124228e-01 9.05625403e-01 -2.73421556e-01 -1.00896537e+00
-3.14580351e-01 -3.46164614e-01 -6.80847883e-01 8.09505343e-01
-1.64108396e+00 -1.11579061e+00 -2.46091515e-01 4.19363528e-01
1.07818723e+00 -5.29029593e-02 9.80762601e-01 2.75642462e-02
-8.99614453e-01 7.59014010e-01 3.18773955e-01 -1.20706677e-01
3.16666394e-01 -1.12496114e+00 5.96415937e-01 8.02083135e-01
2.01598436e-01 8.74265015e-01 2.05189720e-01 -5.34258544e-01
-1.41461170e+00 -1.17971563e+00 1.10611868e+00 1.21365182e-01
5.93913198e-01 -3.42648357e-01 -1.17165375e+00 6.52618110e-01
3.05764884e-01 -2.47327849e-01 9.42669094e-01 5.57634652e-01
-3.11526269e-01 2.50437465e-02 -6.27227485e-01 4.59834993e-01
8.79001975e-01 -4.83282387e-01 -9.60600972e-01 1.52609527e-01
8.10965955e-01 -1.59524798e-01 -6.34370923e-01 3.78247015e-02
6.35024607e-01 -8.43158722e-01 6.07337654e-01 -4.46171582e-01
5.28723121e-01 8.53400528e-02 2.13675812e-01 -9.35127795e-01
-6.53275251e-01 -6.78948760e-01 -6.26889646e-01 1.25969803e+00
2.78305113e-01 -5.99476218e-01 8.84483278e-01 1.50885284e-01
-4.86924708e-01 -1.03765380e+00 -6.77596450e-01 -8.70630980e-01
-1.07754901e-01 -4.15494680e-01 7.27444649e-01 6.04112804e-01
6.17342293e-01 6.58051550e-01 -2.81472951e-01 -1.66074559e-01
2.09469259e-01 6.64793074e-01 4.44813907e-01 -1.13585985e+00
-3.19023699e-01 -6.62023425e-01 -1.29007936e-01 -1.62740529e+00
3.09743494e-01 -1.07460964e+00 1.11659676e-01 -1.44909978e+00
3.29957485e-01 -4.43073392e-01 -4.85202193e-01 4.68616992e-01
-4.65182036e-01 -1.33465677e-01 4.63147499e-02 4.76802856e-01
-7.94188857e-01 7.77852952e-01 8.74857306e-01 -5.63304164e-02
-3.57365072e-01 -1.03706151e-01 -8.17661405e-01 6.39794230e-01
7.41008997e-01 -3.43729079e-01 -5.37352085e-01 -7.07422376e-01
1.94340572e-02 -1.99518085e-01 -3.95166636e-01 -9.01577711e-01
5.51130474e-01 1.03003062e-01 1.39737308e-01 -7.56927788e-01
-3.52492630e-02 -5.39081275e-01 -3.99893761e-01 4.56967235e-01
-6.08144939e-01 1.79110169e-01 2.22592115e-01 7.36074746e-01
-5.82914174e-01 -4.94158924e-01 5.93712151e-01 -6.11313283e-02
-8.33109796e-01 2.75601655e-01 -4.68709171e-01 -1.18338570e-01
1.09779298e+00 -4.90768015e-01 -8.10571685e-02 -3.50706349e-03
-4.79725003e-01 3.08236718e-01 1.52694702e-01 4.99955654e-01
5.57884455e-01 -1.00343430e+00 -4.25850391e-01 6.70857906e-01
1.89740211e-01 1.36020124e-01 4.64680821e-01 7.98218787e-01
-3.51774812e-01 9.49291706e-01 -1.57047193e-02 -3.72908980e-01
-1.24957991e+00 8.62796664e-01 -3.21664810e-01 -7.80182540e-01
-1.15604401e+00 7.98867047e-01 2.96882421e-01 4.53472659e-02
3.02840412e-01 -3.54642421e-01 -3.91804338e-01 1.98813602e-01
1.07820523e+00 3.41620475e-01 3.98536891e-01 -3.84138525e-01
-1.16105817e-01 3.84226441e-01 -7.15722442e-01 2.69367635e-01
1.34497154e+00 -3.61423612e-01 -2.80913979e-01 3.82964224e-01
1.09009314e+00 -3.61550972e-02 -6.67867303e-01 -2.88338989e-01
4.45847303e-01 -1.90816030e-01 3.29221457e-01 -1.91797137e-01
-8.27622414e-01 8.95943165e-01 -1.31346419e-01 5.96413314e-01
1.17107284e+00 -2.50061542e-01 1.25168812e+00 5.60931802e-01
1.46190211e-01 -1.31315053e+00 -8.94785896e-02 8.86266887e-01
6.18796647e-01 -8.11089039e-01 -8.95325001e-03 -4.00932312e-01
-5.36070764e-01 1.18687487e+00 5.29236794e-01 9.01213754e-03
5.12299657e-01 3.44190419e-01 -2.41271943e-01 -2.09281132e-01
-9.92079616e-01 -1.91858947e-01 5.43999784e-02 3.60920936e-01
7.67544031e-01 -1.08152203e-01 -4.39869165e-01 8.77541542e-01
-1.76397264e-01 1.01921640e-01 3.22877020e-01 1.15250325e+00
-7.03063548e-01 -1.01366639e+00 -1.76511437e-01 6.72643661e-01
-6.75134957e-01 -2.78666675e-01 -6.35805205e-02 4.71738994e-01
-2.61412501e-01 7.99082220e-01 4.23267037e-01 -1.98078677e-01
6.85009211e-02 3.23226154e-01 1.74749330e-01 -9.08651590e-01
-8.21966827e-01 1.37541015e-02 9.69606414e-02 -4.39376682e-01
-2.51160324e-01 -3.99783671e-01 -1.39464843e+00 -2.28453949e-01
-6.12515867e-01 5.86566627e-01 5.11362731e-01 9.59459484e-01
9.39213097e-01 6.07201397e-01 5.55820763e-01 -8.45474899e-01
-7.47033477e-01 -1.22925150e+00 -7.40520954e-01 2.53658444e-01
1.77380040e-01 -4.17150855e-01 4.58690822e-02 -8.41373652e-02] | [10.850839614868164, 7.695093154907227] |
1c483995-1f74-41da-88e4-62866e4e1c12 | physical-passive-patch-adversarial-attacks-on | 2207.05729 | null | https://arxiv.org/abs/2207.05729v2 | https://arxiv.org/pdf/2207.05729v2.pdf | Physical Passive Patch Adversarial Attacks on Visual Odometry Systems | Deep neural networks are known to be susceptible to adversarial perturbations -- small perturbations that alter the output of the network and exist under strict norm limitations. While such perturbations are usually discussed as tailored to a specific input, a universal perturbation can be constructed to alter the model's output on a set of inputs. Universal perturbations present a more realistic case of adversarial attacks, as awareness of the model's exact input is not required. In addition, the universal attack setting raises the subject of generalization to unseen data, where given a set of inputs, the universal perturbations aim to alter the model's output on out-of-sample data. In this work, we study physical passive patch adversarial attacks on visual odometry-based autonomous navigation systems. A visual odometry system aims to infer the relative camera motion between two corresponding viewpoints, and is frequently used by vision-based autonomous navigation systems to estimate their state. For such navigation systems, a patch adversarial perturbation poses a severe security issue, as it can be used to mislead a system onto some collision course. To the best of our knowledge, we show for the first time that the error margin of a visual odometry model can be significantly increased by deploying patch adversarial attacks in the scene. We provide evaluation on synthetic closed-loop drone navigation data and demonstrate that a comparable vulnerability exists in real data. A reference implementation of the proposed method and the reported experiments is provided at https://github.com/patchadversarialattacks/patchadversarialattacks. | ['Matan Jacoby', 'Chaim Baskin', 'Alex M. Bronstein', 'Yaniv Nemcovsky'] | 2022-07-11 | null | null | null | null | ['drone-navigation'] | ['computer-vision'] | [ 1.09935075e-01 5.11211336e-01 2.52570435e-02 4.46193665e-02
-4.48441327e-01 -1.03746080e+00 5.70842743e-01 -1.92294195e-01
-4.58496541e-01 7.10348845e-01 -4.42843556e-01 -4.23513025e-01
2.23781615e-01 -8.29676867e-01 -1.49259245e+00 -8.71948361e-01
-2.02360898e-01 2.31537342e-01 3.87109399e-01 -6.41617477e-01
-1.88732184e-02 7.66324043e-01 -1.30143940e+00 -4.60111022e-01
5.47259510e-01 8.56639504e-01 -4.47877228e-01 8.94613087e-01
8.35880339e-01 6.74571276e-01 -8.48710001e-01 -3.09439659e-01
8.39284599e-01 -1.26343995e-01 -3.76730144e-01 -2.34940544e-01
7.88634896e-01 -4.38637525e-01 -8.83313954e-01 1.69722784e+00
4.28745121e-01 1.73922330e-01 4.25516993e-01 -1.62303030e+00
-2.85841674e-01 1.03830524e-01 8.40278491e-02 1.29179090e-01
2.04589128e-01 6.89848959e-01 5.27085543e-01 -2.44297355e-01
4.34609771e-01 9.39584851e-01 5.97994268e-01 7.98516333e-01
-1.25722933e+00 -8.86686802e-01 1.75491124e-01 -1.11645684e-01
-1.43686604e+00 -4.12973881e-01 6.28153205e-01 -5.22937477e-01
5.58800697e-01 2.28249654e-01 4.81114626e-01 1.57153642e+00
6.49400413e-01 2.56438643e-01 7.62841403e-01 1.17856555e-01
4.44944084e-01 2.47231256e-02 -1.56441331e-01 6.92238748e-01
5.20508707e-01 9.04457271e-01 -1.14687070e-01 -4.50070918e-01
5.45502067e-01 -8.05613995e-02 -6.53223157e-01 -6.66011572e-01
-9.41469312e-01 7.99466908e-01 9.29785132e-01 -3.99095774e-01
-2.07036957e-01 4.27205473e-01 3.27531576e-01 6.15094602e-01
-1.48606196e-01 5.52247345e-01 -3.10247093e-01 2.07243666e-01
-6.16427511e-02 5.12638390e-01 9.28864121e-01 8.43774259e-01
7.59938180e-01 5.98975718e-01 4.56620365e-01 1.00266397e-01
1.84647068e-01 1.09656417e+00 3.80714506e-01 -8.66442204e-01
4.58573818e-01 1.86047740e-02 2.88209558e-01 -1.33134174e+00
-3.90768051e-01 -3.61087412e-01 -8.00366163e-01 1.00620937e+00
5.53602278e-01 -5.81493497e-01 -9.34801579e-01 2.03133893e+00
4.70381886e-01 4.08512503e-01 4.63995695e-01 1.12239838e+00
3.54301333e-01 4.78819847e-01 -4.85747904e-01 2.46715456e-01
8.81211042e-01 -3.47243547e-01 -4.55991089e-01 -6.91657543e-01
3.24650317e-01 -4.52583671e-01 6.88290775e-01 4.08723354e-01
-5.67511022e-01 -4.21737999e-01 -1.69023442e+00 4.59262788e-01
-5.24732351e-01 -4.65711921e-01 3.13753299e-02 8.42170715e-01
-6.40677094e-01 3.62787366e-01 -9.72180486e-01 -1.95862979e-01
-1.73026752e-02 6.23916805e-01 -6.54710174e-01 1.34528533e-01
-1.64761996e+00 1.16853023e+00 3.71271402e-01 1.88624650e-01
-1.61347961e+00 -3.84664625e-01 -1.12458146e+00 -3.83013695e-01
6.42179608e-01 -7.04752147e-01 1.25986171e+00 -9.76382971e-01
-1.54506409e+00 5.45238137e-01 2.95940697e-01 -9.66529131e-01
6.47515237e-01 -2.30483860e-01 -4.52105701e-01 -1.21222483e-02
-1.56807616e-01 4.42261815e-01 1.37439466e+00 -1.43888032e+00
-3.16563338e-01 -3.78870547e-01 7.84255266e-01 1.52409762e-01
5.77912256e-02 -6.96337879e-01 -3.64659019e-02 -5.93342006e-01
6.40495047e-02 -1.63897014e+00 -4.94791657e-01 1.93122268e-01
-7.31917024e-01 8.91392052e-01 9.26965356e-01 -1.04510836e-01
5.46489000e-01 -2.31452751e+00 1.40863210e-01 3.07501435e-01
1.59981236e-01 4.12377834e-01 1.00032166e-01 4.36388522e-01
-7.26049617e-02 -5.11108525e-02 -2.83730477e-01 -1.23652175e-01
1.09736480e-01 6.12527788e-01 -9.14966941e-01 1.11748672e+00
-4.09216397e-02 5.94415486e-01 -7.00256467e-01 4.22675520e-01
5.10015264e-02 4.39752281e-01 -6.58926904e-01 2.91036785e-01
-9.43838432e-02 4.89378095e-01 -2.33902320e-01 1.80224925e-01
5.84683418e-01 3.87854457e-01 -1.60137236e-01 -1.15146656e-02
3.08840781e-01 1.01342596e-01 -1.22675037e+00 1.13758302e+00
-2.74680436e-01 8.14093471e-01 1.32690728e-01 -8.34401608e-01
7.61259139e-01 2.64402241e-01 -9.64331850e-02 -2.06608266e-01
5.01560450e-01 9.50873122e-02 2.98298508e-01 -2.19531983e-01
6.23868763e-01 -1.64602712e-01 -5.59548676e-01 1.84970796e-01
-1.35871708e-01 -4.34446275e-01 -5.93403637e-01 1.06301591e-01
1.07665539e+00 -2.41133749e-01 4.48721915e-01 -1.71630159e-01
5.63612759e-01 1.03796117e-01 7.50875235e-01 8.38933825e-01
-3.86384875e-01 4.36172187e-01 3.47257465e-01 -4.36106384e-01
-8.05302739e-01 -1.28907299e+00 7.31387660e-02 2.53437936e-01
6.57615662e-01 9.20133367e-02 -6.64077461e-01 -7.86689818e-01
2.08594888e-01 4.57622886e-01 -7.94941425e-01 -8.19006145e-01
-4.71598774e-01 -3.16564709e-01 1.23810136e+00 2.16975078e-01
4.12648529e-01 -4.58113134e-01 -7.23752558e-01 -5.36141992e-02
2.07721412e-01 -1.25782216e+00 -3.36817324e-01 2.92721927e-01
-5.32278657e-01 -1.45817471e+00 -1.51252523e-01 -6.33953631e-01
9.08966184e-01 1.91126674e-01 5.29681504e-01 4.80154678e-02
4.46962900e-02 3.88786137e-01 -1.13742828e-01 -5.53620756e-01
-8.68705511e-01 -2.61306703e-01 8.68665755e-01 4.80504148e-02
-1.71755701e-01 -5.02936900e-01 -3.61443013e-01 5.51075518e-01
-1.21823668e+00 -6.13174438e-01 -3.53343263e-02 9.01627481e-01
2.68761992e-01 1.71705753e-01 9.97006297e-02 -7.86742389e-01
3.55690002e-01 -5.11539757e-01 -1.04264379e+00 -4.98376310e-01
-2.65936881e-01 1.03472844e-01 1.22018099e+00 -7.75882244e-01
-4.29929942e-01 8.49343240e-02 -1.50799409e-01 -6.66512191e-01
-3.19554836e-01 3.39905500e-01 -3.14209849e-01 -7.37660289e-01
1.11986625e+00 1.39045775e-01 2.36724004e-01 8.93953294e-02
2.29846179e-01 2.29504168e-01 8.53026390e-01 -3.52909118e-01
1.63252962e+00 8.04832578e-01 3.35592955e-01 -8.23780715e-01
-3.17897946e-01 8.96487907e-02 -1.59951359e-01 -2.16961905e-01
5.12682557e-01 -1.00443828e+00 -8.74654591e-01 7.02624977e-01
-1.00826800e+00 -4.98803467e-01 -5.75142018e-02 3.32994819e-01
-5.76014161e-01 4.49633092e-01 -3.16903859e-01 -4.61889327e-01
2.18593463e-01 -1.47832870e+00 5.47767758e-01 9.74911451e-02
1.29252344e-01 -9.66910660e-01 2.57738769e-01 -3.95936146e-02
6.91061318e-02 6.17297709e-01 3.26873392e-01 -7.33359635e-01
-7.03699827e-01 -6.50447726e-01 6.01444960e-01 4.93143976e-01
-1.87637173e-02 -1.33823261e-01 -1.08388305e+00 -8.43275845e-01
2.60188758e-01 -2.14043170e-01 4.86116916e-01 6.25338480e-02
3.52790028e-01 -7.70099342e-01 -1.24144211e-01 8.94770205e-01
1.35812998e+00 2.69759417e-01 5.72049439e-01 6.00980461e-01
6.92085505e-01 2.53731400e-01 5.86077213e-01 2.26930156e-01
1.03336066e-01 6.85761988e-01 1.38012135e+00 1.77409381e-01
5.10840952e-01 -3.31644595e-01 7.88757682e-01 -3.15169059e-03
2.95727342e-01 -4.21202898e-01 -7.16157138e-01 2.76799142e-01
-1.65036285e+00 -7.63537288e-01 1.44230381e-01 2.56553411e+00
1.56040534e-01 6.02652669e-01 -1.96590394e-01 5.83253503e-02
6.05377316e-01 4.49776977e-01 -7.45614231e-01 -3.79201502e-01
4.15588990e-02 -2.93473363e-01 1.13935435e+00 9.49020624e-01
-1.30201626e+00 8.69157493e-01 5.21105671e+00 2.55526245e-01
-1.44726741e+00 -7.82590285e-02 -7.43838549e-02 -1.02711700e-01
1.30923733e-01 1.23240128e-01 -7.13094890e-01 3.89531195e-01
8.73011470e-01 -2.11479023e-01 4.49884683e-01 8.90938044e-01
-8.98018628e-02 5.73629849e-02 -1.10656488e+00 6.13335073e-01
1.37786239e-01 -1.06760657e+00 -7.46691823e-02 4.92625624e-01
5.81595421e-01 3.60575497e-01 4.43466246e-01 2.63696194e-01
6.85411036e-01 -7.09696174e-01 7.84067512e-01 2.42911518e-01
4.00462061e-01 -9.89006162e-01 7.57679641e-01 5.93260586e-01
-8.85281801e-01 -4.56315652e-02 -5.09784520e-01 -3.05483848e-01
1.89469606e-01 2.57803500e-02 -8.05754483e-01 4.15528148e-01
5.27998686e-01 2.90837497e-01 -2.64649749e-01 7.22476780e-01
-6.29779994e-01 4.20592576e-01 -4.92859364e-01 2.60441303e-01
4.30829942e-01 6.88336715e-02 1.19357502e+00 5.54983139e-01
1.92060247e-01 -3.07063162e-02 3.20892572e-01 5.33890903e-01
1.08557895e-01 -5.47516346e-01 -1.42798591e+00 3.96271229e-01
4.53252763e-01 6.15980089e-01 -5.00295982e-02 -4.20707241e-02
-4.18956652e-02 1.07111669e+00 1.87101930e-01 5.61146617e-01
-1.08994794e+00 -4.32952106e-01 1.66104150e+00 -1.90441534e-01
2.78936058e-01 -4.79549170e-01 1.34885490e-01 -1.24768412e+00
-7.46622682e-02 -1.18039763e+00 1.48366839e-01 -4.57415313e-01
-9.76565480e-01 7.62059450e-01 -3.27141881e-01 -1.87229097e+00
-4.93216038e-01 -6.16815388e-01 -4.49720472e-01 5.95056057e-01
-1.07758737e+00 -8.06582749e-01 -1.69111758e-01 7.49293983e-01
-1.38367526e-02 -3.55789632e-01 8.10621202e-01 -1.19660497e-01
-5.22867501e-01 7.76713967e-01 5.55676296e-02 4.17395502e-01
6.45384371e-01 -1.02994049e+00 7.63668418e-01 1.58071005e+00
2.35082172e-02 7.23167062e-01 1.37356043e+00 -5.22174239e-01
-1.59107471e+00 -1.26830292e+00 1.23579174e-01 -4.81589139e-01
9.37773108e-01 -3.61794502e-01 -8.11966002e-01 1.06743240e+00
-5.68889827e-02 3.77448201e-01 2.59765625e-01 -5.32381356e-01
-6.22118711e-01 -8.61035958e-02 -1.25965512e+00 1.01738822e+00
6.33937538e-01 -7.83296585e-01 -5.60607433e-01 8.34050551e-02
8.42544794e-01 -1.05294120e+00 -7.44966745e-01 4.19658482e-01
5.19939303e-01 -1.01827276e+00 1.23651493e+00 -6.53701901e-01
-1.08418413e-01 -8.36492598e-01 -4.30374682e-01 -1.71831620e+00
6.60288557e-02 -8.48757267e-01 -3.37191746e-02 5.27153790e-01
2.28052542e-01 -1.28144252e+00 6.51399314e-01 3.34023893e-01
-1.25381932e-01 -1.20691001e-01 -1.25302720e+00 -1.18075192e+00
2.44037569e-01 -6.07410431e-01 4.80509073e-01 8.91376317e-01
-2.07016081e-01 -6.69293925e-02 -7.27858484e-01 1.58846807e+00
6.47110462e-01 -3.79158229e-01 1.39483023e+00 -7.13110685e-01
-5.03594458e-01 -7.89059624e-02 -1.17322302e+00 -1.14613092e+00
3.60279888e-01 -5.55972338e-01 4.88373190e-01 -5.77705622e-01
-8.43552053e-01 -1.94352984e-01 4.13353071e-02 1.34649366e-01
-7.40514919e-02 4.52028096e-01 3.64225715e-01 1.00358641e-02
5.41982614e-03 5.07637978e-01 8.12501907e-01 -5.13585925e-01
3.69720235e-02 4.44373488e-01 -3.45523864e-01 7.21886337e-01
9.11136091e-01 -6.97906077e-01 -5.87785304e-01 -3.96121681e-01
2.23865256e-01 9.86146405e-02 7.12290645e-01 -1.41879857e+00
4.21107352e-01 -5.93141615e-02 -1.74656719e-01 8.65448043e-02
6.40894711e-01 -1.20066845e+00 3.33128959e-01 9.50650871e-01
6.08251728e-02 3.24532315e-02 4.87634808e-01 8.40888262e-01
-3.01850557e-01 -1.09654456e-01 1.03413558e+00 1.75798953e-01
-6.34108126e-01 5.18978536e-01 -7.47002125e-01 3.61117609e-02
1.15560699e+00 -1.53071702e-01 -6.79621935e-01 -6.97499931e-01
-6.26945615e-01 6.90033138e-02 9.12264347e-01 4.69291568e-01
6.29923701e-01 -1.17292976e+00 -2.06855550e-01 5.66033781e-01
3.37776870e-01 1.19997166e-01 2.91030824e-01 5.44866085e-01
-6.66989923e-01 2.03500181e-01 -3.09765786e-01 -5.94393969e-01
-1.03912735e+00 9.24245656e-01 8.96452010e-01 2.78236598e-01
-5.64642608e-01 6.22051418e-01 5.24568796e-01 -6.97724760e-01
2.21544713e-01 -4.23896372e-01 2.11108327e-01 -3.67465347e-01
4.25701350e-01 -4.18254500e-03 -1.09345736e-02 -9.73299265e-01
-3.65801007e-01 4.83323067e-01 1.10978700e-01 -7.77179599e-02
6.51334524e-01 -2.04630226e-01 3.04939002e-01 2.48070851e-01
1.21238422e+00 2.07032844e-01 -1.62262559e+00 -1.50773972e-01
-4.74863529e-01 -3.94169301e-01 -6.49700090e-02 -1.98925555e-01
-1.00166392e+00 5.39010227e-01 5.97479284e-01 2.79957831e-01
7.96818435e-01 -4.86684650e-01 6.88305855e-01 8.95672619e-01
6.96603417e-01 -5.56843162e-01 -1.25956014e-01 8.39364529e-01
8.76262665e-01 -1.29095984e+00 -3.36883485e-01 -2.00650692e-01
-5.32543480e-01 8.92336845e-01 7.61778295e-01 -8.22352052e-01
7.11155534e-01 4.25214708e-01 6.56510234e-01 6.92435279e-02
-5.13029575e-01 2.87554264e-02 8.13305154e-02 7.97012269e-01
-6.94768310e-01 5.82970642e-02 3.66234034e-01 7.56964460e-02
-3.56274247e-01 -5.87392569e-01 1.06583166e+00 1.01007831e+00
-3.25827003e-01 -7.49666691e-01 -8.52174580e-01 -2.03135356e-01
-2.57179767e-01 1.62320375e-01 -2.43044734e-01 1.01964259e+00
-4.26518545e-02 8.50638330e-01 -1.17782041e-01 -8.14172208e-01
5.51880777e-01 -2.90547997e-01 1.40180647e-01 -3.83608043e-01
-4.08476055e-01 -5.71654379e-01 4.89039756e-02 -8.60677361e-01
-6.71956167e-02 -4.70521361e-01 -1.09745181e+00 -4.46116030e-01
-5.73266996e-03 -3.32586803e-02 5.42177618e-01 8.34796309e-01
1.84640884e-01 4.50567305e-01 7.50822186e-01 -1.12649775e+00
-8.27101290e-01 -5.69584012e-01 -4.37584430e-01 1.56531394e-01
1.17066145e+00 -8.99727643e-01 -8.43762696e-01 -1.47375420e-01] | [5.436560153961182, 7.822946071624756] |
76f86bce-1521-426f-92aa-1a199d5bd4ec | ecg-heartbeat-classification-using-multimodal | 2107.09869 | null | https://arxiv.org/abs/2107.09869v1 | https://arxiv.org/pdf/2107.09869v1.pdf | ECG Heartbeat Classification Using Multimodal Fusion | Electrocardiogram (ECG) is an authoritative source to diagnose and counter critical cardiovascular syndromes such as arrhythmia and myocardial infarction (MI). Current machine learning techniques either depend on manually extracted features or large and complex deep learning networks which merely utilize the 1D ECG signal directly. Since intelligent multimodal fusion can perform at the stateof-the-art level with an efficient deep network, therefore, in this paper, we propose two computationally efficient multimodal fusion frameworks for ECG heart beat classification called Multimodal Image Fusion (MIF) and Multimodal Feature Fusion (MFF). At the input of these frameworks, we convert the raw ECG data into three different images using Gramian Angular Field (GAF), Recurrence Plot (RP) and Markov Transition Field (MTF). In MIF, we first perform image fusion by combining three imaging modalities to create a single image modality which serves as input to the Convolutional Neural Network (CNN). In MFF, we extracted features from penultimate layer of CNNs and fused them to get unique and interdependent information necessary for better performance of classifier. These informational features are finally used to train a Support Vector Machine (SVM) classifier for ECG heart-beat classification. We demonstrate the superiority of the proposed fusion models by performing experiments on PhysioNets MIT-BIH dataset for five distinct conditions of arrhythmias which are consistent with the AAMI EC57 protocols and on PTB diagnostics dataset for Myocardial Infarction (MI) classification. We achieved classification accuracy of 99.7% and 99.2% on arrhythmia and MI classification, respectively. | ['Naimul Khan', 'Ling Guan', 'Anika Tabassum', 'Zeeshan Ahmad'] | 2021-07-21 | null | null | null | null | ['heartbeat-classification'] | ['medical'] | [ 4.98179585e-01 -3.67355943e-01 2.37268880e-01 -2.75205046e-01
-9.03356731e-01 -4.55018252e-01 2.13220686e-01 3.56656492e-01
-3.38728309e-01 8.98741722e-01 -8.77171084e-02 -5.93575597e-01
-3.04063320e-01 -6.19305730e-01 -3.50310981e-01 -8.00896883e-01
-1.62355125e-01 2.31664598e-01 -5.00022769e-01 1.06773913e-01
3.38702537e-02 7.46081531e-01 -1.16616011e+00 5.71441352e-01
7.39260018e-01 1.27318347e+00 -2.77190834e-01 1.22879088e+00
3.06313664e-01 4.80402172e-01 -6.38400495e-01 -1.92992523e-01
4.76988181e-02 -7.69400716e-01 -7.09584355e-01 -1.89250529e-01
2.07374413e-02 -1.20493263e-01 -2.49091864e-01 7.07500458e-01
1.05627966e+00 -2.82979161e-01 8.86853814e-01 -1.00599039e+00
-5.12676910e-02 5.69266856e-01 -5.22559166e-01 5.28479159e-01
2.06036314e-01 1.30077470e-02 4.51708376e-01 -9.11872625e-01
3.63778889e-01 6.24444127e-01 9.33250070e-01 2.26561964e-01
-9.11954284e-01 -4.33369935e-01 -7.19759345e-01 1.94783062e-01
-1.36119390e+00 -1.44665495e-01 9.38602448e-01 -5.09985387e-01
6.71949863e-01 5.61730742e-01 8.10636461e-01 7.79805899e-01
6.59453154e-01 3.16909254e-01 1.06347013e+00 -4.09131765e-01
-2.11729065e-01 -2.58826077e-01 2.46953055e-01 7.54983425e-01
2.71262258e-01 4.16424200e-02 -3.94157946e-01 -4.97909456e-01
7.55201459e-01 3.21220875e-01 -4.72084284e-01 1.64708465e-01
-1.67270851e+00 5.76443791e-01 2.36983553e-01 5.43207526e-01
-7.67957628e-01 2.33791415e-02 5.11438727e-01 3.93273860e-01
2.95046754e-02 7.54026175e-02 -3.46713424e-01 -6.47817105e-02
-8.97255301e-01 1.27578929e-01 5.76150954e-01 6.33384287e-02
6.29351974e-01 -4.56448719e-02 -3.28579694e-01 7.07508266e-01
3.75601470e-01 6.69668853e-01 5.97347617e-01 -8.32011938e-01
3.68628383e-01 6.61517322e-01 -2.08661810e-01 -1.09040809e+00
-7.97383845e-01 -7.78788388e-01 -1.57018375e+00 9.25987363e-02
1.48737863e-01 -4.78418618e-01 -7.85901904e-01 1.46727312e+00
4.20498028e-02 2.92869687e-01 3.86433303e-01 8.63364816e-01
1.19691026e+00 2.59512335e-01 -1.93545409e-02 -2.62191147e-01
1.52543080e+00 -2.13814363e-01 -7.01568723e-01 4.08345163e-01
8.89816463e-01 -8.02352369e-01 3.33130479e-01 5.15081823e-01
-1.04572022e+00 -7.62719214e-01 -1.25708795e+00 2.53244013e-01
-1.47179767e-01 4.83987212e-01 3.51516336e-01 6.39108479e-01
-9.24226046e-01 7.33078539e-01 -6.56268716e-01 -5.56617603e-02
6.33699238e-01 5.85599303e-01 -6.44390583e-01 1.71178490e-01
-1.35915685e+00 8.48542929e-01 3.41567129e-01 4.84538227e-01
-7.76619971e-01 -4.96905833e-01 -8.23173881e-01 -5.07205836e-02
-3.95656437e-01 -1.18012393e+00 7.14088500e-01 -8.10435593e-01
-1.20380139e+00 8.18378806e-01 1.57235965e-01 -5.23991108e-01
3.47020149e-01 -7.74620101e-02 -4.49484944e-01 5.03986597e-01
-1.76686719e-01 5.38153410e-01 8.77800047e-01 -1.10061860e+00
-5.77330291e-01 -4.46600020e-01 -2.48357788e-01 1.65023506e-01
-1.28338590e-01 -1.32392868e-01 2.11163878e-01 -6.55361116e-01
2.54300863e-01 -7.56708443e-01 -4.60371599e-02 -5.17201781e-01
-4.41058844e-01 1.23714097e-01 8.88065398e-01 -9.03747857e-01
1.31689167e+00 -2.24625516e+00 3.95752311e-01 3.24137598e-01
7.25661278e-01 3.63781869e-01 9.14452076e-02 2.14122891e-01
-4.00171816e-01 1.14619441e-01 -5.37008703e-01 -2.56331265e-01
-4.28693980e-01 1.21348888e-01 1.22833006e-01 5.90040445e-01
2.08789364e-01 1.21007466e+00 -4.86754894e-01 -6.89920485e-01
4.65292871e-01 6.99760258e-01 -1.01306632e-01 7.91030303e-02
6.55626237e-01 7.46204078e-01 -2.86405027e-01 6.33918881e-01
6.54914558e-01 -2.50905991e-01 1.88903183e-01 -8.04540396e-01
1.51104078e-01 -3.32642198e-01 -9.11268175e-01 1.75784671e+00
-2.40728870e-01 2.25934461e-01 -1.98096633e-01 -1.29126585e+00
9.12143767e-01 7.94547975e-01 8.52322340e-01 -2.36057058e-01
5.47272623e-01 2.71058828e-01 3.13048989e-01 -5.87186754e-01
-1.82668746e-01 -3.40693176e-01 -6.72269166e-02 4.80190068e-01
3.59893203e-01 2.57933766e-01 -8.97311792e-02 -4.39533517e-02
1.07807779e+00 -1.35597959e-01 3.73925269e-01 -2.54371583e-01
7.47610986e-01 -4.17526543e-01 4.20903772e-01 8.45080137e-01
-3.40990514e-01 9.91430998e-01 3.08416814e-01 -8.50103676e-01
-6.48506403e-01 -1.11824596e+00 -2.72911787e-01 1.57224432e-01
-1.04986474e-01 -2.95546114e-01 -5.87365329e-01 -7.18885303e-01
-2.95471370e-01 -1.54407278e-01 -4.42396700e-01 -2.57416010e-01
-6.32979631e-01 -1.15899491e+00 1.04641855e+00 6.04339898e-01
7.38303900e-01 -1.04314458e+00 -1.07848346e+00 2.75105715e-01
-5.06624281e-01 -8.81300807e-01 5.19375950e-02 3.20062757e-01
-9.02572155e-01 -1.18594599e+00 -9.14667428e-01 -5.20507634e-01
2.62391806e-01 -2.19523564e-01 9.60695684e-01 2.05970138e-01
-6.91558182e-01 1.67662412e-01 -3.45392644e-01 -4.49500829e-01
-3.87834340e-01 6.37759715e-02 -8.26473236e-02 4.53997970e-01
-1.17975578e-01 -7.52525032e-01 -8.68449390e-01 -2.49758244e-01
-7.97996461e-01 1.61615252e-01 7.56950259e-01 9.29330409e-01
2.99365014e-01 -2.43501827e-01 9.79578078e-01 -5.22895515e-01
6.71757281e-01 -3.07395875e-01 -7.97331035e-02 2.09126562e-01
-2.70369768e-01 -3.88459772e-01 3.23061585e-01 -5.29601574e-02
-7.43115008e-01 2.09046528e-01 -1.91341594e-01 -5.66783607e-01
-3.87165010e-01 9.92812455e-01 7.71820545e-02 -8.11876580e-02
6.79248691e-01 1.63503349e-01 6.29887655e-02 -1.30215734e-01
-7.45496452e-02 8.31688941e-01 7.96192169e-01 -3.14190477e-01
3.63291323e-01 3.11315775e-01 5.25162280e-01 -5.97119093e-01
-5.01096964e-01 -3.16166639e-01 -9.01839018e-01 -4.27126169e-01
1.26720166e+00 -8.13427210e-01 -8.62219274e-01 7.46976912e-01
-1.18485558e+00 3.25295776e-01 -7.69100059e-03 7.63053000e-01
-3.65113109e-01 6.09668612e-01 -7.38837481e-01 -7.12982655e-01
-8.37540925e-01 -1.03678846e+00 1.03866541e+00 1.83013573e-01
-2.58593500e-01 -1.16164482e+00 1.25556499e-01 3.82090420e-01
3.05467874e-01 9.85651314e-01 1.17048299e+00 -5.78247130e-01
-8.46548378e-02 -4.24973786e-01 -2.20466495e-01 6.19577348e-01
3.23334962e-01 -4.48217131e-02 -1.07683921e+00 -1.04241438e-01
2.11846218e-01 -3.26940268e-02 9.14378941e-01 5.90404809e-01
8.85064304e-01 3.20736051e-01 -2.65166640e-01 7.34873414e-01
1.26418042e+00 4.05153900e-01 7.16387153e-01 -1.16113044e-01
8.97819400e-01 -3.95890558e-03 1.18462972e-01 6.04880750e-01
3.12404126e-01 2.62504905e-01 2.93387502e-01 -6.15755141e-01
8.69379193e-03 5.06830156e-01 1.01438731e-01 8.99919331e-01
-4.85293150e-01 -6.20772243e-02 -1.17925286e+00 3.16215187e-01
-1.92960835e+00 -8.21220815e-01 -3.11375767e-01 2.17169929e+00
6.10938191e-01 -9.09770206e-02 -5.80460541e-02 7.22814560e-01
6.59573138e-01 -2.85222381e-01 -8.22241083e-02 -2.58806378e-01
-3.57649118e-01 6.34031773e-01 1.24873944e-01 1.76398769e-01
-1.44763947e+00 7.02870414e-02 5.52604628e+00 3.40088278e-01
-1.44958735e+00 1.70184314e-01 7.71882653e-01 3.61604720e-01
1.51507124e-01 -2.50203252e-01 -1.03276499e-01 1.97553247e-01
1.00829995e+00 1.42659768e-01 7.49126226e-02 -2.64707375e-02
-2.14847922e-02 4.15097848e-02 -8.33099782e-01 1.60011375e+00
-2.43433546e-02 -1.53014421e+00 5.77912433e-03 -1.05540693e-01
4.24657434e-01 5.45915551e-02 -3.76139209e-02 5.65948077e-02
-6.11467004e-01 -1.17191625e+00 2.15487778e-01 1.04477108e+00
9.58109558e-01 -7.88684547e-01 1.38220918e+00 1.66036710e-01
-1.10546398e+00 -4.22796831e-02 2.01361790e-01 8.15093070e-02
1.75068781e-01 7.15884387e-01 -5.39068699e-01 1.28650832e+00
6.21200144e-01 9.55594182e-01 -5.43737292e-01 9.38117325e-01
4.27779287e-01 5.67902386e-01 -2.66671777e-01 5.72636724e-01
-8.23525563e-02 -4.09443378e-02 4.95641619e-01 1.04853320e+00
5.42170644e-01 9.62388217e-02 1.46958709e-01 5.87427199e-01
-3.82013917e-02 1.00852624e-01 -5.86488068e-01 -4.74154875e-02
-6.53509200e-02 1.54845548e+00 -8.33057761e-01 -5.50029099e-01
-2.71369338e-01 8.39890778e-01 -2.99022287e-01 2.80369192e-01
-1.00078273e+00 -6.34715736e-01 2.22645346e-02 -2.26367593e-01
-3.89467105e-02 1.41145959e-01 -7.06999004e-01 -1.27162480e+00
-9.34408307e-02 -7.88418531e-01 4.79954898e-01 -5.24108469e-01
-1.12433469e+00 8.80783856e-01 -4.40149270e-02 -1.28589678e+00
-1.80910736e-01 -4.26708251e-01 -7.32311368e-01 1.24760473e+00
-1.23236930e+00 -1.35360956e+00 -5.37407637e-01 7.16910839e-01
-1.29126847e-01 -3.56568366e-01 1.21710944e+00 7.14695156e-01
-4.26371664e-01 3.47369224e-01 -3.28097969e-01 2.75970519e-01
5.58100224e-01 -1.08689702e+00 -2.28835568e-01 5.72940886e-01
-4.13020663e-02 3.61019373e-01 1.90576300e-01 -4.91255343e-01
-1.22497892e+00 -1.09817660e+00 8.99562061e-01 -2.57206827e-01
4.79506664e-02 2.75251955e-01 -6.16953552e-01 2.24663213e-01
3.41568381e-01 2.70259142e-01 9.73248899e-01 -2.98167557e-01
1.31216198e-01 -1.23508960e-01 -1.00294256e+00 1.22728527e-01
4.37257946e-01 -6.10552192e-01 -4.79024708e-01 1.98670074e-01
4.37700264e-02 -3.20725501e-01 -1.35894132e+00 1.10363746e+00
7.42099345e-01 -1.06262970e+00 8.87665749e-01 -4.83819604e-01
3.53305757e-01 -4.40535575e-01 -1.88602954e-01 -1.03633547e+00
-1.22450933e-01 -5.36543548e-01 -1.35611696e-02 8.13642263e-01
3.18555832e-01 -6.88346326e-01 2.35621601e-01 -7.69333169e-02
-3.33959311e-01 -1.02388871e+00 -1.06346965e+00 -1.14933863e-01
-1.42389402e-01 -3.54755193e-01 4.24642175e-01 1.06095469e+00
-4.79028374e-02 4.29548323e-01 -3.03461850e-01 4.26376686e-02
5.55240631e-01 1.91078335e-01 3.72536629e-01 -1.46198583e+00
-1.37415722e-01 -3.59411836e-01 -6.82254016e-01 8.51793736e-02
-9.99321789e-02 -1.17175472e+00 -4.46216524e-01 -1.41672611e+00
3.34904760e-01 -2.99552828e-01 -1.01159644e+00 5.96685827e-01
-2.31815338e-01 7.80945003e-01 2.38504052e-01 9.34776664e-02
-1.96854070e-01 9.83670801e-02 1.04333246e+00 1.74964429e-03
-1.85701534e-01 4.95354831e-03 -4.74367797e-01 6.77119732e-01
8.85888875e-01 -3.47097754e-01 -1.74704492e-01 -1.38622997e-02
9.67280790e-02 7.26269305e-01 7.93929935e-01 -1.44362605e+00
-1.58818409e-01 4.67925727e-01 7.65215814e-01 -7.61330307e-01
2.32198820e-01 -6.24944329e-01 5.29535472e-01 6.71403944e-01
-8.02998543e-02 2.49451846e-01 1.58110678e-01 8.69113505e-02
-5.84110439e-01 6.87067434e-02 7.15273798e-01 5.71335591e-02
2.47283150e-02 2.94956952e-01 -4.90071625e-01 -3.34180921e-01
9.60084558e-01 -9.80580598e-02 -1.67923436e-01 -1.74764171e-01
-1.18635058e+00 -2.32579753e-01 -1.62788346e-01 5.40312454e-02
1.00039887e+00 -1.49450564e+00 -9.94723976e-01 3.47466886e-01
-6.47524148e-02 -2.88050562e-01 6.84133351e-01 1.68191779e+00
-8.77124786e-01 2.90555239e-01 -4.79925543e-01 -9.52566326e-01
-1.34315634e+00 1.65188059e-01 6.35058999e-01 -2.36014500e-01
-5.54012835e-01 4.20837879e-01 -3.20692137e-02 -8.35115463e-02
-1.24700852e-01 -4.59194303e-01 -4.20299560e-01 1.87459722e-01
2.75382578e-01 2.55941391e-01 3.54742169e-01 -7.25896537e-01
-5.25833011e-01 7.41966188e-01 2.41328925e-01 -4.09708768e-02
1.05256891e+00 7.94286206e-02 -3.04541975e-01 3.49973291e-01
1.27911127e+00 -4.50976521e-01 -5.37285745e-01 2.39132285e-01
-3.21076900e-01 1.55944332e-01 1.96025744e-01 -9.48129356e-01
-1.24403179e+00 1.27838194e+00 1.08880281e+00 1.57025278e-01
1.48500288e+00 -3.08655292e-01 8.75658572e-01 2.37607524e-01
-1.11091733e-01 -4.61192131e-01 -3.45017195e-01 1.26747176e-01
7.13323116e-01 -9.78661120e-01 -1.20081410e-01 -7.93101043e-02
-6.84186220e-01 1.48793375e+00 -5.65045476e-02 -5.25820181e-02
1.08353293e+00 2.90124625e-01 4.51453328e-01 -3.01469952e-01
-5.18322587e-01 -5.35523742e-02 4.95819777e-01 4.53905821e-01
5.91489613e-01 2.02506661e-01 -4.44757789e-01 8.95680785e-01
2.16605410e-01 3.57888341e-01 2.93571740e-01 1.09469116e+00
-1.65577054e-01 -9.95392799e-01 -3.27063829e-01 6.32793188e-01
-8.53127420e-01 -1.07691452e-01 -1.59263834e-01 4.11978900e-01
5.89415252e-01 8.04002821e-01 -3.31207991e-01 -6.46707118e-01
-4.70288098e-02 4.24793065e-01 6.83806419e-01 -2.53414065e-01
-1.05645037e+00 3.30134362e-01 -1.18417896e-01 -1.62777647e-01
-6.90354407e-01 -5.98096728e-01 -1.28046954e+00 2.45065525e-01
-1.43086940e-01 4.95379306e-02 6.22321963e-01 1.19215441e+00
4.94975686e-01 8.97580504e-01 5.28486967e-01 -8.83179605e-01
-2.28676349e-01 -1.07522094e+00 -4.68971312e-01 4.49995786e-01
6.07710302e-01 -5.05864322e-01 -8.84003490e-02 4.08819735e-01] | [14.275814056396484, 3.2620701789855957] |
49e3bb73-b2ac-4631-8dea-0e77186117c8 | pmaa-a-progressive-multi-scale-attention | 2303.16565 | null | https://arxiv.org/abs/2303.16565v1 | https://arxiv.org/pdf/2303.16565v1.pdf | PMAA: A Progressive Multi-scale Attention Autoencoder Model for High-Performance Cloud Removal from Multi-temporal Satellite Imagery | Satellite imagery analysis plays a vital role in remote sensing, but the information loss caused by cloud cover seriously hinders its application. This study presents a high-performance cloud removal architecture called Progressive Multi-scale Attention Autoencoder (PMAA), which simultaneously leverages global and local information. It mainly consists of a cloud detection backbone and a cloud removal module. The cloud detection backbone uses cloud masks to reinforce cloudy areas to prompt the cloud removal module. The cloud removal module mainly comprises a novel Multi-scale Attention Module (MAM) and a Local Interaction Module (LIM). PMAA establishes the long-range dependency of multi-scale features using MAM and modulates the reconstruction of the fine-grained details using LIM, allowing for the simultaneous representation of fine- and coarse-grained features at the same level. With the help of diverse and multi-scale feature representation, PMAA outperforms the previous state-of-the-art model CTGAN consistently on the Sen2_MTC_Old and Sen2_MTC_New datasets. Furthermore, PMAA has a considerable efficiency advantage, with only 0.5% and 14.6% of the parameters and computational complexity of CTGAN, respectively. These extensive results highlight the potential of PMAA as a lightweight cloud removal network suitable for deployment on edge devices. We will release the code and trained models to facilitate the study in this direction. | ['Yachao Cui', 'Pin Tao', 'Junliang Xing', 'Kai Li', 'Xuechao Zou'] | 2023-03-29 | null | null | null | null | ['cloud-removal', 'cloud-detection'] | ['computer-vision', 'computer-vision'] | [ 5.92694990e-03 -4.28164750e-01 2.47892320e-01 -1.98941585e-02
-5.68294048e-01 -3.41021478e-01 5.47710240e-01 -3.26817751e-01
-2.16329843e-01 5.46244144e-01 2.28102580e-02 -3.33688766e-01
-1.00328691e-01 -1.08362198e+00 -5.79458654e-01 -1.26899660e+00
-1.44119352e-01 -3.16897221e-02 1.01052625e-02 -2.42007405e-01
-3.14820409e-01 8.02401125e-01 -1.61676180e+00 1.55463144e-01
1.17189133e+00 1.16895759e+00 7.47670114e-01 5.32423735e-01
5.17937280e-02 6.58437848e-01 -4.12994593e-01 7.10255951e-02
4.42390561e-01 -1.45582125e-01 -2.37628371e-01 4.60889973e-02
3.85066539e-01 -5.25440037e-01 -4.22232538e-01 1.06739271e+00
8.47003460e-01 -3.76005769e-02 3.22118312e-01 -1.14119661e+00
-6.46155894e-01 1.43671915e-01 -7.86559165e-01 4.59304392e-01
-5.71140528e-01 2.64125645e-01 7.97918856e-01 -1.06048381e+00
2.47625887e-01 6.89873576e-01 6.62079513e-01 1.97708428e-01
-7.77796984e-01 -9.71824527e-01 3.39656591e-01 1.46779686e-01
-1.72242093e+00 -2.45440632e-01 5.38828731e-01 -3.98110121e-01
1.09507644e+00 3.05670738e-01 8.95452559e-01 4.58376616e-01
2.03622356e-01 5.82551956e-01 1.27481163e+00 -1.22293457e-01
1.79228976e-01 -1.58769339e-02 -9.45874602e-02 2.49294758e-01
3.83642346e-01 1.85208902e-01 -2.05317393e-01 7.52408132e-02
7.35610664e-01 3.70711118e-01 -5.13106346e-01 3.29185367e-01
-5.97832561e-01 8.21786642e-01 9.72978830e-01 3.46538961e-01
-9.43392277e-01 2.14127854e-01 8.21562558e-02 3.68097238e-02
8.87496114e-01 -4.48525175e-02 -4.75086153e-01 2.64431953e-01
-1.12503016e+00 1.07338896e-03 2.25248918e-01 9.82265174e-01
6.93520129e-01 6.02969110e-01 -2.08114028e-01 4.58184361e-01
2.15222657e-01 1.11303818e+00 2.40605146e-01 -4.88811433e-01
2.59263366e-01 5.65370262e-01 1.41382217e-01 -8.56047451e-01
-2.25529194e-01 -1.03444397e+00 -1.33079529e+00 4.23471034e-01
-5.14924526e-01 -2.56129444e-01 -9.50750232e-01 1.31130981e+00
3.31586987e-01 5.81560135e-01 6.03461079e-02 1.14303434e+00
9.62249100e-01 8.87509763e-01 1.04427822e-01 5.15408106e-02
1.40848899e+00 -1.05550671e+00 -5.15030265e-01 -2.07612991e-01
1.58161774e-01 -4.19663578e-01 7.88832664e-01 -5.95271401e-02
-7.72068560e-01 -4.12533641e-01 -9.20510709e-01 5.70767038e-02
-7.00644553e-01 3.42372656e-01 6.94902241e-01 3.89861554e-01
-1.31672525e+00 4.06882972e-01 -6.66508794e-01 -1.79262131e-01
6.24789178e-01 2.07735941e-01 5.23776151e-02 -2.09558308e-01
-1.12623775e+00 6.22283161e-01 8.17631632e-02 5.79920888e-01
-9.71758127e-01 -7.86070347e-01 -5.79286754e-01 6.31483078e-01
-3.23927030e-02 -7.54124105e-01 6.75316572e-01 -1.08071148e+00
-1.25120735e+00 5.17350852e-01 -5.78009859e-02 -4.90183979e-01
1.67291403e-01 -3.36088806e-01 -5.32788396e-01 3.59792799e-01
7.56976753e-02 8.15592825e-01 1.28056276e+00 -1.18534219e+00
-8.07365716e-01 -2.70186096e-01 -5.02071455e-02 5.18835187e-01
-3.69645983e-01 6.55937493e-02 -4.16211128e-01 -7.15526521e-01
-2.01920137e-01 -8.66186440e-01 -1.78055719e-01 6.26374735e-03
-8.16284344e-02 2.67247260e-01 1.14315069e+00 -6.37735605e-01
1.06120837e+00 -2.24891186e+00 -8.89221653e-02 1.59754623e-02
4.41525996e-01 5.96935749e-01 -3.93409342e-01 1.76594347e-01
-1.38369575e-01 2.72088438e-01 -2.83896476e-01 -3.18286896e-01
-3.52991968e-01 1.79927703e-02 -3.47774625e-01 5.10136008e-01
3.10185820e-01 8.84669721e-01 -7.97278345e-01 -1.14220724e-01
3.65368009e-01 1.00369537e+00 -2.94125259e-01 4.57448699e-02
-8.33800435e-03 3.42487484e-01 -4.73002911e-01 1.07868850e+00
1.32129729e+00 -3.51846844e-01 -6.15468360e-02 -1.38130367e-01
-5.30605853e-01 -9.06600431e-02 -7.98462152e-01 9.77094829e-01
-4.84510839e-01 8.41557562e-01 5.74396074e-01 -4.27714258e-01
6.11991882e-01 2.63573021e-01 4.24026370e-01 -8.89398694e-01
1.54334813e-01 6.50514662e-02 -2.56213784e-01 -3.74216408e-01
6.61415935e-01 -2.29813159e-01 3.30500603e-01 1.25031501e-01
-3.37239295e-01 -1.75095778e-02 -4.99761999e-01 1.41831309e-01
8.55562687e-01 -1.54598951e-01 2.62955148e-02 -3.37657660e-01
2.91674167e-01 -6.05948493e-02 5.08802772e-01 6.47972345e-01
-2.92971909e-01 5.87109447e-01 -2.45536163e-01 -4.42215234e-01
-9.63392973e-01 -7.13271797e-01 -1.20974407e-01 8.10553432e-01
1.51996076e-01 -2.11295471e-01 -5.45750499e-01 -5.31000555e-01
1.23473316e-01 3.79796773e-01 -5.61257780e-01 3.75240631e-02
-2.58304715e-01 -1.12319815e+00 3.80544633e-01 6.61999464e-01
9.94814038e-01 -1.12505627e+00 -6.57171905e-01 -3.57002988e-02
-1.88976660e-01 -1.18918753e+00 -2.21313685e-01 9.66077745e-02
-9.80937064e-01 -8.44220579e-01 -6.39736593e-01 -4.38649654e-01
5.83765626e-01 1.02406251e+00 1.04852271e+00 4.15911853e-01
-1.17580518e-01 8.76159817e-02 -5.91852307e-01 -5.54336965e-01
2.82162666e-01 1.47891030e-01 -1.85112461e-01 1.42145932e-01
2.20492378e-01 -8.15420508e-01 -8.07384551e-01 -1.23963235e-02
-1.06717718e+00 7.44621754e-02 9.57841575e-01 8.78730357e-01
6.62988186e-01 5.01962066e-01 2.23733693e-01 -6.18924320e-01
1.91308647e-01 -6.58661246e-01 -8.30443680e-01 -1.26873747e-01
-6.44164503e-01 -5.51139772e-01 4.72558200e-01 -2.01910995e-02
-1.01602578e+00 7.73317069e-02 3.40287127e-02 -8.07102501e-01
-2.23691389e-02 6.57453299e-01 -3.44385743e-01 -3.90725136e-01
2.39310250e-01 5.51555336e-01 -4.39164400e-01 -4.95086879e-01
2.12056652e-01 8.51452470e-01 3.85794044e-01 3.62199917e-02
1.12104881e+00 8.51562738e-01 -2.13481069e-01 -1.14860070e+00
-6.28129363e-01 -5.54325283e-01 -4.74881351e-01 -3.66519779e-01
8.70855331e-01 -1.71479142e+00 -4.50009197e-01 9.06108141e-01
-8.88556242e-01 -5.92578828e-01 -2.29657248e-01 2.86471456e-01
2.83240348e-01 -2.26982161e-02 -4.92506176e-01 -9.57113028e-01
-1.05069363e+00 -7.25090802e-01 1.33359551e+00 4.26476985e-01
6.94260597e-01 -6.22062147e-01 -3.47070187e-01 2.30137303e-01
9.28531170e-01 2.38026574e-01 5.16244590e-01 7.83549845e-02
-1.24762142e+00 -5.28789498e-02 -7.13894546e-01 5.21257579e-01
4.21132296e-02 -5.00317197e-03 -1.36956167e+00 -6.43833518e-01
-2.24273764e-02 1.05630860e-01 1.21333408e+00 6.51204646e-01
1.28903222e+00 -6.80777356e-02 -2.45124146e-01 1.03444242e+00
1.82816327e+00 3.63847092e-02 1.01376510e+00 5.12974262e-01
9.42173898e-01 -6.46225689e-03 4.16126043e-01 5.66275358e-01
4.43505347e-01 2.03645512e-01 1.03904343e+00 -6.32545233e-01
-2.89937615e-01 2.89625734e-01 2.85470665e-01 8.84140730e-01
-3.37558389e-01 -3.28376681e-01 -6.54214323e-01 8.46177995e-01
-1.58114433e+00 -1.01600730e+00 -2.31284559e-01 1.90132892e+00
3.20447773e-01 -2.45610088e-01 -3.56521845e-01 -1.74443364e-01
6.57912731e-01 4.82287973e-01 -6.02408588e-01 1.61926940e-01
-5.43910980e-01 4.44392949e-01 8.39207828e-01 4.72666562e-01
-1.23132193e+00 1.07229304e+00 5.49663353e+00 8.52207124e-01
-1.46763372e+00 4.81538922e-01 4.50940311e-01 -1.22461118e-01
-2.26291448e-01 -1.78395938e-02 -6.89581811e-01 7.53925800e-01
8.32164705e-01 3.51237208e-01 4.50681359e-01 7.91937590e-01
4.29678053e-01 -1.77522898e-01 -1.10648528e-01 8.75921309e-01
-8.46576020e-02 -1.34452689e+00 -3.32519226e-02 1.96134269e-01
8.09711635e-01 8.77098322e-01 7.68037885e-02 3.67264211e-01
5.20060584e-02 -8.04285705e-01 5.76212406e-01 6.10727847e-01
1.09700000e+00 -7.48567581e-01 9.01522040e-01 1.39083594e-01
-1.61885285e+00 -2.61469543e-01 -3.76461118e-01 -2.02655792e-01
-1.65874913e-01 1.06516194e+00 -2.23689571e-01 8.20358813e-01
1.04160368e+00 8.31189573e-01 -5.13307154e-01 1.08058834e+00
-3.25983852e-01 6.46253109e-01 -4.41405714e-01 3.96104604e-01
3.29650879e-01 -3.94549549e-01 6.18391812e-01 1.25629938e+00
4.30810839e-01 4.33179259e-01 2.81516537e-02 6.67531192e-01
-2.80802667e-01 -2.27446571e-01 -4.01593566e-01 1.36134580e-01
7.36079276e-01 1.48125184e+00 -4.97304350e-01 -5.63277543e-01
-4.58419919e-01 1.09501898e+00 4.25230013e-03 3.97093385e-01
-9.68274534e-01 -2.82860756e-01 1.08578944e+00 2.88555548e-02
8.11472774e-01 -4.30226810e-02 -2.69011140e-01 -1.42239869e+00
1.11047193e-01 -8.31043065e-01 1.77511841e-01 -1.29527724e+00
-1.09185493e+00 1.01948941e+00 -3.40780616e-01 -1.44954455e+00
4.66123164e-01 -3.82506728e-01 -9.32453096e-01 1.17724872e+00
-2.35596728e+00 -1.43664825e+00 -1.13674617e+00 5.80689728e-01
4.59135562e-01 -2.35583913e-03 9.10011411e-01 5.22958279e-01
-8.54373336e-01 2.59255111e-01 4.39800113e-01 -4.08171788e-02
2.75931865e-01 -9.33135986e-01 4.03878838e-01 1.24028075e+00
-3.24611396e-01 1.82851359e-01 3.39690745e-01 -7.33112395e-01
-1.25386178e+00 -1.80709040e+00 6.65371835e-01 2.66994387e-02
3.89915347e-01 -2.63756394e-01 -7.78510809e-01 5.74970186e-01
4.01106298e-01 4.32968527e-01 2.84845173e-01 -2.84081876e-01
-1.94616929e-01 -4.75279689e-01 -1.10057545e+00 2.77208477e-01
7.45802462e-01 -7.18132675e-01 2.33756825e-01 3.33536237e-01
7.28659928e-01 -4.05046761e-01 -6.82883739e-01 4.46143240e-01
2.53047496e-01 -9.36890721e-01 6.93114161e-01 -5.51007614e-02
3.77144724e-01 -6.97802544e-01 -2.87216395e-01 -1.34649277e+00
-8.52995872e-01 -3.01325530e-01 -2.67536104e-01 1.14528263e+00
8.14801604e-02 -6.92703843e-01 6.46999300e-01 1.94525838e-01
-3.75882477e-01 -6.69621944e-01 -8.52106452e-01 -7.58470476e-01
-7.87783489e-02 -3.15760702e-01 9.07485127e-01 1.22209430e+00
-8.94621432e-01 1.52289644e-01 -3.80608797e-01 9.28165913e-01
5.26677728e-01 5.45766354e-01 4.77527291e-01 -1.14761913e+00
-5.25380820e-02 -2.70116478e-01 -2.11421430e-01 -7.38619328e-01
-5.84835820e-02 -6.67819440e-01 -1.50660366e-01 -1.31987560e+00
3.52395117e-01 -4.71403480e-01 -3.45640689e-01 6.69677734e-01
-3.39682043e-01 5.43561697e-01 4.49714452e-01 3.71223569e-01
-2.88268387e-01 9.88088906e-01 9.59876359e-01 -1.97524652e-01
-1.81251198e-01 -1.03929467e-01 -6.36424184e-01 5.42441070e-01
1.05087626e+00 -6.50779366e-01 -6.89338446e-02 -7.98953652e-01
-3.53932306e-02 -4.01683837e-01 6.67971134e-01 -1.17824066e+00
1.25222325e-01 2.65396032e-02 5.99162221e-01 -8.04307222e-01
3.89850467e-01 -1.06108356e+00 1.57055870e-01 3.54381770e-01
5.92597127e-01 7.78876394e-02 5.53317666e-01 5.28266251e-01
-2.60552198e-01 2.56139040e-01 8.40435863e-01 -5.57829887e-02
-7.91852593e-01 6.68479919e-01 -5.13343811e-01 -4.18179572e-01
8.30266178e-01 -7.27137700e-02 -3.71029615e-01 -3.76674771e-01
-6.00252926e-01 3.66002709e-01 3.74492705e-01 1.57012776e-01
6.87092483e-01 -1.10262990e+00 -9.36598957e-01 3.53165537e-01
3.62110813e-03 2.66078293e-01 7.38804460e-01 9.45207834e-01
-5.25349259e-01 2.27126032e-01 -7.40869716e-02 -5.62003970e-01
-1.22556829e+00 3.67180824e-01 6.21501327e-01 -2.47673333e-01
-8.00328612e-01 7.06569076e-01 2.70613581e-01 -1.22231580e-01
-1.87295452e-01 -2.67350823e-01 -9.72286984e-03 7.82160647e-03
6.39248192e-01 2.60361016e-01 2.75131255e-01 -6.78670287e-01
-3.46464336e-01 3.87751788e-01 1.72714576e-01 2.80240268e-01
1.80911171e+00 -3.56301367e-01 -2.57454842e-01 -1.78185791e-01
7.18920231e-01 3.86777855e-02 -1.44691706e+00 -3.57666969e-01
-7.27863789e-01 -5.81345499e-01 8.40498984e-01 -1.04912066e+00
-1.83584869e+00 8.25484455e-01 9.75322664e-01 4.69170474e-02
1.78460169e+00 -2.64384449e-01 8.24733734e-01 -3.18073630e-02
1.34214014e-01 -7.48720109e-01 -3.87050718e-01 6.70594811e-01
9.46485519e-01 -1.25490594e+00 2.11506754e-01 -3.66892248e-01
-5.98086834e-01 5.97374201e-01 6.10147417e-01 -1.13984093e-01
8.48609686e-01 4.42946970e-01 1.01864159e-01 -5.14395773e-01
-6.23866737e-01 -6.46789134e-01 4.01739068e-02 5.91356933e-01
-5.70995919e-02 4.89545971e-01 3.27096701e-01 4.89537299e-01
1.91961795e-01 -1.19519152e-01 3.87647897e-01 7.84449458e-01
-2.66347021e-01 -5.96175253e-01 -4.24009711e-01 5.11303306e-01
-2.75914341e-01 -9.11705971e-01 -1.46240979e-01 9.38812017e-01
4.40851182e-01 8.68771493e-01 2.50392497e-01 -5.14670253e-01
-2.43063876e-03 -1.99368730e-01 -1.31161526e-01 -3.67273867e-01
-9.09113944e-01 3.85370493e-01 -2.14902416e-01 -5.30003846e-01
-5.34497559e-01 -6.25771523e-01 -8.65300357e-01 -6.04850829e-01
-7.06173778e-01 -1.08351991e-01 8.01640928e-01 7.03378916e-01
1.00716925e+00 7.43885398e-01 8.36022019e-01 -1.23081577e+00
-4.36530672e-02 -1.15763855e+00 -8.76324713e-01 -3.12083345e-02
6.92533076e-01 -3.71337563e-01 -2.87110597e-01 -1.17745854e-01] | [9.894438743591309, -1.798379898071289] |
214c5eef-3cbd-4c10-be9e-ab65834b0bff | lightweight-attribute-localizing-models-for | 2306.09822 | null | https://arxiv.org/abs/2306.09822v1 | https://arxiv.org/pdf/2306.09822v1.pdf | Lightweight Attribute Localizing Models for Pedestrian Attribute Recognition | Pedestrian Attribute Recognition (PAR) deals with the problem of identifying features in a pedestrian image. It has found interesting applications in person retrieval, suspect re-identification and soft biometrics. In the past few years, several Deep Neural Networks (DNNs) have been designed to solve the task; however, the developed DNNs predominantly suffer from over-parameterization and high computational complexity. These problems hinder them from being exploited in resource-constrained embedded devices with limited memory and computational capacity. By reducing a network's layers using effective compression techniques, such as tensor decomposition, neural network compression is an effective method to tackle these problems. We propose novel Lightweight Attribute Localizing Models (LWALM) for Pedestrian Attribute Recognition (PAR). LWALM is a compressed neural network obtained after effective layer-wise compression of the Attribute Localization Model (ALM) using the Canonical Polyadic Decomposition with Error Preserving Correction (CPD-EPC) algorithm. | ['Andrzej Cichocki', 'Ahmed Mohamed Khedr', 'Thar Baker', 'Zaher Al Aghbari', 'Imran Junejo', 'Naveed Ahmed', 'Salman Ahmadi-Asl', 'Anh Huy Phan', 'Konstantin Sobolev', 'Dimitrii Ermilov', 'Ashish Jha'] | 2023-06-16 | null | null | null | null | ['pedestrian-attribute-recognition', 'person-retrieval', 'neural-network-compression', 'neural-network-compression'] | ['computer-vision', 'computer-vision', 'methodology', 'miscellaneous'] | [ 3.95701051e-01 -5.07723272e-01 -1.97603360e-01 -4.54341859e-01
-2.35998183e-01 -7.22107068e-02 7.06724644e-01 3.66544873e-01
-5.15006959e-01 5.38858771e-01 1.66047230e-01 -9.58422720e-02
-9.94067565e-02 -9.18090880e-01 -5.58616519e-01 -9.89875853e-01
-1.22757077e-01 3.00470769e-01 -1.83807492e-01 -2.23176870e-02
1.50911972e-01 6.79812908e-01 -1.85362136e+00 4.20067847e-01
7.57779419e-01 1.49965584e+00 -1.95210084e-01 3.77765745e-01
-7.17228428e-02 5.94893992e-01 -2.57524580e-01 -1.14534080e+00
3.57973993e-01 -1.94253754e-02 -7.36943424e-01 -7.29647726e-02
7.07100272e-01 -3.05836588e-01 -6.71309888e-01 1.29739344e+00
6.31703317e-01 1.74011394e-01 6.18421137e-01 -1.44380593e+00
-6.08265519e-01 2.62312174e-01 -3.30759853e-01 9.01409388e-02
1.91044167e-01 -1.50206491e-01 7.64056921e-01 -9.90434229e-01
1.24774590e-01 1.48888493e+00 9.44014966e-01 5.75080931e-01
-1.08857274e+00 -3.26392323e-01 -2.28625178e-01 1.01512444e+00
-1.71191347e+00 -4.46536273e-01 7.41479695e-01 -2.08790049e-01
8.84070516e-01 5.53410709e-01 7.95593560e-01 1.08376420e+00
1.64776504e-01 8.68271410e-01 8.30678105e-01 -2.56991327e-01
2.41673931e-01 -1.57611996e-01 1.89273775e-01 4.94933695e-01
4.92306679e-01 -1.56533509e-01 -6.74342394e-01 -2.65377223e-01
6.23148203e-01 2.36069500e-01 4.91245016e-02 -1.64447010e-01
-1.15281951e+00 6.42871737e-01 3.55332553e-01 -2.04476211e-02
-6.64238632e-01 3.75210010e-02 9.54606295e-01 2.40305662e-01
3.26370984e-01 -2.48685814e-02 -1.22292645e-01 -2.71316350e-01
-8.20644617e-01 1.98358208e-01 5.37791193e-01 9.47231412e-01
4.68625128e-01 1.64080933e-01 2.93876939e-02 1.05525923e+00
6.57582730e-02 4.64809477e-01 4.74758118e-01 -8.84012759e-01
5.46803117e-01 7.66354561e-01 -2.95026630e-01 -1.47482061e+00
-2.29424715e-01 -3.48029852e-01 -1.77696598e+00 -1.63000509e-01
4.01473671e-01 3.40003490e-01 -4.87800092e-01 1.29936147e+00
3.28700721e-01 1.48093805e-01 5.31197377e-02 8.46200526e-01
6.88454270e-01 5.90482354e-01 4.43134636e-01 1.97230950e-01
1.65796947e+00 -5.96717417e-01 -5.35200298e-01 1.15196314e-02
6.12092555e-01 -6.65385008e-01 7.48254240e-01 5.37642479e-01
-9.14178669e-01 -8.40067685e-01 -1.25696564e+00 -2.35774487e-01
-5.90618610e-01 3.35231036e-01 7.77528703e-01 1.00577009e+00
-9.20806885e-01 5.39846122e-01 -5.72790086e-01 -4.03402984e-01
7.55491972e-01 7.35200644e-01 -7.49315977e-01 -3.56087893e-01
-1.26344514e+00 6.29740596e-01 3.54748636e-01 4.64704335e-01
-5.72112560e-01 -2.72964567e-01 -6.86030507e-01 1.06792368e-01
-7.67051578e-02 -8.03651810e-01 4.88518596e-01 -7.01228440e-01
-1.18747222e+00 8.20349574e-01 -2.28679910e-01 -8.04372430e-01
1.83035642e-01 -2.72662818e-01 -4.85538810e-01 2.24332154e-01
-1.06564291e-01 5.49258113e-01 9.51597214e-01 -5.48294961e-01
-4.79831636e-01 -7.76907086e-01 -1.75688177e-01 8.01567584e-02
-8.69995892e-01 2.01448575e-01 -1.12974077e-01 -8.05291533e-01
4.22011852e-01 -8.73819828e-01 -1.20905414e-01 6.27879500e-02
-5.70212305e-01 -3.57490331e-01 9.28504527e-01 -9.42841232e-01
1.01433194e+00 -2.09930420e+00 1.43579036e-01 3.39028955e-01
3.99354100e-01 7.00571597e-01 -7.85804614e-02 6.07527494e-02
-1.56974778e-01 -9.50043872e-02 -6.30827397e-02 -6.60022795e-01
-6.34967834e-02 2.97884732e-01 -9.81056243e-02 6.68422997e-01
-3.97231728e-02 6.27260625e-01 -3.37172955e-01 -7.42726564e-01
2.42092311e-01 8.71002972e-01 -4.13097531e-01 5.18854521e-03
1.77829117e-01 1.79606289e-01 -3.59196782e-01 9.88160133e-01
1.17740917e+00 1.79484889e-01 -1.33950055e-01 -5.46459794e-01
7.25140497e-02 -1.59604400e-01 -1.31721675e+00 1.34722996e+00
-1.19784638e-01 7.58574963e-01 -5.66847529e-03 -1.43796933e+00
1.13126910e+00 3.40350091e-01 5.88449597e-01 -6.93854272e-01
1.28806517e-01 3.88356149e-01 -1.94523260e-01 -6.10364795e-01
7.89004207e-01 4.34346825e-01 5.63829839e-02 1.54056296e-01
5.37554314e-03 9.61013019e-01 1.65801104e-02 7.67627954e-02
7.59367168e-01 -2.92511225e-01 2.43790910e-01 -1.51518822e-01
1.18172324e+00 -3.90376061e-01 6.58612847e-01 6.25573754e-01
-3.73539716e-01 3.22557807e-01 2.59137094e-01 -1.24380767e+00
-1.71848524e+00 -6.01203561e-01 -3.54799271e-01 9.77632642e-01
-1.52322724e-01 -5.75098753e-01 -1.02267575e+00 -2.04965681e-01
-1.52139381e-01 1.64815187e-01 -3.33013058e-01 -3.22065890e-01
-7.91724265e-01 -9.65469539e-01 1.11343443e+00 7.12271869e-01
1.46765363e+00 -6.29463613e-01 -4.52550590e-01 2.23128960e-01
-4.21395272e-01 -1.37995338e+00 -2.02434972e-01 -3.17645252e-01
-8.51100564e-01 -8.02791715e-01 -8.42805684e-01 -7.13719308e-01
6.77535236e-01 1.56526417e-01 8.66601348e-01 1.54867023e-01
-2.41523236e-01 2.78897345e-01 -2.37736732e-01 -1.65670887e-01
-2.08075061e-01 1.09711662e-01 6.66113794e-01 6.87789738e-01
9.30452526e-01 -6.37445569e-01 -7.18634307e-01 3.66248727e-01
-6.23993576e-01 1.38701294e-02 6.48950040e-01 9.37552094e-01
6.19833350e-01 5.92642576e-02 1.63746223e-01 -3.05375695e-01
6.27329230e-01 -5.24621345e-02 -4.99142230e-01 2.28832424e-01
-3.35023463e-01 1.57722924e-02 8.30747366e-01 -4.77606416e-01
-6.84009731e-01 4.40842211e-02 -3.73719424e-01 -3.01005244e-01
-3.82466733e-01 1.83211595e-01 -6.30886614e-01 -4.38183635e-01
4.13839579e-01 6.36530221e-01 1.24257781e-01 -5.50040901e-01
-2.20242050e-02 8.87255907e-01 7.23852277e-01 -7.19851375e-01
4.64995503e-01 3.98263335e-01 7.03347743e-01 -1.09163523e+00
-3.60190868e-01 -1.74502909e-01 -7.62359202e-01 -2.45792374e-01
8.76529396e-01 -7.33751059e-01 -1.30221653e+00 9.32715952e-01
-1.38916874e+00 5.13118565e-01 -6.54649511e-02 5.01682281e-01
-3.20949972e-01 7.95738995e-01 -6.28457606e-01 -7.88029671e-01
-7.35529721e-01 -1.19852304e+00 7.06863225e-01 4.13260125e-02
1.05942033e-01 -5.62542975e-01 -5.59684873e-01 4.72678870e-01
6.61275029e-01 2.70221978e-01 1.05552495e+00 -7.57884145e-01
-6.28120542e-01 -6.07239783e-01 -5.10654390e-01 6.30042493e-01
-4.18037772e-01 -5.03895760e-01 -9.61091995e-01 -3.04781049e-01
-2.84623671e-02 -2.63650306e-02 7.55572081e-01 3.59980553e-01
1.56542814e+00 -7.86819160e-01 -2.12183282e-01 9.22876656e-01
1.13100684e+00 -1.00237563e-01 8.87784779e-01 6.97425306e-01
8.80581141e-01 5.14244616e-01 2.72070050e-01 5.40439665e-01
2.93418825e-01 8.11847806e-01 4.30440456e-01 2.60509729e-01
-3.29267196e-02 2.56257262e-02 1.63263857e-01 8.76187086e-01
-7.43604541e-01 -3.05885170e-02 -9.24564779e-01 2.26459756e-01
-1.78957224e+00 -1.01195717e+00 -2.23954931e-01 2.46154571e+00
3.45957488e-01 -2.81681009e-02 1.09308504e-01 6.52727783e-01
8.85697901e-01 -1.01432130e-02 -4.71300572e-01 -5.05201757e-01
-5.58510721e-01 -1.10160150e-01 4.85264808e-01 1.08253129e-01
-1.30132520e+00 5.32949567e-01 5.07243681e+00 1.09702337e+00
-6.40633821e-01 1.52457118e-01 7.90114939e-01 3.81390870e-01
2.41617337e-01 -1.83730423e-01 -1.06614852e+00 6.49223268e-01
1.16444147e+00 4.05534327e-01 3.51425320e-01 9.89901066e-01
-1.24654479e-01 1.47761166e-01 -1.09221470e+00 1.49612153e+00
1.42225698e-01 -1.30939746e+00 4.85359639e-01 2.75272429e-01
1.00257076e-01 -4.07834291e-01 3.65611941e-01 4.60888911e-03
-4.57457691e-01 -9.87121344e-01 3.05814773e-01 5.92206836e-01
9.17494714e-01 -1.06098592e+00 1.14616024e+00 8.96253362e-02
-1.46448171e+00 -2.59182960e-01 -9.69707251e-01 4.69753407e-02
-1.64747119e-01 6.01842940e-01 -4.46705192e-01 4.60243672e-01
8.49859118e-01 6.09285712e-01 -7.42151856e-01 1.05860460e+00
4.01368558e-01 4.95261937e-01 -3.42659682e-01 -6.20832630e-02
1.27556369e-01 -4.38031167e-01 7.70814896e-01 1.07102478e+00
3.53122652e-01 -4.01060004e-03 -2.17937693e-01 4.90642130e-01
-1.29539177e-01 9.93627757e-02 -4.81687188e-01 1.42268077e-01
6.26759887e-01 9.56984460e-01 -3.60829294e-01 -3.34422022e-01
-2.55716383e-01 1.02782655e+00 -1.77076496e-02 1.27066508e-01
-5.10790348e-01 -5.01825631e-01 8.93278897e-01 1.08029589e-01
-8.83558393e-02 -4.66951370e-01 -4.14803743e-01 -1.00424254e+00
1.98772326e-01 -9.64295685e-01 4.19314921e-01 -3.26964706e-01
-1.26956177e+00 7.20181286e-01 -1.03293985e-01 -1.46880150e+00
-9.39390659e-02 -7.62991846e-01 -2.23554105e-01 7.76411712e-01
-1.15734577e+00 -1.49315417e+00 -4.70319450e-01 9.12887990e-01
3.45769584e-01 -8.07726800e-01 1.02066386e+00 8.83117676e-01
-7.03839839e-01 1.12875986e+00 1.30260631e-01 3.58437181e-01
3.64782125e-01 -8.45850945e-01 4.92805988e-01 9.04831171e-01
-3.60821515e-01 7.41410017e-01 6.04288876e-01 -3.74116719e-01
-1.58678174e+00 -1.12815189e+00 1.12263644e+00 -4.43761386e-02
1.39528677e-01 -3.05213779e-01 -8.86297882e-01 2.90689468e-01
-2.62534302e-02 2.08985567e-01 9.31727231e-01 -4.65047657e-02
-5.88247001e-01 -5.27756274e-01 -1.33665192e+00 4.28602278e-01
8.85734618e-01 -6.57260120e-01 7.64188841e-02 3.34674031e-01
4.35946196e-01 -1.25301212e-01 -1.00113606e+00 2.29643792e-01
6.50988758e-01 -9.14547563e-01 1.54736233e+00 -6.17399037e-01
2.19687656e-01 -2.81762511e-01 -6.53495014e-01 -6.72596395e-01
-3.15887392e-01 -4.16074812e-01 -3.67880702e-01 1.31858397e+00
-2.43340835e-01 -6.55651152e-01 9.39954102e-01 9.38361526e-01
-3.49601433e-02 -8.31560075e-01 -1.38861609e+00 -6.73492670e-01
-3.13383996e-01 -3.66535276e-01 9.57390606e-01 7.87040412e-01
-5.13120890e-01 6.35813689e-03 -6.96002424e-01 1.38005838e-01
1.22218835e+00 -4.84949708e-01 6.06215656e-01 -1.58143842e+00
1.51179388e-01 -2.91265607e-01 -1.30321574e+00 -9.56156015e-01
4.51477394e-02 -7.25893199e-01 -6.79150641e-01 -1.07008457e+00
2.84929395e-01 -4.48475301e-01 -1.15739755e-01 2.79934436e-01
2.99620092e-01 4.50446397e-01 1.72114551e-01 4.70073313e-01
-5.74243367e-01 6.57325923e-01 7.38926530e-01 -4.67183232e-01
2.53290713e-01 1.44927889e-01 -2.37663373e-01 7.19060957e-01
7.88325369e-01 -1.95899442e-01 8.19976181e-02 -4.65056419e-01
2.89196581e-01 -2.58879870e-01 7.77705014e-01 -1.55552590e+00
6.88122988e-01 3.54993165e-01 8.16353381e-01 -7.73260534e-01
7.24885643e-01 -9.05782759e-01 7.26862624e-02 4.70127761e-01
-3.32172602e-01 6.72401965e-01 -2.13549957e-02 4.94812816e-01
-4.19619262e-01 -1.64949924e-01 6.87912405e-01 8.03690180e-02
-8.98601115e-01 5.71546972e-01 -4.74770784e-01 -5.71556330e-01
7.93525755e-01 -6.45392120e-01 -3.47806811e-01 -1.91820547e-01
-5.76123536e-01 -2.81818837e-01 7.19590336e-02 7.90756494e-02
1.16228843e+00 -1.83937645e+00 -8.39758515e-01 4.90026146e-01
-9.77697745e-02 -5.65359741e-02 5.24337351e-01 7.67047524e-01
-6.76325083e-01 6.08556330e-01 -5.42749226e-01 -6.45267606e-01
-1.41379952e+00 4.92016524e-01 2.04747319e-01 -6.64059371e-02
-6.85455382e-01 8.20702374e-01 -4.58308280e-01 -1.71096370e-01
4.35720712e-01 4.00150150e-01 -4.30270702e-01 -5.96789457e-02
6.95957124e-01 8.36978674e-01 2.76878506e-01 -1.12063563e+00
-2.59789914e-01 5.88967383e-01 -2.63093650e-01 3.04011852e-01
1.27074623e+00 -1.06156811e-01 -5.95543981e-01 -2.16145605e-01
1.39717841e+00 -6.81812704e-01 -7.65417933e-01 -5.13324916e-01
-1.01712979e-01 -5.40506184e-01 6.14026748e-02 2.92024557e-02
-1.00576317e+00 1.12952709e+00 9.47540462e-01 -4.02973732e-03
1.22443795e+00 -4.89948124e-01 1.23137712e+00 8.89273703e-01
3.41696173e-01 -1.35976267e+00 -2.80254055e-02 4.21447277e-01
5.94654679e-01 -9.27811086e-01 1.50770992e-01 -2.06533968e-01
-3.27524245e-01 1.37046146e+00 4.38049406e-01 3.80568467e-02
5.93070865e-01 -9.64905620e-02 -6.17093444e-01 1.62410140e-01
-1.80308938e-01 2.23474473e-01 2.47776255e-01 7.74760306e-01
1.97774485e-01 1.87258899e-01 -1.92229360e-01 5.28422713e-01
-2.17367083e-01 -4.31403548e-01 2.09897667e-01 6.12485170e-01
-2.43803591e-01 -1.28477323e+00 -7.32835352e-01 6.61926508e-01
-3.51059675e-01 -1.63371116e-01 3.46892253e-02 2.30039135e-01
3.46511990e-01 7.26297617e-01 7.90792480e-02 -9.77325499e-01
1.72528867e-02 4.93305363e-02 2.75835216e-01 3.60801429e-01
-2.44071662e-01 -5.36969483e-01 1.06153218e-02 -5.52562654e-01
-5.54873586e-01 -8.45637321e-01 -5.69032669e-01 -9.20501888e-01
1.62938118e-01 -4.94405860e-03 8.70236278e-01 8.48601043e-01
5.21198273e-01 1.85717538e-01 4.58467424e-01 -7.85264075e-01
-6.08954787e-01 -8.03761721e-01 -4.14608747e-01 4.80558068e-01
2.24535197e-01 -5.92424631e-01 9.06411633e-02 7.71606192e-02] | [14.571295738220215, 1.0010404586791992] |
c34238c5-c275-4ba7-8c4d-acd5a6c92283 | extracting-fine-grained-knowledge-graphs-of-1 | null | null | https://aclanthology.org/2021.emnlp-main.381 | https://aclanthology.org/2021.emnlp-main.381.pdf | Extracting Fine-Grained Knowledge Graphs of Scientific Claims: Dataset and Transformer-Based Results | Recent transformer-based approaches demonstrate promising results on relational scientific information extraction. Existing datasets focus on high-level description of how research is carried out. Instead we focus on the subtleties of how experimental associations are presented by building SciClaim, a dataset of scientific claims drawn from Social and Behavior Science (SBS), PubMed, and CORD-19 papers. Our novel graph annotation schema incorporates not only coarse-grained entity spans as nodes and relations as edges between them, but also fine-grained attributes that modify entities and their relations, for a total of 12,738 labels in the corpus. By including more label types and more than twice the label density of previous datasets, SciClaim captures causal, comparative, predictive, statistical, and proportional associations over experimental variables along with their qualifications, subtypes, and evidence. We extend work in transformer-based joint entity and relation extraction to effectively infer our schema, showing the promise of fine-grained knowledge graphs in scientific claims and beyond. | ['Scott Friedman', 'Ian Magnusson'] | null | null | null | null | emnlp-2021-11 | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [ 4.60477732e-02 5.09186745e-01 -1.07576263e+00 -1.66101336e-01
-5.38354993e-01 -8.80283773e-01 1.01614296e+00 1.20613277e+00
-2.08780274e-01 1.17427719e+00 7.09627926e-01 -7.74984896e-01
-5.83316147e-01 -9.92106915e-01 -9.07719672e-01 -1.19181857e-01
-8.17270875e-02 5.33125281e-01 1.42507747e-01 1.82779685e-01
5.06962538e-01 5.58786809e-01 -8.68052661e-01 2.23126680e-01
6.55682445e-01 5.83380163e-01 -4.69815910e-01 2.62834549e-01
-2.90507287e-01 1.24975157e+00 -6.60181522e-01 -9.73257303e-01
-5.96578240e-01 -2.24161491e-01 -9.76488829e-01 -5.38097620e-01
5.77639222e-01 3.29250127e-01 -4.77454573e-01 8.47716689e-01
3.15969080e-01 -3.57900113e-01 9.26251173e-01 -1.26941717e+00
-1.03845835e+00 1.31447208e+00 -7.43298471e-01 5.27346253e-01
5.15024364e-01 -1.95869282e-01 1.40695131e+00 -3.60197395e-01
1.57061696e+00 1.11159623e+00 7.79181898e-01 1.33574262e-01
-1.48578644e+00 -8.07501793e-01 -9.27145183e-02 2.08376229e-01
-1.17339814e+00 -3.04785669e-01 2.60887712e-01 -9.95589018e-01
9.66686964e-01 2.58732468e-01 5.19331753e-01 1.21408629e+00
8.32227618e-03 -6.01543151e-02 1.31791949e+00 -2.75974303e-01
8.25433806e-03 3.45147401e-02 5.60308158e-01 9.55699444e-01
9.77496207e-01 -2.43011355e-01 -6.99018896e-01 -6.63404107e-01
5.44222057e-01 -2.00935423e-01 -4.69935276e-02 1.20293096e-01
-1.45476878e+00 4.12830442e-01 2.51598001e-01 3.30847919e-01
-4.38741893e-01 6.85805753e-02 4.32381809e-01 -9.52986926e-02
4.79249328e-01 9.81313467e-01 -7.96473026e-01 1.35542914e-01
-8.83082151e-01 2.60406941e-01 9.79969800e-01 1.21930063e+00
5.87055624e-01 -6.55917585e-01 -5.79097211e-01 6.22527540e-01
-2.62345187e-02 1.17015420e-02 -9.27171260e-02 -1.26610160e+00
3.81510556e-01 1.01481771e+00 6.32492406e-03 -1.24754870e+00
-7.83772051e-01 -5.61592281e-01 -5.57463527e-01 -3.87117356e-01
2.93297857e-01 -4.16765548e-02 -7.40362823e-01 1.72339952e+00
3.66981983e-01 1.85811982e-01 -3.02199572e-01 2.66165972e-01
1.64136469e+00 -1.35637447e-01 8.26865613e-01 -1.87950701e-01
1.87123406e+00 -2.05313817e-01 -1.05674517e+00 3.05029333e-01
9.13088918e-01 -4.89807039e-01 4.23930138e-01 -1.44670233e-01
-9.73828256e-01 2.12500751e-01 -7.78655052e-01 -4.28730935e-01
-8.52218866e-01 -1.94041327e-01 1.06287622e+00 1.96344674e-01
-7.95058072e-01 6.02689147e-01 -2.49528185e-01 -1.11149997e-01
9.42964971e-01 1.16226532e-01 -6.66372418e-01 2.18350381e-01
-1.21720111e+00 1.07948732e+00 2.27428421e-01 -4.72772837e-01
-2.11042166e-01 -1.52282953e+00 -8.13256025e-01 1.81988716e-01
8.84970725e-01 -1.01243138e+00 4.60059106e-01 5.15923798e-01
-4.22092706e-01 1.27122009e+00 -1.20168224e-01 -2.85253316e-01
-1.30155951e-01 3.84354174e-01 -6.59310699e-01 -5.26327826e-02
4.65225369e-01 3.20505559e-01 -2.33833387e-01 -8.28008890e-01
-3.64963502e-01 -5.37033737e-01 2.09316418e-01 -2.55936801e-01
-2.03050315e-01 3.80971998e-01 -3.77299726e-01 -4.94865239e-01
-2.34497890e-01 -7.87783921e-01 -5.42585403e-02 -1.19373970e-01
-1.00530994e+00 -6.77790284e-01 3.54492456e-01 -4.55880553e-01
1.53216505e+00 -1.70299482e+00 2.16195017e-01 1.77859828e-01
1.02714407e+00 -3.58474493e-01 3.21853608e-01 5.25562286e-01
-3.00140053e-01 9.73831952e-01 1.02089070e-01 1.47225663e-01
1.02384247e-01 1.22757487e-01 -3.09042782e-02 3.69730562e-01
1.83550194e-01 1.31052041e+00 -1.04630220e+00 -1.06555820e+00
-4.15346116e-01 1.14250757e-01 -3.84187669e-01 -4.87086356e-01
-1.49187893e-01 6.92815557e-02 -8.02102387e-01 7.70249248e-01
1.39332697e-01 -7.41963327e-01 4.43219304e-01 -7.52767324e-01
-2.70169139e-01 7.81926095e-01 -4.59277838e-01 1.33953381e+00
7.00572506e-02 6.58462167e-01 -3.48745496e-03 -8.44997466e-01
6.06674492e-01 2.08235592e-01 6.93026066e-01 -2.28827000e-01
-6.89515397e-02 8.15859064e-02 -1.33440527e-03 -5.19597471e-01
2.54274309e-01 6.25695512e-02 -3.35088909e-01 3.63239676e-01
1.69407666e-01 6.36321492e-03 5.40854156e-01 8.48199546e-01
1.66064870e+00 1.78368971e-01 4.49331999e-01 -5.12922645e-01
2.50563294e-01 3.01514298e-01 6.73426032e-01 6.28800809e-01
3.13888580e-01 9.69979819e-03 1.23504484e+00 3.32021452e-02
-9.50044096e-01 -7.68209875e-01 -6.68925464e-01 7.20708370e-01
-3.27680677e-01 -8.98598373e-01 -2.21898660e-01 -7.45582461e-01
5.05033195e-01 6.73126459e-01 -1.26562274e+00 5.32780625e-02
-2.05122307e-01 -9.92245257e-01 8.28250527e-01 2.33046666e-01
1.31028686e-02 -7.19159305e-01 1.33779511e-01 -6.18760660e-02
-1.77258089e-01 -1.50295091e+00 -2.18143106e-01 1.59724638e-01
-5.10334551e-01 -1.59586310e+00 -1.40611529e-01 -2.94716597e-01
5.00570059e-01 -4.11656737e-01 1.54573274e+00 4.50565070e-02
-5.06067455e-01 -5.68447225e-02 3.20047550e-02 -6.61420941e-01
-3.46657783e-01 -5.01283742e-02 -3.48874122e-01 -7.07638800e-01
4.38804775e-01 -5.53210199e-01 -3.39669585e-01 7.48371780e-02
-5.89257181e-01 1.65180769e-02 5.08542538e-01 6.99615777e-01
6.32262290e-01 -1.50343388e-01 7.40883231e-01 -1.65851617e+00
6.15744650e-01 -9.40817416e-01 -3.03983837e-01 5.50123334e-01
-1.03959334e+00 1.96892202e-01 7.42468343e-04 -1.83709830e-01
-7.11537421e-01 -6.92536891e-01 2.50206560e-01 3.75922859e-01
-1.39540145e-02 1.03494275e+00 9.98915080e-03 -5.51329069e-02
7.66725302e-01 -6.14841044e-01 -2.35003725e-01 -3.63166243e-01
6.53266490e-01 4.45451587e-01 4.95221883e-01 -9.99967098e-01
3.65474820e-01 2.64014989e-01 7.46230423e-01 -3.18425298e-01
-1.08690834e+00 -2.97547370e-01 -4.79628623e-01 2.29003832e-01
8.59696090e-01 -7.66799390e-01 -1.49979627e+00 -2.42943257e-01
-9.56656456e-01 2.31413059e-02 -4.05244738e-01 5.97843051e-01
1.93207059e-02 1.70210406e-01 -9.52416480e-01 -2.94332296e-01
1.07251080e-02 -8.74522686e-01 9.51686323e-01 -2.51106955e-02
-6.05885088e-01 -1.25091767e+00 1.36685699e-01 5.73276162e-01
8.67462903e-03 8.04096699e-01 1.50454938e+00 -7.86155224e-01
-3.35082620e-01 -1.01417065e-01 -6.74570620e-01 -7.40816116e-01
1.38979450e-01 2.56372869e-01 -3.24626267e-01 4.81099248e-01
-7.75613129e-01 -1.30973667e-01 8.79080951e-01 1.86168537e-01
1.31790233e+00 -5.24703622e-01 -9.01339054e-01 3.12563151e-01
1.22761548e+00 -9.12560299e-02 3.31635475e-01 3.33058238e-01
1.09473765e+00 8.46837699e-01 -1.32928476e-01 7.98385888e-02
6.68399632e-01 8.10821831e-01 3.69951613e-02 -1.90737367e-01
-4.23763037e-01 -2.61363745e-01 -3.12203288e-01 4.35512900e-01
-4.02935475e-01 -2.13157028e-01 -9.21450436e-01 5.25411606e-01
-1.44646978e+00 -1.07203531e+00 -9.18504775e-01 1.79532015e+00
1.64236629e+00 2.16829240e-01 -5.88771291e-02 -2.35115573e-01
6.33995295e-01 -1.57233268e-01 -4.34190989e-01 -1.74738750e-01
-4.94381011e-01 2.63095975e-01 8.59271407e-01 2.71389991e-01
-5.74748874e-01 8.07131588e-01 6.86122131e+00 9.87626791e-01
-4.82870281e-01 1.62537217e-01 6.31329000e-01 -3.10197263e-03
-8.92527163e-01 3.24962139e-01 -8.26943994e-01 3.68502229e-01
1.02967966e+00 -3.84397298e-01 6.95135072e-02 5.15061915e-02
2.75229421e-02 -6.90008998e-02 -1.04088557e+00 3.91293526e-01
-1.98202759e-01 -2.14356971e+00 1.20673887e-02 4.87046063e-01
8.12945485e-01 -1.88341841e-01 -3.09699118e-01 -1.94463674e-02
9.20210361e-01 -1.12819266e+00 5.66578448e-01 6.99561656e-01
1.23142648e+00 -8.45934898e-02 5.98508239e-01 -1.89436585e-01
-6.23024940e-01 2.98109144e-01 6.36502504e-02 1.25019372e-01
-1.54424652e-01 1.13870692e+00 -8.13731432e-01 1.02242446e+00
5.29247940e-01 9.85667288e-01 -7.40626395e-01 7.18914986e-01
-4.17139024e-01 8.68317485e-01 6.55315071e-02 -1.05566181e-01
-1.97717562e-01 -3.95513438e-02 2.70591557e-01 1.56181061e+00
1.85238533e-02 5.29853940e-01 -2.61658460e-01 1.12078881e+00
-8.21091235e-01 -5.80161996e-02 -5.86608112e-01 -7.53410757e-01
7.60855258e-01 1.43660367e+00 -8.09693456e-01 -6.97807550e-01
-3.90099853e-01 -5.85524291e-02 5.74796081e-01 3.92063677e-01
-5.34331441e-01 -2.50615835e-01 4.29147810e-01 3.54423910e-01
-2.00226650e-01 -1.15175165e-01 -6.48642719e-01 -9.08613384e-01
-5.12541950e-01 -5.02443612e-01 7.98966408e-01 -9.52404141e-01
-1.48692048e+00 1.61951974e-01 1.71098396e-01 -3.98713768e-01
1.30573079e-01 -4.94887710e-01 -5.60771450e-02 9.30164933e-01
-9.64459300e-01 -1.25129330e+00 1.41085060e-02 2.28733942e-01
-2.71581620e-01 7.93395936e-02 9.09238100e-01 4.22255784e-01
-7.75514364e-01 4.76688325e-01 -3.90879780e-01 2.27202997e-01
9.91571724e-01 -1.37181199e+00 3.10270041e-01 1.37636764e-02
9.56307873e-02 9.77741659e-01 8.46006632e-01 -1.21257949e+00
-1.18662119e+00 -7.52552629e-01 1.26454401e+00 -1.13443553e+00
1.39594448e+00 -3.44340891e-01 -9.37351644e-01 7.87510335e-01
1.22013338e-01 -7.51347542e-02 9.88660812e-01 9.27079797e-01
-9.34188306e-01 2.91446418e-01 -1.05045784e+00 5.51954746e-01
1.65601790e+00 -5.53937018e-01 -5.70360065e-01 6.35215402e-01
8.57947290e-01 -3.39120716e-01 -1.70984268e+00 4.62475508e-01
3.67832512e-01 -1.75197527e-01 1.10184824e+00 -1.21846151e+00
8.84575427e-01 -6.79290071e-02 2.80519456e-01 -9.56360340e-01
-6.60265982e-01 -3.75808954e-01 -2.66358733e-01 1.50626707e+00
1.11032856e+00 -5.07466376e-01 5.45260966e-01 6.17432892e-01
-3.03200811e-01 -9.46780324e-01 -4.37075973e-01 -3.17067653e-01
1.58273712e-01 -5.99326603e-02 6.56047940e-01 1.67243361e+00
3.54680181e-01 6.01541102e-01 3.05485904e-01 -2.03255430e-01
8.74703407e-01 3.58296901e-01 1.96679249e-01 -1.71043611e+00
-1.17444523e-01 -6.18451476e-01 -4.52824980e-01 -9.46373940e-02
2.57004023e-01 -1.22080922e+00 -6.85774326e-01 -2.00166726e+00
8.65402579e-01 -8.21623683e-01 -2.10676432e-01 9.13889229e-01
-2.62328207e-01 1.75013840e-01 -4.93271589e-01 1.97436661e-01
-5.97100258e-01 -2.70137303e-02 1.32940578e+00 -2.55739599e-01
3.00131500e-01 -6.96781993e-01 -1.16898930e+00 4.87640917e-01
2.93138415e-01 -6.65273607e-01 -6.81198686e-02 -1.35103539e-01
8.49787951e-01 2.17907086e-01 5.61295629e-01 -2.18317941e-01
6.19751215e-01 -5.05041003e-01 3.50244761e-01 -2.65038282e-01
-8.98259282e-02 -1.73527494e-01 6.73455000e-01 6.98603094e-02
-1.01051366e+00 -2.23913372e-01 2.16598555e-01 6.22651637e-01
1.41543463e-01 1.42230438e-02 -2.43029054e-02 -6.36710152e-02
-5.94165325e-02 2.92485595e-01 1.15195252e-01 4.53529209e-01
6.52214706e-01 3.94956857e-01 -1.46022046e+00 9.48214233e-02
-8.62538397e-01 7.53709376e-02 3.55549693e-01 3.58030200e-01
-5.44485217e-03 -1.04054153e+00 -1.03256476e+00 -6.59583271e-01
2.07595035e-01 -5.43057680e-01 8.37793648e-02 1.08511519e+00
1.87193233e-04 3.77472609e-01 -3.47193405e-02 1.69898242e-01
-1.20368862e+00 8.74957621e-01 -2.33746231e-01 -3.60941619e-01
-3.40574056e-01 7.33713686e-01 1.70378700e-01 -1.05346464e-01
-1.15084440e-01 -5.48746735e-02 -5.26787043e-01 2.52802730e-01
8.03126022e-03 4.55793560e-01 -6.48281276e-02 -4.32216167e-01
-6.74955010e-01 3.98742080e-01 -8.14401507e-02 -1.75365694e-02
1.44242823e+00 2.01593518e-01 -8.64575148e-01 6.00120187e-01
1.00510585e+00 8.32325876e-01 -3.45905036e-01 -2.27414101e-01
3.68217051e-01 1.57323942e-01 6.70176372e-02 -1.28754413e+00
-7.52620876e-01 4.02580082e-01 -6.19469345e-01 3.31698626e-01
4.06615227e-01 6.32221639e-01 3.08338981e-02 -1.06530070e-01
1.85770273e-01 -7.40261912e-01 -4.28158760e-01 1.37199536e-01
8.55659068e-01 -7.28570819e-01 8.76552761e-01 -1.13302827e+00
-9.55025777e-02 7.08788514e-01 2.27881894e-01 3.65883917e-01
5.52755177e-01 5.72362602e-01 -6.83199346e-01 -1.19640434e+00
-8.12330246e-01 -3.52042429e-02 6.15178108e-01 1.34770140e-01
8.90516818e-01 2.29808867e-01 -8.68829131e-01 8.39643836e-01
-1.53263122e-01 -5.28010465e-02 5.83618879e-01 6.06398761e-01
1.10781461e-01 -1.11846817e+00 -7.39501119e-02 1.06299198e+00
-1.01061141e+00 -4.61403519e-01 -8.65345478e-01 7.21527100e-01
4.29473281e-01 8.31459403e-01 -2.18895003e-01 -2.88289040e-01
3.40490252e-01 -2.42691845e-01 5.25017321e-01 -7.36257493e-01
-6.83660030e-01 -3.08924735e-01 9.19928968e-01 -2.77535081e-01
-7.38557875e-01 -6.41313553e-01 -1.47930276e+00 -5.12021124e-01
-3.33826751e-01 4.87457365e-01 7.41844535e-01 9.69522536e-01
8.73426914e-01 1.05535710e+00 -1.11376651e-01 1.77691862e-01
2.17567965e-01 -8.81720722e-01 -4.78104681e-01 4.67344999e-01
-1.15117617e-01 -9.71595645e-01 -2.82812268e-01 1.22025758e-01] | [8.897939682006836, 8.455231666564941] |
97fe974f-798f-4e8b-89d2-4a5a1d1d3a08 | space-time-recurrent-memory-network | 2109.06474 | null | https://arxiv.org/abs/2109.06474v2 | https://arxiv.org/pdf/2109.06474v2.pdf | Space Time Recurrent Memory Network | Transformers have recently been popular for learning and inference in the spatial-temporal domain. However, their performance relies on storing and applying attention to the feature tensor of each frame in video. Hence, their space and time complexity increase linearly as the length of video grows, which could be very costly for long videos. We propose a novel visual memory network architecture for the learning and inference problem in the spatial-temporal domain. We maintain a fixed set of memory slots in our memory network and propose an algorithm based on Gumbel-Softmax to learn an adaptive strategy to update this memory. Finally, this architecture is benchmarked on the video object segmentation (VOS) and video prediction problems. We demonstrate that our memory architecture achieves state-of-the-art results, outperforming transformer-based methods on VOS and other recent methods on video prediction while maintaining constant memory capacity independent of the sequence length. | ['Fuxin Li', 'Chanho Kim', 'Hung Nguyen'] | 2021-09-14 | space-time-recurrent-memory-network-1 | https://openreview.net/forum?id=TYqb6EXphrr | https://openreview.net/pdf?id=TYqb6EXphrr | null | ['video-prediction'] | ['computer-vision'] | [ 1.37615442e-01 -2.16553867e-01 -6.75870180e-01 -3.98344934e-01
-6.40281260e-01 -2.24124715e-01 2.14632571e-01 -2.93245852e-01
-4.62537467e-01 3.59940112e-01 -1.18583813e-01 -4.28232491e-01
5.16611971e-02 -6.38213277e-01 -1.14567864e+00 -5.64925790e-01
-2.11337522e-01 4.10138458e-01 7.38322198e-01 5.52873075e-01
1.83634594e-01 1.32644162e-01 -1.52456737e+00 4.69885379e-01
3.96942049e-01 1.75443280e+00 6.24089777e-01 7.62378335e-01
-1.06453367e-01 1.30826497e+00 -1.35906592e-01 -2.64479876e-01
2.15437457e-01 -1.30624995e-01 -1.19196463e+00 2.27308705e-01
7.26377964e-01 -4.99063939e-01 -7.72323430e-01 8.17883551e-01
9.31589827e-02 3.91819477e-01 4.33642238e-01 -1.22161746e+00
-6.77853048e-01 6.20415866e-01 -4.44989294e-01 4.74897861e-01
-8.34049210e-02 -8.38744566e-02 1.09303963e+00 -8.66863608e-01
4.73095983e-01 1.15719056e+00 6.92246377e-01 5.16138554e-01
-9.28415895e-01 -5.85372150e-01 5.91663241e-01 8.83533955e-01
-1.29628789e+00 -4.72935200e-01 5.61137259e-01 -6.18875802e-01
1.31107044e+00 -1.37221040e-02 7.97073305e-01 8.79657865e-01
1.44364506e-01 1.30321586e+00 4.77255136e-01 -5.45912683e-02
1.60337672e-01 -1.83514908e-01 4.75221723e-02 1.05118358e+00
-4.59555864e-01 -7.28896558e-02 -9.40185606e-01 2.36538172e-01
9.67989445e-01 8.57922584e-02 -1.69034868e-01 -3.46255660e-01
-9.47902799e-01 7.84618556e-01 3.37031275e-01 -5.42467535e-02
-1.13461986e-01 6.73729956e-01 5.33730268e-01 3.48386198e-01
5.91409147e-01 -1.85832232e-01 -7.04275668e-01 -3.14709097e-01
-1.35914862e+00 -1.10714972e-01 6.14175200e-01 1.02868438e+00
4.89364326e-01 1.47789985e-01 -3.46044809e-01 7.87746549e-01
5.08461893e-01 4.27648902e-01 4.05891389e-01 -1.33688450e+00
5.59625864e-01 2.66318530e-01 1.22611364e-02 -9.02722955e-01
-2.94885665e-01 -3.09800297e-01 -8.71243834e-01 -8.38570967e-02
2.82695204e-01 9.00910869e-02 -1.07766223e+00 1.59611356e+00
1.28334045e-01 8.96187484e-01 -4.43469316e-01 9.44129884e-01
7.52511203e-01 9.17574048e-01 7.60345906e-02 -3.71803820e-01
1.14637697e+00 -1.42935336e+00 -5.57369828e-01 -5.00723422e-01
3.58526498e-01 -4.87331748e-01 9.18006837e-01 2.66622335e-01
-1.27883506e+00 -7.49245584e-01 -8.68327439e-01 -1.93594933e-01
7.42194951e-02 1.43624470e-01 4.62974072e-01 3.90060663e-01
-1.17562854e+00 7.76379943e-01 -1.32612944e+00 8.28607231e-02
6.53512597e-01 4.30680573e-01 7.18026906e-02 1.41618744e-01
-1.03754532e+00 6.74851477e-01 3.94979864e-01 2.62702912e-01
-1.27205050e+00 -6.45653725e-01 -7.98365533e-01 2.39206299e-01
5.17694473e-01 -6.98688030e-01 1.39776373e+00 -9.29189503e-01
-1.65259326e+00 7.16355860e-01 -5.46518385e-01 -8.60412955e-01
3.28014135e-01 -3.68403047e-01 -9.22999233e-02 2.27030680e-01
-1.61430702e-01 8.41979921e-01 1.25959218e+00 -6.59425437e-01
-8.15105319e-01 -1.52907357e-01 1.10820696e-01 -8.05909839e-03
-5.22828102e-01 -1.50882348e-01 -9.15607572e-01 -6.75059795e-01
4.49006408e-02 -8.86848927e-01 -1.33486286e-01 2.35405892e-01
-1.14997484e-01 -3.40601474e-01 9.39306617e-01 -8.40864122e-01
1.50555146e+00 -1.97056675e+00 4.34218615e-01 6.45477846e-02
2.37851515e-01 3.11785996e-01 -1.42412603e-01 -1.53386697e-01
2.38140509e-01 8.35980996e-02 1.06822811e-01 -4.44308966e-01
1.61014557e-01 3.88535768e-01 -5.04770160e-01 3.31976295e-01
-1.42682428e-02 8.89753819e-01 -7.11293995e-01 -9.08649027e-01
1.62616998e-01 2.86315203e-01 -7.85355568e-01 4.33869004e-01
-5.20243406e-01 2.66124487e-01 -2.67542571e-01 4.99223143e-01
3.64028841e-01 -9.48037863e-01 3.66461575e-01 -3.37774396e-01
3.61953117e-03 4.69935745e-01 -9.02384698e-01 1.67751145e+00
-4.15890366e-01 8.69152784e-01 -2.31511369e-01 -1.31835949e+00
5.05104303e-01 3.68161440e-01 5.40514588e-01 -7.54069269e-01
-4.47001271e-02 -9.53478441e-02 -3.97855908e-01 -6.98508143e-01
4.12637532e-01 8.76955390e-02 3.24624002e-01 3.75792742e-01
2.69959956e-01 4.70803946e-01 3.21911633e-01 -1.34254605e-01
9.13938820e-01 3.20118785e-01 -2.44514123e-01 1.06447970e-03
5.08059621e-01 -3.07954073e-01 6.67619407e-01 7.65530765e-01
-1.32438347e-01 1.93180546e-01 5.33236146e-01 -6.46980524e-01
-1.08922672e+00 -1.07955420e+00 5.74799962e-02 1.62672853e+00
2.23594308e-01 -5.56853890e-01 -7.87787080e-01 -4.97911692e-01
-1.95788816e-01 2.96989828e-01 -4.94070202e-01 -3.50540131e-02
-7.65560091e-01 -3.63094598e-01 3.42115998e-01 9.53772247e-01
6.02312028e-01 -9.51463580e-01 -8.35145295e-01 2.09375069e-01
-3.90823036e-01 -1.38330626e+00 -7.12084293e-01 -9.89917479e-03
-1.04037774e+00 -1.01823401e+00 -5.89416027e-01 -9.89813685e-01
3.20496023e-01 -6.99500879e-03 1.25272977e+00 -1.64743625e-02
-2.02434212e-02 5.14833510e-01 -8.73897001e-02 1.94067553e-01
-2.46908106e-02 2.00217605e-01 -2.50439227e-01 3.20463814e-02
1.05341367e-01 -4.65488821e-01 -7.61385083e-01 3.80117506e-01
-7.15196908e-01 3.55548114e-01 2.35950872e-01 9.13031399e-01
8.30583274e-01 -1.51471570e-01 1.31380066e-01 -5.64721465e-01
3.05960644e-02 -4.56954569e-01 -9.30447638e-01 5.42188525e-01
-4.60747898e-01 2.42926344e-01 2.87436455e-01 -6.22796893e-01
-8.76246512e-01 2.36871660e-01 -1.93062052e-02 -9.72917080e-01
3.93262416e-01 6.23285770e-01 1.78153381e-01 -1.73344016e-01
-8.79250392e-02 4.01754856e-01 -2.31916234e-01 -5.40777743e-01
9.63067934e-02 3.65994453e-01 3.48967105e-01 -4.03356642e-01
4.04452652e-01 2.93722868e-01 1.12816645e-03 -5.88958919e-01
-1.21213794e+00 -3.39888424e-01 -6.42836273e-01 -2.88153499e-01
9.64745700e-01 -9.55124140e-01 -1.08937955e+00 5.90363979e-01
-1.27697635e+00 -7.18449712e-01 3.91207263e-02 4.11760956e-01
-7.96794415e-01 4.37932432e-01 -1.12911201e+00 -6.55987024e-01
-2.81897098e-01 -1.23063588e+00 1.05178618e+00 7.53742363e-03
7.86279365e-02 -1.20734227e+00 -9.83460695e-02 5.62253296e-01
3.67638677e-01 -2.69520700e-01 9.21580434e-01 -2.35699601e-02
-1.24159670e+00 2.98304677e-01 -3.45337242e-01 3.84055912e-01
-2.78125733e-01 -2.08841816e-01 -7.19864964e-01 -3.51341665e-01
-9.72918123e-02 -3.43782157e-01 1.22442019e+00 6.45852387e-01
1.72845662e+00 -5.26907384e-01 -3.80315542e-01 8.57895076e-01
1.15344489e+00 1.63001671e-01 3.96731555e-01 1.93131372e-01
7.82988071e-01 1.32800028e-01 6.26384318e-01 5.43419540e-01
5.40465057e-01 7.85589218e-01 4.74436820e-01 1.76333115e-01
-4.73315790e-02 -2.08175495e-01 5.61994612e-01 8.85870934e-01
-1.53518796e-01 -3.65728796e-01 -8.45547497e-01 6.75831437e-01
-2.27351665e+00 -1.02760792e+00 3.38991255e-01 2.16776347e+00
8.75443757e-01 1.79024592e-01 1.27981737e-01 -1.93493038e-01
6.30636752e-01 3.98902416e-01 -7.75836110e-01 -2.21196502e-01
6.32783324e-02 6.45417720e-02 6.80086792e-01 4.80346799e-01
-1.44157255e+00 1.14467192e+00 7.44152164e+00 7.86636949e-01
-1.31854296e+00 3.58849108e-01 7.76762962e-01 -5.21868467e-01
1.89466979e-02 -3.15325201e-01 -8.40326607e-01 5.69927335e-01
1.20106852e+00 6.57171234e-02 6.70314372e-01 8.61186981e-01
-5.27683552e-03 5.69462553e-02 -1.27302420e+00 1.30932665e+00
2.40699528e-03 -1.70469618e+00 -7.16917962e-02 -2.21084848e-01
6.15948260e-01 4.64985281e-01 3.13714355e-01 1.95616916e-01
4.32778709e-03 -1.13896716e+00 9.07897115e-01 6.20685816e-01
7.10282743e-01 -4.49114472e-01 4.65036035e-01 2.33673364e-01
-1.42250896e+00 -2.72080570e-01 -4.43668336e-01 5.01343124e-02
3.76110852e-01 2.45144650e-01 -4.75379646e-01 -1.91061899e-01
1.06879723e+00 9.10538673e-01 -3.62542540e-01 1.06806362e+00
-1.27827842e-02 8.57506692e-01 -3.63232523e-01 2.28851020e-01
4.09572035e-01 8.01599324e-02 2.15415582e-01 1.24896729e+00
4.20556009e-01 -8.14848915e-02 5.35565078e-01 6.24666989e-01
-3.41920912e-01 -2.32315660e-01 -1.01994015e-01 -1.56206876e-01
3.95331949e-01 7.14534879e-01 -5.93179524e-01 -5.79085052e-01
-6.89904928e-01 9.10894096e-01 5.86884320e-01 4.66374815e-01
-1.25473809e+00 1.17411196e-01 6.17723584e-01 1.93046123e-01
1.09190786e+00 -2.88803488e-01 -5.62158972e-02 -1.01475680e+00
2.42006883e-01 -4.36839908e-01 5.04465699e-01 -6.18660629e-01
-1.04373300e+00 5.14781833e-01 -1.24864139e-01 -1.19020009e+00
-4.00584012e-01 -7.84765601e-01 -1.50907785e-01 3.30432981e-01
-1.47292650e+00 -1.11349022e+00 -2.06656549e-02 7.28927553e-01
6.37137115e-01 -1.77548021e-01 5.59623599e-01 5.96436143e-01
-5.69667101e-01 7.75698304e-01 1.67628035e-01 3.12271118e-01
2.89220840e-01 -9.24430072e-01 2.85594612e-01 7.61596918e-01
3.13647062e-01 2.46494293e-01 4.69321877e-01 -4.13722515e-01
-1.46760261e+00 -1.31032419e+00 7.76077509e-01 -2.13819921e-01
8.79557490e-01 -2.25844815e-01 -9.21858072e-01 1.14715672e+00
1.23108700e-01 4.34236377e-01 6.37342870e-01 1.35975242e-01
-5.36977708e-01 -2.19964549e-01 -6.07668817e-01 3.91290396e-01
1.29129934e+00 -8.62231195e-01 -3.57840687e-01 3.83164287e-01
7.99964011e-01 -7.39116549e-01 -1.11123323e+00 2.89406359e-01
6.85759366e-01 -7.78557897e-01 9.92697775e-01 -8.36617470e-01
5.06605208e-01 -2.52102762e-01 -1.64486066e-01 -7.32677042e-01
-4.88708913e-01 -5.53644359e-01 -8.21865022e-01 7.99487352e-01
3.72359514e-01 -1.51074573e-01 1.12328863e+00 5.38947582e-01
-1.12218365e-01 -8.70501459e-01 -1.22822678e+00 -7.42812157e-01
-1.62921786e-01 -6.94603324e-01 2.26331532e-01 4.30605143e-01
-9.07781795e-02 3.36341739e-01 -9.01044786e-01 1.52308390e-01
6.21376336e-01 5.27500212e-01 2.13814154e-01 -9.42019105e-01
-6.08120918e-01 -3.57006758e-01 -4.38663214e-01 -1.91136146e+00
4.77085501e-01 -7.59598792e-01 1.49472356e-01 -1.29159570e+00
4.72903252e-01 -2.52912670e-01 -6.06756389e-01 5.30984044e-01
1.01398796e-01 3.10050070e-01 4.33295667e-01 2.65766770e-01
-1.42692232e+00 5.42876840e-01 1.15210581e+00 -4.03112590e-01
1.52792901e-01 4.22734804e-02 5.47498316e-02 8.18316400e-01
4.92235482e-01 -3.83165985e-01 -4.80727434e-01 -9.62768912e-01
3.05518329e-01 3.56852829e-01 4.55547273e-01 -9.39208627e-01
5.23533821e-01 -2.63448000e-01 3.63678724e-01 -8.42332006e-01
6.57740235e-01 -7.55884469e-01 3.90163697e-02 3.81682724e-01
-5.05768478e-01 -1.69230793e-02 1.24152459e-01 6.52107120e-01
-3.82536739e-01 -6.46118373e-02 7.30967343e-01 -1.67461950e-02
-1.09281075e+00 8.17762911e-01 -5.60358286e-01 1.59477577e-01
8.55716765e-01 -2.17810869e-01 -1.28588527e-01 -3.78422350e-01
-9.69883680e-01 4.54004347e-01 1.86656713e-01 5.36924064e-01
8.39484274e-01 -1.38386953e+00 -3.31183463e-01 6.06691204e-02
-2.16287017e-01 -1.42429352e-01 4.92611110e-01 9.11792755e-01
-3.99429291e-01 6.74923241e-01 -2.08188996e-01 -8.75588953e-01
-1.38368583e+00 7.50693142e-01 4.09811616e-01 -3.71484578e-01
-4.15717155e-01 1.11297131e+00 1.67275935e-01 2.38518249e-02
6.62615418e-01 -4.59896207e-01 -1.73680529e-01 -9.96840447e-02
6.08806014e-01 4.25984085e-01 -2.75689125e-01 -6.10661685e-01
-2.93004423e-01 6.79107010e-01 6.62536547e-02 2.54091084e-01
1.28722668e+00 -3.29321116e-01 -1.40575290e-01 6.20669246e-01
1.26601648e+00 -6.57544732e-01 -1.77482820e+00 -6.27461195e-01
2.19747469e-01 -4.22209442e-01 2.93691725e-01 -2.95571625e-01
-1.55194378e+00 1.08855724e+00 7.98533082e-01 -1.96564812e-02
1.09011781e+00 1.72552362e-01 1.17056322e+00 5.16987085e-01
2.48882815e-01 -1.33635998e+00 7.01153502e-02 8.32314909e-01
5.46753883e-01 -1.25945985e+00 -2.18767717e-01 -3.11411560e-01
-3.68217766e-01 1.14703727e+00 5.72917879e-01 2.26480467e-03
9.28763688e-01 3.28673124e-01 -3.26422393e-01 4.49764729e-02
-1.19440031e+00 2.42733248e-02 5.93450010e-01 3.82699877e-01
5.53516567e-01 -1.30847085e-03 2.43271634e-01 4.02832031e-01
1.13035209e-01 2.99195677e-01 -7.35647753e-02 6.96287572e-01
-5.07501900e-01 -9.25477147e-01 -2.01673117e-02 5.05630255e-01
-5.28425455e-01 -2.33564153e-01 2.31004998e-01 2.75561631e-01
3.82639356e-02 6.78407550e-01 5.16430259e-01 -4.28786188e-01
-2.66550273e-01 7.92255253e-02 7.95568168e-01 -2.39881098e-01
-1.87913254e-01 1.40166059e-01 -3.19064558e-02 -9.13311541e-01
-6.32903576e-01 -4.90026623e-01 -1.16561353e+00 -4.25781846e-01
-1.11980081e-01 -6.61443770e-02 2.52784848e-01 1.17388427e+00
3.90967995e-01 4.49922204e-01 2.93695956e-01 -8.56033087e-01
-4.06717926e-01 -7.53838956e-01 -3.90594602e-01 2.08494157e-01
3.53614807e-01 -7.02519655e-01 2.35722959e-01 4.36744571e-01] | [9.041831016540527, 0.2798403203487396] |
711cf9f2-ce4b-43b7-808e-5fa978977c7c | educational-game-on-cryptocurrency-investment | 2301.10541 | null | https://arxiv.org/abs/2301.10541v3 | https://arxiv.org/pdf/2301.10541v3.pdf | Educational Game on Cryptocurrency Investment: Using Microeconomic Decision Making to Understand Macroeconomics Principles | Gamification is an effective strategy for motivating and engaging users, which is grounded in business, marketing, and management by designing games in nongame contexts. Gamifying education, which consists of the design and study of educational games, is an emerging trend. However, the existing classroom games for understanding macroeconomics have weak connections to the microfoundations of individual decision-making. We design an educational game on cryptocurrency investment for understanding macroeconomic concepts in microeconomic decisions. We contribute to the literature by designing game-based learning that engages students in understanding macroeconomics in incentivized individual investment decisions. Our game can be widely implemented in online, in-person, and hybrid classrooms. We also reflect on strategies for improving the user experience for future educational game implementations. | ['Luyao Zhang', 'Jiasheng Zhu'] | 2023-01-25 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [-8.53720129e-01 1.90883368e-01 -2.89475322e-01 1.80980518e-01
-1.68882355e-01 -4.67873335e-01 3.61180425e-01 1.73565850e-01
-3.15011531e-01 2.77163655e-01 2.71308422e-01 -1.29835427e+00
1.62583571e-02 -1.42982471e+00 -1.46457806e-01 -3.01633805e-01
2.06814170e-01 8.64116922e-02 -1.22357300e-02 -6.61896706e-01
6.85969055e-01 -8.67087245e-02 -1.66272676e+00 -3.02312583e-01
9.98460889e-01 2.48278961e-01 1.35549411e-01 6.84474409e-01
-2.48938978e-01 1.36866367e+00 -5.78628659e-01 -8.33387017e-01
1.98502377e-01 -8.44987392e-01 -3.93681854e-01 -4.67332453e-01
-3.88336241e-01 -7.13049710e-01 -4.54234984e-03 9.42000449e-01
6.41953409e-01 2.80281931e-01 3.33972842e-01 -1.24254394e+00
-7.74541736e-01 7.77219594e-01 -3.56530994e-01 3.97505939e-01
1.65197358e-01 3.74444306e-01 7.42520750e-01 -1.72668234e-01
2.60318398e-01 7.42006123e-01 3.83650541e-01 5.47202468e-01
-5.39340258e-01 -1.06491137e+00 1.51179701e-01 5.91877811e-02
-7.66237319e-01 1.32311761e-01 5.70332527e-01 -5.72965264e-01
8.11666191e-01 1.92056730e-01 1.71710289e+00 3.84574831e-01
2.58874863e-01 6.22378528e-01 1.47253060e+00 -6.01076245e-01
6.25552237e-01 3.07644099e-01 1.17715061e-01 3.74538928e-01
5.61318338e-01 4.90125090e-01 -4.04463589e-01 2.05182076e-01
1.28075063e+00 1.02978796e-01 3.32982779e-01 1.83311641e-01
-5.64759791e-01 1.44303083e+00 -1.20961003e-01 1.91209435e-01
-6.70966744e-01 5.43458462e-01 2.53136277e-01 4.05018628e-01
5.67158103e-01 6.87967002e-01 -1.97564855e-01 -1.20677984e+00
-6.20656729e-01 5.86564302e-01 1.23461151e+00 4.10368651e-01
3.59381557e-01 3.32859635e-01 2.99682859e-02 3.67705703e-01
8.52428257e-01 2.03745589e-01 3.69724751e-01 -1.13387454e+00
-3.41034420e-02 5.02314508e-01 2.39068821e-01 -7.38067031e-01
5.80192693e-02 -2.61625439e-01 -3.84495128e-03 7.77489841e-01
5.36551893e-01 -9.10258412e-01 -2.73594111e-01 1.47867000e+00
5.38230538e-01 3.88273895e-01 -1.12401053e-01 6.52463734e-01
1.18288684e+00 3.37032378e-01 7.87400663e-01 2.32845157e-01
1.62386620e+00 -6.31359637e-01 -6.76657200e-01 6.09965585e-02
7.34737635e-01 -7.44116545e-01 1.19858217e+00 1.73891112e-01
-1.33250403e+00 -1.79345980e-01 -7.75171876e-01 1.80288151e-01
-5.75507581e-01 -6.10225976e-01 1.20159721e+00 2.04530692e+00
-1.20599830e+00 4.64438677e-01 -8.70841086e-01 -1.07635446e-01
4.06195074e-01 3.24381649e-01 7.13486135e-01 3.36134911e-01
-1.36683381e+00 8.53178203e-01 -1.96105585e-01 -5.87004542e-01
-4.44629610e-01 -1.34233940e+00 -6.64872706e-01 5.57430446e-01
3.15187633e-01 -9.59385753e-01 1.66450942e+00 -5.49655497e-01
-2.15472531e+00 7.92707264e-01 7.50968218e-01 -6.64869994e-02
5.42917669e-01 1.98013157e-01 2.13212475e-01 -2.46480420e-01
1.88003257e-02 3.40372473e-01 -1.17527224e-01 -5.07864773e-01
-8.50382149e-01 -4.51854467e-02 6.79092526e-01 8.32912207e-01
3.13740335e-02 6.52044654e-01 4.63104039e-01 -5.26464224e-01
-3.61976951e-01 -1.62748158e-01 -6.63272619e-01 -7.91039228e-01
5.72979808e-01 -4.19899821e-01 2.59762406e-01 -5.98701119e-01
1.23699558e+00 -1.81291270e+00 -1.02742863e+00 1.46662351e-03
4.53440130e-01 -9.12350416e-02 3.75613958e-01 8.13034713e-01
2.45351195e-01 7.46659994e-01 1.10625076e+00 1.32960469e-01
8.62637639e-01 -3.72539580e-01 1.14962973e-01 -1.20418489e-01
-5.40315688e-01 9.99702811e-01 -1.18770897e+00 1.47878423e-01
4.85875309e-01 1.59480363e-01 -9.08563614e-01 2.13705659e-01
1.94833279e-01 1.34013742e-01 -5.17285466e-01 3.49186540e-01
6.94319069e-01 5.97106817e-04 2.66841918e-01 1.08110356e+00
-6.29181921e-01 9.99832451e-01 -1.31078339e+00 8.12179446e-01
-7.05578625e-01 5.89021504e-01 -1.25954330e-01 -7.65690327e-01
5.77673256e-01 3.49267602e-01 5.19946039e-01 -9.21523094e-01
2.68052369e-01 1.38688520e-01 4.59158927e-01 -3.70511472e-01
7.70292699e-01 -2.77274877e-01 -4.79905680e-02 1.46633637e+00
-3.29183459e-01 -7.52752185e-01 3.62623215e-01 2.59234458e-01
8.03997934e-01 2.01800942e-01 5.16166925e-01 -5.67189038e-01
-5.47191739e-01 -5.29262051e-02 2.15508476e-01 8.26166511e-01
-4.69214410e-01 -2.36287862e-01 7.56484807e-01 -2.80317992e-01
-6.51584089e-01 -6.31240189e-01 3.11759263e-01 1.63667679e+00
1.25090480e-02 -5.77757835e-01 -1.09481406e+00 -1.92116812e-01
-7.58768171e-02 9.49152172e-01 -3.46718043e-01 2.84652919e-01
-4.50387374e-02 -9.85303521e-01 -4.83010001e-02 4.52574521e-01
6.98915660e-01 -1.11522925e+00 -1.20187664e+00 5.23573101e-01
1.56560883e-01 -4.43635583e-01 -4.12858754e-01 -9.64749083e-02
-5.99167645e-01 -1.13108623e+00 -3.54578763e-01 -6.00052357e-01
1.46944880e-01 6.01458192e-01 1.25953639e+00 5.94083786e-01
3.39832231e-02 7.83184171e-01 -3.66551816e-01 -1.28442132e+00
-3.85777652e-01 -1.44905776e-01 -3.11818600e-01 -1.10331893e+00
9.37480390e-01 -8.11562479e-01 -9.91323292e-01 1.61587462e-01
-3.59223038e-01 5.08884490e-01 -3.78734916e-02 3.83664906e-01
-2.17847332e-01 1.69030383e-01 7.63304830e-01 -9.62450087e-01
1.25004554e+00 -7.23263800e-01 -8.22260678e-01 -3.47677886e-01
-9.13161635e-01 -7.03746796e-01 1.48338825e-01 -4.40964967e-01
-9.87426758e-01 -1.20416212e+00 -4.20125097e-01 6.62617266e-01
2.95329913e-02 8.60638678e-01 1.05809122e-01 -4.50737238e-01
2.63466477e-01 -2.63014734e-01 3.44194062e-02 1.23076424e-01
1.30395561e-01 3.65241498e-01 -2.86062568e-01 -8.42749357e-01
2.28910580e-01 -3.80994141e-01 -4.03546661e-01 -5.52564204e-01
2.38099605e-01 -1.32647201e-01 2.90494293e-01 -8.27283740e-01
5.27368367e-01 -1.16904819e+00 -1.61285961e+00 7.20455885e-01
-3.35684478e-01 -1.07372117e+00 -7.07188010e-01 8.49856913e-01
-5.06752849e-01 -3.01567614e-01 -8.90956819e-01 -1.11395288e+00
5.97120263e-02 -1.13923883e+00 1.35641202e-01 1.31655228e+00
-1.77271351e-01 -1.53024292e+00 3.86081040e-01 8.48379135e-01
8.73000920e-01 3.65230106e-02 6.12974644e-01 -4.87769753e-01
-8.53720784e-01 -3.13263610e-02 1.94142610e-01 -2.68026471e-01
-2.70793200e-01 4.09599543e-01 -5.26532769e-01 3.55077714e-01
6.98206946e-02 -2.41323873e-01 1.89370140e-02 8.16210330e-01
4.34503734e-01 -3.35204363e-01 1.60086438e-01 9.97672752e-02
1.48059368e+00 7.97532022e-01 8.25531781e-01 7.65295744e-01
-1.02169022e-01 6.96762383e-01 5.74337244e-01 6.28816962e-01
1.26050615e+00 8.93023014e-02 1.22196048e-01 -1.15462616e-01
4.08786178e-01 -5.44050634e-01 3.72831255e-01 9.19500828e-01
-7.53327310e-01 2.54757077e-01 -1.12502575e+00 6.24798059e-01
-1.63152206e+00 -1.18769908e+00 -9.80603397e-02 1.93537462e+00
7.05885291e-01 1.60480723e-01 6.96627200e-01 -1.92260638e-01
3.33378226e-01 -1.64178267e-01 4.05760296e-02 -9.89440620e-01
5.33966541e-01 1.10253668e+00 3.59030068e-01 6.18469834e-01
-4.62008238e-01 1.05001700e+00 6.81921291e+00 7.19134510e-01
-8.86797845e-01 3.60185415e-01 1.20486426e+00 -3.05246580e-02
-8.34650099e-01 2.18068406e-01 -7.27053940e-01 4.78184879e-01
1.06365490e+00 -1.15722442e+00 2.54363954e-01 1.04180145e+00
8.96688461e-01 -5.26260138e-01 -4.18052346e-01 3.76555145e-01
-7.80425608e-01 -1.79778814e+00 -8.50887597e-01 6.69758320e-01
1.17280090e+00 -4.54794437e-01 2.45052561e-01 8.76702011e-01
1.55306721e+00 -1.05046403e+00 6.00671589e-01 -1.71297163e-01
1.54460281e-01 -1.03424323e+00 7.70881832e-01 2.76824892e-01
-8.71769190e-01 1.87105998e-01 -9.88741219e-02 -1.34429359e+00
-2.63475031e-01 -4.76920493e-02 -5.70421159e-01 4.94610034e-02
6.78696573e-01 1.20041668e-01 2.92128175e-02 1.15582728e+00
-3.39727193e-01 9.21779633e-01 -9.89484936e-02 -7.76174247e-01
3.09505641e-01 -1.01306963e+00 -2.13470653e-01 6.07771516e-01
3.88942450e-01 8.99009943e-01 -6.79039489e-03 9.26987171e-01
3.12083930e-01 6.57460749e-01 -4.79466528e-01 -2.71369755e-01
5.78360736e-01 1.10242784e+00 -1.21863031e+00 7.10199624e-02
-5.45123279e-01 9.32226777e-02 -3.12131107e-01 3.42911631e-01
-6.35513723e-01 -4.22610492e-01 1.27402937e+00 2.61845976e-01
1.05784789e-01 -8.36269259e-02 -1.02586591e+00 -7.69147933e-01
-8.07805181e-01 -1.03732383e+00 7.25305453e-03 -4.53923762e-01
-8.65571141e-01 -4.80740190e-01 -1.25300527e-01 -5.65807879e-01
-5.91702819e-01 -3.02742094e-01 -1.58822608e+00 1.13843238e+00
-1.14050877e+00 -9.24418330e-01 -2.41898149e-01 -8.04520473e-02
3.06360275e-01 1.41066030e-01 5.92730403e-01 1.55807525e-01
-4.89364535e-01 4.11213189e-01 2.74486784e-02 -1.83454737e-01
7.87612349e-02 -1.31577122e+00 6.83345139e-01 7.61188507e-01
-2.25352004e-01 7.97858953e-01 6.97086692e-01 -5.09980083e-01
-9.41229582e-01 -1.90125704e-01 8.00675988e-01 -1.30427822e-01
8.44272614e-01 -2.97765374e-01 -2.90548116e-01 3.62555683e-01
7.35526800e-01 -9.53216076e-01 1.41325247e+00 4.15266395e-01
4.52638716e-01 3.35890770e-01 -1.19140375e+00 1.15924323e+00
8.88327360e-01 -5.24791360e-01 -8.31189156e-02 8.48756507e-02
5.26024103e-01 -6.76506698e-01 -8.68040621e-01 -5.61240554e-01
6.27212405e-01 -1.22907734e+00 9.18334782e-01 -7.35454500e-01
9.11990345e-01 4.31709737e-01 2.84349084e-01 -1.46159756e+00
-2.63436109e-01 -1.23385310e+00 3.23884696e-01 1.18362272e+00
1.00283340e-01 -9.36043680e-01 1.39584434e+00 1.39639461e+00
8.97233039e-02 -7.36469269e-01 -4.36051190e-01 -2.99291462e-01
8.38690698e-01 -6.14465475e-01 1.16008973e+00 1.13305426e+00
7.56433785e-01 -3.41991037e-01 9.55893844e-02 -4.10788268e-01
4.30626631e-01 -1.09911546e-01 1.04922056e+00 -9.28048491e-01
-5.29951274e-01 -1.16245639e+00 -3.88547808e-01 -8.26729655e-01
-2.87143260e-01 -1.41685948e-01 -4.74171579e-01 -1.38029587e+00
1.09339260e-01 -6.21193290e-01 1.54169604e-01 -1.62832126e-01
-6.13811612e-01 2.00088888e-01 4.74269420e-01 -4.39280331e-01
-1.67121202e-01 1.24897659e-01 1.45175195e+00 5.69940388e-01
-5.41487873e-01 3.38459313e-01 -1.60667288e+00 5.91818035e-01
1.00660634e+00 -1.36317447e-01 -5.96409917e-01 2.63159573e-01
4.78864670e-01 1.62428901e-01 -1.99305415e-02 -4.82033849e-01
3.89191359e-01 -1.23359919e+00 -2.27206618e-01 -9.95637700e-02
-4.93297935e-01 -3.59894931e-01 3.13157916e-01 6.44967437e-01
1.27214253e-01 2.21545786e-01 2.29497626e-01 -3.73517424e-01
1.19766772e-01 -4.82622266e-01 3.15874100e-01 -3.64356130e-01
-4.05082405e-01 8.65752250e-02 -1.17527628e+00 3.40996861e-01
1.30983794e+00 -6.78531885e-01 -4.54530805e-01 -1.10377526e+00
-3.93295586e-01 2.97848046e-01 5.46488285e-01 -1.83002636e-01
1.80437088e-01 -1.01308489e+00 -5.97709894e-01 5.44361360e-02
-6.28566444e-01 -3.05573761e-01 4.00233299e-01 4.91106838e-01
-1.14785111e+00 4.17504758e-01 -5.80371439e-01 3.44361484e-01
-1.05728459e+00 -1.72345668e-01 6.43292248e-01 -7.04621911e-01
-2.47375980e-01 8.95031214e-01 1.79540902e-01 -3.27569127e-01
-3.77567969e-02 -2.50342876e-01 -5.25336385e-01 1.79348692e-01
7.55400956e-01 7.05782771e-01 -2.81272590e-01 3.13923389e-01
2.33109131e-01 -1.35212108e-01 3.24818254e-01 -4.06676233e-01
1.38529825e+00 -3.05757195e-01 1.63478822e-01 1.61710769e-01
2.57388204e-01 2.84154356e-01 -1.19455564e+00 4.18842047e-01
-1.46069348e-01 -8.16193044e-01 2.41271153e-01 -6.76852822e-01
-1.04298699e+00 6.41911387e-01 3.41165006e-01 5.90593040e-01
9.19875622e-01 -5.30124903e-01 2.63200641e-01 -4.30052489e-01
5.37182570e-01 -1.30607224e+00 1.13419190e-01 3.20873708e-01
-8.61875899e-03 -1.03805459e+00 -1.86624527e-01 -4.60529685e-01
-6.42859221e-01 1.02452874e+00 7.81759739e-01 -2.67754376e-01
1.32903421e+00 5.31586349e-01 2.13754758e-01 -2.89150000e-01
-8.93076122e-01 -3.69664818e-01 -4.82392371e-01 1.08414757e+00
7.65582383e-01 6.85780048e-01 -9.98422325e-01 1.18121064e+00
-7.33835816e-01 4.53315496e-01 1.18789172e+00 1.00442922e+00
-4.08856928e-01 -1.23070204e+00 -4.40721303e-01 4.67560798e-01
-1.02345991e+00 -6.09265208e-01 -6.96483999e-02 1.05879128e+00
1.47742769e-02 1.11526096e+00 3.26144904e-01 -1.94336429e-01
-1.35595324e-02 -1.20353088e-01 4.22871888e-01 -7.29621232e-01
-1.53771973e+00 -7.96942040e-02 1.75931573e-01 -1.90880336e-02
-3.00034378e-02 -9.69116032e-01 -8.18745792e-01 -1.65676224e+00
-1.87140137e-01 4.01335984e-01 6.63378239e-01 7.79803514e-01
1.45443693e-01 5.83362937e-01 4.79987383e-01 -3.42062652e-01
-4.28161144e-01 -7.23439574e-01 -1.10163653e+00 -3.10168862e-01
-5.15059829e-01 -5.00625491e-01 -2.44448006e-01 -5.51055431e-01] | [10.026567459106445, 7.061843395233154] |
1f177952-f1f9-4b82-a940-8f4b79a9ca8a | campari-camera-aware-decomposed-generative | 2103.17269 | null | https://arxiv.org/abs/2103.17269v1 | https://arxiv.org/pdf/2103.17269v1.pdf | CAMPARI: Camera-Aware Decomposed Generative Neural Radiance Fields | Tremendous progress in deep generative models has led to photorealistic image synthesis. While achieving compelling results, most approaches operate in the two-dimensional image domain, ignoring the three-dimensional nature of our world. Several recent works therefore propose generative models which are 3D-aware, i.e., scenes are modeled in 3D and then rendered differentiably to the image plane. This leads to impressive 3D consistency, but incorporating such a bias comes at a price: the camera needs to be modeled as well. Current approaches assume fixed intrinsics and a predefined prior over camera pose ranges. As a result, parameter tuning is typically required for real-world data, and results degrade if the data distribution is not matched. Our key hypothesis is that learning a camera generator jointly with the image generator leads to a more principled approach to 3D-aware image synthesis. Further, we propose to decompose the scene into a background and foreground model, leading to more efficient and disentangled scene representations. While training from raw, unposed image collections, we learn a 3D- and camera-aware generative model which faithfully recovers not only the image but also the camera data distribution. At test time, our model generates images with explicit control over the camera as well as the shape and appearance of the scene. | ['Andreas Geiger', 'Michael Niemeyer'] | 2021-03-31 | null | null | null | null | ['3d-aware-image-synthesis'] | ['computer-vision'] | [ 3.46988201e-01 1.78994790e-01 9.80689302e-02 -2.16770500e-01
-5.46455920e-01 -1.00512302e+00 9.47954834e-01 -8.21891129e-01
1.67425182e-02 3.38825583e-01 2.59922177e-01 -4.54147309e-02
2.71800905e-01 -7.42376983e-01 -9.16237473e-01 -8.78529191e-01
5.69016516e-01 6.40011013e-01 -2.67947074e-02 1.59592390e-01
9.21653360e-02 7.51996458e-01 -1.48647416e+00 -5.83022237e-02
4.84222621e-01 4.93818223e-01 1.41410485e-01 1.17178428e+00
2.19855055e-01 6.27908766e-01 -5.62098265e-01 -4.60937768e-01
7.36335576e-01 -7.36057758e-01 -3.68033201e-01 9.56093609e-01
7.56926060e-01 -6.51289523e-01 -6.79711819e-01 9.48986411e-01
2.12475508e-01 -1.14554606e-01 6.53830886e-01 -1.30342066e+00
-7.85752654e-01 -2.02617213e-01 -6.82142317e-01 -4.01855707e-01
2.28347227e-01 3.99558604e-01 9.21038210e-01 -6.35663033e-01
7.05965459e-01 1.18576157e+00 3.59775245e-01 5.58371902e-01
-1.89769292e+00 -2.85736531e-01 6.32688403e-02 -4.30296749e-01
-1.23374295e+00 -5.69095612e-01 1.13160789e+00 -5.07719278e-01
4.88112003e-01 2.87142426e-01 8.22529852e-01 1.28107142e+00
2.26387516e-01 5.90575218e-01 1.06827641e+00 -4.37145859e-01
1.97603822e-01 2.30145022e-01 -4.78413790e-01 5.87234259e-01
3.87997240e-01 2.23140955e-01 -5.46625555e-01 -1.22200362e-02
1.30356300e+00 5.40577509e-02 -3.51913840e-01 -1.14870131e+00
-1.14667904e+00 6.20747566e-01 1.71936586e-01 -7.20448792e-02
-3.44950020e-01 2.24822044e-01 -3.04941803e-01 -1.32055044e-01
3.28970224e-01 6.39032185e-01 4.50036973e-02 2.18730316e-01
-9.88835633e-01 5.26733100e-01 9.00229335e-01 1.26060224e+00
9.66054261e-01 2.17866421e-01 1.45751953e-01 2.52606839e-01
3.76477689e-01 8.15204978e-01 7.46985897e-02 -1.50774157e+00
1.92210600e-01 4.62210059e-01 2.16056734e-01 -1.06463432e+00
1.74111482e-02 -4.20404434e-01 -7.18401015e-01 5.90999126e-01
4.82644409e-01 -2.66183913e-02 -1.13085341e+00 1.96818960e+00
3.07060927e-01 1.49301350e-01 5.48767224e-02 9.66871202e-01
2.85060018e-01 6.64367914e-01 -5.07675290e-01 1.07229784e-01
1.02692103e+00 -5.49543560e-01 -3.51052523e-01 -5.48284769e-01
2.80445963e-02 -9.09758627e-01 1.00932801e+00 3.63441736e-01
-1.41692495e+00 -2.75420099e-01 -1.09531820e+00 -3.13076079e-01
1.57585844e-01 -5.67723475e-02 3.70526344e-01 7.16498435e-01
-1.10469627e+00 6.81432784e-02 -8.86481524e-01 -2.88618386e-01
3.13112944e-01 7.25920871e-02 -4.70992237e-01 -2.75651962e-01
-4.75564986e-01 8.14781785e-01 9.43867788e-02 -6.96363598e-02
-1.09843922e+00 -7.01420963e-01 -9.05380845e-01 -1.65220760e-02
2.29113609e-01 -1.26610374e+00 1.19714296e+00 -9.93336439e-01
-1.58677900e+00 1.12598765e+00 -2.89227128e-01 -5.93086779e-02
7.00027525e-01 4.59117219e-02 1.41878724e-01 1.68817684e-01
-3.21140140e-02 6.59713507e-01 1.13266706e+00 -1.93271327e+00
-1.80889219e-01 -3.64816099e-01 3.81819218e-01 4.57892239e-01
4.94906455e-02 -4.40794230e-01 -7.16230690e-01 -4.45365101e-01
2.44971827e-01 -1.11246240e+00 -2.07944900e-01 2.40431026e-01
-5.31232178e-01 6.54945493e-01 8.36819291e-01 -5.64972937e-01
5.38079143e-01 -2.21153688e+00 4.38041806e-01 4.26738635e-02
3.95808965e-01 -1.05276860e-01 -1.66043982e-01 3.89349818e-01
-1.33557558e-01 5.84860146e-02 -2.11425245e-01 -7.74245262e-01
7.21076652e-02 3.55094075e-01 -4.77811605e-01 7.45196283e-01
3.04740280e-01 8.16252351e-01 -6.94598317e-01 -2.69356549e-01
5.03796220e-01 7.28682220e-01 -1.00826311e+00 4.38913941e-01
-3.69181484e-01 6.71138048e-01 -2.13574961e-01 2.76713878e-01
8.62486064e-01 -3.69210720e-01 2.49068290e-01 -3.21157932e-01
-2.38620117e-02 -1.04737766e-02 -1.16590559e+00 1.79437602e+00
-4.91148859e-01 6.28337681e-01 2.55620778e-01 -7.29756594e-01
7.05728590e-01 2.27013245e-01 5.01651645e-01 -5.26820183e-01
1.46338895e-01 -2.02867687e-02 -1.05154760e-01 -2.88419783e-01
4.94095832e-01 -5.33531487e-01 5.70638478e-02 5.31468689e-01
5.93240336e-02 -8.95147979e-01 -1.83118537e-01 4.62763071e-01
8.38049412e-01 5.74071646e-01 -7.57800648e-03 -1.36084780e-01
-9.74690989e-02 1.01664819e-01 4.63235497e-01 5.18005610e-01
3.32720608e-01 1.20743930e+00 5.92501342e-01 -2.68645942e-01
-1.60258114e+00 -1.40857923e+00 2.16817349e-01 2.97852010e-01
2.08271116e-01 -1.71725959e-01 -9.16414917e-01 -2.00536922e-01
-1.69808611e-01 8.70503664e-01 -6.52252674e-01 -2.97994852e-01
-5.95004261e-01 -6.41426861e-01 2.02509686e-01 1.23394750e-01
3.79413247e-01 -3.06856036e-01 -6.94061935e-01 -1.04602814e-01
-8.44426453e-03 -1.25972843e+00 -6.46255732e-01 -6.13212511e-02
-7.18525887e-01 -1.05070400e+00 -7.27565050e-01 -4.00040418e-01
9.98632014e-01 6.46575928e-01 1.17487776e+00 -3.58242273e-01
-3.75395656e-01 6.87367797e-01 9.05587301e-02 -2.15525836e-01
-5.76023936e-01 -4.92386609e-01 -1.35467753e-01 3.59698862e-01
-1.61990047e-01 -8.08207989e-01 -6.28188789e-01 2.51855314e-01
-1.04783142e+00 4.42734361e-01 5.31114578e-01 6.85453713e-01
5.73266089e-01 2.38192137e-02 -2.31931046e-01 -8.37535858e-01
6.43724278e-02 -1.33118391e-01 -9.12189722e-01 1.17539309e-01
-3.28709245e-01 1.67038769e-01 5.37848175e-01 -5.08322418e-01
-1.33380342e+00 4.40476656e-01 3.42356265e-01 -9.65038419e-01
-2.12356225e-01 -1.27636626e-01 -5.89909613e-01 4.28807884e-02
7.05396771e-01 3.06176931e-01 1.44805074e-01 -3.29081982e-01
6.51422203e-01 1.86774537e-01 6.51425123e-01 -6.52571023e-01
1.20854044e+00 8.50255311e-01 2.93064177e-01 -1.01899505e+00
-7.15704739e-01 -2.72640120e-02 -8.14463615e-01 -7.53353834e-02
9.41610277e-01 -1.02398503e+00 -3.99252415e-01 5.04315376e-01
-1.26166582e+00 -4.18919474e-01 -5.26314080e-01 4.76417601e-01
-8.59750688e-01 1.66119486e-01 -1.30732790e-01 -7.63460457e-01
3.27676266e-01 -1.21574521e+00 1.39721537e+00 1.01746628e-02
-2.20359355e-01 -9.35143232e-01 5.44053800e-02 2.69731224e-01
2.35225931e-01 3.96276414e-01 9.57446992e-01 1.55177757e-01
-1.22709990e+00 -2.95289427e-01 -1.72153965e-01 3.54435414e-01
1.11458577e-01 1.97123557e-01 -1.23446918e+00 -9.60941166e-02
3.23075086e-01 -7.29127526e-02 4.55361813e-01 4.35240149e-01
8.31541002e-01 -2.98693269e-01 -9.82799008e-02 1.02355921e+00
1.45031857e+00 -7.62082590e-03 6.77086830e-01 -1.38856634e-01
1.00880682e+00 6.73709512e-01 -1.07387468e-01 2.86404759e-01
4.59603608e-01 8.09873283e-01 6.38565063e-01 -2.81446397e-01
-4.93447840e-01 -6.89194977e-01 2.26150051e-01 4.68734562e-01
7.01408461e-02 -5.06223261e-01 -6.98289454e-01 3.18908751e-01
-1.51501346e+00 -1.05461264e+00 3.70030552e-02 2.44604874e+00
6.05002463e-01 -1.45633101e-01 9.60299447e-02 -1.49324328e-01
4.87043887e-01 2.43864983e-01 -8.02094877e-01 3.59505825e-02
-3.32931280e-01 -1.08418614e-02 6.22875094e-01 7.55614400e-01
-7.24994183e-01 6.77966774e-01 6.63410616e+00 3.42847794e-01
-1.36296129e+00 -2.18376040e-01 6.05405748e-01 -3.86037350e-01
-8.28946114e-01 4.33795035e-01 -5.27966440e-01 2.11970091e-01
4.45762992e-01 -3.10099781e-01 6.69233143e-01 4.68112975e-01
1.20974235e-01 -8.51053447e-02 -1.33364892e+00 1.20395172e+00
3.48879606e-01 -1.35945988e+00 2.30847225e-01 6.11940980e-01
7.89954722e-01 -2.25941211e-01 1.93452671e-01 -2.82293141e-01
5.26718915e-01 -8.19038093e-01 1.12456441e+00 6.53310478e-01
8.13644528e-01 -5.06869376e-01 -3.10916752e-02 6.31384790e-01
-6.48594141e-01 3.35842252e-01 -2.55536199e-01 3.63207236e-02
3.04516196e-01 5.30474067e-01 -7.06410885e-01 3.50900590e-01
4.05193716e-01 6.29733324e-01 -4.03603882e-01 7.60858357e-01
-3.30192804e-01 3.51371646e-01 -3.71330857e-01 4.13429618e-01
-4.38576117e-02 -6.06556058e-01 6.80999041e-01 6.90686107e-01
5.53498626e-01 2.35369533e-01 2.38331035e-02 1.44814742e+00
-6.45980090e-02 -5.08607268e-01 -9.70180392e-01 -1.47455931e-01
1.94737822e-01 1.16877449e+00 -6.94838405e-01 9.40126106e-02
-3.19272250e-01 1.16656661e+00 3.41895595e-02 7.16079950e-01
-7.58557141e-01 7.26753846e-02 7.04510450e-01 4.13432509e-01
2.29477301e-01 -5.89038610e-01 -3.34580392e-01 -1.52400017e+00
2.77063921e-02 -7.98289120e-01 -2.52222806e-01 -1.13295329e+00
-1.14346611e+00 2.50852108e-01 1.36730656e-01 -1.21211433e+00
-3.29045892e-01 -7.33206391e-01 -4.98367995e-01 1.05805779e+00
-1.10216784e+00 -1.35314846e+00 -2.43734181e-01 5.73366702e-01
3.34721833e-01 1.79799572e-01 7.16587603e-01 3.92698087e-02
-3.74210387e-01 2.37448737e-01 2.16499269e-02 -2.63402704e-02
5.91914654e-01 -1.22019899e+00 5.25519907e-01 1.05182207e+00
4.30729896e-01 6.86974168e-01 9.06858981e-01 -3.31386417e-01
-2.02466941e+00 -8.18578780e-01 4.69668835e-01 -8.78599823e-01
4.34030056e-01 -7.89608598e-01 -5.79357088e-01 8.16665769e-01
2.80134201e-01 9.83846188e-03 3.23620945e-01 -2.53894925e-01
-4.96746868e-01 -4.01644967e-02 -9.70890999e-01 8.27765822e-01
9.82814074e-01 -6.83989525e-01 -1.46072760e-01 2.00163364e-01
6.35657251e-01 -6.40311956e-01 -5.35428882e-01 -1.07274704e-01
5.47632158e-01 -1.19366074e+00 1.08251894e+00 -3.20632756e-01
5.18686950e-01 -4.19721067e-01 -4.80392963e-01 -1.34032309e+00
-2.34412193e-01 -7.55499721e-01 5.62414713e-02 1.08656478e+00
3.11768889e-01 -5.32311738e-01 9.36843455e-01 1.14910126e+00
-5.68901338e-02 -2.99639493e-01 -5.52542806e-01 -6.98936760e-01
1.07988566e-01 -3.21674854e-01 5.49649298e-01 8.24587822e-01
-6.21868730e-01 5.46900332e-01 -5.79953849e-01 4.89407986e-01
9.17054832e-01 3.95175874e-01 1.29212689e+00 -9.50149655e-01
-6.53650403e-01 -3.88487488e-01 -4.13951516e-01 -1.29721141e+00
6.15861639e-02 -5.85002661e-01 7.02550188e-02 -1.30980563e+00
3.58325005e-01 -3.47354650e-01 6.45841897e-01 -4.19981852e-02
9.30384398e-02 2.14292899e-01 3.18538427e-01 3.25429916e-01
1.15058571e-02 5.11888444e-01 1.48729336e+00 1.33840322e-01
1.25154108e-02 -1.85368210e-02 -8.22582841e-01 7.66140997e-01
5.25577664e-01 -1.64445356e-01 -7.93309510e-01 -8.99524629e-01
2.56545275e-01 2.34679952e-02 8.98879528e-01 -6.11882210e-01
1.47320375e-01 -3.83242011e-01 6.89747572e-01 -3.49518627e-01
7.98061371e-01 -8.90245438e-01 6.71147883e-01 -3.53500359e-02
-1.27180189e-01 -1.41033694e-01 -3.38990912e-02 5.87482452e-01
-2.51678880e-02 -7.10812509e-02 8.99832606e-01 -3.30746502e-01
-2.31786877e-01 4.51979816e-01 -2.48344447e-02 9.04182643e-02
9.48417962e-01 -4.32402998e-01 -2.88695186e-01 -6.74372554e-01
-4.54906940e-01 -2.82699019e-01 1.23096633e+00 1.94633588e-01
4.40564454e-01 -1.25641119e+00 -5.60299397e-01 6.14337444e-01
-5.49837062e-03 5.20780683e-01 1.70131058e-01 2.96881288e-01
-6.16356909e-01 1.21338725e-01 1.38183897e-02 -9.00760710e-01
-9.60380912e-01 6.18206024e-01 5.73256314e-01 3.38688016e-01
-6.86329365e-01 6.67121053e-01 8.91219676e-01 -4.94697005e-01
-4.02055448e-03 -3.65897529e-02 6.31869555e-01 -3.13358814e-01
2.36557007e-01 6.17797002e-02 -3.15714449e-01 -8.02327693e-01
1.92188621e-02 8.60310197e-01 2.43328899e-01 -5.24410248e-01
1.15825903e+00 -2.02827111e-01 3.56696360e-02 3.07699919e-01
1.30555153e+00 2.99977034e-01 -1.90069020e+00 -1.72627166e-01
-6.17778242e-01 -9.01496291e-01 2.17946544e-01 -4.22529697e-01
-1.15777051e+00 1.01688790e+00 1.04240276e-01 1.43884867e-01
1.04642653e+00 1.35759026e-01 3.82870287e-01 6.70094490e-02
2.92928636e-01 -5.68404913e-01 2.18576714e-01 1.51232108e-01
8.37648094e-01 -1.04603469e+00 2.48160705e-01 -3.55595887e-01
-6.12170577e-01 9.44243014e-01 4.52389926e-01 -3.54049981e-01
4.84965235e-01 3.73753309e-01 -9.59695503e-03 -2.03624472e-01
-4.79455948e-01 8.53760988e-02 2.89032102e-01 7.03782856e-01
1.96684152e-01 -6.26265397e-03 7.31720269e-01 -1.13162749e-01
-3.34220201e-01 -2.73433626e-01 6.97039068e-01 7.14473128e-01
-1.61126912e-01 -1.18340087e+00 -5.17734706e-01 -6.67350143e-02
9.79450941e-02 1.27539247e-01 -5.73778510e-01 8.75150740e-01
8.36691912e-03 6.37655675e-01 1.41830578e-01 -9.98451337e-02
3.82011205e-01 -1.61531195e-01 9.89368081e-01 -7.80841887e-01
2.06864849e-01 5.12617290e-01 -3.00720930e-01 -6.54847682e-01
-4.33878601e-01 -7.98034310e-01 -8.05472493e-01 -4.11003441e-01
-1.63278788e-01 -2.42698267e-01 6.81667566e-01 5.82739532e-01
4.82685506e-01 1.73960194e-01 7.61098802e-01 -1.38351035e+00
-4.21438366e-01 -3.47784370e-01 -6.21091068e-01 3.69280845e-01
6.59555376e-01 -4.93948400e-01 -4.86367553e-01 4.66952920e-01] | [9.26497745513916, -3.1114513874053955] |
612f3230-12c3-42c3-881c-753587dc90c5 | calibrated-feature-decomposition-for | 2111.13945 | null | https://arxiv.org/abs/2111.13945v1 | https://arxiv.org/pdf/2111.13945v1.pdf | Calibrated Feature Decomposition for Generalizable Person Re-Identification | Existing disentangled-based methods for generalizable person re-identification aim at directly disentangling person representations into domain-relevant interference and identity-relevant feature. However, they ignore that some crucial characteristics are stubbornly entwined in both the domain-relevant interference and identity-relevant feature, which are intractable to decompose in an unsupervised manner. In this paper, we propose a simple yet effective Calibrated Feature Decomposition (CFD) module that focuses on improving the generalization capacity for person re-identification through a more judicious feature decomposition and reinforcement strategy. Specifically, a calibrated-and-standardized Batch normalization (CSBN) is designed to learn calibrated person representation by jointly exploring intra-domain calibration and inter-domain standardization of multi-source domain features. CSBN restricts instance-level inconsistency of feature distribution for each domain and captures intrinsic domain-level specific statistics. The calibrated person representation is subtly decomposed into the identity-relevant feature, domain feature, and the remaining entangled one. For enhancing the generalization ability and ensuring high discrimination of the identity-relevant feature, a calibrated instance normalization (CIN) is introduced to enforce discriminative id-relevant information, and filter out id-irrelevant information, and meanwhile the rich complementary clues from the remaining entangled feature are further employed to strengthen it. Extensive experiments demonstrate the strong generalization capability of our framework. Our models empowered by CFD modules significantly outperform the state-of-the-art domain generalization approaches on multiple widely-used benchmarks. Code will be made public: https://github.com/zkcys001/CFD. | ['Zheng-Jun Zha', 'Liang Li', 'Wei Wu', 'Jiawei Liu', 'Kecheng Zheng'] | 2021-11-27 | null | null | null | null | ['generalizable-person-re-identification'] | ['computer-vision'] | [-1.21324062e-02 -3.36439312e-01 -1.71488568e-01 -5.28453887e-01
-6.12947762e-01 -7.05634654e-01 6.64036632e-01 -1.54123427e-02
-4.56103027e-01 7.08022952e-01 4.70085740e-01 2.75728881e-01
-6.44760370e-01 -7.15004146e-01 -2.62876958e-01 -9.50103104e-01
1.65567487e-01 4.45990980e-01 -3.12525809e-01 -3.70438427e-01
-1.95946962e-01 1.44658342e-01 -1.36525953e+00 -7.00148642e-02
1.09883142e+00 6.66761160e-01 -2.17418283e-01 -1.25101153e-02
2.33720109e-01 1.51263654e-01 -5.29805362e-01 -5.92228115e-01
5.01857579e-01 -5.59317112e-01 -5.02626479e-01 5.73507585e-02
3.32099885e-01 -2.98294604e-01 -5.68898618e-01 1.34337902e+00
6.29399836e-01 2.95061618e-01 5.51067531e-01 -1.29767466e+00
-1.01205838e+00 4.78186727e-01 -5.31079829e-01 3.05904537e-01
5.54898024e-01 3.91564488e-01 8.33348095e-01 -7.64629960e-01
2.81224251e-01 1.26590645e+00 3.79362375e-01 7.68062949e-01
-1.49362671e+00 -1.32107460e+00 4.86629009e-01 6.11717477e-02
-1.57694197e+00 -3.89280885e-01 1.18316209e+00 -4.83711988e-01
4.06283945e-01 4.92062896e-01 5.74213922e-01 1.60491562e+00
-3.74658316e-01 5.06097078e-01 1.18349087e+00 -6.68461397e-02
-1.13778003e-01 3.05131793e-01 5.29404998e-01 4.71557796e-01
7.38017917e-01 3.76878232e-01 -5.54600775e-01 -1.89233378e-01
9.11433637e-01 3.32481802e-01 -4.01583940e-01 -4.26233023e-01
-1.44470274e+00 5.53344429e-01 4.50731605e-01 2.08766967e-01
-2.45311752e-01 -4.80703086e-01 4.03726935e-01 3.21353793e-01
2.83581793e-01 4.93010908e-01 -4.42157537e-01 1.42775022e-03
-6.07294023e-01 5.48606515e-01 4.44030851e-01 9.11807239e-01
9.41293120e-01 -2.60675639e-01 -5.86560667e-01 7.92100310e-01
5.74839450e-02 5.78118205e-01 7.13202477e-01 -6.45890355e-01
7.47264802e-01 1.11660492e+00 1.35693237e-01 -1.09230065e+00
-3.92519563e-01 -8.57726872e-01 -1.26168334e+00 -1.73453540e-01
6.18900955e-01 -9.26589817e-02 -5.70244610e-01 2.24399114e+00
3.73099148e-01 3.63197356e-01 1.02504149e-01 1.17403173e+00
9.23962831e-01 1.47813335e-02 1.40111640e-01 -2.10310463e-02
1.69670582e+00 -5.90328872e-01 -4.18324232e-01 -2.72368103e-01
1.12005666e-01 -2.85129607e-01 9.66025770e-01 1.08347408e-01
-7.75103986e-01 -7.43234158e-01 -1.10258996e+00 -3.30138505e-02
-2.57609606e-01 1.48137793e-01 7.44583368e-01 7.32326210e-01
-2.38545090e-01 4.24728453e-01 -2.95654386e-01 -1.51413441e-01
5.54171741e-01 6.62609518e-01 -9.19081330e-01 -2.29285076e-01
-1.56594992e+00 4.36990142e-01 4.17305022e-01 1.07169300e-01
-5.93209267e-01 -9.09984469e-01 -7.84966469e-01 1.34734496e-01
3.95762652e-01 -1.07791400e+00 6.13259077e-01 -7.42965937e-01
-1.33971119e+00 9.88372684e-01 -2.41927169e-02 3.87420952e-02
5.83664358e-01 1.66751258e-02 -7.57808626e-01 -9.39856768e-02
3.18313539e-01 1.96633026e-01 8.59903097e-01 -1.17483056e+00
-4.71982688e-01 -7.60313153e-01 2.64859200e-01 3.92886490e-01
-6.97883308e-01 -7.79725388e-02 -2.97180295e-01 -8.31597149e-01
1.95802420e-01 -6.17493510e-01 3.26206349e-02 -3.78849238e-01
-5.94731092e-01 -1.99675351e-01 1.18417077e-01 -7.61207819e-01
1.15977633e+00 -2.24018002e+00 3.88985753e-01 3.46602082e-01
6.48672700e-01 1.51805386e-01 -2.73484409e-01 2.03504428e-01
-4.51053321e-01 -7.70331547e-03 -7.46228695e-02 -4.69409466e-01
1.10812902e-01 -3.35752368e-02 -2.43805259e-01 6.08610272e-01
1.24575675e-01 9.12481368e-01 -1.04608488e+00 -1.72817409e-01
7.68951699e-02 3.60574633e-01 -5.53372800e-01 1.26244798e-01
4.20388967e-01 9.78971839e-01 -6.45978868e-01 6.37959361e-01
9.89125192e-01 -2.76905019e-02 1.72983348e-01 -3.70191097e-01
1.92877024e-01 1.56833038e-01 -1.32363451e+00 1.81225097e+00
-2.23552110e-03 -1.19547121e-01 -8.62622634e-02 -1.06412745e+00
1.09700048e+00 5.50739616e-02 4.90818769e-01 -7.61045933e-01
1.04441218e-01 -3.83017473e-02 1.89292058e-02 -2.77281731e-01
2.42061600e-01 -5.97786829e-02 -4.07541156e-01 3.52051526e-01
1.97232902e-01 6.91092253e-01 -1.70451283e-01 1.86718225e-01
6.99139655e-01 -8.47550575e-03 4.36430186e-01 -4.37859833e-01
1.00336814e+00 -3.35586369e-01 1.02047038e+00 5.91000140e-01
-5.32724023e-01 5.18942714e-01 4.50214744e-01 -3.29028875e-01
-8.02901685e-01 -1.17137051e+00 -1.07288219e-01 1.05523705e+00
6.96930707e-01 -4.13093060e-01 -5.81107557e-01 -8.12170506e-01
3.68623734e-01 3.68557513e-01 -8.85881364e-01 -5.76637745e-01
-3.66419941e-01 -9.19314682e-01 6.83085501e-01 5.92350781e-01
7.90760577e-01 -4.85404611e-01 2.68693954e-01 -1.63361683e-01
-3.50338519e-01 -8.15401375e-01 -7.90409863e-01 -1.96460798e-01
-5.37966907e-01 -1.09067070e+00 -7.89856553e-01 -4.47307557e-01
8.61589015e-01 4.62095648e-01 7.62667954e-01 -2.74474005e-04
-6.68270811e-02 3.56439948e-01 -1.86905116e-01 -4.38035727e-02
1.54969469e-01 1.04940124e-03 5.71779549e-01 4.49445307e-01
8.35570157e-01 -8.62999856e-01 -7.83232331e-01 5.16487002e-01
-5.60208917e-01 -9.68453959e-02 4.95036125e-01 1.21826661e+00
2.78995365e-01 1.15907744e-01 6.38828814e-01 -7.00288296e-01
8.08588028e-01 -6.21749759e-01 -2.04677358e-01 2.30966777e-01
-6.03080034e-01 1.01298608e-01 5.83883226e-01 -8.07658553e-01
-1.13471746e+00 -6.03757314e-02 2.52242893e-01 -2.73351997e-01
-3.16124976e-01 1.86452702e-01 -8.78284752e-01 -1.19816996e-02
7.17889309e-01 4.83217716e-01 -6.88445792e-02 -6.13295197e-01
2.85744578e-01 5.27937472e-01 8.53111863e-01 -9.81304109e-01
1.27347612e+00 4.56363559e-01 -2.70954877e-01 -1.29779950e-01
-7.86543012e-01 -6.26820683e-01 -7.89795995e-01 1.38921589e-01
4.66567576e-01 -1.29281664e+00 -1.02116752e+00 5.12872219e-01
-8.67571592e-01 2.25968078e-01 -4.01647747e-01 5.64529240e-01
-5.60323298e-02 5.62248349e-01 -2.98810422e-01 -5.82332671e-01
-2.29826316e-01 -8.99722457e-01 9.34512675e-01 6.22437537e-01
-1.37117505e-01 -7.29893267e-01 1.42764822e-01 5.48335969e-01
1.48019731e-01 2.15114698e-01 6.39782369e-01 -1.02525806e+00
-2.97099173e-01 -4.53916699e-01 -5.11099756e-01 3.09737384e-01
3.31120908e-01 -7.66521156e-01 -1.09352458e+00 -5.14246106e-01
-1.40666217e-01 -1.12056926e-01 8.86788070e-01 -1.06723502e-01
1.01837099e+00 -3.21823448e-01 -3.52895111e-01 9.79575932e-01
1.08134830e+00 -2.45620012e-01 3.48158181e-01 4.00105894e-01
9.95939910e-01 5.41176081e-01 2.57298708e-01 5.81935048e-01
7.56974816e-01 6.85188949e-01 1.29407807e-03 8.11321586e-02
-1.80853587e-02 -4.72972274e-01 2.13005915e-01 3.93777043e-01
-6.49208248e-01 2.26933822e-01 -4.26240206e-01 2.33640984e-01
-1.86745405e+00 -1.18294442e+00 1.53936252e-01 2.45962667e+00
8.68728757e-01 -1.37082636e-01 4.91531938e-01 1.42541498e-01
7.85627604e-01 3.39787155e-02 -7.85040617e-01 4.53718245e-01
-4.40429866e-01 -1.38387326e-02 3.16044986e-01 3.19604129e-01
-1.19997931e+00 6.75617516e-01 4.52967882e+00 8.62915039e-01
-6.70302987e-01 2.29242012e-01 2.47582927e-01 -1.57404333e-01
-4.42208081e-01 -6.88380077e-02 -9.21722054e-01 6.63582802e-01
2.30039075e-01 -5.62417567e-01 7.29542077e-01 7.10893273e-01
-5.72675318e-02 3.72124940e-01 -1.26950002e+00 1.41125262e+00
1.65764317e-02 -8.63108039e-01 9.20401812e-02 3.05983961e-01
5.64939618e-01 -7.07002699e-01 2.97579765e-01 5.61745644e-01
1.10297374e-01 -9.06535804e-01 6.23755038e-01 7.26095676e-01
9.42047596e-01 -8.45428109e-01 6.98507369e-01 2.56139129e-01
-1.29564345e+00 -3.76689821e-01 -4.71564353e-01 -2.57010639e-01
-3.26526165e-01 5.81291020e-01 -1.35665134e-01 1.02842867e+00
6.10894918e-01 8.80641580e-01 -7.49860168e-01 6.75731659e-01
-2.37942874e-01 2.06580311e-01 2.31767446e-02 4.57403988e-01
-3.19407195e-01 -3.61184865e-01 7.26762295e-01 1.06827188e+00
5.37074320e-02 3.30263376e-01 3.31714779e-01 1.25436628e+00
-4.54063416e-02 -8.39843750e-02 -2.18103886e-01 2.49101132e-01
6.32543385e-01 1.32039654e+00 -1.57376543e-01 -2.76905239e-01
-3.11066300e-01 1.32507873e+00 3.62857521e-01 7.40676820e-01
-7.61605859e-01 -2.03290567e-01 1.25041950e+00 -1.71467468e-01
-8.80815014e-02 2.20614276e-03 -3.58132660e-01 -1.80647159e+00
2.76764482e-01 -1.05321920e+00 6.31850064e-01 -1.01206139e-01
-2.00295639e+00 3.79597038e-01 6.03367165e-02 -1.49854124e+00
-4.22824845e-02 -4.08870518e-01 -5.32428205e-01 1.48054373e+00
-1.43970275e+00 -1.56961918e+00 -6.64919257e-01 1.18306065e+00
-2.50162389e-02 -5.65954447e-01 9.69512105e-01 4.98885125e-01
-1.18592072e+00 1.28034520e+00 -5.84233142e-02 4.67311591e-01
1.04032755e+00 -1.04743266e+00 1.20329112e-01 9.02798951e-01
-2.06406638e-01 1.38678932e+00 4.11083132e-01 -6.60380721e-01
-1.34401906e+00 -9.41276371e-01 6.63669944e-01 -5.97464919e-01
4.48346108e-01 -7.00969100e-01 -9.52998936e-01 6.31597221e-01
-3.75185907e-01 1.48255423e-01 9.12600696e-01 4.31430817e-01
-1.02119863e+00 -4.23128694e-01 -1.19649863e+00 5.20331144e-01
1.48702085e+00 -8.87410760e-01 -7.30134130e-01 1.21368162e-01
6.33602560e-01 -2.45085940e-01 -9.53650713e-01 2.82733917e-01
7.66681015e-01 -8.57307792e-01 1.46352887e+00 -8.79469514e-01
9.84178111e-02 -3.79976928e-01 -9.01780203e-02 -1.17279756e+00
-9.69466567e-01 -6.97807074e-01 -1.61382556e-01 1.65502965e+00
-8.80255476e-02 -1.10349023e+00 6.46183252e-01 8.82184684e-01
2.74271637e-01 -3.66439879e-01 -8.80332887e-01 -1.01079106e+00
1.41618460e-01 -2.31258404e-02 1.20957947e+00 1.37792814e+00
2.43469760e-01 2.80938715e-01 -5.90790629e-01 4.24123019e-01
9.59971964e-01 2.41929367e-01 8.27700198e-01 -1.51583648e+00
-3.86995554e-01 -6.08267725e-01 -5.46969235e-01 -8.77327681e-01
2.26354226e-01 -1.16452503e+00 -5.64730704e-01 -8.28200221e-01
6.98879421e-01 -5.47581553e-01 -7.99816072e-01 4.91603971e-01
-6.56752169e-01 5.27405068e-02 2.31071427e-01 3.41968298e-01
-3.85190487e-01 6.60206616e-01 1.31494284e+00 -3.12666565e-01
-2.83395559e-01 -4.69643436e-02 -1.47671652e+00 4.23106164e-01
7.01007664e-01 -3.46668214e-01 -3.97639006e-01 -1.66122824e-01
-1.24680139e-01 -3.58142048e-01 9.67215180e-01 -9.75672424e-01
3.86002481e-01 -2.34651849e-01 8.37848961e-01 -4.02658358e-02
3.60800445e-01 -5.89534283e-01 8.89415890e-02 2.12650001e-01
-2.44597867e-01 -3.32574546e-01 -5.70405573e-02 7.15920389e-01
-4.82616648e-02 2.72055745e-01 4.70697314e-01 -4.63348962e-02
-7.08252192e-01 6.20730281e-01 4.81654942e-01 5.28682582e-02
8.82338583e-01 -2.31156304e-01 -6.10095739e-01 -1.19550951e-01
-5.96304238e-01 1.93409950e-01 4.40647602e-01 5.80477953e-01
5.05189300e-01 -1.56164062e+00 -9.46971953e-01 6.87367499e-01
4.31778282e-01 -3.08037043e-01 7.12190688e-01 6.94390535e-01
3.42542857e-01 2.31837377e-01 -4.85994816e-01 -2.58504570e-01
-8.94850492e-01 7.97994852e-01 4.29253787e-01 -3.45803440e-01
-5.08729756e-01 9.88600612e-01 6.62834764e-01 -6.67324245e-01
1.79083887e-02 7.53376558e-02 -2.89352655e-01 1.03054591e-01
7.93449342e-01 4.86247420e-01 -3.46847922e-01 -8.43637228e-01
-7.39340365e-01 6.73438430e-01 -5.37736416e-02 7.13828355e-02
1.13684952e+00 -2.99541593e-01 -1.18635252e-01 6.61724880e-02
9.52098727e-01 2.04782739e-01 -1.24238300e+00 -5.68766296e-01
-3.86366665e-01 -6.12365603e-01 -2.81900764e-01 -8.90920579e-01
-9.51520145e-01 6.47378385e-01 5.99631786e-01 -1.53749958e-01
1.28942990e+00 -1.10696219e-01 5.74070692e-01 6.71649799e-02
3.58978629e-01 -7.76793480e-01 -3.43564115e-02 2.31228814e-01
8.14029276e-01 -1.32209504e+00 6.74629360e-02 -5.43574750e-01
-6.04257345e-01 8.71018648e-01 7.73882926e-01 -2.11339593e-01
3.76620203e-01 -2.15724617e-01 -3.06691736e-01 -3.67057174e-02
-5.74461073e-02 -2.61665016e-01 7.57139683e-01 9.68667746e-01
1.12354988e-02 4.03405964e-01 -1.72418401e-01 1.77868271e+00
-2.40844876e-01 -1.82081535e-01 -1.40390411e-01 2.02353761e-01
1.11344241e-01 -1.21128738e+00 -5.32524705e-01 2.06073910e-01
-3.80439833e-02 -1.25560045e-01 -3.58084351e-01 7.56391585e-01
6.79462552e-01 8.39655697e-01 -1.20729581e-01 -8.49673271e-01
5.17214954e-01 -4.37943302e-02 5.16966879e-01 -4.23051536e-01
-6.50530159e-01 -2.36073643e-01 -8.48936066e-02 -6.32446170e-01
-2.58986324e-01 -7.54691720e-01 -9.91109014e-01 -5.05909503e-01
-2.90544750e-03 -5.79723753e-02 3.16443518e-02 9.51193273e-01
4.84999716e-01 4.31238383e-01 7.09664583e-01 -7.74683714e-01
-8.06208313e-01 -8.44211280e-01 -8.01724732e-01 9.39633787e-01
4.34616715e-01 -9.74072993e-01 -3.22156638e-01 -1.23620860e-01] | [14.725126266479492, 1.0242007970809937] |
bd1444cf-86bc-4df3-9ead-3f30c0d18bc6 | abstractive-sentence-summarization-with-1 | 2108.05123 | null | https://arxiv.org/abs/2108.05123v3 | https://arxiv.org/pdf/2108.05123v3.pdf | ICAF: Iterative Contrastive Alignment Framework for Multimodal Abstractive Summarization | Integrating multimodal knowledge for abstractive summarization task is a work-in-progress research area, with present techniques inheriting fusion-then-generation paradigm. Due to semantic gaps between computer vision and natural language processing, current methods often treat multiple data points as separate objects and rely on attention mechanisms to search for connection in order to fuse together. In addition, missing awareness of cross-modal matching from many frameworks leads to performance reduction. To solve these two drawbacks, we propose an Iterative Contrastive Alignment Framework (ICAF) that uses recurrent alignment and contrast to capture the coherences between images and texts. Specifically, we design a recurrent alignment (RA) layer to gradually investigate fine-grained semantical relationships between image patches and text tokens. At each step during the encoding process, cross-modal contrastive losses are applied to directly optimize the embedding space. According to ROUGE, relevance scores, and human evaluation, our model outperforms the state-of-the-art baselines on MSMO dataset. Experiments on the applicability of our proposed framework and hyperparameters settings have been also conducted. | ['Lu Zheng', 'Qian Zhang', 'Jing Xiao', 'Youxin Chen', 'Chang Shu', 'Zijian Zhang'] | 2021-08-11 | null | null | null | null | ['abstractive-sentence-summarization'] | ['natural-language-processing'] | [ 4.84067053e-01 -1.64603442e-01 -1.91358954e-01 -3.22080761e-01
-1.02281582e+00 -3.21490884e-01 9.01742697e-01 3.19025815e-01
-4.16614950e-01 3.07945460e-01 7.18565345e-01 1.71567842e-01
-8.35417509e-02 -4.31786984e-01 -5.65233052e-01 -7.03720272e-01
3.49462926e-01 1.97025254e-01 5.94785847e-02 -1.54704511e-01
5.14564633e-01 -8.45174268e-02 -1.50859797e+00 6.85255408e-01
1.15553808e+00 7.57359564e-01 4.52057987e-01 4.92209077e-01
-5.31408310e-01 5.53157151e-01 -4.67282593e-01 -7.69203126e-01
-1.83044195e-01 -6.80843830e-01 -8.09535503e-01 2.57963121e-01
6.37844026e-01 1.81266423e-02 -2.56654352e-01 1.27276671e+00
7.09662855e-01 1.50013104e-01 7.70889640e-01 -9.53268170e-01
-8.47705722e-01 7.72720695e-01 -9.24807131e-01 1.51588783e-01
4.15473014e-01 -1.62937008e-02 1.36463594e+00 -1.04778886e+00
4.81768072e-01 1.34025800e+00 4.04850423e-01 3.87535930e-01
-1.15481806e+00 -2.71006286e-01 4.91876572e-01 5.64381003e-01
-1.27503145e+00 -5.86202323e-01 1.03062725e+00 -1.67335302e-01
9.05992806e-01 4.39036936e-01 5.43004036e-01 1.03737533e+00
6.27572685e-02 1.18135905e+00 7.12807238e-01 -6.35153830e-01
-8.77796113e-02 2.35904604e-01 2.02489227e-01 6.50454283e-01
6.09977134e-02 -4.61248636e-01 -7.75233209e-01 9.25643668e-02
2.99109429e-01 -1.90871432e-01 -3.51445764e-01 -2.51001000e-01
-1.43385005e+00 6.09397471e-01 3.70776236e-01 4.20021683e-01
-4.88545448e-01 -6.36455417e-02 5.67640245e-01 2.14563340e-01
4.34538811e-01 3.37069273e-01 1.26735494e-01 5.63879088e-02
-1.06797874e+00 1.66190445e-01 3.14958274e-01 7.41855323e-01
6.42409980e-01 -3.43420178e-01 -6.00597739e-01 1.22480416e+00
4.82389808e-01 2.46851623e-01 6.64398909e-01 -7.79622436e-01
9.59129214e-01 8.27334583e-01 -2.49662757e-01 -1.28669965e+00
-1.38903856e-01 -5.21739900e-01 -1.05146885e+00 -4.76642191e-01
-1.69417143e-01 2.07315519e-01 -5.92868209e-01 1.65989661e+00
2.67769217e-01 1.78402394e-01 2.90861309e-01 9.23261762e-01
1.01933169e+00 8.12132120e-01 -4.52420190e-02 -3.95895928e-01
1.54701090e+00 -1.38616669e+00 -9.83717501e-01 -3.14382523e-01
4.11315918e-01 -1.10171616e+00 1.22986293e+00 1.55119702e-01
-1.35356116e+00 -5.88038921e-01 -1.14273524e+00 -3.03535193e-01
-1.43313363e-01 4.22189057e-01 2.58962542e-01 2.38293126e-01
-9.42590058e-01 3.50595713e-01 -6.49675906e-01 -5.25947690e-01
3.44592780e-01 3.05423737e-02 -2.10898206e-01 -3.22187915e-02
-1.12606585e+00 8.68906319e-01 4.48566854e-01 3.94109935e-01
-4.53116119e-01 -3.42494816e-01 -7.05956221e-01 1.28055334e-01
3.58656615e-01 -1.14121830e+00 9.94232595e-01 -9.12970245e-01
-1.38974106e+00 8.85552585e-01 -3.47193897e-01 -4.59920228e-01
3.42773497e-01 -4.79572028e-01 -2.64551848e-01 3.62829149e-01
3.37536111e-02 7.67293453e-01 9.11144197e-01 -1.45098734e+00
-7.00786531e-01 -3.38203698e-01 -2.52503268e-02 9.55219448e-01
-6.62634432e-01 -3.76418121e-02 -7.71873653e-01 -8.81066918e-01
3.48209977e-01 -6.96992517e-01 1.78396963e-02 -4.37827885e-01
-5.51205039e-01 -2.58152366e-01 6.65680110e-01 -7.46772647e-01
1.61886239e+00 -2.13990021e+00 7.22611606e-01 -1.59291595e-01
-1.13305161e-02 2.35304385e-01 -4.16209280e-01 5.19202232e-01
1.24325804e-01 -7.44486749e-02 -2.51484096e-01 -7.22082615e-01
1.30361274e-01 -1.02699228e-01 -4.35332417e-01 1.81991294e-01
2.41692185e-01 1.01073110e+00 -6.66462719e-01 -8.74045372e-01
1.46322042e-01 4.42906231e-01 -4.65265840e-01 2.69890934e-01
-1.42582119e-01 2.81195700e-01 -3.70114475e-01 5.64676464e-01
5.76136053e-01 -2.03428283e-01 1.81374718e-02 -7.62184858e-01
8.65631774e-02 2.48115093e-01 -8.35742056e-01 2.27054739e+00
-4.01902705e-01 5.73843956e-01 -1.65158421e-01 -1.25838077e+00
7.39398599e-01 2.28842333e-01 2.69362062e-01 -6.77856326e-01
1.36485755e-01 -3.77505422e-02 -2.24577129e-01 -7.26651013e-01
9.29257333e-01 1.60131902e-01 -1.28901144e-02 3.22564036e-01
1.17869750e-01 -4.08575796e-02 2.01390892e-01 4.31116462e-01
7.07462609e-01 2.31466576e-01 2.28419662e-01 4.31912169e-02
8.63423705e-01 -7.39744976e-02 2.90928572e-01 5.63591421e-01
-9.82623622e-02 7.73276567e-01 4.34962928e-01 3.09002411e-04
-9.12626982e-01 -9.45800066e-01 1.28352091e-01 1.13390863e+00
5.11164010e-01 -5.78094304e-01 -8.35318983e-01 -5.85316896e-01
-3.86093616e-01 7.80573368e-01 -5.76013982e-01 -2.86018819e-01
-5.84298134e-01 -8.69943738e-01 5.21767080e-01 2.36744925e-01
6.14148796e-01 -8.91363502e-01 -3.84681553e-01 1.10891744e-01
-8.24675798e-01 -1.17437351e+00 -6.72348022e-01 -2.72045255e-01
-8.87304664e-01 -8.55756462e-01 -7.87517369e-01 -8.66094708e-01
6.28227115e-01 6.62064970e-01 1.06966603e+00 -3.73920090e-02
-8.73413458e-02 5.43929636e-01 -5.35941005e-01 -1.26258522e-01
-2.96433806e-01 2.51667082e-01 -4.40237731e-01 2.31210366e-01
2.00486332e-01 -5.25391102e-01 -7.77153552e-01 1.73549578e-01
-1.01840711e+00 4.01132584e-01 8.94943833e-01 9.37000096e-01
5.61499357e-01 -1.42271116e-01 5.02414703e-01 -5.09959280e-01
9.34808969e-01 -3.52567375e-01 -2.30440661e-01 8.70847583e-01
-4.06968474e-01 1.86049864e-01 2.77762979e-01 -3.83262843e-01
-1.34239125e+00 -2.71233290e-01 2.63334095e-01 -4.28797424e-01
2.47683674e-02 7.48394370e-01 -3.66112888e-01 4.91469771e-01
2.26675853e-01 5.29342592e-01 -7.37083778e-02 -3.34282696e-01
6.78400755e-01 7.46756911e-01 6.93669200e-01 -6.17061675e-01
4.34064209e-01 3.81946981e-01 -2.59637147e-01 -8.11298966e-01
-9.37324643e-01 -4.78579074e-01 -6.25633538e-01 -2.73280054e-01
9.67408895e-01 -1.04507625e+00 -4.02583718e-01 4.33120519e-01
-1.44036853e+00 3.62474471e-01 -2.64647696e-02 4.53347713e-01
-3.47897500e-01 7.00988591e-01 -2.92557806e-01 -5.83497047e-01
-6.31830335e-01 -1.18115163e+00 1.29729223e+00 2.13443890e-01
-9.10425261e-02 -9.01110709e-01 1.28444120e-01 8.03643227e-01
3.16425920e-01 -1.94337949e-01 9.56282973e-01 -4.65097338e-01
-4.99640107e-01 -5.54554649e-02 -3.15735281e-01 2.96182066e-01
6.10927530e-02 1.13018779e-02 -8.02434206e-01 -1.95370436e-01
4.57757302e-02 -3.84841770e-01 1.22610283e+00 3.17662179e-01
1.04386938e+00 -3.55809778e-01 -2.87257910e-01 2.61991054e-01
1.08923447e+00 -1.90090835e-01 7.85416782e-01 3.53979468e-01
7.52344668e-01 9.20706987e-01 6.38596833e-01 3.16670805e-01
7.59399712e-01 7.95518219e-01 2.95602769e-01 8.10326636e-02
-3.02685231e-01 -2.38819256e-01 4.89011377e-01 1.39234447e+00
4.95315641e-02 -4.51647788e-01 -5.15822113e-01 5.97295702e-01
-2.15183020e+00 -1.17397535e+00 3.04102208e-02 2.09771538e+00
8.88216078e-01 8.35815328e-04 -1.07091814e-01 -1.99333265e-01
9.47942436e-01 4.36789989e-01 -2.85787314e-01 -1.71406209e-01
-3.40514451e-01 -3.55276555e-01 -5.15488014e-02 3.89115840e-01
-1.08164144e+00 9.75594938e-01 5.22876596e+00 1.09046233e+00
-1.06753457e+00 -1.49912117e-02 4.83705491e-01 -1.58485174e-01
-6.38820291e-01 1.11768819e-01 -5.86307585e-01 3.82941633e-01
4.22330409e-01 -1.17603391e-01 3.91423672e-01 1.02452770e-01
2.70895381e-02 -2.07235157e-01 -1.02188766e+00 1.12489593e+00
5.44103920e-01 -1.20155585e+00 4.60123837e-01 -1.60251692e-01
7.79658258e-01 -2.33546570e-01 2.46736720e-01 1.53221309e-01
-1.39714867e-01 -8.18325937e-01 6.85417295e-01 7.25357473e-01
2.76882172e-01 -6.89902067e-01 7.90072203e-01 1.99562922e-01
-1.22207451e+00 1.46429524e-01 -2.81600237e-01 4.31213170e-01
2.39542708e-01 4.97675329e-01 -6.20501280e-01 1.06294155e+00
4.38321084e-01 7.44038522e-01 -8.12424958e-01 8.62639070e-01
-1.96468264e-01 2.96748549e-01 -3.44940908e-02 4.25492711e-02
3.08157444e-01 -3.10830206e-01 7.34829426e-01 1.26339328e+00
4.51376349e-01 -1.92800030e-01 -6.40835799e-03 7.30490327e-01
-9.31661353e-02 3.85606945e-01 -3.10841054e-01 -3.99543904e-02
5.36495805e-01 1.18862188e+00 -4.32043940e-01 -4.13305253e-01
-4.34140503e-01 1.24481368e+00 3.89723122e-01 3.24534059e-01
-8.51300955e-01 -2.89367557e-01 2.02511802e-01 -2.97966361e-01
4.06977147e-01 -3.02093453e-03 -3.82790059e-01 -1.47279537e+00
3.72018903e-01 -9.58892703e-01 4.93899465e-01 -7.15844035e-01
-1.22646928e+00 5.38582623e-01 1.40650630e-01 -1.45185935e+00
-2.76989117e-02 6.98145702e-02 -6.02296293e-01 6.55488431e-01
-1.55006492e+00 -1.40641427e+00 -1.69992775e-01 4.49068457e-01
9.35716271e-01 -2.78621048e-01 6.38386488e-01 2.89896488e-01
-7.73439169e-01 6.57564342e-01 9.45303123e-04 -2.45419949e-01
8.83297503e-01 -1.01520944e+00 1.23515889e-01 9.56505775e-01
4.10960138e-01 6.76876903e-01 8.20616722e-01 -3.84116471e-01
-1.35346675e+00 -9.37784791e-01 9.04984534e-01 -7.65165538e-02
4.95149463e-01 -8.85891691e-02 -1.13772953e+00 3.29475760e-01
1.01899743e+00 -4.65198010e-01 6.92707241e-01 1.74722910e-01
-3.96543324e-01 -2.24290252e-01 -6.66953087e-01 8.47538054e-01
8.35355401e-01 -4.64946508e-01 -9.09884095e-01 8.32187310e-02
6.90907478e-01 -1.90862685e-01 -6.12260461e-01 4.12331492e-01
4.98105764e-01 -9.70494866e-01 8.69484425e-01 -3.19093436e-01
7.05145657e-01 -3.67452592e-01 -3.10162693e-01 -1.35215008e+00
-7.43967071e-02 -6.50411129e-01 -1.15138248e-01 1.67483163e+00
3.01344305e-01 -1.54266745e-01 3.47094297e-01 1.53998733e-01
-2.24197716e-01 -6.55606747e-01 -8.71745050e-01 -3.21898788e-01
-4.06346619e-02 -2.24391207e-01 5.23028612e-01 9.25770283e-01
2.11438909e-01 1.02278721e+00 -4.47894275e-01 1.63160235e-01
6.43064737e-01 2.90329307e-01 7.80579329e-01 -7.72174776e-01
-1.07525527e-01 -9.22873557e-01 -2.05037415e-01 -1.16781700e+00
2.61703670e-01 -8.39731932e-01 -2.90454086e-03 -1.74372089e+00
6.74170613e-01 5.66379018e-02 -3.33167970e-01 2.90979147e-01
-5.85851908e-01 4.29067016e-02 3.58599752e-01 4.37040150e-01
-1.10045397e+00 1.10040247e+00 1.18091977e+00 -3.94567698e-01
-1.51081473e-01 -3.41908336e-01 -7.95620918e-01 5.71961284e-01
5.87474346e-01 -1.33922368e-01 -5.42847991e-01 -8.16899478e-01
3.27156097e-01 3.21878083e-02 3.51756364e-01 -8.64071369e-01
5.09761453e-01 4.16942611e-02 -1.13226175e-02 -9.69688594e-01
3.49450827e-01 -6.79925799e-01 -7.96855241e-02 2.18530893e-02
-7.59895265e-01 2.90230662e-01 -2.43647750e-02 6.44338965e-01
-6.28703177e-01 -3.19682986e-01 5.54793358e-01 4.18509878e-02
-5.77732563e-01 -1.68697014e-01 -5.98720983e-02 1.02943256e-01
8.13281894e-01 -1.94013491e-01 -5.35614133e-01 -2.87640810e-01
-4.42319065e-01 4.06957656e-01 3.20403188e-01 5.61314344e-01
8.29694092e-01 -1.50976574e+00 -9.97964263e-01 -7.48417601e-02
2.31837183e-01 -6.66277856e-03 6.73046172e-01 1.00832772e+00
-1.56358957e-01 3.83900881e-01 -4.64433208e-02 -7.36879826e-01
-1.49668109e+00 4.01244670e-01 6.37347922e-02 -5.56854010e-01
-4.02186245e-01 7.42783904e-01 3.25788915e-01 -2.75887072e-01
4.37613964e-01 -4.86872196e-02 -4.86891896e-01 3.70440245e-01
5.51510155e-01 3.80045801e-01 7.80556798e-02 -7.91207910e-01
-3.17931116e-01 7.36593306e-01 -4.46840525e-01 -2.95722246e-01
1.17243052e+00 -7.03034699e-01 -3.59068722e-01 4.80549067e-01
1.21836436e+00 -1.74673215e-01 -8.89108300e-01 -5.44005156e-01
1.06453970e-02 -3.58610719e-01 1.14733418e-02 -6.97746694e-01
-9.12439108e-01 9.37604368e-01 5.06856441e-01 7.27975965e-02
1.34087515e+00 1.25689968e-01 7.52647460e-01 3.61743242e-01
-3.04486513e-01 -1.24653375e+00 5.05442977e-01 3.79430115e-01
1.10296071e+00 -1.25716555e+00 2.80967772e-01 -3.19312066e-01
-9.96139705e-01 9.59865212e-01 5.67920685e-01 2.08447888e-01
3.71657051e-02 -2.77177215e-01 -3.15193161e-02 -2.04073653e-01
-1.12300611e+00 -1.45553499e-01 7.82234907e-01 2.39898145e-01
6.02806509e-01 -1.95052817e-01 -6.03063107e-01 4.78045583e-01
-4.01271135e-03 -3.43689352e-01 2.32105225e-01 6.37381792e-01
-5.11233926e-01 -1.08289599e+00 -3.95816803e-01 1.50172651e-01
-4.08747435e-01 -2.29114458e-01 -4.51690644e-01 4.28742498e-01
-1.74046367e-01 9.82164681e-01 1.14365472e-02 -2.60829628e-01
4.33347046e-01 -7.05944225e-02 6.31222665e-01 -2.59225875e-01
-5.62570691e-01 5.14007092e-01 -5.52971736e-02 -3.02014112e-01
-9.22961593e-01 -7.60433555e-01 -9.16206241e-01 4.40975465e-02
-5.08238971e-01 1.72223702e-01 6.77451313e-01 9.93470907e-01
5.44623494e-01 7.02404201e-01 6.41544282e-01 -7.32991278e-01
-5.49793959e-01 -1.13892043e+00 -1.21718287e-01 5.50257325e-01
9.99013558e-02 -3.76649380e-01 -1.55975476e-01 1.55207589e-01] | [10.702460289001465, 0.9827584028244019] |
d274ce51-31e2-4297-a908-1bbb9a6f99b9 | object-detection-based-handwriting | 2106.14989 | null | https://arxiv.org/abs/2106.14989v1 | https://arxiv.org/pdf/2106.14989v1.pdf | Object Detection Based Handwriting Localization | We present an object detection based approach to localize handwritten regions from documents, which initially aims to enhance the anonymization during the data transmission. The concatenated fusion of original and preprocessed images containing both printed texts and handwritten notes or signatures are fed into the convolutional neural network, where the bounding boxes are learned to detect the handwriting. Afterwards, the handwritten regions can be processed (e.g. replaced with redacted signatures) to conceal the personally identifiable information (PII). This processing pipeline based on the deep learning network Cascade R-CNN works at 10 fps on a GPU during the inference, which ensures the enhanced anonymization with minimal computational overheads. Furthermore, the impressive generalizability has been empirically showcased: the trained model based on the English-dominant dataset works well on the fictitious unseen invoices, even in Chinese. The proposed approach is also expected to facilitate other tasks such as handwriting recognition and signature verification. | ['Suting Miao', 'Yucheng Hu', 'Yuli Wu'] | 2021-06-28 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 5.28886497e-01 -4.09683771e-02 1.09841421e-01 -4.04725969e-01
-6.44208848e-01 -8.91353250e-01 5.32389998e-01 -1.73649695e-02
-6.08706057e-01 4.17253762e-01 -7.19342753e-02 -1.44052580e-01
1.66139379e-02 -7.03252792e-01 -8.74415576e-01 -9.65789020e-01
1.27761140e-01 3.86096001e-01 -1.50925234e-01 2.93634832e-01
5.65732121e-01 1.15670061e+00 -1.34695756e+00 5.58670700e-01
7.50662863e-01 1.25607002e+00 -6.83814809e-02 8.46847773e-01
2.31802687e-01 7.98587024e-01 -9.52517450e-01 -9.59338188e-01
5.33864975e-01 5.74143603e-02 -3.62311661e-01 1.31629363e-01
6.63933277e-01 -1.08334422e+00 -7.69513667e-01 1.14255762e+00
3.90566200e-01 -2.01224640e-01 2.06428811e-01 -8.49569380e-01
-1.12969232e+00 6.93785548e-01 -6.32937491e-01 -3.70777875e-01
9.67688039e-02 1.67381212e-01 6.42890811e-01 -8.79924476e-01
7.14475751e-01 9.42297161e-01 4.93904680e-01 3.06357861e-01
-5.68454921e-01 -5.64594030e-01 -3.30784380e-01 9.51592028e-02
-1.44808507e+00 -3.32310557e-01 6.88874662e-01 2.52465233e-02
3.45349163e-01 5.26493788e-01 1.50418073e-01 1.16754556e+00
-1.75361484e-02 1.30759442e+00 9.46328282e-01 -3.40585321e-01
-1.05817899e-01 4.64024425e-01 4.64246184e-01 7.57051945e-01
5.17882049e-01 -3.68138812e-02 -5.43826640e-01 -2.06393539e-03
6.95657670e-01 2.13151649e-01 -3.41044664e-01 1.72413468e-01
-1.18556070e+00 3.59173566e-01 2.24546358e-01 1.61207780e-01
-3.21249962e-01 -5.31231649e-02 6.22878730e-01 1.26105949e-01
1.43599376e-01 1.96050867e-01 2.46951059e-02 -5.74177057e-02
-1.19394326e+00 8.45804363e-02 6.97727442e-01 1.31394577e+00
5.67885041e-01 1.83744263e-02 -4.38648462e-01 5.56807995e-01
1.02531668e-02 5.69574952e-01 2.53133267e-01 -3.67430776e-01
9.06138301e-01 6.57164156e-01 3.87923777e-01 -1.23624551e+00
-5.12191132e-02 -3.95474344e-01 -1.10681415e+00 1.18368506e-01
6.58348739e-01 -1.62396818e-01 -8.34438682e-01 6.71411812e-01
6.48772791e-02 -2.70902157e-01 4.96233255e-02 1.25812423e+00
5.11048377e-01 5.89340746e-01 -4.11004305e-01 3.84892046e-01
1.54633534e+00 -9.69992638e-01 -8.25150251e-01 2.94144392e-01
3.71408343e-01 -9.93455470e-01 8.79217982e-01 6.76045835e-01
-8.94603729e-01 -7.64380932e-01 -1.15744269e+00 -2.56975502e-01
-4.05962676e-01 1.13493776e+00 3.88555616e-01 1.25942874e+00
-7.32422173e-01 6.50317907e-01 -7.10621655e-01 2.14048661e-02
9.46278811e-01 5.91248989e-01 -4.54346776e-01 -1.16426922e-01
-8.09894443e-01 3.06332648e-01 5.50903499e-01 9.36370134e-01
-8.69150639e-01 -3.89442027e-01 -5.81356108e-01 1.77908108e-01
1.02141887e-01 1.82171822e-01 5.96291721e-01 -1.10093236e+00
-1.63161874e+00 1.00501299e+00 1.17724622e-02 -6.07676387e-01
1.16966963e+00 -3.59002739e-01 -5.13511956e-01 3.08599204e-01
-4.89957154e-01 1.46571785e-01 1.30170023e+00 -1.11163795e+00
-6.78977549e-01 -6.50725842e-01 -3.81849945e-01 -3.46597470e-02
-8.10228169e-01 4.64772493e-01 -7.94676423e-01 -9.78241026e-01
1.85234882e-02 -8.27351213e-01 2.98795670e-01 1.13705687e-01
-8.69506657e-01 2.62313634e-01 1.19923031e+00 -1.32589686e+00
7.55414009e-01 -2.29269361e+00 -2.70871192e-01 6.38529062e-01
9.57912020e-03 5.91344714e-01 -2.45396066e-02 1.05800070e-01
1.91035509e-01 -2.28041977e-01 -8.10959339e-02 -4.56874818e-01
1.86940998e-01 -2.81182051e-01 -8.12718868e-01 8.53954971e-01
2.55399883e-01 1.02965343e+00 -4.46581990e-01 -2.37845868e-01
9.62632447e-02 4.51904774e-01 3.68647464e-02 2.95754641e-01
1.55628681e-01 2.69545704e-01 -3.95172685e-01 9.23043191e-01
1.47723532e+00 1.97730482e-01 1.65416583e-01 -1.58850849e-01
1.03232628e-02 -1.73807159e-01 -1.26513910e+00 1.27284622e+00
-9.72240046e-02 6.94539964e-01 4.49885458e-01 -5.37652731e-01
1.36092663e+00 -7.07611144e-02 8.61645639e-02 -5.51973939e-01
1.52063400e-01 2.98557848e-01 -3.25715572e-01 -4.60299462e-01
1.19036615e+00 6.05149388e-01 1.91471900e-03 8.65765393e-01
-3.71204704e-01 3.93586725e-01 -3.36015970e-01 9.89111587e-02
7.27806151e-01 2.36314818e-01 -4.18797761e-01 -4.34131101e-02
7.59046555e-01 -1.14484534e-01 2.70089269e-01 8.67693186e-01
-2.19509732e-02 5.90591252e-01 3.04348707e-01 -5.18171668e-01
-1.31071603e+00 -6.50208890e-01 -1.61618620e-01 7.90287793e-01
1.98475793e-01 2.77136743e-01 -9.77283835e-01 -8.04299593e-01
1.70008853e-01 3.82703632e-01 -3.07750791e-01 1.21975064e-01
-8.03969204e-01 -4.42445934e-01 1.29744816e+00 7.41365373e-01
8.98058712e-01 -1.02000356e+00 -4.97862905e-01 6.10881066e-03
2.77922094e-01 -1.03489137e+00 -6.26758337e-01 -9.14688781e-02
-4.56420898e-01 -7.83386827e-01 -1.11356914e+00 -8.18397880e-01
1.02844536e+00 -5.06272241e-02 1.54636368e-01 1.35287523e-01
-4.68864441e-01 8.13269839e-02 -2.07592338e-01 -2.52595127e-01
-4.98238891e-01 -2.45879162e-02 -4.33716811e-02 8.72985780e-01
5.60851693e-01 1.37274101e-01 -5.99594057e-01 2.72456706e-01
-9.60255206e-01 -3.92709017e-01 6.64917052e-01 8.19263577e-01
4.06159908e-01 1.50405452e-01 8.50484446e-02 -9.48123991e-01
4.39668685e-01 2.09245980e-01 -1.04182470e+00 6.72550023e-01
-8.65623578e-02 -1.50667563e-01 9.49720979e-01 -3.66588831e-01
-1.39299667e+00 2.49464437e-01 2.66223587e-02 -4.71674740e-01
-3.55288208e-01 -1.20396286e-01 -4.25628424e-01 -4.37203199e-01
2.84166515e-01 6.20596468e-01 -1.89697981e-01 -4.46061999e-01
2.76004225e-01 1.44048488e+00 8.94078314e-01 -5.13027728e-01
7.97370255e-01 7.41458237e-01 -1.99764878e-01 -8.84200513e-01
-5.12184687e-02 -1.92321971e-01 -7.91286588e-01 -1.27759665e-01
6.79572403e-01 -6.73996210e-01 -1.19378603e+00 1.17040074e+00
-1.50545883e+00 1.98343441e-01 -3.11923977e-02 4.35327530e-01
-2.06265658e-01 1.04503834e+00 -1.00204432e+00 -1.11260939e+00
-6.56898379e-01 -1.06487978e+00 1.38276041e+00 2.84548640e-01
3.17182504e-02 -5.16860664e-01 -5.87282121e-01 5.97214520e-01
1.93100065e-01 3.94858718e-02 8.82760286e-01 -9.51436162e-01
-1.05781817e+00 -9.58899617e-01 -6.70083761e-01 5.47652185e-01
-8.60436335e-02 1.99265972e-01 -1.34253299e+00 -4.29786563e-01
-1.28027634e-03 -3.33267897e-01 8.70195270e-01 -9.74218473e-02
1.66691411e+00 -5.36232531e-01 -7.44365379e-02 8.42938423e-01
1.19371092e+00 1.82762966e-01 1.01316500e+00 1.92305699e-01
8.33164871e-01 6.55540168e-01 7.04746842e-01 6.77571118e-01
-2.89215595e-01 2.64808923e-01 3.59260589e-01 -2.58096606e-01
1.13097057e-01 -2.04148322e-01 2.36315250e-01 5.51842332e-01
-1.76561698e-01 -5.27613640e-01 -6.21176064e-01 3.81432772e-01
-1.58156669e+00 -7.70312905e-01 -2.96840578e-01 2.29960155e+00
4.55989093e-01 -2.07242787e-01 -2.15940684e-01 8.33118781e-02
1.16999996e+00 1.70126572e-01 -5.71633518e-01 -5.87692618e-01
-4.97305155e-01 2.43306503e-01 1.04571390e+00 1.74148083e-01
-1.29640853e+00 9.20640051e-01 5.43104315e+00 9.05916095e-01
-1.18423104e+00 -3.11515927e-01 8.18490922e-01 1.45422667e-01
-7.41721392e-02 -4.09857661e-01 -1.02999711e+00 5.62196553e-01
6.96098208e-01 2.39512220e-01 4.31683391e-01 8.12811673e-01
1.20379720e-02 1.57021433e-01 -1.02090287e+00 9.94511962e-01
2.73639888e-01 -1.40144563e+00 1.99558631e-01 3.80865633e-02
6.30102694e-01 -5.44463038e-01 4.94110465e-01 -2.70095944e-01
1.23314962e-01 -9.64637101e-01 9.20259118e-01 7.38023758e-01
1.06597888e+00 -9.45675135e-01 1.03565764e+00 2.65956789e-01
-6.73531175e-01 -2.93349266e-01 -7.12419450e-01 4.47105825e-01
-1.39649019e-01 5.66322505e-01 -9.49804187e-01 8.47788870e-01
5.30349433e-01 5.05402446e-01 -4.23598319e-01 6.95424795e-01
-2.79791206e-01 4.65010434e-01 -2.15993732e-01 -2.56255597e-01
2.52122998e-01 -4.06887680e-01 3.43991011e-01 1.37145329e+00
3.88330668e-01 -1.48364902e-01 -3.77141386e-01 1.10486782e+00
-5.14526010e-01 1.58612877e-02 -3.19073409e-01 -4.86360878e-01
2.99106956e-01 1.44797480e+00 -7.06258714e-01 -3.68834108e-01
-1.68210059e-01 1.61483407e+00 6.44734353e-02 3.50244790e-01
-9.80451703e-01 -1.13805640e+00 2.16412887e-01 -1.40634254e-01
5.89373946e-01 -9.29700136e-02 -6.73395514e-01 -1.18695033e+00
6.00283861e-01 -9.15636539e-01 1.47008777e-01 -6.42753422e-01
-1.08884430e+00 5.31812966e-01 -9.56438184e-01 -1.14512765e+00
2.69647717e-01 -1.00269055e+00 -4.79765415e-01 1.14699626e+00
-1.23224378e+00 -1.35366035e+00 -6.44651830e-01 5.97512841e-01
9.49840620e-02 -5.17203391e-01 7.09040761e-01 3.99782509e-01
-8.41035664e-01 1.21153200e+00 6.48484409e-01 8.37453485e-01
6.38199627e-01 -8.35506678e-01 5.61989725e-01 1.11409998e+00
-2.54456252e-02 8.43009531e-01 1.49661470e-02 -8.36897790e-01
-1.77968359e+00 -1.44034386e+00 7.01249301e-01 -4.56088930e-01
3.78653318e-01 -7.55721688e-01 -8.42648268e-01 7.00485468e-01
7.19588920e-02 -2.34942034e-01 4.19448197e-01 -5.81356943e-01
-3.95359129e-01 -3.57208014e-01 -1.39740360e+00 4.66890126e-01
6.87624574e-01 -8.00108790e-01 -2.78070658e-01 4.95576859e-01
2.45539948e-01 -5.54497123e-01 -8.29480171e-01 -2.11078450e-01
7.48770177e-01 -6.70783520e-01 7.42803872e-01 -6.63951337e-01
5.60873091e-01 -2.33121142e-01 -5.85332215e-02 -3.03409249e-01
2.51245312e-02 -9.28184390e-01 -9.58507583e-02 1.58907139e+00
1.47284418e-01 -5.45841634e-01 1.22614717e+00 9.59120572e-01
1.34173259e-01 -2.45698333e-01 -7.30913520e-01 -8.74802709e-01
-3.42623413e-01 -1.11155018e-01 1.03302932e+00 8.16253066e-01
-3.36522013e-01 -5.39385438e-01 -7.15391517e-01 5.52870989e-01
8.69172931e-01 2.61134654e-01 8.37006330e-01 -6.55861616e-01
-1.66141748e-01 -1.11992322e-02 -5.28764427e-01 -1.30347824e+00
2.35085487e-01 -8.78855526e-01 -1.27591699e-01 -6.74627960e-01
1.22273073e-01 -4.29651588e-01 -1.71914235e-01 4.35421407e-01
1.74128246e-02 2.91615427e-01 1.42353073e-01 2.15887055e-01
-2.57438242e-01 2.66457856e-01 9.77042854e-01 -5.31199098e-01
2.29133256e-02 2.56120622e-01 -3.73691529e-01 2.72955000e-01
5.77811122e-01 -2.76742607e-01 3.82398695e-01 -5.80240071e-01
-2.44999766e-01 -1.90816537e-01 5.88007271e-01 -6.53668702e-01
5.01270294e-01 3.65111679e-01 9.47677374e-01 -9.30496991e-01
7.45030493e-02 -1.18156421e+00 -2.31360331e-01 5.20783365e-01
-4.97835428e-01 -2.04622388e-01 1.01381734e-01 6.32534802e-01
-1.99666560e-01 -4.64348197e-01 6.12333357e-01 2.27284044e-01
-5.03229916e-01 2.91764796e-01 -1.44515142e-01 -7.58592308e-01
1.07366872e+00 -5.67806900e-01 -4.35699314e-01 -5.85868582e-02
-3.19772601e-01 -2.48165324e-01 4.30468529e-01 2.47316346e-01
8.00811827e-01 -1.10447788e+00 -7.64125586e-01 5.75901270e-01
-1.06741481e-01 -1.94395378e-01 4.86961991e-01 5.24052978e-01
-1.16671336e+00 6.37863278e-01 -3.88868511e-01 -4.99481052e-01
-1.40136611e+00 7.51724601e-01 8.44017938e-02 -6.23978712e-02
-6.21929526e-01 8.04446459e-01 -3.11410248e-01 -3.12852472e-01
6.34796619e-01 -3.08228016e-01 1.50211349e-01 -1.51114896e-01
7.98018277e-01 7.89619386e-01 3.78082573e-01 -4.49104279e-01
-3.61222327e-01 3.99486750e-01 -5.96285880e-01 1.38599068e-01
1.33111882e+00 2.74516851e-01 -4.38666344e-01 -4.65175331e-01
1.44233692e+00 2.80259758e-01 -1.51146829e+00 -1.89505607e-01
2.21128948e-02 -7.58562565e-01 -2.71476626e-01 -6.20411277e-01
-1.21425319e+00 1.02414596e+00 5.55059195e-01 -2.78494626e-01
1.03056967e+00 -6.17933989e-01 8.72021973e-01 7.54480481e-01
4.29531425e-01 -1.33388042e+00 -2.90179580e-01 2.13466823e-01
8.10229480e-01 -8.23814988e-01 1.49406731e-01 -2.16156900e-01
-8.06792140e-01 1.57605183e+00 1.67484790e-01 -2.23026365e-01
1.19779781e-01 5.43360889e-01 3.82392853e-02 1.51270896e-01
1.92854285e-01 5.55635929e-01 8.09606984e-02 6.24599993e-01
-5.00502139e-02 -2.32597534e-03 6.84934333e-02 7.70998895e-01
-1.60467580e-01 -1.20848052e-01 4.97448057e-01 8.21279466e-01
5.06213419e-02 -7.98405826e-01 -6.85966015e-01 3.41720968e-01
-5.55467725e-01 -1.07524693e-02 -9.13246751e-01 5.34257472e-01
2.17949366e-03 5.16821682e-01 2.51137316e-01 -2.62018591e-01
2.87684262e-01 -1.62105963e-01 3.02675307e-01 -4.39195102e-03
-1.28412020e+00 -1.84178323e-01 -1.75259545e-01 -4.69614148e-01
6.90045729e-02 -4.90163028e-01 -1.04489136e+00 -5.69052696e-01
-4.16946769e-01 -1.79739818e-01 8.90244246e-01 5.09596169e-01
4.71312791e-01 1.23275213e-01 9.15228367e-01 -5.27909398e-01
-9.02649164e-01 -5.55463672e-01 -9.61847305e-01 5.02429307e-01
3.35381448e-01 3.24941486e-01 -5.43359183e-02 5.67945763e-02] | [12.25903034210205, 1.804038166999817] |
eafc0539-1d5e-4b81-93a9-880e85d5d6a3 | approaching-small-molecule-prioritization-as | 1911.10241 | null | https://arxiv.org/abs/1911.10241v2 | https://arxiv.org/pdf/1911.10241v2.pdf | Cross-modal representation alignment of molecular structure and perturbation-induced transcriptional profiles | Modeling the relationship between chemical structure and molecular activity is a key goal in drug development. Many benchmark tasks have been proposed for molecular property prediction, but these tasks are generally aimed at specific, isolated biomedical properties. In this work, we propose a new cross-modal small molecule retrieval task, designed to force a model to learn to associate the structure of a small molecule with the transcriptional change it induces. We develop this task formally as multi-view alignment problem, and present a coordinated deep learning approach that jointly optimizes representations of both chemical structure and perturbational gene expression profiles. We benchmark our results against oracle models and principled baselines, and find that cell line variability markedly influences performance in this domain. Our work establishes the feasibility of this new task, elucidates the limitations of current data and systems, and may serve to catalyze future research in small molecule representation learning. | ['Samuel G. Finlayson', 'Scott L. Lipnick', 'Matthew B. A. McDermott', 'Alex V. Pickering', 'Isaac S. Kohane'] | 2019-11-22 | null | null | null | null | ['cross-modal-information-retrieval'] | ['miscellaneous'] | [ 8.10486257e-01 -2.28093818e-01 -6.90935910e-01 -3.32583129e-01
-1.34862542e+00 -9.87503409e-01 5.63821256e-01 5.63534558e-01
-1.86752796e-01 1.20262814e+00 4.00191814e-01 -4.18777376e-01
-2.16243610e-01 -6.14423335e-01 -1.05675387e+00 -1.09856749e+00
-2.31935363e-03 5.18427551e-01 -4.64509279e-01 -8.44431669e-02
3.04113060e-01 6.96939707e-01 -8.63267004e-01 6.00003302e-01
6.90718234e-01 5.85828483e-01 -1.23620965e-01 4.72164750e-01
3.20310116e-01 6.92846656e-01 -4.82705563e-01 -1.31764129e-01
1.12836650e-02 -5.07422984e-01 -9.16713893e-01 -5.07420778e-01
6.29663110e-01 1.84207499e-01 -3.56116593e-01 7.88670003e-01
1.04415536e+00 1.65995374e-01 1.09061635e+00 -6.94172621e-01
-6.26921237e-01 8.23848471e-02 -3.87988359e-01 2.45018870e-01
3.83954436e-01 1.89191207e-01 1.20986378e+00 -6.25515044e-01
8.55143487e-01 1.00138807e+00 4.65013117e-01 7.45021880e-01
-1.79563296e+00 -6.16937697e-01 -8.49531440e-04 1.02262795e-02
-1.35310709e+00 -6.75927579e-01 5.22409976e-01 -5.35695076e-01
1.32830477e+00 3.23229045e-01 2.02927113e-01 1.47118235e+00
6.18148386e-01 5.94397664e-01 8.29338253e-01 -6.68154657e-02
2.42920533e-01 -3.57904255e-01 -2.48692632e-02 7.44226038e-01
2.73877174e-01 -1.48175834e-02 -4.51536179e-01 -6.51930273e-01
2.63004899e-01 8.39838933e-04 -4.71833795e-01 -4.15336788e-01
-1.13180494e+00 8.42714310e-01 3.19909364e-01 2.69280255e-01
-2.96619296e-01 3.74016672e-01 3.63565922e-01 2.21882761e-01
5.92631876e-01 1.17975640e+00 -9.58808720e-01 9.77776945e-02
-6.31895781e-01 2.55591601e-01 8.82668495e-01 4.37954456e-01
5.44289768e-01 -2.86812901e-01 -3.11995596e-01 8.19883704e-01
1.69696391e-01 3.60108078e-01 3.30447078e-01 -5.75516224e-01
2.55041301e-01 3.54654044e-01 -7.91671351e-02 -8.68454218e-01
-6.23474181e-01 -4.78511065e-01 -6.03890419e-01 -2.14633599e-01
3.33956838e-01 -1.00726105e-01 -8.22480798e-01 1.93894815e+00
1.30648702e-01 3.44902962e-01 1.35936916e-01 4.54389066e-01
8.03721607e-01 7.00071156e-01 5.41633844e-01 -4.82816130e-01
1.06409633e+00 -6.75604224e-01 -4.13023919e-01 6.89417943e-02
9.73022521e-01 -6.71243429e-01 7.45037973e-01 2.95054555e-01
-8.68225455e-01 -7.86282774e-03 -1.11146343e+00 -1.08070068e-01
-5.40283442e-01 -6.36721104e-02 9.59602952e-01 4.35888648e-01
-5.96991122e-01 1.00763083e+00 -8.45244646e-01 -1.58397749e-01
7.24972606e-01 7.94017792e-01 -6.16495609e-01 -6.29458129e-02
-1.20012999e+00 1.12958539e+00 2.71256566e-01 -6.74623027e-02
-1.11178768e+00 -1.13850927e+00 -6.25209749e-01 1.85017541e-01
1.09366968e-01 -1.05250895e+00 9.29854453e-01 -5.04247308e-01
-1.52327347e+00 9.07131195e-01 -3.20959568e-01 -1.66974023e-01
-9.78499744e-03 9.44000706e-02 -3.09969276e-01 -1.07818246e-01
8.74088407e-02 4.14186567e-01 2.13341177e-01 -9.10623491e-01
-1.14318781e-01 -6.13116324e-01 2.31237002e-02 1.97145462e-01
-5.46186455e-02 -1.18180864e-01 -1.72310010e-01 -4.55556780e-01
-3.15963358e-01 -9.86079991e-01 -6.89890385e-01 -5.28719842e-01
-6.27835631e-01 -9.70965177e-02 -2.28736941e-02 -1.54389247e-01
1.07876635e+00 -1.72809303e+00 6.73217773e-01 3.40054929e-01
3.19643468e-01 1.32116452e-01 -4.48944837e-01 6.98106110e-01
-5.70272446e-01 4.84784395e-01 -1.23043686e-01 1.60223976e-01
-2.43904725e-01 -2.71404237e-01 -3.34358364e-01 7.64899492e-01
3.14629167e-01 1.19528162e+00 -1.02846646e+00 -1.00923344e-01
-3.06992084e-01 5.41097045e-01 -5.89010656e-01 1.31507456e-01
-5.62573850e-01 5.37021458e-01 -7.93572366e-01 8.55352163e-01
3.50269049e-01 -5.53544044e-01 4.79022801e-01 -4.61251527e-01
2.52502352e-01 4.92603034e-01 -3.82916480e-01 1.78414035e+00
-2.94088393e-01 2.05137104e-01 -4.21098113e-01 -1.10817385e+00
5.04474819e-01 4.38942313e-01 9.52962816e-01 -5.77878773e-01
2.76726801e-02 1.38360396e-01 1.77978903e-01 -3.69409829e-01
-5.39111644e-02 -3.99630278e-01 1.76232204e-01 3.26322585e-01
1.97507083e-01 -1.20224543e-01 2.77477186e-02 -1.24578565e-01
1.21686387e+00 1.49999753e-01 6.46846354e-01 -1.92106247e-01
5.23824036e-01 -6.91896752e-02 4.38175291e-01 5.87715387e-01
1.21427663e-01 4.70976651e-01 7.95387089e-01 -4.65071499e-01
-9.10638452e-01 -8.59989524e-01 -4.34503734e-01 1.14943564e+00
-1.47825748e-01 -5.79196692e-01 -3.81516546e-01 -8.35742176e-01
1.07647248e-01 2.58011669e-01 -9.37080324e-01 -4.98270214e-01
-3.50544810e-01 -1.44816291e+00 6.69535577e-01 3.07822108e-01
-3.53646010e-01 -6.67833686e-01 2.77394712e-01 3.14211071e-01
7.36497864e-02 -7.34903336e-01 -5.03250122e-01 7.10906982e-01
-8.28423738e-01 -1.38649952e+00 -5.96522033e-01 -6.92373335e-01
5.67412794e-01 -3.28191705e-02 1.12633777e+00 -2.16410100e-01
-3.64605069e-01 8.14095289e-02 1.21387072e-01 -6.23204947e-01
-3.36717665e-01 4.61496264e-01 4.44691330e-02 -1.63313657e-01
3.71246159e-01 -7.63688147e-01 -7.83410251e-01 7.47658461e-02
-8.55335057e-01 -2.72367775e-01 7.71341681e-01 1.06615555e+00
1.11821091e+00 -4.24564868e-01 7.88352072e-01 -1.41428113e+00
8.56465578e-01 -6.74267530e-01 -4.20809388e-01 4.52959150e-01
-6.50248885e-01 4.84030813e-01 6.79515421e-01 -1.93389311e-01
-5.86866319e-01 3.76149476e-01 -3.06305587e-01 -9.68412682e-02
-1.43569872e-01 9.83217478e-01 -4.68639731e-01 -2.50101656e-01
8.52393210e-01 3.21933299e-01 -1.53200015e-01 -5.08059561e-01
3.23875874e-01 3.40011895e-01 6.32532313e-02 -8.59227419e-01
3.50825399e-01 2.73378521e-01 6.23815596e-01 -5.72965562e-01
-1.02097905e+00 -4.89612460e-01 -4.80516076e-01 5.96898437e-01
7.87269413e-01 -9.40280914e-01 -9.61667657e-01 2.64270473e-02
-1.02940261e+00 -2.90013283e-01 9.71479490e-02 3.76527965e-01
-7.95520186e-01 3.39413404e-01 -5.49097836e-01 -7.90397823e-02
-5.79376578e-01 -1.26586246e+00 1.30040395e+00 -1.90098107e-01
-2.40461111e-01 -1.07384217e+00 7.41850674e-01 3.93196940e-01
1.14426427e-01 5.38944960e-01 1.46717310e+00 -1.03756380e+00
-5.84856153e-01 -2.15662956e-01 2.84093153e-02 5.06848376e-03
4.95003730e-01 4.74329144e-02 -1.03574264e+00 -4.44585204e-01
-4.11639124e-01 -6.01596236e-01 1.12658644e+00 5.70273340e-01
1.39854348e+00 -2.22937107e-01 -7.92709351e-01 9.76860046e-01
1.29938877e+00 3.33688438e-01 6.22995973e-01 1.90400690e-01
7.68319309e-01 2.73915589e-01 3.24552715e-01 1.14609219e-01
-1.23650983e-01 7.44373739e-01 2.24369913e-01 -3.87286186e-01
2.25507364e-01 -2.07326308e-01 1.36212111e-01 3.29169631e-01
-2.78953433e-01 -4.87325460e-01 -7.17384875e-01 9.05145109e-02
-1.71819007e+00 -1.15933681e+00 1.96452245e-01 2.13572264e+00
1.32217395e+00 -2.13698804e-01 -2.11693719e-02 -5.54305017e-01
2.34485254e-01 2.09440380e-01 -8.72274935e-01 -2.54400074e-01
-3.10529351e-01 7.55326867e-01 5.69329381e-01 4.75829124e-01
-1.27169180e+00 8.94560456e-01 7.47157860e+00 9.37008619e-01
-1.31290329e+00 -3.59520704e-01 1.02954161e+00 -1.24511026e-01
-5.46034873e-01 -3.26978832e-01 -6.71319366e-01 1.76388845e-01
1.13452196e+00 -4.18379873e-01 2.74155736e-01 5.44197679e-01
3.20818543e-01 3.98472279e-01 -1.86412513e+00 7.92988181e-01
2.54821237e-02 -1.75620687e+00 3.42569143e-01 3.03143024e-01
8.88036013e-01 3.19754750e-01 2.97881335e-01 -9.27914679e-03
2.08035037e-01 -1.79174602e+00 -2.65640765e-01 4.23138440e-01
7.95643032e-01 -6.79841816e-01 4.57629114e-01 -3.02807987e-02
-7.72059441e-01 2.64397621e-01 -3.57735246e-01 3.09392571e-01
-3.43037397e-01 1.74089640e-01 -1.10820258e+00 6.37841642e-01
-5.25610000e-02 1.01392329e+00 -4.35984999e-01 9.68148768e-01
-1.19692370e-01 5.16289294e-01 -2.55268943e-02 -3.68667021e-02
9.85274389e-02 -2.60493517e-01 2.28039354e-01 1.28157008e+00
-1.24096610e-01 8.85079652e-02 2.80861288e-01 7.66638994e-01
-5.55639684e-01 5.17613769e-01 -7.52857566e-01 -5.59514225e-01
2.20909178e-01 1.07958305e+00 -3.82365972e-01 -1.46733925e-01
-3.34943384e-01 7.98702300e-01 4.35443908e-01 5.94488382e-01
-7.81550407e-01 -1.69186980e-01 1.20432520e+00 -1.35257900e-01
-3.52131985e-02 1.36611745e-01 -2.10616246e-01 -1.30840099e+00
-4.47441608e-01 -1.23188627e+00 4.71398175e-01 -3.63730520e-01
-1.37971568e+00 2.45660841e-01 -3.68796170e-01 -1.06400752e+00
-1.15092151e-01 -8.62903416e-01 -3.54933381e-01 9.96779501e-01
-1.55559468e+00 -1.10492647e+00 3.48616213e-01 3.52438897e-01
2.58720666e-01 -1.23067856e-01 1.46327960e+00 3.78786087e-01
-7.50516832e-01 7.14145124e-01 7.14405298e-01 -2.31779248e-01
1.04079163e+00 -1.15059674e+00 2.22385034e-01 1.07146166e-01
3.27151984e-01 1.06498563e+00 7.51676559e-01 -4.74010110e-01
-1.58135080e+00 -1.16824341e+00 7.14549065e-01 -9.86384988e-01
6.46290779e-01 -2.58670598e-01 -7.04482079e-01 6.14826083e-01
-1.78347781e-01 1.68878008e-02 1.54584610e+00 4.41883087e-01
-5.37018716e-01 1.90673731e-02 -8.80063593e-01 3.92264456e-01
8.79679918e-01 -7.83336461e-01 -1.71899706e-01 8.98848474e-01
6.36999488e-01 -4.44817662e-01 -1.21142030e+00 5.30856073e-01
6.24454379e-01 -2.22180977e-01 1.39275444e+00 -1.76515734e+00
6.69144094e-01 -2.53614217e-01 -2.46226490e-01 -1.60414231e+00
-5.73603809e-01 -4.52128619e-01 1.14765219e-01 7.37783790e-01
1.09837413e+00 -4.23292994e-01 9.84225035e-01 5.88107824e-01
-4.29332852e-02 -1.11760807e+00 -8.64512205e-01 -4.56443965e-01
6.09576404e-01 1.75862610e-01 3.22196126e-01 1.23552907e+00
3.06428254e-01 9.03020918e-01 -3.13062549e-01 1.55164793e-01
1.30919322e-01 3.46043795e-01 5.82141519e-01 -1.16381383e+00
-5.68616748e-01 -4.78328615e-01 -3.35839540e-01 -7.63359785e-01
4.43969756e-01 -1.20784569e+00 -2.98324317e-01 -1.32843125e+00
7.34202743e-01 -3.48789543e-02 -8.38761389e-01 4.94737118e-01
-4.34244722e-01 -1.76705827e-03 -3.83184344e-01 5.90952439e-03
-5.78371227e-01 5.26900351e-01 1.20080101e+00 -7.48442709e-01
-1.63455144e-01 9.05977339e-02 -1.13337386e+00 2.47409388e-01
6.15975022e-01 -5.35562456e-01 -4.68589187e-01 -1.17563996e-02
4.08841819e-01 -8.78987983e-02 -1.25561967e-01 -3.66776019e-01
4.01588250e-03 -5.74018419e-01 6.61124289e-01 -1.27278892e-02
3.43702197e-01 -5.49446642e-01 1.66185796e-01 4.30663884e-01
-7.41641104e-01 -2.24598646e-01 2.75963515e-01 8.77986491e-01
-1.10969178e-01 1.24065422e-01 6.90038681e-01 -1.83157668e-01
-2.51429409e-01 7.04794168e-01 -2.73352176e-01 3.36814448e-02
7.44753242e-01 9.19028744e-02 -3.18831652e-01 -1.12991646e-01
-7.58998513e-01 7.50742853e-02 3.65299821e-01 1.94589749e-01
3.16880465e-01 -1.14124584e+00 -7.26610959e-01 -2.05348670e-01
4.72492307e-01 -2.94067204e-01 -3.77107374e-02 7.33382642e-01
-4.57855016e-01 8.10677767e-01 5.01955897e-02 -1.77971229e-01
-1.47326136e+00 7.76448607e-01 7.20523357e-01 -4.43412006e-01
2.00130180e-01 1.07713902e+00 6.23122513e-01 -3.55030984e-01
7.70850619e-03 -1.07696295e-01 -3.31863850e-01 2.35140696e-02
3.24379593e-01 -4.97311866e-03 3.02107364e-01 -3.78438085e-01
-4.57181185e-01 4.42807853e-01 -4.22291309e-01 6.05967045e-01
1.64339244e+00 5.03759205e-01 -2.52243310e-01 9.75707844e-02
1.64739537e+00 -5.94615601e-02 -7.67598689e-01 -3.24892998e-02
2.30513792e-02 -1.55154482e-01 -6.79860190e-02 -1.04756129e+00
-6.64061010e-01 6.85222685e-01 5.21571875e-01 -5.75024068e-01
7.78201640e-01 1.17295913e-01 4.27175522e-01 1.06347167e+00
2.32536085e-02 -7.46137381e-01 -6.12251982e-02 4.07021701e-01
6.97650194e-01 -1.30306804e+00 3.47482651e-01 -3.21999103e-01
-2.80063063e-01 1.07731390e+00 4.24397469e-01 2.36007825e-01
3.42064261e-01 1.04848884e-01 -1.66034326e-01 -6.33647740e-01
-1.02478802e+00 2.66650915e-02 4.51587945e-01 5.28882027e-01
1.18001747e+00 -3.33271525e-03 -4.42874551e-01 4.01988536e-01
2.94803739e-01 1.15102224e-01 2.07373366e-01 7.29196966e-01
-2.10359558e-01 -1.68292236e+00 1.10017926e-01 4.83761698e-01
-9.63113606e-01 -5.50547421e-01 -9.79423702e-01 4.09752250e-01
-1.79829732e-01 6.04685664e-01 -5.13617575e-01 -2.85115838e-01
3.79505664e-01 2.26796523e-01 7.29995191e-01 -8.71420503e-01
-3.92422885e-01 1.72597066e-01 2.30843455e-01 -5.71384609e-01
-3.76402199e-01 -4.79686737e-01 -1.03312314e+00 -8.58440325e-02
-3.64992350e-01 4.20955390e-01 4.41557586e-01 7.98696697e-01
7.53476381e-01 5.85227311e-01 6.72023714e-01 -4.49901670e-01
-5.81699967e-01 -7.00382054e-01 -4.25140381e-01 3.23014975e-01
3.36983979e-01 -4.31100398e-01 1.64159406e-02 2.00255781e-01] | [5.106541633605957, 5.827070236206055] |
cb81b604-6795-495f-9972-40f94edb8a74 | threat-model-agnostic-adversarial-defense | 2207.08089 | null | https://arxiv.org/abs/2207.08089v1 | https://arxiv.org/pdf/2207.08089v1.pdf | Threat Model-Agnostic Adversarial Defense using Diffusion Models | Deep Neural Networks (DNNs) are highly sensitive to imperceptible malicious perturbations, known as adversarial attacks. Following the discovery of this vulnerability in real-world imaging and vision applications, the associated safety concerns have attracted vast research attention, and many defense techniques have been developed. Most of these defense methods rely on adversarial training (AT) -- training the classification network on images perturbed according to a specific threat model, which defines the magnitude of the allowed modification. Although AT leads to promising results, training on a specific threat model fails to generalize to other types of perturbations. A different approach utilizes a preprocessing step to remove the adversarial perturbation from the attacked image. In this work, we follow the latter path and aim to develop a technique that leads to robust classifiers across various realizations of threat models. To this end, we harness the recent advances in stochastic generative modeling, and means to leverage these for sampling from conditional distributions. Our defense relies on an addition of Gaussian i.i.d noise to the attacked image, followed by a pretrained diffusion process -- an architecture that performs a stochastic iterative process over a denoising network, yielding a high perceptual quality denoised outcome. The obtained robustness with this stochastic preprocessing step is validated through extensive experiments on the CIFAR-10 dataset, showing that our method outperforms the leading defense methods under various threat models. | ['Michael Elad', 'Alex Bronstein', 'Bahjat Kawar', 'Roy Ganz', 'Tsachi Blau'] | 2022-07-17 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 6.40732050e-01 1.97358206e-02 2.84228355e-01 -5.75402305e-02
-7.31289744e-01 -8.59571815e-01 9.36004639e-01 -1.97368428e-01
-4.51966554e-01 4.71952021e-01 -1.06938206e-01 -3.52832139e-01
-8.38048309e-02 -8.17674637e-01 -1.05415380e+00 -1.25109899e+00
-2.00154439e-01 -5.78166768e-02 1.84486896e-01 -3.02146465e-01
1.52686104e-01 7.54214644e-01 -1.06498814e+00 8.38003531e-02
6.33739173e-01 1.07966650e+00 -3.45058948e-01 6.89812183e-01
3.87791663e-01 6.38861239e-01 -8.93701315e-01 -6.67308390e-01
5.05156636e-01 -3.93925339e-01 -4.75758791e-01 -7.11212456e-02
3.16157490e-01 -3.10923100e-01 -5.01128137e-01 1.56822622e+00
4.78854090e-01 7.87482876e-03 7.44590163e-01 -1.06493342e+00
-8.73995483e-01 5.36729038e-01 -4.51254249e-01 2.15390399e-01
5.77098392e-02 3.91728431e-01 4.83848453e-01 -4.80920374e-01
4.66393471e-01 1.34279525e+00 5.75423241e-01 9.36128497e-01
-1.41835928e+00 -6.26374841e-01 -7.79648498e-02 4.72180136e-02
-1.27612913e+00 -2.88274437e-01 1.05035222e+00 -3.98334891e-01
3.83970052e-01 1.45858184e-01 1.48104146e-01 2.02903080e+00
5.86888909e-01 3.58858287e-01 1.36664736e+00 -2.49182567e-01
5.89858115e-01 1.79300055e-01 -1.35765254e-01 2.09675610e-01
2.04024002e-01 4.64710444e-01 -1.76264495e-01 -3.40003252e-01
3.98482114e-01 -1.59536209e-02 -5.70047438e-01 -1.87795937e-01
-5.67819655e-01 9.35832977e-01 6.66332483e-01 2.46440783e-01
-6.01437330e-01 2.35306367e-01 4.38217461e-01 3.68172556e-01
4.38961625e-01 2.50041097e-01 1.67645458e-02 2.76745945e-01
-8.37302327e-01 1.58735514e-01 7.68382609e-01 2.54614294e-01
2.98367769e-01 5.92633247e-01 -8.12942311e-02 5.05385220e-01
3.48108172e-01 5.67498326e-01 3.11226875e-01 -5.76588631e-01
2.82776058e-01 3.64859030e-02 -1.46268979e-01 -1.17099881e+00
1.25477575e-02 -8.19638371e-01 -1.24895704e+00 7.70968556e-01
3.08243811e-01 -3.38417053e-01 -1.05955422e+00 2.03286409e+00
3.49721104e-01 2.84243733e-01 3.78616124e-01 6.79021358e-01
1.62153766e-01 5.04447639e-01 1.44473791e-01 6.66798055e-02
1.03806984e+00 -2.11184070e-01 -4.71106321e-01 -1.39068022e-01
-6.67770393e-04 -4.98387367e-01 7.27333665e-01 7.88513422e-01
-7.88842678e-01 -4.11333174e-01 -1.32566583e+00 6.37405515e-01
-4.03922051e-01 -4.32777047e-01 9.38132107e-02 1.32094657e+00
-1.09949625e+00 7.47518539e-01 -8.26581061e-01 -2.82976851e-02
7.01376855e-01 2.64271230e-01 -4.49109882e-01 -2.26564169e-01
-1.38515568e+00 7.57800400e-01 3.42480689e-01 3.58965218e-01
-1.53340316e+00 -6.24987423e-01 -5.86657882e-01 -1.53093904e-01
1.53907552e-01 -6.30478740e-01 7.49365330e-01 -1.20707250e+00
-1.56560409e+00 6.75292850e-01 4.13156539e-01 -1.14947486e+00
8.86761844e-01 -3.63433629e-01 -3.92008305e-01 4.26978171e-01
-3.74700010e-01 3.41373026e-01 1.69535148e+00 -1.53213120e+00
5.80001324e-02 -4.53528941e-01 1.61198661e-01 -3.06286484e-01
-7.07307279e-01 2.33695120e-01 6.79639056e-02 -9.86773431e-01
-1.32891327e-01 -1.06465435e+00 -3.77642900e-01 -2.81954527e-01
-7.07806945e-01 4.20644343e-01 9.37680900e-01 -4.28887248e-01
7.81834424e-01 -2.31693339e+00 2.91432142e-01 3.93676370e-01
2.17525333e-01 5.49935460e-01 -2.26162486e-02 3.95452410e-01
-3.91797513e-01 2.08013326e-01 -4.94467974e-01 -5.07490933e-01
1.59194518e-03 2.57349342e-01 -1.04214811e+00 7.72076607e-01
3.00548643e-01 7.14657605e-01 -5.52770376e-01 1.20876536e-01
1.75622568e-01 9.27449524e-01 -5.11528432e-01 1.41566917e-01
-1.84125394e-01 5.02498984e-01 -4.44649071e-01 3.46870512e-01
8.84369910e-01 2.88900077e-01 -2.70340979e-01 -3.79745327e-02
3.74440283e-01 -3.63917321e-01 -9.21830535e-01 1.09749770e+00
-2.77505219e-01 3.66532147e-01 1.98223427e-01 -1.22385550e+00
8.50943923e-01 2.94420183e-01 1.38126865e-01 2.41384543e-02
5.08383989e-01 1.25917837e-01 2.43294075e-01 -3.16843897e-01
-7.21753240e-02 -2.38187522e-01 -9.48270559e-02 1.96119905e-01
1.19102798e-01 -1.55185193e-01 -3.28318298e-01 3.25403690e-01
1.36171997e+00 -1.27124503e-01 -9.64050293e-02 -1.86861292e-01
6.70716584e-01 -3.74323308e-01 3.38027716e-01 9.78810132e-01
-2.14289412e-01 5.62637985e-01 5.18389404e-01 -3.51593524e-01
-7.99198329e-01 -1.16026723e+00 -6.88877106e-02 4.48907018e-01
-1.33824855e-01 1.71970248e-01 -1.32921576e+00 -8.46504271e-01
-1.89002082e-01 7.64734328e-01 -1.00384998e+00 -7.82567739e-01
-3.68527502e-01 -1.05916929e+00 1.06779850e+00 2.64612406e-01
6.15585566e-01 -9.62412238e-01 -5.52387357e-01 2.27780789e-02
4.37862247e-01 -1.25259662e+00 -5.88493198e-02 4.09998089e-01
-7.23296463e-01 -7.87062228e-01 -7.61733770e-01 -2.93057561e-01
6.09661162e-01 -1.64732620e-01 5.91938138e-01 -1.56431183e-01
-2.50132769e-01 4.56753820e-01 -3.64703625e-01 -4.95409399e-01
-9.27091002e-01 -2.99354732e-01 3.99486274e-01 6.18285775e-01
-1.24865010e-01 -9.33511496e-01 -4.54366535e-01 1.58235766e-02
-1.58481884e+00 -6.83434486e-01 6.19959176e-01 7.95233130e-01
2.99818754e-01 4.19592530e-01 5.12465000e-01 -8.34259450e-01
6.94789827e-01 -6.24833047e-01 -5.20960927e-01 -3.75863016e-02
-3.53710592e-01 1.15897685e-01 1.15974247e+00 -8.99413049e-01
-8.99594188e-01 -9.84391719e-02 -5.32693923e-01 -9.76777554e-01
-4.15216625e-01 3.74249667e-01 -5.27953565e-01 -4.97534215e-01
9.85994220e-01 4.87958491e-01 -1.84317380e-01 -1.86387688e-01
4.36490238e-01 2.94627547e-01 7.96613991e-01 -5.09464443e-01
1.59746182e+00 6.88683331e-01 3.48001271e-01 -7.85657585e-01
-6.77987754e-01 3.68745297e-01 -2.28289664e-01 -2.08025321e-01
8.28964174e-01 -4.55422908e-01 -3.82627219e-01 9.69342947e-01
-1.07680714e+00 -2.05064476e-01 -1.07342668e-01 2.68588305e-01
-3.86749834e-01 4.45640445e-01 -6.01476252e-01 -8.92053783e-01
-4.12020892e-01 -1.30351090e+00 5.04566371e-01 2.22550333e-01
1.74324036e-01 -8.97221565e-01 1.22245271e-02 1.25907823e-01
6.93344414e-01 6.77369416e-01 8.77718270e-01 -1.00497568e+00
-5.10446906e-01 -6.49721801e-01 1.77577466e-01 1.02959025e+00
-8.35476369e-02 9.40787345e-02 -1.12839842e+00 -3.76082450e-01
7.42243171e-01 -3.50507289e-01 9.01299000e-01 3.52691263e-01
1.21648550e+00 -4.43933308e-01 7.47448794e-05 9.46485102e-01
1.50052810e+00 2.49755532e-01 8.15384328e-01 4.52479303e-01
6.22186840e-01 4.76629436e-01 -1.25824008e-03 2.23837197e-01
-3.51505637e-01 4.28858727e-01 1.13213575e+00 4.42555696e-02
7.63487816e-02 -3.19657288e-02 5.60888588e-01 1.25565603e-01
3.55234109e-02 -4.71497357e-01 -7.47124374e-01 2.54104286e-01
-1.40200448e+00 -1.09381175e+00 8.42659697e-02 2.18463397e+00
7.18780518e-01 6.37436569e-01 -1.43770039e-01 3.74795228e-01
5.60279191e-01 3.20982397e-01 -7.11760163e-01 -3.30379695e-01
-3.02923203e-01 4.25701499e-01 5.69732249e-01 4.25505370e-01
-1.29825187e+00 8.67921352e-01 5.44657135e+00 9.35493290e-01
-1.40243638e+00 1.58640668e-01 7.70961404e-01 -1.69852395e-02
-6.58325031e-02 -2.99852163e-01 -4.62527305e-01 4.60941851e-01
1.04267132e+00 -8.04597791e-03 3.60407203e-01 8.73480380e-01
1.09911546e-01 3.36044103e-01 -7.62378216e-01 6.91865921e-01
2.82813668e-01 -1.15298331e+00 2.89432138e-01 2.13846162e-01
6.35189950e-01 -6.46293089e-02 7.21288204e-01 6.67080954e-02
5.05979955e-01 -1.16620064e+00 7.26400673e-01 5.91858625e-01
4.70912844e-01 -1.04928255e+00 6.61224008e-01 4.90665346e-01
-5.22321999e-01 -1.71148300e-01 -3.09082150e-01 4.23819155e-01
5.44515513e-02 7.78559923e-01 -4.32911128e-01 5.95530033e-01
7.51138270e-01 1.90524116e-01 -3.48501831e-01 5.45583487e-01
-5.78179002e-01 9.69977856e-01 -1.98408097e-01 3.09307396e-01
3.00808489e-01 -9.64904279e-02 9.58216190e-01 1.11608863e+00
1.60694093e-01 -1.93797216e-01 -1.06641904e-01 1.02726352e+00
-1.50803983e-01 -3.35800231e-01 -8.32285345e-01 1.27744019e-01
2.05066115e-01 1.20613587e+00 -7.07542717e-01 -5.35661392e-02
3.04894764e-02 1.23795676e+00 5.38012050e-02 5.47110796e-01
-1.15442395e+00 -3.80587429e-01 7.21176803e-01 -1.98107228e-01
3.89289975e-01 -2.22824678e-01 -2.34360412e-01 -9.29265618e-01
6.48692623e-02 -1.22006214e+00 1.35664850e-01 -3.90706152e-01
-1.40992880e+00 1.12260318e+00 -6.26744032e-02 -1.08216691e+00
-2.83306897e-01 -5.68208158e-01 -8.43245149e-01 9.69190955e-01
-1.30842638e+00 -1.19008374e+00 8.50865692e-02 7.85239995e-01
2.94361115e-01 -3.19548607e-01 7.07344472e-01 1.18215447e-02
-7.55835116e-01 8.01984131e-01 3.12327705e-02 2.22903267e-01
4.61420834e-01 -1.16022623e+00 4.42670673e-01 1.45799983e+00
2.05714613e-01 8.54156852e-01 1.15240324e+00 -5.27030826e-01
-1.37005687e+00 -1.27291191e+00 4.79662120e-02 -3.91379684e-01
9.55736578e-01 -4.64314133e-01 -1.11679518e+00 5.63736856e-01
4.29768294e-01 3.29439253e-01 5.05238712e-01 -6.32191181e-01
-7.48907983e-01 -1.04086690e-01 -1.41600490e+00 7.21106887e-01
7.44934857e-01 -5.66631794e-01 -5.95316410e-01 1.03595853e-01
6.82858586e-01 -1.75344288e-01 -6.34856164e-01 3.85280132e-01
2.27801383e-01 -1.02513599e+00 1.27122593e+00 -5.98815143e-01
4.37945753e-01 -2.84422547e-01 -3.23092222e-01 -1.52804899e+00
-1.19136600e-02 -1.00517106e+00 -2.80794740e-01 1.20921290e+00
1.49773061e-01 -8.94553363e-01 6.30838096e-01 4.22170937e-01
-4.06953655e-02 -7.08328247e-01 -1.07600951e+00 -1.00447249e+00
3.68051529e-01 -6.02615118e-01 2.57479548e-01 7.77481198e-01
-7.06353486e-01 -2.62804627e-01 -4.19887245e-01 7.53413498e-01
1.04141152e+00 -5.03378391e-01 7.02068329e-01 -8.99543464e-01
-4.28252399e-01 -3.92191857e-01 -5.82978427e-01 -5.48886180e-01
4.40357208e-01 -5.25595009e-01 -8.15784850e-04 -7.08956838e-01
-2.97340065e-01 3.59032303e-02 -3.22833568e-01 3.01195353e-01
-2.99210995e-01 5.48624754e-01 7.21677691e-02 1.34670913e-01
6.48694932e-02 5.40139437e-01 6.35572672e-01 -2.01718912e-01
1.39522761e-01 2.39655316e-01 -6.91603363e-01 8.42035234e-01
8.78800929e-01 -7.98700809e-01 -4.60301310e-01 -1.68203950e-01
-3.53291258e-02 -3.23210776e-01 8.24686110e-01 -1.36674464e+00
4.26660515e-02 7.96759799e-02 2.94679344e-01 -4.97911349e-02
4.14041847e-01 -9.48553920e-01 1.18098311e-01 7.26417720e-01
-2.66448766e-01 -1.88473418e-01 2.60203570e-01 8.77056122e-01
-1.31383911e-01 -3.14084142e-01 1.18269670e+00 7.20520392e-02
-2.47988403e-01 3.39517921e-01 -4.06370163e-01 -3.46611105e-02
1.19367492e+00 -4.20490801e-02 -2.23740876e-01 -4.14084285e-01
-7.67538369e-01 -4.99640763e-01 4.15607125e-01 2.70538688e-01
6.13638699e-01 -1.01391709e+00 -7.73853838e-01 3.09965372e-01
-3.06156844e-01 -3.59555244e-01 4.10287678e-01 6.65679574e-01
-2.74807662e-01 -1.48473606e-01 -1.33953184e-01 -5.64149261e-01
-1.16366887e+00 9.38570440e-01 5.73891699e-01 -2.63038427e-01
-5.61777711e-01 9.37183022e-01 2.60353208e-01 6.83986768e-02
1.50677919e-01 6.14150167e-02 -8.25681910e-02 -1.52932867e-01
5.80837131e-01 -3.08219325e-02 1.19780488e-01 -6.64468467e-01
-2.34071970e-01 3.95701557e-01 -1.22786239e-01 -3.29243273e-01
1.40513968e+00 1.36026755e-01 1.04749449e-01 2.41928324e-02
1.14727318e+00 1.41107306e-01 -1.38839698e+00 -1.09945923e-01
-3.54264975e-02 -3.50755095e-01 1.96040079e-01 -6.04756057e-01
-1.27679062e+00 9.14183259e-01 6.23726666e-01 6.48289800e-01
1.28370738e+00 -3.30272615e-01 6.38270557e-01 1.68182015e-01
2.71855831e-01 -4.43808109e-01 8.87566432e-02 3.40227455e-01
9.58469808e-01 -8.76366198e-01 -4.05798733e-01 -1.36284202e-01
-4.84556764e-01 9.53115642e-01 1.59391582e-01 -5.98441601e-01
7.30082631e-01 4.67413366e-01 2.87308425e-01 2.35526990e-02
-4.39425945e-01 1.95317969e-01 2.52607703e-01 8.46077263e-01
-2.98928261e-01 -2.47686908e-01 4.32649031e-02 5.24944901e-01
-1.61731809e-01 -3.19678813e-01 5.55786788e-01 6.99829459e-01
-8.53724629e-02 -1.05416632e+00 -8.23369920e-01 9.43102874e-03
-8.45684230e-01 -1.42969176e-01 -4.30987418e-01 5.37763357e-01
8.13778266e-02 1.07113242e+00 -5.85274220e-01 -5.34419715e-01
1.46105349e-01 9.29792598e-02 3.12827975e-01 -2.97616988e-01
-8.61857593e-01 -7.35735446e-02 -4.43071842e-01 -5.20942926e-01
-1.75600067e-01 -7.05082357e-01 -7.37330914e-01 -1.44157887e-01
-1.17594406e-01 -4.36630324e-02 7.96820760e-01 9.28112388e-01
1.73674807e-01 5.98177493e-01 9.68233407e-01 -1.12534463e+00
-1.24636889e+00 -7.92623699e-01 -4.58622128e-01 6.22939825e-01
5.15708864e-01 -4.99850094e-01 -8.95463943e-01 -3.90658975e-02] | [5.572411060333252, 7.882796764373779] |
24a92b66-2d7b-4325-9b42-c108d4c620f2 | polyvoice-language-models-for-speech-to | 2306.02982 | null | https://arxiv.org/abs/2306.02982v2 | https://arxiv.org/pdf/2306.02982v2.pdf | PolyVoice: Language Models for Speech to Speech Translation | We propose PolyVoice, a language model-based framework for speech-to-speech translation (S2ST) system. Our framework consists of two language models: a translation language model and a speech synthesis language model. We use discretized speech units, which are generated in a fully unsupervised way, and thus our framework can be used for unwritten languages. For the speech synthesis part, we adopt the existing VALL-E X approach and build a unit-based audio language model. This grants our framework the ability to preserve the voice characteristics and the speaking style of the original speech. We examine our system on Chinese $\rightarrow$ English and English $\rightarrow$ Spanish pairs. Experimental results show that our system can generate speech with high translation quality and audio quality. Speech samples are available at https://speechtranslation.github.io/polyvoice. | ['Yuxuan Wang', 'Siyuan Feng', 'Tom Ko', 'Mingxuan Wang', 'Yuping Wang', 'Zejun Ma', 'Lu Lu', 'Xi Chen', 'Ye Bai', 'Fengpeng Yue', 'Tang Li', 'Qiao Tian', 'Xuxin Cheng', 'Kexin Wang', 'Yunlong Zhao', 'Chen Xu', 'Zhiying Huang', 'Qianqian Dong'] | 2023-06-05 | null | null | null | null | ['speech-to-speech-translation', 'speech-synthesis'] | ['speech', 'speech'] | [ 0.04373016 0.29986015 -0.17800465 -0.11767933 -1.2105147 -0.5518887
0.5374795 -0.41505408 -0.00967555 0.56001574 0.56226903 -0.747115
0.70962435 -0.6073567 -0.60213727 -0.4870477 0.48869896 0.30410632
0.07205168 -0.4697137 -0.31280223 0.2300678 -1.3078116 0.5446299
0.8909824 0.51023924 0.50584453 0.9585576 -0.3311452 0.7200123
-0.6070298 -0.38206708 0.16315927 -0.98936874 -0.7762353 0.14216633
-0.08472401 -0.22912629 -0.46735236 0.8638003 0.75651157 0.08025409
0.47552133 -1.0807933 -0.5233321 1.0248852 0.11225823 -0.09594116
0.56626844 0.05271063 0.99538934 -1.1748328 0.67493004 1.4743317
0.26166213 0.7626441 -1.2697457 -0.7404413 -0.07470326 -0.02764081
-1.5422527 -1.221925 0.61806244 -0.21339637 0.97347355 0.50647396
0.57301706 1.3175395 0.19484027 0.8860701 0.97827166 -0.79485023
0.21294178 0.2865046 -0.36632693 0.3941951 -0.66724855 0.24301122
-0.7923981 -0.06324483 0.54066473 -0.3851105 -0.34779328 0.09007516
-1.2419987 0.6778385 -0.06931806 0.32015508 -0.20010693 0.10211542
0.4617602 0.7770934 0.56589115 -0.02500068 -0.23012213 -0.37411946
-1.0307406 0.07128394 0.8045476 1.4809444 0.48784766 0.548793
-0.0879434 1.2164339 0.42270908 0.98260415 0.8302746 -1.0562851
0.56963825 -0.1562431 -0.07402243 -0.42017418 0.20096566 -0.16040593
-0.69556946 -0.05283673 -0.15805861 -0.11918735 -0.77695733 1.5712384
0.09552282 -0.11852058 0.39515904 0.5775144 0.9226074 1.096209
-0.34007996 -0.5243731 1.1286352 -1.3408638 -1.086015 -0.03657074
0.54342717 -1.3162986 1.463199 0.23937234 -1.4036058 -0.64467275
-0.80026686 -0.09870017 0.0512578 0.3178502 -0.04567896 0.7929599
-1.3988919 0.24033438 -0.91693413 -0.32376304 -0.28588185 0.01374836
-0.18021919 0.2786894 -1.190259 0.42408028 0.06933763 -0.19485386
-0.9950432 -0.16690128 -1.044755 0.02100795 0.15917425 -0.48135164
1.8800982 -0.9994647 -2.2834299 0.5240748 -0.7219175 -0.36461577
0.39384645 0.04107397 -0.75670975 0.24560605 0.08392906 0.5504267
1.0166454 -1.2374812 -0.5502612 0.11472274 -0.51757675 0.25493222
-0.272443 0.469098 -0.61119026 -1.0492692 0.10811659 -1.099867
0.07807219 -0.16499835 -0.48709932 0.02767394 0.48135233 -0.95270044
1.5224305 -2.3922372 0.15077294 0.05494734 -0.29613653 0.19647598
-0.35912457 0.813944 0.03407764 0.12296162 -0.21570797 -0.8333144
0.02729107 0.31905133 -0.61427957 -0.10157972 -0.0237079 0.74848914
-0.77342516 -0.36661234 0.05928032 0.45173794 -0.61255354 0.46280867
-0.10502322 0.5029631 -0.12410319 0.6254552 0.31924573 0.39234546
0.25939134 0.32220384 -0.24983598 1.0244443 -0.8578058 1.8081264
-0.9515327 0.59980243 0.47558406 -0.46382615 1.1313822 0.91663164
0.18474129 -0.6385623 0.0673285 0.6396348 -0.04793461 -0.383482
0.60940903 -0.21604544 0.03314253 0.5450629 0.20323098 -0.8241345
0.10008094 0.12555452 0.73562795 -0.04249192 0.1463586 -0.25144246
0.5225385 -0.30180696 0.5308221 0.31739166 -0.15157525 0.81763965
0.11626382 0.16103767 -1.1881663 -1.2733979 0.17577183 1.2079272
-0.22000869 -0.7703274 -0.95731413 -0.21972004 -0.560939 0.95007706
0.08096526 -0.12371456 -0.70898974 0.04403437 0.8630928 0.2837648
0.09635793 -1.2546988 0.24413045 0.45717013 -0.76409435 -1.2144297
-1.1015036 -0.04851476 -0.59597784 -0.3404096 -0.86112344 -1.1553378
0.2422358 0.192732 1.0039419 -0.13414252 0.36434177 0.2570013
-0.60245633 -0.3401482 -1.3750626 0.14578126 0.47010294 0.11186935
0.02727507 -0.7593428 -0.15311621 0.35029852 -0.84723765 0.03058715
0.30174297 0.76078176 0.6620542 -0.24088411 0.5824646 -0.39211646
0.8620902 -0.16167378 -0.29953784 0.09180652 -0.45525748 -0.16697933
0.9882739 -0.3747216 -0.99368405 -0.11019231 -0.70341116 -0.4577023
-0.05349706 0.24158518 -0.48542836 0.35579824 0.55699307 0.6916699
0.12650907 -0.6472568 0.5838538 1.4335015 0.7243027 -0.58822304
0.72240996 0.13474734 -0.75866675 -1.1954592 0.01230325 -0.38306758
-0.39148116 -0.05343232 0.5499149 -1.1857415 -0.13810585 0.42591745
-1.3173933 -0.636144 -0.3894562 0.6237372 -0.87084967 0.36191472
-0.8703463 -0.87585634 -0.5486591 -1.4812684 1.2678921 -0.2364225
-0.29698035 -0.6919537 0.20415887 0.30047828 0.37835968 -0.52717257
0.7261555 -0.45230925 -0.31449386 0.362166 0.338236 0.6672407
0.4201566 0.1295625 -0.92018044 -0.39270994 -0.08792713 -0.14881486
0.5464779 0.09367634 0.5268529 -0.5671641 0.06040019 0.5341717
0.754086 0.37909198 0.6335803 -0.2331691 0.408225 0.39069778
0.35912216 0.34209037 0.4782067 0.98496866 -0.20465636 -0.11886019
-0.6442185 -0.60065365 1.1979018 1.9072032 0.28792945 -0.48011395
-0.91233987 0.8469275 -1.533595 -0.79654765 -0.04389054 1.9793042
1.1683621 -0.05989876 0.18307965 0.2696131 0.6358733 0.01632513
0.06624192 -1.0076499 0.05658591 0.3688853 -0.01945292 0.988966
-0.6266456 1.3512669 6.3737407 1.1674553 -1.2866834 0.32108873
0.2673941 -0.03528633 -0.6848826 0.01905897 -0.5782446 0.5199621
1.42756 -0.32580736 0.8063287 0.5211432 0.79957306 0.3899995
-0.96525264 1.0354568 0.0111685 -1.078381 0.16863722 -0.14085147
0.63107204 0.20499153 0.02946127 0.3344482 0.26613638 -0.77766275
1.2724363 0.1469312 1.221623 -0.6287439 0.2738512 0.5066101
-1.4210745 0.2298471 -0.11980449 0.1402552 0.44594696 0.36777362
-0.9248072 0.67133826 0.5880223 0.39093563 -0.03309178 0.47930208
-0.5580425 1.0698837 -0.22380829 0.06574935 -0.00945405 -0.29132754
0.85730016 1.3071097 0.7594431 -0.1119001 0.41406417 0.586312
-0.12388702 0.52953446 -0.6589607 -0.18913895 0.7413267 0.72937804
-0.3501266 -0.51491517 -0.25388345 1.2501974 -0.15183406 0.48206902
-0.5375163 -0.5069062 0.7253957 0.05174979 0.23658633 -0.42595527
-0.05768984 -1.4142059 0.19719844 -1.4113703 -0.21189417 -0.8975606
-0.7951113 1.1260747 -0.2860144 -1.3413575 -0.881658 -0.1751037
-0.49504384 0.9396946 -1.349749 -1.2097661 0.24824645 0.6142059
1.0148761 -0.5049159 1.121171 0.40487203 -0.30241928 0.7430482
0.36934337 0.13643728 0.84596914 -1.0027303 0.74864155 0.96494997
0.43920094 0.5186622 0.72975165 -0.5021254 -1.1853415 -1.1703012
1.5205208 -0.22544019 0.61612153 -0.7697359 -0.6671109 0.6359638
0.65299255 -0.3345472 0.784593 -0.28737316 -0.19611482 -0.13141172
-0.78858024 0.8661091 1.0667613 -1.0290335 -0.5200203 0.1536247
1.2440891 -0.27890572 -0.75314254 0.00751415 0.44154546 -0.63130945
0.5773025 -0.17402877 0.20826817 -0.28932956 -0.4670758 -1.7532616
0.07672609 -1.3407913 0.07043327 1.476922 0.7314829 -0.7052543
0.09969892 -0.08008979 -0.5301804 -0.27281463 -1.0748948 -1.11078
0.31987765 -0.5944774 0.8726209 0.61909837 0.28741163 0.5251253
-0.42569444 0.09944587 0.12317495 0.12342251 0.7830188 -0.5388739
-0.59702814 -0.35405195 0.20131303 -1.2305807 0.28403628 -1.0723325
0.34276563 -1.2651182 -0.29557672 -0.24032985 0.08167416 0.5690346
0.15087853 0.1603942 0.3487301 0.23235792 0.02172339 1.0442592
1.1337296 -0.18039553 -0.59311706 0.1809084 -0.41428363 0.33903545
1.090449 -0.4562472 -0.49234208 -0.5849897 -0.34744856 0.38395742
-0.12024705 -0.7711114 -0.01174302 -0.11508937 -0.50605506 -0.31307876
0.38150033 -0.5228204 0.1653025 0.3554125 -0.33404112 0.23092537
0.14595036 0.11593947 -0.62757325 0.06967649 0.686059 0.14219476
-0.14184757 0.19284704 -0.89469653 -0.15906179 0.41661406 0.22850904
0.05271295 -0.8101645 -0.92120796 -0.08883487 0.49409634 0.73377615
0.67694503 -1.635009 -1.0450513 0.5803541 0.19501437 -0.40147418
-0.24794778 0.607934 -0.4632906 0.44502854 0.24216019 -0.5688988
-1.2101527 0.37253657 0.23768266 0.11238557 -0.52349263 0.6173812
0.04673183 -0.876153 0.06795508 -0.58629 0.23513798 -0.36063838
0.5685158 0.19495586 0.33150947 -1.0046083 -0.17468913 0.10800558
0.18506774 -0.905325 0.94823116 -0.5054084 -0.11201876 0.55596846
1.0408481 0.5525491 -0.67473716 -0.31254867 -0.3661598 -0.2924769
0.03642616 -0.55620456 -0.67254096 1.0257431 0.35479394 0.18902035
1.124294 -0.07229757 1.1123827 0.27395123 0.4649519 -1.2817208
-0.1774381 0.83775544 1.0418823 -0.75598514 -0.75024617 -0.51387864
-0.55256665 0.93945044 0.13947159 0.33399642 0.5759906 0.48362854
0.6753198 0.5137916 -0.91531 -0.30419847 0.14960481 0.6188021
0.664055 0.44947714 -0.41568464 0.6476123 -0.80912113 -0.26461208
0.5419512 0.7093644 -0.4487864 -1.726966 -0.5343821 -0.21333297
-0.32734662 -0.47315183 -0.5234584 0.2673904 -0.18047737 1.4709406
-0.14608029 -0.52291137 0.42618123 0.4369772 0.11880005 -0.80584025
-0.49602553 0.753448 0.31256533 -0.59104985 -0.11305989 -0.57830876
-1.356796 -0.38216695 -0.19004096 0.3891244 0.6560266 0.6329846
0.47422615 0.5944237 1.1695042 -0.62110335 -0.59792405 -1.0770218
-0.5654962 0.06614573 0.5067562 0.06919753 -0.2037611 0.39917058] | [14.686823844909668, 6.853146076202393] |
566c8554-5ef9-40ac-88a8-8a0f4a762c18 | forged-image-detection-using-sota-image | 2211.15196 | null | https://arxiv.org/abs/2211.15196v1 | https://arxiv.org/pdf/2211.15196v1.pdf | Forged Image Detection using SOTA Image Classification Deep Learning Methods for Image Forensics with Error Level Analysis | The advancement in the area of computer vision has been brought using deep learning mechanisms. Image Forensics is one of the major areas of computer vision application. Forgery of images is sub-category of image forensics and can be detected using Error Level Analysis. Using such images as an input, this can turn out to be a binary classification problem which can be leveraged using variations of convolutional neural networks. In this paper we perform transfer learning with state-of-the-art image classification models over error level analysis induced CASIA ITDE v.2 dataset. The algorithms used are VGG-19, Inception-V3, ResNet-152-V2, XceptionNet and EfficientNet-V2L with their respective methodologies and results. | ['Pandharinath Ghonge', 'Nandan Kanvinde', 'Abhishek Gupta', 'Raunak Joshi'] | 2022-11-28 | null | null | null | null | ['image-forensics'] | ['computer-vision'] | [ 1.52382687e-01 -1.09449483e-01 1.55622795e-01 7.51548307e-03
-4.94139791e-01 -3.19928974e-01 7.19003201e-01 1.08227193e-01
-5.75296402e-01 7.22568333e-01 -3.91579002e-01 -8.01328540e-01
-1.34496495e-01 -8.96251082e-01 -8.43092382e-01 -5.87645173e-01
-6.08462729e-02 -1.80386811e-01 4.51646060e-01 -2.45760337e-01
6.95758045e-01 7.16154814e-01 -1.28766227e+00 7.43479192e-01
2.36819535e-01 1.41652071e+00 -3.47940564e-01 9.85946238e-01
-1.25325292e-01 1.25640690e+00 -8.52392554e-01 -6.77546382e-01
3.71916205e-01 9.49091185e-03 -9.94331360e-01 1.24780484e-01
6.94887817e-01 -2.62675762e-01 -6.58516169e-01 1.46503639e+00
4.42119390e-01 -1.95296451e-01 6.87239587e-01 -1.62649405e+00
-8.90164196e-01 1.76077083e-01 -7.28457749e-01 1.26216042e+00
-2.42026970e-01 3.85256827e-01 1.56357586e-02 -9.71717060e-01
4.98016298e-01 1.24628472e+00 1.03114510e+00 8.01334009e-02
-9.11547840e-01 -9.12510395e-01 -7.53675580e-01 9.28131104e-01
-1.00882506e+00 -7.88694546e-02 8.31507564e-01 -5.72439790e-01
1.12051308e+00 -2.59425268e-02 2.60863930e-01 1.00606894e+00
5.61719775e-01 6.45889938e-01 1.54158270e+00 -5.89165151e-01
1.11689396e-01 4.31612097e-02 4.89319324e-01 6.10758007e-01
5.49423337e-01 5.38649619e-01 -4.52936709e-01 3.41021031e-01
5.08603215e-01 -7.96999708e-02 7.47518390e-02 2.58976877e-01
-3.95886153e-01 1.09626210e+00 8.45164478e-01 3.80044937e-01
-2.82241732e-01 5.07204771e-01 8.06592762e-01 6.59158289e-01
4.37333584e-01 1.74701095e-01 -1.87225789e-01 1.14597909e-01
-1.22909153e+00 2.61415690e-01 9.46904793e-02 5.94606459e-01
5.45106173e-01 5.07590950e-01 -5.01933359e-02 4.33774024e-01
1.13308415e-01 7.72063562e-04 5.73297799e-01 -7.58120835e-01
2.77301192e-01 7.17059135e-01 -2.30681315e-01 -1.25929987e+00
-3.91370431e-02 -3.57312024e-01 -8.67089093e-01 1.08676505e+00
3.73160154e-01 1.62336364e-01 -1.49653924e+00 8.34291101e-01
1.54661775e-01 3.84924173e-01 1.33207180e-02 5.36455691e-01
9.02038097e-01 4.53772336e-01 1.60185888e-01 3.21992576e-01
1.48517454e+00 -7.39864826e-01 -5.80291986e-01 -6.21179603e-02
4.02317047e-01 -6.79581344e-01 3.77750725e-01 8.09394300e-01
-8.45221996e-01 -7.76102185e-01 -1.40263271e+00 -2.29727432e-01
-1.10387611e+00 -1.03829820e-02 4.18644428e-01 1.07167852e+00
-9.88044739e-01 8.84109795e-01 -5.09913862e-01 -7.93426111e-02
1.36060190e+00 3.91756356e-01 -6.27886355e-01 -7.47863855e-03
-9.85259891e-01 9.96279597e-01 7.31027365e-01 -1.23068787e-01
-1.10832703e+00 -7.40107298e-01 -5.33077419e-01 -9.10579339e-02
-1.85668375e-02 1.03276536e-01 8.66315722e-01 -1.02564144e+00
-7.16367602e-01 1.52925503e+00 8.29600811e-01 -1.22953475e+00
7.94756949e-01 5.28925136e-02 -7.29515135e-01 5.35041392e-01
8.97909328e-02 5.80053806e-01 1.25171161e+00 -1.08987558e+00
-7.91975856e-01 -4.86386180e-01 -2.48285934e-01 -7.37829864e-01
-8.33699778e-02 4.06943738e-01 1.48317307e-01 -6.66587055e-01
-2.76716381e-01 -5.21030843e-01 3.43737662e-01 1.38641626e-01
-5.05036950e-01 -1.04541719e-01 1.59549010e+00 -9.67153430e-01
8.87793541e-01 -2.04372978e+00 -6.24142766e-01 -5.25364876e-02
3.70029092e-01 1.09347558e+00 1.53986350e-01 1.52341202e-01
-6.30918741e-01 4.44127411e-01 -2.47208282e-01 -6.82554096e-02
-5.51356077e-01 -1.87593997e-01 -3.60036343e-01 9.93981600e-01
4.68763649e-01 9.36965287e-01 -6.47828043e-01 -5.80952644e-01
7.32527912e-01 2.83122003e-01 -8.79657418e-02 3.49617787e-02
2.10138813e-01 6.65650368e-02 -1.16461486e-01 8.36740851e-01
1.13418090e+00 1.38651609e-01 -6.45787001e-01 -4.97520924e-01
-3.75046507e-02 -4.52565014e-01 -7.87133336e-01 1.15655708e+00
-7.38834962e-02 1.36927974e+00 -1.01144716e-01 -1.61437833e+00
8.10194612e-01 4.63614874e-02 2.56258041e-01 -7.55313814e-01
7.32092321e-01 9.92619097e-02 1.35119751e-01 -8.97493899e-01
3.84387165e-01 -1.32896274e-01 2.72124708e-01 1.80102810e-01
6.53713226e-01 2.48118520e-01 -1.69154666e-02 1.33922040e-01
1.26862144e+00 -4.03833278e-02 1.04799882e-01 -1.92220852e-01
6.46665633e-01 1.45579055e-01 1.41362146e-01 6.96968853e-01
-7.49985695e-01 6.40550196e-01 4.37971503e-01 -8.03388715e-01
-1.48359466e+00 -7.22757578e-01 -4.11978304e-01 3.26202095e-01
-8.80555660e-02 1.67904288e-01 -1.08205175e+00 -7.70859361e-01
2.29164124e-01 7.10027277e-01 -1.09937203e+00 -2.67388463e-01
-6.40565038e-01 -7.26770282e-01 1.36323619e+00 4.40010071e-01
1.26398456e+00 -1.46853542e+00 -1.21076393e+00 1.21775702e-01
4.14016038e-01 -1.35965943e+00 2.07638279e-01 2.60976344e-01
-7.94766724e-01 -1.72634184e+00 -3.87500852e-01 -5.55653811e-01
2.51048625e-01 2.82192975e-01 1.11496270e+00 4.21507955e-01
-1.17575490e+00 2.05218166e-01 -5.96065521e-01 -7.46758103e-01
-4.81615126e-01 -3.84856462e-01 -3.56827796e-01 2.27701977e-01
7.01195180e-01 -4.87841785e-01 -8.01715910e-01 -7.60373175e-02
-1.31550455e+00 -3.83385450e-01 5.76663315e-01 8.66815627e-01
1.68243140e-01 2.36402392e-01 4.80017245e-01 -9.24299181e-01
5.09558856e-01 -7.28910983e-01 -4.99682486e-01 2.56723315e-01
-7.08283246e-01 -1.56224713e-01 3.22007447e-01 -2.67312467e-01
-7.22696722e-01 -3.50886852e-01 1.66865721e-01 -1.07720482e+00
-2.30869442e-01 1.27727926e-01 1.83251482e-02 -4.81714785e-01
9.67081189e-01 2.60035038e-01 4.57817782e-03 -2.27302104e-01
2.75742024e-01 9.30102050e-01 1.07720697e+00 -9.99298617e-02
6.28701150e-01 6.77577615e-01 3.07427585e-01 -9.18589652e-01
-3.54294360e-01 -4.33104306e-01 -5.93020618e-01 -4.69301313e-01
1.10898316e+00 -5.10106385e-01 -5.32978475e-01 1.11606634e+00
-1.21643794e+00 -2.59069446e-02 7.78197348e-02 -1.69207573e-01
-3.61294180e-01 4.86397386e-01 -5.61725795e-01 -1.00014198e+00
-6.96283758e-01 -1.25788152e+00 8.07785451e-01 2.32656285e-01
4.66589600e-01 -9.00248885e-01 -2.96363294e-01 5.27702749e-01
4.91796404e-01 8.64431322e-01 6.83459878e-01 -8.95669818e-01
-3.48773420e-01 -5.11170030e-01 -7.34914839e-01 8.62778127e-01
-1.45533934e-01 -8.16584453e-02 -1.15683019e+00 -2.13209867e-01
1.36694402e-01 -6.56851053e-01 1.20000029e+00 3.61026973e-01
1.52325809e+00 -8.70622098e-02 -1.15324005e-01 8.15013230e-01
1.89393866e+00 2.35870928e-01 1.38293743e+00 8.46288145e-01
5.09315431e-01 1.89667732e-01 2.58923680e-01 5.35467789e-02
-3.22596610e-01 3.22263479e-01 9.73999918e-01 1.80285461e-02
-7.74324000e-01 -9.29083824e-02 8.12033657e-03 -2.07721666e-01
1.49473980e-01 -2.25951955e-01 -1.02700329e+00 6.28142178e-01
-1.46581972e+00 -1.27810276e+00 -5.51630616e-01 1.69496179e+00
3.07213992e-01 2.34970510e-01 -2.46400777e-02 8.62193644e-01
9.79346752e-01 -3.93909365e-02 -5.13742328e-01 -5.20412922e-01
-2.62416244e-01 5.75246453e-01 9.77484167e-01 -1.00767903e-01
-1.41912103e+00 1.06625271e+00 5.91735458e+00 1.04577947e+00
-1.22684824e+00 5.32272518e-01 1.23104882e+00 3.72337162e-01
7.14314997e-01 -1.21831931e-01 -5.17242730e-01 8.17848623e-01
1.03152287e+00 5.47340453e-01 9.94621739e-02 7.41571546e-01
-6.99904785e-02 -4.11110222e-01 -6.88273787e-01 1.38843441e+00
3.18464845e-01 -1.79752052e+00 2.11208109e-02 1.70403376e-01
5.09361446e-01 2.14782748e-02 3.83495778e-01 2.63319016e-01
6.54970482e-02 -1.56657052e+00 6.85304284e-01 1.68048039e-01
1.01816034e+00 -1.02238798e+00 8.90287817e-01 4.87951189e-02
-5.85900605e-01 -3.61818790e-01 -6.55423045e-01 4.14697349e-01
-3.00689787e-02 4.00250018e-01 -5.35266221e-01 4.25111830e-01
1.33599329e+00 6.13199353e-01 -8.11663866e-01 1.08781540e+00
-1.62280705e-02 9.07096565e-01 1.46363720e-01 5.11726081e-01
4.63790596e-01 5.11657409e-02 4.46236253e-01 1.33882022e+00
2.47312352e-01 3.09294984e-02 -3.45068574e-01 1.04171944e+00
-3.69445980e-01 -4.56813604e-01 -7.02278018e-01 9.13047194e-02
2.19875842e-01 1.38293493e+00 -9.83713806e-01 -4.74609971e-01
-1.65606081e-01 8.43816102e-01 2.14111298e-01 -1.61428556e-01
-7.02603877e-01 -4.59250420e-01 2.29264289e-01 3.53260100e-01
5.66702843e-01 1.31111369e-01 -3.24754030e-01 -8.27831864e-01
-5.27071059e-01 -9.88104999e-01 4.37267095e-01 -9.29869831e-01
-1.24976707e+00 6.53687358e-01 1.43787965e-01 -1.07325351e+00
1.44282535e-01 -1.36802828e+00 -9.19933975e-01 5.14235377e-01
-1.75670493e+00 -1.45082760e+00 -4.88223255e-01 9.26427186e-01
8.37646306e-01 -9.29590583e-01 2.53782570e-01 4.11680430e-01
-3.78464222e-01 6.12109959e-01 -1.17023863e-01 7.35652566e-01
4.28583056e-01 -1.19139898e+00 2.91371405e-01 1.26115596e+00
8.72384831e-02 -1.53915212e-01 6.57207191e-01 -7.67068148e-01
-1.03366733e+00 -1.32674837e+00 4.11158614e-02 -4.47734386e-01
6.57926023e-01 1.31610483e-01 -9.78035748e-01 4.55562383e-01
6.37427628e-01 7.51271188e-01 4.01595861e-01 -9.26532269e-01
-6.27944767e-01 1.55754045e-01 -1.85385740e+00 -1.74854591e-01
5.78021407e-01 -5.23506582e-01 -6.78697348e-01 2.17338756e-01
7.57044256e-02 -2.26851493e-01 -6.60330772e-01 1.14017129e-01
9.56882238e-02 -1.37283850e+00 1.20854771e+00 -8.19992125e-01
6.39262676e-01 -1.27476111e-01 -2.29229361e-01 -9.65679824e-01
2.55656689e-02 -3.83272856e-01 -2.16010123e-01 1.14107430e+00
-1.98434576e-01 -1.79474115e-01 9.90800738e-01 5.50229847e-02
-6.76574139e-03 -3.12276036e-01 -1.09114230e+00 -8.91099393e-01
4.95454937e-01 -6.82369888e-01 3.08680743e-01 8.55365634e-01
-6.85668766e-01 -2.48646438e-01 -4.14144754e-01 7.89799914e-02
1.11155617e+00 -4.48848188e-01 6.35251641e-01 -1.05447292e+00
-2.74954829e-02 -4.74093378e-01 -1.24007022e+00 1.74127176e-01
-1.70945134e-02 -8.20410788e-01 -4.17300165e-01 -1.06375277e+00
-8.00686181e-02 -1.63212821e-01 -3.39561015e-01 4.32562500e-01
3.46734449e-02 7.49223471e-01 2.87181199e-01 2.44815588e-01
-1.63391203e-01 -1.66148897e-02 6.85392499e-01 -4.60280836e-01
4.78186548e-01 -1.53983667e-01 -2.76733667e-01 6.15236819e-01
9.90297079e-01 -7.48561800e-01 -2.04916716e-01 -8.44078809e-02
2.63935607e-02 -2.59956867e-01 9.35200036e-01 -1.42980039e+00
3.10351789e-01 2.94527799e-01 9.10570681e-01 -7.51961410e-01
1.66599989e-01 -6.57244325e-01 3.83164808e-02 8.18203926e-01
-1.41904950e-01 4.47415084e-01 4.25319344e-01 6.39832258e-01
-3.43426377e-01 -5.83375037e-01 1.26663542e+00 -4.66857523e-01
-1.04672050e+00 2.19252005e-01 -3.73510987e-01 -4.40864600e-02
1.32257175e+00 -6.84719503e-01 -6.20072186e-01 -1.03564123e-02
-4.12997991e-01 -3.00644755e-01 1.17524266e-01 3.68712008e-01
1.03264606e+00 -1.14987755e+00 -8.39607060e-01 1.16632255e-02
9.77737233e-02 -4.35973525e-01 2.22871512e-01 4.81011033e-01
-9.96811330e-01 1.66198343e-01 -1.00130999e+00 -5.60845971e-01
-1.45783198e+00 8.97465706e-01 7.05256104e-01 -8.81362334e-02
-6.27643406e-01 7.18077362e-01 -4.80392754e-01 2.16250956e-01
-8.94537494e-02 9.10662636e-02 -4.04361129e-01 -5.51182739e-02
8.48081827e-01 7.81750798e-01 4.21223164e-01 -8.00874174e-01
-3.72053057e-01 2.02060267e-01 -2.89715767e-01 1.89672396e-01
1.52787542e+00 2.28886187e-01 -2.36012265e-02 -1.86976850e-01
1.45780110e+00 -7.62012064e-01 -1.05525887e+00 4.16992679e-02
2.86501497e-01 -7.77398050e-01 7.19797671e-01 -9.23757493e-01
-1.64730597e+00 1.25476050e+00 1.42999899e+00 1.75823182e-01
1.10227966e+00 -3.75037998e-01 6.73989892e-01 -2.30213888e-02
1.58806518e-01 -1.33110726e+00 5.99857509e-01 1.46770496e-02
8.14470470e-01 -1.67424953e+00 1.25028715e-01 -3.04536019e-02
-4.09016818e-01 1.44124866e+00 5.45778632e-01 -7.89677680e-01
9.37273681e-01 2.40680888e-01 -1.18305072e-01 -4.84830916e-01
-1.88272849e-01 -3.63987088e-02 -1.64100225e-03 1.06585670e+00
7.77262300e-02 -1.69197381e-01 -1.00732990e-01 1.72206163e-01
2.81568915e-02 2.55934685e-01 8.08688998e-01 9.98310208e-01
-4.39990908e-01 -8.98301065e-01 -6.90458179e-01 6.54450119e-01
-1.09607697e+00 4.03582789e-02 -3.06934595e-01 1.01988578e+00
5.37031949e-01 1.15179789e+00 -4.21160683e-02 -4.54300702e-01
-1.07619993e-01 -1.78570837e-01 4.75806355e-01 -1.17505535e-01
-9.60300863e-01 -6.45496845e-01 -2.46321842e-01 -5.65571010e-01
-4.27282274e-01 -4.32128459e-01 -9.73771691e-01 -4.82023567e-01
-2.48856619e-01 -4.07745302e-01 1.07263494e+00 9.80936885e-01
2.99443424e-01 5.78361034e-01 2.78933913e-01 -8.25999320e-01
-5.61514020e-01 -1.05785191e+00 -4.65723783e-01 7.39314020e-01
5.04079938e-01 -8.69032621e-01 -3.15383852e-01 2.81677842e-01] | [12.413532257080078, 1.0247831344604492] |
d8f32c8e-9359-483e-8730-88f7e8ab34eb | multi-scale-relational-graph-convolutional | 2212.08781 | null | https://arxiv.org/abs/2212.08781v1 | https://arxiv.org/pdf/2212.08781v1.pdf | Multi-Scale Relational Graph Convolutional Network for Multiple Instance Learning in Histopathology Images | Graph convolutional neural networks have shown significant potential in natural and histopathology images. However, their use has only been studied in a single magnification or multi-magnification with late fusion. In order to leverage the multi-magnification information and early fusion with graph convolutional networks, we handle different embedding spaces at each magnification by introducing the Multi-Scale Relational Graph Convolutional Network (MS-RGCN) as a multiple instance learning method. We model histopathology image patches and their relation with neighboring patches and patches at other scales (i.e., magnifications) as a graph. To pass the information between different magnification embedding spaces, we define separate message-passing neural networks based on the node and edge type. We experiment on prostate cancer histopathology images to predict the grade groups based on the extracted features from patches. We also compare our MS-RGCN with multiple state-of-the-art methods with evaluations on both source and held-out datasets. Our method outperforms the state-of-the-art on both datasets and especially on the classification of grade groups 2 and 3, which are significant for clinical decisions for patient management. Through an ablation study, we test and show the value of the pertinent design features of the MS-RGCN. | ['Septimiu Salcudean', 'Ali Bashashati', 'Martin Gleave', 'Larry Goldenberg', 'Ladan Fazli', 'Roozbeh Bazargani'] | 2022-12-17 | null | null | null | null | ['multiple-instance-learning'] | ['methodology'] | [ 1.81114912e-01 4.57551211e-01 -6.02265485e-02 -3.91282290e-01
-8.08059752e-01 -2.98908800e-01 3.00672531e-01 7.33745515e-01
-2.17533067e-01 2.29106829e-01 1.25148043e-01 -3.89355779e-01
-2.88743913e-01 -1.14437664e+00 -6.07252955e-01 -6.94825709e-01
-5.88714361e-01 2.04978079e-01 3.04746568e-01 -2.24829942e-01
-6.24085702e-02 7.40987003e-01 -8.60429049e-01 7.58511782e-01
3.84069592e-01 9.28412855e-01 3.44438665e-02 9.08375442e-01
-1.23420022e-01 6.90561593e-01 -4.13066089e-01 -5.78242183e-01
2.48558328e-01 -1.84524909e-01 -7.38103569e-01 -6.75743371e-02
6.49750650e-01 -1.05475731e-01 -5.15135944e-01 9.54731286e-01
5.29311359e-01 -4.99480277e-01 5.96851408e-01 -9.72060859e-01
-8.61593068e-01 6.44029260e-01 -8.07619035e-01 3.95873636e-01
-2.17133671e-01 1.82094187e-01 9.15173650e-01 -5.21183133e-01
7.45256245e-01 9.58399832e-01 1.10323012e+00 2.76001543e-01
-1.31652343e+00 -5.14235020e-01 4.75732982e-03 1.34571776e-01
-1.34805143e+00 1.26901165e-01 9.01806295e-01 -3.96107435e-01
7.77987897e-01 2.65662938e-01 8.06222498e-01 8.71391416e-01
8.89352322e-01 3.24887484e-01 1.10385048e+00 -2.06443012e-01
-6.90511465e-02 -6.12259246e-02 1.66959435e-01 1.24385011e+00
3.07242453e-01 -1.05485417e-01 -2.23023310e-01 -1.92414224e-01
1.09698093e+00 4.38856006e-01 -2.09917605e-01 -1.50127947e-01
-1.33467364e+00 9.90629494e-01 1.31172752e+00 5.11220932e-01
-2.69741476e-01 4.09715772e-01 3.14278722e-01 2.76594639e-01
5.89290798e-01 4.28544790e-01 -1.12405345e-01 6.99343443e-01
-7.43288100e-01 -3.21288779e-02 7.01097012e-01 4.81155097e-01
7.95432568e-01 -4.54615206e-01 -4.30893481e-01 6.17039740e-01
4.20143843e-01 2.39225794e-02 4.01862562e-01 -3.90758008e-01
2.73047805e-01 1.23429370e+00 -7.42999017e-01 -1.32030940e+00
-1.05102837e+00 -8.00170660e-01 -1.43728447e+00 1.64372653e-01
3.15086842e-01 1.45929068e-01 -1.18985975e+00 1.37443435e+00
2.67163247e-01 3.78731906e-01 -7.58167207e-02 6.24921918e-01
1.37871027e+00 2.34484836e-01 5.16210385e-02 3.38877961e-02
1.66754818e+00 -8.91340673e-01 -5.25948048e-01 -2.41843849e-01
1.01621222e+00 -4.06028718e-01 6.47002697e-01 2.72289589e-02
-8.48591685e-01 -4.77804601e-01 -1.07745862e+00 -1.53042167e-01
-7.81976879e-01 3.99584323e-01 9.05152857e-01 3.15216243e-01
-1.42158854e+00 8.76985013e-01 -1.05309272e+00 -3.97817612e-01
6.08264267e-01 5.46804845e-01 -6.18935227e-01 1.13043249e-01
-1.09188688e+00 6.46931052e-01 -2.77235415e-02 3.65423113e-01
-6.22144938e-01 -1.02310431e+00 -8.71127248e-01 6.31941855e-02
-1.48805559e-01 -9.42285895e-01 7.37934291e-01 -4.81990725e-01
-8.69062901e-01 1.05868828e+00 3.20919603e-01 -3.11890692e-01
4.81123507e-01 6.62743986e-01 -3.30589503e-01 4.14161652e-01
-2.98335217e-02 8.20383906e-01 3.91052574e-01 -1.09624755e+00
-4.75858748e-01 -6.08421862e-01 3.30900729e-01 3.27346101e-02
-4.37474549e-01 -3.86259496e-01 -4.14092034e-01 -6.49353385e-01
1.84653148e-01 -9.57790971e-01 -5.70978999e-01 3.13426822e-01
-5.19574583e-01 1.10354543e-01 6.34260416e-01 -7.33339846e-01
1.10388279e+00 -2.16944575e+00 1.69184729e-01 3.44025165e-01
8.01135361e-01 -2.41393849e-01 -1.92257702e-01 1.53982550e-01
-1.71280593e-01 4.38649476e-01 2.46177278e-02 -5.11686563e-01
-3.05444300e-01 1.78279355e-01 3.54500711e-01 6.72648191e-01
4.05369192e-01 1.13554251e+00 -8.60188603e-01 -7.55238056e-01
8.11032727e-02 7.02040136e-01 -5.65176308e-01 1.00995854e-01
3.60642046e-01 2.70075232e-01 -2.92024583e-01 8.54467154e-01
4.74776477e-01 -9.31835949e-01 2.69167334e-01 -8.42911124e-01
3.75116050e-01 -4.06158090e-01 -7.95077205e-01 1.64148998e+00
-3.35086673e-01 4.39188749e-01 2.47188453e-02 -7.05871761e-01
6.37431502e-01 1.12864971e-01 6.09345257e-01 -3.83846879e-01
1.69553742e-01 1.29474670e-01 1.10259615e-01 -4.17514712e-01
3.12904716e-01 -1.04587369e-01 -1.01557709e-01 5.50166182e-02
2.17688337e-01 1.15028232e-01 1.79351479e-01 2.84150153e-01
1.61276257e+00 -5.68327010e-01 3.24720800e-01 -2.47430339e-01
2.80104667e-01 -2.30033413e-01 2.01349929e-01 6.66169643e-01
-2.17482284e-01 8.52421999e-01 8.65337789e-01 -6.22225821e-01
-6.94993019e-01 -9.44888949e-01 -1.46054596e-01 6.84762955e-01
1.86823264e-01 -3.99717689e-01 -4.12222952e-01 -8.87774408e-01
2.94225454e-01 -2.40729898e-01 -1.40700400e+00 -2.64120281e-01
-5.75954080e-01 -1.09296763e+00 5.44836104e-01 7.15174496e-01
2.61098266e-01 -8.45024943e-01 -2.03382388e-01 1.19778365e-01
1.74933165e-01 -9.82777357e-01 -2.90418923e-01 3.30550641e-01
-9.24522638e-01 -1.32430685e+00 -6.59794331e-01 -9.36269522e-01
1.17137313e+00 1.58324286e-01 1.14847434e+00 4.18723106e-01
-7.03198671e-01 2.66798317e-01 -1.84237704e-01 -2.94597179e-01
-5.53271234e-01 3.40293229e-01 -4.96817142e-01 1.58294499e-01
-1.94418684e-01 -6.32372499e-01 -1.00007391e+00 6.67082146e-02
-1.12391245e+00 2.47716844e-01 1.19087243e+00 8.58411014e-01
8.05807233e-01 -2.33354606e-02 2.63490677e-01 -1.15894461e+00
6.38395727e-01 -4.08975899e-01 -2.42573410e-01 4.61570948e-01
-5.47168911e-01 2.54577752e-02 3.08139920e-01 -3.17279100e-01
-3.56273323e-01 -1.71444297e-01 1.46143600e-01 -4.95029718e-01
1.84268564e-01 7.97800422e-01 2.65983850e-01 -6.28547430e-01
7.12797523e-01 -3.29044670e-01 1.64565295e-01 -1.95127577e-01
2.11201042e-01 2.78799564e-01 3.63474011e-01 -4.85664681e-02
6.53683841e-01 5.77306926e-01 5.97869277e-01 -3.85848492e-01
-6.56544805e-01 -3.75782639e-01 -5.82980812e-01 -2.96955496e-01
1.02860224e+00 -7.71757543e-01 -6.68839455e-01 5.16249895e-01
-9.86108720e-01 -3.30738336e-01 -2.04118520e-01 1.30482435e-01
-4.08114903e-02 1.52254179e-01 -1.19611001e+00 -3.69812436e-02
-5.29891968e-01 -1.36705494e+00 1.48298228e+00 3.43841314e-01
1.34349093e-01 -1.38744628e+00 5.61146513e-02 1.06189914e-01
5.02836704e-01 7.19410777e-01 1.18624783e+00 -6.39187098e-01
-5.74278772e-01 -5.22398770e-01 -4.83379275e-01 -1.56201627e-02
3.04373175e-01 4.60039303e-02 -7.81820357e-01 -4.88881499e-01
-5.19408941e-01 -2.38730088e-02 1.07149887e+00 4.81970966e-01
1.44734991e+00 -1.70474648e-01 -6.99955165e-01 8.65497291e-01
1.73729014e+00 -2.80568480e-01 5.41190803e-01 3.64346892e-01
9.72040117e-01 4.14816499e-01 -8.49196911e-02 -4.10723239e-02
4.86475587e-01 2.96435386e-01 6.36233926e-01 -7.99458742e-01
-3.87996048e-01 1.31972566e-01 -1.57812044e-01 7.13231385e-01
-1.02774374e-01 -1.63582899e-02 -1.01242208e+00 4.86821473e-01
-1.71455491e+00 -4.88219231e-01 -2.12157182e-02 1.68106115e+00
7.35989809e-01 2.26167619e-01 -3.03740710e-01 -1.40479729e-01
7.57496238e-01 3.20349693e-01 -2.77028680e-01 -2.36713752e-01
6.59973994e-02 2.26627618e-01 7.06164777e-01 2.17337295e-01
-1.17175198e+00 1.68398961e-01 5.94978189e+00 5.12789547e-01
-1.35921907e+00 1.22637868e-01 1.07221401e+00 -1.67429186e-02
-1.71365187e-01 -2.53443986e-01 -6.33310556e-01 4.09564041e-02
7.87015378e-01 -3.61293666e-02 1.96539819e-01 4.78666812e-01
-1.74946964e-01 2.17431724e-01 -1.20542967e+00 7.33153164e-01
7.76174963e-02 -1.72021759e+00 2.01771200e-01 4.46278185e-01
5.85253358e-01 3.15144621e-02 7.47937188e-02 3.68894249e-01
2.09650442e-01 -1.12009120e+00 4.58031939e-03 5.37757158e-01
8.72439802e-01 -4.62907016e-01 1.11935174e+00 -2.57733077e-01
-1.39402688e+00 -1.04180366e-01 -2.88162291e-01 4.58219469e-01
-2.32153550e-01 6.64685071e-01 -1.00152636e+00 1.00171185e+00
5.70285976e-01 8.33228707e-01 -1.33934593e+00 8.72376084e-01
9.95073095e-02 3.11562300e-01 -1.99394748e-02 1.67020112e-01
2.48212278e-01 2.66409367e-01 1.55311331e-01 1.43608034e+00
4.12542373e-01 -9.20339748e-02 1.62930548e-01 7.12703347e-01
-3.00214499e-01 7.38155320e-02 -4.34568018e-01 -5.85332438e-02
-8.14426020e-02 1.94206369e+00 -1.11443126e+00 -1.51407093e-01
-4.83865172e-01 6.11729383e-01 5.73480904e-01 1.89140737e-01
-7.03977644e-01 -3.30450088e-01 2.82930642e-01 3.91822964e-01
1.48690298e-01 1.41994372e-01 -1.53688505e-01 -9.59902823e-01
-1.67066738e-01 -3.64369124e-01 7.53659606e-01 -6.03906155e-01
-1.78142321e+00 7.54408658e-01 -1.60548538e-01 -1.37078583e+00
1.29480347e-01 -8.41688216e-01 -7.06311464e-01 6.05775177e-01
-1.67109537e+00 -1.76948607e+00 -5.94041109e-01 4.32936758e-01
-5.20004332e-02 -2.80774944e-02 9.36374426e-01 2.07109287e-01
-4.17753667e-01 6.90995276e-01 -2.27271721e-01 6.10671818e-01
8.24858487e-01 -1.49737144e+00 1.47041544e-01 3.21604818e-01
-2.50594199e-01 5.94487846e-01 1.39555529e-01 -6.85174108e-01
-1.44521403e+00 -1.57032120e+00 4.72318649e-01 -1.37859911e-01
8.71630251e-01 -3.79605323e-01 -8.42454851e-01 7.10462868e-01
9.55576748e-02 7.17849791e-01 9.76719916e-01 8.82901773e-02
-2.59246767e-01 -3.87667060e-01 -1.24070168e+00 5.37095308e-01
8.71158004e-01 -4.56646860e-01 -5.08892499e-02 3.82997960e-01
8.43646824e-01 -7.28727162e-01 -1.48114800e+00 8.00424695e-01
4.29551363e-01 -7.78345585e-01 9.85593379e-01 -3.62116218e-01
7.15343118e-01 -1.69457063e-01 -9.20047425e-03 -1.44398272e+00
-7.54915476e-01 6.43945038e-02 5.31451292e-02 9.19264376e-01
5.50635815e-01 -6.27471447e-01 8.48172903e-01 1.76587701e-01
-2.44465470e-01 -1.31656098e+00 -9.97253478e-01 -2.29642510e-01
3.84678394e-02 6.91246167e-02 5.08296013e-01 9.08946693e-01
-1.52263805e-01 1.83184683e-01 2.09477141e-01 5.37656844e-01
6.84720278e-01 1.76942095e-01 4.25911725e-01 -1.21263778e+00
-3.79930347e-01 -6.33986473e-01 -1.12263334e+00 3.37111428e-02
-1.41110912e-01 -1.36227608e+00 -4.84956920e-01 -1.91448736e+00
4.51667935e-01 -2.03023776e-01 -7.69912899e-01 7.19689369e-01
-3.45372230e-01 7.14744151e-01 7.19468594e-02 1.04759708e-01
-5.48924744e-01 2.74761058e-02 1.57869887e+00 -7.22120583e-01
-1.76226464e-03 -3.88668269e-01 -9.40755486e-01 4.62426931e-01
5.58610976e-01 -2.58406878e-01 -1.50354043e-01 -2.18897820e-01
4.19745743e-01 1.30829543e-01 7.42827833e-01 -1.18965101e+00
3.31404418e-01 1.33193910e-01 8.13018799e-01 -4.24533427e-01
1.16292819e-01 -7.27376580e-01 4.23023075e-01 5.88125110e-01
-5.19295335e-01 5.43040298e-02 3.09924424e-01 8.05065811e-01
-7.06334710e-02 2.16649234e-01 5.98711193e-01 -3.29102188e-01
-1.88030481e-01 6.06050968e-01 6.32183701e-02 -5.56008399e-01
9.85072315e-01 -2.07384974e-01 -7.68596351e-01 9.51289013e-03
-1.01833582e+00 2.15910167e-01 3.10562998e-01 3.14036548e-01
6.99469686e-01 -1.50514543e+00 -7.67137408e-01 2.96309382e-01
4.55858618e-01 2.13240668e-01 5.47921360e-01 1.06566417e+00
-6.32511616e-01 2.57983543e-02 -2.41452426e-01 -1.01150501e+00
-1.37975311e+00 3.78312230e-01 6.84117258e-01 -1.19159424e+00
-5.23530483e-01 6.97246552e-01 4.06562954e-01 -4.68981236e-01
1.02962047e-01 -7.74969816e-01 -5.30760944e-01 -3.29570659e-02
3.89524937e-01 -5.70911281e-02 5.52596807e-01 -1.73282340e-01
-3.03708851e-01 4.81984109e-01 -4.08298582e-01 4.04430240e-01
1.50137901e+00 1.99025571e-01 -4.71916527e-01 4.32410598e-01
1.55662501e+00 -2.23846897e-01 -7.96783864e-01 -1.93133980e-01
-3.97960186e-01 -1.36497810e-01 2.47239977e-01 -7.48449922e-01
-1.58545256e+00 5.82348526e-01 1.08271706e+00 4.32077646e-01
1.06659007e+00 1.49367169e-01 4.50297207e-01 1.26388445e-01
2.33179703e-01 -5.01229644e-01 3.08187395e-01 6.99335337e-02
9.16003644e-01 -1.28630757e+00 2.71273404e-01 -5.31100988e-01
-1.31879345e-01 1.36255181e+00 5.26780069e-01 -2.35454202e-01
9.44639683e-01 4.55286831e-01 1.39965817e-01 -8.33897531e-01
-6.50221527e-01 1.58886477e-01 4.15630877e-01 3.16140711e-01
5.90939522e-01 3.57570559e-01 -6.35483563e-02 6.79936111e-01
-5.35075925e-02 8.63853171e-02 5.53608239e-01 9.53567863e-01
5.40671824e-03 -9.09597695e-01 -1.40160829e-01 1.06053400e+00
-5.34390330e-01 -6.69181673e-03 -3.97252262e-01 9.09148037e-01
8.05366486e-02 5.18638194e-01 4.79038730e-02 -5.88262498e-01
3.90104443e-01 -5.43710053e-01 5.41440129e-01 -6.23826385e-01
-1.11556447e+00 5.94780296e-02 -1.13732941e-01 -5.11714995e-01
-3.07899684e-01 -1.97307020e-01 -1.17673206e+00 -1.72358245e-01
-2.09440961e-01 -2.16483206e-01 5.81629336e-01 5.27982771e-01
4.02659744e-01 1.26906943e+00 6.12100840e-01 -9.44816411e-01
-3.52617830e-01 -9.26689088e-01 -7.28007436e-01 4.72764224e-01
4.81548965e-01 -3.91995877e-01 -4.52284396e-01 -1.95049554e-01] | [15.140381813049316, -2.978750705718994] |
d37938db-83d4-40c8-87a6-1d7b62c6058b | extensible-motion-based-identification-of-xr | 2302.07517 | null | https://arxiv.org/abs/2302.07517v4 | https://arxiv.org/pdf/2302.07517v4.pdf | Extensible Motion-based Identification of XR Users using Non-Specific Motion Data | In this paper, we combine the strengths of distance-based and classification-based approaches for the task of identifying extended reality users by their movements. For this we explore an embedding-based model that leverages deep metric learning. We train the model on a dataset of users playing the VR game ``Half-Life: Alyx'' and conduct multiple experiments and analyses using a state of the art classification-based model as baseline. The results show that the embedding-based method 1) is able to identify new users from non-specific movements using only a few minutes of enrollment data, 2) can enroll new users within seconds, while retraining the baseline approach takes almost a day, 3) is more reliable than the baseline approach when only little enrollment data is available, 4) can be used to identify new users from another dataset recorded with different VR devices. Altogether, our solution is a foundation for easily extensible XR user identification systems, applicable to a wide range of user motions. It also paves the way for production-ready models that could be used by XR practitioners without the requirements of expertise, hardware, or data for training deep learning models. | ['Christian Rack', 'Marc Erich Latoschik', 'Andreas Hotho', 'Tamara Fernando', 'Konstantin Kobs'] | 2023-02-15 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [-3.95713039e-02 -2.36156881e-01 -2.60919034e-01 -2.45557547e-01
-9.43063259e-01 -7.60435164e-01 3.66955966e-01 1.08802207e-01
-6.48668408e-01 3.39041501e-01 1.92118529e-02 -4.39542800e-01
-2.55674124e-01 -5.82808733e-01 -3.60702425e-01 -9.12168473e-02
-2.20811173e-01 2.63357997e-01 2.37700760e-01 -5.42019069e-01
2.31383696e-01 7.23735571e-01 -1.64749539e+00 5.96443340e-02
4.45808202e-01 1.05981314e+00 4.68377024e-02 8.24496150e-01
4.50074971e-01 8.50857317e-01 -6.44790530e-01 -8.04460496e-02
4.04820621e-01 -7.78377429e-02 -7.03708887e-01 -3.98838758e-01
7.16916442e-01 -7.95205534e-01 -7.49804020e-01 3.49006593e-01
9.95327294e-01 4.74479795e-01 2.74656624e-01 -1.11335742e+00
-8.65454137e-01 2.22329751e-01 -3.10217142e-01 3.30536515e-01
9.99481142e-01 7.76543841e-02 9.08172727e-01 -5.86391032e-01
5.18197119e-01 6.25547707e-01 9.99188781e-01 7.56685913e-01
-9.99262214e-01 -5.96551359e-01 -1.95199978e-02 1.65226113e-04
-1.53596377e+00 -3.72244805e-01 4.28723514e-01 -6.51977837e-01
8.83623600e-01 7.29929090e-01 6.24535561e-01 1.64485836e+00
-4.00765866e-01 8.83944631e-01 5.06223619e-01 -1.82010695e-01
2.57695675e-01 2.51201808e-01 2.62698531e-01 6.52937770e-01
9.27357823e-02 1.49788424e-01 -3.83206159e-01 -2.57726848e-01
8.11005116e-01 1.33979008e-01 -4.10578579e-01 -3.35773617e-01
-1.24539554e+00 8.14338028e-01 2.44879439e-01 7.67442226e-01
-1.86497092e-01 1.18455470e-01 4.77770239e-01 4.21781421e-01
1.73900589e-01 8.81318331e-01 -6.68149352e-01 -8.90436590e-01
-9.91192281e-01 4.25866246e-01 6.09988511e-01 9.01432514e-01
1.91318586e-01 -1.40650913e-01 3.52924615e-02 6.92052722e-01
1.84894413e-01 1.09914914e-01 7.59096026e-01 -9.76171315e-01
2.51473606e-01 5.39582670e-01 3.66693944e-01 -9.09347117e-01
-5.32842457e-01 -1.55685216e-01 -1.76601112e-01 1.55497700e-01
4.20181900e-01 -2.47541904e-01 -6.25752628e-01 1.63164258e+00
1.36590660e-01 5.03531814e-01 -3.57162595e-01 7.73684502e-01
9.64323282e-01 1.76929116e-01 -4.19141322e-01 4.03820932e-01
1.32516873e+00 -6.64124846e-01 -4.37299997e-01 1.10085215e-02
9.57947969e-01 -4.93618488e-01 1.45721769e+00 5.37796974e-01
-9.52856719e-01 -1.06512618e+00 -1.27984023e+00 -1.65359691e-01
-5.16239285e-01 4.80564564e-01 7.74444818e-01 1.44912219e+00
-1.30963683e+00 8.43402088e-01 -1.04477179e+00 -5.78579426e-01
2.28853896e-01 8.93939555e-01 -5.04097760e-01 9.50142965e-02
-9.44940269e-01 7.39065707e-01 -5.61591499e-02 4.22018543e-02
-3.69285524e-01 -7.12368548e-01 -8.72538269e-01 -9.92563367e-02
1.36976942e-01 -3.64694536e-01 1.42455065e+00 -5.10993242e-01
-1.64067662e+00 6.78647578e-01 1.00009166e-01 1.48961442e-02
6.47231281e-01 -5.64429402e-01 -6.69092834e-01 -1.12374298e-01
-4.21229303e-02 2.70538539e-01 5.42443812e-01 -8.63258421e-01
-6.57874405e-01 -4.45557624e-01 4.33368534e-01 -2.67555788e-02
-6.10579610e-01 -9.38358810e-03 -5.60985208e-01 -4.13058758e-01
-1.70796454e-01 -1.24776495e+00 -5.77724800e-02 -1.19402096e-01
-7.59267434e-02 -1.50900781e-01 8.29736590e-01 -7.07142651e-01
1.31544805e+00 -2.26366973e+00 1.86886847e-01 2.19625190e-01
4.47858393e-01 2.94183612e-01 -3.91786732e-02 4.56146836e-01
-8.97001252e-02 4.90794145e-02 2.43240565e-01 -5.12763083e-01
1.10036649e-01 -2.56157398e-01 -1.88298330e-01 4.84424829e-01
-3.41122240e-01 6.76957726e-01 -8.99978757e-01 1.54822424e-01
7.91144073e-01 6.91850305e-01 -5.31302333e-01 2.89470911e-01
4.98257995e-01 5.86716950e-01 -2.57778257e-01 5.09206414e-01
4.13486719e-01 -1.03702858e-01 1.74620703e-01 -1.21339932e-01
-2.25197133e-02 2.62780068e-03 -1.46627593e+00 1.94660723e+00
-7.43731916e-01 1.16265953e+00 -4.59844828e-01 -7.04708397e-01
7.93622077e-01 3.25941682e-01 7.46633053e-01 -4.85182285e-01
3.82363290e-01 1.74241602e-01 -2.45648339e-01 -6.38933778e-01
8.05811286e-01 5.97725630e-01 -2.79839605e-01 5.75612307e-01
6.86987862e-02 3.93959254e-01 -1.22935154e-01 4.02234588e-03
1.50769818e+00 1.70613676e-01 -4.23958991e-03 2.64695376e-01
4.31892544e-01 -5.01897156e-01 7.37721995e-02 8.80263627e-01
-3.41542691e-01 7.47627914e-01 -5.99050447e-02 -5.76843917e-01
-7.35432923e-01 -1.06011617e+00 -1.29878595e-02 1.44309962e+00
1.77024692e-01 -5.07294297e-01 -5.46467245e-01 -6.15291655e-01
-2.07391044e-04 4.86871123e-01 -8.60233128e-01 -1.83982089e-01
-6.24907911e-01 -4.65691924e-01 7.80379713e-01 9.76291239e-01
1.77457601e-01 -7.10065424e-01 -7.59917021e-01 2.62837000e-02
-1.86898261e-01 -1.18825018e+00 -5.65998852e-01 -3.58057469e-01
-6.12828076e-01 -1.05326056e+00 -7.88930893e-01 -7.02360094e-01
1.85038507e-01 2.38471195e-01 8.70728731e-01 2.15843916e-01
-3.61323178e-01 8.89963150e-01 -5.41979194e-01 -5.31274118e-02
1.44060314e-01 4.36177641e-01 4.63006228e-01 -4.79466319e-02
7.43382454e-01 -4.07137066e-01 -8.11397076e-01 3.62190515e-01
-4.71514255e-01 -5.52893639e-01 1.60000309e-01 2.68815219e-01
-8.51505324e-02 -2.29942858e-01 4.20305490e-01 -7.17462838e-01
8.54164720e-01 -4.66664881e-01 -3.64463657e-01 6.99741254e-03
-4.23743129e-01 -2.71123827e-01 3.47223192e-01 -7.92837381e-01
-5.08995712e-01 7.84693807e-02 -4.62222487e-01 -2.75909334e-01
-2.36701936e-01 9.44564193e-02 5.01228981e-02 -1.89053595e-01
9.71236110e-01 -4.00631800e-02 -4.45490539e-01 -6.07367754e-01
4.27982539e-01 1.23340416e+00 5.87042451e-01 -3.87667209e-01
5.04269719e-01 4.57877338e-01 -5.24462163e-01 -9.77137983e-01
-1.13066971e-01 -8.18880737e-01 -7.96535134e-01 -2.11610228e-01
9.34954703e-01 -9.39188659e-01 -1.17635322e+00 2.90349513e-01
-8.59136462e-01 -4.63538378e-01 -2.20544443e-01 8.45768452e-01
-5.57097733e-01 4.02784616e-01 -6.13478661e-01 -9.38545942e-01
6.96295649e-02 -1.12545347e+00 1.16528845e+00 2.51013190e-01
-7.58916497e-01 -9.26364481e-01 3.84665042e-01 4.64179695e-01
4.07296360e-01 4.74855393e-01 1.97863072e-01 -7.13414788e-01
-3.93922836e-01 -8.54620576e-01 1.25018239e-01 5.07437438e-02
3.91530424e-01 5.77720739e-02 -1.09798837e+00 -5.62352479e-01
-1.54759690e-01 -3.53303462e-01 1.92199305e-01 2.59274989e-01
1.29932976e+00 3.08296651e-01 -4.77186531e-01 5.94969392e-01
1.16872668e+00 2.95742631e-01 7.45187402e-01 4.31936800e-01
9.75266397e-01 3.05815339e-01 4.10509557e-01 4.64171737e-01
6.27757967e-01 1.20088649e+00 1.49707809e-01 -1.70229539e-01
4.24285680e-01 -3.27297710e-02 1.17655955e-01 5.10255158e-01
-5.87913752e-01 -1.19325392e-01 -7.43359923e-01 5.19836307e-01
-1.90193081e+00 -1.15026307e+00 2.69185696e-02 2.37569284e+00
2.28617892e-01 -8.90756622e-02 6.83338940e-01 3.01813960e-01
3.61851037e-01 4.10569012e-02 -4.60731447e-01 -3.41254026e-01
6.47854328e-01 6.10551536e-01 4.70047593e-01 3.46030802e-01
-1.00908709e+00 5.51943600e-01 6.59292364e+00 3.28157544e-01
-1.10640466e+00 2.68438011e-01 6.91875964e-02 -3.09374094e-01
1.48784697e-01 -4.42264497e-01 -5.95571339e-01 6.13641739e-01
1.19871533e+00 4.89377439e-01 5.80264390e-01 1.08430183e+00
1.69548556e-01 1.01828456e-01 -1.56924617e+00 1.54693925e+00
1.85597375e-01 -1.17175925e+00 -6.05052412e-01 3.25423449e-01
2.42184922e-01 1.55917138e-01 2.46989787e-01 7.27918506e-01
1.12027191e-01 -1.06339490e+00 5.51271319e-01 4.93693888e-01
9.32454884e-01 -7.00907230e-01 7.08199799e-01 2.64831156e-01
-1.33995247e+00 -2.86646992e-01 -2.98774485e-02 -4.34960648e-02
1.89743161e-01 -2.88019359e-01 -4.37574804e-01 4.32088345e-01
9.81429338e-01 5.08432090e-01 -7.34173775e-01 8.48948777e-01
3.34135205e-01 2.72802234e-01 -1.86721087e-01 1.95225388e-01
1.17375869e-02 -4.81365249e-02 3.47863231e-03 8.93081069e-01
5.85218191e-01 1.09972447e-01 1.53953284e-02 1.97890744e-01
-2.94772964e-02 -3.67598422e-02 -8.21260512e-01 -1.26863241e-01
5.64410925e-01 1.05764031e+00 -6.51568413e-01 -8.90125036e-02
-5.85440636e-01 1.44631982e+00 2.22202331e-01 1.48323223e-01
-1.13985837e+00 -7.02986240e-01 7.40012228e-01 2.37375110e-01
3.88036698e-01 -5.06733298e-01 -8.34715590e-02 -1.19310367e+00
1.06457970e-03 -9.67841029e-01 2.21916169e-01 -5.98901212e-01
-9.31701899e-01 8.61305654e-01 -1.53718308e-01 -1.46461606e+00
-6.64688706e-01 -8.45048249e-01 -3.13969910e-01 6.77934766e-01
-8.69900405e-01 -7.76704431e-01 -6.69413030e-01 5.68729043e-01
2.69679040e-01 -4.43154961e-01 1.13132858e+00 7.85014987e-01
-5.53299725e-01 1.04880226e+00 2.09266514e-01 3.82423609e-01
5.45995414e-01 -1.29360712e+00 6.85343921e-01 4.19619709e-01
5.26303649e-01 1.04577267e+00 6.17890120e-01 -1.54355556e-01
-1.35639441e+00 -6.84101760e-01 5.65712452e-01 -1.54575217e+00
3.03429544e-01 -5.93939066e-01 -6.17919445e-01 1.06719220e+00
-1.97128177e-01 5.71255721e-02 1.50086284e+00 5.64525127e-01
-2.15963021e-01 2.05736123e-02 -1.15855086e+00 5.80674767e-01
1.23131859e+00 -9.36064482e-01 -4.84510750e-01 2.68002898e-01
4.87751037e-01 -7.48470843e-01 -1.15160453e+00 1.70396209e-01
1.02068162e+00 -1.00727427e+00 1.22796202e+00 -5.56669295e-01
-1.96473941e-01 -1.59305632e-01 -3.68380606e-01 -8.89546216e-01
-3.39374959e-01 -7.71766961e-01 -4.95702505e-01 7.42624283e-01
3.45897377e-01 -4.05948728e-01 8.43060195e-01 9.03666973e-01
1.12328045e-01 -7.72280455e-01 -7.13098764e-01 -8.30922067e-01
-4.13128197e-01 -9.18395579e-01 9.58450556e-01 1.13028836e+00
1.45334318e-01 3.04070920e-01 -5.89103103e-01 2.11789384e-01
5.67125250e-03 -3.29630136e-01 1.13609064e+00 -1.42086291e+00
-7.65138745e-01 -2.66770691e-01 -8.23692203e-01 -1.14820004e+00
-1.16876379e-01 -4.86639827e-01 -3.97976339e-01 -1.38416600e+00
-3.23298067e-01 -5.90446949e-01 -3.36743474e-01 3.68971735e-01
1.44389510e-01 6.41094387e-01 2.31482923e-01 1.02224067e-01
-4.44936693e-01 2.99689502e-01 5.51331937e-01 1.27304867e-01
-7.92077363e-01 3.51505429e-01 -9.59120631e-01 4.67742115e-01
5.93277752e-01 -1.92614615e-01 -7.35278010e-01 -4.06565875e-01
3.08283955e-01 -1.64047539e-01 1.13377735e-01 -1.33571863e+00
-9.41520110e-02 3.34629476e-01 4.96883631e-01 -3.13582778e-01
7.64655471e-01 -8.19916487e-01 1.26100212e-01 2.76689261e-01
-7.20246956e-02 3.64524722e-01 2.87352353e-01 4.53854680e-01
3.69356871e-01 -8.61358047e-02 2.79992908e-01 1.04744151e-01
-8.16559970e-01 3.93222421e-01 -4.33670253e-01 -8.38435516e-02
1.19304883e+00 -6.10684752e-01 1.52314320e-01 -4.84324127e-01
-9.07474101e-01 -1.03447117e-01 5.66295505e-01 9.20154095e-01
5.01140416e-01 -1.51272547e+00 -1.95781484e-01 2.21293777e-01
2.85337985e-01 -5.65300465e-01 1.68130487e-01 4.34078932e-01
-6.78991616e-01 3.05674940e-01 -1.61210880e-01 -7.29154170e-01
-1.31635690e+00 4.23377782e-01 2.69174963e-01 9.19431895e-02
-6.67319775e-01 7.13711619e-01 -4.61657554e-01 -9.02838528e-01
3.36432606e-01 -4.94996101e-01 -4.43521887e-01 1.33320212e-01
8.96429837e-01 6.30386770e-01 1.47599250e-01 -8.82671416e-01
-5.24139345e-01 4.55272585e-01 -1.16647638e-01 -4.41194028e-01
1.20172310e+00 -1.61949135e-02 6.94499195e-01 7.97624230e-01
1.32448924e+00 2.17977598e-01 -1.10622501e+00 2.57633645e-02
-1.42581433e-01 -8.66641164e-01 1.58959314e-01 -4.06604320e-01
-8.61862421e-01 7.66867518e-01 1.21948457e+00 2.55966336e-01
8.16844881e-01 -1.26620233e-01 1.04919827e+00 3.39574099e-01
6.02761269e-01 -1.00614321e+00 3.00125241e-01 1.80671915e-01
3.34356785e-01 -1.22166979e+00 -1.82932794e-01 9.08287093e-02
-6.16681218e-01 9.28620577e-01 5.70749462e-01 -4.52190675e-02
7.07386494e-01 -6.03738204e-02 1.89375579e-01 -2.43442819e-01
-9.39795673e-02 -2.59304404e-01 4.21516478e-01 8.78524840e-01
6.09086514e-01 2.30417587e-02 1.22267164e-01 6.67346895e-01
-4.67652023e-01 1.92400113e-01 5.68357170e-01 1.02825320e+00
1.48966923e-01 -1.20652461e+00 -3.79183367e-02 3.84563833e-01
-2.86619902e-01 2.31427640e-01 -2.42036477e-01 1.04370546e+00
2.81415820e-01 9.13401961e-01 9.76598039e-02 -9.61745322e-01
6.89102471e-01 -2.87244648e-01 8.03097188e-01 -8.39532733e-01
-6.50515735e-01 -6.11982703e-01 -1.97601214e-01 -8.53107095e-01
-1.58467337e-01 -7.14505196e-01 -9.19358015e-01 -6.39769197e-01
-4.13862348e-01 -1.94113478e-01 6.69766307e-01 8.67625713e-01
5.23498714e-01 4.98198688e-01 6.23247087e-01 -1.06608129e+00
-2.95749635e-01 -7.92897344e-01 -6.10741973e-01 2.71828592e-01
4.78591025e-01 -6.69396400e-01 -1.20673500e-01 -4.56002951e-01] | [7.5521240234375, 0.6816166043281555] |
84cb7969-bc28-49c6-a8bd-a37cd674d260 | ensemble-clustering-based-on-evidence | null | null | https://www.sciencedirect.com/science/article/pii/S0031320319301281 | https://www.sciencedirect.com/science/article/pii/S0031320319301281 | Ensemble clustering based on evidence extracted from the co-association matrix | The evidence accumulation model is an approach for collecting the information of base partitions in a clustering ensemble method, and can be viewed as a kernel transformation from the original data space to a co-association matrix. However, cluster structure information may be partially lost in this transformation; hence, some methods proposed in the literature try to find the lost information and return it to the ensemble process. In this paper, an interesting phenomenon is introduced: remove some evidences from the co-association matrix, which can result in more accurate clustering results. The intuitive explanation for this is that some evidences in the original co-association matrix could be noise, with negative effects on the final clustering. However, it is difficult to detect those evidences practically, let alone remove them from the matrix. To remedy this problem, we remove multiple level evidences having low occurrence frequencies, because negative evidences do not normally occur regularly in the base partitions. Subsequently, we use normalized cut to achieve multiple clustering results. To discriminate the optimal ensemble result, an internal validity index, which uses only the co-association matrix, is specially designed for the clustering ensemble. The experimental results on 16 datasets demonstrate that the proposed scheme outperforms some state-of-the-art clustering ensemble approaches. | [] | 2019-08-01 | null | null | null | null | ['spectral-graph-clustering', 'clustering-ensemble'] | ['graphs', 'graphs'] | [ 5.51315024e-02 -2.57763058e-01 6.25222083e-03 -5.20661734e-02
-5.89324795e-02 -3.16655248e-01 8.01566690e-02 4.38513458e-01
-3.38283569e-01 7.79930949e-01 -1.99231520e-01 9.29395482e-02
-6.44826412e-01 -9.42319930e-01 -2.61627376e-01 -1.33418036e+00
-8.18829611e-02 2.21285850e-01 2.32221708e-01 -2.62686256e-02
3.59445184e-01 1.88117325e-01 -1.90854764e+00 2.70979226e-01
1.56344235e+00 6.99337244e-01 1.02358595e-01 -7.00135529e-02
-1.89540893e-01 5.38728893e-01 -9.33141828e-01 -6.68462440e-02
3.99929695e-02 -6.47904098e-01 -3.81542832e-01 1.45153925e-01
-4.08192873e-01 2.21968919e-01 2.35697716e-01 1.43898964e+00
2.58098245e-01 7.90659711e-02 8.38733733e-01 -1.48411477e+00
-2.78238326e-01 6.57440484e-01 -8.76245916e-01 -3.41843851e-02
1.38391629e-01 -1.72319084e-01 7.57897258e-01 -8.87249231e-01
3.30695063e-01 1.05984426e+00 3.64545703e-01 -2.44791787e-02
-1.14661884e+00 -8.54152203e-01 7.75844455e-02 5.74505210e-01
-1.86090219e+00 -1.35257542e-01 8.77538383e-01 -3.46928418e-01
2.36086562e-01 5.43437302e-01 7.73260832e-01 5.46178579e-01
2.84878671e-01 5.41669607e-01 1.48597431e+00 -4.22032744e-01
2.34714597e-01 3.54460359e-01 2.75401026e-01 3.68833959e-01
7.76451290e-01 -2.51294047e-01 -2.81012684e-01 -1.39318898e-01
2.09618241e-01 1.45101219e-01 -6.17660642e-01 -3.34094226e-01
-1.02150536e+00 3.93777549e-01 3.69639397e-01 7.59721696e-01
-5.04668713e-01 -6.79343343e-01 2.13650718e-01 1.73026562e-01
2.43358657e-01 -1.24263084e-02 -1.94305897e-01 3.00062537e-01
-8.71586740e-01 3.19753811e-02 5.83671749e-01 3.79384249e-01
9.96146798e-01 -2.71433562e-01 5.32211736e-02 8.02010536e-01
4.16387677e-01 4.52231348e-01 5.97346187e-01 -5.40136218e-01
3.68007213e-01 1.12805295e+00 -3.98655124e-02 -1.51012242e+00
-2.72804141e-01 -6.37197733e-01 -1.31207919e+00 7.05901608e-02
3.31812322e-01 -3.33305076e-02 -5.22470713e-01 1.43668771e+00
5.33119857e-01 4.58048373e-01 1.06458105e-01 9.55111980e-01
7.18462408e-01 6.25068903e-01 -8.31860211e-03 -8.54929984e-01
1.26307666e+00 -4.57253307e-01 -9.68645751e-01 2.32841730e-01
2.37876773e-01 -7.82774806e-01 5.94671011e-01 7.60156631e-01
-6.45297945e-01 -6.52176797e-01 -1.28090608e+00 7.78561592e-01
-3.48098814e-01 3.11336637e-01 3.57822508e-01 5.14107049e-01
-5.94811141e-01 5.16116500e-01 -4.76594031e-01 -2.03755885e-01
-6.41361484e-03 2.13168010e-01 -3.31631571e-01 -2.11893529e-01
-1.23685694e+00 7.52418578e-01 8.83501232e-01 3.92285019e-01
-8.55757913e-04 -2.04005569e-01 -3.02069634e-01 8.06980804e-02
6.84549987e-01 -3.43037993e-01 3.46721053e-01 -9.12861466e-01
-8.53615403e-01 2.61726677e-01 -2.52469540e-01 1.40777186e-01
2.30486810e-01 1.65998787e-01 -9.15855587e-01 7.04782754e-02
1.80711195e-01 -6.58688620e-02 6.73321247e-01 -1.85070944e+00
-7.15693176e-01 -5.42999446e-01 -5.46078324e-01 1.48758084e-01
-5.67382932e-01 -3.39412540e-01 -4.62846458e-01 -5.53614199e-01
7.37753272e-01 -8.45125020e-01 -3.03150922e-01 -8.44389737e-01
-5.04545510e-01 -3.21501285e-01 9.32803273e-01 -6.97019815e-01
1.70571673e+00 -2.10447645e+00 1.63050622e-01 9.37093973e-01
1.80694044e-01 2.29092896e-01 3.10216337e-01 3.11061561e-01
-2.09871456e-01 2.14071363e-01 -5.36053836e-01 2.55842894e-01
-4.04998362e-01 2.00096235e-01 3.23418714e-02 2.84351051e-01
-4.30861302e-02 1.51137382e-01 -8.82951081e-01 -8.88867378e-01
2.51133502e-01 2.55173087e-01 -1.48287550e-01 -3.64553183e-02
2.17702940e-01 4.57794875e-01 -4.19701368e-01 5.03737509e-01
1.16255617e+00 1.58122048e-01 6.07660472e-01 -4.52245563e-01
-2.48882055e-01 -2.50546366e-01 -1.84453261e+00 1.00900948e+00
1.26640484e-01 1.00521281e-01 1.88786194e-01 -1.34043849e+00
1.19631743e+00 3.96180302e-01 8.55489135e-01 -2.86705315e-01
2.96252191e-01 3.07173580e-01 4.37710047e-01 -5.50165176e-01
4.30285186e-01 -1.52738139e-01 2.04427391e-01 3.50636601e-01
-3.36257786e-01 3.87396157e-01 5.93954623e-01 1.08704500e-01
6.98217869e-01 -3.01709741e-01 3.46642733e-01 -3.37474614e-01
1.09216619e+00 4.11508605e-03 1.01052904e+00 5.10168552e-01
1.30514905e-01 2.46646717e-01 3.28039765e-01 -1.05869822e-01
-7.06474900e-01 -9.55541611e-01 -2.51140326e-01 3.15096527e-01
5.40793955e-01 -6.22648120e-01 -8.21776271e-01 -6.48124278e-01
-1.59204349e-01 5.14120340e-01 -3.54736358e-01 -3.41752470e-01
-2.64047563e-01 -1.23366845e+00 3.23082238e-01 1.67413473e-01
8.21123719e-01 -9.80965912e-01 -2.41892233e-01 4.07032847e-01
-5.95914423e-01 -4.65300024e-01 5.53142391e-02 -4.93028499e-02
-9.99833584e-01 -1.47170150e+00 -2.04768047e-01 -4.48058397e-01
8.92630875e-01 6.26168191e-01 6.48613572e-01 7.84070969e-01
6.23227023e-02 -9.53894928e-02 -6.45278454e-01 -4.25483465e-01
-5.15040338e-01 -8.29689726e-02 2.25012243e-01 4.11402047e-01
6.43063426e-01 -6.40530050e-01 -4.79293972e-01 5.62994003e-01
-1.04936230e+00 -2.78781354e-01 6.74605131e-01 8.74038577e-01
4.32391226e-01 1.25146139e+00 7.29211390e-01 -5.21459877e-01
7.81634510e-01 -5.28209150e-01 -2.61616975e-01 3.87458652e-01
-8.26662719e-01 -6.35750890e-02 7.74398685e-01 -2.38215715e-01
-1.19706166e+00 -2.37632245e-01 2.06030101e-01 -3.65899146e-01
-2.02881664e-01 8.08091521e-01 -6.16277039e-01 3.29637349e-01
3.28901142e-01 4.59169477e-01 1.07614629e-01 -3.71981889e-01
-2.50065066e-02 7.55342066e-01 3.88456404e-01 -5.72812855e-01
8.37495863e-01 4.60491836e-01 1.62264600e-01 -7.42113888e-01
-4.69463766e-01 -5.80828011e-01 -8.08590770e-01 -4.90763277e-01
6.03163660e-01 -5.75601578e-01 -7.26596117e-01 2.63582736e-01
-9.47955847e-01 6.28040075e-01 9.97940730e-03 7.23299146e-01
3.44753265e-02 6.33223116e-01 -2.42819846e-01 -1.05327082e+00
-5.34435548e-02 -1.05353081e+00 3.24023753e-01 3.92168701e-01
-1.80426076e-01 -6.76355660e-01 -4.91713546e-02 1.54120460e-01
-1.19438060e-01 1.93957075e-01 9.27760661e-01 -6.03839040e-01
-3.84915233e-01 -1.54633924e-01 1.26942499e-02 4.94886637e-01
2.95124143e-01 5.02331018e-01 -6.84415638e-01 -2.18009457e-01
1.88044265e-01 3.39164913e-01 1.01967287e+00 1.64881125e-01
1.14867687e+00 -2.07562551e-01 -6.97926342e-01 2.14741915e-01
1.35529602e+00 6.19880736e-01 8.70715082e-01 3.93290758e-01
5.58167100e-01 7.63673723e-01 1.04494500e+00 4.33462620e-01
1.84413880e-01 4.45884854e-01 2.15111926e-01 -9.28559378e-02
4.00528312e-01 -2.86232401e-03 1.31823704e-01 1.27696347e+00
-7.36329019e-01 -2.42398277e-01 -6.74751759e-01 2.24562034e-01
-2.13655686e+00 -1.19311643e+00 -8.28325391e-01 2.38968658e+00
6.78683221e-01 1.34839341e-01 -2.79060882e-02 9.25487757e-01
1.10587144e+00 -1.74438193e-01 -1.92443728e-01 -1.63826663e-02
-3.58823985e-01 -1.33176818e-01 1.49791867e-01 1.85894907e-01
-1.00121188e+00 3.50127816e-01 5.47546005e+00 1.07014406e+00
-9.28795695e-01 -3.89462523e-02 3.58288735e-01 2.64730126e-01
-3.23093265e-01 2.11057931e-01 -4.74399149e-01 8.34258497e-01
3.82389545e-01 -2.31033057e-01 8.43982100e-02 4.66206789e-01
3.42568427e-01 -4.21321034e-01 -5.68290889e-01 8.07961345e-01
-3.73676196e-02 -5.99548101e-01 -6.04007393e-02 3.59367490e-01
8.00831378e-01 -8.08050156e-01 -2.63479382e-01 1.03099838e-01
2.32210755e-01 -7.06665158e-01 3.20636570e-01 7.82867551e-01
3.09701622e-01 -1.21278691e+00 1.19468009e+00 6.99232101e-01
-1.34399366e+00 -1.92813963e-01 -5.36019385e-01 -1.61787169e-03
-3.15162204e-02 1.38516593e+00 -6.70799613e-01 1.37029207e+00
8.67027760e-01 7.02215314e-01 -6.47785068e-01 1.24409997e+00
-2.67091393e-01 6.29422665e-01 -4.09873605e-01 1.35990739e-01
-1.91456944e-01 -9.09542382e-01 7.80896187e-01 8.99852693e-01
7.10158706e-01 3.62781048e-01 2.06894174e-01 7.49583900e-01
1.68090343e-01 4.97183830e-01 -5.50404608e-01 4.61700082e-01
6.98988259e-01 1.29772627e+00 -1.14178240e+00 -5.65199077e-01
-1.00991182e-01 7.20215142e-01 -3.42138223e-02 3.24713409e-01
-7.60906518e-01 -2.97209859e-01 2.22173959e-01 3.98638379e-03
7.44293258e-02 2.42709145e-02 -3.68028522e-01 -1.15239882e+00
1.48636147e-01 -9.64675784e-01 5.86799681e-01 -4.98605937e-01
-1.28687978e+00 3.04397613e-01 1.08337887e-01 -1.61159706e+00
1.25935748e-01 -2.04870120e-01 -7.86483169e-01 6.19065285e-01
-8.51884782e-01 -5.75489819e-01 -4.73904878e-01 5.93478799e-01
8.08366463e-02 -6.97750449e-02 6.09562814e-01 3.96694511e-01
-8.63717139e-01 3.92214149e-01 4.99316871e-01 -1.41499564e-01
8.02293122e-01 -1.03066432e+00 -8.41259718e-01 9.90608454e-01
-4.08538245e-02 9.17947888e-01 6.16397202e-01 -9.78811979e-01
-7.93784142e-01 -7.83175766e-01 6.21101558e-01 -1.77890718e-01
2.52991289e-01 7.89141059e-02 -1.24362695e+00 1.80157080e-01
1.64198130e-01 -4.56185251e-01 7.92810738e-01 1.97814494e-01
1.28938705e-02 -4.03263569e-01 -1.21039093e+00 6.12922609e-01
7.04738855e-01 9.46483612e-02 -6.34884953e-01 -2.20850199e-01
1.77821070e-01 2.86608428e-01 -9.65450644e-01 7.36474276e-01
5.64203560e-01 -1.27226591e+00 9.22646642e-01 -4.80395257e-02
2.54776120e-01 -1.12509477e+00 1.64048940e-01 -1.61246467e+00
-4.66817975e-01 1.01323508e-01 6.93880320e-02 1.77553988e+00
1.13753863e-01 -6.82231486e-01 4.21460658e-01 2.62677968e-02
1.67852387e-01 -5.30342817e-01 -7.76124895e-01 -7.28439569e-01
-3.16319168e-01 -1.37870476e-01 8.34149301e-01 1.33769572e+00
1.85382187e-01 3.17976296e-01 -2.19666570e-01 2.56239593e-01
8.72174501e-01 1.55644730e-01 5.92876732e-01 -1.71876872e+00
1.28139958e-01 -4.45443064e-01 -2.60088801e-01 -3.75813931e-01
-9.05479193e-02 -7.72301674e-01 5.18034808e-02 -1.47787476e+00
3.28399777e-01 -7.26484239e-01 -6.15111411e-01 2.37863109e-01
-5.42337298e-01 1.78873420e-01 1.19199537e-01 6.76189125e-01
-4.66314107e-01 5.01728117e-01 1.23654723e+00 -1.74873471e-01
-3.90945077e-01 -4.43592817e-02 -7.28956401e-01 8.44392717e-01
8.40799809e-01 -5.63315034e-01 -4.07769829e-01 3.79727662e-01
-4.25216407e-02 -1.75933674e-01 8.10466260e-02 -1.18975270e+00
3.90902072e-01 -2.51234531e-01 6.27216697e-01 -1.00911891e+00
-6.85414970e-02 -1.24949813e+00 6.05117619e-01 4.77519304e-01
3.01800966e-01 -1.22896306e-01 -4.53404337e-02 5.04762530e-01
-4.79064941e-01 -4.04140651e-01 5.31165361e-01 -6.17874041e-02
-4.93122399e-01 -1.36803225e-01 -5.74408293e-01 -5.36443949e-01
1.13388264e+00 -5.62048972e-01 -1.79929093e-01 -1.78655181e-02
-8.65997672e-01 3.85054439e-01 5.47240198e-01 2.01474547e-01
5.88642776e-01 -1.58057189e+00 -7.92997599e-01 1.00913420e-01
-8.35192110e-03 1.22322768e-01 5.20361483e-01 1.18503308e+00
-2.14967102e-01 1.52616709e-01 -1.47175699e-01 -7.16696560e-01
-1.47355509e+00 6.51611030e-01 -2.85575651e-02 -3.73967797e-01
-2.58497834e-01 1.32716700e-01 -6.66112229e-02 -2.85450399e-01
-2.34408036e-01 1.28030568e-01 -7.39610970e-01 4.88824427e-01
4.42074537e-01 7.42462099e-01 2.18852013e-01 -6.82897508e-01
-5.40247142e-01 6.91225290e-01 1.78812355e-01 8.27402100e-02
1.15773618e+00 -2.48069048e-01 -7.96828330e-01 4.66511846e-01
8.14154565e-01 3.29802066e-01 -5.78505754e-01 1.10921087e-02
2.97976658e-02 -5.09279907e-01 -2.20732242e-01 -5.30797839e-01
-1.01260066e+00 4.96595174e-01 4.52415824e-01 6.14461005e-01
1.45608544e+00 -1.80337861e-01 3.11600029e-01 3.04832458e-01
3.08786601e-01 -1.35147846e+00 -1.40435129e-01 1.25142127e-01
6.08489513e-01 -9.84528482e-01 2.88813591e-01 -8.98211598e-01
-4.21342552e-01 1.10602152e+00 7.80424178e-01 7.93666616e-02
6.92803800e-01 9.30420458e-02 -1.84670925e-01 -1.84307009e-01
-4.26950693e-01 -1.58621252e-01 1.99176013e-01 4.88571733e-01
2.46135488e-01 1.65689975e-01 -1.03083062e+00 9.04956579e-01
-1.08315989e-01 -1.58863664e-01 5.13503850e-01 6.54035449e-01
-6.30091429e-01 -1.20231414e+00 -1.12677073e+00 5.71217835e-01
-2.60299802e-01 2.99113333e-01 -3.54798168e-01 6.59181535e-01
7.37184048e-01 1.42113161e+00 -3.02519333e-02 -7.01921284e-01
4.04002458e-01 2.55093426e-01 2.30039433e-01 -1.66407824e-01
-2.77451158e-01 2.40928903e-01 -1.95522234e-01 -1.53894469e-01
-7.51300871e-01 -5.79188645e-01 -1.35713816e+00 -4.22120392e-01
-7.81154454e-01 8.01208973e-01 3.69400710e-01 8.73495698e-01
2.34817490e-02 7.16874301e-01 7.85807431e-01 -4.50010061e-01
-1.01710921e-02 -1.05916023e+00 -7.50374496e-01 6.19983613e-01
-2.14744151e-01 -9.77618694e-01 -6.98280931e-01 4.74448651e-02] | [7.6503167152404785, 4.557351112365723] |
4598ca7f-c667-474e-a1c7-c5e3c511d9cd | gnep-based-dynamic-segmentation-and-motion | 2307.02595 | null | https://arxiv.org/abs/2307.02595v2 | https://arxiv.org/pdf/2307.02595v2.pdf | GNEP Based Dynamic Segmentation and Motion Estimation for Neuromorphic Imaging | This paper explores the application of event-based cameras in the domains of image segmentation and motion estimation. These cameras offer a groundbreaking technology by capturing visual information as a continuous stream of asynchronous events, departing from the conventional frame-based image acquisition. We introduce a Generalized Nash Equilibrium based framework that leverages the temporal and spatial information derived from the event stream to carry out segmentation and velocity estimation. To establish the theoretical foundations, we derive an existence criteria and propose a multi-level optimization method for calculating equilibrium. The efficacy of this approach is shown through a series of experiments. | ['David Sayre', 'Harbir Antil'] | 2023-07-05 | null | null | null | null | ['motion-estimation'] | ['computer-vision'] | [ 3.66687745e-01 -2.79517323e-01 -3.82739365e-01 -2.38956232e-02
-4.07907933e-01 -7.21319914e-01 6.35098755e-01 1.52199911e-02
-8.29191566e-01 5.73501825e-01 -1.36352956e-01 -2.40513429e-01
-3.08039576e-01 -4.97752696e-01 -5.50782442e-01 -5.84590733e-01
-3.46136838e-01 -1.56141967e-01 5.09440839e-01 1.97547689e-01
3.95845622e-01 5.67276418e-01 -1.33903348e+00 -1.52044162e-01
5.49623072e-01 9.27525520e-01 7.59638250e-02 1.27495456e+00
7.54997581e-02 1.12915480e+00 -5.11844814e-01 -2.10125804e-01
4.43341821e-01 -6.77107036e-01 -6.52139187e-01 6.02303982e-01
-1.05499573e-01 -6.78900778e-01 -3.73610556e-01 8.07890773e-01
2.20568806e-01 2.89380223e-01 3.61447304e-01 -1.58576298e+00
1.21679775e-01 1.06203385e-01 -6.51576936e-01 7.99157858e-01
6.10248268e-01 1.57686234e-01 7.52512991e-01 -4.57909316e-01
9.83598411e-01 8.24691772e-01 4.26835686e-01 4.80485827e-01
-9.40472186e-01 -6.89496174e-02 2.23670319e-01 2.74982512e-01
-1.17637920e+00 -3.88376594e-01 9.24059868e-01 -4.95791912e-01
6.31672800e-01 2.60755986e-01 1.07032442e+00 8.71265888e-01
3.01649123e-01 1.05169582e+00 9.37795103e-01 -5.49439013e-01
4.46606964e-01 -4.01305705e-01 -7.01058060e-02 4.31666493e-01
1.24319524e-01 4.21322465e-01 -8.73679698e-01 -7.38122240e-02
1.22293484e+00 -6.14974126e-02 -3.69414777e-01 -3.80006254e-01
-1.33339560e+00 6.66808665e-01 5.61891906e-02 1.70534141e-02
-6.53307498e-01 5.25673211e-01 2.91829795e-01 7.76238069e-02
3.06924462e-01 -6.64837435e-02 1.95152789e-01 -3.67907196e-01
-1.18566918e+00 2.14626446e-01 7.40383804e-01 7.89141417e-01
4.61672157e-01 1.82206202e-02 -1.01157285e-01 -1.23428114e-01
4.40156400e-01 1.48021355e-01 2.11385787e-01 -1.77040696e+00
1.35613427e-01 2.46208295e-01 5.38764596e-01 -1.02286959e+00
-5.99764325e-02 1.55253261e-01 -4.65822846e-01 3.33595216e-01
6.25459850e-01 -4.20749217e-01 -4.67440337e-01 1.48123431e+00
5.89545786e-01 9.64184403e-01 9.60239023e-02 8.53877723e-01
4.89834845e-02 8.73538494e-01 -6.95665600e-03 -9.14059997e-01
7.97073901e-01 -6.28529072e-01 -8.35699141e-01 3.81884038e-01
1.59653172e-01 -4.59588319e-01 5.86667955e-01 3.90547216e-01
-1.37300217e+00 -1.32096067e-01 -8.13018799e-01 4.49647486e-01
-5.45688495e-02 -4.09173310e-01 5.02728224e-01 7.10184932e-01
-1.17441094e+00 4.00025070e-01 -1.15840006e+00 -3.54142606e-01
2.84651995e-01 4.52454060e-01 -7.73309618e-02 1.87862158e-01
-6.61384583e-01 3.88549417e-01 3.14165384e-01 1.96947142e-01
-1.09205484e+00 -4.64916348e-01 -7.38685846e-01 -1.87316954e-01
5.68826079e-01 -6.21001184e-01 1.45912635e+00 -1.20307100e+00
-1.71183348e+00 6.52859449e-01 -3.49402279e-01 -6.80962384e-01
9.37283158e-01 2.66214516e-02 2.02890366e-01 8.04977655e-01
1.37326777e-01 4.97776687e-01 7.99110472e-01 -1.11034071e+00
-1.03677237e+00 -8.23138803e-02 3.23380738e-01 4.35375869e-01
-1.89624816e-01 2.02830255e-01 -5.16118109e-01 -4.72285748e-01
-3.07207048e-01 -7.48663843e-01 -5.98453104e-01 -8.75877291e-02
-2.47402221e-01 1.43651918e-01 7.47138262e-01 -1.15794472e-01
1.24129856e+00 -1.97547090e+00 8.53356719e-02 2.33768657e-01
4.49398488e-01 -1.14493847e-01 1.76329881e-01 5.43924987e-01
3.18568081e-01 3.98974419e-02 -1.86777506e-02 -2.28583649e-01
-3.11556160e-01 1.82777762e-01 -2.22307920e-01 7.61354387e-01
1.49230704e-01 9.37403679e-01 -1.25417459e+00 -8.49030554e-01
7.65948832e-01 2.77446300e-01 -3.97010148e-01 1.85067341e-01
-5.95634282e-02 6.50554061e-01 -6.03934705e-01 5.15366256e-01
2.80176848e-01 -2.75503486e-01 2.94928282e-01 2.45406672e-01
-5.53637505e-01 -4.17153865e-01 -1.41391814e+00 1.49767673e+00
-1.31839752e-01 1.04997492e+00 2.61791855e-01 -1.02895010e+00
1.92423075e-01 5.09980500e-01 1.22364926e+00 -3.93334538e-01
4.22847807e-01 -1.99028984e-01 -3.97386402e-01 -5.43600798e-01
6.57697320e-01 4.07423787e-02 -1.08599439e-01 4.76210117e-01
-1.35220379e-01 1.60450146e-01 3.81092221e-01 4.11640376e-01
1.22455227e+00 1.78199455e-01 4.55205619e-01 -9.17597190e-02
4.85492647e-01 3.78178090e-01 5.72109938e-01 9.39315021e-01
-8.42031419e-01 4.02967602e-01 4.49451983e-01 -2.61489809e-01
-9.01892662e-01 -1.15505779e+00 1.61837786e-01 6.41584873e-01
7.93160915e-01 -3.44985753e-01 -9.30714130e-01 -3.43901902e-01
-3.01941633e-01 1.38649896e-01 -5.39577186e-01 1.60459161e-01
-5.30193567e-01 -7.54293025e-01 3.18126172e-01 5.86317420e-01
4.10619944e-01 -7.79631376e-01 -1.48827934e+00 3.98987353e-01
-2.50352889e-01 -1.39487827e+00 -4.66235906e-01 -5.81348352e-02
-8.85849118e-01 -1.23297250e+00 -6.03596270e-01 -3.67378712e-01
5.21150947e-01 4.99824077e-01 9.50240552e-01 8.45135003e-02
-2.57343411e-01 1.23485768e+00 -2.74493724e-01 -2.46604815e-01
-2.25709081e-01 -3.22352529e-01 -9.96232256e-02 3.92987698e-01
-2.80674174e-02 -3.30960482e-01 -7.79769659e-01 1.37079462e-01
-1.12714863e+00 -2.27125734e-02 2.38431618e-02 5.05027056e-01
6.48435295e-01 9.08740014e-02 4.32053775e-01 -4.27938044e-01
5.93538582e-01 -5.75396061e-01 -1.20260143e+00 7.50537217e-02
-3.10229778e-01 -4.88753378e-01 3.46676588e-01 -5.08428097e-01
-1.11920273e+00 5.00135958e-01 4.70913976e-01 -6.27229393e-01
-4.40792404e-02 3.54886979e-01 1.95441678e-01 -2.61479884e-01
1.16636634e-01 1.65138751e-01 -1.87275410e-02 2.89589852e-01
4.74161565e-01 4.28346604e-01 8.58046591e-01 -4.10781413e-01
5.03046572e-01 1.15075207e+00 -4.03223857e-02 -9.05957639e-01
-2.24802837e-01 -7.80999660e-01 -5.35084128e-01 -1.08531272e+00
1.09803045e+00 -6.74199522e-01 -1.30507600e+00 6.17539525e-01
-1.34440529e+00 -4.73588705e-01 -6.25898182e-01 5.78001976e-01
-8.90786409e-01 5.11740446e-01 -8.34726870e-01 -1.24632895e+00
3.15664768e-01 -1.14277029e+00 9.76293385e-01 5.89453757e-01
-4.82687466e-02 -1.20979548e+00 1.94770694e-01 2.10523590e-01
7.48373475e-03 4.91914302e-01 -6.53638542e-02 -1.67579472e-01
-1.26819503e+00 -2.53423989e-01 1.09594744e-02 -1.39034212e-01
-1.72700509e-01 4.85454977e-01 -6.62428796e-01 -7.32913092e-02
1.72096059e-01 2.27551416e-01 4.14946824e-01 1.09359074e+00
6.95704341e-01 1.54301301e-01 -3.63175929e-01 6.01591408e-01
1.84753227e+00 5.37392557e-01 6.06222570e-01 5.54908514e-01
4.57025677e-01 4.19168711e-01 6.25997126e-01 8.49812329e-01
3.80119801e-01 5.81750154e-01 5.29805243e-01 7.87876993e-02
3.22880447e-01 -1.28194196e-02 4.43589002e-01 5.10636091e-01
-3.19582075e-01 -6.09738648e-01 -6.97055697e-01 7.69544303e-01
-2.17668033e+00 -1.34342337e+00 -1.92277774e-01 2.22432184e+00
3.68274629e-01 -7.88181871e-02 4.30814058e-01 1.62400797e-01
9.21162605e-01 1.05717562e-01 -3.81702125e-01 -2.04038233e-01
1.38072036e-02 -1.65164590e-01 8.30729365e-01 5.74965000e-01
-1.04762220e+00 8.49970043e-01 7.94368839e+00 4.77733076e-01
-9.13268805e-01 -3.85996737e-02 5.97338259e-01 -2.49554098e-01
-6.47297734e-03 1.66222975e-01 -4.93409604e-01 5.69641232e-01
1.08136892e+00 -5.15018523e-01 3.91595513e-01 3.49354118e-01
8.94616067e-01 -6.87023461e-01 -9.54090297e-01 9.57524776e-01
-1.34212881e-01 -1.70465112e+00 -1.20184071e-01 2.72198319e-01
7.01844692e-01 -1.43473417e-01 -1.48049206e-01 -6.08049273e-01
4.24190432e-01 -6.19726777e-01 8.99732709e-01 4.64712471e-01
3.39732707e-01 -6.42577112e-01 1.88860908e-01 2.71959156e-01
-1.31266236e+00 -1.14423737e-01 2.85549968e-01 -3.80328119e-01
8.53901863e-01 2.56018668e-01 -3.82404596e-01 6.29956663e-01
4.71376508e-01 7.70010471e-01 6.10937998e-02 1.24270499e+00
1.90012217e-01 5.72259545e-01 -5.09472907e-01 2.08108097e-01
3.00259173e-01 -4.74629611e-01 9.83852863e-01 1.10240817e+00
1.19540162e-01 2.89541155e-01 2.18195125e-01 6.40786111e-01
-7.83543661e-03 -1.98714137e-01 -8.12861443e-01 1.36373669e-01
3.88740271e-01 1.15878785e+00 -1.39899945e+00 -3.47497255e-01
-6.71348333e-01 1.05042255e+00 -2.91545868e-01 6.45455003e-01
-1.04330945e+00 -2.04774514e-01 3.14712137e-01 -1.38533950e-01
2.55426854e-01 -6.84910119e-01 -1.48276314e-01 -1.31524456e+00
-2.52056308e-02 -4.63476390e-01 4.45854485e-01 -5.99424303e-01
-6.67803109e-01 2.08642855e-01 4.31537330e-01 -1.24941063e+00
-6.56750441e-01 -3.30224276e-01 -8.78233671e-01 2.76787847e-01
-1.61902833e+00 -6.85446799e-01 -3.85093898e-01 7.76521444e-01
7.13514209e-01 1.92397490e-01 -4.08354588e-02 1.60584375e-01
-6.69103980e-01 -1.09116897e-01 1.91466674e-01 2.60350257e-01
-8.60632882e-02 -1.25982583e+00 -2.38003489e-02 1.47018981e+00
1.73116073e-01 8.05015713e-02 5.95246196e-01 -5.09987235e-01
-1.58365977e+00 -6.53751433e-01 5.02083361e-01 -3.96672517e-01
7.87399530e-01 1.35345487e-02 -3.33975941e-01 8.08373511e-01
2.89476663e-01 1.04735404e-01 3.72801572e-01 -7.24240065e-01
6.03539824e-01 -3.10439654e-02 -9.11786556e-01 6.82220161e-01
8.35418046e-01 -3.96804035e-01 -2.65707493e-01 -6.01538569e-02
3.59406024e-01 -4.06734467e-01 -5.80431163e-01 9.60687101e-02
5.02989411e-01 -1.04315460e+00 1.07973588e+00 -3.18377584e-01
2.94486552e-01 -4.23188508e-01 -1.29051194e-01 -6.87549353e-01
2.19740823e-01 -1.27704024e+00 -1.59191772e-01 1.00803339e+00
-1.76029012e-01 -4.83829170e-01 1.17567909e+00 5.95322490e-01
2.67330587e-01 -3.07822347e-01 -1.22681439e+00 -8.07435870e-01
-3.16061139e-01 -7.94716358e-01 -5.72889447e-02 6.50353789e-01
4.02531624e-02 -6.92279264e-02 -2.65342236e-01 1.51201814e-01
9.35663879e-01 -9.15295035e-02 5.14092863e-01 -8.79986107e-01
-1.01342142e-01 -3.48506093e-01 -7.12920606e-01 -1.14353144e+00
1.72615573e-01 -2.86103070e-01 1.86375275e-01 -1.20952356e+00
1.10060252e-01 1.06241114e-01 -2.14432091e-01 -3.13006133e-01
6.87184334e-02 2.80637980e-01 3.88379812e-01 2.68751025e-01
-9.89074826e-01 2.92180181e-01 8.79862845e-01 2.70301104e-01
-3.25302184e-01 8.54736287e-03 -1.26949444e-01 7.93962300e-01
5.20758092e-01 -3.91190082e-01 -6.55468762e-01 -2.95652747e-01
2.34066427e-01 5.83304286e-01 8.03870499e-01 -8.07239354e-01
6.35235488e-01 -5.77981710e-01 1.04676276e-01 -4.46836889e-01
2.74359882e-01 -1.03156090e+00 2.55672842e-01 5.32839239e-01
-3.29790026e-01 2.92597502e-01 1.91453695e-02 1.13642740e+00
-2.81455100e-01 -1.15286000e-01 6.92628443e-01 -1.06001839e-01
-1.02023709e+00 2.26286903e-01 -9.62920129e-01 5.48855849e-02
1.66322637e+00 -8.07948887e-01 6.22978620e-02 -5.95986724e-01
-6.33308530e-01 1.77259102e-01 5.67712545e-01 -1.37301147e-01
7.14904249e-01 -1.06370425e+00 -3.50026786e-01 -9.12661478e-02
-4.24979448e-01 -2.42744118e-01 9.58346352e-02 1.06316972e+00
-7.81405747e-01 1.98154315e-01 -1.92733213e-01 -9.76253569e-01
-1.07488978e+00 3.48377496e-01 5.94546020e-01 -1.33097872e-01
-5.44464350e-01 3.88429344e-01 1.09333992e-01 5.21964550e-01
-2.68775485e-02 -2.48978436e-01 -3.65462787e-02 7.55980238e-02
3.65696609e-01 7.89702058e-01 -3.52084011e-01 -6.08089924e-01
-3.31340402e-01 4.14983600e-01 3.85923535e-01 -8.18976402e-01
1.01165032e+00 -8.36090624e-01 1.22545719e-01 5.47918022e-01
7.91192770e-01 -1.93538919e-01 -1.78756404e+00 1.74487740e-01
2.87732273e-01 -6.80297792e-01 2.20139518e-01 -9.08505768e-02
-1.04191601e+00 4.86556500e-01 4.50493455e-01 4.96467471e-01
1.28704572e+00 -2.51389593e-01 8.06585968e-01 -8.34357142e-02
3.56463939e-01 -1.24114501e+00 7.41402954e-02 1.26302913e-01
6.30066777e-03 -1.07639730e+00 -2.46406913e-01 -5.20039439e-01
-3.82983685e-01 1.01202583e+00 1.42807826e-01 -3.19337487e-01
5.43304503e-01 6.40967011e-01 5.69052016e-03 -7.34215900e-02
-7.88334846e-01 -3.72355163e-01 -2.12560087e-01 7.20528901e-01
9.66724977e-02 -2.35725537e-01 -4.72576410e-01 1.31486863e-01
6.51547253e-01 4.93565500e-01 1.13643861e+00 1.39924276e+00
-1.83409810e-01 -8.91360521e-01 -3.94736320e-01 1.33247271e-01
-7.28344679e-01 2.39017621e-01 -2.25104287e-01 8.76644611e-01
-3.90152097e-01 1.09049964e+00 2.14910939e-01 2.78121755e-02
1.27379358e-01 -3.10575873e-01 5.22406340e-01 -1.47970125e-01
-3.84322405e-01 1.37230560e-01 -2.56426215e-01 -7.81382680e-01
-1.13646221e+00 -8.59341919e-01 -1.16245854e+00 -4.27140146e-01
-4.44476455e-01 7.07731172e-02 4.77946043e-01 9.19397056e-01
2.86850095e-01 4.12205726e-01 8.46713364e-01 -1.24094534e+00
-1.92083176e-02 5.90648279e-02 -4.70208883e-01 2.86797404e-01
5.01901805e-01 -3.68608773e-01 -3.74700963e-01 7.13820696e-01] | [8.687732696533203, -1.2013835906982422] |
4ab1129b-7bc0-4d42-9cb0-9708fa3171ec | zuo-zhuan-ancient-chinese-dataset-for-word | null | null | https://aclanthology.org/2022.naacl-srw.17 | https://aclanthology.org/2022.naacl-srw.17.pdf | Zuo Zhuan Ancient Chinese Dataset for Word Sense Disambiguation | Word Sense Disambiguation (WSD) is a core task in Natural Language Processing (NLP). Ancient Chinese has rarely been used in WSD tasks, however, as no public dataset for ancient Chinese WSD tasks exists. Creation of an ancient Chinese dataset is considered a significant challenge because determining the most appropriate sense in a context is difficult and time-consuming owing to the different usages in ancient and modern Chinese. Actually, no public dataset for ancient Chinese WSD tasks exists. To solve the problem of ancient Chinese WSD, we annotate part of Pre-Qin (221 BC) text Zuo Zhuan using a copyright-free dictionary to create a public sense-tagged dataset. Then, we apply a simple Nearest Neighbors (k-NN) method using a pre-trained language model to the dataset. Our code and dataset will be available on GitHub. | ['Mamoru Komachi', 'Teruaki Oka', 'Hongfei Wang', 'Xiaomeng Pan'] | null | null | null | null | naacl-acl-2022-7 | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 5.04343547e-02 -4.67519820e-01 -4.34900261e-02 -3.37282985e-01
-9.17791069e-01 -7.92462707e-01 4.72547024e-01 3.64349395e-01
-1.05510843e+00 1.10894263e+00 3.92771274e-01 -3.48094374e-01
-1.79759711e-02 -7.94930458e-01 -8.49452689e-02 -4.73392695e-01
2.59538889e-01 5.87162912e-01 2.79509932e-01 -5.98334789e-01
5.99611163e-01 1.76313743e-01 -1.12348115e+00 1.89618826e-01
1.31225634e+00 4.35509920e-01 9.27368939e-01 8.40781331e-02
-6.09742403e-01 7.65139014e-02 -9.67128992e-01 -3.98937255e-01
2.61114657e-01 -4.35421348e-01 -1.22009134e+00 -5.61319649e-01
-3.57057489e-02 3.22543859e-01 -1.12680038e-02 1.22683024e+00
8.46640706e-01 2.95711547e-01 3.45853120e-01 -5.22290945e-01
-1.00897253e+00 1.00557220e+00 -2.45866254e-01 3.61235291e-01
4.49586928e-01 -4.81830239e-01 1.11143839e+00 -8.01439464e-01
9.11176443e-01 9.47296321e-01 1.58212766e-01 5.52826107e-01
-8.15932035e-01 -6.99237466e-01 -1.32811561e-01 4.36219305e-01
-1.76828277e+00 -4.12741080e-02 4.87574339e-01 2.17091828e-03
1.02820718e+00 1.37683019e-01 9.96305287e-01 1.05122447e+00
8.63350630e-02 8.20154250e-01 1.20660233e+00 -6.75896287e-01
3.64634752e-01 -2.03609616e-01 4.95410487e-02 6.98220581e-02
2.76051641e-01 -3.77493232e-01 -6.89738214e-01 -2.09146544e-01
2.48918682e-01 -2.11050272e-01 -2.53180414e-01 5.79488337e-01
-1.35363328e+00 6.45227432e-01 2.82430381e-01 1.01986313e+00
-6.80179819e-02 -4.15417969e-01 3.60885590e-01 3.23224872e-01
6.32034659e-01 8.23980451e-01 -7.61698544e-01 -2.34373286e-01
-1.02217793e+00 2.73074210e-01 6.58090651e-01 1.04863632e+00
8.03368568e-01 -4.26444501e-01 1.14679918e-01 1.26322246e+00
-1.42433739e-03 5.98716378e-01 8.44095230e-01 -4.49926496e-01
4.16916549e-01 6.21486127e-01 -3.50880772e-02 -1.02856457e+00
-3.57138425e-01 -1.85135379e-01 -6.76226258e-01 -5.44300318e-01
2.32727736e-01 -1.77996844e-01 -6.77736700e-01 1.41637886e+00
3.27206999e-01 9.95793641e-02 2.73035407e-01 1.01348972e+00
7.21265018e-01 6.10441983e-01 2.33899623e-01 -9.95581001e-02
1.62447298e+00 -2.27169618e-01 -6.38869166e-01 -5.79383016e-01
5.73128164e-01 -1.24035430e+00 1.23224783e+00 3.03718269e-01
-4.08153564e-01 -1.81916177e-01 -1.14616859e+00 -5.47637582e-01
-7.48546839e-01 1.63467601e-01 5.08014560e-01 4.00720775e-01
-6.92255914e-01 3.00334066e-01 -5.20582020e-01 -8.65974486e-01
1.48957819e-01 -1.01937093e-01 -3.59317750e-01 -3.65104765e-01
-1.85413897e+00 1.17756546e+00 1.06711543e+00 2.24272981e-02
-3.54812413e-01 -7.31727362e-01 -9.13061082e-01 -3.96573365e-01
7.30819881e-01 -3.88146490e-01 9.99006808e-01 -6.05876684e-01
-1.06837702e+00 1.27172589e+00 -2.93737650e-01 -2.52902389e-01
1.06593944e-01 -5.67047536e-01 -9.36536133e-01 -1.54771030e-01
6.34276330e-01 3.19914579e-01 2.61990100e-01 -8.99932027e-01
-8.58237684e-01 -3.62962097e-01 -1.62725538e-01 8.23484659e-02
-4.15439643e-02 1.96731657e-01 -7.11978555e-01 -1.12020159e+00
2.43200585e-01 -9.76574004e-01 -3.99697214e-01 -3.70979041e-01
-4.16973293e-01 -5.35555482e-01 5.44559598e-01 -8.92984688e-01
1.78159463e+00 -2.12774730e+00 -4.23018694e-01 2.47231260e-01
-9.84444320e-02 5.28775096e-01 -1.46452352e-01 5.47856450e-01
1.03588954e-01 3.16609263e-01 -6.16478205e-01 1.70336634e-01
-1.42407969e-01 5.26352704e-01 -1.61214098e-01 1.71056598e-01
1.71593040e-01 5.85587621e-01 -1.48380268e+00 -6.72220409e-01
8.60008299e-02 2.37085298e-01 -1.02191135e-01 -1.08714566e-01
-2.52148122e-01 1.90038621e-01 -8.24818492e-01 6.25605643e-01
6.07596040e-01 1.97156161e-01 6.33014798e-01 1.05712498e-02
-5.60095727e-01 6.59855247e-01 -9.83994067e-01 2.27780747e+00
-5.54721475e-01 4.24574852e-01 -5.16942739e-01 -7.06564665e-01
1.02317894e+00 1.07077576e-01 2.35142961e-01 -9.84769344e-01
4.38993722e-02 5.83622634e-01 -1.23865910e-01 -5.35953879e-01
1.00306654e+00 -2.71553457e-01 -4.79106873e-01 2.85984635e-01
-1.79669470e-01 -3.57688338e-01 5.34861982e-01 2.22568080e-01
1.01994669e+00 7.95428008e-02 8.23087692e-01 -7.71312237e-01
4.92344469e-01 5.90059757e-01 1.03420830e+00 5.42520821e-01
5.76939210e-02 7.39390075e-01 1.20277524e-01 -4.96239305e-01
-9.58059251e-01 -9.17763710e-01 -3.28006178e-01 7.37419546e-01
1.55014589e-01 -8.18760157e-01 -4.99907851e-01 -3.66796613e-01
-2.58016288e-01 9.85145152e-01 -4.05793697e-01 1.35263592e-01
-6.92666531e-01 -8.40894878e-01 7.58761823e-01 1.29468545e-01
7.43052959e-01 -1.04653943e+00 -3.74838322e-01 4.91126239e-01
-5.46942234e-01 -1.07513225e+00 -6.90492094e-01 -8.20476376e-03
-2.59277523e-01 -1.25415993e+00 -7.35746503e-01 -9.47845638e-01
3.29307824e-01 2.41967559e-01 1.04546106e+00 5.17192967e-02
-3.80860239e-01 -3.15682709e-01 -7.41238356e-01 -7.63670146e-01
-2.44302556e-01 3.52476478e-01 2.11510226e-01 -5.59977233e-01
1.07846665e+00 -1.62100881e-01 -4.26785409e-01 4.51521389e-02
-1.12335527e+00 -3.13501030e-01 3.67566317e-01 5.16139686e-01
7.24659264e-01 7.73276836e-02 4.70026731e-01 -1.00591803e+00
7.03023255e-01 -4.97443736e-01 -7.36250699e-01 5.54861844e-01
-5.33159435e-01 1.31928027e-01 4.59277928e-01 -1.14567250e-01
-1.16159928e+00 2.76169693e-03 -5.28802395e-01 3.96641821e-01
-1.61937382e-02 9.31007028e-01 -3.04777324e-01 4.95482832e-01
6.29616737e-01 1.75289899e-01 -5.68460166e-01 -6.42677784e-01
3.62216830e-01 1.02879250e+00 5.79917967e-01 -7.71517813e-01
5.07897794e-01 1.34826571e-01 -3.47377777e-01 -1.05578554e+00
-1.26112902e+00 -5.87584913e-01 -9.78513062e-01 2.62618989e-01
1.09770966e+00 -8.28441501e-01 -4.32011217e-01 2.66974509e-01
-1.08629394e+00 -3.07289995e-02 1.11698642e-01 5.33390224e-01
2.16147736e-01 6.58240616e-01 -2.01554805e-01 -4.33559299e-01
-5.30100644e-01 -7.40346611e-01 7.25233436e-01 2.59764761e-01
-6.15860343e-01 -7.99496233e-01 3.24485481e-01 2.01296389e-01
2.57763267e-01 1.03663892e-01 1.09466898e+00 -8.98907363e-01
-2.23677754e-02 2.97643966e-03 -1.77338626e-02 6.03371002e-02
5.20959258e-01 -4.65852991e-02 -6.11939013e-01 1.04900680e-01
-1.70693770e-01 -6.74231797e-02 6.65230036e-01 6.94127530e-02
8.60557795e-01 -9.66724679e-02 -3.07278514e-01 1.50681272e-01
1.58813238e+00 3.45835626e-01 5.96685529e-01 8.04457247e-01
4.23257709e-01 4.30249602e-01 9.33488309e-01 5.81079066e-01
5.95208764e-01 4.81257200e-01 -3.97956669e-01 3.05488080e-01
6.48325235e-02 -2.14393482e-01 1.84871569e-01 9.34047163e-01
3.67038958e-02 -3.17064434e-01 -1.41547227e+00 9.59405303e-01
-1.51320815e+00 -7.06784964e-01 -8.41837898e-02 2.22228241e+00
1.31355870e+00 1.73402708e-02 -3.50422412e-01 -4.77316603e-02
7.47748017e-01 2.88272146e-02 -1.44480094e-01 -1.46849886e-01
-6.54419899e-01 5.83618104e-01 3.00850719e-01 2.92166024e-01
-9.23645377e-01 1.40701032e+00 5.74794388e+00 1.22013974e+00
-1.06426132e+00 1.09072953e-01 1.17037080e-01 1.17298905e-02
-4.63846534e-01 2.76761085e-01 -7.48476446e-01 5.81590116e-01
4.93350208e-01 -4.37144339e-01 1.88754782e-01 6.18727922e-01
9.62560400e-02 -2.68310964e-01 -6.08694851e-01 1.08019996e+00
8.48441347e-02 -1.20168591e+00 -2.50865109e-02 -2.52096653e-01
6.15190923e-01 3.82265896e-01 -6.60208046e-01 1.15825780e-01
4.73735034e-01 -6.60701215e-01 5.76754391e-01 3.44172776e-01
8.10873806e-01 -8.82197976e-01 7.49000967e-01 3.17503989e-01
-1.25974476e+00 3.99146765e-01 -7.84529626e-01 1.21153310e-01
2.29686603e-01 1.16943777e+00 -6.73527837e-01 7.42100239e-01
8.59701693e-01 8.30690086e-01 -5.33433795e-01 8.15970778e-01
-6.87774122e-01 6.45050287e-01 -4.05570984e-01 -2.31342182e-01
2.04200611e-01 -2.32961118e-01 5.42131305e-01 1.29881883e+00
6.28718793e-01 5.25140405e-01 2.62722492e-01 6.93301022e-01
4.05711792e-02 4.34820890e-01 -3.09875190e-01 -2.60650605e-01
8.15259099e-01 1.09583700e+00 -7.58800387e-01 -3.11029404e-01
-3.51292908e-01 1.03420424e+00 3.32890600e-01 1.65946677e-01
-2.40066275e-01 -6.86006546e-01 8.69485915e-01 -1.65889606e-01
-1.51676489e-02 -4.33763146e-01 -2.91838914e-01 -1.40537798e+00
-1.63779091e-02 -7.54572988e-01 8.72106314e-01 -5.65318823e-01
-1.67155576e+00 7.35557556e-01 -3.04932334e-02 -1.24944437e+00
1.52086943e-01 -4.66369122e-01 -3.43426913e-01 1.03889024e+00
-1.47824609e+00 -9.72710073e-01 -1.72293838e-02 3.81547093e-01
4.81922716e-01 -9.61292759e-02 9.92019653e-01 3.57009023e-01
-4.78618354e-01 3.45750988e-01 1.06640823e-01 3.74325782e-01
1.08300281e+00 -1.30618584e+00 6.23152852e-01 9.81916249e-01
4.20079045e-02 1.06379652e+00 6.44680381e-01 -1.14163852e+00
-1.08653426e+00 -8.60628903e-01 1.71770215e+00 -3.27485770e-01
8.63028705e-01 -2.96393871e-01 -1.11123693e+00 3.23353469e-01
3.91629428e-01 -3.23597044e-01 1.09279096e+00 8.24063197e-02
-1.46816745e-01 2.60708015e-02 -8.97177517e-01 7.86559284e-01
1.13899827e+00 -7.21049011e-01 -1.13441467e+00 1.69647634e-01
6.75921917e-01 -3.03525358e-01 -9.20804679e-01 -1.47718261e-03
3.84783357e-01 -2.75396615e-01 7.13008583e-01 -2.77025282e-01
1.96627244e-01 -7.74673402e-01 -1.80092156e-01 -1.58631623e+00
-7.28382319e-02 -6.23954654e-01 8.28335702e-01 1.49075520e+00
6.96160734e-01 -5.71262777e-01 3.15323412e-01 6.55107856e-01
-2.29446962e-01 -3.46364141e-01 -9.96420205e-01 -6.58150613e-01
2.05903620e-01 -6.87261403e-01 9.06193614e-01 1.38674271e+00
2.77938902e-01 4.75159883e-01 -2.23932207e-01 6.03897423e-02
2.83189058e-01 1.80404231e-01 1.52892947e-01 -1.11434722e+00
3.70608777e-01 -2.20416203e-01 -1.90700516e-01 -8.02696645e-01
3.45947981e-01 -1.17288613e+00 1.50954604e-01 -1.61481571e+00
-4.86835884e-03 -7.51240849e-01 -2.80734569e-01 8.05736244e-01
-6.14416838e-01 7.78026804e-02 1.12316713e-01 2.06399471e-01
-3.24488133e-01 4.91598964e-01 1.05302358e+00 -2.35029291e-02
-1.87437296e-01 -4.00734365e-01 -7.63293922e-01 3.74452561e-01
8.18875432e-01 -9.19284105e-01 -5.77681288e-02 -6.81650817e-01
6.95990741e-01 -3.98170590e-01 -2.10517451e-01 -6.72510743e-01
3.31898421e-01 -4.40857977e-01 3.06849718e-01 -7.21204937e-01
-1.77427635e-01 -6.45460129e-01 8.14529322e-03 3.89670461e-01
-1.54815435e-01 2.87040532e-01 9.35046151e-02 1.47673205e-01
-3.96082669e-01 -3.38072836e-01 4.44087684e-01 -4.24243122e-01
-1.29013550e+00 9.95191112e-02 -5.77701151e-01 4.71555740e-01
7.20347345e-01 6.66769892e-02 -1.91686913e-01 2.33263597e-01
-3.16389054e-01 5.21803379e-01 4.66752380e-01 7.41345465e-01
4.62091863e-01 -1.20037174e+00 -8.71972859e-01 -1.36399478e-01
4.82335329e-01 1.90788597e-01 2.44495824e-01 1.73573464e-01
-7.73363113e-01 2.55008876e-01 -1.19561829e-01 -5.83174936e-02
-9.28622305e-01 5.15422106e-01 -1.78327233e-01 1.63544938e-02
-5.07130921e-01 6.10557616e-01 -3.10540289e-01 -6.14220202e-01
-3.27891171e-01 -1.85709745e-01 -4.04558152e-01 1.19494922e-01
8.37015986e-01 2.38007039e-01 1.74954742e-01 -5.65225899e-01
-6.95523322e-01 5.72023094e-01 -1.35941315e-03 -3.15835662e-02
1.27389669e+00 -3.95984530e-01 -5.50305665e-01 3.06126297e-01
9.98348117e-01 2.29015633e-01 -2.42148191e-01 -4.78068441e-01
6.88544214e-01 -5.37243545e-01 -2.69633353e-01 -8.28730881e-01
-6.72065020e-01 4.22107905e-01 2.29346111e-01 -2.77140141e-01
1.13151741e+00 1.44626021e-01 1.19980097e+00 7.21583843e-01
5.68459153e-01 -1.61529613e+00 -5.60682774e-01 1.10712779e+00
8.27747285e-01 -1.23499691e+00 6.12119511e-02 -2.69839406e-01
-7.86170602e-01 9.55484271e-01 4.05230999e-01 1.83930382e-01
1.01660550e+00 1.25061288e-01 5.78140020e-01 -1.41884387e-01
-2.80899912e-01 -5.29215932e-01 3.40115815e-01 5.67682564e-01
7.67454684e-01 6.95146248e-02 -1.21219528e+00 8.66008103e-01
-7.38370061e-01 7.80445933e-02 2.87770689e-01 1.09593093e+00
-3.04350227e-01 -1.70041203e+00 -6.48810267e-02 3.14757735e-01
-6.54214859e-01 -6.42563403e-01 -4.73193526e-01 6.13922358e-01
4.35754716e-01 1.01504731e+00 -1.87173501e-01 -2.55272090e-01
1.72858357e-01 9.46099591e-03 9.96367633e-02 -9.04331267e-01
-5.26911020e-01 4.21874039e-02 2.10980639e-01 -2.28279784e-01
-6.45095468e-01 -6.01916492e-01 -1.62759793e+00 -2.59466052e-01
-1.77354530e-01 3.52468938e-01 4.50507760e-01 1.40580833e+00
2.26688445e-01 -1.21652111e-01 1.70223594e-01 -1.74106359e-01
2.69651767e-02 -8.90302837e-01 -7.87926853e-01 4.07163888e-01
-2.03330442e-01 -3.62119824e-01 7.07403868e-02 -1.01929486e-01] | [10.252280235290527, 9.412559509277344] |
e8f6d9d5-ad05-46d2-a4ba-3891ebf598fe | inconsistent-few-shot-relation-classification | 2110.08254 | null | https://arxiv.org/abs/2110.08254v1 | https://arxiv.org/pdf/2110.08254v1.pdf | Inconsistent Few-Shot Relation Classification via Cross-Attentional Prototype Networks with Contrastive Learning | Standard few-shot relation classification (RC) is designed to learn a robust classifier with only few labeled data for each class. However, previous works rarely investigate the effects of a different number of classes (i.e., $N$-way) and number of labeled data per class (i.e., $K$-shot) during training vs. testing. In this work, we define a new task, \textit{inconsistent few-shot RC}, where the model needs to handle the inconsistency of $N$ and $K$ between training and testing. To address this new task, we propose Prototype Network-based cross-attention contrastive learning (ProtoCACL) to capture the rich mutual interactions between the support set and query set. Experimental results demonstrate that our ProtoCACL can outperform the state-of-the-art baseline model under both inconsistent $K$ and inconsistent $N$ settings, owing to its more robust and discriminate representations. Moreover, we identify that in the inconsistent few-shot learning setting, models can achieve better performance with \textit{less data} than the standard few-shot setting with carefully-selected $N$ and $K$. In the end of the paper, we provide further analyses and suggestions to systematically guide the selection of $N$ and $K$ under different scenarios. | ['Kam-Fai Wong', 'Gabriel Pui Cheong Fung', 'Jiarun Cao', 'Zhijing Jin', 'Hongru Wang'] | 2021-10-13 | null | null | null | null | ['few-shot-relation-classification', 'few-shot-relation-classification'] | ['methodology', 'natural-language-processing'] | [ 1.39929473e-01 1.21503528e-02 -5.71677268e-01 -5.64928114e-01
-8.66939247e-01 -4.62904423e-02 3.12620163e-01 3.53230909e-02
-4.05722171e-01 5.67925811e-01 -3.46260905e-01 -1.10431928e-02
-4.22457844e-01 -8.81216288e-01 -5.68083584e-01 -5.90128958e-01
1.19870445e-02 5.07578313e-01 3.30885500e-01 -1.94625914e-01
3.29742357e-02 -2.04117671e-02 -1.95789957e+00 2.43884102e-01
6.49438322e-01 1.15262163e+00 1.38980709e-02 3.52444261e-01
-6.90727085e-02 8.28544974e-01 -6.06315851e-01 -5.63179195e-01
1.87812075e-01 -7.68933058e-01 -7.96983182e-01 -1.35192260e-01
3.33025753e-01 -1.76095933e-01 -3.35776299e-01 1.02699995e+00
6.87752545e-01 6.14979923e-01 7.79574692e-01 -1.46081650e+00
-8.34011555e-01 7.53026605e-01 -5.49285769e-01 6.10247254e-01
-5.84273785e-02 1.96038440e-01 1.26572406e+00 -1.28260243e+00
6.71422541e-01 9.73976016e-01 5.29909015e-01 6.58814013e-01
-1.18347716e+00 -9.74809706e-01 2.82751739e-01 7.62008965e-01
-1.65412521e+00 -6.19891942e-01 8.61874580e-01 -2.49523431e-01
1.13809419e+00 2.14932352e-01 3.34055632e-01 9.95045841e-01
-2.04873249e-01 6.68428540e-01 5.81538618e-01 -6.71401501e-01
5.49122930e-01 9.01292637e-02 7.92636156e-01 7.35077381e-01
3.27843249e-01 2.86497846e-02 -6.56345427e-01 -2.03362986e-01
2.92322844e-01 1.41545579e-01 -1.46866009e-01 -5.57487845e-01
-5.47975302e-01 9.71839428e-01 2.17197716e-01 4.28624004e-01
1.20925285e-01 -1.02568250e-02 4.44905102e-01 4.49435502e-01
5.50575137e-01 5.08278012e-01 -5.74478745e-01 -7.22728074e-02
-6.41164899e-01 -4.06937711e-02 6.85232997e-01 1.29235387e+00
9.83283520e-01 2.23079622e-02 -5.44121742e-01 1.32520533e+00
-7.82475621e-02 -5.29794283e-02 4.85007584e-01 -8.40588868e-01
5.04466534e-01 6.33602381e-01 -1.15171187e-01 -6.59636557e-01
-1.91180304e-01 -3.27315599e-01 -7.78149247e-01 -1.31763041e-01
1.91415429e-01 3.57435048e-02 -9.14599776e-01 2.02644801e+00
2.83823252e-01 5.08347631e-01 7.43386820e-02 6.64989531e-01
1.02926397e+00 2.23109186e-01 1.18236303e-01 -6.27182066e-01
1.37641180e+00 -1.04439616e+00 -7.70099819e-01 -5.20685256e-01
8.59423339e-01 -2.69750774e-01 1.32533026e+00 -7.13061020e-02
-1.01434064e+00 -5.84037244e-01 -1.20613658e+00 -4.83465679e-02
-6.29856944e-01 -1.99259639e-01 6.45187497e-01 6.25164032e-01
-5.56649745e-01 5.83196819e-01 -5.08236825e-01 -3.27008277e-01
6.29282534e-01 1.40265107e-01 -1.22881182e-01 -5.39401114e-01
-1.52379739e+00 8.37803662e-01 2.60804772e-01 -1.32381424e-01
-6.66121840e-01 -7.65954792e-01 -1.17611277e+00 2.68578321e-01
9.79769409e-01 -2.57227957e-01 1.08322704e+00 -3.01545858e-01
-1.03933418e+00 9.10913706e-01 -3.25670063e-01 -2.24349186e-01
4.68633510e-02 -1.24489307e-01 -3.28516841e-01 4.76412997e-02
1.74351573e-01 3.44040483e-01 4.97591347e-01 -9.86617565e-01
-4.88598049e-01 -5.00662208e-01 -1.27211273e-01 8.00370052e-02
-2.68601090e-01 4.46360409e-02 -6.74059451e-01 -5.45393527e-01
1.15836099e-01 -5.64014137e-01 1.20235026e-01 -8.83813202e-02
-2.74887592e-01 -6.65080786e-01 7.66123235e-01 1.45014107e-01
1.45389819e+00 -2.25938296e+00 -9.50443093e-03 -8.93446580e-02
1.76863059e-01 5.10930538e-01 -2.29213357e-01 2.22463042e-01
-2.93948293e-01 1.69372857e-01 -2.35762492e-01 -3.32550824e-01
-3.59257199e-02 2.35340014e-01 5.96265420e-02 3.86281401e-01
2.39689946e-01 1.02905381e+00 -8.68353307e-01 -7.24398017e-01
-2.29910500e-02 1.67013556e-01 -2.89595991e-01 2.84649014e-01
-1.01741172e-01 -2.69550800e-01 -2.62686670e-01 8.85249436e-01
4.73246157e-01 -6.21044159e-01 2.19451725e-01 -1.58255130e-01
3.14838946e-01 -1.30428195e-01 -1.07638073e+00 1.59744918e+00
-1.96540132e-01 3.32827657e-01 -3.48095059e-01 -1.40963602e+00
8.90893161e-01 3.01888198e-01 3.30378920e-01 -7.88250446e-01
3.50146204e-01 1.32599683e-03 2.63351679e-01 -5.15432239e-01
1.90118536e-01 -3.37014735e-01 -2.22809151e-01 6.86360240e-01
3.19647610e-01 8.46301317e-02 2.75667518e-01 6.24529608e-02
1.36900985e+00 -3.87443393e-01 5.88867605e-01 5.55122569e-02
2.73996983e-02 -2.94440478e-01 1.07183969e+00 1.03654599e+00
-8.36860061e-01 4.84187096e-01 4.80225384e-01 -2.70838320e-01
-7.33768165e-01 -7.47738063e-01 -1.67891771e-01 1.54337180e+00
4.87156123e-01 -2.68979073e-01 -4.93379265e-01 -7.92691112e-01
-9.10792798e-02 1.02723217e+00 -7.65306771e-01 -8.02212536e-01
-4.42322016e-01 -9.61772323e-01 4.02352482e-01 6.06990516e-01
5.64674914e-01 -1.14521456e+00 -6.26210392e-01 2.05970630e-02
-1.47639513e-01 -9.86188829e-01 -4.27520305e-01 6.82309866e-01
-4.80647206e-01 -1.33351183e+00 -3.75057966e-01 -9.43364739e-01
3.35808992e-01 5.55330932e-01 1.14947903e+00 7.61069208e-02
-6.28864527e-01 1.80712834e-01 -6.86269999e-01 -3.28845620e-01
2.36058503e-01 -6.83237165e-02 -1.01614565e-01 -1.20212652e-01
1.01512611e+00 -4.15120870e-01 -5.01300097e-01 5.21542013e-01
-7.71099627e-01 -2.49439612e-01 4.62316453e-01 1.18663740e+00
7.53317118e-01 1.63882181e-01 8.02269757e-01 -1.31235790e+00
4.67703611e-01 -7.25802600e-01 -8.60046595e-03 7.40990758e-01
-9.02410388e-01 -9.44406018e-02 6.40159547e-01 -6.99355841e-01
-8.88479829e-01 -4.72112089e-01 2.02379972e-01 -8.36324632e-01
6.58172667e-02 5.47905445e-01 -2.41818711e-01 2.13990912e-01
7.99634576e-01 1.01139806e-01 -2.09337145e-01 -2.83033520e-01
2.67870545e-01 6.06005728e-01 1.95220381e-01 -3.63921523e-01
5.49643576e-01 2.76432812e-01 -3.01604003e-01 -5.69219828e-01
-1.16003323e+00 -5.85672021e-01 -5.08779585e-01 1.56664532e-02
6.73388124e-01 -7.57738709e-01 -3.39712679e-01 3.91206056e-01
-7.74442673e-01 -4.82140511e-01 -6.05874538e-01 3.19368362e-01
-6.20358765e-01 1.05345562e-01 -6.34797156e-01 -8.91227722e-01
-2.98399478e-01 -9.62242484e-01 8.35263848e-01 6.40903115e-02
-1.89041898e-01 -5.67447662e-01 4.70073149e-03 3.87577593e-01
2.29469389e-01 -9.38128084e-02 1.38193202e+00 -9.65587914e-01
-2.16976061e-01 -4.09385026e-01 -4.13941890e-01 3.50336172e-02
1.77077547e-01 -3.13617408e-01 -1.14129817e+00 -2.87212610e-01
1.29689753e-01 -8.60926330e-01 1.02401507e+00 2.62097955e-01
1.31266475e+00 -5.33560663e-02 -4.24443603e-01 5.03136933e-01
1.29613769e+00 4.50800747e-01 5.82726002e-01 -1.55899510e-01
5.44598043e-01 6.03597701e-01 1.06309223e+00 4.11222398e-01
2.73328334e-01 6.57718897e-01 2.78766453e-01 2.36434206e-01
-1.40097901e-01 3.90189178e-02 -1.79018512e-01 5.74068248e-01
1.04074739e-01 -3.26424003e-01 -8.70059729e-01 4.58431989e-01
-1.95158708e+00 -1.10838795e+00 6.42105699e-01 2.04864550e+00
9.40245450e-01 1.88376099e-01 -1.74143434e-01 1.77107900e-01
1.03033710e+00 3.21231723e-01 -9.96264756e-01 -9.03096497e-02
9.36781894e-03 6.00910842e-01 -6.15374595e-02 3.21609110e-01
-1.01689661e+00 1.06568503e+00 5.80157518e+00 1.21307671e+00
-7.83250511e-01 3.50033849e-01 8.63212824e-01 -5.21924436e-01
-1.25398859e-01 -1.06474049e-01 -1.01599383e+00 4.97071773e-01
9.92776930e-01 -9.06767324e-02 1.91682130e-01 1.01833332e+00
-2.40194753e-01 -8.44443887e-02 -1.49809468e+00 1.12208211e+00
4.90620881e-01 -1.35608041e+00 -1.97868064e-01 -2.58603394e-01
5.39825261e-01 -1.69418052e-01 -4.81959395e-02 1.01532757e+00
3.71569365e-01 -8.87714982e-01 3.93673569e-01 2.49746576e-01
1.06989729e+00 -7.18282640e-01 7.48160779e-01 5.04082322e-01
-1.30517673e+00 -2.83116221e-01 -5.38611710e-01 1.90914422e-02
-2.80997038e-01 2.40056857e-01 -2.96775609e-01 3.07198554e-01
8.78742456e-01 5.50299466e-01 -4.41415578e-01 5.65498710e-01
1.09981425e-01 3.21688414e-01 1.75738558e-01 -2.93235630e-01
-1.96314335e-01 1.51593447e-01 1.05011754e-01 8.47453415e-01
1.59301087e-01 5.00789523e-01 2.64950752e-01 7.86609590e-01
-1.80241928e-01 -2.16614474e-02 -6.74575806e-01 4.47064862e-02
1.02759945e+00 8.20918560e-01 -6.32303357e-01 -5.09562850e-01
-4.47813869e-01 7.35280275e-01 9.42969918e-01 4.22087312e-01
-8.19849551e-01 -6.88942254e-01 5.37761211e-01 -1.99301243e-01
5.56192696e-01 3.08607727e-01 -8.49408582e-02 -1.24280465e+00
-9.50323325e-03 -5.40270329e-01 7.79528260e-01 -5.30899763e-01
-1.62555456e+00 4.64008540e-01 1.43170282e-01 -1.17231858e+00
-1.58291504e-01 -2.99907237e-01 -7.30572462e-01 6.29905701e-01
-1.40339291e+00 -7.98771143e-01 -1.80772126e-01 4.88271862e-01
7.89938629e-01 -2.56907910e-01 9.91394579e-01 2.46962354e-01
-1.16925359e+00 1.11351132e+00 6.03292845e-02 1.84553906e-01
8.72392893e-01 -8.78513753e-01 3.19253840e-02 6.76049471e-01
9.95719582e-02 6.94832265e-01 5.98816156e-01 -5.06959975e-01
-1.26583600e+00 -1.23181248e+00 8.22184265e-01 -1.36585727e-01
4.01082098e-01 -3.89701694e-01 -1.06010914e+00 6.36643827e-01
-1.40817925e-01 6.21788383e-01 1.29405046e+00 4.67804641e-01
-7.56226122e-01 -2.88249463e-01 -1.25479043e+00 4.80266362e-01
1.36439955e+00 -6.16211712e-01 -5.30573606e-01 1.39165848e-01
1.05767846e+00 3.54661164e-03 -7.56488740e-01 6.68173790e-01
4.44073975e-01 -9.72762585e-01 9.10259545e-01 -7.95400321e-01
2.62403518e-01 2.09452257e-01 -5.44485152e-01 -1.08521986e+00
-5.15127659e-01 -8.82976204e-02 -4.31261510e-01 1.05046356e+00
5.82878709e-01 -4.65641648e-01 9.06872809e-01 8.15257847e-01
-7.41164386e-02 -1.14483845e+00 -1.17111850e+00 -9.14871812e-01
-3.67064797e-03 -6.28765583e-01 5.09268939e-01 1.28080976e+00
2.36131966e-01 6.12570941e-01 -4.24569219e-01 -2.00336859e-01
5.08408785e-01 3.54475200e-01 4.65836018e-01 -1.25969946e+00
-5.32155216e-01 -2.58511722e-01 -2.58065909e-01 -5.88581920e-01
3.33880514e-01 -8.18296969e-01 2.27788091e-01 -1.09735429e+00
6.62317634e-01 -5.58800459e-01 -6.84773564e-01 6.29547596e-01
-6.26002848e-01 2.05216676e-01 5.27903140e-02 1.99321285e-01
-1.13710737e+00 7.15550244e-01 8.33180964e-01 -1.77041024e-01
-2.41139174e-01 -1.56674549e-01 -9.06134844e-01 5.11680603e-01
6.68745458e-01 -3.56541157e-01 -6.56450808e-01 -1.18543871e-01
2.22816747e-02 1.25903934e-01 6.32278174e-02 -7.51915693e-01
2.99026191e-01 -2.70167083e-01 2.75642961e-01 -4.90739763e-01
6.08450770e-01 -5.06166399e-01 -3.82795721e-01 2.56620735e-01
-6.46751583e-01 -3.86795193e-01 -1.26365885e-01 8.35605025e-01
-7.55467117e-02 -3.92402112e-01 9.29268897e-01 -1.84828222e-01
-8.42156112e-01 6.16475940e-01 1.22916892e-01 4.71927285e-01
1.33630419e+00 -2.51404852e-01 -7.33821750e-01 -2.12719977e-01
-7.66191840e-01 4.14270133e-01 1.93661526e-01 3.82999927e-01
7.69149601e-01 -1.41684675e+00 -3.09316754e-01 3.40854287e-01
7.96128809e-01 1.17046103e-01 5.58084905e-01 6.07185781e-01
1.70982704e-01 1.63335487e-01 6.65906966e-02 -2.86344349e-01
-1.27516890e+00 8.27920675e-01 1.77977204e-01 -3.38355809e-01
-3.29800457e-01 1.11338747e+00 -3.49512473e-02 -3.08325768e-01
6.52042270e-01 1.13536194e-01 -1.95497572e-01 3.48123848e-01
6.42290652e-01 4.32181925e-01 -7.02426508e-02 -3.46357822e-01
-3.11418951e-01 3.67670000e-01 -5.23565710e-01 1.09957799e-01
1.22515941e+00 6.80697933e-02 1.31649092e-01 8.69283736e-01
1.25619233e+00 -6.78268909e-01 -1.12779617e+00 -6.17480755e-01
-1.32554352e-01 -4.85029101e-01 -2.64403015e-01 -5.08072495e-01
-9.71503198e-01 8.50240588e-01 6.71091378e-01 1.12284988e-01
8.68733048e-01 4.74233210e-01 4.83552366e-01 5.49200237e-01
4.94857460e-01 -1.33672130e+00 3.51227313e-01 5.15006900e-01
4.06048924e-01 -1.55972409e+00 -7.18631521e-02 -5.11389554e-01
-5.72625458e-01 5.59373379e-01 1.17906749e+00 1.17378876e-01
8.53607237e-01 -5.68638705e-02 -9.73242819e-02 -5.62694192e-01
-1.10780239e+00 -3.60920608e-01 7.52481744e-02 6.05266333e-01
3.85473788e-01 -5.55570573e-02 -2.06170365e-01 9.61148560e-01
1.13709785e-01 -2.06849620e-01 1.67497247e-01 1.28911006e+00
-6.01631343e-01 -8.60162854e-01 4.78740446e-02 8.84707808e-01
-2.44910046e-01 -1.53170466e-01 -3.85412395e-01 7.57626355e-01
3.25791597e-01 1.30540335e+00 1.81932479e-01 -6.47697031e-01
3.12676579e-01 4.89015251e-01 2.40068182e-01 -1.16422915e+00
-3.27527344e-01 -2.56511103e-02 -2.40646582e-02 -5.77410936e-01
-3.79160792e-01 -3.88067126e-01 -1.13379800e+00 -1.67092234e-01
-7.76004553e-01 1.99995160e-01 -6.26412556e-02 1.10078692e+00
3.89925599e-01 5.50261974e-01 6.22277319e-01 -5.07585108e-01
-6.97905779e-01 -1.16414082e+00 -1.02083349e+00 4.69953150e-01
3.85443680e-03 -1.04530215e+00 -4.90790129e-01 -4.06576812e-01] | [9.995759010314941, 3.0285911560058594] |
b9466d4a-566e-4035-8b55-2de8613a3abd | taa-gcn-a-temporally-aware-adaptive-graph | 2305.08779 | null | https://arxiv.org/abs/2305.08779v1 | https://arxiv.org/pdf/2305.08779v1.pdf | TAA-GCN: A Temporally Aware Adaptive Graph Convolutional Network for Age Estimation | This paper proposes a novel age estimation algorithm, the Temporally-Aware Adaptive Graph Convolutional Network (TAA-GCN). Using a new representation based on graphs, the TAA-GCN utilizes skeletal, posture, clothing, and facial information to enrich the feature set associated with various ages. Such a novel graph representation has several advantages: First, reduced sensitivity to facial expression and other appearance variances; Second, robustness to partial occlusion and non-frontal-planar viewpoint, which is commonplace in real-world applications such as video surveillance. The TAA-GCN employs two novel components, (1) the Temporal Memory Module (TMM) to compute temporal dependencies in age; (2) Adaptive Graph Convolutional Layer (AGCL) to refine the graphs and accommodate the variance in appearance. The TAA-GCN outperforms the state-of-the-art methods on four public benchmarks, UTKFace, MORPHII, CACD, and FG-NET. Moreover, the TAA-GCN showed reliability in different camera viewpoints and reduced quality images. | ['Scott T. Acton', 'Peter Young', 'Matthew Korban'] | 2023-05-15 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [-2.22562119e-01 -2.77496479e-03 7.30234459e-02 -3.91517907e-01
1.33464500e-01 7.87765682e-02 3.07706445e-01 -4.26953584e-02
-2.78324664e-01 4.07308429e-01 5.12715317e-02 2.72164285e-01
2.71785334e-02 -8.92644882e-01 -4.82702792e-01 -8.29952121e-01
-3.54139388e-01 2.02569261e-01 1.82525650e-01 -2.21364379e-01
-1.17888227e-01 6.93722010e-01 -1.69930792e+00 -7.32559264e-02
8.35356176e-01 1.53633749e+00 -3.53762388e-01 3.24801832e-01
8.59338939e-02 6.59655452e-01 -1.41507253e-01 -9.43230808e-01
2.39320561e-01 -1.13270991e-01 -3.19075406e-01 2.66776443e-01
7.12805688e-01 -5.29460549e-01 -5.56400180e-01 9.12513256e-01
6.13379419e-01 2.46505991e-01 5.47184944e-01 -1.63015521e+00
-7.63289332e-01 2.56822139e-01 -1.06797600e+00 -1.79936290e-01
2.19610780e-01 2.19208887e-03 4.30627048e-01 -6.59889340e-01
5.86471260e-01 1.66226709e+00 7.67961442e-01 6.41534746e-01
-4.48204547e-01 -8.66828620e-01 7.76043296e-01 5.07521749e-01
-1.38362813e+00 -1.37730911e-01 9.39077318e-01 -3.47350299e-01
4.52093273e-01 -1.01311253e-02 1.02093506e+00 1.21568525e+00
3.39808434e-01 6.61185086e-01 9.81056273e-01 -1.38561606e-01
3.33649069e-02 -6.49295032e-01 -8.46418217e-02 1.43030715e+00
7.70944580e-02 3.17853899e-03 -6.43777847e-01 2.68030409e-02
8.45836341e-01 6.45644516e-02 3.09160203e-02 -3.43744218e-01
-7.04053044e-01 5.35433471e-01 4.53017682e-01 -2.93393582e-02
-4.12620693e-01 3.02348316e-01 5.24478793e-01 1.39961377e-01
8.94617736e-01 -5.14277406e-02 -4.15009141e-01 1.04443938e-01
-6.43949151e-01 1.20991103e-01 2.78527677e-01 9.00012910e-01
4.61715013e-01 3.47366065e-01 -2.04164028e-01 7.81444013e-01
3.31142247e-01 5.51566601e-01 3.38506699e-01 -9.23763156e-01
1.93691716e-01 8.36073697e-01 -3.34482133e-01 -1.30267274e+00
-8.52376163e-01 -2.48713985e-01 -1.16121042e+00 9.05775204e-02
4.52473968e-01 -1.93832610e-02 -1.34346116e+00 2.13451266e+00
5.54612219e-01 1.18716292e-01 -3.97548527e-01 8.42276335e-01
1.23949182e+00 3.79043162e-01 2.13354856e-01 -4.68890458e-01
1.12654293e+00 -1.09734511e+00 -7.83987105e-01 -1.98195353e-01
3.05084944e-01 -4.73210543e-01 4.91399825e-01 4.57626581e-01
-1.05622804e+00 -7.24089324e-01 -7.74611354e-01 8.81465152e-03
-3.75462353e-01 3.27846259e-01 1.02558112e+00 5.86775839e-01
-1.45927715e+00 6.62354589e-01 -7.62876511e-01 -6.36935472e-01
3.69685680e-01 5.86159289e-01 -6.65630460e-01 -2.58795232e-01
-1.00900328e+00 4.20265079e-01 1.97254539e-01 4.65544969e-01
-8.25099826e-01 -2.29522794e-01 -1.21240461e+00 -8.07672963e-02
3.60031009e-01 -8.76071811e-01 7.76617944e-01 -1.42822945e+00
-1.50220501e+00 8.71822238e-01 -1.61010101e-01 7.85806254e-02
5.29501319e-01 -3.61926466e-01 -8.29653025e-01 2.31313989e-01
-2.79160649e-01 7.52084494e-01 1.23453355e+00 -8.26941073e-01
-1.99130341e-01 -9.87842500e-01 1.47928549e-02 1.85558423e-01
-4.64859426e-01 6.12747371e-02 -9.27690029e-01 -9.35613573e-01
2.22655937e-01 -1.01371765e+00 2.00384036e-02 4.85949188e-01
-2.63583697e-02 -3.89075249e-01 9.15176094e-01 -1.02166104e+00
1.47505784e+00 -2.01617718e+00 5.02094686e-01 1.17482997e-01
5.08963168e-01 3.18952292e-01 -2.74465233e-01 1.95322320e-01
-1.41038924e-01 -2.06627116e-01 1.31521836e-01 -5.60261309e-01
-3.05699497e-01 8.02497566e-02 5.45696378e-01 4.66016710e-01
5.21722026e-02 7.20387280e-01 -8.92668009e-01 -8.85493517e-01
2.04604074e-01 4.28368509e-01 -3.00006837e-01 2.02532157e-01
-1.34669319e-01 2.73905158e-01 -2.40113959e-01 1.04829109e+00
8.93731833e-01 -9.46251228e-02 8.60219523e-02 -4.38486338e-01
9.12886932e-02 -7.55492032e-01 -1.09470654e+00 1.72723973e+00
-4.20908257e-02 1.61490858e-01 9.48663428e-02 -5.40196180e-01
1.16124928e+00 1.75498769e-01 6.62355900e-01 -7.81503439e-01
5.21015108e-01 -4.43349080e-03 -2.07388714e-01 -6.07059479e-01
4.05816823e-01 4.33261871e-01 1.90281764e-01 -6.02196865e-02
3.11036497e-01 6.15555823e-01 3.94279629e-01 3.21414024e-01
8.50730777e-01 3.20246458e-01 4.87094074e-02 -1.77693501e-01
6.92563951e-01 -9.73012567e-01 1.01455116e+00 -2.00672075e-02
-4.93270725e-01 5.05613089e-01 4.74653304e-01 -9.84709024e-01
-6.86330736e-01 -1.01698625e+00 3.99050385e-01 1.05668592e+00
2.25887984e-01 -5.52102149e-01 -7.39727318e-01 -8.55703890e-01
-3.51511650e-02 5.45838401e-02 -1.18071461e+00 -3.85295153e-01
-4.33964282e-01 -5.45383990e-01 1.03955992e-01 8.26529622e-01
9.11358893e-01 -9.37127829e-01 -1.46009587e-02 -1.10504977e-01
-1.74754009e-01 -1.28751540e+00 -8.80038798e-01 -5.27222335e-01
-9.44638312e-01 -1.23542559e+00 -7.98404992e-01 -6.43604696e-01
9.56573069e-01 -7.15851113e-02 1.00076139e+00 2.69854844e-01
-2.20093787e-01 5.71341276e-01 -4.42571014e-01 -3.47261935e-01
9.06434506e-02 -1.50366887e-01 5.89920022e-02 5.34579575e-01
1.70616716e-01 -6.82439864e-01 -7.49491096e-01 2.99652666e-01
-5.56884766e-01 3.82811666e-01 5.10055006e-01 6.72696829e-01
4.48132008e-01 -4.47114296e-02 4.09652710e-01 -7.13722944e-01
2.31423721e-01 -2.68911868e-01 -5.76488614e-01 3.34770620e-01
-7.64541745e-01 -6.19617850e-02 3.99682432e-01 -5.51378667e-01
-1.10453844e+00 9.18878540e-02 -1.56121805e-01 -6.87940836e-01
1.75147697e-01 2.52327561e-01 -4.62458372e-01 -2.30488211e-01
1.12726696e-01 -1.01738945e-01 2.85527200e-01 -3.39044571e-01
2.40137711e-01 1.16885751e-01 5.84977567e-01 -5.94083488e-01
7.12145627e-01 3.15532506e-01 4.74063247e-01 -7.53009319e-01
-5.94832838e-01 -1.87271997e-01 -7.49793887e-01 -7.55294323e-01
1.06011403e+00 -9.82717276e-01 -6.22702777e-01 1.52118301e+00
-1.06470442e+00 -3.15097362e-01 2.62435675e-01 3.63987476e-01
-2.93576658e-01 6.55485094e-01 -9.29894567e-01 -8.24169695e-01
-8.04468393e-01 -9.90928948e-01 1.23463607e+00 7.23836839e-01
1.69582874e-01 -9.17894363e-01 -3.44591558e-01 3.67411673e-01
1.65249646e-01 9.07298207e-01 9.46426332e-01 -2.99537182e-02
-2.00742692e-01 -1.61751416e-02 -3.10578018e-01 3.95350456e-01
2.21438006e-01 4.95386750e-01 -8.72904122e-01 -5.38100183e-01
-7.63941944e-01 -3.48402083e-01 6.68669224e-01 6.00633323e-01
1.51368999e+00 -2.76482612e-01 -3.92015055e-02 8.93708110e-01
1.15962732e+00 1.78182349e-01 8.19935918e-01 1.80697124e-02
9.96781826e-01 4.59508836e-01 7.30457842e-01 5.89255095e-01
6.76806271e-01 5.61440587e-01 8.83109212e-01 -4.82875437e-01
-2.74751157e-01 -3.31662893e-02 3.80278617e-01 1.10086083e+00
-7.56364584e-01 -1.81741402e-01 -3.90494585e-01 5.58407247e-01
-2.01262426e+00 -5.98522902e-01 -1.10649236e-01 1.90824783e+00
4.04312640e-01 -9.32190493e-02 5.13939738e-01 -1.02039725e-01
8.67439806e-01 4.27769989e-01 -6.07382834e-01 -4.51869339e-01
-1.33604422e-01 2.77346849e-01 3.13556612e-01 -1.80898532e-01
-1.11054432e+00 8.06444228e-01 4.97693682e+00 6.32343292e-01
-1.11000228e+00 -1.34032503e-01 8.57404232e-01 1.61312923e-01
2.05788791e-01 -5.44664383e-01 -4.92915183e-01 5.12514591e-01
4.95324105e-01 5.05301617e-02 4.89755839e-01 9.30912554e-01
-6.24768026e-02 1.56714782e-01 -8.26896429e-01 1.24293375e+00
4.90654141e-01 -7.65276730e-01 9.32828933e-02 -1.40072644e-01
5.77927291e-01 -4.90541190e-01 1.65309623e-01 3.49897921e-01
5.18161543e-02 -8.91650140e-01 7.57094622e-01 6.27905250e-01
1.10423279e+00 -9.38120544e-01 9.14671838e-01 -3.08105260e-01
-1.70579529e+00 -1.88020736e-01 -1.34149656e-01 1.29581103e-03
-1.63389206e-01 3.73551905e-01 -2.12137340e-04 1.00922918e+00
1.07420278e+00 8.78754914e-01 -1.15673172e+00 8.73942912e-01
-2.94929236e-01 2.09214672e-01 -6.81494847e-02 2.66933858e-01
6.92376792e-02 -9.07328054e-02 1.78981781e-01 8.27888608e-01
5.10607064e-01 1.16040692e-01 1.67158797e-01 5.85280955e-02
-1.79487169e-01 1.79341674e-01 -2.32531011e-01 -1.49831682e-01
2.07337305e-01 1.53491724e+00 -8.16250324e-01 -8.77478719e-02
-5.34550071e-01 1.17999566e+00 4.57992196e-01 3.08137536e-01
-9.86921787e-01 -1.09416924e-01 6.79612815e-01 2.11888716e-01
1.43086731e-01 -2.71106124e-01 4.20790344e-01 -1.02049875e+00
-5.71789593e-02 -9.29411411e-01 7.45990455e-01 -1.17162991e+00
-1.27793860e+00 7.71141768e-01 -7.11032934e-03 -8.46792877e-01
6.06511794e-02 -6.97874010e-01 -5.11459231e-01 5.03222167e-01
-1.14859939e+00 -1.87077355e+00 -9.86580908e-01 9.86813307e-01
3.65121990e-01 -1.44352615e-01 6.37175620e-01 5.15208006e-01
-8.88967276e-01 1.01479065e+00 -5.12331843e-01 4.14784700e-01
9.23750818e-01 -1.00280440e+00 3.70269954e-01 1.07355666e+00
-3.44445437e-01 1.97915763e-01 2.33348593e-01 -8.61176908e-01
-1.43259001e+00 -1.45642114e+00 5.38751900e-01 1.04618482e-01
2.90717542e-01 -2.13918224e-01 -7.55259335e-01 6.83417797e-01
3.57393585e-02 4.07055497e-01 2.20715389e-01 1.50256723e-01
-6.71202779e-01 -6.21300757e-01 -1.18756759e+00 6.27147436e-01
1.41037977e+00 -6.37329593e-02 1.92227259e-01 5.33340611e-02
9.29267585e-01 -6.54619455e-01 -1.02923226e+00 6.43724203e-01
8.81406665e-01 -9.51287150e-01 7.03544497e-01 -3.66379708e-01
3.60896349e-01 -1.78551778e-01 7.66089931e-02 -1.17081833e+00
-6.34706974e-01 -6.73135638e-01 -4.57197160e-01 1.38207841e+00
-2.34309301e-01 -5.16218424e-01 7.31387079e-01 5.18104255e-01
-2.45957166e-01 -1.07204247e+00 -7.86471128e-01 -7.67596185e-01
-4.37445760e-01 -1.40272146e-02 9.34817851e-01 7.89376855e-01
-7.26905823e-01 9.33821797e-02 -6.69678450e-01 3.01398665e-01
7.65311182e-01 -2.41744161e-01 7.13579953e-01 -1.40303349e+00
7.08274022e-02 -2.65064806e-01 -9.04324234e-01 -3.96179408e-01
9.95702893e-02 -4.48310614e-01 -3.70002121e-01 -1.53157067e+00
1.31691158e-01 -2.50473946e-01 -4.94534910e-01 6.31826758e-01
-3.26539725e-01 2.14631215e-01 1.58436090e-01 -4.17452306e-01
-8.62922966e-01 7.16356337e-01 1.44381404e+00 -2.28033662e-01
3.72673199e-02 -9.01604742e-02 -3.39013427e-01 8.37145209e-01
6.99041247e-01 -6.65645972e-02 -5.43292999e-01 -4.72431123e-01
2.52679259e-01 -5.85048236e-02 1.40945628e-01 -1.17422616e+00
5.28646484e-02 -2.85543829e-01 7.22321689e-01 -7.03966856e-01
5.65002918e-01 -7.50545800e-01 4.24732924e-01 4.25332755e-01
3.78673702e-01 6.68219090e-01 2.70165175e-01 5.65255761e-01
2.37225126e-02 3.54089200e-01 9.36025560e-01 -1.48609653e-01
-1.10057819e+00 1.07519782e+00 7.14643300e-02 -2.05143616e-01
1.15110552e+00 -3.49077433e-01 -5.44385910e-01 -5.37935197e-01
-6.54896498e-01 3.01681101e-01 6.06660366e-01 7.21756101e-01
7.34856546e-01 -1.57893765e+00 -5.78747511e-01 3.22679281e-01
2.85078079e-01 -1.44503206e-01 7.80440390e-01 8.90171468e-01
-5.26568174e-01 -2.52620727e-01 -5.56453586e-01 -4.74816650e-01
-1.66348696e+00 8.43485117e-01 2.14962482e-01 -4.29533511e-01
-3.21614146e-01 9.45609510e-01 4.28747386e-01 -1.00341618e-01
4.35278654e-01 -1.52471900e-01 -3.20526391e-01 7.14787142e-03
2.37298787e-01 5.51280618e-01 3.65669690e-02 -8.49783957e-01
-4.13578182e-01 7.89342105e-01 -5.42629734e-02 3.93116415e-01
1.27713823e+00 -1.63966313e-01 -3.70874226e-01 2.20951006e-01
8.20156634e-01 -2.07389995e-01 -1.23345101e+00 -1.83389053e-01
-3.38453084e-01 -3.82143795e-01 -7.79877976e-02 -7.01148093e-01
-1.94422710e+00 6.46294415e-01 9.61257219e-01 -2.88797647e-01
1.76781178e+00 -2.55645126e-01 9.56772268e-01 -1.96186170e-01
5.15699863e-01 -1.19127202e+00 5.03140688e-01 2.96144664e-01
9.43299770e-01 -9.45581794e-01 3.64101142e-01 -4.78859514e-01
-4.80326980e-01 1.35959113e+00 1.25342894e+00 2.26261392e-01
7.04504669e-01 -2.42029965e-01 -1.22100310e-02 -1.83338940e-01
-5.15251100e-01 -1.64112344e-01 5.80190599e-01 7.94291019e-01
2.29440674e-01 4.57645692e-02 -3.84067029e-01 6.46423578e-01
8.73055458e-02 -8.45850334e-02 1.12726726e-01 5.76308846e-01
8.17871764e-02 -9.16720927e-01 -8.06629136e-02 4.08964336e-01
-3.61529589e-01 2.32803345e-01 -4.31228399e-01 9.48833942e-01
5.41035891e-01 7.53570139e-01 4.12074775e-02 -7.77990282e-01
8.71767327e-02 -1.60871968e-01 8.39917362e-01 -1.75351232e-01
-5.53065896e-01 2.37919409e-02 2.46058717e-01 -9.92813230e-01
-6.48687482e-01 -4.98030633e-01 -9.13811266e-01 -5.05892158e-01
-2.47303680e-01 -3.01737219e-01 5.67623258e-01 6.40608191e-01
4.31351542e-01 6.81612909e-01 5.09211838e-01 -7.23915637e-01
2.64920592e-01 -1.00888669e+00 -7.62962639e-01 5.78976572e-01
-3.41339558e-02 -1.04685700e+00 -3.71795893e-03 -3.23624164e-02] | [13.531078338623047, 0.8468642830848694] |
dbb64f8b-64e6-440e-9920-6957e1fbec65 | dynamic-persistent-homology-for-brain | 2201.00087 | null | https://arxiv.org/abs/2201.00087v2 | https://arxiv.org/pdf/2201.00087v2.pdf | Dynamic Topological Data Analysis for Brain Networks via Wasserstein Graph Clustering | We present the novel Wasserstein graph clustering for dynamically changing graphs. The Wasserstein clustering penalizes the topological discrepancy between graphs. The Wasserstein clustering is shown to outperform the widely used k-means clustering. The method applied in more accurate determination of the state spaces of dynamically changing functional brain networks. | ['H. Hill Goldsmith', 'Vince D. Calhoun', 'Ian C. Carroll', 'Shih-Gu Huang', 'Moo K. Chung'] | 2022-01-01 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [-6.00359552e-02 2.19327092e-01 3.86352301e-01 -3.86656970e-01
8.32692236e-02 -5.31294882e-01 3.25943977e-01 1.50774911e-01
-2.93925673e-01 5.21022260e-01 -2.72419214e-01 -3.01855445e-01
-8.70006025e-01 -4.11088556e-01 -1.79308236e-01 -8.94158840e-01
-1.01944518e+00 5.32466054e-01 5.93540490e-01 -7.28250891e-02
1.09851569e-01 6.29549444e-01 -7.29851365e-01 -3.81200224e-01
6.75188422e-01 5.73047161e-01 3.50877583e-01 9.68858480e-01
2.33063564e-01 7.51946345e-02 -1.45950690e-01 4.02653404e-02
2.95172632e-01 -5.36554933e-01 -5.95572174e-01 -2.67753694e-02
3.41679990e-01 2.63321549e-01 -5.18138409e-01 1.34621072e+00
3.52289766e-01 6.60471499e-01 8.66452038e-01 -1.39801311e+00
-5.72051466e-01 7.42719233e-01 -5.88741779e-01 9.53405857e-01
1.87379159e-02 -1.89867303e-01 8.35261047e-01 -3.03785056e-01
9.59887743e-01 1.44406044e+00 5.86703777e-01 4.68725920e-01
-1.42801380e+00 -4.61884290e-01 4.71200570e-02 4.93004054e-01
-1.33357358e+00 -2.21813977e-01 8.57413888e-01 -6.10027492e-01
1.05248713e+00 1.90665230e-01 9.41766798e-01 6.12733960e-01
9.30791020e-01 6.52662739e-02 1.16191506e+00 -2.43153498e-01
3.78683865e-01 -6.11435950e-01 7.13577628e-01 9.34092462e-01
4.99413311e-01 5.41282855e-02 -9.97106172e-03 -4.61936384e-01
8.55384350e-01 -2.50244230e-01 -1.11702777e-01 -9.09575820e-01
-1.23863173e+00 3.43059063e-01 4.86272216e-01 8.70769858e-01
-2.65257180e-01 7.25968957e-01 3.54702592e-01 7.19379365e-01
3.46166134e-01 2.56762087e-01 -8.02195296e-02 1.92111712e-02
-8.27592492e-01 -1.60973355e-01 3.86277646e-01 8.17006111e-01
4.11539912e-01 2.26936936e-01 2.90708899e-01 3.07701856e-01
7.87005663e-01 4.68433112e-01 2.51584142e-01 -1.13417661e+00
-9.60679911e-03 4.80800122e-01 -3.52120072e-01 -1.17086458e+00
-8.29750657e-01 5.65760098e-02 -1.04238975e+00 1.52227515e-02
4.46544796e-01 -1.77373692e-01 -6.91404402e-01 1.40849757e+00
2.99559414e-01 2.55770743e-01 -2.72414178e-01 4.83997107e-01
2.73252010e-01 1.95736498e-01 -8.50099325e-02 -6.70871973e-01
3.63436162e-01 -3.79781634e-01 -1.03441286e+00 3.86182576e-01
4.20126736e-01 -1.35476217e-01 3.72833937e-01 1.91302791e-01
-9.27943110e-01 -8.88045356e-02 -9.66149628e-01 4.56268758e-01
-3.63813698e-01 -7.67492592e-01 3.97052139e-01 7.70324588e-01
-1.90249300e+00 1.22944534e+00 -1.32100821e+00 -5.71102560e-01
2.45676003e-02 3.80603731e-01 -5.25365531e-01 2.49745876e-01
-1.05598235e+00 8.60091090e-01 3.55990648e-01 3.87161165e-01
-7.16878176e-01 -2.82065183e-01 -8.68692517e-01 -5.76500408e-02
-2.89844930e-01 -1.62985250e-01 4.83925432e-01 -3.24600399e-01
-1.38113856e+00 6.53484106e-01 2.34086424e-01 -5.20947814e-01
8.50255370e-01 8.72359514e-01 -2.71381199e-01 3.90974075e-01
-2.01793328e-01 4.20838773e-01 8.79972994e-01 -8.46164525e-01
5.28708696e-01 -5.71830332e-01 -8.02051246e-01 1.51288182e-01
-1.62374526e-01 -2.82480508e-01 3.62942278e-01 -2.78412223e-01
5.40932715e-01 -9.44164753e-01 -5.03914535e-01 -2.80119836e-01
-2.81829983e-01 -5.15930355e-01 9.26316381e-01 -4.30411518e-01
9.97236013e-01 -2.09083891e+00 3.36533785e-01 7.84210920e-01
9.80524302e-01 -2.99486697e-01 -5.73591031e-02 3.96296650e-01
-4.95319366e-01 1.23332426e-01 -1.71563238e-01 1.26105160e-01
4.56694625e-02 1.89983562e-01 2.91180015e-01 1.15559363e+00
-5.30134678e-01 7.86392987e-01 -1.30729091e+00 -5.71586549e-01
2.28000581e-01 4.81306732e-01 -2.11222082e-01 -3.19557965e-01
4.55052942e-01 2.35903099e-01 -2.99411595e-01 3.91979069e-02
4.76185292e-01 -1.72472179e-01 8.20878506e-01 -4.32262123e-02
2.21132189e-01 -5.36231220e-01 -9.25549865e-01 1.18198347e+00
4.83716130e-01 9.67657864e-01 9.06791985e-02 -1.40366256e+00
9.81578231e-01 3.92274171e-01 1.04230750e+00 -1.15436688e-01
3.37128907e-01 -1.93466589e-01 4.94517058e-01 -2.07500726e-01
-3.35826248e-01 -1.93373069e-01 1.59251660e-01 6.27337098e-01
2.32747495e-01 -2.66335875e-01 2.81725079e-01 7.16616333e-01
1.51062679e+00 -5.54359853e-01 2.51341075e-01 -1.26412463e+00
2.41546810e-01 -4.42516387e-01 3.13479573e-01 5.67090750e-01
-1.03783274e+00 5.42998910e-02 8.51477265e-01 -3.72342527e-01
-7.95059204e-01 -1.86221480e+00 -2.81730771e-01 4.74482119e-01
2.61592388e-01 -2.28010520e-01 -1.15804625e+00 -5.19574046e-01
1.00732416e-01 2.60761291e-01 -8.91727984e-01 -3.65588069e-01
-1.33616239e-01 -7.91702271e-01 2.61536807e-01 1.22697622e-01
1.32969841e-01 -9.32005167e-01 -5.01197994e-01 4.54619378e-02
1.98625997e-01 -7.03196526e-01 -6.18075013e-01 9.70862508e-02
-1.30840158e+00 -1.52038574e+00 -4.91923243e-01 -8.39952052e-01
9.55583751e-01 1.34912178e-01 5.78933716e-01 2.29484901e-01
-6.74033046e-01 1.01723897e+00 -1.21041968e-01 2.26689547e-01
-4.59225148e-01 -4.29444790e-01 6.75232172e-01 -1.07596755e-01
2.15585679e-01 -9.12636042e-01 -5.57512879e-01 4.68178600e-01
-5.32933414e-01 -4.46181446e-01 -4.32216376e-02 5.64136922e-01
5.86901307e-01 5.00592291e-01 5.24149060e-01 -7.99227238e-01
1.08869624e+00 -4.02838558e-01 -5.60106218e-01 3.16022933e-01
-1.18675005e+00 2.14665905e-01 2.02322856e-01 -7.00413108e-01
-4.69597071e-01 -5.17348386e-03 6.42317891e-01 -2.65075237e-01
1.00737460e-01 2.24881440e-01 3.51929069e-01 -5.68352699e-01
6.85458720e-01 -9.25860032e-02 3.32101375e-01 -1.57838270e-01
4.74924982e-01 2.00776160e-01 2.66827941e-01 -2.40188465e-01
6.89254403e-01 4.95528787e-01 2.20861375e-01 -1.02230620e+00
2.83786356e-01 -5.73972583e-01 -1.30549383e+00 -1.01890063e+00
8.80481899e-01 1.31683156e-01 -6.70375526e-01 2.50845522e-01
-7.19111264e-01 -7.24483669e-01 -2.13569537e-01 4.51045424e-01
-8.82035375e-01 7.94153988e-01 -7.98951924e-01 -7.25404084e-01
-2.27098048e-01 -8.11798096e-01 3.82209331e-01 -3.09577554e-01
-2.83153415e-01 -1.78608763e+00 7.32577026e-01 -2.04466149e-01
2.13217929e-01 8.02199483e-01 8.82786512e-01 -3.16852689e-01
8.76347870e-02 -1.99717417e-01 -5.61664766e-03 -8.70117173e-02
5.04235141e-02 5.26390195e-01 -1.92314684e-01 -5.56479931e-01
3.87301333e-02 4.10212427e-01 4.91372019e-01 8.35781276e-01
4.84171391e-01 8.95055532e-02 -5.56369305e-01 2.69978512e-02
1.35559356e+00 4.70284194e-01 3.74291718e-01 -1.46396205e-01
5.87943256e-01 7.11275697e-01 -1.19279601e-01 -3.84880938e-02
2.06919685e-01 -4.60811034e-02 1.67508438e-01 4.28297549e-01
1.56501219e-01 2.35288873e-01 2.15654284e-01 1.38285971e+00
-1.77101061e-01 2.68557400e-01 -1.42606521e+00 6.18609846e-01
-2.01007652e+00 -1.04844189e+00 -4.61131305e-01 1.62333393e+00
3.56600076e-01 2.92947173e-01 1.39519110e-01 3.39857861e-02
1.15827739e+00 1.19525656e-01 -7.06785977e-01 -3.46584380e-01
3.10026467e-01 -3.08069736e-01 6.20857000e-01 8.58879864e-01
-4.88461167e-01 7.12845266e-01 8.88591766e+00 1.78248271e-01
-3.03713888e-01 3.80780011e-01 2.17738912e-01 -6.60237074e-02
-9.79573131e-02 -6.72974512e-02 1.28815919e-01 3.91814202e-01
1.15712643e+00 -7.24760234e-01 7.86863685e-01 4.22986180e-01
4.71185505e-01 -5.40490858e-02 -8.96402478e-01 1.14430904e+00
-1.36613503e-01 -7.48819172e-01 -3.19612861e-01 3.24803293e-01
8.45309138e-01 2.67657012e-01 -9.77996141e-02 -1.44020423e-01
7.75935411e-01 -6.03885293e-01 2.96206057e-01 7.43569434e-01
6.10120416e-01 -5.26136577e-01 2.42065370e-01 9.93019119e-02
-1.43781769e+00 2.37483650e-01 -7.37611711e-01 1.26543224e-01
1.68870419e-01 8.82917166e-01 -6.32809103e-01 -9.23118889e-02
7.80628145e-01 1.07069099e+00 -6.38175666e-01 9.71403599e-01
2.03050852e-01 1.92170918e-01 -2.61501461e-01 -2.73959816e-01
1.84334874e-01 -7.17032671e-01 9.25426364e-01 9.11556542e-01
3.51123326e-02 1.50066093e-01 -1.01381801e-01 7.04488695e-01
1.65601790e-01 4.56267968e-02 -9.88264322e-01 -2.52941608e-01
2.24556834e-01 1.26146078e+00 -1.57405102e+00 -3.30347747e-01
3.99836004e-01 7.29421198e-01 4.63718921e-01 2.97804296e-01
-3.10506314e-01 -4.10866767e-01 5.36170959e-01 2.32150003e-01
-1.55591413e-01 -8.82220387e-01 2.34373361e-02 -9.93806064e-01
-4.89829719e-01 1.35942787e-01 4.69336957e-01 -7.60807931e-01
-1.26621819e+00 5.14171302e-01 4.03369099e-01 -5.87469280e-01
-1.12417519e-01 -6.67001665e-01 -6.49247169e-01 2.83726990e-01
-8.12989771e-01 -1.02171833e-02 -1.22274905e-01 7.30149865e-01
-1.47716962e-02 -1.65296897e-01 4.41275239e-01 -1.35689855e-01
-5.53284824e-01 -3.96029092e-02 6.50949836e-01 3.14259380e-02
4.99670655e-01 -1.86031508e+00 4.20726359e-01 8.89132500e-01
-2.62241997e-02 6.32887125e-01 8.05398345e-01 -1.02738607e+00
-1.24802661e+00 -7.56672323e-01 4.94004369e-01 -4.11017567e-01
1.40061164e+00 -4.37924534e-01 -8.15149188e-01 7.32596993e-01
2.75646478e-01 2.60262072e-01 4.49707776e-01 -1.86551467e-01
-8.02517533e-02 4.22640704e-03 -1.51669812e+00 4.39365596e-01
1.34704828e+00 -5.08674622e-01 -4.58849311e-01 7.15363383e-01
2.36434028e-01 2.20806688e-01 -1.38719642e+00 7.34663103e-03
4.15266216e-01 -7.37449527e-01 6.30957007e-01 -8.10609162e-01
-6.59269571e-01 -1.31249741e-01 2.02557698e-01 -1.76115131e+00
-8.88846993e-01 -1.01215720e+00 -8.93829986e-02 5.30218780e-01
2.17411697e-01 -8.11829150e-01 6.72743022e-01 7.13891625e-01
4.38591003e-01 -2.13395357e-01 -1.32922685e+00 -1.11363101e+00
1.12939395e-01 -5.29913418e-03 -4.47298214e-02 1.31358945e+00
1.00561821e+00 1.18263468e-01 3.16794276e-01 -2.50447452e-01
1.38122118e+00 -4.49197054e-01 -2.02205747e-01 -1.56296229e+00
2.75051117e-01 -7.76036143e-01 -1.19580150e+00 -3.50756734e-03
4.48365003e-01 -1.09947276e+00 2.22994536e-01 -1.65258384e+00
5.59652969e-02 -1.02769539e-01 -6.81442976e-01 2.04268564e-02
2.24005148e-01 1.16484657e-01 -1.29022777e-01 1.06322765e-01
-1.09541893e+00 2.59499609e-01 1.25617516e+00 -8.34607780e-02
-1.92936823e-01 -4.88111287e-01 -6.33643866e-02 4.52744663e-01
9.34766471e-01 -4.60235476e-01 -7.20518529e-01 6.00418188e-02
2.48687983e-01 1.25561863e-01 -4.12044600e-02 -9.58709836e-01
3.69756252e-01 1.52128533e-01 1.04855247e-01 -2.71625519e-01
-3.17128271e-01 -1.00328791e+00 4.39697891e-01 8.06511939e-01
-3.58774096e-01 5.99846005e-01 -3.93013567e-01 1.28584695e+00
2.34527364e-01 -6.83375373e-02 1.34453046e+00 -9.45293382e-02
-2.93158859e-01 5.70559025e-01 -8.04164290e-01 9.34187695e-02
1.36095572e+00 -2.67262131e-01 -1.85668081e-01 -2.64042825e-01
-1.57528377e+00 3.87467772e-01 3.28740835e-01 8.05017576e-02
7.39552855e-01 -1.44932950e+00 -2.55826890e-01 -5.14331870e-02
-3.80683362e-01 -5.92456400e-01 2.33574942e-01 1.18158793e+00
-6.67649686e-01 9.08423439e-02 -5.83671689e-01 -6.38234794e-01
-1.42497730e+00 6.04730666e-01 6.96500063e-01 -1.64282337e-01
-7.84464896e-01 4.42618579e-01 -3.94009024e-01 -3.02110523e-01
5.40056303e-02 -3.89871806e-01 -2.86249638e-01 8.01365450e-02
2.43648499e-01 1.01633346e+00 -1.57017484e-01 -5.05606353e-01
-5.51074386e-01 3.06713521e-01 2.99947590e-01 -1.47535101e-01
1.52632487e+00 -4.54137206e-01 -6.36271596e-01 9.23147857e-01
1.29095125e+00 -8.43054235e-01 -1.25408709e+00 -7.04542035e-03
4.12896007e-01 -5.34933545e-02 1.99265033e-01 -2.32749090e-01
-1.00482154e+00 7.31041789e-01 8.51441205e-01 1.03426981e+00
6.42723382e-01 1.01851523e-01 3.98149729e-01 5.50544679e-01
5.42956114e-01 -1.30957210e+00 -4.82728519e-02 2.80990660e-01
7.25900769e-01 -7.36312449e-01 1.13463402e-01 -1.90096334e-01
-2.50206381e-01 1.20611238e+00 2.07763925e-01 -8.23364377e-01
1.35884202e+00 4.04135466e-01 -2.16772482e-01 -7.32059777e-01
-5.96664786e-01 3.22359055e-01 3.31753701e-01 1.02332008e+00
1.40279070e-01 3.66149157e-01 -2.56936252e-01 -2.85157382e-01
-1.46505103e-01 -7.17944801e-01 7.00011075e-01 5.88161707e-01
-6.69830441e-01 -5.23518085e-01 -9.65255648e-02 7.98541069e-01
1.29112422e-01 1.04697950e-01 -8.34053934e-01 5.87128580e-01
-6.54404640e-01 8.25623155e-01 1.89090386e-01 -1.79467380e-01
1.98644996e-01 2.99990058e-01 9.84437227e-01 -3.32181215e-01
-1.17406949e-01 3.78233306e-02 -4.57752794e-01 -6.09404624e-01
-5.76939046e-01 -1.19154000e+00 -1.15641510e+00 -5.74566305e-01
-4.75927919e-01 3.46545249e-01 5.15545607e-01 6.29974008e-01
1.07737601e-01 5.68094313e-01 5.50616443e-01 -8.31431270e-01
-1.71960294e-01 -1.03317976e+00 -1.19193113e+00 4.01704013e-01
3.09981495e-01 -9.25064683e-01 -7.33757913e-01 5.20998687e-02] | [7.098659038543701, 5.251951694488525] |
5961faf0-20e6-49e7-8d6f-263858b960f9 | reducing-bias-in-modeling-real-world-password | 2010.12269 | null | https://arxiv.org/abs/2010.12269v5 | https://arxiv.org/pdf/2010.12269v5.pdf | Reducing Bias in Modeling Real-world Password Strength via Deep Learning and Dynamic Dictionaries | Password security hinges on an in-depth understanding of the techniques adopted by attackers. Unfortunately, real-world adversaries resort to pragmatic guessing strategies such as dictionary attacks that are inherently difficult to model in password security studies. In order to be representative of the actual threat, dictionary attacks must be thoughtfully configured and tuned. However, this process requires a domain-knowledge and expertise that cannot be easily replicated. The consequence of inaccurately calibrating dictionary attacks is the unreliability of password security analyses, impaired by a severe measurement bias. In the present work, we introduce a new generation of dictionary attacks that is consistently more resilient to inadequate configurations. Requiring no supervision or domain-knowledge, this technique automatically approximates the advanced guessing strategies adopted by real-world attackers. To achieve this: (1) We use deep neural networks to model the proficiency of adversaries in building attack configurations. (2) Then, we introduce dynamic guessing strategies within dictionary attacks. These mimic experts' ability to adapt their guessing strategies on the fly by incorporating knowledge on their targets. Our techniques enable more robust and sound password strength estimates within dictionary attacks, eventually reducing overestimation in modeling real-world threats in password security. Code available: https://github.com/TheAdamProject/adams | ['Massimo Bernaschi', 'Giuseppe Ateniese', 'Marco Cianfriglia', 'Dario Pasquini'] | 2020-10-23 | null | null | null | null | ['security-studies'] | ['miscellaneous'] | [-4.07505482e-02 -4.49339122e-01 1.31253123e-01 1.02594392e-02
-5.30730844e-01 -1.30369627e+00 4.07649428e-01 3.59091640e-01
-5.69705784e-01 4.61324275e-01 -9.24657509e-02 -1.21657979e+00
3.23412754e-02 -8.65193427e-01 -4.83707041e-01 -2.67851919e-01
9.34544206e-02 3.98171842e-01 -1.65598281e-02 -6.38898492e-01
5.30337393e-01 6.06588840e-01 -9.46221173e-01 1.15778163e-01
5.99551797e-01 3.47504705e-01 -7.48661458e-02 1.00869465e+00
3.09292287e-01 3.92085493e-01 -8.38101983e-01 -6.01715565e-01
4.75395501e-01 -3.86359096e-02 -5.72398245e-01 -6.88639164e-01
1.99950308e-01 -8.38103652e-01 -1.35051295e-01 1.03066063e+00
7.38426089e-01 -2.07912982e-01 3.58540744e-01 -1.01610279e+00
-4.40520465e-01 9.28032398e-01 -1.16975985e-01 4.13340867e-01
7.44992554e-01 8.44594836e-01 6.36742055e-01 -4.82873201e-01
4.01898846e-02 8.83827746e-01 9.63936031e-01 4.61019725e-01
-1.42144907e+00 -1.01891494e+00 -4.51687761e-02 -1.56225711e-01
-1.57319355e+00 -5.43195009e-01 6.20937526e-01 -4.03922856e-01
9.14006233e-01 2.43795402e-02 5.33966660e-01 1.61907053e+00
6.40222728e-02 -3.28005664e-02 1.24914920e+00 -4.79387164e-01
2.22840026e-01 6.41542256e-01 1.31271988e-01 2.69026160e-01
7.37783790e-01 5.32002270e-01 -1.73624277e-01 -7.59910882e-01
6.92199588e-01 -5.17256651e-03 -5.85705400e-01 -1.91642940e-01
-9.68038678e-01 8.36686373e-01 -4.12179977e-02 4.73694205e-01
-4.76726800e-01 -1.10806830e-01 3.23230535e-01 4.80277240e-01
-4.06633526e-01 1.12905145e+00 -6.98958933e-01 -6.43559635e-01
-8.09820831e-01 3.36200804e-01 1.21858943e+00 4.16215897e-01
5.48822105e-01 3.65769058e-01 4.98655558e-01 3.81178141e-01
2.94330180e-01 7.72774041e-01 4.70093071e-01 -5.81986666e-01
3.83112341e-01 2.61584461e-01 6.02263153e-01 -1.16451228e+00
-2.13255808e-01 -3.37148070e-01 -3.79630119e-01 2.98148066e-01
8.80764782e-01 -2.42899477e-01 -7.37728596e-01 1.78842533e+00
4.68287505e-02 1.69836745e-01 -9.18527693e-02 3.73021990e-01
-1.35545373e-01 1.74003854e-01 3.46914887e-01 -2.31370199e-02
1.44490707e+00 1.61165431e-01 -3.06036651e-01 -1.51892021e-01
5.71843445e-01 -8.72393489e-01 1.69745982e+00 6.72759414e-01
-1.03180206e+00 -4.24688995e-01 -1.31717491e+00 6.77020788e-01
-5.87477028e-01 -6.09455168e-01 6.20728910e-01 1.69455981e+00
-1.00440490e+00 5.64001203e-01 -9.64219630e-01 -2.45867714e-01
1.17967650e-01 6.77965462e-01 -3.43738019e-01 2.60415345e-01
-1.71730602e+00 1.22869718e+00 4.07595366e-01 -1.83202811e-02
-1.21049213e+00 -7.42375970e-01 -7.29839683e-01 1.50846928e-01
2.20973641e-01 -4.51824099e-01 1.48193729e+00 -6.11057699e-01
-1.45682347e+00 3.97866428e-01 2.71866292e-01 -3.30835670e-01
2.64562637e-01 6.02546483e-02 -2.47222185e-01 -8.08885694e-02
-3.94521564e-01 -1.82227388e-01 9.76573527e-01 -1.41533005e+00
-7.27346167e-02 -1.09650172e-01 5.20632982e-01 5.16879596e-02
-4.74414438e-01 -1.75952911e-02 2.51231819e-01 -4.55059260e-01
-5.04416645e-01 -8.18278730e-01 -3.54872346e-01 -5.81458271e-01
-5.11427969e-02 4.89295155e-01 6.50372386e-01 -7.18644917e-01
1.68141341e+00 -1.96805751e+00 -2.69279271e-01 5.99101007e-01
3.98597330e-01 8.08962941e-01 -2.91796401e-02 9.29394841e-01
-4.15034331e-02 5.39958417e-01 1.12757564e-01 -2.86584079e-01
1.72813684e-01 2.40582451e-02 -7.09689379e-01 3.57094765e-01
-3.68310750e-01 5.41178882e-01 -9.04811859e-01 -3.73474173e-02
3.34908992e-01 5.30440927e-01 -8.12149048e-01 3.88142467e-01
-1.89466041e-03 4.69163418e-01 -3.42879117e-01 7.30030715e-01
6.31213069e-01 -1.20787814e-01 6.25717163e-01 -6.55121207e-02
1.67183384e-01 5.53389072e-01 -1.36208689e+00 1.06073427e+00
-8.10945868e-01 -1.32811880e-02 -1.21472012e-02 -6.17345810e-01
5.57384491e-01 4.46641147e-01 -6.84366450e-02 -4.29043204e-01
4.84119087e-01 4.51222450e-01 3.73457074e-01 -1.19835071e-01
4.38032895e-01 -1.52583525e-01 -2.73835689e-01 1.04473162e+00
-9.31392312e-02 -2.16182336e-01 -2.18046218e-01 3.18983525e-01
9.88761961e-01 -5.61240613e-01 5.33900797e-01 6.73751011e-02
4.17515397e-01 -3.66184562e-01 2.54707903e-01 1.28213942e+00
-4.24871624e-01 7.92841464e-02 3.46885979e-01 -5.38261890e-01
-1.06230485e+00 -1.16569507e+00 -1.29739176e-02 9.12512541e-01
-1.78529575e-01 -5.05123377e-01 -1.05091095e+00 -7.13896096e-01
-9.24509987e-02 9.49897587e-01 -5.00838757e-01 -5.30666471e-01
-6.84574544e-01 -6.24428570e-01 1.27743804e+00 4.85291034e-01
2.48021975e-01 -9.22658682e-01 -6.97038054e-01 2.72624016e-01
-6.88059032e-02 -9.68426466e-01 -5.55378079e-01 9.40565541e-02
-5.05622089e-01 -1.21181953e+00 -5.73462062e-02 -2.74026066e-01
4.88650560e-01 1.85779423e-01 1.11412811e+00 7.50969350e-01
9.96580198e-02 7.69928098e-01 -3.23424906e-01 -5.03074467e-01
-8.72824430e-01 2.50056177e-01 5.98367691e-01 -2.98829734e-01
6.96149647e-01 -7.56817400e-01 -7.21276879e-01 4.72824603e-01
-9.46780860e-01 -7.22227514e-01 5.61317742e-01 8.37172449e-01
-6.14118353e-02 2.24233046e-01 3.51500541e-01 -9.19957697e-01
1.29516935e+00 -4.99555439e-01 -6.83076620e-01 1.70210510e-01
-9.73670661e-01 -7.17851296e-02 1.00309992e+00 -1.14141393e+00
-6.11861467e-01 -2.38761902e-01 -5.02339065e-01 -1.50951475e-01
-3.75818700e-01 3.66296768e-01 -3.61009926e-01 -5.90956926e-01
1.14379013e+00 4.48922336e-01 -1.38511375e-01 -2.99054503e-01
2.12436453e-01 8.22766423e-01 8.54970738e-02 -1.14345276e+00
1.24469495e+00 2.49463413e-02 -6.20016098e-01 -6.22057438e-01
-2.64684409e-01 3.80853787e-02 -4.51338172e-01 6.20878115e-02
2.12014958e-01 -5.38554013e-01 -1.19890940e+00 8.48399818e-01
-8.01854551e-01 -7.45340586e-01 1.29959315e-01 1.64876491e-01
-2.85772413e-01 7.82773674e-01 -5.20607233e-01 -8.38582397e-01
-3.96380663e-01 -1.50496697e+00 5.31748593e-01 2.63189077e-01
-6.29198551e-01 -1.24844635e+00 2.26660252e-01 3.91374588e-01
9.02393878e-01 -5.61335050e-02 7.04189956e-01 -1.19113457e+00
-3.37165326e-01 -5.49277484e-01 3.41578722e-01 4.36193258e-01
2.89576679e-01 5.88731794e-03 -9.24701989e-01 -6.90534651e-01
3.09622526e-01 -2.77575999e-01 -1.60309032e-01 -1.19684018e-01
8.22955370e-01 -5.21350861e-01 -4.03696559e-02 5.64705253e-01
1.29797208e+00 3.55937749e-01 5.69592595e-01 4.85905468e-01
3.89172673e-01 1.35352403e-01 1.63285822e-01 5.19373715e-01
5.35256326e-01 6.27400219e-01 1.41951367e-01 2.26380944e-01
5.96791685e-01 -6.99077308e-01 2.18879819e-01 2.58322537e-01
-3.87946926e-02 -6.74510002e-02 -1.20469201e+00 3.27408224e-01
-9.96451020e-01 -1.05838931e+00 4.16426986e-01 2.67492723e+00
1.17016697e+00 6.35414422e-01 2.72457510e-01 5.98060966e-01
3.47311765e-01 -2.32838131e-02 -2.34153017e-01 -8.13903451e-01
3.43308181e-01 6.46535218e-01 6.08554602e-01 9.16133285e-01
-7.85369873e-01 1.06761181e+00 6.37124252e+00 3.85264188e-01
-1.19840491e+00 4.12219688e-02 3.23064148e-01 2.67614275e-01
-3.37388605e-01 4.03012276e-01 -8.22354674e-01 6.93753719e-01
1.43343079e+00 -1.06177010e-01 1.00386226e+00 5.59481859e-01
2.21865222e-01 7.83166811e-02 -9.05127823e-01 6.02487087e-01
-1.63976386e-01 -9.64039981e-01 9.18290168e-02 1.41881675e-01
-4.67233211e-02 -4.77306783e-01 5.28172374e-01 4.84656483e-01
7.47973919e-01 -1.08298457e+00 3.83605599e-01 1.29588038e-01
8.97298574e-01 -7.33435392e-01 6.69988394e-01 3.95908177e-01
-9.04818416e-01 -2.77958184e-01 -9.14684087e-02 -3.24066013e-01
1.11183293e-01 2.26074625e-02 -1.28965878e+00 -3.35387811e-02
3.38030428e-01 -1.55352280e-01 -4.74497497e-01 5.42158127e-01
-3.70220333e-01 1.05634689e+00 -5.65296173e-01 3.73871066e-03
5.48972599e-02 3.04045141e-01 1.00247957e-01 1.04292464e+00
2.21179426e-02 3.12623113e-01 1.84858739e-01 4.96144354e-01
2.53657341e-01 -1.19780235e-01 -6.98206544e-01 -3.87904167e-01
1.15800679e+00 1.05897093e+00 -4.75464910e-01 7.62845129e-02
-8.96441415e-02 6.18285596e-01 1.47287756e-01 4.94510323e-01
-8.11036110e-01 -3.86210978e-01 9.10796404e-01 3.51718634e-01
-2.11918708e-02 -7.19345093e-01 -3.64206761e-01 -1.20873964e+00
-3.89302522e-01 -1.91923761e+00 3.62140894e-01 -3.12147051e-01
-1.28934240e+00 6.00209653e-01 6.74759820e-02 -8.65329206e-01
-2.84920365e-01 -5.82346618e-01 -6.15267038e-01 1.24214935e+00
-1.14708865e+00 -1.11649311e+00 -6.57372922e-02 8.68665516e-01
-5.89089431e-02 -5.43479361e-02 1.10907185e+00 4.12764430e-01
-2.90967435e-01 1.27108467e+00 -6.24834776e-01 4.26084161e-01
7.48720765e-01 -1.22984910e+00 8.09613705e-01 8.35825801e-01
1.04332753e-01 1.74889195e+00 1.08076990e+00 -6.30102396e-01
-1.31627107e+00 -2.79655963e-01 5.90571105e-01 -9.98551011e-01
1.02215755e+00 -4.63503838e-01 -1.16887546e+00 8.43370914e-01
-1.14994861e-01 -3.98926318e-01 1.08173466e+00 1.10089302e-01
-8.54565024e-01 2.47046962e-01 -1.39752698e+00 8.20010006e-01
4.59100425e-01 -1.05931664e+00 -5.36820173e-01 9.46672186e-02
3.73719811e-01 -4.68772143e-01 -9.48149085e-01 9.22133476e-02
7.49826312e-01 -8.63375902e-01 1.16611862e+00 -8.19303453e-01
-5.63984573e-01 -3.12239945e-01 -2.30126008e-01 -1.00953913e+00
-7.87376091e-02 -9.74431992e-01 -7.99671859e-02 9.01970983e-01
4.35304344e-01 -1.03064430e+00 6.50083840e-01 8.69354129e-01
6.21466756e-01 -5.71927190e-01 -5.38549960e-01 -4.99182791e-01
4.78032827e-01 -6.14238381e-01 8.79295468e-01 1.24445581e+00
5.25187291e-02 -9.75435302e-02 -4.22546118e-01 7.44953275e-01
6.82552576e-01 -6.61667347e-01 9.31240857e-01 -9.38575864e-01
-7.37559795e-01 -3.93401027e-01 -3.07030469e-01 -6.87681198e-01
-1.33297294e-01 -2.70278513e-01 -4.44573313e-01 -6.00032389e-01
-1.86909541e-01 -8.07228923e-01 -2.82003373e-01 5.60643792e-01
-3.52698982e-01 1.78636149e-01 1.22186102e-01 1.83834732e-01
-4.95383106e-02 -9.20409262e-02 5.14219463e-01 1.27065510e-01
-3.91849250e-01 2.84928918e-01 -1.21192300e+00 7.02470601e-01
1.15604889e+00 -5.36546111e-01 -6.37599528e-01 -1.05694719e-01
4.54664260e-01 -4.22817329e-03 2.74731725e-01 -9.59399521e-01
1.68114275e-01 -2.90818751e-01 2.09013596e-01 2.79843479e-01
1.01517878e-01 -8.13548088e-01 2.55104572e-01 7.10383117e-01
-5.05274087e-02 5.94423115e-01 4.38779384e-01 2.76222974e-01
3.91877472e-01 -3.38571608e-01 8.80704284e-01 -2.45165318e-01
-1.66763350e-01 1.99778616e-01 -5.62548995e-01 6.46144003e-02
8.41818869e-01 -4.97351050e-01 -3.02124172e-01 -6.02050722e-01
-6.06837571e-01 -2.09767714e-01 1.09089017e+00 1.37329042e-01
6.19989276e-01 -5.85762024e-01 -3.53744686e-01 4.69586194e-01
-1.77678123e-01 -6.47696555e-01 1.65746018e-01 4.56184685e-01
-7.80447841e-01 5.00342436e-02 -1.75515756e-01 -1.70178294e-01
-1.17242944e+00 7.61613548e-01 8.10476065e-01 -5.38078964e-01
-7.63512999e-02 6.23696685e-01 -4.99979407e-01 -5.02222419e-01
1.45921379e-01 1.29132956e-01 -5.84639423e-02 -3.26382220e-01
8.16036224e-01 -1.66711107e-01 6.40521348e-02 -2.05533057e-01
-3.35530430e-01 2.01126963e-01 -5.62934816e-01 -2.20462173e-01
7.39756644e-01 1.14639759e-01 2.72876859e-01 -2.08776793e-03
5.47726810e-01 4.76457149e-01 -8.36563885e-01 -3.64961481e-04
-7.02435374e-02 -7.25903511e-01 -3.05802196e-01 -1.06282699e+00
-5.93609333e-01 8.04577708e-01 5.35052598e-01 4.27543640e-01
1.07268775e+00 -7.74483323e-01 8.95262718e-01 5.97846150e-01
8.62198174e-01 -6.39855385e-01 8.28053504e-02 3.22529227e-01
3.09328914e-01 -9.78105366e-01 -8.00540596e-02 2.57698923e-01
-5.20808160e-01 9.52486217e-01 5.43193817e-01 9.62882042e-02
8.08056831e-01 6.10814035e-01 5.20004630e-01 -1.77377984e-02
-4.75689083e-01 4.11196947e-01 -3.72695059e-01 9.09329057e-01
1.37389019e-01 8.11351538e-02 -1.16996489e-01 7.43011713e-01
-2.55833954e-01 -2.60343879e-01 8.79149377e-01 1.04460812e+00
-2.02258870e-01 -1.64643872e+00 -6.80338323e-01 1.03628613e-01
-6.05927527e-01 -4.73977745e-01 -4.43474501e-01 6.94848180e-01
-1.45561159e-01 1.18882084e+00 -6.07161105e-01 -8.30506444e-01
4.26180184e-01 1.09893329e-01 3.21166694e-01 -6.00566328e-01
-1.05363524e+00 -6.03945076e-01 -2.74470933e-02 -3.45709145e-01
2.90991873e-01 -5.35073757e-01 -7.29985118e-01 -8.90061200e-01
-3.44774514e-01 3.57348561e-01 5.94636023e-01 6.60795152e-01
3.19952011e-01 -5.83010986e-02 7.35076308e-01 -7.81117797e-01
-1.28111672e+00 -8.31531823e-01 -3.13346177e-01 5.01986802e-01
1.43308073e-01 -7.43776441e-01 -6.60957336e-01 -2.76576459e-01] | [5.834367752075195, 7.606562614440918] |
34eb5b05-1ed2-4752-83c2-635ca8daa96c | explainable-artificial-intelligence-4 | 2303.14615 | null | https://arxiv.org/abs/2303.14615v1 | https://arxiv.org/pdf/2303.14615v1.pdf | Explainable Artificial Intelligence Architecture for Melanoma Diagnosis Using Indicator Localization and Self-Supervised Learning | Melanoma is a prevalent lethal type of cancer that is treatable if diagnosed at early stages of development. Skin lesions are a typical indicator for diagnosing melanoma but they often led to delayed diagnosis due to high similarities of cancerous and benign lesions at early stages of melanoma. Deep learning (DL) can be used as a solution to classify skin lesion pictures with a high accuracy, but clinical adoption of deep learning faces a significant challenge. The reason is that the decision processes of deep learning models are often uninterpretable which makes them black boxes that are challenging to trust. We develop an explainable deep learning architecture for melanoma diagnosis which generates clinically interpretable visual explanations for its decisions. Our experiments demonstrate that our proposed architectures matches clinical explanations significantly better than existing architectures. | ['Mohammad Rostami', 'Ruitong Sun'] | 2023-03-26 | null | null | null | null | ['melanoma-diagnosis'] | ['computer-vision'] | [ 1.76577494e-01 4.90653992e-01 -6.01167977e-01 -1.97986171e-01
-3.22002888e-01 -4.51855093e-01 4.90739256e-01 2.86725909e-01
1.30124137e-01 6.97605252e-01 3.58055919e-01 -7.75407612e-01
6.54889420e-02 -7.10454464e-01 -2.13331550e-01 -7.62264013e-01
2.48876378e-01 3.36330354e-01 -1.89783424e-01 -1.18119895e-01
1.08917942e-02 6.60553098e-01 -1.00026739e+00 8.97302926e-01
8.50598514e-01 7.87402391e-01 -7.05002844e-02 1.02735722e+00
-4.46620435e-01 1.02493322e+00 -4.56664383e-01 -6.19932175e-01
-9.64281708e-02 -4.02025312e-01 -9.52239096e-01 2.53664017e-01
2.44087517e-01 -3.53839874e-01 -1.46720603e-01 9.52138007e-01
6.20884309e-03 -9.20068622e-01 1.17590082e+00 -1.37345600e+00
-8.52999032e-01 5.55162542e-02 -7.12547660e-01 -2.13585496e-02
2.52854705e-01 3.30448389e-01 6.99830890e-01 -6.88679934e-01
4.07964736e-01 1.11814761e+00 8.64776611e-01 1.19813764e+00
-8.47577035e-01 -3.07677716e-01 1.13781288e-01 4.36442226e-01
-9.19235349e-01 -6.88049868e-02 1.93954095e-01 -5.65657794e-01
6.52619779e-01 7.08058119e-01 9.60350513e-01 1.35279858e+00
9.51187432e-01 8.12989950e-01 1.19903791e+00 -2.77582616e-01
1.46758750e-01 2.03624554e-02 2.62633771e-01 9.60392594e-01
3.65820646e-01 2.08857149e-01 -2.18702644e-01 1.67897746e-01
8.22259009e-01 7.57879734e-01 -2.92801559e-01 1.65252730e-01
-8.12157452e-01 9.07491565e-01 1.05976140e+00 -8.72395635e-02
-3.89006943e-01 3.30081701e-01 2.01945305e-01 2.34009743e-01
1.17636256e-01 2.44883537e-01 -1.69786751e-01 2.32810125e-01
-4.12188023e-01 -2.05539927e-01 5.07783175e-01 2.55573034e-01
3.08670253e-01 -3.74553427e-02 -2.35828176e-01 4.61499274e-01
4.22371298e-01 2.74153441e-01 5.91133893e-01 -3.16841066e-01
-3.08527738e-01 1.03341734e+00 -2.40996003e-01 -7.60297358e-01
-5.51074862e-01 -5.02508581e-01 -1.51956916e+00 7.01668918e-01
5.29260576e-01 2.96985228e-02 -1.54898500e+00 1.14173281e+00
-1.37158819e-02 -1.79316532e-02 1.41082540e-01 9.95232642e-01
1.02824605e+00 2.53704429e-01 5.27641416e-01 1.69749036e-01
1.42564166e+00 -1.08543861e+00 -7.57398784e-01 -4.95244682e-01
4.59637403e-01 -5.66741765e-01 8.57869685e-01 4.73493874e-01
-6.52613580e-01 -2.78340131e-01 -8.60918760e-01 -2.13709354e-01
-3.51061761e-01 1.87147036e-01 1.27514577e+00 5.70016444e-01
-1.10861659e+00 2.92648673e-01 -8.58689070e-01 -7.84467101e-01
9.08113360e-01 2.88333744e-01 -5.22530735e-01 -8.64906088e-02
-5.87193549e-01 1.07471037e+00 7.16667250e-02 3.47176373e-01
-1.22202635e+00 -5.43726921e-01 -5.79578817e-01 -1.65394500e-01
6.54185563e-02 -1.03572392e+00 1.11397481e+00 -1.41686690e+00
-8.81575406e-01 1.16716301e+00 -4.65997338e-01 -6.06666267e-01
5.74018896e-01 7.98298419e-02 -4.99700755e-01 7.02506080e-02
-3.62567812e-01 7.59873450e-01 1.10434628e+00 -1.17525363e+00
-7.52271473e-01 -3.26575100e-01 8.52665976e-02 8.72274023e-03
-8.82381350e-02 -4.68554139e-01 -2.13554371e-02 -4.90193218e-01
-1.07731661e-02 -8.93858433e-01 -7.13565350e-01 7.34276474e-01
-9.09959376e-01 4.44850214e-02 8.77454638e-01 -7.47123480e-01
9.21596050e-01 -1.73900044e+00 -7.22910836e-02 -1.17605634e-01
1.14063513e+00 3.44834626e-01 3.32468189e-02 2.75671005e-01
-2.95130253e-01 5.67093670e-01 7.64384866e-02 2.42629796e-02
-3.91930073e-01 3.58487159e-01 -1.50963843e-01 2.30562165e-01
5.69544971e-01 1.16682971e+00 -9.48399603e-01 -2.84567297e-01
3.54616344e-01 6.69741333e-01 2.96075810e-02 2.36423850e-01
-2.21618369e-01 3.52387756e-01 -3.00848037e-01 9.46012795e-01
5.42048395e-01 -6.72986090e-01 1.88064814e-01 -1.26298115e-01
3.90856951e-01 -1.79699615e-01 -3.65359753e-01 1.05144727e+00
-1.90503597e-01 1.11800528e+00 -3.09802622e-01 -5.60235679e-01
4.33913887e-01 2.95697421e-01 -4.76538688e-02 -2.24341691e-01
1.84900001e-01 -6.59881383e-02 3.80043596e-01 -7.74097264e-01
-1.07109599e-01 -5.01769543e-01 4.14502352e-01 2.22138435e-01
-6.20220959e-01 7.73775578e-03 -3.54180038e-01 1.25241458e-01
1.32615125e+00 -4.74135250e-01 8.70354891e-01 7.54418373e-02
4.46174711e-01 2.87993670e-01 7.50660241e-01 4.16012973e-01
-3.27763110e-01 7.28567898e-01 8.70885372e-01 -1.32584250e+00
-7.31826186e-01 -1.18927646e+00 -1.13695122e-01 2.26350889e-01
4.26084101e-02 -6.28073886e-02 -5.87182999e-01 -1.12430835e+00
1.36284873e-01 4.72131371e-01 -1.27290678e+00 -2.08431602e-01
1.03529468e-01 -7.75623977e-01 3.67753536e-01 8.14647377e-01
3.41113478e-01 -6.32620037e-01 -4.85800147e-01 -1.66656926e-01
3.13503414e-01 -6.20721340e-01 1.10786289e-01 2.17940226e-01
-7.32115984e-01 -1.80959117e+00 -8.28732312e-01 -6.95085526e-01
1.39413571e+00 4.34290498e-01 1.11456966e+00 7.96226382e-01
-8.73416722e-01 5.93389086e-02 -1.60808057e-01 -9.33875918e-01
-9.47809935e-01 -2.46088460e-01 -1.98842004e-01 2.67524403e-02
6.49295092e-01 2.18946502e-01 -9.93003666e-01 7.04091862e-02
-8.52875948e-01 7.04605758e-01 1.04123807e+00 1.15360057e+00
6.27127230e-01 8.68949369e-02 2.53722876e-01 -1.20091128e+00
7.47662485e-01 -6.18289173e-01 2.40314286e-02 4.26260352e-01
-5.61298072e-01 -1.21115193e-01 6.51607811e-01 -1.30371109e-01
-8.75815034e-01 -2.46914905e-02 -1.35199547e-01 -1.94046527e-01
-5.72232783e-01 3.59310180e-01 2.67013043e-01 -2.45593771e-01
7.71021843e-01 -3.12077086e-02 3.33007604e-01 -1.56372324e-01
-8.99485424e-02 6.92349374e-01 2.64222264e-01 3.79966587e-01
7.00107932e-01 6.29471779e-01 4.58366990e-01 -7.13972986e-01
-1.09580362e+00 -1.11790515e-01 -4.36622053e-01 -3.25041771e-01
1.06450999e+00 -7.58681536e-01 -6.43058956e-01 4.27726358e-01
-1.08489680e+00 -2.68316120e-01 -1.56499838e-04 3.08307502e-02
1.04710303e-01 2.80631155e-01 -7.48297036e-01 -6.05676472e-01
-2.75220186e-01 -1.09261346e+00 9.66345847e-01 5.34050047e-01
-7.42766023e-01 -1.54382014e+00 -1.52964190e-01 3.70070219e-01
3.28798085e-01 6.76103294e-01 1.29636002e+00 -4.00452316e-01
-4.98654217e-01 -2.94123650e-01 -6.02965713e-01 1.01490647e-01
5.13104737e-01 4.45147365e-01 -1.18552768e+00 -2.10165158e-01
-4.49489683e-01 -2.36166850e-01 9.95925307e-01 5.79716504e-01
1.45619929e+00 -1.92604586e-01 -8.72754574e-01 6.68410122e-01
1.42315960e+00 9.77749079e-02 5.51615775e-01 1.28195882e-01
6.09563649e-01 5.33109963e-01 3.41130376e-01 8.84492174e-02
3.85967612e-01 1.19754739e-01 9.86272991e-01 -9.32889462e-01
-5.11663616e-01 -1.67461947e-01 -8.25540647e-02 2.45246783e-01
-2.26427436e-01 -3.10396761e-01 -1.32269931e+00 4.77066427e-01
-1.77632344e+00 -7.77179420e-01 -7.42734790e-01 1.90729928e+00
5.01395464e-01 3.26845460e-02 -2.59822696e-01 2.06911832e-01
4.34368134e-01 -3.10507774e-01 -6.33358955e-01 -8.18899691e-01
-8.17975551e-02 -4.46229242e-02 3.09967190e-01 4.11642045e-01
-9.79623318e-01 6.29976451e-01 7.50489187e+00 3.70720834e-01
-1.52760077e+00 -1.32302999e-01 1.19808006e+00 1.44529313e-01
-4.92491633e-01 -1.81055814e-01 -4.80780035e-01 3.61906469e-01
5.36321163e-01 -8.22799057e-02 -2.25916535e-01 6.36571050e-01
2.37067983e-01 -1.64183572e-01 -1.53426969e+00 1.00787318e+00
-1.53580144e-01 -1.72978330e+00 4.51067060e-01 2.74682283e-01
6.72289908e-01 -4.77764040e-01 4.41740274e-01 -1.26730978e-01
3.96368623e-01 -1.86662292e+00 2.03124154e-02 7.50753760e-01
1.06864202e+00 -5.69441438e-01 1.04248786e+00 1.57497123e-01
-7.00729966e-01 -2.44735718e-01 -5.11668026e-01 -2.78210849e-01
-3.41318369e-01 3.85765344e-01 -1.68755722e+00 9.25546959e-02
3.59471381e-01 8.52117062e-01 -8.87689412e-01 1.17645490e+00
-7.28556573e-01 5.45151711e-01 2.76464254e-01 -1.04869053e-01
3.52276891e-01 3.91589075e-01 2.26276800e-01 1.00607145e+00
3.65429074e-01 -1.41630828e-01 -1.56725436e-01 8.00364494e-01
2.32831508e-01 -2.23485395e-01 -7.76571870e-01 -8.12135637e-02
1.95251908e-02 1.39046812e+00 -6.58466876e-01 -6.72670826e-02
-3.94659042e-01 1.25654209e+00 2.43802354e-01 2.30158150e-01
-5.76660931e-01 2.87709981e-01 8.96772265e-01 3.24143797e-01
-4.29281294e-01 5.00834525e-01 -5.79956174e-01 -1.13315296e+00
-4.96727854e-01 -8.56649876e-01 5.44634163e-01 -8.74555767e-01
-1.40820193e+00 7.25019991e-01 -7.93675661e-01 -1.29691195e+00
-3.24578106e-01 -1.13061595e+00 -1.00452423e+00 7.58402288e-01
-1.78447986e+00 -1.75836742e+00 -9.15812731e-01 4.28055972e-01
7.13629365e-01 -2.52161413e-01 1.43248427e+00 -4.69371796e-01
-7.44760275e-01 5.28877258e-01 -3.15309733e-01 1.36816025e-01
5.57323575e-01 -1.67683005e+00 1.56328753e-01 6.21697366e-01
-1.13748752e-01 3.00996035e-01 6.03559911e-01 -5.85697353e-01
-1.22171271e+00 -1.11416066e+00 7.65163720e-01 -5.10176420e-01
5.10846913e-01 1.67950511e-01 -7.82842815e-01 5.59106231e-01
4.71523792e-01 -1.08088292e-01 1.29196894e+00 1.62809476e-01
-1.71331808e-01 -6.15999065e-02 -1.09219944e+00 9.07643557e-01
5.06459534e-01 -3.51610750e-01 -3.93699646e-01 6.17152512e-01
1.50235504e-01 -2.94506162e-01 -5.18855274e-01 1.63607836e-01
5.42672992e-01 -1.08821821e+00 7.48849750e-01 -8.82856488e-01
7.96067059e-01 -2.14180276e-01 4.31907594e-01 -1.55683124e+00
-4.88741308e-01 -3.20566058e-01 -2.28662249e-02 5.53601027e-01
5.39432526e-01 -4.24150825e-01 1.21822262e+00 8.82104158e-01
1.83443338e-01 -1.16076291e+00 -5.35477638e-01 -1.66669741e-01
2.49140989e-02 -1.55285716e-01 4.36356753e-01 1.04439008e+00
-1.27287924e-01 1.12456396e-01 -1.43021166e-01 3.24436992e-01
7.29765832e-01 -4.63683717e-02 4.62034822e-01 -1.28654993e+00
-9.10506845e-02 -5.73773861e-01 -8.16634774e-01 -3.71920675e-01
-1.90713271e-01 -8.19722950e-01 -5.67330003e-01 -1.86202848e+00
5.37206590e-01 -2.51836896e-01 -4.69047248e-01 7.93117940e-01
-3.85674596e-01 3.13179612e-01 -9.56682861e-02 7.31633306e-02
-5.65878525e-02 -9.50616375e-02 1.48216045e+00 -5.29819787e-01
2.88666219e-01 1.00733466e-01 -1.08042109e+00 1.15871835e+00
6.99077308e-01 -1.34974688e-01 -3.84502202e-01 -6.70005143e-01
3.65223646e-01 1.19005024e-01 8.45842957e-01 -7.81454980e-01
2.48915970e-01 -4.65572655e-01 1.01120412e+00 -3.80140096e-01
1.76678210e-01 -9.55395758e-01 3.17335218e-01 1.01412201e+00
-3.33738625e-01 -2.08907634e-01 2.23467723e-01 8.42747927e-01
-3.28755766e-01 -8.18654969e-02 7.75196195e-01 -2.44097203e-01
-1.05186462e+00 4.05332565e-01 -7.26044178e-01 -5.65523863e-01
1.41819286e+00 -4.85124439e-01 -5.78487277e-01 -6.15441859e-01
-8.32951307e-01 2.42360547e-01 7.26473331e-01 3.44276994e-01
1.10550892e+00 -1.20575941e+00 -7.08713531e-01 1.84676707e-01
2.31687590e-01 -9.76940468e-02 3.10466617e-01 7.85118520e-01
-1.02032948e+00 3.98738712e-01 -5.43962002e-01 -6.72619283e-01
-1.65454197e+00 2.92894572e-01 6.82824314e-01 -2.96729803e-01
-5.54794133e-01 1.19636726e+00 5.88092029e-01 1.69431657e-01
3.22612017e-01 -3.32162529e-01 -3.08069795e-01 -2.67767608e-01
7.84205794e-01 7.22962916e-02 -3.13300908e-01 -1.09666988e-01
-2.86989808e-01 4.97760475e-01 -5.45202136e-01 6.99243903e-01
1.16093969e+00 1.46474481e-01 -9.23870280e-02 1.77822560e-01
6.16735220e-01 -3.89463931e-01 -1.20201147e+00 2.80519426e-01
-3.08242887e-01 -5.92593849e-01 2.01663300e-01 -1.28995872e+00
-1.30482793e+00 1.42984092e+00 7.93773830e-01 3.60241622e-01
1.15365064e+00 -2.48573422e-01 5.87142229e-01 1.27474785e-01
3.72733474e-02 -6.11466229e-01 -6.98465900e-03 -1.72390178e-01
9.26347733e-01 -1.77072227e+00 4.84996960e-02 -6.79199278e-01
-8.38609278e-01 1.65243471e+00 8.18703949e-01 9.98297706e-02
4.10320431e-01 3.14890027e-01 6.52321815e-01 -3.51789773e-01
-1.16936338e+00 -1.75756574e-01 5.25396764e-01 8.82387578e-01
6.58980310e-01 4.12597090e-01 -2.25116382e-03 5.27549446e-01
3.45581651e-01 4.28830869e-02 7.06483066e-01 5.78074634e-01
-2.03969777e-01 -1.19832826e+00 -3.26565325e-01 7.68008351e-01
-5.96984625e-01 -4.01744656e-02 -1.05066538e+00 8.94764721e-01
2.50514686e-01 7.16082692e-01 1.35082332e-02 -3.77311110e-01
-2.32737839e-01 -8.66152123e-02 2.76629061e-01 -6.08601630e-01
-4.63645428e-01 -2.68112928e-01 1.30831465e-01 -4.57942069e-01
-5.60560971e-02 -1.20067246e-01 -1.29854584e+00 -3.51947874e-01
2.37900451e-01 -3.66662294e-01 4.43841726e-01 8.77771199e-01
2.76555628e-01 6.93843067e-01 3.87713879e-01 -6.83629215e-02
-2.33016819e-01 -5.03337741e-01 -5.47966003e-01 6.03303730e-01
6.66096270e-01 -2.77924925e-01 -1.94920242e-01 2.65182376e-01] | [15.476926803588867, -2.778426170349121] |
2fd6e05a-4702-4530-a90e-8ce4279c4462 | increased-confidence-adversarial-examples-for | 2005.06023 | null | https://arxiv.org/abs/2005.06023v2 | https://arxiv.org/pdf/2005.06023v2.pdf | Increased-confidence adversarial examples for deep learning counter-forensics | Transferability of adversarial examples is a key issue to apply this kind of attacks against multimedia forensics (MMF) techniques based on Deep Learning (DL) in a real-life setting. Adversarial example transferability, in fact, would open the way to the deployment of successful counter forensics attacks also in cases where the attacker does not have a full knowledge of the to-be-attacked system. Some preliminary works have shown that adversarial examples against CNN-based image forensics detectors are in general non-transferrable, at least when the basic versions of the attacks implemented in the most popular libraries are adopted. In this paper, we introduce a general strategy to increase the strength of the attacks and evaluate their transferability when such a strength varies. We experimentally show that, in this way, attack transferability can be largely increased, at the expense of a larger distortion. Our research confirms the security threats posed by the existence of adversarial examples even in multimedia forensics scenarios, thus calling for new defense strategies to improve the security of DL-based MMF techniques. | ['Rongrong Ni', 'Benedetta Tondi', 'Wenjie Li', 'Mauro Barni'] | 2020-05-12 | null | null | null | null | ['image-forensics'] | ['computer-vision'] | [-1.15785562e-01 -1.06481828e-01 2.26273075e-01 1.44166619e-01
-6.37574196e-01 -1.13457048e+00 7.69987762e-01 1.90974846e-01
-5.81486106e-01 5.09782732e-01 -2.96774626e-01 -7.82885492e-01
-3.32792327e-02 -1.04600322e+00 -1.00720835e+00 -6.97539449e-01
-2.37128973e-01 6.41791299e-02 5.01949668e-01 -3.61942291e-01
4.22205657e-01 1.12831736e+00 -1.20305538e+00 3.64579648e-01
2.48646885e-01 4.84783024e-01 -3.02319437e-01 8.56497943e-01
-6.08150326e-02 7.63298988e-01 -1.06551766e+00 -9.74810541e-01
7.36655653e-01 -5.19904718e-02 -8.29479039e-01 -2.24323735e-01
5.37587821e-01 -6.28448784e-01 -6.76537931e-01 1.31596899e+00
6.12131834e-01 -3.11946034e-01 4.65882450e-01 -1.55270219e+00
-2.47353539e-01 5.20399392e-01 -4.25265878e-01 5.89700580e-01
3.49073857e-01 5.78550220e-01 3.33972722e-01 -2.69643426e-01
6.42463386e-01 1.44555628e+00 4.59506094e-01 3.54838699e-01
-9.15873706e-01 -8.88290703e-01 -1.94010302e-01 4.29686099e-01
-1.15195262e+00 -2.88764656e-01 7.18684793e-01 -2.82700062e-01
4.87282991e-01 1.27409041e-01 1.31243199e-01 1.32012594e+00
2.10924357e-01 2.87042975e-01 1.14115548e+00 -6.14524007e-01
1.87046170e-01 4.06451344e-01 9.02968124e-02 4.90013421e-01
7.08786011e-01 2.63747156e-01 -2.03343198e-01 -4.95549798e-01
6.81215167e-01 -4.04199839e-01 -2.83401579e-01 -1.64897740e-01
-8.61169159e-01 1.01421463e+00 4.13909048e-01 5.48366249e-01
-2.01910332e-01 1.81348190e-01 9.94569242e-01 5.23649871e-01
1.77955285e-01 6.75425589e-01 -1.07720688e-01 8.11468065e-02
-8.89209151e-01 2.53665417e-01 9.29058969e-01 6.54396534e-01
5.84123552e-01 3.15567791e-01 2.28579760e-01 4.35137264e-02
-1.87804457e-02 3.11945200e-01 1.31169170e-01 -9.01010871e-01
4.65554535e-01 2.97558606e-01 -1.62134305e-01 -1.12562299e+00
1.23396881e-01 -1.77212387e-01 -4.11563963e-01 7.00106621e-01
8.48086596e-01 -1.89243063e-01 -3.57669860e-01 1.53014541e+00
3.61919463e-01 3.40546072e-01 1.48738742e-01 5.66019356e-01
2.80272633e-01 5.04395843e-01 1.26839012e-01 6.91746175e-02
1.28573513e+00 -3.14091563e-01 -4.86368895e-01 1.48858771e-01
5.47255814e-01 -9.98951018e-01 9.27647531e-01 5.30077577e-01
-9.23199594e-01 -3.94109696e-01 -1.26994753e+00 4.36299354e-01
-8.06137204e-01 -5.06532371e-01 4.09668982e-01 1.48784876e+00
-7.40809500e-01 7.82893181e-01 -5.38792610e-01 -1.93124741e-01
4.64938343e-01 4.99877781e-01 -5.62545896e-01 -1.58542451e-02
-1.43213761e+00 9.60964799e-01 4.86809134e-01 -1.19467698e-01
-1.14627147e+00 -5.40304363e-01 -4.75891531e-01 6.67505786e-02
3.99809450e-01 -1.42204568e-01 8.66948843e-01 -8.27382445e-01
-9.84009147e-01 1.02595472e+00 4.86569792e-01 -8.72734368e-01
1.09897649e+00 -2.04917900e-02 -4.88136679e-01 4.86512423e-01
-1.03072807e-01 2.18905330e-01 1.05340254e+00 -1.23636556e+00
-1.87998742e-01 -2.85798818e-01 7.80493855e-01 -4.85320419e-01
-8.44500065e-01 3.60571563e-01 1.33292153e-02 -4.99146342e-01
-7.35338867e-01 -9.64335442e-01 -1.26158848e-01 1.91167414e-01
-4.82925445e-01 2.89600462e-01 1.19347346e+00 -5.21574080e-01
7.77267039e-01 -2.23098421e+00 -3.59675884e-01 2.52002865e-01
4.91437949e-02 9.24124837e-01 -6.11032993e-02 5.50428748e-01
-3.12992573e-01 4.04831827e-01 -4.33669612e-03 1.24200203e-01
-6.26517236e-02 5.17116636e-02 -5.91266990e-01 1.02155423e+00
2.59010464e-01 5.33441842e-01 -8.32002521e-01 -5.47551751e-01
5.06585538e-01 5.04121602e-01 -3.29870254e-01 1.69097856e-01
-5.80109283e-02 4.12220687e-01 -3.62331569e-01 3.36305231e-01
9.91031051e-01 2.13801607e-01 1.11323446e-01 -1.89995453e-01
3.90969999e-02 -3.84775409e-03 -1.28088498e+00 9.60601807e-01
-4.59896356e-01 9.75360572e-01 4.41909693e-02 -9.56239820e-01
8.05893958e-01 3.48320514e-01 5.20216599e-02 -5.17740548e-01
1.88741878e-01 2.89016128e-01 2.17541829e-01 -5.75094521e-01
3.39945197e-01 -1.64544344e-01 7.27746785e-02 6.00825429e-01
1.25269860e-01 3.64270151e-01 2.26209208e-01 2.17914015e-01
1.26927912e+00 -7.38454983e-02 1.04955778e-01 -1.23381115e-01
9.32819545e-01 -2.44552195e-01 -2.05140765e-04 1.02797246e+00
-2.91116685e-01 2.75010675e-01 6.89760685e-01 -4.35009360e-01
-1.27277088e+00 -1.04243541e+00 1.51058044e-02 6.38649881e-01
2.99533717e-02 -2.48374581e-01 -1.05953479e+00 -1.04697418e+00
-2.53862619e-01 7.29332864e-01 -5.72215259e-01 -3.34269702e-01
-9.12724435e-01 -5.76811194e-01 1.35892200e+00 3.24786723e-01
6.55527174e-01 -1.03249848e+00 -6.58771873e-01 -8.89156759e-03
9.65500921e-02 -1.56901240e+00 1.62145600e-01 -8.21982399e-02
-7.38565087e-01 -1.50554097e+00 -6.08280778e-01 -2.51226574e-01
3.82037222e-01 3.15400690e-01 7.93271661e-01 4.12586838e-01
-3.72839987e-01 4.04713035e-01 -4.72872585e-01 -4.35744196e-01
-1.21463025e+00 -8.40322524e-02 -1.01836167e-01 -1.87841401e-01
2.00645030e-01 -5.89877069e-01 -3.45948488e-01 2.72659540e-01
-1.41605186e+00 -9.61538553e-01 3.64173859e-01 3.50479960e-01
-2.02514261e-01 4.80991811e-01 4.28150624e-01 -8.82197320e-01
5.34743071e-01 -5.34174442e-01 -7.89788544e-01 1.49672046e-01
-2.88468301e-01 -7.32340217e-02 9.40550625e-01 -7.79572487e-01
-7.40062833e-01 -4.08115089e-01 -2.67256171e-01 -6.95611596e-01
-4.36711341e-01 7.71285594e-02 -4.46687222e-01 -6.99350417e-01
9.61185396e-01 2.42613535e-02 -1.21544905e-01 -2.82224566e-01
3.46526712e-01 3.81850749e-01 4.83307451e-01 -6.32296741e-01
1.38346446e+00 6.74472868e-01 2.97924101e-01 -8.43840718e-01
-2.79338241e-01 -1.21454708e-01 -4.51719522e-01 -4.38428015e-01
5.54997683e-01 -4.55819637e-01 -8.34625661e-01 5.54128468e-01
-1.41368830e+00 4.43863608e-02 -1.38995890e-02 2.17495620e-01
-3.91148925e-01 1.10366917e+00 -6.59932494e-01 -7.97250628e-01
-8.58613849e-02 -1.56003213e+00 6.01221740e-01 -1.39384344e-01
7.61808753e-02 -1.10617256e+00 -2.03480184e-01 3.36713791e-01
4.16754425e-01 6.11986518e-01 9.90716398e-01 -1.08026576e+00
-5.77822864e-01 -5.55846930e-01 -1.81110084e-01 6.55631900e-01
-1.53190389e-01 1.69058383e-01 -1.16830981e+00 -4.30335432e-01
3.74327362e-01 -2.82878056e-02 4.25416797e-01 -1.64105415e-01
1.00564170e+00 -4.40219074e-01 1.63782984e-01 4.18669045e-01
1.64246058e+00 4.85164113e-02 1.25824499e+00 9.89936113e-01
5.07716000e-01 7.02998579e-01 4.77636099e-01 2.49097064e-01
-4.13064331e-01 6.21690512e-01 9.47688937e-01 -2.69622579e-02
-1.01204701e-01 1.85971901e-01 5.97779572e-01 -1.79228615e-02
4.99931499e-02 -4.15392160e-01 -6.70055747e-01 2.64352202e-01
-1.17622733e+00 -1.25237632e+00 -3.52430731e-01 2.40734506e+00
4.80901212e-01 5.33185780e-01 3.10394019e-01 6.09379828e-01
1.18099272e+00 2.61073172e-01 -1.47636086e-01 -7.89686143e-01
-1.55968562e-01 3.67589027e-01 7.76834190e-01 2.02713609e-01
-1.13435924e+00 8.18751037e-01 6.14930630e+00 1.00005305e+00
-1.27262592e+00 2.80875504e-01 3.41932535e-01 1.42655984e-01
1.66790545e-01 2.21031860e-01 -5.61451614e-01 6.36501372e-01
1.33878529e+00 -1.71567887e-01 1.07231781e-01 8.87940705e-01
4.53909002e-02 2.03275964e-01 -8.61697793e-01 4.83746260e-01
-3.61402258e-02 -1.48695123e+00 1.60115600e-01 5.61526537e-01
6.49042279e-02 -4.48787123e-01 1.60302714e-01 1.34307191e-01
7.92101249e-02 -8.58208954e-01 6.36129022e-01 -1.12766176e-01
4.77268010e-01 -1.06064188e+00 8.87045681e-01 2.54368246e-01
-6.50131762e-01 -4.44278680e-02 -4.77808803e-01 2.52250493e-01
1.97597221e-01 4.59820360e-01 -9.74504173e-01 6.48342848e-01
5.07043362e-01 -1.00720078e-01 -6.59311831e-01 9.73409832e-01
-1.98269412e-01 8.40524435e-01 -2.34194010e-01 3.35037708e-01
5.15188992e-01 1.58849329e-01 7.68633902e-01 1.49641109e+00
2.19609037e-01 -4.59795028e-01 -2.53255904e-01 7.83234358e-01
-1.31146148e-01 6.24557808e-02 -1.00222743e+00 -2.29896586e-02
3.83814216e-01 1.10390890e+00 -9.72219586e-01 -7.20070256e-03
-2.55291134e-01 7.89094329e-01 -1.37272224e-01 -4.38269526e-02
-1.00708032e+00 -2.38904744e-01 4.36368346e-01 4.79318619e-01
3.91072571e-01 -2.24548161e-01 2.02768579e-01 -9.14846778e-01
4.84215394e-02 -1.24942350e+00 4.39361989e-01 -3.87938678e-01
-1.31683600e+00 5.28265953e-01 2.06121996e-01 -1.33841538e+00
-1.74128756e-01 -8.62324297e-01 -8.89403999e-01 5.21906614e-01
-1.57255661e+00 -1.30025470e+00 1.01744965e-01 8.75919521e-01
2.64727920e-01 -4.66433585e-01 6.81999266e-01 5.65764964e-01
-2.22038865e-01 8.84210050e-01 -2.40232050e-01 4.85303372e-01
7.80161142e-01 -9.15146410e-01 4.69311595e-01 1.25947928e+00
1.05637692e-01 7.35083520e-01 1.08110559e+00 -6.17630005e-01
-1.33448768e+00 -9.80383158e-01 2.38633499e-01 -3.07394743e-01
9.85093534e-01 -3.68745238e-01 -1.11754882e+00 4.79973823e-01
3.13923448e-01 1.30499508e-02 6.81608856e-01 -3.39652717e-01
-8.66406024e-01 1.57582134e-01 -1.56196344e+00 5.98752379e-01
2.83226907e-01 -7.13769555e-01 -5.00525534e-01 5.41014612e-01
7.81368315e-01 1.05788507e-01 -7.68938780e-01 1.57676622e-01
3.57593805e-01 -1.33459449e+00 1.42717314e+00 -6.94597542e-01
3.66024315e-01 -2.51867741e-01 -2.36595258e-01 -6.65803552e-01
3.03328365e-01 -7.10625827e-01 -1.84162349e-01 1.46972263e+00
-8.59962478e-02 -6.43210948e-01 5.61899781e-01 1.97032720e-01
3.16932946e-01 -1.01807810e-01 -1.11902761e+00 -1.28367710e+00
5.55474520e-01 -6.01165056e-01 6.01155221e-01 1.15430033e+00
-6.02027357e-01 -3.45663935e-01 -3.08629394e-01 5.40602505e-01
7.84399569e-01 -4.78174835e-01 9.23969328e-01 -9.59498882e-01
-3.98265362e-01 -2.81513989e-01 -1.01144373e+00 -1.59537103e-02
3.12377095e-01 -5.09056151e-01 -5.77780783e-01 -5.20630360e-01
-6.41464517e-02 -2.61530608e-01 -1.84585929e-01 9.85354856e-02
2.08804179e-02 3.96430194e-01 6.63750052e-01 2.19304591e-01
-2.47523651e-01 -1.48989320e-01 7.73103058e-01 -5.08299433e-02
3.28927845e-01 1.04143033e-02 -4.90449190e-01 8.21861625e-01
7.50708282e-01 -9.06236410e-01 -2.13683724e-01 1.31864203e-02
1.51757196e-01 -2.76697814e-01 8.20567966e-01 -1.20973146e+00
1.03486009e-01 -8.41766968e-02 3.78096476e-02 -2.71534890e-01
2.09164903e-01 -1.02264059e+00 4.60750312e-02 6.61710680e-01
-1.24551140e-01 1.13976590e-01 3.85372609e-01 5.13398230e-01
2.97010858e-02 -6.92763686e-01 9.83471692e-01 -3.35549027e-01
-5.80424190e-01 3.40228826e-02 -5.30048966e-01 -6.51242882e-02
1.19364429e+00 -3.76020491e-01 -5.99996686e-01 -5.08795023e-01
-4.91982013e-01 -4.82511997e-01 7.41866648e-01 3.40786092e-02
3.65257889e-01 -9.70510244e-01 -6.20025933e-01 -2.32225448e-01
-1.26544446e-01 -7.12557435e-01 2.38727048e-01 3.89029115e-01
-8.38200569e-01 1.31286718e-02 -5.78255892e-01 -2.13062897e-01
-1.42904186e+00 1.22704935e+00 3.11666250e-01 -3.50743026e-01
-5.56667864e-01 2.89890647e-01 9.63653922e-02 -1.14214547e-01
1.86296657e-01 6.73497438e-01 -1.91482492e-02 3.49863730e-02
7.59837508e-01 6.18513346e-01 2.37626687e-01 -6.42776191e-01
-3.54238480e-01 1.04330130e-01 -9.46309492e-02 -1.22937791e-01
1.32516229e+00 1.67378068e-01 -9.53667760e-02 -7.70739987e-02
1.36929095e+00 3.76178175e-01 -8.53954315e-01 1.04301989e-01
1.42449709e-02 -7.83974111e-01 -6.69791102e-02 -4.86000061e-01
-1.38517761e+00 1.14530492e+00 6.95135355e-01 5.76790392e-01
1.00985789e+00 -2.39091009e-01 7.83460021e-01 4.33450937e-01
3.98184389e-01 -7.22434163e-01 2.44643122e-01 2.41624434e-02
5.42714655e-01 -8.43371987e-01 1.29338667e-01 -4.18672591e-01
-2.34416619e-01 1.55292475e+00 2.67713189e-01 -5.56576252e-01
3.69059622e-01 6.35497928e-01 -2.47445516e-02 -1.23166554e-01
-8.27249438e-02 1.39037117e-01 -2.76071727e-01 8.50832164e-01
8.53098184e-02 -3.20463181e-01 -2.07044899e-01 -7.05971057e-03
1.42996265e-02 -3.44299585e-01 1.11236322e+00 1.07071364e+00
-1.75227612e-01 -1.46371984e+00 -1.06805027e+00 -2.09886774e-01
-9.01922882e-01 6.12919964e-02 -5.35026371e-01 1.34369552e+00
2.09890112e-01 8.90782714e-01 -3.25415283e-01 -2.62250334e-01
1.63170964e-01 -6.17640242e-02 4.88254964e-01 -7.69178942e-02
-1.03099310e+00 -1.26675993e-01 3.56334671e-02 -5.27740538e-01
-3.97829026e-01 -5.54959059e-01 -7.61520386e-01 -8.55001569e-01
-5.02913117e-01 -8.63670707e-02 9.15803790e-01 9.30604875e-01
7.60296062e-02 3.02727818e-01 7.51477182e-01 -7.34201968e-01
-7.87624300e-01 -4.39801544e-01 -4.14990604e-01 4.90405202e-01
1.90814644e-01 -6.35147631e-01 -6.54682636e-01 1.70597732e-02] | [5.623366832733154, 7.763838291168213] |
5462215e-9e05-4e7d-8da6-31161c115e16 | supervoice-text-independent-speaker | 2205.14496 | null | https://arxiv.org/abs/2205.14496v1 | https://arxiv.org/pdf/2205.14496v1.pdf | SuperVoice: Text-Independent Speaker Verification Using Ultrasound Energy in Human Speech | Voice-activated systems are integrated into a variety of desktop, mobile, and Internet-of-Things (IoT) devices. However, voice spoofing attacks, such as impersonation and replay attacks, in which malicious attackers synthesize the voice of a victim or simply replay it, have brought growing security concerns. Existing speaker verification techniques distinguish individual speakers via the spectrographic features extracted from an audible frequency range of voice commands. However, they often have high error rates and/or long delays. In this paper, we explore a new direction of human voice research by scrutinizing the unique characteristics of human speech at the ultrasound frequency band. Our research indicates that the high-frequency ultrasound components (e.g. speech fricatives) from 20 to 48 kHz can significantly enhance the security and accuracy of speaker verification. We propose a speaker verification system, SUPERVOICE that uses a two-stream DNN architecture with a feature fusion mechanism to generate distinctive speaker models. To test the system, we create a speech dataset with 12 hours of audio (8,950 voice samples) from 127 participants. In addition, we create a second spoofed voice dataset to evaluate its security. In order to balance between controlled recordings and real-world applications, the audio recordings are collected from two quiet rooms by 8 different recording devices, including 7 smartphones and an ultrasound microphone. Our evaluation shows that SUPERVOICE achieves 0.58% equal error rate in the speaker verification task, it only takes 120 ms for testing an incoming utterance, outperforming all existing speaker verification systems. Moreover, within 91 ms processing time, SUPERVOICE achieves 0% equal error rate in detecting replay attacks launched by 5 different loudspeakers. | ['Eric J. Hunter', 'Li Xiao', 'Ying Zhu', 'Nikolay Ivanov', 'Qiben Yan', 'Hanqing Guo'] | 2022-05-28 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [ 1.19298734e-01 -3.30096215e-01 1.46682963e-01 -1.38619408e-01
-7.14152098e-01 -7.37382412e-01 3.30939472e-01 -1.88475236e-01
-2.11942479e-01 2.61883706e-01 3.33587676e-01 -5.91741979e-01
3.39252293e-01 -4.32975084e-01 -2.81468928e-01 -6.33034348e-01
-6.95661157e-02 -3.30962211e-01 1.57880858e-01 4.02276916e-03
-7.12203095e-03 4.36947197e-01 -1.70384812e+00 1.89558536e-01
6.21285856e-01 1.09807765e+00 -7.50096068e-02 1.07107663e+00
1.06087968e-01 8.57924223e-02 -1.30348504e+00 -2.79178053e-01
2.63272338e-02 -2.77022123e-01 -4.91836071e-01 -1.60283968e-01
6.55303225e-02 -5.19949079e-01 -4.51391369e-01 1.22690022e+00
1.11638808e+00 -1.99354924e-02 8.98134783e-02 -1.32658195e+00
-3.62639040e-01 7.85990477e-01 -6.45395070e-02 4.21857327e-01
8.41389716e-01 2.48843998e-01 6.14164412e-01 -7.17031240e-01
-8.24555978e-02 1.35213864e+00 5.65897405e-01 7.25275457e-01
-8.08715820e-01 -1.15156424e+00 -4.13394690e-01 1.35889590e-01
-1.46403933e+00 -1.03014100e+00 9.92943704e-01 -1.93234548e-01
7.22960055e-01 6.29208326e-01 2.68777013e-01 1.54676938e+00
1.90304562e-01 2.69172847e-01 7.77772546e-01 -3.45168293e-01
1.42395407e-01 3.55617225e-01 2.75208414e-01 2.42123410e-01
-1.92639828e-02 3.32819015e-01 -6.21644616e-01 -5.94826519e-01
2.16001928e-01 -1.09907530e-01 -7.47833431e-01 5.80136299e-01
-1.28799081e+00 6.39736474e-01 -1.56037182e-01 4.03278530e-01
-1.83492780e-01 -2.53416121e-01 5.65912187e-01 3.58008236e-01
-1.55626805e-02 6.56587034e-02 -1.61276266e-01 -3.21442157e-01
-6.25579119e-01 -2.23640501e-01 1.02306533e+00 6.48345888e-01
8.32201913e-02 5.86459160e-01 1.63751319e-01 9.04593289e-01
6.37779832e-01 9.72077906e-01 9.53360736e-01 -4.71229017e-01
4.21445698e-01 -1.94102421e-01 -4.90964316e-02 -1.10930455e+00
-1.79231137e-01 -9.58673880e-02 -6.21416688e-01 -2.51155913e-01
1.77127406e-01 -4.74191576e-01 -5.17693102e-01 1.56940043e+00
4.94819492e-01 4.36406314e-01 2.93014914e-01 7.32918024e-01
1.02820528e+00 7.46859848e-01 -3.15281719e-01 -4.20480341e-01
1.75839734e+00 -7.20324993e-01 -1.14642322e+00 1.42363220e-01
3.97832226e-03 -1.18625522e+00 1.30106258e+00 5.19395888e-01
-6.23414218e-01 -8.52904677e-01 -1.08491337e+00 4.81257975e-01
-2.20059067e-01 -1.33448064e-01 -4.20560688e-02 1.66257143e+00
-7.74145603e-01 9.41036493e-02 -5.59421480e-01 3.85014415e-02
-6.30279928e-02 3.06881875e-01 -3.16749603e-01 3.76838624e-01
-1.53101385e+00 2.06492335e-01 -1.74948469e-01 1.01071801e-02
-1.02794373e+00 -5.69688618e-01 -8.51361096e-01 1.22013323e-01
1.26818623e-02 -1.23621441e-01 1.47588742e+00 -4.84184474e-01
-1.94028592e+00 3.81551713e-01 -3.97043824e-01 -4.27753866e-01
1.82220727e-01 -2.17726782e-01 -1.59883654e+00 3.17296296e-01
-1.54678553e-01 -8.47394094e-02 1.38079190e+00 -8.89685929e-01
-3.43326986e-01 -1.21849664e-01 -4.30742621e-01 -3.16488177e-01
-5.73465765e-01 5.95771432e-01 3.29890490e-01 -6.42163396e-01
7.54788592e-02 -7.94126332e-01 4.35955495e-01 -6.01453781e-01
-6.71120882e-01 6.40166774e-02 1.27958190e+00 -8.62828732e-01
1.33249664e+00 -2.63947034e+00 -6.48595035e-01 2.42885038e-01
-4.31833714e-02 7.05846906e-01 7.02040344e-02 2.64019996e-01
-6.02242015e-02 3.92679334e-01 4.11341861e-02 -3.17682534e-01
-6.24038801e-02 -1.53863683e-01 -7.44495809e-01 5.80816984e-01
-2.36768007e-01 2.76783764e-01 -5.85830092e-01 -1.54259235e-01
1.63241714e-01 8.49807262e-01 -3.35815430e-01 2.37487048e-01
6.28304183e-01 3.47794741e-01 -2.02476129e-01 8.04245234e-01
7.91248381e-01 4.99850124e-01 -1.80600509e-01 -1.40838012e-01
-6.12337291e-02 8.41403902e-01 -1.34825683e+00 1.03917372e+00
-4.25295860e-01 7.17256367e-01 5.85003674e-01 -5.04601836e-01
1.02952397e+00 1.05054641e+00 1.10847831e-01 -1.75639063e-01
4.53541487e-01 3.69634688e-01 1.76683068e-01 -6.44649029e-01
3.01954210e-01 9.99619998e-03 -3.45921963e-02 5.08208513e-01
-1.31609170e-02 -1.01997092e-01 -4.45441961e-01 -2.01879770e-01
1.07574761e+00 -8.58771265e-01 9.21698511e-02 -1.23544931e-02
9.74102139e-01 -9.84404385e-01 3.70270163e-01 6.21480048e-01
-8.04274142e-01 4.80693340e-01 -1.07083797e-01 1.25721022e-01
-4.04023796e-01 -1.12169170e+00 -2.26800010e-01 8.89458716e-01
-4.12891284e-02 -4.39509451e-01 -9.41889107e-01 -5.30173242e-01
-1.50011301e-01 6.27667010e-01 1.13895610e-01 -3.15782040e-01
-5.19083619e-01 -1.98263213e-01 1.45768619e+00 7.19199628e-02
6.28533959e-01 -1.21537781e+00 -3.68796438e-01 2.17781678e-01
-3.57587487e-01 -1.38868618e+00 -9.27937508e-01 -1.76226035e-01
-3.49666387e-01 -7.98689961e-01 -4.64681804e-01 -8.13487053e-01
4.03968245e-02 5.59245229e-01 5.55551112e-01 -1.92980707e-01
-1.68851405e-01 4.28969413e-01 -3.37250441e-01 -6.49632752e-01
-9.07381177e-01 -1.02235250e-01 8.57287884e-01 2.60609120e-01
4.77006197e-01 -5.80243409e-01 -2.41780534e-01 7.31336355e-01
-7.37625718e-01 -8.70689869e-01 -5.14309630e-02 6.14459455e-01
-8.82453620e-02 4.99351352e-01 8.99280906e-01 -3.72010432e-02
1.09633255e+00 -3.62555474e-01 -2.86864579e-01 -9.73895788e-02
-1.90553993e-01 -5.67365825e-01 6.33924425e-01 -9.29477453e-01
-9.80585575e-01 -2.55383164e-01 -6.04993701e-01 -2.75484592e-01
-5.67173123e-01 2.12030366e-01 -5.99850655e-01 -9.33416560e-02
6.81644619e-01 4.66445357e-01 2.73221374e-01 -4.54669595e-01
4.67586555e-02 1.71449780e+00 6.98528230e-01 -1.76161140e-01
8.70774508e-01 1.83967739e-01 -8.30058634e-01 -1.59015954e+00
2.10419092e-02 -5.53364635e-01 1.05425715e-01 -2.06382766e-01
5.46679378e-01 -8.96785200e-01 -1.30515492e+00 1.02851081e+00
-1.33741653e+00 2.90994614e-01 1.59605488e-01 9.08861220e-01
-6.41513336e-03 7.09446609e-01 -8.54608655e-01 -1.20870900e+00
-7.24495292e-01 -1.33776999e+00 8.71448576e-01 2.07988337e-01
-4.10429478e-01 -4.46397394e-01 -1.47188783e-01 6.40088439e-01
6.39582634e-01 -3.25114995e-01 3.90241832e-01 -1.01506972e+00
1.37474194e-01 -2.43931815e-01 4.24834251e-01 6.06800675e-01
7.03416526e-01 5.86344860e-02 -1.58007562e+00 -2.51169384e-01
7.12895393e-01 -5.14072133e-04 2.53750026e-01 2.53852338e-01
8.82764399e-01 -5.84361434e-01 -3.75674292e-02 3.77843678e-01
5.09911716e-01 7.40524352e-01 6.16798162e-01 -2.00132132e-01
5.10159731e-01 5.75663388e-01 7.95106888e-02 2.76446640e-01
-5.35333529e-02 5.16684294e-01 2.79794306e-01 3.90668184e-01
-1.14976466e-01 -1.05748579e-01 9.73165154e-01 1.15298593e+00
3.51793081e-01 -3.96729380e-01 -7.21809089e-01 3.39179128e-01
-7.47838259e-01 -1.02822137e+00 -7.01946393e-02 2.39840436e+00
8.29541028e-01 1.91448361e-01 2.62707323e-01 9.08266246e-01
1.22842038e+00 3.42451602e-01 -3.56890559e-01 -6.03906572e-01
1.28988221e-01 2.37372369e-01 1.95662692e-01 7.51968324e-01
-9.97398019e-01 5.18011332e-01 6.18999577e+00 6.30139828e-01
-1.61123824e+00 1.25112280e-01 1.88529953e-01 -6.48039102e-04
-1.39327362e-01 -6.48302376e-01 -7.99574018e-01 7.82606781e-01
1.48001492e+00 2.01339144e-02 5.24845600e-01 9.07308400e-01
5.08316278e-01 3.66989762e-01 -8.39653313e-01 1.15201330e+00
2.03376055e-01 -7.49339700e-01 -1.79930508e-01 2.40031570e-01
-5.60797937e-02 -9.43865720e-03 2.05965012e-01 1.56018257e-01
-1.79966971e-01 -1.00415134e+00 7.49768615e-01 -1.35699868e-01
7.61413217e-01 -7.97200799e-01 7.66016543e-01 3.14419538e-01
-1.28924000e+00 -8.92171934e-02 -2.38282718e-02 -2.33656634e-02
3.94403160e-01 6.84549093e-01 -1.19101191e+00 3.67865175e-01
6.73401475e-01 -6.84153102e-03 -6.68351948e-02 7.31292844e-01
-1.48339972e-01 1.12556350e+00 -4.63096827e-01 -1.59877345e-01
-3.21109980e-01 3.26133430e-01 9.10770595e-01 1.19896758e+00
5.53022921e-01 -3.97148691e-02 -2.52251923e-01 5.87510824e-01
-3.83637706e-03 -1.50228903e-01 -6.36927903e-01 -1.23449944e-01
1.05415499e+00 1.00868499e+00 -2.82038808e-01 -5.58390245e-02
-7.75903165e-02 8.19841981e-01 -7.31319547e-01 4.51419085e-01
-1.04605794e+00 -9.78166223e-01 9.16731238e-01 -1.57604381e-01
5.45956455e-02 -3.17544043e-01 3.16342227e-02 -9.75357175e-01
1.55549034e-01 -1.33627701e+00 -5.89339510e-02 -4.58053231e-01
-1.10233212e+00 7.16985226e-01 -4.88494098e-01 -1.17605388e+00
-3.10698599e-01 -4.14920807e-01 -8.02892148e-01 9.59120691e-01
-1.22181141e+00 -5.05799472e-01 -1.98860928e-01 8.47332716e-01
5.61032474e-01 -3.53841305e-01 1.08501041e+00 6.46875262e-01
-6.19925797e-01 9.11759079e-01 -2.16158196e-01 4.54870611e-01
6.56686664e-01 -6.55529797e-01 8.12675297e-01 8.48433435e-01
1.33659348e-01 1.01008344e+00 7.02370405e-01 -3.73014271e-01
-1.38880599e+00 -7.27380931e-01 1.10103297e+00 -1.79340973e-01
4.91455466e-01 -5.49793959e-01 -9.25762951e-01 2.35414460e-01
1.03788488e-01 7.14038834e-02 1.08488750e+00 -4.12528455e-01
-5.10956526e-01 -2.18182474e-01 -1.46162212e+00 3.88025671e-01
5.87619483e-01 -1.13814735e+00 -7.12996006e-01 -1.42319933e-01
1.10227704e+00 -1.43732190e-01 -6.78466558e-01 1.01689510e-01
7.71266222e-01 -9.66458380e-01 1.02484584e+00 -1.93187878e-01
-2.22647294e-01 -3.78215879e-01 -3.77345383e-01 -1.13231719e+00
2.11335972e-01 -1.21080029e+00 -2.02926740e-01 1.83922732e+00
2.43073642e-01 -1.16626167e+00 3.29343349e-01 2.31177121e-01
6.52989000e-02 -7.94037431e-02 -1.18834972e+00 -9.38416600e-01
-4.05667633e-01 -8.42960298e-01 1.01632869e+00 9.00457084e-01
2.70271748e-01 2.88654804e-01 -4.04720575e-01 6.39420152e-01
4.44391280e-01 -6.71026647e-01 7.39987791e-01 -9.93976831e-01
-1.35655925e-01 -2.74070382e-01 -3.50837648e-01 -1.13278806e+00
1.00415766e-01 -4.36210126e-01 8.94960165e-02 -6.14803135e-01
-6.43859267e-01 -8.90477821e-02 -1.76509261e-01 3.36602181e-02
4.94736023e-02 -7.54228160e-02 6.86513558e-02 -2.20207125e-02
3.01907092e-01 5.49719930e-01 5.94440162e-01 -3.28882068e-01
-4.04348075e-01 5.62026381e-01 -4.02922243e-01 8.54252219e-01
8.68427455e-01 -3.84135514e-01 -4.73263234e-01 -5.94263822e-02
-5.27440608e-01 2.94115335e-01 2.75667876e-01 -1.11428320e+00
2.22435102e-01 2.24614099e-01 -1.37777746e-01 -3.29392254e-01
4.56484079e-01 -9.16549385e-01 6.47628382e-02 6.84009254e-01
-1.59739509e-01 -1.91932786e-02 1.44701079e-01 3.97088826e-01
-2.95432508e-01 -1.73957482e-01 5.89711189e-01 2.39555225e-01
-1.03497274e-01 -1.30297512e-01 -7.78032780e-01 -2.73124814e-01
6.45526528e-01 -1.95171893e-01 -4.29955393e-01 -6.07833743e-01
-4.53134090e-01 -4.38817501e-01 -4.73184027e-02 5.69783092e-01
7.91208684e-01 -1.17793238e+00 -3.92640293e-01 8.36875260e-01
-2.86618590e-01 -3.28447700e-01 2.41209149e-01 5.38841844e-01
-1.57227278e-01 4.51789886e-01 2.15680793e-01 -6.80709243e-01
-1.75092554e+00 4.21595871e-01 4.33858007e-01 4.58513677e-01
-3.77509713e-01 8.06722581e-01 -3.26650441e-01 -4.56447572e-01
5.92230082e-01 -4.56728727e-01 -2.23439470e-01 1.35029599e-01
1.04772615e+00 6.12154186e-01 3.37451965e-01 -8.24911714e-01
-7.30940580e-01 2.39726916e-01 1.65416211e-01 -4.74896967e-01
5.57400167e-01 -2.43973091e-01 1.06692798e-01 3.21795017e-01
1.38277435e+00 7.48199999e-01 -5.56977749e-01 -5.32674752e-02
-2.49549523e-01 -3.79746705e-01 -1.84258353e-02 -4.27196264e-01
-8.51985395e-01 7.92989910e-01 8.58377218e-01 8.49191666e-01
8.64849269e-01 -3.38069946e-01 1.30557954e+00 2.51080871e-01
4.23414707e-01 -6.90865040e-01 1.74722020e-02 4.50240821e-01
6.73493564e-01 -9.44990337e-01 -6.21645689e-01 -5.21114945e-01
-4.22194332e-01 1.04719889e+00 1.87798619e-01 3.36003780e-01
9.53012884e-01 4.19544607e-01 5.29603779e-01 2.11587459e-01
-2.13682860e-01 3.35354894e-01 8.97164047e-02 8.10556173e-01
2.76801914e-01 1.61149606e-01 5.86160384e-02 9.20612335e-01
-9.73194301e-01 -3.42883646e-01 6.58395171e-01 6.90889478e-01
-4.44346219e-01 -8.49762321e-01 -1.01816916e+00 3.31166969e-03
-1.02893519e+00 -1.37271613e-01 -3.26908857e-01 1.07578471e-01
-2.28261039e-01 1.97202766e+00 -5.45889884e-02 -9.33690131e-01
3.25542063e-01 3.01378429e-01 -3.16606879e-01 -3.14082861e-01
-6.82008505e-01 2.22131461e-01 5.66052012e-02 -3.88449728e-01
-2.87991226e-01 -5.86156011e-01 -1.23590612e+00 -5.78914106e-01
-6.21313989e-01 4.30437684e-01 1.04404199e+00 8.02812696e-01
3.51877213e-01 6.22804582e-01 1.16331017e+00 -6.66739106e-01
-9.93402779e-01 -1.13835227e+00 -7.22224891e-01 -5.28091714e-02
1.00227261e+00 -2.90209413e-01 -1.01285815e+00 2.36567743e-02] | [14.083671569824219, 5.888390064239502] |
0b450840-7b2e-4bc6-b197-7e6e1fc16892 | a-natural-language-corpus-of-common-grounding | 1907.03399 | null | https://arxiv.org/abs/1907.03399v1 | https://arxiv.org/pdf/1907.03399v1.pdf | A Natural Language Corpus of Common Grounding under Continuous and Partially-Observable Context | Common grounding is the process of creating, repairing and updating mutual understandings, which is a critical aspect of sophisticated human communication. However, traditional dialogue systems have limited capability of establishing common ground, and we also lack task formulations which introduce natural difficulty in terms of common grounding while enabling easy evaluation and analysis of complex models. In this paper, we propose a minimal dialogue task which requires advanced skills of common grounding under continuous and partially-observable context. Based on this task formulation, we collected a largescale dataset of 6,760 dialogues which fulfills essential requirements of natural language corpora. Our analysis of the dataset revealed important phenomena related to common grounding that need to be considered. Finally, we evaluate and analyze baseline neural models on a simple subtask that requires recognition of the created common ground. We show that simple baseline models perform decently but leave room for further improvement. Overall, we show that our proposed task will be a fundamental testbed where we can train, evaluate, and analyze dialogue system's ability for sophisticated common grounding. | ['Akiko Aizawa', 'Takuma Udagawa'] | 2019-07-08 | null | null | null | null | ['goal-oriented-dialog', 'dialogue-understanding'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.38576069e-01 6.98771894e-01 1.23794898e-01 -4.77627873e-01
-6.51802301e-01 -7.54887402e-01 1.11673617e+00 3.96615863e-01
-2.88034737e-01 8.96023333e-01 4.88033354e-01 -6.85581803e-01
1.64263561e-01 -5.20599008e-01 -4.24744010e-01 -1.74124211e-01
-2.88290903e-02 7.83282161e-01 2.65259564e-01 -9.09490943e-01
2.54123569e-01 -1.01880701e-02 -9.74119008e-01 4.55068588e-01
9.45428014e-01 5.27705193e-01 2.47688785e-01 9.52522039e-01
-1.01806901e-01 1.14204729e+00 -1.07037103e+00 -6.80359244e-01
-7.21965879e-02 -7.14298666e-01 -1.75899577e+00 3.63729615e-03
4.72987115e-01 -6.89276755e-01 -2.24996626e-01 8.32050383e-01
3.08603197e-01 2.63139993e-01 3.81856441e-01 -1.37378001e+00
-2.11511865e-01 7.77347147e-01 2.80768722e-01 1.74187750e-01
8.84046733e-01 2.96836138e-01 1.49392009e+00 -3.70990694e-01
6.68335199e-01 1.20554709e+00 7.45376825e-01 9.33673739e-01
-1.23034918e+00 -1.95634231e-01 1.65093422e-01 -1.71122238e-01
-6.97031379e-01 -8.15781713e-01 3.29416245e-01 -3.74037921e-01
1.27260697e+00 4.19178069e-01 4.87392753e-01 1.26635969e+00
1.03319548e-01 8.99171174e-01 1.15794885e+00 -5.45706868e-01
-1.79605424e-01 2.20779195e-01 7.85793722e-01 8.84368479e-01
-1.98330544e-02 -2.33765826e-01 -8.28106403e-01 -3.05874139e-01
4.85757411e-01 -5.90729475e-01 -3.98445725e-01 -1.08790686e-02
-1.22399032e+00 7.98935711e-01 6.41041249e-03 4.82980996e-01
3.35296169e-02 -1.14675373e-01 3.86195481e-01 6.78498447e-01
1.25650600e-01 1.16787195e+00 -5.93160272e-01 -7.33486593e-01
-5.07453024e-01 4.45771813e-01 1.66876757e+00 8.72208059e-01
6.04327321e-01 -3.41853321e-01 -2.74117708e-01 7.70378172e-01
1.61871642e-01 8.67447928e-02 3.96078914e-01 -1.26796448e+00
7.24144816e-01 8.04200768e-01 3.27299833e-01 -1.00393808e+00
-8.34999681e-01 -4.10477608e-01 -5.81791759e-01 -2.83613086e-01
8.67128253e-01 -3.41986239e-01 -2.42316976e-01 2.09951186e+00
2.28444915e-02 -1.69550270e-01 4.73997474e-01 4.15648431e-01
1.00644231e+00 4.65919733e-01 -7.39260018e-02 -1.51380956e-01
1.27602112e+00 -9.89103496e-01 -9.71800566e-01 -2.90469289e-01
1.16755033e+00 -5.66617191e-01 1.34230733e+00 4.02007908e-01
-1.31418598e+00 -2.83767939e-01 -1.16062343e+00 -2.71302074e-01
-1.03406340e-01 -2.07676649e-01 1.01226723e+00 5.89564562e-01
-1.30586410e+00 7.36184359e-01 -5.39556801e-01 -5.80133557e-01
-3.07755142e-01 1.71050355e-01 -4.49151397e-01 4.12611395e-01
-1.54281151e+00 1.56716144e+00 3.49803478e-01 2.39042059e-01
-8.37226927e-01 -2.67067581e-01 -1.07641506e+00 -9.54000056e-02
6.46916032e-01 -6.18163466e-01 1.96921527e+00 -3.90516281e-01
-1.64330995e+00 9.79550898e-01 2.88668578e-03 -6.49341524e-01
3.43301952e-01 -4.59357649e-01 -6.12543412e-02 4.81130229e-03
1.77768588e-01 3.82402152e-01 1.82461012e-02 -1.21406400e+00
-5.98133743e-01 3.51365656e-02 8.00584972e-01 3.54657233e-01
-2.37625867e-01 -6.18245974e-02 -7.34734088e-02 -3.43656540e-01
8.07260200e-02 -8.73545110e-01 -7.73344338e-02 -4.49950248e-01
-6.43399954e-01 -3.71954083e-01 3.22252423e-01 -7.92063296e-01
1.32977879e+00 -1.81449699e+00 4.13135052e-01 -6.59629703e-02
6.27400398e-01 2.82038480e-01 -3.47475559e-02 9.02903855e-01
2.59701222e-01 3.05197984e-01 -2.47311503e-01 -5.56191206e-01
4.23105866e-01 3.81490380e-01 -4.35473502e-01 3.27864997e-02
1.98321924e-01 8.65697086e-01 -9.76537526e-01 -2.31819153e-01
-3.74847129e-02 -2.18352705e-01 -7.11662948e-01 6.73288167e-01
-7.04034150e-01 3.93335253e-01 -1.26132995e-01 1.38906822e-01
-1.27866507e-01 -4.84400690e-01 3.69239122e-01 9.76919457e-02
1.38826296e-01 1.08531630e+00 -9.64299738e-01 2.02475548e+00
-5.61192989e-01 7.39950478e-01 2.98244327e-01 -7.27226675e-01
8.32624018e-01 3.05647433e-01 8.34327787e-02 -4.53835279e-01
1.62388563e-01 7.64244944e-02 5.39517939e-01 -4.54181731e-01
9.67990339e-01 -9.76248831e-02 -2.82599896e-01 1.15903378e+00
3.42408091e-01 -6.05527103e-01 1.76129073e-01 7.71679819e-01
1.17538607e+00 -2.89070845e-01 2.16965392e-01 -2.52494127e-01
3.17961514e-01 1.19887069e-01 1.35977507e-01 1.02560842e+00
-3.97446185e-01 2.58496702e-02 8.36318195e-01 -3.83817196e-01
-6.04758978e-01 -7.10668445e-01 3.14948410e-02 1.27895200e+00
4.47141081e-02 -8.61791432e-01 -8.08091283e-01 -6.61684871e-01
-5.28234303e-01 9.08862531e-01 -4.00110036e-01 -1.82473823e-01
-5.72551906e-01 -4.59594876e-01 9.76605237e-01 1.67378768e-01
8.70870888e-01 -1.01723707e+00 -4.95933980e-01 2.65151680e-01
-7.13136494e-01 -1.30144525e+00 -1.31151006e-01 1.82054549e-01
-5.77245176e-01 -1.32743847e+00 -2.24467605e-01 -5.36472917e-01
1.77734762e-01 3.54579419e-01 1.49393463e+00 6.27748489e-01
2.18992621e-01 7.14536190e-01 -3.63289624e-01 -2.57323921e-01
-9.55381095e-01 4.79107052e-01 -1.22834355e-01 -3.43382150e-01
1.04415558e-01 -4.17287380e-01 -8.63979757e-02 3.71554554e-01
-5.17051935e-01 3.04953188e-01 1.95038587e-01 1.13704932e+00
-4.03475106e-01 -1.04857855e-01 5.43626845e-01 -1.08732772e+00
1.46590769e+00 -3.53989422e-01 -2.71107614e-01 3.07712764e-01
-3.07993114e-01 1.50847942e-01 2.20656350e-01 1.09161176e-01
-1.20500994e+00 -5.95017791e-01 -2.91865259e-01 5.91749847e-01
-8.60634744e-02 6.94168568e-01 -8.75551775e-02 1.82229549e-01
7.92931259e-01 5.62212020e-02 1.37195185e-01 -2.75858045e-01
4.87346917e-01 7.41151333e-01 6.79657459e-01 -1.22695708e+00
4.65292901e-01 -1.11181347e-03 -3.35243434e-01 -1.00370669e+00
-1.04354227e+00 -4.81387556e-01 -5.94681382e-01 -8.54013138e-04
7.26773560e-01 -7.99502194e-01 -7.89405406e-01 5.97310126e-01
-1.61804974e+00 -1.05285263e+00 -7.58853257e-02 5.83077110e-02
-6.44671917e-01 6.49101436e-01 -9.52140152e-01 -7.50662446e-01
-3.03694338e-01 -9.43502188e-01 8.46034825e-01 3.10474262e-02
-1.07079184e+00 -1.17463040e+00 2.34837443e-01 6.56059682e-01
4.44589436e-01 -7.45002329e-02 1.06999183e+00 -1.14116240e+00
-4.04260248e-01 5.80540411e-02 1.95220243e-02 3.52425128e-01
2.02095106e-01 -1.56194493e-01 -1.01429355e+00 -2.35602617e-01
1.86183438e-01 -9.05857801e-01 4.56380665e-01 -3.34902108e-01
5.21684706e-01 -5.02793133e-01 4.33696760e-03 3.36467847e-02
4.72172171e-01 1.55836672e-01 6.16335213e-01 2.20132723e-01
3.42914611e-01 1.16610706e+00 4.50355858e-01 3.68556380e-01
9.87633348e-01 7.43029654e-01 1.19391404e-01 -1.32622495e-01
1.44688025e-01 -1.98681101e-01 2.55021691e-01 1.07912016e+00
3.49345244e-02 -3.12502354e-01 -1.13224900e+00 4.94781464e-01
-1.86944354e+00 -1.02570677e+00 -6.88043982e-02 1.86276269e+00
1.39471686e+00 3.26180637e-01 1.10733092e-01 1.15924940e-01
6.42176807e-01 2.76029795e-01 -3.63514237e-02 -7.17193186e-01
-6.53962512e-03 -1.79717708e-02 -4.48321313e-01 1.14422810e+00
-8.45032990e-01 1.12736511e+00 6.89374208e+00 1.06622361e-01
-7.38793671e-01 7.11501837e-02 3.63652498e-01 4.05916363e-01
-4.40016031e-01 3.04108784e-02 -6.89364791e-01 6.99504837e-02
8.83734882e-01 -5.70765324e-02 4.00028855e-01 5.09400129e-01
-1.65330261e-01 -1.89080253e-01 -1.60771275e+00 5.44725776e-01
-5.84184602e-02 -1.16875172e+00 -2.18746558e-01 -1.25007257e-01
5.14439762e-01 -1.32851869e-01 -2.50532925e-01 5.43120623e-01
9.57426548e-01 -1.08906460e+00 4.30697858e-01 1.15422197e-01
3.54287863e-01 3.36802402e-03 7.50128865e-01 7.52615988e-01
-4.87770945e-01 2.77012289e-01 -7.48453736e-02 -6.17048979e-01
2.78650317e-04 5.98277152e-02 -1.20987856e+00 4.01039869e-01
1.41056150e-01 4.52938616e-01 -4.30479079e-01 3.74649435e-01
-5.46394646e-01 4.10529643e-01 -3.47673535e-01 -1.59355417e-01
3.28712642e-01 1.61295727e-01 4.16880727e-01 1.18398178e+00
-2.75413722e-01 3.13822240e-01 3.85806382e-01 6.76993310e-01
-1.52072996e-01 -1.19412921e-01 -7.91270792e-01 -2.75148958e-01
4.26421404e-01 1.01429439e+00 -3.62614393e-01 -4.88555193e-01
-3.37367386e-01 8.37595940e-01 5.66466391e-01 1.19606055e-01
-3.07808340e-01 -1.99125484e-01 4.89238292e-01 -2.26260126e-01
-4.11736846e-01 -6.21794164e-01 -1.10564172e-01 -1.50693691e+00
1.34520009e-01 -1.39216423e+00 4.24797267e-01 -5.58320701e-01
-1.21353436e+00 7.34983087e-01 3.53313126e-02 -4.27822053e-01
-7.74980724e-01 -5.71828783e-01 -7.93274343e-01 7.91523337e-01
-1.20988810e+00 -1.06269884e+00 -4.23047870e-01 7.04455733e-01
5.48724949e-01 -1.14283010e-01 1.27556992e+00 -1.10349149e-01
-4.80937719e-01 5.44753134e-01 -5.18660009e-01 2.29448572e-01
7.45705128e-01 -1.49192262e+00 6.09151721e-01 7.37164259e-01
3.25769365e-01 1.15574598e+00 8.88983309e-01 -5.81617177e-01
-1.27380848e+00 -4.92436171e-01 1.29352939e+00 -8.45787466e-01
1.04035175e+00 -6.09050333e-01 -1.07907259e+00 1.04732811e+00
7.05270290e-01 -7.06059992e-01 9.45295870e-01 8.20178568e-01
-4.29979771e-01 3.50912273e-01 -9.17141497e-01 7.77609348e-01
1.20605302e+00 -8.94002914e-01 -1.25297570e+00 4.97320354e-01
8.38041127e-01 -6.76871657e-01 -9.11581337e-01 1.12845629e-01
4.12632942e-01 -9.82723951e-01 4.46514934e-01 -1.04334509e+00
5.11552334e-01 1.29501551e-01 -3.14069003e-01 -1.45505679e+00
2.08723202e-01 -1.38715196e+00 4.81606508e-03 1.31230247e+00
6.92607939e-01 -6.64851367e-01 6.35743320e-01 1.20937538e+00
-3.78459334e-01 -5.36183119e-01 -7.66675353e-01 -5.05584061e-01
2.98062831e-01 -4.70802307e-01 5.28252780e-01 1.14903724e+00
8.13768685e-01 9.20723796e-01 -2.95109481e-01 -4.69161905e-02
1.79255724e-01 -2.62739439e-03 1.30838513e+00 -1.16866374e+00
-5.06734669e-01 -3.97523344e-01 -6.96701035e-02 -1.43456352e+00
4.12972957e-01 -6.64757371e-01 1.67695761e-01 -1.62086093e+00
-7.17919543e-02 -5.47379375e-01 1.85620382e-01 5.02614915e-01
-1.93003595e-01 -2.41364524e-01 3.19171309e-01 1.88570917e-01
-8.96086633e-01 5.73928595e-01 1.18010974e+00 -2.40162194e-01
-2.49214292e-01 6.14474434e-03 -1.04268014e+00 7.83311009e-01
8.62382591e-01 1.31614208e-01 -6.24419391e-01 -6.78491116e-01
4.04937923e-01 2.92205423e-01 3.23583394e-01 -6.40642941e-01
5.08325219e-01 -1.42276317e-01 -4.67451960e-01 -1.83145821e-01
5.28586745e-01 -4.73285258e-01 -4.35785055e-01 5.25290966e-01
-6.99157655e-01 1.52646050e-01 2.57700592e-01 8.04021060e-02
-2.47551456e-01 -3.96450073e-01 2.52467096e-01 -3.66179436e-01
-7.24447429e-01 -2.78625041e-01 -5.44592917e-01 5.68022847e-01
7.48665035e-01 1.41677290e-01 -8.42883825e-01 -8.59862447e-01
-8.15102637e-01 4.36817259e-01 3.46291631e-01 2.64789879e-01
3.88086587e-01 -8.80495310e-01 -7.40937889e-01 -1.49087667e-01
3.98224071e-02 3.05606127e-02 -1.27031684e-01 3.67405713e-01
-5.10520101e-01 6.54526174e-01 -1.63556933e-01 -5.05363226e-01
-1.15910745e+00 1.71477981e-02 4.72693026e-01 -6.61072493e-01
-4.32898730e-01 8.92450452e-01 2.50088811e-01 -6.69803798e-01
3.71022314e-01 -4.62061852e-01 -2.95874238e-01 3.25042605e-02
5.14732838e-01 9.10690203e-02 1.20579593e-01 -4.71856117e-01
-5.97877149e-03 -2.06857529e-02 -3.84093940e-01 -4.15846497e-01
1.04197955e+00 -3.58662248e-01 -3.05815727e-01 6.23765230e-01
7.62421250e-01 -1.30003676e-01 -8.55324447e-01 -4.27850455e-01
2.12504685e-01 -1.64133042e-01 -4.72917825e-01 -1.06642711e+00
-1.94844976e-01 1.11665606e+00 -2.84847051e-01 6.41767561e-01
5.36908448e-01 9.08335373e-02 7.33932555e-01 1.15961206e+00
6.62173390e-01 -8.56518984e-01 4.05307263e-01 1.18006814e+00
1.16349757e+00 -1.32589674e+00 -7.53130540e-02 -4.51144785e-01
-6.90893292e-01 8.37632418e-01 9.98967350e-01 4.64988351e-01
1.07038170e-01 1.07898854e-01 2.50850648e-01 -5.04411042e-01
-1.15934002e+00 -1.12086579e-01 -1.22814797e-01 5.26670992e-01
4.72173423e-01 2.38183537e-03 -1.40554532e-01 7.29921758e-01
-6.72480643e-01 -5.23155868e-01 8.16621602e-01 9.10222828e-01
-4.44459796e-01 -1.18757582e+00 1.74958333e-02 2.37817064e-01
-4.09337431e-01 -1.45967677e-01 -9.57399011e-01 8.90234828e-01
-4.37324643e-01 1.41421521e+00 -2.36255333e-01 -5.48737347e-01
3.81549358e-01 2.13963673e-01 5.55569232e-01 -1.08995926e+00
-9.55538213e-01 -6.62142038e-01 1.00551200e+00 -5.09955287e-01
-3.01025629e-01 -3.52698535e-01 -1.02361619e+00 -5.51001668e-01
-5.53191543e-01 5.96265554e-01 3.42456758e-01 1.34695137e+00
7.97629580e-02 2.28744656e-01 2.75616646e-01 -4.16014761e-01
-8.09154928e-01 -1.31137156e+00 -1.89790130e-01 5.68383873e-01
3.01114887e-01 -5.64385235e-01 -2.11522266e-01 -7.25559667e-02] | [12.640085220336914, 8.01456069946289] |
5a88431d-21c1-4059-ab7b-093da8c61ad8 | context-prediction-for-unsupervised-deep | 1901.08396 | null | https://arxiv.org/abs/1901.08396v2 | https://arxiv.org/pdf/1901.08396v2.pdf | Self-Supervised Deep Learning on Point Clouds by Reconstructing Space | Point clouds provide a flexible and natural representation usable in countless applications such as robotics or self-driving cars. Recently, deep neural networks operating on raw point cloud data have shown promising results on supervised learning tasks such as object classification and semantic segmentation. While massive point cloud datasets can be captured using modern scanning technology, manually labelling such large 3D point clouds for supervised learning tasks is a cumbersome process. This necessitates methods that can learn from unlabelled data to significantly reduce the number of annotated samples needed in supervised learning. We propose a self-supervised learning task for deep learning on raw point cloud data in which a neural network is trained to reconstruct point clouds whose parts have been randomly rearranged. While solving this task, representations that capture semantic properties of the point cloud are learned. Our method is agnostic of network architecture and outperforms current unsupervised learning approaches in downstream object classification tasks. We show experimentally, that pre-training with our method before supervised training improves the performance of state-of-the-art models and significantly improves sample efficiency. | ['Jonathan Sauder', 'Bjarne Sievers'] | 2019-01-24 | self-supervised-deep-learning-on-point-clouds | http://papers.nips.cc/paper/9455-self-supervised-deep-learning-on-point-clouds-by-reconstructing-space | http://papers.nips.cc/paper/9455-self-supervised-deep-learning-on-point-clouds-by-reconstructing-space.pdf | neurips-2019-12 | ['3d-point-cloud-linear-classification', 'point-cloud-pre-training', 'unsupervised-3d-point-cloud-linear-evaluation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.81747305e-01 4.14513618e-01 -3.08374375e-01 -8.41502130e-01
-6.04327440e-01 -6.00786448e-01 5.72487473e-01 1.76358879e-01
-5.30852199e-01 2.18742803e-01 -5.99735975e-01 -4.37073112e-01
7.81056583e-02 -8.29642355e-01 -1.32832754e+00 -3.05519998e-01
7.53846345e-03 1.34173810e+00 5.88199079e-01 8.78859237e-02
3.11145306e-01 1.09518552e+00 -1.92332017e+00 1.26196712e-01
6.85163140e-01 1.02883172e+00 4.64674532e-01 4.13674384e-01
-6.89744294e-01 2.54811913e-01 -2.54438341e-01 -1.36267632e-01
6.52345777e-01 2.17577368e-01 -1.00872421e+00 4.09948111e-01
6.06586039e-01 -3.08691591e-01 8.78227502e-03 9.59059417e-01
9.70588401e-02 7.16931149e-02 6.99518442e-01 -1.23732781e+00
-6.74006566e-02 2.45480821e-01 -4.95243728e-01 -9.92803201e-02
-3.22188944e-01 2.93559194e-01 8.18865657e-01 -9.37481046e-01
5.27608871e-01 1.05588293e+00 8.50497425e-01 5.84341347e-01
-1.15837550e+00 -6.86376035e-01 6.10836372e-02 -1.62902653e-01
-9.72302675e-01 -4.58907276e-01 9.23139155e-01 -6.45955443e-01
9.98083055e-01 7.77355507e-02 8.13620269e-01 5.33891499e-01
-3.39997530e-01 9.59210157e-01 6.31424904e-01 -2.63316363e-01
4.66995537e-01 1.07966013e-01 1.47164762e-01 6.96852505e-01
3.19319308e-01 -8.97362828e-03 -1.40093565e-01 9.53367501e-02
9.83631611e-01 5.16110301e-01 3.18743706e-01 -9.92056727e-01
-9.80543911e-01 8.54574025e-01 8.02273095e-01 7.14139268e-02
-3.39514941e-01 4.17487562e-01 2.71578014e-01 5.56844547e-02
7.05079019e-01 6.00958824e-01 -8.64451051e-01 8.74084979e-02
-1.18084300e+00 2.16268644e-01 4.82740492e-01 1.22063518e+00
1.25060344e+00 -1.32786334e-01 5.58938146e-01 8.16872954e-01
2.73809344e-01 3.79272848e-01 2.68905759e-01 -1.17554247e+00
2.05569685e-01 1.07835817e+00 2.81722676e-02 -4.91141617e-01
-4.70174968e-01 -4.26694959e-01 -6.25950098e-01 6.22913361e-01
3.41520905e-01 2.44807377e-01 -1.45827568e+00 9.05085683e-01
4.33485717e-01 3.07362258e-01 -2.44164750e-01 8.38221014e-01
7.29055822e-01 5.09106219e-01 8.83651152e-03 3.26516181e-01
9.50714707e-01 -7.30138183e-01 7.19119236e-02 -5.88281989e-01
7.45057046e-01 -3.51465732e-01 9.28672612e-01 4.36144143e-01
-9.83556211e-01 -5.54639697e-01 -9.00597692e-01 -2.02556193e-01
-4.65440392e-01 -1.35726864e-02 9.36666191e-01 3.27480286e-01
-8.45728576e-01 9.11710620e-01 -1.12570918e+00 -2.38680154e-01
1.39362061e+00 8.41673553e-01 -3.81184697e-01 -2.24446952e-01
-3.08874846e-01 6.80623293e-01 5.79004288e-01 6.31062165e-02
-9.17590380e-01 -7.85267472e-01 -7.81871915e-01 4.17851284e-02
2.73828179e-01 -7.16378093e-01 1.46330833e+00 -1.14530396e+00
-1.42625999e+00 1.29326129e+00 -9.94582847e-02 -6.77193999e-01
4.07964259e-01 -2.78703302e-01 1.96883753e-01 2.21711755e-01
2.41136864e-01 1.32256401e+00 1.15885484e+00 -1.64615810e+00
-6.98018610e-01 -5.33344150e-01 -2.03415543e-01 -9.99464095e-02
1.59690931e-01 -3.87304306e-01 -4.97778445e-01 -1.78232379e-02
5.49822152e-01 -1.01649380e+00 -5.73928952e-01 3.19946676e-01
-2.79184312e-01 -5.57678461e-01 1.17290866e+00 -7.78387338e-02
4.27126363e-02 -2.07317996e+00 -1.95348918e-01 3.15692276e-01
3.78397733e-01 4.70579892e-01 -1.52358226e-02 -1.46164015e-01
-2.80542150e-02 7.07086846e-02 -6.15499079e-01 -4.35387701e-01
-3.84069942e-02 5.65253377e-01 -1.82781667e-01 5.15744805e-01
6.18924320e-01 1.20385671e+00 -9.48400795e-01 -3.91927630e-01
5.96856594e-01 1.97284162e-01 -6.08940899e-01 1.17089488e-01
-8.62266243e-01 6.22695267e-01 -3.41073245e-01 6.71959639e-01
7.02188373e-01 -5.27488232e-01 -1.83512509e-01 2.48717628e-02
7.21889511e-02 4.45016325e-01 -7.38305032e-01 2.13948154e+00
-5.74697793e-01 7.30673552e-01 -8.42367262e-02 -1.47262216e+00
9.79549825e-01 -4.22906466e-02 7.42161930e-01 -5.10794699e-01
2.17442259e-01 2.99449682e-01 -2.17736453e-01 -4.75713938e-01
3.57387513e-01 -3.51604670e-01 6.98103085e-02 2.16656834e-01
3.58977199e-01 -7.53139257e-01 -1.78139523e-01 -6.91429675e-02
9.95298862e-01 2.79132962e-01 -3.01833719e-01 -1.38004497e-01
-1.82150751e-02 6.89374745e-01 3.56303632e-01 6.11796379e-01
5.47829233e-02 6.42220974e-01 2.21925810e-01 -7.90832400e-01
-1.35344243e+00 -9.17271912e-01 -3.04372132e-01 8.92085850e-01
2.11530209e-01 6.15007244e-02 -5.66430628e-01 -7.25471377e-01
2.18837142e-01 5.30011654e-01 -2.95178622e-01 1.88130960e-02
-5.57743788e-01 -2.81520128e-01 2.12833092e-01 6.14691973e-01
3.44157279e-01 -1.06542206e+00 -6.82791829e-01 1.30080998e-01
5.28327823e-01 -1.28525579e+00 5.50730526e-01 4.37662363e-01
-1.51142323e+00 -1.05422354e+00 -3.86866599e-01 -9.21591103e-01
1.05753219e+00 5.03347874e-01 1.39211559e+00 2.59048253e-01
-3.31999779e-01 2.77397782e-01 -1.71501160e-01 -9.13333774e-01
-1.44673035e-01 4.45746034e-01 -1.22476116e-01 -1.46972761e-01
6.91119254e-01 -7.29054511e-01 -4.86408889e-01 1.06402792e-01
-9.19574738e-01 1.35704845e-01 8.24706554e-01 6.11021280e-01
9.23605561e-01 1.85806125e-01 4.09376591e-01 -1.27756882e+00
-2.81075798e-02 -4.63433325e-01 -8.90673280e-01 -4.29370940e-01
-2.56357193e-01 1.72643021e-01 3.78190905e-01 -1.35071382e-01
-6.65143549e-01 8.03791285e-01 -4.48804379e-01 -9.73965883e-01
-7.01300859e-01 1.43738627e-01 -9.55686066e-03 -1.65459916e-01
6.16580963e-01 -1.19625293e-01 2.01746374e-01 -5.77966332e-01
4.79416132e-01 5.96576095e-01 5.59009910e-01 -4.75991189e-01
9.85582530e-01 8.88584852e-01 1.58907965e-01 -9.43393528e-01
-9.04453397e-01 -8.29948962e-01 -1.20842636e+00 -1.48494378e-01
9.38197136e-01 -8.67935240e-01 -5.90189874e-01 2.37269953e-01
-1.28947675e+00 -5.85207701e-01 -5.85220575e-01 2.55010188e-01
-5.94197690e-01 -4.72037792e-02 -1.57548204e-01 -6.27272606e-01
-1.05760418e-01 -1.04222727e+00 1.54753935e+00 3.63519713e-02
-6.40321821e-02 -8.64137053e-01 -1.46389842e-01 4.37556684e-01
-2.29655895e-02 1.57505736e-01 7.64150441e-01 -7.77892351e-01
-1.08767283e+00 -4.56920981e-01 -2.74513900e-01 3.67860079e-01
-2.11859178e-02 -2.02498183e-01 -1.10277295e+00 -2.21817512e-02
-2.20407784e-01 -4.15587425e-01 9.35567975e-01 2.77799755e-01
1.76772213e+00 -9.03232917e-02 -7.03773975e-01 7.26542711e-01
1.30044079e+00 -1.57296866e-01 5.17609954e-01 2.43158460e-01
9.51273382e-01 7.21204281e-01 3.58325243e-01 -3.50716673e-02
6.74358234e-02 2.46907994e-01 8.01255763e-01 -2.15849832e-01
2.82640517e-01 -3.11193615e-01 -2.99171805e-01 3.62098664e-01
4.36249971e-02 2.38038480e-01 -1.28100383e+00 7.65371442e-01
-1.77410388e+00 -7.38220334e-01 -2.58807361e-01 1.95987093e+00
4.93488342e-01 5.49919426e-01 7.45972991e-02 3.18849027e-01
3.84457409e-01 -1.82567716e-01 -9.04893100e-01 -1.22335084e-01
4.08661097e-01 6.66085839e-01 7.44096220e-01 2.71281034e-01
-1.21081138e+00 1.27409756e+00 5.86273384e+00 4.91709203e-01
-1.30960882e+00 9.12622213e-02 4.86387193e-01 -8.83333534e-02
-2.42257029e-01 -8.03622380e-02 -6.70840442e-01 3.19455326e-01
8.64014924e-01 5.86196780e-01 1.41669035e-01 1.18597317e+00
9.43832546e-02 -4.65672128e-02 -1.36068308e+00 1.16787469e+00
-3.32737416e-01 -1.76496053e+00 4.29811329e-02 1.88965276e-01
6.69489086e-01 6.93841934e-01 -9.27350074e-02 2.95879304e-01
5.77868581e-01 -1.17436981e+00 5.01822650e-01 3.61012250e-01
8.11656058e-01 -7.06381559e-01 5.35710573e-01 8.61982167e-01
-6.58153594e-01 1.28135070e-01 -7.05499530e-01 -4.26931605e-02
2.78755720e-03 9.88260984e-01 -1.23443139e+00 2.21857518e-01
8.00146818e-01 8.78447711e-01 -4.91190076e-01 1.12555659e+00
-5.46847098e-02 6.56448722e-01 -6.95511818e-01 1.40956908e-01
3.86437088e-01 -2.65375018e-01 3.65138471e-01 9.21868622e-01
4.09977660e-02 8.63846317e-02 1.25397518e-01 1.04928303e+00
-3.01365286e-01 -4.03991550e-01 -9.45757151e-01 -7.68739209e-02
4.03832614e-01 1.17980897e+00 -1.14025450e+00 -3.51097584e-01
-1.81262046e-01 7.26094186e-01 5.46133220e-01 8.38711672e-03
-2.43108675e-01 -1.85686260e-01 6.10411704e-01 3.95023793e-01
5.44133186e-01 -6.15425169e-01 -7.13559926e-01 -8.78203094e-01
-2.02544201e-02 -2.26773158e-01 -1.89398438e-01 -8.50065231e-01
-1.24135804e+00 3.07378769e-01 -2.84876977e-03 -1.36593354e+00
-3.04228365e-01 -8.72866452e-01 -6.17797971e-01 5.66378355e-01
-1.63169074e+00 -1.11531842e+00 -2.92408526e-01 3.38814855e-01
7.42295682e-01 -1.50437474e-01 4.39586043e-01 1.30716503e-01
-2.33657941e-01 -9.14104935e-03 -5.64833498e-03 3.51094335e-01
2.37501375e-02 -1.15678978e+00 7.59279668e-01 5.92104077e-01
4.06443745e-01 4.79981214e-01 3.55069071e-01 -5.38496614e-01
-1.50285769e+00 -1.48018193e+00 3.79791260e-01 -6.07348621e-01
1.90716788e-01 -6.49846673e-01 -1.02739584e+00 8.09129775e-01
-4.63626444e-01 5.96372545e-01 4.29637611e-01 1.08586334e-01
-2.22092196e-01 -1.06711000e-01 -1.18090224e+00 1.33351207e-01
1.20357311e+00 -4.43875581e-01 -6.05303824e-01 6.45202696e-01
7.39203274e-01 -4.42184925e-01 -5.39080918e-01 4.77742046e-01
1.41406760e-01 -8.28085244e-01 1.10057068e+00 -8.08924675e-01
4.52438474e-01 -3.04351449e-01 1.42910078e-01 -1.18487060e+00
-7.98250139e-02 -3.10262173e-01 -6.15274645e-02 7.49394059e-01
3.52520347e-01 -5.06983876e-01 1.54824531e+00 6.50963902e-01
-5.88825524e-01 -6.99941456e-01 -8.57096136e-01 -6.46920204e-01
1.76846802e-01 -6.98144853e-01 6.83066428e-01 8.82402837e-01
-6.03909373e-01 4.48101103e-01 4.23622817e-01 2.69006461e-01
7.88188756e-01 2.41183653e-01 1.05005348e+00 -2.04935384e+00
8.99701416e-02 -3.56307417e-01 -7.27541566e-01 -1.27039123e+00
5.66708922e-01 -1.10235870e+00 1.88869461e-01 -1.78574407e+00
-3.40183377e-01 -1.14330578e+00 3.25434178e-01 6.33958697e-01
2.49770030e-01 3.44872385e-01 -3.80965285e-02 4.93622839e-01
-6.14518285e-01 3.46142918e-01 9.81168211e-01 -3.70857716e-01
-2.88829535e-01 2.84411132e-01 -3.31474543e-01 9.17249143e-01
8.72019529e-01 -6.90900028e-01 -3.79215509e-01 -8.58978689e-01
2.91429758e-01 -2.90382683e-01 6.58443093e-01 -1.13681316e+00
2.19401240e-01 1.67802442e-02 5.85197210e-01 -9.73682523e-01
4.66727227e-01 -1.26326919e+00 -4.10203524e-02 2.26029560e-01
-1.15057424e-01 -4.71946239e-01 2.71386415e-01 5.46130955e-01
-5.09470999e-02 -4.63053048e-01 7.46489823e-01 -5.57802916e-01
-8.75152171e-01 6.02322936e-01 -1.24119774e-01 -2.48667970e-01
9.68207359e-01 -5.54668725e-01 2.26883441e-01 1.57202661e-01
-6.76243424e-01 1.62417203e-01 6.76079392e-01 3.22790086e-01
6.83780670e-01 -1.01950431e+00 -3.27233255e-01 4.10950512e-01
1.10664636e-01 1.40150774e+00 -2.50157386e-01 2.69059449e-01
-8.68036628e-01 4.02882516e-01 -1.72510639e-01 -1.35480165e+00
-7.35763967e-01 5.42724252e-01 3.42718989e-01 1.64897576e-01
-8.96776974e-01 9.00824249e-01 2.08984176e-03 -1.02053618e+00
5.86041398e-02 -4.74701792e-01 8.57752338e-02 -2.80355394e-01
9.48342681e-03 7.17347786e-02 5.35722911e-01 -3.88196826e-01
-1.61820069e-01 5.69434881e-01 -1.15909971e-01 1.13495715e-01
1.74165833e+00 2.99197078e-01 -1.14293829e-01 4.68245119e-01
1.43786705e+00 -6.83584332e-01 -1.56906116e+00 -3.06325108e-01
2.88942128e-01 -4.50972795e-01 2.41617575e-01 -3.10004056e-01
-1.10699689e+00 1.26532006e+00 4.47440058e-01 1.43442526e-01
7.07240880e-01 3.81662011e-01 8.72073829e-01 8.40458274e-01
7.06676185e-01 -8.98600340e-01 -5.57104684e-02 6.19276702e-01
4.57624286e-01 -1.62737215e+00 2.36723982e-02 -5.13066888e-01
-1.64893016e-01 1.05541301e+00 4.56073165e-01 -5.80163360e-01
7.22270489e-01 1.41665012e-01 -1.30607069e-01 -5.99412799e-01
-5.06670058e-01 -3.66167396e-01 2.51161724e-01 7.81507909e-01
-2.07840353e-02 -9.07440186e-02 6.31042182e-01 1.31349877e-01
-3.43062013e-01 6.15548678e-02 1.25846133e-01 1.15147507e+00
-7.78255522e-01 -9.19569969e-01 -2.44262293e-01 9.31898415e-01
-9.97080654e-02 2.85743743e-01 -4.83764291e-01 6.42422736e-01
2.58665442e-01 5.33098340e-01 7.85651982e-01 -1.81984127e-01
2.20549718e-01 1.54689983e-01 4.78710145e-01 -1.06050014e+00
-3.26389909e-01 -2.78328776e-01 -1.51181117e-01 -4.97085780e-01
-5.83101571e-01 -5.78057289e-01 -1.54353023e+00 -9.09704622e-03
-2.94516206e-01 1.03774443e-02 1.12644458e+00 1.18447888e+00
5.21604717e-01 3.68263423e-01 5.17429173e-01 -1.76108038e+00
-1.53763905e-01 -7.01685309e-01 -2.39626661e-01 4.84685898e-01
4.36527044e-01 -7.26032257e-01 -7.45519921e-02 2.71756321e-01] | [8.05838394165039, -3.1912994384765625] |
a33e7066-7b23-41ca-aa85-77faf019e690 | extending-layered-models-to-3d-motion | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Dong_Lao_Extending_Layered_Models_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Dong_Lao_Extending_Layered_Models_ECCV_2018_paper.pdf | Extending Layered Models to 3D Motion | We consider the problem of inferring a layered representa-tion, its depth ordering and motion segmentation from a video in whichobjects may undergo 3D non-planar motion relative to the camera. Wegeneralize layered inference to the aforementioned case and correspond-ing self-occlusion phenomena. We accomplish this by introducing a flat-tened 3D object representation, which is a compact representation of anobject that contains all visible portions of the object seen in the video,including parts of an object that are self-occluded (as well as occluded)in one frame but seen in another. We formulate the inference of such flat-tened representations and motion segmentation, and derive an optimiza-tion scheme. We also introduce a new depth ordering scheme, which isindependent of layered inference and addresses the case of self-occlusion.It requires almost no computation given the flattened representations.Experiments on benchmark datasets show the advantage of our methodcompared to existing layered methods, which do not model 3D motionand self-occlusion. | ['Dong Lao', 'Ganesh Sundaramoorthi'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 9.50861424e-02 5.44623911e-01 -2.78682888e-01 -2.12350085e-01
-5.83017111e-01 -5.19855618e-01 6.48291528e-01 -9.42371264e-02
-1.14848077e-01 5.32558918e-01 1.07766911e-01 -2.12372467e-01
-3.78159503e-03 -6.32472992e-01 -1.10115826e+00 -7.47091711e-01
-1.78043857e-01 7.16003656e-01 7.18517125e-01 2.59878546e-01
7.94713199e-02 7.57444978e-01 -1.67014778e+00 5.04592896e-01
6.50123060e-01 7.10695326e-01 1.88519761e-01 9.79564190e-01
-2.59367377e-01 9.32146132e-01 -5.18460095e-01 -1.59110844e-01
5.17776981e-02 -1.01615056e-01 -1.07315147e+00 1.15228415e+00
8.73641312e-01 -4.89154100e-01 -4.25540179e-01 8.08791995e-01
-2.34658554e-01 3.77960913e-02 7.41902590e-01 -1.21027231e+00
-2.75165215e-03 1.32155353e-02 -1.00269771e+00 2.27955431e-01
5.67210019e-01 -2.92648196e-01 6.34231806e-01 -9.15338993e-01
9.39920902e-01 1.56445360e+00 3.81419569e-01 4.20873404e-01
-1.34893072e+00 1.07694767e-01 7.39370763e-01 5.09759299e-02
-1.04016542e+00 -5.05341947e-01 7.14255810e-01 -7.37284660e-01
7.06962109e-01 5.09925425e-01 4.46090221e-01 5.66075623e-01
2.25932360e-01 1.16495049e+00 1.04814172e+00 -3.69768381e-01
2.46426672e-01 -1.45400986e-01 4.11718458e-01 8.39507341e-01
5.80039263e-01 -2.07062095e-01 -3.00805271e-01 -1.96061030e-01
9.74108338e-01 4.21908870e-03 -2.91927487e-01 -8.24560285e-01
-1.34589064e+00 3.06210697e-01 1.28218696e-01 2.42959913e-02
-4.01307285e-01 1.14661470e-01 9.77622867e-02 -3.82910892e-02
6.72540784e-01 -1.93780541e-01 -3.32365662e-01 1.13531157e-01
-1.13168752e+00 4.30821329e-01 9.66461480e-01 8.88370812e-01
1.10208106e+00 1.45718500e-01 1.90618917e-01 2.83214509e-01
3.14550102e-01 3.29170436e-01 1.47516936e-01 -1.24139786e+00
6.35383785e-01 6.76283240e-01 3.39328051e-01 -8.26508760e-01
-2.95719147e-01 4.20713015e-02 -6.12171412e-01 6.41782880e-01
4.45113152e-01 1.88932456e-02 -1.27550411e+00 1.33959901e+00
8.13533664e-01 4.57671285e-01 1.36674255e-01 8.58888507e-01
9.33870018e-01 7.61493385e-01 -2.86601275e-01 -3.02964479e-01
1.41283906e+00 -8.93710196e-01 -6.56611025e-01 -1.59619257e-01
5.04700899e-01 -3.64187419e-01 5.17176151e-01 4.15170580e-01
-1.52926409e+00 -4.95163292e-01 -1.11771238e+00 -2.08407104e-01
5.97737767e-02 -2.56395280e-01 5.06201029e-01 2.75881797e-01
-1.10947049e+00 5.08715928e-01 -1.20316970e+00 -2.36930400e-01
2.73900121e-01 4.62046206e-01 -7.98757970e-01 -1.63492840e-02
-4.89097178e-01 6.18279696e-01 4.78327036e-01 2.57364154e-01
-9.72815335e-01 -4.55065608e-01 -1.12789619e+00 -9.73928720e-02
5.92259109e-01 -8.82250011e-01 1.23309791e+00 -1.08236253e+00
-9.95051563e-01 1.07193196e+00 -5.82325459e-01 -1.74602196e-01
5.34957230e-01 -1.22500002e-01 8.31997544e-02 4.56468552e-01
3.12357321e-02 5.13657451e-01 8.76889944e-01 -1.52174115e+00
-4.66641635e-01 -7.91984975e-01 7.09402263e-01 5.96095145e-01
2.38493472e-01 -1.48628950e-01 -8.48270416e-01 -5.82202911e-01
9.11988497e-01 -9.71470833e-01 -3.20439875e-01 2.93757558e-01
-7.21996903e-01 -8.65030736e-02 1.35441077e+00 -5.85184097e-01
9.10260618e-01 -2.08621764e+00 5.71737945e-01 -1.45753637e-01
3.21326375e-01 3.90553772e-02 1.24992296e-01 1.44549757e-01
-8.74730051e-02 8.49817991e-02 -4.36372370e-01 -7.14175761e-01
-2.20930442e-01 7.04778016e-01 -1.68558568e-01 7.82014132e-01
1.75345749e-01 5.39137542e-01 -7.83766806e-01 -8.24653208e-01
3.22747350e-01 3.62151444e-01 -6.59304142e-01 5.17920971e-01
-4.83960003e-01 3.97832245e-01 -6.05090141e-01 7.68777430e-01
8.61474395e-01 -1.53937116e-01 1.99676659e-02 -1.81454364e-02
-2.86299050e-01 1.37476534e-01 -1.57801116e+00 1.75131977e+00
-8.95267278e-02 4.37343061e-01 4.39637423e-01 -9.21554506e-01
4.83877122e-01 4.70908463e-01 6.36645973e-01 5.96139301e-03
-4.75988612e-02 -1.68147951e-01 -4.87661451e-01 -8.06151807e-01
4.35730934e-01 -9.62808803e-02 1.03798121e-01 2.30583951e-01
-1.46430448e-01 -2.02599362e-01 1.79643825e-01 3.51824194e-01
8.72254550e-01 5.41301489e-01 1.63133442e-01 -3.49533260e-01
5.37139297e-01 7.55930394e-02 6.76535785e-01 6.23609900e-01
6.77849650e-02 8.13076496e-01 7.49780059e-01 -4.78080094e-01
-8.37574184e-01 -1.01865911e+00 -1.22578867e-01 5.95421135e-01
4.96979266e-01 -3.79711866e-01 -8.88298154e-01 -5.66164613e-01
-1.01683050e-01 2.50434041e-01 -7.32935011e-01 4.06795174e-01
-7.11139619e-01 -5.16919911e-01 -4.34925035e-03 5.01832187e-01
4.86346811e-01 -5.81872225e-01 -1.00383019e+00 1.41694754e-01
-4.59745944e-01 -1.26338649e+00 -2.06750736e-01 -6.58163801e-03
-1.30245161e+00 -1.13952720e+00 -7.04322755e-01 -7.89235115e-01
9.83678341e-01 5.26702702e-01 1.09184980e+00 1.02556296e-01
-3.56295377e-01 5.67758143e-01 5.85879833e-02 -7.86762908e-02
-2.41680786e-01 -3.44782650e-01 -1.95324153e-01 1.37270778e-01
-1.15493856e-01 -3.79335642e-01 -4.71732318e-01 2.00668395e-01
-9.08382177e-01 4.38032031e-01 1.85381457e-01 4.53156739e-01
9.17082608e-01 1.78201735e-01 -2.69474775e-01 -1.10430706e+00
-3.44716728e-01 -4.54440564e-01 -7.02017844e-01 1.92016229e-01
2.18838528e-01 9.38337818e-02 8.40403587e-02 -3.50021482e-01
-1.34652078e+00 3.20566118e-01 8.57382491e-02 -6.87636495e-01
-3.49464148e-01 1.71718106e-01 -5.30985177e-01 1.40363321e-01
1.49660707e-01 -8.79524350e-02 -2.51010329e-01 -5.96360087e-01
3.98347557e-01 3.30797881e-01 5.43912828e-01 -7.03444183e-01
6.99211836e-01 1.16852427e+00 3.19453388e-01 -1.18256235e+00
-7.30252683e-01 -5.29159963e-01 -1.16743493e+00 -1.83831215e-01
1.04944587e+00 -8.86135101e-01 -5.15888989e-01 3.88738573e-01
-1.54187632e+00 -1.36626780e-01 -2.98028618e-01 4.34886277e-01
-8.12253356e-01 6.89775825e-01 -5.64677060e-01 -1.01090062e+00
3.32842559e-01 -1.16173339e+00 1.34450746e+00 -4.21963669e-02
-1.79276839e-01 -1.04371381e+00 8.41826871e-02 3.58402580e-01
-4.51580435e-01 7.85476744e-01 1.05419290e+00 -1.75984845e-01
-1.13889027e+00 -7.40625113e-02 -1.10708689e-02 3.95436771e-02
-7.48455226e-02 9.59798619e-02 -9.49120700e-01 -2.26772368e-01
2.00168937e-01 3.67866494e-02 8.27632904e-01 4.67249244e-01
1.01058996e+00 -4.12820131e-01 -4.60693479e-01 7.05368996e-01
1.44406831e+00 1.97676182e-01 5.96267998e-01 2.84611940e-01
8.93241465e-01 8.29895437e-01 5.24477482e-01 2.76774287e-01
4.63515788e-01 7.96102643e-01 9.09131944e-01 -3.69617850e-01
-1.88689843e-01 1.69583820e-02 1.78536639e-01 4.47434664e-01
-3.47139746e-01 -3.16572994e-01 -5.54625928e-01 5.37590265e-01
-1.91812360e+00 -9.43527460e-01 -6.44526660e-01 2.27189779e+00
3.43471199e-01 1.24663718e-01 1.88882694e-01 2.70296097e-01
6.45637572e-01 4.27821547e-01 -6.41969979e-01 -4.27763492e-01
-9.94310081e-02 -3.42192650e-01 2.33546406e-01 7.75925934e-01
-1.26355565e+00 5.91188312e-01 6.57386732e+00 8.83457959e-02
-7.23476410e-01 -4.26244959e-02 6.72876894e-01 1.81363955e-01
-4.76804793e-01 1.91340595e-01 -8.50356698e-01 8.33651945e-02
5.33856511e-01 -5.62548488e-02 -2.02895924e-02 7.48054028e-01
1.78930491e-01 -3.48673135e-01 -1.30881310e+00 5.75757325e-01
3.53100479e-01 -1.14296389e+00 1.59787104e-01 2.31039718e-01
8.34581554e-01 -4.22230512e-01 -1.34714961e-01 -3.39326784e-02
2.62036100e-02 -5.24134755e-01 8.73214185e-01 4.43386048e-01
5.51095724e-01 -4.49254930e-01 2.61719257e-01 6.25154793e-01
-1.31759191e+00 2.00301349e-01 -2.75408417e-01 -3.63037959e-02
3.51290315e-01 4.79381770e-01 -4.99301642e-01 6.73463821e-01
5.29286146e-01 6.65933788e-01 -9.95351970e-02 1.05918157e+00
-4.05723490e-02 2.20167190e-01 -2.41921544e-01 6.25329494e-01
2.92110354e-01 -3.36277366e-01 9.22608674e-01 1.17171776e+00
1.51382044e-01 4.72012401e-01 4.67557520e-01 5.06502330e-01
1.69544309e-01 -3.64233196e-01 -6.97200596e-01 3.21589261e-01
-3.56500074e-02 8.32475901e-01 -9.47473526e-01 -6.60103858e-01
-7.52734721e-01 1.18769181e+00 2.17755482e-01 7.11734474e-01
-7.34568298e-01 1.84128266e-02 6.40021861e-01 6.30316019e-01
5.58754444e-01 -4.17139590e-01 -4.02394198e-02 -1.41802740e+00
1.70781255e-01 -4.23388571e-01 4.86932129e-01 -9.67750728e-01
-6.07870698e-01 6.33167505e-01 5.95981777e-01 -1.16779912e+00
-3.83354962e-01 -5.83096743e-01 -3.67323726e-01 5.24975598e-01
-1.49696243e+00 -1.02519941e+00 -2.70813286e-01 5.85391223e-01
9.68407333e-01 5.45470297e-01 5.38673639e-01 -5.60957864e-02
-4.34101850e-01 -2.81916231e-01 -1.97294369e-01 -1.56758547e-01
-8.11819211e-02 -1.34724844e+00 1.86416954e-01 1.02805650e+00
1.49742095e-02 3.64035100e-01 7.01170981e-01 -7.40203440e-01
-1.52169752e+00 -1.16386437e+00 8.94426286e-01 -4.74822253e-01
4.14742410e-01 -4.73221958e-01 -1.16602409e+00 1.08474219e+00
-8.61281678e-02 1.64123088e-01 1.86589837e-01 -1.59635589e-01
-2.41489470e-01 1.69834688e-01 -9.43948030e-01 4.85301733e-01
1.05489540e+00 -3.86578798e-01 -8.78454089e-01 3.76873076e-01
6.42443955e-01 -8.75147402e-01 -5.31264246e-01 5.33558071e-01
4.80200440e-01 -1.20510447e+00 1.29043067e+00 -6.64928317e-01
2.05982596e-01 -3.99618238e-01 -9.81543735e-02 -6.91631198e-01
-1.07458390e-01 -5.21982431e-01 -6.24692023e-01 7.67854929e-01
-1.04822040e-01 -1.86064750e-01 1.27287710e+00 5.89717925e-01
-3.39053541e-01 -8.56535494e-01 -9.90886152e-01 -8.21223438e-01
-6.16735406e-02 -2.01292172e-01 2.80909777e-01 8.54690909e-01
-2.59373218e-01 -1.63093172e-02 -2.36854956e-01 6.02730334e-01
9.37847137e-01 2.66879171e-01 8.85777414e-01 -1.39534569e+00
-2.50762582e-01 1.84193924e-01 -7.59951770e-01 -1.47074568e+00
3.86313468e-01 -4.97353822e-01 1.39036000e-01 -1.90859437e+00
2.10150048e-01 -2.09280234e-02 8.96015912e-02 2.41352767e-01
-8.63239989e-02 1.00545324e-01 1.27284512e-01 3.94226402e-01
-6.87936246e-01 2.72135526e-01 1.40270305e+00 -1.17193066e-01
-1.35825723e-01 7.57609531e-02 -2.25905702e-01 1.30522025e+00
3.57048631e-01 -4.63881642e-01 -6.94083095e-01 -7.22909391e-01
-2.67419636e-01 7.12893307e-01 5.82169712e-01 -8.04517388e-01
9.22857672e-02 -2.13820264e-01 4.14855659e-01 -1.25954401e+00
9.71641302e-01 -9.51699317e-01 4.06215131e-01 4.76647973e-01
-6.00182153e-02 2.36366093e-01 1.51970148e-01 7.13944018e-01
-1.38316423e-01 -3.38657320e-01 5.94918132e-01 -5.26615262e-01
-8.13322306e-01 4.34581071e-01 -6.22747958e-01 -7.43393451e-02
1.32199728e+00 -8.16336393e-01 4.75718454e-02 -1.70824170e-01
-1.20867670e+00 4.43454087e-02 6.58740938e-01 1.18888669e-01
6.67530894e-01 -1.03710473e+00 -6.93871677e-01 2.96189934e-01
-1.60948291e-01 4.26355183e-01 4.04314816e-01 6.15088642e-01
-8.18106413e-01 5.02631485e-01 2.20111907e-01 -1.00362742e+00
-1.54334915e+00 5.76784372e-01 2.05677167e-01 -1.16582792e-02
-9.41079021e-01 6.80811048e-01 9.81903493e-01 -8.98082033e-02
4.08554286e-01 -6.16675258e-01 -2.19709039e-01 -1.59579203e-01
5.63721836e-01 5.39338291e-01 -2.27764603e-02 -1.02928376e+00
-2.95806259e-01 8.70487809e-01 -3.91910188e-02 -1.23250388e-01
9.87469792e-01 -2.82040030e-01 -2.13812381e-01 8.25658381e-01
1.11086047e+00 -2.30437115e-01 -1.66220880e+00 -1.51205629e-01
-1.35714158e-01 -8.28239679e-01 -9.92502794e-02 -3.96155454e-02
-9.73219872e-01 9.07315195e-01 1.12650603e-01 1.81476876e-01
8.11379611e-01 2.33607128e-01 2.69346893e-01 2.45921642e-01
4.96068627e-01 -4.16440129e-01 1.60701692e-01 4.10277426e-01
7.67336249e-01 -9.32659328e-01 2.47582778e-01 -8.44604909e-01
-4.07159984e-01 1.15341914e+00 6.30758584e-01 -2.22762331e-01
6.00918412e-01 2.52952576e-01 -2.21048549e-01 -2.76222676e-01
-7.31296241e-01 -1.17953241e-01 2.77727365e-01 5.87736189e-01
3.05222541e-01 -1.46578401e-01 -8.76680642e-05 -1.87150836e-01
2.49839246e-01 -1.39354929e-01 6.70944273e-01 1.19083989e+00
-4.69763368e-01 -6.19485259e-01 -6.80724740e-01 6.08076416e-02
-3.78367186e-01 3.58103096e-01 -1.75595626e-01 1.06384778e+00
1.96114361e-01 8.15848410e-01 4.74531472e-01 2.05476359e-01
3.31688434e-01 -2.50712067e-01 6.89759612e-01 -9.03815091e-01
9.02353004e-02 4.26761121e-01 1.78289309e-01 -6.50959253e-01
-7.69275784e-01 -8.53203475e-01 -1.34533072e+00 -1.17377765e-01
-2.95088708e-01 1.56617910e-01 7.29412675e-01 9.69084978e-01
-2.56636348e-02 4.86044854e-01 4.75558698e-01 -1.33145368e+00
-5.95202670e-02 -3.11279774e-01 -6.73989832e-01 3.26457769e-01
8.15998435e-01 -6.49939358e-01 -3.83923441e-01 4.33153421e-01] | [8.85463809967041, -2.020390510559082] |
73471be7-7e14-48e2-83f0-548cb5bd529b | thermal-spread-functions-tsf-physics-guided | 2304.00696 | null | https://arxiv.org/abs/2304.00696v1 | https://arxiv.org/pdf/2304.00696v1.pdf | Thermal Spread Functions (TSF): Physics-guided Material Classification | Robust and non-destructive material classification is a challenging but crucial first-step in numerous vision applications. We propose a physics-guided material classification framework that relies on thermal properties of the object. Our key observation is that the rate of heating and cooling of an object depends on the unique intrinsic properties of the material, namely the emissivity and diffusivity. We leverage this observation by gently heating the objects in the scene with a low-power laser for a fixed duration and then turning it off, while a thermal camera captures measurements during the heating and cooling process. We then take this spatial and temporal "thermal spread function" (TSF) to solve an inverse heat equation using the finite-differences approach, resulting in a spatially varying estimate of diffusivity and emissivity. These tuples are then used to train a classifier that produces a fine-grained material label at each spatial pixel. Our approach is extremely simple requiring only a small light source (low power laser) and a thermal camera, and produces robust classification results with 86% accuracy over 16 classes. | ['Oliver Cossairt', 'Ashok Veeraraghavan', 'Aggelos Katsaggelos', 'Florian Willomitzer', 'Emma Alexander', 'Vishwanath Saragadam', 'Aniket Dashpute'] | 2023-04-03 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Dashpute_Thermal_Spread_Functions_TSF_Physics-Guided_Material_Classification_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Dashpute_Thermal_Spread_Functions_TSF_Physics-Guided_Material_Classification_CVPR_2023_paper.pdf | cvpr-2023-1 | ['material-classification'] | ['computer-vision'] | [ 5.70531309e-01 -5.86972713e-01 5.08503653e-02 -2.30885133e-01
-4.22400057e-01 -6.96468234e-01 7.56027877e-01 -4.50456627e-02
-4.64364588e-01 6.19775474e-01 -5.38548648e-01 -7.92265013e-02
7.68505186e-02 -8.19959104e-01 -4.81672466e-01 -1.59506428e+00
5.50543427e-01 7.28907704e-01 4.91503775e-01 4.26387429e-01
5.77909052e-01 7.98326313e-01 -1.61486387e+00 -2.85861194e-02
6.04734898e-01 1.27675354e+00 3.28693748e-01 7.90233254e-01
7.63881207e-03 6.26879334e-01 -1.08813189e-01 2.60411531e-01
2.33596489e-01 -4.43492323e-01 -1.03269303e+00 1.79159343e-01
3.62483770e-01 -3.06462705e-01 -4.53697853e-02 7.21575260e-01
1.01119794e-01 5.46372950e-01 1.36491764e+00 -7.67836094e-01
-3.83334130e-01 4.53954116e-02 -8.89173985e-01 2.01095730e-01
5.42488173e-02 2.50266463e-01 6.63420498e-01 -4.41935152e-01
4.67887074e-01 7.66751409e-01 4.55452859e-01 7.72111416e-01
-1.34449792e+00 -2.58175969e-01 -2.46257007e-01 1.30856913e-02
-1.10512531e+00 -4.26080644e-01 9.92932141e-01 -8.47437024e-01
8.04539025e-01 3.51126581e-01 7.32357681e-01 6.46450818e-01
5.54917812e-01 1.35221228e-01 1.58453143e+00 -4.99156892e-01
7.45812237e-01 2.39626795e-01 9.95674655e-02 7.60304451e-01
5.44253830e-03 2.36406088e-01 -3.40984017e-01 -1.62093684e-01
6.58204854e-01 -2.92600840e-02 -7.97864050e-02 -1.90412194e-01
-1.05153537e+00 3.99545461e-01 3.97158414e-01 1.04026161e-01
-3.46202642e-01 3.12525779e-01 -5.81245637e-03 -6.50336891e-02
7.51753151e-01 4.92102236e-01 -1.41799837e-01 -1.06654398e-01
-1.00726950e+00 8.39468613e-02 5.51684678e-01 1.51079610e-01
1.02027035e+00 -3.16964179e-01 1.26114339e-01 8.51238668e-01
2.34819680e-01 1.06342852e+00 5.72458990e-02 -1.12424445e+00
-2.53930956e-01 4.23943996e-01 9.38357934e-02 -5.53257048e-01
-1.76885933e-01 1.53015796e-02 -5.67772567e-01 6.14102006e-01
6.14142537e-01 8.40777084e-02 -1.16409576e+00 1.33624172e+00
7.56474197e-01 2.08533630e-01 -2.18422562e-01 1.03286421e+00
3.89328361e-01 8.14022839e-01 9.23215970e-02 -4.13940191e-01
1.19118285e+00 -6.93790376e-01 -1.44399583e-01 -1.69684663e-01
3.61603320e-01 -7.00693369e-01 8.94003510e-01 2.53519088e-01
-1.01258528e+00 -1.92807913e-01 -9.04147387e-01 -1.81550886e-02
-2.62669623e-01 -1.26027256e-01 4.99023229e-01 5.05780876e-01
-5.42212605e-01 7.12713718e-01 -1.06235075e+00 -3.85532409e-01
4.12124872e-01 1.53537661e-01 -2.09701825e-02 -1.34203508e-01
-3.49612027e-01 8.62905264e-01 4.21698764e-02 5.61066903e-02
-8.27283621e-01 -1.04909515e+00 -2.80836344e-01 -6.18609726e-01
1.58583030e-01 -8.46374273e-01 1.00057590e+00 -6.15918577e-01
-1.86473715e+00 1.07425630e+00 -3.97456825e-01 8.03838372e-02
6.25799716e-01 -1.39532313e-02 1.66416705e-01 4.86947894e-01
2.19653174e-02 2.31390104e-01 1.18921459e+00 -1.54173338e+00
-2.93139875e-01 -4.32912797e-01 -2.84062833e-01 9.21203941e-03
5.60687818e-02 1.95397884e-02 -1.12828776e-01 -5.38665839e-02
3.47218841e-01 -1.17392874e+00 6.15821965e-02 9.20509472e-02
-3.72669786e-01 6.98608309e-02 1.11152661e+00 -2.96646148e-01
4.41014856e-01 -2.10490847e+00 1.73533320e-01 1.38635665e-01
1.30198598e-01 -1.69201165e-01 1.39740944e-01 8.89559388e-02
2.49607295e-01 -1.37995481e-01 -7.76407897e-01 -4.91526127e-02
-5.46289206e-01 -2.01812029e-01 -2.74474829e-01 8.09951901e-01
1.52355269e-01 8.79278958e-01 -9.54543650e-01 -5.20825028e-01
5.98979890e-01 6.05449378e-01 -1.99536353e-01 1.41094372e-01
-5.29194593e-01 7.40770221e-01 -5.72139144e-01 5.67325115e-01
9.34718370e-01 -3.74108739e-02 -1.92556575e-01 -4.07902062e-01
-4.58242387e-01 2.88982186e-02 -7.54710138e-01 1.32103133e+00
-5.72785497e-01 7.00306296e-01 2.43839726e-01 -8.92294705e-01
8.46371651e-01 -4.72842492e-02 9.29379761e-01 -6.70351446e-01
5.35981774e-01 3.61132145e-01 -3.95148754e-01 -5.10809243e-01
3.39737982e-01 -7.15743899e-01 2.29293644e-01 9.87351060e-01
-4.34437454e-01 -7.92329311e-01 -1.83025256e-01 -9.57478676e-03
1.17911899e+00 2.34462306e-01 -2.65019506e-01 -4.61686373e-01
3.44326168e-01 1.62798941e-01 5.12625538e-02 5.79449713e-01
-6.79691136e-02 6.07504725e-01 -4.38541472e-02 -4.92236078e-01
-1.24471092e+00 -1.23245871e+00 -3.68531615e-01 8.08763683e-01
3.48062277e-01 2.50092983e-01 -9.52940047e-01 -2.22179413e-01
1.81783527e-01 3.18341643e-01 -7.20906317e-01 -4.40813005e-01
-5.91736019e-01 -8.35498333e-01 -1.53927296e-01 4.06770676e-01
6.28119648e-01 -1.08726227e+00 -9.86925781e-01 2.29404122e-02
-1.27119213e-01 -1.15862405e+00 -1.75247863e-02 2.17332602e-01
-9.26369727e-01 -1.02726316e+00 -4.59494114e-01 -2.56212473e-01
8.51572990e-01 3.47330421e-01 1.01014864e+00 6.35932907e-02
-8.37954760e-01 5.23916125e-01 -1.74158081e-01 -2.13790104e-01
-3.32239956e-01 -3.63700241e-02 9.71007496e-02 2.43152544e-01
1.39059424e-01 -3.82583857e-01 -7.30377913e-01 2.35484824e-01
-6.37252033e-01 6.96290731e-02 3.63043517e-01 3.39138001e-01
6.23423576e-01 4.60214287e-01 -1.49561763e-01 -7.70492971e-01
9.33413804e-02 -2.24672154e-01 -7.41623521e-01 7.83777386e-02
-3.95671844e-01 4.44919541e-02 4.51442033e-01 -5.67690194e-01
-1.34161842e+00 1.02484629e-01 2.46614829e-01 -2.70929217e-01
-2.71649301e-01 -3.11092213e-02 1.28470749e-01 -3.61524045e-01
6.68836892e-01 2.49188572e-01 -9.21159238e-02 -2.65925139e-01
2.38218203e-01 5.31287730e-01 6.81854963e-01 -9.03458238e-01
1.06531990e+00 1.03154433e+00 5.18178880e-01 -1.43747807e+00
-6.81157529e-01 -4.63813841e-01 -8.59714627e-01 -6.99376583e-01
9.20548916e-01 -3.45222831e-01 -1.12770820e+00 8.95773232e-01
-7.57360935e-01 -7.48386502e-01 -4.46584940e-01 4.68395084e-01
-5.66467047e-01 1.35947600e-01 -5.17064750e-01 -1.19712698e+00
-2.97486275e-01 -8.48989785e-01 1.12625754e+00 3.05557251e-01
-7.67386928e-02 -1.18532646e+00 5.37939891e-02 6.92438245e-01
4.48781282e-01 2.51010686e-01 1.12538445e+00 5.94248414e-01
-8.76862884e-01 -1.13235511e-01 -3.62298816e-01 2.72832245e-01
3.42726439e-01 4.37273353e-01 -1.34060109e+00 -1.14614308e-01
3.09650004e-01 -3.87036711e-01 1.16203713e+00 8.95754337e-01
1.18778265e+00 3.37962240e-01 -5.25427520e-01 6.91638827e-01
1.79702401e+00 9.48181301e-02 2.10411251e-01 1.41446099e-01
7.89191186e-01 6.72407925e-01 5.33270121e-01 2.60044307e-01
-8.59341919e-02 5.05353630e-01 2.64698714e-01 8.98892246e-03
-1.57803595e-01 3.47869933e-01 2.16010615e-01 4.06831771e-01
-5.77590525e-01 -3.43764834e-02 -9.00777698e-01 3.01216692e-01
-1.26708782e+00 -9.76882935e-01 -4.14380729e-01 2.42042351e+00
8.75743508e-01 -1.14715822e-01 6.53114766e-02 9.73580927e-02
4.29942876e-01 1.05394222e-01 -7.89804637e-01 -2.52115279e-01
2.47500420e-01 3.45274776e-01 7.90043652e-01 7.18745947e-01
-8.52695227e-01 7.98606157e-01 7.15973234e+00 4.22923386e-01
-1.65705359e+00 -1.29166573e-01 5.22910058e-01 -3.20696414e-01
-2.92487442e-01 1.43775605e-02 -4.42041576e-01 6.01975918e-01
9.34943676e-01 1.97585911e-01 6.79955959e-01 3.94658744e-01
2.35860720e-01 -9.19127703e-01 -1.10384452e+00 7.90504456e-01
-2.50111639e-01 -1.07929289e+00 -3.04422498e-01 1.79661036e-01
5.69154799e-01 2.11508930e-01 1.22923329e-01 -6.00957036e-01
1.04911804e-01 -5.56482077e-01 5.88623047e-01 7.12279141e-01
9.78355527e-01 -2.22444102e-01 1.26048505e-01 2.25115463e-01
-1.14201272e+00 3.09667904e-02 -3.67170453e-01 5.48849106e-02
9.15411487e-02 1.23991656e+00 -4.88270372e-01 1.49151385e-01
7.19890296e-01 4.50041741e-01 -2.45732293e-01 5.69037735e-01
1.81408767e-02 6.57716811e-01 -7.18610823e-01 -3.79945077e-02
-6.45761192e-02 -6.98549151e-01 3.52242321e-01 9.78266895e-01
1.25020027e-01 3.99347365e-01 -1.23270825e-01 1.15361011e+00
8.86735395e-02 -3.06956172e-01 -3.78552914e-01 8.11231732e-02
2.79922485e-01 1.51918173e+00 -1.22911012e+00 -2.26293325e-01
-7.34994980e-03 1.00013161e+00 1.27774760e-01 3.90047878e-01
-5.45084953e-01 5.94881624e-02 4.75463569e-01 3.24211091e-01
-7.28623569e-02 -3.20742518e-01 -4.42622989e-01 -8.92902374e-01
2.86679238e-01 1.00946695e-01 -2.27309495e-01 -9.59092617e-01
-1.21337759e+00 1.28807396e-01 7.97163099e-02 -8.12214971e-01
1.72507569e-01 -7.70475328e-01 -7.30318367e-01 9.86097217e-01
-1.59086323e+00 -1.11450982e+00 -6.49243474e-01 3.90797228e-01
2.32396811e-01 3.68486226e-01 6.12940431e-01 -1.97120756e-01
-5.35477877e-01 -1.66658401e-01 2.93659449e-01 -2.39841640e-01
6.04065537e-01 -1.24397087e+00 -7.25975186e-02 4.77956742e-01
-4.62964833e-01 3.23582411e-01 8.76385093e-01 -5.85408866e-01
-1.73575985e+00 -1.01296484e+00 3.54368091e-01 -6.85781538e-01
5.60395777e-01 -4.91751939e-01 -8.90518188e-01 1.73255235e-01
-6.10408634e-02 3.59833330e-01 5.45393229e-01 -2.28930354e-01
-3.68328393e-01 -2.36361593e-01 -1.42166924e+00 1.54642642e-01
7.83638597e-01 -8.60660613e-01 -2.87947834e-01 4.75930870e-01
1.22429989e-01 -9.46255326e-02 -5.92889547e-01 1.30975306e-01
8.51527691e-01 -7.62658298e-01 7.56299436e-01 -2.06013694e-01
6.09617889e-01 -1.51913553e-01 -4.11127917e-02 -1.10499597e+00
-3.93638968e-01 -2.19221741e-01 1.55589610e-01 1.04149926e+00
3.99978936e-01 -6.59794509e-01 1.10756350e+00 9.92277741e-01
7.93294013e-02 -5.75642347e-01 -6.49468780e-01 -7.67023385e-01
1.42704025e-01 -4.31162268e-01 8.94649103e-02 8.11546803e-01
-3.19812834e-01 1.03139291e-02 5.44938147e-02 1.62488446e-01
1.05493772e+00 3.66859227e-01 3.54830325e-01 -1.47449207e+00
3.94708663e-02 -1.18701950e-01 -3.85245346e-02 -7.04409957e-01
1.94641441e-01 -7.57191956e-01 3.90358031e-01 -1.33213949e+00
6.42737210e-01 -7.67483830e-01 -7.80027285e-02 8.68467465e-02
9.91123542e-02 4.89919275e-01 -3.29665810e-01 5.13415873e-01
-4.41415049e-02 2.61693567e-01 1.04005301e+00 -2.77194291e-01
-2.76570678e-01 -1.81766048e-01 -5.23127466e-02 5.27769685e-01
6.58261478e-01 -4.05444473e-01 -2.54969507e-01 -3.53499085e-01
1.34078383e-01 -3.22400063e-01 6.59800649e-01 -1.16453934e+00
4.11720686e-02 -6.94234669e-01 3.26969147e-01 -4.22834516e-01
5.88574767e-01 -9.38663542e-01 6.12605736e-02 4.17933136e-01
-1.09103128e-01 -6.86761618e-01 1.58460755e-02 2.79463857e-01
4.42338616e-01 -3.98491144e-01 1.33315086e+00 -2.41016284e-01
-2.76868165e-01 3.30294698e-01 -4.87903804e-01 -2.79191881e-01
1.29804778e+00 -2.73955077e-01 -4.62316006e-01 2.45785594e-01
-3.24217469e-01 -3.86090428e-01 9.85770226e-01 -9.18316245e-02
3.87120724e-01 -1.15599871e+00 -4.41724032e-01 2.77336538e-01
-1.40771016e-01 1.48281991e-01 3.89599711e-01 8.00892115e-01
-7.45573580e-01 9.36839804e-02 -6.85145557e-02 -1.02746367e+00
-1.12315929e+00 3.69833171e-01 6.56762600e-01 3.22386533e-01
-6.33147359e-01 1.05307198e+00 -5.57676330e-03 -4.25537415e-02
-4.78714973e-01 -3.13981593e-01 3.38270366e-01 -6.32068515e-02
3.34729582e-01 5.74589372e-01 1.82502031e-01 -5.52978516e-01
-4.85756844e-01 1.38253641e+00 -2.54342053e-02 -2.61535883e-01
1.37542450e+00 -1.00749455e-01 -5.40853798e-01 7.96798229e-01
1.29934347e+00 -2.23866314e-01 -1.48229957e+00 -2.44119883e-01
-4.76258129e-01 -5.60027659e-01 5.42357802e-01 -5.99349856e-01
-1.12669206e+00 9.22109783e-01 4.86289173e-01 3.92781496e-01
1.15033495e+00 4.57881004e-01 6.80862129e-01 1.29616976e-01
2.41278976e-01 -1.43098652e+00 2.78772097e-02 1.75687432e-01
3.74588072e-01 -1.20277953e+00 4.29981768e-01 -6.46191418e-01
-9.46576446e-02 1.00239086e+00 3.67825627e-01 -8.54931325e-02
7.15771198e-01 5.21941066e-01 1.88211724e-01 -3.93305987e-01
-6.85724378e-01 1.87619582e-01 1.51110753e-01 5.70183277e-01
3.29944938e-01 1.69838846e-01 2.14778647e-01 -6.72106445e-01
-6.17358610e-02 -9.79408249e-02 3.05982530e-01 1.04417038e+00
-6.52351797e-01 -6.92232370e-01 -3.82186741e-01 6.31750703e-01
-4.65602502e-02 2.88259894e-01 -4.96204793e-01 9.88037065e-02
1.18169434e-01 8.60564530e-01 3.90407920e-01 -2.45971531e-01
-8.20069239e-02 1.62311018e-01 9.12647963e-01 -5.16301692e-01
2.43750010e-02 -1.13206180e-02 -4.21065718e-01 -5.27208507e-01
-8.75579000e-01 -9.53906775e-01 -1.33215308e+00 -4.80428398e-01
-3.80931020e-01 -1.25311792e-01 1.08479357e+00 1.18848205e+00
-1.77102000e-01 3.06825221e-01 1.00972712e+00 -1.15370488e+00
-9.79294330e-02 -7.90829122e-01 -8.32202315e-01 4.48410362e-01
6.32639885e-01 -8.90687823e-01 -7.76529729e-01 3.31380993e-01] | [9.81561279296875, -2.870887279510498] |
411be1d2-3d8c-41ef-bce0-9b90e5cbcc2f | semantic-relatedness-based-re-ranker-for-text | 1909.0795 | null | https://arxiv.org/abs/1909.07950v2 | https://arxiv.org/pdf/1909.07950v2.pdf | Semantic Relatedness Based Re-ranker for Text Spotting | Applications such as textual entailment, plagiarism detection or document clustering rely on the notion of semantic similarity, and are usually approached with dimension reduction techniques like LDA or with embedding-based neural approaches. We present a scenario where semantic similarity is not enough, and we devise a neural approach to learn semantic relatedness. The scenario is text spotting in the wild, where a text in an image (e.g. street sign, advertisement or bus destination) must be identified and recognized. Our goal is to improve the performance of vision systems by leveraging semantic information. Our rationale is that the text to be spotted is often related to the image context in which it appears (word pairs such as Delta-airplane, or quarters-parking are not similar, but are clearly related). We show how learning a word-to-word or word-to-sentence relatedness score can improve the performance of text spotting systems up to 2.9 points, outperforming other measures in a benchmark dataset. | ['Lluís Padró', 'Francesc Moreno-Noguer', 'Ahmed Sabir'] | 2019-09-17 | semantic-relatedness-based-re-ranker-for-text-1 | https://aclanthology.org/D19-1346 | https://aclanthology.org/D19-1346.pdf | ijcnlp-2019-11 | ['text-spotting'] | ['computer-vision'] | [ 5.76168716e-01 -1.92130953e-01 1.78269729e-01 -5.75353742e-01
-6.92430973e-01 -4.71036077e-01 1.15321648e+00 7.82269835e-01
-4.77871418e-01 1.36308879e-01 5.03900468e-01 -1.74382731e-01
-5.65966070e-02 -5.00719547e-01 -4.86763597e-01 -5.38795769e-01
2.83200502e-01 4.23029780e-01 3.75975743e-02 -3.85552496e-01
9.17239428e-01 2.41992325e-01 -1.58351517e+00 4.88822728e-01
2.83139974e-01 9.03175473e-01 3.51560652e-01 6.44115508e-01
-5.63612461e-01 6.41521513e-01 -8.40875566e-01 -5.35459220e-01
-7.39576817e-02 -5.47978878e-01 -7.77728200e-01 5.43595515e-02
8.92435312e-01 3.04790493e-03 -3.28696728e-01 1.22154486e+00
3.15470219e-01 1.76997766e-01 8.98349285e-01 -1.08862054e+00
-7.01215863e-01 2.89319754e-01 -6.68328881e-01 2.71304339e-01
3.96306336e-01 -3.75045627e-01 1.34382498e+00 -1.14928579e+00
5.65760434e-01 1.41581130e+00 6.22634530e-01 2.86675900e-01
-1.00196004e+00 -2.78481245e-01 -6.03631958e-02 5.52274406e-01
-1.33738458e+00 -1.88037366e-01 7.47136116e-01 -4.52673018e-01
8.39294314e-01 3.42898846e-01 1.05576523e-01 1.17429090e+00
1.41093567e-01 7.94113696e-01 7.32573271e-01 -6.97133541e-01
2.38930985e-01 6.60026148e-02 3.61449301e-01 5.76394320e-01
3.16370994e-01 -5.14002264e-01 -6.63733184e-01 -7.57525712e-02
4.13691774e-02 2.28001013e-01 5.98201901e-03 -2.95337707e-01
-1.13574755e+00 1.22169113e+00 5.07305324e-01 6.25201523e-01
-1.66109189e-01 1.69810474e-01 7.45654225e-01 1.97080150e-01
4.22686666e-01 8.08419406e-01 -3.03689484e-02 4.66916151e-02
-1.18139637e+00 1.98798761e-01 8.60999465e-01 6.39576197e-01
8.03936958e-01 -5.80988824e-01 -8.54233652e-02 8.91417444e-01
1.18555106e-01 3.41788024e-01 8.20183158e-01 -6.78734183e-01
6.51707470e-01 4.60989892e-01 -1.17394954e-01 -1.58750713e+00
-3.11837643e-01 4.47602011e-02 -4.50409293e-01 6.17383160e-02
2.62129366e-01 8.70195683e-03 -7.68303275e-01 1.25476623e+00
3.04497276e-02 -6.70485478e-03 1.59794644e-01 9.11321223e-01
7.24703610e-01 5.76871395e-01 -6.25443012e-02 1.00625470e-01
1.74184442e+00 -9.60717797e-01 -6.18429422e-01 -7.83249140e-01
9.41388488e-01 -1.29829597e+00 1.10864878e+00 1.27217993e-02
-5.07057428e-01 -2.70894855e-01 -1.12218344e+00 -2.14074552e-01
-9.04571116e-01 -2.73561068e-02 -7.87468478e-02 5.01757205e-01
-6.85742438e-01 8.43977451e-01 -1.98574245e-01 -1.09735680e+00
2.17003256e-01 -1.23506859e-01 -5.26416123e-01 -2.53123254e-01
-1.13989365e+00 1.08436120e+00 6.11656725e-01 -4.11514550e-01
-3.30057114e-01 -1.94084793e-01 -9.75536585e-01 1.14073187e-01
2.27887318e-01 -4.89825875e-01 9.31503892e-01 -1.23600197e+00
-6.85048938e-01 1.39006031e+00 -2.47932285e-01 -6.68967962e-01
1.45870924e-01 -3.15189064e-01 -6.47850990e-01 4.58824635e-01
5.08500159e-01 6.52019024e-01 1.47956800e+00 -1.04840815e+00
-6.75256073e-01 -5.56416690e-01 -5.74728027e-02 4.05616313e-01
-6.20930135e-01 4.55515653e-01 -1.15594059e-01 -7.90899754e-01
2.46660322e-01 -9.13256645e-01 4.35516872e-02 9.21974238e-03
-4.53827769e-01 -4.35447633e-01 1.42825341e+00 -7.57590413e-01
9.96976376e-01 -2.25997090e+00 -2.80048281e-01 1.15386494e-01
2.82216519e-01 3.36891919e-01 -2.03398392e-01 8.70952725e-01
-2.85292000e-01 9.76579934e-02 -2.98841238e-01 -3.67352456e-01
1.26571923e-01 1.68202087e-01 -3.26155663e-01 5.21267951e-01
2.91979015e-01 7.38974273e-01 -8.87972414e-01 -6.15616977e-01
1.92125380e-01 2.90592819e-01 -4.48768176e-02 -1.57165453e-01
3.38337086e-02 -1.47026807e-01 -3.96661192e-01 2.71290869e-01
3.35743666e-01 -4.27993029e-01 1.28293037e-02 -4.41520691e-01
1.82936326e-01 3.83697510e-01 -8.37252557e-01 1.59136224e+00
-4.85756367e-01 1.40712607e+00 -2.73478508e-01 -1.29767442e+00
9.64387715e-01 -1.89508572e-02 2.06165418e-01 -7.93864727e-01
3.07700932e-02 2.18012836e-03 -2.70995617e-01 -7.60037184e-01
8.11869919e-01 -1.29913902e-02 -8.00291747e-02 6.63921118e-01
-1.83598042e-01 -4.08648811e-02 1.55178517e-01 3.82259876e-01
1.27301466e+00 -5.00081956e-01 3.10610205e-01 -2.23265976e-01
5.47570288e-01 1.79681927e-01 -2.95201451e-01 6.69673860e-01
-2.80292183e-01 7.76288331e-01 6.28877819e-01 -4.45668936e-01
-1.26557124e+00 -6.99326456e-01 -2.23464649e-02 1.19296062e+00
3.38890851e-01 -8.09984684e-01 -7.17184424e-01 -8.73713315e-01
3.86710577e-02 1.08350670e+00 -6.98721826e-01 -4.36236888e-01
-5.63028932e-01 -3.31967056e-01 5.25142431e-01 3.03557545e-01
1.38111606e-01 -8.92460585e-01 -6.75738513e-01 1.31141907e-02
-2.92132765e-01 -1.12493026e+00 -5.78379273e-01 2.45997608e-01
-6.87251329e-01 -9.19394195e-01 -8.22272420e-01 -9.37428355e-01
6.72637999e-01 6.10409796e-01 1.18068671e+00 -4.74466123e-02
-5.01155972e-01 5.15966237e-01 -4.42128092e-01 -1.26869574e-01
-4.35152948e-01 -2.50282794e-01 -1.91292483e-02 1.86839953e-01
8.38580370e-01 -2.89453328e-01 -7.22049236e-01 1.96535021e-01
-1.04394603e+00 -4.28059906e-01 4.12658691e-01 9.21130478e-01
5.52198626e-02 1.51727498e-01 1.92829132e-01 -5.43456078e-01
9.19924259e-01 -4.55932051e-01 3.77299711e-02 2.42051914e-01
-6.81535006e-01 1.99339703e-01 3.59148115e-01 -3.29262227e-01
-5.07090271e-01 9.80629623e-02 7.14443401e-02 -5.32805026e-01
-5.23993015e-01 5.29028416e-01 2.24101886e-01 -5.94803803e-02
9.69765425e-01 2.87335545e-01 5.29772695e-03 -3.41799170e-01
5.39875388e-01 9.13461506e-01 4.06981409e-01 -9.63913798e-02
6.63353801e-01 6.50114179e-01 1.37930876e-02 -1.10804498e+00
-8.59461963e-01 -1.04183483e+00 -4.15917188e-01 -8.21940973e-02
9.33525324e-01 -6.30292594e-01 -4.20020789e-01 -9.52522904e-02
-1.39664054e+00 5.06706655e-01 -5.95197715e-02 4.27714676e-01
-4.87571865e-01 8.28330934e-01 -2.96347320e-01 -4.63503510e-01
-2.22147703e-01 -7.66205370e-01 1.31438053e+00 -4.03474681e-02
-6.36846244e-01 -9.53735709e-01 2.37292368e-02 4.89980131e-01
1.66404396e-01 -1.67410955e-01 1.14188528e+00 -1.36883473e+00
1.59069911e-01 -5.14013946e-01 -5.10594070e-01 2.74613678e-01
3.16500545e-01 -2.97826350e-01 -8.55779827e-01 -1.32401377e-01
2.21372664e-01 -1.79693982e-01 9.52795684e-01 1.67163670e-01
9.90769565e-01 -3.02288383e-01 -3.57595354e-01 -3.38870361e-02
1.15014780e+00 -1.36221290e-01 6.40730500e-01 6.02650702e-01
8.24181259e-01 1.26593137e+00 6.21116281e-01 3.84034604e-01
1.82695612e-02 8.31623793e-01 5.22903562e-01 1.94498211e-01
-1.61486208e-01 -8.30150917e-02 3.17973942e-01 3.13909203e-01
8.23574722e-01 -3.10325414e-01 -1.10117769e+00 6.69732273e-01
-1.94675457e+00 -9.52551067e-01 -4.00536656e-01 2.16917801e+00
5.46234965e-01 3.58591862e-02 -9.80573520e-02 1.33701921e-01
1.18927872e+00 2.76577532e-01 -2.66815513e-01 -5.05628467e-01
-7.66896382e-02 -2.07378834e-01 4.98951554e-01 2.86552876e-01
-1.21532166e+00 1.02350807e+00 5.90951967e+00 1.10868549e+00
-1.08328700e+00 -5.27827777e-02 5.08935630e-01 -4.54150811e-02
1.89483851e-01 -6.16420619e-02 -7.80241311e-01 5.72834551e-01
9.46932435e-01 -6.51535466e-02 3.48543525e-02 8.09674740e-01
1.38413031e-02 -1.18427366e-01 -1.32130623e+00 1.41777396e+00
9.26827371e-01 -1.26815057e+00 1.35367587e-01 -8.74659494e-02
4.06772792e-01 -1.26009688e-01 1.11634284e-01 -6.28847182e-02
-1.18606910e-01 -9.63653445e-01 4.87271130e-01 2.10796416e-01
3.08792859e-01 -7.83194005e-01 6.09166682e-01 6.17856532e-02
-8.19963157e-01 2.06123024e-01 -4.30793434e-01 2.44861275e-01
8.33300427e-02 6.65344238e-01 -1.09567165e+00 1.37358427e-01
7.33682275e-01 9.24923480e-01 -9.14337397e-01 9.25654292e-01
-1.46798626e-01 2.86172241e-01 -2.02402249e-01 -3.58093202e-01
6.42642617e-01 -5.01564741e-02 7.55310774e-01 1.38414848e+00
4.40367311e-01 -4.19090539e-01 -1.02091245e-01 7.42824197e-01
-3.30249928e-02 3.75031710e-01 -1.03129816e+00 -4.09884214e-01
3.65940034e-02 1.36118019e+00 -9.99995351e-01 -3.76107812e-01
-4.36726898e-01 1.52508473e+00 2.02885643e-02 1.46084681e-01
-5.79502225e-01 -9.35639441e-01 7.02786505e-01 -9.34584439e-02
6.79349601e-01 -1.82313807e-02 -1.78193063e-01 -7.54942715e-01
2.62653619e-01 -6.75031662e-01 2.84860790e-01 -1.10136116e+00
-1.59989393e+00 4.13442373e-01 -3.48902315e-01 -1.37921381e+00
-4.40666795e-01 -7.25957215e-01 -6.26673937e-01 6.87313914e-01
-1.16080737e+00 -9.75476384e-01 -2.24464893e-01 6.85957551e-01
7.72090971e-01 -3.44686925e-01 6.81575894e-01 1.72666475e-01
-1.23536959e-01 3.72380078e-01 3.59052569e-01 2.93559968e-01
1.00048125e+00 -1.10075927e+00 7.40419686e-01 7.98914850e-01
7.75295198e-01 4.80171710e-01 1.10032248e+00 -6.58902705e-01
-9.80501652e-01 -9.51615572e-01 1.43026757e+00 -6.34371102e-01
1.20908666e+00 -2.25307137e-01 -9.43452239e-01 1.94114551e-01
3.40425879e-01 -2.73284942e-01 6.33286238e-01 -2.81941630e-02
-8.08271527e-01 2.01003730e-01 -9.25432682e-01 9.42791045e-01
6.62273943e-01 -1.04418862e+00 -1.06313968e+00 8.94975781e-01
7.03519523e-01 8.63315538e-02 -2.55776346e-01 -2.61568278e-01
1.72983944e-01 -7.90063560e-01 1.11524391e+00 -9.82169569e-01
8.03348899e-01 -1.64072096e-01 -4.16671783e-01 -1.14099658e+00
-1.17216920e-02 -3.15072507e-01 5.03183976e-02 1.11518562e+00
9.54006091e-02 -3.33468825e-01 7.72466183e-01 3.56460869e-01
3.14586237e-02 2.46675964e-02 -1.01627493e+00 -9.82383370e-01
-7.40881488e-02 -3.36378336e-01 2.00471863e-01 1.19175839e+00
1.03080064e-01 8.39556515e-01 -2.27476656e-01 -5.81276342e-02
5.13028085e-01 -2.06290036e-02 3.49108696e-01 -1.23226607e+00
6.44200668e-02 -7.46605456e-01 -9.40916479e-01 -9.09042478e-01
3.51776987e-01 -9.93071496e-01 2.00100526e-01 -1.35273683e+00
1.82794780e-02 1.05015911e-01 -1.37516350e-01 1.34301767e-01
-1.10061906e-01 3.21776241e-01 3.12628180e-01 2.61984497e-01
-6.42090559e-01 3.73163313e-01 6.47081971e-01 -5.34607172e-01
2.63632804e-01 -1.06057063e-01 -5.58194578e-01 8.62565100e-01
7.33682871e-01 -7.77052104e-01 -5.09289205e-02 -2.80293792e-01
4.50047106e-01 -2.52063274e-01 6.40497327e-01 -9.26626921e-01
4.34173942e-01 2.75200516e-01 1.36448696e-01 -6.04093254e-01
4.64637101e-01 -1.17503333e+00 -5.59988976e-01 3.59358996e-01
-7.45183408e-01 2.72792995e-01 -8.05012137e-02 9.61309612e-01
-3.79661351e-01 -1.03200233e+00 4.48705584e-01 -5.83465211e-02
-9.36129689e-01 -3.11795324e-01 -4.94181424e-01 7.47705922e-02
9.25083339e-01 -3.67625713e-01 -4.20917392e-01 -5.91087878e-01
-4.43326622e-01 -1.01412319e-01 3.30209047e-01 7.53903568e-01
7.67042279e-01 -1.25809646e+00 -7.13124096e-01 -1.40538156e-01
7.86122441e-01 -7.59559274e-01 -1.14585139e-01 6.86906099e-01
-4.00875270e-01 4.75569516e-01 2.00854465e-02 -6.19581640e-01
-1.65144515e+00 7.01158822e-01 1.05618581e-01 5.42636625e-02
-8.02357495e-01 6.70214534e-01 2.20476091e-01 -2.32779030e-02
2.90387332e-01 1.10119641e-01 -3.40808839e-01 4.71372545e-01
8.22055995e-01 2.55573303e-01 2.25212574e-01 -9.14904237e-01
-4.97784078e-01 7.96416044e-01 -3.49360257e-01 -7.53526986e-02
9.27090108e-01 -3.55254382e-01 -2.54795849e-01 4.94563729e-01
1.75901628e+00 -3.78151059e-01 -6.37842834e-01 -4.90022153e-01
5.23663282e-01 -6.05938494e-01 2.61920005e-01 -5.40913641e-01
-7.14188814e-01 1.14734030e+00 7.93056309e-01 5.47672331e-01
7.55959511e-01 3.74983490e-01 9.88971889e-01 1.00709999e+00
-2.07364574e-01 -1.35935247e+00 4.66985226e-01 7.23736584e-01
7.87393510e-01 -1.60464120e+00 -6.38742745e-02 1.02912430e-02
-7.93928325e-01 1.37313282e+00 9.45255607e-02 -9.48927328e-02
3.85803074e-01 -1.19596832e-01 -4.53718603e-02 -5.34038484e-01
-3.40110123e-01 -4.23656613e-01 6.30503237e-01 5.70948601e-01
2.84409970e-01 -1.94360048e-01 -2.86030769e-01 -5.29625267e-02
-1.27068803e-01 -7.52405167e-01 3.41517389e-01 9.58382607e-01
-7.29187548e-01 -9.26144779e-01 -5.42339027e-01 6.08257473e-01
-5.22872388e-01 -4.79415387e-01 -8.51249456e-01 4.05542046e-01
-3.37466478e-01 9.43192422e-01 3.68928701e-01 -2.34897435e-01
-1.39619512e-02 2.27268606e-01 7.56856427e-02 -6.52083158e-01
-5.16149104e-01 -1.70276508e-01 9.16094109e-02 -3.95798683e-01
-3.98060501e-01 -6.39836788e-01 -1.15486264e+00 -2.34954044e-01
-2.06897348e-01 -1.56706646e-02 1.03386664e+00 1.20012271e+00
2.47475758e-01 7.41662756e-02 6.18730068e-01 -6.66420937e-01
-2.29140759e-01 -7.25488007e-01 -6.04733527e-01 7.95015216e-01
4.01192755e-01 -3.69532794e-01 -5.85170984e-01 2.74510175e-01] | [11.66940689086914, 2.2171733379364014] |
ac0d8b5a-02d7-47b2-97e8-5d335c70c10b | curved-text-detection-in-natural-scene-images | 1908.0999 | null | https://arxiv.org/abs/1908.09990v1 | https://arxiv.org/pdf/1908.09990v1.pdf | Curved Text Detection in Natural Scene Images with Semi- and Weakly-Supervised Learning | Detecting curved text in the wild is very challenging. Recently, most state-of-the-art methods are segmentation based and require pixel-level annotations. We propose a novel scheme to train an accurate text detector using only a small amount of pixel-level annotated data and a large amount of data annotated with rectangles or even unlabeled data. A baseline model is first obtained by training with the pixel-level annotated data and then used to annotate unlabeled or weakly labeled data. A novel strategy which utilizes ground-truth bounding boxes to generate pseudo mask annotations is proposed in weakly-supervised learning. Experimental results on CTW1500 and Total-Text demonstrate that our method can substantially reduce the requirement of pixel-level annotated data. Our method can also generalize well across two datasets. The performance of the proposed method is comparable with the state-of-the-art methods with only 10% pixel-level annotated data and 90% rectangle-level weakly annotated data. | ['Weiping Wang', 'Xugong Qin', 'Yu Zhou', 'Dongbao Yang'] | 2019-08-27 | null | null | null | null | ['curved-text-detection'] | ['computer-vision'] | [ 2.48632088e-01 2.84718961e-01 -2.57624835e-01 -4.39711213e-01
-1.33316958e+00 -6.26787543e-01 3.68611693e-01 1.13734506e-01
-6.16511524e-01 4.61748719e-01 -1.98955774e-01 -1.98110685e-01
8.73394370e-01 -5.63812494e-01 -6.61089718e-01 -5.37562788e-01
5.24037480e-01 6.23955309e-01 1.26119924e+00 7.65071809e-02
1.86662167e-01 2.08758846e-01 -1.26706755e+00 3.86218369e-01
9.42195415e-01 8.30288351e-01 2.84122705e-01 6.97517216e-01
-5.24889827e-01 5.40765584e-01 -3.75061274e-01 -3.52540880e-01
4.42220747e-01 -2.41288558e-01 -8.04749787e-01 6.21472836e-01
7.02711046e-01 -3.99135381e-01 -1.21115193e-01 1.00524712e+00
3.77468765e-01 -8.67921934e-02 6.28395498e-01 -7.95308650e-01
-2.72285163e-01 5.64269483e-01 -1.22463179e+00 -8.32872316e-02
8.59152377e-02 8.47068951e-02 9.44129169e-01 -1.31765497e+00
5.67285657e-01 9.29385185e-01 7.13876426e-01 5.71323454e-01
-1.10339344e+00 -5.32692611e-01 5.22850573e-01 -4.42394078e-01
-1.49593019e+00 -8.33166484e-03 7.96840072e-01 -5.14694989e-01
3.53831232e-01 8.71201232e-03 3.72937143e-01 7.99553216e-01
-2.90162206e-01 1.24923241e+00 1.12507808e+00 -7.19543636e-01
1.89387128e-01 2.55626976e-01 2.89420694e-01 1.09028876e+00
3.21678966e-01 -3.91006947e-01 -4.42623466e-01 -3.37215513e-02
6.76280737e-01 -1.83472186e-01 -1.94686398e-01 -4.39488769e-01
-1.34785831e+00 6.04394615e-01 2.13256717e-01 3.80166024e-01
9.78593677e-02 -8.34286958e-02 3.72809768e-01 -4.83141720e-01
8.00275862e-01 -4.87151816e-02 -4.78014618e-01 1.02979489e-01
-1.53823757e+00 -1.01596519e-01 6.51398718e-01 1.34244108e+00
7.31614888e-01 3.74398418e-02 -2.43841156e-01 9.19099808e-01
2.49565125e-01 4.94832933e-01 4.79050308e-01 -5.20692706e-01
9.17361796e-01 9.04256642e-01 2.27563053e-01 -4.33871180e-01
-4.50820684e-01 -9.56295952e-02 -5.46298087e-01 1.79055169e-01
8.96286905e-01 -3.86069268e-01 -1.41625214e+00 9.57109690e-01
4.61931288e-01 -1.48768961e-01 -3.66561204e-01 8.05005968e-01
7.44638681e-01 5.48105419e-01 -2.17273161e-01 4.21099849e-02
1.25016415e+00 -1.38702965e+00 -6.88762128e-01 -4.83878702e-01
8.55125964e-01 -1.01594758e+00 1.43388522e+00 3.69320869e-01
-1.09761536e+00 -6.28453314e-01 -1.07649720e+00 -3.19239646e-01
-4.65427548e-01 7.46374965e-01 2.49228239e-01 9.20366228e-01
-9.07190681e-01 1.85155854e-01 -9.06929433e-01 -5.08405983e-01
6.91032290e-01 3.28613967e-01 -1.05896629e-01 7.55520761e-02
-5.14855325e-01 4.58920568e-01 6.84219301e-01 1.54890055e-02
-6.64233923e-01 -3.96288425e-01 -7.20407605e-01 -1.49082452e-01
6.94876611e-01 -9.86276343e-02 1.33631790e+00 -8.37892890e-01
-1.46564019e+00 1.19576073e+00 -1.08785689e-01 -2.03041911e-01
1.00248635e+00 -3.23934793e-01 -7.67281651e-02 3.46303940e-01
1.63532749e-01 9.43022490e-01 8.88809443e-01 -1.50985765e+00
-1.03314352e+00 -3.18575531e-01 -4.16004509e-01 1.26635849e-01
-4.03126299e-01 -4.28950116e-02 -1.16579270e+00 -8.45140517e-01
3.20740581e-01 -9.60484028e-01 -3.04569542e-01 4.45680171e-01
-8.25698853e-01 -2.52805233e-01 1.36284292e+00 -4.96891469e-01
1.16988587e+00 -1.83070898e+00 -4.64508981e-01 2.60717534e-02
7.20206276e-02 2.92326272e-01 -2.12747883e-02 3.00693493e-02
2.18366474e-01 2.21375197e-01 -4.84861851e-01 -5.62246680e-01
-1.36319131e-01 3.01948860e-02 -2.50006407e-01 5.73531449e-01
2.65951842e-01 8.72881174e-01 -6.45563841e-01 -1.26242125e+00
4.56394404e-01 1.86192885e-01 -2.82799780e-01 7.50054270e-02
-5.27556777e-01 3.07336450e-01 -6.60726964e-01 1.02180731e+00
8.35069895e-01 -3.61891329e-01 2.18753889e-02 -2.02110767e-01
-2.07074642e-01 -2.23642692e-01 -1.26389754e+00 1.66925490e+00
-7.97585994e-02 8.35083783e-01 1.41912997e-01 -8.89107049e-01
9.16153848e-01 1.12827145e-01 3.42038393e-01 -2.37873003e-01
3.13263506e-01 2.36055851e-01 -3.36953998e-01 -4.56120849e-01
6.39252901e-01 2.32914656e-01 -1.95132419e-01 4.65548307e-01
-2.04775959e-01 -6.12994432e-01 2.50761300e-01 2.50134557e-01
8.90694678e-01 5.24557948e-01 1.50186375e-01 -4.24319178e-01
6.67056680e-01 2.90602505e-01 5.08243680e-01 8.95326853e-01
-3.30594838e-01 1.03123498e+00 3.30882341e-01 -3.87521803e-01
-1.30187201e+00 -7.04676807e-01 -3.75914097e-01 1.33744931e+00
3.74785870e-01 -3.37756783e-01 -1.33479178e+00 -1.14546633e+00
-3.10532987e-01 4.39491123e-01 -7.16666818e-01 5.93874753e-01
-7.04147100e-01 -7.54748344e-01 5.77546000e-01 9.84536767e-01
8.71924818e-01 -7.90913641e-01 -3.73940587e-01 -8.16938058e-02
-6.73837513e-02 -1.62864280e+00 -7.87552416e-01 4.27031517e-01
-9.80495632e-01 -9.13952470e-01 -9.81601059e-01 -1.05464721e+00
1.11694336e+00 1.70231670e-01 1.03770423e+00 9.80211347e-02
-4.50529665e-01 1.71115220e-01 -4.31274623e-01 -2.89069742e-01
-2.43543431e-01 1.60969615e-01 -3.97842735e-01 4.60783131e-02
2.40976080e-01 3.46265405e-01 -6.10102415e-01 7.58095562e-01
-1.01557934e+00 1.58106744e-01 6.69146419e-01 8.50652456e-01
9.67148066e-01 -2.65053771e-02 1.81770921e-01 -1.40602517e+00
1.31667927e-01 3.04130793e-01 -8.25494945e-01 1.95163831e-01
-3.18652451e-01 4.91970126e-03 5.50615251e-01 -5.92886984e-01
-1.33294964e+00 8.16928208e-01 1.07622193e-02 -1.31648406e-01
-3.81458849e-01 -1.12592310e-01 -2.22115040e-01 -6.09541386e-02
8.23879957e-01 2.34175157e-02 -6.10324800e-01 -4.31335181e-01
6.00634456e-01 9.36494172e-01 6.45668149e-01 -6.75391972e-01
9.56346750e-01 1.05059624e+00 -2.34615803e-01 -9.21153486e-01
-1.21161890e+00 -8.82719696e-01 -1.30098045e+00 -1.09488644e-01
1.14864635e+00 -7.84034848e-01 -7.88578168e-02 5.59570730e-01
-9.54393923e-01 -7.87715793e-01 -2.77630329e-01 2.34261945e-01
-4.38762516e-01 6.92247689e-01 -6.59291089e-01 -7.06085384e-01
-3.75963598e-01 -1.08674240e+00 1.75160348e+00 2.26821437e-01
7.38927051e-02 -7.50306904e-01 -2.61843503e-01 6.51668906e-01
-2.99737632e-01 2.03523412e-01 4.49902356e-01 -6.45880997e-01
-5.48858583e-01 -6.32170975e-01 -5.84144711e-01 2.50466853e-01
-9.43759084e-02 1.30243048e-01 -1.08353782e+00 -3.67554985e-02
-5.10419607e-01 -6.41431868e-01 9.56118703e-01 4.04086411e-01
1.54544151e+00 2.63665199e-01 -6.65258586e-01 2.93533832e-01
1.25004888e+00 -1.14388652e-01 3.42012078e-01 4.35748324e-02
9.92853045e-01 6.44471824e-01 9.71009254e-01 2.30032071e-01
1.74894661e-01 4.89513487e-01 1.15302093e-01 -6.13734484e-01
-1.88659564e-01 -2.29063436e-01 2.97634341e-02 4.40159529e-01
1.32288903e-01 -4.59055930e-01 -1.18439353e+00 6.55062199e-01
-1.95173538e+00 -5.77722549e-01 -6.62420154e-01 1.91523635e+00
1.02237892e+00 5.88773906e-01 2.94345409e-01 3.02248269e-01
1.05180538e+00 -7.93643966e-02 -5.95253229e-01 3.52724418e-02
-4.16377820e-02 1.13220640e-01 8.35085988e-01 3.73259157e-01
-1.50530362e+00 1.39138985e+00 6.63078356e+00 1.01429546e+00
-8.82064581e-01 1.33147553e-01 8.80157471e-01 1.31975040e-01
2.63179928e-01 -1.98982000e-01 -1.21876478e+00 3.00338894e-01
4.10848916e-01 3.10502470e-01 -2.42600828e-01 1.11274171e+00
3.14215451e-01 -5.56103647e-01 -1.09652257e+00 8.91684830e-01
3.85580868e-01 -9.64740574e-01 -2.50877798e-01 -1.15923725e-01
1.30776596e+00 6.80327713e-02 -1.41695917e-01 1.63710639e-01
4.77485150e-01 -6.82500958e-01 7.66136050e-01 1.50848836e-01
1.02223957e+00 -3.65399957e-01 6.04016781e-01 5.68395734e-01
-1.44551003e+00 2.92331964e-01 -3.96589518e-01 4.02477443e-01
7.60164410e-02 6.48072243e-01 -8.79017353e-01 8.31261650e-02
7.59577751e-01 5.74048758e-01 -8.95534039e-01 8.62959146e-01
-4.25045580e-01 8.97236168e-01 -5.91701388e-01 -2.88990408e-01
3.61658216e-01 -7.15317875e-02 3.50999720e-02 1.57959330e+00
-1.28829554e-02 2.74100471e-02 7.74864614e-01 7.67278314e-01
-3.64426494e-01 3.73289198e-01 -2.21529961e-01 1.94051832e-01
8.64158422e-02 1.49129903e+00 -1.56997287e+00 -6.29705846e-01
-8.02676558e-01 1.20597613e+00 1.45725563e-01 3.52575362e-01
-7.76566029e-01 -3.85141730e-01 -5.23417830e-01 3.85657281e-01
5.93250871e-01 -9.88664478e-02 -7.76424110e-01 -1.13604462e+00
1.32990539e-01 -4.17475015e-01 3.33293229e-01 -1.07563710e+00
-1.13054049e+00 5.47616899e-01 -1.83228210e-01 -1.36205900e+00
2.03513920e-01 -9.24142301e-01 -5.94656348e-01 6.62613213e-01
-1.46669936e+00 -1.52148938e+00 -6.59822822e-01 4.88318413e-01
1.22772479e+00 1.51938975e-01 3.36605251e-01 -5.52380309e-02
-6.77805185e-01 6.21152401e-01 2.01692104e-01 7.17905164e-01
8.39170933e-01 -1.51053691e+00 3.83041829e-01 9.75139320e-01
2.08693549e-01 1.06194206e-01 4.90537345e-01 -7.54398227e-01
-8.02740097e-01 -1.25908291e+00 4.47319716e-01 -5.62081218e-01
6.57738149e-01 -7.49059796e-01 -9.94589031e-01 6.68054223e-01
5.20963781e-02 6.41300440e-01 4.21583414e-01 -2.71009743e-01
-1.52034387e-01 1.19627126e-01 -1.18475235e+00 5.99663436e-01
8.36191297e-01 -2.70424813e-01 -5.51990986e-01 6.21566594e-01
4.87630665e-01 -7.76709676e-01 -4.03667659e-01 3.31850052e-01
4.43843037e-01 -8.13227654e-01 5.50651610e-01 -1.06677078e-01
2.87994951e-01 -4.48137194e-01 1.67406008e-01 -6.50723219e-01
3.31848234e-01 -4.82917547e-01 1.50283396e-01 1.30293739e+00
6.80564225e-01 -1.76734313e-01 1.40904105e+00 5.50794601e-01
-2.63840884e-01 -6.94628775e-01 -6.39242172e-01 -6.14719152e-01
3.83641362e-01 -4.25435394e-01 -1.06070861e-01 7.74711967e-01
-1.70676205e-02 1.84877113e-01 -5.78839779e-02 6.27633557e-02
5.53571045e-01 2.75698930e-01 9.02918518e-01 -1.30073285e+00
5.61613552e-02 -2.04991594e-01 -2.40128234e-01 -1.47430360e+00
1.71367899e-01 -6.02325976e-01 6.19782448e-01 -1.54491735e+00
3.18175077e-01 -2.79928178e-01 2.86407441e-01 6.66568100e-01
-4.62422222e-01 6.55409455e-01 -1.33272603e-01 1.25908911e-01
-8.32097530e-01 2.49736965e-01 1.31915998e+00 -2.46103317e-01
-1.98774815e-01 1.22687034e-02 -1.52209431e-01 1.30858564e+00
7.29236186e-01 -4.76059288e-01 -2.50961930e-01 -1.46714419e-01
-1.26070067e-01 -2.48446032e-01 1.35771722e-01 -8.01749587e-01
3.12988490e-01 2.42716074e-02 6.49064600e-01 -1.42916918e+00
6.68838397e-02 -9.84034538e-01 -7.98660994e-01 6.16897531e-02
-3.57547909e-01 -3.93221855e-01 2.19567448e-01 7.36523867e-01
8.91841426e-02 -5.22730231e-01 1.16992915e+00 -9.29169506e-02
-6.68707967e-01 1.72703579e-01 -3.91067594e-01 4.87941563e-01
1.34824729e+00 -3.88431489e-01 -2.41113007e-01 -7.64709488e-02
-7.64791131e-01 3.04426849e-01 6.70247614e-01 1.85469121e-01
3.43632430e-01 -8.96841228e-01 -4.91743982e-01 -2.41169166e-02
3.65956366e-01 4.54959929e-01 -1.12876713e-01 5.76157331e-01
-8.32329452e-01 5.50149083e-01 3.40983182e-01 -1.01844716e+00
-1.26635194e+00 5.43794870e-01 2.73892850e-01 -3.49662989e-01
-8.22420657e-01 7.19519019e-01 2.16191381e-01 -3.52410465e-01
3.70569587e-01 -7.71188676e-01 1.95510820e-01 -1.41491964e-01
2.62070894e-01 2.35926434e-01 1.31361380e-01 -6.12212837e-01
-1.84627280e-01 1.06068170e+00 -2.31985226e-01 -2.88605183e-01
7.25162029e-01 -3.44726816e-02 3.85750800e-01 4.92346168e-01
8.10398400e-01 2.65785575e-01 -1.55054283e+00 -3.64713132e-01
1.87090963e-01 -3.84476215e-01 2.79324025e-01 -6.76584899e-01
-1.06349993e+00 1.03055644e+00 5.13462603e-01 3.24005969e-02
8.98812592e-01 1.79606140e-01 6.86543703e-01 4.15730596e-01
2.61371106e-01 -1.69362533e+00 5.52173257e-01 8.37058127e-02
3.82993877e-01 -1.65733421e+00 2.26008937e-01 -9.64455903e-01
-7.59219646e-01 1.16982627e+00 9.65733111e-01 -2.13855579e-02
4.90904808e-01 7.96850026e-01 3.07933807e-01 6.04951754e-02
-1.82621226e-01 -5.53561211e-01 4.02106553e-01 5.28214037e-01
6.90575421e-01 -2.97318757e-01 -4.46939766e-01 4.69911873e-01
2.78822839e-01 -1.42787859e-01 5.65835416e-01 1.04141295e+00
-6.36161506e-01 -1.08727431e+00 -6.98704541e-01 5.54038465e-01
-5.80748975e-01 -9.77405310e-02 -6.78850710e-01 1.04723763e+00
-5.98951941e-03 9.03721333e-01 1.86280265e-01 2.19072565e-01
2.09955841e-01 2.10039675e-01 1.41807631e-01 -8.47965062e-01
-4.15932059e-01 6.78592682e-01 -1.19985342e-01 -2.30125964e-01
-5.04794180e-01 -5.68838775e-01 -1.66962993e+00 2.80388564e-01
-9.16410446e-01 -1.75975859e-01 4.31913793e-01 7.45100260e-01
-6.92494661e-02 3.70600104e-01 3.70625407e-01 -9.21783149e-01
-1.83650002e-01 -1.05930007e+00 -6.03589654e-01 3.26714814e-01
7.53048658e-02 -3.33044946e-01 -3.43933016e-01 7.54975617e-01] | [12.021356582641602, 2.2584116458892822] |
45952f9f-488f-46cd-9d7d-35d7fb93c745 | bbbd-bounding-box-based-detector-for | 2204.12841 | null | https://arxiv.org/abs/2204.12841v1 | https://arxiv.org/pdf/2204.12841v1.pdf | BBBD: Bounding Box Based Detector for Occlusion Detection and Order Recovery | Occlusion handling is one of the challenges of object detection and segmentation, and scene understanding. Because objects appear differently when they are occluded in varying degree, angle, and locations. Therefore, determining the existence of occlusion between objects and their order in a scene is a fundamental requirement for semantic understanding. Existing works mostly use deep learning based models to retrieve the order of the instances in an image or for occlusion detection. This requires labelled occluded data and it is time consuming. In this paper, we propose a simpler and faster method that can perform both operations without any training and only requires the modal segmentation masks. For occlusion detection, instead of scanning the two objects entirely, we only focus on the intersected area between their bounding boxes. Similarly, we use the segmentation mask inside the same area to recover the depth-ordering. When tested on COCOA dataset, our method achieves +8% and +5% more accuracy than the baselines in order recovery and occlusion detection respectively. | ['Zoltan Vamossy', 'Kaziwa Saleh'] | 2022-04-27 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [ 2.92359143e-01 -1.82877213e-01 -1.45110205e-01 -3.59833270e-01
-3.09871167e-01 -7.49800622e-01 2.02328399e-01 4.36303020e-01
-2.33317703e-01 1.39105886e-01 -1.87100321e-01 -2.44606152e-01
2.11711928e-01 -8.84057343e-01 -8.79782140e-01 -5.38657606e-01
1.34387612e-01 7.16157913e-01 7.19173312e-01 2.51233220e-01
3.34386796e-01 7.04782009e-01 -1.53291368e+00 4.99925315e-01
9.10272598e-01 1.12744606e+00 4.43835318e-01 4.52411622e-01
-4.64944392e-01 4.47734565e-01 -6.20968342e-01 -1.06154844e-01
4.70016479e-01 3.11013479e-02 -7.95455456e-01 5.79335630e-01
7.21672595e-01 -6.92842722e-01 -2.47580931e-01 9.12586987e-01
1.10637203e-01 -1.34018481e-01 6.31015360e-01 -1.12088728e+00
-1.93746462e-01 1.34428203e-01 -1.01980424e+00 2.24993169e-01
2.67987162e-01 -1.26122609e-01 7.98480332e-01 -7.75558889e-01
5.00323832e-01 1.24989319e+00 2.75412083e-01 9.32775512e-02
-1.06929994e+00 -6.34103656e-01 4.02139008e-01 2.92537928e-01
-1.26787889e+00 -2.48180658e-01 7.03980744e-01 -5.66213131e-01
4.43075269e-01 3.15680683e-01 4.87439215e-01 3.19142640e-01
-3.17107797e-01 1.14379275e+00 9.03454244e-01 -5.00000119e-01
1.78158164e-01 1.93496197e-01 3.31623912e-01 4.98397499e-01
2.84276992e-01 -3.83034468e-01 -4.83335331e-02 2.12943166e-01
8.77824783e-01 1.46881610e-01 -8.87681991e-02 -4.52562779e-01
-8.18705082e-01 6.91864312e-01 6.11006618e-01 4.37646778e-03
-1.50502533e-01 1.91832259e-02 2.14676455e-01 -2.52140880e-01
3.33175629e-01 2.08141819e-01 -5.79828918e-01 3.93372357e-01
-1.02358031e+00 1.82283774e-01 7.68737853e-01 1.02111709e+00
1.08513093e+00 -4.53526437e-01 -1.56106725e-01 6.23867333e-01
2.40902632e-01 1.24132432e-01 -1.29292935e-01 -7.65068829e-01
5.36994278e-01 1.03084064e+00 2.64051050e-01 -1.31514800e+00
-4.28056508e-01 -4.59503263e-01 -5.01965463e-01 -2.23872941e-02
6.58815145e-01 1.55834824e-01 -1.33287704e+00 1.17558205e+00
9.15295959e-01 2.31757879e-01 -3.33056509e-01 1.03078830e+00
9.20057774e-01 6.38912141e-01 7.69685879e-02 4.95364750e-03
1.75589788e+00 -1.17869759e+00 -6.05310738e-01 -6.39729857e-01
5.71237624e-01 -1.13899016e+00 8.78268838e-01 2.88641334e-01
-8.68002176e-01 -5.90537727e-01 -1.06368899e+00 -5.25126219e-01
-3.19854081e-01 4.88100022e-01 9.26168323e-01 2.91156113e-01
-3.94526690e-01 3.10513586e-01 -7.89756715e-01 -9.58857015e-02
6.67787313e-01 3.27247947e-01 -1.90430120e-01 -3.32025141e-01
-6.88637793e-01 5.46486437e-01 4.16902214e-01 1.31069124e-01
-7.59429276e-01 -5.47293186e-01 -8.84787798e-01 1.12155259e-01
5.59102595e-01 -1.58067718e-01 9.03750420e-01 -7.57411182e-01
-7.58176029e-01 1.01511049e+00 -4.62813467e-01 -3.67633224e-01
6.15254998e-01 -6.17547870e-01 2.57053059e-02 1.26018345e-01
3.35228473e-01 8.10780764e-01 6.69027567e-01 -1.29416227e+00
-8.40343416e-01 -5.05695999e-01 3.65124732e-01 3.14415514e-01
8.01804126e-04 -4.98446077e-02 -9.82102990e-01 -3.61790836e-01
9.11452293e-01 -7.69995332e-01 -1.31885648e-01 2.69000739e-01
-6.28022671e-01 -3.88917357e-01 1.27651548e+00 -9.25877213e-01
1.15578008e+00 -2.15827036e+00 -2.71231622e-01 7.01904967e-02
1.79080307e-01 3.12044561e-01 1.15890466e-01 1.40647262e-01
1.32197738e-01 1.84888393e-01 -2.00842664e-01 -2.61413306e-01
-9.46502313e-02 2.87624121e-01 -3.97942454e-01 4.96302128e-01
1.96327016e-01 6.61761642e-01 -6.36897385e-01 -7.24209249e-01
4.90456730e-01 4.00135487e-01 -5.70809424e-01 2.18822792e-01
-4.16000694e-01 5.60813427e-01 -4.13847536e-01 7.38100410e-01
1.15453494e+00 -1.47414595e-01 1.83544114e-01 -5.04999995e-01
-6.36781231e-02 4.11752492e-01 -1.65611064e+00 1.46039319e+00
-4.24272269e-01 7.07491040e-01 -5.15560955e-02 -9.72288549e-01
7.21634686e-01 1.71410979e-03 2.96718717e-01 -8.33532095e-01
1.42810062e-01 2.29676545e-01 1.09049097e-01 -6.73261285e-01
2.47970879e-01 4.54637975e-01 1.76447093e-01 3.20634991e-01
-4.83041108e-01 5.57772703e-02 5.61657369e-01 2.96170488e-02
6.96277916e-01 1.39254645e-01 2.08511606e-01 -1.10041335e-01
4.79476213e-01 -3.39409858e-02 8.29696357e-01 6.08303845e-01
-1.04753993e-01 7.40799606e-01 5.71491361e-01 -7.61378407e-01
-8.84815753e-01 -7.48046219e-01 -4.28303987e-01 1.01055050e+00
7.62340724e-01 -1.79710895e-01 -7.94216573e-01 -5.93066633e-01
-4.21465794e-03 4.45876330e-01 -4.93249446e-01 3.63181442e-01
-1.00780571e+00 -6.61547959e-01 -7.43281543e-02 6.17265344e-01
7.78463125e-01 -9.30308759e-01 -7.92644382e-01 1.06942549e-01
-5.07127166e-01 -1.47136867e+00 -4.44916964e-01 1.02577642e-01
-7.88625896e-01 -1.31473386e+00 -2.14468911e-01 -9.58940327e-01
9.08625484e-01 5.50717771e-01 1.07607436e+00 3.02778393e-01
-6.60732150e-01 -4.47250277e-01 -9.55008268e-02 -4.35155481e-01
1.74290255e-01 1.08372815e-01 -5.12570739e-01 -2.30498705e-02
2.67300308e-01 -4.00807500e-01 -9.58266735e-01 6.21520519e-01
-8.45238149e-01 2.63178706e-01 5.32166064e-01 2.24096894e-01
8.16303194e-01 1.53409913e-01 1.28244355e-01 -9.73172367e-01
-1.51618168e-01 -1.56418234e-01 -8.07301760e-01 2.86510855e-01
-1.49241770e-02 -3.99497487e-02 1.90134704e-01 -3.21777076e-01
-7.67089546e-01 3.88043433e-01 1.32250190e-01 -3.32253605e-01
-5.39612710e-01 -4.45256867e-02 -5.50163090e-01 1.68027192e-01
1.94213673e-01 -1.38471037e-01 -6.47635221e-01 -7.14664042e-01
2.87601084e-01 5.06586432e-01 4.05651331e-01 -3.07248414e-01
5.21294713e-01 8.69756579e-01 -6.68852776e-02 -7.25207865e-01
-1.11911690e+00 -8.59567881e-01 -9.78941321e-01 7.95913115e-02
9.81861055e-01 -7.55542874e-01 -5.66430807e-01 2.48234272e-01
-1.54354501e+00 -4.97207232e-02 1.32308885e-01 2.13368341e-01
4.82204966e-02 3.50452751e-01 -4.64454263e-01 -7.76525140e-01
-2.30893835e-01 -1.31063354e+00 1.35299599e+00 3.45856041e-01
1.78062299e-03 -5.71472168e-01 -2.97657847e-01 5.00471175e-01
-2.75200643e-02 2.70522684e-01 9.83411908e-01 -5.15971959e-01
-1.15319562e+00 -4.01867867e-01 -7.07448721e-01 -7.43398294e-02
2.99470037e-01 5.07077985e-02 -9.54705298e-01 -1.08720055e-02
-1.75217658e-01 3.43745351e-02 8.46204758e-01 4.65617388e-01
1.41492271e+00 -3.62340838e-01 -6.57733858e-01 7.06036150e-01
1.43675244e+00 3.52710068e-01 6.45564377e-01 3.49416703e-01
9.08469260e-01 8.68855953e-01 7.16782570e-01 1.98396355e-01
2.66592622e-01 8.25319052e-01 6.20539665e-01 -3.74610364e-01
-2.72716254e-01 -1.28112435e-01 -2.44850665e-01 3.07799399e-01
2.68193543e-01 -3.94332737e-01 -8.22531462e-01 5.80088198e-01
-1.73408937e+00 -4.79893357e-01 -4.40616697e-01 2.25622582e+00
6.84351563e-01 1.86176375e-01 -5.21828867e-02 2.35429049e-01
9.92731392e-01 7.15427175e-02 -7.15083003e-01 -2.45012715e-01
1.24106772e-01 1.33422121e-01 4.82239008e-01 5.44385850e-01
-1.44116426e+00 1.13391662e+00 5.92157030e+00 8.34117770e-01
-9.84943211e-01 -8.69113505e-02 7.83430934e-01 -5.48348911e-02
-1.36142522e-02 2.55897790e-01 -1.11019123e+00 3.51111978e-01
-8.12105313e-02 5.66211998e-01 1.51186034e-01 8.40445757e-01
1.49922639e-01 -6.53300405e-01 -1.13165629e+00 8.91268134e-01
-1.19329086e-02 -8.15992534e-01 -1.03750825e-01 -4.04909849e-02
7.67766118e-01 -4.41331446e-01 -9.87761319e-02 -4.12425958e-02
-4.96051386e-02 -1.00700736e+00 6.99683666e-01 -7.19428286e-02
3.06442648e-01 -5.91885269e-01 5.76396644e-01 4.16131496e-01
-1.37662089e+00 1.73066422e-01 -4.91876185e-01 -1.33365616e-01
1.58973426e-01 8.13511193e-01 -8.66191566e-01 2.62131631e-01
6.47900283e-01 4.53240693e-01 -6.05364203e-01 1.39357698e+00
-5.40912747e-01 4.04459029e-01 -5.69525003e-01 3.97563696e-01
1.22815341e-01 -2.54614264e-01 3.52112800e-01 1.01378953e+00
-8.94964933e-02 -5.59107773e-02 7.22298920e-01 8.67281735e-01
-6.42366484e-02 9.65161249e-02 -1.43953234e-01 3.15371484e-01
6.22290909e-01 1.45739174e+00 -1.26396775e+00 -4.00215119e-01
-3.16516548e-01 9.84850705e-01 3.21215808e-01 3.32711518e-01
-1.11634862e+00 -3.09880733e-01 3.80156606e-01 3.86973262e-01
4.62988555e-01 -2.26433039e-01 -5.02593696e-01 -8.21411610e-01
4.06047612e-01 -5.78258693e-01 2.61388123e-01 -4.77390796e-01
-7.49680340e-01 3.97342533e-01 -6.27175793e-02 -1.08818948e+00
2.83076346e-01 -4.47501272e-01 -5.43516517e-01 6.47747040e-01
-1.49555922e+00 -1.01288331e+00 -7.13417709e-01 1.46111980e-01
8.76793146e-01 7.55067468e-01 3.13790023e-01 7.01162577e-01
-7.64651299e-01 3.02718580e-01 -3.28348428e-01 3.33470285e-01
4.20038044e-01 -1.06248271e+00 2.76204616e-01 9.15180147e-01
2.44790912e-01 4.01168585e-01 7.19860256e-01 -6.19328320e-01
-8.58938456e-01 -1.09693229e+00 6.79710805e-01 -8.80205780e-02
8.57967287e-02 -5.78855932e-01 -1.04065883e+00 6.02393985e-01
6.10029362e-02 9.11847055e-02 2.36046463e-01 1.45581998e-02
-3.18674892e-01 -1.02985926e-01 -9.25321221e-01 5.42070687e-01
1.08184564e+00 -2.15504244e-01 -2.20045969e-01 6.40986204e-01
7.85947561e-01 -7.36225247e-01 -3.30223233e-01 7.13584006e-01
6.34549618e-01 -1.01881146e+00 1.13574016e+00 -3.71642619e-01
3.33128303e-01 -6.76592946e-01 -4.27873321e-02 -4.63775605e-01
-7.72838993e-03 -1.24302499e-01 3.43413092e-03 1.17995167e+00
2.19020367e-01 -2.30559677e-01 9.78558242e-01 6.07602596e-01
3.82880978e-02 -7.83169746e-01 -5.51016271e-01 -3.56297672e-01
-3.72713685e-01 -3.05499017e-01 5.65660894e-01 7.69028187e-01
-7.04504907e-01 2.39288464e-01 -1.24658987e-01 5.69946945e-01
4.93468851e-01 6.26994550e-01 9.45000350e-01 -1.19779098e+00
-1.48367332e-02 -1.46968067e-01 -3.56754810e-01 -1.44387901e+00
-1.88847333e-01 -5.40146172e-01 6.97506517e-02 -1.88504624e+00
2.27906212e-01 -6.80809796e-01 6.19915351e-02 4.65366095e-01
-2.57772058e-01 3.71229738e-01 -1.09104542e-02 3.08630645e-01
-7.08724320e-01 1.36162639e-01 1.27104962e+00 -5.72476201e-02
-4.19619679e-01 4.92742956e-02 -3.78042698e-01 1.16180229e+00
7.80388534e-01 -4.61903840e-01 -2.94130772e-01 -8.49857152e-01
-3.82690094e-02 -2.27061316e-01 4.35137004e-01 -1.04589069e+00
1.89994857e-01 -1.06937014e-01 6.53973520e-01 -1.28907394e+00
3.42660755e-01 -8.90641034e-01 -4.55828048e-02 4.44643706e-01
-9.73492786e-02 -1.24191530e-01 2.29339644e-01 4.22896951e-01
9.21695959e-03 -5.17973721e-01 6.78559065e-01 -3.60780030e-01
-8.40176582e-01 3.80409092e-01 -1.76765975e-02 -9.98372734e-02
1.18793082e+00 -3.28751415e-01 -2.01948881e-01 -2.37649474e-02
-7.56376684e-01 3.64716768e-01 3.86546433e-01 3.60501677e-01
5.28759778e-01 -9.47501361e-01 -4.17935073e-01 2.74006724e-01
-9.26929191e-02 6.15296602e-01 2.31632471e-01 7.20323622e-01
-8.83500695e-01 4.59909976e-01 2.00368404e-01 -8.74450862e-01
-1.38591480e+00 5.61195433e-01 2.83986986e-01 -1.52784482e-01
-5.18192947e-01 8.30533624e-01 9.24629927e-01 -1.71095401e-01
3.79291862e-01 -3.62406790e-01 -3.03518862e-01 1.02505989e-01
5.03117740e-01 4.17952389e-01 -5.55491308e-03 -4.72149909e-01
-3.96282971e-01 8.30225587e-01 -3.65768760e-01 1.94091141e-01
8.21046472e-01 -8.58698785e-02 -2.92340189e-01 1.50021687e-01
1.19296741e+00 1.92958396e-02 -1.25281262e+00 -2.91942805e-01
6.20185733e-02 -7.50794470e-01 -6.75203502e-02 -4.31383282e-01
-1.22272921e+00 1.13962173e+00 5.76269388e-01 2.11088449e-01
9.78649020e-01 2.56340414e-01 7.98914671e-01 6.09158957e-03
1.53419629e-01 -9.09558773e-01 6.59312978e-02 2.62401044e-01
6.41655385e-01 -1.34358954e+00 3.06399852e-01 -1.25470376e+00
-1.56261802e-01 8.61089945e-01 9.99073446e-01 -1.89059466e-01
3.79201055e-01 1.64684042e-01 -5.41148745e-02 -2.61718839e-01
-2.42834106e-01 -4.64790553e-01 4.81454134e-01 3.41825038e-01
3.45268786e-01 3.40483673e-02 -3.79822880e-01 5.45055121e-02
-5.81443720e-02 -5.81266761e-01 1.29454806e-01 9.02883947e-01
-6.95229471e-01 -9.32673514e-01 -3.77389014e-01 3.70207787e-01
-5.05541801e-01 -1.38503060e-01 -2.34715924e-01 7.49171615e-01
5.53341091e-01 1.03058016e+00 3.90362859e-01 1.70286939e-01
2.50009060e-01 -2.28475735e-01 3.62854481e-01 -8.00354242e-01
-2.84131244e-02 4.65180784e-01 -1.06245592e-01 -5.52183509e-01
-4.00527954e-01 -5.64770937e-01 -1.46238697e+00 1.06476806e-01
-7.73436427e-01 -1.76449016e-01 7.05563545e-01 1.01612377e+00
3.00050586e-01 4.88871545e-01 5.43687880e-01 -6.82060242e-01
7.16646016e-02 -8.18407595e-01 -2.76652217e-01 5.48582852e-01
1.67659491e-01 -7.90552199e-01 -2.23426029e-01 2.34929956e-02] | [9.382972717285156, 0.46351781487464905] |
980f7494-7e00-42a7-91f7-f92e97edd52a | another-vertical-view-a-hierarchical-network | 2304.05106 | null | https://arxiv.org/abs/2304.05106v1 | https://arxiv.org/pdf/2304.05106v1.pdf | Another Vertical View: A Hierarchical Network for Heterogeneous Trajectory Prediction via Spectrums | With the fast development of AI-related techniques, the applications of trajectory prediction are no longer limited to easier scenes and trajectories. More and more heterogeneous trajectories with different representation forms, such as 2D or 3D coordinates, 2D or 3D bounding boxes, and even high-dimensional human skeletons, need to be analyzed and forecasted. Among these heterogeneous trajectories, interactions between different elements within a frame of trajectory, which we call the ``Dimension-Wise Interactions'', would be more complex and challenging. However, most previous approaches focus mainly on a specific form of trajectories, which means these methods could not be used to forecast heterogeneous trajectories, not to mention the dimension-wise interaction. Besides, previous methods mostly treat trajectory prediction as a normal time sequence generation task, indicating that these methods may require more work to directly analyze agents' behaviors and social interactions at different temporal scales. In this paper, we bring a new ``view'' for trajectory prediction to model and forecast trajectories hierarchically according to different frequency portions from the spectral domain to learn to forecast trajectories by considering their frequency responses. Moreover, we try to expand the current trajectory prediction task by introducing the dimension $M$ from ``another view'', thus extending its application scenarios to heterogeneous trajectories vertically. Finally, we adopt the bilinear structure to fuse two factors, including the frequency response and the dimension-wise interaction, to forecast heterogeneous trajectories via spectrums hierarchically in a generic way. Experiments show that the proposed model outperforms most state-of-the-art methods on ETH-UCY, Stanford Drone Dataset and nuScenes with heterogeneous trajectories, including 2D coordinates, 2D and 3D bounding boxes. | ['Xinge You', 'Qinmu Peng', 'Beihao Xia', 'Conghao Wong'] | 2023-04-11 | null | null | null | null | ['trajectory-prediction'] | ['computer-vision'] | [-3.94080609e-01 -5.35551012e-01 -1.56986073e-01 -7.09981248e-02
-1.21485926e-01 -7.11082995e-01 7.60460794e-01 -1.79556519e-01
-1.88540310e-01 7.29770422e-01 2.80418515e-01 -9.04569179e-02
-2.52416819e-01 -9.85785007e-01 -6.50294185e-01 -8.76759410e-01
-5.27600050e-01 3.81382078e-01 5.27077615e-01 -5.96644521e-01
-1.08266577e-01 4.14741844e-01 -1.68129432e+00 1.26921356e-01
7.86476612e-01 6.46687806e-01 -2.04006713e-02 3.20762187e-01
3.93656790e-02 5.27772665e-01 -4.57659662e-01 -5.04236042e-01
3.11805785e-01 -3.10549498e-01 -3.36511642e-01 -1.47005081e-01
-6.94349483e-02 -3.92309964e-01 -5.63934565e-01 8.81822348e-01
2.48012260e-01 4.20903742e-01 7.89669812e-01 -1.61082721e+00
-3.81286979e-01 5.64087093e-01 -6.33715093e-01 1.19100362e-01
5.31560063e-01 2.22241879e-01 6.95843518e-01 -5.42231858e-01
7.31289625e-01 1.35088646e+00 9.13068473e-01 4.96724576e-01
-1.01641154e+00 -8.52047205e-01 6.99193478e-01 5.29439092e-01
-1.38982260e+00 1.51859835e-01 7.65500903e-01 -7.75862634e-01
7.08978057e-01 3.86623204e-01 9.80515003e-01 1.45208836e+00
-5.29860221e-02 8.52531612e-01 7.43106127e-01 3.99637520e-01
-5.89179769e-02 -1.43543631e-01 2.99942493e-02 3.77330184e-01
-8.01869854e-02 3.77603948e-01 -3.19611691e-02 -2.77388990e-01
4.88457829e-01 3.99647713e-01 -3.23998243e-01 -8.81220624e-02
-1.67069554e+00 6.81202829e-01 3.24879795e-01 2.15484127e-01
-2.80753523e-01 -7.23336115e-02 4.14721966e-01 -8.27835174e-04
4.91647273e-01 -1.16674090e-02 -4.12734061e-01 -4.02077913e-01
-7.36481190e-01 9.40778255e-01 5.30502260e-01 1.25504851e+00
6.71504498e-01 3.55139300e-02 -3.13764840e-01 6.49105608e-01
2.64160275e-01 6.66616023e-01 4.44783837e-01 -7.13506043e-01
7.09197819e-01 8.09216201e-01 4.43486959e-01 -1.53132224e+00
-8.26440036e-01 -2.24693581e-01 -1.17401671e+00 -2.88345426e-01
4.42908734e-01 -5.16515374e-01 -5.06666481e-01 1.98740077e+00
7.05965161e-01 8.25351357e-01 -4.04472388e-02 1.07570779e+00
7.77698815e-01 9.99465704e-01 -7.56342933e-02 -3.56730253e-01
1.15466535e+00 -9.90406692e-01 -5.84763646e-01 4.56853628e-01
7.59773612e-01 -5.01441598e-01 8.19343925e-01 2.42222294e-01
-9.12219167e-01 -6.90838754e-01 -6.22038484e-01 2.72419810e-01
-4.56721008e-01 4.10537452e-01 6.19159043e-01 3.76598656e-01
-6.19151771e-01 5.98265171e-01 -9.38010991e-01 -3.17973197e-01
1.31831486e-02 1.59028098e-01 -1.37097299e-01 2.33062655e-01
-1.57205880e+00 6.59906566e-01 3.36677939e-01 2.04536587e-01
-7.32550621e-01 -6.88843369e-01 -7.18250275e-01 -1.19035870e-01
2.37451464e-01 -6.35918796e-01 7.73575664e-01 -3.09446782e-01
-1.46484375e+00 1.52810872e-01 -2.04008326e-01 -2.67994970e-01
8.13371599e-01 6.78098947e-02 -7.14565873e-01 -1.26080200e-01
2.99699958e-02 4.75515813e-01 6.49768054e-01 -1.02729905e+00
-8.83533955e-01 -3.25226188e-01 2.90549695e-01 1.83457151e-01
-2.68424779e-01 -1.57191068e-01 -4.31118876e-01 -1.04527187e+00
-1.78986460e-01 -1.48386765e+00 -3.53838682e-01 -4.17266756e-01
-3.27589601e-01 -5.21991551e-01 1.01092672e+00 -4.78922576e-01
1.65385878e+00 -2.02666259e+00 6.13954246e-01 2.81011835e-02
1.26071081e-01 2.18892485e-01 2.93553714e-02 7.02160478e-01
1.38758034e-01 6.44159466e-02 -2.72649616e-01 -3.33283544e-01
2.76055872e-01 -2.52526645e-02 -5.57189941e-01 5.43399990e-01
-3.44422273e-02 6.73261642e-01 -1.07655799e+00 -3.51147860e-01
2.58062720e-01 3.81036431e-01 -7.31185973e-01 8.55077207e-02
-3.24433893e-01 8.51776481e-01 -6.47162139e-01 3.78521234e-01
8.04479480e-01 -3.18569653e-02 -1.64425358e-01 -1.33081943e-01
-4.57572728e-01 -2.51874864e-01 -1.17171681e+00 1.62063217e+00
-2.16755643e-01 4.00207341e-01 -1.34495109e-01 -9.16872263e-01
7.02129066e-01 3.18899006e-01 9.85750854e-01 -1.05689600e-01
-3.49741131e-02 4.83311992e-03 1.33003984e-02 -7.78068721e-01
5.57838559e-01 5.73142581e-02 -3.62496734e-01 2.77243733e-01
-3.29366535e-01 1.60778373e-01 4.14042652e-01 3.09578925e-02
9.57470238e-01 4.78153378e-01 -1.70065060e-01 -3.39092277e-02
6.21781170e-01 2.76419312e-01 7.12372005e-01 2.89813727e-01
-1.47867814e-01 4.73847240e-01 3.40575069e-01 -6.40460134e-01
-9.21849906e-01 -8.70219350e-01 -2.19293073e-01 9.25140023e-01
7.08856344e-01 -6.56181872e-01 -7.40297735e-01 -6.36473715e-01
-3.17234062e-02 3.97758275e-01 -6.07517660e-01 4.26450139e-03
-8.31009686e-01 -1.13446617e+00 6.31758809e-01 3.52421731e-01
6.24685109e-01 -8.59229684e-01 -5.31454206e-01 2.65796632e-01
-4.39421713e-01 -1.08742130e+00 -5.83527207e-01 -6.34102225e-01
-5.70602477e-01 -9.44860339e-01 -1.01063585e+00 -3.15628737e-01
4.52395290e-01 3.81798923e-01 5.91685295e-01 -6.24558367e-02
1.54637313e-03 2.46086180e-01 -6.77628100e-01 -2.03587472e-01
1.70879886e-01 -8.05700198e-03 3.73614490e-01 2.55917460e-01
2.23811820e-01 -7.69025803e-01 -7.56950915e-01 7.27289975e-01
-7.93610394e-01 1.11528955e-01 1.21624216e-01 7.05427170e-01
3.03770930e-01 4.21883911e-01 5.83783090e-01 -6.20635927e-01
6.26667619e-01 -1.09287024e+00 -4.25436080e-01 1.03826270e-01
1.00160979e-01 -3.20233703e-01 1.11859143e+00 -8.98101270e-01
-1.01515508e+00 3.31132673e-02 -2.80752331e-01 -8.08558524e-01
-4.31946814e-01 4.34839725e-01 -1.23499379e-01 2.64928609e-01
4.75360394e-01 3.04018348e-01 -3.72106284e-01 -1.23784125e-01
4.48902667e-01 3.42240572e-01 9.11706686e-02 -6.73117340e-01
9.72689509e-01 6.71734929e-01 1.12191379e-01 -6.92967355e-01
-2.18241081e-01 -3.27837765e-01 -6.99481368e-01 -4.90991533e-01
9.94070768e-01 -8.84536147e-01 -1.22524548e+00 6.36965930e-01
-1.30656087e+00 -1.65540978e-01 1.03534199e-01 8.49858522e-01
-6.20817780e-01 5.11186182e-01 -5.49753785e-01 -7.35596836e-01
1.91683426e-01 -1.40140104e+00 1.26341891e+00 1.04740977e-01
-1.91223815e-01 -8.86933804e-01 3.05516392e-01 7.79068843e-02
1.65441111e-01 6.55503869e-01 7.92913616e-01 -4.33842570e-01
-5.70363224e-01 -4.14296798e-02 1.95093285e-02 -3.19344312e-01
-4.50782999e-02 1.41090378e-01 -4.68882650e-01 -1.81266323e-01
-1.41154960e-01 -2.78011933e-02 6.91962421e-01 2.84155756e-01
1.07147288e+00 -5.30316770e-01 -7.59233236e-01 7.35665083e-01
8.32580388e-01 4.81848985e-01 4.13107991e-01 4.19279002e-02
8.43908370e-01 1.07646072e+00 8.58549535e-01 6.18800223e-01
8.18389475e-01 1.10207546e+00 3.88539612e-01 2.97525674e-01
3.65910381e-01 -3.86991888e-01 5.58986843e-01 9.29192901e-01
-6.35589659e-01 -4.70510244e-01 -7.87828743e-01 4.76900905e-01
-2.37055469e+00 -1.43851602e+00 -4.14544672e-01 1.93850684e+00
2.90394068e-01 -2.74890155e-01 7.75552034e-01 -1.95094496e-01
9.12899256e-01 2.48580351e-01 -6.84271634e-01 1.22676983e-01
1.56118184e-01 -6.93847597e-01 1.62837088e-01 2.34079301e-01
-1.21446896e+00 8.49315226e-01 5.41238689e+00 1.20274460e+00
-1.04708028e+00 -4.83066738e-02 1.88671991e-01 -1.85926214e-01
-4.44553524e-01 -1.89254314e-01 -1.05293179e+00 1.12525642e+00
6.87275767e-01 -3.14106606e-02 6.03729188e-01 7.29095221e-01
3.45473856e-01 3.68935168e-01 -9.13210154e-01 1.13797402e+00
-1.81603879e-01 -1.18994904e+00 1.23893134e-01 1.29970044e-01
7.34631658e-01 -2.12850988e-01 1.18276633e-01 6.51485980e-01
3.49780411e-01 -8.83351505e-01 8.34167838e-01 6.81813776e-01
4.32621062e-01 -7.66952634e-01 4.51388747e-01 9.32754576e-01
-1.83172607e+00 -9.95665342e-02 -4.67259169e-01 -8.82896930e-02
5.67509353e-01 1.64621875e-01 -3.83658439e-01 9.64257061e-01
8.60038757e-01 1.36689198e+00 -7.68390223e-02 9.49011624e-01
3.33508074e-01 4.09719437e-01 -4.63931769e-01 -1.23697311e-01
3.67253870e-01 -6.69298530e-01 7.37372279e-01 1.09282625e+00
9.33397651e-01 6.64853454e-01 3.86718363e-01 8.36662114e-01
3.89234990e-01 1.40417293e-02 -8.65487754e-01 1.05238602e-01
4.82603282e-01 1.05816770e+00 -6.95257366e-01 -2.24392310e-01
-4.86477643e-01 6.06643736e-01 6.73872083e-02 5.19657433e-01
-1.44410992e+00 -1.37162626e-01 7.99539626e-01 1.96296230e-01
1.74031809e-01 -4.71952915e-01 3.25788528e-01 -1.56624126e+00
-7.34110102e-02 -5.67028821e-01 1.95554286e-01 -3.52013379e-01
-1.45575845e+00 8.98739636e-01 5.17591894e-01 -2.12324643e+00
-4.02187616e-01 -3.40701938e-01 -6.39769733e-01 5.32887459e-01
-1.06581759e+00 -1.21842909e+00 -2.40460634e-01 9.20249462e-01
5.15284956e-01 -1.35495797e-01 5.23135543e-01 6.77757025e-01
-7.77907133e-01 3.56148511e-01 1.36832491e-01 -1.18888587e-01
3.63159388e-01 -8.50884616e-01 3.33662122e-01 5.87930739e-01
7.39393979e-02 5.94425917e-01 6.58421576e-01 -7.95374930e-01
-1.22655880e+00 -1.30504060e+00 4.86484051e-01 -4.55004543e-01
7.67302513e-01 -3.85622442e-01 -7.10989833e-01 7.43623435e-01
-1.13461286e-01 -9.11660194e-02 5.82503617e-01 -2.28722282e-02
5.76723963e-02 1.04482174e-02 -9.05782521e-01 9.37050879e-01
1.54235911e+00 -1.33647755e-01 -3.24719638e-01 4.43593204e-01
8.38703394e-01 -4.92126435e-01 -1.00145006e+00 8.32257390e-01
6.66117430e-01 -1.09417486e+00 1.18385971e+00 -5.83282888e-01
2.67274201e-01 -6.75216019e-01 -8.33716767e-04 -1.39686787e+00
-2.85175860e-01 -5.00287831e-01 -1.45894349e-01 1.11353135e+00
1.82704270e-01 -5.47969818e-01 7.66301930e-01 2.30489746e-01
-3.02251130e-01 -1.01824880e+00 -9.24332261e-01 -7.52956033e-01
2.94176377e-02 -5.82105815e-01 1.21181214e+00 1.00495982e+00
1.85109064e-01 -1.63025841e-01 -8.84112239e-01 2.10304454e-01
2.41958022e-01 4.13910389e-01 9.64147568e-01 -1.19408071e+00
-1.63448557e-01 -5.50362051e-01 -3.79344970e-01 -1.50885665e+00
4.04460937e-01 -7.18983650e-01 -1.41984344e-01 -1.09840918e+00
-1.12170905e-01 -6.31518602e-01 2.96403497e-01 -2.53078304e-02
-5.32052964e-02 3.30013670e-02 3.14362496e-01 5.02024472e-01
-4.78825688e-01 1.01198530e+00 1.41798091e+00 -1.62911683e-01
-3.26423585e-01 3.97609830e-01 1.27445376e-02 9.94017661e-01
5.17528772e-01 -2.59199709e-01 -8.22524965e-01 -3.00652921e-01
1.93975270e-01 4.89078045e-01 3.92128021e-01 -1.06275451e+00
3.90274495e-01 -5.85824370e-01 1.67537451e-01 -8.99609447e-01
6.88733339e-01 -1.04371357e+00 6.71136618e-01 2.49940515e-01
-1.11537769e-01 3.06899011e-01 1.56816006e-01 9.33826923e-01
-2.84152359e-01 1.87640503e-01 2.15053186e-01 -1.91922620e-01
-5.30830562e-01 9.16832209e-01 -5.78200340e-01 -1.41630501e-01
1.58847642e+00 -3.14761579e-01 -4.27520752e-01 -3.64690363e-01
-9.71418440e-01 5.79942822e-01 4.44911599e-01 5.05267322e-01
4.89085436e-01 -1.67338562e+00 -5.75343251e-01 3.55174020e-02
-9.72411186e-02 1.12535052e-01 9.48101938e-01 9.90872920e-01
-2.97022820e-01 4.33136791e-01 -2.90230751e-01 -6.98741853e-01
-9.54915285e-01 8.46690357e-01 2.00724795e-01 -3.63589972e-01
-6.89083993e-01 6.37443364e-01 4.03911769e-01 -6.35803699e-01
1.20829344e-01 -4.60846663e-01 -6.55049324e-01 3.69056910e-01
3.75751674e-01 6.84478939e-01 -4.42864060e-01 -1.20010138e+00
-2.43185684e-01 1.08624089e+00 3.54102045e-01 3.11248936e-02
1.09141827e+00 -2.50184655e-01 1.14576943e-01 5.18956304e-01
1.00125754e+00 1.30500449e-02 -1.37015903e+00 9.99703482e-02
-2.15523526e-01 -4.21972573e-01 -6.94850981e-01 -2.94845790e-01
-1.09840286e+00 1.03572333e+00 1.84017956e-01 5.70562124e-01
9.93745327e-01 -2.88015038e-01 1.26229620e+00 1.51652977e-01
8.50985885e-01 -7.58679450e-01 -1.25353485e-01 5.41538537e-01
8.45686734e-01 -8.94328296e-01 -2.94230580e-01 -4.91143227e-01
-8.62536192e-01 1.03338194e+00 7.47942090e-01 -2.21756309e-01
9.25707579e-01 -6.34819418e-02 -3.92955810e-01 7.42677748e-02
-6.03493810e-01 -1.89890519e-01 3.92130315e-01 5.90273321e-01
5.31999841e-02 2.29563370e-01 -3.43893379e-01 9.41320837e-01
-4.17786896e-01 -2.18330100e-01 2.70686448e-01 2.95839906e-01
-4.14612517e-02 -1.14992678e+00 -3.71924877e-01 2.46862784e-01
-9.95981619e-02 2.85409808e-01 6.39511421e-02 7.87875712e-01
7.59640276e-01 1.00613284e+00 -2.99589746e-02 -1.00359142e+00
4.05268013e-01 -3.11132252e-01 1.90819249e-01 -4.56564754e-01
-5.53491533e-01 -6.67280480e-02 -1.11620478e-01 -5.31133473e-01
-7.29821801e-01 -8.84254932e-01 -1.28064644e+00 -8.16352844e-01
-7.69769028e-02 2.04927847e-01 1.04684167e-01 9.10796165e-01
3.74228925e-01 2.82365680e-01 5.91874301e-01 -1.20536768e+00
-1.50296316e-01 -8.35663736e-01 -5.34108996e-01 5.11130571e-01
2.59473175e-01 -1.11330462e+00 -2.65580237e-01 -2.50330213e-02] | [6.294097900390625, 0.9958465099334717] |
ebede5ac-8e3b-4c4f-bf42-206b42e3157f | me-gcn-multi-dimensional-edge-embedded-graph | null | null | https://openreview.net/forum?id=JFEPWILzYU | https://openreview.net/pdf?id=JFEPWILzYU | ME-GCN: Multi-dimensional Edge-Embedded Graph Convolutional Networks for Semi-supervised Text Classification | Compared to sequential learning models, graph-based neural networks exhibit excellent ability in capturing global information and have been used for semi-supervised learning tasks, including citation network analysis or text classification. However, most GCNs are designed with the single-dimensional edge feature and neglected to utilise the rich edge information about graphs. In this paper, we introduce the ME-GCN (Multi-dimensional Edge-enhanced Graph Convolutional Networks) for semi-supervised text classification. A text graph for an entire corpus is firstly constructed to describe the undirected and multi-dimensional relationship of word-to-word, document-document, and word-to-document. The graph is initialised with corpus-trained multi-dimensional word and document node representation, and the relations are represented according to the distance of those words/documents nodes. Then, the generated graph is trained with ME-GCN, which considers the edge features as multi-stream signals, and each stream performs a separate graph convolutional operation. Our ME-GCN can integrate a rich source of graph edge information of the entire text corpus. The results have demonstrated that our proposed model has significantly outperformed the state-of-the-art methods across eight benchmark datasets. | ['Anonymous'] | 2021-10-16 | null | null | null | acl-arr-october-2021-10 | ['semi-supervised-text-classification-1'] | ['natural-language-processing'] | [ 3.35490666e-02 -1.46491593e-02 -2.26588592e-01 -1.06797032e-01
-1.46552250e-02 -2.53819853e-01 8.22754264e-01 4.85621303e-01
-1.60984159e-01 1.64405331e-01 2.54561931e-01 -4.43816632e-01
-1.92635670e-01 -1.15641820e+00 -2.65100867e-01 -5.73043048e-01
-3.75105143e-01 3.87556046e-01 1.75210118e-01 -1.84161633e-01
3.27669114e-01 2.85657972e-01 -1.14997077e+00 2.00714190e-02
5.49518585e-01 1.02115738e+00 2.36385062e-01 9.28747058e-01
-9.45951402e-01 7.73641884e-01 -4.59146917e-01 -2.95916200e-01
-2.82917529e-01 -5.44161320e-01 -6.20276093e-01 5.34422882e-02
1.40676424e-01 -1.89901702e-02 -9.18589294e-01 1.22712159e+00
4.07189190e-01 1.90783158e-01 7.11459279e-01 -1.37543273e+00
-9.04699087e-01 8.40479374e-01 -5.37524462e-01 2.19001099e-01
3.54995489e-01 -3.40075731e-01 1.36919677e+00 -8.38595152e-01
7.82972693e-01 1.31459963e+00 4.29509491e-01 9.81785953e-02
-6.70892239e-01 -5.90543091e-01 4.26035672e-01 1.32418081e-01
-1.29603994e+00 1.01318814e-01 1.14638627e+00 -2.89892077e-01
1.21865129e+00 -8.65386426e-02 9.42348242e-01 9.55632389e-01
6.79501772e-01 7.66757786e-01 2.52351254e-01 -3.58478665e-01
-7.25375041e-02 -4.08099413e-01 6.74093962e-01 1.10952950e+00
3.81492525e-01 -4.49347556e-01 -5.71425021e-01 3.94525640e-02
7.08065987e-01 2.42811695e-01 -3.49320173e-01 -7.48825371e-02
-1.09860313e+00 9.07644629e-01 7.18198776e-01 6.70239329e-01
-3.21202189e-01 1.44266158e-01 6.87421024e-01 3.79431009e-01
8.07544947e-01 1.40618673e-02 -6.95269406e-02 -3.64724267e-03
-7.51546919e-01 -1.37119070e-01 1.02671015e+00 1.16582453e+00
6.24313474e-01 1.05520010e-01 -2.06757024e-01 5.50714552e-01
6.94887400e-01 2.15757653e-01 7.64571369e-01 1.02543913e-01
7.99540758e-01 1.30007207e+00 -8.20082426e-01 -1.74136126e+00
-7.96290874e-01 -7.01790214e-01 -1.46109998e+00 -4.64709103e-01
-1.31186053e-01 -1.65020660e-01 -8.01972747e-01 1.20473886e+00
6.31302148e-02 1.76736951e-01 6.99589495e-03 6.23284459e-01
1.58156538e+00 9.21247959e-01 -1.84876323e-01 -2.69206822e-01
1.19139206e+00 -1.11592197e+00 -1.03113866e+00 -2.37924755e-01
8.49636018e-01 -2.86897421e-01 8.10044169e-01 9.57018137e-02
-7.64988363e-01 -4.67604131e-01 -9.89639819e-01 -1.18076265e-01
-8.22066665e-01 -5.68677150e-02 7.14745343e-01 1.38709545e-01
-1.13326955e+00 4.28553730e-01 -5.08418381e-01 -3.46699804e-01
4.86563057e-01 1.85208946e-01 -3.75411391e-01 -1.01161651e-01
-1.44558394e+00 2.71782130e-01 6.59975648e-01 3.09980392e-01
-3.93016905e-01 -5.45801111e-02 -1.14944267e+00 5.16226172e-01
1.76067740e-01 -5.11509836e-01 5.17041385e-01 -5.48832774e-01
-1.23193359e+00 5.03317773e-01 -2.29860358e-02 -1.62609488e-01
1.55641697e-02 3.67119819e-01 -8.90960753e-01 2.67392397e-01
-3.77365202e-02 1.90413550e-01 7.44492114e-01 -8.98676515e-01
-3.70870113e-01 -5.37063181e-01 -3.23590517e-01 4.72198159e-01
-9.06056702e-01 -1.63119957e-01 -9.62248981e-01 -6.87501669e-01
3.12166423e-01 -5.33130109e-01 7.45342150e-02 -3.39500695e-01
-7.17913330e-01 -6.11927569e-01 1.08725190e+00 -3.75951022e-01
1.72130811e+00 -2.12213373e+00 2.55244642e-01 3.90103400e-01
8.34948242e-01 1.92327470e-01 -4.05194163e-01 7.05225348e-01
5.12949415e-02 1.35039538e-01 1.37126595e-02 -4.61986125e-01
-1.04527377e-01 7.42830411e-02 7.00033680e-02 4.59945500e-01
-6.56391829e-02 1.33382106e+00 -1.06777442e+00 -7.91343749e-01
-6.15877472e-03 5.95006585e-01 9.44906175e-02 1.38484582e-01
-2.11720526e-01 -1.87884733e-01 -7.19755113e-01 3.84027421e-01
5.50290644e-01 -7.72014618e-01 2.50917315e-01 -8.18619505e-02
3.06207091e-01 1.19881421e-01 -1.11785662e+00 1.71632063e+00
-1.84202403e-01 9.40960884e-01 -1.57929838e-01 -1.18293262e+00
1.32455122e+00 9.34753269e-02 3.39626342e-01 -5.89297175e-01
5.06539762e-01 -7.54120201e-02 -3.24551314e-02 -4.39016908e-01
5.62166810e-01 4.82766002e-01 -2.37429491e-03 6.05655372e-01
4.33789998e-01 -1.14959970e-01 3.81600559e-01 8.02146971e-01
1.05433130e+00 -3.95616502e-01 8.96848589e-02 -2.16932148e-01
7.39052296e-01 -2.70260513e-01 -1.23822831e-01 3.42758179e-01
2.22861528e-01 4.05754864e-01 8.94544959e-01 -2.65699089e-01
-4.63864207e-01 -5.35428762e-01 1.90147966e-01 9.75463271e-01
2.31665537e-01 -8.34779680e-01 -4.55186605e-01 -8.48057866e-01
5.65801412e-02 4.52002496e-01 -6.07740879e-01 -4.01589036e-01
-2.69010901e-01 -6.32335961e-01 3.20790946e-01 4.38879043e-01
6.21250391e-01 -1.07130814e+00 1.87292546e-01 3.27811748e-01
1.97428733e-01 -9.66065586e-01 -8.73150468e-01 1.66802943e-01
-7.89544284e-01 -1.24079561e+00 -7.36014903e-01 -1.37014127e+00
8.10522318e-01 4.35706615e-01 1.08789492e+00 4.04324651e-01
-2.37051606e-01 3.70833158e-01 -6.02952957e-01 -1.68538436e-01
-1.00808993e-01 3.83669168e-01 -4.37397361e-01 2.17239290e-01
4.39303160e-01 -4.16106939e-01 -5.02681017e-01 -6.40722141e-02
-9.03122604e-01 -6.49669841e-02 3.02962571e-01 7.74824440e-01
4.37503397e-01 3.77637625e-01 6.38032615e-01 -1.14479089e+00
1.46476877e+00 -6.54067814e-01 -3.93743843e-01 3.26499134e-01
-9.01733339e-01 3.37060727e-02 9.06813622e-01 -4.27718252e-01
-5.80263615e-01 -2.78659731e-01 1.80720121e-01 -2.57919401e-01
3.15365553e-01 1.21710253e+00 -8.90722126e-02 -2.08260901e-02
3.05674255e-01 3.98592860e-01 -6.71724323e-03 -1.95120618e-01
6.75755978e-01 1.08417797e+00 3.26267421e-01 -1.15869142e-01
4.29840595e-01 1.31366879e-01 3.47088635e-01 -7.94492006e-01
-4.31602001e-01 -5.96698642e-01 -6.15221918e-01 -4.06724095e-01
8.92842591e-01 -5.92468381e-01 -5.91238439e-01 8.37071240e-01
-1.22011817e+00 -2.31279209e-01 1.61794007e-01 5.50000072e-01
1.17638208e-01 5.98643303e-01 -7.91930914e-01 -5.37747741e-01
-8.05439591e-01 -7.49474227e-01 1.06358969e+00 2.81489581e-01
2.94737846e-01 -1.69996321e+00 -3.49645279e-02 -2.52616614e-01
3.91300261e-01 1.66412398e-01 1.06538677e+00 -1.09625113e+00
-4.89008352e-02 -5.50754368e-01 -5.15398085e-01 5.31214215e-02
3.01030934e-01 6.07887506e-02 -4.35840935e-01 -4.15156335e-01
-4.07309681e-01 -9.65710804e-02 9.76730764e-01 3.88275623e-01
1.19428706e+00 -2.54518569e-01 -6.62977576e-01 5.63924432e-01
1.36046863e+00 2.92165801e-02 2.18188345e-01 1.36724040e-01
1.28096247e+00 3.86112869e-01 -2.90946886e-02 2.74475098e-01
6.41383052e-01 4.53636516e-03 5.86971462e-01 -1.85616300e-01
-2.28566155e-01 -1.97353005e-01 -1.78795867e-02 1.47155011e+00
1.79053023e-01 -1.10831022e+00 -9.84635592e-01 4.60048050e-01
-2.00325584e+00 -5.14279544e-01 -7.15296149e-01 1.57767344e+00
2.72675991e-01 4.77727413e-01 -6.89668581e-02 1.36684582e-01
1.05755901e+00 6.58733487e-01 -5.52346587e-01 -2.12415054e-01
-2.34799385e-01 3.89575101e-02 3.21728826e-01 2.92889446e-01
-8.40490818e-01 9.56190526e-01 5.64081049e+00 6.14797473e-01
-1.15614307e+00 -3.22609574e-01 3.81592333e-01 2.40713432e-01
-3.70981514e-01 -3.48903537e-01 -5.27427971e-01 4.35977429e-01
8.76708388e-01 -5.64550459e-01 3.33085001e-01 5.84800243e-01
-2.65432477e-01 3.80499691e-01 -8.86789501e-01 1.08872092e+00
3.67667615e-01 -1.36489594e+00 4.32927459e-01 2.18759961e-02
6.00710332e-01 1.41516924e-01 -2.30470747e-01 2.54009128e-01
3.67988050e-01 -9.87072051e-01 1.69149563e-01 6.24356925e-01
9.36375737e-01 -8.20317090e-01 9.06096816e-01 4.12368059e-01
-1.99342132e+00 1.27678633e-01 -3.91399860e-01 -3.30610536e-02
-2.69697933e-03 7.64887869e-01 -5.59706271e-01 1.05876756e+00
3.19396466e-01 1.38070762e+00 -6.92778647e-01 5.50253332e-01
-1.15924709e-01 5.58920860e-01 -5.24999127e-02 -7.87239671e-01
5.60856462e-01 -4.25269485e-01 3.38962138e-01 1.44062769e+00
3.73001307e-01 -1.42142670e-02 1.55338928e-01 5.66496193e-01
-5.96963763e-01 4.02084053e-01 -7.12327242e-01 -6.87061965e-01
3.98936242e-01 1.56671572e+00 -1.08095300e+00 -5.31652927e-01
-6.98469162e-01 9.61648643e-01 5.38730919e-01 5.06415248e-01
-3.52227658e-01 -1.09781563e+00 -5.11939786e-02 -2.74197251e-01
2.67795295e-01 -4.11801636e-01 2.30377391e-02 -1.08411872e+00
-9.35091823e-02 -3.45805675e-01 7.57790089e-01 -8.47170770e-01
-1.47270072e+00 9.94562209e-01 -3.36350650e-01 -9.99349296e-01
-2.22024545e-01 -7.60702431e-01 -1.04873228e+00 9.69691098e-01
-1.37852514e+00 -1.17947555e+00 -6.45918727e-01 8.95140588e-01
3.54110718e-01 -5.38596034e-01 7.72066772e-01 -5.99474683e-02
-6.86949193e-01 4.93693113e-01 1.85437962e-01 6.08603895e-01
2.91263252e-01 -1.30286050e+00 8.29095721e-01 6.81586325e-01
3.78081739e-01 4.65438962e-01 1.87044824e-03 -9.51120198e-01
-1.84243774e+00 -1.29502439e+00 8.69447052e-01 1.79066345e-01
1.02954841e+00 -5.23272574e-01 -9.92358863e-01 6.87638402e-01
3.40701550e-01 4.70178336e-01 3.86120230e-01 3.67964245e-03
-2.64830172e-01 1.14213139e-01 -7.19304442e-01 5.00110805e-01
1.24807978e+00 -6.53311193e-01 -3.69759440e-01 5.37330449e-01
1.07948649e+00 -4.05693203e-01 -9.67135310e-01 -1.93095431e-01
2.11420774e-01 -3.68000776e-01 5.86023092e-01 -6.76437318e-01
5.23514032e-01 5.49690202e-02 1.60284325e-01 -1.54394007e+00
-6.39018297e-01 -4.84220356e-01 -4.75192636e-01 1.23427773e+00
3.35411489e-01 -8.85936737e-01 6.35223031e-01 -7.51560777e-02
-1.76150784e-01 -8.46686065e-01 -6.69839025e-01 -5.04422426e-01
-1.04638003e-01 -3.03378493e-01 7.31973827e-01 1.14949024e+00
4.24840808e-01 9.66187596e-01 1.08070232e-01 -1.09945737e-01
5.74725568e-01 1.92348525e-01 4.20530826e-01 -1.61616862e+00
1.07253760e-01 -9.14915264e-01 -7.15954065e-01 -1.27536941e+00
5.01796782e-01 -1.54998481e+00 -3.35630894e-01 -2.32543945e+00
6.96020797e-02 2.44771112e-02 -2.90076733e-01 2.45966971e-01
-3.51582676e-01 -2.97220379e-01 -1.91952541e-01 1.58611283e-01
-7.12759018e-01 6.08615935e-01 1.63502872e+00 -5.35238087e-01
-2.63320208e-01 -1.92996383e-01 -5.34168184e-01 2.07452089e-01
6.26883209e-01 -2.95347720e-01 -6.49711788e-01 -4.83652174e-01
5.24540484e-01 9.49706510e-02 -8.45718384e-03 -6.19136453e-01
9.03179705e-01 2.25695029e-01 3.27947706e-01 -7.26632953e-01
-1.58160105e-01 -8.04401636e-01 -2.99824178e-01 3.72802109e-01
-4.13047612e-01 9.15481597e-02 -1.06538221e-01 1.01600659e+00
-3.61403674e-01 -5.81769235e-02 9.38189328e-02 -1.58824965e-01
-4.67940360e-01 8.55738282e-01 -1.43362984e-01 7.72169605e-02
6.93226814e-01 -3.30814086e-02 -5.83243310e-01 -4.81593966e-01
-4.06489551e-01 5.37727594e-01 -1.17895693e-01 5.79646289e-01
8.42536449e-01 -1.41032612e+00 -7.09984004e-01 3.30585271e-01
2.27935106e-01 4.26074624e-01 5.15044294e-02 3.36758643e-01
-3.99647921e-01 4.78242278e-01 1.73103735e-01 -5.29975235e-01
-1.01729727e+00 7.02392995e-01 1.14149056e-01 -4.83661056e-01
-9.71913517e-01 8.64380240e-01 -1.82886526e-01 -3.32260340e-01
4.05772299e-01 -3.31290126e-01 -7.84881771e-01 3.42872292e-01
1.71014771e-01 2.82998830e-01 1.60394251e-01 -6.99804306e-01
-3.17041755e-01 6.32964432e-01 -1.18689761e-01 2.04430088e-01
1.40797472e+00 -5.89790531e-02 -4.44996178e-01 5.49720645e-01
1.68620002e+00 -3.58558923e-01 -6.68117106e-01 -6.17600918e-01
-1.01132810e-01 -3.55727179e-03 6.03102088e-01 -1.82805151e-01
-1.45778239e+00 9.94095623e-01 1.04374520e-01 8.21278572e-01
9.86369848e-01 9.68830511e-02 8.52119863e-01 5.35028815e-01
-5.73039651e-02 -9.43719506e-01 3.10169518e-01 7.37878025e-01
7.43624210e-01 -9.89108741e-01 1.66938946e-01 -3.43435794e-01
-5.03194928e-01 1.70752382e+00 4.88855720e-01 -1.63240805e-01
1.18987203e+00 1.26857027e-01 -2.67209113e-01 -9.26712275e-01
-6.04790986e-01 -1.36921301e-01 6.30954802e-01 4.35732782e-01
4.28348303e-01 9.54724289e-03 -1.99743420e-01 6.01023436e-01
-1.72258005e-01 -2.06417724e-01 2.41113335e-01 6.85527980e-01
-3.30187976e-01 -7.31060684e-01 1.47021994e-01 1.05151582e+00
1.73524953e-02 -3.82059306e-01 -6.48673415e-01 6.62286878e-01
-5.76677024e-01 9.93502140e-01 4.48191941e-01 -6.08246326e-01
1.16187677e-01 3.20993215e-02 6.96274713e-02 -6.55999899e-01
-4.92020458e-01 7.55877979e-03 1.82560533e-02 -2.35152125e-01
-2.58720636e-01 -2.21559510e-01 -1.66156626e+00 -2.30574101e-01
-6.42894566e-01 2.13190228e-01 7.69532502e-01 8.62409115e-01
4.66775358e-01 1.11117017e+00 8.00755203e-01 -6.79157019e-01
7.22063333e-02 -1.20396852e+00 -9.96022046e-01 3.21557313e-01
2.94222355e-01 -3.03161323e-01 -5.98391593e-01 -2.90772110e-01] | [9.975107192993164, 6.745555400848389] |
676b5bd6-fd7a-4076-8d4a-c97f8f53f4d6 | multi-target-domain-adaptation-via | 2103.1397 | null | https://arxiv.org/abs/2103.13970v1 | https://arxiv.org/pdf/2103.13970v1.pdf | Multi-Target Domain Adaptation via Unsupervised Domain Classification for Weather Invariant Object Detection | Object detection is an essential technique for autonomous driving. The performance of an object detector significantly degrades if the weather of the training images is different from that of test images. Domain adaptation can be used to address the domain shift problem so as to improve the robustness of an object detector. However, most existing domain adaptation methods either handle single target domain or require domain labels. We propose a novel unsupervised domain classification method which can be used to generalize single-target domain adaptation methods to multi-target domains, and design a weather-invariant object detector training framework based on it. We conduct the experiments on Cityscapes dataset and its synthetic variants, i.e. foggy, rainy, and night. The experimental results show that the object detector trained by our proposed method realizes robust object detection under different weather conditions. | ['Francis Ng', 'Jinlin Chen', 'Ting Sun'] | 2021-03-25 | null | null | null | null | ['robust-object-detection', 'multi-target-domain-adaptation'] | ['computer-vision', 'computer-vision'] | [ 1.30670100e-01 -6.39135003e-01 -8.95310342e-02 -6.16581976e-01
-1.15483634e-01 -6.13177538e-01 5.25134683e-01 -2.84812927e-01
-4.80445683e-01 6.45151019e-01 -2.40824163e-01 -1.79559842e-01
1.71304360e-01 -8.65828454e-01 -4.19796646e-01 -9.48288441e-01
4.31618005e-01 3.23633850e-01 1.05184937e+00 -4.22944605e-01
2.32095644e-01 4.04875040e-01 -1.57391465e+00 -5.45534380e-02
1.12763524e+00 7.23183632e-01 5.87274611e-01 5.20732582e-01
-2.36738667e-01 2.51961678e-01 -8.77715111e-01 3.47244740e-01
4.67922032e-01 -2.91121602e-01 -1.81823790e-01 3.57606262e-01
1.59008563e-01 -4.02961552e-01 -3.84889722e-01 1.29360867e+00
4.75162417e-01 2.55877882e-01 8.79669368e-01 -1.40967453e+00
-6.39512122e-01 -1.68235913e-01 -4.98250604e-01 6.65691733e-01
-2.07841843e-01 1.22335494e-01 1.44432276e-01 -7.33994007e-01
3.25819075e-01 1.16546416e+00 2.24950016e-01 7.41000295e-01
-8.33350420e-01 -1.20530295e+00 4.44206059e-01 6.56238496e-01
-1.41288292e+00 -2.86570072e-01 8.94287586e-01 -5.21371126e-01
2.88398832e-01 -2.45416969e-01 2.50488758e-01 8.80376220e-01
2.04341888e-01 5.34121633e-01 1.42221534e+00 -1.87700883e-01
4.18258160e-01 5.43187380e-01 2.30689228e-01 2.16885477e-01
6.87820613e-01 3.81561607e-01 -1.63626507e-01 1.86989084e-01
3.64708424e-01 7.73864239e-02 -1.92752436e-01 -4.90214199e-01
-8.29096138e-01 7.31571734e-01 4.46243972e-01 9.93319452e-02
1.61322206e-02 -5.38384676e-01 3.29301417e-01 4.64259773e-01
4.43997175e-01 2.53877640e-02 -3.85241061e-01 3.90888870e-01
-4.86492485e-01 3.18445355e-01 5.11086822e-01 1.31214821e+00
9.82141793e-01 3.89029205e-01 -1.16523448e-02 8.39152992e-01
3.24943274e-01 1.16701782e+00 6.41496062e-01 -3.75957280e-01
3.05817962e-01 6.54402554e-01 4.67265338e-01 -6.86274469e-01
-5.16178131e-01 -2.93817520e-01 -3.23311657e-01 5.81019342e-01
2.56725788e-01 -2.24131092e-01 -1.27557158e+00 1.37017965e+00
5.81070185e-01 3.21923107e-01 6.27536654e-01 1.13787436e+00
1.15904570e+00 7.84108818e-01 1.29378185e-01 -8.54097456e-02
1.14091849e+00 -7.70069718e-01 -6.82612360e-01 -6.80016220e-01
3.79847765e-01 -6.84098423e-01 9.94750619e-01 1.66318804e-01
-1.90006286e-01 -9.13447142e-01 -1.40048683e+00 4.33086395e-01
-5.07770181e-01 1.10134400e-01 1.05368011e-02 8.20136726e-01
-2.31528655e-01 -1.80645138e-01 -4.88667637e-01 -6.43975079e-01
2.61415333e-01 1.42117172e-01 -2.70853698e-01 -3.90447855e-01
-1.20221126e+00 1.06367636e+00 1.01423919e+00 -1.77671224e-01
-1.17325938e+00 -1.42900556e-01 -7.93737650e-01 -2.87528515e-01
2.14917272e-01 -7.09825605e-02 1.20866525e+00 -1.04025733e+00
-1.34320307e+00 7.81397641e-01 -2.55662650e-01 -4.23062176e-01
1.78698942e-01 -4.91842069e-02 -1.24511850e+00 -2.92728171e-02
2.71273762e-01 2.74015933e-01 1.11080790e+00 -1.25853229e+00
-1.21232748e+00 -3.25374097e-01 -2.87991047e-01 3.03608626e-01
-3.37037891e-01 -6.23265244e-02 -7.65297115e-02 -3.47343177e-01
1.12718046e-01 -9.52541828e-01 1.40159642e-02 -1.24358326e-01
2.23136536e-04 -2.09929436e-01 1.64422834e+00 -3.46028626e-01
7.12467670e-01 -2.50200415e+00 -4.14909720e-01 1.60898864e-02
-2.91823924e-01 6.25003099e-01 -9.03410390e-02 -2.87394356e-02
-9.28356778e-03 -3.79980147e-01 -3.27822745e-01 2.56799549e-01
-3.71261388e-01 5.79823971e-01 -3.27187598e-01 5.72822869e-01
3.05888653e-01 2.86892205e-01 -7.61913478e-01 -5.63436091e-01
1.91991389e-01 -4.03565019e-02 -5.00292964e-02 3.17828566e-01
-1.60756275e-01 7.52808690e-01 -6.02390110e-01 5.40232420e-01
1.23530555e+00 2.44259164e-01 -1.49513781e-01 2.24614963e-01
-2.50446737e-01 3.23728248e-02 -1.32481742e+00 1.09062338e+00
-2.27140129e-01 6.50088131e-01 -5.24686612e-02 -1.08579445e+00
1.54162800e+00 1.03834964e-01 -1.26443982e-01 -9.44223344e-01
1.00462765e-01 3.10806125e-01 2.48517483e-01 -6.27387702e-01
3.47852319e-01 -3.48332286e-01 -1.92503050e-01 -1.01874046e-01
-2.02615455e-01 -2.28466108e-01 1.47818565e-01 -2.48720348e-01
7.39857197e-01 2.99689034e-03 2.82682091e-01 -3.47229570e-01
8.35331619e-01 3.80075127e-01 1.17651725e+00 6.19852662e-01
-7.04574943e-01 4.15371954e-01 -8.50139335e-02 -4.73675936e-01
-1.07549858e+00 -1.20867395e+00 -4.10005927e-01 1.19197917e+00
8.92240465e-01 4.73525822e-01 -3.34526956e-01 -7.89324701e-01
1.28526881e-01 7.72736788e-01 -3.06537777e-01 -6.74855769e-01
-5.98104179e-01 -8.84430408e-01 3.93988073e-01 5.89493573e-01
1.16630602e+00 -9.59898055e-01 -5.54487944e-01 1.79015398e-01
-1.44178554e-01 -1.26984656e+00 -2.23536655e-01 1.18946493e-01
-6.84002101e-01 -8.93839836e-01 -6.28088593e-01 -1.25371993e+00
5.46340227e-01 6.94878578e-01 6.45694911e-01 -3.25035363e-01
1.28019378e-01 2.25923285e-02 -5.25874913e-01 -1.05771542e+00
-5.64212143e-01 -9.63135660e-02 3.58519763e-01 2.86293834e-01
7.84053326e-01 -3.80843878e-01 -5.70835531e-01 8.65968823e-01
-1.04972708e+00 -4.74373847e-01 6.63815498e-01 7.60416865e-01
4.12665755e-01 4.31800097e-01 8.70555043e-01 -6.72130108e-01
3.10432494e-01 -5.34359336e-01 -9.84183609e-01 7.08638132e-02
-6.48418248e-01 3.44746746e-02 5.81265628e-01 -6.34408057e-01
-1.50309849e+00 2.84317911e-01 3.38281870e-01 -1.76608950e-01
-5.90915501e-01 5.86336777e-02 -4.89519358e-01 -1.71240941e-01
1.08537066e+00 4.29682255e-01 -1.16058744e-01 -4.20514643e-01
1.43704921e-01 9.70712602e-01 8.01343203e-01 -2.75008291e-01
1.28528714e+00 6.86210454e-01 -8.43877420e-02 -9.29851234e-01
-4.97415185e-01 -9.92359817e-01 -5.30040741e-01 -1.53052613e-01
8.32972467e-01 -1.18855941e+00 -3.43432128e-02 5.35661578e-01
-9.47769701e-01 -2.11235285e-01 4.28789258e-02 6.04716301e-01
-2.27580771e-01 2.99292624e-01 1.39436379e-01 -7.24715531e-01
-1.65435504e-02 -8.78902555e-01 7.52352059e-01 6.84112549e-01
4.46639121e-01 -7.85662293e-01 2.05636933e-01 -5.75828366e-02
2.15483412e-01 7.95469955e-02 6.72899544e-01 -7.71277428e-01
-2.73925275e-01 -2.36069813e-01 -3.70862305e-01 4.60442811e-01
3.16764146e-01 -1.76205918e-01 -1.05819237e+00 -3.25398028e-01
1.64635956e-01 7.00916946e-02 9.69132125e-01 6.48304746e-02
4.87378836e-01 -2.37262417e-02 -6.60932183e-01 4.61145550e-01
1.42083645e+00 7.30992317e-01 6.12325549e-01 7.28453875e-01
3.68436575e-01 3.94857258e-01 1.24020958e+00 2.25191742e-01
2.29342803e-01 5.54515541e-01 4.21509117e-01 -9.23296958e-02
-3.05283703e-02 -2.67825965e-02 4.35422838e-01 9.04622823e-02
3.66038740e-01 -1.64536104e-01 -1.00536203e+00 8.78818810e-01
-1.88677084e+00 -7.73291767e-01 -2.12132007e-01 2.18206954e+00
4.46103156e-01 4.97221053e-01 1.63434833e-01 6.61446750e-02
1.00031579e+00 -8.98666978e-02 -8.37359011e-01 -2.89628416e-01
-2.14737684e-01 -6.85672387e-02 8.11353981e-01 1.41209051e-01
-1.40285850e+00 1.00971639e+00 6.19112921e+00 4.97585654e-01
-1.25555217e+00 1.92545444e-01 -4.13247012e-02 1.51379213e-01
1.94567770e-01 -3.92598398e-02 -1.02338111e+00 5.58156431e-01
7.97266066e-01 -3.38241667e-01 -1.17995173e-01 1.32617378e+00
3.14713597e-01 -1.72259003e-01 -5.91846764e-01 8.72388422e-01
-3.49896140e-02 -5.82033873e-01 -2.02328026e-01 -2.55167365e-01
7.82277107e-01 6.85714185e-02 6.04434721e-02 6.15148485e-01
3.13243955e-01 -4.80843991e-01 4.55608100e-01 1.65886268e-01
6.07956946e-01 -6.13133550e-01 7.23719299e-01 6.24078810e-01
-1.15878928e+00 -4.27519202e-01 -8.45855176e-01 -2.07913846e-01
-9.57443640e-02 4.40266401e-01 -1.20037520e+00 2.92652369e-01
9.58925664e-01 6.61674798e-01 -5.62593043e-01 1.35487533e+00
-2.91203707e-01 7.69263268e-01 -3.01974177e-01 3.57290241e-03
1.04183465e-01 -5.18576689e-02 8.41505468e-01 1.12582862e+00
2.63396531e-01 1.43257543e-01 4.95780587e-01 5.40835321e-01
3.42310488e-01 2.16973536e-02 -1.00138891e+00 3.99683177e-01
5.65755308e-01 1.10066414e+00 -5.80358803e-01 -5.48616707e-01
-3.99654865e-01 8.13727975e-01 -1.48037016e-01 5.01460314e-01
-9.92010891e-01 -6.94622219e-01 8.12584281e-01 9.95400324e-02
5.97041786e-01 -4.10435826e-01 -3.26560140e-01 -1.01155794e+00
3.74102667e-02 -5.45699298e-01 4.69669342e-01 -8.04107547e-01
-1.12134302e+00 3.88204992e-01 2.43248314e-01 -1.69541311e+00
5.96543914e-03 -7.29531527e-01 -7.93670535e-01 8.41374636e-01
-1.97338676e+00 -8.22665155e-01 -8.88934970e-01 7.47987926e-01
5.20542502e-01 -4.56930518e-01 4.87232864e-01 3.42872769e-01
-5.09714246e-01 5.19678950e-01 4.90789413e-01 2.08538592e-01
1.13334036e+00 -9.85334337e-01 2.07493991e-01 1.19272387e+00
-2.63189882e-01 7.61366338e-02 9.23737884e-01 -6.56307936e-01
-7.47744024e-01 -1.59693444e+00 3.59569401e-01 -1.85785174e-01
2.37663403e-01 -2.52317190e-01 -1.08925140e+00 4.74047303e-01
1.43596239e-03 1.39834166e-01 2.21320972e-01 -3.07037055e-01
-5.29325187e-01 -7.11446762e-01 -1.35291600e+00 3.30217123e-01
7.04015493e-01 -9.53125060e-02 -8.77184033e-01 4.29023832e-01
8.17668080e-01 -5.08785427e-01 -3.64281923e-01 6.50801182e-01
2.00063661e-01 -9.07768250e-01 7.31968701e-01 -4.46471900e-01
-2.20744163e-01 -9.20000970e-01 -8.09169486e-02 -1.39147282e+00
-5.71027875e-01 -4.26566079e-02 3.50864351e-01 1.12622046e+00
1.93744570e-01 -1.06499243e+00 7.25526571e-01 6.19773529e-02
-3.76230091e-01 3.60657990e-01 -8.35745096e-01 -1.35542703e+00
9.72618088e-02 -1.09948240e-01 6.44297838e-01 6.50634587e-01
-3.46237659e-01 3.19859385e-01 -9.87733081e-02 7.04159617e-01
5.00452399e-01 2.15391740e-01 1.05661118e+00 -1.42873943e+00
6.85636327e-02 -1.95871037e-03 -7.30590999e-01 -9.39148307e-01
8.83456990e-02 -5.30463815e-01 6.86601102e-01 -1.38061821e+00
-1.50460377e-01 -4.78236467e-01 -4.03128326e-01 3.35404873e-01
-2.72384584e-01 2.55077213e-01 -1.15075022e-01 4.04110193e-01
-3.87169480e-01 4.62269753e-01 1.12097526e+00 -3.57739896e-01
-4.36827600e-01 2.32439488e-01 -4.27600294e-01 6.99786782e-01
1.16241288e+00 -6.67675138e-01 -5.05631685e-01 -3.80827308e-01
-4.22128826e-01 -4.64352876e-01 3.71958792e-01 -1.50726223e+00
3.02165337e-02 -5.37785530e-01 4.15159285e-01 -5.52423477e-01
3.05801537e-02 -9.94182110e-01 -3.35792512e-01 5.44375241e-01
2.48225838e-01 -2.38217190e-01 5.19093454e-01 8.33069980e-01
-3.07120919e-01 -2.27409825e-01 1.29209113e+00 2.05701664e-02
-1.58899581e+00 1.20419957e-01 -4.06627595e-01 5.87387942e-02
1.30605721e+00 -3.63781333e-01 -5.88715136e-01 -1.72489688e-01
-4.64485526e-01 3.53962362e-01 4.38735694e-01 6.99960649e-01
6.59473956e-01 -1.44817424e+00 -8.89859259e-01 3.72949481e-01
7.80334234e-01 9.79756117e-02 -5.64408729e-05 4.47461933e-01
-3.73615414e-01 2.67606735e-01 -4.87190932e-01 -8.82185519e-01
-1.34840643e+00 8.50421250e-01 4.37188774e-01 3.29668939e-01
-5.53081214e-01 4.39743519e-01 6.45721793e-01 -3.32168847e-01
-9.36002582e-02 -2.95170426e-01 -4.01900977e-01 -3.63021761e-01
5.79684913e-01 2.25405902e-01 1.07903875e-01 -6.93955123e-01
-4.64245260e-01 6.64111078e-01 -6.49426952e-02 1.03087835e-02
7.95480490e-01 -1.89319476e-01 3.36392999e-01 4.02203530e-01
8.46706688e-01 -3.19132924e-01 -1.39154887e+00 -6.37027442e-01
-9.59796458e-02 -4.12909865e-01 -2.17776969e-01 -5.96812904e-01
-6.75710380e-01 9.02297020e-01 1.26653659e+00 5.93112148e-02
1.32668233e+00 -2.30619252e-01 6.72048211e-01 5.64832449e-01
3.76843750e-01 -1.33265257e+00 7.69566372e-02 6.38075829e-01
5.88527501e-01 -1.57282782e+00 -1.17626652e-01 -2.75158882e-01
-7.50701725e-01 8.76596868e-01 9.76577222e-01 -3.50888371e-01
6.70579553e-01 -6.72638491e-02 3.45570028e-01 8.31215084e-02
-3.10639888e-01 -6.37951434e-01 2.20823839e-01 1.12833881e+00
-2.33956337e-01 1.58032164e-01 -2.84581840e-01 2.77712196e-01
1.29672289e-01 -4.96339910e-02 4.63644087e-01 1.04662824e+00
-1.23697960e+00 -9.02842283e-01 -8.83144379e-01 1.15997858e-01
8.55512097e-02 3.29867005e-01 -2.64945179e-01 8.40718508e-01
5.80315590e-01 1.26950598e+00 1.38891250e-01 -5.19239426e-01
6.74683034e-01 1.97468102e-01 7.93062150e-02 -7.11179614e-01
1.72132984e-01 -1.34889632e-01 -1.06547609e-01 3.62758301e-02
-4.86084640e-01 -6.14893258e-01 -1.42618489e+00 1.27022997e-01
-3.83943826e-01 1.00240909e-01 6.04297340e-01 9.48856056e-01
1.48019239e-01 3.46372187e-01 7.90803671e-01 -4.15063351e-01
-3.46957684e-01 -1.02236998e+00 -7.12276280e-01 5.59029162e-01
4.86203820e-01 -1.07654130e+00 -2.75976270e-01 2.06279382e-01] | [9.870905876159668, 2.112945318222046] |
2144ddca-2bb1-45ef-a6cd-114c38325501 | tatum-level-drum-transcription-based-on-a | 2010.03749 | null | https://arxiv.org/abs/2010.03749v1 | https://arxiv.org/pdf/2010.03749v1.pdf | Tatum-Level Drum Transcription Based on a Convolutional Recurrent Neural Network with Language Model-Based Regularized Training | This paper describes a neural drum transcription method that detects from music signals the onset times of drums at the $\textit{tatum}$ level, where tatum times are assumed to be estimated in advance. In conventional studies on drum transcription, deep neural networks (DNNs) have often been used to take a music spectrogram as input and estimate the onset times of drums at the $\textit{frame}$ level. The major problem with such frame-to-frame DNNs, however, is that the estimated onset times do not often conform with the typical tatum-level patterns appearing in symbolic drum scores because the long-term musically meaningful structures of those patterns are difficult to learn at the frame level. To solve this problem, we propose a regularized training method for a frame-to-tatum DNN. In the proposed method, a tatum-level probabilistic language model (gated recurrent unit (GRU) network or repetition-aware bi-gram model) is trained from an extensive collection of drum scores. Given that the musical naturalness of tatum-level onset times can be evaluated by the language model, the frame-to-tatum DNN is trained with a regularizer based on the pretrained language model. The experimental results demonstrate the effectiveness of the proposed regularized training method. | ['Kazuyoshi Yoshii', 'Eita Nakamura', 'Ryo Nishikimi', 'Ryoto Ishizuka'] | 2020-10-08 | null | null | null | null | ['drum-transcription'] | ['music'] | [ 3.69023502e-01 -4.71247286e-01 4.53889742e-03 -1.07412621e-01
-9.85069215e-01 -5.16823590e-01 -5.85368499e-02 -2.28440925e-01
-3.64780456e-01 4.10696387e-01 2.14306921e-01 -2.61170473e-02
-1.71470329e-01 -5.39269388e-01 -6.52827382e-01 -7.81484663e-01
-4.24963199e-02 2.24341094e-01 -6.83398247e-02 -1.67566657e-01
2.54301280e-01 6.58351034e-02 -1.56335783e+00 5.89439213e-01
4.76363897e-01 9.57026780e-01 4.72409993e-01 1.00586545e+00
-1.42861500e-01 8.40442002e-01 -8.88600111e-01 -2.28351410e-02
2.88363129e-01 -9.85562027e-01 -2.42163017e-01 -1.59165785e-01
7.30867088e-02 -2.79721081e-01 -2.49836519e-01 1.15573883e+00
6.32265449e-01 3.22065532e-01 7.10931659e-01 -4.48493302e-01
-2.32814342e-01 1.27537298e+00 -5.44664323e-01 1.99843451e-01
1.57601640e-01 -1.39474317e-01 1.34536004e+00 -1.02872181e+00
3.09498250e-01 1.01159978e+00 8.39758992e-01 1.98320135e-01
-1.02224493e+00 -7.16616213e-01 -1.36720210e-01 1.58655465e-01
-1.68165731e+00 -2.75576472e-01 1.01925981e+00 -5.05570412e-01
7.93775141e-01 2.37412229e-01 5.93562782e-01 8.78248930e-01
4.83701751e-02 8.47427309e-01 6.39988422e-01 -8.33160400e-01
1.64644923e-02 -6.81292951e-01 3.75708491e-02 2.59944528e-01
-4.59309667e-01 -2.24668104e-02 -7.64393806e-01 6.89271465e-02
1.30971467e+00 -1.19126052e-01 -2.88325429e-01 4.83758330e-01
-1.13360214e+00 6.89553559e-01 1.42531916e-01 6.63645506e-01
-6.00149214e-01 4.10825342e-01 8.22166204e-01 1.30943805e-01
1.56450137e-01 3.88624907e-01 -2.62726843e-01 -4.83158439e-01
-1.42011106e+00 2.10243389e-01 4.84995067e-01 6.27156138e-01
4.76095229e-01 7.05144167e-01 -4.06215400e-01 9.88995254e-01
7.18680471e-02 2.32381240e-01 8.75372231e-01 -8.99601996e-01
3.15846473e-01 1.60773784e-01 4.48865891e-02 -8.97268414e-01
3.00683372e-04 -6.41173542e-01 -1.12132692e+00 -7.71508664e-02
6.07199073e-01 -1.78598449e-01 -7.17339694e-01 1.89637554e+00
-1.83053717e-01 4.78836566e-01 -1.07929930e-02 1.05733573e+00
4.18380499e-01 1.23998415e+00 -2.21033111e-01 -5.77018738e-01
1.36620438e+00 -5.68641007e-01 -1.02449083e+00 8.73037800e-02
3.69356185e-01 -1.18725097e+00 1.22877657e+00 7.34624386e-01
-9.98969793e-01 -1.06191182e+00 -9.70615923e-01 -3.20315883e-02
3.31767112e-01 6.75870478e-01 2.08991408e-01 2.73463368e-01
-5.05796432e-01 8.33484709e-01 -7.39809811e-01 2.80171722e-01
-3.30632389e-01 7.30271786e-02 3.58553946e-01 4.64896470e-01
-1.40159345e+00 3.51109833e-01 8.18593442e-01 4.87198949e-01
-8.74872029e-01 -3.01038563e-01 -5.19031584e-01 1.29423797e-01
1.73154891e-01 -2.06848562e-01 1.60895228e+00 -1.18557727e+00
-1.72318721e+00 6.29373491e-01 -3.48076552e-01 -6.44122958e-01
1.24653965e-01 -3.59863639e-01 -4.00369585e-01 1.08731762e-01
-1.43085018e-01 9.20289308e-02 1.12605548e+00 -7.23461807e-01
-5.00060916e-01 7.68970400e-02 -3.78230929e-01 2.54307002e-01
-9.24153104e-02 2.85322070e-01 -3.62727970e-01 -1.39751327e+00
3.09664547e-01 -8.77398431e-01 2.62029767e-02 -6.55172586e-01
-5.72035909e-01 -4.19467509e-01 3.51510286e-01 -9.29872990e-01
1.90326023e+00 -2.31871390e+00 2.46734202e-01 2.10338250e-01
-2.29699537e-01 2.22581938e-01 1.88224595e-02 4.36800152e-01
-2.65555918e-01 -3.84375572e-01 -2.50790507e-01 -2.38864049e-01
4.21315506e-02 1.93025485e-01 -6.83135509e-01 1.34963572e-01
-1.55650422e-01 4.82727885e-01 -9.10292506e-01 -2.99352884e-01
5.95039129e-02 6.02425039e-01 -2.67596751e-01 4.50090259e-01
-3.79111052e-01 5.88248849e-01 -4.27314341e-02 2.91068941e-01
1.03277281e-01 3.60033602e-01 2.42192477e-01 -3.79128486e-01
-3.56704533e-01 7.63581395e-01 -1.29490054e+00 1.84578180e+00
-3.00998241e-01 5.92584729e-01 -2.88319528e-01 -9.06125247e-01
1.16208863e+00 8.15147340e-01 5.23780882e-01 -2.03121290e-01
2.79768467e-01 4.07283664e-01 5.09876966e-01 -2.83746511e-01
6.66990876e-01 -5.73917150e-01 -1.50809661e-01 5.42006910e-01
-7.45681953e-03 -2.06511661e-01 2.50715971e-01 -4.62717026e-01
8.82567704e-01 3.58922094e-01 3.74842405e-01 1.41400322e-01
2.03724310e-01 -3.72040689e-01 8.32992017e-01 5.86461186e-01
2.48705581e-01 1.00630701e+00 4.28998739e-01 -5.13060153e-01
-1.16411746e+00 -9.58439648e-01 5.31704798e-02 1.46540546e+00
-4.56467927e-01 -7.46850133e-01 -9.17812824e-01 9.15121287e-02
-5.66419303e-01 6.89976871e-01 -2.26375878e-01 -1.31510111e-04
-9.98266518e-01 -5.96511900e-01 1.04081416e+00 5.28216660e-01
8.93732905e-02 -1.45086896e+00 -5.53311765e-01 7.56531298e-01
-3.69739085e-01 -8.19503486e-01 -8.37215662e-01 4.36711639e-01
-8.20833743e-01 -5.86304367e-01 -9.46801066e-01 -8.96827579e-01
1.60400406e-01 -1.61750346e-01 8.50353897e-01 -2.26492092e-01
3.44990306e-02 -4.31764871e-01 -5.23165405e-01 -4.18659985e-01
-3.67515028e-01 2.50326674e-02 4.21448708e-01 2.68870980e-01
3.21961045e-01 -8.99138272e-01 -4.76710737e-01 2.46107757e-01
-1.03497040e+00 1.00779586e-01 6.22237980e-01 7.40453541e-01
1.08366060e+00 3.84498835e-01 7.13992119e-01 -5.93516767e-01
8.38040709e-01 8.56716633e-02 -4.18942094e-01 -3.67662162e-02
-4.38950099e-02 1.22929318e-02 8.85239244e-01 -9.20288563e-01
-5.68369985e-01 1.78856134e-01 -4.05075878e-01 -8.77135992e-01
1.31411746e-01 8.13618660e-01 -2.12623198e-02 6.43310070e-01
7.78205395e-01 3.95908445e-01 -4.17246640e-01 -6.06728792e-01
2.27940068e-01 6.79007888e-01 1.00514877e+00 -8.17009330e-01
6.44211113e-01 -1.95664987e-01 -3.36789429e-01 -7.62799382e-01
-1.02885532e+00 -4.00702566e-01 -6.88159108e-01 -3.18235397e-01
7.51246512e-01 -9.75723863e-01 -6.38649702e-01 4.33934599e-01
-1.42786980e+00 -1.83312267e-01 -4.62129354e-01 7.71515489e-01
-7.53555775e-01 4.32953894e-01 -1.10409689e+00 -1.23953116e+00
-5.04678547e-01 -9.87851441e-01 8.45134556e-01 3.87438163e-02
-6.91317856e-01 -5.01254141e-01 1.90144852e-01 -1.70112178e-01
4.42741178e-02 2.99344957e-01 1.01521742e+00 -3.37361366e-01
-3.45148742e-01 -3.63482714e-01 2.81717151e-01 6.19700253e-01
3.10762465e-01 1.53327852e-01 -1.11883593e+00 1.60773774e-03
3.19788754e-01 -1.26355290e-01 7.09627032e-01 8.23083639e-01
1.23724365e+00 -3.18246245e-01 4.13188517e-01 4.81964350e-01
9.95303333e-01 4.41599041e-01 6.13637507e-01 -4.22490984e-02
7.19352424e-01 1.58581167e-01 5.02057970e-01 6.44219995e-01
-1.51844621e-01 7.25222290e-01 4.02664058e-02 4.27171230e-01
-1.68094248e-01 -5.00240684e-01 7.05821097e-01 1.83902693e+00
-5.68810761e-01 -2.74377018e-02 -6.37356162e-01 5.61813772e-01
-1.93932462e+00 -1.06158161e+00 -1.62085459e-01 2.44258094e+00
1.24213767e+00 4.46166426e-01 2.12670997e-01 6.47549331e-01
1.05029428e+00 1.54804602e-01 -5.06075144e-01 -5.27777612e-01
-1.93659782e-01 5.60687423e-01 -5.14809340e-02 2.67126054e-01
-8.31189156e-01 8.66791546e-01 5.55745888e+00 1.39766002e+00
-1.28408444e+00 9.57771577e-03 2.77704358e-01 -2.39341199e-01
-8.77240673e-02 -3.33973467e-02 -6.86189592e-01 4.31457758e-01
1.09679186e+00 1.30217910e-01 6.55390680e-01 6.92090809e-01
5.98472536e-01 2.86788553e-01 -1.26506817e+00 1.28361940e+00
-1.89967394e-01 -1.00709522e+00 1.54041196e-03 -2.76531368e-01
6.60023332e-01 -3.60567272e-01 1.29571512e-01 4.21123415e-01
-9.54115540e-02 -1.09313989e+00 1.05406666e+00 7.28588402e-01
1.10261595e+00 -1.00016880e+00 4.76356715e-01 5.09029627e-01
-1.69106734e+00 1.78747341e-01 -5.07280707e-01 -3.34818840e-01
2.90619701e-01 8.81047070e-01 -1.13419628e+00 4.96578336e-01
3.25848371e-01 6.11637354e-01 1.14592902e-01 9.95194674e-01
-4.76377487e-01 1.20509744e+00 -4.78143424e-01 7.79770985e-02
1.85503736e-01 -2.69732714e-01 6.07828796e-01 1.33094501e+00
9.11645353e-01 -2.05414128e-02 -1.66627485e-02 8.39912355e-01
-1.57762751e-01 1.65015399e-01 -4.05505121e-01 -4.01406974e-01
4.98595238e-01 9.02246356e-01 -7.64794409e-01 -1.15238912e-01
1.64892435e-01 9.15418029e-01 4.29808721e-02 3.62081677e-01
-7.06429005e-01 -4.95713234e-01 2.91502327e-01 2.20477413e-02
3.35549444e-01 -5.14610231e-01 -2.93698639e-01 -9.43755090e-01
-4.40114066e-02 -8.79968882e-01 1.77006364e-01 -8.35554123e-01
-1.24567235e+00 8.63869071e-01 -3.34346950e-01 -1.71641123e+00
-8.77427459e-01 -3.54025662e-01 -7.17586935e-01 1.06781232e+00
-8.60537708e-01 -7.36089826e-01 5.56853354e-01 4.72492665e-01
8.19499433e-01 -1.29919648e-01 1.05858815e+00 2.31567919e-01
-3.92531902e-01 4.70667571e-01 1.03204109e-01 4.84899491e-01
6.44630432e-01 -1.20919359e+00 1.80366740e-01 8.25639129e-01
7.61964023e-01 8.68158281e-01 8.13259482e-01 -5.43569505e-01
-1.04937112e+00 -9.41774249e-01 7.80156374e-01 -2.51972646e-01
7.22445786e-01 -3.77397180e-01 -1.01049256e+00 5.71547151e-01
2.27376819e-01 -2.80737281e-01 7.38120496e-01 1.43424988e-01
-1.28289685e-01 -2.24054709e-01 -2.16620609e-01 5.32545745e-01
5.96002042e-01 -1.05926120e+00 -1.01958513e+00 1.17681094e-01
6.35117233e-01 -5.45953155e-01 -6.85749233e-01 4.60692316e-01
7.52939284e-01 -9.14843559e-01 7.78446734e-01 -2.79176414e-01
4.24614936e-01 -6.97206795e-01 -2.15361714e-01 -1.21742606e+00
-2.00589120e-01 -1.06965387e+00 -3.41137379e-01 1.19018269e+00
1.69121489e-01 2.74333030e-01 4.83778179e-01 -9.17567685e-02
-3.62717152e-01 -3.89149189e-01 -1.14944375e+00 -6.44985855e-01
-2.99006075e-01 -9.32697773e-01 1.96174935e-01 8.09667349e-01
-1.58874869e-01 5.48821390e-01 -8.03733289e-01 1.68970913e-01
3.59458357e-01 3.01176965e-01 4.82449979e-01 -1.09299147e+00
-7.08717525e-01 -4.37944144e-01 -1.58123195e-01 -1.22531879e+00
-4.85446081e-02 -7.59945154e-01 4.87299442e-01 -1.01367998e+00
-2.69539028e-01 -1.25740826e-01 -8.33397746e-01 2.62613595e-01
-1.97820500e-01 1.74797252e-01 2.69586831e-01 3.53531152e-01
-2.15432897e-01 5.05805612e-01 1.07053375e+00 -1.43750757e-02
-5.36526442e-01 4.44882900e-01 -1.04792990e-01 1.13760698e+00
6.13594472e-01 -5.71809590e-01 -3.60297263e-01 -3.79618824e-01
3.86285335e-01 5.96589148e-01 -5.92885688e-02 -1.25398207e+00
2.87673682e-01 1.51549920e-01 1.80815861e-01 -1.09018016e+00
6.76755905e-01 -4.68030483e-01 2.46648461e-01 2.71605670e-01
-5.12791157e-01 3.08155298e-01 1.17864311e-01 3.79562587e-01
-5.48213959e-01 -6.03286445e-01 5.01848578e-01 -1.28608227e-01
-3.60079706e-01 -8.49572476e-03 -6.66761220e-01 -1.55250862e-01
3.24922562e-01 -1.41635969e-01 5.31117439e-01 -4.25407559e-01
-7.61996448e-01 -6.00436151e-01 -1.67179897e-01 9.60853025e-02
5.68866432e-01 -1.62982845e+00 -6.76773369e-01 1.53443620e-01
-7.17127994e-02 2.13636711e-01 3.58321637e-01 4.98714119e-01
-4.39986050e-01 1.23978332e-01 9.03818160e-02 -5.74442148e-01
-8.33014429e-01 3.04428488e-01 2.17082784e-01 -3.24214041e-01
-6.30236924e-01 1.03748274e+00 2.21811429e-01 -1.31441385e-01
5.70660830e-01 -9.62957799e-01 -1.32783741e-01 1.28806204e-01
5.81332803e-01 -1.50586376e-02 -7.57453814e-02 -6.42268479e-01
1.43810898e-01 6.61529779e-01 2.46344939e-01 -4.24929798e-01
1.29273248e+00 1.34367585e-01 -2.17604071e-01 1.35924768e+00
7.19933331e-01 3.15066874e-01 -1.10468733e+00 -2.38366187e-01
1.20843515e-01 1.16491020e-02 -6.37089908e-02 -4.47145820e-01
-7.28243470e-01 1.14240348e+00 3.74849528e-01 2.95637667e-01
1.18376434e+00 -5.78817785e-01 1.11933041e+00 4.31236953e-01
1.77006200e-01 -1.30610037e+00 5.94764836e-02 8.31115782e-01
9.04769242e-01 -4.49733943e-01 -4.33958560e-01 6.87088668e-02
-4.82199967e-01 1.50160098e+00 2.51670659e-01 -4.34577465e-01
4.85040456e-01 3.38868350e-01 8.94771740e-02 1.83672279e-01
-3.99253249e-01 -1.70910954e-01 3.90296221e-01 1.26219302e-01
7.45258749e-01 1.83400616e-01 -3.11100721e-01 9.89881575e-01
-7.65548646e-01 3.01993098e-02 3.37845027e-01 4.25576061e-01
-5.36090255e-01 -1.23536050e+00 -7.78261185e-01 1.12304166e-01
-7.07081854e-01 -4.62003499e-01 -1.24130361e-01 2.62717843e-01
2.99520105e-01 7.22789526e-01 4.00307216e-02 -6.36808157e-01
1.79174870e-01 4.56895798e-01 4.14065063e-01 -7.75919974e-01
-6.78592980e-01 1.05326509e+00 -7.57088065e-02 -1.28487600e-02
-4.50432628e-01 -4.59664524e-01 -1.44716442e+00 2.11328894e-01
-2.00170144e-01 3.75082552e-01 4.61141706e-01 8.74506652e-01
-2.13332266e-01 9.85356569e-01 5.94459474e-01 -1.08554709e+00
-3.80272985e-01 -1.39937150e+00 -9.79420841e-01 2.69891322e-01
2.03989208e-01 -1.90359518e-01 -1.15713395e-01 3.78079593e-01] | [15.802563667297363, 5.485580921173096] |
e98ec003-d82e-442a-abbc-7345f1601f80 | updated-headline-generation-creating-updated | null | null | https://aclanthology.org/2022.acl-long.446 | https://aclanthology.org/2022.acl-long.446.pdf | Updated Headline Generation: Creating Updated Summaries for Evolving News Stories | We propose the task of updated headline generation, in which a system generates a headline for an updated article, considering both the previous article and headline. The system must identify the novel information in the article update, and modify the existing headline accordingly. We create data for this task using the NewsEdits corpus by automatically identifying contiguous article versions that are likely to require a substantive headline update. We find that models conditioned on the prior headline and body revisions produce headlines judged by humans to be as factual as gold headlines while making fewer unnecessary edits compared to a standard headline generation model. Our experiments establish benchmarks for this new contextual summarization task. | ['Mark Dredze', 'Adrian Benton', 'Sheena Panthaplackel'] | null | null | null | null | acl-2022-5 | ['headline-generation'] | ['natural-language-processing'] | [ 7.21132159e-01 8.91955972e-01 -3.68868828e-01 -3.86989325e-01
-1.26666164e+00 -6.79019034e-01 1.03413582e+00 9.77203727e-01
-5.02108157e-01 1.42682397e+00 1.13217866e+00 -2.61559606e-01
1.49137020e-01 -4.39019382e-01 -1.01283026e+00 2.35153303e-01
2.78162301e-01 6.27124071e-01 2.50962555e-01 -4.74009365e-01
8.53089154e-01 2.50776168e-02 -1.08741713e+00 5.06345987e-01
1.20654655e+00 3.89093429e-01 4.64376599e-01 9.79212344e-01
5.05381525e-02 4.81820911e-01 -1.16585135e+00 -7.29667127e-01
-3.77356298e-02 -8.11011374e-01 -1.06151223e+00 1.98468640e-01
1.09094810e+00 -2.37661004e-01 -9.94602963e-02 8.48883331e-01
3.80305380e-01 3.08804482e-01 8.31072628e-01 -7.47189760e-01
-7.98697352e-01 1.41000676e+00 -5.76437354e-01 5.91838777e-01
9.95957732e-01 -3.18171531e-01 1.33487487e+00 -6.75932407e-01
1.28209591e+00 1.10371065e+00 7.67507732e-01 2.02720538e-01
-1.24918222e+00 -6.66518286e-02 4.21679705e-01 -3.97774801e-02
-3.86124194e-01 -8.13885808e-01 3.26271892e-01 -3.06116760e-01
1.29564106e+00 5.51295400e-01 6.13269448e-01 1.05996060e+00
9.88940716e-01 5.72062850e-01 2.88208991e-01 -6.51614010e-01
-2.74597835e-02 -5.87448813e-02 5.53711534e-01 3.09420317e-01
6.73550248e-01 -5.13838708e-01 -4.08908665e-01 -3.55154961e-01
-9.81631204e-02 -5.11059999e-01 -3.64995778e-01 4.45297182e-01
-1.25828350e+00 6.89622521e-01 -2.46961359e-02 1.44264251e-01
-7.52920091e-01 1.68203264e-02 6.13958418e-01 4.19970930e-01
7.04381943e-01 1.32714200e+00 -4.67932045e-01 -2.08019972e-01
-1.31790757e+00 7.25528240e-01 1.15102339e+00 1.31464803e+00
5.94528794e-01 -4.58928674e-01 -6.61659598e-01 7.28812277e-01
-1.42063424e-01 4.48227465e-01 5.30542672e-01 -8.92618895e-01
8.74009371e-01 2.68819124e-01 4.15906370e-01 -1.05227673e+00
-3.65438312e-01 -7.07205832e-01 -4.40044224e-01 -4.90533978e-01
-3.39093059e-01 -4.84203517e-01 -4.34128910e-01 1.30007076e+00
-8.43896717e-02 -4.44995970e-01 5.11012152e-02 -8.75478685e-02
8.35321069e-01 8.88765395e-01 -3.93135250e-01 -1.06532824e+00
1.02589035e+00 -1.19313800e+00 -1.08097613e+00 -3.83408159e-01
7.46531427e-01 -1.36077499e+00 9.17827189e-01 2.62495279e-01
-1.46996260e+00 -4.93182957e-01 -1.34881449e+00 -1.44487113e-01
7.46701658e-02 1.97746947e-01 1.11465100e-02 -1.71783090e-01
-8.84703577e-01 7.72297025e-01 -2.78728157e-01 -7.01418459e-01
-4.36002202e-02 -5.00726461e-01 -3.07863206e-01 3.37684870e-01
-1.19582462e+00 1.12907898e+00 5.47821820e-01 -2.95410722e-01
-3.18345726e-01 -9.69976902e-01 -9.79864120e-01 -1.70495212e-01
5.11584878e-01 -1.03773999e+00 2.07973504e+00 -6.56026423e-01
-1.02669013e+00 6.19370639e-01 -5.13447762e-01 -8.94569039e-01
6.20814562e-01 -7.32757330e-01 -5.58635354e-01 3.42710642e-03
7.45209098e-01 3.11525315e-01 9.36237633e-01 -1.03043401e+00
-1.08191180e+00 4.42561395e-02 -1.69738054e-01 4.13993925e-01
1.95697412e-01 1.90664425e-01 -5.57468295e-01 -1.27840483e+00
-3.80681098e-01 -9.24129665e-01 -2.24978179e-01 -8.46764088e-01
-1.20364666e+00 -1.54297352e-01 4.16807711e-01 -1.12889218e+00
2.14548922e+00 -1.44164729e+00 2.08203986e-01 -5.26866727e-02
1.95091888e-01 -1.85043961e-01 -8.23311731e-02 8.13258648e-01
6.77781254e-02 4.84998226e-01 -2.37006113e-01 -3.57431233e-01
-4.15776856e-02 -2.90596932e-01 -8.07867706e-01 7.35130459e-02
-2.50080854e-01 8.10250103e-01 -8.96988809e-01 -5.15998960e-01
-5.90946794e-01 -6.40360832e-01 -7.89626896e-01 2.22966418e-01
-6.64855540e-01 -1.43305779e-01 -9.58784223e-02 -4.83865738e-02
1.78541243e-01 -1.52697906e-01 -1.35389686e-01 -3.43868643e-01
-2.85476714e-01 7.18663394e-01 -6.41416490e-01 1.40177298e+00
-4.52538908e-01 9.77321029e-01 -2.74036527e-01 -8.46468732e-02
6.58788800e-01 1.47934064e-01 -1.16712518e-01 -6.31562591e-01
-1.74789965e-01 1.26907788e-03 -2.08955124e-01 -5.46418309e-01
1.51084363e+00 7.95799345e-02 -5.44191241e-01 1.03606045e+00
-1.08068161e-01 -4.87070173e-01 8.33689749e-01 1.06781209e+00
8.99313569e-01 -2.08537847e-01 8.58941257e-01 7.70652890e-02
1.40591219e-01 2.96701282e-01 4.26352888e-01 1.33020282e+00
3.45694900e-01 4.78409588e-01 6.22889757e-01 -2.00396627e-02
-1.25070405e+00 -6.55602276e-01 5.26014976e-02 1.04849172e+00
-1.92947369e-02 -1.22806871e+00 -8.02825928e-01 -8.56776834e-01
-6.62948117e-02 1.57280326e+00 -7.36722171e-01 -1.71929821e-01
-7.07301795e-01 -2.15720415e-01 3.99010658e-01 1.88973859e-01
1.69214770e-01 -9.32929635e-01 -5.36981106e-01 6.09294713e-01
-4.64529276e-01 -7.85675704e-01 -1.25126147e+00 -1.12377666e-01
-6.09049141e-01 -1.07055902e+00 -6.32253468e-01 -7.59966314e-01
6.87370896e-01 -8.90803933e-02 1.25751126e+00 1.79837793e-01
1.92732990e-01 2.31866658e-01 -5.12675881e-01 -6.26599014e-01
-1.14335358e+00 5.08697689e-01 1.08004428e-01 -5.28411686e-01
-3.51226687e-01 1.82194114e-02 -1.27145797e-01 -1.58051059e-01
-8.81138980e-01 4.28708881e-01 4.93595183e-01 8.32473397e-01
4.70735967e-01 -2.15238675e-01 1.03043199e+00 -1.73437238e+00
1.44185686e+00 -5.18230915e-01 -1.30926713e-01 6.59226894e-01
-8.20174754e-01 1.75445110e-01 7.19254732e-01 9.98183787e-02
-1.50860238e+00 -5.41472733e-01 -2.11326733e-01 5.64091802e-01
5.44446893e-02 1.12705576e+00 2.81424910e-01 8.74942303e-01
9.44784403e-01 1.03831649e-01 -2.39560142e-01 -6.02560401e-01
6.46776021e-01 4.80372339e-01 9.40969169e-01 -3.83772701e-01
7.54392862e-01 -7.90010691e-02 -7.38188505e-01 -8.53804410e-01
-1.17522943e+00 -2.12658182e-01 -7.30780065e-01 -3.02936047e-01
5.37961900e-01 -5.63798666e-01 3.03463101e-01 1.71996504e-01
-1.60176635e+00 3.56412753e-02 -5.57819426e-01 1.32442430e-01
-3.93955618e-01 8.06174099e-01 -4.16119844e-01 7.21855313e-02
-7.23403037e-01 -7.91209638e-01 7.67034471e-01 3.24908972e-01
-1.39398432e+00 -1.01386583e+00 4.88322943e-01 1.30173609e-01
2.07958475e-01 2.43619949e-01 1.34230530e+00 -1.12201619e+00
1.95796490e-01 -5.06520689e-01 2.12161273e-01 2.28904694e-01
3.91461372e-01 6.77852929e-01 -8.35382938e-02 -1.62791371e-01
-2.13145629e-01 -1.19288936e-02 8.60617220e-01 5.26964843e-01
5.42563379e-01 -1.24075937e+00 -5.81631541e-01 1.54682994e-01
6.74930215e-01 1.28636703e-01 3.94179583e-01 5.70064723e-01
3.74885470e-01 3.41884881e-01 8.42867255e-01 6.89948618e-01
3.51805866e-01 6.64136708e-01 -2.03780711e-01 2.39766791e-01
-1.80961698e-01 -7.15585768e-01 3.58457416e-01 1.30906498e+00
5.87991834e-01 -8.49726498e-01 -5.32838047e-01 6.56175971e-01
-1.85395586e+00 -1.22280192e+00 -2.58947998e-01 1.94554472e+00
1.28549278e+00 6.48103714e-01 -6.12856001e-02 -3.28122854e-01
8.94685149e-01 3.79493833e-01 -3.97823721e-01 -7.87651122e-01
-3.28526109e-01 -1.07742958e-01 2.89778858e-01 9.91811752e-01
-9.16750669e-01 8.33614409e-01 7.44821453e+00 5.27169287e-01
-6.56012952e-01 -1.98409706e-01 2.90694952e-01 -2.30704457e-01
-8.93592179e-01 1.46961033e-01 -1.18908846e+00 6.79858387e-01
9.46878493e-01 -1.21819675e+00 -2.21872136e-01 5.39015591e-01
6.28737271e-01 -4.17203933e-01 -1.41223633e+00 3.90803397e-01
6.05425596e-01 -1.90648401e+00 8.03525805e-01 -3.52515996e-01
1.23243499e+00 -3.19579989e-01 -2.66874403e-01 1.55007094e-01
3.01288605e-01 -5.49353302e-01 1.13851857e+00 7.78457463e-01
6.17387414e-01 -5.57904541e-01 8.48090231e-01 3.97085637e-01
-5.83013892e-01 3.89089555e-01 -1.83848202e-01 -1.97201576e-02
8.35801840e-01 6.74556375e-01 -1.19409645e+00 5.56191623e-01
7.77788535e-02 1.15111959e+00 -1.05620849e+00 1.21617794e+00
-3.86419863e-01 5.64119756e-01 1.30291268e-01 -1.12377360e-01
4.05248851e-02 -1.05065167e-01 1.15517485e+00 1.54225135e+00
2.76663393e-01 -1.50796443e-01 1.04058124e-01 3.70334595e-01
-6.16460800e-01 2.21585825e-01 -5.96529543e-01 -2.24982888e-01
7.63066471e-01 8.68294418e-01 -4.07114953e-01 -8.20259094e-01
2.94049717e-02 1.06987584e+00 1.12253033e-01 2.51313180e-01
-7.50690579e-01 -1.03973103e+00 5.19314557e-02 1.21042803e-01
1.22979403e-01 1.25739560e-01 -3.60132068e-01 -1.04415631e+00
1.04884334e-01 -1.02609670e+00 4.57223475e-01 -1.02797806e+00
-1.02793050e+00 7.98660696e-01 1.00722350e-01 -1.09059787e+00
-5.46839714e-01 2.52616763e-01 -1.14152551e+00 4.75164711e-01
-9.79094923e-01 -4.26671416e-01 6.58261105e-02 -2.18766466e-01
1.20489836e+00 -7.37801194e-02 4.39363450e-01 -1.45363152e-01
-6.97441161e-01 7.18389034e-01 2.14733258e-01 -2.21166819e-01
1.14581072e+00 -1.28625047e+00 9.43463326e-01 1.27856469e+00
-2.76742298e-02 1.04759645e+00 1.39490139e+00 -1.40770185e+00
-7.74910331e-01 -1.23965180e+00 1.59271336e+00 -5.19083977e-01
8.11090469e-01 1.73751876e-01 -7.45954514e-01 1.09579730e+00
9.39504623e-01 -1.41113484e+00 6.23371720e-01 5.67573030e-03
-3.77202183e-02 2.89798796e-01 -6.16950989e-01 9.19558346e-01
8.53711486e-01 -3.45871568e-01 -1.83860910e+00 7.53356278e-01
1.19119883e+00 -6.00440860e-01 -6.50529623e-01 5.34648262e-02
3.12804222e-01 -3.33073169e-01 2.95337945e-01 -9.46255684e-01
9.77843225e-01 -1.99892431e-01 3.63880485e-01 -1.98626316e+00
-4.49815840e-01 -1.27748299e+00 -1.07670158e-01 1.38306236e+00
1.18681300e+00 -4.29963678e-01 2.03108899e-02 4.56225991e-01
-7.96977401e-01 -3.71889919e-01 -6.01943076e-01 -5.66299200e-01
-2.66909450e-01 -6.63537979e-02 4.20232952e-01 6.01158559e-01
5.10038078e-01 9.81136382e-01 -3.13143015e-01 -4.01219755e-01
2.66216964e-01 5.20280115e-02 1.01320839e+00 -1.13750291e+00
-1.10004731e-01 -6.28565371e-01 5.01747847e-01 -1.30975759e+00
1.79867208e-01 -9.84047055e-01 2.48740539e-01 -2.14447856e+00
3.88278723e-01 1.00297861e-01 4.37802464e-01 1.65495917e-01
-5.01118779e-01 -3.64451706e-01 6.78687394e-02 4.69841897e-01
-9.06149626e-01 4.82568055e-01 9.82961476e-01 -2.61119962e-01
-4.22104895e-01 1.06230050e-01 -1.34374678e+00 7.53705204e-01
4.34377491e-01 -5.49365819e-01 -2.24968597e-01 -2.36425877e-01
7.55108416e-01 5.84328398e-02 -2.74877518e-01 -6.66478336e-01
5.06750464e-01 -2.15238258e-01 2.26585031e-01 -9.55651760e-01
-2.94245839e-01 -1.95994854e-01 1.75800964e-01 2.81796992e-01
-1.28941965e+00 8.65672350e-01 1.68479547e-01 7.09454894e-01
-1.73072189e-01 -6.57065511e-01 4.58670735e-01 5.47058335e-05
-5.45278728e-01 -1.03763498e-01 -7.86384642e-01 6.97356105e-01
9.18063104e-01 -2.89530277e-01 -9.14464712e-01 -7.09028006e-01
-5.12196839e-01 2.61387259e-01 7.76298761e-01 5.02786219e-01
4.97684747e-01 -8.59222531e-01 -1.05826247e+00 -2.58351594e-01
3.15142602e-01 -2.30269700e-01 -1.93570703e-02 5.86895704e-01
-7.05935538e-01 6.02464259e-01 -1.49777881e-03 2.27326136e-02
-1.28939414e+00 2.39769429e-01 -1.81244150e-01 -3.96454692e-01
-8.17093134e-01 6.94333196e-01 -1.30877078e-01 -1.87203810e-02
1.55477062e-01 -6.77998662e-01 -3.35687876e-01 6.92050457e-01
8.36002827e-01 5.00502408e-01 4.43884075e-01 -4.63870019e-01
1.32703722e-01 1.16667055e-01 -9.54797626e-01 -2.45660111e-01
1.22733343e+00 -5.41125894e-01 -2.18561500e-01 6.33036137e-01
1.15175462e+00 6.61878526e-01 -7.44139254e-01 -2.63151407e-01
3.39880347e-01 -1.27621427e-01 -2.60248661e-01 -9.71351266e-01
-1.50679037e-01 -5.74477129e-02 -6.40546620e-01 3.93119246e-01
5.10014832e-01 2.04851285e-01 1.08828783e+00 7.97392786e-01
-1.29816895e-02 -1.49939263e+00 7.27917487e-03 9.18737411e-01
1.70127642e+00 -7.44724572e-01 6.75368845e-01 5.19172587e-02
-9.02303576e-01 1.07892680e+00 5.57270408e-01 1.60772547e-01
3.03740412e-01 -2.50963625e-02 -1.54834330e-01 -3.63497078e-01
-1.17995572e+00 5.09020925e-01 4.69365001e-01 2.40060553e-01
4.53997791e-01 -7.27235153e-02 -6.60004914e-01 6.34599745e-01
-8.69932890e-01 -3.07300001e-01 1.43592834e+00 1.02967429e+00
-6.61923647e-01 -5.66614389e-01 -1.76444635e-01 1.20466900e+00
-4.66753185e-01 -2.86588132e-01 -9.47334409e-01 5.21866858e-01
-5.73086739e-01 9.63779151e-01 1.84677601e-01 -1.88501120e-01
6.47508025e-01 -4.24533300e-02 1.84476659e-01 -1.21984005e+00
-9.20836985e-01 5.18464148e-02 8.68275583e-01 -9.15560871e-02
-5.07551916e-02 -8.11860621e-01 -1.24571252e+00 -3.74546826e-01
-3.32800120e-01 6.61435783e-01 3.42526048e-01 9.89953518e-01
6.56677365e-01 6.19994760e-01 4.65368301e-01 -6.67973459e-01
-5.96080482e-01 -1.26155674e+00 -1.64926514e-01 6.20919406e-01
4.58413005e-01 -1.03480332e-01 -3.45591515e-01 5.06622314e-01] | [12.372573852539062, 9.414962768554688] |
dd9e30da-3155-477a-ba52-ae4989820fdd | language-relatedness-and-lexical-closeness | null | null | https://aclanthology.org/2021.wat-1.26 | https://aclanthology.org/2021.wat-1.26.pdf | Language Relatedness and Lexical Closeness can help Improve Multilingual NMT: IITBombay@MultiIndicNMT WAT2021 | Multilingual Neural Machine Translation has achieved remarkable performance by training a single translation model for multiple languages. This paper describes our submission (Team ID: CFILT-IITB) for the MultiIndicMT: An Indic Language Multilingual Task at WAT 2021. We train multilingual NMT systems by sharing encoder and decoder parameters with language embedding associated with each token in both encoder and decoder. Furthermore, we demonstrate the use of transliteration (script conversion) for Indic languages in reducing the lexical gap for training a multilingual NMT system. Further, we show improvement in performance by training a multilingual NMT system using languages of the same family, i.e., related languages. | ['Pushpak Bhattacharyya', 'Nikhil Saini', 'Jyotsana Khatri'] | null | null | null | null | acl-wat-2021-8 | ['transliteration'] | ['natural-language-processing'] | [-2.20730249e-02 -5.38941808e-02 -5.23174465e-01 -3.55761051e-01
-1.28967476e+00 -7.58919835e-01 5.76052070e-01 -3.05160910e-01
-6.29049778e-01 1.09056008e+00 3.39298040e-01 -1.04211199e+00
6.67002201e-01 -3.47249210e-01 -1.19270182e+00 -4.62685041e-02
2.93892503e-01 8.10865045e-01 -5.97692013e-01 -4.14393455e-01
-4.04306471e-01 -1.23580471e-01 -1.48887649e-01 7.21324027e-01
1.04813910e+00 3.06635827e-01 2.87182897e-01 5.00327766e-01
-8.90438929e-02 6.72594786e-01 -4.97857720e-01 -8.67509604e-01
5.89353383e-01 -6.98234379e-01 -7.57270515e-01 -5.20198524e-01
5.97404480e-01 -1.73260674e-01 -3.95730436e-01 8.44524264e-01
5.88924527e-01 -4.79176760e-01 5.08965313e-01 -8.05036604e-01
-1.19113982e+00 1.42711854e+00 -5.24064183e-01 2.26898268e-01
-2.89766993e-02 -6.89552864e-03 1.30147910e+00 -1.30402529e+00
9.31966603e-01 1.29555202e+00 7.45328426e-01 5.13030946e-01
-1.39753008e+00 -6.98710084e-01 -1.91004902e-01 1.83336258e-01
-1.42719364e+00 -6.56833649e-01 1.44544393e-01 -1.74482927e-01
1.81864107e+00 1.56959947e-02 3.28798413e-01 1.42195892e+00
6.96994483e-01 9.45508242e-01 1.33806658e+00 -7.64156103e-01
-4.08218652e-01 3.47013295e-01 -2.25530148e-01 6.47969425e-01
-9.83717814e-02 2.90741995e-02 -5.63344240e-01 2.12199345e-01
4.67613667e-01 -7.46436000e-01 1.99306995e-01 3.56199652e-01
-1.66176498e+00 8.40329826e-01 9.57491770e-02 4.62247044e-01
-2.40171820e-01 2.27082610e-01 8.10629487e-01 8.87217462e-01
7.13374436e-01 5.19579530e-01 -8.50746453e-01 -1.43894285e-01
-8.36357594e-01 -4.39020187e-01 7.58106947e-01 1.37546706e+00
6.22427046e-01 4.75666583e-01 -2.53654391e-01 1.26607406e+00
-7.23737329e-02 9.22470748e-01 6.77745521e-01 -5.43135762e-01
1.11183798e+00 6.35094941e-02 -3.59126568e-01 -1.67453915e-01
-1.67575970e-01 -5.33909738e-01 -6.19024754e-01 -5.26000917e-01
6.78410605e-02 -6.73376143e-01 -7.70504951e-01 1.82170582e+00
-2.26162344e-01 -3.10843110e-01 6.00680470e-01 4.16418672e-01
3.49373519e-01 1.04592896e+00 -7.88218994e-03 -4.79804501e-02
1.08509123e+00 -1.56340671e+00 -7.73473799e-01 -4.63393837e-01
1.09087706e+00 -1.29798007e+00 1.05488694e+00 9.81489420e-02
-1.17045629e+00 -4.67076331e-01 -1.03694916e+00 -3.85218799e-01
-4.37347621e-01 7.15977490e-01 4.49645013e-01 3.18351001e-01
-1.14685655e+00 3.13309819e-01 -8.29241753e-01 -6.23274028e-01
-2.08211154e-01 3.45769614e-01 -4.79345769e-01 -2.13208884e-01
-1.64567709e+00 1.65567303e+00 4.65144664e-01 2.62916293e-02
-8.94062519e-01 -6.38407707e-01 -7.90001035e-01 -4.05562311e-01
-3.22729051e-01 -6.22455060e-01 1.38292456e+00 -1.04981589e+00
-1.61904085e+00 9.18237746e-01 -2.75763810e-01 -6.65381730e-01
4.56254184e-01 -3.98657620e-01 -7.18869925e-01 -2.90008932e-01
4.83079106e-01 8.44925880e-01 4.03203487e-01 -6.04266524e-01
-5.99777579e-01 -7.54844397e-02 -3.12267989e-01 4.17893887e-01
-5.87917328e-01 6.74065113e-01 -4.34130192e-01 -7.27278769e-01
-4.24843937e-01 -1.17851579e+00 -1.29984468e-01 -8.20397973e-01
-6.00243032e-01 -1.43869236e-01 4.86572385e-01 -1.44934857e+00
1.14103281e+00 -1.90268552e+00 4.17243093e-01 -1.98528513e-01
-4.04333740e-01 1.02981627e-01 -5.48633993e-01 6.51920617e-01
-1.46061644e-01 2.98502773e-01 -2.83149891e-02 -6.37857080e-01
1.13887966e-01 4.90853667e-01 -3.18876177e-01 2.46940061e-01
4.87252146e-01 1.34803319e+00 -6.71005368e-01 -4.61703598e-01
-6.00822410e-03 4.93503898e-01 -2.13385612e-01 2.72826459e-02
-8.46290588e-02 5.36993206e-01 6.77553862e-02 5.39693475e-01
1.67480037e-01 1.86632723e-01 3.66002649e-01 -1.22269906e-01
-4.02621597e-01 9.91949677e-01 -3.15060407e-01 2.02303982e+00
-1.01761365e+00 7.96987712e-01 -2.22388893e-01 -5.35527408e-01
7.57597446e-01 7.47362316e-01 1.98629230e-01 -9.53934610e-01
4.52761576e-02 9.18298781e-01 2.56286055e-01 -1.74316674e-01
5.78720093e-01 -9.74226370e-02 -3.68429899e-01 7.34206617e-01
5.78091741e-01 -4.27988470e-02 3.53622019e-01 -1.05924338e-01
7.46702909e-01 3.67296189e-01 1.85140014e-01 -4.97827917e-01
5.80425411e-02 2.47198820e-01 4.37623709e-01 3.50138515e-01
9.05243754e-02 3.13672811e-01 -2.73218239e-03 -5.35425305e-01
-1.71995413e+00 -1.04355431e+00 -2.46491194e-01 1.25943971e+00
-6.92274809e-01 -5.03013730e-01 -7.74308503e-01 -7.10696220e-01
-2.25288078e-01 8.99813116e-01 -3.21066648e-01 3.24771069e-02
-1.27907121e+00 -9.77107048e-01 1.24956560e+00 4.06187207e-01
2.02931628e-01 -8.98972332e-01 2.54884690e-01 5.07722914e-01
-7.08791554e-01 -1.40430892e+00 -9.13138390e-01 7.11948872e-01
-6.45500183e-01 -3.35580289e-01 -7.03124881e-01 -1.17968047e+00
2.61809975e-01 -1.63790479e-01 1.32661378e+00 -4.59718436e-01
1.42759636e-01 -1.53054714e-01 -1.30898710e-02 -2.63982832e-01
-9.27891076e-01 8.48524094e-01 3.97190481e-01 -4.13710952e-01
4.26774830e-01 -3.73200446e-01 1.64250374e-01 1.97502092e-01
-4.21193212e-01 4.28131878e-01 7.09502220e-01 8.15059721e-01
5.24055243e-01 -6.97408378e-01 4.75383103e-01 -7.15139687e-01
7.19928205e-01 -4.61006373e-01 -4.13552701e-01 6.18128419e-01
-6.03298187e-01 2.42549017e-01 9.03578758e-01 -4.26819503e-01
-7.94766426e-01 -2.38674760e-01 -2.49659508e-01 -1.63040325e-01
3.83034796e-01 7.14436531e-01 -7.23605677e-02 2.17283443e-01
7.13459313e-01 3.52939278e-01 -3.81797165e-01 -5.10584235e-01
7.12343693e-01 7.92164922e-01 4.92303371e-01 -7.44762003e-01
6.44852400e-01 -3.56125444e-01 -4.43636656e-01 -4.17375714e-01
-5.41013420e-01 1.17316082e-01 -8.93428445e-01 1.58831820e-01
8.61789703e-01 -1.37280464e+00 1.00053430e-01 1.79226637e-01
-1.53945506e+00 -5.79940975e-01 -1.07472993e-01 7.32677400e-01
-4.95292336e-01 -1.64522856e-01 -1.36542702e+00 -5.71169145e-02
-7.66363502e-01 -1.31348550e+00 8.54244351e-01 -3.49198103e-01
-4.15842235e-01 -1.30859268e+00 3.67720187e-01 1.70649633e-01
5.87481201e-01 -2.81812102e-01 1.20959485e+00 -6.48343205e-01
-3.02371591e-01 -8.44935793e-03 -1.60489857e-01 6.15350664e-01
7.46292621e-02 -1.88461214e-01 -5.36731422e-01 -4.97524261e-01
-3.88070583e-01 -5.85433304e-01 5.55357575e-01 1.83601975e-01
8.34860355e-02 -4.43595290e-01 -1.10771812e-01 6.92583859e-01
1.33898389e+00 6.70699924e-02 2.89951980e-01 5.32634079e-01
9.65426624e-01 3.44635546e-01 9.66516882e-02 -2.03223005e-01
7.14349866e-01 7.94989347e-01 -3.88536304e-01 -4.86350149e-01
-5.66082776e-01 -5.29536963e-01 1.11523545e+00 1.92175770e+00
1.84600562e-01 -1.97881535e-01 -1.04623127e+00 6.40005171e-01
-1.67336774e+00 -3.18866014e-01 -3.15320998e-01 1.87762630e+00
1.37351632e+00 -1.24295913e-01 -2.89713144e-01 -7.02016115e-01
5.03140509e-01 -1.85282618e-01 -3.98161411e-01 -9.95481312e-01
-7.07018793e-01 5.16556680e-01 8.43607664e-01 7.86443710e-01
-8.42742622e-01 1.70384526e+00 7.13182878e+00 8.58724356e-01
-1.27022529e+00 6.93317652e-01 5.27675867e-01 -5.46924621e-02
-2.88185179e-01 1.83444470e-01 -1.04306400e+00 2.26038307e-01
1.61339903e+00 -3.18976641e-01 9.08902586e-01 4.56538916e-01
1.68064982e-01 5.12944162e-01 -1.19503069e+00 4.95578229e-01
1.43592536e-01 -1.04691970e+00 1.75044179e-01 -2.67733950e-02
1.10500276e+00 9.79069531e-01 8.50798711e-02 6.69498682e-01
8.97566557e-01 -1.12190139e+00 8.89511287e-01 4.40724827e-02
1.30377936e+00 -7.63100088e-01 6.48054481e-01 2.25267723e-01
-1.08261299e+00 2.90271908e-01 -4.58734900e-01 1.32442966e-01
2.14454725e-01 2.82265961e-01 -1.14226270e+00 8.73321652e-01
3.91297996e-01 9.06984210e-01 -6.02751672e-01 3.60962451e-01
-6.43459260e-01 1.00663841e+00 -2.76531518e-01 3.54637563e-01
4.76776332e-01 -3.34708601e-01 4.44900781e-01 1.68094313e+00
6.27420366e-01 -9.67449963e-01 2.85962284e-01 5.32268584e-01
-4.56387341e-01 6.41523421e-01 -8.00495028e-01 -2.74728507e-01
3.32787991e-01 1.01445091e+00 -1.12201041e-02 -4.10636008e-01
-5.45531452e-01 1.40169442e+00 7.06984043e-01 3.40623796e-01
-7.67270565e-01 -1.58392116e-01 4.81048048e-01 -3.83222699e-01
-1.38942348e-02 -5.34711659e-01 -3.58141243e-01 -1.50284028e+00
-4.62409072e-02 -1.24590790e+00 7.40522668e-02 -7.42264211e-01
-1.23476911e+00 1.11572433e+00 -2.79183954e-01 -1.05444574e+00
-6.95809484e-01 -5.91665089e-01 -1.32574379e-01 1.51346326e+00
-1.48646629e+00 -1.77207005e+00 8.86019826e-01 4.19122219e-01
5.83080292e-01 -5.02664328e-01 1.22320497e+00 8.75788629e-01
-7.15908587e-01 1.12043321e+00 4.36758250e-01 5.03927231e-01
1.27337587e+00 -1.31070757e+00 1.06860328e+00 9.81806755e-01
4.55445379e-01 8.68135154e-01 3.36616576e-01 -6.82009876e-01
-1.42382729e+00 -1.46296930e+00 1.71629488e+00 -6.36241496e-01
1.03921449e+00 -6.56371355e-01 -5.14423132e-01 1.31538868e+00
1.25551295e+00 -5.23323298e-01 6.89449131e-01 2.90606022e-01
-5.05502403e-01 1.32487774e-01 -5.12032926e-01 8.05853486e-01
6.82759225e-01 -9.21991825e-01 -3.95684808e-01 6.27222717e-01
1.05735970e+00 -3.94851923e-01 -1.22161317e+00 2.67960638e-01
6.41891837e-01 5.74242175e-02 6.92099094e-01 -7.54596710e-01
5.81645429e-01 -1.42956629e-01 -4.41261739e-01 -1.91338921e+00
-1.98494077e-01 -6.68592036e-01 2.42901608e-01 1.04396415e+00
1.25011516e+00 -6.08932316e-01 4.21189219e-02 -3.28555018e-01
-4.89854664e-01 -5.18755257e-01 -1.14888346e+00 -1.05846894e+00
8.71164382e-01 -3.71605456e-01 3.04131150e-01 1.26100135e+00
7.51687214e-02 1.05744874e+00 -8.26816559e-01 -1.62278101e-01
3.05251360e-01 -2.17513442e-01 5.42577803e-01 -5.18473446e-01
-5.25629342e-01 -1.85654700e-01 2.49180719e-01 -8.45898986e-01
3.84865642e-01 -1.77502263e+00 5.24663106e-02 -1.43424237e+00
2.73115098e-01 -2.32116476e-01 -2.80998111e-01 9.82834518e-01
-9.63481367e-02 5.14214814e-01 2.22734451e-01 4.45677578e-01
-2.22539976e-01 3.65894049e-01 1.07272339e+00 -2.14635611e-01
1.27048820e-01 -3.70794743e-01 -3.65884662e-01 1.95593402e-01
1.00381947e+00 -7.19171584e-01 1.03753984e-01 -1.43925261e+00
3.60888630e-01 6.24723695e-02 -3.22296202e-01 -7.33598411e-01
-1.32393790e-02 1.43221930e-01 2.58274972e-01 -3.82036448e-01
1.72117591e-01 -3.93876851e-01 1.27046183e-01 4.23049152e-01
-4.64628488e-01 9.29154873e-01 5.02648592e-01 -2.93145031e-01
-1.76269934e-01 -3.21249617e-03 5.42188466e-01 -1.60179630e-01
-2.33391568e-01 1.17265105e-01 -6.67495728e-01 8.27673674e-02
4.24201488e-01 3.85076791e-01 -3.82981151e-01 -7.43556172e-02
-5.09313226e-01 2.18211636e-01 3.68982196e-01 7.65517950e-01
3.98280807e-02 -1.52827513e+00 -1.37218428e+00 2.41207346e-01
1.85332790e-01 -8.49292576e-01 -4.73334074e-01 1.00807464e+00
-3.47615272e-01 6.92669630e-01 -3.73160779e-01 -3.45827043e-01
-9.34630692e-01 1.75063297e-01 5.10848641e-01 -7.18761742e-01
-2.23731607e-01 6.21220171e-01 -2.40892082e-01 -1.24190891e+00
-2.41310909e-01 -2.08545119e-01 3.68662536e-01 -2.11754844e-01
2.47365132e-01 5.35870902e-02 4.27174568e-01 -7.34476566e-01
-1.80893764e-01 3.11281800e-01 -3.23010534e-01 -6.74612224e-01
1.09436035e+00 -1.16886139e-01 -5.49065590e-01 6.49143219e-01
1.39326286e+00 2.12746620e-01 -5.62367260e-01 -4.87410188e-01
2.49773979e-01 3.43450159e-01 -1.58031076e-01 -1.33817041e+00
-6.52570248e-01 1.07337439e+00 3.71677727e-01 -5.74416816e-01
7.49394536e-01 -2.97929645e-01 1.11982656e+00 5.45226038e-01
4.85562027e-01 -1.19672441e+00 -5.67488611e-01 1.34642661e+00
7.20749855e-01 -1.19718659e+00 -4.05689716e-01 1.66778877e-01
-6.51266694e-01 1.11088622e+00 4.80184287e-01 1.57020763e-01
1.41872838e-01 5.81446111e-01 5.53618610e-01 3.61420274e-01
-1.00227296e+00 1.76714346e-01 3.07792991e-01 2.11971030e-01
1.00462186e+00 6.20673478e-01 -6.84055328e-01 3.33423883e-01
-5.98097503e-01 -2.74793446e-01 2.19493181e-01 7.30502725e-01
1.00445785e-01 -1.70170569e+00 -1.49926454e-01 1.08129315e-01
-7.23472595e-01 -1.06774116e+00 -4.91096973e-01 7.21338689e-01
2.03340098e-01 8.32041800e-01 -2.27349043e-01 -5.51929533e-01
1.71054125e-01 4.81458127e-01 6.39010847e-01 -7.61287987e-01
-9.41418886e-01 2.45781273e-01 4.74203765e-01 -2.17988044e-01
1.55943349e-01 -6.56421125e-01 -8.73850465e-01 -2.48292044e-01
-2.14745272e-02 2.47907802e-01 1.06562757e+00 8.93930435e-01
3.00676823e-01 5.41310787e-01 5.85521996e-01 -3.64582539e-01
-4.90139455e-01 -1.50259757e+00 4.89478409e-02 -2.01612234e-01
-3.07731703e-02 1.39676154e-01 7.40937889e-02 2.16613263e-01] | [11.581209182739258, 10.40008544921875] |
d4d9752c-1baa-4720-ad70-0512daeb6885 | dmrnet-learning-discriminative-features-with | null | null | https://ieeexplore.ieee.org/document/9944858 | https://www.zdzheng.xyz/files/Han_TPAMI22.pdf | DMRNet++: Learning Discriminative Features with Decoupled Networks and Enriched Pairs for One-Step Person Search | Person search aims at localizing and recognizing query persons from raw video frames, which is a combination of two sub-tasks, i.e., pedestrian detection and person re-identification. The dominant fashion is termed as the one-step person search that jointly optimizes detection and identification in a unified network, exhibiting higher efficiency. However, there remain major challenges: (i) conflicting objectives of multiple sub-tasks under the shared feature space, (ii) inconsistent memory bank caused by the limited batch size, (iii) underutilized unlabeled identities during the identification learning. To address these issues, we develop an enhanced decoupled and memory-reinforced network (DMRNet++). First, we simplify the standard tightly coupled pipelines and establish a task-decoupled framework (TDF). Second, we build a memory-reinforced mechanism (MRM), with a slow-moving average of the network to better encode the consistency of the memorized features. Third, considering the potential of unlabeled samples, we model the recognition process as semi-supervised learning. An unlabeled-aided contrastive loss (UCL) is developed to boost the identification feature learning by exploiting the aggregation of unlabeled identities. Experimentally, the proposed DMRNet++ obtains the mAP of 94.5% and 52.1% on CUHK-SYSU and PRW datasets, which exceeds most existing methods. | ['Yi Yang', 'Nong Sang', 'Changxin Gao', 'Zehuan Yuan', 'Dongdong Yu', 'Kai Su', 'Zhedong Zheng', 'Chuchu Han'] | 2022-11-10 | null | null | null | ieee-transactions-on-pattern-analysis-and-24 | ['person-re-identification', 'pedestrian-detection', 'person-search'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-6.32661954e-02 -4.51703250e-01 -5.16090281e-02 -4.31992441e-01
-6.42256260e-01 -3.75499368e-01 6.62532210e-01 -3.64601284e-01
-8.53576303e-01 6.72225654e-01 8.22175741e-02 1.48064584e-01
1.09512143e-01 -5.40872395e-01 -6.74927294e-01 -6.85892522e-01
2.07785055e-01 4.12734270e-01 3.01805019e-01 3.08874637e-01
1.13372430e-02 2.97351748e-01 -1.83967233e+00 6.91783577e-02
8.82810473e-01 1.12760460e+00 1.54859990e-01 4.32220936e-01
5.99441268e-02 7.48218000e-01 -3.38133037e-01 -6.15969956e-01
2.88388908e-01 -8.22607428e-02 -7.83835292e-01 1.41304329e-01
5.65275908e-01 -6.34587526e-01 -5.38048267e-01 1.12008214e+00
8.12206566e-01 3.54488224e-01 4.00114119e-01 -1.25510526e+00
-7.65991628e-01 2.75711775e-01 -7.42558241e-01 3.28864485e-01
2.64175266e-01 3.72942954e-01 5.41504145e-01 -1.22596562e+00
3.70468616e-01 1.46478772e+00 5.65630734e-01 7.62122214e-01
-9.37316895e-01 -8.85307252e-01 4.30250496e-01 5.05176544e-01
-1.82111073e+00 -7.13697731e-01 2.84935832e-01 -4.06133056e-01
7.57259727e-01 3.40501070e-01 4.41858381e-01 1.04765379e+00
-5.51086307e-01 9.85013902e-01 7.69826114e-01 -3.30697119e-01
-9.14231166e-02 3.31200391e-01 6.03036463e-01 6.61049902e-01
2.66420960e-01 3.10956657e-01 -8.12632799e-01 2.21505538e-02
5.76792121e-01 2.73420125e-01 -1.13020159e-01 -1.00880720e-01
-9.80749011e-01 4.72334683e-01 2.60199904e-01 -4.26072329e-02
-3.29934776e-01 -1.82399437e-01 3.61823827e-01 5.06971180e-02
2.48726189e-01 -1.71610832e-01 -2.14243114e-01 6.86242804e-02
-9.54150677e-01 2.04675451e-01 4.50776905e-01 1.00292301e+00
8.77339900e-01 -5.39146736e-02 -4.83809024e-01 9.15958941e-01
5.67900598e-01 6.39418125e-01 4.73727077e-01 -6.07956231e-01
5.05560517e-01 6.53042614e-01 3.04312855e-01 -8.32062185e-01
-3.22774619e-01 -6.42820239e-01 -8.55133891e-01 -2.05712616e-01
4.06759262e-01 -1.79723620e-01 -8.67394388e-01 1.97883248e+00
4.50472027e-01 5.82520604e-01 -8.38702023e-02 1.23803437e+00
9.82353508e-01 3.86584491e-01 3.40200007e-01 -1.06750302e-01
1.51484847e+00 -1.48130190e+00 -5.04627764e-01 -3.47008288e-01
3.17200631e-01 -4.69188660e-01 6.98084414e-01 -8.93655419e-03
-1.02768016e+00 -1.05765080e+00 -1.00750065e+00 -9.93356928e-02
-4.67729062e-01 5.81507027e-01 3.55715871e-01 8.52068484e-01
-1.08977294e+00 1.07733101e-01 -5.26157558e-01 -3.65555614e-01
4.67696071e-01 6.44605577e-01 -4.01911676e-01 -1.81640059e-01
-1.18484998e+00 6.56068861e-01 3.36262912e-01 5.80529273e-01
-7.52575696e-01 -5.24282277e-01 -8.53014350e-01 6.80940077e-02
5.05727053e-01 -7.81542480e-01 8.89332354e-01 -7.93298841e-01
-1.24202061e+00 1.02798891e+00 -5.06811559e-01 -4.67222065e-01
7.09274113e-01 -4.40210283e-01 -5.22637129e-01 -1.58430472e-01
1.74506605e-01 7.72131443e-01 6.93063080e-01 -1.00930309e+00
-9.27724957e-01 -6.45329893e-01 -1.76820323e-01 4.19341058e-01
-5.99922359e-01 1.82338238e-01 -1.09051919e+00 -5.80184042e-01
-1.45305514e-01 -8.75230610e-01 -1.02979662e-02 -3.34475236e-03
-3.60759497e-01 -3.25595349e-01 5.12066841e-01 -9.06866610e-01
1.40737534e+00 -2.31452250e+00 -1.60900073e-03 1.46133721e-01
2.96905816e-01 5.59868217e-01 -2.03121260e-01 -1.70692116e-01
5.40034473e-02 -1.62129417e-01 2.30501547e-01 -8.84539545e-01
7.47392252e-02 -1.00764938e-01 3.73796150e-02 4.01415557e-01
1.29998773e-01 1.09003532e+00 -7.58377612e-01 -6.16589069e-01
1.86649978e-01 3.52674365e-01 -2.14723825e-01 2.32189745e-01
3.04195613e-01 4.57616925e-01 -3.24590176e-01 9.16334212e-01
1.00377989e+00 -3.61152530e-01 -4.75749969e-02 -3.50812018e-01
-2.07933441e-01 -2.07151920e-01 -1.56290972e+00 1.56282210e+00
6.53713942e-02 2.10412934e-01 1.33029178e-01 -8.42225671e-01
7.72349000e-01 6.16398454e-02 2.86409467e-01 -8.49811733e-01
3.60431485e-02 1.32055178e-01 -3.35820943e-01 -3.87566537e-01
5.17005980e-01 4.88973796e-01 1.74341321e-01 4.56704766e-01
2.14228705e-01 1.24464607e+00 1.10052913e-01 5.74976206e-02
7.96275556e-01 6.24184795e-02 7.96163157e-02 -8.45784768e-02
9.61518645e-01 -2.38450095e-01 7.76910424e-01 1.06761289e+00
-5.33116221e-01 4.54159021e-01 -2.14889452e-01 -6.38310850e-01
-7.43776441e-01 -9.91534531e-01 -7.27629662e-02 1.38011265e+00
5.12120783e-01 -1.35522917e-01 -6.76772535e-01 -6.40433192e-01
7.26467595e-02 3.72089632e-02 -4.59653497e-01 -2.13849798e-01
-5.59219062e-01 -1.07126892e+00 7.35653460e-01 6.50893152e-01
9.41142976e-01 -8.37794960e-01 -2.53785253e-01 2.43583433e-02
-2.23361269e-01 -1.11367691e+00 -9.64674950e-01 -2.29423359e-01
-3.12520623e-01 -1.05880630e+00 -9.24586177e-01 -9.23496127e-01
8.53079617e-01 5.32601833e-01 9.62567270e-01 2.85191596e-01
-2.27184951e-01 4.03139859e-01 -1.42100453e-01 1.11384159e-02
2.40087986e-01 6.97851032e-02 4.57935184e-01 5.94386816e-01
7.60022402e-01 -1.47452995e-01 -8.15378368e-01 5.84285438e-01
-5.43920755e-01 8.77272114e-02 6.55484855e-01 1.09179902e+00
6.27288938e-01 -3.36982012e-02 5.51735640e-01 -4.69439894e-01
2.83744693e-01 -3.25835109e-01 -5.27107537e-01 7.19489157e-01
-7.33185828e-01 -1.70552433e-01 3.05584669e-01 -7.63853073e-01
-1.26206982e+00 3.43300849e-01 1.72759313e-02 -4.98184413e-01
-8.92879292e-02 2.94047184e-02 -4.23431158e-01 -2.84852624e-01
2.83663988e-01 6.89683318e-01 -2.65640486e-02 -4.94862974e-01
2.02329889e-01 7.75825024e-01 8.99316430e-01 -5.33936143e-01
7.94858515e-01 4.15630490e-01 -4.27352518e-01 -4.47294235e-01
-5.25120854e-01 -6.99535251e-01 -6.08305097e-01 -3.61774415e-01
7.89535284e-01 -1.42060339e+00 -1.00904524e+00 9.71274614e-01
-1.03670180e+00 -1.85131077e-02 6.26106858e-02 4.83946353e-01
2.04160623e-02 5.68506777e-01 -6.24171257e-01 -1.06754935e+00
-6.81632400e-01 -1.06594217e+00 9.59318757e-01 7.65315056e-01
1.48742080e-01 -5.39035201e-01 -2.22161084e-01 5.98039567e-01
3.70947272e-01 -2.96762615e-01 3.08470011e-01 -8.16197157e-01
-6.83761358e-01 -2.36592501e-01 -6.87907100e-01 2.39858121e-01
-5.40874675e-02 -5.03844798e-01 -1.20872271e+00 -4.29978073e-01
-3.47531378e-01 -3.67886484e-01 1.02649188e+00 2.47313887e-01
9.37288761e-01 -1.93704754e-01 -4.82867181e-01 7.48887002e-01
1.04990232e+00 1.16474189e-01 4.49407548e-01 3.60480368e-01
6.95597351e-01 5.55629015e-01 5.00560880e-01 4.80894476e-01
7.06642866e-01 8.30094695e-01 -1.90043356e-02 -6.97857067e-02
-2.35111237e-01 -2.32197762e-01 2.74615228e-01 4.97556657e-01
-1.36242047e-01 -9.26578715e-02 -6.49276137e-01 3.64436448e-01
-2.29444456e+00 -9.67328250e-01 8.64776447e-02 2.39667106e+00
5.46372652e-01 -2.48726141e-02 3.74452621e-01 -2.97596037e-01
1.13868892e+00 5.43676224e-03 -8.43148530e-01 5.11754036e-01
-3.45896363e-01 -1.48881525e-01 4.04632121e-01 4.18401122e-01
-1.38357782e+00 8.50447416e-01 5.82311058e+00 8.42562497e-01
-9.08773780e-01 3.70927811e-01 7.25284874e-01 -3.21206212e-01
3.25511098e-01 -3.11484337e-01 -1.34987247e+00 8.60464454e-01
5.53415477e-01 2.65233722e-02 5.51216781e-01 7.82955766e-01
2.74781268e-02 2.24236492e-02 -1.05447972e+00 1.47667408e+00
2.48413056e-01 -1.03564787e+00 3.05456109e-02 -1.20641723e-01
3.71860683e-01 -1.32740214e-01 2.17404664e-01 5.36499918e-01
5.91457710e-02 -7.89455354e-01 8.50928426e-01 6.92014337e-01
7.54615486e-01 -6.88843071e-01 8.25621128e-01 4.01498675e-01
-1.66387641e+00 -2.94081569e-01 -3.68730456e-01 2.18529120e-01
2.45443270e-01 2.93052465e-01 -2.39897653e-01 5.67978203e-01
8.83163452e-01 6.72178984e-01 -8.48104835e-01 1.14313948e+00
1.80713326e-01 2.05871210e-01 -3.63621652e-01 2.54322261e-01
-8.56905058e-02 -6.38685897e-02 2.34805062e-01 1.24724126e+00
1.30966499e-01 -1.11593291e-01 4.22231197e-01 8.72291148e-01
-5.70497438e-02 -1.30412281e-01 -2.32322644e-02 3.25606167e-01
7.85790503e-01 1.20902276e+00 -4.22319353e-01 -4.63221878e-01
-6.10716343e-01 1.18446517e+00 5.02674520e-01 5.08352041e-01
-9.35031235e-01 1.07095681e-01 5.25911570e-01 -2.21693411e-01
1.48676202e-01 7.65743405e-02 9.24879033e-03 -1.35730267e+00
3.80562097e-01 -7.98405707e-01 6.06040061e-01 -3.18092287e-01
-1.40699756e+00 5.70809126e-01 -2.37652957e-01 -1.15114534e+00
-6.85579628e-02 -3.94538969e-01 -3.97148609e-01 1.18204165e+00
-1.63277721e+00 -1.39259994e+00 -6.55900478e-01 7.77499676e-01
5.73710561e-01 -4.72908080e-01 6.51077211e-01 9.62398887e-01
-1.26863825e+00 1.14834917e+00 -1.96100160e-01 3.90949607e-01
9.41795409e-01 -8.51271272e-01 2.86737740e-01 1.18609250e+00
-1.70409992e-01 8.05732191e-01 1.13137655e-01 -6.63068175e-01
-1.29706335e+00 -1.29561734e+00 8.83999348e-01 -3.55669171e-01
1.53156281e-01 -3.08168411e-01 -8.71639907e-01 5.47825575e-01
-2.65517801e-01 2.64823616e-01 6.51554942e-01 -1.76202115e-02
-4.44417179e-01 -2.38782406e-01 -1.08513498e+00 3.95970941e-01
1.39780307e+00 -5.66228271e-01 -1.43791854e-01 1.73956186e-01
5.00422001e-01 -3.07990164e-01 -5.17780840e-01 3.14321309e-01
8.38462174e-01 -8.83997500e-01 1.34427834e+00 -5.37087381e-01
-3.32985371e-01 -5.45810997e-01 -8.29894319e-02 -6.05627894e-01
-6.12550497e-01 -3.98631901e-01 -4.78602529e-01 1.53698969e+00
1.61972880e-01 -6.94721639e-01 8.82274628e-01 1.07124972e+00
2.87076056e-01 -5.77378154e-01 -9.64913249e-01 -8.42188418e-01
-5.30559659e-01 4.45959484e-03 6.67982399e-01 6.55175149e-01
-3.73337746e-01 2.31525555e-01 -8.38779211e-01 4.17907625e-01
9.10732210e-01 -1.67279065e-01 7.48286366e-01 -1.13781917e+00
-2.97509581e-01 -3.82864237e-01 -2.62702793e-01 -1.40890062e+00
1.72001317e-01 -6.17347956e-01 -3.62109803e-02 -9.96525526e-01
7.47286022e-01 -4.48673695e-01 -5.97859919e-01 4.31729674e-01
-5.93466520e-01 1.12011284e-01 1.50095209e-01 5.92077732e-01
-1.19435537e+00 4.66414869e-01 6.86471939e-01 -2.17641398e-01
-2.19476774e-01 5.25421053e-02 -5.86008370e-01 4.41076815e-01
4.08066601e-01 -1.19708702e-01 -2.70186156e-01 -6.72845185e-01
-2.85234928e-01 -2.59622782e-01 6.89553678e-01 -1.12198007e+00
8.93373489e-01 1.46768346e-01 7.31044769e-01 -7.55431235e-01
3.95430535e-01 -6.22358501e-01 2.41600692e-01 4.63135511e-01
-1.64339960e-01 2.16286212e-01 -6.55262992e-02 6.63736522e-01
-3.19625109e-01 2.34698392e-02 7.06031501e-01 -4.25427407e-02
-1.19352102e+00 5.96204042e-01 1.21563204e-01 -1.96516171e-01
9.80216384e-01 -4.19826418e-01 -4.71920252e-01 -9.19877812e-02
-6.10235035e-01 6.97855115e-01 2.55558401e-01 4.98581380e-01
6.25072300e-01 -1.47816157e+00 -6.45246267e-01 3.84950042e-01
1.08809650e-01 -1.60263274e-02 7.36052513e-01 8.20628047e-01
-8.33175033e-02 4.76310790e-01 -1.01290047e-01 -5.55691719e-01
-1.33324718e+00 5.51905453e-01 4.27985877e-01 -3.95371616e-01
-2.54970253e-01 1.01648390e+00 3.04948360e-01 -4.94716674e-01
6.26835167e-01 5.11709094e-01 -3.93972188e-01 9.77223814e-02
9.62610364e-01 6.62477076e-01 -1.57380283e-01 -8.20869386e-01
-6.62218690e-01 4.60409164e-01 -3.40954125e-01 1.50114715e-01
9.31831062e-01 -5.01329184e-01 -1.95685681e-02 -1.68263644e-01
9.40496027e-01 -4.55366969e-01 -1.42750382e+00 -6.44443035e-01
3.27254348e-02 -3.86256665e-01 -2.23154798e-01 -6.46991909e-01
-9.75506306e-01 4.63690132e-01 1.09268510e+00 -3.39700580e-01
1.09354043e+00 -1.79457635e-01 8.41579080e-01 3.97225171e-01
3.95519078e-01 -1.24221265e+00 -7.35827908e-02 3.51928264e-01
3.44596118e-01 -1.46275043e+00 -1.65780067e-01 -4.05880123e-01
-3.91833812e-01 7.99804151e-01 1.09340656e+00 3.22807878e-01
5.30229926e-01 2.50675622e-02 -1.71765443e-02 2.08982855e-01
-4.81437743e-01 -4.10171270e-01 4.45287555e-01 5.53637743e-01
-2.94242203e-02 -8.21057633e-02 -9.83853862e-02 1.16489577e+00
3.20352048e-01 1.78274229e-01 -3.43411535e-01 7.04047620e-01
-2.87585855e-01 -9.54185128e-01 -5.55437803e-01 3.27665895e-01
-2.03031436e-01 -7.80595168e-02 -1.91183075e-01 3.95188689e-01
5.25329828e-01 1.03643978e+00 2.48025786e-02 -6.88462257e-01
2.29744852e-01 1.52417362e-01 9.69370827e-02 -3.87862355e-01
-5.46837986e-01 -4.43666875e-02 -1.07363395e-01 -5.31941652e-01
-4.60064292e-01 -5.55879176e-01 -8.30488145e-01 -3.61925691e-01
-4.64554906e-01 -1.97845027e-02 1.62847593e-01 1.03593588e+00
6.41055644e-01 2.28684768e-01 5.20555317e-01 -7.39777625e-01
-7.44142890e-01 -8.98118913e-01 -5.36195934e-01 5.82242846e-01
1.52508035e-01 -8.23705375e-01 -6.09382130e-02 2.88771633e-02] | [14.799138069152832, 0.8417012691497803] |
ceead40d-3d84-474d-8036-2b5b6cedb388 | pointclustering-unsupervised-point-cloud-pre | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Long_PointClustering_Unsupervised_Point_Cloud_Pre-Training_Using_Transformation_Invariance_in_Clustering_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Long_PointClustering_Unsupervised_Point_Cloud_Pre-Training_Using_Transformation_Invariance_in_Clustering_CVPR_2023_paper.pdf | PointClustering: Unsupervised Point Cloud Pre-Training Using Transformation Invariance in Clustering | Feature invariance under different data transformations, i.e., transformation invariance, can be regarded as a type of self-supervision for representation learning. In this paper, we present PointClustering, a new unsupervised representation learning scheme that leverages transformation invariance for point cloud pre-training. PointClustering formulates the pretext task as deep clustering and employs transformation invariance as an inductive bias, following the philosophy that common point cloud transformation will not change the geometric properties and semantics. Technically, PointClustering iteratively optimizes the feature clusters and backbone, and delves into the transformation invariance as learning regularization from two perspectives: point level and instance level. Point-level invariance learning maintains local geometric properties through gathering point features of one instance across transformations, while instance-level invariance learning further measures clusters over the entire dataset to explore semantics of instances. Our PointClustering is architecture-agnostic and readily applicable to MLP-based, CNN-based and Transformer-based backbones. We empirically demonstrate that the models pre-learnt on the ScanNet dataset by PointClustering provide superior performances on six benchmarks, across downstream tasks of classification and segmentation. More remarkably, PointClustering achieves an accuracy of 94.5% on ModelNet40 with Transformer backbone. Source code is available at https://github.com/FuchenUSTC/PointClustering. | ['Tao Mei', 'Lusong Li', 'Zhaofan Qiu', 'Ting Yao', 'Fuchen Long'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['point-cloud-pre-training', 'deep-clustering', 'philosophy', 'deep-clustering'] | ['computer-vision', 'miscellaneous', 'miscellaneous', 'natural-language-processing'] | [-1.49326427e-02 1.37540996e-01 -5.14619648e-01 -7.31557786e-01
-7.61910617e-01 -7.22335994e-01 5.77674747e-01 3.05895954e-01
-1.05806932e-01 -1.09133795e-01 -7.30221672e-03 -2.69183159e-01
-2.34095857e-01 -8.56917500e-01 -1.23666608e+00 -7.06954181e-01
-8.40636566e-02 5.27059376e-01 2.56715357e-01 -1.52927473e-01
2.43396148e-01 1.01676738e+00 -1.42802608e+00 4.41149712e-01
7.85364270e-01 9.77722824e-01 -4.50780913e-02 2.84646243e-01
-1.61853239e-01 3.34173560e-01 -1.63651332e-01 -2.88628917e-02
5.07233918e-01 2.66013056e-01 -8.68784785e-01 2.15373755e-01
6.33554220e-01 1.62542224e-01 -3.77161264e-01 8.94027829e-01
9.47091877e-02 3.28340709e-01 8.48663330e-01 -1.34199297e+00
-9.94289219e-01 4.90129620e-01 -7.55433917e-01 1.42265871e-01
-1.99992612e-01 2.55519241e-01 1.19082761e+00 -1.10100305e+00
3.99987280e-01 1.25228727e+00 8.06629777e-01 2.78653324e-01
-1.32018661e+00 -6.30005002e-01 3.25808108e-01 5.28848134e-02
-1.38765943e+00 -1.34499103e-01 8.82335365e-01 -5.36138475e-01
1.01116788e+00 2.77385533e-01 4.82867628e-01 8.76991391e-01
2.90002316e-01 8.07352006e-01 7.15289176e-01 -2.15123385e-01
1.64357588e-01 -1.87946215e-01 3.44513655e-01 5.57209015e-01
6.20986968e-02 8.14158916e-02 -2.55500346e-01 9.95453969e-02
1.01693797e+00 5.74424982e-01 -4.10798043e-02 -9.12838340e-01
-1.36412644e+00 9.26001668e-01 1.20302367e+00 1.22838065e-01
-1.29872665e-01 4.44563597e-01 4.70710754e-01 4.28947955e-01
4.26883072e-01 2.95619905e-01 -7.14217484e-01 1.95644423e-01
-6.56077564e-01 -9.43951011e-02 2.58045942e-01 1.47965109e+00
1.15367067e+00 1.09541319e-01 -1.11954391e-01 8.73541236e-01
3.85294765e-01 5.68728268e-01 5.38945913e-01 -8.29239905e-01
4.59346116e-01 9.70048130e-01 -6.37447834e-01 -1.10596859e+00
-5.47379434e-01 -7.46087551e-01 -1.03978360e+00 2.00203955e-01
-1.85034983e-02 2.40534633e-01 -1.27916741e+00 1.69979000e+00
2.40062490e-01 3.55104506e-01 -1.40119657e-01 7.12405920e-01
9.31474209e-01 5.47423184e-01 4.72123660e-02 3.89314294e-01
1.15132272e+00 -9.37772870e-01 -4.81989011e-02 7.61170015e-02
7.65833080e-01 -5.67335784e-01 1.47593868e+00 9.78528336e-02
-9.06057775e-01 -7.24245846e-01 -9.50984478e-01 -1.39352456e-01
-5.45917273e-01 1.96247129e-03 7.30754137e-01 2.68515229e-01
-1.13815820e+00 6.29214168e-01 -1.24017191e+00 -6.05385959e-01
8.38971794e-01 7.10053444e-01 -5.15989304e-01 -7.27568492e-02
-5.03807783e-01 4.56910014e-01 3.19043040e-01 -2.56019741e-01
-6.11012995e-01 -1.21936536e+00 -1.00179219e+00 9.27893296e-02
-1.70273427e-02 -7.15895355e-01 9.61168170e-01 -8.71679723e-01
-1.31627834e+00 1.13230777e+00 -1.14053205e-01 -4.00844067e-01
2.23371878e-01 -7.63313621e-02 -2.36461192e-01 7.34550357e-02
5.10971427e-01 8.43291879e-01 1.04980266e+00 -1.29637718e+00
-5.66635430e-01 -6.11822128e-01 -2.09671885e-01 2.98949536e-02
-2.21503198e-01 -4.17785585e-01 -7.22833693e-01 -7.09269345e-01
6.89987600e-01 -1.13503695e+00 -2.15346813e-01 9.27959308e-02
-6.54357970e-01 -3.97116333e-01 1.08235788e+00 -3.72300483e-02
5.62733889e-01 -2.43148589e+00 1.50147863e-02 6.43486321e-01
2.35070735e-01 -8.04829001e-02 -2.27513239e-01 1.48284271e-01
-6.03723049e-01 2.11033061e-01 -5.01659989e-01 -2.15851963e-01
1.25355586e-01 3.98482740e-01 -4.72639561e-01 7.51571238e-01
4.68811601e-01 1.22157836e+00 -6.96309865e-01 -2.99344987e-01
6.48820341e-01 4.36644256e-01 -8.19773555e-01 -9.22109783e-02
-8.68988857e-02 5.92498362e-01 -4.86022502e-01 8.57851684e-01
8.92193317e-01 -3.81213874e-01 -5.16337097e-01 -3.65577102e-01
-1.00335330e-01 -1.50233228e-02 -7.26707816e-01 2.20995069e+00
-4.06367660e-01 4.91703361e-01 -3.00256550e-01 -1.25519121e+00
1.04882526e+00 -8.91941711e-02 8.69220734e-01 -6.84331536e-01
-7.91864097e-03 8.38997290e-02 -1.62338302e-01 -1.83301404e-01
2.17888996e-01 9.24741924e-02 -2.36198559e-01 3.20211232e-01
3.28437686e-01 -2.40048617e-01 -3.25032115e-01 1.05474249e-01
1.09906530e+00 2.37762868e-01 1.22436211e-02 -4.54682469e-01
3.11227262e-01 1.46727383e-01 4.00999874e-01 6.29542172e-01
-7.79207982e-03 8.78337443e-01 9.43878293e-02 -4.50122476e-01
-9.61796403e-01 -1.57619250e+00 -5.57646632e-01 1.35752475e+00
2.89314359e-01 -3.70544910e-01 -4.56818074e-01 -6.31087840e-01
3.80748630e-01 6.07195735e-01 -8.02916586e-01 -5.20994425e-01
-5.41075706e-01 -4.89775181e-01 4.32298034e-01 8.30130696e-01
4.88862425e-01 -8.71523321e-01 -2.36120582e-01 -1.07905731e-01
3.31786066e-01 -9.12634313e-01 -4.91233170e-01 4.30938840e-01
-1.19399619e+00 -9.00285721e-01 -2.60356963e-01 -8.52599680e-01
8.76707554e-01 3.62964928e-01 1.17541647e+00 1.81039810e-01
-1.93224862e-01 7.43523419e-01 -4.21946704e-01 -3.71399879e-01
6.86980039e-02 6.17147148e-01 -1.29083497e-02 4.40280735e-02
4.39783812e-01 -1.02057874e+00 -6.70518100e-01 3.48131806e-01
-9.22754169e-01 -3.22013050e-01 6.31607652e-01 5.22517562e-01
1.17338610e+00 -2.48095974e-01 2.41781816e-01 -9.90521729e-01
1.47413552e-01 -5.99733293e-01 -2.70404935e-01 -1.28056426e-02
-3.63447100e-01 3.48990560e-02 6.04305625e-01 -1.43776849e-01
-6.26626670e-01 3.86811256e-01 -8.41857642e-02 -9.41028118e-01
-4.34579402e-01 2.78391868e-01 -3.95090342e-01 -2.64966667e-01
8.17302108e-01 2.67399907e-01 -3.66829266e-03 -4.64914292e-01
7.98086584e-01 2.52473027e-01 7.75438607e-01 -9.13673222e-01
1.32964599e+00 9.96863663e-01 7.47891143e-02 -8.21407914e-01
-6.89437091e-01 -7.86908329e-01 -1.28835356e+00 2.57681813e-02
8.39865208e-01 -8.52831960e-01 -5.88945627e-01 -4.06327322e-02
-9.21058118e-01 -3.50769550e-01 -8.23984981e-01 2.02848688e-01
-8.63973618e-01 1.28701046e-01 -2.44442597e-01 1.79261286e-02
-2.96512038e-01 -9.93618071e-01 1.30639362e+00 -8.16701204e-02
-3.75173539e-02 -1.23191619e+00 -8.77676010e-02 -1.06364027e-01
1.58839822e-01 6.46081984e-01 1.05299759e+00 -7.77448118e-01
-7.69625723e-01 -2.43779257e-01 -2.52702355e-01 2.92112768e-01
3.49407643e-01 5.96063845e-02 -9.69636142e-01 -5.73052168e-01
-6.59718066e-02 -9.43992883e-02 9.31895912e-01 4.67234015e-01
1.75023878e+00 -7.29467124e-02 -5.25625944e-01 1.51098979e+00
1.40141964e+00 -2.05980748e-01 6.08360887e-01 5.74280024e-01
1.26872790e+00 3.69179666e-01 3.30984741e-01 1.65489256e-01
3.51700097e-01 5.55426836e-01 6.88995004e-01 -4.04170990e-01
-1.41066313e-01 -3.36598366e-01 3.21520455e-02 8.69201779e-01
-1.31526172e-01 1.85751051e-01 -1.01236045e+00 4.66991752e-01
-1.90414631e+00 -6.60479724e-01 -2.44817048e-01 1.92046082e+00
3.15708339e-01 -1.53400619e-02 -1.45272203e-02 3.94768789e-02
7.15296566e-01 3.06401122e-02 -9.01818454e-01 -2.90005565e-01
-1.22047178e-01 3.61529022e-01 7.24961221e-01 1.83567196e-01
-1.36033702e+00 1.25006425e+00 5.46003056e+00 7.06912041e-01
-1.30931997e+00 1.53062463e-01 4.50751871e-01 3.90933715e-02
-3.29201072e-01 -5.25140390e-02 -4.30217922e-01 2.73996919e-01
5.05482793e-01 -4.27197218e-02 5.81230409e-02 1.07210767e+00
1.11203805e-01 8.43611896e-01 -1.42027569e+00 9.47431028e-01
-1.35290146e-01 -1.45256507e+00 4.35215056e-01 5.66874631e-02
7.30111361e-01 6.36561513e-01 5.30753314e-01 4.44828957e-01
4.00703341e-01 -1.12302041e+00 6.96179032e-01 4.60676134e-01
9.65150118e-01 -8.21068048e-01 3.75346541e-01 1.01079322e-01
-1.35557914e+00 1.03148200e-01 -6.62911892e-01 2.79412091e-01
-2.19270006e-01 3.06858480e-01 -7.09460378e-01 6.36907935e-01
1.04246235e+00 1.46115494e+00 -7.67419040e-01 9.74339783e-01
-1.30038783e-01 6.93374515e-01 -5.62569797e-01 6.48283720e-01
4.91890937e-01 -2.29354620e-01 5.46456158e-01 1.36809003e+00
1.47190824e-01 -1.30592063e-01 2.39571586e-01 1.06809771e+00
-1.50641531e-01 1.12205006e-01 -6.63458049e-01 5.55012882e-01
4.86956209e-01 1.18934095e+00 -9.18602288e-01 -2.56994832e-02
-3.50440711e-01 8.15181673e-01 4.78368998e-01 4.72548097e-01
-7.95896351e-01 -2.78121293e-01 8.05979133e-01 2.53552258e-01
5.54348469e-01 -2.83391804e-01 -7.17759848e-01 -1.17060697e+00
-3.08907554e-02 -4.87598747e-01 5.57937264e-01 -4.64309871e-01
-1.49932027e+00 4.93313879e-01 9.45914686e-02 -1.64616835e+00
1.88060790e-01 -8.47009063e-01 -1.05489790e+00 4.01734799e-01
-1.65654981e+00 -1.64127207e+00 -4.83478934e-01 1.08626628e+00
5.83297968e-01 -3.36923748e-01 6.13859296e-01 8.20663124e-02
-5.10116577e-01 9.54622149e-01 2.93970406e-01 2.92370498e-01
4.74465400e-01 -1.41897345e+00 6.94508672e-01 6.55086100e-01
3.62710178e-01 8.76098275e-01 7.98398331e-02 -3.66820395e-01
-1.44808125e+00 -1.82393038e+00 1.59548998e-01 -8.36680532e-01
6.56307280e-01 -4.61202651e-01 -1.22385490e+00 1.12388229e+00
-2.49053016e-01 4.67227668e-01 6.31481409e-01 1.64996013e-01
-6.31548345e-01 -2.62909591e-01 -1.01536918e+00 4.41958249e-01
1.33386362e+00 -6.09391928e-01 -6.23871326e-01 6.63988173e-01
1.07915974e+00 -5.52416384e-01 -1.23925054e+00 5.74758768e-01
4.27848585e-02 -7.29750574e-01 1.31045568e+00 -6.92779481e-01
1.57563657e-01 -2.74999619e-01 -4.08937633e-01 -1.27305365e+00
-8.64962995e-01 -4.02351379e-01 1.00225717e-01 1.14998639e+00
5.01700044e-01 -8.35308433e-01 9.74562466e-01 3.17212641e-01
-7.29623795e-01 -7.41637945e-01 -9.58652794e-01 -9.47975695e-01
6.80272281e-01 -7.46222019e-01 1.03744650e+00 1.21377337e+00
-3.64705473e-01 1.89783588e-01 4.35738176e-01 5.29460073e-01
6.51542127e-01 1.77092671e-01 9.24743116e-01 -1.37103224e+00
1.73776433e-01 -5.78129470e-01 -8.74532819e-01 -1.09971046e+00
4.87316519e-01 -1.64025652e+00 -1.98119670e-01 -1.42531812e+00
-1.62956178e-01 -7.59257197e-01 -5.50144732e-01 7.32771218e-01
3.13189656e-01 2.81867802e-01 2.11586937e-01 8.97849500e-01
-6.08147323e-01 6.33243859e-01 1.26153100e+00 -4.69026715e-01
-3.07909369e-01 1.87853262e-01 -7.35270381e-01 8.04368734e-01
8.96466553e-01 -4.48104799e-01 -5.41427255e-01 -7.08397388e-01
-1.87889189e-01 -7.24582314e-01 5.31183541e-01 -1.11566806e+00
3.54479641e-01 -1.01327039e-01 4.42517281e-01 -8.00798297e-01
2.03955621e-01 -1.08775234e+00 -1.17620900e-01 2.59429842e-01
-4.13512021e-01 1.72865376e-01 4.01091099e-01 6.78684533e-01
-1.13419272e-01 2.68831611e-01 8.64736259e-01 -4.90574203e-02
-8.79635215e-01 8.74559045e-01 1.71440765e-01 8.47556740e-02
9.89549637e-01 -5.64147413e-01 -4.95560989e-02 8.24262053e-02
-8.32529485e-01 2.31002182e-01 5.99555016e-01 6.28628969e-01
7.90616333e-01 -1.54492581e+00 -6.19922340e-01 5.44326663e-01
5.22170961e-01 6.74025834e-01 3.73108648e-02 7.15298474e-01
-5.58067143e-01 3.69303077e-01 -2.23754466e-01 -1.51835585e+00
-7.53143668e-01 5.89967906e-01 3.64420354e-01 2.78066397e-01
-1.13231742e+00 8.01845610e-01 7.19523430e-01 -1.06852639e+00
1.56873032e-01 -9.01902676e-01 3.45991105e-02 -4.41100627e-01
1.50671871e-02 8.55452791e-02 3.43598336e-01 -9.00648952e-01
-4.83652085e-01 1.09452093e+00 -1.91747800e-01 3.37873697e-01
1.47836745e+00 1.04223199e-01 -2.12595642e-01 4.92474198e-01
1.65929484e+00 -5.16599596e-01 -1.33899677e+00 -4.42043513e-01
1.46797925e-01 -3.87992471e-01 5.75263388e-02 -3.12416941e-01
-1.34552121e+00 8.95680428e-01 5.93330204e-01 8.24743882e-03
9.26037490e-01 4.02120829e-01 4.84491378e-01 5.85314214e-01
2.13066474e-01 -7.95671821e-01 6.01374730e-02 6.73537076e-01
1.11728632e+00 -1.26719201e+00 -8.51485059e-02 -5.53763330e-01
-4.48179305e-01 1.02761233e+00 7.79430687e-01 -5.95263660e-01
1.05406320e+00 -5.41375726e-02 -1.44535571e-01 -6.36515856e-01
-4.52744424e-01 -8.80495235e-02 6.20837867e-01 9.54607964e-01
1.84564427e-01 2.58775622e-01 5.62240601e-01 4.62655365e-01
-6.28796816e-01 -5.37171483e-01 -1.09466970e-01 7.85810947e-01
-3.98539662e-01 -7.30623901e-01 -3.94258082e-01 5.37067354e-01
1.31433159e-01 9.64128822e-02 -4.00575072e-01 1.10923493e+00
2.76825488e-01 4.83834267e-01 5.80423355e-01 -4.98504251e-01
7.89823592e-01 -3.11598450e-01 2.38511845e-01 -9.99590456e-01
-4.56670463e-01 -1.32109791e-01 -7.24256814e-01 -7.89773703e-01
-3.06519628e-01 -6.09676480e-01 -1.68343759e+00 -3.04506451e-01
-3.07748914e-02 -1.56371463e-02 5.18279791e-01 7.83393323e-01
6.57931924e-01 5.55051208e-01 8.49608064e-01 -1.05658603e+00
-3.66384745e-01 -7.58191228e-01 -4.21218455e-01 7.41786242e-01
3.31781626e-01 -6.99346483e-01 -3.28377098e-01 1.50468618e-01] | [7.9609527587890625, -3.535647392272949] |
b2110c2c-82ef-41e9-8d6d-cb6e80096c62 | unsupervised-video-domain-adaptation-a | 2208.07365 | null | https://arxiv.org/abs/2208.07365v2 | https://arxiv.org/pdf/2208.07365v2.pdf | Unsupervised Video Domain Adaptation for Action Recognition: A Disentanglement Perspective | Unsupervised video domain adaptation is a practical yet challenging task. In this work, for the first time, we tackle it from a disentanglement view. Our key idea is to handle the spatial and temporal domain divergence separately through disentanglement. Specifically, we consider the generation of cross-domain videos from two sets of latent factors, one encoding the static information and another encoding the dynamic information. A Transfer Sequential VAE (TranSVAE) framework is then developed to model such generation. To better serve for adaptation, we propose several objectives to constrain the latent factors. With these constraints, the spatial divergence can be readily removed by disentangling the static domain-specific information out, and the temporal divergence is further reduced from both frame- and video-levels through adversarial learning. Extensive experiments on the UCF-HMDB, Jester, and Epic-Kitchens datasets verify the effectiveness and superiority of TranSVAE compared with several state-of-the-art methods. The code with reproducible results is publicly accessible. | ['Yi Ren', 'Jing Jiang', 'Zhiqiang Xu', 'Xiang Yin', 'Xinghua Qu', 'Lingdong Kong', 'Pengfei Wei'] | 2022-08-15 | null | null | null | null | ['video-domain-adapation'] | ['computer-vision'] | [ 2.50863343e-01 -2.82008886e-01 -1.61961243e-01 -9.69393030e-02
-8.19229543e-01 -7.20913112e-01 7.56310999e-01 -5.05743504e-01
-2.95444101e-01 7.93058932e-01 2.55438060e-01 8.20269585e-02
-7.50272572e-02 -3.37764174e-01 -6.68410003e-01 -1.04707778e+00
5.18121310e-02 1.37199640e-01 2.27400675e-01 -5.67627400e-02
-2.20812365e-01 1.57218784e-01 -1.25935733e+00 1.17060028e-01
9.67070758e-01 9.07666206e-01 1.82355911e-01 4.76771861e-01
2.96864629e-01 7.69247234e-01 -4.14868742e-01 -3.97613883e-01
3.70196134e-01 -5.65717101e-01 -2.40010574e-01 2.85618007e-01
2.48544142e-01 -4.87801611e-01 -6.22043073e-01 1.08330178e+00
5.18724263e-01 2.25206971e-01 5.63199759e-01 -1.32154834e+00
-5.70851624e-01 7.85027370e-02 -4.99732047e-01 2.39612460e-02
2.56762326e-01 2.06205010e-01 8.71406078e-01 -7.45938003e-01
6.92264974e-01 1.15924335e+00 2.54982322e-01 6.10635757e-01
-1.38439941e+00 -8.67949843e-01 4.15876478e-01 3.10952425e-01
-1.33621907e+00 -5.38887918e-01 1.00930715e+00 -7.59011745e-01
3.92028749e-01 6.16460480e-02 4.99141335e-01 1.61590922e+00
-1.10010810e-01 9.00564611e-01 1.02068269e+00 -2.42619559e-01
1.89761773e-01 -1.01316400e-01 -5.36921144e-01 5.09000957e-01
-1.64988875e-01 3.57934445e-01 -5.68032563e-01 -9.51857865e-02
8.53569210e-01 -3.17283534e-02 -6.01176679e-01 -8.23474586e-01
-1.29675031e+00 7.96090305e-01 5.00590950e-02 1.59486309e-02
-1.38109937e-01 -1.88280314e-01 5.05903363e-01 2.58780241e-01
5.09723127e-01 7.34028593e-02 -4.63860750e-01 -1.50470257e-01
-8.44824374e-01 3.49507958e-01 5.40257454e-01 1.13503242e+00
4.93239582e-01 3.02243948e-01 -3.69785011e-01 8.96649063e-01
3.14200282e-01 4.96618330e-01 5.86635351e-01 -7.96100676e-01
6.66778982e-01 1.41793385e-01 1.89443707e-01 -1.07644892e+00
8.29633027e-02 -1.94646955e-01 -9.68045950e-01 1.21357456e-01
4.19712007e-01 -2.80775666e-01 -9.58600760e-01 2.13282299e+00
3.70808274e-01 7.17736363e-01 9.62919742e-02 1.02123058e+00
5.51125646e-01 8.33270729e-01 6.78155646e-02 -4.72781092e-01
1.12730455e+00 -1.08831775e+00 -9.28372622e-01 -2.01966465e-01
1.11740351e-01 -7.21504450e-01 7.83839643e-01 3.03286970e-01
-9.25259531e-01 -7.71748066e-01 -1.18222702e+00 2.25773547e-02
-7.23817339e-03 2.94776797e-01 1.96825609e-01 2.98787534e-01
-5.79112709e-01 2.92023987e-01 -1.05812442e+00 -8.35494883e-03
2.10377231e-01 1.96472242e-01 -3.93054128e-01 -1.08205765e-01
-1.36749351e+00 4.93118793e-01 4.30566818e-01 -2.71898154e-02
-1.07068694e+00 -5.23753762e-01 -9.84073460e-01 -7.56333098e-02
6.25650048e-01 -6.92930102e-01 1.06574416e+00 -1.07378495e+00
-1.77015316e+00 6.02183402e-01 -4.06842530e-02 -2.96701670e-01
8.13619792e-01 -2.60195374e-01 -4.91965115e-01 2.90262610e-01
9.75303501e-02 4.31250572e-01 1.29659402e+00 -1.24361598e+00
-4.75371003e-01 -2.58152872e-01 2.05451287e-02 3.76468122e-01
-4.08624560e-01 -1.38353989e-01 -8.73601317e-01 -1.17426121e+00
-2.13727638e-01 -1.05246580e+00 1.13007471e-01 8.65586922e-02
-1.47234991e-01 2.00413510e-01 8.80923331e-01 -1.02313614e+00
1.07327640e+00 -2.46624088e+00 8.97070348e-01 -2.71109164e-01
1.55226424e-01 4.08038199e-01 -2.91668296e-01 1.91817254e-01
-8.41222852e-02 -2.10540354e-01 -3.74239832e-01 -5.08639514e-01
-9.32364259e-03 3.00097376e-01 -4.16696429e-01 4.64766115e-01
2.18495145e-01 6.66332781e-01 -1.05076933e+00 -3.79479289e-01
1.65134981e-01 5.95149994e-01 -6.43734992e-01 5.44870198e-01
-2.89926529e-01 9.68025088e-01 -4.35537398e-01 3.95067155e-01
7.88242877e-01 7.24670067e-02 3.44461858e-01 -2.73409069e-01
1.01722308e-01 4.31303829e-02 -1.09417665e+00 1.93990755e+00
-2.18591020e-01 4.41115022e-01 2.51767963e-01 -1.08742356e+00
6.67243600e-01 5.73758602e-01 5.42220771e-01 -7.01201022e-01
-2.02038232e-02 8.06772485e-02 -3.23691145e-02 -5.91551125e-01
2.73305506e-01 -2.97028244e-01 -1.72150880e-01 5.89039847e-02
4.64515060e-01 1.40021488e-01 1.25374958e-01 3.25939655e-02
7.77330756e-01 6.28262758e-01 2.98088789e-01 -3.19706276e-02
6.47345126e-01 -3.88187230e-01 1.06979156e+00 3.05389941e-01
-3.76721382e-01 8.13799679e-01 5.85837603e-01 -8.62157270e-02
-9.91568506e-01 -1.36538911e+00 1.87530041e-01 8.16848218e-01
3.00502121e-01 -3.56529355e-01 -6.53073251e-01 -8.88373196e-01
-4.65489514e-02 4.42943811e-01 -5.61463058e-01 -2.69565403e-01
-6.36602521e-01 -4.41215396e-01 4.41456914e-01 4.87307101e-01
5.10997355e-01 -7.67197430e-01 -2.05103904e-01 1.16052806e-01
-4.79327857e-01 -1.43048561e+00 -8.02558839e-01 -1.12250827e-01
-5.97284257e-01 -8.34036052e-01 -1.04602969e+00 -5.88562548e-01
3.70114744e-01 1.90852046e-01 8.11689973e-01 -3.66707027e-01
-5.41821606e-02 1.80715710e-01 -4.71619189e-01 -5.27251102e-02
-3.50732744e-01 7.50890176e-04 1.80009618e-01 4.54156965e-01
3.55692231e-03 -7.95594990e-01 -5.09780467e-01 4.18015838e-01
-1.05217612e+00 2.61731535e-01 4.27310109e-01 1.04886270e+00
6.70575321e-01 3.08552422e-02 5.09855509e-01 -6.96424544e-01
2.90791988e-01 -5.68867207e-01 -6.63963199e-01 1.37842149e-01
-1.01770610e-01 1.78655870e-02 7.90050626e-01 -7.96445847e-01
-1.38303459e+00 1.04917407e-01 2.47310065e-02 -9.08946455e-01
-2.05239877e-01 2.20270544e-01 -8.33708286e-01 3.38508159e-01
2.31035262e-01 4.76270020e-01 -8.68780687e-02 -5.65629601e-01
4.27650839e-01 3.84797812e-01 5.89404583e-01 -7.26466417e-01
1.04592431e+00 6.02252603e-01 -3.03571969e-01 -6.68924868e-01
-6.56399310e-01 -3.49854529e-01 -6.58327579e-01 -1.04059435e-01
1.13280129e+00 -1.20415819e+00 -2.11922929e-01 6.76900446e-01
-1.02352536e+00 -4.36992854e-01 -2.03984201e-01 4.83852476e-01
-8.28060031e-01 5.58787107e-01 -4.97877121e-01 -4.93342668e-01
1.08082801e-01 -1.16969442e+00 1.00226665e+00 1.06979735e-01
9.10667554e-02 -9.87135172e-01 2.57182956e-01 3.78185689e-01
7.92944152e-03 4.87490922e-01 6.92375302e-01 -4.86339480e-01
-6.18028998e-01 3.53789702e-02 -7.07506761e-02 5.77306449e-01
3.27743918e-01 -1.80511177e-01 -7.16957450e-01 -5.41897655e-01
1.89845249e-01 -1.84837505e-01 8.77491951e-01 1.05984211e-01
9.61026013e-01 -4.39372659e-01 -2.11604312e-01 9.29161310e-01
1.27571917e+00 3.45571637e-01 5.51555574e-01 2.80738860e-01
9.23835933e-01 3.84944201e-01 6.57538295e-01 4.79887187e-01
2.53853112e-01 9.74057078e-01 1.87771082e-01 1.39064500e-02
-1.63058594e-01 -3.92041504e-01 5.10698020e-01 9.41409171e-01
-8.80894512e-02 -4.71567899e-01 -5.76894343e-01 5.42579114e-01
-2.07052088e+00 -9.81692076e-01 3.60449314e-01 2.12273216e+00
8.04084241e-01 -3.85963991e-02 2.41141766e-01 -1.03283590e-02
7.14316785e-01 5.33064485e-01 -6.15502536e-01 2.30667382e-01
-1.33741185e-01 -1.63855389e-01 1.52010277e-01 3.62962335e-01
-1.28185308e+00 7.77337611e-01 5.20344210e+00 9.01233077e-01
-1.02057993e+00 1.84230566e-01 2.31975287e-01 -2.49652594e-01
-1.08233437e-01 -2.90880334e-02 -4.90321279e-01 8.32827747e-01
5.07374346e-01 -3.77010331e-02 5.26524305e-01 5.50645590e-01
1.31604478e-01 2.89251626e-01 -1.14188027e+00 8.65924716e-01
1.14464700e-01 -8.70531321e-01 5.79590835e-02 2.34222319e-02
8.27710450e-01 -2.97726959e-01 1.90073028e-01 3.87892574e-01
1.60551280e-01 -6.51297271e-01 9.96280313e-01 6.38283312e-01
9.24798429e-01 -6.32499456e-01 3.98074657e-01 4.23177689e-01
-1.24426699e+00 6.35098070e-02 -8.81544501e-02 1.45197958e-01
3.77448350e-01 3.54602069e-01 -1.24150507e-01 9.36732531e-01
4.81626093e-01 1.01230633e+00 -2.77359009e-01 7.21247494e-01
-4.50471729e-01 4.58665907e-01 -3.40623185e-02 6.60828352e-01
-3.64254490e-02 -3.55625749e-01 8.95142376e-01 1.08446431e+00
3.92620206e-01 2.15249017e-01 2.33617470e-01 8.03366065e-01
1.88542958e-02 -2.25542188e-01 -4.75939214e-01 -6.18205741e-02
3.15354377e-01 1.06169772e+00 -2.92254061e-01 -3.09466422e-01
-6.05418026e-01 1.42987442e+00 2.96597749e-01 6.99175119e-01
-1.10466695e+00 -1.46895364e-01 9.03744340e-01 -1.40897974e-01
6.30224884e-01 -4.41422462e-01 1.31700337e-01 -1.89225221e+00
2.04525054e-01 -1.13185728e+00 3.57792854e-01 -5.86830914e-01
-1.30328023e+00 5.77032864e-01 2.03183591e-01 -1.58693945e+00
-4.45978165e-01 -4.30503190e-01 -2.77727365e-01 8.32749784e-01
-1.50430250e+00 -1.30501652e+00 -2.74431944e-01 8.03112805e-01
6.22672021e-01 -3.84073138e-01 6.13390625e-01 4.37784940e-01
-7.53755093e-01 7.32738853e-01 3.64100575e-01 2.09775582e-01
9.36259866e-01 -9.74074543e-01 1.93277076e-01 1.14948595e+00
7.98052773e-02 3.18896502e-01 5.95404565e-01 -5.33296108e-01
-1.18722022e+00 -1.13808692e+00 4.65079993e-01 -3.27366918e-01
6.70338213e-01 -6.48680210e-01 -1.06721735e+00 6.43731654e-01
1.14526756e-01 9.48432609e-02 6.10524476e-01 -3.51168245e-01
-6.81718886e-01 -1.38751581e-01 -8.43773425e-01 6.63497269e-01
9.78353679e-01 -6.74506664e-01 -5.29772580e-01 -3.38770822e-02
9.15324688e-01 -5.63587308e-01 -7.50218868e-01 5.43506145e-01
6.00983560e-01 -1.04495180e+00 1.08675003e+00 -3.56243312e-01
5.68650484e-01 -4.38760191e-01 -3.08304280e-01 -1.41686058e+00
-2.67878443e-01 -7.45646060e-01 -4.66681182e-01 1.56687462e+00
-1.02736428e-01 -4.83136475e-01 5.19817114e-01 2.73790389e-01
8.71597752e-02 -3.29134494e-01 -9.78958189e-01 -1.07234800e+00
9.57648531e-02 -1.37576073e-01 4.77343678e-01 1.03570211e+00
-2.38737956e-01 4.12620515e-01 -9.20434117e-01 2.38688290e-01
6.19683087e-01 3.13480832e-02 7.84464478e-01 -8.17965329e-01
-7.17874050e-01 -1.76948130e-01 -3.40155602e-01 -1.33158386e+00
3.51100624e-01 -5.89351296e-01 2.49206960e-01 -1.02091634e+00
2.20706761e-01 -8.65536183e-02 -3.48031014e-01 1.71900004e-01
-3.59125525e-01 -5.26308315e-03 5.23698926e-01 3.90060157e-01
-5.10984242e-01 1.05370235e+00 1.41922343e+00 -1.22960910e-01
-1.25640064e-01 -2.26419225e-01 -4.92930502e-01 6.48136258e-01
5.76629221e-01 -3.81217778e-01 -6.62653089e-01 -6.39295280e-01
-3.00871819e-01 3.97504777e-01 4.24775332e-01 -1.01118386e+00
-3.91712226e-02 -2.99229056e-01 1.58624128e-01 -2.70354241e-01
4.97907400e-01 -7.92922199e-01 2.14541197e-01 1.86124548e-01
-1.98776051e-01 -1.78941905e-01 3.10656577e-01 1.03383601e+00
-3.85757655e-01 1.97483152e-01 8.64446402e-01 1.18571088e-01
-6.19809508e-01 5.09999275e-01 -1.88045561e-01 1.59190536e-01
1.03692067e+00 6.14476688e-02 -9.62531194e-02 -4.46329802e-01
-8.48040700e-01 1.36916369e-01 5.09397864e-01 6.34445727e-01
5.39991856e-01 -1.61682475e+00 -7.61078894e-01 5.05017281e-01
2.23196089e-01 -8.80185980e-03 5.54799020e-01 7.40143538e-01
-7.15895221e-02 1.63228929e-01 -2.77976692e-01 -4.96119171e-01
-1.16000295e+00 7.69106507e-01 1.36507571e-01 -4.21753585e-01
-4.80666816e-01 6.36711240e-01 7.29187906e-01 -2.79335052e-01
1.77913427e-01 4.86783534e-02 -8.60905051e-02 7.08906278e-02
4.24294353e-01 2.89106309e-01 -3.03186804e-01 -9.59280670e-01
-3.23307455e-01 5.76221466e-01 -4.54675080e-03 -3.46444845e-01
1.13096607e+00 -3.74794543e-01 2.77458459e-01 3.93143564e-01
1.40187585e+00 6.87991157e-02 -1.80992794e+00 -3.33283067e-01
-3.50254893e-01 -6.94376886e-01 -2.28098124e-01 -6.27028644e-01
-1.10599923e+00 9.50952411e-01 5.59429407e-01 1.02656052e-01
1.47353113e+00 -2.17246667e-01 7.98674107e-01 -2.03650862e-01
1.32055134e-01 -8.96189392e-01 3.01965415e-01 3.55046421e-01
9.92446780e-01 -1.14611292e+00 -2.03560382e-01 -3.76942873e-01
-7.11355209e-01 7.94291317e-01 6.06804311e-01 -2.33525515e-01
4.92752880e-01 2.09225655e-01 -1.26301885e-01 2.79323548e-01
-5.91590285e-01 -1.50720507e-01 5.97885609e-01 7.50953794e-01
1.58837497e-01 4.02863473e-02 -1.94540441e-01 7.97109187e-01
2.04930797e-01 3.78001072e-02 2.20675066e-01 7.64488995e-01
1.09263614e-01 -1.28135264e+00 -2.01795518e-01 -1.73181921e-01
-4.56295431e-01 6.40980452e-02 -1.07911855e-01 8.55309367e-01
3.33473235e-01 6.90793037e-01 -8.84910300e-02 -4.18012291e-01
3.44739705e-01 -3.51979909e-03 4.50836837e-01 -4.11062002e-01
1.44223275e-03 5.36046743e-01 -1.22492902e-01 -4.59981441e-01
-6.21765852e-01 -7.81334817e-01 -7.97037244e-01 -9.76744574e-03
-1.50826946e-01 -3.10217291e-02 1.43374130e-01 8.99305284e-01
4.19316620e-01 6.50293350e-01 6.29366338e-01 -1.00283575e+00
-5.70295155e-01 -7.47371674e-01 -5.83367109e-01 6.98137343e-01
6.08324289e-01 -8.31618607e-01 -4.86100852e-01 5.62839746e-01] | [10.728326797485352, -0.01215418428182602] |
1c0cc511-2a6c-4ba6-a220-87dc75a94053 | mrfalign-protein-homology-detection-through | 1401.2668 | null | http://arxiv.org/abs/1401.2668v2 | http://arxiv.org/pdf/1401.2668v2.pdf | MRFalign: Protein Homology Detection through Alignment of Markov Random Fields | Sequence-based protein homology detection has been extensively studied and so
far the most sensitive method is based upon comparison of protein sequence
profiles, which are derived from multiple sequence alignment (MSA) of sequence
homologs in a protein family. A sequence profile is usually represented as a
position-specific scoring matrix (PSSM) or an HMM (Hidden Markov Model) and
accordingly PSSM-PSSM or HMM-HMM comparison is used for homolog detection. This
paper presents a new homology detection method MRFalign, consisting of three
key components: 1) a Markov Random Fields (MRF) representation of a protein
family; 2) a scoring function measuring similarity of two MRFs; and 3) an
efficient ADMM (Alternating Direction Method of Multipliers) algorithm aligning
two MRFs. Compared to HMM that can only model very short-range residue
correlation, MRFs can model long-range residue interaction pattern and thus,
encode information for the global 3D structure of a protein family.
Consequently, MRF-MRF comparison for remote homology detection shall be much
more sensitive than HMM-HMM or PSSM-PSSM comparison. Experiments confirm that
MRFalign outperforms several popular HMM or PSSM-based methods in terms of both
alignment accuracy and remote homology detection and that MRFalign works
particularly well for mainly beta proteins. For example, tested on the
benchmark SCOP40 (8353 proteins) for homology detection, PSSM-PSSM and HMM-HMM
succeed on 48% and 52% of proteins, respectively, at superfamily level, and on
15% and 27% of proteins, respectively, at fold level. In contrast, MRFalign
succeeds on 57.3% and 42.5% of proteins at superfamily and fold level,
respectively. This study implies that long-range residue interaction patterns
are very helpful for sequence-based homology detection. The software is
available for download at http://raptorx.uchicago.edu/download/. | ['Zhiyong Wang', 'Sheng Wang', 'Jianzhu Ma', 'Jinbo Xu'] | 2014-01-12 | null | null | null | null | ['multiple-sequence-alignment'] | ['medical'] | [ 6.02279186e-01 -2.18213499e-01 -3.13222498e-01 -3.76101166e-01
-6.76540971e-01 -4.65143532e-01 2.84526259e-01 5.85438907e-01
-4.03055668e-01 1.16293490e+00 -3.74799907e-01 -5.04775941e-01
-3.33494097e-02 -4.64451104e-01 -7.11172640e-01 -1.07148337e+00
-1.71872169e-01 5.47805309e-01 6.89897776e-01 -4.35087562e-01
4.57238168e-01 6.88934386e-01 -1.48113549e+00 4.37749565e-01
8.50653887e-01 5.33453107e-01 8.86981666e-01 7.37280011e-01
-1.10334530e-01 1.75101534e-01 -4.04744685e-01 -1.49960205e-01
-2.27997378e-01 -7.58197188e-01 -8.82357538e-01 -2.28587955e-01
4.17198129e-02 3.67660284e-01 4.16427374e-01 9.07289207e-01
3.62676710e-01 -6.56625405e-02 8.46086204e-01 -9.02568340e-01
-1.07283428e-01 -4.39502858e-03 -4.93510246e-01 4.68448587e-02
8.72617602e-01 2.44506255e-01 9.75980043e-01 -9.11771655e-01
9.77345407e-01 1.13184822e+00 6.40858889e-01 1.25036448e-01
-1.46759093e+00 -3.01532179e-01 -3.25643897e-01 5.68702519e-01
-1.27668011e+00 1.82534695e-01 1.42394111e-01 -6.09020293e-01
1.55303466e+00 4.86350864e-01 5.93125224e-01 6.56442285e-01
6.37115538e-01 2.60787219e-01 1.25021744e+00 -4.48778182e-01
2.34468266e-01 -3.10020745e-01 3.09714466e-01 5.11786520e-01
-7.08578452e-02 -1.31464198e-01 -2.78034747e-01 -8.68999779e-01
5.54067671e-01 3.40981334e-02 -3.40269297e-01 -4.44437653e-01
-1.18833077e+00 8.62472296e-01 3.82344685e-02 4.53763872e-01
-6.20041490e-01 -4.89291400e-01 2.84511387e-01 2.85878509e-01
-1.39513269e-01 3.54817212e-01 -5.42875528e-01 -2.16346458e-01
-7.71122932e-01 3.99483740e-01 7.47596681e-01 4.92627680e-01
9.31580305e-01 -5.24997294e-01 5.78379989e-01 8.13215494e-01
2.93953061e-01 3.87980580e-01 6.40098393e-01 -5.84939122e-01
-6.95093125e-02 6.40162587e-01 2.44379133e-01 -9.18265283e-01
-4.88331944e-01 2.19159439e-01 -7.05981314e-01 2.72010982e-01
5.10794222e-01 5.16704500e-01 -7.35853076e-01 1.55687213e+00
4.22686428e-01 -1.83997303e-02 2.70857036e-01 8.29581678e-01
4.90832180e-01 8.59458983e-01 2.01383919e-01 -6.17739379e-01
1.60441756e+00 -5.43480873e-01 -4.57764298e-01 2.45284304e-01
7.26257324e-01 -1.14513791e+00 7.54354179e-01 4.10488784e-01
-1.07925034e+00 -6.23223364e-01 -1.10454893e+00 3.23634744e-01
-3.33135277e-01 -2.53390837e-02 1.19722769e-01 1.55299500e-01
-8.31665277e-01 9.13612306e-01 -9.86368835e-01 -6.02136075e-01
-3.60011786e-01 5.26109159e-01 -7.73890018e-01 2.38943622e-01
-1.24045897e+00 1.28161669e+00 7.08803833e-01 -3.65383714e-01
-3.28472167e-01 -2.45313004e-01 -4.98071223e-01 -1.24604078e-02
-3.07976574e-01 -3.83062810e-01 9.55283582e-01 -7.71458030e-01
-1.26326561e+00 1.02122426e+00 -6.29054904e-01 -4.87233013e-01
1.20259210e-01 1.11367367e-01 -4.38288897e-01 3.66751075e-01
-6.13641515e-02 4.71793324e-01 4.05710936e-01 -8.03814411e-01
-5.88445306e-01 -1.84852555e-01 -4.96587515e-01 4.95031625e-02
4.22694594e-01 5.91120899e-01 2.36135200e-01 -4.61795807e-01
4.41838771e-01 -9.98559296e-01 -5.38915455e-01 -3.88071299e-01
-1.01306222e-01 -3.41974199e-01 5.81455708e-01 -9.77789700e-01
1.15594316e+00 -1.73088968e+00 4.47231263e-01 4.81135041e-01
-1.95785556e-02 6.01643264e-01 -8.93333182e-03 1.12627637e+00
-8.57453167e-01 -1.80549026e-01 -5.56976318e-01 5.40553749e-01
-2.50573933e-01 1.92801386e-01 -9.38648432e-02 6.95001900e-01
2.35757768e-01 4.41746712e-01 -7.92594850e-01 -3.76480937e-01
3.28695446e-01 5.31611741e-01 -7.10556507e-02 1.56700790e-01
-1.19264923e-01 4.39437389e-01 -1.45799622e-01 4.87079799e-01
9.03798521e-01 -4.16975915e-01 9.09102023e-01 -8.89975503e-02
-3.17078084e-01 4.27228570e-01 -1.00579727e+00 1.14630985e+00
1.07480563e-01 1.87401012e-01 -3.14890236e-01 -1.08916664e+00
1.38415372e+00 5.05087733e-01 5.56354463e-01 -2.10885689e-01
-3.40319484e-01 5.23648858e-01 4.26605523e-01 -2.52802789e-01
1.00194886e-01 -1.57749606e-03 3.22181135e-01 7.44930506e-02
-2.08304841e-02 3.74302953e-01 3.62285584e-01 5.76903522e-02
1.32206464e+00 2.56046653e-01 8.91053021e-01 -3.26781422e-01
1.07129109e+00 1.72565714e-01 6.65177166e-01 1.08922936e-01
-7.21611604e-02 3.58896822e-01 5.01774728e-01 -4.47478622e-01
-1.24760818e+00 -9.99047279e-01 -5.39978862e-01 7.84961045e-01
1.21710978e-01 -6.62805259e-01 -7.56500781e-01 -4.47388232e-01
-5.98503016e-02 2.97909200e-01 4.75430042e-02 -1.16529427e-01
-7.37445533e-01 -1.06888950e+00 2.31164724e-01 7.93415383e-02
6.04646280e-02 -1.18310165e+00 -5.47450662e-01 5.48245788e-01
-1.92826614e-01 -5.56966126e-01 -1.94581077e-01 4.50096220e-01
-9.27383363e-01 -1.23240805e+00 -9.00794744e-01 -1.00530517e+00
4.50046033e-01 3.41950893e-01 9.34515834e-01 1.12602629e-01
-3.32706541e-01 -5.44097424e-01 -4.81949776e-01 2.50847906e-01
-9.68203247e-01 -2.31882289e-01 2.74349004e-01 -2.63838917e-01
7.10570693e-01 -9.34944212e-01 -4.37276602e-01 9.96420443e-01
-6.56597257e-01 6.17325753e-02 8.32417369e-01 1.09026754e+00
8.85013998e-01 -2.32723042e-01 5.16456604e-01 -5.62399089e-01
1.64134637e-01 -4.24791783e-01 -6.45025253e-01 4.73166585e-01
-5.97808778e-01 -1.00671709e-01 5.10902584e-01 -5.85873663e-01
-6.32198393e-01 5.83979130e-01 -4.89056975e-01 -3.79599705e-02
-4.16580379e-01 4.30220515e-01 -3.32631350e-01 -2.00150266e-01
7.04970479e-01 5.99849582e-01 3.24268758e-01 -7.01712251e-01
-2.21044064e-01 6.98037028e-01 5.54174364e-01 -3.95727068e-01
4.32501704e-01 1.69721153e-03 1.82637393e-01 -9.69702065e-01
1.26137823e-01 -9.92992699e-01 -8.05189669e-01 8.54232162e-02
9.89558041e-01 -6.27298117e-01 -9.16816354e-01 5.09744644e-01
-1.20238566e+00 8.06378424e-02 5.55942714e-01 7.08988070e-01
-8.85977864e-01 1.12240505e+00 -7.06328571e-01 -6.20067656e-01
-1.77110419e-01 -1.40813875e+00 9.06781197e-01 3.48384418e-02
-5.86250961e-01 -6.32967532e-01 3.85610282e-01 2.59119004e-01
-1.16148993e-01 4.36278999e-01 1.18296349e+00 -8.56071532e-01
-1.39629811e-01 -1.26707271e-01 5.80091476e-02 1.61649764e-01
1.35255039e-01 2.12235421e-01 -3.80669326e-01 -3.19017887e-01
-2.99742430e-01 -4.15591374e-02 5.83112180e-01 2.06488565e-01
4.21554178e-01 -1.07488729e-01 -5.92640460e-01 2.29325116e-01
1.40177035e+00 7.02107310e-01 9.29200888e-01 6.68306470e-01
1.84080705e-01 6.74741209e-01 1.25560462e+00 3.24473053e-01
-3.02463531e-01 1.27417076e+00 3.79887849e-01 -1.16121501e-01
3.72524410e-01 -5.69581874e-02 4.78810787e-01 4.71865654e-01
-4.27088112e-01 -3.53771709e-02 -9.73717272e-01 9.31178033e-02
-2.13265300e+00 -1.12831259e+00 -7.85120904e-01 2.25183558e+00
1.03776145e+00 3.82147282e-02 5.58654666e-01 2.41530180e-01
1.10041952e+00 -3.28592360e-01 -4.45681870e-01 -7.15488911e-01
-3.63827974e-01 2.30723426e-01 4.80470151e-01 3.57796878e-01
-9.36898589e-01 5.96605003e-01 5.85328770e+00 9.79327679e-01
-8.29674125e-01 -3.04541737e-01 1.90123040e-02 6.46643877e-01
1.87211782e-01 3.70764226e-01 -7.98381329e-01 7.96346962e-01
1.18850541e+00 1.36615317e-02 7.05720217e-04 9.06968534e-01
1.99215516e-01 -2.13690788e-01 -7.85293102e-01 8.69872570e-01
-6.89651012e-01 -1.32884443e+00 -1.07353605e-01 3.76770645e-01
2.61404663e-01 -9.14005861e-02 -4.19115007e-01 -2.90917546e-01
1.86224312e-01 -8.09768081e-01 2.20070541e-01 4.78912503e-01
4.71070796e-01 -8.52819860e-01 9.07258868e-01 4.38999027e-01
-1.26374555e+00 5.05671740e-01 -7.01743066e-01 1.39710501e-01
5.26181161e-01 5.87267995e-01 -1.04762971e+00 6.86715841e-01
5.76836586e-01 5.39882064e-01 -1.49352551e-01 9.70032811e-01
-2.33506113e-01 5.63523948e-01 -2.82982349e-01 -5.05860746e-02
9.73711312e-02 -5.59818745e-01 5.59795499e-01 1.39590120e+00
1.03101730e-01 1.73113704e-01 3.97812307e-01 2.86575824e-01
6.78039074e-01 5.46191990e-01 -6.87450618e-02 -1.02519885e-01
3.05677235e-01 1.00753164e+00 -6.86554015e-01 -4.48148876e-01
-3.95756572e-01 1.07056153e+00 3.83487679e-02 7.23973215e-02
-8.14429641e-01 -5.42636991e-01 1.14897120e+00 1.13328248e-01
5.07153988e-01 -4.04485278e-02 4.68109637e-01 -8.57930422e-01
-1.76534668e-01 -1.17406452e+00 4.29663956e-01 -7.48163998e-01
-1.13493812e+00 6.05055928e-01 -6.70648664e-02 -1.24760032e+00
-5.68071663e-01 -6.84436679e-01 -3.97324085e-01 1.45603430e+00
-1.12433898e+00 -7.88397431e-01 1.80172473e-01 2.02864856e-01
4.03625011e-01 -1.12952821e-01 1.20609987e+00 -1.08010387e-02
-3.44251156e-01 3.29210460e-01 5.93878925e-01 -4.72090095e-01
7.56663322e-01 -1.19499803e+00 5.59340119e-01 3.26594740e-01
-3.93995613e-01 8.54650378e-01 1.02160835e+00 -9.88753676e-01
-9.78584051e-01 -6.33048654e-01 1.29596424e+00 -6.68570101e-02
6.09016955e-01 -5.05906455e-02 -1.49486732e+00 2.84439474e-01
-9.41658765e-02 -5.32425106e-01 1.03270626e+00 -3.49116236e-01
-2.63686895e-01 5.38814425e-01 -1.23294687e+00 2.48191133e-01
7.52237618e-01 -3.12027663e-01 -6.79074168e-01 5.59337735e-01
3.86633515e-01 -1.74280107e-01 -1.54393339e+00 5.93120277e-01
8.71281028e-01 -1.24723387e+00 1.12985611e+00 -7.51840770e-01
8.23235735e-02 -7.72720933e-01 -2.76888102e-01 -7.07079053e-01
-6.22666478e-01 -5.66241741e-01 6.09810874e-02 9.13560629e-01
4.42045540e-01 -7.14413822e-01 6.77720785e-01 -9.03096125e-02
-2.05520570e-01 -9.23121452e-01 -8.68736029e-01 -9.54640269e-01
-1.31410331e-01 2.71762580e-01 4.85258013e-01 8.52190554e-01
2.48465568e-01 1.99164376e-01 -2.45912001e-01 1.74087569e-01
2.69679785e-01 3.17678452e-01 6.08200431e-01 -1.12709510e+00
-5.65687537e-01 -2.59052187e-01 -9.21048939e-01 -7.18802929e-01
-1.16874920e-02 -7.22550750e-01 -6.86025247e-02 -1.13807845e+00
3.57315838e-01 2.90116630e-02 -2.45896980e-01 4.76352185e-01
-6.10710569e-02 8.43389556e-02 -1.46783113e-01 6.01048231e-01
-1.81888655e-01 1.65086463e-02 6.34211481e-01 2.67826557e-01
-1.54688954e-01 1.84367284e-01 2.01327309e-01 7.28953183e-01
9.22311187e-01 -3.80214036e-01 1.23036608e-01 8.45999837e-01
2.37473808e-02 2.48496309e-01 1.14576779e-01 -7.02420771e-01
-2.78958291e-01 -4.62445885e-01 2.88772047e-01 -9.56553519e-01
4.24848795e-01 -5.43005109e-01 8.98959994e-01 1.21306932e+00
-2.81471331e-02 4.43325400e-01 -1.51872620e-01 4.59341764e-01
-3.08815837e-01 -4.83539850e-01 1.06604707e+00 -9.99857206e-03
-8.49669456e-01 -1.94308847e-01 -8.75934660e-01 -8.39949012e-01
1.03176510e+00 -5.28551996e-01 -2.21156031e-01 -1.10320352e-01
-1.02243340e+00 -2.71791548e-01 1.05316627e+00 6.07454544e-03
6.63443744e-01 -9.84740138e-01 -5.11177242e-01 1.45050168e-01
2.88728952e-01 -6.88905656e-01 1.78879529e-01 1.06036866e+00
-8.92805755e-01 5.68462014e-01 -5.16547859e-01 -8.49446177e-01
-2.11586666e+00 7.52666593e-01 9.07111689e-02 -3.34230512e-01
-3.06959659e-01 4.92979109e-01 1.98867545e-01 -2.82057911e-01
-1.33834362e-01 -4.13993783e-02 -2.70580590e-01 -2.91352153e-01
5.65560937e-01 3.85106534e-01 1.58231109e-01 -1.17006516e+00
-6.66394472e-01 8.35305274e-01 -1.87672257e-01 2.62992322e-01
1.14222467e+00 2.28610188e-02 -4.79654223e-01 9.67250988e-02
1.28627503e+00 -3.05235803e-01 -7.64865160e-01 5.82083501e-02
4.80623573e-01 -2.41715744e-01 -6.27960205e-01 -7.71442592e-01
-1.71547487e-01 5.91224551e-01 7.55841792e-01 5.80911338e-02
1.10765338e+00 8.77582207e-02 8.08123052e-01 3.73704404e-01
4.87635911e-01 -5.81833243e-01 -4.63163614e-01 3.53750199e-01
6.61453724e-01 -9.51203942e-01 -8.62050429e-02 -7.03394234e-01
-5.24450541e-01 1.24557686e+00 2.34119475e-01 6.14794828e-02
1.45338893e-01 -4.19114009e-02 -1.46535933e-01 -5.51464595e-02
-8.28755558e-01 -1.04895830e-01 2.29419440e-01 5.76755524e-01
7.16668844e-01 2.51967255e-02 -1.17441154e+00 2.65551776e-01
-1.49748594e-01 -2.34643757e-01 1.34349182e-01 1.18486500e+00
-9.29347038e-01 -1.77714837e+00 -5.31233430e-01 -4.51140292e-02
-4.49151963e-01 2.34455876e-02 -6.87691271e-01 6.36177182e-01
-5.42102158e-02 6.96078420e-01 -2.80274510e-01 -4.29780275e-01
8.92057046e-02 3.35781217e-01 4.12288040e-01 -3.52620870e-01
-5.90307772e-01 3.29302251e-01 1.76807389e-01 -4.92834181e-01
-5.25849164e-01 -8.58184993e-01 -1.58986211e+00 -2.94644594e-01
-5.29417276e-01 7.72275627e-01 8.40480924e-01 6.72427833e-01
2.34032273e-01 -1.44200519e-01 6.59464180e-01 -6.11249685e-01
-5.78936815e-01 -9.19045806e-01 -6.74190462e-01 2.73265332e-01
-8.26303381e-04 -6.53987288e-01 -2.34404743e-01 1.65371940e-01] | [4.793907165527344, 5.29660177230835] |
8c6ed46e-06eb-4abf-96d2-ebbb668eee0e | meld-a-multimodal-multi-party-dataset-for | 1810.02508 | null | https://arxiv.org/abs/1810.02508v6 | https://arxiv.org/pdf/1810.02508v6.pdf | MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversations | Emotion recognition in conversations is a challenging task that has recently gained popularity due to its potential applications. Until now, however, a large-scale multimodal multi-party emotional conversational database containing more than two speakers per dialogue was missing. Thus, we propose the Multimodal EmotionLines Dataset (MELD), an extension and enhancement of EmotionLines. MELD contains about 13,000 utterances from 1,433 dialogues from the TV-series Friends. Each utterance is annotated with emotion and sentiment labels, and encompasses audio, visual and textual modalities. We propose several strong multimodal baselines and show the importance of contextual and multimodal information for emotion recognition in conversations. The full dataset is available for use at http:// affective-meld.github.io. | ['Soujanya Poria', 'Navonil Majumder', 'Erik Cambria', 'Devamanyu Hazarika', 'Rada Mihalcea', 'Gautam Naik'] | 2018-10-05 | meld-a-multimodal-multi-party-dataset-for-1 | https://aclanthology.org/P19-1050 | https://aclanthology.org/P19-1050.pdf | acl-2019-7 | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [-1.56925768e-01 -1.50942847e-01 2.40873341e-02 -6.72877550e-01
-1.25556982e+00 -7.59462595e-01 6.54036462e-01 8.98959264e-02
-3.17247599e-01 6.23815298e-01 8.17933798e-01 2.89570987e-01
5.33791661e-01 -7.68454671e-02 -1.94743231e-01 -4.60670680e-01
-1.15418613e-01 1.94734991e-01 -5.70418119e-01 -4.43902254e-01
-1.45209581e-01 1.65784746e-01 -1.38923752e+00 7.56927669e-01
4.17452514e-01 1.46698451e+00 -1.99861437e-01 1.03512311e+00
-1.21554390e-01 8.22772563e-01 -6.23126447e-01 -9.33456302e-01
-3.39378417e-01 -4.97341216e-01 -8.98000419e-01 2.66644448e-01
7.01040477e-02 -3.04502130e-01 -2.28502974e-01 7.01259494e-01
9.70902562e-01 2.60457337e-01 5.07241011e-01 -1.62449729e+00
-1.69729948e-01 3.27069908e-01 -3.47203791e-01 -1.59323275e-01
8.98298502e-01 -1.13582328e-01 1.29798388e+00 -1.22174096e+00
7.80301273e-01 1.31983209e+00 4.73919690e-01 7.38422573e-01
-8.77593517e-01 -4.64391500e-01 -6.01565316e-02 1.53661475e-01
-1.12126315e+00 -8.88783514e-01 9.98672068e-01 -1.85481310e-01
9.16798651e-01 4.78565842e-01 6.26989543e-01 1.77444696e+00
-3.10992718e-01 1.33080232e+00 1.10339904e+00 -4.52362895e-01
4.07468490e-02 4.17820990e-01 6.68319613e-02 3.82239699e-01
-1.03354990e+00 -5.23768842e-01 -9.85817909e-01 -4.60744798e-01
2.05559641e-01 -3.48600179e-01 -3.64818931e-01 -3.37042351e-04
-8.53679061e-01 8.77305150e-01 -9.69021097e-02 3.07181776e-01
-4.14472908e-01 -2.78175205e-01 1.10965216e+00 6.00887537e-01
7.52068281e-01 1.54119596e-01 -4.80043620e-01 -1.15039265e+00
-3.49635452e-01 -1.19975254e-01 1.12417114e+00 8.11167896e-01
5.62588632e-01 -2.78090805e-01 1.52073428e-01 1.49511886e+00
1.37308419e-01 5.35424173e-01 1.77458882e-01 -1.21794629e+00
4.15717930e-01 5.35463512e-01 1.27601445e-01 -1.27643526e+00
-7.48519838e-01 4.87807035e-01 -8.40985000e-01 -4.31045979e-01
1.84599102e-01 -7.12372899e-01 -8.53592008e-02 1.69728136e+00
4.40747947e-01 -2.90209204e-01 3.53165299e-01 9.54617620e-01
1.43707418e+00 7.82446146e-01 -2.07926217e-03 -2.68896252e-01
1.56986284e+00 -9.76560950e-01 -1.12054598e+00 -4.69084941e-02
4.15045917e-01 -1.04418445e+00 1.13677680e+00 5.16035140e-01
-1.08761299e+00 -2.15034142e-01 -5.60457885e-01 -1.41883731e-01
-5.50915062e-01 6.76814020e-02 7.83906102e-01 6.77166045e-01
-9.54659343e-01 -2.01965451e-01 -3.48208427e-01 -5.70957124e-01
5.35704829e-02 3.67862023e-02 -6.33718550e-01 1.18666068e-01
-1.44563985e+00 7.25303531e-01 -2.20799372e-01 3.29181671e-01
-6.57196343e-01 -9.90192816e-02 -1.11276984e+00 -2.32550263e-01
2.87188858e-01 3.70566212e-02 1.63782477e+00 -1.19621551e+00
-1.96494496e+00 9.87201810e-01 -5.29980361e-01 1.21641420e-01
1.76720843e-01 -2.44548336e-01 -8.15479040e-01 5.87309718e-01
-3.40663403e-01 7.42806196e-01 6.82941437e-01 -1.12067223e+00
-4.17269409e-01 -2.61759788e-01 6.43558875e-02 4.50843722e-01
-6.29076302e-01 6.08505726e-01 -9.40158308e-01 -3.40085089e-01
-3.92259687e-01 -1.07103050e+00 1.02797799e-01 -5.04501879e-01
-6.30138695e-01 -2.03884035e-01 7.34632850e-01 -6.83418691e-01
9.82995808e-01 -2.53855896e+00 2.45620832e-01 7.20299557e-02
1.51212048e-02 -3.13522071e-01 -3.12177241e-01 8.57998550e-01
-3.91141921e-02 3.20771560e-02 2.14381218e-02 -9.50598240e-01
4.78455871e-01 6.61327913e-02 -1.39334321e-01 2.71529943e-01
1.05004638e-01 9.25276995e-01 -6.31185174e-01 -5.64746678e-01
1.95222780e-01 6.44654155e-01 -2.06329837e-01 5.24541438e-01
-5.83829693e-02 6.45070732e-01 -3.12195212e-01 8.69072080e-01
5.78406632e-01 -1.32589489e-01 2.65137464e-01 -2.94613212e-01
-1.27011418e-01 1.02231480e-01 -8.17128181e-01 1.70644653e+00
-5.44979930e-01 1.15235937e+00 6.05909288e-01 -5.81506729e-01
7.82795548e-01 9.56829369e-01 7.32492507e-01 -5.92421710e-01
4.92120177e-01 -2.47055385e-02 -5.34134626e-01 -8.13829064e-01
9.34561253e-01 -2.08735883e-01 -8.42847645e-01 4.54713672e-01
3.18147928e-01 -2.05333635e-01 1.44346669e-01 4.37868267e-01
8.15540493e-01 -4.03669894e-01 8.38790312e-02 4.45624262e-01
4.64666158e-01 -9.39931795e-02 4.17993158e-01 2.85241455e-01
-4.85469609e-01 5.64453244e-01 8.76964748e-01 -9.32889432e-03
-5.41886628e-01 -6.29720926e-01 -2.30635807e-01 1.54848218e+00
-9.69370455e-02 -6.87480628e-01 -5.29114187e-01 -4.60685790e-01
-4.05036569e-01 3.14566344e-01 -6.58633590e-01 2.42692024e-01
-2.94650700e-02 -5.32813966e-01 7.97784030e-01 1.97465092e-01
4.82508659e-01 -1.20099461e+00 -1.03209808e-01 -5.85532635e-02
-1.02334619e+00 -1.39134586e+00 -6.43385351e-01 5.04377633e-02
-2.06687212e-01 -9.02005672e-01 -8.23905587e-01 -6.02667451e-01
1.56268224e-01 -4.77356873e-02 1.33092737e+00 -4.68333542e-01
-1.84077144e-01 1.06541812e+00 -7.03404486e-01 -2.77881771e-01
-4.03621912e-01 5.22665009e-02 -8.51266906e-02 3.00552666e-01
4.81349528e-01 -1.62151188e-01 -2.93447703e-01 4.20923054e-01
-5.77643275e-01 -8.65755305e-02 1.22035243e-01 8.03654253e-01
1.95678607e-01 -3.52036417e-01 8.05416405e-01 -5.90875745e-01
1.00212109e+00 -5.06752729e-01 -9.44344029e-02 1.31761804e-01
4.37116414e-01 -7.51189888e-01 2.59442806e-01 -2.91816205e-01
-1.37401950e+00 -1.04735516e-01 -5.53596258e-01 -2.16062710e-01
-5.64241648e-01 7.48327851e-01 -3.77165675e-01 2.49772251e-01
-2.31526773e-02 -1.89192072e-01 -1.17812000e-01 -3.65337104e-01
4.99322802e-01 1.17261398e+00 3.76561970e-01 -6.33531749e-01
-1.23735867e-01 2.57249296e-01 -5.67692399e-01 -1.33196318e+00
-6.11431897e-01 -6.69282794e-01 -2.79497176e-01 -8.72632205e-01
9.70461071e-01 -1.14952314e+00 -1.26994073e+00 7.00555444e-01
-1.18834054e+00 -3.33303362e-01 1.29603177e-01 4.99775261e-01
-5.03433645e-01 2.94483900e-01 -1.06486344e+00 -1.17796016e+00
-2.89871275e-01 -9.53549087e-01 1.10834157e+00 1.29754812e-01
-6.62202477e-01 -9.84882057e-01 1.82118982e-01 5.52061200e-01
1.45834893e-01 3.20153862e-01 4.10817027e-01 -6.23124838e-01
3.47136408e-01 -3.74492288e-01 -4.85845357e-02 3.80729795e-01
1.49258031e-02 2.10922956e-01 -1.29891992e+00 7.51490965e-02
-1.50027841e-01 -1.38948679e+00 5.20780921e-01 1.87336192e-01
7.42282569e-01 -7.06840530e-02 9.67073143e-02 -1.19489536e-01
7.92962730e-01 2.25495890e-01 4.54624116e-01 -8.14045146e-02
3.08486789e-01 1.08586860e+00 6.42118037e-01 9.11083877e-01
7.89091170e-01 6.16410494e-01 2.93968290e-01 -2.40802094e-01
5.11569500e-01 7.94823021e-02 5.00080943e-01 1.35717273e+00
4.82452437e-02 -5.32162368e-01 -8.90196562e-01 5.16832829e-01
-1.75086081e+00 -1.09893501e+00 -4.68319282e-02 1.65344441e+00
9.31013346e-01 -3.16504985e-01 4.56481397e-01 -9.89793837e-02
6.29813313e-01 3.56222302e-01 -4.36129183e-01 -7.26123810e-01
-5.59005558e-01 -3.36681694e-01 -4.25554395e-01 5.32426715e-01
-1.23318768e+00 7.05729425e-01 6.13957882e+00 6.34701073e-01
-9.64665353e-01 1.69841275e-01 9.70785618e-01 -2.23213002e-01
-1.83731914e-01 -5.25314093e-01 -5.10188460e-01 2.04002559e-01
1.14583218e+00 1.10291883e-01 3.82438064e-01 7.52947688e-01
2.81410158e-01 -3.46991897e-01 -1.03664649e+00 1.30322921e+00
3.57912123e-01 -8.45074475e-01 -5.18065155e-01 -1.94437817e-01
6.73036873e-01 1.74003124e-01 1.00445479e-01 6.39939904e-01
6.70142248e-02 -8.39478791e-01 4.09608603e-01 5.04335880e-01
6.84327126e-01 -1.00112772e+00 8.88591349e-01 2.00992543e-02
-1.13132811e+00 2.17941403e-01 2.10449934e-01 1.44915476e-01
5.52856386e-01 3.54438156e-01 -5.36911786e-01 3.96691620e-01
9.52839911e-01 8.68552744e-01 -3.02168459e-01 4.76692557e-01
6.09968752e-02 5.80250442e-01 -2.95155674e-01 -2.47057587e-01
2.39165470e-01 -2.91287482e-01 4.13266093e-01 1.58294821e+00
7.88951442e-02 2.93164015e-01 1.01318143e-01 9.11190137e-02
-5.53393781e-01 3.60422552e-01 -5.19497931e-01 -5.39791882e-01
4.22531873e-01 1.68749630e+00 -3.36904854e-01 -3.38377714e-01
-7.53352880e-01 1.19307292e+00 1.69919312e-01 4.65174139e-01
-8.41697812e-01 -3.49891156e-01 7.54903555e-01 -9.17162180e-01
-1.32570341e-01 -2.73352135e-02 1.48161784e-01 -1.24954796e+00
-1.55918807e-01 -1.26003337e+00 5.38271487e-01 -8.86490941e-01
-1.56995988e+00 7.96462357e-01 -4.19076771e-01 -1.15161669e+00
-4.68472689e-01 -4.51792866e-01 -3.68274301e-01 5.41339517e-01
-1.06760001e+00 -9.58755970e-01 -4.13339198e-01 8.77116561e-01
6.27575934e-01 -2.58440580e-02 1.33908355e+00 5.13265669e-01
-7.63113558e-01 5.86838305e-01 6.80228844e-02 3.15807968e-01
1.24653637e+00 -1.10625982e+00 -3.16455811e-01 -1.23870097e-01
-1.94820017e-01 2.31766194e-01 7.59516537e-01 -2.57808529e-02
-1.62661493e+00 -5.47997177e-01 1.07994401e+00 -4.59607631e-01
8.02942574e-01 -7.30926096e-01 -6.25519514e-01 6.24161243e-01
1.14199293e+00 -5.49258053e-01 1.32992148e+00 4.82050061e-01
-1.23294361e-01 1.25990584e-01 -1.02628732e+00 7.19058037e-01
4.32162672e-01 -9.79420900e-01 -2.19306380e-01 2.39528134e-01
4.67678845e-01 -4.08174336e-01 -1.35402822e+00 2.06958950e-01
7.16787338e-01 -1.06699181e+00 6.38653100e-01 -3.45493138e-01
5.96303046e-01 3.74214768e-01 -4.71199244e-01 -1.46467340e+00
4.88010138e-01 -8.39769542e-01 3.83560993e-02 1.79189336e+00
6.07229233e-01 -4.79230046e-01 4.70856488e-01 9.67629135e-01
-9.68266651e-02 -6.30986691e-01 -1.01270366e+00 -1.46360591e-01
-2.66470373e-01 -7.43081570e-01 3.45090628e-01 1.23003018e+00
9.16157782e-01 7.92622089e-01 -9.06127751e-01 -1.11463755e-01
1.12394921e-01 2.57541388e-01 1.05398285e+00 -7.82513499e-01
1.77151263e-02 -3.81508201e-01 -9.64398775e-03 -1.12089157e+00
4.89555597e-01 -3.63247842e-01 1.31842345e-01 -1.20409393e+00
2.63104111e-01 1.09951146e-01 1.00349523e-02 4.80951458e-01
1.53400689e-01 4.32983547e-01 1.24972351e-01 -1.49207085e-01
-1.20275462e+00 1.10206378e+00 1.09280646e+00 -1.86443582e-01
-1.15218744e-01 -2.25468993e-01 -4.88987386e-01 7.90688097e-01
1.02316368e+00 3.54310200e-02 3.80922072e-02 -3.51106119e-03
3.63235265e-01 6.05426252e-01 7.42716789e-02 -3.57716203e-01
2.36791670e-02 1.13198802e-01 2.09059566e-01 -8.39557469e-01
1.26534295e+00 -8.77930939e-01 -7.90191144e-02 -3.67269099e-01
-6.06016219e-01 1.23993464e-01 5.30569434e-01 3.28435719e-01
-8.01504552e-01 5.48717752e-03 3.54675293e-01 6.69336319e-02
-7.39014208e-01 1.67854615e-02 -9.44247425e-01 2.20986053e-01
6.47630930e-01 2.88556904e-01 -4.04655069e-01 -1.20503020e+00
-1.03167903e+00 4.93555814e-01 2.08918467e-01 6.66380346e-01
6.00123048e-01 -1.41615307e+00 -5.63399792e-01 -2.81768501e-01
4.57088262e-01 -6.52767837e-01 7.68855214e-01 1.07836771e+00
1.49518371e-01 2.86608547e-01 -9.41745490e-02 -4.18106139e-01
-1.69278204e+00 9.53357518e-02 1.63354278e-01 5.75668877e-03
-6.98172078e-02 7.43474722e-01 -7.74292275e-02 -7.96534479e-01
4.90738302e-01 2.43592709e-01 -3.98800135e-01 6.55587733e-01
6.35305643e-01 3.38041812e-01 -1.68789268e-01 -1.10611320e+00
-3.43116999e-01 2.29865834e-01 3.30564052e-01 -6.19669080e-01
1.23904037e+00 -8.51910174e-01 -3.83704752e-01 1.11295044e+00
1.54529238e+00 2.22957313e-01 -9.02496755e-01 -1.44193754e-01
-1.34292632e-01 -2.74290621e-01 -2.76580065e-01 -8.36047649e-01
-7.84938157e-01 9.06514823e-01 3.50273758e-01 5.43334484e-01
1.26457536e+00 2.10433245e-01 8.22256446e-01 4.65290487e-01
4.91535366e-02 -1.39546931e+00 4.85040784e-01 9.52742040e-01
1.25423563e+00 -1.67266572e+00 -5.76260507e-01 -2.73839146e-01
-1.48563743e+00 1.04425514e+00 5.50516605e-01 6.55344665e-01
5.49570680e-01 1.89508155e-01 6.25237525e-01 -3.23799223e-01
-1.01137006e+00 -5.26679307e-02 1.03522785e-01 2.76071697e-01
7.76246130e-01 2.71802619e-02 7.88166896e-02 9.37189698e-01
-1.72099888e-01 -3.64201963e-01 3.17363292e-01 7.61635959e-01
1.13530885e-02 -9.34032798e-01 -3.72912228e-01 1.06599107e-01
-5.95740736e-01 -1.25054732e-01 -8.82074475e-01 5.69535017e-01
-3.98224711e-01 1.68317616e+00 -1.53559130e-02 -4.76714492e-01
4.15082663e-01 4.16271538e-01 6.32026643e-02 -6.10088594e-02
-7.73065865e-01 4.63177085e-01 9.24870670e-01 -5.79005539e-01
-8.40285003e-01 -8.05482805e-01 -1.05083990e+00 -3.79470766e-01
-2.43083388e-02 4.53949660e-01 7.70410597e-01 5.51988542e-01
5.30749977e-01 2.21522748e-01 9.74138737e-01 -1.28490388e+00
2.07708955e-01 -1.13238704e+00 -6.78026974e-01 4.69202846e-01
3.22990030e-01 -3.83345306e-01 -5.14553070e-01 6.98700249e-02] | [13.093202590942383, 5.686499118804932] |
ce4ef6c3-1a01-49a7-a152-2b8a1ab7f446 | keys-to-better-image-inpainting-structure-and | 2208.03382 | null | https://arxiv.org/abs/2208.03382v2 | https://arxiv.org/pdf/2208.03382v2.pdf | Keys to Better Image Inpainting: Structure and Texture Go Hand in Hand | Deep image inpainting has made impressive progress with recent advances in image generation and processing algorithms. We claim that the performance of inpainting algorithms can be better judged by the generated structures and textures. Structures refer to the generated object boundary or novel geometric structures within the hole, while texture refers to high-frequency details, especially man-made repeating patterns filled inside the structural regions. We believe that better structures are usually obtained from a coarse-to-fine GAN-based generator network while repeating patterns nowadays can be better modeled using state-of-the-art high-frequency fast fourier convolutional layers. In this paper, we propose a novel inpainting network combining the advantages of the two designs. Therefore, our model achieves a remarkable visual quality to match state-of-the-art performance in both structure generation and repeating texture synthesis using a single network. Extensive experiments demonstrate the effectiveness of the method, and our conclusions further highlight the two critical factors of image inpainting quality, structures, and textures, as the future design directions of inpainting networks. | ['Humphrey Shi', 'Ning Yu', 'Yuqian Zhou', 'Jitesh Jain'] | 2022-08-05 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 4.72656876e-01 2.20697984e-01 7.92496949e-02 -8.88108462e-02
-6.30298555e-01 -1.55021876e-01 5.81929266e-01 -5.18693745e-01
1.54468521e-01 8.26655447e-01 3.39439541e-01 1.41003132e-01
1.92171067e-01 -1.13514447e+00 -1.14311659e+00 -6.67632699e-01
1.04213305e-01 1.64959267e-01 1.37004778e-01 -4.72798228e-01
1.15960710e-01 5.64613581e-01 -1.57632136e+00 6.71870828e-01
9.64262486e-01 1.32236183e+00 3.65120471e-01 5.28454781e-01
-1.91980764e-01 9.69159484e-01 -6.02659881e-01 -5.07617652e-01
4.76366878e-01 -7.98643172e-01 -4.35665458e-01 2.62585133e-01
5.14064014e-01 -6.67490840e-01 -4.55856353e-01 8.50606501e-01
2.39343405e-01 -2.34256655e-01 5.63524485e-01 -8.47531319e-01
-1.00385833e+00 5.01470387e-01 -7.64014542e-01 -3.34692538e-01
2.06437424e-01 3.08072180e-01 6.53016210e-01 -9.06429768e-01
7.74497688e-01 1.28690016e+00 7.31766105e-01 5.45664966e-01
-1.27971995e+00 -6.40648127e-01 -1.53766260e-01 2.36070007e-02
-1.13083005e+00 -3.54485601e-01 1.19448924e+00 -1.93850949e-01
6.35963500e-01 1.99418679e-01 5.64960420e-01 1.08970881e+00
6.09168530e-01 5.81036925e-01 1.10633135e+00 -4.70195532e-01
8.90100002e-02 -3.29506308e-01 -8.37623179e-01 8.24326813e-01
8.17857161e-02 4.24080998e-01 -5.44344962e-01 2.77171582e-01
1.52966452e+00 6.07968913e-03 -3.73744190e-01 2.02211738e-02
-1.09612620e+00 5.90893328e-01 6.74635053e-01 2.47848302e-01
-3.97304952e-01 4.94768798e-01 3.27330641e-02 2.71339446e-01
7.09714532e-01 6.92487001e-01 1.11803701e-02 3.39380130e-02
-1.18406105e+00 4.31082040e-01 2.96493709e-01 1.15406740e+00
9.20127094e-01 4.09650832e-01 -3.47756803e-01 8.81385386e-01
-5.31319119e-02 3.33783805e-01 2.53683090e-01 -1.19437623e+00
2.43856788e-01 3.34482491e-01 2.24407569e-01 -1.06332386e+00
5.63522018e-02 -3.74819666e-01 -1.22334743e+00 4.01289195e-01
2.67669894e-02 -5.46441972e-03 -9.98902917e-01 1.35356224e+00
9.98481512e-02 1.65041789e-01 -3.42400551e-01 6.46390498e-01
7.36426890e-01 7.91150808e-01 -3.40969563e-01 6.58493862e-03
1.23994935e+00 -1.43776059e+00 -9.38263357e-01 -1.28053501e-01
1.11747213e-01 -1.14805424e+00 9.65577602e-01 3.56139362e-01
-1.38544512e+00 -9.57312584e-01 -1.11462522e+00 -2.46057555e-01
2.54565448e-01 8.13219473e-02 6.21654987e-01 2.28736132e-01
-1.12981236e+00 1.08856797e+00 -5.76255262e-01 1.24890156e-01
6.94563031e-01 -1.46179408e-01 -2.30478421e-01 -9.75012109e-02
-8.31324637e-01 5.29496431e-01 1.20873734e-01 1.59353122e-01
-8.21501195e-01 -1.00474083e+00 -7.43382394e-01 1.28983960e-01
1.45851150e-01 -1.02904224e+00 1.17373824e+00 -1.11696041e+00
-1.78520811e+00 8.15290570e-01 -1.80067912e-01 -4.31187332e-01
7.44859278e-01 -1.74635544e-01 -3.33219230e-01 4.36011314e-01
8.55300874e-02 8.52346301e-01 1.28396690e+00 -1.50538087e+00
-5.75841248e-01 8.92120898e-02 -2.29610220e-01 -1.92219496e-01
9.96299386e-02 -2.39038974e-01 -2.31345534e-01 -1.41256595e+00
-1.51173035e-02 -5.04552186e-01 -2.83562869e-01 4.36514378e-01
-3.04515809e-01 1.39046997e-01 1.17624080e+00 -8.74725759e-01
1.13648045e+00 -2.23519039e+00 -3.17144655e-02 -7.67904520e-02
2.66258955e-01 1.47992641e-01 -3.28202397e-01 5.12570620e-01
-3.52988392e-02 2.24615261e-01 -2.38944158e-01 -5.09587288e-01
-2.22323567e-01 1.22829735e-01 -7.69704700e-01 -6.40815450e-03
6.07340991e-01 1.12252223e+00 -6.40828133e-01 -1.31982818e-01
2.04110786e-01 5.76060355e-01 -5.44674397e-01 4.91560191e-01
-5.25410235e-01 6.51213586e-01 -4.38695639e-01 6.60016596e-01
8.43966842e-01 -1.78550005e-01 -2.35017642e-01 -3.59629422e-01
-1.58110678e-01 1.79931268e-01 -7.54597664e-01 1.84862542e+00
-5.80267787e-01 5.81824422e-01 1.29472017e-01 -6.21010005e-01
1.20140719e+00 1.13437437e-01 2.65587956e-01 -1.05186617e+00
2.25733239e-02 4.13426667e-01 -2.20789820e-01 -3.61104935e-01
7.11920917e-01 -1.70171589e-01 3.32746685e-01 3.35658729e-01
-1.75762773e-01 -5.78945160e-01 5.30314483e-02 -7.04459995e-02
7.90427208e-01 3.72968465e-01 -2.32975453e-01 -2.13195696e-01
1.02922574e-01 -3.10066283e-01 3.18664789e-01 7.68683195e-01
4.10064161e-01 1.31785274e+00 5.56600928e-01 -7.18162656e-01
-1.68501985e+00 -9.51906145e-01 -9.71089751e-02 4.54162896e-01
1.01583593e-01 -3.59551221e-01 -1.04331970e+00 -2.71450937e-01
-2.43740708e-01 4.58196521e-01 -8.19372833e-01 -2.06898749e-01
-8.64494205e-01 -3.09930593e-01 4.07295287e-01 6.10904157e-01
1.11372364e+00 -1.33788764e+00 -5.02877414e-01 4.50260758e-01
-2.88175404e-01 -1.11280799e+00 -6.03313506e-01 -3.50765139e-01
-1.06267941e+00 -8.24905157e-01 -1.05507815e+00 -1.04621267e+00
5.99890649e-01 2.18568131e-01 1.32459807e+00 3.56769353e-01
-3.61598074e-01 -2.19433486e-01 -4.67410713e-01 -1.73042595e-01
-7.51564145e-01 -1.90386295e-01 -3.47570658e-01 2.05465421e-01
-5.18052161e-01 -9.82795894e-01 -1.01596522e+00 3.41717780e-01
-1.29232717e+00 6.88724816e-01 7.78734386e-01 1.00759876e+00
7.42646635e-01 2.60672271e-01 3.38930905e-01 -7.41074800e-01
4.07210767e-01 5.46604171e-02 -3.75378460e-01 8.99383426e-02
-3.52105051e-01 2.18684465e-01 8.34782779e-01 -3.73805523e-01
-1.26168275e+00 -3.05552930e-01 -4.56439763e-01 -5.48683465e-01
-6.87542781e-02 1.69855922e-01 -1.32815272e-01 -3.31475258e-01
5.84480286e-01 4.11203951e-01 8.01294073e-02 -5.78722179e-01
3.41416687e-01 2.57845014e-01 5.73793471e-01 -9.44561601e-01
7.89933980e-01 7.64951885e-01 7.76721761e-02 -6.91252768e-01
-8.40428293e-01 4.03370589e-01 -2.52486199e-01 -1.65673524e-01
7.12116897e-01 -8.87533903e-01 -3.96894664e-01 8.44958544e-01
-1.28580856e+00 -5.60548306e-01 -6.77741408e-01 -1.38437852e-01
-7.22922087e-01 4.25044507e-01 -9.30411816e-01 -4.18351471e-01
-5.33982217e-01 -1.19638383e+00 1.42511654e+00 1.99513659e-01
-1.55763291e-02 -7.09743083e-01 -2.10809857e-01 2.79395103e-01
7.75075376e-01 6.52415276e-01 1.18166721e+00 6.44792795e-01
-9.95439410e-01 2.56015331e-01 -3.60985667e-01 4.85820681e-01
3.33754212e-01 6.70469031e-02 -8.80389094e-01 -1.79810002e-01
2.11911336e-01 -2.36652672e-01 1.06427395e+00 4.83387947e-01
1.49581838e+00 -4.51723307e-01 -1.23389959e-01 1.03870916e+00
1.48260331e+00 2.66797543e-01 1.37738681e+00 2.83880502e-01
8.69953811e-01 4.26599264e-01 2.62171507e-01 3.15762371e-01
-8.35193470e-02 7.09183216e-01 2.87167221e-01 -4.91209030e-01
-8.24633479e-01 -6.93889797e-01 1.75086513e-01 6.41863883e-01
-2.95455396e-01 -1.82439685e-01 -5.00125706e-01 3.38296980e-01
-1.59908187e+00 -9.84997690e-01 -1.48513734e-01 1.86232877e+00
1.01739895e+00 1.10331871e-01 -1.59621298e-01 -7.62849972e-02
6.26259446e-01 2.28693724e-01 -3.94986272e-01 -2.19584376e-01
-3.97100538e-01 7.42334545e-01 2.15906486e-01 5.11666477e-01
-7.65236557e-01 9.84970808e-01 6.53831911e+00 1.43278456e+00
-1.26962495e+00 -1.76381379e-01 1.30905974e+00 1.90830350e-01
-5.18878341e-01 -3.99450026e-02 -5.78829408e-01 5.97062647e-01
4.26616430e-01 1.84617087e-01 5.15542269e-01 7.04146981e-01
1.87655628e-01 -9.46801156e-02 -1.01899922e+00 9.72165108e-01
5.06809503e-02 -2.07353282e+00 5.81415296e-01 -6.68072188e-03
1.27996302e+00 -3.91910851e-01 3.63316536e-01 -1.60293579e-01
-7.28185326e-02 -1.25049543e+00 1.02947819e+00 6.81473970e-01
1.26859057e+00 -7.86184430e-01 3.02800357e-01 3.16221826e-02
-1.16916347e+00 1.60858080e-01 -3.38735759e-01 -6.85173972e-03
2.37317458e-01 9.53682780e-01 -2.59317219e-01 6.44760549e-01
7.64403403e-01 7.85911322e-01 -4.35355276e-01 7.94694006e-01
-4.64022756e-01 4.83636826e-01 -1.27604872e-01 3.86353254e-01
2.05948696e-01 -3.49134535e-01 2.12287471e-01 8.68185699e-01
7.08641350e-01 3.26273553e-02 -1.17026754e-01 1.55873573e+00
-3.11033219e-01 -2.44951144e-01 -5.43304324e-01 -1.31232794e-02
1.92305937e-01 1.16731834e+00 -7.36491919e-01 -3.70876014e-01
-2.46788889e-01 1.07750690e+00 1.45358369e-01 3.36043864e-01
-8.32848430e-01 -5.07405102e-01 5.62553704e-01 6.08602524e-01
5.76425016e-01 -2.58795947e-01 -4.19599086e-01 -1.12207580e+00
2.65920460e-01 -9.32791471e-01 -3.02296460e-01 -1.05822289e+00
-1.33264780e+00 8.83788645e-01 -4.15561080e-01 -1.38547528e+00
3.02929506e-02 -5.54274380e-01 -6.93817139e-01 8.50235105e-01
-1.55568814e+00 -1.31296527e+00 -4.13528025e-01 3.41036290e-01
5.63682139e-01 7.31531382e-02 6.60657704e-01 2.57709980e-01
-1.44705772e-01 5.48281193e-01 2.05590129e-01 7.93215781e-02
6.13512874e-01 -6.72236800e-01 6.74101055e-01 9.60026681e-01
-2.16435771e-02 2.20614418e-01 5.26458025e-01 -7.69687712e-01
-1.21237981e+00 -1.27446008e+00 5.21164894e-01 -4.49244641e-02
2.80435413e-01 -3.27344000e-01 -8.91280413e-01 3.37491810e-01
4.14859921e-01 -1.08675577e-01 5.04951775e-02 -4.69525754e-01
-3.36754203e-01 -1.95896029e-01 -1.22035277e+00 7.34362483e-01
1.24597394e+00 -2.98215359e-01 -2.10251257e-01 2.41456538e-01
8.05306137e-01 -5.47935069e-01 -6.59482956e-01 5.97599626e-01
4.99399275e-01 -1.41830099e+00 7.95669079e-01 -3.23477536e-02
1.33801496e+00 -4.29278344e-01 1.17783003e-01 -1.34578526e+00
-5.09091556e-01 -9.97275710e-01 -3.90785635e-02 1.04481101e+00
1.32184684e-01 -2.80799329e-01 9.39142823e-01 1.77407607e-01
-3.75458181e-01 -9.80745256e-01 -6.71614945e-01 -6.98905170e-01
1.11711174e-01 -1.65482625e-01 9.67678487e-01 6.81193709e-01
-6.92568421e-01 6.03975430e-02 -6.66670859e-01 -2.63165236e-01
5.00065684e-01 3.95905465e-01 7.01932728e-01 -8.24048698e-01
-5.36936760e-01 -4.95913982e-01 -1.45953983e-01 -1.39600718e+00
-1.40880436e-01 -3.53770882e-01 2.77424306e-02 -1.37993574e+00
-3.47645879e-02 -5.11818826e-01 3.31895113e-01 2.72777647e-01
-9.07581747e-02 6.93882644e-01 2.05479115e-01 2.15159789e-01
1.82873034e-03 9.69453156e-01 2.02212787e+00 -1.63678885e-01
-7.21043441e-03 -3.53758156e-01 -7.63464272e-01 6.99284315e-01
5.44182897e-01 -2.11005807e-01 -2.85503268e-01 -8.35733116e-01
2.91648299e-01 1.70051172e-01 4.26143199e-01 -1.16735220e+00
-1.10681295e-01 -2.47364063e-02 7.53399789e-01 -3.18287492e-01
2.87862599e-01 -6.50494218e-01 5.73294580e-01 3.94215673e-01
-1.42420799e-01 -7.10287690e-02 1.75653651e-01 4.28005427e-01
-5.81418693e-01 1.00617804e-01 1.12480545e+00 -3.63277048e-01
-4.41111743e-01 4.54419941e-01 -7.75873885e-02 -5.43191433e-02
7.76352286e-01 -4.26116407e-01 -3.07204217e-01 -5.61978221e-01
-4.21387345e-01 -3.34175736e-01 6.86562538e-01 4.50566053e-01
9.00211573e-01 -1.54069364e+00 -8.29204917e-01 6.24172390e-01
-2.07119912e-01 3.49694401e-01 4.52620715e-01 3.84057045e-01
-1.09420407e+00 -1.39666116e-02 -4.66342866e-01 -5.01369417e-01
-7.28169620e-01 3.01446736e-01 2.11331308e-01 -3.18378955e-01
-8.50197315e-01 9.06676531e-01 5.87970793e-01 1.18608713e-01
5.12008648e-03 -4.65514004e-01 3.47642988e-01 -3.94397885e-01
6.64712191e-01 1.40204489e-01 5.39456122e-02 -2.36034840e-01
3.48567933e-01 7.42468059e-01 -2.40500540e-01 1.62782416e-01
1.45994616e+00 1.89363554e-01 -3.31723839e-01 -1.21436886e-01
8.99447918e-01 5.93129918e-02 -1.80811799e+00 -1.63366664e-02
-6.26880229e-01 -6.94373667e-01 -1.57639608e-01 -6.99906230e-01
-1.45450699e+00 8.13710332e-01 2.88941771e-01 3.14037651e-02
1.34690738e+00 -2.31493011e-01 1.39254975e+00 -2.76998393e-02
6.71626389e-01 -8.71314466e-01 4.86917436e-01 3.48536193e-01
1.43917572e+00 -7.24653959e-01 -1.17872782e-01 -8.28870058e-01
-2.51796037e-01 1.24794495e+00 6.07282460e-01 -6.28770530e-01
5.49801111e-01 5.19209683e-01 -5.77178299e-02 2.94851102e-02
-5.12490332e-01 3.29354465e-01 2.57237494e-01 6.35973990e-01
4.02762324e-01 -2.71011814e-02 -8.80206227e-02 2.99756676e-01
-4.76882488e-01 -1.59604829e-02 5.02312124e-01 6.87423408e-01
-2.14239076e-01 -1.27222848e+00 -3.61009628e-01 4.20292169e-01
-3.69923234e-01 -2.64696717e-01 -1.27953842e-01 6.31092131e-01
3.61698806e-01 6.47965252e-01 1.29952222e-01 -2.76925266e-01
3.05455059e-01 -4.52805758e-01 8.15508723e-01 -3.85867029e-01
-4.06885773e-01 2.37537533e-01 -7.72811994e-02 -8.85491371e-01
-3.00143361e-01 -2.22973861e-02 -8.61864507e-01 -3.95511001e-01
-8.93758163e-02 -1.90050378e-01 3.18969578e-01 6.08406246e-01
7.97386587e-01 6.97571218e-01 7.08770871e-01 -1.12685108e+00
-7.40635544e-02 -9.43807721e-01 -7.56583154e-01 5.76026678e-01
2.67731637e-01 -4.64638352e-01 -3.09944719e-01 2.66482353e-01] | [11.410807609558105, -1.0743569135665894] |
93255f72-8e04-499c-ae4e-d13811f642c9 | strong-baselines-for-parameter-efficient-few | 2304.01917 | null | https://arxiv.org/abs/2304.01917v1 | https://arxiv.org/pdf/2304.01917v1.pdf | Strong Baselines for Parameter Efficient Few-Shot Fine-tuning | Few-shot classification (FSC) entails learning novel classes given only a few examples per class after a pre-training (or meta-training) phase on a set of base classes. Recent works have shown that simply fine-tuning a pre-trained Vision Transformer (ViT) on new test classes is a strong approach for FSC. Fine-tuning ViTs, however, is expensive in time, compute and storage. This has motivated the design of parameter efficient fine-tuning (PEFT) methods which fine-tune only a fraction of the Transformer's parameters. While these methods have shown promise, inconsistencies in experimental conditions make it difficult to disentangle their advantage from other experimental factors including the feature extractor architecture, pre-trained initialization and fine-tuning algorithm, amongst others. In our paper, we conduct a large-scale, experimentally consistent, empirical analysis to study PEFTs for few-shot image classification. Through a battery of over 1.8k controlled experiments on large-scale few-shot benchmarks including Meta-Dataset (MD) and ORBIT, we uncover novel insights on PEFTs that cast light on their efficacy in fine-tuning ViTs for few-shot classification. Through our controlled empirical study, we have two main findings: (i) Fine-tuning just the LayerNorm parameters (which we call LN-Tune) during few-shot adaptation is an extremely strong baseline across ViTs pre-trained with both self-supervised and supervised objectives, (ii) For self-supervised ViTs, we find that simply learning a set of scaling parameters for each attention matrix (which we call AttnScale) along with a domain-residual adapter (DRA) module leads to state-of-the-art performance (while being $\sim\!$ 9$\times$ more parameter-efficient) on MD. Our extensive empirical findings set strong baselines and call for rethinking the current design of PEFT methods for FSC. | ['Soheil Feizi', 'Shell Xu Hu', 'Daniela Massiceti', 'Samyadeep Basu'] | 2023-04-04 | null | null | null | null | ['few-shot-image-classification'] | ['computer-vision'] | [ 5.58755934e-01 -3.90116870e-02 -1.18537366e-01 -4.62471217e-01
-9.76347268e-01 -4.80691552e-01 8.28726649e-01 -1.61062062e-01
-6.23893023e-01 4.54839647e-01 2.26219699e-01 -2.98959613e-01
-1.81494251e-01 -5.89521527e-01 -9.84346092e-01 -6.16304874e-01
2.36226216e-01 3.97748053e-01 4.99230146e-01 -3.59051168e-01
1.88078970e-01 1.53299674e-01 -1.93330979e+00 2.91902423e-01
5.71159661e-01 9.95810747e-01 1.57675102e-01 7.59204924e-01
1.09437987e-01 6.69181526e-01 -6.05333626e-01 -5.40873766e-01
4.04029071e-01 -6.19119108e-01 -8.36002648e-01 8.85496661e-03
7.13761628e-01 -2.60183126e-01 -1.04415476e-01 8.38391781e-01
6.27376497e-01 4.17489409e-01 8.53207767e-01 -1.21819305e+00
-7.41547942e-01 5.91162920e-01 -4.94331986e-01 4.27538276e-01
-3.55283260e-01 8.88335168e-01 1.06384718e+00 -8.02960455e-01
6.33587241e-01 8.67391109e-01 9.61614013e-01 8.66656899e-01
-1.37573838e+00 -7.01417387e-01 -2.12490916e-01 2.40040332e-01
-1.18402743e+00 -8.37081671e-01 3.51950318e-01 -4.37009186e-01
1.42044806e+00 8.76524895e-02 5.12414932e-01 1.26195633e+00
5.81239350e-02 4.47425544e-01 1.16687596e+00 -6.14499331e-01
5.32901943e-01 3.29020679e-01 3.70332271e-01 5.87344170e-01
3.92155647e-02 2.76128352e-01 -3.50166559e-01 1.24764768e-03
4.26343352e-01 -7.76523724e-02 -1.47710636e-01 -4.66948003e-01
-9.04146075e-01 1.04578269e+00 4.06090111e-01 1.92294449e-01
-5.07078134e-02 4.57557797e-01 7.32516229e-01 4.89894807e-01
3.39422226e-01 8.66538763e-01 -5.15311897e-01 -4.46840763e-01
-9.15841997e-01 -6.39057308e-02 5.65107465e-01 9.34327245e-01
1.04510772e+00 2.06587523e-01 -5.53786516e-01 1.09820259e+00
-4.14275438e-01 3.53497028e-01 8.85927379e-01 -9.32733655e-01
2.39670753e-01 2.76995361e-01 -2.76848197e-01 -3.97294074e-01
-2.47221962e-01 -5.39367616e-01 -4.85015899e-01 2.82982349e-01
3.27619731e-01 -1.65552050e-01 -1.32878339e+00 1.85152376e+00
1.23383470e-01 3.40566218e-01 3.52557898e-02 7.19421208e-01
6.34040415e-01 5.13713837e-01 1.31589547e-01 -6.17969781e-02
1.31750166e+00 -1.09156501e+00 -1.09714352e-01 -3.28259259e-01
6.77157521e-01 -5.91077268e-01 1.77578104e+00 1.73672020e-01
-9.10145521e-01 -6.82976425e-01 -1.20372331e+00 -2.25475011e-03
-5.36408067e-01 -2.74768710e-01 6.90079391e-01 7.37105548e-01
-8.96154523e-01 8.59966993e-01 -7.67763495e-01 -7.04469740e-01
7.28003204e-01 3.71791840e-01 -8.44258219e-02 -2.12078258e-01
-9.03615534e-01 9.57582116e-01 2.48405010e-01 -5.08251011e-01
-1.24241030e+00 -1.11233020e+00 -5.98384559e-01 2.98026830e-01
5.16484499e-01 -8.17164838e-01 1.33498454e+00 -9.11634684e-01
-1.52373040e+00 8.05749357e-01 9.66760442e-02 -6.79823875e-01
1.30739748e-01 4.38274220e-02 -2.63496876e-01 5.95648624e-02
6.69414131e-03 8.50164115e-01 1.29315996e+00 -9.80815470e-01
-6.14694417e-01 -1.75949171e-01 1.68978512e-01 -4.29361202e-02
-5.87681353e-01 -7.30409175e-02 -5.05733311e-01 -5.74328482e-01
-4.70166504e-01 -8.05699944e-01 6.04479238e-02 -2.07566231e-01
-1.59268398e-04 -1.58711806e-01 6.95462167e-01 -7.16914684e-02
1.07177365e+00 -2.18399382e+00 -3.00524365e-02 -1.75837010e-01
1.52673379e-01 6.24335587e-01 -4.67711747e-01 2.53960878e-01
-1.42597511e-01 -3.54314074e-02 -3.29101712e-01 -4.69997406e-01
1.49646953e-01 3.26951027e-01 -4.21046346e-01 2.89906949e-01
4.11362588e-01 1.23686767e+00 -6.67816520e-01 -2.26957694e-01
2.77426660e-01 3.45994025e-01 -7.08639145e-01 -5.86859183e-03
-2.45116875e-01 -1.15353689e-01 1.17857317e-02 4.89303768e-01
2.90909380e-01 -5.19703209e-01 -3.10968876e-01 -2.48053685e-01
3.65581848e-02 -3.95024382e-03 -8.97306263e-01 1.69256926e+00
-5.48316658e-01 5.33674300e-01 -5.59011996e-01 -1.10107851e+00
5.91444492e-01 8.87634885e-03 5.11261970e-02 -9.36624408e-01
3.07043254e-01 1.50778428e-01 9.89343971e-02 -4.01344925e-01
3.83088917e-01 -5.15047312e-01 -3.12585719e-02 6.09741628e-01
7.82820702e-01 -7.27590322e-02 1.93002388e-01 2.36906275e-01
1.53802013e+00 -9.82066542e-02 2.55655587e-01 -2.59528428e-01
-3.62730846e-02 1.59423500e-01 4.96432781e-01 9.63750780e-01
-4.13206369e-01 7.68310249e-01 3.44852328e-01 -1.92976341e-01
-1.33367479e+00 -9.75667775e-01 -2.20698133e-01 1.51244783e+00
-2.24968314e-01 -4.00883526e-01 -8.29054058e-01 -7.12332666e-01
-4.18315828e-02 7.98049867e-01 -9.20848846e-01 -6.54994726e-01
-2.27062255e-01 -8.98115218e-01 6.86562598e-01 7.09054351e-01
3.96337777e-01 -9.61027265e-01 -9.17005360e-01 2.55301803e-01
3.49797636e-01 -1.02522171e+00 -5.88750184e-01 7.30430722e-01
-6.72253132e-01 -1.01205707e+00 -8.15163493e-01 -5.30269206e-01
3.80914152e-01 5.04716396e-01 1.08296895e+00 -2.44046096e-02
-4.54102546e-01 5.61174154e-01 -5.45004010e-01 -5.17134786e-01
-1.94073305e-01 3.27760726e-01 1.36043236e-01 -4.49104868e-02
5.31976163e-01 -7.11520791e-01 -5.88542521e-01 3.56186539e-01
-8.82177651e-01 -2.39408817e-02 6.49613559e-01 1.15386975e+00
3.21412653e-01 -1.23599134e-01 6.34629369e-01 -1.16818702e+00
4.50415820e-01 -3.76500487e-01 -5.50237834e-01 3.71679306e-01
-8.92952085e-01 3.61473680e-01 6.29610300e-01 -7.07653522e-01
-9.67241704e-01 -2.23187670e-01 4.89980541e-02 -8.25080276e-01
-4.50751595e-02 1.96116939e-01 2.37262309e-01 -2.35612169e-01
1.27645886e+00 2.14868143e-01 -8.45649168e-02 -3.02129686e-01
4.37849462e-01 5.15583754e-01 6.57028556e-01 -5.85515320e-01
9.71206069e-01 3.57564986e-01 -4.09628779e-01 -7.99238682e-01
-1.13292277e+00 -4.88323331e-01 -4.70652610e-01 3.73538546e-02
8.42757463e-01 -9.50450361e-01 -3.53561938e-01 3.89609039e-01
-6.14382267e-01 -9.19424713e-01 -7.53364861e-01 2.66310602e-01
-6.25322163e-01 -1.84651315e-01 -4.33261335e-01 -4.82618362e-01
-4.15196091e-01 -1.09166110e+00 9.13764358e-01 3.07937264e-01
-2.95395732e-01 -9.28746045e-01 2.02193022e-01 3.92447680e-01
7.83314943e-01 -5.08483090e-02 1.00940740e+00 -6.26589954e-01
-4.21965718e-01 2.06908658e-01 -3.77864063e-01 4.63636816e-01
-1.54779837e-01 -2.01441318e-01 -1.46262527e+00 -4.39638525e-01
-1.60626978e-01 -8.07248890e-01 1.21567607e+00 3.35991561e-01
1.01819634e+00 -1.50305688e-01 -8.28004777e-02 1.09362566e+00
1.69032717e+00 -1.54540343e-02 6.19053364e-01 3.95082414e-01
5.69315910e-01 1.19390607e-01 4.59901303e-01 4.39802587e-01
9.80773419e-02 6.89167082e-01 9.25806686e-02 1.67235598e-01
-5.74072719e-01 -4.02745046e-02 3.49780321e-01 5.17152131e-01
-2.24049374e-01 3.80678438e-02 -7.81317592e-01 6.10993028e-01
-1.59156120e+00 -1.06913733e+00 4.64215279e-01 2.26561642e+00
1.03052664e+00 4.33362752e-01 2.11304128e-01 2.21217591e-02
4.13360745e-01 1.18097551e-01 -9.26234901e-01 -4.41012770e-01
5.44765126e-03 9.14216518e-01 6.69864833e-01 2.05493778e-01
-9.54308510e-01 1.12573528e+00 5.95646095e+00 1.16113675e+00
-1.31202054e+00 4.01658505e-01 4.89374787e-01 -5.24390936e-01
-1.32737219e-01 5.09459786e-02 -1.11046600e+00 3.22028935e-01
1.24670506e+00 -2.05464989e-01 8.52209926e-01 9.36426580e-01
-3.16949129e-01 -4.30794433e-02 -1.27511477e+00 1.01780427e+00
3.19229364e-01 -1.67236698e+00 -1.61566123e-01 5.73135214e-03
9.79350209e-01 5.91337025e-01 2.56948024e-01 9.65742826e-01
4.54442561e-01 -9.05584335e-01 7.73092270e-01 1.59261838e-01
1.22909212e+00 -5.35862029e-01 5.29348969e-01 1.60419643e-01
-8.12577188e-01 -3.48195612e-01 -7.07530141e-01 -2.12274343e-02
-1.86703175e-01 2.91995645e-01 -8.21945786e-01 5.95998727e-02
8.38702440e-01 6.05703115e-01 -9.88054097e-01 8.55804801e-01
4.00497504e-02 9.00155306e-01 -8.23260397e-02 1.21511303e-01
3.88992935e-01 2.64977008e-01 2.71178305e-01 1.10303724e+00
1.72430173e-01 2.25353122e-01 -3.05410296e-01 8.03061604e-01
-1.62325710e-01 -3.10637444e-01 -4.09051895e-01 -1.82187647e-01
3.41181546e-01 1.22551513e+00 -7.37638414e-01 -5.19767880e-01
-5.12044907e-01 9.25923109e-01 5.92288315e-01 2.83267081e-01
-9.93897259e-01 -6.20050669e-01 6.65894628e-01 3.77342105e-02
7.25058496e-01 1.84000194e-01 -3.05482507e-01 -1.25902927e+00
-1.31068826e-01 -9.56445396e-01 4.82063234e-01 -7.34279394e-01
-1.37188506e+00 4.46988225e-01 2.44983118e-02 -1.11740041e+00
-1.93491086e-01 -5.21986663e-01 -8.54835749e-01 6.12274051e-01
-1.57472384e+00 -1.01000476e+00 -2.86545932e-01 7.10522711e-01
6.72074616e-01 -3.94196540e-01 8.93870175e-01 1.61933333e-01
-6.93551481e-01 1.08185101e+00 2.51524240e-01 -1.71381027e-01
9.16992486e-01 -1.14321494e+00 4.81966585e-01 7.37492561e-01
2.29693606e-01 6.89316392e-01 7.66072452e-01 -2.63635725e-01
-1.46774089e+00 -1.15970397e+00 3.47988129e-01 -6.16146386e-01
7.85342872e-01 -3.74928445e-01 -1.03063405e+00 6.45026863e-01
1.21023037e-01 2.70442963e-01 6.82972670e-01 3.37339938e-01
-7.31398046e-01 -2.71281958e-01 -1.01287782e+00 3.76013875e-01
1.18575191e+00 -6.34133875e-01 -7.24336803e-01 1.08443908e-01
9.30870891e-01 -7.45093822e-02 -7.51776874e-01 2.09807813e-01
5.74062824e-01 -1.04594588e+00 8.81258488e-01 -7.61670470e-01
5.39960146e-01 8.26858655e-02 -3.67430866e-01 -1.52344275e+00
-5.11656642e-01 -4.48694915e-01 -9.89008322e-02 1.26216090e+00
4.17356491e-01 -5.39872348e-01 7.15988874e-01 6.72678232e-01
-4.09459502e-01 -7.31208384e-01 -5.78219354e-01 -1.06194627e+00
8.81736502e-02 -5.65971911e-01 4.64982033e-01 9.15586591e-01
-1.98862210e-01 7.26566792e-01 -2.35511258e-01 -1.48406297e-01
5.92798889e-01 -1.56426013e-01 1.00430214e+00 -1.12015390e+00
-7.93860137e-01 -4.89352077e-01 -3.70392084e-01 -4.32975739e-01
-1.28869236e-01 -8.45087707e-01 8.15318972e-02 -1.16529870e+00
6.10743642e-01 -4.69999373e-01 -4.71244216e-01 8.67649794e-01
-1.66204542e-01 3.74950558e-01 3.57093692e-01 1.98504418e-01
-6.54566646e-01 6.19976699e-01 9.04507816e-01 -3.04647475e-01
-2.02526003e-01 -1.77016288e-01 -9.16008115e-01 4.65564549e-01
5.52213788e-01 -5.51121294e-01 -6.56240046e-01 -4.23655242e-01
1.57768466e-02 -3.72951299e-01 4.84466642e-01 -1.34821975e+00
2.09025249e-01 -1.62576512e-01 2.81207472e-01 -6.19346127e-02
5.15964329e-01 -4.11747456e-01 -2.43906766e-01 3.84745806e-01
-3.04913104e-01 -3.70875448e-01 3.91766429e-01 5.63605905e-01
1.94963187e-01 -4.58713531e-01 1.16168261e+00 -1.30899489e-01
-1.23912275e+00 2.95259774e-01 1.43238874e-02 5.41860998e-01
9.66493845e-01 -3.19530606e-01 -6.53958440e-01 -4.95715849e-02
-4.57423657e-01 -6.11338764e-02 5.80858946e-01 2.95969784e-01
3.26035231e-01 -1.14093542e+00 -4.53983009e-01 3.12211394e-01
6.02617085e-01 -2.64000595e-01 3.09854239e-01 7.89310098e-01
-1.72241163e-02 2.50992388e-01 -3.26134443e-01 -5.50630629e-01
-1.12254333e+00 6.91215038e-01 1.37622342e-01 -7.22308606e-02
-7.19094336e-01 1.15548897e+00 1.84113115e-01 -2.55413920e-01
2.60773391e-01 -2.66428798e-01 1.78883493e-01 1.35150626e-01
6.70063078e-01 2.19648570e-01 1.77155867e-01 -1.72908276e-01
-1.09986961e-01 5.93457341e-01 -2.19588667e-01 -1.09636597e-01
1.68354928e+00 1.14663012e-01 4.85907316e-01 5.63904285e-01
1.43217432e+00 -5.12538373e-01 -1.58920860e+00 -4.38533098e-01
-2.66085088e-01 -3.30361038e-01 1.75154418e-01 -8.62120688e-01
-8.43433201e-01 9.85539556e-01 6.88794255e-01 -1.31884143e-01
1.17702293e+00 1.10014684e-01 7.70683169e-01 6.23720944e-01
3.88392329e-01 -1.31931818e+00 3.12423259e-01 6.67009652e-01
4.64202493e-01 -1.46241331e+00 -1.59704626e-01 1.70625269e-01
-8.34660590e-01 7.89170504e-01 7.20061243e-01 -1.16285063e-01
4.71687943e-01 1.02715887e-01 -1.41230226e-01 -3.32705587e-01
-1.12274504e+00 -4.40464169e-01 2.22407088e-01 6.74699128e-01
1.03118584e-01 -1.84164360e-01 2.11765602e-01 5.56886435e-01
-3.51920873e-01 1.03650883e-01 3.62208843e-01 7.68020034e-01
-5.85173428e-01 -7.95691848e-01 -8.05856362e-02 7.65698135e-01
-5.94107024e-02 -2.62667537e-01 2.43832078e-02 8.01842809e-01
2.19307840e-01 6.77022755e-01 1.60518184e-01 -5.02108932e-01
4.42549765e-01 3.12653571e-01 6.78455114e-01 -1.00123096e+00
-6.26247346e-01 -2.72438139e-01 -7.40433037e-02 -5.74361086e-01
-2.12268412e-01 -6.46264851e-01 -9.30304348e-01 -2.51943916e-01
-3.20562273e-01 -4.36818339e-02 4.94777828e-01 1.14962959e+00
4.76773441e-01 6.43180847e-01 4.81809676e-01 -8.46759319e-01
-9.96162713e-01 -1.03520453e+00 -5.42453945e-01 4.99661624e-01
1.32700488e-01 -7.85795212e-01 -4.56300050e-01 1.43732607e-01] | [9.876113891601562, 2.8590681552886963] |
2824fc42-65a2-4fc0-9a13-68b04b11a871 | self-attention-amortized-distributional | 2301.04791 | null | https://arxiv.org/abs/2301.04791v2 | https://arxiv.org/pdf/2301.04791v2.pdf | Self-Attention Amortized Distributional Projection Optimization for Sliced Wasserstein Point-Cloud Reconstruction | Max sliced Wasserstein (Max-SW) distance has been widely known as a solution for less discriminative projections of sliced Wasserstein (SW) distance. In applications that have various independent pairs of probability measures, amortized projection optimization is utilized to predict the ``max" projecting directions given two input measures instead of using projected gradient ascent multiple times. Despite being efficient, Max-SW and its amortized version cannot guarantee metricity property due to the sub-optimality of the projected gradient ascent and the amortization gap. Therefore, we propose to replace Max-SW with distributional sliced Wasserstein distance with von Mises-Fisher (vMF) projecting distribution (v-DSW). Since v-DSW is a metric with any non-degenerate vMF distribution, its amortized version can guarantee the metricity when performing amortization. Furthermore, current amortized models are not permutation invariant and symmetric. To address the issue, we design amortized models based on self-attention architecture. In particular, we adopt efficient self-attention architectures to make the computation linear in the number of supports. With the two improvements, we derive self-attention amortized distributional projection optimization and show its appealing performance in point-cloud reconstruction and its downstream applications. | ['Nhat Ho', 'Dang Nguyen', 'Khai Nguyen'] | 2023-01-12 | null | null | null | null | ['point-cloud-reconstruction'] | ['computer-vision'] | [ 1.59973744e-02 -1.29012480e-01 7.57099465e-02 -6.19666636e-01
-1.01760745e+00 -4.27489758e-01 4.16147232e-01 -1.37801602e-01
-6.80140853e-01 6.47502422e-01 1.74138650e-01 -5.56234062e-01
-4.73185629e-01 -8.44377220e-01 -8.62002552e-01 -8.93500924e-01
1.28268555e-01 6.26831293e-01 2.99610049e-02 1.57274365e-01
3.53220224e-01 4.13123220e-01 -1.29338074e+00 -4.67446074e-02
1.08361351e+00 9.09786105e-01 4.65504795e-01 5.99982679e-01
-3.80085051e-01 -1.04420729e-01 -1.35752689e-02 -5.04867315e-01
6.76864564e-01 -3.59071225e-01 -6.78903878e-01 -3.75401080e-01
5.10546505e-01 -4.02081579e-01 -9.28656533e-02 1.37193763e+00
5.17312050e-01 3.21202487e-01 1.03736508e+00 -1.43047094e+00
-1.05366194e+00 6.50113642e-01 -9.85902965e-01 3.53764683e-01
-1.64482743e-02 3.37896869e-02 1.46637428e+00 -1.41085732e+00
2.73239255e-01 1.26707971e+00 7.09464133e-01 3.04682195e-01
-1.40946889e+00 -4.96218145e-01 2.57784873e-01 3.50198388e-01
-1.43417728e+00 1.59176484e-01 5.94209015e-01 -2.96121120e-01
8.77509117e-01 6.02415740e-01 6.25296354e-01 7.58389711e-01
8.87363702e-02 9.38432813e-01 9.23133135e-01 -1.14532597e-01
2.96667337e-01 2.15410843e-01 1.91926450e-01 5.33271194e-01
4.86137569e-01 -3.45999897e-01 -2.85620719e-01 -3.01167399e-01
4.60683316e-01 4.29657757e-01 -3.51654977e-01 -5.04859984e-01
-1.07699382e+00 8.61348510e-01 3.83489937e-01 -6.11864887e-02
-2.83984214e-01 2.43037656e-01 1.52871951e-01 2.00481638e-01
4.44289029e-01 3.49676818e-01 -3.89174044e-01 -2.18119904e-01
-9.99046445e-01 3.81657273e-01 4.19257015e-01 9.74781394e-01
9.87466753e-01 -1.76680163e-01 -3.07209194e-01 6.94455504e-01
5.17183959e-01 1.05215406e+00 6.37264490e-01 -8.74589384e-01
6.71839356e-01 5.04432082e-01 1.56020135e-01 -1.10250998e+00
-2.85573810e-01 -1.65816844e-01 -8.35650086e-01 3.32199931e-02
3.91403377e-01 1.61226895e-02 -8.13950956e-01 1.92978668e+00
3.21859926e-01 -8.68879929e-02 -2.89594531e-01 1.06087995e+00
1.07378714e-01 8.17959189e-01 -2.43442997e-01 -1.30290136e-01
1.06501722e+00 -6.96453333e-01 -4.27876353e-01 5.78373000e-02
6.55958354e-01 -5.61059415e-01 1.67388296e+00 1.12718262e-01
-9.83054399e-01 -2.51930445e-01 -1.00526798e+00 -2.53737420e-01
-2.04507828e-01 -3.43883425e-01 2.44941920e-01 6.32155776e-01
-8.68319273e-01 9.33655143e-01 -9.59172666e-01 -2.08876491e-01
5.71221888e-01 1.00947410e-01 -2.35697493e-01 -1.34356096e-01
-8.65306735e-01 6.79002762e-01 6.13674335e-02 4.53853533e-02
-6.12679422e-01 -9.75254536e-01 -6.07662082e-01 2.98555374e-01
8.34868923e-02 -5.61880708e-01 7.22462654e-01 -2.49798223e-01
-1.35011661e+00 5.36047876e-01 -1.87419802e-01 -4.56055880e-01
7.62845933e-01 -4.19055015e-01 5.37712984e-02 -9.93787572e-02
2.15384007e-01 3.01357329e-01 6.64016306e-01 -5.99289834e-01
-2.75529504e-01 -7.67054796e-01 -2.11396530e-01 3.78191233e-01
-7.63436079e-01 -2.65593737e-01 4.13004272e-02 -4.97955680e-01
5.49730599e-01 -9.57060039e-01 -1.67406842e-01 1.49434000e-01
-6.63689554e-01 -2.82964408e-01 7.39371419e-01 -3.08180362e-01
1.27226329e+00 -2.16359472e+00 1.63183257e-01 1.40642539e-01
9.32957083e-02 1.87890694e-01 -2.31881544e-01 3.24167192e-01
-4.29967046e-02 2.14424431e-01 -6.56472743e-01 -4.51662064e-01
3.05287898e-01 1.91120803e-01 -4.60953265e-01 7.57226110e-01
2.44627520e-01 6.79261982e-01 -9.01634753e-01 -5.30459642e-01
1.37650948e-02 2.31417254e-01 -7.92253911e-01 1.59704834e-01
8.16666558e-02 -9.19257477e-02 -2.90460587e-01 2.16844216e-01
1.22983754e+00 -1.77361250e-01 -1.02983922e-01 -2.80528396e-01
-7.69471228e-02 2.45095212e-02 -1.19262493e+00 1.61529028e+00
-4.09703702e-01 3.94690305e-01 -2.90381283e-01 -1.04618120e+00
9.70248342e-01 -2.57997006e-01 5.98034382e-01 -3.56025755e-01
-6.55378997e-02 3.71957034e-01 -1.36244655e-01 -3.91782880e-01
4.52267826e-01 -3.65572661e-01 1.48938492e-01 8.59980047e-01
-1.52813509e-01 1.33565580e-02 2.26965383e-01 1.86830699e-01
1.02630043e+00 2.69432902e-01 -3.96456644e-02 -6.53107405e-01
4.10353541e-01 -2.54979342e-01 7.79991686e-01 5.29569149e-01
-2.52885073e-01 1.07644022e+00 5.20299256e-01 -4.03916389e-01
-1.10022569e+00 -1.61966598e+00 -4.00203943e-01 1.02791381e+00
2.21706569e-01 -3.28427881e-01 -5.22111952e-01 -9.53578949e-01
2.38231510e-01 9.49862480e-01 -5.62682390e-01 -1.67477265e-01
-4.44694698e-01 -1.11460805e+00 3.16666335e-01 7.23926067e-01
2.97602445e-01 -4.24842715e-01 -7.96539485e-01 -3.02695297e-02
5.16012125e-02 -4.49236095e-01 -9.83585954e-01 2.22529233e-01
-8.49494517e-01 -9.01529729e-01 -1.22480059e+00 -3.30667555e-01
8.09737027e-01 5.74750066e-01 6.61612153e-01 -7.51796007e-01
-5.21331988e-02 -2.07927148e-03 -1.13986768e-01 -4.74018991e-01
3.54930848e-01 -2.27373168e-01 3.08550388e-01 1.17944308e-01
6.75203443e-01 -7.42817879e-01 -8.47744226e-01 3.05284381e-01
-8.40194583e-01 -2.91169465e-01 5.45510411e-01 9.47800398e-01
9.94654655e-01 -2.83415079e-01 2.64167637e-01 -7.62567461e-01
6.46098495e-01 -6.23530865e-01 -5.58244765e-01 2.68487006e-01
-8.55146587e-01 5.49551964e-01 8.33811462e-01 -3.17555398e-01
-8.29477310e-01 -2.64834106e-01 7.37812445e-02 -9.24620211e-01
4.79746550e-01 1.99463233e-01 -1.47277236e-01 3.90754223e-01
4.72147673e-01 3.27993333e-01 8.84626359e-02 -4.72515345e-01
4.93559390e-01 6.48368478e-01 7.08922297e-02 -6.08798146e-01
5.38790286e-01 6.91634536e-01 1.79253340e-01 -6.37219429e-01
-8.73335838e-01 -4.34054852e-01 -3.71874362e-01 1.89562753e-01
8.61294508e-01 -4.34848577e-01 -6.64784789e-01 1.29580647e-01
-1.11054742e+00 7.22297058e-02 -6.01332784e-01 7.84821570e-01
-6.31329060e-01 4.88750815e-01 -1.96403325e-01 -1.12702262e+00
-5.49108267e-01 -1.08623827e+00 9.99766290e-01 -1.27409145e-01
-6.64473325e-02 -7.51535296e-01 3.43401015e-01 -1.69519544e-01
4.44485873e-01 9.78567675e-02 9.40722287e-01 -7.16176331e-01
-2.97609180e-01 -1.81145370e-01 -5.57432353e-01 5.22790968e-01
-1.12490222e-01 -1.48046836e-01 -6.96357489e-01 -3.38132828e-01
1.46809414e-01 -3.93663310e-02 9.13359165e-01 4.51405793e-01
1.45174611e+00 -4.57758367e-01 3.71036679e-02 9.27861869e-01
1.47749436e+00 -1.30476177e-01 2.34437153e-01 1.44926548e-01
7.90244937e-01 5.37841797e-01 5.01653314e-01 6.04640245e-01
3.89007092e-01 2.96748817e-01 4.59285259e-01 3.29978645e-01
1.47207692e-01 -4.09270555e-01 4.29889560e-01 9.96003211e-01
5.22974916e-02 -4.85657379e-02 -8.52929890e-01 5.91348469e-01
-1.77828884e+00 -1.01747262e+00 -1.60048783e-01 2.49276972e+00
3.46196771e-01 -1.64993610e-02 1.82655454e-01 2.47245729e-02
8.64101768e-01 2.35208973e-01 -8.68774772e-01 -6.81797981e-01
-1.73880368e-01 5.65207452e-02 6.17471635e-01 3.44442755e-01
-6.81696713e-01 3.51419508e-01 5.39907837e+00 8.94528687e-01
-9.01111960e-01 3.07931334e-01 5.42661726e-01 -5.47446489e-01
-9.70929682e-01 -1.87596500e-01 -7.24609196e-01 8.03165793e-01
6.60794497e-01 -3.80754977e-01 2.71434158e-01 1.07611895e+00
-4.91255373e-02 1.24953270e-01 -1.21407580e+00 1.10113084e+00
-3.38087268e-02 -1.03842485e+00 1.85879409e-01 2.85911560e-01
6.33955598e-01 3.53861660e-01 3.76862735e-01 2.05866128e-01
1.64322048e-01 -6.95311427e-01 7.28084087e-01 3.32442820e-01
7.58009911e-01 -8.52346897e-01 7.15581417e-01 3.00911874e-01
-9.79972839e-01 -8.11493322e-02 -9.78528678e-01 3.09224665e-01
3.45577747e-01 7.90272355e-01 -5.61293125e-01 3.29221368e-01
8.54149997e-01 4.70567882e-01 1.12790177e-02 8.31812322e-01
1.67849571e-01 3.65786344e-01 -8.00882876e-01 -3.99034023e-01
4.36679840e-01 -7.74825335e-01 8.75504851e-01 1.06668341e+00
9.52152729e-01 -2.89333701e-01 -1.11993633e-01 1.09772301e+00
1.97015688e-01 2.61303574e-01 -5.99595129e-01 -1.11366630e-01
5.50900996e-01 8.67350399e-01 -4.86681223e-01 -8.35949406e-02
-3.23356271e-01 9.60219562e-01 4.08709437e-01 3.11412007e-01
-8.94177318e-01 -6.65750980e-01 1.00768077e+00 3.73109490e-01
5.65833926e-01 -3.01087856e-01 -5.44364810e-01 -1.11249495e+00
3.20470452e-01 -1.90864414e-01 5.13164878e-01 -4.86138076e-01
-1.61425889e+00 6.82451129e-01 9.68360342e-03 -1.38346493e+00
2.31124967e-01 -5.62490404e-01 -7.54383206e-01 9.57307935e-01
-1.41151750e+00 -8.75642478e-01 3.84707972e-02 4.85822380e-01
4.13132727e-01 -3.52786779e-02 6.75300062e-01 2.42071867e-01
-6.87585831e-01 7.78043509e-01 4.82228935e-01 -3.55025560e-01
6.19745851e-01 -1.63094938e+00 4.01808202e-01 7.82649338e-01
2.38571793e-01 6.90732837e-01 8.15416276e-01 -4.41399783e-01
-1.57309473e+00 -1.20865762e+00 5.87800205e-01 -2.25607514e-01
7.68860459e-01 -1.89147577e-01 -8.65440667e-01 4.80084598e-01
-1.80499420e-01 2.19672769e-01 7.74482489e-01 1.20164491e-01
-5.56395710e-01 -4.07036245e-01 -1.34219980e+00 6.44032896e-01
1.17053568e+00 -4.47534055e-01 -5.60752928e-01 3.72991681e-01
7.60654807e-01 1.56591728e-01 -6.85687125e-01 4.16571289e-01
5.53885818e-01 -1.12345207e+00 8.03382754e-01 -6.69178367e-01
4.25745815e-01 -2.88526028e-01 -5.82789123e-01 -1.21214271e+00
-4.44542438e-01 -5.73741436e-01 -2.51728892e-01 9.43243861e-01
5.02533197e-01 -8.23402345e-01 9.43916738e-01 4.85828161e-01
-5.16882166e-02 -1.27536154e+00 -1.08144057e+00 -1.04345071e+00
4.06519711e-01 -5.97409248e-01 8.35877597e-01 6.20789409e-01
3.75795588e-02 7.76852220e-02 -3.42713892e-01 2.90264301e-02
8.25073779e-01 4.72337544e-01 5.24067640e-01 -9.63926852e-01
-4.51104492e-01 -5.35983205e-01 -5.18102169e-01 -1.23950398e+00
-8.63091499e-02 -1.29560411e+00 7.93702155e-03 -1.33541310e+00
3.41600835e-01 -2.63671100e-01 -6.82443380e-01 4.78822738e-02
-3.22873443e-01 7.12998435e-02 2.07288101e-01 3.00419837e-01
-4.37692285e-01 9.40565050e-01 1.10807550e+00 -3.42445187e-02
-8.74301419e-02 6.29365072e-02 -6.37126684e-01 6.70045137e-01
6.11239910e-01 -6.01368546e-01 -5.64030528e-01 -7.33356774e-01
3.99071366e-01 -3.35076720e-01 -3.31134461e-02 -7.66939700e-01
1.38404399e-01 -2.19287470e-01 1.81370020e-01 -8.23700666e-01
4.18209493e-01 -6.77950799e-01 -3.03719550e-01 4.55420256e-01
-2.90254891e-01 5.25484800e-01 -3.46856892e-01 8.89382839e-01
6.68487400e-02 -2.98270017e-01 9.77798939e-01 3.34055386e-02
-8.57432559e-02 6.76378191e-01 -2.23380495e-02 1.41560957e-01
1.11976767e+00 -2.81574905e-01 -2.35024132e-02 -1.36647820e-01
-3.19391727e-01 1.92663878e-01 3.77641261e-01 -5.45838475e-02
9.11656380e-01 -1.56293392e+00 -7.99021065e-01 3.72010857e-01
6.72435313e-02 2.50409275e-01 3.86024743e-01 1.06163430e+00
-4.16259140e-01 2.25679919e-01 -1.41145751e-01 -6.59239709e-01
-7.66863286e-01 7.02877879e-01 -1.72899216e-01 -1.32356301e-01
-6.57986104e-01 1.16652024e+00 3.05570900e-01 -5.59880972e-01
-1.14671797e-01 -2.98214942e-01 3.40303183e-01 8.88999328e-02
5.97929418e-01 8.45916748e-01 -7.68845603e-02 -3.15882802e-01
-4.31007177e-01 6.20061576e-01 -1.22111119e-01 -2.78083980e-01
1.49389601e+00 -2.57063299e-01 -2.93452069e-02 6.30818725e-01
1.74767005e+00 -8.72141942e-02 -1.43274963e+00 -9.98478532e-02
-8.74080360e-02 -8.50438237e-01 -1.81056093e-02 -4.81863543e-02
-8.81288946e-01 1.23521531e+00 5.00233173e-01 3.15249771e-01
8.71850193e-01 -2.66809225e-01 1.16361725e+00 4.32296693e-01
1.62552133e-01 -1.24072707e+00 -1.03054263e-01 3.72745931e-01
9.47368383e-01 -1.04957116e+00 -1.81589406e-02 2.70548221e-02
-5.88545203e-01 9.45651710e-01 6.06576025e-01 -3.37843210e-01
8.76116276e-01 1.50167003e-01 -2.42146090e-01 2.57462114e-02
-3.94077480e-01 -2.44376771e-02 3.24190140e-01 3.41683209e-01
1.84352145e-01 9.87854749e-02 -6.22147858e-01 3.78425837e-01
-3.51984888e-01 -4.78559077e-01 2.24984288e-01 5.57772994e-01
-4.41103727e-01 -7.58661926e-01 -1.17867805e-01 7.53802240e-01
-2.38277197e-01 -1.07157253e-01 6.40940070e-02 3.88783336e-01
-2.15738237e-01 4.18653548e-01 3.09619606e-01 -5.07076621e-01
1.90457597e-01 3.20011564e-02 3.78799498e-01 -4.28909332e-01
1.20362721e-01 -1.32881403e-01 -4.88831371e-01 -6.25728488e-01
1.27404332e-02 -8.01497698e-01 -1.13629699e+00 -3.60069811e-01
-4.04465705e-01 1.42032936e-01 7.93289661e-01 8.16860676e-01
4.99563396e-01 2.73954645e-02 8.68121088e-01 -5.86968958e-01
-1.30617046e+00 -9.63992000e-01 -7.62009859e-01 5.82504869e-01
4.00743932e-02 -5.61484039e-01 -5.62427521e-01 -6.22956812e-01] | [7.638640403747559, 4.116988182067871] |
c7c845c7-fbb2-49db-9f9a-e509d32a2902 | two-stream-consensus-network-for-weakly-1 | 2010.11594 | null | https://arxiv.org/abs/2010.11594v1 | https://arxiv.org/pdf/2010.11594v1.pdf | Two-Stream Consensus Network for Weakly-Supervised Temporal Action Localization | Weakly-supervised Temporal Action Localization (W-TAL) aims to classify and localize all action instances in an untrimmed video under only video-level supervision. However, without frame-level annotations, it is challenging for W-TAL methods to identify false positive action proposals and generate action proposals with precise temporal boundaries. In this paper, we present a Two-Stream Consensus Network (TSCN) to simultaneously address these challenges. The proposed TSCN features an iterative refinement training method, where a frame-level pseudo ground truth is iteratively updated, and used to provide frame-level supervision for improved model training and false positive action proposal elimination. Furthermore, we propose a new attention normalization loss to encourage the predicted attention to act like a binary selection, and promote the precise localization of action instance boundaries. Experiments conducted on the THUMOS14 and ActivityNet datasets show that the proposed TSCN outperforms current state-of-the-art methods, and even achieves comparable results with some recent fully-supervised methods. | ['Gang Hua', 'Junsong Yuan', 'Qilin Zhang', 'Wei Tang', 'Le Wang', 'Yuanhao Zhai'] | 2020-10-22 | two-stream-consensus-network-for-weakly | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4602_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123510035.pdf | eccv-2020-8 | ['weakly-supervised-action-localization', 'weakly-supervised-temporal-action'] | ['computer-vision', 'computer-vision'] | [ 7.30902791e-01 1.31276190e-01 -7.33840346e-01 -3.24041009e-01
-8.15810978e-01 -5.44009916e-02 4.96474892e-01 -2.50762969e-01
-5.68084240e-01 7.67811179e-01 3.83765519e-01 1.61823004e-01
2.24657223e-01 -2.14323655e-01 -6.97853386e-01 -6.97107315e-01
3.01437220e-03 1.41642243e-01 8.87264073e-01 1.85481012e-01
2.14131415e-01 -8.27767625e-02 -1.42654228e+00 6.75925672e-01
7.13640273e-01 1.09406579e+00 2.45967641e-01 4.15330648e-01
2.02489913e-01 1.61741447e+00 -4.44158375e-01 -1.02297761e-01
2.18618676e-01 -6.83639765e-01 -8.78947496e-01 4.40547377e-01
6.59404993e-01 -5.50994515e-01 -4.04803067e-01 8.20792794e-01
2.96521366e-01 3.30382079e-01 1.77709743e-01 -1.44540358e+00
-2.48856992e-01 5.12768328e-01 -6.92054689e-01 4.24428135e-01
3.36001039e-01 5.29694557e-01 9.23898995e-01 -1.01221418e+00
7.68809080e-01 1.07488334e+00 5.41131914e-01 7.78618038e-01
-1.02199221e+00 -6.93591177e-01 6.18726790e-01 5.65843403e-01
-1.27889526e+00 -6.29850924e-01 7.43839920e-01 -2.54982650e-01
8.60082805e-01 -5.61629608e-02 7.96874702e-01 1.32812226e+00
5.13078757e-02 1.40840542e+00 8.62822652e-01 -1.72759369e-01
3.41996133e-01 -5.06752670e-01 -2.67158180e-01 8.45682740e-01
-1.39376119e-01 -1.25449806e-01 -9.55388904e-01 1.34731874e-01
9.77889359e-01 6.13429174e-02 -3.87397081e-01 -3.91418755e-01
-1.64630735e+00 3.02826047e-01 3.04858208e-01 3.31128269e-01
-6.15074992e-01 4.24743176e-01 4.57641363e-01 -3.24916601e-01
5.47350466e-01 2.43118227e-01 -4.78748590e-01 -5.58059394e-01
-1.19417536e+00 8.93599093e-02 5.87604456e-02 8.50048423e-01
5.57332695e-01 2.08383888e-01 -8.74267399e-01 5.03697097e-01
8.93573016e-02 1.40461773e-01 5.61237574e-01 -1.46187770e+00
6.63360298e-01 7.63487577e-01 4.06273156e-01 -7.51450181e-01
-1.80746809e-01 -3.36298347e-01 -6.94992006e-01 2.68337931e-02
4.35876667e-01 3.39885391e-02 -9.70946670e-01 1.69336104e+00
4.01425689e-01 9.77787733e-01 -1.59717187e-01 1.22361016e+00
5.24639308e-01 4.91856098e-01 3.55360895e-01 -3.31091821e-01
9.87201035e-01 -1.47363353e+00 -9.49438572e-01 -3.70872736e-01
7.26450324e-01 -3.34697336e-01 1.05529630e+00 3.37026387e-01
-1.11110377e+00 -8.25356066e-01 -8.42794597e-01 4.85565737e-02
3.25294673e-01 5.87306917e-01 5.82997262e-01 5.35038859e-02
-7.73302138e-01 5.90478480e-01 -1.24166906e+00 -2.29477361e-01
9.74094331e-01 5.78670874e-02 -4.78692621e-01 -6.75024390e-02
-1.03205526e+00 6.41503334e-01 5.48780441e-01 3.17471713e-01
-1.36526227e+00 -6.19976103e-01 -9.14632797e-01 -6.20178692e-02
8.06288183e-01 -4.22965258e-01 1.44762850e+00 -1.35004687e+00
-1.58126140e+00 5.82079053e-01 -3.97701591e-01 -7.71036565e-01
6.68226838e-01 -3.33498895e-01 -2.05886841e-01 4.64643091e-01
3.50100487e-01 1.07040596e+00 9.59828019e-01 -9.28019464e-01
-1.04112208e+00 -4.66750227e-02 1.87356874e-01 3.20303977e-01
-1.56194851e-01 1.46732554e-02 -7.76318014e-01 -7.32814729e-01
1.68371215e-01 -6.58272445e-01 -3.54574591e-01 2.72341996e-01
-2.18108714e-01 -5.32948017e-01 8.78705561e-01 -5.32217324e-01
1.34349287e+00 -2.12400460e+00 1.77622840e-01 -2.97855705e-01
1.97177172e-01 3.87722492e-01 -2.37180948e-01 -1.06718838e-01
-5.03537953e-02 -2.17921525e-01 -2.27535903e-01 -6.15902841e-01
-2.46221960e-01 3.09056729e-01 -1.02214590e-01 5.69810212e-01
5.75101316e-01 8.74116421e-01 -1.31410182e+00 -8.16075385e-01
4.86667812e-01 2.42524624e-01 -5.35061538e-01 1.73149154e-01
-4.62729275e-01 6.83710098e-01 -5.31248391e-01 8.51504803e-01
1.67153955e-01 -3.91492039e-01 1.19884714e-01 -3.55177999e-01
-1.42580464e-01 2.02825412e-01 -1.10538816e+00 2.01877642e+00
-9.00301635e-02 5.47836065e-01 -1.91923261e-01 -1.15384305e+00
5.17946482e-01 4.23185885e-01 7.90930927e-01 -6.98108554e-01
1.69509113e-01 -9.53354985e-02 -1.20173462e-01 -4.93050367e-01
2.95110822e-01 1.48952648e-01 1.56008065e-01 3.05356324e-01
3.31264794e-01 4.71689820e-01 4.43065166e-01 2.74579406e-01
1.47188556e+00 8.52654457e-01 2.63911188e-01 1.09900720e-01
8.32469821e-01 1.42935691e-02 1.31012976e+00 7.35449612e-01
-8.12764108e-01 6.78430676e-01 4.39834297e-01 -6.17568016e-01
-7.35723972e-01 -7.20770955e-01 2.87806988e-01 1.23772550e+00
4.31097716e-01 -5.32233417e-01 -8.39595377e-01 -1.19185436e+00
-4.61974651e-01 6.27838910e-01 -6.38576567e-01 -2.70091772e-01
-7.30335116e-01 -3.13125134e-01 4.76316839e-01 8.47659290e-01
8.76257718e-01 -1.30496228e+00 -7.59099007e-01 4.28394854e-01
-7.40841866e-01 -1.47311068e+00 -7.34553576e-01 -8.53162333e-02
-8.60642672e-01 -1.29008889e+00 -7.23073661e-01 -5.71274936e-01
8.69719565e-01 2.52229929e-01 8.88822496e-01 -2.96513038e-03
-9.27893370e-02 1.75727487e-01 -6.33202910e-01 -6.12362288e-02
-1.04488164e-01 -1.79440036e-01 7.82029927e-02 4.14372891e-01
3.96345377e-01 -2.86201775e-01 -7.54710913e-01 7.10651338e-01
-7.26351619e-01 3.64544183e-01 6.41225219e-01 7.59593725e-01
8.34462523e-01 -8.33659098e-02 5.29734850e-01 -5.70357800e-01
1.08852861e-02 -6.24112338e-02 -3.78829509e-01 3.13606590e-01
-2.85919696e-01 -1.03692718e-01 3.83278459e-01 -6.65406346e-01
-1.22795522e+00 5.04646480e-01 1.35967240e-01 -9.11223471e-01
-1.52855426e-01 1.35031372e-01 -1.16087750e-01 9.52365249e-02
5.39957583e-01 4.07259285e-01 -2.85552859e-01 -2.21804813e-01
2.69845784e-01 1.87134132e-01 7.92816520e-01 -3.56535971e-01
5.31723082e-01 6.90346658e-01 -3.42477679e-01 -4.03656304e-01
-1.33460474e+00 -5.08616447e-01 -8.64107430e-01 -5.87210298e-01
1.13186717e+00 -1.02592218e+00 -5.04067063e-01 5.66335261e-01
-1.23448455e+00 -6.78038359e-01 -4.98058349e-01 6.17997408e-01
-8.65047693e-01 4.04467374e-01 -4.69928294e-01 -7.76796520e-01
-1.73377499e-01 -9.64983046e-01 1.29679799e+00 1.12061471e-01
-2.45438024e-01 -5.42378128e-01 -3.24831903e-02 6.60237372e-01
1.60246715e-02 1.59957528e-01 1.06208935e-01 -2.94018507e-01
-7.50044167e-01 -1.28581733e-01 -1.15729079e-01 3.85592580e-01
1.98368564e-01 -1.02165677e-01 -7.02286899e-01 -5.42532951e-02
-2.39991859e-01 -5.59389472e-01 1.04743373e+00 5.12223184e-01
1.53359354e+00 -2.22172558e-01 -3.00304860e-01 4.83125240e-01
8.48447919e-01 6.17827997e-02 7.14575887e-01 1.56942308e-01
6.75030410e-01 1.91540033e-01 1.27930689e+00 6.19935870e-01
2.62441188e-01 8.20522904e-01 5.81408799e-01 -2.30220869e-01
-2.49011010e-01 -4.57602978e-01 5.59557259e-01 2.23305508e-01
-5.57550788e-01 -1.91758484e-01 -4.83701199e-01 6.34917140e-01
-2.35967875e+00 -1.41411889e+00 3.57266776e-02 2.04459834e+00
8.72608423e-01 3.73344332e-01 2.21813992e-01 1.28797621e-01
9.14659381e-01 4.69521970e-01 -7.10860133e-01 4.83380109e-01
-4.64898199e-02 2.86228936e-02 4.27882016e-01 2.54293263e-01
-1.49990904e+00 1.18276405e+00 5.79836941e+00 9.44177330e-01
-8.28637540e-01 4.68879104e-01 6.48691177e-01 -3.36317748e-01
3.30641717e-01 2.06704829e-02 -6.95837021e-01 4.69017118e-01
4.94944662e-01 5.69233187e-02 5.81896268e-02 9.25115466e-01
8.21067750e-01 -2.99631178e-01 -1.12902951e+00 9.66544867e-01
2.46419966e-01 -1.40498638e+00 -1.49748385e-01 -3.64650249e-01
8.89422178e-01 -8.39517117e-02 -2.54987866e-01 4.63053763e-01
2.43696719e-01 -5.39866507e-01 9.92025673e-01 4.82329190e-01
5.76407909e-01 -4.26581889e-01 6.93191350e-01 3.58487040e-01
-1.51948535e+00 -2.77410865e-01 -1.32425964e-01 -1.04100943e-01
5.44487357e-01 2.86830097e-01 -5.03089905e-01 3.88912916e-01
7.06803858e-01 1.37979007e+00 -4.42297161e-01 9.74541366e-01
-6.89357996e-01 7.77574003e-01 -5.77554591e-02 2.86237955e-01
4.77163941e-01 2.30850592e-01 3.90354484e-01 8.85076821e-01
1.22548819e-01 3.26268554e-01 4.89863783e-01 7.05827117e-01
-2.69948300e-02 -2.20447421e-01 -7.87332430e-02 -1.14876926e-02
1.58211634e-01 1.24327135e+00 -7.85059929e-01 -6.07435882e-01
-4.05145407e-01 1.21308231e+00 3.48398536e-01 3.71302813e-01
-1.33925700e+00 1.48024783e-01 5.17382622e-01 2.65429288e-01
3.45084608e-01 -5.33637814e-02 1.32444605e-01 -1.41326225e+00
1.07340008e-01 -7.24808276e-01 5.10136127e-01 -9.79688048e-01
-8.44679415e-01 2.37985700e-01 -1.30670488e-01 -1.66266811e+00
-3.82071431e-03 -1.80494711e-01 -6.80721998e-01 1.83137804e-01
-1.29030883e+00 -1.12699664e+00 -4.02907282e-01 6.39349222e-01
1.05388498e+00 2.48140525e-02 3.33820999e-01 3.99283826e-01
-8.23192954e-01 4.06941026e-01 -5.83045542e-01 2.49174356e-01
7.06145763e-01 -9.78587091e-01 4.89568152e-02 1.18165326e+00
1.95412129e-01 1.94432810e-01 4.63066548e-01 -8.41647089e-01
-1.01127887e+00 -1.57715237e+00 7.31936216e-01 -4.58359510e-01
5.88611901e-01 -1.31435588e-01 -7.75150478e-01 8.31115127e-01
-6.89004362e-02 5.20608187e-01 2.91624874e-01 -3.51142794e-01
-5.55675589e-02 -2.10885912e-01 -9.03220296e-01 4.88096178e-01
1.47780478e+00 -2.34101519e-01 -6.16502821e-01 5.48371673e-01
6.11862540e-01 -4.83647138e-01 -5.64579487e-01 5.22461832e-01
5.23733258e-01 -9.82875884e-01 7.62293220e-01 -5.90142608e-01
5.31641304e-01 -6.93912029e-01 1.98976353e-01 -8.18955839e-01
-3.91909003e-01 -6.27896726e-01 -4.52835828e-01 1.07592320e+00
1.59206763e-01 1.52263949e-02 1.08600962e+00 3.79825175e-01
-3.72952968e-01 -7.19139457e-01 -1.25175118e+00 -7.15077102e-01
-7.53797710e-01 -6.31988287e-01 1.26538187e-01 8.06684196e-01
7.89959654e-02 1.35610998e-01 -8.19475353e-01 1.10471047e-01
5.88080585e-01 -2.28372976e-01 7.04907358e-01 -7.28123784e-01
-1.26940206e-01 -2.11937517e-01 -6.35561824e-01 -1.34228098e+00
3.74117017e-01 -4.45568353e-01 5.12709975e-01 -1.44204438e+00
2.99956053e-01 -6.05887696e-02 -5.47335804e-01 9.86098647e-01
-2.60615408e-01 5.24581790e-01 1.10053673e-01 1.72561273e-01
-1.52462149e+00 8.95102620e-01 1.38083529e+00 -2.09939942e-01
-1.90263465e-01 9.35551375e-02 -2.28731364e-01 9.69384968e-01
6.01844788e-01 -5.17621934e-01 -5.24790943e-01 -3.28040063e-01
-2.11848140e-01 1.27973512e-01 5.49763203e-01 -1.32397842e+00
3.13581675e-01 -5.35162985e-01 2.78144121e-01 -8.83960068e-01
3.17757189e-01 -6.72810674e-01 -2.13074803e-01 5.48200846e-01
-4.85943288e-01 -3.57130796e-01 -1.96003586e-01 8.51114273e-01
-3.18954468e-01 1.10712335e-01 8.71311903e-01 -5.39121963e-02
-1.17256582e+00 6.14103258e-01 -3.85903001e-01 6.48977011e-02
1.40368652e+00 -3.23687047e-01 -2.22367257e-01 -2.52102435e-01
-7.62010038e-01 4.85202909e-01 2.57616401e-01 6.10126972e-01
6.97263598e-01 -1.46908784e+00 -6.07773185e-01 5.36268502e-02
3.32377374e-01 8.51486027e-02 4.05161351e-01 1.17887485e+00
-1.71282575e-01 2.97570974e-01 -1.14856020e-01 -7.27570534e-01
-1.21824098e+00 3.42943698e-01 3.94779116e-01 -2.48081639e-01
-7.47630835e-01 8.92819703e-01 2.30411410e-01 6.51172474e-02
4.29767579e-01 -5.50458729e-01 -2.07076982e-01 -2.39360899e-01
8.10343802e-01 3.67099166e-01 -2.34321296e-01 -7.42107153e-01
-4.91445929e-01 2.99609363e-01 8.53469372e-02 4.33353782e-02
1.14589667e+00 -1.78005900e-02 1.91215068e-01 2.59699762e-01
6.07579112e-01 -6.74951434e-01 -2.04979658e+00 -3.65736455e-01
-1.01679541e-01 -7.08491564e-01 4.96538095e-02 -6.27606750e-01
-1.22400713e+00 6.91233575e-01 4.65724915e-01 -4.16468024e-01
1.17220175e+00 -6.70454511e-03 7.49124527e-01 3.09901625e-01
4.31839079e-01 -1.31715119e+00 6.34561360e-01 3.55330139e-01
6.59778714e-01 -1.25089884e+00 -3.81895564e-02 -3.83660048e-01
-8.15535128e-01 7.76792765e-01 1.26883090e+00 -7.71328211e-02
1.19085282e-01 1.05229825e-01 -2.49199674e-01 3.53947505e-02
-9.84709501e-01 -2.97818482e-01 2.27915660e-01 5.66458583e-01
2.23651588e-01 -3.98304582e-01 -3.66284162e-01 5.72204888e-01
6.99381173e-01 6.28466606e-01 3.53945732e-01 9.74286497e-01
-5.32654524e-01 -8.54171157e-01 -1.39934942e-01 4.32534099e-01
-4.19962794e-01 1.44948915e-01 -2.35462457e-01 5.08149862e-01
4.25573051e-01 8.86050522e-01 5.07231876e-02 -3.86729538e-01
2.78910398e-01 4.05301750e-02 4.41609859e-01 -7.00104654e-01
-2.76166558e-01 1.19799785e-01 1.23200662e-01 -1.20896673e+00
-1.12515378e+00 -7.09180176e-01 -1.41845965e+00 2.28641853e-01
-3.83905232e-01 -3.01367007e-02 -3.71362194e-02 1.21661365e+00
3.33690524e-01 6.79111600e-01 4.29855198e-01 -1.20122516e+00
-1.94824785e-01 -1.00217807e+00 -3.83645684e-01 4.62002754e-01
1.76051259e-01 -9.39189494e-01 -3.34883392e-01 5.32901227e-01] | [8.449689865112305, 0.5860508680343628] |
37497341-a600-48ae-962f-c6f3d7220184 | unsupervised-domain-adaptation-through-shape | 2207.02529 | null | https://arxiv.org/abs/2207.02529v1 | https://arxiv.org/pdf/2207.02529v1.pdf | Unsupervised Domain Adaptation through Shape Modeling for Medical Image Segmentation | Shape information is a strong and valuable prior in segmenting organs in medical images. However, most current deep learning based segmentation algorithms have not taken shape information into consideration, which can lead to bias towards texture. We aim at modeling shape explicitly and using it to help medical image segmentation. Previous methods proposed Variational Autoencoder (VAE) based models to learn the distribution of shape for a particular organ and used it to automatically evaluate the quality of a segmentation prediction by fitting it into the learned shape distribution. Based on which we aim at incorporating VAE into current segmentation pipelines. Specifically, we propose a new unsupervised domain adaptation pipeline based on a pseudo loss and a VAE reconstruction loss under a teacher-student learning paradigm. Both losses are optimized simultaneously and, in return, boost the segmentation task performance. Extensive experiments on three public Pancreas segmentation datasets as well as two in-house Pancreas segmentation datasets show consistent improvements with at least 2.8 points gain in the Dice score, demonstrating the effectiveness of our method in challenging unsupervised domain adaptation scenarios for medical image segmentation. We hope this work will advance shape analysis and geometric learning in medical imaging. | ['Yongyi Lu', 'Alan Yuille', 'Wei Shen', 'Yan Wang', 'Zongwei Zhou', 'Fengze Liu', 'Yuan YAO'] | 2022-07-06 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [ 1.09940499e-01 4.18521941e-01 -4.61474210e-02 -6.81807578e-01
-8.99306059e-01 -6.69834375e-01 2.99977362e-01 4.15945947e-01
-4.39470679e-01 3.35232973e-01 2.27064908e-01 -1.77693844e-01
7.06662834e-02 -7.25964129e-01 -9.20389235e-01 -7.95928895e-01
1.70100942e-01 9.63089943e-01 1.38690054e-01 1.96738303e-01
-5.18398322e-02 4.55754906e-01 -7.28767574e-01 1.75975665e-01
1.15516627e+00 9.63815331e-01 2.21912086e-01 4.93218303e-01
-1.64963141e-01 4.57626462e-01 -1.67968526e-01 -4.39083725e-01
2.53275871e-01 -2.69505322e-01 -9.13097382e-01 3.82968545e-01
2.38020614e-01 -4.30298656e-01 -4.37446386e-02 1.05328739e+00
5.35922050e-01 -1.51042659e-02 1.03933418e+00 -6.57876849e-01
-3.66474330e-01 7.00628340e-01 -3.45663339e-01 4.55280729e-02
-2.62957484e-01 2.10255086e-01 9.16342139e-01 -5.30259848e-01
7.36883998e-01 8.05105031e-01 9.38857436e-01 4.75671113e-01
-1.46113145e+00 -1.76208332e-01 -1.50989518e-01 -2.18108237e-01
-1.04034090e+00 -1.95818871e-01 8.76909792e-01 -6.83017433e-01
7.59855032e-01 1.28401741e-02 7.77951300e-01 6.75988078e-01
3.14288110e-01 1.16069925e+00 8.20925891e-01 -2.41025299e-01
1.53581604e-01 1.98952869e-01 -7.32858479e-02 7.81889617e-01
1.47297382e-01 -2.76426002e-02 2.11779803e-01 -8.36317837e-02
9.89778638e-01 -1.71412975e-01 -1.31640255e-01 -9.39344525e-01
-1.11023223e+00 1.06496155e+00 7.42889822e-01 1.27803698e-01
-7.26577818e-01 6.01378232e-02 4.01337266e-01 -1.36365503e-01
4.25272226e-01 3.94332856e-01 -4.07085508e-01 2.17949480e-01
-9.51040328e-01 3.44644077e-02 7.19938338e-01 5.58031917e-01
5.20293295e-01 -1.75374582e-01 -2.83880681e-01 1.01426744e+00
5.93439639e-01 2.55492598e-01 3.75732988e-01 -9.98775005e-01
-2.07735747e-02 6.80503190e-01 -2.91886389e-01 -7.01043963e-01
-6.36002481e-01 -4.38822120e-01 -5.96756577e-01 2.16370866e-01
7.22156644e-01 -1.28668293e-01 -1.30281913e+00 1.53029466e+00
7.23272860e-01 8.84869248e-02 -4.09575664e-02 1.08847451e+00
1.04066432e+00 2.47340307e-01 2.28056535e-01 2.17701539e-01
1.30144393e+00 -8.41819525e-01 -4.51077849e-01 9.22569856e-02
6.44788325e-01 -7.89379001e-01 7.79142141e-01 4.76496786e-01
-1.13804305e+00 -1.31030291e-01 -8.29927087e-01 -5.81731647e-02
1.38386682e-01 1.53403074e-01 7.22345352e-01 6.96932077e-01
-7.74064720e-01 8.14221263e-01 -1.53902054e+00 -5.90907522e-02
1.03308642e+00 5.26802063e-01 -8.60708654e-02 1.52890747e-02
-6.18136346e-01 7.67552018e-01 3.62740904e-01 -1.33328304e-01
-7.25690663e-01 -1.04862821e+00 -9.24094737e-01 -4.79465723e-02
1.64080888e-01 -9.86366212e-01 1.16510892e+00 -1.10745001e+00
-1.56952810e+00 1.08347833e+00 2.42087081e-01 -7.56118655e-01
8.30619037e-01 8.79161060e-02 8.01536441e-02 4.43453580e-01
-5.80750667e-02 1.16900396e+00 7.32493579e-01 -1.25799775e+00
-1.89604446e-01 -3.79100293e-01 -3.49487573e-01 2.42166579e-01
-1.37982413e-01 -3.91706616e-01 -4.39688832e-01 -7.07376301e-01
2.26946056e-01 -9.44810569e-01 -4.62747335e-01 3.32424223e-01
-1.95956931e-01 -1.45036295e-01 4.18412447e-01 -9.82723296e-01
5.80820799e-01 -2.01670027e+00 8.57193321e-02 3.96333843e-01
2.28712857e-01 1.80482194e-01 3.41017134e-02 -3.75076115e-01
2.33555734e-01 -8.80910605e-02 -6.83793128e-01 -1.67538941e-01
-1.31778657e-01 5.29346645e-01 1.26632228e-01 6.13095641e-01
3.25360894e-01 1.21384656e+00 -7.21459329e-01 -7.89914310e-01
4.20372277e-01 7.00126767e-01 -1.07297671e+00 2.43908942e-01
-4.36766326e-01 9.40646768e-01 -4.35790062e-01 4.13846821e-01
7.76348650e-01 -4.08053368e-01 1.75983876e-01 -3.98932189e-01
2.70322233e-01 6.98632970e-02 -7.45606840e-01 1.95487273e+00
-2.74942487e-01 2.24444091e-01 2.36099482e-01 -1.15204871e+00
7.99216866e-01 3.84293705e-01 9.93934870e-01 -5.21457076e-01
3.80534649e-01 2.45606497e-01 1.77249104e-01 -4.65430945e-01
-4.02495861e-02 -4.93973434e-01 3.99734229e-01 2.24844694e-01
4.14666593e-01 -4.08710837e-01 -5.93416346e-03 -6.97297305e-02
6.09134316e-01 4.08257008e-01 6.45423532e-02 -4.66795534e-01
2.88657695e-01 2.54114687e-01 7.08918631e-01 4.26686376e-01
-4.17360753e-01 1.00575244e+00 4.19066906e-01 -5.09413004e-01
-1.17134392e+00 -1.12857282e+00 -5.36565483e-01 7.94566631e-01
4.69639972e-02 1.51587367e-01 -9.65451598e-01 -1.01133406e+00
1.02709373e-02 7.60831594e-01 -6.13078952e-01 4.66427989e-02
-6.98229969e-01 -9.58884120e-01 4.48625535e-01 7.69418836e-01
1.66415155e-01 -1.13402510e+00 -6.79356098e-01 4.16405112e-01
-1.91674396e-01 -1.00296628e+00 -5.75457454e-01 1.36668980e-01
-1.23932874e+00 -1.09756434e+00 -1.17257023e+00 -9.99655068e-01
8.74443233e-01 -4.61260140e-01 1.27913797e+00 -1.71642378e-02
-5.21548510e-01 4.89207059e-01 -7.91271180e-02 -4.87981737e-01
-6.87434196e-01 1.53313745e-02 -4.17484403e-01 -1.08324379e-01
2.69943863e-01 -3.47991139e-01 -9.45228815e-01 1.46162346e-01
-9.17322040e-01 -1.74015611e-02 6.32117867e-01 1.02604949e+00
9.29950833e-01 -3.20554048e-01 2.91834921e-01 -1.00312781e+00
1.95824534e-01 -4.24099594e-01 -7.17811525e-01 1.10294759e-01
-7.48895943e-01 2.40737244e-01 2.82769918e-01 -3.59826595e-01
-1.14033449e+00 4.31026280e-01 -5.75658679e-01 -5.53977668e-01
-3.70420754e-01 6.21891320e-01 1.27990633e-01 -8.82533789e-02
4.93647188e-01 1.89993650e-01 5.27747154e-01 -3.22756797e-01
1.03012770e-01 1.46076098e-01 4.13525432e-01 -5.30681252e-01
2.62423992e-01 5.49948931e-01 6.35550916e-02 -5.44415116e-01
-4.94402081e-01 -4.93901163e-01 -8.03377628e-01 -1.96519811e-02
1.13906312e+00 -7.73900807e-01 -5.17080307e-01 2.06910551e-01
-7.27674603e-01 -5.87612092e-01 -3.10623288e-01 6.78929865e-01
-6.83766484e-01 6.30915821e-01 -7.38983870e-01 -3.70326817e-01
-5.58070242e-01 -1.58666003e+00 1.01617169e+00 2.29469121e-01
-7.94441625e-02 -1.45069063e+00 1.48617744e-01 5.18137634e-01
4.44537878e-01 4.53007817e-01 1.02220118e+00 -9.39736426e-01
-4.10158068e-01 7.78962001e-02 -2.02010259e-01 4.07878578e-01
-5.31165861e-02 -2.09172279e-01 -7.12494433e-01 -2.82686263e-01
9.35410848e-04 -2.20535055e-01 1.10082412e+00 1.21103239e+00
1.23638546e+00 -1.27452269e-01 -2.44909465e-01 8.13560784e-01
1.33670521e+00 -7.94127360e-02 3.59398186e-01 8.14050883e-02
7.63500810e-01 4.66160238e-01 2.28189856e-01 3.54676217e-01
3.89494389e-01 4.09940988e-01 5.13562918e-01 -4.60179538e-01
-2.80736089e-01 -6.50062039e-02 -6.20181598e-02 7.52660513e-01
1.79078117e-01 1.11681506e-01 -1.17740631e+00 8.17538321e-01
-1.48250687e+00 -3.20269108e-01 -1.48861557e-01 1.93403816e+00
1.04754341e+00 4.49410044e-02 3.05982888e-01 -4.13145810e-01
4.39843655e-01 -3.11381578e-01 -7.12719917e-01 -1.33076295e-01
1.33276016e-01 2.41052404e-01 6.20105982e-01 4.61823046e-01
-1.41038918e+00 8.81773710e-01 5.92895079e+00 5.66175163e-01
-1.34920931e+00 -5.34977950e-02 8.23620439e-01 2.60388970e-01
-3.70685518e-01 -2.45125487e-01 -1.65924191e-01 2.13999689e-01
5.22880673e-01 2.65557587e-01 2.69284904e-01 8.29622746e-01
-9.66695882e-03 2.83766836e-02 -1.08973157e+00 7.48809934e-01
-1.57637566e-01 -1.21040285e+00 6.45051301e-02 1.26249388e-01
8.07171762e-01 2.79423803e-01 1.55221060e-01 2.11231887e-01
2.95600027e-01 -1.17345250e+00 4.77632374e-01 4.33472484e-01
4.79889989e-01 -6.61956370e-01 8.23603809e-01 1.49295390e-01
-7.44166195e-01 3.70772779e-01 -2.60999650e-01 6.01143777e-01
1.44886896e-01 6.24616861e-01 -1.48252237e+00 4.41132367e-01
5.00155985e-01 6.10488176e-01 -3.66339803e-01 1.33715141e+00
2.53697094e-02 9.39325154e-01 -5.25128305e-01 2.70590872e-01
2.56451279e-01 -2.79882640e-01 6.80917561e-01 1.20547283e+00
-1.20471213e-02 1.10915221e-01 3.42528701e-01 1.30897355e+00
-1.66348785e-01 2.80134976e-01 -1.57763422e-01 1.93965156e-02
-1.89879499e-02 1.29139996e+00 -1.15547311e+00 -2.96095669e-01
-2.75303930e-01 8.85632992e-01 7.60803977e-03 3.14938687e-02
-6.71238422e-01 8.74329284e-02 3.24229419e-01 -2.71660695e-03
5.98479986e-01 2.50973217e-02 -7.26599813e-01 -1.14001822e+00
-1.90974757e-01 -7.47329652e-01 6.12227261e-01 -2.47005701e-01
-1.47365785e+00 2.87891895e-01 -2.65434355e-01 -1.02507627e+00
-9.14274827e-02 -7.25085139e-01 -4.91791248e-01 5.75323403e-01
-1.59318578e+00 -1.29729640e+00 -1.44270122e-01 2.92703807e-01
4.77127671e-01 3.67879719e-02 8.23068678e-01 2.40856186e-01
-1.99063256e-01 6.47041142e-01 1.52681440e-01 6.03605151e-01
7.03105390e-01 -1.67879093e+00 9.08936560e-02 4.09714818e-01
2.63126910e-01 3.64491373e-01 6.23062909e-01 -6.19527161e-01
-1.15072727e+00 -1.02191794e+00 3.23869735e-01 -4.77817774e-01
2.13158071e-01 4.40789759e-02 -1.15200472e+00 6.97656870e-01
1.05259627e-01 2.91942447e-01 8.61362159e-01 -1.24045357e-01
-3.14680599e-02 3.42663616e-01 -1.57623315e+00 2.80453891e-01
4.79601055e-01 3.69878300e-02 -6.77643895e-01 2.60863930e-01
3.38716000e-01 -8.11247289e-01 -1.39129293e+00 7.53099859e-01
5.87796390e-01 -7.53289223e-01 1.10920835e+00 -5.79418659e-01
6.06356800e-01 5.35864476e-03 1.08228922e-01 -1.32586432e+00
-1.96113557e-01 -6.28283322e-02 1.49015868e-02 8.80833149e-01
4.54433262e-01 -3.17843169e-01 1.01810992e+00 7.45737135e-01
-3.00559163e-01 -8.97766054e-01 -7.53616810e-01 -3.40850621e-01
8.07508707e-01 -3.38737309e-01 3.32888812e-01 1.03416812e+00
-3.39664936e-01 -9.27044749e-02 1.71170637e-01 1.58784762e-01
8.96565676e-01 7.57120177e-02 5.32502294e-01 -1.33556545e+00
-3.91504407e-01 -8.48500788e-01 -2.67027557e-01 -1.00123560e+00
9.28104818e-02 -1.46674025e+00 1.54232979e-01 -1.70424366e+00
4.56389725e-01 -5.74332416e-01 -4.15711910e-01 4.24129725e-01
-2.30929315e-01 2.27858990e-01 -1.25046089e-01 1.24045379e-01
-5.70598066e-01 5.97643733e-01 1.50119746e+00 -2.74937719e-01
-2.62973577e-01 7.22411796e-02 -4.43028212e-01 7.87980139e-01
6.93939567e-01 -4.96986985e-01 -5.24632894e-02 -5.20596683e-01
-3.24690044e-01 1.04513660e-01 5.07475913e-01 -6.70795083e-01
9.09567401e-02 8.38854760e-02 7.57461190e-01 -5.20755529e-01
-7.60928839e-02 -8.67311954e-01 -2.46382251e-01 5.80329299e-01
-4.49756294e-01 -6.58125758e-01 3.32378805e-01 5.32317102e-01
-1.16013624e-01 -2.64432371e-01 1.10566163e+00 -1.87096089e-01
-4.30536330e-01 3.86557579e-01 -2.08724439e-01 3.18477571e-01
7.72818446e-01 -1.66073233e-01 5.53311229e-01 -1.60855085e-01
-1.08241510e+00 3.84819955e-01 3.53423268e-01 -1.08422413e-02
5.44842541e-01 -1.16882777e+00 -1.00316238e+00 2.70096421e-01
1.71791911e-02 4.81031448e-01 2.66996711e-01 1.17263305e+00
-8.40438783e-01 2.63979107e-01 -2.44625539e-01 -1.12235391e+00
-1.04829717e+00 2.82348096e-01 6.47029579e-01 -4.63666916e-01
-9.49084342e-01 9.76224005e-01 2.47325614e-01 -8.48867834e-01
1.92514539e-01 -5.70352077e-01 -1.99584097e-01 -1.55869424e-01
-1.76846031e-02 6.55884147e-02 1.41904846e-01 -5.28275490e-01
-3.01000744e-01 5.34525156e-01 -1.29050136e-01 -1.57843009e-02
1.58210289e+00 -1.87061657e-03 1.17712252e-01 5.33794574e-02
1.20329928e+00 -2.07990333e-01 -1.65047801e+00 -3.70393485e-01
2.18752712e-01 -1.59485087e-01 4.57324088e-01 -1.06816149e+00
-1.34526157e+00 8.81934106e-01 8.33332300e-01 -2.53896594e-01
1.01069224e+00 1.82252795e-01 1.03057015e+00 -2.18253173e-02
3.58290263e-02 -1.05016565e+00 2.90378258e-02 3.33021581e-01
5.98015666e-01 -1.62334549e+00 -3.48124914e-02 -3.33147138e-01
-8.77178371e-01 1.24429524e+00 2.56728470e-01 -3.06053400e-01
7.73908734e-01 2.92744428e-01 3.86909842e-01 -2.01827630e-01
-2.92266935e-01 -2.66073961e-02 7.92993009e-01 5.02965510e-01
6.37069046e-01 1.62101179e-01 -2.34821096e-01 6.97324634e-01
-4.69059423e-02 -1.92206502e-02 2.14996636e-01 6.70780778e-01
-3.14460188e-01 -1.07834923e+00 -2.09790170e-01 5.19209445e-01
-8.79316747e-01 9.08382833e-02 -8.54851753e-02 4.83701080e-01
-4.35619056e-02 3.64151329e-01 1.63506135e-01 2.42519423e-01
9.79303718e-02 1.12603419e-01 7.36454844e-01 -4.32115585e-01
-7.34607518e-01 5.41444480e-01 -4.18922991e-01 -3.68294984e-01
-1.64581865e-01 -8.35496366e-01 -1.65599096e+00 1.71414137e-01
-2.55243093e-01 -1.50267586e-01 6.90530062e-01 8.26668024e-01
1.42269090e-01 6.69302046e-01 1.93949685e-01 -7.50298917e-01
-7.01316893e-01 -8.24193478e-01 -3.28459352e-01 7.85594881e-01
1.93978220e-01 -5.23841381e-01 6.94721118e-02 2.38770127e-01] | [14.5833158493042, -2.4046952724456787] |
c70d70c8-517a-4ee4-b4bc-ac5a61d797f2 | future-video-synthesis-with-object-motion | 2004.00542 | null | https://arxiv.org/abs/2004.00542v2 | https://arxiv.org/pdf/2004.00542v2.pdf | Future Video Synthesis with Object Motion Prediction | We present an approach to predict future video frames given a sequence of continuous video frames in the past. Instead of synthesizing images directly, our approach is designed to understand the complex scene dynamics by decoupling the background scene and moving objects. The appearance of the scene components in the future is predicted by non-rigid deformation of the background and affine transformation of moving objects. The anticipated appearances are combined to create a reasonable video in the future. With this procedure, our method exhibits much less tearing or distortion artifact compared to other approaches. Experimental results on the Cityscapes and KITTI datasets show that our model outperforms the state-of-the-art in terms of visual quality and accuracy. | ['Jaesik Park', 'Rongrong Gao', 'Qifeng Chen', 'Yue Wu'] | 2020-04-01 | future-video-synthesis-with-object-motion-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Wu_Future_Video_Synthesis_With_Object_Motion_Prediction_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Wu_Future_Video_Synthesis_With_Object_Motion_Prediction_CVPR_2020_paper.pdf | cvpr-2020-6 | ['predict-future-video-frames'] | ['computer-vision'] | [ 1.60222903e-01 -3.08955330e-02 3.15282762e-01 -2.99011320e-01
-8.17384347e-02 -4.38087404e-01 8.06965530e-01 -6.34318352e-01
-6.32054284e-02 5.41384339e-01 6.78437129e-02 2.64409035e-02
4.02152210e-01 -4.86262172e-01 -8.35809231e-01 -6.41571820e-01
-2.14373901e-01 -2.71377042e-02 8.61861050e-01 6.85705021e-02
1.41778395e-01 5.12270451e-01 -1.73923862e+00 5.18971086e-01
4.97911900e-01 8.23947728e-01 3.97531360e-01 1.09995770e+00
1.54913679e-01 1.09193861e+00 -1.81081370e-01 -6.92140102e-01
6.67911112e-01 -5.78436732e-01 -7.69248426e-01 1.07326925e+00
7.82852530e-01 -8.20737362e-01 -9.75445092e-01 1.04748905e+00
-1.51401192e-01 2.33547449e-01 2.82181263e-01 -9.55908000e-01
-5.29568493e-01 -3.65579240e-02 -6.47712946e-01 2.93751627e-01
6.68281794e-01 3.58049691e-01 4.67800617e-01 -1.01771486e+00
1.09364951e+00 1.31534791e+00 3.72691989e-01 5.69216907e-01
-1.21166682e+00 -2.31241331e-01 7.01161385e-01 4.82280731e-01
-1.29159617e+00 -7.83836901e-01 9.07772005e-01 -4.46521819e-01
8.20366740e-01 2.60986745e-01 9.12861049e-01 1.06510401e+00
3.64558846e-01 7.94357300e-01 6.12114072e-01 -4.11323935e-01
3.61823440e-02 -6.09637126e-02 -2.82831788e-01 6.78920150e-01
6.27620593e-02 3.55220735e-01 -3.49868476e-01 9.60499942e-02
9.08645928e-01 2.25502580e-01 -3.63379270e-01 -4.52740192e-01
-1.46983778e+00 4.54026833e-02 -9.61668715e-02 1.50356647e-02
-5.98285317e-01 1.14405192e-01 -1.15070395e-01 1.93872377e-01
6.60372615e-01 -8.04252252e-02 -3.54532301e-01 -4.52525169e-02
-1.11992824e+00 3.58224750e-01 5.27047157e-01 1.11640215e+00
5.79175174e-01 4.74087358e-01 2.35164613e-01 2.56322265e-01
2.01985434e-01 5.12151182e-01 1.70240283e-01 -1.53075492e+00
4.32628095e-01 1.60644814e-01 5.53305268e-01 -1.21675813e+00
1.12216353e-01 1.74662806e-02 -5.82116544e-01 3.78748029e-01
4.36419636e-01 -2.27034330e-01 -1.14754224e+00 1.26466477e+00
4.12981153e-01 8.57338786e-01 2.32197233e-02 7.70879269e-01
4.17421609e-01 1.08618116e+00 -3.11639458e-01 -6.55134678e-01
7.90870249e-01 -1.23705792e+00 -8.91963720e-01 -3.68503898e-01
-2.64987350e-02 -9.56505239e-01 4.64838237e-01 6.65477037e-01
-1.56698811e+00 -9.89470482e-01 -8.92889738e-01 5.66767715e-02
5.80854826e-02 -5.40864803e-02 4.03940260e-01 2.90656179e-01
-1.26780128e+00 7.61074483e-01 -1.12062383e+00 -3.59739035e-01
4.50059064e-02 6.65331706e-02 -4.21745420e-01 4.34607789e-02
-5.64849317e-01 8.10115218e-01 4.01353210e-01 6.14221282e-02
-9.51346695e-01 -6.25759959e-01 -8.06312561e-01 -1.87901966e-02
3.11744928e-01 -7.11976409e-01 1.37618244e+00 -1.51424015e+00
-1.62109244e+00 7.01402307e-01 -5.17215669e-01 -4.65282917e-01
8.91607106e-01 -3.08752030e-01 -7.16971934e-01 4.06122744e-01
-4.05187279e-01 6.38986349e-01 9.60304260e-01 -1.52028525e+00
-1.10427618e+00 2.79103797e-02 1.44356325e-01 4.17658299e-01
-1.75251756e-02 -1.13294244e-01 -1.10084105e+00 -6.72440529e-01
1.80784449e-01 -1.10868883e+00 -3.15891743e-01 2.71287054e-01
1.49451494e-01 4.89079177e-01 1.27887809e+00 -8.14742148e-01
1.16546762e+00 -2.17722631e+00 1.72243923e-01 -3.10175210e-01
1.71424299e-01 2.13405222e-01 -2.71042466e-01 2.62319058e-01
-3.00925732e-01 -2.16442004e-01 2.85073310e-01 -4.38852161e-01
-5.52303016e-01 2.40595683e-01 -6.32681370e-01 6.97452426e-01
1.65846288e-01 6.52073145e-01 -1.12126744e+00 -4.81252939e-01
6.89888418e-01 5.59969306e-01 -5.67846358e-01 4.67058778e-01
-3.41907620e-01 6.51335478e-01 -3.18320245e-01 4.58650470e-01
8.19686651e-01 -2.76847064e-01 2.81272441e-01 -1.27187267e-01
-1.44778386e-01 -3.09066772e-02 -1.26002371e+00 1.59423912e+00
2.23907623e-02 7.91391432e-01 -1.55628994e-01 -6.21812522e-01
6.05543315e-01 4.42899317e-01 6.62118256e-01 -5.13254464e-01
-1.77280098e-01 -2.61442125e-01 -8.46987963e-02 -8.94885361e-01
5.66560090e-01 9.37560052e-02 4.51541036e-01 4.76391688e-02
-8.90226737e-02 -7.95645565e-02 3.79415512e-01 2.09739879e-01
6.86701179e-01 7.32351840e-01 1.52751684e-01 6.55374154e-02
5.88706315e-01 -4.52823043e-02 7.20319033e-01 3.01490307e-01
-2.24533796e-01 9.52489138e-01 6.12008162e-02 -9.80898738e-01
-1.37166810e+00 -1.22405696e+00 3.71778697e-01 5.78810334e-01
4.91559923e-01 -2.95250267e-01 -7.60488331e-01 -3.64928335e-01
-4.67724562e-01 7.22102821e-01 -4.98137444e-01 1.21417888e-01
-8.78031790e-01 -4.17228818e-01 -2.86818504e-01 3.27745467e-01
4.45416689e-01 -8.28425288e-01 -7.15295792e-01 2.98410147e-01
-4.87893999e-01 -1.64861429e+00 -7.39521027e-01 -7.14576483e-01
-9.90094960e-01 -1.01055467e+00 -6.90515876e-01 -7.63873160e-01
7.95962274e-01 8.46792161e-01 1.09998417e+00 1.34627894e-01
-1.61194369e-01 5.30990303e-01 -2.06059977e-01 -8.58842582e-02
-5.76487958e-01 -9.07359660e-01 1.93967745e-01 5.83874166e-01
-1.22478522e-01 -5.93315125e-01 -9.19330478e-01 1.52589560e-01
-9.21072721e-01 7.83570528e-01 2.33981565e-01 5.81634223e-01
6.39899254e-01 4.57504392e-01 -3.83487552e-01 -5.81317127e-01
-3.24790448e-01 -2.58299470e-01 -8.60500932e-01 4.17512774e-01
-2.70041049e-01 -2.86193430e-01 8.58050823e-01 -7.85797417e-01
-1.73594320e+00 4.60372061e-01 3.10427606e-01 -7.57101536e-01
-2.33985573e-01 -1.28975511e-01 2.04416271e-02 2.10371077e-01
2.62072206e-01 3.69861275e-01 -2.26628169e-01 -4.47017342e-01
3.14098358e-01 1.64894402e-01 9.94756520e-01 -2.63586819e-01
8.84503067e-01 9.65553522e-01 -9.61789936e-02 -9.52355623e-01
-6.55531466e-01 -2.98805714e-01 -1.02025819e+00 -6.86887681e-01
8.93861771e-01 -1.06283116e+00 -4.15810376e-01 6.64670587e-01
-1.43171895e+00 -2.32066840e-01 -1.63504675e-01 5.06733179e-01
-6.42874599e-01 6.55876458e-01 -5.96574605e-01 -8.90086830e-01
9.29312110e-02 -1.09587467e+00 9.31275487e-01 2.94519097e-01
-1.48844533e-02 -9.83764231e-01 1.91044305e-02 6.27668202e-02
-3.91464084e-02 3.87479424e-01 3.95272285e-01 9.85230282e-02
-1.09233546e+00 -1.20385088e-01 5.50112389e-02 2.28246734e-01
1.58274591e-01 7.23535359e-01 -8.33186984e-01 -3.57452482e-01
2.33304948e-01 4.94861990e-01 5.24966002e-01 5.09811819e-01
1.09200096e+00 -5.00983834e-01 -3.52580577e-01 8.35931301e-01
1.58406246e+00 7.31338024e-01 8.38159919e-01 2.98327029e-01
6.55425370e-01 6.43144608e-01 7.62292564e-01 4.34860379e-01
3.95016849e-01 7.44242668e-01 4.66609418e-01 -1.86345100e-01
-3.98575157e-01 -2.82920510e-01 6.34862781e-01 7.22957313e-01
-4.68382806e-01 -4.03461993e-01 -7.84306526e-01 6.17821813e-01
-2.01001716e+00 -1.51939964e+00 -3.22674721e-01 2.01268792e+00
1.80861339e-01 5.16970418e-02 -4.89162579e-02 -1.76414281e-01
6.91351652e-01 2.11086631e-01 -5.14563620e-01 -1.08107157e-01
-2.23784015e-01 -4.45849776e-01 3.71549010e-01 6.02867544e-01
-1.08776569e+00 8.31173241e-01 7.68385029e+00 1.86231062e-01
-1.10206199e+00 -1.92571118e-01 8.91825497e-01 -1.70217037e-01
8.52822959e-02 2.57373005e-01 -6.54130638e-01 4.34751660e-01
9.39216375e-01 -4.00958240e-01 6.11632645e-01 6.26511395e-01
5.60966492e-01 -1.32214904e-01 -1.14049888e+00 9.60494757e-01
2.01676190e-01 -1.56788445e+00 2.22503558e-01 -1.07049204e-01
1.10214913e+00 -5.54113425e-02 1.25128955e-01 -2.06295326e-01
2.66313583e-01 -4.86008018e-01 1.36491442e+00 9.53720093e-01
5.32486260e-01 -4.15301740e-01 3.35646361e-01 3.38521749e-01
-1.23173392e+00 -1.67701468e-01 -2.67333210e-01 -3.02011669e-01
6.48072600e-01 1.76163241e-01 -4.46098983e-01 4.96863931e-01
9.00083184e-01 9.73507881e-01 -6.65628731e-01 9.70957279e-01
1.03657283e-01 3.63877147e-01 7.23167434e-02 6.59489751e-01
5.45168109e-02 -5.66694736e-01 6.36291087e-01 1.12270987e+00
5.10531485e-01 5.65876901e-01 2.86166817e-01 5.50955534e-01
3.29728931e-01 -3.12079519e-01 -6.53947532e-01 2.45597407e-01
2.57790908e-02 9.69994485e-01 -7.10126936e-01 -8.44971657e-01
-8.27677727e-01 1.41911376e+00 -2.55926475e-02 7.81318128e-01
-9.74801660e-01 3.30945730e-01 7.05606163e-01 3.72412354e-01
6.42716408e-01 -4.17904943e-01 1.44970015e-01 -1.71279740e+00
2.65846491e-01 -8.08888555e-01 1.81227520e-01 -1.18484175e+00
-9.28060710e-01 8.31665337e-01 1.89424545e-01 -1.77553165e+00
-4.94168758e-01 -5.59399664e-01 -6.63292170e-01 4.43481386e-01
-1.37210786e+00 -1.12443602e+00 -2.06087351e-01 5.10279715e-01
1.14366221e+00 1.34132355e-01 4.92976755e-01 1.46611286e-02
-3.55720580e-01 -1.59212649e-01 2.59612888e-01 -4.12216410e-02
4.86820161e-01 -8.62477958e-01 8.60147536e-01 1.54536140e+00
3.47782016e-01 2.66683638e-01 1.13733292e+00 -7.04026699e-01
-1.41602898e+00 -1.03783977e+00 7.77618945e-01 -5.01466393e-01
6.35470927e-01 -1.55736446e-01 -8.57205629e-01 9.18605924e-01
3.29265624e-01 3.05337965e-01 7.62620047e-02 -7.25404322e-01
-2.66813543e-02 -1.80932075e-01 -7.33950138e-01 9.50501561e-01
1.14177597e+00 -2.70596653e-01 -5.69558620e-01 2.07618371e-01
7.58555233e-01 -6.94861591e-01 -5.63055694e-01 2.73680031e-01
8.79724026e-01 -1.23087776e+00 1.26799524e+00 -5.57838023e-01
4.42254514e-01 -5.38092196e-01 -1.61233306e-01 -9.59375739e-01
-5.49331546e-01 -9.03663933e-01 -4.86115813e-01 8.75550926e-01
3.65576483e-02 -1.24612369e-01 8.46779227e-01 1.08498144e+00
1.26972541e-01 -3.72402072e-01 -5.82865894e-01 -7.65642762e-01
-3.66809249e-01 -2.59382725e-01 4.31539267e-01 6.97847843e-01
-2.27761567e-01 1.40660116e-02 -9.88134742e-01 4.34366107e-01
7.22825408e-01 2.80817479e-01 1.09251583e+00 -9.38694656e-01
-4.84912992e-01 -4.83519472e-02 -5.23100734e-01 -1.28076899e+00
2.67569255e-03 -1.26362175e-01 -9.93024558e-02 -1.26647139e+00
2.55435228e-01 3.22942227e-01 1.75893717e-02 -1.24764889e-01
-2.55395770e-01 1.63815513e-01 4.53974128e-01 2.12576613e-01
-7.75937080e-01 3.87556493e-01 1.53312254e+00 -4.95281070e-02
-2.29279548e-01 -1.12815663e-01 -6.35698736e-02 1.29201376e+00
3.93244952e-01 -2.08620474e-01 -5.46692193e-01 -7.55814373e-01
-2.98540920e-01 5.51855803e-01 3.94032151e-01 -1.02856147e+00
2.04724342e-01 -5.82389832e-01 7.84830749e-01 -8.10658038e-01
6.85686648e-01 -1.10962820e+00 9.47209835e-01 4.26329941e-01
-4.68166620e-02 4.39334512e-01 1.30402014e-01 7.90645242e-01
-1.39219508e-01 2.28469804e-01 7.93150663e-01 -2.17344269e-01
-1.09891784e+00 5.36108553e-01 -4.85487759e-01 -4.18292463e-01
1.29545939e+00 -7.01377213e-01 -7.76979849e-02 -6.39348567e-01
-9.24923241e-01 -2.58754361e-02 8.74850333e-01 7.20025599e-01
8.67582679e-01 -1.16389143e+00 -6.72673881e-01 3.72906536e-01
-2.39263967e-01 -2.08317876e-01 5.19276738e-01 5.50079226e-01
-8.57640564e-01 3.02253455e-01 -2.36522734e-01 -6.37474000e-01
-1.64624202e+00 9.48032916e-01 3.10624033e-01 -6.99239150e-02
-1.02991152e+00 4.61086422e-01 7.36512959e-01 4.25350040e-01
2.16149226e-01 -2.78688759e-01 -6.69911280e-02 -4.30307925e-01
9.30999100e-01 3.93671930e-01 -3.25774938e-01 -1.13999927e+00
-4.47671413e-02 5.18195391e-01 -1.03583105e-01 -2.45554522e-01
1.54689896e+00 -8.76044929e-01 1.51020050e-01 4.68760252e-01
9.65947628e-01 -1.45014420e-01 -2.06736636e+00 -1.81265980e-01
-1.91276476e-01 -1.06023526e+00 -1.50092542e-01 -4.61050421e-01
-1.17773557e+00 5.54900587e-01 5.06502867e-01 2.20603459e-02
1.35393918e+00 -2.62280613e-01 7.26995707e-01 6.82248399e-02
5.02504945e-01 -8.62371743e-01 5.18796258e-02 2.88727999e-01
8.31517220e-01 -1.02267051e+00 2.82181650e-01 -5.14556229e-01
-7.81041384e-01 1.34726012e+00 6.88132048e-01 -1.67688146e-01
7.56200433e-01 2.75938034e-01 1.13133870e-01 1.19667478e-01
-1.23352444e+00 7.47678652e-02 5.02160192e-01 5.75428307e-01
2.15964571e-01 -3.21128637e-01 1.25040025e-01 -7.82244578e-02
1.14767633e-01 1.99742224e-02 8.66476893e-01 7.64215589e-01
-2.13675812e-01 -9.16398823e-01 -5.91106176e-01 -2.76839696e-02
-5.63558996e-01 1.42157286e-01 -1.22723073e-01 6.82569206e-01
1.44602969e-01 1.04304981e+00 1.93728939e-01 -2.26563945e-01
4.01981950e-01 -1.45630881e-01 6.77991450e-01 -2.65853375e-01
-6.86066523e-02 6.01226032e-01 -2.32894090e-03 -9.00404274e-01
-8.14472616e-01 -8.79605770e-01 -9.47768927e-01 -4.19756055e-01
-2.38254309e-01 -3.36318195e-01 2.17006460e-01 7.14057744e-01
2.73197621e-01 3.83977532e-01 7.03192115e-01 -1.38935375e+00
7.51326159e-02 -4.63660419e-01 -4.12472099e-01 8.82054925e-01
6.67531371e-01 -3.78851086e-01 -2.22901866e-01 1.23323584e+00] | [9.631426811218262, -2.053640842437744] |
ae9ecaf3-57f5-456b-9d5a-b8074564d8d5 | contrastnet-a-contrastive-learning-framework | 2305.09269 | null | https://arxiv.org/abs/2305.09269v1 | https://arxiv.org/pdf/2305.09269v1.pdf | ContrastNet: A Contrastive Learning Framework for Few-Shot Text Classification | Few-shot text classification has recently been promoted by the meta-learning paradigm which aims to identify target classes with knowledge transferred from source classes with sets of small tasks named episodes. Despite their success, existing works building their meta-learner based on Prototypical Networks are unsatisfactory in learning discriminative text representations between similar classes, which may lead to contradictions during label prediction. In addition, the tasklevel and instance-level overfitting problems in few-shot text classification caused by a few training examples are not sufficiently tackled. In this work, we propose a contrastive learning framework named ContrastNet to tackle both discriminative representation and overfitting problems in few-shot text classification. ContrastNet learns to pull closer text representations belonging to the same class and push away text representations belonging to different classes, while simultaneously introducing unsupervised contrastive regularization at both task-level and instance-level to prevent overfitting. Experiments on 8 few-shot text classification datasets show that ContrastNet outperforms the current state-of-the-art models. | ['Jie Xu', 'Yongyi Mao', 'Richong Zhang', 'Junfan Chen'] | 2023-05-16 | null | null | null | null | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 4.91419196e-01 -7.23061990e-03 -3.57562751e-01 -5.19785166e-01
-5.91547668e-01 1.89081728e-01 8.66632402e-01 6.84479415e-01
-4.69407618e-01 6.87194586e-01 1.78585947e-01 2.84626812e-01
-2.42584303e-01 -7.76987195e-01 -2.47764558e-01 -6.35281563e-01
4.17237878e-01 5.24386525e-01 3.72089952e-01 -2.43825123e-01
3.82855654e-01 -1.90532714e-01 -1.93957746e+00 6.82966590e-01
9.73282218e-01 8.69191706e-01 3.59511785e-02 3.00479144e-01
-6.41402185e-01 1.24646986e+00 -6.04976833e-01 -3.71114910e-01
-1.99116036e-01 -7.47120917e-01 -7.80954182e-01 4.51650564e-03
5.72640121e-01 8.39142129e-02 -1.53529018e-01 1.10896873e+00
6.32976592e-01 7.08099067e-01 8.77759278e-01 -1.40538955e+00
-5.88587403e-01 8.30664396e-01 -4.51209843e-01 3.00733596e-01
-1.30852744e-01 -1.18944094e-01 1.16193736e+00 -9.95340466e-01
7.36202776e-01 1.05036771e+00 7.51031578e-01 8.94945145e-01
-1.25338602e+00 -7.00885713e-01 3.59817445e-01 2.93535739e-01
-1.07062232e+00 -4.26613808e-01 9.82867181e-01 -4.55220133e-01
1.07429767e+00 5.94738312e-02 3.25105250e-01 1.42263675e+00
3.82580943e-02 1.02058887e+00 8.42128277e-01 -5.59496880e-01
4.79814708e-01 2.39893422e-01 7.51142323e-01 4.44362313e-01
1.08889841e-01 -2.62224466e-01 -6.05981052e-01 -1.19175293e-01
-1.00678476e-02 5.24857938e-01 -3.66868898e-02 -5.92972517e-01
-7.02170610e-01 1.07333004e+00 3.62341106e-01 8.99552643e-01
-1.00799106e-01 -1.28112212e-01 9.22866702e-01 4.13434148e-01
1.00786996e+00 4.40687805e-01 -5.13800740e-01 -1.06697343e-01
-1.03054881e+00 5.39495051e-02 6.79556966e-01 9.80087996e-01
8.24441433e-01 6.27380088e-02 -5.42523682e-01 1.33194923e+00
-1.59095332e-03 -1.03307776e-01 1.17478597e+00 -1.57350034e-01
6.14364505e-01 1.00270247e+00 -3.61407816e-01 -8.84272277e-01
-5.04099190e-01 -4.60618734e-01 -7.93059409e-01 -2.20129583e-02
1.09028772e-01 -1.15249112e-01 -1.01073265e+00 1.53527749e+00
1.05856143e-01 4.23352003e-01 1.90403283e-01 5.08429050e-01
1.25567329e+00 7.10834265e-01 5.17510176e-01 -4.17534471e-01
1.21707940e+00 -1.22110283e+00 -7.36218631e-01 -5.87108076e-01
1.31204021e+00 -3.25928777e-01 1.20391273e+00 -4.52026911e-02
-8.04271519e-01 -6.48590982e-01 -1.07541668e+00 6.73982222e-03
-8.38035822e-01 -2.49220043e-01 2.60201097e-01 4.84297276e-01
-3.06223571e-01 8.62536430e-01 -3.11773270e-01 -5.25949419e-01
6.81295156e-01 -3.12047508e-02 -1.31209657e-01 -1.89386919e-01
-1.37337422e+00 8.74432266e-01 7.94060409e-01 -4.89650577e-01
-7.89871931e-01 -9.25647199e-01 -8.53094161e-01 3.46779019e-01
5.94388068e-01 -5.08611679e-01 1.13003433e+00 -1.13049304e+00
-1.25785995e+00 1.07440436e+00 1.98396817e-01 -4.35826749e-01
5.06515205e-01 -5.66949397e-02 -5.73701978e-01 -1.62300467e-01
2.86210388e-01 3.87381703e-01 1.01228702e+00 -1.15376484e+00
-8.66231740e-01 -5.71644247e-01 -3.15059692e-01 4.17834312e-01
-8.20162714e-01 -1.82200864e-01 1.13855982e-02 -6.12156808e-01
-1.58485487e-01 -4.64364439e-01 -3.76229286e-02 -3.83703858e-01
-2.05242485e-01 -6.84117317e-01 1.11119330e+00 1.54507738e-02
1.41755629e+00 -2.02882671e+00 1.24233373e-01 -4.21054512e-01
2.12625116e-01 5.46407223e-01 -2.07382172e-01 4.19292867e-01
-2.46492460e-01 -1.11003824e-01 -5.24937250e-02 -4.28306520e-01
-8.88109282e-02 1.16534732e-01 -3.55639637e-01 2.14851201e-01
-6.52303323e-02 7.94295251e-01 -1.26103437e+00 -5.54778099e-01
2.67237365e-01 1.44444257e-01 -1.54850706e-01 2.64293313e-01
-3.39538872e-01 8.10570549e-03 -4.11505759e-01 5.26536524e-01
2.96541899e-01 -2.85679728e-01 9.14129540e-02 8.40983540e-02
7.76981115e-02 2.73452669e-01 -8.34249258e-01 1.78655291e+00
-5.23407102e-01 5.70070267e-01 -5.08199751e-01 -1.57785785e+00
1.07323813e+00 4.15538579e-01 4.45499003e-01 -7.94464111e-01
4.40978795e-01 1.66886270e-01 -7.40813464e-02 -6.45247400e-01
4.73897308e-01 -6.35494411e-01 -7.48091340e-02 5.67367673e-01
6.39282942e-01 1.57433457e-03 2.69959182e-01 6.99697584e-02
9.46934819e-01 8.93772095e-02 5.60388207e-01 -1.83638901e-01
2.83514917e-01 1.21527292e-01 4.91210371e-01 9.95153725e-01
-4.77214634e-01 3.74267220e-01 3.76044333e-01 -6.79150939e-01
-9.23606813e-01 -5.45236528e-01 -2.91967064e-01 1.99365103e+00
2.65761837e-02 -5.44029176e-01 -3.75903457e-01 -1.08482337e+00
-7.92209283e-02 1.14246297e+00 -1.10660255e+00 -6.95150375e-01
-1.98232010e-01 -8.47956359e-01 3.68998080e-01 6.32206321e-01
3.65920454e-01 -1.25917864e+00 -6.69183016e-01 4.59545851e-01
1.29321022e-02 -7.31238008e-01 -1.09384127e-01 8.33245575e-01
-1.14517498e+00 -1.12052619e+00 -9.54095781e-01 -1.04683733e+00
5.38060129e-01 5.42244613e-01 1.25617790e+00 -1.45170069e-03
-4.27330017e-01 3.17575149e-02 -6.77215576e-01 -6.18782997e-01
-4.48784769e-01 3.43608290e-01 -1.99603200e-01 1.40230078e-02
9.26502168e-01 -4.39481109e-01 -2.23653182e-01 2.69061983e-01
-9.23716545e-01 -6.35519740e-04 -1.75765343e-02 1.30275857e+00
1.58846140e-01 7.72744045e-02 6.96711183e-01 -1.33022690e+00
6.94411814e-01 -8.25592995e-01 -5.64546250e-02 4.16757196e-01
-8.17132354e-01 -1.35389566e-01 8.73645425e-01 -6.78431332e-01
-1.16463280e+00 -3.71007085e-01 3.56340647e-01 -4.66978818e-01
-2.80241996e-01 5.33680379e-01 1.06277555e-01 4.13274288e-01
1.12403810e+00 3.41530323e-01 -1.51188210e-01 -2.41870865e-01
2.01685190e-01 7.19558597e-01 -1.24714896e-01 -3.30262899e-01
2.50784039e-01 3.94724220e-01 -3.35316330e-01 -8.74289930e-01
-1.64407670e+00 -8.34459305e-01 -6.53314412e-01 -1.44892141e-01
7.02391088e-01 -8.41953516e-01 -4.65292782e-02 4.21503603e-01
-7.80754447e-01 -3.24053824e-01 -5.97159445e-01 2.36266166e-01
-6.18042588e-01 3.07629071e-02 -3.60168248e-01 -7.00787842e-01
-4.64377642e-01 -8.22802961e-01 8.15601230e-01 2.58012235e-01
-2.76103020e-01 -1.23190999e+00 4.06296521e-01 1.84943900e-01
2.86342204e-01 -1.02934413e-01 1.02911258e+00 -1.38881707e+00
4.30264235e-01 -3.84909540e-01 -8.71896446e-02 1.68065310e-01
2.04270333e-01 -2.37582684e-01 -1.22506642e+00 -4.32867438e-01
7.82191604e-02 -8.60696912e-01 1.23771203e+00 4.32903618e-01
8.98873925e-01 -7.61101544e-02 -5.76183736e-01 3.80590558e-01
1.35258472e+00 2.02876136e-01 5.05984485e-01 5.47985196e-01
6.79805040e-01 7.84281433e-01 7.58502841e-01 3.90246332e-01
-1.09886102e-01 5.20771742e-01 1.31838307e-01 3.26647282e-01
3.29036750e-02 -2.43680850e-01 -4.39741649e-02 6.84498966e-01
9.49373543e-02 -3.20625275e-01 -9.95203137e-01 4.29256082e-01
-2.19972563e+00 -1.19823539e+00 1.07761122e-01 1.92891681e+00
6.88964725e-01 2.99190581e-01 5.74166924e-02 1.88986585e-01
1.06010497e+00 4.44402486e-01 -5.46245635e-01 -2.36950696e-01
3.48545909e-02 -6.61229640e-02 -1.75045788e-01 -8.66443589e-02
-1.35314322e+00 1.10564315e+00 5.31869698e+00 1.07624233e+00
-9.78742182e-01 3.78998101e-01 5.45366406e-01 -2.71308243e-01
3.69636179e-03 1.88084308e-03 -1.02045953e+00 4.22113597e-01
9.71866369e-01 -3.38210255e-01 -2.45316595e-01 1.18721855e+00
-3.67303729e-01 2.36108691e-01 -1.07898545e+00 9.15872216e-01
4.38451707e-01 -1.30097091e+00 2.51687169e-01 -3.35967958e-01
1.06383872e+00 -2.30305027e-02 -1.42377824e-01 1.10653841e+00
2.19968662e-01 -8.27645063e-01 3.75359833e-01 3.59797060e-01
4.07644033e-01 -7.75254786e-01 8.36493433e-01 5.85117817e-01
-9.84223366e-01 -4.42776471e-01 -9.41262543e-01 -7.57499412e-02
-1.40789121e-01 4.47366029e-01 -7.89578199e-01 3.37444633e-01
5.90951622e-01 1.10710156e+00 -4.83480275e-01 8.34086895e-01
-1.45429550e-02 4.21443969e-01 3.38102102e-01 -3.99156958e-01
5.62218130e-01 5.67449220e-02 3.48651260e-01 1.09906185e+00
1.19468085e-01 -1.99959259e-02 3.07847112e-01 7.13482022e-01
-1.95392206e-01 5.65480530e-01 -8.21829498e-01 -1.24503672e-01
2.36222908e-01 1.14565718e+00 -7.90471077e-01 -7.34354854e-01
-5.54530680e-01 8.59430492e-01 7.41927803e-01 1.20064281e-01
-5.35237491e-01 -6.87806189e-01 2.55757928e-01 -5.59594035e-02
2.72971958e-01 5.19747198e-01 -3.12202901e-01 -1.37750924e+00
-3.54566842e-01 -6.79033577e-01 8.73885810e-01 -5.30974030e-01
-1.55783379e+00 4.82946068e-01 -1.63348526e-01 -1.50303710e+00
-1.63408101e-01 -4.26188767e-01 -8.51088285e-01 4.36157018e-01
-1.36326909e+00 -1.02071035e+00 -3.00280005e-01 5.51986337e-01
1.13176501e+00 -5.80712676e-01 1.10559559e+00 5.59824221e-02
-5.81559658e-01 6.14791989e-01 4.95790392e-01 1.14124611e-01
9.22189832e-01 -1.03830278e+00 1.19292617e-01 3.07301909e-01
1.16622910e-01 2.89913476e-01 6.19489551e-01 -6.57668173e-01
-7.16112316e-01 -1.12995195e+00 9.55076337e-01 -2.80625015e-01
7.89751589e-01 -1.68827146e-01 -1.35854900e+00 5.87896764e-01
-3.60974856e-02 3.44748288e-01 1.04404640e+00 4.48064148e-01
-6.17591441e-01 -4.28026164e-04 -9.60299730e-01 5.94736516e-01
9.00691807e-01 -5.55554867e-01 -1.02749372e+00 5.09957969e-01
3.95706385e-01 4.03281972e-02 -5.40534854e-01 2.42176205e-01
2.92982310e-01 -9.55781758e-01 6.51376724e-01 -1.26516962e+00
6.90198720e-01 3.90220135e-01 -8.35711285e-02 -1.65357244e+00
-4.48681146e-01 -9.27617922e-02 -3.03874612e-01 1.30783749e+00
2.46008024e-01 -3.02904308e-01 8.05452168e-01 2.18414053e-01
-2.44436160e-01 -4.74097759e-01 -9.57041025e-01 -9.26560342e-01
3.96217287e-01 -3.10387909e-01 1.54640883e-01 1.66633177e+00
6.24221206e-01 1.01545811e+00 -5.21644175e-01 -5.12624860e-01
6.15061045e-01 1.11430019e-01 4.67028916e-01 -1.86918867e+00
8.88378546e-02 -7.47332036e-01 -5.00347674e-01 -4.95182961e-01
5.22980154e-01 -1.20459032e+00 2.82985121e-01 -1.39041054e+00
5.70920765e-01 -2.12756261e-01 -6.17784083e-01 4.90103811e-01
-4.76998091e-01 2.46168803e-02 1.40999213e-01 2.79132575e-01
-9.19961274e-01 7.88653195e-01 9.97765899e-01 -5.58197856e-01
-2.78310210e-01 -2.56977919e-02 -5.14468074e-01 7.65421748e-01
8.56040597e-01 -9.07041252e-01 -5.89763880e-01 -1.00669749e-01
2.07934707e-01 -3.47502157e-02 7.59383887e-02 -1.12475240e+00
3.43829602e-01 -7.51825795e-02 3.69862765e-01 -4.78396714e-01
1.59794524e-01 -7.85059094e-01 -4.95696068e-01 5.60941219e-01
-9.55584168e-01 -5.65217257e-01 4.30545360e-02 7.14930832e-01
-2.06788227e-01 -9.25951242e-01 1.13337517e+00 -3.85958463e-01
-8.86602581e-01 2.64467031e-01 -3.37016582e-01 5.45197546e-01
1.18389153e+00 -2.54584968e-01 -5.21358311e-01 -9.88481045e-02
-8.54913652e-01 1.40713245e-01 1.71741918e-01 6.68282270e-01
5.15759349e-01 -1.27370870e+00 -7.15045154e-01 1.16362683e-01
8.69771957e-01 -3.65145564e-01 6.19766593e-01 5.90956569e-01
1.11115798e-01 3.13534290e-01 -1.85757011e-01 -4.81256276e-01
-1.17070305e+00 8.24399114e-01 3.48601639e-01 -5.18975198e-01
-8.60956728e-01 9.24407899e-01 1.01080105e-01 -5.08442760e-01
4.58216697e-01 1.16542004e-01 -5.22176564e-01 6.75597847e-01
8.38253975e-01 5.86462080e-01 1.58215657e-01 -3.33806813e-01
-2.86975466e-02 2.47263581e-01 -6.95969641e-01 4.52678323e-01
1.41865373e+00 2.31003035e-02 2.89387494e-01 1.07963681e+00
1.34874439e+00 -8.29925597e-01 -1.13140452e+00 -5.90349376e-01
2.12941945e-01 -2.28552580e-01 1.88859701e-01 -6.59539700e-01
-8.27540278e-01 1.19104409e+00 6.36289120e-01 4.68508005e-01
5.53056359e-01 -1.12581328e-01 5.54170012e-01 6.33220196e-01
2.80641057e-02 -1.77504158e+00 4.79260534e-01 8.65316570e-01
3.10480535e-01 -1.59582078e+00 -2.03365483e-03 1.02686271e-01
-7.57162213e-01 1.30235422e+00 8.73256624e-01 -5.72139025e-02
7.86181092e-01 -1.58693761e-01 6.47659553e-03 -3.72917682e-01
-1.05258811e+00 -2.29908735e-01 3.57171983e-01 5.17020226e-01
7.42084265e-01 -2.24766746e-01 -3.10924560e-01 6.84562266e-01
4.69015330e-01 3.86450626e-02 2.91325390e-01 1.38727188e+00
-9.44132328e-01 -7.29255319e-01 3.83808613e-02 7.67229378e-01
-3.65122885e-01 -2.01216489e-01 -3.03542465e-01 7.33527720e-01
-7.04737753e-02 8.02129507e-01 3.69402915e-01 -2.43119001e-01
4.29831862e-01 7.03967035e-01 1.99296981e-01 -1.14957309e+00
-7.64784753e-01 -1.13821842e-01 -5.30398674e-02 -1.11937515e-01
-5.94016314e-01 -2.79466033e-01 -1.05922198e+00 6.22885861e-02
-5.26268363e-01 3.46847057e-01 2.38070354e-01 1.15261555e+00
1.69392228e-01 7.75393963e-01 5.40920973e-01 -7.71204412e-01
-8.49278212e-01 -1.36922717e+00 -8.91275227e-01 7.39578545e-01
3.08289621e-02 -7.98965156e-01 -5.45335174e-01 -1.89034134e-01] | [10.274604797363281, 3.637838125228882] |
a5865a83-7d7b-40e9-81f3-93f022f18f15 | a-three-player-gan-generating-hard-samples-to | 1903.03496 | null | http://arxiv.org/abs/1903.03496v1 | http://arxiv.org/pdf/1903.03496v1.pdf | A Three-Player GAN: Generating Hard Samples To Improve Classification Networks | We propose a Three-Player Generative Adversarial Network to improve
classification networks. In addition to the game played between the
discriminator and generator, a competition is introduced between the generator
and the classifier. The generator's objective is to synthesize samples that are
both realistic and hard to label for the classifier. Even though we make no
assumptions on the type of augmentations to learn, we find that the model is
able to synthesize realistically looking examples that are hard for the
classification model. Furthermore, the classifier becomes more robust when
trained on these difficult samples. The method is evaluated on a public dataset
for traffic sign recognition. | ['Luc van Gool', 'Bert de Brabandere', 'Simon Vandenhende', 'Davy Neven'] | 2019-03-08 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 5.79676867e-01 5.53167462e-01 -6.86581954e-02 -2.88441449e-01
-9.56872761e-01 -7.55375087e-01 6.77276969e-01 -6.81916654e-01
-2.74416327e-01 1.03456473e+00 -2.47863173e-01 -3.83743286e-01
4.28476870e-01 -9.31509912e-01 -7.91475773e-01 -8.18377316e-01
1.87022284e-01 5.98407626e-01 7.98818991e-02 -1.66459814e-01
-1.49161264e-01 5.49740732e-01 -1.56613433e+00 5.35083950e-01
9.65913832e-01 9.37842190e-01 -4.05179679e-01 1.13405085e+00
4.75809216e-01 1.19499719e+00 -1.10112131e+00 -6.93008304e-01
7.19152927e-01 -7.44614482e-01 -6.72691226e-01 1.36964366e-01
7.35292077e-01 -2.93488413e-01 -3.92773747e-01 1.12090075e+00
3.50492746e-01 5.52365631e-02 9.38004971e-01 -1.83903456e+00
-5.74381113e-01 4.36666131e-01 3.18782441e-02 2.84070503e-02
2.40918081e-02 6.84637427e-01 9.50679183e-01 -5.19827783e-01
7.02226102e-01 9.81946886e-01 6.36564612e-01 1.18660092e+00
-1.22123456e+00 -9.50207770e-01 -1.27113611e-01 4.34241854e-02
-1.06695259e+00 -5.66880941e-01 6.95185423e-01 -3.85501295e-01
4.40778017e-01 3.93607110e-01 5.43715715e-01 1.54360652e+00
9.35684219e-02 7.79559553e-01 1.21338630e+00 -3.06803763e-01
2.88051277e-01 4.80173707e-01 -1.04314223e-01 4.09106821e-01
2.94781029e-01 5.93799770e-01 1.83749422e-02 -7.28952959e-02
5.07631421e-01 -4.62915540e-01 -9.24743488e-02 -2.31118351e-01
-6.24452174e-01 9.46775556e-01 2.46416047e-01 -1.75527990e-01
-4.68609065e-01 3.16228479e-01 1.84253395e-01 4.43392873e-01
1.18992753e-01 7.49198020e-01 -1.46677280e-02 -6.41461834e-02
-8.75679672e-01 3.83764535e-01 9.69402492e-01 8.30863357e-01
4.04026419e-01 4.69723046e-01 -3.56004357e-01 4.75699842e-01
1.29772201e-01 6.05696499e-01 2.66057700e-01 -1.05588889e+00
3.85194093e-01 3.17057043e-01 -1.55925937e-02 -4.64860231e-01
2.05104485e-01 -4.35494334e-01 -5.63321412e-01 8.56374741e-01
8.72745216e-01 -4.10970509e-01 -1.35911071e+00 1.78468168e+00
-7.97425359e-02 4.23357904e-01 2.15944052e-01 7.08388984e-01
7.22268224e-01 4.68713075e-01 7.08962977e-02 2.78320670e-01
8.80037189e-01 -8.43415380e-01 -3.15963179e-01 -3.75159234e-01
5.40670335e-01 -6.06273413e-01 9.53373194e-01 3.22607517e-01
-1.13458276e+00 -5.77420950e-01 -1.13787043e+00 2.69386113e-01
-5.27854383e-01 -9.94400904e-02 5.94499409e-01 1.15481722e+00
-8.33242774e-01 4.69517082e-01 -4.60112363e-01 1.22033827e-01
8.24118733e-01 2.66826063e-01 -3.07406694e-01 -2.66802013e-01
-1.15709233e+00 1.10255444e+00 2.01902539e-01 1.00838855e-01
-1.18994284e+00 -4.24075007e-01 -7.25376189e-01 -4.88464274e-02
9.51042026e-03 -6.24622285e-01 1.44014823e+00 -1.57088149e+00
-1.39773393e+00 9.08793747e-01 1.43398657e-01 -4.75190759e-01
1.09225273e+00 5.75422227e-01 -4.05999333e-01 -9.47391987e-02
3.06638191e-03 8.47286344e-01 9.53549325e-01 -1.57215416e+00
-7.24568486e-01 2.69628912e-01 2.41788015e-01 3.93299107e-03
2.75981843e-01 -2.49747157e-01 7.63162449e-02 -7.54184842e-01
-4.21352863e-01 -1.01628816e+00 -2.12282687e-01 1.46816194e-01
-5.98840833e-01 -1.76209938e-02 1.04037178e+00 -5.32438159e-01
4.87621397e-01 -2.03429937e+00 -3.06991130e-01 5.33544064e-01
1.28943726e-01 6.53053582e-01 -5.02865851e-01 -4.47078701e-03
-3.22834462e-01 4.53403741e-01 -2.05212131e-01 -2.09809899e-01
6.00304035e-03 5.08431673e-01 -3.73341501e-01 2.30410054e-01
7.21581340e-01 1.05199730e+00 -8.37038457e-01 -1.75473809e-01
-1.96087241e-01 2.57455170e-01 -7.17895627e-01 2.20263973e-01
-2.27875456e-01 5.21891773e-01 -3.26336324e-01 4.79105979e-01
6.31468058e-01 -7.08340248e-03 -1.88845500e-01 2.33216390e-01
4.31606501e-01 2.88312972e-01 -1.16785920e+00 6.13660932e-01
-3.22027415e-01 8.05857658e-01 -1.24965392e-01 -1.06968403e+00
8.36855531e-01 1.46673754e-01 -9.15254746e-03 -4.65354562e-01
1.90236151e-01 1.92688540e-01 5.73950171e-01 -3.76672208e-01
2.79798716e-01 -6.33291125e-01 -1.01823032e-01 6.89151883e-01
-1.50791472e-02 -4.20306236e-01 4.37354028e-01 9.37057957e-02
1.29062700e+00 -7.78623223e-02 -1.90745771e-01 1.88956410e-01
3.89991075e-01 4.27266061e-02 7.35035241e-01 8.96923482e-01
-3.68842363e-01 6.59794033e-01 6.35099649e-01 -2.83021420e-01
-1.22921598e+00 -1.17699289e+00 1.03362598e-01 6.75338984e-01
-6.83729276e-02 2.27682278e-01 -7.93629467e-01 -1.18714559e+00
4.66889888e-02 9.49412346e-01 -8.42072487e-01 -5.33970833e-01
-5.45267761e-01 -4.95284557e-01 1.13030016e+00 7.53750145e-01
4.27770942e-01 -1.14745617e+00 -3.12288523e-01 1.19833313e-02
-2.18048482e-03 -1.03080356e+00 -5.03895223e-01 1.90519467e-01
-5.64012110e-01 -1.39973736e+00 -3.70013356e-01 -8.62400413e-01
9.35615957e-01 -2.56997049e-01 1.10900021e+00 1.92235529e-01
-1.79582849e-01 1.74671352e-01 -3.64968568e-01 -7.93835104e-01
-1.35535693e+00 -8.82543921e-02 -3.11444402e-01 3.27759236e-02
1.99630916e-01 -3.18977416e-01 -3.52067173e-01 5.78218400e-01
-8.91336322e-01 3.98279503e-02 4.77444500e-01 1.18942714e+00
1.53937206e-01 1.72384217e-01 7.36932278e-01 -9.09435689e-01
6.52131796e-01 -3.94237161e-01 -4.78712678e-01 1.59491226e-01
-3.97661060e-01 1.29023954e-01 9.48871255e-01 -8.43815327e-01
-9.77399409e-01 1.65194467e-01 -2.71370858e-01 -3.12867075e-01
-2.77807444e-01 4.05433849e-02 -2.83331841e-01 -4.22518432e-01
1.01801658e+00 1.38560578e-01 1.93622172e-01 -9.67148915e-02
3.58558118e-01 5.04390597e-01 6.25547826e-01 -5.96793234e-01
1.29993749e+00 2.12421075e-01 1.72262918e-02 -5.57783902e-01
-5.06953657e-01 4.08065945e-01 -4.53385562e-01 -3.82845432e-01
6.09841347e-01 -6.03337944e-01 -5.56439757e-01 7.27664709e-01
-9.70582545e-01 -6.93054974e-01 -9.69474256e-01 2.91741520e-01
-6.69998527e-01 -2.09272578e-02 -4.73822415e-01 -8.03142309e-01
7.20777288e-02 -1.25149405e+00 5.25423348e-01 4.90790457e-01
-2.99265325e-01 -9.55900609e-01 -2.31382355e-01 5.41587293e-01
4.47066098e-01 5.33250332e-01 8.36088538e-01 -9.29451704e-01
-8.07473421e-01 -7.10359156e-01 -1.32382870e-01 1.01941729e+00
1.41446292e-01 3.77981931e-01 -1.26607335e+00 -8.54582489e-02
-2.65428960e-01 -7.72319973e-01 8.70555222e-01 4.89460379e-02
1.08943880e+00 -4.21041340e-01 -9.57584381e-02 4.82607156e-01
8.70530903e-01 5.44998705e-01 1.14327216e+00 6.43995777e-02
4.73143071e-01 4.10478383e-01 3.09691221e-01 -2.00838774e-01
1.62360013e-01 4.79268253e-01 3.73697937e-01 -2.55116224e-01
-3.83067012e-01 -4.83493060e-01 3.99763465e-01 7.89856091e-02
-1.48036093e-01 -5.09206176e-01 -9.44732070e-01 2.22802624e-01
-1.28644633e+00 -1.33685935e+00 6.07206561e-02 2.07504463e+00
8.94366205e-01 4.33806568e-01 2.31761605e-01 4.83304113e-01
6.30435884e-01 -4.98252772e-02 -5.11735201e-01 -5.51968813e-01
-1.77156776e-01 7.47390389e-01 4.82288897e-01 5.92444599e-01
-1.06598628e+00 9.20294344e-01 7.66113472e+00 7.59980321e-01
-1.22525096e+00 -1.03079550e-01 8.70887995e-01 7.32954964e-03
-2.59268671e-01 -8.79532471e-02 -6.61811769e-01 6.76660419e-01
7.58356631e-01 -9.19328928e-02 4.50500637e-01 8.29131722e-01
-7.81720206e-02 3.00916668e-04 -1.38200021e+00 4.59220082e-01
1.02665946e-01 -1.21855199e+00 9.27977413e-02 1.02162771e-01
8.44777942e-01 -2.38160938e-01 3.53372067e-01 5.86443603e-01
7.82387376e-01 -1.46663189e+00 8.75183880e-01 5.40656149e-01
7.78699696e-01 -8.61433625e-01 8.48226428e-01 4.79913205e-01
-6.29756331e-01 -1.84360355e-01 -4.50564623e-02 -7.05276951e-02
-2.70270929e-02 1.30966142e-01 -1.08308232e+00 6.44878894e-02
8.23760498e-03 1.12057313e-01 -6.34824932e-01 1.22692287e+00
-7.19646335e-01 8.24792862e-01 -1.81166381e-01 4.06410061e-02
1.12900145e-01 5.16644903e-02 5.08962870e-01 1.05256891e+00
8.11808482e-02 3.24489027e-02 1.01420321e-01 1.05590892e+00
-3.57681930e-01 -4.44352150e-01 -8.04041505e-01 -2.49303043e-01
3.75571966e-01 9.76993263e-01 -5.31674206e-01 -4.93139654e-01
1.51600270e-03 8.68310511e-01 4.14913110e-02 5.30967116e-01
-8.79142284e-01 -3.42532575e-01 6.90662742e-01 1.17088053e-02
1.50920868e-01 1.73077419e-01 -5.25176585e-01 -1.10917830e+00
1.33872941e-01 -1.36956334e+00 2.69302666e-01 -7.36697972e-01
-1.61528885e+00 5.34176111e-01 -2.60699749e-01 -1.27044415e+00
-6.95101500e-01 -8.05838764e-01 -1.14014602e+00 1.27334416e+00
-1.31381261e+00 -1.24352419e+00 -2.47604311e-01 4.18787599e-01
2.32468143e-01 -3.93025935e-01 7.91609883e-01 2.63391465e-01
-4.43300039e-01 1.09344161e+00 -2.51598120e-01 5.75736821e-01
3.62775564e-01 -1.13371432e+00 6.58422410e-01 1.04990435e+00
-4.21181433e-02 2.16628686e-01 7.44087219e-01 -6.19360924e-01
-8.69538665e-01 -1.21879768e+00 7.72109270e-01 -6.85445905e-01
3.24206114e-01 -4.38782215e-01 -7.60052145e-01 6.13284111e-01
-2.24578366e-01 2.65327632e-01 6.61757946e-01 -2.94733614e-01
-6.82239354e-01 1.35105792e-02 -1.59051311e+00 4.80106413e-01
8.57336283e-01 -5.80468893e-01 -4.82636750e-01 1.83028877e-01
7.92928338e-02 -5.46366334e-01 -4.78201151e-01 2.06494406e-01
6.66634977e-01 -5.23393989e-01 9.96860266e-01 -1.12355828e+00
5.31276464e-01 -2.11166158e-01 -8.26770365e-02 -1.66468334e+00
-1.25017464e-01 -3.22574526e-01 4.00538854e-02 1.16881728e+00
8.54682982e-01 -8.56013179e-01 1.32751036e+00 8.19705844e-01
1.09184466e-01 -5.26311100e-01 -9.56198871e-01 -1.27114797e+00
4.75820035e-01 -5.42970955e-01 4.94638503e-01 9.58732426e-01
-3.69661063e-01 2.44604304e-01 -4.00255591e-01 2.06474513e-02
6.38093591e-01 -2.94487536e-01 9.98183489e-01 -1.31661379e+00
-5.70828378e-01 -6.27329826e-01 -6.50050163e-01 -6.59013033e-01
2.28641406e-01 -1.14056230e+00 2.58751899e-01 -1.11470711e+00
6.72123581e-02 -7.20677078e-01 -4.27761637e-02 5.99439323e-01
-2.94683814e-01 6.02316916e-01 4.80865002e-01 -1.12187624e-01
-2.95584486e-03 3.04864854e-01 1.45994306e+00 -3.38725209e-01
-4.99997996e-02 4.46315050e-01 -8.11713994e-01 5.24111688e-01
8.97891343e-01 -6.92406714e-01 -3.69436949e-01 3.48373652e-02
2.76324281e-04 -2.67439216e-01 7.09795356e-01 -9.12275732e-01
2.57196967e-02 -1.72648206e-01 5.69008410e-01 -9.12809223e-02
4.17449296e-01 -5.66163480e-01 1.91665426e-01 5.38506985e-01
-4.99167591e-01 -4.18027639e-01 1.34790227e-01 2.92393148e-01
-8.44732821e-02 -4.26651388e-01 1.39884770e+00 1.32619396e-01
-2.02941418e-01 2.87327647e-01 -4.90966827e-01 4.82267648e-01
1.05047286e+00 -3.18160027e-01 -5.17301679e-01 -7.95659661e-01
-7.07583487e-01 2.43031576e-01 5.45853257e-01 3.27234805e-01
4.70514089e-01 -1.56523955e+00 -8.83863330e-01 4.48900938e-01
2.38879789e-02 -9.99677926e-02 -1.78946987e-01 3.22440386e-01
-3.62062722e-01 -7.16665983e-02 -4.65163380e-01 -2.14969084e-01
-1.30832124e+00 2.40544587e-01 7.70797133e-01 -3.58190924e-01
-2.39724427e-01 8.37366700e-01 -1.81337930e-02 -3.40477139e-01
3.71988863e-01 -2.63124287e-01 1.20633565e-01 3.14771906e-02
4.77481574e-01 3.24804097e-01 -2.29420647e-01 -6.18758380e-01
-4.64090407e-02 -2.92068832e-02 9.05725081e-03 -1.52565673e-01
1.09476125e+00 6.32286012e-01 4.38936710e-01 1.59035519e-01
9.96447861e-01 -1.67784795e-01 -1.39242995e+00 6.08414523e-02
-3.37697387e-01 -4.95729506e-01 -2.54274487e-01 -1.07130992e+00
-1.20870304e+00 8.61205816e-01 5.16592085e-01 3.02488774e-01
8.26216817e-01 -4.35889393e-01 5.98180175e-01 4.58579719e-01
2.40842074e-01 -1.04041779e+00 1.14742354e-01 5.33846200e-01
1.00712502e+00 -1.25692201e+00 -4.38869566e-01 -4.52469677e-01
-7.83630431e-01 1.04294777e+00 7.31927395e-01 -3.28500569e-01
3.04118454e-01 4.41447765e-01 3.15264225e-01 1.54345885e-01
-6.60158575e-01 -9.40833464e-02 3.47676873e-01 1.19693577e+00
1.67498514e-02 8.52257833e-02 1.39130831e-01 3.97622347e-01
-4.90267634e-01 8.80002528e-02 8.48221540e-01 8.10091376e-01
-1.62445620e-01 -1.28384233e+00 -4.37467158e-01 5.95767558e-01
-2.49340385e-01 7.88327679e-02 -7.40717828e-01 1.04296684e+00
4.51893568e-01 9.69389677e-01 -1.27656031e-02 -5.52652836e-01
3.46058965e-01 4.19449061e-01 4.49537814e-01 -6.61161065e-01
-5.95597804e-01 -6.92179084e-01 4.73731488e-01 -2.67574698e-01
3.09145097e-02 -7.19573200e-01 -9.58850443e-01 -3.24187368e-01
-2.88736135e-01 5.88812716e-02 4.79093581e-01 8.30723166e-01
-1.42921209e-01 5.41718721e-01 1.00006390e+00 -6.78335428e-01
-8.52802634e-01 -7.91786253e-01 -4.44698364e-01 5.73633850e-01
3.22240412e-01 -5.05530953e-01 -5.45580924e-01 1.18070535e-01] | [11.47636604309082, -0.2606852948665619] |
dcc3ad52-9e5d-43f0-9a71-f26c81c1ab41 | uncertainty-in-real-time-semantic | 2301.01201 | null | https://arxiv.org/abs/2301.01201v3 | https://arxiv.org/pdf/2301.01201v3.pdf | Uncertainty in Real-Time Semantic Segmentation on Embedded Systems | Application for semantic segmentation models in areas such as autonomous vehicles and human computer interaction require real-time predictive capabilities. The challenges of addressing real-time application is amplified by the need to operate on resource constrained hardware. Whilst development of real-time methods for these platforms has increased, these models are unable to sufficiently reason about uncertainty present. This paper addresses this by combining deep feature extraction from pre-trained models with Bayesian regression and moment propagation for uncertainty aware predictions. We demonstrate how the proposed method can yield meaningful uncertainty on embedded hardware in real-time whilst maintaining predictive performance. | ['Clinton Fookes', 'Ethan Goan'] | 2022-12-20 | null | null | null | null | ['real-time-semantic-segmentation'] | ['computer-vision'] | [ 2.57836401e-01 7.20016718e-01 -3.44924122e-01 -9.75639582e-01
-9.48056996e-01 -2.29862273e-01 7.14814484e-01 1.68012619e-01
-4.22129035e-01 7.37047791e-01 -4.88940358e-01 -4.77338165e-01
-1.38642743e-01 -6.05727732e-01 -7.99316525e-01 -3.47128958e-01
-2.21336395e-01 8.46947312e-01 7.14053571e-01 2.68288851e-01
2.96589673e-01 5.13396382e-01 -1.73497152e+00 1.07036129e-01
6.34299219e-01 1.41353464e+00 4.58097970e-03 9.16664302e-01
1.04668893e-01 5.11277854e-01 -3.88099074e-01 7.06853643e-02
1.31832123e-01 2.41346240e-01 -4.44919884e-01 -2.75911719e-01
-3.31309512e-02 -5.08635163e-01 -1.80768773e-01 1.04570138e+00
7.92405382e-02 -2.07490027e-02 5.89251935e-01 -1.64017546e+00
3.46488982e-01 7.47320950e-01 -4.81881708e-01 9.72502977e-02
-4.98376526e-02 1.96918651e-01 4.34336126e-01 -3.97623509e-01
1.97300225e-01 1.38002276e+00 6.57894492e-01 4.49488223e-01
-1.01615107e+00 -7.54794955e-01 1.65830161e-02 1.24968953e-01
-1.31220353e+00 -7.25007474e-01 4.23154563e-01 -3.45922917e-01
1.40973139e+00 -2.83004977e-02 1.72422707e-01 7.83218980e-01
9.49164808e-01 8.95292759e-01 8.99682760e-01 -3.12039703e-01
6.92806184e-01 3.53101999e-01 1.28940940e-01 4.27672267e-01
3.79659504e-01 6.46346033e-01 -5.35524905e-01 -1.62223712e-01
1.96277350e-01 -4.07182515e-01 5.22140145e-01 -1.12318836e-01
-6.04787230e-01 7.76323617e-01 1.85406767e-02 -1.46532536e-01
-3.82775873e-01 8.88987780e-01 5.24279416e-01 -2.87511885e-01
4.23905522e-01 2.09611729e-01 -8.09505641e-01 -4.36909020e-01
-1.49504423e+00 2.71658242e-01 9.29520190e-01 1.31280363e+00
5.64977467e-01 2.56547868e-01 1.98060885e-01 1.81703363e-03
1.09181750e+00 6.18828952e-01 1.77069724e-01 -1.18422592e+00
4.57511321e-02 7.78065398e-02 1.42627224e-01 -7.01766312e-01
-7.20529437e-01 -2.21962750e-01 -1.79721043e-01 5.54181159e-01
3.94500941e-02 -1.46524295e-01 -1.36119616e+00 1.37648690e+00
2.99470037e-01 4.36292887e-01 3.38726938e-01 4.31584239e-01
2.81068355e-01 5.83880007e-01 3.32053363e-01 -1.63067633e-03
1.21068299e+00 -3.99886280e-01 -6.68077111e-01 -5.90560496e-01
2.78771430e-01 -6.31288469e-01 6.17796518e-02 6.83906794e-01
-7.90043950e-01 -1.97115213e-01 -1.58119369e+00 1.16423562e-01
-4.32332531e-02 -3.58328700e-01 8.52413714e-01 1.01669765e+00
-8.29042912e-01 7.18026042e-01 -1.66383672e+00 -1.49848774e-01
7.37392902e-01 7.30351508e-01 1.84556946e-01 1.86627865e-01
-1.14157426e+00 1.24366081e+00 6.77916825e-01 2.01491013e-01
-1.00932419e+00 -4.92416382e-01 -1.07469428e+00 -2.31366724e-01
2.72117913e-01 -2.23824173e-01 1.67895460e+00 -6.56111240e-01
-1.83394182e+00 2.45480314e-01 4.48890351e-05 -1.11023462e+00
5.34103394e-01 -3.99167538e-01 -5.31145275e-01 1.07001159e-02
-4.21203002e-02 9.13228035e-01 1.19126105e+00 -1.07094991e+00
-9.52834308e-01 -2.90759087e-01 -3.23338956e-01 -2.26490140e-01
3.15606415e-01 -2.17309117e-01 -2.54822195e-01 2.80315369e-01
1.72742456e-01 -1.18777573e+00 -7.30757654e-01 -2.00867593e-01
-6.27963617e-02 5.80428801e-02 1.24464405e+00 -5.50070763e-01
6.57345176e-01 -1.74862051e+00 -7.04531729e-01 3.90579641e-01
-1.28868520e-01 1.47337869e-01 3.32743496e-01 -9.66114625e-02
6.12263799e-01 -3.48220728e-02 -1.34238526e-01 -2.49267425e-02
2.22062781e-01 5.54608583e-01 -5.48566520e-01 6.45573199e-01
6.99404657e-01 6.24115705e-01 -8.56072903e-01 -9.24365103e-01
6.78914368e-01 5.44012666e-01 -4.45666671e-01 -6.28356412e-02
-8.38517666e-01 1.50594756e-01 -8.46298695e-01 6.18827164e-01
7.71958053e-01 1.72516882e-01 1.97408199e-01 1.29221827e-01
-1.13008067e-01 4.29528952e-01 -1.14713001e+00 1.54693222e+00
-6.56717300e-01 7.54702508e-01 1.09953980e-04 -6.78700805e-01
8.65166306e-01 4.39672582e-02 2.86376894e-01 -5.44224203e-01
6.69062734e-01 2.28119418e-01 -5.77286556e-02 -6.71576187e-02
9.77745235e-01 -1.74495503e-01 -3.80912572e-01 1.25284031e-01
-5.34393825e-03 -8.10786486e-01 -4.38275754e-01 3.94663885e-02
1.08234072e+00 7.66909003e-01 2.30078310e-01 -5.53540766e-01
-1.04589976e-01 4.31431383e-01 8.12220752e-01 4.66439337e-01
-7.09693491e-01 3.16500664e-01 2.58440822e-01 -1.62766188e-01
-9.97013927e-01 -8.91678691e-01 -4.85113919e-01 5.67814171e-01
3.73127371e-01 -1.98346704e-01 -6.70145869e-01 -4.18378770e-01
9.30228531e-02 1.43335187e+00 -2.81333208e-01 -2.88473964e-01
-1.69655472e-01 -6.55954838e-01 6.10491574e-01 7.31397927e-01
2.33359262e-01 -4.09455866e-01 -1.48144066e+00 5.99466920e-01
4.83856589e-01 -1.41649544e+00 3.78547460e-01 6.17125273e-01
-9.17491615e-01 -5.41470885e-01 3.37815046e-01 -4.11531292e-02
1.34700716e-01 -3.60041529e-01 1.03223407e+00 -3.17771107e-01
-5.93917966e-01 3.19961429e-01 2.73358673e-02 -1.00988662e+00
-6.40470564e-01 -2.73348659e-01 1.22926049e-01 -6.88820899e-01
5.12316406e-01 -3.91757190e-01 -3.71691316e-01 3.07364762e-01
-7.32194662e-01 3.94820757e-02 5.23225009e-01 6.83333218e-01
7.06382513e-01 5.82809329e-01 5.86015165e-01 -7.76892006e-01
1.63607121e-01 -6.07138216e-01 -1.21958601e+00 1.16298892e-01
-1.13858020e+00 5.18084764e-01 9.53385904e-02 -2.11949021e-01
-1.31132185e+00 5.70119441e-01 -9.42377374e-02 -2.11034492e-01
-1.47881284e-01 3.20603937e-01 -3.84520963e-02 6.64244443e-02
6.84484184e-01 -6.29771233e-01 -4.01778296e-02 2.95497328e-01
4.89339501e-01 9.24629331e-01 5.74592233e-01 -6.25652254e-01
4.49192613e-01 5.03771603e-01 3.29549253e-01 -7.51403749e-01
-4.13560480e-01 -1.43385842e-01 -3.36242795e-01 -2.03589559e-01
6.56075895e-01 -1.04827273e+00 -6.03603065e-01 1.19917266e-01
-1.03763664e+00 -3.07777315e-01 -2.01226309e-01 6.16852999e-01
-9.29279745e-01 1.42603979e-01 -1.68459281e-01 -1.40096855e+00
-2.66047925e-01 -1.33533192e+00 1.13839996e+00 6.03054285e-01
-5.51496804e-01 -6.87166214e-01 -3.25962484e-01 2.29337007e-01
4.75186050e-01 2.93880165e-01 4.20595706e-01 -7.95675039e-01
-8.91960502e-01 -7.04126120e-01 -1.47967204e-01 5.99612556e-02
-3.49700540e-01 3.73264283e-01 -1.41238737e+00 1.81805953e-01
1.55777678e-01 -3.70737165e-01 5.98958910e-01 6.03824139e-01
9.14078236e-01 2.01712474e-01 -7.02005684e-01 7.94238448e-02
1.25559342e+00 4.00050759e-01 4.80752975e-01 2.49972805e-01
1.22951530e-01 6.67672932e-01 1.18999040e+00 6.15890980e-01
4.08103347e-01 4.67399657e-01 6.54980361e-01 7.12570131e-01
2.79999405e-01 -1.68549493e-01 4.48927432e-02 2.57999837e-01
5.14708996e-01 -2.09307432e-01 -1.31096256e+00 5.87201476e-01
-2.22705603e+00 -4.90928441e-01 -1.65895354e-02 2.12910724e+00
5.82182705e-01 8.53569746e-01 -4.81037498e-01 -3.73591930e-02
5.08125186e-01 -2.15231180e-01 -1.07175469e+00 -7.20480323e-01
4.98171061e-01 5.35308197e-03 1.18962169e+00 5.78195095e-01
-9.64265883e-01 1.02995741e+00 6.74262381e+00 9.75810349e-01
-9.80453312e-01 7.87638128e-02 8.68789732e-01 -1.74947567e-02
-3.93940926e-01 3.52356553e-01 -1.11150491e+00 1.99722737e-01
1.93498802e+00 -1.80833936e-01 -1.45529080e-02 1.30850875e+00
1.46069363e-01 -8.70384753e-01 -1.20657885e+00 6.93190396e-01
-2.58004814e-01 -1.03545427e+00 -5.89785635e-01 -4.30249460e-02
5.34073412e-01 2.66217977e-01 1.97910443e-01 3.47959489e-01
8.30473483e-01 -1.25000608e+00 8.98864627e-01 4.85731661e-01
5.61047554e-01 -1.13036835e+00 8.04657221e-01 4.04635608e-01
-6.52783751e-01 8.11463892e-02 -3.80394906e-01 -1.57363340e-02
2.96192586e-01 9.17488575e-01 -1.52399671e+00 2.48275727e-01
5.97528338e-01 1.23845920e-01 -1.52069256e-01 7.96061218e-01
4.69461456e-02 6.72477007e-01 -1.01838350e+00 -1.63998649e-01
1.32588655e-01 2.47266933e-01 2.78640747e-01 1.14233971e+00
3.22476298e-01 1.38026506e-01 1.09277703e-02 7.30860591e-01
5.85890591e-01 -8.18252683e-01 -4.03686404e-01 -3.09220582e-01
5.25378227e-01 9.82900798e-01 -1.21254790e+00 -1.36127889e-01
-8.92655253e-02 5.54757178e-01 -1.62426814e-01 -1.98434576e-01
-1.18248403e+00 -1.66133180e-01 6.83412433e-01 3.99469808e-02
1.64879784e-01 -6.14120424e-01 -9.50104475e-01 -4.69896227e-01
-3.24657530e-01 -4.01492387e-01 7.93733895e-02 -7.04975843e-01
-7.94264734e-01 6.72099590e-01 4.08452570e-01 -7.61180341e-01
-9.25982118e-01 -7.32928455e-01 -2.10064933e-01 7.83163726e-01
-1.29435110e+00 -1.24540627e+00 2.15623647e-01 -3.45900767e-02
6.35290802e-01 -5.45199700e-02 8.13211739e-01 -2.71369725e-01
-2.07852989e-01 3.46513838e-01 5.57759311e-03 -6.86930716e-01
3.84531200e-01 -9.25771296e-01 6.30825222e-01 8.86696458e-01
-2.19275117e-01 1.58385694e-01 1.46778274e+00 -9.67433691e-01
-1.50517523e+00 -1.11002612e+00 3.99605900e-01 -6.25583589e-01
7.43909597e-01 -1.43502548e-01 -5.84001362e-01 6.50109887e-01
-7.41718933e-02 2.10421190e-01 2.74837077e-01 1.83688223e-01
-1.49060324e-01 1.25526100e-01 -1.59006608e+00 3.96024525e-01
3.86214405e-01 -4.05938238e-01 -5.02255142e-01 1.39397997e-02
8.09707403e-01 -8.10723007e-01 -7.19184399e-01 8.19450736e-01
5.21436989e-01 -6.59127355e-01 5.49610972e-01 -1.55319706e-01
8.64104256e-02 -2.37197116e-01 -3.94187570e-01 -7.88384318e-01
3.15379262e-01 -7.26159573e-01 -2.41184801e-01 1.12921381e+00
6.67826176e-01 -5.50063133e-01 1.07461715e+00 1.63180125e+00
-3.11183572e-01 -5.97836018e-01 -1.34702837e+00 -7.47453511e-01
-1.39419362e-01 -1.41721129e+00 6.41481936e-01 1.24459386e-01
-4.06350242e-03 -1.64160088e-01 6.28928514e-03 6.39004111e-01
8.15882087e-01 -1.80721879e-01 3.36130202e-01 -1.05299795e+00
-7.77409300e-02 -8.35696235e-02 -1.10454953e+00 -4.54826772e-01
3.79425377e-01 -2.01496676e-01 8.87679040e-01 -1.36525130e+00
-4.52894578e-03 -8.12285602e-01 -4.44093272e-02 8.63256902e-02
2.96849966e-01 -1.35716170e-01 -2.91473448e-01 -1.46874905e-01
-7.61834025e-01 5.79616129e-01 2.10834444e-01 1.87614895e-02
5.28555829e-03 2.03412428e-01 -2.80058324e-01 1.09129679e+00
9.76706326e-01 -6.18280947e-01 -7.49086916e-01 -1.09728128e-01
3.72321308e-01 1.42843530e-01 2.36494273e-01 -1.42950070e+00
4.75398988e-01 -2.98505247e-01 3.81019294e-01 -7.97730565e-01
4.62822765e-01 -1.11611700e+00 2.59120852e-01 3.82323056e-01
-4.69941087e-02 -2.74504662e-01 8.95721436e-01 1.05933511e+00
7.73553923e-02 -4.46541905e-01 9.55822170e-01 2.19108149e-01
-1.15513098e+00 5.07895499e-02 -8.15121353e-01 -2.67648131e-01
1.37463415e+00 -1.81580126e-01 -4.32086065e-02 -2.34996632e-01
-4.87612844e-01 4.61292773e-01 3.64307910e-01 5.04429400e-01
6.25412047e-01 -8.13617647e-01 -1.65075451e-01 1.89041533e-02
-8.78936797e-02 4.83092427e-01 2.03955382e-01 2.50198305e-01
-7.52059519e-01 6.25641108e-01 -2.05571055e-01 -9.75205183e-01
-8.87127042e-01 3.37444693e-01 2.55232036e-01 -3.41335265e-03
-3.24471563e-01 8.91872346e-01 -2.89633155e-01 -1.62765443e-01
2.62219667e-01 -3.71271998e-01 2.27246419e-01 -4.73507285e-01
3.68870199e-01 2.09513277e-01 1.95193678e-01 -4.06181216e-01
-5.78133285e-01 3.44830230e-02 -1.16333768e-01 -6.85491920e-01
1.07706773e+00 -4.69409436e-01 3.00596029e-01 6.43630028e-01
6.99664712e-01 -5.27718008e-01 -1.83626819e+00 1.28601506e-01
5.78567088e-01 -2.11414754e-01 7.06333518e-01 -9.40623283e-01
-5.71001709e-01 9.14053619e-01 9.51618135e-01 -3.41704249e-01
7.16629982e-01 -1.62334323e-01 7.65893698e-01 2.74945170e-01
8.93756628e-01 -1.52575397e+00 -4.89594370e-01 4.35849637e-01
3.35515469e-01 -1.58053088e+00 4.64717954e-01 -4.49662983e-01
-6.13029957e-01 1.07053220e+00 6.86924160e-01 -1.86830997e-01
1.10891354e+00 9.90664959e-01 -1.11381657e-01 -3.24093737e-02
-1.01771927e+00 5.19268475e-02 1.04428262e-01 8.69605541e-01
1.79259166e-01 2.89625227e-01 1.55801877e-01 7.65036464e-01
-1.75504163e-01 5.02955839e-02 6.07740223e-01 1.21069098e+00
-5.97569168e-01 -8.16924512e-01 -4.23922420e-01 4.67150480e-01
-5.14264584e-01 1.56990752e-01 1.74922809e-01 5.66746652e-01
1.06180027e-01 1.37190843e+00 1.56171083e-01 -5.45595407e-01
-2.91142642e-01 1.08501147e-02 4.13590282e-01 -5.56782186e-01
3.92917951e-04 -5.19080879e-03 5.18755376e-01 -7.20573962e-01
7.90890772e-03 -7.53510296e-01 -1.83663201e+00 -6.69676587e-02
-6.18595064e-01 -7.37470686e-02 1.66331685e+00 1.12567413e+00
4.42469656e-01 7.08111286e-01 8.52441713e-02 -1.14818883e+00
-7.99264193e-01 -7.03803957e-01 -4.43628669e-01 -6.35065615e-01
1.01525247e-01 -7.72755504e-01 -1.28589168e-01 -1.10419206e-01] | [6.966721534729004, -0.34941309690475464] |
b6335061-cde3-4e46-82a5-c461edf06650 | dataset-reproducibility-and-ir-methods-in | null | null | https://aclanthology.org/2020.lrec-1.218 | https://aclanthology.org/2020.lrec-1.218.pdf | Dataset Reproducibility and IR Methods in Timeline Summarization | Timeline summarization (TLS) generates a dated overview of real-world events based on event-specific corpora. The two standard datasets for this task were collected using Google searches for news reports on given events. Not only is this IR method not reproducible at different search times, it also uses components (such as document popularity) that are not always available for any large news corpus. It is unclear how TLS algorithms fare when provided with event corpora collected with varying IR methods. We therefore construct event-specific corpora from a large static background corpus, the newsroom dataset, using differing, relatively simple IR methods based on raw text alone. We show that the choice of IR method plays a crucial role in the performance of various TLS algorithms. A weak TLS algorithm can even match a stronger one by employing a stronger IR method in the data collection phase. Furthermore, the results of TLS systems are often highly sensitive to additional sentence filtering. We consequently advocate for integrating IR into the development of TLS systems and having a common static background corpus for evaluation of TLS systems. | ['Leo Born', 'Katja Markert', 'Maximilian Bacher'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['timeline-summarization'] | ['natural-language-processing'] | [ 1.70278654e-01 -1.46979373e-02 -2.70899266e-01 -3.47044080e-01
-1.60599256e+00 -9.00892496e-01 1.09209812e+00 8.28653932e-01
-8.01935673e-01 7.74778485e-01 9.32020962e-01 -2.20414147e-01
-1.56949490e-01 -5.21370411e-01 -5.99807680e-01 -1.85784489e-01
3.76641542e-01 7.32046247e-01 6.72207475e-01 -5.99193156e-01
6.49885118e-01 5.81267998e-02 -1.35165465e+00 5.72754443e-01
6.19379759e-01 5.83147943e-01 1.82749227e-01 8.35176349e-01
-1.77825645e-01 8.95469785e-01 -1.30865002e+00 -2.38120198e-01
1.44325376e-01 -5.46899021e-01 -9.70422685e-01 -3.84605020e-01
3.25315297e-01 2.16201663e-01 -1.75059676e-01 6.27993822e-01
6.32603943e-01 2.19235748e-01 5.57299256e-01 -6.50896966e-01
-2.92426378e-01 1.09543598e+00 -4.20675874e-01 7.97392964e-01
8.08197975e-01 -2.28966460e-01 1.26097763e+00 -3.12743992e-01
1.04314435e+00 1.08967495e+00 4.34890181e-01 -5.02074324e-03
-1.09346151e+00 -2.28531793e-01 8.19872543e-02 -1.36511952e-01
-1.07448709e+00 -8.06501627e-01 5.74196339e-01 -3.36805612e-01
8.69460464e-01 8.20140898e-01 3.24784219e-01 1.17611456e+00
6.18462898e-02 7.59084105e-01 8.58227193e-01 -7.47221708e-01
2.75919765e-01 2.02534378e-01 5.00553250e-01 -1.02257051e-01
3.56139511e-01 -5.63247144e-01 -7.16020286e-01 -7.34245121e-01
-1.40166909e-01 -4.49146748e-01 -4.22170490e-01 3.06617677e-01
-1.34549260e+00 8.04080248e-01 -1.07322089e-01 7.52977729e-01
-4.47292745e-01 -2.18173593e-01 8.97883594e-01 6.32486045e-01
9.79934990e-01 8.40643883e-01 -5.76709211e-01 -4.63929862e-01
-1.15316534e+00 7.11274207e-01 1.15059400e+00 6.96013153e-01
1.80965289e-01 -4.33538824e-01 -4.79671001e-01 8.70467126e-01
-1.29242897e-01 1.46083370e-01 7.43586540e-01 -8.09930980e-01
7.48658240e-01 4.30257499e-01 5.68713784e-01 -9.11732495e-01
-4.80178475e-01 -2.64867634e-01 -1.50736958e-01 -2.93915451e-01
5.29600382e-01 -1.54422969e-01 -3.88495266e-01 1.71535945e+00
2.75935769e-01 -4.65357482e-01 2.86550671e-01 5.14042199e-01
9.96486545e-01 8.68533731e-01 3.75486016e-02 -7.71425426e-01
1.41871667e+00 -4.99844074e-01 -7.53536284e-01 -3.29571664e-01
8.01474929e-01 -1.09442723e+00 1.08150840e+00 3.22144747e-01
-1.16043532e+00 -9.55321714e-02 -1.01676941e+00 -5.06707728e-02
-3.33610177e-01 -8.15099552e-02 4.53689814e-01 3.96024883e-01
-1.03281832e+00 3.75676513e-01 -5.21668732e-01 -8.00776541e-01
-3.24330240e-01 -4.71272804e-02 -1.66669667e-01 2.47572154e-01
-1.34049904e+00 1.17370045e+00 5.02168655e-01 -5.15745759e-01
-5.91288656e-02 -5.08572519e-01 -4.74077404e-01 -8.47991109e-02
6.84498310e-01 -5.10515809e-01 1.85835040e+00 -8.63847613e-01
-1.17750108e+00 6.95647776e-01 -1.96860939e-01 -5.28479338e-01
4.79318976e-01 -2.02205092e-01 -5.58918417e-01 1.55632749e-01
5.13197482e-01 -3.40679400e-02 3.99535030e-01 -9.30942118e-01
-6.91093326e-01 -2.26828694e-01 6.50383979e-02 4.83464599e-01
-9.21096280e-02 7.18668103e-01 -3.39779466e-01 -6.12204373e-01
-2.51349151e-01 -5.65287948e-01 -1.55578062e-01 -8.33705842e-01
-2.78008193e-01 -5.20370662e-01 4.69436854e-01 -6.17328703e-01
1.70828056e+00 -1.89736164e+00 1.43668801e-02 -8.98042768e-02
-2.12303653e-01 -4.56238240e-02 -5.69670834e-03 8.82689178e-01
2.08910525e-01 3.46632928e-01 1.81488782e-01 5.62995896e-02
-1.95985198e-01 -1.75563604e-01 -5.26041925e-01 2.94460088e-01
-2.99842179e-01 5.36518872e-01 -1.05200541e+00 -6.77864015e-01
-2.64018297e-01 -1.55973300e-01 -3.28440160e-01 -9.30336639e-02
-4.37274516e-01 7.30952844e-02 -7.13721633e-01 3.52591872e-01
-1.31489187e-01 -1.53230771e-01 7.79966563e-02 5.66406995e-02
-6.04702055e-01 9.32852745e-01 -8.99503827e-01 1.85704696e+00
-3.49561960e-01 6.93411946e-01 -2.85277158e-01 -8.17514479e-01
6.99510217e-01 5.46562016e-01 5.75886309e-01 -7.59273529e-01
5.36007248e-02 1.91977143e-01 -2.53357172e-01 -3.26149970e-01
1.14708054e+00 5.89616410e-03 -5.69713831e-01 8.12469959e-01
-9.46780946e-03 -1.39991060e-01 8.84732544e-01 7.23577976e-01
1.36180639e+00 -2.86080033e-01 4.59120870e-01 -4.48195159e-01
9.94804054e-02 6.36942506e-01 4.94711518e-01 1.07778096e+00
1.99874848e-01 7.94317603e-01 6.13446832e-01 -3.33180934e-01
-9.06018913e-01 -5.98408580e-01 -4.05493021e-01 1.28150415e+00
-3.37509671e-04 -7.84521997e-01 -4.68694985e-01 -6.65760517e-01
-5.15851855e-01 1.10335207e+00 -5.14278293e-01 6.56621996e-03
-4.22862172e-01 -1.09228718e+00 6.09778941e-01 -1.51893705e-01
1.02414258e-01 -1.01926970e+00 -7.30539143e-01 6.28489912e-01
-7.13017881e-01 -1.04450977e+00 -5.81143141e-01 1.12250127e-01
-5.31309783e-01 -1.02307415e+00 -5.41080594e-01 -2.13712901e-01
2.10048437e-01 3.76425713e-01 1.49453259e+00 -1.57154262e-01
7.30075613e-02 4.47399110e-01 -7.90235341e-01 -7.74056494e-01
-8.49208057e-01 4.55828965e-01 7.71450996e-02 -2.57693857e-01
3.55053991e-01 -3.91979456e-01 -3.18035901e-01 7.53418431e-02
-1.03032613e+00 -2.68280208e-01 2.40439638e-01 5.40338039e-01
9.60975438e-02 -4.99016158e-02 7.08384931e-01 -1.16548216e+00
1.16180003e+00 -6.86901569e-01 -3.17465842e-01 5.12354374e-01
-5.20958781e-01 8.41209441e-02 3.77407759e-01 -4.46393728e-01
-1.19391358e+00 -4.27998692e-01 9.87644196e-02 3.63208711e-01
1.27895579e-01 8.28774512e-01 3.10280114e-01 6.86388969e-01
1.37209082e+00 -4.13962826e-02 -4.07573104e-01 -6.01824701e-01
1.31216809e-01 8.30174506e-01 5.43012559e-01 -6.21088088e-01
4.84354943e-01 2.41992533e-01 -8.39920342e-01 -8.75967979e-01
-1.14095819e+00 -8.87329400e-01 -1.11220777e-01 -2.70398051e-01
3.99726450e-01 -9.82717156e-01 -9.03870314e-02 9.05311480e-02
-1.25953734e+00 -1.15045793e-01 -5.70991814e-01 4.93916273e-01
-3.01503837e-01 1.72872931e-01 -4.24081802e-01 -6.96559012e-01
-6.08370781e-01 -7.45752633e-01 9.79587734e-01 1.58548981e-01
-9.33774829e-01 -7.52830386e-01 7.07761049e-01 3.49954695e-01
2.20308647e-01 2.81798333e-01 5.47720015e-01 -1.17633295e+00
1.78653494e-01 -4.97238010e-01 1.80047974e-01 -2.62820810e-01
6.25601485e-02 3.26195240e-01 -7.75371015e-01 -1.39195636e-01
-6.69512013e-03 -4.82283831e-01 7.63995528e-01 4.74586844e-01
4.29451048e-01 -5.87848306e-01 -2.29042545e-01 -2.00207993e-01
1.35163546e+00 1.76282078e-01 4.90199059e-01 8.79008114e-01
1.35651991e-01 6.01097822e-01 7.73620069e-01 4.90569919e-01
2.98958868e-01 7.89117217e-01 -2.56632358e-01 1.97131246e-01
2.68632829e-01 -1.60094500e-01 6.42183185e-01 9.66716766e-01
-1.07142672e-01 -5.86780250e-01 -8.04958820e-01 7.27801383e-01
-1.91639149e+00 -1.37067068e+00 -8.19307864e-02 2.28695059e+00
1.15179336e+00 4.58805025e-01 4.23093617e-01 9.53145251e-02
6.37244463e-01 4.50217545e-01 -2.44381651e-01 -4.73964602e-01
-2.79006511e-01 -7.77696893e-02 4.88965809e-01 2.83565253e-01
-8.24177563e-01 5.71317673e-01 6.74013948e+00 6.85414076e-01
-8.83171380e-01 9.52577740e-02 4.33196872e-01 -4.39024478e-01
-4.90582019e-01 1.29000813e-01 -7.57555544e-01 5.86705029e-01
1.44904780e+00 -8.84568870e-01 -7.15975463e-02 7.24804878e-01
4.43030596e-01 -4.99396473e-01 -9.12652731e-01 5.02276123e-01
1.63347483e-01 -1.51990759e+00 -1.50308803e-01 -1.69854179e-01
4.71208483e-01 4.90357935e-01 -5.08173585e-01 4.02098596e-01
7.33124912e-01 -3.83469611e-01 1.14084399e+00 9.29098725e-02
5.80911517e-01 -4.71894830e-01 7.03746736e-01 5.50236106e-01
-6.85803592e-01 2.85479069e-01 -2.34200522e-01 7.96444565e-02
2.81693518e-01 5.61228812e-01 -7.61233687e-01 8.63813937e-01
7.32150614e-01 4.40256298e-01 -7.07593739e-01 8.39161873e-01
2.09024221e-01 8.46968651e-01 -7.07709789e-01 -1.47984013e-01
2.02073589e-01 6.81784749e-02 8.83399785e-01 1.57186747e+00
2.33179107e-01 1.25465065e-01 1.55094281e-01 1.97240114e-01
-4.46501113e-02 4.96525526e-01 -7.10364699e-01 -1.08933821e-02
5.84055960e-01 1.15556455e+00 -8.76483321e-01 -4.88292962e-01
-4.67500329e-01 4.72167373e-01 1.71382591e-01 2.05316305e-01
-4.70023602e-01 -2.15213180e-01 1.05587579e-01 3.72571975e-01
-1.31071404e-01 6.98038787e-02 -3.24992351e-02 -1.38398230e+00
1.01294845e-01 -1.01842070e+00 9.70201373e-01 -8.22029114e-01
-1.03483677e+00 8.44360471e-01 5.01264930e-01 -1.11787224e+00
-5.86487174e-01 1.21632256e-01 -8.06574941e-01 4.05122727e-01
-1.15433085e+00 -5.15140057e-01 1.80452794e-01 3.54498088e-01
8.61905992e-01 -2.01472417e-02 5.69927096e-01 1.57619447e-01
-4.48633909e-01 1.58151388e-01 1.56065404e-01 -8.85269716e-02
1.12120163e+00 -1.14523220e+00 4.43280190e-01 9.01809275e-01
5.65848589e-01 6.72672570e-01 1.33488131e+00 -6.15613282e-01
-1.04958892e+00 -6.98343992e-01 1.27852726e+00 -7.31753945e-01
8.07750106e-01 -4.47290577e-02 -8.81315053e-01 5.58396161e-01
4.35285538e-01 -6.39092028e-01 4.88804221e-01 6.28486931e-01
-2.51159430e-01 -5.81028126e-03 -8.59800637e-01 5.09812653e-01
7.06942201e-01 -2.77656615e-01 -1.26001084e+00 7.01777637e-01
7.61622608e-01 -4.33061838e-01 -5.58371961e-01 2.07040712e-01
2.44092390e-01 -5.95236957e-01 6.71162069e-01 -5.35255790e-01
4.14692134e-01 -2.88771898e-01 -9.15369242e-02 -1.31517792e+00
-1.33443445e-01 -6.85853839e-01 2.74087131e-01 1.44407177e+00
7.93306410e-01 -4.59700555e-01 1.43887162e-01 7.23970711e-01
-3.55042741e-02 -1.70376152e-01 -8.74271989e-01 -6.38709545e-01
-1.15977377e-01 -4.86732930e-01 2.60063320e-01 1.02330899e+00
3.84734273e-01 8.29887927e-01 -7.50223473e-02 -8.09998438e-02
1.89527720e-01 9.42631811e-02 8.01760733e-01 -1.51170599e+00
-9.43001434e-02 -2.91992813e-01 -1.07113793e-02 -3.88865858e-01
-2.63807654e-01 -6.78528130e-01 7.84255192e-03 -1.49558198e+00
3.65875006e-01 -2.84390181e-01 -2.58118898e-01 3.09670120e-01
-2.46308863e-01 -2.21540313e-02 -3.12102735e-02 5.42618811e-01
-9.76082146e-01 1.44697115e-01 5.04715800e-01 2.48296224e-02
-6.52170897e-01 -1.49045640e-03 -1.04899776e+00 7.37684190e-01
8.17055881e-01 -6.71028137e-01 -4.48364198e-01 -2.93427140e-01
6.63810849e-01 2.11470515e-01 -9.72288474e-02 -8.44218314e-01
5.01236260e-01 -2.80221164e-01 -1.35414943e-01 -5.94647348e-01
-2.02870116e-01 -2.32138887e-01 3.37049305e-01 1.76307578e-02
-6.66746438e-01 3.27241719e-01 1.46285012e-01 6.25514865e-01
-2.93449402e-01 -3.59853774e-01 4.73344117e-01 -4.17378128e-01
-4.94065583e-01 -2.02246368e-01 -5.45040667e-01 6.34531200e-01
6.71866894e-01 -7.79756457e-02 -3.87016565e-01 -4.82177943e-01
-2.00649709e-01 4.24412861e-02 4.89916652e-01 5.68754375e-01
-1.16527401e-01 -8.06649208e-01 -1.23043621e+00 -5.95478952e-01
3.63560200e-01 1.01215996e-01 -1.15287296e-01 5.74057937e-01
-3.13907564e-01 3.27923179e-01 2.53746301e-01 -3.33442777e-01
-1.22512543e+00 2.52560556e-01 -2.80228466e-01 -5.83584428e-01
-7.93949962e-01 4.11089331e-01 -9.06461626e-02 -4.24078479e-02
1.32143088e-02 -3.33527885e-02 -4.74187046e-01 6.35175347e-01
9.24536467e-01 1.68970786e-02 4.26562637e-01 -5.12947202e-01
-2.02515885e-01 -6.23344583e-03 -4.91289318e-01 -6.42849565e-01
1.56386685e+00 -2.50870317e-01 1.77984476e-01 9.03895557e-01
7.40834296e-01 3.65817875e-01 -5.42877555e-01 -3.77387583e-01
4.52688068e-01 -1.38687775e-01 1.96108207e-01 -7.50354588e-01
-5.59333622e-01 -2.09764436e-01 -2.80399919e-01 6.64293826e-01
9.57063735e-01 2.44604647e-01 5.32560110e-01 5.19089520e-01
3.96231979e-01 -1.39073980e+00 -3.27604651e-01 2.92963475e-01
1.15358901e+00 -1.09736407e+00 5.81794202e-01 8.36947858e-02
-8.48459303e-01 9.45981383e-01 3.77768949e-02 1.77400455e-01
1.29899859e-01 1.56962827e-01 2.20204026e-01 -4.14104939e-01
-1.18688524e+00 -1.75373983e-02 2.14298680e-01 -1.70459107e-01
5.60852647e-01 -2.97972322e-01 -1.03880203e+00 5.57387590e-01
-4.02590394e-01 -1.51666731e-01 9.02450562e-01 1.12029040e+00
-4.05194521e-01 -1.14120245e+00 -4.24498260e-01 6.15164757e-01
-1.04757440e+00 2.92255972e-02 -6.94776356e-01 9.17483151e-01
-4.97777551e-01 1.05492413e+00 -2.65243016e-02 -1.12869985e-01
4.71478909e-01 1.45402446e-01 5.57392649e-02 -7.69080579e-01
-9.08387065e-01 4.09859642e-02 6.94851637e-01 -2.71423489e-01
-6.42839730e-01 -1.17828596e+00 -7.82856047e-01 -2.72288650e-01
-4.83594030e-01 8.48750889e-01 7.91518867e-01 8.96375537e-01
2.84698278e-01 1.87112093e-01 5.99101007e-01 -6.21572018e-01
-4.99651521e-01 -1.10735774e+00 -2.61282384e-01 4.59520042e-01
4.66282330e-02 -3.54668349e-01 -5.20111859e-01 1.07462972e-01] | [12.397475242614746, 9.420120239257812] |
d51880ed-be7e-4775-bb4e-5e487df648c3 | space-time-separable-graph-convolutional-1 | 2110.04573 | null | https://arxiv.org/abs/2110.04573v1 | https://arxiv.org/pdf/2110.04573v1.pdf | Space-Time-Separable Graph Convolutional Network for Pose Forecasting | Human pose forecasting is a complex structured-data sequence-modelling task, which has received increasing attention, also due to numerous potential applications. Research has mainly addressed the temporal dimension as time series and the interaction of human body joints with a kinematic tree or by a graph. This has decoupled the two aspects and leveraged progress from the relevant fields, but it has also limited the understanding of the complex structural joint spatio-temporal dynamics of the human pose. Here we propose a novel Space-Time-Separable Graph Convolutional Network (STS-GCN) for pose forecasting. For the first time, STS-GCN models the human pose dynamics only with a graph convolutional network (GCN), including the temporal evolution and the spatial joint interaction within a single-graph framework, which allows the cross-talk of motion and spatial correlations. Concurrently, STS-GCN is the first space-time-separable GCN: the space-time graph connectivity is factored into space and time affinity matrices, which bottlenecks the space-time cross-talk, while enabling full joint-joint and time-time correlations. Both affinity matrices are learnt end-to-end, which results in connections substantially deviating from the standard kinematic tree and the linear-time time series. In experimental evaluation on three complex, recent and large-scale benchmarks, Human3.6M [Ionescu et al. TPAMI'14], AMASS [Mahmood et al. ICCV'19] and 3DPW [Von Marcard et al. ECCV'18], STS-GCN outperforms the state-of-the-art, surpassing the current best technique [Mao et al. ECCV'20] by over 32% in average at the most difficult long-term predictions, while only requiring 1.7% of its parameters. We explain the results qualitatively and illustrate the graph interactions by the factored joint-joint and time-time learnt graph connections. Our source code is available at: https://github.com/FraLuca/STSGCN | ['Fabio Galasso', 'Luca Franco', 'Alessio Sampieri', 'Theodoros Sofianos'] | 2021-10-09 | space-time-separable-graph-convolutional | http://openaccess.thecvf.com//content/ICCV2021/html/Sofianos_Space-Time-Separable_Graph_Convolutional_Network_for_Pose_Forecasting_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Sofianos_Space-Time-Separable_Graph_Convolutional_Network_for_Pose_Forecasting_ICCV_2021_paper.pdf | iccv-2021-1 | ['human-pose-forecasting'] | ['computer-vision'] | [-2.45884255e-01 1.49307072e-01 -6.49958327e-02 1.28624411e-02
-1.95226237e-01 -4.10180598e-01 5.74634314e-01 -1.37042150e-01
-3.68453681e-01 4.22247529e-01 2.45673865e-01 -1.25872910e-01
-3.94503027e-01 -4.46810633e-01 -8.12050641e-01 -4.99634832e-01
-8.70500624e-01 5.87424695e-01 2.81618595e-01 -5.50963700e-01
-3.31255436e-01 3.76965284e-01 -1.25440776e+00 -1.74529240e-01
4.29337442e-01 8.44210446e-01 -2.28392538e-02 8.83024931e-01
5.05680561e-01 6.55975580e-01 -6.73375502e-02 -3.93538862e-01
2.72703499e-01 -3.40974420e-01 -7.09803820e-01 -6.91693276e-02
5.23062646e-01 -2.92899329e-02 -8.86362135e-01 4.35102135e-01
4.97585535e-01 3.69282097e-01 4.10323292e-01 -1.60211670e+00
-3.02333206e-01 3.51212651e-01 -5.20982206e-01 1.44568518e-01
3.15202862e-01 3.46161485e-01 9.76215839e-01 -5.90949953e-01
8.13637137e-01 1.43785453e+00 1.05787420e+00 5.14786243e-01
-1.19867206e+00 -4.38508242e-01 4.34669346e-01 4.29196239e-01
-1.34399259e+00 1.78241432e-01 8.60397518e-01 -6.58827245e-01
1.28838980e+00 2.06884161e-01 1.22552991e+00 1.39993751e+00
7.56738722e-01 8.52878690e-01 6.73563182e-01 3.42126898e-02
-4.72554117e-02 -1.04256260e+00 6.32250980e-02 9.53569770e-01
-2.84027867e-02 2.21533135e-01 -6.59088194e-01 -4.63133715e-02
9.16230023e-01 1.23854965e-01 -2.04111233e-01 -6.88689709e-01
-1.41705537e+00 5.79377115e-01 8.51141870e-01 2.17291906e-01
-3.41571838e-01 6.92833602e-01 4.98106629e-01 1.53006911e-01
5.90239465e-01 7.66266584e-02 -6.01406276e-01 -4.02987629e-01
-5.92212796e-01 6.32498503e-01 9.21536446e-01 7.92480469e-01
3.65797698e-01 1.14213303e-02 7.36843888e-03 5.42702854e-01
3.57161939e-01 5.60022056e-01 1.09089307e-01 -8.05560350e-01
6.08491480e-01 5.03659487e-01 -1.90155998e-01 -1.40126121e+00
-1.15443778e+00 -7.10085750e-01 -1.21863067e+00 6.14931760e-03
6.33851111e-01 -1.89542770e-01 -1.09316647e+00 2.02350640e+00
4.36290205e-01 4.61206704e-01 -5.24059117e-01 1.20737422e+00
4.06844795e-01 5.82616925e-01 7.94281214e-02 1.47318989e-01
1.39684916e+00 -1.13027227e+00 -5.90099096e-01 -4.10753399e-01
8.09083998e-01 -4.05922025e-01 7.36818492e-01 3.33290786e-01
-9.20209646e-01 -8.00684929e-01 -9.51922834e-01 -1.18307635e-01
-1.90763593e-01 -4.82200831e-02 7.06762552e-01 1.63037345e-01
-1.18901730e+00 8.96955431e-01 -1.41705227e+00 -4.46060479e-01
4.60155085e-02 5.41183054e-01 -5.39231002e-01 -5.01625240e-02
-1.40410352e+00 8.64948630e-01 7.06238598e-02 5.25331974e-01
-6.20541632e-01 -7.41624594e-01 -9.43447888e-01 -1.62278965e-01
4.97140825e-01 -1.09507012e+00 9.54397500e-01 -4.36513215e-01
-1.30602825e+00 4.25829977e-01 2.09402695e-01 -6.66342914e-01
9.04161870e-01 -5.92272103e-01 -4.06432092e-01 7.71812275e-02
-7.85401314e-02 6.49785876e-01 5.29383540e-01 -7.74998903e-01
-2.95088917e-01 -5.50940037e-01 -6.15072139e-02 3.87980789e-01
5.78079410e-02 -4.88304645e-01 -8.63245487e-01 -8.72095704e-01
1.20349407e-01 -1.65908718e+00 -3.65729332e-01 5.58242127e-02
-4.49426919e-01 -1.40509024e-01 6.22184336e-01 -9.57419574e-01
1.35491645e+00 -1.97276223e+00 9.23723698e-01 1.06901526e-01
5.59231400e-01 1.25759080e-01 -1.88192874e-01 8.02089334e-01
-1.90369189e-01 -2.03144357e-01 -1.69406757e-01 -6.75192416e-01
7.45610818e-02 4.40161377e-01 2.05987766e-01 8.07899892e-01
-1.39858983e-02 1.27214336e+00 -9.18169022e-01 -2.20505908e-01
3.41668189e-01 8.28756869e-01 -5.40942609e-01 -4.98777442e-02
-1.21374533e-01 6.23241782e-01 -2.41267920e-01 2.91567445e-01
1.81659445e-01 -3.80367279e-01 2.04004988e-01 -3.90499771e-01
1.17952943e-01 3.81761082e-02 -1.04212499e+00 2.05750227e+00
-1.60838410e-01 6.06787264e-01 -1.04646929e-01 -8.76444817e-01
6.13160789e-01 2.76188701e-01 1.06166279e+00 -7.70183742e-01
1.51650250e-01 5.33214249e-02 2.45470405e-01 -3.36857796e-01
3.98351252e-01 2.15984493e-01 -2.91102231e-01 4.40276749e-02
2.31246680e-01 2.39560992e-01 1.92455515e-01 4.40184474e-01
1.38775826e+00 5.09634674e-01 -4.81913164e-02 -3.09330910e-01
1.69106960e-01 -1.08373143e-01 5.00703990e-01 3.59513551e-01
-1.91589966e-01 6.75782263e-01 4.90354747e-01 -5.41966438e-01
-9.88546014e-01 -1.13899684e+00 4.75267977e-01 9.57533956e-01
1.87382460e-01 -8.02492380e-01 -6.51734412e-01 -5.14416099e-01
2.84142345e-01 2.09234536e-01 -9.83357668e-01 -3.85602683e-01
-9.41450894e-01 -3.75771910e-01 4.83561665e-01 6.16937995e-01
2.49412775e-01 -7.82406509e-01 -6.33335114e-01 3.98758471e-01
-3.34073186e-01 -1.16072083e+00 -7.32892573e-01 1.34227462e-02
-8.51754487e-01 -1.07548547e+00 -7.41452336e-01 -3.47736180e-01
1.75607443e-01 5.84884025e-02 1.13648140e+00 1.30683303e-01
-6.00428879e-01 6.32592797e-01 -3.50425065e-01 -3.15412916e-02
6.59088343e-02 2.47215971e-01 1.91441536e-01 1.70678068e-02
-1.19925879e-01 -9.00782168e-01 -9.04648840e-01 3.64967674e-01
-6.23056054e-01 3.54963779e-01 4.14509058e-01 8.82862926e-01
4.84423190e-01 -3.87297153e-01 1.52092144e-01 -4.77884084e-01
2.69163877e-01 -5.39167821e-01 -2.74649352e-01 -1.38647035e-02
-4.73352700e-01 -2.13933066e-02 2.92021900e-01 -6.20601296e-01
-5.04974544e-01 1.98941752e-01 -4.65286262e-02 -7.13348866e-01
1.59692690e-01 6.98599160e-01 1.21268168e-01 6.91879019e-02
5.02435565e-01 5.36262877e-02 2.51675665e-01 -5.75790584e-01
5.47407389e-01 -1.57132968e-01 7.05587029e-01 -4.85126853e-01
8.41304302e-01 4.48615551e-01 4.26062077e-01 -7.23769546e-01
-5.26754260e-01 -6.04740500e-01 -9.78919864e-01 -6.17706537e-01
1.05266941e+00 -9.57589567e-01 -7.73211241e-01 7.51611590e-01
-1.03751516e+00 -7.82579303e-01 -6.43895194e-02 6.62866354e-01
-7.88548589e-01 4.36517328e-01 -9.66800869e-01 -6.06175244e-01
-1.63849637e-01 -8.25699687e-01 1.26128125e+00 -2.08652973e-01
-4.81161863e-01 -1.13515544e+00 2.42756262e-01 2.39432663e-01
7.07311481e-02 7.90372252e-01 6.52518511e-01 -2.65745610e-01
-4.00860608e-01 -2.68448800e-01 1.26760498e-01 -8.29541404e-03
-1.24207892e-01 -1.98361307e-01 -3.72936189e-01 -5.55365443e-01
-4.34096038e-01 1.76413115e-02 7.35639632e-01 5.65498173e-01
7.15873241e-01 -6.07940182e-02 -4.74433750e-01 7.63372719e-01
1.07664180e+00 -1.35577604e-01 5.23280919e-01 1.80215657e-01
1.37640285e+00 5.31023085e-01 4.83086467e-01 2.41798580e-01
7.41686285e-01 9.64962959e-01 5.36100090e-01 -1.52092651e-01
-3.12216729e-01 -3.73692840e-01 3.78431112e-01 1.20955622e+00
-6.37552679e-01 -2.57373303e-01 -1.14770722e+00 4.83237326e-01
-2.48434591e+00 -7.77633131e-01 -5.05039394e-01 1.92982829e+00
3.08987021e-01 1.19554423e-01 4.89932418e-01 2.74111032e-02
4.16659117e-01 3.32232833e-01 -5.81804633e-01 -3.73978689e-02
1.35651127e-01 3.25485617e-02 5.73137999e-01 5.40320635e-01
-1.10604608e+00 7.70096362e-01 4.97296715e+00 5.75331569e-01
-9.69006181e-01 4.21041101e-02 3.17650139e-01 -3.57196957e-01
1.86768949e-01 -7.41922557e-02 -5.50369084e-01 3.51341695e-01
9.67811644e-01 5.11007942e-02 4.56961840e-01 4.97649133e-01
4.80608493e-02 9.18782949e-02 -1.23917139e+00 1.02511048e+00
-9.03740525e-02 -9.98714089e-01 -2.82720387e-01 3.74759674e-01
3.26429307e-01 3.45889509e-01 -7.81889632e-02 3.92446220e-01
2.02866748e-01 -9.57522988e-01 9.60592806e-01 7.24086583e-01
4.71822172e-01 -6.53565586e-01 4.50310200e-01 3.52812201e-01
-1.76919818e+00 -2.82453317e-02 1.63466737e-01 -2.02845201e-01
4.56625521e-01 4.59454000e-01 -2.49771044e-01 9.11658287e-01
8.33188057e-01 1.05952561e+00 -6.11443043e-01 7.85707891e-01
2.66026729e-03 5.95962226e-01 -5.10166824e-01 1.28870651e-01
3.69262725e-01 -2.14399457e-01 7.08356619e-01 1.12673044e+00
2.03704223e-01 9.00110379e-02 4.94045258e-01 5.10875344e-01
3.19917738e-01 -2.31288522e-01 -4.56782520e-01 -6.67841733e-02
-6.04008213e-02 1.07301950e+00 -8.41406226e-01 -1.01854568e-02
-2.80710459e-01 1.11662781e+00 4.43031639e-01 4.34552997e-01
-1.06919909e+00 -2.47817058e-02 7.32279420e-01 2.92853475e-01
2.93617576e-01 -8.68120432e-01 7.74829313e-02 -1.11829126e+00
2.91845858e-01 -6.43530965e-01 5.06558239e-01 -5.75903893e-01
-1.23409486e+00 5.24441421e-01 2.67574102e-01 -1.19674039e+00
-4.87010956e-01 -6.61221981e-01 -1.41073048e-01 7.22919285e-01
-8.12669635e-01 -1.42991734e+00 -2.87949353e-01 8.05544734e-01
3.63926321e-01 2.37697899e-01 5.70562005e-01 3.37072045e-01
-4.71809179e-01 4.52714622e-01 -3.67220253e-01 2.65655041e-01
4.67702806e-01 -1.23408926e+00 1.01099801e+00 6.72797501e-01
1.27914295e-01 4.41064745e-01 7.33290136e-01 -1.00812769e+00
-1.82213557e+00 -9.79109526e-01 8.01778793e-01 -5.97333908e-01
9.65355992e-01 -7.85449505e-01 -9.41136122e-01 7.63625503e-01
-3.07164766e-04 3.13896030e-01 2.11323962e-01 2.06700549e-01
-4.47244346e-01 7.73726851e-02 -4.96934444e-01 8.13334227e-01
1.71501923e+00 -3.73532236e-01 -2.38311350e-01 4.19896275e-01
8.24122846e-01 -7.44039536e-01 -1.05529308e+00 6.38143659e-01
9.37039912e-01 -8.58426929e-01 1.10752273e+00 -6.20881796e-01
3.57528657e-01 -2.18131348e-01 -1.40452925e-02 -1.36178088e+00
-6.57757759e-01 -6.62113130e-01 -5.29701054e-01 6.41681969e-01
2.37864628e-01 -4.85047042e-01 7.37993002e-01 3.87792170e-01
-2.15353072e-01 -1.14320588e+00 -1.10291493e+00 -1.04880428e+00
-2.03966703e-02 -5.39458156e-01 2.13411152e-01 8.07495356e-01
-1.99890167e-01 4.55145597e-01 -7.95842350e-01 4.05809544e-02
6.52794778e-01 -2.57214189e-01 8.83639157e-01 -1.25034845e+00
-5.17032444e-01 -4.16132540e-01 -8.36458087e-01 -1.07810032e+00
-9.88471881e-02 -7.36118138e-01 6.43488690e-02 -1.55893195e+00
-9.96479914e-02 9.62723047e-03 -1.92299247e-01 4.05380011e-01
-1.10415347e-01 2.18157470e-02 5.48795521e-01 9.88932252e-02
-5.30236185e-01 6.34231985e-01 1.32965422e+00 -3.38742547e-02
-5.80494776e-02 -1.64551973e-01 1.01784401e-01 5.23450792e-01
6.91832781e-01 -2.14487165e-01 -6.59739494e-01 -3.52970272e-01
3.93334687e-01 3.91262650e-01 7.67536759e-01 -1.38643479e+00
3.42149734e-01 5.45493178e-02 3.05876315e-01 -6.66776359e-01
6.19989932e-01 -7.55404472e-01 6.60861969e-01 7.67363131e-01
-1.22441694e-01 5.99912822e-01 3.04948449e-01 1.08982110e+00
3.30477133e-02 8.16793740e-01 3.80164623e-01 6.10838048e-02
-6.93869472e-01 7.27022052e-01 -2.32686430e-01 -5.94860874e-03
8.79412353e-01 -1.64753169e-01 -1.51462302e-01 -5.42203069e-01
-1.09107125e+00 4.08861935e-01 2.35677585e-01 8.78213108e-01
3.63427281e-01 -1.46716428e+00 -5.84953785e-01 -3.00198840e-03
2.01863777e-02 -3.12038921e-02 7.60884643e-01 1.27667105e+00
-4.87917781e-01 3.34739357e-01 -2.92518735e-01 -9.99719501e-01
-1.14161444e+00 5.38167655e-01 3.64745557e-01 -5.90380788e-01
-1.05046022e+00 7.73366153e-01 1.84581369e-01 -4.73213136e-01
2.02966422e-01 -4.31662858e-01 3.70247029e-02 -1.90383732e-01
-1.41239529e-02 4.95153427e-01 -9.27680135e-02 -9.55094576e-01
-4.89901632e-01 8.08143675e-01 3.18398356e-01 -9.96380225e-02
1.28270447e+00 -1.84025452e-01 1.18550165e-02 7.50290930e-01
1.35113883e+00 -5.03196537e-01 -1.46023023e+00 -1.49133712e-01
-6.42049238e-02 -9.98371020e-02 -2.23706380e-01 -6.96890771e-01
-1.27767968e+00 8.67671669e-01 5.56167185e-01 1.35973066e-01
8.76681924e-01 -1.68659598e-01 9.77908432e-01 1.46022081e-01
6.46102846e-01 -9.06804442e-01 1.63536176e-01 7.70901740e-01
1.21441531e+00 -8.12901676e-01 1.37574822e-01 -5.56684017e-01
-4.57734764e-01 9.77308273e-01 7.12957263e-01 -4.82955605e-01
9.56719041e-01 1.41203418e-01 -1.33785129e-01 -4.36811119e-01
-8.54013681e-01 -1.12951711e-01 8.05829644e-01 5.22685945e-01
4.47934389e-01 3.99587631e-01 -2.93686360e-01 4.78102833e-01
-3.12491357e-01 -1.16537862e-01 -1.51183099e-01 9.24170196e-01
7.94933289e-02 -9.92635131e-01 1.36748385e-02 2.14908630e-01
-7.32701793e-02 2.12086961e-01 -4.18403774e-01 1.25504839e+00
2.57326681e-02 5.91973364e-01 -6.40496090e-02 -9.06724632e-01
5.88352501e-01 -3.86065133e-02 5.67457199e-01 -3.10638279e-01
-6.63149536e-01 3.20738226e-01 2.40589961e-01 -1.23026967e+00
-3.64345014e-01 -7.00432837e-01 -1.27758467e+00 -4.58894581e-01
8.57764482e-03 -1.29863113e-01 4.77870882e-01 8.69263411e-01
6.13574982e-01 8.30378413e-01 9.15481448e-02 -1.34071350e+00
-2.58921862e-01 -9.68959689e-01 -5.98779321e-01 5.72674155e-01
3.44705105e-01 -1.00832295e+00 -2.01112494e-01 -8.40511844e-02] | [7.287280082702637, -0.3189002275466919] |
3254edf1-7bb4-4efa-9f8b-7cb461598345 | towards-moocs-for-lip-reading-using-synthetic | 2208.09796 | null | https://arxiv.org/abs/2208.09796v2 | https://arxiv.org/pdf/2208.09796v2.pdf | Towards MOOCs for Lipreading: Using Synthetic Talking Heads to Train Humans in Lipreading at Scale | Many people with some form of hearing loss consider lipreading as their primary mode of day-to-day communication. However, finding resources to learn or improve one's lipreading skills can be challenging. This is further exacerbated in the COVID19 pandemic due to restrictions on direct interactions with peers and speech therapists. Today, online MOOCs platforms like Coursera and Udemy have become the most effective form of training for many types of skill development. However, online lipreading resources are scarce as creating such resources is an extensive process needing months of manual effort to record hired actors. Because of the manual pipeline, such platforms are also limited in vocabulary, supported languages, accents, and speakers and have a high usage cost. In this work, we investigate the possibility of replacing real human talking videos with synthetically generated videos. Synthetic data can easily incorporate larger vocabularies, variations in accent, and even local languages and many speakers. We propose an end-to-end automated pipeline to develop such a platform using state-of-the-art talking head video generator networks, text-to-speech models, and computer vision techniques. We then perform an extensive human evaluation using carefully thought out lipreading exercises to validate the quality of our designed platform against the existing lipreading platforms. Our studies concretely point toward the potential of our approach in developing a large-scale lipreading MOOC platform that can impact millions of people with hearing loss. | ['C. V Jawahar', 'Vinay Namboodiri', 'Rudrabha Mukhopadhyay', 'Bipasha Sen', 'Aditya Agarwal'] | 2022-08-21 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [-1.07815817e-01 3.59288365e-01 -1.93196699e-01 -2.15162393e-02
-1.09024537e+00 -3.72335523e-01 4.39105183e-01 -3.76016885e-01
-5.03055632e-01 8.27870488e-01 8.11971486e-01 -2.97819942e-01
1.09556213e-01 -2.62654454e-01 -5.35344005e-01 -8.82911608e-02
6.02341950e-01 5.65204024e-01 2.17942163e-01 -4.69440997e-01
1.48560345e-01 9.80339274e-02 -2.05422664e+00 3.18182021e-01
1.02800834e+00 4.32640702e-01 5.67511261e-01 9.23193097e-01
-2.31327608e-01 7.23282039e-01 -7.65504420e-01 -5.36573887e-01
-2.21244525e-02 -3.78526509e-01 -6.26773417e-01 1.64484695e-01
5.66543639e-01 -6.32886589e-01 -3.68547648e-01 8.13026607e-01
1.19811201e+00 -4.34223376e-02 1.01909742e-01 -1.25766838e+00
-4.51728612e-01 5.67956805e-01 5.07217878e-03 -1.93837583e-01
8.88388932e-01 7.36211777e-01 5.56933403e-01 -8.44626486e-01
6.01470232e-01 1.11512971e+00 5.90676725e-01 8.51656735e-01
-1.02716768e+00 -5.94849169e-01 -2.91408032e-01 3.85420680e-01
-1.13976765e+00 -1.22272873e+00 5.49401224e-01 -4.35711712e-01
8.73940587e-01 1.02546357e-01 9.11406934e-01 1.51275194e+00
-5.57282627e-01 5.76022148e-01 9.54368651e-01 -8.27300906e-01
6.89335242e-02 5.35848081e-01 -2.14843512e-01 4.11850959e-01
9.10979521e-04 -3.00336093e-01 -1.20623839e+00 2.59247839e-01
2.08990335e-01 -3.01073998e-01 -7.30944455e-01 3.26551571e-02
-1.12553048e+00 5.39694548e-01 -2.60359734e-01 1.19665340e-01
-4.67066728e-02 -7.15772957e-02 3.71576220e-01 2.19997957e-01
2.01520100e-01 9.32145268e-02 -2.04233065e-01 -7.41278708e-01
-1.14355433e+00 1.60673022e-01 8.92041564e-01 9.26368952e-01
2.85950243e-01 -3.00494283e-01 -2.34465197e-01 1.43062317e+00
3.87124509e-01 4.16399211e-01 8.25786352e-01 -1.05403900e+00
4.52910811e-01 3.90915602e-01 -1.84264611e-02 -2.89892703e-01
-2.19744414e-01 2.01804005e-02 -2.34271660e-01 3.78172874e-01
6.83447182e-01 -2.89321989e-01 -8.30694497e-01 1.79056394e+00
1.67559519e-01 -9.84280631e-02 -2.46680900e-01 7.14199483e-01
8.48255515e-01 1.55883104e-01 -8.91111568e-02 -1.41735032e-01
1.30648351e+00 -1.14502251e+00 -8.42463672e-01 -2.40824744e-01
4.30079818e-01 -9.74528432e-01 1.60649848e+00 3.76687348e-01
-1.41197562e+00 -3.27150971e-01 -6.22797966e-01 -1.92512482e-01
-2.01099202e-01 1.22338962e-02 3.92699718e-01 1.10614765e+00
-1.53866553e+00 1.95844397e-01 -3.13463897e-01 -7.06875980e-01
5.02110720e-01 2.92840481e-01 -6.19740307e-01 -3.14374328e-01
-1.05851948e+00 7.54639685e-01 -1.60899401e-01 -3.89136434e-01
-6.59320772e-01 -8.67902339e-01 -7.43981004e-01 3.78371850e-02
1.87828019e-01 -6.67325675e-01 1.44023061e+00 -8.97306800e-01
-1.86962640e+00 1.02400231e+00 -2.10631862e-01 -1.35006651e-01
9.00651276e-01 -1.84105486e-01 -3.36522251e-01 2.69328862e-01
-3.43086645e-02 8.87753665e-01 9.37855005e-01 -6.37942851e-01
-5.65258443e-01 -1.36860430e-01 1.98394228e-02 2.89241761e-01
-5.06879926e-01 3.87504339e-01 -3.64060938e-01 -3.51490438e-01
-4.45122808e-01 -7.49794185e-01 5.04786491e-01 3.43468130e-01
-3.52459759e-01 -8.53458270e-02 6.85745835e-01 -1.20536172e+00
8.54366064e-01 -2.26646805e+00 -2.19122961e-01 -2.82954842e-01
2.44360894e-01 4.19458479e-01 9.54890251e-02 5.04857421e-01
3.04047734e-01 2.86943376e-01 8.19706172e-02 -5.94797611e-01
2.76747525e-01 -2.24321172e-01 6.30499870e-02 1.04897171e-01
-2.40896076e-01 5.63220561e-01 -6.86381876e-01 -5.10641992e-01
1.23131834e-01 7.80743897e-01 -7.45431006e-01 3.44499290e-01
-1.61748424e-01 3.29439700e-01 5.19007444e-01 6.93574429e-01
3.81895095e-01 2.57433578e-02 1.94838122e-02 3.80237967e-01
-2.23665923e-01 4.72062826e-01 -9.94986832e-01 1.74409199e+00
-7.61145711e-01 1.07272279e+00 3.83196145e-01 -3.56920838e-01
4.06600744e-01 4.89131063e-01 1.82112530e-01 -2.77771980e-01
2.92344615e-02 2.45446414e-01 6.32179081e-02 -9.30796921e-01
2.59320408e-01 -1.84199631e-01 4.90270138e-01 3.02784920e-01
1.60864547e-01 -3.42018336e-01 1.50226608e-01 8.30645114e-02
1.11723423e+00 -1.45607129e-01 -7.04692528e-02 1.53931305e-01
4.51192170e-01 -3.34898651e-01 1.42308429e-01 5.55134475e-01
-6.19267166e-01 8.47893953e-01 3.67109537e-01 3.13873351e-01
-1.10492861e+00 -9.87910271e-01 6.61509410e-02 1.28775418e+00
-5.99819601e-01 -3.27971697e-01 -1.11592948e+00 -7.99246132e-02
-3.78260911e-01 6.59807920e-01 6.01590164e-02 1.48285434e-01
3.25170234e-02 7.63236657e-02 8.30128670e-01 1.61935091e-01
6.19427443e-01 -1.39761984e+00 -4.20886695e-01 7.09424019e-02
-5.23374796e-01 -1.28618491e+00 -8.79520774e-01 -7.01509476e-01
5.40719135e-04 -7.68643022e-01 -1.27001417e+00 -1.05697775e+00
3.05577964e-01 1.85453773e-01 7.25789309e-01 -1.50136411e-01
-4.63853329e-01 6.64793372e-01 -2.55727589e-01 -6.57921314e-01
-7.24499166e-01 3.33786085e-02 3.91566008e-01 -1.04321837e-01
2.39762336e-01 -6.30300522e-01 -6.75551593e-01 3.26953441e-01
-6.41235530e-01 2.51133263e-01 4.51613396e-01 6.90136492e-01
-7.72663206e-02 -2.65062451e-01 8.51414979e-01 -1.69647634e-01
1.03759491e+00 -2.14748487e-01 -4.62843597e-01 2.03438789e-01
-3.58183801e-01 -3.79797369e-01 2.00635463e-01 -7.87328005e-01
-9.71023440e-01 -1.85677424e-01 -3.01160425e-01 -3.57585609e-01
-2.67859429e-01 8.80103633e-02 -3.95619333e-01 1.94036290e-02
5.29298007e-01 2.12207049e-01 5.87464988e-01 -4.54390764e-01
3.12993884e-01 1.54912913e+00 6.53015256e-01 -1.63595945e-01
4.83728826e-01 2.07280353e-01 -7.05882668e-01 -1.36698949e+00
-2.49715433e-01 -5.18599987e-01 -1.52767703e-01 -5.99745870e-01
7.38574207e-01 -9.68829215e-01 -1.18950009e+00 1.07891679e+00
-1.13114858e+00 -6.79735124e-01 -2.20761284e-01 4.79578644e-01
-6.59270942e-01 2.88060546e-01 -2.60053098e-01 -8.83062661e-01
-3.12202722e-01 -1.37129903e+00 8.79537165e-01 4.35345501e-01
-4.20058966e-01 -6.74746573e-01 1.44206062e-01 1.17021704e+00
6.95480347e-01 -4.89570975e-01 5.43572664e-01 -4.13606763e-01
-4.37591851e-01 -2.04910472e-01 -1.17774017e-01 6.05024159e-01
3.88840288e-02 -3.20034698e-02 -1.19344127e+00 -2.62134701e-01
-3.95508945e-01 -7.97975421e-01 2.56959051e-01 3.55318606e-01
8.70938241e-01 -4.10402060e-01 -5.42772301e-02 3.21499437e-01
8.43243122e-01 -3.16999137e-01 7.82799363e-01 1.29929423e-01
4.22841817e-01 1.07028317e+00 1.28648400e-01 4.52127248e-01
6.01547182e-01 9.12956715e-01 -4.47472148e-02 5.24606332e-02
-7.84874856e-01 -5.31138659e-01 7.71328866e-01 8.57945859e-01
1.16290823e-01 -1.58037812e-01 -1.12172401e+00 8.18833470e-01
-1.21050167e+00 -1.17434883e+00 3.14251721e-01 2.39245605e+00
1.12577021e+00 -6.06455691e-02 5.25988281e-01 3.18335086e-01
9.49043453e-01 -3.15488338e-01 -3.76458257e-01 -2.50061333e-01
-4.31831293e-02 2.33055010e-01 2.23129421e-01 6.75058603e-01
-2.80969381e-01 9.79075015e-01 6.12578487e+00 7.15607285e-01
-1.23437858e+00 3.13348085e-01 2.90367186e-01 -5.82248807e-01
-1.06717624e-01 -2.97885448e-01 -8.20558071e-01 6.91699982e-01
1.11226010e+00 -2.67384015e-03 9.00489032e-01 5.84998310e-01
6.14090800e-01 -1.49073422e-01 -8.80078256e-01 1.21570587e+00
3.50753099e-01 -1.26150334e+00 -2.99868017e-01 1.54236808e-01
2.66697377e-01 2.93807834e-01 -5.30493492e-03 2.66960144e-01
1.56368110e-02 -1.17923331e+00 8.48817587e-01 4.25352544e-01
1.42191684e+00 -4.03953850e-01 3.60315472e-01 3.75337183e-01
-6.84905946e-01 -1.15131550e-01 -2.16527861e-02 1.14080690e-01
2.95502663e-01 2.25118101e-01 -1.25081503e+00 -3.10189992e-01
7.60030389e-01 -5.24912775e-02 -4.05534357e-01 1.31195176e+00
-2.77852148e-01 5.61558783e-01 -4.57805306e-01 -2.52930403e-01
-4.45787400e-01 3.78202468e-01 6.80879056e-01 9.54843223e-01
4.82118785e-01 -2.54629880e-01 -2.99955398e-01 6.19249284e-01
-3.58767360e-01 2.97953457e-01 -7.09479511e-01 -3.35125290e-02
6.91126585e-01 9.26246405e-01 -2.88411546e-02 1.49344549e-01
-6.26423001e-01 8.02741826e-01 1.20778479e-01 1.64089561e-01
-5.17716289e-01 -5.78718960e-01 9.07011390e-01 7.97188401e-01
-1.89004257e-01 -2.64031310e-02 5.64114423e-03 -9.89814103e-01
2.19124466e-01 -1.51660252e+00 -1.50904253e-01 -1.09392834e+00
-7.59670436e-01 2.56232113e-01 -4.25526470e-01 -9.55264986e-01
-2.94607460e-01 -5.94164073e-01 -5.67141891e-01 9.57402885e-01
-1.30392754e+00 -1.18728173e+00 -6.16671801e-01 8.17869663e-01
8.79104316e-01 -5.56004226e-01 6.10197008e-01 4.17427123e-01
-4.15717393e-01 1.04402423e+00 -1.87127948e-01 3.21621932e-02
1.12673676e+00 -6.58768773e-01 -5.42206392e-02 5.83553553e-01
-1.24048583e-01 4.32743579e-01 6.93233907e-01 -4.39863235e-01
-1.06073332e+00 -5.42073905e-01 1.08238029e+00 -3.58348638e-01
7.05087364e-01 -7.27011561e-01 -5.84275663e-01 3.38689983e-01
3.67587179e-01 -4.26662773e-01 7.52280533e-01 -2.05991510e-02
-4.54226658e-02 -1.77375749e-01 -1.23241198e+00 8.83439481e-01
1.36597800e+00 -8.68893147e-01 -2.59315312e-01 5.01544714e-01
9.82615590e-01 -1.01685271e-01 -5.94362617e-01 -1.11619808e-01
7.88129210e-01 -1.10157275e+00 7.18448997e-01 -2.69584090e-01
3.72292072e-01 1.74514815e-01 -9.30402614e-03 -1.22078800e+00
4.97224659e-01 -1.41531003e+00 3.73695642e-01 1.77278399e+00
4.84609395e-01 -7.28170097e-01 8.15678000e-01 8.62806022e-01
-6.35801181e-02 -3.13898087e-01 -1.08158362e+00 -6.49265289e-01
-7.59773934e-03 -6.90467238e-01 4.93759453e-01 5.48862040e-01
6.64606750e-01 1.51266783e-01 -4.05179560e-01 -1.63526133e-01
3.65055561e-01 -8.98580849e-01 1.14944577e+00 -1.11285651e+00
-5.09414375e-01 -4.88646209e-01 -4.94756907e-01 -5.21511912e-01
3.52860212e-01 -7.23519146e-01 1.19282223e-01 -1.51651704e+00
1.81677476e-01 -8.73230770e-02 4.31259662e-01 4.72701550e-01
9.24542323e-02 -1.57888997e-02 2.43222713e-01 1.27338558e-01
-6.00241721e-02 2.76167989e-01 1.09554076e+00 -1.93103641e-01
-4.36646700e-01 7.68734291e-02 -7.40453005e-01 6.52514756e-01
8.23318183e-01 -1.54307544e-01 -6.25527918e-01 -3.49934936e-01
1.80919707e-01 1.12360984e-01 3.42510283e-01 -1.13465846e+00
4.16144222e-01 2.17888895e-02 -2.90294737e-01 9.98963416e-03
5.06556630e-01 -3.28112751e-01 -9.60606411e-02 1.88426822e-01
-2.77278751e-01 -2.99323887e-01 1.14117302e-01 9.86344442e-02
-1.88528020e-02 -2.46046022e-01 9.35680151e-01 -4.92827706e-02
-1.60460204e-01 -1.70313986e-03 -6.71818376e-01 3.78372341e-01
8.12443137e-01 -3.15377325e-01 -6.03887141e-01 -1.06847954e+00
-7.12584257e-01 -1.50560243e-02 7.13641644e-01 4.29862082e-01
5.89662790e-01 -7.21423149e-01 -8.13103855e-01 2.42384776e-01
1.23480611e-01 -1.73642233e-01 1.54737830e-01 8.32769811e-01
-7.05092847e-01 3.34876299e-01 -4.06325072e-01 -2.51925379e-01
-1.50013578e+00 1.62552670e-01 4.53335047e-01 8.39749277e-02
-7.05833375e-01 9.16284084e-01 -2.36314222e-01 -3.34779382e-01
6.69649422e-01 7.73986503e-02 -1.40223242e-02 6.21980242e-02
7.21450984e-01 6.22577786e-01 -3.74465324e-02 -5.13171315e-01
-1.44531950e-01 2.38186732e-01 9.47651789e-02 -7.44465947e-01
1.00372195e+00 -2.79154450e-01 1.19682081e-01 2.02238381e-01
7.90329516e-01 6.34510994e-01 -1.05633080e+00 2.79763937e-01
-4.48320597e-01 -3.89474988e-01 -6.81651235e-02 -8.96809876e-01
-6.09303296e-01 9.99461055e-01 7.29263425e-01 -1.26272187e-01
1.00505567e+00 1.62715346e-01 7.94808269e-01 1.38228759e-01
3.48961562e-01 -1.33540046e+00 2.06471428e-01 1.77627355e-01
9.89695549e-01 -1.31758702e+00 -5.56678832e-01 -3.10598850e-01
-5.74652016e-01 7.00195014e-01 4.68611836e-01 7.54134417e-01
5.66085815e-01 1.51622564e-01 3.87940556e-01 3.53511661e-01
-5.82636178e-01 -4.66801018e-01 -6.72942549e-02 1.13386846e+00
3.75342607e-01 -3.58026735e-02 -2.04799235e-01 5.23402333e-01
-7.07572281e-01 3.38024259e-01 9.03304458e-01 4.46630001e-01
-2.72846907e-01 -1.17803395e+00 -6.05088830e-01 1.69982031e-01
-5.44346213e-01 -2.80702978e-01 -3.78923714e-01 5.47171533e-01
1.01156920e-01 1.40774870e+00 -1.72398180e-01 -1.69002160e-01
2.24748939e-01 5.90474427e-01 3.21680188e-01 -5.91932416e-01
-3.07528108e-01 -1.71095192e-01 3.40019137e-01 -1.99969545e-01
1.62233502e-01 -8.46736670e-01 -8.90520811e-01 -5.06485820e-01
-1.40496552e-01 -4.95452255e-01 1.14175963e+00 7.95850694e-01
4.98142779e-01 2.50304580e-01 2.41124526e-01 -8.78112614e-01
-4.26049501e-01 -1.29740608e+00 -3.44586849e-01 3.34750742e-01
3.70802999e-01 -5.18418252e-01 -5.73348284e-01 1.13887295e-01] | [14.276481628417969, 4.927881717681885] |
32ce5dd8-59c7-4db3-9231-2b5041186bc6 | graphfc-customs-fraud-detection-with-label | 2305.11377 | null | https://arxiv.org/abs/2305.11377v1 | https://arxiv.org/pdf/2305.11377v1.pdf | GraphFC: Customs Fraud Detection with Label Scarcity | Custom officials across the world encounter huge volumes of transactions. With increased connectivity and globalization, the customs transactions continue to grow every year. Associated with customs transactions is the customs fraud - the intentional manipulation of goods declarations to avoid the taxes and duties. With limited manpower, the custom offices can only undertake manual inspection of a limited number of declarations. This necessitates the need for automating the customs fraud detection by machine learning (ML) techniques. Due the limited manual inspection for labeling the new-incoming declarations, the ML approach should have robust performance subject to the scarcity of labeled data. However, current approaches for customs fraud detection are not well suited and designed for this real-world setting. In this work, we propose $\textbf{GraphFC}$ ($\textbf{Graph}$ neural networks for $\textbf{C}$ustoms $\textbf{F}$raud), a model-agnostic, domain-specific, semi-supervised graph neural network based customs fraud detection algorithm that has strong semi-supervised and inductive capabilities. With upto 252% relative increase in recall over the present state-of-the-art, extensive experimentation on real customs data from customs administrations of three different countries demonstrate that GraphFC consistently outperforms various baselines and the present state-of-art by a large margin. | ['Shou-De Lin', 'Meeyoug Cha', 'Cheng-Te Li', 'Yu-Che Tsai', 'Karandeep Singh'] | 2023-05-19 | null | null | null | null | ['fraud-detection'] | ['miscellaneous'] | [-8.61492082e-02 -1.95723642e-02 -3.78527373e-01 -4.37163830e-01
-3.45302373e-01 -4.73925322e-01 3.03825289e-01 5.79296291e-01
-4.90286380e-01 7.24121451e-01 -3.32956284e-01 -8.67305279e-01
-2.12321430e-02 -1.10847485e+00 -6.93248391e-01 -1.36810482e-01
-1.28976479e-01 5.60368061e-01 -1.43911034e-01 -5.57508647e-01
2.33574793e-01 3.95038188e-01 -6.90023899e-01 3.55740696e-01
9.79051888e-01 1.15567482e+00 -3.92979592e-01 2.75436372e-01
9.64116976e-02 8.45090628e-01 -6.02110684e-01 -1.22613311e+00
6.76909685e-01 -1.26643687e-01 -5.84583044e-01 -2.12673038e-01
3.00785065e-01 -7.08669484e-01 -4.05491471e-01 1.34835672e+00
-1.12467088e-01 -2.22166300e-01 7.56979048e-01 -1.19136226e+00
-7.56083667e-01 5.96129537e-01 -9.60305870e-01 4.96190935e-01
2.86314756e-01 -1.48416951e-01 1.21040511e+00 -6.81274354e-01
5.68394423e-01 7.46379197e-01 1.25443339e+00 5.45933880e-02
-1.06686604e+00 -1.06991708e+00 1.09360449e-01 -1.97127864e-01
-1.25223637e+00 -1.04528613e-01 7.34770775e-01 -4.68531370e-01
1.43661129e+00 1.54444680e-01 5.46327412e-01 9.69369292e-01
2.41163731e-01 6.30801380e-01 9.21974719e-01 -4.71906751e-01
2.77317408e-02 1.12261966e-01 2.47897848e-01 1.29171610e+00
1.23999810e+00 -1.12358063e-01 -5.51510215e-01 -5.94682574e-01
9.03048933e-01 3.67047548e-01 2.50392824e-01 -1.17000699e-01
-6.25220835e-01 1.22158539e+00 3.26799452e-01 3.05535048e-01
-4.82058167e-01 1.09088831e-01 7.79503703e-01 6.54183388e-01
6.43014491e-01 4.48895186e-01 -6.25547171e-01 6.61547482e-02
-1.26580071e+00 3.10445666e-01 1.00721538e+00 9.86374915e-01
6.92280352e-01 3.45674872e-01 5.84867895e-01 7.16365695e-01
3.61140221e-01 2.86445379e-01 3.07859093e-01 -5.74519992e-01
1.23226988e+00 1.26055312e+00 1.93437949e-01 -1.55846488e+00
-6.49719357e-01 -3.28532100e-01 -1.24389184e+00 -6.16519526e-02
6.76396370e-01 -3.09046358e-01 -7.60050058e-01 1.18537891e+00
3.33948694e-02 -6.08993411e-01 -2.01984897e-01 7.07968473e-01
2.56466150e-01 1.31281018e-01 -2.51393253e-03 -1.66825846e-01
1.09422493e+00 -5.30792475e-01 -5.30777693e-01 -2.15801865e-01
7.90247917e-01 -4.38259602e-01 7.69849896e-01 4.59971368e-01
-8.03576171e-01 1.61098301e-01 -1.11424911e+00 2.92344451e-01
-7.66005695e-01 6.16133548e-02 1.30906725e+00 9.75889385e-01
-4.55862552e-01 7.71340609e-01 -8.27086151e-01 -2.44514495e-01
7.33566344e-01 4.66118574e-01 -4.65936840e-01 -2.92750299e-02
-1.28780437e+00 8.70987773e-01 5.06741524e-01 2.60414690e-01
-2.80809820e-01 -3.80830526e-01 -9.79917705e-01 -1.05932184e-01
4.13747102e-01 -3.09152901e-01 1.04275370e+00 -1.02353537e+00
-5.89847565e-01 1.00140560e+00 4.83587384e-01 -1.08621466e+00
7.82491267e-01 1.39240623e-01 -7.39911318e-01 5.16086482e-02
3.68809819e-01 -1.00804612e-01 7.76775301e-01 -8.98919523e-01
-7.30057538e-01 -4.38115835e-01 8.49884376e-02 -2.40155429e-01
-3.87362123e-01 1.65303454e-01 4.78053614e-02 -1.13511479e+00
3.37896377e-01 -6.38683498e-01 -1.31065831e-01 -5.13260126e-01
-5.45003772e-01 -1.51045442e-01 4.73854095e-01 -1.06115723e+00
1.51373124e+00 -1.78948236e+00 -6.05871558e-01 5.39545596e-01
2.33149558e-01 4.61962432e-01 4.40488398e-01 6.86737180e-01
-1.53451174e-01 3.26385170e-01 -5.85030973e-01 -1.64522439e-01
8.64258707e-02 2.55329728e-01 -1.27087504e-01 6.49991751e-01
1.68742627e-01 6.66738272e-01 -8.93341482e-01 -2.77412385e-01
-1.28180068e-02 -3.24417800e-01 -4.62764502e-01 -1.00287423e-01
-8.50641429e-02 -4.14534211e-01 -4.65440542e-01 1.57650113e+00
5.76885879e-01 -3.25010896e-01 6.65652156e-01 -3.53558660e-02
2.22929493e-01 2.05590516e-01 -1.18630934e+00 1.51262248e+00
1.92882121e-02 2.17676461e-01 6.06488734e-02 -1.57464564e+00
8.31729949e-01 1.21025704e-02 5.20403266e-01 -8.48312199e-01
6.16957918e-02 5.36309958e-01 -1.24772765e-01 -4.19259250e-01
6.59965515e-01 -2.55661637e-01 -5.94000340e-01 7.75597572e-01
1.78156886e-02 2.94254363e-01 4.07209903e-01 4.96405721e-01
1.40639973e+00 -5.96613921e-02 6.12905979e-01 -9.28723440e-02
1.09816484e-01 4.05631810e-01 6.74152195e-01 7.78345883e-01
-3.53012383e-01 1.60248965e-01 5.98586261e-01 -9.32525754e-01
-1.04461610e+00 -7.74767935e-01 1.94281772e-01 1.15221989e+00
-3.18162948e-01 -1.38016596e-01 -6.39594078e-01 -1.30301058e+00
5.61632752e-01 5.03796220e-01 -4.43881840e-01 8.51325616e-02
-6.31963670e-01 -1.23682761e+00 1.01275194e+00 4.47543144e-01
6.54218674e-01 -1.05559170e+00 -5.47174692e-01 4.73682910e-01
-1.55237705e-01 -9.42590475e-01 -2.91841686e-01 3.27089041e-01
-8.40388060e-01 -1.51185656e+00 -3.21447194e-01 -5.80347538e-01
9.56860721e-01 -5.73643073e-02 1.05273628e+00 2.13760510e-01
-3.85617346e-01 -1.43088445e-01 -8.74290690e-02 -4.96640205e-01
-4.75980759e-01 9.19197127e-02 3.91447872e-01 -6.32769391e-02
7.01242507e-01 -6.68753088e-01 -4.01753932e-01 -2.05352217e-01
-8.37970495e-01 -2.12145463e-01 5.42061329e-01 8.84708166e-01
1.87346131e-01 3.08269501e-01 9.02354479e-01 -1.29335499e+00
8.62984061e-01 -6.84726477e-01 -8.35031569e-01 2.02028841e-01
-1.22310758e+00 -2.78399169e-01 8.31397951e-01 2.24141013e-02
-9.17250991e-01 -3.04013401e-01 3.40336233e-01 -2.46867269e-01
-2.47679856e-02 8.99191618e-01 3.72187346e-01 2.59234849e-03
6.85913086e-01 -1.25043973e-01 -2.25892156e-01 -5.89466035e-01
4.61046696e-02 8.64850819e-01 5.63816607e-01 -2.82710284e-01
8.81779075e-01 4.63544041e-01 -8.10183585e-02 -4.12540555e-01
-5.08428812e-01 -6.06383085e-01 -5.73741913e-01 2.02286919e-03
2.21063703e-01 -9.99370813e-01 -8.44914556e-01 7.00082064e-01
-1.07043505e+00 -5.78710586e-02 7.72357732e-02 1.83000743e-01
-1.48528218e-01 6.38021648e-01 -9.96630073e-01 -1.13881755e+00
-6.27502024e-01 -5.15504301e-01 4.52022910e-01 -1.22094542e-01
-3.79463106e-01 -1.00252330e+00 -1.72037736e-01 7.15445936e-01
1.39739007e-01 7.77689755e-01 9.43036973e-01 -1.06513691e+00
-2.52752244e-01 -7.29814947e-01 -6.57413363e-01 7.11189568e-01
3.49126250e-01 -3.29488188e-01 -3.46189380e-01 -4.12397265e-01
-1.25192970e-01 -4.10620391e-01 8.93363297e-01 -9.12995264e-02
6.52267694e-01 -9.98683810e-01 -1.30307540e-01 3.80792797e-01
1.79873455e+00 3.14682633e-01 2.67984957e-01 6.36643589e-01
6.76771343e-01 4.33434516e-01 4.85012263e-01 7.25175083e-01
5.21733761e-01 6.52959570e-02 4.14640009e-01 -1.00116991e-01
4.63207394e-01 -4.12738740e-01 2.32235759e-01 5.94815910e-01
-3.23965639e-01 -8.92940685e-02 -1.20183110e+00 7.77386308e-01
-1.95232499e+00 -9.47195768e-01 -2.51426250e-01 1.96350181e+00
7.93400049e-01 3.79636824e-01 2.05948800e-01 2.54205585e-01
6.72747552e-01 2.09528934e-02 -5.44274867e-01 -7.87247539e-01
3.36744264e-02 6.48362264e-02 1.13740146e+00 -8.12661424e-02
-1.20532382e+00 8.38195145e-01 6.19146252e+00 6.22962534e-01
-6.59333169e-01 6.12176843e-02 5.98208010e-01 -9.89669114e-02
9.06544030e-02 -1.97408959e-01 -5.12382269e-01 6.58125341e-01
8.09367418e-01 1.05442323e-01 6.31318331e-01 8.07094395e-01
9.26471949e-02 -1.88698515e-01 -7.33035743e-01 8.44067752e-01
1.10359140e-01 -1.61197186e+00 1.58785790e-01 2.45715946e-01
8.06179166e-01 2.42789686e-01 -3.06134850e-01 2.96626598e-01
6.31463170e-01 -7.97560692e-01 7.13426828e-01 2.86490917e-01
9.13896620e-01 -7.99033642e-01 9.14746881e-01 3.53072584e-01
-9.18132901e-01 -3.26907754e-01 -1.29597366e-01 -4.09875780e-01
-4.06568050e-02 7.84693301e-01 -8.88458252e-01 9.70476627e-01
8.36209834e-01 4.80666846e-01 -3.19102317e-01 3.22870970e-01
6.42365292e-02 6.10382795e-01 -3.27800751e-01 1.02460884e-01
4.95698124e-01 -4.92547065e-01 1.55359253e-01 1.62166429e+00
1.06021566e-02 9.92342085e-02 1.83999270e-01 7.08730578e-01
-6.55681729e-01 -1.32784992e-02 -8.17699909e-01 -3.36999416e-01
2.53041804e-01 9.52128470e-01 -7.31225431e-01 -4.33814704e-01
-6.28399432e-01 8.69355500e-01 5.49682081e-01 2.62963474e-01
-6.61153018e-01 -5.88509619e-01 3.61318618e-01 4.44876164e-01
1.83438390e-01 -2.82787174e-01 -3.85274142e-01 -1.24962997e+00
3.58255506e-01 -9.88322854e-01 9.73766029e-01 -1.87290579e-01
-1.52977872e+00 3.92289236e-02 -1.42458677e-01 -9.41601932e-01
-2.96594441e-01 -6.08341336e-01 -2.94426441e-01 5.58688164e-01
-1.58624017e+00 -1.26907468e+00 1.48242787e-01 6.88880622e-01
1.63782775e-01 -6.55359328e-01 7.58353949e-01 4.33410913e-01
-5.68578839e-01 6.58607304e-01 2.64091998e-01 8.31337988e-01
4.37539488e-01 -1.17929721e+00 6.46708608e-01 1.01901436e+00
7.97398854e-03 8.17851961e-01 3.47151756e-01 -1.11902177e+00
-1.25863039e+00 -1.21986520e+00 1.14515066e+00 -2.19551921e-01
1.08321095e+00 -4.05541986e-01 -7.18318522e-01 1.07467818e+00
-1.16799302e-01 -1.54284641e-01 6.83815181e-01 1.74045473e-01
-7.39524007e-01 -1.35661483e-01 -1.74468923e+00 1.59316346e-01
1.05564964e+00 -6.00901902e-01 -7.68206477e-01 6.37371838e-01
1.52907863e-01 -2.91077375e-01 -9.72483754e-01 1.97810099e-01
6.54177189e-01 -1.13390660e+00 7.41534770e-01 -9.04116094e-01
2.08578080e-01 4.19610627e-02 -2.48074755e-01 -7.23347008e-01
-1.45688325e-01 -5.43275297e-01 -3.22891593e-01 1.10941803e+00
4.62979198e-01 -9.69425201e-01 8.95742178e-01 7.18289137e-01
1.60396636e-01 -5.46980381e-01 -1.15889692e+00 -9.49341893e-01
-2.39163384e-01 -4.85634774e-01 4.67924178e-01 1.73023331e+00
3.61596078e-01 -1.30631536e-01 -4.83743817e-01 7.17620701e-02
8.36419165e-01 1.30219042e-01 5.13418913e-01 -1.22402585e+00
-8.81910920e-02 -2.82094806e-01 -4.03162956e-01 -2.38701075e-01
-5.94198704e-02 -8.98947477e-01 -6.17294967e-01 -1.56733370e+00
1.25655249e-01 -6.19363844e-01 -4.13391709e-01 9.57631707e-01
3.87397289e-01 3.32906753e-01 5.69351530e-03 3.35895717e-01
-5.09101689e-01 -1.50314778e-01 5.08687615e-01 -3.30586582e-01
-7.21679479e-02 -9.37267095e-02 -8.93273890e-01 1.09559906e+00
8.15452099e-01 -6.76556408e-01 1.32251546e-01 -3.73661578e-01
5.70457518e-01 -1.57749895e-02 3.33048314e-01 -5.60785174e-01
2.47287735e-01 -3.08417559e-01 4.30874169e-01 -5.23227692e-01
-1.98783427e-01 -7.04243362e-01 5.39549403e-02 7.38129199e-01
1.78044870e-01 4.78711873e-01 -5.90878651e-02 7.40606666e-01
-4.64801155e-02 -1.61286324e-01 5.42425156e-01 -5.61555326e-01
-3.70721102e-01 2.34931499e-01 -4.13054615e-01 -3.01694535e-02
7.44026721e-01 -2.55544871e-01 -2.94775814e-01 -7.13348538e-02
-3.50391179e-01 -5.33389598e-02 4.58942682e-01 3.17770123e-01
3.60988289e-01 -1.18063402e+00 -5.93494713e-01 2.37137899e-01
7.01572374e-02 3.40442471e-02 -1.45999968e-01 6.36495709e-01
-8.05667222e-01 4.99883503e-01 -8.39173570e-02 4.76167761e-02
-7.21960545e-01 4.77507472e-01 3.57775360e-01 -7.97591269e-01
-4.18448538e-01 5.97870946e-01 -4.82250988e-01 -4.52212125e-01
5.81009500e-02 -6.50002897e-01 1.77231416e-01 2.04517737e-01
1.20742217e-01 6.55988514e-01 3.13904017e-01 -4.60366696e-01
-5.08029163e-01 -2.21209675e-01 -3.38930309e-01 3.84613872e-01
1.57676268e+00 1.62814602e-01 -2.75684923e-01 7.57632479e-02
1.01657617e+00 -1.58464521e-01 -6.79050803e-01 -6.07833266e-02
3.61463487e-01 -5.10978818e-01 -2.73367256e-01 -1.13473833e+00
-1.16898751e+00 1.98092803e-01 1.05359927e-01 6.33686364e-01
8.80616307e-01 -5.28829217e-01 8.67452621e-01 7.63950467e-01
7.67477036e-01 -1.61395884e+00 -2.19162330e-01 1.85508013e-01
5.82123041e-01 -1.27411211e+00 3.65975678e-01 -2.24759773e-04
-5.64557433e-01 9.12526429e-01 3.20988297e-01 -4.18107897e-01
3.86928052e-01 1.88076049e-02 -1.54993832e-01 -5.50161421e-01
-1.55730769e-01 2.42724478e-01 -2.53181249e-01 5.59069455e-01
1.16451770e-01 3.28551888e-01 -4.65201706e-01 8.54393363e-01
-3.06399949e-02 1.42549425e-02 6.66977346e-01 1.00033712e+00
-1.82478979e-01 -9.56293106e-01 -3.17227036e-01 1.18181968e+00
-8.46096456e-01 -3.29221666e-01 -5.04934847e-01 1.16189730e+00
3.44889492e-01 9.52162147e-01 -1.37292877e-01 -6.73118532e-02
5.79780400e-01 2.08117679e-01 1.62903890e-01 -3.40738714e-01
-1.06912005e+00 -3.63862783e-01 4.38309848e-01 -5.27054429e-01
-3.67009968e-01 -9.62324619e-01 -1.38876271e+00 -7.81883121e-01
-4.26481605e-01 -8.00222605e-02 6.30253971e-01 7.46096611e-01
1.08131066e-01 1.80294309e-02 5.81797957e-01 -1.69481650e-01
-6.49638653e-01 -9.31973219e-01 -1.00393558e+00 7.29742050e-01
1.89470351e-01 -5.04681230e-01 -3.51156056e-01 -6.03045151e-02] | [7.306859970092773, 5.828357219696045] |
3737d1d7-e232-488f-a4c2-202d525bbe9f | effective-conditioned-and-composed-image | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Baldrati_Effective_Conditioned_and_Composed_Image_Retrieval_Combining_CLIP-Based_Features_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Baldrati_Effective_Conditioned_and_Composed_Image_Retrieval_Combining_CLIP-Based_Features_CVPR_2022_paper.pdf | Effective Conditioned and Composed Image Retrieval Combining CLIP-Based Features | Conditioned and composed image retrieval extend CBIR systems by combining a query image with an additional text that expresses the intent of the user, describing additional requests w.r.t. the visual content of the query image. This type of search is interesting for e-commerce applications, e.g. to develop interactive multimodal searches and chatbots. In this demo, we present an interactive system based on a combiner network, trained using contrastive learning, that combines visual and textual features obtained from the OpenAI CLIP network to address conditioned CBIR. The system can be used to improve e-shop search engines. For example, considering the fashion domain it lets users search for dresses, shirts and toptees using a candidate start image and expressing some visual differences w.r.t. its visual content, e.g. asking to change color, pattern or shape. The proposed network obtains state-of-the-art performance on the FashionIQ dataset and on the more recent CIRR dataset, showing its applicability to the fashion domain for conditioned retrieval, and to more generic content considering the more general task of composed image retrieval. | ['Alberto del Bimbo', 'Tiberio Uricchio', 'Marco Bertini', 'Alberto Baldrati'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['composed-image-retrieval'] | ['computer-vision'] | [ 2.52593547e-01 -4.19871449e-01 -3.31557781e-01 -3.64250153e-01
-9.02395010e-01 -9.90887284e-01 6.84921741e-01 1.51507765e-01
-4.99068767e-01 6.28050268e-02 -5.09264283e-02 -6.09071255e-02
-2.24482387e-01 -4.32838023e-01 -5.01652658e-01 -6.54951632e-01
1.88865528e-01 9.20873642e-01 1.55934647e-01 -9.18264091e-01
2.66440958e-01 7.01706529e-01 -1.79623556e+00 1.06833053e+00
3.30236256e-01 1.49230957e+00 5.85206807e-01 6.35765493e-01
-2.22602278e-01 7.81407952e-01 -5.02705812e-01 -4.61069703e-01
1.10598572e-01 -2.12388143e-01 -9.31610584e-01 3.10778707e-01
8.33835363e-01 -3.32225174e-01 -2.08845273e-01 8.22196841e-01
3.81401300e-01 3.97334486e-01 5.85762143e-01 -1.21543097e+00
-1.10425115e+00 3.83001894e-01 -2.40762800e-01 1.36596605e-01
1.00175834e+00 2.80022770e-02 1.30259013e+00 -8.05520475e-01
1.15249956e+00 1.38037586e+00 -5.35021648e-02 4.16343033e-01
-1.41065753e+00 -4.18932050e-01 6.61003888e-02 5.95762014e-01
-1.36258388e+00 -1.50287703e-01 1.06338084e+00 -1.47186086e-01
5.72549820e-01 8.82304847e-01 5.99876463e-01 1.04833221e+00
-1.62745625e-01 1.37575769e+00 1.02427161e+00 -6.49893045e-01
2.04880554e-02 6.35473430e-01 2.45489851e-02 5.19248188e-01
-5.98445892e-01 -1.55984730e-01 -3.66901845e-01 5.60680106e-02
6.17648959e-01 1.45353585e-01 4.61426144e-03 -5.20293057e-01
-1.09540284e+00 1.01314759e+00 7.93993533e-01 7.78210759e-01
-1.01885639e-01 -2.00043935e-02 3.48134905e-01 6.76940799e-01
4.42142129e-01 6.84799373e-01 -3.24235409e-01 3.58539313e-01
-8.02020669e-01 3.98602307e-01 8.82665038e-01 1.14878678e+00
6.65363193e-01 -5.48432112e-01 -5.18337727e-01 1.34570014e+00
1.98267698e-01 5.24446309e-01 1.64956510e-01 -8.50480318e-01
2.48702615e-01 5.01453936e-01 6.51998892e-02 -1.13533986e+00
-3.59120965e-01 4.35767807e-02 -6.80847228e-01 -1.77242905e-02
3.56849372e-01 4.34373140e-01 -7.42208004e-01 1.19694328e+00
8.28714743e-02 -6.65318191e-01 -2.00770795e-01 1.37602270e+00
1.30094421e+00 8.89611959e-01 5.82510158e-02 1.00516930e-01
1.89158118e+00 -1.09134769e+00 -8.20285380e-01 -1.93674311e-01
2.23435581e-01 -1.27614403e+00 1.19406903e+00 5.61383009e-01
-1.06655693e+00 -8.83888543e-01 -8.12175810e-01 -4.27412391e-01
-9.26805973e-01 6.32949173e-01 2.40490571e-01 3.28938752e-01
-1.14423501e+00 3.10580373e-01 1.39272109e-01 -7.38528430e-01
-1.81403279e-01 4.68564093e-01 -3.91625464e-01 -5.21353602e-01
-1.20166588e+00 8.35276306e-01 2.73490041e-01 2.07527757e-01
-5.04535675e-01 -3.21112305e-01 -8.53331029e-01 1.04114912e-01
7.91424811e-01 -4.25968260e-01 1.14075470e+00 -1.28635550e+00
-1.42987275e+00 1.22903585e+00 2.73775548e-01 -5.11158220e-02
4.31303412e-01 -4.82824035e-02 -6.96323633e-01 5.64054549e-01
1.51620790e-01 1.00063336e+00 9.88454878e-01 -1.53080034e+00
-5.07281125e-01 -5.10385454e-01 4.40172553e-01 3.12166095e-01
-2.19315737e-01 4.64664161e-01 -1.19799221e+00 -6.26618564e-01
-1.62969545e-01 -1.28800404e+00 -8.61100107e-02 2.76072383e-01
-3.76961887e-01 -4.87951726e-01 1.05720699e+00 -7.34922469e-01
7.70455658e-01 -2.23669791e+00 3.43658835e-01 5.83558559e-01
-2.99294814e-02 1.55194953e-01 -6.20660722e-01 5.76732039e-01
1.46011531e-01 1.52009428e-02 4.26441729e-01 -3.46390307e-02
1.62125558e-01 -3.47339250e-02 9.39421505e-02 1.60806239e-01
4.36362475e-02 9.69598353e-01 -7.44468391e-01 -7.19659626e-01
5.33598244e-01 4.06073570e-01 -2.22455263e-01 2.01474667e-01
-4.77166027e-01 2.47184068e-01 -4.78700817e-01 5.58624268e-01
5.97776115e-01 -2.84109145e-01 2.51643747e-01 -7.43924260e-01
-7.85097331e-02 -3.83565187e-01 -1.15250540e+00 1.93052125e+00
-6.37238860e-01 8.19177568e-01 2.16034576e-01 -8.03424478e-01
1.20110250e+00 1.36853144e-01 6.17296755e-01 -1.17792404e+00
1.80277810e-01 1.00039385e-01 -3.96229357e-01 -6.83797836e-01
9.14856911e-01 2.89276004e-01 -4.63254333e-01 3.71340513e-01
1.50479808e-01 -3.75921041e-01 5.18144906e-01 2.03406349e-01
5.73878706e-01 9.83255953e-02 2.54056193e-02 -1.63006470e-01
8.16856027e-01 2.08208963e-01 -5.14613032e-01 1.00357926e+00
1.88756973e-01 4.66653913e-01 1.17812999e-01 -6.10331655e-01
-1.06955361e+00 -9.09771979e-01 4.89228405e-02 1.68007517e+00
5.97235441e-01 -1.34873077e-01 -1.29247412e-01 -4.27656949e-01
-1.29168510e-01 4.59161282e-01 -5.34468532e-01 -9.02042016e-02
-4.94629622e-01 4.63087223e-02 2.03704014e-01 6.95574656e-02
5.46395242e-01 -1.42779422e+00 -4.21011984e-01 -3.04421433e-03
-5.17451525e-01 -1.10412753e+00 -7.06767321e-01 8.83468613e-02
-3.44024211e-01 -6.57781243e-01 -1.15897012e+00 -1.45857251e+00
3.40332121e-01 3.77640039e-01 1.23595369e+00 1.25293881e-01
-8.61341476e-01 8.63376260e-01 -7.52298057e-01 1.20971836e-01
-5.17291307e-01 1.79968745e-01 -5.47932625e-01 2.07436368e-01
2.17361450e-01 3.50293845e-01 -6.72099888e-01 7.10842788e-01
-1.18223250e+00 3.02211121e-02 4.82183009e-01 7.45503783e-01
4.15085912e-01 -1.99587166e-01 1.55876711e-01 -6.01665437e-01
8.77053916e-01 -2.81848371e-01 -4.82010424e-01 8.51290584e-01
-2.07964361e-01 1.76826213e-02 3.22007656e-01 -7.70602107e-01
-1.14166415e+00 2.26082012e-01 -9.86235738e-02 -3.37177902e-01
-4.53413188e-01 3.79247934e-01 5.71967699e-02 -1.08627617e-01
5.71473718e-01 1.71548411e-01 -1.14489503e-01 -6.50372565e-01
8.02621543e-01 6.69759870e-01 5.63494444e-01 -2.85959721e-01
4.22910303e-01 2.72035986e-01 -1.91052988e-01 -8.74954998e-01
-4.27908510e-01 -1.15082526e+00 -6.11418486e-01 -6.92347407e-01
1.06290185e+00 -5.75049162e-01 -1.14724886e+00 -2.52204835e-02
-1.27272499e+00 5.08886576e-02 -8.56905431e-02 1.31829336e-01
-6.48857296e-01 2.05536678e-01 -8.31782758e-01 -6.35116696e-01
-3.55228812e-01 -1.20537090e+00 1.40313447e+00 5.31971976e-02
5.80005087e-02 -7.24365175e-01 -1.91081688e-01 7.94698417e-01
3.56274843e-01 -2.34922290e-01 1.00574315e+00 -8.21137130e-01
-5.79791903e-01 -5.17784417e-01 -4.32941228e-01 1.18935809e-01
4.85556684e-02 -2.51279861e-01 -6.99061334e-01 -1.79507807e-01
-5.34178674e-01 -6.88858688e-01 8.33199799e-01 7.51113892e-02
1.21861482e+00 -2.94149965e-01 -1.74661845e-01 3.50204855e-02
1.51401043e+00 5.25853157e-01 6.45432651e-01 2.85421520e-01
3.00680727e-01 8.12858999e-01 7.66020417e-01 2.78435051e-01
4.63171117e-02 1.46388078e+00 2.97155499e-01 -5.72119713e-01
-1.66977361e-01 -1.39349662e-02 8.62576663e-02 2.20550030e-01
-3.34525593e-02 -2.87873834e-01 -5.33672869e-01 2.20629752e-01
-1.67456663e+00 -1.03096962e+00 -8.98237824e-02 1.73800302e+00
6.48109138e-01 -4.34188783e-01 2.79871851e-01 -2.90231675e-01
6.80450261e-01 1.35979131e-01 -4.50001806e-01 -6.01571560e-01
-1.21129833e-01 1.53106466e-01 1.30525827e-01 2.15642735e-01
-1.09013796e+00 1.06242824e+00 5.46763515e+00 1.29086363e+00
-1.00981116e+00 -3.53311375e-02 7.40196645e-01 2.00563788e-01
-2.50584483e-01 -3.80703777e-01 -5.96257150e-01 -4.81867418e-02
1.82688415e-01 2.85738826e-01 9.07588422e-01 8.44491184e-01
-2.88776169e-03 -2.88856298e-01 -1.27901864e+00 1.23964214e+00
6.49132848e-01 -1.28401601e+00 1.54772773e-01 -1.78996071e-01
4.02239919e-01 -3.54363501e-01 4.23666060e-01 4.11924392e-01
-2.23152675e-02 -6.93543315e-01 7.07334876e-01 4.95137185e-01
8.06535721e-01 -4.23546076e-01 7.32243121e-01 9.89881456e-02
-1.25735354e+00 -7.86476508e-02 -1.21467985e-01 4.77991194e-01
7.45672919e-03 -3.50312740e-01 -7.73978949e-01 4.69549060e-01
8.64477456e-01 4.49536651e-01 -9.80464578e-01 9.65157986e-01
1.41870156e-01 -1.93799913e-01 -4.39706415e-01 -5.96951187e-01
4.12949055e-01 -2.82976419e-01 3.68724376e-01 1.39344704e+00
1.33535862e-01 -2.11104199e-01 4.16909218e-01 9.44955647e-01
-6.59492612e-02 6.97310209e-01 -6.30893350e-01 1.19658291e-01
-1.48881942e-01 1.53238583e+00 -9.10146415e-01 -4.08495843e-01
-2.46177182e-01 1.46420336e+00 -3.31553705e-02 5.01176476e-01
-6.40009642e-01 -4.35808480e-01 -1.54417679e-02 -6.90908134e-02
6.08502209e-01 2.47691751e-01 3.85607034e-01 -9.84473646e-01
-1.09889105e-01 -1.07440436e+00 6.11233830e-01 -1.35099673e+00
-1.37940383e+00 8.50712419e-01 2.57074773e-01 -1.33847058e+00
-3.55271041e-01 -7.23211169e-01 -1.70121893e-01 4.87563044e-01
-1.22246003e+00 -1.48218119e+00 -3.21165234e-01 1.05464065e+00
1.01003599e+00 -1.14355259e-01 6.92046046e-01 4.65263933e-01
2.71412358e-02 4.48707938e-01 7.04882815e-02 5.84434122e-02
8.56582701e-01 -9.83592689e-01 -2.35550061e-01 -6.40678182e-02
7.66819894e-01 2.95260608e-01 8.17139983e-01 -3.26222748e-01
-1.65717232e+00 -5.96227050e-01 7.88157701e-01 -1.00786395e-01
4.97049212e-01 -5.29744923e-01 -4.06088442e-01 2.76026607e-01
6.35249078e-01 -1.72716007e-01 5.12716770e-01 2.13661551e-01
-3.64393175e-01 -4.50003594e-01 -1.12319100e+00 6.05441689e-01
4.95425314e-01 -5.53414643e-01 -3.78074110e-01 6.49882376e-01
5.28310359e-01 -1.71548665e-01 -9.07436311e-01 4.01267894e-02
9.39564705e-01 -6.10769629e-01 1.39871764e+00 -5.82936049e-01
3.90367508e-01 1.74788311e-02 -2.68971294e-01 -1.07090855e+00
-3.61990869e-01 -1.55910030e-01 6.76686704e-01 9.68449414e-01
5.83531260e-01 2.88950264e-01 4.97635037e-01 5.72498262e-01
2.22493321e-01 -1.29118651e-01 -8.30446422e-01 -5.53444028e-01
-4.03377473e-01 -4.81902868e-01 3.12385648e-01 6.67644203e-01
1.85869277e-01 5.18618464e-01 -8.77800941e-01 -3.27501625e-01
2.54887044e-01 4.71913278e-01 6.76177263e-01 -1.14834452e+00
-2.55595237e-01 -4.46285069e-01 -2.96963960e-01 -1.32403660e+00
-3.03186364e-02 -1.00220978e+00 2.31311083e-01 -1.46610510e+00
3.01360369e-01 -3.45726877e-01 -2.55187631e-01 2.43224218e-01
3.39271516e-01 8.16855907e-01 1.02936637e+00 1.65313661e-01
-1.24934447e+00 8.77120718e-02 1.41305971e+00 -7.84656167e-01
-4.33953941e-01 -3.84463579e-03 -3.06580245e-01 2.92965710e-01
6.94578767e-01 -9.16300416e-02 -1.88758090e-01 -7.86141753e-02
3.91494870e-01 4.27748054e-01 6.88117802e-01 -5.67263246e-01
2.44181395e-01 5.39570935e-02 4.72830147e-01 -8.43496621e-01
8.47281933e-01 -1.40818000e+00 8.77587348e-02 3.03134441e-01
-9.16736782e-01 3.02806273e-02 4.67938520e-02 5.34495890e-01
-3.34674895e-01 -4.23890382e-01 3.14177752e-01 -4.30723786e-01
-9.79962587e-01 -2.47094929e-02 -6.99043393e-01 -4.17987972e-01
7.13408470e-01 -1.12113774e-01 -2.37528205e-01 -8.62865925e-01
-1.25266278e+00 2.46396005e-01 4.90472652e-02 9.38765109e-01
9.22766268e-01 -1.59930396e+00 -4.10018623e-01 1.39756799e-01
7.22302496e-01 -7.25307167e-01 4.22185034e-01 5.29287398e-01
-4.21520501e-01 6.40217841e-01 -2.87146807e-01 -7.96447635e-01
-1.72948098e+00 1.01387072e+00 -6.64934292e-02 -4.04810220e-01
-3.55722070e-01 4.48609263e-01 1.03576444e-01 -2.52321571e-01
3.95298630e-01 -3.15639645e-01 -5.71272850e-01 3.64714116e-01
4.32514906e-01 1.28181845e-01 -3.19920182e-02 -7.97985673e-01
-1.52443781e-01 6.37589633e-01 -1.61334008e-01 -2.35896587e-01
1.12821627e+00 -3.66062880e-01 -1.32809028e-01 5.26416242e-01
1.58708251e+00 -4.23496634e-01 -4.34880704e-01 -4.54967558e-01
-8.09424836e-03 -4.62082237e-01 3.26289311e-02 -1.26902688e+00
-1.21271718e+00 6.02396548e-01 1.06179917e+00 5.15324295e-01
1.28319025e+00 7.41906345e-01 3.41903239e-01 8.44129741e-01
3.43174905e-01 -1.33491993e+00 5.80087245e-01 1.04047075e-01
1.49457097e+00 -1.42206395e+00 -4.14485112e-02 -3.12793314e-01
-9.73801374e-01 1.45667470e+00 4.23851907e-01 2.40746625e-02
5.76587856e-01 -2.85296857e-01 1.27343431e-01 -4.79000807e-01
-5.79386711e-01 -5.35755634e-01 9.26789284e-01 4.71978307e-01
2.48757213e-01 5.35197183e-02 -2.77939975e-01 8.78585353e-02
2.79974312e-01 -3.49614955e-02 -1.00437775e-01 6.20323479e-01
-2.86173731e-01 -1.14330685e+00 -6.01404130e-01 3.72825056e-01
-9.15896595e-02 -4.74808328e-02 -7.19070971e-01 8.98315668e-01
1.30867973e-01 1.26034033e+00 1.32788271e-01 -2.24524528e-01
4.85320300e-01 -9.66417715e-02 5.28524458e-01 -2.20861375e-01
-9.89332199e-01 6.89145803e-01 2.84684032e-01 -5.45490444e-01
-7.34262228e-01 -5.60958982e-01 -3.76992017e-01 2.04757117e-02
-4.34243143e-01 -5.26325665e-02 8.65533948e-01 6.52246714e-01
-1.08377680e-01 1.19185813e-01 5.61878145e-01 -8.59081805e-01
-1.96383432e-01 -9.30505455e-01 -7.53606498e-01 9.53885734e-01
2.80838400e-01 -3.59693855e-01 1.49171755e-01 2.76544303e-01] | [10.82569694519043, 1.1655329465866089] |
af06f050-2426-4d54-b52e-c5cd8a2d734e | colt5-faster-long-range-transformers-with | 2303.09752 | null | https://arxiv.org/abs/2303.09752v2 | https://arxiv.org/pdf/2303.09752v2.pdf | CoLT5: Faster Long-Range Transformers with Conditional Computation | Many natural language processing tasks benefit from long inputs, but processing long documents with Transformers is expensive -- not only due to quadratic attention complexity but also from applying feedforward and projection layers to every token. However, not all tokens are equally important, especially for longer documents. We propose CoLT5, a long-input Transformer model that builds on this intuition by employing conditional computation, devoting more resources to important tokens in both feedforward and attention layers. We show that CoLT5 achieves stronger performance than LongT5 with much faster training and inference, achieving SOTA on the long-input SCROLLS benchmark. Moreover, CoLT5 can effectively and tractably make use of extremely long inputs, showing strong gains up to 64k input length. | ['Sumit Sanghai', 'Yun-Hsuan Sung', 'Yi Tay', 'James Lee-Thorp', 'Mandy Guo', 'David Uthus', 'Yury Zemlyanskiy', 'Siddhartha Brahma', 'Santiago Ontañón', 'Michiel de Jong', 'Tao Lei', 'Joshua Ainslie'] | 2023-03-17 | null | null | null | null | ['long-range-modeling'] | ['natural-language-processing'] | [ 9.36911926e-02 2.38107994e-01 -7.61854574e-02 -3.57129693e-01
-9.25258994e-01 -7.16879129e-01 6.39262497e-01 4.21539634e-01
-7.28035510e-01 6.66987062e-01 5.34844339e-01 -8.71917725e-01
1.23198368e-01 -1.03338766e+00 -9.47620928e-01 -2.52253681e-01
-2.85729412e-02 8.58776748e-01 1.10114798e-01 -8.07525143e-02
1.95014223e-01 1.35121435e-01 -1.02251947e+00 7.50455379e-01
7.40264177e-01 8.07326078e-01 1.88634053e-01 9.32836950e-01
-5.14357388e-01 8.57190847e-01 -2.95473248e-01 -8.25988293e-01
2.60669380e-01 3.03743035e-01 -1.28580201e+00 -5.41107535e-01
5.61546683e-01 -4.81252015e-01 -7.97165781e-02 7.05173075e-01
3.37460190e-01 2.60146074e-02 5.39136052e-01 -7.61208057e-01
-1.03957868e+00 1.56923461e+00 -5.25612354e-01 2.94725388e-01
1.10065565e-01 3.37643087e-01 1.84966600e+00 -1.42746735e+00
2.72545427e-01 1.48312545e+00 7.23117530e-01 2.52437174e-01
-1.26394844e+00 -4.69302446e-01 5.91273367e-01 -3.07005597e-03
-8.70078564e-01 -3.14432353e-01 8.78021345e-02 -6.69337809e-02
1.70772302e+00 1.30040973e-01 5.51093757e-01 8.19850266e-01
3.22302163e-01 1.38429594e+00 8.68256748e-01 -4.55064088e-01
-3.01731676e-01 -1.23545587e-01 7.15188563e-01 6.95690095e-01
2.27882892e-01 -4.07420367e-01 -5.07119060e-01 1.12885922e-01
4.76827353e-01 2.71152914e-01 1.86014883e-02 5.60080469e-01
-1.32388020e+00 5.56616247e-01 4.70019877e-01 2.84201115e-01
-3.89337480e-01 6.86074197e-01 5.49299955e-01 6.42079234e-01
4.17392820e-01 6.61296427e-01 -9.50348735e-01 -5.02770126e-01
-7.25732744e-01 3.30824442e-02 6.33308172e-01 1.10185671e+00
6.39383614e-01 -8.08010995e-02 -6.63844824e-01 6.56108141e-01
-5.61914258e-02 7.94871569e-01 3.75893593e-01 -6.69129610e-01
9.09883797e-01 5.44191957e-01 -1.09653331e-01 -4.11543995e-01
-1.83826223e-01 -4.50930864e-01 -8.73226166e-01 -1.37129948e-01
6.28892124e-01 -2.39636660e-01 -1.07398558e+00 1.65317786e+00
-3.80175859e-01 -4.42799687e-01 -2.27601063e-02 3.99677694e-01
3.58958602e-01 9.01056767e-01 2.26283848e-01 2.83554584e-01
1.52785301e+00 -1.10726869e+00 -4.28148746e-01 -6.57591939e-01
7.21919179e-01 -7.09072769e-01 1.75382686e+00 6.01076066e-01
-1.56707394e+00 -3.03538293e-01 -6.86704993e-01 -9.05664027e-01
-5.96641064e-01 1.20490722e-01 1.01995015e+00 3.19519401e-01
-1.18314672e+00 5.91116011e-01 -8.80903840e-01 -3.10055465e-02
6.18743002e-01 4.37196702e-01 -1.84850335e-01 -4.14013118e-01
-1.29110038e+00 1.00104928e+00 1.96359381e-01 2.47956529e-01
-6.91510201e-01 -1.07153893e+00 -7.05177724e-01 8.22792292e-01
2.48959646e-01 -6.97187722e-01 1.63593626e+00 -5.05141079e-01
-1.38825107e+00 3.89044404e-01 -5.63659906e-01 -7.66368032e-01
4.93017375e-01 -8.49312484e-01 2.79644758e-01 -1.72032699e-01
-6.93065897e-02 7.29780257e-01 5.42987645e-01 -4.67166513e-01
-4.11510050e-01 -1.66211024e-01 4.05971676e-01 1.11875318e-01
-6.85363889e-01 7.35855997e-02 -5.96863091e-01 -5.34304023e-01
-2.52312779e-01 -6.99895740e-01 -2.30219126e-01 -6.25683218e-02
-5.64410090e-01 -5.93718052e-01 3.15446794e-01 -5.01151860e-01
1.08850741e+00 -1.84263480e+00 -1.11780368e-01 1.08670190e-01
4.62721258e-01 2.93608606e-01 -3.66333395e-01 4.22078818e-01
2.32641436e-02 4.04426426e-01 -1.27699777e-01 -5.36889493e-01
4.17557418e-01 3.43687981e-01 -8.24635208e-01 -1.79339200e-01
4.62690532e-01 1.41854882e+00 -9.68904793e-01 -3.07997048e-01
-3.67949009e-02 2.26842925e-01 -8.37764263e-01 -2.70971447e-01
-4.22199368e-01 -5.26600540e-01 -2.73527652e-01 5.22105157e-01
4.71064776e-01 -5.41303039e-01 9.52569172e-02 1.68022126e-01
-8.69852155e-02 9.70204413e-01 -6.23809993e-01 1.38407588e+00
-1.03146732e+00 8.92992258e-01 -1.72235355e-01 -6.16326690e-01
3.74741346e-01 2.68175453e-01 1.32421657e-01 -6.97854400e-01
8.44805613e-02 1.93789110e-01 -1.60268649e-01 -7.51573294e-02
8.85754526e-01 -3.95235270e-02 -2.42717192e-01 9.99857306e-01
1.63420573e-01 -9.36552137e-02 3.90595943e-01 8.38366270e-01
1.17211974e+00 -3.42902899e-01 -1.40593648e-01 -2.87744343e-01
1.62101552e-01 -3.61886114e-01 3.02301496e-01 1.00272739e+00
3.14333230e-01 2.43459567e-01 1.12868881e+00 -4.58677024e-01
-1.14008486e+00 -1.14096618e+00 5.80813289e-02 1.49872756e+00
-4.13716108e-01 -8.38331401e-01 -4.97140646e-01 -6.91932857e-01
3.15424681e-01 7.43394971e-01 -6.94682419e-01 -8.17083493e-02
-6.21694505e-01 -6.72347188e-01 8.18867087e-01 1.14492273e+00
1.93088010e-01 -1.15004051e+00 -4.73430812e-01 2.86237150e-01
-6.04719073e-02 -1.03519237e+00 -5.39948642e-01 7.55748808e-01
-8.55083942e-01 -7.15089798e-01 -7.09442139e-01 -5.54461777e-01
6.67924702e-01 1.13620862e-01 1.49455416e+00 1.70213178e-01
-8.78237262e-02 -8.96489695e-02 -6.59355968e-02 -6.14202321e-01
-2.98362374e-02 4.72618967e-01 -1.24213390e-01 -5.17548323e-01
4.79846060e-01 -5.75495958e-01 -2.29465708e-01 -5.49238473e-02
-5.35974085e-01 1.47879764e-01 1.05705106e+00 1.05762792e+00
3.72965574e-01 -1.14771359e-01 3.63017261e-01 -1.27186227e+00
7.80706942e-01 -3.07851464e-01 -5.14023304e-01 4.00058359e-01
-7.21017778e-01 5.26776493e-01 1.02892292e+00 -2.52011150e-01
-9.70683932e-01 -4.93420333e-01 -3.30960602e-01 -1.39692321e-01
3.76993626e-01 7.13737667e-01 1.28401577e-01 3.18467170e-01
4.91876572e-01 2.53038257e-01 -3.54560763e-01 -5.36127865e-01
6.24255836e-01 3.88223320e-01 4.28285956e-01 -1.06276965e+00
6.44537568e-01 1.64842770e-01 -3.65384400e-01 -5.87564051e-01
-1.01430035e+00 -2.04373270e-01 -4.82268363e-01 4.03995603e-01
3.65080863e-01 -8.09703171e-01 -9.14467096e-01 5.47028184e-01
-1.38537705e+00 -7.11046457e-01 -4.00248915e-01 2.60223061e-01
7.84005821e-02 9.51457620e-02 -1.32311857e+00 -6.35215282e-01
-9.33336258e-01 -9.89431620e-01 9.17948365e-01 -1.56483427e-01
-1.77552581e-01 -9.07182693e-01 -4.14932519e-01 -3.77903990e-02
6.25536978e-01 -6.01103604e-01 1.26467466e+00 -5.25256455e-01
-8.19422960e-01 -2.87512809e-01 -5.41570127e-01 4.44672227e-01
-3.49442244e-01 5.05422913e-02 -1.05955958e+00 -1.03076994e-01
-6.68576002e-01 -6.38821840e-01 1.42773139e+00 2.34690025e-01
1.40477872e+00 -5.33193052e-01 -1.56804472e-01 4.23595250e-01
1.20963740e+00 -4.06987637e-01 4.75699812e-01 -1.22412182e-01
7.85830855e-01 2.40356117e-01 1.82731375e-01 2.44092122e-01
4.92497325e-01 1.94212440e-02 6.30178526e-02 -1.13110520e-01
1.75804808e-03 -3.77819896e-01 5.86668968e-01 1.00891292e+00
-1.26843881e-02 -3.23845506e-01 -1.03771484e+00 8.10459912e-01
-1.47922599e+00 -8.89328897e-01 -1.77610442e-01 2.01118755e+00
1.19780183e+00 7.55415797e-01 -3.50829035e-01 4.22611475e-01
-8.09408650e-02 -1.78470705e-02 -4.53920990e-01 -8.73844683e-01
-1.13901705e-01 9.25838828e-01 6.99220896e-01 7.11707592e-01
-6.83149397e-01 1.16557324e+00 7.39313602e+00 8.27639997e-01
-1.09385407e+00 -6.16147555e-02 7.16122448e-01 -5.34591138e-01
-7.39636123e-01 -9.09066871e-02 -9.28758025e-01 1.66498825e-01
9.91893828e-01 -2.68619269e-01 4.19550717e-01 8.03205967e-01
-1.77382559e-01 -3.97289321e-02 -1.25193596e+00 4.05196279e-01
-3.82398158e-01 -1.30807090e+00 3.52534622e-01 -1.39008269e-01
6.64938629e-01 4.82712299e-01 3.08734775e-01 5.88855863e-01
9.20905828e-01 -1.17240000e+00 7.22921193e-01 8.18360671e-02
9.63006616e-01 -8.17559123e-01 7.35602736e-01 2.05885142e-01
-1.25163591e+00 -9.56401080e-02 -6.29009128e-01 -3.80098164e-01
5.80326596e-04 1.00996983e+00 -1.04393399e+00 9.08773169e-02
6.77718699e-01 7.17853963e-01 -6.00447476e-01 6.20951891e-01
-8.04826021e-01 8.90397847e-01 -6.43889248e-01 -4.07588512e-01
6.18429661e-01 2.85908788e-01 2.90139560e-02 1.67864764e+00
1.58212230e-01 -1.56234745e-02 -2.91677248e-02 8.18802118e-01
-4.81810451e-01 -7.22295642e-02 -4.53013331e-01 -2.80751526e-01
6.26646101e-01 1.26318014e+00 -4.07976478e-01 -8.01050484e-01
-4.47670639e-01 7.47571945e-01 7.12372005e-01 4.69629943e-01
-5.93443632e-01 -8.45018089e-01 6.79798841e-01 4.80631888e-02
4.72198278e-01 -3.29597324e-01 -7.33918071e-01 -1.04076779e+00
1.01650469e-01 -6.31404936e-01 4.13174152e-01 -7.77427137e-01
-1.10586584e+00 5.81815898e-01 -2.28217378e-01 -4.76867735e-01
-1.68906361e-01 -9.65953469e-01 -7.63550401e-01 1.21569264e+00
-1.76540005e+00 -1.25084722e+00 1.61360830e-01 6.51891112e-01
3.74269128e-01 3.51070285e-01 7.98299730e-01 4.25991446e-01
-3.93883854e-01 1.10981977e+00 2.54071481e-03 3.59516472e-01
5.77017784e-01 -1.76461995e+00 1.10184407e+00 8.99588466e-01
2.41065428e-01 1.27865338e+00 3.76821607e-01 -1.94511294e-01
-1.66521287e+00 -9.71582472e-01 1.61043847e+00 -6.66821897e-01
8.98083091e-01 -7.98556924e-01 -7.38274336e-01 1.26696157e+00
5.76960206e-01 -6.67167976e-02 4.91952151e-01 9.38212216e-01
-7.77586043e-01 -5.79724982e-02 -7.16136754e-01 6.85148001e-01
9.27554786e-01 -8.87151122e-01 -8.03977728e-01 3.45021725e-01
1.01987123e+00 -2.91092068e-01 -7.95924187e-01 -1.73435003e-01
6.92908823e-01 -4.48456615e-01 9.64734435e-01 -6.75101638e-01
9.35353279e-01 1.67619824e-01 6.86030686e-02 -1.31655288e+00
-4.70360458e-01 -7.48732328e-01 -1.69134811e-01 9.97760296e-01
9.72812593e-01 -7.09117174e-01 7.04468668e-01 5.58736861e-01
-2.25358009e-01 -9.65293944e-01 -5.69241345e-01 -5.51264167e-01
6.32777631e-01 -8.70694876e-01 7.61354685e-01 7.36012638e-01
4.06040758e-01 7.28471994e-01 -1.72776058e-01 -1.82391003e-01
2.75719613e-01 2.18903676e-01 5.31112671e-01 -9.84336078e-01
-3.83917898e-01 -6.28179014e-01 2.81472683e-01 -1.59542942e+00
-8.59650150e-02 -1.18408680e+00 8.95422027e-02 -1.56880033e+00
2.31039330e-01 -6.66720569e-01 -4.93076563e-01 1.04311109e+00
-4.25674021e-01 1.56498805e-01 2.37357438e-01 -2.28936225e-01
-6.24287307e-01 4.01891500e-01 1.31101584e+00 -4.28544343e-01
2.55675167e-01 -2.48544678e-01 -9.86221433e-01 6.40232086e-01
4.63379204e-01 -2.66849339e-01 -3.24863344e-01 -1.27573931e+00
8.55850399e-01 -2.84619927e-01 1.65947661e-01 -8.16176236e-01
3.17226529e-01 -5.12326211e-02 4.51395810e-01 -8.42268288e-01
2.47947112e-01 -4.83099133e-01 -7.82807708e-01 4.33296770e-01
-7.42825329e-01 4.30663049e-01 3.40300769e-01 3.25841248e-01
-8.82493556e-02 -4.49940376e-02 3.88734430e-01 -3.73814046e-01
-3.44583660e-01 4.60131973e-01 -5.45983613e-01 3.16737771e-01
3.61002803e-01 1.24531247e-01 -5.10326445e-01 -6.03505187e-02
-4.23679054e-01 5.44916093e-01 -2.14321259e-02 1.62976682e-01
4.97180045e-01 -1.01194119e+00 -7.05462635e-01 3.80526602e-01
-1.32046938e-01 3.91203612e-01 -4.18521203e-02 7.11865723e-01
-3.83531600e-01 9.46406484e-01 2.38241330e-02 -1.96187258e-01
-1.00904632e+00 5.61346114e-01 -9.47551131e-02 -8.60712290e-01
-6.88165426e-01 1.34468079e+00 1.89800799e-01 -4.92098987e-01
2.97777534e-01 -1.27147245e+00 2.10730270e-01 6.14720583e-02
7.06732213e-01 2.69010775e-02 8.63803849e-02 3.19058388e-01
-3.40282857e-01 3.55595410e-01 -5.53202629e-01 -1.75396010e-01
1.25699234e+00 2.55180418e-01 -2.14405239e-01 1.83432013e-01
1.18559110e+00 2.52432317e-01 -8.95417094e-01 -6.44722462e-01
4.16499339e-02 -2.26097405e-01 3.31689278e-03 -1.03216529e+00
-1.11709559e+00 1.49895883e+00 -4.00597990e-01 7.72232637e-02
8.29463720e-01 -3.07837725e-01 1.20583057e+00 1.17514193e+00
3.46134871e-01 -9.91999447e-01 1.30577192e-01 1.24629724e+00
7.65196025e-01 -7.56597936e-01 -4.20950502e-02 -2.23420821e-02
-5.28499126e-01 1.02515602e+00 5.08678138e-01 -1.41598150e-01
5.93346596e-01 7.13419616e-01 -2.83549219e-01 -7.39176497e-02
-1.57407725e+00 1.41154036e-01 -1.10466238e-02 1.84575170e-01
7.56494224e-01 8.51749927e-02 2.00003814e-02 6.74889863e-01
-3.90284657e-01 6.36808127e-02 2.86337823e-01 8.74621451e-01
-4.18366790e-01 -1.16363084e+00 -4.82629985e-02 8.17258179e-01
-6.54473603e-01 -1.02367496e+00 -2.02599615e-01 4.81607795e-01
-2.22364664e-01 5.62733114e-01 2.21881986e-01 -2.29142144e-01
1.53409764e-01 2.54871458e-01 3.40318859e-01 -6.19025826e-01
-9.96962786e-01 -3.38323981e-01 1.58206061e-01 -4.23846126e-01
4.03774023e-01 -5.24902165e-01 -1.35459471e+00 -9.90374029e-01
-1.68709859e-01 2.50090122e-01 3.42918843e-01 8.71879935e-01
4.74825531e-01 7.95975029e-01 2.33549327e-01 -5.08585632e-01
-9.08527017e-01 -1.23961973e+00 -2.97547400e-01 -1.25676766e-02
3.60728532e-01 -1.15776494e-01 -1.23673648e-01 -8.13066736e-02] | [10.90000057220459, 7.669509410858154] |
7c2a2dad-eec0-49a7-b8c4-134fda82530b | neural-graph-matching-networks-for-fewshot-3d | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Michelle_Guo_Neural_Graph_Matching_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Michelle_Guo_Neural_Graph_Matching_ECCV_2018_paper.pdf | Neural Graph Matching Networks for Fewshot 3D Action Recognition | We propose Neural Graph Matching (NGM) Networks, a novel framework that can learn to recognize a previous unseen 3D action class with only a few examples. We achieve this by leveraging the inherent structure of 3D data through a graphical representation. This allows us to modularize our model and lead to strong data-efficiency in few-shot learning. More specifically, NGM Networks jointly learn a graph generator and a graph matching metric function in a end-to-end fashion to directly optimize the few-shot learning objective. We evaluate NGM on two 3D action recognition datasets, CAD-120 and PiGraphs, and show that learning to generate and match graphs both lead to significant improvement of few-shot 3D action recognition over the holistic baselines. | ['Li Fei-Fei', 'De-An Huang', 'Edward Chou', 'Serena Yeung', 'Michelle Guo', 'Shuran Song'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['3d-human-action-recognition'] | ['computer-vision'] | [ 4.66152310e-01 2.74768829e-01 -4.07577187e-01 -4.54581827e-01
-8.06110501e-01 -3.76907229e-01 7.88314939e-01 -1.80095285e-01
-5.71957901e-02 8.74423459e-02 4.88149792e-01 -9.68772992e-02
8.69558230e-02 -7.43472159e-01 -8.12208593e-01 -2.99143583e-01
-1.20471418e-01 5.29434621e-01 2.35316530e-01 6.52437508e-02
-5.28515177e-03 5.76557338e-01 -1.50822580e+00 4.03395951e-01
4.50481743e-01 9.07167256e-01 -4.97184806e-02 7.97686100e-01
9.70473662e-02 1.20110166e+00 -2.35084236e-01 -4.30080682e-01
6.54493928e-01 -6.14669561e-01 -6.22008622e-01 5.59132457e-01
1.15368259e+00 -7.65256822e-01 -1.09200072e+00 7.58093417e-01
5.99293590e-01 5.24303257e-01 7.99928725e-01 -1.37044084e+00
-8.09859931e-01 3.92351806e-01 -4.88477379e-01 1.79502979e-01
4.38806564e-01 7.20351934e-01 1.13400674e+00 -7.81782568e-01
8.07889342e-01 1.47964394e+00 6.42086089e-01 9.08738256e-01
-1.29781318e+00 -3.72618854e-01 9.57065895e-02 2.30121002e-01
-9.42096472e-01 -4.98867065e-01 7.77669668e-01 -4.09011245e-01
1.57963705e+00 -3.69493663e-01 8.49058032e-01 1.47016001e+00
1.42646760e-01 1.02443779e+00 4.72038716e-01 -3.84370834e-01
3.26656252e-01 -8.15622985e-01 3.26513857e-01 1.04710746e+00
5.73049970e-02 4.03821260e-01 -8.76082003e-01 2.05950454e-01
7.24192023e-01 2.60979474e-01 9.27518457e-02 -1.07227588e+00
-7.59925127e-01 7.31605768e-01 5.19735217e-01 -1.32012859e-01
-1.20414615e-01 6.00822330e-01 4.51388866e-01 3.50003272e-01
3.97614390e-01 3.26795459e-01 9.15025081e-03 -3.29607040e-01
-4.49029416e-01 2.38282010e-01 5.90246677e-01 1.04412222e+00
7.27097631e-01 2.43718803e-01 -4.97100621e-01 5.99408269e-01
-5.63478060e-02 5.66861212e-01 2.58633375e-01 -1.01108563e+00
6.14209116e-01 9.37183380e-01 -3.07238579e-01 -7.76303887e-01
-3.96158874e-01 -2.06517220e-01 -5.76033354e-01 5.31281471e-01
4.21315581e-01 7.12036639e-02 -1.52503884e+00 1.89929664e+00
2.86119580e-01 3.52414161e-01 1.51474878e-01 7.47159004e-01
8.69605780e-01 2.25057587e-01 5.11157401e-02 4.78030890e-01
8.29166770e-01 -1.11667120e+00 -2.14160696e-01 -5.97060502e-01
1.01262236e+00 -4.02929969e-02 1.30179298e+00 -1.52503893e-01
-9.45966959e-01 -6.63975120e-01 -1.12934518e+00 -2.89154053e-01
-2.80128002e-01 -2.61823058e-01 6.87268496e-01 2.97280163e-01
-8.20210636e-01 9.81576502e-01 -1.15269291e+00 -5.99611163e-01
1.04368591e+00 8.58517736e-02 -6.00676060e-01 -5.10803282e-01
-6.58726156e-01 7.44300783e-01 4.49591130e-01 -3.89537483e-01
-1.50724876e+00 -6.58973575e-01 -1.24325168e+00 6.19749688e-02
6.46987259e-01 -9.86931264e-01 1.31711805e+00 -4.78603542e-01
-1.31673920e+00 1.17893267e+00 2.86310822e-01 -7.62338281e-01
4.13376927e-01 -2.03647286e-01 -1.24097809e-01 3.17945153e-01
3.11762933e-02 7.28084028e-01 7.11922884e-01 -5.42122602e-01
-3.46372485e-01 -5.36184847e-01 3.08747500e-01 1.73594743e-01
-1.04669094e-01 -3.40303808e-01 -4.45508629e-01 -6.19593740e-01
-6.62156269e-02 -8.42112422e-01 -2.67894149e-01 4.06278193e-01
-2.24052042e-01 -1.49225980e-01 6.91521406e-01 -2.83755034e-01
6.19326949e-01 -2.15922022e+00 3.24224800e-01 -1.70096487e-01
3.74370128e-01 2.30131671e-01 -7.52006054e-01 4.93624151e-01
-1.03214055e-01 -3.33851963e-01 -3.58670801e-01 -4.03225780e-01
1.47340924e-01 5.04704177e-01 -1.31538317e-01 3.81351918e-01
4.78228629e-01 1.64717829e+00 -1.13813102e+00 -2.19254240e-01
3.43165338e-01 3.27558786e-01 -6.12861097e-01 3.52122456e-01
-4.05304849e-01 2.08248273e-02 -2.50996381e-01 9.25098956e-01
2.14144275e-01 -5.91988325e-01 1.04103141e-01 -1.54948503e-01
6.58823907e-01 4.33827676e-02 -8.69748056e-01 2.35658979e+00
-2.21318409e-01 4.09062028e-01 -4.75327551e-01 -1.09733307e+00
9.21065032e-01 -2.58737970e-02 4.18950230e-01 -8.43654573e-01
2.55590200e-01 -1.43022388e-01 -2.58815825e-01 -4.79631305e-01
6.14733920e-02 -1.61193967e-01 -2.31137946e-01 6.84739590e-01
7.96343207e-01 -8.93319771e-03 1.40909493e-01 5.15272319e-01
1.71050715e+00 4.25655156e-01 4.03173149e-01 2.03999802e-01
-1.11441694e-01 -7.28722736e-02 5.39246261e-01 1.05345654e+00
-2.97700107e-01 5.55568874e-01 6.13856018e-01 -6.43915772e-01
-9.74112153e-01 -1.38110816e+00 7.58290470e-01 1.00722027e+00
2.67136320e-02 -5.18812120e-01 -5.12511790e-01 -1.28664184e+00
3.50913912e-01 6.55189395e-01 -8.45701277e-01 -8.02471638e-01
-5.01190186e-01 3.12115550e-02 3.89975518e-01 9.80840981e-01
4.40755278e-01 -9.07421112e-01 -7.47854173e-01 6.29058555e-02
3.89842600e-01 -1.33135736e+00 -6.79094911e-01 3.07325751e-01
-1.15144157e+00 -1.41481054e+00 -6.16320193e-01 -5.89919150e-01
5.98688900e-01 5.19025564e-01 1.41334105e+00 -5.92627078e-02
-6.14165068e-01 6.87304854e-01 -5.12515426e-01 -5.47533594e-02
-5.62700272e-01 -5.75221740e-02 -1.08166151e-01 -9.24545452e-02
6.67985737e-01 -1.01298046e+00 -3.43471944e-01 3.72616351e-02
-6.77983046e-01 9.89054516e-02 6.53622687e-01 7.80413508e-01
5.98260164e-01 -5.27839005e-01 3.36231828e-01 -7.27944076e-01
2.70740002e-01 -1.65145993e-01 -5.84909618e-01 3.21178824e-01
-4.68802571e-01 3.28580737e-01 4.21581835e-01 -3.73998374e-01
-5.70717216e-01 4.18581933e-01 1.63147658e-01 -1.12872028e+00
-3.00296277e-01 2.58886933e-01 -4.24760938e-01 -2.54465699e-01
8.55235279e-01 1.80050775e-01 4.09699559e-01 -6.25606477e-01
8.55775476e-01 1.19328871e-01 8.48787665e-01 -2.59943783e-01
9.53218877e-01 4.03779864e-01 3.96294564e-01 -4.93853956e-01
-1.19288957e+00 -4.06524211e-01 -9.87582624e-01 -3.78312081e-01
1.03068972e+00 -1.01474833e+00 -5.03057659e-01 6.22125268e-01
-7.20822692e-01 -8.23300600e-01 -7.40726352e-01 3.84772569e-01
-1.03258443e+00 4.45057184e-01 -5.20242274e-01 -5.26155710e-01
-2.92165309e-01 -6.87093496e-01 1.35165882e+00 1.56234773e-02
-2.13335440e-01 -7.21139193e-01 2.47319698e-01 2.25403368e-01
-6.04332127e-02 7.02751040e-01 9.63654935e-01 -6.53553367e-01
-6.48981392e-01 -3.81351650e-01 -2.79939383e-01 4.54893321e-01
1.21922247e-01 -4.50068474e-01 -8.14860582e-01 -3.62165570e-01
-4.51240927e-01 -1.06538188e+00 1.22562325e+00 7.34171048e-02
8.77531290e-01 -1.52654439e-01 -2.28985220e-01 8.02729785e-01
1.36272788e+00 -5.16711064e-02 6.35829568e-01 -3.32922995e-01
1.08818793e+00 1.66225791e-01 4.48645145e-01 4.42414731e-01
4.87716310e-02 6.08967066e-01 5.08321345e-01 1.71728283e-01
-7.13879228e-01 -9.07291949e-01 3.31367284e-01 3.45701724e-01
8.15784782e-02 -1.47239536e-01 -8.76140833e-01 4.22422081e-01
-2.04790115e+00 -1.21283662e+00 4.82640058e-01 2.01850247e+00
4.58574533e-01 2.75138468e-01 4.10648227e-01 -1.22709125e-01
5.04346490e-01 6.76521242e-01 -9.43909883e-01 -6.05477951e-03
-9.44702476e-02 4.87490892e-01 3.76873046e-01 3.23361307e-01
-1.10145962e+00 1.17303658e+00 6.56588221e+00 6.43099606e-01
-5.68864524e-01 -2.00113922e-01 2.37779215e-01 -4.73576248e-01
-8.87129456e-02 2.99948640e-02 -6.37025774e-01 2.56871898e-02
7.41813123e-01 -1.74430087e-01 4.02352095e-01 1.05656850e+00
-1.89505383e-01 3.12787801e-01 -1.56939912e+00 1.33638144e+00
2.75979489e-01 -1.61905801e+00 4.51997310e-01 1.59397289e-01
7.34153748e-01 3.38789344e-01 -3.06651324e-01 5.60724616e-01
7.60356069e-01 -9.69900727e-01 4.04213548e-01 6.04730427e-01
6.51855350e-01 -6.43810868e-01 1.03554577e-01 2.71631896e-01
-1.19245303e+00 -2.38025635e-01 -4.78753150e-01 -2.71653116e-01
3.61690879e-01 2.17869923e-01 -7.67208338e-01 4.21715081e-01
2.79333383e-01 1.34655762e+00 -6.37078762e-01 1.03956783e+00
-4.71320719e-01 2.85881281e-01 8.26985613e-02 -2.62906961e-02
4.96854544e-01 2.74789650e-02 6.96787477e-01 6.81700647e-01
1.79084942e-01 2.76642382e-01 4.16367292e-01 1.01619995e+00
-4.43841428e-01 -2.94348150e-01 -1.15179205e+00 -5.51461279e-01
1.68897137e-01 8.87168050e-01 -4.55626607e-01 -4.11869705e-01
-5.37919104e-01 1.22749293e+00 7.14725375e-01 2.60589719e-01
-6.00186408e-01 -3.40460896e-01 8.06910217e-01 -6.19989000e-02
5.55839181e-01 -4.04603362e-01 1.17526233e-01 -1.28806615e+00
3.16332243e-02 -1.02040029e+00 7.05891013e-01 -8.15835238e-01
-1.52017999e+00 1.37837425e-01 -1.28301203e-01 -1.33223534e+00
-5.52127123e-01 -7.01112390e-01 -7.02075958e-01 4.24675703e-01
-1.13901520e+00 -1.32380319e+00 -5.52578747e-01 7.16784656e-01
5.39410949e-01 -3.80101502e-01 7.63944566e-01 2.22509149e-02
-3.25339198e-01 7.84256160e-01 -3.69982421e-01 4.35458720e-01
4.36522514e-01 -1.30020428e+00 1.29153907e+00 8.53876114e-01
8.00441206e-01 -4.71067317e-02 2.39230543e-01 -7.43575633e-01
-1.62090421e+00 -1.25083375e+00 5.62266171e-01 -7.74639904e-01
5.73689818e-01 -7.10988820e-01 -8.02756906e-01 7.60846555e-01
-1.09765202e-01 3.82452667e-01 5.18958688e-01 1.64522901e-01
-8.94817591e-01 -3.43380347e-02 -9.40867841e-01 6.33708715e-01
2.02955270e+00 -8.64138067e-01 -8.14152598e-01 4.50371325e-01
9.05149341e-01 -3.59771281e-01 -8.01586211e-01 3.80391032e-01
5.76947033e-01 -9.52914119e-01 9.64500844e-01 -1.44222939e+00
4.59039778e-01 8.10072124e-02 -4.71972704e-01 -1.26840580e+00
-3.41865152e-01 -8.05249989e-01 -6.59847498e-01 7.08504856e-01
2.18401715e-01 -2.62250185e-01 1.20094645e+00 5.43706775e-01
-2.82864302e-01 -7.61265099e-01 -8.08726728e-01 -1.38506341e+00
-1.32786721e-01 -4.16311175e-01 5.54497540e-01 6.13259792e-01
-4.73305136e-02 5.58687449e-01 -5.84476054e-01 -1.50892735e-01
1.07101011e+00 2.66863167e-01 1.28145111e+00 -1.12933993e+00
-7.16115475e-01 -3.77378464e-01 -1.16683161e+00 -1.31366611e+00
3.34619164e-01 -1.25213587e+00 -4.93225828e-02 -1.64027107e+00
3.33617002e-01 2.71055728e-01 -3.63457978e-01 8.85126770e-01
-5.16603254e-02 2.69708842e-01 3.55672508e-01 -1.18942298e-01
-8.59254122e-01 8.37456763e-01 1.22991860e+00 -4.30927932e-01
-1.34597898e-01 -2.07487464e-01 -4.81741846e-01 4.97654706e-01
3.92260373e-01 -4.87233400e-01 -7.25656331e-01 -6.10211074e-01
-1.85458392e-01 1.24979943e-01 7.65857518e-01 -1.25928640e+00
8.16167817e-02 -7.61906132e-02 3.07161719e-01 -6.26503825e-01
4.73160118e-01 -5.12427390e-01 -2.24738255e-01 4.16542590e-01
-4.68180269e-01 -2.92792380e-01 9.75960791e-02 1.00219977e+00
1.17938243e-01 1.12711877e-01 7.74510503e-01 -2.83719003e-01
-1.18270326e+00 9.07462358e-01 2.28544012e-01 5.60027957e-01
9.32542324e-01 -4.59220260e-01 -3.43185216e-01 -4.12609220e-01
-8.13701034e-01 2.14876473e-01 6.54497564e-01 6.61323249e-01
7.46914923e-01 -1.75281012e+00 -4.48849082e-01 1.85799181e-01
5.64240396e-01 -2.50640899e-01 3.20328981e-01 5.01035452e-01
-1.15662411e-01 4.52316776e-02 -4.95323032e-01 -6.58028185e-01
-1.20176506e+00 7.39875317e-01 4.85667437e-01 -1.23353608e-01
-1.15815806e+00 9.05208707e-01 5.87234870e-02 -3.65271211e-01
6.36363149e-01 -3.03517967e-01 4.25276995e-01 -2.36974850e-01
5.97979307e-01 3.15097153e-01 -2.96954989e-01 -2.68533230e-01
-4.03831035e-01 5.95853806e-01 2.46357489e-02 4.85051833e-02
1.45116615e+00 2.73166597e-01 5.80059946e-01 4.36878651e-01
1.53014004e+00 -7.47097790e-01 -1.84622538e+00 -3.87959749e-01
-6.30019456e-02 -7.53188848e-01 -1.32483274e-01 -7.20435739e-01
-1.12879586e+00 9.30555522e-01 4.82917070e-01 -2.59302050e-01
8.35272312e-01 2.88570106e-01 9.35647070e-01 8.65033269e-01
2.99691737e-01 -9.78031397e-01 7.56476462e-01 5.35018384e-01
7.34996557e-01 -1.34006608e+00 -1.03515908e-01 4.21723425e-02
-4.92936343e-01 1.04333341e+00 6.54026806e-01 -4.24865127e-01
4.82802361e-01 -1.53659852e-02 -7.33568743e-02 -6.15725279e-01
-8.60031545e-01 -6.03430271e-01 5.25546491e-01 9.87935722e-01
-1.88694358e-01 -2.04598546e-01 4.42936599e-01 3.27572733e-01
2.51680315e-01 2.37991557e-01 3.00339669e-01 1.11091065e+00
-5.27216554e-01 -1.00110316e+00 5.03282189e-01 4.51522529e-01
1.30335346e-01 1.36370361e-01 -9.01003420e-01 9.95456576e-01
-4.97513264e-01 6.79747581e-01 1.00941122e-01 -7.61055112e-01
6.32075906e-01 3.56077343e-01 8.41794252e-01 -8.65032136e-01
-2.44663330e-03 -4.23106968e-01 3.31000298e-01 -1.25697744e+00
-2.98587859e-01 -3.82211804e-01 -1.10549986e+00 -7.02585950e-02
-3.06716505e-02 -4.05755252e-01 1.57605261e-01 9.38608944e-01
7.75003433e-01 4.00674909e-01 5.22536039e-01 -8.06442261e-01
-9.77439523e-01 -7.21575797e-01 -7.05106378e-01 7.93984890e-01
9.98721123e-02 -8.78260314e-01 -3.65078479e-01 -1.76259041e-01] | [8.692953109741211, 0.9577859044075012] |
15c253e8-1ff2-48a0-952e-25bbff1c536b | refining-image-categorization-by-exploiting | 1703.05451 | null | http://arxiv.org/abs/1703.05451v1 | http://arxiv.org/pdf/1703.05451v1.pdf | Refining Image Categorization by Exploiting Web Images and General Corpus | Studies show that refining real-world categories into semantic subcategories
contributes to better image modeling and classification. Previous image
sub-categorization work relying on labeled images and WordNet's hierarchy is
not only labor-intensive, but also restricted to classify images into NOUN
subcategories. To tackle these problems, in this work, we exploit general
corpus information to automatically select and subsequently classify web images
into semantic rich (sub-)categories. The following two major challenges are
well studied: 1) noise in the labels of subcategories derived from the general
corpus; 2) noise in the labels of images retrieved from the web. Specifically,
we first obtain the semantic refinement subcategories from the text perspective
and remove the noise by the relevance-based approach. To suppress the search
error induced noisy images, we then formulate image selection and classifier
learning as a multi-class multi-instance learning problem and propose to solve
the employed problem by the cutting-plane algorithm. The experiments show
significant performance gains by using the generated data of our way on both
image categorization and sub-categorization tasks. The proposed approach also
consistently outperforms existing weakly supervised and web-supervised
approaches. | ['Xian-Sheng Hua', 'Wankou Yang', 'Yazhou Yao', 'Fumin Shen', 'Zhenmin Tang', 'Jian Zhang'] | 2017-03-16 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 5.74516177e-01 2.58109383e-02 -4.32141095e-01 -4.71999049e-01
-9.64338899e-01 -5.70924163e-01 4.60468799e-01 2.50658363e-01
-5.82358539e-01 4.95241731e-01 6.94827884e-02 6.48515895e-02
-3.22810143e-01 -8.27607274e-01 -4.65671390e-01 -7.01248765e-01
5.83453119e-01 4.65065122e-01 5.09436905e-01 -4.13561463e-02
6.71090066e-01 3.76871899e-02 -2.35558438e+00 6.37287378e-01
9.21704948e-01 1.30005813e+00 5.45205414e-01 1.51892722e-01
-5.29668510e-01 4.40823972e-01 -1.51534155e-01 -2.84536064e-01
2.04655811e-01 -2.25793734e-01 -1.27180910e+00 7.97273099e-01
5.12901902e-01 1.10611625e-01 3.99233013e-01 1.52038920e+00
1.51080430e-01 1.60196088e-02 7.56605446e-01 -1.19219780e+00
-5.12567520e-01 3.44444901e-01 -7.03120232e-01 -1.06446799e-02
1.25338342e-02 -3.31550151e-01 1.04198277e+00 -9.36919630e-01
8.02939951e-01 1.26563215e+00 4.54199642e-01 5.34104884e-01
-1.15905046e+00 -6.79514647e-01 4.10606891e-01 5.05807936e-01
-1.56643558e+00 -1.76177874e-01 7.36748099e-01 -5.66184342e-01
3.55490327e-01 3.22266221e-01 1.53331652e-01 7.75338948e-01
-3.37448299e-01 6.19729042e-01 1.37951279e+00 -7.54794300e-01
1.67128891e-01 7.81738281e-01 6.15569353e-01 6.17393017e-01
1.32102773e-01 -4.10077572e-01 -2.14701474e-01 -3.04283388e-02
2.34282047e-01 1.85622387e-02 -1.60938427e-01 -4.26092148e-01
-1.10092640e+00 9.25445974e-01 3.91252279e-01 4.71858114e-01
-3.47148865e-01 -2.57899910e-01 4.69462186e-01 1.85742334e-01
7.87712395e-01 2.80195147e-01 -6.55614197e-01 5.96229076e-01
-7.90560007e-01 2.26915218e-02 4.87408042e-01 1.17303109e+00
9.50468659e-01 -5.54976225e-01 8.14761221e-02 1.42816150e+00
3.12386721e-01 2.34946907e-01 7.30084419e-01 -8.31489742e-01
5.10254502e-01 8.55206728e-01 -1.73087507e-01 -1.02615285e+00
-3.54819208e-01 -3.42416435e-01 -7.01276958e-01 -6.20895103e-02
3.91414136e-01 3.51523399e-01 -1.14405441e+00 1.40500581e+00
5.03686845e-01 -1.25434011e-01 1.10723220e-01 8.97021592e-01
1.18933928e+00 5.63645065e-01 6.06031239e-01 -3.56624901e-01
1.80354011e+00 -1.08744431e+00 -6.08842075e-01 -1.75630748e-01
3.75248879e-01 -8.35401773e-01 1.05790806e+00 3.73378307e-01
-7.09861815e-01 -8.77190053e-01 -8.04537654e-01 7.72980377e-02
-8.31797421e-01 3.54119301e-01 3.52911264e-01 4.57685351e-01
-8.78930926e-01 2.14133367e-01 -6.96659982e-02 -7.36199856e-01
5.02893388e-01 2.60802627e-01 -3.78836989e-01 -2.70168632e-01
-1.00718784e+00 5.93831718e-01 8.20832372e-01 -3.17210644e-01
-4.99377221e-01 -4.74182606e-01 -8.11327517e-01 -1.16577838e-03
7.80762732e-01 -3.39440495e-01 7.82266140e-01 -1.52060890e+00
-8.93608212e-01 1.35823584e+00 -1.37548283e-01 -1.86049044e-01
1.01607502e-01 2.61460990e-01 -3.09700489e-01 3.74217272e-01
5.99187791e-01 9.50407684e-01 8.71991932e-01 -1.78000152e+00
-1.36125016e+00 -6.83248103e-01 -4.64182012e-02 4.53965038e-01
-6.17344379e-01 -1.85716450e-01 -8.37724864e-01 -7.84239948e-01
6.00063801e-01 -9.74492788e-01 -1.86060429e-01 -3.04881305e-01
-3.47350299e-01 -6.98941648e-01 7.20445514e-01 -5.71613133e-01
9.92992163e-01 -2.10655069e+00 5.67603745e-02 3.11104208e-01
7.86084905e-02 -2.13899501e-02 -2.32329458e-01 -4.79862466e-02
2.49489695e-02 3.07379365e-01 -4.85860072e-02 1.54406615e-02
-3.55130672e-01 1.59817591e-01 -1.09737769e-01 1.21698007e-01
-1.85761601e-01 5.44739664e-01 -7.59949744e-01 -1.07145667e+00
2.86745220e-01 6.28372878e-02 -4.28865612e-01 -2.18326617e-02
-2.51507550e-01 3.23809147e-01 -5.25842607e-01 8.34503531e-01
7.09934831e-01 -1.93865016e-01 2.71164775e-01 -6.46929741e-01
-2.33622473e-02 -1.26983792e-01 -1.26220727e+00 1.46733034e+00
-5.57856560e-01 1.25892118e-01 -4.05199528e-02 -1.49453080e+00
7.69116461e-01 5.13319746e-02 5.87039411e-01 -8.05470705e-01
3.66406254e-02 4.26828533e-01 -4.54142004e-01 -8.29367578e-01
3.19077909e-01 -2.02756450e-01 -1.21634156e-01 2.30751663e-01
3.70413840e-01 -1.54053330e-01 5.00192821e-01 1.10902591e-02
4.38340008e-01 1.68755978e-01 2.52273381e-01 -5.89021385e-01
9.12959158e-01 6.15333676e-01 4.39850658e-01 8.76785994e-01
-4.73357961e-02 6.69403434e-01 5.21598868e-02 -3.84124130e-01
-8.74620438e-01 -6.65335834e-01 -3.52010339e-01 1.47314978e+00
6.55830562e-01 -2.39069283e-01 -9.81293678e-01 -1.09032953e+00
-1.09444886e-01 5.73786616e-01 -4.75630671e-01 1.25048131e-01
-2.32311919e-01 -1.00918710e+00 2.21073609e-02 2.46970281e-01
7.08944380e-01 -1.15461481e+00 -1.91160023e-01 6.04875721e-02
-6.58194125e-01 -1.18457055e+00 -3.28307778e-01 2.78750449e-01
-7.45491505e-01 -1.39887011e+00 -5.33332229e-01 -1.32943416e+00
8.24201584e-01 6.77733660e-01 1.05699766e+00 2.91173935e-01
-3.71629685e-01 5.51328182e-01 -7.24043012e-01 -2.46630028e-01
-2.73401856e-01 2.24970624e-01 -2.17558637e-01 2.76269555e-01
6.95236802e-01 -5.87588735e-02 -4.50512230e-01 5.57700872e-01
-1.03122926e+00 1.56324729e-01 6.02550924e-01 8.74959707e-01
8.97386432e-01 6.22491300e-01 7.01366782e-01 -1.12936425e+00
2.92779326e-01 -5.17643750e-01 -4.88427043e-01 4.34675455e-01
-1.04485512e+00 8.20337012e-02 3.78343493e-01 -4.16999221e-01
-1.35354471e+00 5.47788143e-01 9.40237287e-03 1.81998923e-01
-5.13312578e-01 3.06331992e-01 -1.90214396e-01 -1.44376829e-01
4.98260140e-01 4.05444354e-01 -1.80553034e-01 -5.32372475e-01
3.66285890e-01 1.10662985e+00 3.55246603e-01 -4.97864634e-01
5.18355429e-01 6.34970605e-01 -2.16426164e-01 -1.05677879e+00
-1.27861428e+00 -1.19594288e+00 -8.01835597e-01 -3.81572187e-01
1.12710595e+00 -9.95232224e-01 -2.91898996e-01 1.99715033e-01
-9.34049129e-01 2.21645907e-01 -1.16470590e-01 2.31425062e-01
-4.23166543e-01 6.40476227e-01 -3.60097826e-01 -5.35601079e-01
-2.63079673e-01 -1.28459156e+00 1.31375933e+00 1.90609455e-01
5.98513260e-02 -6.02012336e-01 -4.01749194e-01 9.54714715e-01
6.42819628e-02 -2.06594512e-01 1.20452106e+00 -6.83030963e-01
-3.68208200e-01 -8.76726210e-02 -6.58998311e-01 5.25131106e-01
4.66042273e-02 -6.25060916e-01 -9.34341013e-01 -2.77968403e-02
9.17417184e-03 -5.25023818e-01 1.05105281e+00 2.80726850e-01
1.48092806e+00 -3.32437992e-01 -4.85841215e-01 3.15556794e-01
1.84958231e+00 1.15454793e-01 2.62169093e-01 6.66810095e-01
7.53635466e-01 1.19038939e+00 9.42692518e-01 1.35031328e-01
4.25494134e-01 6.67211533e-01 3.50136846e-01 -3.42489481e-02
-2.55371183e-01 7.00506056e-03 -2.79675007e-01 6.64003968e-01
-8.04918166e-03 -1.12096474e-01 -7.22356021e-01 6.82443142e-01
-1.95119917e+00 -7.45339334e-01 -4.24962878e-01 1.88395488e+00
1.05913115e+00 8.26525912e-02 1.61313650e-03 1.58480749e-01
1.06741917e+00 -2.02995196e-01 -2.34139532e-01 1.65100157e-01
8.00252485e-04 1.36713246e-02 7.07880437e-01 3.33753794e-01
-1.56758201e+00 1.01139176e+00 5.39510727e+00 1.11657631e+00
-6.78677559e-01 3.75743449e-01 8.71104062e-01 4.54809427e-01
-1.91907175e-02 -8.80684406e-02 -9.28700328e-01 4.24477100e-01
4.22051430e-01 1.89032644e-01 1.52566433e-01 1.06603539e+00
-1.66971996e-01 -3.19830835e-01 -7.11942792e-01 1.19386125e+00
4.17342335e-01 -1.03309596e+00 4.95582819e-01 -8.73403773e-02
8.25396657e-01 -2.55884260e-01 -1.99403971e-01 2.12259829e-01
7.36640096e-02 -5.96491635e-01 9.63588774e-01 6.49382547e-02
8.04457307e-01 -5.66270411e-01 9.11879599e-01 3.82406414e-01
-1.25910521e+00 -3.71725291e-01 -5.40856898e-01 3.60061944e-01
-3.65373135e-01 4.61673707e-01 -4.03557897e-01 4.48863059e-01
1.00582051e+00 5.84526122e-01 -8.88844192e-01 8.33329797e-01
4.79933731e-02 4.84477609e-01 -8.47968385e-02 8.08124393e-02
3.61935943e-01 -2.90078849e-01 5.71767576e-02 1.27114034e+00
7.30603486e-02 1.43173650e-01 4.41345900e-01 4.75508302e-01
3.43730226e-02 3.84539545e-01 -2.98944443e-01 3.58552486e-01
1.81819648e-01 1.49022937e+00 -1.36362803e+00 -5.43496788e-01
-4.47100937e-01 9.55546737e-01 3.05608474e-03 3.45147282e-01
-5.08441746e-01 -3.87511045e-01 8.84393752e-02 1.65423527e-01
2.37718612e-01 2.26373523e-01 -4.25088286e-01 -1.09488225e+00
-2.03312617e-02 -8.03073049e-01 8.85800719e-01 -6.56572759e-01
-1.33293772e+00 7.60623574e-01 2.91496366e-01 -1.21872759e+00
4.29501683e-02 -7.09331691e-01 1.86245993e-01 3.65155220e-01
-1.64550126e+00 -1.37175739e+00 -7.26212740e-01 5.73171020e-01
1.24720013e+00 -3.02377734e-02 6.30817473e-01 4.89308983e-01
-1.92081764e-01 1.08222455e-01 1.04485042e-01 -4.40154644e-03
6.87275708e-01 -1.23015881e+00 -4.17880625e-01 5.97840428e-01
1.33317679e-01 2.14483470e-01 4.18082625e-01 -5.58246255e-01
-7.92201698e-01 -1.20079255e+00 9.42742288e-01 -2.44720340e-01
3.34425151e-01 -3.31728965e-01 -7.56872714e-01 1.77192822e-01
3.74808609e-02 -1.27430126e-01 4.90066230e-01 -1.06424004e-01
-4.50212866e-01 -2.07526490e-01 -1.17296302e+00 3.09328407e-01
1.04236782e+00 -3.27501059e-01 -5.98414302e-01 6.88238025e-01
5.36637425e-01 1.63601503e-01 -4.24829006e-01 4.95195389e-01
2.95225531e-01 -6.10006690e-01 1.15110803e+00 -4.75355715e-01
4.60052967e-01 -3.20510745e-01 -2.38819286e-01 -1.12184262e+00
-3.20232183e-01 3.32378328e-01 7.22362578e-01 1.40677738e+00
3.44197184e-01 -2.59934098e-01 7.22713530e-01 1.53935090e-01
2.01897129e-01 -3.31781149e-01 -7.48044074e-01 -5.48232079e-01
-1.55229762e-01 -4.24102396e-01 1.52968556e-01 8.03113937e-01
-1.43409789e-01 4.57675517e-01 -4.52046171e-02 3.82956713e-02
9.66645896e-01 3.69194597e-01 2.68787682e-01 -1.41483700e+00
2.42031723e-01 -3.64372492e-01 -2.28414521e-01 -6.48618460e-01
3.04494232e-01 -1.07706571e+00 5.17680764e-01 -1.55646169e+00
7.37617254e-01 -7.14973927e-01 -1.79427266e-01 4.12583768e-01
-1.97784558e-01 7.79427648e-01 1.40931815e-01 5.62768638e-01
-1.11017430e+00 9.68896300e-02 8.45729172e-01 -4.96432632e-01
5.49919307e-02 2.07079295e-02 -8.36663485e-01 9.45866227e-01
7.65636623e-01 -5.95805287e-01 -4.73268390e-01 -1.44692004e-01
1.96565110e-02 -3.74218583e-01 4.08166796e-01 -8.32034647e-01
1.68374047e-01 -3.12123209e-01 1.54683441e-01 -5.94075322e-01
3.79164657e-03 -1.11402094e+00 -1.39407709e-01 3.62736464e-01
-5.72522283e-01 -4.45153713e-01 -1.22537963e-01 7.58563459e-01
-4.24590677e-01 -7.65264809e-01 1.05452728e+00 -6.25015616e-01
-1.23413408e+00 1.62337236e-02 -4.39589351e-01 3.04590221e-02
1.07999074e+00 -1.39766470e-01 -7.22658932e-02 -4.27494422e-02
-8.56002867e-01 9.99307781e-02 9.99988467e-02 5.68943799e-01
4.38495159e-01 -1.10921383e+00 -4.88789558e-01 -2.37070359e-02
5.64751923e-01 -1.71161965e-01 2.46900797e-01 5.06715119e-01
-2.92590737e-01 4.91049409e-01 -6.19615205e-02 -7.13524222e-01
-1.58422053e+00 8.15844536e-01 1.02692842e-01 -8.58384743e-02
-2.92768389e-01 5.45571685e-01 4.27860439e-01 -6.12587452e-01
4.85711306e-01 2.19077934e-02 -8.31526995e-01 5.29881120e-01
3.43972266e-01 2.18159914e-01 1.87808648e-01 -8.96182656e-01
-3.30583960e-01 9.39374506e-01 3.61618474e-02 1.60287172e-01
1.34832680e+00 -6.12712920e-01 -3.79769802e-01 1.28655300e-01
1.28969467e+00 -4.66018379e-01 -6.77174270e-01 -4.63339180e-01
5.02844751e-01 -3.05856884e-01 2.10626528e-01 -8.28398883e-01
-1.15227258e+00 5.21721303e-01 9.01131332e-01 4.41475123e-01
1.30808043e+00 3.60293865e-01 5.80838084e-01 2.90152162e-01
4.15319651e-01 -1.51930845e+00 1.50251001e-01 1.89033478e-01
6.89410210e-01 -1.58231986e+00 3.43851075e-02 -1.07900655e+00
-7.29941010e-01 1.12290084e+00 7.65698969e-01 1.21117085e-01
8.35364461e-01 -1.73172265e-01 8.41399357e-02 -2.66098857e-01
-2.65404224e-01 -7.36968935e-01 6.38764679e-01 5.95847905e-01
1.66036457e-01 -1.60139173e-01 -7.35447824e-01 6.63571000e-01
3.80028933e-01 -7.20819235e-02 2.42929965e-01 6.60448611e-01
-7.23003805e-01 -1.06602263e+00 -5.40751100e-01 6.47018850e-01
-5.04815161e-01 -1.11464486e-01 -2.36321762e-01 6.77801490e-01
6.61976099e-01 1.23237789e+00 -3.68385343e-03 -3.17228287e-01
3.14475417e-01 8.14767331e-02 1.78598493e-01 -8.27717602e-01
-4.16005731e-01 1.82645604e-01 -5.07347733e-02 -3.95196021e-01
-9.83223200e-01 -4.35726523e-01 -1.04856777e+00 3.27749699e-01
-6.85275555e-01 2.95419276e-01 8.60551655e-01 9.80239928e-01
1.53569832e-01 3.61209661e-01 5.73256731e-01 -7.89997101e-01
-5.37660837e-01 -9.03328419e-01 -5.22108138e-01 1.01844919e+00
-7.00404048e-02 -7.07838297e-01 -6.11877322e-01 5.81041932e-01] | [9.463478088378906, 2.684633731842041] |
51167a46-3b67-451e-9e7f-250454dfd20b | abdomenct-1k-is-abdominal-organ-segmentation | 2010.14808 | null | https://arxiv.org/abs/2010.14808v2 | https://arxiv.org/pdf/2010.14808v2.pdf | AbdomenCT-1K: Is Abdominal Organ Segmentation A Solved Problem? | With the unprecedented developments in deep learning, automatic segmentation of main abdominal organs seems to be a solved problem as state-of-the-art (SOTA) methods have achieved comparable results with inter-rater variability on many benchmark datasets. However, most of the existing abdominal datasets only contain single-center, single-phase, single-vendor, or single-disease cases, and it is unclear whether the excellent performance can generalize on diverse datasets. This paper presents a large and diverse abdominal CT organ segmentation dataset, termed AbdomenCT-1K, with more than 1000 (1K) CT scans from 12 medical centers, including multi-phase, multi-vendor, and multi-disease cases. Furthermore, we conduct a large-scale study for liver, kidney, spleen, and pancreas segmentation and reveal the unsolved segmentation problems of the SOTA methods, such as the limited generalization ability on distinct medical centers, phases, and unseen diseases. To advance the unsolved problems, we further build four organ segmentation benchmarks for fully supervised, semi-supervised, weakly supervised, and continual learning, which are currently challenging and active research topics. Accordingly, we develop a simple and effective method for each benchmark, which can be used as out-of-the-box methods and strong baselines. We believe the AbdomenCT-1K dataset will promote future in-depth research towards clinical applicable abdominal organ segmentation methods. The datasets, codes, and trained models are publicly available at https://github.com/JunMa11/AbdomenCT-1K. | ['Yuhui Li', 'Yunpeng Wang', 'Shangqing Liu', 'Qi Zhang', 'Shucheng Cao', 'Xin Liu', 'Qiyuan Wang', 'Congcong Wang', 'Xingle An', 'Yichi Zhang', 'Cheng Ge', 'Xiaoping Yang', 'Jian He', 'Cheng Zhu', 'Song Gu', 'Yao Zhang', 'Jun Ma'] | 2020-10-28 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [-1.66006386e-01 8.62131303e-04 -6.71565413e-01 -3.80867511e-01
-9.53097582e-01 -7.29015529e-01 9.63416398e-02 4.61222142e-01
-2.73864925e-01 5.47727764e-01 1.89686641e-01 -5.12873054e-01
-7.59435296e-02 -4.01809990e-01 -5.85473716e-01 -8.13471735e-01
-3.23930949e-01 6.96719885e-01 1.46866128e-01 2.05282852e-01
-3.60874474e-01 2.23384857e-01 -2.53343284e-01 2.38636374e-01
8.74557972e-01 8.32118213e-01 -1.04151711e-01 5.92851698e-01
-6.26267642e-02 6.07382774e-01 -2.10617408e-01 -4.07761008e-01
4.29752856e-01 -6.65633619e-01 -9.15235043e-01 -5.19993762e-03
2.85289437e-01 -3.31317306e-01 -2.88499951e-01 7.96777666e-01
7.17534065e-01 -4.53290939e-01 5.04855156e-01 -9.67831612e-01
-5.74968457e-01 8.66256356e-01 -7.06575096e-01 2.64155298e-01
4.79899161e-02 3.58498156e-01 6.15447760e-01 -4.38904613e-01
3.50829929e-01 4.98573065e-01 1.02546322e+00 5.44583380e-01
-1.04033506e+00 -7.23321855e-01 4.74153161e-02 -5.78964233e-01
-1.15420187e+00 1.20012634e-01 2.91754812e-01 -4.45713162e-01
2.04961628e-01 3.71713221e-01 8.11031103e-01 9.77468252e-01
2.85310239e-01 1.05846250e+00 9.96326685e-01 -6.21166229e-02
-5.86213171e-02 -7.57418126e-02 2.05379367e-01 1.02900815e+00
6.19351983e-01 -1.66238815e-01 2.06682429e-01 -2.68345773e-01
1.01850188e+00 2.89700836e-01 -5.52106202e-01 -5.58104217e-01
-1.83873701e+00 7.39048421e-01 7.39757419e-01 3.27540189e-01
-3.74069512e-01 6.92346171e-02 7.44050622e-01 1.10358037e-01
2.28580981e-01 4.24748600e-01 -6.73400998e-01 1.77868262e-01
-8.60104084e-01 -1.07695810e-01 9.77099895e-01 1.09840024e+00
2.20662430e-01 -2.39886060e-01 -2.96271801e-01 6.80998147e-01
3.16999227e-01 3.04852873e-01 7.78657615e-01 -5.15077889e-01
5.00546634e-01 5.59721291e-01 -3.34159285e-01 -4.02944684e-01
-9.27694142e-01 -6.59663498e-01 -1.22710204e+00 -4.62109357e-01
6.17760122e-01 -4.72171724e-01 -1.26274180e+00 1.47834384e+00
5.30959010e-01 2.20824286e-01 1.46461613e-02 1.26770651e+00
1.45392835e+00 3.63897443e-01 3.57444763e-01 -1.64081812e-01
1.60685265e+00 -1.31214046e+00 -3.71302634e-01 6.93930387e-02
1.04400384e+00 -7.52574921e-01 8.00040066e-01 2.58767515e-01
-1.03173459e+00 -2.01951101e-01 -5.92114091e-01 1.89193338e-01
-9.42807794e-02 3.15332621e-01 1.08723259e+00 6.47266209e-01
-6.98054850e-01 2.80334264e-01 -1.15314186e+00 -5.67747712e-01
6.12960100e-01 4.80627000e-01 -2.78796494e-01 -2.14250073e-01
-9.63025153e-01 6.01469576e-01 3.75551432e-01 3.63980001e-03
-1.03398693e+00 -1.03496242e+00 -8.10604572e-01 -2.14490443e-01
5.26690006e-01 -9.73484039e-01 1.23738289e+00 -6.25180423e-01
-1.29920256e+00 1.18732786e+00 4.58514720e-01 -3.92538786e-01
5.30992448e-01 1.33889452e-01 -2.86971271e-01 2.84955680e-01
1.11514442e-01 5.67487836e-01 2.63935894e-01 -1.07969141e+00
-3.17232966e-01 -3.32449317e-01 -2.24159837e-01 8.00588131e-02
-2.12948974e-02 -1.46742523e-01 -7.70412564e-01 -8.55440915e-01
3.76355290e-01 -1.32880139e+00 -5.84765553e-01 2.09969714e-01
-6.37712479e-01 6.93756044e-02 4.50723618e-02 -5.86328268e-01
1.13210058e+00 -2.21958733e+00 -5.76250814e-02 -5.05291186e-02
2.64649987e-01 1.25951245e-01 3.18282959e-03 -3.52671593e-02
-2.19140321e-01 2.98244208e-01 -2.90429413e-01 -7.61811212e-02
-2.67622113e-01 1.66619241e-01 2.35932410e-01 6.52937412e-01
-2.18380049e-01 1.14069831e+00 -9.58202720e-01 -9.76559937e-01
1.78986862e-01 -7.18773007e-02 -3.72858196e-01 2.20276117e-01
9.34628025e-02 8.21998477e-01 -8.31121266e-01 1.10153806e+00
5.98045468e-01 -8.73440802e-01 1.57219574e-01 -3.59782040e-01
8.16008821e-02 -1.63286984e-01 -6.94659472e-01 2.18166447e+00
-1.21340625e-01 -3.66685875e-02 1.51660040e-01 -9.55152392e-01
4.21687961e-01 5.62685609e-01 9.70387757e-01 -5.91186225e-01
1.51238486e-01 5.07708192e-01 1.51121587e-01 -6.10290825e-01
1.83777418e-02 -2.16848895e-01 -2.20294252e-01 2.27588922e-01
1.48468345e-01 -7.92634338e-02 3.85474682e-01 1.30450964e-01
8.53974223e-01 -2.67866105e-01 3.47487897e-01 -5.10007441e-01
2.11346313e-01 3.64084244e-01 5.65320551e-01 7.46824801e-01
-6.78107202e-01 1.02246916e+00 4.68050867e-01 -6.96997643e-01
-6.87944114e-01 -1.26895654e+00 -6.13781214e-01 8.59571159e-01
4.41484749e-01 -8.13213289e-02 -5.34655929e-01 -1.13884199e+00
1.20093256e-01 7.33881071e-02 -6.60317123e-01 2.71215618e-01
-6.29063129e-01 -1.41642249e+00 8.03747952e-01 6.36393547e-01
3.97738546e-01 -7.37809360e-01 -3.38645548e-01 2.67953247e-01
-3.47378433e-01 -1.21318579e+00 -5.91287613e-01 3.28401327e-01
-1.20930624e+00 -1.37522280e+00 -1.53271937e+00 -9.70499098e-01
7.39232600e-01 7.40052536e-02 1.38326621e+00 1.67570218e-01
-4.58508849e-01 2.77859777e-01 -2.16523677e-01 -2.85604417e-01
-2.53803432e-01 3.06482375e-01 -3.22898388e-01 -4.04541403e-01
-8.19131657e-02 9.22704041e-02 -1.04433978e+00 4.51590478e-01
-7.25535154e-01 1.11680672e-01 8.22296500e-01 8.92697513e-01
8.24528217e-01 -5.26590526e-01 4.37723607e-01 -1.07352006e+00
2.27463007e-01 -8.73004556e-01 -3.97419631e-01 5.03731191e-01
-3.68032455e-01 -3.06390196e-01 4.74866241e-01 -4.27834183e-01
-7.70417571e-01 4.01110984e-02 -1.92469001e-01 -2.81473994e-01
-3.01878780e-01 7.15420604e-01 5.42210817e-01 -1.94037840e-01
5.76853514e-01 2.55701363e-01 -2.11536750e-01 -3.61584365e-01
-2.08436362e-02 2.82459289e-01 4.11806256e-01 -7.60671198e-01
4.37900275e-01 4.98307824e-01 -8.54180101e-03 -4.16572899e-01
-1.00308645e+00 -7.51285076e-01 -5.20379722e-01 1.40974149e-01
1.01654875e+00 -1.22001767e+00 -2.40256906e-01 4.97475386e-01
-6.41106009e-01 -6.93138719e-01 -2.25306526e-01 8.83224845e-01
-1.21971913e-01 5.92429399e-01 -1.33557284e+00 -6.59671947e-02
-7.79589236e-01 -1.51316845e+00 9.62334394e-01 4.83132213e-01
-1.37174785e-01 -1.27396739e+00 1.43504754e-01 4.73095059e-01
4.91969258e-01 5.50188124e-01 8.74825239e-01 -9.33564663e-01
-6.26566350e-01 -2.01566279e-01 -3.39271069e-01 -4.13644537e-02
1.62052870e-01 -1.27771229e-01 -3.00043017e-01 -5.61057627e-01
-3.17008078e-01 -5.58935761e-01 7.14916468e-01 8.17117631e-01
1.60659134e+00 3.05644460e-02 -4.74413604e-01 1.16130185e+00
1.33165407e+00 2.31416095e-02 2.01662645e-01 2.08541676e-01
7.19757020e-01 -2.98067083e-04 4.43595737e-01 4.73341554e-01
6.02529466e-01 1.75858006e-01 4.37034756e-01 -8.32334697e-01
3.58531922e-02 3.27360108e-02 -3.39361995e-01 1.05630612e+00
-1.08901958e-03 -2.69877255e-01 -1.19832230e+00 7.48588622e-01
-1.51137018e+00 -2.77330279e-01 -3.59696209e-01 1.94817698e+00
9.55651641e-01 -7.84138069e-02 1.88400790e-01 -4.84523356e-01
5.22928119e-01 2.04690751e-02 -5.81488073e-01 9.08192247e-02
2.30917871e-01 -3.11788581e-02 6.45348191e-01 -2.51304030e-01
-1.48343027e+00 3.55970472e-01 6.25305843e+00 6.77444160e-01
-1.47413611e+00 5.33584543e-02 1.18040478e+00 1.10924825e-01
8.71160403e-02 -2.56837040e-01 -6.10736966e-01 6.95484877e-01
6.19441509e-01 1.22350588e-01 -7.36208707e-02 9.14157152e-01
-2.54081011e-01 9.19337198e-03 -1.27439368e+00 9.20784414e-01
-5.05046658e-02 -1.22506118e+00 -1.65307954e-01 -1.75580531e-02
8.67522418e-01 5.96337616e-01 2.89056804e-02 5.39471209e-01
1.19930819e-01 -9.58402157e-01 1.19629353e-01 8.31330195e-02
9.37839329e-01 -1.44830614e-01 1.11531866e+00 2.47985646e-01
-1.18348360e+00 2.16752112e-01 -1.04817420e-01 6.10103071e-01
1.09188333e-01 5.69690824e-01 -5.81104577e-01 8.12457025e-01
8.25493395e-01 6.83286190e-01 -5.99240899e-01 1.43203318e+00
9.23069045e-02 8.60052347e-01 -5.40300846e-01 1.89104542e-01
4.24024522e-01 -7.07529262e-02 1.04296811e-01 1.44279850e+00
3.37566435e-01 3.73631924e-01 5.41134834e-01 6.63270414e-01
-3.18251580e-01 3.49996686e-01 -1.05090067e-01 -2.32181512e-02
-2.56140307e-02 1.60618043e+00 -1.25304425e+00 -5.17461538e-01
-6.68491840e-01 6.84724808e-01 -1.80703536e-01 2.30799571e-01
-1.15610802e+00 1.59402683e-01 4.21611145e-02 -2.12850072e-03
-6.11490048e-02 2.17490211e-01 -4.45062011e-01 -1.56289554e+00
-3.28059703e-01 -8.54011655e-01 1.14023423e+00 -3.97060692e-01
-1.60894811e+00 3.75472724e-01 3.87598425e-02 -1.31097603e+00
2.07849011e-01 -5.26124954e-01 -5.48770666e-01 5.49022436e-01
-1.66740537e+00 -1.13494134e+00 -7.60216892e-01 5.35370648e-01
4.47516441e-01 4.65859175e-02 7.66810119e-01 4.47903514e-01
-7.22717881e-01 7.72786200e-01 1.74500436e-01 8.14921260e-01
8.56085837e-01 -1.25877500e+00 3.12981531e-02 2.89948791e-01
-2.62080461e-01 5.70246816e-01 1.06648281e-01 -4.20523018e-01
-1.45742166e+00 -1.19935203e+00 2.44951665e-01 -4.14419532e-01
6.16979659e-01 3.16170976e-02 -7.68976688e-01 8.63052070e-01
2.08952352e-01 7.11443841e-01 1.12918031e+00 -3.17647606e-02
1.45247802e-01 7.01492056e-02 -1.27362275e+00 4.06828880e-01
6.58284962e-01 1.44984677e-01 -4.39367920e-01 6.36205018e-01
6.75358236e-01 -1.20614719e+00 -1.44046736e+00 8.39915514e-01
4.63707894e-01 -7.07281411e-01 1.12537873e+00 -3.76392007e-01
5.99386334e-01 5.77833839e-02 2.57113010e-01 -1.12634158e+00
-1.60650790e-01 -3.02973717e-01 1.86300743e-02 8.08117628e-01
5.32589078e-01 -6.40910208e-01 9.60560143e-01 6.61003947e-01
-5.05949199e-01 -1.21320939e+00 -5.84690750e-01 -5.30273736e-01
5.11297762e-01 1.59760430e-01 4.49145049e-01 1.42182600e+00
-1.76520981e-02 1.41256105e-03 -8.55767727e-02 5.20928539e-02
7.72707999e-01 6.53779387e-01 6.99428141e-01 -8.83086026e-01
-3.87804896e-01 -4.06122565e-01 3.44188623e-02 -1.08738697e+00
-4.63489950e-01 -1.05873334e+00 -1.18504636e-01 -1.68696451e+00
7.49994159e-01 -7.73510218e-01 -5.43009698e-01 6.18583560e-01
-4.22544360e-01 2.75564611e-01 1.07549436e-01 3.54633957e-01
-9.71759140e-01 2.03294471e-01 1.81082094e+00 -5.42077608e-02
-7.12193409e-03 1.82885990e-01 -5.96998930e-01 7.30834663e-01
7.62001216e-01 -5.35540700e-01 -1.03772178e-01 -4.15811270e-01
-3.87805581e-01 5.66330612e-01 2.55840629e-01 -8.80387127e-01
2.28645518e-01 -1.99944541e-01 6.57899320e-01 -6.90465689e-01
-1.18573450e-01 -7.76113510e-01 5.97970597e-02 9.92400348e-01
-2.96309739e-01 2.91169956e-02 2.93529540e-01 2.88270622e-01
-3.06317300e-01 -8.94295871e-02 9.68041182e-01 -7.58630633e-01
-5.02428532e-01 8.84172559e-01 8.78306925e-02 6.67771935e-01
1.18675745e+00 -3.57425585e-02 -2.18025297e-01 2.44710278e-02
-7.50458539e-01 7.98462093e-01 2.64145136e-01 6.96328580e-02
2.36712977e-01 -9.73430812e-01 -9.42866683e-01 -1.38164470e-02
2.13456735e-01 7.12699056e-01 4.78527248e-01 1.41932678e+00
-1.03522694e+00 5.56301117e-01 9.39289387e-03 -1.07010853e+00
-9.31644738e-01 6.07813060e-01 5.20873368e-01 -6.82048678e-01
-7.73294628e-01 8.13684821e-01 5.41996419e-01 -7.99728215e-01
1.90904960e-01 -8.16375554e-01 2.54839540e-01 -2.70746529e-01
-6.08807765e-02 2.98564956e-02 -5.24559133e-02 -2.29581103e-01
-3.98220569e-01 5.15194893e-01 -1.32336125e-01 7.34884381e-01
1.15728521e+00 2.92907096e-02 8.86212736e-02 1.56832933e-01
1.12843919e+00 -2.48865038e-01 -8.96192729e-01 -2.56675184e-01
-1.75274372e-01 -1.42076671e-01 -1.80354834e-01 -1.01361716e+00
-1.43847287e+00 7.31380761e-01 6.63052559e-01 4.05862508e-03
1.17317963e+00 1.43395916e-01 1.09530354e+00 1.23937172e-03
3.10283452e-01 -5.88289201e-01 3.77441011e-02 2.48734772e-01
4.97923702e-01 -1.78706729e+00 3.13114911e-01 -5.06768286e-01
-8.03775609e-01 1.02959061e+00 7.37120152e-01 5.34776822e-02
7.21005619e-01 4.32381392e-01 4.31402057e-01 -2.56749392e-01
-3.74387115e-01 8.13057795e-02 2.24283308e-01 2.71821674e-02
8.10788155e-01 2.94169128e-01 -3.65892708e-01 7.98540413e-01
1.64525032e-01 9.57196578e-02 2.56294638e-01 8.84997308e-01
5.96056180e-03 -7.74135828e-01 -2.48523667e-01 5.89431882e-01
-9.89239395e-01 -1.37751713e-01 4.10019420e-02 1.21584260e+00
-9.91786942e-02 2.88190722e-01 -3.50906909e-01 2.63355374e-01
2.59750456e-01 -2.31868163e-01 4.64982480e-01 -5.66521108e-01
-9.09129202e-01 2.60365516e-01 -2.70918071e-01 -4.05190438e-01
-2.83202052e-01 -4.98952329e-01 -1.51637387e+00 -9.62147489e-02
-3.91310900e-01 1.50700241e-01 5.19998074e-01 6.60356998e-01
4.00827043e-02 6.18575633e-01 4.49114203e-01 -5.51390648e-01
-7.87341774e-01 -8.07590306e-01 -3.67151678e-01 6.09353960e-01
2.70980835e-01 -1.81431159e-01 -2.57421821e-01 -9.57346335e-02] | [14.674345970153809, -2.4450056552886963] |
3737eb6e-4cc9-42e4-8c5c-59c0b49652d7 | docred-a-large-scale-document-level-relation | 1906.06127 | null | https://arxiv.org/abs/1906.06127v3 | https://arxiv.org/pdf/1906.06127v3.pdf | DocRED: A Large-Scale Document-Level Relation Extraction Dataset | Multiple entities in a document generally exhibit complex inter-sentence relations, and cannot be well handled by existing relation extraction (RE) methods that typically focus on extracting intra-sentence relations for single entity pairs. In order to accelerate the research on document-level RE, we introduce DocRED, a new dataset constructed from Wikipedia and Wikidata with three features: (1) DocRED annotates both named entities and relations, and is the largest human-annotated dataset for document-level RE from plain text; (2) DocRED requires reading multiple sentences in a document to extract entities and infer their relations by synthesizing all information of the document; (3) along with the human-annotated data, we also offer large-scale distantly supervised data, which enables DocRED to be adopted for both supervised and weakly supervised scenarios. In order to verify the challenges of document-level RE, we implement recent state-of-the-art methods for RE and conduct a thorough evaluation of these methods on DocRED. Empirical results show that DocRED is challenging for existing RE methods, which indicates that document-level RE remains an open problem and requires further efforts. Based on the detailed analysis on the experiments, we discuss multiple promising directions for future research. | ['Zheng-Hao Liu', 'Jie zhou', 'Yuan Yao', 'Deming Ye', 'Peng Li', 'Yankai Lin', 'Maosong Sun', 'Zhiyuan Liu', 'Xu Han', 'Lixin Huang'] | 2019-06-14 | docred-a-large-scale-document-level-relation-1 | https://aclanthology.org/P19-1074 | https://aclanthology.org/P19-1074.pdf | acl-2019-7 | ['document-level-relation-extraction'] | ['natural-language-processing'] | [-1.56081870e-01 3.61925870e-01 -4.29730564e-01 -3.65837932e-01
-6.12460196e-01 -6.39010131e-01 7.38004506e-01 3.99413347e-01
-3.62743020e-01 9.59820688e-01 6.26461685e-01 -1.50756478e-01
-2.84952730e-01 -7.83645570e-01 -3.99664074e-01 -1.52714849e-01
-1.62580788e-01 4.89350766e-01 4.08179134e-01 -5.24059713e-01
-4.97211050e-03 1.56017616e-01 -1.15516448e+00 4.44853038e-01
1.08468318e+00 7.74220407e-01 -3.34528759e-02 3.81542981e-01
-3.01127553e-01 1.06510663e+00 -6.61505699e-01 -1.01244199e+00
2.14706417e-02 -3.41696918e-01 -1.43021166e+00 -3.37646045e-02
1.99411750e-01 -1.01720132e-01 -4.61138785e-01 9.57776845e-01
4.22116727e-01 -1.67065226e-02 6.49168968e-01 -1.06598651e+00
-1.01527476e+00 1.29683912e+00 -4.02006149e-01 2.77218848e-01
6.72683954e-01 -5.60243845e-01 1.59540129e+00 -1.07625794e+00
9.40407753e-01 1.06189835e+00 6.03761554e-01 1.42124489e-01
-6.15746856e-01 -5.36370158e-01 3.48618418e-01 1.20365627e-01
-1.62220466e+00 -4.65796918e-01 5.64741492e-01 -1.83003366e-01
1.45404148e+00 3.73964041e-01 2.95554221e-01 8.99802089e-01
-1.71746567e-01 8.93150568e-01 8.96523714e-01 -5.68789840e-01
-3.12500000e-01 1.32676646e-01 4.95248854e-01 4.40236181e-01
5.17870009e-01 -5.94246030e-01 -3.84169996e-01 -4.13981415e-02
2.38637000e-01 -2.37129182e-01 -4.68525618e-01 -3.48928804e-03
-1.41653001e+00 3.52153569e-01 4.04106945e-01 4.95649159e-01
-1.36267453e-01 -5.63487232e-01 6.22599900e-01 1.74244985e-01
6.66781604e-01 7.44877636e-01 -8.11702192e-01 -1.16456904e-01
-6.08193159e-01 2.44672716e-01 1.11518204e+00 1.52437901e+00
6.05196416e-01 -7.59504795e-01 -2.81682670e-01 9.47744548e-01
2.14550629e-01 3.10999662e-01 4.45045412e-01 -3.26934814e-01
1.24452519e+00 1.15004408e+00 1.29518777e-01 -1.31562757e+00
-4.28058207e-01 -4.01741803e-01 -9.77292299e-01 -8.97089899e-01
-1.17379382e-01 -3.92606556e-01 -3.43884349e-01 1.20543051e+00
2.65241355e-01 -4.97002825e-02 6.20600641e-01 5.73417187e-01
1.74985337e+00 4.63321626e-01 -1.87923387e-01 -4.32513505e-01
1.56326056e+00 -1.31411695e+00 -1.07313168e+00 -3.44066918e-01
1.05685782e+00 -6.66218698e-01 7.05567718e-01 -1.07442560e-02
-7.31249332e-01 -2.31826872e-01 -9.92343128e-01 -3.53470147e-01
-7.31962919e-01 6.17398441e-01 7.54918396e-01 6.75505251e-02
-5.66394150e-01 3.34974706e-01 -7.00066924e-01 -3.94239813e-01
1.66193530e-01 4.11354080e-02 -7.45323896e-01 -1.30448475e-01
-1.77156365e+00 1.10404217e+00 8.02524745e-01 5.21984816e-01
-1.88516617e-01 -3.89504552e-01 -1.22307873e+00 8.56978372e-02
8.07172000e-01 -5.74183583e-01 9.90055263e-01 -7.86100477e-02
-1.01913166e+00 9.35396016e-01 -3.23873729e-01 -1.83138102e-01
1.14300705e-01 -5.18972218e-01 -9.30035233e-01 -5.28450087e-02
2.51839578e-01 -1.02024898e-02 -5.65130934e-02 -1.21730602e+00
-6.46091402e-01 -4.05544311e-01 2.91602731e-01 3.33869040e-01
-5.23000479e-01 5.14687359e-01 -8.23390484e-01 -7.29849815e-01
9.72783491e-02 -7.91450620e-01 -8.16918095e-04 -8.11793745e-01
-1.20196998e+00 -7.38540053e-01 6.44532025e-01 -6.67630792e-01
1.93462825e+00 -1.86372936e+00 2.08272785e-01 -1.36182398e-01
5.17680347e-01 5.37552953e-01 -1.34002781e-02 1.01776302e+00
-1.74090698e-01 4.06940877e-01 -2.20146388e-01 -3.86672378e-01
-3.31808589e-02 1.37430891e-01 -3.73128206e-01 -7.08599538e-02
2.40820721e-01 1.19086778e+00 -1.10437751e+00 -8.09802830e-01
-4.88625407e-01 8.37698132e-02 -2.25345488e-03 4.40950722e-01
8.26687887e-02 -8.43511708e-03 -7.50983119e-01 6.26101971e-01
6.65797651e-01 -3.77287686e-01 4.96258616e-01 -5.50242841e-01
1.63187176e-01 8.98001373e-01 -1.33068049e+00 1.32529950e+00
-2.16699809e-01 4.32042986e-01 -3.94020796e-01 -8.78541589e-01
9.92499888e-01 3.35864663e-01 3.07991147e-01 -3.65441412e-01
-1.43841639e-01 3.00332874e-01 -2.20518604e-01 -9.71034944e-01
8.15512657e-01 2.23413199e-01 -3.51115048e-01 3.19445968e-01
1.42731205e-01 -5.18043600e-02 7.78459132e-01 7.83921957e-01
1.32962859e+00 2.25441884e-02 5.85790098e-01 1.59607809e-02
7.26583421e-01 -2.83107776e-02 7.09351540e-01 3.95860285e-01
1.27902985e-01 3.61210763e-01 6.05592191e-01 -2.36997828e-02
-5.06829858e-01 -6.76015079e-01 -1.95077777e-01 7.48037279e-01
3.50848794e-01 -1.30080926e+00 -4.22677726e-01 -1.32109129e+00
-1.40608922e-01 5.01652181e-01 -5.01811147e-01 1.86169684e-01
-6.40348256e-01 -8.79834473e-01 7.51425505e-01 6.52234614e-01
7.52563417e-01 -8.97937834e-01 3.79829973e-01 1.14087448e-01
-5.73915839e-01 -1.77045941e+00 -3.38973939e-01 1.48579627e-01
-4.59383249e-01 -1.32335865e+00 -4.00874823e-01 -1.04869342e+00
8.23620260e-01 3.40130180e-01 1.55902004e+00 1.68469831e-01
2.43286103e-01 3.47153246e-02 -9.01830375e-01 -1.95673138e-01
-2.08235815e-01 6.13359213e-01 2.00651009e-02 -5.24709821e-01
7.60074914e-01 -4.71267790e-01 -2.13324174e-01 4.04590964e-01
-7.02590585e-01 1.12574250e-01 5.73350549e-01 7.76899159e-01
3.54597688e-01 4.97583717e-01 7.02189565e-01 -1.56922114e+00
1.05668545e+00 -6.10019147e-01 -1.53740659e-01 8.29401195e-01
-6.91973448e-01 7.18653668e-03 7.03669846e-01 -4.55412306e-02
-1.20604944e+00 -3.63609135e-01 -8.13530572e-03 4.24107492e-01
1.21706277e-01 1.21212411e+00 -3.19263339e-01 5.73652864e-01
4.60848719e-01 6.49129674e-02 -6.21760845e-01 -5.47621846e-01
5.48627555e-01 1.19924343e+00 6.64375126e-01 -6.88912451e-01
9.24790502e-01 -5.86392451e-03 -3.09430331e-01 -4.05777156e-01
-1.77985847e+00 -6.86401844e-01 -1.09556770e+00 2.84646273e-01
6.12050354e-01 -1.20294547e+00 -3.66064101e-01 3.21975976e-01
-1.33158410e+00 8.75912309e-02 -6.73127640e-03 3.99498880e-01
2.09443986e-01 4.98713881e-01 -9.23582852e-01 -4.53779012e-01
-4.71329093e-01 -6.94182158e-01 1.12849975e+00 2.98540026e-01
-2.21476153e-01 -1.16457641e+00 2.21520379e-01 6.54736578e-01
-3.87713872e-02 2.86222156e-03 6.94578886e-01 -1.01891828e+00
-2.74102479e-01 -3.03100526e-01 -4.85459119e-01 3.28392446e-01
5.92207849e-01 2.25706398e-01 -5.22079706e-01 2.41253227e-02
-5.31708419e-01 -6.32141709e-01 7.27271140e-01 -5.25664926e-01
8.96930456e-01 -4.81284589e-01 -6.49952233e-01 1.52284339e-01
9.07551765e-01 -2.70827830e-01 6.19252622e-01 4.79027838e-01
9.72295344e-01 7.13339806e-01 1.05197346e+00 3.33365798e-01
1.18189204e+00 6.37815952e-01 1.68979131e-02 -1.99601397e-01
-1.20827116e-01 -4.50313151e-01 1.29882306e-01 1.58260214e+00
-1.84468895e-01 -4.98777986e-01 -8.76534224e-01 6.92373574e-01
-2.04220295e+00 -9.49807644e-01 -4.73056823e-01 1.48549128e+00
1.47698474e+00 1.45396575e-01 -9.40007865e-02 6.67934865e-02
7.66562939e-01 1.95096478e-01 -5.94227277e-02 -3.61368731e-02
-4.81097966e-01 -6.06243238e-02 1.51593432e-01 1.87079489e-01
-1.26528382e+00 1.22959721e+00 6.19366980e+00 7.90679812e-01
-5.42866647e-01 -1.94763720e-01 1.28422379e-01 4.98994887e-01
-3.87306631e-01 2.71164447e-01 -1.39166963e+00 3.79571617e-01
5.68608940e-01 -4.59752560e-01 -5.97475236e-03 7.33596623e-01
-2.71308124e-01 9.17089283e-02 -1.31881714e+00 7.02821493e-01
2.39314869e-01 -1.27239347e+00 -7.07929730e-02 -1.35636076e-01
8.34439695e-01 4.91587222e-02 -5.45732379e-01 6.82910264e-01
5.29110610e-01 -1.03649163e+00 1.20113686e-01 2.73828059e-01
6.21014059e-01 -5.24339378e-01 1.32997322e+00 5.48089206e-01
-1.39956081e+00 2.37944365e-01 -3.35693747e-01 -5.75390719e-02
2.38849401e-01 1.00388181e+00 -7.49786139e-01 1.24145138e+00
6.72396362e-01 1.32951868e+00 -9.22091722e-01 6.08747244e-01
-8.37208748e-01 5.29128194e-01 -1.99746281e-01 -1.36515081e-01
-4.07469757e-02 -5.28110228e-02 2.22179115e-01 1.72125900e+00
1.31958127e-01 2.56609291e-01 2.16199964e-01 5.48650801e-01
-5.81673920e-01 2.94149190e-01 -5.01431584e-01 -2.41630644e-01
8.65925729e-01 1.60250449e+00 -4.43934023e-01 -3.46708596e-01
-5.69882631e-01 9.15639102e-01 9.14821446e-01 2.52092302e-01
-4.67969954e-01 -9.29639757e-01 2.52965212e-01 -1.49373472e-01
3.75064671e-01 -3.69086325e-01 7.83957317e-02 -1.77419055e+00
5.52071214e-01 -7.63308883e-01 6.20724261e-01 -8.06283474e-01
-1.63457131e+00 9.31484818e-01 1.07473783e-01 -1.36692023e+00
-2.30954945e-01 -3.86715740e-01 -3.23529989e-01 6.09648168e-01
-1.70219839e+00 -1.37785482e+00 -2.74872184e-01 2.97847867e-01
2.08202049e-01 -9.13268477e-02 9.37979758e-01 5.41469514e-01
-1.04193306e+00 6.57475054e-01 -1.84483096e-01 8.02736461e-01
9.64020491e-01 -1.34751463e+00 5.43501496e-01 1.03776062e+00
4.14696217e-01 1.05048883e+00 3.33866715e-01 -8.90189528e-01
-1.24983633e+00 -1.16734493e+00 1.65169024e+00 -8.56662929e-01
1.00399101e+00 -3.97397012e-01 -9.41351056e-01 1.08111572e+00
2.98998445e-01 3.52865487e-01 9.35103834e-01 7.55466282e-01
-4.22330648e-01 -2.07504705e-02 -7.73915648e-01 6.14688635e-01
1.35220551e+00 -6.44029081e-01 -1.07638288e+00 4.88995194e-01
8.71567369e-01 -6.86997890e-01 -1.36597455e+00 7.48001933e-01
1.58324884e-03 -5.53164423e-01 6.68638885e-01 -7.53401935e-01
7.08715737e-01 -4.58373487e-01 4.24456671e-02 -1.38017726e+00
-1.55576065e-01 -4.93258446e-01 -7.76859105e-01 2.13584042e+00
9.20728922e-01 -4.90090013e-01 3.74259084e-01 7.46178508e-01
-3.37962918e-02 -1.08448911e+00 -4.06691909e-01 -8.16347599e-01
-1.28284812e-01 -2.92555124e-01 6.60206139e-01 1.23735726e+00
5.58192730e-01 9.21544075e-01 -3.20289552e-01 3.31575215e-01
-1.05851376e-02 3.60038757e-01 7.58981347e-01 -1.01094985e+00
-1.01764321e-01 6.77095801e-02 -2.87932962e-01 -1.43637204e+00
5.77746332e-01 -1.00063634e+00 -7.24352896e-02 -1.91123343e+00
6.15578711e-01 -7.43179500e-01 -6.14172630e-02 6.24089897e-01
-6.38282955e-01 1.45707012e-03 -1.14870675e-01 4.49776858e-01
-1.09143233e+00 7.56304264e-01 1.31333852e+00 -1.70576312e-02
-1.31479409e-02 -1.37846813e-01 -1.14912903e+00 6.89559460e-01
3.01539540e-01 -5.05433381e-01 -3.92712176e-01 -3.77165228e-01
7.18258381e-01 -7.26754367e-02 -2.61855513e-01 -2.73272306e-01
5.46938002e-01 -2.54288882e-01 2.08683070e-02 -6.92004561e-01
-9.76056010e-02 -6.67296469e-01 -4.70162511e-01 -3.01144004e-01
-4.29575354e-01 7.90042877e-02 -2.67332017e-01 3.53479654e-01
-6.94668114e-01 -3.41305822e-01 3.08548957e-02 3.81157696e-02
-5.88728964e-01 1.94021374e-01 9.73603502e-02 7.64485657e-01
8.08960676e-01 3.17965239e-01 -8.37458909e-01 -2.22397849e-01
-4.21246707e-01 6.42952681e-01 7.72517622e-02 6.03452563e-01
3.63882035e-01 -1.39113951e+00 -9.13755894e-01 -4.33702230e-01
5.37393093e-01 5.38399756e-01 -3.70895094e-03 6.73508227e-01
-2.25582093e-01 4.95504797e-01 5.06595969e-01 -1.14728831e-01
-1.45331466e+00 5.33815265e-01 -3.63674422e-04 -8.81957769e-01
-6.73579395e-01 8.65895510e-01 -3.08395386e-01 -7.91905880e-01
-3.16836946e-02 -2.76826710e-01 -7.96847761e-01 1.20476983e-01
5.44663191e-01 1.97686497e-02 3.16656381e-01 -6.13060594e-01
-6.60728872e-01 3.60801578e-01 -5.32325268e-01 3.09916407e-01
1.54853559e+00 -4.19420391e-01 -7.16475725e-01 4.11370426e-01
1.08933270e+00 6.68883801e-01 -5.20379126e-01 -6.68025553e-01
4.37283605e-01 -2.14995578e-01 -2.04843834e-01 -9.06772316e-01
-9.29579973e-01 4.00860816e-01 -6.43797576e-01 4.47280586e-01
9.89463687e-01 3.99132699e-01 1.01501834e+00 8.42049837e-01
2.87111729e-01 -1.11956978e+00 -1.57193273e-01 8.90543163e-01
9.37540770e-01 -1.33906651e+00 6.48385584e-01 -1.21132779e+00
-8.74306262e-01 1.06591225e+00 9.31145787e-01 3.17636430e-01
8.42302561e-01 4.68909651e-01 -8.53119045e-02 -3.91895473e-01
-7.68379748e-01 -3.18248302e-01 5.72917461e-01 3.66676182e-01
9.35739160e-01 -1.26483232e-01 -5.27329445e-01 1.26171732e+00
-4.07093197e-01 -6.88516200e-02 4.78967786e-01 1.00341654e+00
1.18543550e-01 -1.40042722e+00 3.46813738e-01 3.72929603e-01
-3.76892686e-01 -4.33736742e-01 -8.03173065e-01 6.33078516e-01
9.05094892e-02 1.08910990e+00 -4.02245134e-01 -3.98750544e-01
4.98161495e-01 -2.59222358e-01 2.20918149e-01 -1.05903769e+00
-5.15824974e-01 -6.31629109e-01 9.22226071e-01 -1.00319788e-01
-7.49369681e-01 -4.46785837e-01 -1.45984566e+00 -3.23422328e-02
-7.68741310e-01 4.74102199e-01 2.34199032e-01 1.36001706e+00
6.31191790e-01 5.90719759e-01 6.64211810e-01 -1.56145066e-01
-2.55595863e-01 -1.16780376e+00 -5.29395401e-01 5.09726286e-01
-6.67130128e-02 -6.25792980e-01 -4.47678447e-01 -8.54702070e-02] | [9.312660217285156, 8.682700157165527] |
b86da363-d8b8-4229-83d4-67bf3a3a8767 | measuring-few-shot-extrapolation-with-program | null | null | https://openreview.net/forum?id=UZTzGNV_64a | https://openreview.net/pdf?id=UZTzGNV_64a | Measuring few-shot extrapolation with program induction | Neural networks are capable of learning complex functions, but still have problems generalizing from few examples and beyond their training distribution. Meta-learning provides a paradigm to train networks to learn from few examples, but it has been shown that its most popular benchmarks require very limited generalization capabilities. Program induction lies at the opposite end of the spectrum: programs are capable of extrapolating from very few examples, but we still do not know how to efficiently search for complex programs. We propose a common benchmark for both communities, measuring extrapolation from few examples coming from the execution of small programs. These are obtained by leveraging a C++ interpreter on codes from programming competitions; extracting small sub-codes with their corresponding input-output pairs. Statistical analysis and preliminary human experiments show the potential of this benchmark for enabling progress in few-shot extrapolation. | ['Leslie Pack Kaelbling', 'Tomas Perez', 'Joshua B. Tenenbaum', 'Javier Lopez-Contreras', 'Ferran Alet'] | 2020-10-13 | null | null | null | neurips-workshop-cap-2020-12 | ['program-induction'] | ['computer-code'] | [ 1.18570246e-01 -1.02943502e-01 -6.84464872e-01 -4.91132081e-01
-7.45639443e-01 -4.35765982e-01 4.82678771e-01 3.19238901e-01
-5.12066960e-01 4.06563491e-01 2.95469281e-03 -6.14607096e-01
8.44803303e-02 -7.80856133e-01 -1.05185449e+00 -2.44755000e-01
-4.07507390e-01 4.64421839e-01 2.16572806e-01 -4.67274427e-01
3.87254596e-01 1.89187214e-01 -1.86704385e+00 9.87732947e-01
8.63481164e-01 5.25406957e-01 8.80949646e-02 1.11232698e+00
-2.77509481e-01 1.19886410e+00 -6.13262773e-01 -2.43405387e-01
-4.30490188e-02 -3.39862764e-01 -7.65035033e-01 -6.10677183e-01
6.16798699e-01 -2.89607167e-01 -2.71872967e-01 1.14296794e+00
2.33952433e-01 2.12606966e-01 5.51225007e-01 -1.29173398e+00
-8.76668811e-01 9.02125180e-01 -1.85282752e-01 4.79774714e-01
5.96865416e-01 5.23579061e-01 1.09500086e+00 -8.43640745e-01
6.44435048e-01 7.65981436e-01 1.12688768e+00 9.08437371e-01
-1.60441601e+00 -3.12967330e-01 -4.64368045e-01 1.23095356e-01
-9.64671612e-01 -4.42082375e-01 4.87923443e-01 -5.35546899e-01
1.64737737e+00 3.25044245e-01 5.35855055e-01 1.19883192e+00
-1.11769363e-01 1.05252147e+00 1.10223806e+00 -5.67812264e-01
3.23840857e-01 4.46972281e-01 4.94297832e-01 8.97419810e-01
-9.42836255e-02 5.53606212e-01 -3.18101853e-01 -5.37508130e-01
1.34885341e-01 1.67198047e-01 -4.11193013e-01 -3.50497842e-01
-1.05022192e+00 8.98854494e-01 4.26741242e-01 4.33952481e-01
1.82510633e-02 4.41372544e-01 1.06814396e+00 7.12575018e-01
2.43159726e-01 6.08486176e-01 -5.47449291e-01 -5.10423481e-01
-1.12152863e+00 4.21064734e-01 1.25266457e+00 1.46139729e+00
8.03574800e-01 1.37428001e-01 -7.10290149e-02 7.16793656e-01
-4.00852084e-01 -9.67713073e-03 7.66561925e-01 -9.11098659e-01
5.93242466e-01 7.55359650e-01 -2.67031223e-01 -4.57823098e-01
-1.10989340e-01 -1.73113614e-01 -5.01322389e-01 6.00854099e-01
4.77122545e-01 -3.48203510e-01 -6.58145130e-01 1.45947981e+00
-3.15463722e-01 6.34222627e-02 -5.84640317e-02 3.14749718e-01
7.43317366e-01 7.27113843e-01 9.17976052e-02 -1.71472076e-02
7.52541244e-01 -1.16383874e+00 2.55487263e-01 -5.06348491e-01
1.15154898e+00 -1.12462573e-01 1.67415011e+00 5.41572928e-01
-1.07916832e+00 -8.34011078e-01 -1.14420140e+00 1.59945980e-01
-5.21932125e-01 -2.22836301e-01 9.48901355e-01 7.82411039e-01
-1.10549796e+00 1.04706633e+00 -7.28838086e-01 -2.01798350e-01
6.95073307e-01 2.59927660e-01 -1.64740607e-01 -1.54736698e-01
-6.81114435e-01 9.89467442e-01 7.86899388e-01 -4.61883634e-01
-1.12056267e+00 -9.39276040e-01 -7.91612923e-01 3.30459684e-01
1.29405588e-01 -4.81996536e-01 1.44666481e+00 -1.31187546e+00
-1.06776333e+00 1.02700889e+00 9.49726477e-02 -6.42654896e-01
4.86089468e-01 -1.54178981e-02 -4.38219458e-01 -2.20532537e-01
-1.93479776e-01 4.65549618e-01 5.66686690e-01 -8.78571749e-01
-6.83170617e-01 -1.24914192e-01 2.05392912e-01 -1.99957952e-01
-5.62793374e-01 1.39881209e-01 1.22510843e-01 -3.05525273e-01
-4.65741456e-01 -8.81629705e-01 -3.01890105e-01 -9.01532844e-02
-1.00709364e-01 -4.77880001e-01 4.99515086e-01 -5.50516546e-01
1.21297872e+00 -2.14176965e+00 -5.81843406e-03 -9.24848765e-02
1.38008431e-01 3.80989820e-01 -9.25626904e-02 3.19937319e-01
-3.97526026e-01 2.36001045e-01 -2.81008393e-01 2.63872772e-01
1.43240973e-01 -6.03003846e-03 -4.38251704e-01 2.74479508e-01
1.25329301e-01 9.83842790e-01 -1.22051382e+00 -3.45274270e-01
-1.03533119e-01 -4.75133434e-02 -9.22810078e-01 1.20461732e-01
-8.28289807e-01 -1.72182292e-01 -1.46674514e-01 7.55963147e-01
2.30812192e-01 -2.12712094e-01 -2.47256771e-01 3.05348992e-01
1.54091790e-01 4.72142734e-02 -5.72303891e-01 1.89293814e+00
-7.83326447e-01 1.04804814e+00 -4.56995100e-01 -1.33151186e+00
8.82562339e-01 1.23572208e-01 -5.05381376e-02 -2.79038608e-01
-7.27123469e-02 3.19576383e-01 3.32678020e-01 -9.43165302e-01
5.13613701e-01 -7.69721121e-02 -2.78963566e-01 4.69802588e-01
5.30155540e-01 -3.29938471e-01 5.00523090e-01 4.96361442e-02
1.47414637e+00 3.20426285e-01 2.98743069e-01 -2.38379240e-01
3.24049056e-01 7.08627284e-01 1.00591771e-01 1.31474495e+00
-2.15514839e-01 3.90934616e-01 5.21012187e-01 -1.03621924e+00
-1.41690505e+00 -8.91767502e-01 1.08747676e-01 1.84330690e+00
-5.44107974e-01 -6.07080221e-01 -7.64268160e-01 -8.98415208e-01
8.60232487e-02 1.33163273e+00 -6.53598249e-01 -2.88272679e-01
-6.65395975e-01 -6.23537242e-01 7.47877657e-01 7.66237080e-01
1.39797583e-01 -1.23881030e+00 -7.71542668e-01 2.72232890e-01
3.16815585e-01 -6.76113129e-01 -1.63088068e-01 7.58678198e-01
-1.16580331e+00 -1.02009487e+00 -6.64823771e-01 -1.04604185e+00
6.89777136e-01 -2.23159522e-01 1.85547292e+00 5.78047335e-01
-5.16243815e-01 2.55854696e-01 -1.07312679e-01 -5.95814168e-01
-1.15742433e+00 1.71699658e-01 -2.11531594e-01 -1.00788009e+00
7.93601096e-01 -9.42449868e-01 -9.81253237e-02 -1.92206621e-01
-4.80304241e-01 -1.62559509e-01 6.17559850e-01 1.25492597e+00
2.37790793e-02 -9.41149238e-03 4.95764822e-01 -1.38642585e+00
8.53159010e-01 -8.47196281e-01 -6.17685378e-01 3.37609172e-01
-6.15401745e-01 1.27247274e-01 1.34058785e+00 -8.11914980e-01
-9.47915614e-01 1.05844900e-01 -1.67259127e-01 -3.49786222e-01
-5.97414851e-01 5.91564000e-01 3.20563763e-01 -9.08964947e-02
1.68923974e+00 5.71826398e-01 -2.34379530e-01 -1.28026739e-01
5.79726696e-01 5.67411125e-01 6.70007288e-01 -1.15541923e+00
5.13294935e-01 -3.32777156e-03 -3.82947236e-01 -7.45956600e-01
-5.80220759e-01 -3.54094088e-01 -4.28703040e-01 9.50999781e-02
1.11516938e-01 -6.79286778e-01 -6.83277786e-01 6.97845519e-02
-1.19576275e+00 -7.07164764e-01 -4.69559669e-01 9.67545882e-02
-9.04607177e-01 8.40210617e-02 -8.59144211e-01 -4.50622916e-01
-3.45810205e-01 -1.01026750e+00 4.27317351e-01 1.66400939e-01
-5.67653716e-01 -9.55668211e-01 2.13330030e-01 -1.43249780e-01
6.09139919e-01 1.86724752e-01 1.47527194e+00 -1.20931828e+00
-4.30424064e-01 -4.99408573e-01 1.50292695e-01 6.26281857e-01
-4.13480788e-01 1.01383440e-01 -1.00774264e+00 -1.90359429e-01
4.75067906e-02 -8.67380679e-01 9.39057291e-01 -2.20376421e-02
1.44940460e+00 -3.82580280e-01 -3.80818188e-01 6.88670635e-01
1.70107794e+00 2.47914605e-02 5.61909795e-01 2.72277951e-01
4.74673003e-01 3.57706070e-01 2.93422848e-01 1.71874374e-01
-2.11011782e-01 1.19933143e-01 1.27884328e-01 4.72281903e-01
3.65346521e-02 -2.76039809e-01 4.62965280e-01 7.25869179e-01
-1.44381508e-01 3.78571630e-01 -1.41382825e+00 7.04034925e-01
-1.49127090e+00 -1.28068829e+00 3.82238105e-02 2.09543180e+00
1.18248355e+00 7.92958319e-01 1.95111424e-01 -1.55436411e-01
5.17440975e-01 -6.58899173e-02 -7.02472329e-01 -7.13724494e-01
2.89820552e-01 5.30920625e-01 1.83667079e-01 1.43818513e-01
-7.41072893e-01 6.76860511e-01 7.12033129e+00 9.63184834e-01
-8.15813243e-01 1.28155956e-02 7.47800827e-01 -2.12232433e-02
-1.44769162e-01 1.76708251e-01 -8.24023843e-01 4.82055724e-01
1.47136796e+00 -5.32039344e-01 7.53720105e-01 1.82657790e+00
-6.04985476e-01 -1.01930641e-01 -1.82714117e+00 7.65044212e-01
1.36845708e-01 -1.65240693e+00 -2.19599098e-01 -4.34425950e-01
1.09137487e+00 6.01276577e-01 6.43215328e-03 1.28756809e+00
6.68798804e-01 -1.31738341e+00 4.32403624e-01 4.17987674e-01
7.10778058e-01 -8.64292741e-01 3.63341153e-01 1.06451404e+00
-7.03426898e-01 -4.91930932e-01 -5.42460978e-01 -1.62234351e-01
-6.78742290e-01 3.04326683e-01 -1.06599295e+00 -1.93082184e-01
5.93912899e-01 4.67690825e-01 -9.08165872e-01 1.08406746e+00
-1.36090696e-01 6.24967396e-01 -2.58184653e-02 -6.96495295e-01
1.00596629e-01 7.54372030e-02 3.66047055e-01 1.55783832e+00
3.89482826e-01 -1.77510440e-01 3.66522674e-03 1.49986386e+00
-2.04528376e-01 -1.29302010e-01 -1.10626864e+00 -2.78014302e-01
4.49759007e-01 9.73626614e-01 -4.55136865e-01 -7.79676795e-01
-8.22300911e-01 7.16580033e-01 5.85639179e-01 2.36878783e-01
-7.88564324e-01 -8.75498712e-01 9.76660252e-02 -2.96722740e-01
8.67838860e-02 -3.50178080e-03 -2.89639682e-01 -1.17989039e+00
-2.94825854e-03 -1.52731347e+00 3.61655563e-01 -7.69775748e-01
-1.04862058e+00 6.37249410e-01 -1.05581293e-02 -9.38052952e-01
-7.69246161e-01 -7.93922663e-01 -9.06249523e-01 9.04747784e-01
-9.66981232e-01 -7.66415477e-01 -4.27059919e-01 3.18685383e-01
7.57524014e-01 -5.07556498e-01 1.08848679e+00 3.00086420e-02
-1.07480422e-01 6.56986892e-01 2.46021733e-01 3.06699008e-01
2.48972088e-01 -1.49834895e+00 8.68820012e-01 3.69803220e-01
2.88049132e-01 8.07181180e-01 9.37004209e-01 -3.40479434e-01
-1.58788514e+00 -9.34158444e-01 4.90746588e-01 -8.57803643e-01
7.44462907e-01 -2.45998949e-01 -1.25213802e+00 9.55298901e-01
5.89721352e-02 4.15604502e-01 5.31129181e-01 2.98567295e-01
-6.93943799e-01 1.85880542e-01 -9.61046875e-01 6.85671568e-01
9.97124374e-01 -9.45584714e-01 -9.04523849e-01 5.23365676e-01
6.43147230e-01 -4.04544026e-01 -7.97819972e-01 2.50144362e-01
5.08476555e-01 -1.38890135e+00 1.06103408e+00 -1.07972050e+00
1.11505771e+00 3.11746389e-01 -2.21899524e-01 -1.50739288e+00
1.36106595e-01 -4.77996230e-01 -4.69106704e-01 1.15146184e+00
5.36507607e-01 -1.54181316e-01 1.28943634e+00 6.18214071e-01
-2.88142174e-01 -7.86985755e-01 -2.28339493e-01 -1.11724031e+00
5.03219783e-01 -6.25533581e-01 4.50815886e-01 8.87699127e-01
5.64073026e-01 3.32296759e-01 1.11211397e-01 -3.60288233e-01
6.81511402e-01 3.43875498e-01 8.98624003e-01 -1.07094026e+00
-9.20657754e-01 -8.42925191e-01 -2.37579986e-01 -8.11999857e-01
4.76187527e-01 -1.34584415e+00 1.75267801e-01 -9.17806923e-01
5.81320941e-01 -3.07128131e-01 -1.00963093e-01 3.14603508e-01
-7.73219541e-02 -2.72856593e-01 -1.44222602e-01 7.54490048e-02
-5.00412107e-01 3.26745883e-02 7.28931904e-01 -4.20634776e-01
-2.12395489e-01 3.17866504e-01 -5.34036219e-01 1.08481061e+00
8.10679436e-01 -6.99434638e-01 -4.34113234e-01 -1.55244380e-01
4.80065078e-01 3.03238839e-01 3.57098788e-01 -1.39586365e+00
5.22234917e-01 -2.51805812e-01 3.66865516e-01 -3.07525665e-01
-8.57270434e-02 -6.34163558e-01 -1.55974612e-01 7.65260279e-01
-8.28414619e-01 3.10383923e-02 3.57711196e-01 5.59785843e-01
-1.15213748e-02 -1.21669304e+00 8.06698918e-01 -8.64853978e-01
-1.23867619e+00 8.23765174e-02 -1.31422460e-01 6.41494095e-01
8.08707416e-01 -1.22818403e-01 -4.08962399e-01 -1.46341175e-01
-7.71239579e-01 -1.98638454e-01 6.53150916e-01 1.98719546e-01
6.10921443e-01 -1.15119135e+00 -5.78559458e-01 2.12072790e-01
4.82693613e-01 -4.16846812e-01 -1.28721651e-02 5.40712774e-01
-6.22480273e-01 1.18002519e-01 -2.86436468e-01 -6.31198287e-01
-1.19897687e+00 1.03957856e+00 3.49876881e-01 -5.63822016e-02
-5.42860866e-01 9.83441472e-01 -3.71963292e-01 -9.37918961e-01
5.30655742e-01 -4.27689523e-01 2.96904892e-01 -4.30312186e-01
7.60807633e-01 3.13172936e-01 5.20418882e-02 8.67600888e-02
8.23123157e-02 -1.34761363e-01 -1.97059676e-01 2.78350115e-01
1.70476782e+00 8.12037110e-01 -1.68429568e-01 7.23023951e-01
1.51243615e+00 -3.58052433e-01 -1.07032859e+00 -2.92712778e-01
6.63177609e-01 -4.82717454e-01 -4.06669974e-01 -7.14717090e-01
-4.64672744e-01 1.24192953e+00 4.10930783e-01 3.17703158e-01
7.84048378e-01 -1.56198606e-01 5.73335648e-01 1.17774785e+00
5.33140242e-01 -1.14733040e+00 3.54440600e-01 6.95324719e-01
5.63536108e-01 -1.40230525e+00 -5.33552803e-02 1.25129521e-01
-2.28373751e-01 1.60038292e+00 9.01574135e-01 -4.69693303e-01
2.43444800e-01 5.74653685e-01 -5.33953905e-01 -1.78751349e-01
-1.07780409e+00 2.82591254e-01 9.73368902e-03 8.62481415e-01
6.30209625e-01 -9.60248634e-02 3.90524626e-01 4.82252747e-01
-1.57322228e-01 3.05357933e-01 6.64667189e-01 1.10378778e+00
-7.16051340e-01 -7.75886297e-01 -1.72516078e-01 9.61118698e-01
-3.78790557e-01 -4.81828988e-01 7.15979636e-02 9.61095870e-01
-4.10931744e-02 3.40998113e-01 -2.59766310e-01 -5.30351937e-01
3.07911754e-01 4.61273044e-01 7.10973382e-01 -1.03647316e+00
-7.95253932e-01 -7.01671958e-01 2.51051635e-01 -4.65171874e-01
3.60174850e-02 -4.68950748e-01 -8.95828545e-01 -5.42440891e-01
-3.28027695e-01 1.65411785e-01 5.94845355e-01 6.40626252e-01
-2.28476152e-02 4.26783383e-01 4.59953725e-01 -9.62638795e-01
-1.49211478e+00 -8.42513740e-01 -3.47732961e-01 4.07902598e-01
6.01501942e-01 8.84389728e-02 -5.90660870e-01 1.95099100e-01] | [8.181859016418457, 7.513160705566406] |
ed2f7665-beea-4e55-b177-b35e7129700b | augmented-policy-gradient-methods-for | null | null | https://openreview.net/forum?id=S1gN8yrYwB | https://openreview.net/pdf?id=S1gN8yrYwB | AUGMENTED POLICY GRADIENT METHODS FOR EFFICIENT REINFORCEMENT LEARNING | We propose a new mixture of model-based and model-free reinforcement learning
(RL) algorithms that combines the strengths of both RL methods. Our goal is to reduce the sample complexity of model-free approaches utilizing fictitious trajectory
rollouts performed on a learned dynamics model to improve the data efficiency of
policy gradient methods while maintaining the same asymptotic behaviour. We
suggest to use a special type of uncertainty quantification by a stochastic dynamics
model in which the next state prediction is randomly drawn from the distribution
predicted by the dynamics model. As a result, the negative effect of exploiting
erroneously optimistic regions in the dynamics model is addressed by next state
predictions based on an uncertainty aware ensemble of dynamics models. The
influence of the ensemble of dynamics models on the policy update is controlled
by adjusting the number of virtually performed rollouts in the next iteration according to the ratio of the real and virtual total reward. Our approach, which we
call Model-Based Policy Gradient Enrichment (MBPGE), is tested on a collection of benchmark tests including simulated robotic locomotion. We compare our
approach to plain model-free algorithms and a model-based one. Our evaluation
shows that MBPGE leads to higher learning rates in an early training stage and an
improved asymptotic behaviour. | ['Frank Hees', 'Rene Vossen', 'Christoph Henke', 'Gregor Roering', 'Kai Lagemann'] | 2019-09-25 | null | null | null | null | ['policy-gradient-methods'] | ['methodology'] | [-9.15634483e-02 2.41883680e-01 -3.22113484e-01 6.02157414e-02
-3.55958015e-01 -3.88526946e-01 7.74836004e-01 1.58906728e-01
-1.01805747e+00 1.45415318e+00 6.58578891e-03 -8.32672566e-02
-3.02557081e-01 -8.83333504e-01 -8.83358896e-01 -8.29761744e-01
-1.72450274e-01 7.26672411e-01 5.16506135e-01 -4.72562104e-01
5.04143119e-01 4.48479623e-01 -1.56980753e+00 -1.79700747e-01
8.69631648e-01 8.21487486e-01 1.94867596e-01 5.89197934e-01
-5.82491718e-02 7.63269365e-01 -5.41394472e-01 1.57444134e-01
2.84452975e-01 -4.38855380e-01 -3.56793970e-01 -2.26107031e-01
-4.34074730e-01 -2.79546231e-01 1.30412564e-01 8.62502038e-01
5.96274793e-01 5.32500923e-01 5.94919860e-01 -8.76045763e-01
3.06915551e-01 7.69201636e-01 -3.72822434e-01 1.21152818e-01
1.22358561e-01 4.66216177e-01 3.83691192e-01 -1.39511883e-01
8.43074381e-01 1.21195328e+00 3.89433384e-01 6.52676404e-01
-1.35654831e+00 -4.33805376e-01 4.25027698e-01 1.16879374e-01
-9.07465458e-01 -1.86879024e-01 5.81661940e-01 -3.26772302e-01
8.34509790e-01 -7.55230039e-02 8.45990717e-01 8.75441790e-01
4.55557108e-01 6.13756657e-01 1.28230095e+00 -4.54233140e-01
1.07696712e+00 1.86787620e-01 -3.68224531e-01 7.27342367e-01
2.09134012e-01 8.17481816e-01 -3.58408988e-01 -3.28125238e-01
7.02590585e-01 -4.43909913e-01 -1.67445317e-01 -9.28103030e-01
-8.93141985e-01 8.07311952e-01 3.24019939e-01 1.47143230e-01
-6.89149797e-01 4.28736031e-01 4.71620232e-01 2.78120160e-01
3.20601612e-01 8.35483253e-01 -4.74584609e-01 -3.88957083e-01
-8.24704409e-01 7.69688427e-01 8.86353135e-01 2.10988894e-01
6.43998682e-01 2.29760140e-01 -2.98270285e-01 5.17522216e-01
1.73549622e-01 2.78700858e-01 9.23359096e-01 -1.07640064e+00
1.33475408e-01 5.71033239e-01 6.28475547e-01 -4.85440940e-01
-3.63407284e-01 -5.28037727e-01 -2.22829312e-01 8.52199018e-01
4.89827394e-01 -5.40607989e-01 -7.69206226e-01 1.81937039e+00
6.09388113e-01 1.02364406e-01 6.43848628e-02 7.64313996e-01
-2.95575887e-01 5.87951541e-01 8.98755789e-02 -3.60566229e-01
4.81517375e-01 -7.29903698e-01 -5.53325355e-01 1.23679213e-01
7.54553199e-01 -6.28784001e-02 1.00344467e+00 6.32335365e-01
-9.85874534e-01 -3.72131884e-01 -1.14205730e+00 8.03605914e-01
-2.86982298e-01 8.91074985e-02 1.71657279e-01 4.95469630e-01
-8.31932902e-01 1.41222727e+00 -1.06199527e+00 -8.85165930e-02
1.48271695e-01 4.30696070e-01 2.06625015e-02 4.30105269e-01
-1.02149642e+00 1.21869910e+00 7.62673736e-01 -3.15949887e-01
-9.91931081e-01 -4.00241196e-01 -4.79179621e-01 -1.18605129e-01
6.84157908e-01 -4.86058980e-01 1.34196258e+00 -9.91931438e-01
-2.30444574e+00 2.13288933e-01 2.99418807e-01 -1.01219869e+00
1.16337073e+00 -2.57068068e-01 2.75868118e-01 -1.98303536e-01
-2.35279098e-01 6.40027404e-01 9.65406477e-01 -1.47123098e+00
-6.95243716e-01 -1.01137854e-01 -5.83187975e-02 3.77949387e-01
9.14414041e-03 -7.09516048e-01 1.58166453e-01 -2.10701033e-01
-3.14744055e-01 -1.23839045e+00 -5.28174222e-01 -2.59331316e-01
9.92971007e-03 -1.05636470e-01 5.86264133e-01 -2.72266895e-01
1.24901986e+00 -1.63592827e+00 5.78657687e-01 3.99771601e-01
-2.66771674e-01 4.04516608e-01 1.50011942e-01 5.64626575e-01
3.25591534e-01 -6.33674338e-02 -3.52243781e-01 -1.61254629e-01
2.14297287e-02 4.58107978e-01 -5.36635995e-01 2.66571969e-01
-1.45160437e-01 3.76092106e-01 -1.28714049e+00 -2.74702460e-01
4.12110537e-01 -4.96333167e-02 -6.51865423e-01 3.40367109e-01
-7.52461672e-01 6.56100035e-01 -5.75383067e-01 2.06477437e-02
2.76209116e-01 2.08474502e-01 3.69965166e-01 3.64325851e-01
-4.36843753e-01 8.60955864e-02 -1.12962079e+00 1.27753019e+00
-4.56102014e-01 -2.30080411e-02 -2.49231994e-01 -9.61032569e-01
1.04724813e+00 5.93501441e-02 4.56724107e-01 -4.30388153e-01
1.58055380e-01 3.00093293e-01 9.63104367e-02 -3.31310123e-01
3.72745961e-01 -4.93892394e-02 1.67578727e-01 4.06318784e-01
3.30941565e-02 -4.23569649e-01 3.41285586e-01 -1.62405610e-01
1.02977538e+00 1.05775094e+00 5.38620591e-01 -4.14122492e-01
5.03201067e-01 4.05450389e-02 6.16069138e-01 9.48439538e-01
-3.34497482e-01 -2.35747676e-02 7.08736658e-01 -4.09541726e-01
-9.97757435e-01 -8.36734056e-01 2.28492290e-01 9.32990849e-01
3.66138890e-02 -2.36699224e-01 -6.26637042e-01 -8.83998156e-01
2.58688897e-01 1.15941417e+00 -8.69607449e-01 -4.36854243e-01
-7.81508744e-01 -8.04188192e-01 1.42960832e-01 1.87926233e-01
3.95127445e-01 -1.20720422e+00 -1.22002769e+00 5.45374870e-01
1.56790689e-01 -5.73540807e-01 1.51494563e-01 3.54853481e-01
-1.06227887e+00 -8.84416342e-01 -6.79547191e-01 7.36214891e-02
2.71697700e-01 -6.79925799e-01 8.44749212e-01 -1.67364880e-01
3.33524436e-01 3.49372894e-01 -4.68098462e-01 -4.57049221e-01
-7.86830544e-01 -2.94648726e-02 2.78739989e-01 -2.91719973e-01
-2.11404175e-01 -5.86097121e-01 -4.08346325e-01 1.76089615e-01
-6.10336185e-01 -1.03033729e-01 4.00985062e-01 1.03533792e+00
5.81684411e-01 -2.60753721e-01 7.69286871e-01 -5.57224870e-01
8.48925352e-01 -5.08191586e-01 -9.66921926e-01 1.97776332e-01
-8.87836754e-01 8.55143249e-01 8.61001849e-01 -7.87242591e-01
-1.08784533e+00 5.36720715e-02 -3.87953632e-02 -3.87649179e-01
4.50644232e-02 2.48243436e-01 4.00643498e-01 -7.17298314e-02
8.19954813e-01 3.26478660e-01 3.87071759e-01 -3.41544062e-01
3.10400724e-01 7.89050013e-02 9.25966650e-02 -9.12301838e-01
1.76160425e-01 2.85452425e-01 2.33133644e-01 -4.14453536e-01
-2.95960099e-01 5.27263917e-02 -2.45430976e-01 -6.86567128e-01
2.99076647e-01 -2.41415232e-01 -7.94338346e-01 3.93965304e-01
-6.98413253e-01 -8.85175228e-01 -8.44810545e-01 6.45048320e-01
-1.17456877e+00 2.68818378e-01 -1.84365466e-01 -1.22689009e+00
-1.92613468e-01 -1.10873187e+00 5.98172307e-01 3.24370414e-01
-5.46771847e-02 -8.19622993e-01 7.26362944e-01 -5.05669117e-01
4.15931851e-01 5.01793146e-01 8.09670091e-01 -5.75123847e-01
-2.08463773e-01 -1.59055591e-01 4.25302595e-01 3.52828175e-01
-2.27100372e-01 1.03973644e-02 -6.32146597e-01 -2.52370417e-01
-7.87121877e-02 -5.14777660e-01 9.04621959e-01 4.68594372e-01
7.18120754e-01 -3.39007914e-01 -4.92815226e-01 1.85077954e-02
1.46179533e+00 3.82743090e-01 4.78550434e-01 6.22473419e-01
2.26564761e-02 5.73433042e-01 9.40315425e-01 9.29679573e-01
-2.49078348e-02 8.30315948e-01 6.85285568e-01 6.67295277e-01
2.57197976e-01 -4.15225327e-01 5.44846594e-01 7.20480084e-02
-2.22853690e-01 -1.89004973e-01 -6.69102132e-01 4.17586774e-01
-2.27252460e+00 -8.61049294e-01 5.12744188e-01 2.68350744e+00
8.16112399e-01 4.53079730e-01 5.55306554e-01 8.41305032e-02
4.71468300e-01 -9.30018574e-02 -7.61628449e-01 -6.28930509e-01
3.48894417e-01 -6.29954860e-02 6.13260150e-01 6.81903064e-01
-7.70231187e-01 8.95244300e-01 5.78518152e+00 9.59195435e-01
-1.10367596e+00 -2.35609382e-01 3.68400455e-01 -2.67265558e-01
2.65528001e-02 8.70306119e-02 -7.66146541e-01 7.80727863e-01
1.07295632e+00 -3.23547482e-01 6.24171674e-01 1.04072917e+00
5.00741780e-01 -6.64745927e-01 -8.05790961e-01 2.60668725e-01
-5.08894682e-01 -1.24030745e+00 -3.99168432e-02 8.94913375e-02
5.81553519e-01 -2.53525414e-02 -1.69835106e-01 6.20481193e-01
6.31246626e-01 -4.74078417e-01 9.22973871e-01 8.92068863e-01
2.61972636e-01 -9.74221230e-01 7.64153004e-01 7.94420183e-01
-5.81318736e-01 -4.78905797e-01 -2.32876316e-01 -2.53553361e-01
1.33455053e-01 5.01002908e-01 -9.57231820e-01 5.02419055e-01
2.95247018e-01 2.20337257e-01 -4.76216711e-02 1.13692927e+00
-9.16981995e-02 6.22861385e-01 -6.25974059e-01 -6.82995319e-01
5.42143464e-01 -3.36108834e-01 9.80845153e-01 8.25021505e-01
1.61085337e-01 -3.07794750e-01 1.51489571e-01 7.68149436e-01
4.67326164e-01 8.20054933e-02 -7.51182377e-01 1.38116539e-01
3.69719595e-01 8.38470817e-01 -7.85596550e-01 -3.60601604e-01
5.61468005e-01 5.73078811e-01 5.71327150e-01 1.32944118e-02
-7.27234066e-01 2.36568805e-02 2.39834085e-01 -6.95100874e-02
3.17160577e-01 -1.36384845e-01 1.00727767e-01 -8.61656010e-01
-1.78777367e-01 -5.35077989e-01 1.78473592e-01 -3.55103135e-01
-8.35500300e-01 3.93389374e-01 4.60983634e-01 -1.15976691e+00
-1.06767607e+00 -2.99970597e-01 -5.29921472e-01 5.69449306e-01
-1.21301389e+00 -2.61792988e-01 1.69540092e-01 4.35499996e-02
3.04322600e-01 -1.59348413e-01 7.66987264e-01 -3.35974187e-01
-2.28267819e-01 1.46122009e-01 4.91823733e-01 -5.57717443e-01
3.84774297e-01 -1.44508445e+00 -8.72373953e-02 3.85288775e-01
-4.60540920e-01 1.39159605e-01 1.20016456e+00 -8.46130073e-01
-7.10276723e-01 -8.29764068e-01 3.30145121e-01 -1.52445450e-01
4.36949819e-01 5.38183860e-02 -7.62307942e-01 2.04371691e-01
-2.26830423e-01 -2.32566014e-01 3.56044881e-02 -1.66976884e-01
3.12182039e-01 -6.97074085e-02 -1.42471170e+00 6.93410873e-01
5.86162746e-01 2.80988455e-01 -4.97451246e-01 -3.24627124e-02
4.01368946e-01 -3.21310639e-01 -6.03701115e-01 6.81886375e-01
7.84275949e-01 -1.08837271e+00 5.38802922e-01 -6.12184823e-01
1.44889325e-01 -1.19873531e-01 6.32995069e-02 -1.74781501e+00
6.32685944e-02 -5.99340260e-01 -5.01798093e-01 7.28502929e-01
1.34017602e-01 -7.42784560e-01 8.02268445e-01 2.17099518e-01
1.44697919e-01 -1.22212350e+00 -1.10887325e+00 -1.02655208e+00
2.31460512e-01 2.26560496e-02 3.83202404e-01 3.68649930e-01
7.25139827e-02 -1.84217870e-01 -4.54799950e-01 -3.22429597e-01
5.83022654e-01 1.04788793e-02 7.63264775e-01 -8.11037600e-01
-5.95726073e-01 -5.46790898e-01 -1.68904886e-01 -6.67010128e-01
1.22082017e-01 -2.77360797e-01 2.85837978e-01 -1.16121602e+00
-3.11409205e-01 -4.50133532e-01 -3.13662648e-01 1.29446477e-01
-8.92778635e-02 -5.04499614e-01 2.78317213e-01 9.97504070e-02
-4.50810552e-01 1.01897323e+00 1.10321689e+00 4.55179274e-01
-8.42153311e-01 3.54104638e-01 9.15327445e-02 8.11468005e-01
1.01396286e+00 -6.63801551e-01 -5.24966598e-01 3.56494367e-01
2.16296017e-01 3.62096846e-01 1.39688507e-01 -1.29780114e+00
-7.92142302e-02 -3.35885912e-01 1.00660399e-01 -3.82788420e-01
1.91532925e-01 -4.53484416e-01 -4.91058007e-02 1.08785343e+00
-6.23710454e-01 -5.38742542e-03 2.80575395e-01 9.24510300e-01
2.84397006e-01 -4.43339974e-01 9.47462380e-01 -3.17660451e-01
-5.75292289e-01 -2.10526556e-01 -6.48337781e-01 -1.79745838e-01
1.06280863e+00 1.43707348e-02 9.04020593e-02 -4.17481959e-01
-8.39152396e-01 2.48764440e-01 4.16954607e-01 1.57787949e-01
2.95686960e-01 -7.71533012e-01 -3.32948059e-01 -6.07194640e-02
-1.29174650e-01 -4.57096756e-01 3.35960388e-02 6.21801972e-01
-4.00022358e-01 2.46365324e-01 -5.22670567e-01 -2.39978388e-01
-7.31263280e-01 5.80521703e-01 8.06943595e-01 -8.94311070e-01
-4.47820395e-01 4.35907483e-01 -4.52794820e-01 -6.04751229e-01
3.61054718e-01 -4.00119066e-01 -2.96580255e-01 -2.22525239e-01
1.91090390e-01 7.47411489e-01 -1.22246996e-01 -2.43731607e-02
-1.09726436e-01 1.73877463e-01 1.36158645e-01 -6.61313832e-01
1.15902007e+00 6.25573993e-02 3.85043770e-01 7.22606778e-01
4.05623168e-01 -1.44966781e-01 -1.85197663e+00 1.24033712e-01
1.89146191e-01 -1.59539565e-01 7.79120326e-02 -1.17665291e+00
-4.27944303e-01 4.55057114e-01 9.01387215e-01 8.68161768e-02
8.08596969e-01 -5.86629868e-01 1.98435739e-01 6.45837307e-01
7.94696689e-01 -1.45809388e+00 -9.66119487e-03 4.08017039e-01
8.71128857e-01 -9.65798616e-01 1.84514731e-01 3.46434504e-01
-8.34224045e-01 1.02253091e+00 7.44533241e-01 -6.57012045e-01
4.59859818e-01 1.84440300e-01 -2.88421631e-01 3.07241648e-01
-9.36801493e-01 -3.20972383e-01 -1.93340123e-01 4.37584847e-01
-2.34039828e-01 1.50420712e-02 -8.01937044e-01 2.55887002e-01
9.52872485e-02 4.96620715e-01 4.41230834e-01 1.25058997e+00
-8.17646086e-01 -1.21801734e+00 -2.70229101e-01 2.93112606e-01
-1.74867943e-01 4.36300874e-01 -2.03716099e-01 1.00019455e+00
-1.78088099e-02 5.74159920e-01 -1.25555128e-01 -3.26663315e-01
1.07879937e-01 2.13285387e-01 6.52494252e-01 -2.60172725e-01
-5.38136303e-01 -6.31204396e-02 1.29785001e-01 -7.64334679e-01
-1.64925560e-01 -6.05114877e-01 -1.33069861e+00 -4.00937200e-02
-2.70068407e-01 3.79557759e-01 8.92686546e-01 1.03598821e+00
4.34437782e-01 3.59858245e-01 5.82434773e-01 -1.18066657e+00
-1.31519115e+00 -9.30565417e-01 -4.70417917e-01 -3.67161185e-02
2.41032019e-01 -1.17353380e+00 -3.59599113e-01 -4.37403619e-01] | [4.29914665222168, 2.1455538272857666] |
e44ce7b9-a4b3-4716-94c4-9c6e7fdd8f40 | guided-stereo-matching-1 | 1905.10107 | null | https://arxiv.org/abs/1905.10107v1 | https://arxiv.org/pdf/1905.10107v1.pdf | Guided Stereo Matching | Stereo is a prominent technique to infer dense depth maps from images, and deep learning further pushed forward the state-of-the-art, making end-to-end architectures unrivaled when enough data is available for training. However, deep networks suffer from significant drops in accuracy when dealing with new environments. Therefore, in this paper, we introduce Guided Stereo Matching, a novel paradigm leveraging a small amount of sparse, yet reliable depth measurements retrieved from an external source enabling to ameliorate this weakness. The additional sparse cues required by our method can be obtained with any strategy (e.g., a LiDAR) and used to enhance features linked to corresponding disparity hypotheses. Our formulation is general and fully differentiable, thus enabling to exploit the additional sparse inputs in pre-trained deep stereo networks as well as for training a new instance from scratch. Extensive experiments on three standard datasets and two state-of-the-art deep architectures show that even with a small set of sparse input cues, i) the proposed paradigm enables significant improvements to pre-trained networks. Moreover, ii) training from scratch notably increases accuracy and robustness to domain shifts. Finally, iii) it is suited and effective even with traditional stereo algorithms such as SGM. | ['Davide Pallotti', 'Stefano Mattoccia', 'Matteo Poggi', 'Fabio Tosi'] | 2019-05-24 | guided-stereo-matching | http://openaccess.thecvf.com/content_CVPR_2019/html/Poggi_Guided_Stereo_Matching_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Poggi_Guided_Stereo_Matching_CVPR_2019_paper.pdf | cvpr-2019-6 | ['stereo-matching'] | ['computer-vision'] | [ 5.41427970e-01 1.00774616e-01 9.10657421e-02 -3.97592962e-01
-7.05370843e-01 -3.31062138e-01 7.05047667e-01 9.63387415e-02
-6.40218496e-01 8.54782820e-01 1.06651828e-01 1.77497208e-01
-1.32589579e-01 -8.81935656e-01 -9.87426996e-01 -6.69939399e-01
-1.60704646e-02 5.00667989e-01 5.19842923e-01 -3.30420524e-01
2.46664912e-01 5.96889794e-01 -2.05937767e+00 1.79136246e-01
9.11146462e-01 1.24761987e+00 6.11596704e-01 2.44924679e-01
-8.71284530e-02 5.49927056e-01 -1.21835716e-01 -3.23509753e-01
6.92444324e-01 1.11019500e-01 -5.55759728e-01 9.63539556e-02
9.76887107e-01 -7.15219378e-01 -6.56944692e-01 8.50847661e-01
5.66510856e-01 4.50261645e-02 2.48884678e-01 -9.16465461e-01
-2.76630878e-01 7.06998631e-02 -4.51165318e-01 -1.21442892e-01
2.84420460e-01 3.47128093e-01 8.03065658e-01 -8.82568061e-01
7.83417106e-01 9.05744433e-01 8.83079946e-01 4.03427601e-01
-1.34078538e+00 -4.94812995e-01 2.06698775e-01 2.08119869e-01
-1.10469341e+00 -6.77802861e-01 9.55398560e-01 -3.17418158e-01
9.39386785e-01 -1.83480799e-01 6.57479465e-01 1.16802788e+00
-2.27974609e-01 6.83056414e-01 1.29894066e+00 -2.49352634e-01
1.97752118e-01 7.06201792e-02 -1.58611730e-01 5.87138832e-01
1.77510351e-01 5.60753584e-01 -7.67227590e-01 2.04641119e-01
8.01929593e-01 5.15187308e-02 -3.03407580e-01 -9.09854472e-01
-1.15958369e+00 6.70326233e-01 8.51645052e-01 1.19352274e-01
-5.31088710e-01 -1.22578163e-02 2.05121249e-01 2.94412583e-01
4.12558287e-01 3.41467917e-01 -4.43225294e-01 -4.46840897e-02
-1.09953904e+00 2.65754610e-01 3.73165280e-01 8.05301785e-01
1.42676950e+00 6.65113628e-02 2.65194803e-01 7.44468212e-01
-2.54167132e-02 5.39445519e-01 3.57354254e-01 -1.16432726e+00
6.71725392e-01 5.34904540e-01 -8.50993320e-02 -9.83667254e-01
-4.11580563e-01 -5.38657963e-01 -1.00272763e+00 5.05407274e-01
5.94245732e-01 8.80053118e-02 -1.08533108e+00 1.86583161e+00
2.70986140e-01 3.32034171e-01 1.91846546e-02 8.48437726e-01
5.77726483e-01 2.43491724e-01 -4.66509163e-01 1.08938858e-01
7.74709702e-01 -7.66754389e-01 -1.02220275e-01 -7.91488707e-01
3.14919680e-01 -6.09104216e-01 8.44229817e-01 6.65649056e-01
-1.06446815e+00 -7.71265388e-01 -1.04380131e+00 -3.29920769e-01
-4.41657990e-01 -3.73820692e-01 7.59952366e-01 2.45563030e-01
-1.37510037e+00 9.33354020e-01 -7.91354120e-01 -3.58329743e-01
6.57642603e-01 6.03645205e-01 -7.08622813e-01 -6.32261574e-01
-9.70446527e-01 7.43568301e-01 3.08562487e-01 1.75693363e-01
-8.49500358e-01 -7.96705425e-01 -1.03117633e+00 -1.11267097e-01
3.58684629e-01 -9.73985791e-01 1.01373756e+00 -9.66671646e-01
-1.50493658e+00 9.69166160e-01 -2.49177039e-01 -6.12874150e-01
6.90742195e-01 -4.04781550e-01 1.11577243e-01 3.96892548e-01
3.05186898e-01 1.08599246e+00 9.48716402e-01 -1.43392718e+00
-7.85124362e-01 -6.38122439e-01 3.22412878e-01 8.03686231e-02
-3.12902957e-01 -5.28727949e-01 -4.22581971e-01 -3.25324595e-01
4.20544833e-01 -8.75080168e-01 -3.24554861e-01 2.64940500e-01
-2.03640044e-01 3.17057550e-01 6.69555187e-01 -5.55724859e-01
5.07253766e-01 -2.01758718e+00 1.97257921e-01 1.52658030e-01
1.60189644e-01 3.21117371e-01 -2.12486222e-01 4.75710094e-01
2.48964690e-02 -4.01018858e-01 -4.88732785e-01 -5.81160188e-01
-1.50498986e-01 3.80798280e-01 -1.46696210e-01 4.81904000e-01
3.34050477e-01 6.96815908e-01 -9.67868030e-01 -1.86184168e-01
6.35102928e-01 6.25495374e-01 -6.94527090e-01 2.88758904e-01
-7.18075112e-02 8.52605104e-01 -1.42878413e-01 4.95134890e-01
9.34717894e-01 -1.27994463e-01 -1.07761197e-01 -1.78630129e-01
-2.26322174e-01 3.99332702e-01 -1.26657677e+00 2.33200860e+00
-8.18176508e-01 5.97972631e-01 1.01971880e-01 -1.13991892e+00
1.01391840e+00 -7.30492398e-02 3.60427618e-01 -1.02789772e+00
-1.12605495e-02 5.75772703e-01 -1.12944983e-01 -3.57952029e-01
4.06947643e-01 -1.79073453e-01 2.21964031e-01 -8.94755423e-02
2.13419661e-01 -2.87139326e-01 1.62957177e-01 -5.95262386e-02
1.04978478e+00 3.22841495e-01 2.52489775e-01 -9.09875184e-02
6.33240342e-01 -1.66341245e-01 5.62898457e-01 6.60643220e-01
5.29160090e-02 8.39358330e-01 1.20918892e-01 -5.51137149e-01
-9.58516896e-01 -1.01593900e+00 -2.55486459e-01 6.92247689e-01
2.85164922e-01 -5.34310453e-02 -3.92804533e-01 -5.43588281e-01
2.97945172e-01 2.40623295e-01 -5.28889656e-01 3.25105302e-02
-6.91466093e-01 -2.36298829e-01 1.61752596e-01 6.54496729e-01
8.59983861e-01 -8.00206423e-01 -7.97734618e-01 2.92492092e-01
-7.06144720e-02 -1.40686715e+00 1.72828645e-01 4.10198927e-01
-1.19177783e+00 -1.04801953e+00 -8.60008657e-01 -5.96536636e-01
5.61469555e-01 6.47203803e-01 1.15003347e+00 -7.35915974e-02
-4.86918762e-02 2.86368459e-01 -1.20683901e-01 -2.18179300e-02
-1.07792281e-01 1.10530399e-01 -1.17724717e-01 1.22551816e-02
1.23460233e-01 -1.08650565e+00 -8.47849786e-01 1.53649122e-01
-8.60962212e-01 1.62365749e-01 7.65958011e-01 9.66778219e-01
6.00888133e-01 -1.92233488e-01 3.39418620e-01 -8.33932281e-01
-3.67573015e-02 -1.90243527e-01 -7.24753797e-01 -3.93175840e-01
-5.69075644e-01 2.29535311e-01 7.22361803e-01 3.07649635e-02
-1.04933512e+00 1.92943484e-01 -4.44141954e-01 -5.21780133e-01
-4.90926385e-01 3.96368414e-01 -1.61100537e-01 -3.00572395e-01
6.71021044e-01 1.74439609e-01 1.12058923e-01 -6.49307430e-01
3.12507987e-01 4.11973447e-01 8.29667509e-01 -3.68638337e-01
1.05242753e+00 1.06875896e+00 3.38235319e-01 -7.74110556e-01
-9.03459370e-01 -5.95303953e-01 -1.02309096e+00 6.11169674e-02
4.53267932e-01 -1.11768818e+00 -3.69419098e-01 7.67010272e-01
-9.99907970e-01 -4.04095024e-01 -3.42879027e-01 3.41199756e-01
-6.25772715e-01 5.75815856e-01 -3.18105459e-01 -5.40624738e-01
-8.80349055e-02 -1.01089442e+00 1.30431330e+00 3.79920900e-02
5.48966974e-02 -9.77638662e-01 1.10428808e-02 3.95497024e-01
3.96737069e-01 3.85225773e-01 4.80458260e-01 -2.07401276e-01
-9.81767416e-01 -1.82966948e-01 -4.31916326e-01 4.64718521e-01
8.86999816e-02 -2.91822881e-01 -1.47753644e+00 -4.81685400e-01
3.01362351e-02 -4.62130398e-01 1.22083652e+00 1.88629523e-01
1.08559859e+00 5.06940857e-02 -1.64408937e-01 1.04231739e+00
1.64935887e+00 -2.77513564e-01 6.97650731e-01 7.69823551e-01
7.70427763e-01 8.41385365e-01 5.59714139e-01 3.56146961e-01
4.97783691e-01 8.04483593e-01 8.87508333e-01 -1.87685996e-01
-1.70928344e-01 -3.62837791e-01 4.93810214e-02 4.45161581e-01
-6.26233146e-02 2.26649374e-01 -8.96525204e-01 6.84710026e-01
-1.81454766e+00 -7.74660647e-01 -1.22206602e-02 2.44464111e+00
6.38783276e-01 3.94253165e-01 -1.29917145e-01 3.21878850e-01
2.70357609e-01 3.64403993e-01 -7.39196181e-01 4.78887552e-04
-4.26563114e-01 5.59208333e-01 5.55082023e-01 3.94162416e-01
-1.01344717e+00 8.47957790e-01 5.15310526e+00 4.75741982e-01
-1.39635885e+00 -1.15761325e-01 3.13536137e-01 -6.29984215e-02
-3.51111561e-01 -6.37590736e-02 -6.38733506e-01 3.80171806e-01
5.83909273e-01 1.98768795e-01 4.89476442e-01 8.24102283e-01
1.56188264e-01 -3.44310850e-01 -1.32375562e+00 1.22972596e+00
5.18183596e-02 -1.37338793e+00 -6.60243854e-02 2.08956465e-01
8.50281239e-01 4.63806897e-01 1.34459347e-01 2.11734951e-01
1.49276736e-03 -8.12460065e-01 6.16745710e-01 2.13517025e-01
7.44646847e-01 -6.23931766e-01 8.44540656e-01 5.59147179e-01
-9.35579598e-01 -1.75710142e-01 -5.92508316e-01 -3.36977452e-01
1.07476018e-01 8.43424439e-01 -5.68284452e-01 8.15179348e-01
7.12382972e-01 1.04402757e+00 -6.19215190e-01 1.02161539e+00
-3.66432726e-01 -3.96089349e-03 -5.47571301e-01 6.65950835e-01
4.68227744e-01 -8.64665359e-02 3.63261729e-01 9.06107843e-01
4.80602056e-01 -4.14730608e-01 3.91689688e-02 6.94902003e-01
-9.82066542e-02 -2.13544905e-01 -1.03101516e+00 5.31145155e-01
2.74574578e-01 1.20284295e+00 -4.05446023e-01 -2.03618750e-01
-6.10334396e-01 1.04409575e+00 5.96722186e-01 3.84168863e-01
-3.89127344e-01 4.29208428e-02 8.14733863e-01 3.34835082e-01
5.68679750e-01 -3.48042101e-01 -3.73640269e-01 -1.13779950e+00
2.18188345e-01 -8.13717782e-01 1.05663724e-01 -6.83420122e-01
-1.23660374e+00 5.58085203e-01 -1.81987211e-01 -1.38967526e+00
-5.78973353e-01 -6.75940037e-01 -3.91376853e-01 9.97732043e-01
-2.45290446e+00 -1.22636008e+00 -7.45280266e-01 8.39882672e-01
4.31883961e-01 5.33094071e-03 6.53335690e-01 3.97578478e-01
-1.38827652e-01 3.38497698e-01 1.43927783e-01 -1.29193485e-01
8.23189139e-01 -1.21741986e+00 3.71602923e-01 9.60874617e-01
1.43005475e-01 4.42261159e-01 5.39256573e-01 -2.60106623e-01
-1.35841632e+00 -8.96674693e-01 6.90816283e-01 -2.24237055e-01
5.03669739e-01 -3.27481925e-01 -9.82728958e-01 3.19354802e-01
5.81251606e-02 2.35831946e-01 2.29138002e-01 1.45778447e-01
-4.10630047e-01 -4.56204116e-01 -1.02179253e+00 4.33098882e-01
1.31202197e+00 -7.42693961e-01 -5.31654954e-01 1.52931422e-01
4.92366135e-01 -6.70151472e-01 -5.29474735e-01 6.43341422e-01
5.34222245e-01 -1.74485517e+00 1.05818284e+00 -1.03748612e-01
6.70543790e-01 -1.41861841e-01 -3.09155822e-01 -1.35396469e+00
-1.67561099e-01 -5.53400695e-01 -1.21179812e-01 9.85098660e-01
7.27913082e-02 -7.54834771e-01 1.10674846e+00 4.88875598e-01
-4.13303941e-01 -7.10041165e-01 -1.10803258e+00 -8.98941219e-01
3.55654620e-02 -4.76728201e-01 5.21067142e-01 7.59599984e-01
-4.13537651e-01 2.03405857e-01 -3.63972634e-01 3.30651551e-01
8.39689672e-01 3.17014724e-01 1.13258517e+00 -1.42745113e+00
-4.92554694e-01 -2.33946756e-01 -5.99337399e-01 -1.42506242e+00
1.27049685e-01 -7.75072515e-01 3.61397155e-02 -1.53993464e+00
-1.18463404e-01 -5.31608820e-01 -2.51159996e-01 3.48495901e-01
-1.42214641e-01 4.09771621e-01 2.81201243e-01 1.65374607e-01
-1.01505980e-01 6.54454827e-01 1.16873586e+00 -9.32406783e-02
-9.51758698e-02 -6.76065385e-02 -5.42277336e-01 8.91697884e-01
6.56499624e-01 -3.00796807e-01 -4.65494633e-01 -7.38995135e-01
2.42476404e-01 -2.52227858e-02 7.19661415e-01 -1.31310189e+00
2.02637896e-01 2.28855774e-01 3.68395537e-01 -6.08410895e-01
6.58731222e-01 -1.01619947e+00 -5.79440780e-02 2.81372130e-01
2.67753843e-02 -3.39179635e-01 3.43965173e-01 6.00435197e-01
-5.03011048e-01 -1.91814139e-01 8.87357831e-01 -2.01732159e-01
-1.14601541e+00 4.11031991e-01 2.20747128e-01 9.63774621e-02
7.00232029e-01 -7.17155874e-01 -2.29542896e-01 -3.27920824e-01
-5.08197606e-01 1.44279018e-01 8.48614037e-01 1.97289869e-01
6.33433163e-01 -1.07725298e+00 -4.62170511e-01 5.78380167e-01
2.27308229e-01 5.99779487e-01 3.38010252e-01 8.77932250e-01
-4.29721385e-01 4.14117783e-01 -5.08628190e-01 -9.44063127e-01
-9.38420177e-01 4.94823784e-01 2.18611911e-01 -2.28620753e-01
-8.51322234e-01 8.90706956e-01 5.35832167e-01 -4.25670981e-01
1.65869236e-01 -4.38816637e-01 7.41536766e-02 8.52869600e-02
2.39414960e-01 1.18493743e-01 3.92683268e-01 -4.80864912e-01
-3.27548981e-01 8.28957021e-01 3.32071148e-02 -5.66736460e-02
1.55489671e+00 -2.34008849e-01 1.87965468e-01 2.99722075e-01
1.30963647e+00 -2.56131999e-02 -1.80518770e+00 -5.22536933e-01
-1.10213697e-01 -9.00629044e-01 2.57708460e-01 -5.55431664e-01
-1.20866704e+00 1.24602318e+00 5.33454657e-01 -1.19181626e-01
1.23259592e+00 -3.48984122e-01 9.54483509e-01 4.56370652e-01
7.90240884e-01 -8.85401905e-01 2.21993448e-03 5.14898300e-01
7.08097994e-01 -1.61798728e+00 -1.16934165e-01 -3.35871577e-01
-2.48471871e-01 1.15301049e+00 4.85116839e-01 -1.81792393e-01
5.94619215e-01 1.50646016e-01 4.85528335e-02 -3.37387025e-02
-4.16905701e-01 -5.94440937e-01 6.94678202e-02 8.84686172e-01
1.42257288e-01 -4.75254834e-01 2.78852880e-01 -3.30037214e-02
-1.06346466e-01 1.88163027e-01 3.47827852e-01 9.38816190e-01
-4.42455679e-01 -1.14176893e+00 -1.54682979e-01 3.22223097e-01
-8.14077705e-02 -2.54914314e-01 -1.87421277e-01 9.70536292e-01
2.99436897e-01 7.30276585e-01 5.65892719e-02 -2.16496989e-01
5.31318724e-01 -2.25667760e-01 5.37531376e-01 -6.56735957e-01
-3.40525746e-01 -1.14581041e-01 4.68193926e-02 -9.37628031e-01
-6.65076315e-01 -7.35455930e-01 -9.26578104e-01 -2.21757859e-01
-3.00044734e-02 -4.06920582e-01 5.40222466e-01 9.31167185e-01
4.07483011e-01 3.53042543e-01 6.96309328e-01 -1.40289700e+00
-4.20037150e-01 -6.92413568e-01 -4.47174430e-01 5.35462379e-01
5.74266374e-01 -8.76722932e-01 -4.36999708e-01 -9.98375416e-02] | [8.767915725708008, -2.3717997074127197] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.